Haplo-identical HSCT (haplo-HSCT) has substantially broadened the utility of HSCT for patients without HLA-matched donors. However, haplo-HSCT largely relies on triple immunosuppression with PT-Cy + tacrolimus + MMF, which are associated with toxicities and with defects in post-HSCT immune reconstitution. Therefore, a targeted, less globally immunosuppressive regimen could have a substantial impact on the success of HSCT. CD137 (4-1BB) is a major costimulation receptor that is rapidly upregulated on donor T cells. The immediate and restricted expression of CD137 on alloreactive T cells makes this an ideal drug target for AGVHD prevention. We have developed an anti-CD137-antibody drug conjugate (CD137-ADC) with a short T1/2 (2 days), that selectively depletes activated CD137+ T cells following a single dose. We have demonstrated the ability of this ADC to control AGVHD and enable immune reconstitution with tolerogenic subsets of donor T cells.

Methods: Haplo-HSCT in non-human primates, including the assessment of donor chimerism and detailed clinical monitoring, were performed as previously reported (Tkachev et al. Science Transl. Med. 2021). We compared HSCT without GVHD prophylaxis ('NoRx') to that performed using a single dose of CD137-ADC (0.3mg/kg) given 4 hours post-HSCT. The immunologic impact of CD137-ADC was determined by flow cytometry and single-cell (sc)RNA-Seq.

Results: (1) Engraftment: All CD137-ADC HSCT recipients engrafted with donor cells, with chimerism >90% by Day 28, and neutrophil and platelet engraftment at 12, and 11 days, respectively. (2) GVHD: All NoRx animals developed severe AGVHD necessitating euthanasia, with median onset of 8 days post-HSCT. In contrast, the median survival of the CD137-ADC cohort was 97 days (p<0.01). Three of 4 recipients were long-term survivors, with GVHD limited to spontaneously resolving skin AGVHD early after HSCT (Panel 1). (3) Toxicities: Importantly, no treatment-associated systemic toxicities were detected, including no CMV reactivation or end-organ disease, and no PTLD. (4) Immune impact: Flow cytometry revealed depletion of activated PD1+ CD4 central and effector-memory T cells in the CD137-ADC vs NoRx cohorts (Tcm/Tem; 1.2% vs 10.9% and 0.6 vs 7.6%, p<0.01) and PD1+ CD8 Tem (5% vs 24.3%, p<0.01), consistent with these being direct targets of the ADC. In contrast, the CD137-ADC spared OX40+PD1- CD4 Tcm/Tem compared to NoRx (11.7 vs 3%, and 12.1 vs 2.6%, p<0.01), OX40+PD1- CD8 Tem (4.1 vs 1.1%, p<0.01), CD4 naïve (9.3 vs 1.8, p<0.01) and CD8 memory stem T cells (6.3 vs 1.5%, p<0.01). Importantly, by Day 14 post-HSCT, there was restored reconstitution of proliferative OX40+ CD4 Tcm/Tem, CD28+CD127+OX40+ and CD28+CD127- CD8 transitional/effector-memory T cells (Ttm/em), followed by reconstitution of CD8 Tem, CD4 and CD8 effector T cells. However, unlike in the NoRx cohort, this reconstitution did not induce GVHD. To identify the unique T cell subtypes/states that correlated with GVHD-sparing immune reconstitution, we performed scRNA-Seq. This analysis identified 36 discrete T cell states across NoRx and CD137-ADC cohorts, with several cohort-specific clusters (Panel 2). For NoRx, we identified enrichment for a T cell cluster (Cluster 1) characterized by expression of the chemokines PF4 (CXCL4), PPBP (CXCL7), CCL23 and CCL5 along with Granzymes A/B/K and other cytotoxic molecules (KLRC2, NKG7, GNLY). In contrast, the CD137-ADC cohort showed enrichment for a CD8 T cell cluster (Cluster 2) expressing EOMES and Granzymes K/M, along with other naïve/early T cell memory-associated genes (SELL, CCR7, CXCR5, TCF7) and effector/exhaustion-associated genes (TBX21, TOX). The CD137-ADC cohort was also enriched for T cells expressing TCR gamma-9 and delta-2 chains (Cluster 3), suggesting enhanced reconstitution of the protective g9d2 T cell subset in these recipients.

Conclusions: We have identified CD137-ADC as a novel targeted strategy to deplete CD137+ activated T cells early after haplo-HSCT, protecting against AGVHD with a single dose, in the absence of systemic toxicities or viral reactivation. The T cells that expand after CD137-ADC demonstrate a unique flow cytometric and transcriptomic profile, suggesting tolerogenic rather than pathogenic T cell immune reconstitution with this approach. To our knowledge, this is the first use of an ADC to control GVHD and introduces an important new prophylaxis paradigm for haplo-HSCT.

Gerdemann:AlloVir: Divested equity in a private or publicly-traded company in the past 24 months, Patents & Royalties. Lanieri:Magenta Therapeutics: Current Employment. Olson:Abbvie: Current equity holder in publicly-traded company, Ended employment in the past 24 months, Patents & Royalties; Magenta Therapeutics: Current Employment, Current equity holder in publicly-traded company; Enthera: Current holder of stock options in a privately-held company, Membership on an entity's Board of Directors or advisory committees; Lassen Therapeutics: Current holder of stock options in a privately-held company, Membership on an entity's Board of Directors or advisory committees. Kean:Bluebird Bio: Research Funding; Kite/Gilead: Research Funding; Magenta: Research Funding; EMD-Serono: Research Funding; HiFiBio: Consultancy; Mammoth Biosciences: Current equity holder in private company; Bristol Myers Squibb: Research Funding; Vertex: Consultancy.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution