Original Article

Synergistic action of Smad4 and Pten in suppressing pancreatic ductal adenocarcinoma formation in mice

  • Oncogene volume 29, pages 674686 (04 February 2010)
  • doi:10.1038/onc.2009.375
  • Download Citation


Mutations of SMAD4/DPC4 are found in about 60% of human invasive pancreatic ductal adenocarcinomas (PDACs); yet, the manner in which SMAD4 deficiency enhances tumorigenesis remains elusive. Using a Cre-LoxP approach, we generated a mutant mouse carrying a targeted deletion of Smad4 in the pancreas. We showed that the absence of Smad4 alone did not trigger pancreas tumor formation; however, it increased the expression of an inactivated form of Pten, suggesting a role of Pten in preventing Smad4–/– cells from undergoing malignancy. To investigate this, we disrupted both Pten and Smad4. We showed that Pten deficiency initiated widespread premalignant lesions, and a low tumor incidence that was significantly accelerated by Smad4-deficiency. The absence of Smad4 in a Pten-mutant background enhanced cell proliferation and triggered transdifferentiation from acinar, centroacinar and islet cells, accompanied by activation of Notch1 signaling. We showed that all tumors developed in the Smad4/Pten-mutant pancreas exhibited high levels of pAKT and mTOR, and that about 50 and 83% of human pancreatic cancers examined showed increased pAKT and pmTOR, respectively. Besides the similarity in gene expression, the pAKT and/or pmTOR-positive human PDACs and mouse pancreatic tumors also shared some histopathological similarities. These observations indicate that Smad4/Pten-mutant mice mimic the tumor progression of human pancreatic cancers that are driven by activation of the AKT–mTOR pathway, and uncovered a synergistic action of Smad4 and Pten in repressing pancreatic tumorigenesis.


The pancreas suffers from two important diseases, diabetes and pancreatic cancer. Diabetes affects at least 30 million people worldwide and pancreatic ductal adenocarcinoma (PDAC) is the fourth leading cause of cancer-related death in Western societies (Deramaudt and Rustgi, 2005; Hezel et al., 2006; Maitra and Hruban, 2008). PDAC is diagnosed in approximately 1 in 10 000 people in the United States, with a 3–5% 5-year survival rate after diagnosis (Warshaw and Fernandez-del Castillo, 1992; Hezel et al., 2006; Maitra and Hruban, 2008). Poor survival rate is primarily because of the lack of early detection and frequent metastasis of primary tumors into lymph nodes and surrounding organs, such as the liver and stomach (Yeo et al., 2002; Farnell et al., 2005; Yachida and Iacobuzio-Donahue, 2009).

PDAC develops through a defined series of precursor stages before advancing to a more aggressive and devastating cancer, accompanied by signature mutations of a number of tumor-suppressor genes and oncogenes (Bardeesy and DePinho, 2002; Maitra et al., 2006; Maitra and Hruban, 2008). Mutations are frequently found in Trp53 (75%), SMAD4/DPC4 (60%) and BRCA2 (7 to 19%) (Hahn et al., 1996a, 2003; Rozenblum et al., 1997). Loss of function due to mutation, deletion or promoter hypermethylation also occurs in p16 (90%) and PTEN (60%) (Asano et al., 2004; Attri et al., 2005). Activation mutations of Kras occur early and are found in more than 90% of invasive cancers (Almoguera et al., 1988; Rozenblum et al., 1997). Activation of AKT, a downstream target of PTEN, through its phosphorylation (pAKT-Ser473) or overexpression, is found in about 20–70% of PDAC cases analysed in different patient populations (Altomare et al., 2003; Semba et al., 2003; Michl and Downward, 2005). Although some of these alterations seem to occur in a temporal sequence in progressive stages (Bardeesy and DePinho, 2002), the specific pathogenic roles and cellular events associated with these mutations are poorly understood.

A dozen animal models have been generated that carry genetically engineered signature mutations found in human pancreatic cancer (reviewed in Hruban et al., 2006). An analysis of these mutant mice has contributed significantly to our understanding of the functions of individual disease-related genes, and their genetic interactions during cancer initiation and progression through precursor stages to invasive carcinomas (Bardeesy and DePinho, 2002). It has been shown that activation mutation of Kras (KrasG12D) is sufficient to initiate pancreatic tumorigenesis, although most tumors are limited primarily to the pancreatic intraepithelial neoplasia (PanINs) stages (Aguirre et al., 2003; Hingorani et al., 2003). On the other hand, the sole inactivation of p16 (Ink4a), p19 (Arf) or p53 fails to produce neoplastic lesions in the pancreas (Bardeesy et al., 2002; Aguirre et al., 2003). Notably, a combination of KrasG12D expression with loss-of-function mutations in Ink4a, Arf, p53, type II TGF-β receptor or Smad4 results in an earlier appearance of PanIN lesions. These lesions progress rapidly to highly invasive and metastatic cancers (Bardeesy et al., 2002; Aguirre et al., 2003; Ijichi et al., 2006; Izeradjene et al., 2007). Similar cooperative roles in pancreatic ductal cancer progression has also been established between Ink4a/Arf and/or p53 and oncogenic signaling of transforming growth factor-α (TGF-α) (Bardeesy et al., 2002). Recently, tumor-suppressor Pten was disrupted in the pancreas using the Cre-LoxP-mediated approach (Stanger et al., 2005). Pten-mutant mice displayed a progressive replacement of the acinar pancreas with highly proliferative ductal structures, primarily because of the expansion of centroacinar cells rather than transdifferentiation of acinar cells. Although the absence of Pten quickly triggers these precancerous lesions, only a small fraction of mutant mice developed pancreatic ductal carcinoma. These data suggest that other collaborating factors are required for full malignant transformation, although mutations of other common signature mutations, such as p53, p19, p16 and Kras, have not been detected (Stanger et al., 2005).

