Targeting the NPY/NPY1R signaling axis in mutant p53–dependent pancreatic cancer impairs metastasis

Abstract

Pancreatic cancer (PC) is a highly metastatic malignancy. More than 80% of patients with PC present with advanced-stage disease, preventing potentially curative surgery. The neuropeptide Y (NPY) system, best known for its role in controlling energy homeostasis, has also been shown to promote tumorigenesis in a range of cancer types, but its role in PC has yet to be explored. We show that expression of NPY and NPY1R are up-regulated in mouse PC models and human patients with PC. Moreover, using the genetically engineered, autochthonous KP^R172HC mouse model of PC, we demonstrate that pancreas-specific and whole-body knockout of Npy1r significantly decreases metastasis to the liver. We identify that treatment with the NPY1R antagonist BIBO3304 significantly reduces KP^R172HC migratory capacity on cell-derived matrices. Pharmacological NPY1R inhibition in an intrasplenic model of PC metastasis recapitulated the results of our genetic studies, with BIBO3304 significantly decreasing liver metastasis. Together, our results reveal that NPY/NPY1R signaling is a previously unidentified antimetastatic target in PC.

Targeting NPY1R reduces mutant p53–dependent pancreatic cancer metastasis.

INTRODUCTION

Pancreatic cancer (PC) remains one of the most lethal forms of human cancer worldwide, with an estimated 5-year survival rate of only 13% (1). PC exhibits widespread invasion in combination with early-stage metastatic events, which leads to >45% of patients with PC presenting with distant metastatic disease at diagnosis and precludes surgical resection for most patients (~80%) (1, 2). The 5-year survival rate of PC declines markedly from 44% for localized disease to 3% for distant disease (1), which underscores the critical need to understand which signaling pathways regulate PC metastatic colonization and to identify new ways to target these mechanisms to improve outcomes for patients with metastatic disease (3, 4). While advances have been made in the treatment of PC in recent years with the addition of nab-paclitaxel (Abraxane) to gemcitabine chemotherapy (8.5-month median survival for gemcitabine/Abraxane versus 6.8 months for gemcitabine alone) and FOLFIRINOX chemotherapy (11.1-month median survival), almost all patients eventually relapse with metastatic disease and, therefore, better therapeutic options are urgently required in the clinical management of this aggressive cancer (5, 6).

Pancreatic ductal adenocarcinoma (PDAC) is the most common form of PC, representing >90% of human patients. Most PDAC cases present with an activating mutation in KRAS (>90% of cases), resulting in the formation of precursor lesions known as pancreatic intraepithelial neoplasms (7–9). Progression to advanced stages of disease is typically accompanied by alterations in tumor suppressor genes, the most common being p53 (TRP53; 50 to 75% of patients with PC) (10, 11). Metastasis is the result of a complex interplay between both cancer cell–autonomous functions and feedback from the tumor microenvironment (12–15). Cancer cell–autonomous metastatic processes can be driven by the genetic landscape of PC, exemplified by the gain-of-function mutation p53^R172H, which promotes significantly increased metastasis relative to a genetic loss of p53 in Kras^G12D mutant mice (11, 16). Moreover, alterations in signaling pathways that regulate epithelial-to-mesenchymal transition and components that promote cancer stemness also influence the cancer cell phenotype, leading to increased metastasis (17, 18). Last, pancreatic tumors are characterized by a poorly vascularized (19–22), highly fibrotic (12, 19, 23–25), hypoxic (26–29), and immunosuppressive tumor microenvironment (30), where interaction with both cellular and noncellular elements can stimulate the acquisition of invasive and motile properties within PC cells (31–35). This can lead to early and pervasive cancer cell dissemination (36). As metastasis is a key factor in PC-related mortality, therapies that target the cell-intrinsic and cell-extrinsic pathways that lead to metastatic spread would be of great therapeutic value.

Neuropeptide Y (NPY), together with peptide YY (PYY) and pancreatic polypeptide (PPY), form the NPY family of peptides, which in humans signals through four different G protein–coupled receptors (NPY1R, NPY2R, NPY4R, and NPY5R) (37, 38). NPY is the most common neuropeptide in the central and peripheral nervous system where it regulates appetite and energy homeostasis, while PYY and PPY are gut-derived peptides acting in an endocrine and paracrine fashion to control satiety and endocrine functions (39, 40). NPY signaling has also been implicated in numerous biological processes that are commonly deregulated over the course of cancer progression, including cell proliferation (41, 42), immune cell function (43, 44), fibrosis (45), neural invasion (46), and angiogenesis (47, 48). NPY family members have been shown to increase the motility and chemotaxis of breast and prostate cancer cells in vitro (49, 50) and have also been associated with metastatic spread of sarcoma in vivo (51), two properties that are also characteristic of PC tumorigenesis. There is also emerging evidence suggesting that NPY contributes to tumorigenesis in several different cancer types including neuroblastoma (NB) (52), prostate (50, 53), liver (54), colon (55), and breast cancer (56, 57). However, its role in PC tumor development is yet to be assessed.

Here, we identify that NPY and its receptor, Npy1r, are up-regulated in the highly metastatic genetically engineered KP^R172HC mouse model of PC (11, 16) and demonstrate that both ligand and receptor are expressed in the primary tumor as well as liver metastases. Pancreas-specific and whole-body Npy1r genetic ablation in the autochthonous KP^R172HC model led to a significant reduction in liver metastases at the study end point. Critically, using the selective NPY1R antagonist BIBO3304, we show reduced motility of KP^R172HC cells when migrating on cell-derived matrices (CDMs). Furthermore, we recapitulated the decrease in metastasis observed upon Npy1r knockout in the genetically engineered KP^R172HC mouse model using an intrasplenic model of PC metastasis, where BIBO3304 significantly decreased metastatic burden within the liver. Overall, we reveal that NPY/NPY1R signaling is a previously unidentified target in mutant p53–dependent metastasis in PC and its inhibition may represent a potential novel antimetastatic strategy in this highly aggressive and lethal cancer.

RESULTS

NPY expression is up-regulated in PC

Given the role of NPY in promoting disease progression in various other cancers, we aimed to determine whether NPY signaling could also affect pancreatic tumor development and progression. Therefore, we initially assessed the expression of NPY ligands and their receptors in two autochthonous, genetically engineered mouse models (GEMMs) of PC: the low metastatic KP^floxC model (Pdx1-Cre; LSL-Kras^G12D/+; Trp53^flox/+; Fig. 1A) (16, 58) and the highly metastatic KP^R172HC model (Pdx1-Cre; LSL-Kras^G12D/+; LSL-Trp53^R172H/+; Fig. 1A) (11, 16, 58). In both mouse models, PC tumorigenesis is driven by a point mutation in Kras^G12D, but the GEMMs are distinct in their alteration status of the tumor suppressor p53 (Trp53). The KP^floxC model exhibits a loss of function of p53, while the KP^R172HC model expresses a gain-of-function p53^R172H mutation, which we, and others, have previously shown drives a metastatic program to the liver relative to KP^floxC mice (Fig. 1A) (16, 35, 59). Here, quantitative real-time polymerase chain reaction (Q-RT-PCR) analysis identified that Npy mRNA was significantly up-regulated in KP^R172HC tumors relative to the normal pancreas (Fig. 1Bi), while expression of its sister peptides, peptide YY (Pyy) and pancreatic polypeptide (Ppy), was either significantly down-regulated or unchanged (Fig. 1, Bii and Biii). Next, we assessed NPY protein expression in primary PDAC tumors isolated from end-stage KP^R172HC and KP^floxC mice as well as in normal pancreas tissue. Immunohistochemistry (IHC) analysis revealed that NPY protein expression is significantly increased in tumors from both the KP^floxC and KP^R172HC GEMMs (16) relative to the age-matched normal pancreas (Fig. 1C). In addition, NPY is also expressed in liver metastases of the highly metastatic KP^R172HC model (Fig. 1D), suggesting that its elevated expression is conserved during the metastatic cascade.

