Senescence-induced vascular remodeling creates therapeutic vulnerabilities in pancreas cancer

KRAS mutant pancreatic ductal adenocarcinoma (PDAC) is characterized by a desmoplastic response that promotes hypovascularity, immunosuppression, and resistance to chemo- and immunotherapies. We show that a combination of MEK and CDK4/6 inhibitors that target KRAS-directed oncogenic signaling can suppress PDAC proliferation through induction of retinoblastoma (RB) protein-mediated senescence. In preclinical mouse models of PDAC, this senescence-inducing therapy produces a senescence-associated secretory phenotype (SASP) that includes pro-angiogenic factors that promote tumor vascularization, which in turn enhances drug delivery and efficacy of cytotoxic gemcitabine chemotherapy. In addition, SASP-mediated endothelial cell activation stimulates the accumulation of CD8+ T cells into otherwise immunologically “cold” tumors, sensitizing tumors to PD-1 checkpoint blockade. Therefore, in PDAC models, therapy-induced senescence can establish emergent susceptibilities to otherwise ineffective chemo- and immunotherapies through SASP-dependent effects on the tumor vasculature and immune system. Overall design: For RNA-seq analysis of the transcriptional profiles of tumor cells from KPCmut transplant PDAC tumors treated for 10 days with vehicle or combined trametinib (1 mg/kg body weight), and palbociclib (100 mg/kg), total RNA was extracted from GFP+ tumor cells (sorted on a FACSAria (BD Biosciences)) using the RNeasy Mini Kit (Qiagen). Purified polyA mRNA was subsequently fragmented, and first and second strand cDNA synthesis performed using standard Illumina mRNA TruSeq library preparation protocols. Double stranded cDNA was subsequently processed for TruSeq dual-index Illumina library generation. For sequencing, pooled multiplexed libraries were run on a HiSeq 2500 machine on RAPID mode. Approximately 10 million 76bp single-end reads were retrieved per replicate condition. Resulting RNA-Seq data was analyzed by removing adaptor sequences using Trimmomatic (Bolger et al., 2014), aligning sequencing data to GRCh37.75(hg19) with STAR (Dobin et al., 2013), and genome wide transcript counting using HTSeq (Anders et al., 2015) to generate a RPKM matrix of transcript counts. Genes were identified as differentially expressed using R package DESeq2 with a cutoff of absolute log2FoldChange = 1 and adjusted p-value < 0.05 between experimental conditions (Love et al., 2014). Functional enrichments of these differential expressed genes were performed with enrichment analysis tool Enrichr (Kuleshov et al., 2016) and the retrieved combined score (log(p-value) * z-score) was displayed.