Oral Cytarabine Ocfosfate in Acute Myeloid Leukemia and Non-Hodgkin’S Lymphoma – Phase I/II Studies and Pharmacokinetics

Home , Cytarabine

Leukemia (1998) 12, 1618–1626  1998 Stockton Press All rights reserved 0887-6924/98 $12.00 http://www.stockton-press.co.uk/leu Oral cytarabine ocfosfate in acute myeloid leukemia and non-Hodgkin’s lymphoma – phase I/II studies and pharmacokinetics

J Braess1, M Freund2, A Hanauske3, G Heil4, C Kaufmann1, W Kern1, M Schu¨ssler5, W Hiddemann1 and E Schleyer1

1Department of Hematology and Oncology, University of Go¨ttingen, Germany; 2Department of Hematology and Oncology, University of Rostock, Germany; 3Department of Oncology, Universitair Ziekenhuis Gasthuisberg, Leuven, Belgium; 4Department of Hematology and Oncology, MHH Hannover, Germany; and 5ASTA Medica, Frankfurt/M, Germany

Cytosine arabinoside (AraC) is rapidly inactivated via systemic logical malignancies. Especially in the treatment of acute mye- deamination with half-lives ranging from 1 h (i.v.) to 4 h (s.c.) – logenous leukemias AraC constitutes the basis of most and cannot be applied orally due to its hydrophilic properties. 1–3 These limitations might be overcome by YNK01 – a lipophilic currently used combination regimens. prodrug of AraC – that is resistant to deoxycytidine deaminase Due to its hydrophilic properties, AraC cannot be applied and can be applied orally. In the present study the therapeutic via the oral route but must be given by intravenous (i.v.) or activity, side-effects and pharmacokinetics of YNK01 were subcutaneous (s.c.) injection. Following i.v. or s.c. application evaluated in a phase I/II study including patients with relapsed AraC is rapidly deaminated by the ubiquitous cytidine deami- or refractory acute myeloid leukemia (AML) (n = 23) or low- = nase to its non-cytotoxic metabolite uracil arabinoside grade non-Hodgkin’s lymphoma (NHL) (n 20). YNK01 was 4–6 given by 14 day cycles with escalating doses starting with a (AraU) resulting in short plasma half-lives of approximately daily dose of 50 mg/m2 (equivalent to 20 mg/m2 AraC on a molar 1 (i.v.) to 4 (s.c.) h only. In an attempt to overcome these limi- basis). The maximum tolerated dose was reached at the tations a series of lipophilic AraC analogues was synthesized.7 600 mg/m2 dose level with WHO grade 3–4 diarrhoea as the Of these, cytarabine ocfosfate (YNK01) with a C18 chain as main toxicity. In the 23 patients with AML two complete the lipid moiety revealed the highest antileukemic activity in remissions, four partial remissions and three patients with vitro and in animal models and was therefore further pur- stable disease were observed. In the 23 patients with AML two 8 complete remissions, four partial remissions and three patients sued. In preclinical investigations (H Takayama, 1990; and with NHL two cases reached partial remission and six other H Takayama and Y Esumi 1990, unpublished results) it was patients mainained stable disease. Pharmacokinetic evalu- shown that the drug is absorbed from the intestine following ations were performed during 34 treatment cycles in 28 oral application. It is subsequently taken up into hepatocytes patients. The data suggest that YNK01 was absorbed in the dis- and slowly metabolized to AraC by a cytochrome P450- tal part of the small intestine and taken up into hepatocytes. dependent oxydative enzyme system.9 The cytotoxic nucleo- After hepatic ␸ and subsequent ␤-oxydation of YNK01 the released AraC (and its deamination product AraU) appeared in side AraC is subsequently released into the systemic circu- the systemic circulation. Time of maximum concentration (h), lation over a prolonged period of time. Aside from the advan- half-life (h) and area under the curve (ng·h/ml, at the 1200 mg tage of a convenient oral application these pharmacokinetic dose level) were as follows (VC variation coefﬁcient) YNK01: properties also suggest a prolonged systemic exposure to AraC 1.0 (0.58), 10.1 (0.43), 12622 (0.65); AraC: 23.2 (0.57), 22.6 (0.36), due to long-lasting plasma levels. To further evaluate these 3496 (0.76); AraU: 19.2 (0.58) 22.3 (0.33) 15441 (0.66). Of the perspectives and their clinical relevance, a phase I/II study of total dose of YNK01 15.8% was absorbed and metabolized to AraC and AraU, deﬁning the metabolic bioavailability of this cytarabine ocfosfate was initiated evaluating the side-effects prodrug. A linear relationship was observed between YNK01 and therapeutic activity but particularly the pharmacokinetics dose and YNK01 AUC and AraC AUC over the whole dose range of the mother drug and its two metabolites (AraC, AraU).10 tested. AraC was released from hepatocytes over a prolonged period of time resulting in long lasting plasma levels similar to a continuous i.v. infusion. After administration of YNK01 at a Patients and methods dosage of 100–150 mg/m2 plasma levels of AraC were compara- ble to those achieved after low-dose AraC treatment (20 mg/m2) while at doses of YNK01 of 450–600 mg/m2 concentrations of Patients standard-dose AraC (100 mg/m2) were obtained. We conclude that YNK01 shows considerable activity against relapsed and Patients with either advanced low-grade NHL or released or refractory AML and NHL and that its pharmacokinetic proper- refractory AML were eligible for the present study. For ties offers advantages in comparison to (standard) i.v. or s.c. inclusion a WHO performance status grade 3 or better, age AraC in clinical practice. у Keywords: pharmacokinetics; oral application; AML; NHL; cyto- 18 years as well as normal liver and kidney function were sine arabinoside required. All patients were required to be ineligible for standard second/third-line treatment as judged by the local inves- tigator in order to qualify for the investigational drug treat- Introduction ment. Following information about the investigational nature of the studies all patients gave their written informed consent. Cytarabine ocfosfate (YNK01, 1-␤-d-arabinofuranosylcytosine Prior to its initiation the study was approved by the local eth- 5Ј-stearylphosphate) is a lipophilic derivative of cytosine arab- ics committee at the participating institutions. The therapeutic inoside (AraC). The nucleoside analogue cytosine arabinoside and investigational aspects of the study are in accordance with is one of the most effective agents in the treatment of hemato- the Declaration of Helsinki.

