Sequential Oral Hydroxyurea and Intravenous Cytosine Arabinoside in Refractory Childhood Acute Leukemia: A Pediatric Oncology Group Phase I Study

Summary

At concentrations >0.1 mM, Hydroxyurea (HU) enhances the accumulation of cytosine arabinoside (ara-C) in leukemia cells in vitro. This study of children with refractory acute leukemia was designed to take advantage of this biochemical modulation. A fixed dose of HU and an escalating dose of ara-C were used. Oral HU, 1200 mg/m² was followed 2 hours later by ara-C, 250-3100 mg/m² intravenously in 15 minutes. The combination was given on days 1,2,3 and 8,9,10. Thirty-three children (26 ALL, 7 ANLL) were treated; 29 received at least one full course. All patients developed grade 4 cytopenias. Other grade 3-4 toxicities included: hyperbilirubinemia (2), elevated transaminases (3), transient gait disturbance (1), stomatitis (3), typhlitis (1), nausea/vomiting (9) and marrow aplasia >4 weeks (1). Three patients had intracranial bleeds while thrombocytopenic. Only liver toxicities and nausea/vomiting exhibited any dosage effect. The maximum tolerated dose of ara-C was 2400 mg/m². There were 6 complete responses (5ALL), 5 partial responses (3 ALL), and 19 patients with no response or progressive disease. There was no dosage effect for response with 2 complete responses occurring at the lowest ara-C level. Responses were transient (1-3 months). 20/26 patients achieved a peak serum HU level >0.5 mM by 2 hours after the HU dose. The mean level at 2 hours was 0.57 mM (range 0.21-0.99 mM). This combination of HU and ara-C is tolerable and has efficacy in refractory leukemias. Responses at the lowest ara-C dose level suggests synergism.

INTRODUCTION

Cytosine arabinoside (ara-C) is an active drug for the treatment of acute leukemias. The activity of this S-phase agent depends upon the extent of its uptake and biochemical conversion to the active nucleotide, ara-CTP, as well as the cell cycle profile of the target population at the time of drug exposure. The current study was based on two premises: 1. pre-exposure to hydroxyurea (HU) would lead to enhancement of ara-CTP formation and 2. each dosing of HU/ara-C would favorably affect the cell cycle kinetics of the leukemia cells (via synchronization and recruitment) leading to enhanced cytotoxicity with the next dosing. This phase I study had three objectives: 1. to determine the maximum tolerated dose of ara-C when given 2 hours following a dose of HU on days 1,2,3 and 8,9,10; 2.to determine the pharmacokinetics of HU in this regimen and 3. to obtain preliminary evidence for anti-leukemic effects of this drug combination.

MATERIALS AND METHODS

Patient selection

Thirty-three heavily pretreated children with refractory acute leukemia (26 acute lymphocytic leukemia [ALL], 7 acute non-lymphocytic leukemia [ANLL]; 27 with 2 or more relapses) entered this study. There were 16 males and 17 females. Median age was 8 (range 2-20) years. Eligibility requirement included: >25% blasts in the marrow, no uncontrolled serious infection, adequate nutritional status, total bilirubin <1.7 mg/dl, and creatinine <1.2 mg/dl. No patient had been on therapy (except steroids) for at least one week. Supportive care (transfusions, infection prophylaxis, other) was at the discretion of individual institutions. Written informed consent was obtained as per local institutional guidelines.

Therapy

Patients were allowed only clear liquids on each morning of treatment. HU, 1200 mg/m², was given orally at time zero. Dissolving the drug in juice or punch was permitted. Ara-C was administered intravenously over 15 minutes starting 2 hours after the HU. Each course of treatment consisted of the above sequence on days 1,2,3 and 8,9,10. Responders could have additional courses every 21 days (or longer if required for recovery from toxicities). Three to five patients were treated at each ara-C dose level from 250-2400 mg/m². Two patients received 3100 mg/m². Twenty-nine patients who received at least one full course and were evaluable for determination of the maximum tolerated dose (MTD).

Toxicities

were graded according to standard NCI criteria as used by the Pediatric Oncology Group. Data from all patients were used for determination of dose-limiting toxicities.

