ICANCER RESEARCH 50, 1220-1225, February 15. 1990]
`
`Phase I and Pharmacokinetic
`NSC 281272)
`
`Study of Arabinofuranosyl-5-azacytosine
`
`(Fazarabine,
`
`Antonella Surbone, Harry Ford, Jr., James A. Kelley, Noa Ben-Baruch, Rose V. Thomas, Robert Fine, and
`Kenneth H. Cowan1
`
`Medicine Branch ¡A.S., N. B-B., R. V. T., R. F., K. H. C.J and Laboratory of Medicinal Chemistry [H. F., J. A. K.J, National Cancer Institute, Bethesda, Maryland 20892
`
`ABSTRACT
`
`(ara-
`A Phase I clinical trial of l-i8-D-arabinofuranosyl-5-azacytosine
`AC or fazarabine) given as a 72-h continuous infusion on a 21-day cycle
`was conducted in 27 adult patients with refractory cancer. The major
`toxicity was reversible granulocytopenia and thrombocytopenia. Dose-
`limiting toxicity was observed at a dose rate of 1.96 mg/mz/h in which
`Grade IV leukopenia (WBC < 1.0(1(1,mm')occurred in 4 of 11 patients
`and Grade IV thrombocytopenia (platelets <25,000/mm') occurred in 3
`of 11 patients. Plasma steady-state levels ranged from 0.13 to 0.6 /JMfor
`doses of 1.25 to 5.94 mg/m2/h. Mean total body clearance was 647 ml/
`min nr. Minor clinical responses were seen in one patient with testicular
`cancer, one patient with colon cancer, one patient with breast cancer, and
`one patient with acute nonlymphocytic leukemia. Another patient with
`adenocarcinoma of unknown primary had stable disease during 13 cycles
`of therapy. Based on the results of this study, the recommended dose for
`Phase II studies of l-/9-r>-arabinofuranosyl-5-azacytosine administered
`as a 72-h continuous infusion is 2.0 mg/m2/h (48 mg/m2/day).
`
`to 5-AC (which undergoes intracellular phosphoryla
`resistant
`tion by this enzyme) but sensitive to both ara-C and ara-AC
`(10).
`Although the mechanism of ara-AC cytotoxicity is not well
`understood,
`this analogue is incorporated into cellular DNA in
`a dose- and time-dependent manner
`(5, 6, 11). Furthermore,
`ara-AC inhibition of DNA synthesis
`in LI210 cells occurs at
`drug doses that do not affect RNA or protein synthesis. Thus,
`the mechanism of ara-AC cytotoxicity,
`like that of ara-C,
`is
`apparently related to its ability to inhibit DNA synthesis.
`One of the most interesting features of ara-AC is its relatively
`broad spectrum of activity in both in vivo and in vitro preclinical
`screens (8, 12-15).
`In contrast
`to both parent compounds,
`ara-
`C and 5-AC, which show little if any activity against
`solid
`tumors, ara-AC demonstrated
`significant activity in a National
`Cancer Institute panel of human tumor xenografts.
`In addition
`to the above-mentioned
`activity against human and mouse
`leukemia cell
`lines, ara-AC displayed activity against human
`colon,
`lung, breast, and ovarian cancer xenografts.
`studies have
`Preclinical pharmacokinetic
`and toxicological
`been performed on both mice and dogs (12, 15) and limiting
`toxicities appeared to be myelosuppression
`and gastrointestinal
`distress, with the occurrence of bloody diarrhea
`and emesis
`(15). Other
`toxicities
`that were not dose limiting included
`lethargy,
`tremors,
`and ataxia. Histological
`evidence of drug-
`related toxicity included bone marrow depletion,
`lymphoid
`organ depletion, necrosis of gastrointestinal
`epithelium,
`and
`maturation arrest with necrosis of the seminiferous
`tubules of
`the testis (15).
`activity, we
`its broad spectrum of preclinical
`Because of
`initiated a Phase I clinical study of this agent. Preclinical studies
`indicated that
`the activity of this agent
`is schedule dependent
`and that continuous
`infusion and frequent
`intermittent
`admin
`istration are apparently more effective than single-dose sched
`ules (12-15). Thus,
`in this Phase I trial ara-AC was adminis
`tered as a 72-h infusion to patients with clinically refractory
`cancer.
`
`MATERIALS AND METHODS
`
`INTRODUCTION
`ara-AC2 (fazarabine; NSC 281272)
`is a synthetic pyrimidine
`features of two effective
`nucleoside containing
`the structural
`antineoplastic
`agents, ara-C and 5-AC (Fig. 1). ara-AC com
`bines the arabinose sugar of ara-C (with its inverted hydroxyl
`group at
`the 2'-position) with the triazine base of 5-AC,
`in
`which nitrogen
`is substituted
`for
`the C-5 of cytidine. This
`triazine ring is highly susceptible to reduction and nucleophilic
`reactions. Thus, ara-AC,
`like 5-AC, undergoes decomposition
`in aqueous solution to inactive products
`through the addition
`of water to its 5-6 double bond with concomitant
`ring opening.
`In contrast, when ara-AC is dissolved in organic solvents such
`as DMSO,
`its stability is markedly enhanced and there is little
`hydrolysis of the triazine ring (3, 4).
`ara-AC,
`like ara-C, apparently enters cells via a nucleoside
`transport mechanism and is subsequently
`activated through
`phosphorylation
`by deoxycytidine
`kinase (5, 6). Degradative
`pathways
`for both ara-AC and ara-C involve phosphatases
`which can cleave phosphate groups,
`from the arabinofuranosyl
`moiety. While ara-C can be converted to an inactive metabolite,
`l-/3-D-arabinofuranosyluridine,
`by cytidine deaminase (7), ara-
`AC is a poor substitute for this enzyme and is relatively refrac
`tory to inactivation by deamination (5).
`The similarity in activation via deoxycytidine kinase suggests
`that cross-resistance between the two nucleoside analogues, ara-
`C and ara-AC, can occur (8, 9). Indeed, P388 murine leukemia
`cells that are resistant
`to ara-C because of a decrease in deoxy
`cytidine kinase activity are also cross-resistant
`to ara-AC (9).
`In contrast, HL-60 cells deficient
`in uridine/cytidine
`kinase are
`
`revised 10/20/89; accepted 11/9/89.
`Received 4/17/89;
`The costs of publication of this article were defrayed in pan by the payment
`of page charges. This article must
`therefore be hereby marked advertisement
`in
`accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
`' To whom requests for reprints
`should be addressed, at Medicine Branch.
`National Cancer Institute. NIH, Bldg. 10, Rm. 12N226. Bethesda. MD 20892.
`2The abbreviations used are: ara-AC,
`l-#-D-arabinofuranosyl-5-azacytosine;
`ara-C,
`I-0-D-arabinofuranosylcytosine;
`5-AC, 5-azacytidine; HPLC, high-per
`formance liquid chromatography; C„plasma steady-state concentration; DMSO,
`dimethyl sulfoxide.
`
`I.
`are shown in Table
`characteristics
`Patient Selection. Patient
`ranging in age from
`Twenty-seven (26 évaluablefor toxicity) patients,
`25 to 72 years were entered into the study. All patients had a histolog-
`ically proved diagnosis of malignant disease with definitive evidence of
`metastatic spread and/or of inoperable local recurrence. One patient
`with treatment-induced
`acute nonlymphoblastic
`leukemia was given
`ara-AC under a special exemption;
`this patient was included in the
`pharmacokinetic
`evaluation but not in the toxicity evaluation since he
`was neutropenic and thrombocytopenic before entering this study. The
`patients in this study were, in general, heavily pretreated. All but four
`patients had previously received chemotherapy.
`In addition to chemo
`therapy,
`13 patients
`(48%) were treated previously with radiation
`therapy.
`Prior to beginning treatment with ara-AC, each patient underwent a
`comprehensive evaluation including complete history, physical exami
`nation, and evaluation of measurable disease by appropriate
`radio
`CELGENE 2121
`1220
`APOTEX v. CELGENE
`IPR2023-00512
`
`© 1990 American Association for Cancer Research.
`
`

`

`PHASE I ARA-AC (FAZARABINE)
`
`resulting in a final ara-AC concentra
`in water (Tera Pharmaceuticals)
`tion of 25 mg/ml. A 24-h supply of drug was diluted to a final volume
`of 12 ml with 70% DMSO and placed in a 12-ml Luer-lock syringe
`(Monoject, Model 512936). The syringe was replaced every 24 h with
`one containing freshly reconstituted drug solution. The drug was deliv
`ered by an Autosyringe infusion pump (Travenol Laboratories; Model
`AS2F) as a "piggyback" into an i.v. solution of 5% dextrose in water
`and infused at 20 ml/h. A polyolefin-lined extension set (Pacesetter
`Infusion, Ltd.; Model 126) was used to connect
`the syringe to the i.v.
`set. This diluted solution of ara-AC was administered through a pe
`ripheral vein or a central
`line.