Fig. 1. NPY expression is up-regulated in both mouse models of PC and human patients with PC.

(A) Schematic representation of the two genetically engineered PC mouse models, KP^floxC and KP^R172HC (16). (B) Q-RT-PCR results of the NPY signaling ligands (i) Npy, (ii) Pyy, and (iii) Ppy from whole tissue samples of normal pancreas and KP^R172HC tumors (n = 3). (C) Representative images and quantification of IHC analysis of NPY protein expression of normal pancreas and KP^floxC and KP^R172HC tumors (n = 3). Scale bars, 200 μm. (D) Representative image of NPY IHC of KP^R172HC liver metastases. Scale bar, 200 μm. (E) Representative images of normal pancreas and KP^floxC and KP^R172HC tumors stained with RNAscope for Npy (red) and Npy1r (blue). Scale bars, 25 μm. (F) Representative images of KP^floxC tumor and KP^R172HC tumor with matched liver metastases stained with RNAscope for Krt19 (red) and Npy (blue, top panel) or Npy1r (blue, bottom panel). Scale bars, 50 μm for the primary tumor and 25 μm for the liver metastases. Zoomed insets are 10 μm by 10 μm. (G and H) Gene expression data from the TCGA database assessed through OncoDB (60) for (G) NPY in human PDAC relative to normal pancreas and (H) NPY1R, NPY2R, NPY4R, and NPY5R expression from human PDAC. Normal pancreas (n = 200) and PDAC (n = 178). (I) (i) Representative images and (ii) quantification of IHC analysis of NPY protein expression of human PDAC tumors relative to patient matched normal pancreas (n = 4). Scale bars, 200 μm. Means ± SEM. ns, P ≥ 0.05; *P < 0.05; **P < 0.01; ****P < 0.0001 by an unpaired parametric t test or a one-way ANOVA with multiple comparisons.

Moreover, this increase in NPY expression was further confirmed by chromogenic RNAscope, where staining for Npy (Fig. 1E, red) and its receptor Npy1r (Fig. 1E, blue) was readily detected in KP^floxC and KP^R172HC tumors relative to the normal pancreas (Fig. 1E). Next, we investigated the spatial localization of Npy and Npy1r using RNAscope by costaining with the epithelial cancer cell marker Keratin 19 (Krt19). Npy (Fig. 1F, top panel) and Npy1r (Fig. 1F, bottom panel) were predominantly colocalized in cancer cells in both the KP^floxC and KP^R172HC primary tumors (Fig. 1F). Both Npy and Npy1r expression was retained in KP^R172HC primary tumor and matched liver metastases, suggesting that it is maintained during metastatic spread (Fig. 1F). Overall, these data show that expression of Npy/NPY and its receptor Npy1r is up-regulated in PC mouse models compared to the normal pancreas and warrant further investigation into the role of NPY signaling through NPY1R in PC tumorigenesis.

To determine whether NPY expression was also up-regulated in tumors of human patients with PC, and to assess the clinical relevance of targeting NPY, we next interrogated the publicly available The Cancer Genomic Atlas (TCGA) dataset via OncoDB (60) and found that NPY mRNA expression was significantly elevated in PDAC tumors relative to the normal pancreas (Fig. 1G). This suggests that NPY could also play a role in tumor development in human patients with PC. Furthermore, we identified that of the four different NPY receptors (NPY1R, NPY2R, NPY4R, and NPY5R), NPY1R was the most highly expressed NPY receptor in human PDAC samples (Fig. 1H). Last, to confirm that NPY was also expressed at the protein level in human patients with PC, we performed IHC on human PDAC samples with the matched adjacent normal pancreas and found that NPY protein expression was significantly higher in tumors relative to the adjacent normal pancreas (Fig. 1Ii; quantified in Fig. 1Iii). Together, these results show that NPY and its receptor NPY1R are up-regulated in both mouse and human PC tumors, suggesting a role for NPY signaling in PC tumor development and progression.

Conditional and whole-body knockout of Npy1r in the KP^R172HC model reduces metastasis to the liver

Given that Npy/NPY and Npy1r expression is up-regulated in the highly metastatic KP^R172HC mouse model and NPY1R was the most highly expressed receptor in human PDAC, we next wanted to investigate the effect of genetic ablation of Npy1r on disease progression in the genetically engineered KP^R172HC mouse model of PC. To distinguish between PC cell–autonomous and nonautonomous NPY1R functions, we crossed KP^R172HC mice with Npy1r^flox/flox mice for conditional Npy1r knockout in the pancreatic epithelium as well as with Npy1r^−/− mice to generate a whole-body Npy1r knockout in the KP^R172HC model. In the pancreas-specific model, Npy1r knockout was driven under the Pdx1-Cre promoter, causing Npy1r to be depleted in all cells also expressing Kras^G12D and Trp53^R172H mutations. We generated long-term cohorts of KP^R172HC Npy1r wild-type mice [Pdx1-Cre, LSL-Kras^G12D/+, LSL-Trp53^R172H/+, Npy1r^+/+ (Npy1r WT)], heterozygous pancreas-specific Npy1r knockout KP^R172HC mice [Pdx1-Cre, LSL-Kras^G12D/+, LSL-Trp53^R172H/+, Npy1r^flox/+ (Npy1r^flox/+)], and homozygous pancreas-specific Npy1r knockout KP^R172HC mice [Pdx1-Cre, LSL-Kras^G12D/+, LSL-Trp53^R172H/+, Npy1r^flox/flox (Npy1r^flox/flox; Fig. 2A)]. Animals were closely monitored for palpable tumors and were euthanized upon reaching humane end points, as previously described (16). Overall, conditional Npy1r knockout was well tolerated with no significant change in body weight observed between the different genotypes (fig. S1A).

Fig. 2. Pancreas-specific Npy1r genetic ablation decreases metastasis to the liver in the KP^R172HC mouse model.

(A) Schematic showing the pancreas-specific Npy1r knockout KP^R172HC model (knockout tissue in blue). (B) Kaplan-Meier analysis of survival of KP^R172HC mice for the three different pancreas-specific Npy1r knockout genotypes, Npy1r WT, Npy1r^flox/+, and Npy1r^flox/flox (n ≥ 20 mice per genotype). (C) Quantification of visible macrometastases in the liver at the study end point. (D) (i) Representative images and (ii) H&E images of sections through the livers at the study end point of the pancreas-specific Npy1r knockout. Metastases are outlined in green. Data for Npy1r WT were the same for both pancreas-specific and whole-body Npy1r knockout survival studies and are also shown in Fig. 3D. Scale bars, 5 mm. Means ± SEM. ns, P ≥ 0.05; **P < 0.01; ***P < 0.001 by a one-way ANOVA or Kaplan-Meier survival analysis.

While Kaplan-Meier analysis of survival showed that conditional Npy1r knockout in the pancreas did not affect the overall survival [median survival of 168 days for Npy1r WT versus 147 days for Npy1r^flox/+ versus 174 days for Npy1r^flox/flox (Fig. 2B)], we observed a significant decrease in metastasis to the liver at the study end point upon both heterozygous and homozygous pancreas-specific knockout of Npy1r relative to KP^R172HC Npy1r WT mice (Fig. 2, C, Di, and Dii). This suggests that NPY1R signaling in PC cells may be required for metastasis to the liver in the KP^R172HC model, which is the most common site of metastasis in human patients with PC. This observation did not occur at other PC metastatic sites, e.g., lung, diaphragm, or peritoneum, following conditional Npy1r ablation (fig. S1B); however, metastasis to these organs was comparatively low across all genotypes when compared with the liver, consistent with this model and clinical features in the human setting.

