Abstract
Background:
Haploidentical (Haplo) allogeneic HCTs (alloHCT) have been used more frequently over the last decade as survival is similar to HLA-matched related donor (MRD) alloHCTs.
Objective:
We aimed to identify donor and recipient immune signatures before alloHCT that are associated with clinically meaningful outcomes in MRD vs Haplo alloHCT recipients.
Study Design:
This retrospective cohort study of 165 MRD (n=132) and Haplo (n=33) alloHCT recipients and their related donors between 2007–2019 with paired peripheral blood samples immunophenotyped for T-cell, B-cell, NK cell and dendritic cell (DC) subsets. Immune cells were quantified before alloHCT in donors and recipients; calculations of immune cell ratios were classified as high, intermediate, and low and analyzed with alloHCT outcomes.
Results:
Haplo donors were younger than MRD donors (median: 35 vs 51 years), whereas Haplo recipients were older than MRD recipients (median: 68 vs 54 years), were more likely to have a Karnofsky Performance Score ≤70 (76% vs 57%), 3+ comorbidities (54% vs 47%), and were in complete remission prior to alloHCT (58% vs 42%).
In MRD alloHCT, a lower ratio of CD4+ to CD8+ effector memory cells in the donor was associated with lower 4-yr overall survival (OS; 25% v 61%; p=0.009), lower 4-yr progression free survival (PFS; 25% v 58%; p=0.014) and higher incidence of 1-yr transplant-related mortality (TRM; 39% v 7%; p=0.009) in recipients. A higher ratio of CD8+ effector memory to total NK cells measured in MRD recipients was associated with a higher incidence of grade II-IV aGvHD (63% v 37%; p=0.004) but was not statistically significant for III-IV aGvHD (23% v 12%).
In Haplo alloHCT, a lower ratio of total T-regulatory to CD4+ central memory cells in the donor was associated with lower 4-yr PFS (22% v 60%; p=0.0091). A higher ratio of CD4+ effector memory to CD8+ effector memory cells measured in Haplo recipients pre-alloHCT was associated with lower 4-yr OS (25% v 88%; p=0.0039). In both MRD and Haplo recipients, a higher ratio of CD4+ naïve to CD4+ central memory cells was associated with a higher incidence of grade II-IV aGvHD (64% v 38%; p=0.04).
Conclusion:
Evaluation of pre-alloHCT immune signatures of the donor and recipient may influence clinically meaningful patient outcomes in both MRD and Haplo transplants.
Keywords: immune profiles, allogeneic transplant, haploidentical transplant, matched related donor, flow cytometry
Background
Over the past two decades, haploidentical (Haplo) allogeneic hematopoietic cell transplantation (alloHCT) has become a well-established alternative to human leukocyte antigen (HLA)-matched related donor (MRD) alloHCT. Haplo alloHCT expands the available donor pool to patients who do not have MRDs or HLA-matched unrelated donors. In addition, the Haplo strategy allows patients who have higher risk/worse disease and more comorbidities to undergo alloHCT.1–7
Our prior work demonstrated that higher CD19+ B-cell and γδ T-cell counts in multiple myeloma patients treated with autologous HCT correlated with improved progression-free survival (PFS) and overall survival (OS).8 In NHL patients treated with autologous HCT, a higher proportion of CD8+ effector memory T-cells was associated with worse PFS and OS.9 Prior studies investigating immune repertoires and alloHCT outcomes have focused on evaluating the donor allograft composition in the HLA-matched setting.10,11 Studies on donor immune composition in the setting of Haplo HCT have been more limited and focused on specific cell populations.4,12
We hypothesized that immune signatures in both recipients and donors measured before alloHCT are associated with clinically meaningful outcomes, and that they differ by Haplo and MRD alloHCT. This is the first comprehensive immune profile analysis of both donors and recipients, which is unique in 1) describing qualitative differences in the immune phenotypes of Haplo and MRD donors and recipients and 2) examining donor and recipient baseline immune features in association with outcomes after Haplo and MRD alloHCT. The clinical implications of such associations could potentially change algorithms for donor selection and/or guide therapeutic strategies to manipulate the immune milieu in the recipient.
Methods
Patient population
This retrospective cohort was comprised of 165 MRD (n=132) and Haplo (n=33) alloHCT recipient-donor pairs treated at Roswell Park Comprehensive Cancer Center, from August 2007 to June 2019 (Supplemental Figure 1). A patient was considered an MRD if they had a sibling donor matched at HLA-A, -B and -DRB1. Haplo was defined as a related donor who was not a sibling and HLA-mismatched with the recipient. This study was approved by the Roswell Park Institutional Review Board and research was performed in accordance with the Declaration of Helsinki.