SMAD4 serves as a common mediator of TGF-β signals and has important functions in many biological processes (Heldin et al., 1997; Massague, 1998; Weinstein et al., 2000; Weinstein and Deng, 2006). Mutations of SMAD4 have also been detected in pancreas cancer, colon cancer, cholangiocarcinoma cancer and gastric adenocarcinomas (Tamura et al., 1996; Hahn et al., 1996a, 1996b; Howe et al., 1998; Friedl et al., 1999; Kang et al., 2002; Suto et al., 2002), suggesting an important tumor-suppressor role of SMAD4 in these tissues/organs. In mice, loss of Smad4 results in early embryonic lethality because of impaired extraembryonic membrane formation and decreased epiblast proliferation (Sirard et al., 1998; Yang et al., 1998), whereas Smad4 heterozygous mice developed gastric polyposis and cancer because of haploinsufficiency (Takaku et al., 1999; Xu et al., 2000). Using tissue-specific knockout of Smad4 (Yang et al., 2002), we, along with others, showed that Smad4 deficiency could cause tumor formation in mammary tissue (Li et al., 2003), skin (Yang et al., 2005; Qiao et al., 2006), liver (Xu et al., 2006), forestomach (Teng et al., 2006) and colon (Kim et al., 2006). A recent study revealed that loss of Smad4 alone does not initiate pancreatic cancer formation, although it accelerates tumor formation on activation of oncogenic signaling, such as Kras (Izeradjene et al., 2007). However, the role of Smad4 in pancreatic tumorigenesis and its potential interaction with other tumor suppressor/oncogenes has not been determined.

In this study, we performed pancreas-specific knockout of Smad4 and/or Pten in mice to assess their collaborative activities in pancreas cancer formation. Our data revealed that absence of Smad4 and Pten synergistically induced transdifferentiation of all three major cell lineages in the pancreas and activation of Notch signaling that may be responsible for tumorigenesis. Further data indicated that all tumors developed in the Smad4/Pten-mutant pancreas showed high levels of pAKT and pmTOR, and exhibited some similarities with human PDACs that carry high levels of pAKT and pmTOR.


Targeted disruption of Smad4 and/or Pten in mouse pancreas

Pancreas-specific deletion of Smad4 and/or Pten was achieved by crossing Smad4 conditional (Yang et al., 2002) and/or Pten conditional (Groszer et al., 2001) mutant mice with Pdx1-Cre mice (Gu et al., 2002). The Pdx1 gene is highly expressed throughout the entire pancreatic epithelium, including both endocrine and exocrine cells, as well as subpopulations of duodenal and gastric enteroendocrine cells in the early stages of embryogenesis. In adults, Pdx1 is most abundant in insulin-producing-β cells and at a lower level in acinar cells, but is not detected in duct cells. Using Rosa26 mouse as a reporter (Soriano, 1999), we detected Cre-mediated recombination in all three types of pancreatic tissues: exocrine, endocrine and duct (Supplementary Figures 1a–d) tissues. This expression pattern is consistent with a previous study showing that cells expressing Pdx1 in early embryonic developmental stages give rise to all three types of pancreatic tissues in postnatal animals (Gannon et al., 2000; Gu et al., 2002). PCR analysis detected Cre-mediated deletion of Smad4 and Pten in pancreas and duodenum, but not in a number of other organs/tissues isolated from Smad4Co/Co;Pdx1-Cre and PtenCo/Co;Pdx1-Cre mice (Supplementary Figure 1e).

Absence of Smad4 accelerates pancreas ductal abnormalities and carcinoma initiated by Pten deficiency

Next, we studied Smad4Co/Co;Pdx1-Cre, PtenCo/Co;Pdx1-Cre and Smad4Co/Co;PtenCo/Co;Pdx1-Cre mice for possible pancreatic lesions and tumorigenesis. The pancreas of wild-type and Smad4Co/Co;Pdx1-Cre mice seemed normal during a study period of 14 months, except for one Smad4Co/Co;Pdx1-Cre animal that exhibited ductal hyperplasia (Table 1). In contrast, all PtenCo/Co;Pdx1-Cre mice exhibited an increased size of the pancreas, primarily because of increased number and size of islets than that of control mice (Figures 1a–c, Supplementary Figure 2a–d). The increased number and size of islets were associated with Ki67-positive cells (Supplementary Figures 2e and f), suggesting that they were caused by increased cell proliferation. The pancreas of Smad4Co/Co;PtenCo/Co;Pdx-Cre mice had similar increases in the number and size of islets compared with that of PtenCo/Co;Pdx-Cre mice (Supplementary Figure 1a–d). However, their pancreas was bigger than that of PtenCo/Co;Pdx-Cre mice because of more severe ductal abnormalities and tumor formation (Figure 1d, and Table 1). These observations indicate that the absence of Smad4 significantly increased pancreas premalignant lesions and tumor incidence in Smad4Co/Co;PtenCo/Co;Pdx-Cre mice.

Table 1: Summary of phenotypesa of mice with various genotypes
Figure 1
Figure 1

Absence of Smad4 and Pten synergistically promotes pancreas ductal abnormalities and carcinoma. (a–d) Whole mount view of the entire pancreas of 6-month-old mice with different genotypes (wild type: WT, Smad4Co/Co;Pdx1-Cre: Sm4-/-, PtenCo/Co;Pdx1-Cre: Pt-/-, and Smad4Co/Co;PtenCo/Co;Pdx1-Cre: Sm4-/-Pten-/-). Arrows point to nodules. (e–h) Histological sections of the pancreas of 2-month-old mice. Arrows point to islets. (i–k, n) Histological sections of the pancreas of 4-month-old PtenCo/Co;Pdx1-Cre (i) and Smad4Co/Co;PtenCo/Co;Pdx1-Cre (j, k, n) mice. (l, m) Immunohistochemical staining of the pancreas sections of 4-month-old Smad4Co/Co;PtenCo/Co;Pdx1-Cre mice using antibodies to CK-19 and Mucin 5. (o) Percentage of pancreas to body weight of 6- to 8-month-old mice with different genotypes. At least six mice for each genotype were used. Standard deviation is not included for Smad4Co/Co;PtenCo/Co;Pdx1-Cre mice because the variation is too large due to significantly different sizes of tumors in their pancreases. *represents significant difference (P<0.05) by Student's t-test. Magnification: × 3 for A–D, × 100 for (e–h), × 50 for (i, j), × 200 for (k–n).

