KMT2D depletion promotes KRAS-induced pancreatic carcinogenesis independent of TP53 Authors: Ezequiel J. Tolosa, Glancis L. Raja Arul, Ashley N. Sigafoos, Victoria Ferrero, David R. Pease, Luciana L. Almada, Renzo E. Vera, Nicole M. Pena Ruiz, Tara L. Hogenson, Kayla C. LaRue-Nolan, Brooke R. Tader, Rondell P. Graham, Mathew Brown, Catherine E. Hagen, Aleksey V. Matveyenko, Ryan M. Carr, Kai Ge, Martin E. Fernandez-Zapico Published: 2026 ABSTRACT Despite extensive efforts identifying disease drivers, therapeutics developed against them have so far failed to improve patient outcomes in pancreatic ductal adenocarcinoma (PDAC), the most common histological subtype of pancreatic cancer. Thus, focus has turned to the characterization of molecules contributing to the drivers’ ability to promote the development of this disease. Here, we demonstrate a novel role for KMT2D, a histone methyltransferase, in PDAC biology. Analysis of exome-seq and RNA-Seq data shows that KMT2D is highly mutated and significantly downregulated in human PDAC. Further, low KMT2D expression associates with lower survival in human PDAC. Next, we examined the role of Kmt2d in PDAC development using two independent genetically engineered mouse models driven by mutant Kras in the presence and absence of the tumor suppressor Tp53. Loss of Kmt2d increases tumor incidence mostly with a poorly differentiated phenotype and reduces survival in both models even in the absence of Tp53. Similarly, in human PDAC, the loss of TP53 does not affect the survival of KMT2D low cases. Of note, the sole Kmt2d depletion did not affect normal pancreas development or mice survival. RNA-Seq analysis demonstrates the enrichment of pathways controlling cell growth in Kmt2df/+ PDAC cells. Interestingly, inactivation of Kmt2d resulted in increased levels of Ki67-positive cells in vivo. Together, these findings define the role of this methyltransferase in Kras-driven PDAC biology as tumor suppressor candidate for this dismal condition. Pancreatic ductal adenocarcinoma (PDAC) remains a fatal disease that is projected to become the second cause of cancer-related death in the United States by 2040 (1). PDAC has well-established highly recurrent mutations in four driver genes including the oncogene KRAS and the tumor suppressors TP53, CDKN2A, and SMAD4. Unfortunately, except KRAS, these genetic drivers are not currently therapeutically actionable (2, 3, 4, 5, 6, 7, 8, 9, 10, 11). Further, despite the extensive development of systemic and targeted therapies, the 5-years survival rate for metastatic PDAC, the most frequent form of presentation of this malignancy, has remained almost the same in the last 5 decades (12, 13, 14). Extensive sequencing studies have identified additional drivers in PDAC; however, efforts to translate them into druggable targets have largely failed to improve disease outcomes. In the absence of additional targetable mutations, chemotherapy remains the mainstay of treatment for most patients. Thus, consideration must be given to nongenetic drivers to enhance our understanding of PDAC pathogenesis to inform new treatment strategies (15, 16, 17, 18, 19, 20). Although the role of the above driver mutations on PDAC initiation/early development is well-established, these mutations alone cannot account for PDAC heterogeneity nor discern early from advanced disease. For example, two consensus PDAC subtypes with significantly different prognosis, classical and basal, are defined by differential gene expression (21, 22, 23, 24) as opposed to mutational profile (25, 26, 27). In addition, somatic variants can rarely distinguish between PDAC primary tumors and metastases (2, 28). In fact, it has been shown that PDAC metastasis is driven by enhancer reprogramming rather than secondary somatic drivers (19). Thus, key PDAC pathogenic features are largely conferred by the regulators of the epigenetic landscape rather than the genetic make-up. In our study, we have found a novel role for the histone methyltransferase KMT2D in pancreatic cancer. KMT2D belongs to the 6-member histone-lysine N-methyltransferase 2 (KMT2) family that catalyzes the methylation of H3K4 to promote genome accessibility and subsequent transcriptional activation (29, 30, 31). We have found that KMT2D carries the highest incidence of pathogenic mutations in PDAC cases. Our assessment of biological role of KMT2D demonstrates that the complete or partial loss of Kmt2d expression reduces survival and increases tumor incidence in vivo, highlighting its tumor suppressive activity in PDAC. Further, sequencing analyses reveal a significant enrichment for proliferative pathways in the absence of Kmt2d. Our data are of particular significance in PDAC given the high mutational incidence and lower levels of KMT2D, thereby potentially impacting the development of new therapeutic approaches for this dismal malignancy. Results and discussion KMT2D is highly mutated and downregulated in PDAC To assess the mutation and expression landscape of KMT2s in PDAC, we queried ∼200 patient studies and databases (Tables S1–S3), including The Cancer Genome Atlas (TCGA), International Cancer Genome Consortium, Queensland Center for Medical Genomics, National Cancer Institute's Clinical Proteomic Tumor Analysis Consortium, University of Texas Southwestern databases, and the Cancer Cell Line Encyclopedia (32, 33, 34, 35, 36, 37). Across all cancers, 36.66% of the queried cases harbored alterations in KMT2s, among which KMT2D was the most mutated, at 33.38%, followed by KMT2C (28.76%), KMT2A (13.59%), KMT2B (12.56%), and SETs (SETD1A and SETD1B- 11.71% combined) (Fig. 1A, Table S1). This trend largely holds even for PDAC cases, with KMT2C being highly mutated (40.59%), closely followed by KMT2D (34.65%), and similar breakdowns for mutations in KMT2A (9.40%), SETs (SETD1A, SETD1B combined- 10.89%), and KMT2B (4.46%) (Fig. 1B, Table S1). We then used ClinVar to determine disease significance of the variants identified in the KMT2D-altered samples across all cancers and PDAC cases. Over 300 KMT2D mutations were predicted to be pathogenic in pan-cancer cases (Fig. 1C, Table S1), which was also observed in the case of PDAC, where a majority of KMT2D mutations were annotated as potentially pathogenic (Fig. 1D, Table S1). Further, KMT2D mutations both in pan-cancer and in PDAC cases were largely missense or truncating mutations (Fig. S1, A and B). Mutational analysis of other KMT2 through ClinVar yielded no significant results. We compared KMT2C and KMT2D mutations annotated through ClinVar, as these are more frequent in PDAC. When compared to paralog KMT2C, KMT2D mutations are predominantly pathogenic in nature in a majority of patient samples (Fig. S1B, Table S1). We noted that only 13 pathogenic pan-cancer KMT2C mutations were identified in this dataset. No pathogenic KMT2C mutations were identified in PDAC cases (Table S1).Figure 1Low levels of KMT2D associated with worse survival in human PDAC.A, pie chart representing percentage of PanCancer cases with KMT2 mutations (patient samples and cancer cell lines) and breakdown based on subtype (KMT2A-D, SETD1A, SETD1B). B, pie chart representing percentage of PDAC cases with KMT2 mutations (patient samples and cancer cell lines) and breakdown based on subtype (KMT2A-D, SETD1A, SETD1B). C, graphical representation of predicted disease phenotypes associated with PanCancer KMT2D mutations annotated using ClinVar (n = 34,727). D, graphical representation of predicted disease phenotypes associated with PDAC KMT2D mutations annotated using ClinVar (n = 1274). E, Kaplan-Meier curve showing overall survival from TCGA-PDAC cases grouped based on KMT2D expression reveals a statistically significant difference in survival between KMT2D low (n = 78) and KMT2D high (n = 98) expressing tumors. Statistical significance calculated using the Log-rank test. p-value less than 0.05 was designated statistically. CS: conflicting significance, US: uncertain significance, LP: likely pathogenic, P: pathogenic, LB: likely benign, B: benign. PDAC, pancreatic ductal adenocarcinoma; TCGA, The Cancer Genome Atlas. Low levels of KMT2D associated with worse survival in human PDAC.A, pie chart representing percentage of PanCancer cases with KMT2 mutations (patient samples and cancer cell lines) and breakdown based on subtype (KMT2A-D, SETD1A, SETD1B). B, pie chart representing percentage of PDAC cases with KMT2 mutations (patient samples and cancer cell lines) and breakdown based on subtype (KMT2A-D, SETD1A, SETD1B). C, graphical representation of predicted disease phenotypes associated with PanCancer KMT2D mutations annotated using ClinVar (n = 34,727). D, graphical representation of predicted disease phenotypes associated with PDAC KMT2D mutations annotated using ClinVar (n = 1274). E, Kaplan-Meier curve showing overall survival from TCGA-PDAC cases grouped based on KMT2D expression reveals a statistically significant difference in survival between KMT2D low (n = 78) and KMT2D high (n = 98) expressing tumors. Statistical significance calculated using the Log-rank test. p-value less than 0.05 was designated statistically. CS: conflicting significance, US: uncertain significance, LP: likely pathogenic, P: pathogenic, LB: likely benign, B: benign. PDAC, pancreatic ductal adenocarcinoma; TCGA, The Cancer Genome Atlas. Next, we assessed KMT2 mRNA expression levels in TCGA PDAC patient samples. We observed a notable downregulation of KMT2 mRNA expression, with 62% of the samples showing KMT2 downregulation. In particular, KMT2C was downregulated in 49.71% while KMT2D was downregulated in 54.23% of the samples (Fig. S2A, Table S4). We validated these results in our PDAC patient-derived organoid models using RNA-Seq (38, 39). In line with the PDAC TCGA results, we observed downregulation of all KMT2s across a majority of our PDAC patient-derived organoids. We found that KMT2D was downregulated in 60% of the samples (Fig. S2B, Table S5). Further, categorization as high or low KMT2D expression based on RNA-Seq Z-Score (expression above 0 is high and expression below 0 is low) showed no association with PDAC transcriptomic subtypes (Table S5). Finally, we demonstrated that PDAC cases with low levels of KMT2D were associated with decreased survival compared to cases with high levels of KMT2D expression (Fig. 1E, Table S4). Taken together, this data demonstrates that KMT2D has the highest pathogenic mutational incidence amongst all KMT2s in PDAC cases. Future studies should be aimed at defining the functionality of KMT2D mutations and variants of uncertain significance to determine their biological significance for PDAC pathogenesis. We have also observed significant downregulation of KMT2D in PDAC samples. Characterizing the mechanism underlying the decreased KMT2D expression in PDAC as well as the interplay with genetic mutations of this methyltransferase will have significant implications for patients as nearly 10% of them harbor mutations and ∼50% of the cases have decreased expression in KMT2D. Loss of Kmt2d accelerates Kras-dependent