Nonmyeloablative, HLA-mismatched unrelated peripheral blood transplantation with high-dose posttransplantation cyclophosphamide

Abstract

Background:

High-dose posttransplantation cyclophosphamide (PTCy) is an effective platform for prevention of severe graft-versus-host disease (GVHD) after allogeneic bone marrow (BM) transplantation with mismatched unrelated donors (mMUD). Prior studies evaluating PTCy with mMUD favored BM allografts over peripheral blood stem cell transplantation (PBSCT) due to concerns that PBSCT may be associated with an increased risk of acute and chronic GVHD. In addition, haploidentical PBSCT is associated with high rates of cytokine release syndrome (CRS), and this was another concern with mMUD PBSCT.

Objective:

Determine the feasibility and safety of using mMUD PBSCT with PTCy as GVHD prophylaxis.

Study Design:

Patients who received mMUD PBSCT using a PTCy-based GVHD prophylaxis at Johns Hopkins Hospital as part of a prospective clinical trial of mMUD and non–first-degree relative haploidentical transplantation with PTCy (NCT01203722) were included. All patients underwent T cell–replete PBSCT between November 2012 and August 2020. Statistical analyses were performed using the Kaplan-Meier method and proportional subdistribution hazard regression model for competing risks.

Results:

Of the 29 patients in the study, the median age was 54 years, with 10 patients (34%) being age ≥60 years. Nineteen grafts (66%) were matched for 9/10, 6 (21%) for 8/10, and 4 (14%) for 7/10 HLA loci. No primary or secondary graft failure occurred. The median time to neutrophil recovery (≥500/µL) and platelet recovery (≥20,000/µL) were 17 and 28 days, respectively. Full donor chimerism was achieved in all patients by day 60. The cumulative incidence (CuI) of grade II-IV acute GVHD at 180 days was 15% (90% CI: 3%, 26%). There were no cases of severe chronic GVHD, 3 cases of mild chronic GVHD, one case of moderate chronic GVHD. The CuI of nonrelapse mortality (NRM) was 7% (90% CI: NA, 18%) at one year. Eighteen (62%) patients experienced mild CRS (grade 1–2) and one (3%) experienced severe CRS (grade 3–5). The CuI of relapse at one year was 29% (90% CI: 8%, 50%). The OS at one year was 93% (90% CI: 85%, 100%). PFS at one year was 64% (90% CI: 46%, 88%). GVHD-free relapse-free survival at one year was 41% (90% CI: 23%, 73%). Chronic GVHD-free relapse-free survival at one year was 64% (90% CI: 46%, 88%).

Conclusion:

mMUD PBSCT using PTCy-based GVHD prophylaxis is safe and feasible. All patients engrafted and rates of NRM (7%) and acute GVHD (15%) at one year were low. There was only one case (3%) of severe CRS. Compared with previously published outcomes, mMUD PBSCT using PTCy-based GVHD prophylaxis has a safety and efficacy profile that may not be different from PBSCT from matched donors. These results further solidify that all patients who require blood or bone marrow transplantation should be able to find an acceptable donor.

Introduction

High-dose posttransplantation cyclophosphamide (PTCy) has emerged as an effective platform for prevention of severe graft-versus-host disease (GVHD) after allogeneic blood or marrow transplantation [1–3]. Historically, increasing numbers of human leukocyte antigen (HLA) mismatches were associated with an increased risk for GVHD and graft rejection [4, 5]. However, with PTCy-based GVHD prophylaxis, the use of HLA-haploidentical (haplo) donors or matched unrelated donors (MUD) leads to outcomes indistinguishable from those achieved with HLA-matched sibling transplantation [6–10]. Further studies showed that the approach could be extended to partially mismatched unrelated donors (mMUD) and haplo related donors who are not first-degree relatives [11–15]. Recent prospective studies of mMUD transplantation used bone marrow (BM) allografts exclusively due to concerns for increased rates of acute and chronic GVHD with peripheral blood stem cell transplantation (PBSCT) [11, 12, 16]. However, data are limited with mMUD PBSCT. A comparator group of mMUD PBSCT identified through the National Marrow Donor Program/Be the Match (NMDP/BTM) initiative showed no differences in overall survival, GVHD, relapse rates, or non-relapse mortality (NRM) for mMUD PBSCT in comparison with the prospectively studied mMUD BM transplantations [13]. Likewise, a retrospective study comparing PTCy mMUD vs MUD using primarily PBSCT (92%) found no difference in acute GVHD, NRM, or survival. However, there was a lower incidence of survival free of chronic GVHD and relapse in the PTCy mMUD group at 2 years [15]. Finally, a soon-to-be-published, prospective single-center study of 39 mMUD PBSCT had encouraging outcomes with 1-year overall survival (OS) and GVHD-free relapse-free survival (GRFS) of 87% (95% CI: 71–94) and 68% (95% CI: 51–81), respectively [14].

