SEL1L expression in pancreatic adenocarcinoma parallels SMAD4 expression and delays tumor growth in vitro and in vivo

Article metrics


Recent data suggest that SEL1L may play an important role in pancreatic carcinoma, similar to breast cancer, where the expression of SEL1L has been associated with a reduction in both proliferative activity in vitro and clinical tumor aggressiveness. To investigate this possibility, we examined the expression of Sel1L in a series of primary pancreatic carcinomas by immunohistochemistry and characterized the effects of Sel1L overexpression both in vitro and in vivo. In 74 pancreatic cancers analysed, 36% lacked Sel1L expression, although there was no significant correlation between the expression of Sel1L and any clinicopathologic parameter, including survival. However, immunohistochemical reactivity for Sel1L and Dpc4/Smad4 was concordant in 69% of cases (χ2 test P<0.004). Overexpression of SEL1L in stably transfected pancreatic cancer cells caused both a decrease in clonogenicity and anchorage-independent growth as well as a significant increase in the levels of activin A and SMAD4. When implanted in nude mice, Suit-2-SEL1L-overexpressing clones displayed a considerably reduced rate of tumor growth. Thus, it can be hypothesized that Sel1L plays an important function in the growth and aggressiveness of pancreatic carcinoma. Moreover, our data provide evidence that SEL1L has an impact on the expression of genes involved in regulation of cellular growth, possibly through the TGF-β signaling pathway.


The SEL1L gene (Biunno et al., 1997; Biunno et al., 2000) is the human ortholog of the Caenorhabditis elegans sel-1(suppressor-enhancer-lin) gene, which encodes for a protein that acts as a negative regulator of lin-12 (Grant and Greenwald, 1996,1997). The lin-12/Notch genes encode for proteins that play crucial roles in determining developmental choices in several precursor cell types both in C. elegans and mammals (Weinmaster, 1997; Greenwald, 1998; Artavanis-Tsakonas et al., 1999). Notably, Notch signaling appears to be critical in pancreatic exocrine and endocrine cell fate determination of the supposed common pluripotent endodermal progenitor (Apelqvist et al., 1999; Jensen et al., 2000). Notch signaling has been implicated in cell proliferation and apoptosis, and two vertebrate lin-12/Notch homologues, murine int-3 and human Tan-1, have been associated with several cancers (Ellisen et al., 1991; Jhappan et al., 1992; Robbins et al., 1992; Capobianco et al., 1997; Shelly et al., 1999). Comparative sequence analysis of SEL1L in different species has shown that it is highly conserved, suggesting that it plays an important role in cellular functions (Biunno et al., 2002).

SEL1L has been reported as highly expressed in human developing and normal adult pancreas with much lower levels detected in other tissues (Biunno et al., 1997; Donoviel et al., 1998; Harada et al., 1999; Cattaneo et al., 2000). SEL1L has been also found downregulated or completely repressed in a proportion of primary adenocarcinomas of the pancreas (Biunno et al., 1997; Donoviel et al., 1998). Furthermore, the SEL1L promoter showed preferential activity in cancer cell lines of pancreatic origin (Cattaneo et al., 2001). Taken together, these data suggest that SEL1L might play a role in pancreatic cancer. Indeed, the involvement of SEL1L in oncogenesis has been recently demonstrated in breast cancer, where the expression of SEL1L is associated with a reduction in both proliferative activity in vitro and aggressiveness in patients (Orlandi et al., 2002b).

In the present paper, we report on the expression of SEL1L in a series of primary pancreatic adenocarcinomas using a monoclonal antibody raised against the recombinant protein (Orlandi et al., 2002a). We further investigated the effect of inducible SEL1L expression in pancreatic cancer in vitro and in vivo using stable transfectants of Suit-2 cells, which express endogenous SEL1L at low levels and have weak endogenous SEL1L promoter activity (Cattaneo et al., 2001). Our results suggest that SEL1L may play a role in pancreatic tumor growth and aggressiveness, possibly through the TGF-β signaling pathway.


Immunohistochemical expression of Sel1L in normal pancreas and primary pancreatic adenocarcinomas: correlation with clinicopathological data and expression of Dpc4/Smad4

A total of 74 cases of primary pancreatic adenocarcinoma were examined in the present study. Pertinent clinicopathological and Sel1L expression data are reported in Table 1. Clinicopathological data, including stage, grade, presence of vascular and/or perineurial invasion were available for all cases, while follow-up was available for 50 patients. Data relative to the combined immunohistochemical expression of Sel1L and Dpc4/Smad4 is shown in Table 2.

Table 1 Clinicopathological features and expression of Sel1L in 74 pancreatic adenocarcinomas
Table 2 Expression of Sel1L and Smad4 in pancreatic carcinomas

Representative examples of immunostaining of Sel1L are shown in Figure 1. In normal pancreas, Sel1L was strongly expressed in acini, while islet cells showed somewhat weaker staining. Positivity was typically granular cytoplasmic diffuse, with frequent perinuclear staining. Ductal cells were negative with the exception of those lining smaller ducts. The latter were often stained with a relatively weak and sometime only focal positivity. Of the 74 carcinomas, 47 cases (64%) had positive staining in virtually all cancer cells, while 27 cases (36%) were negative for expression of Sel1L. In positive cases, neoplastic cells showed cytoplasmic staining with occasional nuclear localization. The staining intensity of cancer cells was as strong as in normal pancreatic acini in 20 cases, while the remaining 27 cases showed less-intense reactivity in neoplastic cells. The cancer-associated stromal cells were generally unstained with occasional positivity in few spindle-shaped cells, while among reactive cells only plasma cells were positive, when present. Only those tumors showing staining of normal pancreatic structures and absence of reactivity in cancer cells were scored as negative.

Figure 1

SEL1L immunostaining in normal pancreas and pancreatic adenocarcinoma. In normal pancreas (a), Sel1L is expressed in acinar and islet cells. The staining is cytoplasmic granular diffuse with some perinuclear staining in acinar cells (a1), while endocrine cells show a somewhat weaker stain (a2). Cells in normal ducts are negative (a3), while the smaller ducts show a weak positivity (a4). Examples of pancreatic adenocarcinomas with and without Sel1L expression are shown in panels b and c, respectively. The general concordance between Sel1L and Dpc4 immunohistochemical expression in pancreatic cancer is exemplified by the serial sections of the cases shown in panels d/e and f/g, where the cancer cells invading a perineural space are stained for both Sel1L (d) and Dpc4 (e), while the Sel1L-negative adenocarcinoma associated with the islet of Langerhans (f) is also negative for Dpc4 (g).

The immunohistochemical status of SMAD4 was either already available (Moore et al., 2001a) or determined for all cases. The results of immunostaining for Smad4 in primary pancreatic adenocarcinoma have been previously described in detail (Moore et al., 2001a). Smad4 is expressed in all normal pancreatic acinar, ductal, and islet cells, where strong to intermediate cytoplasmic reactivity with variable nuclear labeling is observed. The cytoplasmic staining of islet cells is usually strong, and the nuclear immunoreactivity is variable from relatively few to virtually all nuclei (Scarpa et al., 2002). A total of 38 cases scored positive for Smad4 expression and showed diffuse cytoplasmic and occasional nuclear staining in the vast majority of tumor cells (>90%).

There was concordant immunohistochemical reactivity for Sel1L and Smad4 in 51 of the 74 cancers (69%). In particular, 31 cases (41%) showed the expression of both proteins, whereas 20 cases (27%) lacked immunohistochemical expression of either. Of the 23 cases showing discordant staining, 16 showed Sel1L expression and lack of Dpc4, while seven cases had positive staining for Smad4 and lack of Sel1L.

Both univariate and multivariate analysis was carried out including all the above reported conventional factors as well as Sel1L and Smad4 status as determined by immunohistochemistry. None of the clinicopathologic parameters including survival showed significant correlation with expression of either Sel1L or Smad4. Interestingly, the only significant association was found between the expression of Sel1L and Smad4 (χ2 test P<0.004).

Stable transfection of SEL1L in Suit-2 cells and genomic PCR analysis

To examine the role of SEL1L in regulating human pancreatic cancer cell growth, Suit-2 cells were transfected with a vector containing the entire SEL1L cDNA driven by an inducible promoter and selected with G-418. The Suit-2 cell line was chosen for these studies as it expresses relatively low levels of endogenous SEL1L RNA and protein (data not shown) and has weak endogenous SEL1L promoter activity (Cattaneo et al., 2001). This cell line possesses mutations in the K-ras, p53, and p16 genes, but is wild type for SMAD4 (Moore et al., 2001b). Stable transfection of these cells with either SEL1L-pDEX.1 or the empty vector pDEX.1 yielded equivalent numbers of G-418-resistant clones, indicating the lack of growth inhibitory or toxic effects of the SEL1L gene under normal culture conditions.

In total, 22 clones were analysed for the presence and integrity of the construct. In particular, genomic insertion of the construct and the correct junction between the promoter vector and SEL1L cDNA was verified by PCR amplification of DNA from Suit-2-SEL1L-pDEX.1 and Suit-2-pDEX.1 cells using primers amplifying a fragment spanning the pDEX.1 promoter (sense pDEX.1) and the 3′-portion of the SEL1L cDNA (antisense IBD8). Suit-2-SEL1L 10, 14, and 37 clones retained the entire SEL1L cDNA in frame with the promoter as shown by the presence of 2.8 kb DNA fragment (Figure 2, lanes 2–4). The cells stably transfected with vector alone, designated Suit-2-pDEX.1 (lane 5), showed no such amplification product.

Figure 2

Genomic PCR analysis of Suit-2-SEL1L clones. High molecular weight DNA extracted from Suit-2-SEL1L clones 10, 14, and 37 (lanes 2–4) and from the control Suit-2-pDEX.1 clone (lane 5) was PCR amplified using pDEX.1 and IBD8 primers. The 2.8 kb DNA fragment indicates the integrity of the promoter-cDNA junction. M is the marker; positive (transfected plasmid) and negative (water) controls are shown in lanes 1 and 6, respectively.

Transcription and Western blot analysis of Suit-2-SEL1L clones after DEX induction

To determine the presence of recombinant SEL1L transcript, reverse transcription–PCR (RT–PCR) was carried out on RNA extracted from Suit-2-SEL1L and Suit-2-pDEX.1 cells untreated or treated with DEX for 1 and 2 weeks. PCR amplification was performed using a primer set recognizing the 5′-end of the SEL1L cDNA (antisense IB13R) and the promoter vector downstream of the transcription initiation site (sense pDEX.1). As shown in Figure 3a, Suit-2-pDEX.1 cells showed no SEL1L signal either before or after treatment with DEX (lanes 1–3), whereas Suit-2-SEL1L clones 14, 37, and 10 revealed the presence of an appreciable increase in the 900 bp transcript after induction by dexamethasone (DEX) (lanes 5, 6, 8, 9, 11, 12). A weak signal in uninduced samples is indicative of promoter leakage, a phenomenon previously described (Orlandi et al., 2002b). The same amount of unretrotranscribed RNA was amplified to control for contaminating genomic DNA (data not shown). Hypoxanthine phosphoribosyltransferase (HPRT) expression was used to monitor the efficiency and quality of the newly synthesized cDNA.