Flow cytometric analysis
Immune profiles were tested at a median of 21 days (range 6–62 days) before alloHCT in recipients and their respective donors using a comprehensive immunophenotyping panel via methods previously described.8,9 CD4+ cell count was collected as a measure of immune reconstitution around Day +100 post-alloHCT.13 The flow cytometrist was blinded to treatments and outcomes. Briefly, blood drawn into heparinized tubes was processed within 24 hours by washing three times with a flow cytometry buffer (0.5% BSA, 0.04 g/l Na2EDTA, and 0.1% sodium azide in PBS pH 7.2). Blood was then immunophenotyped using a stain-then-lyse technique and analyzed using a FACSCanto II flow cytometer (Becton Dickinson Biosciences, San Jose, CA, USA) with 50,000 events for each determination.
Lymphocytes were broadly defined using combined regions based on CD45 versus side scatter, and forward scatter versus side scatter. T-cells were separated from this population into CD3+ CD4+ and CD3+ CD8+ subsets. Naïve CD4+ and CD8+ T-cells were defined as CD3+CD4+ or CD3+CD8+ cells expressing CD45RA+ and either CD27+ (panel #1) or CD28+ (panel #2); Central Memory CD4+ or CD8+ T-cells were defined as CD45RA- and either CD27+ CD45RO- CD27+ (panel #1) or CD28+ (panel #2); Effector Memory CD4+ or CD8+ T-cells were defined as CD45RA- and either CD27- (panel #1) or CD28- (panel #2). Total T-regulatory (Treg) cells were defined as CD3+ CD4+ CD25+ CD127(dim) and as activated Treg cells if they were HLADR+. CD19+ B-cells were separated into naïve B-cells expressing CD27- (panel #1) or IgM+ CD27- (panel #2) and into memory B-cells expressing CD27+. Myeloid and plasmacytoid dendritic cells were initially defined as CD45+ cells falling within a forward versus side scatter mononuclear cell region containing both lymphocytes and monocytes. Myeloid dendritic cells were defined as CD11c+ HLADr+ CD123+/− and CD3- CD19- CD56- and plasmacytoid dendritic cells were defined as CD11c- HLADr+ CD123+ and CD3- CD19- CD56-. The gating strategy can be viewed in Supplemental Figure 2.
Data were analyzed using WinList version 6 (Verity Software House, Topsham, ME). These methods were conducted for panel #1 from August 2007 to May 2018, and expanded to panel #2 from May 2018 through August 2019. The concordance of the 2 panels by immune cell subset is depicted in Supplemental Table 1.
Statistical analyses
Clinical characteristics were analyzed using Pearson Χ2 test or Fisher’s exact test. The values for each immune cell were predefined at the <25, 25–75, and >75 percentiles of the entire cohort for panels 1 and 2, independently. Recipients (and donors) were then combined based on these predefined categories. Specifically, recipients who had an immune cell subset below the 25th percentile for panel #1 were grouped with recipients who had an immune cell subset below the 25th percentile for panel #2 as shown in Supplemental Table 1 in order to normalize the data across 2 panels and standardize the analysis. This was repeated for each predefined category (<25%, 25–75%, >75%) and each cell marker. Some of the predefined categories were collapsed based on data frequencies. Ratios were calculated by dividing immune cell subset counts within each patient, such as CD4+ naïve by CD4+ total, and classified as high, intermediate, and low for analysis of alloHCT outcomes.
OS was calculated from the date of allogeneic cell infusion (Day 0) to the date of death due to any cause and censored if alive at last follow-up. PFS was calculated from Day 0 to the date of first disease progression/relapse, death due to any cause, and censored if alive and progression-free at last follow up. Four-year estimates of OS and PFS are provided. Kaplan-Meier survival curves were calculated for OS and PFS, and significance was assessed using the log-rank test. GvHD-free, relapse-free survival (GRFS), defined as the absence of grade III–IV acute GvHD, chronic GvHD requiring systemic treatment, relapse, and death was evaluated using X2 tests.
Transplant-related mortality (TRM) was defined as death due to any cause other than disease progression. Acute graft-versus-host-disease (aGvHD) maximum grade was categorized as none and grade I versus grade II to IV for “grade II to IV aGvHD”. Grade III to IV aGvHD was defined as no aGvHD and grade I to II versus grade III to IV. Chronic GvHD (cGvHD) at 1-year was defined as ‘yes’ for any patient who developed moderate or severe cGvHD by 1-year post-HCT. Cumulative incidence curves were calculated for TRM and aGvHD. Analyses were stratified by age, gender, disease, Karnofsky performance score (KPS), sex match, disease status at HCT, HCT comorbidity index,14 disease risk index (DRI)15 and a DRI by age interaction term. All statistical tests were conducted using a two-sided p-value.
To balance small sample size with multiple comparisons, we reported all findings with a p-value below p=0.10 and defined statistical significance as p<0.01. Analyses were performed using SAS 9.4 (Cary, NC).
Results
Baseline characteristics of recipients and donors are presented in Table 1 (characteristics by panel are presented in Supplemental Table 2). Haplo donors were younger than MRD donors (median age 35 vs 51 years; p<0.0001), whereas Haplo recipients were older than MRD recipients (median age 68 vs 54 years; p<0.0001). Substantial differences did not exist by gender or race for donors or recipients nor were there any substantial differences by recipient diagnosis. Treatment differences between Haplo and MRD existed for stem cell source (p<0.0001), HCT regimen (p<0.0001), and GvHD prophylaxis (p<0.0001), as expected. Haplo and MRD recipient outcomes are compared in Table 2 (Outcomes by panel are presented in Supplemental Table 3). Haplo recipients were more likely to be in complete remission by Day +100 than MRD recipients (64% vs 48%, respectively) and less likely to develop grade III-IV aGvHD (3% vs 15%) and moderate/severe cGvHD (6% vs 30%). OS at 1-year was similar (60% vs 64%, respectively). Individual cell analyses are presented in the supplemental files. Immune cell ratio results are presented by MRD vs Haplo and Recipient vs Donor Immune cell subsets (Figure 1, Supplemental Table 4).
Table 1.
Clinical characteristics of recipients and donors
| Haplo (N=33) n (%) |
MRD (N=132) n (%) |
p-values | |
|---|---|---|---|
| Donor age at recipient HCT, years | <0.0001 | ||
| Median (range) | 35 (14 – 66) | 51 (17 – 73) | |
| Donor gender | 0.6 | ||
| Female | 14 (42) | 66 (50) | |
| Male | 19 (58) | 66 (50) | |
| Donor race/ethnicity | 0.3 | ||
| Non-Hispanic White | 30 (91) | 126 (95) | |
| Non-Hispanic Black | 2 (6) | 4 (3) | |
| Non-Hispanic Native American | 0 (0) | 1 (1) | |
| Hispanic | 0 (0) | 1 (1) | |
| Non-Hispanic Other | 1 (3) | 0 (0) | |
| Recipient age at HCT, years | <0.0001 | ||
| Median (range) | 68 (39 – 77) | 54 (21 – 72) | |
| 19–29 | 0 (0) | 8 (6) | |
| 30–39 | 1 (3) | 12 (9) | |
| 40–49 | 3 (9) | 26 (20) | |
| 50–59 | 4 (12) | 51 (39) | |
| 60–69 | 14 (42) | 33 (25) | |
| 70–77 | 11 (33) | 2 (2) | |
| Recipient age at HCT, years | <0.0001 | ||
| ≤60 | 8 (24) | 103 (78) | |
| >60 | 25 (76) | 29 (22) | |
| Recipient gender | 0.6 | ||
| Female | 15 (45) | 50 (38) | |
| Male | 18 (55) | 82 (62) | |
| Recipient race/ethnicity | 0.4 | ||
| Non-Hispanic White | 30 (91) | 125 (95) | |
| Non-Hispanic Black | 3 (9) | 4 (3) | |
| Non-Hispanic Native American | 0 (0) | 1 (1) | |
| Hispanic | 0 (0) | 2 (2) | |
| Diagnosis | 0.4 | ||
| ALL | 4 (12) | 19 (14) | |
| AML | 16 (48) | 58 (44) | |
| MDS | 8 (24) | 26 (20) | |
| MPN | 2 (6) | 8 (6) | |
| NHL | 3 (9) | 21 (16) | |
| KPS at HCT | 0.1 | ||
| ≤70 | 25 (76) | 75 (57) | |
| 80 | 6 (18) | 37 (28) | |
| 90 | 2 (6) | 20 (15) | |
| Prior HCT | |||
| AutoHCT | 3 (9) | 13 (10) | 0.99 |
| No HCT | 30 (91) | 119 (90) | |
| Stem cell source | <0.0001 | ||
| BM | 21 (64) | 9 (7) | |
| PB | 12 (36) | 123 (93) | |
| Conditioning intensity | 0.0008 | ||
| Myeloablative | 0 (0) | 40 (30) | |
| Reduced intensity | 33 (100) | 92 (70) | |
| HCT Regimen | <0.0001 | ||
| BuCy | 0 (0) | 22 (17) | |
| CyTBI | 0 (0) | 17 (13) | |
| FluCy | 0 (0) | 1 (1) | |
| FluCyTBI | 33 (100) | 0 (0) | |
| FluMel | 0 (0) | 49 (37) | |
| FluMelTBI | 0 (0) | 42 (32) | |
| EtoposideTBI | 0 (0) | 1 (1) | |
| Sex Match | 0.9 | ||
| Matched | 16 (48) | 60 (45) | |
| Mismatched | 17 (52) | 72 (55) | |
| Donor/Recipient Sex Match | 0.8 | ||
| Male/Female | 9 (27) | 28 (21) | |
| Female/Male | 8 (24) | 44 (33) | |
| Male/Male | 10 (30) | 38 (29) | |
| Female/Female | 6 (18) | 22 (17) | |
| HLA match | <0.0001 | ||
| 6/6 | 0 (0) | 115 (87) | |
| 10/10 | 0 (0) | 17 (13) | |
| 5/10 | 23 (70) | 0 (0) | |
| 6/10 or 7/10 | 8 (24) | 0 (0) | |
| 3/6 or 4/6 | 2 (6) | 0 (0) | |
| GvHD prophylaxis | <0.0001 | ||
| TacroMtxMMF | 0 (0) | 114 (87) | |
| TacroMMFCy | 29 (88) | 1 (1) | |
| TacroMMF or TacroMtx | 0 (0) | 14 (11) | |
| SiroMMFCy | 4 (12) | 1 (1) | |