PDAC development Based on the above results, we turned our focus to KMT2D to determine its biological role in PDAC development. We first determined if the deletion of Kmt2d had an impact on pancreas development using genetically engineered mouse models (GEMMs) with pancreas-specific Kmt2d knockout driven by Cre recombinase under the control of the promoter of the pancreas-associated transcription factor 1a (Ptf1a) (40) (Fig. S3A). Cre recombination was confirmed using genomic DNA from the pancreas tissue (Fig. S3B, left panel). Cre mice only showed WT Kmt2d fragment at 1230 bp while recombination was observed at heterozygous and homozygous knockdown of Kmt2d with fragments at 1232 bp/490 bp or just 490 bp, respectively. We further validated the knockdown and measured Kmt2d mRNA expression in pancreas tissue from all three genotypes, seeing a significant reduction in Cre Kmt2df/f cohort (Fig. S3B, right panel). Representative H&E images of pancreas tissue from each of the three genotypes at two different time points show consistent morphology with normal pancreas despite the partial or complete loss of Kmt2d (Fig. 2A). Using immunofluorescence staining for insulin, glucagon, and amylase in tissue sections from Cre and Cre Kmt2df/+ mouse pancreata, we showed no significant alterations in the endocrine or exocrine compartments in these mice (Fig. 2B). Quantification of β cells area in the Cre and Cre Kmt2df/+ mice also showed no significant difference between the two cohorts (Fig. 2C). Finally, we noted no effect on the overall survival of both Cre Kmt2df/+ and Cre Kmt2df/f mice compared to the Cre controls (Fig. 2, D and E). Thus, our findings demonstrated that the knockout of Kmt2d had no significant impact on pancreatic development.Figure 2Kmt2d KO does not alter survival or pancreatic development.A, representative H&E images of Cre, Cre Kmt2df/+, Cre Kmt2df/f at 90days timepoint and Kaplan-Meier (K-M) timepoint, scale bar represents 200 μm. B, representative IF images of mouse pancreatic tissue from Cre and Cre Kmt2df/+ mice, staining for insulin (red), glucagon (white), amylase (green), and DAPI (blue), scale bar represents 20 μm. C, quantification of relative β-cell area in Cre and Cre Kmt2df/+ mice showed no significant differences in the percent of β-cell area between the two cohorts. Statistical analysis, Mann-Whitney test, mean ± SD. D, Kaplan-Meier curve of Cre (n = 10) and Cre Kmt2df/+ mice (n = 9) showed no statistical significance in overall survival between the two cohorts. E, Kaplan-Meier curve representing overall survival in Cre (n = 10) and Cre Kmt2df/f mice (n = 10) showed no statistical significance between the two cohorts. Statistical significance for Kaplan-Meier curves was calculated using the log-rank (Mantel-Cox) test and p-value less than 0.05 was designated statistically significant. IF, immunofluorescence. Kmt2d KO does not alter survival or pancreatic development.A, representative H&E images of Cre, Cre Kmt2df/+, Cre Kmt2df/f at 90days timepoint and Kaplan-Meier (K-M) timepoint, scale bar represents 200 μm. B, representative IF images of mouse pancreatic tissue from Cre and Cre Kmt2df/+ mice, staining for insulin (red), glucagon (white), amylase (green), and DAPI (blue), scale bar represents 20 μm. C, quantification of relative β-cell area in Cre and Cre Kmt2df/+ mice showed no significant differences in the percent of β-cell area between the two cohorts. Statistical analysis, Mann-Whitney test, mean ± SD. D, Kaplan-Meier curve of Cre (n = 10) and Cre Kmt2df/+ mice (n = 9) showed no statistical significance in overall survival between the two cohorts. E, Kaplan-Meier curve representing overall survival in Cre (n = 10) and Cre Kmt2df/f mice (n = 10) showed no statistical significance between the two cohorts. Statistical significance for Kaplan-Meier curves was calculated using the log-rank (Mantel-Cox) test and p-value less than 0.05 was designated statistically significant. IF, immunofluorescence. Next, we generated a KPC mouse model containing Ptf1a Cre-dependent activation of oncogenic Kras and loss of tumor suppressor Tp53; KrasG12D; p48-Cre-Tp53f/+. The KPC mouse is a well-established GEMM to model PDAC progression as it mirrors pertinent clinical and histopathological features commonly observed in human PDAC, including ascites, cachexia, poor vascularity, and fibrosis (41, 42). These mice and Cre Kmt2df/f mice were subsequently crossed to generate KPC Kmt2df/+ and KPC Kmt2df/f mice, where Kmt2d expression was either partially or completely knocked out (Fig. S4A). Similar to the Cre model, recombination and expression of Kmt2d was confirmed in the cross (Fig. S4B). The loss of Kmt2d significantly impacted overall survival in a dose-dependent fashion in the KPC cohort. Survival analysis revealed a significant reduction in overall survival in both the KPC Kmt2df/+ and KPC Kmt2df/f mice compared to the KPC control mice (Fig. 3A). Representative H&E images showing adenocarcinoma observed in all genotypes are shown in Figure 3B. Histological analysis found an increased number of poorly differentiated tumors in the KPC Kmt2df/+ and KPC Kmt2df/f mice compared to KPC animals (Fig. S4F). Of note, tumor incidence was higher in mice with a partial loss of Kmt2d compared to WT mice (100% vs. 78%). Immunohistochemistry (IHC) of phospho-ERK (pERK), a downstream marker of active KRAS, was conducted in all three KPC groups to show the increase in active Kras compared to the Cre control animal (Fig. 3C). Further analysis showed elevation of