The ability to safely engraft HLA-mismatched donors with improved GVHD prophylaxis has increased our awareness of cytokine release syndrome (CRS) 1 to 3 days after allograft infusion. Our initial experience with haploidentical PBSCT was associated with cytokine release syndrome (CRS) in 90% of recipients and severe CRS in in 17% [17]. Rates of CRS may be higher with PBSCT compared with BM allografts because PBSCT have an approximately 8-fold increase in the number of T lymphocytes [18]. Therefore, CRS was anticipated to be a problem with mMUD PBSCT. Despite concerns for increased GVHD and CRS with mMUD PBSCT, there are advantages of using PBSCT over BM allografts. Certain diseases, such as myelodysplastic syndrome or myeloproliferative neoplasms may have improved engraftment and possibly enhanced antitumor alloreactivity with PBSCT [19]. In addition, stem cell quality and quantity may be more reliable with PBSCT [20]. In light of these advantages, we conducted a prospective trial to address the safety and feasibility of mMUD PBSCT with PTCy.

Materials and Methods

A prospective clinical trial of mMUD and non–first-degree relative haploidentical transplantation with PTCy was conducted at The Johns Hopkins Hospital (NCT01203722). Results from patients who received mMUD BM grafts and patients who received non-first degree relative haplo-related BM grafts have previously been published [11, 12]. Twenty-nine consecutive patients who received mMUD PBSCT and PTCy are presented here. Eligible patients were ages 0.5 to 75 years, lacked at least an available HLA haplo first-degree related donor or a MUD, and had adequate organ function as previously published [11]. The 5 HLA matching loci were A, B, C, DRB1, DQB1. Other loci evaluated were DRB3, DRB4, DRB5, DQA1, DPB1. Donors having the fewest mismatched loci were prioritized for younger age[21], major ABO compatibility, and matched cytomegalovirus IgG serostatus.

With the exception of patients over the age of 70, no patients were routinely admitted to the hospital as part of the transplant process. Patients over the age of 70 were typically admitted from the beginning of the prep regimen until engraftment.

All patients received T cell–replete PBSCT between November 2012 and August 2020. Conditioning consisted of cyclophosphamide (14.5 mg/kg IV on days −6 and −5), fludarabine (30 mg/m² IV on days −6 to −2, with an adjustment for renal function), and total body irradiation (200 cGy on day −1). GVHD prophylaxis consisted of high-dose PTCy (50 mg/kg IV daily on days 3 and 4) with mesna, mycophenolate mofetil 15 mg/kg by mouth three times a day (maximum 3 g/day) from day 5 to day 35, and sirolimus from day 5 to day 180 (target trough, 5–12 ng/mL) [11]. In the absence of GVHD, sirolimus was discontinued without a taper. One patient received tacrolimus initially due to a recent stroke and concern for sirolimus-associated hyperlipidemia. Filgrastim began on day 5 and was continued through neutrophil engraftment.

Neutrophil engraftment was defined as the first of 3 consecutive laboratory values on different days with an absolute neutrophil count ≥ 500/mm³. Platelet engraftment was defined as the first of 3 consecutive laboratory values on different days with a platelet count ≥ 20,000/mm³ without platelet transfusion in the preceding 7 days. Graft failure was defined as ≤5% donor chimerism in peripheral blood or bone marrow after day 60 in the absence of bone marrow relapse. Donor chimerism was evaluated at approximately days 30, 60, 180, and 365, CRS was graded according to published criteria[22]. The database was locked on November 20, 2020.

