Low-dose antithymocyte globulin enhanced the efficacy of tacrolimus and mycophenolate for GVHD prophylaxis in recipients of unrelated SCT

Abstract

We performed a retrospective analysis of the outcome of 197 consecutive unrelated donor transplant recipients who received GVHD prophylaxis either TM regimen (tacrolimus and mycophenolate) (121 patients) or TM/ATG-G regimen (TM with low-dose antithymocyte globulin (ATG) of 4.5 mg/kg, ATG-G, Genzyme) (76 patients). Cumulative incidences of grade II-IV acute GVHD for the TM and TM/ATG-G cohorts were 49% and 61% (P = 0.11) and grade III-IV acute GVHD for the TM and TM/ATG-G cohorts were 27% and 14% (P = 0.02), respectively. There was no difference in the incidence of relapse or disease progression between TM and TM/ATG-G—16% and 23% (P = 0.64). TM/ATG-G cohort had lower incidence of non-relapse mortality (NRM; 37% vs 20%, P = 0.01), chronic GVHD (56% vs 43%, P<0.001) and more favorable global chronic GVHD severity (P<0.001). Univariate analyses showed improved OS and PFS of patients who received TM/ATG-G. Multivariate analysis confirmed TM/ATG-G had a favorable influence on OS (P = 0.05) but not on PFS (P = 0.07). We concluded that low-dose ATG of 4.5 mg/kg given in conjunction with TM improved GVHD prophylaxis without increased risk of relapse. Lower NRM, lower incidence and severity of chronic GVHD could potentially improve survival.

INTRODUCTION

The biological basis of T-cell depletion in animal models for the prevention of GVHD has been established since 1968.¹ Using the Center for International Blood and Marrow Transplant Research (CIBMTR) database, Champlin et al.² evaluated various T-cell depletion strategies in 870 patients with leukemia who received T-cell-depleted transplants from unrelated or HLA-mismatched related donors from 1982 to 1994. Patients receiving narrow-specificity antibodies for T-cell depletion experienced higher leukemia-free survival than recipients of T-cell depleted SCTs by other techniques but similar to those for non-T-cell-depleted transplants.² In vivo administration of T-cell- and B-cell-depleting agents as GVHD prophylaxis is a much simpler strategy and therefore much more generalizable in its application.

Over the past decades, the Food and Drug Administration has approved several T-cell- and B-cell-depleting antibodies for various non-hematopoietic-transplant indications that gradually found their roles in hematopoietic cell transplantation.^3–5 A randomized phase III trial using rabbit antithymocyte globulin (ATG) produced by immunizing with Jurkat cells (ATG-F, Fresenius Biotech, Grafelfing, Germany) given in combination with CYA and MTX in unrelated donor transplants conditioned with myeloablative regimens had resulted in a lower incidence of acute and chronic GVHD without increase in relapse and non-relapse mortality (NRM), although there was no difference in OS.⁶

Another preparation of rabbit ATG-G (Thymoglobulin, Genzyme, Cambridge, MA, USA) derived from immunization with human thymocytes has unique biological properties and has been extensively reported in transplant literature.⁷ Addition of ATG-G to well-established pharmacological regimens of GVHD prophylaxis could possibly improve clinical outcome. Here we now report our experience with the combination of ATG-G, tacrolimus and mycophenolate mofetil (MMF) in unrelated donor SCT (UHSCT).

MATERIALS AND METHODS

Subjects

We studied 197 consecutive patients who received allo-SCT from unrelated donor from June 2008 to December of 2012. Attending physicians evaluating patient eligibility for SCT assigned preparative regimens according to the underlying diagnosis. These patients received either tacrolimus and MMF (TM) or TM with ATG-G (TM/ATG-G regimen) for GVHD prophylaxis. Patients receiving the TM regimen were transplanted earlier in the study period while all except one patient in the TM/ATG-G group received their transplant in 2011 or after, when we adopted the inclusion of low-dose ATG-G as a standard of care. This decision was based on the conclusion of our prospective trial of ATG-G suggesting the favorable outcome when ATG-G was included in the GVHD prophylaxis.⁸

