The use of microbial iron chelators and polyamine analogues as antineoplastics

MISSING IMAGE

Material Information

Title:
The use of microbial iron chelators and polyamine analogues as antineoplastics
Physical Description:
vi, 158 leaves : ill. ; 29 cm.
Language:
English
Creator:
Ingeno, Michael Joseph, 1953-
Publication Date:

Subjects

Subjects / Keywords:
Iron Chelates -- therapeutic use   ( mesh )
Polyamines -- therapeutic use   ( mesh )
Antineoplastic Agents   ( mesh )
Medicinal Chemistry thesis Ph.D   ( mesh )
Dissertations, Academic -- Medicinal Chemistry -- UF   ( mesh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1988.
Bibliography:
Bibliography: leaves 151-157.
Statement of Responsibility:
by Michael Joseph Ingeno.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001071589
oclc - 20379920
notis - AFF6025
sobekcm - AA00004809_00001
System ID:
AA00004809:00001

Full Text











THE USE OF MICROBIAL IRON CHELATORS AND POLYAMINE
ANALOGUES AS ANTINEOPLASTICS














By

MICHAEL JOSEPH INGENO


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


1988















This work is dedicated to my family, with love.














ACKNOWLEDGMENTS

I would like to express my sincere appreciation to my research

advisor, Dr. Raymond J. Bergeron, for his generous guidance, expert

advice and friendship over the past four years. His enthusiasm for

science has been an inspiration.

I would also like to thank Dr. Richard R. Streiff for his assist-

ance in the completion of the animal studies and Dr. Raul Braylan for

his advice on flow cytometry and cell cycle analysis.

I am especially grateful for the assistance and dedication of

Jack Kramer Austin, Jr., Gabriel Luchetta, Annette Zaytoun, and Dexter

E. Beck whose time and effort have helped make this work possible, and

for the technical assistance of Joe Kayal, Stacia Goldey, B. Wesson

Young, and Rosa Rosado.

I wish to acknowledge the other coworkers, past and present, in

Dr. Bergeron's laboratory who have been helpful in the completion of my

many projects and helped make my learning experience an enjoyable one.

I also wish to acknowledge Irma Smith for her expert advice and

assistance in the preparation of this manuscript.

Finally, I wish to thank my wonderful family: I am eternally

grateful to Mr. and Mrs. James Ingeno for a lifetime of love and

encouragement, they were always there when I needed them; I am blessed

with my daughter, Leah, who could make me smile when all else failed;

and I am constantly indebted to my beautiful wife, Deborah, for her

love, tenderness, patience and support throughout the completion of

this work.















TABLE OF CONTENTS

Page

ACKNOWLEDGMENTS. . . iii

ABSTRACT . . .. . v

CHAPTERS

I. INTRODUCTION . . .. 1

Catecholamide Iron Chelators . 1

Polyamine Analogues. . 11

II. CATECHOLAMIDE CHELATORS. . 21

Materials and Methods. . ... 21

Results. . ... 30

Discussion . . 63

III. POLYAMINE ANALOGUES. . . 76

Materials and Methods. . 76

Results. . ... 87

Discussion . . 127

IV. CONCLUSIONS. . . ... 145

REFERENCES . . .. 151

BIOGRAPHICAL SKETCH. . . .. 158















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

THE USE OF MICROBIAL IRON CHELATORS AND POLYAMINE
ANALOGUES AS ANTINEOPLASTICS

By

Michael Joseph Ingeno

December 1988

Chairman: Raymond J. Bergeron
Major Department: Medicinal Chemistry

This research focuses on the development of compounds which are

effective in controlling the growth of cancer cells. The two classes

of agents studied are the microbial iron catecholamide chelators

(siderophores) and the N-alkyl polyamine analogues.

The biological activity of the catecholamide chelators, parabac-

tin and vibriobactin, has been assigned, at least in part, to their

ability to chelate iron and inhibit the iron dependent enzyme ribonuc-

leotide reductase. These siderophores are potent inhibitors of neo-

plastic cell growth, with 50% inhibitory concentrations in the micro-

molar range. The ligands have pronounced effects on cell cycle kinet-

ics producing a GI-S cell cycle block. On short term exposure to the

chelators, this block in DNA synthesis is reversed simply by washing

the chelator away. The cells remained synchronized for at lease three

cycles after release of the block. The cell cycle block and synchroni-

zation induced by the chelators may be exploited to potentially enhance

cytotoxicity when used in combination with other antineoplastics.

v











The biological activity of the N-alkylated polyamine derivatives

has been postulated to be a consequence of their similarity to the

natural intracellular polyamines, spermidine and spermine in terms of

uptake, while functionally they are unable to fulfill or interfere with

important roles of the natural polyamines. Inhibition of cell growth

may result as a consequence of alterations in the polyamine biosyn-

thetic network. Preliminary studies of the structural boundary condi-

tions required for significant antiproliferative activity of the poly-

amine analogues have been carried out. The most active analogues, the

terminally diethylated tetramines N1,N12-diethylspermine and N1,N14-

diethylhomospermine, have been shown to be potent inhibitors of neo-

plastic cell growth, with 50% inhibitory concentrations in the micro-

molar range in vitro and significant in vivo activity. They have been

found to have only incidental effects on nuclear DNA, while having pro-

found effects on mitochondrial DNA.

Results are presented as to the action of these drugs both in

vitro and in vivo. From these studies a great deal has been determined

about the growth inhibitory effects, mechanisms of actions, broadness

of activity against various cell lines, structure activity relation-

ships, and effectiveness of these compounds either alone or in combina-

tion with other agents.














CHAPTER I
INTRODUCTION

This research focuses on the development of compounds which are

effective in controlling the growth of cancer cells.

Catecholamide Iron Chelators

The first part of the dissertation is concerned with the evalua-

tion of catecholamide iron chelators as a means of controlling the pro-

liferative processes. Over the past decade there has been a dramatic

increase in the attention given to the crucial role which iron plays in

proliferative processes. Although iron has oxidation states between -2

and +6, the equilibrium between the +2 and +3 oxidation states, an

equilibrium which is very sensitive to both ligation and pH (1), best

characterizes the biological use of this metal. This transition metal

plays a critical role in a variety of biological redox systems within

the cell, e.g., the cytochromes, peroxidases, catalases, and ribonucle-

otide reductase (2). Even though virtually all organisms are auxo-

tropic for Fe(III), because of the insolubility of ferric hydroxide

(Fe(OH)3), the predominant form of the metal under physiological con-

ditions (Ksp=10-38M4), nature has had to develop rather sophisticated

mechanisms for accessing the metal (3).

Microorganisms produce low-molecular weight ligands (sidero-

phores) which chelate and facilitate transport of iron into the cell

(2). Most of these siderophores fall into two general structural

classes, hydroxamates and catecholamides (4). The catecholamides,

1











2
typified by enterobactin (Fig. 1-la), generally form tighter iron

complexes (metal-ligand formation constant (Kf=1048-1 ) (5) than the
hydroxamates, exemplified by desferrioxamine (Fig. 1-lb) (Kf=1031M-1)

(6).

Mammalian cells have developed larger, more complicated mole-

cules, iron binding proteins for the metal's transport and utilization:

e.g., the iron shuttle proteins, transferring (7), lactoferrin (8),

uteroferrin (9), and the iron storage protein, ferritin (10). It has

been suggested that the virulence of pathogenic bacteria may be par-

tially related to the efficiency with which its siderophore removes

iron from these iron binding proteins (11).

Transferrin, the major iron transport protein in animals, binds

two iron atoms in the +3 oxidation state, each with a slightly differ-

ent metal-protein formation constant. This binding is very sensitive

to the oxidation state of the metal and to pH with an apparent iron

binding constant at physiological pH of Kf=1028M-1 (12). Once the

transferring is bound to the transferring receptor on the cell's surface,

it is internalized via endocytotic vesicles. The iron-protein-receptor

complex enters a nonlysosomal acidic vacuole where conditions favor

release of the iron for utilization or storage as an iron-ferritin com-

plex. The apotransferrin-receptor complex then returns to the cell

surface where the apotransferrin is released for further use (13).

The number of transferring receptors on a cell membrane has been

directly linked to the proliferative state of the cell and/or the



















NH

CH HI


O

CO

CH
,NH
'CO



OH


O

CH2
I


HCs
NH
CH2~.o CO 'CO



OH
a.


OH OH
OH H 0 0
NN NN N N.. NH2
0 OH H


b.

Figure 1-1. The microbial iron chelators


a. Enterobactin
b. Desferrioxamine











availability of iron (14). For example, it has been observed that

highly proliferative cells have significantly greater numbers of trans-

ferrin receptors than resting cells (15). Alternatively, if iron is

withheld from cells there is a substantial increase in the number of

transferring receptors (16). This increase is due to stimulation of

synthesis and translation of the transferring receptor mRNA, and can be
inhibited by the addition of iron (17). Conversely, if iron is added

to the cells in a utilizable form there is a dramatic decrease in

transferring receptor expression, and an increase in cellular ferritin

iron as a means of defense against bacterial infection and neoplasms.

It is interesting to observe how the body alters iron metabolism

in response to a bacterial infection. In the course of an infection,

the body attempts to withhold iron from the invading organism by

decreasing gastrointestinal absorption of iron, lowering the serum

transferring iron saturation, and increasing liver ferritin levels and

iron ferritin storage in reticuloendothelial systems (18). The same

response is seen in febrile states, neoplasia, and on injection of var-

ious sham substances (19). It has been demonstrated that mice fed a

low iron diet developed smaller and slower growing tumors than mice

maintained on a normal iron diet (20). Conversely, iatrogenic iron

overload as well as idiopathic hemochromotosis, a genetic disease char-

acterized by iron overload, have been associated with increased inci-
dence of neoplasia (21,22). These findings suggest continued study of

the role of iron metabolism in malignancy and the possible adverse

effects of iron supplementation in patients with cancer.

There is evidence to indicate that activated macrophage cytotox-
icity may be mediated by iron removal from the target cells (23). A










5
selective inhibition of iron dependent DNA replication and mitochondri-

al respiration is observed. The authors speculate that a macrophage-
derived iron chelator could be involved, leading to cytostasis. Later

studies (24) found that aconitase, a citric acid cycle enzyme contain-

ing iron-sulfur clusters, activity declined simultaneously with arrest

of DNA synthesis. The cytotoxic activated macrophage-induced inhibi-

tion of aconitase is shown to be due to loss of iron from the iron-

sulfer cluster. More recently (25), a newly discovered cyclic hydrox-
amate siderophore, bisucaberin, was found to render tumor cells sus-

ceptible to cytolysis mediated by macrophages and showed specific

inhibition by addition of ferric iron.

One of iron's more interesting roles in cells is its association
with ribonucleotide reductase (Fig. 1-2), an enzyme which catalyses the

production of deoxyribonucleotides in the rate limiting step of DNA
synthesis (26). In mammalian cells, ribonucleotide reductase consists

of two nonidentical dimer subunits, M1 and M2, analogous to the B1 and

B2 subunits of Escherichia coli (E. coli) (27). The M2 dimer contains
a nonheme iron center. The metal stabilizes a tyrosyl free radical at

the active site and, due to the rapid turnover of the enzyme, a con-
stant supply of iron is essential for the enzyme's activity (28).

Ribonucleotide reductase levels are elevated in neoplastic cells and
have been shown to be transformation-linked and progression-linked in a
series of hepatomas (29).

If the theory that withholding iron as a defense mechanism
against neoplasia is correct, then artificially inducing a state of

iron deprivation by means of the siderophores may prove to be effective
in controlling neoplastic cell growth. Iron chelators could control

















Substrate
Specificity
(ATP, dATP
dTTP, dGTP)


Activity
(ATP, dATP)


Figure 1-2.
Ref. 26).


Bl-subunit







H




B2-subunit


Model of E. coli ribonucleotide reductase (from










7

cellular proliferation by completing extracellular iron (including that

bound to transferring "free" cytoplasmic iron, and/or iron bound to

critical proteins or enzymes.

There are several rationales for the use of the microbial cate-

cholamide iron chelators as a means of controlling neoplastic growth.

1. The formation constants for the catecholamide iron complexes

(Kf=1046-1052 M-1) are far greater than those observed for chelators

previously used as antiproliferatives (4). Considering how effectively

the catecholamide siderophores bind iron, one would expect these che-

lators to remove iron from several enzyme bound sites, e.g., ribonucle-

otide reductase, mitochondrial enzymes, as well as from transferring.

However, in spite of the thermodynamic binding advantage these ligands

have over various proteins, if the ligand cannot access the protein

bound iron, it obviously cannot remove it. Thus, removal of iron is a

kinetic problem.

2. Previously low biological yields from microorganisms and ted-

ious isolation procedures of these compounds limited their use. Chemi-

cal schemes have now been devised for the efficient synthesis of the

catecholamide microbial siderophores as well as a number of synthetic

analogues (30-32). The natural products are 2,3-dihydroxybenzoylamide

derivatives of polyamines, e.g., N1,N8-bis(2,3-dihydroxybenzoyl)sperm-

idine (compound II) and parabactin, first isolated from cultures of

Paracoccus denitrificans by Tait (33) and vibriobactin, isolated from

Vibrio cholerae by Neilands et al. (34) (Fig. 1-3).

3. It has been postulated that the transferring receptor is a

potential marker for the identification of proliferating cells (35).

















HHO



H N N
H H
a.





O ROH

0N O RIR
N. OR
RO
O CH3 0 o 0


H
b. R=H
c. R=CH3



RO RO


S OR
N O N O0
OR

O CH3 O CH3 0

N N N
H H
d. R=H
e. R=CH3


Figure 1-3. The catecholamide iron chelators.

a. Compound II
b. Parabactin
c. Tetramethyl parabactin
d. Vibriobactin
e. Permethylated vibriobactin











9

In a study using monoclonal antibodies to the transferring receptor, it

was found that in all cases of breast carcinomas transferring receptors

could be detected while in samples of normal breast tissue little or no

evidence of staining was exhibited (36). The evidence that neoplastic

cells, as indicated by their uninhibited proliferation and increased

means of iron acquisition, have higher iron requirements in comparison

to host cells provides a rationale for the potential selectivity of

iron removal by the microbial iron chelators as a target in antineo-

plastic strategy.

