Effects of polyamine analogs on mitochondrial DNA and cell growth


Material Information

Effects of polyamine analogs on mitochondrial DNA and cell growth
Physical Description:
vii, 157 leaves : ill. ; 29 cm.
Bortell, Rita Bellis, 1953-
Publication Date:


Subjects / Keywords:
DNA, Mitochondrial -- drug effects   ( mesh )
Polyamines   ( mesh )
Cell Division   ( mesh )
Pharmacology and Therapeutics thesis Ph.D   ( mesh )
Dissertations, Academic -- Pharmacology and Therapeutics -- UF   ( mesh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph.D.)--University of Florida, 1987.
Bibliography: leaves 145-156.
Statement of Responsibility:
by Rita Bellis Bortell.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000983932
oclc - 20379902
notis - AEW0099
sobekcm - AA00006106_00001
System ID:

Full Text







This dissertation is dedicated to David Alan Bortell, for the support and
encouragement he gave me during this endeavor.

I would like to thank my supervisor and mentor, Dr. Allen Neims, for his
careful guidance of my graduate training. I also want to thank the members of

my committee: Drs. Stephen Baker, William Hauswirth, Margaret James,
Warren Ross, and Thomas Rowe. I would like to extend my appreciation to the
fellow members of our laboratory: Lynn Raynor, Gurmit Singh, Lori Lim, Debra

Stinson, Sukanya Kanthawatana, Daniel Danso, and Mary Anne Kelly. I
particularly want to thank Mike Ingeno, Bonnie O'Brien, David Bortell, Debra,
and Lynn for the generous help they gave me. Special thanks go to Shari
McArdle, with whom I most closely shared the graduate school experience. A
final debt of gratitude goes to Dr. Allen Neims and Lynn Raynor, who greatly
enriched my training and made the experience fun.


ACKNOWLEDGMENTS ........................... iii

ABSTRACT ..... ................................ vi

INTRODUCTION .... ....... ........................ 1

LITERATURE REVIEW ................................
Mitochondrial DNA (MtDNA).......................... 3
Nuclear/Mitochondrial Regulation ................. ....... 9
Polyamines... ................. ................ 24
MGBG (Methylglyoxal bis[guanylhydrazone]) ............. ... 31
Specific Aims .... .................. ............. 43

MATERIALS AND METHODS ...........................
Cell Culture ........................ ............ 45
Cell Counts .... ................ .............. 45
Determination of IC50 .................... ......... 46
Cell Viability ........................... ....... 46
Clonogenic Assay .................... ............ 46
Cell Size ..................................... 47
Flow Cytometry ............................ ........... 47
Rhodamine 123 Uptake .............................. 47
[14C]-MGBG Stock Solution Preparation . ..... 48
Uptake Protocol ................................. 48
Cloning of MtDNA Probe ................. ............ 48
Nick Translation of MtDNA Probe ................ ...... 50
Gel Electrophoresis .............................. 50
Southern Blot ................ .... .... ..... .. 51
DotBlot...... .............. ......... .........51
Hybridization with [35S]-labeled MtDNA Probe . .... 52

Characterization of L1210 Cells....................... 53
Effects of Polyamine Analogs on Cell Division and MtDNA
Accum ulation ..... ..... ... .. .. .. .. ..... 59
Resistance to MGBG and Recovery MtDNA Synthesis and Copy
Num ber . . . . 84
Effects of MtDNA Content on Mitochondrial and Cell Functions 11
UptakeofMGBG. .................................. 113

DISCUSSIO N .. . ... .... .. ... .. .. ..
Determination of the MtDNA Copy Number in L1210 Cells ....... .123
Decrease in Cell Growth and MtDNA/Cell with MGBG and DES
Exposure . . . . 124
Mitochondrial Function During MGBG Exposure . .... 127

Increased MtDNA/Cell with Low Concentrations of MGBG ...... .129
Recovery of MtDNA Replication . . ... 130
Implications for Nuclear/Mitochondrial Interactions . ... 132

SUMMARY ...................... ............ 141

REFERENCES ................... .............. 145

BIOGRAPHICAL SKETCH.................. ............. 157

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



Rita Bellis Bortell

December 1987

Chairman: Allen H. Neims
Major Department: Pharmacology and Therapeutics
The average mouse leukemia L1210 cell contains 1450 copies of
mitochondrial DNA (mtDNA). Somehow this number is maintained during
exponential growth. We have studied the effects of the polyamine analogs,
methylglyoxal bis[guanylhydrazone] (MGBG) and diethylspermine (DES), in
order to explore this relationship because appropriate doses of the drugs

selectively inhibit mtDNA replication without causing a decrease in cell
viability, at least in the case of MGBG.
Exponentially-growing L1210 cells were exposed to 1.0-3.3 jiM MGBG
for 1-31 days. Cells were collected on various days for dot blot quantitation of
mtDNA by hybridization of cell lysate to a [35S]-labeled, full-length mouse
mtDNA probe. At >1.3 jiM MGBG, mtDNA replication was completely inhibited
for variable times depending on dose (64 hr at 3.3 PjM). Since cell division
continued, albeit somewhat slower, the mtDNA content per cell was diluted
with each cell division. Whereas cell size decreased when the mtDNA copy

number per cell was low, the potential-dependent uptake of rhodamine 123

uptake into mitochondria was not affected by MGBG treatment.

Diethylspermine had similar initial effects on mtDNA accumulation and

cell growth. With exposure to either drug, the amount of mtDNA in viable cells
never decreased below 10% of the control value because cell division

ceased. Although the effects of DES on mtDNA accumulation were not

reversible, mtDNA synthesis of MGBG-exposed cells resumed after 3 days.

When mtDNA synthesis resumed, the mtDNA content per cell doubled every
12 hours whether or not drug had been removed from the medium. The fact

that this mtDNA synthesis occurred when the rate of cell division was at its

lowest resulted in repletion of the cellular content of mtDNA. When the mtDNA

was nearly repleted, the rate of cell division recovered to the control value of

12 hours. The observation that the doubling of mtDNA in the virtual absence

of cell division required the same 12 hours suggests that replication of mtDNA
may proceed at its maximal rate in untreated exponentially-growing L1210

cells and actually limit how rapidly these cells can divide.

The mechanism of recovery of mtDNA synthesis in these cells was not

due to genetic mutation, but rather to a phenotypic adaptation. The L1210
cells initially concentrated [14C]-MGBG more than 1000-fold relative to the

medium. When mtDNA synthesis resumed, the apparent intracellular

concentration of the drug had begun to decrease to a new steady state value,

although the concentration was still higher than that earlier associated with
inhibition. The mtDNA copy number of cells continuously exposed to MGBG
had returned to the control value at 31 days.


The number of copies of mitochondrial DNA (mtDNA) seems to be
characteristic for a given cell type, whether that cell is dividing or not (1, 2). It
can be appreciated that maintenance of the characteristic copy number in
dividing cells is especially complex. For this amount of mtDNA per cell to be
maintained, the mtDNA must double with each cell division. Because the
nucleus and mitochondria are distinct organelles, it is not understood how this

regulation occurs.

This dissertation focuses on the basic question of how the characteristic
copy number of mitochondrial DNA (mtDNA) is maintained in dividing cells.

My basic approach is to perturb rapidly dividing cells with a drug or compound
which selectively inhibits mtDNA replication and use dot blot analysis and

measurements of cell number to follow the resultant effects on the copy
number of mtDNA and on cell division.

The drug I chose to use was methylglyoxal bis(guanylhydrazone), or
MGBG. This drug is cationic, metabolically stable, and was developed as a
polyamine congener (3). It has been used clinically as an antiproliferative
agent and to treat trypanosomiasis (4, 5). I selected this drug because it
displayed two important characteristics. First, it inhibits mtDNA replication
without damaging the mtDNA at concentrations which initially have only a
minimal effect on nuclear DNA replication (6, 7). Second, preliminary studies
in our laboratory suggested that long-term treatment (up to 31 days) of L1210
cells with MGBG did not substantially affect cell viability. Consequently, this


drug could be used to study the cell's response, over time, to differential

perturbation of mtDNA and cell replication.


The literature review will be divided into four sections. They are 1) an
overview of the literature on mtDNA; 2) a more detailed review of the
interaction between the nuclear and mitochondrial genetic systems; 3) a brief
survey of polyamine research; and 4) a synopsis of the literature on MGBG
and related polyamine analogs. This chapter will conclude with a summary of
the specific goals of my research.

Mitochondrial DNA
Mitochondria are the organelles of eukaryotic cells which are
responsible for oxidative phosphorylation, among other functions. According
to the endosymbiotic theory of the evolutionary origin of the eukaryotic cell, the
mitochondrion evolved from a free-living prokaryote, probably the purple
nonsulfur bacteria (Rhodospirillaceae) (8), which was incorporated into an
ancestral eukaryotic cell (9). Prior to this, the protoeukaryote had depended

on fermentation for the production of ATP. In time the prokaryote suffered a
progressive loss of autonomy: its rate of proliferation became linked with that
of the host cell, many of its biosynthetic capabilities were lost or taken over by
the host, and its metabolic functions became integrated with that of the host.
This theory implies that the inner mitochondrial membrane and its
invaginations (the cristae) are homologous with the plasma membrane of
present-day bacteria, whereas the outer mitochondrial membrane is derived
from the protoeukaryote.

Mitochondria are self-replicating, arising only by growth and division of

preexisting mitochondria (10). The existence of DNA within mitochondria was
first observed by Nass and Nass in 1963 (11). Mitochondrial DNA is a double-
stranded circular genome (Figure 1); exceptions are the linear mtDNA
molecules of the ciliated protests Tetrahymena and Paramecium (10). The
genome's two strands contain different proportions of purines, and thus can be
separated by density gradient centrifugation into heavy strand (H-strand) and
light strand (L-strand). Immediately after replication of mtDNA, the two
daughter molecules exist in a relaxed circular state. Within one hour under
normal conditions, however, a mitochondrial topoisomerase introduces
approximately 100 supercoils into each molecule (12).
This supercoiled form may subsequently be converted into a

displacement loop (D-loop) form which is unique to the mitochondrial genome.
This triplex D-loop structure is formed by the synthesis of a short daughter H-
strand that remains stably associated with the parental closed circle; the
parental H-strand is consequently displaced as a single strand. The D-loop
region varies in length between different species, and sometimes within the
same animal (13). In general, the D-loop is approximately 1000 base pairs

(bp) in length (14).
Mitochondrial DNA of human, murine, and bovine species has been

completely sequenced and consists of approximately 16,000 bp (14). In
addition, these sequences have been completely mapped for restriction sites
(15) and locations of promoters for replication and transcription of both heavy
and light strands of mtDNA (16, 17). The coding regions of the genome have
been mapped, and the majority of the proteins (12 of 13) and transfer RNAs
(tRNAs) (16 of 22), and both ribosomal RNAs (rRNAs) are coded for by the H-
strand (Figure 1). The only noncoding areas of mammalian mtDNA are a short

A\V." "

"\ .l

Figure 1. Schematic of human mtDNA. Hatched segments show the extent of
genes for known proteins, filled segments show the genes for
transfer RNA's, and crossed segments show the genes for
ribosomal RNA's.

sequence outside the D-loop region which encompasses the origin of L-strand

replication, and the D-loop region which contains the promoter sites for H-

strand replication and both H- and L-strand transcription (14).

There are two current models for mitochondrial transcription in

mammalian cells. One model proposes that transcription begins near the

origin of replication and continues as a single polycistronic transcript around

the entire genome, but with a high probability of premature termination

immediately following the 12S-16S region (18). The alternative model

suggests that there are two promoters for the initiation of RNA synthesis on the

H-strand near the origin of replication (19, 20). Transcription from one site

circumscribes the entire mitochondrial genome, while transcription from the

other site terminates after the two rRNA cistrons which are immediately

downstream. Once transcription has occurred, the polycistronic precursors

are cleaved into mature messenger RNAs (mRNAs), probably before the

genome is completely transcribed (21). Since there are very few nucleotides

between adjacent genes on mammalian mtDNA (and sometimes none), the

signals for RNA processing may not be specific sequences er se (22). Most

of the mRNAs and both rRNAs are flanked by tRNA genes, and the secondary

structure of these tRNAs in the transcript may provide the signals for cleavage

by the RNA processing enzyme (23).

The mitochondrial genome is thought to be associated with the inner

membrane at its matrix face (24). Molecules of mtDNA are frequently seen to

be attached to fragments of membrane or proteinaceous material after gentle

lysis of the organelles (9). In some cases there is good evidence that such

associations are not artifacts; for example, in one study it was shown that

membrane fragments were always attached to approximately the same

position on HeLa cell mtDNA (10). Thus, it is probable that mtDNA molecules

are attached to organelle membranes at least some of the time.