Preparative regimens

The intensity of the regimen was categorized according to published CIBMTR consensus.⁹ Reduced intensity conditioning (RIC) regimens given to patients with lymphoid malignancies were rituximab in combination with BEAM (BCNU 300 mg/m², etoposide 100 mg/m² every 12 h × 8 doses, ara-C 100 mg/m² every 12 h × 12 doses and melphalan 140 mg/m² × 1) or Flu/Mel/TBI (fludarabine 30 mg/m²/day × 5, melphalan 140 mg/m² × 1 and TBI 100 cGy × 1). Myeloablative regimen (MA) for lymphoid malignancies were cyclophosphamide 60 mg/kg × 2 days or etoposide 60 mg/kg given in conjunction with TBI 1200 cGy given in two daily fractions over 3–4 days. Patients with myeloid malignancies received BU 130 mg/m² daily × 4 and fludarabine 30 mg/m² daily × 5 for MA regimen or BU 130 mg/m² daily × 2, fludarabine 30 mg/m² daily × 5 and TBI 200 cGy × 1 for RIC regimen.

GVHD prophylaxis

Patient received either TM or TM/ATG-G regimen for GVHD prophylaxis, assigned by physicians who first saw patients in the consultation. TM regimen had been the preceding standard of care and later ATG-G was introduced into our practice, mainly for older patients who had high-risk disease index and mostly considered unfit for MA regimen. Tacrolimus and MMF were started on day − 3. The dose and blood level monitoring of tacrolimus and dose adjustment were previously described.¹⁰ MMF was administered i.v. at the dose of 10 mg/kg every 8 h and later converted to oral route when able to tolerate oral medication. Blood levels of MMF were not monitored. MMF was discontinued on day 30 if patients did not develop acute GVHD. ATG-G total dose of 4.5 mg/kg was given daily by i.v. infusion over 3 days: day − 3, 0.5 mg/kg; day − 2, 1.5 mg/kg; and day − 1, 2.5 mg/kg). All patients received premedication with acetaminophen, diphenhydramine and steroids prior to ATG-G infusion. The stem cell graft was infused within 24 h of the last administration of ATG-G.

Statistical analyses

Patient characteristics and demographics between patients receiving TM vs TM/ATG-G were compared using the chi-square test or Fisher's exact test for categorical variables and the Kruskal–Wallis test for continuous variables. Probability of OS and PFS was estimated according to the method of Kaplan and Meier.¹¹ Death from any cause was considered an event for OS estimate and relapse or death was considered an event for relapse-free survival. Surviving patients were censored at last follow-up.

We graded the severity of acute GVHD according to consensus criteria and chronic GVHD according to NIH Consensus.^12,13 Probabilities of acute grade II–IV GVHD, acute grade III– IV GVHD, chronic GVHD, relapse and NRMs were calculated using the cumulative-incidence function.¹⁴ For GVHD end points, death without GVHD was the competing event. For NRM, disease relapse was the competing event. For relapse, death without relapse was the competing event. Data on patients without either competing event were censored at last follow-up. All P-values were two-sided, and a value of <0.05 was considered statistically significant.

Cox's proportional hazards model was used to evaluate covariates with independent effect on the end point of interest.¹⁵ Variables included in the models were: age at transplantation, disease-risk category, hematopoietic cell transplantation-specific comorbidity index (HCT-CI),¹⁶ regimen intensity, and GVHD prophylaxis. All the analyses were completed using the IBM SPSS Statistics 20 software (IBM, Armonk, NY, USA) except for competing risk analyses where R version 2.15.0 (2012-03-30, ‘cmprsk package’, author: Bob Gray, gray@jimmy.harvard.edu) was used. Cumulative incidence of relapse, NRM, acute and chronic GVHD between TM and TM/ATG-G were compared using the Gray's test.¹⁷

RESULTS

Study subjects

From June 2008 to December 2012, we performed UHSCT in 197 consecutive patients with hematological malignancies at Karmanos Cancer Institute. Characteristics and demographics of patients who received GVHD prophylaxis with TM and TM/ATG-G regimen are shown in Table 1. Patients who received TM/ATG-G were significantly older (P = 0.04). Patients with high-risk disease more frequently received a RIC (P = 0.04, data not shown). Distribution of HCT-CI was similar between patients who received MA and RIC regimens (P = 0.41, data not shown), as was disease-risk index (P = 0.72, data not shown). There was a higher proportion of mismatched transplant recipients who received TM/ATG-G. The remaining characteristics were evenly distributed between TM and TM/ATG-G. The median follow-up for survivors was 20 months (range 4–59 months); 30 months (range 5–59 months) and 12 months (range 5–23 months) for the TM and TM/ATG-G groups, respectively.

