Background
HLA-G is highly expressed in various types of cancers. It binds to the inhibitory receptor LILRB1 (leukocyte Ig-like receptor B1) or LILRB2 (leukocyte Ig-like receptor B2) on the cell surface of immune cells including NK, T, and dendritic cells, and inhibits their anticancer function.
LILRB1, an HLA-G receptor, is expressed in both resting and activated T cells, while programmed cell death protein 1 (PD1) is mainly expressed in activated T cells. It is thought that these play different immune cell suppression roles in regulating antitumor response. HLA-G is thus a potentially new immune checkpoint and a novel immunotherapy target.
Monoclonal antibodies that target HLA-G have so far elicited limited anticancer efficacy. We developed AR092, an anti-HLA-G/CD3 bispecific antibody (BsAb), that was shown to synergistically increase T cell activation through the unique functioning of HLA-G/LILRB1 blockade and a HLA-G-specific T cell redirection effect, thereby maximizing cancer cell killing efficacy. In addition, we sought to overcome the safety hurdle of CD3 BsAbs by strategically lowering the affinity to CD3 and minimize the immune-mediated toxicities including cytokine release syndrome (CRS).
Methods
AR092 was manufactured and produced based on our proprietary Antibody-Like Cell Engager (ALiCE) format. We analyzed the HLA subtype binding, binding affinity for HLA-G and CD3, and thermal stability characteristics of AR092, and evaluated it using various in vitro assays including T cell activation, tumor cell lysis, and cytokine release analysis. We confirmed its anticancer activity in vivo using cell-derived xenograft (CDX) and patient-derived xenograft (PDX) mouse models.
Results
AR092 specifically binds to HLA-G without any nonspecific binding to other HLA subtypes. The binding affinity of the anti-HLA-G/CD3 BsAb to CD3 was 100 times lower compared to the parental anti-CD3 antibody. T cell activation and tumor cell lysis was dose dependent when assessed using different HLA-G positive tumor cell lines. In addition, proinflammatory cytokines such as TNF-α, IFN-, IL-2 and IL-6 were induced at lower levels compared to benchmark. Finally, AR092 exhibited anti-tumor activity in a 4T1/hB2M-hHLA-G breast tumor syngeneic mouse model, in a A375/hB2M-hHLA-G melanoma CD34+ huHSC-NCG CDX mouse model, and in a 148T ColvM colon tumor PDX mouse model.
Conclusions
AR092 demonstrated significant antitumor effects in vitro and in vivo and may overcome the limitations of existing immunotherapy treatments by increasing the therapeutic potential and mitigating toxicities.