pERK in the KPC Kmt2df/+ animals compared to the KPC mice. Conversely, KPC Kmt2df/f mice showed lower levels of pERK compared to the KPC mice (Fig. 3, C and D). Cell lines derived from heterozygous or homozygous Kmt2d knockouts exhibit no differences in pERK levels compared to cells derived from KPC mice (Fig. S4, C and D). Kmt2d expression in the mice-derived cell lines was confirmed by qPCR (Fig. S4E).Figure 3Loss of Kmt2d accelerates Kras-dependent PDAC development.A, Kaplan-Meier curve representing overall survival in KPC (n = 10) and KPC Kmt2df/+ mice (n = 11) demonstrate a significant reduction in survival in KPC Kmt2df/+ mice relative to KPC mice (p < 0.05) (top panel). Kaplan-Meier curve of KPC (n = 10) and KPC Kmt2df/f (n = 16) mice highlight a drastic reduction in the survival of KPC Kmt2df/f compared to KPC mice (p < 0.0001) (lower panel). Statistical significance for Kaplan-Meier curves was calculated using the log-rank (Mantel-Cox) test and p-value less than 0.05 was designated statistically significant. B, representative H&E images of adenocarcinoma in pancreas tissue from KPC, KPC Kmt2df/+, and KPC Kmt2df/f mice. Scale bar represents 200 μm. C, representative IHC of pERK in KPC, KPC Kmt2df/+, and KPC Kmt2df/f mice verifying Kras signal activation. Scale bar represents 200 μm (top row). Dashed box indicates area zoomed in and displayed in bottom row. Scale bar represents 100 μm (bottom row). D, quantitative violin plot of percent pERK for each genotype in (C) with red dotted line indicating percent-positive pERK of Cre cohort. Mann-Whitney test was performed to determine statistical significance for quantifications. E, Kaplan-Meier curve showing overall survival from TCGA-PDAC cases with low KMT2D expression based on TP53 mutation status reveals no difference in survival between TP53 WT (N = 28) and TP53 mutant (N = 50) expressing tumors. Statistical significance calculated using the Log-rank test. p-value less than 0.05 was designated statistically significant. IHC, immunohistochemistry; PDAC, pancreatic ductal adenocarcinoma; pERK, phospho-ERK; TCGA, The Cancer Genome Atlas. Loss of Kmt2d accelerates Kras-dependent PDAC development.A, Kaplan-Meier curve representing overall survival in KPC (n = 10) and KPC Kmt2df/+ mice (n = 11) demonstrate a significant reduction in survival in KPC Kmt2df/+ mice relative to KPC mice (p < 0.05) (top panel). Kaplan-Meier curve of KPC (n = 10) and KPC Kmt2df/f (n = 16) mice highlight a drastic reduction in the survival of KPC Kmt2df/f compared to KPC mice (p < 0.0001) (lower panel). Statistical significance for Kaplan-Meier curves was calculated using the log-rank (Mantel-Cox) test and p-value less than 0.05 was designated statistically significant. B, representative H&E images of adenocarcinoma in pancreas tissue from KPC, KPC Kmt2df/+, and KPC Kmt2df/f mice. Scale bar represents 200 μm. C, representative IHC of pERK in KPC, KPC Kmt2df/+, and KPC Kmt2df/f mice verifying Kras signal activation. Scale bar represents 200 μm (top row). Dashed box indicates area zoomed in and displayed in bottom row. Scale bar represents 100 μm (bottom row). D, quantitative violin plot of percent pERK for each genotype in (C) with red dotted line indicating percent-positive pERK of Cre cohort. Mann-Whitney test was performed to determine statistical significance for quantifications. E, Kaplan-Meier curve showing overall survival from TCGA-PDAC cases with low KMT2D expression based on TP53 mutation status reveals no difference in survival between TP53 WT (N = 28) and TP53 mutant (N = 50) expressing tumors. Statistical significance calculated using the Log-rank test. p-value less than 0.05 was designated statistically significant. IHC, immunohistochemistry; PDAC, pancreatic ductal adenocarcinoma; pERK, phospho-ERK; TCGA, The Cancer Genome Atlas. We then set to assess the loss of Kmt2d on a second PDAC GEMM driven by mutant Kras carrying WT Tp53. To do this, we generated a KC mouse (LSL-KrasG12D; Ptf1a-Cre) expressing mutant Kras and developing PDAC (43). This KC mouse was crossed with the Cre Kmt2df/+ mouse to generate KC Kmt2df/+ and KC Kmt2df/f mice (Fig. S5A). Cre recombination and reduced mRNA expression of Kmt2d was confirmed as described in the methods section (Fig. S5B). We noted a dose-dependent effect on overall survival. While we saw no significant difference in survival between the KC and KC Kmt2df/+ cohorts (Fig. 4A), we observed a significant reduction in survival in the KC Kmt2df/f cohort compared to the KC cohort (Fig. 4B). Like the KPC model, tumor incidence was higher in mice with a partial loss of Kmt2d than in Kmt2d WT mice (82% vs. 40%). Complete loss of Kmt2d expression in both the KPC and KC GEMMs significantly reduced overall survival. While the partial loss of Kmt2d did not affect overall survival in the KC GEMM, we observed a significant decrease in overall survival in the KPC GEMM (Fig. 3, A and B). Interestingly, similar to KPC mice, in the KC cohort, we found an increased number of poorly differentiated tumors in the KC Kmt2df/+ and Kmt2df/f mice (Fig. S5C). Representative H&E images of tumors for these crosses are shown in Figure 4C. IHC of pERK was performed in all three KC groups to confirm increase in active KRAS (Fig. 4D), but unlike the KPC cohort, all Kmt2d KO genotypes showed lower levels of pERK (Fig. 4E).Figure 4Depletion of Kmt2d accelerates Kras-dependent PDAC development in the presence of WT Tp53.A, Kaplan-Meier curve of KC (n = 15) and KC Kmt2df/+ (n = 11) showed no statistical significance