Comparative spatial transcriptomics of pancreatic cancer with ductal and acinar origins in mouse models

Abstract

Dear Editor,

Pancreatic ductal adenocarcinoma (PDAC) is a lethal malignancy, and while both acinar and ductal cells can contribute to its origin, their roles in defining PDAC subtypes remain unclear.¹ Investigating the cellular origin of PDAC may provide valuable insights into the biological processes of carcinogenesis and inform novel clinical classification schemes, enabling more precise and effective diagnostic and therapeutic strategies. Acinar cell-derived PDAC, often through acinar-to-ductal metaplasia (ADM), has been well-studied using models such as Pdx1-Cre or Ptf1a-Cre.^2-4 Ductal cells can also give rise to precursor lesions, including intraductal papillary mucinous neoplasms (IPMN), which is an imaging-recognizable lesion that is helpful for the early diagnosis of PDAC.⁵ However, models focusing on ductal cell-derived PDAC remain limited. Transcription factor Sox9, a hallmark marker of pancreatic ductal cells, enables lineage-specific gene editing via Sox9-CreER.^{6, 7} In this study, we induced the carcinogenesis of PDAC by conditionally activating Kras^G12D and deleting Trp53 in ductal or acinar cells. Compared to the well-established KPPC (Kras^LSL-G12D/+;Trp53^fl/fl;Pdx1-CreER) mouse model, the KPPS (Kras^LSL-G12D/+;Trp53^fl/fl;Sox9-CreER) mouse model produces a substantial proportion of IPMN with varying pathological grades. Spatial transcriptomics further revealed partially shared, yet distinct, molecular and tumour microenvironment features between the KPPC and KPPS models, which were validated in human datasets.

We developed the KPPS mouse model and harvested pancreatic tissues between weeks 4 and 24 post-tamoxifen induction (Figure 1A,B and Figure S1A). H&E staining was performed on pancreatic formalin-fixed paraffin-embedded (FFPE) tissue sections from both KPPS and KPPC models (Figure 1C and Figure S1B). Histological analysis of tissues collected from KPPS mice revealed progressive development of IPMN, including low-grade (LG), high-grade (HG), and IPMN-associated invasive carcinoma (IPMN-IC), which exhibited characteristic tubular adenocarcinoma features (Histological criteria are elaborated in the supplementary information; Figure 1C,D). By 24 weeks post-Tamoxifen injection, all KPPS mice had progressed to invasive carcinoma, with IPMN-IC being the predominant phenotype, accounting for more than 80% of the cases, accompanied by conventional PDAC (Figure 1C, D). Immunohistochemical analysis of mucins (MUC) revealed that IPMN lesions from KPPS mice exhibited high expression of MUC1, weak expression of MUC5AC, and minimal to no expression of MUC2 (Figure 1E and Figure S1C), suggesting a non-intestinal IPMN subtype (including gastric and pancreatobiliary types). Notably, the IPMN-IC lesions predominantly exhibited features of the pancreatobiliary subtype. Survival analysis indicated that the median survival of KPPS mice was 20 weeks post-Tamoxifen (n = 46), significantly longer than KPPC mice (13 weeks, n = 25; p < 0.0001; Figure 1F). Whole-exome sequencing (WES) was performed on pancreatic tissues evaluated as normal morphology (KPPS5000 and KPPS5038), IPMN-IC (KPPS6711 and KPPS5294), and conventional PDAC (KPPC1705 and KPPC2604), with spleen tissue used as a control to identify somatic mutations (Figure 1G). Mutations identified in cancer-related signalling pathways, such as PI3K-Akt and MAPK signalling pathways, were more frequent in the KPPC compared to the KPPS model (Figure 1G). A ductal-derived KPPS cell line (KPPS5333) was established, demonstrated stable epithelial phenotype and robust growth with a doubling time of ∼31 h in proliferation assays (Figure S2B,C). To create a subcutaneous tumour model, 5 × 10⁶ KPPS5333 cells were injected into the scapular region of 8-week C57BL/6J mice. Tumour growth was monitored, and by day 14, most tumours exceeded 10 mm in diameter (Figure S2D,E).