Disease-free survival (DFS), overall survival (OS), GVHD-free relapse-free survival (GRFS), and chronic GVHD-free relapse-free survival (CRFS) were estimated by the Kaplan-Meier method [23]. DFS failure was defined as the time lapse from transplant until relapse or death. GRFS failures were defined as the time from transplant until grade III-IV acute GVHD, systemic treatment of chronic GVHD, or a DFS failure[23]. CRFS failures included moderate or severe chronic GVHD or DFS failure[24]. The severity of acute GVHD was determined using the 1994 consensus criteria on acute GVHD grading, and the severity of chronic GVHD was determined using the NIH consensus criteria [25, 26]. Median follow-up was calculated by the reverse Kaplan-Meier method. Cumulative incidence (CuI) of relapse, non-relapse mortality (NRM), count recovery, GVHD, and the start of immunosuppression for acute GVHD were estimated with the proportional subdistribution hazard regression model for competing risks [27]. Relapse was a competing risk for NRM. Relapse and graft failure were competing risks for GVHD.

Results

Baseline patient characteristics for these 29 patients are shown in Table 1. The median age was 54 years (range: 22–74) with 10 patients (34%) being age ≥60 years. Of note, no patients eligible for this protocol were turned away for lack of a suitable donor. With regard to allografts, 19 (66%) were matched for 9/10, 6 (21%) for 8/10, and 4 (14%) for 7/10 HLA loci (Table 2). The HLA-loci listed in the mismatch columns of Table 2 represent one antigen mismatch unless designated as allele or (2) antigen mismatch. The estimated median follow-up was 8 months.

Table 1:

Patient and Transplant Characteristics

Patients		N=29	(Range), %
	Median patient age in years	54	(22–74)
	Male Sex	14	48%
	Diagnosis
	Acute myeloid leukemia	11	38%
	MDS or MPN	6	21%
	Non-Hodgkin Lymphoma	5	17%
	Acute lymphooblastic leukemia	4	14%
	Chronic myeloid leukemia	2	7%
	Histiocytic sarcoma	1	3%
	CR at transplant	23	79%
	Post-HSCT maintenance therapy*	5	17%
Unrelated Grafts
	HLA matches
	9/10	19	66%
	8/10	6	21%
	7/10	4	14%
	ABO incompatibility
	Major	2	7%
	Minor	8	28%
	Both major and minor	0
	CMV serostatus
	CMV IgG mismatch	8	28%
	Patient CMV seropositivity	23	80%
	Donor age, yr, median (range)	27	(20–43)
	Female donor/male recipient	5	17%
	Cell dose infused, median (IQR)
	Total nucleated cells x 10^8/kg	10.7	(8.3,14.4)
	CD34+ cells x 10^6/kg	10.7	(7.8,17.2)
	CD3+ cells x 10^7/kg	39	(29,49)

CMV: cytomegalovirus, CR: complete remission, DRI: Disease Risk Index, IgG: immunoglobulin G, IQR: interquartile range, MDS: myelodysplastic syndrome, MPN: myeloproliferative neoplasm

*gilteritinib (x2), sorafenib, blinatumomab, APR-246 with azacitadine

Table 2:

HLA matching

HLA match 5 loci	Mismatches 5 loci	Mismatches other loci
7/10	A, DRB1 allele, DQB1	DQA1, DPB1
7/10	C allele, DRB1, DQB1	DRB3, DQA1, DPB1 (2)
7/10	A allele, DRB1, DQB1	DRB4, DQA1, DPB1 (2)
7/10	B, C, DRB1 allele	DPB1
8/10	DRB1, DQB1	DQA1, DPB1
8/10	DRB1, DQB1	DRB3, DPB1, DQA1 allele
8/10	DRB1, DQB1	DRB3, DQA1 allele, DPB1
8/10	DRB1, DQB1	DPB1 (2)
8/10	B, C	DPB1
8/10	DRB1, DQB1	DRB3, DQA1, DQA1
9/10	DRB1	DQA1 allele, DPB1
9/10	B	DPB1 (2)
9/10	B	DRB3, DPB1
9/10	A	-
9/10	B	DPB1
9/10	B	DPB1
9/10	B allele	DPB1
9/10	B allele	-
9/10	C	-
9/10	A	DPB1
9/10	-	DPB1
9/10	A	-
9/10	A	DPB1
9/10	A	DPB1
9/10	B	DPB1
9/10	A	DPB1
9/10	A	-
9/10	A	DPB1
9/10	A	DPB1