Next, we performed a histological analysis of the pancreas of mice of various genotypes. Although the pancreas of wild-type and Smad4Co/Co;Pdx-Cre mice seemed normal (Figures 1e and f), one out of five PtenCo/Co;Pdx1-Cre mice started to show premalignant lesions at 1 month of age, and more animals exhibited abnormalities in their pancreas as they became older (Figure 1g and Table 1). In contrast, two of three Smad4Co/Co;PtenCo/Co;Pdx-Cre mice at 15 days of age showed multifocal abnormal ductal structures (Table 1). At 2 months of age, all five mice that were examined developed multifocal abnormal ductal structures and tumor foci (Figure 1h, and Table 1). The ductal structures gradually increased in quantity and size (Figures 1i and j). They underwent malignant transformation (Figure 1k) and spread throughout the pancreas (Figure 1d, data not shown). The epithelial nature of these ducts was confirmed by an antibody, CK19, which specifically marks duct epithelium (Figure 1l), and by an antibody to Mucin 5, a marker for early stages of ductal carcinoma (Figure 1m). PDAC was eventually developed in 3 out of 12 PtenCo/Co;Pdx-Cre mice and in 18 out of 19 Smad4Co/Co;PtenCo/Co;Pdx-Cre mice examined between 4 and 8 months of age (Figure 1n, Table 1). This difference in the tumor formation profile is quite impressive, and suggests a synergistic action between Smad4 and Pten deficiency in PADC formation. Furthermore, we found that the pancreas of Smad4Co/Co;PtenCo/Co;Pdx-Cre mice contained a much wider spread of tumor foci and more solid tumor masses. This is also reflected by the much higher ratio of pancreas to body weight of Smad4Co/Co;PtenCo/Co;Pdx-Cre mice in comparison with that of PtenCo/Co;Pdx-Cre mice (Figure 1o). On the basis of these data, we conclude that the absence of Smad4 significantly accelerates pancreas ductal abnormalities and PADCs initiated by Pten deficiency.

An H&E section revealed that, in at least five mice, tumor cells invaded into several other organs around the pancreas, including liver, guts and/or stomach (Table 1). We showed earlier that Cre-mediated deletion of Smad4 and Pten was detected in the duodenum (Supplementary Figure 1g). Consistent with this, Smad4Co/Co;Pdx-Cre and Smad4Co/Co;PtenCo/Co;Pdx-Cre mice showed tumor formation in the duodenum. This lesion is not a focus of this study, but will be investigated separately.

Absence of Smad4 increases cell proliferation in Pten-deficient pancreas

Next, we studied cell proliferation and death in animals with different phenotypes. Our data indicated that loss of Smad4 does not affect cell proliferation compared with wild-type controls (Figures 2a, b and e). In contrast, the pancreas of PtenCo/Co;Pdx-Cre mice showed an increased (yet not statistically significant) number of BrdU-positive acinar cells compared with controls when examined at P15, despite the lack of histological abnormalities at this time point (Figures 2c and e). Loss of Smad4 further increased the proliferation of Pten-deficient cells (Figures 2d and e). We next examined cell death using TUNEL assay. We detected a graded increase in apoptosis from wild-type, Smad4Co/Co;Pdx-Cre, PtenCo/Co;Pdx-Cre to Smad4Co/Co;PtenCo/Co;Pdx-Cre mice (Figure 2f–h). But such increases did not reach a statistically significant level. Thus, the absence of Smad4 increased cell proliferation in Pten-deficient pancreas without causing a further increase in cell death.

Figure 2
Figure 2

Genetic interaction between Smad4 and Pten. (a–e) Immunohistochemical staining of the pancreas of P15 mice showing BrdU-positive cells, mostly in the (a–d), and percentage of these cells in mice with different genotypes (e). Majority BrdU-positive cells are acinar cells, whereas ductal cells were occasionally observed (arrow in a). (f, g, h) TUNEL assay of 2-month-old wild-type (f) and PtenCo/Co;Pdx1-Cre (g) mice, and percentage of apoptotic cells in mice with different genotypes (h). (i–m) Immunohistochemical staining of the pancreas of 4-month-old wild-type (i, k) and PtenCo/Co;Pdx1-Cre (j, l, m) mice using an antibody against BMP4 (i, j) and pSmad1/5/8 (k, l, m). (n, o) Immunohistochemical staining of the pancreas of 4-month-old wild-type (n) and Smad4Co/Co;Pdx1-Cre (o) mice. Multiple sections from three to five mice were used for each analysis. (p) Western blot analysis showing levels of Smad4 and Pten in wild-type, Smad4Co/Co;Pdx1-Cre, and Smad4Co/Co;PtenCo/Co;Pdx1-Cre mice. *represents significant difference (P<0.05) by Student's t-test. Magnification: × 300 for (a–d) and (f–i), × 200 for (j), × 300 for (k–m) and × 150 for (n, o).

Genetic interactions between Smad4 and Pten during pancreatic cancer formation

To investigate the mechanism underlying the synergistic effect between Smad4 and Pten in repressing pancreas tumor formation, we examined the expression levels of a number of genes in the Pten and Smad4 pathways. We detected increased levels of Bmp4 (Figures 2i and j), pSMAD1/5/8 (Figures 2k–m) and Smad4 (Figure 2p) in the pancreas of PtenCo/Co;Pdx-Cre mice compared with that of control mice. These data indicate that the deletion of Pten triggers an activation of the BMP-Smad1/5/8-Smad4 pathway. On the other hand, these data also revealed increased levels of Pten protein in the pancreas of Smad4Co/Co;Pdx-Cre mice compared with control mice (Figures 2o and p).

