Citation
Chemical stability and in vitro cytotoxicity of melphalan.

Material Information

Title:
Chemical stability and in vitro cytotoxicity of melphalan.
Creator:
Stout, Susan Ann, 1959-
Publisher:
University of Florida
Publication Date:
Language:
English
Physical Description:
xiii, 107 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Chlorides ( jstor )
Dosage ( jstor )
Drug evaluation ( jstor )
Hydrolysis ( jstor )
In vitro fertilization ( jstor )
Ions ( jstor )
Kinetics ( jstor )
Neuroblastoma ( jstor )
pH ( jstor )
Pharmacokinetics ( jstor )
Antineoplastic Agents -- pharmacokinetics ( mesh )
Dissertations, Academic -- Pharmaceutics -- UF ( mesh )
Melphalan -- pharmacokinetics ( mesh )
Melphalan -- pharmacology ( mesh )
Neoplasms, Experimental ( mesh )
Pharmaceutics thesis, Ph.D. ( mesh )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1987.
Bibliography:
Bibliography: leaves 101-106.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Susan Ann Stout.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
030319365 ( ALEPH )
18128626 ( OCLC )

Aggregation Information

UFIR:
Institutional Repository at the University of Florida (IR@UF)
UFETD:
University of Florida Theses & Dissertations
IUF:
University of Florida

Downloads

This item has the following downloads:


Full Text












CHEMICAL STABILITY AND IN VITRO CYTOTOXICITY OF MELPHALAN







BY


SUSAN ANN STOUT



























A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY



UNIVERSITY OF FLORIDA 1987
















FOR MY PARENTS




John and Rosamond Stout




for supporting me when I needed it, encouraging me when I was down, and

believing in me all the while.
















ACKNOWLEDGEMENTS


I would like to extend my sincere appreciation to my major professors, Dr. Christopher Riley and Dr. Hartmut Derendorf, for their exceptional guidance, their constant encouragement, and most of all, their friendship. I would also like to thank the members of my supervisory committee, Dr. Adrian Gee, Dr. John Dorsey, Dr. Stephen Curry and Dr.

John Perrin, as well as Dr. John Graham-Pole, for their advice throughout my graduate career. And a special word of thanks goes to Ms. Diana Worthington-White for her patience and assistance with the neuroblastoma studies. And last, but not least, a sincere thank you is extended to Dr. Linda Adair for giving me confidence when I needed it the most.

























111
















TABLE OF CONTENTS



PAGE

ACKNOWLEDGEMENTS. . 1iii

LIST OF TABLES. . vi

LIST OF FIGURES. . vii

ABSTRACT. . xi

CHAPTER

I. INTRODUCTION. . 1

The Chemotherapy of Cancer . . . . . 1
The Treatment of Neuroblastoma with
Melphalan . . . . . . . . 2
The Mechanism of Action of Melphalan . . . 4 The Chemistry of Melphalan . . . . . 7
The Pharmacokinetics of Melphalan After I.V.
Administration . . . . . . . 10
The Relationship Between Pharmacokinetics
and Pharmacodynamics. . . 12
The In Vitro Evaluation of Chemosensitivity 14
Objectives . . . . . . . . . 17

II. EXPERIMENTAL PROCEDURES. . . 19

Chemicals and Reagents. . . 19 Ultraviolet Spectroscopy. . 21 High Performance Liquid Chromatography . . 21 Kinetic Experiments. . 23
Chloride Determinations. . 24
Data Analysis. . 24 Cell Cultures. . 25 Iododeoxyuridine Assay. . . 25 In Vitro Dynamic Experiments. . 26

III. RESULTS AND DISCUSSION. . . 28

High Performance Liquid Chromatography . . 28 Chemical Stability of Melphalan. . 36
Influence of pH. . . 36 Influence of Chloride. . . 47

iv









The Production of Chloride. . 58 Influence of Temperature . . . . . 62 Influence of Ionic Strength . . . . 65
The Stability of Melphalan in Infusion Media . 65 In Vitro Pharmacokinetics of Melphalan . . 68 Iododeoxyuridine Assay. . . . . 70 In Vitro Pharmacodynamics of Melphalan . . 79
Relationship Between the In Vitro
Pharmacokinetics and Pharmacodynamics
for Melphalan. . . . 88

IV. CONCLUSIONS. . . 94

LIST OF REFERENCES. . 101

BIOGRAPHICAL SKETCH. . . . 107











































V
















LIST OF TABLES



TABLE PAGE

1. Summary of In Vivo Pharmacokinetic Parameters for
Melphalan Following I.V. Administration . . 11

2. Microscopic Rate Constants for the Hydrolysis of
Melphalan and Kinetically Determined pKa
Values at 370C (u=0.5). . . . . 42

3. Rate Constants for the Hydrolysis of Melphalan at
pH 6.0 (u=0.5). . . . . . 50

4. Kinetic Data for the Hydrolysis of Melphalan
Showing the Effects of pH and Buffer
Com 9 osition in the Presence of 0.3 M Chloride
(50 C ) . . . . . . . . . . . 55

5. Activation Energies for Melphalan at Various pHs
(u=0 .5) . . . . . . . . . . 64

6. The Half Life of Melphalan in Infusion Media
(250C) . . . . . . . . . . . 67

























vi
















LIST OF FIGURES



FIGURE PAGE

1. Structure of melphalan showing three ionizable
groups. . . . . . . . . . . 5

2. Proposed mechanism for guanine alkylation by
melphalan according to Price, 1975. . 6

3. Proposed pathway for the hydrolysis of
melphalan. . . . . . . . . . 8

4. Chromatogram of an aqueous solution of melphalan
(peak 3) and its degradation products (peaks
1 and 2). . 29

5. Chromatogram of (a) water containing melphalan
peak (1), monohydroxymelphalan (peak 2), and
dihydroxymelphalan (peak 3), (b) blank
culture medium and (c) culture medium
containing melphalan and its hydrolysis
products. . . 30

6. Calibration curve of peak area versus
concentration for melphalan in culture medium
obtained using the conditions in figure 5. . 32

7. UV spectra for (a) melphalan anj (b)
dihydroxymelphalan (both 10- M) in HPLC
mobile phase (pH 2.3). . . . 33

8. Peak area versus time data for melphalan and
monohydroxymelphalan in water at 370C. The
curve for monohydroxymelphalan was simulated
using equation 2. . . 35

9. The degradation of melphalan as a function of
time at three initial concentrations (pH
7.47, u=0.5, 370C, r>O.999). . . . . 37

10. Representative first order plots for the
degradation of melphalan (initial
concentration=100 ug/ml, 370C, u=0.5).
Numbers represent pH values (r>0.999). . . 39


vii









11. Log koh5-pH profile for melphalan at 370C
(u= .5). The line has been simulated using
equation 13 and the constants in table 2 and
the symbols represent experimental values. . 40

12. UV absorbance of melphalan (5x10-5 M) at 260 nm
as a function of pH according to equation 21
(r=0.999). . 45

13. UV absorbance of dihydroxymelphalan (10-4 M) at
260 nm as a function of pH according to
equation 21 (r=0.997). . . . . . . 46

14. Relationship between the reciprocal first order
rate constant for the degradation of
melphalan and the concentration of chloride
according to equation 28 (pH 6.0, u=0.5,
500C, r=0.999). . . . . . . . . 49

15. Proposed scheme for the conversion of melphalan
(M) to monohydroxymelphalan (MOH). . . 52

16. First order plots of % remaining vs. time (500C,
pH 6.55, [Cl"]=0.3 M) showing the effects of
ionic strength (open u=0.6, closed u
unadjusted) and phosphate on the hydrolysis
of melphalan (r>0.998). . . .53

17. Relationship between ko /(k -kobs) for
melphalan and the bu er concentration plotted according to equation 30 (500C,
u=0.6). . . . . 56

18. Relationship between (k2'+k'"[OH-])/ k-l[ClV]
and pH for melphalan (50 C, u=0.6). . . . 57

19. Calibration curve for voltage (E) vs. log
chloride concentration (M) at pH 6.0 and 370C
(u=0.5). . . . 59

20. Relationship between the ratio of chloride
concentration to the initial melphalan
concentration and time (u=0.5). Lines were
simulated using equation 36 and the constants
in table 3 (r>0.997). . . . 61

21. Arrhenius plots of kobs vs reciprocal absolute
temperature for the hydrolysis of melphalan. . 63

22. Relationship between kobs and ionic strength (u)
for the hydrolysis of melphalan at pH 6.0 and
370C. . . . . . . . . . . 66



viii









23. Concentration dependencX for melphalan in
culture medium at 37 C and three initial
concentrations and in the presence of 106
neuroblastoma cells/ml (open) (r>0.993). . . 69

24. Concentration versus time profiles for melphalan
(M), monohydroxymelphalan (MOH), and
dihydroxymelphalan (M(OH)2) in culture medium
according to equations 40-42 (370C, r>0.993). . 71

25. Calibration curve for ny oblastoma cells (IMR
32) in which cpm of IUdR incorporated is
plotted against cell concentration. Verticle
bars represent standard deviations for the
mean of three measurements. . . 72

26. Initial response of neuroblastoma cells treated
for four hours with melphalan. . . . 74

27. Percent survival -f neurob astoma cells treated
with 10- 10- or 10- M melphalan for two
hours and assayed immediately and up to 5
days following treatment. . . 75

28. Cell survival vs. melphalan concentration for
roblastoma cells obtained using the
IUdR performed (a) immediately, and (b) 3
days following treatment and (c) a clonogenic
assay (Worthington-White, 1986). 77

29. Percent survival vs. exposure time for
neuroblastoma cells treated with 10-5 M
melphalan and assayed on 0, 1, 2, 3, and 4
days following treatment. . . 80

30. Percent survival vs. melphalan dose for
neuroblastoma cells treated for 20 minutes
and assayed 3 days following exposure. . . 82

31. Percent survival vs. exposure time for
neuroblastoma cells treated with 2x107 M
melphalan and assayed 3 days following
treatment. . . 84

32. Percent survival vs. melphalan dose for
neuroblastoma cells treated for 20, 40, 60,
90, 120, and 240 minutes and assayed 3 days
following treatment. . . 86

33. Percent survival vs. exposure time for
neuro lastoma cells treated with 10-7
2x10" 5x10-7, 10-6, 5x106 and 10-5M
melphalan and assayed 3 days later. . . 87


ix









34. Kill rate (a) and melphalan concentration (b) as
a function of exposure time for neuroblastoma
cells treated with an nitial drug
concentration of 2x10' M and assayed 3 days
following treatment. . . 91

35. Kill rate as a function of exposure time for
three nitial melpha an concentrations
[2x10 M (a), 5x10 M (b), and 10- M (c)]
and assayed 3 days following treatment. . . 92
















































x















Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy



CHEMICAL STABILITY AND IN VITRO CYTOTOXICITY OF MELPHALAN By


Susan Ann Stout


August 1987


Chairman: Hartmut Derendorf
Major Department: Pharmaceutics



Melphalan is an alkylating agent which has been used for a number of years in the treatment of a variety of cancers.

The compound undergoes rapid hydrolysis to form first monohydroxy- and then dihydroxymelphalan. The hydrolysis kinetics of melphalan have been investigated as a function of pH (1-13), chloride concentration (0-0.5 M), temperature (25-500C), and ionic strength (0.12-1.0) using HPLC.

Experimental data support the hypothesis that the rate determining reaction is the unimolecular conversion of melphalan to an ethyleneimmonium ion. Melphalan was found

to be more stable at low pH due to an increase in the overall positive charge on the molecule and a decreased



xi









tendency to take on an additional positive charge to form the ethyleneimmonium ion. The stability of melphalan was found to increase in the presence of chloride as a result of competition between chloride and water for the intermediate species.

The in vitro pharmacokinetics and pharmacodynamics of melphalan were studied using a model system of human neuroblastoma cells suspended in culture medium. The

degradation of melphalan in culture medium, monitored by HPLC, was found to be first order (half life= 1.2 h) and independent of the initial concentration (10-5-10-3 M) both in the presence and absence of cells. To evaluate the response of cells treated with melphalan, a short term assay based upon the cellular incorporation of [1251]iododeoxyuridine was used. Cell survival was determined by comparing the cell concentration for a treated sample to that for an untreated control. Cell response was

evaluated as a function of assay delay time (0-4 days), initial melphalan concentration (10-7-10-3 M), and duration of melphalan exposure (0-240 minutes). Cell survival was found to decrease dramatically during the four days following initial drug treatment indicating a delay between drug-cell interaction and apparent pharmacodynamic response. The dose response curve for the delayed effect was found to be very steep. The maximum rate of cell death for all effective doses was seen during the first 20 minutes of drug




xii









exposure. There was no direct correlation between drug concentration and kill rate in the investigated dose range.






















































xiii















CHAPTER I
INTRODUCTION


Cancer is one of the most frustrating and devastating diseases of our time. To the patient, the term cancer has

become almost synonymous with death. To the clinician, cancer represents a puzzle which is only beginning to be understood. Treatment options are generally limited to surgery, radiation therapy or chemotherapy, the application

of which varies depending upon the type of cancer being treated. Surgical methods are usually limited to cancers which exist as well defined tumors or those localized in a particular organ. Radiation therapy and chemotherapy are often quite effective in killing rapidly dividing neoplastic cells; however, these methods are nonspecific and toxic to normal cells as well. In some cases, the combination of

treatment options has led to the realistic expectation of long term survival or cure.



The Chemotherapy of Cancer


Most anticancer agents act by interfering with one or

more phases of the cell cycle thereby preventing normal cell division and growth. Consequently, cells most susceptible to the toxic effects of anticancer agents are those which



1







2

are in the process of division. Since, most anticancer agents are not specific for neoplastic cells, they are toxic to rapidly dividing normal cells as well. The problem of toxicity presents one of the major obstacles encountered in the chemotherapy of cancer.

Another barrier which frequently arises in the

chemotherapy of cancer is the development of drug resistant tumor cells. A number of approaches have been investigated

over the years to combat resistant cells. One of the earliest attempts was to use combination chemotherapy but it was soon recognized that many cells rapidly develop a cross resistance to a variety of drugs. Another approach has been the use of very high doses of chemotherapy which appears to overcome some of the mechanisms of resistance. This

approach, however, is limited by toxicity to normal cells, most notably bone marrow. To compensate for the toxicity which is encountered with the use of high dose chemotherapy,

this treatment is frequently followed by bone marrow transplantation. Drugs which exhibit myelotoxicity as the major dose limitation are the most reasonable candidates for this type of therapy.



The Treatment of Neuroblastoma with Melphalan


Neuroblastoma, the third most common childhood

malignancy, arises from the sympathetic ganglion cells and

frequently metastasizes to the bone, bone marrow, lymph







3

nodes, skin and liver (Stanfill and Hayes, 1986). It has been reported that survival rates are about 10% at three years for children diagnosed at 1-5 years of age and about 25% for those diagnosed at age 6 or older (D'Angio et al., 1985). Aggressive chemotherapeutic treatment, sometimes in combination with total body irradiation, has been somewhat successful in the treatment of neuroblastoma but this type of therapy is limited by bone marrow suppression.

During the past few years, many investigators have

focused on the use of high doses of chemotherapy, with or without radiation therapy, supported by autologous or allogeneic bone marrow transplantation for the treatment of

a variety of refractory cancers, including neuroblastoma. Early reports by Fernbach et al. (1968) and Evans (1968) indicated that the alkyating agent, melphalan, was ineffective in the treatment of neuroblastoma. However, in 1979, encouraging results were published by McElwain et al.

for the use of high dose melphalan followed by autologous bone marrow transplantation for the treatment of neuroblastoma. Unlike the more conventional drugs used to treat neuroblastoma, such as the anthracyclines and vinca alkaloids, melphalan shows no specific organ toxicity. The major dose limitation of melphalan is myelotoxicity (Pritchard et al., 1982). Since 1979, several investigations have been reported illustrating the usefulness of high dose melphalan therapy followed by bone







4

marrow rescue for the treatment of refractory neuroblastoma (Pritchard et al., 1982; Graham-Pole et al., 1982; Corringham et al., 1983; Lazarus et al., 1983; Strandjord et al., 1983; Graham-Pole et al., 1984; D'Angio et al., 1985).

Unfortunately, it is difficult to compare the various clinical studies due to a wide range of drugs being given concurrently with melphalan as well as diverse dosing regimens. It appears, however, that the incorporation of bone marrow rescue allows for a 3-5 fold increase in drug dosage which may lead to increased tumor cell kill (GrahamPole et al., 1984). At the present time, no standard dosage

regimen exists for high dose melphalan therapy in the treatment of neuroblastoma.



The Mechanism of Action of Melphalan


Melphalan (figure 1) is an alkylating agent which was

first synthesized in 1954 by Bergel and Stock. The compound is a nitrogen mustard derivative of L-phenylalanine and is often referred to as L-phenylalanine mustard. Melphalan is

believed to act by forming covalent linkages between the alkyl side chains of the compound and cellular constituents, primarily DNA. The mechanism of alkylation has been previously proposed (Price, 1975) and is represented in figure 2. The first step in the alkylation process is the

formation of an ethyleneimmonium ion (M ) which may react with a nucleophilic species to form a covalent bond.





5









3

NH CH2-CH2-Cl
2 2
CH-CH2 / CHCHCI
COOH

2







Figure 1: Structure of melphalan showing three ionizable
groups.






6


,CH2- CH -Cl COOH
R-N 2 CHC1
C-CH-CI 1 2
2 2 NH2


GUANINE in DNA
0

N CH
CH-CH-CI 2 N
2 2 R


OH CH 2CH2N-CH 2-CH2CI
N, N 1\
CH R
HN J/
2 N N MONO-ALKYLAT\ON R'


IMMONIUM ION GUANINE



CH-CH2N-CHCH2 OH 2 2 OH
N+ R +N
N /
CH HC
H N N N CROSS N N NH
2 I LINKAGE 1 2
R' R'



Figure 2: Proposed mechanism for guanine alkylation by
melphalan (Price, 1975).







7

Bifunctional alkylating agents, such as melphalan, may then form a second ethyleneimmonium ion which subsequently reacts with another nucleophilic species. One of the most biologically significant reactions appears to be the alkylation of the 7-nitrogen atom of guanine residues of DNA

(Calabresi and Parks, 1980). This reaction has several important consequences. First, the alkylation results in the formation of the enol tautomer whereas the keto tautomer is favored under normal circumstances. The keto tautomer of guanine forms a base pair with a cytosine residue, whereas, the enol tautomer forms a base pair with a thymine residue leading to miscoding. Secondly, alkylation of the

7-nitrogen of guanine can result in splitting of the imidazole ring. And finally, alkylation of two guanine residues may lead to both inter- and intra-strand

crosslinking which is believed to be the most serious cytotoxic effect of alkylating agents (Pratt and Ruddon, 1979).



The Chemistry of Melphalan


Melphalan is a very unstable compound which forms

monohydroxymelphalan (MOH) and dihydroxymelphalan (M(OH)2) in aqueous solution (figure 3). The hydrolysis of melphalan

has been studied in both plasma (Furner et al., 1976) and buffer at pH 7.4 (Flora et al., 1979; Chang et al., 1978).

This reaction is also believed to be the major route of






8







,CH CHC 2-C1 k, X-k5 CH2 CH2- X
RN2 Cl-+ R-N+ : R-N
kCH2-CHC CH2 CH2C] CH2CH-Cl

k 2 H20

CCH CH2-CH-OH
R=CCH H' + R-N MO

j k 3 1 -3



H2-CH20H kCH2-CH2-OH CH2 -C H
H+ + R-N R -N+ R-N
CHf-CH2-OH '6 CH-CHf X

Cl MOH.X









Figure 3: Proposed pathway for the hydrolysis of melphalan.







9

biotransformation of the drug (Alberts et al., 1979). It has been reported that the stability of melphalan is improved by decreasing the pH, decreasing the temperature, and by the addition of chloride (Flora et al., 1979; Chang et al., 1979). Detailed information however, on these effects and the mechanism of hydrolysis of melphalan is limited.

The degradation of melphalan (figure 3) has been

resonably assumed (Chang et al., 1978; Flora et al., 1979) to proceed via the mechanism proposed earlier (Golumbic et al., 1946; Fruton and Bergmann, 1946; Bartlett et al., 1947)

for the transformation of methyl-di(2-chloroethyl)amine in water. The rate limiting step for the transformation of 2-chloroethylamines is the cyclization reaction to give corresponding ethyleneimmonium ions with the loss of chloride (Bartlett and Swain, 1949). Ethyleneimmonium ions

are highly susceptible to substitution by nucleophiles which, in the case of water, results in hydroxylation. The

hydroxylated analogues of bifunctional alkylating agents, such as melphalan, may react further to produce the corresponding dihydroxy analogues. Side reactions with nucleophiles (X) other than water may occur giving rise to

MX, MXC1, MXOH, or MX2. Figure 3 shows the generalized mechanism of the nucleophilic substitution of melphalan in aqueous solution.







10

The Pharmacokinetics of Melphalan After I.V. Administration


With few exceptions, authors have found that

concentration-time profiles for melphalan following i.v. administration are best fit by a two compartment body model. Table 1 summarizes pharmacokinetic data in the literature following i.v. administration. The general expression for a two compartment body model in which drug is introduced into the systemic circulation is given by equation 1.

(1) C =ae-at + be
p

In this expression, Cp is the plasma drug concentration, t is the time following administration, a and 3 are the hybrid constants for the distribution and elimination phases, respectively, and a and b are the y-axis intercepts for the

extrapolated distribution and elimination portions of the curve, respectively.

The distribution phase for melphalan is rapid with a half life ranging from 5 to 13 minutes. The elimination phase half life varies from about 0.6 to 1.7 hours after i.v. injections of 15 to 180 mg/m2. The one exception to this is a study conducted by Ninane et al. (1985) which reported a

terminal half life of 3 hours following a 180 mg/m2 dose. The data indicate that melphalan pharmacokinetics is dose
2
independent for i.v. injections up to 180 mg/m In

addition, after normalizing for body surface area, there appears to be no significant difference between pharmacokinetic parameters measured in children and adults.







11

TABLE 1

Summary of In Vivo Pharmacokinetic Parameters for Melphalan
Following I.V. Administration


Dose Age tl/2's tl/2,g Cltot Vd,

[mg/m2] Group [min] [h] [ml/min/m2] [1/m2] Reference


15 A 7.7 1.38 203 24.0 Bosanquet (1982)
23 A 12.6 1.44 192 24.2 Alberts (1979)
140 6.2 0.89 *** *** Davis (1982)
140 C ** 1.64 274 26.1 Taha (1983)
140 C 8.8 0.72 257 16.1 Gouyette (1986)
140 A 6.9 0.83 525 37.5 Gouyette (1986)
140 C 7.0 0.61 447 23.5 Ardiet (1986)
140 A 5.9 0.79 372 24.8 Ardiet (1986)
180 C 6.6 3.00 357 85.5 Ninane (1985)
180 C 10.5 0.80 498 35.6 Gouyette (1986)
180 A 5.9 0.68 532 31.3 Gouyette (1986)


A=adults
C=children
* ages not reported
** one compartment body model used
*** data not available







12

The Relationship Between Pharmacokinetics and Pharmacodynamics


Holford and Sheiner (1986, p. 189) described the terms pharmacokinetics and pharmacodynamics by saying that

"pharmacokinetics is what the body does to the drug; pharmacodynamics is what the drug does to the body." Investigators have recently recognized the necessity of relating the pharmacokinetics of a drug with an elicited pharmacodynamic response. Gibaldi and Perrier (1982)

described two factors which determine the relationship between drug concentration and a given response. First, the drug concentration may be directly or indirectly related to a response and, second, the drug may interact reversibly or

irriversibly with a receptor. Considerable progress has been made in describing the kinetics of reversible pharmacologic responses (Gibaldi and Perrier, 1982); and as a result, many investigators are conducting clinical studies to correlate pharmacokinetics and pharmacodynamics for drugs which act in this manner.