Furthermore, to assess non–cancer cell–autonomous contributions of NPY signaling through NPY1R, we crossed the KP^R172HC model with Npy1r^−/− whole-body knockout mice. For this, we generated long-term cohorts of Npy1r wild-type KP^R172HC mice [Pdx1-Cre, LSL-Kras^G12D/+, LSL-Trp53^R172H/+, Npy1r^+/+ (Npy1r WT)], Npy1r heterozygous knockout KP^R172HC mice [Pdx1-Cre, LSL-Kras^G12D/+, LSL-Trp53^R172H/+, Npy1r^+/− (Npy1r^+/−)], and Npy1r homozygous knockout KP^R172HC mice [Pdx1-Cre, LSL-Kras^G12D/+, LSL-Trp53^R172H/+, Npy1r^−/− (Npy1r^−/−; Fig. 3A)]. As before, all animals were closely monitored for palpable tumors and all animals were euthanized upon reaching humane end points. Here, we also observed no overall changes in body weight upon whole-body Npy1r knockout (fig. S1C). We found that Npy1r whole-body knockout in the KP^R172HC GEMM mimicked our results from the conditional Npy1r knockout, exhibiting no overall changes in survival [median survival of 168 days for Npy1r WT versus 198 days for Npy1r^+/− versus 190 days for Npy1r^−/− (Fig. 3B; compared to Fig. 2B)] but again showing a significant decrease in metastasis to the liver relative to KP^R172HC Npy1r WT mice (Fig. 3, C, Di, and Dii). Moreover, there were no significant changes in metastasis to organ sites outside of the liver (fig. S1D), consistent with the Npy1r conditional knockout model.

Fig. 3. Whole-body Npy1r genetic ablation decreases metastasis to the liver in the KP^R172HC mouse model.

(A) Schematic showing the whole-body Npy1r knockout KP^R172HC model (knockout tissue in blue). (B) Kaplan-Meier analysis of survival of KP^R172HC mice for the three different whole-body Npy1r knockout genotypes, Npy1r WT, Npy1r^+/−, and Npy1r^−/− (n ≥ 19 mice per genotype). (C) Quantification of visible macrometastases in the liver at the study end point. (D) (i) Representative images and (ii) H&E images of sections through the livers at the study end point of the whole-body Npy1r knockout. Metastases are outlined in green. Data for Npy1r WT were the same for both pancreas-specific and whole-body Npy1r knockout survival studies and are also shown in Fig. 2D. Scale bars, 5 mm. Means ± SEM. ns, P ≥ 0.05; *P < 0.05; **P < 0.01 by a one-way ANOVA or Kaplan-Meier survival analysis.

We also determined the time to palpable tumor and time from palpable tumor to the end stage for conditional and whole-body Npy1r knockout animals and found no significant difference compared to KP^R172HC Npy1r WT mice (fig. S2, A and B), suggesting that the decrease in liver metastasis observed upon Npy1r knockout is not due to a change in overall disease progression. Moreover, because of the known role of NPY/NPY1R signaling in energy homeostasis control (40, 61, 62), we quantified the weights of key organs and adipose and muscle tissues of our mice at the end stage (fig. S3, A and B). Here, we found that while Npy1r knockout did not affect most organs measured, we observed significant increases in white adipose tissue weight and muscle weight, which may provide an additional benefit in PC, which is commonly associated with tissue wasting (63–65).

The pancreatic tumor microenvironment is a complex ecosystem consisting of extracellular matrix (ECM) as well as a plethora of different cell types including cancer-associated fibroblasts, endothelial cells, and immune cells, which have all previously been shown to regulate cancer metastasis to distant sites (12, 19, 35, 66–69). We therefore assessed the pancreatic tumor microenvironment in end-stage tumors of the whole-body Npy1r knockout using IHC. This revealed that genetic Npy1r ablation did not significantly alter the abundance of stromal cell populations or ECM including cancer-associated fibroblasts [αSMA (ACTA2) and PDGFRB; fig. S4, A and B], endothelial cells [CD31 (PECAM1); fig. S4C], and fibrillar collagens assessed by transmitted and polarized light imaging of Picrosirius Red (fig. S5, A and B). Furthermore, no significant difference in the abundance of T cell subsets (CD4, CD8, and FOXP3; fig. S6, A to C), macrophages, or neutrophils (F4/80, ELANE, and MPO; fig. S7, A to C) was detected upon Npy1r knockout. These data therefore suggest that the decreased metastasis we observed upon Npy1r knockout was not due to changes in the overall composition of the primary tumor microenvironment.

We next assessed changes in gene expression upon Npy1r knockout using RNA sequencing (RNA-seq) and mass spectrometry (MS) proteomics. This revealed that 107 transcripts and 427 proteins were differentially expressed between end-stage tumors of KP^R172HC Npy1r WT and KP^R172HC Npy1r^−/− mice (figs. S8, A and B, and S9, A and B, and tables S1 and S2), which may contribute to the different metastatic capacity observed for both genotypes. Down-regulated transcripts and proteins included genes previously described to be regulated by NPY (70), such as CCN2, whose inhibition has been shown to decrease bone metastases of MDA-MB-231 breast cancer cells (71), as well as CIRBP, which was previously found to be required for metastasis in a xenograft mouse model of bladder cancer (72). Similarly, we found significant down-regulation of LAMA3 mRNA and protein, whose expression has previously been correlated with liver metastasis in patients with PC (73). Up-regulated transcripts and proteins included LGALS2, which has been implicated to restrain tumor progression in colorectal and breast cancer (74, 75), as well as CRABP2, whose overexpression has been associated with decreased invasiveness of ER+ breast cancer cells (76). Gene set enrichment analysis (GSEA) enrichment also identified a down-regulation in transforming growth factor–β signaling, coagulation, and hypoxia hallmarks (figs. S8B and S9B and tables S1 and S2), which have previously been associated with metastasis (77–80). Further studies are needed to elucidate how these alterations in gene expression programs contribute to the decrease in metastasis following Npy1r knockout.

Together, our GEMM studies support the notion that NPY signaling through NPY1R is required for the metastatic phenotype in KP^R172HC cancer cells with no further reduction in metastases to the liver upon whole-body Npy1r knockout compared to the pancreas-specific Npy1r knockout in KP^R172HC mice (Fig. 2C compared with Fig. 3C). Heterozygous Npy1r knockout in both the conditional and whole-body settings was sufficient to decrease metastasis to the liver, thereby suggesting that pharmacological inhibition of NPY1R function instead of complete Npy1r ablation may be a feasible approach to decrease PC metastasis. These results suggest that the NPY/NPY1R signaling axis could be a novel target in PC metastasis and is required for mutant p53–dependent metastatic spread to the liver in KP^R172HC mouse models, highlighting NPY1R as a previously unidentified target in PC.

Pharmacological NPY1R inhibition reduces the motility of KP^R172HC cells

Following the observation that pancreas-specific, genetic ablation of Npy1r in the KP^R172HC mouse model significantly reduced metastasis to the liver, which was not further improved upon by whole-body Npy1r knockout, we next aimed to elucidate how PC cell–autonomous NPY1R inhibition might be driving this antimetastatic effect. Previous studies have shown that NPY signaling stimulates cancer cell motility and invasive potential in vitro in breast and prostate cancer cell lines (49, 50, 80, 81), warranting further investigation of the role of NPY1R in PC cell migration. We first assessed the expression of Npy in previously characterized primary cancer lines isolated from the low metastatic KP^floxC mouse model compared to the highly metastatic KP^R172HC mouse model (Fig. 4A) (16, 82). To continue the assessment of NPY1R function in an immunocompetent setting, for the KP^R172HC cell line, we used the well-characterized syngeneic TB32043 cell line, which was isolated from KP^R172HC mice backcrossed with the C57BL/6J background (11, 83, 84). We found that Npy mRNA expression was significantly increased in primary cancer cells of the C57BL/6J syngeneic KP^R172HC model relative to primary cancer cells of the KP^floxC model via Q-RT-PCR (Fig. 4Bi), suggesting that Npy is important for the maintenance of the highly metastatic phenotype of the KP^R172HC cancer cells. Notably, the up-regulation of Npy expression in KP^R172HC cells compared to KP^floxC cells (Fig. 4Bi) was enhanced compared to our immunohistochemical results in tumors (Fig. 1C), suggesting that Npy expression may depend on the local tumor microenvironment and/or an enhancement of Npy expression in KP^R172HC cells upon backcross with the C57BL/6J background. Expression of its sister peptides Pyy and Ppy was unchanged or undetectable in KP^R172HC cancer cells relative to KP^floxC cancer cells (Fig. 4, Bii and Biii), further confirming that Npy, and not its sister peptides, may have a role in the metastatic spread in the KP^R172HC mouse model of PC.

