Data from: STAT3 sustains tumorigenicity following mutant KRAS ablation
Data files
Sep 17, 2025 version files 63.16 MB
-
C14PST_vs_A5PST_peak_compare_table.tsv.xls
333.82 KB
-
C14PST_vs_B6PST_peak_compare_table.tsv.xls
302.85 KB
-
fpkm_genename_DKO4_DKO3_KRASKO_DKO2_640_KPCprtl.xlsx
4.41 MB
-
fpkm_genename_KPCprtl_DKO4_KRASKO_STAT3KO.xlsx
3.83 MB
-
gene_fpkm_FOXA1_FOXA2.xlsx
6.06 MB
-
gene_fpkm_PANC1_DKO_STAT3KO_KRASKO.xlsx
6.63 MB
-
README.md
7.48 KB
-
scRNAseq_iKRAS_Ctrl-doxday4.xlsx
10.83 MB
-
scRNAseq_iKRAS-Ctrl_dox.xlsx
11.05 MB
-
scRNAseq_ST3KO__dox.xlsx
9.99 MB
-
scRNAseq_ST3KO_-doxday4.xlsx
9.73 MB
Abstract
Mechanisms underpinning the residual tumorigenicity of mutant KRAS cancers following KRAS inactivation remain to be completely understood. Our studies show that KRAS ablation in pancreatic ductal adenocarcinoma (PDAC) prevents tumor growth contingent on the concomitant inactivation of the transcription factor STAT3. Mechanistically, the incurred losses of KRAS and STAT3 disrupt a temporal balance between tumor cell differentiation, proliferation, and self-renewal. Our findings identify a specific role for STAT3 in supporting cancer cell fitness with a particular focus on KRAS-inhibited tumors and provide a rationale for developing therapies targeting both mutant KRAS and STAT3. RNA sequencing data are presented for PDAC cells following CRISPR-mediated STAT3 and KRAS double knockouts compared to single knockouts, and with double knockouts following enhanced expression of the FOXA1 and FOXA2 genes. Chromatin immunoprecipitation analyses are presented with antibodies to phosphorylated STAT3 in parental and KRAS knockout cells and control antibodies in parental cells. Excel files of sequencing results prepared by Novogene Corp. are provided. In addition, Excel files are included for single-cell RNA-seq analyses of tumors formed by doxycycline-induced mutant KRAS (iKRAS) pancreatic cells and derived STAT3 knockout cells in mice treated with doxycycline and following removal of doxycycline for 4 days. The data was acquired by Cold Spring Harbor Laboratory Genome Center from the Loupe browser software.
https://doi.org/10.5061/dryad.p5hqbzkxs
Description of the data and file structure
Description (1) (2): We evaluated the contribution of STAT3 as a cancer dependency in a mutant KRAS murine pancreatic ductal adenocarcinoma cell line. Murine PDAC (KPC) cells (KRASG12D, TP53R172H) (Hingorani et al. 2005 Cancer Cell) were depleted for mutant KRAS (KRASKO) and STAT3 (STAT3KO). Double knockout (DKO) cells for KRAS and STAT3 were generated from tumorigenic KRASKO cells. For isogenic CRISPR/Cas9-mediated murine knockout cells, we used murine sgKRAS RNA (5′-gtggttggagctgatggcgt-3′) and sgSTAT3 RNA (5’-gcagctggacacacgctacc-3’ or 5’-gtacagcgacagcttcccca-3’) cloned into LentiCRISPRv2 puro (Addgene). Several independent knockout cell clones were isolated and evaluated. Alentiviral-adapted pWPXL/pEF1a vector encoding the hyperactive STAT3Y640F mutation was introduced into the double knockout DKO4.
