de vita 19: agosto 2011

lunedì 22 agosto 2011

19_08

Chapter 19: Pharmacology of Cancer Chemotherapy
19.8: Miscellaneous Chemotherapeutic Agents

Bruce D. Cheson

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

Homoharringtonine

Homoharringtonine (HHT) and its congener, harringtonine, are
cephalotaxine esters isolated from the evergreen tree Cephalotaxus
fortunei Hook F, which is distributed throughout southern and
northeastern China. The two esters differ by only a single methylene
group, and they have similar activity against murine leukemia. [ref: 1]
The greater availability of HHT led to its further testing in the
United States.
The primary action of HHT appears to be inhibition of protein
synthesis, with degradation of polyribosomes, delayed inhibition of
initiation of protein synthesis, and inhibition of chain elongation by
interference with peptide bond formation. [ref: 2] DNA effects may also
be important, with a block in progression of cells from G(1) into S
phase and from G(2) phase into M phase. [ref: 3] Prolonged drug
exposure is necessary for maximal antileukemic effect in vitro.
Preclinical toxicology identified toxicities of the bone marrow,
gastrointestinal tract, kidneys, and heart. [ref: 4,5]
Radiolabeled HHT exhibits a triphasic plasma decay, with a terminal
half-life of 65.3 hours and an apparent volume of distribution of 2.4
L/kg. [ref: 6] In early phase I studies, HHT was administered as a
daily 10- to 360-minute infusion for up to 10 days. [ref: 7] Dose-
limiting cardiovascular toxicity with hypotension began 4 or more hours
after drug administration, which is presumed secondary to
vasodilatation with a compensatory increase in cardiac output.
Hypotension is ameliorated by interrupting the infusion or by fluid
administration, or both, and prolonging the duration of administration.
The major dose-limiting toxicity with currently used infusion schedules
is myelosuppression.
Initial clinical studies with HHT conducted in China identified
activity against acute myeloid leukemia and chronic-phase chronic
myelogenous leukemia (CML). [ref: 8] Clearing of central nervous system
blasts occurred after intrathecal administration. Variable activity was
observed in the initial series of phase II trials in U.S. pediatric and
adult patients with acute leukemia or myelodysplastic syndrome. [ref:
8-10] In late chronic-phase CML, a continuous intravenous infusion of
HHT, 2.5 mg/m**2/d, for 10 to 14 days each month induces complete
hematologic remission in 72% of cases, with cytogenetic response in
31%. [ref: 11] HHT is being combined with interferon or cytarabine in
early chronic-phase CML. O'Brien et al. [ref: 12] reported that HHT
followed by interferon did not achieve a higher cytogenetic response
rate than previously observed with HHT alone.
Nonmyelosuppressive toxicities have been minimal, including diarrhea,
hyperglycemia, nausea and vomiting, tachycardia, chest pain, headache,
and fatigue.

Suramin

Suramin is a polysulfonated napthylurea first used for the treatment of
onchocerciasis and trypanosomiasis in the 1920s. Its use against
parasites and discoid lupus erythematosus was abandoned because of the
availability of more effective drugs. Its inhibition of reverse
transcriptase and other RNA polymerases led to trials in patients with
autoimmune deficiency syndrome. [ref: 13] However, initial clinical
enthusiasm was not substantiated by additional study. [ref: 14,15]
The precise mechanism of suramin's antitumor action is unknown. The
drug binds nonspecifically to a wide variety of plasma proteins and
enzymes. It inhibits the binding and mitogenic action of many
polypeptide autocrine growth factors, including platelet-derived growth
factor, fibroblast growth factor, transforming growth factor-alpha and
-beta, epidermal growth factor, insulin-like growth factor-1 and -2,
interleukin-2, transferrin, and nerve growth factor. [ref: 16] It is
capable of dissociating growth factors from their receptors, with
higher affinity to heparin binding growth factors. It interferes with
glycosaminoglycan catabolism, leading to an accumulation in the liver
and blood of heparan sulfate and dermatan sulfate, which are thought to
be related to cell proliferation. Suppression of bone resorption may
contribute to the decreased pain reported in patients with prostate
cancer.
Suramin exhibits antitumor activity against a number of cell lines,
notably growth factor-responsive tumors, but low doses induce
proliferation in some cell lines. It inhibits the growth of malignant,
but not normal, prostate cells. Early clinical trials suggested
activity against adrenal, renal, and other cancers. [ref: 16] Activity
in prostate cancer has led to phase III studies.

Clinical Pharmacology

Suramin has limited absorption from the gastrointestinal tract. The
intravenous route is recommended because of better bioavailability.
The original dosing schedule of 1 g weekly for 6 weeks resulted in
plasma concentrations that fell over the first few hours after
administration but gradually increased over time, with increasing
trough levels before each injection. [ref: 17] The pharmacokinetics
were described by a two-compartment model, with an initial
(distribution) half life (t(1/2)alpha) of 2 days and an elimination
half life (t(1/2)beta) of 48 days (range, 44 to 54 days). Suramin is
99.7% protein bound, primarily to albumin, and may persist in the blood
for 3 months after administration, with no evidence of metabolism and
80% renal clearance. The total body clearance is only 0.41 mL/min, with
little interpatient variability. Suramin does not cross the blood-brain
barrier. It may displace other highly protein-bound drugs.

Dose and Schedule

The optimal schedule of administration is still being determined. [ref:
18-24] Earlier schedules used adaptive control feedback [ref: 25] in
which the timing and calculated dose were pharmacologically computed
for individual patients to maintain plasma concentrations in the range
of 200 to 300 ug/mL. Labor-intense pharmacologic monitoring was used
because of concern that the severe neurologic toxicity with suramin was
directly correlated with high blood levels. More recently, other
pharmacokinetic correlations have been postulated, including time above
a threshold concentration, total dose, and others. [ref: 26] The
relative importance of free drug concentration is unknown. However,
several phase I studies have determined that concentrations in the 200
to 300 ug/mL range are better tolerated overall. Phase I studies have
demonstrated little pharmacokinetic variability, making complex
adaptive control algorithms unnecessary. This observation has led to
investigation of a wide variety of schedules, including a 14-day
continuous infusion, intermittent short infusions, and intermittent
bolus administration. [ref: 20,27-29]
Suramin has modest activity in patients with prostate cancer. [ref:
19,27,28,30,31] Combinations with other agents have been studied in
prostate cancer and other solid tumors, but without clear additive
benefit. [ref: 32-37] The future of this drug is uncertain.

Toxicity

The most serious toxic effect of suramin is a polyneuropathy, which may
begin within several weeks of therapy and peaks in severity 3 to 6
months after the drug is discontinued. It ranges from mild stocking-
glove paresthesia to paralysis requiring mechanical ventilation, and it
is an indication to discontinue treatment. [ref: 38,39] At 350
mg/m**2/d by continuous infusion, a Guillain-Barre syndrome occurred in
11% of patients; the incidence increased to 40% with levels of more
than 350 ug/mL . [ref: 38] Suramin may lead to a progressive,
reversible myopathy; hyperesthesia of palms and soles; headache; and
altered taste. Adrenal insufficiency is very common and may be
irreversible. All patients receive concurrent corticosteroids until
normal adrenal function can be documented.
Infections are frequent with suramin therapy because the drug induces
lymphocytotoxicity and myelosuppression and inhibits phagocytosis and
bacterial killing, which is compounded by the addition of
hydrocortisone.
Other common toxicities include renal dysfunction, transaminase
elevations, and coagulopathy. [ref: 40,41] Prophylactic vitamin K has
been used to minimize the contribution from other causes. Bleeding is
managed by replacement of blood and plasma. Heparin can be given safely
using careful monitoring. An increase in serum creatinine or the
development of a coagulopathy necessitates interruption of therapy.
Other serious toxicities include supraventricular arrhythmias,
especially atrial fibrillation, pericardial effusions, and deep venous
thromboses.
Rash has been reported, occasionally with desquamation or toxic
epidermolysis as well as keratoacanthomas and superficial actinic
keratoses. Vortex keratopathy, which resolves after therapy, also has
been reported. Metabolic consequences include hyponatremia,
hypokalemia, hypocalcemia, hypermagnesemia, hypophosphatemia,
hypouricemia, and elevations in amylase and lipase. Rash and renal
dysfunction may not recur if the drug is resumed.

Bleomycin

The bleomycins are a group of glycopeptides originally extracted from a
strain of Streptomyces verticillus from culture broths obtained from
the soil of a Japanese coal mine. [ref: 42] The most active agent was a
mixture of peptides now known as bleomycin, with a molecular weight of
1200.
The primary action of bleomycin is to produce single- and double-
strand DNA breaks, which result from the production of free radicals by
an Fe(II)-bleomycin complex intercalated between opposing strands of
DNA. It is ineffective in producing strand breaks of native RNA or
synthetic ribonucleotide polymers. Cells are most sensitive to
bleomycin during the G(2) and M phases and least sensitive in the G(1)
phase. [ref: 43] Noncycling cells may be more sensitive than cycling
cells. The observation that cells were killed during G(2) suggested an
advantage for a continuous infusion, which was not supported by
clinical trials.

Cellular Pharmacology

Bleomycin is taken up by cells slowly and inactivated by an
aminohydrolase found in normal and malignant cells. [ref: 44] Hydrolase
levels are higher in animal species resistant to the pulmonary toxicity
of bleomycin and is low in lung and skin, the two organs most
susceptible to bleomycin toxicity. Levels in tumor cell lines do not
appear to correlate with drug resistance.

Clinical Pharmacology

Using a 4- to 5-day continuous intravenous infusion, the steady-state
concentration is approached approximately 12 hours after initiation of
infusion and ranged from 0.132 to 0.312 mu/mL. After an intravenous
bolus of 15 U/m**2, peak plasma concentrations reach 1 to 10 mu/mL,
with a rapid two-phase disappearance from plasma with a half-life of
elimination of approximately 3 hours. Approximately two-thirds of
excretion is renal, and the half-life increases rapidly in patients
with a creatinine clearance of less than 25 to 35 mL/min. There is
increased pulmonary toxicity with renal insufficiency, but no formal
guidelines for dose reduction have been determined. Bleomycin is
absorbed rapidly after intramuscular administration, resulting in peak
plasma concentrations approximately one-third to one-half of those
obtained after rapid intravenous administration. One hour after
intramuscular injection, maximum serum levels range from 0.13 to 0.35
mu/ml, with no drug detectable in the serum 24 hours after injection.
Absorption after subcutaneous injection has not been clearly defined.
Intracavitary administration of bleomycin achieves levels 10- to 22-
fold higher than simultaneous plasma levels and is effective in the
control of malignant effusions. [ref: 45-49] Approximately 45% of the
intracavitary dose is absorbed into systemic circulation. Bleomycin
also has been applied topically. [ref: 50] No pharmacologic advantage
to intraperitoneal administration has been reported.

Toxicity

A test dose of 1 mg of bleomycin is generally administered before a
weekly or twice-weekly dose of 5 to 15 U/m**2 because of the risk of
hypersensitivity with urticaria, periorbital edema, and bronchospasm.
The dose-limiting toxicity of bleomycin is pulmonary fibrosis of
uncertain pathogenesis, which occurs in 10% of patients and is more
common in patients older than 70 years, with doses of more than 400 U,
or in those with a history of chest radiotherapy and in the
postoperative period. [ref: 51-59] The onset is usually delayed, and
the initial physical examination and chest x-ray may be normal.
Eventually, rales, rhonchi, and pleural friction rubs are noted, and
abnormal pulmonary function, with decreased lung capacity and increased
lung stiffness, is seen. Clinical parameters better predict outcome
than pulmonary function studies. Chest x-ray may reveal increased
interstitial markings, patchy reticulonodular infiltrates,
consolidation, or nodules indistinguishable from metastatic lesions,
which may cavitate. Biopsy findings are nonspecific. After the drug is
discontinued, reversal may take months, and fibrosis may be only
partially reversible and may be fatal. A number of investigational
approaches are being evaluated to prevent or reduce the severity of
this complication. [ref: 54,55,60-62]
Myelosuppression and immunosuppression are not prominent. Fevers
occur in 20% to 50% of patients, occasionally with hyperpyrexia.
Mucocutaneous toxicities are common, with mucositis; alopecia; and
hyperpigmentation, erythema, induration, hyperkeratosis, and peeling
that may progress to ulceration. The digits, hands, joints, and areas
of prior radiation or surgery are most affected. Acute arthritis may
occur.

L-Asparaginase

The growth of malignant and normal cells depends on the availability of
specific nutrients and cofactors required for protein synthesis. Some
nutrients can be synthesized within the cell, whereas others, such as
essential amino acids, require exogenous sources. L-asparagine is a
nonessential amino acid synthesized by the transamination of L-aspartic
acid by a reaction catalyzed by the enzyme L-asparagine synthetase. The
ability to synthesize asparagine is notably lacking in malignancies of
lymphoid origin. In 1953, Kidd [ref: 63] first reported that the growth
of transplantable lymphomas of rat and mouse was inhibited by guinea
pig serum, and subsequent experiments demonstrated that the responsible
factor was L-asparaginase. Subsequent purification from Escherichia
coli and Erwinia carotovora permitted production of large quantities of
the enzyme for clinical use. The purified bacterial enzyme has a
molecular weight of 133,000 to 141,000 daltons and is composed of four
subunits, each with one active site. The enzymes are specific for the
l-isomer. Asparaginase catalyzes the conversion of L-asparagine to
aspartic acid and ammonia. The enzyme does not enter cells, instead
degrading circulating asparagine to aspartic acid, which cannot be
converted to asparagine by the cancer cell. In contrast, most normal
cells can synthesize asparagine from aspartic acid by induction of
asparagine synthetase. This metabolic difference is not absolute, as
demonstrated by the toxicity profile of the agent. Resistance occurs
through increased expression of the asparagine synthetase gene, which
is transcriptionally silent in most tissues and leads to increased
enzyme synthesis in response to a decrease in intracellular L-
asparagine levels. [ref: 64] Resistance may also be mediated by the
formation of asparaginase antibodies that alter asparaginase
pharmacokinetics.