Engraftment and GVHD

No primary or secondary graft failure occurred. All evaluable patients achieved neutrophil and platelet engraftment. One patient was unevaluable for platelet recovery due to death. The median time to neutrophil engraftment was 17 (range, 13–28) days. The median time to ≥ 20,000/mm³ platelet recovery was 28 (range, 16–60) days. The median time to ≥ 50,000/mm³ platelet recovery was 29.5 (range, 19–153) days with 90% by day 60. Full donor chimerism was achieved in 29 (100%) patients by day 60 as assessed by unsorted peripheral blood cells.

The CuI of acute GVHD grade II-IV at days 100 and 180 were 11% (90% CI: 1%, 21%) and 15% (90% CI: 3%, 26%) respectively (Figure 1). There were no cases of grade IV acute GVHD. All acute GVHD was limited to the skin (7 cases of grade I, 3 cases of grade II, and 2 cases of grade III). The CuI of acute GVHD grade II-IV at day 180 was not different among the 10 patients with ≥ 2 mismatched HLA loci (20%) versus the 19 patients with 1 mismatched HLA locus (12%) (Figure 1). The CuI of the start of immunosuppression for treatment of acute GVHD was 33% (90% CI: 18%, 49%) at six months and 46% (90% CI: 21%, 70%) at 12 months.

Figure 1: Probability of aGVHD and chronic GVHD.

(A) Probability of acute grade II-IV GVHD by match. High match = 9/10 match. Low match < 9/10 match. There was no difference in CuI of aGVHD at day 180 among the 10 patients with ≥ 2 mismatched HLA loci [20%, 90% CI (NA, 42%)] versus the 19 patients with 1 mismatched HLA locus [12%, 90% CI (NA, 25%)].

(B) Probability of chronic GVHD and acute GVHD in all patients. The CuI of acute GVHD grade II-IV at day 180 was 15% (90% CI: 3%, 26%) and CuI of chronic GVHD at 12 months was 23% (90% CI 1%, 46%). aGVHD: acute graft versus host disease, CuI: cumulative incidence, CI: confidence interval

There have been no cases of severe chronic GVHD according to NIH consensus criteria [25]. There were 3 cases of mild chronic GVHD of the skin. One case was successfully treated with prednisone. The other 2 cases required immunosuppression until the time of patient death or last follow up. There was one case of mild chronic GVHD of the skin that progressed to moderate involving the eyes, mouth, and skin. This case was successfully treated with prednisone and tacrolimus. The CuI of chronic GVHD at 6 and 12 months were 4% and 23%, respectively.

Morbidity and NRM

The CuI of NRM was 7% (90% CI: NA, 18%) at six and twelve months (Figure 2). Of 2 patients experiencing NRM, one died of cardiac arrest at day 46 and the other died from multiorgan failure and multiple infections including central nervous system toxoplasmosis at day 115. By day 100, 22 patients (76%) required at least one unplanned admission for toxicity, including neutropenic fever, bacteremia, CRS, cytomegalovirus viremia, BK virus, Clostridioides difficile colitis, dehydration, and pleural effusions. Nine (31%) patients were transitioned to tacrolimus due to toxicity attributed to sirolimus (pleural effusions, fevers, stomatitis, diarrhea, and hyperlipidemia).

Figure 2: Probabilities of Non-relapse Mortality and Relapse.

The CuI of non-relapse mortality at 12 months was 7%, 90% CI (NA, 16%). The CuI of relapse at 12 months was 29%, 90% CI (8%, 50%).

CuI: cumulative incidence, CI: confidence interval

Cytokine Release Syndrome

Of the 29 patients transplanted, 18 (62%) experienced mild CRS (grade 1–2) and one (3%) experienced severe CRS (grade 3–5). The patient with severe CRS was mismatched for one DRB1, one DRB3, one DQA1, one DQB1, and one DPB1 allele. The patient developed hypotension requiring 2 vasopressors. All cases of CRS, including the one severe case, resolved after administration of PTCy between days 3 and 4.