Apart from this attachment, isolated organelle genomes generally are

not completed with histones or other proteins in the manner of true

chromosomes in the nucleus; exceptions are the highly condensed mtDNA

molecules of the slime molds which appear to be completed with a basic

protein (25), and the mtDNA of Xenopus oocytes, which can be isolated in

association with protein in the form of structures resembling the nucleosomes

of chromatin (10). Also, recent evidence suggests that the displaced single

strands (D-loop) of mtDNA, but not the double-stranded segments of the

genome, are coated with DNA-binding proteins apparently unique to mtDNA

(26, 27). These proteins are thought to play a critical role in maintaining the

integrity of these triplex replication loops.

In mammals, the mitochondrial genome codes for 13 proteins and

some, but not all, of the components necessary for protein synthesis; this

includes two rRNAs and 22 tRNAs (28). The known human mtDNA-coded

proteins include 3 subunits of cytochrome oxidase, one of ATPase, 8 of NADH

dehydrogenase, and one of cytochrome b (29, 30). Antibodies raised to

synthetic peptides predicted from the DNA sequence have been used to

establish the existence of translated products from all previously unassigned

reading frames (URFs) (31). Attardi and his coworkers (32) recently identified

the last URF of human mtDNA as a subunit of NADH dehydrogenase. All of

the mtDNA-coded proteins are integral components of the inner mitochondrial

membrane and play key roles in oxidative phosphorylation. It follows,

therefore, that the mitochondrial genetic system is indispensable for the

biogenesis of the aerobic energy-generating system of all eukaryotic cells.

Whereas mtDNA codes for 13 mitochondrial polypeptides, the
organelle contains over 270 different proteins (29, 33). The remaining

polypeptides, which account for approximately 90% of total mitochondrial
protein, are encoded in nuclear DNA and synthesized on cytoplasmic

ribosomes. Available evidence indicates that each of these proteins contains

a leader sequence enriched in basic (positively-charged) amino acids which

directs the protein to the mitochondria, possibly to the point of adherence

between the inner and outer membranes (34). The protein is then imported
into mitochondria via a set of processes dependent on the existence of a

mitochondrial membrane potential (35).

If the majority of mitochondrial proteins are imported from the

cytoplasm, then why do cells expend more than a hundred nuclear genes just

to allow the mitochondrial genetic system to transcribe and translate about a

dozen polypeptides in situ? Why, in fact, does the mitochondrial genome

retain the genes for any of the particular proteins that it has? It has been

suggested that mitochondrial translation products are so hydrophobic that they

are best made in situ (36). However, other hydrophobic proteins are

successfully transported in the cell (37), and some mitochondrially coded

proteins are as hydrophilic as some soluble proteins (38).

Alternatively, it may be that all eukaryotes are in the evolutionary

process of transferring mitochondrial genes to the nucleus. The fact that the

more highly evolved eukaryotes have the smallest mitochondrial genome is

consistent with this idea. The mammalian mitochondrial genome has about
16,000 kb; yeast mtDNA is 5 times longer largely because 1) it has more
genes than mammalian mtDNA, 2) several genes have introns, and 3) it

contains long noncoding stretches between genes (14). This finding suggests

that there has been a dynamic net flux of mitochondrial genetic material into

the nucleus (39). Indeed, firm evidence has emerged that mtDNA can be

stably integrated into nuclear genomes (40). Rearranged parts of the

mitochondrial genes var 1, cob/box and ori/rep have been found in the nuclear

genome of yeast (41). Homologues of mtRNA genes occur in locust nuclei

(42) and mtDNA sequences have been reported in sea urchin (43) and rat

nuclear DNA (44). In addition, transposition of mtDNA into nuclear sequences

may be more than just an evolutionary phenomenon, as this process has been

reported to occur as a normal part of senescence in Podospora (45). Recent

evidence, however, suggests that this latter observation may be artifactual

(46). In any event, it seems increasingly likely that normal or abnormal

mtDNA, released into the cytosol during mitochondrial breakdown or the
normal organelle fusion/segregation cycles, could become incorporated into

nuclear DNA (or vice versa).

However, a possible limitation in the transfer of functional genes from

the mitochondrion to the nucleus is the slightly different genetic code used by

the two systems. One of the triplicate codons in mitochondria codes for

tryptophan, but is recognized as a stop codon in the nucleus (47). The 13

critical protein-encoding genes that remain in the mammalian mitochondia

have such a codon (48). This would make further transfer of usable genes to

the nucleus very improbable.
Nuclear/Mitochondrial Regulation

Regulation of Mitochondrial Enzymes in Differentiated Cells
Because most of the mitochondrial proteins are encoded in nuclear

genes and synthesized on cytoplasmic ribosomes, interactions between the
nucleus and mitochondria must play an important role in mitochondrial

regulation. Respiratory enzymes and mitochondrial membranes are normally

synthesized and/or degraded in a coordinated fashion as the proportions of

the various components are characteristic for a given cell (49). The content of

mitochondrial enzymes in mammalian cells is regulated by several

environmental conditions including oxygen tension and hormones (50, 51,

52). One major question is whether the mitochondrial proteins, or even just

those of the inner membrane, are regulated individually or as a set. Because

various mitochondrial components have distinct turnover times and different

degrees of nuclear and mitochondrial control including specific genomic

expressions, individual proteins appear to be individually regulated (53, 54).

But, as discussed below, there are exceptions to this general rule.

During hypoxic incubation of mouse lung macrophages and rat skeletal

muscle L8 cells, the activities of various mitochondrial enzymes (both nuclear

and mitochondrially coded) decreased about 50-60% in tandem with no

statistical difference between the various activities (55). The kinetics of this

decrease and of the recovery of activity after reexposure to normoxia were

also similar for the different enzyme activities. This suggests that, at least in

certain cases, mitochondrial enzymes can be regulated as a unit.

It is known that nuclear gene expression is frequently regulated at the

transcriptional level (56); however, the levels) at which mtDNA gene

expression is controlled is unknown. Some reports have addressed this

question from the point of view of gene dose or mtDNA copy number, i.e. is the

increase or decrease in mtDNA gene expression proportional to changes in

the content of mtDNA? In the experiment described above, although there

were uniform decreases in mitochondrial enzyme activities after exposure to

hypoxia, there was no change in the mtDNA copy number, a result that speaks

against regulation of mitochondrial gene expression only by gene dosage.

Another experiment which sought to answer whether mtDNA gene
dosage controls gene expression was conducted by Williams (57). He
determined the concentrations of mtDNA, mitochondrial rRNA, and
cytochrome b mRNA (a mitochondrial gene product) in normal versus
electrically-stimulated rabbit striated muscles. This tissue was chosen
because the oxidative capacity of mammalian striated muscles can vary nearly
10-fold, a finding which presumably reflects major differences in the

expression of genes (including mtDNA) that encode enzymes of oxidative
metabolism (58). After electrical stimulation of the muscle, mtDNA, mtRNA,
and the mitochondrially coded gene product cytochromee b) were observed to
vary in direct proportion to the oxidative capacity of the tissue. Because the
expression of mitochondrial genes in mammalian striated muscle was

proportional to their copy number, these results do support the hypothesis that
amplification of the mitochondrial genome relative to chromosomal DNA, i.e.

gene dosage, was responsible for the enhanced expression of mitochondrial
genes in highly oxidative tissues (57).

This interpretation, of course, differs from that of the hypoxia experiment
described above. The different conclusions reached in the two experiments

may relate to the different tissues and/or animals used. Also, the differences
may reflect the different experimental designs; e.g., there were no changes in
mtDNA copy number after 4 days of hypoxia, but neither were there changes
in mtDNA content after the same period of muscle stimulation (the changes
occurred after 7 days). In any event, the regulation of mitochondrial gene
expression in differentiated cells, to date, is not well understood.
Reoguation of Mitochondrial Enzymes in Transformed Cells
The regulation of mitochondrial gene expression in transformed
mammalian cells is also not well understood (59). The pioneering manometric

studies of Warburg in the 1920s (60) first revealed that tumor cells are

defective in respiration and have abnormally high rates of aerobic glycolysis.

Warburg suggested that the cancer cell originates from heritable injury to the

mitochondria in that the respiratory system of all cancer cells was thought to

be damaged (61, 62). By 1960, this hypothesis had lost much of its impact

because elevated rates of glycolysis could not be associated in any consistent

manner with changes in the structure or function of the mitochondrion (63, 64).

Nevertheless, numerous publications in the last two decades have

catalogued a wide range of abnormalities in tumor cell mitochondria. For

example, tumor cells apparently contain fewer mitochondria than normal cells

and the organelles in the tumor cells may be structurally aberrant (65, 66).

Many deficiencies of energy-linked functions have also been reported, and

they include reduced respiratory control (ADP-stumulated state 3 respiration)

(67). Inner membrane alterations are also seen in tumor mitochondria, such

as changes in the amount and properties of ATPase (68). The levels of

mitochondrial-associated enzymes such as hexokinase are also different in

tumor cells (69, 70). It is possible that these complex multiple changes in

membrane characteristics result from rearrangements of mtDNA (71) or from

topological aberrations of the organellar genome (72). Indeed, aberrations of

mtDNA (mainly associated with an increase in the percentage of dimers and

catenanes) have been found in tumor cells (73).

Relationship Between Mitochondrial Protein and Gene Dosage

Almost all eukaryotic cells contain many molecules of mtDNA (10). The

exceptions are a few fungi and more than a thousand species of protozoa that
have no mitochondria (46). Moreover, under given metabolic conditions,

different cells appear to have a characteristic mitochondrial and mtDNA

content (1). How is this regulated? There have been attempts to relate

mtDNA content to the number of mitochondria or the mass of mitochondrial

protein isolated. The studies reviewed by Nass (74) and by Borst and Kroon

(75) indicate a yield of 0.2 to 1.8 upg of mtDNA per mg of mitochondrial protein,

or a content of roughly 2 to 10 mtDNA molecules per organelle, depending

upon the animal cell or tissue examined. Thus, multiple mitochondrial

genomes are packaged into each mitochondrion, and there are as few as one

mitochondrion to more than several thousand mitochondria per cell,

depending on the type.

However, these older studies were somewhat suspect because it was

difficult to selectively assay mtDNA, given the need to first separate it from a

several hundred-fold excess of nuclear DNA. Consequently, Bogenhagen

and Clayton (1) studied the relationship of mtDNA content to mitochondrial

volume in cells lacking the ability to incorporate exogenous thymidine into

nuclear DNA (because they had lost the major cellular thymidine kinase, TK-).

Such cell lines retain a mitochondrial-specific activity which allows mtDNA

labeling with exogenous radioactive thymidine. Stereological analysis (76) of

thin sections of two of these cell types (LMTK- and LDTK-) revealed that both

contained 5+1% of the total cell volume within mitochondria. Even after

considering the larger mitochondrial content of HeLaTK- cells (7% of cell

volume), the mtDNA content of HeLaTK- cells was at least four times larger

than that of either TK- L cell line. Thus, the mtDNA content did not simply

relate to the volume of mitochondria. This result was consistent, however, with

the qualitative observation that the mitochondria of HeLaTK- cells were

uniformly more well developed (i.e. contained more cristae per unit volume)

than the mitochondria of L cells.

Cell Division and Regul-ition of MtDNA Copy Number
As noted above, an added degree of complexity is that a given cell type

seems to maintain a characteristic copy number of mtDNA even when dividing

(1). This implies that the mtDNA content must double before each cell

division. Yet, the replication of mtDNA in dividing cells is not linked to the S

phase of the cell cycle, rather it apparently occurs continuously throughout the

cell cycle (77, 78). In addition, mtDNA may replicate in the absence of nuclear

DNA replication, as in cellular hypertrophy (24).

How is the mtDNA copy number regulated during cell division? For the

copy number to be maintained during exponential cell growth, the number of
mtDNA molecules in a culture, N, must parallel the increase in cell number,

N=Noext, where No is the number of molecules present at t=0, and t equals the

cell generation time when N/No=2 (79). However, the rates of mitochondrial

and nuclear DNA synthesis may each be affected differently by exposure to

various chemicals, thereby conceivably altering the mtDNA copy number. The

antiproliferative agent, MGBG, is one such compound. It inhibits mtDNA

synthesis at concentrations which have little effect on nuclear DNA synthesis.

This and other aspects of the drug will be considered in more detail in a later


The phenanthrdine dye, ethidium bromide, is a well-known inhibitor of

mtDNA replication and transcription (80, 81). When human VA2-B cells were

treated with 20 ng/ml ethidium bromide, a progressive dilution of the mtDNA
content, down to 10% of control cell values after three doublings, has been
reported (82). This large reduction is consistent with a logarithmic dilution in

the number of preexisting mtDNA molecules per cell, which in the absence of

any new synthesis or turnover should result, after 3 cell doublings, in a level

which is 1/8 or about 12% of normal. As yet, no study dealing with the long-

term effect of the drug on the mtDNA content of mammalian cells in culture has

been reported, mainly because ethidium bromide limits the growth capacity of

these cells to between 3 and 4 cell generations (82, 83).