Figure 3

Transcriptional and Western blot analysis of Suit-2-SEL1L clones. (a) RT-PCR analysis of RNA extracted from the Suit-2-pDEX.1 control clone uninduced and DEX-induced for 1 and 2 weeks (lanes 1–3), Suit-2-SEL1L clones 14 (lanes 4–6), 37 (lanes 7–9) and 10 (lanes 10–12). M is the marker. The induction of the 900 bp recombinant SEL1L transcript is evident as is the increased expression level of DPC4/Smad4 (565 bp) and activin A (395 bp) after DEX treatment, while the activin receptor II expression level remains unaltered. Equal amounts of synthesized cDNA were tested by HPRT primers. (b) Protein expression of SEL1L from Suit-2-pDEX.1 control clone before and after 1 and 2 weeks of DEX induction (lanes 1–3), from Suit-2-SEL1L clones 14 (lanes 4–6), 37 (lanes 7–9) and 10 (lanes 10–12). The lower band is likely to be an artifact observed in Western analysis as it is not induced and remains constant in level under all experimental conditions

Western blot analysis using the Msel1 monoclonal antibody (MAb) was performed to verify that the increased amounts of Sel1L RNA led to a concomitant increase of the protein. As expected, control Suit-2-pDEX.1 cells contained similar levels of SEL1L protein before and after DEX addition (Figure 3b lanes 1–3), whereas Suit-2-SEL1L clones 14, 37, and 10 showed increased levels of SEL1L after DEX addition (lanes 5, 6, 8, 9, 11, 12).

The effect of SEL1L expression on anchorage-dependent and -independent growth of Suit-2-SEL1L clones

Suit-2-SEL1L clones 14 and 37 as well as Suit-2-pDEX.1 cells were grown in the absence or presence of DEX for 7 days. Cell proliferation rates were determined daily for each clone by cell enumeration of trypsinized suspensions using trypan blue exclusion. Neither uninduced nor induced SEL1L clones showed any difference in replication times (Figure 4), indicating the lack of growth inhibitory effects of the overexpressed SEL1L under normal culture conditions.

Figure 4

In vitro proliferation of Suit-2-SEL1L clones. Suit-2-pDEX.1 and Suit-2-SEL1L clones 14 and 37 were cultured, harvested at the indicated times and counted in duplicate by trypan blue exclusion. Data represent mean values obtained from three independent experiments; bars, ±s.d. There was no significant difference between the uninduced and induced cells

Colony-formation analysis, however, revealed that there were fewer and smaller colonies deriving from DEX-treated SEL1L clones with respect to those from untreated cells or from DEX-treated or untreated Suit-2-pDEX.1 transfected control clones (Figure 5a). Suit-2-SEL1L clones 14 and 37 exhibited a 61.3 and 53.4% reduction in the ability to form adherent colonies after induction, respectively (P<0.00001; P<0.001). In contrast, the plating efficiency of control Suit-2-pDEX.1 cells was unchanged after treatment with DEX. The number of colonies formed before and after DEX addition is shown numerically in Figure 5b.

Figure 5

Colony-forming assay. Suit-2-pDEX.1 and Suit-2-SEL1L clones 14 and 37 were pretreated with 1 μ M of DEX for 7 days or grown uninduced in culture medium, trypsinized, and seeded in duplicate at density of 300 cells/well (35 mm diameter plate) in the presence or absence of DEX, respectively. Colonies were scored after 10 days (a). Data are given as mean values obtained from three independent experiments (b); bars, ±s.d.

These same clones were also analysed for their ability to grow and form colonies in soft agar (soft agar colony assay). Suit-2-SEL1L clones induced with DEX showed a significant decrease (≈80%) in their cloning efficiency, whereas Suit-2-pDEX.1 cells were unaffected by DEX treatment (Figure 6). These results suggest that although increased levels of SEL1L were not sufficient to affect in vitro growth rates, it was adequate to alter tumor growth.

Figure 6

Anchorage-independent clonogenicity in soft agar. Suit-2-pDEX.1 and Suit-2-SEL1L clone 14 were pretreated with 1 μ M of DEX for 7 days or grown uninduced and seeded in 0.6% soft agar in duplicate at a density of 15 000 cells/well. Colonies were scored after 14 days. Nine fields of each duplicate determination were examined under the microscope and colonies counted. Data are given as mean values obtained from three independent experiments; bars, ±s.d.

Induction of SEL1L is associated with increased expression of SMAD4 and activin A

It has been recently reported that in rat RSMG-1 cells (a submandibular gland epithelial cell line) SEL1L mRNA expression is induced by activin A (Furue et al., 2001), a member of the TGF-β superfamily. Activin A modulates cell differentiation and proliferation by effecting a broad range of cellular targets (Chen et al., 2002). In order to identify which molecular pathways may be involved in signaling by SEL1L, we investigated the endogenous expression level of activin A and DPC4/SMAD4, two genes induced by TGF-β.

RT-PCR analysis was carried out on Suit-2-SEL1L clones 14 and 37, treated with DEX for 1 or 2 weeks. These cells showed significantly increased expression levels of endogenous activin A and DPC4/Smad4 compared to either untreated control cells or DEX-treated Suit-2-pDEX.1 cells (Figure 3a). No differences in activin receptor II expression were observed in any of the clones after DEX treatment, demonstrating the specificity of this effect.

