19_05

Chapter 19: Pharmacology of Cancer Chemotherapy
19.5: Antimetabolites

Edward Chu
Augusto C. Mota
Miklos C. Fogarasi

Cancer: Principles and Practice of Oncology, 6th Edition
Published by Lippincott Williams & Wilkins, Copyright 2001

Methotrexate

Aminopterin was the first antimetabolite to demonstrate clinical
activity in the treatment of patients with malignancy. This antifolate
analogue was used to induce remissions in children with acute leukemia
in the 1940s. [ref: 1] Aminopterin has since been replaced by
methotrexate (MTX), the 4-amino, 10-methyl analogue of folic acid. MTX
remains the most widely used antifolate in cancer chemotherapy, with
documented activity against a wide range of human malignancies,
including leukemia, breast cancer, colorectal cancer, head and neck
cancer, lymphoma, osteogenic sarcoma, urothelial cancer, and
choriocarcinoma. Antifolates have also been used to treat a host of
nonmalignant disorders, including psoriasis, rheumatoid arthritis,
graft-versus-host disease, bacterial and plasmodial infections, and
parasitic infections associated with the acquired immunodeficiency
syndrome. [ref: 2] This class of agents represents the best-
characterized and most versatile of all chemotherapeutic drugs in
current clinical use.

Mechanism of Action

MTX is a tight-binding inhibitor of dihydrofolate reductase (DHFR), a
critical enzyme in folate metabolism (Fig. 19.5_1). [ref: 2] The
importance of DHFR stems from its role in maintaining the intracellular
folate pool in its fully reduced form as tetrahydrofolates. These
compounds serve as one-carbon carriers required for the synthesis of
thymidine-5'-monophosphate (thymidylate), purine nucleotides, and
certain amino acids. Thymidylate synthase (TS) catalyzes the formation
of thymidine-5'-monophosphate from 2'-deoxyuridine 5'monophosphate
(deoxyuridylate, dUMP) (Fig. 19.5_2). This reaction uses 5,10-
methylenetetrahydrofolate as a methyl donor and results in the
oxidation of the reduced folate to dihydrofolate. The activity of the
TS reaction thus creates the requirement for DHFR to maintain the
intracellular reduced folate pool needed for one-carbon transfer
reactions. The reduced folate, 10-formyltetrahydrofolate, serves as a
substrate for two folate-dependent enzymes of de novo purine synthesis,
glycinamide ribonucleotide (GAR) transformylase and aminoimidazole
carboxamide ribonucleotide transformylase. An intact DHFR pathway is
therefore necessary for continued de novo thymidylate and purine
nucleotide biosynthesis.
The precise mechanism by which MTX produces metabolic inhibition
remains a subject of ongoing debate. The long-held view has been that
inhibition of DHFR results in an accumulation of oxidized folates at
the expense of reduced folates owing to the continued synthetic
function of TS. Ultimate depletion of the required reduced folates
would result in cessation of de novo thymidylate and purine
biosynthesis as well as inhibition of protein synthesis. However,
several investigators have demonstrated that after exposure of
malignant cells to inhibitory concentrations of MTX, intracellular
reduced folates are depleted by only 50% to 70%, a level presumably
insufficient to account for the observed inhibition of DNA synthesis.
[ref: 2,3] Additional metabolic effects of MTX result from its
transformation to polyglutamate forms (see Fig. 19.5_1). MTX and
physiologic folate polyglutamates are formed by the enzyme
folylpolyglutamyl synthetase, which adds up to five to seven glutamyl
groups in a gamma-peptide linkage. Polyglutamation is a time- and
concentration-dependent process that occurs in tumor cells and, to a
lesser extent, in normal tissues. [ref: 4] These polyglutamate
metabolites have a prolonged intracellular half-life and allow for
prolonged drug action in malignant cells. The relative difference in
polyglutamate formation in normal versus malignant cells may account
for the selective activity of the drug. As much as 80% of MTX found in
malignant tissues is in the polyglutamated forms, and these metabolites
are potent, direct inhibitors of several folate-dependent enzymes,
including DHFR, TS, and aminoimidazole carboxamide ribonucleotide and
GAR transformylases. [ref: 1-3,5,6] Thus, metabolic inhibition
resulting from MTX exposure is a multifactorial process and may depend
on several factors, including partial depletion of reduced folates and
direct inhibition of folate-dependent enzymes by the polyglutamates of
both MTX and dihydrofolate that accumulate after inhibition of DHFR.
The precise mechanism by which MTX induces cytotoxicity remains an
area of continued investigation. MTX-induced depletion of thymidine
triphosphate (dTTP) and purine nucleotides interferes with the cellular
capacity to repair DNA, resulting in DNA strand breaks. [ref: 7,8]
Furthermore, the need for repair is accentuated by an intracellular
accumulation of dUMP resulting from the inhibitory effects of MTX on
TS. dUMP can be converted to the triphosphate nucleotide form (dUTP),
which is then incorporated into DNA, resulting in inhibition of chain
elongation and DNA synthesis. Excision repair of the DNA containing
these misincorporated dUTP moieties by the enzyme uracil DNA
glycosylase results in further DNA fragmentation.
Novel mechanisms by which MTX may exert its cytotoxic action have
been described. Treatment with MTX results in a significant dose-
dependent reduction in methionine synthase enzyme activity. [ref: 9,10]
This enzyme catalyzes the folate-dependent reaction in which 5-
methyltetrahydrofolate serves as a critical one-carbon carrier methyl
donor and mediates the conversion of homocysteine to methionine. Thus,
inhibition of methionine synthase leads to inhibition of a number of
key downstream pathways, including transmethylation reactions,
polyamine biosynthesis, protein synthesis, or all three. [ref: 11]
Treatment of cultured Ehrlich ascites tumor cells with MTX resulted in
up to a 3.5-fold increase in 5'-phosphoribosyl-1-pyrophosphate (PRPP)
levels, which was associated with a significant suppression in the rate
of glucose transport. Coadministration of MTX and hypoxanthine
completely protected against the growth-inhibitory action of MTX and
reversed the effect on PRPP production and on the rate of glucose
transport. [ref: 12] Thus, MTX may exert its anticancer effect, in
part, through inhibition of critical glucose transport mechanisms,
thereby starving the cancer cell of essential nutrients required to
maintain cellular metabolism and growth.
MTX is most active against rapidly proliferating cells, as its
cytotoxic effects occur primarily during the S phase of the cell cycle.
During longer periods of drug exposure, a higher fraction of cells can
enter the S phase of the cell cycle, resulting in greater cell kill. In
addition, MTX polyglutamate formation is substantially enhanced with
longer periods of drug exposure, thereby increasing cytotoxicity. The
cytotoxic effects of MTX are also greater with increasing drug
concentrations. Therefore, MTX cytotoxicity is highly dependent on the
absolute drug concentration and the duration of drug exposure.
MTX enters cells by the same active transport mechanisms used by
physiologic reduced folates. In general, intracellular drug
concentrations reach steady state in less than 30 minutes. Folate
transport is a complex process with at least two carrier-mediated,
energy-dependent mechanisms existing in mammalian cells. The first is
the classic reduced folate carrier (RFC) system that has a relatively
low affinity for MTX and reduced folates such as leucovorin (LV), with
affinity constants in the micromolar range. [ref: 2,13-17] The RFC
system has a large capacity and is primarily responsible for transport
of MTX into cells at pharmacologic concentrations. The human RFC gene
has been mapped to the long arm of chromosome 1, and it encodes a
protein with a predicted molecular weight of 59 to 68 kD.
A second folate transport system involves a high-affinity, membrane-
bound folate receptor-binding protein with affinity constants for
reduced folates and folic acid in the nanomolar range. [ref: 18,19] MTX
is a relatively poor substrate for this folate-binding protein, and its
affinity is 10- to 30-fold lower than that of physiologic reduced
folates. The human folate receptor protein (FR) is a 38- to 40-kD
glycoprotein bound to the cellular membrane via a C-terminal,
glycosylphosphatidylinositol tail. [ref: 20] It is expressed on the
surface of various normal tissues, including human placenta, choroid
plexus, renal tubules, and fallopian tubes. Of note, this receptor is
also highly expressed on the surface of a number of epithelial tumors,
including ovarian cancer, but not on normal ovarian tissue, making it
an attractive target for antigen-directed anticancer therapies. [ref:
21] Alterations in the tissue expression of FR can be induced by
changes in the exogenous folate concentration or by alteration in
normal physiology such as in pregnancy. At least three different
isoforms of the human FR have been described to date, and they are
classified as FR-alpha, FR-beta, and FR-gamma. [ref: 17,22] These
isoforms have unique folate-binding affinities and variable expression
in specific tissues. FR-alpha is highly expressed in human epithelial
tissues and in some cancers such as ovarian cancer, whereas FR-beta is
expressed in human placenta and other nonepithelial tissues. Although
the FR-alpha and FR-beta isoforms share 70% to 80% amino acid sequence
homology, they differ significantly in their respective affinities and
stereospecificities for reduced folates. Human FR-gamma lacks a
glycosylphosphatidylinositol tail, and this isoform most likely
represents a secretory protein.
An additional MTX transport system has been described in murine L1210
leukemic cells that is completely distinct from either the RFC or the
FR systems. [ref: 23,24] Further studies are under way to characterize
the role of this transporter as a determinant of MTX cytotoxicity. It
is likely that the relative function of each of these distinct
transport systems depends on the extracellular folate concentration,
and their expression may vary significantly among different cell lines.
Their interrelationship and role in MTX transport remains an active
area of research. Nonclassic antifolate compounds, such as trimetrexate
and trimethoprim, do not rely on specific transport systems for
cellular entry. Such analogues are active against various malignant
cell lines resistant to MTX on the basis of decreased transport
capacity. In addition to these transport systems, two energy-dependent
MTX efflux transport systems have been described in murine leukemic
L1210 cells using inside out membrane vesicles. [ref: 25] These two
systems appear to be functionally distinct and sensitive to a different
range of chemical inhibitors. The major efflux transporter is identical
to a glutathione conjugate membrane pump and accounts for nearly 70% of
MTX efflux.
Reduced folates, such as 5-formyltetrahydrofolate (LV), prevent,
rescue, or both prevent and rescue cells from the toxic effects of MTX.
The predominant species of reduced folate in human plasma, 5-
methyltetrahydrofolate, circulates with levels in the range of 5 to 50
nM, a concentration inadequate to rescue cells. Administration of
appropriate doses of LV after high-dose MTX therapy can prevent
toxicity to the bone marrow and gastrointestinal epithelium, the two
most rapidly dividing cells in the body. The dose of LV required to
rescue normal tissues is dependent on the antifolate concentration at
the time of antidote administration. [ref: 26] The competitive nature
of this rescue suggests that LV does more than simply replete
intracellular reduced folate pools. LV is converted to intracellular
folates that can compete with both MTX and dihydrofolate polyglutamates
to overcome the inhibition of TS and de novo purine synthesis. In
addition, MTX and reduced folates compete with one another for
transport into cells and for subsequent intracellular polyglutamation.
Presumably, rescue from MTX-associated metabolic inhibition occurs only
when adequate levels of dihydrofolate have accumulated after LV
administration. The administration of exogenous thymidine may also be
used to decrease MTX toxicity. This approach appears to be less
effective than LV, as inhibition of the de novo purine pathway by MTX
remains unaffected by its use. Administration of the recombinant
bacterial enzyme carboxypeptidase G(2), which hydrolyzes MTX to
inactive metabolites, is currently undergoing clinical testing as an
alternative strategy to rescue from high-dose MTX therapy. [ref: 27]

Mechanisms of Resistance

The development of cellular resistance to MTX remains a major obstacle
to its effective clinical use. In experimental systems, resistance to
antifolates may result from several mechanisms, including an alteration
in antifolate transport due to either a defect in the RFC or FR
systems, [ref: 28-31] decreased capacity to polyglutamate MTX through
either decreased expression of folylpolyglutamyl synthetase or
increased expression of the catabolic enzyme gamma-glutamyl hydrolase,
[ref: 32-35] and alterations in the target enzyme DHFR through either
increased expression of the wild-type protein or overexpression of a
mutant protein with reduced binding affinity for MTX. [ref: 2,36,37]
Amplification of the DHFR gene is one of the most common forms of MTX
resistance observed in experimental systems. [ref: 2] The amplified
gene may be stably integrated into chromosomal DNA in the form of a
homogeneously staining region, or it may exist in extrachromosomal
pieces of DNA known as double-minute chromosomes. [ref: 38]
Homogeneously staining region-mediated gene amplification is associated
with the development of stable resistance to MTX. In contrast, double-
minute chromosomes are unequally distributed during cell division, and,
in the absence of continued selective pressure of MTX, cells eventually
revert to a sensitive phenotype with wild-type levels of DHFR
expression. It was shown that resistant human leukemic HL-60 cells
coamplify both DHFR and hMSH3, the human mutS homologue 3 gene. [ref:
39] Overproduction of hMSH3 results in virtually complete sequestration
of the nuclear hMSH2 mismatch repair (MMR) protein. The net effect of
this protein-protein interaction is a marked reduction in the
efficiency of base-base MMR. As MMR deficiency has been implicated as a
potential mechanism of resistance to the platinum analogues cisplatin
and carboplatin, DNA methylating agents, and doxorubicin (Adriamycin),
it is conceivable that this same resistance phenotype may contribute to
the development of resistance to MTX and other antifolate analogues.
An alternative mechanism of resistance has been ascribed to mutations
that result in a DHFR protein product with an altered binding affinity
for MTX. There is evidence that naturally occurring DHFR alleles with
differing affinities to MTX may exist in cells and provide a mechanism
for the rapid emergence of MTX resistance. [ref: 40] In several in
vitro experimental model systems, the levels of DHFR enzyme activity
acutely increase after exposure to MTX, other antifolate analogue
compounds, or both. [ref: 41,42] This acute induction of DHFR in
response to drug exposure is mediated, in part, by a translational
regulatory mechanism. [ref: 43] DHFR protein, in its unbound or free
state, is capable of specifically repressing the translation of its own
messenger RNA (mRNA). However, when DHFR protein is bound to an
antifolate inhibitor, it is unable to repress DHFR mRNA translation,
and the rate of new DHFR protein synthesis increases. [ref: 43,44]
Thus, induction of DHFR may represent a clinically relevant mechanism
for the acute development of cellular drug resistance.
Apoptosis, the process of programmed cell death, is a critical event
during normal development and in the pathogenesis of several disease
states, including cancer, autoimmune disorders, viral infection, and
neurodegenerative diseases. Bcl-2 can repress cell death triggered by a
wide array of stimuli, including chemotherapy and gamma-irradiation.
Bcl-XL, a structural homologue of Bcl-2, has also been shown to provide
protection against a wide range of anticancer agents. These prosurvival
proteins presumably act at some common final step to prevent or
overcome the cell death pathway induced by various anticancer agents.
It has been shown that murine lymphoid FL5.12 cells transduced with
Bcl-XL as compared with another antiapoptotic gene Bcl-2, become
resistant to the cytotoxic effects of MTX. [ref: 45] Thus, the
expression of Bcl-XL may represent an important indicator for
predicting chemosensitivity to MTX and other antifolate analogues.
Despite many years of active investigation, the relative contribution
of each of these mechanisms as a determinant of MTX resistance remains
unclear. However, there is growing evidence to support the concept that
the emergence of MTX resistance, in the clinical setting, is a
multifactorial process. In fact, DHFR gene amplification, defective
transport, and decreased polyglutamate formation have all been observed
in clinical specimens taken from MTX-resistant patients. [ref: 46-49]

Clinical Pharmacology and Pharmacokinetics

Accurate monitoring of MTX concentrations in plasma is essential for
the safe and optimal use of this agent in cancer chemotherapy,
particularly with high-dose regimens. At least four methods are
presently available for the clinical monitoring of MTX drug levels,
including the DHFR enzyme inhibition assay, a competitive protein-
binding assay, a fluorescence-polarization radioimmunoassay technique,
and an enzyme-multiplied immunoassay system. [ref: 2]
The absorption of oral MTX is saturable and erratic at higher doses,
such that oral doses should be kept to less than 25 mg/m**2. The drug
is usually administered intravenously. The volume of distribution of
MTX approaches that of total body water, and approximately 60% of the
drug is bound to serum albumin at pharmacologic drug concentrations.
[ref: 2] Although plasma pharmacokinetics are variable, MTX metabolism
generally follows a three-phase pattern. The initial distribution
phase, which lasts for only a few minutes, is followed by a second
phase lasting 12 to 24 hours, during which time the drug is eliminated
with a half-life of 2 to 3 hours. The final phase of drug clearance has
a half-life of 8 to 10 hours. The last two phases of drug elimination
are considerably lengthened in patients with renal dysfunction. There
is substantial evidence that a more rapid systemic clearance of drug is
associated with a high risk of relapse in children receiving MTX for
maintenance therapy of acute lymphocytic leukemia.
The distribution of MTX into third-space fluid collections, such as
pleural effusions and ascitic fluid, can substantially alter MTX
pharmacokinetics. The slow release of accumulated MTX from these third
spaces over time prolongs the terminal half-life of the drug, leading
to potentially increased clinical toxicity. [ref: 50] Although no
strict guidelines exist for the treatment of patients with ascites or
pleural effusions, it is advisable to evacuate these fluid collections
before treatment and monitor plasma drug concentrations closely. In
addition, patients with bladder cancer who have undergone cystectomy
and ileal conduit loop diversion may experience a significant increase
in toxicity secondary to MTX treatment. [ref: 51] Thus, caution should
be given when beginning therapy with MTX in this particular subset of
patients.
Elimination of MTX occurs primarily through renal excretion. MTX is
filtered by the glomerulus and is actively secreted in the proximal
tubule. Renal clearance usually equals or exceeds creatinine clearance.
However, rates of drug clearance may vary widely, and they do not
precisely parallel renal function. Renal excretion of MTX is inhibited
by probenecid, penicillins, cephalosporins, aspirin, and nonsteroidal
antiinflammatory drugs. [ref: 2] The combination of MTX and
nonsteroidal antiinflammatory drugs has been associated with severe
toxicity in patients receiving high-dose MTX. Patients with impaired
renal function (creatinine clearance less than 60 mL/min) should not be
treated with high-dose MTX. Moreover, standard doses of MTX should be
reduced in proportion to reductions in creatinine clearance.
The introduction of high-dose MTX regimens led to the identification
of at least two MTX metabolites. 7-Hydroxymethotrexate (7-OH-MTX)
constitutes 20% to 46% of drug excreted in urine from 12 to 24 hours
after the start of a high-dose infusion. It is formed through the
action of aldehyde oxidase in the liver and is a weak inhibitor of
DHFR. 7-OH-MTX is a substrate for folylpolyglutamyl synthetase, and the
resulting polyglutamate metabolites are inhibitors of the folate-
dependent enzymes TS and aminoimidazole carboxamide ribonucleotide
transformylase, with a potency similar to that of MTX polyglutamates.
[ref: 52] A second metabolite, 2,4 diamino-N10-methyl pteroic acid
(DAMPA), a product of bacterial degradation of MTX in the gut lumen, is
inactive and constitutes approximately 25% of the excreted drug at 24
to 48 hours after drug infusion. The exact role of these metabolites in
producing MTX toxicity or enhancing therapeutic activity remains
uncertain.
Biliary excretion of MTX represents approximately 10% of overall MTX
drug clearance. [ref: 53] However, in the presence of renal
dysfunction, enterohepatic circulation may represent an important
pathway of drug elimination. Most MTX excreted in bile is reabsorbed as
intact drug, but an undefined fraction is converted by intestinal flora
to DAMPA. Intestinal binding of drug with oral charcoal or the anion-
exchange resin cholestyramine enhances nonrenal drug excretion. Given
the relatively minor role of biliary excretion in drug elimination, no
adjustments in MTX dose are necessary for patients with hepatic
dysfunction.

