Search

BRCA1-mediated homologous recombination is an important DNA repair mechanism that is the target of FDA-approved PARP inhibitors, yet details of BRCA1-mediated functions remain to be fully elucidated. Similarly, immune checkpoint molecules are targets of FDA-approved cancer immunotherapies, but the biological and mechanistic consequences of their application are incompletely understood. We show here that the immune checkpoint molecule PD-L1 regulates homologous recombination in cancer cells by promoting BRCA1 nuclear foci formation and DNA end resection. Genetic depletion of tumor PD-L1 reduced homologous recombination, increased nonhomologous end joining, and elicited synthetic lethality to PARP inhibitors olaparib and talazoparib in vitro in some, but not all, BRCA1 wild-type tumor cells. In vivo, genetic depletion of tumor PD-L1 rendered olaparib-resistant tumors sensitive to olaparib. In contrast, anti-PD-L1 immune checkpoint blockade neither enhanced olaparib synthetic lethality nor improved its efficacy in vitro or in wild-type mice. Tumor PD-L1 did not alter expression of BRCA1 or its cofactor BARD1 but instead coimmunoprecipitated with BARD1 and increased BRCA1 nuclear accumulation. Tumor PD-L1 depletion enhanced tumor CCL5 expression and TANK-binding kinase 1 activation in vitro, similar to known immune-potentiating effects of PARP inhibitors. Collectively, these data define immune-dependent and immune-independent effects of PARP inhibitor treatment and genetic tumor PD-L1 depletion. Moreover, they implicate a tumor cell–intrinsic, immune checkpoint–independent function of PD-L1 in cancer cell BRCA1-mediated DNA damage repair with translational potential, including as a treatment response biomarker. Significance: PD-L1 upregulates BRCA1-mediated homologous recombination, and PD-L1–deficient tumors exhibit BRCAness by manifesting synthetic lethality in response to PARP inhibitors, revealing an exploitable therapeutic vulnerability and a candidate treatment response biomarker. See related commentary by Hanks, p. 2069

BackgroundTumor intracellular programmed cell death ligand-1 (PDL1) mediates pathologic signals that regulate clinical treatment responses distinctly from surface-expressed PDL1 targeted by αPDL1 immune checkpoint blockade antibodies.MethodsWe performed a drug screen for tumor cell PDL1 depleting drugs that identified Food and Drug Administration (FDA)-approved chlorambucil and also 9-[2-(phosphonomethoxy)ethyl] guanine. We used in vitro and in vivo assays to evaluate treatment and signaling effects of pharmacological tumor PDL1 depletion focused on chlorambucil as FDA approved, alone or plus αPDL1.ResultsPDL1-expressing mouse and human ovarian cancer lines and mouse melanoma were more sensitive to chlorambucil-mediated proliferation inhibition in vitro versus corresponding genetically PDL1-depleted lines. Orthotopic peritoneal PDL1-expressing ID8agg ovarian cancer and subcutaneous B16 melanoma tumors were more chlorambucil-sensitive in vivo versus corresponding genetically PDL1-depleted tumors. Chlorambucil enhanced αPDL1 efficacy in tumors otherwise αPDL1-refractory, and improved antitumor immunity and treatment efficacy in a natural killer cell-dependent manner alone and plus αPDL1. Chlorambucil-mediated PDL1 depletion was relatively tumor-cell selective in vivo, and treatment efficacy was preserved in PDL1KO hosts, demonstrating tumor PDL1-specific treatment effects. Chlorambucil induced PDL1-dependent immunogenic tumor cell death which could help explain immune contributions. Chlorambucil-mediated PDL1 reduction mechanisms were tumor cell-type-specific and involved transcriptional or post-translational mechanisms, including promoting PDL1 ubiquitination through the GSK3β/β-TRCP pathway. Chlorambucil-mediated tumor cell PDL1 depletion also phenocopied genetic PDL1 depletion in reducing tumor cell mTORC1 activation and tumor initiating cell content, and in augmenting autophagy, suggesting additional treatment potential.ConclusionsPharmacological tumor PDL1 depletion with chlorambucil targets tumor-intrinsic PDL1 signaling that mediates treatment resistance, especially in αPDL1-resistant tumors, generates PDL1-dependent tumor immunogenicity and inhibits tumor growth in immune-dependent and independent manners. It could improve treatment efficacy of selected agents in otherwise treatment-refractory, including αPDL1-refractory cancers, and is rapidly clinically translatable.

