MYELOABLATIVE INTRAVENOUS BUSULFAN/FLUDARABINE CONDITIONING DOES NOT FACILITATE RELIABLE ENGRAFTMENT OF DUAL UMBILICAL CORD BLOOD GRAFTS IN ADULT RECIPIENTS

Abstract

The efficacy of once daily intravenous busulfan with fludarabine as a preparative regimen for partially matched umbilical cord blood transplantation has not been formally studied. We randomized 10 adult patients with myeloid malignancies to receive either concurrent or sequential administration of intravenous busulfan 130mg/m² once daily × 4 days and fludarabine 40mg/m² daily × 4 days followed by dual umbilical cord blood transplantation. The median combined cryopreserved total nucleated cell dose was 3.6 × 10⁷/kg recipient body weight (range 2.8 to 4.5 × 10⁷/kg). Graft versus Host disease prophylaxis was provided by tacrolimus and mycophenolate mofetil. Donor-derived neutrophil recovery was observed in only 2 of 10 patients resulting in premature closure of the study as per graft failure stopping rules. We conclude that the myeloablative conditioning regimen of once daily intravenous busulfan with fludarabine provides insufficient immunosuppression to allow for engraftment of partially matched, dual umbilical cord blood grafts.

INTRODUCTION

Since the publication by de Lima and colleagues in 2004, allogeneic stem cell transplant (SCT) conditioning consisting of once daily intravenous busulfan and fludarabine (BU/FLU) has gained widespread popularity (1). Enthusiasm for this regimen is derived from the fact that it provides a myeloablative dose of busulfan with exceptionally low early treatment-related mortality (3%). Pharmacokinetic studies demonstrate predictable busulfan blood levels with an AUC comparable to dose adjusted oral administration (1). The regimen facilitates reliable engraftment of allogeneic peripheral blood or bone marrow grafts from HLA-identical siblings or matched unrelated donors. Widespread use of umbilical cord blood as a stem cell source for adult patients in need of allogeneic SCT has been hindered by treatment-related mortality as high as 63% following myeloablative conditioning and high rates of graft failure due to an inadequate stem cell dose or immunologic rejection (2). Preliminary evidence suggests that infusion of two umbilical cord blood grafts improves engraftment in adult recipients (3–5). This prompted us to conduct a prospective study of BU/FLU preparation followed by dual umbilical cord blood transplantation. Given its non-specific alkylating properties, we hypothesize that busulfan could interfere with deoxycytidine kinase mediated accumulation of intracellular F-ara-ATP thereby attenuating the immunosuppressive properties of the fludarabine. We therefore randomized patients to sequential versus concurrent administration of IV busulfan and fludarabine.

PATIENTS, MATERIALS, METHODS

Patients

Ten patients, median age 41, with myeloid malignancies (acute myelogenous leukemia [AML], chronic myelogenous leukemia [CML] or myelodysplastic syndrome [MDS]) provided informed consent for participation in this randomized phase II study. Of the AML patients, subject #4 had 4% residual blasts at the time of transplantation. The remainder had no detectable disease in the bone marrow. All MDS-RAEB-2 patients were downgraded to RA or RAEB with pre-transplantation therapy. The CML patients were in chronic phase at the time of transplantation. The study was approved by the institutional review board of the Duke University School of Medicine. Patient and disease characteristics are demonstrated in Table 1.

Table 1.

Patient Characteristics

Patient No.	Age y/Sex	Weight (kg)	Disease	Karyotype	Status at Transplantation
1	40M	101	AML	Inv16	CR2
2	21M	76	MDS (Hypocell.)	−7	Aplastic
3	29M	96*	CML(BC)	Ph+	CP
4	49F	77	RAEB-2	Complex	RAEB
5	32M	95	AML	Normal	PR**
6	44F	68	MDS RAEB-2	Inv3	RA
7	50F	93	AML	Complex	CR1
8	41M	92	AML	Normal	CR2
9	54F	76	CML	Ph−	CP
10	34M	67	RAEB-2	Complex	RA