between the cohorts (p = n.s.). B, Kaplan-Meier curve between KC (n = 11) and KC Kmt2df/f (n = 12) display statically decreased survival of KC Kmt2df/f compared to KC cohort (p < 0.01). Statistical significance for Kaplan-Meier curves was calculated using the log-rank (Mantel-Cox) test. C, representative H&E images of adenocarcinoma in pancreas tissue from KC, KC Kmt2df/+, and KC Kmt2df/f mice. Scale bar represents 200 μm. D, representative IHC images for pERK in Cre, KC, KC Kmt2df/+, and KC Kmt2df/f mice verifying Kras signal activation. Scale bar represents 200 μm (top row). Dashed box indicates area zoomed in and displayed in bottom row. Scale bar represents 100 μm (bottom row). E, quantitative violin plot of percent pERK for each genotype in (D) with reddotted line indicating percent-positive pERK of Cre cohort. Mann-Whitney test was performed to determine statistical significance for quantifications. p-value less than 0.05 was designated statistically significant. IHC, immunohistochemistry; PDAC, pancreatic ductal adenocarcinoma; pERK, phospho-ERK. Depletion of Kmt2d accelerates Kras-dependent PDAC development in the presence of WT Tp53.A, Kaplan-Meier curve of KC (n = 15) and KC Kmt2df/+ (n = 11) showed no statistical significance between the cohorts (p = n.s.). B, Kaplan-Meier curve between KC (n = 11) and KC Kmt2df/f (n = 12) display statically decreased survival of KC Kmt2df/f compared to KC cohort (p < 0.01). Statistical significance for Kaplan-Meier curves was calculated using the log-rank (Mantel-Cox) test. C, representative H&E images of adenocarcinoma in pancreas tissue from KC, KC Kmt2df/+, and KC Kmt2df/f mice. Scale bar represents 200 μm. D, representative IHC images for pERK in Cre, KC, KC Kmt2df/+, and KC Kmt2df/f mice verifying Kras signal activation. Scale bar represents 200 μm (top row). Dashed box indicates area zoomed in and displayed in bottom row. Scale bar represents 100 μm (bottom row). E, quantitative violin plot of percent pERK for each genotype in (D) with reddotted line indicating percent-positive pERK of Cre cohort. Mann-Whitney test was performed to determine statistical significance for quantifications. p-value less than 0.05 was designated statistically significant. IHC, immunohistochemistry; PDAC, pancreatic ductal adenocarcinoma; pERK, phospho-ERK. From our results so far, we observe that even a partial loss of Kmt2d, in conjunction with mutant Kras and Tp53 expression, is sufficient to significantly impact overall survival and induce severe tumor formation in vivo. This tumor suppressive role persists in the KC model, where even in the presence of WT Tp53, Kmt2d loss reduced survival in a dose-dependent manner. This is particularly interesting as we demonstrated that mutations in TP53 do not impact the survival of cases with low KMT2D expression (Fig. 3E, Table S4) while mutations in TP53 resulted in lower survival in cases with high levels of KMT2D (Fig. S2C, Table S4), thus supporting the notion that its tumor suppressor function is independent of TP53. KMT2D regulates genes controlling tumor proliferation To define the underlying mechanisms mediating Kmt2d biological effects, we performed RNA-Seq on cell lines derived from our KPC and KPC Kmt2df/+ GEMMs. Principal component analysis demonstrated distinct clustering for the different genotypes (Fig. S6A). Additionally, we observed differential expression profile between KPC and KPC Kmt2df/+ based on the Z-score analysis (Fig. S6B). Next, we looked at DESeq2 analysis, which found significantly differentially expressed genes in KPC (202 genes) and KPC Kmt2df/+ (151 genes) (Fig. S6C). We then performed gene set enrichment analysis (GSEA) using our RNA-Seq dataset by assessing enrichment of these custom gene signatures in our KPC and KPC Kmt2df/+ cells (Table S6). We observed an enrichment in proliferative signatures including MYC and E2 targets and G2/M transition in KPC Kmt2df/+ cells and inflammatory responses including TNFα and IFNγ signaling as well IL6/STAT3 pathway in the KPC cells (Figs. 5, A and B, S6D, Table S7). We have validated these RNA-Seq results in a separate cohort of tumor samples showing that the depletion of Kmt2d upregulated the expression of proliferative genes including Cdk2, Birc5, Ccnd1, and downregulated antitumoral genes (i.e., Fas1) (Fig. S6E). In agreement with these findings, we showed that tumors from KPC Kmt2df/+ cells have higher levels of Ki67-positive cells compared to the tumors from KPC mice (Fig. 5C), with control mice Cre and Cre Kmt2df/+ exhibiting little to no positive Ki67 cells (Fig. S7A). Interestingly, these proliferative changes were not accompanied by global changes in the inflammatory response as indicated by levels of F4/80 showing comparable percent positive cells for both KPC and KPC Kmt2df/+ (Fig. S7B). Conversely, the KPC group had significantly more apoptosis (TUNEL-positive cells) than Kmt2df/+ and f/f mice (Fig. S8, A–D). Further analysis using Masson trichrome staining showed reduced levels of fibrosis in KPC Kmt2df/+ mice compared to KPC mice (Fig. S7C), and increased levels of angiogenesis, indicated by CD31 expression, were observed on the KPC Kmt2d KO mice compared to the KPC animals (Fig. S9, A–D).Figure 5The inactivation of Kmt2d leads cell growth transcriptomic signature and enhances tumor proliferation.A, gene-set enrichment analysis (GSEA) of mouse hallmark gene sets showing NES and FDR in mouse cell lines with KPC and KPC Kmt2df/+; n = 8. B, GSEA plots of mouse hallmark genes sets significantly enriched in