Treatment schedule Correspondence: J Braess, Georg-August-University, Department of Hematology and Oncology, Robert-Koch-Strasse 40, 37075 Go¨ttingen, Germany; Fax: 49 551 39 2914 Preceding the ﬁrst cycle of treatment in every patient a single Received 17 April 1998; accepted 26 June 1998 lead dose of the respective dose level was applied for determi- Cytarabine ocfosfate in ANL and NHL – pharmacokinetics J Braess et al 1619 nation of baseline pharmacokinetic parameters. Seventy-two column was performed. Separation modalities involved ion-

hours following the lead dose cytarabine ocfosfate was given pairing on a reversed-phase C6H5 phenyl column with a by oral administration over 14 consecutive days. The drug was mobile phase of 50% acetonitril and 50% 0.04 m sodium applied as capsules of 100 mg kindly provided by ASTA Med- phosphate buffer for isocratic elution. An UV detector was ica, Frankfurt/M, Germany. In order to facilitate comparative used for signal detection at 275 nm. Recovery following sam- evaluations drug doses were assessed on the basis of calcu- ple preparation was between 52 and 64%. lated square meter concentrations rather than on the absolute An HPLC assay was developed for the detection of AraU. levels that were clinically administered. Starting at an initial Plasma samples were prepared by precipitation of proteins 2 dose level of 50 mg/m /day a dose escalation was performed using HClO4 and speed-vac drying of the supernatant. The with initial steps of 50 mg up to 200 mg/m2. Subsequently, specimen was then taken up in a 10-fold higher concentration doses were escalated in steps of 150 mg/m2 starting at compared to the original sample in order to achieve a limit 300 mg/m2 up to 900 mg/m2. A minimum of three patients of detection of 20 ng/ml. Urine samples were prepared by

was included at each dose level and doses were escalated adding HClO4 to the sample followed by 60 min of heating to the next level if no non-hematological toxicity WHO Ͼ2 at 90°C. Samples were then eluted over SCX anion exchange was observed. columns for final cleansing. As separation modalities a 3 ␮ ␮ C18 reversed phase column and a 5 SA cation exchange column with a 0.1 m sodium phosphate buffer as the mobile Response and toxicity criteria phase were used. A limit of detection of 20 ng/ml and 500 ␮g/ml for plasma and urine, respectively, was reached. Response in AML was assessed according to CALGB criteria The coefficients of variance were moderately low at 14.2% (1979). A complete remission (CR) required a bone marrow and 16.5% for plasma and urine, respectively, at the detection smear with less than 5% blasts and a normal peripheral blood limit and approximately 10% at higher concentrations. count. A partial remission (PR) was defined as a marrow infil- For AraC detection, a radio-immunoassay first described by tration of 6–25%, less than 5% blasts in the peripheral blood, Enomoto (unpublished report 1990) was used in order to a leukocyte count of 2000–4000/␮l(Ͼ500/␮l granulocytes), reach a sufficiently low limit of detection. Conditions were hemoglobin Ͼ8 g/dl and thrombocytes 25 000–100 000/␮l. modified in order to optimize sample preparation by precipit- Patients were considered as having stable disease (SD) if PR ating plasma samples with a low volume of acetonitril and criteria were not met and no progression of disease had speed-drying the supernatant. A limit of detection of 100 occurred. pg/ml and 1 ng/ml for plasma and urine was reached, respect- In NHL CR was defined as the disappearance of all disease ively. The intra-assay coefficient of variation was 18% at the manifestations. A PR was diagnosed if measured lesions had detection limit. regressed by more than 50% of the initial size. The bone mar- All data were analyzed by using the TOPFIT pharmacoki- row had to be negative. Stable disease (SD) was diagnosed if netic computer program which allowed an optimized adap- criteria of PR were not met but no progression of the disease tation of variation coefficients between the observed and cal- had occurred. culated data, respectively.11 The data of all three substances Toxicities were assessed according to WHO criteria. were fitted to a linear one-compartment model. The plasma decay curve of cytarabine ocfosfate, AraC and AraU and the cumulative renal elimination of the latter two substances were Pharmacokinetic analysis determined independently.

For pharmacokinetic analyses no distinction was made between AML and NHL. Blood samples were taken prior to and 1, 2, 4, 6, 8, 12, 16, 20, 24, 28, 32, 48, 62 and 76 h after Results the lead dose and after the last dose of the regular 14 day regimen to reliably establish plasma half-lives of cytarabine ocfosfate, AraC and AraU. During the treatment cycle six Clinical efficacy further samples were taken prior to the 1st, 2nd, 3rd, 6th, 9th, 12th, 14th dose in order to estimate the pseudo steady-state levels of all three substances. In order to determine the renal Twenty-three patients with AML and 20 patients with NHL excretion of all three substances urine was collected for 72 h were included in the study and were evaluable for clinical following the lead dose and the last dose of the cycle. response. In the 23 AML patients, two CRs, four PRs and three Blood samples were collected in heparinized vacutainers to SDs were observed. CRs were achieved at daily doses of which 250 mg of tetrahdyrouridine (THU) were added for the 300 mg/m2 and 450 mg/m2 in relapsed patients (prior CR Ͻ1 inhibition of cytidine deaminase. Samples were immediately year). PRs were observed in four patients (relapsed and centrifuged, plasma was removed and instantly frozen at refractory) at doses of 450 mg/m2 and 600 mg/m2. SD was −20°C until analysis. Urine specimens were similarly pre- maintained in three patients (relapsed and elderly/unfit) at pared. After measuring the total urine volume and stirring tho- doses of 150, 200 and 600 mg/m2. roughly an aliquot of 50 ml was frozen at −20°C. From the total of 20 NHL two patients responded with a PR Cytarabine ocfosfate concentrations in plasma and urine and six patients experienced a SD. PRs were achieved in two were determined by HPLC analysis (Ramsauer et al 1994). In patients with centrocytic lymphoma and B-CLL and were brief, sample clean-up was performed using tubes filled with maintained for 15 and Ͼ24 months at doses between 2 2 C18 material. In order to reach a low limit of quantification in 450 mg/m and 750 mg/m . SD was obtained in patients with the range of 5 ng/ml (variation coefficient (VC) 6.8%) and of centrocytic lymphoma, B-CLL (three patients), lymphoplasmo- 500 pg/ml (VC 7.3) for plasma and urine, respectively, enrich- cytoid lymphoma and centroblastic–centrocytic lymphoma at ment on C18 cartridges prior to elution on to the analytical daily doses of 50, 100, 300 and 450 mg/m2. Cytarabine ocfosfate in ANL and NHL – pharmacokinetics J Braess et al 1620 Toxicities

The maximum tolerated dose (MTD) was reached at the 600 mg/m2 dose level with two of six patients not completing the first cycle due to diarrhoea WHO grades 3 and 4, respectively. Other non-hematological toxicities did not exceed Cytarabine ocfosfate WHO grades 1 and 2. Major hematological toxicities were experienced by one AML patient who went off study after one cycle due to non-recovery from hematological toxicity within On average cytarabine ocfosfate first appeared after a lag-time 50 days after the start of treatment. In the NHL group throm- of 1.0 h. Maximum plasma concentrations were reached after bocytopenia (WHO grades 1 and 2) was the main hematolog- an average of 6.5 h. Plasma levels then declined with an aver- 2 ical toxicity with one out of six patients at the 450 mg/m level age t1/2 of 10.1 h. No urinary excretion of cytarabine ocfosfate experiencing WHO grade 4 thrombocytopenia and a fall in was noted over the whole dose range. hemoglobin (WHO grade 3). Patient 21 who received the During dose escalation from 50 mg/m2 to 900 mg/m2/day a highest dose of 900 mg/m2 in the course of an intraindividual linear relation between the cytarabine ocfosfate dose and its dose escalation also experienced WHO grade 4 diarrhea as AUC was observed (r = 0.78, (p Ͻ 0.0001), Figure 2). No pla- the limiting toxicity. Hematological toxicity at this level was teauing of the AUC during dose escalation was noted. This predominantly thrombocytopenia (WHO grade 3). finding is consistent with linearity of absorption of the mother drug during dose escalation. Pharmacokinetic results Considerable variability in cytarabine ocfosfate AUC at a given dose was observed: still, interindividual differences in Pharmacokinetic analyses were performed in 28 patients from cytarabine ocfosfate AUC were more pronounced than three centers that agreed to participate in this part of the study intraindividual differences. During intraindividual dose escal- and that provided the required material. Inclusion/exclusion ations, which could be performed in two patients, correlation requirements for the pharmacokinetic part of the study did not coefficients of 0.97 (patient 3: 346 ng·h/ml at 50 mg/m2 dose differ. In two patients (both B-CLL) intraindividual dose escal- level, 1690 at 150 mg/m2, 4635 at 300 mg/m2) and 0.98 ation was performed with patient 3 receiving subsequently (patient 21: 7460 at 450 mg/m2, 12 800 and 13 000 at doses of 50, 150 and 300 mg/m2/day and patient 21 receiving 750 mg/m2, 14 100 at 900 mg/m2) were found (as compared cycles of 450, 600, 750, 900 and 750 mg/day. All other to r = 0.78 for the whole group). The values of these two patients were treated on a single dose level. Pharmacokinetic patients who contributed pharmacokinetic data on several evaluations were performed on all the aforementioned cycles dose levels did not differ significantly from the mean values of of patients 3 and 21 and on all first cycles of the remaining the whole group (Tables 1–3) and therefore did not adversely 26 patients. A total of 34 cycles has therefore been analyzed influence the correlation of YNK01 dose and AUC (Figure 2). for pharmacokinetic purposes. The pharmacokinetic charac- Omitting these two patients results in a moderately higher teristics of all three substances are given in Tables 1–3. correlation (r = 0.86 (P Ͻ 0.0001)). The average plasma concentrations and the respective standard deviations of cytarabine ocfosfate at a dose level of Dose level 50 100 150 200 300 450 600 750 900 600 mg/m2 mg are demonstrated in Figure 1a. This graph clearly depicts the pseudo steady state during the 14 days Cycles 637153531 cycle. No accumulation was noted.