In contrast to other vertebrate cells studied so far, chicken embryo

fibroblasts have been found to be inherently resistant to the growth-inhibitory

effect of ethidium bromide when supplied with exogenous pyrimidine

nucleosides or nucleotides (84, 85). Long-term ethidium bromide treatment

resulted in chicken embryo fibroblast cells which had no measurable mtDNA

and which were respiration deficient. This phenotype was maintained

whether or not the cells were transferred to drug-free medium (86). This was

the first and, as yet, only demonstration that vertebrate cells of the rhoo

phenotype (such as in Saccharomvces cerevisiae) can proliferate in culture.

Study of growth parameters indicated that no lag or adaptation period was

necessary for pyrimidine-supplemented chick cell populations to proliferate in

the presence of ethidium bromide (87).

Whereas ethidium bromide inhibits mtDNA synthesis more than nuclear

DNA synthesis, the reverse is true for treatment of cultured mammalian cells

with either 5-fluorodeoxyuridine or methotrexate (88). These drugs inhibit the

enzyme thymidylate synthetase, so that thymidine must be supplied

exogenously. Certain cells (such as LMTK-) lack the major cellular thymidine

kinase but do contain a genetically distinct mitochondrial enzyme.

Consequently, when these cells are treated with methotrexate and supplied

with exogenous thymidine, only the mitochondria can synthesize DNA. When

the concentration of exogenous thymidine is >20 i.M, the rate of mtDNA

synthesis is at least 50% of the control rate while nuclear DNA synthesis is

<4% of control. The resultant effect on the mtDNA copy number was not

measured, however.

"Natural" Homoplasmy and/or Heteroplasmy of MtDNA
The genotype of mtDNA tends to be the same within a given cell; i.e.,

most cells are homoplasmic (that is, they carry only one type of mtDNA). In
addition, the mitochondrial genotype between different tissues of a single
individual generally appears to be the same (89). This is somewhat surprising

since it has been suggested that mtDNA evolves at a rate 10-fold faster than
that of single-copy nuclear DNA (90, 91). Heteroplasmy has been observed in

only a few species, and usually analyzed only in a single individual (92, 93).
This seeming disparity between the rapid rate of mtDNA mutation and the
observation that single variant forms of the mitochondrial genome appear to
dominate the cell population of a single individual (94) can be explained partly

by the fact that mammalian oocyte mitochondria are derived from only a small
number of progenitors (see below) (95).
Unlike nuclear DNA, mtDNA is inherited maternally in mammals, i.e.,

only the egg contributes mtDNA to the zygote. The fate of paternal
mitochondrial genomes remains obscure (96, 97). There is 100 times more
mtDNA in follicular oocytes than in somatic cells, suggesting at least a 100-

fold amplification during oogenesis (98, 99, 100). Because there is no

additional mtDNA synthesis after fertilization until about the 64-cell blastocyst
stage, there is a rapid dilution of the mtDNA copy number to 1-2 per
mitochondrion in zona-encased oocytes. A higher content of mitochondrial
genomes per organelle is probably not recovered until after early cell divisions
of the embryo (101). The eventual mitochondrial genotype of the individual is
derived from only the few of these 64 cells that serve to form the growing
embryo. This may explain, in part, the seemingly incongruous reports that
mtDNA has a rapid rate of mutation and yet most individuals are homoplastic.

In addition, while alleles of nuclear genes segregate only at meiosis,
mitochondrial genes can "segregate" during mitotic divisions of the eukaryotic
cell. This vegetative segregation may explain how an initially heteroplasmic
cell can give rise to homoplasmic daughter cells during somatic cell division
(102). How might this occur? The "average" haploid yeast cell, for example,
contains about 50 copies of the mtDNA molecules and thus about 50 copies of

each mitochondrial gene (103). A mutational event, such as for resistance to
erythromycin (an inhibitor of mitochondrial protein synthesis) is unlikely to

affect more than one copy of a gene at a time and would therefore result
initially in a cell with 1 mutant and 49 wild type alleles. This heteroplasmic cell

eventually produces homoplasmic mutant progeny cells in which all mtDNA
molecules carry the mutant allele. Experiments have shown that selection
plays the major role in determining the fate of erythromycin-resistant genomes
in the presence of erythromycin on a nonfermentable carbon source, i.e. under
"selective" conditions. This was primarily intracellular selection, acting in the
absence of cell division to make cells homoplasmic for the erythromycin-
resistant allele. Hydroxyurea, which inhibits nuclear and mtDNA synthesis

(104), blocked this selection. The simplest interpretation is that selection
requires mtDNA synthesis.
In addition, intercellular selection may play a role in determining the

fate of new erythromycin-resistant mutations in that cells with many copies of
the mutant allele may begin to grow and divide, while homoplasmic
erythromycin-sensitive cells or cells with few copies of the erythrcmycin-
resistant allele cannot divide. Although the mechanism for the intracellular
selection of mutant alleles in this case was unknown, there is some evidence,
at least for the petite mutation in yeast, that the presence of certain nucleotide

sequences and other features of the mutant genome may endow it with a

replicative advantage over the wild type genome, the expression of which can

become suppressed (105).

Segregation of MtDNA in Hybrids and Cvbrids

To further explore how a given mtDNA genotype is selected for

replication and/or expression, numerous experiments with interspecies

hybrids (cell/cell) and cybrids (cell/cytoplasm) have been performed (106,

107). Ideally, the two parental lines are different enough for the DNA, RNA,

and protein from the hybrid (or cybrid) cell mitochondria to be characterized as

belonging to one or both species. Analysis of the mtRNA and protein can

determine whether the two types of mtDNA function completely independently

or interact in a complementary fashion, or whether only one genome is

responsible for all of the mitochondrially coded products.

Segregation patterns of mtDNA in cybrids and hybrids have been

divided into 3 categories, namely 1) segregation of foreign mtDNA (or

chromosome-dependent segregation); 2) stochastic segregation; and 3)

segregation of host mtDNA (108, 109). In interspecies hybrids, such as

human-mouse hybrids, when there is loss of chromosomes from one parental

cell, the mtDNA from that parental cell is also lost (chromosome-dependent

segregation) (110, 111). These observations suggest that an incompatibility

exists between different species of nuclear and mitochondrial genomes and is

responsible for the "segregation of foreign mtDNA" (112).

In contrast, in intraspecies cybrids, there should be no incompatibility

between nuclear and mitochondral genomes. Mitochondrial DNA of

cytoplasts can be propagated in the host cells and both parental mtDNAs are
found to be codominant at an early stage after cybrid isolation (108).

However, after cultivation for several months, some individual cells in the

population of cybrids were observed to have a different ratio of parental

mtDNA (112). A similar phenomenon occurs in mouse-rat hybrids in which

both parental chromosomes are stably retained (109). In these cases, since

there is no selective pressure from the nuclear genomes on either parental

mitochondrial genomes, both parent mtDNAs segregate randomly into

daughter cells and may, in a sufficient number of divisions, become pure for

either parental type of mtDNA (112). This type of mtDNA segregation is

named "stochastic segregation."

Although stochastic segregation of mtDNA should occur in intraspecies

cybrids, in human cybrids constructed by fusion of tumorigenic HeLaTG cells

with cytoplasts of normal primary fibroblasts, host cell mtDNA (HeLaTG

mtDNA) was segregated from the cybrids (113). If stochastic segregation

occurred, some subclones of the cybrids should have contained HeLaTG

mtDNA, but this was never observed. Thus, there must be some differences

between the mitochondrial genomes of HeLaTG cells and fibroblasts that are

responsible for the preferential segregation of HeLaTG mtDNA. This

segregation pattern is referred to as "segregation of host mtDNA."

Recombination of MtDNA and/or its Encoded Products

Recently, metabolic complementation between mitochondria in somatic

cell hybrids or cybnds was suggested from the observation that

chloramphenicol (CAP)-sensitive mtDNA was maintained in cells even when

they were cultivated continuously for a long time with CAP (114, 115). Oliver

and Wallace (116) showed that a marker polypeptide encoded on CAP-

sensitive mtDNA could be synthesized in the presence of CAP. These data
were interpreted to mean that mitochondrial cooperation arises from

mitochondrial fusion or intermitochondrial exchange of mRNA, a finding which

suggests, in turn, that mitochondrial genomes may interact with each other

and form recombinant molecules.

To date, however, there is no direct evidence to suggest that
recombination of mtDNA occurs, at least in mammals. Recombination of
mammalian mtDNA has been examined using mouse X rat and mouse X
hamster somatic cell hybrids and rat cybrids (117, 118). Genetic and physical
analyses showed that the mtDNAs of the hybrids and cybrids were simple
mixtures of the two parental mtDNAs. These observations suggest that, in

contrast to the case with yeast or plant mtDNA, recombination of mammalian
mtDNA occurs rarely, if at all.
Effects of MtDNA on Expression of Nuclear DNA
Saccharomvces cerevisiae cells seem to respond to the quality and

quantity of mtDNA and modulate the levels of nuclear-encoded RNA's,
perhaps as a means of intergenomic regulation (119). These results suggest
a hitherto undescribed type of nuclear-mitochondrial interaction whereby
nuclear DNA sequences can respond to the state of the mitochondrial
genome. For example, the ability of yeast cells to utilize certain sugars varies
in different petites (119), suggesting that the expression of some pathways of

sugar utilization can in some way be influenced by different, defective
mitochondrial genomes. Another related observation is that a linear, double-
stranded DNA killer plasmid from Kluyveromyces lactis, when introduced into
Saccharomyces cerevisiae, is unstable in both rho+ and mit- cells, but is stable
in rhoo petites (120). The abundance of nuclear-encoded RNAs is increased
in one or more of these respiratory-deficient petites, as if these cells were
attempting to compensate for their respiratory-deficient defect or for the
particular mtDNA lesion that they harbor. Perhaps a decreased mtDNA copy
number stimulates the transcription of nuclear genes necessary for mtDNA

How might expression of some nuclear genes be affected differently in

these respiratory-deficient cells? One possibility is that a mitochondrial gene
product is exported and functions outside the organelle, for example, as a
negative (or positive) regulatory element (119, 121). According to this model,

control of expression of some nuclear genes would be maintained as long as
cells contained a functional mitochondrial genome. Upon conversion to

petites, however, the cells could no longer synthesize the putative regulator

since petites have lost the ability to carry out mitochondrial protein synthesis,

and they lack most, and in some cases all, of the mitochondrial transcripts

found in rho+ or mit- cells. Although protein or RNA export from mitochondria
has not been demonstrated directly, there are some indications that such

export may occur.
For example, sporulation in MATalpha/MATa diploids of

Saccharomyces cerevisiae cells is accompanied by a novel pattern of protein

synthesis, as shown by the disappearance of some 'mitotic' polypeptides and

by the appearance of a new set of 'meiotic' polypeptides (103). Inhibition of

mitochondrial protein synthesis by erythromycin caused the disappearance of

several 'meiotic' polypeptides within one hour. Fractionation of extracts of the

cytosolic and mitochondrial components showed that those proteins that were
sensitive to erythromycin were localized within the cytosol and that they were

also sensitive to cycloheximide, an indication that they were not mitochondrial

translational products. However, in a mitochondrially inherited erythromycin-
resistant mutant, which had altered mitochondrial ribosomes, neither
sporulation nor in vivo protein synthesis were affected by erythromycin (122).
This mutant synthesized all of the meiotic polypeptides under sporulation
conditions in the presence of erythromycin, ruling out the possibility that the

drug could act at the level of the cytosolic protein synthesis system during

sporulation. This evidence suggests that mitochondrial protein synthesis is

needed for the expression of some nuclear genes during sporulation.
There is now evidence that mitochondrial gene products may regulate
nuclear gene expression in mammals as well. This comes primarily from the
observations that a particular antigen, although specified in part by a gene in
the major histocompatibility complex, is maternally inherited (123). This

maternally-transmitted antigen (Mta) is a murine cell-surface molecule whose
phenotype is determined solely by that of the mother (124). Sex-linked
transmission is ruled out because the phenotypes of F1 males and females in

a single litter are identical (125). The Mta phenotype cannot be modified by
fostering, embryo transfer or transfer of bone marrow cells to lethally irradiated
mice, making the involvement of a conventional infectious agent unlikely
(123). This interpretation is further supported by the finding that the maternal
lineage of the Mta+ phenotype is stable for at least 11 generations of

backcrossing to Mta- males (126). The expression of Mta in somatic cell
hybrids requires functional mitochondria from the Mta+ parent cell line.
Pretreatment of the Mta* parent with the mitochondrial poison Rhodamine 6G

(127, 128) resulted in hybrids which were Mta-, or diminished in Mta

expression. This is the first evidence for mitochondrial control of the
expression of a cell membrane molecule in eukaryotes. Whereas Mta is
encoded by a structural gene in the major histocompatibility complex, its

expression seems to be controlled by a mtDNA gene product.
The techniques of somatic cell genetics also provide a powerful tool for
investigating whether mitochondrial genomes play a role in the activation
and/or suppression of the phenotypic expression of tumorigenicity. Several
investigators have shown that tumorigenicity can sometimes be suppressed

by fusing tumor cells with cytoplasts from nontumorigenic cells (113, 129).

