The Fully Human Anti-CD47 Antibody SRF231 Has Dual-Mechanism Antitumor Activity Against Chronic Lymphocytic Leukemia (CLL) Cells and Increases the Activity of Both Rituximab and Venetoclax

Background

CD47 is over-expressed by many tumor types and protects tumor cells from destruction via tumor-intrinsic and -extrinsic means. The fully human anti-CD47 monoclonal antibody (mAb) SRF231 has previously been shown to block the "don't eat me" CD47/signal regulatory protein alpha (SIRPα) interaction and induce macrophage-mediated phagocytic uptake of CD47-expressing tumor cells, either alone or in the presence of anti-CD20 mAb. Furthermore, SRF231 inhibited tumor growth in preclinical models of aggressive non-Hodgkin lymphoma (Holland P, et al. ASH 2016). Here, we explored the activity of SRF231 against CLL cells for the first time, both as monotherapy and in combination with rituximab or venetoclax (VEN).

Methods

Peripheral blood mononuclear cells from 24 CLL patients were evaluated for CD47 surface expression by flow cytometry. Primary CLL or Jurkat target cells were treated ex vivo with SRF231 or isotype control and evaluated in phagocytosis and cell death assays. Human monocyte-derived macrophages were cocultured with fluorescently-labeled target tumor cells and exposed to SRF231 and/or rituximab (commercial supply). BH3 profiling was performed by gently permeabilizing primary CLL cells and measuring the release of cytochrome C (Cyto-C) in response to BH3-only peptides by flow cytometry. Priming for apoptosis was measured by Cyto-C release in response to BIM BH3 peptide, and pro-survival protein dependencies were measured by response to specific BH3-only sensitizer peptides. Statistical analyses were by unpaired and paired t-test with a two-tailed nominal p ≤ 0.05 considered as significant. In vivo antitumor activity was assessed using tumor xenograft studies in CB17 SCID mice. Mice with established, subcutaneous Ri-1 tumors were randomized and treated with either isotype control, SRF231, VEN (Medkoo), or combination of SRF231 and VEN.

Results

CD47 was expressed in all primary CLL cells (n = 24, median mean fluorescence intensity [MFI] 7913, range 3575-18,329) with a slightly higher expression in unmutated CLL (U-CLL) vs mutated CLL (M-CLL) samples (U-CLL median MFI = 9106, n = 8 vs M-CLL, median MFI = 7713, n = 14, 2 unknown, p = 0.047). Primary CLL cells were significantly more susceptible to phagocytosis upon ex vivo treatment with SRF231 in combination with rituximab (median % increase in phagocytosis over isotype control of 32.28% in the combination vs 11.78% with rituximab alone, n = 24, p < 0.0001). Upon coculture of Jurkat cells with macrophages, SRF231 not only induced phagocytosis (EC_50, 332 ± 65 ng/mL, n = 3), but also induced cell death of non-phagocytosed target tumor cells (EC₅₀, 295 ± 43 ng/mL, n = 3). While soluble SRF231 did not induce significant target tumor cell killing, immobilized SRF231 induced Jurkat cell and primary CLL cell death (median % alive of 34.6% in SRF231 treated cells vs 64.4% in controls, n = 24, p < 0.0001). To assess the mechanism of cell death induction, tumor cells were pretreated with a pan-caspase inhibitor, Z-VAD-FMK, which revealed that SRF231-mediated tumor cell death is caspase-independent. In primary CLL cells, BH3 profiling confirmed that SRF231 did not alter mitochondrial priming for apoptosis or pro-survival Bcl-2 family protein dependencies. Pre-treatment with the phospholipase C (PLC) inhibitor U73122 prior to SRF231 exposure partially blocked the ability of SRF231 to kill CLL cells (median % alive of 45.4% in pre-treated cells vs 25.4% in controls, n = 6, p = 0.0029). In addition to these in vitro studies, SRF231 displayed profound antitumor activity in a xenograft model of B-cell lymphoma as a single agent, and led to complete and durable tumor regression in combination with VEN.

Conclusion

Ex vivo treatment of primary CLL cells with SRF231 led to dual antitumor effects of tumor cell-extrinsic plus -intrinsic mechanisms by augmenting rituximab-induced phagocytosis and inducing tumor cell death. SRF231 induced death of tumor cells through a caspase-independent mechanism that depends at least partially on PLC. In vivo, SRF231 in combination with VEN led to complete and durable tumor regression in a xenograft model. SRF231 is currently being evaluated across multiple tumor types in a Phase 1 clinical trial (NCT03512340).