The kinetics of irreversible pharmacologic responses have

been studied in theory and in vitro (Gibaldi and Perrier, 1982), however, clinical pharmacokinetic-pharmacodynamic

evaluations for drugs in this category (e.g. certain antibiotics and anticancer agents) are limited. In the field of antibiotic therapy, in vitro susceptibility testing of microorganisms has been quite successful as a basis for







13

rational antibiotic selection and in the determination of minimum inhibitory concentrations. Garrett (1978)

summarized a series of extensive investigations which focused on the relationship between microbial generation rate and antibiotic concentration in vitro. Only recently

has in vitro pharmacodynamic data for antibiotics been integrated with clinical pharmacokinetics to design appropriate dosing regimens (Schentag et al., 1986).

In the field of cancer chemotherapy, very few attempts have been made to relate pharmacokinetic and pharmacodynamic measurements for anticancer agents. Chemotherapeutic agents may be grouped into two broad classes. One class includes

agents, such as methotrexate, cytosine arabinoside, and vincristine, which act only at specific phases of the cell

cycle and are thus referred to as cell-cycle specific. Another class consists of agents, such as melphalan and cyclophosphamide, which exert their pharmacologic action at any phase of the cell cycle and are termed phasenonspecific. In 1971, Jusko presented a pharmacodynamic model, which was later modified by Gibaldi and Perrier (1982), for phase-nonspecific chemotherapeutic agents. Both of these models predict that the logarithm of the fraction

of surviving cells is linearly related to drug dosage provided that the pharmacodynamic measurement is made after all of the drug has been eliminated. The models, however,

do not take into account the pharmacokinetics of the drug







14

and assume that events such as absorption, distribution, and

elimination are complete before the cytotoxic effects are apparent. Jusko concluded that the cytotoxic effect will only be a function of the total dose, independent of the route of administration. There is limited information in the literature to experimentally validate this hypothesis.

Pharmacodynamic investigations of this type are especially suited for in vitro cell culture systems provided that the pharmacokinetics of the drug can be modeled.



The In Vitro Evaluation of Chemosensitivity


Treatment of cells with cytotoxic agents may result in

(1) immediate cell death, (2) delayed cell death in which the cell may undergo several divisions before its eventual death, (3) cell insult in which the metabolic processes are temporarily halted but the cell maintains indefinitely its ability to reproduce, or (4) no effect at all. It is generally accepted that the permanant loss of proliferative capacity is the most relevant criterion for evaluating cell kill and, therefore, the lethal effect of a chemotherapeutic agent (Roper and Drewinko, 1976).

The validity of results obtained from in vitro

chemosensitivity studies relies upon the ability of the assay to accurately assess cell death as well as the type of assay which is used. There are basically two types of in vitro assays which may be used to assess cell kill following







15

drug treatment. The first type is a long term assay which

measures the proliferative capacity of the cell following drug treatment. The second type is a short term assay which is performed immediately following treatment and, therefore,

measures only the immediate response of the cell to the drug. The conclusions drawn from in vitro chemosensitivity studies may be quite different depending upon which type of

assay is used to assess viability since the two types of assays measure very different responses.

One of the most popular of the long term assays is a clonogenic stem cell assay published in 1977 by Hamburger and Salmon. Stem cells are defined as those cells within a population having unlimited growth potential. In the clonogenic assay, cell death is measured by the inability of

the drug treated cells to form colonies as compared to untreated cells. The concept that clonogenic cells in vitro

represent stem cells in vivo and that stem cells are the most important target for chemotherapeutic agents has been the basis for the widespread acclaim of clonogenic assays (Weisenthal and Lippman, 1985). However, it has been suggested that, while stem cells are generally the most sensitive, they are not the only relevant target for chemotherapeutic agents (Drewinko et al., 1981; Weisenthal et al., 1984) and it is possible that clonogenic cells in vitro may not reflect stem cells in vivo (Weisenthal and Lippman, 1985). In addition, clonogenic assays are







16

extremely time consuming requiring 10-14 days of incubation for the production of a sufficient number of colonies. With

this type of assay, it is not practical to perform a calibration curve for the determination of the number of viable cells due to the large number of cells required. Furthermore, the assays lack the sensitivity necessary to evaluate thoroughly the toxic effects of drugs and the results are often highly variable.

Many investigators have chosen to use short term assays, such as dye exclusion, 51Cr release, or precursor incorporation, to quantitate cell viability in vitro due to

the many practical problems associated with clonogenic assays (Roper and Drewinko, 1976; Weisenthal et al., 1984;

Weisenthal et al., 1983; Bhuyan et al., 1976; Rupniak et al., 1983; Roper and Drewinko, 1979; Sondak et al., 1984). Recently, an assay was published by Gee et al. (1985) which is based upon the cellular incorporation of a radiolabelled DNA precursor [125IIiododeoxyuridine (125IUdR). The method was used to detect cell death following in vitro treatment of leukemia cells with antibody and complement and was shown to be superior in terms of speed, sensitivity, and reproducibility to both a clonogenic assay and a 51Cr release assay. The technique also allows for the inclusion

of a calibration curve in each experiment to accurately assess the number of viable cells remaining following treatment.







17

When using short term assays, such as the 125IUdR assay,

to measure in vitro chemosensitivity, there are several important points which must be considered. First, short term assays measure responses of the total cell population

in contrast to clonogenic assays which measure only the small, clonogenic fraction. Secondly, short term assays performed immediately following drug exposure only measure the immediate cell response while delayed cell responses (death and/or recovery) are unaccounted for. These points make it difficult to make comparisons of results obtained using the two types of assays. Short term assays may, however, be delayed after drug treatment making it possible

and practical to measure both immediate and delayed cell responses.



Objectives


The objective of this project is to develop a thorough understanding of the chemistry of melphalan as well as its cytotoxic activity in vitro. Melphalan is known to be a very unstable compound, however, detailed information on the

mechanism of its hydrolysis is not available in the literature. Therefore, the first part of this project is an in depth evaluation of the hydrolysis kinetics of melphalan

as a function of pH, chloride concentration, temperature, and ionic strength.







18

The pharmacological activity of melphalan can be assessed by examining its lethal effect on cancer cells. The second

part of this project is an investigation of the in vitro cytotoxicity of melphalan as a function of exposure time and

drug concentration. A model system consisting of human neuroblastoma cells suspended in culture medium will be used to conduct these studies.

The chemical hydrolysis of melphalan is believed to be the major route of biotransformation. It is, therefore, an

ideal candidate for an in vitro evaluation since the concentration changes with time at a rate which is comparable to that in vivo. By following both the

concentration of melphalan as well as its cytotoxic activity as a function of time, it may be possible to characterize a

relationship between the in vitro pharmacokinetics and pharmacodynamics. Such a relationship could provide useful information for proposing a dosage regimen which would lead to the optimum therapeutic response.
















CHAPTER II
EXPERIMENTAL PROCEDURES



Chemicals and Reagents


Melphalan was obtained from Sigma Chemical Co. (St.

Louis, MO) and was used as received. Dihydroxymelphalan was prepared by heating an aqueous solution of melphalan at 600C for 2 hours (Furner et al., 1976). HPLC grade methanol and

sodium dodecyl sulfate and ACS grade buffer constituents were obtained from Fisher Scientific Co. (Fair Lawn, NJ). Sodium hydroxide (1 M) was prepared from Dilut-it Analytical Concentrate supplied by J.T. Baker Chemical Co.

(Phillipsburg, NJ). Deionized water was filtered through 0.45 um nylon filters (Rainin Instrument Co., Inc., Woburn, MA) before use.

Kinetic experiments conducted in the absence of chloride

were performed using McIlvaine buffers as described by Elving et al. (1956) in the range of pH 2.2 to 8.0. Below

pH 2.2, nitric acid was used. Between pH 8.8 and 10.2, boric acid-sodium hydroxide buffers were used as described

by Perrin and Dempsey (1974) and above pH 10.2, sodium hydroxide was used. For these studies, the ionic strength

(u) was adjusted to 0.5 M with sodium nitrate unless otherwise stated.


19







20

To evaluate the effects of pH and buffer constituents on

the hydrolysis of melphalan in the presence of chloride, acetic acid-sodium acetate buffers (0.025-0.2 M) were used between pH 3.7 and 4.7. For solutions between pH 5.8 and 7.7, potassium phosphate monobasic-sodium phosphate dibasic buffers (0.025-0.1 M) were used. Boric acid-sodium

hydroxide buffers (0.015-0.05 M) were used in the pH range of 8.66 to 10.0 and sodium hydroxide was used above pH 10.0. The ionic strength was adjusted with sodium nitrate. In all

experiments, the pH of each solution was measured at the temperature of the experiment with a Fisher MicroProbe Combination Electrode and a digital pH millivolt meter model 611 (Orion Research, Cambridge, MA).

Culture medium (RPMI 1640), fetal bovine serum (FBS), trypsin containing ethylenediaminetetraacetic acid (EDTA) and Hankes balanced salt solution were obtained from GIBCO (Grand Island, NY). [125I]Iododeoxyuridine (125IUdR) was obtained from Amersham Corporation (Arlington Hts, IL) and

was diluted to an activity of 10 uCi/ml in RPMI 1640 containing 20% FBS and 0.2% gelatin (Sigma Chemical Co.) immediately prior to use. Bovine serum albumin (BSA, Sigma Chemical Co.) was adjusted to a concentration of 0.5% w/v in normal saline and stored at 40C. Trichloroacetic acid (TCA,

Sigma Chemical Co.) was prepared as a 20% w/v aqueous solution.







21

Ultraviolet Spectroscopy


Ultraviolet absorption data were obtained for freshly prepared melphalan (10-4 M) and dihydroxymelphalan (10-4 M) in HPLC mobile phase consisting of 40% methanol and 60% buffer (pH 2.0) containing 0.015 M sodium dodecyl sulfate.

Spectra were recorded using a Cary 219 Spectrophotometer (Varian, Palo Alto, CA). The absorption of freshly prepared melphalan and dihydroxymelphalan was further investigated in the pH range of -0.3 to 4.54 and 1.2 to 7.6, respectively, in which the pH was adjusted using either hydrochloric acid or sodium hydroxide.



High Performance Liquid Chromatography


The high performance liquid chromatography (HPLC) system

consisted of a Constametric IIG pump (LDC, Riviera Beach, FL), a Negretti model 190 injector (20 ul sample loop) (HPLC Technology, Palos Verdes Estates, CA) and a Spectromonitor D variable wavelength detector (LDC) operated at 260 nm. Some studies were conducted using a Micrometrics 728 autosampler

(Micrometrics, Norcross, GA) with a Valco RC6U injection valve fitted with a 20 ul sample loop (Valco Instruments, Houston, TX) in which the sample vials were filled immediately prior to injection. Peak areas were measured

using either a Hewlett Packard 3392 Integrator (Hewlett Packard Co., Avondale, PA) or a data system consisting of an







22

HP87 microcomputer (Hewlett Packard Co., Palo Alto, CA),a Nelson Analytical Interface with software model 366, version

2.2 (Cupertino, CA), and a Hewlett Packard dual 3 1/2 inch disc drive (model 9121). The chromatograms were displayed on a Recordall 5000 strip chart recorder (Fisher Scientific Co.).

The loss of melphalan from aqueous solutions was

monitored by reversed phase HPLC using an ODS Hypersil column (HETP, Sutton, UK) (150 mm x 4.6 mm i.d.) prepared as described by Bristow et al. (1977), and a mobile phase of 50% methanol and 50% acetate buffer (pH 4.7, 0.1 M). A flow

rate of 1.8 ml/min and ambient temperature were used throughout. Aqueous solutions of melphalan were injected directly onto the HPLC column.

To monitor the change in the concentration of melphalan, monohydroxymelphalan and dihydroxymelphalan in culture

medium, an HPLC system consisting of a cyanopropylsilane column (Dupont Instruments, Wilmington, DE) (150 mm x 4.6 mm i.d.), and a mobile phase of 40% methanol and 60% buffer (pH

2.0 adjusted with phosphoric acid and containing 0.015 M sodium dodecyl sulfate) were used. The mobile phase flow

rate was 1.0 ml/min and ambient temperatures were used throughout. Culture medium containing melphalan was incubated at 370C and aliquots were withdrawn at appropriate intervals and diluted 1 to 10 with mobile phase immediately prior to injection.







23

Kinetic Experiments


The influence of pH (1-13, u=0.5, 370C), temperature (25-50'C, pH 2, 6, 9, 11, u=0.5) and ionic strength (0.15-1.0, pH 6, 370C) on the degradation of melphalan was investigated. Melphalan (100 ug/ml) was dissolved in the appropriate buffer with the aid of sonification and placed in a thermostated (+- 0.10C) water bath. The solution was allowed to equilibrate to temperature for 10 minutes before

the first sample (100 ul) was taken and injected (20 ul) directly onto the HPLC column. The first injection was taken as t=0 since in all cases the loss of melphalan was

first order and independent of its initial concentration. Subsequent samples were injected at appropriate intervals and the peak areas found for melphalan were recorded as a percentage of that found at t=0. Peak areas of freshly prepared standards were found to be linearly related to the concentration injected (0.1-200 ug/ml). Single

determinations were made for each set of samples and sampling was continued for at least 4 half lives.

Kinetic experiments were also carried out in the presence

of chloride (0.3 M) to determine the influence of pH (3.7-10), buffer composition (acetate, phosphate or borate)

and buffer concentration (15-200 mM) on the rate of hydrolysis of melphalan at 500C. Duplicate solutions were

prepared and analyzed for each set of conditions and sampling was continued for at least 4 half lives.







24

Chloride Determinations


A combination chloride electrode model 96-17B and a digital pH/milivolt meter model 611 (both Orion Research, Cambridge, MA) were used to monitor the production of chloride due to the degradation of melphalan (figure 3). A Haake Circulator (Saddle Brook, NJ) was used for temperature control (t 0.10C). Standard solutions were prepared in the pre-determined linear range of 4.4 x 10-4 to 10-1 M chloride in phosphate buffer (pH 6.0). Standard curves of voltage

(E) vs. log chloride concentration were constructed at each temperature under study. Melphalan (50 mg) was dissolved in a mixture of ethanol (1.5 ml) and nitric acid (20 ul). This

solution was immediately transferred to a 100 ml beaker containing 48.5 ml of buffer (pH 6.0, u=0.5), equilibrated to temperature. Voltage readings were taken as a function

of time (up to 6 hours) and compared with the calibration curve to obtain the chloride concentration. Readings were taken after 48 hours and the initial melphalan concentration

was calculated as the chloride concentration at this time divided by 2 (figure 3).



Data Analysis


Two nonlinear least squares regression programs were used

in these studies to obtain estimates of the various parameters based upon the experimental data (SASNLIN,







25

version 83.4, SAS Institute, Inc., Cary, NC, and PCNONLIN, 1985, C.M. Metzler and D.L. Weiner, Statistical Consultants, Inc., Lexington, KY).



Cell Cultures


A human neuroblastoma cell line (IMR 32) was used in these experiments and was obtained from the American Type Culture Collection (Rockville, MD). The cells were maintained as monolayers at 370C in a humidified atmosphere with 5% CO2 in tissue culture flasks (150 cm2) containing RPMI 1640 supplemented with 20% FBS. Before use, the cells were removed from the flasks using 10 ml trypsin (5 g/l + 0.2 g/l EDTA in balanced salt solution) and washed three times in culture medium. Viabilities were determined using trypan blue exclusion.



Iododeoxyuridine Assay


Following drug treatment and washing, each cell pellet was exposed to 100 ul of 125IUdR (10 uCi/ml) for 2 hours at 370C. At this time, the cells were washed three times in culture medium to remove residual, unincorporated 125IUdR and the proteins were precipitated with 100 ul BSA (added as a co-precipitate) and 2 ml TCA. The precipitate was allowed

to form overnight at 40C at which time the tubes were centrifuged for 10 minutes at 2000 rpm. The supernatant was







26

decanted and each cell pellet was analyzed in an automatic

gamma counter (1275 Minigamma Gamma Counter, LKB, Turku, Finland).



In Vitro Dynamic Experiments


Viable cells were suspended in RPMI 1640 + 20% FBS at the desired concentrations. A calibration curve (1x104 to 2x106 cells/ml) was included in each experiment to establish the relationship between 125IUdR incorporation (counts per minute, cpm) and the number of viable cells. Blanks were also included in each experiment to allow for correction of

the gamma counter background noise and residual 125IUdR. The mean cpm for the blank was subtracted from the mean cpm for the standards and samples. Each sample contained 1x106 cells/ml and was treated with either culture medium alone or culture medium containing melphalan. The cells were

incubated at 370C for the desired time interval, washed three times with culture medium and subjected to the 125IUdR assay. Percent cell survival was calculated as the ratio of

the cell concentration of the treated sample to the cell concentration of the control multiplied by 100.

For the delayed assay studies, control and treated cells were plated in culture flasks (25 cm2) containing RPMI 1640 + 20% FBS and maintained at 370C in a humidified atmosphere containing 5% CO2 for various lengths of time (up to 4 days) prior to the 125IUdR assay. Control and treated cells were







27

trypsinized and washed three times in culture medium and the

entire contents of each flask (supernatant and trypsinized cells) were subjected to the assay.
















CHAPTER III
RESULTS AND DISCUSSION



High Performance Liquid Chromatography


High performance liquid chromatography (HPLC) was

utilized to monitor the loss of melphalan from aqueous solution and culture medium. Figure 4 displays a

chromatogram of an aqueous solution containing melphalan (peak 3) and degradation products (peaks 1 and 2). Under the conditions employed, melphalan eluted with a retention time of 3.6 minutes and was well separated from the degradation products. In order to monitor the changing concentration of the hydrolysis products in addition to that

of melphalan, the HPLC conditions had to be modified. Figure 5 displays a chromatogram of (a) water containing melphalan (peak 1), monohydroxymelphalan (peak 2), and dihydroxymelphalan (peak 3), (b) blank culture medium and

(c) culture medium containing melphalan, monohydroxymelphalan, and dihydroxymelphalan. No interferences were observed between the peaks corresponding

to melphalan and monohydroxymelphalan and those resulting from material present in the culture medium. There was, however, a small peak present in culture medium which eluted




28







29




0.3
3





Column: ODS Hypersil (5 micron,
15 cm x 4.6 mm, i.d.)
Mobile Phase: 50% Methanol + 50%
Acetate Buffer.
(pH 4.7, 0.1M)
Flow Rate: 1.8 ml/min 02
UV Detection: 260 nm Temperature: Ambient



C

0




0.1









2


0L




0 2 4 6 Time (min)





Figure 4: Chromatogram of an aqueous solution of melphalan
(peak 3) and its degradation products (peaks 1
and 2).







30



2 a




3


Column: Cyanoproplysilane (5
micron, 15 cm x 4.6 mm,
i.d.)
Mobile Phase: 40% Methanol + 60%
Buffer (pH 2.0
+ 0.015 M sodium dodecyl sulfate) Flow Rate: 1.0 ml/min UV Detection: 260 nm Temperature: Ambient




0 10 20
mnn



2 c
b












o 10 20 0 10 20
min
min




Figure 5: Chromatogram of (a) water containing melphalan
peak (1), monohydroxymelphalan (peak 2), and
dihydroxymelphalan (peak 3), (b) blank culture
medium and (c) culture medium containing
melphalan and its hydrolysis products.







31

just prior to the peak corresponding to dihydroxymelphalan

making it difficult to measure low concentrations of the compound. Figure 6 represents a calibration curve of peak

area versus concentration for melphalan in culture medium obtained using the conditions listed in figure 5.

In order to calculate the concentrations of

monohydroxymelphalan and dihydroxymelphalan using the

melphalan calibration curve, the absorptivities of these compounds had to be ascertained in HPLC mobile phase (pH 2.3). Figure 7 displays the UV spectra for melphalan and dihydroxymelphalan in HPLC mobile phase (pH 2.3). The molar absorptivity at 260 nm (in HPLC mobile phase, pH 2.3) was determined to be 18,800 for melphalan and 4200 for dihydroxymelphalan.

The molar absorptivity of monohydroxymelphalan could not be evaluated by UV spectroscopy because a pure sample of the compound was not available. However, it was possible to estimate the molar absorptivity of monohydroxymelphalan using HPLC data. In order to use this approach, it was necessary to assume that the area of the HPLC peak corresponding to monohydroxymelphalan is proportional to the concentration of that compound in the range being investigated (10-6 to 10-5 M). It was also necessary to assume that the molar absorptivity of monohydroxymelphalan is different from that of melphalan (in HPLC mobile phase, pH 2.3). This is a reasonable assumption based on the






32





50

0 sIope=4.37x10 (- 2.2x10 )

4 4
40 int.=1.29x10 (-1.1x1O
2
r =0.999
If) 4
1 S =2.lxlO
C 30 xy







0 2
ci)
nL 100
0 1 2 3 4 5 6 7 8 9 10

Melphalan [M x 106


Figure 6: Calibration curve of peak area versus
concentration for melphalan in culture medium
obtained using the conditions in figure 5.







2.0 33








1.6








1.2 a





0


0.8








0.4

b






0 r i
220 260 300 nm



Figure 7: UV spectra for (a) melphalan 4and (b)
dihydroxymelphalan (both 10- M) in HPLC mobile
phase (pH 2.3).







34

difference in the absorptivity of dihydroxymelphalan and melphalan at the pH of the mobile phase. Figure 8 displays

the HPLC peak area versus time data for melphalan and monohydroxymelphalan in in water at 370C. The data suggest that the rate constants for the formation and degradation of monohydroxymelphalan are equal. The HPLC peak area (PA) for

monohydroxymelphalan as a function of time can then be written

-k t
(2) PA = rPA k te 1
MOH M 1
0

where r represents the ratio of the molar absorptivity for monohydroxymelphalan to that for melphalan in mobile phase, k1 is the rate constant for the degradation of melphalan as well as the formation and degradation of monohydroxymelphalan and the subscripts, MOH and Mo refer to monohydroxymelphalan and melphalan at time 0, respectively. The values of r and k1 may be estimated by fitting the HPLC

peak area versus time data for monohydroxymelphalan to equation 2 using nonlinear regression. The value of r multiplied by the molar absorptivity of melphalan in mobile phase gives an estimate of 14,690 as the molar absorptivity for monohydroxymelphalan in mobile phase. Correction factors of 1.28 and 4.55 were then calculated for

monohydroxymelphalan and dihydroxymelphalan, respectively, by taking the ratio of the molar absorptivity of melphalan to that for each degradation product. The correction factor

was multiplied by the concentration obtained using the






35





100






0

x

o 10


40

LL AMOH





0 1 2 3 4 5 6

Time [h]


Figure 8: Peak area versus time data for melphalan and monohydroxymelphalan in water at 370C. The curve
for monohydroxymelphalan was simulated using
equation 2.







36

melphalan calibration curve to determine the concentration of each hydrolysis product.



Chemical Stability of Melphalan


At constant pH, ionic strength and temperature, the

overall loss of melphalan was first order over at least four half lives and independent of the initial melphalan concentration (figure 9). These observations were

consistent with previous reports on the degradation of melphalan (Flora et al., 1979; Chang et al., 1978, 1979). The pseudo first order rate constants (kobs) were calculated by least square linear regression from the slope of linear

plots of the logarithms of the percentage of melphalan remaining against time.



Influence of pH

Melphalan has three ionizable functional groups (figure 1) and up to four species may exist in aqueous solution such that

2++
(3) [M] = [MH ] + [MH2] + [MH] + [M-]
3

The appropriate dissociation constants for melphalan are given by equations 4-6, respectively.