Schedules of Administration

The safe use of high-dose MTX with LV rescue requires a thorough
understanding of MTX pharmacokinetics. High-dose MTX therapy is used in
the treatment of high-grade lymphomas, osteogenic sarcoma, and acute
leukemia. These regimens use otherwise lethal infusions of MTX given
over 6 to 42 hours in doses of 500 mg/m**2 or higher. High-dose MTX can
be safely administered to patients provided that careful attention is
paid to intravenous fluid hydration, urinary alkalinization, plasma
drug level monitoring, and adequate administration of LV.
During infusion of high-dose MTX, rapid renal excretion results in
high urinary drug concentrations. Urinary MTX concentrations
approaching 10 mM exceed solubility, resulting in intratubular
precipitation and acute renal failure with potentially disastrous
consequences. This complication can be avoided by vigorous hydration (3
L fluid/m**2/24 hours, beginning 12 hours before infusion and
continuing for 36 hours), and urinary alkalinization to increase drug
solubility. Administration of MTX should not begin until urine flow is
100 mL/h and urine pH is 7 or greater, and these parameters should be
carefully monitored during the course of therapy. High-dose MTX therapy
should not be used in patients with impaired renal function (creatinine
clearance less than 60 mL/min).
Close monitoring of MTX plasma levels is essential for guiding the
duration and amount of LV required to prevent severe MTX-associated
toxicity. Given the competitive interaction between MTX and LV, the
dose of the rescue agent must be increased in proportion to the plasma
concentration of MTX. If the MTX level exceeds 0.5 uM 48 hours from the
start of the infusion, the LV dose may be adjusted to 15, 100, or 200
mg/m**2 every 6 hours for MTX levels of 0.5, 1.0, and 2.0 uM,
respectively. Drug levels should be rechecked every 24 hours and the LV
dose adjusted until the drug concentration is less than 50 nM. However,
clinicians should be aware that overzealous use of LV may counteract
the cytotoxic effects of MTX in tumor cells as well as in host cells.
[ref: 54] For this reason, it is important to use doses of LV that are
adequate but not excessive, so that normal but not tumor cells will be
rescued. In patients with delayed MTX excretion, LV is usually given
intravenously, because its oral bioavailability is decreased at total
doses higher than 40 mg.
Despite careful attention to detail, persistent elevations of plasma
MTX levels may sometimes occur. Plasma MTX levels higher than 10 uM at
48 hours are poorly rescued even with high doses of LV. Hemodialysis
and peritoneal dialysis are ineffective in removing MTX, with clearance
rates of only 35 to 40 mL/min. Experimental approaches to reduce toxic
levels of MTX include hemoperfusion over a charcoal column, oral
administration of activated charcoal or cholestyramine to increase
enterohepatic drug loss, and intravenous infusion of the degradative
enzyme carboxypeptidase G(2). [ref: 27,55]
MTX penetrates poorly into the cerebrospinal fluid (CSF), and CSF
levels are 30-fold lower than plasma levels at equilibrium. [ref: 56]
However, after high-dose MTX therapy, peak CSF levels greater than the
therapeutic threshold of 1 uM can be achieved. Systemic high-dose MTX
therapy has been used to prevent meningeal leukemia and lymphoma.
Intrathecal injection of MTX can also be used for prophylaxis. For
treatment of meningeal carcinomatosis, injection of MTX through an
indwelling Ommaya reservoir is recommended because drug administered
into the CSF via the lumbar space circulates poorly into the
ventricles, resulting in inadequate CSF drug levels. A total
intrathecal dose of 12 mg is advised for all persons older than 3 years
of age. In normal patients, the CSF half-life is approximately 12
hours, but it may be prolonged in patients with active meningeal
disease. Delayed clearance from the CSF has been associated with an
increased risk of MTX neurotoxicity.

Toxicity

The primary toxic effects of MTX therapy are myelosuppression and
gastrointestinal mucositis. The occurrence of these adverse effects and
other toxicities depends on the dose, schedule, and route of drug
administration. Mucositis usually appears 3 to 7 days after MTX therapy
and precedes the decrease in granulocyte and platelet count by several
days. Myelosuppression and mucositis are usually completely reversed
within 14 days, unless drug elimination mechanisms are impaired. In
patients with compromised renal function, even small doses of MTX may
result in serious toxicity.
MTX-induced nephrotoxicity is thought to result from the intratubular
precipitation of MTX and its metabolites, 7-OH-MTX and DAMPA, in acidic
urine. Antifolates may also exert a direct toxic effect on the renal
tubules. Vigorous hydration and urinary alkalinization have greatly
reduced the incidence of renal failure in high-dose regimens. [ref: 2]
MTX is associated with both acute and chronic hepatotoxicity. Acute
elevations in hepatic enzyme levels, as well as hyperbilirubinemia, are
often observed during high-dose therapy, but these usually return to
normal within 10 days. Chronic administration of daily oral MTX, as has
been used in the treatment of psoriasis, is associated with the
development of hepatic fibrosis in as many as 25% of patients.
Cirrhosis of the liver has also been described in this group.
Intermittent, weekly MTX therapy, rather than continuous daily
treatment, is associated with a lower incidence of hepatotoxicity.
Although the precise mechanism of MTX hepatotoxicity is not known,
liver biopsies of patients with drug-induced liver disease reveal
increased lipid deposition in the liver.
MTX causes a poorly defined, self-limited pneumonitis characterized
by fever, cough, and interstitial pulmonary infiltrates. Lung biopsies
have not revealed consistent pathologic findings. Although a
hypersensitivity reaction has been proposed as a possible explanation,
rechallenge with MTX does not uniformly result in a return of symptoms.
With the increasing use of chronic, low-dose MTX therapy for rheumatoid
arthritis, there is now a growing number of cases of MTX-associated
lung damage. No specific therapy for MTX pneumonitis is recommended
other than withholding MTX therapy during the acute episode.
Three distinct neurotoxic syndromes are associated with intrathecal
MTX therapy. [ref: 57] The most common syndrome is an acute chemical
arachnoiditis that arises immediately after intrathecal drug
administration. This syndrome is characterized by severe headaches,
nuchal rigidity, vomiting, fever, and an inflammatory cell infiltrate
in the CSF. A subacute form of neurotoxicity is seen in approximately
10% of patients and usually occurs after the third or fourth course of
intrathecal therapy. It is most common in adults with active meningeal
leukemia and consists of motor paralysis, cranial nerve palsies, and
seizures or coma, or both. A change in therapy is absolutely indicated,
because continued intrathecal MTX therapy may result in death. The
third syndrome is a chronic, demyelinating encephalopathy, typically
occurring in children months to years after receiving intrathecal MTX.
Patients present with dementia, limb spasticity, and, in advanced
cases, coma. Computed tomography scan reveals ventricular enlargement,
cortical thinning, and diffuse intracerebral calcifications. [ref: 83]
High-dose systemic MTX therapy is occasionally associated with an
acute, transient cerebral dysfunction with symptoms of paresis,
aphasia, and behavioral abnormalities, and seizures have also been
described in 4% to 15% of patients receiving high-dose MTX. [ref: 58]
Symptoms occur within 6 days of MTX treatment and usually completely
resolve within 48 to 72 hours. In addition, a chronic form of
neurotoxicity is manifested as an encephalopathy with dementia and
motor paresis developing in the second or third month after treatment.
At present, the underlying mechanism of CNS toxicity from MTX remains
unknown. There is no evidence to support the therapeutic use of LV in
patients who develop neurotoxic symptoms.
True anaphylactic reactions to MTX are exceedingly rare. There have
been a few reported cases of toxic skin erythema and desquamation of
the hands after high-dose MTX therapy. In men treated with high-dose
MTX, a reversible defect in spermatogenesis may occur. However, no
alterations in reproductive function have been reported in women
treated with MTX.

New Antifolates

Trimetrexate

In the 1990s, several new antifolates were developed in an attempt to
circumvent some of the known mechanisms of resistance to MTX, to target
folate-dependent enzymes other than DHFR, or both. The more lipid-
soluble quinazoline antifolate trimetrexate (TMTX, Neutrexin) differs
from MTX by not requiring the RFC system for cellular transport and by
lacking the potential for polyglutamation. Like MTX, it is a potent
inhibitor of DHFR and produces metabolic inhibition through mechanisms
similar to MTX. It was thought that malignant cells that had become
resistant to MTX, by virtue of either reduced membrane transport or
deficient polyglutamation, would remain sensitive to this antifolate
analogue. Because TMTX is not polyglutamated, its intracellular half-
life is shorter than MTX, necessitating frequent or continuous dosing
schedules. In contrast to MTX, TMTX can serve as a substrate for the P-
glycoprotein-associated efflux pump. As a result, cross-resistance to
TMTX and a host of natural products, including anthracyclines, taxanes,
and vinca alkaloids, can develop in multidrug resistant cancer cells
overexpressing the P170 glycoprotein.
TMTX has been tested using a variety of dose schedules. It is highly
protein bound (more than 90%) and cleared principally from the body by
hepatic metabolism. [ref: 59] The terminal plasma half-life of the drug
ranges from 12 to 20 hours. TMTX has been tested in the phase II
setting for the treatment of a variety of malignancies, principally
with a regimen of 8 to 12 mg/m**2 daily for 5 days every 3 weeks. [ref:
60,61] The dose-limiting toxicity has been myelosuppression. Other
toxicities include rash, mucositis, fever, nausea and vomiting, and
reversible transaminasemia. Antitumor responses have ranged from 13% to
26%, with the highest rates in head and neck (26%) and non-small cell
lung (19%) cancers. As a single agent, the compound has shown little
activity against gastrointestinal cancers. However, when used in
combination with 5-fluorouracil (5-FU) and LV, response rates on the
order of 30% to 35% have been reported in patients with previously
untreated metastatic colorectal cancer. [ref: 62]

Tomudex

Raltitrexed (ZD1694, Tomudex) is a quinazoline antifolate that is a
potent and specific inhibitor of TS. Like MTX, raltitrexed is
transported into the cell via the RFC. To exert its cytotoxic activity,
this compound needs to be metabolized to its higher polyglutamated
forms. The polyglutamates of raltitrexed are approximately 100-fold
more potent than the parent compound with regard to their affinity for
TS, and they exhibit prolonged intracellular retention. [ref: 63] The
principal mechanism of action of this compound is inhibition of TS,
resulting in depletion of key nucleotide precursors required for DNA
repair and synthesis. Mechanisms of resistance to raltitrexed include
reduced transport, decreased polyglutamation, and overexpression of the
target enzyme, TS. [ref: 64,65]
Raltitrexed has undergone phase I testing, and the recommended dosing
schedule is 3 mg/m**2 given as a 15-minute intravenous infusion every 3
weeks. [ref: 60,61] The drug is cleared from the body principally by
renal excretion, and its clearance follows a three-compartment
elimination model with a terminal half-life of 10 to 22 hours. The
major toxicities include an anorexia and fatigue syndrome, diarrhea,
myelosuppression, and reversible transaminasemia. Several phase II
clinical trials have investigated the activity of this drug in patients
with a wide spectrum of malignancies, including advanced colorectal,
breast, hepatocellular, platinum-resistant ovarian, non-small cell
lung, gastric, and pancreas cancers. [ref: 66] Response rates have
ranged from 14% for patients with pancreas cancer to 25% for those with
breast cancer. The largest reported experience has been for first-line
treatment of patients with advanced colorectal carcinoma. An overall
response rate of 26% was observed in a phase II trial involving 176
patients. Randomized, phase III trials to compare the efficacy of
raltitrexed versus LV-modulated 5-FU have been completed. [ref: 67,68]
Raltitrexed has comparable response rates and similar palliative and
survival benefits when compared with 5-FU/LV. Reduced toxicity was
observed with this novel antifolate analogue, however, and it was
associated with a more convenient administration schedule. This drug is
approved as first-line therapy for patients with advanced colorectal
cancer in Australia, Canada, and several European countries. Further
studies are ongoing to determine the role of this antifolate TS
inhibitor compound, either alone or in combination with other
anticancer agents, in the therapy of colorectal cancer in North
America.

Other Antifolates

Several pteroyl glutamate analogues have shown superior preclinical
activity compared with MTX, and this enhanced effect is thought to be
due to enhanced transport, more avid polyglutamation, or a unique site
of action. 10-Ethyl-5-deaza-aminopterin (Edatrexate), a potent
inhibitor of DHFR with enhanced transport and more efficient
polyglutamation relative to MTX, has demonstrated activity against non-
small cell lung cancer. [ref: 60,61,69] Mucositis is the dose-limiting
toxicity. 5,10-Dideazatetrahydrofolate (Lometrexol) is a new antifolate
that impairs de novo purine synthesis by virtue of its direct
inhibitory effects on GAR transformylase. Phase I studies have been
performed, and this agent is currently undergoing phase II clinical
investigation.
LY231514 was originally developed as a pure antifolate inhibitor of
TS. However, preclinical studies suggest that it is not entirely
specific for TS and that it inhibits other folate-dependent enzymes,
including DHFR and GAR transformylase. [ref: 60,61,70] Like
raltitrexed, this analogue requires polyglutamation for its potent
inhibitory effects on TS, and it uses the RFC for entry into cells.
This compound has been investigated in several phase I studies, and the
major toxicities include neutropenia, anorexia and fatigue syndrome,
gastrointestinal toxicity, and reversible transaminasemia. [ref: 71]
Promising antitumor activity has been noted in a wide variety of tumor
types, and phase II studies have confirmed activity against non-small
cell lung cancer and mesothelioma. [ref: 72]

5-Fluoropyrimidines

The rationale for the synthesis of the fluoropyrimidines stemmed from
the observation that rat hepatoma cells use uracil more efficiently
than normal rat intestinal mucosa. [ref: 73] This finding suggested
that uracil metabolism might represent a potential target for cancer
chemotherapy.
The chemical structures of the fluoropyrimidines in common clinical
practice are shown in Figure 19.5_3. 5-FU has a fluorine atom
substituted in place of hydrogen at the carbon-5 position of the
pyrimidine ring (see Fig. 19.5_3). The deoxyribonucleoside
derivative, 5-fluoro-2'-deoxyuridine (FUdR) is limited in its clinical
use because of its rapid degradation in normal and tumor tissues. It
has mainly been used for hepatic arterial infusions. Ftorafur and 5'-
deoxyfluorouridine represent two fluoropyrimidine analogues that have
been incorporated in oral prodrug forms of 5-FU, and they both
demonstrate promising clinical activity.
5-FU has antitumor activity against a wide spectrum of solid tumors,
including epithelial malignancies arising in the breast,
gastrointestinal tract, head and neck, and ovary, with single-agent
response rates of 10% to 30%. [ref: 74] Significant efforts have
focused on identifying agents that can biochemically modulate the
cytotoxic effects of 5-FU. Such modulators of 5-FU include other
antineoplastic agents such as MTX, cisplatin, and CPT-11, ionizing
radiation, cytokines such as the interferons, and agents that by
themselves have little to no activity such as LV and 5-ethynyluracil
(EU). [ref: 74,75] Thus, in the clinical setting, 5-FU is most often
incorporated as part of a combination regimen.

