Abstract
Background
Myeloid-derived suppressor cells (MDSCs) are potent immunoregulatory cells. Their role in modulating acute graft-versus-host disease (aGVHD) following allogeneic hematopoietic stem cell transplantation (allo-HSCT) remains unclear. This study aimed to investigate the impact of MDSC levels in granulocyte colony-stimulating factor (G-CSF)–mobilized peripheral blood stem cell (PBSC) grafts on the incidence of II–IV° aGVHD.
Results
We retrospectively analyzed 170 allo-HSCT recipients. Employing an exposure-based analytical framework, patients were stratified into high- and low-dose groups based on the infused dose of polymorphonuclear MDSCs (PMN-MDSCs). In Fine-Gray competing risk analyses, a high dose of PMN-MDSCs per kilogram (> 17.5 × 10⁶/kg) was an independent protective factor against aGVHD (subdistribution hazard ratio [sHR] 0.25, 95% CI 0.07–0.88, P = 0.039). A clinically applicable optimal cut-off was identified at 11.3 × 10⁶/kg. Patients receiving a PMN-MDSC dose > 11.3 × 10⁶/kg had significantly superior 300-day overall survival (92.4% vs. 74.9%, P = 0.005) and GVHD-free relapse-free survival (76.3% vs. 40.4%, P < 0.001) compared to the low-dose group. In the malignant disease subset (n = 147), a high dose of the activated LOX-1⁺ PMN-MDSC subset was associated with a markedly lower cumulative incidence of relapse (1.37% vs. 11.42%, P = 0.015). Longitudinal monitoring revealed that the peak incidence of aGVHD (median day + 29) closely followed the peak and subsequent decline of circulating MDSC levels.
Conclusions
The absolute dose of PMN-MDSCs in G-CSF-mobilized grafts is an independent determinant of II–IV°aGVHD risk and survival outcomes, with a defined threshold of 11.3 × 10⁶/kg. These findings position PMN-MDSC dose as a promising biomarker for risk stratification and a potential lever for optimizing graft composition in allo-HSCT.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12865-025-00794-5.
Keywords: Polymorphonuclear MDSCs, Acute graft-versus-host disease, Peripheral blood stem cell, Granulocyte colony-stimulating factor, Allogeneic hematopoietic stem cell transplantation
Introduction
Allogeneic hematopoietic stem cell transplantation (allo-HSCT) has become an essential therapeutic approach for malignant hematologic disorders and severe immunodeficiencies, with substantial progress achieved in recent years [1]. However, acute graft-versus-host disease (aGVHD) remains one of the most common complications following allo-HSCT and continues to limit transplant success and long-term quality of life [2]. Identifying factors associated with aGVHD is therefore critical to guide prophylaxis and improve patient outcomes.
Myeloid-derived suppressor cells (MDSCs) are known to play pivotal roles in tumor biology, chronic inflammation, and immune modulation after allo-HSCT [3]. The immunosuppressive effects of MDSCs can influence post-transplant immunological tolerance and can inhibit the function of effector T cells to reduce severity of aGVHD [4, 5]. Nevertheless, the precise functions of donor-derived peripheral blood stem cell (PBSC)–infused MDSCs remain incompletely understood, and therapeutic strategies targeting MDSCs are still under exploration. This study aimed to investigate the association between infused PBSC-derived MDSCs and the incidence of aGVHD in allo-HSCT recipients, and more importantly, to provide novel insights into potential biomarkers and therapeutic strategies for improving transplant outcomes.
Methods
Patients and donors
A total of 170 patients who underwent allo-HSCT at Beijing Lu Daopei Hospital between July 2022 and October 2023 were retrospectively analyzed, including 140 related donor transplants and 30 unrelated donor transplants. Diagnoses of acute leukemia and aplastic anemia were defined according to the World Health Organization (WHO) classification of hematolymphoid tumors [6], while myelodysplastic syndrome (MDS) was diagnosed based on the Chinese Guidelines for the Diagnosis and Treatment of MDS (2019 edition) [7]. This study was approved by the Ethics Committee of Beijing Lu Daopei Hospital (ethics No. DPEC-M-202309) and conforms to the principles outlined in the Declaration of Helsinki (Clinical trial number: not applicable).
Sample collection and mobilization
Donors received recombinant human granulocyte colony-stimulating factor (rhG-CSF; 5–10 µg/kg/day, administered subcutaneously twice daily) for mobilization. PBSCs collection was processed through a blood separator, which is able to separate and extract different components of the blood based on their specific characteristics. The entire collection process generally lasts 2 to 4 h, depending on the efficiency of the mobilization and the equipment to achieve the target stem cell count for the recipient. Donor grafts were collected into sodium citrate–anticoagulated bags, and recipient peripheral blood samples were obtained at day + 7, +14, + 21, +28, + 42, +60, and + 90 after transplantation.
Flow cytometric analysis of MDSCs
Mononuclear cells were isolated by density-gradient centrifugation. MDSCs and their subsets were identified by multicolor flow cytometry using the following antibody panel: CD45, CD14, CD11b, CD16, CD33, CD15, HLA-DR, CD123, and LOX-1. Data were acquired on a BD flow cytometer (FACSDiva v8.0.2) and analyzed using Kaluza v2.1 software. For each sample, at least 1 × 106 events were recorded. Absolute counts were determined using BD Trucount tubes and CD45/CD14 staining according to the manufacturer’s instructions [8, 9]. Mononuclear cells were isolated by density-gradient centrifugation. MDSCs and their subsets were identified as follows: Total MDSCs were defined as CD45⁺CD11b⁺CD33⁺HLA-DR⁻/low cells. These were further subdivided into: Monocytic MDSCs (M-MDSCs, CD14⁺CD15⁻) and Polymorphonuclear MDSCs (PMN-MDSCs, CD14⁻CD15⁺). The activated subset of PMN-MDSCs was identified by the expression of LOX-1 (LOX-1⁺PMN-MDSCs). Furthermore, to maximize clarity and reproducibility, we have included a representative gating strategy for the flow cytometric analysis as Supplementary Figure S1.