RAEB- Refractory anemia with excess blasts

RA- Refractory Anemia

BC- Blast Crisis

CP- Chronic Phase

CR- Complete Remission; no morphologic evidence of disease

^*Adjusted ideal body weight

^**Partial Remission; 4% blasts with aberrant phenotype

Treatment plan

Bone marrow conditioning consisted of fludarabine (Berlex Laboratories, Wayne, NJ), 40mg/m²/d × 4 days and IV busulfan (PDL Biopharma, Redwood City CA) 130mg/m²/d × 4 days. Patients were randomized to receive the drugs sequentially (fludarabine days -9 to -6; Busulfan days -5 to -2) or concurrently (days -5 to -2) (1). Six patients received the drugs sequentially, and 4 concurrently. Busulfan pharmacokinetic measurements were performed on the first and fourth dose using a high-pressure chromatography assay as previously described (6). Patients were transplanted with two partially matched umbilical cord blood grafts that were at least 4 of 6 HLA-matched with the recipient and 4 of 6 HLA matched between grafts (low resolution class I, high resolution class II) (Table 2). Graft versus Host disease prophylaxis was provided by tacrolimus dosed to achieve serum levels of 10-20 ng/ml for 6 months and mycophenolate mofetil 1000mg twice daily until 60 days following transplantation. Donor chimerism was determined by quantitative PCR amplification of informative microsatellites using DNA isolated from peripheral blood or bone marrow CD3+ and CD15+ cells. Primary graft failure was defined as a patient who never had detectable donor hematopoiesis. Secondary graft failure was defined as a patient who had detectable early donor hematopoiesis that was subsequently lost. Patients with early disease relapse were deemed not evaluable for graft failure.

Table 2.

Graft Characteristics, Engraftment, and Outcome

Patient No.	Combined Cord TNC/kgf (×107)	HLA-Matching† (Cord#1 + Cord #2)	Busulfan/Fludarabine Administration	Maximum %Donor Myeloid Chimerism (Cord #1/#2)	Maximum %Donor Lymphoid Chimerism (Cord #1/#2)	Current Status	Cause of Death
1	3.9	4/6 + 4/6	Concurrent	0/0	0/0	CR* day 700	---
2	3.4	4/6 + 4/6	Sequential	0/0	0/0	Dead day 89*	Infection
3	2.8	4/6 + 4/6	Concurrent	0/0	0/0	Dead day 134*	Infection
4	3.7	4/6 +4/6+	Sequential	6/6^		Dead day 137	Relapsed AML
5	3.9	6/6 + 6/6¶	Sequential	14/77	9/76	Dead day 118	Relapsed AML
6	4.5	4/6 + 4/6¶	Concurrent	25/52	5/20	Dead day 85	Relapsed AML
7	3.6	5/6 + 5/6	Sequential	0/9^		Dead day 33	Relapsed AML
8	3.5	4/6 + 4/6	Sequential	0/0	0/0	CR day 485√	---
9	3.1	4/6 + 4/6	Concurrent	0/0	0/0	Dead day 108*	Infection
10	3.9	4/6 + 4/6	Sequential	0/0	0/0	Dead day 88	Recurrent AML

^fCryopreserved

^†Low resolution Class I, High resolution Class II

^*Post Haploidentical second transplant

^{^}Unselected bone marrow sample

^¶Cord #1 and #2 are Class I and Class II allele-level matched (probable siblings)

^√Post-Autologous bone marrow rescue

⁺5/6 in direction of rejection

Statistical methods

A sample size of 24 for each arm of the study was expected to provide enough power to detect a difference in rate of engraftment. Separate stopping rules for primary or secondary graft failure were constructed for each treatment arm using the Bayesian approach that was based on an expected graft failure rate of less than or equal to 20%. The parameters assumed α=0.8 and β=3.2. Three graft failure events among the first 6 patients enrolled in each arm triggered early stopping. Early stopping rules were activated for both treatment arms.