Quantitative reverse transcription (qRT–PCR) was used to validate the differences in the expression levels of SEL1L, SMAD4, and activin A in Suit-2-SEL1L clones 14 and 37 after 1 or 2 weeks induction with DEX. In clone 14, SEL1L mRNA levels were increased by 5.1- and 4.6-fold after 1 and 2 weeks of induction with DEX, respectively. The same clone showed increases of 2.7- and 4.5-fold in SMAD4 mRNA and increases of 3.4- and 6.9-fold after 1 and 2 weeks of induction, respectively. Similarly, clone 37 showed levels of SEL1L mRNA levels that were increased by 2.7- and 4.1-fold after 1 and 2 weeks of induction with DEX, respectively. Increases of 2.8- and 4.9-fold were observed in SMAD4 mRNA and increases of 3.3 and 7.2-fold in activin A after 1 and 2 weeks of induction, respectively.

Suppression of tumorigenicity by SEL1L in immunodeficient mice

To investigate the effects of induced Sel1L overexpression on tumor growth, immunosuppressed nu/nu mice were implanted with either Suit-2-pDEX.1 or Suit-2-SEL1L clones that had been pretreated with 100 nM DEX for 14 days; tumor growth was monitored for 3 weeks. Representative tumors are shown in Figure 7, along with tumor volume plotted as a function of time. Untreated Suit-2-pDEX.1 clones showed the highest rate of growth and were already visible at 3 days postimplantation. The Suit-2-Sel1L clones pretreated with DEX displayed a retarded rate of tumor growth, and nodules were not visible until 15 days postimplantation. Untreated Suit-2-SEL1L cells and DEX-treated Suit-2-pDEX.1 cells showed intermediate growth rates. The well-known antiproliferative effect of glucocorticoids on a variety of cell types both in vitro and in vivo most likely explains the difference in proliferation between treated and untreated clones (Cook et al., 1988; Thompson, 1989; Norman et al., 1994; Kudawara et al., 2001). The observed differences between untreated Suit-2-pDEX.1 and Suit-2-SEL1L cells may be explained by promoter leakage and, thus, the antiproliferative effect of increased SEL1L expression. The addition of DEX to Suit-2-SEL1L cells further reduces proliferation due to the combined antiproliferative effects of DEX and the induction of SEL1L by the glucocorticoid. No significant histological differences were observed in any of the four treatment groups.

Figure 7

Tumor growth of Suit-2 clones in nu/nu mice. In (a), representative examples of tumors from mice implanted with untreated Suit-2-pDEX.1 cells (1) and Suit-2-Sel1L clone 37 treated with DEX prior to implantation (2) are shown in the upper panels, both at 3 weeks following subcutaneous injection. Their histologic appearance showed no significant difference in any of the treatment groups. In (b), average tumor volume plotted as a function of time for the four indicated different treatment groups


The results of the present study may be summarized as follows: (a) in normal pancreas, SEL1L is expressed in acinar and endocrine cells, while ductal cells were negative with the exception of smaller ducts that were frequently positive showing a relatively weak and sometimes only focal stain; (b) 36% of primary pancreatic adenocarcinomas lack immunohistochemical expression of Sel1L protein, a finding that significantly correlates with the lack of Dpc4 expression; (c) stable, inducible overexpression of SEL1L causes a decrease in clonogenicity and anchorage-independent growth of pancreatic cancer cells; (d) induction of SEL1L in Suit-2-SEL1L clones significantly increased expression levels of endogenous activin A and SMAD4 compared to either untreated or DEX-treated Suit-2-pDEX.1 cells; (e) Suit-2-SEL1L clones pretreated with DEX displayed a significantly reduced rate of tumor growth when implanted in nude mice compared to untreated cells or Suit-2-pDEX.1 cells.

While the precise function of SEL1L is unknown, its homology to the C. elegans gene sel-1 and yeast HRD3 genes suggests that it may be involved in a secretory/endocytic pathway (Hampton et al., 1996; Grant and Greenwald, 1997). Previous immunohistochemical studies have shown that transfected Sel1L is localized in cytoplasmic vesicles (Grant and Greenwald, 1996; Orlandi et al., 2002b). Moreover, in normal pancreas, Sel1L is also located in the cytoplasm and the finding that Sel1L is expressed in both pancreatic endocrine and acinar cells and absent in large ducts is further consistent with the hypothesis that Sel1L has a role in protein trafficking.

By RT–PCR analysis on bulk pancreatic adenocarcinoma, it has been previously reported that SEL1L is lacking in a proportion of cases (Biunno et al., 1997; Donoviel et al., 1998). However, the identity of the cell population(s) giving rise to this expression, or which was responsible for the decreased expression in primary cancers was unclear given the scarce cancer cellularity that is characteristic of pancreatic adenocarcinoma. Our immunohistochemical analysis of Sel1L on a series of primary pancreatic adenocarcinomas established that SEL1L is indeed expressed in neoplastic cells and that its expression is abolished in tumor cells in 36% of the cases.

A similar situation was found in breast cancer in that Sel1L expression is likewise absent in about 30% of the cases. In vitro studies in MCF-7 breast cancer cells showed that Sel1L is associated with a reduction in proliferative activity with possible involvement in cell–matrix interactions (Orlandi et al., 2002b). The increased aggressiveness of Sel1L-negative carcinomas was clinically evident since the overall survival of breast cancer patients correlated significantly with the downregulation of Sel1L (Orlandi et al., 2002b).

