Impact of Rituximab and Host/Donor Fc Receptor Polymorphisms after Allogeneic Hematopoietic Cell Transplantation for CD20⁺ B Cell Malignancies

Abstract

We previously reported a 24% 1-year relapse rate in 93 older or medically unfit patients with CD20⁺ B cell malignancies after allogeneic hematopoietic cell transplantation (HCT) with low-intensity conditioning. The current prospective study tested the hypothesis that disease relapse could be reduced and overall survival (OS) improved by peritransplantation administration of rituximab (RTX). Sixty-three patients received RTX (375 mg/m²/day) on days −3, +10, +24, and +38 along with 2 to 3 Gy total body irradiation with or without fludarabine (30 mg/m² for 3 days). Median RTX levels of >25 μg/mL were achieved through day +84 after transplantation, but RTX level was not correlated with relapse or graft-versus-host disease (GVHD). HCT recipients with F/F and V/F FCγRIIIa polymorphisms showed a trend toward a higher relapse rate compared with those with V/V polymorphism (P= .15). No difference in outcome was found based on V/V donor pairing. Five-year relapse rates were similar between RTX-treated patients and historical controls (32% versus 28%; P= .94). RTX-treated patients had greater 5-year OS (47% versus 38%; P = .13) and progression-free survival (41% versus 32%; P = .12) compared with historical controls who underwent HCT without RTX, although the difference was not statistically significant. The incidence of acute GVHD was similar in the 2 groups (grade II-IV, 57% versus 56%; grade III-IV, 13% versus 17%), but the 5-year incidence of chronic GVHD was higher among RTX-treated patients (62% versus 47%). In patients with relapsed or refractory non-Hodgkin lymphoma, peritransplantation RTX neither reduced relapse nor improved GVHD. The role of donor-recipient pairing by FCγRIIIa polymorphisms in outcomes remains to be determined.

INTRODUCTION

Non-Hodgkin lymphoma (NHL) is an increasingly common cancer, with an estimated 74,200 new cases in the United States in 2019 (Surveillance Epidemiology and End Results database, United States National Cancer Institute: https://seer.cancer.gov/statfacts/html/nhl.html). The worldwide incidence of NHL has increased rapidly in recent decades, particularly in the elderly.

Indolent NHL remains incurable with currently available conventional therapies, and many patients with aggressive NHL relapse after conventional treatment or high-dose therapy with autologous hematopoietic cell transplantation (HCT). Myeloablative allogeneic HCT can produce long-term disease-free survival in patients with relapsed or refractory NHL; however, this treatment is generally limited to younger and healthier patients and is associated with high rates of transplantation-related mortality (TRM) [1–6]. The median age at diagnosis for the most common types of NHL ranges from 59 to 64 years [7], and patients may be considerably older by the time that allogeneic HCT is contemplated. Thus, many patients with relapsed NHL are ineligible for this potentially curative therapy as a result of advanced age, heavy pretreatment (including previous autologous HCT), or medical comorbidities.

Reduced-intensity and nonmyeloablative (NMA) conditioning regimens for allogeneic HCT have been developed that exploit immunologic graft-versus-tumor effects and allow extension of allogeneic HCT to include patients who are ineligible for myeloablative conditioning. A NMA conditioning regimen of 2 Gy total body irradiation (TBI) and 90 mg/m²fludarabine facilitates allogeneic engraftment and development of the graft-versus-tumor effect [8] with decreased rates of hospitalization and regimen-related toxicity compared with myeloablative conditioning [9–11].

We have reported long-term survival of patients with relapsed or refractory indolent and aggressive NHL using this approach [12,13]; however, disease recurrence has been a major cause of treatment failure. This prospective clinical trial studied the addition of the chimeric monoclonal anti-CD20 antibody rituximab (RTX) to existing NMA allogeneic HCT approaches in patients with CD20⁺ B cell malignancies. RTX was administered immediately before HCT and in the weeks following HCT, with the primary goal of reducing the rate of disease relapse. We compared the results of this phase II trial with previously published data for patients conditioned with the same regimen but not given RTX.

The pharmacokinetics and optimal dosing of RTX in the setting of allogeneic HCT are undefined, and thus we collected pharmacokinetic data to evaluate and optimize our dosing regimen. We also collected data on donor and recipient polymorphisms of the FCγRIIIa (CD16) receptor, which have been shown in some, but not all, studies to impact the antitumor activity of RTX [14–18]. We hypothesized that recipients with a change to a more favorable polymorphism phenotype (V/V) would have less risk of relapse.

METHODS

Patient Selection

This prospective study (ClinicalTrials.gov identifier NCT00867529) included patients undergoing NMA allogeneic HCT for relapsed or refractory CD20⁺ B cell malignancies at Fred Hutchinson Cancer Research Center (Fred Hutch) between September 2009 and September 2013. The median duration of follow-up was 84 months (range, 25 to 107 months).

All patients received unmodified peripheral blood mononuclear cell (PBMC) grafts from an appropriate HLA-matched related or unrelated donor. Exclusion criteria were use of an HLA-haploidentical donor, pregnancy, cardiac ejection fraction <40%, pulmonary diffusion capacity <30% of predicted value, decompensated liver disease, renal insufficiency, Karnofsky Performance Status <50% to 60%, serologic evidence of infection with HIV, or active bacterial or fungal infection unresponsive to medical therapy. The clinical trial was approved by the Fred Hutch Institutional Review Board. All patients signed informed consent forms also approved by the Fred Hutch Institutional Review Board.

We compared the outcomes of 156 patients who underwent transplantation at our institution between 1998 and 2006, including 63 patients who received RTX and 93 who did not receive RTX (historical controls).

Pretransplantation Patient Characteristics

Patients were diagnosed with indolent or aggressive NHL. Indolent lymphomas included follicular lymphoma (n = 10), chronic lymphocytic leukemia (n = 9), Hodgkin-like B cell lymphoma (n = 1), and lymphoplasmacytic lymphoma (n = 1). Aggressive NHL included transformed diffuse large cell lymphoma (n = 15), de novo diffuse large cell lymphoma (n = 16), mantle cell lymphoma (n = 9), histiocytic sarcoma (n = 1), and composite lymphoma (n = 1).

Chemotherapy-sensitive disease was defined by partial response (PR) or complete response (CR), according to standard criteria [19], with the chemotherapy regimen immediately preceding HCT. Pretransplantation comorbidities were assessed retrospectively using the HCT Comorbidity Index [20,21 ]. Patients and their donors were matched for HLA-A, -B, -C, -DRB1, and -DQB1 [22]. A single allele or antigen disparity was allowed for HLA-A, -B, or -C as defined by high-resolution typing.

Study Design and Treatment

The treatment protocol is shown in Figure 1. In this single-arm phase II study, treatment was initiated in the outpatient setting, and patients were admitted to inpatient services only as medically necessary for control of transplantation-related complications [11].

Figure 1.

Peritransplantation RTX protocol. RTX serum levels (R level) assays were performed on days −3, +10, +24, +38, +60, +84, and +180. Polymorphisms of CD16 (FCγRIIIa) receptor were assayed-HCT and on day +84.

Day 0 was defined as the day on which PBSCs were infused. Conditioning began 4 days before HCT. From days −4 to −2, a majority of patients received fludarabine (30 mg/m²/day i.v.) before 2 to 3 Gy of TBI on day 0. Peripheral blood mononuclear cells were infused as soon as possible following TBI. Of note, 1 patient received clofarabine (30 mg/m²/day × 5 days) instead of fludarabine [23]. All study patients received RTX at a dose of 375 mg/m² on day −3 before and days +10, +24, and +38 after HCT. Patients underwent bone marrow aspiration and peripheral blood cell sorting on days +28, +56, and +84 after HCT to assess chimerism and disease status.

Graft-versus-host disease (GVHD) prophylaxis consisted of cyclosporine (CSP) or tacrolimus (Tac) combined with mycophenolate mofetil (MMF) with or without sirolimus, as described previously [24–26]. Supportive care, including antimicrobial, varicella zoster virus, and cytomegalovirus prophylaxis, followed institutional standard practice guidelines. Acute and chronic GVHD were graded as described previously [27].

