Pooled CRISPR screening in pancreatic cancer cells implicates co-repressor complexes as a cause of multiple drug resistance via regulation of epithelial-to-mesenchymal transition

Abstract

Background: Pancreatic ductal adenocarcinoma (PDAC) patients suffer poor outcomes, including a five-year survival of below 10%. Poor outcomes result in part from therapeutic resistance that limits the impact of cytotoxic first-line therapy. Novel therapeutic approaches are needed, but currently no targeted therapies exist to treat PDAC.

Methods: To assess cellular resistance mechanisms common to four cytotoxic chemotherapies (gemcitabine, 5-fluorouracil, irinotecan, and oxaliplatin) used to treat PDAC patients, we performed four genome-wide CRISPR activation (CRISPR_act) and CRISPR knock-out (CRISPR_ko) screens in two common PDAC cell lines (Panc-1 and BxPC3). We used pathway analysis to identify gene sets enriched among our hits and conducted RNA-sequencing and chromatin immunoprecipitation-sequencing (ChIP-seq) to characterize top hits from our screen. We used scratch assays to assess changes in cellular migration with HDAC1 overexpression.

Results: Our data revealed activation of ABCG2, a well-described efflux pump, as the most consistent mediator of resistance in each of our screens. CRISPR-mediated activation of genes involved in transcriptional co-repressor complexes also conferred resistance to multiple drugs. Expression of many of these genes, including HDAC1, is associated with reduced survival in PDAC patients. Up-regulation of HDAC1 in vitro increased promoter occupancy and expression of several genes involved in the epithelial-to-mesenchymal transition (EMT). These cells also displayed phenotypic changes in cellular migration consistent with activation of the EMT pathway. The expression changes resulting from HDAC1 activation were also observed with activation of several other co-repressor complex members. Finally, we developed a publicly available analysis tool, PancDS, which integrates gene expression profiles with our screen results to predict drug sensitivity in resected PDAC tumors and cell lines.

Conclusion: Our results provide a comprehensive resource for identifying cellular mechanisms of drug resistance in PDAC, mechanistically implicate HDAC1, and co-repressor complex members broadly, in multi-drug resistance, and provide an analytical tool for predicting treatment response in PDAC tumors and cell lines.

Keywords: ABCG2; CRISPR; Drug resistance; Genome-wide screen; HDAC1; Pancreatic cancer; Precision oncology.

Fig. 1

CRISPR screen reveals drug resistance genes. A Schematic describing our screening protocol. B A scatterplot shows the mean L2FC sum for all four drugs assayed in each of two cell lines compared to the log₁₀ p-value for the association of the same gene’s expression with with patient survival. ABCG2 stands out as the most highest L2FC over all four drugs. C Boxplots indicate the sgRNA fold change in counts per million comparing treated cells to control cells for each replicate and each cell line for the top ABCG2 sgRNA. Circles represent data from Panc-1 cells. Squares represent BXPC-3

Fig. 2

A Scatterplot shows ABCG2 stands out as the single gene showing a highly significant change in expression with use of the ABCG2_A and ABCG2_B sgRNAs to activate ABCG2 expression. These data are generated from RNA-sequencing of MiaPaca-2 and Panc-1 cells over-expressing ABCG2. B Cell survival curves display the fraction of cells surviving after a series of irinotecan doses for Panc-1 cells stably expressing an ABCG2-targeting sgRNA (blue) or non-targeting control sgRNAs (red and black). C Each bar represents a ratio of ABCG2 over-expressing cells compared to a non-targeting control with the indicated drug treatments (+ indicates a treatment was added). Treatment of MiaPaca2 cells over-expressing ABCG2 using the ABCG2 guide shows no significant cell death with inhibitor only, where as ABCG2 alone confers drug resistance. The combination of 3uM sorafenib (green) or KO143 (maroon) with ABCG2 overexpression restores normal sensitivity to Irinotecan. D Boxplots showing the normalized expression levels of ABCG2 in Kirby et al. for patients surviving < 300 days or > 900 days. P-value is not significant

Fig. 3

A Using weighted averages derived from the screen data we predicted likely sensitivity to gemcitabine based on expression of resistance-associated genes (PKG). There is a significant difference in survival between patients with predicted high versus low gemcitabine sensitivity (p = 0.01, Chi-squared test). B Predicted Gemcitabine Sensitivity was calculated using a weighted average of gene expression for resistance-associated genes based on expression profiles for 18 pancreatic cancer cell lines. These data were plotted compared to experimentally measured IC₅₀ groups (Low: IC₅₀ < 20uM, Moderate: 100uM < IC₅₀ < 300; High: IC50 > 300uM. The Wilcoxon P-value between the most resistant group of cell lines and the most sensitive is 0.097. C Scatterplot showing observed irinotecan IC₅₀ values compared to predicted sensitivity for 18 PDAC cells assayed by the Cancer Cell Line Encyclopedia. (Rho = 0.44, p = 0.06)

Fig. 4

A Scatter plot showing pathways enriched for multi-drug resistance in our CRISPR_act and CRISPR_ko screens and their association with patient survival in the TCGA cohort. Patient survival is represented by the size and color of the circle. B A volcano plot shows that activation of HDAC1 expression using a dCas9-activation approach results in strong overexpression of HDAC1 based on RNA-sequencing. C Pathway enrichment analysis shows that HDAC1 overexpression especially affects epithelial-to-mesenchymal transition (EMT), cell efflux, apoptosis, autophagy, and DNA repair. P-values reported are derived from Fisher’s exact test comparing observed versus expected number of genes in each pathway. A full list is available in Supplemental Table S5. D Overexpression of each of the target genes listed on top right quadrant leads to a similar pattern of overexpression of genes shown on the lower right

Fig. 5

A ChIP-seq analysis reveals ChIP-seq peaks in the control (CTL) MiaPaCa-2 cells overlap significantly with MiaPaCa-2 cells over-expressing HDAC1 (red). An additional 17,501 peaks are identified with overexpression of HDAC1. B Cumulative distribution plot showing that HDAC1 binding sites identified upon overexpression of HDAC1 (blue) are nearby transcription start sites (TSS) of differentially expressed genes. C Scratch assay shows that overexpression of HDAC1 leads to increased migration compared to control cells. D Quantification of the scratch assays shows a significant difference with HDAC1 overexpression (Student’s T-test)