Loss of Parp7 increases type I interferon signalling and reduces pancreatic tumour growth by enhancing immune cell infiltration

Abstract

Background: Pancreatic ductal adenocarcinoma (PDAC) is one of the most lethal forms of cancer, and despite low incidence rates, it remains the sixth leading cause of cancer related deaths worldwide. Immunotherapy, which aims to enhance the immune system's ability to recognize and eliminate cancer cells, has emerged as a promising approach in the battle against PDAC. PARP7, a mono-ADP-ribosyltransferase, is a negative regulator of the type I interferon (IFN-I) pathway and has been reported to reduce anti-tumour immunity.

Methods: We used murine pancreatic cancer cells, CR705, CRISPR/Cas9, in vivo tumour models and spectral flow cytometry to determine the role of PARP7 in pancreatic tumour growth.

Results: Loss of Parp7 elevated the levels of interferon stimulated gene factor 3 (ISGF3) and its downstream target genes, even in the absence of STING. Cancer cells knocked out for Parp7 (CR705Parp7KO) produced smaller tumours than control cells (CR705Cas9) when injected into immunocompetent mice. Transcriptomic analyses revealed that CR705Parp7KO tumours had increased expression of genes involved in immunoregulatory interactions and interferon signalling pathways. Characterization of tumour infiltrating leukocyte (TIL) populations showed that CR705Parp7KO tumours had higher proportions of natural killer cells, CD8+ T cells and a lower proportion of anti-inflammatory macrophages (M2). The overall TIL profile of CR705Parp7KO tumours was suggestive of a less suppressive microenvironment.

Conclusions: Our data show that loss of Parp7 reduces PDAC tumour growth by increasing the infiltration of immune cells and enhancing anti-tumour immunity. These findings provide support to pursue PARP7 as a therapeutic target for cancer treatment.

Keywords: CRISPR/Cas9; Poly-ADP-ribose polymerase 7; pancreatic cancer; tumour infiltrating leukocytes; type I interferon.

Figure 1

PARP7 inhibition differentially affects AHR and IFN-I signalling in CR705, K8484 and BxPC3 cells. Relative PARP7 and AHR protein levels in (A) CR705, (B) K8484 and (C) BxPC3 cells treated with or without RBN-2397 (100 nM) treatment for 24h. Cyp1a1 levels are increased with PARP7i in all cell lines. (D) CR705, (E) K8484 and (F) BxPC3 cells were treated with 10 nM FICZ, 100 nM RBN-2397 and FICZ+RBN-2397 for 4h. RBN-2397 treatment differentially induced Ifnb levels in (G) CR705, (H) K8484 and (I) BxPC3 cells. PARP7i resulted in decreased cell proliferation of (J) CR705 but not (K) K8484 nor (L) BxPC3 cells. Cells were treated with 100 nM of RBN-2397 for 72h. *p<0.05 significance compared with DMSO, ^# p<0.05 significance compared with FICZ. ns, not significant.