Pharmacokinetics Analysis of Rituximab

RTX serum levels were assessed before the first RTX dose, after each dose, and on days +60, +84, and +180. RTX levels were measured using a murine anti-idiotype (anti-Id) monoclonal antibody (18C9) that binds specifically to RTX, to the murine parent antibody used to construct RTX, and to Fab’ fragments of these antibodies, which do not cross-react with other human IgG1 antibodies or human serum (Supplementary Figure S1). The 18C9 anti-Id was used in an enzyme-linked immunosorbent assay to measure serum RTX concentrations. This assay can quantify RTX levels in fresh, refrigerated, or frozen human serum to <1 μg/mL. Therapeutic serum RTX level was defined as RTX >25 μg/mL. This cutoff was chosen based on a reported correlation of trough levels >25 μg/mL with disease response [28].

Analysis of FCγRIIIa Polymorphisms

Polymorphisms of FCγRIIIa (CD16) receptor assays were performed pre-HCT and on day +84. FCγRIIIa was genotyped by PCR and single-strand confirmation polymorphism (SSCP). A region of exon 4 of FCγRIIIa was amplified by PCR, and the 125-bp product was verified on a 2% agarose gel and by sequencing. The PCR products were denatured and then electrophoresed to separate the different strands (Supplementary Figure S2). This method sensitively identified FCγRIIIa polymorphisms at position 158 (phenylalanine/phenylalanine [F/F], valine/phenylalanine [V/F], and valine/valine [V/V]) from small amounts of DNA isolated from patient cells.

Historical Controls

For historical comparisons, we included data from all patients treated for B cell malignancies on previous prospective and registered trials of allogeneic HCT after NMA conditioning with 30 mg/m²/d fludarabine and 2 Gy TBI between 1998 and 2006 [12,13]. These patients served as the historical controls.

Statistical Analysis

The study was originally designed to evaluate whether addition of RTX would reduce the rate of relapse/progression compared with patients treated without RTX. However, owing to the lengthy accrual period and overall changes in transplantation outcomes over time, comparisons to historical controls should not be interpreted as an evaluation of RTX. The analysis used all available patients, and the sample size was not based on power calculations to test any specific hypothesis. Overall survival (OS) and progression-free survival (PFS) were estimated by the Kaplan-Meier method. Diagnoses of relapse, defined as recurrence of malignancy, and progression were based on previously published criteria [29]. Nonrelapse mortality (NRM) was defined as death without relapse or progression. Cumulative incidences of relapse, NRM, and acute and chronic GVHD were estimated by competing-risk methods, with death as a competing risk for relapse and GVHD and relapse as a competing risk for NRM. For all time-to-event endpoints, comparisons between RTX-treated patients and historical controls, and analyses of risk factors among RTX-treated patients, were based on hazard ratio analysis using Cox regression. All cited P values relating to survival and cumulative incidence refer to the hazard ratio analyses, not to specific time points.

Comparisons of patient characteristics in Table 1 were based on the chi-squared test for grouped variables and on the Wilcoxon rank-sum test for continuous variables. The comparisons of serum RTX levels shown in Figure 4 were based on the Wilcoxon rank-sum test. All Pvalues are 2-sided and unadjusted for multiple comparisons. All analyses were carried out using SAS software (SAS Institute, Cary, NC).

Table 1.