Statistical analysis
An initial exploratory analysis was performed to identify potential associations between graft composition and clinical outcomes. In this analysis, patients were grouped based on the post-transplant development of II–IV° aGVHD (GVHD group, n = 37) or the absence of such an event (no-GVHD group, n = 133). The proportions and absolute counts of MDSC subsets in the donor grafts were compared between these outcome-defined groups using the Mann-Whitney U test. The results of these exploratory comparisons are provided in Supplementary Figures S2, S3, and Supplementary Table S1. It is important to note that these analyses were descriptive and intended for hypothesis generation; the primary analysis for this study was the exposure-based analysis of MDSC dose on time-to-event outcomes, as detailed below.
The primary, exposure-based analyses were conducted as follows: baseline data were summarized using medians (IQR) for continuous variables and frequencies (percentages) for categorical variables. The cumulative incidence of II-IV0 aGVHD, with death treated as a competing risk, was estimated using the Fine-Gray proportional subdistribution hazards model to calculate subdistribution hazard ratios (sHRs) and corresponding 95% confidence intervals (CIs). Variables with a P value < 0.10 in univariate analyses or those deemed clinically relevant were entered into multivariate models to identify independent prognostic factors. The optimal prognostic cut-off value for the infused PMN-MDSC dose was determined based on this model and internally validated using bootstrap resampling with 1,000 iterations.
OS, GRFS, LFS, and RI were compared using the log-rank test or Gray’s test, as appropriate. Longitudinal MDSC levels were evaluated by two-way ANOVA. All statistical analyses were performed using SPSS version 26.0 (IBM, Armonk, NY, USA) and R software (version 4.5.0), and graphical plots were generated with GraphPad Prism version 9.0 (GraphPad Software, San Diego, CA, USA). A two-tailed P < 0.05 was considered statistically significant.
Results
To initially investigate the potential role of MDSCs in aGVHD, we compared the composition of donor grafts between patients who did and did not develop grade II–IV aGVHD (exploratory analysis). This comparison revealed that grafts infused into patients who later developed aGVHD contained significantly lower levels of total MDSCs and their PMN-MDSC subset (Supplementary Figure S2, S3; Supplementary Table S1). Based on this observation, we hypothesized that the absolute dose of infused MDSCs, particularly PMN-MDSCs, might be a causal protective factor against aGVHD. We therefore proceeded to test this hypothesis using an exposure-based analytical framework, stratifying all patients (N = 170) by the dose of PMN-MDSCs they received (high vs. low).
Baseline characteristics
This study analyzed a cohort of 170 patients who underwent allogeneic hematopoietic stem cell transplantation. The cohort consisted predominantly of patients with acute myeloid leukemia (51.8%), most of whom were in complete remission (80.0%) at transplantation. The majority received transplants from related donors (82.4%), primarily haploidentical donors (82.9%), following busulfan-based conditioning regimens (80.0%). The median infused doses of CD34 + cells and mononuclear cells were 4.65 × 10⁶/kg and 5.48 × 10⁸/kg, respectively. Regarding the key investigational parameters, the grafts contained median doses of 33.75 × 10⁶/kg total MDSCs, with the polymorphonuclear subset (PMN-MDSCs) measuring 17.50 × 10⁶/kg (Table 1).
Table 1.
Baseline Characteristics and Transplant-Related Information of 170 Patients
| Characteristics | Total (n = 170) |
|---|---|
| Donor gender | |
| Male | 112 (65.9) |
| Female | 58 (34.1) |
| Patient gender | |
| Male | 86 (50.6) |
| Female | 84 (49.4) |
| HSCT type | |
| Related donors | 140 (82.4) |
| Unrelated donors | 30 (17.6) |
| Weight, M (Q₁, Q₃) | 58.60 (47.12, 69.00) |
| Type of disease | |
| ALL | 28 (16.5) |
| AML | 88 (51.8) |
| AA | 23 (13.5) |
| MDS | 20 (11.8) |
| Others | 11 (6.4) |
| Disease status pre-HSCT | |
| CR | 136 (80.0) |
| PR/NR | 34 (20.0) |
| Donor G-CSF Dose, µg/kg, M (Q₁, Q₃) | 8.98 (8.24, 9.86) |
| Donor type | |
| HID | 141 (82.9) |
| MSD | 29 (17.1) |
| Conditioning Regimen | |
| BU-based | 136 (80.0) |
| TBI-based | 34 (20.0) |
| MNC cells,×108/kg, M (Q₁, Q₃) | 5.48 (4.13, 7.43) |
| CD34+ cells,×106/kg, M (Q₁, Q₃) | 4.65 (3.66, 6.03) |
| WBC engraftment (days), M (Q₁, Q₃) | 13.00 (12.00, 16.00) |
| PLT engraftment (days), M (Q₁, Q₃) | 13.00 (11.00, 18.00) |
| GVHD prophylaxis regimen | |
| CSA-based | 143 (84.1) |
| FK506-based | 27 (15.9) |
| M-MDSCs(units/kg), M (Q₁, Q₃) | 7.05 (2.30, 18.45) |
| PMN-MDSCs(units/kg), M (Q₁, Q₃) | 17.50 (9.98, 46.00) |
| MDSCs(units/kg), M (Q₁, Q₃) | 33.75 (14.62, 68.62) |
| LOX1 + PMN-MDSCs(units/kg), M (Q₁, Q₃) | 2.55 (0.80, 7.17) |
| MDSCs(%), M (Q₁, Q₃) | 0.05 (0.03, 0.08) |
| M-MDSCs(%), M (Q₁, Q₃) | 0.01 (0.00, 0.03) |
| PMN-MDSCs(%), M (Q₁, Q₃) | 0.03 (0.02, 0.06) |
| LOX1 + PMN-MDSCs(%), M (Q₁, Q₃) | 0.16 (0.06, 0.23) |
| MDSCs(units), M (Q₁, Q₃) | 14108.10 (6613.02, 24350.72) |
| M-MDSCs(units), M (Q₁, Q₃) | 2615.83 (947.27, 8520.55) |
| PMN-MDSCs(units), M (Q₁, Q₃) | 8563.16 (3779.54, 16540.51) |
| LOX1 + PMN-MDSCs(units), M (Q₁, Q₃) | 961.78 (352.70, 2484.63) |
| Median follow up time, days, M (Q₁, Q₃) | 518 (260, 584) |
Abbreviation: GVHD graft-versus-host disease, HSCT hematopoietic stem cell transplantation, ALL acute lymphoblastic leukemia, AML acute myeloid leukemia, AA aplastic anemia, MDS myelodysplastic syndromes, CR complete remission, PR/NR partial or no response, G-CSF granulocyte colony-stimulating factor, HID haploidentical donor, MSD matched sibling donor, BU busulfan, TBI total body irradiation, MNC mononuclear cells, PBSC peripheral blood stem cells, CSA cyclosporine, M Median, Q₁ 1st Quartile, Q₃ 3st Quartile
Cumulative incidence of aGVHD stratified by MDSC dose
To assess the impact of graft MDSC composition on the risk of severe aGVHD, we stratified all patients (N = 170) into high- and low-dose groups based on the median infused dose of each MDSC subset and analyzed the cumulative incidence of grade II-IV aGVHD using competing risk analysis. As shown in Fig. 1, stratification by the median dose of each MDSC subset revealed consistent trends in the cumulative incidence of grade II-IV aGVHD. For absolute cell counts, high-dose groups demonstrated lower incidence rates compared to low-dose groups: 28.9% vs. 14.9% (P = 0.027) for total MDSCs (Fig. A2), 29.3% vs. 14.0% (P = 0.012) for monocytic MDSCs (M-MDSCs) (Fig. B2), and 27.5% vs. 16.7% (P = 0.074) for polymorphonuclear MDSCs (PMN-MDSCs) (Fig. C2). Analyses based on cellular proportions showed similar directional trends (Fig. A1, B1, C1). No significant associations were observed for the LOX-1+ PMN-MDSCs subset by either percentage or absolute count (Fig. D1, D2).
Fig. 1.
Association Between Infused MDSC Dose and the Cumulative Incidence of Grade II–IV aGVHD. A1 The Cumulative Incidence of Grade II-IV aGVHD between high and low dose of MDSCs (%); A2 The Cumulative Incidence of Grade II-IV aGVHD between high and low dose of MDSCs (units); B1 The Cumulative Incidence of Grade II-IV aGVHD between high and low dose of M-MDSCs (%); B2 The Cumulative Incidence of Grade II-IV aGVHD between high and low dose of M-MDSCs (units); C1 The Cumulative Incidence of Grade II-IV aGVHD between high and low dose of PMN-MDSCs (%); C2 The Cumulative Incidence of Grade II-IV aGVHD between high and low dose of PMN-MDSCs (units); D1 The Cumulative Incidence of Grade II-IV aGVHD between high and low dose of LOX-1+ PMN-MDSCs (%); D2 The Cumulative Incidence of Grade II-IV aGVHD between high and low dose of LOX-1+ PMN-MDSCs (units)
Association between infused MDSC dose per kilogram and aGVHD
We further evaluated the impact of the body weight-normalized doses of MDSC subsets. As shown in Fig. 2, the cumulative incidence of II-IV0 aGVHD was significantly lower in patients receiving a high dose of total MDSCs per kilogram (12.94%; 95% CI: 7.46–22.46) compared to the low-dose group (30.59%; 95% CI: 22.21–42.13; P = 0.006; Fig. 2A). This protective effect was also significant for PMN-MDSCs, with incidences of 14.12% in the high-dose group versus 29.41% in the low-dose group (P = 0.014; Fig. 2C). A strong, non-significant trend was observed for M-MDSCs (16.47% vs. 27.06%, P = 0.071; Fig. 2B) and for the activated LOX-1+ PMN-MDSCs subset (16.47% vs. 27.06%,P = 0.086; Fig. 2D).
Fig. 2.