`Treatment Plan. The starting dose was 0.2 mg/m2/h administered as
`a continuous infusion over 72 h. This starting dose was determined on
`the basis of the tolerated dose in mice during 72-h continuous infusion
`(10 mg/m2/h) with an additional consideration given to the potential
`increased metabolic activation of this drug in humans. Some studies
`suggest
`that human cells may contain higher
`levels of activity of
`deoxycytidine kinase,
`the enzyme that converts ara-AC to its toxic
`metabolites (15). Indeed, previous clinical experience had shown that
`fludarabine, an agent
`that
`is also converted to toxic species by deoxy
`cytidine kinase, was considerably more toxic in humans that
`it was in
`mice and other animals (15). Based on this,
`the starting dose of ara-
`AC in this Phase I clinical study was chosen to be 0.2 mg/m2/h for 72
`h. Dose escalation was allowed in individual patients. A minimum of
`three patients were treated at each dose level before escalation to the
`subsequent dose level was permitted in new patients. The ara-AC dose
`was initially escalated to 0.4 mg/m2/h and then to 0.8 mg/m2/h for 72
`h with subsequent doses being escalated by 25% of the previous dose
`level (as indicated in Table 3). ara-AC was administered every 21 days
`if the patients had recovered from toxicity of the previous cycle;
`otherwise, administration of the next dose was delayed for 1 week. The
`objective of the study was to determine the maximally tolerated dose
`of ara-AC when given as a 72-h continuous
`infusion and to determine
`the pharmacokinetics of this agent in humans. The maximally tolerated
`dose was defined as the dose at which greater than 50% of the patients
`treated at a given dose experienced Grade III or greater
`toxicity.
`Toxicity was graded according to the guidelines recommended by the
`Cancer Therapy Evaluation Program, National Cancer Institute.
`In Vitro Stability. An appropriate volume of 1.0 x 10~3M ara-AC in
`
`DMSO was added to fresh heparinized human plasma (pH 7.6) from a
`normal volunteer and to 0.1 M phosphate-buffered saline solution (pH
`7.2) at 37°Cto give a final ara-AC concentration of 1 Mg/m' (4.1 JUM).
`A 0.5-ml aliquot was taken from each sample before incubation (t = 0),
`and the remainder of the sample was then incubated at 37 ±0.2°C.At
`predetermined intervals 0.5-ml aliquots were removed from each sam
`ple, and ara-AC concentrations were measured by the method described
`below. Sampling was continued for a minimum of three half-lives.
`Pharmacokinetic Studies. Blood samples were obtained in 10 ml
`heparinized tubes both prior to administration
`of ara-AC and at 24-h
`intervals after the initiation of therapy. Each sample was maintained
`on ice to retard hydrolysis of the parent drug, and plasma was separated
`from cells as soon as possible by centrifugation at 1000 x g for 10 min.
`Because of limited sample stability, an aliquot of plasma was processed
`and analyzed within 4 h using a modification of a reverse-phase high-
`performance liquid chromatography assay (16).
`Briefly, 2 ¿ilof 2.25 x IO"3 M (1.0 ¿ig)2'-deoxy-5-azacytidine were
`added to a 0.5-ml aliquot of plasma as an internal standard. The sample
`was vortexed for 10s to mix the internal standard. A 1.0-ml phenyl-
`boronic acid solid-phase extraction cartridge (Bond Elut PBA; Analy-
`tichem International, Harbor City, CA) was activated by washing with
`1 ml of methanol
`followed by 1 ml of 0.01 M phosphate buffer, pH 8.O.
`Attached to this by an adaptor was a 3.0 ml disposable cartridge to
`which the plasma sample was transferred. The sample was then slowly
`pushed through the cartridge and the eluant was collected. The cartridge
`was rinsed with 0.5 ml of 0.01 M phosphate buffer, pH 8.0, and this
`was then ultrafiltered by centrifugation at 1000 x g for 30 min at 4°C
`in a Centrifree partition system (Amicon Corp., Danvers, MA).
`A 100-fil aliquot was analyzed using a Model 204W liquid chroma
`tography system (Walters Associate, Milford, MA) consisting of a U6K
`injector, a Model 6000 A solvent delivery system, and a Model 116
`1221
`
`Ara-C
`5-AC
`Fig. l. Chemical structures of S-AC, Ara-C. Ara-AC. D, changes in structure
`compared to cytidine.