However, there have been no reports of the induction of tumorigenicity in

nontumorigenic cells by the addition of cytoplasts from tumorigenic cells (130,

131). These findings suggest that cytoplasmic elements and/or mitochondrial
genomes may sometimes contribute to the suppression of tumorigenicity (but

this is not consistently true).

Partitioning of Regulatory Molecules
The notion of partitioning of regulatory molecules between the

mitochondria and the nucleus is not restricted, of course, to products of the
mitochondrial genome. There is now good evidence for the partitioning of

some nuclear gene products between the mitochondria and

extramitochondrial compartments (132, 133). Thus, nuclear-encoded

regulatory factors could partition between the mitochondria and the nucleus,

and their relative distributions could be determined by the presence or

absence, or the amounts, of mtDNA (or RNA) sequences. This mechanism

would allow for regulation of nuclear gene expression in response to the

quality as well as the quantity of mtDNA or RNA sequences. In that way, the
nuclear genome could "sense" the amount or kind of mtDNA in the cell.

Indeed, Clayton e al. (134) have recently shown that a135-nucleotide RNA

species, not encoded in the mitochondrial genome, is the RNA moiety

necessary for activity of the site-specific endoribonuclease involved in primer
RNA metabolism in mammalian mitochondria. This finding implies transport of

a nucleus-encoded RNA, essential for organelle DNA replication, to the
mitochondria matrix.

All organisms contain significant amounts of the polyamines,
spermidine and spermine, and their precursor, putrescine. In mammals the
concentrations of these aliphatic amines vary considerably between different
cell types (135). Thus, cells heavily involved in polynucleotide and protein

synthesis such as proliferating or protein secreting cells contain large amounts
of polyamines (in the mM range), while concentrations in metabolically less
active cells are lower (136). Polyamines are cations at physiological pH and,
in vitro can bind through ionic forces to the negatively-charged groups of
nucleic acids, proteins and phospholipids (137, 138). Polyamines can also
form hydrogen bonds, and the aliphatic hydrocarbon part of their structure may

allow some interactions with hydrophobic environments such as those
occurring in membranes. These properties could contribute to the multiple
actions attributed to the polyamines in many biological processes.
The biosynthetic pathway for putrescine, spermidine and spermine in
mammalian cells has been well established (Figure 2). Two 'decarboxylases'
and two 'aminopropyltransferases' are involved. The two decarboxylases are

of particular interest as they are present in mammalian cells in very small
amounts, have very short half-lives, and are highly inducible (137). Because
of these properties the decarboxylases can regulate polyamine synthesis and
enable cells to respond to a variety of stimuli with a rapid increase in

polyamine levels.
Ornithine decarboxylase (ODC) is a pyridoxal phosphate-dependent
enzyme (136). It is thought to play the key role in polyamine synthesis, and its
half-life of about 10 minutes is the shortest reported for an enzyme in
mammalian cells (139). Ornithine decarboxylase is present in very small

amounts in quiescent cells, and its activity increases rapidly and dramatically




C02 1






---- > Methylthioadenosine
| (MTA)



-- D-SAM

H 2 N-(C H2)3-N H-(CH2)4-N H-(CH2)3-NH2


Figure 2. Polyamine synthesis in mammalian cells. Enzymes involved are:
1) ornithine decarboxylase; 2) S-adenosylmethionine
decarboxylase; 3) spermidine synthase; and 4) spermine synthase.

within a few hours of exposure to trophic stimuli such as hormones, certain
drugs, and growth factors (140, 141, 142).
S-Adenosylmethionine decarboxylase (SAM-DC) also has a relatively
short half-life (~1 hr) and has the added pecularity that it utilizes as cofactor
covalently bound pyruvate instead of pyridoxal phosphate, the usual cofactor
for decarboxylases (136). Like ODC, it is present in mammmalian tissues in

very small amounts, and its activity is regulated by many hormones and other
growth-promoting stimuli. Once decarboxylated, S-adenosylmethionine is
committed to polyamine production in that no other substantial reactions
utilizing decarboxylated S-adenosylmethionine are known. The

decarboxylation of S-adenosylmethionine is the rate-limiting step in
spermidine formation.

The transfer of an aminopropyl group from decarboxylated S-adenosyl-
methionine to putrescine is catalyzed by spermidine synthase, one of the
aminopropyltransferases. The other aminopropyl group needed to convert
spermidine into spermine, is also derived from decarboxylated S-

adenosylmethionine, a reaction catalyzed by the second
aminopropyltransferase, spermine synthase (136). As the activities of

spermidine synthase and spermine synthase are much higher than those of

the two decarboxylases, putrescine and decarboxylated S-
adenosylmethionine concentrations within the cell tend to be very low in
comparison to spermine and spermidine (137).
An active transport system for the polyamines distinct from those for
amino acids has been found in all cells tested to date (136). The activity of this
system is increased in polyamine-depleted cells, and it can maintain cellular
polyamine levels in the presence of very low extracellular concentrations. The
role of the transport system under normal circumstances is unclear, however,

because intracellular synthesis is used to provide polyamines, and
extracellular polyamine concentrations are usually low. The transport system
can also take up other basic substances, including drugs like MGBG; this will

be discussed in greater detail in the next section. In addition, some cultured
cells efflux polyamines when cell growth has been restricted by inhibitors,
confluence, or lack of growth factors (136).

The first step in the catabolism of these polycations is acetylation (143).
The degradation of spermine into spermidine or of spermidine into putrescine,
processes which reverse the normal biosynthetic pathway for polyamines,
involves the sequential activity of two enzymes, N1-acetyltransferase and

polyamine oxidase (Figure 2). Putrescine may be acetylated and eventually
oxidized in a reaction catalyzed by monoamine oxidase, or directly oxidized in

a reaction catalyzed by diamine oxidase, which is induced by growth stimuli.
Polyamines and their acetylated derivatives can also be oxidized in a reaction
catalyzed by plasma amine oxidase (137).

Evidence from studies on polyamine-deficient mutant cells and on
drugs that inhibit polyamine biosynthesis indicates that polyamines are

necessary for cell division and differentiation (136, 144). Indeed, proliferating
tissues are particularly rich in polyamines and their biosynthetic enzymes
(145). Although it is apparent that polyamines play important biological roles,
their exact functions) is unclear. Polyamines have a wide spectrum of effects,
particularly in cell-free systems. Such actions include control of the initiation
of translation, stimulation of ribosomal subunit association, stabilization of
tRNA and DNA structure, and stimulation of RNA and DNA synthesis (135).
Polyamines have also been shown to enhance reactions catalyzed by DNA

polymerase (146), RNA polymerase (147), polynucleotide ligase and kinase

(148), and gyrase (149).

In vitro, polyamines bind preferentially to double-stranded nucleic

acids. They bind in the minor groove of the double helix and, by interacting
with phosphate groups, stabilize the double helix (147). A high concentration

of polyamines induces condensation of nucleic acids. This association with
DNA may represent a major function of polyamines, involving both molecular

stabilization and control of gene expression (150). However, it is prudent to

recognize that many of these actions may only reflect the fact that polyamines
interact with acidic macromolecules in vitro and that physiological relevancy
has not been established.

The polyamines, spermine and spermidine, stimulate the rate of
phosphorylation of selected proteins by several types of cyclic nucleotide-

independent protein kinases (151, 152). For example, spermine and

spermidine increase the activity of cytosolic and nuclear type II casein kinases

(153). The enhancement of certain of these protein kinase reactions by

polycations seems to relate primarily to their interaction with the protein

substrate, yielding more favorable conformations for phosphorylation by the

protein kinase, rather than by an effect on the enzyme per se (152).

Polyamines have also been shown to inhibit certain enzyme reactions.
For example, physiological concentrations of spermine completely inhibit

phosphorylation by nuclear protein kinase NI (151). Polyamines, especially

spermine, inhibit cardiac phospholipid-sensitive, calcium-dependent protein
kinase and myosin light chain kinase, a calmodulin sensitive protein kinase
(154). Even protein phosphatases, including a type isolated from heart, may

be influenced by polyamines with the effect dependent on the substrate used

Moreover, spermidine or spermine, at near-physiological
concentrations of 1.5 mM, stimulates by 8-fold the phosphorylation of

phosphoinisitol (PI) in membranes from A431 cells. The reaction product is

almost exclusively PI-4-phosphate (156). Also, spermidine or spermine can
increase by 4-fold the rate of hydrolysis of synthetic phospholipid analogs by

phospholipase A2 (157). In addition, polyamines are substrates for

transglutaminases, are precursors of gamma-aminobutyric acid, and uncouple

the parietal cell H+, K+-ATPase system (135). Polyamines also appear to be

second messengers in mediating Ca++ fluxes and neurotransmitter release in

potassium-depolarized synaptosomes (158).
Addition of micromolar concentrations of polyamines to cultured

myocardial cells results in an increase of cGMP and a parallel decrease of

cAMP levels (137). Conversely, depletion of intracellular polyamines by

polyamine synthesis inhibitors leads to an accumulation of cAMP and a

depletion of cGMP (159). These effects are mediated by modifications of the

activities of the enzymes involved in cyclic nucleotide metabolism, i.e.

adenylate and guanylate cyclase and cAMP- and cGMP-phosphodiesterase.

Considering that cGMP and cAMP are usually regarded as positive and

negative signals, respectively, for the induction of cell growth, the property of

polyamines to affect the metabolism of the two nucleotides in opposite

directions correlates well with their suggested role in cell growth (137).

Polyamines also have many actions on mitochondria. Mitochondria

can "buffer" cytosolic free Ca++ concentrations, but the physiological

significance of this uptake has been questioned because it is triggered only by
Ca++ concentrations that exceed those found in normal cells. In the presence

of physiological concentrations of Mg++, the extramitochondrial free Ca++
which the mitochondria will buffer is 0.6 to 1 piM (160), but the physiological

free Ca++ concentration in the cytosol is only 0.15 to 0.25 I.M (161, 162). It

now appears that spermine, which had been lost upon permeabilization of


cells, will restore the sensitivity of mitochondria to external Ca++ (160).
Spermine both increases the rate and affinity of Ca++ uptake into mitochondria
and decreases the set-point to which isolated mitochondria buffer free Ca++

concentration. Spermine stimulates Ca++ uptake by mitochondria but not by

microsomes (163). The half maximally effective concentration of spermine (50

gM) is in the range of physiological concentrations of this polyamine in the

cell. Spermidine is 5 times less potent; putrescine is inactive. The stimulation
of mitochondrial Ca++ uptake by spermine is inhibited by Mg++ in a

concentration-dependent manner. Spermine is thus an activator of the
mitochondrial Ca++ uniporter, and Mg++ is an antagonist. Spermine also

inhibits the efflux of Ca++ from mitochondria induced by ruthenium red.

Polyamines may thus confer to the mitochondria an important role in the
regulation of the free Ca++ concentration in the cytoplasm as well as in the

mitochondrial matnx.

At physiological concentrations spermine may also be transported into

rat liver mitochondrial matrix space, provided that mitochondria are energized

and inorganic phosphate is available for concurrent transport (164).
Furthermore, mersalyl, a known inhibitor of phosphate transport, prevents both

spermine uptake and release. Magnesium ions inhibit the transport of
spermine, conceivably by competing for the same binding sites on the

mitochondrial membrane.
Fixed negative charges of anionic sites have previously been proposed

as possible bind:ng sites for large organic cations within the mitochondrial
membranes (165). These binding sites are probably composed mainly of
ionized phospholipids and may be involved in transport of small cations and
anions across the mitochondrial membrane. The phospholipid-containing

binding sites are reportedly localized mainly at the outer face of inner

mitochondrial membrane.
In 1960, Tabor (166) reported that low concentrations of spermine and

spermidine inhibited swelling of mitochondria in hypotonic media. More

recently, polyamines were shown to improve the respiratory control and

prevent loss of control in heat-aged mitochondria, protect oxidative

phosphorylation, and prevent a fall in the mitochondrial membrane potential

(167, 168, 169). At concentrations comparable to those attained

intracellularly, the polyamines inhibited state 4 respiration of rat liver

mitochondria, but they had much less effect on state 3 or uncoupled

respiration (125 = 7.5 and 7.0 mM for spermidine and spermine, respectively)

(167). Electron microscopy revealed that polyamines caused the outer

mitochondrial compartment to collapse, bringing the inner and outer

membranes into apparent contact with one another (167). Recent evidence

suggests that the import of nuclear-coded proteins into mitochondria occurs at

sites where the inner and outer mitochondrial membranes are juxtaposed to

one another (170). It seems possible, therefore, that polyamines may affect

this process.
MGBG (Methylglyoxal bisa[uanylhydrazonel)

History and Clinical Utility

MGBG has been recognized as a potent antiproliferative agent since

1958 when Freedlander and French (171) reported that it inhibited the growth
of L1210 leukemia cells in mice. Clinical studies disclosed that MGBG (also
known as methyl-GAG) could induce remissions in some patients with acute
myelocytic leukemia, malignant lymphoma, and certain other neoplasms (4,

5). MGBG has also been used in the treatment of trypanosomiasis (172).