Patient Characteristics

Characteristic	Peri-HCT RTX (N = 63)	Historical Controls (N = 93)	P Value
Age, yr, median (range)	56 (26-71)	53 (17-66)	.08
Sex, n (%)
Female	20 (32)	34 (36)	.54
Male	43 (68)	59 (63)
Diagnosis, n (%)
CLL	9 (14)	1 (1)	<.0001
Diffuse large B cell lymphoma	31 (49)	44 (47)
Follicular lymphoma	10 (16)	40 (43)
Mantle cell lymphoma	9 (14)	0
SLL	0	5 (5)
Other*	4 (6)	8 (9)
Disease type, n (%)
Indolent	21 (33)	48 (52)	.02
Aggressive	42 (67)	45 (48)
Time from diagnosis to HCT, n (%)
1-24 mo	22 (35)	19 (22)	.14
25-60 mo	21 (33)	41 (47)
>60 mo	20 (32)	27 (31)
Previous regimens, n (%)†
1-3	27 (43)	20 (22)	.004
4+	36 (57)	73 (78)
Status at HCT, n (%)
CR ± MRD	30 (48)	31 (33)	.06
PR/stable	12 (19)	33 (35)
Relapsed/refractory	20 (32)	29 (31)
Response to last regimen, n (%)
Chemosensitive	35 (57)	62 (67)	.04
Chemorefractory	27 (43)	23 (25)
Untested	0	8 (8)
HCT-CI, n (%)
0	7 (11)	35 (35)	.002
1-2	22 (35)	24 (34)
3+	34 (54)	22 (31)
Conditioning, n (%)
2 Gy TBI	13 (21)	14 (15)	.0009
2 Gy TBI + Flu‡	42 (67)	79 (85)
3 Gy TBI + Flu	8 (13)	0
Donor, n (%)
Related	23 (37)	55 (59)	.007
	Matched	22	54
	1-antigen MM	1	1
Unrelated	40 (63)	38 (41)
	Matched	30	23
	1-antigen MM	8	7
	1-allele MM	2	8
Previous autologous HCT, n (%)
No	15 (25)	42 (45)	.0009
Planned tandem HCT	24 (38)	13 (14)
Other autologous HCT	24 (38)	38 (41)
CD34 cell dose, × 106 cells/kg, median (range)	8.9 (2.8-22.7)	8.0 (1.0-26.3)	.28

^*Among RTX-treated patients: composite lymphoma, n = 1; lymphoplasmacytic lymphoma, n = 1; Hodgkin-like B cell lymphoma, n = 1; histiocytic sarcoma, n = 1. Among historical controls: Burkitt’s lymphoma, n = 1; marginal zone lymphoma, n = 2.

^†All patients received RTX treatment as part of a previous regimen.

^‡Includes 1 patient receiving clofarabine.

Figure 4.

Median RTX level by diagnosis group: NHL, n = 54; CLL, n = 9. Day −3, P = .17; day +10, P= .005; day +24, P= .03; day +38, P= .03; day +64, P= .0007; day +84, P= .003.

RESULTS

Patient Characteristics

Demographic information on the study patients (n = 63) contrasted with those of the historical controls (n = 93) is provided in Table 1. More than one-half (54%) of RTX-treated patients and 31% of historical controls had serious comorbidities, with HCT-CI score of ≥3 (P = .002). The median age was 56 years for RTX-treated patients and 53 years for controls (P = .08). Underlying diseases varied between the patient groups. RTX-treated patients had indolent NHL in 33% of cases and aggressive NHL in 67% of cases. In comparison, controls had more indolent disease (52%) and less aggressive disease (48%). Controls more frequently received grafts from related donors (59% versus 37%; P = .007). Among the 63 RTX-treated patients, 48 (76%) had previous autologous HCT to allogeneic HCT and 36 (57%) had ≥4 previous lines of therapy, whereas fewer historical controls had undergone prior autologous HCT (55%) and were more heavily pretreated (78%). All patients received RTX treatment in previous lines of therapy pretransplantation.

Outcomes

Relapse.

Thirty-six RTX-treated patients (57%) died, 16 of them from relapse. The median time to relapse was 5.2 months (range, .7 to 28.8 months). Figure 2A shows that among RTX-treated patients, those with indolent NHL had a lower 5-year relapse rate compared with those with aggressive NHL (19% versus 38%). In addition, chemorefractory patients had a slightly higher 5-year relapse rate than chemosensitive patients (33% versus 29%; P = .43), and patients with disease bulk had a comparable 5-year relapse rate as patients with no disease bulk at transplantation (32% versus 33%; P = .92). In these groups, the RTX-treated patients and historical controls had comparable 5-year relapse rates (38% versus 38%; 19% versus 19%; P = .7) (Figure 2A).

Figure 2.

Kaplan-Meir curves of study patients versus historical controls by disease diagnosis. (A) Relapse. (B) OS. (C) NRM. (D) PFS.

There was no statistically significant association between therapeutic RTX serum levels before transplantation and relapse (P = .80). A test for trend across patient groups by increasing RTX plasma concentrations (<25, 26 to 100, 101 to 200, and >200 μg/mL) also showed no association with relapse rate. This analysis was limited by the relatively small number of patients and events.

GVHD.