Spatial transcriptomics analysis was performed on FFPE tumour samples harvested from KPPS and KPPC mouse models. The tumour epithelium of both models exhibited significant differences in pathway enrichment compared to normal pancreatic tissue (Figure 2A). Differential gene analysis was conducted for the tumour epithelium and stroma in both models (Figure 2B,C). Among these differentially expressed genes, some were specifically associated with pancreatic ductal and acinar cells (Figure 2D,E). In the pancreas, Tm4sf4 and Cdh17 are primarily expressed in ductal cells, and studies have shown that Tm4sf4 co-localizes with Sox9 expression.^{8, 9} Our results indicate that Tm4sf4 and Cdh17 are significantly upregulated in KPPS tumour (Figure 2D), which may serve as biomarkers for identifying ductal cell-derived PDAC. Additionally, markers specific to acinar cells, including Cel, Tff2, Ctrb1, Cela3b, and Reg1, were expressed in the KPPC tumour (Figure 2D,E, and Figure S4D). Notably, compared to KPPS model, the surrounding normal pancreatic tissue in KPPC tumours exhibited upregulation of Krt19 and downregulation of acinar cell markers, suggesting that the pancreatic tissue adjacent to KPPC tumours may be undergoing ADM (Figure 2D). Additionally, Muc5ac was upregulated in the tumour epithelium of KPPC mice but nearly absent in that of KPPS mice (Figure 2D). The regulatory intensity of transcription factors exhibited significant differences between the epithelial components of KPPS and KPPC tumours (Figure S5). Immune profiling using CIBERSORT showed that both models harboured immunosuppressive microenvironments (Figure 2F,G; Figures S6 and S7). Compared to KPPC, the KPPS tumour microenvironment showed a pronounced enrichment of dendritic cells (DCs), particularly resting DCs (both p < 0.0001), accompanied by a significant reduction of plasma cells (both p < 0.0001, Figure 2G).

To further assess clinical relevance, 178 human PDAC datasets from The Cancer Genome Atlas (TCGA) were clustered into ductal-dominant and acinar-dominant groups based on ductal/acinar markers (Figure 3A). Transcriptomic analysis revealed enrichment of pathways related to KRAS, JAK-STAT, and epithelial-mesenchymal transition (EMT) in acinar-dominant PDAC, while ductal-dominant PDAC showed upregulation of MTOR, MYC, glycolysis, and antigen presentation pathways (Figure 3B,C). CIBERSORT and immunophenoscore (IPS) analyses confirmed enrichments of resting DCs and antigen presentation in ductal-dominant tumours (Figure 3D,E).

Our study demonstrates the dual origins of PDAC, with distinct histological, molecular, and immunological profiles, showing similar results with X.⁷ We observed that KPPS models predominantly developed IPMN-IC, displaying a characteristic of the pancreatobiliary subtype, supporting its utility as a robust model for studying PDAC. Gene expression profiling revealed upregulation of ductal markers (Tm4sf4, Cdh17) in KPPS, and acinar markers (Cel, Tff2, etc.) in KPPC. This distinction offers new insights into the molecular subtypes of PDAC, which may contribute to precise diagnosis, prognosis prediction and therapeutic strategy.

Conceptualization: Ming Cui, Jialu Bai, Ya Hu and Yupei Zhao. Methodology: Ming Cui, Jialu Bai, Xiaoyan Chang, Ruiling Xiao and Shengwei Mo. Investigation: Ming Cui and Jialu Bai. Data Analysis and Curation: Ming Cui, Jialu Bai, Ruiling Xiao and Sen Yang. Visualization: Ming Cui and Jialu Bai. Writing-Original Draft: Ming Cui and Jialu Bai. Writing-Review & Editing: Kevin C Soares, Sen Yang, Lei You, Quan Liao, Jin He, Ya Hu and Yupei Zhao. Supervision and funding acquisition: Ming Cui, Jialu Bai, Ya Hu and Yupei Zhao.

The authors declare no conflicts of interest.

This work was supported by funding from the National Natural Science Foundation of China (grant no. 82302076), Fundamental Research Funds for the Central Universities (grant no. 3332024199), the CAMS Innovation Fund for Medical Sciences (CIFMS) (grant no. 2023-I2M-2-002), National High Level Hospital Clinical Research Funding (grant no. 2022-PUMCH-D-001), Beijing Natural Science Foundation (grant no. 7224340), Beijing Science and Technology Innovation Foundation for University or College students (grant no. 2024dcxm057), Milstein Medical Asian American Partnership (MMAAP) foundation and Peking Union Medical College Hospital Talent Cultivation Program Category D UHB12625.

The ethical aspects of the research were reviewed and approved by the Institutional Animal Care and Use Committee, Beijing Vitalstar Biotechnology Co., Ltd. before initiation (Ethical approval number: VST-SY-20210108).