(4) K = [H ][MH]/[MH 2+
a,1 2 3
(5) Ka,2 = [H ][MHI/[MH2]
a,3 =2
(6) K a,3 =[H+][M-]/[MH]





37



100.0
5
M [Mx 10 ]
0 39.7 A 8.2 10.0 1.6


- 10.0 xA



0
1.0





0.1 .
0 60 120 180 240 300
time [min


Figure 9: The degradation of melphalan as a function of
time at three initial concentrations (pH 7.47,
u=0.5, 370C).







38

The fractions of each species present in aqueous solution may be given by equations 7-10.

(7) f(MH2+ + 3/B
3
(7) f(MH3 ) = [H ]2K/B
(8) f(MH) = [H ] K /B
2 a,1

(9) f(MH) = HI+1K lKal/

(10) f(M) = Ka,1Ka,2Ka,3/B


where


(11) B = [H ]3 + [H ]2Ka,l + [H ]Ka,l Ka,2 + Ka,l Ka,2 Ka,3


The influence of pH on the degradation of melphalan was investigated over the range of 0.91 to 13.0 at 370C and an ionic strength of 0.5 (figure 10). There was no evidence of any general or specific acid/base catalysis. Instead the data were consistent with a unimolecular reaction with M- M (figure 3) being the rate determining step. The form of the

log kobs-pH profile (figure 11) was consistent with the different rates at which the four ionic species of melphalan (MH32+, MH2 MH, M-) are converted to their corresponding ethyleneimmonium ions. The respective micro rate constants for these reactions are kl(MH3 2+), kl(MH2 ), kl(MH), kl(M-). It follows that kobs is given by

MH 2 2+ MH MH M
(12) kob k 3 f(MH+) + k 2f (MH ) + k f(MH) + k f(M )
obs 1 3 1 2 1 1






39



100





0.91
C
C

E
10-1.5

C


ci)


4.86 1.96
8.75 1,3.0


0 1 2 3 4 5 6

Time [h]




Figure 10: Representative first order plots for the
degradation of melphalan (initial
concentration=100 ug/ml, 370C, u=0.5). Numbers
represent pH values (r>0.999).







40



10.0













L 1.0


0










0 .1 -,

0 2 4 6 8 10 12 14


pH






Figure 11: Log k s-pH profile for melphalan at 370C (u=0.?3. The line has been simulated using
equation 15 and the constants in table 2 and the
symbols represent experimental values.







41

in which f represents the fraction of each species present. Substitution of equations 7-10 into equation 12 gives

(13) kobs
2+ +kMH 2+3 MH + 2 MH M
k3 [H]3 +k kM2[HI]K +k kM[H ]K K +kM K K K
+11 al 1 a,l a,2 1 a,l a,2 a,3


[H ]3 [H' ]2Ka,l + [H a,1Ka,2 + Ka, Ka,2 Ka,3

The kinetic data were fit to equation 13 using nonlinear least squares regression to obtain the values for the rate constants and the dissociation constants. An initial

fitting of the data to equation 13 gave a value of 109 for kl(MH32+) indicating that the contribution of the reaction to equation 13 is negligible. This is not surprising since a protonated 2-chloroethylamine is unlikely to form a cyclic ethyleneimmonium ion. The kinetic data (figure 11) were reanalyzed using a value of 0 for kl(MH32+) and the estimates for the rate constants and pKa values are given in table 2. The pKa values of 2.75 and 9.17 (u=0.5, 370C) compare well with the thermodynamic pKas (Merck Index, 1983) of 2.59 and 9.25 for the carboxylic acid and amino groups of phenylalanine, respectively. Thus the pKas of 1.42, 2.59, and 9.25 may be ascribed to positions 1,2, and 3 in figure 1, respectively.

The values of k1 listed in table 2 increase with decreasing protonation of the melphalan species such that
2+ +
MH 2 MH MH M
k 13 < k 2 < k < k







42

TABLE 2

Microscopic Rate Constants for the Hydrolysis of Melphalan
and Kinetically Determined pKa Values at 370C (u=0.5)


Parameter Value

kl(MH2 ) 0.74 h-i

kl(MH) 0.98 h-i

kl(M-) 2.02 h-i

pKa,1 1.42

pKa,2 2.75

pKa,3 9.17







43

This order reflects the magnitude of the positive charge on the various species and the tendency of the species to take on an additional positive charge and form the corresponding ethyleneimmonium ion.

The tertiary amine pKa values of both melphalan and dihydroxymelphalan were evaluated by spectrophotometry at the wavelength of maximum absorption, 260 nm. This could be accomplished because the unprotonated nitrogen is associated

with the chromophore by conjugation of the lone pair electrons with the aromatic system. Protonation of the nitrogen disrupts the conjugation and thereby causes a change in the UV absorption. At a given wavelength, the total absorbance (AT) can be expressed as

(14) AT = A a+ Ab

where Aa and Ab are the absorbances of the acidic and basic species, respectively. The absorbance of any species at a given wavelength is given by

(15) A = ECl


where E is the molar absorptivity at that wavelength, C is the molar concentration of the species, and 1 is the cell path length. Equation 14 can then by rewritten as

(16) AT = C Cl + EbCbl

in which the subscripts a and b represent the acidic and basic species, respectively. The total concentration (CT) is given by

(17) CT = Ca + Cb







44

and the acid dissociation constant (Ka) can be expressed as

(18) K a = [H ]Cb/Ca


After rearrangements and substitutions of equations 17 and 18, it can be shown that

(19) Ca = C T[H ]/([H ] + K )
a T a

and

(20) Cb = C K /([H ] + K

Equations 19 and 20 can then be substituted into equation 16 to give equation 21.

(21) A = (e aC [H ]1 + bC K a1)/([H ] + Ka Absorbance and pH data were fitted to equation 21 using nonlinear regression to obtain estimates of the pKa values.

Figures 12 and 13 display the absorbance of melphalan and dihydroxymelphalan, respectively, as a function of pH. A pKa value of 1.4 was obtained for the tertiary amine of melphalan which agrees well with the value obtained from the kinetic experiments (pKa=1.42). The pKa of the tertiary amine of dihydroxymelphalan was estimated to be 3.9

indicating that the presence of the two chloride atoms in melphalan strongly influences the acidity of the functional group.





45


1.0



0.8



V 0.6C
0
on 0.4



0.2



0.0 I I *
-1 0 1 2 3 4 5

pH



Figure 12: UV absorbance of melphalan (5x10-5 M) at 260 nm
as a function of pH according to equation 21
(r=0.999).





46


1.6

1.4

1.2

0 1.00
C

-Q 0.60.4

0.2


0 1 2 3 4 5 6 7 8
pH




Figure 13: UV absorbance of dihydroxymelphalan (10-4 M) at
260 nm as a function of pH according to equation
21 (r=0.997).







47

Influence of Chloride

Chang et al. (1979) have shown that the degradation of melphalan is inhibited by the addition of sodium chloride.

This may be attributed to the competition between the chloride ions and water for the ethyleneimmonium ion (M+, figure 3) thereby introducing a significant contribution from the reverse reaction (k-1). The effect of chloride concentration (0-0.49 M) on the loss of melphalan was studied at pH 6.0 (u=0.5) and 500C. At this pH the zwitterionic form (MH) of melphalan predominates, and the contributions from the reactions of MH32+, MH2 and M to the overall loss of melphalan may be considered negligible (i.e. [M]=[MH]). Thus, in this section the nomenclature has been simplified by eliminating the species designation (i.e. kl= kl(MH)). The total loss of melphalan at pH 6 may be described by

(22) -dEM + M ]/dt = -d[M]/dt d[M ]/dt where

(23) -d[M]/dt = kl[M] k_ 1 EM]Cl1] and


(24) -d[M ]/dt = k2[M ] k1[MI + k-[M ][cl ] At steady state it may be assumed that


(25) -d[M ]/dt = 0







48

and an expression (equation 26) to describe the concentration of the intermediate ([M I) may be derived

(26) [Me] = k1[M]/(k l [C1 ] + k2 Substituting equations 23, 24, and 26 into equation 22 gives

(27) -d[MI total /dt = k1k2[M]/(k-1 [c] + k2) which may be solved to give an equation (28) relating the pseudo first order rate constant (kobs) to the chloride concentration.

(28) l/kobs = k 1[Cl 1/k1k2 + 1/k1


Figure 14 shows the linear relationship between l/kobs and the chloride concentration (equation 28). The slope (k-1/klk2) and intercept (1/kl) coefficients of this relationship (figure 14) were calculated by linear regression to obtain values of 4.74 h-1 and 10.47 M-1 for k1 and the ratio, k-1/k2, respectively. The ratio, k-1/k2, may be taken as the competition factor describing the relative reactivity of the intermediate (M ) with chloride compared with water. Values of k1 and k-1/k2 were determined at 25

and 370C (equation 37) from the values of kobs in the presence of 0 and 0.4 M chloride. The value of the competition factor (k-1/k2) was found to increase with increasing temperature (table 3).

The influence of pH (3.7-13), ionic strength, buffer composition (acetate, phosphate, borate) and buffer





49


1.4

1.2

1.0

r1 0.8U)
0.6


0.4


0.2

0.0 I
0.0 0.1 0.2 0.3 0.4 0.5

Chloride [M]



Figure 14: Relationship between the reciprocal first order
rate constant for the degradation of melphalan and the concentration of chloride according to
equation 28 (pH 6.0, u=0.5, 500C, r=0.999).







50

TABLE 3

Rate Constants for the Hydrolysis of Melphalan at pH 6.0 (u=0.5)


Rate Constant 250C 370C 500C


k1 (h- ) 0.19 0.92 4.74

k3 (h-1 ) 0.13 0.39 2.38

k-1/k2 (M-1) 6.38 8.01 10.47







51

concentration (15-200 mM) on the rate of degradation of melphalan in the presence of chloride (0.3 M) was

investigated at 500C. In the presence of chloride, the reverse reaction (k-1, figure 3) between chloride and the ethyleneimmonium ion becomes significant and other nucleophiles, such as buffers, present in solution would be expected to compete with chloride for the cyclic intermediate. Thus, k2 (figure 3) must be expanded to take into account other nucleophiles and the hydroxide ion as follows

(29) k2 = k2' + k 2n[OH] + k21 ''[X] where k2', k2'', and k2" '. are the rate constants for the reaction of the ethyleneimmonium ion with water, hydroxide ion (OH-), or another nucleophile (X-) (figure 15).

Additional terms should be added to equation 29 for each additional nucleophile present in the system. Substitution of equation 28 into equation 29 gives, after rearrangement

k /(k k ) =
(30) obs 1 obs
(k2' + k2n[OH ])/k1 [c] + k2 .[X ]/k1 ]


Equation 30 predicts that the rate of hydrolysis of melphalan in the presence of chloride will be accelerated by

increases in pH and the concentration of nucleophilic substances, such as buffer components, in the solution.

Figure 16 shows that the hydrolysis of melphalan at pH

6.55 is independent of ionic strength but strongly dependent






52


2CH2- CH2
R-N
CH2-CH2-C[ k lk






CH2-CH2-CL +H20 k2 X

2 /+


/CH2CH 2OH H-C H2- X
R-N MOH R-N

CH f- CH-Cl CH2-CHf-C


COACH
CH-CH2 /= R NH2







Figure 15: Proposed scheme for the conversion of melphalan
(M) to monohydroxymelphalan (MOH).







53



100

*,O 25 mM MIO 50 mM A,A 100 mM CD



E
Q) 10C



ci)

ci)






1
0 1 2 3

Time [h]



Figure 16: First order plots of % remaining vs. time (500C,
pH 6.55, [Cl1]=0.3 M) showing the effects of
ionic strength (open u=0.6, closed u unadjusted)
and phosphate on the hydrolysis of melphalan
(r>O.998).







54

on the concentration of phosphate buffer in the presence of 0.3 M chloride. The effects of pH and buffer composition on the rate of degradation of melphalan in the presence of 0.3 M chloride were studied at 500C and an ionic strength of 0.6 M. The results of these experiments are listed in table 4 and it can be seen that kobs is significantly influenced by pH and the compositions of buffers used. The data in table 4 were plotted according to equation 30 and the results can be seen in figure 17. In the case of phosphate buffers (pH

5.8, 6.5, 7.7), the slopes of the plots (k2" '/k-1[C1~]) increase with increasing pH indicating that HPO42- competes more strongly than H2PO4j for the ethyleneimmonium ion. The

values of kobs and, hence kobs/(kl-kobs), were much less dependent on the concentration of acetate and borate (figure

17), indicating the weak nucleophilicity of these ions compared with phosphate. The intercepts of these relationships (k2'+k2''[OHf])/(k-1[ClV]) were independent of pH over the range 3.7 to 7.7 but increased with increasing pH above pH 7.7 (table 4 and figure 17).

The degradation of melphalan was also monitored at pH 12

and 13 ([Cl1]=0.3, u=0.6, 500C) to further define the influence of pH. Figure 18 illustrates that the competition

between chloride, water and the hydroxide ion for the ethyleneimmonium ion changes with pH. The function, (k2'+k2" [OH-])/k-1[C1~], remains constant up to pH 7.7 and then increases with increasing pH. Between pH 3.7 and 7.7,







55

TABLE 4

Kinetic Data for the Hydrolysis of Melphalan Showing the Effects of pH and Buffer Composition in the Presence of 0.3 M Chloride (500C)


k1 kobs kobs k2'+k2"[OH] Buffer
Concentration pH (h-1) (h-1) ki-kobs k-1[Cl]
(mM) (a) (b) (c)

Acetate

25 3.70 4.61 1.05 0.295 0.280
100 3.70 4.61 1.06 0.299
200 3.70 4.61 1.06 0.299

25 4.60 4.71 1.11 0.308 0.290
100 4.60 4.71 1.30 0.381
200 4.60 4.71 1.41 0.427

Phosphate

25 5.80 4.71 1.11 0.308 0.273
50 5.80 4.71 1.25 0.361
100 5.80 4.71 1.52 0.476

25 6.55 4.72 1.17 0.331 0.275
50 6.55 4.72 1.64 0.532
100 6.55 4.72 2.10 0.801

25 7.70 4.85 1.73 0.554 0.332
50 7.70 4.85 2.18 0.816
100 7.70 4.85 2.72 1.280

Borate

15 8.66 5.76 1.89 0.488 0.462
25 8.66 5.76 1.90 0.492
50 8.66 5.76 2.02 0.540

15 10.0 9.02 3.20 0.550 0.532
25 10.0 9.02 3.21 0.552
50 10.0 9.02 3.37 0.596

Sodium Hydroxide

10 12.0 9.70 4.41 0.834 0.834
100 13.0 9.70 7.90 4.400 4.400

(a) u=0.6
(b) Mean of two determinations
(c) Intercept of equation 30






56

1.2- pH 4.7
Cn *pH 3.7
-0
0
0.8


v0.4
0

0.0 I
0 50 100 150 200 Acetate [mM]

1.2- pH 7.7 pH 8.7
*pH6.Z6 A pH 10.0
.0 A pH 5.88
0.8



0

0.0 .
0 25 50 75 100 0 10 20 30 40 50
Phosphate [mM] Borate [mM]

Figure 17: Relationship between kobs/(kl-kobs) for melphalan and the buffer concentration plotted
according to equation 32 (500C, u=0.6).






57


10.0






O


1.0
0







0 .1 *
2 4 6 8 10 12 14

pH



Figure 18: Relationship between (k2' +k2 "[OHf])/ k-1[Cl] and pH for melphalan (50C, u=0.6).







58

the only competition which significantly influences the hydrolysis of melphalan (excluding buffer effects) is that between chloride and water for the ethyleneimmonium ion and the value of the competition factor (k-1/k2', equations 29

and 30) is 11.9 M-1. Above pH 11, competition between hydroxide and chloride dominates and the value of the competition factor (k-1/k2'', equations 29 and 30) was calculated to be 0.081. This indicates that hydroxide is a stronger nucleophile than chloride which is consistent with published indices of nucleophilicity (Edwards, 1956; Belluco et al., 1965).



The Production of Chloride

Another approach for investigating the degradation of melphalan was to monitor the production of chloride arising

from the hydrolysis of melphalan. These studies were conducted using an ion specific electrode at 25, 37, and 500C and pH 6.0 (u=0.5). Figure 19 displays a calibration

curve of voltage (E) versus the log of the chloride concentration at pH 6.0 and 370C. With no added chloride and a low initial concentration of melphalan ([M]o=3.28x10-4

M) the contributions from the reverse reactions may be considered negligible and the rate equation for the production of chloride is given by






59




400



360



320
Q)


0 280slope=-61.69 ( 0.43)
int.=159.81 (-1.06)
240 r 2=0.9997
S =0.43
xy
2 0 0 - I . I .
-4 -3 -2 -1
10 10 10 10

Chloride [M]

Figure 19: Calibration curve for voltage (E) vs. log
chloride concentration (M) at pH 6.0 and 370C
(u=0.5).







60

The solutions of equation 31 may be derived as follows. Since the loss of melphalan is first order, it may be written

-k t
(32) [M] = [M] e 1


The rate equation for the production of monohydroxymelphalan (MOH) is

(33) d[MOH]/dt = k2 [M ] k3 [MOH] and the steady state concentration of the ethyleneimmonium ion (M ) in the absence of added chloride is given by

(34) [M+] = k 1[M ek 1 t/k2


Substitution of equation 34 into 33 gives an expression which may be integrated to give equation 35.

-k t -k t
(35) [MOHI = k [M0(e 1 e 3 )/(k3 k1 Substituting equation 35 into 31 gives an expression which may be integrated to give equation 36.

(36) [Cl1]/[M]0 = 2 + ((k1 2k 3)ek 1t + k ek 3 t)/(k k )

Equation 36 is consistent with two unimolecular reactions in which the 2-chloroethylamino functional groups of melphalan and monohydroxymelphalan are transformed to their

respective ethyleneimmonium ions with the loss of two chloride ions. Figure 20 shows that the ratio of chloride to the initial melphalan concentration approaches a value of 2 with time, consistent with equation 36. Using previously determined values of k the data shown in figure 20 were





61


2.0 -E



1.6 500 C



0 1.2- 37 C



0.8



0.4 1



0.0
0 1 2 3 4 5 6 7

Time [h]




Figure 20: Relationship between the ratio of chloride
concentration to the initial melphalan
concentration and time (u=0.5). Lines were
simulated using equation 36 and the constants in
table 3 (r>0.997).







62

fitted to equation 36 by nonlinear regression to obtain the values of k3 at 25, 37, and 500C. The values of the various rate constants obtained by studying the role of chloride in the hydrolysis of melphalan are summarized in table 3.



Influence of Temperature

The influence of temperature on the degradation of melphalan was studied at three temperatures (25, 37, and 500C), four pHs (2.0, 6.0, 9.0, and 11.0) and a constant ionic strength of 0.5 (figure 21). The values of kobs were

fitted to the Arrhenius equation (equation 37) by linear regression to obtain the activation energies (Ea) at the chosen pH (table 5).

(37) log kobs = log A E a/2.303RT The pH values were chosen so that the individual activation energies for the reactive species could be determined from

the appropriate simultaneous equations. However, it was found that the activation energies were independent of pH over the range of 2 to 11 as indicated by the parallel Arrhenius plots (figure 21 and table 5). The consistency of the activation energy indicates that the nature of the rate determining step does not change with changing pH (Jencks, 1969). Additionally, the microscopic rate constants can simply be estimated at any temperature from equation 38 using a mean activation energy of 24 kcal/mole (table 5).

(3 ) og Ek(2) /k(1) E T(2) (1)1/ 0 R (1) T(2)
(38) log [ka /k1 ] = Ea[T T ]/2.303RT T





63


10.0
pH 1 1.0 pH 9.0 0
pH 6.0 A pH 2.0 0

r1

5 1.0

0






0.1
3.05 3.15 3.25 3.35

1 /T [K-1 x 1000



Figure 21: Arrhenius plots of kob vs reciprocal absolute
temperature for the hydrolysis of melphalan.







64

TABLE 5

Activation Energies for Melphalan at Various pHs (u=0.5)


pH Ea (kcal/mole)


2.0 24.96
6.0 24.39
9.0 23.84
11.0 22.81







65

Influence of Ionic Strength

The influence of ionic strength (u=0.12 to 1.0, adjusted

with sodium nitrate) on the degradation of melphalan was studied at 370C and pH 6.0. The data (figure 22) was fit by

linear regression to equation 39 where k is the pseudo first order rate constant at u=0 and a is the slope.
1
(39) log k ob a + log k0


Although the equation has been derived for dilute solutions (Martin et al., 1983), it did prove useful as an empirical

method of describing the effect of ionic strength on the degradation of melphalan. Overall, the effect of ionic strength was small (a=0.15) consistent with the rate determining step being unimolecular (figure 3).



The Stability of Melphalan in Infusion Media


The stability of melphalan in various pharmaceutical diluents was measured since the drug is often administered

by i.v. infusion. Table 6 displays the half life of melphalan at 250C in 0.9% sodium chloride, 5% dextrose, and 0.9% sodium chloride-5% dextrose (1:1). Due to the presence

of chloride, the half life of melphalan is significantly longer in normal saline (18.1 hours) than in dextrose (5.4 hours). For that reason, 0.9% sodium chloride would appear to be the best diluent for the i.v. infusion of melphalan.







66




10.0













C ~1.0


-O
0









0 .1 I I
0.0 0.2 0.4 0.6 0.8 1.0








Figure 22: Relationship between kob and ionic strength (u)
for the hydrolysis of me phalan at pH 6.0 and
370C.







67

TABLE 6

The Half Life of Melphalan in Infusion Media (250C)




Medium Half Life [h]


0.9% sodium chloride 18.1

5% dextrose 5.4

0.9% sodium chloride- 12.5
5% dextrose (1:1)







68

In Vitro Pharmacokinetics of Melphalan


The rate of degradation of melphalan in culture medium was evaluated at 370C in the presence and absence of neuroblastoma cells and at three initial melphalan concentrations. Figure 23 illustrates the ratio of the measured concentration to the initial concentration of melphalan as a function of time. The data were tested using a two way analysis of variance and it was found that there

was no significant influence of the initial concentration (P>0.05) or of the presence of cells (P>O.05) on the rate constant for the degradation of melphalan. In addition, extracellular concentrations of melphalan were not significantly different from concentrations measured in the absence of cells.

The appearance and disappearance of monohydroxymelphalan as well as the appearance of dihydroxymelphalan in culture medium were also monitored by HPLC. It was found that the concentration-time profiles were best fit to equations 40-42 in which the rate constant for the formation of monohydroxymelphalan is equal to the rate constant for its disappearance (k').

(40) [M] = EM] ek't
0
(41) [MOHI = f'[M] k'tekt

(42) [M(OH) 2 1 n[M[0 e-k't (k't + 1)]






69



1.00

o
c- 0.500

E0.20
C)


S 0.100

o0 10 M
0.05 -4 M
CO:3& A 10 MI k
o -5
E O N 10 M
0.02


0.01- I I I
0 1 2 3 4 5 6

Time [h]


Figure 23: Concentration dependency for melphalan in culture medium at 370C and three initial concentrations and in the presence of 106
neuroblastoma cells/ml (open) (r>O.993).







70

In these equations, f' and f" represent the fraction of melphalan (M) converted to monohydroxymelphalan (MOH) and dihydroxymelphalan (M(OH)2), respectively. In media

containing additional nucleophiles (e.g. culture media), these factors have values less than one (f'=0.54, f"=0.43) due to the production of various side products. Figure 24 displays the concentration-time profiles for melphalan and its hydrolysis products in culture medium at 370C. The curves were simulated using equations 40-42. The half life for melphalan measured in vitro (1.12 hours) compares well with the terminal half life measured in vivo (0.6-1.7 hours, table 1).