| Siro or Tacro | 0 (0) | 2 (2) | |
| Status at HCT | 0.1 | ||
| CR1 | 19 (58) | 55 (42) | |
| CR>1 | 4 (12) | 16 (12) | |
| HI | 0 (0) | 2 (2) | |
| Relapse 1 | 0 (0) | 11 (8) | |
| Relapse >1 | 0 (0) | 2 (2) | |
| PIF | 9 (27) | 23 (18) | |
| Untreated | 1 (3) | 23 (17) | |
| CMV status recipient/donor | 0.2 | ||
| neg/neg | 13 (39) | 53 (40) | |
| neg/pos | 1 (3) | 21 (16) | |
| pos/neg | 13 (39) | 40 (30) | |
| pos/pos | 6 (18) | 18 (14) | |
| Cytogenetics | 0.4 | ||
| Adverse | 6 (18) | 16 (12) | |
| Intermediate | 19 (58) | 63 (48) | |
| Favorable | 0 (0) | 1 (1) | |
| N/A | 8 (24) | 52 (39) | |
| ABO match | 0.2 | ||
| Matched | 21 (64) | 77 (58) | |
| Major mismatch | 3 (9) | 30 (23) | |
| Minor mismatch | 9 (27) | 25 (19) | |
| Disease risk index | 0.8 | ||
| 0 | 2 (6) | 13 (10) | |
| 1 | 22 (67) | 85 (64) | |
| 2 | 8 (24) | 27 (20) | |
| 3 | 1 (3) | 7 (5) | |
| HCT-CI | 0.4 | ||
| 0 | 4 (12) | 32 (24) | |
| 1–2 | 11 (33) | 38 (29) | |
| 3–4 | 14 (42) | 42 (32) | |
| ≥5 | 4 (12) | 20 (15) |
Abbreviations: ALL - acute lymphoblastic leukemia; AML – acute myeloid leukemia; AutoHCT – autologous HCT; BM – bone marrow; Bu – busulfan; CMV – Cytomegalovirus; CR- complete remission; Cy – cyclophosphamide; Tacro – tacrolimus; Flu – fludarabine; Haplo – haploidentical donor; HCT - hematopoietic cell transplant; HCT-CI – HCT comorbidity index; HI – hematologic improvement; MDS – myelodysplastic syndrome; Mel – melphalan; MMF - mycophenolate mofetil; MRD – matched related donor; MPN – myeloproliferative neoplasm; Mtx – methotrexate; N/A – not applicable; NHL – non-Hodgkin lymphoma; PB – peripheral blood; PIF – primary induction failure; Siro – sirolimus; TBI – total body irradiation.
Table 2.
Recipient outcomes By Donor Match
| Haplo (n=33)% | MRD (n=132)% | p-value | |
|---|---|---|---|
| Best response at Day +100 | 0.4 | ||
| Continuous complete remission | 21 (64) | 64 (48) | |
| Complete remission | 4 (12) | 38 (29) | |
| Stable disease | 2 (6) | 6 (5) | |
| Progressed | 4 (12) | 15 (11) | |
| Not evaluable/early death | 2 (6) | 9 (7) | |
| Chimerism in those who did not relapse/die | |||
| Day 30 Chimerism in those who did not relapse/die by Day +30 | 0.03 | ||
| Complete Lymphoid/Myeloid | 26 (90) | 97 (77) | |
| Complete Myeloid only | 1 (3) | 24 (19) | |
| Incomplete | 1 (3) | 0 (0) | |
| Missing | 1 (3) | 5 (4) | |
| Day 100 Chimerism in those who did not relapse/die by Day +100 | 0.1 | ||
| Complete Lymphoid/Myeloid | 27 (93) | 89 (85) | |
| Complete Lymphoid only | 0 (0) | 3 (3) | |
| Complete Myeloid only | 0 (0) | 11 (10) | |
| Incomplete | 1 (3) | 0 (0) | |
| Missing | 1 (3) | 2 (2) | |
| 1 Year Chimerism in those who did not relapse/die by 1 year | 0.1 | ||
| Complete Lymphoid / Myeloid | 15 (94) | 67 (88) | |
| Complete Lymphoid only | 0 (0) | 1 (1) | |
| Complete Myeloid only | 0 (0) | 1 (1) | |
| Missing | 1 (6) | 7 (9) | |
| Maximum grade of acute GVHD | 0.01 | ||
| None | 13 (39) | 63 (48) | |
| I | 9 (27) | 9 (7) | |
| II | 10 (30) | 38 (29) | |
| III | 1 (3) | 9 (7) | |
| IV | 0 (0) | 11 (8) | |
| NE | 0(0) | 2(1) | |
| Maximum grade/severity of chronic GVHD | |||
| None | 22 (66) | 31 (24) | <0.0001 |
| Limited | 3 (9) | 14 (11) | |
| Extensive | 6 (18) | 69 (52) | |
| Not evaluable/early death | 2 (6) | 18 (14) | |
| None | 22 (66) | 31 (24) | <0.0001 |
| Mild Moderate/severe | 7 (21) | 43 (33) | |
| None | 2 (6) | 40 (30) | |
| Not evaluable/early death | 2 (6) | 18 (14) | |
| Follow up time in months | 48.8 (24.6 – 90.9) | 93 (31.9–154.6) | <0.0001 |
| Status at 1 year | 0.96 | ||
| Alive | 20 (61) | 84 (64) | |
| Dead | 13 (39) | 48 (36) | |
| Cause of death | |||
| Died due to Disease | 9 (27) | 25 (19) | |
| Died due to Infection | 2 (6) | 9 (7) | |
| Died due to GVHD | 1 (3) | 9 (7) | |
| Died due to other causes | 1 (3) | 5 (4) |
Abbreviations: GvHD – graft-versus-host disease; Haplo – haploidentical donor; HCT - hematopoietic cell transplant; MRD – matched related donor.
Figure 1. Significant immune cell ratios by MRD vs Haplo and Patient vs Donor cells for HCT outcomes.
Larger dots represent a lower p-value with the largest representing p<0.01. Blue dots represent improved outcomes whereas red dots represent worse outcomes. Darker blue or reds dots represent an intermediate ratio between immune cell populations as opposed to a higher immune cell ratio associated with HCT outcomes. aGvHD – acute graft-versus-host disease, CM – central memory, DC – dendritic cells, EM – effector memory, NK – natural killer cells, NS – not significant, OS – overall survival, PFS – progression free survival, Treg – T regulatory cells, TRM – transplant-related mortality.
Donor immune cell ratios and OS, PFS, TRM, and aGvHD
For MRD recipients, a lower ratio of CD4+ effector memory to CD8+ effector memory cells in the donor was significantly associated with lower 4-year OS (25% vs 45% vs 61%; p=0.009; Figure 2A), lower 4-year PFS (25% vs 39% vs 58%; p=0.014; Figure 2B) and a higher incidence of 1-year TRM (36% vs 45% vs 7%; p=0.009; Figure 2C). In Haplo recipients, a lower ratio of Total Treg to CD4+ central memory cells in the donor was associated with lower 4-year PFS (22% vs 60%; p=0.0091; all other results mentioned in the text are presented in Supplemental Figure 3).
Figure 2.
In MRD recipients, the ratio of CD4+ effector memory to CD8+ effector memory cells in the donor associated with A) 4-year Overall Survival, B) 4-year Progression-Free Survival, and C) and 1-year incidence of Transplant-Related Mortality
The following cell types found in donors were only associated with a higher incidence of grade II-IV aGvHD in MRD: a higher ratio of CD8+naïve to CD8+ central memory cells (52% vs 34%; p=0.03) and a lower ratio of CD8+ central memory to NK cells (59% vs 38%; p=0.03). A higher ratio of CD4+ naïve to CD4+ central memory cells in donors was associated with a higher incidence of grade II-IV aGvHD in Haplo and MRD recipients (64% vs 38%; p=0.04), whereas no associations were found with grade III-IV aGvHD. In donor cells, an intermediate ratio of total Tregs to CD4+ central memory cells (P=0.0079) in the donor was associated with higher GRFS in the recipient.
Recipient immune cell ratios pre-HCT and OS, PFS, TRM, aGvHD, and GRFS
In Haplo recipients, a higher CD4+ effector memory to CD8+ effector memory ratio was associated with lower 4-year OS (88% vs 25%; p=0.0039). MRD and Haplo recipients with a higher ratio of CD4+ naïve to CD4+ central memory cells were associated with grade II-IV aGvHD (70% vs 31%; p<0.0001) but not grade III-IV aGvHD (18% vs 8%; p=0.08).
In Haplo recipients, a lower CD19+ naïve to CD19+ memory B-cell ratio was associated with lower 4-year OS (20% vs 40% vs 69%; 0.045; Figure 3A), 4-year PFS (20% vs 29% vs 69%; p=0.045; Figure 3B), and a higher incidence of grade II-IV aGvHD (80% vs 57% vs 10%; p=0.003; Figure 3C) but not grade III-IV (14% vs 0% vs 0%; p=0.18; Figure 3D). A higher CD8+ effector memory to NK cell ratio (63% vs 37%; p=0.004) in MRD recipients was associated with an increased incidence of grade II-IV aGvHD but not grade III-IV (22% vs 11%; p=0.1). In MRD recipients, a higher ratio of CD4+ naïve to CD8+ naïve cells measured in the recipient was associated with a higher incidence of cGvHD (54% vs 28%; p=0.0048).
Figure 3.