Cumulative Incidence of Grade II–IV aGVHD Stratified by Body Weight-Normalized MDSC Subset Dose. A The Cumulative Incidence of Grade II-IV aGVHD between high and low dose of MDSCs (units/kg); B The Cumulative Incidence of Grade II-IV aGVHD between high and low dose of M-MDSCs (units/kg); C The Cumulative Incidence of Grade II-IV aGVHD between high and low dose of PMN-MDSCs (units/kg); D The Cumulative Incidence of Grade II-IV aGVHD between high and low dose of LOX-1+ PMN-MDSCs (units/kg)
Independent prognostic factors and optimal dose threshold
According to Fine-Gray competing risk model analyses (Table 2), univariable analysis demonstrated that higher infused doses per kilogram of body weight were significantly associated with a lower risk of grade II-IV aGVHD, including total MDSCs (> 33.3 vs. ≤ 33.3 × 10⁶/kg: sHR 0.38, 95% CI 0.19–0.75, P = 0.006) and PMN-MDSCs (> 17.5 vs. ≤ 17.5 × 10⁶/kg: sHR 0.43, 95% CI 0.22–0.85, P = 0.014). Similarly, a higher absolute count of M-MDSCs per microliter (> 2600 vs. ≤ 2600/µL: sHR 0.42, 95% CI 0.21–0.82, P = 0.012) was also associated with protection. In the multivariable analysis adjusting for potential confounders, a high dose of PMN-MDSCs (> 17.5 × 10⁶/kg) remained an independent protective factor against aGVHD (sHR 0.25, 95% CI 0.07–0.88, P = 0.039). A high absolute count of M-MDSCs (> 2600/µL) also retained its independent association (sHR 0.28, 95% CI 0.10–0.78, P = 0.015). The dose of total MDSCs was not independently associated in the multivariable model. To establish a clinically applicable threshold for PMN-MDSCs, the optimal cut-off value was determined directly from the Fine-Gray model, which identified 11.3 × 10⁶/kg. The robustness of this cut-off was supported by bootstrap resampling (median: 12.13 × 10⁶/kg; 95% CI: 9.98–89.04). Based on this threshold, patients were stratified into a PMN-MDSC low-dose group (≤ 11.3 × 10⁶/kg; n = 52) and a high-dose group (> 11.3 × 10⁶/kg; n = 118). Baseline characteristics of this two groups are summarized in Supplementary Table S2. Consistent with this stratification, the cumulative incidence of II–IV0 aGVHD was significantly lower in the PMN-MDSC high-dose group compared with the low-dose group (14.41%; 95% CI, 9.28–22.36 vs. 38.46%; 95% CI, 27.27–54.24; P < 0.001).
Table 2.
Fine-Gray Competing Risk Regression Analysis of Factors Associated with II–IV0 aGVHD
| Characteristic | Univariable | Multivariable | ||
|---|---|---|---|---|
| sHR (95% CI) | p-value | sHR (95% CI) | p-value | |
| MDSCs(units/kg) | ||||
| ≤33.3 | reference | reference | ||
| ＞33.3 | 0.38 (0.19 -0.75) | 0.006 | 0.71 (0.21 -2.46) | 0.590 |
| M-MDSC (units/kg) | ||||
| ≤7 | reference | reference | ||
| ＞7 | 0.55 (0.29 -1.05) | 0.071 | 0.84 (0.31 -2.31) | 0.730 |
| PMN-MDSCs(units/kg) | ||||
| ≤17.5 | reference | reference | ||
| ＞17.5 | 0.43 (0.22 -0.85) | 0.014 | 0.25 (0.07 -0.88) | 0.039 |
| LOX1+PMN-MDSCs(units/kg) | ||||
| ≤2.5 | reference | reference | ||
| ＞2.5 | 0.56 (0.29 -1.09) | 0.086 | 1.16 (0.46 -2.92) | 0.750 |
| MDSCs (%) | ||||
| ≤5 | reference | reference | ||
| ＞5 | 0.53 (0.28 -1.03) | 0.060 | 0.92 (0.36 -2.35) | 0.850 |
| M-MDSCs (%) | ||||
| ≤1 | reference | reference | ||
| ＞1 | 0.54 (0.27 -1.04) | 0.066 | 0.96 (0.31 -2.96) | 0.940 |
| PMN-MDSCs (%) | ||||
| ≤3 | reference | |||
| ＞3 | 0.64 (0.34 -1.24) | 0.187 | ||
| LOX+PMN-MDSCs (%) | ||||
| ≤16 | reference | |||
| ＞16 | 1.53 (0.80 -2.94) | 0.198 | ||
| MDSCs absolute count,/μL | ||||
| ≤14000 | reference | reference | ||
| ＞14000 | 0.47 (0.24 -0.92) | 0.027 | 0.92 (0.21 -4.11) | 0.910 |
| M-MDSCs absolute count,/μL | ||||
| ≤2600 | reference | reference | ||
| ＞2600 | 0.42 (0.21 -0.82) | 0.012 | 0.28 (0.10 -0.78) | 0.015 |
| PMN-MDSCs absolute count,/μL | ||||
| ≤8000 | reference | reference | ||
| ＞8000 | 0.55 (0.29 -1.06) | 0.074 | 0.59 (0.30 -1.21) | 0.150 |
| LOX1+PMN-MDSCs absolute count,/μL | ||||
| ≤900 | reference | |||
| ＞900 | 0.83 (0.44 -1.57) | 0.571 | ||
| Gender | ||||
| Female | reference | |||
| Male | 0.59 (0.30 -1.15) | 0.120 | ||
| Age,years | ||||
| ≤18 | reference | reference | ||
| ＞18 | 0.48 (0.25 -0.93) | 0.029 | 0.54 (0.26 -1.15) | 0.110 |