We showed previously that Smad4 and Pten regulate each other through a novel feedback mechanism in the liver and that the absence of Smad4 increases Pten transcription (Xu et al., 2006). In the pancreas, it was shown that overexpression of TGFβ1 reduces Pten expression (Ebert et al., 2002). Thus, it is conceivable that the impaired TGF-β signaling due to the absence of Smad4 in the pancreas could trigger an upregulation of Pten. Interestingly, levels of phosphorylated Pten, which is an inactivated form of the protein, were also increased in the pancreas of Smad4Co/Co;Pdx-Cre mice (Figures 3a–d). The increased levels of total Pten and pPten suggest that Pten may serve as a barrier in preventing Smad4-deficient cells from malignant transformation, which may account for the absence of tumor formation in Smad4Co/Co;Pdx-Cre mice.

Figure 3
Figure 3

Molecular alterations associated with tumorigenesis in the pancreas of PtenCo/Co;Pdx1-Cre and Smad4Co/Co;PtenCo/Co;Pdx1-Cre mice. Immunohistochemical staining using antibodies to pPten (a–d), pAKT (e–h), pmTOR (i–l), cyclin D1 (m–p) and c-Myc (q–t). Genotypes are indicated at the top of the figure. Arrows point to islets. Three to five mice were used for each analysis. Magnification: × 150 for (a–t).

We next checked the expression of genes downstream of the Pten pathway. We showed that pAKT was undetectable in wild-type and Smad4 mutant pancreas, and it was highly expressed in all PDACs derived from both PtenCo/Co;Pdx-Cre and PtenCo/Co;Smad4Co/Co;Pdx-Cre mice (Figures 3e–h). A strong expression of the phosphorylated form of the mammalian target of rapamycin (mTOR) was also observed (Figure 3i–l). AKT and mTOR signals promote cell cycle progression, cell growth and proliferation through triggering the activation of some oncoproteins. To characterize this, we used an antibody to cyclin D1 that is overexpressed in numerous types of cancers. Our data indicated that cyclin D1 was undetectable in the pancreas of wild-type and Smad4 mutant mice (Figures 3m and n). Cyclin D1 was also present at low levels in premalignant ducts (left, Figure 3o), and at high levels in the cancers of PtenCo/Co;Pdx-Cre (right, Figure 3o) and PtenCo/Co;Smad4Co/Co;Pdx-Cre (Figure 3p) mice. A similar expression pattern was also detected for c-Myc (Figures 3q–t). These data revealed that tumors developed in PtenCo/Co;Pdx-Cre and PtenCo/Co;Smad4Co/Co;Pdx-Cre mice showed the activation of AKT–mTOR signaling and a number of additional oncogenes.

Transdifferentiation from acinar and islet cells in pancreas tumors

It was previously shown that tumors developed in Pten pancreas-deficient mice were originated from amplification of centroacinar cells. Using the lectin dolichos biflorus agglutinin (DBA), a marker of ductal cells and centroacinar cells, we observed an increased number of DBA-positive cells in the Pten-deficient pancreas than in wild type and Smad4 mutant mice (Figures 4a and b). This observation is consistent with a previous finding that the absence of Pten causes amplification of this population of cells, which develop into abnormal ductal cells through metaplasia (Stanger et al., 2005). Notably, we found that in the pancreas of Smad4 and Pten double mutant mice, the number of centroacinar cells significantly increased (Figure 4c). This is consistent with the synergistic activity of these genes in pancreas tumorigenesis.

Figure 4
Figure 4

Molecular alterations associated with transdifferentiation in the pancreas of PtenCo/Co;Pdx1-Cre and Smad4Co/Co;PtenCo/Co;Pdx1-Cre mice. (a–c) Staining using DBA in the pancreas of wild-type (a), PtenCo/Co;Pdx1-Cre (b) and Smad4Co/Co;PtenCo/Co;Pdx1-Cre (c) mice. (d, e) Histological sections of the pancreas of Smad4Co/Co;PtenCo/Co;Pdx1-Cre mice. (f–n) Staining of the pancreas of Smad4Co/Co;PtenCo/Co;Pdx1-Cre mice with antibodies for CK19 (f, g, i), amylase (h, i), Pdx-1 (j–l), Glut2 (m) and insulin (n). Three to five mice were used for each analysis. Magnification: × 300 for (a–c), (d) (left), (e–k), (m) (left) and (n). × 200 for (m) (right).

Our histological analysis indicated that many newly formed small tubes maintained mixed acinar and ductular cells (Figures 4d and e). We also found that some ducts emerged from the islet cells (Figure 4f) and were positive for CK19 (Figure 4j). These phenotypes were observed in the pancreas of Pten mice but they were much more severe in double mice. These data suggest that some of these abnormal duct structures may be produced from transdifferentiation from acini and/or islet cells.

To provide molecular evidence, we performed lineage marker analysis. We first used an antibody to amylase (Figure 4g), which was produced by acinar cells specifically, in combination with an antibody to CK19, which marks ductal cells (Figure 4h). Our data detected many double-positive cells in the ducts (Figure 4i). The expression of amylase was transit and disappeared as the tubular structures enlarged. The amylase and CK19 double staining may represent a stage of the transdifferentiation process from acinar cells to ductal cells.

Next, we performed immunohistochemical staining for ducts that emerged from the islet, using an antibody to Pdx-1 (Figure 4j). We showed that many newly formed duct cells were highly positive for Pdx-1 (Figure 4k). We also found some cells that were still in the morphological transition state (arrow, Figure 4l). Because Pdx-1 is not expressed in duct cells in postnatal mice (Gu et al., 2002), the positive staining for Pdx-1 suggests that these duct cells originated from islet cells. This notion was confirmed by using two additional markers for islet cells, glucose transporter type 2 (Glut-2) and insulin. Both Glut-2 and insulin are normally expressed in mature β-cells (Figures 4m and n), and are associated with early endocrine cell development. In both cases, the proteins were detected in some premalignant ducts, and their levels were significantly lower than those in normal islet cells and they also became nuclear localized, suggesting that these duct cells were in a transition state and no longer produced normal amounts of these proteins.