Response definitions

Complete response included < 5% marrow blasts with normal hematopoiesis, transfusion independence and absence of signs/symptoms/physical findings of leukemia elsewhere in the body. Partial response required 6-25 % marrow blasts. No response or progressive disease required > 25% marrow blasts with increasing blasts in the blood or marrow.

Pharmacologic studies

Blood samples were obtained at baseline and 30, 60, 90 120, 180, 240 and 360 minutes after HU administration. Serum was frozen and sent on dry ice by overnight mail to a central laboratory. 0.5 ml of serum was mixed, on ice, with 1.0 ml of 1M perchloric acid for 20 minutes and spun 2500 RPM for 15 minutes. The supernatant was processed in a colorimetric assay previously described¹ based on the method of Belt, et al.² The HU half-life for individual patients was calculated by fitting a single exponential to the decaying portion of each patient's serum concentration curve using PCNONLIN (Statistical Consultants).

RESULTS

Toxicities

Toxicities are summarized by type and dose levels in Table 1.

Table 1.

Toxicilies by ana C dose level (mg/m²). For each organ system, each number indicates the degree of toxicity for a single patient.

Dose	# Courses	GI	Liver	Neurologic	Skin	Mucosa	Infection	Hematologic
250	6	2,2	3	2,4,4,4	1	-	2,4	4,4,4,4,4,5
500	3	2	-	2	-	3	2	all
750	4	-	-	-	-	-	2,4	all
1000	3	3	-	-	-	-	-	all
1350	9	3,3,3	2	-	-	1,3	-	all
1800	7	1,2	3	2	-	-	3	all
2400	9	3,3,3	3,4	-	-	2.3	3,3,3	all
3100	2	1,2	3	4	-	-	5	-

Hematologic

All patients who received a full course of treatment became pancytopenic (platelets < 25,000/ul, neutrophil count < 500/ul). Three patients (all treated with 250 mg/m² of ara-C) had subarachnoid bleeding while severely thrombocytopenic. One of these patients, refractory to platelet transfusions prior to starting treatment, died on day 24 with non-responding disease.

Dermatologic

One patient developed dry desquamation. He had previously received total body irradiation.

Gastrointestinal

Nausea and vomiting tended to be worse with higher ara-C doses. They were usually less severe after the first day of treatment. One patient developed diarrhea. Another patient died with probable typhlitis on day 16; he was receiving treatment for gram negative septicemia at the start of chemotherapy. Three patients (treated with 500, 1350 and 2400 mg/m² of ara-C) developed grade 3 stomatitis.

Hepatic

Three patients had transient transaminase elevations. One patient had transient hyperbilirubinemia (maximal total bilirubin 2.0 mg/dl). A second patient died at home on day 19 with a total bilirubin of 6.9 mg/dl and normal transaminases. Septicemia was suspected but blood cultures and an autopsy were not performed.

Neurologic

One patient (treated with 1800 mg/m² of ara-C) had a transient gait disturbance. The three patients with subarachnoid bleeds are discussed above.

Early deaths (prior to day 22)

the three deaths from subarachnoid hemorrhage, typhlitis (probable) and sepsis (probable) are discussed above.

Pharmacokinetics

Data is available for 26 patients, 19 of whom had a complete set of data points. Twenty patients achieved a peak HU level > 0.5mM by 2 hours. The mean peak level was 0.81 (range 0.27-1.50 ) mM. All but 2 patients achieved their peak level by 30 to 120 minutes after HU; the two outliers peaked at 180 minutes. The mean level at 2 hours was 0.57 (range 0.21-0.99) mM. There was no correlation of peak levels of HU with responses. The patient with the lowest peak level achieved a complete remission.

The serum half-life was calculated for each patient separately. The mean half-life was 146 +/− 9 (SE) minutes (median 141 minutes, range 94-264 minutes). Serum levels are summarized in Table 2.

Table 2.

Hydroxyurea serum levels (in mM)

Time	Mean	S.D.	Range
30 min	.48	.41	<.05 - 1.50
60 min	.67	.35	.17 - 1.37
90 min	.61	.26	.24 - 1.12
2 hr	.57	.20	.21 - 0.99
3 hr	.47	.15	.16 - 0.76
4 hr	.34	.13	.15 - 0.63
6 hr	.20	.09	.07 - 0.51

Response

Responses by ara-C dose level and diagnosis are summarized in Tables 3 and 4. Six patients (5 ALL) had a complete response and five (3 ALL) had a partial response. One patient with myelomonocytic ANLL had marrow hypoplasia lasting longer than 4 weeks; this patient had had experienced prolonged marrow suppression after conventional high dose ara-C.