Iododeoxyuridine Assay


In order to evaluate the response of neuroblastoma cells

following in vitro treatment with melphalan, a short term assay based upon incorporation of 125IUdR into cellular DNA, was investigated. Figure 25 represents a calibration curve

for neuroblastoma cells in which the log of 125IUdR incorporated (cpm) is plotted against the log of the viable cell concentration (cells/ml). The slope of the calibration

curve illustrated is 1.06 with the standard error of the slope being 0.03. Calibration curves were routinely found to be linear in the range of 1x104 to 2x106 cells/ml. These

plots were used to determine the concentration of viable cells remaining after drug treatment. Calibration curves






71




100j




0





o 10-4-j


2~ QM
o A MOH
E M (O H)2



0 1 2 3 4 5 6

Time [h]


Figure 24: Concentration versus time profiles for melphalan
(M), monohydroxymelphalan (MOH), and
dihydroxymelphalan (M(OH)2) in culture medium according to equations 40-42 (370C, r>0.993).





72



105

slope=1.06 (+ .01)
+
int.=-2.09 ( .07)
I -2
o r =0.9977
a4
10 S =0.03
o0 xy
C)






(9




2
1 0 I,,. I F II

104 10 5 106

Cells/ml



Figure 25: Calibration curve fo 12euroblastoma cells (IMR 32) in which cpm of IUdR incorporated is
plotted against cell concentration. Verticle
bars represent standard deviations for the mean
of three measurements.







73

were included in each experiment to account for variations in detector efficiency. The linearity of the calibration curves indicated that the detector efficiency was constant during the counting of a group of samples and standards.

The variability of the assay was determined by analyzing aliquots of a given neuroblastoma sample at three cellular concentrations. At 1x104 cells/ml, the coefficient of variation was 6.1% (n=8), at 1x105 cells/ml it was 5.3% (n=7), and at 1x106 cells/ml it was 13.2% (n=12). The limit of detection for neuroblastoma cells was determined by the lower end of the calibration curve to be 1x104 cells/ml.

The ability of the assay to detect the effect of

melphalan on neuroblastoma cells can be seen in figure 26.

Cells were treated for four hours with initial melphalan concentrations ranging from 10-7 to 10-3 M and assayed immediately following treatment. A dose response

relationship can be seen as an increase in melphalan concentration results in a decrease in percent survival.

One of the major critisisms of short term

chemosensitivity assays is that they are generally performed immediately following drug treatment and do not account for delayed cell death and/or recovery. It has been suggested

that short term assays should be postponed for 4-5 days following treatment to allow delayed responses to become evident (Weisenthal and Lippman, 1985). Figure 27 displays the results from an experiment in which cells were treated





74




100










10
4-1

CL








-7 -6 -5 -4 -3
10 10 10 10 10

Melphalan Concentration [M]

Figure 26: Initial response of neuroblastoma cells treated
for four hours with melphalan.





75







100.0 -F

50.0 [







C) 5.0OVV
X X






1.0X
0 .-4---)





01 2 34 5

Assay Delay Timne [da y s]


Figure 27: Percent- urvival of neu oblastoma cells treated
with 10" (filled), 10- (cross-hatch), or 10(open) M melphalan for two hours and assayed
immediately and up to 5 days following
treatment.







76

with 10-3, 10-5, or 10-7 M melphalan for two hours and assayed immediately and up to 5 days following treatment. For the 10-5 M treatment, there is a dramatic delayed response as the percent cell survival drops from 80 to 3% during the four days following treatment. There does not appear to be a significant difference between responses measured on day 3 and day 4.

In comparing the 125IUdR assay with a clonogenic assay,

differences in cell survival determined using the two methods should be expected since the two assays measure very

different responses. While the 125IUdR assay performed immediately following treatment measures only the immediate

response of the cells to the drug, a clonogenic assay measures the proliferative capacity of the cells following drug treatment. However, by delaying the 125IUdR assay for a given time following drug treatment, it should be possible to detect delayed cell responses. Figure 28 compares the results of the present study obtained using the 125IUdR assay with unpublished results obtained using a clonogenic

assay (Worthington-White, 1986). In the present study, neuroblastoma cells were treated with melphalan for one hour

and assayed (a) immediately and (b) three days following treatment. In the study of Worthington-White, the same neuroblastoma cell line was exposed to melphalan for one hour and assayed using the clonogenic assay of Hamburger and Salmon (1977). It is obvious that by delaying the 125IUdR






77


100

A c

OA




CT)
- 10
CI)


a 9 O



bO



10-7 10-6 10-5 10-4 10

Melphalan Concentration [M]


Figure 28: Cell survival vs. melphalan concentrati U5for
neuroblastoma cells obtained using the IUdR
performed (a) immediately, and (b) 3 days
following treatment and (c) a clonogenic assay
(Worthington-White, 1986).







78

assay, it is possible to detect delayed cell responses (figure 28, a and b). However, differences in cell survival

determined using the 125IUdR and clonogenic assays are increased when the 125IUdR assay is delayed (figure 28, b and c).

There are a number of possible explanations for the differences seen with the two assays. First, even though the 125IUdR assay has been delayed for three days following

treatment, the method still does not evaluate long term proliferative capacity as does the clonogenic assay. Secondly, the two assays measure different cell populations

which may very well have different sensitivities to the drug. Incorporation of 125IUdR should occur for cells which are synthesizing DNA. In contrast, clonogenic cells usually represent a very small fraction of the total cell population. For the neuroblastoma cell line used in these studies, 5x105 untreated cells produced only 70 colonies after incubation for 10 days (Worthington-White, 1986). The

extreme differences in the response of cells treated with the same drug prompt questions as to the significance of results obtained with these types of methods. However, it is not necessarily a question of which is right and which is wrong, but a question of what each of the two assays reveals about the effect of the drug. This question, unfortunately, has not been adequately addressed in the literature but is

vital to the interpretation of results based on these methods.







79

In Vitro Pharmacodynamics of Melphalan


Although nitrogen mustards are capable of reacting with many cellular constituents, it appears that cytotoxicity is primarily the result of the interaction between the drug and DNA (Pratt and Ruddon, 1979; Conners, 1983). Geiduschek (1961) demonstrated that DNA treated with nitrogen mustard

is converted to a reversibly denaturable form suggesting that the drug forms cross-links between the two strands of

the double helix. Strong correlations have since been observed between DNA cross-linking and cytotoxicity for cells treated with melphalan (Murnane and Byfield, 1981; Ducore et al., 1982). Unless DNA repair enzymes can correct

the damage prior to the next division, the cross-links inhibit DNA replication and cytotoxicity results as the cell attempts to divide (Calabresi and Parks, 1980; Hemminki and Ludlum, 1984). This mechanism of action implies that there is a time delay between alkylation and cytotoxicity.

The time delay between drug-cell interaction and

pharmacologic effect (i.e. cell death) is demonstrated in figure 29. Cells treated with 10-5 M melphalan for time periods up to four hours display a dramatic decrease in percent survival during the four days following the initial drug exposure. For example, a two hour exposure predicts a survival of 80% when the pharmacodynamic evaluation is made immediately following treatment (day 0) in comparison to a

survival of 4% when the evaluation is made three days






80




1 00 p


50 -A

20 A-A
.> 20


4-j 10C

0 5(1.




2



0 60 120 180 240

Exposure Time [min]

Figure 29: Percent survival vs. exposure time f r neuroblastoma cells treated with 10 M
melphalan and assayed on 0 (0 ), 1 (A), 2 (E),
3 (y), and 4 (*) days following treatment.







81

following treatment. These results stress the importance of delaying pharmacodynamic measurements following treatment in order to avoid underestimating the effect of the drug.

One of the first parameters which must be evaluated in a pharmacodynamic study is the relationship between drug dose and response. Jusko (1971) presented a mathematical model

for the evaluation of a pharmacodynamic response as a function of drug dose for phase-nonspecific chemotherapeutic agents. This model was later expanded by Gibaldi (1982) and

predicts that the logarithm of the fraction of surviving cells (Sf) is linearly and inversely related to dose (D) according to the following equation




In this equation, ks and kr are the rate constants for natural mitotic growth and normal cell degradation,

respectively, and KL is a function of the affinity of the target cell for the drug, the elimination rate constant for

the drug, and the rate constants responsible for the appearance and disappearance of the drug at the effective site of action. This model is based on the assumption that the evaluation of Sf is made after all of the drug is eliminated.

Dose-response data for cells treated for 20 minutes with initial melphalan concentrations ranging from 10-7 to 10-5 M and assayed three days following drug exposure can be seen in figure 30. Over this concentration range, there does not






82




100





0
O







e
10


50


0.1 1.0 10.0

Dose [Mx 106


Figure 30: Percent survival vs. melphalan dose for
neuroblastoma cells treated for 20 minutes and
assayed 3 days following exposure.







83

appear to be a linear relationship between the log cell survival and dose. The data, however, indicate an extremely steep dose-response relationship as the survival dropped to 20% with a change in dose from 10-7 to 2x10-7 M melphalan. Frei (1979) reported that in vivo data in the literature suggest that the dose response curve for phase-nonspecific agents, such as melphalan, is steep. In some cases, a two fold difference in dose has been shown to cause a substantial difference in in vivo response.

In addition to dose, another parameter which is expected

to influence the pharmacodynamic response is the time for which the cells are exposed to the drug. In traditional in vitro studies, the concentration of a chemically stable drug remains constant during the experiment. However, as previously pointed out, this is not the case for melphalan as the compound is chemically unstable and the concentration changes with time. In addition, the chemical instability of melphalan is involved in the effect of the drug as the rate determining step for hydrolysis and alkylation is believed to be the same. Therefore, another objective of the present

study was to evaluate cell survival as a function of melphalan exposure time. Figure 31 displays the percent survival as a function of exposure time for cells treated

with 2x10-7 M melphalan and assayed three days following treatment. The largest decrease in cell survival was seen within the first 20 minutes followed by a gradual decrease







84





100






0J


L.
Cn
10 C
a










0 60 120 180 240

Exposure Time [min]



Figure 31: Percent survival vs. exposure time for
neuroblastoma cells treated with 2x10-7 M
melphalan and assayed 3 days following
treatment.







85

in survival. There does not appear to be a significant difference in the response after a two hour exposure in comparison to that after a four hour exposure.

These two parameters, the applied dose and the duration of exposure, will simultaneously affect the pharmacodynamic response. Graphical representations of these two factors are seen in figures 32 and 33. In both figures, it is evident that for all doses investigated above 10-7 M, a 20 minute exposure resulted in the highest cell survival while the lowest survival was observed after the longest exposure

time (2-4 hours). In all cases, there was a gradual decrease in survival after 20 minutes but the change in survival between two and four hours is not significant.

It is necessary to consider the possible activity of the

hydrolysis products of melphalan since the compound is rapidly hydrolyzed both in vitro and in vivo. It is

unlikely that the dihydroxylated analogues of alkylating agents have antitumor activity since these compounds are stable and do not form reactive intermediates. This is supported by the experimental data in figure 33 as there appears to be no significant decrease in cell survival between two and four hours when the concentration of dihydroxymelphalan is increased. Monofunctional alkylators

may react with cellular constituents, however, there is no indication that these compounds result in significant cytotoxicity in comparison to bifunctional species. Connors







86




100







CO


10 O


CL










0.1 1.0 10.0


Dose [Mx 10]



Figure 32: Percent survival vs. melphalan dose for
neuroblastoma cells treated for 20 (0), 40 (g),
60 (n), 90 (A), 120 ([]), and 240 (U) minutes and
assayed 3 days following treatment.







87



100k











c- 10ci)

n
a : OI








0 60 120 180 240

Exposure Time [min]



Figure 33: Percent survival vs. exposure time f~r
neuro lastoma cel s treated with 10 (
2x1 ? (0), xo1- (A), 10- (n), 5x1o- (U) and
10- ([]) M melphalan and assayed 3 days later.




Full Text

PAGE 1

CHEMICAL STABILITY AND IN VITRO CYTOTOXICITY OF MELPHALAN BY SUSAN ANN STOUT A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1987

PAGE 2

FOR MY PARENTS John and Rosamond Stout for supporting me when I needed it, encouraging me when I was down, and believing in me all the while.

PAGE 3

ACKNOWLEDGEMENTS I would like to extend my sincere appreciation to my major professors, Dr. Christopher Riley and Dr. Hartmut Derendorf, for their exceptional guidance, their constant encouragement, and most of all, their friendship. I would also like to thank the members of my supervisory committee, Dr. Adrian Gee, Dr. John Dorsey, Dr. Stephen Curry and Dr. John Perrin, as well as Dr. John Graham-Pole, for their advice throughout my graduate career. And a special word of thanks goes to Ms. Diana Worthington-White for her patience and assistance with the neuroblastoma studies. And last, but not least, a sincere thank you is extended to Dr. Linda Adair for giving me confidence when I needed it the most. iii

PAGE 4

TABLE OF CONTENTS PAGE ACKNOWLEDGEMENTS LIST OF TABLES LIST OF FIGURES ABSTRACT iii vi vii xi CHAPTER I. INTRODUCTION 1 The Chemotherapy of Cancer 1 The Treatment of Neuroblastoma with Melphalan 2 The Mechanism of Action of Melphalan 4 The Chemistry of Melphalan. 7 The Pharmacokinetics of Melphalan After I.V. Administration. 10 The Relationship Between Pharmacokinetics and Pharmacodynamics 12 The In Vitro Evaluation of Chemosensitivity 14 Objectives. 17 II. EXPERIMENTAL PROCEDURES 19 Chemicals and Reagents 19 Ultraviolet Spectroscopy 21 High Performance Liquid Chromatography 21 Kinetic Experiments 23 Chloride Determinations 24 Data Analysis 24 Cell Cultures 25 Iododeoxyuridine Assay 25 In Vitro Dynamic Experiments 26 III. RESULTS AND DISCUSSION 28 High Performance Liquid Chromatography 28 Chemical Stability of Melphalan 36 Influence of pH 36 Influence of Chloride 47 iv

PAGE 5

The Production of Chloride 58 Influence of Temperature 62 Influence of Ionic Strength 65 The Stability of Melphalan in Infusion Media 65 In Vitro Pharmacokinetics of Melphalan 68 Iododeoxyuridine Assay. 70 In Vitro Pharmacodynamics of Melphalan 79 Relationship Between the In Vitro Pharmacokinetics and Pharmacodynamics for Melphalan 88 IV. CONCLUSIONS 94 LIST OF REFERENCES 101 BIOGRAPHICAL SKETCH 107 V

PAGE 6

LIST OF TABLES TABLE 1. Summary of In Vivo Pharmacokinetic Parameters for Melphalan Following I.V. Administration 2. Microscopic Rate Constants for the Hydrolysis of Melphalan and Kinetically Determined pKa Values at 37c (u=0.5) 3. Rate Constants for the Hydrolysis of Melphalan at pH 6.0 (u=0.5) 4. Kinetic Data for the Hydrolysis of Melphalan Showing the Effects of pH and Buffer Comgosition in the Presence of 0.3 M Chloride ( 50 C) . . 5. Activation Energies for Melphalan at Various pHs (u=0.5) 6. The Half Life of Melphalan in Infusion Media (25C) vi PAGE 11 42 50 55 64 67

PAGE 7

LIST OF FIGURES FIGURE 1 Structure of melphalan showing three ionizable groups. 2. Proposed mechanism for guanine alkylation by melphalan according to Price, 1975. 3. Proposed pathway for the hydrolysis of melphalan. 4. Chromatogram of an aqueous solution of melphalan (peak 3) and its degradation products (peaks 1 and 2). 5. Chromatogram of (a) water containing melphalan peak (1), monohydroxymelphalan (peak 2), and dihydroxymelphalan (peak 3), (b) blank culture medium and (c) culture medium containing melphalan and its hydrolysis products. 6. Calibration curve of peak area versus concentration for melphalan in culture medium obtained using the conditions in figure 5. 7. UV spectra for (a) melphalan ang (b) dihydroxymelphalan (both 10M) in HPLC mobile phase (pH 2.3). 8. Peak area versus time data for melphalan and monohydroxymelphalan in water at 37c. The curve for monohydroxymelphalan was simulated using equation 2. 9. The degradation of melphalan as a function of time at three initial concentrations (pH 7.47, u=0.5, 37c, r >0.999). 10. Representative first order plots for the degradation of melphalan (initial concentration=lOO ug/ml, 37c, u=0.5). Numbers represent pH values (r>0.999). vii PAGE 5 6 8 29 30 32 33 35 37 39

PAGE 8

11. 12. Log kobs-pH profile for melphalan at 37c (u=0.5). The line has been simulated using equation 13 and the constants in table 2 and the symbols represent experimental values. UV absorbance of melphalan (5xlo-5 M) at 260 nm as a function of pH according to equation 21 (r=0.999). 13. UV absorbance of dihydroxymelphalan (10-4 M) at 260 nm as a function of pH according to equation 21 (r=0.997). 14. Relationship between the reciprocal first order rate constant for the degradation of melphalan and the concentration of chloride according to equation 28 (pH 6.0, u=0.5, 50c, r=0.999). 15. Proposed scheme for the conversion of melphalan (M) to monohydroxymelphalan (MOH). 16. First order plots of% remaining vs. time (50c, pH 6.55, [Cl-]=0. 3 M) showing the effects of ionic strength (open u=0.6, closed u unadjusted) and phosphate on the hydrolysis of melphalan (r>0.998). 17. 18. 19. 20. 21. 22. Relationship between melphalan and the plotted according u=0.6). kob~/(k1-kobs) for bufter concentration to equation 30 (50c, Relationship between (k2'+k6"[0H-])/ k_1[cl-] and pH for melphalan (50 C, u=0.6). Calibration curve for voltage (E) vs. log chloride concentration (M) at pH 6.0 and 37c (u=0.5). Relationship between the ratio of chloride concentration to the initial melphalan concentration and time (u=0.5). Lines were simulated using equation 36 and the constants in table 3 (r>0.997). Arrhenius plots of kobs vs reciprocal absolute temperature for the hydrolysis of melphalan. Relationship between kobs and ionic strength (u) for the hydrolysis of melphalan at pH 6.0 and 37c .................. viii 40 45 46 49 52 53 56 57 59 61 63 66

PAGE 9

23. Concentration dependenc6 for melphalan in culture medium at 37 C and three initial concentrations and in the presence of 106 neuroblastoma cells/ml (open) (r>0.993). 24. Concentration versus time profiles for melphalan (M), monohydroxymelphalan (MOH), and dihydroxymelphalan (M(OH) 2 ) in culture medium according to equations 40-42 (37c, r >0.993). 25. Calibration curve for nl~Soblastoma cells (IMR 32) in which cpm of IUdR incorporated is plotted against cell concentration. Verticle bars represent standard deviations for the mean of three measurements. 26. Initial response of neuroblastoma cells treated for four hours with melphalan. 27. Percent sur3ival gt neurob~astoma cells treated with 10, 10; or 10M melphalan for two hours and assayed immediately and up to 5 days following treatment. 28. Cell survival vs. melphalan concentration for ~ 2groblastoma cells obtained using the IUdR performed (a) immediately, and {b) 3 days following treatment and (c) a clonogenic assay (Worthington-White, 1986). 29. Percent survival vs. exposure time for neuroblastoma cells treated with 10-5 M melphalan and assayed on 0, 1, 2, 3, and 4 days following treatment. 30. Percent survival vs. melphalan dose for neuroblastoma cells treated for 20 minutes and assayed 3 days following exposure. 31. Percent survival vs. exposure time for neuroblastoma cells treated with 2xlo-7 M melphalan and assayed 3 days following treatment. 32. Percent survival vs. melphalan dose for n euroblastoma cells treated for 20, 40, 60, 90, 120, and 240 minutes and assayed 3 days following treatment. 33. Percent survival vs. exposure time for neuroblastoma cells treated with 10-7, 2xlo-7 5xlo-7 10-6 5xl0-6 and 10-S M melphalan and assayed 3 days later. ix 69 71 72 74 75 77 80 82 84 86 87

PAGE 10

34. Kill rate (a) and melphalan concentration (b) as a function of exposure time for neuroblastoma cells treated with an ~nitial drug concentration of 2x10Mand assayed 3 days following treatment. 91 35. Kill rate as a function of exposure time for three_~nitial melph~!an concentrati~gs [2x10 M (a), SxlO M (b), and 10 M (c)] and assayed 3 days following treatment. 92 X

PAGE 11

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy CHEMICAL STABILITY AND IN VITRO CYTOTOXICITY OF MELPHALAN By Susan Ann Stout August 1987 Chairman: Hartmut Derendorf Major Department: Pharmaceutics Melphalan is an alkylating agent which has been used for a number of years in the treatment of a variety of cancers. The compound undergoes rapid hydrolysis to form first monohydroxy-and then dihydroxymelphalan. The hydrolysis kinetics of melphalan have been investigated as a function of pH (1-13), chloride concentration (0-0.5 M), temperature (25-S0C), and ionic strength (0.12-1.0) using HPLC Experimental data support the hypothesis that the rate determining reaction is the unimolecular conversion of melphalan to an ethyleneimmonium ion. Melphalan was found to be more stable at low pH due to an increase in the over al 1 positive charge on the molecule and a decreased xi

PAGE 12

tendency to take on an additional positive charge to form the ethyleneimmonium ion. The stability of melphalan was found to increase in the presence of chloride as a result of competition between chloride and water for the intermediate species. The in vitro pharmacokinetics and pharmacodynamics of melphalan were studied using a model system 0 human neuroblastoma cells suspended in culture medium. The degradation of melphalan in culture medium, monitored by HPLC, was found to be first order (half life= 1.2 h) and independent of the initial concentration (10-5-10-3 M) both in the presence and absence of cells. To evaluate the response of cells treated with melphalan, a short term assay based upon the cellular incorporation of [125I]iododeoxyuridine was used. Cell survival was determined by comparing the cell concentration for a treated sample to that for an untreated control. Cell response was evaluated as a function of assay delay time ( 0-4 days), initial melphalan concentration (10-7-10-3 M), and duration of melphalan exposure (0-240 minutes). Cell survival was found to decrease dramatically during the four days following initial drug treatment indicating a delay between drug-cell interaction and apparent pharmacodynamic response. The dose response curve for the delayed effect was found to be very steep. The maximum rate of cell death for all effective doses was seen during the first 20 minutes of drug xii

PAGE 13

exposure. There was no direct correlation between drug concentration and kill rate in the investigated dose range. xiii

PAGE 14

CHAPTER I INTRODUCTION J Cancer is one of the most frustrating and devastating diseases of our time. To the patient, the term cancer has become almost synonymous with death. To the clinician, cancer represents a puzzle which is only beginning to be understood. Treatment options are generally limited to surgery, radiation therapy or chemotherapy, the application of which varies depending upon the type of cancer being treated. Surgical methods are usually limited to cancers which exist as well defined tumors or those localized in a particular organ. Radiation therapy and chemotherapy are often quite effective in killing rapidly dividing neoplastic cells; however, these methods are nonspecific and toxic to normal cells as well. In some cases, the combination of treatment options has led to the realistic expectation of long term survival or cure. The Chemotherapy of Cancer Most anticancer agents act by interfering with one or more phases bf the cell cycle thereby preventing normal cell division and growth. Consequently, cells most susceptible to the toxic effects of anticancer agents are those which 1

PAGE 15

2 are in the process of division. Since, most anticancer agents are not specific for neoplastic cells, they are toxic to rapidly dividing normal cells as well. The problem of toxicity presents one of the major obstacles encountered in the chemotherapy of cancer. Another barrier which frequently arises in the chemotherapy of cancer is the development of drug resistant tumor cells. A number of approaches have been investigated over the years to combat resistant cells. One of the earliest attempts was to use combination chemotherapy but it was soon recognized that many cells rapidly develop a cross resistance to a variety of drugs. Another approach has been the use of very high doses of chemotherapy which appears to overcome some of the mechanisms of resistance. This approach, however, is limited by toxicity to normal cells, most notably bone marrow. To compensate for the toxicity which is encountered with the use of high dose chemotherapy, this treatment is frequently followed by bone marrow transplantation. Drugs which exhibit myelotoxicity as the major dose limitation are the most reasonable candidates for this type of therapy. The Treatment of Neuroblastoma with Melphalan Neuroblastoma, the third most common childhood malignancy, arises from the sympathetic ganglion cells and frequently metastasizes to the bone, bone marrow, lymph

PAGE 16

3 nodes, skin and liver (Stanfill and Hayes, 1986). It has been reported that survival rates are about 10% at three years for children diagnosed at 1-5 years of age and about 25% for those diagnosed at age 6 or older (D'Angio et al., 1985). Aggressive chemotherapeutic treatment, sometimes in combination with total body irradiation, has been somewhat successful in the treatment .of neuroblastoma but this type of therapy is limited by bone marrow suppression. During the past few years, many investigators have focused on the use of high doses of chemotherapy, with or without radiation therapy, supported by autologous or allogeneic bone marrow transplantation for the treatment of a variety of refractory cancers, including neuroblastoma. Early reports by Fernbach et al. (1968) and Evans (1968) indicated that the alkyating agent, melphalan, was ineffective in the treatment of neuroblastoma. However, in 1979, encouraging results were published by McElwain et al. for the use of high dose melphalan followed by autologous bone marrow transplantation for the treatment of neuroblastoma. Unlike the more conventional drugs used to treat neuroblastoma, such as the anthracyclines and vinca alkaloids, melphalan shows no specific organ toxicity. The major dose limitation of melphalan is myelotoxicity (Pritchard et al., 1982). Since 1979, several investigations have been reported illustrating the usefulness of high dose melphalan therapy followed by bone

PAGE 17

4 marrow rescue for the treatment of refractory neuroblastoma (Pritchard et al., 1982; Graham-Pole et al., 1982; Corringham et al., 1983; Lazarus et al., 1983; Strandjord et al., 1983; Graham-Pole et al., 1984; D'Angio et al., 1985). Unfortunately, it is difficult to compare the various clinical studies due to a wide range of drugs being given concurrently with melphalan. as well as di verse dosing regimens. It appears, however, that the incorporation of bone marrow rescue allows for a 3-5 fold increase in drug dosage which may lead to increased tumor cell kill (GrahamPole et al., 1984). At the present time, no standard dosage regimen exists for high dose melphalan therapy in the treatment of neuroblastoma. The Mechanism of Action of Melphalan Melphalan (figure 1) is an alkylating agent which was first synthesized in 1954 by Bergel and Stock. The compound is a nitrogen mustard derivative of L-phenylalanine and is often referred to as L-phenylalanine mustard. Melpha,lan is believed to act by forming covalent linkages between the alkyl side chains of the compound and cellular constituents, primarily DNA. The mechanism of alkylation has been previously proposed (Price, 197 5) and is represented in figure 2. The first step in the alkylation process is the formation of an ethyleneimmonium ion (M+) which may react with a nucleophilic species to form a covalent bond.

