Targeting CD47-SIRPa axis shows potent preclinical anti-tumor activity as monotherapy and synergizes with PARP inhibition

Abstract

The objective was to correlate CD47 gene expression with resistance to immune checkpoint inhibitors (ICI) in tumor tissue of gynecological cancer (GC). Further, we sought to assess the efficacy of targeting CD47 pathway alone and in combination in pre-clinical ovarian cancer (OC) models. We performed transcriptomic analyses in GC treated with ICI. Signaling pathway enrichment analysis was performed using Ingenuity Pathway Analysis. Immune cell abundance was estimated. CD47 expression was correlated with other pathways, objective response, and progression-free survival (PFS). Anti-tumor efficacy of anti-CD47 therapy alone and in combination was investigated both in-vitro and in-vivo using cell-line derived xenograft (CDX) and patient-derived xenograft (PDX) models. High CD47 expression associated with lower response to ICI and trended toward lower PFS in GC patients. Higher CD47 associated negatively with PDL1 and CTLA4 expression, as well as cytotoxic T-cells and dendritic cells but positively with TGF-β, BRD4 and CXCR4/CXCL12 expression. Anti-CD47 significantly enhanced macrophage-mediated phagocytosis of OC cells in-vitro and exhibited potent anti-tumor activity in-vivo in OC CDX and PDX models. In-vitro treatment with PARPi increased CD47 expression. Anti-CD47 led to significantly enhanced in-vitro phagocytosis, enhanced STING pathway and synergized in-vivo when combined with PARP inhibitors in BRCA-deficient OC models. This study provides insight on the potential role of CD47 in mediating immunotherapy resistance and its association with higher TGF-β, BRD4 and CXCR4/CXCL12 expression. Anti-CD47 showed potent anti-tumor activity and synergized with PARPi in OC models. These data support clinical development of anti-CD47 therapy with PARPi in OC.

Fig. 1. Different immune cell abundance for cohort of 49 patients with gynecologic cancer who received immunotherapy with immune checkpoint inhibitors and correlation of CD47 with immune checkpoint and other genes involved in immune response and surveillance.

A Heatmap of estimated immune cell abundance, sorted by responders and non-responders. B Pie chart of the prevalence of each immune cell abundance among all samples. C Correlation plot between CD47 and different immune cells as well as other markers including genes involved in immune response and surveillance. The correlation plot was made using “corrplot” R package, with the color of the circles shows the absolute value of corresponding correlation coefficients and area of the circles shows the significance. D, E CD47 expression in correlation with objective response to immunotherapy and progression-free survival while on immunotherapy. Box inside the violin plot visualizes summary statistics including the lower and upper hinges correspond to the first and third quartiles (the 25th and 75th percentiles) and center line as the median. D: CD47 RNA expression among responders to immunotherapy compared to those who did not respond in the entire cohort. CD47 was significantly higher among responders compared to non-responders (p = 0.03) (Top). Kaplan–Meier curves of PFS based on CD47 expression in entire cohort (Bottom). E CD47 RNA expression among responders to immunotherapy compared to those who did not respond in ovarian cancer patients only. CD47 was higher among responders compared to non-responders (p = 0.09) (Top). Kaplan–Meier curves of PFS based on CD47 expression in ovarian cancer patients (Bottom). CD47 group “High” or “low” was defined by median RNA expression of CD47. F The estimated importance of all relevant variables for predicting ICI responses using a Random Forest Model.

Fig. 2. Preclinical in-vitro and in-vivo activity of anti-CD47 therapy in ovarian cancer cells.

A in-vitro phagocytosis of ovarian cancer cells (OVCAR3 and TOV-21G) with macrophages in the presence of 1uM TTI-621 or control Fc. B in-vivo anti-tumor activity of anti-CD47 therapy with Hu-5F9 compared to control and Olaparib in PDX model of high grade serous ovarian cancer (chemotherapy and PARPi resistant model). PDX tumor was implanted subcutaneously, treatment was started when tumor was established. Anti-CD47 was given intraperitoneally twice a week for four weeks. Olaparib was given 50 mg/kg oral gavage 5 times a week till end of the experiment. C NOD/SCID mice were implanted (IP) with TOV-21G cells and treated with TTI-621 (IP 10 mg/kg starting day 7 till day 35). Tumor growth measured by BLI. Kaplan-Meier plots of survival. TTI-621 significantly inhibited tumor growth by both BLI (p = 0.025) and animal survival (p < 0.0001). D TTI-621 treatment schedule for NOD/SCID mice implanted intraperitoneally (IP) with luciferase expressing OVCAR-3 cells. Representative bioluminescence images (BLI) of mice at randomization and day 50. Tumor burden measured by BLI at randomization and day 78. TTI-621 significantly inhibited tumor growth (p = 0.002). No toxicity was observed for mice treated with TTI-621 as measured by body weight change. E NOD/SCID mice were implanted (IP) with SKOV-3 cells and treated with TTI-621. Representative BLI images at randomization (day 6) and day 28. Tumor growth measured by BLI. Kaplan-Meier plots of survival. TTI-621 significantly inhibited tumor growth by both BLI (p = 0.001) and animal survival (p = 0.0003). *≤0.05, **≤0.01, ***≤0.001, ****≤0.0001. Two-Tailed t-Test Assuming Equal Variances. Error bars represent the mean +/− standard error of the mean.

