“Incidence and significance of donor-specific antibodies in haploidentical stem cell transplantation”

PGF is a devastating complication after allogeneic transplant. We retrospectively analyzed our haploidentical transplant registry to report the incidence and impact of DSA and anti-HLA on engraftment. 107 patients were identified. Median recipient-age of 22, median donor-age of 31. Sixty-two patients had AML (58%), 29 had ALL (27%), 16 (15%) had other malignancies. Sixty-one recipients (57%) had positive anti-HLA, 56 of them had the DSA results available, of these 17 patients had DSAs (15% of the total number of patients, or 28% of patients who have anti-HLA antibodies). The median cumulative MFI was 2062. Sixty-three percent of the DSA were against class-II HLA antigens. The OS, CIR, aGvHD, and cGvHD did not differ between patients with and without anti-HLA antibodies, nor between patients with and without DSA. The gender of the recipient and donor, as well as the gender mismatch between recipient and donor, were statistically associated with the incidence of anti-HLA antibodies. Three patients only developed GF (2.8%), one was primary (0.9%) and the other two secondary GF (1.9%). None of the GF cases was in patients with anti-HLA antibodies or DSA. The presence of anti-HLA or DSAs did not affect the outcomes including the incidence of PGF.


INTRODUCTION
With the better and easier control of the bidirectional alloreactivity, haploidentical stem cell transplantation (haplo-HSCT) from a first-degree related haplotype-mismatched donor (siblings, children, and parents) became a standard option in the absence of a matched donor [1]. The availability of multiple motivated family donors, the ease of donor procurement and the lower cost compared to the use of matched unrelated donors made it an attractive option and many centers adopted haplo-HSCT as a default alternate if no matched sibling donor is available [2]. The turning point in the history of HLA-mismatched transplant was the use of unmanipulated grafts with post-transplant cyclophosphamide. This major accomplishment circumvented the issues of ex-vivo T cell depletion and using mega doses of CD34 selected grafts, which were problematic and limited the wide adoption of this modality [3]. The human major histocompatibility complex or the human leukocyte antigen (HLA) complex encodes a large number of genes [4]. The exposure to non-self HLA antigens triggers the development of anti-HLA antibodies. These antibodies are frequently caused by transfusions, pregnancy or previous transplant. When these antibodies target specifically the donor HLA, these are called donor-specific antibodies (DSA). In the setting of HLA-mismatched transplantation, the prevalence of anti-HLA antibodies can be up to 40% and the prevalence of DSA can be up to 25%. The presence of DSA correlates with poor graft function and graft failure (GF) in the setting of mismatched transplantation [1,[5][6][7][8][9][10][11][12][13][14][15][16][17]. Despite the availability of multiple potential donors, transplant physicians still face scenarios where the recipient has DSA against the appropriate donor. If alternative donors are not available, desensitization to remove the alloantibodies from the serum of recipient ahead of transplant is recommended to minimize the risk of GF. We herein report the incidence and impact of DSA and anti-HLA from a single-center cohort of myeloablative haploidentical transplant for malignant diseases.

PATIENTS AND METHODS Patients
107 consecutive patients with malignant hematologic disorders received a myeloablative haplo-HSCT at King Faisal Specialist Hospital and Research Center (KFSH&RC) between January 2013 and December 2020. Eligibility criteria for haplo-HSCT include left ventricular ejection fraction >50%, serum creatinine <130 mmol/L, serum bilirubin <35 mmol/L, liver enzymes <2 times the upper limit of normal value, serum albumin >30 gm/L, pulmonary function test values >50% predicted, Karnofsky score >70%, signed informed consent for both donor and recipient. Exclusion criteria include failure to meet any of the eligibility criteria, absence of donor or recipient consent forms, human immunodeficiency virus positivity, major psychiatric disorder, and a positive pregnancy test. The standard haplo-HSCT platform at KFSH&RC uses rabbit antithymoglobulin (1-1.5 mg/kg/ day on day −3 and day −2) post-transplantation cyclophosphamide (50 mg/kg/day on day +3 and +5) and calcineurin inhibitors starting day +4 for graft versus host disease (GvHD) prophylaxis. The standard stem cell source is primed bone marrow (BM). The standard conditioning is thiotepa (5 mg/kg/day on day −8 and −7), busulfan (3.2 mg/kg/day from day −6 to day −4) with fludarabine (50 mg/m 2 from day −6 to day −4) for myeloid malignancies, and fludarabine (30 mg/m 2 from day −7 to day −4) with 10 gray TBI (2 gray twice daily on day −10 and day −9 and once on day −8) for lymphoid malignancies. The standard graft source is primed BM. The Institutional Review Board approved the study.