RESULTS AND DISCUSSION

All patients with AML were in remission at the time of protocol entry. Patient 3 was in second chronic phase following induction therapy for lymphoid blast crisis. All subjects with MDS with increased blasts underwent induction chemotherapy and achieved a partial response before protocol therapy. The median combined umbilical cord blood cell dose was 3.6 × 10⁷ total nucleated cells/kg (Table 2). All grafts were at least 4 of 6 HLA matched with the recipient and each other. The mean and median daily AUC’s for busulfan were 4185 and 4225 μmole-min (range, 2634–5278 μmole-min), respectively. These values are nearly identical to those reported by de Lima and colleagues who, with the same preparative regimen, achieved 100% donor engraftment using peripheral blood or bone marrow grafts (1). Eight of 10 patients experienced primary or secondary graft failure, although 2 patients were not evaluable for engraftment due to early disease relapse. Only 2 of 10 patients achieved significant, yet incomplete donor hematopoietic chimerism (one on sequential, the other on the concurrent treatment arms). Stopping rules for graft failure were met for both arms of the trial and thus there was no appreciable difference in the probability of engraftment between sequential or concurrent BU/FLU administration. Hematopoiesis was reconstituted in 5 patients following a second transplant (4 using haploidentical grafts, 1 using an autologous back-up graft), two of whom have achieved long-term disease-free survival (Table 2). Three patients died of infection-related complications and 5 from relapsed AML.

The poor engraftment that we observed with BU/FLU conditioning is in contrast to what has been reported in adult recipients using dual umbilical cord blood grafts and more immunosuppressive myeloablative or non-myeloablative conditioning regimens (3–5). This occurred despite providing an umbilical cord blood graft of comparable HLA-matching and size. In the study by Ballen and colleagues, engraftment was achieved in 18 of 21 patients following reduced intensity preparation with melphalan 100mg/m², fludarabine 150mg/m2 and anti-thymocyte globulin (4). Brunstein and colleagues transplanted 110 patients, most of whom required 2 umbilical cord blood units to achieve a target cell dose of 3.0 × 10⁷/kg. With a non-myeloablative preparative regimen of cyclophosphamide 50mg/kg, fludarabine 200mg/m² and total body irradiation 200cGy, the cumulative incidence of sustained engraftment was 85% (5). In an earlier study, the same group reported inferior engraftment when non-myeloablative doses of oral busulfan (8mg/kg) were used instead of cyclophosphamide (7).

Following myeloablative, single agent dosing of busulfan (16mg/kg total dose) with autologous stem cell support, Peters and colleagues demonstrated sparing of mature lymphoid cells in the peripheral blood. No significant change was observed in CD3+, CD4+ or CD8+ T-cell counts during the 26-day post-transplant monitoring period. B-cell immune function was also spared (8). While Gandhi and Plunkett demonstrate that fludarabine potentiates alkylator-induced cytotoxicity (9), the lack of busulfan induced lymphocytotoxicity negates this potentially beneficial synergy and raises significant questions surrounding the immunosuppressive properties of the BU/FLU regimen. This is particularly important in the setting of umbilical cord blood transplantation given the low stem cell dose and donor/recipient HLA-discordance.

Sequencing in drug administration has been shown to affect the efficacy of many combination therapies, including those with fludarabine and busulfan (9, 10). The cytotoxicity of alkylating agents such as busulfan is generally attributed to the formation of DNA crosslinks, however these drugs are also known to react with many other biomolecules such as proteins, enzymes and thiols (11, 12). Fludarabine induced lymphocytotoxicity is contingent on intracellular accumulation of F-ara ATP, which is in turn dependent on the activity of intracellular deoxycytidine kinase (9). Busulfan-induced interference of deoxycytidine kinase would therefore compromise the immunosuppressive properties of fludarabine. It is for this reason that we constructed a two-armed trial, establishing a cohort of patients who would receive fludarabine whose activity could not be affected by coadministration of busulfan. Verification of busulfan-induced deoxycytidine kinase inhibition is planned through measurement of F-ara ATP levels from samples collected following drug administration. However, given the lack of reliable engraftment in both arms of the trial, the clinical significance of this interaction is in question.

Limitations of the study include the small sample size and the fact that despite being in remission at the time of enrollment, half of the subjects had high-risk cytogenetic abnormalities. While early relapse may serve to inflate the graft failure rate, it should be noted that 4 of 5 patients with standard risk cytogenetic abnormalities also experienced graft failure. The M.D. Anderson group added antithymocyte globulin (ATG) to BU/FLU as added immunosuppression for recipients of unrelated donor grafts (1). Due to concerns over impaired post-transplant immune recovery, susceptibility to opportunistic infections and prolonged platelet recovery, we opted not to include ATG (13, 14).

The BU/FLU regimen is gaining popularity as an attractive option to standard myeloablative conditioning for allogeneic SCT. However, our data suggest that given in the doses and schedules we tested, BU/FLU does not provide adequate immunosuppression to facilitate engraftment of partially matched umbilical cord blood grafts and should not be used in this setting.