Table 1 Pharmacokinetic parameters of cytarabine ocfosfate (mean and variation coefﬁcient)

2 Dose in mg/m Cycles AUC in ng·h/ml t1/2 in h Lag-time in h Concmax in ng/ml tmax in h Clearanceto t in ml/min

50 6 1099 11.9 1.0 53.8 5.0 1588 0.31 0.27 0.66 0.25 0.41 0.35 100 3 2629 7.2 0.8 154.8 6.8 1288 0.37 0.18 0.76 0.46 0.33 0.36 150 7 2754 9.1 1.1 160.0 5.5 2339 0.63 0.34 0.55 0.59 0.24 0.50 200 1 1975 4.1 0.8 147.0 6.8 2840 — 300 5 10044 12.8 0.9 502.3 6.8 1292 0.41 0.57 0.83 0.73 0.27 0.45 450 3 10767 13.2 1.5 406.7 9.2 1354 0.29 0.35 0.69 0.25 0.72 0.43 600 5 12622 8.1 1.0 603.6 8.0 1736 0.65 0.30 0.04 0.43 0.56 0.83 750 3 16767 8.9 1.0 754.7 7.8 1630 0.40 0.58 0.53 0.33 0.43 0.324 900 1 14100 10.2 0.5 762.3 3.9 2130

Mean 34 — 10.1 1.0 — 6.5 1730 VC 0.43 0.58 0.48 0.52 Cytarabine ocfosfate in ANL and NHL – pharmacokinetics J Braess et al 1621 Table 2 Pharmacokinetic parameters of AraC (mean and variation coefﬁcient) following cytarabine ocfosfate application

2 Dose in mg/m Cycles AUC in ng·h/ml t1/2 in h Concmax in tmax in h Eliminationrenal Clearrenal in ml/min ng/ml in % of dose

50 6 227 21.9 9.1 8.2 1.2 41.2 0.88 0.38 1.17 0.51 0.92 0.61 100 3 446 21.8 6.7 30.6 0.6 30.9 0.91 0.31 0.38 0.31 0.49 1.20 150 7 783 26.1 16.7 25.5 0.6 23.9 0.94 0.47 0.78 0.55 0.84 1.48 200 1 456 33.4 4.2 49.2 2.8 135.7 ————— — 300 5 1560 16.8 31.4 24.8 2.3 70.5 0.61 0.31 0.72 0.35 0.72 0.95 450 3 2838 23.8 36.7 34.7 1.5 77.5 1.14 0.26 1.01 0.35 0.39 0.78 600 5 3496 23.9 68.7 23.4 3.7 101.1 0.76 0.32 0.39 0.66 0.51 0.36 750 3 6600 20.3 132.7 20.2 1.77 22.7 0.57 0.28 0.65 0.18 0.97 0.74 900 1 2250 20.3 61.5 15.2 2.7 145.2 ————— —

Mean 34 — 22.6 — 23.2 1.7 57.2 VC 0.36 0.57 0.90 0.89

Table 3 Pharmacokinetic parameters of AraU (mean and variation coefﬁcient) following cytarabine ocfosfate application

2 Dose in mg/m Cycles AUC in ng·h/ml t1/2 in h Concmax in tmax in h Eliminationrenal Clearrenal in ml/min ng/ml in % of dose

50 6 1767 18.2 31.7 24.1 24.4 86.2 0.84 0.33 0.31 0.41 0.83 0.49 100 3 2909 25.1 78.2 15.3 17.1 94.1 0.90 0.43 1.08 0.45 0.57 0.25 150 6 3279 24.4 79.4 13.0 8.8 66.2 0.56 0.28 0.68 0.55 0.31 0.52 200 1 6350 17.8 288.0 1.0 11.6 48.6 ————— — 300 6 4741 24.3 90.9 17.5 8.4 69.9 0.58 0.46 0.51 0.32 0.75 0.50 450 3 7349 20.8 176.0 14.1 9.7 86.5 0.24 0.18 0.33 0.90 0.39 0.59 600 5 15441 22.2 241.0 29.1 21.6 126.6 0.66 0.31 0.49 0.51 0.60 0.42 750 3 11070 24.5 185.0 21.0 8.8 93.2 0.51 0.33 0.06 0.56 0.84 0.79 900 1 5390 15.7 87.8 31.6 2.8 63.1 ————— —

Mean 34 — 22.3 — 19.2 14.4 85.2 VC 0.33 0.58 0.86 0.52

AraC ocfosfate dose and the AraC AUC during dose escalation (r = 0.68) (eg 227 ng·h/ml at 50 mg/m2 dose level, 783 at AraC is formed from cytarabine ocfosfate through hepatic 150 mg/m2, 2838 at 450 mg/m2). The variability of the AraC metabolism. Maximum AraC plasma levels were reached after AUC (Table 2) at a given dose level of cytarabine ocfosfate an average of 23.2 h. AraC then gradually decreased with an was more pronounced (VC = 0.57) than the respective cytara- average half life of 22.6 h. AraC levels reached a pseudo ste- bine ocfosfate AUC (Table 1). This ﬁnding is consistent with ady state as depicted in Figure 1b for the 600 mg/m2 dose additive causes of variability during the course of absorption group (mean and standard deviation). Renal excretion of AraC and metabolism of cytarabine ocfosfate. As a consequence, was also measured. An average of 1.7% (on a molar basis) of variability of AraC plasma concentrations after YNK01 is mod- the total given dose of cytarabine ocfosfate was excreted as erately higher than during continuous i.v. AraC applications AraC via the kidneys. A renal clearance of 57.3 ml/min was (VC = 0.40–0.46 (standard- and intermediate-dose AraC), derived from these data. unpublished results). A linear relation was observed between the cytarabine Cytarabine ocfosfate in ANL and NHL – pharmacokinetics J Braess et al 1622

Figure 1 (a) Plasma concentration curve (mean and standard deviation) of cytarabine ocfosfate at the 600 mg/m2 dose level (n = 5). (b) Plasma concentration curve (mean and standard deviation) of AraC at the 600 mg/m2 dose level (n = 5). (c) Plasma concentration curve (mean and standard deviation) of AraU at the 600 mg/m2 dose level (n = 5). Cytarabine ocfosfate in ANL and NHL – pharmacokinetics J Braess et al 1623

Figure 2 Correlation of YNK01 AUC and YNK01 dose.

AraU Discussion

The main metabolite of AraC uracil arabinoside (AraU) fea- Cytarabine ocfosfate is a recently developed lipophilic ana- tured a t of 19.2 h. Plasma levels of AraU declined with an max logue of cytosine arabinoside intended for oral administration. average half-life of 22.3 h. Renal excretion of AraU occurred Reports of a high antineoplastic activity in preclinical trials8 at an average of 14.4% of the total given dose of cytarabine and preliminary data suggesting favorable pharmacokinetic ocfosfate. A renal clearance of 85.2 ml/min was calculated properties10 as compared to the mother compound have from these data. raised interest about the clinical potential of this drug. Avail- During dose escalation a linear relation between the cytara- able clinical data consist of studies with a limited number of bine ocfosfate dose and the AraU AUC was observed cases12,13 and case reports mostly on patients with AML or (r = 0.61) (eg 1767 ng·h/ml at 50 mg/m2 dose level, 3279 at myelodysplastic syndromes.14–28 150 mg/m2, 15441 at 600 mg/m2). Still, the variability of the Within the limits of a phase I/II setting clinical efficacy AraU AUC at a given dose level of cytarabine ocfosfate was could also be demonstrated in this trial despite the fact that even more pronounced than in the case of AraC. patients were heavily pretreated/refractory and/or elderly and The sum of the AraC AUC and the AraU AUC constituted not eligible for more intensive therapy. In the AML study two the metabolic product of cytarabine ocfosfate. A linear relapsed and prognostically unfavorable patients achieved a relationship between this sum of AraC AUC and AraU AUC CR with further patients reaching a PR (four) or SD (three) and cytarabine ocfosfate AUC was observed (r = 0.72 respectively. In the NHL group no CRs were reached but two (P Ͻ 0.0001), Figure 3). This finding is consistent with linearity long lasting PRs were noteworthy with one patient, who had of metabolism of cytarabine ocfosfate following absorption of previously proved to be refractory to chlorambucile/ the drug. prednisone, radiation therapy, mitoxantrone/prednimustine and fludarabine maintaining a PR which now exceeds 2 years.29 SD was maintained in a further six patients. Dose lim- Bioavailability iting toxicity was encountered at the seventh dose escalation level (600 mg/m2) with two out of six patients going off study Usually bioavailability of a drug following oral administration due to grades 3 and 4 diarrhea. Another patient suffered from is determined by measuring its AUC after intravenous vs oral grade 3 diarrhea at the 900 mg/m2 dose level which was administration. In the case of cytarabine ocfosfate this reached in the course of an intraindividual dose escalation. approach could not be followed due to intravascular hemo- Symptoms disappeared after dose reduction to 750 mg/m2. lysis at the injection site after i.v. administration in toxicology Other toxicities such as nausea/vomiting or stomatitis did not studies in rodents. As a substitute the ‘metabolic bioavailabili- exceed WHO grades 1 and 2. Thrombocytopenia was the ty’ was determined by measuring the renal excretion of AraC main hematologic toxicity with one patient at the 450 and and AraU: since more than 90% of cytarabine ocfosfate that another at the 600 mg/m2 dose level experiencing WHO is absorbed after oral administration is metabolized to AraC grade 4 toxicity. and AraU which are almost completely excreted via the kid- Even though clinical efficacy and toxicities needed to be neys measurement of renal excretion allows calculation of the established for the further development of YNK01 it is neces- metabolic bioavailability. Using this concept, bioavailability sary to stress that the prodrug itself is not active (unpublished of cytarabine ocfosfate amounted to an average of 15.8% (on results) but that its cytotoxicity is mediated by AraC. Instead, a molar basis). the novel aspects of YNK01 pharmacology are its pharmacoki- Cytarabine ocfosfate in ANL and NHL – pharmacokinetics J Braess et al 1624

Figure 3 Correlation of the sum of AraC and AraU AUC and YNK01 AUC.

netics which are of particular interest because its molecular YKN01, uptake into hepatocytes, oxydative metabolism to structure attempted to improve on several limitations of the AraC and final release into the circulation, maximum AraC mother compound AraC. Hence, the stearate chain renders concentrations are reached late after 23.2 h. The half-life of the molecule resistant to enzymatic deamination AraC (22.6 h) following YKN01 application is significantly (=inactivation) by cytidine deaminase which otherwise causes longer than following i.v. or s.c. application of AraC itself a rapid inactivation of AraC within the systemic circu- (about 1 and 2–4 h, respectively). Nevertheless, the mean resi- lation.Furthermore, instead of being a hydrophilic nucleoside dence time of an individual AraC molecule is independent like AraC, the cytarabine ocfosfate molecule has distinct lipo- regardless of whether application was via oral YNK01, i.v. philic properties due to the addition of the fatty acid chain. AraC infusion or s.c. AraC injection. Instead, the prolonged As a consequence, cytarabine ocfosfate was expected to be persistence of AraC plasma levels following YNK01 can be quantitatively absorbed following oral administration. In fact, explained by the existence of a pool of AraC precursors in the data of the present study show that cytarabine ocfosfate hepatocytes which is gradually metabolized. Flip-flop kinetics was absorbed with a lag-time of 1.0 h and reached maximum are the explanation for this unusual pharmacokinetic behav- plasma levels after 6.5 h. The rather long lag-time and the late iour: for most agents it can be assumed that the drug is pro-