In Haplo recipients, the ratio of CD19+ naïve to CD19+ memory B-cells in the patient was associated with A) 4-year Overall Survival, B) 4-year Progression-Free Survival, C) grade II-IV acute GvHD, and D) not associated with grade III-IV acute GvHD.
Age was the only variable that differed by donor match (MRD vs Haplo) and OS. None of the single cell subsets or ratios investigated with OS by transplant type were statistically significant (please refer to supplemental methods for additional details). No immune cell ratios measured in the recipient were associated with GRFS.
Donor to recipient immune cell ratios pre-HCT and OS, PFS, TRM, and aGvHD
To investigate any imbalances between donor and recipient immune cells, ratios were calculated to compare the relative frequency of immune cell ratios between donors and their recipients. In Haplo alloHCT, patients had a lower 4-year OS (0% vs 49%; p=0.011; Figure 4A) and PFS (0% vs 41%; p=0.026; Figure 4B) if the CD4+ to CD8+ naïve cell ratio was high in the donor but low in the recipient, despite no significant associations in the donors or recipients. In MRD alloHCT, patients had a higher TRM (32% vs 11%; p=0.005; Figure 5A) and grade III-IV aGvHD (20% vs 9%; p=0.027; Figure 5B) if the CD4+ to CD8+ effector memory ratio was low in the donor and high in the recipient. In MRD and Haplo alloHCT, if the activated Treg to CD4+ effector memory ratio was high in the donor but low in the recipient, this yielded higher TRM (28% vs 11%; p=0.012) and grade III-IV aGvHD (22% vs 7%; p=0.017). If the activated Treg to CD4+ effector memory ratio was low in the donor but high in the recipient, it had no effect on grade III-IV aGvHD in Haplo alloHCT (n=0/23).
Figure 4.
In Haplo recipients, the donor to recipient CD4+ naïve to CD8+ naïve ratio was associated with A) 4-year Overall Survival and B) Progression-Free Survival.
Figure 5.
In MRD recipients, the donor to patient CD4+ effector memory to CD8+ effector memory ratio was associated with A) Transplant-Related Mortality and B) grade III-IV acute GvHD.
Analyses by Risk
To investigate whether the significant CD4+ to CD8+ effector memory ratio had differential effects by underlying known risk factors for poor OS and higher TRM, results were stratified by recipient and donor age, HCT-CI, KPS, disease status at transplant, and DRI (Supplemental Figure 4). For MRD alloHCT, a high CD4+ to CD8+ effector memory ratio measured in donors conveyed better OS in both low- and high-risk groups defined by HCT-CI (0–2 vs 3+), disease status at HCT (CR vs not in CR), KPS (<80 vs ≥80), and DRI (high/very high vs low/intermediate). However, this ratio measured in donors had no impact on OS for recipients ≤50 years or when the donor was ≤40 years. Similarly, a high CD4+ to CD8+ effector memory ratio measured in donors conveyed lower TRM in HCT-CI, KPS, and DRI low and high-risk groups, however it only had an impact on TRM in high-risk groups: older recipients, older donors, and patients not in CR at alloHCT.
In Haplo alloHCT, a higher CD4+ to CD8+ effector memory ratio measured in recipients demonstrated lower OS in both low- and high-risk groups defined by HCT-CI, recipient age, donor age, disease status at HCT, and KPS (Supplemental Figure 5). A higher ratio significantly associated with OS in patients with low or intermediate risk DRI and could not be evaluated in high/very high risk DRI with only one patient in that group.
Analyses by Immune Reconstitution
Post-alloHCT immune reconstitution measured as the day +100 CD4+ count >50 cells/µL was significantly associated with improved OS, PFS, GvHD and TRM (Supplemental Figure 6). None of the pre-alloHCT immune cell counts were associated with the Day +100 CD4+ cell count, either overall or stratified by Haplo and MRD, possibly due to small sample sizes. However, when the analyses were limited to those patients who did not progress at Day +100, more CD4+ naïve cells measured in the donor were significantly associated with a lower CD4+ count in Haplo patients at Day +100 (p=0.0075; Supplemental Table 5).
Discussion
In this retrospective cohort study, we determined that a variety of immune cell ratios in related allogeneic donors are associated with OS, PFS, TRM and GvHD. There were marked differences in the specific ratios which correlated with outcomes in recipients of MRD and Haplo alloHCT. These could be due to inherent differences due to degree of HLA match/mismatch, demographic differences due to donor age and availability, conditioning regimens, and GvHD prophylaxis with post-transplant cyclophosphamide (PTCy).
It is assumed that donor cell engraftment post-alloHCT would drive greater effects on outcomes by using baseline donor immune profiles instead of those in recipients. However, we observed associations between multiple recipient immune cell ratios and outcomes after alloHCT in patients with full donor chimerism. As with patterns of donor immune cell ratios in the MRD and Haplo settings, the patterns of associations between recipient immune cell ratios and outcomes differed by MRD and Haplo.