Activation of Notch1 signaling is associated with transdifferentiation

Next, we investigated possible reasons for transdifferentiation. It has been shown that Notch signaling has an important role in differentiation of all cell lineages in the pancreas, as well as in tumorigenesis (Hald et al., 2003; Murtaugh et al., 2003; Roy et al., 2007; Sawey et al., 2007). Using an antibody to the activated form of Notch1, which localizes to the nucleus on proteolytic cleavage of the membrane-bound inactive form (Schroeter et al., 1998), we showed that the activated Notch1 was undetectable in the pancreas of wild-type and Smad4Co/Co;Pdx-Cre mice (Figure 5a). In contrast, clusters of Notch1-positive cells were detected in the nucleus of the acini of both PtenCo/Co;Pdx-Cre mice and Smad4Co/Co;PtenCo/Co;Pdx-Cre mice before the morphological appearance of abnormal ducts (Figure 5b). Notch1-positive cells became more abundant during the transition process from acini to ducts (Figure 5c), and they completely populated all ducts (Figure 5d), including enlarged cyst-like structures (Figure 5e). Because Notch1-positive cells were not observed in the pancreas of Smad4Co/Co;Pdx-Cre mice, we believe that Smad4 deficiency does not trigger the activation of Notch1. However, absence of Pten triggered Notch1 expression, which facilitates the amplification of Notch1-positive cells in the pancreas of Smad4Co/Co;PtenCo/Co;Pdx-Cre mice.

Figure 5
Figure 5

Activation of Notch1 signaling in the pancreas of PtenCo/Co;Pdx1-Cre and Smad4Co/Co;PtenCo/Co;Pdx1-Cre mice. (a–e) Immunohistochemical staining of the pancreas of wild-type (a) and Smad4Co/Co;PtenCo/Co;Pdx1-Cre (b–e) mice using an antibody to activated Notch1. Arrows point to Notch1-positive acinar (b) and ductal (c) cells. (f, g) Immunohistochemical staining of the pancreas of Smad4Co/Co;PtenCo/Co;Pdx1-Cre mice using an antibody to Hes-1. The boxed area is amplified in (g). Three to five mice were used for each analysis. Magnification: × 400 for (a) and (b), × 500 for (c), × 200 for (d–f).

Next we checked expression of Hes-1, a downstream target of activated Notch signaling (Jarriault et al., 1998). We observed a similar pattern of Hes-1 expression in abnormal duct structures and tumors (Figures 5f and g), confirming that the Notch 1 pathway is activated. It was shown that expression of activated Notch1 in Pdx1-expressing progenitor cells caused an accumulation of ductal structures that were completely devoid of acinar cells (Hald et al., 2003). Thus, the activated Notch signaling could have a significant role in the transdifferentiation observed in PtenCo/Co;Pdx-Cre and Smad4Co/Co;PtenCo/Co;Pdx-Cre mice.

Human pancreatic cancer displayed activated AKT signaling

The PDAC developed in both Pten and Smad4/Pten double mutant mice exhibits high levels of pAKT and pmTOR (Figure 3), suggesting that activation of the AKT–mTOR pathway serves as a driving force for tumorigenesis and progression. To investigate whether this pathway is activated in human sporadic PDACs, we checked human pancreatic cancers using antibodies to activated AKT and mTOR. We found that 13 of 26 human PDACs expressed pAKT in the ducts (Figure 6a). pAKTs were also detected in acinar, islet and stroma cells (Figures 6b–d and j). We next checked the expression of pmTOR, and detected its expression in 55% (16 of 29) of ductal carcinoma examined (Figure 6e). pmTOR was also detected in acinar cells of 34% (10 of 29) tumors (Figure 6f). Interestingly, pmTOR was also detected in centroacinar cells in at least three tumors (Figures 6g and h).

Figure 6
Figure 6

Expression of pAKT and pmTOR in human sporadic pancreatic cancers. (a–h) Immunohistochemical staining of human PDACs (provided by Dr Tadashi Yoshino, Okayama University Graduate School of Medicine and Dentistry, Okayama, Japan) with antibodies to pAKT (a–d) and pmTOR (e–h). pAKT is detected in ducts (a), acinar (b), islets (c) and stroma (d) cells. pmTOR is detected in ducts (e), acini (f) and centroacinar cells (arrows, g, h). (i) Summary of staining of pAKT and pmTOR in human cancers. (j) Histopathological comparison between mouse and human PDACs. Magnification: × 200 for (a–h) and × 150 for (j).

Besides the similarity in gene expression, the pAKT and/or mTOR-positive human PDACs and tumors developed in our mutant mice also displayed similarities in histopathology (Figure 6j). Furthermore, histological analysis also revealed a similar histopathology between many mouse tumors and human tumors during tumor initiation and progression (Supplementary Figure 3). These observations indicate that our mutant mice mimic the tumor progression of human pancreatic cancers that are driven by activation of the AKT–mTOR pathway.


Pancreas tumorigenesis occurs through multiple stages of progression, involving genetic alterations of many genes (Bardeesy and DePinho, 2002; Hruban et al., 2006; Maitra et al., 2006; Maitra and Hruban, 2008; Yachida and Iacobuzio-Donahue, 2009). Mutations of SMAD4, which are found in approximately 60% of human pancreas cancers, are suspected to be late events during tumor progression (Hahn et al., 1996a, 1996b; Bardeesy and DePinho, 2002). In this study, we showed that the mutation of mouse Smad4 alone does not cause cancer formation. However, this mutation does accelerate tumor formation in Pten-mutant mice. This observation uncovers a synergistic role of Smad4 and Pten in pancreas cancer. We further showed that the absence of Smad4 in a Pten-deficient background enhanced cell proliferation and induced transdifferentiation. This enables tumor formation from multiple cell lineages and allows for tumorigenesis to escape the inhibition effects of the host defense system.

It was shown that overexpression of TGFβ1 in the pancreas and in transgenic mice reduced Pten expression (Ebert et al., 2002). Interestingly, we show here that Smad4 deficiency causes an upregulation of the Pten total protein and of the inactivated protein (pPten). These data suggest that Smad4-mediated TGFβ signaling inhibits Pten expression, therefore, the loss of Smad4 triggers an increased expression of Pten. Meanwhile, high levels of Pten may inhibit cell proliferation; therefore, it is conceivable that the increased level of Pten triggers a feedback response to reduce the active form of Pten protein to maintain a balance. Consequently, Smad4 deficiency alone is not sufficient enough for causing pancreatic tumor formation.