Tumor PD-L1 mediates diverse cell-intrinsic signals that increase cancer pathogenesis including the DNA damage response (DDR). Although immune consequences of tumor-extrinsic PD-L1 blockade are known, novel therapeutic vulnerabilities from inhibiting tumor-intrinsic PD-L1 signals are little studied. To assess tumor-intrinsic PD-L1 DDR control, we depleted PD-L1 by CRISPR/Cas9 or shRNA (PD-L1KO) in mouse B16 (melanoma), 4T1 (breast), ID8agg (ovarian), and human bladder RT4 cancer lines. We identified specific DDR signaling defects in PD-L1KO versus control (con) cells after exposure to distinct DNA damaging agents. 24h gemcitabine (or 2 Gy X-ray) in vitro in PD-L1KO vs. con RT4 cells elevated γH2AX (DNA damage marker) and reduced Chk2 DDR gene expression but not ATM, ATR, Chk1, or BRCA1 (immunoblots), and decreased BRCA1/p-RPA32 nuclear foci formation (confocal) altogether suggesting tumor-intrinsic PD-L1 promotes specific DDR/DNA repair pathways. PD-L1 loss could thus predict synthetic lethality to DDR inhibitors (DDRi) Chk1i or PARPi. PD-L1KORT4, ID8, and 4T1 cells were more sensitive to small molecule Chk1i in vitro (>15 fold vs. con, MTT) and in vivo in NSG mice. Chk1i and PARPi led to γH2AX increase in PD-L1KO B16, RT4, and 4T1 vs. sham treated tumor cells as predicted by DDR defects. Genetic cellular PD-L1 depletion, but not anti-PD-L1 (extrinsic), strongly sensitized B16 and 4T1 tumors to PARPi in vivo. Strikingly, PARPi had no effect on PD-L1KO B16 in RAG2KO mice despite treating WT mice, indicating a strong immune component to DDRi treatment efficacy. Thus, DDRi plus inhibiting tumor intrinsic PD-L1 signals could improve immunotherapy in immunotherapy-sensitive or resistant tumors.

Aging is the biggest risk factor for cancer, yet little is known about how it affects cancer immunotherapy. Melanoma response to αPD-1/αPD-L1 correlates with CD8+ TCF-1+ T cell stem cell (TCSC) generation. We tested αPD-1 (100 μg/mouse), αPD-L1 (100 μg/mouse) or αPD-L2 (200 μg/mouse) in aged (18–24 months) and young (3–8 months) mice challenged orthotopically with B16 melanoma (SQ) or ID8agg ovarian cancer (IP). Tumors and draining lymph nodes (TDLN) were analyzed by flow. We previously reported that αPD-1 treats young and aged with B16 and αPD-L1 only treats young. αPD-L2 treated B16 in aged but not young, the first single agent cancer immunotherapy exhibiting this property. Preliminary data indicate survival benefit of αPD-L2 against ID8agg in aged but not young mice. Aged mice tolerated all treatments comparable to young. B16 efficacy in young (αPD-1, αPD-L1) and aged (αPD-1, αPD-L2) correlated with increased TCSC and total TIL, but TCSC differed by age and treatment. PD-1 interacts with PD-L1 and PD-L2 while PD-L1 interacts with PD-1 and CD80. Aged mice expressed significantly more T-cell PD-1 and PD-L2 and up to 40-fold more PD-L2, PD-L1 and CD80 vs. young on myeloid cells and NK cells. We generated bone marrow-derived DC with GM-CSF. Aged DC expressed much high PD-L2 versus young, suggesting a cell-intrinsic PD-L2 age bias. Treatment differences in aged versus young could depend on these immune checkpoint or TCSC differences. We are now identifying mechanisms for increased PD-L2 and contributions to αPD-L2 efficacy in aged, and testing TCSC effects on treatments. Our work can improve cancer immunotherapy in aged hosts and provide models for treatment resistance in young hosts.