We found no correlation between the expression of Sel1L and any clinicopathological parameter, including prognosis, which is at variance with the observation that the lack of Sel1L expression was correlated with a worse prognosis in breast cancer (Orlandi et al., 2002b). However, in breast cancer patients, the correlation between Sel1L expression and survival was only apparent for those patients surviving for at least 3 years (Orlandi et al., 2002b). As the average survival time for resected pancreatic cancer patients is roughly 15 months, it is not surprising that no statistically significant effects of Sel1L expression were found in this highly aggressive cancer type. We also found no significant correlation between the expression of Smad4 and any clinical parameter, in contrast to a previous report that suggested that patients undergoing a Whipple resection for PDC show more favorable survival when the cancer expresses Smad4 (Tascilar et al., 2001). The former report analysed a panel of 249 cases, 30 of which were cancers smaller than 3 cm, and the correlation between Smad4 expression and survival was less evident in these neoplasms. Even though T1–T2 stage cancers were clearly under-represented in our set of cases, we nonetheless observed no such statistical trend between Smad4 expression and patient survival.

In order to further study a possible involvement of Sel1L in pancreatic cancer, we stably transfected Suit-2 cells, previously found to constitutively express low levels of SEL1L, with a construct containing the SEL1L entire open reading frame under control of an inducible promoter. Increased SEL1L expression levels in Suit-2 cells were adequate to strongly reduce their ability to form adherent colonies and nearly completely abolished the ability of these cells to grow in soft agar. Interestingly, induced SEL1L expression did not affect the proliferation rate of Suit-2 cells. In this respect, it is worth mentioning that the re-expression of SMAD4 in colon cancer as well as in pancreatic cancer cells was sufficient to suppress tumor formation in nude mice without any influence on the proliferative activity (Schwarte-Waldhoff et al., 1999; Schwarte-Waldhoff et al., 2000). Additionally, mice implanted with Suit-2 clones overexpressing SEL1L showed significant reduction in tumor development compared to those implanted with uninduced cells or cells with only endogenous Sel1L.

The correlation between Smad4 and Sel1L expression in primary cancers prompted us to examine this association further, where it was found that ectopic overexpression of Sel1L in Suit2 cells led to increased levels of both activin A and SMAD4. Both these genes are involved in the TGF-β signal transduction pathway and their role in tumor development and progression has been investigated (Schutte et al., 1996; Roberts, 1999; Massague et al., 2000; Rooke and Crosier, 2001; Chen et al., 2002). Activin A exerts its effect on a wide range of cellular targets by modulation of cell differentiation and proliferation (Chen et al., 2002). In two in vitro breast cancer systems (MCF-7 and T47D), activin A appears as a potent cell growth inhibitor (Cocolakis et al., 2001; Liu et al., 1996). SMAD4 is inactivated by mutation or homozygous deletion in about one-half of all pancreatic cancers (Rozenblum et al., 1997; Moore et al., 2001a) as well as in a smaller percentage of colon and breast cancers (Lei et al., 1996; Schutte et al., 1996). Moreover, it was reported that SMAD4 re-expression in pancreatic carcinoma cell line induced the cell to shift from potently angiogenic to an antiangiogenic phenotype in vitro and in vivo by modulating regulators of angiogenesis (Schwarte-Waldhoff et al., 2000). The biological significance of Sel1L-induced tumor suppression is likely to be associated with increased activin A and SMAD4 expression levels, at least in vitro.

In conclusion, the expression of Sel1L, SMAD4, or both genes is lacking in about 60% of pancreatic cancers, further stressing the importance of the TGF-β pathway in pancreatic tumorigenesis. Moreover, our data provide evidence that SEL1L has an impact on the expression of TGF-β target genes and underline the utility of further investigating the role of Sel1L in pancreatic carcinoma. Similar to breast cancer, it can be suggested that Sel1L plays an important role in signal transduction in pancreatic carcinoma.

Materials and methods

All studies were approved by the Ethics Committee of Verona University (approval AIRC_482, January 2002).

Normal pancreas and primary pancreatic adenocarcinomas

Samples of normal adult pancreas and 74 primary pancreatic ductal carcinomas were from the files of the Department of Pathology, University of Verona, Italy. All cases have been characterized for stage, grade, presence of vascular and/or perineurial invasion. For 50 cases, follow-up data was also available. No blood-borne metastatic cancer was present in any of the present cases, which were all obtained from patients undergoing pancreatic resections. Only a few cancers were well differentiated (three cases) or undifferentiated (two cases). The vast majority of cases were either moderately (37 cases) or poorly (29 cases) differentiated. Three cancers were of the colloid histological type. The TNM staging system was used to classify pancreatic neoplasms (Kløppel et al., 2000). T1 are those tumors limited to the pancreas with a greatest dimension of 2 cm or less; T2 are those tumors limited to the pancreas and are larger than 2 cm; T3 are those tumors extending directly into the duodenum or bile duct or peripancreatic tissues; T4 are tumors extending directly into the stomach, spleen, colon, or large adjacent vessels. N, nodal metastasis, is a or b if one or more regional lymph nodes were involved.