Table 1.

Patient characteristics and demographics

Characteristics	TM	TM/ATG-G	P-value
Number	121	76
Median age (range), years	52 (20–70)	56 (21–70)	0.04
Sex			0.48
F	62 (51%)	35 (46%)
M	59 (49%)	41 (54%)
Median follow-up for survivors (range)	30 months (5–59 months)	12 months (5–23 months)	< 0.001
Stem cell graft			0.63
Marrow	6 (5%)	5 (7%)
Peripheral blood	115 (95%)	71 (93%)
Diagnosis			0.72
ALL	12 (10%)	12 (16%)
AML	50 (41%)	35 (46%)
CLL	12 (10%)	4 (5%)
CML	5 (4%)	3 (4%)
Hodgkin disease	1 (1%)	1 (1%)
MDS/MPN	21 (17%)	14 (18%)
Myeloma	1 (1%)	0 (0%)
Non-Hodgkin lymphoma	18 (15%)	7 (9%)
Prolymphocytic leukemia	1 (1%)	0 (0%)
HLA-match			0.004
10/10	119 (98%)	67 (88%)
9/10	2 (2%)	9 (12%)
Disease-risk category			0.004
Early	30 (25%)	36 (47%)
Intermediate	16 (13%)	8 (11%)
Advanced	75 (62%)	32 (42%)
Regimen intensity			0.005
Myeloablative	73 (60%)	30 (39%)
Reduced intensity	48 (40%)	46 (61%)
HCT-CI 20			0.24
Low	15 (12%)	9 (12%)
Intermediate	69 (57%)	35 (46%)
High	37 (31%)	32 (42%)

Abbreviations: F = female; HCT-CI = hematopoietic cell transplantation-specific comorbidity index; M = male; MDS = myelodysplastic syndrome; MPN = myeloproliferative neoplasms. Disease risk category: AML, ALL and CML are classified as: Early phase (first complete remission (CR1) or first chronic phase (CP1)), Intermediate phase (second or subsequent CR or CP or accelerated phase (AP)), or Advanced phase (primary induction failure, active disease or blastic phase) disease. MDS is divided into: Early phase (refractory anemia (RA) or refractory anemia with ringed sideroblasts (RARS)), and Advanced phase (refractory anemia with excess blasts (RAEB) or chronic myelomonocytic leukemia (CMML)) disease. Lymphoma is classified according to sensitivity to prior chemotherapy: Early phase CR1 (chemosensitive); Intermediate phase ≥ CR2 (chemosensitive); and Advanced phase: residual disease after salvage therapy (chemoresistant). Patients with CLL, myeloma and prolymphocytic leukemia were categorized as advanced disease as all these patients had failed multiple regimens prior to transplantation.

Acute GVHD

Cumulative incidences of grade II-IV acute GVHD for the TM and TM/ATG-G cohorts were 49% (95% confidence interval (CI), 39–59%) and 61% (95% CI, 49–73%) (P = 0.11), respectively. Cumulative incidences of grade III-IV acute GVHD for the TM and TM/ATG-G cohorts were 27% (95% CI, 19–35%) and 14% (95% CI, 4–24%) (P = 0.02), respectively (See Figures 1a and b).

Figure 1.

(a) Cumulative incidence of grade II-IV acute GVHD; (b) cumulative incidence of grade III-IV acute GVHD.

Univariate analyses for other covariates on the cumulative incidence of acute GVHD including preparative regimen (MA vs RIC), HLA-matching (9/10 vs 10/10) did not show a difference in the incidence of grade II-IV or grade III-IV acute GVHD (data not shown). There were no significant differences in the distribution of the organ stage between patients receiving TM or TM/ATG-G for GVHD prophylaxis (See Supplementary Material). There was no liver GVHD in the TM/ATG-G group and 8 out of the 121 patients in the TM group had liver GVHD (P = 0.12). In patients with gastrointestinal GVHD, there were 37 out of the 121 (31%) in the TM group who developed stage 2–4 acute GVHD, whereas only 10 out of the 76 (13%) patients in the TM/ATG-G group (P = 0.06 for the distribution of all stages of gastrointestinal GVHD).