4. Evidence suggested that the catecholamide chelators could

penetrate cellular membranes. A correlation between antiproliferative

activity and octanol/saline partition coefficients for a series of

siderophores has been established (37). It was found that the ratios

of the ligand partition coefficients between octanol and phosphate

buffered saline for the catecholamides versus dihydroxybenzoic are

closer to the ratios of the ligand's antileukemic activity than are the

ratios of their binding constants. This suggested that cellular pene-

tration as well as iron-chelation potential may determine the chelat-

or's antiproliferative activity.

5. Parabactin and compound II have pronounced inhibitory effects

on the growth of L1210 cells in vitro with IC50 values in the micro-

molar range (38). These values are in the range of antineoplastics

currently being used and are at concentrations attainable as serum

levels in vivo.

6. There is good evidence that the catecholamide siderophores,

unlike previously studied chelators, are relatively nontoxic to the











10

host. It has been observed that compound II, the biological precursor

to parabactin, has an LD50 in mice greater than 800 mg/kg when given as

a single intraperitoneal (i.p.) injection (39). It has been further

observed that compound II and parabactin at concentrations up to 2 mM

(100 times the 50% inhibitory concentration for murine L1210 leukemia

in vitro) appeared to be nontoxic to a confluent monolayer of monkey

kidney cells (37).

7. The iron dependent enzyme ribonucleotide reductase is a suit-

able target for antineoplastic activity of the catecholamide chelators.

Hydroxyurea, an inhibitor of ribonucleotide reductase (40), has found

clinical usefulness as an anticancer agent. A number of other less

avid iron chelators such as desferrioxamine (41), a-picolinic acid (42)

as well as certain toxic thiosemicarbazones (43) have been shown to

inhibit tumor growth by interfering with ribonucleotide reductase

activity.

Parabactin and compound II have been shown to be potent inhibi-

tors of ribonucleotide reductase activity and to block DNA biosynthesis

at the Gi-S boundary of the cell cycle. However, RNA and protein syn-

thesis were inhibited to a much lesser extent, and DNA polymerase activ-

ity was essentially unaffected (38). Alterations in intracellular deox-

ynucleoside triphosphate levels revealed elevated dTTP pool and lower-

ing of dATP pool size as is characteristic of other ribonucleotide

reductase inhibitors. Cytidine diphosphate (CDP) reductase activity

was inhibited by >97% in L1210 cells treated for 4 h with 5 pM parabac-

tin. The cell cycle block seen in L1210 cells after a 4 h exposure to 5

UM parabactin could be partially reversed by the addition of exogenous











Fe(C1)3 and the bulk of the cells cascade into S-phase 3 h later.

Catecholamide chelators have also demonstrated pronounced effects

against the DNA virus Herpes simplex type I (37), another ribonucleo-

tide reductase dependent system, while they were inactive against the

RNA virus, Vesicular stomatitis.

It is clear that there is sufficient evidence to suggest that

iron withholding should be pursued as a means of controlling the pro-

liferative process. The microbial catecholamide siderophores and their

synthetic derivatives may represent an important class of pharmacologi-

cal agents with potential value in the treatment of proliferative

disorders. They may prove useful in cancer chemotherapy as single

agents or in combination with other agents.

Polyamine Analogues

The second part of this dissertation deals with the evaluation of

the N-alkylated derivatives of the polyamines as potential antiprolif-

eratives. The role of polyamines in proliferative processes has been

extensively reviewed in recent years (44-50). The polyamines spermi-

dine, spermine, and the diamine putrescine (Fig. 1-4) are present in

all mammalian cells and are required for normal cellular maintenance,

proliferation, and differentiation. These basic amines are extensively

protonated at physiological pH and capable of electrostatic interac-

tions with a variety of macromolecules including nucleic acids and pro-

teins. Although many roles have been ascribed to the polyamines, their

exact functions in cellular physiology are not yet well understood.

The intracellular concentrations of these polyamines are main-

tained through a highly regulated metabolic system (49,50) (Fig. 1-5).


















H2N vNH2

a.






H
H2N__N NH2


b.


H

H2.N N2
H
C.


Figure 1-4.


The polyamines.


Putrescine
Spermidine
Spermine







13








ARGININE

1 UREA


ORNITHINE

3 --- CO2


PUTRESCINE
DECARBOXYLATED 7
S-ADENOSYLMETHIONINE > 4
2 N -ACETYLSPERMIDINE
2 -CO2 5'-METHYLTHIO-
S-ADENOSYL- ADENOSINE
METHIONINE
2CO SPERMIDINE
2 CO2 7
DECARBOXYLATED
N'-ACETYLSPERMINE
S-ADENOSYLMETHIONINE 5

5'-METHYLTHIO- 6
ADENOSINE

SPERMINE



Figure 1-5. Pathway of polyamine synthesis and intercon-
version in mammalian cells. Enzymes involved are: (1) arginase,
(2) S-adenosylmethionine decarboxylase, (4) Spermidine synthase,
(5) Spermine synthase, (6) Spermidine N -acetyltransferase, and (7)
Polyamine oxidase (49).











14
In mammalian cells putrescine is derived from ornithine by the pyridox-

al phosphate-dependent enzyme ornithine decarboxylase (ODC). Ornithine

is available in the plasma or formed fran arginine by the action of

arginase. Levels of ODC are normally low in quiescent cells, and its

activity can be increased manyfold within hours after exposure to such

stimuli as hormones, drugs, tissue regeneration and growth factors

(47).

To convert putrescine to spermidine an aminopropyl group must be

added. The aminopropyl moiety is derived from S-adenosylmethionine

(SAM) which is committed to polyamine biosynthesis as an aminopropyl

donor once it is decarboxylated by S-adenosylmethionine decarboxylase

(AdoMetDC). Levels of AdoMetDC are regulated by the cell's need for

spermidine and the availability of putrescine as a substrate for sperm-

idine synthesis. Its activity is also regulated by growth-promoting

stimuli (49).

The transfer of the aminopropyl group from decarboxylated

S-adenosylmethionine to putrescine is catalyzed by the action of the

enzyme spermidine synthase to yield spermidine. In a similar fashion,

a second aminopropyl transfer to spermidine, catalyzed by spermine

synthase, yields spermine. Despite the similarity of these reactions,

spermidine synthase and spermine synthase are substrate specific

enzymes. The two synthases are thought to be regulated by the avail-

ability of their substrates as well as factors stimulating cell growth.

The other product of the aminopropyl transfer reaction is 5'-methyl-

thioadenosine (MTA). The MTA produced by polyamine synthesis is










15

rapidly degraded and converted by a salvage mechanism into 5'-AMP and

methionine (50).

Spermine may also be interconverted into spermidine and spermi-

dine into putrescine by the action of the enzymes N1-acetyltransferase

and polyamine oxidase. The acetyltransferase uses acetyl-CoA to con-

vert spermine into spermidine and spermidine into putrescine. The N1

acetylated polyamines are good substrates for polyamine oxidase, which

cleaves at the internal nitrogen yielding 3-acetamidopropanal and

putrescine or spermidine. Normally, only small amounts of the acetyl-

ated derivatives are present due to the much greater activity of poly-

amine oxidase (49).

An active transport system distinct from those for amino acids is

also present (49,50). It is unclear why cells would maintain a poly-

amine transport system since intracellular synthesis is normally used

to provide polyamines and extracellular polyamine levels are normally

low. However, uptake and efflux may be important mechanisms by which

cellular polyamine levels are maintained. The uptake pathways for the

polyamines have not been fully characterized at the biochemical level

and there may be multiple transport mechanisms (50).

Rapidly proliferating tissues have high levels of polyamines and

their biosynthetic enzymes (47), and increased quantities of poly-

amines are found in the urine and serum of humans and animals bearing

tumors (51). These facts suggest the potential of impacting on poly-

amine metabolism as a useful target for antiproliferative therapy.

Interruption of the polyamine biosynthetic network has been par-

tially successful as a means of controlling the proliferative process












including the growth of cancer cells (52-54). The antineoplastic

drugs, a-difluoromethylornithine (DFMO) and methylglyoxyal-bis(guanyl

hydrazone) (MGBG) (Fig. 1-6), are both potent inhibitors of enzymes

which are critical to polyamine biosynthesis. DFMO is an enzyme-

activated irreversible inhibitor of ODC which causes depletion of

intracellular polyamine pools (55). It is postulated that DFMO is

recognized by ODC as a substrate and that its decarboxylation yields a

highly reactive intermediate which binds irreversibly to the active

site of the enzyme. The action of MGBG, however, is not specific.

MGBG induces spermine/spermidine acyl-transferase, inhibits diamine

oxidase and AdoMetDC, and has other effects (50). Although it is not

clear which of MGBG's roles is most important in inhibiting the growth

of cancer cells, uptake of the drug by the polyamine transport appar-

atus seems critical. The drug has been regarded as a structural ana-

logue of spermidine, and since its uptake involves an active transport

mechanism cells, can actually concentrate the drug so that millimolar

quantities of MGBG can accumulate intracellularly (56).

Although DFMO and MGBG have shown impressive results against

transplantable murine tumor models, the success of both in a clinical

setting, unfortunately, has been somewhat marginal (50,57). The action

of DFMO on most neoplasms is cytostatic, and the rapid rate of synthe-

sis of new ODC (half-life of less than 1 h in many cell lines) makes it

essential to maintain high drug concentrations since enzyme levels and

cell growth are restored rapidly once drug is removed (49). MGBG has

produced unacceptable host toxicity in early trials.


















CHF2
COOH
H2N
NH2

a.









CHa


,NH2


Figure 1-6. The inhibitors of polyamine biosynthesis.

a. DFMO
b. MGBG











18

The combination of MGBG and DFMO has been reported to have syner-

gistic activity (57). It was speculated that the simultaneous inhibi-

tion of AdoMetDC by MGBG and reduction of ODC levels by DFMO treatment

would enhance tumor cell kill. Furthermore, DFMO depletion of intra-

cellular polyamine pools increases uptake of MGBG by the polyamine

uptake apparatus (58). Since DFMO is relatively nontoxic (53), it was

hoped that potential tumor cell selectivity would be increased while

decreasing MGBG's host toxicity. This does not seem to be the case in

several trials (50). However, the fact that these agents do have anti-

cancer activity, albeit limited, supports the aim of impacting on poly-

amine metabolism as a rational approach for controlling the prolifera-

tive process.

The initial design, synthesis and testing of polyamine deriva-

tives which utilize the polyamine uptake apparatus and exhibit antipro-

liferative properties was conducted by Porter, Bergeron and Stolowich

(59). There are several properties which may be desirable for the

design of polyamine analogues as antiproliferatives: (1) the analogues

may utilize the polyamine uptake apparatus as a means of delivery, and

such uptake may be enhanced by polyamine depletion, (2) the analogues

may fulfill some but not all of the functions of polyamines and thereby

interfere with normal polyamine metabolism and/or function, and (3) the

analogues may act to regulate the polyamine-biosynthetic pathway lead-

ing to depletion of cellular polyamines and their biosynthetic en-

zymes.

In a systematic approach, a number of spermidine analogues were

synthesized and studied for their structure/function relationships









19

(59-65) and some were found to meet the above requirements. It was

found that, in terms of uptake, retention of charge is of major import-

ance since N-acetylation as opposed to N-alkylation at either the cen-

tral or the terminal nitrogens decreased the ability of the analogue to

compete with spermidine for uptake into the cell (65). While the

N1,N8-alkyl derivatives (65) were poorer competitors for spermidine

uptake than N4-alkyl derivatives, they accumulated to high concentra-

tions during longer incubations (66). The antileukemic activity of

several N4-alkyl spermidine analogues was enhanced by pretreatment of

cells with DFMO (67). Structural analogues of spermidine differing in

aliphatic chain length separating the amines were also studied (68).

Several of the homologues were found to support cell growth and no more

than a one-carbon extention was tolerated for biological activity.

Certain of the N1,N8-alkyl spermidine derivatives, particularly

N1,N8-diethylspermidine (BESPD), were found to inhibit L1210 cell

growth, diminish ODC activity, reduce AdoMetDC, and deplete intracell-

ular polyamine pools (65,66,69). It has been concluded that BESPD

affects polyamine biosynthesis, at least, at one critical point by

regulating ODC at a translational and/or post-translational level in a

manner similar to the natural polyamines. However, the analogue is

apparently incapable of fulfilling certain critical functions of natur-

al polyamines necessary for normal cellular function and proliferation

(65). Depletion of intracellular polyamine pools by BESPD may also

occur by displacement of the natural polyamines from their intracell-

ular binding sites. In addition, potent induction of the spermidine/

spermine acetyl transferase activity by BESPD (70) could lead to the











20

back-conversion of spermine and spermidine to putrescine and efflux of

the polyamines from the cell.

The effects of these polyamine analogues may vary considerably

from one cell type to another depending on the cellular requirements

for polyamines. Despite inhibition of cell growth, BESPD was not toxic

towards L1210 cells (69). However, BESPD at concentrations of greater

than 5 vM was found to be cytotoxic to human large-cell lung cancer

cells normally quite resistant to DFMO (71).

The design, synthesis, evaluation, and study of the biochemical

and physiological effects of this new class of antiproliferatives is

currently the area of intense investigation.















CHAPTER II
CATECHOLAMIDE CHELATORS

Materials and Methods

The catecholamide chelators and methylated analogues were synthe-

sized by methods previously described (30-32,72). Stock solutions (2

mM) of the catecholamide chelators were made in 50% (v/v) ethanol and

water. Stock solutions of Adriamycin and ara-C were prepared in water.