Methods
HLA typing. HLA-A, -B, -C, -DR, -DQ typing of all recipients and donors was done by DNA molecular typing using reverse sequence specific oligonucleotide probes according to the manufacturer's instructions (One Lambda, Canoga Park, CA, USA) and/or Next Generation Sequencing (Omixon, Budapest, Hungary). In patients with anti-DP antibodies, the corresponding donor was typed for HLA-DP using the same platforms.
Detection of HLA-antibodies. Patients' sera were screened for the presence or absence of HLA antibodies. Sensitized patients were subsequently tested for class I (HLA-A, -B, and -C) and class II (HLA-DR, -DQ, and -DP) HLA antibodies using single antigen beads (SAB) on a Luminex platform according to the manufacturer's instructions (One Lambda, Canoga Park, CA, USA). Patients with a DSA level above 500 mean fluorescence intensity (MFI) were considered positive for DSA.
C1q assay. Serum samples from patients with positive DSA were tested for C1q-binding ability using SAB-C1q assay (One Lambda, Canoga Park, CA, USA) according to the manufacturer's instructions. For DSA and C1qfixing DSA, a MFI > 500 was defined as positive.
Flow crossmatch. Three-color flow cytometry crossmatches were performed. Patient historic and current serum were incubated with donor's pronase-treated T and B cells for 30 min. T and B lymphocytes were stained for 10 min with phycoerythrine (PE) and peridinin-chlorophyll-protein (PerCp) conjugated monoclonal antibodies specific for CD3 and CD19, respectively. The presence of bound antibodies was determined using a fluorescein isothiocyanate-conjugated anti-human IgG. Flow cytometric analysis was performed using a FACSCanto II instrument and FACSDiva software (BD, PharMingen, San Jose, CA). Flow crossmatch results were analyzed based on median channel shift (MCS) over background. A positive crossmatch was reported if the MCS value was more than 2.5 standard deviations of control serum.
Desensitization. With the limited data that was available when we started our Haplo-HCT program, the first patient with DSA was desensitized with IVIG only and the second patient was desensitized despite having relatively low cumulative MFI (3983). With the emergence of literature, we adopted the following protocol as our standard: alternate day one-volume plasma exchange (3 sessions) using albumin replacement to start a week before starting conditioning therapy; one day after the third-and-last plasma exchange session, one dose of intravenous immunoglobulin (IVIG) 1 g/kg is given followed by one dose of rituximab 375 mg/m 2 the next day after IVIG. We reserved buffy coat addition to patients with persistent cumulative MFI above 5000 after receiving the plasma exchange, IVIG and rituximab.
Buffy coat. The buffy coat was prepared from the donor on day −2 of transplant. Whole blood unit was separated by centrifugation within 8 h of collection. The plasma component then underwent a second centrifugation to separate the buffy coat. The final buffy coat was then crossmatched, irradiated, and infused on day −1 of transplant.
Definitions. Myeloid engraftment was defined as the first of 3 consecutive days with an absolute neutrophil count (ANC) > 0.5 × 10 9 /L. Platelet engraftment was defined as platelet count maintained above 20 × 10 9 /L without platelet transfusion for 7 days. Primary GF was defined as failure to achieve an ANC > 0.5 × 10 9 /L for three consecutive days by day 28. Secondary GF was defined as subsequent development of persistent ANC < 0.5 ×10 9 /L after achieving adequate engraftment with BM hypocellularity and loss of donor chimerism, or loss of donor chimerism in the setting of autologous recovery or progressive loss of chimerism with cytopenias necessitating another infusion of stem cells to reverse it. Acute GvHD (aGvHD) and chronic GVHD (cGvHD) were defined according to the standard criteria. Non-relapse mortality (NRM) was defined as mortality from any cause without evidence of relapse. Therapy related mortality was defined as mortality from any cause without evidence of relapse within the first 100 days post-transplant. Overall survival (OS) was calculated from time of transplant until death from any cause. Event free survival was calculated from time of transplant until death or GF. The TNC (total nucleated count) and CD34 (cluster of differentiation) doses were calculated as the total infused number of TNC or CD34 cells per kilogram (Kg) of recipient weight.
Statistics. For categorical variables, frequencies and percentages were used to summarize the patients' characteristics. For continuous variables, medians and interquartile ranges were used. The Chi-square test was used to compare categorical variables, and the Mann-Whitney U test was used to compare continuous variables. Acute GvHD (aGvHD) and cGvHD were defined as the time required to develop aGvHD and cGvHD, respectively, and were summarized using cumulative incidence, with death without GvHD as competing events. Cumulative incidence of relapse was defined as the time to relapse accounting for NRM as a competing event and was summarized by cumulative incidence. Using the Gray test, the cumulative incidence curves of aGvHD, cGvHD, relapse, and NRM were compared between patients with anti-HLA antibodies and those without, and between DSA patients and those without. OS was defined as time of death from any cause. Patients who were alive were censored at the time of last contact. Probabilities of OS was summarized using Kaplan-Meier estimator with variance calculated using Greenwood formula. Survival curves were compared using log-rank test. P value < 0.05 was considered significant. Analysis was conducted using RStudio. Version 1.4.1106 © 2009-2021 RStudio, PBC.
The median age of the 17 patients who had DSA was 36 (range: 8-48), 10 were females and seven males. Recipient/donor pairs were concordant in 12 cases. Nine patients had AML, three had MDS, three had ALL, one had CML and one had bilineage leukemia. Twelve recipient/donor pairs were ABO matched. Twelve recipients had a BM graft. The median cumulative MFI was 2062 (IQR: 1038-6500). The median cumulative MIFI in females (5320.5; IQR: 633-23017) was higher than the median MIFI in males (1725; IQR: 804-5926) however the difference did not reach statistical significance (p = 0.475). Sixty-three percent of the DSA were against class II HLA antigens, while 37% were against class I HLA antigens. In regards of recipient-donor relationship, eight of the cases were sibling-to-sibling, five were child-to-parent and four were parent-tochild. Table 2 is a detailed description of the cases with DSA.