In contrast, these data showed the absence of Pten-affected gene expression in the Smad4 pathway, including upregulation of Bmp4 and Smad4 expression. These findings reveal a reciprocal regulation of Smad4 and Pten. Smad4 is a well-established tumor suppressor and inhibitor of tumor formation in multiple tissues, including pancreas, colon, liver, skin, mammary gland and stomach (reviewed in (Weinstein et al., 2000; Weinstein and Deng, 2006)). Because Pten loss results in widespread premalignant lesions, many of which do not advance to fully malignant tumors, it is possible that the increased Bmp4-Smad1/Smad5/Smad8-Smad4 signaling is a mechanism to counterattack the malignant effect of Pten deficiency, thereby inhibiting tumorigenesis. Our observation that deletion of Smad4 significantly enhances tumor formation indicates that this may indeed be the case. A previous investigation revealed that overexpression of Smad4 inhibited the proliferation of a pancreatic cancer cell line in vitro and reduced xenografted tumorigenesis through induction of cell cycle inhibitors p15ink4b and p21 (Peng et al., 2002). We detected increased proliferation in Smad4 and Pten double mutant pancreas compared with Pten-mutant pancreas at premalignant stages, highlighting an important cooperative effect of Smad4 in promoting cell proliferation in Pten-deficient cells.

It has been shown that pancreas-specific deletion of Pten resulted in ductal metaplasia because of the expansion of centroacinar cells, leading to PDAC formation in a small fraction of animals studied (Stanger et al., 2005). A similar expansion of centroacinar cells is observed in our Smad4Co/Co;PtenCo/Co;Pdx-Cre mice (Figure 4c). In addition, we observed extensive transdifferentiation in Smad4Co/Co;PtenCo/Co;Pdx-Cre mice compared with Pten alone, involving multiple cell origins: acinar, islets and ducts. Previous studies indicated that activation of oncogenes, such as Kras, could induce pancreatic ‘ductal’ neoplasia from nonductal exocrine cells (Guerra et al., 2007). Our data provide evidence that a similar phenomenon can also be caused by inactivation of tumor suppressors. Because there is no effective way to treat PDAC, a better characterization of the cellular origin of PDAC should provide useful information regarding tumor initiation and early diagnosis, leading to the development of better therapeutic treatment.

Our data revealed that tumorigenesis in Smad4Co/Co;PtenCo/Co;Pdx-Cre animals is associated with the activation of Notch signaling. The activation of Notch signaling was reported to result in the transdifferentiation of acinar cells to ducts (Hald et al., 2003; Sawey et al., 2007). Consistent with the activation of Notch signaling, we detected an increased expression of Hes-1, a downstream target of activated Notch signaling (Jarriault et al., 1998). Indeed, our further analysis indicates that expression of many genes that are involved in various aspects of the Notch-signaling pathway is significantly increased in Smad4Co/Co;PtenCo/Co;Pdx-Cre mice than in PtenCo/Co;Pdx-Cre mice (X Xu and CX Deng, unpublished observation). However, it is currently unclear how the absence of Smad4 and Pten causes the activation of the Notch-signaling pathway. This interesting issue deserves further investigation.

Activation of AKT occurs in about 20–70% of human PDACs analysed in different populations (Altomare et al., 2003; Semba et al., 2003; Michl and Downward, 2005). Among about 30 sporadic human PDACs from a Japanese population, 50 and 83% showed increased pAKT and pmTOR. These data highlight an important role of PTEN/AKT/mTOR signaling in PDAC formation. These data also provide useful information for the development of an effective therapeutic regime for human PDACs. Treatment of human PDACs with inhibitors of mTOR, such as rapamycin, alone or in combination with different drugs such as those against activated Notch signaling, may be an effective approach for killing AKT/mTOR-positive cancer cells. Our Smad4/Pten-mutant mouse model can be used for testing various therapeutic treatment options in the near future.

Materials and methods


Mice carrying Smad4 (Yang et al., 2002) and Pten (Groszer et al., 2001) conditional knockout alleles were crossed with a Pdx-Cre transgenic mouse (Gu et al., 2002) to generate cohorts of Smad4Co/Co;Pdx-Cre, PtenCo/Co;Pdx-Cre and Smad4Co/Co;PtenCo/Co;Pdx-Cre mice. The average genetic background of these mice is 25% 129; 50% FVB; 25% Black Swiss. Genotyping of Smad4 conditional and Pten conditional mice was carried out as described (Groszer et al., 2001; Yang et al., 2002). Pdx-Cre transgenic (Gu et al., 2002) and Rosa26-β-gal reporter (R26R) (Soriano, 1999) mice were genotyped as described (Xu et al., 2006). All experiments were approved by the Animal Care and Use Committee of the National Institute of Diabetes, Digestive and Kidney Diseases (ACUC, NIDDK).

Staining of pancreas tissues and cancer with X-gal

For X-gal staining, pieces of pancreas tissues were dissected and placed into OCT-freezing medium. The samples were gradually frozen in dry ice and then transferred to −80 °C for permanent storage. Sections of 8-μm thickness were prepared using a cryostat and stained as described previously (Xu et al., 2006).

Histology, immunohistochemical staining and western blot analysis

Tissue was fixed in 10% neutral-buffered formalin (Sigma) at 4 °C overnight, dehydrated through a graded alcohol series of xylene and paraffin, and then embedded in paraffin. Sections of 5 μm thickness were prepared for H&E and antibody staining using regular procedures. The following antibodies were used for immunohistochemical staining: CK-19 (Dako, Glostrup, Denmark), c-Myc, SMAD4 and cyclin D1 (Santa Cruz, Santa Cruz, CA, USA), BrdU (BrdU staining kit, Zymed Laboratories Inc., South San Francisco, CA, USA), PTEN, AKT, phospho-AKT (Ser 473), phospho-Pten, Notch1, SK6 and phospho-mTOR (Ser2448) (Cell Signaling Inc., Danvers, MA, USA), insulin (Sigma, St Louis, MO, USA), Bmp4 and Glut2 (Chemicon, Billerica, MA, USA), Pdx-1 (Upstate, Lake Placid, NY, USA), Mucin 5 (Nococastra, Newcastle upon Tyne, UK). DBA staining was performed on the basis of the manufacturer's protocol (Zymed Laboratories Inc.).

