Time for iMDSCs: off-the-shelf MDSCs to control GVHD Free

In this issue of Blood, Ma et al¹ report a novel approach to generate off-the-shelf, myeloid-derived suppressor cells (MDSCs) from human-induced pluripotent stem cell (iPSC)–derived CD34⁺ cells. This product overcomes several limitations that have hindered the clinical translation of MDSCs.

Graft-versus-host disease (GVHD) remains a major cause of morbidity in allogeneic hematopoietic stem cell transplantation recipients.² Although several new therapies have improved outcomes in recent years, adoptive cellular therapies hold promise for patients with refractory disease.³ Regulatory T cells and mesenchymal stromal/stem cells (MSCs) are currently under clinical investigation.³ MDSCs are another immunosuppressive cell type with in vitro and in vivo efficacy.⁴ However, challenges such as low yield from peripheral blood (PB), inconsistent product quality, and functional loss in inflammatory conditions have impeded MDSC development. Ma et al now describe an efficient platform to generate large numbers of MDSCs from iPSC-derived CD34⁺ cells. These cells, termed iPSC-derived MDSCs (iMDSCs), not only expand robustly but also retain their suppressive function under proinflammatory conditions, likely due to the sustained expression of phosphoglycerate dehydrogenase (PHGDH), a purine-metabolizing enzyme.

MDSCs are a heterogeneous group of immature myeloid cells distinct from terminally-differentiated monocytes and neutrophils.⁵ In humans, early-stage MDSCs, monocytic MDSCs (M-MDSCs), and polymorphonuclear MDSCs (PMN-MDSCs) represent phenotypically- and morphologically-distinct subsets.⁶ Early-stage MDSCs are Lin^–/HLA-DR^–/CD33⁺, M-MDSCs are CD11b⁺CD14⁺HLA-DR^low/–CD15^–, and PMN-MDSCs are CD14^–CD11b⁺CD15⁺(or CD66b⁺).⁶ Their shared hallmark is the suppression of T-, B-, and natural killer-cell activation, commonly assessed in vitro by the inhibition of T-cell proliferation and interferon gamma production.⁶ Although partially overlapping in function, M-MDSCs and PMN-MDSCs differ in transcription factors and suppressive mechanisms: M-MDSCs use NOS-2–generated nitric oxide and cytokines such as interleukin-10 (IL-10) and transforming growth factor β, whereas PMN-MDSCs produce reactive oxygen species (ROS), arginase 1, and prostaglandin E₂.⁷ 

Despite being implicated in tumor progression through immune suppression and perimetastatic niche formation, MDSC transfer may benefit conditions such as GVHD.⁵ The investigators previously generated mouse MDSCs from bone marrow cells using granulocyte colony–stimulating factor (CSF) and granulocyte-macrophage CSF (GM-CSF), which reduced lethality via arginase-1–mediated arginine depletion.⁴ Furthermore, MDSCs were found to inhibit GVHD development in preclinical models by the induction of T helper 2 responses.⁸ However, a later study revealed that inflammasome activation in the inflammatory GVHD setting impaired MDSC function.⁹ 

The current study addresses key barriers to off-the-shelf MDSC use. First, the authors present a novel method for the expansion of MDSCs. iPSC-derived CD34⁺ cells were cultured in cytokine-enriched media on OP9 stromal cells expressing the Notch ligand DLL4. Over 19 days (including a 7-day IL-3– and M-CSF–based expansion phase) cell numbers increased >450-fold, far exceeding PB-MDSC yields. iMDSCs suppressed T-cell proliferation comparably to PB-MDSCs, and this effect was cell-contact dependent. Most iMDSCs were CD14⁺ (M-MDSC–like), but a CD14^– PMN-MDSC–like subset was also present. Notably, CD14⁺ cells had significantly higher suppressive activity than CD14^– cells, correlating with the increased expression of PU.1, CEBPβ, IRF8, pSTAT1, and pSTAT3. Transcriptomic analysis revealed the enrichment of ROS-related genes in CD14⁺ iMDSCs, suggesting a functional mechanism. Notably, blocking Siglec-9, Galectin-9, or TIM-3 did not impair suppressive function, indicating these are not essential mediators.

Given that MDSCs lose function in inflammatory environments via inflammasome activation, Ma et al evaluated iMDSC sensitivity to this mechanism. iMDSCs retained suppressive activity after adenosine triphosphate + lipopolysaccharide exposure, unlike PB-MDSCs, which became dysfunctional. RNA sequencing revealed PHGDH, a key enzyme in serine biosynthesis, as a candidate regulator. Although PHGDH expression declined in PB-MDSCs after inflammasome activation, iMDSCs maintained expression. Serine is essential for nucleic acid and protein synthesis, as well as one-carbon metabolism. Thus, it is conceivable that impaired de novo serine synthesis due to the loss of PHGDH might compromise MDSC function. Notably, small interfering RNA–mediated knockdown of PHGDH in iMDSCs impaired their suppressive capacity, supporting its role in maintaining MDSC activity under stress.

iMDSCs were also tested in a xenogeneic GVHD model. These cells preferentially accumulated in the lungs and liver, with limited intestinal homing. The administration of iMDSCs to irradiated immunodeficient mice receiving human PB mononuclear cells (PBMCs) improved survival and reduced GVHD-associated clinical scores and weight loss. Importantly, iMDSCs did not impair PBMC-mediated antitumor responses in a Nalm6 B-cell leukemia model.

Further studies will be needed to optimize the immunosuppressive function of iMDSCs upon prolonged cryopreservation period. In the current study, longer-term (2 weeks) cryopreservation resulted in >50% loss of immunosuppressive capacity, potentially due to reduced pSTAT1 and pSTAT3 expression. Furthermore, a constant percentage of functional CD14⁺ iMDSCs, which showed higher immunosuppressive capacity than CD14^– MDSCs, will be required. However, an attempt to increase the CD14⁺ cell yield by adding GM-CSF between day 12 and day 19 was not successful. Furthermore, in the investigated model, the homing efficiency to the intestine, the major target organ of therapy-refractory GVHD, was lower than to the lungs and liver. Transduction with an E-cadherin chimeric antigen receptor (CAR) did not change the in vivo life span of iMDSCs or result in an outcome benefit, unlike previous studies with E-cadherin CAR–transduced MSCs.¹⁰ The clinical development of iMDSCs will require optimization of both cell dose and treatment schedule. Moreover, even though the current study demonstrated that high-dose dexamethasone did not alter iMDSC phenotype or in vitro suppressive function, the impact of other GVHD therapies on iMDSC activity remains to be elucidated.

In summary, the study by Ma et al provides a promising and scalable platform for generating iPSC-derived MDSCs with preserved immunosuppressive function in inflammatory environments, offering a viable path toward further exploration of iMDSC efficacy in GVHD.

Conflict-of-interest disclosure: The author declares no competing financial interests.