tmax are in accordance with ﬁndings in animal studies, which vided to the organism in a short period of time (eg i.v. appli- demonstrated that radioactively labelled YNK01 was mainly cation, rapid absorption after oral application). In such a case,

absorbed in the distal part of the small intestine (terminal t1/2 represents the sum of all processes of elimination (eg distri- ileum). During dose escalation from 50 to 900 mg/m2 a linear bution, renal or hepatic excretion, metabolism). Given the

relationship was found between YNK01 dose and AUC (Figure case that absorption is distinctly slower than elimination, t1/2 2) demonstrating that absorption of the prodrug is not satu- represents the rate limiting process of absorption. In the case

rated in the tested dose range. This phenomenon is most likely of cytarabine ocfosfate, t1/2 does therefore not represent the caused by the structural properties of YNK01 that feature a elimination of AraC from the systemic circulation but the pre- hydrophilic ‘head’ (nucleoside) and a hydrophobic tail ceding events of intestinal absorption of cytarabine ocfosfate, (stearate chain) resulting in micell formation and passive its uptake into hepatocytes, the subsequent oxydative metab- absorption. An average of 15.8% of the total dose was olism to AraC and the final release of the nucleoside into the absorbed and metabolized to AraC and AraU which consti- systemic circulation. This phenomenon also explains the vir- tutes the metabolic bioavailability of YNK01. Biliary excretion tually identical half-lives of AraC and AraU (22.6 h and of unchanged YNK01 was not tested in our patients but prior 22.3 h, respectively). Even though a further step (deamination) animal studies showed that only minute amounts (1–2% of is needed for the production of AraU this rapid process does the dose) were excreted unchangedly into the bile. not significantly alter the resulting half-life of AraU. Similarly, no renal excretion of YNK01 was noted. This was The AUCs of YNK01 and of AraC following YNK01 admin- to be expected due to the lipophilic properties of the prodrug istration show considerable interindividual differences with that presumably allow complete tubular reabsorption follow- variation coefficients ranging from 0.31 to 0.67 for YNK01 ing glomerular filtration. The half-life of 10.1 h for YNK01, and from 0.57 to 1.14 for AraC at the nine tested dose levels. therefore, mainly represents the uptake of the drug into hepa- The most likely cause for these variabilities is the cytochrome tocytes and its subsequent metabolism to AraC and AraU. In p450-dependent oxidative metabolism of YNK01. Several contrast to AraC, deamination is of no quantitative relevance antineoplastic agents such as epidophyllotoxins, ifosfamide, to YNK01 half-life. In vitro studies have shown that YNK01 is tamoxifen, paclitaxel and vinca alkaloids have been shown to no substrate for cytidine deaminase (unpublished data). Due be metabolized (and partially activated) by this enzyme sys- to the additive processes of intestinal passage, absorption of tem.30 Due to large interindividual differences in enzyme Cytarabine ocfosfate in ANL and NHL – pharmacokinetics J Braess et al 1625 acticities – partly determined by ethnic origin,31 partly ine arabinoside and mitoxantrone in patients with refractory acute induced by other xenobiotics – wide variations in cytotoxic leukemias. Results of a clinical phase II study. Cancer 1997; 79: drug metabolism are observed, eg with a factor of 4.4 between 59–68. 32 5 Breithaupt H, Pralle H, Eckhardt T, von Hattingberg M, Schick J, high and low metabolizers of vinblastin and of 8.9 for docet- Loffler H. Clinical results and pharmacokinetics of high-dose cyto- 33 axel. Different rates of cytochrome-dependent metabolism sine arabinoside (HD ARA-C). Cancer 1982; 50: 1248–1257. may contribute to wide variations in drug plasma levels as 6 Capizzi RL, White JC, Powell BL, Perrino F. Effect of dose on the exemplified by the anthrachinone mitoxantrone, also a sub- pharmacokinetic and pharmacodynamic effects of cytarabine. strate of cytochrome p450,34 which exhibits a factor of 13 Semin Hematol 1991; 28: 54–69. between the lowest and highest AUC following body-surface 7 Saneyoshi M, Moruzumi M, Kodama K, Machida K, Kuminaka A, 35 Yoshino H, Synthetic nucleosides and nucleotides. XVI. Synthesis adjusted i.v. drug application. Mitoxantrone, therefore, is an and biological evaluation of a series of 1–beta-d-arabinofuranosyl- example of a high variability in plasma levels following i.v. cytosine-5Ј-alkyl or arylphosphates. Chemical Pharm Bull 1980; administration which even exceeds the variations observed in 28: 2915–2923. other, orally applied, drugs such as YNK01. 