Intracellular Metabolism and Mechanism of Action

Intracellular activation is required for the fluoropyrimidines to exert
their cytotoxic effects. 5-FU readily enters cells via the facilitated
uracil transport mechanism, whereas FUdR is a substrate for the
facilitated nucleoside transport system. These compounds are anabolized
to cytotoxic forms by several biochemical pathways (Fig. 19.5_4). 5-
FU is converted to FUdR by thymidine phosphorylase. Subsequent
phosphorylation of FUdR by thymidine kinase results in formation of the
active 5-FU metabolite, 5-fluoro-2'-deoxyuridine monophosphate (FdUMP).
In the presence of the reduced folate cofactor, 5,10-
methylenetetrahydrofolate, FdUMP forms a stable covalent complex with
TS. [ref: 76] TS catalyzes the sole intracellular de novo formation of
thymidine-5'-monophosphate from dUMP (see Fig. 19.5_2). Inhibition of
TS leads to depletion of dTTP, thus interfering with DNA biosynthesis
and repair. 5-FU may be anabolized to fluorouridine monophosphate
through the sequential action of uridine phosphorylase and uridine
kinase. In the presence of PRPP, orotic acid phosphoribosyltransferase
directly converts 5-FU to fluorouridine monophosphate. This metabolite
is further metabolized to fluorouridine diphosphate and then to the
triphosphate form (FUTP), which is subsequently incorporated into RNA.
An alternate pathway for FdUMP synthesis is via conversion of
fluorouridine diphosphate to fluorodeoxyuridine diphosphate (FdUDP) by
ribonucleotide reductase (RR).
Inhibition of TS by FdUMP is one of the principal mechanisms of 5-FU
action. The TS-FdUMP-folate ternary complex is slowly dissociable, and
the intracellular level of 5,10-methylenetetrahydrofolate is critical
for ternary complex formation as well as for maintaining enzyme
inhibition. Depletion of intracellular reduced folate pools prevents
ternary complex formation in various tissue culture systems. [ref: 74]
Pharmacologic concentrations of LV enhance the cytotoxicity of 5-FU by
expanding the intracellular pools of 5,10-methylenetetrahydrofolate,
thereby increasing the extent and duration of TS inhibition. Randomized
clinical trials in advanced colorectal cancer indicate that the
addition of LV to bolus 5-FU significantly improves the response rate
compared with bolus 5-FU alone. [ref: 77,78] However, the actual
benefit in patient survival is marginal and on the order of only 2 to 3
months.
5-FU is extensively incorporated into both nuclear and cytoplasmic
RNA species, and this process interferes with normal RNA processing and
function. [ref: 79-84] The extent of 5-FU-RNA incorporation correlates
with cytotoxicity in some in vitro tissue culture and in vivo models.
The following effects have been described as a consequence of 5-FU-RNA
incorporation: alteration of the secondary structure of RNA; inhibition
of mRNA polyadenylation, which decreases mRNA stability; inhibition of
the conversion of high-molecular-weight nuclear RNA species to lower
molecular-weight ribosomal RNA; quantitative and qualitative changes in
protein synthesis; and incorporation into uracil-rich small nuclear RNA
species, which interferes with normal splicing. 5-FU-containing
transfer RNA forms covalent complexes with enzymes involved in
posttranslational modification of uracil residues, thereby inhibiting
their function. Despite the significant progress made in this area of
research, it remains unclear as to how these 5-FU-mediated RNA effects
translate into cytotoxicity.
Inhibition of TS leads not only to depletion of dTTP, but to
accumulation of dUMP. Both FdUMP and dUMP may be subsequently
metabolized to their respective triphosphate forms. Incorporation of
dUTP and FdUTP into cellular DNA, with resultant inhibition of DNA
synthesis and function, may represent another mechanism of
cytotoxicity. Two enzymatic mechanisms limit the DNA incorporation of
(F)dUTP: dUTP nucleotidohydrolase degrades the triphosphate forms, and
uracil-DNA glycosylase removes uracil residues from DNA. [ref: 85,86]
The combined effects of dTTP depletion and (F)dUTP-DNA incorporation
are associated with inhibition of nascent DNA chain elongation, altered
DNA stability, production of DNA single-strand breaks, and interference
with DNA repair. The genotoxic stress resulting from TS inhibition may
activate programmed cell death pathways in susceptible cells, resulting
in induction of parental DNA fragmentation. Both internucleosomal DNA
laddering typical of classic apoptosis and very high-molecular-weight
DNA fragmentation (50 kb or greater) have been described. [ref: 87]
Factors operating downstream from TS (e.g., bcl-2 and p53 status) may
influence the cellular response to such genotoxic stress. [ref: 88,89]
Some studies suggest that in some human colon cancer cells, the process
of thymineless death resulting from 5-FU therapy may be mediated via
the Fas-signaling pathway. [ref: 90]
The relative contribution of each of these mechanisms of action to
the cytotoxicity of 5-FU remains unclear and may depend on the specific
patterns of intracellular 5-FU metabolism, which vary among different
normal tissues and tumor types. The concentration of drug and the
duration of exposure play crucial roles in determining the ultimate
mechanism of cytotoxicity. However, there is now increasing support for
the critical role of TS as a therapeutic target. The specific lines of
evidence for this view include the enhanced activity of LV-modulated 5-
FU in the therapy of both early stage and advanced colorectal cancer,
the strong correlation between low TS content in tumor tissue and
response to 5-FU-based therapy, the correlation between level of TS
enzyme inhibition in tumors after 5-FU administration and response to
5-FU therapy, and the clinical activity of selective TS inhibitors such
as Tomudex. [ref: 74,91,92]

Mechanisms of Resistance

Given the multiple sites of cytotoxic action of 5-FU and the various
metabolic pathways required for its activation, it is not surprising
that several mechanisms of resistance have been identified in
experimental and clinical settings. [ref: 74,92] However, the relative
contribution of each of these mechanisms in the development of cellular
resistance to 5-FU in the actual clinical setting remains unclear.
In human and murine tumor cells selected in vitro for resistance to
5-FU, a variety of mechanisms have been described. Deletion or
diminished activity of thymidine or uridine kinase, thymidine or
uridine phosphorylase, and orotate phosphoribosyl transferase
interferes with metabolic activation. Decreased accumulation of FUTP,
FdUMP, and (F)dUTP may result from increased activity of catabolic
enzymes [acid and alkaline phosphatases, dUTP hydrolase, and
dihydropyrimidine dehydrogenase (DPD)]. [ref: 93] Expansion of
intracellular cytidine triphosphate (CTP) pools associated with
elevated CTP synthase activity results in feedback inhibition of
uridine kinase, thus decreasing the metabolism of 5-FU to
ribonucleotide forms. Decreased incorporation of 5-FU into both RNA and
DNA may result in decreased sensitivity. [ref: 92,94] A relative
deficiency of the reduced folate substrate 5,10-
methylenetetrahydrofolate may also compromise the cytotoxic action of
FdUMP on TS. This may result from low extracellular levels of reduced
folates, decreased membrane transport of reduced folates, or reduced
activity of folylpolyglutamate synthase, thereby preventing its
polyglutamation. [ref: 95]
Alterations in the target enzyme TS represent the most commonly
described mechanism of resistance to 5-FU. A decrease in binding
affinity of the 5-FU metabolite FdUMP to the TS target has resulted
from point mutations in the protein-coding region of the TS gene. [ref:
96] In vitro, in vivo, and clinical model systems have shown a strong
correlation between the levels of TS enzyme activity and TS protein and
chemosensitivity to 5-FU. In this regard, cell lines, tumors with
higher levels of TS, or both are relatively more resistant to 5-FU.
This increase in TS protein content is usually associated with
amplification of the TS gene. [ref: 97] In cell lines made resistant to
cisplatin or doxorubicin, cross-resistance to 5-FU may develop on the
basis of increased TS expression as a consequence of increased
transcription of the TS gene. [ref: 98] In several in vitro and in vivo
model systems, the levels of TS enzyme activity and TS protein have
been shown to acutely increase after exposure to 5-FU, other specific
TS inhibitor compounds, or both. [ref: 74,99,100] Moreover, acute
increases in the expression of TS protein have been identified in the
clinical setting in paired tumor tissue biopsies obtained from patients
before and during therapy with 5-FU. [ref: 101] This acute induction of
TS protein in response to drug exposure is mediated by a translational
regulatory mechanism. [ref: 102] TS protein, in its unbound or free
state, is capable of specifically repressing the translation of its own
mRNA. However, when TS protein is bound to either nucleotide,
antifolate inhibitors, or both, it is unable to repress TS mRNA
translation, and the rate of new TS protein synthesis increases. [ref:
103] Thus, induction of TS may represent an efficient and clinically
relevant mechanism for the acute development of drug resistance.

Clinical Pharmacology

An understanding of 5-FU pharmacokinetics is important given the wide
choice of routes and schedules of administration available, each of
which has advantages in terms of differing spectrum of host toxicity.
The most widely used methods for quantitating 5-FU in biologic fluids
are high-pressure liquid chromatography (HPLC) and gas chromatography-
mass spectrometry. [ref: 104] Nuclear magnetic resonance imaging with
**19F offers the potential for noninvasive monitoring of intratumoral
accumulation of 5-FU; trapping of 5-FU within tumor tissue has been
associated with clinical response.
5-FU is administered by either intravenous bolus or continuous
infusion. The volume of distribution is slightly larger than the
extracellular space, and 5-FU readily penetrates into tissues, CSF, and
extracellular third-space accumulations such as ascites or pleural
effusions. After intravenous bolus doses of 370 to 720 mg/m**2, peak
plasma concentrations reach 300 uM to 1 mM; thereafter, metabolic
elimination is rapid, with a primary half-life of 8 to 14 minutes.
Plasma levels of 5-FU fall below 1 uM within 2 hours, an approximate
threshold for cytotoxic effects. A prolonged third elimination phase of
5-FU after intravenous bolus (half-life, approximately 5 hours),
detected by a sensitive gas chromatography-mass spectrometry method,
may reflect release of 5-FU from tissues. Less than 10% of parent drug
is excreted in the urine, whereas the balance is cleared through
metabolic pathways (catabolism is greater than intracellular
anabolism). 5-FU is enzymatically inactivated to dihydrofluorouracil by
DPD. [ref: 105] Although the liver expresses the highest levels of DPD
in the body, this enzyme is widely distributed in other tissues,
including gastrointestinal mucosa and peripheral lymphocytes.
Dihydropyrimidinase subsequently converts dihydrofluorouracil to alpha-
fluoro-ureidopropionic acid, then beta-alanine synthase catalyzes the
irreversible formation of alpha-fluoro-beta-alanine with the release of
ammonia and CO(2). 5-FU and its catabolites undergo biliary excretion
and enterohepatic circulation.
5-FU clearance is saturable; total body clearance decreases with
increasing doses and deviates from a linear relationship above a
certain dose (depending on schedule of administration). For example,
clearance ranges from approximately 350 to 850 mL/min/m**2 with bolus
doses of 720 and 370 mg/m**2. It may be difficult to predict plasma
concentrations or the risk of severe toxicity at high doses.
Rare patients with inherited DPD deficiency may have life-threatening
or fatal toxicity if treated with fluoropyrimidine-based chemotherapy.
[ref: 106,107] Because affected individuals are otherwise in good
health, the first indication of the presence of this inborn error of
metabolism usually follows an unexpectedly severe reaction to 5-FU
chemotherapy. Careful testing of DPD-deficient patients has revealed an
autosomal recessive pattern of inheritance. It is now estimated that as
many as 3% to 5% of adult cancer patients may have this pharmacogenetic
syndrome. Several molecular defects, including point mutations and
deletions due to exon skipping, have been identified in DPD-deficient
patients who experience severe toxicity to 5-FU. Further studies are in
progress to establish the precise relationship between the level of DPD
activity and 5-FU host toxicity.
Frequently used continuous infusion schedules include protracted
venous infusion (300 mg/m**2/d when given alone), 96- to 120-hour
infusion every 3 weeks (1000 mg/m**2/d), and a weekly 24-hour infusion
(2600 mg/m**2/d). 5-FU clearance is faster with constant infusion,
varying from approximately 2000 to 3000 mL/min/m**2 depending on the
dose rate. The incidence of serious clinical toxicity tends to increase
with higher systemic exposure (steady-state plasma concentrations
during constant infusion and total area under the concentration time
curve with bolus administration). Pharmacologic monitoring with
intracycle dose modifications offers the potential to avoid serious
toxicity. Variation in 5-FU steady-state plasma levels according to
time of day have been reported during constant infusion, although the
time of day at which peak and trough plasma values occurred has not
been consistent between studies.
Hepatic metastases obtain their blood supply predominantly from the
arterial circulation via the hepatic artery. In patients with
metastases confined to the liver, hepatic arterial infusion of 5-FU or
FUdR provides high local drug concentrations to the tumor. The first-
pass extraction of FUdR by the normal hepatic parenchyma approaches
95%, and little drug enters the systemic circulation. In contrast, the
first-pass extraction of 5-FU ranges from 20% to 50%. [ref: 108] The
recommended dosage of FUdR is 0.2 mg/kg/d for up to 14 days (7
mg/m**2/d), whereas 440 to 555 mg/m**2/d of 5-FU may be given for up to
14 days; lower 5-FU doses (total of 300 mg/d) may be tolerated for
longer periods. The response rates reported for previously untreated
colorectal carcinoma approach 50% in selected series. [ref: 109]
Although the time to hepatic disease progression is significantly
longer compared with systemic therapy with single-agent 5-FU or 14-day
intravenous infusion of FUdR, the time to extrahepatic disease
progression and overall survival remain unchanged. Approximately 30% of
patients with liver-only metastases who have failed to respond to
systemic 5-FU may respond to hepatic arterial infusion FUdR. Systemic
toxicities are usually dose limiting with hepatic arterial infusion of
5-FU, presumably because a higher concentration of drug reaches the
systemic circulation. These side effects include oral mucositis and
gastrointestinal symptoms such as nausea, vomiting, and diarrhea.
Myelosuppression occurs less often. Local and regional toxicities
include peptic ulceration and chemical hepatitis (usually mild). In
contrast, systemic toxicities are less common with FUdR, whereas local
and regional toxicities predominate. Hepatic toxicity is dose limiting,
and gastritis, duodenitis, or frank ulcers occur in 20% to 25% of
patients.
5-FU and FUdR may also be given by the intraperitoneal route. Both
drugs are absorbed primarily through the portal circulation and are
subject to first-pass clearance in the liver before reaching the
systemic circulation. In early trials, dialysate concentrations of 5 mM
or less 5-FU maintained by intermittent exchanges of fluid for up to 5
days were tolerated, and the ratio of intraperitoneal to plasma 5-FU
levels was approximately 300. 5-FU clearance is slower from the
peritoneal cavity than from plasma (half-life approximately 1.5 hours).
Up to 20 mM 5-FU given intraperitoneally for 4 hours with 90 mg/m**2
cisplatin every 4 weeks is tolerated, except for mild nausea and
vomiting and sometimes diarrhea, whereas granulocytopenia occurs with
higher dialysate concentrations. FUdR, 3000 mg given in 2 L of
dialysate for 4 hours daily for 3 days, is well tolerated. The major
systemic toxicity is nausea and vomiting, and this is usually well
controlled. The peritoneal to plasma FUdR ratio is approximately 2700,
suggesting a potential pharmacologic advantage for the use of
intraperitoneal FUdR over 5-FU.