Table 3.

Response by Ara-C dose

Dose	CR	PR	NR	PD	Early Death	Total #
250	2	0	2	1	0	5
500	0	0	2	1	0	3
750	1	0	2	0	0	3
1000	0	0	3	0	0	3
1350	2	0	3	0	0	5
1800	1	2	2	1	0	6
240	0	3	2	0	1	6
3100	0	0	0	0	2	2
	--	--	--	--	--	--
Total #	6	5	16	3	3	33

Tab1e 4.

Response by diagnosis

Diagnosis	CR	PR	NR	PD	Early Death	Total #
non-T ALL	3	2	11	2	3	21
T-cell ALL	2	1	1	1	0	5
ANLL	1	2*	4	0	0	7

^*One of the PR's with ANLL achieved an MI marrow, but had residual Auer rods.

Duration of response was noted for the six children with complete remission. One patient was removed from study after one course because of a subarachnoid hemorrhage. A second patient with subarachnoid hemorrhage expired on day 43 in complete remission after two courses of treatment. A third patient was taken off study in complete remission after one month because of a transient gait disturbance. Two patients relapsed after receiving a second course of treatment. A final patient had an M2 marrow after a third course and frank relapse after the fourth course of ara-C.

DISCUSSION

Hydroxyurea, an inhibitor of ribonucleotide reductase, has been in clinical use for more than 20 years. Phase I/II trials performed in the early 1960's using daily schedules demonstrated only mild activity in childhood leukemias as measured by objective responses.^3,4 Despite the lack of significant single agent activity for remission induction, HU has been successfully used to lower blast counts in CML blast crisis⁵ and acute leukemia.⁶

The pharmacokinetics of intravenous and oral HU have been studied in adults with solid tumors.² The mean plasma half-lives of HU following discontinuation of a 72 hour continuous infusion (2-3 mg/m²/min) ranged from 220 to 267 minutes. Plasma levels after single oral doses of 500 and 800 mg/m² peaked from 30 to 120 minutes after the dose, with mean peak plasma concentrations of 0.54 and 0.80 mM after the two doses respectively. The children in our study achieved a mean peak serum level of 0.81mM after a dose of 1200 mg/m² and demonstrated a shorter half-life (mean 146 minutes) than in adults. This may represent differences in metabolism in the two populations. Metabolism of HU is strongly suggested by the finding that only 26-61 % of the total IV dose was recoverable in the urine of adult patients.² Other studies also have shown a great variability in the amount of untransformed drug recovered in urine.⁷

High dose ara-C is an effective remission induction agent for adults with relapsed ANLL. Complete remission rates as high as 61 % have been achieved using 3 gm/m² over 2 hours, every 12 hours for 12 doses.⁸ Capizzi achieved complete remission in 45 % of children with relapsed leukemias using 3 gm/m² over 3 hours, every 12 hours for 4 doses followed by asparaginase.⁹

Several agents facilitate an increase in intracellular levels of ara-CTP. Walsh demonstrated that simultaneous 4 hour intravenous infusions of HU and ara-C in L1210-bearing mice produced a two-fold increase in ara-CTP in the tumor cells when compared to ara-C infusion alone.¹² No increase in ara-CTP content was seen in brain, ileum or spleen cells. Howell reported that a 4 hour exposure to 0.1, 1.0 and 10 mM HU caused a tripling in ara-CTP levels in HL-60 cells compared to controls with an associated 69 % decrease in the dCTP pool in the 1.0 mM HU-treated cells.¹³ Tanaka reported a four-fold increase in ara-CTP, with a 50 % reduction in dCTP, in CCRF-CEM cells following 1 mM HU.¹⁴ Rauscher reported that HU induced enhancement of ara-CTP formation and cytotoxicity, associated with decreased intracellular dCTP, in both HL-60 and L1210 cell lines.¹⁵ He emphasized the importance of cytokinetic ad well as biochemical modulatory effects in the interaction of the two drugs. Decrease in intracellular dCTP may not always be associated with enhanced ara-CTP formation. Grant showed synergism between HU and ara-C without significant change in ara-CTP levels using a highly ara-C resistant, deoxycytidine kinase deficient HL-60 cell line.¹⁶