| 2nd HSCT | ||||
| No | reference | |||
| Yes | 0.59 (0.18 -1.93) | 0.380 | ||
| Disease status pre-HSCT | ||||
| CR | reference | |||
| PR/NR | 0.64 (0.25 -1.66) | 0.360 | ||
| HSCT type | ||||
| Related donors | reference | reference | ||
| Unrelated donors | 0.42 (0.13 -1.34) | 0.140 | 0.64 (0.13 -3.04) | 0.570 |
| Donor type | ||||
| HID | reference | reference | ||
| MSD | 0.40 (0.13 -1.24) | 0.110 | 0.33 (0.09 -1.27) | 0.110 |
| MNC ×108/kg | ||||
| ≤6 | reference | |||
| ＞6 | 1.30 (0.68 -2.49) | 0.420 | ||
| CD34 ×106/kg | ||||
| ≤4 | reference | reference | ||
| ＞4 | 1.30 (0.68 -2.49) | 0.091 | 0.70 (0.33 -1.48) | 0.350 |
sHR subdistribution hazard ratio
Impact of PMN-MDSC dose on survival outcomes and relapse
With a median post-transplant follow-up of 518 days (IQR, 260-584 days), the PMN-MDSC high-dose group exhibited significantly superior 300-day overall survival (OS) compared to the low-dose group (92.37% [95% CI, 87.58-97.16] vs. 74.91% [62.44-87.38], P = 0.005; Figure 3A). The benefit was even more pronounced for GVHD-free relapse-free survival (GRFS), with rates of 76.27% [68.60-83.95] versus 40.38% [26.37-54.40] (P < 0.001; Figure 3B). To investigate the potential impact on the graft-versus-leukemia (GVL) effect, we analyzed outcomes in the subset of patients with malignant diseases (n=147). While a high dose of PMN-MDSCs (>8,000/μL) was associated with a strong trend towards lower relapse and significantly better 300-day leukemia-free survival (LFS) (Table 3), the most striking effect was observed for the LOX-1⁺ PMN-MDSC subset. Infusion of a high dose of LOX-1⁺ PMN-MDSCs (>900/μL) was associated with a significantly lower cumulative incidence of relapse (1.37% vs. 11.42%, P=0.015) and a markedly higher 300-day LFS (90.41% vs. 68.55%, P=0.002) compared to the low-dose group (Table 3).
Fig. 3.
Survival Outcomes Stratified by the Infused Dose of PMN-MDSCs. A Overall survival between PMN-MDSC high-dose and low-dose groups; B GVHD-free relapse-free survival between PMN-MDSC high-dose and low-dose groups
Table 3.
Univariable Analysis of the Association between Infused MDSC Subsets and Relapse or Leukemia-Free Survival in Patients with Malignant Diseases
| Characteristic | RI(95% CI) | p-value (Gray's Test) | 300-day LFS(95% CI) | p-value (Logrank Test) |
|---|---|---|---|---|
| MDSCs,/μL | 0.277 | 0.007 | ||
| ≤14000 | 8.8% (4.1-19.1) | 69.68% (58.65-80.71) | ||
| ＞14000 | 4.1% (1.3-12.3) | 89.19% (82.11-96.26) | ||
| M-MDSCs,/μL | 0.695 | 0.500 | ||
| ≤2600 | 7.49% (3.18-17.66) | 76.31% (65.93-86.68) | ||
| ＞2600 | 5.41% (2.08-14.02) | 82.42% (73.76-91.10) | ||
| PMN-MDSCs,/μL | 0.271 | 0.005 | ||
| ≤8000 | 8.73% (4.04-18.85) | 69.68% (58.73-80.64) | ||
| ＞8000 | 4.00% (1.32-12.12) | 89.33% (82.35-96.32) | ||
| LOX-1+ PMN-MDSCs,/μL | 0.015 | 0.002 | ||
| ≤900 | 11.42% (5.90-22.11) | 68.55% (57.43-79.67) | ||
| ＞900 | 1.37% (0.20-9.59) | 90.41% (83.66-97.17) |
LFS leukemia-free survival, RI relapse incidence
Post-transplant kinetics of circulating MDSCs and temporal association with aGVHD onset
The longitudinal reconstitution kinetics of circulating MDSCs were monitored after transplantation. The levels peaked at day +30 (median, 27/μL; IQR, 9–59) and subsequently declined to a median of 9/μL (IQR, 3–25) by day +90 (Figure 4A). While patients in the high-dose PMN-MDSC group consistently showed numerically higher MDSC levels across most time points compared to the low-dose group, this difference was not statistically significant in a two-way ANOVA accounting for time and group (main effect of group: P = 0.464) (Figure 4B). MDSC levels varied significantly over time (main effect of time: P = 0.027), with both groups exhibiting parallel kinetic trajectories (group-by-time interaction: P = 0.932). We then evaluated the temporal relationship between this immunologic reconstitution and clinical events. Among the 37 patients who developed grade II-IV aGVHD, the median time to onset was day +29 (IQR, 21–56). Notably, this peak incidence of aGVHD occurred immediately following the peak in circulating MDSC levels (around day +28) and coincided with the onset of their subsequent decline (Figure 4A).
Fig. 4.