Clinical Toxicity

The primary effects of 5-FU are exerted on rapidly dividing tissues,
specifically gastrointestinal mucosa and bone marrow. The spectrum of
toxicities associated with 5-FU varies considerably according to the
dose, schedule, and route of administration. The most frequently used
dose and schedules of single-agent 5-FU given by bolus injection are
600 mg/m**2/week and 425 mg/m**/d for 5 days every 3 to 4 weeks. The
dose of 5-FU generally should be reduced when used in combination with
LV, the magnitude of which varies according to the schedule and LV
dose. Epithelial ulceration may occur throughout the gastrointestinal
tract and may manifest as mucositis, pharyngitis, dysphagia,
esophagitis, gastritis, colitis, or proctitis. Diarrhea may be watery
or bloody, and the combination of nausea, vomiting, and profuse
diarrhea can lead to profound dehydration and hypotension. Disruption
of the integrity of the gut lining may permit access of enteric
organisms into the blood stream, with the potential for overwhelming
sepsis, particularly if the granulocyte nadir coincides with diarrhea.
5-FU should be withheld in the face of ongoing mucositis or diarrhea,
even if mild, and subsequent doses should be reduced when the patient
has fully recovered. If diarrhea occurs, supportive care and vigorous
hydration should be given. Antidiarrheal agents, such as diphenoxylate
and loperamide, may help control mild diarrhea, but they are generally
ineffective in controlling diarrhea of greater severity. [ref: 110] In
this setting, the somatostatin analogue octreotide seems to have
greater efficacy. Mouth cooling (oral cryotherapy) with oral ice chips
for 30 minutes starting immediately before bolus 5-FU substantially
reduces the severity of mucositis. Nausea and vomiting may occur but
are usually controlled with antiemetics. Myelosuppression may also be
observed, with granulocytopenia occurring more than thrombocytopenia.
With the schedule of a daily dose for 5 days, the granulocyte and
platelet nadirs tend to occur during the second or third week of
treatment. In contrast, myelosuppression generally occurs after the
fourth weekly dose of the weekly bolus 5-FU schedule.
Continuous intravenous infusion of 5-FU at doses of 1000 mg/m**2/d
for 5 days every 3 weeks results in only minor myelosuppression,
whereas stomatitis and diarrhea are the principal dose-limiting
toxicities. With protracted venous infusion of 5-FU (300 mg/m**2/d),
serious myelosuppression is less common. The infusion can be
interrupted at the first signs of mouth soreness or diarrhea, thus
limiting the severity of toxicity. However, palmar-plantar
erythrodysesthesia (hand-foot syndrome) is a more subacute toxicity
that may eventually be dose limiting. [ref: 111] With a weekly 24-hour
infusion of 2600 mg/m**2 5-FU, neurotoxicity and gastrointestinal
toxicity are dose limiting.
Other dermatologic toxicities associated with 5-FU therapy include
alopecia, changes in the fingernails, and dermatitis that varies from a
pruritic erythematous rash followed by scaling to more severe cases
with vesicle formation. 5-FU enhances the cutaneous toxicity of
radiation, and reactions usually occur within 7 days of radiation.
Photosensitivity reactions, increased pigmentation over the veins into
which 5-FU has been administered, as well as more generalized
hyperpigmentation, and atrophy are possible. Hand-foot syndrome most
often occurs with protracted infusion schedules of 5-FU, but may also
be seen in patients receiving LV-modulated 5-FU. Ocular toxicity
includes blepharitis, epiphora, tear duct stenosis, and acute and
chronic conjunctivitis. The acute inflammatory response is reversible
when the drug is discontinued early in the treatment course, but
progression may require surgical correction of ectropion and tear duct
stenosis.
Acute neurologic symptoms, including somnolence, cerebellar ataxia,
and upper motor signs, are primarily seen in patients receiving
intracarotid infusions for head and neck tumors, but neurologic
toxicity may also occur with weekly schedules (24-hour infusion is
greater than bolus). The premise that 5-FU catabolites play a role in
the neurotoxicity is supported by preclinical models of neurotoxicity.
However, patients with DPD deficiency may experience severe neurologic
toxicity in the absence of myelosuppression and gastrointestinal
toxicity after 5-FU therapy. [ref: 112] Thus, the exact 5-FU metabolite
responsible for 5-FU neurotoxicity remains unclear.
A syndrome of chest pain, cardiac enzyme elevations, and
electrocardiographic changes consistent with myocardial ischemia may be
seen in temporal association with 5-FU administration. In some
patients, coronary angiography revealed no abnormalities, suggesting
vasospasm as a possible mechanism. This toxicity has been attributed to
parent drug and to the catabolites, fluoro-beta-alanine and
fluoroacetate. Concentration-dependent vasoconstriction occurs when
isolated vascular smooth muscle rings are exposed in vitro to 5-FU, and
this can be reversed with nitrates.
Intrahepatic administration of FUdR is complicated mainly by
cholestatic jaundice and biliary sclerosis. These adverse side effects
are thought to result from direct perfusion of the blood supply to the
gallbladder and upper bile duct with high local drug concentrations. Of
note, this complication occurs much less frequently with hepatic
arterial infusion of 5-FU. Biliary sclerosis typically occurs by the
third cycle of treatment. Therapy with FUdR may be reinstituted at a
lower dose after normalization of serum hepatic enzyme levels, but most
patients become progressively less tolerant. Catheter-related
complications include thrombosis of the catheterized vessel, hemorrhage
or infection at the site of insertion, and slippage of the catheter
into the gastroduodenal artery with resultant necrosis of the
intestinal epithelium, hemorrhage, and perforation.

Drug Interactions

A host of drug interactions have been investigated in an attempt to
enhance the cytotoxicity and therapeutic selectivity of
fluoropyrimidine chemotherapy. The interaction of 5-FU with MTX is of
particular interest as both drugs are often used in combination
chemotherapy for breast and colorectal cancer. When given before 5-FU,
MTX-mediated inhibition of DHFR results in accumulation of PRPP. [ref:
74] Increased availability of PRPP promotes formation of fluorouridine
monophosphate via the reaction catalyzed by orotic acid
phosphoribosyltransferase, with enhanced FUTP incorporation into RNA.
In contrast, administration of 5-FU before MTX antagonizes the
antipurine effects of MTX by preventing the accumulation of
dihydrofolate and maintaining the pool of reduced folates needed for de
novo purine synthesis. A 24-hour interval is superior to a 1-hour
interval in some preclinical and clinical studies. [ref: 113] The dose
of MTX in clinical trials has varied, but doses of 100 mg/m**2 or more
are usually followed by LV rescue. A metaanalysis of randomized trials
of MTX and 5-FU revealed a higher response rate compared with single-
agent bolus 5-FU. [ref: 114]
Synergy between 5-FU and cisplatin has been noted in several
preclinical models. Possible mechanisms of interaction include
cisplatin-mediated increases in the intracellular content of the
reduced folate pool, enhanced DNA damage, and interference with repair
of cisplatin-DNA adducts. In some preclinical models, concurrent
exposure to both drugs produces synergy, whereas other models suggest
that preexposure to 5-FU before cisplatin administration is superior to
the opposite sequence. Clinical synergy between infusional 5-FU and
cisplatin is evident in tumor types that are sensitive to both drugs,
such as squamous cell cancers of the head and neck and esophagus,
whereas randomized trials in colorectal cancer show no benefit with the
addition of cisplatin. The influence of sequence of cisplatin and 5-FU
on therapeutic outcome has not been carefully studied in clinical
trials.
The salvage enzyme thymidine kinase converts thymidine to thymidine
monophosphate, thereby bypassing the TS-mediated inhibition of de novo
thymidylate production. In tissue culture models, thymidine (10 to 20
uM) is frequently used to replete dTTP pools and thus negate its
potential contribution to toxicity. In some models, pharmacologic
concentrations of thymidine promote 5-FU RNA incorporation (by feedback
inhibition of thymidine kinase and RR). [ref: 74] In vivo, however,
thymidine (and its catabolite thymine) increased the half-life of 5-FU
by interfering with the catabolism of 5-FU to dihydrofluorouracil,
leading to a marked increase in 5-FU toxicity with no improvement in
antitumor activity. Simultaneous exposure to pharmacologic
concentrations of uridine may antagonize the RNA-directed cytotoxicity
of 5-FU by decreasing its activation to the ribonucleotide level by
uridine kinase; furthermore, expanded UTP pools may decrease FUTP
incorporation into RNA. Delayed administration of uridine increases the
rate of recovery from 5-FU-associated inhibition of both RNA and DNA
synthesis in some models. [ref: 115] Delayed administration of oral
uridine to patients decreases the myelosuppression associated with
weekly bolus 5-FU, but the effect on therapeutic activity has yet to be
determined. [ref: 116]
Interferons (IFNs) enhance the in vitro and in vivo activity of 5-FU
and FUdR in a cell-line-dependent manner. [ref: 117] Heterogeneity
exists among cancer cell lines as to the specific type of IFN that
enhances fluoropyrimidine toxicity. IFN-alpha may increase the activity
of thymidine and uridine phosphorylases, and increased FdUMP formation
has been reported in some cell lines. In other models, IFN-alpha may
enhance fluoropyrimidine-mediated DNA damage in the absence of a direct
effect on FdUMP pools or the extent of TS inhibition. In a human colon
cancer cell line, IFN-gamma abrogated the acute increase in TS protein
accompanying 5-FU exposure, and in so doing, enhanced the cytotoxic
effects of 5-FU. [ref: 118] IFN-alpha may decrease the clearance of 5-
FU in some individuals in a dose- and schedule-dependent manner,
particularly with consecutive daily administration of IFN-alpha in
conjunction with bolus 5-FU. Initial clinical trials appeared
promising, although the toxicity of IFN-alpha-modulated 5-FU was
substantial. The final results from several randomized trials
evaluating IFN-alpha-modulated 5-FU alone or with LV in advanced
colorectal cancer reveal no benefit of IFN to 5-FU therapy in terms of
overall response rate and survival. [ref: 119,120] Moreover, toxicity
was significantly increased with the addition of IFN. A large
randomized trial testing the contribution of IFN-alpha to a daily
schedule for 5 days of bolus 5-FU/LV has been conducted by the National
Surgical Adjuvant Breast and Bowel Project in the adjuvant treatment of
colon cancer. The results from this trial have been published, and as
in the case of advanced disease, IFN-alpha therapy is of no added
benefit in the adjuvant setting. [ref: 121]
Preexposure to inhibitors of de novo pyrimidine synthesis, such as N-
(phosphonoacetyl)-l-aspartic acid and brequinar, results in depletion
of UTP, CTP, dUMP, and dCTP pools. These biochemical effects are
associated with enhanced anabolism of 5-FU to FUTP, increased 5-FU-RNA
incorporation, and enhanced DNA-directed cytotoxicity. Although
promising results were initially reported with a weekly schedule of
low-dose N-(phosphonoacetyl)-l-aspartic acid (250 mg/m**2
intravenously) given 1 day before a 24-hour infusion of 5-FU (2600
mg/m**2), a randomized phase II trial conducted by the Southwest
Oncology Group failed to demonstrate an improvement in response rate
compared with the same schedule of 5-FU alone. [ref: 122] Results from
other multiinstitutional phase III trials are pending.
Preclinical studies show that 5-FU enhances the cytotoxicity of
ionizing radiation, and both preclinical and clinical studies have
revealed that radiosensitization appears to be enhanced with prolonged
exposure. The underlying mechanisms for this synergistic interaction
may include increased DNA damage, inhibition of DNA repair, and
accumulation of cells in S phase. [ref: 74,123,124] Some work suggests
that the G(1)/S checkpoint may play a critical role in determining the
ability of 5-FU to enhance the cytotoxic effects of radiation therapy.
Moreover, this effect may not depend on normal p53 function, but
instead, may rely on intact cyclin E-dependent kinase activity. One
example highlighting the successful clinical application of this
approach is the use of protracted infusional 5-FU during pelvic
radiation in the adjuvant treatment of rectal cancer. [ref: 125]

Strategies to Permit Oral Administration of Fluoropyrimidines

Tegafur, Uracil, 5-Fluorouracil

Administration of 5-FU by the oral route has generally been avoided
because of erratic bioavailability. Several strategies to allow oral
dosing by decreasing the catabolism of 5-FU are being explored. One
approach involves the drug tegafur, uracil, 5-fluorouracil (UFT), a
combination of Ftorafur (1-[2-tetrahydrofuryl]-5-fluorouracil, tegafur;
see Fig. 19.5_3), a 5-FU prodrug, in a 1:4 molar ratio with uracil.
Preclinical studies indicated that UFT resulted in significantly higher
tumor-to-serum 5-FU ratios than Ftorafur alone. UFT is administered
orally in divided doses, and it has been given daily for 5 to 28 days.
With oral doses ranging from 50 to 300 mg/m**2, maximum plasma levels
of Ftorafur and 5-FU occur between 0.6 and 2.1 hours. Ftorafur levels
(2.7 to 20.0 ug/mL) exceed 5-FU levels (0.025 to 0.9 ug/mL, 0.2 to 7.0
uM), and Ftorafur clearance is approximately 70 mL/min. Combined phase
II data from 438 patients treated with UFT in Japan reveal the drug has
single-agent activity (19% to 32% response rate) in adenocarcinomas
arising in the gastrointestinal tract and breast. Hematologic toxicity
is mild. Gastrointestinal toxicity includes anorexia (24%), nausea and
vomiting (12.5%), and diarrhea (12%). UFT in combination with LV has
been tested here in the United States in both phase I and phase II
studies. When UFT was used at a dose of 300 mg/m**2/d with 150 mg/d LV
in patients with advanced colorectal cancer, an overall response rate
of 42% was observed. [ref: 126] This treatment was well tolerated, and
there was only a 24% incidence of grade 3 gastrointestinal toxicity, as
manifested by diarrhea, vomiting, and abdominal cramps, and asthenia
and fatigue. The results of two randomized, phase III studies comparing
oral UFT (300 mg/m**2/d) and oral LV (75 or 90 mg/d) for 28 days every
35 days with the Mayo Clinic regimen of 5-day bolus 5-FU (425
mg/m**2/d) and LV (20 mg/m**2/d) given every 35 days in previously
untreated patients with metastatic colorectal cancer show that the oral
and intravenous regimens have similar response rates, time to disease
progression, and overall survival. However, the oral regimen was
associated with a significantly lower incidence of severe neutropenia
and need for hospitalization. [ref: 127] Thus, UFT/LV is an acceptable
alternative to intravenous 5-FU/LV for treatment of advanced colorectal
cancer. Studies are ongoing to determine its potential role in the
adjuvant therapy of early-stage colorectal cancer.

Capecitabine (Xeloda)

Capecitabine (Xeloda) represents another oral prodrug of 5-FU that was
designed with the rationale of generating selective 5-FU activation in
tumor tissue. [ref: 128] When administered orally, it is absorbed
intact through the intestinal mucosa, metabolized by a carboxylesterase
enzyme in the liver to 5'-deoxy-5-fluorocytidine, and then converted to
5'-deoxy-5-fluorouridine by cytidine deaminase, an enzyme principally
located in the liver. 5'-Deoxy-5-fluorouridine is then converted
directly at the tumor site by the thymidine-metabolizing enzyme,
thymidine phosphorylase, a protein that has been shown to function as
an angiogenic factor.
Preclinical studies have shown that tumor 5-FU concentrations are
significantly higher than those measured in plasma. Moreover, this oral
prodrug has shown activity against a broad spectrum of human tumor
xenografts with relatively mild host toxicity. Currently, this agent is
approved for use as a third-line agent in the treatment of
anthracycline- and taxane-resistant advanced breast cancer. [ref: 129]
Studies are ongoing to determine the role of this compound in the
therapy of advanced colorectal cancer. Preliminary results of a
randomized phase III North American trial comparing capecitabine (2500
mg/mg**2/d for 14 days every 3 weeks) with a Mayo Clinic regimen of a
daily dose for 5 days show that overall response rates are
significantly higher for capecitabine (23.2%) versus 5-FU/LV (15.5%).
[ref: 130] Both the duration of response and progression-free survival
are similar. Although survival data remains immature at this time,
capecitabine appears to have a more favorable toxicity profile than
intravenous 5-FU/LV and the most common toxicities observed were hand-
foot syndrome (17.7%) and diarrhea (16.3%), with myelosuppression being
relatively uncommon.

S1

S1 is an oral formulation composed of the 5-FU prodrug Ftorafur, the
DPD inhibitor 5-chloro-2,4-dihydroxypyridine, and oxonic acid. [ref:
75] 5-Chloro-2,4-dihydroxypyridine is a competitive, reversible
inhibitor of DPD that helps to significantly prolong the half-life of
5-FU and improve oral bioavailability. Oxonic acid is an inhibitor of
pyrimidine phosphoribosyltransferase, and it acts to prevent 5-FU
phosphorylation and subsequent incorporation of 5-FU metabolites into
the RNA of normal tissues in the gastrointestinal tract. Given its
mechanism of action, the goal of oxonic acid is to protect against 5-
FU-mediated gastrointestinal toxicity. Phase I studies revealed that
the dose-limiting toxicity was myelosuppression mainly in the form of
neutropenia. In Japan, where this drug was initially developed, phase
II trials have been conducted in several malignancies, including
gastric, colorectal, breast, and head and neck cancer. Response rates
have ranged between 30% and 50%, and toxicity has been generally mild,
with myelosuppression predominating. Phase I and II studies are under
way in the United States to confirm its clinical activity.

5-Ethynyluracil

The uracil analogue 5-ethynyluracil (Eniluracil, EU, GW776, 776C85) is
a potent, mechanism-based inhibitor of DPD. [ref: 75] On binding of EU
to DPD (apparent K(m) approximately 2 uM), an unstable intermediate is
formed, following which the drug becomes covalently linked to the
enzyme. Administration of EU to both animals and humans results in
complete inhibition of total body DPD, as evidenced by inhibition of
DPD enzyme activity in peripheral mononuclear cells and by up to 100-
fold elevations of plasma uracil levels. [ref: 131,132] Treatment with
EU completely inhibits DPD enzyme activity in both the tumors and
peripheral lymphocytes of patients with colorectal cancer, a finding
that confirms the use of peripheral mononuclear cells as an accurate
and reliable surrogate marker for tumoral DPD activity. Preclinical
studies in vivo show that treatment with EU significantly improves 5-FU
pharmacokinetics, resulting in a marked elevation of both the plasma
half-life and area under the concentration time curve of 5-FU.
This work has been extended to the clinical setting in which phase I
studies investigating the combination of 5-FU and EU have been
performed. EU significantly increased 5-FU plasma half-life by up to
20-fold and decreased 5-FU clearance by 18-fold, leading to a prolonged
half-life of 4 to 6 hours. [ref: 133] Moreover, in the presence of EU,
the oral bioavailability of 5-FU approaches 100%. This then allows for
effective oral administration of 5-FU. The 5-FU/EU combination has
subsequently been advanced to the phase II setting in patients with
metastatic colorectal cancer. When 5-FU and EU were administered for 28
days on an every 35-day cycle at the respective doses of 1.0 and 10.0
mg/m**2/d, a 29% objective response rate was observed. Treatment was
relatively well tolerated, and grade 3 to 4 diarrhea was noted in 16%
of patients, stomatitis and mucositis in 4%, and neutropenia in 4%.
EU is a promising biochemical modulator of 5-FU, and clinical studies
suggest that it may help to improve the antitumor activity and host
toxicity profile of 5-FU chemotherapy. Currently, there are two large
randomized, phase III studies comparing 5-FU and EU with either the
Mayo Clinic daily for 5 days regimen of 5-FU and LV or with protracted
infusional 5-FU. The results from these studies will provide critical
insights as to the precise role of the 5-FU and EU combination in the
therapeutic armamentarium of patients with metastatic colorectal
cancer. Moreover, this combination may find clinical application in the
treatment of other solid tumors such as advanced breast cancer. [ref:
133a]

Cytarabine

Cytarabine (1-beta-D-arabinofuranosylcytosine, Ara-C) is one of several
arabinose nucleosides isolated from the sponge Cryptothethya crypta,
[ref: 134] differing from its physiologic counterpart by virtue of a
stereotypic inversion of the 2'-hydroxyl group of the sugar moiety
(Fig. 19.5_5). Ara-C is regarded as one of the most important drugs
in the treatment of acute myelogenous leukemia (AML). A regimen of Ara-
C combined with an anthracycline, given as a 5- or 7-day continuous
infusion is considered the standard induction treatment for AML.
Furthermore, Ara-C is used in the treatment of other hematologic
malignancies, such as non-Hodgkin's lymphoma, chronic myelogenous
leukemia, and acute lymphocytic leukemia.