BCNU was prepared in 50% (v/v) ethanol and water. All solutions were

passed through a 0.2 um filter prior to use. Reagents for cell culture

were obtained from Gibco (Grand Island, NY). Cell culture flasks, 25

and 75 cm2, were purchased from Corning (Corning, NY). Adriamycin

was obtained from Adria (Columbus, OH), carmustine (BCNU) from Bristol

(Syracuse, NY), hydroxyurea from Aldrich Chemical Co. (Milwaukee, WI)

and cytosar (ara-C) from UpJohn (Kalamazoo, MI). Diamidino-phenylin-

dole (DAPI) and propidium iodide (PI) were obtained from Sigma (St.

Louis, MO). Desferrioxamine was provided by Ciba-Geigy (Basel, Swit-

zerland). Iron atomic absorption standard solution, 1003 ug/mL Fe in

2% HNO3 (Aldrich) was used as an Fe(III) source. Cremophore RH-40

was provided by BASF-Wyandot Corp. (Parsipany, NJ). Rhodamine-123

(Rh-123) was purchased from (Eastman Kodak Company, Rochester, NY).

[methyl-3H]Thymidine was obtained from New England Nuclear (Boston,

MA). DBA/2 mice were obtained from Jackson Laboratories (Bar Harbor,

ME).













Cell Culture.--Murine L1210 leukemia cells and human Burkitt

lymphoma (Daudi) cells were maintained in exponential growth as

suspension cultures at 0.3 x 104 to 1.5 x 106 cells/mL and 1 x 105 to

2 x 106 cells/mL, respectively. Chinese hamster ovary (CHO) cells and

murine B-16 melanoma cells were grown as monolayers to near confluency

in 48 h after reseeding at 2.5 x 105 cells/25 cm2 flask. Human fore-

skin fibroblasts (HFF) cells were grown as monolayers to near conflu-

ency in 96 h after a 1:4 dilution of cells in fresh medium. Monolayer

cells were lifted as single cell suspensions with 0.05% trypsin and

0.02% ethylenediaminetetraacetic acid tetrasodium salt (2 mL/25 cm2)

and diluted with Hanks' balanced salt solution. All cells were grown

in complete medium containing RPMI-1640, 2% 4-(2-hydroxyethyl)-l-piper-

azinethanesulfonic acid/3-(N- morpholino)propanesulfonic acid (HEPES--

MOPS buffer), and 10% fetal bovine serum unless otherwise stated.

Cells were grown in 25 cm2 tissue culture flasks in a total volume of

10 mL under a humidified 5% CO2 atmosphere at 370C.

IC50 Value Determinations.--Cultures in logarithmic growth were

treated with compounds of interest at concentrations ranging from 10-3

to 10-7 M. Cells were counted by electronic particle analysis (Coulter

Counter, Model ZB1, Coulter Electronics, Hialeah, FL) and confirmed

periodically with hemocytometer measurements. Cell viability was

assessed by trypan blue dye exclusion.

The percentage of control growth was determined as follows:


% of control growth = 100 x [Final treated cell no.-initial
inoculum]/[Final untreated cell no.-initial inoculum]











23
The 50% inhibitory concentration (IC50) value is defined as the

concentration of compound necessary to reduce cell growth to 50% of

control growth after defined intervals of exposure. The IC50 values

were determined as the chelator concentration necessary to inhibit cell

growth to 50% of the control growth at 48 h for all cell lines except

HFF cells, for which the IC50 values were determined at 96 h.

Preformed chelator-iron chelates were formed by the addition of

equimolar amounts of chelator and iron to the media 30 min prior to the

addition of the cells.

A 50% ethanol and water solution in appropriate amounts was used

to treat cells in all control flasks. Addition of iron and ethanol

alone in the concentrations used had no significant effect on growth

studies.

Stock suspensions of L1210 cells were maintained for 48 h, with

normal 10-12 h doubling times, in complete mediums containing 5%, 10%,

and 20% fetal bovine serum. The IC50 value of parabactin was deter-

mined for cells grown in each medium.

Drug Combinations.--L1210 cells in suspension culture were incu-

bated in the presence of parabactin and antineoplastics at various con-

centrations and combinations. Two types of experiments were performed

for each drug combination.

a. To obtain median effect plots for simultaneous drug combina-

tions, cells at 3 x 104 cells/mL were exposed to a constant ratio of

chelator to antineoplastic, a ratio in which absolute concentrations

were varied. The absolute concentrations of the drugs were chosen

in the neighborhood of their IC50 values. This gave the ratio of








24

parabactin to ara-C of 100:1, of parabactin to Adriamycin of 100:1, and

of parabactin to BCNU of 1:1. The growth was monitored at 48 h, and an

IC50 value was determined. A separate experiment was run to determine
the IC50 value of parabactin, ara-C, Adriamycin, and BCNU.

b. In a second experiment, cells at 3 x 105/mL were first incu-

bated with 5 pM parabactin for 5 h at 370C, then washed twice with

fresh medium, resuspended at 3 x 104 cells/mL, allowed to incubate at

37C for 3 h (allowing the cells to cascade into S phase), and then

treated with varying concentrations of the antineoplastic. Growth was

monitored every 12 h for 96 h.

Additional experiments were carried out with various combinations

of parabactin and Adriamycin, ara-C or BCNU treatments.

c. Cells at 3 x 105 cells/mL were treated with 5 .M parabactin

for 3 h, and then BCNU was added at varying concentrations. The flasks

were allowed to incubate for an additional 5 h. The cells were next

washed with fresh medium and resuspended at 3 x 104 cells/mL, then

their growth was monitored for up to 96 h.

d. Cells at 3 x 105 cells/mL were treated with 5 yM parabactin

for 4 h, and then Adriamycin was added. The flasks were allowed to

incubate for an additional 2 h. The cells were next washed with fresh

medium and resuspended at 1 x 105 cells/mL, then their growth was mon-

itored for 48 h.

e. Cells at 3 x 105 cells/mL were treated with 2 yM parabactin

for 4 h, and then ara-C was added and the flasks were allowed to

incubate for an additional 12 h. The cells were next washed with fresh

medium and resuspended at 5 x 104 cells/mL, then their growth was mon-

itored for up to 120 h.











25
f. Cells at 3 x 105 cells/mL were simultaneously treated with 5

pM parabactin and ara-C. The cells were allowed to incubate for 5 h,

washed with fresh medium, resuspended at 5 x 104 cells/mL, and their

growth was monitored for up to 96 h.

In each case, the drug concentrations and treatment times are

indicated on the figure legends.

Data Analysis for Simultaneous Drug Combinations.--Briefly, this

procedure involves (a) determination of an IC50 value for each drug and

(b) determination of an IC50 value for a constant ratio of two drugs.

The molar ratio of the drugs is maintained while the absolute concen-

tration of the drugs is changed. The cell growth data from these

experiments are plotted as log Fa/Fu versus concentrations (log [c]),

where Fa is fraction of cells affected, Fu is fraction of cells unaf-

fected, and c is molar concentration. The data from the median effect

plots were analyzed by the method of Chou and Talalay (73) according to

the following equation.


A/DA + B/DB = CI (A)

The D values are the IC50 concentrations of each of the drugs alone,

the A and B values are the concentrations of each drug at the IC50

value of the constant ratio combination, and CI is the combination in-

dex. When the CI is in excess of 1, the combination is "antagonistic."

When it is less than one, the combination is "synergistic"; and when

the CI is equal to 1, the combination is "additive."










26

Regrowth Studies.--L1210 cells in suspension culture at approxi-

mately 3 x 105 cells/mL were incubated in the presence of chelator for

5 h at 370C. The cells were then washed twice in fresh complete medium,

resuspended at a final concentration of 5 x 104 cells/mL, and incubated

at 370C. Cell samples during logarithmic growth were counted for up to

60 h after ligand treatment and compared to controls.

Cloning Assay.--Cells at 3 x 105 cells/mL were incubated with 5

uM parabactin for 5 h and washed with fresh medium. Treated cells were

then plated in triplicate 96-well microtiter plates at 0.4 cells/well

with each well containing 100 yM of sample. The plates were incubated

at 370C in a humidified incubator in an atmosphere of 5% CO2 and 95%

air. The plates were examined with an inverted phase microscope at

X100 magnification. The final number of colonies per plate was quanti-

tated at 7 days after plating. Groups of 50 or more cells per well were

identified as having been cloned from a single viable cell.

Viability Assay.--Cell viability was assayed by two dye exclusion

methods. In one method cells were diluted 1:1 with 0.4% trypan blue in

phosphate-buffered saline and counted in a hemacytometer using a light

microscope. In the second method 106 cells were diluted in 1 mL of a

1.12% sodium citrate solution of propidium iodide (50 ug/mL) and count-

ed on a fluorescent microscope.

Viability determination by Rh-123 uptake (74) was assayed by add-

ing the mitochondrial dye Rh-123 (10 yg/mL) to a aliquot of 106 cells.

The samples were incubated at 37C for 10 min and washed once with med-

ium. Trypan blue was added, prior to counting, to give a final concen-

tration of 0.2 ug/mL. Rh-123 and trypan blue stained cultures were












observed in a hemocytometer using a Zeiss epifluorescence Axioscop

microscope. The excitation wavelength was 485 nm.

Flow Cytometric Analysis.--Cell analysis was performed with a

RAT-COM flow cytometer (RATCOM, Inc., Miami, FL) interfaced with a

microcomputer (IBM-XT). Cell samples of 106 cells were taken at vari-

ous intervals and stained with DAPI (10 pg/mL) in a nuclear isolation

media (NIM-DAPI) (75). DNA distributions were obtained from the fluor-

escence and analysis of the DAPI-stained cells.

Alternately, cell analysis was performed with a FACS II flow

cytometer (Becton Dickinson FACS Systems, Sunnyvale, CA) interfaced

with a microcomputer (Hewlett Packard 45B, Fort Collins, CO). Samples

of 106 cells were removed and stained with PI and then exposed to

RNase. DNA distributions were obtained from analysis of the red fluor-

escence from the PI-stained DNA (76).

Effect of Percent Fetal Bovine Serum in the Medium on Cell Cycle

Kinetics.--Stocks of L1210 cells were maintained for 48 h with normal

10-12 h doubling times in complete mediums containing 5%, 10%, 15% and

20% fetal bovine serum. The time course for the effect of 5 pM para-

bactin on the cell cycle kinetics of cells grown in each of the mediums

was followed for up to 8 h.

Bromodeoxyuridine (BrdUrd) Incorporation Studies.--Cultured L1210

cells were treated with 10 UM vibriobactin, washed, resuspended, and

incubated as in the regrowth studies. At times, 0, 5, 10, 15 and 20 h

of regrowth cell samples were removed for dual parameter flow cytomet-

ric measurements of cellular DNA content and amount of BrdUrd incorpor-

ated into cellular DNA by methods presented elsewhere (77,78).











28
Radiolabeled Thymidine Incorporation.--Cells treated with 5 uM

parabactin for 5 h were washed, resuspended in fresh complete medium,

and incubated at 370C for 30 min with [3H]thymidine (specific activ-

ity, 80.9 Ci/mmol; 1 yCi/mL) in triplicate tubes containing 106

cells in a total volume of 1 mL of complete medium. Labeling was

halted by the addition of 0.5 mL of ice-cold Hanks' balanced salt

solution (HBSS) containing thymidine (1 mg/mL) to each tube. The cells

were washed, resuspended in 1 mL of 10% trichloroacetic acid in HBSS

containing thymidine (1 mg/mL), and allowed to sit on ice for 30 min.

The acid-precipitable material was filtered and the filter washed.

Then filters were air dried and counted in Biofluor (New England

Nuclear) scintillation fluid. Background labeling was evaluated in

cell samples pulsed for 30 min with [3H]thymidine at 50C.


% of control incorporation = treated cpm treated background cpm x 100
control cpm control background cpm

Cr51Release Assay.--L1210 cells were treated with vibriobactin

(10 pM) for 5 h, washed with fresh media and regrown for 20 h. Approx-

imately 2 x 106 cells were centrifuged and 150 UCi Cr51 added to the

pelleted cells. The pellet was incubated for 45 min at 370C in a 5%

CO2 atmosphere. Sample radiation was counted with an automatic gamma

counter (LKB-Wallac RiaGamma 1274, Wallac Oy, Finland).

A microwell assay using triplicate samples of 200 yL containing

1 x 105 cells was employed as described elsewhere (79). The percent











29

Cr51 release for the control cells was compared to that for the vibrio-

bactin treated cells.

Animal Studies.--The murine L1210 leukemia cells were maintained

in DBA/2J mice. Cells from a single mouse which was injected i.p. with

1.25 x 104 cells 7 days earlier were harvested and diluted with cold

saline so that an inoculum of 105 cells could be administered by a

0.25 mL i.p. injection. In each study, mice were injected with L1210

cells on day 0.

The catecholamide chelators were first solubilized in 20% Cremo-

phor RH-40 in 0.9% saline with sonication and gentle heating (<600C)

and then diluted with an equal volume of 0.9% saline. The chelator was

administered by i.p. injection according to the appropriate dosing

schedule. Concentrations of chelator were adjusted so that the mice

were injected with a volume of 1-2 mL/100 g of animal/dose (i.e.: a 25

g mouse was injected with 0.25-0.5 mL of drug solution).

The in vivo effectiveness of parabactin alone and combinations of

parabactin with ara-C or Adriamycin against the L1210 ascites tumor

were studied. Ara-C and Adriamycin were diluted in 0.9% saline so that

the mice were injected i.p. with a volume of 1 mL/100 g of animal/dose.

Due to the possible toxicity or antitumor activity of Cremophor RH-40,

groups of mice treated with 10% Cremophor RH-40 as well as untreated

mice served as controls. Groups of 6 mice were used in each treatment

schedule.

One parameter used for treatment evaluation was mean animal sur-

vival time (percent increased life span, % ILS).