Figure 2

CR705^Parp7KO cells do not express STING but have increased levels of ISGF3 and selected ISGs. (A) FICZ-induced Cyp1a1 levels were increased in CR705^Parp7KO compared with CR705^Cas9 cells. Cells were treated with 10 nM FICZ for 4h. (B) Expression levels of Ifnb1 did not significantly differ across the samples tested. CR705^Cas9 and CR705^Parp7KO cells were treated with 100 nM RBN-2397 for 24 h and/or 10 µg/mL DMXAA for 4 h, and expression levels were determined by RT-qPCR. (C) Loss of Parp7 cause a slight, but significant decrease in the proliferation of CR705 cells after 72h. (D) CR705 cells did not express STING. CR705 and EO771 cells were treated with 100 nM RBN-2397 for 24 h and their STING levels detected by western blotting. (E) Inhibition of PARP7 with 100 nM RBN-2397 for 24 h enabled visualization of PARP7 in CR705^Cas9 cells. Protein levels of STAT1, STAT2 and IRF9 were increased in CR705^Parp7KO cells. (F) STAT1 protein levels after treatment with 100 nM RBN-2397 for 24, 48 and 72h. (G-I) Expression of Stat1, Stat2 and Irf9 mRNA levels and (J-L) the ISGF3 target genes Cxcl10, Isg15 and Usp18 after 24 h treatment with 100 nM RBN-2397 in control and Parp7^KO cells. (M) Treatment with 1000 U/mL IFNβ for 1 h induced phosphorylation of STAT1. (N) Expression levels of Cxcl10 were significantly elevated in response to IFNβ and was further increased in CR705^Parp7KO and CR705^Cas9 cells treated with RBN-2397. (O, P) Expression levels of Isg15 and Usp18 were significantly elevated in response to IFNβ, and this was further increased in the CR705^Parp7KO cells. Cells were treated with 100 nM RBN-2397 for 24 h and exposed to 1000 U/mL of IFNβ for 4 h (L-N). *p<0.05 compared with DMSO, ^# p<0.05 significance due to Parp7 deficiency, and ^a p<0.05 significance due to PARP7 inhibition compared with IFNβ treatment alone.

Figure 3

Loss of Parp7 affects expression levels of ARTD family members. (A) Heatmap showing expression levels of ARTD family members in CR705^Cas9 and CR705^Parp7KO cells after RBN-2397 treatment for 24h. The data are presented as log2-fold changes compared to the DMSO-treated control cells. (B-D) Levels of Parp9, Parp10 and Parp14 were significantly increased in the CR705^Parp7KO cells, but unaffected by RBN-2397. (E-G) Levels of Parp2, Tnks2 and Parp13 were significantly lower in CR705^Parp7KO cells. Parp2 and Parp13 were decreased after PARP7 inhibition in the control cells. * denotes statistical significance (p<0.05) from the DMSO treated samples, while ^# denotes significance (p<0.05) due to loss of Parp7.

Figure 4

CR705^Parp7KO cells form smaller tumours when injected into immunocompetent mice. (A) CR705^Parp7KO cells gave rise to smaller tumors when injected into immunodeficient NSG mice. n=6. (B) CR705^Parp7KO cells gave rise to smaller tumours when injected into right flank of female C57BL/6 mice. n=8-9. (C) Representative images of Ki67 staining. (D) Quantification of Ki67⁺ cells, displayed as positive cells per mm². (E) Representative images of CD3 staining. (F) CR705^Parp7KO tumours had increased levels of infiltrating CD3⁺ cells compared with CR705^Cas9 tumours. Quantification of CD3⁺ cells, displayed as positive cells per mm². (G) Representative images of CD8 staining. (H) CR705^Parp7KO tumours had increased levels of infiltrating CD8⁺ cells compared with control. *p<0.05, ** p<0.001, ***p<0.0001. ns, not significant.

Figure 5

Spectral flow cytometry identifies differences in the proportion of tumour infiltrating leukocytes (TILs) between CR705^Cas9 and CR705^Parp7KO tumours. (A) t-stochastic neighbour embedding (t-SNE) plot of TIL cell populations. Density plots of TIL cell populations from (B) CR705^Cas9 and (C) CR705^Parp7KO tumours. Red is higher density whereas blue is lower density. Changes in tumour infiltrating (D) lymphocytes and (E) myeloid lineage cells as a percentage of total leukocytes in CR705^Cas9 and CR705^Parp7KO tumours. (F) Changes in M2 macrophages as a percentage of total macrophages in CR705^Cas9 and CR705^Parp7KO tumours. Cell counts and population frequencies determined using FlowJo V10 software. Differences in population frequencies of different TIL populations between CR705^Cas9 and CR705^Parp7KO tumours were compared using Mann-Whitney test or mixed-effects analysis with Šídák’s multiple comparisons test. *p<0.05, **p<0.01, ***p<0.001.