Fig. 3. Expression analyses of different ovarian cancer cells.

Gene expression analysis of CD47 and CCL2 was with RT-QPCR of ovarian cancer cells and CD47 protein expression analysis with flow cytometry were performed. For gene expression analyses, ovarian cancer cells were treated in-vitro with olaparib for 24 h using IC50. For flow cytometry, ovarian cancer cells were treated for 24, 48 and 72 h. A CCL2 expression in ovarian cancer cells treated with olaparib vs. untreated. B CD47 expression in ovarian cancer cells treated with olaparib vs. untreated. C CD47 expression on surface of cancer cells was assessed by flow cytometry after 24 h, 48 h and 72 h of treatment with olaparib compared to control. For analyses, two tailed p-value was calculated assuming equal variance (*≤0.05, **≤0.01, ***≤0.001, ****≤0.0001). Error bars represent the mean +/− standard error of the mean.

Fig. 4. Expression analyses of different ovarian cancer cells.

Gene expression analysis with RT-QPCR of ovarian cancer cells and cytokine analyses using meso-scale assay. A Ovarian cancer cells were treated in-vitro with olaparib, anti-CD47 (Hu-5F9) or both for 24 h using IC50 and RNA was extracted after that. Gene expression of markers of STING pathway (TBK1, IRF3 and IFNB) in PEO4, BRCA2 mutated chemotherapy resistant ovarian cancer cells and in OVCAR8, BRCA2 wild type ovarian cancer cells. B Cytokines levels in supernatant of in-vitro experiment using PEO4 cancer cells after 96 h treatment with olaparib, anti-CD47 (Hu-5F9) or both compared to control. C Ovarian cancer cells were treated in-vitro with olaparib, anti-CD47 (Hu-5F9) or both for 24 h using IC50 and RNA was extracted after that. Gene expression of markers of immunoupressoe myeloid phenotype were assessed including CSF1R and TGFB1R in PEO4 ovarian cancer cells. D Co-culture experiment of UWB1.289.BRCA1 ovarian cancer cells (BRCA1 mutated with reversion mutation) with monocytes using indirect trans-well culture system with 48 h treatment followed by RNA extraction from monocytes. Gene expression of markers of STING pathway (TBK1and IFNB) were assessed. For analyses, two tailed p-value was calculated assuming equal variance (*≤0.05, **≤0.01, ***≤0.001, ****≤0.0001). Error bars represent the mean +/− standard error of the mean.

Fig. 5. Preclinical in-vitro and in-vivo activity of anti-CD47 therapy with PARP inhibition in ovarian cancer cells.

A in-vitro phagocytosis of Veliparib, Niraparib or untreated UWB1.289 BRCA-1WT and BRAC-1Null cells with macrophages in the presence of 1 uM TTI-621 or control Fc. B In-vivo anti-tumor activity of anti-CD47 therapy with Hu-5F9 combined with Olaparib compared to control and Olaparib in CDX model of high grade serous ovarian cancer with BRCA2 mutation (PEO2 ovarian cancer cells). Ovarian cancer cells were implant subcutaneously; treatment was started when tumor was established. Anti-CD47 was given intraperitoneally twice a week for four weeks. Olaparib was given 50 mg/kg oral gavage 5 times a week till end of the experiment. C BRCA1 RNA expression in SKOV3 BRCA1 wild type and BRCA1 KD cell lines. RNA was extracted from cell lines, followed by cDNA synthesis by reverse transcription. Expressions of BRCA1 were quantified using qPCR, and relative expressions of BRCA1 WT and BRCA1-KD cells to the parental SKOV3 Luc+ cells are shown. D NOD/SCID mice implanted (IP) with SKOV-3 (Luc+BRCA1 KD) and treated with TTI-621, niraparib, or TTI-621+niraparib. TTI-621 significantly inhibited tumor growth and prolonged survival (***p < 0.001) compared to vehicle control. TTI-621 + niraparib is superior to either monotherapy (p = 0.0436 compared to TTI-621 alone, and p = 0.0105 compared to niraparib alone). Body weight change of animals with no difference between groups. *≤0.05, **≤0.01, ***≤0.001, ****≤0.0001. Two-Tailed t-Test Assuming Equal Variances. Error bars represent the mean +/− standard error of the mean.