In MRD alloHCT which has faster and more robust T-cell immune reconstitution than in Haplo,16 a lower CD4+ to CD8+ effector memory T-cell ratio in the donor resulted in lower OS, PFS and higher TRM but not aGvHD. A lower CD4+ to CD8+ effector memory T-cell ratio in donors may represent a weaker donor immune response that fails to adequately protect the recipient from leukemia relapse and infection, especially when the recipient has a high CD4+ to CD8+ effector memory T-cell ratio.17–19 In the Haplo HLA-mismatched setting and PTCy, a higher CD4+ to CD8+ effector memory T-cell ratio in the recipient resulted in lower OS, PFS and higher aGvHD. HLA Class I and II mismatches trigger robust CD4+ and CD8+ naïve donor T-cell responses and cytokine production after primary exposure to recipient antigens are eliminated, or at least rendered dysfunctional, after treatment with PTCy.20 However, recipient T-cells may survive both conditioning and PTCy exposure, especially those residing in tissue, and may contribute to GvHD but less so GVL as they migrate from host organs and lymphoid structures early post-transplant.21 This phenomenon may support our observation that recipient CD4+ effector memory T-cell skewing in Haplo alloHCT correlated with lower PFS and higher rates of aGvHD. Additionally, in both MRD and Haplo alloHCT, donor CD4+ naïve skewing was associated with higher incidence of aGvHD. A possible explanation for this association could be that a pool of antigen naïve cells in the graft would turn alloreactive following infusion into the recipient leading to aGvHD.
Reestablishing immune reconstitution post-alloHCT is critical to generate positive long-term outcomes after alloHCT,13 and our results support this. We were unable to establish an association between pre-alloHCT immune cell subsets and Day 100 CD4+ counts, although this lack of association is likely due to small sample sizes rather than the lack of a true effect.
The results of our retrospective analysis should be validated in larger cohorts and investigated in MRD alloHCT with PTCy GvHD prophylaxis regimens to distinguish the effects of HLA-mismatch vs PTCy on our findings. Pre-clinical animal models are needed to elucidate the biological mechanisms that contribute to the association of each outcome and specific immune cell ratio. Additional studies are needed to assess the potential impact of therapeutic modification of immune cell ratios in HCT grafts ex vivo and/or in patients pre-transplant on outcomes.
This is the first longitudinal study on pre-HCT donor and recipient samples comparing the immune profiles of MRD and Haplo donors and recipients on HCT outcomes. These data suggest that in either the MRD or Haplo setting the evaluation of donor immune cell ratios could be impactful on clinically meaningful patient outcomes. These results could guide therapeutic strategies to manipulate the immune milieu in the recipient and improve alloHCT outcomes. Future directions include validating this in other alloHCT populations, and expanding this research to unrelated donors, which may ultimately help select preferred donors.
Supplementary Material
Highlights.
Comprehensive immunophenotyping panel of both recipients and related donors
Pre-allogeneic transplant immune profiles influence transplant outcomes
Related donor immune cells are associated with mortality outcomes
Recipient immune cells are associated with survival and graft-versus-host disease
Acknowledgements
This work was supported by National Cancer Institute (NCI) grant P30CA016056 involving the use of Roswell Park Comprehensive Cancer Center’s Flow & Image Cytometry Shared Resource.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Financial Disclosures: Dr. Philip McCarthy has received honoraria from, and participated in Advisory Boards for, Bristol-Myers Squibb, Bluebird, Celgene, Janssen, Juno Karyopharm, Magenta Therapeutics, Sanofi, The Binding Site, and Takeda. Dr. Shernan Holtan received research funding from Vitrac Therapeutics and Incyte, clinical trial adjudication for CSL Behring, and consults for Ossium, Sanofi, and MaaT Pharma. Dr. Brian Betts received research support from CTI BioPharma and Vitrac and has IP licensing for CD83 CAR T: CRISPR Tx 2021–2023.