Fig. 4. Pharmacological NPY1R inhibition decreases KP^R172HC cell motility but not KP^floxC motility in vitro.

(A) Schematic showing the low metastatic KP^floxC model and highly metastatic KP^R172HC PC mouse models, from which cancer cells were derived and used in the following experiments. (B) Quantification of (i) Npy, (ii) Pyy, and (iii) Ppy expression using Q-RT-PCR of cancer cells isolated from the KP^floxC and KP^R172HC models. (C) Schematic showing CDM generation, decellularization, and cancer cell seeding in vitro. (D) Representative binary images of KP^R172HC cells migrating on CDMs upon BIBO3304 or vehicle treatment. (E) Quantification of KP^R172HC cell total distance traveled, average speed, average distance from the origin, and maximum distance from the origin treated with vehicle or 1 μM BIBO3304 over 8 hours (n = 3). (F) Representative x-y tracks of KP^R172HC cells over 8 hours upon treatment with vehicle or 1 μM BIBO3304 (n = 3, 10 cell tracks per polar plot). (G) Representative binary images of KP^floxC cells migrating on CDMs upon BIBO3304 or vehicle treatment. (H) Quantification of KP^floxC cell total distance traveled, average speed, average distance from the origin, and maximum distance from the origin treated with vehicle or 1 μM BIBO3304 over 8 hours (n = 3). (I) Representative x-y tracks of KP^floxC cells over 8 hours upon treatment with vehicle or 1 μM BIBO3304 (n = 3, 10 cell tracks per polar plot). Scale bars, 100 μm. Means ± SEM. ns, P ≥ 0.05; *P < 0.05; **P < 0.01; ***P < 0.001 by an unpaired parametric t test.

Next, CDMs were generated as previously described (85, 86), where fibroblasts are treated with ascorbic acid to induce ECM deposition (Fig. 4C). After 7 days, the CDMs were decellularized and primary KP^R172HC cancer cells seeded onto the CDMs to monitor their migratory patterns on a three-dimensional ECM scaffold (Fig. 4, C and D; see movie S1). Using primary KP^R172HC cancer cells, we tracked cancer cell movement on CDMs upon pharmacological NPY1R inhibition using the NPY1R antagonist BIBO3304 (Fig. 4, C to F) (87) to assess how inhibiting the NPY signaling axis might affect PC cell motility. BIBO3304 is a selective small-molecule NPY1R antagonist exhibiting high affinity and specificity for NPY1R (87–89). The total distance the cancer cells traveled, the average speed, the average distance from the point of origin, and the maximum distance from the point of origin were all significantly decreased upon NPY1R inhibition using BIBO3304 (Fig. 4, E and F), confirming that NPY1R may regulate the motility of KP^R172HC cancer cells. Moreover, BIBO3304 treatment of KP^floxC cells migrating on CDMs did not result in a significant change in any of the parameters assessed (total difference traveled, average speed, and average/maximum distance from the point of origin; Fig. 4, G to I, and movie S2), suggesting that the low basal level of migration observed in KP^floxC cells does not require NPY/NPY1R signaling. Together, these results reveal a decrease in mutant p53–dependent PC cell motility upon NPY1R inhibition, which could, in part, be contributing to the antimetastatic effect observed upon conditional and whole-body Npy1r knockout within the KP^R172HC mouse model (Figs. 2 and 3).

Pharmacological NPY1R inhibition reduces metastasis to the liver in vivo

We next sought to interrogate whether pharmacological NPY1R inhibition using BIBO3304 would recapitulate the observed decrease in liver metastasis found upon genetic Npy1r deletion. Therefore, we used an intrasplenic model of PC metastasis where the syngeneic KP^R172HC cancer cells were injected into the spleens of C57BL/6J mice to drive metastasis to the liver. In this model, the flux of cancer cells through the portal vein of the spleen results in robust metastasis to the liver, as achieved previously (33, 90). Mice were subjected to daily intraperitoneal injections with BIBO3304 (1 mg/kg) or vehicle before and after intrasplenic injection, resulting in pharmacological NPY1R inhibition being present during the transit and liver colonization phase of KP^R172HC cancer cell metastasis (Fig. 5A). Mice were euthanized 12 days postsurgery (Fig. 5A). Consistent with the genetic Npy1r knockout models, pharmacological NPY1R inhibition resulted in significantly decreased visible metastases within the liver (Fig. 5, B and Di), further confirming that NPY1R is required for metastatic spread in PC. Livers were also stained for KRT19 to detect PC cells to further validate the antimetastatic effect of NPY1R inhibition. From this, total metastatic burden within the liver (KRT19⁺ area) and the number of large metastases (metastases >50,000 μm²) were assessed and were found to be significantly decreased upon BIBO3304 treatment (Fig. 5, C and Dii). These results show that pharmacological NPY1R inhibition using BIBO3304 reduces PC metastasis to the liver, highlighting the possibility that NPY1R inhibition could represent a novel future therapeutic strategy to counteract PC metastasis in combination with standard-of-care therapies, such as gemcitabine/Abraxane or FOLFIRINOX (5, 6).

Fig. 5. Pharmacological NPY1R inhibition decreases metastatic burden within the liver in vivo.

(A) Schematic showing the treatment schedule for the intrasplenic xenograft experiment with BIBO3304 or vehicle control (saline). (B) Quantification of the number of visible liver metastases per mouse (n ≥ 13 mice per treatment). (C) Quantification of the KRT19⁺ surface area and the number of metastases >50,000 μm² per tissue depth upon treatment with vehicle or BIBO3304. (D) (i) Representative images (top panel) and (ii) IHC images (bottom panel) of KRT19-stained livers, outlining liver tissue in pink and metastases in green, upon treatment with vehicle or BIBO3304. Scale bars, 5 mm. Means ± SEM. *P < 0.05; **P < 0.01 by an unpaired parametric t test.

To interrogate whether BIBO3304 also affects early metastatic outgrowth, we repeated the intrasplenic model and assessed metastatic burden on day 5 (4 days after intrasplenic cancer cell injection; Fig. 6A). We found that in this early intrasplenic model, BIBO3304 also significantly reduced liver metastases as assessed by both visual inspection (Fig. 6B) and KRT19 IHC (Fig. 6, C and D). Moreover, to assess whether BIBO3304 treatment affects cell survival or proliferation in these early metastases, we stained our samples for cleaved caspase-3 and Ki67. While we did not observe any changes in cell survival (Fig. 6E), we observed a significant reduction in Ki67 following BIBO3304 treatment (Fig. 6F), suggesting that BIBO3304 may reduce metastasis by decreasing cell proliferation. Collectively, these results demonstrate a requirement for NPY1R in PC metastasis and warrant further assessment of the NPY/NPY1R signaling axis as a cotarget in conjunction with current and contemporary standard-of-care therapies in PC (5, 6).

Fig. 6. Early NPY1R inhibition decreases metastatic burden within the liver in vivo.