8 Kodama K, Moruzumi M, Saitoh K, Kuninaka A, Yoshino A, Sane- Due to the long t of AraC after oral cytarabine ocfosfate yoshi M. Antitumor activity and pharmacology of 1-beta-d-arabin- 1/2 Ј application long-lasting plasma levels of AraC are maintained ofuranosyl-5 -stearylphosphate: an orally active derivate of 1-beta- d mimicking a continuous i.v. infusion. When daily doses -arabinofuranosylcytosine. Jpn J Cancer Res 1989; 80: 679–685. 2 9 Yoshida Y, Yamada Y, Watanabe T, Suga T, Takayama H. Partici- between 100 and 200 mg/m of cytarabine ocfosfate were pation of the peroxisomal beta-oxidation system in the chain short- applied the resulting AraC plasma concentrations were equiv- ening of PCA16, a metabolite of the cytosine arabinoside prodrug alent to those achieved during low-dose (20 mg/m2 s.c.) AraC YNK01 in rat liver. Biochem Pharmacol 1990; 39: 1505–1512. therapy. In the 450 mg/m2 to 600 mg/m2 dose range of cytara- 10 Schleyer E, Braess J, Ramsauer B, Unterhalt M, Kaufmann C, Wilde bine ocfosfate AraC plasma concentrations were in the range S, Schussler M, Hiddemann W. Pharmacokinetics of Ara-CMP- or even exceeded those plasma concentrations achieved dur- Stearate (YNK01): phase I study of the oral Ara-C derivative. 2 Leukemia 1995; 9: 1085–1090. ing continuous infusion of standard-dose AraC (100 mg/m ) 11 Heinzel G, Bozler G, Van Rossum J (eds). Pharmacokinetics dur- such as in the TAD-9 regimen. During i.v. administration of ing Drug Development: Data Analysis and Evaluation Techniques. conventional AraC moderate variations in the AUC are Curve Fitting Programs, 1982, pp. 207–208. observed with vc ranging from 0.40 (standard-dose AraC) to 12 Tatsumi N, Yamada K, Ohshima T, Nakamura T, Ohno R, 0.46 (intermediate-dose AraC). In comparison, variability fol- Masaoka T, Kimura I, Kimura K. Phase II study of YNK01 (1-beta- Ј lowing oral YNK01 was only slightly higher with a vc of 0.57, d-arabinofuranosylcytosine-5 -stearylphosphate) on hematolog- most likely reflecting the additional causes of variability such ical malignancies. Gan To Kagaku Ryoho 1990; 17: 2387–2395. 13 Fukuoka M, Miyazaki T, Yoshida Y, Hattori M, Niitani H, Naka- as resorption and hepatic metabolism. mura T, Hirota Y, Ohta K, Tanaka M, Suzuoki Y et al. Phase I In conclusion, cytarabine ocfosfate represents an interesting study of YNK01 (1-beta-d-arabinofuranosylcytosine-5Ј- prodrug of AraC which has been demonstrated to have clini- stearylphosphate). Gan To Kagaku Ryoho 1990; 17; 2213–2219. cal efficacy in poor-risk patients with AML and NHL. Side- 14 Kiyama Y, Suzuki G, Masauzi N, Ohizumi H, Kobayashi N, Ogas- effects were manageable and mainly consisted of diarrhea and awara M, Naohara T, Saito M, Higa T, Kasai M. Treatment of (AraC typical) hematotoxicity. The drug possesses attractive refractory hematologic malignancies by combination of cytarabine ocfosfate and etoposide. Gan To Kagaku Ryoho 1996; 23: pharmacokinetic features comprising oral applicability and 1717–1720. long-lasting plasma levels of AraC. Its unique pharmacokinetic 15 Nishi T, Funasako M, Hata N, Andoh K, Itoh Y, Ono H, Itoh K, properties lead to a prolonged systemic exposure of AraC Sato Y, Sato H, Taniji M, Nakai S, Ueda A, Fujimoto H, Ohata M. mimicking a continuous i.v. infusion. This prolonged exposure Five cases of myelodysplastic syndrome treated with cytarabine allows maximization of the probability of malignant cells ocfosfate. Gan To Kagaku Ryoho 1996; 23: 1331–1334. passing through the vulnerable phase of the cell cycle and 16 Aoki S, Koyama S, Goto T, Takahashi H, Shibata A. Continuation might therefore contribute to overcoming drug resistance. of complete remission by oral administration of cytarabine ocfosfate in a patient with M0, who achieved remission by small doses Plasma levels such as in low-dose or standard-dose AraC ther- of cytosine arabinoside with G-CSF. Rinsho Ketsueki 1995; 36: apy can be maintained during the whole treatment cycle. 40–44. Phase II studies are now needed to establish the clinical 17 Kadowaki I, Sasaki O, Sasaki T, Ishizawa K, Kimura J, Nomura J, efficacy of this drug and to define its most appropriate Furuyama K, Harigae H, Shishido T, Okuda M et al. Complete indications. remission by cytarabine ocfosfate plus G-CSF therapy in a patient with hypoplastic RAEB-T. Gan To Kagaku Ryoho 1995; 22: 145–147. 18 Hamaoka R, Jozaki K, Amano T, Itoh H, Imai Y, Nishikawa M, References Kurokawa M, Yonezawa T, Chinen Y. Low-dose cytarabine ocfosfate therapy in an elderly acute myelogenous leukemia. Gan To 1 Bishop JF, Matthews JP, Young GA, Szer J, Gillett A, Joshua D, Kagaku Ryoho 1995; 22: 819–822. Bradstock K, Enno A, Wolf MM, Fox R et al. A randomized study 19 Wake A, Takazawa A, Serino Y, Tonai S, Nakanishi M, Murakami of high-dose cytarabine in induction in acute myeloid leukemia S, Ogawa R, Nagata K, Mori N, Nakata K et al. Successful treat- (see comments). Blood 1996; 87: 1710–1717. ment of acute myelomonocytic leukemia developed from MDS 2 Mayer RJ, Davis RB, Schiffer CA, Berg DT, Powell BL, Schulman with cytarabine ocfosfate (SPAC). Gan To Kagaku Ryoho 1995; P, Omura GA, Moore JO, McIntyre OR, Frei E III. Intensive postre- 22: 395–398. mission chemotherapy in adults with acute myeloid leukemia. 20 Takeuchi M, Kojima K, Takaba S, Isokawa M, Tanimizu M, Kimura Cancer and Leukemia Group B (see comments) New Engl J Med F, Ohmoto E, Harada M. Successful treatment of refractory anemia 1994; 331: 896–903. with excess of blasts in transformation with cytarabine ocfosfate. 3 Wiernik PH, Banks PL, Case DC Jr, Arlin ZA, Periman PO, Todd Gan To Kagaku Ryoho 1995; 22: 141–144. MB, Ritch PS, Enck RE, Weitberg AB. Cytarabine plus idarubicin 21 Masumoto A, Ohbayashi Y, Kawada H, Arimoni K, Sasao T, or daunorubicin as induction and consolidaztion therapy for pre- Ogawa Y, Fujiwara H, Fukuda R, Watanabe S, Umeda Y et al. viously untreated adult patients with acute myeloid leukemia. Effects of cytarabine ocfosfate on colony-stimulating factor in mye- Blood 1992; 79: 313–319. lodysplastic syndrome with monosomy 7. Rinsho Ketsueki 1994; 4 Kern W, Schleyer E, Unterhalt M, Wormann B, Buchner T, Hidde- 35: 603–608. mann W. High antileukemic activity of sequential high dose cytos- 22 Shikoshi K, Niitsu N, Takada M, Umeda M. Therapy-related acute Cytarabine ocfosfate in ANL and NHL – pharmacokinetics J Braess et al 1626 non-lymphocytic leukemia (M2) with 7;11 chromosome translo- araC, stearyl-ara-CMP (YNK01). Gan No Rinsho 1990; 36: cation induced into complete remission by low dose cytosine ara- 2204–2209. binoside and cytarabine ocfosfate therapy. Nippon Ronen Igakkai 29 Braess J, Kern W, Unterhalt M, Kaufmann CC, Ramsauer B, Zasshi 1994; 31: 468–471. Schussler M, Kaeser Frohlich A, Hiddemann W, Schleyer E. 23 Okuyama Y, Kitazume K, Iwabe K, Usuki K, Urabe A. Hypoplastic Response to cytarabine ocfosfate (YNK01) in a patient with leukemia successfully treated by oral administration with cytarab- chronic lymphocytic leukemia refractory to treatment with ine ocfosfate. Rinsho Ketsueki 1994; 35: 593–597. chlorambucil/prednisone, fludarabine, and prednimustine/ 24 Arai T, Narimiya S, Hayakawa K, Takaya T, Toshima S, Araki M, mitoxantrone. Ann Hematol 1996; 73: 201–204. Uchida A, Yoshimi N, Shibuya C, Yasuda Y. Successful treatment 30 Kivisto KT, Kroemer HK, Eichelbaum M. The role of humancytoch- of myelodysplastic syndrome with cytarabine ocfosfate. Gan To rome P450 enzymes in the metabolism of anticancer agents: Kagaku Ryoho 1994; 21: 1085–1088. implications for drug interactions. Br J Clin Pharmacol 1995; 40: 25 Ohno R, Tatsumi N, Hirano M, Imai K, Mizoguchi H, Nakamura 523–530. T, Kosaka M, Takatsuki K, Yamaya T, Toyama K et al. Treatment 31 Relling RV. Polymorphic drug metabolism. Clin Pharm 1989; 8: of myelodysplastic syndromes with orally administered 1-beta-d- 852–863. arabinofuranosylcytosine-5Ј-stearylphosphate. Oncology 1991; 32 Zhou Pan XR, Seree E, Zhou XJ, Placidi M, Maurel P, Barra Y, 48: 451–455. Rahmani R. Involvement of human liver cytochrome P450 3A in 26 Inaba T, Shimazaki C, Tatsumi T, Yamagata N, Hirata T, Goto H, vinblastine metabolism: drug interactions. Cancer Res 1993; 53: Fujita N, Nakagawa M, Miyazaki S et al. Successful treatment of 5121–5126. acute myelogenous leukemia in an elderly patient with cytarabine 33 Marre F, Sanderink GJ, de Sousa G, Gaillard C, Martinet M, Rah- ocfosfate. Gan To Kagaku Ryoho 1994; 21: 535–538. mani R. Hepatic biotransformation of docetaxel (Taxotere) in vitro: 27 Ueda T, Kamiya K, Urasaki Y, Wataya S, Kawai Y, Tsutani H, involvement of the CYP3A subfamily in humans. Cancer Res Sugiyama M, Nakamura T. Clinical pharmacology of 1-beta-d-ara- 1996; 56: 1296–1302. binofuranosylcytosine-5Ј-stearylphosphate, an orally administered 34 Mewes K, Blanz J, Ehninger G, Gebhardt R, Zeller KP. Cytochrome long-acting derivative of low-dose 1-beta-d-arabinofuranosylcyto- P-450-induced cytotoxicity of mitoxantrone by formation of elec- sine. Cancer Res 1994; 54: 109–113. trophilic intermediates. Cancer Res 1993; 53: 5135–5142. 28 Nishikawa M, Morita K, Komada F, Uemura Y, Kageyama S, Min- 35 Schleyer E, Kamischke A, Kaufmann CC, Unterhalt M, Hiddemann ami N, Deguchi K, Shirakawa S. A case of complete remission W. New aspects of the pharmacokinetics of mitoxantrone and its from acute unclassified leukemia achieved by using a prodrug of two major metabolites. Leukemia 1994; 8: 435–440.