`
`Ara-AC
`
`27
`26
`82(79)
`
`52
`25-74
`
`17/10
`82
`
`24
`10
`1
`12
`
`Table 1 Patient characteristics
`No. of patients entered
`No. évaluable
`Total no. of cycles (évaluable)
`
`Age. yr
`Mean
`Range
`
`Male/female
`Mean performance status (Karnofsky)
`
`No. of patients with prior therapy
`Chemotherapy only
`Radiation therapy only
`Chemotherapy and radiation therapy
`
`Diagnosis
`Acute nonlymphocytic leukemia
`Adenocarcinoma
`liver, unknown primary
`Adrenal cancer
`Breast cancer
`Colon cancer
`Diffuse lymphoma
`Embryonal cell carcinoma
`Fibrohistiocytoma
`Glioma
`Nodular lymphoma
`Non-small cell lung cancer
`Osteogenic cancer
`Ovarian cancer
`Pancreatic carcinoma
`Renal cell carcinoma
`Small cell lung cancer
`Squamous cell cancer, anus
`Uterine sarcoma
`
`graphic studies. Patients in this study had to have a performance status
`of greater than 50% on a Karnofsky scale. Pretreatment evaluation also
`included complete blood cell count, with WBC differential,
`serum
`chemistries, coagulation studies, serum creatinine, and creatinine clear
`ance. All patients had adequate liver (bilirubin <1.5 mg/100 ml and
`serum transaminases
`less than twice the normal
`range) and renal
`(creatinine clearance of 45 mi/min or a serum creatinine <1.5 mg/100
`ml) function, as well as normal serum electrolytes and normal coagu
`lation studies. All patients evaluated for toxicity had adequate periph
`eral blood cell counts (WBC > 3,000/mm3, platelets > 100,000/mm')
`prior
`to initiation of study. Complete blood cell counts and liver
`function tests were done weekly while the patients received therapy and
`they were seen at least once every 3 weeks prior to a new cycle at which
`time all laboratory tests were repeated. At each visit, patients were
`asked to report and subjectively rate every possible drug-related side
`effect. All patients gave written informed consent prior to therapy in a
`manner consistent with institutional guidelines. Patients who developed
`progressive disease after two cycles were removed from study.
`Drug Formulation and Dosage. ara-AC was supplied by the Division
`of Cancer Treatment, National Cancer Institute (Bethesda, MD), as a
`sterile lyophilized powder in 250-mg vials. Because of limited aqueous
`stability,
`the vials were reconstituted with 9.9 ml of 70% (v/v) DMSO
`
`© 1990 American Association for Cancer Research.
`
`

`

`PHASE l ARA-AC (FAZARABINE)
`
`Table 2 Pharmacokinetic parameters of ara-AC
`Clearance*
`(ml/min/m2)
`
`Dose
`(mg/mVh)
`1.25
`1.75
`2.44
`3.83
`4.75
`5.94
`
`(ng/ml)
`32
`50
`57
`71
`108
`137
`
`651
`583
`713
`899
`733
`729
`
`Cycle
`
`578
`
`10
`II
`13
`
`Patient
`W. P.
`
`33127
`
`1.962.441.96
`
`724761627l
`
`1.962.96
`
`527695
`
`
`
`1143122111.961.562.441.562.441.961.962.442.442.447437
`6270455654685862441703454865536
`65637190458360S598701656
`
`Q.W.J.
`
`F.R.
`
`D.M.
`
`T.K.
`
`S.E.W.V.
`
`N.P.S.D.
`
`W.H.W.T.
`
`S.R.
`
`J.32
`°Measured values of 30-69 ng/ml are in the range of less-certain quanlitation.
`" Mean ±SD. 647 ±141 (N = 21).
`
`V = 3 146 + 21.7X
`r = 0.9660
`
`150
`
`100
`
`1
`
`2345
`
`6
`
`-
`E
`
`UV detector (Gilson Medical Electronics. Middleton, WI). The sample
`was separated on a 4.6- x 250-mm, 5-^m Ultrasphere octadecylsilane
`analytical column (Altex/Beckman. Berkeley. CA) with a mobile phase
`of 0.5% CHjCN in 0.01 M phosphate buffer. pH 6.8. with a flow rate
`of l ml/min. The detector
`signal was integrated by a dual channel
`Spectra-Physics 4200 computing and recording integrator, which was
`interfaced with a ChromStation-AT data system (Compaq 386 version;
`Spectra-Physics, Santa Clara, CA) for further data reduction and stor
`age.
`Standard curves of ara-AC in plasma were prepared for each patient's
`samples by addition of know n amounts of ara-AC to the corresponding
`pretreatment
`sample. These spiked standards were processed in the
`same manner as above. A standard curve was prepared for analysis each
`day for the range 0 to 400 ng/ml. This cune was the best straight
`line
`defined by least squares regression analysis and possessed a correlation
`coefficient >0.997. Ratios of peak height of ara-AC standard to peak
`height of internal
`standard were obtained for a minimum of four
`concentrations each day. Samples were run concurrently, and concen
`trations were calculated from the appropriate
`standard curve. ara-AC
`peaks below the automatic detection threshold of the integrator were
`measured using the data system software. UV detection at 240 nm
`allowed a limit of quantitation (S/N >5) of 70 ng/ml
`(0.3 MM)and a
`limit of detection of 30 ng/ml
`(0.12 //M). ara-AC concentrations
`be
`tween these two limits could also be estimated; but, since these values
`have a greater error,
`they are indicated as falling in the range of less-
`certain quantitation (17) in Table 2.
`Kinetic Calculations. For determination of in vitro stability, half-lives
`were calculated from the best straight
`line obtained by linear regression
`analysis of the logarithm of ara-AC concentration versus time. C«was
`the average of measured 24-h ara-AC plasma concentrations,
`since ara-
`AC levels were observed to reach a plateau by 8 h in patients given
`higher doses on a 24-h infusion schedule (16). The apparent
`total body
`clearance was then defined as the rate of infusion divided by Ca.
`
`RESULTS
`
`In Vitro Stability. ara-AC was unstable in both phosphate-
`buffered saline and human plasma at the clinically achievable
`concentration
`of 1 pg/m\
`(4.1 ^M) (18). The disappearance
`of
`ara-AC in these in vitro experiments was apparent
`first order,
`with the half-life of 2.5 ±0.2 h (n = 2) in fresh human plasma
`incubated at 37°Cbeing much more rapid than the 9.8 ±0.3 h
`(n = 3) observed during incubation in phosphate-buffered
`saline
`under the same conditions.
`Pharmacokinetics.
`ara-AC was measured in the plasma of 13
`patients receiving 21 cycles of therapy at doses above 1.25 mg/
`nr/h. Css ranged from 32 ng/ml
`(0.13 //M) to 137 ng (0.6 JUM)
`(Table 2). For Patient W. P., who received 13 cycles of therapy
`and had the most extensive blood level study, a good linear
`relationship of measured Css with dose was calculated (Fig. 2).
`A relatively rapid mean clearance of 647 ml/min/m2 was ob
`served for all patients (Table 2).
`Toxicity. Twenty-six évaluablepatients were treated with a
`total of 79 cycles of ara-AC. The dose-limiting
`toxicity was
`reversible leukopenia
`and thrombocytopenia
`(Table 3). Grade
`IV leukopenia was observed in 1 of 7 patients
`(1 of 7 cycles)
`given ara-AC at 1.25 mg/nr/h,
`in 1 of 8 patients
`(2 of 11
`cycles) at 1.56 mg/nr/h,
`and in 4 of 11 patients (5 of 14 cycles)
`at 1.96 mg/nr/h. Overall, Grade III or Grade IV leukopenia
`occurred in 6 of 11 patients treated at 1.96 mg/nr/h
`and in 4
`of 7 treated at 2.44 mg/nr/h.
`This was not a reflection of
`cumulative drug toxicity,
`inasmuch as 2 of 4 new patients who
`entered treatment with ara-AC at 1.96 mg/nr/h
`and 4 of 4 new
`patients who entered at 2.44 mg/nr/h
`experienced Grade III
`or IV neutropenia. Thrombocytopenia
`also occurred but overall
`was less prominent
`than leukopenia. Grade IV thrombocyto
`penia was observed in 1 of 7 patients (1 of 7 cycles) treated at
`1222
`
`Dose (mg/m?/hl
`(ng/ml or UM)at
`Fig. 2. Correlation between ara-AC plasma concentrations
`steady state and dose (mg/m2/h)
`in a single patient (W. P.) given multiple cycles
`of therapy.
`
`1.25 mg/m2/h, 1 of 8 patients (2 of 11 cycles) treated with 1.56
`mg/nr/h,
`and 3 of 11 patients (3 of 14 cycles) treated with 1.96
`mg/nr/h
`for 72 h. The median time to WBC nadir was 18 days
`(range, 7 to 31), while the median time to the corresponding
`platelet nadir was 13 days (range, 7 to 22).
`There were five episodes of fever and neutropenia
`
`requiring
`
`© 1990 American Association for Cancer Research.
`
`

`

`PHASE [ ARA-AC (FAZARAB1NE)
`
`Table 3 Hemalological
`
`loxicily of ara-AC
`nadir*IV111411I33113II
`
`of
`patients
`of
`Dose
`III12ii31IV113
`
`(new/total)3/33/53/61/64/74/84/114/70/30/10/10/1No.II111
`cycles357771114174112WBCI
`(mg/inVh)0.20.40.81.01.251.561.962.443.053.834.755.94No.
`
`21
`22
`211nadir"III222323Platelet
`
`leukocytes/mm3: Grade I 3.000-3.999; Grade II. 2.000-2.999; Grade
`"Total
`III, 1,000-1,999; Grade IV, <1,000.