PAGE 18

3 1 ~H2 TH-CH2 COOH t 2 1 Figure 1: Structure of melphalan showing three ionizable groups. 5

PAGE 19

.,..CH-CH-er R-N, 2 2 CH-CH-er 2 2 ClGUANINE in DNA 0 H, II N N:.x \\ CH H N~ N/ 2 N 1 R' lMMONIUM ION GUANINE \ Figure 2: Proposed mechanism for guanine alkylation by melphalan (Price, 1975). 6

PAGE 20

7 Bifunctional alkylating agents, such as melphalan, may then form a second ethyleneimmonium ion which subsequently reacts with another nucleophilic species. One of the most biologically significant reactions appears to be the alkylation of the 7-nitrogen atom of guanine residues of DNA ( Calabresi and Parks, 1980) This reaction has several important consequences. First, the alkylation results in the formation of the enol tautomer whereas the keto tautomer is favored under normal circumstances. The keto tautomer of guanine forms a base pair with a cytosine residue, whereas, the enol tautomer forms a base pair with a thymine residue leading to miscoding. Secondly, alkylation of the 7-ni trogen of guanine can result in splitting of the imidazole ring. And finally, residues may lead to both alkylation of two guanine inter-and intra-strand crosslinking which is believed to be the most serious cytotoxic effect of alkylating agents (Pratt and Ruddon, 1979). The Chemistry of Melphalan Melphalan is a very unstable compound which form_s. monohydroxymelphalan (MOH) and dihydroxymelphalan (M(OH) 2 ) in aqueous solution (figure 3). The hydrolysis of melphalan has been studied in both plasma (Furner et al., 1976) and buffer at pH 7.4 (Flora et al., 1979; Chang et al., 1978). This reaction is also believed to be the major route of

PAGE 21

CCOH R= lH-CHOI 2 NH2 8 ,,CH2 CH 2 X R-N CH2CH2Cl !Mxl 1CHi C~2-0H R-N CHi Cr-:2 X jMOH.Xj Figure 3: Proposed pathway for the hydrolysis of melphalan.

PAGE 22

biotransformation of the drug (Alberts et al., 1979). It has been reported that the stability of melphalan is improved by decreasing the pH, decreasing the temperature, and by the addition of chloride (Flora et al., 1979; Chang et al. 1979) Detailed information however, on these effects and the mechanism of hydrolysis of melphalan is limited. 9 The degradation of melphalan ( figure 3) has been resonably assumed (Chang et al., 1978; Flora et al., 1979) to proceed via the mechanism proposed earlier (Golumbic et al., 1946; Fruton and Bergmann, 1946; Bartlett et al., 1947) for the transformation of methyl-di(2-chloroethyl)amine in water. The rate limiting step for the transformation of 2-chloroethylamines is the cyclization reaction to give corresponding ethyleneimmonium ions with the loss of chloride (Bartlett and Swain, 1949). Ethyleneimmonium ions are highly susceptible to substitution by nucleophiles which, in the case of water, results in hydroxylation. The hydroxylated analogues of bifunctional alkylating agents, such as melphalan, may react further to produce the corresponding dihydroxy analogues. Side reactions with nucleophiles (X) other than water may occur giving rise to MX, MXCl, MXOH, or MX2 Figure 3 shows the generalized mechanism of the nucleophilic substitution of melphalan in aqueous solution.

PAGE 23

10 The Pharmacokinetics of Melphalan After I.V. Administration With few exceptions, authors have found that concentration-time profiles for melphalan following i. v. administration are best fit by a two compartment body model. Table 1 summarizes pharmacokinetic data in the literature following i.v. administration. The general expression for a two compartment body model in which drug is introduced into the systemic circulation is given by equation 1. ( 1) C p -at = ae -13 t + be In this expression, Cp is the plasma drug concentration, t is the time following administration, a and 13 are the hybrid constants for the distribution and elimination phases, respectively, and a and bare the y-axis intercepts for the extrapolated distribution and elimination portions of the curve, respectively. The distribution phase for melphalan is rapid with a half life ranging from 5 to 13 minutes. The elimination phase half life varies from about 0.6 to 1.7 hours after i.v. injections of 15 to 180 mg/m2 The one exception to this is a study conducted by Ninane et al. (1985) which reported a terminal half life of 3 hours following a 180 mg/m2 dose. The data indicate that melphalan pharmacokinetics is dose independent for i. v. injections up to 180 mg/m2 In addition, after normalizing for body surface area, there appears to be no significant difference between pharmacokinetic parameters measured in children and adults.

PAGE 24

11 TABLE 1 Summary of In Vivo Pharmacokinetic Parameters for Melphalan Following I.V. Administration Dose Age tl/2 ,a t1;2 ,13 Cltot V d,13 [mg/m2 ] Group [min] [ h] [ml/min/m2 ] [l/m2J Reference 15 A 7.7 1. 38 203 24.0 Bosanquet (1982) 23 A 12.6 1. 44 192 24.2 Alberts (1979) 140 6.2 0.89 *** *** Davis (1982) 140 C ** 1. 64 274 26.1 Taha (1983) 140 C 8.8 0.72 257 16.1 Gouyette (1986) 140 A 6.9 0.83 525 37.5 Gouyette (1986) 140 C 7.0 0.61 447 23.5 Ardiet (1986) 140 A 5.9 0.79 372 24.8 Ardiet (1986) 180 C 6.6 3.00 357 85.5 Ninane (1985) 180 C 10.5 0.80 498 35.6 Gouyette (1986) 180 A 5.9 0.68 532 31.3 Gouyette (1986) A=adults C=children ages not reported ** one compartment body model used *** data not available

PAGE 25

The Relationship Between Pharmacokinetics and Pharmacodynamics 12 Holford and Sheiner (1986, p. 189) described the terms pharmacokinetics and pharmacodynamics by saying that "pharmacokinetics is what the body does to the drug; pharmacodynamics is what the drug does to the body." Investigators have recently. recognized the necessity of relating the pharmacokinetics of a drug with an elicited pharmacodynamic response. Gibaldi and Perrier (1982) described two factors which determine the relationship between drug concentration and a given response. First, the drug concentration may be directly or indirectly related to a response and, second, the drug may interact reversibly or irri versibly with a receptor. Considerable progress has been made in describing the kinetics of reversible pharmacologic responses (Gibaldi and Perrier, 1982); and as a result, many investigators are conducting clinical studies to correlate pharmacokinetics and pharmacodynamics for drugs which act in this manner. The kinetics of irreversible pharmacologic responses have been studied in theory and in vitro {Gibaldi and Perrier, 1982), however, clinical pharmacokinetic-pharmacodynamic evaluations for drugs in this category (e.g. certain antibiotics and anticancer agents) are limited. In the field of antibiotic therapy, in vitro susceptibility testing of microorganisms has been quite successful as a basis for

PAGE 26

13 rational antibiotic selection and in the determination of minimum inhibitory concentrations. Garrett (1978) summarized a series of extensive investigations which focused on the relationship between microbial generation rate and antibiotic concentration in vitro. Only recently has in vitro pharmacodynamic data for antibiotics been integrated with clinical pharmacokinetics to design appropriate dosing regimens (Schentag et al., 1986). In the field of cancer chemotherapy, very few attempts have been made to relate pharmacokinetic and pharmacodynamic measurements for anticancer agents. Chemotherapeutic agents may be grouped into two broad classes. One class includes agents, such as methotrexate, cytosine arabinoside, and vincristine, which act only at specific phases of the cell cycle and are thus referred to as cell-cycle specific. Another class consists of agents, such as melphalan and cyclophosphamide, which exert their pharmacologic action at any phase of the cell cycle and are termed phasenonspecific. In 1971, Jusko presented a pharmacodynamic model, which was later modified by Gibaldi and Perrier (1982), for phase-nonspecific chemotherapeutic agents. Both of these models predict that the logarithm of the fraction of surviving cells is linearly related to drug dosage provided that the pharmacodynamic measurement is made after all of the drug has been eliminated. The models, however, do not take into account the pharmacokinetics of the drug

PAGE 27

14 and assume that events such as absorption, distribution, and elimination are complete before the cytotoxic effects are apparent. Jusko concluded that the cytotoxic effect will only be a function of the total dose, independent of the route of administration. There is limited information in the literature to experimentally validate this hypothesis. Pharmacodynamic investigations of this type are especially suited for in vitro cell culture systems provided that the pharmacokinetics of the drug can be modeled. The In Vitro Evaluation of Chemosensitivity Treatment of cells with cytotoxic agents may result in ( 1) immediate cell death, ( 2) delayed cell death in which the cell may undergo several divisions before its eventual death, (3) cell insult in which the metabolic processes are temporarily halted but the cell maintains indefinitely its ability to reproduce, or (4) no effect at all. It is generally accepted that the permanant loss of proliferative capacity is the most relevant criterion for evaluating cell kill and, therefore, the lethal effect of a chemotherapeutic agent (Roper and Drewinko, 1976). The validity of results obtained from in vitro chemosensi ti vi ty studies relies upon the ability of the assay to accurately assess cell death as well as the type of assay which is used. There are basically two types of in vitro assays which may be used to assess cell kill following

PAGE 28

drug treatment. 15 The first type is a long term assay which measures the proliferative capacity of the cell following drug treatment. The second type is a short term assay which is performed immediately following treatment and, therefore, measures only the immediate response of the cell to the drug. The conclusions drawn from in vitro chemosensitivity studies may be quite different depending upon which type of assay is used to assess viability since the two types of assays measure very different responses. One of the most popular of the long term assays is a clonogenic stem cell assay published in 1977 by Hamburger and Salmon. Stem cells are defined as those cells within a population having unlimited growth potential. In the clonogenic assay, cell death is measured by the inability of the drug treated cells to form colonies as compared to untreated cells. The concept that clonogenic cells in vitro represent stem cells in vivo and that stem cells are the most important target for chemotherapeutic agents has been the basis for the widespread acclaim of clonogenic assays (Weisenthal and Lippman, 1985). However, it has been suggested that, while stem cells are generally the most sensitive, they are not the only relevant target for chemotherapeutic agents (Drewinko et al., 1981; Weisenthal et al., 1984) and it is possible that clonogenic cells in vitro may not reflect stem cells in vivo (Weisenthal and Lippman, 1985). In addition, clonogenic assays are

PAGE 29

16 extremely time consuming requiring 10-14 days of incubation for the production of a sufficient number of colonies. With this type of assay, it is not practical to perform a calibration curve for the determination of the number of viable cells due to the large number of cells required. Furthermore, the assays lack the sensitivity necessary to evaluate thoroughly the toxic effects of drugs and the results are often highly variable. Many investigators have chosen to use short term assays, such as dye exclusion, 51cr release, or precursor incorporation, to quantitate cell viability in vitro due to the many practical problems associated with clonogenic assays (Roper and Drewinko, 1976; Weisenthal et al., 1984; Weisenthal et al., 1983; Bhuyan et al. 1976; Rupniak et al., 1983; Roper and Drewinko, 1979; Sondak et al., 1984). Recently, an assay was published by Gee et al. (1985) which is based upon the cellular incorporation of a radiolabelled DNA precursor [125r]iododeoxyuridine (125rudR). The method was used to detect cell death following in vitro treatment of leukemia cells with antibody and complement and was shown to be superior in terms of speed, sensi ti vi ty, and reproducibility to both a clonogenic assay and a 51cr release assay. The technique also allows for the inclusion of a calibration curve in each experiment to accurately assess the number of viable cells remaining following treatment.

PAGE 30

17 When using short term assays, such as the 125IUdR assay, to measure in vitro chemosensi ti vi ty, there are several important points which must be considered. First, short term assays measure responses of .the total cell population in contrast to clonogenic assays which measure only the small, clonogenic fraction. Secondly, short term assays performed immediately following drug exposure only measure the immediate cell response while delayed cell responses (death and/or recovery) are unaccounted for. These points make it difficult to make comparisons of results obtained using the two types of assays. Short term assays may, however, be delayed after drug treatment making it possible and practical to measure both immediate and delayed cell responses. Objectives The objective of this project is to develop a thorough understanding of the chemistry of melphalan as well as its cytotoxic activity in vitro. Melphalan is known to be a very unstable compound, however, detailed information on the mechanism of its hydrolysis is not available in the literature. Therefore, the first part of this project is an in depth evaluation of the hydrolysis kinetics of melphalan as a function of pH, chloride concentration, temperature, and ionic strength.

PAGE 31

18 The pharmacological activity of melphalan can be assessed by examining its lethal effect on cancer cells. The second part of this project is an investigation of the in vitro cytotoxicity of melphalan as a function of exposure time and drug concentration. A model system consisting of human neuroblastoma cells suspended in culture medium will be used to conduct these studies. The chemical hydrolysis of melphalan is believed to be the major route of biotransformation. It is, therefore, an ideal candidate for an in vitro evaluation since the concentration changes with time at a rate which is comparable to that in vivo. By following both the concentration of melphalan as well as its cytotoxic activity as a function of time, it may be possible to characterize a relationship between the in vitro pharmacokinetics and pharmacodynamics. Such a relationship could provide useful information for proposing a dosage regimen which would lead to the optimum therapeutic response.

PAGE 32

CHAPTER II EXPERIMENTAL PROCEDURES Chemicals and Reagents Melphalan was obtained from Sigma Chemical Co. (St. Louis, MO) and was used as received. Dihydroxymelphalan was prepared by heating an aqueous solution of melphalan at 60c for 2 hours (Furner et al., 1976). HPLC grade methanol and sodium dodecyl sulfate and ACS grade buffer constituents were obtained from Fisher Scientific Co. ( Fa i r Lawn NJ ) Sodium hydroxide (1 M) was prepared from Dilut-it Analytical Concentrate supplied by J.T. Baker Chemical Co. (Phillipsburg, NJ). Deionized water was filtered through 0.45 um nylon filters (Rainin Instrument Co., Inc., Woburn, MA) before use. Kinetic experiments conducted in the absence of chloride were performed using Mc I l vaine buffers as described by Elving et al. (1956) in the range of pH 2.2 to 8.0. Below pH 2.2, nitric acid was used. Between pH 8.8 and 10.2, boric acid-sodium hydroxide buffers were used as described by Perrin and Dempsey (1974) and above pH 10.2, sodium hydroxide was used. For these studies, the ionic strength (u) was adjusted to O. 5 M with sodium nitrate unless otherwise stated. 19

PAGE 33

20 To evaluate the effects of pH and buffer constituents on the hydrolysis of melphalan in the presence of chloride, acetic acid-sodium acetate buffers (0.025-0.2 M) were used between pH 3.7 and 4.7. For solutions between pH 5.8 and 7.7, potassium phosphate monobasic-sodium phosphate dibasic buffers (0.025-0.1 M) were used. Boric acid-sodium hydroxide buffers (0.015-0.05 M) were used in the pH range of 8.66 to 10.0 and sodium hydroxide was used above pH 10.0. The ionic strength was adjusted with sodium nitrate. In all experiments, the pH of each solution was measured at the temperature of the experiment with a Fisher MicroProbe Combination Electrode and a digital pH millivolt meter model 611 (Orion Research, Cambridge, MA). Culture medium (RPMI 1640), fetal bovine serum (FBS), trypsin containing ethylenedi aminetetraacetic acid ( EDTA) and Hankes balanced salt solution were obtained from GIBCO (Grand Island, NY). [ 125r] Iododeoxyuridine ( 125IUdR) was obtained from Amersham Corporation (Arlington Hts, was di luted to an activity of 10 uCi/ml in RPMI IL) and 1640 containing 20% FBS and 0.2% gelatin (Sigma Chemical Co.) immediately prior to use. Bovine serum albumin (BSA, Sigma Chemical Co.) was adjusted to a concentration of 0.5% w / v in normal saline and stored at 4c. Trichloroacetic acid (TCA, Sigma Chemical Co. ) was prepared as a 20% w / v aqueous solution.

PAGE 34

21 Ultraviolet Spectroscopy Ultraviolet absorption data were obtained for freshly prepared melphalan (10-4 M) and dihydroxymelphalan (10-4 M) in HPLC mobile phase consisting of 40% methanol and 60% buffer (pH 2.0) containing 0.015 M sodium dodecyl sulfate. Spectra were recorded using a Cary 219 Spectrophotometer (Varian, Palo Alto, CA). The absorption of freshly prepared melphalan and dihydroxymelphalan was further investigated ih the pH range of -0.3 to 4.54 and 1.2 to 7.6, respectively, in which the pH was adjusted using either hydrochloric acid or sodium hydroxide. High Performance Liquid Chromatography The high performance liquid chromatography (HPLC) system consisted of a Constametric IIG pump (LDC, Riviera Beach, FL), a Negretti model 190 injector (20 ul sample loop) (HPLC Technology, Palos Verdes Estates, CA) and a Spectromonitor D variable wavelength detector (LDC) operated at 260 nm. Some studies were conducted using a Micrometrics 728 autosampler (Micrometrics, Norcross, GA) with a Valeo RC6U injection valve fitted with a 20 ul sample loop (Valeo Instruments, Houston, TX) in which the sample vials were filled immediately prior to injection. Peak areas were measured using either a Hewlett Packard 3392 Integrator (Hewlett Packard Co., Avondale, PA) or a data system consisting of an

PAGE 35

22 HP87 microcomputer (Hewlett Packard Co., Palo Alto, CA),a Nelson Analytical Interface with software model 366, version 2.2 (Cupertino, CA), and a Hewlett Packard dual 3 1/2 inch disc drive (model 9121). The chromatograms were displayed on a Recordall 5000. strip chart recorder (Fisher Scientific Co.). The loss of melphalan from aqueous solutions was monitored by reversed phase HPLC using an ODS Hypersi 1 column (HETP, Sutton, UK) (150 mm x 4.6 mm i.d.) prepared as described by Bristow et al. (1977), and a mobile phase of 50% methanol and 50% acetate buffer (pH 4.7, 0.1 M). A flow rate of 1. 8 ml/min and ambient temperature were used throughout. Aqueous solutions of melphalan were injected directly onto the HPLC column. To monitor the change in the concentration of melphalan, monohydroxymelphalan and dihydroxymelphalan in culture medium, an HPLC system consisting of a cyanopropylsilane column (Dupont Instruments, Wilmington, DE) (150 mm x 4.6 mm i.d.), and a mobile phase of 40% methanol and 60% buffer (pH 2.0 adjusted with phosphoric acid and containing 0.015 M sodium dodecyl sulfate) were used. The mobile phase flow rate was 1. 0 ml/min and ambient temperatures were used throughout. Culture medium containing melphalan was incubated at 37c and aliquots were withdrawn at appropriate intervals and diluted 1 to 10 with mobile phase immediately prior to injection.

PAGE 36

23 Kinetic Experiments The influence of pH ( 1-13, u=O. 5, 3 7c), temperature (25-5o0c, pH 2, 6, 9, 11, u=0.5) and ionic strength (0.15-1.0, pH 6, 37c) on the degradation of melphalan was investigated. Melphalan (100 ug/ml) was dissolved in the appropriate buffer with the aid of sonification and placed in a thermostated ( 0. 1 c) water bath. The solution was allowed to equilibrate to temperature for 10 minutes before the first sample (100 ul) was taken and injected (20 ul) directly onto the HPLC column. The first injection was taken as t=O since in all cases the loss of melphalan was first order and independent of its initial concentration. Subsequent samples were injected at appropriate intervals and the peak areas found for melphalan were recorded as a percentage of that found at t=O. Peak areas of freshly prepared standards were found to be linearly related to the concentration injected (0.1-200 ug/ml). Single determinations were made for each set of samples and sampling was continued for at least 4 half lives. Kinetic experiments were also carried out in the presence of chloride ( 0. 3 M) to determine the influence of pH (3.7-10), buffer composition (acetate, phosphate or borate) and buffer concentration ( 15-200 mM) on the rate of hydrolysis of melphalan at 50c. Duplicate solutions were prepared and analyzed for each set of conditions and sampling was continued for at least 4 half lives.