%ILS = 100 x [mean survival time treated animals-
mean survival time controls]/[mean survival time controls]

Another parameter used for treatment evaluation was tumor burden.
The L1210 cells were harvested from the animal's peritoneum by two

washings with 5 mL of 0.9% saline. The total number of cells recovered
was determined by counting on a hemocytometer.
The in vivo effect of a single i.p. injection of parabactin (100

mg/kg) on L1210 cell cycle kinetics was studied in mice which had been
injected i.p. with 1 x 105 L1210 cells 6 days earlier. Mice were sac-
rificed by cervical dislocation each hour after injection of the che-
lator and the cells were harvested from the peritoneum with HBSS. The
cells were stained with NIM-DAPI and DNA content measured by flow cyto-
metric analysis.
Results

Inhibition of Cell Growth
The effectiveness of various iron chelators as inhibitors of

L1210 cell growth was assessed by comparison of their IC50 values
(Table 2-1). The bidentate compound, dihydroxybenzoic acid, while
being a good iron chelator (Kf = 1036M-1) (5,37) is not very effective
cell growth inhibitor, requiring millimolar concentrations. Hydroxy-
urea, which structurally resembles a bidentate hydroxamate chelator,
was only fair at inhibiting cell growth, with an IC50 value of 40 yM.
The hexacoordinate hydroxamate iron chelator, desferrioxamine, and the
tetracoordinate catecholamide chelator, compound II, showed similar
activity against L1210 cell growth with IC50 values of 8 UM and 7 pM
respectively. The most effective of the compounds tested were the
hexacoordinate catecholamide chelators, vibriobactin and parabactin,











TABLE 2-1
IC50 VALUES OF VARIOUS CHELATORS AGAINST
CULTURED L1210 LEUKEMIA CELLS



Chelators IC50 Values at 48 H


Dihydroxybenzoic Acid 2.8 mM

Hydroxyurea 40 yM

Desferrioxamine 8 yM

Compound II 7 yM

Vibriobactin 2 yM

Parabactin 1.5 yM










32

having IC50 values of 2 UM and 1.5 vM respectively. The dose-effect
curves generated in these studies indicate linear inhibitory effects
for parabactin and vibriobactin within a very narrow range of ligand
concentrations (Fig. 2-1). For example, at concentrations of greater
than 5 vM, cells uniformly grow to only about 10% of controls, while
below 1 pM approximately 90% control growth is obtained.
The time dependence of growth inhibition of cultured L1210 cells

by the siderophores is shown in Figures 2-2 and 2-3. The concentration
of 10 .M was selected for both parabactin and vibriobactin as a value
at approximately 5 times their IC50 values. The onset of growth
inhibition by both siderophores was almost immediate. Both chelators
eventually caused a total cessation cell growth. In either case, the
growth inhibition could be prevented with the addition of an equimolar
amount (10 uM) of Fe(III) to the culture medium at the same time as the
chelator.
Methylation of the ligand catechol hydroxyl groups, as in tetra-
methyl parabactin (Fig. 1-3c) and permethylated vibriobactin (Fig.
1-3e), greatly reduced the growth inhibitory activity of both chelat-
ors. Total methylation of the three catechol groups of vibriobactin

completely eliminated its activity at a concentration 10 UM (Fig. 2-2).
In fact, it showed no inhibitory activity at concentrations of up to 50
vM. Tetramethyl parabactin (10 yM), while having greatly reduced
inhibitory activity, did have some effect on L1210 growth (Fig. 2-3).

In fact, it had an IC50 value of 30 uM.
The IC50 values for parabactin were determined against L1210
cells grown in culture mediums with varying concentrations of fetal









33







100





S---0- Parabactin
1 0 Vibriobactin
3 10-









S1
0 T







0 10 20 30 40 50 60 70 80 90 100

PERCENT CONTROL GROWTH

Figure 2-1. Dose-response curves illustrating the effects of
increasing concentrations of parabactin and vibriobactin on the
growth of L1210 cells.
















1000000













S 100000
..J




----0-- Control
Vibriobactin 10uM + Fe(III) 10uM
Permethylated vibriobactin 10uM
Vibriobactin 10uM


10000 I
0 10 20 30 40 50

TIME (hours)

Figure 2-2. Comparison of the effects of 10 pM vibriobactin,
in the presence or absence of Fe(III), and 10 pM permethylated
vibriobactin on the growth of L1210 cells.
















1000000















100000
..J







--o- Control
-Parabactin 10uM + Fe(lll) 10uM
-- Tetramethyl parabactin 10uM
-- Parabactin 10uM

10000 ,---
0 10 20 30 40 50 60

TIME (hours)

Figure 2-3. Comparison of the effects of 10 yM parabactin,
in the presence or absence of Fe(III), and 10 uM tetramethyl
parabactin on the growth of L1210 cells.











bovine serum (Fig. 2-4). Cells grown in 5, 10 and 20% FBS had IC50
values of 0.8, 1.2, and 1.6 yM respectively.
Parabactin was also found to have potent growth inhibitory ef-
fects on cultured B16-melanoma and CHO cells exhibiting a 48 h IC50
value of 1.5 VM for both cell lines. Both vibriobactin and parabactin
exhibited a 48 h IC50 value of 2 pM for Daudi cells and parabactin was
active, with a 96 h IC50 value of 2 uM, against HFF cells. The IC50
value for HFF cells was determined at 96 h due to their relatively slow
doubling time (45-50 h) compared to the other cell lines studied (12-24
h).
Treatment Reversibility
When cells were treated for 5 h with 5 UM parabactin, the growth-

inhibitory effects of parabactin were found to be reversible on washing
away the ligand. Viability of parabactin-treated cells, taking cell
samples every 5 h for 35 h after washing away of the ligand, was
verified by dye exclusion and Rh-123 uptake methods. At 10 h after
washing, the viability decreased to a minimum of 88% by the dye

exclusion methods and only 85% by Rh-123, the most sensitive of the
methods used. These results were in agreement with the cloning assay
which indicated 91% of colonies were cloned or single cell after
short-term exposure (5 h) to the chelator compared to single control
cells. The number of cells per colony from the parabactin-treated
cells was consistently less than from the control cells.
Cells having been treated for 5 h with 10 yM vibriobactin showed
similar reversibility of inhibitory effects on washing away of the lig-

and and reseeding the cells in fresh medium. A 51Cr release assay per-
formed on vibriobactin treated cells at 15 h after removal of the lig-
and showed no increase in 51Cr release relative to control cells,

















2.00


1.75
z
O
I 1.50

I-
Z 1.25
0
z
O
8 1.00


0 0.75

I-
Z 0.50-

o
'n 0.25


0.00 *
0 10 20


% FETAL BOVINE SERUM

Figure 2-4. The effect of varying amounts of fetal bovine
serum in the L1210 growth medium on the 50% inhibitory concentra-
tion of parabactin.









38

establishing little, if any, cytotoxicity due to short term (less than

5 h) 10 uM vibriobactin.

However, when cells were treated for an extended period of time

with 5 uM parabactin, cytocidal activity was observed (Table 2-2). For

example, at 60 h, only 30% of the cells were viable as determined by

trypan blue exclusion. Dye exclusion of cells treated with 10 yM vib-

riobactin for 48 h indicated only 10% viability, and only 20% viability

of cells exposed to 100 1M desferrioxamine for 60 h. Surprisingly, 300

pM hydroxyurea was the least toxic of the compounds tested with only

25% cytotoxicity after 60 h of treatment. Although in dividing HFF

cells, growth was inhibited by parabactin (IC50 value of 2 yM at 96

h), treatment of a confluent, nondividing, monolayer of HFF cells with

20 pM parabactin for 24 h appeared to be nontoxic by light microscopy.

Also, the treated confluent cells diluted and regrown in drug-free med-

ium grew at the same rate as control cells.

Effects of the Catecholamide Chelators
on Cell Cycle Kinetics

Flow cytometric analysis was employed as the method to follow the

effects of the chelators on the cell cycle progression. The time

course for the effect of 10 yM vibriobactin on the DNA content of L1210

cells is seen in Figure 2-5. Flow cytometric analysis of untreated

L1210 cells reveals high S (50%) and G2-M (15%) phase components of

the cell population with about 35% of the cells having G1 phase DNA

content. After only 2 h alterations in the DNA content of the vibrio-

bactin treated cells were observed. There was a substantial effect on

the cell cycle kinetics with a clear block at the G1-S border after 3










TABLE 2-2
CYTOCIDAL ACTIVITY OF CHELATORS


Percent of Activitya

Ligand 0 12 h 24 h 48 h 60 h


Parabactin (5 pM) <2 5 15 49 70

Vibriobactin (10 pM) <2 4 36 90 -

Desferrioxamine (100 pm) <2 5 20 60 80

Hydroxyurea (300 yM) <2 15 25

aThe percentage of L1210 cell kill in cells treated with parabactin,
vibriobactin, desferrioxamine and hydroxyurea with treatment time.
Cell viability determined by trypan blue exclusion.



















A





S G2

I I


0.5 +


-M


160 200


0 40 80 120 160 200


40 80 120 160 200


DNA CONTENT


Figure 2-5.


Flow cytometric analysis of L1210 cells.


Control cells.
Cells treated with 10 pM vibriobactin for
Cells treated with 10 pM vibriobactin for
Cells treated with 10 pM vibriobactin for


0.5 4


0 40 80 120 160 200


40 80 120


3 h.
4 h.
5 h.


"' -


1.0 4


1.0 +











h. Within 5 h, cells with G2-M phase content were eliminated, and

cells with S phase content were reduced to 30%. The cells were pre-

vented from entering S phase and 70% of the cells have G1 phase con-

tent. A comparable cell cycle block in DNA synthesis was observed in

the DNA histograms of cells treated with 5 UM parabactin for 5 h (Fig.

2-6). Cells treated for up to 24 h with 5 UM parabactin, however,

showed no greater block in cell cycle kinetics than cells treated for 5

h, and cells treated with less than 4 vM parabactin did not appear to

reduce S phase to the extent that the 5 yM treatment did. For further

studies it was therefore indicated that the minimum concentration and

time for parabactin to effectively eliminate DNA synthesis in L1210

cells was 5 pM for 5 h.

In a study to determine the ability of 5 vM parabactin to cause a

cell cycle block in L1210 cells grown in culture mediums with various

percentages of fetal bovine serum (5%, 10%, 15% and 20%), it was found

that there was no significant alteration in time to reach complete cell

cycle block (5 h) for the cells in any of the different culture medi-

ums.

In a comparative study, parabactin was shown to be a far more

effective cell cycle blocking agent than either hydroxyurea or desfer-

rioxamine (Fig. 2-6). For example, to generate cell cycle blocks simi-

lar to that produced by 5pM parabactin (IC50, 1.5 pM) at 5 h, L1210

cells required 100 yM desferrioxamine (IC50, 7.5 uM) for 5 h or 300 UM

hydroxyurea (C150, 50 UM) for 5 h. When comparing the IC50 value of

each of the ligands, it is clear that the compounds with the higher

IC50 values are also the poorer cycle blocking agents.







42








1.0 1.0-

A I B

C 0.5- 0.5-

z

o .
U
1.0- 1.0
C D
,-t
S0.5- 0.5

ix 0 0

O ,, 0 ,,,
0 120 240 360 420 0 120 240 360 420


DNA CONTENT

Figure 2-6. Flow cytometric analysis of L1210 cells.

A. Control cells.
B. Cells treated with 5 uM parabactin for 5 h.
C. Cells treated with 100 UM desferrioxamine for 5 h.
D. Cells treated with 300 uM hydroxyurea for 5 h.











Synchronization Effects

Flow cytometry analysis of the cell cycle kinetics of cells

treated with 10 vM vibriobactin for 5 h indicated a clear block at the

G1-S border of the cell cycle. Washing and reseeding the treated cells

in fresh complete medium caused pronounced changes in the cell cycle

phases with time (Fig. 2-7). The block in DNA synthesis is released

and at 5 h after removal of the drug the majority of cells have S phase

DNA content. The progression through the cell cycle phases continued

for up to 20 h. Cell cycling was further substantiated by BrdUrd

incorporation into DNA (Table 2-3). The percent BrdUrd incorporation

at various times after removal of the chelator was roughly consistent

with changes in S phase DNA content as determined by flow cytometry.

Cells incubated with 5 yM parabactin for 5 h exhibited a block in

DNA synthesis at the G1-S border (by flow cytometry) with a greatly

decreased S and G2 phase (Figs. 2-6 and 2-8). Concentrations of 100 yM

desferrioxamine and 300 uM hydroxyurea for 5 h were required to produce

similar blocks. On removal of the chelators by simple washing and

reseeding the cells in drug-free culture medium, the block in DNA syn-

thesis was removed and the cells cascaded into S phase. Parabactin-

treated cells maintained a synchronous population of cells for 3 cell

cycles. (Figure 2-8 only includes cells moving into the third cycle.)

However, desferrioxamine- and hydroxyurea-treated cells showed normal-

ized DNA histograms after 12 h.

Radiolabeled Thymidine Incorporation.--Incorporation of [3H]thy-

midine into L1210 cells after the cells were treated with 5 uM

























0 40 80 120 160 200 240 0 40 80 120 160 200 240


1.0 1.0

C


05 05




0 40 80 120 160 200 240 0


40 80 120 160 200 240


0 40 80 120 160 200 240


DNA CONTENT

Figure 2-7. Flow cytometric DNA content analysis of L1210
washed and regrown in fresh medium after a 5-h treatment with
vibriobactin.

Analysis performed after removal of ligand.


cells
10 pM











TABLE 2-3
PERCENT BROMODEOXYURIDINE INCORPORATION INTO L1210
CELLS TREATED WITH 10 pM VIBRIOBACTIN FOR 5 H THEN
WASHED AND RESEEDED IN FRESH MEDIUM



Percent BrdUrd Incorporation

Time after Wash Control Treated


0 44 16

5 50 47

10 53 45

15 47 16

20 49 32




























Figure 2-8. Cell synchronization of L1210 cells by 5 uM
parabactin. DNA content determined by flow cytometric analysis.