Immunohistochemistry for Sel1L and Dpc4

Immunohistochemistry was performed on paraffin sections using the monoclonal antibodies anti-Sel1L (MSel1) (Orlandi et al., 2002a) and anti-Dpc4/Smad4 (Santa Cruz Smad4 clone B-8 from Autogen Bioclear, Calne, UK). The anti-Sel1L and Smad4 monoclonal antibodies were used at a dilution of 1 : 900 and 1 : 200, respectively, following antigen retrieval by microwaving (750 W) for three times 10 min each in 10 mM citrate buffer, pH 6. Endogenous peroxidase activity was quenched by incubation of slides in 3% hydrogen peroxide. Nonspecific binding of antibodies was blocked by incubating sections in a Protein Blocking Agent (UltraTech HRP, Immunotech, Marseille, France) for 10 min. After removal of excess blocking solution, the appropriate antibody was added. Sections were incubated for 120 min and then washed three times with PBS. Sections were incubated for a further 10 min with a Polyvalent Biotinylated Antibody (UltraTech HRP), followed by washing three times with PBS. Slides were then incubated with streptavidin–peroxidase reagent, followed by diaminobenzidine substrate. After washing with distilled water, slides were counterstained with Mayer's hematoxylin, dehydrated and mounted with Entellan. Negative controls were incubated without MAb MSel1.

Reproducibility of the immunohistochemistry analysis was assessed in the preliminary set up of the immunoperoxidase assays with MAb MSel1, which included selection of MAb concentration, scoring system with intra- and interobserver evaluation, and reproducibility on serial slides of the same cases. Normal pancreatic structures in the same sections served as positive controls. Cases scored as positive for Sel1L showed diffuse cytoplasmic and/or nuclear staining in the vast majority of cancer cells; cases scored as negative showed no stain at all or occasional positive cells. Cases scored as positive for Smad4 showed diffuse cytoplasmic and/or nuclear staining in the vast majority of cancer cells; cases scored as negative showed no stain at all or occasional positive cells, as previously described (Moore et al., 2001a).

SEL1L-pDEX.1 construct and transfection of Suit-2 cells

The SEL1L-pDEX.1 construct was previously described (Orlandi et al., 2002b). The human pancreatic cancer cell line Suit-2, derived from a liver metastasis of human pancreatic carcinoma (Moore et al., 2001b), was grown in RPMI (Microbiological Associates, Walkersville, MD, USA) supplemented with 10% fetal bovine serum (Hyclone), and penicillin/streptomycin (100 IU/ml) in a humidified chamber (95% air and 5% CO2) at 37°C. Suit-2 cells were stably transfected with SEL1L-pDEX.1 or pDEX.1 by electroporation (Bio-Rad gene pulsar) as previously described (Orlandi et al., 2002b). Individual neomycin-resistant clones were isolated and grown for further analysis. To control the presence of the construct in cancer cells, DNA was PCR amplified using a set of primers specific for the promoter (pDEX.1, 5′-IndexTermCCTGGTCATCATCCTGCCTT-3′) and the 3′ regions of SEL1L (IBD8, 5′-IndexTermGCTGGATCCAGTGCCTATTACTGTGG-3′) with an expected product size of 2.8 kb. Amplification conditions used were as previously described (Orlandi et al., 2002b).

In vitro induction of SEL1L transcription

Suit-2 selected clones containing the SEL1L-pDEX.1 construct in a stable form were treated with 1 μ M of DEX for 7–14 days after which the cells were harvested and analysed by RT–PCR and Western blotting.

RT–PCR Analysis of SEL1L, DPC4, activin A, and activin receptor II

Total RNA was extracted from cells using the SV 96 Total RNA Isolation System kit (Promega) following the manufacturer's instructions. Total RNA (1 μg) was used in each reaction containing 5 μ M MgCl2, 1 × reaction buffer (50 mM Tris-HCl (pH 8.8), 8 nM MgCl2, 30 mM KCl, 1 mM dithiothreithol), 1 μ M dNTPs, 5 U of RNase inhibitor (RNAsin), 0.8 μg oligo-p(dT)15 primer, 1.6 μg random primer, and 15 units of avian myeloblastosis virus reverse transcriptase (Amersham). The reaction mixture was incubated for 10 min at 25°C and for 60 min at 42°C. The enzyme was denatured at 99°C for 5 min and chilled on ice. PCR amplifications were performed using 2 μl of RT product per reaction. PCR conditions consisted of 3 min at 94°C followed by 20–32 cycles (see below) at 94°C for 1 min, annealing temperature for 1 min (Ta noted below), extension at 72°C for 1 min followed by a final extension of 5 min at 72°C. PCR conditions used to detect the expression of the genes of interest were: SEL1L (exogenous) 30 cycles, Ta 60°C; activin A 32 cycles, Ta 62°C; activin receptor II 24 cycles, Ta 55°C; SMAD4 32 cycles, Ta 58°C; HPRT 20 cycles, Ta 60°C. All the PCR products were electrophoresed on 1.5% agarose gel and stained with ethidium bromide. Sequencing of the amplified DNA products confirmed the corresponding human genes. Oligonucleotide primer sequences were:

Quantitative RT–PCR

Gene expression levels of SEL1L, SMAD4, and activin A were evaluated in Suit-2-SEL1L clones 14 and 37 after 1 and 2 weeks induction with DEX using SYBR-Green-based real-time qRT-PCR. β-Actin was used as internal reference in each reaction containing 1X SYBR-Green master mix buffer (Applied Biosystems), 200 nM forward and reverse gene-specific primers, and 40 ng of cDNA. Assays were performed in triplicate using an Applied Biosystems Model 7700 instrument. The amount of PCR product was measured in real time as the increase in SYBR Green fluorescence. cDNA aliquots were quantified for target genes using the threshold cycle (C(t)) method normalized to β-actin. Control wells containing SYBR-Green PCR master mix and primers without sample cDNA emitted no fluorescence after 40 cycles.