Kmt2d-deficient mouse cell lines including MYC Targets V1, E2F Targets, and G2M Transition. C, representative IHC of proliferative marker Ki67 in Cre, KPC, and KPC Kmt2df/+ with quantitative violin plot of percent Ki76 for each genotype with reddotted line indicating percent-positive Ki67 of Cre cohort. Mann-Whitney test was performed to determine statistical significance for quantifications. p-value less than 0.05 was designated statistically significant. NES = Normalized Enrichment Score; IHC, immunohistochemistry. The inactivation of Kmt2d leads cell growth transcriptomic signature and enhances tumor proliferation.A, gene-set enrichment analysis (GSEA) of mouse hallmark gene sets showing NES and FDR in mouse cell lines with KPC and KPC Kmt2df/+; n = 8. B, GSEA plots of mouse hallmark genes sets significantly enriched in Kmt2d-deficient mouse cell lines including MYC Targets V1, E2F Targets, and G2M Transition. C, representative IHC of proliferative marker Ki67 in Cre, KPC, and KPC Kmt2df/+ with quantitative violin plot of percent Ki76 for each genotype with reddotted line indicating percent-positive Ki67 of Cre cohort. Mann-Whitney test was performed to determine statistical significance for quantifications. p-value less than 0.05 was designated statistically significant. NES = Normalized Enrichment Score; IHC, immunohistochemistry. Our study demonstrates the tumor suppressive activity of Kmt2d through the use of disease-relevant GEMMs driven by KRAS, where even the partial loss of Kmt2d drastically reduced cell survival, induced adenocarcinoma formation, and drove disease progression. Interestingly, this tumor suppressor regulatory mechanism was independent of Tp53 as Kmt2d loss can accelerate tumor progression in the presence of WT Tp53. Through RNA-Seq, we further demonstrate the enrichment of proliferative pathways in KPC Kmt2df/+ cells including gene signatures for the activation of oncogenic transcription factors MYC and E2F. These gene expression changes are associated with increased levels of proliferation in tumors from the KPC Kmt2df/+. Altogether, our findings define a new tumor suppressor mechanism working in parallel to know regulatory pathways controlling cancer growth and thus open new avenues for the development of new treatment and diagnostic approaches for PDAC. Experimental procedures Animals Animals were housed in pathogen-free conditions in facilities approved by the American Association for Accreditation of Laboratory Animal Care in accordance with current regulations and standards of the United States Department of Agriculture, Department of Health and Human Services, and National Institutes of Health/IACUC. Ptf1a-Cre, LSL-KrasG12D, and LSL-KrasG12D; Tp53f/+ animals were generated as previously described (40, 41, 42, 43). Ptf1a-Cre and LSL-KrasG12D mice were crossed to generate Ptf1a-Cre; LSL-KrasG12D (KC) mice. Ptf1a-Cre and LSL-KrasG12D; Tp53f/+ mice were crossed to produce Ptf1a-Cre; LSL-KrasG12D; Tp53f/+ (KPC) mice. Ptf1a-Cre and Kmt2df/+ mice were crossed to produce Pft1a-Cre; Kmt2df/f (Cre Kmt2df/f) mice. KC and Cre Kmt2df/f mice were further crossed to produce the following: (1) Ptf1a-Cre; LSL-KrasG12D; Kmt2df/f (2) Ptf1a-Cre; LSL-KrasG12D; Kmt2df/+ mice. KPC and Cre Kmt2df/f mice were crossed to produce the following: (1) Ptf1a-Cre; LSL-KrasG12D; Tp53f/+; Kmt2df/f; (2) Ptf1a-Cre; LSL-KrasG12D; Tp53f/+; Kmt2df/+ mice. KC and KPC mice were used as controls for subsequent experiments. A set of mice were aged for survival and Kaplan-Meier curves were generated to evaluate the effect of the Kmt2d knockout in PDAC transgenic mouse models. Statistical significance was calculated by Log-rank (Mantel-Cox), with a p-value less than 0.05 designated as significant. Cell lines, cell culture, and transfections Murine cell lines were derived from KPC, KPC Kmt2df/+, and KPC Kmt2df/f GEMMs. Cells were cultured in Dulbecco’s modified Eagle’s medium (10-013-CV, Corning) supplemented with 10% heat-inactivated fetal bovine serum (35-010-CV, Corning) and 1X Penicillin-Streptomycin (15140122, Gibco). Fig. S4E includes expression control showing the reduction of Kmt2d in heterozygous and homozygous knockouts cell lines. We have evaluated the levels of H3K4me1, H3K4me3 and H3K27Ac in KPC, and Kmt2df/+ cell lines. The data included in Fig. S10 demonstrate that the depletion of Kmt2d did not affect the levels of H3K27Ac but shows slight but not significant decrease in H3K4me1 and increase in H3K4me3 (also not significant). Cells were maintained at 37 °C and 5% CO2 and routinely screened for mycoplasma using the ATCC universal mycoplasma detection kit (30-1012K, ATCC). In vivo genotyping Tail DNA was extracted using 180 μl of 50 mM NaOH with proteinase K, incubated initially at 37 °C for 30 min and then at 95 °C for 10 min, followed by the addition of 20 μl of 1M Tris EDTA (pH 7.4). Two microliters of the extracted crude DNA was used for PCR amplification with Terra PCR Direct Red Dy Premix (639286, TaKaRa) and genotype was assessed for these alleles; Ptf1a-Cre, KrasG12D, Tp53f/+ and Kmt2d using the following primers Ptf1a-Cre 5′-agatgttcgcgattatcttc-3′ (sense), 5′-agctacaccagagacgg-3′(antisense); KrasG12D 5′-agctagccaccatggcttgagtaagtctgca-3′ (sense), 5′-cctttacaagcgcacgcagactgtaga-3′ (antisense); Tp53 5′-aaggggtatgagggacaagg-3′ (sense), 5′-gaagacagaaaaggggaggg-3′ (antisense); and Kmt2D 5′-gttcactcagtggggctgtg-3′ (sense), 5′-gcagaagcctgctatgtcca-3′ (antisense). For Kmt2D recombination, the following primers was used: 5′-aggggtaccccaggaaatttg-3′ (sense), 