A. Control cells.
B. Cells treated with 5 pM parabactin for 5 h.

Cells were then washed, resuspended in drug-free medium, and
samples removed at the following times.

C. 4 h.
D. 6 h.
E. 10 h.
F. 12 h.
G. 14 h.
H. 16 h.
I. 18 h.
J. 22 h.


































0.

2BD0



140


A



G1 S G62-M
I S

0 60 120 180 240 300 350 420480





0 60 1 180 240B300 3 200



O 60 120 180 240300 350 420480


S I I I I I I I I I I I I- 1
0 60 120 180 240300 30 420480
460-

D

220-



O- L .I.
0 60 120 180 240 300 350 420480

140

E

370



O 'I 10 10 2300 350 4204I 0
0 60 120 180 240 300 360 420480


2s0
o- F

175




0 60 120 180 240 300 360 420480
27-







0 ,
0 60 120 180 240 300 350 420480
170










520:
: \ H

260-




0 60 120 180 240 300 360 420480

I







0 60 120 180 240 300 3i0 420480
3D -



40




0 60 120 180 240300 350 420480


DNA CONTENT


C


YY


"









48

parabactin for 5 h and washed (Fig. 2-9) revealed cycling of incorpor-

ation in a time frame consistent with the normal 10 h to 12 h doubling

time of L1210 cells, and with the cycling of DNA content as seen in the

flow cytometric studies (Fig. 2-8).

Drug Combination Studies

Adriamycin-treated Cells.--The Adriamycin IC50 under our experi-

mental conditions was 0.027 UM at 48 h, while the parabactin IC50 was

1.34 vM at 48 h. However, when L1210 cells were treated with para-

bactin and Adriamycin in simultaneous combination, the results indicat-

ed that the combination was antagonistici" (Fig. 2-10). Following the

method of analysis of Chou and Talalay, the cells were exposed to a

constant molar ratio of parabactin to Adriamycin of 100:1. The micro-

molar concentrations at this constant ratio were 2:0.02, 1.75:0.0175,

1.5:0.015, 1.25:0.0125, 1.0:0.01, and 0.5:0.005. The IC50 of the com-

bination was 1.42 vM (CI = 1.57). Chou and Talalay's analysis of the

data (see Materials and Methods) clearly revealed the combination to be

"antagonistic" at the 100:1 ratio.

In a second experiment, cells were treated with 5 yM parabactin

for 5 h, washed free of the ligand, placed in fresh medium to allow the

cells to cascade into the S phase of the cell cycle, and grown with

0.02 yM or 0.03 pM Adriamycin (Fig. 2-11). In this experiment the

parabactin clearly potentiated the activity of Adriamycin (see Discus-

sion). Finally, in a third experiment cells were treated with 5 yM

parabactin for 4 h and then Adriamycin 0.2 pM was added for an addi-

tional 2 h. Cells were washed, reseeded and grown in fresh drug-free

















140


120
Z
0
100-
O
0.
cM 80-
0
z











2 4 6 8 10 12 14
-i 60
0
I-
z
o 40
0

20-



0 2 4 6 8 10 12 14


TIME (hours)

Figure 2-9. Time course for the incorporation of [3H]thy-
midine into the acid-precipitable fraction of L1210 cells grown in
fresh medium after treatment with 5 yM parabactin for 5 h.



























L.
0


a~
O 4) c
o~
S- 0

X
E
U


cl
-o~


CC
C C0
U o E
-Jo 00





0~ A (V



4JE

0 0
u L.
c c+"F
E N CU 4-

co uro
( E V fa
0Ld -0








10 130L.) (U.
aj



.(0-



<
cn<

LL-




0o 4D v M0
; ds c; 6



flIIBI !Doi




















10





o
o
o


V




E 1
_J
-J



----- Control
--- Parabactin 5uM-5h pretreatment
---- Adriamycin 0.02uM
a Adriamycin 0.03uM
Parabactin pretreatment followed by Adriamycin 0.02uM
Parabactin pretreatment followed by Adriamycin 0.03uM

0 10 20 30 40 50 60 70 80 90 100

TIME (hours)
Figure 2-11. The effects of Adriamycin on the growth of
L1210 cells, and Adriamycin on the growth of cells which have been
pretreated with 5 UM parabactin for 5 h then washed to remove the
block in DNA synthesis. The growth of control cells and cells
exposed to 5 pM parabactin for 5 h and resuspended in drug-free
medium is shown for comparison.









52
medium (Fig. 2-12). The results indicated additivity of the two drug

effects.

Ara-C-treated Cells.--The IC50 of ara-C in our L1210 cell assay

system was 0.033 UM, L1210 cells were treated with parabactin and

ara-C simultaneously at a constant molar ratio of 100:1. The concen-

trations at a constant ratio of parabactin to ara-C were 1.75:0.0175,

1.5:0.015, 1.25:0.0125, 1.0:0.01, 0.75:0.0075, and 0.5:0.005. The IC50

of the combination was 1.39 pM (CI = 1.33). Again, Chou and Talalay

analysis of the growth data indicated "antagonism" (Fig. 2-13).

However, when cells were first treated with 5 UM parabactin for 5

h, washed with fresh medium, and followed by treatment with ara-C, as

with the Adriamycin:parabactin combination, the parabactin potentiated

the activity of the ara-C (Fig. 2-14).

When cells were simultaneously treated with 5 UM parabactin and

1 pM or 2 yM ara-C for 5 h (Fig. 2-15), or first treated with 2 yM

parabactin for 4 h then 1 yM or 2 UM ara-C for an additional 12 h, then

washed, reseeded and grown in fresh drug-free medium (Fig. 2-16), the

results indicated additivity of the two drug effects for both combina-

tions.

BCNU-treated Cells.--The IC50 for BCNU in our system was 4.13 PM.

When cells were treated simultaneously with parabactin and BCNU at a

molar ratio of 1:1, the IC50 of the combination was 2.28 pM (CI = 1.1),

indicating additivity (Fig. 2-17). The absolute micromolar drug

concentrations to which cells were exposed were 3:3, 2:2, 1:1, 0.5:0.5,

and 0.25:0.25.



















10




O
O
O















Adriamycin 0.2uM-2h
0
0
x




1
.J


---o--- Control

Parabactin 5uM-6h

Adriamycin 0.2uM-2h

-- Parabactin 5uM-4h + parabactin 5uM/Adriamycin 0.2uM-2h



.1 -i I I *
0 10 20 30 40 50

TIME (hours)

Figure 2-12. The effects of 0.2 pM Adriamycin exposure for 2
h, parabactin 5 UM treatment for 6 h, and 5 yM parabactin for 4 h
plus the simultaneous exposure of 0.2 yM Adriamycin for an addi-
tional 2 h on the growth of L1210 cells. Treated cells were washed
and resuspended in drug-free medium at time = 0 h.











54













C)

0 u

Lm



00S

(AU

Lo
w( 0~
0




C'j


c .2

cl 0
0'U 0












77- 0.



t5 m 4- (A
S-
c U) 0D

c 0






f/ c 0
-4 to 0r
C 1-0
'iE 1
,a) m S
S- 0.




go


Lq C 4J M
9

n~j/e.1 E ol

















10





















-Control
S Parabactin 5uM-5h pretreatment
Ara-C 0.03uM
A Ara-C 0.04uM
Parabactin pretreatment followed by ara-C 0.03uM
Parabactin pretreatment followed by ara-C 0.04uM

0 10 20 30 40 50 60 70 80 90 100

TIME (hours)

cells, and ara-C on the growth of cells which have been retreated
with 5 uM parabactin for 5 h then washed to remove the block in DNA
synthesis. The growth of control cells and cells exposed to 5 uM
parabactin for 5 h and resuspended in drug-free medium is shown for
comparison.



















10














E 1

u.I
---o-- Control
Parabactin 5uM-5h
--o-- Ara-C 1 uM-5h
-- Ara-C 2uM-5h
Parabactin 5uM + ara-C 1 uM -5h
Parabactin 5uM + ara-C 2uM -5h




0 10 20 30 40 50 60 70 80 90 100


TIME (hours)

Figure 2-15. The effects of ara-C exposure for 5 h,
parabactin 5 VM treatment for 5 h, and the simultaneous exposure of
5 uM parabactin plus ara-C for 5 h on the growth of L1210 cells.
Treated cells were washed and resuspended in drug-free medium at
time-0 h.























10











_J
LU

--- Control
---*- Parabactin 2uM-16h
--- Ara-C 1uM-12h
Parabactin 2uM-4h + parabactin 2uM/ara-C 1 uM-12h


.1 I I I I l

0 20 40 60 80 100 120

TIME (hours)

Figure 2-16. The effects of ara-C exposure for 12 h on the
growth of L12110 cells, and 2 UM parabactin treatment for 4 h plus
exposure to ara-C and parabactin for an additional 12 h on the
growth of L1210 cells. The growth of control (untreated) cells and
cells exposed to 2 yM parabactin for 16 h is shown for comparison.
Treated cells were washed and resuspended in drug-free medium at
time-0 h.


























-CZ
00



z

C)~
to o


CU
Z t

curr



~0
U aJcc
cu c



4 u 0

-4

~ q~W
Rr E
t 0 0ou
00 =
4@ OO




C


4-)
'4-)
0 U)
o a E





.4- c



Z41



'I- 0
,-4 oU0










0
0li -04

& -


L. :D 0
O = E










to to
4-)








0s 0; 9O




fl/1z1 ~o1












When cells were first treated with 5 1M parabactin for 5 h,

washed with fresh medium, and treated with BCNU immediately, an

additive effect was again observed (Fig. 2-18).

However, when cells were first exposed to 5 uM parabactin for 3

h, then presented with varying concentrations of BCNU for an additional

5 h, washed, and resuspended in fresh medium, cell growth indicated

parabactin potentiation of BCNU activity (Fig. 2-19).

Animal Studies.--Ascites L1210 tumor cells in DBA/2 mice were

blocked and synchronized by a single i.p. injection of parabactin 100

mg/kg (Fig. 2-20). The DNA histogram of the untreated ascites tumor

shows a DNA content distribution similar to that of the cultured cells.

At seven hours after the parabactin treatment a block in DNA synthesis

at the GI-S border of the cells is evident by the reduction in G2 phase

content and the broadening of the GI-S content peak. At 8 h cells

which were blocked at the GI-S border begin to enter S phase, and by

9 h after the injection a large majority of the ascites cells have S

phase DNA content indicating a synchronous population of cells.

There was no significant increased life span in mice given

1 x 105 L1210 cells i.p. and then treated with a daily (OD) injection

of parabactin 100 mg/kg on days 1-6 compared to untreated mice or mice

treated with only the 10% Cremophore drug vehicle. However, there was

an 80% reduction in the ascites tumor burden of mice treated with para-

bactin 100 mg/kg every 8 h (98 h) for 5 injections beginning on day 5,

compared to untreated mice. This dosage schedule could not be








60











10




o
o;

4
71,





-j
-JI
w
O


-a--- Control
-- Parabactin 5uM-5h pretreatment
--o- BCNU 5uM
-- BCNU 6uM
Parabactin pretreatment followed by BCNU 5uM
A Parabactin pretreatment followed by BCNU 6uM


0 10 20 30 40 50 60 70 80 90 100

TIME (hours)

Figure 2-18. The effects of BCNU on the growth of L1210
cells, and BCNU on the growth of cells which have been pretreated
with 5 pM parabactin for 5 h then washed to remove the block in DNA
synthesis. The growth of cells exposed to 5 yM parabactin for 5 h
and resuspended in drug-free medium is shown for comparison.


















10













E1
-J
-J
w

0- Control
0 Parabactin 5uM-8h
-- BCNU 2uM-5h
a, BCNU 3uM-5h
-- Parabactin 5uM-3h + parabactin 5uM/BCNU 1 uM-5h
-- Parabactin 5uM-3h + parabactin 5uM/BCNU 2uM-5h
Parabactin 5uM-3h + parabactin 5uM/BCNU 3uM-5h

0 10 20 30 40 50 60 70 80 90 100110120

TIME (hours)

Figure 2-19. The effects of BCNU exposure for 5 h,
parabactin 5 uM treatment for 8 h, and 5 UM parabactin for 3h plus
the simultaneous exposure to BCNU for an additional 5 h on the
growth of L1210 cells. Treated cells were washed and resuspended
in drug-free medium at time-0 h.



















1500. 1300

A B

750 650



e oo ... o.. .. .. ...

940 260
C D


470- 130



0 00
0 60 120 180 240 300 360 420 480 0 60 120 180 240 300 360 420 41

DNA CONTENT


Figure 2-20. Flow cytometric analysis of ascites L1210 cells
from DBA/2 mice after a single i.p. injection of parabactin 100
mg/kg.

A. Oh
B. 7 h
C. 8 h
D. 9 h









63

continued for more than the 5 doses in the mice due to the observed

toxic effects (constipation and death) of the Cremophore vehicle when

administered alone at concentrations necessary to solubilize this dose

of parabactin.

Various combinations of parabactin with ara-C or Adriamycin in

mice with L1210 tumors (Table 2-4) did not disclose any significant

increases in life span beyond the additive effects of the individual

treatments.

Discussion

Iron chelators have been shown to have profound effects on the

growth properties of various tumor cell lines (37,41-43). The antipro-

liferative activity of the siderophores against L1210 cells as

assessed by their IC50 values indicated that, when ranked according to

their ability to sequester iron, the compounds with the higher iron

binding constants generally demonstrated better antileukemic activity

(37). The most active of the compounds tested here (the hexacoordinate

catecholamide chelators, vibriobactin and parabactin) have an enormous

affinity for iron and exhibited the most potent antiproliferative

activity with IC50 values against L1210 cells of 2 yM and 1.5 UM,

respectively. Both compounds were able to cause nearly complete cessa-

tion of cell growth in 24 h at concentrations above 5 yM while below 1

1M little growth inhibition was seen.