Relapse and survival

The CuI of relapse was 18% (90% CI: 4%, 32%) at six months and 29% (90% CI: 8%, 50%) at twelve months (Figure 3). The median OS has not been reached. The OS at one year was 93% (90% CI: 85%, 100%). The median DFS has not been reached and DFS at one year was 64% (90% CI: 46%, 88%). At this time, DFS and CRFS outcomes are equivalent (Figure 3).

Figure 3:

Kaplan Meir estimates of overall survival, disease-free survival, and GVHD-free relapse-free survival. OS at 1 year was 93%, 90% CI (85%, 100%). DFS at 1 year was 64%, 90% CI (46%, 88%). GRFS at 1 year was 41%, 90% CI (23%,73%).

OS: overall survival, DFS: disease-free survival, GRFS: acute GVHD-free-relapse-free-survival, CI: confidence interval.

The median GRFS was 10.6 months. GRFS at six months was 72% (90% CI: 58%, 88%), and at twelve months it was 41% (90% CI: 23%, 73%) (Figure 3). The median CRFS was not reached at the time of this analysis. CRFS at six months was 74% (90% CI: 61%, 91%), and at twelve months it was 64% (90% CI: 46%, 88%).

Discussion

This study demonstrates the feasibility and safety of using mMUD donors and extends those results to include mMUD PBSCT with PTCy. With regard to feasibility, we found that a donor could be identified for every patient and we did not have to resort to 6/10 or 5/10 matches in any instance. With regard to safety there have been two broad concerns: GVHD and CRS. We observed little acute GVHD and no gut GVHD—a finding that is similar to our previous report of mMUD with PTCy using BM as a graft source where there was 1 case of ungradable GI GVHD (9/10 mismatched donor) among 20 patients [11]. Similarly, we observed only a single instance of severe CRS and note that severe CRS occurred in 17% of haplo PBSCT recipients [17]. CRS may be more common in patients with a more extensive mismatch and we note that mismatch at DRB1 and DPB1 has been linked to fever following transplant with haploidentical donors [17, 28]. However, with only a single case, we hesitate to make inferences.

Previous studies have suggested that PTCy has advantages over an alternative approach involving anti-thymocyte globulin (ATG) [29, 30]. In a single center retrospective study of 76 mMUD transplant (96% with PBSCT) there was no difference in survival, relapse or GRFS but the incidences of acute GVHD, chronic GVHD, and NRM were lower in the PTCy group [29]. Furthermore, PBSCT using PTCy demonstrated a lower incidence of severe acute GVHD and superior survival in matched-pair analysis of registry data [30].

Rates of relapse (29%) were similar to that seen in prior s with non-myeloablative transplantation [12, 31, 32]. With the goal of reducing the risk of relapse, studies evaluating post-transplantation maintenance are underway. In the current trial, 5 patients received post-transplantation maintenance therapy on separate clinical trials: two gilteritinib, one sorafenib, one blinatumomab, and one APR-246 (targeting TP53) with 5-azacitadine).

Patients on this trial began immunosuppression on day 5 after PBSCT with mycophenolate mofetil until day 35 and sirolimus until day 180, or longer if there was evidence of GVHD. In other recent investigations, our group explored shortened immunosuppression in patients receiving BM or PBSCT [19, 33] from related haploidentical donors, but we have not explored shortened duration immunosuppression in the context of BM or PBSCT from mMUDs. While sirolimus was the preferred immunosuppressant for this trial, when concerns arose about possible adverse effects of sirolimus, tacrolimus substitution was allowed. Ten of 29 patients (34%) were started on or transitioned to tacrolimus.

These results demonstrate that mMUD PBSCT using PTCy has a safety and efficacy profile that may not be different from PBSCT from matched donors [34, 35]. There were high rates of engraftment (100%), and low rates of acute GVHD (15%) and non-relapse mortality (7%). CRS occurred in one patient and appeared less common than with use of haploidentical PBSCT, perhaps because 9/10 donors were able to be identified for most patients. Thus, where the underlying disease, donor preference, or logistical considerations suggest that PBSCT may be a more appropriate allograft source than BM, there is no compelling contraindication to its use when considering mMUDs.

Highlights.

• PBSC transplantation from mismatched unrelated donors is safe and effective

• Engraftment and GVHD rates are similar to those with matched donor transplantation

• Severe cytokine release syndrome was uncommon