References
- 1.Rashidi A, Hamadani M, Zhang MJ, et al. Outcomes of haploidentical vs matched sibling transplantation for acute myeloid leukemia in first complete remission. Blood Adv 2019;3(12):1826–1836. 10.1182/bloodadvances.2019000050. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Meybodi MA, Cao W, Luznik L, et al. HLA-haploidentical vs matched-sibling hematopoietic cell transplantation: a systematic review and meta-analysis. Blood Adv 2019;3(17):2581–2585. 10.1182/bloodadvances.2019000614. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Wieduwilt MJ, Metheny L, Zhang MJ, et al. Haploidentical vs sibling, unrelated, or cord blood hematopoietic cell transplantation for acute lymphoblastic leukemia. Blood Adv 2022;6(1):339–357. 10.1182/bloodadvances.2021004916. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Gagelmann N, Bacigalupo A, Rambaldi A, et al. Haploidentical Stem Cell Transplantation With Posttransplant Cyclophosphamide Therapy vs Other Donor Transplantations in Adults With Hematologic Cancers: A Systematic Review and Meta-analysis. JAMA Oncology. 2019;5(12):1739–1748. 10.1001/jamaoncol.2019.3541. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Brunstein CG, Fuchs EJ, Carter SL, et al. Alternative donor transplantation after reduced intensity conditioning: results of parallel phase 2 trials using partially HLA-mismatched related bone marrow or unrelated double umbilical cord blood grafts. Blood. 2011;118(2):282–288. 10.1182/blood-2011-03-344853. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ciurea SO, Zhang MJ, Bacigalupo AA, et al. Haploidentical transplant with posttransplant cyclophosphamide vs matched unrelated donor transplant for acute myeloid leukemia. Blood. 2015;126(8):1033–1040. 10.1182/blood-2015-04-639831. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Kanate AS, Mussetti A, Kharfan-Dabaja MA, et al. Reduced-intensity transplantation for lymphomas using haploidentical related donors vs HLA-matched unrelated donors. Blood. 2016;127(7):938–947. 10.1182/blood-2015-09-671834. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ho CM, McCarthy PL, Wallace PK, et al. Immune signatures associated with improved progression-free and overall survival for myeloma patients treated with AHSCT. Blood Adv 2017;1(15):1056–1066. 10.1182/bloodadvances.2017005447. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Herr MM, Torka P, Zhang Y, et al. Immune profiling in diffuse large B-cell lymphoma and mantle cell lymphoma patients treated with autologous hematopoietic cell transplant. Bone Marrow Transplant. 2020;55(1):77–85. 10.1038/s41409-019-0591-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Podgorny PJ, Pratt LM, Liu Y, et al. Low counts of B cells, natural killer cells, monocytes, dendritic cells, basophils, and eosinophils are associated with postengraftment infections after allogeneic hematopoietic cell transplantation. Biol Blood Marrow Transplant. 2016;22(1):37–46. 10.1016/j.bbmt.2015.09.003. [DOI] [PubMed] [Google Scholar]
- 11.Moiseev IS, Babenko EV, Epifanovskaya OS, et al. High prevalence of CD3, NK, and NKT cells in the graft predicts adverse outcome after matched-related and unrelated transplantations with post transplantation cyclophosphamide. Bone Marrow Transplant. 2020;55(3):544–552. 10.1038/s41409-019-0665-3. [DOI] [PubMed] [Google Scholar]
- 12.Cieri N, Oliveira G, Greco R, et al. Generation of human memory stem T cells after haploidentical T-replete hematopoietic stem cell transplantation. Blood. 2015;125(18):2865–2874. 10.1182/blood-2014-11-608539. [DOI] [PubMed] [Google Scholar]
- 13.Bertaina A, Abraham A, Bonfim C, et al. An ISCT Stem Cell Engineering Committee Position Statement on Immune Reconstitution: the importance of predictable and modifiable milestones of immune reconstitution to transplant outcomes. Cytotherapy. 2022;24(4):385–392. 10.1016/j.jcyt.2021.09.011. [DOI] [PubMed] [Google Scholar]
- 14.Sorror ML, Maris MB, Storb R, et al. Hematopoietic cell transplantation (HCT)-specific comorbidity index: a new tool for risk assessment before allogeneic HCT. Blood. 2005;106(8):2912–2919. 10.1182/blood-2005-05-2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Armand P, Kim HT, Logan BR, et al. Validation and refinement of the Disease Risk Index for allogeneic stem cell transplantation. Blood. 2014;123(23):3664–3671. 10.1182/blood-2014-01-552984. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Rieger MJ, Stolz SM, Müller AM, et al. Haploidentical transplant with posttransplant cyclophosphamide vs matched related and unrelated donor transplant in acute myeloid leukemia and myelodysplastic neoplasm. Bone Marrow Transplant. 2023;58(10):1121–1129. 10.1038/s41409-023-02042-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Laidlaw BJ, Craft JE, Kaech SM. The multifaceted role of CD4+ T cells in CD8+ T cell memory. Nature Reviews Immunology 2016;16(2):102–111. 10.1038/nri.2015.10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Caserta S, Borger JG, Zamoyska R. Central and effector memory CD4 and CD8 T-cell responses to tumor-associated antigens. Crit Rev Immunol 2012;32(2):97–126. 10.1615/critrevimmunol.v32.i2.10. [DOI] [PubMed] [Google Scholar]
- 19.Gasper DJ, Tejera MM, Suresh M. CD4 T-cell memory generation and maintenance. Crit Rev Immunol 2014;34(2):121–146. 10.1615/critrevimmunol.2014010373. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Fletcher RE, Nunes NS, Patterson MT, et al. Posttransplantation cyclophosphamide expands functional myeloid-derived suppressor cells and indirectly influences Tregs. Blood Adv 2023;7(7):1117–1129. 10.1182/bloodadvances.2022007026. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Sacirbegovic F, Günther M, Greco A, et al. Graft-versus-host disease is locally maintained in target tissues by resident progenitor-like T cells. Immunity 2023;56(2):369–385.e366. 10.1016/j.immuni.2023.01.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.