Western Blotting for Sel1L

Cells were lysed using a buffer containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 10% glycerol, 0.5% NP40, 10 μg/ml aprotinin, and 1 μ M phenylmethylsulfonyl fluoride. Protein concentration was determined by the Bradford assay; 40 μg of each sample was resolved on 10% SDS–polyacrylamide gel, and Western blotting was performed as previously described (Orlandi et al., 2002a, b). MAb MSel1 was used at 10 μg/ml.

Cells growth rate analysis

Cells were seeded in 48-well plates three (3000 cells/well) in 200 μl culture medium in the absence or presence of 1 μ M DEX and grown for 7 days. Cell proliferation was assessed by direct cell counting at specific time intervals using trypan blue. Each assay point was performed in duplicate and the experiment repeated over four times.

Colony-forming assay

Suit-2-pDEX.1 or Suit-2-SEL1L-pDEX.1 cells, treated and untreated with 1 μ M DEX for 1 week, were trypsin-detached and seeded in duplicate in six-well plates at a density of 300 cells/well. After 10 days of growth in the absence or presence of DEX, the cells were fixed with methanol and stained with 10% Giemsa. Experiments were repeated at least three times. The Student's t-test was used to determine statistical significance.

Anchorage-independent clonogenicity assays

Suit-2-pDEX.1 or Suit-2-SEL1L-pDEX.1 cells treated with 1 μ M DEX for 1 week were trypsin-detached and seeded in duplicate in six-well plates at a density of 15 000 cells/well in semisolid medium containing 0.3% Bacto-Agar (Difco) over 0.8% agarose layer containing DEX when indicated. Fresh medium was added weekly. Colony formation was scored after 14 days of growth.

In vivo growth assay

For the tumorigenicity assay, suspensions of 1.5 × 106 cells in a volume of 0.1 ml of PBS were injected s.c. into the right flank of six-week-old male athymic mice (Crl:CD-1-nu/nu BR, Charles River, Lecco, Italy). Four sets of five mice were each injected with the Suit-2-pDEX.1 and Suit-2-SEL1L-pDEX.1 cell lines, each untreated and treated for 14 days with DEX. Tumor diameter was measured in two dimensions with a caliper.

Statistical analysis

Univariate analysis was performed using χ2 test or Fisher's exact test to evaluate categorical variables, whereas the Mann–Whitney U test was used for continuous variables. The Spearman correlation test was used to evaluate the association between variables. For survival analysis, the primary statistical outcome in this study was overall survival measured from the date of surgery. Death from cancer was the end point. Overall survival distribution was calculated by the product-limit method and analysed using the Mantel–Cox test. Multivariate survival analysis was performed using Cox's proportional-hazard model. The SPSS release 11.0 (SPSS Inc., Chicago, IL, USA) statistical program was used.



monoclonal antibody




reverse transcription–PCR


hypoxanthine phosphoribosyltransferase


  1. Apelqvist A, Li H, Sommer L, Beatus P, Anderson DJ, Honjo T, Hrabe de Angelis M, Lendahl U and Edlund H . (1999). Nature, 400, 877–881.

  2. Artavanis-Tsakonas S, Rand MD and Lake RJ . (1999). Science, 284, 770–776.

  3. Biunno I, Appierto V, Cattaneo M, Leone BE, Balzano G, Socci C, Saccone S, Letizia A, Della Valle G and Sgaramella V . (1997). Genomics, 46, 284–286.

  4. Biunno I, Bernard L, Dear P, Cattaneo M, Volorio S, Zannini L, Bankier A and Zollo M . (2000). Hum. Genet., 106, 227–235.

  5. Biunno I, Castiglioni B, Rogozin IB, DeBellis G, Malferrari G and Cattaneo M . (2002). Omics., 6, 187–198.

  6. Capobianco AJ, Zagouras P, Blaumueller CM, Artavanis-Tsakonas S and Bishop JM . (1997). Mol. Cell. Biol., 17, 6265–6273.

  7. Cattaneo M, Orlandi R, Ronchini C, Granelli P, Malferrari G, Menard S and Biunno I . (2000). Int. J. Biol. Markers, 15, 26–32.

  8. Cattaneo M, Sorio C, Malferrari G, Rogozin IB, Bernard L, Scarpa A, Zollo M and Biunno I . (2001). DNA Cell. Biol., 20, 1–9.

  9. Chen YG, Lui HM, Lin SL, Lee JM and Ying SY . (2002). Exp. Biol. Med. (Maywood), 227, 75–87.

  10. Cocolakis E, Lemay S, Ali S and Lebrun JJ . (2001). J. Biol. Chem., 276, 18430–18436.

  11. Cook PW, Swanson KT, Edwards CP and Firestone GL . (1988). Mol. Cell. Biol., 8, 1449–1459.

  12. Donoviel DB, Donoviel MS, Fan E, Hadjantonakis A and Bernstein A . (1998). Mech. Dev., 78, 203–207.

  13. Ellisen LW, Bird J, West DC, Soreng AL, Reynolds TC, Smith SD and Sklar J . (1991). Cell., 66, 649–661.

  14. Furue M, Zhang Y, Okamoto T, Hata RI and Asashima M . (2001). Biochem. Biophys. Res. Commun., 282, 745–749.

  15. Grant B and Greenwald I . (1996). Genetics, 143, 237–247.

  16. Grant B and Greenwald I . (1997). Development, 124, 637–644.

  17. Greenwald I . (1998). Genes. Dev., 12, 1751–1762.

  18. Hampton RY, Gardner RG and Rine J . (1996). Mol. Biol. Cell., 7, 2029–2044.

  19. Harada Y, Ozaki K, Suzuki M, Fujiwara T, Takahashi E, Nakamura Y and Tanigami A . (1999). J. Hum. Genet., 44, 330–336.

