AN ELECTROPHYSIOLOGICAL ANALYSIS OF CHRONIC ETHANOL EFFECTS
ON SYNAPTIC DISTRIBUTION AND FUNCTION IN RAT HIPPOCAMPUS
BY
WICKLIFFE C. ABRAHAM
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA 1981
This dissertation is dedicated to
Stuart B. and Ida Jeanne Dagger Abraham.
Young man, take care lest you find what you are looking for.
Granit
ACKNOWLEDGEMENTS
Once in a long while, a friendship is formed that interweaves
personalities in a unique way. Many thanks to Jim for keeping me out of school my first year and keeping me in school for the remaining years.
Friendships are multifaceted but in part serve to resolve the past, enliven the present and give hope to the future. Those instrumental in enlivening my graduate years include Brian, Denise, Dick, Lia, Lynn, Mary Margaret and FVC.
The love and support from Elise has been sustaining and instructive.
My supervisory committee--Chuck Vierck, Don Walker, Floyd Thompson and Keith Berg--has been thoughtful, helpful and supportive. Two in particular stand out as guiding lights and resourceful friends--Steve Zornetzer and Bruce Hunter. Thank you, Steve and Bruce.
Paul Manis has made crucial contributions to the data analysis in these experiments. Thanks also to Bill Brownell for use of his computer facilities. My research has relied greatly on the technical support of Dot Robinson, Larry Ezell, Pat Burnett and NERDC. I have been supported financially by an NSF predoctoral fellowship and an NIAAA predoctoral fellowship.
iv
TABLE OF CONTENTS
PAGE
ACKNOWLEDGEMENTS . . iv
INDEX OF ABBREVIATIONS vii
ABSTRACT viii
CHAPTER I. BACKGROUND 1
Alcohol-related Amnesia: Clinical Studies . . . . 2 Hippocampal Dysfunction and Amnesia: Clinical Studies. 5
Alcohol-related vs. Hippocampus-related Amnesia:
Laboratory Studies . . . . . . . . 9
Hippocampal Anatomy and Physiology. . . . . . 13
Animal Models of Chronic Alcohol Consumption. . . . 25 Rationale 26
CHAPTER II. GENERAL METHODS . . . . . . . . . 28
Ethanol Administration . . . . . . . . . 28
Electrophysiological Methods. . . . . . . . 29
Experimental Protocol 30
Data Analysis 32
CHAPTER III. CHRONIC ETHANOL EXPOSURE AND SYNAPTIC DISTRIBUTION
IN CA1 OF RAT HIPPOCAMPUS: CURRENT-SOURCE DENSITY
ANALYSIS 35
Introduction 35
Methods 36
Results . 37
Discussion 50
CHAPTER IV. AUGMENTATION OF SHORT-TERM PLASTICITY IN CA1 OF
RAT HIPPOCAMPUS AFTER CHRONIC ETHANOL TREATMENT. .... 59
Introduction 59
Methods 60
Results 62
Discussion 77
CHAPTER V. ELECTROPHYSIOLOGICAL ANALYSIS OF CHRONIC ETHANOL
NEUROTOXICITY IN THE DENTATE GYRUS: ENTORHINAL
AFFERENTS 83
v
PAGE
Introduction . . . . . . . . . . . 83
Methods . . . . . . . . . . . . . 85
Results 86
Discussion. . . . . . . . . . . . . 113
CHAPTER VI. GENERAL DISCUSSION . . . . . . . . 121
APPENDIX. CURRENT-SOURCE DENSITY ANALYSIS . . . . . 129
REFERENCES . . . . . . . . . . . . . 132
BIOGRAPHICAL SKETCH . . . . . . . . . . . 145
vi
INDEX OF ABBREVIATIONS
AB--angular bundle I/O--input/output AH--Ammon's horn IPI--interpulse interval ALV--alveus KF--Korsakoff ANOVA--analysis of variance LTM--long-term memory CAl-4--hippocampal subfields LTP--long-term potentiation CNS--central nervous system PS--population spike COM--commissural PPP--paired-pulse potentiation CSD--current-source density PTP--posttetanic potentiation DG--dentate gyrus SCH--Schaffer collaterals DRL--differential reinforcement SG--stratum granulosum
of low rate responding
SM--stratum moleculare
EP--evoked potential
SO--stratum oriens
EPSP--excitatory postsynaptic
potential SP--stratum pyramidale FP--frequency potentiation SR--stratum radiatum GABA--y-aminobutyric acid STM--short-term memory HF--hippocampal fissure TBR--to-be-remembered HIL--dentate hilus
vii
Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
AN ELECTROPHYSIOLOGICAL ANALYSIS OF CHRONIC ETHANOL EFFECTS
ON SYNAPTIC DISTRIBUTION AND FUNCTION IN RAT HIPPOCAMPUS
By
Wickliffe C. Abraham
June, 1981
Chairman: Steven F. Zornetzer
Major Department: Neuroscience
Chronic alcoholism in man results in a syndrome marked by a variety of neuropsychological and neurophysiological abnormalities. In advanced stages, the central impairment is a severe anterograde amnesia. The hippocampus, long thought to be involved in memory consolidation, could be a major target of alcohol neurotoxicity. Recent studies using a rodent model of chronic ethanol consumption have described not only a number of cognitive impairments but also neuropathology in the hippocampus, including cell loss and decreased spine density. The present experiments were conducted to provide electrophysiological correlates of the morphological data.
Rats consumed an ethanol-containing liquid diet for 20 weeks. Controls were either pair-fed a similar diet with sucrose isocalorically substituted for ethanol or received standard laboratory chow. Eight weeks after ethanol withdrawal, the rats were prepared for electrophysiological field potential recordings in the hippocampus using standard extracellular recording techniques.
viii
The first two experiments examined the synaptic distribution and function of CA3 afferents to CAl. Evoked potential profiles, orthogonal to the layering of afferent terminals, were obtained and analyzed with current-source density techniques. The CA3 afferents to stratum radiatum of CAI exhibited a bimodal distribution of current density, possibly reflecting a separation of the ipsilateral and contralateral CA3 terminals. Ethanol treatment produced an overall shrinkage of the synaptic zone independent of tissue volume changes. The shrinkage was specific to the hypothesized commissural input, proximal to the cell layer. The hypothesized ipsilateral fibers showed evidence indicative of a compensatory enhancement of synaptic efficacy.
The CA3-CAl synapses exhibited normal input/output relationships for all measures in ethanol-treated rats. Furthermore, ethanol had no effects on long-term potentiation. However, ethanol treatment did lead to an augmentation of short-term plasticity. The population spike was differentially enhanced in both dual-pulse and repetitive stimulation paradigms. These findings may be due to an ethanol-related reduction in the efficacy of recurrent inhibition.
The entorhinal cortex-dentate gyrus synapses were similarly studied. Laminar analyses through the dentate dorsal and ventral blades again revealed in ethanol-treated animals a shrinkage of the synaptic fields. In both blades, the shrinkage was restricted to the outer molecular layer. These results are suggestive of preferential damage to the distal terminals and/or dendrites. In contrast to the CAI findings, shortterm plasticity was refractory to ethanol treatment. However, input/ output analysis revealed that ethanol-treated rats had smaller population spike amplitudes per synaptic potential at asymptotic stimulus currents.
ix
These data probably reflect the loss of granule cells that has been documented previously. Finally, of the two known components of longterm potentiation, the ethanol regimen apparently compromised only the potentiation of neural excitability.
The experiments presented here have electrophysiologically confirmed the loss of afferent input to CA1 subsequent to chronic ethanol treatment. The loss is apparently confined to an as yet unidentified subset of the CA3 afferents. It is evident that the afferent drive to the dentate is also reduced, particularly in distal dendritic regions. Confirmatory evidence for the loss of granule cells is presented. The remaining data indicate that chronic ethanol consumption can significantly alter short-term plasticity in CA1 and long-term potentiation in the dentate. These alterations may underlie some of the ethanol-related behavioral impairments. The electrophysiological data also compare favorably with similar findings in aged animals, suggesting that there may be a common mode of action.
x
CHAPTER I
BACKGROUND
Long-term consumption of alcohol leads to a variety of neuropsychological impairments in both humans and laboratory animals. One of the more profound and debilitating aspects of the chronic alcohol syndrome is the impairment of memory processing capabilities. Converging lines of evidence suggest that damage to the hippocampus, a central nervous system (CNS) telencephalic structure, may underlie the decline in mnemonic functioning. In spite of some studies directly examining the histology of the hippocampus (see below), much of the inference linking alcohol consumption to hippocampal damage is based on the similarity of the behavioral alterations resulting from chronic alcohol exposure and those produced by surgically-induced hippocampal damage. Clearly, such an inferential approach has many shortcomings. Most importantly, brain structure-function relationships are not so well understood that similarities (or dissimilarities) of behavioral deficits necessarily imply similarities (or dissimilarities) of underlying brain damage. Nonetheless, clues are needed which point to where in the CNS alcohol may be exerting its toxic effects. A correspondence of the behavioral symptoms of the chronic alcohol and hippocampal syndromes would at least give us sufficient reason to suspect that the hippocampus is damaged by chronic alcohol ingestion.
An important consideration in the study of alcohol-related dementia is the extent to which alcohol serves as the primary toxic agent.
1
2
Nutritional or metabolic disturbances secondary to chronic ethanol intoxication are now considered to underlie several dementing diseases formerly thought to be due primarily to alcohol toxicity. These include Wernicke syndrome, pellagra, Marchiafava-Bignami disease, hepatic encephalopathy, and central pontine myelinosis (Victor and Banker, 1978). Although some authors also maintain that Korsakoff syndrome has a nutritional etiology, i.e., thiamin deficiency (Victor and Banker, 1978; Victor et al., 1971), the issue is not yet resolved. No cases of persistent Korsakoff syndrome have yet been documented to be induced by malnutrition in the absence of concomitant ethanol consumption (Freund, 1973). Furthermore, animal models have shown that profound memory disturbances persist after chronic alcohol consumption in spite of adequate nutrition (Freund, 1970; Walker and Freund, 1973; Walker and Hunter, 1978). Thus, the evidence favors the view that alcohol consumption per se can directly produce dementia.
Alcohol-related Amnesia: Clinical Studies
Korsakoff (KF) syndrome is a "pure" amnesia in that memory functions are profoundly disturbed while other cognitive functions of the patients remain relatively intact (Victor et al., 1971). The symptomatology includes a limited retrograde amnesia, profound anterograde amnesia, apathy, indifference to environmental events and limited insight into the disability. Confabulation is occasionally observed but is probably secondary to the loss of memory for recent events. The deficits are relatively permanent, even after alcohol withdrawal (Brion, 1969; Victor et al., 1971). Although the term KF psychosis or syndrome has been used to categorize amnesias of a variety of etiologies (Brion,
3
1969), we will reserve this term exclusively for alcohol-related amnesias.
As previously mentioned, the salient presenting sign of KF syndrome is a severe anterograde amnesia. The exact nature of this deficit has been the subject of considerable controversy. Research in this area has focused on 1) whether short-term memory is affected and 2) whether the long-term memory deficits are more specific to the storage or the retrieval of information. The major protagonists involved in the latter issue are Butters and Cermak's group in Boston that supports the encoding deficit hypothesis and Warrington and Weiskrantz's group in Oxford that supports the retrieval deficit hypothesis.
In many respects, short-term memory (STM) is relatively unimpaired in KF patients. In particular, the initial registration of information into STM appears intact since KF patients show normal digit spans, normal recency effects in verbal free recall and normal memory for visual location (Brooks and Baddeley, 1976; Kinsbourne and Wood, 1975).
On the other hand, there is considerable evidence that maintenance or retrieval of information from STM may be impaired. When evaluated on the Peterson-Peterson paradigm, KF patients' STM was observed to be acutely susceptible to the disruptive influence of interpolated tasks (performed between stimulus presentation and recall) (Cermak et al., 1971). This retroactive interference is modality-independent but more pronounced when the modality is the same for both the distractor and the to-be-remembered (TBR) items (DeLuca et al., 1975; DeLuca et al., 1976; Strauss arid Butler, 1978). The STM of KF patients has also been shown to be extremely susceptible to proactive interference, which becomes evident as these STM tests are repeated (Cermak et al., 1976; Parkinson, 1979).
4
The next class of experiments were generated by the debate concerning storage vs. retrieval hypotheses of KF-related long-term memory (LTM) deficits. The conventional mid-1960s view of KF-related anterograde amnesia was that the amnesics were unable to store new experiences into LTM. Retrieval processes per se appeared intact since the patients showed only a limited retrograde amnesia. The Oxford group has forced a revision of this early notion. These researchers have been able to demonstrate successful memory consolidation in KF patients, particularly when employing perceptuomotor tasks (Brooks and Baddeley, 1976; Warrington and Weiskrantz, 1970). The Oxford researchers instead focus on enhanced interference at the time of recall as the major element of amnesia. Winocurhas suggested that the inability to distinguish multiple word lists is related to an impaired time-tagging or discrimination of contextual cues (Winocur and Kinsbourne, 1978; Winocur and Weiskrantz, 1976).
On the other side of the storage-retrieval controversy is the
Boston group which concentrates on encoding deficits in the KF syndrome. The emphasis of their work focuses on the failure of patients to spontaneously semantically process TBR items. Much of the support for this thesis comes from analysis of KF patients' error patterns; i.e., KF patients are not afflicted by interference in the semantic dimension, presumably due to their failure to process information semantically in the first place (Cermak and Moreines, 1976; Cermak et al., 1973). Forced use of semantic processing can improve the patients' performance to a limited extent (Cermak and Reale, 1978). Additionally, cues given at the time of recall to facilitate KF patients' performance are notoriously ineffective. Only when the cues are also available at the time
5
of encoding does the whole cuing procedure give positive results (Cermak et al., 1980).
Given the above findings, the storage vs. retrieval controversy of KF amnesia may no longer be an issue. The fact that learning and memory can be demonstrated in a variety of situations indicates that KF patients can consolidate if interference does not act so rapidly as to eliminate the TBR items from STM. The amnesia seems more precisely associated with an encoding difficulty. This difficulty will be manifested at the time of retrieval when previously learned material interferes with extraction of desired items from the memory banks. Note that retrieval per se is operative since KF patients make errors of commission, not errors of omission. Thus, both storage and retrieval are still functional but the "mnemons" are no longer distinctive. As Winocur suggests, this may reflect inadequate evaluation of context or perhaps an inability to distinguish items in the temporal dimension.
Hippocampal Dysfunction and Amnesia: Clinical Studies
Let us now consider the human hippocampal amnesic syndrome so that we may later compare this syndrome with KF psychosis. The widespread interest in the hippocampus and memory processing originated with the medial temporal lobe resections in man performed by Scoville and his colleagues (Penfield and Milner, 1958; Scoville, 1954; Scoville and Milner, 1956). On the surface, the amnesia resulting from these surgical procedures was very similar to that seen in KF psychosis, i.e., a profound anterograde amnesia coupled with a relatively intact STM.
One patient in particular (H.M.) has been the focus of attention and his memory disorder has been well characterized. We will discuss
6
the case of H.M. in detail since this case is considered to represent the classic amnesic syndrome and serves as a standard against which other clinical syndromes as well as animal studies are compared.
In 1953, H.M. underwent surgery to alleviate his debilitating
epileptic attacks. His resection was one of the most radical performed by Scoville, extending 8 cm posterior to the temporal pole and involving the anterior 2/3 of the hippocampus bilaterally. Immediately following recovery from surgery, H.M. exhibited a severe anterograde amnesia although no other neurological deficits were evident. Rigorous analyses confirmed many of the early anecdotal observations concerning H.M.'s memory disorder. Such cognitive functions as concept formation and instruction-following were intact as were his dealings with spatial relationships and spatial orientation. His STM was within normal range when tested by forward digit span, backward digit span and Seashore Tonal Memory tests (Milner et al., 1968). His STM decay curve was not significantly different from that of normal subjects (Wickelgren, 1968).
On the other hand, H.M. exhibited dramatically poor LTM for both verbal and nonverbal tasks. These tasks included supraspan digit sequences, standard verbal learning tasks, delayed matching-to-sample, delayed recall of a complex geometric design and recognition of recurrent nonsense patterns (reviewed in Milner et al., 1968). The patient could learn visually-guided mazes only if the number of choice points was kept close to STM span length. He was severely retarded in acquiring a tactually-guided maze even when the number of choice points was within short-term span.
In a manner analogous to KF patients, H.M. could show significant or normal savings on certain tasks. His acquisition and retention of
7
motor skills was essentially normal (Corkin, 1968). Although he showed retarded learning of very simple mazes, he retained the information reasonably well once learned (Milner et al., 1968). In addition, H.M. demonstrated normal acquisition of an incomplete-figure task and significant savings (although less than controls) on an unexpected retest either one hour or four months later (Milner, 1970). It should be noted that cuing at the time of testing was never shown to facilitate H.M.'s performance on any task except a face recognition task. Additionally, his remote memory was excellent. Thus, there is little evidence for a retrieval failure underlying H.M.'s amnesia.
How does KF amnesia compare to medial temporal lobe amnesia? The similarities are striking. Profound anterograde amnesia is seen in both verbal and nonverbal domains. Remote memories are essentially intact with any differences explainable by the differences in the time course of lesion development. On the other hand, recent memories are extremely susceptible to the disruptive influences of interpolated tasks. Both kinds of amnesics show good learning on perceptuomotor or partial information tasks. Some differences in symptomatology are also evident. Hippocampal amnesics have normal STM decay constants while KF patients showenhanced forgetting under distractor conditions. Korsakoff amnesics show considerable interference effects when retrieving from LTM but this phenomenon has not been demonstrated for patients such as H.M. Unlike KF patients, temporal lobe patients are not aided by cuing procedures during recall. Thus, although these two syndromes do not mirror one another exactly, overall there are enough similarities between them to suspect hippocampal damage in the KF patients. The more radical but localized damage in medial temporal lobe patients may be responsible for what appears to be a more pure amnesic syndrome.
8
We might expect that histopathological examination of autopsied brain tissue would clarify the extent to which the hippocampus is involved in KF syndrome. Unfortunately, the major published work on this subject does little to clarify the issue (Victor et al., 1971). Wernicke-Korsakoff patients showed discontinuous lesions of specific nuclear groups. The damage was generally bilateral and concentrated in paraventricular nuclei. The hippocampus itself was found to be damaged in only 36% of the cases. However, a major hippocampal efferent target, the medial mammillary nucleus, was involved in nearly all cases. Despite these findings, Victor et al. (1971) suggested that lesions of the dorsomedial nucleus of the thalamus were best correlated with the amnesia. It has recently been suggested that the amnesia may occur only when both the dorsomedial nucleus and the mammillary bodies are damaged (Mair et al., 1979).
It seems that the above pathology data provide little support for the hippocampal hypothesis of KF amnesia. Two points are relevant here. First, many cellular changes could have gone undetected by exclusive use of light microscopy. Dendritic and synaptic alterations as well as biochemical changes are obvious possibilities. Secondly, significant destruction of major inputs and outputs to a structure may produce behavioral changes similar to those produced by lesion of the structure alone. The fact that the effects of necrosis of both the medial mammillary body and the dorsomedial nucleus produce a syndrome comparable to hippocampal lesions alone suggests that such a possibility is quite reasonable. Unfortunately, Victor et al. (1971) did not give a detailed account of the nature of extent of the damage to cerebral cortical areas more closely afferent to the hippocampus such as cingulate, entorhinal
9
or subicular cortex. We conclude that a resolution of the question of hippocampal involvement in alcohol-related amnesias requires more detailed experimental analyses which are aimed directly at the issue.
Alcohol-related vs. Hippocampus-related Amnesia: Laboratory Studies
The need to clarify the role of the hippocampus in amnesia has led to a massive attempt to create animal models of the human amnesic syndrome. This section will briefly outline and compare the behavioral effects of hippocampal damage and chronic alcohol consumption in laboratory animals.
Despite extensive research efforts, no satisfactory model of the human hippocampal amnesic syndrome has yet been achieved. Lesions of the hippocampus produced specific behavioral alterations but they were not found in all tasks and they were not always performance deficits. A complete review of this voluminous literature would not be appropriate here. There already exist several exhaustive reviews (Douglas, 1967; Horel, 1978; Jarrard, 1973; Kimble, 1968; O'Keefe and Nadel, 1978). To summarize the findings, the major behavioral changes that follow hippocampal damage include: 1) chance or worse levels of spontaneous alternation, 2) decreased habituation to novel stimuli, 3) poor performance in discrimination reversal tasks, 4) prolonged extinction of discrimination responses, 5) impaired acquisition of mazes based on the use of spatial cues and 6) impaired passive avoidance but facilitated shuttle avoidance performance (Black et al., 1977; Douglas, 1967; Kimble, 1968; Kimble, 1975). Conspicuously absent from this list is a multimodal deficit in memory consolidation. Instead, the hippocampal syndrome in animals is characterized by a marked performance deficit in situations where
10
environmental contingencies require the cessation of learned or prepotent behavior and/or the initiation of a different behavior (or strategy).
There are two classes of data that are particularly relevant to the human amnesic syndrome and thus deserve special attention. The first is relevant to the suggestion that hippocampal-lesioned animals are particularly deficient in context-dependent retrieval (or encoding) of memories. The first indication of this deficit came from the demonstration that the impairments in both passive avoidance and extinction performance commonly associated with hippocampal damage were alleviated if external goal-box stimuli associated with the change in reinforcement contingencies were detectable by the animal early in the runway (Winocur and Bindra, 1976). In a manner similar to human amnesics, hippocampal rats could learn a visual discrimination task as well as controls, but were significantly impaired if a high interference task involving similar stimulus materials was interposed between training and testing (Winocur, 1979). Finally, Winocur has demonstrated that reversal-learning deficits can be alleviated when contextual cues dissociate the different reward contingencies (Winocur and Olds, 1978). These and other data (Hirsch et al., 1978) not only support a context hypothesis of hippocampal functions but also nicely interface with data from human amnesics (see above review of KF amnesia).
The other set of data relevant to the human literature arises from tasks employing a delay between stimulus and response, such as delayed response, delayed alternation, go-no-go discrimination and differential reinforcement of low rate responding (DRL) tasks (Isseroff, 1979; Iversen, 1976; O'Keefe and Nadel, 1978). Interestingly, in the situations reasonably devoid of spatial information (go-no-go temporal
II
alternations and DRL), hippocampal animals show increasing deficits with increasing times between trials (Walker and Means, 1973; Walker et al., 1972). Furthermore, a lesioned animal's performance is severely impaired at short intervals if interfering tasks are interposed (Walker and Means, 1973). These data show a remarkable similarity to those seen in the human amnesic literature.
Animal models of chronic alcohol consumption have also been developed. The pattern of behavioral deficits associated with these models is highly reminiscent of the performance of hippocampal-lesioned animals. For example, a nurber of studies reported that chronic alcohol ingestion causes impairments in avoidance acquisition long after ethanol withdrawal. Mice placed on a special, ethanol-containing liquid diet for four months were deficient in acquisition and retention of an inhibitory avoidance task (Freund, 1974). Memory for passive avoidance training could also be disrupted by posttraining ethanol consumption. One-way avoidance by hamsters was unaffected by consumption of ethanol in aqueous solution (Harris et al., 1979b).
Ethanol-consuming rats are deficient when performing several other tasks including ones appetitively motivated. Six months of ethanol consumption by rats led to deficits in acquisition of Hebb-Williams closed field mazes and in acquisition and retention of an aversively motivated moving belt task (Fehr et al., 1976). Timing behavior, DRL and temporal shock discrimination were also disrupted by chronic ethanol ingestion (Denoble and Begleiter, 1979; Smith et al., 1979; Walker and Freund, 1973). Preliminary experiments in our laboratory indicate that alcoholic rats are deficient in spontaneous alternation. Finally, temporal single-alternation, go-no-go behavior of alcoholic rats is
12
almost identical to that of rats with hippocampal damage. These rats are impaired by either long intertrial intervals or short intertrial intervals filled by an interfering task.
Chronic ethanol consumption by rats or mice produces one deficit not produced by hippocampal lesions--impaired shuttle avoidance. Hippocampal lesions normally facilitate this behavior. Nonetheless, this task has been employed for rigorous analyses of ethanol-related performance deficits. Time gradients of consumption indicate that 3, 5, 7 and 9 months but not 1.5 months of ethanol consumption produced the behavioral deficits. These deficits persisted as long as has been tested postwithdrawal, up to 4.5 months (Freund, 1979; Freund and Walker, 1971b). In add:: ion, chronic phenobarbital consumption does not produce similar behavioral effects, thus ruling out confounding sedation-related variables such as reduced sensory input or behavioral activity (Freund, 1974).
There is now good histopathological evidence that both the hippocampus and cerebellum are severely affected by chronic ethanol consumption. Five months exposure to the ethanol-containing liquid diet developed by Freund and Walker led to dramatic changes in mouse hippocampal neuronal morphology as observed with Golgi stains. The CA1 pyramidal cell basilar dendrites were severely attenuated while the apical shafts were fairly well preserved except for the most distal arborization. Dentate granule cells showed relatively good preservation of general neuronal morphology. Both subfields exhibited severe decreases in spine density (Riley and Walker, 1978). Subsequent work in rats has revealed a 15-20% loss of CA1 and CA3 pyramidal cells, dentate granule cells and cerebellar granule cells (Walker et al., 1980).
13
In summary, animal models of chronic ethanol consumption have provided data relevant to two important issues. First, the direct neurotoxic effects of ethanol have been confirmed since the models generally employ nutritionally adequate diets. Secondly, there is now more conclusive neuropathological evidence that the hippocampus is adversely affected by chronic ethanol consumption. The success of the models on these two points encourages further research aimed at determining more directly how hippocampal functioning is altered by chronic ethanol exposure.
Hippocampal Anatomy and Physiology
The hippocampus is a large, 3-layered allocortical structure sandwiched (in rat) between the neocortex and thalamus/midbrain. Rostrally, it lies just posterior to the septal nuclei. It arches first posteriorly, then ventrolaterally and then again rostrally until it rests temporally just behind the amygdala. Thus, we can assign septal and temporal poles to this structure. The course of the hippocampus between the two poles is termed the longitudinal axis.
The hippocampus is divided into two subregions, Ammon's horn and the dentate gyrus. In cross-section, these two regions can be viewed as two interlocking Cs. The dorsal or lateral arm of Ammon's horn (Ramon y Cajal's regio superior, 1911) adjoins the periallocortial subiculum while the ventral or medial arm (Ramon y Cajal's regio inferior) fits into the concavity (or hilus) of the dentate gyrus. Regio superior is largely segregated from regio inferior and the dentate gyrus by the hippocampal fissure. The dentate gyrus is also subdivided into two segments--the inner blade (facing the fissure) and the outer blade (facing the outside of the brain).
14
The dentate gyrus (DG) is trilayered: a cell-poor molecular layer, the granule cell layer and the polymorph layer. The tightly packed granule cells send their highly ramified, spiny dendrites out across the whole molecular layer while their axons perforate the polymorph layer and hilus to innervate Ammon's horn (AH) (Ramon y Cajal, 1911). All other cell types project only within the DG. The polymorph layer contains most of these short axon cells, including basket cells. Both their dendrites and axons extend across the granule cell layer into the molecular layer (Lorente de No, 1934). The remainder of the hilar region represents a transition from the polymorph cells to the pyramidal cells of AH. Recent studies have shown that these deep hilar cells project only to the ipsilateral and contralateral dentate molecular layer (Amaral, 1978; Fricke and Cowan, 1978; Swanson et al., 1978).
Ammon's horn is also trilayered; a relatively thin core of packed projection (pyramidal) cells is flanked on either side by cell-poor molecular layers. However, the organization of AH is sufficiently complex to warrant subdividing this region into several more laminae (Fig. 1-1). Starting with the deep layers and moving superficially toward the fissure are: 1) alveus, myelinated afferent and efferent fibers,2) stratum oriens, containing axons running horizontally and at right angles to the basilar dendrites of the pyramidal cells,3) stratum pyramidale,4) stratum lucidum,5) stratum radiatum, and 6) stratum lacunosum-moleculare. The last three laminae are all invaded by pyramidal cell apical dendrites and are distinguished by the afferents which terminate specifically in each layer. Finally, the pyramidal cell layer has been divided into a four part scheme (CAl-4) based on cell morphology and afferent connections which we will describe later (Lorente de No, 1934).
Figure 1-1. A schematic hippocampal lamella diagramming the major excitatory trisynaptic
circuit. Stippled areas indicate the pyramidal and granule cell layers. The heavy
black lines indicate the two major afferent pathways that were studied in the present
experiments. The electrodes placements for the CAl experiments are shown. The recording electrode site was just dorsal to the pyramidal cell layer while the stimulation electrode position was in stratum radiatum near the CAl-CA2 border. Abbreviations: COM, commissural fibers; hi, hilus, MF, mossy fibers; PP, perforant path; SCH, Schaffer
collaterals; sg, stratum granulosum; sm, stratum moleculare; so, stratum oriens; sp.
stratum pyramidale; sr, stratum radiatum.
CAl
st$
DG CA3
17
The flow of information in the hippocampus is marked by three
salient features. The first is the one-way flow of information from entorhinal cortex to DG, CA3, CA1 and then subiculum. Each of these afferent fibers involved provides a powerful excitatory synaptic drive onto the next region (Andersen et al., 1966a). The second feature is the lamination of the synaptic fields. Each afferent pathway involved in this flow of information terminates on specific portions of the postsynaptic dendrites. The last organizational feature is the remarkable lamellar orientation of these connections. Slices, 300-500 Pm thick and transverse to the longitudinal axis, will contain many of the major fiber tracts and connections intact.
The intrinsic synaptic organization of the hippocampus is reasonably well known. Axons from layer two of entorhinal cortex (medial and lateral) course in the angular bundle (perforant path) and enter the hippocampus posteriorly. These fibers terminate in the outer 2/3 of the molecular layer of the DG and distribute topographically onto granule cell dendrites (Hjorth-Simonsen and Jeune, 1972; McNaughton and Barnes, 1977; Steward and Scoville, 1976). The proximal 1/4 of granule cell dendrites is innervated by the ipsilateral and contralateral deep hilar neurons (Fricke and Cowan, 1978; West et al., 1979).
The granule cell axons form mossy fibers and exit through the hilus. Collaterals of these fibers innervate the polymorphic and hilar cells while the main axons stream into CA3 and terminate in stratum lucidum on the most proximal portions of the pyramidal cell apical dendrites (Lorente de No, 1934; Swanson et al., 1978). The mossy fibers are excitatory and do not extend past CA3.
18
The pyramidal cells of CA3 have the most diverse projections of any hippocampal neurons. The longitudinal association pathway courses through the ipsilateral stratum radiatum of CA3 and CA2, innervating CA3 pyramidal cells throughout the longitudinal axis (Raisman et al., 1965; Swanson et al., 1978). The second major projection is bilateral via fimbria and fornix to the lateral septal nucleus (DeFrance et al., 1973; Meibach and Seigel, 1977). Branches of these axons, known as Schaffer collaterals, perforate the pyramidal cell layer and terminate most densely in stratum radiatum of CAl. A less dense innervation of CA1 stratum oriens has also been observed (Hjorth-Simonsen, 1973; Lorente de No, 1934; Swanson et al., 1978). Commissural fibers cross in the ventral psalterium and innervate CAl-CA3 contralaterally. These fibers overlap the ipsilateral Schaffer collateral projection by terminating in both stratum oriens and radiatum (Gottlieb and Cowan, 1973). A branching axon from a single CA3 neuron may contribute to the septal, commissural and the Schaffer collateral projections (Swanson et al., 1980).
Axons arise from the basal portion of the CA1 pyramidal cells, descend into the alveus and bifurcate. The rostrally directed axons exit via the fimbria and fornix and terminate ipsilaterally in the lateral septal nucleus (Meiback and Seigel, 1977; Swanson and Cowan, 1977). The caudally directed axons terminate in the subiculum, parasubiculum and entorhinal cortex (Finch and Babb, 1980; Lorente de No, 1934). No commissural pathways originate from CA1 (Blackstad, 1956; Gottlieb and Cowan, 1973; Laurberg, 1979).
Besides the entorhinal cortex, there are three major sources of
extrahippocampal afferents: medial septal nucleus and vertical limb of
19
the diagonal band, scattered diencephalic nuclei and the brainstem monoamine nuclei. These afferents exert a complex mixture of excitatory and inhibitory influences on all regions of the hippocampus. However, since they have little bearing on the current experiments, they will not be discussed further.
Although the hippocampus is highly organized and structured anatomically, its synaptic efficacy is extremely labile. There are several lines of research investigating these special plastic properties of hippocampal synapses. The research relevant to this work is concerned with how the efficacy of selected hippocampal synaptic connections is modified by prior synaptic activity.
One striking aspect of research in this area is the almost exclusive use of extracellular field potential analyses of synaptic activity. These macropotential-based analyses provide sensitive and informative data concerning the status of specific connections within the hippocampus. This is possible because the highly stratified inputs and the homogeneous arrangement of the postsynaptic cellular elements produce very structured but simple source-sink relationships. Accordingly, electrical stimulation of a single afferent pathway will activate synaptically only a restricted dendritic region. If the input is excitatory, a "sink" will be generated by current flowing into the cells. To complete the circuit, current will flow out of the cell at the "source" (neighboring cellular regions) and through the extracellular space back to the site of the sink. The flow of current through the resistive extracellular fluid generates a voltage which is easily detected by a nearby, extracellularly-placed microelectrode. At the sink, the intracellular positive potential will be recorded as a
20
negative voltage extracellularly. In a similar manner, the electrode will record a positive voltage at the source. The size, packing density and homogeneous orientation of dendrites and somata are such that quite large evoked potentials can be recorded, i.e., in the millivolt range (Andersen and Lomo, 1970; Teyler et al., 1977). Separating the source and sink regions is an isopotential zone commonly referred to as the inversion point. Lomo (1971a) showed that extracellularly recorded potentials had the same onset as intracellularly recorded excitatory postsynaptic potentials (EPSPs). Thus, he termed this extracellular potential the population EPSP (hereafter referred to simply as the EPSP).
Two other extracellular potentials are commonly recorded in the
hippocampus. The first is the compound action potential representative of synchronous activity of the afferent fibers. This di- or triphasic event, termed the presynaptic volley, is small relative to the synaptic potentials and can usually be recorded only in the layer containing the afferent fibers (Andersen et al., 1978; Lomo, 1971a). At high stimulation intensities, a negative deflection can be observed superimposed on the extracellular EPSP recorded in the cell layer (Fig. 1-2). Lomo (1971a) demonstrated that this sharp negativity is in fact a compound spike potential representative of the synchronous discharge of postsynaptic neurons. He termed this potential the population spike
(PS) in agreement with Lorente de No (1947). More detailed analyses have confirmed the relationship of the PS and unit activity (Andersen et al., 1971b).
The plasticity exhibited at hippocampal synapses has been subdivided into broad categories based on the frequency and duration of
Figure 1-2. A CA1 field potential evoked by high intensity Schaffer collateral/commissural
stimulation and recorded just dorsal to the pyramidal cell layer. The labelled features of the waveform are: a, calibration pulse (1 mv, 2 msec, positive up) injected during baseline recording prior to stimulation; b, stimulus artifact; c, onset of the
population EPSP (seen as positive or source since the recording was taken remotely from
the active excitatory synaptic region or sink); d, population spike superimposed on the EPSP and indicative of synchronously active pyramidal cells. Quantification of
these potentials is described in the text. Calibration: 1 mv, 2 msec.
CAl EP
b d lmv
2 msec
23
afferent stimulation as well as the duration of response change. The categories include paired-pulse potentiation, frequency potentiation, posttetanic potentiation and long-term potentiation. Each of the phenomena will be briefly discussed.
Paired-pulse potentiation is identical to the condition-test technique described in other systems. In this situation, the response evoked by the first (conditioning) stimulus is compared to the response to the delayed second (test) stimulus. Such double-pulse stimulation of excitatory hippocampal afferents results in a potentiated EPSP and PS. Although the exact characteristics of potentiation differ for the PS and EPSP, potentiation can be observed with interpulse intervals as long as several hundreds of milliseconds (Assaf and Miller, 1978; Lomo, 1971b; Steward et al., 1977).
Frequency potentiation refers to an enhanced response, both EPSP and PS, during a train of stimuli to an afferent pathway. Unlike with paired pulses, qualitative differences are found among hippocampal subfields. Frequency potentiation is typical of mossy fiber-CA3 synapses as well as Schaffer collateral-CAl synapses. The optimal frequencies for potentiation are approximately 8-12 Hz (Dunwiddie and Lynch, 1978). Stimulation at these frequencies often leads to the appearance of double or triple population spikes, a phenomenon poorly understood at this time (Bliss and Lomo, 1973). On the other hand, the perforant path to dentate synapses often becomes depressed with repetitive stimulation (White et al., 1979). Detailed analyses of this occurrence indicated that the parametric features of the response decrement conformed to previously established criteria for "habituation" (Harris et al., 1979a; Teyler and Alger, 1976; Thompson and Spencer, 1966). Curiously,
24
habituation is observed in the hippocampus only at the perforant pathgranule cell synapse (Teyler and Alger, 1976).
The first event that follows tetanic stimulation is posttetanic potentiation (PTP), generally lasting only a few seconds. Careful analysis of this short-lasting augmentation in the dentate gyrus (McNaughton, 1977) indicated that its decay is described by two exponential equations with time constants consistent with those found at frog neuromuscular junction (Magleby and Zengel, 1976). Generally, PTP is found superimposed on a period of response depression which can last for several minutes (Bliss and Lomo, 1973; Deadwyler et al., 1978; Teyler et al., 1977). Subsequently, if the frequency and duration of stimulation is adequate, a later period of enhancement develops. This late-developing enhancement is unusual in that it can persist for extremely long periods of time, on the order of days or weeks (Buzsaki, 1980; Bliss and Gardner-Medwin, 1973; Douglas, 1977; Douglas and Goddard, 1975). Thus, it is commonly referred to as long-term potentiation (LTP). Long-term potentiation differs from frequency potentiation in that the optimal frequency for LTP is 100 Hz or more (Dunwiddie and Lynch, 1978).
In summary, the laminar organization of the hippocampus lends itself well to extracellular monitoring of synaptic activity. The relative ease of data collection and interpretation plus the lability of its synapses not only facilitateselectrophysiological analysis of hippocampal functioning but also makes this structure an exciting model of synaptic plasticity. Studies of the effects of chronic ethanol consumption on these processes would significantly enhance our understanding of ethanol's toxic effects on the hippocampus.