Reconstitution Kinetics of Circulating MDSCs (median with interquartile ranges) and Timing of aGVHD Onset. A The longitudinal reconstitution kinetics of circulating MDSCs; B Comparison of MDSC levels at different time points after allo-HSCT between the PMN-MDSC high-dose and low-dose groups
Discussion
Leukemia is a malignant hematologic disorder characterized by abnormal clonal proliferation of hematopoietic stem cells, and allo-HSCT has become an important therapeutic strategy [10, 11]. In recent years, the number of patients undergoing allo-HSCT has increased steadily, with nearly 1.5 million procedures successfully performed in over 1,500 transplant centers worldwide [1]. G-CSF has been widely used in clinical practice as the primary mobilizing agent [12]. As a member of the hematopoietic growth factor family, G-CSF binds to its specific receptor (G-CSFR), activating downstream signaling pathways that promote myeloid progenitor proliferation, granulocytic differentiation, and the mobilization of hematopoietic stem cells from the bone marrow into peripheral circulation [13]. Although the mechanisms of G-CSF in immune regulation have not yet been fully clarified, emerging evidence suggests that it may reshape the donor immune microenvironment and influence both donor and recipient immune states [14, 15]. Our previous study demonstrated that G-CSF mobilization significantly increased donor MDSCs and revealed preliminary associations between extreme PMN-MDSC doses and post-transplant complications [9]. Building on these findings, we sought to systematically determine whether the infused dose of donor MDSCs, within a broader clinical range, directly influences the risk of severe acute GVHD, utilizing a larger cohort and longitudinal design.
The immune microenvironment is composed of diverse immune cells and soluble factors, in which MDSC is a newly discovered, heterogeneous, immature bone marrow cell population with significant immunosuppressive functions [16–18], capable of suppressing innate and adaptive immune responses. Its main functions include inhibiting the proliferation and activation of T cells, regulating inflammatory responses, participating in hematopoietic reconstitution and promoting the growth and metastasis of tumors. MDSCs exert immune suppression function through a variety of mechanisms [19, 20]: (1) degrading arginine via arginase and inducible nitric oxide synthase (iNOS), thereby reducing T-cell CD3ζ expression, impairing proliferation, and inhibiting T-cell activity; (2) producing reactive oxygen species (ROS), which exert cytotoxic effects on immune cells and block T-cell activation, while also secreting immunosuppressive enzymes such as indoleamine-2,3-dioxygenase, cyclooxygenase, and heme oxygenase; (3) releasing transforming growth factor-β and IL-10 to promote regulatory T cell (Treg) generation and induce M2 macrophage polarization [21]; and (4) expressing galectin-9, which interacts with T-cell TIM-3 to attenuate CD8+ T-cell effector function [22]. As GVHD is mainly induced by dysregulated immune response, the immunosuppressive MDSCs can reduce both the incidence and severity of aGVHD [23]. Consistent with this, our Fine-Gray competing risk analyses identified the infused dose of PMN-MDSCs per kilogram of body weight as an independent protective factor against aGVHD (sHR 0.25, P = 0.039). Crucially, we established a clinically actionable threshold of 11.3 × 10⁶/kg, derived directly from the survival model, which robustly stratified patient risk. A pivotal finding is that this PMN-MDSC dose threshold conferred significant survival benefits. Patients receiving a dose > 11.3 × 10⁶/kg experienced superior overall survival and markedly improved GVHD-free relapse-free survival.
LOX-1+ PMN-MDSCs represent a functionally enhanced subset of PMN-MDSCs, which have been shown to expand in multiple cancers and correlate with poor prognosis [24]. These cells contribute to the tumor-associated immunosuppressive milieu by inhibiting effector T-cell activity, promoting Treg induction, and secreting pro-inflammatory cytokines [25]. Their immunosuppressive potency is closely linked to the inhibitory signals they express [26]. LOX-1+ PMN-MDSCs exert stronger suppressive effects through mechanisms such as endocytosis and inflammatory mediator release, which modulate T-cell function [24]. In our cohort, the absolute counts of LOX-1+ PMN-MDSCs were reduced in patients with GVHD (Supplementary Figures S2), and showed the importance of both MDSC abundance and functional status in immune regulation, although the underlying mechanisms warrant further investigation. Moreover, analysis within the malignant disease subset revealed a potential “GVHD-GVL separation” effect associated with specific MDSC subsets. While high-dose PMN-MDSCs were linked to better leukemia-free survival, the most striking reduction in relapse incidence was associated with a high dose of the activated LOX-1⁺ PMN-MDSC subset. This in vivo clinical observation aligns with and extends the pre-clinical findings reported for iPSC-derived MDSCs (iMDSCs) [27]. For instance, recent international studies have shown that “off-the-shelf” iMDSCs can serve as a novel therapeutic strategy for aGVHD [27]. In GVHD mouse models, infusion of iMDSCs significantly improved survival while preserving graft-versus-leukemia (GVL) activity, thus achieving a favorable “separation effect” mechanistically. RNA sequencing and knockdown studies identified phosphoglycerate dehydrogenase (PHGDH), a key enzyme in the serine biosynthesis pathway, as essential for protecting iMDSCs from inflammasome activation, thereby maintaining their “inflammation-resistant” phenotype. The potent effect of LOX-1⁺ cells underscores that both the quantity and the functional activation state of MDSCs may be critical for optimal immune modulation after transplant.
We also monitored the post-transplant kinetics of circulating MDSCs. While levels varied significantly over time, the reconstitution trajectory did not differ significantly between the high- and low-dose PMN-MDSC groups.
