Pancreatic ductal adenocarcinoma (PDAC) has one of the lowest 5-year survival rates of all cancers, and limited treatment options exist. Immunotherapy is effective in some cancer types, but the immunosuppressive tumor microenvironment (TME) of PDAC is a barrier to effective immunotherapy. CXCR2+ myeloid-derived suppressor cells (MDSCs) are abundant in PDAC tumors in humans and in mouse models. MDSCs suppress effector cell function, making them attractive targets for restoring anti-tumor immunity. In this study, we show that the most abundant soluble factors released from a genetically diverse set of human and mouse PDAC cells are CXCR2 ligands, including CXCL8, CXCL5, and CXCL1. Expression of CXCR2 ligands is at least partially dependent on mutant KRAS and NFκB signaling, which are two of the most commonly dysregulated pathways in PDAC. We show that MDSCs are the most prevalent immune cells in PDAC tumors. MDSCs expressed high levels of CXCR2, and we found that myeloid cells readily migrate toward conditioned media (CM) prepared from PDAC cultures. We designed CXCR2 ligand-Fc fusion proteins to modulate the CXCR2 chemotactic signaling axis. Unexpectedly, these fusion proteins were superior to native chemokines in binding and activation of CXCR2 on myeloid cells. These “superkines” were potent inhibitors of PDAC CM-induced myeloid cell migration and were superior to CXCR2 small-molecule inhibitors and neutralizing antibodies. Our findings suggest that CXCR2 superkines may disrupt myeloid cell recruitment to PDAC tumors, ultimately improving immunotherapy outcomes in patients with PDAC.

Cutaneous T-cell lymphoma (CTCL) is a rare T-cell malignancy characterized by inflamed and painful rash-like skin lesions that may affect large portions of the body’s surface. Patients experience recurrent infections due to a compromised skin barrier and generalized immunodeficiency resulting from a dominant Th2 immune phenotype of CTCL cells. Given the role of the unfolded protein response (UPR) in normal and malignant T-cell development, we investigated the impact of UPR-inducing drugs on the viability, transcriptional networks, and Th2 phenotype of CTCL. We found that CTCL cells were >5-fold more sensitive to the proteasome inhibitor bortezomib (Btz) and exhibited a distinct signaling and transcriptional response compared to normal CD4+ cells. The CTCL response was dominated by the induction of the HSP70 family member HSPA6 (HSP70B’) and, to a lesser extent, HSPA5 (BiP/GRP78). To understand the significance of these two factors, we used a novel isoform selective small-molecule inhibitor of HSPA5/6 (JG-023). JG-023 induced pro-apoptotic UPR signaling and enhanced the cytotoxic effects of proteasome inhibitors and other UPR-inducing drugs in CTCL but not normal T cells. Interestingly, JG-023 also selectively suppressed the production of Th2 cytokines in CTCL and normal CD4+ T cells. Conditioned media (CM) from CTCL were immunosuppressive to normal T cells through an IL-10-dependent mechanism. This immunosuppression could be reversed by JG-023, other HSP70 inhibitors, Btz, and combinations of these UPR-targeted drugs. Our study points to the importance of the UPR in the pathology of CTCL and demonstrates the potential of proteasome and targeted HSPA5/6 inhibitors for therapy.

The treatment of acute myeloid leukemia (AML) remains a challenge due to disease heterogeneity, which undermines efforts to develop targeted therapeutics, leaving conventional chemotherapy as the standard of care (SOC). Sialic acid binding Ig-like lectin 3 (CD33) is a myeloid cell surface glycoprotein that is highly expressed on AML blasts. However, while ~90% of AML cases express CD33, 50% of these patients harbor a single nucleotide polymorphism (SNP) that eliminates the antibody binding epitope for existing CD33-targeted antibody therapeutics. In this study, we developed an immunotherapy (M2T-CD33) that targets CD33 directly to MHC class II (MHCII) molecules on antigen presenting cells for enhanced presentation to the immune system. We found that M2T-CD33 induces a robust polyclonal anti-CD33 humoral response composed of the full immunoglobulin repertoire. M2T-CD33 induced an anti-AML response in a syngeneic mouse model that was dependent on CD4+ and CD8+ T cells. The immune response was elicited against both the full length and spliced version of CD33 and showed no evidence of toxicity at concentrations 40-fold higher than the efficacious dose. Finally, M2T-CD33 was enhanced by combinations with anti-PD-1 therapy. These experiments demonstrate the preclinical potential of M2T-CD33 in AML and emphasize the importance of MHCII for cancer immunotherapy.