25
Animal Models of Chronic Alcohol Consumption
Effective animal models of chronic alcohol consumption must fulfill certain conditions. For example, the ethanol must be delivered in a manner that is practical for extended periods of time. The typical means of accomplishing this is through oral administration, although intragastric intubation techniques have been used. The animals must also receive a fairly large proportion of their calories as ethanol yet maintain adequate nutrition and normal body weight gain. This difficult though necessary requirement is not met by all paradigms. For instance, some studies employ aqueous ethanol solutions plus normal lab chow diets for the experimental animals. However, ingestion of ethanol-derived calories may lead to decreased consumption of lab chow, promoting malnutrition and reduced weight gain.
Our laboratory has previously developed a feeding procedure which overcomes the above limitations (Walker and Freund, 1971). Experimental animals receive a special liquid diet as their sole source of calories. The solution contains a large percentage of ethanol-derived calories (35-39%) yet is fortified with essential vitamins and minerals. Control animals are either pair-fed a solution with sucrose isocalorically substituted for ethanol or are maintained on standard lab chow and tap-water diets. As noted earlier, animals restricted to the ethanol-containing diets for extended periods of time show both severe behavioral deficits and pathological changes in hippocampal neuronal morphology (Riley and Walker, 1978; Walker and Hunter, 1978).
26
Rationale
The similarities of alcohol-related and hippocampus-related amnesic syndromes strongly suggest that chronic ethanol ingestion may be toxic to the hippocampus. Synaptic plasticity is a fundamental feature of the hippocampus and such plasticity is believed to have important behavioral implications. The purpose of the present research was to investigate whether ethanol has a detrimental effect on the functional plasticity of hippocampal synapses.
Studies of aging animals have supported the logic underlying these experiments. Both behavioral deficits and neuropathology in the hippocampus are prominent in aged animals (Barnes, 1979; Bondareff, 1979; Freund and Walker, 1971a; Bondareff and Geinisman, 1976; Geinisman et al., 1978; Scheibel et al., 1976). These findings have led other researchers to study the aged hippocampus electrophysiologically. Landfield et al. (1978) reported that the Schaffer collateral to CA1 synapses were essentially normal in all animals when given single or paired stimuli. However, when this system was challenged with high frequency stimulation, measures of both long- and short-term potentiation revealed diminished enhancement in aged animals.
More quantitative studies (Barnes, 1979; Barnes and McNaughton, 1980) have demonstrated a different pattern of results for the perforant path to dentate gyrus system. First, single stimuli at a fixed current produced smaller synaptic responses in aged animals. However, if one controlled for the number of fibers actually stimulated, the existing synapses in the aged animals proved to be more powerful. When a high frequency tetanus was delivered, aged animals showed normal PTP and LTP. Repeated tetani elevated the LTP asymptote of the
27
evoked response in younger animals, but they had no effects on the aged animals beyond those produced by the first tetanus. Barnes (1979) confirmed Landfield's finding that the groups did not differ in their capacity for paired-pulse potentiation. These exciting findings in aged animals emphasize the need for similar research in alcoholic animals. The following experiments represent the first attempt to investigate the long-term consequences of chronic ethanol consumption on hippocampal synaptic function.
CHAPTER II
GENERAL METHODS
Ethanol Administration
Male Long-Evans hooded rats (200-250 g) were matched by weight and assigned to the following treatment groups: 1) an experimental group (Group E) which received an ethanol-containing liquid diet, 2) a control group (Group S) which was pair-fed a sucrose-containing liquid diet, and in some experiments 3) a second control group (Group LC) which received pelleted laboratory chow and water ad libitum. The preparation, contents and nutritional adequacy of the liquid diet procedure have been documented by Walker and Freund (1971). The ethanol liquid diet had 35-39% ethanol-derived calories (8.1-9.4% v/v, ethanol) and was prepared by mixing an ethanol stock solution (63.3% v/v with Sustacal (Mead-Johnson Co.). The control liquid diet was identical except sucrose was isocalorically substituted for ethanol. Both diets were fortified with Vitamin Diet Fortification Mixture, 0.3 g/100 ml, and Salt mixture XIV, 0.5 g/100 ml (Nutritional Biochemicals Co.). The liquid diets provide 1.3 kcal/ml. The rats were kept individually housed in stainless steel cages'in a colony room with 0700-1900 hr light cycle. All liquid diets were prepared fresh each day and administered in calibrated drinking bottles with stainless steel drinking tubes. Diet consumption was measured daily.
The liquid diets were administered for a period of 20 weeks. The percentage of ethanol- or sucrose-derived calories was increased by 1%
28
29
every four weeks (35-39%). Group S rats were pair-fed with Group E rats in order to equalize caloric and nutrient intake during the 20 week period. At the end of the 20 week treatment period, all rats received laboratory chow and water ad libitum. Electrophysiological recordings were obtained within a 10 week period commencing eight weeks after discontinuation of the liquid diet treatment. This extended abstinence period was used to eliminate or at least minimize the residual effects of ethanol intoxication or acute ethanol withdrawal. Moreover, both the duration of ethanol treatment (20 weeks) and ethanol abstinence (eight weeks) were chosen to facilitate comparisions with previous morphological studies (cf. Walker et al., 1980). Rats from each group were coded in order to prevent experimenter bias in final electrode placement and data collection. The code was not broken until data analysis was completed.
Electrophysiological Methods
Recordings were made in urethane-anesthetized preparations (1.5 g/ kg ip.) with supplemental doses of urethane administered as required. Each animal was placed in a Kopf stereotaxic instrument (located within a screen-shielded cage) and the skull exposed. The skull tilt was adjusted so that lambda and bregma were level on the horizontal plane. All stereotaxic coordinates were referenced to bregma, midline and the skull surface. A heating pad was used to maintain rectal temperature at 37 0.50 C. Stimulating and recording electrodes were remotely controlled via hydraulic microdrives and placed in the brain through small burr holes in the skull.
30
Each animal was implanted with four electrodes. A screw electrode was placed anterior to the brain to ground the preparation. A 75 Pm platinum-iridium wire served as the reference electrode and was placed in the neocortex close to the hippocampal recording site. The stimulating electrode was a concentric bipolar (Rhodes Medical Instruments) with a 250 pm outer diameter and a 100 pm tip bared for 25 pm and extending 75 pm beyond the main shaft. The electrodes had an impedance of 20-30 kohms at 1 khz. A fiber-filled glass micropipette filled with 4 M NaCl (1-2 -pm tip diameter, 1-3 Mohms at 1 khz) was used for extracellular field potential recordings. Field potentials were amplified by a Grass P511 differential AC preamplifier, filtered at 0.3-10 khz and either recorded immediately on an X-Y plotter or stored on magnetic tape for later analysis. Potentials were in some cases averaged by a Dagan 4800 Signal Averager. Electrical stimulation was delivered by a Nuclear Chicago constant current stimulator and consisted of monophasic square wave pulses (0.1 msec duration). During high frequency stimulation, biphasic pulses (0.1 msec each half phase) were usually employed to reduce electrode polarization. A 1-2 my (2 msec) calibration pulse was routinely given just prior to each stimulus. Plotted field potentials were quantified with a Numonics Digitizer.
Experimental Protocol
The following experiments include studies of both CA1 and the DG.
Since these two regions have somewhat different physiological characteristics, the experimental protocol was somewhat different for each region. Nonetheless, the basic data collection procedures will be outlined here and deviations from the protocol will be noted in the appropriate chapters.
31
At the beginning of a recording session, the recording electrode was lowered into the appropriate hippocampal subfield while monitoring extracellular unit activity. The stimulation electrode was then placed to produce maximal activation of the afferent fibers of interest as revealed by the amplitude of the PS. Following determination of the optimal electrode configuration, synaptic distribution and function were evaluated through five separate stimulation procedures. 1) Input/ output (I/0) curves were generated by systematic variation of the stimulus current (10-1000 pa) in order to evaluate synaptic potency. Stimulus pulses were delivered at frequencies ranging from 0.1-0.03 Hz and four responses were averaged at each current intensity. 2) A laminar analysis was conducted by stepping the recording electrode in 25 pm increments through the subfield along an axis parallel to the orientation of the pyramidal or granule dendrites. The stimulus current was fixed to give an EPSP amplitude 50% of that seen at the PS threshold. At each 25 pm step and after a 25 sec recovery period, four to eight field potential responses to 0.1 Hz stimulation were recorded for subsequent averaging. 3) Paired-pulse potentiation (PPP) was evaluated at low (subthreshold for a PS) and high (suprathreshold for a PS) stimulus current intensities. Interpulse intervals (IPI) varied from 15-180 msec. Stimulus pairs were delivered at 0.1-0.03 Hz and four responses were averaged at each IPI. 4) From three to five frequency potentiation
(FP) series were conducted at varying stimulation frequencies and durations: 1 Hz/5 sec, 1 Hz/25 sec, 5 Hz/5 sec, 10 Hz/2.5 sec and 10 Hz/5 sec. The stimulus current was always suprathreshold for the production of a PS. This design allowed for comparisons of FP at equivalent tetanus duration (5 sec) or at equal numbers of stimuli within the
32
train (25 pulses). Every third response within each tetanus was measured and compared to the baseline response generated prior to each tetanus. 5) Following each FP series, test pulses were systematically delivered for 10-30 min in order to assess the development of LTP. In addition, LTP was produced by high frequency stimulation at the end of each experiment and monitored for 30 min.
Following data collection, small electrolytic lesions (10 Pa, 10 sec, anodal) were made at the stimulation site. Similar lesions were made through the recording electrode at the isopotential point and at a fixed distance from the isopotential point along the electrode tract. Measurement of the distance between the recording electrode lesions permitted correction for tissue shrinkage normally occurring during histological preparation. Electrode placements were verified in all animals through microscopic examination of the lesions in myelinstained sections.
Data Analysis
A major difficulty in making quantitative comparisons between treatment groups is the extreme within-group variance in evoked responses normally observed in the intact hippocampus (Barnes, 1979; Landfield et al., 1978). Several standardization procedures were employed to reduce this variability. First, following the laminar analysis, the recording electrode was placed at a fixed distance (125-150 pm) from the inversion point. All I/0 curves, PPP, FP and LTP sequences were performed at this fixed reference point in each animal. Secondly, stimulus current strengths for PPP, FP and LTP were all normalized with respect to the individual I/O curve. These values
33
will be detailed later for each study. The stimulus current was readjusted prior to tetanus delivery when long-term effects of the previous tetanus were observed. Finally, all electrodes were carefully selected to fall within a narrow range of tip size, impedance and shaft taper requirements.
Analysis of the evoked responses generated by the laminar analysis does not allow a precise localization of current sources and sinks. In order to enhance the spatial and temporal resolution of the analysis, a one-dimensional current-source density (CSD) analysis was applied to the data (see Appendix). Current densities were derived from the extracellular evoked responses by calculation of the second spatial derivative of the potentials using the D4 smoothing of Freeman and Nicholson (1975). The CSD computed with this smoothing was found to have low noise, yet still exhibit all major current sources and sinks. Since tissue conductivity was not determined, the results are expressed in mv/mm2. Due to the laminar and lamellar organization of the hippocampus, the onedimensional CSD as employed here assumes that the major synaptic currents flow parallel to the long axis of the pyramidal cells and that no large conductivity gradients exist in this dimension (Freeman and Stone, 1969; Nicholson and Freeman, 1975). A recent study employing CSD analysis of the Schaffer/commissural (SCH/COM) afferents to CA1 supports the validity of this approach (Leung, 1979), provided that the field of excitation sampled by the microelectrode is confined to the center of the lamella.
In order to minimize errors arising from violations of these assumptions, and to reduce response variability, the following standardization procedures were employed. First, the orientation of the stimulating
34
and recording electrodes was adjusted to correspond to the orientation of the lamellae for the SCH/COM-CAl path (Andersen et al., 1971a). The lamellae are oriented saggitally (relative to midline) with a slight shift anteriorly. In order to insure the intralamellar localization of the electrodes, the location of the recording electrode was adjusted until evoked potentials were obtained with EPSP and PS thresholds that fell within a relatively narrow range. There is less concern about the lamellar organization of the DG since the stimulating electrode was placed in the angular bundle and thus activated fibers which project all along the longitudinal axis of the hippocampus. The EPSP and PS thresholds were assessed for each animal at a fixed recording site relative to the inversion point of the laminar analysis. Finally, the stimulus current used to generate the laminar profiles was chosen according to the standardization procedure described above.
Following data analysis, the code was broken and the data grouped by treatment. The potentiation data were expressed as either difference from baseline response (latency measures) or percentage of baseline response (amplitude measures). The data were typically analyzed by 2-way analysis of variance (ANOVA) with repeated measures on one factor (treatment by stimulus condition). Individual comparisons were performed using Duncan's multiple range test and student t-tests. Withinanimal comparisons were accomplished through use of paired t-tests.
CHAPTER III
CHRONIC ETHANOL EXPOSURE AND SYNAPTIC DISTRIBUTION IN CA1
OF RAT HIPPOCAMPUS: CURRENT-SOURCE DENSITY ANALYSIS
Introduction
The pathological deterioration associated with chronic ethanol
abuse in man is normally attributed to a variety of coexisting conditions, most often malnutrition. However, experiments in laboratory animals have provided convincing evidence that ethanol, even in the presence of a nutritionally adequate diet, exerts toxic effects in the CNS. The neurotoxic actions are manifest both indirectly as acquisition deficits in a variety of tasks (Denoble and Begleiter, 1979; Fehr et al., 1976; Walker and Freund, 1971; Walker and Hunter, 1978) and directly as morphological deterioration in such brain regions as hippocampus and cerebellum (Riley and Walker, 1978; Walker et al., 1980). In the rat hippocampus, chronic ethanol treatment produces a 15-20% loss of granule cells of the dentate gyrus as well as pyramidal cells of CA1 and CA2-4. In addition, studies in Golgi material have provided evidence of dendritic atrophy and spine loss in the mouse hippocampus (Riley and Walker, 1978).
If the morphological deterioration observed in the hippocampus is to be related to the behavioral deficits in ethanol-treated animals, it is imperative that synaptic function be studied directly. In the present study, we examined electrophysiologically the persistent effects of chronic ethanol treatment on synaptic function in CA1 of the
35
36
rat hippocampus. Electrical stimulation within stratum radiatum simultaneously activates SCH and COM fibers (from ipsilateral and contralateral CA3, respectively) which course together and terminate on the pyramidal cells in stratum radiatum of CAl. The stratification of the various afferents and the homogeneous arrangement of cellular elements in CA1 allow meaningful extracellular analysis of synaptic potentials generated by the pyramidal cells. However, the extracellular potentials generated under these conditions can be rather large and volumeconducted for great distances, hampering the utility of this analysis. We have therefore combined extracellular evoked potential (EP) and current-source density analysis in order to more precisely define the distribution of synaptic currents induced by SCH/COM inputs to CA1 after chronic ethanol treatment.
Methods
The animals, treatment conditions and general electrophysiological methodology have been described previously in Chapter II. Following placement of the electrodes, the synaptic distribution of SCH/COM afferents to CA1 was examined by laminar analysis and assessment of I/0 relations. Initially, the extracellular isopotential point of the laminar analysis was determined and the recording electrode adjusted to a fixed point 125 pm more dorsally. Thresholds for both the EPSP and the PS were determined at this fixed reference point. The stimulus current was then adjusted to produce an EPSP which was 50% of the EPSP amplitude observed at the PS threshold. The recording electrode was moved to the point 350 pm dorsal to the inversion point and a laminar analysis was conducted in 25 pm increments at the fixed stimulus
37
current over a 1000 pm descent through the hippocampus. At each 25 jim step and after a 25 sec recovery period, four EP responses to
0.1 Hz stimulation were recorded for subsequent averaging and CSD analysis by a PDP 11/40 laboratory computer. At the end of the laminar analysis, the recording electrode was withdrawn to the point 125 Pm dorsal to the inversion point. Input/output functions were then obtained by systematically varying the stimulus current (20-1000 pa).
At the termination of each experiment, an electrolytic lesion was placed at the stimulation site. Electrolytic lesions (10 pa, 10 sec) through the recording electrode were made at the inversion point of the laminar analysis and 650 pm more ventrally. The location of the lesions was verified in all animals by light microscopy in myelinstained sections.
Results
The results are based on a total sample of 14 rats which were distributed as follows: E = 6, S = 4 and LC = 4. Because of the small sample and the fact that Groups S and LC rats did not differ on any measure, they were combined into a single control group (C = 8). Group E rats consumed a mean daily ethanol dosage of 13.91 g/kg during the 20 week treatment period. While blood ethanol concentrations were not evaluated in these rats, both the pattern and magnitude of daily ethanol intake were comparable to previous experiments in which behavioral or morphological changes were observed (Walker and Hunter, 1978; Walker et al., 1980). Mean body weights at the termination of the experiment were comparable across groups: E = 531 g, C = 513 g.
38
Histological analysis. One consequence of significant cell loss could be a selective shrinkage of hippocampal volume relative to the remaining brain mass. If this occurred, the relative position of electrodes placed within the hippocampus using fixed stereotaxic coordinates could be differentially affected by ethanol treatment. However, three-dimensional plots of electrode tip positions relative to the septal pole of the hippocampus, the brain surface and the midline revealed no significant differences between groups. Stimulating electrode sites were localized to stratum radiatum usually near the CAl-CA2 border, while recording electrode tracts were consistently found 7001000 pm more posteriorly in CAl. All electrode placements were confined to the dorsal hippocampus. Lesions at the isopotential point of the laminar analysis were always near the border of stratum pyramidale and radiatum (range, relative to stratum pyramidale: 0-44 pm into stratum radiatum). Detailed measures (corrected for histological shrinkage) were made of the major laminae of CAI including stratum oriens, pyramidale, radiatum and lacunosum-moleculare. These measures were made along the actual trajectory of the electrode track through CAl. Significant differences in the width of the laminae were not observed between groups (Table 3-1).
I/O relationships. Chronic ethanol treatment failed to significantly alter basic synaptic responses to SCH/COM stimulation. Neither EPSP thresholds (E = 70.0 4.8 pa; C = 52.5 6.5 pa) nor PS thresholds (E = 260.0 46.0 pa; C = 218.7 40.3 pa), nor EPSP amplitude at PS threshold (E = 2.96 0.39 my; C = 2.91 0.34 my) was significantly different between groups (data presented as mean SEM). Finally, 2-way ANOVAs revealed no statistically significant group differences over the
39
TABLE 3-1
Chronic Ethanol Effects on Widths of CA1 Laminaea
Laminab Control Ethanol A A% SO width (pm) 128.8 7.0 141.5 4.0 12.7 10.0 SP width (pm) 43.2 2.9 49.8 2.5 6.6 15.3 SR width (pm) 279.9 7.0 270.3 8.2 -9.6 -3.4 SM width (pm) 86.6 3.9 80.6 4.6 -6.0 -6.9 CA1 width (pm) 538.5 10.7 541.5 12.7 3.0 0.6 aMean + SEM
bSO, stratum oriens; SP, stratum pyramidale; SR, stratum radiatum; SM, stratum moleculare
40
range of stimulus currents examined (20-1000 pa) for latency to EPSP onset, latency to peak of the PS, EPSP amplitude or PS amplitude.
Field potential and CSD analysis. Stimulation of the SCH/COM
pathway produced a surface-positive, deep-negative dipole field. These potentials reflect summated excitatory synaptic activity (current sinks) in stratum radiatum, the terminal region of SCH/COM fibers, and an outward (largely capacitative) current source in stratum pyramidale and the proximal portions of stratum oriens. Figure 3-1 shows a typical EP profile and associated CSD distribution plotted as a function of depth from the ventral alvear surface.
The spatial distributions of the field potentials and current densities within CA1 are more easily understood when amplitude profiles are generated at fixed latencies from stimulus onset. Figure 3-2 illustrates EP and CSD profiles at 8.0 msec (from stimulus) for both a control and an ethanol-treated animal. It is clear from this figure that the CSD plots provide considerably more localization of current flow than the EP plots. In both animals, the major current sink occurred in stratum radiatum, and was bounded dorsally and ventrally by current sources. A major current source was observed in stratum pyramidale and proximal portions of stratum oriens. A smaller current source extended through most of stratum moleculare to the hippocampal fissure. These results compare favorably with previous descriptions of current densities in the SCH/COM-CAI path (Schubert and Mitzdorf, 1979). However, in contrast, the current sink in stratum radiatum always exhibited two components (Fig. 3-2) which were separated by near zero current density or, more rarely, by a minor source in the middle of stratum radiatum. These two component sinks may reflect a separation of the COM and
Figure 3-1. Full one-dimensional laminar analysis through CA1 for a single Group E animal.
Field potentials are represented on the left, current densities on the right. Depths indicate distance ventral to the ventral alvear border. In this animal, the stratum
pyramidale-stratum radiatum border is found about 175 pm from the alveus. Stimulation
was applied at time T1 but the artifact has been suppressed in this figure. The T2 bar (8.0 msec from stimulus) indicates the slice of time used for Figures 3-2 and 3-3. The
EP waveforms include a calibration pulse (1 mv, 2 msec, positive up) prior to the stimulus. For the CSD waveforms, an upward deflection from baseline represents a
source; a downward deflection represents a sink.
DEPTH EP CSD
(Mm) T T2 I 2
0
5 0 _. ._. 00
CAL
. ..... .....-- m / 2
150- .2mV 500 2 mV/mm2
200250- -----------300350
400- -/-- \- ----
450- --500
525
0 8.0 16.0 24.0 0 8.0 16.0 24.0
TIME (ms) TIME (ms)
Figure 3-2. Individual laminar profiles of EPs and CSDs at 8.0
msec after stimulus delivery showing EP and CSD amplitudes at a fixed latency from stimulus illustrates more clearly
the change in responses recorded across the width of CA1 (compare with Fig. 3-1). The y-axes are scaled to indicate depth from the ventral border of the alveus. Note
the different x-axis scalings. Data from an ethanol animal is shown on the top half of the figure; a control
animal is presented on the bottom half. The field potentials change polarity at about the stratum pyramidaleradiatum border. Positive currents are sources; negative
currents are sinks.
44
FP CSD O
100
; . . . . . . . . . . . .
200-: .
. ....... . . .
. . . i . . ; . i . . . . . . .
300- ......
ETHANOL
.. . i . . . . . . . . . ; . . .. . . . .
400
] ~. . . . . . ;
500
-70 -35 0 35 70 -2000 -1000 0 1000 2000 V(mV) Im (mv/ mm2)
E
100
...... ............. . C O N T RO L
300
400
500-:
. .... . . . . . . . . .
. . . . . .. ... . i . . . . i . . .
. . . . . .. . i . . . . . . . . . .
S . . . .. . . . . I . . . . I . . . . I . . . . . . . . | . . . . .
-50 -25 0 2.5 5.0 -600 -300 0 300 600
V (my) Im (my/mm2)
45
SCH terminal fields within stratum radiatum, a hypothesis that will be considered in detail below.
Group comparisons of the spatial distribution of the field potentials and current densities required normalization, since the absolute measures of the major laminae of CA1 varied across animals (Table 3-1). The following normalization was employed. First, EP and CSD profiles during the rising phase (4.0 msec) and at the peak (8.0 msec) of current flow were obtained for each animal (e.g., Fig. 3-2). Only those values contained between the ventral alvear border and the hippocampal fissure (as determined from electrode track reconstruction) were considered. This distance was then divided into equal increments of 4% of the total distance. Since the mean total distance was approximately 540 im for each group (Table 3-1), each equivalent step (4%) reflected a mean distance of 21.6 pm. The amplitude of the EP and CSD for each animal was obtained at these percentage increments using linear interpolation between data points. This analysis was performed with a Numonics Digitizer.
Group laminar profiles (at 8.0 msec) are presented in Figure 3-3 relative to a sample CA1 pyramidal cell and the mean widths of the major CA1 laminae. While the spatial resolution of the current sinks is diminished in the control group plot, two distinct peaks of inward current are still found at 71.3 Pm and 228.3 pm from the pyramidal cell layer. It can be readily seen that while chronic ethanol treatment produced only subtle alterations in the EP profiles, the ethanol treatment dramatically altered the spatial distribution and density of the synaptic currents as revealed by CSD analysis.
Figure 3-3. Group laminar profiles of normalized EPs and CSDs at 8.0 msec after stimulus
delivery. On the left is a representative pyramidal cell and demarcations of the major
CAl laminae (adapted from Scheibel, 1979). The distance in microns through CAl is an
average for all animals (there were no group differences in this measure). The normalization procedure for both EP and CSD data is described in the text. The EP laminar
profile reveals positive potentials in SO and SP which invert just past the cell layer to negative potentials throughout SR and SM. There are no significant ethanol effects except for the negative potential half-width (see text). The CSD profile resolves the
current sources (positive Im) and current sinks (negative Im) much more precisely.
Passive current sources are observed in SO/SP and in SM. Note the two peaks of inward
current in SR of control animals (filled circles). The ethanol group (open circles)
shows an overall shrinkage of the sink in SR, a reduced component sink proximal to the
cell layer and an expanded component sink more distally from the cell layer. Abbreviations: ALV, alveus; SO, stratum oriens; SP, stratum pyramidale; SR, stratum radiatum;
SM, stratum moleculare; HF, hippocampal fissure.
ALV DEPTH (UM) EP CSD so ,,.
100 - 100C 00' .-oSP -all/J o S200 -- -200
,
SR 300 - 300 o
/ *CONTROL pp
o o ALCOHOL
0
-400- -400sm 0
SM 'o 500- -500
HF
-60 -50 -40 -30 -20 -,o 0 I 20 -40 -30 -20 -10 0 10 20 30 40 50 60 AMPLITUDE (my) AMPLITUDE (I: mv/mm2 102)
48
The amplitude and distribution of field potentials and current
densities in stratum radiatum were quantified at both 4.0 and 8.0 msec following stimulus onset. Since the spatial patterns of current flow are similar at the two time points, only the results at 8.0 msec will be described (Table 3-2). There were no significant group differences for the majority of EP measures. However, Group E showed a significant reduction of the spatial half-width of the negative potential region. (Spatial half-width was defined as the distance across stratum radiatum between the two points whose amplitudes lay half way between the peak negative voltage and the asymptotic negative voltage.) Quantitative comparisons of Group C and Group E CSD profiles more precisely characterized the changes resulting from the chronic ethanol exposure. Ethanol treatment significantly reduced the spatial extent of the overall current sink in stratum radiatum. This shrinkage of the SCH/ COM synaptic field occurred despite procedures which normalized the individual distance from alveus to hippocampus fissure. Separate sink analyses revealed that the overall shrinkage was totally accounted for by a shrinkage of the proximal current sink. Interestingly, the distal sink of Group E was somewhat enlarged, with a significantly greater amplitude and area, but only a weak trend towards increased spatial extent. These data are suggestive of an increased efficacy of distally located synapses. However, since we did not directly measure tissue conductivity, the amplitude and area measures are only proportional to the actual CSD. Thus, group differences in these measures could reflect a change in tissue conductivity instead of alterations in evoked synaptic current.
49
TABLE 3-2
Laminar Analysis Quantification (8.0 msec from stimulus)a
Chronic Ethanol Effects on Field Potential and Current Density Distributions
Measure Control Ethanol A% EP
Peak voltage (mv) -5.5 0.8 -6.3 0.5 14.1 Peak latency (msec) 8.9 0.4 8.0 0.5 -10.2 Spatial half-width (pm) 206.5 8.9 172.4 9.0 -16.5c Peak distance to cell layer (pm) 196.8 14.0 184.6 6.3 -6.2 Inversion distance to cell
layer (pm) 6.9 5.4 8.2 3.5 18.8 CSD
Total sink width (pm) 283.8 14.1 246.4 10.0 -13.2b Distance from inter-sink
minimum to cell layer (pm) 163.5 19.0 104.0 8.6 -36.4c Length along dendritic tree (pm)
proximal sink 141.3+ 14.4 88.9 8.8 -37.1c distal sink 141.2 18.3 168.0 16.4 19.0 Peak distance to cell layer (pm)
proximal sink 71.3 16.1 56.4 6.0 -20.9 distal sink 228.3 14.5 181.1 9.2 -20.7b Peak amplitude (mv/mm2)
proximal sink 511.2 83.9 436.2151.3 -14.7
distal sink 418.1 99.7 673.6 78.4 61.1b Sink area (mv/mm)
proximal sink 44.7 7.2 27.2 8.1 -39.1
distal sink 40.0 11.6 70.9 7.2 77.2b aMean SEM
p < .05 1 tail
cp < .05 2 tail
50
Discussion
The results of this experiment further our understanding of the
pattern of synaptic distribution of SCH/COM afferents to stratum radiatum of CA1 in both control and ethanol-treated rats. The CSD analysis of laminar profiles through CA1 indicated the presence of a major excitatory current sink in stratum radiatum but with two spatially distinct components. Chronic ethanol treatment produced a persistent reduction in the spatial extent of the current sink proximal to stratum pyramidale with a less profound expansion of the more distal current sink. The significance of each of these findings will be considered in turn below.
Synaptic distribution in stratum radiatum of CAl. It is currently well documented that the major afferents to stratum radiatum and stratum oriens of CA1 arise from both ipsilateral and contralateral CA3 (Blackstad, 1956; Hjorth-Simonsen, 1973; Laurberg, 1979; Swanson et al., 1978). However, within stratum radiatum the density and spatial distribution of terminals from each source has not been satisfactorily characterized. The major unresolved question is whether or not these two afferents terminate differentially across the width of stratum radiatum. Early Golgi studies by Schaffer (1892) and Lorente de No (1934) showed that the Schaffer collaterals run in a dense band (stratum lacunosum) distal to stratum radiatum and presumably terminate in this region and the neighboring portion of stratum radiatum. This laminar organization is particularly prevalent in the rabbit (Lorente de No, 1934). Subsequent physiological evidence supporting a bimodal distribution of SCH/COM afferents to stratum radiatum has emerged from the study of laminar profiles of evoked field potential in rabbits, where
51
peak responses to activation of COM afferents were found more proximally in stratum radiatum (Andersen et al., 1966a; Stanley et al., 1979) than peak responses to SCH stimulation (Andersen et al., 1966a).
Anatomical studies in the rat, employing degeneration (Blackstad, 1956; Hjorth-Simonsen, 1973), autoradiographic (Laurberg, 1979; Swanson et al., 1978) and horseradish peroxidase (Laurberg, 1979) tracing techniques, have been unable to qualitatively distinguish between the spatial distribution of associational and commissural afferents to CAl. In contrast, others using degeneration and autoradiographic techniques have been able to identify a bimodal yet overlapping distribution of CA3 afferents to the apical dendrites of CAI pyramidal cells in the rat (Gottlieb and Cowan, 1973; Raisman et al., 1965). Gottlieb and Cowan (1973), by counting silver grains in autoradiographs, have presented the only quantitative anatomical evidence that COM terminals concentrate most heavily proximally to stratum pyramidale.
While the available anatomical and physiological evidence is
presently inconclusive (particularly in the rat), there is some support for the notion that SCH/COM afferents distribute in an overlapping yet bimodal fashion in stratum radiatum. We have therefore hypothesized that the dual current sinks observed in stratum radiatum in the present experiment reflect a separation of the COM and SCH synaptic fields. However, three important methodological issues must be considered in relation to this hypothesis. First, the utility of EP recordings in mapping studies, such as those mentioned above (cf. Andersen et al., 1966a), can be considerably enhanced through the use of CSD analysis. The CSD analysis provides a significant improvement in both spatial and temporal resolution of current density in the region of the recording
52
electrode. Thus, current sinks (current entering the cells) and current sources (current entering the extracellular space) can be precisely localized (Freeman and Stone, 1969; Nicholson and Freeman, 1975). However, in contrast to our results, previous application of CSD analysis to the distribution of synaptic currents following stimulation of stratum radiatum have revealed only a single current sink in stratum radiatum of CA1 (Leung, 1979; Schubert and Mitzdorf, 1979). These latter studies used recording intervals of 100 pm as compared to the 25 pm intervals employed in the present experiment. Since the stratum radiatum is approximately 300-400 pm in the rat (Table 3-1), we believe this discrepancy can be accounted for by the enhanced spatial resolution provided by the 25 pm sampling interval used in our experiments.
A second important factor is the lack of tissue conductivity measurements in the hippocampus. For this reason, the conductivity and gradient-dependent terms (Freeman and Stone, 1969; Nicholson and Freeman, 1975) remain unevaluated in the CSD calculations. These terms, which are necessary for an accurate quantitative evaluation of transmembrane currents, have been shown to contribute relatively little to the overall current densities in the optic tectum of pigeons and teleosts (Freeman and Stone, 1969; Vanegas et al., 1979) and in the anuran cerebellum (Nicholson and Freeman, 1975). Although the results of the calculations have been expressed in mv/mm2 (units which are only proportional to transmembrane current), the existence of significant conductivity gradients can confound interpretation of source-sink distributions in the region of the gradient (see Haberly and Shepard, 1973). While this is a concern in a laminated structure such as the hippocampus, there is
53
no compelling evidence from the anatomical organization to suggest the presence of significant conductivity gradients in the axis parallel to the pyramidal cell dendrites. Nonetheless, it remains possible that careful evaluation of these terms could change the form of the CSD distribution observed in stratum radiatum after SCH/COM activation.
Another possible source of artifact which may account for the appearance of two current sinks in stratum radiatum is related to the width of the activated synaptic population. Nicholson and Freeman (1975) have calculated radial CSDs for different ratios of the activated population width versus length. If the activated population is narrow, even recordings on the axis of symmetry can produce a distorted CSD distribution, with artifactual bimodal sink peaks in the true sink region (see their Fig. 6). Stimulation of the SCH/COM system in rabbit hippocampus, however, produces lamellar synaptic activity with a half-width of about 4 mm (Andersen et al., 1971a). This corresponds to a width-length ratio of about 4, for which artifactual multipeaked distributions are not theoretically expected. In addition, the procedures used in our study for placement of recording and stimulating electrodes should insure recordings along the axis of symmetry of the activated population.
It should be noted that the spatial distribution of current sinks does not necessarily quantitatively reflect the density of synapses along the dendritic tree. This discrepancy arises because, when considering the effective transmembrane currents at any one point along the dendritic tree, contributions not only from synapses at that point (if excitatory, then the effective current sink), but also the passive currents from remote synapses (in this case, sources) must be considered.
54
The passive currents will depend on the membrane properties of the dendritic tree. The result is that the observed current-source density at any given point in the tissue may be approximated by the convolution integral of the density of activated synapses over the entire dendritic tree and the spatial distribution of currents produced by any one synapse. Proper evaluation of the synaptic density distribution given the CSD distribution requires development of a reasonably accurate model representing the dendritic-somatic membrane and deconvolution of the integral. Neither task is trivial.
While we recognize these limitations in our interpretation of current source-sink distributions, the working hypothesis that the two sink regions indicate the presence of two overlapping but spatially separate synaptic inputs corresponding to the SCH and COM afferents will be used below in the interpretation of the effects of chronic ethanol treatment on synaptic distribution in stratum radiatum. We are currently attempting to verify this hypothesis through the use of more discrete microstimulation techniques.
Chronic ethanol treatment. Chronic ethanol treatment produced persistent alterations in the synaptic distribution of SCH/COM afferents to stratum radiatum of CAl. Ethanol treatment produced a significant reduction (13%) of the overall spatial extent of the major current sink within stratum radiatum. This presumed loss of SCH/COM synapses correlates well with the 15-20% cell loss observed in all hippocampal subfields at identical durations of ethanol exposure (Walker et al., 1980). This reduction does not merely result from morphological shrinkage of the laminae of CA1 (which might be expected after 15-20% cell loss) since our normalization procedure controlled for group differences
55
in the width of CAl. Rather, ethanol treatment appeared to produce an underlying shrinkage of the extent of excitatory current flow within stratum radiatum independent of any alterations in CA1 layer thicknesses. The failure to detect changes in the widths of the laminae of CA1 (Table 3-1) is of questionable validity at present, since these measures were not made in anatomically matched sections but rather along the trajectory of the electrode penetrations, a procedure which greatly increases variability.
Ethanol treatment also produced differential effects on the two
components of the major current sink in stratum radiatum. Measures obtained at the peak of the synaptic response (8.0 msec) showed the spatial extent of the proximal current sink to be reduced by nearly 40% in ethanol-treated rats with little or no change in the peak amplitude or area. On the other hand, the distal current sink exhibited a somewhat expanded spatial extent and a significantly greater peak amplitude and area. Since our CSD analysis did not include measures of conductivity, these latter changes in the magnitude of excitatory current flow may only reflect group differences in tissue conductivity, perhaps arising from changes in tissue morphology. However, the facts that the I/0 functions (which reflect the combined SCH/COM input) did not differ across groups and the distal current sink did exhibit a slight enlargement combine to strongly suggest that an increase in synaptic efficacy among the afferents contributing to the distal current sink occurred in ethanol-treated rats in order to compensate for the loss of synaptic drive among synapses more proximal to stratum pyramidale. According to our working hypothesis, these results suggest that chronic ethanol consumption leads to a selective reduction of COM afferents to CAl. This
56
selective deafferentation may provide the stimulus for a compensatory enhancement of synaptic drive among SCH afferents, which could reflect sprouting and formation of additional synapses among SCH afferents (Goldowitz et al., 1979) or an increase in the efficacy of existing synapses (Barnes, 1979).
It is difficult to specify the mechanisms underlying the relatively selective effects of chronic ethanol treatment on COM afferents. A recent experiment, employing retrograde double-labelling techniques, has strongly suggested that SCH and COM afferents to CA1 arise as axon collaterals from the same population of cells in CA3 (Swanson et al., 1980). Thus, it is unlikely that ethanol treatment selectively reduces a subpopulation of cells in CA3 which serves as the source of commissural afferents to CAl. In order for chronic ethanol treatment to selectively reduce COM afferents to CA1, apparently it must damage COM while sparing SCH axon collaterals even though both may arise from the same pyramidal cell. This requirement would be satisfied if long-tract myelinated fibers were particularly sensitive to the toxic actions of ethanol. Such a possibility not only explains the destruction of myelinated COM fibers but also the relative sparing of the SCH axons which are only weakly myelinated in the rat (Andersen et al., 1978). Interestingly, Marchiafava-Bignami disease, a somewhat rare complication of alcoholism in man, has been characterized by demyelination within the major hemispheric commissures, notably the corpus callosum and anterior commissure (Dreyfus, 1974). This hypothesis would predict that the commissural fibers to CA3 and the DG should be similarly reduced by ethanol treatment. Alternatively, ethanol treatment could act within CA1 to selectively destroy axon terminals in the proximal portion of
57
stratum radiatum. This hypothesis would be valid irrespective of the source of afferents contributing to the dual currents sinks. However, because ethanol most likely distributes uniformly throughout the brain (Sunahara et al., 1978), the hypothesis would require that differences exist between the proximal and distal stratum radiatum in some morphological feature such as the extent of vascular supply or proximity to the ventricular system, neither of which presently provides a compelling explanation of our results (cf. Coyle, 1978). Finally, it is possible that both SCH and COM afferents are damaged by ethanol exposure but only the SCH fibers have the capacity for compensatory regeneration. Unfortunately, on the basis of existing evidence, it is not possible to choose between these many alternatives.