`* Platelets/mm3: Grade 1, 75,000-90,000: Grade II. 50.000-74.000; Grade III,
`25,000-49,000; Grade IV, <25,000.
`
`during this study.
`treatment
`for i.v. antibiotic
`hospitalization
`In three cases no source of infection was detected. One patient
`developed Candida and herpes esophagitis during the period of
`fever and neutropenia. One patient with adrenocortical
`cancer
`developed Listeria meningitis
`at which time the WBC was
`1400/mm3. She was treated successfully and recovered without
`sequelae. Another patient became neutropenic during his 12th
`cycle of ara-AC therapy. He subsequently developed staphylo-
`coccal septicemia
`and Candida
`endocarditis
`on a prosthetic
`mitral valve (he had a previous history of Candida endocarditis).
`Although his WBC returned to normal, he subsequently died
`of cardiac complications. This was the only treatment-related
`death on the study.
`Increases in liver function tests were observed in 14 patients
`treated with ara-AC. This was predominantly manifested by
`increases in alkaline phosphatase although there were also some
`minor increases in transaminases. Overall,
`there was no appar
`ent relationship between ara-AC dose and liver toxicity. Grade
`I liver toxicity (increased alkaline phosphatase
`or
`transami
`nases) was seen in 19 cycles (occurring in doses ranging from
`0.4 to 2.44 mg/m2). Grade II in 13 cycles (in doses ranging
`from 0.8 to 5.94 mg/m2), Grade III in one cycle (at 1.56 mg/
`m2), and Grade IV toxicity in 2 cycles (at 1.96 mg/m2). In each
`case,
`the hepatic abnormality was transient
`and returned to
`baseline prior to the next cycle. One-half (7 of 14) of the patients
`who developed elevated liver function tests in this study had
`evidence of liver involvement with tumor,
`thus complicating
`the evaluation for hepatotoxicity.
`There were few other side effects reported in this trial. Since
`ara-AC was reconstituted
`in DMSO in order
`to enhance
`its
`stability during prolonged infusions,
`the odor of DMSO was
`noticeable to everyone who entered the patients'
`rooms. How
`ever, the patients themselves did not complain of the odor, nor
`did they complain of any garlic-like taste. Nausea and vomiting
`were uncommonly
`(7 of 27 patients)
`reported
`in this trial.
`Overall,
`there were only 5 episodes of Grade I and 2 Grade II
`nausea and vomiting. One patient
`received a total of 13 cycles
`with an escalation from 0.2 mg/m2/h to 5.94 mg/m2/h for 72
`h. This 44-year-old patient complained
`of sleep disturbances
`(nightmares
`and excessive tiredness) beginning after Cycle 4
`and of
`impotence
`after Cycle 6. These toxicities were not
`observed in any other patient.
`Clinical Activity. In this Phase I study, both prolonged disease
`stabilization in one patient and minor clinical responses in four
`other patients were observed. One patient
`(W. P.) with an
`adenocarcinoma
`of the liver of unknown primary had stable
`disease without any evidence of progression by abdominal com
`
`scan during 13 cycles of therapy
`tomographic
`puterized axial
`with ara-AC. Of the four patients who were noted to have minor
`clinical
`responses, one (E. W.) entered the study with rapidly
`progressing liver métastasesfrom colon carcinoma. Following
`two cycles of ara-AC the size of the liver as noted by physical
`examination
`diminished
`from 28 cm to 19 cm. This patient
`developed disease progression in the retroperitoneum after four
`cycles of therapy. One patient
`(V. L.) with metastatic breast
`cancer had some shrinkage in the size and character of extensive
`chest wall
`lesions after
`three cycles of ara-AC but developed
`progressive disease after the fourth cycle. One patient
`(T. R.)
`previously treated with chemotherapy and radiation therapy for
`Hodgkin's disease had developed acute nonlymphocytic
`leuke
`mia. Following treatment with ara-AC,
`there was a diminution
`in the number of circulating blasts, a decrease in láclatedehy-
`drogenase,
`a decrease
`in fever, and a feeling of subjective
`improvement. After two cycles of therapy,
`there was a complete
`disappearance
`of biopsy-proved
`leukemic skin lesions. How
`ever, bone marrow biopsies failed to show a decrease in the
`proportion of blasts, and the patient's
`therapy was stopped after
`three cycles.
