An optimized MSC product promotes GvHD prophylaxis through exosome-mediated regulatory T cell induction Free

We are currently performing clinical studies (NCT05589896 and ossiumhealth.com/HOPE) with cryopreserved HPC, Marrow recovered from organ donor vertebral bodies (VBs) and banked for subsequent allogeneic hematopoietic cell transplant (alloHCT). As reported elsewhere, outcomes have been positive using highly HLA mismatched (4-6/8) grafts with post-transplant cyclophosphamide (PTCy). A low incidence of mild to moderate steroid responsive acute Graft-vs-Host-Disease (aGvHD) was observed. No incidences of chronic GVHD has been reported thus far. Although aGVHD was manageable with standard treatment, the resulting morbidity could still impact a patient's quality of life. Thus, we are developing a complementary VB-derived mesenchymal stem/stromal cell (MSC) GVHD prophylaxis. Potency has been enhanced by priming MSCs with interferon-γ (IFNγ) and synchronizing cells prior to cryopreservation, which maintains immediate post-thaw immunomodulatory function. This enhanced MSC cellular therapy has been termed γSMCs (IFNγ-primed, synchronized MSCs).

We have conducted in vitro and in vivo studies in GvHD models. In vitro suppression of T cell activation by γSMCs was significantly better (p<0.01) than with unmodified MSCs isolated from the same donor and otherwise cultured in the same manner. It was discovered that a major mechanism by which γSMCs suppress T cell activation is through expression of bioactive extracellular vesicles (EVs). Purified EVs exhibited dose-dependent induction of regulatory T cells (Tregs). Isolated peripheral blood T cells from 4 donors were independently stimulated with CD3/CD28 antibodies for 3 d in the absence or presence of purified EVs (3.2x10¹⁰ – 2x10¹¹). Treg induction was dose-dependent; more than doubling at the high EV dose (p<0.05, one-way ANOVA, Dunnett's test). Regulatory interleukin 10 (IL10) cytokine levels increase by >8-fold to >2000 pg/ml (p<0.001). Furthermore, EV-treated T cells dose-dependently suppressed HLA-mismatched allogeneic T cell activation, with complete inhibition at the highest ratio (1:2) tested. Visualization of EV uptake by T cells using a FlowSight imaging cytometer suggested a direct Treg induction mechanism. EV-associated transforming growth factor-β (TGFβ; 645 ±45 pg/ml per 10¹¹ EVs), a potent immunomodulatory cytokine, was shown to be the principal mediator of Treg induction. Pretreating EVs with a neutralizing TGFβ antibody abrogated Treg induction, diminishing to 3.9±0.5% compared to 13.1±0.6% with isotype control (p<0.001; one-way ANOVA, Tukey's test).

Based on these encouraging in vitro data, we conducted a study in an allogeneic mouse model of moderate aGvHD. Irradiated (2.5 Gy) BALB/c (H-2Kd MHC haplotype) mice (n=15/group) were transplanted with bone marrow and splenocytes from C57BL/6 (H-2Kb) donors. Treatments were either (1) vehicle, (2) unmodified MSCs, (3) γSMCs or (4) purified EVs, all twice weekly by tail vein injection beginning at 3 d. Doses were 2x10⁶ cells or 5x10⁹ EVs. Kaplan-Meier survival curves indicated a clear survival benefit by human γSMCs treatment with 60% of mice surviving to 28 d compared to only 40% in both the control (vehicle) and unmodified MSC groups. Treatment with EVs delayed GVHD deaths by ~5 days and the median time of survival was significantly (p<0.05; ANOVA, Dunnett's) longer for mice that died in the EV-treated (21d, IQR: 17-23.5) versus the control group (10 d, IQR: 10-17). However, mortality at 28 d (47%) was similar to control, suggesting that EVs may have only partially inhibited murine effector T cell activation. The protective mechanism of human EVs in this xenogeneic model is unclear given that Treg levels in surviving mice did not differ.

Given that γSMCs are being developed initially for GvHD prophylaxis in the scenario of alloHCT with mMUD and PTCy, we wished to determine the optimal timing of treatment that did not diminish PTCy effects. Interestingly, EVs promoted memory T (Tmem) and naïve T (Tn) cell proliferation when added to T cells at the same time as treatment with the Cy analog mafosfamide (MAF). Conversely, adding EVs 6 h after MAF only slightly stimulated Tmem and Tn populations while profoundly enhancing Tregs.

These studies indicate that γSMCs, in part through the action of secreted EVs, is a potential aGvHD prophylaxis agent. Additional studies will be performed to assess efficacy, mechanisms and optimal dose and timing in a humanized mouse GvHD model as well as a bone marrow-on-a-chip system.