The results of the present experiment coupled with our previous anatomical evidence indicate that chronic ethanol treatment in the presence of a nutritionally adequate diet induces a complex sequence of structural and functional changes in the hippocampus. Chronic ethanol treatment produces a loss of both hippocampal pyramidal and dentate gyrus granule cells (Walker et al., 1980). Because of the intrinsic and commissural connections of the hippocampus, such cell loss produces partial deafferentation of neurons in the hippocampal formation, a result which could explain the reduction in dendritic spines and dendritic branching observed at comparable durations of ethanol treatment (Riley and Walker, 1978; McMullen et al., 1980). The present study provides physiological evidence for deafferentation in CAl of the hippocampus, but deafferentation which is selective to the COM afferents. These findings agree with preliminary observations in Golgi material in which
58
chronic ethanol treatment attenuated apical dendritic branching proximal to the CA1 pyramidal cell layer (McMullen et al., 1980).
Deafferentation in the hippocampus is known to stimulate synaptic reorganization whose extent depends upon the specific afferents and the hippocampal subfield deafferented (Goldowitz et al., 1979; Nadler et al., 1980a; Nadler et al., 1980b). While we have noted a reduction of dendritic spines in CA1 (Riley and Walker, 1978), shorter durations of ethanol treatment have been reported to increase dendritic spines in CA1 (Kunz et al., 1976). The present results also suggest that compensatory synaptic reorganization may occur in association with deafferentation induced by chronic ethanol treatment. It is not yet clear whether the apparent compensatory increase in SCH afferent synaptic drive observed in this experiment began during the course of ethanol treatment or during the extended period of ethanol abstinence (eight weeks) prior to sampling. In either case, synaptic reorganization has important implications for studies of functional recovery from the cognitive impairment produced by chronic alcoholism in man. The available evidence now indicates that chronic ethanol treatment can induce numerous changes in the hippocampus including cell loss, deafferentation and synaptic reorganization. Further research will be required to understand the contribution of each of these changes to the cognitive impairments produced by chronic ethanol exposure.
CHAPTER IV
AUGMENTATION OF SHORT-TERM PLASTICITY IN CAI OF RAT
HIPPOCAMPUS AFTER CHRONIC ETHANOL TREATMENT
Introduction
Chronic alcoholism is often associated with pathological deterioration in many organ systems including the CNS. The most severe CNS disorder, Wernicke-Korsakoff syndrome, is associated with widespread neuropathology and a variety of neurological symptoms, the hallmark of which is adebilitating and permanent anterograde amnesia in the absence of a general cognitive decline (Courville, 1966; Talland, 1965; Victor et al., 1971). This pathological deterioration has been attributed to several coexisting conditions, especially malnutrition and thiamine deficiency (Victor et al., 1971). However, it seems likely that ethanol exerts direct neurotoxic effects in the CNS, since brain damage and neuropsychological deterioration have been observed in chronic alcoholic patients with no history of malnutrition, head trauma or exposure to other toxic agents (Epstein et al., 1977; Haug, 1968; Smith et al., 1973; Tumarkin et al., 1955). Moreover, animal studies have shown that chronic ethanol exposure in the presence of a nutritionally adequate diet results in 1) retarded acquisition of a variety of behavioral tasks in rodents including shuttlebox avoidance (Freund and Walker, 1971b; Sotzing and Brown, 1976; Walker and Freund, 1971), DRL (Denoble and Begleiter, 1979; MacDonell and Marcuella, 1978; Walker and Hunter, 1978), go-no-go discrimination (Walker and Hunter, 1978) and complex maze
59
60
acquisition (Fehr et al., 1976), 2) a 15-20% loss of granule cells in the dentateand pyramidal cells in CA1 and CA2-4 in rat hippocampus (Walker et al.,1980, and 3) spine loss and dendritic atrophy in granule cells and the basilar dendrites of CAI pyramidal cells of the rodent hippocampus (Riley and Walker, 1978). On the basis of these morphological results, physiological studies of hippocampal synaptic connections of rats chronically exposed to ethanol might reveal impaired function commensurate with the neuropathology. However, other outcomes are possible, particularly since deafferentation within the hippocampus leads to considerable reorganization of existing connections (Goldowitz et al., 1979; Nadler et al., 1980a; Nadler et al., 1980b).
The present study examined electrophysiologically the persistent effects of chronic ethanol consumption on synaptic function in stratum radiatum of CA1 in the rat hippocampus. Repetitive stimulation of SCH/ COM afferents to CA1 evokes large extracellular field potentials which are highly labile, exhibiting both short- and long-lasting changes in response to relatively brief tetanic stimuli (Alger and Teyler, 1976; Creager et al., 1980; Dunwiddie and Lynch, 1978; Schwartzkroin and Wester, 1975). We have examined such plasticity in order to assess synaptic function within the hippocampus following chronic ethanol treatment.
Methods
The animals, treatment conditions and general electrophysiological methodology have been described previously in Chapter II. However, since certain details of the protocol differ between this study and
61
the following one in the DG, procedures unique to this experiment are presented below
The stimulating and recording electrodes were placed in the brain through small burr holes in the skull. The coordinates for the stimulating and recording electrodes were (relative to bregma): 3.2 mm posterior and 2.8 mm lateral, and 4.2 mm posterior and 3.0 mm lateral, respectively. Once the stimulating electrode was placed to produce optimal activation of SCH/COM fibers (near the CAl-CA2 border), the recording electrode was stationed 125 im dorsal to the inversion point.
Initially, I/0 curves were generated using currents ranging from 20-1000 pa. The PPP was evaluated at low (EPSP amplitude 50% of that at PS threshold) and high (PS amplitude 40% of the asymptotic PS amplitude) current intensities. The IPIs varied from 20-180 msec. Stimulus pairs were delivered at 0.1 Hz. A total of five FP series were conducted at various stimulation frequencies and durations: 1 Hz/5 sec, 1 Hz/25 sec, 5 Hz/5 sec, 10 Hz/2.5 sec and 10 Hz/5 sec. Test stimuli (0.03 Hz) were then delivered posttetanus for 15-30 min to examine the amount of any LTP that may have been produced by the stimulation. When LTP was produced, the stimulation current was readjusted to give the standard PS response, 20% of asymptote, prior to proceeding with the next stimulus train. The final procedure was to stimulate at 100 Hz for 10 sec to insure the production of LTP. Responses to test stimuli were followed for 30 min posttetanus. Animals exhibiting afterdischarge posttetanus were not included in this analysis.
Four EP measures were commonly assessed: EPSP onset latency, PS peak latency, EPSP amplitude and PS amplitude. Because the precise onset of the EPSP was difficult to gauge due to its shallow initial
62
slope, the EPSP onset was arbitrarily chosen to be that point where the EP rose 200 pv above baseline. In an attempt to avoid contamination by the PS, EPSP amplitude was measured 0.75 msec following EPSP onset. The PS peak latency was measured from EPSP onset. Finally, PS amplitude was assessed by averaging the amplitudes from the peak negativity to the preceding and following positive peaks (cf. Alger and Teyler, 1976). Although others have used only the amplitude from the first positive peak to the negative peak of the PS (e.g., Lomo, 1971a) as the measure of PS amplitude, these two methods give quite similar results. For this reason, we chose the more conventional averaging method.
Results
The following results were based on a total sample of 29 rats which were distributed as follows: Group E = 12, Group S = 10 and Group LC =
7. Because Group S and Group LC rats did not differ on any measure, these groups were combined into a single control group (Group C = 17). Group E animals consumed a mean daily ethanol dosage (14.14 g/kg) comparable to previous experiments in which either associative deficits or neuronal loss in hippocampus has been observed. Body weights did not differ among the three groups at any point during the experiment. Mean ( SEM) body weights at the end of the experiment were: Group E = 553.1 15.4 g, Group C = 519.1 10.0 g.
Histological analysis. Chronic ethanol treatment has been shown to produce cell loss in the rat hippocampus (Walker et al., 1980). If one consequence of this cell loss is a selective shrinkage of the volume of the hippocampus, then the relative positions of electrodes placed within the hippocampus using fixed stereotaxic coordinates might be
63
differentially altered in Group E. In order to address this issue, electrode tracts were localized and plotted in three dimensions relative to the brain surface, midline and septal pole of the hippocampus. This analysis revealed no differences among the three groups. Stimulating electrode sites were in stratum radiatum near the CAlI-CA2 border. Recording electrode tracts were localized to dorsal hippocampal CA1 (see Fig. 1-1). Electrolytic lesions placed at the inversion point of the laminar analysis were consistently found at the stratum pyramidaleradiatum border. The location of the inversion point, with respect to the cell layer, also did not differ across groups. Finally, detailed measures (corrected for tissue shrinkage) were made of the major laminae of CA1 including stratum oriens, pyramidale and radiatum. These measures were made along the trajectory of the electrode track and also did not differ between groups.
I/0 relationships. Chronic ethanol treatment did not produce statistically significant alterations in the basic synaptic responses to single pulse stimulation (Fig. 4-1). The threshold current (Pa) required to elicit an EPSP was: Group E = 61.7 3.8; Group C = 61.2 6.1. The PS threshold current values were also virtually identical across groups (Group E = 239.5 23.9; Group C = 251.8 27.9). Moreover, the EPSP amplitude at PS threshold (Group E = 2.96 0.39 my; Group C = 2.91 0.34 my) and the PS amplitude at asymptote (Group E = 5.78 1.04 my; Group C = 6.23 0.62 my) did not differ between groups. These results are important, since these values were used to standardize the stimulus currents used during subsequent potentiation series in each rat. Finally, 2-way ANOVA revealed no statistically significant group differences over the range of stimulus currents examined (20-1000 pa) for latency
Figure 4-1. I/0 curves (mean SEM) for ethanol-treated (filled symbols) and control animals
(open symbols). The data are standardized by expressing values in terms of stimulus current steps from EPSP threshold (C) or PS threshold (A,B) as described in the text.
A. Plots of EPSP onset latencies (circles) and PS peak latencies (squares) by stimulus
current. While ethanol had no statistically significant effects, it did produce a small
trend toward reduced PS and EPSP latencies. B. Plot of PS amplitude by stimulus current.
Ethanol produced a small but nonsignificant trend toward reduced responses at high currents. C. EPSP peak amplitude at currents subthreshold for the production of a PS.
Again, there were no significant treatment effects.
70- 70
A B. 60 60 SEPSP ALCOHOL
UPS
50 0 EPSP CONTROL 5 U r 40- 40- C 30- O'
z <
S30 30- 2040 120
0 0
0 1200 300 400 500 0 100 200 300 400 500 CURRENT (pA) CURRENT (PA)
66
to EPSP onset, latency to peak of the PS, EPSP amplitude or PS amplitude.
Paired-pulse potentiation. Chronic ethanol treatment failed to alter the pattern of response to paired-pulse stimulation. In both groups, facilitation of the test pulse peak EPSP amplitude (at current levels subthreshold for PS) was maximal at an interpulse interval of 30 msec, where potentiation 150-175% of control was observed (Fig. 4-2B). In contrast, test pulse PS amplitudes were dramatically inhibited at these short IPIs, but exhibited facilitation at pulse intervals of 80 msec or greater (Fig. 4-2B). This differential action of paired stimuli on EPSP and PS amplitude is identical to that observed in previous studies of CA1 and the DG (Creager et al., 1980; Landfield et al., 1978; Lomo, 1971b; Steward et al., 1977). While the pattern of response to paired stimuli was preserved, chronic ethanol treatment did produce significant changes in the magnitude of PPP. Two-way ANOVA revealed a significant treatment effect for PS amplitude (F(1,39) = 6.58, p < .02) and a significant group X IPI interaction for the latency to PS (F(1,39) = 2.47, p < .03). Subsequent t-tests revealed these effects to be maximal at 100-175 msec IPIs. The chronic ethanol-induced enhancement of PPP is illustrated in Figure 4-2A in representative rats from ethanol and control groups. Although there was a trend toward enhancement of the test pulse EPSP amplitude at short IPIs in ethanoltreated rats, the ANOVA was not statistically significant (F(1,44) =
1.30, p > .2). These effects could not be related to differences in the conditioning pulse current levels across groups. The standardization procedure designed to produce equivalent baseline responses across animals resulted in baseline EPSP (Group E = 1.23 0.17 mv; Group C =
Figure 4-2. Paired-pulse potentiation in CA1 after chronic
ethanol treatment. A. Recordings from individual alcohol
and control rats at varying interpulse intervals (IPI).
For each rat, responses to the conditioning stimulus are superimposed. Calibration: 1.0 my, 5.0 msec. B. Group
comparisions (mean SEM) of PS and EPSP amplitude as a
function of IPI. Note the different scales for EPSP and PS. Asterisks denote statistically significant (p < .05)
group differences as indicated by student t-tests.
68
S Cond. Test
Pulse Pulse
I
IP I 20 40 60 80 100 125 150 175
z 0
_J 0 I 0
._j
B.
400
* EPSP
0 PS Alcohol
; o EPSP T
0 0 PS Control x 300
C .0
0
0
W 200 200 x<
C 1 100 -150
Q.
I..
w
, ,i 0
S 20 40 60 80 00 120 40 160 180 0
CONDITION-TEST INTERVAL(msec.)
69
1.39 0.13 my) and PS (Group E = 1.59 0.41 my; Group C = 1.56 0.16 my) amplitudes which did not differ across groups.
Frequency potentiation. As was the case with PPP, chronic ethanol treatment failed to alter the frequency-dependent pattern of responses during FP. At 1 Hz, potentiation of the PS rapidly grew to asymptotic values (within four stimuli) and remained at a stable level (400% of control) thereafter. Both 5 and 10 Hz stimulation produced more robust levels of potentiation (5 Hz > 10 Hz) including the development of multiple population spikes. However, FP at these frequencies was more unstable, exhibiting a definite waxing and waning which was clearly more related to the number of stimuli presented than to the frequency of stimulation (Fig. 4-3). Since both 5 and 10 Hz tetani lead to the development of multiple population spikes, separate analyses were conducted on either the amplitude of the first spike alone or on the grand sum of all the spikes. Identical results were obtained with each measure. Only the data for the combined PS measures will be presented.
Figure 4-3 compares 1 Hz, 5 Hz and 10 Hz FP across groups at identical numbers of stimuli (25 pulses). Chronic ethanol treatment did not influence FP at 1 Hz stimulation. However, a 2-way ANOVA revealed that FP was enhanced in ethanol-treated rats at 10 Hz (F(1,44) = 4.33, p < .05). Stimulation at 5 Hz produced intermediate results although the effect at this frequency did not reach statistical significance. Thus, chronic ethanol treatment augments FP at higher frequencies sampled over a relatively narrow range.
No measures other than the amplitude of the PS were significantly influenced by ethanol treatment. Neither the latency to EPSP onset or to PS peak was affected. Since EPSP amplitude was measured at a fixed
Figure 4-3. Frequency potentiation in CA1 comparing (mean SEM)
alcohol and control groups as a function of stimulus frequency (1, 5 and 10 Hz). A statistically significant treatment effect was observed only at 10 Hz. Intermediate effects
were observed at 5 Hz, whereas no group differences were
noted at 1 Hz.
71
I Hz
400
1
200
* Alcohol
0 Control
L I 4 7 IO 13 16 19 22 25
2
J 800 w 5 Hz V)
m
S600
O
400
u
CL a.
0
400
z
L)
0
Ld a
n
L 200 I-
0- I 4 7 10 13 16 19 22 25
600S 10 HZ T
400
200
4 7 10 13 16 19 22 25
STIMULUS NUMBER WITHIN TETANUS
72
latency (0.75 msec) from onset, this measure was often contaminated by the PS. Especially during FP, the measures of EPSP amplitude and PS amplitude were often inversely correlated. Frequency potentiation produced a dramatic reduction in PS latency, even to the point of obliterating the EPSP completely. We therefore abandoned the EPSP amplitude measure in both the FP and LTP series.
Long-term potentiation. We investigated the long-term effects of repetitive stimulation by presenting six separate tetani which were systematically varied by frequency and duration. The long-term effects of each of the stimulus trains on PS amplitude are presented in Figures 4-4 and 4-5. The pattern of long-term effects was dependent on both the frequency and duration of the tetanus. Stimulation at 1 Hz failed to produce LTP. Rather, when of sufficient duration (25 sec), 1 Hz stimulation produced a brief (1-2 min) depression of PS amplitude immediately following the tetanus (Fig. 4-4). Stimulation at 5, 10 and 100 Hz (Fig. 4-5) produced LTP. These results agree reasonably well with the parametric in vitro study of Dunwiddie and Lynch (1978). Stimulation at 5, 10 (5 sec) and 100 Hz produced characteristic LTP of the PS which reached a peak at five minutes posttetanus. In contrast, 10 Hz (2.5 sec) produced a robust posttetanic potentiation but showed little evidence of LTP 15 minutes following tetanus (Fig. 4-5). These data suggest that a rather complicated interaction may exist between the frequency and duration of the tetanus in producing LTP in CA1 of the rat hippocampus. Under the conditions of this experiment, considerable decay of LTP was observed by 30 minutes following even the 100 Hz tetanus.
Figure 4-4. Posttetanic depression of PS amplitude after low
frequency stimulation. Response depression was observed
only after 25 sec of 1 Hz stimulation. Ethanol treatment
significantly reduced the magnitude of this posttetanic depression. Asterisks denote statistically significant
group differences (p < .05).
74
200
I Hz 5 sec.
S150
I) 00 ...
C
CD
b. 50 Alcohol %,- 0 Control
0
w 0,
o
S20 40 60 2 5 10 15 I
-J I Hz 25 sec. a
< 100
50
* Alcohol 0 Control
0I I I I
20 40 60 2 5 I10 15
Sec. M in.
TIME POST-TETANUS
Figure 4-5. Long-term potentiation in alcohol and control groups compared (mean SEM)
as a function of the frequency and duration of the tetanus. When the duration of the tetanus was varied at a constant frequency (10 Hz), strikingly different patterns of potentiation were observed in both groups. Ethanol treatment failed to
significantly affect long-term potentiation at the frequencies tested.
10 Hz 2.5 sec. 10 Hz 5sec.
300 -* Alcohol00 O Control
200 -200z
100 100
2 O ~a l_ s u s . O111 ; I
M
(n
D 0 I0
S20 40 60 2 5 0 15 20 40 6 22 5 10 15
sec. min. sec. min. .
600 5 Hz 5 sec. 600 100 Hz 10 sec. 13.
U
0
p. 500 500
z
W
S400 400
D
300 300
S200 200 (L. 200
100- 100O0 IOO.. . . . .,
20 40 60 2 5 15 30 20 40 60 2 5 15 30
sec. min. sec. man.
TIME POST-TETANUS
77
Chronic ethanol treatment also produced effects which were dependent on the frequency and duration of stimulation. The transient depression of the PS induced by 1 Hz stimulation (25 sec) was significantly reduced in ethanol-treated animals (F(1,44) = 4.33, p < .05). Subsequent t-tests revealed that this effect was present only at time points immediately following the tetanus (20-50 sec). Ethanol treatment failed to alter the pattern of LTP produced by 5, 10 or 100 Hz stimulation. Analyses of variance were not statistically significant for any of the results shown in Figure 4-5. However, at 100 Hz stimulation, chronic ethanol treatment, while failing to alter the development of LTP of the PS, nevertheless appeared to produce a trend toward a reduction in the magnitude of LTP starting approximately five minutes posttetanus (Fig. 4-5).
Discussion
We have used a nutritionally controlled liquid diet preparation to produce an animal model of chronic ethanol exposure. At comparable durations of exposure, chronic ethanol results in a 15-20% loss of granule and pyramidal cells of the rat hippocampal formation (Walker et al., 1980) as well as dendritic atrophy and reduction of dendritic spines of CA1 pyramidal cells and dentate granule cells of the mouse hippocampus (Riley and Walker, 1978). Taken together, these results suggest that chronic ethanol exposure induces a progressive morphological deterioration of neurons in the hippocampus ultimately ending in cell death. Despite these findings, 20 weeks of ethanol exposure failed to produce significant alterations in basic waveforms, EPSP and PS thresholds, I/0 functions or the production of LTP of CA3-derived fibers
78
terminating in stratum radiatum of CAl. The predominant effect of chronic ethanol treatment was an enhancement of the potentiation of PS responses to paired stimuli or repetitive stimulation at 5 and 10 Hz in the absence of changes in the synaptic response (EPSP) when measurable. Even though a trend toward enhancement of EPSP potentiation was observed during PPP (Fig. 4-2B), this enhancement was only apparent at short IPIs. Both ethanol and control groups reached a similar asymptote at longer IPIs where ethanol treatment produced its greatest enhancement of PS amplitude.
The existence of alterations of potentiation of the PS without
concomitant changes in the EPSP suggests that ethanol did not act primarily on the SCH/COM-CAI synapses themselves. Rather, ethanol treatment appeared to enhance the ability of CAl pyramidal cells to fire synchronously in response to repetitive stimulation. While a number of hypotheses could account for these results, recent evidence indicates that recurrent inhibition plays a significant role in paired-pulse facilitation of the PS. Experimental treatments which directly reduce recurrent inhibition in the hippocampus (anoxia, picrotoxin and enkephalins) enhance the facilitation of PS amplitude to paired stimuli and in some cases directly facilitate PS responses to single shock stimuli (Andersen, 1960; Dunwiddie et al., 1980; Lee et al., 1980) without significantly altering the synaptic response. Thus, under ordinary conditions recurrent inhibition antagonizes paired-pulse facilitation within the hippocampus. This evidence, coupled with the present results, supports the view that chronic ethanol treatment reduces the effectiveness of recurrent feedback inhibition of pyramidal cells of CAl. Facilitation of the PS amplitude to paired stimuli was enhanced in
79
ethanol-treated rats in the absence of significant effects on potentiation of the synaptic response. Further, chronic ethanol treatment enhanced FP of the PS only at those frequencies (5 and 10 Hz) in which recurrent inhibition would exert an influence. Spike responses to 1 Hz stimulus trains were identical across groups. Finally, ethanol treatment significantly reduced the posttetanic depression produced by 1 Hz stimulation (Fig. 4-4). Although ethanol treatment failed to influence the more brief depression produced by higher frequency stimulation (Fig. 4-5), it is noteworthy that such depression may have causes unrelated to recurrent inhibition, for example, transmitter depletion (Yamamoto et al., 1980).
Recurrent inhibition in the hippocampus is mediated by y-aminobutyric acidergic (GABAergic) neurons which receive axon collaterals and project back onto the parent pyramidal and granule cells. These interneurons are believed to be basket cells (Andersen et al., 1964a; Andersen et al., 1964b; Lorente de No, 1934). Since relatively few basket cells exist in the hippocampus (relative to principle cell types), the powerful and widespread recurrent inhibition in the hippocampus must reflect a great divergence of synaptic contacts. Even a modest alteration in basket cells would be expected to produce effects which are more potent than the loss of principle cell types (Walker et al., 1980). Thus, ethanol treatment could reduce recurrent inhibition by reducing the population of basket cells or otherwise altering GABAergic neurotransmission. Chronic ethanol treatment has been reported to decrease GABA concentrations by some (Patel and Lal, 1973; Volicer et al., 1977), but not all investigators (Sutton and Simmonds, 1973), and to reduce the density of low affinity GABA binding sites (Lilijequist and Engel, 1979;
80
Ticku, 1980; Ticku and Burch, 1980). Moreover, chronic ethanol treatment decreases the affinity and density of receptor binding for benzodiazepines (Freund, 1980), which appear to augment GABAergic inhibition in widespread brain areas. However, in many of these studies the duration of ethanol treatment was substantially shorter (2-3 weeks) than in the present study and GABAergic function was assessed within 24 hours of ethanol withdrawal. There is little doubt that these effects are dependent upon the duration of ethanol exposure. For example, while 15-19 days of ethanol exposure failed to alter benzodiazepine receptor binding (Freund, 1980; Karobath et al., 1980), seven months of exposure decreased the density of benzodiazepine receptors for at least one month following abstinence (Freund, 1980). Thus, it remains unclear whether longer durations of ethanol exposure would decrease GABAergic neurotransmission after eight weeks of ethanol abstinence, the time period assayed in this study.
While reduced recurrent inhibition may be best explained by a
direct action of chronic ethanol treatment on basket cells, at least one alternative should be considered. Anesthetic agents including ethanol have been reported to augment recurrent inhibition of single unit activity in the hippocampus (Newlin et al., 1979; Tsuchiya and Fukushima, 1978; Wolf and Haas, 1977). Since ethanol exhibits crosstolerance to other depressants (Kalant et al., 1971), the apparent reduction of recurrent inhibition could result from differential group responses to general anesthesia. This hypothesis must be seriously considered since chronic ethanol treatment has recently been shown to produce tolerance of the response to ethanol of SCH/COM-CAl synapses in a hippocampal slice preparation (Carlen and Corrigall, 1980). However,
81
we consider this alternative unlikely for the following reasons. Neither the initial nor supplemental doses of urethane required to produce and maintain anesthetic levels differed between groups. Behavioral tolerance (Kalant et al., 1971) to ethanol normally dissipates over a brief time course (2-3 days) and ethanol tolerance in hippocampal slices was only observed immediately upon ethanol withdrawal (Carlen and Corrigall, 1980). Since our data were collected at least eight weeks after ethanol withdrawal, it is likely that the residual effects of ethanol tolerance would have dissipated. This possibility, while still unlikely, cannot be completely ruled out (cf. Begleiter et al., 1980).
The effects of chronic ethanol treatment on PPP, FP and posttetanic depression are significant insofar as they persisted for at least eight weeks after ethanol abstinence. Nevertheless, the failure to observe significant alterations in basic synaptic responses (thresholds, I/0 curves) is surprising since both the pre- and postsynaptic elements of the SCH/COM-CAl path are reduced by chronic ethanol treatment (Walker et al., 1980). However, it is unlikely that a simple relationship exists between the number of CA1 pyramidal cells and the size of the extracellular EPSP and PS. Further, it is plausible that the varied effects of ethanol treatment could interact. For example, the basket cells probably tonically inhibit the pyramidal cells (Alger and Nicoll, 1980). Picrotoxin, a GABA receptor blocker, has also been shown to facilitate PS responses to single shock stimuli (Dunwiddie et al., 1980). Thus, it is possible that a loss of recurrent inhibition may have masked a reduction in PS responses (I/0 curves) produced by deafferentation of CAl. Alternatively, the failure to observe changes in
82
basal synaptic response strength may reflect morphological reorganization and recovery, since reactive synaptogenesis has been documented in CA1 following destruction of afferents originating in CA3 (Nadler et al., 1980a; Nadler et al., 1980b).
At present, a connection between LTP in the hippocampus and normal memory formation is clearly speculative. However, the issue may be addressed through correlative analysis of synaptic potentiation in memory deficient animals. Studies in aged, memory deficient rats have indicated that LTP is reduced in both CA1 (Landfield et al., 1978) and the dentate gyrus (Barnes, 1979) of the hippocampus. The similarity of the mnemonic deficits associated with alcoholism and aging have led to the hypothesis that alcohol accelerates the aging process (Beck et al., 1979; Ryan and Butters, 1980). Studies of chronic ethanoltreated and aged animals indicate a reasonable similarity in the nature of morphological deterioration in the hippocampus (Bondareff, 1979; Brizzee and Ordy, 1979; Scheibel, 1979). However, aged animals exhibit a reduction in FP and LTP in SCH/COM-CAl connections (Landfield et al., 1978) whereas ethanol-treated rats exhibit enhanced PPP and FP. Insofar as the hippocampus is involved in memory formation (Milner et al., 1968; O'Keefe and Nadel, 1978), these results suggest that very different mechanisms may underlie the mnemonic deficits associated with alcoholism and aging. This conclusion must be considered preliminary, particularly since a trend toward a greater decay of LTP was observed in ethanoltreated rats (Fig. 4-5), and a recent study with hippocampal slices has reported an impairment of LTP formation after chronic ethanol treatment assayed immediately following ethanol withdrawal (Durand et al., 1980).
CHAPTER V
ELECTROPHYSIOLOGICAL ANALYSIS OF CHRONIC ETHANOL NEUROTOXICITY
IN THE DENTATE GYRUS: ENTORHINAL AFFERENTS
Introduction
Long-term consumption of ethanol leads to numerous neuropsychological impairments in both humans and laboratory animals. Perhaps the hallmark of the cognitive dysfunction is a profound impairment in the subject's ability to acquire and store new information (Talland, 1965; Victor et al., 1971; Walker et al., 1981). The debilitating nature of the alcoholic syndrome has prompted many researchers to investigate the loci and nature of ethanol-related neurotoxicity in the brain. The neuropathology in human Korsakoff patients centers around the dorsomedial nucleus of the thalamus and the mammillary bodies, with other brain regions more variably affected (Mair et al., 1979; Victor et al., 1971). However, studies in laboratory animals, through careful control of nutrition, genetics, amount and duration of ethanol consumption and other environmental variables, should be better able to localize those brain regions particularly sensitive to ethanol-induced neurotoxicity.
Using the liquid diet procedure of Walker and Freund (1971), our laboratory has demonstrated considerable neuropathology in ethanolconsuming rats and mice. Cerebellar vermal Purkinje and granule cells are reduced in number by 15-20% following five months of ethanol treatment (Walker et al., 1981). In addition, surviving Purkinje cells show a 25-30% reduction in dendritic arborization (Riley and Walker, 1981).
83
84
Importantly for this study, similar pathology has been demonstrated in the hippocampus. There is approximately a 15% loss of both hippocampal pyramidal and granule cells (Walker et al., 1980) in addition to a reduction in spine density along both pyramidal basilar dendrites and granule cell dendrites (Riley and Walker, 1978). It should be noted that the histological analyses were performed at least one month following ethanol withdrawal, implying that these are lasting changes in CNS morphology.
The above histological data suggest that chronic ethanol treatment produces deleterious effects on hippocampal cytoarchitecture. It is imperative thatwe correlate these findings with electrophysiological analyses of synaptic and cellular activity if we are to understand how chronic ethanol exposure affects hippocampal function. Previously (Chapter III), we electrophysiologically confirmed a shrinkage of the SCH/COM terminal field in CA1 of ethanol-treated rats (20 weeks on ethanol, eight weeks off). Although these afferents still responded normally to single-shock stimulation, they did show evidence of an enhanced pairedpulse and frequency potentiation of the extracellularly recorded PS (Chapter IV). No changes in LTP were observed. The present study assessed the ubiquity of these alterations by analyzing the functional integrity of the angular bundle-DG synaptic connections. The angular bundle (AB) fibers originate in the medial and lateral entorhinal cortices and distribute topographically onto the outer 2/3 of the granule cell dendrites of both dorsal and ventral blades (McNaughton and Barnes, 1977; Steward and Scoville, 1976). Excitatory synapses are formed by boutons en passage. As in CA1, intra- and extracellular recordings have shown that these synapses can exhibit a wide range of potentiation
85
phenomena, including PPP, FP and LTP (Bliss and Gardner-Medwin, 1973; Bliss and Lomo, 1973; Lomo, 1971b; White et al., 1979), in response to repetitive activation. The present study examines with extracellular recording techniques the effects of ethanol on the laminar distribution of the AB-DG synapses as well as their responsivity to single and repetitive stimuli.
Methods
The animals, treatment conditions and general electrophysiological methodology have been described earlier in Chapter II. As before, rats were maintained on liquid diets for 20 weeks and withdrawn for eight weeks prior to electrophysiological study. The major procedural alteration was that, since Groups S and LC have rarely differed on a variety of histological, behavioral and electrophysiological measures in numerous experiments, Group LC was dropped from this experiment. Certain of the physiological characteristics of the AB-DG synapses differ from those seen in CA1; therefore, the experimental protocol of the earlier experiments required modification.
The electrode placements were based on the following stereotaxic
coordinates: recording electrode (tip broken back to a 3-5 pm diameter),
3.8 mm posterior to bregma, 2.4 mm lateral; stimulating electrode, 8.1 mm posterior to bregma, 5.0 mm lateral. The recording electrode was initially positioned in the dentate hilus and used to guide the stimulating electrode placement by monitoring the PS amplitude and threshold. The recording electrode was then moved to a fixed (except during the laminar analysis) recording site 150 pm ventral to the dorsal blade inversion. At the end of a recording session, lesions were made through
86
the recording electrode at the inversion points of both the dorsal and ventral blades, approximately 650 pm apart.
In this experiment, the I/O curve (10-1000 pa) was taken prior to the laminar analysis. During the laminar analysis, the recording electrode was stepped in 25 pm increments completely across both blades of the DG. Eight responses at 0.1 Hz were recorded for off-line averaging and analysis by a PDP 11/40 laboratory computer. The electrode was then returned to the fixed recording site. At this point, PPP (15-175 msec IPI) was evaluated at both low (EPSP amplitude 25% of that at PS threshold) and high (PS amplitude 50% of asymptote) stimulus intensities. Stimulus pairs were given at 0.03 Hz. Frequency potentiation was examined with the following tetani: 1 Hz/25 sec, 5 Hz/5 sec and 10 Hz/5 sec at a current that produced a PS 50% of asymptote. Posttetanus test stimuli were given only for 10 min following each tetanus since these frequencies and durations of stimulation did not produce LTP in the DG. To produce LTP, 10 stimulus trains (400 Hz, 20 msec duration) were applied at a frequency of one train per minute. This procedure has been shown to produce robust LTP in the DG (Douglas, 1977). Test pulses were given 25 and 50 sec following each tetanus and the responses averaged to provide an indication of the development of LTP. Following the tenth train, test pulses were given (0.05-0.03 Hz) for the next 30 min.
Results
This experiment employed a total of 17 rats which were distributed among the groups as follows: Group E = 9; Group S = 8. As mentioned previously, a lab chow control group was not included in the study. Group E rats consumed a mean daily ethanol dosage of 13.11 g/kg, which
87
is comparable to the consumption observed not only in the previous electrophysiological studies but also in the studies demonstrating either associative deficits or neuronal loss in the hippocampus. The ethanol rats showed normal weight gain throughout the treatment period and did not differ from sucrose rats in weight at the time of electrophysiological study: Group E = 525.2 34.5 g; Group S = 497.1 18.9 g.
Histological analysis. As in the previous experiments (Chapters III and IV),there was some concern that the ethanol treatment may have produced sufficient changes in brain volume that the use of fixed stereotaxic coordinates would have resulted in differential electrode placement between the groups. To check for this possibility, both the recording and stimulating electrode sites were plotted with respect to both the septal pole of the hippocampus and midline. This analysis revealed that there were no group differences in gross placement of the electrodes. Stimulation electrode tips were localized to the AB, just posterior to the hippocampus and approximately 4 mm posterior to the recording sites. The recording electrode lesions were found in the dorsolateral DG, approximately 50-70 lm (25-35% of the total stratum moleculare) distal to the granule cell layer of each blade (Fig. 5-1).
Detailed measures of the major DG laminae were made along the actual electrode trajectory. The use of two recording electrode lesions spaced a known distance apart (based on microdrive readings) allowed us to correct for the tissue shrinkage that normally occurs during histological fixation and celloidin embedding. As in CA1 (Chapter III), there were no differences between the groups in laminae widths except in the dorsal blade where Group E actually showed a significantly larger (t =
4.57, p < .01) molecular layer (Table 5-1). For Group S only, the
Figure 5-1. Photomicrographs of the dorsal hippocampal dentate
gyrus (DG) where recordings for the present experiment were
made. Arrows indicate the lesions that were made through
the recording electrode at the inversion point for the dorsal blade (top half of figure) and the ventral blade (bottom half of figure) as produced by AB stimulation. Note that in each blade the inversion point is in the proximal
portion of the molecular layer. The sections are 30 pm
thick and stained with Weil myelin stain.
89
L
90
TABLE 5-1
Chronic Ethanol Effects on Widths (pm) of DG Laminaea
Lamina Sucrose Ethanol A A% DSM width 185.7 3.0 228.4 8.1 42.7 23.0b DSG width 55.0 4.4 69.4 6.3 14.4 26.2 HIL width 409.8 19.6 361.3 36.1 -48.5 -11.8 VSG width 62.9 5.5 65.8 5.1 2.9 4.6 VSM width 228.4 15.4 255.6 18.5 27.2 11.9 bMean SEM
p < .01 student's t-test
Abbreviations: DSM, dorsal stratum molecular; DSG, dorsal stratum granulosum; HIL, hilus; VSG, ventral stratum granulosum; VSM, ventral stratum moleculare
|
PAGE 1
AN ELECTROPHYSIOLOGICAL ANALYSIS OF CHRONIC ETHANOL EFFECTS ON SYNAPTIC DISTRIBUTION AND FUNCTION IN RAT HIPPOCAMPUS BY WICKLIFFE C. ABRAHAM A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1981
PAGE 2
This dissertation is dedicated to Stuart B. and Ida Jeanne Dagger Abraham.