(A) Schematic showing the treatment schedule for the early intrasplenic xenograft experiment with BIBO3304 or vehicle control (saline). (B) Quantification of the number of visible liver metastases per mouse (n ≥ 8 mice per treatment). (C) Quantification of the KRT19⁺ surface area and the number of metastases >10,000 μm² per tissue depth upon treatment with vehicle or BIBO3304. (D) Representative IHC images of KRT19-stained livers, outlining liver tissue in pink and metastases in green, upon treatment with vehicle or BIBO3304. Scale bars, 5 mm. (E) Quantification and representative images of cleaved caspase-3 (CC3) IHC in liver metastases on day 5 following intrasplenic KP^R172HC cancer cell injection and treatment with vehicle or BIBO3304 (n ≥ 8 mice per treatment). (F) Quantification and representative images of Ki67 IHC in liver metastases on day 5 following intrasplenic KP^R172HC cancer cell injection and treatment with vehicle or BIBO3304 (n ≥ 8 mice per treatment). Scale bars, 150 μm. Means ± SEM. ns, P ≥ 0.05; *P < 0.05; ***P < 0.001 by an unpaired parametric t test.

DISCUSSION

The role of the NPY signaling axis in tumorigenesis has been established in multiple cancers. For example, in NB, NPY increases tumor growth in mice xenografts (52), and high serum NPY correlates with decreased survival, increased metastasis, and disease recurrence in human patients with NB (91, 92). In breast cancer, NPY1R expression was confirmed in 85% of primary human breast carcinomas and 100% of lymph node metastases exhibiting a switch from NPY2R expression in normal breast tissue to NPY1R dominant expression during cancer progression (56). In addition, high NPY1R expression in circulating tumor cells isolated from peripheral blood has also been shown to correlate with poor survival and lymph node metastasis in patients with breast cancer (57). Furthermore, NPY signaling was shown to regulate angiogenesis in colon cancer and tumor growth and fibrosis in liver cancer (54, 55). However, NPY’s function in PC tumorigenesis has yet to be assessed. Therefore, we sought to characterize the role of NPY in PC and elucidate whether NPY is playing a role in PC tumorigenesis and progression. In this study, we found that NPY is up-regulated at both the mRNA and protein levels in the highly metastatic KP^R172HC model of PC relative to the normal pancreas and is predominantly colocalized with cancer cells in the primary tumor and liver metastases. Furthermore, NPY expression is increased in the tumors of human patients with PDAC relative to the normal pancreas, and expression analysis confirmed that Npy1r is up-regulated in KP^R172HC tumors compared with the normal pancreas. Moreover, NPY1R is the most highly expressed NPY receptor in human patients with PDAC, supporting a role of NPY signaling in murine and human PC tumorigenesis.

PC metastasis is a key factor in human patients’ mortality, with the 5-year survival of patients presenting with metastatic disease being only 3% (1). Here, we show that both heterozygous and homozygous pancreas-specific and whole-body knockout of Npy1r results in significantly decreased metastasis to the liver in the highly metastatic KP^R172HC mouse model. The liver is the most common site of metastasis in human patients, with ~80% of patients with PC presenting with hepatic metastases at autopsy (93). Heterozygous Npy1r knockout in both the pancreas-specific and whole-body knockout models was sufficient to decrease metastatic burden within the liver, suggesting that complete Npy1r depletion is not required to impair this metastatic program. Our observations are consistent with other cancer types where NPY has been shown to play a role in metastasis (51, 57, 80). In a recent study investigating Ewing sarcoma, it was found that enhanced metastasis to the bone was driven by a hypoxia-induced activation of NPY signaling components (51). This observation might not be limited to Ewing sarcoma tumors, as hypoxia also induces NPY1R and NPY5R expression in breast cancer cells, causing them to migrate more upon NPY stimulation in vitro (80). Moreover, human patients with breast cancer whose circulating tumor cells were NPY1R positive exhibited decreased overall survival and the level of NPY1R also correlated with late-stage disease and lymph node metastases (57). Last, immunohistochemical analysis of human patients with prostate cancer found that NPY, NPY1R, NPY2R, and NPY5R expression was all significantly increased at the invasive border relative to the bulk tumor mass (50). These data, alongside our observation of decreased liver metastasis upon Npy1r ablation in the KP^R172HC mouse model of PC, highlight a new role for the NPY signaling axis in PC aggressiveness.

Transcriptomic and proteomic assessment via RNA-seq and MS, respectively, identified ~500 differentially expressed transcripts and proteins in pancreatic tumors following Npy1r knockout. While some of these differentially expressed genes/proteins have already been shown to have a role in cancer metastasis, such as CCN2 (71), CIRBP (72), and LAMA3 (73), many are underexplored in PC, warranting further studies based on our datasets.

We also investigated the cancer cell–intrinsic features of NPY1R inhibition in KP^R172HC cells and found that their three-dimensional motility was decreased upon NPY1R inhibition when migrating upon CDMs. This is consistent with studies that have found a role for NPY in regulating the motility and chemotaxis of cancer cells. For example, two breast cancer cell lines MDA-MB-231 and MCF-7 exhibit increased motility and invasion in response to NPY, which could be blocked with pharmacological NPY1R and NPY5R antagonists (49). Moreover, NPY was shown to regulate chemotaxis in the highly aggressive LNCaP prostate cancer cell line (50). The observed reduction in the motility of KP^R172HC cells upon BIBO3304 treatment suggests a role for NPY in cancer cell movement and could represent one of the modes through which NPY1R inhibition decreases metastasis to the liver.

Our study focused on NPY1R because of its significantly increased expression in human patients with PDAC relative to NPY2R, NPY4R, and NPY5R. However, other NPY receptors have been implicated in promoting tumorigenicity in other cancers; for example, NPY2R is up-regulated in vascular endothelial growth factor A–depleted orthotopic models of colon cancer, and NPY2R antagonists inhibited angiogenesis and tumor growth in these models (55). Furthermore, in NB, NPY has been shown to promote cell motility via increased cytoskeleton remodeling (94), and NPY5R expression was significantly up-regulated in NB cells adjacent to blood vessels, suggesting preferential NPY5R expression in angioinvasive cancer cells (92). Therefore, other NPY receptors could be playing a role in PC and future studies could aim to elucidate their functions.

Last, we were able to recapitulate the observed decrease in liver metastases seen during Npy1r genetic ablation using pharmacological NPY1R inhibition in an intrasplenic model of metastasis. Here, using pharmacological NPY1R inhibition via the small-molecule antagonist BIBO3304, we show a significant decrease in metastatic burden in the liver compared with vehicle-treated mice, highlighting a potential future therapeutic strategy to reduce metastatic burden for patients. BIBO3304 exhibits subnanomolar affinity for human NPY1R (0.69 ± 0.16 nM), which is comparable to that of rodent NPY1R (0.72 ± 0.42 nM), and therefore can bind and inhibit human NPY1R (87) but has yet to be used in human clinical trials. The fact that BIBO3304 is unable to cross the blood-brain barrier and thus cannot affect the central action of other NPY1R-mediated functions, such as stimulating appetite and reducing energy expenditure, represents a distinct advantage in a cancer setting, since PC progression is often associated with tissue wasting (63–65). Here, we observed an increase in adipose and muscle tissue mass in Npy1r knockout settings, providing an additional benefit on top of reducing metastasis in PC.

Collectively, our findings establish NPY1R as a previously unidentified drug target in PC metastasis. We demonstrate that genetic and pharmacological inhibition of NPY1R reduces metastasis in the highly aggressive KP^R172HC model. Targeting this pathway may therefore represent a highly effective novel antimetastatic strategy for future assessment in conjunction with other standard-of-care approaches, such as gemcitabine/Abraxane or FOLFIRINOX chemotherapies (5, 6), to improve outcomes for patients with PC.