However, MGBG accumulated to extremely high levels in certain

normal cells (6, 173), and the consequent severe toxicity set practical
limitations on its use (3). Mucositis was the dose-limiting toxicity in Phase I-II

clinical trials (4). Fatigue, anorexia, mild nausea and vomiting, myalgias, and
neuropathy were also seen. The subacute toxicity of MGBG involved both
proliferative (gastrointestinal, bone marrow, and lymphoid) and
nonproliferative (hepatic, renal, and cardiac) tissues (174). However, more

recent investigations involving variation in dosage schedules and use in
combination with other anti-proliferative agents have rekindled interest in the

clinical use of MGBG (3, 175). Because no metabolites of MGBG have ever
been detected in urine, feces, or various tissues, and virtually no radioactive

CO2 was expired after in vivo administration of [14C]-MGBG, it is believed that
MGBG does not undergo biotransformation in higher animals (175).

Relationship to Polyamines
French and his colleagues were the first to point out that MGBG had

structural resemblances to spermidine (Figure 3), and they suggested that
MGBG might interfere with the biological functions of this naturally-occurring

polyamine (3). MGBG is a basic molecule consisting of 2 aminoguanidine
groups separated by an aliphatic chain. MGBG has pKa's of about 7.5 and 9.2
at 250C (3), presumably due to resonance stabilization of the protonated forms
of the aminoguanidine moieties at both ends of the molecule. Spermidine,
whose molecular contours are quite similar to those of MGBG in its most
extended conformations, has pKa values of about 8.4, 9.8 and 10.8 at 250C

MGBG is a potent (<1 .LM) competitive inhibitor of S-adenosyl-

methionine decarboxylase (SAM-DC), which is a key enzyme in the
biosynthesis of the polyamines, spermidine and spermine (176, 177).





Figure 3. Structures of spermidine and its analog, MGBG.

However, the inhibitory effects of MGBG on polyamine biosynthesis in animals

are transient because of a striking secondary increase in SAM-DC. It is

hypothesized that binding of MGBG to the enzyme may stabilize it, and hence
prolong the normal one hour half-life of SAM-DC (178).
MGBG is also a potent noncompetitive inhibitor of diamine oxidase

(179), an enzyme involved in putrescine degradation. In addition, MGBG
causes an increase in the activities of ornithine decarboxylase (ODC), which

synthesizes putrescine, and in spermidine/spermine N1-acetyl-transferase

(SAT), an enzyme which participates in the degradation of spermine and
spermidine to putrescine (176). The increases of both ODC and SAT activities
provoked by MGBG in cultured human erythroid leukemia K562 cells were

blocked by treatment with cycloheximide, a finding that suggests that the
increase in enzyme activity required synthesis of protein. The putrescine

content in cells treated with MGBG increased 20-fold, whereas the levels of

spermidine and spermine were depressed by 35 and 50%, respectively. The
marked increase in putrescine levels presumably resulted from inhibition of
SAM-DC and diamine oxidase, and induction of ODC and spermidine and

spermine acetyltransferases. All of these factors act to increase putrescine
Thus, in cells treated with MGBG, putrescine pools increase while those

of spermidine and spermine decrease slowly (180). Nonetheless, a number of
laboratories (181, 182, 183) have been unable to correlate these changes in
polyamine concentration with inhibition of cell growth. It seems that the
antiproliferative effect produced by MGBG correlates much better with the
intracellular MGBG concentrations than with the depletion of polyamines (183,
184). The ability of spermidine to prevent the antiproliferative effects of MGBG,
initially a critical observation in linking MGBG to polyamines, is now attributed

to competition for cellular uptake rather than replenishment of MGBG-depleted

spermidine pools (185). Moreover, difluoromethylornithine (DFMO), a highly
specific inhibitor of polyamine biosynthesis, is cytostatic (186), whereas MGBG
is cytotoxic, a result suggesting that the two drugs may have different
mechanisms of action.
Effects on Mitochondria
Indeed, another pharmacological action for MGBG has been identified.

In 1966, Pine and DiPaolo (187) found that MGBG and certain other cationic

compounds inhibited the respiration of ascites L1210 cells, and uncoupled
phosphorylation in isolated mitochondria. In common with 2,4-dinitrophenol,

the drugs selectively inhibited acetate incorporation into lipid and, to some

extent, reduced the cellular pools of ATP. In a different study, pyruvate
oxidation, another measure of mitochondrial function, was significantly
decreased 4 hours after exposure of ascites L1210 leukemia cells to MGBG;
growth inhibition did not occur until after 10-12 hours (185). Depletion of
spermidine pools to levels comparable to those attained during DFMO-

induced cytostasis did not occur even after 24 hours of exposure. Also, in

studies with intact A30 cells, the rate of oxygen uptake in MGBG-treated cells
was decreased to as low as 10% of control (188). The degree of inhibition of

oxygen consumption was dependent on the concentration of MGBG. When
spermidine was added to the MGBG-containing medium, however, respiratory
activity was restored and approached that of normal cells.
Pathak lt al. (189) have demonstrated that exposure of cultured

leukemia L1210 cells to 0.1 to 10 pM MGBG resulted in a concentration-

dependent inhibition of cellular proliferation which became apparent after
about 1 to 2 generation times (12 to 24 hr). Ultrastructural examination of cells
exposed to at least 1.0 iM MGBG for 24 hr revealed extensive swelling of

mitochondria, deterioration and eventual loss of cristae and increase in matrix

density. This ultrastructural damage was selective for the mitochondria; there

were no concomitant changes in nuclear ultrastructure (190). In addition, the

mitochondrial damage preceded growth inhibition by about 12 hr and did not

immediately affect cell viability. Because damage was apparent in all of the

mitochondria of each affected cell, the organelle population was probably

homogeneous in its sensitivity to the drug. Moreover, exposure to 10 g.M

MGBG for 24 hr caused all of the cells to be affected. Whether the

mitochondrial structural damage was associated with functional deterioration

and whether the damage was reversible were not demonstrated.

Similar damage has been observed in ascites L1210 cells treated in

vivo with a single dose (75 mg/kg) of either MGBG or ethidium bromide (190).

After 24 hr, the mitochondria were swollen, lost their inner structure, and, in the

case of treatment with ethidium bromide, developed numerous electron

densities within the matrix. Analysis of the nucleotide pools of these cells by

HPLC revealed that treatment with either MGBG or ethidium bromide depleted

the ATP pools to 52 and 16% of control, respectively. The overall adenylate

energy charge of the cell (1/2[ADP+2ATP] [AMP+ADP+ATP]) was also

reduced to 68 and 58% of control, respectively. The measurement of ATP

levels suggested that, whereas MGBG caused substantial mitochondrial

damage, the resultant energy depletion oer se was not sufficient to account for

growth inhibition (191).

When MGBG-treated ceils were harvested, washed, and reinoculated

into untreated mice, all cells reattained nearly normal ultrastructure after 48 hr,

whereas ethidium bromide-treated cells did not, even after 96 hr (190).

Analysis of the nucleotide pools of the MGBG-treated cells indicated that the

mitochondria had recovered their functional capabilities as well. The

adenylate energy charge for these cells was essentially the same as that for

untreated cells. Also, the cells previously treated with MGBG displayed

unaltered leukemogenicity in that the animals died at nearly the same time as

they did after inoculation with untreated cells.

MGBG also causes ultrastructural damage to mitochondria in several

other murine and human cell types including P288 mouse leukemia, L-cell

mouse fibroblast. C3H/10T1/2 mouse embryo fibroblast, and NALM-1 human

chronic myelocytic leukemia cells (176). Interestingly, the onset of

mitochondrial damage in these cell lines related inversely to the generation

time of the particular cell lines. This relationship between mitochondrial

damage and the rate of cell proliferation was tested in two separate cell

systems (174). In cultures of human lymphocytes stimulated with

phytohemagglutinin, only those cells undergoing blastogenesis were bound to

be affected by MGBG. Similarly, MGBG treatment of confluent and

subconfluent cultures of C3H/10T1/2 mouse embryo fibroblasts affected only

the dividing, subconfluent cells.

The effects of MGBG on isolated rat liver mitochondria have also been

studied (191). At drug concentrations comparable to those attained

intracellularly, MGBG significantly inhibited state 4 respiration (125=6 mM), but

had less of an inhibitory effect on state 3 or uncoupled respiration (125=16-20

mM). This may be due to the fact that in the absence of ADP (state 4),

mitochondria generate a significant electrochemical gradient across their

inner membrane. By election microscopy, MGBG-treated mitochondria

appeared swollen, and the spaces between cristae membranes or inner and

outer membranes were collapsed, obliterating the intermembrane space


Since the inhibition of mitochondrial respiration in isolated
mitochondria was prevented by potassium cations and enhanced by
valinomycin, the drug may compete for potassium-binding sites, possibly
membrane phosphoplipids (191). Pretreatment of mitochondria with MGBG
protected against the nonspecific swelling effects of Triton X-100, and the
electrophoretic mobility of mitochondria was markedly slowed by MGBG.

Overall, the data suggest that MGBG neutralizes the net negative surface

potential of isolated mitochondria by binding to sites (possibly phospholipids)
at the inner mitochondrial membrane; any subsequent interference with cation
binding and/or transport may result in inhibition of bioenergetic functions.

Fixed negative charges of anionic sites have been proposed as

possible binding sites for large organic cations within the mitochondrial

membranes (165). Neutralization of this surface potential with compounds

such as polyamines can affect many mitochondrial activities, as well as
ultrastructure. Similar or identical binding sites at the inner mitochondrial

membrane have been suggested for binding certain bis[guanylhydrazones]

(191). Moreover, it has been shown that cationic drugs need not penetrate

across inner mitochondrial membranes because binding sites are probably

fixed at the outer surface (165). In addition, the effects of MGBG on isolated rat
liver mitochondria are readily prevented or reversed by polyamines, and these

interactions are also affected by the mitochondrial transmembrane potential
(192). Magnesium cations enhance the protective action of polyamines.
These data indicate that competition exists between MGBG and polyamines
for low affinity, negatively-charged binding sites at the outer surface of inner
mitochondria! membranes.
Evidence which implicates mitochondrial damage as important in the

antiproliferative effects of MGBG comes from studies of other polyamine

analogs. For example, 4,4'-diacetyldiphenylurea-bis(guanylhydrazone) is an
aromatic bis-(guanylhydrazone) with potent antiproliferative activity. The drug
causes profound ultrastructural damage to mitochondria (174), but has no
effect on polyamine biosynthesis (193). Similarly, ultrastructural damage also
precedes detectable inhibition of cell growth (189). Damage to mitochondrial
ultrastructure is selective for proliferating cells. This has been observed
among cultured cells (174) and more recently in the intestinal epithelium
(194), a site of MGBG toxicity, where damage only occurs to mitochondria of
the dividing crypt cells but not to the nondividing villous cells. Although uptake
studies were not done, Diala e al. (195) made the interesting observation that
wild-type yeast are growth-inhibited by MGBG while petite mutants, lacking
mitochondria, are not. Growth inhibition by MGBG was antagonized by
spermidine, but neither the polyamine nor MGBG had any effect on growth
inhibition by ethidium bromide. MGBG had little or no effect on cell viability
throughout the 2-day exposure.
While the numerous effects of MGBG on mitochondria provide

substantial indication that these actions are responsible for the
antiproliferative activity of the drug, this relationship has not yet been
established. Several workers have demonstrated that normal and malignant
cells possess high-affinity and saturable transport systems that promote
intracellular accumulation of putrescine, spermidine, spermine, and certain
agents such as MGBG (196, 197). Because a saturable carrier is involved, the
system seems to be specific for certain structural features of the polyamines
and their analogs. In general, it appears that putrescine, spermidine,
spermine, and MGBG are transported by common membrane transport
systems that are: 1) distinguishable from other known transport systems (e.g.
for various amino acids); 2) saturable and temperature-dependent, with

maximal rates at 370C; 3) inhibited by uncouplers of oxidative phosphorylation
and certain respiratory poisons; and 4) dependent on the proliferative and, in
some instances, hormonal status of the cells (198). Like the polyamine-
biosynthetic enzymes, the transport system seems to be highly regulated,
responding rapidly to polyamine antimetabolites such as DFMO, an
irreversible inhibitor of ODC, and MGBG, an inhibitor of SAM-DC (199, 200,
201). Agents such as MGBG, which utilize the polyamine transport system,
may be concentrated intracellularly more than 1000-fold relative to the
medium (196, 202).
Seppanen e al. (203) have found that cells can concentrate MGBG