5′- atgagataatggatgtgaaaacgct-3′ (antisense). PCR cycling conditions used were as follows: 95 °C for 5 min, 95 °C for 30 s, 58 °C or 63 °C for 45 s, and 72 °C for 1 min, for 34 cycles. Amplified PCR products were resolved on 1 to 2% agarose gels for genotype validation. Immunofluorescence To assess endocrine and exocrine pancreatic morphology, mice pancreata (Cre: female, 94 days; Cre Kmt2df/+: female, 82 days) were harvested, fixed in 10% formalin, and embedded in paraffin. Sections were then immunostained for insulin in β-cells (Guinea pig polyclonal, 1:100, ab7842, Abcam), glucagon in α-cells (Mouse monoclonal, 1:1000, G2654, Sigma-Aldrich), and α-amylase in acinar cells (Rabbit polyclonal A8273, 1:100, Sigma-Aldrich), with Vectashield-DAPI mounting medium (H-1000, Vector Laboratories). Images were acquired at 5-20× magnification by the Zeiss AXIO Observer Z1 microscope (Carl Zeiss Microscopy, LLC.) and analyzed using ZenPro software (Carl Zeiss Microscopy, LLC). IHC and tissue staining IHC staining for pERK 1/2 was performed at the Pathology Research Core (Mayo Clinic) using the Leica Bond RX stainer (Leica). Formalin-Fixed, Paraffin-Embedded tissues were sectioned at 5 microns and IHC staining was performed on-line. Slides for pERK 1/2 stain were retrieved for 20 min using BOND Epitope Retrieval Solution 1 (AR9961, Leica). The pERK 1/2 primary antibody (4370, Cell Signaling) is a rabbit monoclonal and it was diluted to 1:1200 in Background Reducing Diluent (Dako) and incubated for 15 min. The detection system used was Polymer Refine Detection System (Leica). This system includes the hydrogen peroxidase block, polymer reagent, 3,3′-diaminobenzidine (DAB), and hematoxylin. Immunostaining visualization was achieved by incubating slides for 10 min in DAB and DAB buffer (1:19 mixture) from the Bond Polymer Refine Detection System. To this point, slides were rinsed between steps with 1X bond wash buffer (Leica). Slides were counterstained for 5 min using Schmidt hematoxylin and molecular biology grade water (1:1 mixture), followed by several rinses in 1X Bond wash buffer and distilled water, this is not the hematoxylin provided with the Refine kit. Once the immunochemistry process was completed, slides were removed from the stainer and rinsed in tap water for 5 min. Slides were dehydrated in increasing concentrations of ethyl alcohol and cleared in three changes of xylene prior to permanent coverslipping in xylene-based medium. IHC for Ki67 and F4/80 were done using IHC Application Solutions Kit (13079, Cell Signaling) and manufacturers protocol was followed. Briefly, slides were rehydrated with xylene and decreasing concentrations of ethyl alcohol. Antigen retrieval was performed using sodium citrate for 20 min followed by a hydrogen peroxidase block for 10 min and animal free blocking solution for 1 h. Slides were then incubated overnight at 4 °C with primary antibody Ki67 at 1:500 (12202S, Cell Signaling) or F4/80 at 1:200 (70076S, Cell Signaling). Following day slides were rinsed, incubated with DAB for 5 to 10 min, and counterstained with hematoxylin blued with 1% acetic acid solution for 30 s and lastly dehydrated again with increasing concentration of ethyl alcohol and cleared with xylene prior to permanent mounting with coverslip. IHC for CD31 was also done using IHC Application Solutions Kit. Antigen retrieval on the rehydrated slides was performed using TE buffer (10 mM Tris/1 mM EDTA, pH 9.0) for 30 min. Slides were incubated overnight at 4 °C with primary antibody CD31 (PECAM-1, 77699S, Cell Signaling) at 1:200 dilution. Following day slides were rinsed, incubated with DAB for 10 min, and counterstained with hematoxylin blued with 1% acetic acid solution for 30 s and lastly dehydrated again with increasing concentration of ethyl alcohol and cleared with xylene prior to permanent mounting with coverslip. H&E staining was performed at the Pathology Research Core (Mayo Clinic). Pancreas/Spleen tissues were collected from mice and fixed in 10% neutral-buffered formalin. Tissue slides were baked and run through xylene to remove paraffin. Tissue was then rehydrated and stained with hematoxylin and blued with 1% acetic acid solution for 30 s. After this, tissue was stained with eosin, dehydrated, and cleared. Trichrome Masson staining was performed at the Pathology Research Core (Mayo Clinic). Unstained slides were rehydrated and placed in Bouin’s solution at 58 °C for 15 min and rinsed until solution ran clear. Slides were stained with hematoxylin for 5 min followed by Biebrich scarlet-acid fuchsin solution for 10 min. After rinsing, slides were placed in phosphotungstic-phosphomolybdic acid solution for 10 min and blued using aniline blue solution for 5 min, dehydrated, and mounted with coverslip. Tissue quantification IHC and Masson Trichrome staining were analyzed using QuPath v.0.5.1 software for positive cell detection and pixel classification, respectively. For each staining, eight to ten images were analyzed per mouse with each group containing three to five mice per genotype. TUNEL assay Apoptotic cells in tissue sections corresponding to the KC and KPC cohorts were detected using the TUNEL assay kit (25879, Cell Signaling) following manufacturer’s instructions for paraffin-embedded samples. Briefly, sections were deparaffinized in xylene and rehydrated through a graded ethanol series. Antigen retrieval was performed by incubating the sections in 20 μg/ml proteinase K for 30 min. After washes with PBS, sections were first incubated with TUNEL