Intracellular Metabolism and Mechanism of Action

As with other nucleoside analogues and their physiologic counterparts,
Ara-C enters cells via nucleoside-specific transmembrane transport
proteins, the most important one being the es (equilibrative inhibitor-
sensitive) receptor. Studies with blasts from patients with acute
leukemias have shown a strong correlation between expression of es
transporters and in vitro sensitivity to Ara-C. [ref: 135]
Once within the cytoplasm, Ara-C requires activation for its
cytotoxic effects. The first metabolic step is the conversion of Ara-C
to ara-cyyidine monophosphate (Ara-CMP) by the enzyme deoxycytidine
kinase (dCK) (Fig. 19.5_6). This enzyme is the rate-limiting step in
intracellular anabolism of Ara-C. This metabolic step is saturable, and
the K(m) is approximately 20 uM. [ref: 136] Ara-CMP is subsequently
phosphorylated to ara-cytidine diphosphate (Ara-CDP) and Ara-CTP by the
enzymes pyrimidine monophosphate kinase and pyrimidine diphosphate
kinase, respectively. Ara-CTP competes with the native substrate
deoxycytidine triphosphate (dCTP) for DNA incorporation by DNA-directed
DNA polymerase (K(i) approximately 1 uM). The incorporated Ara-CTP
residue is a potent inhibitor of DNA polymerase alpha (involved in
Okazaki fragment synthesis on the lagging strand of the replication
fork), DNA polymerase delta (the leading strand replicase), and DNA
polymerase beta (involved in the repair of chemically induced DNA
damage). [ref: 137-139] Inhibition of DNA polymerases, in turn,
interferes with DNA chain elongation during both semiconservative DNA
replication and DNA repair. [ref: 140] The incorporated Ara-CTP residue
functions as a relative DNA chain terminator, and interference with
chain elongation is influenced by sequence specificity of the DNA
template. [ref: 138,140] Initiation of new DNA replication
intermediates continues, however, with accumulation of nascent DNA
fragments. [ref: 140] This process may result in abnormal duplication
of limited portions of DNA, thus increasing the possibility of
recombination, crossover, and gene amplification. Over time, the
nascent DNA strand can be extended beyond the arabinosylnucleotide
residue, and digestion of DNA reveals the presence of Ara-CMP in the
internucleotide (internal) position. [ref: 138-140] In some human
leukemic cell lines, Ara-C-mediated DNA damage is accompanied by a
pattern of internucleosomal DNA fragmentation typical of apoptosis
(programmed cell death). [ref: 141] There is evidence suggesting that
Ara-C metabolism in AML blasts differs from that in normal bone marrow
mononuclear cells and CD34+ hematopoietic stem cells. [ref: 142] The
total levels of phosphorylated Ara-C metabolites, including Ara-CMP,
Ara-CDP, Ara-CTP, Ara-CDP-choline, and Ara-UMP, were two- to fourfold
higher in leukemic blast cells when compared with normal bone marrow
cells, both at standard and high doses of Ara-C. The most striking
difference was found with the Ara-CDP-choline metabolite in the setting
of Ara-C dose escalation. The level of this metabolite was 4.3-fold
higher in leukemic blast cells. [ref: 142]
In animal studies, Ara-CMP inhibits transfer of galactose, N-
acetylglucosamine, and sialic acid to cell surface glycoproteins. Ara-
CTP (0.1 to 1.0 mM) inhibits synthesis of CMP-acetylneuraminic acid,
and Ara-CDP choline can be incorporated into membranes. These effects
on phospholipid synthesis and incorporation into membranes may possibly
affect membrane structure and function.
Catabolism of Ara-C involves two key enzymes, cytidine deaminase and
deoxycytidylate deaminase. They convert Ara-C and Ara-CMP into their
respective nontoxic two metabolites, Ara-U and Ara-UMP. Other catabolic
enzymes that may affect Ara-C metabolism include dCTP pyrophosphatase,
5'-nucleotidase, and alkaline and acid phosphatases. The balance
between intracellular activation and degradation is critical in
determining the amount of drug that is ultimately converted to Ara-CTP
and, thus, its subsequent cytotoxic and antitumor activity (see
Fig. 19.5_6).

Mechanisms of Resistance

Ara-C is most active during the S phase of the cell cycle. [ref: 134]
The rate of DNA synthesis influences Ara-C cytotoxicity, with maximum
effects observed when cells are exposed to Ara-C during periods of
rapid DNA synthesis. Longer exposures allow a greater proportion of
cells to enter S phase and are associated with enhanced incorporation
of Ara-CTP into DNA. [ref: 143] Therefore, the duration of Ara-C
exposure seems to be a critical determinant of its cytotoxicity.
Several mechanisms of resistance to Ara-C have been described.
Impaired transmembrane transport, decreased rate of anabolism,
increased rate of catabolism, or all three may result in the
development of Ara-C resistance. In vitro studies have demonstrated
that amplification of the cytidine deaminase gene with resultant
overexpression of the corresponding protein product leads to Ara-C
resistance. [ref: 144] The level of cytidine deaminase enzyme activity
has been shown to correlate with clinical response in patients with AML
undergoing induction chemotherapy with Ara-C-containing regimens. [ref:
145] Blasts from patients who attained complete remission and from
those with previously untreated leukemia had significantly higher
levels of cytidine deaminase than blasts from patients with refractory
disease.
Deletion of the gene encoding deoxycytidine kinase, expansion of CTP
and dCTP pools, overexpression of bcl-2, and decreased intracellular
half-life of Ara-CTP after drug removal are mechanisms that have been
implicated in Ara-C resistance. [ref: 143,146,147] The cytotoxicity of
Ara-C in leukemic cells isolated from patients correlates well with
both the extent of DNA incorporation and the intracellular retention of
Ara-CTP after drug exposure. [ref: 148]
Cellular sensitivity to Ara-C can also be influenced by the
concomitant use of other drugs. For example, all-trans-retinoic acid
was found to enhance Ara-C cytotoxicity, as well as Ara-C-induced
apoptosis in HL-60 human leukemia cells. [ref: 149] However, the
mechanism by which this sensitization is mediated remains unknown. The
sensitivity of human leukemic cell lines to Ara-C has also been tested
in the presence of stem cell factor. The addition of stem cell factor
to a suspension culture system leads to a significant increase in the
toxicity of Ara-C to self-renewing blast progenitors, especially when
associated with high concentrations of Ara-C. [ref: 150] Although
limited in its current clinical application, 6-mercaptopurine (6-MP)
has shown a positive interaction with Ara-C. By inhibiting cytidine
deaminase enzyme activity in L1210 murine leukemic cells, 6-MP is able
to maintain high concentrations of Ara-C in the culture medium. This
effect results in enhanced incorporation of Ara-C into cells and
subsequent activation to Ara-CTP. [ref: 141]

Clinical Pharmacology and Pharmacokinetics

Ara-C has poor oral bioavailability (approximately 20%) due to
extensive deamination within the gastrointestinal tract. Consequently,
Ara-C is administered via the intravenous route. The drug can also be
given subcutaneously. After intravenous bolus administration, Ara-C is
rapidly cleared with biphasic elimination: The initial half-life is
approximately 12 minutes, whereas the terminal half-life is
approximately 2 hours. Within 24 hours, 78% of a bolus dose is excreted
in the urine (71% as Ara-U, 7% as Ara-C). During continuous intravenous
infusion, steady-state plasma levels of Ara-C increase linearly to 5 to
10 uM, and drug clearance is approximately 1000 mL/min/m**2.
Thereafter, deamination is saturated, and plasma levels can increase
unpredictably. [ref: 151] With continuous infusion of doses from 100 to
200 mg/m**2/d, steady-state plasma levels range from 0.2 to 1.0 uM, and
CSF levels are approximately 50% of the plasma levels. [ref: 134,152]
When administered as a high-dose (greater than 2 g/m**2) intravenous
infusion over 1 to 3 hours, the plasma elimination is triphasic: alpha,
beta, and gamma half-lives are 16 minutes, 1.8 hours, and 6 hours,
respectively. [ref: 152,153] The mean plasma concentration at the end
of infusion ranges from approximately 60 to 150 uM, but 12 hours later
falls to less than 0.5 uM. Ara-C crosses the blood-brain barrier when
used at high doses, with CSF levels between 7% to 14% of plasma levels,
reaching peak levels of up to 10 uM. Because cytidine deaminase enzyme
activity is nearly completely absent in CSF, the drug displays a longer
half-life in the CSF (approximately 2 to 4 hours). [ref: 153]
Ara-C can also be given intrathecally as prophylaxis against CNS
tumor involvement and to treat leptomeningeal disease of both
hematologic and solid malignancies. The usual dose is anywhere from 30
to 60 mg in 5 to 10 mL diluent twice weekly until documentation of
three consecutively negative CSF cytology results. Intrathecal
administration of 50 mg/m**2 Ara-C yields peak concentrations of 1 mM,
and cytotoxic concentrations (0.4 uM or above) are maintained for 24
hours. [ref: 134] Of note, the diluent supplied with commercial
formulations of Ara-C contains 0.945% benzyl alcohol. Given the
potential toxicity of benzyl alcohol, diluents containing this
preservative should not be used for intrathecal administration in
neonates or with high-dose regimens. In these situations, preservative-
free 0.9% sodium chloride injection or other isotonic buffered diluents
should be used to reconstitute the drug. Finally, Ara-C can also be
used intraperitoneally. This approach is commonly used as second-line
treatment for ovarian cancer patients presenting primarily with
intraperitoneal disease, and it is usually given in combination with
cisplatin.

Toxicity

The toxicity profile of Ara-C is highly dependent on the dose and
schedule of administration. Myelosuppression is the dose-limiting
toxicity with a standard regimen of 100 to 200 mg/m**2/d for 7 days.
Leukopenia and thrombocytopenia are the most severe cytopenias, with
the nadir occurring between days 7 and 14 after drug administration.
However, the duration of the nadir can be significantly influenced by
the concomitant use of other cytotoxic agents and also by previous
treatment with chemotherapy.
Gastrointestinal toxicity commonly manifests as a mild to moderate
degree of anorexia, nausea, and vomiting. Mucositis, diarrhea, ileus,
and abdominal pain can also be observed. Less commonly, epithelial
ulceration can occur, ranging from superficial ulceration to intramural
hematoma formation and perforation. Transient hepatic dysfunction,
manifested as elevation of liver enzymes, may also occur with Ara-C
given at conventional doses. Acute pancreatitis has been associated
with Ara-C, mostly when given as a continuous infusion. The Ara-C
syndrome has been described in pediatric patients receiving Ara-C for
hematologic malignancies and is characterized by fever, myalgia, bone
pain, maculopapular rash, conjunctivitis, malaise, and occasional chest
pain. These symptoms usually begin within 12 hours after Ara-C
infusion. This syndrome most likely represents an allergic reaction to
Ara-C, as patients usually develop symptoms months after the first
dose, and corticosteroids can prevent its onset. [ref: 154]
Administration of Ara-C at high doses (2 to 3 g/m**2 intravenously
over 1 to 3 hours, every 12 hours; 100 mg/m**2/h for 24 hours) produces
severe myelosuppression, sometimes with prolonged nadirs. Severe
gastrointestinal toxicity in the form of mucositis, diarrhea, or both,
is also frequently observed. Neurologic toxicity is significantly more
common with high-dose Ara-C than with standard doses. The clinical
manifestations of neurologic toxicity are diverse and include seizures,
cerebral and cerebellar dysfunction, peripheral neuropathy, bilateral
rectus muscle palsy, aphasia, and Parkinsonian symptoms. [ref: 155]
Clinical signs of cerebellar dysfunction occur in up to 15% of patients
within 8 days and include dysarthria, dysdiadochokinesia, dysmetria,
and ataxia. [ref: 156] Change in alertness and cognitive ability,
memory loss, and frontal lobe release signs reflect cerebral toxicity.
Despite discontinuation of therapy, clinical recovery is incomplete in
up to 30% of affected patients. The severity of peripheral neuropathy
increases with higher cumulative Ara-C doses. Electromyography and
nerve conduction test results suggest a demyelinating polyneuropathy
with axonal degeneration. Significant neurotoxicity appears uncommon at
cumulative doses of 36 g/m**2 or less. Neurotoxicity may also be
reduced by prolonged intravenous administration (over 3 hours or more).
Patients older than 50 years and patients with elevated serum
creatinine levels are particularly susceptible to neurologic toxicity.
Other side effects are less commonly seen. Pulmonary complications
may include noncardiogenic pulmonary edema, acute respiratory distress,
and pneumonia, resulting from Streptococcus viridans infection. [ref:
157] Other side effects associated with high-dose Ara-C include
conjunctivitis (often responsive to topical corticosteroids), a painful
hand-foot syndrome, and, rarely, anaphylactic reactions. A fatal case
of toxic epidermal necrolysis has been described. Neutrophilic eccrine
hydradenitis, an unusual cutaneous reaction manifested as plaques or
nodules, can occur during the second week after high-dose Ara-C.
Intrathecal administration of Ara-C is usually uneventful. However,
it may produce fever, seizures, and alterations in mental status within
the first 24 hours of administration. Ara-C is teratogenic in animals.
Although Ara-C produces chromosomal breaks in both cultured cells and
bone marrow, it is not an established carcinogen in humans.

Drug Interactions

In vitro studies and animal tumor model systems have provided evidence
for synergistic activity between Ara-C and alkylating agents, platinum
compounds, purine analogues, antifolates, and fluoropyrimidines. More
recently, synergism has also been observed with Ara-C and other agents,
such as bryostatin 1, fludarabine, and paclitaxel. The use of IFN-alpha
combined with Ara-C in low doses is useful in the treatment of patients
with early chronic phase chronic myelogenous leukemia, despite the
existence of in vitro data suggesting a negative interaction for these
two drugs. [ref: 158] Specific biochemical, cellular, or both, kinetic
mechanisms have been described for each of these interactions, and the
sequence of drug administration seems to be critical in some cases.
Ara-C enhances the cytotoxicity of various DNA-damaging agents,
including alkylating agents (cyclophosphamide, carmustine),
topoisomerase II inhibitors (amsacrine, etoposide), cisplatin, and
ionizing radiation. The mechanism underlying these positive
interactions appears to be interference by Ara-C with the process of
DNA repair, resulting in enhanced DNA damage.
The metabolism of Ara-U, the main catabolic by-product of Ara-C, is
important for the metabolism and toxicity of Ara-C. Ara-U is mainly
excreted in the urine. High concentrations of Ara-U can decrease
deamination of Ara-C through feedback inhibition of cytidine deaminase,
thus resulting in increased intracellular levels of Ara-CTP. [ref:
146,178] Ara-U also increases the fraction of murine leukemic cells
entering S phase, thereby enhancing Ara-C cytotoxicity. [ref: 159]
Accumulation of high levels of Ara-U may occur in both plasma and CSF
in patients receiving high-dose Ara-C, with a possible increase in Ara-
CTP formation in brain tissue. [ref: 153,160] These observations may
explain the increased risk of neurotoxicity with high-dose Ara-C,
especially in those with impaired renal function.
Interference with the DNA incorporation of Ara-C (e.g., by
pretreatment with TS inhibitors) may antagonize its cytotoxicity. [ref:
161] MTX pretreatment, however, may increase Ara-CTP formation.
Antimetabolites that decrease the competing pools of dCTP may enhance
Ara-C anabolism, DNA incorporation, and its subsequent cytotoxicity.
Such agents include inhibitors of RR (fludarabine, hydroxyurea, and
high-dose thymidine), and inhibitors of CTP synthase (the
investigational drugs acivicin, cyclopentenol cytosine, and 3-
deazauridine).
Interactions between various cytokines and Ara-C may have potential
clinical implications. A 24-hour exposure of human myeloid leukemic
cells to pIXY 321, a fusion protein combining granulocyte-macrophage
colony-stimulating factor and interleukin-3, enhances high-dose Ara-C-
mediated induction of apoptosis. [ref: 162]

Gemcitabine

Gemcitabine (2',2'-difluorodeoxycytidine, dFdC, Gemzar) is a
difluorinated analogue of deoxycytidine (see Fig. 19.5_5). This
compound has shown significant preclinical and clinical activity
against several human solid tumors, including cancer of the pancreas,
small cell and non-small cell lung cancer, and bladder cancer. In
contrast to Ara-C, the spectrum of antitumor activity of gemcitabine is
much broader, despite the similarities in structure, metabolism, and
mechanism of action. [ref: 163] The most commonly used schedule in
clinical practice is 1000 mg/m**2 intravenously administered weekly for
3 weeks, followed by a 1-week rest. This schedule seems to provide the
most acceptable toxicity profile with the greatest dose intensity. This
compound has been moved rapidly from phase I/II studies into phase III
studies, mostly in combination with other established anticancer
agents.