PAGE 37

24 Chloride Determinations A combination chloride electrode model 96-l 7B and a digital pH /milivolt meter model 611 (both Orion Research, Cambridge, MA) were used to monitor the production of chloride due to the degradation of melphalan (figure 3). A Haake Circulator (Saddle Brook, NJ) was used for temperature control (~ 0.1c). Standard solutions were prepared in the pre-determined linear range of 4.4 x 10-4 to 10-l M chloride in phosphate buffer (pH 6 .0). Standard curves of voltage (E) vs. log chloride concentration were constructed at each temperature under study. Melphalan (50 mg) was dissolved in a mixture of ethanol (1.5 ml) and nitric acid (20 ul). This solution was immediately transferred to a 100 ml beaker containing 48.5 ml of buffer (pH 6.0, u=0.5), equilibrated to temperature. Voltage readings were taken as a function of time (up to 6 hours) and compared with the calibration curve to obtain the chloride concentration. Readings were taken after 48 hours and the initial melphalan concentration was calculated as the chloride concentration at this time divided by 2 (figure 3). Data Analysis Two nonlinear least squares regression programs were used in these studies to obtain estimates of the various parameters based upon the experimental data ( SASNLIN,

PAGE 38

25 version 83.4, SAS Institute, Inc., Cary, NC, and PCNONLIN, 1985, C.M. Metzler and D.L. Weiner, Statistical Consultants, Inc., Lexington, KY). Cell Cultures A human neuroblastoma cell line {IMR 32) was used in these experiments and was obtained from the American Type Culture Collection (Rockville, MD). The cells were maintained as monolayers at 37c in a humidified atmosphere with 5% co2 in tissue culture flasks (150 cm 2 ) containing RPMI 1640 supplemented with 20% FBS. Before use, the cells were removed from the flasks using 10 ml trypsin (5 g/1 + 0.2 g/1 EDTA in balanced salt solution) and washed three times in culture medium. trypan blue exclusion. Viabilities were determined using Iododeoxyuridine Assay Following drug treatment and washing, each cell pellet was exposed to 100 ul of 125IUdR (10 uCi/ml) for 2 hours at 37C. At this time, the cells were washed three times in culture medium to remove residual, unincorporated 125IUdR and the proteins were precipitated with 100 ul BSA (added as a co-precipitate) and 2 ml TCA. The precipitate was allowed to form overnight at 4c at which time the tubes were centrifuged for 10 minutes at 2000 rpm. The supernatant was

PAGE 39

26 decanted and each cell pellet was analyzed in an automatic gamma counter (1275 Minigamma Gamma Counter, LKB, Turku, Finland). In Vitro Dynamic Experiments Viable cells were suspended in RPM! 1640 + 20% FBS at the desired concentrations. A calibration curve (lxl04 to 2x106 cells/ml) was included in each experiment to establish the relationship between 125 IUdR incorporation ( counts per minute, cpm) and the number of viable cells. Blanks were also included in each experiment to allow for correction of the gamma counter background noise and residual 125IUdR. The mean cpm for the blank was subtracted from the mean cpm for the standards and samples. Each sample contained lx106 cells/ml and was treated with either culture medium alone or culture medium containing melphalan. The cells were incubated at 37c for the desired time interval, washed three times with culture medium and subjected to the 125rudR assay. Percent cell survival was calculated as the ratio of the cell concentration of the treated sample to the cell concentration of the control multiplied by 100. For the delayed assay studies, control and treated cells were plated in culture flasks (25 cm2 ) containing RPM! 1640 + 20% FBS and maintained at 37c in a humidified atmosphere containing 5% co2 for various lengths of time (up to 4 days) prior to the 125IUdR assay. Control and treated cells were

PAGE 40

27 trypsinized and washed three times in culture medium and the entire contents of each flask (supernatant and trypsinized cells) were subjected to the assay.

PAGE 41

CHAPTER III RESULTS AND DISCUSSION High Performance Liquid Chromatography High performance liquid chromatography (HPLC) was utilized to monitor the loss of melphalan from aqueous solution and culture medium. Figure 4 displays a chromatogram of an aqueous solution containing melphalan (peak 3) and degradation products (peaks 1 and 2). Under the conditions employed, melphalan eluted with a retention time of 3. 6 minutes and was well separated from the degradation products. In order to monitor the changing concentration of the hydrolysis products in addition to that of melphalan, the HPLC conditions had to be modified. Figure 5 displays a chromatogram of (a) water containing melphalan (peak 1), monohydroxymelphalan (peak 2), and dihydroxymelphalan (peak 3), (b) blank culture medium and ( C ) culture medium containing melphalan, monohydroxymelphalan, and dihydroxymelphalan. No interferences were observed between the peaks corresponding to melphalan and monohydroxymelphalan and those resulting from material present in the culture medium. There was, however, a small peak present in culture medium which eluted 28

PAGE 42

Column: ODS Hypersil (5 micron, 15 cm x 4.6 mm, i.d.) Mobile Phase: 50% Methanol+ 50% Acetate Buffer. (pH 4.7, O.lM) Flow Rate: 1.8 ml/min UV Detection: 260 nm Temperature: Ambient (1) (.) C: ro J:l ... 0 (/) J:l 0 3 0 2 -r0.1 0 29 J 2 L I l I I 0 2 4 6 Time (min) Figure 4: Chromatogram of an aqueous solution of melphalan (peak 3) and its degradation products (peaks 1 and 2).

PAGE 43

Column: Cyanoproplysilane (5 micron, 15 cm x 4.6 mm, i. d.) Mobile Phase: 40% Methanol+ 60% Buffer (pH 2.0 + 0.015 M sodium dodecyl sulfate) Flow Rate: 1.0 ml/min UV Detection: 260 nm Temperature: Ambient 0 10 20 min b 0 0 2 3 1 0 20 m1 n 2 3 10 20 min Figure 5: Chromatogram of (a) water containing melphalan peak (1), monohydroxymelphalan (peak 2), and dihydroxymelphalan (peak 3), (b) blank culture medium and (c) culture medium containing melphalan and its hydrolysis products. 30 a C

PAGE 44

31 just prior to the peak corresponding to dihydroxymelphalan making it difficult to measure low concentrations of the compound. Figure 6 represents a calibration curve of peak area versus concentration for melphalan in culture medium obtained using the conditions listed in figure 5. In order to calculate the concentrations of monohydroxymelphalan and dihydroxymelphalan using the melphalan calibration curve, the absorptivities of these compounds had to be ascertained in HPLC mobile phase (pH 2.3). Figure 7 displays the UV spectra for melphalan and dihydroxymelphalan in HPLC mobile phase (pH 2.3). The molar absorptivity at 260 nm (in HPLC mobile phase, pH 2.3) was determined to be 18,800 for melphalan and 4200 for dihydroxymelphalan. The molar absorptivity of monohydroxymelphalan could not be evaluated by UV spectroscopy because a pure sample of the compound was not available. However, it was possible to estimate the molar absorptivity of monohydroxymelphalan using HPLC data. In order to use this approach, it was necessary to assume that the area of the HPLC peak corresponding to monohydroxymelphalan is proportional to the concentration of that compound in the range being investigated ( 10-6 to 10-5 M). It was also necessary to assume that the molar absorptivity of monohydroxymelphalan is different from that of melphalan (in HPLC mobile phase, pH 2. 3). This is a reasonable assumption based on the

PAGE 45

32 50 11 + 9 slope=4.37x10 (-2.2x10 ) 4 + 4 40 int.= 1 .29x10 (-1.1x10) 2 r--i r =0.999 l[) 4 I s =2.1x10 0 30 xy rX L-..J 0 Cl) 20 L <{ .:Y. 0 Cl) o_ 10 0 0 1 2 3 4 5 6 7 8 9 10 Melphalan [M x 1 o6J Figure 6: Calibration curve of peak area versus concentration for melphalan in culture medium obtained using the conditions in figure 5.

PAGE 46

Figure 7: 2.0 1.6 1.2 a a, (J C: ta .Q ... 0 II'> .Q ct 0.8 220 260 300 nm UV spectra for (a) rnelphalan and (b) dihydroxyrnelphalan (both 10-4 M) in HPLC mobile phase (pH 2.3). 33

PAGE 47

34 difference in the absorptivity of dihydroxymelphalan and melphalan at the pH of the mobile phase. Figure 8 displays the HPLC peak area versus time data for melphalan and monohydroxymelphalan in in water at 37c. The data suggest that the rate constants for the formation and degradation of monohydroxymelphalan are equal. The HPLC peak area (PA) for monohydroxymelphalan as a function of time can then be written ( 2 ) where r represents the ratio of the molar absorptivity for monohydroxymelphalan to that for melphalan in mobile phase, k 1 is the rate constant for the degradation of melphalan as well as the formation and degradation of monohydroxymelphalan and the subscripts, MOH and M0 refer to monohydroxymelphalan and melphalan at time 0, respectively. The values of rand k 1 may be estimated by fitting the HPLC peak area versus time data for monohydroxymelphalan to equation 2 using nonlinear regression. The value of r multiplied by the molar absorptivity of melphalan in mobile phase gives an estimate of 14,690 as the molar absorptivity for monohydroxymelphalan in mobile phase. Correction factors of 1.28 and 4.55 were then calculated for monohydroxymelphalan and dihydroxymelphalan, respectively, by taking the ratio of the molar absorptivity of melphalan to that for each degradation product. The correction factor was multiplied by the concentration obtained using the

PAGE 48

35 100 n I 0 .X 0 10 (l) L <( ..Y eM 0 (l) Q_ .A MOH 1 -y 0 1 2 3 4 5 6 Time [h] Figure 8: Peak area versus time data for melphalan and monohydroxymelphalan in water at 37c. The curve for monohydroxymelphalan was simulated using equation 2.

PAGE 49

36 melphalan calibration curve to determine the concentration of each hydrolysis product. Chemical Stability of Melphalan At constant pH, ionic strength and temperature, the overall loss of melphalan was first order over at least four half lives and independent of the initial melphalan concentration (figure 9). These observations were consistent with previous reports on the degradation of melphalan (Flora et al., 1979; Chang et al., 1978, 1979). The pseudo first order rate constants (kobs) were calculated by least square linear regression from the slope of linear plots of the logarithms of the percentage of melphalan remaining against time. Influence of pH Melphalan has three ionizable functional groups (figure 1) and up to four species may exist in aqueous solution such that The appropriate dissociation constants for melphalan are given by equations 4-6, respectively. ( 4 ) ( 5 ) ( 6 ) K = [H+][MH+]/[MH2+J a,l 2 3 K = [H+][MH]/[MH+2 J a,2 K = [H+][M-]/[MH] a,3

PAGE 50

37 100.0 M [M x 1 o5J 0 39.7 ti .. 8.2 l{) 1 .6 0 10.0 X 2 L-...J C 0 -0 ..c 1 .0 0.. Q) 2 0.1 0 60 120 180 240 300 time [min] Figure 9: The degradation of melphalan as a function of time at three initial concentrations (pH 7.47, u=0.5, 37c).

PAGE 51

38 The fractions of each species present in aqueous solution may be given by equations 7-10. ( 7) f(MH 2+) 3 = [H+]3/B ( 8) .J.. f(MH2 ) = [H+]2 K /B a,l ( 9) f(MH) + = [H ]K 1 K 1/B a, a ( 10) f(M ) = K K K /B a,l a,2 a,3 where ( 11) The influence of pH on the degradation of melphalan was investigated over the range of 0.91 to 13.0 at 37c and an ionic strength of 0.5 (figure 10). There was no evidence of any general or specific acid/base catalysis. Instead the data were consistent with a unimolecular reaction with M-M+ (figure 3) being the rate determining step. The form of the log kobs-pH profile ( figure 11) was consistent with the different rates at which the four ionic species of melphalan 2+ + (MH3 MH2 MH, M) are converted to their corresponding ethyleneimmonium ions. The respective micro rate constants It follows ( 12) k obs that kobs is given by 2+ + = krH3 f(MH~+) + krH2f(MH;) + krHf(MH) + k~ f(M-)

PAGE 52

39 100 0) 0.91 C C -0 E (l) 10 1.56 0::: ...+-J C (l) u L (l) o_ 4.86 1. 9 6 8.75 13.0 1 0 1 2 3 4 5 6 Time [h] Figure 10: Representative first order plots for the degradation of melphalan (initial concentration=lOO ug/ml, 37c, u=0.5). Numbers represent pH values (r>0.999).

PAGE 53

.-, I ...c 1 .0 L.......J U) _a 0 ..Y. 0.1 Figure 11: 0 40 2 4 6 8 10 12 14 pH Log k bs-pH profile for melphalan at 37c (u=O.g). The line has been simulated using equation 15 and the constants in table 2 and the symbols represent experimental values.

PAGE 54

41 in which f represents the fraction of each species present. Substitution of equations 7-10 into equation 12 gives [H+J3 + [H+]2 K [ +]K 1 + H lK 2 + K lK 2K 3 a, a, a, a, a, a The kinetic data were fit to equation 13 using nonlinear least squares regression to obtain the values for the rate constants and the dissociation constants. An initial fitting of the data to equation 13 gave a value of 10-9 for k 1(MH3 2+) indicating that the contribution of the reaction to equation 13 is negligible. This is not surprising since a protonated 2-chloroethylamine is unlikely to form a cyclic ethyleneimmonium ion. The kinetic data (figure 11) were reanalyzed using a value of O for k 1 (MH3 2+) and the estimates for the rate constants and pKa values are given in table 2. The pKa values of 2.75 and 9.17 (u=0.5, 37C) compare well with the thermodynamic pKas (Merck Index, 1983) of 2.59 and 9.25 for the carboxylic acid and amino groups of phenylalanine, respectively. Thus the pKas of 1.42, 2.59, and 9.25 may be ascribed to positions 1,2, and 3 in figure 1, respectively. The values of k 1 listed in table 2 increase with decreasing protonation MH2+ kl 3 < of the melphalan species such that + kMH2 < 1 <

PAGE 55

42 TABLE 2 Microscopic Rate Constants for the Hydrolysis of Melphalan and Kinetically Determined pKa Values at 37c (u=O.S) Parameter Value + k 1(MH2 ) 0.74 h-1 k 1(MH) 0.98 h-1 k 1(M-) 2.02 h-1 pKa,l 1. 42 pKa,2 2.75 pKa,3 9.17

PAGE 56

43 This order reflects the magnitude of the positive charge on the various species and the tendency of the species to take on an additional positive charge and form the corresponding ethyleneimmonium ion. The tertiary amine pKa values of both melphalan and dihydroxymelphalan were evaluated by spectrophotometry at the wavelength of maximum absorption, 260 nm. This could be accomplished because the unprotonated nitrogen is associated with the chromophore by conjugation of the lone pair electrons with the aromatic system. Protonation of the nitrogen disrupts the conjugation and thereby causes a change in the UV absorption. At a given wavelength, the total absorbance (AT) can be expressed as where Aa and Ab are the absorbances of the acidic and basic species, respectively. The absorbance of any species at a given wavelength is given by (15) A = e:cl where e: is the molar absorptivity at that wavelength, C is the molar concentration of the species, and 1 is the cell path length. Equation 14 can then by rewritten as ( 16) A = e: C l + e: C l T a a b b in which the subscripts a and b represent the acidic and basic species, respectively. is given by ( l 7) C a The total concentration (CT)

PAGE 57

44 and the acid dissociation constant (Ka) can be expressed as ( 18) After rearrangements and substitutions of equations 17 and 18, it can be shown that and Equations 19 and 20 can then be substituted into equation 16 to give equation 21. Absorbance and pH data were fitted to equation 21 using nonlinear regression to obtain estimates of the pKa values. Figures 12 and 13 display the absorbance of melphalan and dihydroxymelphalan, respectively, as a function of pH. A pKa value of 1.4 was obtained for the tertiary amine of melphalan which agrees well with the value obtained from the kinetic experiments (pKa=l.42). The pKa of the tertiary amine of dihydroxymelphalan was estimated to be 3. 9 indicating that the presence of the two chloride atoms in melphalan strongly influences the acidity of the functional group.

PAGE 58

Q) u C 0 _Q L 0 (f) _Q <( 0.8 0.6 0.4 0.2 0.0 4==::::;::::......:;:....--,---y-~,----,---,--~--.-~~___J -1 0 1 2 pH 3 4 5 45 Figure 12: UV absorbance of melphalan (Sxlo-5 M) at 260 nm as a function of pH according to equation 21 (r=0.999).

PAGE 59

46 1 6 1 .4 1.2 Q) u 1 .0 C 0 _Q 0.8 L 0 (/) _Q 0.6 <( 0.4 0.2 0.0 0 1 2 3 4 5 6 7 8 pH Figure 13: UV absorbance of dihydroxymelphalan (lo-4 M) at 260 nm as a function of pH according to equation 21 (r=0.997).

PAGE 60

47 Influence of Chloride Chang et al. (1979) have shown that the degradation of melphalan is inhibited by the addition of sodium chloride. This may be attributed to the competition between the chloride ions and water for the ethyleneimmonium ion (M+, figure 3) thereby introducing a significant contribution from the reverse reaction ( k_ 1) The effect of chloride concentration ( 0-0. 49 M) on the loss of melphalan was studied at pH 6.0 (u=0.5) and so0c. At this pH the zwitterionic form (MH) of melphalan predominates, and the contributions from the reactions of MH3 2+, MH2+, and Mto the overall loss of melphalan may be considered negligible (i.e. [M]=[MH]). Thus, in this section the nomenclature has been simplified by eliminating the species designation (i.e. k 1= k 1 (MH)). described by The total loss of melphalan at pH 6 may be (22) -d[M + M+]/dt = -d[M]/dt d[M+]/dt where and (24) -d[M+]/dt = k 2 [M+] k 1 [M] + k_1[M+][Cl-] At steady state it may be assumed that (25) -d[M+]/dt = 0

PAGE 61

and an expression (equation 26) to describe the concentration of the intermediate ([M+]) may be derived 48 Substituting equations 23, 24, and 26 into equation 22 gives which may be solved to give an equation (28) relating the pseudo first order rate constant ( kobs) to the chloride concentration. Figure 14 shows the linear relationship between 1 /kobs and the chloride concentration (equation 28). The slope (k_1 / k 1 k 2 ) and intercept (l/ k 1 ) coefficients of this relationship (figure 14) were calculated by linear regression to obtain values of 4.74 h-l and 10.47 M-l for k 1 and the ratio, k_1 / k 2 respectively. The ratio, k_1 ; k 2 may be taken as the competition factor describing the relative reactivity of the intermediate (M+) with chloride compared with water. Values of k 1 and k_1 / k 2 were determined at 25 and 37c ( equation 37) from the values of kobs in the presence of O and O. 4 M chloride. The value of the competition factor ( k_ 1/k2 ) was found to increase with increasing temperature (table 3). The influence of pH (3. 7-13), ionic strength, buffer composition (acetate, phosphate, borate) and buffer

PAGE 62

49 1 4 -,-----------------1.2 1.0 r-, 0.8 ....c t..__J (/) ....0 0.6 0 / ..Y. 0.4 0.2 0.0 -t-~~-.----.~-,-------.~-.---,-~-.----.~-J 0.0 0.1 0.2 0.3 0.4 0.5 Chloride [M] Figure 14: Relationship between the reciprocal first order rate constant for the degradation of melphalan and the concentration of chloride according to equation 28 (pH 6.0, u=0.5, so0c, r=0.999).

PAGE 63

50 TABLE 3 Rate Constants for the Hydrolysis of Melphalan at pH 6.0 (u=0.5) Rate Constant 2s0c 37c so0c k1 (h-1) 0.19 0.92 4.74 k3 (h-1) 0.13 0.39 2.38 k_1/k2 (M-1) 6.38 8.01 10.47

PAGE 64

51 concentration ( 15-200 mM) on the rate of degradation of melphalan in the presence of chloride ( 0 3 M) was investigated at so0c. In the presence of chloride, the reverse reaction (k_1 figure 3) between chloride and the ethyleneimmonium ion becomes significant and other nucleophiles, such as buffers, present in solution would be expected to compete with chloride for the cyclic intermediate. Thus, k 2 (figure 3) must be expanded to take into account other nucleophiles and the hydroxide ion as follows where k 2', k 2 ', and k 2''' are the rate constants for the reaction of the ethyleneimmonium ion with water, hydroxide ion (OH-), or another nucleophile (X-) (figure 15). Additional terms should be added to equation 29 for each additional nucleophile present in the system. Substitution of equation 28 into equation 29 gives, after rearrangement ( 30) kb /(kl -kb) = 0 S O S (k2' + k2"[0H-])/k_l[Cl-] + k/' '[X-]/k_l[Cl-] Equation 30 predicts that the rate of hydrolysis of melphalan in the presence of chloride will be accelerated by increases in pH and the concentration of nucleophilic substances, such as buffer components, in the solution. Figure 16 shows that the hydrolysis of melphalan at pH 6.55 is independent of ionic strength but strongly dependent

PAGE 65

/CH2-CH2 Cl R-~ [BJ CH2-CH2-ct k '11 k -1 J~ 1 ,..CH2-CH2-x R-N '\ CH2-CHfCl R 52 Figure 15: Proposed scheme for the conversion of melphalan (M) to monohydroxymelphalan (MOH).

PAGE 66

01 C C -0 E Q) 0: 4-' C Q) u L Q) 0.... 100 10 1 0 1 ,o 25 mM ,o 50 mM A,'6. 100 mM 2 Time [h] 53 3 Figure 16: First order plots of% remaining vs. time (S0c, pH 6.55, [Cl-]=0.3 M) showing the effects of ionic strength (open u=0.6, closed u unadjusted) and phosphate on the hydrolysis of melphalan (r>0.998).

PAGE 67

54 on the concentration of phosphate buffer in the presence of 0.3 M chloride. The effects of pH and buffer composition on the rate of degradation of melphalan in the presence of 0.3 M chloride were studied at so0c and an ionic strength of 0.6 M. The results of these experiments are listed in table 4 and it can be seen that kobs is significantly influenced by pH and the compositions of buffers used. The data in table 4 were plotted according to equation 30 and the results can be seen in figure 17. In the case of phosphate buffers (pH 5.8, 6.5, 7.7), the slopes of the plots (k2"'/k_1[cl-]) increase with increasing pH indicating that HPo 4 2 competes more strongly than H 2Po4 for the ethyleneimmonium ion. The values of kobs and, hence kobs/(k1-k0bs), were much less dependent on the concentration of acetate and borate (figure 17), indicating the weak nucleophilici ty of these ions compared with phosphate. The intercepts of these relationships (k2'+k2 '[OH-] )/(k_1[cl-]) were independent of pH over the range 3.7 to 7.7 but increased with increasing pH above pH 7.7 (table 4 and figure 17). The degradation of melphalan was also monitored at pH 12 and 13 ( [ Cl -l =O. 3, u=O. 6, so0c) to further define the influence of pH. Figure 18 illustrates that the competition between chloride, water and the hydroxide ion for the ethyleneimmonium ion changes with pH. The function, then increases with increasing pH. Between pH 3.7 and 7.7,

PAGE 68

55 TABLE 4 Kinetic Data for the Hydrolysis of Melphalan Showing the Effects of pH and Buffer Composition in the Presence of 0.3 M Chloride (50C) Buffer k1 kobs kobs k2 I + k2 II [ OH] Concentration pH (h-1) (h-1) k1-kobs k_1[cl-] (mM) (a) ( b) ( C) Acetate 25 3.70 4.61 1. 05 0.295 0.280 100 3.70 4.61 1. 06 0.299 200 3 .70 4.61 1. 06 0.299 25 4.60 4.71 1.11 0.308 0.290 100 4.60 4.71 1. 30 0.381 200 4.60 4.71 1. 41 0.427 Phosphate 25 5.80 4.71 1.11 0.308 0.273 50 5.80 4.71 1.25 0.361 100 5.80 4.71 1. 52 0.476 25 6.55 4.72 1.17 0.331 0.275 50 6.55 4.72 1. 64 0.532 100 6.55 4.72 2.10 0.801 25 7.70 4.85 1. 73 0.554 0 .332 50 7 .70 4.85 2.18 0.816 100 7.70 4.85 2.72 1. 280 Borate 15 8.66 5.76 1. 89 0.488 0.462 25 8.66 5.76 1. 90 0.492 50 8.66 5.76 2.02 0.540 15 10.0 9.02 3.20 0.550 0.532 25 10.0 9.02 3 .21 0.552 50 10.0 9.02 3.37 0.596 Sodium Hydroxide 10 12.0 9 .70 4.41 0.834 0.834 100 13.0 9.70 7 .90 4.400 4.400 (a) u=0. 6 ( b) Mean of two determinations ( C) Intercept of equation 30

PAGE 69

Cl) ..0 0 .:::t. I 1 .2 0.8 0.4 1 .2 pH 4.7 ,,,..-.... pH 3.7 U) ....0 0 .:::t. 0.8 I .:::t. '--" "' U) 0.4 ....0 0 .:::t. 0.0-+---~~~~--.----.---f "'f pH 7.7 + pH 6.6 A pH 5.8 0 50 1 00 1 50 200 Acetate [mM] 56 e pH 8.7 A pH 10.0 0 25 50 75 100 0 Phosphate [mM] 10 20 30 40 50 Borate [mM] Figure 17: Relationship between k0b5/(k1-kobs) for melphalan and the buffer concentration plotted according to equation 32 (S0c, u=0.6).

PAGE 70

,,,--.... .-, I u L-...J rI "-" "' ,,,--.... .-, 1.0 I I 0 L-...J N + N "-" 0.1 Figure 18: 2 4 6 8 pH 10 12 14 Relationship between (k2~+k2"~0H-])/ k_1[cl-] and pH for melphalan (50 C, u-0.6). 57

PAGE 71

58 the only competition which significantly influences the hydrolysis of melphalan (excluding buffer effects) is that between chloride and water for the ethyleneimmonium ion and the value of the competition factor (k_1/k2', equations 29 and 30) is 11.9 M-1 Above pH 11, competition between hydroxide and chloride dominates and the value of the competition factor ( k_ 1/k2 equations 29 and 30) was calculated to be 0.081. This indicates that hydroxide is a stronger nucleophile than chloride which is consistent with published indices of nucleophilicity (Edwards, 1956; Belluco et al., 1965). The Production of Chloride Another approach for investigating the degradation of melphalan was to monitor the production of chloride arising from the hydrolysis of melphalan. These studies were conducted using an ion specific electrode at 25, 37, and 5o0c and pH 6.0 (u=0.5). Figure 19 displays a calibration curve of voltage ( E) versus the log of the chloride concentration at pH 6.0 and 37c. With no added chloride and a low initial concentration of melphalan ([M]0=3.28xl0-4 M) the contributions from the reverse reactions may be considered negligible and the rate equation for the production of chloride is given by

PAGE 72

r, w L..J Q) CJ) 0 -+-' 0 > 360 320 280 slope=-61 .69 (+0.43) + int.=159.81 (-1.06) 240 r2=0.9997 S =0.43 xy 59 -4 10 -3 10 -2 10 -1 10 Chloride [M] Figure 19: Calibration curve for voltage (E) vs. log chloride concentration (M) at pH 6.0 and 37c (u=O. 5).

PAGE 73

The solutions of equation 31 may be derived as follows. Since the loss of melphalan is first order, it may be written ( 3 2) [M] 60 The rate equation for the production of monohydroxymelphalan (MOH) is and the steady state concentration of the ethyleneimmonium ion (M+) in the absence of added chloride is given by ( 3 4) Substitution of equation 34 into 33 gives an expression which may be integrated to give equation 35. Substituting equation 35 into 31 gives an expression which may be integrated to give equation 36. ( 3 6 ) Equation 36 is consistent with two unimolecular reactions in which the 2-chloroethylamino functional groups of melphalan and monohydroxyrnelphalan are transformed to their respective ethyleneimmonium ions with the loss of two chloride ions. Figure 20 shows that the ratio of chloride to the initial melphalan concentration approaches a value of 2 with time, consistent with equation 36. Using previously determined values of k 1 the data shown in figure 20 were

PAGE 74

61 1. 6 50 C II 1 .2 37 C 0 2 L__J "' II 0.8 u L__J 25 C 0.4 0 1 2 3 4 5 6 7 Time [h] Figure 20: Relationship between the ratio of chloride concentration to the initial melphalan concentration and time (u=0.5). Lines were simulated using equation 36 and the constants in table 3 (r>0.997).

PAGE 75

62 fitted to equation 36 by nonlinear regression to obtain the values of k 3 at 25, 37, and 50c. The values of the various rate constants obtained by studying the role of chloride in the hydrolysis of melphalan are summarized in table 3. Influence of Temperature The influence of temperature on the degradation of melphalan was studied at three temperatures (25, 37, and 50c), four pHs (2.0, 6.0, 9.0, and 11.0) and a constant ionic strength of 0.5 (figure 21). The values of kobs were fitted to the Arrhenius equation (equation 37) by linear regression to obtain the activation energies (Ea) at the chosen pH (table 5). ( 3 7) log kb = log A -E /2.303RT o s a The pH values were chosen so that the individual activation energies for the reactive species could be determined from the appropriate simultaneous equations. However, it was found that the activation energies were independent of pH over the range of 2 to 11 as indicated by the parallel Arrhenius plots (figure 21 and table 5). The consistency of the activation energy indicates that the nature of the rate determining step does not change with changing pH (Jencks, 1969). Additionally, the microscopic rate constants can simply be estimated at any temperature from equation 38 using a mean activation energy of 24 kcal/mole (table 5). ( 38)

PAGE 76

63 10.0 pH 11 .0 pH 9.0 D pH 6.0 A pH 2.0 0 .-, ..-...c 1.0 L_J (f) ...0 0 ..:x. 0. 1 -t-1--t---1----l---1--i------1r-----l 3.05 3.15 3.25 3.35 1 /T [ K 1 ] x 1 0 0 0 Figure 21: Arrhenius plots of k0bQ vs reciprocal absolute temperature for the hydrolysis of melphalan.

PAGE 77

TABLE 5 Activation Energies for Melphalan at Various pHs (u=0.5) pH 2.0 6.0 9.0 11. 0 Ea (kcal/mole) 24.96 24.39 23.84 22.81 64

PAGE 78

65 Influence of Ionic Strength The influence of ionic strength (u=0.12 to 1.0, adjusted with sodium nitrate) on the degradation of melphalan was studied at 37c and pH 6.0. The data (figure 22) was fit by _linear regression to equation 39 where k0 is the pseudo first order rate constant at u=O and a is the slope. ( 39) log kb = 0 S a2 + log k 0 Although the equation has been derived for dilute solutions (Martin et al., 1983), it did prove useful as an empirical method of describing the effect of ionic strength on the degradation of melphalan. O verall, the effect of ionic strength was small ( a=O .15) consistent with the rate determining step being unimolecular (figure 3). The Stability of Melphalan in Infusion Media The stability of melphalan in various pharmaceutical diluents was measured since the drug is often administered by i. v. infusion. Table 6 di splays the half life of melphalan at 25c in 0.9% sodium chloride, 5% dextrose, and 0.9% sodium chloride-5% dextrose (la). Due to the presence of chloride, the half life of melphalan is significantly longer in normal saline (18.l hours) than in dextrose (5. 4 hours). For that reason, 0.9% sodium chloride would appear to be the best diluent for the i.v. infusion of melphalan.

PAGE 79

_c 1.0 U1 ...0 0 66 0.1 0.0 0.2 0.4 0.6 0.8 1 .0 Figure 22: I Relationship between kobs and ionic strength (u) for the hydrolysis of melphalan at pH 6.0 and 37c.

PAGE 80

TABLE 6 The Half Life of Melphalan in Infusion Media (25c) Medium 0.9% sodium chloride 5% dextrose 0.9% sodium chloride-5% dextrose (1:1) Half Life [h] 18.1 5.4 12.5 67

PAGE 81

68 In Vitro Pharmacokinetics of Melphalan The rate of degradation of melphalan in culture medium was evaluated at 37c in the presence and absence of neuroblastoma cells and at three initial melphalan concentrations. Figure 23 illustrates the ratio of the measured concentration to the initial concentration of melphalan as a function of time. The data were tested using a two way analysis of variance and it was found that there was no significant influence of the initial concentration (P>0.05) or of the presence of cells (P>0.05) on the rate constant for the degradation of melphalan. In addition, extracellular concentrations of melphalan were not significantly different from concentrations measured in the absence of cells. The appearance and disappearance of monohydroxymelphalan as well as the appearance of dihydroxymelphalan in culture medium were also monitored by HPLC. It was found that the concentration-time profiles were best fit to equations 40-42 in which the rate constant for the formation of monohydroxymelphalan is equal to the rate constant for its disappearance (k' ). ( 40) ( 41) ( 42) [M] = [M] e-k't 0 [MOH] = f' [M] k'te-k't 0 [M(OH)2] = f"[M]o[l e-k't(k't + l)]

PAGE 82

1 .00 u C 0.50 0 u 0 +-' C 0.20 "' u C 0 u -0 (1) L :J (J) 0 (1) 2 0.10 0.05 0, 6.' D 11 0.02 0.01 0 1 -3 10 M 10 10 2 -4 M -5 M 3 Time [h] 4 5 Figure 23: Concentration dependency for melphalan in culture medium at 37c and three initial concentrations and in the presence of 106 neuroblastoma cells/ml (open) (r>0.993). 69 6

PAGE 83

70 In these equations, f' and f" represent the fraction of melphalan (M) converted to monohydroxymelphalan (MOH) and dihydroxymelphalan (M(OH) 2), respectively. In media containing additional nucleophiles (e.g. culture media), these factors have values less than one (f'=0.54, f"=0.43) due to the production of various side products. Figure 24 displays the concentration-time profiles for melphalan and its hydrolysis products in culture medium at 37c. The curves were simulated using equations 40-42. The half life for melphalan measured in vitro (1.12 hours) compares well with the terminal half life measured in vivo (0.6-1.7 hours, table 1). Iododeoxyuridine Assay In order to evaluate the response of neuroblastoma cells following in vitro treatment with melphalan, a short term assay based upon incorporation of 125IUdR into cellular DNA, was investigated. Figure 25 represents a calibration curve for neuroblastoma cells in which the log of 125IUdR incorporated (cpm) is plotted against the log of the viable cell concentration (cells/ml). The slope of the calibration curve illustrated is 1.06 with the standard error of the slope being 0.03. Calibration curves were routinely found to be linear in the range of lxl04 to 2xl06 cells/ml. These plots were used to determine the concentration of viable cells remaining after drug treatment. Calibration curves

PAGE 84

71 1 00 --......---------------------, tn1 0 ...--X 2 L--' C 10 0 .. 0 L .. C (l) u M C 0 A MOH u II M(OH) 2 1 0 1 2 3 4 5 6 Time [h] Figure 24: Concentration versus time profiles for melphalan (M), monohydroxymelphalan (MOH), and dihydroxymelphalan (M(OH) 2 ) in culture medium according to equations 40-42 (37c, r >0.993).

PAGE 85

105 -0 (1) ..+-J 0 L 0 104 Q_ L 0 u C fr: -0 :::::J LO 103 N 2 0... u Figure 25: slope= 1.06 int.=-2.09 2 r =0.9977 s =0.03 xy (+ .01) + (-.07) 105 Cells/ml 72 .J.. Calibration curve fo1 geuroblastoma cells (IMR 32) in which cpm of 2 IUdR incorporated is plotted against cell concentration. Verticle bars represent standard deviations for the mean of three measurements.

PAGE 86

73 were included in each experiment to account for variations in detector efficiency. The linearity of the calibration curves indicated that the detector efficiency was constant during the counting of a group of samples and standards. The variability of the assay was determined by analyzing aliquots of a given neuroblastoma sample at three cellular concentrations. At lxl04 cells/ml, the coefficient of variation was 6. 1% ( n=8), at lxl05 cells/ml it was 5. 3% (n=7), and at lxl06 cells/ml it was 13.2% (n=l2). The limit of detection for neuroblastoma cells was determined by the lower end of the calibration curve to be lxl04 cells/ml. The ability of the assay to detect the effect of melphalan on neuroblastoma cells can be seen in figure 26. Cells were treated for four hours with initial melphalan concentrations ranging from 10-7 to 10-3 M and assayed immediately following treatment. A dose response relationship can be seen as an increase in melphalan concentration results in a decrease in percent survival. One of the major critisisms of short term chemosensitivity assays is that they are generally performed immediately following drug treatment and do not account for delayed cell death and/or recovery. It has been suggested that short term assays should be postponed for 4-5 days following treatment to allow delayed responses to become evident (Weisenthal and Lippman, 1985). Figure 27 displays the results from an experiment in which cells were treated

PAGE 87

0 > -> L ::J U) +-' C Q) u L Q) o_ 74 1 00 -----------------, 10 1 -7 1 0 -6 10 -5 10 -4 10 Melphalan Concentration [M] -3 10 Figure 26: Initial response of neuroblastoma cells treated for four hours with rnelphalan.

PAGE 88

0 > > L :=J U) -+-' C Q) u I..._ U) o_ 75 r ------1 00. 0 l I 1 5 0 0 l 10.0 5.0 1 .0 o.s---, -1, I ----I ---l 0 1 2 3 4 5 Ass a y De I a y Ti r n e [days J Figure 27: Percent ~urvival of neu5oblastoma cells treate9 with 10(filled), 10(cross-hatch), or 10(open) M melphalan for two hours and assayed immediately and up to 5 days following treatment.

PAGE 89

76 with 10-3 10-5 or 10-7 M melphalan for two hours and assayed immediately and up to 5 days following treatment. For the 10-5 M treatment, there is a dramatic delayed response as the percent cell survival drops from 80 to 3% during the four days following treatment. There does not appear to be a significant difference between responses measured on day 3 and day 4. In comparing the 125IUdR assay with a clonogenic assay, differences in cell survival determined using the two methods should be expected since the two assays measure very different responses. While the 125IUdR assay performed immediately following treatment measures only the immediate response of the cells to the drug, a clonogenic assay measures the proliferative capacity of the cells following drug treatment. However, by delaying the 125IUdR assay for a given time following drug treatment, it should be possible to detect delayed cell responses. Figure 28 compares the results of the present study obtained using the 125IUdR assay with unpublished results obtained using a clonogenic assay (Worthington-White, 1986). In the present study, neuroblastoma cells were treated with melphalan for one hour and assayed (a) immediately and (b) three days following treatment. In the study of Worthington-White, the same neuroblastoma cell line was exposed to melphalan for one hour and assayed using the clonogenic assay of Hamburger and Salmon (1977). It is obvious that by delaying the 125IUdR

PAGE 90

0 > -> L ::J (/) 4--J C Q) 0 L Q) Q__ 100 10 1 Figure 28: 77 0 a Melphalan Concentration [M] Cell survival vs. melphalan concentrati~~ for neuroblastoma cells obtained using the 5rudR performed (a) immediately, and (b) 3 days following treatment and (c) a clonogenic assay (Worthington-White, 1986).

PAGE 91

78 assay, it is possible to detect delayed cell responses (figure 28, a and b). However, differences in cell survival determined using the 125IUdR and clonogenic assays are increased when the 125IUdR assay is delayed (figure 28, b and c). There are a number of possible explanations for the differences seen with the two assays. First, even though the 125IUdR assay has been delayed for three days following treatment, the method still does not evaluate long term proliferative capacity as does the clonogenic assay. Secondly, the two assays measure different cell populations which may very well have different sensitivities to the drug. Incorporation of 125IUdR should occur for cells which are synthesizing DNA. In contrast, clonogenic cells usually represent a very small fraction of the total cell population. For the neuroblastoma cell line used in these studies, Sxl05 untreated cells produced only 70 colonies after incubation for 10 days (Worthington-White, 1986). The extreme differences in the response of cells treated with the same drug prompt questions as to the significance of results obtained with these types of methods. However, it is not necessarily a question of which is right and which is wrong, but a question of what each of the two assays reveals about the effect of the drug. This question, unfortunately, has not been adequately addressed in the literature but is vital to the interpretation of results based on these methods.

PAGE 92

79 In Vitro Pharmacodynamics of Melphalan Although nitrogen mustards are capable of reacting with many cellular constituents, it appears that cytotoxicity is primarily the result of the interaction between the drug and DNA ( Pratt and Ruddon, 1979; Conners, 1983) Geiduschek (1961) demonstrated that DNA treated with nitrogen mustard is converted to a reversibly denaturable form suggesting that the drug forms cross-links between the two strands of the double helix. Strong correlations have since been observed between DNA cross-linking and cytotoxici ty for cells treated with melphalan (Murnane and Byfield, 1981; Ducore et al., 1982). Unless DNA repair enzymes can correct the damage prior to the next division, the cross-links inhibit DNA replication and cytotoxicity results as the cell attempts to divide (Calabresi and Parks, 1980; Hemminki and Ludlum, 1984) This mechanism of action implies that there is a time delay between alkylation and cytotoxicity. The time delay between drug-cell interaction and pharmacologic effect (i.e. cell death) is demonstrated in figure 29. Cells treated with 10-S M melphalan for time periods up to four hours display a dramatic decrease in percent survival during the four days following the initial drug exposure. For example, a two hour exposure predicts a survival of 80% when the pharmacodynamic evaluation is made immediately following treatment (day 0) in comparison to a survival of 4% when the evaluation is made three days

PAGE 93

0 > > L :J if) +-' C (I.) u L (I.) (L 100 50 20 10 5 2 1 80 0 60 120 180 240 Exposure Tirne [min] Figure 29: Percent survival vs. exposure time gr neuroblastoma cells treated with 10M melphalan and assayed on O ( e ) 1 ( ), 2 ( ) 3 ( 9), and 4 (.) days following treatment.

PAGE 94

81 following treatment. These results stress the importance of delaying pharmacodynamic measurements following treatment in order to avoid underestimating the effect of the drug. One of the first parameters which must be evaluated in a pharmacodynamic study is the relationship between drug dose and response. Jusko (1971) presented a mathematical model for the evaluation of a pharmacodynamic response as a function of drug dose for phase-nonspecific chemotherapeutic agents. This model was later expanded by Gibaldi (1982) and predicts that the logarithm of the fraction of surviving cells {Sf) is linearly and inversely related to dose (D) according to the following equation In this equation, ks and kr are the rate constants for natural mitotic growth and normal cell degradation, respectively, and KL is a function of the affinity of the target cell for the drug, the elimination rate constant for the drug, and the rate constants responsible for the appearance and disappearance of the drug at the effective site of action. This model is based on the assumption that the evaluation of Sf is made after all of the drug is eliminated. Dose-response data for cells treated for 20 minutes with initial melphalan concentrations ranging from 10-7 to 10-5 M and assayed three days following drug exposure can be seen in figure 30. Over this concentration range, there does not

PAGE 95

0 > -> L ::J (/) -+-' C (1) u L (1) 0.... 10 ~ -----5 -t-----.-----r--,----y----r-T""-r-..--r----r--..----.----.---.~~ 82 0. 1 1.0 10.0 6 Dose [M x 10 J Figure 30: Percent survival vs. melphalan dose for neuroblastoma cells treated for 20 minutes and assayed 3 days following exposure.

PAGE 96

83 appear to be a linear relationship between the log cell survival and dose. The data, however, indicate an extremely steep dose-response relationship as the survival dropped to 20% with a change in dose from 10-7 to 2xlo-7 M melphalan. Frei (1979) reported that in vivo data in the literature suggest that the dose response curve for phase-nonspecific agents, such as melphalan, is steep. In some cases, a two fold difference in dose has been shown to cause a substantial difference in in vivo response. In addition to dose, another parameter which is expected to influence the pharmacodynamic response is the time for which the cells are exposed to the drug. In traditional in vitro studies, the concentration of a chemically stable drug remains constant during the experiment. However, as previously pointed out, this is not the case for melphalan as the compound is chemically unstable and the concentration changes with time. In addition, the chemical instability of melphalan is involved in the effect of the drug as the rate determining step for hydrolysis and alkylation is believed to be the same. Therefore, another objective of the present study was to evaluate cell survival as a function of melphalan exposure time. Figure 31 displays the percent survival as a function of exposure time for cells treated with 2xlo-7 M melphalan and assayed three days following treatment. The largest decrease in cell survival was seen within the first 20 minutes followed by a gradual decrease

PAGE 97

100 --------------------, 0 > -> L :J U) -+-J 10 C Q) u L Q) (L 1 0 Figure 31: \ ~ ~ ----60 120 180 Exposure Time [min] Percent survival vs. exposure time for neuroblastoma cells treated with 2xlo-7 M melphalan and assayed 3 days following treatment. 240 84

PAGE 98

in survival. There does not appear to be a significant difference in the response after a two hour exposure in comparison to that after a four hour exposure. 85 These two parameters, the applied dose and the duration of exposure, will simul~aneously affect the pharmacodynamic response. Graphical representations of these two factors are seen in figures 32 and 33. In both figures, it is evident that for all doses investigated abov e 10-7 M, a 20 minute exposure resulted in the highest cell survival while the lowest survival was observed after the longest exposure time (2-4 hours). In all cases, there was a gradual decrease in survival after 20 minutes but the change in survival between two and four hours is not significant. It is necessary to consider the possible activity of the hydrolysis products of melphalan since the compound is rapidly hydrolyzed both in vitro and in vivo. It is unlikely that the dihydroxylated analogues of alkylating agents have antitumor activity since these compounds are stable and do not form reactive intermediates. This is supported by the experimental data in figure 33 as there appears to be no significant decrease in cell survival between two and four hours when the concentration of dihydroxymelphalan is increased. Monofunctional alkylators may react with cellular constituents, however, there is no indication that these compounds result in significant cytotoxicity in comparison to bifunctional species. Connors

PAGE 99

0 > -> I.... ::J U) ..+.J 100 C 10 Q) u I.... Q) 0... 1 0.1 Dose 1.0 6 [M x 1 0 ] 86 10.0 Figure 32: Percent survival vs. melphalan dose for neuroblastoma cells treated for 20 (Q), 40 (e), 60 (6), 90 (A), 120 (0), and 240 (a) minutes and assayed 3 days following treatment.

PAGE 100

0 > -> L ::J (f) -+-' C Q) u L Q) o_ 87 C --====-100 i l 10 1 0 60 120 180 240 Exposure Time [min] Figure 33: Percent survival vs. exposure time f9r neur~91astoma ce!7s treate9 with 10-{f), 2xlg (Q), SxlO (6), 10 6 (6), SxlO (a) and 10(0) M melphalan and assayed 3 days later.

PAGE 101

88 ( 1979) reported that the mono functional derivative of melphalan has no antitumor activity and that two alkylating moieties are necessary for cytotoxic action. In a similar manner, bifunctional sulfur mustard has been shown to be as much as 100 times more active than the mono functional compound (Pratt and Ruddon, 1979). As previously stated, the primary mechanism of action of most alkylating agents is believed to be DNA cross-linking which supports the hypothesis that bifunctionality is necessary for cytotoxicity. It has been suggested that, while monofunctional alkylators do not significantly contribute to cytotoxicity, they do result in carcinogenic effects (Connors, 1979). Relationship Between the In Vitro Pharmacokinetics and Pharmacodynamics for Melphalan The importance of determining whether or not there is a relationship between the pharmacokinetics of a drug and the pharmacodynamic response has been recognized within the past few years. The establishment of such a relatidnship can lead to a more thorough understanding of the pharmacology of the drug as well as provide a rational foundation for the design of a dosing regimen. In searching for a pharmacokinetic-pharmacodynamic relationship, it is first necessary to identify a response which can be measured and then to determine how that response changes with drug

PAGE 102

89 concentration. Since it is often difficult to quantitate pharmacodynamic responses in vivo, in vitro systems can provide a valuable alternative for conducting such experiments. One potential problem, however, frequently encountered when using in vitro systems is that the drug concentration remains constant during the course of the experiment. This presents an artificial situation which does not adequately reflect in vivo pharmacokinetic events such as metabolism and elimination. In the present study, the pharmacokinetics of melphalan are represented by the in vitro degradation of the drug. Cytotoxicity, as measured by the decreased ability of drug treated cells to incorporate a radiolabeled DNA precursor in comparison to untreated control cells, represents the pharmacodynamic response. Melphalan is an interesting compound for such an investigation since the chemical hydrolysis is believed to be the major route of biotransformation. The relationship between the pharmacokinetics and pharmacodynamics of melphalan is complicated by the fact that the drug-cell interaction does not result in an immediate response but rather one which only becomes apparent after a certain delay. In the previous dose-response figures (32 and 33), drug concentration was related to the cumulative cytotoxicity. To evaluate the drug effect at single time points, the data were transformed and the rate of cell kill (i.e. the change

PAGE 103

90 in percent survival, b%, divided by the time interval, bt) was calculated. If there exists a direct relationship between drug concentration and effect, plots of drug concentration and kill rate versus time should be similar. However, as can be seen in figure 34, the time course of kill rate (a) and drug concentration (b) are not parallel. Whereas the melphalan concentration decreases in a first order manner, the maximum kill rate is observed during the first 20 minute interval after which the kill rate rapidly drops to zero. The lack of a direct correlation between kill rate and concentration is even more obvious in figure 35 where the kill rate as a function of time is displayed for (a) 2xlo-7 (b) Sxlo-6 and (c) 10-S M melphalan doses. For a concentration range which varies SO-fold, plots of kill rate versus time are nearly identical. There are many possible explanations for this observation. First, the dose-response relationship may be so steep that is is difficult to evaluate the effect of drug concentration. Secondly, the relationship between drug-cell interaction and cytotoxicity may not be proportional. And finally, the cell population being investigated is not homogeneous and at any time consists of cells in different phases of the cycle. While melphalan is believed to be phase-nonspecific in that alkylation may occur during any phase of the cycle, it has been shown that the sensitivity to alkylating agents varies depending upon which phase the

PAGE 104

5.0 -al 91 4.0 3.0 4-J
PAGE 105

92 5.0 5.0 A b a 4.0 4.0 3.0 3.0 2.0 2.0 \ 1.0 1.0 \ ......... 0.0 -0.0 A-A-A -1.0 0 30 60 Figure 35: 90 -+J
PAGE 106

cells are in during exposure (Meyn and Murray, 1984). 93 It is therefore likely that the most sensitive cells are rapidly killed which might account for the early, maximum kill rate which was observed.