Taqman–PCR analyses and quantitative RT–PCR (qRT–PCR)

Total RNA was extracted from cells using RNA STAT-60 (Tel-Test Inc., Friendswood, TX, USA). QRT–PCR was achieved by combining in vitro reverse transcription with quantitative PCR, which was performed in a Stratagene Mx3000P thermal cycler (Stratagene, La Jolla, CA, USA). Briefly, 5 μg of purified total RNA from each sample was treated with DNAase (Ambion, Austin, TX, USA) and was then subjected to a reverse-transcription reaction using AMV reverse transcriptase (Roche, Indianapolis, MN, USA). The real-time PCR was performed using an ABI PRISM 7000 Sequence Detection System (Applied Biosystems, Foster City, CA, USA).

Conflict of interest

The authors declare no conflict of interest.


  1. , , , , , et al. (2003). Activated Kras and Ink4a/Arf deficiency cooperate to produce metastatic pancreatic ductal adenocarcinoma. Genes Dev 17: 3112–3126.

  2. , , , , , . (1988). Most human carcinomas of the exocrine pancreas contain mutant c-K-ras genes. Cell 53: 549–554.

  3. , , , , , et al. (2003). Frequent activation of AKT2 kinase in human pancreatic carcinomas. J Cell Biochem 88: 470–476.

  4. , , , , , . (2004). The PI 3-kinase/Akt signaling pathway is activated due to aberrant Pten expression and targets transcription factors NF-kappaB and c-Myc in pancreatic cancer cells. Oncogene 23: 8571–8580.

  5. , , , , . (2005). Alterations of tumor suppressor gene p16INK4a in pancreatic ductal carcinoma. BMC Gastroenterol 5: 22.

  6. , . (2002). Pancreatic cancer biology and genetics. Nat Rev Cancer 2: 897–909.

  7. , , , , , et al. (2002). Obligate roles for p16(Ink4a) and p19(Arf)-p53 in the suppression of murine pancreatic neoplasia. Mol Cell Biol 22: 635–643.

  8. , . (2005). Mutant KRAS in the initiation of pancreatic cancer. Biochim Biophys Acta 1756: 97–101.

  9. , , , , , et al. (2002). Reduced PTEN expression in the pancreas overexpressing transforming growth factor-beta 1. Br J Cancer 86: 257–262.

  10. , , , , , et al. (2005). A prospective randomized trial comparing standard pancreatoduodenectomy with pancreatoduodenectomy with extended lymphadenectomy in resectable pancreatic head adenocarcinoma. Surgery 138: 618–628; discussion 628–630.

  11. , , , , , et al. (1999). Frequent 4-bp deletion in exon 9 of the SMAD4/MADH4 gene in familial juvenile polyposis patients. Genes Chromosomes Cancer 25: 403–406.

  12. , , . (2000). Mosaic Cre-mediated recombination in pancreas using the pdx-1 enhancer/promoter. Genesis 26: 143–144.

  13. , , , , , et al. (2001). Negative regulation of neural stem/progenitor cell proliferation by the Pten tumor suppressor gene in vivo. Science 294: 2186–2189.

  14. , , . (2002). Direct evidence for the pancreatic lineage: NGN3+ cells are islet progenitors and are distinct from duct progenitors. Development 129: 2447–2457.

  15. , , , , , et al. (2007). Chronic pancreatitis is essential for induction of pancreatic ductal adenocarcinoma by K-Ras oncogenes in adult mice. Cancer Cell 11: 291–302.

  16. , , , , , et al. (2003). BRCA2 germline mutations in familial pancreatic carcinoma. J Natl Cancer Inst 95: 214–221.

  17. , , , , , et al. (1996a). Homozygous deletion map at 18q21.1 in pancreatic cancer. Cancer Res 56: 490–494.

  18. , , , , , et al. (1996b). DPC4, a candidate tumor suppressor gene at human chromosome 18q21.1. [see comments]. Science 271: 350–353.

  19. , , , , , . (2003). Activated Notch1 prevents differentiation of pancreatic acinar cells and attenuate endocrine development. Dev Biol 260: 426–437.

  20. , , . (1997). TGF-beta signalling from cell membrane to nucleus through SMAD proteins. Nature 390: 465–471.

  21. , , , , . (2006). Genetics and biology of pancreatic ductal adenocarcinoma. Genes Dev 20: 1218–1249.

  22. , , , , , et al. (2003). Preinvasive and invasive ductal pancreatic cancer and its early detection in the mouse. Cancer Cell 4: 437–450.

  23. , , , , , et al. (1998). Mutations in the SMAD4/DPC4 gene in juvenile polyposis [see comments]. Science 280: 1086–1088.

  24. , , , , , et al. (2006). Pathology of genetically engineered mouse models of pancreatic exocrine cancer: consensus report and recommendations. Cancer Res 66: 95–106.

  25. , , , , , et al. (2006). Aggressive pancreatic ductal adenocarcinoma in mice caused by pancreas-specific blockade of transforming growth factor-beta signaling in cooperation with active Kras expression. Genes Dev 20: 3147–3160.

  26. , , , , , et al. (2007). Kras(G12D) and Smad4/Dpc4 haploinsufficiency cooperate to induce mucinous cystic neoplasms and invasive adenocarcinoma of the pancreas. Cancer Cell 11: 229–243.

  27. , , , , , et al. (1998). Delta-1 activation of notch-1 signaling results in HES-1 transactivation. Mol Cell Biol 18: 7423–7431.

  28. , , . (2002). Expression of G1-S modulators (p53, p16, p27, cyclin D1, Rb) and Smad4/Dpc4 in intrahepatic cholangiocarcinoma. Hum Pathol 33: 877–883.