Chronic GVHD

Cumulative incidence of chronic GVHD was 43% (95% CI, 40–72%) in patients who received TM/ATG-G and 56% (95% CI, 46–66%) in patients who received TM, respectively (P<0.001). Although the pattern of time to chronic GVHD suggest delayed development of chronic GVHD in the TM/ATG-G group, there was a significant difference in the distribution of the severity of chronic GVHD between TM and TM/ATG-G cohorts. Patients who received TM/ATG-G had a significantly lower global severity of chronic GVHD (P<0.001; Figure 2 and Table 2).

Figure 2.

Cumulative incidence of chronic GVHD.

Table 2.

Distribution of global chronic GVHD severity

Severity	TM	TM/ATG-G	Total
None	54 (44.6%)	53 (69.7%%)	107
Mild	26 (21.5%)	15 (19.7%)	41
Moderate	22 (18.2%)	5(6.6%)	32
Severe	19 (15.7%)	3 (3.9%)	27

P < 0.001.

Cytomegalovirus (CMV) reactivation/infection and thrombotic microangiopathy

Forty-eight of the 121 patients (39.7%) in the TM cohort and 28 patients (36.8%) of the 76 patients in the TM/ATG-G developed CMV reactivation or infection, and the risk of CMV reactivation/infection was not different between the two GVHD prophylaxis group (P = 0.40). Five patients developing thrombotic microangiopathy all received TM. There were no sinusoidal obstruction syndromes encountered in the entire study cohort.

Disease progression/relapse and NRM

There was no difference in the cumulative incidence of relapse or disease progression at 20 months between the TM and TM/ATG-G cohorts, 16% (95% CI, 8–24%) vs 23% (95% CI, 0–47%) (P = 0.64), respectively (Figure 3a). There was no difference in the risk of relapse between patients who received MA or RIC regimens, 16% (95% CI, 8–24%) vs 19% (95% CI, 11–27%) (P = 0.38), respectively. NRM was significantly lower in the TM/ATG-G cohort, 20% (95% CI, 10–30%) vs 37% (95% CI, 29–45%) (P = 0.01), respectively (Figure 3b). The intensity of preparative regimen had no influence on relapse rate at 50 months: MA vs RIC, 16% (95% CI, 12–20%) and 19% (95% CI, 15–23%) (P = 0.38); and there was no difference in NRM, 34% (95% CI, 30–38%) and 31% (95% CI, 28–34%), respectively. In comparing subgroups of patients who received MA or RIC vs TM or TM/ATG-G, there were no difference in the relapse rate or NRM among the four subgroups of patients, although there was a trend for lower NRM in patients receiving TM/ATG-G regardless of the intensity of the preparative regimen (Figures 4a and b).

Figure 3.

(a) Cumulative incidence of disease progression or relapse; (b) cumulative incidence of NRM.

Figure 4.

(a) Cumulative incidence of relapse at 20 months for the four subgroups—MA or RIC vs TM or TM-ATG-G; (b) Cumulative incidence of NRM at 20 months for the four subgroups—MA or RIC vs TM or TM-ATG-G.

Survival

Kaplan–Meier estimates for PFS and OS for the TM and TM/ATG-G cohorts are shown in Figures 5a and b, respectively. Disease-risk status at the time of transplant affected OS at 4 years: 72% (95% CI, 60–84%, N = 66)) for low-risk, 51% (95% CI, 26–76%, N = 24) for intermediate risk, and 34% (95% CI, 20–48%, N = 107) for high-risk disease, respectively (P<0.001). OS at 2 years was significantly better in the TM/ATG-G cohort: 64% (95% CI, 48–80%, N = 76) vs 49% (95% CI, 39–59%) (P = 0.01) in the TM cohort. There were no significant differences among patients who received a 9/10 or 10/10 matched donor stem cell products, the intensity of preparative regimen and HCT-CI. Multivariate analysis revealed unfavorable impact of high-risk disease but improved OS in receiving TM/ATG-G for GVHD prophylaxis (Table 3).