Outcomes
The OS, CIR, aGvHD, and cGvHD did not differ between patients with and without anti-HLA antibodies, nor between patients with and without DSA (Figs. 1a-f, and 2a-f). The gender of the recipient and donor, as well as the gender mismatch between recipient and donor, were statistically associated with the incidence of anti-HLA antibodies, but not with DSA. There was a tendency for an association between the recipient/donor relationship and the incidence of anti-HLA and DSA, but it did not achieve statistical significance (Table 1). Of the 17 patients with DSA, five patients underwent desensitization. Post desensitization, testing was carried out on four of them. One patient converted to negative DSA, the second patient had reduction in the intensity of antibodies however stayed positive by crossmatch, the third patient had reduction in the intensity of the antibodies but C1q, crossmatch were not done in his case, and the fourth patient was crossmatch positive before desensitization and turned to negative crossmatch after desensitization. These results are summarized in   Table 4 is a summary of the engraftment and GF data.

DISCUSSION
The presence of DSA against the unshared haplotype in the recipient is a known mechanism of GF after Haplo-HSCT [5]. However, the risk of GF is usually multifactorial (baseline disease, cell dose and source, presence of DSA, intensity of conditioning, transplant platform used, use of lymphodepletion, direction of mismatch and measures taken to decrease the risk among others) where the presence or absence of these factors will have a positive or negative impact on the engraftment. Meticulous risk stratification of these patients, in the context of all these factors, is essential to avoid undesirable outcomes. We analyzed 107 consecutive patients transplanted with our myeloablative haploidentical transplant platform for malignant diseases. The incidence of anti-HLA antibodies and DSA was similar to what was reported in the literature [5,[7][8][9][10][18][19][20][21][22]. While in a previous study the majority of DSA were directed against class I antigens (71%), in our study the majority were directed against class II antigens (63%) [20]. We evaluated the association between different baseline characteristics and the incidence of anti-HLA and DSA. The gender of the recipient and donor, as well as the gender mismatch between recipient and donor, were statistically associated with the incidence of anti-HLA antibodies, but not with DSA. There was a tendency for an association between the recipient/donor relationship and the incidence of anti-HLA and DSA, but it did not achieve statistical significance. These associations are consistent with the reported literature where female recipients were associated with higher incidence of anti-HLA and DSA [7,20]. The median intensity of the DSA was higher in female recipients (but not statistically significant) and that is in line of the published literature [7,8]. Sensitization during pregnancies by offspring's HLA antigens is well known to cause alloimmunization, and the other well-known factor is transfusion [5,8,[23][24][25][26]. Unfortunately, these variables were not recorded in our registry so we could not analyze their effects. As mentioned previously, DSA are a known risk factor for primary GF [5,7,9,10,12,13,20,27]. This risk depends on many other risk factors and has ranged between 1% with myeloablative conditioning and 8% with non-myeloablative regimens [7,22,[28][29][30][31][32][33][34]. In our study, one patient developed primary GF (0.9%), two developed secondary GF (1.9%), and interestingly none of the 17 recipients with DSA developed GF. The median time to ANC and platelets engraftment were similar in the anti-HLA group, DSA group and patients with no antibodies (time for platelet engraftment was longer in the DSA group but did not reach statistical difference). We assume this low incidence of primary GF is attributed to the use of a myeloablative regimen, the use of ATG in our preparative regimen and the fact that all  [35,36], but we could not study this in our cohort due to the low number of events. The OS, CIR and GvHD were not statistically different between the anti-HLA group, DSA group and patients with no antibodies in our study. In a previously reported study, there was a trend for better OS and CIR in favor of the non-DSA group. The authors postulated that this was probably related to the fact that GF was higher in the DSA group [20]. Desensitization was carried out in five patients (range of cumulative DSA MFI: 3983-23017), one of them was transplanted and engrafted only after mild reduction in the intensity of DSA despite persistent positive crossmatch; methods and results detailed in Table 3. Twelve patients were not  desensitized before transplant and all engrafted (range of cumulative DSA MFI: 633-5926). The European Society for Blood and Marrow Transplantation published guidelines for managing patients with DSA in the setting of haplo-HSCT and recommended desensitization for patients with DSA levels >5000 MFI [21].

CONCLUSION
Primary GF is a dreadful complication after allogeneic transplant where every attempt should be made to prevent it. The presence of DSA in the setting of mismatched transplant in general, and specifically in haplo-HSCT has been associated with higher incidence of PGF. In our cohort, the incidence of PGF was low and there was no cases of GF reported in patients with DSA. Our study suffers from the usual limitations of retrospective analysis; however, it highlights the multidimensional decision-making importance when thinking about GF. Unfortunately, this issue is difficult to address in prospective trials, and larger datasets are need to examine the impact of each risk factor (baseline disease, cell dose and source, presence of DSA, intensity of conditioning, transplant platform used, use of lymphodepletion, direction of mismatch and measures taken to decrease the risk of GF, etc.). We recommend analyzing large registries aiming to come up with a stratification tool based on the risk factors so that the treating teams can risk-stratify individual patients and recommend the appropriate risk-mitigation strategy.

DATA AVAILABILITY
Raw data were generated at King Faisal Specialist Hospital and Research Center, Riyadh, KSA. Derived data supporting the findings of this study are available from the corresponding author [RE] upon request.