MATERIALS AND METHODS

Experimental design

This study assesses the function of NPY signaling in PC. Q-RT-PCR analysis of in vitro assays of cell lines and CDMs was conducted in independent biological triplicates with three technical replicates per repeat for each genotype or treatment group. Q-RT-PCR and IHC analysis of mouse tumor tissue and normal pancreas were performed with tumors/tissues isolated from three independent mice per genotype. IHC analysis of human PDAC tissue and matched normal pancreas was performed with tumors and tissues isolated from four patients with PC, taking four representative regions of interest (ROIs) per tissue type per patient. Human NPY and NPY1R expression was assessed via analysis of mRNA expression from the TCGA datasets through OncoDB (60). Mouse numbers used in in vivo experiments are outlined in the corresponding figure legends.

Statistical analysis

Unless stated otherwise, P values were determined using unpaired parametric t test (comparison between two groups) or an unpaired one-way analysis of variance (ANOVA) assuming Gaussian distribution with Tukey’s multiple comparisons test (comparison between more than two groups). Kaplan-Meier survival curves were evaluated using a log-rank Mantel-Cox test. All statistical analysis was performed using GraphPad Prism (GraphPad Software Inc., CA) with statistical significance defined as ns, P ≥ 0.05; *P < 0.05; **P < 0.01; ***P < 0.001.

Human ethics

Human ethics approval for acquisition of data and biological material was obtained from North Shore Local Health District human research ethics committee (2023/ETH02130).

Animals

Animal experiments were conducted in compliance with the Australian code of practice for the care and use of animals for scientific purposes and the Garvan/St. Vincent’s Animal Ethics committee (guidelines 19/06, 19/10, 19/13, 22/08, 22/09, and 22/10). All experimental end points were in accordance with Garvan/St. Vincent’s Animal Ethics committee (guidelines 19/06, 19/10, 19/13, 22/08, 22/09, and 22/10). Mice were kept in ventilated cages on a 12-hour light/dark cycle with 24-hour access to food and water.

Quantitative real-time polymerase chain reaction (Q-RT-PCR)

Q-RT-PCR on tumors from end-stage KP^R172HC mice and normal pancreas was performed by collecting tissue samples isolated from three independent mice per genotype and immediately snap freezing the tissues before homogenization in 500 μl of RNA Prep Buffer using metal bead lysing matrix Eppendorf tubes (MP Biomedicals) in a FastPrep Tissue Homogenizer (MP Biomedicals). Q-RT-PCR on cell lines (KP^R172HC TB32043 and KP^floxC cancer cells) was performed by seeding three repeats of three replicates of 500,000 cells into 10-cm dishes for 48 hours. The cells were then washed twice with phosphate-buffered saline (PBS), lifted with a cell scraper, transferred to an Eppendorf tube, and spun at 13,000 rpm for 5 min at 4°C. The supernatant was removed, and the cell pellet was snap frozen and stored at −80°C. RNA extraction was performed using the RNA Qiagen Spin Column Extraction protocol (Qiagen). cDNA was synthesized using the Roche transcriptor first strand cDNA synthesis kit’s protocol (Roche). Q-RT-PCR was performed using the Universal ProbeLibrary System (UPL; Roche) on a LightCycler 480 System (Roche) using the forward and reverse primers listed in Table 1. Gapdh was used as control for tissues and Rplp0 used as control for cell lines.

Table 1. UPL probe number and forward and reverse primer sequences for all genes assessed during Q-RT-PCR analysis.

Genes	UPL probe #	Forward	Reverse
Npy	96	5′-GAA AGC ACA GAA AAC GCC CCC AG-3′	5′-AAA TGG GGC GGA GTC CAG CCT A-3′
Pyy	17	5′-CCT ACC CTG CCA AAC CAG-3′	5′-GGA CAT CTC TTT TTC CAT ACC G-3′
Ppy	1	5′-TGG CTT GAT TCC CTG CTC-3′	5′-ACG GGC TGA AGA CAA GAG AG-3′
Gapdh	52	5′-GGG TTC CTA TAA ATA CGG ACT GC-3′	5′-CCA TTT TGT CTA CGG GAC GA-3′
Rplp0	9	5′-ACT GGT CTA GGA CCC GAG AAG-3′	5′-CTC CCA CCT TGT CTC CAG TC-3′

Immunohistochemistry

Hematoxylin and eosin (H&E) and IHC staining was performed on 4-μm sections of formalin-fixed paraffin-embedded tissue with sections for H&E placed onto plain glass slides and sections for IHC placed onto positively charged slides. The slides were incubated for 2 hours in a 60°C oven for maximum adhesion. H&E sections were deparaffinized and stained following standard H&E procedures on the Leica ST5010 Autostainer XL with hematoxylin [Hematoxylin Harris nontoxic (acidified), Australian Biostain] and eosin (Eosin Phloxine Alcoholic 1%, Australian Biostain). All IHC was performed on a Leica Bond RX machine using the Bond Polymer Refine Detection Kit (DS9800), where slides were first dewaxed using Bond Dewax Solution (Leica, AR2992) before heat-induced epitope retrieval at 100°C with either epitope retrieval solution 1 (pH 6, Leica AR9961) or epitope retrieval solution 2 (pH 9, AR9640) for 20 to 40 min (please see Table 2 for details). The slides were then incubated with primary antibody for 60 min before being washed, incubated with secondary antibody (polymer), and visualized with 3,3′-diaminobenzidine. Please see Table 2 for primary antibody dilution, host source, and retrieval protocol for the antibodies used. For NPY staining of KP^R172HC tumors, KP^floxC tumors, and normal pancreas, tumor/tissue sections from three independent mice per genotype were stained. For human tissues, sections of tumor and adjacent normal pancreas on the same slide and isolated from five independent patients with PC were stained, and four ROIs were taken from epithelial cancer regions versus four ROIs from normal pancreas regions and averaged. For IHC analysis of KP^R172HC Npy1r WT and KP^R172HC Npy1r^−/− tumors (from 10 animals with the highest number of liver metastases per genotype), 10 ROIs per tumor were analyzed. For IHC analysis of Ki67 and cleaved caspase-3 of the early intrasplenic study (day 5), 10 metastases per animal were analyzed. Slides were imaged on the Hamamatsu NanoZoomer S210 Digital slide scanner and quantified in QuPath.

Table 2. List of primary antibodies used for IHC with their supplier or company, catalog number, level of dilution, host source, staining system, and retrieval protocol.

Antibody	Supplier/company	Catalog no.	Antibody dilution	Host	Staining system	Retrieval
NPY	Abcam	Ab30914	1:2000	Rabbit	Leica BOND IHCF	pH 9, 20 min
KRT19	Abcam	Ab133496	1:1000	Rabbit	Leica BOND IHCF	pH 6, 40 min
Ki67	Thermo Fisher Scientific	Ab-4 Neomark	1:500	Rabbit	Leica BOND IHCF	pH 9, 30 min
CC3	Cell Signaling Technology	9661	1:200	Rabbit	Leica BOND IHCF	pH 9, 20 min
αSMA	Abcam	Ab5694	1:100	Rabbit	Envision Rabbit	pH 9, 30 min
PDGFRB	Cell Signaling Technology	3169	1:100	Rabbit	Leica BOND IHCF	pH 9, 30 min
CD31	DIA-310	Dianova	1:100	Rat	Leica BOND IHCF	pH 6, 40 min
CD4	Cell Signaling Technology	25229	1:100	Rabbit	Leica BOND IHCF	pH 9, 20 min
CD8	Cell Signaling Technology	98941	1:200	Rabbit	Leica BOND IHCF	pH 9, 20 min
FOXP3	Cell Signaling Technology	12653	1:400	Rabbit	Leica BOND IHCF	pH 9, 20 min
F4/80	Cell Signaling Technology	70076	1:100	Rabbit	Envision Rabbit	pH 9, 30 min
ELANE	Abcam	Ab68672	1:500	Rabbit	Leica BOND IHCF	pH 6, 40 min
MPO	Agilent	A039829-2	1:2000	Rabbit	Leica BOND IHCF	pH 9, 30 min

Picrosirius Red histological staining and transmitted and polarized light imaging and analysis

Formalin-fixed paraffin-embedded tumors (from 10 animals with the highest number of liver metastases per genotype) were cut at 4-μm section thickness and allowed to adhere to Superfrost Plus slides. Slides were dewaxed on a Leica ST5010 Autostainer XL and manually stained using 0.1% Picrosirius Red and 0.02% phosphomolybdic acid for fibrillar collagens (Australian Biostain). Following a rinse in acidified water and dehydration in a graded ethanol series, slides were covered with coverslips using a Leica Coverslipper (CV5030) and imaged on a Slideview VS200 slide scanner (Olympus) in bright-field and polarized (birefringence) mode. The Picrosirius Red signal was analyzed in QuPath, and the birefringence signal was analyzed with ImageJ as described previously (19, 33) with 10 ROIs per animal. To assess the red-orange, yellow, and green birefringence signals, images were analyzed using ImageJ, as described previously (19, 33). Briefly, hue-saturation balance thresholding was applied (high birefringence/red-orange 0 > H < 27 | 0 > S < 255 | 70 > B < 255, medium birefringence/yellow 28 > H < 47 | 0 > S < 255 | 70 > B < 255, and low birefringence/green 48 > H < 140 | 0 > S < 255 | 70 > B < 255).

RNAscope

RNAscope was performed using RNAscope 2.5 HD Duplex Detection kit (ADV322500), as previously described (95, 96). Sections (4 μm) of formalin-fixed paraffin-embedded tissues were cut and then dehydrated, and then an ImmEdge Hydrophobic Barrier Pen (ADV310018) was used to draw a hydrophobic barrier around the tissue of interest. The tissue was then pretreated with RNAscope H₂O₂ & Protease Plus Reagent (ADV322330), followed by submergence in RNAscope Target Retrieval Reagent (ADV322000) and then, lastly, the addition of Protease Plus (ADV322340). RNA was then amplified and detected using probes for Npy (ACDBIO, 313321), Npy1r (ACDBIO, 427021), and Krt19 (ACDBIO, 402941) using RNAscope 2.5 HD Duplex Reagent Kit (ADV322500). For RNAscope, a normal pancreas, KP^floxC end-stage tumor, KP^R172HC end-stage tumor, and the matched liver were stained. Five field-of-view images were acquired at 63× magnification on a Leica DM4000 bright-field microscope for each tissue and representative image shown.

Expression analysis in OncoDB

The OncoDB database (https://oncodb.org/) was used to assess NPY expression in patients with PDAC. RNA-seq data of pancreatic tissue obtained from patients with PDAC (n = 178) and normal pancreas controls (n = 200) were analyzed for NPY expression [transcripts per million (TPM)] and subjected to an unpaired parametric t test. NPY1R, NPY2R, NPY4R, and NPY5R expression (TPM) was analyzed in patients with PDAC (n = 178), and an unpaired one-way ANOVA assuming Gaussian distribution with Tukey’s multiple comparisons test was performed.

Survival studies with genetically engineered mice

KP^R172HC mice (Pdx1-Cre, LSL-Kras^G12D/+, LSL-Trp53^R172H/+) (11, 16) were crossed with Npy1r^flox mice (89, 97) for conditional knockout of Npy1r in pancreatic epithelial cells. Cohorts of KP^R172HC Npy1r wild-type mice (Pdx1-Cre, LSL-Kras^G12D/+, LSL-Trp53^R172H/+, Npy1r^+/+), heterozygous pancreas-specific Npy1r knockout KP^R172HC mice (Pdx1-Cre, LSL-Kras^G12D/+, LSL-Trp53^R172H/+, Npy1r^flox/+), and homozygous pancreas-specific Npy1r knockout KP^R172HC mice (Pdx1-Cre, LSL-Kras^G12D/+, LSL-Trp53^R172H/+, Npy1r^flox/flox) were generated. Furthermore, KP^R172HC mice were crossed with Npy1r whole-body knockout mice (89, 97) to generate cohorts of KP^R172HC Npy1r wild-type mice (Pdx1-Cre, LSL-Kras^G12D/+, LSL-Trp53^R172H/+, Npy1r^+/+), whole-body Npy1r heterozygous knockout KP^R172HC mice (Pdx1-Cre, LSL-Kras^G12D/+, LSL-Trp53^R172H/+, Npy1r^+/−), and homozygous Npy1r knockout KP^R172HC mice (Pdx1-Cre, LSL-Kras^G12D/+, LSL-Trp53^R172H/+, Npy1r^−/−). KP^R172HC Npy1r WT animals were obtained from a pool of mice generated from the progeny of both genetic crosses. All mice were bred at Australian BioResources, and genotyping was performed at the Garvan Molecular Genetics facility. Mice with the appropriate genotypes were enrolled into the study at ~6 weeks of age and weighed, palpated, and monitored once weekly until detection of a palpable tumor, after which monitoring was increased to at least 3× weekly. The time to palpable tumor and time from palpable tumor to the end point were determined for all animals where tumor could be clearly detected via palpation. Mice were euthanized upon reaching the study end point, which included the development of signs of ascites, overnight weight loss of ≥10% or weight loss of ≥20% compared to the maximum body weight, hunching posture, and signs of pain. Tissues and organs were harvested from end-point animals and weighed unless determined not feasible (e.g., because of tissues/organs excessively infiltrated/invaded by the tumor). Mice were censored from the study if they had to be euthanized because of unspecific end points not related to PC, including presentation of severe prolapses, papilloma, and lymphomas.

RNA-seq of KP^R172HC tumors

Tumors were isolated from KP^R172HC Npy1r WT and Npy1r^−/− mice at the end stage, snap frozen, and stored at −80°C. RNA extraction was performed on 10 KP^R172HC Npy1r WT and 10 KP^R172HC Npy1r^−/− end-stage tumors (from 10 animals with the highest number of liver metastases per genotype) using the QIAGEN RNeasy Mini Kit (no. 74104) according to the manufacturer’s instructions. RNA concentration and integrity were assessed using the Agilent 4200 TapeStation system and a Qubit 3.0 Fluorometer (Thermo Fisher Scientific). Library preparation was performed using the Illumina Stranded mRNA Library Preparation Kit according to the manufacturer’s protocol, and paired-end sequencing was performed using the NovaSeq S4 Flow Cell (300 cycles) Standard.

RNA-seq data were processed using Nextflow pipeline nf-core/rnaseq (3.16.0) (98). Briefly, sequence reads were aligned to mouse reference genome assembly GRCm39 using STAR (2.7.10a) (99) and expression counts estimated for Gencode basic gene annotation M35 using RSEM (1.3.1) (100). Downstream analysis was performed in R (4.4.2) using Bioconductor packages DESeq2 (1.44.0) (101) for differential expression analysis and fgsea (1.30.0) for GSEA. P values were adjusted using the Benjamini-Hochberg procedure to control the false discovery rate.

MS proteomics sample preparation

Tumors were isolated from KP^R172HC Npy1r WT and Npy1r^−/− at the end stage, snap frozen, and stored at −80°C. Samples were prepared as previously reported (19). Briefly, samples were lysed with a bead beater in 1% SDS and 100 mM tris (pH 8.5) for 2 min at 50 Hz. Cell debris was removed by centrifugation, and the supernatant was denatured with 100 mM tris(2-carboxyethyl)phosphine hydrochloride and alkylated with 40 mM chloroacetic acid for 5 min at 45°C. Protein concentration was determined using Qubit (Protein Assay Kit, Life Technologies). Proteins were coupled to a 1:1 mix of hydrophilic:hydrophobic Sera-Mag SpeedBeads (Cytvia) as per the manufacturer’s instructions at a ratio of 1 μg of protein:10 μg of beads with incubation at room temperature for 8 min. Beads were washed three times with 80% ethanol, dried, and resuspended in 100 μl of 10% tetrafluoroethylene in 100 mM tris-HCl (pH 7.5). Proteins were then digested with the addition of 1:50 sequencing-grade trypsin (Sigma-Aldrich) and 1:50 LysC sequencing grade LysC (Wako, Japan) (micrograms of protease:micrograms of protein) overnight at 37°C and 600 rpm in a ThermoMixer. Digestion was stopped by acidification to 1% trifluoroacetic acid (TFA), and peptides were desalted using SDB-RPS (styrenedivinylbenzene reverse phase sulfonate) StageTips. Briefly, StageTips were equilibrated with 100% acetonitrile (ACN), 30% methanol/1% TFA, and 0.2% TFA. Peptides were added and then washed with 99% ethyl acetate and 1% TFA. Peptides were eluted with 5% ammonium hydroxide/80% ACN. Peptides were dried for 1 hour at 45°C in a Speedvac and stored at −30°C before liquid chromatography–tandem MS (LC-MS/MS) analysis.

DIA LC-MS/MS proteomics

Peptides were reconstituted in 3% ACN and 0.1% formic acid (FA). One microgram of peptide was separated on an in-house packed column (150 μm by 150 mm, 1.9-μm ReproSil Pur 120 C18, Dr. Maisch GmbH, Germany) with a Vanquish Neo Liquid Chromatography (Thermo Fisher Scientific) with mobile phases A [0.1% (v/v) FA] and B [80% (v/v) ACN and 0.1% (v/v) FA]. Peptides were eluted at a flowrate of 1.2 μl/min and a gradient of 5% B to 35% B in 35 min, 35% B to 60% B in 1 min, and 60% B to 98% B in 30 s and then washed with 95% B for 3 min. The LC was coupled to an Orbitrap Exploris 480 mass spectrometer (Thermo Fisher Scientific) with a spray voltage of 2600 V, a radio frequency lens of 30%, and an ion transfer tube heated to 320°C. The Orbitrap Exploris 480 was operated in data-independent acquisition (DIA) mode as previously reported (19, 102) comprising of a survey MS scan acquired as a profile with a maximum injection time of 54 ms and a standard automatic gain control target across a scan range from 350 to 1650 m/z with an Orbitrap resolution of 30,000. For MS/MS analysis, peptides were isolated with 34 variable width overlapping windows for a normalized automatic gain control target of a 1000% maximum isolation time of 30 ms (Table 3). These were fragmented with stepped higher-energy collisional dissociation collision energies of 25, 27, and 30 and analyzed across a mass range from 300 to 2000 m/z with an Orbitrap resolution of 15,000. Data files were searched with Spectronaut version 19.1 using a Mus musculus database (UniProt release 2024_04) and Trypsin and Lys-C/P enzymatic cleavage rules. The default settings were used for the remaining parameters. Raw data were normalized using log₂(x) transformation and median subtraction. To assess for differentially abundant candidate proteins between two conditions, filters for >1.25-fold change and a q value (adjusted P value) of <0.05 were applied. GSEA analysis was performed using fgsea (1.30.0).

Table 3. Quadrupole isolation windows used for DIA MS analysis.

Window number	Window start (m/z)	Window end (m/z)
1	350	375
2	374	397
3	396	413
4	412	428
5	427	442
6	441	455
7	454	468
8	467	480
9	479	492
10	491	505
11	504	518
12	517	532
13	531	546
14	545	560
15	559	574
16	573	589
17	588	604
18	603	620
19	619	636
20	635	653
21	652	670
22	669	688
23	687	707
24	706	728
25	727	749
26	748	776
27	775	803
28	802	834
29	833	865
30	864	904
31	903	963
32	962	1032
33	1031	1101
34	1100	1650

Cell culture

Primary KP^R172HC (TB32043) (11, 83, 84) and KP^floxC cancer cells (16, 35) were previously isolated from the tumors of end-stage KP^R172HC mice (Pdx1-Cre, LSL-Kras^G12D/+, LSL-Trp53^R172H/+) and KP^floxC mice (Pdx1-Cre, LSL-Kras^G12D/+, Trp53^flox/+) (16). Telomerase-immortalized fibroblasts (TIFs) were also generated previously (103). Both cancer cell lines and TIFs were cultured in Dulbecco’s modified Eagle’s medium (DMEM; high glucose, pyruvate, Gibco) supplemented with 10 mM Hepes (Thermo Fisher Scientific), 10% fetal bovine serum (HyClone), and 1% penicillin/streptomycin (Thermo Fisher Scientific) at 37°C with 20% O₂ and 5% CO₂. All cell lines were regularly tested for mycoplasma (all negative results).

Cell-derived matrices (CDMs)

CDMs were generated as described previously (85, 86). Before CDM generation, the glass surface was coated with 0.2% gelatin and allowed to set at 37°C for 2 hours before being rinsed twice with Dulbecco’s PBS and formalin fixed for 30 min at room temperature. After a further two washes with PBS, the fixed gelatin cross-links were quenched in 1 M sterile glycine for 30 min at room temperature, and coated plates were rinsed twice with PBS and once with DMEM before TIF cell seeding. To generate the CDMs, on day 0, 1.5 × 10⁵ TIFs were seeded per well in a glass-bottom 24-well plate (Corning). On days 1, 3, and 5, cells were treated with ascorbic acid (50 μg/ml) in DMEM supplemented with 10 mM Hepes, 10% fetal bovine serum, and 1% penicillin/streptomycin. On day 7, the CDMs were denuded, using an extraction buffer [0.5% (v/v) Triton X-100, 20 mM NH₄OH, and 0.01% sodium deoxycholate, made up in PBS], and rinsed twice with PBS, before addition of deoxyribonuclease I (10 μg/ml; Roche) in PBS containing calcium and magnesium, to digest DNA residue. This was followed by two PBS rinses before 4000 TB32043 KP^R172HC or 4000 KP^floxC cancer cells were seeded and allowed to adhere and grow for 16 hours. BIBO3304 (1 μM) or vehicle (water) was added to the cells before subsequent live cell imaging for 8 hours using a Leica Live Cell system with a Leica DMi6000 inverted microscope or an IncuCyte S3 Live-Cell Analysis Instrument (Sartorius). Tracking and quantification of cell migration were performed on binary images using a MATLAB plug-in (CellTracker) (104).

Intrasplenic injections

For intrasplenic experiments, 5 × 10⁵ TB32043 KP^R172HC cells were injected into the spleens of C57BL/6JAusb mice in 50 μl of Hanks Balanced Salt Solution as previously described (33, 90). Mice were first anesthetized using gaseous isoflurane, and a left subcostal incision was made through the skin. The cell suspension was then injected into the spleen during open laparotomy, followed by application of cyanoacrylate to the injection site to prevent hemorrhage from the spleen. The surgery site was closed using 5-mm resorbable vicryl sutures on the peritoneal wall and clips for the skin, and mice were treated with analgesia bupivacaine (topically, 8 mg/kg) and buprenorphine (subcutaneously, 0.075 mg/kg). BIBO3304 (1 mg/kg) or vehicle (saline) was administered daily via intraperitoneal injection (day 0 to day 11). Mice were euthanized either on day 5 or on day 12, and the number of visible metastases in the liver was quantified by first separating the liver into its various lobes and then counting the number of metastases on the front and back on the lobes. Livers were subsequently fixed in formalin, dehydrated, and paraffin embedded for histological analysis. Animals with surgical complications were euthanized immediately and excluded from the study.

Supplementary Materials

The PDF file includes:

Figs. S1 to S9

Legends for tables S1 and S2

Legends for movies S1 and S2

Consortium members of the Australian Pancreatic Cancer Genome Initiative (APGI)

Consortium members of the Avner Australian Pancreatic Cancer Matrix Atlas (APMA)

Other Supplementary Material for this manuscript includes the following:

Tables S1 and S2

Movies S1 and S2