`cell
`(K. S.) in this study had embryonal
`Another patient
`carcinoma
`and had been previously treated with two different
`salvage combination
`chemotherapy
`regimens. At
`the time of
`entering this trial,
`this patient had bilateral adrenal masses, a
`testicular mass, pleural effusions, and ascites which required
`twice-weekly paracentesis. After
`the first cycle of ara-AC, he
`no longer required paracentesis, and after the fourth cycle, there
`was no evidence of ascites by abdominal
`computerized
`axial
`tomographic
`scan. He was subsequently treated with an addi
`tional 9 cycles of ara-AC. Despite clinical
`improvement
`there
`was no apparent change in the size of his adrenal or testicular
`masses and his «-fetoprotein increased during this time from
`600 to >11,000. An orchiectomy was performed which showed
`an embryonal cell carcinoma with yolk sac elements, which had
`not been a prominent
`feature of his previous biopsies.
`It
`is
`possible that ara-AC promoted the differentiation of the tumor
`toward a yolk sac tumor
`resulting in enhanced «-fetoprotein
`production.
`Indeed, ara-AC does induce differentiation of HL-
`60 cells (8), a property presumed to be related to its ability to
`inhibit DNA methylation (11). It is also possible that the testes
`represented a sanctuary site. Following orchiectomy,
`this pa
`tient's a-fetoprotein
`returned to normal. The patient developed
`neutropenia
`during a subsequent
`cycle of ara-AC which was
`complicated by staphylococcal
`sepsis and Candida endocarditis
`on a prosthetic mitral valve (the patient had developed Candida
`endocarditis
`during earlier
`treatments
`for testicular
`cancer).
`The patient
`subsequently
`died of cardiac complications. At
`autopsy,
`there were only necrotic masses in the abdomen and
`adrenals without evidence of tumor. He did, however, have two
`small sites of residual embryonal cancer in his lung.
`
`DISCUSSION
`
`In this Phase I trial, ara-AC was given as a 72-h infusion to
`adult patients with refractory cancer. The dose-limiting toxicity,
`which was reversible leukopenia, was reached at 1.96 mg/nr/h
`for 72 h in this heavily pretreated patient population. Neutro
`penia was more prominent
`than thrombocytopenia. Other
`tox
`icities included mild elevations in liver function abnormalities.
`Hepatotoxicity was difficult
`to evaluate in this patient popula
`tion since many of the patients had metastatic disease in the
`liver. Nausea and vomiting were not prominent
`side effects of
`this drug, despite the obvious odor produced by the DMSO
`1223
`
`© 1990 American Association for Cancer Research.
`
`

`

`PHASE l ARA-AC (FAZARABINE)
`
`REFERENCES
`
`of
`10 months) was seen in one patient with adenocarcinoma
`unknown origin involving the liver. Minor clinical
`responses
`were observed in one patient with acute nonlymphocytic
`leu
`kemia,
`in one patient with colon cancer metastatic to liver, in
`one patient with breast cancer, and in one patient with em
`bryonal
`testicular cancer. This wide range of possible clinical
`activity mirrors the breadth of observed preclinical activity.
`ara-AC is thus a new antineoplastic
`agent which appears to
`combine the chemical and clinical features of ara-C and 5-AC.
`The 70% DMSO formulation
`of ara-AC is well tolerated by
`patients when it
`is piggy-backed into a rapidly dripping i.v.
`solution and given continuously for 72 h. Although all of our
`patients have been treated in the hospital,
`it should be possible
`to deliver ara-AC as an outpatient
`regimen. The only limiting
`feature to outpatient
`therapy may be the noxious odor produced
`by the DMSO. The recommended dose for Phase II studies in
`heavily pretreated
`patients
`is 2.0 mg/nr/h
`for 72 h. Higher
`dose levels should be tolerated by individuals who have received
`little or no prior chemotherapy.
`Preclinical
`studies and this
`Phase I study suggest
`that ara-AC may have a broad spectrum
`of activity against human solid tumors.
`
`which was required for the effective formulation of the clinical
`dose.
`The mean clearance for the 21 cycles of therapy for which
`ara-AC could be measured was 647 ±141 ml/min/nr,
`which
`is not statistically different
`from the mean clearance (571 ml/
`min/nr) observed in the Phase I pediatrie trial of this agent for
`a dose of 15 mg/m2/h over 24 h (18). These clearances also fell
`into the 480-
`to 1157-ml/min/m2
`range observed in non human
`primates
`for ara-AC doses of 200 mg/kg given over either 15
`min or l h (19). This rapid clearance of ara-AC is about one-
`half that of ara-C (20), the nucleoside that ara-AC most closely
`resembles biochemically. While ara-C is rapidly metabolized by
`cytidine deaminase, araAC is refractory to catabolism by this
`enzyme (5). Thus, any clearance of ara-AC through deamina-
`tion, even considering the high levels of this enzyme in human
`liver and kidney (21), is expected to be minimal as is the case
`for 2',3'-dideoxycytidine,
`another cytosine nucleoside which is
`a poor substrate for mammalian cytidine deaminase (22). Based
`on in vitro stability, however, ara-AC probably undergoes sub
`stantial hydrolytic degradation in vivo. If one assumes a steady-
`state volume of distribution equivalent
`to total body water (42
`liters for a 70-kg person), which is what has been observed for
`ara-AC in monkeys (19), a maximum hydrolytic contribution
`to plasma clearance of 110 ml/min/m2 can be calculated (23).