In an effort to ascertain the broadness of activity of the cate-

cholamide chelators, they were tested against various other cultured

cell lines. Parabactin and vibriobactin were found to be potent growth

inhibitors of various animal and human tumor cells with 48 h IC50









64










-- CM O .-4 0 No .-- CJ --4 L O
v-

o 0
(-4
C\J >1
_m c "0


S* o
0*
tf * 0-

z > +1 4-
+1 +1 +1 +1 +1 +1 +1 +1 1 +1 +1 +1 +1 +1+1
W> -
z 3E LO n ko1 r- o LO r~. 0o 0-o LoLo 0: -4-40 -
C-= L/) eV a) 0 c 0 0C00 L Oc4 C;0 00!, iCO a
I- *:0 -*- *4 *-* a*



,C\ I .-0I '-



1-) N a I -A al )

0O I. V C C-4 1 > (a
LIJ o( A .c r(. 'aUo *'a > i

Z U 0' 0 -4 ,-4 'V 000 000 -. .M -
0 00 =OO I 0 000 000 C i
Co m DD a a l )DD) CD Co

c< .- 4- (0 >


I 4-) C 4 I
o (M n V) 0) a),-- ai t I
Ll 0 W =

UJ C e E < e < CD
c : 4- EE E EE E E
-0 00 0 o O on0*
S OM O 0 0 L
SLLJ I I --4 -4 I-




** -,
C C .O

) *- *f- *- 'V '0
4I 4- C L. -0 .0
'V a 'V 4. +-> i
CC C C .0 X> U L'V'
*i I C I 1- M* CL E
4.. 4J. S.. S-- X 0)
-4 7 'V 'M 0-. S-4- a) O)
I- "I XI ea 4j 4-
S-- C r C r' C CC CC 0- 4-4-
Su a. u. U >)>) u >> U >, (v CM C
m< O 'EE (a E 4 -'-4
cL) x0 L.> L) .0 ( fu .0'0 (a 0 4- Cr 'r-V



r- *r- *F-
S- C l-
L- tawCC


u I O EE
l- I I -
SC.0 C *) L L.
0. (WE OkiC n- Ls-VV
x E =
Lu Z 'to.0 U









65

values of approximately 2 yM for all cell lines tested. Parabactin was

also found to be a potent inhibitor of actively-dividing normal HFF

cell growth with a 96 h IC50 value of 2 yM.

That growth inhibition is due to iron chelation seems apparent

from studies in which iron, added at the same time as the chelators,

prevents the effects. This is consistent with previous findings (37)

in which the antileukemic and antiviral viral activities of compound II

could similarly be prevented. Also, methylation of the ligand's cate-

chol groups essentially blocking the siderophore's chelating function-

alities (Fig. 1-3) eliminated the antileukemic activity of vibriobactin

and greatly reduced the activity of parabactin (Figs. 2-2 and 2-3).

The residual activity of tetramethyl parabactin is likely due to the

remaining central oxazoline nitrogen and phenolic hydroxyl group's

ability to chelate iron. It has also been shown that parabactin was

able to suppress the growth of a variety of bacteria, while the addi-

tion of the tetramethylated parabactin analogue or the preformed

ferric-parabactin chelate was unable to inhibit microbial prolifer-

ation (72).

Cytostasis (apparent by dye exclusion, Rh-123 mitochondrial stain-

ing, 51Cr release and clonagenic assays) is observed on a short-term

exposure (5 h) of L1210 cells to 5 UM parabactin or 10 UM vibriobactin,

and cidal activity when the cells are exposed for longer periods of

time (Table 2-2). The stasis is reversible by simply washing the lig-

and away from the cells within the first 5 h. However, a nondividing

confluent monolayer of HFF cells exposed to 20 pM parabactin (10 times

its IC50 value) for 24 h exhibited no apparent toxic effects.









66

The effects of vibriobactin and parabactin on L1210 cell growth

are even more apparent after examining their impact on cell cycle kin-

etics. Both ligands hold the cells at the GI-S border until they are

released by washing the chelator away, at which point cells cascade

into S phase (Figs. 2-7 and 2-8). This synchronous population of cells

can be followed through the cell cycle with time, and in the case of

parabactin a portion of the cells remained synchronized for 3 cell

cycles. When parabactin treated cells are reseeded in fresh medium

there is approximately a 14 h lag in growth relative to untreated con-

trols. At this point cells appear to grow normally with a doubling

time of 12 h (Figs. 2-11, 2-14, 2-18.)

Cycling of [3H]thymidine incorporation into DNA (Fig. 2-9) during

the first 12 h after washing parabactin from treated cells, or cycling

of BrdUrd incorporation into DNA after vibriobactin pretreatment (Table

2-3) was consistent with the time frame of changes in S-phase followed

by flow cytometry. However, the percentage of control incorporation

during certain phases in the cell cycle reveals an actual DNA synthetic

rate which does not correspond to the levels expected from the percent-

ages of cells with S-phase DNA content by flow cytometry (Fig. 2-8).

For example, flow cytometry indicated that at 8 h after washing the

percentage of cells having S-phase DNA content is approximately twice

that (200%) of control; however, [3H]thymidine incorporation corre-

sponds to a DNA synthetic rate of only 120% of control. From these

studies it seems that while a sizeable fraction of the cells moves

synchronously through the cell cycle in 10 to 12 h, there is also a

fraction of cells which are moving more slowly through the cell cycle.

Together, the fraction of cells moving more slowly through the cell









67
cycle along with the decrease in viability of cells after short-term

parabactin treatment may account for the 14 h delay in cell growth

relative to controls.

The phase specificity of the effects of the catecholamides is

similar to that seen with hydroxyurea (80-81) which is believed to

inhibit cell growth through interference with the enzyme ribonucleotide

reductase by a free radical scavaging mechanism (26). The hydroxamate

iron chelator desferrioxamine has also been shown to inhibit DNA syn-

thesis through interference with the same enzyme. Interfering with

this enzyme's activity blocks DNA synthesis, and cells accumulate at

the G1-S boundary of the cell cycle (81). It has been demonstrated

that parabactin is a potent inhibitor of ribonucleotide reductase

(38).

It is highly unlikely that the growth inhibition and cell cycle

effects of the catecholamides are a consequence of extracellular iron

chelation. It has been shown that the removal of iron from transferring

by siderophores occurs at a rate greater than 10% per hour only at high

ligand to transferring ratios (82), and removal is impeded in the pres-

ence of serum proteins (83). Transferrin is supplied to the cell

growth medium along with other growth factors as a 10% serum supple-

ment. Unless the removal of iron from the extracellular medium by the

chelator occurred at a rate far in excess of 10% per hour it would not

account for the early onset of inhibitory effects of the siderophores

on DNA synthesis (3 h) (Fig. 2-5) or cell growth (12 h) (Fig. 2-3)

since L1210 cells grown in medium with 5% or 10% serum proliferate at

essentially the same rate. Also, there was not a proportional increase

in IC50 values or significant change in the time to obtain a complete









68

cell cycle block when parabactin treated cells are grown in 5%, 10% or

20% fetal calf serum. If cell growth inhibition required an estab-

lished concentration of chelator to limit extracellular iron, there

should be a proportional response of IC50 values in relation to changes

in serum concentration in the medium. Hence, if deprivation of iron to

the cells caused by chelator-induced deferration of transferring in the

medium was the mechanism by which the catecholamides inhibit DNA syn-

thesis, the times necessary to see a reduction in cell growth (5 pM

parabactin for 12 h) and DNA content (5 pM parabactin for 3 h) would be

expected to be much greater. Finally, there exists a good correlation

between the antiproliferative activity and lipophilicity of various

catecholamide chelators examined (37). This correlation would not be

expected if extracellular chelation were taking place.

The effects of the parabactin on the cell cycle kinetics of L1210

cells were compared with the effects of desferrioxamine and hydroxyurea

(Fig. 2-6). Cells exposed to 5 yM parabactin experienced a block at

the Gi-S border in 5 h, while desferrioxamine was required to be in

excess of 100 UM for a 5 h exposure. Finally, hydroxyurea needed to be

in excess of 300 vM concentration for the same 5 h exposure time. It

is interesting that the concentration of desferrioxamine required to

block cells at the GI-S border is in excess of 13 times the 48 h IC50

value of desferrioxamine and 7.5 times the 48 h IC50 value of hydroxy-

urea. The cell cycle blocking ability of parabactin at a concentra-

tion of 3.5 times its IC50 value was therefore far superior to that









69

of hydroxyurea or desferrioxamine. The release of the desferrioxamine

(84) and hydroxyurea (85) blocks results in less effective cell cycle

synchronizations, as the DNA histograms are normalized following one

passage through the cell cycle.

As mentioned above, the reversibility of the effects of parabac-

tin is dependent on how long the cells are exposed to the ligand. For

example, when the cells are exposed to 5 1M parabactin for 48 h 59% of

the cells die and 10 yM desferrioxamine kills 60% (Table 2-2).

The potent cell cycle blocking ability of parabactin and its

reversibility suggested its application in combination with other

phase-specific antineoplastic drugs.

Parabactin-treated cells offer two opportunities for potentiation

of growth inhibition.

a. Agents acting at the G1-S border should be more effective

since the cells are held at this phase.

b. Agents acting during S phase should have improved activity

since a larger percentage of the cell population will be synchronized

into the S phase when the parabactin block is released.

When utilized in combination with ara-C, Adriamycin, or BCNU,

parabactin can produce "antagonistic," "additive," or "synergistic"

potentiationn) effects depending on the time frame during which the

drugs are applied.

Two basic experiments were run in the combination chemotherapy

studies.

a. Parabactin was used in simultaneous combination for 48 h with

BCNU, ara-C, or Adriamycin.









70

b. Cells were first blocked with parabactin and held at the GI-S

border, then the parabactin was washed away allowing cells to cascade

into S phase. The cells were then exposed to BCNU, ara-C, or Adriamy-

cin.

The latter experiments relied heavily on cell cytometry tech-

niques to evaluate the fraction of cells in each phase of the cell

cycle after treatment with the catecholamide chelator.

A third experiment was performed with RCNU. The cells were

treated with parabactin for 3 h and then with BCNU for 5 h. Then,

after both drugs were washed away, cells were resuspended and regrown.

When cells are exposed to parabactin and ara-C simultaneously for

48 h, the effect of the combination is an "antagonistic" one (Fig.

2-13). This is in keeping with the previous observations that ara-C,

an inhibitor of DNA polymerase (86), kills cells best in S phase (87)

and that parabactin holds cells at the GI-S border. This parabactin

block thus diminishes the S-phase component of the cell population

reducing the number of cells most susceptible to ara-C. Finally, the

"antagonism" is apparent from the analysis of drug-induced growth

inhibition. The IC50 value for the ara-C direct combination with para-

bactin at a 100:1 ratio was 1.33 yM with a CI>1 indicating an

"antagonistic" effect.

However, when cells were first exposed to 5 UM parabactin for 5

h, washed free of the ligand releasing the block, and then treated with

ara-C, the parabactin potentiates the activity of ara-C particularly at

0.04 UM ara-C (Fig. 2-14). Inspection of Figure 2-14 reveals that the

growth of parabactin-treated cells is approximately 14 h behind the









71
growth of the control cells. Consequently, when comparing the effects

of the parabactin:ara-C-treated cells with ara-C treated cells, the 14

h lag for the parabactin-treated cells must be considered. Therefore,

in evaluating the combination curve relative to the ara-C curve, one

must look at a point on the combination line 14 h ahead of the ara-C

line. Even at this, it is clear that parabactin potentiates the activ-

ity of ara-C.

The same experiments were carried out with Adriamycin. The sim-

ultaneous drug combination experiment with 100:1 ratio of parabactin to

Adriamycin indicated "antagonism." Applying Equation A to the data

from Figure 2-10 resulted in a CI value of 1.57 indicating antagonism.

However, again when the cells were treated first with parabactin to

initiate a block at Gi-S and then washed free of the ligand and treated

with Adriamycin, the effect of Adriamycin was potentiated particularly

at 0.03 uM (Fig. 2-11). Again, one must consider the 14-h lag time in

growth generated by the parabactin. However, even after accounting for

this, reduction in growth rate is about twice that observed for the

Adriamycin itself. In view of the toxicity of Adriamycin, this may be

of some practical significance. As both Adriamycin (88) and ara-C are

proposed to operate best on cells in the S phase, the observed results

are as expected.

The final drug studied in this evaluation was BCNU. When cells

were exposed to parabactin and BCNU simultaneously for 48 h the results

indicated the combination to be additive with a CI approximately equal

to 1 (Fig. 2-17).

When the cells were first treated with 5 yM parabactin for 5 h,

and then released and immediately treated with BCNU (Fig. 2-18), the










72

effects were still additive. However, when the cells were treated with

5 pM parabactin for 3 h, then exposed to BCNU for an additional 5 h,

and then washed and resuspended in culture, there is clearly a potenti-

ation of the combined drug effects (Fig. 2-19). After accounting for

the parabactin-induced lag the effect of the BCNU was two times as

great as it was in the absence of parabactin. It was interesting to

note that 1 pM BCNU in combination with parabactin was more active than

3 yM BCNU alone. In fact, cells treated with 1 uM BCNU (not shown)

grew at the same rate as controls. The literature suggests that BCNU

operates best at the G1-S border and in G2 of the cell cycle, while

cells in S phase are more resistant (89). The BCNU potentiation

effects of the drug combination were consistent with the observation

that cells treated with parabactin were held at the G1-S border. It

remains somewhat unclear, however, why we did not observe synergistic

effects when the drugs were used in direct simultaneous combination

without washing.