  29. , , , , , et al. (2006). Smad4 signalling in T cells is required for suppression of gastrointestinal cancer. Nature 441: 1015–1019.

  30. , , , , , et al. (2003). Squamous cell carcinoma and mammary abscess formation through squamous metaplasia in Smad4/Dpc4 conditional knockout mice. Development 130: 6143–6153.

  31. , . (2008). Pancreatic cancer. Annu Rev Pathol 3: 157–188.

  32. , , . (2006). Molecular pathogenesis of pancreatic cancer. Best Pract Res Clin Gastroenterol 20: 211–226.

  33. . (1998). TGF-beta signal transduction. Ann Rev Biochem 67: 753–791.

  34. , . (2005). Mechanisms of disease: PI3K/AKT signaling in gastrointestinal cancers. Z Gastroenterol 43: 1133–1139.

  35. , , , . (2003). Notch signaling controls multiple steps of pancreatic differentiation. Proc Natl Acad Sci USA 100: 14920–14925.

  36. , , , , , et al. (2002). Suppression of tumorigenesis and induction of p15(ink4b) by Smad4/DPC4 in human pancreatic cancer cells. Clin Cancer Res 8: 3628–3638.

  37. , , , , , . (2006). Hair follicle defects and squamous cell carcinoma formation in Smad4 conditional knockout mouse skin. Oncogene 25: 207–217.

  38. , , . (2007). The multifaceted role of Notch in cancer. Curr Opin Genet Dev 17: 52–59.

  39. , , , , , et al. (1997). Tumor-suppressive pathways in pancreatic carcinoma. Cancer Res 57: 1731–1734.

  40. , , . (2007). Matrix metalloproteinase 7 controls pancreatic acinar cell transdifferentiation by activating the Notch signaling pathway. Proc Natl Acad Sci USA 104: 19327–19332.

  41. , , . (1998). Notch-1 signalling requires ligand-induced proteolytic release of intracellular domain. Nature 393: 382–386.

  42. , , , . (2003). Phosphorylated Akt/PKB controls cell growth and apoptosis in intraductal papillary-mucinous tumor and invasive ductal adenocarcinoma of the pancreas. Pancreas 26: 250–257.

  43. , , , , , et al. (1998). The tumor suppressor gene Smad4/Dpc4 is required for gastrulation and later for anterior development of the mouse embryo. Genes Dev 12: 107–119.

  44. . (1999). Generalized lacZ expression with the ROSA26 Cre reporter strain. Nat Genet 21: 70–71.

  45. , , , , , et al. (2005). Pten constrains centroacinar cell expansion and malignant transformation in the pancreas. Cancer Cell 8: 185–195.

  46. , , , , , et al. (2002). Allelotype analysis of the PTEN, Smad4 and DCC genes in biliary tract cancer. Anticancer Res 22: 1529–1536.

  47. , , , , , . (1999). Gastric and duodenal polyps in Smad4 (Dpc4) knockout mice. Cancer Res 59: 6113–6117.

  48. , , , , , et al. (1996). Allelotype of adenoma and differentiated adenocarcinoma of the stomach. J Pathol 180: 371–377.

  49. , , , , , et al. (2006). Synergistic function of Smad4 and PTEN in suppressing forestomach squamous cell carcinoma in the mouse. Cancer Res 66: 6972–6981.

  50. , . (1992). Pancreatic carcinoma. N Engl J Med 326: 455–465.

  51. , . (2006). Genetic disruptions within the murine genome reveal the numerous roles of the Smad gene family in development, disease, and cancer, vol. 5. Springer: Dordrecht, 151–176pp.

  52. , , . (2000). Functions of mammalian Smad genes as revealed by targeted gene disruption in mice. Cytokine Growth Factor Rev 11: 49–58.

  53. , , , , , et al. (2000). Haploid loss of the tumor suppressor Smad4/Dpc4 initiates gastric polyposis and cancer in mice. Oncogene 19: 1868–1874.

  54. , , , , , et al. (2006). Induction of intrahepatic cholangiocellular carcinoma by liver-specific disruption of Smad4 and Pten in mice. J Clin Invest 116: 1843–1852.

  55. , . (2009). The pathology and genetics of metastatic pancreatic cancer. Arch Pathol Lab Med 133: 413–422.

  56. , , , , , et al. (2005). Targeted disruption of Smad4 in mouse epidermis results in failure of hair follicle cycling and formation of skin tumors. Cancer Res 65: 8671–8678.

  57. , , , . (2002). Generation of Smad4/Dpc4 conditional knockout mice. Genesis 32: 80–81.

  58. , , , . (1998). The tumor suppressor SMAD4/DPC4 is essential for epiblast proliferation and mesoderm induction in mice. Proc Natl Acad Sci USA 95: 3667–3672.

  59. , , , , , et al. (2002). Pancreaticoduodenectomy with or without distal gastrectomy and extended retroperitoneal lymphadenectomy for periampullary adenocarcinoma, part 2: randomized controlled trial evaluating survival, morbidity, and mortality. Ann Surg 236: 355–366; discussion 366–358.

Download references


We thank Dr Douglas Melton for providing Pdx-Cre mice; Dr Hong Wu for Pten conditional mutant mice; and members of the Deng lab for critically reading the paper. This research was supported by the Intramural Research Program of the National Institute of Diabetes, Digestive and Kidney Diseases, National Institutes of Health, USA.

Author information


  1. Genetics of Development and Disease Branch, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD, USA

    • X Xu
    • , B Ehdaie
    •  & C -X Deng
  2. Department of Pathology, Okayama University Graduate School of Medicine and Dentistry, Okayama, Japan

    • N Ohara
    •  & T Yoshino


  1. Search for X Xu in:

  2. Search for B Ehdaie in:

  3. Search for N Ohara in:

  4. Search for T Yoshino in:

  5. Search for C -X Deng in:

Corresponding author

Correspondence to C -X Deng.

Supplementary information

Word documents

  1. 1.

    Supplementary Materials

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)