equilibration buffer and then with TdT enzyme and TUNEL reaction buffer for 60 min protected from light. Sections were counterstained with DAPI to visualize nuclei. Images were acquired using the Echo Revolution microscope with 10× magnification. Positive cells were quantified in eight random fields per section in QuPath version 5.1 using positive cell detection. Negative controls were prepared by omitting the TdT enzyme from the reaction mix. Western blotting Whole cell extracts were prepared using RIPA lysis buffer supplemented with Complete inhibitor cocktail (77809900, Roche Applied Science,) and phosphatase inhibitors (PI78420, Thermo Fisher Scientific). Cell lysate was sonicated on ice at 10% frequency for 3 s for four cycles with 10 s in between each cycle. Additionally, the samples were passed through a 27 ½-gauge needle with a 1 ml syringe. After the lysates were cleared at 15,000×g for 10 min, supernatants were collected. Protein concentration was determined using a Pierce BCA Protein Assay Kit (23227, Thermo Fisher Scientific) and equal amounts of protein (20–50 μg/lane) were separated by electrophoresis using a 4 to 20% polyacrylamide protein gradient gel (4561094, Bio-Rad). After transferring to polyvinylidene fluoride membrane, blots were probed overnight with the following primary antibodies: ERK 1/2 (9107S, Cell signaling), p-Erk/P-p44/42 MAPK (4370S, Cell signaling), H3K4me3 (ab8580, abcam), H3 (3638S, Cell Signaling), H3K27ac (ab4729, abcam), H3K4me1(5326S, Cell signaling). Blots were then incubated with anti-rabbit (ALI3403, Biosource, 1∶2000 dilution) or anti-mouse (A3673, Sigma-Aldrich, 1∶2000 dilution) secondary antibodies conjugated with horseradish peroxidase, and signals were visualized by chemiluminescence (34580, SuperSignal West Pico PLUS Chemiluminescent Substrate, Thermo Fisher Scientific) and the ChemiDoc Imagin System (Bio-Rad). RNA-sequencing For RNA-Seq analysis in KPC and KPC Kmt2df/+ cell lines, RNA was extracted from 1 × 106 cells/group in triplicate with TRIzol reagent and then isolated using RNeasy kit per manufacturer instructions (74004, Qiagen). Library preparation and sequencing was outsourced to the Genome Analysis Core from the Mayo Clinic. RNA-Seq FastQ files were processed and trimmed using Trim Galore 0.6.4 (44). FastQC was used to assess read quality (http://www.bioinformatics.babraham.ac.uk/projects/fastqc). Reads were then aligned with STAR 2.7.3a, using mm10 genome as reference through the Mayo Biocluster (45). The read counts were generated with HTSeq 0.9.1 (46). BAM files for biological replicates were merged using SAMtools (47, 48). DESeq2 was used for differential gene expression analysis using default parameters (Galaxy Version 2.11.40.6) (49). Upregulation or downregulation was determined by Down < 0 > Up. Significance was determined by < 0.05 adjusted p-value. GSEA was completed using GSEA 4.2.3. Visualization was done using ComplexHeatmap in R Studio (2023.12.1). Quantitative real-time PCR Analysis of mRNA expression in cell lines was performed using TRIzol reagent (15596018, Invitrogen) per manufacturer protocol. Two micrograms of total RNA was reverse-transcribed using the High-Capacity cDNA Synthesis Kit (4368814, Applied Biosystems). A portion of the total complementary DNA was amplified by real-time quantitative PCR, using 1X IQ SYBR Green Supermix (1708880, Bio-Rad) and expression for Kmt2d was assessed using the following primers: Kmt2d 5′-cctctgatggagttaccgct-3′ (sense), 5′-ggtatggggccgtttatagtgt-3′(antisense). As housekeeping, the following primers were used: Prt 5′-aagtgtttattcctcatggactga-3′ (sense), 5′-ctcccatctccttcatgacatc-3′ (antisense) and Tbp 5′-gaagttccctataaggctggaag-3′ (sense), 5′-aggagaacaattctgggtttga-3′ (antisense). Amplification was performed using the C1000 Thermal Cycler under the following reaction conditions: 95 °C for 3 min followed by 40 cycles of 30 s, 30 s at 60 °C, and 20 s at 72 °C. Each mRNA level was normalized by comparison with Prt and Tbp. The results were calculated following the ΔΔCT method. Statistical analysis Results are presented as a mean value from a minimum of three experimental replicates ± standard deviation. All experiments were performed in triplicate, unless otherwise specified. Graphs were generated and statistical analysis performed using GraphPad Prism version 10.5.0. ANOVA was used to assess differences between more than two group means followed by multiple group comparison. For nonparametric analysis, Kruskal Wallis was used followed by multiple group comparison. For two group testing, student’s t test was applied to evaluate significance across biological replicates. Mann-Whitney test was used in the case of nonparametric data. Log-rank (Mantel-Cox) test was used to assess significance in overall survival. Data availability The RNA-Seq data generated in this publication is available on the Gene Expression Omnibus database: GSE328371. RNA-Seq data from PDOs was accessed from GSE185335 (39). Human data for Kaplan-Meier curves was accessed from TCGA via dbGaP (Accession Number: phs000178). (50) Patient data to assess the mutation and expression landscape of KMTs in PDAC was accessed from cBioPortal and the Cancer Cell Line Encyclopedia (32, 33, 34, 35, 36, 37). Any additional information required to reanalyze the data reported in this work is available upon request. Supporting information This article contains supporting information. Conflicts of interest The authors declare that they have no conflicts of interests with the contents of this article.