03.05 DSP216 blocks HLA-G and CD47 signaling toward immune cells and restores anti-cancer activity

Background

CD47 checkpoint inhibition has yielded promising results in clinical trials, particularly when combined with other immunotherapeutics. Therefore, we designed Dual Signaling Protein 216 (DSP216), an Fc-fusion protein combining CD47 with HLA-G checkpoint inhibition. DSP216 is composed of sequences selected from the SIRPα and LILRB2 extracellular domains fused to human IgG1. The SIRPα domain of DSP216 binds to and blocks CD47, a don’t eat me signal, frequently overexpressed on cancer cells. The LILRB2 domain binds to and blocks HLA-G, a pivotal regulator of immune tolerance during pregnancy, that is upregulated in many cancers. HLA-G binds to inhibitory receptors LILRB1 (ILT2) and LILRB2 (ILT4) expressed on immune cells and thereby, suppresses anti-cancer immunity in a multifold manner. Blocking HLA-G binding to its receptors reverses this immune suppression. To avoid off tumor toxicity, DSP216 was designed to bind cells in an ‘AND-gate’ fashion, requiring the expression of both targets for effective binding and minimizing binding to cells that express only one of the targets. Here, we report on the preclinical in vitro characterization of DSP216.

Materials and Methods

Binding of DSP216 to HLA-G⁺ CD47⁺ and HLA-G^– CD47⁺ cancer cells, PBMCs and red blood cells (RBCs) was tested by flow cytometry. To test binding specificity of DSP216, HLA-G⁺ CD47⁺ cancer cells were incubated with HLA-G and/or CD47 blocking antibodies prior to DSP216 binding. The effect of DSP216 on the polarization of M0 macrophages co-cultured with HLA-G⁺ CD47⁺ cancer cells was tested by monitoring the expression of CD163 by flow cytometry. Macrophage phagocytosis was tested by incubation of CTV-stained cancer cells with macrophages, followed by CD11b staining and identification of double positive cells by flow cytometry. NK cytotoxicity was tested by co-culture studies of NK cells with cancer cells at different ratios and subsequent staining of cancer cells with Annexin V and read out by flow cytometry.

Results

DSP216 bound to HLA-G⁺ CD47⁺ cancer cells, but not to HLA-G^– CD47⁺. When HLA-G⁺ CD47⁺ cancer cells were pre-incubated with HLA-G and CD47 blocking antibodies, binding was completely abrogated. When cancer cells were mixed with RBCs and PBMCs, DSP216 did not bind to single positive CD47⁺ RBCs and only very weakly to CD47⁺ PBMCs. In co-culture of M0 macrophages with HLA-G⁺ CD47⁺ cancer cells, DSP216 prevented HLA-G induced upregulation of the M2 specific marker CD163. DSP216 with silenced and active Fc domain significantly increased phagocytosis of CD47⁺ HLA-G ⁺ cells, but not of HLA-G^– CD47⁺ cells. Similarly, DSP216 reversed HLA-G dependent inhibition of NK cytotoxicity. DSP216 with active Fc domain combined checkpoint blockade with ADCP/ADCC and utilized both mechanisms to eliminate HLA-G⁺ CD47⁺cancer cells by phagocytosis and NK cytotoxicity.

Conclusions

DSP216 binds with high affinity to HLA-G⁺ CD47⁺ cells and does not or marginally bind to single positive cells, like RBCs and PBMCs. DSP216 prevents HLA-G mediated upregulation of the M2 marker CD163 on macrophages and eliminates HLA-G⁺ CD47⁺ cancer cells by promoting macrophage-mediated phagocytosis and NK cell-mediated cytotoxicity. DSP216 is a novel bifunctional therapeutic that has the potential to reverse immunosuppressive signaling and unleash innate anti-cancer immune responses.

L.J. Jacob: C. Other Research Support (supplies, equipment, receipt of drugs or other in-kind support); Significant; KAHR medical.