PAGE 107

CHAPTER IV CONCLUSIONS Melphalan is an anticancer agent which exerts its primary pharmacological effect by alkylating and crosslinking strands of DNA. The compound has been used for a number of years in the treatment of multiple myeloma and cancers of the breast, ovary, and testis. No major organ toxicity has been associated with melphalan therapy and the primary dose limitation appears to be myelosuppression. During the past 10 years, high dose melphalan therapy followed by bone marrow rescue has been investigated for the treatment of various refractory cancers including neuroblastoma, Ewing' s sarcoma, and osteosarcoma. The rationale for high dose melphalan therapy stems primarily from in vitro studies demonstrating that higher initial concentrations of the drug lead to increased tumor cell kill. Melphalan is known to be a very unstable compound both in vitro and in vivo, however, studies have not been conducted to relate the change in drug concentration with the effect. Since 1954 when melphalan was first synthesized, studies concerning the chemical stability of the compound have been reported in the literature. Rigorous investigations, however, relating the effects of chloride and pH to the 94

PAGE 108

95 mechanism of hydrolysis have not been previously conducted. Consequently, the first part of this project was a thorough examination of the hydrolysis kinetics of melphalan. The hydrolysis of melphalan results in the formation of monohydroxymelphalan and its subsequent conversion to dihydroxymelphalan. In the proposed mechanism, the rate determining step in the hydrolysis is the formation of a cyclic ethyleneimmonium ion with loss of a chloride ion. The cyclic intermediate is rapidly attacked by nucleophiles which, in the case of water, yields monohydroxymelphalan. An increased level of chloride leads to competition between chloride and water for the intermediate and a significant stabilization of the parent molecule. The half life of melphalan is. dramatically increased in media containing chloride. Nucleophilic species, such as buffer constituents and hydroxide ion, may compete with chloride for the intermediate thereby reducing the stability of melphalan, however, the half life can never be shorter than that in the absence of chloride. The reverse reaction between chloride and the intermediate is negligible in comparison to the reaction of the intermediate with water in the absence of added chloride. Under these circumstances, the degradation of melphalan is unaffected by nuc leophi les, other than water, which are present in the system. This observation supports the hypothesis that the formation of the ethyleneimmonium ion, rather than its reaction with water,

PAGE 109

is the rate determining step in the reaction. The hydrolysis of melphalan follows apparent first order kinetics in all of the experiments which have been performed. 96 In the absence of chloride, pH affects the hydrolysis of melphalan only by influencing the ionization of the three functional groups. As demonstrated by the rate-pH profile, the degradation rate constant for melphalan is significantly reduced at pH values less than 3. The increased stability of the molecule at low pH is attributed to the increase in the overall positive charge on the molecule and a decreased tendency to take on an additional positive charge to form the ethyleneimmonium ion. The fully protonated form of melphalan should not degrade since the protonated tertiary amining could not form the cyclic intermediate. From the kinetic experiments, estimates of the pKa values for the tertiary amino, the carboxyl, and the primary amino functional groups of melphalan have been obtained. The investigations conducted on the hydrolysis kinetics of melphalan have practical significance in addition to their theoretical value. First, if the drug is being administered clinically by i. v. infusion, it would be beneficial to prepare melphalan in normal saline rather than another diluent in order to increase its stability. In addition, when collecting samples to be analyzed at a later time, the addition of a small volume of concentrated

PAGE 110

97 hydrochloric acid would stabilize the drug by increasing the chloride content as well as lowering of the pH. And finally, reducing the temperature of both infusion media and samples containing melphalan would help to retard the degradation of the molecule. To evaluate the pharmacodynamic response of cells treated with melphalan, a short term assay based upon the incorporation of a radiolabeled DNA precursor was investigated. The assay was chosen for this study because it is simple and rapid and results are less variable than those obtained using clonogenic assays. In addition, the method allows for the inclusion of a calibration curve in each experiment in order to accurately assess cell viability. There is a dramatic difference in results measured when the assay is performed immediately following drug treatment in comparison to the results observed after a given delay. This emphasizes that for melphalan, there is a significant time delay between drug-cell interaction and pharmacologic response. This aspect should be further investigated in order to utilize the method to examine other drug-cell interactions. It is quite difficult to make comparisons of results obtained with different in vitro chemosensi ti vi ty assays since the the two types of assays ( long and short term) measure different cytotoxic responses. A further complication results as the two types of assays most likely

PAGE 111

98 evaluate different cell populations which may have different sensitivities to cytotoxic agents. While short term assays, such as the one described.in this report, may provide a more quantitative evaluation of cell survival, long term clonogenic assays may reveal qualitative information concerning the proliferative capacity of the cells following drug exposure. The ultimate question is which type of method most accurately reflects the in vivo response of cancer cells to chemotherapeutic agents. There are isolated reports in the literature concerning in vitro/in vivo correlations for the various assays, however, a standard method for evaluating in vitro chemosensitivity has not been established. One objective of the current project was to determine if there is a relationship between the in vitro pharmacokinetics and pharmacodynamics of melphalan. Since the in vitro rate of degradation of melphalan is representative of the in vivo elimination of the drug, the system provides a reasonable model for studying such a relationship. The data indicate a very, steep dose response relationship which makes the evaluation of the effect of drug concentration difficult. Over the dose range evaluated, there did not appear to be a direct, linear relationship between drug concentration and effect. At all effective concentrations evaluated, the maximum rate of cell kill was observed during the first 20 minutes of drug

PAGE 112

99 exposure. An additional complication is that the cell population being investigated is probably not homogeneous in terms of sensitivity to the drug. This aspect could be exploited by conducting studies using synchronized cell populations. Unfortunately, there is limited information available in the literature with which to compare the results of the present study. As previously emphasized, comparisons are complicated by the use of different assays as well as different cell lines. In a study conducted by Hill and Whelan (1981), a linear, inverse relationship was observed between the log cell survival, determined using CHP 100 neuroblastoma cells and a colony forming assay, and the dose of melphalan administered. These results agree with the model presented by Jusko (1971). However, in a study reported by Worthington-White et al. (1986), the response of cells was found to increase even though the cumulative dose of melphalan decreased. In that study, the authors administered the same total dose of melphalan to cultured neuroblastoma cells in four different dosing schedules and evaluated cell survival using a clonogenic assay. At the end of each dosing interval, the investigators removed the residual drug prior to the next application so that there was no accumulation. As a result, the area under the melphalan concentration versus time curve decreased with increasing administration. The authors reported that the

PAGE 113

100 cell survival decreased with increasing administration, however, the cumulative dose of melphalan administered, evaluated using the concentration versus time curve, also decreased. One possible explanation for this observation is that the time above a given concentration, rather than the cumulative dose of drug, is related. to cell survival. In summary, three different investigations on the same type of cancer cell using the same drug reveal three different pharmacokinetic-pharmacodynamic relationships. The differences are most likely the result of different in vitro assays used to evaluate cell responses. In vitro investigations provide a practical means of analyzing the correlation between pharmacokinetics and pharmacodynamics. These methods have many potential applications in the area of cancer chemotherapy where little is known about the relationship of drug concentration and effect. However, these systems are often much more complex than might first appear and it is imperative that the variables influencing in vitro pharmacodynamic responses are thoroughly investigated. In order to rationally interpret the results of such in vitro studies, the methodologies must be further researched with particular emphasis on in vivo correlations. Finally, in order to determine the significance of in vitro chemosensitivity studies, it is ultimately necessary to compare the results with clinical pharmacokinetic-pharmacodynamic data.

PAGE 114

LIST OF REFERENCES Alberts, D.S., Chang, S.Y., Chen, H.S.G., Moon, T.E., Evans, T .L., Furner, R.L., Himmelstein, K. and Gross, J .F. "Kinetics of intravenous melphalan." Clin. Pharmacol. Ther. (1979) 26, 73-80. Ardiet, C., Tranchand, B., Biron, P., Rebattu, P. and Philip, T. "Pharmacokinetics of high-dose melphalan in children and adults with forced diuresis." Cancer Chemother. Pharmacol. (1986) 16, 300-305. Bartlett, P.D., Ross, S.D. and Swain, C.G. "Kinetics and mechanisms of the reactions of tertiary B chloroethylamines in solution. I. Methyl-bischloroethylamine." J. Am. Chem Soc. (1947) 69, 2971-2977. Bartlett, P.D. and Swain, C G. "Kinetics of hydrolysis and displacement reactions of B,B'dichlorodiethyl sulfide (mustard gas) and of B-chlorohydroxydiethyl sulfide (mustard chlorohydrin)." J. Am. Chem. Soc. (1949) 71, 1406-1415. Belluco, U., Cattalini, L., Basolo, F., Pearson, R.G. and Turco, A. "Nucleophilic constants and substrate discrimination. Factors for substitution reactions of platinum (II) complexes." J. Am. Chem. Soc. ( 1965) 87, 241-246. Bergel, F. and Stock, J.A. "Cyto-active amino acid and peptide derivatives. Part I. Substituted phenylalanines." J. Am. Chem. Soc. (1954) 76, 2409-2417. Bhuyan, B.K., Loughman, B.E., Fraser, T.J. and Day, K.J. "Comparison of different methods of determining cell viability after exposure to cytotoxic compounds." Exp. Cell Res. (1976) 97, 275-280. Bosanquet, A.G. and Gilby, E.D. "Pharmacokinetics of oral and intravenous melphalan during routine treatment of multiple myeloma." Eur. J. Cancer Clin. Oncol. ( 1982) 18, 4, 355-362. 101

PAGE 115

102 Bristow, P.A., Brittain, P.N., Riley, C.M., and Williamson, B.F. "Upward slurry packing of liquid chromatography columns." J. Chrornatogr. (1977) 13, 57-64. Calabresi, P. and Parks, R.E. "Antiproliferative agents and drugs used for irnrnunosuppresion." in The Pharmacological Basis of Therapeutics, 6th ed. New York: Macmillan Publishing Co., Inc., 1980, pp. 1256-1257. Chang, S.Y., Alberts, D.S., Farquhar, D., Melnick, L.R., Walson, P.D. and Salmon, S.E. "Hydrolysis and protein binding of rnelphalan." J. Pharm. Sci. ( 1978) 67, 682-684. Chang, S.Y.; Evans, T.L. and Alberts, D.S. "The stability of rnelphalan in the presence of chloride ion." J. Pharrn. Pharmacol. (1979) 31, 853-854. Conners, T.A. "Alkylating drugs, nitrosourea and dialkyltriazenes." in Cancer chemotherapy, annual 1. New York: Elsevier Science Publishing Co., Inc., 1979, pp. 26-28. Conners, T.A. "Alkylating drugs, nitrosourea and dialkyltriazenes.'' in Cancer chemotherapy, annual 5. New York: Elsevier Science Publishing Co., Inc., 1983, pp. 32-33. Corringham, R., Gilmore, M., Prentice, H.G. and Boesen, E. "High dose rnelphalan with autologous bone marrow transplant." Cancer (1983) 52, 1783-1787. D'Angio, G.J., August, C., Elkins, W., Evans, A.E., Seeger, R., Lenarsky, C., Feig, S., Wells, J., Ramsay, N., Kirn, T., Woods, W., Krivit, W., Strandjord, S., Coccia, P. and Novak, L. "Metastatic neuroblastoma managed by supralethal therapy and bone marrow reconstitution (BMRc). Results of a four-institution children's cancer study group pilot study." in Advances in Neuroblastoma Research. New York: Alan R. Liss, Inc., 1985, pp. 557-563. Davis, T.P., Peng, Y.M., Goodman, G.E. and Alberts, D.S. "HPLC, MS, and pharrnacokinetics of rnelphalan, bisantrene, and 13-ci s retinoic acid." J. Chroma to gr. Sci. ( 1982) 20, 511-516. Drewinko, B., Patachen, M., Yang, L.Y. and Barlogie, B. "Differential killing efficacy of twenty antiturnor drugs on proliferating and nonproliferating human tumor cells." Cancer Res. ( 1981) 41, 2328-2333.

PAGE 116

103 Ducore, J.M., Erickson, L.C., Zwelling, L.A., Laurent, G. and Kohn, K.W. "Comparative studies of DNA cross-linking and cytotoxicity in Burkitt's lymphoma cell lines treated with cis-diamminedichloroplatinum (II) and Lphenylalanine mustard." Cancer Res. ( 1982) 42, 897-902. Edwards, J.O. character." "Polarizability, basicity and nucleophilic J. Am. Chem. Soc. (1956) 78, 1819-1820. Elving, P.J., Markowitz, J.M. and Rosenthal, I. "Preparation of buffer systems of constant ionic strength." Anal. Chem. (1956) 28, 1179-1180. Evans, A.E. "Melphalan therapy for children with metastatic neuroblastoma." Cancer Treat. Rep. (1968) 52, 293-296. Fernbach, D.J., Haddy, T.B., Holcomb, T.M., Stuchy, W.J.Jr., Sullivan, M.P. and Watkins, W.L. "L-sarcolysin (NSC 8806) therapy for children with metastatic neuroblastoma." Cancer Chemother. Rep. (1968) 52, 293-296. Flora, K.P., Smith, S.L. and Cradock, J.C. "Applications of a simple high performance liquid chromatographic method for the determination of melphalan in the presence of its hydrolysis products." J. Chromatogr. (1979) 177, 91-97. Frei, E., III. "Antitumor agents-dose response curve clinical and experimental considerations." Exp. Hematol. (1979) 7, 262-264. Fruton, J.S. and Bergmann, M. "Chemical reactions of the nitrogen mustard gases. III. The transformations of ethyl-bis($-chloroethyl)amine in water." J. Org. Chem. (1946) 11, 543-549. Furner, R.L., Mellett, L.B., Brown, P.K. and Duncan, G. "A method for the measurement of L-phenylalanine mustard in the mouse and dog by high pressure liquid chromatography." Drug Metab. Dispos. (1976) 4, 577-583. Garrett, E.R. "Kinetics of microbial action." Scand. J. Infect. Dis. (1978) 14, 54-85. Gee, A.P., Rolfe, A.E., Worthington-White, D., Graham-Pole, J. and Boyle, M.D.P. "A rapid alternative to the clonogenic assay for measuring antibody and and complement-mediated killing of tumor cells." Clin. Immunol. Immunopathol. (1985) 34, 263-271. Geiduschek, E.P. "Reversible DNA." Proc. Natl. Acad. Sci. (1961) 47, 950-955.

PAGE 117

104 Gibaldi, M. and Perrier, D. "Kinetics of pharmacologic response." in Pharmacokinetics, 2nd. New York: Marcel Dekker, Inc., 1982, pp. 221-269. Golumbic, C., Fruton, J.S. and Bergmann, M. "Chemical reactions of the nitrogen mustard gases. I. The transformations of methyl-bis(~-chloroethylamine) in water." J. Org. Chem. (1946) 11, 518-536. Gouyette, A., Hartmann, 0. and Pico, J.L. "Pharmacokinetics of high-dose melphalan in children and adults." Cancer Chemother. Pharmacol. ( 1986) 16, 184-189. Graham-Pole, J., Gross, S., Herzig, R., Lazarus, H., Weiner, R. and Coccia, P. "High dose melphalan and autologous bone marrow transplantation for resistant neuroblastoma and Ewing's sarcoma." Blood (1982) 60, suppl. 1, 168a. Graham-Pole, J., Lazarus, H.M., Herzig, R.H., Gross, S., Coccia, P., Weiner, R. and Strandjord, S. "High dose melphalan therapy for the treatment of children with refractory neuroblastoma and Ewing's sarcoma." Am. J. Pediatr. Hem. Oncol. (1984) 6, 1, 17-26. Hamburger, A.W. and Salmon, S.E. "Primary bioassay of human tumor stem cells." Science (1977) 197, 461-463. Hemminki, K. and Ludlum, D.B. by antineoplastic agents." (1984) 73, 1021-1028. "Covalent modification of DNA J. Natl. Cancer Insti. Hill, B.T. and Whelan, R.D.S. "Assessments of the sensitivities of cultured human neuroblastoma cells to anti-tumour drugs." Pediatr. Res. (1981) 15, 1117-1122. Holford, N.H.G. and Sheiner, L.B. "Kinetics of pharmacologic response." in Pharmacokinetics: theory and methodology. New York: Pergamon Press, 1986, p. 189. Jusko, W.J. "Pharmacodynamics of chemotherapeutic effects: dose-time-response relationships for phase-nonspecific agents." J. Pharm. Sci. (1971) 60, 892-895. Lazarus, H.M., Herzig, R.H., Graham-Pole, J., Wolf, S.N., Phillips, G.L., Strandjord, S., Hurd, D., Forman, W., Gordon, E.M., Coccia, P., Gross, S. and Herzig, G.P. "Intensive melphalan chemotherapy and cryopreserved autologous bone marrow transplantation for the treatment of refractory cancer." J. Clin. Oncol. ( 1983) 6, 1, 359-367.

PAGE 118

105 McElwain, T.J., Hedley, D.W., Gordon, M.Y., Jarman, M. and Pritchard, J. "High dose melphalan and non-cryopreserved autologous bone marrow treatment of malignant melanoma and neuroblastoma." Exp. Haematol. ( 1979) 7, suppl. 5, 360-371. Martin, A., Swarbrick, J. and Cammarata, A. "Kinetics." in Physical Pharmacy, Physical Chemical Principals in the Pharmaceutical Sciences, 3rd ed. Philadelphia: Lea and Febiger, 1983, p. 376. Meyn, R.E. and Murray, D. "Cell cycle effects of alkylating agents." Pharmac. Ther. (1984) 24, 147-163. Murnane, J.P. and Byfield, J.E. "Irreparable DNA crosslinks and mammalian cell lethality with bifunctional alkylating agents." Chem.-Biol. Interactions (1981) 38, 75-86. Ninane, J., Baurain, R., de Selys, A., Trouet, A. and Cornu, G. "High dose melphalan in children with advanced malignant disease." Cancer Chemother. Pharmacol. (1985) 15, 263-267. Perrin, D.D. Control. and Dempsey, B. Buffers for 2!1 and Metal Ion London: Chapman and Hall, 1947, p. 147. Pratt, W.B. and Ruddon, R.W. "The alkylating agents." in The Anticancer Drugs. New York: Oxford University Press, 1979, pp. 64-74. Price, C.C. "Chemistry of alkylation." in Antineoplastic and Immunosuppressive Agents, Part II. Berlin: SpringerVerlag, 1975, pp. 1-5. Pritchard, J., McElwain, T.J. and Graham-Pole, J. "High dose melphalan with autologous marrow for treatment of advanced neuroblastoma." Br. J Cancer (1982) 45, 86-94. Roper, P.R. and Drewinko, B. "Comparison of in vitro methods to determine drug induced cell lethality." Cancer Res. (1976) 36, 2182-2188. Roper, P.R. and Drewinko, B. "Cell survival following treatment with anti tumor drugs." Cancer Res. ( 1979) 39, 1428-1430. Rupniak, H.T., Dennis, L.Y. and Hill B.T. "An intercomparison of in vitro assays for assessing cytotoxicity after a 24 hour exposure to anti-cancer drugs." Tumori (1983) 69, 37-42.

PAGE 119

Schentag, J.J., DeAngelis, C. and Swanson, D.J. "Dual individualization with antibiotics." in Applied pharmacokinetics, 2nd ed. Spokane, WA: Applied Therapeutics, Inc., 1986, pp. 463-492. 106 Sondak, V.K., Bertelsen, C.A., Tanigawa, N., HildebrandZanki, S.U., Morton, D.L., Korn, E.L. and Kern, D.H. "Clinical correlations with chemosensitivities measured with a rapid thymidine incorporation assay." Cancer Res. (1984) 44, 1725-1728. Stanfill, P. and Hayes, F.A. "Neuroblastoma and related tumors." in Pediatric Oncology and Hematology. St. Louis: The C.V. Mosby Company, 1986, p. 92. Strandjord, S., Gordon, Erlinda, Gordon, Elizabeth, GrahamPole, J Novak, L., Shina, D., Lazarus, H., Herzig, R. and Coccia, P. "High dose melphalan (L-PAM) and fractionated total body irradiation (F-TBI) as preparation for bone marrow transplantation (BMT) in children with recurrent stage IV neuroblastoma (NB)-A preliminary report." Proc. Am. Assoc. Cancer Res. ( 1983) 24, 159. Taha, I.A.-K., Ahmad, R.A., Pritchard, J. and Rogers, H.J. "Pharmacokinetics of melphalan in children following high-dose intravenous injection." Cancer Chemother. Pharmacol. (1983) 10, 212-216. Weisenthal, L.M., Dill, P.L., Kurnick, N.B. and Lippman, M .E. "Comparison of dye exclusion assays with a clonogenic assay in the determination 0 drug-induced cytotoxici ty." Cancer Res. ( 1983) 43, 258-264. Weisenthal, L.M. and Lippman, M.E. "Clonogenic and nonclonogenic in vitro chemosensitivity assays." Cancer Treat. Rep. (1985) 69, 615-631. Weisenthal, L.M., Shoemaker, R.H., Marsden, J.A., Dill, P.L., Baker, J.A. and Moran, E.M. "In vitro chemosensitivity assay based on the concept of total tumor cell kill." Rec. Results Cancer Res. (1984) 94, 161-173. Windholz, M., ed. The Merck Index, 10th ed. Rahway, NJ: Merck & Co., Inc., 1983, p. 7144. Worthington-White, D.A. unpublished results, 1986. Worthington-White, D.A., Graham-Pole, J.R., Stout, S.A. and Riley, C.M. "In yitro studies with melphalan and pediatric neoplastic and normal bone marrow cells." Int. J. Cancer (1986) 37, 819-823.

PAGE 120

BIOGRAPHICAL SKETCH The author was born in 1959 and was raised in Panama City, Florida. She graduated from Bay High School in 1977 and received an Associate of Arts degree in 1979 from Gulf Coast Community College, Panama City, Florida. She then transferred to Florida State University in Tallahassee where she received a Bachelor of Science degree in chemistry in 1981. She was employed as a chemist in the College of Pharmacy at the University of Florida in Gainesville from 1982 until January of 1984 when she entered graduate school in the Department of Pharmaceutics at the University of Florida. She received the degree of Doctor of Philosophy in 1987. 107


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 EFZMQ8IW1_N4YEMR INGEST_TIME 2017-06-05T17:06:21Z PACKAGE AA00055792_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES