19_06

Chapter 19: Pharmacology of Cancer Chemotherapy
19.6: Topoisomerase Interactive Agents

Clinton F. Stewart
Mark J. Ratain

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

Introduction

Isolation and characterization of the topoisomerase enzymes have
provided a basis for development of anticancer drugs that contribute to
the treatment of patients with a wide range of malignancies. [ref: 1]
Simply stated, the topoisomerase enzymes control and modify the
topologic states of DNA. [ref: 2] The mechanisms of these enzymes
involve DNA cleavage and strand passage through the break, followed by
religation of the cleaved DNA. The precise manner by which these
complicated events occur in a single cell is the source of intense
research, and the results of this research promise to provide
additional targets for anticancer drug therapy.
The length and intricate structure of DNA in eukaryotic cells
requires efficient organization to fit in the nucleus. This
organization presents a challenge to the cell to access and segregate
portions of DNA required for cellular functions, especially because it
is done within an extremely complex and dense intracellular milieu. The
DNA topoisomerases are nuclear enzymes that assist in these cellular
functions. These enzymes reduce DNA twisting and supercoiling that
occur in selected regions of DNA as a result of essential cellular
processes such as transcription, replication, and repair recombination.
All known DNA topoisomerases possess two characteristics: to cleave and
reseal the phosphodiester backbone of DNA; and to form a covalent
enzyme-DNA linkage, which allows the passage of another single- or
double-strand DNA through the nicked DNA. Thus, topoisomerases enable
the DNA within a cell to be tightly packed and yet still accessible for
processes necessary for proper cellular function.
The first DNA topoisomerase was discovered in the early l970s. [ref:
3] Since then, many different DNA topoisomerases have been found in
eukaryotic and prokaryotic cells. In mammalian cells, these enzymes
have been differentiated into two types--type I and type II--based on
their mechanistic and physical properties. The characteristics of each
topoisomerase are summarized and compared in Table 19.6_1. [ref: 4]
The type I topoisomerase (top I) found in mammals is a monomeric
protein encoded by a single-copy gene located on chromosome 20; its
activity is adenosine triphosphate (ATP)-independent. [ref: 5] This
enzyme binds preferentially to double-stranded DNA and cleaves one of
the DNA strands of the duplex, simultaneously forming an enzyme-DNA
covalent bond between a tyrosine residue and the 3'-phosphate of the
cleaved DNA. Through a swivel mechanism, the unbroken strand can pass
through this enzyme-mediated nick and release the torsional stress of
the DNA double helix. [ref: 6] Top I has been shown to be regulated at
the transcriptional, translational, and posttranslational levels.
Phosphorylation-dephosphorylation and poly(adenosine-
diphosphoribo)sylation are important mechanisms of top I regulation in
vitro, but the therapeutic implications of these findings are not
known.
Numerous reports have shown malignant tissue to contain higher levels
of top I than its normal counterpart. Specifically, this was observed
for colon and ovarian carcinoma, chronic lymphocytic leukemia, and
diffuse histiocytic lymphoma. [ref: 7,8] This initial observation of
increased levels of top I in malignant tissues suggested to
investigators that use of top I interactive agents might lead to a
selective antitumor effect; however, results from subsequent clinical
trials have not supported this hypothesis. In fact, more recent studies
in vitro suggest that the apparent high level of top I in malignant
tissue might be related to differences in malignant compared to normal
tissue. Other factors, such as rate of DNA synthesis, repair of drug-
induced double-strand breaks, or presence of drug transporters, are
also potential determinants of drug activity in the cell.
Two top II isoenzymes have been identified in humans. [ref: 9] The
alpha form, which has an apparent molecular weight of 170 kd, is
encoded by a single-copy gene located on chromosome 17q21-22. The beta
form has a molecular weight of 180 kd and has been mapped to chromosome
3p4. Top II alpha and beta have different subnuclear distributions and
DNA binding patterns, supporting the hypothesis that each isoform has a
specific cellular function, but these precise functions are not known.
Whereas top IIbeta is relatively constant over cell and growth cycles,
top IIalpha increases in rapidly proliferating cells. [ref: 10]
In contrast to type I topoisomerase, the function of type II
topoisomerases is ATP-dependent. [ref: 11] Once top II binds to duplex
DNA, nucleophilic reactions sequentially cleave the two complementary
strands of DNA 4 base pairs apart, and the resulting 5'-phosphoryl
groups become covalently linked to a pair of tyrosine groups, one in
each half of the dimeric top II enzyme. Once the double-strand break
has been made, the cleaved ends must be moved apart by at least 2 nm
(the diameter of a double-stranded DNA helix) and a second double-
strand segment of DNA passed through the break. DNA strand passage is
completely dependent on the binding of magnesium and ATP. Once strand
passage is complete, the cleaved DNA is religated. As with top I, top
II is a phosphoprotein, and casein kinase II and protein kinase C can
phosphorylate it in vitro, resulting in an enhancement of enzyme
activity.

Mechanism of Action of Topoisomerase Interactive Agents

Although much is known about the biochemical effects of the
topoisomerase interactive agents, very little is known about the actual
mechanism by which cell death mediated by these agents occurs.
Presumably, the damage done to DNA ultimately leads to necrosis or
apoptosis through a series of cellular processes, such as cell cycle
perturbations (possibly involving cyclins or cyclin-dependent kinases)
or DNA repair deficiency. Figure 19.6_1 is a diagrammatic depiction
of the normal activities of the topoisomerases and the effect of
topoisomerase interactive agents.
Top I interactive agents, thus far consisting primarily of
camptothecin analogues, interact with the enzyme-DNA complex. [ref: 12]
This interaction prevents the resealing of the top I-mediated DNA
single-strand breaks. However, these breaks result in cell death only
if DNA synthesis is ongoing. A collision between the advancing
replication fork and the drug-stabilized single-strand break in DNA
results in replication fork breakage and double-strand breaks in the
DNA. Treatment of mammalian cells with top I inhibitors induces
inhibition of DNA synthesis, cell cycle arrest in G(2), and cell death
by apoptosis. Drug-induced G(2) arrest has been associated with a
failure to activate cdc2 kinase. Because the cytotoxicity associated
with top I interactive agents is highly dependent on DNA synthesis, any
deregulation of cyclins, cell cycle-regulated kinases, or phosphatases
may influence the cytotoxicity of top I interactive agents (see
Fig. 19.6_1). Several other types of chemotherapeutic agents, such as
doxorubicin or actinomycin D, inhibit both top I and top II by a
similar mechanism, but at different sites in the DNA.
Top II is the molecular target for many anticancer drugs, such as the
aminoacridines, anthracyclines, and epipodophyllotoxins. [ref: 13]
These drugs inhibit religation of DNA cleaved by top II and induce
protein-linked breaks in the DNA, as documented by DNA alkaline elution
assays. When drug is removed, these breaks are reversible. The
demonstration that drug-induced, protein-associated strand breaks were
mediated through interaction with top II occurred when Chen and Liu
[ref: 3] were able to show that the protein covalently bound to DNA
fragments induced by these drugs was top II. They used the term
cleavable complex to refer to these lesions because the enzyme-DNA
complex could be isolated (cleaved) after protein denaturation. The
cleavable complex in situ is a covalent topoisomerase-DNA complex.
In addition to the top II interactive agents that stabilize top II-
DNA complexes, other inhibitors apparently inhibit the enzyme before
covalent binding to DNA occurs. However, these agents have not yet
proven useful clinically. Except for the epipodophyllotoxins, all
mammalian top II inhibitors are DNA intercalators that insert a planar
moiety between two adjacent base pairs in duplex DNA. However, as more
is learned about the biochemical mechanisms of the top II interactive
agents, it becomes clear that the historical characterization of top II
interactive agents based simply on DNA-binding properties is no longer
an appropriate means of classifying their mechanism of action. Further
study is required to understand the molecular mechanisms by which DNA
strand breaks lead to antitumor activity.

Epipodophyllotoxins

Etoposide and Teniposide

The historical details of the development of the epipodophyllotoxins
for clinical use have been reviewed. [ref: 14] The path that led to the
final development of etoposide and teniposide began in 1820 with the
inclusion of podophyllin in the United States Pharmacopeia Drug
Information. Although extracts of the Podophyllum peltatum (May apple,
mandrake plants) had been used for years by natives of the Himalayas
and the Americas as cathartics and anthelminthics, it was not until
1942, when the curative effect of podophyllin in condylomata acuminata
was demonstrated, that a number of derivatives were isolated and
synthesized. Early clinical trials of constituents of the resinous
extract of podophyllin included the derivative podophyllotoxin, but
clinical responses were poor, with excessive toxicity. Further chemical
modification of podophyllotoxin by addition of the carbohydrate moiety
beta-D-thenylidene glucoside led to the compound teniposide, which was
first introduced into clinical trials in 1970 (Fig. 19.6_2). The
second derivative formed by addition of beta-D-ethylidene glucoside to
the podophyllotoxin molecule led to the compound etoposide, which,
because of advantages in formulation, has been studied more thoroughly.
The differences in physicochemical properties of these two compounds
are presented in Table 19.6_2.
Etoposide (VP-16-213), a semisynthetic podophyllotoxin derivative,
was introduced into clinical trials in 1973 and within 1 year had a
major role in the treatment of small cell lung cancer (SCLC) and
lymphoma. Clinical trials of teniposide began in the United States in
1967 and, because antitumor effects were seen in early trials,
generated enthusiasm. Since then more aggressive teniposide dosing
regimens have been evaluated, and teniposide has shown activity as a
single agent in the treatment of SCLC. In 1983 etoposide was approved
in the United States for combination therapy of refractory testicular
tumors and SCLC; however, it is now used in frontline combination
therapy for many malignancies. In 1993, teniposide was approved for use
in combination with other approved anticancer drugs for induction
therapy in patients with refractory acute lymphoblastic leukemia. The
enhanced role of epipodophyllotoxins in anticancer therapy has resulted
primarily from an improved understanding of their clinical
pharmacology. Through rational application of pharmacologic principles,
the role of these agents in cancer therapy can be further refined.
Future clinical studies must include biochemical pharmacology studies
to identify and exploit potential pharmacologic differences between
teniposide and etoposide. These studies should also focus on schedule
dependency, rational drug combinations, new disease targets, central
nervous system (CNS) penetration, and toxicity considerations. [ref:
15]
Therapeutic success with oral epipodophyllotoxin therapy, primarily
etoposide, has stimulated an increase in the understanding of etoposide
and teniposide oral absorption (Table 19.6_3). Because of solubility
concerns, the commercially available form of etoposide has a special
formulation; however, in some pediatric studies of oral etoposide and
all studies of oral teniposide, the injectable formulation has been
used orally. Although delayed absorption of etoposide (e.g., longer
than 4 hours) has been observed in patients receiving concurrent
narcotics, the basis for this observation is unknown. [ref: 16] Marked
intrapatient and interpatient variability characterizes both etoposide
and teniposide oral bioavailability. [ref: 17] Oral absorption of
etoposide and teniposide is nonlinear, with less than a proportional
increase in etoposide in the area under the plasma concentration-time
curve (AUC) with increased oral dose. The cause for these findings is
unknown, but they suggest that more frequent oral administration of low
doses are preferable to less frequent administration of high doses to
increase dose intensity.
Etoposide and teniposide have comparable distribution properties
within the body (Table 19.6_4); however, the extent of plasma protein
binding is different between the two drugs. Etoposide is highly bound
to plasma proteins, with approximately 6% to 8% of the total drug
concentration not bound to plasma proteins; [ref: 18,19] whereas less
than 1% of the total teniposide concentration is unbound. [ref: 20] The
etoposide binding ratio measured in patients correlated directly with
serum albumin concentration, consistent with in vitro studies that show
etoposide primarily binds to albumin. [ref: 21] Hyperbilirubinemia
(e.g., total bilirubin >10 mg/dL) is associated with an increased
etoposide percent unbound. In vitro studies confirmed that bilirubin
displaces etoposide from binding sites on albumin, leading to an
increased percent unbound. A model to predict etoposide percent
unbound, based on serum albumin and total bilirubin, has been
prospectively validated in cancer patients. [ref: 22] This model
provides clinicians and investigators with a method of estimating a
patient's etoposide plasma protein binding using easily accessible
patient data.
The activity of etoposide against tumors of the CNS in preclinical
and clinical studies is intriguing because etoposide penetrates poorly
into the cerebrospinal fluid (CSF) after standard and high intravenous
doses or oral doses (e.g., CSF to plasma concentration ratio of
approximately 2%). [ref: 23] The CSF penetration of teniposide is lower
(i.e., CSF to plasma concentration ratio of 0.03% to 0.55%) [ref: 24];
however, after intravenous doses of 50 to 100 mg/m**2 significant
amounts of teniposide were found in cerebral tumor samples. Teniposide
demonstrated a greater penetration into tumor tissue than peritumoral
tissue, suggesting further evaluation of teniposide for use in
treatment of patients with brain tumors, [ref: 25,26] although more
recent work suggests that presence of blood in tissue specimens may
hinder the accurate assessment of teniposide penetration into brain
tumor tissue. [ref: 27]
Nonrenal clearance accounts for 70% to 90% of etoposide and
teniposide elimination, respectively. The exact extent and clinical
relevance of this route of elimination are unknown. Numerous
metabolites of etoposide have been identified, although they comprise
only a minor percentage of the administered dose and have little, if
any, inherent cytotoxic activity. Similarly, only trace quantities of
teniposide metabolites have been found in humans. [ref: 28] More recent
work with in vitro incubation of etoposide or teniposide with human
liver microsomes has shown that O-demethylation by cytochrome P-450
enzymes leads to formation of a catechol metabolite. [ref: 29] Further
studies with a panel of prototypical substrates and inhibitors
demonstrated the catechol formation is catalyzed by human CYP3A4. [ref:
30] Formation of these reactive metabolites may have important clinical
consequences because they covalently bind to DNA and cellular protein
and have intrinsic cytotoxic activity. [ref: 31,32] Renal clearance of
etoposide is greater than teniposide. Approximately 10% or 50% of an
administered dose is recovered in the urine as unchanged teniposide or
etoposide, respectively.
The disposition of etoposide in patients with renal and hepatic
dysfunction has been extensively studied, whereas little is known about
the effect of organ dysfunction on teniposide. [ref: 33-39] From early
reports, etoposide clearance in patients with chronic renal failure was
not altered by hemodialysis [ref: 40,41]; however, this may be
attributed to poor dialyzer efficiency, lack of penetration of
etoposide into the dialysis membrane, or high protein binding.
Etoposide systemic clearance is significantly correlated with
creatinine clearance [ref: 42]; however, few of these studies have
reported the clinical relevance of altered renal function on etoposide
toxicity. In a study of patients with normal albumin and hepatic
transaminases, an increase in etoposide AUC was seen in patients with
serum creatinine greater than 1.4 mg/dL. [ref: 43] Based on increased
myelosuppression seen in the patients with elevated creatinine, the
authors recommend a 30% decrease in etoposide dose for patients with
serum creatinine greater than 1.4 mg/dL. Other approaches to adjusting
etoposide dosage for altered renal function include use of measured
renal function [e.g., **51Cr-EDTA clearance], [ref: 44,45] or adaptive
control dosing based on measured plasma concentrations. [ref: 46] These
approaches provide more accurate methods of determining patient-
specific dose reductions for renal dysfunction; furthermore, they allow
for dose escalation in patients with unusually rapid renal function.
Thus far, all dosage adjustment recommendations for etoposide in
patients with renal dysfunction are based on hematopoietic toxicity;
whether adjusting etoposide dosage compromises antitumor efficacy is
unknown.
Although etoposide is renally excreted (approximately 50%), a
significant nonrenal or metabolism component exists (primarily
etoposide glucuronide formed in the liver and excreted in the urine).
[ref: 47] However, data from three studies show that total etoposide
clearance and half-life are not significantly altered in patients with
hepatic dysfunction (i.e., total bilirubin 2 to 32 mg/dL and elevated
transaminases) compared with controls. This seeming contradiction of no
change in etoposide total systemic clearance in patients with hepatic
dysfunction could be explained by concomitant increases in unbound drug
(e.g., bilirubin displacement or hypoalbuminemia) countering a
reduction of nonrenal clearance (e.g., hepatocellular metabolism of
unbound drug). The relationship between reduced unbound clearance and
increased fraction unbound, leading to offsetting changes in total drug
clearance, is depicted in Figure 19.6_3.
Etoposide dosage adjustment in patients with organ dysfunction should
not be made solely on changes in etoposide pharmacokinetics, although
they can aid in the understanding of altered drug toxicity and
therapeutic efficacy associated with impairment of renal and hepatic
function. Quantitation of the relation between pharmacokinetics and
clinical outcome (pharmacodynamics) provides the clinician with a model
to optimize etoposide dosage for an individual patient. The three
primary determinants of etoposide disposition, and presumably
pharmacologic effect, include excretion (renal), metabolism (hepatic),
and protein binding. In a clinical situation in which excretion or
metabolism is altered, the pharmacologic effect of etoposide could also
be affected. In the absence of prospectively validated dosing
recommendations, any recommendations for dosage alterations can only be
considered as guidelines for initial dosing. Clinical effects and
patient tolerance should guide subsequent etoposide dosing. Other
clinical considerations include the current condition of the patient
(e.g., performance status), previous cytotoxic chemotherapy, and the
therapeutic goal (palliation vs. cure). Arbitrary reductions in
etoposide dosage based solely on estimates of renal and hepatic
function without regard to pharmacologic effect may lead to overdosing
and toxicity or to underdosing and inadequate antitumor effects.
Table 19.6_5 summarizes what effect different clinical conditions may
have on etoposide protein binding and suggested dosage adjustments.
The results of in vitro cell culture studies and murine tumor models
have firmly established a relation between epipodophyllotoxin exposure
and cytotoxicity [ref: 48]; however, most clinical pharmacodynamic
studies focus on the relation between etoposide exposure (e.g., AUC,
steady-state drug concentration (Cp(ss)), time above a particular
concentration, or trough concentration) and toxicity (e.g., neutropenia
or thrombocytopenia). One clinical pharmacodynamic study of continuous
infusion teniposide in children with acute leukemia, lymphoma, or
neuroblastoma, showed the median Cp(ss) was higher in responding
patients (24 vs. 8 uM), presumably due to a greater teniposide
clearance in the nonresponder group. [ref: 49] From this study, a range
of teniposide systemic exposure associated with a high probability of
response and an acceptable degree of reversible myelotoxicity was
defined. In another study, teniposide unbound systemic exposure was
significantly correlated with percent decrease in white blood cell
count (WBC), whereas total systemic exposure was not as well
correlated. [ref: 50]
Inclusion of patient-specific variables, such as serum albumin in
pharmacodynamic models, [ref: 51,52] provides indirect evidence for the
importance of protein binding to etoposide pharmacodynamics. Because
only unbound drug is active, systemic exposure to unbound drug should
be more informative of response in a patient population in which
protein binding might be considered to be variable. In a number of
clinical trials, a statistically significant mathematical relation
between systemic exposure to unbound etoposide and myelosuppression has
been observed. [ref: 53-55] Thus, in patients with anticipated variable
etoposide protein binding (e.g., hypoalbuminemia), measured unbound
etoposide systemic exposure could be a more informative measure of drug
effect than total systemic exposure.
The effect of schedule on etoposide activity has been the source of
intense investigation. Although an early study [ref: 56] suggested
prolonged schedules of etoposide were more effective in patients with
SCLC, convincing evidence for schedule dependency came from two studies
of single-agent intravenous etoposide in previously untreated SCLC.
[ref: 57,58] In the first study, 500 mg/m**2 was given either as a 24-
hour infusion or in five daily 2-hour infusions. The systemic exposure
to total etoposide was identical between the two treatment arms, but
the duration of exposure to putatively cytotoxic concentrations (1
ug/mL) was 95 hours in the daily arm compared to 46 hours in the 24-
hour infusion arm. Furthermore, the overall response rate in the 5-day
arm was 89%, compared with only 10% in the 24-hour arm. Confirmation of
this observation was derived from a randomized trial of 500 mg/m**2
given intravenously over either a 5-day or an 8-day schedule. [ref: 58]
No difference in antitumor activity was noted between the two
schedules, although hematologic toxicity was more severe in the 5-day
schedule.
Etoposide total systemic clearance is increased in the presence of
the anticonvulsants (e.g., phenytoin or phenobarbital) often
administered with high-dose chemotherapy. Because etoposide undergoes
hepatic metabolism, induction of hepatic enzymes by anticonvulsants may
be of clinical relevance. Median etoposide systemic clearance in adults
receiving phenytoin was approximately 40% greater than in patients not
receiving anticonvulsants as part of their bone marrow conditioning
regimen. [ref: 59] In children, median etoposide systemic clearance was
77% higher in those receiving phenobarbital or phenytoin than in those
not taking anticonvulsants. [ref: 60] The increased systemic clearance
associated with anticonvulsant coadministration translates into a lower
systemic exposure at the same dose. Thus, patients receiving
anticonvulsants or other drugs known to induce hepatic enzymes need a
higher dose of etoposide to achieve a similar systemic exposure to that
attained in the absence of this interaction. In contrast, etoposide
systemic clearance is decreased in patients receiving cyclosporine[ref:
61] or its analogue valspodar, [ref: 62] suggesting inhibition of P-450
metabolism, disruption of P glycoprotein function, or modulation of
other mechanisms of etoposide elimination. [ref: 63]
Etoposide and cisplatin are widely used to treat solid tumors.
Because cisplatin causes both acute and chronic decreases in renal
function, numerous studies have investigated the potential for it to
alter etoposide excretion. The concurrent infusion of cisplatin with
high-dose etoposide (i.e., 350 mg/m**2/d for 5 consecutive days) did
not alter the pharmacokinetics of cisplatin or etoposide. [ref: 64,65]
However, etoposide systemic clearance was significantly lower in the
first 48 hours after cisplatin, as compared to 21 days later. [ref: 66]
Early studies suggested cumulative cisplatin exposure was associated
with lower etoposide systemic clearance [ref: 67]; however, data from
more patients have failed to demonstrate a persistent decrease in
clearance with up to 360 mg/m**2 cisplatin. Whether etoposide excretion
would be affected by larger cumulative cisplatin doses remains to be
determined.
Phenylbutazone and sodium salicylate (at pharmacologic
concentrations) were able to displace etoposide from plasma protein
binding sites. [ref: 68] Other drugs (e.g., ifosfamide, indomethacin,
nafcillin) were able to displace etoposide, but only at
suprapharmacologic plasma concentrations. Unpublished data suggest that
therapeutically relevant concentrations of tolbutamide, sodium
salicylate, and sulfamethiazole can displace protein-bound teniposide
in fresh human serum. The clinical relevance of this displacement is
unknown.
Myelosuppression is the dose-limiting toxicity for etoposide and
teniposide, and only at very high doses is mucositis dose-limiting
(i.e., >1000 mg/m**2). Granulocyte nadir counts occur between 5 and 15
days after intravenous drug administration, and recovery is usually
complete by day 28. After continuous oral administration, the nadir
granulocyte count occurs between day 21 and 28, and in most patients,
recovery is sufficient by day 35 for retreatment. [ref: 69-71]
Thrombocytopenia occurs less often, with nadir counts observed within 9
to 16 days after drug administration. Regardless of route of
administration, myelosuppression is reversible and usually not
cumulative. To avoid the potential of severe myelosuppression,
consideration should be given to reducing the etoposide dosage for
patients who have received extensive prior myelosuppressive
chemotherapy or radiation to marrow-bearing areas of the skeleton. Mild
to moderate nausea and vomiting occur in approximately 30% to 40% of
patients and may be more frequent with oral than intravenous
administration. Other gastrointestinal toxicities, including
constipation, diarrhea, stomatitis, and anorexia, have been reported
but are infrequent at standard intravenous doses; however, with
divided-dose oral etoposide, diarrhea and mucositis were dose-limiting.
At etoposide dosages used in bone marrow transplantation regimens,
mucositis occurs more frequently, and as the dosage increases, the
severity also increases. Two patients receiving high-dose etoposide
(cumulative dose at least 6.8 g/m**2) had elevations in bilirubin,
alkaline phosphatase, and aminotransferase levels that reversed within
12 weeks. [ref: 72]
Hypersensitivity reactions (including vasomotor changes), symptoms
related to the gastrointestinal tract, and pulmonary symptoms are
observed after therapy with etoposide. Although the rate reported in
adults is less than 3%, children with acute lymphocytic leukemia have
an incidence as high as 51% [ref: 73]; however, children appear to
develop more frequent reactions to etoposide than adults. [ref: 74]
Premedication with histamine (H(1) and H(2)) blockers and a slower
infusion rate may reduce the risk of further hypersensitivity reactions
upon rechallenge with etoposide, although patients developing
bronchospasm, urticaria, and severe hypotension or in whom symptom
resolution was slow probably should not be rechallenged.
Therapy-associated acute nonlymphocytic leukemia (t-ANLL) has been
described after epipodophyllotoxin-containing therapy for both solid
tumors and acute lymphocytic leukemia. [ref: 75-80] The incidence of t-
ANLL varies widely from 1.6% to as high as 25%. These leukemias appear
relatively early after diagnosis of the primary tumor (<5 years),
present in overt leukemia without preceding myelodysplasia, and are
usually French-American-British (FAB) subtypes of M4 and M5.
Relationships between the cumulative dose and schedule of etoposide
therapy and development of t-ANLL have been described. Cumulative
etoposide doses greater than 2 g/m**2 or 3 g/m**2 have been associated
with a greater incidence of t-ANLL, although this association has not
been found in all studies. A relationship between schedule of
administration and development of t-ANLL also has been suggested, with
frequent administration of high-dose intravenous epipodophyllotoxin
associated with an increased incidence of t-ANLL. Reports suggest that
t-ANLL may also occur with oral etoposide therapy. [ref: 81,82] If the
metabolism of the epipodophyllotoxins has prognostic significance to
the development of t-ANLL, as suggested by some investigators, [ref:
83] then pharmacogenetic differences may be important. [ref: 84]
A somewhat unexpected adverse event associated with teniposide is
characterized by somnolence, hypotension, and metabolic acidosis and
has been described in three children receiving more than 500 mg/m**2 of
intravenous teniposide over 4 hours. [ref: 85] Due to poor water
solubility, teniposide is formulated with polyoxyethylated castor oil
(Cremophor EL) and 42.7% (volume to volume ratio) dehydrated ethanol,
and in these patients, clinically significant (i.e., >60 mg/dL) ethanol
concentrations were detected at the time of the adverse event. To avoid
high ethanol and teniposide concentrations, teniposide doses of more
than 500 mg/m**2 should be given over 8 hours. With the exception of
this acute, vehicle-related reaction, the pattern of toxicity for
teniposide and etoposide are identical.

Etoposide Phosphate

Because of poor water solubility, etoposide is formulated with modified
polysorbate 80 (Tween 80), polyethylene glycol 300, and ethanol.
Preparation of an etoposide prodrug by modification of the etoposide
molecule to add a phosphate group at the 4 position in the E ring led
to a more water-soluble compound. [ref: 86] Etoposide phosphate is
rapidly and completely converted by endogenous phosphatases to
etoposide. Initial studies have evaluated parenteral administration,
but etoposide phosphate can also be given orally. [ref: 87-89]
In early phase I studies, etoposide generated from etoposide
phosphate showed the same pharmacokinetic and toxicity pattern as
etoposide, but with a number of advantages over the parent compound.
[ref: 90-92] Excipients known to be toxic are not found in etoposide
phosphate. The drug can be given by intravenous bolus (i.e., 5 minutes
vs. 30 to 60 minutes), reducing the cost of drug preparation and
increasing patient convenience. Etoposide phosphate is more stable and
can be given at high concentrations, making it ideal for high-dose
therapy or continuous infusion regimens. Whether these advantages will
translate into improved clinical efficacy remains to be answered.

Camptothecin Analogues

The antitumor activity of 20(S)-camptothecin, a plant alkaloid isolated
from Camptotheca acuminata, has been recognized for more than 20 years.
[ref: 93,94] Although 10-hydroxycamptothecin demonstrated activity in
studies conducted primarily in China, its use was associated with
severe and unpredictable toxicity. Several camptothecin derivatives
have been evaluated in clinical trials, including 9-amino-20(S)-
camptothecin (9-AC), 9-nitrocamptothecin (rubitecan), lurtotecan (GI
147211), 9-dimethylaminomethyl-10-hydroxycamptothecin (topotecan), and
7-ethyl-10-(4-[1-piperidino]-1-piperidino)-carbonyloxy-camptothecin
(irinotecan) (Fig. 19.6_4). The two camptothecin analogues approved
for clinical use (topotecan and irinotecan) contain the camptothecin
pentacyclic structure with a lactone (closed ring) moiety in the E
ring. This lactone is essential for cytotoxicity because the open ring,
or hydroxy acid form, is inactive. Because of logistical difficulties
in stabilizing the lactone ring before analysis, many investigators
have chosen to acidify the plasma sample, which makes all hydroxy acid
convert back to the lactone form. Thus, these investigators measure the
sum of lactone and hydroxy acid, or total drug. The clinical
pharmacokinetics of the camptothecin analogues are summarized in
Table 19.6_6. Many different doses, routes, and schedules of
administration have been evaluated for the camptothecin analogues, and
controversy exists over which is optimal.
Studies of low-dose, protracted topotecan in mice bearing xenografts
of human tumors have shown less toxicity and equal or greater antitumor
activity over shorter, more intense courses, [ref: 95-97] stimulating
interest in oral administration of camptothecin analogues. Clinical
studies of oral topotecan show a variable time to peak concentration
[ref: 98,99] and have marked interpatient variability. However, in a
study of either 15 or 21 days of oral topotecan in children,
intrapatient variability for the oval topotecan AUC and bioavailability
(F) was smaller than interpatient variability. [ref: 100] In most
clinical trials, topotecan is administered on an empty stomach,
although in one study, co-administration of topotecan gelatin capsules
with a high-fat meal led to a small decrease in the rate, but not
extent, of absorption. [ref: 101] Pharmacokinetic data from a study of
the oral absorption of intravenous irinotecan showed that irinotecan
was rapidly absorbed, and peak irinotecan concentrations were observed
within 1 to 2 hours of adminstration. [ref: 102] The relative extent of
conversion of irinotecan to SN-38 was high, with mean ratios of 0.8 and
0.7 on days 1 and 5, respectively. This compares with values of 0.02
and 0.07 after intravenous administration [ref: 103] and suggests that
presystemic formation of SN-38 occurs after oral administration.
Preliminary results of a small pediatric study suggest that the
absolute bioavailability of intravenous irinotecan administered orally
is approximately 10%, which is in agreement with preclinical studies in
the mouse. [ref: 104]
Topotecan and irinotecan both have a large steady-state volume of
distribution (Vd(ss)), consistent with either extensive plasma protein
or tissue binding. Topotecan and irinotecan are approximately 20% and
50% bound to plasma protein, respectively, thus the high Vd(ss) is most
likely due to extensive tissue binding; SN-38 is approximately 98%
bound to plasma proteins. [ref: 105] The impact of human serum albumin
on the conversion of the lactone to carboxylate form is an important
aspect of camptothecin drug-protein interactions. [ref: 106,107]
Variations in the camptothecin molecular structure may alter these
interactions, with SN-38 lactone having enhanced stability in the
presence of human serum albumin, compared with irinotecan or topotecan.
The delta-lactone ring moiety of 9-aminocamptothecin hydrolyzes almost
immediately (more than 99.5%) in the presence of human serum albumin,
substantially reducing the pharmacologically active lactone form in
human blood. In contrast, topotecan does not associate with human serum
albumin, resulting in a higher level of lactone stability in human
plasma. [ref: 108] The clinical significance of variability in serum
albumin concentrations on camptothecin toxicity and antitumor activity
is presently unknown.
Topotecan CSF penetration in children was determined from the ratio
of CSF to plasma AUC during a 24- and 72-hour continuous topotecan
infusion. [ref: 109] The median CSF penetration of topotecan lactone
was 29% (range, 10% to 59%) and 42% (range, 11% to 97%) for the 24- and
72-hour continuous infusions, respectively. The degree of penetration
is consistent with that reported for the nonhuman primate model. [ref:
110]
The primary route of elimination from the body varies between
topotecan and irinotecan. For topotecan, 50% to 65% of a dose is
recovered in the urine; thus, renal excretion is a major route of
elimination. [ref: 111-113] Although topotecan has been measured in
human bile, the importance of biliary excretion is unknown.
Approximately 30% to 40% of topotecan is eliminated by nonrenal
pathways, and the N-desmethyl metabolite of topotecan has been isolated
from urine of patients receiving topotecan. [ref: 114] This metabolite
has antitumor activity equal to that of the parent compound, and its
formation is catalyzed by the cytochrome P-450 system. The maximum
plasma concentration of total (sum of lactone and hydroxy acid) N-
desmethyl topotecan is only 0.5% of total topotecan, and the average
urinary recovery is less than 5% of the administered dose. Two other
metabolites, topotecan O-glucuronide and N-desmethyl topotecan O-
glucuronide have been reported; however, they accounted for only 13.5%
of the urinary recovery of the administered dose. Thus, other nonrenal
routes (e.g., metabolism) of topotecan elimination are yet to be
identified. [ref: 115]
In contrast, irinotecan itself has little in vitro cytotoxicity [ref:
116] and requires conversion, by the carboxylesterase enzyme, to 7-
ethyl-10-hydroxycamptothecin (SN-38) for antitumor activity. In vitro
studies have shown decreased carboxylesterase activity may be a
mechanism of cellular resistance to irinotecan. [ref: 117] Furthermore,
transfection of carboxylesterases into tumor cells increases the
activation and cytotoxicity of irinotecan. [ref: 118,119] SN-38 is
conjugated to glucuronic acid at the C(10) position by UGT1A1, and this
metabolite has no intrinsic antitumor activity. [ref: 120] The extent
of conversion of SN-38 to its glucuronide has been inversely correlated
with the risk of severe diarrhea, because the other major route of SN-
38 excretion is biliary excretion by canalicular multispecific organic
anion transporter (cMOAT) [ref: 121,122] (presumably leading to mucosal
injury). [ref: 123] In addition to SN-38 and SN-38 glucuronide, 7-
ethyl-10-[4-N-(5-aminopentanoic acid)-1-
piperidino]carbonyloxycamptothecin (APC) and 7-ethyl-10-(4-amino-1-
piperidino)carbonyloxycamptothecin (NPC) are oxidative metabolites of
irinotecan formed by CYP3A4. [ref: 124,125] Renal excretion is a minor
route of elimination for irinotecan and SN-38.
The disposition of topotecan in patients with renal and hepatic
dysfunction has been studied in adults receiving intravenous topotecan
daily for 5 consecutive days. [ref: 126-128] Patients with renal
dysfunction were placed into three groups according to their creatinine
clearance (20 or below, 21 to 40, or 41 to 60 mL/min), and hepatic
dysfunction was defined as a total bilirubin greater than 1.5 mg/dL. A
control group of patients with normal renal and hepatic function was
also studied. Topotecan disposition and hematologic toxicity were not
significantly altered in patients with hepatic dysfunction (as defined
by elevated total bilirubin); therefore, no dosage alteration is
recommended for these patients. Severe neutropenia was observed in
patients with moderate to severe renal dysfunction treated at one-third
of the adult maximal tolerated dosage (i.e., 0.5 mg/m**2/d). Based on
the results of this study, the investigators recommended an initial
topotecan dosage of 0.75 mg/m**2/d when given daily for 5 consecutive
days to patients with moderate renal dysfunction (i.e., creatinine
clearance <39 mL/min). In another study of patients with renal
dysfunction, no correlation was found between total topotecan clearance
and a more specific measure of glomerular filtration rate [technetium
Tc 99m DPTA clearance]. Furthermore, one patient studied with a
technetium clearance of 19 mL/min/m**2 had a normal topotecan total
clearance, suggesting topotecan may undergo renal tubular secretion in
addition to glomerular filtration. Results of a study in mice showed
that probenecid would inhibit renal tubular secretion of topotecan and
decrease topotecan renal and systemic clearance, leading to an increase
in topotecan lactone systemic exposure. [ref: 129] Irinotecan undergoes
significant hepatic metabolism, and preliminary pharmacokinetic and
toxicity results of a study of irinotecan administration in adults with
liver dysfunction suggest that the irinotecan dosage should be reduced
by one-third in patients with total bilirubin more than three times the
upper limit of normal. [ref: 130]
As with the epipodophyllotoxins, results of preclinical studies of
the camptothecin analogues show a definite relationship between dose
and antitumor effect in mice bearing human tumor xenografts. These
studies have evaluated a variety of schedules and routes of
administration, and the results suggest that the camptothecin analogues
are highly schedule-dependent. [ref: 131] Early clinical studies of the
relationship between systemic exposure and pharmacologic effect
reported a statistically significant and clinically relevant
relationship between drug exposure and myelosuppression (i.e.,
percentage change in absolute neutrophil count and platelets). [ref:
132,133] A significant correlation was reported between systemic
exposure to 9-aminocamptothecin lactone and the extent of neutropenia
observed after a 72-hour continuous infusion. [ref: 134] Few studies
have looked at the relationship between drug exposure and antitumor
efficacy; however, one study of continuous infusion topotecan in
children with relapsed acute leukemia found a correlation between
topotecan lactone systemic exposure and toxicity and topotecan lactone
systemic exposure and oncolytic effect (Fig. 19.6_5). [ref: 135]
The presence of the pH-sensitive lactone ring present in the two
commercially available camptothecin analogues, topotecan and
irinotecan, raises the question of whether lactone or total systemic
exposure is a better representation of pharmacologic effect. The
advantage of measuring total drug is the elimination of the more
cumbersome determination of the lactone concentrations, which require
immediate sample processing. Total concentrations (lactone plus hydroxy
acid) may be determined without immediate processing, and assays may be
performed several weeks or months later, thus, making it feasible to do
large population studies. Use of total drug as a surrogate for systemic
exposure to the active lactone has potential drawbacks (e.g., presence
of active metabolites, intrapatient variability), although the lactone
and hydroxy acid exist in a pH-dependent equilibrium that varies little
at physiologic pH.
Reversible myelosuppression, with both neutropenia and
thrombocytopenia, is the dose-limiting toxicity observed with
topotecan. After intravenous dosing of topotecan, the neutrophil nadir
occurs between 8 and 10 days, and recovery is usually complete by day
21. On a schedule of extremely high doses given daily for 5 days,
hemolytic anemia was observed as a dose-limiting toxicity. [ref: 136]
Reversible moderate to severe anemia also has been reported with
topotecan. The use of growth factors to further escalate the topotecan
dose has been used with mixed success. [ref: 137] With oral topotecan,
prolonged oral administration (i.e., twice daily for 21 days every 28
days) resulted in gastrointestinal side effects as the dose-limiting
toxicity, whereas, when given in the short term (i.e., once daily for 5
days every 21 days), myelosuppression was the dose-limiting toxicity.
[ref: 138] For topotecan, the only nonhematologic dose-limiting
toxicity reported is mucositis, but only after a 120-hour continuous
topotecan infusion. [ref: 139,140]
The major dose-limiting toxicities for irinotecan are
myelosuppression and diarrhea. Irinotecan-induced diarrhea is generally
of two types. The first type has an early onset, beginning during or
immediately after the irinotecan infusion. This is often accompanied by
facial flushing and abdominal cramping characteristic of diarrhea
associated with vasoactive compounds. Standard-dose anticholinergic
drugs, such as scopolamine or atropine, can be used to control this
diarrhea, which is caused by the cholinergic effects of irinotecan. The
second type of diarrhea, a cholera-like syndrome unresponsive to
loperamide or codeine, is often dose-limiting. Many therapeutic
approaches have been tried to ameliorate or prevent the diarrhea
associated with irinotecan, including the use of the cyclooxygenase
inhibitor indomethacin and the enkephalinase inhibitor acetorphan, but
none has been universally successful. [ref: 141,142] Modulation of
irinotecan pharmacokinetics by inhibitors of SN-38 biliary excretion or
inducers of SN-38 glucuronidation may be another method to reduce the
severity of irinotecan-associated diarrhea. [ref: 143-145]
Other nonhematologic toxicities seen with the camptothecin analogues
are generally reversible and not dose-limiting. Mild nausea and
vomiting have occurred in approximately 20% to 30% of patients. Low-
grade fever has been observed in approximately 20% of patients, and
alopecia occurs at higher doses. Other mild toxicities observed include
fatigue, anorexia, and skin rash. Top I interactive agents have not
been in clinical use long enough for therapy-associated malignancies to
be reported; however, sister chromatid exchanges and gene deletions or
rearrangements have been induced in vitro. [ref: 146]
Topotecan systemic clearance and the formation of the N-desmethyl
topotecan metabolite are increased in the presence of enzyme-inducing
anticonvulsants (e.g., phenytoin). [ref: 147] Although hepatic
metabolism is a relatively minor component of topotecan disposition,
the steep exposure-response relationship observed with topotecan makes
this interaction clinically relevant. Patients concomitantly
administered enzyme-inducing anticonvulsants may require an increase in
topotecan dose to achieve a similar pharmacologic effect as a patient
not receiving anticonvulsants. This interaction with anticonvulsants
was also observed in a group of patients receiving 9-aminocamptothecin,
a camptothecin analogue that has no known hepatic metabolism. In this
study, patients receiving 9-aminocamptothecin and anticonvulsants did
not have the expected extent of myelosuppression. On further study of
9-aminocamptothecin pharmacokinetics, it was found that the median
steady-state 9-aminocamptothecin plasma concentrations in patients on
anticonvulsants (25.3 nM) was significantly lower than patients not
receiving anticonvulsants (76.5 nM). [ref: 148] Anticonvulsants also
increased irinotecan clearance in a study of patients with glioma,
which led to a decrease in the mean irinotecan and SN-38 AUC. [ref:
149] Although the numbers of patients studied were small, it appears
that phenobarbital may have a different effect from carbamazepine and
phenytoin, in that phenobarbital increased SN-38 glucuronide and APC
AUC by 1.6- and 2.6-fold, respectively. In another study of adults with
malignant glioma receiving irinotecan and anticonvulsants and
dexamethasone, the irinotecan clearance was twofold and the systemic
exposure to irinotecan, SN-38, and SN-38 glucuronide was statistically
lower compared with a historical control group of patients with non-CNS
tumors. [ref: 150]

Anthracyclines and Related Compounds

Anthracyclines, anticancer agents consisting of a pigmented aglycone,
an amino sugar, and a lateral chain (Fig. 19.6_6), have been in
clinical practice since the 1960s and represent one of the most
commonly used classes of anticancer drugs. [ref: 151] The first
anthracyclines in clinical use, doxorubicin and daunorubicin, were
produced by the Streptomyces species, and the anthracyclines have thus
been classified as antitumor antibiotics; however, classification by
mechanism of action is more rational, especially because the second-
generation anthracyclines (e.g., idarubicin, epirubicin) are synthetic.
The first anthracycline, doxorubicin, still remains the most widely
used and is the benchmark against which new analogues are compared.
The anthracyclines induce formation of covalent topoisomerase-DNA
complexes and prevent the enzyme from completing the religation portion
of the ligation-religation reaction. These agents are also DNA
intercalators that insert part of their planar structures between two
adjacent base pairs in DNA, causing single-stranded and double-strand
breaks. The anthracyclines can undergo chemical reduction through
enzymatically catalyzed or iron-catalyzed pathways to yield reactive
free radical intermediates. Through hydrogen peroxide and hydroxyl
radicals, these free radical intermediates can cause oxidative damage
to cellular proteins. Under hypoxic conditions, these free radicals can
rearrange to form metabolites capable of covalently binding to DNA.
Although the anthracyclines are associated with all of these reactions,
it is their interaction with top II that is the most important
mechanism of cytotoxicity. [ref: 152]
With the exception of idarubicin, none of the anthracyclines are
administered orally. Idarubicin, a synthetic analogue of daunorubicin,
has increased lipophilicity compared to daunorubicin, which allows it
to be readily absorbed from the gastrointestinal tract. [ref: 153]
Absorption of idarubicin is erratic and incomplete (Table 19.6_7);
however, higher concentrations of its active metabolite, idarubicinol,
are achieved after oral than intervenous administration because of
first-pass hepatic metabolism.
Anthracyclines as a class are unable to cross the blood-brain barrier
either because of low lipophilicity, the presence of P glycoprotein in
the cells of brain endothelial vessels, or both. [ref: 154] However,
after intravenous administration of idarubicin, its metabolite,
idarubicinol, can be detected in the CSF (1% to 13% simultaneous plasma
concentration) at concentrations associated with in vitro cytotoxicity.
The plasma protein binding for the anthracyclines is probably not
clinically relevant, with the exception of idarubicin and idarubicinol,
as hypoalbuminemia may increase systemic exposure to unbound idarubicin
and idarubicinol.
Several metabolic pathways have been reported for the anthracyclines.
Reduction of the ketone on carbon 13 yields 13S-dihydro derivatives,
which are then named after the parent anthracycline with the suffix -ol
(e.g., doxorubicinol). This reaction is catalyzed by the ubiquitous
aldoketoreductases, which in general convert daunorubicin and
idarubicin more rapidly than doxorubicin and epirubicin. Thus, plasma
concentrations of daunorubicinol and idarubicinol rapidly exceed those
of the parent drug (AUC ratio of metabolite to parent drug, 2 to 5),
compared with doxorubicinol and epirubicinol (AUC ratio, 0.3 to 0.5).
Most 13-dihydro metabolites of the anthracyclines do not have antitumor
activity; however, idarubicinol is an exception, primarily due to its
lipophilicity, which allows entry into the cell. At one time,
deglycosylation was thought to represent a metabolic pathway, but now
the clinical significance of the anthracycline aglycones is unknown.
Epirubicin, an epimer of doxorubicin, is characterized by a unique
metabolic step present only in humans. The equatorial position of the
4' hydroxyl group allows epirubicin to be conjugated to glucuronic
acid. The glucuronide AUC is similar to that of the parent compound,
potentially explaining the lower myelotoxic and cardiotoxic properties
of epirubicin compared with doxorubicin; however, the exact clinical
significance of this metabolic step is unknown.
Elimination of the anthracyclines proceeds primarily through the
bile, with urinary excretion accounting for less than 10% of the total
dose administered. No evidence for enterohepatic recirculation has been
observed. Although epirubicin is excreted in the bile, a larger
proportion of an injected dose is recovered in the urine relative to
the other anthracyclines, due to formation of soluble glucuronides.
[ref: 155]
In patients with hyperbilirubinemia or reduced renal function, dose
reduction recommendations for daunorubicin and doxorubicin are provided
in the package literature. These dose reductions are based on
retrospective data and have not been prospectively validated. The
results of a study of patients with hyperbilirubinemia receiving
doxorubicin raise questions about the validity of adjusting doses in
patients with hyperbilirubinemia. [ref: 156] The systemic clearance of
epirubicin and idarubicin in patients with liver disease is decreased,
and although a dosage reduction is recommended based on bilirubin or
serum aspartate, as with doxorubicin, it is unclear if a dosage
reduction is clinically indicated. Whereas renal impairment reduces the
clearance of epirubicin and idarubicin, only idarubicin has been
adequately studied to provide dosing guidelines.
As with other anticancer drugs, few clinical studies have related
exposure to anthracyclines and antitumor effect. Early studies
suggested doxorubicin concentration, peak or 3 hours after end of
infusion, was associated with outcome of remission induction or
reduction of tumor mass, respectively. These findings have led
investigators to speculate that improved tumor response might be linked
to high initial plasma doxorubicin concentrations. The contribution of
the anthracycline metabolite to the overall effect depends on the
anthracycline under study. As discussed previously, idarubicinol has
significant cytotoxic activity; patients with a low rate of epirubicin
glucuronidation had a lower percent change in neutrophils and better
tumor response. The only pharmacodynamic relation between doxorubicin
systemic exposure and toxicity (e.g., decrease in WBC) was noted after
continuous doxorubicin infusion. [ref: 157] A positive correlation was
noted between the AUC for epirubicin, or epirubicin and epirubicinol,
and logarithm of the WBC survival fraction. [ref: 158]
Although a number of drugs have been reported to interact with
doxorubicin, most of these have been in experimental systems, so their
clinical significance is unknown. Of the drug interactions reported in
patients with cancer, only a few are of clinical consequence. Despite
numerous studies showing pharmacologic prevention of anthracycline-
induced cardiotoxicity, only dexrazoxane has been able to significantly
retard development of cardiotoxicity. The reversal of doxorubicin-
induced multidrug resistance has been attempted using a variety of
pharmacologic modulators, such as trifluoperazine, verapamil, and
cyclosporine, and has met with mixed success.
The dose-limiting acute toxicity of the anthracyclines is
myelosuppression, primarily affecting the neutrophils; in the treatment
of leukemia, however, this may be considered a desirable side effect.
The onset of myelosuppression is usually 7 days after administration,
with maximum effect seen at approximately day 10 to 14. Recovery is
usually complete by day 21 to 28. Gastrointestinal toxicities are
common, including nausea and vomiting, diarrhea, and mucositis.
Although not dose-limiting, alopecia also occurs in almost all
patients. Although not always readily apparent when it occurs,
extravasation of the anthracyclines can lead to severe local tissue
damage and deep ulcerations that progress over weeks. These lesions are
slow to heal and often require skin grafting, although the graft is not
always successful. Once an extravasation has been discovered, the
optimal method of management is unknown, although most agree that local
measures such as ice packs and subcutaneous injections of saline,
steroids, or bicarbonate may be useful. Topical dimethylsulfoxide has
been suggested to be a safe and effective approach to reducing the
tissue damage associated with anthracycline-induced extravasation.
[ref: 159,160] The best approach to avoiding extravasation is to take
all possible precautions when administering an anthracycline, such as
ensuring good blood return on the intravenous line, monitoring the
intravenous site carefully, and good patient education.
The anthracyclines are associated with both acute and chronic cardiac
toxicity. The less common acute cardiac toxicity includes nonspecific
electrocardiographic changes that may be observed during or immediately
after the infusion. In its extreme form, this acute toxicity can
include a pericarditis-myocarditis syndrome with onset of fever,
pericarditis, and congestive heart failure. [ref: 161] No association
between the acute toxicity and later development of chronic toxicity
has been shown. Besides symptomatic management, no specific therapy is
recommended for this relatively rare syndrome.
Anthracyclines also produce a dose-dependent congestive myopathy that
often leads to congestive heart failure. This typically becomes
apparent 4 to 8 weeks after the last anthracycline dose, although it
may occur during treatment or years later. The clinical significance of
this chronic toxicity has been its inability to treat it, leading to a
drug-induced mortality ranging in early reports from 33% to 70%, and in
more recent reports to less than 30%. If a sufficiently high cumulative
dose is administered, all anthracyclines can cause cardiac toxicity;
however, idarubicin and epirubicin are associated with a lower
incidence than doxorubicin and daunorubicin. The mechanisms of the
chronic cardiotoxicity have been reviewed extensively. To summarize,
they include enzymatic-mediated formation of oxygen free radicals that
initiate lipid peroxidation and a nonenzymatic pathway for free radical
formation. [ref: 162] Iron is central to both pathways; it is required
to begin hydroxyl radical production in the first pathway, and to form
an iron-drug complex in the second. The risk of anthracycline-
associated congestive heart failure is increased by the presence of
preexisting heart disease or hypertension, prior radiation to the heart
or mediastinum, age (generally, children younger than 4 years are more
susceptible), prior cyclophosphamide therapy (questionable), and
cumulative anthracycline dose. For doxorubicin, a cumulative dose of
less than 450 mg/m**2, in the absence of other risk factors, is rarely
associated with cardiomyopathy, whereas as the cumulative dose
increases to 550, 600, and 700 mg/m**2, the incidence increases to 7%,
15%, and 30%, respectively. For daunorubicin, the overall incidence of
congestive heart failure is low (approximately 1.2% in a population of
5613 patients), and the incidence was related to the cumulative
daunorubicin dose administered. [ref: 163] At daunorubicin doses of 550
mg/m**2, the incidence of congestive heart failure was 4%, whereas at a
total dose of 1050 mg/m**2, the incidence increased to 14%. A
retrospective study of idarubicin cardiotoxicity found few clinically
significant symptoms at cumulative doses of 290 mg/m**2, although left
ventricular ejection fraction (LVEF) decreased during therapy. A
metaanalysis of studies comparing epirubicin and doxorubicin found
epirubicin as active as doxorubicin (1:1) and less cardiac toxic
(1:1.8). [ref: 164]
Patients receiving anthracyclines should be monitored for the onset
of cardiomyopathy. The most useful noninvasive test is serial
radionuclide angiocardiography, which provides a reproducible measure
of LVEF and is sensitive to subclinical cardiac dysfunction. This
technique and others have done much to allow earlier detection of
subclinical cardiac toxicity and to reduce the mortality associated
with it. Management of the cardiomyopathy involves bed rest and
afterload reduction. With continued conservative management, patients
may experience a gradual improvement; this improvement, however, may
take more than 1 year.
Many different approaches have been tried to prevent anthracycline-
induced cardiac toxicity, including analogue synthesis, altering dosing
schedules, and the use of biochemical antagonists. Although many of the
anthracycline analogues are associated with a lower incidence of
toxicity, they still are associated with decreased LVEF, albeit
asymptomatic. The use of continuous infusion schedules of
anthracyclines has reduced the incidence of cardiac toxicity somewhat,
providing a pharmacokinetic basis for the hypothesis that high peak
concentrations are associated with an increased incidence of
cardiotoxicity. [ref: 165] However, this approach is not widely used
because of concern over compromising antitumor efficacy, unpredictable
toxicities, and logistical issues.
Dexrazoxane (ICRF-187) is the first biochemical antagonist shown in a
randomized clinical trial to dramatically reduce the incidence of
cardiac toxicity in patients with breast cancer, without altering the
antitumor activity of the doxorubicin combination regimen. Although its
mechanism is not fully elucidated, dexrazoxane either acts as a free
radical scavenger or prevents formation of free radicals.
In addition to oral and intravenous administration, doxorubicin and
epirubicin have been administered via the intraarterial route for well-
defined, nonresectable metastatic and primary tumors of the liver.
Consistent with their metabolic profile, the systemic toxicity for
doxorubicin was not reduced using this route of administration, whereas
relatively low levels of epirubicin were produced because of
substantial first-pass effect. Intrapleural doxorubicin has been used
to treat malignant pleural effusions, and intravesical administration
is an effective therapy for recurrent, superficial transitional cell
bladder carcinoma.

Mitoxantrone and Losoxantrone

Cumulative cardiotoxicity limits the clinical use of the
anthracyclines, and much work has been directed at synthesizing related
compounds less prone to undergo reductive metabolism to form free
radicals. [ref: 166-168] Mitoxantrone, a member of the anthracenedione
class of synthetic antitumor compounds, is an example of one effort. It
lacks the sugar moiety of the anthracycline drugs, but retains the
planar polycyclic aromatic ring structure that permits its
intercalation into DNA. More important, mitoxantrone lacks the ability
to produce the quinone-type free radicals thought to be responsible for
anthracycline-associated cardiac toxicity. Losoxantrone (biantrazole;
CI-941) is a member of theanthrapyrazolegroup of antitumor agents,
which are structurally similar to the anthracyclines. The
anthrapyrazoles were synthesized to have antitumor activity similar to
the anthracyclines, but without cardiotoxicity. The rationale
underlying the development of the anthrapyrazoles was to make the
electron-deficient quinone chromophore of the anthracyclines more
resistant to enzymatic reduction by forming a quasi-iminoquinone. The
planar conformation and cationic nature of the anthracyclines,
necessary for intercalative binding to DNA, is retained in the
anthrapyrazoles. Losoxantrone can also cause cytotoxicity by inhibition
of top II. [ref: 169] Preclinical studies demonstrated that a number of
anthrapyrazoles had a spectrum of activity and potency similar to
doxorubicin and were superior to mitoxantrone.
The clinical pharmacokinetics of mitoxantrone and losoxantrone are
summarized in Table 19.6_8. Both compounds have very large
distribution volumes and are highly protein bound (percentage unbound,
<5%). [ref: 170,171] Both compounds undergo hepatic metabolism, but the
clinical relevance of these metabolites is unknown, as is whether the
dosage of either compound should be reduced in patients with hepatic
dysfunction. The low urinary recovery for both compounds suggests no
dosage adjustment is necessary for patients with mild to moderate renal
dysfunction.
The primary dose-limiting toxicity of both compounds is
myelosuppression, primarily neutropenia, although thrombocytopenia has
been reported. [ref: 172,173] Cumulative myelosuppression was not
noted. At higher doses, moderate nausea and vomiting have been
reported, but only mild nausea and vomiting at lower doses. Although
mitoxantrone is associated with less cardiotoxicity than the
anthracyclines, it can cause decreases in LVEF. For losoxantrone, data
for 183 patients showed mild to severe cardiotoxicity in 25 and
congestive heart failure in 2 patients.

Dactinomycin (Actinomycin D)

Dactinomycin is a member of a class of compounds first isolated from
Streptomyces parvullus. The only member to be clinically useful,
dactinomycin is highly effective in the treatment of Wilms' tumor,
Ewing's sarcoma, embryonal rhabdomyosarcoma, and gestational
choriocarcinoma. Responses have also been reported in testicular
cancer, Kaposi's sarcoma, and lymphoma. [ref: 174]
Dactinomycin is composed of a planar tricyclic ring chromophore
(phenoxazone) to which two identical cyclic polypeptides are attached.
[ref: 175] The compound binds to DNA by intercalation, depending on a
specific interaction between the polypeptide chains and deoxyguanosine.
This interaction blocks the ability of DNA to act as a template for RNA
and DNA synthesis in a concentration-dependent manner. Low drug
concentrations inhibit RNA synthesis more than higher drug
concentrations, which block both RNA and DNA syntheses. Dactinomycin
can also cause topoisomerase-mediated single-strand breaks in DNA,
although the contribution of these breaks to cytotoxicity is unclear.
[ref: 176] In vitro studies suggest that concomitant actinomycin D and
hyperthermia (e.g., 42**o or 43**oC) resulted in an increased uptake
and prolonged retention of actinomycin D. [ref: 177]
After an intravenous bolus, dactinomycin has a rapid disappearance
phase followed by a long elimination half-life of 36 hours. [ref: 178]
Dactinomycin is minimally metabolized, is concentrated in nucleated red
blood cells, and does not cross the blood-brain barrier. Dactinomycin
is primarily eliminated by renal and biliary excretion, although only
30% has been recovered in the urine and stool over the week after a
dose. Inadequate data exist to formulate dosage recommendations for use
of dactinomycin in patients with renal or hepatic dysfunction.
The dose-limiting toxicity for dactinomycin is myelosuppression,
affecting both platelets and neutrophils. Severe nausea and vomiting
occur during the first few hours after administration, but these side
effects are responsive to antiemetic therapy. Gastrointestinal toxicity
manifests primarily as stomatitis, accompanied by pain, and diarrhea
may occur. Extravasation of this drug after intravenous use may result
in severe soft tissue damage and ulceration. Dactinomycin is a
radiation sensitizer, leading to enhanced skin and gastrointestinal
toxicity when administered concurrently with radiation. [ref: 179] In
addition, late radiation damage to lung and liver appears increased,
potentially because of the ability of dactinomycin to block repair of
radiation-mediated DNA damage. This recall reaction can be observed
even when months separate radiation therapy and dactinomycin
administration.