PAGE 3
Young man, take care lest you find what you are looking for. Granit
PAGE 4
ACKNOWLEDGEMENTS Once in a long while, a friendship is formed that interweaves personalities in a unique way. Many thanks to Jim for keeping me out of school my first year and keeping me in school for the remaining years. Friendships are multifaceted but in part serve to resolve the past, enliven the present and give hope to the future. Those instrumental in enlivening my graduate years include Brian, Denise, Dick, Lia, Lynn, Mary Margaret and FVC. The love and support from Elise has been sustaining and instructive. My supervisory committee--Chuck Vierck, Don Walker, Floyd Thompson and Keith Berg--has been thoughtful, helpful and supportive. Two in particular stand out as guiding lights and resourceful friends -Steve Zornetzer and Bruce Hunter. Thank you, Steve and Bruce. Paul Manis has made crucial contributions to the data analysis in these experiments. Thanks also to Bill Brownell for use of his computer facilities. My research has relied greatly on the technical sup port of Dot Robinson, Larry Ezell, Pat Burnett and NERDC. I have been supported financially by an NSF predoctoral fellowship and an NIAAA predoctoral fellowship. iv
PAGE 5
ACKNOWLEDGEMENTS INDEX OF ABBREVIATIONS ABSTRACT. CHAPTER I. BACKGROUND. TABLE OF CONTENTS Alcohol-related Amnesia: Clinical Studies ..... Hippocampal Dysfunction and Amnesia: Clinical Studies Alcohol-related vs. Hippocampus-related Amnesia: Laboratory Studies ........... Hippocampal Anatomy and Physiology ..... Animal Models of Chronic Alcohol Consumption .. Rationale . . ... CHAPTER I I. GENERAL METHODS Ethanol Administration .... Electrophysiological Methods. Experi mental Protocol . Data Analysis ........ CHAPTER III. CHRONIC ETHANOL EXPOSURE AND SYNAPTIC DISTRIBUTION IN CAl OF RAT HI PPOCAMPUS: CURRENT-SOURCE DENSITY PAGE iv vii viii 1 2 5 9 13 25 26 28 28 29 30 32 ANALYSIS 35 Introduction. Methods Results . Discussion. CHAPTER IV. AUGMENTATION OF SHORT-TERM PLASTICITY IN CAl OF 35 36 37 50 RAT HIPPOCAMPUS AFTER CHRONIC ETHANOL TREATMENT. 59 Introduction. 59 Methods 60 Results 62 Discussion. 77 CHAPTER V. ELECTROPHYSIOLOGICAL ANALYSIS OF CHRONIC ETHANOL NEUROTOXICITY IN THE DENTATE GYRUS: ENTORHINAL AFFERENTS. . . . . 83 V
PAGE 6
Introduction. Methods ..... Results ...... Discussion .. CHAPTER VI. GENERAL DISCUSSION .. APPENDIX. CURRENT-SOURCE DENSITY ANALYSIS. REFERENCES ..... BIOGRAPHICAL SKETCH vi PAGE 83 85 86 113 121 129 132 145
PAGE 7
INDEX OF ABBREVIATIONS AB--angular bundle AH--Ammon's horn ALV--alveus ANOVA--analysis of variance CAl-4--hippocampal subfields CNS--central nervous system COM--commissural CSD--current-source density DG--dentate gyrus DRL--differential reinforcement of low rate responding EP--evoked potential EPSP--excitatory postsynaptic potential FP--frequency potentiation GABA--y-aminobutyric acid HF--hippocampal fissure HIL--dentate hilus vii I/0--input/output IPI--interpulse interval KF--Korsakoff LTM--long-term memory LTP--long-term potentiation PS--population spike PPP--paired-pulse potentiation PTP--posttetanic potentiation SCH--Schaffer collaterals SG--stratum granulosum SM--stratum moleculare SO--stratum oriens SP--stratum pyramidale SR--stratum radiatum STM--short-term memory TBR--to-be-remembered
PAGE 8
Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulf i llment of the Requirements for the Degree of Doctor of Philosophy AN ELECTROPHYSIOLOGICAL ANALYSIS OF CHRONIC ETHANOL EFFECTS ON SYNAPTIC DISTRIBUTION AND FUNCTION IN RAT HIPPOCAMPUS By Wickliffe C. Abraham June, 1981 Chairman: Steven F. Zornetzer Major Department: Neuroscience Chronic alcoholism in man results in a syndrome marked by a variety of neuropsycho1ogica1 and neurophysio1ogica1 abnormalities. In advanced stages, the central impairment is a severe anterograde amnesia. The hippocampus, long thought to be involved in memory consolidation, could be a major target of alcohol neurotoxicity. Recent studies using a rodent model of chronic ethanol consumption have described not only a number of cognitive impairments but also neuropathology in the hippo campus, including cell loss and decreased spine density. The present experiments were conducted to provide electrophys i ological correlates of the morphological data. Rats consumed an ethanol-containing liquid d iet for 20 weeks. Controls were either pair-fed a similar diet with sucrose isocalorically substituted for ethanol or received standard laboratory chow. Eight weeks after ethanol withdrawal, the rats were prepared for electrophysiological field potential recordings in the hippocampus using standard extracellular recording techniques. viii
PAGE 9
The first two experiments examined the synaptic distribution and function of CA3 afferents to CAl. Evoked potential profiles, orthogo nal to the layering of afferent terminals, were obtained and analyzed with current-source density techniques. The CA3 afferents to stratum radiatum of CAl exhibited a bimodal distribution of current density, possibly reflecting a separation of the ipsilateral and contralateral CA3 terminals. Ethanol treatment produced an overall shrinkage of the synaptic zone independent of tissue volume changes. The shrinkage was specific to the hypothesized commissural input, proximal to the cell layer. The hypothesized ipsilateral fibers showed evidence indicative of a compensatory enhancement of synaptic efficacy. The CA3-CA1 synapses exhibited normal input/output relationships for all measures in ethanol-treated rats. Furthermore, ethanol had no effects on long-term potentiation. However, ethanol treatment did lead to an augmentation of short-term plasticity. The population spike was differentially enhanced in both dual-pulse and repetitive stimulation paradigms. These findings may be due to an ethanol-related reduction in the efficacy of recurrent inhibition. The entorhinal cortex-dentate gyrus synapses were similarly studied. Laminar analyses through the dentate dorsal and ventral blades again revealed in ethanol-treated animals a shrinkage of the synaptic fields. In both blades, the shrinkage was restricted to the outer molecular layer. These results are suggestive of preferential damage to the distal terminals and/or dendrites. In contrast to the CAl findings, shortterm plasticity was refractory to ethanol treatment. However, input/ output analysis revealed that ethanol-treated rats had smaller population spike amplitudes per synaptic potential at asymptotic stimulus currents. ix
PAGE 10
These data probably reflect the loss of granule cells that has been documented previously. Finally, of the two known components of long term potentiation, the ethanol regimen apparently compromised only the potentiation of neural excitability. The experiments presented here have electrophysiologically con firmed the loss of afferent input to CAl subsequent to chronic ethanol treatment. The loss is apparently confined to an as yet unidentified subset of the CA3 afferents. It is evident that the afferent drive to the dentate is also reduced, particularly in distal dendritic regions. Confirmatory evidence for the loss of granule cells is presented The remaining data indicate that chronic ethanol consumption can significantly alter short-term plasticity in CAl and long-term potentiation in the dentate. These alterations may underlie some of the ethanol-related behavioral impairments. The electrophysiological data also compare favorably with similar findings in aged animals, suggesting that there may be a common mode of action. X
PAGE 11
CHAPTER I BACKGROUND Long-term consumption of alcohol leads to a variety of neuropsycho logical impairments in both humans and laboratory animals. One of the more profound and debilitating aspects of the chronic alcohol syndrome is the impairment of memory processing capabilities. Converging lines of evidence suggest that damage to the hippocampus, a central nervous system (CNS) telencephalic structure, may underlie the decline in mnemonic functioning. In spite of some studies directly examining the histology of the hippocampus (see below), much of the inference linking alcohol consumption to hippocampal damage is based on the similarity of the behavioral alterations resulting from chronic alcohol exposure and those produced by surgically-induced hippocampal damage. Clearly, such an inferential approach has many shortcomings. Most importantly, brain structure-function relationships are not so well understood that similarities (or dissimilarities) of behavioral deficits necessarily imply similarities (or dissimilarities) of underlying brain damage. Nonetheless, clues are needed which point to where in the CNS alcohol may be exerting its toxic effects. A correspondence of the behavioral symptoms of the chronic alcohol and hippocampal syndromes would at least give us sufficient reason to suspect that the hippocampus is damaged by chronic alcohol ingestion. An important consideration in the study of alcohol-related dementia is the extent to which alcohol serves as the primary toxic agent. l
PAGE 12
2 Nutritional or metabolic disturbances secondary to chronic ethanol intoxication are now considered to underlie several dementing diseases formerly thought to be due primarily to alcohol toxicity. These include Wernicke syndrome, pellagra, Marchiafava-Bignami disease, hepatic en cephalopathy, and central pontine myelinosis (Victor and Banker, 1978). Although some authors also maintain that Korsakoff syndrome has a nutritional etiology, i.e., thiamin deficiency (Victor and Banker, 1978; Victor et~-, 1971), the issue is not yet resolved. No cases of persistent Korsakoff syndrome have yet been documented to be induced by malnutrition in the absence of concomitant ethanol consumption (Freund, 1973). Furthermore, animal models have shown that profound memory disturbances persist after chronic alcohol consumption in spite of adequate nutrition (Freund, 1970; Walker and Freund, 1973; Walker and Hunter, 1978). Thus, the evidence favors the view that alcohol consumption~ se can directly produce dementia. Alcohol-related Amnesia: Clinical Studies Korsakoff (KF) syndrome is a "pure" amnesia in that memory functions are profoundly disturbed while other cognitive functions of the patients remain relatively intact (Victor et~-, 1971). The symptomatology includes a limited retrograde amnesia, profound anterograde amnesia, apathy, indifference to environmental events and limited insight into the disability. Confabulation is occasionally observed but is probably secondary to the loss of memory for recent events. The deficits are relatively permanent, even after alcohol withdrawal (Brion, 1969; Victor et~-, 1971). Although the term KF psychosis or syndrome has been used to categorize amnesias of a variety of etiologies (Brion,
PAGE 13
3 1969), we will reserve this term exclusively for alcohol-related amnesias. As previously mentioned, the salient presenting sign of KF syndrome is a severe anterograde amnesia. The exact nature of this deficit has been the subject of considerable controversy. Research in this area has focused on 1) whether short-term memory is affected and 2) whether the long-term memory deficits are more specific to the storage or the retrieval of information. The major protagonists involved in the latter issue are Butters and Cermak's group in Boston that supports the encod ing deficit hypothesis and Warrington and Weiskrantz's group in Oxford that supports the retrieval deficit hypothesis. In many respects, short-term memory (STM) is relatively unimpaired in KF patients. In particular, the initial registration of information into STM appears intact since KF patients show normal digit spans, normal recency effects in verbal free recall and normal memory for visual location (Brooks and Baddeley, 1976; Kinsbourne and Wood, 1975). On the other hand, there is considerable evidence that maintenance or retrieval of information from STM may be impaired. When evaluated on the Peterson-Peterson paradigm, KF patients' STM was observed to be acutely susceptible to the disruptive influence of interpolated tasks (performed between stimulus presentation and recall) (Cermak et tl,, 1971). This retroactive interference is modality-independent but more pronounced when the modality is the same for both the distractor and the to-be-remembered (TBR) items (Deluca et tl, 1975; Deluca et tl,, 1976; Strauss and Butler, 1978). The STM of KF patients has also been shown to be extremely susceptible to proactive interference, which becomes evident as these STM tests are repeated (Cermak et .tl_., 1976; Parkinson, 1979).
PAGE 14
4 The next class of experiments were generated by the debate concern ing storage~ retrieval hypotheses of KF-related long-term memory (LTM) deficits. The conventional mid-1960s view of KF-related anterograde amnesia was that the amnesics were unable to store new experiences into LTM. Retrieval processes~ se appeared intact since the patients showed only a limited retrograde amnesia. The Oxford group has forced a revision of this early notion. These researchers have been able to demonstrate successful memory consolidation in KF patients, particularly when employing perceptuomotor tasks (Brooks and Baddeley, 1976; Warrington and Weiskrantz, 1970). The Oxford researchers instead focus on enhanced interference at the time of recall as the major element of amnesia. Winocurhas suggested that the inability to distinguish multiple word lists is related to an impaired time-tagging or discrimination of contextual cues (Winocur and Kinsbourne, 1978; Winocur and Weiskrantz, 1976). On the other side of the storage-retrieval controversy is the Boston group which concentrates on encoding deficits in the KF syndrome. The emphasis of their work focuses on the failure of patients to spon taneously semantically process TBR items. Much of the support for this thesis comes from analysis of KF patients' error patterns; i.e., KF patients are not afflicted by interference in the semantic dimension, presumably due to their failure to process information semantically in the first place (Cermak and Moreines, 1976; Cermak et~., 1973). Forced use of semantic processing can improve the patients' performance to a limited extent (Cermak and Reale, 1978). Additionally, cues given at the time of recall to facilitate KF patients' performance are notoriously ineffective. Only when the cues are also available at the time
PAGE 15
5 of encoding does the whole cuing procedure give positive results (Cermak et ~-, 1980). Given the above findings, the storage~retrieval controversy of KF amnesia may no longer be an issue. The fact that learning and memory can be demonstrated in a variety of situations indicates that KF patients can consolidate if interference does not act so rapidly as to eliminate the TBR items from STM. The amnesia seems more precisely associated with an encoding difficulty. This difficulty will be manifested at the time of retrieval when previously learned material interferes with extraction of desired items from the memory banks. Note that retrieval se is operative since KF patients make errors of commission, not errors of omission. Thus, both storage and retrieval are still functional but the 11mnemons11 are no longer distinctive. As ~Jinocur suggests, this may reflect inadequate evaluation of context or perhaps an inability to distinguish items in the temporal dimension. Hippocampal Dysfunction and Amnesia: Clinical Studies Let us now consider the human hippocampal amnesic syndrome so that we may later compare this syndrome with KF psychosis. The widespread interest in the hippocampus and memory processing originated with the medial temporal lobe resections in man performed by Scoville and his colleagues (Penfield and Milner, 1958; Scoville, 1954; Scoville and Milner, 1956). On the surface, the amnesia resulting from these surgical procedures was very similar to that seen in KF psychosis, i.e., a profound anterograde amnesia coupled with a relatively intact STM. One patient in particular (H.M.) has been the focus of attention and his memory disorder has been well characterized. We will discuss
PAGE 16
6 the case of H.M. in detail since this case is considered to represent the classic amnesic syndrome and serves as a standard against which other clinical syndromes as well as animal studies are compared. In 1953, H.M. underwent surgery to alleviate his debilitating epileptic attacks. His resection was one of the most radical performed by Scoville, extending 8 cm posterior to the temporal pole and involving the anterior 2/3 of the hippocampus bilaterally. Immediately following recovery from surgery, H.M. exhibited a severe anterograde amnesia although no other neurological deficits were evident. Rigorous analyses confirmed many of the early anecdotal observations concerning H.M.'s memory disorder. Such cognitive functions as concept formation and instruction-following were intact as were his dealings with spatial relationships and spatial orientation. His STM was within normal range when tested by forward digit span, backward digit span and Seashore Tonal Memory tests (Milner et tl, 1968). His STM decay curve was not significantly different from that of normal subjects (Wickelgren, 1968). On the other hand, H.M. exhibited dramatically poor LTM for both verbal and nonverbal tasks. These tasks included supraspan digit se quences, standard verbal learning tasks, delayed matching-to-sample, delayed recall of a complex geometric design and recognition of recurrent nonsense patterns (reviewed in Milner et tl-, 1968). The patient could learn visually-guided mazes only if the number of choice points was kept close to STM span length. He was severely retarded in acquiring a tactually-guided maze even when the number of choice points was within short-term span. In a manner analogous to KF patients, H.M. could show significant or normal savings on certain tasks. His acquisition and retention of
PAGE 17
7 motor skills was essentially normal (Corkin, 1968). Although he showed retarded learning of very simple mazes, he retained the information reasonably well once learned (Milner et tl-, 1968). In addition, H.M. demonstrated normal acquisition of an incomplete-figure task and significant savings {although less than controls) on an unexpected retest either one hour or four months later {Milner, 1970). It should be noted that cuing at the time of testing was never shown to facilitate H.M.1s performance on any task except a face recognition task. Additionally, his remote memory was excellent. Thus, there is little evidence for a retrieval failure underlying H.M.1s amnesia. How does KF amnesia compare to medial temporal lobe amnesia? The similarities are striking. Profound anterograde amnesia is seen in both verbal and nonverbal domains. Remote memories are essentially intact with any differences explainable by the differences in the time course of lesion development. On the other hand, recent memories are extremely susceptible to the disruptive influences of interpolated tasks. Both kinds of amnesics show good learning on perceptuomotor or partial information tasks. Some differences in symptomatology are also evident. Hippocampal amnesics have normal STM decay constants while KF patients sho\t1 enhanced forgetting under distractor conditions. Korsakoff amnesics show considerable interference effects when retrieving from LTM but this phenomenon has not been demonstrated for patients such as H.M. Unlike KF patients, temporal lobe patients are not aided by cuing procedures during recall. Thus, although these two syndromes do not mirror one another exactly, overall there are enough similarities be tv.,een them to suspect hippocampal damage in the KF patients. The more radical but localize d damage in medial temporal lobe patients may be responsible for what appears to be a more pure amnesic syndrome.
PAGE 18
8 We might expect that histopathological examination of autopsied brain tissue ~muld clarify the extent to which the hippocampus is involved in KF syndrome. Unfortunately, the major published work on this subject does little to clarify the issue (Victor et~., 1971). Wernicke-Korsakoff patients showed discontinuous lesions of specific nuclear groups. The damage was generally bilateral and concentrated in paraventricular nuclei. The hippocampus itself was found to be damaged in only 36% of the cases. However, a major hippocampal efferent target, the medial marnmillary nucleus, was involved in nearly all cases. Despite these findings, Victor et~. (1971) suggested that lesions of the dorso medial nucleus of the thalamus were best correlated with the amnesia. It has recently been suggested that the amnesia may occur only when both the dorsomedial nucleus and the mammillary bodies are damaged (Mair et ~., 1979). It seems that the above pathology data provide little support for the hippocampal hypothesis of KF amnesia. Two points are relevant here. First, many cellular changes could have gone undetected by exclusive use of light microscopy. Dendritic and synaptic alterations as well as biochemical changes are obvious possibilities. Secondly, significant destruction of major inputs and outputs to a structure may produce be havioral changes similar to those produced by lesion of the structure alone. The fact that the effects of necrosis of both the medial mammillary body and the dorsomedial nucleus produce a syndrome comparable to hippocampal lesions alone suggests that such a possibility is quite reasonable. Unfortunately, Victor et~. (1971) did not give a detailed account of the nature of extent of the damage to cerebral cortical areas more closely afferent to the hippocampus such as cingulate, entorhinal
PAGE 19
9 or subicular cortex. We conclude that a resolution of the question of hippocampal involvement in alcohol-related amnesias requires more detailed experimental analyses which are aimed directly at the issue. Alcohol-related vs. Hippocampus-related Amnesia: Laboratory Studies The need to clarify the role of the hippocampus in amnesia has led to a massive attempt to create animal models of the human amnesic syndrome. This section will briefly outline and compare the behavioral effects of hippocampal damage and chronic alcohol consumption in laboratory animals. Despite extensive research efforts, no satisfactory model of the human hippocampal amnesic syndrome has yet been achieved. Lesions of the hippocampus produced specific behavioral alterations but they were not found in all tasks and they were not always performance deficits. A complete review of this voluminous literature would not be appropriate here. There already exist several exhaustive reviews (Douglas, 1967; Horel, 1978; Jarrard, 1973; Kimble, 1968; O'Keefe and Nadel, 1978}. To summarize the findings, the major behavioral changes that follow hippocampal damage include: 1) chance or worse levels of spontaneous alternation, 2) decreased habituation to novel stimuli, 3) poor performance in discrimination reversal tasks, 4) prolonged extinction of discrimination responses, 5} impaired acquisition of mazes based on the use of spatial cues and 6) impaired passive avoidance but facilitated shuttle avoidance performance (Black et~-, 1977; Douglas, 1967; Kimble, 1968; Kimble, 1975). Conspicuously absent from this list is a multimodal deficit in memory consolidation. Instead, the hippocampal syndrome in animals is characterized by a marked performance deficit in situations where
PAGE 20
10 environmental contingencies require the cessation of learned or prepotent behavior and/or the initiation of a different behavior (or strategy). There are two classes of data that are particularly relevant to the human amnesic syndrome and thus deserve special attention. The first is relevant to the suggestion that hippocampal-lesioned animals are particularly deficient in context-dependent retrieval (or encoding) of memories. The first indication of this deficit came from the demonstration that the impairments in both passive avoidance and extinction performance co1TID1only associated with hippocampal damage were alleviated if external goal-box stimuli associated with the change in reinforcement contingencies were detectable by the animal early in the runway (Hinocur and Bindra, 1976). In a manner similar to human amnesics, hippocampal rats could learn a visual discrimination task as well as controls, but were significantly impaired if a high interference task involving similar stimulus materials was interposed between training and testing (Winocur, 1979). Finally, Winocur has demonstrated that reversal-learning deficits can be alleviated when contextual cues dissociate the different reward contingencies (Winocur and Olds, 1978). These and other data {Hirsch et~-, 1978) not only support a context hypothesis of hippocampal functions but also nicely interface with data from human amnesics (see above review of KF amnesia). The other set of data relevant to the human literature arises from tasks employing a delay between stimulus and response, such as delayed response, delayed alternation, go-no-go discrimination and differential reinforcement of low rate responding (DRL) tasks (Isseroff, 1979; Iversen, 1976; 01Keefe and Nadel, 1978). Interestingly, in the situations reasonably devoid of spatial information (go-no-go temporal
PAGE 21
11 alternations and DRL), hippocampal animals show increasing deficits with increasing times between trials (Halker and Means, 1973; vJalker et tl, 1972). Furthermore, a lesioned animal's performance is severely impaired at short intervals if interfering tasks are interposed (Walker and Means, 1973). These data show a remarkable similarity to those seen in the human amnesic literature. Animal models of chronic alcohol consumption have also been de veloped. The pattern of behavioral deficits associated with these models is highly reminiscent of the performance of hippocampal-lesioned animals. For example, a numje r of studies reported that chronic alcohol ingestion causes impairments in avoidance acquisition long after ethanol withdrawal. Mice placed on a special, ethanol-containing liquid diet for four months were deficient in acquisition and retention of an inhibitory avoidance task (Freund, 1974). Memory for passive avoidance training could also be disrupted by posttraining ethanol consumption. One-way avoidance by hamsters was unaffected by consumption of ethanol in aqueous solution (Harris et~., 1979b). Ethanol-consuming rats are deficient when performing several other tasks including ones appetitively motivated. Six months of ethanol consumption by rats led to deficits in acquisition of Hebb-Williams closed field mazes and in acquisition and retention of an aversively motivated moving belt task (Fehr et~., 1976). Timing behavior, DRL and temporal shock discrimination were also disrupted by chronic ethanol ingestion (Denoble and Begleiter, 1979; Smith et~-, 1979; Walker and Freund, 1973). Preliminary experiments in our laboratory indicate that alcoholic rats are deficient in spontaneous alternation. Finally, temporal single-alternation, go-no-go behavior of alcoholic rats is
PAGE 22
12 almost identical to that of rats with hippocampal damage. These rats are impaired by either long intertrial intervals or short intertrial intervals filled by an interfering task. Chronic ethanol consumption by rats or mice produces one deficit not produced by hippocampal lesions--impaired shuttle avoidance. Hippocampal lesions normally facilitate this behavior. Nonetheless, this task has been employed for rigorous analyses of ethanol-related per formance deficits. Time gradients of consumption indicate that 3, 5, 7 and 9 months but not 1.5 months of ethanol consumption produced the behavioral deficits. These deficits persisted as long as has been tested postwithdrawal, up to 4.5 months (Freund, 1979; Freund and Walker, 1971b). In add: i on, chronic phenobarbital consumption does not produce similar behavioral effects, thus ruling out confounding sedation-related variables such as reduced sensory input or behavioral activity (Freund, 1974). There is now good histopathological evidence that both the hippocampus and cerebellum are severely affected by chronic ethanol consumption. Five months exposure to the ethanol-containing liquid diet developed by Freund and Halker led to dramatic changes in mouse hippocampal neuronal morphology as observed with Golgi stains. The CAl pyramidal cell basilar dendrites were severely attenuated while the apical shafts were fairly well preserved except for the most distal arborization. Dentate granule cells showed relatively good preservation of general neuronal morphology. Both subfields exhibited severe decreases in spine density (Riley and Walker, 1978). Subsequent work in rats has revealed a 15-20 % loss of CAl and CA3 pyramidal cells, dentate granule cells and cerebellar granule cells (Walker et~-, 1980).
PAGE 23
13 In summary, animal models of chronic ethanol consumption have pro vided data relevant to two important issues. First, the direct neuro toxic effects of ethanol have been confirmed since the models generally employ nutritionally adequate diets. Secondly, there is now more con clusive neuropathological evidence that the hippocampus is adversely affected by chronic ethanol consumption. The success of the models on these t\'m points encourages further research aimed at determining more directly how hippocampal functioning is altered by chronic ethanol exposure. Hippocampal Anatomy and Physiology The hippocampus is a large, 3-layered allocortical structure sandwiched {in rat) between the neocortex and thalamus/midbrain. Rostrally, it lies just posterior to the septal nuclei. It arches first posteriorly, then ventrolaterally and then again rostrally until it rests temporally just behind the amygdala. Thus, we can assign septal and temporal poles to this structure. The course of the hippocampus between the two poles is termed the longitudinal axis. The hippocampus is divided into t\'JO subregions, Amman's horn and the dentate gyrus. In cross-section, these two regions can be viewed as two interlocking Cs. The dorsal or lateral arm of Amman's horn (Ramon y Cajal 's regio superior, 1911) adjoins the periallocortial subiculum while the ventral or medial arm (Ramon y Cajal 's regio inferior) fits into the concavity (or hilus) of the dentate gyrus. Regio superior is largely segregated from regio inferior and the dentate gyrus by the hippocampal fissure. The dentate gyrus is also subdivided into two segments--the inner blade (facing the fissure) and the outer blade (facing the outside of the brain).
PAGE 24
14 The dentate gyrus (DG) is trilayered: a cell-poor molecular layer, the granule cell layer and the polymorph layer. The tightly packed granule cells send their highly ramified, spiny dendrites out across the whole molecular layer while their axons perforate the polymorph layer and hilus to innervate Amman's horn (AH) (Ramon y Cajal, 1911). All other cell types project only within the DG. The polymorph layer contains most of these short axon cells, including basket cells. Both their dendrites and axons extend across the granule cell layer into the molecular layer (Lorente de No, 1934). The remainder of the hilar region represents a transition from the polymorph cells to the pyramidal cells of AH. Recent studies have shown that these deep hilar cells project only to the ipsilateral and contralateral dentate molecular layer (Amaral, 1978; Fricke and Cowan, 1978; Swanson et~-, 1978). Ammon's horn is also trilayered; a relatively thin core of packed projection (pyramidal) cells is flanked on either side by cell-poor molecular layers. However, the organization of AH is sufficiently complex to warrant subdividing this region into several more laminae (Fig. 1-1). Starting with the deep layers and moving superficially toward the fissure are: l) alveus, myelinated afferent and efferent fibers,2) stratum oriens, containing axons running horizontally and at right angles to the basilar dendrites of the pyramidal cells, 3) stratum pyramidale,4) stratum lucidum,5) stratum radiatum, and 6) stratum lacunosum-moleculare. The last three laminae are all invaded by pyramidal cell apical dendrites and are distinguished by the afferents which terminate specifically in each layer. Finally, the pyramidal cell layer has been divided into a four part scheme (CAl-4) based on cell morphology and afferent connections which we will describe later (Lorente de No, 1934).
PAGE 25
Figure 1-1. A schematic hippocampal lamella diagramming the major excitatory trisynaptic circuit. Stippled areas indicate the pyramidal and granule cell layers. The heavy black lines indicate the two major afferent pathways that were studied in the present experiments. The electrodes placements for the CAl experiments are shown. The recording electrode site was just dorsal to the pyramidal cell layer while the stimula tion electrode position was in stratum radiatum near the CA1-CA2 border. Abbreviations: COM, commissural fibers; hi, hilus, MF, mossy fibers; PP, perforant path; SCH, Schaffer collaterals; sg, stratum granulosum; sm, stratum moleculare; so, stratum oriens; sp. stratum pyramidale; sr, stratum radiatum.
PAGE 26
G, S9 ~ :~: po ~t:;j!:, "'
PAGE 27
17 The flow of information in the hippocampus is marked by three salient features. The first is the one-way flow of information from entorhinal cortex to DG, CA3, CAl and then subiculum. Each of these afferent fibers involved provides a powerful excitatory synaptic drive onto the next region (Andersen et~-, 1966a). The second feature is the lamination of the synaptic fields. Each afferent pathway involved in this flow of information terminates on specific portions of the post synaptic dendrites. The last organizational feature is the remarkable lamellar orientation of these connections. Slices, 300-500 m thick and transverse to the longitudi n al axis, will contain many of the major fiber tracts and connections intact. The intrinsic synaptic organization of the hippocampus is reasonably well known. Axons from layer two of entorhinal cortex (medial and lateral) course in the angular bundle (perforant path) and enter the hippocampus posteriorly. These fibers terminate in the outer 2/3 of the molecular layer of the DG and distribute topographically onto granule cell dendrites (Hjorth-Simonsen and Jeune, 1972; McNaughton and Barnes, 1977; Steward and Scoville, 1976). The proximal 1/4 of granule cell dendrites is innervated by the ipsilateral and contralateral deep hilar neurons (Fricke and Cowan, 1978; West et~-, 1979). The granule cell axons form mossy fibers and exit through the hilus. Collaterals of these fibers innervate the polymorphic and hilar cells while the main axons stream into CA3 and terminate in stratum lucidum on the most proximal portions of the pyramidal cell apical dendrites (Lorente de No, 1934; Swanson et~-, 1978). The mossy fibers are excitatory and do not extend past CA3.
PAGE 28
18 The pyramidal cells of CA3 have the most diverse projections of any hippocampal neurons. The longitudinal association pathway courses through the ipsilateral stratum radiatum of CA3 and CA2, innervating CA3 pyramidal cells throughout the longitudinal axis (Raisman et~-, 1965; Swanson et~-, 1978). The second major projection is bilateral via fimbria and fornix to the lateral septal nucleus (DeFrance ~~-, 1973; Meibach and Seigel, 1977). Branches of these axons, known as Schaffer collaterals, perforate the pyramidal cell layer and terminate most densely in stratum radiatum of CAl. A less dense innervation of CAl stratum oriens has also been observed (Hjorth-Simonsen, 1973; Lorente de No, 1934; Swanson et~-, 1978). Commissural fibers cross in the ventral psalterium and innervate CA1-CA3 contralaterally. These fibers overlap the ipsilateral Schaffer collateral projection by terminating in both stratum oriens and radiatum (Gottlieb and Cowan, 1973). A branching axon from a single CA3 neuron may contribute to the septal, commissural and the Schaffer collateral projections (Swanson et~-, 1980). Axons arise from the basal portion of the CAl pyramidal cells, descend into the alveus and bifurcate. The rostrally directed axons exit via the fimbria and fornix and terminate ipsilaterally in the lateral septal nucleus (Mei back and Seigel, 1977; Swanson and Cowan, 1977). The caudally directed axons terminate in the subiculum, para subiculum and entorhinal cortex (Finch and Babb, 1980; Lorente de No, 1934). No commissural pathways originate from CAl (Blackstad, 1956; Gottlieb and Cowan, 1973; Laurberg, 1979). Besides the entorhinal cortex, there are three major sources of extrahippocampal afferents: medial septal nucleus and vertical limb of
PAGE 29
19 the diagonal band, scattered diencephalic nucl~i and the brainstem monoamine nuclei. These afferents exert a complex mixture of excitatory and inhibitory influences on all regions of the hippocampus. However, since they have little bearing on the current experiments, they will not be discussed further. Although the hippocampus is highly organized and structured ana tomically, its synaptic efficacy is extremely labile. There are sev eral lines of research investigating these special plastic properties of hippocampal synapses. The research relevant to this work is concerned with how the efficacy of selected hippocampal synaptic connections is modified by prior synaptic activity. One striking aspect of research in this area is the almost exclusive use of extracellular field potential analyses of synaptic activity. These macropotential-based analyses provide sensitive and informative data concerning the status of specific connections within the hippocampus. This is possible because the highly stratified inputs and the homogeneous arrangement of the postsynaptic cellular elements produce very structured but simple source-sink relationships. Accordingly, electrical stimulation of a single afferent pathway will activate synaptically only a restricted dendritic region. If the input is excitatory, a 11sink11 will be generated by current flowing into the cells. To complete the circuit, current will flow out of the cell at the 11source11 (neighboring cellular regions) and through the extracellular space back to the site of the sink. The flow of current through the resistive extracellular fluid generates a voltage which is easily detected by a nearby, extracellularly-placed microelectrode. At the sink, the intracellular positive potential will be recorded as a
PAGE 30
20 negative voltage extracellularly. In a similar manner, the electrode will record a positive voltage at the source. The size, packing density and homogeneous orientation of dendrites and somata are such that quite large evoked potentials can be recorded, i.e., in the millivolt range (Andersen and Loma, 1970; Teyler et tl-, 1977). Separating the source and sink regions is an isopotential zone commonly referred to as the inversion point. Loma (1971a) showed that extracellularly recorded potentials had the same onset as intracellularly recorded excitatory postsynaptic potentials (EPSPs). Thus, he termed this extracellular potential the population EPSP (hereafter referred to simply as the EPSP). Two other extracellular potentials are commonly recorded in the hippocampus. The first is the compound action potential representative of synchronous activity of the afferent fibers. This di-or triphasic event, termed the presynaptic volley, is small relative to the synaptic potentials and can usually be recorded only in the layer containing the afferent fibers (Andersen~~-, 1978; Loma, 1971a). At high stimulation intensities, a negative deflection can be observed superimposed on the extracellular EPSP recorded in the cell layer {Fig. 1-2). Loma (1971a) demonstrated that this sharp negativity is in fact a compound spike potential representative of the synchronous discharge of postsynaptic neurons. He termed this potential the population spike (PS) in agreement with Lorente de No (1947). More detailed analyses have confirmed the relationship of the PS and unit activity (Andersen et tl 1971 b). The plasticity exhibited at hippocampal synapses has been sub divided into broad categories based on the frequency and duration of
PAGE 31
Figure 1-2. A CAl field potential evoked by high intensity Schaffer collateral/commissural stimulation and recorded just dorsal to the pyramidal cell layer. The labelled fea~ tures of the waveform are: a, calibration pulse (1 mv, 2 msec, positive up) injected during baseline recording prior to stimulation; b, stimulus artifact; c, onset of the population EPSP (seen as positive or source since the recording was taken remotely from the active excitatory synaptic region or sink); d, population spike superimposed on the EPSP and indicative of synchronously active pyramidal cells. Quantification of these potentials is described in the text. Calibration: 1 mv, 2 msec.
PAGE 32
CA1 EP a b d 1mvL 2 msec N N
PAGE 33
23 afferent stimulation as well as the duration of response change. The categories include paired-pulse potentiation, frequency potentiation, posttetanic potentiation and long-term potentiation. Each of the phenomena will be briefly discussed. Paired-pulse potentiation is identical to the condition-test tech nique described in other systems. In this situation, the response evoked by the first (conditioning) stimulus is compared to the response to the delayed second (test) stimulus. Such double-pulse stimulation of excitatory hippocampal afferents results in a potentiated EPSP and PS. Although the exact characteristics of potentiation differ for the PS and EPSP, potentiation can be observed with interpulse intervals as long as several hundreds of milliseconds (Assaf and Miller, 1978; Lomo, 1971b; Ste\.'1ard et _tl., 1977). Frequency potentiation refers to an enhanced response, both EPSP and PS, during a train of stimuli to an afferent pathway. Unlike with paired pulses, qualitative differences are found among hippocampal subfields. Frequency potentiation is typical of mossy fiber-CA3 synapses as well as Schaffer collateral-CAl synapses. The optimal frequencies for potentiation are approximately 8-12 Hz (Dunwiddie and Lynch, 1978). Stimulation at these frequencies often leads to the appearance of double or triple population spikes, a phenomenon poorly understood at this time (Bliss and Loma, 1973). On the other hand, the perforant path to dentate synapses often becomes depressed with repetitive stimulation (~Jhite ~_tl., 1979). Detailed analyses of this occurrence indicated that the parametric features of the response decrement conformed to previously established criteria for "habituation" (Harris et .tl_., 1979a; Teyler and Alger, 1976; Thompson and Spencer, 1966). Curiously,
PAGE 34
24 habituation is observed in the hippocampus only at the perforant path granule cell synapse (Teyler and Alger, 1976). The first event that follows tetanic stimulation is posttetanic potentiation (PTP), generally lasting only a few seconds. Careful analysis of this short-lasting augmentation in the dentate gyrus (McNaughton, 1977) indicated that its decay is described by two expo nential equations with time constants consistent with those found at frog neuromuscular junction {Magleby and Zengel, 1976). Generally, PTP is found superimposed on a period of response depression which can last for several minutes (Bliss and Lomo, 1973; Deadwyler et ~q 1978; Teyler et~., 1977). Subsequently, if the frequency and duration of stimulation is adequate, a later period of enhancement develops. This late-developing enhancement is unusual in that it can persist for ex tremely long periods of time, on the order of days or weeks (Buzsaki, 1980; Bliss and Gardner-Medwin, 1973; Douglas, 1977; Douglas and Goddard, 1975). Thus, it is commonly referred to as long-term potentiation (LTP). Long-term potentiation differs from frequency potentiation in that the optimal frequency for LTP is 100 Hz or more (Dunwiddie and Lynch, 1978). In summary, the laminar organization of the hippocampus lends itself well to extracellular monitoring of synaptic activity. The relative ease of data collection and interpretation plus the lability of its synapses not only facilitateselectrophysiological analysis of hippocampal functioning but also makes this structure an exciting model of synaptic plasticity. Studies of the effects of chronic ethanol con sumption on these processes would significantly enhance our understand ing of ethanol's toxic effects on the hippocampus.
PAGE 35
25 Animal Models of Chronic Alcohol Consumption Effective animal models of chronic alcohol consumption must fulfill certain conditions. For example, the ethanol must be delivered in a manner that is practical for extended periods of time. The typical means of accomplishing this is through oral administration, although intragastric intubation techniques have been used. The animals must also receive a fairly large proportion of their calories as ethanol yet maintain adequate nutrition and normal body weight gain. This difficult though necessary requirement is not met by all paradigms. For instance, some studies employ aqueous ethanol solutions plus normal lab chow diets for the experimental animals. However, ingestion of ethanol-derived calories may lead to decreased consumption of lab chow, promoting malnutrition and reduced weight gain. Our laboratory has previously developed a feeding procedure which overcomes the above limitations (Walker and Freund, 1971 ). Experimental animals receive a special liquid diet as their sole source of calories. The solution contains a large percentage of ethanol-derived calories (35-39 % ) yet is fortified with essential vitamins and minerals. Control animals are either pair-fed a solution with sucrose isocalorically substituted for ethanol or are maintained on standard lab chow and tap-water diets. As noted earlier, animals restricted to the ethanol-containing diets for extended periods of time show both severe behavioral deficits and pathological changes in hippocampal neuronal morphology (Riley and Walker, 1978; Walker and Hunter, 1978).