  20. Jensen J, Heller RS, Funder-Nielsen T, Pedersen EE, Lindsell C, Weinmaster G, Madsen OD and Serup P . (2000). Diabetes, 49, 163–176.

  21. Jhappan C, Gallahan D, Stahle C, Chu E, Smith GH, Merlino G and Callahan R . (1992). Genes. Dev., 6, 345–355.

  22. Kløppel G, Hruban R, Longnecker D, Adler G, Kern S and Partanen T . (2000). Pathology and Genetics of Tumours of the Digestive System: WHO lassification of Tumours. Hamilton S, Aaltonen L (eds). IARC press: Lyon, pp 219–250.

  23. Kudawara I, Ueda T, Yoshikawa H, Miyama T, Yamamoto T and Nishizawa Y . (2001). Eur. J. Cancer, 37, 1703–1708.

  24. Lei J, Zou TT, Shi YQ, Zhou X, Smolinski KN, Yin J, Souza RF, Appel R, Wang S, Cymes K, Chan O, Abraham JM, Harpaz N and Meltzer SJ . (1996). Oncogene, 13, 2459–2462.

  25. Liu QY, Niranjan B, Gomes P, Gomm JJ, Davies D, Coombes RC and Buluwela L . (1996). Cancer Res., 56, 1155–1163.

  26. Massague J, Blain SW and Lo RS . (2000). Cell, 103, 295–309.

  27. Moore P, Orlandini S, Zamboni G, Capelli P, Rigaud G, Falconi M, Bassi C, Lemoine N and Scarpa A . (2001a). Br. J. Cancer, 84, 253–262.

  28. Moore P, Sipos B, Orlandini S, Sorio C, Real F, Lemoine N, Gress T, Bassi C, Klöppel G, Kalthoff H, Löhr M and Scarpa A . (2001b). Virch. Arch., 439, 798–802.

  29. Norman J, Franz M, Schiro R, Nicosia S, Docs J, Fabri PJ and Gower Jr WR . (1994). J. Surg. Res., 57, 33–38.

  30. Orlandi R, Cattaneo M, Troglio F, Campiglio M, Biunno I and Menard S . (2002a). Int. J. Biol. Markers, 17, 104–111.

  31. Orlandi R, Cattaneo M, Troglio F, Casalini P, Ronchini C, Menard S and Biunno I . (2002b). Cancer Res., 62, 567–574.

  32. Robbins J, Blondel BJ, Gallahan D and Callahan R . (1992). J. Virol., 66, 2594–2599.

  33. Roberts AB . (1999). Microbes Infect, 1, 1265–1273.

  34. Rooke HM and Crosier KE . (2001). Pathology, 33, 73–84.

  35. Rozenblum E, Schutte M, Goggins M, Hahn SA, Panzer S, Zahurak M, Goodman SN, Sohn TA, Hruban RH, Yeo CJ and Kern SE . (1997). Cancer Res., 57, 1731–1734.

  36. Scarpa A, Orlandini S, Moore P, Lemoine N, Beghelli S, Baron A, Falconi M and Zamboni G . (2002). Virch. Arch., 440, 155–159.

  37. Schutte M, Hruban RH, Hedrick L, Cho KR, Nadasdy GM, Weinstein CL, Bova GS, Isaacs WB, Cairns P, Nawroz H, Sidransky D, Casero Jr RA, Meltzer PS, Hahn SA and Kern SE . (1996). Cancer Res., 56, 2527–2530.

  38. Schwarte-Waldhoff I, Klein S, Blass-Kampmann S, Hintelmann A, Eilert C, Dreschers S, Kalthoff H, Hahn SA and Schmiegel W . (1999). Oncogene, 18, 3152–3158.

  39. Schwarte-Waldhoff I, Volpert OV, Bouck NP, Sipos B, Hahn SA, Klein-Scory S, Luttges J, Kloppel G, Graeven U, Eilert-Micus C, Hintelmann A and Schmiegel W . (2000). Proc. Natl. Acad. Sci. USA, 97, 9624–9629.

  40. Shelly LL, Fuchs C and Miele L . (1999). J. Cell. Biochem., 73, 164–175.

  41. Tascilar M, Skinner HG, Rosty C, Sohn T, Wilentz RE, Offerhaus GJ, Adsay V, Abrams RA, Cameron JL, Kern SE, Yeo CJ, Hruban RH and Goggins M . (2001). Clin. Cancer Res., 7, 4115–4121.

  42. Thompson Jr EA . (1989). Cancer Res., 49, 2259s–2265s.

  43. Weinmaster G . (1997). Mol. Cell Neurosci., 9, 91–102.

Download references


This work was supported by grants from Associazione Italiana Ricerca Cancro, Milan, Italy to AS; Ministero Università (Cofin 2002068231 and 2001068593), Rome, Italy; European Community Grant QLG1-CT-2002-01196. MC is supported by a fellowship from Fondazione Italiana Ricerca sul Cancro (FIRC). PSM is supported by Fondazione Cassa di Risparmio di Verona (Bando 2001), Verona, Italy.

Author information

Correspondence to Ida Biunno or Aldo Scarpa.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Cattaneo, M., Orlandini, S., Beghelli, S. et al. SEL1L expression in pancreatic adenocarcinoma parallels SMAD4 expression and delays tumor growth in vitro and in vivo. Oncogene 22, 6359–6368 (2003) doi:10.1038/sj.onc.1206665

Download citation


  • pancreas
  • adenocarcinoma
  • SEL1L
  • DPC4/Smad4
  • TGFβ-pathway

Further reading