The incidences of grade II-IV acute GVHD (57% versus 56%; P= .65) and of grade III-IV acute GVHD (13% versus 17%; P = .45) were similar in the RTX-treated and historical control patients. The 5-year rate of chronic GVHD was higher in the RTX-treated patients, although the difference was not statistically significant (62% versus 47%, P = .70).

Among RTX-treated patients, there were no significant differences in the rates of acute GVHD grade II-IV or III-IV or of chronic GVHD by diagnosis (P= .86, .96, and .76, respectively). Pretransplantation therapeutic RTX serum level was not significantly associated with GVHD rate (grade II-IV acute GVHD, P = .68; grade III-IV acute GVHD, P = .17; chronic GVHD, P = .16).

OS, PFS, and NRM.

Figure 2B shows comparable 5-year OS rates of RTX-treated patients by disease diagnosis (aggressive NHL, 48%; indolent NHL, 45%). Chemosensitive RTX-treated patients had higher OS than those who were chemorefractory (54% versus 40%), but this difference was not statistically significant (P = .10). There were also no significant associations between disease bulk and OS, PFS, or NRM (P = .53, .92, and .97, respectively).

RTX-treated patients showed a trend toward higher OS and PFS rates compared with historical controls (aggressive NHL, 48% versus 36% [P = .24] and 41% versus 32% [P = .12], respectively), although neither difference was statistically significant (Figure 2B). NRM was significantly lower in RTX-treated patients compared with controls (27% versus 40%; P = .05) (Figure 2C).

RTX Pharmacokinetics

Median RTX plasma concentrations are shown in Figure 3. Therapeutic RTX levels were seen through day +84 after HCT, with the maximum RTX plasma median of 200.5 μg/mL seen at day +64, after the fourth dose. In most study patients, serum concentrations of RTX were detectable after the first infusion of i.v.-administered RTX, and levels increased with subsequent administrations.

Figure 3.

Median RTX levels.

Of note, 2 patients had sustained therapeutic RTX levels >25 μg/mL up to day +180 after HCT; both had aggressive disease and both remained in complete remission at the last follow-up (65 months and 73 months, respectively). In contrast, 3 patients did not achieve therapeutic RTX levels at day +84 after the fourth dose of RTX. All 3 of these patients had a diagnosis of refractory chronic lymphocytic lymphoma (CLL). Two died of treatment related causes, and 1 was alive in remission at last follow-up (86 months). One of the patients who died did not achieve a therapeutic RTX level at any point throughout the course of HCT.

Therapeutic serum RTX levels were achieved by 21 patients before the first dose of scheduled peri-HCT RTX. Median serum concentrations of RTX were significantly higher in patients treated with RTX before HCT than in patients who did not receive such treatment. This was noted after each of the 4 weekly infusions and up to day +84 after treatment (eg, on day +10, P < .0001; day +84, P = .03).

Patients with NHL achieved significantly higher median RTX levels than those with CLL (day +84, P = .003) (Figure 4). Patients with CLL had a lower 5-year relapse mortality compared with those with NHL (22% versus 33%), although no significant association was found between disease diagnosis and relapse rate or time to relapse (P = .57 and .45, respectively).

In addition, 9 patients with high tumor bulk at time of HCT achieved a therapeutic serum RTX level after 2 doses of RTX, but with a lower median serum RTX concentration compared with other study patients throughout the course of HCT.

RTX toxicity was low among the RTX-treated patients. One patient had a pulmonary reaction to RTX at day +24 with low oxygen saturation, which was quickly resolved by oxygen administration. Eight patients experienced late-onset neutropenia (absolute neutrophil count <500 cells/mm³), between days +56 and +95.

Genetic Determinants and Donor-Recipient FCγRIIIa Pairing

Most patients had unfavorable phenylalanine/phenylalanine (F/F) and valine/phenylalanine (V/F) FCγRIIIa-158 polymorphisms, and only 3 patients had favorable valine/valine (V/V) FCγRIIIa-158 polymorphism (59%, 36%, and 5%, respectively). In contrast, more donors presented with favorable polymorphisms; for example, there were twice as many V/V patients (F/F, 39%; V/F, 50% V/V, 11%; Supplementary Table S1). Of note, 1 patient did not engraft and thus had the same F/F FCγRIIIa-158 polymorphism after HCT. FCγRIIIa polymorphism pairing analysis by recipient genotype is described in Table 2.