Intracellular Metabolism and Mechanism of Action

Transport of gemcitabine into cells requires the nucleoside transporter
system. Nucleoside transport-deficient cells are highly resistant to
the drug. Furthermore, the specific type of nucleoside transporter
expressed on the cell surface may be an important determinant of drug
sensitivity. [ref: 164,165] The intracellular concentration of
adenosine triphosphate (ATP) may also be an additional sign of the
sensitivity to gemcitabine. In head and neck cancer cell lines, ATP-
replete cells accumulated significantly less gemcitabine, when compared
with ATP-deplete cells. This finding suggests the existence of an
active efflux mechanism for gemcitabine. [ref: 166] Gemcitabine is
fivefold more lipophilic than Ara-C, a feature that is thought to
contribute to the 65% greater rate of accumulation of gemcitabine in
cells when compared with Ara-C.
Gemcitabine requires intracellular activation for its cytotoxic
effects. The steps involved in its metabolic activation of gemcitabine
are similar to those observed with Ara-C, with both drugs being
activated by the same enzymatic machinery. Deoxycytidine kinase
converts dFdC into dFdCMP. [ref: 166,167] The drug is subsequently
phosphorylated by nucleoside monophosphate and diphosphate kinases to
the respective 5'-diphosphate (dFdCDP) and 5'-triphosphate derivatives
(dFdCTP). [ref: 166,167] dFdCTP is the major cellular metabolite of
dFdC. The intracellular concentration of dFdCTP determines to a large
extent its subsequent metabolism. In cells with lower concentrations of
this metabolite (less than 100 umol/L) the main route of elimination is
by deamination, whereas in cells with higher concentrations (greater
than 100 umol/L) dephosphorylation and urinary excretion predominate.
Furthermore, dFdCTP, by inhibiting dCMP deaminase, establishes a
mechanism of self-potentiation, with a marked prolongation of its
terminal half-life from 3.6 hours to 19.0 hours. This phenomenon may
explain, at least in part, the differences observed between Ara-C and
dFdC in their spectrum of clinical activity.
dFdCDP is an inhibitor of RR, and thus decreases the intracellular
pools of deoxynucleotide triphosphates (dNTPs). Depletion of dCTP, as a
consequence of RR inhibition, leads to decreased feedback inhibition of
deoxycytidine kinase and increased phosphorylation of dFdC. dFdCTP
directly inhibits deoxycytidylate deaminase; as dCTP is a required
activator of this enzyme, dFdCDP-mediated depletion of dCTP also
diminishes deoxycytidylate deaminase activity. dFdCTP competes with
dCTP for incorporation into DNA by DNA polymerase, and depletion of
dCTP favors incorporation of dFdCTP. Inhibition of DNA synthesis may
result from both perturbations of deoxynucleotide pools and
interference with DNA chain elongation. [ref: 168,169] The majority of
incorporated residues are found in the internucleotide linkage
(internal position). Incorporation of [**3H]dFdC into purified RNA has
also been reported. [ref: 170] At equimolar concentrations of Ara-C and
dFdC, dFdCTP formation is greater than Ara-CTP, and dFdCTP is retained
for a much longer period after drug removal. [ref: 166]
In vitro primer extension studies indicate that dFdCTP competes with
dCTP for incorporation into growing DNA strands by purified DNA
polymerases alpha and epsilon (involved in DNA replication and repair).
[ref: 168] The exonuclease activity of DNA polymerase epsilon is unable
to remove the incorporated dFdC residue, and a major pause in the
polymerization process occurs once the primer is extended by one
deoxynucleotide beyond the dFdC residue. Incorporation of dFdC into a
synthetic DNA template strand interferes with base insertion by
bacterial DNA polymerase (Klenow fragment). Thermal denaturation
measurements show that dFdCMP•dGMP base pairs are less stable than
dCMP•dGMP base pairs. In intact HL-60 cells, dFdC markedly slowed
nascent DNA chain elongation as monitored by pH step alkaline elution.
[ref: 170] Pulse-chase experiments with [**3H]dFdC indicate that a
nascent DNA fragment containing [**3H]dFdC progressed over time into
larger DNA intermediates and ultimately into genomic-length DNA. [ref:
170] Thus, dFdC does not function as an absolute chain terminator in
intact cells. Cytidine deaminase converts dFdC to the inactive uridine
metabolite, dFdU. The drug is cell-cycle specific, and blocks cells in
the G(1)/S interface. Cytotoxicity is schedule dependent and increases
with increasing duration of exposure. In a T-cell lymphoblastoid line,
dFdC-DNA incorporation was necessary for induction of apoptosis.

Mechanisms of Resistance

Several mechanisms of resistance to gemcitabine in cell lines have been
described. Nucleoside transport-deficient cells are highly resistant to
gemcitabine. [ref: 164] Furthermore, the efficiency of gemcitabine
uptake can vary significantly according to the specific nucleoside
transporter expressed on the cell surface. [ref: 164]
Several enzymes involved in the intracellular metabolism of
gemcitabine have been implicated in the development of resistance to
this drug. Initial in vitro studies suggested that dCK enzyme activity
deficiency was the most important cause of resistance to gemcitabine.
Gemcitabine-sensitive human ovarian carcinoma cell lines express
tenfold higher dCK enzyme activity than gemcitabine-resistant cells.
[ref: 171] However, experiments using human epidermoid carcinoma KB
cells suggest that the enzyme RR may play an important role as well. RR
is an S-phase-specific, rate-limiting enzyme of the DNA synthesis
pathway. Cells exhibiting resistance to gemcitabine demonstrated a
ninefold overexpression of RR mRNA, a twofold overexpression of RR
protein, and a 2.3-fold higher RR enzyme activity when compared with
gemcitabine-sensitive cells. [ref: 172] The potential role of RR as a
determinant of drug resistance has been confirmed in the human
erythroleukemia K562 cell line, where high RR enzyme activity was
directly correlated with resistance. [ref: 173]
The pattern of cross-resistance between various nucleoside analogues
may have potential clinical implications. For example, gemcitabine was
found to have higher antitumor activity than Ara-C in both Ara-C-
sensitive (L1210 and BCLO) and Ara-C-resistant (LA46 and Bara-C) cell
lines. [ref: 174] In another in vitro experiment, human promyelocytic
leukemia HL-60 cells were made cladribine resistant, resulting in two
resistant sublines (R13 and R23). Neither subline was found to have
cross-resistance to gemcitabine. [ref: 175] Enzymatic characterization
of these sublines revealed that both dCK and 5'-nucleotidase enzymatic
activities are likely to be involved. The ratio of dCK to 5'-
nucleotidase activity was reduced by approximately 65% in both
sublines.

Clinical Pharmacology and Pharmacokinetics

dFdC can be measured in plasma samples by HPLC and an enzyme-linked
immunosorbent assay. As dFdC is rapidly deaminated, blood collection
tubes must contain tetrahydrouridine for accurate determination of
plasma drug levels. After a 30-minute intravenous infusion of dFdC at
doses ranging from 50 to 1000 mg/m**2, the plasma concentration versus
time curve (area under the concentration time curve) increases in a
linear fashion. [ref: 176] The compound is rapidly eliminated from
plasma, mainly by deamination. The median half-life and clearance of
dFdC are 8 minutes and 119 L/h/m**2, respectively. [ref: 176] Renal
clearance of parent drug is less than 10% of the systemic clearance.
dFdU, the main catabolic by-product, is eliminated by biphasic kinetics
characterized by a long terminal phase (half-life of 14 hours). The
pharmacokinetics and toxicity of dFdC in patients with impaired hepatic
and renal function has not yet been determined. The accumulation of
dFdCTP in circulating mononuclear cells appears to be saturated when
plasma levels exceeded 15 to 20 uM; the median half-life for
intracellular retention was 4.7 hours. [ref: 176]
A phase I trial of dFdC in cancer patients explored the maximally
tolerated duration of dFdC infused at a dose rate of 10 mg/m**2/min,
calculated to produce steady-state levels of 20 uM. The maximally
tolerated duration was 8 hours (total dose of 4800 mg/m**2), with mean
steady-state plasma levels of approximately 26 uM and median clearance
of 149 L/h. Cellular pharmacokinetics of dFdCTP in circulating leukemic
cells vary considerably among patients. In eight patients with linear
dFdCTP elimination, the median half-life was 4.6 hours. In seven
patients with biphasic dFdCTP elimination, the median initial and
terminal phase half-life values were 2.5 hours and 6.8 hours.

Toxicity

Although gemcitabine is a relatively well-tolerated drug when used as a
single agent, its toxicity profile can vary significantly according to
the schedule of administration. The most commonly used schedule is a
weekly dose of 800 to 1250 mg/m**2, administered intravenously over 30
minutes, for 3 weeks on an every 4-week cycle. With this schedule,
myelosuppression is the dose-limiting toxicity, and all three lineages
can be affected. A published series of more than 3000 patients treated
with gemcitabine for pancreatic carcinoma revealed that nonhematologic
side effects are relatively uncommon. They include fever (7.3%), pain
(6.8%), asthenia (6.0%), abdominal pain (5.5%), dyspnea (5.0%),
vomiting (3.9%), anorexia (3.6%), and deep venous thrombosis (3.2%).
[ref: 177] A particularly unusual side effect of dFdC is anal pruritus,
which may be prevented with the use of corticosteroids.
Although dyspnea is a relatively uncommon side effect of gemcitabine,
its development during the treatment with the drug may require
discontinuation of the treatment. Continuation of treatment once
dyspnea develops may lead to a fatal outcome. [ref: 178] Patients
usually present with a clinical picture consistent with acute
respiratory distress syndrome, with hypoxemia, pulmonary infiltrates,
and no evidence of left ventricular failure. The onset of acute
respiratory distress syndrome in these patients can take place anywhere
between 2 and 40 days after the first dose of the drug. Thus, close
monitoring of patients for any change in baseline respiratory status is
crucial.
A rare yet potentially fatal complication of dFdC is hemolytic-uremic
syndrome (HUS). [ref: 179] The incidence of this complication has been
estimated to be less than 1%. Early recognition of HUS is important and
should prompt the immediate discontinuation of therapy to prevent death
from HUS-related complications.

Drug Interactions

Gemcitabine has been combined with various chemotherapeutic agents in
the treatment of several solid tumors. Preclinical in vitro studies
have provided evidence of synergism between gemcitabine and various
anticancer agents to support these associations. Cisplatin is one of
the agents most commonly used in combination with gemcitabine. In vitro
studies with different human cancer cell lines have shown synergistic
interaction between gemcitabine and cisplatin. [ref: 179-181] This
synergism is thought to be mainly due to an increase in platinum-DNA
adduct formation, which in turn results from dFdC incorporation into
DNA. [ref: 180] Synergism has also been demonstrated between
gemcitabine and etoposide in human ovarian and lung cancer cell lines.
[ref: 182] Moreover, sequencing of agents in this combination is
important as the synergism is maximum when cells are exposed first to
etoposide and then to gemcitabine. [ref: 182] This may be due to the
fact that cells exposed first to etoposide have low levels of dCTP in
the cytoplasm, which then allow for enhanced phosphorylation of dFdC
and subsequent incorporation of dFdCTP into DNA. [ref: 182]

6-Thiopurines

The development of the purine analogues in cancer chemotherapy began in
the early 1950s with the synthesis of the thiopurines. For this seminal
work, Hitchings and Elion received the Nobel Prize in Medicine in 1988.
The purine analogues, 6-MP and 6-thioguanine (6-TG) continue to be used
principally in the management of acute leukemia. [ref: 183] 6-MP has an
important role in the maintenance therapy of acute lymphoblastic
leukemia (ALL), whereas 6-TG is active in remission induction and in
the maintenance therapy of AML. These analogues have a single
substitution of a thiol group in place of the 6-hydroxyl group of the
purine base (Fig. 19.5_7). 6-MP is a structural analogue of
hypoxanthine, whereas 6-TG is an analogue of guanine. Azathioprine is a
derivative of 6-MP and acts as a prodrug to provide sustained release
of 6-MP.

Intracellular Metabolism and Mechanism of Action

6-MP and 6-TG act similarly with regard to their cellular biochemistry.
In their respective monophosphate nucleotide form, they inhibit de novo
purine synthesis and purine interconversion reactions, whereas the
nucleotide triphosphate metabolites are incorporated directly into
nucleic acids. The relative contribution of each of these actions to
the mechanism of cytotoxicity of these agents is unclear. [ref: 183]
Both 6-MP and 6-TG are converted to their respective monophosphate
forms by hypoxanthine-guanine phosphoribosyl transferase (HGPRT)
(Fig. 19.5_8). These ribonucleotide monophosphates inhibit the first
step of de novo purine synthesis catalyzed by glutamine
phosphoribosylpyrophosphate aminotransferase and block the conversion
of inosinic acid to adenylic acid or to guanylic acid. Inhibition of
purine nucleotide synthesis leads to the build-up of PRPP, which
facilitates the activation of 6-MP and 6-TG to their active nucleotide
forms by HGPRT.
Inhibitors of de novo purine biosynthesis, such as MTX, interact in a
synergistic manner with 6-thiopurines because the MTX-induced block in
purine synthesis expands the PRPP pool required for thiopurine
activation. Both ribonucleotide and deoxyribonucleotide metabolites of
the thiopurines are formed, which can then be incorporated into
cellular RNA and DNA. Incorporation of fraudulent nucleotides into DNA
interferes with DNA replication and results in the formation of DNA
strand breaks. [ref: 184] In some model systems, incorporation of
thiopurine nucleotides into DNA correlates with cytotoxicity.
In addition to 6-MP effects on DNA biosynthesis, there is now
evidence that the glycolytic pathway may also be affected. 6-
Phosphofructo-2-kinase, an essential enzyme for carbohydrate
metabolism, is inhibited by 6-MP. [ref: 185] Finally, this class of
compounds may inhibit the process of angiogenesis, as studies have
shown in vivo antiangiogenic activity of a thiopurine metabolite,
methylmercaptopurine riboside, in a human endometrial adenocarcinoma
xenograft model. [ref: 186]

Mechanisms of Resistance

Biochemical resistance to 6-thiopurines results from a decreased
ability to form cytotoxic nucleotide metabolites. In experimental
systems, resistant cells express either a complete or partial
deficiency of HGPRT. [ref: 187] An alteration in the affinity of HGPRT
for 6-thiopurines has also been described. [ref: 188] Studies have
shown that decreased transmembrane transport of 6-TG can also result in
drug resistance. [ref: 189] In the HHUA, DLD-1, and HCT 166 human
cancer cell lines, MMR-defective cells exhibited higher levels of drug
resistance and increased mutagenic response at the HGPRT locus to 6-TG
when compared with their MMR-proficient counterparts. Thus, the
inability to properly repair damaged or mutant DNA may provide a
selective growth advantage for MMR-defective cells. Moreover, this
finding may provide a mechanism by which 6-TG treatment results in the
development of secondary malignancies. [ref: 190]
In clinical samples derived from patients with AML, drug resistance
has also been associated with either increased concentrations of a
membrane-bound alkaline phosphatase or a conjugating enzyme, 6-
thiopurine methyltransferase (TPMT). [ref: 191] Patients who express
high levels of TPMT activity are unable to form sufficiently high
levels of active nucleotide metabolites after treatment with 6-MP. As
such, they may be more likely to benefit from treatment with 6-TG.
[ref: 192]