However, a critical temporal relationship was observed: the peak incidence of aGVHD (median onset, day + 29) closely followed the peak in circulating MDSC levels and coincided with the onset of their decline. This suggests that the period of waning MDSC-mediated suppression may represent a vulnerable window for aGVHD initiation. Although not statistically significant, the numerically higher MDSC levels in the high-dose group at most time points are consistent with the hypothesis that a larger graft MDSC reservoir might help sustain suppression and mitigate this risk. Other studies [28] have similarly highlighted that delayed M-MDSC reconstitution and insufficient IL-10 production may impair immunosuppressive function, facilitating donor T-cell–mediated tissue damage and increasing aGVHD risk. Conversely, premature or excessive M-MDSC expansion may induce profound immunosuppression, weakening antiviral surveillance and predisposing to CMV reactivation. Notably, patients receiving post-transplant cyclophosphamide (PTCy)-based prophylaxis exhibited earlier M-MDSC recovery, providing further insights into conditioning regimen design. These findings are consistent with our center’s observations and highlight important directions for future research.
Our data demonstrate that the protective effect is specific to the MDSC compartment and not merely a consequence of infusing more total cells, as the CD34⁺ cell dose was not an independent prognostic factor in multivariable analysis. This distinction is clinically vital. Indiscriminately increasing stem cell harvest is not feasible or safe. Instead, future strategies should aim to optimize graft composition—for example, by refining mobilization protocols or developing ex vivo manipulation techniques—to enhance the yield of functional MDSCs, particularly the protective PMN-MDSC subset, without augmenting undesirable cell populations.
This study has certain limitations, including its retrospective, single-center design and the lack of direct functional assessment of MDSC immunosuppressive activity. Further validation in larger prospective cohorts is warranted. Future studies should investigate the phenotypic and functional heterogeneity of MDSCs, their temporal evolution, and interactions with other immune cells and cytokines within distinct microenvironments, to better predict aGVHD risk and optimize personalized therapeutic strategies.
In conclusion, this study demonstrates that the absolute dose of PMN-MDSCs in G-CSF-mobilized grafts is an independent determinant of aGVHD risk and survival outcomes, with a defined threshold of 11.3 × 10⁶/kg. These findings position PMN-MDSC dose as a clinically actionable biomarker for risk stratification and a promising target for optimizing graft composition in allo-HSCT.
Supplementary Information
Acknowledgements
Not applicable.
Abbreviations
- GVHD
Graft-versus-host disease
- HSCT
Hematopoietic stem cell transplantation
- ALL
Acute lymphoblastic leukemia
- AML
Acute myeloid leukemia
- AA
Aplastic anemia
- MDS
Myelodysplastic syndromes
- CR
Complete remission
- PR/NR
Partial remission/ Non-remission
- G-CSF
Granulocyte colony-stimulating factor
- HID
Haploidentical donor
- MSD
Matched sibling donor
- BU
Busulfan
- TBI
Total body irradiation
- MNC
Mononuclear cells
- PBSC
Peripheral blood stem cells
- CSA
Cyclosporine
Authors’ contributions
Man Chen and Xue-qiao Wang conceived and designed the study. Man Chen drafted the initial manuscript and analyzed the data. Yi Li and Wei Zhao reviewed the initial manuscript. Jing Long, Hui Wang, Min-jing Fu collected and provided patient clinical data. All authors contributed to the article and approved the submitted version.
Funding
This study was supported by the 2023 Hebei Provincial Medical Science Research Project (Grant No. 20232042) and the National Natural Science Foundation of China (Grant No. 82472346).
Data availability
All data generated or analysed during this study are included in this article. Data are available upon request from the corresponding author.
Declarations
Ethics approval and consent to participate
This study was approved by the Ethics Committee of Beijing Lu Daopei Hospital (ethics No. DPEC-M-202309) and conforms to the principles outlined in the Declaration of Helsinki (Clinical trial number: not applicable). With written informed consent obtained from all adult participants and from parents of participants under 18 years of age prior to enrollment.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Man Chen and Xue-qiao Wang contributed equally to this work.
Contributor Information
Yi Li, Email: liyiscu@outlook.com.
Wei Zhao, Email: angelina_zw@163.com.
References
- 1.Granot N, Storb R. History of hematopoietic cell transplantation: challenges and progress. Haematologica. 2020;105(12):2716–29. 10.3324/haematol.2019.245688. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Gooptu M, Antin JH, GVHD Prophylaxis. 2020. Front Immunol. 2021;12:605726. 10.3389/fimmu.2021.605726 [DOI] [PMC free article] [PubMed]