some 600- to 1500-fold so that, following exposure to 5-10 I.M drug, the
internal concentration is 4-6 mM. A minimum intracellular concentration of 0.5
to 1 mM was required for growth inhibition to occur. Seppanen g ai. (204)
also found that MGBG uptake is critically dependent on the growth rate of
tumor cells (i.e., slowly dividing cells transport less MGBG than rapidly dividing
cells). Mikles-Robertson et al. (174) found that MGBG cytotoxicity also
correlated with cell growth rate.
Recently, mutants of human fibroblast VA2 cells have been developed

which are 10- to 20-fold more resistant to the growth inhibitory effects of MGBG
than the parent line (205, 206). The variants took up MGBG to a similar extent
as wild type during short-term incubations (less than 5 min), accumulated less
drug during longer incubations (30-120 min), and more readily lost MGBG
when shifted to drug-free medium, results which might indicate decreased
intracellular binding of MGBG. In the absence or presence of MGBG, the
overall pools of polyamines were not appreciably changed from those of the
wild type. Pyruvate oxidation was not significantly inhibited in the mutants
when grown in the presence of MGBG, but was dramatically reduced in wild

type cells. Furthermore, the mitochondrial ultrastructure of the drug-resistant
cell lines was essentially unaffected by culture in MGBG. The resistant cells
took up about 40% less MGBG than their sensitive counterparts. This
difference is considered small among transport mutants (203). Drug
resistance was not cytoplasmically transmitted by cytoplast cell fusion for any
of the four sublines, suggesting that the genes responsible for resistance were
likely of nuclear rather than mitochondrial origin (206). Defective polyamine
and MGBG transport was also found in mutant Chinese hamster ovary and rat
myoblast cells that were resistant to MGBG (207).
Because of the inhibition of cell growth by MGBG, many studies have

addressed the effects of the drug on DNA synthesis. Unlike aromatic
bisguanylhydrazones, which bind effectively to nuclear DNA, MGBG binds
only weakly to DNA and has little effect on DNA polymerase alpha or
thymidine kinase (198, 208). Addition of MGBG to HeLa S3 suspension
cultures caused a decrease in incorporation of [3H]thymidine into nuclear
DNA, but only after putrescine levels had increased and spermidine and
spermine levels had decreased. Since the rate of DNA chain elongation was
only reduced slightly by MGBG, decreased levels of spermidine and spermine
might lead to a decrease in the number of replication units active in DNA
synthesis within each cell.
Effects on MtDNA
The effects of MGBG on mtDNA synthesis have also been investigated

because of the mitochondrial damage elicited by this drug. In exponentially-
growing L1210 leukemia cells prelabeled with [14C]thymidine, the
incorporation of [3H]thymidine into mtDNA was selectively inhibited (55% of
control) when studied 5 hr after exposure of cells to 10 jiM MGBG (7).
Incorporation of label into nuclear DNA, however, was not affected until 8 hr;

the rate of nuclear DNA synthesis was 60% of control after 12 hr. After 16 hr of
exposure, the rates of nuclear and mtDNA synthesis were both about 40% of
control values. Exposure of L1210 cells to 1 IpM MGBG resulted in a mtDNA
synthesis rate 60% of control. At this concentration, growth inhibition was 90%
of control after 18 hr of exposure.
Dye-CsCI gradients of mtDNA indicated that the inhibition of synthesis

occurred in replicative forms of circular DNA (7). Uptake studies excluded the
possibility of drug interference with cellular uptake of thymidine.
Ultrastructural studies revealed a very close correlation between the dose-
response curve for mitochondrial damage and that for MGBG inhibition of
mtDNA synthesis. This correlation suggests a close relationship between
ultrastructural damage and inhibition of mtDNA synthesis, but the possibility
exists that both are epiphenomena of another action of the drug.
A more detailed study of the effects of MGBG on mtDNA synthesis in the
Syrian hamster tumor cell line was done by Nass (6). The fate of mtDNA both
during (24-48 hr) and following (7 hr to 7 days) MGBG treatment was
monitored by ultrastructural, pulse-labeling, and restriction cleavage methods.
MGBG (50 IM) selectively inhibited mtDNA replication prior to significant
inhibiton of nuclear DNA synthesis (73 versus 19%, respectively, at 16 hr).
The drug induced structural alterations, without substantial degradation, of the
closed circular form of mtDNA. Importantly, D-loop strand (7S) initiation was
completely abolished, but mtDNA strands already initiated were able to
complete the circle of replication.
Electron microscopy revealed selective ultrastructural damage to the
mitochondria (including swelling, loss of cristae and matrix components, and
dense inclusions) in up to 96% of cells, while nuclei appeared normal.
Quantitative assays of the uptake and retention of rhodamine 123 in these

MGBG-treated cells revealed that the mitochondrial membrane potential was

maintained. After removal of MGBG, mtDNA resumed replicative active, and

damaged mitochondria recovered near-normal ultrastructure within 1 to 2
Specific Aims

Compounds which, at certain doses, inhibit mtDNA replication more than

cell division would be predicted to deplete the characteristic mtDNA copy
number with each successive cell division. Preliminary studies in our
laboratory with mouse leukemia L1210 cells treated chronically with MGBG at
concentrations which initially inhibited mtDNA replication did decrease the
mtDNA copy number. These studies also demonstrated that there are doses

at which long term survival in MGBG is possible. I wanted to determine if there
was a minimal copy number of mtDNA which cells would tolerate and whether

a new steady state level of mtDNA would be established with chronic MGBG
treatment. I also wished to explore in detail the relative contributions of the
rates of mtDNA replication and cell division to the regulation of the mtDNA
copy number during MGBG treatment. And, finally, I endeavored to determine
how mitochondrial and/or cell function was affected as a consequence of a
change in the mtDNA copy number.
The specific aims of this research were the following.
1) Determine the degree to which mtDNA replication and cell division

are coupled. That is, will cells continue to divide even though their mtDNA
replication is inhibited, thus halving the amount of mtDNA per cell with each
2) If so, determine the minimum copy number of mtDNA per cell which

these cells will tolerate.

3) Once the copy number of mtDNA is diluted by MGBG treatment,

determine whether these cells establish a new steady state with respect to the

content of mtDNA, or return to their characteristic copy number.
4) Determine the relative contributions of the mtDNA doubling rate and

the cell doubling rate to the regulation of the mtDNA copy number in MGBG-
treated cells.
5) Determine if mitochondrial function is normal or impaired

proportionally to the mtDNA copy number, at least as measured by rhodamine

123 uptake.

Cell Culture

Mouse leukemia L1210 cells were grown as suspension cultures in 25
or 75 cm2, canted-neck tissue culture flasks (Fisher Scientific). The cells were

maintained in an incubator (National Appliance Company, Portland, Oregon)
at 370C, and the flasks were kept tightly capped. The medium used was RPMI
1640 supplemented with 16 mM HEPES (3-[N-morpholino] propanesulfonic
acid), 8 mM MOPS (N-2-hydroxyethyl-piperazine-N'-2-ethanesulfonic acid),
2.0 g/L NaHCO3, and10% fetal bovine serum. The starting pH was 7.4. Stock

solutions of MGBG and diethylspermine (DES) were made to concentrations
of 1 mM in water. All ingredients were obtained through Sigma Chemical Co

(except DES, gift of R. Bergeron). The medium and drugs were sterilized by
0.2 l. filtration (Fisher Scientific). Cells were maintained in exponential growth
by reseeding every two days into fresh medium; the starting concentration was
1x105 cells/ml.
Cell Counts

Aliquots (200 pil) of cells to be counted were diluted in 10 ml of Hematall

diluent (Fisher Scientific). The cell number was determined by electronic
particle analysis (Coulter Counter, Model ZF, Coulter Electronics, Hialeah, FL).

Two aliquots of cells were counted from each flask and the mean determined.
Because the cells were passed every two days, a calculation was

devised to plot the counts as a single, continuous line over the entire time
course of an experiment. Cumulative counts were calculated simply as

Cumulative count(n+1) = Cumulative count(n)x Concentration of cells(n+1x
Conrertrtion of viable lls(n)

The mean population doubling time of the cells was determined as
follows: N=NoeXt, where t=time, No is the number of cells present at t=0, and

the doubling time is the value of t when N=2No.

Determination of IC50
The IC50 was defined as the concentration of compound necessary to

decrease cell growth to 50% of control growth at a given time. The percentage
of control growth was determined as follows:
[final cell number (treated group) initial inoculum] X 100
final cell number (control group) initial inoculum
The IC50 of MGBG was determined for the first 2 days of treatment.

Cell Viability
Trypan blue (Eastman Kodak, Rochester, NY) was added to 100 pl
aliquots of cells to a final concentration of 0.06%. The cells were mixed well,

and a 10 pl aliquot was transferred to a hemacytometer (Reichert Scientific
Instruments, Buffalo, NY). At least 100 cells per sample were counted with a

phase contrast microscope (Ernst Leitz, Wetzlar, Germany) at 200X
magnification. The percent viability of the cells was determined as follows:

number of cells excluding trypan blue X 100
total number of cells
Clonogenic Assay
Cells from selected MGBG treatments were counted and diluted to a
concentration of 4 cells/ml. Aliquots of 100 p.l were transferred to triplicate 96-
well culture plates (Fisher Scientific) and incubated at 370C for 1 week. The
culture plates were then examined with an inverse phase microscope (Ze;ss,
West Germany) for numbers of wells which had colonies of cells. Groups of
>50 celis/well were considered as having been cloned from a single viable

cell. The percent cloning efficiency of the drug-treated cells was determined
as follows:
number of wells with colonies (treated group) X 100
number of wells with colonies (control group)
Cell Size

The diameter of the cells was determined directly by the method of
Schwartz t al. (209). Uniform polymeric microspheres (Polyscience,
Warrington, PA) from 4.72 to 10.0 I. in diameter were diluted in Hematall, and
the particle size measured electronically with a FACS Analyzer (Becton-
Dickinson, Sunnyvale, CA) with the amplifier in the log mode. The peak
channel number was plotted against the corresponding calibrated diameter for

each size of microbead to obtain a calibration curve.
Aliquots of 1x106 cells were collected, pelleted, and resuspended in
0.5 ml Hematall for analysis. The peak channel number for the cells was
plotted on the calibration curve to obtain the approximate cell diameter. The
volume of the cells was then calculated from this diameter.
Flow Cytometr
Flow cytometric analysis of nuclear DNA content was done with a
RATCOM flow cytometer (RATCOM Inc., Miami, FL) interfaced with a
microcomputer (IBM-XP). Aliquots of 1x106 cells were resuspended in 0.5 ml
Hematall and were stained with 4',6-diamidino-2-phenylindole (210). The
distribution of nuclear DNA content among the cells was plotted and
integrated by computer.
Rhodamine 123 Uptake

Aliquots of 1x106 cells were pelleted and resuspended in 100 p. of
serum-free medium. Rhodamine 123 (2-[6-amino-3-imino-3H-xanthen-9-yl]-
benzoic acid methyl ester) (Sigma Chemical) was added to each aliquot to

give a final concentration of 1 gg/ml (211). The samples were incubated at
370C for 10 min and washed once with medium. The stained cells were
observed under an epifluorescence microscope (Zeiss) at 400X magnification.
The excitation wavelength was 485 nm.
Preparation of[14C1-MGBG Stock Solution

A stock solution of 1 mM [14C]-MGBG (specific activity 26 mCi/mmol,
Amersham) was prepared aseptically. Aliquots of this stock were added to
media to obtain the indicated final concentration of drug.
Uptake Protocol
Cells were incubated with [14C]-MGBG at concentrations ranging from
1.0 to 3.3 g.M. At the end of each incubation period, an appropriate aliquot of
cells was transferred to a 15-ml tube, centrifuged at 650g for 3 minutes, and
the supernatant was decanted. To remove any labeled drug bound to the
outside of the cell, the cell pellet was washed twice with 4C media containing
1 mM unlabeled MGBG. The cell pellet was resuspended in 1 ml of distilled

water, transferred to a scintillation vial, and the amount of [14C]-label

quantitated by scintillation counting.
Cloning of MtDNA Probe

The 16,295 base pair, full-length murine mtDNA was inserted into the
Sac I site of pSP64 and transfected into E. coli strain HB101 (gift of W.

Hauswirth). A 20 |I1 inoculate of E. coli containing the recombinant plasmid
was aseptically transferred to 250 ml of sterilized Luria broth containing 2.5 g
Bacto-tryptone, 1.25 g Bacto-yeast extract (Difco Laboratories, Detroit, MI), 2.5
g sodium chloride, pH 7.5, and 25 mg ampicillin (Sigma Chemical). The
bacteria were incubated for 16 hrs in a 370C shaking incubator (New
Brunswick Scientific, Edison, NJ) in a 1 L sterilized flask to maximize aeration.

The bacteria were transferred to six 50-ml tubes and centrifuged at

2650g for 15 min at 4C. Pellets were resuspended in 0.5 ml sterile buffer
(25% sucrose, 50 mM Tris, pH 8.0) and combined in a 50 ml centrifuge tube.