The above observations also suggest that proper pretreatment of

cells with parabactin could result in a decrease of the cytoxic effects

of phase-specific antineoplastics by reducing the population of cells

in the phase where the chemotherapeutic exerts its greatest cytotoxic

action. In an attempt to demonstrate the possible "protective" effects

of parabactin treatment against phase-specific toxicity, several exper-

ments were devised. Cells were first treated with parabactin to reduce

the S-phase population of cells. Adriamycin (Fig. 2-12) or ara-C

(Fig. 2-16) was then added for an additional amount of time, and the








73

cells were washed and regrown in drug free medium. Additionally, cells

were treated simultaneously with both parabactin and ara-C for 5 h and

the cells were washed and regrown (Fig. 2-15). The results revealed

additive effects of the combinations in all cases, thus failing to

indicate any protection afforded cells out of synchronization with the

phase specificity of ara-C or Adriamycin by parabactin treatment. The

implications of these results may demonstrate, that while ara-C and

Adriamycin are reported to exert their greatest cytotoxic effects on

cells in S phase, they may enter the cell during any phase of the cell

cycle. These results further imply that studies of simultaneous com-

binations for these types of drugs deserve further consideration.

Analysis of the median effect plots by the method of Chou and Talalay

(73) is only a mathematical manipulation of the 48-h IC50 values of the

drugs at one specific molar ratio. The CI may not be a true reflection

of the activity of the simultaneous combination of two compounds with

such different mechanisms of action and the results must be interpreted

with care.

The ability of parabactin to enhance the chemotherapeutic effects

of various antineoplastic drugs was highly dependent on its ability to

block cell cycle kinetics and to effectively reverse this block releas-

ing the cells into a synchronized cell cycle. A consideration of the

IC50 values of the drugs is probably relevant when considering the

effectiveness of these compounds as cell cycle blocking agents. At 3.5

times its IC50 parabactin was an excellent cell blocking agent; it was

better than desferrioxamine at greater than 13 times its IC50 and the

hydroxyurea at 7.5 times its IC50 (Fig. 2-6). Furthermore, one of

the most critical issues was the fact that parabactin-treated cells

maintained partial synchronization for at least three cell cycles while









74

cell synchronization disappeared after one cell cycle in desferriox-

amine- and hydroxyurea-treated cells.

The antiproliferative effects of parabactin were also studied in

vivo. The L1210 cell tumor model has been used as a useful preclinical

screening system for potential antineoplastics (90). The murine L1210

leukemia is a convenient and very reproducible in vivo model system.

The tumor is very aggressive having an in vivo doubling time of about

10-12 h. Mean survival time of animals injected with 105 cells is

approximately 9.5 days. It has been determined that one viable cell

will kill an animal in about 18 days with a tumor burden of approxi-

mately 1010 cells (91). Therefore, it would require a one log reduc-

tion in cell number (90% cell kill) to see a one day increase in life

span (ILS), and a two-log reduction in cell number (99% cell kill) to

see a 2-3 day ILS, while even with a 99.99% cell kill the animals would

still die in less than 20 days.

It is, therefore, not very surprising that leukemic mice treated

with parabactin, which requires constant exposure of cells for signifi-

cant periods of time to exhibit cytotoxic activity, on a single daily

injection schedule for 6 days, showed no increased life span over

untreated mice. However, leukemic mice treated with parabactin every

6 h for 5 injections did exhibit a reduction in tumor burden of 80%

over controls. Considering the doubling time of this tumor, even this

reduction in cell number would not extend the life of the mice to a

significant degree. A drug having an ILS of greater than 35% in this

system is considered to have significant activity by the National











75

Cancer Institute (92). Extended dosing on this schedule, while exhib-

iting promising activity, is prevented by the low solubility of the

catecholamides in most vehicles suitable for parental administration

(the solubility of parabactin in phosphate buffered saline being on the

order of 1.25 x 10-5 M) (37), and the toxicity of the vehicle alone on

frequent administration.

The ability of parabactin to block and synchronize cells has also

been demonstrated in vivo. Ascites L1210 cells were synchronized by a

single i.p. injection of parabactin (Fig. 2-20) in a manner similar to

that seen in culture. Attempts to demonstrate enhanced cytotoxic

effects of ara-C and Adriamycin on a parabactin synchronized in vivo

cell population resulted in no significant increase in life span of

L1210 mice over the additive effects of the individual treatments

(Table 2-4). Considering the optimum conditions studied with cultured

cells, which exhibited a maximum 2-3 fold potentiation of the cytotoxic

effects of the antineoplastics tested, very little increase in life

span would be expected in vivo by a reduction in tumor burden of

50-75%. Further refinements in the drug dosage and multiple dosing

schedules are currently under investigation.














CHAPTER III
POLYAMINE ANALOGUES

Materials and Methods

The polyamine analogues were synthesized by methods previously

described (30,31). Stock solutions (10 mM) of analogues and dilu-

tions were made in sterile water, and passed through a 0.2 um filter

prior to use. RPMI 1640, fetal bovine serum, HEPES, and MOPS were

obtained from Gibco (Grand Island, NY). Cell culture flasks, 25 and

75 cm2, were purchased from Corning (Corning, NY). Diamidino-

phenylindole (DAPI), propidium iodide (PI), actinomycin-D and pro-

teinase K were obtained from Sigma (St. Louis, MO). Rhodamine-123

(Rh-123) was purchased from Eastman Kodak Company (Rochester, NY).

RNase T1 was obtained from BRL Scientific (Gaithersburg, MD). DBA/2J

and C57B1/6J mice were obtained from Jackson Laboratories (Bar Har-

bor, ME). Albino CD-1 mice were obtained from Harlan Sprague-Dawley

(Indianapolis, IN). Sprague-Dawley rats were obtained from Charles

River (Wilmington, MA).

Cell Culture

All cell culture lines were grown as previously described in

Chapter II. Several additional cell lines utilized in these studies

were maintained in complete medium as before except where noted.

Chinese hamster lung (DC3F) cells were grown as monolayers in 25 cm2

flasks, seeded at 2 x 105 total cells and grown to confluency in

in approximately 96 h. Actinomycin-D resistant Chinese hamster lung

76









77

(DC3F/ADX) cells were similarly maintained with actinomycin-D 10

pg/mL present in the medium. The Chinese Hamster Lung cell lines,

DC3F and DC3F/ADX, were kindly donated by Dr. June Biedler of

Memorial Sloan-Kettering Cancer Center, NY.

Resistant L1210 cell Lines

L1210 cell lines resistant to DESPM (L1210/DES-10) and DEHSPM

(L1210/HDES-1) were selected out by maintaining the cells in increas-

ing concentrations of polyamine analogues over an extended period of

time. Initially, L1210 cells were exposed to either 1 pM DESPM or

0.1 pM DEHSPM for seven days and maintained between 1 x 105 and

1 x 106 cells/mL. Cells were then washed free of the drug and incu-

bated in fresh medium until cell doubling time returned to normal (12

h). These cloned cells were then again exposed to a higher concen-

tration of the analogue until normal cell doubling time resumed.

Analogue concentrations were gradually increased over a period of

approximately one month until cells could be maintained at a concen-

tration of 10 UM DESPM and 1 yM DEHSPM. Resistant cells maintained

at these concentrations of analogue for over six months had normal

L1210 morphology and doubling time.

An Adriamycin resistant L1210 cell line (L1210/DOX-0.6) was

similarly selected out by initially exposing cells to 0.01 UM Adria-

mycin. When approximate normal doubling time resumed, the Adriamycin

concentration was increased until over a period of three months the

cells could be maintained in the presence of 0.6 yM Adriamycin. The

L1210/DOX-0.6 cells maintained for over six months differed from the

normal L1210 cells in that they had irregular shapes and longer doub-

ling times of 14-16 h.











ICQ5 Determinations

The cells were treated while in logarithmic growth (L1210

cells, 3 x 104 cells/mL; Daudi and HL-60, 1 x 105 cells/mL) with the

polyamine derivatives diluted in sterile water and filtered through a

0.2 yM filter immediately prior to use. Following a 48 h incubation

with L1210 cells and a 72 h incubation with Daudi or HL-60 cells,

cells were reseeded (L1210 cells, 3 x 104 cells/mL; Daudi and HL-60

cells, 1 x 105 cells/mL) and incubated in the presence of the poly-

amine derivative for an additional 48 h or 72 h, respectively.

Determination of IC50 values for the most active compounds

(IC50 <5 pm at 96 h) was also performed in the presence of 1 mM

aminoguanidine, a serum diamine oxidase inhibitor (93).

Chinese hamster ovary (CHO) cells and murine B-16 melanoma

cells were seeded at 2 x 105 cells/25 cm2 flask and allowed to attach

for 4 h. At this time the cells were exposed to the polyamine ana-

logue for 48 h. Monolayer cells were not reseeded for additional

exposure times due to problems in plating efficiency after the ini-

tial 48 h treatment.

Cell samples were removed at the indicated time periods for

counting. Cell number was determined by electronic particle analysis

(Coulter Counter, Model ZB1, Coulter Electronics, Hialeah, FL) and

confirmed periodically with hemocytometer measurements.









79
The percentage of control growth was determined as follows:


% of control growth = 100 x [Final treated cell no.-
initial inoculum]/[Final untreated cell no.-initial inoculum]


The IC50 is defined as the concentration of compound necessary

to reduce cell growth to 50% of control growth after defined inter-

vals of exposure.

Analysis of Drug Effects on Mitochondrial DNA

The L1210 cells (in complete media at 1 x 105 cells/mL) were

incubated at 370C in the presence of the compound to be tested.

Every 24 h, for 144 h, cell samples were removed for counting, and

the cells were reseeded in fresh medium and drug at 1 x 105 cells/mL.

Cells were assayed daily for mitochondrial DNA (mtDNA) content.

Mitochondrial DNA assays were performed by R. Bortell and L. Raynor

of Dr. A. Neims' laboratory, Department of Pharmacology and Thera-

peutics, College of Medicine, University of Florida.

Because recovery of the organelles might vary with drug treat-

ment, a dot blot procedure for assay of mtDNA was developed which

involves analysis of cell lysates rather than preparations of mito-

chondria (94). Cells (5 x 104 to 2 x 105/mL) were lysed in 2% SDS in

the presence of proteinase K (5 mg) and RNase TI (100 units). The

total cell lysate was applied to nitrocellulose paper with use of a

96-well, vacuum operated dot-blot apparatus. The blots were hybrid-

ized to a 35S-labeled dATP nick translated probe made by inserting

full-length mouse mtDNA into pSP64 vector at the Sad site. Dot

blots were visualized by autoradiography and cut out, and radioactiv-

ity was determined by scintillation counting.










Flow Cytometric Analysis

Flow cytometric analysis of nuclear DNA (nDNA) content was

performed with the RATCOM flow cytometer (RATCOM Inc., Miami, FL)

interfaced with a microcomputer (IBM-XT). Cultured L1210 cells in

log-phase growth were incubated at 370C with various polyamine ana-

logues and samples removed at 0 h, 48 h, 96 h and 144 h were analyzed

for nDNA content distributions after staining with diamidinophenylin-

dole in a nuclear isolation media (NIM-DAPI) (75).

Alternately, cell analysis was performed with a FACS II flow

cytometer (Becton Dickinson FACS Systems, Sunnyvale, CA) interfaced

with a microcomputer (Hewlett Packard 45B, Fort Collins, CO).

Samples of 106 cells were removed and stained with propidium iodide

(PI) (Calbiochem, San Diego, CA) and then exposed to RNase. DNA

distributions were obtained from analysis of the red fluorescence

from the PI-stained DNA (76).

Cloning Assay

L1210 cells maintained between 105 and 106 cells/mL were incu-

bated with 10 uM DESPM for 96 h. At 24 h intervals treated cells

were washed (2 x 10 mL) and diluted in fresh complete media and

plated in triplicate 96-well microtiter plates at 0.4 cells/well with

each well containing 100 pL of sample. The plates were incubated at

370C in a humidified incubator in an atmosphere of 5% CO2 and 95%

air. The plates were examined with an inverse phase microscope at

100X magnification. The final number of colonies per plate was quan-

titated seven days after plating. Groups of 50 or more cells/well

were identified as having been cloned from a single viable cell.











Regrowth Studies

L1210 cells maintained between 105 and 106 cells/mL were incu-

bated with 10 pM DESPM, 30 yM DENSPM or DEHSPM 0.6 MI for 96 h. At

each 24 h interval, treated cell samples were washed and the cells

resuspended in fresh media at 1 x 105 cells/mL in duplicate 10 mL

flasks. The regrowth of the treated cells was followed for up to

144 h.

Cell Size

Cell size was determined directly by the method of Schwartz, et

al. (95). In brief, uniform polymeric microspheres ranging from

4.72-10.2 uM in diameter (Polysciences, Warrington, PA) were diluted

in Hematall (Fisher Scientific Co.). Electronic size was measured on

the FACS Analyzer (Becton Dickinson, Sunnyvale, CA) with the ampli-

fier in the log mode. The peak channel number for each size micro-

bead was plotted against the corresponding calibrated diameter and

calculated volume to obtain a calibration curve. L1210 cells were

treated with 10 yM DESPM, 30 UM DENSPM or DEHSPM 0.6 UM for 0-144 h

and samples of 106 cells were removed at 24 h intervals and pelleted.

The cells were resuspended in 0.5 mL of Hematall and analyzed. The

peak channel number of the treated cells was plotted on the

calibration curve to obtain the approximate cell size directly.

Viability Assay

The viability of cells which had been treated with 10 uM DESPM

or 0.3 UM DEHSPM and or 30 yM DENSPM maintained between 5 x 104 and









82
1.2 x 106 cells/mL was assayed at appropriate times using two dif-

ferent methods: trypan blue dye exclusion and Rh-123 mitochondrial

staining (74).

Briefly, 1 x 106 cells were pelleted by centrifugation and

resuspended in 1 mL of complete medium containing Rh-123 10 ug/mL,

incubated at 370C for 10 min, and washed once with medium. Trypan

blue was added prior to counting to give a final concentration of 0.2

ug/mL. Rh-123 and trypan blue stained cultures were observed in a

hemocytometer using a Zeiss epifluorescence Axioscop microscope at

630X magnification. The excitation wavelength was 485 nm. Viable

cells show mitochondrial specific uptake of the Rh-123 dye while

dying cells showed diffuse cytoplasmic staining and dead cells did

not fluoresce at all. Cells which did not fluoresce did not exclude

trypan blue. Both methods gave comparable results as to cell death,

however, trypan blue exclusion does not reflect dying cells.