Tumors express intracellular PD-L1 that we showed increases ovarian cancer (OC) and melanoma virulence by increasing mTORC1 and tumor stem cells (TSC). We made PD-L1KO(CRISPR) in ES2 human OC and B16 melanoma cells and re-expressed PD-L1 only on cell surface (PD-L1surf) or in cytoplasm (PD-L1cyto) with specific expression vectors. Subcellular PD-L1 location in all clones was validated by immunoblots, confocal microscopy, and co-IP/mass spec. We selected clones with PD-L1 surf or cyto expression similar to parental cells. PD-L1KO reduced cell diameter 30% which was restored partially in PD-L1surf and fully in PD-L1cyto in ES2 and B16. PD-L1KO significantly decreased viability in all cell lines tested which was fully rescued in PD-L1surf and PD-L1cyto. PD-L1surf ES2 and B16 grew slower in vitro and in vivo than PD-L1cyto and only PD-L1cyto rescued F-actin polymerization phenotype seen in PD-L1KO ES2 and B16. RNA-seq showed that surf and cyto PD-L1 differentially altered canonical signaling pathways, notably mTOR, stemness and gene regulation. PD-L1cyto grew much slower than PD-L1surf B16 in NSG mice but generated more TSC in vitro, and PD-L1surf but not PD-L1cyto B16 responded to αPD-L1 in WT. PD-L1cyto but not PD-L1surf induced mTORC1 activation (p-rpS6, p-P70S6K) through mTORC1 lysosomal docking through LAMTOR1/2 seen by confocal and confirmed by co-IP in ES2 and B16. Imaging CyTOF of human melanomas showed PD-L1cyto association with activated mTOR. Thus, surf versus cyto PD-L1 differentially affects tumor biology, signaling, pathology, TIL and treatment response in OC and melanoma. We define a mechanism for PD-L1-driven mTORC1 activation. Surface vs. cytoplasmic PD-L1 and/or mTORC1 could be response biomarkers for αPD-L1 or αPD-1.

Anti-tumor T cells (Teffs) engage IL2 largely through CD122, vs. CD25 for regulatory T cells (Tregs). In ID8agg ovarian cancer (OC), CD122-selective IL2 complexes (IL2c, made from CD122-blocking ab+IL-2), increased Teff numbers and functions as expected but unexpectedly reduced tumor Treg function, with near-complete tumor rejection in OC. IL2c reduced lineage specific molecules (e.g., Tbet) and cytokines (e.g., IFNg) in tumor Tregs, suggesting IL2c improves Treg differentiation, but not explaining IL2c-driven Treg defects. In vitro, IL-2 and IL-2c produced similar Treg numbers and Foxp3 from naïve CD4+T cells, with functional data pending. αPD-L1 improved IL2c efficacy for full tumor rejection in OC and B16 melanoma. Increased CD8+CXCR5+PD1+Tim3−TCF-1+stem-like T cells (CXCR5+) are associated with favorable αPD-1 or αPD-L1 outcome in mice and humans. IL2c + αPD-L1 markedly increased CXCR5+in B16 melanoma and to a lesser extent in OC. IL2c + αPD-L1 increased tumor CXCR5+perforin expression in B16 vs. either single agent. Injecting tumor-infiltrating CXCR5+into RagKO mice produced CXCR5+Tim3− and CXCR5−Tim3+cells confirming tumor CXCR5+stem cell properties. These data demonstrate several novel IL2c mechanisms to exploit in distinct cancers and treatment combinations. Ongoing work tests IL2c, αPD-L1 and combo treatment effects on CXCR5+ stemness and mechanisms for CXCR5+ induction, and mechanisms for IL2c-driven Treg dysfunction.