Figure 5.

Kaplan-Meier estimates for (a) PFS and (b) OS.

Table 3.

Multivariate analyses of PFS and OS

PFS
TM/ATG-G	0.64	0.40–1.04	0.07
Intermediate-risk disease	2.05	0.99–4.25	0.053
High-risk disease	2.77	1.59–4.83	< 0.001
RIC	0.95	0.62–1.46	0.82
HCT-CI intermediate risk	0.97	0.49–1.93	0.92
HCT-CI high risk	1.33	0.66–2.69	0.42
OS
TM/ATG-G	0.60	0.36–0.99	0.05
High-risk disease	2.75	1.55–4.88	0.001
Intermediate-risk disease	1.62	0.73–3.36	0.24
RIC	0.98	0.63–1.53	0.92
HCT-CI intermediate risk	1.05	0.69–2.25	0.90
HCT-CI high risk	1.69	0.78–3.66	0.18

Abbreviations: HCT-CI = hematopoietic cell transplantation-specific comorbidity index; RIC = reduced intensity conditioning.

In the univariate analysis, disease-risk status affected PFS. PFS at 4 years for the low-risk cohort was 72% (95% CI, 60–84%, N = 66), 44% (95% CI, 24–64%, N = 24) for intermediate risk and 36% (95% CI, 24–48%, N = 107) for high risk disease, respectively (P<0.001). PFS at 2 years was better in the TM/ATG-G cohort compared with the TM cohort: 54% (95% CI, 32–76%, N = 76) vs 43% (95% CI, 33–53%, N = 121) (P = 0.02). Multivariate analysis showed high-risk disease status was independently predictive of poor PFS (P<0.001) and intermediate-risk disease was marginally signifi-cant (P = 0.05). The TM/ATG-G cohort showed a trend toward improved PFS (P = 0.07) (Table 3).

DISCUSSION

In an attempt to overcome rejection of hematopoietic stem cell grafts in a canine model conditioned with sublethal dose of TBI, postgrafting administration of MMF in combination with CYA has been shown to be able to inhibit both graft rejection and GVHD.¹⁸ Clinical trials using MMF and a calcineurin inhibitor confirmed the efficacy of this combination.^19–21 Although MMF in combination with tacrolimus provided similar efficacy for GVHD prevention with less mucosal toxicity compared with MTX and tacrolimus in matched sibling donor transplants, it might be inferior to MTX and tacrolimus in the unrelated donor setting.^22,23

ATG-G used in this study is produced by immunization of New Zealand rabbits with fresh human thymocytes; it had immunological response against various antigens on T cells, B cells, DCs and stromal cells.^7,24 The spectrum of activities of this ATG-G encompassed immune response antigens, adhesion and cell trafficking molecules.⁷ Inhibition of cell trafficking molecule, C-C motif chemokine receptor 5, could be crucial in the pathogenesis of visceral GVHD as shown in a study using maraviroc.²⁵ Furthermore, depletion of B cells might also reduce risk of acute and chronic GVHD.^4,26,27 Bacigalupo et al.²⁸ reported the results of two consecutive randomized trials in the recipients of UHSCT—ATG-G at 7.5 mg/kg vs no ATG-G and ATG-G 15 mg/kg vs no ATG-G. In the trial with ATG-G 15 mg/kg, there was a significant reduction of grade III-IV acute GVHD, chronic GVHD risk and severity but this was offset by increased risk of infection, resulting in no improvement of OS.²⁸ Extended observation confirmed the protection of ATG-G against extensive chronic GVHD, chronic lung dysfuction, late transplant mortality and improvement of quality of life.²⁹ In 2009, a randomized phase III trial of adding ATG-F 20 mg/kg daily × 3 days to CYA and MTX in UHSCT using myeloablative regimens showed lower incidences of acute and chronic GVHD in the ATG-F group without increased relapse or NRM but no improvement of OS.³⁰ The risk and severity of chronic GVHD was significantly reduced with better quality of life in patients receiving ATG-F without increased relapse rate.³¹