PAGE 36
26 Rationale The similarities of alcohol-related and hippocampus-related amnesic syndromes strongly suggest that chronic ethanol ingestion may be toxic to the hippocampus. Synaptic plasticity is a fundamental feature of the hippocampus and such plasticity is believed to have important be havioral implications. The purpose of the present research was to investigate whether ethanol has a detrimental effect on the functional plasticity of hippocampal synapses. Studies of aging animals have supported the logic underlying these experiments. Both behavioral deficits and neuropathology in the hippocampus are prominent in aged animals (Barnes, 1979; Bondareff, 1979; Freund and l1alker, 1971a; Bondareff and Geinisman, 1976; Geinisman et El, 1978; Scheibel et~-, 1976). These findings have led other researchers to study the aged hippocampus electrophysiologically. Landfield et al. (1978) reported that the Schaffer collateral to CAl synapses were essentially normal in all animals when given single or paired stimuli. However, when this system was challenged with high frequency stimulation, measures of both long-and short-term potentiation revealed diminished enhancement in aged animals. More quantitative studies (Barnes, 1979; Barnes and McNaughton, 1980) have demonstrated a different pattern of results for the per forant path to dentate gyrus system. First, single stimuli at a fixed current produced smaller synaptic responses in aged animals. However, if one controlled for the number of fibers actually stimulated, the existing synapses in the aged animals proved to be more powerful. When a high frequency tetanus was delivered, aged animals showed normal PTP and LTP. Repeated tetani elevated the LTP asymptote of the
PAGE 37
27 evoked response in younger animals, but they had no effects on the aged animals beyond those produced by the first tetanus. Barnes (1979) con firmed Landfield's finding that the groups did not differ in their capacity for paired-pulse potentiation. These exciting findings in aged animals emphasize the need for similar research in alcoholic animals. The following experiments represent the first attempt to investigate the long-term consequences of chronic ethanol consumption on hippocampal synaptic function.
PAGE 38
CHAPTER II GENERAL METHODS Ethanol Administration Male Long-Evans hooded rats (200-250 g) were matched by weight and assigned to the following treatment groups: 1) an experimental group (Group E) which received an ethanol-containing liquid diet, 2) a control group (Group S) which was pair-fed a sucrose-containing liquid diet, and in some experiments 3) a second control group (Group LC) which received pelleted laboratory chow and water ad libitum. The preparation, contents and nutritional adequacy of the liquid diet pro cedure have been documented by Walker and Freund (1971). The ethanol liquid diet had 35-39% ethanol-derived calories (8.l-9.4% v/v, ethanol) and was prepared by mixing an ethanol stock solution (63.3% v/v with Sustacal (Mead-Johnson Co.). The control liquid diet was identical except sucrose was isocalorically substituted for ethanol. Both diets were fortified with Vitamin Diet Fortification Mixture, 0.3 g/100 ml, and Salt mixture XIV, 0.5 g/100 ml (Nutritional Biochemicals Co.). The liquid diets provide 1.3 kcal/ml. The rats were kept individually housed in stainless steel cages.in a colony room with 0700-1900 hr light cycle. All liquid diets were prepared fresh each day and ad ministered in calibrated drinking bottles with stainless steel drinking tubes. Diet consumption was measured daily. The liquid diets were administered for a period of 20 weeks. The percentage of ethanolor sucrose-derived calories was increased by 1 % 28
PAGE 39
29 every four weeks (35-39 % ). Group S rats were pair-fed with Group E rats in order to equalize caloric and nutrient intake during the 20 week period. At the end of the 20 week treatment period, all rats received laboratory chow and water ad libitum. Electrophysiological recordings were obtained within a 10 week period commencing eight weeks after discontinuation of the liquid diet treatment. This extended ab stinence period was used to eliminate or at least minimize the residual effects of ethanol intoxication or acute ethanol withdrawal. Moreover, both the duration of ethanol treatment (20 weeks) and ethanol abstinence (eight weeks) were chosen to facilitate comparisions with previous morphological studies (cf. Walker et.!!_., 1980). Rats from each group were coded in order to prevent experimenter bias in final electrode placement and data collection. The code was not broken until data analysis was completed. Electrophysiological Methods Recordings were made in urethane-anesthetized preparations (1.5 g/ kg ip.) with supplemental doses of urethane administered as required. Each animal was placed in a Kopf stereotaxic instrument (located within a screen-shielded cage) and the skull exposed. The skull tilt was adjusted so that lambda and bregma were level on the horizontal plane. All stereotaxic coordinates were referenced to bregma, midline and the skull surface. A heating pad was used to maintain rectal temperature at 37 0.5 C. Stimulating and recording electrodes were remotely controlled via hydraulic microdrives and placed in the brain through small burr holes in the skull.
PAGE 40
30 Each animal was implanted with four electrodes. A screw electrode was placed anterior to the brain to ground the preparation. A 75 m platinum-iridium wire served as the reference electrode and was placed in the neocortex close to the hippocampal recording site. The stimulating electrode was a concentric bipolar (Rhodes Medical Instruments) with a 250m outer diameter and a 100 m tip bared for 25 m and extending 75 m beyond the main shaft. The electrodes had an impedance of 20-30 kohms at 1 khz. A fiber-filled glass micropipette filled with 4 M NaCl (1-2 m tip diameter, 1-3 Mohms at 1 khz) was used for extracellular field potential recordings. Field potentials were amplified by a Grass P511 differential AC preamplifier, filtered at 0.3-10 khz and either recorded immediately on an X-Y plotter or stored on magnetic tape for later analysis. Potentials were in some cases averaged by a Dagan 4800 Signal Averager. Electrical stimulation was delivered by a Nuclear Chicago constant current stimulator and consisted of monophasic square wave pulses (0.1 msec duration). During high frequency stimulation, biphasic pulses (0.1 msec each half phase) were usually employed to reduce electrode polarization. A 1-2 mv (2 msec) calibration pulse was routinely given just prior to each stimulus. Plotted field potentials were quantified with a Numonics Digitizer. Experimental Protocol The following experiments include studies of both CAl and the DG. Since these two regions have somewhat different physiological characteristics, the experimental protocol was somewhat different for each region. Nonetheless, the basic data collection procedures will be out lined here and deviations from the protocol will be noted in the appropriate chapters.
PAGE 41
31 At the beginning of a recording session, the recording electrode was lowered into the appropriate hippocampal subfield while monitoring extracellular unit activity. The stimulation electrode was then placed to produce maximal activation of the afferent fibers of interest as revealed by the amplitude of the PS. Following determination of the optimal electrode configuration, synaptic distribution and function were evaluated through five separate stimulation procedures. l) Input/ output (I/0) curves were generated by systematic variation of the stimulus current (10-1000 a) in order to evaluate synaptic potency. Stimulus pulses were delivered at frequencies ranging from 0.1-0.03 Hz and four responses were averaged at each current intensity. 2) A laminar analysis was conducted by stepping the recording electrode in 25 m increments through the subfield along an axis parallel to the orientation of the pyramidal or granule dendrites. The stimulus current was fixed to give an EPSP amplitude 50% of that seen at the PS threshold. At each 25 m step and after a 25 sec recovery period, four to eight field potential responses to 0.1 Hz stimulation were recorded for sub sequent averaging. 3) Paired-pulse potentiation (PPP) was evaluated at low (subthreshold for a PS) and high (suprathreshold for a PS) stimulus current intensities. Interpulse intervals (IPI) varied from 15-180 msec. Stimulus pairs were delivered at 0.1-0.03 Hz and four responses were averaged at each IPI. 4) From three to five frequency potentiation ifEl series were conducted at varying stimulation frequencies and durations: 1 Hz/5 sec, l Hz/25 sec, 5 Hz/5 sec, 10 Hz/2.5 sec and 10 Hz/5 sec. The stimulus current was always suprathreshold for the production of a PS. This design allowed for comparisons of FP at equivalent tetanus duration (5 sec) or at equal numbers of stimuli within the
PAGE 42
32 train (25 pulses). Every third response within each tetanus was measured and compared to the baseline response generated prior to each tetanus. 5) Following each FP series, test pulses were systematically delivered for 10-30 min in order to assess the development of LTP. In addition, LTP was produced by high frequency stimulation at the end of each experiment and monitored for 30 min. Following data collection, small electrolytic lesions (10 a, 10 sec, anodal) were made at the stimulation site. Similar lesions were made through the recording electrode at the isopotential point and at a fixed distance from the isopotential point along the electrode tract. Measurement of the distance between the recording electrode lesions permitted correction for tissue shrinkage normally occurring during histological preparation. Electrode placements were verified in all animals through microscopic examination of the lesions in myelin stained sections. Data Analysis A major difficulty in making quantitative comparisons between treatment groups is the extreme within-group variance in evoked responses normally observed in the intact hippocampus (Barnes, 1979; Landfield et~-, 1978). Several standardization procedures were employed to reduce this variability. First, following the laminar analysis, the recording electrode was placed at a fixed distance (125-150 m) from the inversion point. All I/0 curves, PPP, FP and LTP sequences were performed at this fixed reference point in each animal. Secondly, stimulus current strengths for PPP, FP and LTP were all normalized with respect to the individual I/0 curve. These values
PAGE 43
33 will be detailed later for each study. The stimulus current was readjusted prior to tetanus delivery when long-term effects of the previous tetanus were observed. Finally, all electrodes were carefully selected to fall within a narrow range of tip size, impedance and shaft taper requirements. Analysis of the evoked responses generated by the laminar analysis does not allow a precise localization of current sources and sinks. In order to enhance the spatial and temporal resolution of the analysis, a one-dimensional current-source density (CSD) analysis was applied to the data (see Appendix). Current densities were derived from the extracellular evoked responses by calculation of the second spatial derivative of the potentials using the 04 smoothing of Freeman and Nicholson (1975). The CSD computed with this smoothing was found to have low noise, yet still exhibit all major current sources and sinks. Since tissue conductivity was not determined, the results are expressed in mv/mm2 Due to the laminar and lamellar organization of the hippocampus, the one dimensional CSD as employed here assumes that the major synaptic currents flow parallel to the long axis of the pyramidal cells and that no large conductivity gradients exist in this dimension (Freeman and Stone, 1969; Nicholson and Freeman, 1975). A recent study employing CSD analysis of the Schaffer/commissural (SCH/COM) afferents to CAl supports the validity of this approach (Leung, 1979), provided that the field of excitation sampled by the microelectrode is confined to the center of the lamella. In order to minimize errors arising from violations of these assumptions, and to reduce response variability, the following standardization procedures were employed. First, the orientation of the stimulating
PAGE 44
34 and recording electrodes was adjusted to correspond to the orientation of the lamellae for the SCH/COM-CAl path (Andersen et~-, 1971a). The lamellae are oriented saggitally (relative to midline) with a slight shift anteriorly. In order to insure the intralamellar localization of the electrodes, the location of the recording electrode was adjusted until evoked potentials were obtained with EPSP and PS thresholds that fell within a relatively narrow range. There is 1ess concern about the lamellar organization of the DG since the stimulating electrode was placed in the angular bundle and thus activated fibers which project all along the 1ongitudinal axis of the hippocampus. The EPSP and PS thresholds were assessed for each animal at a fixed recording site relative to the inversion point of the laminar ana1ysis. Finally, the stimulus current used to generate the 1aminar profiles was chosen according to the standardization procedure described above. Fol1owing data analysis, the code was broken and the data grouped by treatment The potentiation data were expressed as either difference from baseline response (latency measures) or percentage of baseline response (amplitude measures). The data were typically analyzed by 2-way analysis of variance (ANOVA) with repeated measures on one factor (treatment by stimulus condition). Individual comparisons were performed using Duncan's multiple range test and student t-tests. Withinanimal comparisons were accomplished through use of paired t-tests~
PAGE 45
CHAPTER I I I CHRONIC ETHANOL EXPOSURE AND SYNAPTIC DISTRIBUTION IN CAl OF RAT HIPPOCAMPUS: CURRENT-SOURCE DENSITY ANALYSIS Introduction The pathological deterioration associated with chronic ethanol abuse in man is normally attributed to a variety of coexisting conditions, most often malnutrition. However, experiments in laboratory animals have provided convincing evidence that ethanol, even in the presence of a nutritionally adequate diet, exerts toxic effects in the CNS. The neurotoxic actions are manifest both indirectly as acquisition deficits in a variety of tasks (Denoble and Begleiter, 1979; Fehr et~., 1976; Walker and Freund, 1971; Walker and Hunter, 1978} and directly as morphological deterioration in such brain regions as hippocampus and cerebellum (Riley and Walker, 1978; Walker et tl, 1980). In the rat hippocampus, chronic ethanol treatment produces a 15-20 % loss of granule cells of the dentate gyrus as well as pyramidal cells of CAl and CA2-4. In addition, studies in Golgi material have provided evidence of dendritic atrophy and spine loss in the mouse hippocampus (Riley and Walker, 1978). If the morphological deterioration observed in the hippocampus is to be related to the behavioral deficits in ethanol-treated animals, it is imperative that synaptic function be studied directly. In the present study, we examined electrophysiologically the persistent effects of chronic ethanol treatment on synaptic function in CAl of the 35
PAGE 46
36 rat hippocampus. Electrical stimulation within stratum radiatum simul taneously activates SCH and COM fibers (from ipsilateral and contralateral CA3, respectively) which course together and terminate on the pyramidal cells in stratum radiatum of CAl. The stratification of the various afferents and the homogeneous arrangement of cellular elements in CAl allow meaningful extracellular analysis of synaptic potentials generated by the pyramidal cells. However, the extracellular potentials generated under these conditions can be rather large and volumeconducted for great distances, hampering the utility of this analysis. We have therefore combined extracellular evoked potential (EP) and current-source density analysis in order to more precisely define the distribution of synaptic currents induced by SCH/COM inputs to CAl a.fter chronic ethanol treatment. Methods The animals, treatment conditions and general electrophysiological methodology have been described previously in Chapter II. Following placement of the electrodes, the synaptic distribution of SCH/COM afferents to CAl was examined by laminar analysis and assessment of I/0 relations. Initially, the extracellular isopotential point of the laminar analysis was determined and the recording electrode adjusted to a fixed point 125 m more dorsally. Thresholds for both the EPSP and the PS were determined at this fixed reference point. The stimulus current was then adjusted to produce an EPSP which was 50% of the EPSP amplitude observed at the PS threshold. The recording electrode was moved to the point 350 m dorsal to the inversion point and a laminar analysis was conducted in 25 m increments at the fixed stimulus
PAGE 47
37 current over a 1000 m descent through the hippocampus. At each 25 m step and after a 25 sec recovery period, four EP responses to 0.1 Hz stimulation were recorded for subsequent averaging and CSD analysis by a PDP 11/40 laboratory computer. At the end of the laminar analysis, the recording electrode was withdrawn to the point 125 m dorsal to the inversion point. Input/output functions were then ob tained by systematically varying the stimulus current (20-1000 a). At the termination of each experiment, an electrolytic lesion was placed at the stimulation site. Electrolytic lesions (10 a, 10 sec) through the recording electrode were made at the inversion point of the laminar analysis and 650 m more ventrally. The location of the lesions was verified in all animals by light microscopy in myelin stained sections. Results The results are based on a total sample of 14 rats which were distributed as follows: E = 6, S = 4 and LC= 4. Because of the small sample and the fact that Groups Sand LC rats did not differ on any measure, they were combined into a single control group (C = 8). Group E rats consumed a mean daily ethanol dosage of 13.91 g/kg during the 20 week treatment period. While blood ethanol concentrations were not evaluated in these rats, both the pattern and magnitude of daily ethanol intake were comparable to previous experiments in which behavioral or morphological changes were observed (Walker and Hunter, 1978; Walker et~-, 1980). Mean body weights at the termination of the experiment were comparable across groups: E = 531 g, C = 513 g.
PAGE 48
38 Histological analysis. One consequence of significant cell loss could be a selective shrinkage of hippocampal volume relative to the remaining brain mass. If this occurred, the relative position of electrodes placed within the hippocampus using fixed stereotaxic coordi nates could be differentially affected by ethanol treatment. However, three-dimensional plots of electrode tip positions relative to the septal pole of the hippocampus, the brain surface and the midline revealed no significant differences between groups. Stimulating electrode sites were localized to stratum radiatum usually near the CAl-CA2 border, while recording electrode tracts were consistently found 700-1000 m more posteriorly in CAl. All electrode placements were con fined to the dorsal hippocampus. Lesions at the isopotential point of the laminar analysis were always near the border of stratum pyramidale and radiatum (range, relative to stratum pyramidale: 0-44 m into stratum radiatum). Detailed measures (corrected for histological shrink age) were made of the major laminae of CAl including stratum oriens, pyramidale, radiatum and lacunosum-moleculare. These measures were made along the actual trajectory of the electrode track through CAl. Significant differences in the width of the laminae were not observed between groups (Table 3-1 ). I/0 relationships. Chronic ethanol treatment failed to significantly alter basic synaptic responses to SCH/COM stimulation. Neither EPSP thresholds (E = 70.0 4.8 a; C = 52.5 6.5 a) nor PS thresholds (E = 260.0 46.0 a; C = 218.7 40.3 a), nor EPSP amplitude at PS threshold (E = 2.96 0.39 mv; C = 2.91 0.34 mv) was significantly different between groups (data presented as mean SEM). Finally, 2-way ANOVAs revealed no statistically significant group differences over the
PAGE 49
39 TABLE 3-1 Chronic Ethanol Effects on Widths of CAl Laminaea Laminab Control Ethanol b. SO width (l-lm) 128.8 7.0 141.5 4.0 12. 7 SP width (m) 43.2 2.9 49.8 2.5 6.6 SR width (m) 279.9 7.0 270.3 8.2 -9.6 SM width (m) 86.6 3.9 80.6 4.6 -6.0 CAl width (m) 538.5 l 0. 7 541.5 12.7 3.0 aMean SEM bso, stratum oriens; SP, stratum pyramidale; SR, stratum radiatum; SM, stratum moleculare b.% 10.0 15.3 -3.4 -6.9 0.6
PAGE 50
40 range of stimulus currents examined (20-1000 a) for latency to EPSP onset, latency to peak of the PS, EPSP amplitude or PS amplitude. Field potential and CSD analysis. Stimulation of the SCH/COM pathway produced a surface-positive, deep-negative dipole field. These potentials reflect summated excitatory synaptic activity (current sinks) in stratum radiatum, the terminal region of SCH/COM fibers, and an outward (largely capacitative) current source in stratum pyramidale and the proximal portions of stratum oriens. Figure 3-1 shows a typical EP profile and associated CSD distribution plotted as a function of depth from the ventral alvear surface. The spatial distributions of the field potentials and current densities within CAl are more easily understood when amplitude profiles are generated at fixed latencies from stimulus onset. Figure 3-2 illustrates EP and CSD profiles at 8.0 msec (from stimulus) for both a control and an ethanol-treated animal. It is clear from this figure that the CSD plots provide considerably more localization of current flow than the EP plots. In both animals, the major current sink occurred in stratum radiatum, and was bounded dorsally and ventrally by current sources. A major current source was observed in stratum pyramidale and proximal portions of stratum oriens. A smaller current source extended through most of stratum moleculare to the hippocampal fissure. These results compare favorably with previous descriptions of current densities in the SCH/COM-CAl path (Schubert and Mitzdorf, 1979). However, in contrast, the current sink in stratum radiatum always exhibited two components (Fig. 3-2) which were separated by near zero current density or, more rarely, by a minor source in the middle of stratum radiatum. These two component sinks may reflect a separation of the COM and
PAGE 51
Figure 3-1. Full one-dimensional laminar analysis through CAl for a single Group E animal. Field potentials are represented on the left, current densities on the right. Depths indicate distance ventral to the ventral alvear border. In this animal, the stratum pyramidale-stratum radiatum border is found about 175 m from the alveus. Stimulation was applied at time Tl but the artifact has been suppressed in this figure. The T2 bar (8. 0 msec from stimulus) indicates the slice of time used for Figures 3-2 and 3-3. The EP waveforms include a calibration pulse (1 mv, 2 msec, positive up) prior to the stimulus. For the CSD \'Javeforms, an upward deflection from baseline represents a source; a downward deflection represents a sink.
PAGE 52
DEPTH (,um) Tl 0-50-100-150-200-250-300-350-400-450-500-525-I 0 I 8 0 16. 0 TIME ( ms) I 24 0 CAL 2mV t 500 mV/mm2 I I 0 8 0 16. 0 240 TIME (ms)
PAGE 53
Figure 3-2. Individual laminar profiles of EPs and CSDs at 8.0 msec after stimulus delivery showing EP and CSD amplitudes at a fixed latency from stimulus illustrates more clearly the change in responses recorded across the width of CAl (compare with Fig. 3-1). They-axes are scaled to indicate depth from the ventral border of the alveus. Note the different x-axis scalings. Data from an ethanol animal is shown on the top half of the figure; a control animal is presented on the bottom half. The field potentials change polarity at about the stratum pyramidale radiatum border. Positive currents are sources; negative currents are sinks.
PAGE 54
E ::t. :r: I0.. w 0 44 FP 0--------.------100-200300400500. i I -70 -35 0-..... IOO-. ,. 200300400-500-. 1 -50 -25 V V I 0 ... ... I 3 5 (mVl ... ... : ....... . ....... 0 2 5 (mv) ETHANOL I 70 CONTROL . 5 0 CSD I ..i 0 1000 2000 Im (mv/mm2) . .. -600 -300 0 .. I ..... 300 600 Im (mv/mm2)
PAGE 55
45 SCH terminal fields within stratum radiatum, a hypothesis that will be considered in detail below. Group comparisons of the spatial distribution of the field potentials and current densities required normalization, since the absolute measures of the major laminae of CAl varied across animals (Table 3-1). The following normalization was e m ployed. First, EP and CSD profiles during the rising phase (4.0 msec) and at the peak (8.0 msec) of current flow were obtained for each animal (e.g., Fig. 3-2). Only those values contained between the vent r al alvear border and the hippoc ampal fissure (as determined from electrode track reconstruction) were considered. This distance was then divided into equal increments of 4 % of the total distance Since the mean total distance was approximately 540 m for each group (Table 3-1), each equivalent step (4% ) reflected a mean distance of 21.6 m. The amplitude of the EP and CSD for each animal was obtained at these percentage increments using linear interpolation between data points. This analysis was performed with a Numonics Digitizer. Group laminar profiles (at 8.0 msec) are presented in Figure 3-3 relative to a sample CAl pyramidal cell and the mean widths of the major CAl laminae While the spatial resolution of the current sinks is diminished in the control group plot, two distinct peaks of inward current are still found at 71.3 m and 228.3 m from the pyramidal cell layer. It can be readily seen that while chronic ethanol treatment produced only subtle alterations in the EP profiles, the ethanol treatment dramatically altered the spatial distribution and density of the synaptic currents as revealed by CSD analysis.
PAGE 56
Figure 3-3. Group laminar profiles of normalized EPs and CSDs at 8.0 msec after stimulus delivery. On the left is a representative pyramidal cell and demarcations of the major CAl laminae (adapted from Scheibel, 1979). The distance in microns through CAl is an average for all animals (there were no group differences in this measure). The nor malization procedure for both EP and CSD data is described in the text. The EP laminar profile reveals positive potentials in SO and SP which invert just past the cell layer to negative potentials throughout SR and SM. There are no significant ethanol effects except for the negative potential half-width (see text). The CSD profile resolves the current sources (positive Im) and current sinks (negative Im) much more precisely. Passive current sources are observed in SO/SP and in SM. Note the two peaks of inward current in SR of control animals (filled circles). The ethanol group (open circles) shows an overall shrinkage of the sink in SR, a reduced component sink proximal to the cell layer and an expanded component sink more distally from the cell layer Abbreviations: ALV, alveus; SO, stratum oriens; SP, stratum pyramidale; SR, stratum rad i atum; SM, stratum moleculare; HF, hippocampal fissure.
PAGE 57
ALV DEPTH (UM) 0 0 6 EP \ CSD so ~ q '7 100 9 100 ? -o---d / / SP _o/ _o,, / 200 o200 p -// 7 / / / .j:::,, /0/ -...J / o" 300 0/ SR 300 / / / / p ?" CONTRO L I I p 0 o A LCOHOL \ ? 9 I 0 400 400 I b,, 'o --o_ SM 500 500 --q I p / b / HF -60 -50 4 0 3 0 2 0 -10 0 10 20 -40 3 0 -20 -10 0 10 20 30 4 0 50 6 0 AMPLITUDE ( mv) AMPLITUDE (Imv/mm2 J02)
PAGE 58
48 The amplitude and distribution of field potentials and current densities in stratum radiatum were quantified at both 4.0 and 8.0 msec following stimulus onset. Since the spatial patterns of current flow are similar at the two time points, only the results at 8.0 msec will be described (Table 3-2). There were no significant group differences for the majority of EP measures. However, Group E showed a significant reduction of the spatial half-width of the negative potential region. (Spatial half-width was defined as the distance across stratum radiatum between the two points whose amplitudes lay half way between the peak negative voltage and the asymptotic negative voltage.) Quantitative comparisons of Group C and Group E CSD profiles more precisely characterized the changes resulting from the chronic ethanol exposure. Ethanol treatment significantly reduced the spatial extent of the overall current sink in stratum radiatum. This shrinkage of the SCH/ COM synaptic field occurred despite procedures which normalized the individual distance from alveus to hippocampus fissure. Separate sink analyses revealed that the overall shrinkage was totally accounted for by a shrinkage of the proximal current sink. Interestingly, the distal sink of Group E was somewhat enlarged, with a significantly greater amplitude and area, but only a weak trend towards increased spatial extent. These data are suggestive of an increased efficacy of distally located synapses. However, since we did not directly measure tissue conductivity, the amplitude and area measures are only proportional to the actual CSD. Thus, group differences in these measures could reflect a change in tissue conductivity instead of alterations in evoked synaptic current.
PAGE 59
49 TABLE 3-2 Laminar Analysis Quantification (8.0 msec from stimulus)a Chronic Ethanol Effects on Field Potential and Current Density Distributions Measure EP Peak voltage (mv) Peak latency (msec) Spatial half-width (m) Peak distance to cell layer (m) Inversion distance to cell layer (m) CSD Total sink width (m) Distance from inter-sink minimum to cell layer (m) Length along dendritic tree (m) proximal sink distal sink Peak distance to cell layer (m) proximal sink distal sink Peak amplitude (mv/mm2 ) proximal sink distal sink Sink area (mv/mm) proximal sink distal sink ~Mean SEM p < 05 l ta i 1 Cp < .05 2 tail Control -5.5 0.8 8.9 0.4 206.5 8.9 196.8 14. 0 6.9 5.4 283.8 14.1 163. 5 19. 0 141.3 14.4 141.2 18.3 71.3 16.1 228.3 14.5 511.2 83.9 418.1.7 44. 7 7. 2 40.0 11 .6 Ethanol 6% -6.3 0.5 14.1 8.0 0.5 -10.2 172. 4 9. 0 -16. 5C 184. 6 6. 3 -6. 2 8.2 3.5 18.8 246.4 10.0 -13.2b 104.0 8.6 -36.4c 88. 9 8.8 168. 0 16.4 56.4 6.0 181. l 9.2 436. 2 1 51 3 673.6 78.4 27. 2 8. l 70. 9 7. 2 -37. l C 19 .0 -20.9 -20.7b -14.7 61 1 b -39. l 77 .2b
PAGE 60
50 Discussion The results of this experiment further our understanding of the pattern of synaptic distribution of SCH/COM afferents to stratum radiatum of CAl in both control and ethanol-treated rats. The CSD analysis of laminar profiles through CAl indicated the presence of a major excitatory current sink in stratum radiatum but with two spatially distinct components. Chronic ethanol treatment produced a persistent reduction in the spatial extent of the current sink proximal to stratum pyramidale with a less profound expansion of the more distal current sink. The significance of each of these findings will be considered in turn below. Synaptic distribution in stratum radiatum of CAl. It is currently well documented that the major afferents to stratum radiatum and stratum oriens of CAl arise from both ipsilateral and contralateral CA3 (Blackstad, 1956; Hjorth-Simonsen, 1973; Laurberg, 1979; Swanson et~-, 1978). However, within stratum radiatum the density and spatial distribution of terminals from each source has not been satisfactorily characterized. The major unresolved question is whether or not these two afferents terminate differentially across the width of stratum radiatum. Early Golgi studies by Schaffer (1892) and Lorente de No (1934) showed that the Schaffer collaterals run in a dense band (stratum lacunosum) distal to stratum radiatum and presumably terminate in this region and the neighboring portion of stratum radiatum. This laminar organization is particularly prevalent in the rabbit (Lorente de No, 1934). Subsequent physiological evidence supporting a bimodal distribution of SCH/COM afferents to stratum radiatum has emerged from the study of laminar profiles of evoked field potential in rabbits, where
PAGE 61
51 peak responses to activation of COM afferents were found more proximally in stratum radiatum (Andersen et~-, 1966a; Stanley et~-, 1979) than peak responses to SCH stimulation (Andersen et~-, 1966a). Anatomical studies in the rat, employing degeneration (Blackstad, 1956; Hjorth-Simonsen, 1973), autoradiographic (Laurberg, 1979; Swanson et~., 1978) and horseradish peroxidase (Laurberg, 1979) tracing tech niques, have been unable to qualitatively distinguish between the spatial distribution of associational and commissural afferents to CAl. In contrast, others using degeneration and autoradiographic tech niques have been able to identify a bimodal yet overlapping distribution of CA3 afferents to the apical dendrites of CAl pyramidal cells in the rat (Gottlieb and Cowan, 1973; Raisman et~., 1965). Gottlieb and Cowan (1973), by counting silver grains in autoradiographs, have pre sented the only quantitative anatomical evidence that COM terminals concentrate most heavily proximally to stratum pyramidale. While the available anatomical and physiological evidence is presently inconclusive (particularly in the rat), there is some support for the notion that SCH/COM afferents distribute in an overlapping yet bimodal fashion in stratum radiatum. We have therefore hypothesized that the dual current sinks observed in stratum radiatum in the present experiment reflect a separation of the COM and SCH synaptic fields. However, three important methodological issues must be considered in relation to this hypothesis. First, the utility of EP recordings in mapping studies, such as those mentioned above (cf. Andersen et~-, 1966a}, can be considerably enhanced through the use of CSD analysis. The CSD analysis provides a significant improvement in both spatial and temporal resolution of current density in the region of the recording
PAGE 62
52 electrode. Thus, current sinks (current entering the cells) and current sources (current entering the extracellular space) can be precisely localized (Freeman and Stone, 1969; Nicholson and Freeman, 1975). However, in contrast to our results, previous application of CSD analysis to the distribution of synaptic currents following stimulation of stratum radiatum have revealed only a single current sink in stratum radiatum of CAl (Leung, 1979; Schubert and Mitzdorf, 1979). These latter studies used recording intervals of 100 mas compared to the 25 m intervals employed in the present experiment. Since the stratum radiatum is approximately 300-400 min the rat (Table 3-1), we believe this discrepancy can be accounted for by the enhanced spatial resolution provided by the 25 m sampling interval used in our experi ments. A second important factor is the lack of tissue conductivity measurements in the hippocampus. For this reason, the conductivity and gradient-dependent terms (Freeman and Stone, 1969; Nicholspn and Freeman, 1975) remain unevaluated in the CSD calculations. These terms, which are necessary for an accurate quantitative evaluation of transmembrane currents, have been shown to contribute relatively little to the overall current densities in the optic tectum of pigeons and teleosts (Freeman and Stone, 1969; Vanegas et~-, 1979) and in the anuran cerebellum (Nicholson and Freeman, 1975). Although the results of the calculations have been expressed in mv/mm2 (units which are only proportional to transmembrane current), the existence of significant conductivity gradients can confound interpretation of source-sink distributions in the region of the gradient (see Haberly and Shepard, 1973). While this is a concern in a laminated structure such as the hippocampus, there is
PAGE 63
53 no compelling evidence from the anatomical organization to suggest the presence of significant conductivity gradients in the axis parallel to the pyramidal cell dendrites. Nonetheless, it remains possible that careful evaluation of these terms could change the form of the CSD distribution observed in stratum radiatum after SCH/COM activation. Another possible source of artifact which may account for the ap pearance of two current sinks in stratum radiatum is related to the width of the activated synaptic population. Nicholson and Freeman (1975) have calculated radial CSDs for different ratios of the activated population width versus length. If the activated population is narrow, even recordings on the axis of symmetry can produce a distorted CSD distribution, with artifactual bimodal sink peaks in the true sink region (see their Fig~ 6). Stimulation of the SCH/COM system in rabbit hippocampus, however, produces lamellar synaptic activity with a half-width of about 4 mm (Andersen et~-, 1971a). This corresponds to a width-length ratio of about 4, for which artifactual multipeaked distributions are not theoretically expected. In addition, the procedures used in our study for placement of recording and stimulating electrodes should insure recordings along the axis of symmetry of the activated population. It should be noted that the spatial distribution of current sinks does not necessarily quantitatively reflect the density of synapses along the dendritic tree. This discrepancy arises because, when considering the effective transmembrane currents at any one point along the dendritic tree, contributions not only from synapses at that point (if excitatory, then the effective current sink), but also the passive currents from remote synapses (in this case, sources) must be considered.
PAGE 64
54 The passive currents will depend on the membrane properties of the dendritic tree. The result is that the observed current-source density at any given point in the tissue may be approximated by the convolution integral of the density of activated synapses over the entire dendritic tree and the spatial distribution of currents produced by any one synapse. Proper evaluation of the synaptic density distribution given the CSD distribution requires development of a reasonably accurate model representing the dendritic-somatic membrane and deconvolution of the integral. Neither task is trivial. While we recognize these limitations in our interpretation of current source-sink distributions, the working hypothesis that the two sink regions indicate the presence of two overlapping but spatially separate synaptic inputs corresponding to the SCH and COM afferents will be used below in the interpretation of the effects of chronic ethanol treatment on synaptic distribution in stratum radiatum. We are currently attempting to verify this hypothesis through the use of more discrete microstimulation techniques. Chronic ethanol treatment. Chronic ethanol treatment produced persistent alterations in the synaptic distribution of SCH/COM afferents to stratum radiatum of CAl. Ethanol treatment produced a significant reduction (13% ) of the overall spatial extent of the major current sink within stratum radiatum. This presumed loss of SCH/COM synapses correlates well with the 15-20 % cell loss observed in all hippocampal subfields at identical durations of ethanol exposure (Walker et~., 1980). This reduction does not merely result from morphological shrinkage of the laminae of CAl (which might be expected after 15-20 % cell loss) since our normalization procedure controlled for group differences
PAGE 65
55 in the width of CAl. Rather, ethanol treatment appeared to produce an underlying shrinkage of the extent of excitatory current flow within stratum radiatum independent of any alterations in CAl layer thicknesses. The failure to detect changes in the widths of the laminae of CAl (Table 3-1) is of questionable validity at present, since these measures were not made in anatomically matched sections but rather along the trajectory of the electrode penetrations, a procedure which greatly increases variability. Ethanol treatment also produced differential effects on the two components of the major current sink in stratum radiatum. Measures ob tained at the peak of the synaptic response (8.0 msec) showed the spatial extent of the proximal current sink to be reduced by nearly 40% in ethanol-treated rats with little or no change in the peak amplitude or area. On the other hand, the distal current sink exhibited a somewhat expanded spatial extent and a significantly greater peak amplitude and area. Since our CSD analysis did not include measures of conductivity, these latter changes in the magnitude of excitatory current flow may only reflect group differences in tissue conductivity, perhaps arising from changes in tissue morphology. However, the facts that the I/0 functions (which reflect the combined SCH/COM input) did not differ across groups and the distal current sink did exhibit a slight enlargement combine to strongly suggest that an increase in synaptic efficacy among the afferents contributing to the distal current sink occurred in ethanol-treated rats in order to compensate for the loss of synaptic drive among synapses more proximal to stratum pyramidale. According to our working hypothesis, these results suggest that chronic ethanol con sumption leads to a selective reduction of COM afferents to CAl. This
PAGE 66
56 selective deafferentation may provide the stimulus for a compensatory enhancement of synaptic drive among SCH afferents, which could reflect sprouting and formation of additional synapses among SCH afferents (Goldowitz et~-, 1979) or an increase in the efficacy of existing synapses (Barnes, 1979}. It is difficult to specify the mechanisms underlying the relatively selective effects of chronic ethanol treatment on COM afferents. A recent experiment, employing retrograde double-labelling techniques, has strongly suggested that SCH and COM afferents to CAl arise as axon collaterals from the same population of cells in CA3 (Swanson et~-, 1980). Thus, it is unlikely that ethanol treatment selectively reduces a sub population of cells in CA3 which serves as the source of commissural afferents to CAl. In order for chronic ethanol treatment to selectively reduce COM afferents to CAl, apparently it must damage COM while sparing SCH axon co 11 a tera ls even though both may arise from the same pyramidal cell. This requirement would be satisfied if long-tract myelinated fibers were particularly sensitive to the toxic actions of ethanol. Such a possibility not only explains the destruction of myelinated COM fibers but also the relative sparing of the SCH axons which are only weakly myelinated in the rat (Andersen et~-, 1978). Interestingly, Marchiafava-Bignami disease, a somewhat rare complication of alcoholism in man, has been characterized by demyelination within the major hemispheric commissures, notably the corpus callosum and anterior commissure (Dreyfus, 1974). This hypothesis would predict that the commissural fibers to CA3 and the DG should be similarly reduced by ethanol treatment. Alternatively, ethanol treatment could act within CAl to selectively destroy axon terminals in the proximal portion of
PAGE 67
57 stratum radiatum. This hypothesis would be valid irrespective of the source of afferents contributing to the dual currents sinks. However, because ethanol most likely distributes uniformly throughout the brain (Sunahara et~-, 1978), the hypothesis would require that differences exist between the proximal and distal stratum radiatum in some morphological feature such as the extent of vascular supply or proximity to the ventricular system, neither of which presently provides a compelling explanation of our results (cf. Coyle, 1978). Finally, it is possible that both SCH and COM afferents are damaged by ethanol exposure but only the SCH fibers have the capacity for compensatory regeneration. Unfortunately, on the basis of existing evidence, it is not possible to choose between these many alternatives. The results of the present experiment coupled with our previous anatomical evidence indicate that chronic ethanol treatment in the presence of a nutritionally adequate diet induces a complex sequence of structural and functional changes in the hippocampus. Chronic ethanol treatment produces a loss of both hippocampal pyramidal and dentate gyrus granule cells (~Jalker et~-, 1980). Because of the intrinsic and commissural connections of the hippocampus, such cell loss produces partial deafferentation of neurons in the hippocampal formation, a result which could explain the reduction in dendritic spines and dendritic branching observed at comparable durations of ethanol treatment (Riley and 14a 1 ker, 1978; Mc Mu 11 en et ~, 1 980). The present study provides physiological evidence for deafferentation in CAl of the hippocampus, but deafferentation which is selective to the COM afferents. These findings agree with preliminary observations in Golgi material in which
PAGE 68
58 chronic ethanol treatment attenuated apical dendritic branching proximal to the CAl pyramidal cell layer (McMullen ettl., 1980). Deafferentation in the hippocampus is known to stimulate synaptic reorganization whose extent depends upon the specific afferents and the hippocampal subfield deafferented (Goldowitz et tl, 1979; Nadler et~., 1980a; Nadler et tl, 1980b). While we have noted a reduction of dendritic spines in CAl (Riley and Walker, 1978), shorter durations of ethanol treatment have been reported to increase dendritic spines in CAl (Kunz et i!]_., 1976}. The present results also suggest that compensatory synaptic reorganization may occur in association with deafferentation induced by chronic ethanol treatment. It is not yet clear whether the apparent compensatory increase in SCH afferent synaptic drive observed in this experiment began during the course of ethanol treatment or during the extended period of ethanol abstinence (eight weeks) prior to sampling. In either case, synaptic reorganization has important implications for studies of functional recovery from the cognitive impairment produced by chronic alcoholism in man. The available evidence now indicates that chronic ethanol treatment can induce numerous changes in the hippocampus including cell loss, deafferentation and synaptic reorganization. Further research will be required to understand the contribution of each of these changes to the cognitive impairments produced by chronic ethanol exposure.