Figure 6

Spectral flow cytometry identifies differences in the proportion of tumour infiltrating T cell populations between CR705^Cas9 and CR705^Parp7KO tumours. (A) t-stochastic neighbour embedding (t-SNE) plot of T cell populations. Density plots of T cell populations from (B) CR705^Cas9 and (C) CR705^Parp7KO tumours. Red is higher density whereas blue is lower density. (D) Relative abundance of T cell populations in CR705^Parp7KO compared with CR705^Cas9 tumours. (E) Relative proportions of CD4, CD8 and TCRγδ relative to T cells in CR705^Parp7KO tumours. Changes in the percentages of (F) CD4, (G) CD8 and (H) TCRγδ subsets between CR705^Parp7KO compared with CR705^Cas9 tumours. Cell counts and population frequencies determined using FlowJo V10 software. Differences in population frequencies between CR705^Cas9 and CR705^Parp7KO tumours were compared using Mann-Whitney test or mixed-effects analysis with Šídák’s multiple comparisons test. *p<0.05, **p<0.01.

Figure 7

Gene expression profiling of CR705^Parp7KO and CR705^Cas9 tumours reveals an increased number of genes regulating interferon and inflammatory immune signalling. (A) Principal component analysis (PCA) plot of all samples. (B) Heatmap displaying illustrating the relative gene expression profiles of overlapping differentially expressed genes in CR705^Parp7KO and CR705^Cas9 tumours for each replicate. Genes and individual samples are arranged by hierarchical clustering. CR705^Parp7KO and CR705^Cas9 tumours are indicated with coloured bars at the top of the heat map. (C) Volcano plot showing the genes that are significantly upregulated and downregulated genes. Differentially expressed genes were determined using an absolute log2 fold change 1 and an adjusted p value < 0.01. (D) Top 10 pathways identified using the Reactome database that were significantly changed in Parp7^KO compared with Cas9 tumours.

Figure 8

CR705^Parp7KO cells form smaller tumours when injected into Parp7^HA/HA catalytic deficient mice. (A) CR705^Parp7KO cells gave rise to smaller tumours mice compared with CR705^Cas9 cells when injected into female Parp7^HA/HA . n=5-6. (B) At day 21 after injection, CR705^Parp7KO and CR705^Cas9 tumours grew more slowly in Parp7^HA/HA compared with C57BL/6 mice. n=8-11. (C) CD3⁺ T cells and (D) CD4⁺ T cells compared with CR705^Cas9 tumours. (E) CR705^Parp7KO tumours had increased levels of infiltrating CD8⁺ cells compared with control. *p<0.05, ** p<0.001, ***p<0.0001. ns, not significant.

Figure 9

Gene expression profiling of CR705^Parp7KO and CR705^Cas9 tumours in Parp7^HA/HA mice. (A) Volcano plot showing the genes that are significantly upregulated and downregulated genes CR705^Parp7KO vs CR705^Cas9 tumours in Parp7^HA/HA mice. Differentially expressed genes were determined using an absolute log2 fold change 1 and an adjusted p value < 0.01. (B) Top 10 pathways identified using the Reactome database that were significantly changed in Parp7^KO compared with Cas9 tumours. (C) Volcano plot of significantly upregulated and downregulated genes Cas9 tumours in Parp7^HA/HA compared with C57BL/6 mice. (D) Top 10 pathways identified using the Reactome database that were significantly changed in Cas9 tumours in Parp7^HA/HA compared with C57BL/6 mice. (E) Graphical summary of the most significant genes or pathways and the connections among them as determined by Ingenuity pathway analysis of DEGs from Cas9 tumours Parp7^HA/HA compared with C57BL/6 mice. Orange lines specify activation, blue lines specify inhibition, solid lines indicate direct interaction, dashed lines indicate indirect interaction, and dotted lines indicate an inferred relationship.