Clinical Pharmacology

L-asparaginase is administered either intravenously or intramuscularly.
The intramuscular route produces peak blood levels 50% lower than the
intravenous route, but the former may be less immunogenic and is more
commonly used. Three preparations of L-asparaginase are in clinical
use. The most widely used is derived from E coli, and an Erwinia
preparation is available for patients who develop hypersensitivity to
the E coli-derived agent. The usual doses are 6000 IU/m**2 three times
weekly for 3 to 4 weeks, or daily doses of 5000 to 20,000 IU/m**2 for
10 to 20 days. The optimal dose and schedule are unknown. Intermittent
schedules with less frequent administration are associated with reduced
efficacy and increased anaphylaxis. An E coli preparation modified by
the covalent attachment of polyethylene glycol (PEG) has a prolonged
half-life, thus permitting lower doses and less frequent
administration. [ref: 65] The approved dose of PEG-asparaginase is 2500
IU/m**2 every 14 days, either intravenously or intramuscularly.
L-asparaginase concentration in plasma is proportional to a total
dose up to 200,000 IU/m**2 and falls with a primary half-life of 14 to
22 hours after administration. Blood levels of the E coli enzyme are
detectable for 1 to 2 weeks after a single dose, and concentrations of
L-asparagine fall below 1 mmol within minutes of enzyme injection and
remain low for 7 to 10 days after completion of therapy. The half-life
is independent of the dose administered, disease status, renal or
hepatic function, age, or gender. The pharmacokinetics of asparaginase
depend on the preparation. [ref: 66-68] With E coli-derived enzyme, the
t(1/2) (1.14 to 1.35 days) administered by the intramuscular route is
the same irrespective of dose (2500 or 25,000 IU/mL) or with repeated
doses. Peak serum levels are reached in 24 to 48 hours and are no
longer detectable in serum by 10 to 14 days. Extremely low levels are
found in the urine at 24 hours, suggesting clearance of L-asparaginase
by mechanisms other than urinary excretion. Cerebrospinal fluid levels
disappear rapidly. The serum t(1/2) for Erwinia is 0.65 days, and
enzyme was no longer detectable by 7 days. This value is shorter than
the E coli preparation, although similar schedules are often used. The
serum t(1/2) of PEG-modified L-asparaginase as an initial dose was 5.73
days, which is significantly longer than after subsequent doses. [ref:
66]
Patients who experience a hypersensitivity reaction to E coli
asparaginase have a decreased t(1/2) subsequently with PEG
asparaginase. [ref: 66,69] Serum L-asparaginase activity is
undetectable in the week after an anaphylactoid reaction. And even a
"silent" anaphylactoid reaction to E coli may result in neutralizing
antibodies and reduced drug efficacy. [ref: 70]

Toxicities

L-asparaginase has no effect on bone marrow function. Hypersensitivity
is the most serious toxicity, and it occurs in fewer than 10% of
patients. It is manifested by urticaria, nausea, vomiting, and chills,
and less often by a serum sickness-like reaction or by anaphylaxis with
hypotension, laryngospasm, and cardiac arrest, which is fatal in fewer
than 1% of patients. Reactions generally occur during the second week
of treatment or later, and they mandate a change to another
preparation. The risk of hypersensitivity is greater when the drug is
used as a single agent than with concurrent immunosuppressive agents
(steroids, 6-mercaptopurine), at doses higher than 6000 IU/m**2/d
administered by the intravenous route and with repeated courses of
treatment. Neither skin testing nor antibody levels have been
sufficiently predictive. The PEG formulation is the least immunogenic
[ref: 71,72] and may be more cost-effective. [ref: 73] The development
of an allergic reaction does not appear to compromise the efficacy of
the agent. [ref: 74]
Decreased protein synthesis leads to reduced albumin and serum
lipoprotein concentrations. A reduction in vitamin K-dependent clotting
factors, a fall in fibrinogen levels, and decreased platelet
aggregation to collagen may lead to bleeding. Decreases in antithrombin
III, proteins C and S, and increased endogenous thrombin generation are
associated with venous thrombosis and embolism. [ref: 75-81] Other
toxic effects include confusion, aphasia, stupor, or coma in 25% to 33%
of patients[ref: 82,83]; hyperlipidemia [ref: 84,85]; and abnormal
liver enzymes with fatty metamorphosis. L-asparaginase is
contraindicated in patients with a history of pancreatitis because of
the risk of acute pancreatitis. [ref: 86]

Amifostine

Amifostine is a phosphorylated aminothiol prodrug analogue of
cysteamine. It was developed by the Walter Reed Army Medical Institute
(thus the military code name WR-2721) during the cold war as part of a
classified research project to identify an agent that would protect
military personnel from radiation in the event of nuclear war.
Amifostine was found to afford greater protection against radiation
than more than 4000 other compounds screened. Nevertheless, the army
terminated development of this compound in 1988 because of its poor
oral bioavailability and the prohibitive nausea, vomiting, diarrhea,
and abdominal cramps noted with the oral formulation.
When administered intravenously, the pharmacokinetics (PK) of
amifostine varies somewhat with dose. [ref: 87] The clearance from
plasma is rapid (distribution and elimination phases in humans of
t(1/2)alpha, less than 1 minute; t(1/2)beta, 8.8 minutes), with a
plasma half-life of 1 minute and almost all drug cleared by the plasma
within 10 minutes. Bioavailability from the subcutaneous route is high
but variable. [ref: 87,88]
Amifostine is dephosphorylated at the tissue level to its active
metabolite, the free thiol WR-1065, by membrane-bound alkaline
phosphatase. WR-1065 is rapidly taken up by cells and is thought to be
the major cytoprotective metabolite. WR-1065 protects normal cells by
acting as a free radical scavenger and by hydrogen donation to repair
damaged target molecules. [ref: 89,90]
Preclinical studies with amifostine suggested that the agent could
protect normal tissues from radiation and chemotherapy toxicity without
protecting tumors. [ref: 91-97] Phosphorylation of the aminothiol
contributes substantially to the selective uptake of WR-1065 by normal
kidneys, bone marrow, heart, and salivary glands compared with tumor
tissues. Several explanations have been postulated for this
preferential uptake, such as that concentrations of alkaline
phosphatase are higher in normal tissues compared with malignant
tissues. The hypovascular, hypoxic nature of tumors results in
anaerobic metabolism and a low interstitial pH, which are associated
with a low rate of prodrug activation by alkaline phosphatase.
Separate phase I trials of amifostine were conducted in conjunction
with radiotherapy or chemotherapy. A true maximum tolerated dose was
not identified in either setting, but the recommended dose ranges from
740 mg/m**2 to 910 mg/m**2. [ref: 98-101] No clear therapeutic
advantage to the higher doses has been determined. [ref: 100,102,103]
Drug-related toxicity appears to correlate with the duration of the
infusion. [ref: 104]

Chemotherapy-Related Nephrotoxicity

Amifostine has been evaluated for its ability to prevent chemotherapy-
related nephrotoxicity, especially that induced by cisplatin. [ref:
100-102,105] In the only phase III trial, Kemp et al. [ref: 106]
randomized 242 women with advanced ovarian cancer to six cycles of
chemotherapy, with or without amifostine, 910 mg/m**2, before each
cycle. The severity of renal toxicity was reported to be lower in the
group receiving amifostine. Fewer patients discontinued therapy on the
amifostine arm because of toxicity. The response rates and survival
durations were comparable between the two arms. However, the doses of
cisplatin used in this study are higher than the dose currently
recommended, and this regimen is less commonly used than other less
nephrotoxic programs.

Neurologic Toxicities and Ototoxicity

Several small phase II studies and one phase III trial suggest that
amifostine may offer modest protection against the neurologic
toxicities of cisplatin, but with no effect on ototoxicity. [ref:
101,105,106]

Neutropenia and Thrombocytopenia

Various phase I, II, and III trials suggest a myeloprotective effect
from amifostine. Glover et al. [ref: 103] conducted a phase II trial of
amifostine in combination with cyclophosphamide with 21 patients used
as their own controls; 90% had an improved white blood cell count with
the second course of cyclophosphamide compared with the first course.
Whether these findings are clinically meaningful is questionable. [ref:
103,106,107] In a study conducted by the Cancer and Leukemia Group B,
[ref: 108] patients with solid tumors were treated with high-dose
cyclophosphamide with amifostine alone or with amifostine and
granulocyte-macrophage colony-stimulating factor. No additional
protection was noted with the combination.
Preliminary phase I data suggest less thrombocytopenia in patients
treated with carboplatin and amifostine. [ref: 109] However, the
aggregate data from subsequent studies provide less support for
clinically meaningful benefit. [ref: 108,110-113]

Additional Observations

Limited data suggest that amifostine may modulate the cardiotoxicity of
doxorubicin and the pulmonary toxicity of bleomycin. [ref: 54,55,114]

Radioprotection

Amifostine has been evaluated in combination with radiation therapy or
combined modality treatment for patients with head and neck and lung
cancers. [ref: 111,115-124] The suggestion has been made of a reduction
in esophagitis in lung cancer patients and less xerostomia and loss of
taste with amifostine, but with no clear impact on mucositis or
salivary gland function. [ref: 117-119] There is no clear
radioprotective effect in patients with rectal cancer. [ref: 98,125]

Toxicities

The major toxicities associated with amifostine include nausea and
vomiting, hypotension, hypocalcemia, and allergic reactions. The onset
of nausea and vomiting is generally within 15 to 30 minutes of the
start of the infusion, and they resolve spontaneously. Pretreatment
with dexamethasone and a 5-hydroxytryptamine receptor antagonist is
recommended.
Hypotension is a potentially serious side effect. It is generally
systolic, lasting 5 to 15 minutes, without central nervous system,
renal, or cardiovascular consequences. Administration issues that
influence the frequency and severity of hypotension include patient
hydration, infusion duration, position of patient, and antiemetic
pretreatment. Patients should not be receiving agents that potentiate
the potential for hypotension, and the drug should not be administered
to patients who cannot be without antihypertensive medications for at
least 24 hours. Dehydrated patients should not receive the drug until
the problem has been corrected. Patients should be hydrated before
administration of amifostine. Patients should remain supine or
reclining during amifostine therapy.
Hypocalcemia is clinically significant in approximately 1% of
patients and can be managed with oral calcium carbonate and vitamin D.
The drug has been used successfully to treat hypercalcemia. [ref: 126]
Allergic reactions occur in fewer than 1% of patients and are
successfully treated with diphenhydramine.

Drug Administration

Amifostine should be administered over 15 minutes, 5 to 30 minutes
before cytotoxic chemotherapy. The patient should be well hydrated and
in a reclining position, with frequent blood pressure monitoring. The
recommended dose with radiation therapy is 200 mg/m**2/d, as a slow
intravenous push over 3 minutes, 15 to 30 minutes before each radiation
fraction. Bolus schedules have been studied as well. [ref: 127]

Myelodysplastic Syndromes

Amifostine stimulates hematopoiesis in animal models, and in in vitro
studies it stimulates the formation of hematopoietic progenitors from
myelodysplastic syndrome bone marrow. [ref: 128] In a phase I/II study,
[ref: 129,130] the drug was administered at doses of 100, 200, or 400
mg/m**2 three times per week, or 740 mg/m**2 weekly for 3 weeks.
Hematologic improvement was observed in 83% of patients with the
thrice-weekly schedule, including either an increase in neutrophils or
a reduction in red blood cell transfusion requirements. More than 40%
of patients had a rise in their platelet counts. Acceleration to acute
myeloid leukemia was noted in several patients with RAEB-T (refractory
anemia with excess of blasts in transformation). In a subsequent
multicenter trial, [ref: 131] there was single or multilineage
improvement in 35%. A poor response rate was reported using a
continuous schedule of eight uninterrupted thrice-weekly doses of 300
to 450 mg/m**2. [ref: 132] The role of this agent in myelodysplastic
syndrome is being elucidated.

Recommendations for the Use of Amifostine

Based on a careful review of the data, the American Society of Clinical
Oncology made the following recommendations regarding the use of
amifostine[ref: 133]:
- It may be considered for the reduction of nephrotoxicity in
patients receiving cisplatin-based chemotherapy.
- Although it may be considered for the reduction of neutropenia in
patients receiving alkylating agents, chemotherapy dose reduction or
growth factor use should be considered as an alternative to the use of
amifostine.
- Present data are insufficient to recommend the use of amifostine
for protection against thrombocytopenia or the routine use of
amifostine to prevent cisplatin-associated neurotoxicity or
ototoxicity. Similarly, the data were felt to be insufficient to
support the use of amifostine for the prevention of paclitaxel-
associated neurotoxicity.
- The use of amifostine may be considered to decrease the incidence
of acute and late xerostomia in certain patients undergoing
fractionated radiation therapy in the head and neck region, although
the present data are insufficient to recommend the use of amifostine to
prevent radiation therapy-associated mucositis.

19_08

19_07

Chapter 19: Pharmacology of Cancer Chemotherapy
19.7: Antimicrotubule Agents

Eric K. Rowinsky
Anthony W. Tolcher

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

The microtubule is increasingly recognized as a strategic subcellular
target against which to direct therapeutic efforts, owing to the
widespread use of the vinca alkaloids in both curative and palliative
chemotherapeutic regimens and the successful incorporation of the
taxanes in cancer chemotherapeutics. This chapter reviews the vinca
alkaloids, taxanes, and estramustine and other novel antimicrotubule
agents in early development.

Microtubules

Microtubules are integral components of the mitotic spindle, which can
be disrupted by the vinca alkaloids, taxanes, and an increasing number
of both natural products and synthetic compounds, resulting in
metaphase arrest in dividing cells. [ref: 1-3] However, they are also
involved in nonmitotic functions, such as chemotaxis, membrane and
intracellular scaffolding, transport, secretory processes, anchorage of
subcellular organelles and receptors, cell adhesion, and locomotion
transmission of receptor signaling. Antimicrotubule agents may disrupt
a range of these nonmitotic functions. [ref: 1-3]
Microtubules are polymers of dimeric subunits of alpha- and beta-
tubulin (each tubulin subunit consisting of approximately 450 amino
acids with a molecular weight of 50,000 D) that are arranged into 13
protofilaments (Fig. 19.7_1). [ref: 1-4] The dimers are aligned side
by side around an apparently hollow core with the beta subunit of one
dimer in contact with the alpha-tubulin subunit of the next. The
microtubule polymer is in a dynamic equilibrium with the intracellular
pool of tubulin dimers, which results in the simultaneous incorporation
of free dimers into the polymerized structures and release of dimers
into the soluble tubulin pool. The direction of the equilibrium--toward
polymerization or depolymerization--is influenced by several cofactors,
including guanosine triphosphate (GTP), the ionic environment, and
microtubule-associated proteins (MAPs), which is a family of proteins
that regulate tubulin polymerization and microtubule function
(discussed later in the section Microtubule-Associated Proteins and
Microtubule Motors). Microtubule growth occurs spontaneously at the
plus end, resulting in the hydrolysis of GTP, which weakens the binding
affinity of tubulin for adjacent molecules. This, in turn, favors the
opposing process: depolymerization. Net shortening occurs at the
opposite minus end. In essence, microtubules are under the control of
two dynamic processes. The first is dynamic instability, which is the
process whereby microtubule ends switch spontaneously and
stochastically between slowly growing and rapidly shrinking states.
[ref: 4] The rate of dynamic instability is accelerated during some
processes, such as mitosis, so that chromosomes can readily be
"captured" by growing microtubules, thereby leading to the formation of
mitotic spindles; dynamic instability is suppressed, perhaps by MAPs,
during nonproliferative processes (e.g., differentiation). When both
these actions occur simultaneously, the microtubule is said to be
treadmilling, which plays a role in the polar movement of the
chromosomes during anaphase. [ref: 5]
There are at least six isotypes of both alpha- and beta-tubulin in
humans; they are distinguished by slightly different amino acid
sequences and appear to be encoded by different genes. [ref: 6-9] The
C-terminal amino acid sequence of beta-tubulin is the most variable in
terms of amino acid composition, and both posttranslational
modifications, including phosphorylation and glutamylation (which may
account in part for their structural diversity) have been described.
[ref: 9,10] Equivalent isotypes expressed in specific tissues of
different species are highly conserved, indicating that expression of
tubulin isotypes may be important in specific microtubule functions.
[ref: 7-10] Analysis of tubulin isotype expression in various tissues
has demonstrated a complex pattern of isotype distribution, suggesting
functional specificity. [ref: 7-10] In neurons, for example, isotype
segregation within cells, and both differential isotype synthesis and
posttranslational modification during neurite outgrowth, suggest
functional specialization. A third member of the tubulin superfamily,
gamma-tubulin, which is less abundant than the alpha and beta forms,
completes the microtubule-organizing center (MTOC) or centrosome. [ref:
11] Although tubulin can polymerize into microtubules in acellular
preparations, they are polymerized from, and nucleated by, the MTOC,
with minus ends located at the MTOC. [ref: 12] The MTOC in the
cytoplasm of mammalian cells duplicates and separates before cell
division, forming the two poles of the mitotic spindle.

Microtubule-Associated Proteins and Microtubule Motors

The dynamic behavior of microtubules is regulated by a variety of MAPs.
[ref: 1-3] The number of MAPs identified is increasing rapidly, and
these proteins appear to be diverse, differing from species to species
and cell type to cell type. Among the best characterized MAPs are those
that come from mammalian brain, including the tau proteins, MAP1, MAP1c
(an adenosine triphosphatase), MAP2, MAP4, and dynein (a GTPase), which
promote tubulin polymerization and microtubule stability. Some MAPs,
such as the dyneins and kinesins, function as microtubule motors,
transmitting chemical energy to mechanical sliding force and moving
various solutes and subcellular organelles along the microtubule. [ref:
1-3,13] Motor proteins function in many types of cellular events, such
as mitosis, premeiotic events, and organelle transport.

Vinca Alkaloids

The vinca alkaloids are naturally occurring or semisynthetic compounds
that are found in minute quantities in the periwinkle plant
Catharanthus roseus g. Don. [ref: 14-23] The early medicinal uses of
this plant led to the screening of these compounds for their
hypoglycemic activity, which was of little importance as compared to
their cytotoxic effects. Although many vinca alkaloids have been
investigated clinically, only vincristine (VCR), vinblastine (VBL), and
vinorelbine (VRL) are approved for use in the United States. The vinca
alkaloids are dimeric molecules composed of two multiringed units
(Fig. 19.7_2), an indole nucleus (catharanthine) and a dihydroindole
nucleus (vindoline). VCR and VBL are structurally identical except for
a single substitution on the vindoline nucleus, where VCR and VBL
possess formyl and methyl groups, respectively. Despite this small
difference, these two agents significantly differ in their antitumor
and toxicologic profiles. VCR is used more commonly in pediatric
oncology than in adults with cancer, most likely owing to the higher
level of sensitivity of pediatric malignancies to VCR and to the better
tolerance of higher VCR doses in children. VCR is an essential part of
the combination chemotherapeutic regimens used for acute lymphocytic
leukemia and lymphoid blast crisis of chronic myeloid leukemia and
plays an important role in the treatment of both Hodgkin's and non-
Hodgkin's lymphomas. The agent also plays a role in the multimodality
therapy of Wilms' tumor, Ewing's sarcoma, neuroblastoma, and
rhabdomyosarcoma in children, as well as in the treatment of multiple
myeloma and small cell lung cancer in adults. VBL has been an integral
component of chemotherapeutic regimens for germ cell malignancies and
advanced lymphoma and is used in combination with other agents to treat
Kaposi's sarcoma and bladder, brain, and breast cancers. [ref: 14-17]
Deacetyl VBL (vindesine, or VDS), initially identified as a
metabolite of VBL, was introduced in the 1970s. [ref: 18] VDS is
registered in many countries but available only for investigational
purposes in the United States. The agent is most commonly used in
combination with other agents, particularly the platinating agents or
mitomycin C (or both), in treating non-small cell lung cancer, but it
is also active in several hematologic and solid neoplasms. [ref: 17,18]
The semisynthetic VBL derivative VRL (5'-norhydro-VBL), which is
structurally modified on its catharanthine nucleus, is approved in the
United States for treating non-small cell lung cancer as either a
single agent or in combination with cisplatin and has been registered
for advanced breast cancer in many other countries. [ref: 19,20] VRL
has also demonstrated anticancer activity in advanced ovarian carcinoma
and lymphoma; however, a unique role in the therapy of these
malignancies has not been defined.

Mechanism of Action

The principal mechanism of cytotoxicity of the vinca alkaloids is by
interacting with tubulin and disrupting microtubule function,
particularly of microtubules that compose the mitotic spindle
apparatus, leading to metaphase arrest. [ref: 21-25] However, they are
also capable of many other biochemical and biologic activities that may
or may not be related to their effects on microtubules. [ref: 26] In
support of antimicrotubule actions or, more specifically, antimitotic
actions as the principal cytotoxic effect of the vinca alkaloids is
that the dissolution of the mitotic spindle apparatus, appearance of
mitotic figures, and cytotoxicity strongly correlate with both the
duration and concentration of drug treatment. [ref: 26] Nevertheless,
the vinca alkaloids and other antimicrotubule agents also affect both
nonmalignant and malignant cells in the nonmitotic cell cycle, which is
not surprising, as microtubules are involved in many nonmitotic
functions.
The vinca alkaloids bind to sites on tubulin that are distinct from
the binding sites of the taxanes, colchicine, podophyllotoxin, and GTP.
[ref: 20-25] Binding is rapid and readily reversible. There appear to
be two binding sites per mole of tubulin dimer. Vinca alkaloid binding
to tubulin induces tubulin to self-associate into nonmicrotubule
polymers and ordered aggregates through a self-association pathway,
which in turn increases the affinity of one of the binding sites for
the drug. The vinca alkaloid self-association of tubulin can lead to
the formation of paracrystalline structures in vitro, which generally
occurs at high drug concentrations. The vinca alkaloids bind to their
binding sites in intact microtubules with different affinities,
depending on whether the binding sites are located at the microtubule
ends or situated along the microtubule surface. [ref: 25] There are
approximately 16 to 17 high-affinity binding sites per microtubule
(K(d), 1 to 2 umol) located at the ends of each microtubule. Binding of
the vinca alkaloids to these sites disrupts microtubule assembly. The
main effect of low drug concentrations is to decrease the rates of both
growth and shortening at the assembly end of the microtubule, which in
effect produces a "kinetic cap" and suppresses function. [ref: 21-25]
The potent kinetic suppression of tubulin exchange that occurs at low
vinca alkaloid concentrations (<1 umol) is almost certainly due to drug
binding at the high-affinity sites at the microtubule ends. This action
suppresses dynamic instability and increases the time that microtubules
spend in a state of attenuated activity, neither growing nor
shortening. The disruptive effects of the vinca alkaloids on
microtubule dynamics, particularly at the ends of the mitotic spindle,
which leads to metaphase arrest, occur at drug concentrations below
those that decrease microtubule mass. There are also one to two low-
affinity binding sites per mole of tubulin dimer (K(d), 0.25 to 3.0
mmol) along the microtubule surface. [ref: 25] Binding of the vinca
alkaloids to these sites appears to be responsible for the splaying of
microtubules into spiral aggregates or spiral protofilaments, which
leads to microtubule disintegration. This effect occurs at high drug
concentrations (>1 to 2 umol) by a self-propagated mechanism, initially
involving drug binding to a limited number of sites, which
progressively weakens the lateral interactions between the
protofilaments and thereby exposes new sites. Spiral protofilaments may
then associate to form paracrystals.
Despite the wide range of sensitivities of different tissues to the
actions of the vinca alkaloids in vivo, the qualitative effects of
these agents on tubulin, as well as both tubulin-binding and inhibitory
constants, are similar. The differential sensitivities of various
tissues appear to be multifactorial. One of the most likely
explanations is that each tissue type has a distinct tubulin isotype
composition and that vinca alkaloid sensitivity is, in part, tubulin
isotype-dependent. In addition, differences in the tissue content of
cofactors, such as MAPs and GTP, which may influence drug interactions
with tubulin, and variability in cellular permeation and retention may
influence the formation and stability of vinca alkaloid-tubulin
complexes. [ref: 26-36] Differences in the pharmacokinetics between the
vinca alkaloids may also contribute to differential tissue sensitivity.
The vinca alkaloids are rapidly taken up into cells and then
accumulate intracellularly, with intracellular-extracellular
concentration ratios as high as 5- to 500-fold, depending on the cell
type. [ref: 37-39] In murine leukemia cells, the intracellular
concentrations of VCR are 5- to 20-fold higher than the extracellular
concentrations, and this ratio has been reported to range from 150- to
500-fold for other vinca alkaloids in human leukemia cell lines. [ref:
40] In isolated human hepatocytes, VRL is more rapidly taken up and
metabolized than other vinca alkaloids. [ref: 39,40] There are also
marked differences in cellular retention between the vinca alkaloids.
[ref: 31,39,41,42,43] VBL is retained to a much greater degree than
either VCR or VDS. Overall, the most important determinant of the rates
of drug accumulation and retention is lipophilicity. [ref: 39] Drug
uptake and retention may also be tissue-specific as well as drug-
specific, as illustrated by studies indicating that the accumulation
and retention of VRL in neurons are much less than with other vinca
alkaloids. [ref: 31]
It was originally believed that the vinca alkaloids entered cells by
both energy-dependent and temperature-dependent transport processes.
[ref: 39] However, it appears that temperature-independent,
nonsaturable mechanisms, analogous to simple diffusion, account for the
majority of drug transport, and temperature-dependent, saturable
processes are less important. [ref: 39-41] Although the drug
concentration and duration of treatment are important determinants of
both drug accumulation and cytotoxicity, the duration of drug exposure
above a critical threshold concentration appears to be the most
important determinant. [ref: 41,42]

Mechanisms of Resistance

Two mechanisms of resistance to the vinca alkaloids in vitro have been
well characterized. The first is pleiotropic or multidrug resistance
(MDR), which can be innate or acquired. MDR-mediating proteins include
permeability glycoprotein (P-gp), MDR protein, and lung resistance
protein, which are overexpressed in resistant cells and function as
drug efflux pumps. [ref: 43-50 ]The best characterized mechanism is
mediated by the 170-kD P-gp drug efflux pump that is encoded by the
mdr1 gene and results in decreased drug accumulation. The MDR phenotype
confers varying degrees of cross-resistance to other structurally bulky
natural products, such as the taxanes, anthracyclines,
epipodophyllotoxins, and colchicine. [ref: 44-50] The amino acid
sequence of the specific P-gp associated with resistance to the vinca
alkaloids differs slightly from P-gp of cells selected for resistance
to other agents. [ref: 47,48] These proteins also undergo
posttranslational modifications, resulting in further structural
diversity, which may explain the greater degree of resistance for the
specific agent, in which resistance was selected against, and the
variable degrees of resistance to agents aside from that specific
agent. The composition of membrane gangliosides in VCR-resistant cells
has also been demonstrated to be different from wild-type cells, which
may have functional significance. [ref: 49 ]The clinical ramifications
of these mechanisms are not entirely known. In one study in childhood
acute lymphoblastic leukemia, VCR resistance measured in vitro did not
correlate with P-gp overexpression. [ref: 50] Although many types of
agents reverse resistance conferred by P-gp in vitro and the role of
MDR modulators has been a source of great contemporary interest, the
interpretation of clinical studies of resistance modulation has been
confounded by the fact that MDR modulators also enhance drug uptake in
normal cells, decrease biliary elimination and drug clearance, and lead
to enhanced toxicity. [ref: 51,52] Overall, strategies aimed at
reversing resistance to the vinca alkaloids in the clinic with
pharmacologic modulators of MDR have been disappointing. [ref: 52]
Structural and functional alterations in alpha- and beta-tubulins,
resulting from either genetic mutations or posttranslational
modifications, have also been identified in tumor cells with acquired
resistance to the vinca alkaloids. [ref: 53-55] Tubulin alterations may
result in either decreased drug-binding affinity of the altered tubulin
or increased resistance to microtubule disassembly. These "hyperstable"
microtubules are collaterally sensitive to the taxanes, which inhibit
microtubule disassembly (discussed later in Taxanes, Mechanisms of
Resistance). Although the precise mechanisms that lead to cell death
after treatment with the vinca alkaloids are not entirely clear, these
mechanisms appear similar to those that have been elucidated for the
taxanes and involve the action of such genes as p53, bcl-2, and bcl-x
and gene products that trigger programmed cell death or apoptosis after
significant microtubule disruption. [ref: 56,57]

Pharmacology

General Overview

The vinca alkaloids are usually administered intravenously as a brief
infusion, and their pharmacokinetic behavior in plasma is optimally
described by three-compartment models. Table 19.7_1 displays several
pertinent pharmacokinetic features of these agents. At conventional
adult doses, peak plasma concentrations range from 100 to 500 nmol, but
levels of this magnitude are sustained in plasma for only short periods
(alpha half-lives, <5 minutes). [ref: 15,17-19,40,58-61] The vinca
alkaloids share many pharmacokinetic characteristics, including large
volumes of distribution, high clearance rates, and long terminal half-
lives, which reflect the high magnitude and avidity of drug binding in
peripheral tissues. There is also great interindividual and
intraindividual variability in their pharmacologic behaviors, which has
been attributed to many factors, including differences in protein
binding and both hepatic and biliary clearance. [ref: 40] Although it
has been proposed that prolonged infusion schedules may avoid excessive
toxic peak concentrations and increase the duration of drug exposure in
plasma above biologically relevant threshold concentrations for any
given tumor, there is little (if any) evidence to support the notion
that prolonged infusion schedules are more effective than bolus
schedules. This approach has primarily been directed at achieving
plasma concentrations that likely underestimate drug concentrations in
peripheral tissues where binding is high and avid, owing to the
ubiquitous nature of tubulin.
In comparative studies of the vinca alkaloids, VCR had the longest
terminal half-life and the lowest clearance rate, VBL had the shortest
terminal half-life and the highest clearance rate, and VDS had
intermediate characteristics. [ref: 15,18-20,59,60] Comparable values
for VLR overlap with those of VDS and VBL. The longest half-life and
lowest clearance rate of VCR may account for its greater propensity to
induce neurotoxicity, [ref: 58,59] but there are many other
nonpharmacologic determinants of tissue sensitivity (discussed earlier
in the section Mechanism of Action under Vinca Alkaloids).

Vincristine

After conventional doses of VCR (1.4 mg/m**2) given as brief infusions,
peak plasma levels approach 0.4 umol. [ref: 15,17,18,62,63] VCR binds
extensively to both plasma proteins (48%) and formed blood elements,
particularly platelets, which contain high concentrations of tubulin
and led, in the past, to the use of VCR-loaded platelets for treating
disorders of platelet consumption, such as idiopathic thrombocytopenia
purpura. [ref: 17] The platelet count inversely has been demonstrated
to influence drug exposure. [ref: 17,64 ]Penetration of VCR across the
blood-brain barrier is poor, probably because of its large size and the
fact that it is an avid substrate for the multidrug transporter pumps
that maintain the integrity of the blood-brain barrier. [ref:
15,18,40,64-69] Plasma clearance is slow, and terminal half-lives range
from 23 to 85 hours. [ref: 15,17,18,40,58,59,62]
VCR is metabolized and excreted primarily by the hepatobiliary
system. Seventy-two hours after the administration of radiolabeled VCR,
approximately 12% of the radiolabel is excreted in the urine (50% of
which consists of metabolites), and approximately 70% is excreted in
the feces (40% of which consists of metabolites). [ref:
15,18,40,63,65,69,70] The nature of the VCR metabolites identified to
date, as well as the results of metabolic studies in vitro, indicate
that VCR metabolism is mediated by hepatic cytochrome P-450 CYP3A.
[ref: 15,18,40 ]There has been conflicting, albeit sparse, evidence
indicating that peak VCR plasma concentration or systemic exposure
correlates positively with the degree of neurotoxicity. [ref: 15]

Vinblastine

The clinical pharmacology of VBL is similar to that of VCR. Binding of
VBL to plasma proteins and formed elements of blood is extensive. [ref:
40,71,72] Peak plasma drug concentrations are approximately 0.4 umol
after rapid intravenous injections of VBL at standard doses.
Distribution is rapid, and terminal half-lives range from 20 to 24
hours. [ref: 17,40,58,59,66,72,73] Tissue sequestration appears to be
greater for VBL than VCR, with 73% of radioactivity retained in the
body 6 days after treatment with radiolabled drug. [ref: 72] Like VCR,
VBL disposition is principally through the hepatobiliary system. [ref:
40] Fecal excretion of the parent compound is low, indicating that the
metabolism is significant. In vitro studies indicate that the
cytochrome P-450 CYP3A isoform is primarily responsible for the drug
biotransformation. [ref: 40,73] Although the metabolic fate of VBL has
not been fully characterized, 4-deacetyl-VBL, or VDS, which appears to
be as active as the parent compound, is the principal metabolite of
VBL. [ref: 72]

Vindesine

VDS is rapidly distributed to tissues, and terminal half-lives range
from 20 to 24 hours. [ref: 18,40,58,59,69,74-79] The large volume of
distribution, low renal clearance, and long terminal half-life of VDS
suggest that it undergoes extensive tissue binding and delayed
elimination and that drug accumulation may occur with repeated
administration at short intervals. Although peak VDS concentrations
ranging from 0.1 to 1.0 umol are achieved with rapid injections, levels
typically decline to less than 0.1 umol in 1 to 2 hours after
treatment. Plasma levels achieved with rapid injections are
approximately 16-fold higher than those achieved with protracted
infusions; however, prolonged periods of exposure above concentrations
resulting in cytotoxicity in vitro (0.01 to 0.1 umol) are readily
achieved using protracted infusions (1.2 to 2.0 mg/m**2/d for 2 to 5
days). [ref: 18,40,62,76-80] Renal clearance is negligible, accounting
for 1% to 12% of drug disposition. [ref: 18,40,77,79] Similar to the
other vinca alkaloids, VDS disposition is primarily by hepatic
metabolism and biliary clearance, and the cytochrome P-450 isoform
CYP3A plays a major role in drug metabolism. [ref: 18,40,67,81,82]

Vinorelbine

The pharmacologic behavior of VRL is essentially similar to that of the
other vinca alkaloids, with plasma concentrations declining in either a
biexponential or triexponential manner. [ref: 17,19,20,40,83-85] After
intravenous administration, there is a rapid decay of VRL
concentrations followed by a much slower elimination phase (terminal
half-life, 18 to 49 hours). Plasma protein binding has been reported to
range from 80% to 91%, with binding primarily to alpha(1)-acid
glycoprotein, albumin, and lipoproteins, [ref: 19,20,40,86] and drug
binding to platelets is also extensive. The unbound fraction has been
reported to range from 0.09 to 0.20. [ref: 19]
VRL is widely distributed, and high concentrations are found in
virtually all tissues, except brain. [ref: 19,20,40,87,88] The wide
distribution of VRL reflects the agent's lipophilicity, which is among
the highest of the vinca alkaloids. [ref: 22] In fact, drug
concentrations in human lung have been demonstrated to be 300-fold
greater than plasma levels and 3.4- to 13.8-fold higher than lung
concentrations achieved with VDS and VCR, respectively. [ref: 87] As
with other vinca alkaloids, the liver is the principal excretory organ,
and 33% to 80% of the drug is excreted in the feces, whereas urinary
excretion represents only 16% to 30% of total drug disposition, the
bulk of which is unmetabolized VLR. [ref: 18,19,40,89,90] Studies in
humans indicate that 4-O-deacetyl-VRL, 3,6-epoxy-VRL, and several
hydroxy-VRL isomers are the principal metabolites. [ref: 19,20,90]
Although most metabolites are inactive, deacetyl-VRL may be as active
as VRL. The cytochrome P-450 CYP3A isoenzyme appears to be principally
involved in biotransformation. [ref: 19,20,40] Human studies of powder-
and liquid-filled gelatin capsules have shown that bioavailability of
the parent compound is 43% for the powder-filled and 27% for the
liquid-filled capsules. [ref: 85,91] Plasma concentrations peak within
1 to 2 hours after oral treatment, and interindividual variability is
moderate.

Drug Interactions

Methotrexate accumulation in tumor cells is enhanced in vitro by the
presence of VCR or VBL, an effect mediated by a vinca alkaloid-induced
blockade of drug efflux; however, the minimal concentrations of VCR
required to achieve this effect occur only transiently in vivo. [ref:
92-94] The vinca alkaloids also inhibit the cellular influx of the
epipodophyllotoxins in vitro, resulting in less cytotoxicity, but the
clinical ramifications of this effect are unknown. [ref: 95] l-
Asparaginase may reduce the hepatic clearance of the vinca alkaloids,
particularly VCR, which may result in increased toxicity. To minimize
the possibility of this interaction, VCR should be given 12 to 24 hours
before l-asparaginase.
Treatment with the vinca alkaloids has precipitated seizures
associated with subtherapeutic plasma phenytoin concentrations. [ref:
94,96,97] Reduced plasma phenytoin levels have been noted from 24 hours
to 10 days after treatment with both VCR and VBL. Because of the
importance of the cytochrome P-450 CYP3A isoenzyme in vinca alkaloid
metabolism, administration of the vinca alkaloids with erythromycin and
other inhibitors of CYP3A may lead to severe toxicity. [ref: 98]
Concomitantly administered drugs, such as pentobarbital and H(2)-
receptor antagonists, may also influence VCR clearance by modulating
hepatic cytochrome P-450 metabolic processes. [ref: 94,99] Another
potential drug interaction may occur in patients who have Kaposi's
sarcoma related to acquired immunodeficiency syndrome (AIDS) and are
receiving concurrent treatment with 3' azido-3'-deoxythymidine (AZT)
and the vinca alkaloids, as the vinca alkaloids may inhibit
glucuronidation of AZT to its 5'-O-glucuronide metabolite. [ref: 100]

Toxicity

Despite close similarities in structure, the vinca alkaloids differ
significantly in their toxicologic profiles. VCR principally induces
neurotoxicity characterized by a peripheral, symmetric mixed sensory-
motor, and autonomic polyneuropathy. [ref: 15,17-20,101-105] The
primary pathologic effects are axonal degeneration and decreased axonal
transport due to interference with microtubule function. Initially,
only symmetric sensory impairment and paresthesias in a length-
dependent manner (distal extremities first) usually are encountered.
Neuritic pain and loss of deep tendon reflexes may develop with
continued treatment, which may be followed by foot drop, wrist drop,
motor dysfunction, ataxia, and paralysis. Back, bone, and limb pains
occasionally occur. Nerve conduction velocities are usually normal,
although diminished amplitude of sensory and motor nerve action
potentials and prolonged distal latencies, suggesting axonal
degeneration, may be noted. [ref: 15,101,104,105] Cranial nerves may be
affected rarely, resulting in hoarseness, diplopia, jaw pain, and
facial palsies. The uptake of VCR into the brain is low, and central
nervous system effects, such as confusion, mental status changes,
depression, hallucinations, agitation, insomnia, seizures, coma,
inappropriate secretion of antidiuretic hormone (SIADH), and visual
disturbances, are rare. [ref: 15,65-67,106] Acute, severe autonomic
neurotoxicity is uncommon but may arise as a consequence of high-dose
therapy (>2 mg/m**2) or in patients with altered hepatic function.
Toxic manifestations include constipation, abdominal cramps, paralytic
ileus, urinary retention, orthostatic hypotension, and hypertension.
[ref: 107-109] Laryngeal paralysis has also been reported. [ref: 110]
In adults, neurotoxic effects may begin with cumulative doses as
little as 5 to 6 mg, and manifestations may be profound after
cumulative doses of 15 to 20 mg. Children may be less susceptible than
adults, but the elderly are particularly prone. However, the apparent
influence of age may, in fact, be due to previously inadequate dose
calculation by body weight in children and adults and by body surface
area in infants. [ref: 103,104,111,112] In infants, VCR doses are
calculated now according to body weight. Patients with antecedent
neurologic disorders, such as Charcot-Marie-Tooth disease, hereditary
and sensory neuropathy type 1, Guillain-Barre syndrome, and childhood
poliomyelitis, are highly predisposed. [ref: 113,114] VCR treatment in
patients with hepatic dysfunction or obstructive liver disease is
associated with an increased risk of developing neuropathy because of
impaired drug metabolism and delayed biliary excretion.
The only known treatment for VCR neurotoxicity is discontinuation of
the drug or reduction of the dose or frequency of treatment. [ref: 116]
Although a number of antidotes, including thiamine, vitamin B(12),
folinic acid, and pyridoxine, have been used, these treatments have not
been clearly shown to be effective. [ref: 15,17,116] However, results
with several other protective agents appear promising. [ref:
15,17,115,116] In one randomized, double-blind trial, coadministration
of glutamic acid and VCR has been demonstrated to decrease
neurotoxicity. [ref: 15,117] The adrenocorticotropic hormone (4-9)
analogue ORG 2766 has also been shown to protect against VCR-induced
neuropathy both in an animal model and in cancer patients in a double-
blind, placebo-controlled pilot study. [ref: 15] However, the younger
age of the ORG 2766-treated group as compared to the placebo group may
have accounted for this result. Experimental results indicate that
several other agents, such as nerve growth factor, insulin-like growth
factor I, and amifostine, might alter the natural course of drug-
induced neurotoxicity. [ref: 15]
The manifestations of neurotoxicity are similar for the other vinca
alkaloids; however, they are typically less common and severe. [ref:
15,17-20] Severe neurotoxicity is observed infrequently with both VBL
and VDS. VRL has been shown to have a lower affinity for axonal
microtubules than either VCR or VBL, which seems to be confirmed by
clinical observations. [ref: 31,118] Mild to moderate peripheral
neuropathy, principally characterized by sensory effects, occurs in 7%
to 31% of patients, and constipation and other autonomic effects are
noted in 30% of subjects, whereas severe toxicity occurs in 2% to 3%.
Muscle weakness, jaw pain, and discomfort at tumor sites may also
occur. In a study in patients with non-small cell lung cancer randomly
assigned to treatment with either VRL alone, VRL plus cisplatin, or VDS
plus cisplatin, the rate of severe neurotoxicity was lower in both the
single-agent VRL and VRL plus cisplatin arms than in the VDS plus
cisplatin arm. [ref: 119] Furthermore, the addition of cisplatin did
not significantly increase the incidence of severe toxicity in excess
of that observed with VRL alone.
Neutropenia is the principal dose-limiting toxicity of VBL, VDS, and
VRL. Thrombocytopenia and anemia are usually less common and less
severe. The onset of neutropenia is usually 7 to 11 days after
treatment, and recovery is generally by days 14 to 21. Myelosuppression
is not typically cumulative.
Gastrointestinal toxicities, aside from those caused by autonomic
dysfunction, may be caused by all the vinca alkaloids. [ref: 15-20,120]
Gastrointestinal autonomic dysfunction, as manifested by bloating,
constipation, ileus, and abdominal pain, occur most commonly with VCR
or high doses of the other vinca alkaloids. Mucositis occurs more
frequently with VBL than with VRL or VDS and is least common with VCR.
Nausea, vomiting, and diarrhea may also occur to a lesser extent.
Pancreatitis has also been reported with VRL. [ref: 121]
All vinca alkaloids are potent vesicants and may cause significant
tissue damage if extravasation occurs. If extravasation occurs or is
suspected, treatment should be discontinued, and aspiration of any
residual drug remaining in the tissues should be attempted. [ref:
122,123] The application of local heat and the injection of
hyaluronidase, 150 mg subcutaneously, in a circumferential manner
around the needle site are thought to minimize both discomfort and
latent cellulitis. Phlebitis may also occur along the course of an
injected vein, with resultant sclerosis. The risk of phlebitis may
increase if veins are not adequately flushed after treatment.
Mild and reversible alopecia occurs in approximately 10% and 20% of
patients treated with VLR and VCR, respectively. Acute cardiac
ischemia, chest pains without evidence of ischemia, fever without an
obvious source, acute pulmonary effects (alone or in combination with
mitomycin C), Raynaud's phenomenon, hand-foot syndrome, and both
pulmonary and liver toxicity have also been reported with the vinca
alkaloids. [ref: 124-131] All the vinca alkaloids have been implicated
as a cause of SIADH, and patients who are receiving intensive hydration
are particularly prone to severe hyponatremia secondary to SIADH. [ref:
15,17-20] This entity has been associated with elevated plasma levels
of antidiuretic hormone and usual remits in 2 to 3 days. Hyponatremia
generally responds to fluid restriction, as with hyponatremia
associated with SIADH due to other causes.

Administration

VCR is commonly administered to children weighing more than 10 kg as a
bolus intravenous injection at a dose of 1.5 to 2.0 mg/m**2 weekly,
whereas 0.05 to 0.65 mg/kg weekly is commonly used in smaller children.
For adults, the conventional weekly dose is 1.4 mg/m**2. A restriction
of the absolute dose of VCR to 2.0 to 2.5 mg in children and 2.0 mg in
adults (often called capping) has been adopted based on early reports
of substantial gastrointestinal toxicity in small numbers of patients
treated at higher doses. However, this practice is largely unfounded,
and available evidence suggests that it should be reconsidered,
particularly in light of the wide interpatient variability in
pharmacokinetic behavior and tolerance. [ref: 103] There is significant
interpatient variability in the clearance of VCR (as much as 11-fold),
and some patients are able to tolerate much higher doses with little or
no toxicity. Moreover, the safety and efficacy of treatment regimens
that do not employ capping at 2.0 mg have been documented in adults.
[ref: 132] In any case, doses should not be reduced for mild peripheral
neurotoxicity, particularly if the agent is being used in a potentially
curative setting. Instead, doses should be modified for manifestations
indicative of more serious neurotoxicity, including severe symptomatic
sensory changes, motor and cranial nerve deficits, and ileus, until
toxicity resolves. In clearly palliative situations, dose reductions,
lengthened dosing intervals, or selection of an alternative agent may
be justified in the event of moderate neurotoxicity. A routine
prophylactic regimen to prevent severe autonomic toxicity, particularly
severe constipation, is also recommended.
The most common schedule for VBL in combination chemotherapeutic
regimens uses a rapid intravenous injection at a dose of 6 mg/m**2
weekly. Approved dosing recommendations for weekly dosing are 2.5 and
3.7 mg/m**2 for children and adults, respectively, followed by gradual
escalation in increments of 1.8 and 1.25 mg/m**2 weekly based on
hematologic tolerance. It is recommended that weekly VBL doses of 18.5
mg/m**2 in adults and 12.5 mg/m**2 in children not be exceeded as a
single agent; however, these doses are substantially higher than most
patients can tolerate because of myelosuppression, even on less-
frequent schedules of administration. Because the severity of
leukopenia that may occur with identical VBL doses varies widely, VBL
should probably not be given more frequently than once each week.
VDS has been administered intravenously on many schedules, including
weekly and biweekly bolus and prolonged infusion schedules. The agent
has also been given in fractionated doses as either an intermittent or
a continuous infusion over 1 to 5 days. VDS is most commonly
administered as a single intravenous dose of 2 to 4 mg/m**2 every 7 to
14 days. Intermittent or continuous infusion schedules usually employ
VDS doses of 1 to 2 mg/m**2/d for 1 to 2 days or 1.2 mg/m**2/d for 5
days every 3 to 4 weeks. [ref: 62]
VRL is usually administered at a dose of 30 mg/m**2 on a weekly or
biweekly schedule as a 6- to 10-minute intravenous injection through a
side-arm port into a running infusion (alternatively, a slow bolus
injection followed by flushing the vein with 5% dextrose or 0.9% sodium
chloride solutions) or as a short infusion over 20 minutes. [ref: 17-
20] It appears that the more rapid infusions are associated with less
local venous toxicity. Oral doses of 80 to 100 mg/m**2 given weekly are
generally well tolerated, but an acceptable oral formulation has not
yet been approved. Other dosing schedules that have been evaluated
include chronic oral administration of low doses, intermittent high
dose, and prolonged intravenous infusion schedules.
The vinca alkaloids are potent vesicants and should not be
administered intramuscularly, subcutaneously, intravesically, or
intraperitoneally. Direct intrathecal injection of VCR and other vinca
alkaloids, which has occurred as an inadvertent clinical mishap,
induces a severe myeloencephalopathy characterized by ascending motor
and sensory neuropathies, encephalopathy, and rapid death. [ref: 133]
Although it has not been carefully evaluated, the major role of the
liver in the disposition of the vinca alkaloids implies that dose
modifications should be considered for patients with hepatic
dysfunction. [ref: 134] However, firm guidelines have not been
established. A 50% dose reduction is often recommended for patients
with total bilirubin levels between 1.5 and 3.0 mg/dL (50% dose
reduction for bilirubin levels between 2.0 and 3.0 mg/dL is recommended
for VRL), and at least a 75% dose reduction for plasma total bilirubin
levels greater than 3.0 mg/dL. Dose reduction for renal dysfunction is
not indicated.

The Taxanes

The unique chemical structure and mechanism of action of the taxanes,
coupled with their antitumor activities against a broad range of
cancers, has rendered the taxanes one of the most important new classes
of anticancer agents. Interest in the taxanes began in 1963, when a
crude extract of the bark of the Pacific yew tree, Taxus brevifolia,
was shown to have broad activity in preclinical tumor models. In 1971,
paclitaxel was identified as the active constituent of the bark
extract. [ref: 135,136] The initial development of paclitaxel was
hampered by the limited supply of its primary source; the difficulties
inherent in large-scale isolation, extraction, and preparation of bulk
compound for a natural product; and its poor aqueous solubility. [ref:
135-137] Interest was maintained during this time by the
characterization of its novel mechanism of cytotoxic action and the
availability of an adequate drug supply for requisite preclinical and
limited clinical evaluations. The early search for taxanes derived from
more abundant and renewable resources led to the development of
docetaxel, which is synthesized by the addition of a side chain to 10-
deacetylbaccatin III, an inactive taxane precursor found in the needles
and other components of more abundant yew species. [ref: 137,138] The
supply of paclitaxel is no longer a limiting issue because the agent is
also produced semisynthetically from 10-deacetylbaccatin III and other
abundant precursors.
The structures of paclitaxel, docetaxel, and their precursor 10-
deacetylbaccatin III are shown in Figure 19.7_3. The taxanes are
complex esters consisting of a 15-member taxane ring system linked to
an unusual 4-member oxetan ring. The taxane rings of both paclitaxel
and docetaxel (but not 10-deacetylbaccatin III) are linked to an ester
side chain attached to the C13 position of the ring, which is essential
for antimicrotubule and antitumor activity. The structures of
paclitaxel and docetaxel differ in substitutions at the C10 taxane ring
position and on the ester side chain attached at C13.
The most impressive clinical activity of paclitaxel has been in
patients with ovarian and breast cancers. [ref: 135-141] Paclitaxel
initially received regulatory approval in the United States and many
other countries for the treatment of patients with ovarian cancer after
failure of first-line or subsequent chemotherapy. It subsequently
received regulatory approval for patients with advanced breast cancer
after failure of combination chemotherapy or at relapse within 6 months
of adjuvant chemotherapy. Its use in combination with a platinum
compound as primary induction therapy in suboptimally debulked stage
III or IV ovarian cancer and as a component of adjuvant chemotherapy
after primary local treatment in high-risk patients with early-stage
breast cancer has demonstrated a survival advantage in randomized phase
III studies. [ref: 139,141]Paclitaxel has also received regulatory
approval in the United States for second-line treatment of Kaposi's
sarcoma associated with AIDS, in combination with cisplatin as primary
treatment of non-small cell lung cancer, and as a component of adjuvant
chemotherapy in high-risk lymph node-positive breast cancer. [ref: 139-
143]
Docetaxel initially received regulatory approval in the United States
for patients with metastatic breast cancer that has progressed on or
relapsing after anthracycline-based chemotherapy, which was later
broadened to a general second-line indication. [ref: 137,138] Its role
as a component of adjuvant and neoadjuvant chemotherapy after local
treatment of early-stage breast cancer and first-line chemotherapy for
locally advanced or metastatic breast cancer is being evaluated.
Furthermore, docetaxel has received regulatory approval in many
countries for treatment of locally advanced or metastatic non-small
cell lung carcinoma and in the United States for treatment of non-small
cell lung cancer after failure of cisplatin-based therapy. The clinical
antitumor spectra for paclitaxel and docetaxel are similar, with
activity noted in many other diverse tumor types that are generally
refractory to conventional therapies, including lymphoma, and small
cell lung, head and neck, esophageal, endometrial, bladder, and germ
cell carcinomas.

Mechanisms of Action

Schiff et al., [ref: 144] Schiff and Horwitz, [ref: 145] and Manfredi
et al. [ref: 146] initially identified the unique mechanism of action
for paclitaxel in 1979. The taxanes bind to tubulin polymers
(microtubules) at binding sites that are distinct from exchangeable
GTP, colchicine, podophyllotoxin, and the vinca alkaloids. Paclitaxel
binds preferentially to the N-terminal 31 amino acids of the beta-
tubulin subunit, although additional sites of interaction on beta-
tubulin and alpha-tubulin may also be involved. [ref: 147,148] The
binding of paclitaxel to polymerized tubulin is reversible, with a
binding constant of approximately 1 umol. [ref: 146,149] Docetaxel,
which most likely shares the same tubulin-binding site as paclitaxel,
appears to have a 1.9-fold higher affinity for the site. [ref: 149]
Tubulin assembly induced by docetaxel also proceeds with a critical
protein concentration that is 2.1-fold lower than that of paclitaxel.
[ref: 149] However, these differences may not translate into greater
therapeutic indices for docetaxel in the clinic, as greater potency may
also portend more severe toxicity at identical drug concentrations in
vivo. Nevertheless, the results of both preclinical and clinical
studies suggest that the taxanes may not be completely cross-resistant.
[ref: 150,151]
The taxanes stabilize the microtubule against depolymerization,
thereby disrupting normal microtubule dynamics. [ref: 2,25,144-146,152-
157] They profoundly alter the tubulin dissociation rate constants at
both ends of the microtubule, suppressing both treadmilling and dynamic
instability. Association rate constants are not appreciably affected.
The ability of the taxanes to induce polymerization is associated with
stoichiometric drug binding to microtubules, which occurs at
submicromolar concentrations that are readily achieved in the clinic.
At substoichiometric concentrations, the taxanes suppress microtubule
dynamics without increasing the amount of polymerized tubulin. [ref:
152] Taxane-treated microtubules are very stable, resisting
depolymerization by cold, calcium ions, dilution, and other mitotic
drugs. This stability inhibits the dynamic reorganization of the
microtubule network, which is essential for normal function during both
mitosis and interphase.
Both stoichiometric and substoichiometric binding of the taxanes
inhibit the proliferation of cells, principally by inducing a sustained
mitotic block at the metaphase-anaphase boundary; however, the taxanes
also affect interphase microtubules in nonproliferating cells. [ref:
25,144,152] Distinct morphologic evidence that the taxanes affect
microtubules during interphase and mitosis include the formation of
microtubule bundles during the nonmitotic cell cycle phases and
multiple mitotic spindle asters during mitosis. [ref: 154] Many taxane-
induced disturbances in cellular processes lead to apoptosis or
programmed cell death (discussed later in the section Taxanes, Drug
Resistance, and in Chapter 7). [ref: 56,57,157-166] On removal of the
drug after treatment, even at substoichiometric concentrations that do
not increase microtubule mass, cells exit from mitosis but do not
continue to proliferate. Instead, the cells undergo apoptosis, and cell
death ensues in 2 to 3 days. Although the precise mechanism by which
microtubule disturbances lead to apoptosis has not been determined, the
taxanes interact with numerous substances, including regulatory
molecules and oncogenes that bind to the mitotic apparatus. Paclitaxel
has been reported to induce transcription factors and enzymes that
govern proliferation, apoptosis, and inflammation and, interestingly,
some of these effects, such as the induction of tumor necrosis factor-
alpha. [ref: 157,167] The taxanes also inhibit angiogenic activity at
concentrations below those that induce cytotoxicity. [ref: 165,168,169]
Both paclitaxel and docetaxel have been shown to enhance the effects
of ionizing radiation in vitro at clinically achievable concentrations
(<50 nmol) and in vivo, which may related to the inhibition of cell-
cycle progression in the G(2) phase, which is the most radiosensitive
phase of the cell cycle. [ref: 170-172]

Mechanisms of Resistance

The best characterized mechanism of resistance to the taxanes is the
MDR phenotype, mediated by the 170-kD P-gp efflux pump, encoded by the
mdr1 gene (discussed previously in the section Vinca Alkaloids,
Mechanisms of Resistance). [ref: 44-49,173,174] The MDR protein has
been shown to be an efficient transporter of the vinca alkaloids but
not of the taxanes. [ref: 45,175,176] The results of early studies
evaluating the role of MDR in the clinic indicate that cross-resistance
to the taxanes and anthracycline is incomplete, which has significant
clinical ramifications in treating breast cancer. [ref: 140 ]Strategies
aimed at reversing drug resistance in the clinic with various types of
P-gp substrates and inhibitors are also being evaluated, but the
interpretation of the results is confounded by the effects of P-gp
modulators on taxane clearance. [ref: 177,178]
Several taxane-resistant mutant cell lines that have structurally
altered alpha- and beta-tubulin proteins and an impaired ability to
polymerize into microtubules also have been identified (discussed
previously in the section Vinca Alkaloids, Mechanisms of Resistance).
[ref: 54,157,179] These mutants lack normal interpolar mitotic spindles
and have an inherently slow rate of microtubule assembly, which is
normalized in the presence of the drug. Mutants with "hypostable"
microtubules exhibit collateral sensitivity to the vinca alkaloids. A
number of cell lines resistant to tubulin-binding agents, including the
taxanes, have been shown to have alterations in tubulin content,
expression of tubulin isotypes, tubulin polymerization dynamics, or
tubulin isotype content. [ref: 180-185] Mutations of tubulin isotype
genes have also been reported in taxane-resistant cell lines, and beta-
tubulin gene mutations have been reported to be a strong determinant of
paclitaxel resistance in a series of patients with non-small lung
cancer. [ref: 186]
The regulation and integrity of genes that regulate apoptosis, such
as p53, bcl-2, and bcl-x, may be determinants of sensitivity to the
taxanes. [ref: 56,57,157,162-164,187,188] MAPs are also likely to be
involved in these mechanisms of resistance to drug-induced apoptosis,
as illustrated by the fact that MAP4, which is negatively regulated by
wild-type p53, has been shown to increase the sensitivity to
paclitaxel. [ref: 189,190] It has been proposed that paclitaxel induces
apoptosis through two different mechanisms--a p53-independent pathway
in cells blocked in prophase and a p53-dependent mechanism in cells
that accumulate in the G(1 )cell-cycle phase--and requires functional
p53. [ref: 157,191] However, there are conflicting experimental data as
to the role of p53 as a determinant of cell sensitivity to paclitaxel
and other antitumor agents. Several lines of experimental evidence
suggest that the induction of p53 in cells treated with paclitaxel
represents a mechanism of drug resistance. [ref: 192,193] The taxanes
have been also shown to modulate the function of genes involved in
apoptotic regulation and in the disruption of microtubule dynamics by
paclitaxel and other antimicrotubule drugs, and docetaxel results in
the phosphorylation of such regulatory proteins as Bcl-x(L) and Bcl-2,
thereby annulling the antiapoptotic functions of these regulators.
[ref: 194,195]
Interestingly, paclitaxel-resistant cell lines, which have mutations
in tubulin and fail to exhibit phosphorylation of Bcl-x(L) after
paclitaxel treatment, have been described. [ref: 56] These cells
demonstrate Bcl-x(L) phosphorylation in the presence of other
antimicrotubule agents, suggesting that apoptosis mediated by
paclitaxel is related to the drug's ability to interact with
microtubules.

Pharmacology

The taxanes are commonly administered by intravenous infusion at doses
ranging from 175 to 225 mg/m**2 over 3 hours (for paclitaxel) or 75 to
100 mg/m**2 over 1 hour (for docetaxel) every 3 weeks. Various other
administration schedules have been evaluated (discussed later in the
section Administration, Dose, and Schedule). The oral bioavailability
of both paclitaxel and docetaxel is poor, owing in part to the
constitutive overexpression of P-gp by enterocytes or first-pass
metabolism in the liver or intestines (or both). However, biologically
relevant plasma concentrations are transiently achieved if the taxanes
are administered orally after treatment with oral cyclosporin or other
modulators of P-gp and cytochrome P-450 mixed-function oxidases. [ref:
196,197] As shown in Table 19.7_2, paclitaxel and docetaxel share the
following pharmacologic characteristics: large volumes of distribution,
rapid and avid binding to all tissues except for the unperturbed
central nervous system, long terminal half-lives and substantial
hepatic metabolism, biliary excretion, and fecal elimination.

Paclitaxel

Pharmacologic studies of paclitaxel on both long and short
administration schedules have been performed (discussed later in the
section Administration, Dose, Schedule). In early studies that
principally evaluated prolonged (6- and 24-hour) schedules, substantial
interpatient variability was noted, and nonlinear, dose-dependent
behavior was not observed. [ref: 136,137,198] In these studies, drug
disposition was characterized as a biphasic process, with values for
alpha and beta half-lives averaging approximately 20 minutes and 6
hours, respectively. However, more recent studies of paclitaxel
administered on shorter schedules, particularly as a 3-hour infusion,
indicate that the pharmacokinetic behavior of paclitaxel is nonlinear.
[ref: 199-203] Nonlinearity occurs with all administration schedules,
but it is more apparent with shorter infusions that result in higher
plasma paclitaxel concentrations that more effectively saturate both
drug elimination and tissue distribution processes. Both saturable
distribution and elimination processes may be, in part, responsible for
paclitaxel's nonlinear behavior, with tissue distribution becoming
effectively saturated at lower drug concentrations (achieved with
paclitaxel doses <175 mg/m**2 over 3 hours) compared to elimination
processes that are effectively saturated at higher concentrations
(achieved with paclitaxel doses >175 mg/m**2 over 3 hours). The use of
shorter infusion schedules also results in higher plasma concentrations
of paclitaxel's polyoxyethylated castor oil vehicle, which may also be
responsible for this nonlinearity. [ref: 202] This nonlinear profile
may have several important clinical implications, particularly
regarding dose modifications, because dose escalation may result in a
disproportionate increase in drug exposure and hence toxicity, whereas
dose reductions may result in a disproportionate decrease in drug
exposure, thereby decreasing antitumor activity.
Paclitaxel's volume of distribution is much larger than the volume of
total body water, indicating extensive drug binding to plasma proteins
or other tissue elements, probably tubulin. Plasma protein binding is
high (>95%) and readily reversible. [ref: 198] Drug binding to
platelets is extensive and saturable, and animal distribution studies
with radiolabeled paclitaxel indicate extensive drug uptake and
retention by virtually all tissues, except for the normal brain and
testes. [ref: 204] In humans, peak plasma concentrations achieved with
3- to 96-hour schedules (>0.05 to 10 umol) and drug concentrations in
third-space fluid collections, such as ascites (>0.1 umol), are capable
of inducing significant biologic effects in vitro, but drug penetration
into the normal central nervous system is negligible. [ref:
198,204,205]
The liver is the principal organ involved with paclitaxel clearance,
which metabolizes and excretes both paclitaxel and metabolites into the
bile. [ref: 198,206-209] Ninety-eight percent of radioactivity is
recovered from feces collected for 6 days after rats are treated with
radiolabeled paclitaxel, and approximately 71% of an administered dose
of paclitaxel is excreted in the feces over 5 days as either parent
compound or metabolites in humans, with 6alpha-hydroxypaclitaxel being
the largest component and accounting for 26% of the dose. Only 5% is
unchanged paclitaxel. Renal clearance of paclitaxel and metabolites is
minimal, accounting for 14% of the administered dose. [ref: 198] In
humans, cytochrome P-450 mixed-function oxidases are responsible for
the bulk of drug disposition, specifically the isoenzymes CYP2C8, and
CYP3A4, which metabolize paclitaxel to hydroxylated 6alpha-
hydroxypaclitaxel (major) and another hydroxylated metabolite, both of
which are inactive.
Pharmacodynamic analyses as part of individual phase I and II trials
demonstrated that several pharmacokinetic indices of drug exposure can
be related to the various toxicities of paclitaxel, the most important
and consistent of which is the relationship between the severity of
neutropenia and the duration of drug exposure above biologically
relevant plasma concentrations ranging from 0.05 to 0.1 umol. [ref:
198-201,210] However, a prospective analysis of pharmacokinetic
determinants of outcome in several hundred patients with advanced non-
small cell lung cancer treated with the combination of cisplatin and
paclitaxel at either 135 or 250 mg/m**2 over 24 hours demonstrated that
the magnitude of the steady-state plasma paclitaxel concentration
correlated poorly with antitumor activity, disease-free survival, and
overall survival. [ref: 211]

Docetaxel

The pharmacokinetics of docetaxel on a 1-hour schedule are linear at
doses of 115 mg/m**2 or less and optimally fit a three-compartment
model. [ref: 137,138,212-216] Terminal half-lives ranging from 11.1 to
18.5 hours have been reported. In one population study, plasma
concentration data were optimally fit by a three-compartment model, and
the following pharmacokinetic parameters were generated: t(1/2g )of
12.4 hours, clearance of 1 L/h/m**2, and steady-state volume of
distribution of 74 L/m**2. [ref: 212-214] The most important
determinants of docetaxel clearance were the body surface area, hepatic
function, and plasma alpha1-acid glycoprotein concentration, whereas
age and albumin level had significant (albeit minor) influences on
clearance. As with paclitaxel, plasma protein binding is high (>80% to
90%), and binding is primarily to alpha1-acid glycoprotein, albumin,
and lipoproteins. [ref: 213] Docetaxel is also distributed to all
tissues except the central nervous system. [ref: 212,217] In both dogs
and mice treated with radiolabeled drug, fecal excretion accounts for
70% to 80% of total radioactivity, whereas urinary excretion accounts
for 10% or less. [ref: 212,217] The hepatic cytochrome P-450 mixed-
function oxidases, particularly isoforms CYP3A4 and CYP3A5, are
primarily involved in biotransformation that, in contrast to
paclitaxel, principally affects the C13 side chain and not the taxane
ring. [ref: 212,213,218-220]
The main pharmacokinetic determinants of toxicity, particularly the
principal toxicity neutropenia, are drug exposure and the time that
plasma concentrations exceed biologically relevant concentrations.
[ref: 213,214] A population pharmacodynamic analysis of determinants of
outcome in phase II trials of docetaxel revealed that the most
important determinants of the time to progression in patients with
metastatic breast cancer are the pretreatment plasma concentration of
alpha1-acid glycoprotein, number of prior chemotherapeutic regimens,
and number of disease sites, whereas both drug exposure and the
pretreatment plasma concentration of alpha1-acid glycoprotein were
strong positive determinants of time to progression in patients with
advanced lung cancer treated with docetaxel. [ref: 214] Conversely, the
pretreatment plasma level of alpha1-acid glycoprotein was negatively--
albeit significantly--related to the probability of experiencing both
severe neutropenia and febrile neutropenia.

Drug Interactions

Both sequence-dependent pharmacokinetic and toxicologic interactions
between paclitaxel and several other chemotherapy agents have been
noted. [ref: 198] The sequence of cisplatin followed by paclitaxel (24-
hour schedule) induces more profound neutropenia than the reverse
sequence, which is explained by a 33% reduction in the clearance of
paclitaxel after cisplatin. [ref: 221] The least toxic sequence--
paclitaxel before cisplatin--was demonstrated to induce more
cytotoxicity in vitro; therefore, this drug sequence was selected for
further clinical development. [ref: 222] However, sequence dependence
does not appear to be a clinically relevant phenomenon on shorter
schedules. Treatment with paclitaxel on either a 3- or 24-hour schedule
followed by carboplatin has been demonstrated to produce equivalent
neutropenia and less thrombocytopenia as compared to carboplatin as a
single agent, which is not explained by pharmacokinetic interactions.
[ref: 223,224] Although sequence dependence has not been noted with
carboplatin and paclitaxel in clinical studies, this phenomenon has
been noted with other paclitaxel-based chemotherapy combinations, the
most important of which involve the anthracyclines. [ref: 225] Both
neutropenia and mucositis are more severe when paclitaxel on a 24-hour
schedule is administered before doxorubicin, compared to the reverse
sequence, which is most likely due to an approximately 32% reduction in
the clearance of doxorubicin and doxorubicinol when it is administered
after paclitaxel. [ref: 225,226] Although neither sequence-dependent
pharmacologic interactions nor toxicologic interactions between
doxorubicin and paclitaxel on a shorter (3-hour) schedule have been
noted, pharmacologic interactions occur with both sequences, and
combined treatment with paclitaxel (3-hour schedule) and doxorubicin as
a bolus infusion has been associated with a higher frequency of
congestive cardiotoxicity than would have been expected from an
equivalent cumulative doxorubicin dose given without paclitaxel
(discussed later in the section Toxicity). [ref: 227] Similar
decrements in the clearance of epirubicin and its metabolites have also
been noted in studies of paclitaxel combined with epirubicin, but a
similar enhancement of cardiotoxicity has not been observed. [ref: 228]
The precise etiology for these interactions is unclear; however,
competition for the hepatic or biliary P-gp transport of the
anthracyclines with paclitaxel or its polyoxyethylated castor oil
vehicle (or both) is a logical explanation. [ref: 226,229] The vehicle
is suspected because similar effects have not been noted with
docetaxel, which is not formulated in polyoxyethylated castor oil.
Hematologic toxicity has been more profound with the sequence of
cyclophosphamide before paclitaxel (24-hour schedule) than the reverse
sequence. [ref: 230] In human tumor xenografts, both paclitaxel and
docetaxel have been demonstrated to induce thymidine phosphorylase
activity, which may increase the metabolic activation of the oral
fluoropyrimidine prodrug capecitabine. [ref: 231]
Drug interactions may also result from the effects of other classes
of drugs on the cytochrome P-450-dependent metabolism of the taxanes.
Various inducers of cytochrome P-450 mixed-function oxidases, such as
the anticonvulsants phenytoin and phenobarbital, accelerate in the
metabolism of both paclitaxel and docetaxel in human microsomal studies
and in both children and adults who are concurrently receiving
treatment with these anticonvulsants, as manifested by rapid drug
clearance and tolerance of high drug doses. [ref: 209,219,232-234]
Conversely, many types of agents that inhibit cytochrome P-450 mixed-
function oxidases, such as orphenadrine, erythromycin, cimetidine,
testosterone, ketoconazole, fluconazole, midazolam, polyoxyethylated
castor oil, and corticosteroids, interfere with the metabolism of
paclitaxel and docetaxel in human microsomes in vitro; however, the
inhibitory concentrations of these agents exceed those achieved in
clinical practice, and the clinical relevance of these findings is not
known. [ref: 207-209,217-220,235] Although there has been concern that
the use of corticosteroids and different H(2)-receptor antagonists with
variable cytochrome P-450 inhibitory activities as components of
premedication regimens may differentially affect drug clearance and
hence toxicity, neither toxicologic nor pharmacologic differences
between the agents were noted in a randomized clinical trial. [ref:
236]

Toxicity

Myelosuppression is the principal toxicity of paclitaxel and docetaxel.
However, despite similar structures, these agents differ modestly in
their toxicity spectra.

Paclitaxel

Neutropenia is the principal toxicity of paclitaxel. The onset is
usually on days 8 to 10, and recovery is generally complete by days 15
to 21. The main clinical determinant for the severity of neutropenia is
the extent of prior myelosuppressive therapy. Neutropenia is
noncumulative, and the duration of severe neutropenia, even in heavily
pretreated patients, is usually brief. The most important pharmacologic
determinant of the severity of neutropenia is the duration that plasma
concentrations are maintained above biologically relevant levels (0.05
to 0.10 umol; discussed earlier in the section Pharmacology), which may
explain why neutropenia is more severe with longer infusion schedules.
[ref: 237] This does not necessarily mean that longer schedules will
portend optimal antitumor activity in the clinic. Instead, most
randomized clinical data do not indicate that there is an optimal
schedule for any particular tumor, although treatment with higher doses
should be considered if shorter schedules are used. [ref: 238] At
paclitaxel doses exceeding 175 mg/m**2 on a 24-hour schedule and 225
mg/m**2 on a 3-hour schedule, nadir neutrophil counts are typically
less than 500 uL for fewer than 5 days in most courses, even in
untreated patients. Even patients who have received extensive prior
therapy can usually tolerate paclitaxel doses of 175 to 200 mg/m**2
over 3 or 24 hours. More frequent administration schedules (e.g.,
weekly treatment) have been associated with less severe neutropenia as
compared to single-dose schedules (discussed later in the section
Administration, Dose, and Schedule). Severe thrombocytopenia and anemia
are unusual, except in heavily pretreated patients.
Although the incidence of major hypersensitivity reactions in early
phase I trials approached 30%, the incidence is 1% to 3% with effective
prophylaxis. [ref: 135,136,237,239,240] Most major reactions, which are
characterized by dyspnea with bronchospasm, urticaria, and hypotension,
occur within the first 10 minutes after the first (and less frequently
after the second) treatment and resolve completely after stopping
treatment and occasionally occur after treatment with antihistamines,
fluids, and vasopressors. Patients who have major reactions have been
rechallenged successfully after receiving high doses of
corticosteroids, but this approach has not always been successful.
[ref: 241,242] Although minor reactions, such as flushing and rashes,
have been noted in as many as 40% of patients, minor hypersensitivity
reactions do not portend the development of major reactions. The
hypersensitivity reactions are most likely caused by a
nonimmunologically mediated release of histamine or histamine-like
substances, owing to the taxane moiety or, more likely, its
polyoxyethylated castor oil vehicle, possibly through complement
activation. [ref: 243] Although the incidence of major hypersensitivity
reactions is reduced with lower administration rates and longer
infusion durations, the rates of major reactions are low on both 3- and
24-hour schedules when patients are premedicated with corticosteroids
and both H(1)- and H(2)-receptor antagonists (discussed later in the
section Administration, Dose, and Schedule). [ref: 237] In an
assessment of the relative safety of two different paclitaxel schedules
(3 vs. 24 hours), the rates of major reactions were low and similar
(2.1% vs. 1.0%) in patients receiving paclitaxel for 3 or 24 hours,
respectively, with premedication. [ref: 237]
Paclitaxel induces a peripheral neuropathy characterized by sensory
symptoms, such as numbness in a symmetric glove-and-stocking
distribution. [ref: 244-246] Neurologic examination reveals sensory
loss and loss of deep tendon reflexes. Neurophysiologic studies support
a primary disruption of neuronal microtubules resulting in axonal
degeneration and demyelination as the primary pathogenic mechanism;
however, manifestations suggestive of microtubule disruption resulting
in a neuronopathy may be noted, particularly at higher doses or when
combined with other neurotoxic agents, such as cisplatin. [ref: 245]
Severe neurotoxicity is uncommon when paclitaxel is given alone at
doses below 200 mg/m**2 on a 3- or 24-hour schedule every 3 weeks or
below 100 mg/m**2 on a continuous weekly schedule, but almost all
patients experience mild or moderate effects. Symptoms may begin as
soon as 24 to 72 hours after treatment with higher doses (250 mg/m**2
or greater) but usually occur only after multiple courses at 135 to 250
mg/m**2 every 3 weeks. Neurotoxicity is generally more pronounced when
paclitaxel is administered on short infusion schedules, indicating that
peak plasma concentration is a principal determinant. The combination
of paclitaxel on a 3-hour schedule and cisplatin is particularly
neurotoxic. Motor and autonomic dysfunction may occur, especially at
high doses and in patients with preexisting neuropathies due to
diabetes mellitus and alcoholism. Transient myalgia, usually noted 24
to 48 hours after therapy, is also common, and a myopathy has been
described in patients receiving high doses with cisplatin. Although
several measures, such as the administration of amifostine, glutamate,
and pyridoxine, appear to reduce the neurotoxic effects of paclitaxel
in experimental models, there is no convincing clinical evidence that
any specific measure is effective at ameliorating existing
manifestations or preventing the development or worsening or
neurotoxicity. [ref: 244,246] Optic nerve disturbances, manifested by
scintillating scotoma, may also occur. [ref: 247,248] Acute
encephalopathy, which can progress to coma and death, has been reported
after treatment with high doses (600 mg/m**2 or greater). [ref: 249]
Paclitaxel causes cardiac rhythm disturbances, but the clinical
relevance of these effects is not known. [ref: 239,250-252] The most
common rhythm disturbance, a transient bradycardia, was noted in 29% of
patients in one trial. [ref: 239,250,251] Isolated asymptomatic
bradycardia without hemodynamic effects does not appear to be an
indication for discontinuing paclitaxel. More important
bradyarrhythmias, including Mobitz type I (Wenckeback syndrome), Mobitz
type II, and third-degree heart block, have been noted, but the
incidence in a large National Cancer Institute database was only 0.1%.
[ref: 251] Most documented episodes have been asymptomatic. These
events primarily occurred in patients enrolled in early trials that
required continuous cardiac monitoring, indicating that second- and
third-degree heart block are likely underreported because such
monitoring is not usually performed. These bradyarrhythmias are
probably caused by paclitaxel, as related taxanes affect cardiac
automaticity and conduction, and similar disturbances have occurred in
humans and animals who have ingested various species of yew plants.
Myocardial infarction, cardiac ischemia, atrial arrhythmias, and
ventricular tachycardia have been noted, but whether there is a causal
relationship between paclitaxel and these events is uncertain.
There is no evidence that chronic, long-term treatment with
paclitaxel causes progressive cardiac dysfunction. Routine cardiac
monitoring during paclitaxel therapy is not necessary but is advisable
for patients who may not be able to tolerate bradyarrhythmias, such as
those with atrioventricular conduction disturbances or ventricular
dysfunction. Although patients with a wide range of cardiac
abnormalities and cardiac histories were broadly and empirically
restricted from participating in early clinical trials, paclitaxel
treatment has been reported to be well tolerated in a small series of
gynecologic cancer patients with major cardiac risk factors. [ref: 252]
On the other hand, repetitive treatment of patients with the combined
regimen of paclitaxel on a 3-hour schedule and doxorubicin as a brief
infusion is associated with a higher frequency of congestive
cardiotoxicity than would be expected to occur with the same cumulative
doxorubicin dose given without paclitaxel (discussed previously in the
section Drug Interactions). [ref: 226,227] In one study of previously
untreated women with advanced breast cancer treated with escalating
doses of paclitaxel as a 3-hour infusion and doxorubicin, 60 mg/m**2 to
a cumulative dose of 480 mg/m**2, which would be predicted to result in
a less than 5% incidence of congestive cardiotoxicity in patients
treated with doxorubicin alone, the incidence of congestive
cardiotoxicity was approximately 25%. [ref: 227 ]However, the incidence
of cardiotoxicity was less than 5% when similar patients received
identical schedules of paclitaxel and doxorubicin, but the cumulative
doxorubicin dose did not exceed 360 mg/m**2. Both experimental and
early clinical results suggest that dexrazoxane may reduce the
cardiotoxicity of the doxorubicin and paclitaxel combination. [ref:
253,254] The incidence of congestive heart failure was also
significantly higher in breast cancer patients treated with the
combination of trastuzumab and paclitaxel than paclitaxel alone in a
phase III trial; therefore, careful monitoring of patients receiving
this combination is warranted. [ref: 255]
Drug-related gastrointestinal effects, such as vomiting and diarrhea,
are uncommon. Higher paclitaxel doses may cause mucositis, especially
in patients with leukemia who may be more prone to mucosal barrier
breakdown or in patients receiving 96-hour infusions. [ref: 256,257]
Rare cases of neutropenic enterocolitis and gastrointestinal necrosis
have been noted, particularly in patients given high doses of
paclitaxel in combination with doxorubicin or cyclophosphamide. [ref:
230,239,258,259 ]Severe hepatotoxicity and pancreatitis have also been
noted, but these events are rare. [ref: 260,261] Acute bilateral
pneumonitis has been reported in fewer than 1% of patients treated on a
3-hour schedule in one series, and both interstitial and parenchymal
pulmonary toxicity have been reported, but clinically significant
pulmonary effects are uncommon. [ref: 262,263]
Paclitaxel also induces reversible alopecia of the scalp, but all
body hair is usually lost with cumulative therapy. Although the agent
is often not considered a vesicant, extravasation of large volumes can
cause moderate soft tissue injury. Inflammation at the injection site
and along the course of an injected vein may occur. Alopecia occurs in
most patients. Nail disorders have been reported, particularly in
patients treated on weekly schedules. [ref: 264] Recall reactions in
previously irradiated sites have also been noted.

Docetaxel

Neutropenia is the principal toxicity of docetaxel. [ref: 137,138,265]
At a dose of 100 mg/m**2, neutrophil counts are below 500/uL in most
patients. Similar to paclitaxel, the onset of neutropenia occurs on
approximately day 8, and complete resolution typically occurs by days
15 to 21. As with paclitaxel, neutropenia is significantly less when
low doses are administered frequently (i.e., on a weekly schedule;
discussed later in the section Administration, Dose, and Schedule).
The most important determinant of neutropenia is the extent of prior
treatment. Significant effects on platelets and red blood cells are
uncommon.
Although docetaxel is not formulated in polyoxyethylated castor oil,
hypersensitivity reactions have been reported in approximately 31% of
patients receiving docetaxel without premedications in early phase II
studies. [ref: 137,138,265] As with paclitaxel, major reactions
characterized by dyspnea, bronchospasm, and hypotension typically occur
during the first two courses and within minutes after the start of
treatment. Signs and symptoms generally resolve within 15 minutes after
cessation of treatment, and docetaxel is usually able to be
reinstituted without sequelae, occasionally after treatment with an
H(1)-receptor antagonist. However, most hypersensitivity reactions are
minor. Both the incidence and severity of hypersensitivity reactions
appear to be reduced by premedication with corticosteroids and H(1)-
and H(2)-receptor antagonists (discussed later in the section
Administration, Dose, and Schedule). Like paclitaxel, patients who
experience major reactions have been retreated successfully after the
resolution of symptoms and after treatment with corticosteroids and
H(1)-receptor antagonists.
Docetaxel induces a unique fluid retention syndrome characterized by
edema, weight gain, and third-space fluid collection. [ref:
137,138,265-267] Fluid retention is cumulative and does not appear to
be due to hypoalbuminemia or cardiac, renal, or hepatic dysfunction.
Instead, several lines of evidence indicate that it is due to increased
capillary permeability. [ref: 266] Capillary filtration studies in
patients who were not receiving corticosteroid premedication have
revealed a two-stage process, with progressive congestion of the
interstitial space by proteins and water starting between the second
and fourth course, followed by insufficient lymphatic drainage. [ref:
266] In early studies in which prophylactic medication was not used,
fluid retention was not usually significant at cumulative docetaxel
doses below 400 mg/m**2; however, the incidence and severity of fluid
retention increased sharply at cumulative doses of 400 mg/m**2 or
greater and often resulted in the delay or termination of treatment.
Prophylactic treatment with corticosteroids with or without H(1)- and
H(2)-receptor antagonists have been demonstrated to reduce the overall
incidence of fluid retention and increase the number of courses and
cumulative docetaxel dose before the onset of this toxicity (discussed
later in the section Administration, Dose, and Schedule). [ref: 267]
Fluid retention typically resolves slowly after docetaxel is stopped,
with complete resolution occurring several months after treatment in
patients with severe toxicity. Aggressive and early treatment with
progressively more potent diuretics starting with potassium-sparing
diuretics has been successfully used to manage fluid retention. The
incidence of fluid retention appears to be lower in studies using lower
doses (60 to 75 mg/m**2) of docetaxel during each course, but this may
be due to the administration of lower overall cumulative doses, and the
effects of lower doses on antitumor activity are unknown.
Skin toxicity may occur in as many as 50% to 75% of patients [ref:
137,138,265,268]; however, premedication may reduce the overall
incidence of this effect. An erythematous pruritic maculopapular rash
that affects the forearms, hands, or feet is typical. Other cutaneous
effects include desquamation of the hands and feet, palmar-plantar
erythrodysesthesia that may respond to pyridoxine or cooling, [ref:
269,270] and onychodystrophy characterized by brown discoloration,
ridging, onycholysis, soreness, and brittleness and loss of the nail
plate.
Both neurosensory and neuromuscular effects are generally less
frequent and less severe with docetaxel as compared to paclitaxel. Mild
to moderate peripheral neurotoxicity occurs in approximately 40% of
previously untreated patients, [ref: 137,138,265,271,272] and patients
who were previously treated with cisplatin appear to be particularly
susceptible, with the incidence approaching 74% in one trial. [ref:
273] The neurotoxicity is qualitatively similar to that of paclitaxel.
Patients typically complain of paresthesia and numbness, but peripheral
motor effects may also occur. Severe toxicity has been unusual after
repetitive treatment with docetaxel doses less than 100 mg/m**2, except
in patients with antecedent disorders, such as alcohol abuse. Transient
arthralgia and myalgia are occasionally noted within days after
treatment. Malaise or asthenia have been prominent complaints in
patients who have been treated with large cumulative doses,
particularly when docetaxel is administered on a continuous weekly
schedule. [ref: 137,138,265,274] Stomatitis appears to occur more
frequently with docetaxel than paclitaxel, particularly with prolonged
infusions, which are utilized rarely. Mild to moderate conjunctivitis,
which is responsive to topical corticosteroids, may also occur,
particularly with weekly administration. Nausea, vomiting, and diarrhea
have also been observed, but severe gastrointestinal toxicity is rare.

Administration, Dose, and Schedule

Paclitaxel

Many investigations have focused on optimal dosing and scheduling since
the regulatory approval of paclitaxel. [ref: 238] Early clinical
studies were limited to the 24-hour schedule, largely owing to an
apparent increased rate of severe hypersensitivity reactions on shorter
schedules, but the development of effective premedication regimens has
facilitated evaluations of a broad range of dosing schedules. Although
paclitaxel, 135 mg/m**2 on a 24-hour schedule, was initially approved
for patients with refractory and recurrent ovarian cancer, regulatory
approval was subsequently obtained for paclitaxel, 175 mg/m**2 on a 3-
hour schedule. In patients with advanced breast and ovarian cancers,
the cumulative body of randomized study results indicate that both
schedules are equivalent, particularly with regard to event-free
survival and overall survival, although response rates have
occasionally been superior with the 24-hour infusion. [ref: 238,275]
Intriguing results were initially obtained with more protracted
schedules, such as a 96-hour infusion schedule in patients with
advanced breast cancer. [ref: 140,238,276] The development of such
schedules was based on the observation that duration of exposure above
a biologically relevant threshold is one of the most important
determinants of cytotoxicity in vitro (discussed earlier in the section
Pharmacology), but there has been no clear evidence that protracted
infusion schedules are superior to shorter schedules with regard to
clinical efficacy or toxicity. [ref: 238,276-278] The extensive and
rapid distribution of the taxanes to peripheral tissues and the avid
and protracted tissue binding of these agents may explain the lack of
substantial differences in antitumor activity between short and more
protracted administration schedules despite substantial differences in
vitro. There has also been considerable interest in intermittent
schedules, particularly those in which paclitaxel is administered as a
1-hour infusion weekly, which results in substantially less
myelosuppression than conventional 3- and 24-hour every 3-week
schedules. [ref: 279,280] However, the reports that antitumor activity
on weekly schedules is superior to that noted with less frequent
schedules are largely anecdotal, and randomized trials are in progress.
Nevertheless, the weekly schedule may be advantageous for patients who
are at high risk of developing severe myelosuppression.
Paclitaxel is generally administered every 3 weeks at a dose of 175
mg/m**2 over 3 hours or 135 to 175 mg/m**2 over 24 hours. Several phase
III studies in patients with advanced lung, head and neck, and ovarian
cancers have consistently failed to show that paclitaxel doses greater
than 135 to 175 mg/m**2 on a 24-hour schedule are superior to
conventional doses. [ref: 143,238,281] Nearly identical results have
been obtained in a phase III study in patients with metastatic breast
cancer, in which efficacy was not increased in patients treated with
paclitaxel doses greater than 175 mg/m**2 on a 3-hour schedule. [ref:
155,277,326] The following doses have been recommended on less
conventional schedules: 200 mg/m**2 over 1 hour as either a single dose
or 3 divided doses every 3 weeks; 140 mg/m**2 over 96 hours every 3
weeks; and 80 to 100 mg/m**2 weekly. The most common schedules
evaluated in patients with AIDS-associated Kaposi's sarcoma are
paclitaxel, 135 mg/m**2 over 3 or 24 hours every 3 weeks, and 100
mg/m**2 every 2 weeks. [ref: 142] Paclitaxel has also been administered
into the pleural and peritoneal cavities. [ref: 282,283] Biologically
relevant plasma concentrations have been achieved with intraperitoneal
administration, and concentrations in the peritoneal cavity are several
orders of magnitude greater than plasma concentrations. [ref: 282]
The following premedication is recommended to prevent major
hypersensitivity reactions: dexamethasone, 20 mg orally or
intravenously, 12 and 6 hours before treatment; an H(1)-receptor
antagonist (such as diphenhydramine, 50 mg intravenously) 30 minutes
before treatment; and an H(2)-receptor antagonist (such as cimetidine,
300 mg; famotidine, 20 mg; or ranitidine, 150 mg intravenously) 30
minutes before treatment. A single dose of a corticosteroid
(dexamethasone, 20 mg intravenously) administered 30 minutes before
treatment has been reported to confer very effective prophylaxis of
major hypersensitivity reactions. [ref: 284,285] Contact of paclitaxel
with plasticized polyvinyl chloride equipment or devices must be
avoided because of the risk of patient exposures to plasticizers that
may be leached from polyvinyl chloride infusion bags or sets.
Paclitaxel solutions should be diluted and stored in glass or
polypropylene bottles or suitable plastic bags (polypropylene or
polyolefin) and administered through polyethylene-lined administration
sets that include an in-line filter with a microporous membrane not
greater than 0.22 um.
The extensive involvement of hepatic metabolism and biliary excretion
in the disposition of paclitaxel--similar to that of other anticancer
drugs, such as the vinca alkaloids--in which dose modifications are
required indicates that doses should be modified in patients with
hepatic dysfunction. Official recommendations have not been formulated,
but prospective evaluations indicate that patients with moderate to
severe elevations in serum concentrations of hepatocellular enzymes or
bilirubin (or both) are more likely to develop severe toxicity than
patients without hepatic dysfunction. [ref: 286,287] Therefore, it
would be prudent to reduce paclitaxel doses by at least 50% in patients
with moderate or severe hepatic excretory dysfunction
(hyperbilirubinemia) or significant elevations in hepatic
transaminases. Renal clearance contributes minimally to overall
clearance (5% to 10%), and patients with severe renal dysfunction do
not appear to require dose modification. [ref: 288] Based on the
pharmacologic behavior, particularly the wide distributive properties
of the taxanes, dose modifications are not required solely for
peripheral edema and third-space fluid collections.