`This indicates that other as yet to be defined processes such as
`1. Beisler. J. A., Abbasi, M. M.. and Driscoll, J. S. The synthesis and antitumor
`activity of arabinosyl-5-azacytosine. Biochem. Pharmacol.. 26: 2469-2472,
`distribution into other compartments
`(e.g., intracellular
`trans
`1977.
`port),
`renal excretion,
`and anabolic metabolism may account
`2. Beisler, J. A., Abbasi, M. M., and Driscoll, J. S. Synthesis and antitumor
`activity of 5-azacytosine arabinoside. J. Med. Chem., 22: 1230-1234, 1979.
`for the majority of plasma clearance.
`3. Mojaverian, P., and Repta, A. J. Development of an intravenous formulation
`Steady plasma levels of ara-AC ranged from 32 to 137 ng/
`for the unstable investigational cytoloxic nucleosides S-azacytosine arabino
`ml (0.13 to 0.6 MM)for patients receiving drug at rates between
`side (NSC 281272) and 5-azacytidine (NSC 102816). J. Pharm. Pharmacol.,
`Jo: 728-733, 1984.
`1.25 and 5.94 mg/nr/h
`(Table 2; Fig. 2). Studies with ara-AC
`4. Notar!, R. E., and DeYoung, J. L. Kinetics and mechanisms of degradation
`in several
`in vitro systems suggest
`that
`these observed concen
`of the antileukemic agent S-azacytidine in aqueous solutions. J. Pharm. Sci.,
`64: 1148-1156, 1975.
`trations may be sufficient
`for biological activity, especially if
`5. Townsend, A., Ledere, J. M., Dutschman, G., Cooney, D. A., and Cheng,
`they are maintained for the better part of 72 h. Incubation of
`Y. C. Metabolism of l-$-r>arabinosyI-5-azacytosine
`and incorporation into
`Molt-4 human T-lymphoblastic
`leukemia cells with 1.0 ^M ara-
`DNA of human T-lymphoblastic cells (Molt-4). Cancer Res., 45:3522-3528,
`1985.
`AC for 24 h resulted in 98% inhibition in clonogenic assays
`6. Vesely, J., and Piskala, A. Mechanism of action of 1-0-D-arabinofuranosyl-
`(5). Although a direct comparison is not possible because drug
`5-azacytosine and its effects in LI 210 leukemia cells. Neoplasma, 33: 3-10,
`concentrations with respect
`to time are not known for the in
`1986.
`In: B. A. Chabner (éd.).Pharmacologie
`7. Chabner, B. A. Cytosine arabinoside.
`vitro study, drug exposure was probably similar in both cases.
`Principles of Cancer Treatment, pp. 387-401. Philadelphia: W. B. Saunders
`If one takes into account
`the half-life of ara-AC in RPMI 1640
`Co., 1982.
`(19),
`then the average drug concentration
`for
`the 24 h of
`8. Dalai, M., Plowman, J., Breitman, T. R., Schuller, H. M., Del Campo, A.
`A., Vistica, D., Cooney, D. A., and Johns, D. G. Arabinofuranosyl-5-
`incubation is 0.2 ¿ÕM,a level that should be achievable at
`the
`azacytosine: antitumor and cytotoxic properties. Cancer Res., 46: 831-838,
`recommended
`dose of 2.0 mg/m2/h for Phase II studies.
`In
`1986.
`J.
`9. Ahluwalia, G. S., Cohen, M. B., Kang, G.-J., Arnold, S. T., McMahon,
`addition,
`the inhibitory concentration
`for 50% control growth
`B., Dala, M., Wilson, Y. A., Cooney, D. A., Balzarini, J., and Johns, D. G.
`for a 24-h exposure of ara-AC to P388 murine leukemia in vitro
`Arabinosyl-5-azacytosine: mechanisms of native and acquired resistance.
`was found to be 1.9 ^M, and exposure of these same cells to 2
`Cancer Res., 46:4479-4485,
`1986.
`10. Grant, S., Bhalla, K., and Gleyzer, M. Effect of uridine on response of 5-
`fiM ara-AC for 3 day