In addition to this biochemical interaction, both HU and ara-C have individual cell kinetic effects which are relevant to clinical scheduling. Shackney reported the effects of sequential one hour exposures to either of these drugs.¹⁷ The optimal time for a second dose was 8-12 hours, coinciding with the maximal recruitment into S-phase as measured by flow cytometry. Rauscher reported on the clonogenic survival of L1210 and HL-60 cells after a variable HU exposure followed by a one hour exposure to ara-C.¹⁵ Maximal cell kill correlated with minimal intracellular dCTP, maximal intracellular ara-CTP and maximal per cent of cells in S-phase at the start of ara-C exposure. Smets described cell kinetic changes of bone marrow blasts in children with ANLL who were give a single 500-1000 mg/m² intravenous bolus dose of ara-C.¹⁸ Recruitment into S-phase occurred 17-36 hours after the ara-C. The time to maximal recruitment was inversely related to the pretreatment S-phase percentage.

In our study, a single oral dose of HU produced serum levels that are adequate for enhancement of ara-C uptake by leukemia cells in vitro. It was not possible to evaluate cell kinetic perturbation or intracellular ara-CTP levels due to refusal of most families to allow performance of serial bone marrow aspirations.

Stucky reported a trial of HU combined with ara-C in adults with ANLL.¹⁹ HU, 800 mg/m², was given every 4 hours for 16 hours. Eight hours after the last dose (when HU levels were undetectable) ara-C (100 mg/m²) was given as an intravenous bolus. The cycle was repeated daily until marrow blasts were cleared. Twelve of 21 newly diagnosed patients younger than 50 years of age achieved complete remission. The mean treatment time was 11 days and the median time to remission was 31 days. Toxicity was mild. The regimen was designed using a cell kinetic model and did not take advantage of biochemical interactions as described above.

Howell used oral HU (375-1780 mg/m² every 6 hours) and ara-C (100 mg/m² by continuous intravenous infusion for 5 days) in patients with relapsed leukemias and lymphomas.¹³ All patients with leukemia had a decrease in their blast counts. Complete responses were obtained in 1/5 with NHL, 1/2 with CML blast crisis, 0/2 with AML and 0/8 with ALL (6 adults, 2 children). The main toxicities were skin rash and marrow suppression. This regimen was designed for biochemical enhancement without kinetic considerations.

Zittoun gave oral HU (3 grams, days 1 and 4) and ara-C (200 mg/m² by continuous intravenous infusion, days 2,3 and 5,6) to 13 adults (9 previously untreated) with ANLL.²⁰ Three achieved a complete response and 3 a partial remission. Schilsky used oral HU (500 mg every 6 hours for 4 doses).²¹ Twelve hours later, ara-C (100 mg/m² by continuous intravenous infusion for 3 days) was begun. Additional oral HU (500 mg/m² every 4 hours) was given concomitant with the ara-C. Four complete and 5 partial responses were seen in 25 adults with refractory NHL. The primary toxicities were myelosuppression, emesis and rash (20%).

Responses were seen at all ara-C dose levels in our study. This may reflect an ability of hydroxyurea to maximally upregulate intracellular ara-CTP even at low doses of ara-C. This could be evaluated by sequential sampling of marrow or blood leukemic cells during drug administration. The lack of a dose response, however, does suggest an ara-C/hydroxurea interaction rather than a response to ara-C alone.

The HU/ara-C regimen described in this report was tolerable. Although a clear dose-limiting toxicity was not delineated, there were two early deaths at 3100 mg/m² and responses were seen even at the lowest ara-C dose level. Therefore we recommend that an ara-C dose level of 1800-2400 mg/m² be used. Future studies might evaluate the timing of ara-C and hydroxyurea dosing as well as comparing this regimen with other cominations (such as ara-C/clofarabine). In addition the finding of 2 complete and 1 partial response in children with T-cell ALL suggests the need for study in additional patients.