PAGE 69
CHAPTER IV AUGMENTATION OF SHORT-TERM PLASTICITY IN CAl OF RAT HIPPOCAMPUS AFTER CHRONIC ETHANOL TREATMENT Introduction Chronic alcoholism is often associated with pathological deterioration in many organ systems including the CNS. The most severe CNS disorder, Wernicke-Korsakoff syndrome, is associated with widespread neuro pathology and a variety of neurological symptoms, the hallmark of which is a debilitating and permanent anterograde amnesia in the absence of a general cognitive decline (Courville, 1966; Talland, 1965; Victor et~-, 1971). This pathological deterioration has been attributed to several coexisting conditions, especially malnutrition and thiamine deficiency (Victor et~-, 1971). However, it seems likely that ethanol exerts direct neurotoxic effects in the CNS, since brain damage and neuro psychological deterioration have been observed in chronic alcoholic patients with no history of malnutrition, head trauma or exposure to other toxic agents (Epstein et~-, 1977; Haug, 1968; Smith et~-, 1973; Tumarkin et~-, 1955). Moreover, animal studies have shown that chronic ethanol exposure in the presence of a nutritionally adequate diet results in 1) retarded acquisition of a variety of behavioral tasks in rodents including shuttlebox avoidance (Freund and Walker, 1971b; Sotzing and Brown, 1976; Walker and Freund, 1971), DRL (Denoble and Begleiter, 1979; MacDonell and Marcuella, 1978; Walker and Hunter, 1978), go-no-go discrimination (Walker and Hunter, 1978) and complex maze 59
PAGE 70
60 acquisition (Fehr et~-, 1976), 2) a 15-20 % loss of granule cells in thedentateand pyramidal cells in CAl and CA2-4 in rat hippocampus (Walker et~-, 1980, and 3) spine loss and dendritic atrophy in granule cells and the basilar dendrites of CAl pyramidal cells of the rodent hippocampus (Riley and Walker, 1978). On the basis of these morphological results, physiological studies of hippocampal synaptic connections of rats chronically exposed to ethanol might reveal impaired function commensurate with the neuropathology. However, other outcomes are possible, particularly since deafferentation within the hippocampus leads to considerable reorganization of existing connections (Goldowitz et~-, 1979; Nadler et~-, 1980a; Nadler et~-, 1980b). The present study examined electrophysiologically the persistent effects of chronic ethanol consumption on synaptic function in stratum radiatum of CAl in the rat hippocampus. Repetitive stimulation of SCH/ COM afferents to CAl evokes large extracellular field potentials which are highly labile, exhibiting both short-and long-lasting changes in response to relatively brief tetanic stimuli (Alger and Teyler, 1976; Creager et~-, 1980; Dunwiddie and Lynch, 1978; Schwartzkroin and Wester, 1975). We have examined such plasticity in order to assess synaptic function within the hippocampus following chronic ethanol treatment. Methods The animals, treatment conditions and general electrophysiological methodology have been described previously in Chapter II. However, since certain details of the protocol differ between this study and
PAGE 71
61 the following one in the DG, procedures unique to this experiment are presented below The stimulating and recording electrodes were placed in the brain through small burr holes in the skull. The coordinates for the stimulating and recording electrodes were (relative to bregma): 3.2 mm posterior and 2.8 mm lateral, and 4.2 mm posterior and 3.0 rrm lateral, respectively. Once the stimulating electrode was placed to produce optimal activation of SCH/COM fibers (near the CA1-CA2 border), the recording electrode was stationed 125 m dorsal to the inversion point. Initially, I/0 curves were generated using currents ranging from 20-1000 a. The PPP was evaluated at low (EPSP amplitude 50% of that at PS threshold) and high (PS amplitude 40% of the asymptotic PS ampli tude) current intensities. The !Pis varied from 20-180 msec. Stimulus pairs were delivered at 0.1 Hz. A total of five FP series were con ducted at various stimulation frequencies and durations: l Hz/5 sec, 1 Hz/25 sec, 5 Hz/5 sec, 10 Hz/2.5 sec and 10 Hz/5 sec. Test stimuli (0.03 Hz) were then delivered posttetanus for 15-30 min to examine the amount of any LTP that may have been produced by the stimulation. When LTP was produced, the stimulation current was readjusted to give the standard PS response, 20% of asymptote, prior to proceeding with the next stimulus train. The final procedure was to stimulate at 100 Hz for 10 sec to insure the production of LTP. Responses to test stimuli were followed for 30 min posttetanus. Animals exhibiting afterdischarge posttetanus were not included in this analysis. Four EP measures were commonly assessed: EPSP onset latency, PS peak latency, EPSP amplitude and PS amplitude. Because the precise onset of the EPSP was difficult to gauge due to its shallow initial
PAGE 72
62 slope, the EPSP onset was arbitrarily chosen to be that point where the EP rose 200 v above baseline. In an attempt to avoid contamination by the PS, EPSP amplitude was measured 0.75 msec following EPSP onset. The PS peak latency was measured from EPSP onset. Finally, PS amplitude was assessed by averaging the amplitudes from the peak negativity to the preceding and following positive peaks (cf. Alger and Teyler, 1976). Although others have used only the amplitude from the first positive peak to the negative peak of the PS (e.g., Loma, 1971a) as the measure of PS amplitude, these two methods give quite similar results. For this reason, we chose the more conventional averaging method. Results The following results were based on a total sample of 29 rats which were distributed as follows: Group E = 12, Group S = 10 and Group LC= 7. Because Group Sand Group LC rats did not differ on any measure, these groups were combined into a single control group (Group C = 17). Group E animals consumed a mean daily ethanol dosage (14.14 g/kg) comparable to previous experiments in which either associative deficits or neuronal loss in hippocampus has been observed. Body weights did not differ among the three groups at any point during the experiment. Mean ( SEM) body weights at the end of the experiment were: Group E = 553.l 15.4 g, Group C = 519. 1 10.0 g. Histological analysis. Chronic ethanol treatment has been shown to produce cell loss in the rat hippocampus (Halker et~-, 1980). If one consequence of this cell loss is a selective shrinkage of the volume of the hippocampus, then the relative positions of electrodes placed within the hippocampus using fixed stereotaxic coordinates might be
PAGE 73
63 differentially altered in Group E. In order to address this issue, electrode tracts were localized and plotted in three dimensions relative to the brain surface, midline and septal pole of the hippocampus. This analysis revealed no differences among the three groups. Stimulating electrode sites were in stratum radiatum near the CA1-CA2 border. Recording electrode tracts were localized to dorsal hippocampal CAl (see Fig. 1-1). Electrolytic lesions placed at the inversion point of the laminar analysis were consistently found at the stratum pyramidale radiatum border. The location of the inversion point, with respect to the cell layer, also did not differ across groups. Finally, detailed measures (corrected for tissue shrinkage) were made of the major laminae of CAl including stratum oriens, pyramidale and radiatum. These measures were made along the trajectory of the electrode track and also did not differ between groups. I/0 relationships. Chronic ethanol treatment did not produce statistically significant alterations in the basic synaptic responses to single pulse stimulation (Fig 4-1). The threshold current (a) required to elicit an EPSP was: Group E = 61.7 3.8; Group C = 61.2 6 1 The PS threshold current values were also virtually identical across groups (Group E = 239.5 23.9; Group C = 251 .8 27.9) Moreover, the EPSP amplitude at PS threshold (Group E = 2.96 0 .39 mv; Group C = 2.91 0.34 mv) and the PS amplitude at asymptote (Group E = 5.78 1.04 mv; Group C = 6.23 0.62 mv) did not differ between groups. These results are important, since these values were used to standardize the stimulus currents used during subsequent potentiation series in each rat. Finally, 2-way ANOVA revealed no statistically significant group differences over the range of stimulus currents examined (20-1000 a) for latency
PAGE 74
Figure 4-1. I/0 curves (mean SEM) for ethanol-treated (filled symbols) and control animals (open symbols). The data are standardized by expressing values in terms of stimulus current steps from EPSP threshold (C) or PS threshold (A,B) as described in the text. A. Plots of EPSP onset latencies (circles) and PS peak latencies (squares) by stimulus current. While ethanol had no statistically significant effects, it did produce a small trend toward reduced PS and EPSP latencies. B. Plot of PS amplitude by stimulus current. Ethanol produced a small but nonsignificant trend toward reduced responses at high currents. C. EPSP peak amplitude at currents subthreshold for the production of a PS. Again, there were no significant treatment effects.
PAGE 75
70 70 A B 60 60 e EPSP i I I PS ALCOHOL 50 0 EPSP > 50 CONTROL D PS E ---, u w 0) w Cf) 0 u, :? 4 0 :::) 40 I:: >-_J Cl. u :? z w 30 <( 30 I<( _J <( w Cl. 20 20 10 10 4 0 80 120 f-LA 0 0 0 100 200 300 400 500 0 IOO 200 300 400 500 CURRENT (A ) CURRENT (A)
PAGE 76
66 to EPSP onset, latency to peak of the PS, EPSP amplitude or PS ampli tude. Paired-pulse potentiation. Chronic ethanol treatment failed to alter the pattern of response to paired-pulse stimulation. In both groups, facilitation of the test pulse peak EPSP amplitude (at current levels subthreshold for PS) was maximal at an interpulse interval of 30 msec, where potentiation 150-175 % of control was observed {Fig. 4-2B). In contrast, test pulse PS amplitudes were dramatically inhibited at these short !Pis, but exhibited facilitation at pulse intervals of 80 msec or greater (Fig. 4-2B). This differential action of paired stimuli on EPSP and PS amplitude is identical to that observed in pre vious studies of CAl and the DG (Creager et~-, 1980; Landfield et~-, 1978; Lomo, 1971b; Steward et~-, 1977). While the pattern of response to paired stimuli was preserved, chronic ethanol treatment did produce significant changes in the magnitude of PPP. Two-way ANOVA revealed a significant treatment effect for PS amplitude (F(l,39) = 6.58, p < .02) and a significant group X !PI interaction for the latency to PS (F(l ,39) = 2.47, p < .03). Subsequent t-tests revealed these effects to be maximal at 100-175 msec IPis. The chronic ethanol-induced en hancement of PPP is illustrated in Figure 4-2A in representative rats from ethanol and control groups Although there was a trend toward enhancement of the test pulse EPSP amplitude at short !Pis in ethanoltreated rats, the ANOVA was not statistically significant (F(l,44) = 1.30, p > .2). These effects could not be related to differences in the conditioning pulse current levels across groups. The standardization procedure designed to produce equivalent baseline responses across animals resulted in baseline EPSP (Group E = 1.23 0.17 mv; Group C =
PAGE 77
Figure 4-2. Paired-pulse potentiation in CAl after chronic ethanol treatment. A. Recordings from individual alcohol and control rats at varying i nterpulse intervals {!PI). For each rat, responses to the conditioning stimulus are superimposed. Calibration: 1.0 mv, 5.0 msec. B. Group comparisions (mean SEM) of PS and EPSP amplitude as a function of IP!. Note the d ifferent scales for EPSP and PS. Asterisks denote statistically significant (p < .05) group differences as indicated by student t-tests.
PAGE 78
A B 6 Q X C 0 .: vi u ~g u w 0 ::J ::::i a.. ::;;: <( (/) a.. _J 0 a: f z 0 u _J 0 I 0 u _J <{ 400 300 200 100 0 Con d I Test P u lse I P u lse I P l I 2 0 40 60 6 8 80 1 0 0 125 150 175 ~J~j\_~~ h E P S P PS Alcohol 0 E PSP D PS Control 0 Q 200 X C 0 ~ "' -Cl> "CJ 5 u w 0 150 => :J a.. ::;;:
PAGE 79
69 1.39 0.13 mv) and PS (Group E = 1.59 0.41 mv; Group C = 1.56 0.16 mv) amplitudes which did not differ across groups. Frequency potentiation. As was the case with PPP, chronic ethanol treatment failed to alter the frequency-dependent pattern of responses during FP. At l Hz, potentiation of the PS rapidly grew to asymptotic values (within four stimuli) .and remained at a stable level (400% of control) thereafter. Both 5 and 10 Hz stimulation produced more robust levels of potentiation (5 Hz> 10 Hz) including the development of multiple population spikes. However, FP at these frequencies was more unstable, exhibiting a definite waxing and waning which was clearly more related to the number of stimuli presented than to the frequency of stimulation (Fig. 4-3). Since both 5 and 10 Hz tetani lead to the development of multiple population spikes, separate analyses were con ducted on either the amplitude of the first spike alone or on the grand sum of all the spikes. Identical results were obtained with each measure. Only the data for the combined PS measures will be presented. Figure 4-3 compares l Hz, 5 Hz and 10 Hz FP across groups at identical numbers of stimuli (25 pulses). Chronic ethanol treatment did not influence FP at l Hz stimulation. However, a 2-way ANOVA revealed that FP was enhanced in ethanol-treated rats at 10 Hz (F{l,44) = 4.33, p < .05). Stimulation at 5 Hz produced intermediate results although the effect at this frequency did not reach statistical significance. Thus, chronic ethanol treatment augments FP at higher frequencies sampled over a relatively narrow range. No measures other than the amplitude of the PS were significantly influenced by ethanol treatment. Neither the latency to EPSP onset or to PS peak was affected. Since EPSP amplitude was measured at a fixed
PAGE 80
Figure 4-3. Frequency potentiation in CAl comparing (mean SEM) alcohol and control groups as a function of stimulus frequency (1, 5 and 10 Hz). A statistically significant treatment effect was observed only at 10 Hz. Intermediate effects were observed at 5 Hz, whereas no group differences were noted at l Hz.
PAGE 81
71 I H z 400 200 Al co ho l o Co ntrol 0 w 4 7 10 z 13 16 19 22 25 __J 800 w (/)
PAGE 82
72 latency (0.75 msec) from onset, this measure was often contaminated by the PS. Especially during FP, the measures of EPSP amplitude and PS amplitude were often inversely correlated. Frequency potentiation produced a dramatic reduction in PS latency, even to the point of obliterating the EPSP completely. We therefore abandoned the EPSP amplitude measure in both the FP and LTP series. Long-term potentiation. We investigated the long-term effects of repetitive stimulation by presenting six separate tetani which were systematically varied by frequency and duration. The long term effects of each of the stimulus trains on PS amplitude are presented in Figures 4-4 and 4-5. The pattern of long-term effects was dependent on both the frequency and duration of the tetanus. Stimulation at 1 Hz failed to produce LTP. Rather, when of sufficient duration {25 sec), 1 Hz stimulation produced a brief (1-2 min) depression of PS amplitude immediately following the tetanus (Fig. 4-4). Stimulation at 5, 10 and 100 Hz (Fig. 4-5) produced LTP. These results agree reasonably well with the parametric .i!!_ vitro study of Dunwiddie and Lynch (1978). Stimulation at 5, 10 (5 sec) and 100 Hz produced characteristic LTP of the PS which reached a peak at five minutes posttetanus. In contrast, 10 Hz (2.5 sec) produced a robust posttetanic potentiation but showed little evidence of LTP 15 minutes following tetanus (Fig. 4-5). These data suggest that a rather complicated interaction may exist between the frequency and duration of the tetanus in producing LTP in CAl of the rat hippocampus. Under the conditions of this experiment, considerable decay of LTP was observed by 30 minutes following even the l 00 Hz tetanus.
PAGE 83
Figure 4-4. Posttetanic depression of PS amplitude after low frequency stimulation. Response depression was observed only after 25 sec of l Hz stimulation. Ethanol treatment significantly reduced the magnitude of this posttetanic depression. Asterisks denote statistically significant group differences (p < .05).
PAGE 84
74 200 I Hz 5 sec. 150 Q) C Q) 1/) 0 .0 1/) 100 -----::::, C -0 -Q) I Q) '50 Alcohol a. o Control 0 0 w 0 I 0 20 40 60 t 2 5 10 15 => I_J I Hz 25 sec. Q_
PAGE 85
Figure 4-5. Long-term potentiation in alcohol and control groups compared (mean SEM) as a function of the frequency and duration of the tetanus. When the duration of the tetanus was varied at a constant frequency (10 Hz), strikingly different patterns of potentiation were observed in both groups. Ethanol treatment failed to significantly affect long-term potentiation at the frequencies tested.
PAGE 86
w 2 :J w a) (/) ::, 2 I300 200 100 0 I 600 a.. 12 w (..) er ::, 1-...J a.. cl'. (/) a.. 500 400 0 10 Hz 2 5 sec 20 5 Hz 20 40 sec 5 sec. 40 sec Alcohol O Control :_-_-__-Lj --~ 60 2 5 60 2 5 10 min 15 15 min. 30 TIME PO 300 200 100 0 600 500 400 300 200 10 Hz 5sec 20 40 sec 100 Hz 10 sec. 40 sec ST-TETANUS 60 60 2 5 min 15 m in 30
PAGE 87
77 Chronic ethanol treatment also produced effects which were depen dent on the frequency and duration of stimulation. The transient de pression of the PS induced by l Hz stimulation (25 sec) was significantly reduced in ethanol-treated animals (F(l ,44) = 4.33, p < .05). Subse quent t-tests revealed that this effect was present only at time points immediately following the tetanus (20-50 sec). Ethanol treatment failed to alter the pattern of LTP produced by 5, 10 or 100 Hz stimulation. Analyses of variance were not statistically significant for any of the results shown in Figure 4-5. However, at 100 Hz stimulation, chronic ethanol treatment, while failing to alter the development of LTP of the PS, nevertheless appeared to produce a trend toward ~ reduction in the magnitude of LTP starting approximately five minutes posttetanus (Fig. 4-5). Discussion We have used a nutritionally controlled liquid diet preparation to produce an animal model of chronic ethanol exposure. At comparable durations of exposure, chronic ethanol results in a 15-20 % loss of gran ule and pyramidal cells of the rat hippocampal formation (Walker et~ 1980) as well as dendritic atrophy and reduction of dendritic spines of CAl pyramidal cells and dentate granule cells of the mouse hippocampus (Riley and Walker, 1978). Taken together, these results suggest that chronic ethanol exposure induces a progressive morphological deterioration of neurons in the hippocampus ultimately ending in cell death. Despite these findings, 20 weeks of ethanol exposure failed to produce significant alterations in basic waveforms, EPSP and PS thresholds, I/0 functions or the production of LTP of CA3-derived fibers
PAGE 88
78 terminating in stratum radiatum of CAl. The predominant effect of chronic ethanol treatment was an enhancement of the potentiation of PS responses to paired stimuli or repetitive stimulation at 5 and 10 Hz in the absence of changes in the synaptic response (EPSP) when measurable. Even though a trend toward enhancement of EPSP potentiation was ob served during PPP (Fig. 4-28), this enhancement was only apparent at short IPis. Both ethanol and control groups reached a similar asymptote at longer IPis where ethanol treatment produced its greatest enhancement of PS amplitude. The existence of alterations of potentiation of the PS without concomitant changes in the EPSP suggests that ethanol did not act primarily on the SCH/COM-CAl synapses themselves. Rather, ethanol treatment appeared to enhance the ability of CAl pyramidal cells to fire synchronously in response to repetitive stimulation. While a number of hypotheses could account for these results, recent evidence indicates that recurrent inhibition plays a significant role in paired-pulse facilitation of the PS. Experimental treatments which directly reduce recurrent inhibition in the hippocampus (anoxia, picrotoxin and enkephalins) enhance the facilitation of PS amplitude to paired stimuli and in some cases directly facilitate PS responses to single shock stimuli (Andersen, 1960; Dunwiddie et~ 1980; Lee et~ 1980) without significantly altering the synaptic response. Thus, under ordinary conditions recurrent inhibition antagonizes paired-pulse facilitation within the hippocampus. This evidence, coupled with the present results, supports the view that chronic ethanol treatment reduces the effectiveness of recurrent feedback inhibition of pyramidal cells of CAl. Facilitation of the PS amplitude to paired stimuli was enhanced in
PAGE 89
79 ethanol-treated rats in the absence of significant effects on potentiation of the synaptic response. Further, chronic ethanol treatment enhanced FP of the PS only at those frequencies (5 and 10 Hz) in which recurrent inhibition would exert an influence. Spike responses to l Hz stimulus trains were identical across groups. Finally, ethanol treatment significantly reduced the posttetanic depression produced by l Hz stimulation (Fig. 4-4). Although ethanol treatment failed to influence the more brief depression produced by higher frequency stimulation (Fig. 4-5), it is noteworthy that such depression may have causes unrelated to recurrent inhibition, for example, transmitter depletion (Yamamoto et ~. l 98 O) Recurrent inhibition in the hippocampus is mediated by y-amino butyric acidergic (GABAergic) neurons which receive axon collaterals and project back onto the parent pyramidal and granule cells. These interneurons are believed to be basket cells (Andersen et~-, 1964a; Andersen et~-, 1964b; Lorente de No, 1934). Since relatively few basket cells exist in the hippocampus (relative to principle cell types), the powerful and widespread recurrent inhibition in the hippocampus must reflect a great divergence of synaptic contacts. Even a modest alteration in basket cells would be expected to produce effects which are more potent than the loss of principle cell types (\~alker et~., 1980). Thus, ethanol treatment could reduce recurrent inhibition by reducing the population of basket cells or otherwise altering GABAergic neuro transmission. Chronic ethanol treatment has been reported to decrease GABA concentrations by some (Patel and Lal, 1973; Volicer et~., 1977), but not all investigators (Sutton and Simmonds, 1973), and to reduce the density of low affinity GABA binding sites (Liljequist and Engel, 1979;
PAGE 90
80 Ticku, 1980; Ticku and Burch, 1980). Moreover, chronic ethanol treatment decreases the affinity and density of receptor binding for benzo diazepines (Freund, 1980), which appear to augment GABAergic inhibition in widespread brain areas. However, in many of these studies the duration of ethanol treatment was substantially shorter (2-3 weeks) than in the present study and GABAergic function was assessed within 24 hours of ethanol withdrawal. There is little doubt that these effects are de pendent upon the duration of ethanol exposure. For example, while 15-19 days of ethanol exposure failed to alter benzodiazepine receptor binding (Freund, 1980; Karobath et~-, 1980), seven months of exposure decreased the density of benzodiazepine receptors for at least one month following abstinence (Freund, 1980). Thus, it remains unclear whether longer durations of ethanol exposure would decrease GABAergic neurotransmission after eight weeks of ethanol abstinence, the time period assayed in this study. While reduced recurrent inhibition may be best explained by a direct action of chronic ethanol treatment on basket cells, at least one alternative should be considered. Anesthetic agents including ethanol have been reported to augment recurrent inhibition of single unit activity in the hippocampus (Newlin et~-, 1979; Tsuchiya and Fukushima, 1978; Wolf and Haas, 1977). Since ethanol exhibits crosstolerance to other depressants (Kalant et~-, 1971), the apparent reduction of recurrent inhibition could result from differential group responses to general anesthesia. This hypothesis must be seriously con sidered since chronic ethanol treatment has recently been shown to produce tolerance of the response to ethanol of SCH/COM-CAl synapses in a hippocampal slice preparation (Carlen and Corrigall, 1980). However,
PAGE 91
81 we consider this alternative unlikely for the following reasons. Neither the initial nor supplemental doses of urethane required to produce and maintain anesthetic levels differed between groups. Behavioral tolerance (Kalant et~-, 1971) to ethanol normally dissipates over a brief time course (2-3 days) and ethanol tolerance in hippocampal slices was only observed immediately upon ethanol withdrawal (Carlen and Corrigall, 1980). Since our data were collected at least eight weeks after ethanol withdrawal, it is likely that the residual effects of ethanol tolerance \AJould have dissipated. This possibility, while still unlikely, cannot be completely ruled out (cf. Begleiter et~., 1980). The effects of chronic ethanol treatment on PPP, FP and posttetanic depression are significant insofar as they persisted for at least eight weeks after ethanol abstinence. Nevertheless, the failure to observe significant alterations in basic synaptic responses (thresholds, I/0 curves) is surprising since both the pre-and postsynaptic elements of the SCH/COM-CAl path are reduced by chronic ethanol treatment (Halker et~., 1980). However, it is unlikely that a simple relationship exists between the number of CAl pyramidal cells and the size of the extracellular EPSP and PS. Further, it is plausible that the varied effects of ethanol treatment could interact. For example, the basket cells probably tonically inhibit the pyramidal cells (Alger and Nicoll, 1980). Picrotoxin, a GABA receptor blocker, has also been shown to facilitate PS responses to single shock stimuli (Dunwiddie et~-, 1980). Thus, it is possible that a loss of recurrent inhibition may have masked a reduction in PS responses (I/0 curves) produced by deafferentation of CAl. Alternatively, the failure to observe changes in
PAGE 92
82 basal synaptic response strength may reflect morphological reorganization and recovery, since reactive synaptogenesis has been documented in CAl following destruction of afferents originating in CA3 (Nadler et~-, 1980a; Nadler et~-, 1980b). At present, a connection between LTP in the hippocampusand normal memory formation is clearly speculative. However, the issue may be addressed through correlative analysis of synaptic potentiation in memory deficient animals. Studies in aged, memory deficient rats have indicated that LTP is reduced in both CAl (Landfield et~-, 1978) and the dentate gyrus (Barnes, 1979) of the hippocampus. The similarity of the mnemonic deficits associated with alcoholism and aging have led to the hypothesis that alcohol accelerates the aging process (Beck et~-, 1979; Ryan and Butters, 1980). Studies of chronic ethanoltreated and aged animals indicate a reasonable similarity in the nature of morphological deterioration in the hippocampus (Bondareff, 1979; Brizzee and Ordy, 1979; Scheibel, 1979). However, aged animals exhibit a reduction in FP and LTP in SCH/COM-CAl connections (landfield et~-, 1978) whereas ethanol-treated rats exhibit enhanced PPP and FP. Insofar as the hippocampus is involved in memory formation {Milner et~-, 1968; O'Keefe and Nadel, 1978), these results suggest that very different mechanisms may underlie the mnemonic deficits associated with alcoholism and aging. This conclusion must be considered preliminary, particularly since a trend toward a greater decay of LTP was observed in ethanoltreated rats (Fig. 4-5), and a recent study with hippocampal slices has reported an impairment of LTP formation after chronic ethanol treatment assayed immediately following ethanol withdrawal (Durand~~-, 1980).
PAGE 93
CHAPTER V ELECTROPHYSIOLOGICAL ANALYSIS OF CHRONIC ETHANOL NEUROTOXICITY IN THE DENTATE GYRUS: ENTORHINAL AFFERENTS Introduction Long-term consumption of ethanol leads to numerous neuropsychologi cal impairments in both humans and laboratory animals. Perhaps the hallmark of the cognitive dysfunction is a profound impairment in the subject's ability to acquire and store new information (Talland, 1965; Victor et~-, 1971; Walker et~-, 1981). The debilitating nature of the alcoholic syndrome has prompted many researchers to investigate the loci and nature of ethanol-related neurotoxicity in the brain. The neuropathology in human Korsakoff patients centers around the dorso medial nucleus of the thalamus and the mammillary bodies, with other brain regions more variably affected (Mair et ~-, 1979; Victor et ~-, 1971). However, studies in laboratory animals, through careful control of nutrition, genetics, amount and duration of ethanol consumption and other environmental variables, should be better able to localize those brain regions particularly sensitive to ethanol-induced neurotoxicity. Using the liquid diet procedure of Walker and Freund (1971), our laboratory has demonstrated considerable neuropathology in ethanolconsuming rats and mice. Cerebellar vermal Purkinje and granule cells are reduced in number by 15-20 % following five months of ethanol treatment (Walker et~-, 1981). In addition, surviving Purkinje cells show a 25-30 % reduction in dendritic arborization (Riley and Walker, 1981). 83
PAGE 94
84 Importantly for this study, similar pathology has been demonstrated in the hippocampus. There is approximately a 15% loss of both hippocampal pyramidal and granule cells (Walker et tl-, 1980) in addition to a reduction in spine density along both pyramidal basilar dendrites and granule cell dendrites (Riley and Walker, 1978). It should be noted that the histological analyses were performed at least one month follow ing ethanol withdrawal, implying that these are lasting changes in CNS morphology. The above histological data suggest that chronic ethanol treatment produces deleterious effects on hippocampal cytoarchitecture. It is imperative that we correlate these findings with electrophysiological analyses of synaptic and cellular activity if we are to understand how chronic ethanol exposure affects hippocampal function. Previously (Chapter III), we electrophysiologically confirmed a shrinkage of the SCH/COM terminal field in CAl of ethanol-treated rats (20 weeks on ethanol, eight weeks off). Although these afferents still responded normally to single-shock stimulation, they did show evidence of an enhanced paired pulse and frequency potentiation of the extracellularly recorded PS (Chapter IV). No changes in LTP were observed. The present study assessed the ubiquity of these alterations by analyzing the functional integrity of the angular bundle-DG synaptic connections. The angular bundle (AB) fibers originate in the medial and lateral entorhinal cortices and distribute topographically onto the outer 2/3 of the granule cell dendrites of both dorsal and ventral blades (McNaughton and Barnes, 1977; Steward and Scoville, 1976). Excitatory synapses are formed by boutons en passage. As in CAl, intra-and extracellular recordings have shown that these synapses can exhibit a wide range of potentiation
PAGE 95
85 phenomena, including PPP, FP and LTP (Bliss and Gardner-Medwin, 1973; Bliss and Lomo, 1973; Lomo, 1971b; ~Jhite et~-, 1979), in response to repetitive activation. The present study examines with extracellular recording techniques the effects of ethanol on the laminar distribution of the AB-DG synapses as well as their responsivity to single and repetitive stimuli. Methods The animals, treatment conditions and general electrophysiological methodology have been described earlier in Chapter II. As before, rats were maintained on liquid diets for 20 weeks and withdrawn for eight weeks prior to electrophysiological study. The major procedural alteration was that, since Groups Sand LC have rarely differed on a variety of histological, behavioral and electrophysiological measures in numerous experiments, Group LC was dropped from this experiment. Certain of the physiological characteristics of the AB-DG synapses differ from those seen in CAl; therefore, the experimental protocol of the earlier experiments required modification. The electrode placements were based on the following stereotaxic coordinates: recording electrode (tip broken back to a 3-5 m diameter), 3.8 mm posterior to bregma, 2.4 mm lateral; stimulating electrode, 8.1 mm posterior to bregma, 5.0 mm lateral. The recording electrode was initially positioned in the dentate hilus and used to guide the stimulating electrode placement by monitoring the PS amplitude and threshold. The recording electrode was then moved to a fixed (except during the laminar analysis) recording site 150 m ventral to the dorsal blade inversion. At the end of a recording session, lesions were made through
PAGE 96
86 the recording electrode at the inversion points of both the dorsal and ventral blades, approximately 650 m apart. In this experiment, the I/0 curve (10-1000 a) was taken prior to the laminar analysis. During the laminar analysis, the recording electrode was stepped in 25 m increments completely across both blades of the DG. Eight responses at 0.1 Hz were recorded for off-line averaging and analysis by a PDP 11/40 laboratory computer. The electrode was then returned to the fixed recording site. At this point, PPP (15-175 msec IPI) was evaluated at both low (EPSP amplitude 25% of that at PS threshold) and high (PS amplitude 50% of asymptote) stimulus intensities. Stimulus pairs were given at 0.03 Hz. Frequency potentiation was examined with the following tetani: 1 Hz/25 sec, 5 Hz/5 sec and 10 Hz/5 sec at a current that produced a PS 50% of asymptote. Posttetanus test stimuli were given only for 10 min following each tetanus since these frequencies and durations of stimulation did not produce LTP in the DG. To produce LTP, 10 stimulus trains (400 Hz, 20 msec duration) were ap plied at a frequency of one train per minute. This procedure has been shown to produce robust LTP in the DG (Douglas, 1977). Test pulses were given 25 and 50 sec following each tetanus and the responses averaged to provide an indication of the development of LTP. Following the tenth train, test pulses were given (0.05-0.03 Hz) for the next 30 min. Results This experiment employed a total of 17 rats which were distributed among the groups as follows: Group E = 9; Group S = 8. As mentioned previously, a lab chow control group was not included in the study. Group E rats consumed a mean daily ethanol dosage of 13.11 g/kg, which
PAGE 97
87 is comparable to the consumption observed not only in the previous electrophysiological studies but also in the studies demonstrating either associative deficits or neuronal loss in the hippocampus. The ethanol rats showed normal weight gain throughout the treatment period and did not differ from sucrose rats in weight at the time of electrophysiological study: Group E = 525.2 34.5 g; Group S = 497.1 18.9 g. Histological analysis. As in the previous experiments (Chapters III and IV), there was some concern that the ethanol treatment may have produced sufficient changes in brain volume that the use of fixed stereotaxic coordinates would have resulted in differential electrode placement between the groups. To check for this possibility, both the recording and stimulating electrode sites were plotted with respect to both the septal pole of the hippocampus and midline. This analysis revealed that there were no group differences in gross placement of the electrodes. Stimulation electrode tips were localized to the AB, just posterior to the hippocampus and approximately 4 mm posterior to the recording sites. The recording electrode lesions were found in the dorsolateral DG, ap proximately 50-70 m (25-35 % of the total stratum moleculare) distal to the granule cell layer of each blade (Fig. 5-1). Detailed measures of the major DG laminae were made along the ac tual electrode trajectory. The use of two recording electrode lesions spaced a known distance apart (based on microdrive readings) allowed us to correct for the tissue shrinkage that normally occurs during histological fixation and celloidin embedding. As in CAl (Chapter III), there were no differences between the groups in laminae widths except in the dorsal blade where Group E actually showed a significantly larger (t = 4.57, p < .01) molecular layer (Table 5-1 ). For Group Sonly, the
PAGE 98
Figure 5-1. Photomicrographs of the dorsal hippocampal dentate gyrus (DG) where recordings for the present experiment were made. Arrows indicate the lesions that were made through the recording electrode at the inversion point for the dor sal blade (top half of figure) and the ventral blade (bottom half of figure) as produced by AB stimulation. Note that in each blade the inversion point is in the proximal portion of the molecular layer. The sections are 30 m thick and stained with Weil myelin stain.
PAGE 99
89
PAGE 100
90 TABLE 5-1 Chronic Ethanol Effects on Widths (m) of DG Laminaea Lamina Sucrose Ethanol t,. t..% DSM width 185.7 3.0 228.4 8.1 42.7 23.0b DSG width 55.0 4.4 69.4 6.3 14 4 26.2 HIL width 409.8 19. 6 361 .3 36.1 -48.5 -11.8 VSG width 62.9 5.5 65.8 5. 1 2.9 4.6 VSM width 228.4 15 .4 255.6 18.5 27.2 11. 9 aMean SEM bp < .01 student's t-test Abbreviations: DSM dorsal stratum molecular; DSG, dorsal stratum granulosu m ; HIL, hilus; VSG, ventral stratum granulosum; VSM, ventral stratum moleculare
PAGE 101
91 ventral blade was significantly wider than the dorsal blade (paired t = 2.79, p < .05). Since these layer measurements were not made along an axis perfectly perpendicular to the layer orientation nor in anatomically matched sections, group differences or similarities are still open to question. Resolution of this point requires a more careful histological evaluation; such an investigation is currently being conducted in our laboratory. Due to some improper electrode trajectories and an inability to localize certain recording electrode lesions, the group sizes for the above histological analysis and the following EP and CSD analyses were reduced to six and five for Group E and Group S, respectively. Field potential and CSD analysis. Low current stimulation of the AB produced positive-negative dipole fields in the DG. The left panel of Figure 5-2 illustrates these relationships in both dorsal and ventral blades for a sucrose animal. Negative potentials were recorded in the outer 2/3-3/4 of stratum molecular, reflecting summed excitatory synaptic activity (sinks). The anatomical distribution of the negative EP region corresponds well with the known distribution of entorhinal afferents within the DG molecular layer (Hjorth-Simonsen and Jeune, 1972; Steward et~., 1977). The supragranular and granular cell layers exhibit positive potentials, generated by outward capacitative currents (sources). The right panel of Figure 5-2 provides the same data after one-dimen sional CSD analysis. This procedure gives more spatial and temporal resolution of the current sources and sinks. The spatial distributions of the field potential and current densities are better appreciated when amplitude profiles are generated at fixed latencies from stimulus onset. The choice of a fixed latency is appropriate only if there are no group differences in, for example, the
PAGE 102
Figure 5-2. Full one-dimensional laminar analysis through both blades of the DG in a single group S animal. Evoked potentials are presented on the left and the corresponding CSDs are presented on the right. Depths indicate distance from the distal edge of the molecular layer. The stratum moleculare/stratum granulosum border is between 200-250 m for each blade. Stimulation was applied at time Tl but the artifact has been suppressed in this figure. T2 (4.0 msec from stimulus in the dorsal blade and 4.7 msec from stimu lus in the ventral blade) indicates the slice of time de picted in Figure 5-3. The calibration pulse on the EP waveforms is 2 mv, 2 msec. For the CSD waveforms, an upward deflection from baseline represents a source; a downward deflection represents a sink.