Total RNA was isolated from cultured cells (KRASKO, STAT3KO, DKO2, DKO3, DKO,4, and DKO4 reconstituted with STAT3Y640F) as described in Methods. RNA was provided to Novogene Corp. for RNA-seq (cDNA library preparation, sequencing, and bioinformatic analyses). The abundance of RNA transcript reflects the gene expression level directly. Gene expression level was estimated by the abundance of transcripts (count of sequencing) that mapped to tthegenome or an exon. Reads count is proportional to gene expression level, gene length,h, and sequencing depth. FPKM (Fragments Per Kilobase of transcript sequence perMillions base pairs sequenced) was used to estimate gene expression levels, which takes the effects into consideration of both sequencing depth and gene length on counting of fragments.
1) fpkm_genename_DKO4_DKO3_KRASKO_DKO2_640_KPCprtl.xlsx
Excel file (1) compares fpkm RNA transcript levels of the KPC parental cells (KPCprtl) with a KRAS knockout derivative (KRASKO), three KRAS/STAT3 double knockout cell derivatives (DKO2, DKO3, DKO4), and DKO4 reconstituted with STAT3Y640F (ST3640). Column comparisons include the murine gene ID, gene symbol, chromosome, start and end of transcript, positive or negative strand, gene length, gene type, and gene description.
2) fpkm_genename_KPCprtl_DKO4_KRASKO_STAT3KO.xlsx
Excel file (2) compares fpkm RNA transcript levels of the KPC parental cells (KPCPrtl), KRASKO derived cells, STAT3KO derived cells, and the double knockout DKO4 derived cells. Column comparisons include the murine gene ID, gene symbol, chromosome, start and end of transcript, positive or negative strand, gene length, gene type, and gene description.
Description (3): We evaluated the contribution of the genes encoding FOXA1 and FOXA2 to the gene expression of KPC cells. Excel file provides fpkm RNA transcript levels of the KRAS/STAT3 double knockout cell line DKO4 reconstituted with FOXA1 (column D) or FOXA2 (column E) genes by retroviral plasmid transduction (obtained from Addgene). Total RNA was provided to Novogene Corp. for RNA-seq (cDNA library preparation, sequencing, and bioinformatic analyses). Gene expression level was estimated by the abundance of transcripts (count of sequencing) that mapped to the genome or an exon. FPKM (Fragments Per Kilobase of transcript sequence perMillions base pairs sequenced) was used to estimate gene expression levels, which takes the effects into consideration of both sequencing depth and gene length on counting of fragments. Columns include the murine gene ID, gene symbol, chromosome, start and end of transcript, positive or negative strand, gene length, gene type, and gene description
3) gene_fpkm_FOXA1_FOXA2.xlsx
Description (4): We evaluated the contribution of STAT3 as a cancer dependency in a mutant KRAS human pancreatic ductal adenocarcinoma cell line, PANC-1 (KRASG12D, TP53R273H). CRISPR-mediated gene editing was used to generate human knockout cells for KRAS (KRASKO) using sgKRAS RNA (5′-gtagttggagctgatggcgt-3′) and for STAT3 using sgSTAT3 RNA (5’-tgtacagcaccggccgatg-3’). Double knockout (DKO) cells for KRAS and STAT3 were generated from tumorigenic KRASKO cells. Total RNA was isolated from cell lines and provided to Novogene Corp. for RNA-seq (cDNA library preparation, sequencing, and bioinformatic analyses). The Excel file includes gene expression level estimated by the abundance of transcripts (count of sequencing) that mapped to the genome or exon. FPKM (Fragments Per Kilobase of transcript sequence perMillions base pairs sequenced) was used to estimate gene expression levels, which takes the effects into consideration of both sequencing depth and gene length on counting of fragments.