Clinical Pharmacology and Pharmacokinetics

HPLC analysis using the phenylmercury or sulfonated derivatives of the
thiopurines is able to detect plasma drug levels as low as 0.1 mM. An
HPLC method was developed to measure thiopurine levels in erythrocytes
in the 18 to 20 pmol range. [ref: 193] Oral doses of 6-MP of 70 to 100
mg/m**2/d are commonly used in the maintenance therapy of ALL. Oral
absorption of 6-MP is highly erratic, with only 16% to 50% of the
administered dose reaching the systemic circulation. This effect is
mainly due to rapid first-pass metabolism in the liver. [ref: 194] Food
intake and coadministration with the antibiotic cotrimoxazole
significantly reduce drug absorption. The variable bioavailability of
oral 6-MP may be an important determinant of therapeutic outcome,
because low plasma drug concentration over time measurements correlate
with an increased risk of relapse in children with ALL. [ref: 195] 6-MP
bioavailability is increased when combined with high-dose (2 to 5
g/m**2 intravenous) MTX. [ref: 196] Studies have shown that MTX
inhibits xanthine oxidase, an enzyme important in the catabolism of
thiopurines. [ref: 197]
Oral 6-MP is well distributed into most body compartments, with the
exception of the CSF. With high-dose intravenous 6-MP (200 mg/m**2
bolus followed by 800 mg/m**2 over 8 hours), a CSF to plasma ratio of
0.15 is achieved. This schedule is currently being used to prevent CNS
relapse in ALL. [ref: 198] Approximately 30% of the drug binds weakly
to plasma proteins. The plasma half-life is approximately 50 minutes
after intravenous injection and 90 minutes after oral administration.
When studied in children with ALL, plasma concentrations and
erythrocyte thioguanine nucleotide levels are highly variable and
independent from dose. [ref: 199]
The major route of drug elimination is via metabolism by several
enzymatic pathways. 6-MP is oxidized to the inactive metabolite 6-
thiouric acid by xanthine oxidase. Enhanced 6-MP toxicity may result
from the concomitant administration of both oral and intravenous 6-MP
and the xanthine oxidase inhibitor allopurinol. [ref: 200] In patients
receiving both 6-MP and allopurinol, the 6-MP dose should be reduced by
at least 50-75%.
6-MP also undergoes S-methylation by the enzyme TPMT to yield 6-
methylmercaptopurine. After further phosphorylation, 6-
methylmercaptopurine nucleotides are, themselves, capable of inhibiting
de novo purine biosynthesis, but to a lesser extent than thioguanine
nucleotides (TGNs). TPMT plays a similar role in 6-TG and azathioprine
metabolism. [ref: 201] It has been shown that TPMT enzyme activity may
vary considerably between patients. Moreover, the levels of TPMT enzyme
activity correlate inversely with intracellular levels of TGNs and with
the duration of 6-MP-induced cytopenia, suggesting that the level of
inherited TPMT activity may affect directly 6-MP cytotoxicity and host
toxicity. [ref: 191] Due to an autosomal codominant genetic
polymorphism, a series of TPMT phenotypes, with alleles of differing
enzymatic activity, are present in the general population. Point
mutations or loss of alleles of TPMT resulting in altered enzyme
activity correlate with a defect in thiopurine metabolism, thus
defining a true pharmacogenetic syndrome. [ref: 202] A polymerase chain
reaction-based method is widely used for the genetic detection of these
TPMT mutations. [ref: 203] Approximately 0.3% of the white population
expresses either a homozygous deletion or mutation of both alleles of
the TPMT gene. In these patients, grossly elevated TGN concentrations,
profound myelotoxicity with pancytopenia, and extensive
gastrointestinal symptoms are seen after only a brief course of
thiopurine treatment. [ref: 201] An estimated 10% of patients may be at
increased risk for toxicity due to heterozygous loss of the gene or a
mutant allele coding for a less enzymatically active TPMT. [ref:
202,203] Drug interactions affecting TPMT activity have also been
reported with aspirin, sulfasalazine, 5-aminosalicylic acid,
furosemide, and disulfiram. [ref: 201]
Renal excretion of 6-MP is minimal, but at high doses, as much as 20%
to 40% of the drug is removed by the kidneys. [ref: 204] Exceedingly
high doses of 6-MP (more than 1 g/m**2) in children may cause renal
precipitation of drug with hematuria and crystalluria. In patients with
renal dysfunction, dose reductions of 6-MP should be considered.

6-Thioguanine

The main intracellular pathway for 6-TG activation is catalyzed by
HGPRT, with resultant formation of ribonucleotide monophosphates.
Although the main metabolites of 6-TG are TGNs, thioinosine nucleotides
are formed as well. However, the clinical significance of these
respective metabolites as determinants of 6-TG cytotoxicity remains
unclear. [ref: 205] Although higher erythrocyte levels of TGNs are
detected after treatment with maximum tolerated dose levels of 6-TG
than 6-MP, this pharmacodynamic parameter does not clearly correlate
with the myelosuppression associated with 6-TG. [ref: 206]
6-TG is administered orally in doses of 75 to 200 mg/m**2/d for 5 to
7 days in the treatment of AML. Its oral bioavailability is erratic,
with peak plasma levels occurring 2 to 4 hours after ingestion. The
median plasma half-life of 6-TG is approximately 90 minutes. The
catabolism of 6-TG differs from 6-MP in that S-methylation with
subsequent removal of the sulfur atom is an important pathway of drug
elimination. [ref: 183] In a second catabolic pathway, 6-TG undergoes
deamination by the enzyme guanine deaminase (guanase), resulting in 6-
thioxanthene, which is then oxidized by xanthine oxidase to 6-thiouric
acid. In contrast to 6-MP, 6-TG is not a direct substrate for xanthine
oxidase. Because the inhibition of xanthine oxidase results in the
accumulation of 6-thioxanthene, an inactive metabolite, adjustments in
6-TG dosage are not required for patients receiving allopurinol.

Toxicity

The major dose-related toxicities of 6-MP are myelosuppression and
gastrointestinal toxicity. Leukopenia and thrombocytopenia are maximal
7 days after treatment. Full hematologic recovery usually occurs after
14 days. In TPMT-deficient patients, dose reduction to 5% to 25% of the
standard dose (75 mg/m**2/d) is necessary to prevent excessive
toxicity. [ref: 207] Gastrointestinal toxicities include nausea and
vomiting, anorexia, diarrhea, and stomatitis. 6-MP hepatotoxicity
occurs in up to 30% of adult patients and is manifested as mainly
cholestatic jaundice, although elevations of hepatic transaminases may
also be seen. Hepatotoxicity is usually mild and reversible after
discontinuation of 6-MP, but frank hepatic necrosis can occur after
high doses of the drug. Combinations of 6-MP with other known
hepatotoxic agents should be avoided, and liver function test results
should be closely monitored. The mechanism of liver toxicity is not
known, but may relate to the P-450-dependent metabolism of 6-MP to a
hepatotoxic metabolite or accumulation of 6-MP metabolites in the
liver. [ref: 183] 6-TG also causes dose-limiting bone marrow
suppression but is associated with fewer gastrointestinal side effects
and less hepatotoxicity than 6-MP.
As a class of drugs, the 6-thiopurine analogues are potent
suppressors of cell-mediated immunity. As such, prolonged therapy
results in an increased predisposition to bacterial and parasitic
infections. Given their immunosuppressive effects, these agents have
been used to prevent rejection of transplanted organs and to treat
autoimmune diseases, such as Crohn's disease, ulcerative colitis, and
rheumatoid arthritis. Therapeutic immunosuppression occurs at 100 mg/d,
a dose associated with only mild leukopenia. Long-term
immunosuppressive therapy with azathioprine increases the risk of
squamous carcinoma of the skin, non-Hodgkin's lymphoma, and Kaposi's
sarcoma. Chronic 6-MP treatment is associated with teratogenic effects
during the first trimester of pregnancy, and AML has been reported as a
secondary malignancy after 6-MP treatment for Crohn's disease. [ref:
208]

Fludarabine

Fludarabine (9-beta-D-arabinosyl-2-fluoroadenine, Fludara) was
synthesized as part of a rational process to develop more active
analogues of cytarabine (Fig. 19.5_9). [ref: 209] The first compound
in this series was adenine arabinoside (vidarabine; Ara-A). However,
this compound was deaminated to its inactive form to a significant
extent, thereby negating its clinical application. The 2'-fluoro
derivative of Ara-A was subsequently found to be relatively resistant
to deamination, but was difficult to formulate and poorly soluble.
Addition of a 5'-monophosphate moiety to the sugar group yielded
fludarabine, which is relatively resistant to deamination and displays
enhanced solubility.

Mechanism of Action

After intravenous administration, F-Ara-adenosine monophosphate (F-Ara-
AMP) is rapidly dephosphorylated to F-Ara-A, which enters cells by
nucleoside-specific membrane transport mechanisms. The es and ei
nucleoside transporter systems facilitate the cellular uptake of the
hydrophilic nucleoside analogues. [ref: 210] F-Ara-A is then
rephosphorylated by dCK to F-Ara-AMP, which is subsequently metabolized
to the triphosphate form. This nucleotide is the active metabolite of
the drug. It competes with deoxyadenosine triphosphate (dATP) for
incorporation into DNA, and serves as a highly effective chain
terminator. In addition, F-Ara-ATP directly inhibits DNA polymerases
involved in DNA synthesis and repair, such as DNA polymerase alpha and
beta, and inhibits other enzymes involved in DNA synthesis, such as DNA
primase, DNA ligase I and RR. DNA polymerase epsilon is unable to
remove F-Ara-AMP from the 3'-end of DNA even in the presence of excess
enzyme and substrate nucleotides, resulting in the formation of dead-
end complexes. [ref: 211]
Fludarabine is also incorporated into RNA, causing inhibition of RNA
function, processing, and mRNA translation. [ref: 212] In contrast to
other antimetabolites, fludarabine is also active against nondividing
lymphocytes. In fact, the primary effect of fludarabine may result from
activation of apoptosis, as evidenced by the presence of typical
apoptotic fragmentation of DNA into high-molecular-weight fragments,
after drug treatment. [ref: 213] The induction of apoptosis may explain
the activity of this drug in indolent lymphoproliferative diseases with
relatively low S-phase fractions. [ref: 214]

Mechanisms of Resistance

Fludarabine-resistant cell lines, such as JOK-1 [human hairy cell
leukemia (HCL)], K562 (human erythroleukemia), and L1210 (murine
leukemia) have been established. In these resistant lines, nucleoside
transport of F-Ara-A is intact, and no alterations in intracellular
drug accumulation or multidrug-resistant (mdr 1) expression are
observed. However, decreased dCK activity with diminished intracellular
formation of F-Ara-ATP is the principal mechanism of resistance in each
of these cell lines. [ref: 215] Subsequent work has shown that deletion
of one allele of deoxycytidine kinase is sufficient to result in
decreased expression of dCK. Of note, a high degree of cross-resistance
develops to multiple nucleoside analogues requiring activation by dCK,
including Ara-C, 2-CdA, and gemcitabine. [ref: 216] Fludarabine
resistance in WSU-CLL xenografts in SCID mice can be decreased by
pretreatment with bryostatin, a macrocyclic lactone. [ref: 217]
Although the underlying mechanisms for this interaction remain to be
defined, preliminary studies suggest that bryostatin may induce
differentiation of B-CLL cells into HCL-like cells, with expression of
CD11c, CD25, and tartrate-resistant acid phosphatase (TRAP), markers
typically seen in HCL.

Clinical Pharmacology and Pharmacokinetics

Peak concentrations of F-Ara-A are reached 3 to 4 hours after
intravenous or oral administration. Mean plasma levels are proportional
to dose. After intravenous administration, the decline in plasma levels
has been reported to be bi-exponential, with a distribution half-life
of 0.6 to 2.0 hours and a terminal half-life of 6.9 to 19.7 hours.
However, other reports describe a three-compartment model with a
terminal half-life between 10 and 30 hours. [ref: 212,218] The rate-
limiting step in elimination is release from tissues and renal function
affecting clearance. Dose adjustment in the setting of renal impairment
is recommended, and a 30% dose reduction in patients with a serum
creatinine above 1.5 mg/dL or creatinine clearance below 70 mL/min
should be considered.
The median peak concentration of F-Ara-ATP in lymphocytes of leukemic
patients occurs approximately 4 hours after the start of a 30-minute
intravenous infusion of 25 mg/m**2. Intracellular F-Ara-ATP elimination
exhibits a single phase with a dose-dependent terminal phase. The oral
bioavailability of liquid fludarabine is 60% to 80%, leading to
approximately two-thirds of the intracellular F-Ara-ATP levels in
chronic lymphocytic leukemia (CLL) cells achieved with intravenous
administration. [ref: 219]

Treatment

Fludarabine is the most active single agent in the treatment of CLL.
[ref: 220-222] It is also active against indolent non-Hodgkin's
lymphoma, [ref: 223,224] prolymphocytic leukemia, cutaneous T-cell
lymphoma, and Waldenstrom's macroglobulinemia. [ref: 225] This agent
has shown promising activity in approximately one-third of patients
with mantle cell lymphoma, albeit with relatively brief response. [ref:
226] In contrast to its activity in hematologic malignancies, this
compound displays minimal activity against common solid tumors.

Toxicity

In initial trials, fludarabine, when administered as a single 260
mg/m**2 dose or 112 mg/m**2 given daily for 5 days, resulted in
profound myelosuppression. This effect was not initially predicted from
preclinical in vivo studies in mice and beagle dogs, given the
extensive tissue binding and relatively low renal excretion in humans.
In addition, a dose range of 75 to 150 mg/m**2 four times a day for 5
to 7 days resulted in severe prohibitive neurotoxicity characterized by
delayed onset cortical blindness, seizures, coma, and death. [ref: 227]
Subsequent trials demonstrated that fludarabine could be safely
administered at much lower doses of 25 to 30 mg/m**2 daily for 5 days
every 28 days. At standard doses, neurotoxicity occurs in approximately
15% of patients. This toxicity is rarely severe, generally reversible,
and usually presents as headache, somnolence, or peripheral neuropathy.
[ref: 228]
At currently used doses, myelosuppression and immunosuppression are
the major side effects of fludarabine. [ref: 229] Dose-limiting and
possibly cumulative lymphopenia and thrombocytopenia are well
established. [ref: 218] Suppression of the immune system affects T-cell
more than B-cell function. Fevers, often in the setting of neutropenia,
occur in 20% to 30% of patients. Lymphocyte counts, particularly CD4+
cells, decrease rapidly after initiation of therapy, and levels can
drop to as low as 150/uL by approximately 6 months. [ref: 230] CD4+
cell recovery is slow and may take longer than 1 year to recover to
normal levels. Common opportunistic pathogens include herpes zoster,
Candida, and Pneumocystis carinii. [ref: 229] The addition of
prednisone to fludarabine does not improve the response rate or
survival, but significantly increases the risk of opportunistic
infections, notably listeriosis and Pneumocystis carinii. [ref: 230] If
concurrent corticosteroids are necessary, such as in patients with
autoimmune anemia or thrombocytopenia, long-term prophylaxis against
Pneumocystis carinii is mandatory. Hemolytic anemia has been observed,
and in some instances, has resulted in death on rechallenge with
fludarabine. [ref: 231] Thrombocytopenia, precipitation of Evan's
syndrome, and fulminant fatal myelofibrosis have also been reported.
The prolonged immunosuppression experienced with fludarabine has raised
the possibility of an increased incidence of secondary malignancies.
However, this increased risk is now thought to be due to the underlying
immune defects of the malignancy and not to the carcinogenic effects of
the nucleoside analogue. [ref: 232]
Tumor lysis syndrome occurs in less than 1% of patients, and in some
cases, it can be fatal. However, this event does not usually recur on
retreatment with fludarabine. Prophylaxis is not uniformly effective.
Other uncommon toxicities include rash, nausea, vomiting, diarrhea,
stomatitis, anorexia, increased salivation, abdominal cramps, a
metallic taste, transient elevations in hepatic enzymes, and renal
dysfunction. Treatment-associated disseminated skin rash, progressing
to pemphigus-like epidermal necrolysis, has been described. Pulmonary
toxicity, in the form of interstitial pneumonitis, can develop after
multiple courses of treatment. At times, the pulmonary sequelae may be
difficult to distinguish from those associated with opportunistic
infections. This toxicity usually responds to corticosteroids and does
not tend to recur on retreatment.

Drug Interactions

Purine analogues achieve significant response rates in low-grade
lymphomas, presumably due to their ability to induce apoptosis in these
otherwise drug-resistant malignancies. The responses seen, however, are
mostly partial and of short duration. This fact has fostered interest
in identifying drug regimens incorporating fludarabine with enhanced
activity. Fludarabine inhibits the nucleotide excision repair used by
cells to remove the DNA cross-links induced by alkylating agents
(cyclophosphamide, cisplatin). [ref: 233] Complete response rates of
nearly 90% have been observed when fludarabine and cyclophosphamide are
used in combination for patients with previously untreated low-grade
lymphomas. [ref: 234] The combination of fludarabine with the
anthracycline analogue mitoxantrone in the presence or absence of
dexamethasone (FN and FND regimens) has been successfully used to treat
indolent non-Hodgkin's lymphomas. [ref: 235-236] In fact, response
rates in excess of 90%, with half of these being complete responses,
are seen with the triple combination, FND. This compares favorably with
the 60% to 70% response rate observed with single-agent fludarabine.
[ref: 224]
Fludarabine-induced dCTP depletion increases deoxycytidine kinase
activity in K562 human leukemic cells. This enzyme, in addition to
playing a key role in generating active fludarabine metabolites, is
also capable of phosphorylating Ara-C into its active metabolite, Ara-
CTP. The resulting synergistic effect of fludarabine on Ara-C is now
established both in vitro [ref: 237] and in vivo in leukemic blast
cells derived from patients with AML. [ref: 238] This combination has
clinical efficacy in childhood AML in combination with idarubicin,
[ref: 239] in adult CLL, [ref: 240] in refractory or relapsed AML,
[ref: 241,242] and perhaps in patients with myelodysplastic syndromes.
The immunosuppressive effect of fludarabine is being used in a novel,
nonmyeloablative bone marrow transplant preparative regimen called
transplant lite. The goal of this approach is to achieve allogeneic
stem cell engraftment and graft-versus-leukemia/lymphoma effect in
patients with CLL and low-grade lymphomas. In a pilot study,
engraftment was achieved in 11 of 15 patients, with eight showing
complete response. [ref: 243]

Cladribine

2-Chlorodeoxyadenosine (Cladribine, 2-CdA) is a deoxyadenosine purine
nucleoside analogue. A single substitution of a chlorine atom for a
hydrogen atom at the 2 position of the purine ring of deoxyadenosine
renders this compound resistant to adenosine deaminase (ADA) (see
Fig. 19.5_9). It was developed initially as an immunosuppressive
agent. 2-CdA exhibits a dose-dependent in vitro inhibition of lymphoid
neoplasms and human leukemic cell lines, [ref: 244] but has no activity
against solid tumors. Currently, it is the drug of choice in HCL with
activity in low-grade lymphoproliferative disorders as well.

Mechanism of Action

Deoxyadenosine is cleaved within cells by the enzyme ADA, to the
deoxyinosine form. A deficiency of this enzyme leads to toxic
accumulation of deoxyadenosine in lymphocytes, manifesting as the
severe combined immunodeficiency clinical syndrome. 2-CdA enters cells
via the nucleoside transporter system. [ref: 245] Given its resistance
to deamination by ADA, an ADA-deficiency-like state develops, in which
2-CdA accumulates within cells, eventually reaching lymphotoxic levels.
On entry into the cell, it first undergoes conversion to cladribine-
monophosphate (Cld-AMP), which is then eventually metabolized to the
active metabolite, cladribine-triphosphate (Cld-ATP). [ref: 246] The
rate-limiting step is catalyzed by dCK. In contrast, catabolism of 2-
CdA is mediated by a 5'-nucleotidase. The greatest accumulation of
Cld-ATP is observed in cells with high levels of dCK and low 5'-
nucleotidase activity. Cld-ATP competitively inhibits incorporation of
the normal nucleotide dATP into DNA, a process that results in
termination of chain elongation. [ref: 213] Progressive accumulation of
Cld-ATP leads to an imbalance in deoxyribonucleotide pools, thereby
inhibiting further DNA synthesis and repair. [ref: 247] At
concentrations of 0.3 uM, 2-CdA inhibits DNA synthesis by 90% within 30
minutes. [ref: 248] The accumulation of unrepaired DNA breaks over time
may initiate the apoptosis of quiescent, nondividing lymphocytes.
Activation of the caspase-3 proteolytic cascade has been implicated as
a potential mechanism for the onset of apoptosis. [ref: 249] Finally,
2-CdA is a potent inhibitor of RR, and in so doing, it may further
inhibit the synthesis of key nucleotide substrates required for DNA
biosynthesis.

Mechanisms of Resistance

Resistance to 2-CdA has been attributed to altered intracellular
metabolism of the drug. A reduction in dCK activity, the enzyme
responsible for generating cytotoxic nucleotide metabolites, is a major
determinant of acquired resistance. [ref: 250] Cld-AMP and Cld-ATP are
dephosphorylated by the cytoplasmic enzyme, 5'-nucleotidase. WSU-CLL
cells, derived from a patient with CLL, exhibit both low levels of dCK
expression, and high-levels of 5'-nucleotidase, and accordingly, they
are resistant to 2-CdA. Interestingly, restoration of 2-CdA
chemosensitivity by pretreatment of WSU-CLL cells in vitro with
bryostatin has been reported. Bryostatin may induce differentiation of
CLL cells into a hairy cell-like phenotype, evidenced by the induced
expression of TRAP, CD11c, and CD25 in WSU-CLL cells. [ref: 251]


Clinical Pharmacology and Pharmacokinetics

Pharmacokinetic analysis suggests a two-compartment model, with mean
alpha and beta half-lives of 35 +/- 12 minutes and 6.5 +/- 2.5 hours,
respectively. The steady-state concentration after a 2- or 24-hour
infusion of 0.14 mg/kg was 22.5 +/- 11.1 nM. After a 2-hour infusion of
0.14 mg/kg, the mean maximum plasma concentrations were in the range of
100 to 400 nmol/L. The disposition of 2-CdA in plasma remains linear
over a dose range of 0.2 to 2.5 mg/m**2, with limited interindividual
variability. [ref: 252] Although there seems to be a close relationship
between dose and plasma steady-state concentrations, the relationship
between dose and clinical activity remains to be defined. [ref: 253]
When mean plasma concentrations were fitted to a three-compartment
model, the half-lives of the alpha, beta, and gamma phases ranged
between 3 and 12 minutes, 0.7 to 1.5 hours, and 5.7 to 19.0 hours,
respectively.
The dose of 2-CdA used in early clinical trials was 0.09 to 0.10
mg/kg/d administered as a 7-day continuous infusion. As the long
terminal half-life suggested the feasibility of intermittent infusions,
2-CdA has also been tested as a 2-hour infusion of 0.09 to 0.10 mg/kg/d
for 5 to 7 days or a 1-hour infusion at 6 mg/m**2 for 5 days with a 28-
day cycle. A dose-escalation study of bolus daily cladribine
established no dose-limiting nonhematologic toxicity up to 21.5
mg/m**2/d, given on a daily 1-hour intravenous bolus infusion for 5
days to patients with advanced hematologic malignancies. At higher dose
levels, prolonged cytopenias and severe infections define the upper
dose limit of the drug. [ref: 254] In a small series of patients with
HCL, no significant difference in response rate or toxicity was
observed between a 7-day continuous infusion and a daily 2-hour bolus
for 5 days. The daily dose for 5 days appears to be better suited as an
outpatient regimen. After a 2-hour or a continuous infusion of 0.12
mg/kg in patients with CLL, mean intracellular concentrations of 2-CdA
nucleotides are 12.2 and 10.8 umol/L, respectively. Intracellular
concentrations of phosphorylated CdA derivatives thus exceed plasma
concentrations of the metabolites by several hundredfold. [ref: 255]
In circulating leukemic cells of CLL patients treated with 2-CdA for
10 mg/m**2/d for 3 days, Cld-AMP and Cld-ATP median half-lives of 15
and 10 hours were observed, respectively. While maximum plasma 2-CdA
and intracellular Cld-AMP concentrations correlate well, no clear
relationship exists between the level of deoxycytidine kinase activity,
the levels of the intracellular metabolites (Cld-AMP, Cld-ATP), and
response to treatment. [ref: 253] These findings indicate that other,
as yet, unknown determinants of clinical efficacy may be present.
2-CdA is effectively cleared by the kidneys. Renal clearance is
approximately 50%, while 20% to 35% of the drug is excreted unchanged
in the urine. [ref: 252] 2-CdA is able to cross the blood-brain barrier
and penetrates into the CSF. While CSF concentrations in patients, in
the absence of meningeal disease, reach only 25% of detected plasma
levels, CSF levels exceed plasma levels in patients with meningeal
involvement. [ref: 252]
The bioavailability of the drug is almost 100% when given at a dose
of 0.14 mg/kg via the subcutaneous route. The area under the curve
achievable with subcutaneous administration is almost identical to that
of the intravenous route. [ref: 256] Oral 2-CdA reaches lower, but
clinically relevant levels of bioavailability at 37% to 51%. Absorption
of the oral form is decreased by gastric pH values below 2, an effect
that cannot be prevented or reversed with concomitant proton pump
inhibitor use. The bioavailability of oral 2-CdA correlated linearly
with dosing in a small study with oral and intravenous cross-over
design. [ref: 257] Oral dosing showed no cumulative peak concentration
or toxicity, and its pharmacokinetics are well-described by a three-
compartment model. An oral dose of 0.28 mg/kg achieved similar peak
concentrations and area under the concentration time curve as 0.14
mg/kg given either intravenously or subcutaneously. [ref: 256] The
overall feasibility of intermittent intravenous, subcutaneous, or oral
dosing suggests that these different routes of administration may
compete with and possibly replace continuous infusional schedules.

Treatment

A single course of CdA achieves durable complete remissions in 65% to
91% of patients with HCL. [ref: 258-260] Salvage treatment of patients
previously treated with IFN-alpha or splenectomy is as effective as
first-line treatment. Maintenance therapy is not required. Although
minimal residual disease is often found on reexamination of bone marrow
specimens of HCL patients in clinical complete response, relapse rates
are low. Retreatment with cladribine results in complete response in up
to 60% of relapsing patients. [ref: 261] However, it remains unclear
whether cladribine offers any significant long-term survival benefit
over another nucleoside analogue, pentostatin. Responses in patients
with CLL and non-Hodgkin's lymphoma tend to be brief, and salvage of
relapsed or refractory disease is less efficacious than in HCL. [ref:
262,263] 2-CdA achieves high response rates in pediatric patients with
AML, but not in adult patients. 2-CdA has minimal activity against
solid tumors.

Toxicity

At conventional doses, myelotoxicity is dose limiting. Decreased counts
in all three cell lines are typically observed. Thrombocytopenia
usually recovers within 2 to 4 weeks, and neutropenia in 3 to 5 weeks,
after a single course of the drug. Severe, prolonged myelotoxicity is
reported, however, after repeated cycles of cladribine used in the
treatment of CLL or low-grade lymphomas. [ref: 229] Severe autoimmune
hemolytic anemia with fatal bone marrow aplasia has been described in
CLL patients receiving repeated cycles of the drug, as CdA-induced
lymphopenia may exacerbate autoimmune hemolysis. Of note, foci of bone
marrow hypoplasia are seen in cladribine-treated patients, although the
long-term clinical significance of this effect remains unclear.
Immunosuppression accounts for the late morbidity observed in CdA-
treated patients. Lymphocyte counts, particularly CD4+ cells, decrease
within 1 to 4 weeks of drug administration and may remain depressed for
several years. [ref: 264] After discontinuation of cladribine, a median
time of up to 40 months may be required for complete recovery of normal
CD4+ counts. [ref: 229] Fevers occur in 40% to 50% of patients,
typically correlating with the duration of granulocytopenia. These
episodes may be profound, prolonged, and cumulative. [ref: 259]
Opportunistic infections are common, although usually seen less
frequently than with fludarabine. Herpes zoster is most typical, and a
variety of other pathogens are also seen, including Candida,
Pneumocystis, Pseudomonas aeruginosa, Listeria monocytogenes,
Cryptococcus neoformans, Aspergillus, cytomegalovirus, and common
bacterial infections. Infectious complications correlate with decreases
in CD4+ count, and they are more frequent with repeated courses of
therapy. [ref: 229] Treatment-related deaths have been reported in more
than 30% of patients. [ref: 265]
In patients with HCL and CLL, long-term studies have failed to
identify an increase in drug-related mortality. Specifically, initial
concerns about an increased risk for secondary malignancies have not
been confirmed, as the incidence of second cancers is not higher than
what would be expected from the underlying hematologic disorder. [ref:
232,266]
Severe neurotoxicity was encountered at high doses of 2-CdA in phase
I trials, with quadriparesis and paraparesis, proximal neuromyopathy,
and rarely, Brown-Sequard and Guillain-Barre syndromes. At currently
recommended doses, mild to moderate neurotoxicity occurs in 15% of
patients and is, at least, partly reversible with discontinuation of
the drug. [ref: 228]
Tumor lysis syndrome is rare, tends to occur after the first course,
and is generally mild and reversible. However, in rare instances, this
process may be fatal, even in patients with prior therapy.
Cardiotoxicity is uncommon, but cardiac deaths have been reported,
mainly in patients with a prior cardiac history. Pulmonary
complications of 2-CdA therapy are uncommon, but in some cases, they
have been fatal. Rashes, although uncommon, may be severe and can
present as fatal toxic epidermal necrolysis. Mild to severe
gastrointestinal toxicities occur in 15% of patients, with nausea,
vomiting, and diarrhea, and there have been rare reports of anorexia,
severe mucositis, or both. Transient elevations in hepatic enzymes may
occur, with exacerbation of hepatitis B, which may be fatal. Renal
failure occurs only at high doses.

Drug Interactions

Inhibition of RR by 2-CdA depletes intracellular dCTP pools. This
effect leads to compensatory increases in intracellular dCK activity.
dCK activity plays a critical role in the formation of the active
intracellular Ara-C metabolite, Ara-CTP. Synergistic interaction
between 2-CdA and Ara-C was observed in leukemic blast cells isolated
from patients with AML, in which 2-CdA pretreatment led to increases of
intracellular Ara-CTP pools by up to 40%, with sustained inhibition of
DNA synthesis in the circulating leukemia blasts. [ref: 267] Studies
are ongoing to translate the positive interaction between Ara-C and 2-
CdA into the clinical setting.

2'-Deoxycoformycin

2'-Deoxycoformycin (Pentostatin), a fermentation product of
Streptomyces antibioticus, was developed as a potent inhibitor of ADA
(see Fig. 19.5_9). [ref: 268] ADA is present in high concentrations
in lymphoid tissues, and this enzyme plays a vital role in the
differentiation of both T and B cells. Genetic absence of ADA leads to
severe combined immunodeficiency disorder in children, and this
syndrome is characterized by profound T- and B-cell lymphopenia.
Inhibition of this enzyme leads to an accumulation of deoxyadenosine as
there is no alternate route for its metabolic conversion to
deoxyinosine and uric acid. Deoxyadenosine is subsequently
phosphorylated by deoxycytidine kinase to deoxyadenosine monophosphate,
which is then further metabolized to the triphosphate form. This
metabolite accumulates within cells and inhibits RR, a key enzyme in
DNA synthesis. [ref: 212,268]

Mechanism of Action

Pentostatin enters cells via the nucleoside transport system, and once
within the cell, it forms a tight inhibitory complex with ADA. [ref:
268] Exposure of both murine L1210 leukemic cells and normal resting
human lymphocytes results in progressive accumulation of DNA breaks
along with a decrease in RNA synthesis. In response to drug treatment,
DNA repair is activated, resulting in depletion of critical
intracellular nicotinamide-adenine dinucleotide levels. This then leads
to exhaustion of ATP pools, resulting in eventual cell death.

Clinical Pharmacology and Pharmacokinetics

After rapid intravenous infusions (1 to 9 minutes), pentostatin shows
dose-independent first-order elimination, with a biphasic decay
characteristic of a two-compartment open model. The rapid disposition
phase is short, with a mean half-life of 8.72 minutes and a mean
terminal half-life of 4.93 hours. The mean volume of distribution is
23.1 +/- 6.16 L/m**2, and the mean steady-state volume of distribution
is 20.0 +/- 5.31 L/m**2. Nearly 100% of an administered dose is
excreted in the urine, and there is a significant correlation between
plasma levels and creatinine clearance. [ref: 269]

Treatment

Initial clinical trials with high doses of pentostatin in patients with
ALL and HCL revealed prohibitive myelosuppression and neurotoxicity,
the latter manifesting as somnolence, lethargy, confusion, seizures,
and coma. [ref: 270] Current standard regimens consist of 4 mg/m**2
doses every 1 or 2 weeks. [ref: 271] With this schedule, pentostatin
exhibits remarkable activity against HCL. In phase II studies, durable
response rates over 90% are routinely achieved, and maintenance
treatment is not required. On relapse, retreatment with the drug can be
effective. [ref: 272] In a large, prospective phase III trial,
pentostatin achieved significantly higher response rates (79% vs. 38%)
and relapse-free-survival than IFN-alpha(2a). However, this effect was
not translated into overall survival benefit. [ref: 273] Although a
small number of pentostatin-pretreated patients show good response to
cladribine, pentostatin use in patients who have progressed on
cladribine has not been investigated in a systematic manner. Minimal
residual disease on immunohistochemical examination of bone marrow
specimens is detectable in 20% to 40% of HCL patients, who achieve
complete response after pentostatin treatment. Although minimal
residual disease may be associated with an increased risk of relapse,
it remains unclear whether treatment of asymptomatic patients with
minimal residual disease provides clinical benefit. [ref: 274]
Pentostatin is also active in CLL, prolymphocytic leukemia, cutaneous
T-cell lymphoma, indolent non-Hodgkin's lymphoma, chronic myelogenous
leukemia, and Langerhans' cell histiocytosis. [ref: 268,275] Overall
response rates in these malignancies are less than that achieved with
either fludarabine and cladribine or that observed with pentostatin in
HCL. No significant activity is seen in multiple myeloma or in solid
tumors.

Toxicity

The profound immunosuppression associated with pentostatin may persist
for several years after therapy is discontinued. [ref: 276] T-cell
function is affected more than B-cell or natural killer cell function,
perhaps because of higher baseline levels of ADA in T cells. A standard
course of pentostatin is associated with more prolonged
immunosuppression than that observed with a single-course of cladribine
treatment in patients with HCL. The mean time of recovery of CD4+ cells
to normal levels is up to 50 months after completion of treatment.
[ref: 277] Myelosuppression with neutropenia and thrombocytopenia is
commonly described. Thrombotic thrombocytopenic purpura and HUS and
persistent bone marrow failure with myelodysplastic features have been
reported. Neutropenic fever is seen in up to 30% of cases, and
opportunistic infections occur with Candida, herpes zoster,
Pneumocystis carinii, and a variety of other pathogens. [ref: 229]
Although an increased incidence of second cancers is noted in
pentostatin-treated patients, the relative risk is similar to that
observed with the underlying disease in the absence of nucleoside
analogue treatment. [ref: 232]
Ocular complications include conjunctivitis, desquamative keratitis,
or periorbital edema. Dermatologic toxicity in the form of skin rash,
photosensitivity reactions, and a case of fatal erythroderma have been
reported. Nausea and vomiting with pentostatin are dose dependent.
Other uncommon side effects include stomatitis, constipation, diarrhea,
cardiac toxicity, pulmonary toxicity, renal insufficiency, urate
nephropathy, and allergic reactions. Transient and reversible increases
in hepatic enzymes and fulminant hepatic failure have been described.