Five mg of egg white lysozyme (Calbiochem, La Jolla, CA) was added and,
after mixing, the material was allowed to incubate at room temperature for 5
min. One ml of sterile Na2EDTA (0.25 M, pH 8.0) was added, and the

suspension was allowed to incubate at room temperature for 2 min. Eight ml
of a buffer/detergent/ (50 mM Tris and 62.5 mM Na2EDTA, pH 8.0; that was

sterilized before addition of 2% Triton X-100) was added and mixed gently for
2-5 min at room temperature until the viscosity increased markedly. The
mixture was then centrifuged immediately at 31,500g for 25 min at 120C.

The supernatant was transferred to a 50 ml centrifuge tube, the volume
measured, and an equal volume placed in each of 2 ultracentrifuge tubes.

Cesium chloride (0.9 g/ml of supernatant) (BRL Scientific, Gaithersburg, MD)
was added and mixed by inverting the tube. Ethidium bromide (0.5 mg/ml of
supernatant) was added, mixed, and protected from light in all further steps.
The tubes were balanced and centrifuged at 92,000g for 36-40 hrs at 200C.
Bands of material which fluoresced upon illumination at 360 nm were

located with a hand-held UV light (Ultra-Violet Products, Inc., San Gabriel,
CA). The bands were, starting at the top, protein, Form II DNA (nicked

circular), Form I DNA (closed circular), and RNA. Form I DNA was collected
and transferred to a centrifuge tube. The ethidium bromide was removed by
repeated extraction with isopropanol:5M NaCI (1:1) until tne isopropanol was
completely transparent two consecutive times. For each ml of solution, 2.6 ml
of distilled water and 7.2 ml of -200C 95% ethano! were added. The mixture

was then placed in an ethanol/dry ice bath for 30 min and centrifuged at
9000g for 30 min at 4C. The pellet was rinsed with 70% ethanol and vacuum

dried for 30 min. The final pellet was resuspended in 200 p.! of TE buffer (10
mM Tris-HCI containing 1 mM Na2EDTA, pH 8.0).

The molecular weight and concentration of DNA was estimated by gel
electrophoresis of the sample and comparison to marker DNA's, the
concentration and molecular weight of which were known.
Nick Translation of MtDNA Probe

To label the mtDNA probe with [35S] nucleotides, 1 p.g of the cloned
probe was added to 20 gM each of dCTP, dGTP, and dTTP, 50 gCi of
deoxyadenosine 5'(alpha [35S]thio)triphosphate (specific activity 1145
Ci/mmole, Amersham, Arlington Heights, IL), and a mixture of polymerase I
and DNase I (0.2 units and 0.2 ng, respectively) (BRL Scientific) in a final
volume of 50 Il. The mixture was allowed to incubate at 15-170C for 90 min.
Stop solution (30 mM Na2EDTA) was added, and the mixture put on ice.

The nick translated products were loaded onto a Sephadex G-50
column (Pharmacia Inc., Piscataway, NJ) which had been washed previously
with TE buffer and loaded with 3 gg yeast tRNA. The column was centrifuged
at 400g for 4 min and the eluate collected in a 1.5 ml Eppendorf tube.
Gel Electrophoresis
Gels were prepared with 0.7% agarose (Bio-Rad Laboratories,
Richmond, CA) in TBE buffer (1 M Tris, 0.5 M borate, and 50 mM Na2EDTA,

pH 8.0). Molecular weight markers were the Hind III fragments of lambda virus
DNA and a supercoiled pBR322 ladder (BRL Scientific). Tracking dye
(0.025% bromcphenol blue in 2% glycero!) was added to all samples.
Samples were electrophoresed for 150 minutes at 100 volts; the running
buffer was TBE. The gels were stained in the dark with 1 pg!m! ethidium
bromide for 30 min. The DNA was visualized at 360 nm on a UV light box
(Ultra-Violet Products). The gel was photographed under UV light with a

Polaroid camera (F-stop 8, exposure 4 sec) and Type 57 film (Polaroid Corp.,
Cambridge, MA).
Southern Blot
After electrophoresis, the gel was soaked in 100 ml 0.5 M NaOH in 1.5
M NaCI for 1 hr to denature the DNA. The gel was neutralized with 100 ml 1 M
Tris (pH 8.0) in 1.5 M NaCI for 1 hr. The gel was then soaked in 100 ml 10X
SSC (20X SSC = 3 M NaCI, 0.3 M sodium citrate, pH 7.0) for 15 min. The gel
was blotted onto nitrocellulose filters (Bio-Rad Laboratories) by the method of
Southern (212). The filters were incubated at 850C in a vacuum oven
(National Appliance Company) for 2-3 hrs. Immobilized DNA was hybridized
with the full-length mouse mtDNA probe inserted in the Sac I site of pSP64
and labeled by nick translation with [35S]-labeled dATP described above.
Dot Blot
Mitochondrial DNA was quantitated by dot blot analysis of whole cell
lysate by a modification of the method of Barker t al. (213). Aliquots of 0.5-
2.0x105 cells were placed in 1.5 ml Eppendorf tubes. The cells were
sedimented by centrifugation for 10 sec at 13,000g in a microfuge (Fisher
Scientific). The supernatant was decanted, and the cells were resuspended in
4 .l of buffer (250 mM sucrose, 10 mM Tris, and 1 mM Na2EDTA, pH 7.4).

Cells were then incubated in a mixture of 2% sodium dodecyl sulfate, 5 gg
proteinase K (Sigma Chemical), and 100 units RNase T1 (BRL Scientific) in a
final volume of 7.5-8.5 pIl. The initial incubation was for 15 min at 370C, a
procedure which allowed degradation of up to 4 plg of added yeast tRNA
without inactivation of the RNase T1 by the proteinase K (data not shown).
This was followed by incubation for an additional 2-3 hr at 500C.
The cell lysate was then treated with 0.1 M NaOH to denature DNA, and
the mixture was applied to nitrocellulose filters with a dot blot apparatus

(BioRad Laboratories). Immobilized DNA was hybridized with the full-length
mouse mtDNA probe inserted in the Sac I site of pSP64 and labeled by nick
translation with [35S]-labeled dATP described above. The dots were
visualized by autoradiography (SB-5 X-ray film, Eastman Kodak). To assure
equal size of the dots, nitrocellulose paper was cut out with a hole punch, and
quantitated by scintillation counting.
Hybridization with [35S1-labeled MtDNA Probe
Nitrocellulose paper from the Southern or dot blot was placed in a
plastic Seal-a-meal bag (Sears) and a pre-hybridization solution was added
(30 ml 6X SSC, 0.6 ml 50X Denhardt's solution [1% w/v Ficoll 400,
polyvinylpyrrolidone, and bovine serum albumin in distilled water], and 0.9 mg
denatured yeast tRNA). The bag was sealed and allowed to incubate for 8 hrs
at 680C, The pre-hybridization solution was then drained, and the
hybridization solution added (10 ml 6X SSC, 0.2 ml 50X Denhardt's solution,
0.1 mg denatured yeast tRNA, and 50 pl1 denatured [35S]-labeled mtDNA).
The bag was resealed and incubated 18-24 hrs at 680C.
The hybridization solution was drained into a radioactive waste
container, and the blot was removed. The blot was washed first in 100 ml of
6X SSC for 2 hrs at 68C, and then in 100 ml of 2X SSC for another 2 hrs at
680C. The blot was placed between 2 sheets of filter paper and dried in a
vacuum oven (National Appliance Company) for 15 min at 850C. The blot was
placed under autoradiographic film (SB-5, Eastman Kodak) in a film cassette.
The film was exposed for 15-20 hrs and developed witn an X-ray film
processor (Konishiroku Photo Co., Japan).


Characterization of L1210 Cells
Growth Characteristics of Untreated L1210 Cells
To determine the growth characteristics of the untreated murine
leukemia L1210 cell line, cells were seeded to a starting concentration of
2.5x104 cells/ml and incubated at 370 for 5 days without reseeding or adding
fresh medium. The exponential growth, or log phase, of these cells
encompassed a concentration range from 2.5x104 to 2x106 cells/mi (Figure
4). The plateau phase of cell growth occurred at >2x106 cells/ml. Viability, as

measured by trypan blue exclusion, was >98% in log-phase cells but
progressively declined to 40% by 6 days without reseeding. The mean
population doubling time for the log-phase cells under these conditions was
11 hrs. For all further experiments L1210 cells were maintained in
exponential, or log phase, growth by reseeding every 2 days into fresh
medium and with fresh drug, where appropriate.
Copy Number of MtDNA in Untreated L1210 Cells
In order to reliably and accurately assay the mtDNA content of L1210
cells, we found it necessary to routinely incorporate an external (and
occasionally, an internal) standard in the dot blot procedure. The plasmid
pSP64, which had served as the vector for cloning of the mouse mtDNA, was
used as the standard because 1) its quantity (in terms of DNA) is easy to
assess and 2) it hybridizes to the radiolabeiied, nick-translated probe because
the probe contains the appropriate complementary sequences. This meant








-C -
0 -

> 0


>< V
in n3



I 1O

ILU/siuno Ilao

that only one probe (containing both pSP64 and mtDNA sequences) was
required for hybridization, a procedure that minimized variation between the
degree of hybridization of the standard and the mtDNA. A standard curve (ng
DNA vs cpm's [counts per minute]) was generated for each dot blot
experiment. The copy number for mtDNA in L1210 cells was calculated from
this experimental value for amount of mtDNA in a given number of cells and
the molecular weight (107 daltons) of murine mtDNA.
Certain other assumptions were validated in development of this dot
blot. First, the nick-translated probe (containing both plasmid and mouse
mtDNA) hybridized equivalently to both the plasmid standard and cellular
mtDNA, a necessary condition for direct comparison of the counts. Second,
we confirmed that the cell lysate ~r se did not interfere with hybridization. It
was found that varying mixtures of cell lysate and plasmid yield the expected
additive value for DNA which hybridized to the probe. And, finally,
radiolabelled, nick-translated pSP64 plasmid itself did not hybridize with
cellular DNA.
For determination of the mtDNA copy number, known quantities (1-6
ng) of plasmid and known aliquots (1.0-6.0x105) of L1210 cells were added to

a dot blot apparatus and hybridized with a [35S]-labeled full-length murine
mtDNA probe inserted into pSP64. The extent of hybridization of probe to
mtDNA was proportional to aliquot size at least over the range of 1.0x105 to
6.0x105 cells (Figure 5). This asynchronous population of L1210 cells
contained 1450 copies of mtDNA per cell; the mean standard deviation of 3
experiments was 13%.

o W :
tO .1 E
CZ a,

" m 1E

ra < xZ

.C O

<-O 0-


Eo 0)

0 cz
3o c E .2

T .C >
N C a)

.c c a
E o

T ^ C '
^3co y, -o
I r-
nl | ^S
h- oro i y





\ i I 0

co co J c oN

(Bu) VN(Ili 10 lunowuv

Effects of Polvamine Analogs on Cell Division and MtDNA Accumulation

Dose-Ranging Experiments with MGBG

Concentration response for MGBG. Because the intent of these

experiments was to use MGBG to modulate the mtDNA content of L1210 cells

without causing overt cell toxicity, an MGBG concentration response
experiment was done to define the concentration range at which MGBG

inhibited mtDNA accumulation without causing a substantial decrease in cell
viablity. The initial experiment encompassed MGBG concentrations from 0.1

to 100 gM in the medium. Cell counts and viability were determined daily for

two days.

Cell growth over this 2-day period was inhibited by MGBG in a

concentration-dependent manner (Figure 6). After 2 days of incubation with
MGBG, aliquots of cells from each treatment were lysed and the mtDNA

quantitated. When the concentration of MGBG in the medium was <1.0 pM,
there was no effect on the content of mtDNA per cell (Table 1). At MGBG
concentrations 23.3 gIM, mtDNA content per cell was decreased compared to
control, but cell viability was greatly reduced, especially at MGBG
concentrations >10 giM. Furthermore, 55% of cells treated with 10 jIM MGBG

for 2 days had lost the ability to form clones (compared to untreated controls).
The IC50's of MGBG for inhibiting cell growth at 24 and 48 hrs of

treatment were 10 and 2 jIM, respectively.
Decreased mtDNA content at 1.8 to 3.3 uM MGBG. Based on these
previous results, the MGBG concentration range was narrowed to 1.0 to 3.3
jIM for all further experiments. These concentrations provided optimal
differentiation of the effects of MGBG on cell growth and mtDNA, with minimal

overt cellular toxicity. Although MGBG had discernible effects at

concentrations 1.3 JM (see below), attention will first be directed to results


o E


0 .1- ,
- o


c- o


0 )


z a


U 00 C- i- C v-




ilU/suno3o lao

Table 1. Effects of MGBG on the relative amount of mtDNA/cell and the
percent viability in L1210 cells exposed to 0-100 W.M MGBG for 2

Treatment MtDNA/Cell % Cell Viability
(pM MGBG) (% Control)

Control 100 99

0.10 105 98

0.33 94 99

1.0 95 98

3.3 23 96

10 26 80

33 75

100 60

The mtDNA was determined 8 times and the cell viability once for each
experiment; each point represents the mean value of 2 different experiments.