Docetaxel

In the United States, docetaxel is indicated at a dose range of 60 to
100 mg/m**2 and 75 mg/m**2 over 1 hour in patients with breast and non-
small cell lung cancers, respectively, but most early clinical trials
in advanced breast, ovarian, and non-small cell lung cancers evaluated
doses in the higher end of this range (75 to 100 mg/m**2), with scant
data available for patients treated at 60 mg/m**2. [ref: 137,138,265
]Although some untreated or minimally pretreated patients generally
tolerate docetaxel at a dose of 100 mg/m**2 without severe toxicity,
emerging data indicate poorer tolerance in more heavily pretreated
patients in whom 75 mg/m**2 appears to be more reasonable from a
toxicologic perspective. [ref: 289] Like paclitaxel, docetaxel has also
been administered as a 1-hour infusion weekly. Although there are no
clear benefits of chronic weekly drug administration in terms of
antitumor activity, hematologic toxicity is much less than with
conventional dose schedules, with a high incidence of cumulative
asthenia and neurotoxicity noted with docetaxel doses exceeding 36
mg/m**2/wk. [ref: 274] Despite the use of a polysorbate 80 formulation
instead of polyoxyethylated castor oil, which is used to formulate
paclitaxel, a relatively high rate of hypersensitivity reactions and
profound fluid retention in patients who did not receive premedication
has led to the use of several effective premedication regimens, the
most popular of which is dexamethasone, 8 mg orally twice daily for 3
or 5 days starting 1 or 2 days, respectively, before docetaxel, with or
without both H(1)- and H(2)-receptor antagonists given 30 minutes
before docetaxel. [ref: 267]
A retrospective review of docetaxel pharmacokinetics in patients
without hyperbilirubinemia demonstrated that docetaxel clearance is
reduced by approximately 25% in patients with elevations in serum
concentrations of both hepatic transaminases (1.5-fold or greater) and
alkaline phosphatase (2.5-fold or greater), regardless of whether the
elevations are due to hepatic metastases. [ref: 212-214] Therefore,
dose reductions by at least 25% are recommended for such individuals.
More substantial reductions (50% or greater) may be required in
patients who have moderate or severe hepatic excretory dysfunction
(hyperbilirubinemia). [ref: 287] As with paclitaxel (discussed
previously in the section Administration, Dose, and Schedule,
Paclitaxel), there is no rationale for dose modification solely for
renal deficiency or third-space fluid accumulation (or both). Also
similar to the case with paclitaxel, glass bottles or polypropylene or
polyolefin plastic products should be used for preparation and storage,
and docetaxel should be administered through polyethylene-lined
administration sets.

Estramustine Phosphate

Estramustine phosphate (Fig. 19.7_4) is a conjugate of the alkylating
agent nornitrogen mustard linked to 17beta-estradiol by a carbamate
ester. This agent was originally designed so that estramustine would
accumulate specifically in estrogen receptor-bearing breast cancer
cells via the 17beta-estradiol component followed by degradation of the
carbamate ester and release of the alkylating nor-nitrogen mustard
moiety. Estramustine phosphate, however, did not demonstrate useful
anticancer activity in clinical trials in breast cancer and,
thereafter, it was determined that alkylation of DNA did not occur.
[ref: 290] Further investigations later established that preferential
accumulation of radiolabeled estramustine phosphate in the ventral
prostate of rats occurred unrelated to the estrogen receptor. [ref:
291] This selective accumulation was mediated by the presence of a
specific protein in prostate tissue, subsequently labeled estramustine-
binding protein (EMBP). [ref: 291,292] Clinical studies of estramustine
phosphate were initiated in advanced prostate cancer based on this
unique pattern of drug distribution. [ref: 293,294] Anticancer activity
was subsequently demonstrated in prostate cancer patients with disease
refractory to diethylstilbestrol.

Mechanisms of Action

Several mechanisms of cytotoxic activity have been attributed to
estramustine phosphate. The preponderance of data indicates that cell
death is principally mediated through a direct effect on microtubules.
Estramustine is known to inhibit mitotic microtubule networks and to
depolymerize interphase microtubules. [ref: 296,297] Consonant with
other antimicrotubule agents, estramustine-treated cells arrest in the
G(2)/M phase of the cell cycle and then undergo apoptosis. Estramustine
inhibits microtubule function through direct binding to beta-tubulin
independent of MAPs while also inhibiting microtubule function through
an interaction with MAPs. [ref: 298-302] Once bound to tubulin,
estramustine inhibits the dynamic growth and shortening of
microtubules. Like the taxanes, estramustine can also exert an
antiproliferative effect via stabilization of spindle microtubules.
[ref: 300] The binding of estramustine to beta-tubulin, however, occurs
at a unique site distinct from those of the taxanes, colchicine, and
vinca alkaloids. [ref: 303] Finally, the antimicrotubule effects of
estramustine are mediated by the intact conjugate and not the
individual nor-nitrogen or estradiol components. [ref: 304]
The specific binding of estramustine and its metabolite,
estromustine, to EMBP permits tissue selectivity for estramustine
accumulation and action. [ref: 304,305] After exposure to estramustine,
cell lines that contain high levels of EMBP exhibit a greater fraction
of cells arresting in the G(2)/M phase as compared to those with low
levels of EMBP expression. [ref: 304-306] Proteins similar to EMBP have
also been found in other tumors, including gliomas and astrocytomas.
[ref: 307-309] Because estramustine phosphate induces a G(2)/M block,
crosses the blood-brain barrier, and accumulates in gliomas and
astrocytomas, the potential for estramustine selectively to sensitize
central nervous system tumor cells to irradiation is an area of active
investigation. [ref: 310,311]
Other proposed mechanisms of action attributed to estramustine
include interaction and disruption of the nuclear matrix, alterations
of the actin microfilaments of the cytoskeleton, and alterations of ion
flux across the plasma membrane. [ref: 312-314]

Mechanisms of Resistance

Investigations with cell lines made resistant to estramustine have
characterized several mechanisms of acquired drug resistance.
Consistent with its antimicrotubule mechanism of action, resistance to
estramustine can be mediated by alterations at the site of
estramustine-tubulin interaction, increased microtubule stability
through overexpression of specific tubulin isotypes, or alterations in
MAPs. A drug efflux mechanism, distinct from classical MDR has been
described.
The targets of estramustine--beta-tubulins--are composed of multiple
isotypes encoded by separate cellular genes. An increase in beta(III)-
and beta(IVa)-tubulin isotypes relative to other beta-tubulin isotypes
occurs in human prostate cancer cells rendered eight- to ninefold
resistant to estramustine. [ref: 301] Although the precise site of
estramustine binding is not known, microtubules containing beta(III)-
tubulin isotypes appear to bind estramustine less efficiently as
compared to either other beta-tubulins or alpha-tubulin. [ref: 301]
Furthermore, tubulin isotypes differ from one another principally at
MAP binding sites. Because the binding of different beta-tubulin
subtypes to alpha-tubulin alters the dynamic properties of microtubule
growth and stability, a change in the relative beta-tubulin isotypes
may counter the inhibitory and destabilizing effects of estramustine on
microtubule assembly. [ref: 315,316]
Some prostate cancer cell lines with acquired resistance to
estramustine overexpress the MAP tau. The capacity to maintain
microtubule stability and kinetics involves the interaction of tubulin
with MAPs. Exposure to estramustine induces both quantitative and
qualitative changes in tau, leading to a sevenfold increase in
estramustine resistance in some cell lines. [ref: 317] To what extent
alterations in tau or other altered MAPs contribute to clinical
estramustine resistance is not known.
Although estramustine can bind to the classical MDR efflux pump, P-
gp-overexpressing cells are not cross-resistant to estramustine. [ref:
318-320] Estramustine may, in fact, act as a competitive inhibitor of
P-gp action, reducing the efflux of other cytotoxic agents subject to
P-gp-mediated resistance. [ref: 318,320] A drug efflux mechanism
distinct from P-gp has been described that is distinct from P-gp and
can mediate estramustine resistance. [ref: 321] Some cell lines with
acquired estramustine resistance exhibit a sixfold resistance to
estramustine commensurate with the degree of overexpression of the gene
encoding this new efflux pump.

Pharmacology

After oral administration, estramustine phosphate undergoes rapid
dephosphorylation within the gastrointestinal tract, as shown in
Figure 19.7_4. The bioavailability of oral estramustine phosphate is
37% to 75%. [ref: 322,323] The majority of absorbed estramustine is
rapidly metabolized to an oxidized isomer, estromustine, which is the
principal component detected in the plasma. [ref: 324] Maximal
estromustine plasma concentrations are reached within 2 to 4 hours
after oral consumption, and the mean elimination half-life is 14 hours.
[ref: 322] Estromustine pharmacokinetics are linear over the usual
administered oral doses of estramustine phosphate. Peak plasma
concentrations in patients treated chronically with oral estramustine
phosphate at 560 mg/d have been 227 ng/mL for estromustine, 23 ng/mL
for estramustine, 95 ng/mL for estrone, and 9.3 ng/mL for estradiol.
[ref: 324]
Further hydrolysis of the estromustine and estramustine carbamate
linker in the liver yields estrone and estradiol, respectively, and the
nor-nitrogen group. Studies of oral and intravenously administered
radiolabeled estramustine phosphate indicate that estromustine and
estramustine and their metabolites are largely excreted in the feces,
with only small amounts of conjugated estrone and estradiol found in
the urine (<1%). [ref: 322-325]
In contrast to oral administration, intravenous estramustine
phosphate delivers significantly higher plasma concentrations of
estramustine phosphate and metabolites while reducing the marked
interpatient variability noted for the oral route. [ref: 322-325]
Intravenous estramustine phosphate is currently investigational in the
United States.

Drug Interactions

Coadministration of food or dairy products significantly impairs the
absorption of estramustine phosphate. [ref: 326] Calcium-rich foods
appear to lead to the formation of a poorly absorbable calcium complex.
Current recommendations include fasting before the oral administration
of estramustine phosphate and avoidance of calcium-rich foods and
calcium antacids. [ref: 326]
Preliminary evidence suggests that oral estramustine phosphate, when
coadministered with intravenous docetaxel, significantly delays the
clearance of docetaxel, with disproportionate increases in docetaxel
concentrations. [ref: 327] This has led to a reduction in the
recommended dose for docetaxel when combined with estramustine
phosphate despite the fact that, for the most part, these two agents
have nonoverlapping toxicities. The mechanism by which estramustine
impairs docetaxel clearance is not known.

Toxicity

Nausea and vomiting, which are the principal toxicities encountered
with oral estramustine phosphate, may infrequently necessitate
discontinuation. At conventional dosing schedules, nausea and vomiting
can be prevented by antiemetic therapy. Diarrhea has also been observed
in some patients with chronic use. Myelosuppression is not associated
with estramustine phosphate when administered as a single agent.
Commonly observed estrogenic side effects of estramustine therapy
include gynecomastia, nipple tenderness, and fluid retention. Caution
should be exercised in prescribing estramustine phosphate to patients
with congestive heart failure because of the risk for fluid retention
and edema. Thromboembolic complications represent the most hazardous
toxicity of estramustine phosphate therapy and may occur in as many as
10% of patients. These include venous thrombosis, pulmonary emboli, and
cerebrovascular and coronary thrombotic events. Transient elevations in
hepatic transaminases occur in approximately one-third of patients
receiving estramustine phosphate therapy. The rate of hepatic toxicity
is similar to that described for diethylstilbestrol in a randomized
study of estramustine phosphate versus diethylstilbestrol. [ref: 328]

Administration, Dose, and Schedule

Estramustine phosphate is approved for the treatment of metastatic
prostate cancer, particularly hormone-refractory disease. The
recommended daily dose of estramustine phosphate (available as a 140-mg
capsule) is 14 mg/kg of body weight given in three to four divided
doses, though most patients are usually treated in the dosing range of
10 to 16 mg/kg. Patients should be instructed to take estramustine
phosphate with water at least 1 hour before or 2 hours after meals.
Patients are generally treated for 30 to 90 days before assessment of
therapeutic benefit. Chronic oral therapy can be maintained for months
or even years as long as the favorable response continues. Abbreviated
5-day courses of oral estramustine phosphate have been proposed for use
with such chemotherapy agents as docetaxel. This schedule allows for
the concurrent administration of estramustine phosphate with
intravenous chemotherapeutic agents while reducing some of the toxicity
of chronic oral administration.

Novel Compounds Targeting Microtubules

Many other structurally--and functionally--unique antimicrotubule
compounds are the focus of discovery efforts, preclinical development,
and clinical evaluations. Although the majority of efforts are being
directed toward agents that interfere with tubulin, other potential
strategic components of the microtubule, including motor proteins, are
the focus of discovery and developmental efforts. [ref: 13]
The successes with the taxanes have provided the impetus to discover
new chemotypes that work by a similar mechanism but yet have higher
therapeutic indices. Several natural products that are structurally
dissimilar to the taxanes, share their mechanism of action, and show
comparable activities have been identified. For example, rhazinilam,
like paclitaxel, originates from tree bark but is the first nontaxane
identified that induces cold-stable tubulin polymerization in vitro and
microtubule bundling in cells. [ref: 329 ]Unlike paclitaxel, rhazinilam
is capable of inducing tubulin polymerization at 0**oC; however, the
resulting polymerized product is unstable. In contrast, discodermolide,
which originates from a marine sponge, polymerizes tubulin at 37**oC in
vitro more potently and rapidly than does paclitaxel, yielding
polymerization products that are cold-stable, and it polymerizes
tubulin almost as rapidly at 0**oC. [ref: 330] Unlike rhazinilam,
discodermolide-induced tubulin polymers are completely stable to
treatment with calcium ions and are composed of very short microtubules
instead of tubulin spirals. The epothilones A and B, which are derived
by microbial fermentation, appear to be more like the taxanes in their
polymerization products. [ref: 331,332] The microtubules they induce
are relatively long, rigid, and resistant to destabilization by cold
temperature and calcium ions. These epothilones and their analogues are
at least as potent as paclitaxel and cause mitotic arrest and
microtubule bundling. Epothilone B analogues are currently undergoing
clinical evaluation. The marine soft coral-derived natural products--
sarcotidicytins A and B and eleutherobin also promote tubulin
polymerization in a manner analogous to that of paclitaxel. [ref: 333]
All the aforementioned compounds are likely substrates for P-gp to
some extent, expressing varying degrees of cross-resistance against P-
gp-expressing cells. However, other marine-derived, microtubule-
stabilizing cytotoxins, such as laulimalide and isolaulimalide, appear
to be poor substrates for the P-gp drug efflux pump. [ref: 334] Because
eleutherobin, epothilones A and B, and discodermolide competitively
inhibit [**3H]paclitaxel binding to microtubules, a common
pharmacophore has been sought and identified and may enable the
development of hybrid constructs with more desirable biologic
characteristics. [ref: 335]
Other natural products and semisynthetic antimicrotubule compounds
under evaluation interact with tubulin in the vinca alkaloid- or
colchicine-binding domains. Among the most potent are the
cryptophycins, which are a family of cyanobacterial macrolides that
deplete microtubules in intact cells, including cells with the MDR
phenotype. [ref: 336,337] The cryptophycins compete for the binding of
[**3H]VBL, but neither for radiolabeled paclitaxel nor for colchicine,
and inhibit GTP hydrolysis by isolated tubulin. They also have
excellent activity against several types of tumor xenografts, including
tumors resistant to the vinca alkaloids. One semisynthetic analogue,
cryptophycin-52, is currently undergoing initial clinical evaluation.
[ref: 336] The dolastatins constitute a series of oligopeptides
isolated from the sea hare, Dolabela auricularia. [ref: 337-340] Two of
the most potent dolastatins, dolastatin-10 and -15, noncompetitively
inhibit the binding of vinca alkaloids to tubulin, inhibit tubulin
polymerization and tubulin-dependent GTP hydrolysis, stabilize the
colchicine-binding activity of tubulin, and possess cytotoxic activity
in the picomolar to low nanomolar range. Dolastatin-10 and
semisynthetic dolastatin analogues are undergoing preclinical
development and clinical evaluation. [ref: 340] Phomopsin A,
halichondrin B, homohalichondrin B, and spongistatin 1, which interact
with tubulin in the vinca alkaloid-binding domain and with the natural
products combretastatin and steganacin and the synthetic compounds
pyridine and pyridazine, and the pentafluorophenylsulfonamides, which
interact with tubulin at the colchicine-binding domain, are currently
being evaluated in preclinical or early clinical evaluations. [ref:
341,342]