PAGE 103
w 0 <( ....J al ....J <( en a:: 0 0 0 w 0 <( ....J al ....J ,---, <( J \ a:: } \ z w n_ > \ 0 93 EP T l TZ DEPTH ( .m) Tl T2 CSD ~-~-0 50 v100 150.,...,, 200 8.0 16.0 24. 0 0 8 0 16.0 200 ---+-~ 150 ---~--100 --50 0 8.0 16. 0 24.0 + 0 8 0 16.0 Time (msec) 2mv\4oomv1mm2 Time (msec) 24.0 24. 0
PAGE 104
94 latency to peak for the largest amplitude waveforms. Table 5-3 demonstrates that ethanol did not significantly affect peak latencies in either blade for EP or CSD data, although there was a trend in each case for longer latencies in Group E. Table 5-2 shows, however, that for both EPs and CSDs there were differences between the dorsal and ventral blades within groups. Paired t-tests revealed that the latency to 10% of peak amplitude, latency to 30% of peak amplitude and peak latency were all significantly longer in the ventral blade than in the dorsal blade by 0.4-0.9 msec. Only in Group E evoked potentials was this effect not observed. Further analysis of the CSD data showed the 10-90 risetime to be slower in the ventral blade than the dorsal blade. The field potentials did not show any differences in risetime. Based on these observations, the amplitude profiles described below were generated at a 4.0 msec latency in the dorsal blade and a 4.7 msec latency in the ventral blade. Figure 5-3 plots in both blades the average EP and CSD profiles generated at the above-mentioned latencies. Schematic granule cells have been added to facilitate comparisons with the DG cytoarchitecture. Clearly, the CSD analysis localizes spatially the current sources and sinks much more precisely than the simple EP analysis. In both groups, a large current sink occupies most of the outer 2/3 of the molecular layer and is bounded more distally by a small current source and more proximally by a larger current source. The neural basis of the small distal source is unclear at present but may reflect a reduced synaptic density at the distal tips of the granule cells' dendrites. Group comparisons of the spatial distributions of the field potentials and current densities required normalization of each animal 1s data
PAGE 105
95 TABLE 5-2 Latency Differences Between Dorsal and Ventral Blades for EP and CSD Waveformsa Measure Dorsal Blade Ventral Blade Paired t Sucrose EP Latency to 10% 2.24 0.08 2.72.11 2.81 Latency to 30% 2.58 0.07 3.11 0. 08 3.34 Peak Latency 4.09 0.09 4.52 0.17 3 .13 10-90 Risetime 1.22 0.05 1.26.13 0.60 Ethanol EP Latency to 1 0% 2.29 0.08 2.90 0.27 2.55 Latency to 30% 2.76 0.09 3.22 0.24 1. 92 Peak Latency 4.77 0.37 4.66 0.29 0. 21 10-90 Risetime l. 75 0. 21 1.33 0.09 1.57 Sucrose CSD Latency to 10% 2.18 0.05 2.59 0.08 4.49 Latency to 30% 2.58 0.07 3.04 0.10 4.14 Peak Latency 4.15.13 4.77 0.24 2.46 10-90 Risetime 1.39 0.07 1.57 0.09 3.50 Ethanol CSD Latency to 10% 2.28 0.06 2.65 0.20 2.58 Latency to 30% 2.65 0.10 3.14 0.19 4.40 Peak Latency 4.30 0.28 5.19 0.25 6.02 10-90 Risetime 1 38 0. 12 1.85 0.06 4.24 aMean SEM p <.05 <.01 <.05 NS <.05 NS NS NS <. 01 <.02 <.05 <. 01 <.05 <. 01 <.01 <. 01
PAGE 106
Figure 5-3. Group laminar profiles of normalized EPs and CSDs at 4.0 msec (dorsal blade) and 4.7 (ventral blade) after stimulus delivery. The top half of the figure presents the dorsal blade data; the bottom half presents the ventral blade data. Depths indicate the distance from the pia or hippocampal fissure in microns The left-hand panel schematically depicts granule cells drawn to scale and the major dentate laminae. The dotted lines represent the SM/SG borders. The normalization procedure is described in the text. Of the EP profiles (middle panel) no group effects were observed in the dorsal blade but the ethanol group shows more positive values in the ventral blade. The CSD analysis (right-hand panel) more precisely details the group differences. In both blades, the SM sink (negative I) is reduced in the outer part of the layer in the ethanol group which shows a significantly reduced wink width and a shorter distance from cell layer to sink peak. Simple main effects (p < .05) are indicated by asterisks. Abbreviations: DH, dentate hilus; HF, hippocampal fissure; P, pia; sg, stratum granu losum; sm, stratum moleculare.
PAGE 107
DISTANCE(UM} HF 0 0 50 50 w o S M 0 0 p -20 0 2 0 40 -80 -60 -40 -20 0 4 0 6 0 AMPLITUDE (mv) SINK SOURCE AMPLITUDE ( 1 : mv /mm2 xl02)
PAGE 108
98 since the absolute measures of the major laminae varied across animals. The basic normalization procedure was very similar to that used in Chapter III. Separate normalizations were performed for the dorsal and ventral blades. Profiles at the latency appropriate for each blade were generated for individual animals. The distance across the stratum moleculare plus stratum granulosum was divided into 8 % segments. Each 8 % step represented 21.8 min the dorsal blade and 24. 6 min the ventral blade. The amplitude of the EP and CSD was then obtained at these percentage increments using linear interpolation between actual data points through use of a Numonics Digitizer. The data in Figure 5-3 have been normulized and averaged by group. The dorsal blade EP profile appears rather unaltered by ethanol treatment. In contrast, Group E field potentials are uniformly more positive than those of Group Sin the ventral blade. The overall 2-way ANOVA confirmed a significant group main effect in the ventral blade (F(l ,12) = 6.87, p < .05). The results of individual depth comparisons are de picted in the figure (p < .05). The CSD analysis permits a more accurate comparison of sources and sinks between groups. As seen in the righthand panel of Figure 5-3, the CSD distributions over each DG blade are remarkably similar for each particular group. The overall 2-way ANOVA showed a significant group main effect (F(l ,12) = 8.15, p < .02) and group X depth interaction {F{l,12) = 3.59, p < .05) in the dorsal blade and similarly, a nearly significant group main effect (F(l ,12) = 3.70, p = 0.086) and significant group X depth interaction (F(l,12) = 2.04, p < .05) in the ventral blade. Individual comparisons indicated that the group differences in both blades concentrated in the outer molecular layer where the Group S sink was larger and closer to the molecular layer distal border than the Group E sink which had apparently shrunk.
PAGE 109
99 Particular characteristics of the raw waveforms were also individually quantified to clarify the nature of the ethanol effect. The waveforms were analyzed along three dimensions: spatial (sink half-width, distance from cell layer to peak negativity and distance from cell layer to proximal inversion point), magnitude {sink peak amplitude) and temporal (sink peak latency, 10% -90 % risetime and half-width). The distance measures were normalized by expressing them as percent of the total stratum moleculare width. As depicted in Table 5-3, none of these measures of field potentials were affected by ethanol treatment with the exception of ventral blade spatial half-width and distance from cell layer to peak negativity (p < .05). The ethanol effect is more apparent in the CSD distributions. Here, both the dorsal and ventral blades demonstrate a decreased spatial extent of the sink and a shorter distance from cell layer to peak negativity (Table 5-4). The only other statistically significant comparision indicated that Group E had a shorter 10-90 risetime in the ventral blade. There were no treatment-related changes in the maximum amplitudes of the sink. I/0 relationships. In general, chronic ethanol treatment failed to alter the field pot entials evoked by single-shock AB stimulation. The EPSP thresholds usually were less than the minimum current applied (10 a) and thus could not be determined. However, neither PS thresholds (Group E = 146.7 23.1 a; Group S = 120.0 29.7 a), EPSP amplitudes at PS thresholds (Group E = 5.2 0.5 mv; Group S = 5.0 0.4 mv) nor PS amplitudes at asymptote (Group E = 14.6 1.8 mv; Group S = 17.8 1.8 mv) significantly differed between the groups {data presented as mean SEM). Two-way ANOVAs also revealed no significant differences in EPSP onset latency, EPSP slope, PS peak latency or PS amplitude.
PAGE 110
TABLE 5-3 Laminar Analysis Quantificationa Chronic Ethanol Effects on DG Evoked Potential Distribution Dorsal Blade (4.0 msec) Ventral Blade (4.7 msec) M easure Sucrose Ethanol 6% Sucrose Ethanol 6% Sink Half-width -34.ob % SM width 59.5 6.4 48.3 8.0 -18.8 40. 3 3.2 26.6 4.6 SG-Peak Negativity --' -20.3b 0 % SM width 58.7 4.8 49.3 6.3 -16.0 56.8 1.8 45.3 4.1 0 SG-Proximal IV % SM width 23. 1 7.4 21.8 3.6 -6.6 28.5 1.6 32.6 4.9 14.4 Sink Peak Negativity mv -2.83 0.52 -2.97 0.44 4.9 -2 .11 0.25 -1. 37 0.40 -35 .1 Sink Peak Latency mv 4 .09 0.09 4. 77 0.37 7.9 4.52 0.17 4.66 0.29 1.6 10-90 Risetime msec 1.23 0.05 1. 75 0. 21 42.3 1 .38 0.13 1.33 0.09 -3.6 Half-width msec 2.93 0.34 4.72 0.75 61. l 3.06 0.38 2.67 0.36 -12.7 gMean SEM p < .05 student's t -test Abbreviations : SM, stratum moleculare; SG, stratum granulosum; IV, inversion point
PAGE 111
TABLE 5-4 Laminar Analysis Quantificationa Chronic Ethanol Effects on DG Current-Source Density Distribution Dorsal Blade (4.0 msec) Ventral Blade (4. 7 msec) Measure Sucrose Ethanol 11% Sucrose Ethanol 11% Sink ~~idth -19.5b -17.3b % S M width 71.5 1.5 57.6 6.0 61.4 4.2 50.8 3.5 SG-Peak Negativity -26.4b -21 7b % SM width 58.8 2.9 43.3 1.8 53.0 1.5 41.5 4. l ...... 0 ...... SG-Proximal IV % SM width 21.8 3.7 15 .8 2.4 -27.5 22.l 1.5 19. 7 2.0 -10. l Sink Pe~k Amplitude mv/mm -697.4 113.8 -830.4 143.8 19. l -837.0 57.8 -834. l 115. 9 -0.01 Sink Peak Latency msec 4.15 0.13 4.30 0.28 l. 7 4. 77 0.24 5.19 0.25 4.5 l 0-90 Ri setime 18.ob msec l. 39 0.007 l. 38 0. 12 -0. 01 l .57 0.09 l .85 0.06 Half-width msec 3.92 0 .17 3.76 0.42 -4.1 3.83 0.26 4.26 0.33 11. l a Mean SEM bp < .05 student's t-test Abbreviations: SM, stratum moleculare; SG, stratum granulosum; IV, inversion point
PAGE 112
102 Figure 5-4 does indicate a trend toward a shorter PS latency in the ethanol group. Since we successfully obtained an EPSP amplitude measure in this experiment, we performed a transformation of the PS amplitudes by dividing them by the EPSP slope to give a ratio of PS amplitude per unit EPSP slope (PS/EPSP). The resulting data give one an assessment of the relative excitability of the studied neuronal population, in this case the DG granule cells. The data are depicted in Figure 5-4 with current normalized by a increments from the PS threshold current. A 2-way ANOVA indicated a significant group X current interaction (F(l ,18) = 1 .62, p = .05). Figure 5-4 illustrates those current strengths where there were simple main effects. Group E displayed reduced neural excitability (20%), particularly at the higher stimulus currents which produced asymptotic responses. Paired-pulse potentiation. As seen in Figure 5-5 and confirmed by ANOVA, chronic ethanol exerted no effects on PPP in the dentate. Figure 5-5A illustrates PPP of the peak synaptic potential when low current intensities were used at pulse intervals ranging from 15-175 msec. Both groups showed 30-35 % potentiation at the shortest IPI but a progressive decay back to baseline by 125 msec IPL Figure 5-5B also shows no group differences in PS amplitude during PPP. In both groups, the PS was de pressed at the shortest intervals but rapidly potentiated to 50-70 % of control before decaying back to baseline at 175 msec IPI. Due to meas urement difficulties, the EPSP at high current was too variable a measure on which to perform meaningful analyses. The general pattern of de pression and potentiation of the test PS was similar to that described elsewhere (Assaf and Miller, 1978; Lomo, 1971b; Steward et~-, 1977; L~hite et tl, 1979). Despite potentiation of the PS amplitude, the PS
PAGE 113
Figure 5-4. I/0 curves (mean SEM) for ethanol-treated (filled circles) and control animals (open circles). The data are standardized by expressing the values in terms of stimulus current steps from PS threshold. A. PS peak latency. Chronic ethanol produced a strong but nonsignificant trend toward a reduced PS latency, particularly at higher current intensities. B. Ratio of PS amplitude per unit EPSP amplitude. Group E shows a 20% reduction of this measure at the higher current intensities. Asterisks indicate significant simple main effects (p < .05) as revealed by student t-tests.
PAGE 114
l 04 3.2 A -u Q) (/) E 3.0 -(.) 2 6 z w I
PAGE 115
Figure 5-5. Paired-pulse potentiation (percentage of condition ing pulse values) in the dentate gyrus after chronic ethanol treatment. A. Group comparisons (mean SEM) for EPSP peak amplitude by interpulse interval (IPI). The responses were obtained using a stimulus that produced an EPSP 25% of that at PS threshold. B. Group comparisons (mean SEM) for PS amplitude by IPI. Stimulation produced a conditioning spike 50% of that at asymptote. Response depression is seen at short IPis but potentiation rapidly develops. There were no significant treatment effects for either measure. Dotted line indicates baseline values (100%).
PAGE 116
l 06 A ,....... 0 140 0 >
PAGE 117
107 peak latency was lengthened at the same IPis. However, no treatment effects were observed for this measure either. Frequency potentiation. Frequency potentiation and depression were examined over 25 pulses given at each of three frequencies: 1, 5 and 10 Hz. The left half of Figure 5-6 plots the PS amplitude (% of baseline) for every third response. Stimulation at l Hz produced a response depression that reached asymptote at about 50% of baseline values. Such depression has been likened to the habituation described in the spinal cord (Harris et~., 1979a; Teyler and Alger, 1976; Thompson and Spencer, 1966) Stimulation at both 5 and 10 Hz produced a short initial period of spike depression followed by spike enhancement to 200-250% of baseline values. Ethanol had no statistically significant effect on the PS ampli tude at any of the frequencies used. Although a trend toward response suppression is apparent at 5 Hz for Group E, the reverse is true at 1 and 10 Hz. The discrepancies in these trends make it unlikely that the fairly large variances inherent in these data are masking a consistent treatment effect. The other typically assessed EPSP and PS measures also failed to exhibit an ethanol effect. In both groups, EPSP slopes were reduced throughout the tetani, leading to dramatically potentiated PS/EPSP ratios at the higher frequencies. Long-term potentiation. Following each of the FP series, low frequency stimuli were applied for 10 min to test for any long-term changes in synaptic efficacy. The right half of Figure 5-6 reveals that LTP was not produced by any of the stimulus conditions. It should be noted that this was true even for the 10 Hz tetanus which actually was given for 5 sec {50 pulses). Once again, there were no ethanol-related effects on any of the response measures examined. Within a minute following all tetani, the evoked responses essentially reverted to baseline values.
PAGE 118
Figure 5-6. Chronic ethanol effects (mean SEM) on dentate gyrus PS responses to low frequency stimulation (1, 5 and 10 Hz). The data are expressed as percentage of baseline values, which are indicated by dotted lines (100%). The left half of each plot shows the PS amplitude response to every third stimulus within each 25 pulse train. Inhibition (habituation) only is observed at 1 Hz. Depression followed by potentiation is seen at 5 and 10 Hz. The right half of the figure illustrates PS responses to test pulses applied infrequently for 10 min posttetanus. No LTP was observed at any frequency. The ethanol regimen had no effects on the depression or potentiation at any frequency.
PAGE 119
109 120 I Hz 100 80 60 Alcohol w 40 o Sucrose z ...J w Cf) <:t: 7 13 19 25 0 2 5 10 en Cf) ::J 5 Hz z 300 <:t: fw fI w a::: a.. 200 LL 0 0 ---~ :i: iw 100 0 ::J f-...J a.. <:t: 7 13 19 25 0 2 5 10 Cf) a.. 10 Hz 300 200 100 ----~----~I 7 1 3 19 25 0 2 5 10 STIMULUS NUMBER WITHIN TETANUS TIME POST-TETANUS (min)
PAGE 120
110 To insure the production of LTP, 10 high frequency, short duration tetani were given at a frequency of one per minute. Responses were sampled in between tetani to monitor the development of LTP. While considerable potentiation was produced after just one tetanus, succeeding tetani added smaller increments of potentiation until an asymptote was reached after six to eight tetani. Potentiation is evidenced by a de creased EPSP latency, decreased PS latency, increased EPSP slope and increased P S amplitude. On none of these measures did ethanol have a significant influence, although there was a trend for Group E to have longer PS latencies and smaller PS amplitudes (data not shown). These same trends persisted throughout the 30 min posttetani period when test pulses were applied every 30 sec. During this time, significant declines (p < .0001) occurred in nearly all potentiated measures although they did not return completely to baseline. This decay of a maximally potentiated state to a more enduring, but somewhat less potentiated state has been documented previously (Barnes, 1979; Goddard et~-, 1980). Ethanol had no statistically significant influence on the decay and stabilization of the response. However, we also analyzed the influence of ethanol on the "excitability" measure, PS/EPSP (Fig. 5-7). Here we see a strong trend for a group main effect (F(l ,17) = 3.68, p < .08) during LTP buildup. Ethanol apparently produced retarded development of LTP of neuronal excitability. This trend persisted throughout the posttetani period. Interestingly, this was the only measure which did not show a decay of potentiation over time (F(l ,9) = 1.02, p > .4).
PAGE 121
Figure 5-7. Chronic ethanol effects (mean SEM) on the development and maintenance of LTP. LTP progressively developed over a series of 10 high frequency, short duration tetani (details described in text). The left half of this figure plots the PS/EPSP ratio av eraged over the two responses taken 25 and 50 sec posttetanus. Group E (filled circles) shows a nearly significant (p < .08) delayed development of LTP. The EPSP and PS measures individually were unaffected (data not shown). LTP of PS/EPSP was maintained over the 30 min test period (right half of figure) unlike other potentiated measures which showed some decline with time. The group differences remained during this period but again did not reach statistical significance (p > .1). Dotted line represents baseline value of the PS/EPSP measure (100% ).
PAGE 122
250 225 wW a_Z o-_J _J Cf) w 200 Cf) a_
PAGE 123
113 Discussion We have used a nutritionally adequate liquid diet to administer large doses of ethanol to rats. This procedure produced a daily ethanol consumption very similar to that observed in previous studies that have demonstrated behavioral and hippocampal morphological deterioration in ethanol-consuming animals (Freund and Walker, 1971b; Walker and Hunter, 1978; ~folker et tl-, 198 0). The present electrophysiological study pro vides compleme ntary evidence that chronic ethanol ingestion deleteriously affects both synaptic distribution and synaptic function in the hippocampal dentate gyrus. We departed from the usual experimental paradigm by not including a lab chow control group. Errors resulting from this procedural change should be of a conservative Type 2 nature since, although Groups Sand LC have not differed on any behavioral or morphologi cal measures to date, Group S has generally lain between Group E and Group LC (Freund and Halker, 1971b; Riley and Walker, 1978; \valker and Freund, 1973; Walker et ~, 1981 ) The laminar analyses were conducted to explore the effects of ethanol on the synaptic distribution of the entorhinal afferents in the DG molecular layer. It is important to remember that all laminar profile data (EP and CSD) were normalized with respect to the individual molecular layer widths to control for within-group variability in layer widths as well as possible between-group layer effects. Thus, alterations in synaptic fields cannot be simply attributed to gross tissue volume changes. In sucrose animals, a large current sink was found covering nearly the entire outer 2/3-3/4 of stratum moleculare. This region corresponds
PAGE 124
114 precisely with the known terminations of afferents arriving from the lateral and medial entorhinal cortices which topographically project to the outer and middle 1/3 of the molecular layer, respectively (Hjorth Simonsen, 1973; McNaughton and Barnes, 1977). Furthermore, these afferents are known to be excitatory and thus would be expected to produce an active sink (inward current) in the synaptic zone (Andersen et~-, 1966b; Lomo, 1971a; McNaughton and Barnes, 1977). Current will flow passively (capacitatively) primarily out of nonactivated regions of the cells back into the extracellular space. This explains the large current source that appears over the granule cell bodies and proximal dendritic trunks. The existence of a small source in the distal molecular layer was unexpected. One explanation for this source is that synaptic density may be less at the distal edge of the molecular layer. Thus, the dendrites there could serve as a source for the more densely clustered and more proximally located excitatory synapses. Alternatively, the pia at the molecular layer border may insert a significant change in conductivity along the axis of electrode penetration, thus altering the current density calculations for the neighboring neuropil (cf. Chapter III). Finally, it is conceivable that the stimulating electrode was not fully activating the lateral entorhinal fibers. If true, this would suggest a consistently deviated stimulation site since nearly every animal showed evidence of this distal source. It is impossible at this point to choose among the alternatives; indeed, more than one may be valid. Chronic ethanol treatment altered the spatial distribution of the molecular layer sink currents. This effect was suggested by the normal ized EP profile, but was evident only in the ventral blade. The negative field potentials of Group E animals were decreased in amplitude and
PAGE 125
115 covered less distance across the molecular layer. The enhanced resolution provided by CSD analysis uncovered the fact that both blades, in fact, had an extremely similar reduction in sink width over stratum moleculare. Not only was the sink shrunken spatially, but the peak inward current was closer to the cell layer in ethanol-treated animals. The shrinkage was prominent only at the distal edge of the sink. These data suggest a preferential loss of synapses in the outer molecular layer. It cannot be determined from these data whether the primary change was a loss of afferent terminals (derived from lateral entorhinal cortex) or an attenuation of dendrites and spines. However, we do know from other studies that similar ethanol regimens produce granule cell loss (Walker et~-, 1980) and a reduction in spine density, but not dendritic arborization, in the surviving cells (Riley and Walker, 1978). These findings make it unlikely that the sink shrinkage was due to dendritic attenuation, and is best considered a preferential loss of synapses on the distal dendrites. It is interesting that synaptic loss is also thought to occur most prominently in distal dendritic regions in aged animals (Geinisman et~-, 1978). On the other hand, it is possible that the stimulating electrode sites differed slightly within the AB in Group E animals We consider this possibility unlikely. First, the electrodes were placed to generate maximum evoked responses in the dentate hilus. Maximum responses should be produced only by electrodes stimulating as many medial and lateral entorhinal fibers as possible. In addition, neither qualitative nor quantitative evaluations of the stimulation electrode placement uncovered any differences between the groups. For these reasons, we consider the synaptic loss hypothesis the most suitable explanation for the sink shrinkage.
PAGE 126
116 It is interesting to note that the peak in w a r d currents were unchanged in magnitude by the ethanol treatment. This would not be surprising if the loss of synapses were restricted to the distal dendritic tree. Hmvever, spines are also lost more proxi mally (Riley and Walker, 1978) and so the possibility remains that fewer contacts are made through out the dendritic arbor. Overall synaptic drive could be maintained, however, by a compensatory enhancement of synaptic efficacy. Such compensation is apparently present in aged animals where synaptic loss is also evident (Barnes, 1979; Barnes and McNaughton, 1980). Similar reactive changes in synaptic efficacy may have occurred in the alcoholic rats since the I/0 curves for EPSP amplitude did not differ between groups (Fig. 5-4). If compensatory changes v1ere present, one would have to conclude either that distal synapses have less capacity for compensation, or the distal synapses were more heavily damaged and thus could not show adequate com pensation. The regional loss of synaptic drive in CAl of Group E animals (Chapter III) gives plausibility to the second alternative. Synaptic function. The ethanol group exhibited normal I/0 functions of the basic EPSP and PS measures. There were seemingly opposing trends of a decreased PS latency (quicker time to peak from EPSP onset) and a decreased PS amplitude. The somewhat smaller PS in Group E sug gested a possible reduced granule cell excitability since the EPSP I/0 functions were quite similar between groups. This suggestion was con firmed statistically when the PS was expressed in proportion to the EPSP slope. The decreased PS/EPSP was especially apparent at higher stimulus intensities which produced asymptotic responses (and thus less variance).
PAGE 127
117 The change in "excitability" exhibited in the extracellular field potential records raises the question of what cellular mechanisms underlie the phenomenon. As mentioned above, the unaltered EPSP I/0 function, in spite of spine loss and cell loss, may reflect compensatory facilitation of synaptic efficacy, e.g., through enhanced transmitter release or receptor supersensitivity. The decrease in PS/EPSP clearly means that fewer granule cells were activated by a given stimulus (assuming no changes in action potential amplitudes). One explanation for this de crease may be that the granule cells were in fact less excitable and that a lower percentage of cells reached threshold and fired. However, the neurons actually appeared to be more excitable because the latencies to PS peak were somewhat shorter. The other possibility is that there simply may have been fewer granule cells. This possibility is supported by a recent quantitative histological study of granule cell numbers in dorsal hippocampus (Walker et~., 1980). There is a good correlation between the ethanol-induced 15% cell loss observed in that study and the present 20% decrease in PS/EPSP. The trend toward a decreased PS latency can then be viewed as another cellular mechanism of compensation for cell and synapse loss, i.e., enhanced excitability. It is interesting to note in this regard that currents near PS threshold tended to produce larger population spikes in the ethanol group (Fig. 5-4). The parallels between the present I/0 data and corresponding findings in aged rats are striking. Aging results in a significant decrease in synaptic density in the middle molecular layer (Geinisman et~., 1977) although, unlike alcoholic rats, there is apparently only minor, if any, loss of granule cells (Bondareff, 1979; Brizzee and Ordy, 1979; Cotman and Scheff, 1979; Scheibel, 1979). Physiological studies have confirmed
PAGE 128
118 the 1oss of entorhina1-DG synaptic contacts (Barnes, 1979; Barnes and McNaughton, 1980). Both extracellular field potential and unit intracellular recordings have revealed that aged granule cells also exhibit compensatory changes in synaptic efficacy (greater EPSP per fiber volley) as well as action potential latency and threshold (as intracellularly recorded). No changes were observed in action potential amplitudes (Barnes and McNaughton, 1980). The major departure from the ethanol findings is that aged animals actually exhibit larger PS/EPSP ratios than middle-aged controls. However, this finding correlates well with the de creased PS and action potential thresholds plus the maintenance of a normal population of granule cells (Barnes and McNaughton, 1980; Bondareff, l 979). The short-term potentiation functions (PPP and FP) were remarkably unaffected by chronic ethanol treatment. There were no group differences on any of the PS or EPSP measures. The PS PPP in both groups quite closely resembled previously published results in urethane-anesthetized rats (Assaf and Miller, 1978). Urethane-anesthetized rabbits show a more prolonged period of potentiation (Lomo, 1971b). Potentiation of the EPSP has also been reported (Loma, 1971b; Steward et EJ_., 1977) although the phenomenon is apparently difficult to produce in rat hippocampal slices or in freely moving, unanesthetized rats (Barnes, 1979; White et Qj_., 1979). Clearly, these conflicting findings are due to the diverse experimental preparations employed. The FP depression and potentiation were also similar to previously published reports in intact animals (Andersen et EJ_., 1966b; Harris et~., 1979). One novel aspect of FP we have observed is that 5 and 10 Hz tetani produce an unusual sequence of events: increased PS latency, increased PS amplitude but decreased
PAGE 129
119 EPSP slop e when only short bursts of 25 pulses are applied. Obviously, the result is a dramatic potentiation of PS/EPSP. The cellular mechanisms mediating the dynamic processes occurring during FP remain to be determined. Considerable evidence now exists that LTP involves two differentiable processes: an enhancement of synaptic efficacy (increased field EPSP for a given stimulus current or presynaptic fiber volley) and an enhanced neural excitability (enhanced PS/EPSP) (Bliss and Gardner-Medwin, 1973; Douglas and Goddard, 1975; Goddard et _tl., 1980). The t\
PAGE 130
120 group showed changes in short-term potentiation, PPP and FP, although FP has not been so carefully investigated in the aged rats (Barnes, 1979). Both alcoholic and aged rats show impaired LTP processes, but the precise impairment differs between groups. These group differences can be explained tentatively by variations in experimental protocol, but a final judgment awaits direct comparison of the data collection procedures in one group or the other. Some researchers have proposed that chronic alcoholism speeds up the aging process (Beck et~-, 1979; Blusewicz et~-, 1977; Ryan and Butters, 1980). Although the present electrophysiological study compares favorably with similar work in the aged DG (Barnes, 1979; Barnes and McNaughton, 1980), a direct examination of alcoholic rats at different ages would greatly strengthen the argument.
PAGE 131
CHAPTER VI GENERAL DISCUSSION Chronic ethanol treatment is known to exert toxic effects in rodent hippocampus that are observable using quantitative light microscopic techniques. The toxicity is expressed as progressive neuronal deterioration, including dendritic spine loss, ultimately ending in cell death (McMullen et il_.", 1980; Riley and i~alker, 1978; Walker et il_., 1980). Since the hippocampus has a tremendous repertoire of anatomical and physiological compensatory responses to injury (Goldowitz et al., 1979; -Nadler et tl-, 1980a; Steward et il_., 1976), it is important to complement the histological studies with studies of hippocampal physiology in order to understand the functional consequences of chronic ethanol ex posure. The present experiments examined with extracellular recording techniques two monosynaptic hippocampal connections: 1) entorhinal cortex to DG, providing the major afferent input to the hippocampal formation and 2) CA3 to CAl, a hippocampal intrinsic pathway containing fibers originating from both ipsilateral and contralateral CA3. Study of both of these pathways allows comparisons not only of the effects of ethanol on these projections but also of their normal physiology. While the experimental protocols \!Jere not identical between systems, they were similar enough to allow meaningful comparisons. ("DG" will now refer specifically to the entorhinal projection system and "CAl II will refer to the CA3 projection system.) 121
PAGE 132
122 Physiological comparisons of DG and CAl monosynaptic connections. Only one previous study in hippocampal slices has directly compared the DG and CAl responses to single and repetitive shocks of the afferent fibers (Alger and Teyler, 1976). These two systems showed similar I/0 and LTP functions but differed during low frequency stimulation where CAl showed response potentiation but DG exhibited response depression. The present studies confirm many of the Alger and Teyler findings but contradict others; we extended the subfield comparison by examining other stimulation protocols as well. Analysis of the laminar profiles in CAl and DG revealed that EPs in the two regions had markedly different waveforms. Despite the much longer distance traveled by the AB fibers, dentate evoked responses had shorter EPSP onsets, shorter latencies to peak, faster risetimes and shorter half-widths. These differences seem best attributed to the fact that the AB stimulating electrode was activating a narrow column of myelinated fibers, resulting in a more synchronous activation of fastconduction axons than could be achieved in the dispersed, unmyelinated SCH/COM pathway. The DG blade differences in response latency are per plexing since the axonal bifurcations which produce the dorsal and ventral fibers are found in the DG apex molecular layer. It is conceivable that the electrode penetration was not perfectly normal to the DG laminae and thus penetrated more laterally (farther from the bifurcation) in the ventral blade. Paired-pulse potentiation of the EPSP had similar characteristics in CAl and the DG. Maximal potentiation occurred at short !Pis (20-30 msec), but the potentiation had almost completely disappeared by 125-175 msec poststimulus. In contrast, PPP of the PS showed dramatic regional
PAGE 133
123 variation. The PS in CAl was inhibited up to 80 msec following the conditioning stimulus but showed increasing amounts of facilitation even at the longest interval tested (180 msec). The DG PS, on the other hand, recovered from inhibition within 30 msec, showed maximal potentiation at 60-100 msec and declined back to baseline at the 175 msec IPI. Since the EPSP functions were regionally quite alike, the spike differences might be related to variations in spike-generating properties or in recurrent inhibitory interneuron activity (Assaf and Miller, 1978; Dunwiddie et~-, 1980). No studies have yet examined regional differences in such neuronal functions. On the other hand, EPSP potentiation functions are profoundly influenced by the stimulation intensity employed (White et~-, 1979). Thus, the EPSP, which was not measured at the stimulation intensities used to produce PS PPP, may be more related to the PS potentiation than is apparent from the current data. Thus, the cellular basis of the differences in PS potentiation remains unclear. Low frequency stimulation (l-10 Hz) produced robust potentiation of the PS in CAl. The maximal potentiation observed at 5 Hz compares favor ably with the 8 Hz reported by Alger and Teyler (1976). In the DG, very low frequency tetani (l Hz) produced consistent depression of PS and EPSP responses by about 50% While PS and EPSP depress ion was initially observed also with 5 and 10 Hz stimuli, progressively increasing potentiation of the PS only became apparent after the first 4-10 stimuli of the train. These data replicate the early work of Andersen et~ (1966a) in urethane-anesthetized rabbits, but do not agree with the findings of Alger and Teyler in rat hippocampal slices. As in CAl, trains of stimuli at 5 Hz produced somewhat better potentiation. Since it is known that PS FP in CAl (Creager et~ 1980) and DG (present
PAGE 134
124 study) is associated with a declining EPSP, the general potentiation functions at 5 and 10 Hz are reasonably similar in the two subfields. The spike facilitation during a train may be due to a nonspecific mechanism, e.g., extracellular potassium accumulation, since other excitatory afferent inputs to the postsynaptic cells are also potentiated during the train (Creager et~-, 1980). One of the most striking differences between CAl and DG is the minimal stimulation required to produce LTP (cf. Figs. 4-5 and 5-6). Low frequency, short duration tetani (5 Hz/5 sec, 10 Hz/2.5 sec and 10 Hz/5 sec) were able to generate robust LTP in CAl. These frequencies and durations are less than the lowest ones previously reported to produce LTP (Dunwiddie and Lynch, 1978). Interestingly, the same parameters of stimulation produced absolutely no LTP in the DG. At much higher frequencies of stimulation, LTP was easily obtained in the DG, but no comparisons can be made with CAl because of the disparate high frequency stimulation procedures utilized for each structure. Regional differences in LTP production have not been previously observed (Alger and Teyler, 1976). No anatomical or physiological parameters are presently known that can explain the regional dichotomy in synaptic function. The present experiments have identified certain regional differences in synaptic plasticity within the hippocampal formation. These differences are not due to different anesthetic conditions and are still present after chronic ethanol treatment. It is interesting that the input stage to the hippocampus (DG) is somewhat more restricted in synaptic plasticity than regions further downstream within the hippocampus (CAl). These data have an intriguing parallel to the findings from a study employing single unit recording during differential classical conditioning
PAGE 135
125 (Segal and Olds, 1973). Cells in CA3, CAl and DG all showed increased firing in response to the positive conditioned stimulus as learning pro gressed. However, the CA3 and CAl cells exhibited the increase earlier during conditioning and at a shorter latency following the conditioned stimulus than did the DG cells. Interestingly, only the dentate neurons showed differential responding to positive and negative stimuli (cf. Deadwyler et~-, 1979). One might speculate that the differential unit behavior observed during behavioral conditioning is mediated in part by the differential capacities for synaptic plasticity of the hippocampal subfields. Correlative studies such as these may help shore up the hypothesis that synaptic plasticity in the hippocampus is a normally oc curring process in awake behaving animals. Comparisons of ethanol effects in CAl and DG. On the surface, it appears that chronic ethanol treatment resulted in very different alterations in synaptic function in DG and CAl. We observed I/0 changes in dentate but not in CAl, short-term plasticity changes in CAl but not dentate and a clear trend toward reduced LTP in dentate but apparently not in CAl. The chief similarity was a reduction in each of the respective afferent terminal zones, as defined by CSD analysis. Since such disparate findings would imply very different modes of cellular damage, the following section is devoted to integrating the findings into a more cohesive pattern. The I/0 curve was significantly altered by ethanol only in the DG, but a comparison with the data from CAl indicates trends in a similar direction. In the DG, we found trends toward decreased PS latency and decreased PS amplitude at asymptote. It was not until the PS/EPSP ratio was calculated that a significant treatment effect was uncovered. In
PAGE 136
126 CAl, we observed weaker but directionally similar trends of decreased PS latency and decreased PS amplitude at asymptote in Group E animals. Unfortunately, we were unable to obtain a consistent EPSP measure and thus could not calculate the PS/EPSP ratios. These correlations do not necessarily mean that CAl in fact has reduced I/0 function, but the issue should be considered unresolved at this time. The possible I/0 correlation is consistent with the finding that both the DG and CAl exhibited a shrinkage in their respective synaptic fieldsw presumably as a result of synaptic loss. Although chronic ethanol apparently exerted differential regional effects on LTP, we need to be cautious in interpretating this effect. First, the ethanol effect in dentate did not quite reach statistical significance (p < .08). Nonetheless, the effect was most prominent in the PS/EPSP measure, a measure that could not be calculated for CAl responses. An additional caveat is that since such disparate stimulation parameters were used to produce LTP in each structure, meaningful comparisons are severely compromised. Finally, there was in fact a weak trend for PS LTP to be somewhat attenuated after 100 Hz stimulation in CAl. Perhaps the PS/EPSP measure would have revealed a significant effect here. Chronic ethanol has been shown to have deleterious consequences for LTP in CAl studied in hippocampal slices, but so far these studies have been performed only immediately after ethanol withdrawal (Durand et~-, 1980). Because of this procedural flaw, it is not clear whether this was an acute ethanol effect, a withdrawal effect, an adaptation to chronic ethanol exposure or a persistent impairment produced by the ethanol regimen. Thus, while the current stimulation procedures did not uncover an ethanol effect on LTP in CAl, it remains possible that other
PAGE 137
127 stimulus parameters or other response measures are required to observe the phenomenon. Nevertheless, until there is evidence otherwise, we must conclude that there is a regional differentiation of chronic ethanol's influence on LTP. The most interesting puzzle in these data is the dichotomy of chronic ethanol's effects on short-term plasticity. No treatment effects were observed in the DG on either PPP or FP. In CAl, Group E demonstrated an augmentation of PS PPP and, consistently, an augmentation of PS FP at 5 and 10 Hz. We suggested that the CAl effects were best ex plained by a reduction in recurrent inhibition, a hypothesis supported by the reduced inhibition following l Hz stimulation. The fact that PPP and FP behave somewhat differently in normal CAl and DG (see above discussion, this chapter) implies that somewhat different mechanisms are operating in their genesis. Perhaps ethanol has differential effects on these hypothesized different mechanisms of short-term potentiation. Alternatively, other workers in our laboratory have collected preliminary morphological evidence that CAl is somewhat more sensitive to ethanol neurotoxicity than CA3 or DG. This finding allows a more parsimonious explanation; namely, that ethanol has similar effects in CAl and the DG, e.g., reduced recurrent inhibition, but that CAl is in a more advanced state of impairment. Conducting physiological studies after varying durations of ethanol exposure would directly test this hypothesis. The means by which chronic ethanol produces functional impairments remain unclear. It is apparent, however, that chronic ethanol-related impairments in synaptic function are not ubiquitous throughout the hippo campus, but show some subfield differentiation. The changes in I/0 curves and afferent lamination patterns are clearly related to the cell loss and
PAGE 138
128 reduction in spine density that have been observed in similarly treated animals. But until we know more about the cellular mechanisms underly ing the various forms of hippocampal synaptic plasticity, it will remain difficult to determine the causal events by which ethanol exerts its effects on hippocampal potentiation. Comparisons of alcoholism and aging effects on the hippocampus. There exist numerous neuropsychological and neurophysiological similarities in the CNS effects of alcoholism and aging in humans. This fact has led some researchers to hypothesize that chronic alcoholism in fact speeds up the aging process (Beck et~-, 1979; Blusewicz et~-, 1977; Ryan and Butters, 1980). Because of the growing interest in this hypothesis, the effects of chronic ethanol consumption observed in the present studies were compared in many of the previous sections to related findings in aged animals. Without reviewing the specific comparisons, suffice it to say that chronic ethanol and aging largely influence hippocampal synaptic function in a congruent fashion. The major deviation from this correspondence was the enhanced short-term plasticity found in CAl of alcoholic rats. Both the positive and negative correlations should be kept in perspective, however. Researchers have only begun to explore hippocampal physiology in either the alcohol field (Durand et ~-, 1980; present studies) or in the aging field (Barnes, 1979; Barnes and McNaughton, 1980; Landfield and Lynch, 1977; Landfield et~-, 1978). Nevertheless, the reasonable correspondence of the few studies that have been done thus far gives support for a unification of the two syndromes by implicating a common mechanism of action on the CNS.