- 3.Ghimire S, Weber D. Pathophysiology of GvHD and other HSCT-Related major complications. Front Immunol. 2017;8:79. 10.3389/fimmu.2017.00079. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hess NJ, Kink JA, et,al., Exosomes. MDSCs and tregs: A new frontier for GVHD prevention and treatment. Front Immunol. 2023;14:1143381. 10.3389/fimmu.2023.1143381. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Demosthenous C, Sakellari I. The role of Myeloid-Derived suppressor cells (MDSCs) in Graft-versus-Host disease (GVHD). J Clin Med. 2021;10(10):2050. 10.3390/jcm10102050. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Khoury JD, Solary E. The 5th edition of the world health organization classification of haematolymphoid tumours: myeloid and Histiocytic/Dendritic neoplasms. Leukemia. 2022;36(7):1703–19. 10.1038/s41375-022-01613-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Chinese Society of Hematology, Chinese Medical Association. Chinese guidelines for the diagnosis and treatment of myelodysplastic syndromes (2019 edition). Chin J Hematol. 2019;40(2):89–97. 10.3760/cma.j.issn.0253-2727.2019.02.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Fan JW, Fan Y. The ratio of CD226 and TIGIT expression in Tfh and PD-1 + ICOS + Tfh cells are potential biomarkers for chronic Antibody-Mediated rejection in kidney transplantation. J Immunol Res. 2022;2022:5326083. 10.1155/2022/5326083. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wang R, Chen M. A research on the influence of G-CSF mobilization on donor’s peripheral blood MDSCs and its relationship with patient prognosis. Int Immunopharmacol. 2023;124(Pt B):110998. 10.1016/j.intimp.2023.110998. [DOI] [PubMed] [Google Scholar]
- 10.Bewersdorf JP, Abdel-Wahab O. Translating recent advances in the pathogenesis of acute myeloid leukemia to the clinic. Genes Dev. 2022;36(5–6):259–77. 10.1101/gad.349368.122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Stem Cell Application Group, Chinese Society of Hematology, Chinese Medical Association. Chinese expert consensus on allogeneic hematopoietic stem cell transplantation for hematologic diseases (III): acute graft-versus-host disease (2020 edition). Chin J Hematol. 2020;41(7):529–36. 10.3760/cma.j.issn.0253-2727.2020.07.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Chinese Society of Hematology, Chinese Medical Association; Anti-Lymphoma Alliance of Chinese Society of Clinical Oncology. Chinese expert consensus on autologous hematopoietic stem cell mobilization and collection for lymphoma (2020 edition). Chin J Hematol. 2020;41(12):979–83. 10.3760/cma.j.issn.0253-2727.2020.12.003. [Google Scholar]
- 13.Rahi V, Jamwal S, Kumar P. Neuroprotection through G-CSF: recent advances and future viewpoints. Pharmacol Rep. 2021;73(2):372–85. 10.1007/s43440-020-00201-3. [DOI] [PubMed] [Google Scholar]
- 14.Franzke A, Piao W. G-CSF as immune regulator in T cells expressing the G-CSF receptor: implications for transplantation and autoimmune diseases. Blood. 2003;102(2):734–9. 10.1182/blood-2002-04-1200. [DOI] [PubMed] [Google Scholar]
- 15.Zhao X, Peng T. In vivo G-CSF treatment activates the GR-SOCS1 axis to suppress IFN-γ secretion by natural killer cells. Cell Rep. 2022;40(11):111342. 10.1016/j.celrep.2022.111342. [DOI] [PubMed] [Google Scholar]
- 16.Gabrilovich DI, Bronte Vet,al. The terminology issue for myeloid-derived suppressor cells. Cancer Res. 2007;67(1):425. 10.1158/0008-5472.CAN-06-3037. author reply 426. [DOI] [PMC free article] [PubMed]
- 17.Bizymi N, Georgopoulou A. Myeloid-Derived suppressor cells (MDSC) in the umbilical cord blood: biological significance and possible therapeutic applications. J Clin Med. 2022;11(3):727. 10.3390/jcm11030727. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Jain N, Wierda WG, O’Brien S. Chronic lymphocytic leukaemia. Lancet. 2024;404(10453):694–706. 10.1016/S0140-6736(24)00595-6. Epub 2024 Jul 25. [DOI] [PubMed] [Google Scholar]
- 19.Lu J, Luo Y. Myeloid-derived suppressor cells in cancer: therapeutic targets to overcome tumor immune evasion. Exp Hematol Oncol. 2024;13(1):39. 10.1186/s40164-024-00505-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Yin Z, Li C. Myeloid-derived suppressor cells: roles in the tumor microenvironment and tumor radiotherapy. Int J Cancer. 2019;144(5):933–46. 10.1002/ijc.31744. [DOI] [PubMed] [Google Scholar]
- 21.Veglia F, Perego M. Myeloid-derived suppressor cells coming of age. Nat Immunol. 2018;19(2):108–19. 10.1038/s41590-017-0022-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.D’Aveni M, Notarantonio AB. Myeloid-Derived suppressor cells in the context of allogeneic hematopoietic stem cell transplantation. Front Immunol. 2020;11:989. 10.3389/fimmu.2020.00989. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Pawelec G, Picard E. MDSCs, ageing and inflammageing. Cell Immunol. 2021;362:104297. 10.1016/j.cellimm.2021.104297. [DOI] [PubMed] [Google Scholar]
- 24.Coudereau R, Waeckel L. Emergence of immunosuppressive LOX-1 + PMN-MDSC in septic shock and severe COVID-19 patients with acute respiratory distress syndrome. J Leukoc Biol. 2022;111(2):489–96. 10.1002/JLB.4COVBCR0321-129R. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Li X, Liu J. Polymorphonuclear myeloid-derived suppressor cells link inflammation and damage response after trauma. J Leukoc Biol. 2021;110(6):1143–61. 10.1002/JLB.3MA0821-029R. [DOI] [PubMed] [Google Scholar]
- 26.Peterlin P, Béné MC. Assessment of monocytic-myeloid-derived suppressive cells (M-MDSC) before and after allogeneic hematopoietic stem cell transplantation in acute leukemia patients. EJHaem. 2023;4(4):1089–95. 10.1002/jha2.795. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Ma L, Koehn BH. et,al. Inflammasome-resistant IPSC-derived myeloid-derived suppressor cells ameliorate xenogeneic graft-versus-host disease. Blood 2025 Jul 23:blood2025028562. 10.1182/blood.2025028562 [DOI] [PMC free article] [PubMed]
- 28.Thammahong A, Wudhikarn K. Dynamics of monocytic myeloid-derived suppressor cell recovery and interleukin-10 production on graft-versus-host disease and cytomegalovirus reactivation after allogeneic hematopoietic cell transplantation. Transpl Immunol. 2025;92:102263. 10.1016/j.trim.2025.102263. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All data generated or analysed during this study are included in this article. Data are available upon request from the corresponding author.