Table 2.

FCγRIIIa Pairing by Recipient/ Donor Genotype.

FCγRIIIa pre-HCT	FCγRIIIa post-HCT	Patient, n	Relapse, n	5-yr Relapse, %	Overall 5-yr Relapse, %
By recipient genotype
F/F (n = 36)*	F/F	15	4	40	36
	V/F	16	5	38
	V/V	4	1	25
V/F (n = 22)	F/F	7	2	29	27
	V/F	12	4	33
	V/V	3	0	0
V/V (n = 3)	F/F	1	0	0	0
	V/F	2	0	0
	V/V	0	0	–-
No sample drawn (n = 2)	F/F	1	0	0	0
	V/F	1	0	0
	V/V	0	0	–-
By donor genotype
F/F (n = 24)†	F/F	15	4	40	38
	V/F	7	2	29
	V/V	1	0	0
V/F (n = 31)†	F/F	16	5	38	32
	V/F	12	4	33
	V/V	2	0	0
V/V (n = 7)	F/F	4	1	25	14
	V/F	3	0	0
	V/V	0	0	–-
No sample drawn (n = 1)	F/F	1	0	0	0
	V/F	0	0	–-
	V/V	0	0	–-

^*One patient with F/F FCγRIIIa Pre-HCT did not have a post-HCT FCγRIIIa sample.

^†Two patients had no pre-HCT FCγRIIIa sample among post-HCT F/F and V/F FCγRIIIa polymorphisms.

F/F recipients had similar 5-year relapse rates with different donor pairings (F/F, 40% V/F, 38%; V/V, 25%). Similarly, V/F recipients experienced comparable relapse rates with F/F and V/F donor pairings (29% and 33%, respectively), but no relapse was observed with V/V donor pairing. V/V recipients did not experience relapse with each donor pairing (5-year relapse 0%).

In an analysis of FCγRIIIa polymorphism pairing by donor genotype (Table 2), patients with unfavorable FCγRIIIa polymorphisms after engraftment who had mostly pre-HCT unfavorable polymorphisms had similar 5-year relapse rates (F/F, 38%; V/F, 32%). Only 1 V/V patient relapsed, who had F/F polymorphism before HCT.

DISCUSSION

The present phase II study was based on the hypothesis that the previously reported 21% to 41% 3-year incidence of disease recurrence among patients with relapsed or refractory NHL treated with allogeneic HCT [12,13] could be reduced and outcomes improved by peritransplantation RTX. This hypothesis in turn was based on the assumption that RTX would eliminate tumor cells both by direct antitumor activity and via an enhanced graft-versus-tumor effect [30]. Disappointingly, our data do not confirm our hypothesis. We found that RTX failed to reduce relapse rates in a comparison of findings in the current series of 63 patients and in 93 previously reported patients who were not given RTX but underwent transplantation using the same standard conditioning and GVHD prevention regimens. Nevertheless, conclusions from such historical comparisons cannot be considered definitive and can be only hypothesis-generating. The limitations of historical comparisons are exemplified by the finding of lower NRM in the current series compared with the previous series, likely due to gradual improvements in supportive care, as we reported previously [31]. In addition, our conclusions are restricted by small sample size.

The pharmacokinetics of RTX elimination in previous reports were highly variable and remain incompletely described after allogeneic HCT. Our negative findings cannot be explained by insufficient RTX serum levels, because levels of >25 μg/mL were consistently achieved after the first RTX dose and increased further with subsequent dosing up to day +84. Consistent with recent studies, patients with CLL had lower median RTX serum concentrations than those with NHL [32,33]. RTX levels were not correlated with relapse.