PAGE 139
APPENDIX CURRENT-SOURCE DENSITY ANALYSIS The major generators of field potentials in the CNS are in most cases the dendrites. Action potentials in somata and axons do not contribute nearly as much, requiring considerable synchronization to appear in the extracellular records. The interpretation of field potentials is a problem, however, unless additional information is provided. For instance, a negative potential is generally interpreted as a local current "sink," i.e., current flowing from the extracellular space into cells. Likewise, a positive potential is considered to represent a local current "source," i.e., current entering the extracellular space. Two questions immediately arise. First, is a particular sink produced by active current flow (as produced by excitatory synapses) or is it merely passive current from a neighboring active source? (The reverse question can be asked of an observed source.) The answer to this question relies on a knowledge of the anatomy of the system under study and an understanding of the postsynaptic effects of the afferent inputs. The second question is one of spatial resolution. Is a given potential locally produced or volume conducted from more distant sites? This is not a trivial point and is crucial to the correct interpretation of a field potential laminar analysis (as used in the present studies). We can directly address this question, however, through the use of current-source density (CSD) analysis. The chief benefit of CSD analysis is the spatial resolution of the local membrane currents through numerical analysis of the field potential 129
PAGE 140
130 data (see below). Because we were interested in the effect of chronic ethanol on the distribution of synaptic currents within relatively small regions, we employed CSD analysis to give us the necessary resolution. The membrane current (Im) is a function of the extracellularly recorded potential ( ) and the tissue conductivity (a). As long as a is constant and we are not interested in the absolute values of Im' then we only need to calculate Im. Since is a weighted integral of Im' we need to differentiate to obtain I (Freeman and Nicholson, 1975; Nicholson and Freeman, 1975). Let us be more specific. Current flows between points of unequal potential; thus, current (flow) density (J, a vector quantity) is proportional to the derivative of the potential field (i.e., the value of over space). Thus, with reference to extracellular space, the current flows between points on the boundary of the space and cell membranes; e.g., between the regions of sources and sinks. Sources and sinks are scalar quantities, at points in space. Thus, we have: (1) J = dV/dx (current flows between points of uneven potential) and (2) I = dJ/dx (at a point in space) m therefore (3) I = d/dx(dV/dx) = d 2V/dx2 (ignoring conductivity). m The field potential is a continuous function throughout the sampled region, of course, but can only be measured at discrete points by a microelectrode. Thus, the second derivative must be approximated by a finite difference formula (Nicholson, 1979). A full three-dimensional analysis of CSD requires measurement of the potential along all three axes (x, y, and z). If in the lattice spacing
PAGE 141
131 of microelectrode sampling points h = 6x = 6y = 62 and assuming isotropicity of conductivity, then: (4) Im(x,y,z) = ((x+h,y,z) + (x,y+h,z) + (x,y,z+h) + (x-h,y,z) + (x,y-h,z) + (x,y,z-h) 6(x,y,z))/h2 When recording in the center of a synchronously active open field with a width/length ratio at least as great as 2, e.g., when in the middle of an activated hippocampal lamella, one-dimensional analyses may be appropriately used (Nicholson and Freeman, 1975). Equation 4 then simplifies to (recording along the z axis): (5) I (z) = ((z+h) + (z-h) 2(z))/h2 m The differentiation procedure often produces noisy results and thus spatial smoothing operations can be applied to the field potential data to reduce unwanted high frequency spatial noise. Of the various formulae presented by Freeman and Nicholson (1975), we found the 04 smoothing to adequately reduce the noise but still leave intact the major sources and sinks. The 04 smoothing function is presented below: (6) 04 = (l/l00h2 ) (9(r+3h) + 6(r+2h) 5(r+h) 20(r) 5(r-h) + 6(r-2h) + 9(r-3h) where r is the sampling location and his again the sampling interval. Notice that the 04 smoothing function weights the influence of potentials recorded up to three sampling intervals away in either direction from the current recording site. The smoothing function is performed prior to the differentiation of equation 5. The high degree of smoothing introduced by 04 (smearing the spatial resolution of sources and sinks) is partially compensated by using a small sampling interval {h) of 25 m.
PAGE 142
REFERENCES Alger, B.E. and Nicoll, R.A. Spontaneous inhibitory post-synaptic potentials in hippocampus: Mechanism for tonic inhibition. Brain Res. 1980, 200: 195-200. Alger, B.E. and Teyler, T.J. Long-term and short-term plasticity in the CAl, CA3, and dentate regions of the rat hippocampal slice. Brain Res. 1976, 110: 463-480. Amaral, D.G. A Golgi study of cell types in the hilar region of the hippocampus in the rat. ~Comp. Neural. 1978, 182: 851-914. Andersen, P. Interhippocampal impulses II: Apical dendritic activation of CAl neurons. Acta Physiol. Scand. 1960, 48: 178-208. Andersen, P., Blackstad, T.W. and Loma, T. Location and identification of excitatory synapses on hippocampal pyramidal cells. ~Brain Res. 1966a, 1 : 236-248. Andersen, P., Bliss, T.V.P. and Skrede, K.K. Lamellar organization of hippocampal excitatory pathways. ~Brain Res. 1971a, 12: 222-238. Andersen, P., Bliss, T.V.P. and Skrede, K.K. pal population spikes. ~Brain Res. Unit analysis of hippocam197lb, 12: 208-221. Andersen, P., Eccles, J.C. and Loyning, Y. Location of postsynaptic inhibitory synapses on hippocampal pyramids. ~Neurophysiol. 1964a, 27: 592-607. Andersen, P., Eccles, J.C. and Loyning, Y. Pathway of postsynaptic inhibition in the hippocampus. ~Neurophysiol. 1964b, 27: 608-619. Andersen, P., Holmquist, B. and Voorhoeve, P.E. Entorhinal activation of dentate granule cells. Acta Physiol. Scand. 1966b, 66: 448-460. Andersen, P. and Loma, T. Mode of control of h ippocampal pyramidal cell discharges. In R.E. Whalen (Ed.) Neural Control of Behavior. New York: Academic Press, 1970. Andersen, P., Silfrenius, H., Sundberg, H Sveen, 0. and Wigstrom, H. Functional characteristics of unmyelinated fibres in the hippocampal cortex. Brain Res. 1978, 144: 11-18. 132
PAGE 143
133 Assaf, S.Y. and Miller, J.J. Neuronal transmission in the dentate gyrus: Role of inhibitory mechanisms. Brain Res. 1978, 151: 587-592. Barnes, C.A. Memory deficits associated with senescence: A neurophysio logical and behavioral study in the rat. ~Comp. Physiol. Psychol. 1979, 93: 74-104. Barnes, C. and McNaughton, B.L. Physiological compensation for loss of afferent synapses in rat hippocampal pyramidal cells during senescence. i Physiol. 1980, 309: 473-486. Beck, E.C., Dustman, R.E., Blusewicz, M.J. and Cannon, W.G. Cerebral evoked potentials and correlated neuropsychological changes in the human brain during aging: A comparison of alcoholism and aging. In J.M. Ordy and K. Brizzee (Eds.) Sensory Systems and Communication in the Elderly. (Aging, vol. 10) New York: Raven Press, 1979. Begleiter, H., DeNoble, V. and Porjesz, B. Protracted brain dysfunction after alcohol withdrawal in monkey. Adv. Exp. Med. Biol. 1980, 126: 237-250. Black, A.H., Nadel, L. and O'Keefe, J. Hippocampal function in avoidance learning and punishment. Psych. Bull. 1977, 84: 1107-1129. Blackstad, T.W. Commissural connections of the hippocampal region in the rat, with special reference to their mode of termination. i Comp. Neurol. 1956, 105: 417-537. Bliss, T.V.P. and Gardner-Medwin, A.R. Long-lasting potentiation of synaptic transmission in the dentate area of the unanaesthetized rabbit follmving stimulation of the perforant path. i Physiol. 1973, 232: 357-374. Bliss, T.V.P. and Lomo, T. Long-lasting potentiation of synaptic transmission in the dentate area of the anesthetized rabbit following stimulation of the perforant path. iPhysiol. 1973, 232: 331-356. Bondareff, W. Synaptic atrophy in the senescent hippocampus. Mech. Aging Devel. 1979, 9: 163-172. -Bondareff, W. and Geinisman, Y. Loss of synapses in the dentate gyrus of the senescent rat. Am. J. Anat. 1976, 145: 129-136. Brion, S. Korsakoff's syndrome: Clinico-anatomical and physiopathologi cal considerations. In G.A. Talland and N.C. Waugh (Eds.) The Pathology of Memory. New York: Academic Press, 1969. Brizzee, K.R. and Ordy, J.M. Age pigments, cell loss and hippocampal function. Mech. Aging Devel. 1979, 9: 143-162. Brooks, D.N. and Baddeley, A.O. What can amnesic patients learn? Neuropsychol. 1976, 14: 111-122.
PAGE 144
134 Buzsaki, G. Long-term potentiation of the commissural-CAl pyramidal cell response in the hippocampus of the freely moving rat. Neurosci. Lett. 1980, 19: 293-296. Carlen, P.L. and Corrigall, W.A. Ethanol tolerance measured electrophysiologically in hippocampal slices and not in neuromuscular junctions from chronically ethanol-fed rats. Neurosci. Lett. 1980, 17: 95-100. Cermak, L.S., Butters, N. and Gerrein, J. coding ability of Korsakoff patients. 85-94. Cermak, L.S., Butters, N. and Goodglass, H. in Korsakoff patients. Neuropsychol. The extent of the verbal en Neuropsychol. 1973, 11: The extent of memory loss 1971 9: 307-315. Cermak, L.S. and Moreines, I. Verbal retention deficits in aphasic and amnesic patients. Brain Lang. 1976, 3: 16-27. Cermak, L.S., Naus, M.J. and Reale, L. Rehearsal and organizational strategies of alcoholic Korsakoff patients. Brain Lang. 1976, 3: 375-385. Cermak, L.S. and Reale, L. Depth of processing and retention of words by alcoholic Korsakoff patients. ~. ~. Psychol. 1978, 4: 165-174. Cermak, L.S., Uhly, B. and Reale, L. Encoding specificity in the alcoholic Korsakoff patient. Brain Lang. 1980, 11: 119-127. Corkin, S. Acquisition of motor skill after bilateral medial temporal lobe excision. Neuropsychol. 1968, 6: 255-265. Cotman, C.W. and Scheff, S.W. Compensatory synapse growth in aged animals after neuronal death. Mech. Aging Devel. 1979, 9: l 03-116. Courville, C.B. Effects of Alcohol on the Nervous System of Man. Los Angeles: San Lucas Press, 196~ Coyle, P. Spatial features of the rat hippocampal vascular system. ~. Neuro 1 1978, 58: 549-561 Creager, R., Dunwiddie, T. and Lynch, G. Paired-pulse and freuquency facilitation in the CAl region of the in vitro rat hippocampus. ~. Physiol. 1980, 299: 409-424. Deadwyler, S.A., Dunwiddie, T. and Lynch, G. Short-lasting changes in hippocampal neuronal excitability following repetitive synaptic activation. Brain Res. 1978, 147: 384-389. Deadwyler, S.A., \~est, M. and Lynch, G. Activity of dentate granule cells during learning: Differentiation of perforant path input. Brain Res. 1979, 169: 29-41.
PAGE 145
135 DeFrance, J.F., Kitai, S.T. and Shimano, T. Electrophysiological analysis of the hippocampal-septal projections: II. Functional characteristics. ~Brain Res. 1973, 17: 462-470. Deluca, D., Cermak, L.S. and Butters, N. An analysis of Korsakoff patient's recall following varying types of distractor activity. Neuropsychol. 1975, 13: 271-279. Deluca, D., Cermak, L.S. and Butters, N. The differential effects of semantic, acoustic and nonverbal distraction on Korsakoff patient's verbal retention performance. Internat. J. Neurosci. 1976, 6: 279-284. Denoble, V.J. and Begleiter, H. Impairment of acquisition of a DRL schedule following prolonged ethanol consumption. Pharm. Biochem. Behav. 1979, 10: 393-396. Douglas, R.J. The hippocampus and behavior. Psychol. Bull. 1967, 67: 41 6-442. Douglas, R.M. Long-lasting synaptic potentiation in the rate dentate gyrus following brief high frequency stimulation. Brain Res. 1977, 126: 361-365. Douglas, R.M. and Goddard, G. Long-term potentiation of the perforant path-granule cell synapse in the rat hippocampus. Brain Res. 1975, 86: 205-215. Dreyfus, P.M. Diseases of the nervous system in chronic alcoholics. In B. Kissin and H. Begleiter (Eds.) The Biology of Alcoholism, Vol. _1: Clinical Pathology. New York: Plenum Press, 1974. Dunwiddie, T. and Lynch, G. Long-term potentiation and depression of synaptic responses in the rat hippocampus: Localization and frequency dependency. ~Physiol. 1978, 276: 353-368. Dunwiddie, T., Mueller, A., Palmer, M., Stewart, J. and Hoffer, B. Electrophysiological interactions of enkephalins with neuronal circuitry in the rat hippocampus. I. Effects on pyramidal cell activity. Brain Res. 1980, 184: 311-330. Durand, D., Cohen, P.L. and McMullen, P. Impairment of long-term potentiation following chronic ethanol consumption in rats. Soc. Neurosci. Abst. 1980, 89. Epstein, P.S., Psiani, V.D. and Fawcett, J.A. Alcoholism and cer ebral atrophy. Alcoholism: Clin. ~. Res. 1977, l: 61-65. Fehr, K.A., Kalant, H. and LeBlanc, A.E. Residual learning deficit after heavy exposure to cannabis or alcohol in rats. Science. 1976, 192: 1248-1251
PAGE 146
136 Finch, D.M. and Babb, T.L. Neurophysiology of the caudally directed hippocampal efferent system in the rat: Projections to the subicular complex. Brain Res. 1980, 197: 11-26. Freeman, J.A. and Nicholson, C. Experimental optimization of current source-density technique for anuran cerebellum. ~Neurophysiol. 1975, 38: 369-382. Freeman, J.A. and Stone, J. A technique for current density analysis of field potentials and its application to the frog cerebellum. In R. Llinas (Ed.) Neurobiology of Cerebellar Evolution and Devel opment. Chicago: American Medical Association, 1969. Freund, G. Impairment of shock avoidance learning after long-term alcohol ingestion in mice. Science. 1970, 168: 1599. Freund, G. Chronic central nervous system toxicity of alcohol. Ann. Rev. Pharmacol. 1973, 13: 217. Freund, G. Normal shuttle box avoidance learning after chronic phenobarbital intoxication in mice. Psychopharmacol. 1974, 40: 199-203. Freund, G. Effects of chronic alcohol and vitamin-E consumption on aging pigments and learning performance in mice. Life Sci. 1979, 24: 145-152. Freund, G. Benzodiazepine receptor loss in brains of mice after chronic alcohol consumption. Life Sci. 1980, 27: 987-992. Freund, G. and Walker, D.W. The effect of aging on acquisition and retention of shuttle box avoidance in mice. Life Sci. 1971a, 10: 1343-1349. Freund, G. and lrJalker, D.~J. Impairment of avoidance learning by pro longed ethanol consumption in mice. ~Pharacol. ~Ther. 1971b, 179: 284-292. Fricke, R. and Cowan, W.M. An autoradiographic study of the commissural and ipsilateral hippocampo-dentate projections in the adult rat. ~Comp. Neurol. 1978, 181: 253-270. Geinisman, Y., Bondareff, W. and Dodge, J.T. Partial deafferentation of neurons in the dentate gyrus of the senescent rat. Brain Res. 1977, 134: 541-545. Geinisman, Y., Bondareff, W. and Dodge, J.T. Dendritic atrophy in the dentate gyrus of the senescent rat. Amer.~Anat. 1978, 153: 537-544. Goddard, G.V., Bliss, T.V.P., Robertson, A. and Sutherland, R.S. Noradrenaline levels affect long-term potentiation in the hippocampus. Soc. Neurosci. Abst. 1980.
PAGE 147
137 Goldowitz, A., Scheff, S.W. and Cotman, C.H. The specificity of reactive synaptogenesis: A comparative study in the adult rat hippocampal formation. Brain Res. 1979, 170: 427-442. Gottlieb, D.I. and Cowan, W.M. Autoradiographic studies of the commissural and ipsilateral assocational connections of the hippocampus and dentate gyrus of the rat. I. The commissural connections. ~Comp. Neurol. 1973, 149: 393-422. Haberly, L.B. and Shepard, G. Current density analysis of summed evoked potentials in the oppossum propyriform cortex. J. Neurophysiol. 1973, 36: 789-802. Harris, E.W., Lasher, S.S. and Steward, 0. Analysis of the habituationlike changes in transmission in the temporo-dentate pathway of the rat. Brain Res. 1979a, 162: 21-32. Harris, R.A., Krause, W., Goh, E. and Case, J. Behavioral and biochemi cal effects of chronic consumption of ethanol by hamsters. Pharm. Biochem. Behav. 1979b, 10: 343-347. Haug, J.O. Pneumoencephalographic evidence of brain damage in chronic alcoholics. Acta Psychiat. Scand. (Suppl.) 1968, 203: 135-143. Hirsch, R., Leber, B. and Gillman, K. Fornix fibers and motivational states as controllers of behavior: A study stimulated by the con textual retrieval theory. Behav. Biol. 1978, 22: 463-478. Hjorth-Simonsen, A. Some intrinsic connections of the hippocampus in the rat: An experimental analysis. ~. Comp. Neural. 1973, 147: 145-162. Hjorth-Simonsen, A. and Jeune, B. Origin and termination of the hippocampal perforant path in the rat studied by silver impregnation. ~Comp. Neuro l. 1972, 144: 215-232. Horel, J.A. The neuroanatomy of amnesia: A critique of the hippocampal memory hypothesis. Brain. 1978, 101: 403-447. Isseroff, A. Limited recovery of spontaneous alternation after extensive hippocampal damage: Evidence for a memory impairment. ~ Neural. 1979, 64: 284-294. Iversen, S.D. Do hippocampal lesions produce amnesia in animals? Internat. Rev. Neurobiol. 1976, 19: 1-49. Jarrard, L.E. The hippocampus and motivation. Psychol. Bull. 1973, 79: 1-12. Kalant, H., Leblanc, E. and Gibbins, R. Tolerance to, and dependence on, ethanol. In Y. Israel and J. Mardones (Eds.) Biological Basis of Alcoholism. New York: Wiley Interscience, 1971.
PAGE 148
138 Karobath, M., Rogers, J. and Bloom, F.E. Benzodiazepine receptors remain unchanged after chronic ethanol administration. Neuropharmacol. 1980, 19: 125-128 Kimble, O.P. Hippocampus and internal inhibition. Psychol. Bull. 1968, 70: 285-295. Kimble, O.P. Choice behavior in rats with hippocampal lesions. In R.L. Isaacson and K.H. Pribram (Eds.) The Hippocampus. Vol. 2, New York: Plenum Press, 1975. Kinsbourne, M. and Wood, F. Short-term memory processes and the amnesic syndrome. In D. Deutsch and J.A. Deutch (Eds.) Short-term Memory. New York: Academic Press, 1975. Kunz, G., Englisch, H.J. and Wenzel, J. Untersuchungen zur spinesverteilung an pyramidenneuronen der CAl-region des hippocampus der ratte nach langzeitiger oraler alkoholapplikation. J. Hirnforsch. 1976, 17: 351-363. Landfield, P.W. and Lynch, G. Impaired monosynaptic potentiation in in vitro hippocampal slices from aged, memory-deficient rats. J. Geronto 1 1977, 32: 523-533. Landfield, P.W., McGaugh, J.L. and Lynch, G. Impaired synaptic potentiation processes in the hippocampus of aged, memory-deficient rats. Brain Res. 1978, 150: 85102. Laurberg, S. Commissural and intrinsic connections of the rat hippocampus. ~. Comp. Neural. 1979, 184: 685-708. Lee, H.K., Dunwiddie, T. and Hoffer, B. Electrophysiological interactions of enkephalins with neuronal circuitry in the rat hippocampus. II. Effect on interneuron excitability. Brain Res. 1980, 184: 331-342. Leung, L.W.S. Orthodromic activation of hippocampal CAl region of the rat. Brain Res. 1979, 196: 49-64. Liljequist, S. and Engel, J. Attenuation of the y-butyrolactone induced increase in DOPA accumulation by chronic ethanol administration. J. Neural Transm. 1979, 46: 195-204. Lomo, T. Patterns of activation in a monosynaptic cortical pathway: The perforant path input to the dentate area of the hippocampal formation. ~. Brain Res. 1971a, 12: 17-45. Lomo, T. Potentiation of monosynaptic EPSPs in the perforant path dentate granule cell synapse. ~. Brain Res. 1971b, 12: 46-63. Lorente de No, R. Studies on the structure of the cerebral cortex. II. Continuation of the study of the Ammonic system. J. Psychol. Neuro 1 1934, 46: 113-177.
PAGE 149
139 Lorente de No, R. fl Study of Nerve Physiology. New York: The Rockefellar Institute for Medical Research, Vol. 2, 1947. MacDonell, J.S. and Marcuella, H. Short-term memory deficits following prolonged ethanol consumption by rats. In F.A. Seixas (Ed.) Currents in Alcoholism. New York: Grune and Stratton, Vol. 3, 1978. Magleby, K.L. and Zengel, J.E. Long-term changes augmentation, potentiation, and depression of transmitter release as a function of repeated synaptic activity at the frog neuromuscular junction. J. Physiol. 1976, 275: 471-494. Mair, W.G.P., Warrington, E.K. and Weiskrantz, L. Memory disorder in Korsakoff1s psychosis: A neuropathological and neuropsychological investigation of two cases. Brain. 1979, 102: 749-784. McMullen, P.A., Saint-Cyr, J.A., Petit, T.L., Carlen, P.L. and Durand, D. Morphological changes in hippocampal CAl pyramidal cell dendrites after chronic ethanol consumption in the rat-a Golgi study. Soc. Neurosci. Abst. 1980, 736. McNaughton, B.L. Dissociation of short-and long-lasting modification of synaptic efficacy at the terminal of the perforant path. Soc. Neurosci. Abst. 1977. McNaughton, B.L. and Barnes, C.A. Physiological identification and analysis of dentate gyrus cell responses to stimulation of the medial and lateral perforant pathways in the rat. ~Comp. Neural. 1977, 175: 439-454. Meibach, R.C. and Seigel, A. Efferent connections of the hippocampal formation in the rat. Brain Res. 1977, 124: 197-224. Milner, B. Memory and the medial temporal regions of the brain. In K.H. Pribram and G. Broadbent (Eds.) Biology of Memory. New York: Academic Press, 1970. Milner, B., Corkin, S. and Teuber, H.-L. Further analysis of the hippocampal amnesic syndrome: 14-year follow-up study of H.M. Neuropsychol. 1968, 6: 215-234. Nadler, J.V., Perry, B.W. and Cotman, C.W. Selective reinnervation of hippocampal area CAl and the fascia dentata after destruction of CA3-CA4 afferents with kainic acid. Brain Res. 1980a, 182: 1-10. Nadler, J.V., Perry, B.W., Gentry, C. and Cotman, C.W. Loss and reacquisition of hippocampal synapses after selective destruction of CA3-CA4 afferents with kainic acid. Brain Res. 1980b, 191: 387404. Newlin, S.A., French, E.D., Berger, T. and Bloom, F.E. Ethanol prolongs the time course of recurrent inhibition in the hippocampus in response to corrmissural fiber activation. Soc. Neurosci. Abst. 1979.
PAGE 150
140 Nicholson, C. Generation and analysis of extracellular field potentials. In D.R. Humphrey (Ed.) Electrophysiology Techniques. Bethesda: Society for Neuroscience, 1979. Nicholson, C. and Freeman, J.A. Theory of current-source density analysis and determination of conductivity tensor for anuran cerebellum. ~Neurophysiol. 1975, 38: 356-368. O'Keefe, J. and Nadel, L. The Hippocampus as~ Cognitive Map. Oxford: Clarendon Press, 1978. Parkinson, S.R. The amnesic Korsakoff syndrome: A study of selective and divided attention. Neuropsychol. 1979, 17: 67-75. Patel, G.J. and Lal, H. Reduction in brain y-aminobutyric acid and in barbital narcosis during ethanol withdrawal. ~Pharmacol. ~Ther~ 1973, 186: 625-629. Penfield, W. and Milner, B. Memory deficit produced by bilateral lesions in the hippocampal zone. A.M.A. Arch. Neural. Psychol. 1958, 79: 475-497. Raisman, G., Cowan, W.M. and Powell, T.P.S. The extrinsic afferent, commissural and associational afferents to the hippocampal formation. Brain. 1965, 88: 963-996. Ramon y Cajal, S. Histologie de Systeme Nerveux de l 'monne et des Vertebres. Paris: Maloine, Vol. 2, 1911. Riley, J.N. and Walker, D.W. Morphological alterations in hippocampus after long-term alcohol consumption in mice. Science. 1978, 201: 646-648. Riley, J.N. and Walker, D.W. Morphological alterations in cerebellum following chronic alcohol consumption. ~Neurol. 1981, in press. Ryan, C. and Butters, M. Learning and memory impairments in young and old alcoholics. Evidence for the premature-aging hypothesis. Alcoholism: Clin. ~Res. 1980, 4: 288-293. Schaffer, K. Be it rag zur hi stol ogi e der ammonshornformation. Arch. Mikr. Anat. 1892, 39: 611-632. Scheibel, A.B. The hippocampus: Organizational patterns in health and senescence. Mech. Aging Devel. 1979, 9: 89-102. Scheibel, M.E., Lindsay, R.D., Tomiyasu, U. and Scheibel, A.B. Pro gressive dendrite changes in the aging human limbic system. ~Neural. 1976, 53: 420-430. Schubert, P. and Mitzdorf, V. Analysis and quantitative evaluation of the depressive effect of adenosine on evoked potentials in hippocampal slices. Brain Res. 1979, 172: 186-190.
PAGE 151
141 Schwartzkroin, P.A. and Wester, K. Long-lasting facilitation of a synaptic potential following tetanization in the in vitro hippocampal slice. Brain Res. 1975, 89: 107-119. Scoville, W.B. The limbic lobe in man. ~Neurosurg. 1954, 11: 64-66. Scoville, W.B. and Milner, B. Loss of recent memory after bilateral hippocampal lesions. J. Neural. Neurosurg. Psychiat. 1956, 20: 11-21 Segal, M. and Olds, J. Activity of units in the hippocampal circuit of the rat during differential classical conditioning. J. Comp. Physiol. Psychol. 1973, 82: 195-204. Smith, G.J., Myer, J.S., Hull, J.H. and Thomas, D.A. Behavioral effects of prolonged ethanol consumption temporal shock discrimination. Physiol. Behav. 1979, 22: 609-612. Smith, J.W., Burt, D.W. and Chapman, R.F. Intelligence and brain damage in alcoholics: A study of patients of middle and upper social class .Q_. ~Stud. Alcohol. 1973, 34: 414-422. Sotzing, J.H. and Brown, T.S. Chronic intermittent ethyl alcohol inhalation and avoidance learning. Pharmacol. Biochem. Behav. 1976, 5: 417-421. Stanley, J., Taber, K. and Defrance, J.F. A functional analysis of the ventral hippocampal commissural system of the rabbit. Brain Res. 1 979, 167: 168-171 Steward, 0., Cotman, C. and Lynch, G. A quantitative autoradiographic and electrophysiological study of the reinnervation of the dentate gyrus by the contralateral entorhinal cortex following ipsilateral entorhinal lesions. Brain Res. 1976, 114: 181-200. Steward, 0. and Scoville, S.A. Cells of origin of entorhinal cortical afferents to the hippocampus and fascia dentata of the rat. J. Comp. Neurol. 1976, 169: 347-370. Steward, 0., White, W.F. and Cotman, C.W. Potentiation of the excitatory synaptic action of commissural, associational and entorhinal afferents to dentate granule cells. Brain Res. 1977, 134: 551-560. Strauss, E.H. and Butler, R.B. The effect of varying types of interference on haptic memory in the Korsakoff patient. Neuropsychol. 1 978, 16: 81 -90. Sunahara, G.I., Kalant, H., Schofield, M. and Grupp, L. Regional distribution of ethanol in the rat brain. Can.~Physiol. Pharmacol. 1978, 56: 988-992.
PAGE 152
142 Sutton, I. and Simmonds, M.A. Effects of acute and chronic ethanol on the y-aminobutyric acid system in rat brain. Biochem. Pharmacol. 1973, 22: 1685-1692. Swanson, L.W. and Cowan, W.M. An autoradiographic study of the organi zation of the efferent connection of the hippocampal formation in the rat. ~Comp. Neurol. 1977, 172: 49-84. Swanson, L.W., Sawchenko, P.E. and Co1;1an, W.M. Evidence that the commissural, associational and septal projections of the regio inferior arise from the same neurons. Brain Res. 1980, 197: 207-212. Swanson, L.W., Wyss, J.M. and Cowan, W.M. An autoradiographic study of the organization of intrahippocampal association pathways in the rat. J. Comp. Neurol. 1978, 181: 681-716. Talland, G.A. Deranged Memory. New York: Academic Press, 1965. Teyler, T.J. and Alger, B.E. Monosynaptic habituation in the vertebrate forebrain: The dentate gyrus examined in vitro. Brain Res. 1976, 115: 413-426. Teyler, T.J., Alger, B.E., Bergman, T. and Livingston, K. A comparison of long-term potentiation in the in vitro and in vivo hippocampal preparations. Behav. Biol. 1977, 19: 24-34.~Thompson, R.F. and Spencer, W.A. Habituation: A model phenomenon for the study of neuronal substrates of behavior. Psychol. Rev. 1966, 73: 16-43. Ticku, M.J. The effects of acute and chronic ethanol administration and its withdrawal on y-aminobutyric acid receptor binding in rat brain. Brit. J. Pharmacol. 1980, 70: 403-410. Ticku, M.K. and Burch, T. Alterations in y-aminobutyric acid receptor sensitivity following acute and chronic ethanol treatments. J. Neurochem. 1980, 34: 417-423. Tsuchiya, T. and Fukushima, H. Effects of benzodiazepines and pento barbitone on the GABA-ergic recurrent inhibition of hippocampal neurons. Europ. ~Pharmacol. 1978, 48: 421-424. Tumarkin, B., Wilson, J.D. and Collins, G.H. Cerebral atrophy due to alcoholism in young adults. US Armed Forces Med.~1955, 6: 64-74. Vanegas, H., Williams, B. and Freeman, J.A. Responses to stimulation of marginal fibers in the teleostean optic tectum. Exp. Brain Res. 1979, 34: 335-349. Victor, M., Adams, R.D. and Collins, G.H. The Wernicke-Korsakoff Syndrome. Philadelphia: P.A. Davis CO:: 1971.
PAGE 153
143 Victor, M. and Banker, B.Q. Alcohol and dementia. In R. Katzman, R.D. Terry and K.L. Blak (Eds.) Aging. New York: Raven Press, Vol. 7, 1978. Volicer, L., Hurter, B.P., Williams, R., Puri, S.K. and Volicer, B.J. Relationship between brain levels on cyclic nucleotides and y aminobutyric acid during ethanol withdrawal in rats. Drug Ale. Depend. 1977, 2: 317-327. Walker, D.W., Barnes, D.A., Zornetzer, S.F., Hunter, B.E. and Kubanis, P. Neuronal loss in hippocampus produced by prolonged ethanol consumption in rats. Science. 1980, 109: 711-713. Walker, D.W. and Freund, G. Impairment of shuttle box avoidance learning following prolonged alcohol consumption of rats. Physiol. Behav. 1971, 7: 773-778. Walker, D.W. and Freund, G. Impairment of timing behavior after prolonged alcohol consumption in rats. Science. 1973, 182: 597-599. Walker, D.W. and Hunter, B.E. Short-term memory impairment following chronic alcohol consumption in rats. Neuropsychol. 1978, 16: 545-554. Walker, D.W., Hunter, B.E., Barnes, D.E. and Riley, J.N. An animal model of alcohol-induced brain damage: A behavioral and anatomical analysis. In A. Wilkinson (Ed.) Cerebral Deficits in Alcoholism. Addiction Research Foundation, in press, 1981. Walker, D.W. and Means, L.W. Single alternation performance in rats with hippocampal lesions. Disruption by an irrelevant task interposed during the intertrial interval. Behav. Biol. 1973, 9: 93-104. Walker, D.W., Messer, L.G., Freund, G. and Means, L.W. Effect of hippocampal lesions and intertrial interval on single alternation per formance in the rat. J. Comp. Physiol. Psychol. 1972, 80: 469-477. --Warrington, E.K. and Weiskrantz, L. Amnesic syndrome: Consolidation or retrieval. Nature. 1970, 228: 628-630. West, J.R., Nornes, H.0., Barnes, C.L. and Bronfenbenner, M. The cells of origin of the commissural afferents to the area dentata in the mouse. Brain Res. 1979, 160: 203-216. White, W.F., Nadler, J.V. and Cotman, C.W. Analysis of short-term plasticity at the perforant path-granule cell synapse. Brain Res. 1979, 178: 41-53. Wickelgren, W.A. Sparing of short-term memory in an amnesic patient: Implications for strength theory of memory. Neuropsychol. 1968, 6: 235-244.
PAGE 154
144 Winocur, G. Effects of interference on discrimination learning and recall by rats with hippocampal lesions. Physiol. Behav. 1979, 22: 339-345. Winocur, G. and Bindra, D. Effects of additional cues on passive avoidance learning and extinction in rats with hippocampal lesions. Physiol. Behav. 1976, 17: 913-920. Winocur, G. and Kinsbourne, M. amnesics. Neuropsychol. Contextual cuing as an aid to Korsakoff 1978, 16: 671-682. Winocur, G. and Olds, J. Effects of context manipulation on memory and reversal learning in rats with hippocampal lesions. ~. Comp. Physiol. Psychol. 1978, 92: 312-322. Winocur, G. and Weiskrantz, L. An investigation of paired-associate learning in amnesic patients. Neuropsychol. 1976, 14: 97-110. Wolf, P. and Haas, H.L. Effects of diazepines and barbiturates on hippocampal recurrent inhibition. Arch. Pharmacol. 1977, 299: 211-218. Yamamoto, C., Matsumoto, K. and Takagi, M. Potentiation of excitatory postsynaptic potentials during and after repetitive stimulation in thin hippocampal sections. ~Brain Res. 1980, 38: 469-477.
PAGE 155
BIOGRAPHICAL SKETCH Wickliffe C. Abraham was born on September 24, 1954, to Stuart and Ida Jeanne Abraham of Hagerstown, MD. He received his primary and sec ondary schooling in Hagerstown. From 1972-1976 he attended the University of Virginia, graduating with a B.A. in psychology. He received his graduate training (1976-1981) at the Department of Neuroscience, University of Florida. He will receive the Ph.D. degree in June, 1981. His major research interest is the neurobiological basis of learning and memory. Participation in Florida volleyball events kept him physically active and mentally sane. In the fall of 1981, he plans to take a post doctoral position with Dr. Graham Goddard at the University of Otago, New Zealand. 145
PAGE 156
I certify that I have read this study and that in my op1n1on it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. I certify that I have read this study and that in my op1n1on it confor m s to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Don W. Walker Associate Professor of Neuroscience I certify that I have read this study and that in my op1n1on it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. arles J. Vi~(>'.-k', Jr. / n Professor of Neuroscience / // v
PAGE 157
I certify that I have read this study and that in my op1n1on it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. lk/jl FloyJ. omdn ~ Assistant Professor of Neuroscience I certify that I have read this study and that in my op1n1on it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ------~ ~--/ / ~ft e V Keith Berg Associate Professor of Psychology This dissertation was submitted to the Graduate Faculty of the College of Medicine and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. June, 1981 Dean for Graduate Studies and Research
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EQ0NFT2Q5_PDLL2A INGEST_TIME 2017-05-22T18:20:04Z PACKAGE AA00054899_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
|