Animal Studies

The polyamine analogue was diluted in sterile 0.9% saline

within 24 h of use and the unused portion stored at 50C. Concentra-

tions of the spermine analogue tetrahydrochlorides, at each dose were

adjusted so that the mice were injected with a volume of 1 mL/100 g

(i.e. a 25 g mouse was injected with 0.25 mL of drug solution).

Untreated mice served as controls.

The compounds were administered by i.p. injection according to

the appropriate dosing schedule. The murine L1210 leukemia cells

were maintained in DBA/2J mice. Cells from a single mouse which

was injected i.p. with 1.25 x 104 cells seven days earlier were









83
harvested and diluted with cold saline so that an inoculum of 105 or

106 cells could be administered by a 0.25 mL i.p. injection. In

each study mice were injected i.p. with the appropriate number of

L1210 cells on day 0.

Alternately, Alzet mini-osmotic pumps, model 2001 (Alza, Palo

Alto, CA) designed to deliver 1 yL/h for 7 days, were filled with

DEHSPM so that the pump would deliver 20 mg/kg/day. The filled pumps

were placed in sterile saline 0.9% and incubated at 370C for 8 h

prior to placement in the mice. The pumps were implanted, on day 1,

subcutaneously (s.q.) in the dorsal side of the mice through an

incision behind the right flank, and the incision closed with a wound

clip. After 6 days the pumps were removed and the wound closed. The

parameter used for treatment evaluation of activity against L1210

leukemia was mean survival time (percent increased life span, % ILS).


%ILS = 100 x [mean survival time treated animals-
mean survival time controls]/[mean survival time controls]


The murine B-16 melanoma was maintained in C57B1/6J mice. The

solid tumor was excised from a single mouse, which was injected sub-

cutaneously (s.q.) 14 days earlier, and minced with 9 volumes of

Hanks' balanced salt solution (10:1 brei). The brei (0.1 mL) was

administered by s.q. injection in the right flank of the animal on

day 0.

The murine Lewis lung carcinoma was maintained in C57B1/6J

mice. The solid tumor was excised from a single mouse which was

injected s.q. 14 days earlier, and the tumor divided into equal









84
fragments 2-4 mm in size. Each mouse was injected s.q. with a frag-

ment in the sternum region using a 14 g trochar needle on day 0.

The parameter measured for treatment evaluation was median

tumor weight change based on length and width measurements in milli-

meters (92). The tumor weights (mg) were calculated from tumor

dimensions (mm x mm) following the formula for volume of a prolate

ellipsoid:


L x W2 where L is the longer of the two
2 measurements recorded


Median tumor weights were calculated for test (T) and control (C)

groups on days 10, 14 and 17 after inoculation of the tumor.

The activity of the analogue was reported as T/C% on the

measurement day giving the lowest value.

Rates of Uptake of Polyamine Analogues in L1210 Cells

L1210 cells maintained in mid-log phase growth were exposed to

the polyamine analogues for 0, 2, 6, 12 and 24 h, and samples of 106

cells were removed. The cells were washed with fresh medium (2 x 10

mL) and centrifuged to pellet the cells. The supernatant was care-

fully removed and the cells resuspended in 0.6 N perchloric acid

(107 cells/mL). Cells were freeze-fractured by four successive

freeze-thawings using liquid nitrogen. Cell lysates were stored at

-20C until HPLC analysis. For HPLC analysis an aliquot of the solu-

bilized polyamine perchlorates was reacted with dansyl chloride, sod-

ium carbonate and diaminohexane as an internal standard. After con-

centration of the reaction products on a C18 chromatographic plug,

the dansylated polyamines were analyzed on reversed phase analytical









85
C18 column and quantitated using fluorescence detection and elec-

tronic integration. Analysis of HPLC samples were performed by

J.R.T. Vinson, B. Jennings, and V. Andaloro of Dr. R.J. Bergeron's

laboratory, Department of Medicinal Chemistry, College of Pharmacy,

University of Florida.

Pharmacokinetic Studies

Preliminary pharmacokinetic studies to determine half-life of

the analogues in the serum were studied in mice. Blood samples were

collected by exsanguination of CD-1 mice into nonheparinized tubes at

appropriate times after a single i.p. injection of DESPM (25 mg/kg).

The serum was separated from the cells by centrifugation and diluted

with an equal volume of 1.2 N perchloric acid. Analogue levels in

the serum were determined by HPLC analysis by the above method.

The excretion of DEHSPM (25 mg/kg) after a single i.p. injec-

tion was followed in the urine and bile of Sprague-Dawley rats util-

izing bile duct cannulation model. Rat bile duct cannulations were

performed by K. Crist and E. LaGraves of Dr. R.J. Bergeron's labora-

tory, Department of Medicinal Chemistry, College of Pharmacy, Uni-

versity of Florida. The bile duct was cannulated with a segment of

22 ga polyethylene tubing and the bile directed from the rat to frac-

tion collector by a fluid swivel mounted above a metabolic cage.

Normal urine flow was also diverted with a segment of 5 mm tubing

from the bottom of the metabolic cage to the fraction collector.

This system allows the animal to move freely in the cage while

continuous bile and urine samples were being collected in 3 h frac-

tions for 24 h. Sample volumes were measured and diluted with an













equal volume of 1.2 N perchloric acid. Analogue levels in each

fraction were determined by HPLC analysis by the above method.

Toxicity Studies

Both single dose and five-day daily-dose toxicity in CD-1 white

mice were determined for DESPM and DEHSPM. Initially, small groups

of 2-3 mice were used to determine the approximate range of toxicity

for the compound. More accurate determinations of toxicity were made

using groups of greater than 6 mice at each dose of analogue. Dos-

ages ranged so that at the lowest dosage 10% or less of the animals

die and at the highest dosage >90% of the animals die. At least 3

toxicity dosages between these two extremes were determined.

The polyamine analogue was diluted in sterile 0.9% saline with-

in 24 h of use and the unused portion stored at 50C. Concentrations

of the spermine analogue tetrahydrochlorides at each dose were

adjusted so that the mice were injected i.p. with a volume of 1

mL/100 g.

For the single dose toxicity studies each group of mice

received a single i.p. injection of the analogue at an appropriate

dosage. For the 5-day daily-dose toxicity studies each group of mice

received a single i.p. injection of the analogue at an appropriate

dosage daily for 5 days. Animal deaths were checked daily for 30

days. The number of deaths was expressed as percent lethality

[100 x (number of deaths per group in 30 days/number of mice per

group)].












Results
In Vitro Activity
The polyamines which were evaluated for their antiproliferative
activity include the spermidine analogues (Fig. 3-1);
a. N,N -bis(ethyl)-1,8-diaminooctane (BEDAO),
b. N1,N8-bis(propyl)spermidine (BPSPD),
c. N1,N8-bis(ethyl)-N4-(p-nitrophenyl)spermidine (BEPNPSPD),
d. N1,N7-bis(ethyl)norspermidine (BENSPD),
e. N1,N8-bis(ethyl)spermidine (BESPD),
and the spermine analogues (Fig. 3-2);

f. N,N'-diethyl-1,12-diaminododecane (DEDAD),
g. N1,N12-dibenzylspermine (DBSPM),
h. N1,N12-di(2,2,2-trifluoroethyl)spermine (FDESPM),
i. 1,4-bis-[3-(ethylamino)propyl]piperazine (DEPIP),

j. N4,N9-diethylspermine (IDESPM),
k. N1,NI,N12,rl2-tetraethylspermine (TESPM),
and (Fig. 3-3)
1. N1,Nl1-diethylnorspermine (DENSPM),
m. N1,N4,N9,N12-tetraethylspermine (NTESPM),
n. N1,N12-dimethylspermine (DMSPM),
o. 1,20-bis(N-ethylamino)-4,8,13,17-tetraazaeicosane (YANK),
p. N1,N12-dipropylspermine (DPSPM),

q. N1,N12-diethylspermine (DESPM),
r. N1,N2-diethylhomospermine (DEHSPM).
The compounds described in this study are all in some way
related to the natural products spermidine and spermine. They were
















48 h


H
N'N N
BEDAO H

H
N N N
H H
BPSPD

NO2

P H
N BEPN NPSPD
H BEPNPSPD


>100 UM >100 M




> 30 pM > 30 un1







> 10 JM > 10 uM


N >N N
H H H >100 uM
BENSPD
H
N N N 50 pM
H H
BESPD

Figure 3-1. Structure and IC50 values of the spermidine
analogues.


10 uM




10 M1


96 h













48 h


DEDAD H


SH H
N N N N N*

DBSPM

H H
F3C N N N N CF3

FDESPM H H


H H
N NN N_ NJN


Cytotoxic > 25 1M
No inhibition < 12.5 UM




Cytotoxic > 10 UM
No inhibition < 3 uM


>100 JM


>100 PM


> 50 uM > 50 uM


DEPIP


H2 N NH2

IDESPM


H
N N N N
TESPM H


Figure 3-2. Structure and IC50 values of
spermine analogues.


> 25 yM > 25 UM


> 50 uM


5 pM


the less active


96 h











96 h


H H H H
N N N N -
DENSPM

H
N N N
NTESPM H

H H
/N~NNN N/
.00, N N N N
DMSPM H H

H H H
, N ... N N ` N ^' N N
YANK H H H


H H
N N N N
DPSPM H H

H H
N N N N
DESPM H H

H H
NH D M H N
H DEHSPM H


>100 uM 4 yM




100 UM 3 JM




>100 UM 0.75 uM




50 yM 0.5 pM





3 pM 0.2 UM




10 PM 0.1 pM




0.5 uM 0.05 uM


Figure 3-3. Structure and IC50 values of the most active
spermine analogues.


48 h









91
designed to evaluate relationships which might exist between the
structure of various alkylated polaymines and their activity against
tumor cells in culture. The relative antiproliferative activities of
the analogues were assessed by determining their 48-h and 96-h IC50

values. In comparing the spermine compounds with their spermidine
counterparts (Figs. 3-1 to 3-3), it is clear that the spermine com-
pounds are substantially more active against L1210 cells.

Of the spermine analogues tested, DESPM and DEHSPM were the
most effective inhibitors of growth of L1210 cells. DESPM had 48 h

and 96 h IC50 values of 10 yM and 0.1 uM, respectively. DEHSPM

demonstrated superior activity and had 48 h and 96 h IC50 values of
0.5 UM and 0.05 UM, respectively. The IC50 values for the most

active analogues (IC50 value <5 uM at 96 h) determined in the
presence of 100 uM aminoguanidine, an inhibitor of diamine oxidase,
remained the same.

The activity of DESPM against Daudi cells and HL-60 cells is
comparable to that against L1210 cells. Because of the differences
in doubling times between L1210 cells (ca. 10-12 h) and Daudi or
HL-60 (ca. 25 h), the IC50s of DESPM at 48 h and 96 h against L1210

cells are best compared with the IC50s at 72 h and 144 h against

HL-60 (10 yM and 0.3 pM) and Daudi cells (>40 pM and 0.5 UM). Addi-

tionally, DEHSPM was shown to be active against CHO and B-16 melanoma
cells (doubling time for both lines 14-16 h) with IC50 values of
about 0.1 yM at 48 h for both cell lines.

L1210 cells which were selected for their ability to grow
in the presence of Adriamycin (L1210/DOX-0.6 cells) demonstrated

an approximate 50-fold increase in resistance to Adriamycin over the
parental L1210 cell line. The L1210/DOX-0.6 cells were demonstrated

by immunoblotting to have a large increase in the membrane











92

glycoproteins (P150-170 glycoproteins) and considered to be multidrug

resistant cells (96). Both the parental cell line, as well as the

resistant cell line, were colaterally sensitive to DEHSPM with IC50

values of 0.5 pM and 0.05 pM at 48 h and 96 h respectively.

The antiproliferative activity of DEHSPM was tested against

DC3F cells and a proven multidrug-resistant subline, DC3F/ADX cells

(97-99). Both cell lines were determined to be equally sensitive to

DEHSPM having identical 96-h IC50 values of 0.1 pM.

L1210 cells which were developed for their resistance to the

polyamine analogues L1210/DES-10 and L1210/HDES-1 did not demonstrate

an overproduction of the P150-170 glycoproteins (96), indicating that

they had not developed multidrug resistance; both were sensitive to

Adriamycin to the same extent (IC50 value of 0.02 yM at 48 h). Both

cell lines did, however, exhibit cross-resistance to other polyamine

analogues. Both of the resistant cell lines were not inhibited by

treatment with 100 uM DENSPM, and were able to grow to 70% of control

in the presence of either 100 yM DESPM or 100 uM DEHSPM.

The role of the methylene backbone in the biological activity

of the polyamine analogues was studied by comparing DENSPM, DESPM and

DEHSPM. Reduction in the length of the polyamine backbone, decreased

the activity of the analogue. The activity of DENSPM against L1210

cells (IC50 value of >100 pM at 48 h, and 4 yM at 96 h) is relatively

poor when compared to that of DESPM and DEHSPM (Figs. 3-4 and 3-5).

















1000




100



10--






1 I
zI









0 10 20 30 40 50 60 70 80 90 100110
z



z
0
C .1


DENSPM
.01 -- DESPM
S DEHSPM


.001
0 10 20 30 40 50 60 70 80 90 100110

% CONTROL GROWTH
Figure 3-4. Concentration dependence of L1210 cell growth
inhibition by DENSPM, DESPM and DEHSPM at 48 h.
















1000 I



100 -- A DENSPM
--- DESPM
DEHSPM

S 10
z
0


z
UA
z
o .1
0


.01



.001 I
0 10 20 30 40 50 60 70 80 90 100110

% CONTROL GROWTH

Figure 3-5. Concentration dependence of L1210 cell growth
inhibition by DENSPM, DESPM and DEHSPM at 96 h.