In contrast to our present findings, a number of previous studies using peritransplantation RTX have shown encouraging survival outcomes. However, given the retrospective nature, outcome data need to be interpreted with the same caveats that we applied in the present. A preliminary retrospective analysis of Center for International Blood & Marrow Transplant Research registry data showed a lower rate of 5-year relapse in patients receiving RTX during the 6 months before allogeneic HCT compared with those who did not receive RTX [34]. Two recent retrospective Center for International Blood & Marrow Transplant Research reports of reduced-intensity conditioning, peritransplantation RTX, and allogeneic HCT for follicular lymphoma and B cell NHL described 3-year relapse rates of 15% and 24%, respectively [35,36]. Other studies were prospective, but none involved randomized comparisons. Khouri et al [37] reported only 2 relapses among 47 relapsed follicular lymphoma patients undergoing non-myeloablative conditioning with high-dose peritransplantation RTX, largely with HLA-matched related donors; absent a comparator, it is unclear whether the low relapse incidence can be attributed to RTX. Another group reported a 2-year event-free survival of 72% in patients with relapsed or primary refractory B cell NHL undergoing nonmyeloablative conditioned allogenic HCT with peritransplantation RTX. The patients in that study had low HCT-CI score (median, 1) and a limited number of previous therapies [38].

Patients with FCγRIIIa F/F and V/F polymorphisms experienced a higher relapse rate versus those with V/V polymorphism, although this difference was not statistically significant by both recipient and donor FCγRIIIa polymorphisms. This is in line with previous reports indicating an association between FCγRIIIa F/F and V/F polymorphisms and poor clinical response to RTX [16,17,39]. Although the patient numbers were relatively small, we hypothesize that these subsets of patients might benefit from pairing by V/V FCγRIIIa polymorphisms in addition to HLA matching in the setting of allogeneic HCT. In contrast, no differences in 1-year relapse were observed based on donor FCγRIIIa polymorphism in a recently published phase II study of NMA related-donor allogeneic HCT with high-dose post-transplantation cyclophosphamide and post-transplantation RTX for patients with B cell lymphomas (18% V/V, 23% V/F, 17% F/F) [40]. However, this might be attributable to the post-HCT therapy or the short follow-up of this study.

Previous studies have shown that low-affinity polymorphisms of the FCγRIIIa receptor expressed predominantly on natural killer cells, impacted effective RTX antibody-dependent cell-mediated cytotoxicity in vitro [14], and were associated with a low response rate to RTX therapy. Notably, the genotype distributions of unfavorable FCγRIIIa F/F polymorphisms were more prevalent in our study patients than in both healthy donors and in most previously published patients with NHL (59% versus 41% and ≤46%, respectively) [14,16,17,39–48]. This finding might provide another explanation for the relatively high relapse rates in our study patients despite the addition of RTX.

A number of investigators have postulated a contribution of antibodies generated by donor B lymphocytes to the development of both acute and chronic GVHD, and that RTX might be beneficial by suppressing their production [49–51]. A compelling argument against this mechanism has been that, with the exception of 1 report on an antibody against Y chromosome-associated antigens [52], no one has succeeded in raising antibodies against minor histocompatibility antigens in various animal species or in humans. Alternately, RTX-mediated B cell depletion might modulate T cell alloreactivity [53–55]. In apparent agreement with the B cell postulates, several groups have reported successful treatment of acute GVHD and established, refractory chronic GVHD with RTX [51,56,57], although responses were generally limited to cutaneous and musculoskeletal manifestations. In contrast to these reports, current patients experienced virtually identical rates of acute GVHD compared with previous patients, even though therapeutic RTX levels were measured throughout the time period of risk for developing acute GVHD. This finding supports the notion that acute GVHD is a T cell-mediated rather than a B cell-mediated immunologic complication. Given the clinical benefit reported for treating established chronic GVHD, we hypothesized that administration of RTX in the early post-transplantation period might have a prophylactic effect, reducing the incidence of chronic GVHD. Nevertheless, there was no statistically significant difference in the incidence of chronic GVHD among current RTX-treated patients compared with previous patients not given RTX.

In conclusion, incorporation of RTX in the NMA conditioning regimen for allogeneic HCT was not effective in producing survival gains among patients with relapsed or refractory NHL relative to historical controls in this prospective clinical trial. A randomized clinical trial testing a slightly different dosing regimen of the anti-CD20 antibody obinutuzumab is nearing completion and may further clarify conflicting results. Our data suggest, but do not prove, that FC receptor polymorphisms may play a role in determining which patients benefit from peritransplantation RTX. This approach should be further validated using favorable FCγRIIIa polymorphisms in larger cohorts.