*not done.

obtained with MGBG in the range of 1.8 to 3.3 p.M, concentrations at which

there are consistent and striking effects on mtDNA accumulation. Although the
cells were exposed in these experiments to MGBG for 18 days, our initial focus
involves the first 3 days of MGBG treatment because mtDNA synthesis
became resistant to MGBG after that time. This phenomenon will be explored

in more detail in a later section.

Cells were incubated with 1.8 to 3.3 gM MGBG and reseeded every 2
days with fresh medium and drug. At these concentrations of MGBG, cell
growth was inhibited in a concentration-dependent manner, i.e. the greater the
dose, the greater the degree of inhibition (Figure 7). However, this difference
was small for the first 3 days of exposure. Conversely, the effect of MGBG on
mtDNA accumulation did not show a gradation in degree of inhibition, but
rather in the duration of inhibition (Figure 8). Incubation of L1210 cells with

1.8, 2.5, or 3.3 gpM MGBG prevented accumulation of mtDNA for 1, 2, and 3
days, respectively.

Because cell growth continued at a rate which was relatively faster than
the accumulation of mtDNA, a decrease in the amount of mtDNA per cell can

be anticipated with each cell division. The amount of mtDNA per cell did
decrease progressively to only 10% of the control value upon treatment with

2.5 or 3.3 gM MGBG (Figure 9). The time course of this decrease was
consistent with a complete inhibition of mtDNA accumulation coupled with
continued cell division (see theoretical curve in Figure 9), i.e., the amount of
mtDNA per cell was halved with each division.

Determination of mtDNA degradation by MGBG. The inhibition of
mtDNA accumulation by MGBG may reflect inhibition of synthesis, increased
degradation, or both. Consequently, mtDNA was isolated from MGBG treated
cells to determine if there was increased mtDNA damage relative to control.


Lm c ca

co o

0 X

> u)

-0 )D (D

= 0 a.

%- CL

0a) =0)
> 0)


E o)c


E :3-E
=3 C?

slunoo Ila aGA!lelnWuno

om. =L .!.
c co tO (n
0 CM M

52 2
co c co

0) 2
(0 C C



-U) 0' )

x -ci .

2D a) a


0 CD

:3 E -u

.oo .





[(llaN3NaQlu%) X slunoo 113]

uo!elnLun3ooe VNQIIW

CO Li Ci
V- 04 ml

Or -


o $


x "-- 0
cmv 0
a,)i E c"
O >

0 o -
o o
0 ) cE


7, a)


r- x
I--0o = 0

=. =a aR
C- l qU )

(IoJIuo3 %) Ilao/VNal/V

Mitochondrial DNA from equal numbers of control or MGBG treated cells was
isolated and separated into Forms I, II, or III (supercoiled, nicked circular, or

linear, respectively) by gel electrophoresis. The gel was transferred to a
nitrocellulose filter, and the filter was hybridized with a [35S]-labeled mtDNA


Autoradiography of the blot showed that there was no substantial
change in the ratio of intact and nicked forms upon exposure of cells to 1.0 or

3.3 .LM MGBG. However, the total amount of mtDNA at 1.0 gM had decreased
and at 3.3 l.M this decrease was marked. At a higher dose (10 giM) than those

used in the mtDNA accumulation experiments, there was a decrease in the

intact form and a corresponding increase in the linear form of mtDNA.
Interestingly, however, the total content of mtDNA was more than at the 3.3 IM

treatment. This suggests that damaged mtDNA was not degraded and that
cell growth was so inhibited at 10 gM MGBG that the mtDNA copy number was
diluted less than at 3.3 P.M. Taken together, these results are compatible with
the hypothesis that the decrease in mtDNA content, at least in the 1.0 to 3.3

pM range, reflected a lack of synthesis rather than increased degradation.
Effects of Diethylspermine (DES) on Cell Growth and MtDNA

We tested another polyamine analog to determine if it might have
similar effects on mtDNA accumulation and cell growth. This work was done

in collaboration with Dr. Ray Bergeron and Mike Ingeno in the College of
Pharmacy. Diethylspermine (DES, a gift of R. Bergeron, is spermine with ethyl
substituents on each of the 2 terminal nitrogen atoms) was made to a stock
solution of 1 mM. The L1210 cells were reseeded with fresh medium and drug
every 2 days, and incubated with 0, 0.1, 1.0, or 10 pM DES for 6 days. The
growth rates of cells exposed to 0.1 p.M DES were the same as control (data

not shown). Cells exposed to 1.0 and 10 p.M DES for 1 day had normal

growth, but their growth was inhibited after 2 days, and was slowest from 4-6
days (Figure 10). Cell viability (as measured by trypan blue exclusion) was

>90% over the 6-day period.
The mtDNA accumulation was not affected in cells exposed to 0.1 iM
DES (data not shown). However, the mtDNA accumulation of cells exposed to
1.0 or 10 g.M DES was substantially inhibited and showed no signs of

recovery during the 6-day exposure to DES (Figure 11). Whereas the cells

exposed to 0.1 gM DES had a normal content of mtDNA, the mtDNA content of
cells exposed to 1.0 or 10 pM DES decreased to 10-15% of control by 3 days
(Figure 12). The mtDNA content remained at this low level through day 6 of
DES exposure.
Effects of Low Doses of MGBG

Increased mtDNA content at 1.0 to 1.3 uM MGBG. The results obtained
with MGBG concentrations <1.8 gpM were interesting. After 1 day of treatment
with 1.0 or 1.3 gpM MGBG, there was a slight inhibition of mtDNA accumulation

relative to control (Figure 13), in general agreement with the decreasing
duration of inhibition with decreasing MGBG concentration (Figure 8).

Thereafter, the rate of mtDNA accumulation equaled that of control. The rate

of cell growth was slightly inhibited by 24 hrs but, unlike mtDNA accumulation,
continued to be slightly inhibited throughout the 6-day exposure (Figure 14).

By dot blot analysis, 1.0 and 1.3 gM MGBG caused a 12 and 25% decrease,

respectively, in the content of mtDNA per cell at 24 hrs (Figure 15). However,
after this initial decrease, the amount of mtDNA per cell increased and by 4
days actually exceeded the control value. This increase resulted from a
normal rate of mtDNA accumulation in cells that were slightly growth-inhibited.
When the mtDNA content per cell was followed for 18 days at these

concentrations of MGBG, the mtDNA copy number never exceeded 155% of

nr r a)
rt .C 0

6 a)U

o =,
CZU a)
C C4-

- a


c -o
C) >"0

S3 o

L- L-

a) a)

c C

: co c

I6 -c

0%0- c2a

0 U

sluno3o Iao GAilelnwuno


C n

LU -0


75 >.


o ^


Z -0



a >

C -


E (D



-= 0

-0 &


o o




0 o


[(lla0VNaOlw%) X s uno llao3]

uo!lelnwnooe VNGIIAI

_0 cc

o 2o
0 En

CL) 0
x C.-* Ca

0CO t*-

r- a)0
-> o -
cu a)V

co, X.

7- C
o EEn-
Ea -c


0 (0 0

4- o E 0



bS- 2-






0I 0 UO ) l C4

(loiAuoo %) IaO/VNaiNw

(3 C\J
,.a- .


o c-a


C0 C
a --CC

=- o

o E

E-o o
n0 -C 1 -

L- t,-
n-1 a) -0 CL

S(n Z E





1 I
S0 0 0

1 g '" 1-

[(ll8O/VNalw%) x slunoo IioO]

uoilelnwnooB VNaGI/
uoiiejnlunooe VN(IMV


c E
E' QE -

E cL



"0- CZ


) 0
0 | m
c E

0 6.C

-, Q0
m 0 0


-= =L

u V-:



sluno3 I11O a!ljelnLuno

0 -C z
S- (
o > u 1-

o 0

(DO. co


o -c
o) E

0- 0 -


CZ (D f

cn1 I

S0 -CE -

) c CO
a 0 4. o
z .c

CO- a a0)

H- 0 0x



=L =:L
0 cT

a a a a
c(' 0+c

(IOJiuoo %) IlaO/VNaIW,

0 0 0 0
Go T 04

control. Thus, even though the rates of mtDNA accumulation and cell growth
at low concentrations of MGBG were similar, there were subtle differences that

caused the mtDNA content per cell to vary between 75% of control early in
exposure to 155% of control after several days.
Effects on mtDNA of decreased cell growth due to serum depletion.

Because of the unexpected increase in mtDNA per cell following incubation
with low concentrations of MGBG, an experiment was done to determine if this
increase was a nonspecific effect of decreased cell growth. If mtDNA
replication and cell division are not tightly coupled, then it can be

hypothesized that an increased mtDNA copy number may result from any
condition which causes decreased cell growth. To test this idea, mtDNA
content was measured in L1210 cells incubated in media containing limiting
amounts of serum. Decreases in serum concentration slowed cell growth in a

proportional way; the mtDNA content per cell at 3 days, however, was not
different from control (Figure 16).
Resistance to MGBG and Recovery of MtDNA Synthesis and Copy Number
Resistance to MGBG

Resistance of mtDNA synthesis to MGBG. As mentioned earlier, mtDNA

synthesis in L1210 cells resumed 1, 2, or 3 days after the start of exposure to

1.8, 2.5, or 3.3 iM MGBG, respectively. This recovery occurred despite the
continued presence of drug in the medium, and suggested that the cell was
able to overcome the initial effects of MGBG on mtDNA accumulation. In
addition, mtDNA synthesis, once resumed, occurred at the same rate,
regardless of dose of MGBG (Figure 17). Surprisingly, this rate of mtDNA
accumulation was virtually equal to that seen in untreated, growing cells--in
spite of the fact that the treated cells exhibited a decreased growth rate over
this time period (Figure 18).


o3 o) c c

mo co :

0 -0

o -

=- -

E -oL

.z a)
coa E Q E



4 -


: 3

I )



E n

1 0

- C.



0 0
%- 0


a. -
u o



^n >





N oo N I














) o

I > 0C

) 0

, --
a 0

"_ .

a) .

- oa


< 0 (

x x

- o

-o (a


= ::


0 00

[(lleO/VNGlu%) X slunoo 11aO]

uoileinujnooe VNOiw


Cd (0


0 0 0
0 0 0 0
1~o 1

[(I~~Iv~~w% x snoo 103

UO!I~fl~lO~ NGW

0 0c 0


co *-
WL a)

0 ,n "n

S0 I0
-o 4-


m > > o

0 0 c Q.

-< 0 0-

a a )
7.. 0-
a) C"
-c -

>z aC: C
CD CL 00n

U.. ca 0o

E .


ic. =L =
4- CM
e c(* c

o rtD

slunoo IlaO aGIejlnmuno

In contrast to the quantal (on/off) effect of MGBG on mtDNA
accumulation, cell growth exhibited a graded response to varying doses of

MGBG (Figure 18). Cell growth was inhibited in a concentration-dependent
and time-of-treatment-dependent manner.
The net effect of these different rates of mtDNA accumulation and cell
growth was to modulate the mtDNA copy number of these cells (Figure 19).

As noted earlier, there was an initial decrease in the copy number of mtDNA
due to complete inhibition of mtDNA accumulation for up to 3 days. After that,
the copy number increased towards the control value because of the
combined effects of the recovery of mtDNA synthesis and a marked decrease
in the cell growth rate.
Effects of transfer of treated cells to a MGBG-free medium. This result,
in which mtDNA accumulation recovered, but at a rate which never occurred

faster than the 12-hr doubling time of control cells, was unexpected in view of

the reported observation that the mtDNA genome can replicate in vitro in
about 1 hr (214). One possibility was that this 12-hr doubling time
coincidentally represented a continued, but lesser, inhibition of mtDNA

replication by MGBG. To rule out this possibility, a study was done in which
L1210 cells were treated with MGBG and then transferred to drug-free medium

at the time at which recovery of mtDNA synthesis regularly occurred. If the
drug continued to cause partial inhibition of mtDNA synthesis, then those cells
transferred to drug-free medium should have a faster rate of mtDNA
accumulation than those remaining in drug.
The L1210 cells were treated with 3.3 gM MGBG for 4 days. At that

time, haif the cells were maintained at 3.3 pM while the other half were
transferred to drug-free medium for an additional 4 days. Cell counts and dot
blot analysis of mtDNA content were done during the initial exposure to



SM 0 .-
CL0 0>
-CO 0
a) CC)

o- oH
>= 0

a) Za)

0" z 03
-- -
0 -C co
< a)C> a

U) a) a)4-.
0 oa Eo E

a Go W0 "

|0 c c |
l) a) D ( 0

Z -oc < c

.E E E co
E -" -c
1- (D W

F- 0 1" 0) a ) ,





0 0 l 0 I 0 0