4) gene_fpkm_PANC1_DKO_STAT3KO_KRASKO.xlsx
Description (5) (6): We evaluated differences in STAT3 binding to target genes in the murine KPC parental cells, KRAS KO derived cells, STAT3 KO derived cells, or DKO4 cells. Chromatin immunoprecipitation (ChIP) was performed after crosslinking proteins to DNA and following the protocol with Cell Signaling Technology SimpleChIP enzymatic chromatin IP kit. Antibodies to phosphorylated STAT3 (Tyr705)(D3A7) or control rabbit antibodies (2729) were used. Immunoprecipitated DNA was isolated, and libraries were prepared with the NEBNext Ultra II DNA library prep kit for Illumina next-generation sequencing by Novogene Corp. Bioinformatics of genome mapping, peak calling, peak annotation, and differential analysis were performed by Novogene Corp. Narrow peaks with the MACS2 output file, specified as www.genome.ucsc.edu/FAQ/FAQformat.html#format12.
C14PST = ChIP performed with the DKO4 cells and phosphotyrosine STAT3 antibodies
A5PST = ChIP performed with the STAT3KO cells and phosphotyrosine STAT3 antibodies
B6PST = ChIP performed with the KRASKO cells and phosphotyrosine STAT3 antibodies.
Two TSVs were converted to .xls files and should be opened in Excel. Columns specify chromosome, peak start and end, count comparison for each sample, RPM, Fold enrichment, genes, overlap genes, transcription start sites (TSS), etc.
5) C14PST_vs_A5PST_peak_compare_table.tsv.xls
6) C14PST_vs_B6PST_peak_compare_table.tsv.xls
Description (7) (8) (9) (10): We evaluated the influence of STAT3 loss in an inducible mutant KRAS PDAC model. Cells were dispersed from tumors formed in wild-type FVB mice by cells expressing an inducible KRASG12D gene (iKRAS) or iKRAS cells depleted for STAT3 (STAT3 KO) by CRISPR-mediated gene editing. Single-cell RNA-seq was performed on pooled tumors formed in mice treated with doxycycline at day 0 (dox) or following tumor formation and withdrawal from doxycycline for 4 days (doxday4). Single-cell sequencing and analysis were performed by Cold Spring Harbor Laboratory Single Cell Biology Shared Resource (10X Genomics). The Genomics Cell Ranger pipeline was used to convert the file, and secondary analysis was performed with Loupe Cell Browser. Excel files from Loupe Cell Browser data are provided for the multiple Cell Clusters identified by gene expression in each tumor sample. Murine gene ID and expression are presented for each Cell Cluster with an Average, Log2 fold change, and P-value relative to all the cells of the tumor.
7) scRNA-seq_iKRAS_Ctrl_dox
8) scRNA-seq iKRAS_Ctrl-doxday4
9) scRNA-seq_STAT3KO_dox
10) scRNA-seq_STAT3KO_-doxday4
Total RNA was isolated from parental and genetically modified murine pancreatic ductal adenocarcinoma cell lines (KPC) using the PureLink RNA kit (ThermoFisher) and phenol-extracted. Whole exome RNA sequencing with bioinformatics was performed by Novogene Corp, and Excel files are provided.
Data also include chromatin immunoprecipitation results of control KPC cells and cells deleted for mutant KRAS by CRISPR-mediated gene ablation (KRAS KO). These experiments were performed with antibodies to phosphorylated STAT3 and control antibodies. Cell Signaling Technology SimpleChIP Plus enzymatic chromatin IP kit with magnetic beads was used. Briefly, formaldehyde treatment of tissue culture cells was used to crosslink proteins to DNA, chromatin was digested with micrococcal nuclease and sonicated, and crosslinked chromatin was immunoprecipitated overnight with antibodies and reagents from Cell Signaling Technology (control rabbit antibody or phospho-STAT3 antibody (Tyr705). Immunocomplexes were collected with ChIP-grade protein G magnetic beads, stringently washed, reverse cross-linked, and DNA libraries were prepared with NEBNext Ultra II DNA library prep kit for Illumina for next generation sequencing by Novogene Corp. Bioinformatics of genome mapping, peak calling, and peak annotation was performed by Novogene Corp.