The use of ATG in the RIC setting is more controversial. CIBMTR had attempted to evaluate the role of anti-T-cell antibodies in both related and unrelated donor transplantation in patients conditioned with reduced intensity regimens.³² The main concern was the abrogation of graft-versus-tumor effect by anti-T-cell antibody which might increase the relapse rate in the setting of RIC. However, this study was confounded by the use of alemtuzumab, the dose and type of ATG (rabbit—Genzyme vs Fresenius or horse). There was significantly higher relapse with the alemtuzumab and ATG compared with T-cell-replete regimens, with corresponding lower disease-free survival and OS. However, approximately 63% of the patients in this study received rabbit ATG of ≥ 6 mg/kg; this might explain the high relapse rate in patients receiving ATG. More recent studies using rabbit ATG (ATG-G) with the dose between 5 and 7 mg/kg had shown favorable outcome in the RIC setting.^33–35

Previously, we reported the favorable results of ATG-G given in combination with tacrolimus and sirolimus for GVHD prophylaxis in unrelated donor transplantation.⁸ Combination of sirolimus with calcineurin inhibitor with or without MTX appears to improve the efficacy of GVHD prophylaxis in several studies.^36–39 The additive toxicity of sirolimus to tacrolimus and MTX has posed serious concern in other studies.^40,41 In a randomized phase II study of sirolimus/tacrolimus vs MTX/tacrolimus for GVHD prophylaxis, sirolimus/tacrolimus was more effective than MTX/ tacrolimus, resulting in lower grade II-IV acute GVHD and moderate–severe chronic GVHD.⁴² However, the follow-up study of the subjects in this randomized study showed that patients receiving sirolimus had poorer quality of life due to persistent fatigue and nausea.⁴³

In this study, we simplified our previous ATG-G-containing GVHD prophylaxis by replacing sirolimus with MMF. Critical review comparing mycophenolate and MTX in combination with calcineurin inhibitor had shown a favorable toxicity profile for mycophenolate but with a higher rate of acute GVHD in UHSCT.²³ We hypothesized that addition of thymoglobulin to TM could improve the efficacy of GVHD prophylaxis without increased NRM and lessen the complexity of drug monitoring and interaction associated with the administration of sirolimus. Notably, we found that patients who received TM/ATG-G had a lower rate of grade III-IV acute GVHD but no difference in rate of grade II-IV acute GVHD. It is possible that ATG-G might have impaired lymphocyte trafficking through inhibition of C-C motif chemokine receptor 5 (CD195) similar to inhibition by maraviroc, such that visceral GVHD is effectively prevented, resulting in a lower rate of grade III-IV acute GVHD.²⁵ Interestingly, lower rate of grade III-IV acute GVHD and NRM at 20 months was not associated with increased risk of relapse. Furthermore, there is no increased risk of CMV reactivation, indicating that low-dose ATG-G might not have affected the immune recovery as seen in higher dose.⁴⁴

In the univariate analyses, TM/ATG-G significantly improved both OS and PFS. In the multivariate analyses, recipients of TM/ ATG-G exhibited a favorable OS, whereas high-risk disease had unfavorable impact on OS. However, only high-risk disease predicted unfavorable PFS. TM/ATG-G had marginally favorable impact on PFS (Table 3). Taken together, addition of ATG-G to the existing well-tolerated tacrolimus and MMF regimen improved clinical outcome. We observed lower grade III-IV acute GVHD, lower chronic GVHD and better survivals without increased risk of CMV reactivation. Higher dose of ATG-G employed in other studies is clearly associated with abrogation of graft-versus-tumor effect with higher relapse rate and lower survival.^45,46 A Bayesian adaptive randomized study comparing ATG-G of 4.5 mg/kg vs 7.5 mg/kg showed higher treatment success rate (defined as patient being alive, engrafted, in remission and without acute GVHD at day 100) in patients receiving lower dose of ATG-G.⁴⁷ The merits of ATG-G for GVHD prophylaxis were also confirmed by the French Society of Bone Marrow Transplantation and Cellular Therapy group in 242 progressive MDS patients, showing significant reduction of acute GVHD without impact on relapse and survival.⁴⁸ We conclude that the dose of ATG-G at 4.5 mg/kg in combination with calcineurin inhibitor might be an optimal foundation for further development of GVHD prophylaxis. Further confirmatory prospective study to evaluate the ultimate impact of low-dose ATG-G on survival is warranted.