- Email: [email protected]
Increased PDX-1 expression is associated with outcome in patients with pancreatic cancer
Increased PDX-1 expression is associated with outcome in patients with pancreatic cancer Masayuki Koizumi, MD, Ryuichiro Doi, MD, Eiji Toyoda, MD, Toshihiko Masui, MD, Sidhartha S. Tulachan, MD, Yoshiya Kawaguchi, MD, Koji Fujimoto, MD, George K. Gittes, MD, and Masayuki Imamura, MD, Kyoto, Japan, and Kansas City, Mo
Background. During pancreatic development, pancreatic duodenal homeobox gene-1 (PDX-1) is expressed in pancreatic duct cells that have the potential to differentiate into islets. Therefore, PDX-1 is thought to be a marker of de-differentiated cells with the capacity to redifferentiate into several pancreatic cell types. We analyzed PDX-1 expression in human pancreatic cancer specimens, pancreatic cancer cell lines, and the effects of forced expression of PDX-1 in pancreatic cancer cells. Methods. Thirty-five pancreatic adenocarcinomas were immunohistochemically stained with a polyclonal rabbit antibody against mouse PDX-1. Correlations with tumor characteristics were made with chisquared analysis. The influence of clinicopathologic factors on survival was assessed. The expression of PDX-1 in pancreatic cancer cells was examined. Replication-deficient recombinant adenoviruses were constructed by the cosmid-adenoviral DNA terminal protein complex method. PANC-1 cells were infected with Ad-pdx-1 or Ad-LacZ. PANC-1 cells that were infected with adenovirus were used in a cell growth assay and a migration assay and for morphologic analysis. Results. Interestingly, 43% of pancreatic cancers were positive for PDX-1 expression, and 57% of pancreatic cancers were negative (normal pancreatic exocrine tissue shows little or no staining for PDX-1). Lymph node metastasis (P = .02) and histologic grade (P = .04) were correlated significantly with PDX1 expression. Patients with positive PDX-1 had a significantly worse prognosis than those patients with negative PDX-1 (P = .02). Importantly, PDX-1 was an independent variable that effected overall survival (P = .03). Pancreatic cancer cell lines showed no PDX-1 expression. There were no significant differences in cell proliferation or morphologic condition between Ad-pdx-1– and Ad-lacZ–infected PANC-1 cells. However, Ad-pdx-1–infected PANC-1 cells did show a significantly higher migration rate than Ad-lacZ–infected PANC-1 cells. Conclusion. Re-expression of PDX-1 may represent a return to a more de-differentiated state by more aggressive pancreatic cancers and may also represent an important new tumor marker for these aggressive cancers. (Surgery 2003;134:260-6.) From the Department of Surgery and Surgical Basic Science, Kyoto University, Kyoto, Japan; and the Department of Surgical Research, Children’s Mercy Hospital, Kansas City, Mo
PANCREATIC DUODENAL HOMEOBOX GENE-1 (PDX-1) is a transcription factor that is important for the regulation of pancreatic endocrine development and for adult islet β-cell function. PDX-1 expression is detected at e8.5 in endoderm cells in the dorsal gut and at e9.5 in dorsal and ventral pancreatic bud and in the duodenal wall between them.1 Recent Presented at the 64th Annual Meeting of the Society of University Surgeons, Houston, Texas, February 12-15, 2003. Supported by a Grant-in-Aid (#15390395) from the Ministry of Education of Japan. Reprint requests: Ryuichiro Doi, MD, Department of Surgery and Surgical Basic Science, Kyoto University, 54 Shogoinkawaracho, Sakyo, Kyoto 606-8507, Japan. © 2003 Mosby, Inc. All rights reserved. 0039-6060/2003/$30.00 + 0 doi:10.1067/msy.2003.231
260 SURGERY
studies have shown that PDX-1 is an essential factor for pancreatic islet cell differentiation, especially in cells of β-cell lineage.2,3 In adults, PDX-1 is found in both pancreatic islet β-cells and duodenal epithelial cells.4,5 PDX-1 regulates a number of genes that are involved in maintaining β-cell identity and function, including insulin, glucose transporter gene-2, glucokinase, and islet amyloid polypeptide; it also regulates somatostatin in δ-cells.5-7 On the other hand, it has been demonstrated that PDX-1 is re-expressed significantly in proliferating ductal cells during pancreatic regeneration.8 Therefore, PDX-1 is thought to be a marker of redifferentiated cells to regain their pluripotency to differentiate into any pancreatic cell type. Based on these collective findings, we hypothesized that PDX-1 could be reexpressed in pancreatic cancer cells, although the
Koizumi et al 261
Surgery Volume 134, Number 2
histogenesis of pancreatic cancer cell has not been clarified. Here, we first analyzed PDX-1 expression and its correlation with clinicopathologic factors in human pancreatic cancer and investigated the effects of enforced expression of PDX-1 in pancreatic cancer cells that lack PDX-1 expression. MATERIAL AND METHODS Patients and tumor samples. Thirty-five patients with pancreatic cancer (ductal adenocarcinoma) that was confirmed by histologic analysis who underwent pancreatic resection at the Department of Surgery and Surgical Basic Science, Kyoto University, between January 1994 and December 2000 were included in this study. The follow-up data of the patients were updated in August 2002. The patients included in this study were 19 men and 16 women, with ages that ranged from 46 to 76 years at the time of the operation (64.9 ± 8.2 years [mean ± SD]). Patients with other pancreatic malignancies such as intraductal papillary mucinous carcinoma, acinar cell carcinoma, or endocrine tumors were excluded. Tumor specimens were collected after the patients provided informed consent, in accordance with the institutional guidelines. Samples were fixed in 4% paraformaldehyde, embedded in paraffin, and cut into consecutive 4-µm–thick sections. Immunohistochemical analysis. The serial sections were deparaffinized in 3 changes of xylene, rehydrated in descending concentrations of ethanol, and washed 3 times for 5 minutes each with doubledistilled water. After rehydration, the sections were placed in 0.01 mol/L sodium citrate buffer (pH 6.0) for 20 minutes at 95°C and then incubated for 30 minutes at room temperature in 0.3% hydrogen peroxide in methanol to block endogenous peroxidase activity. The sections were then incubated for 30 minutes at room temperature with 0.01 mol/L phosphate-buffered saline solution (pH 7.4) that contained 5% normal goat serum and 1% bovine serum albumin (fraction V; Sigma Chemical Co, St. Louis, Mo) followed by overnight incubation at 4°C with polyclonal rabbit antibody against mouse PDX-1 (a gift from Dr C. V. Wright, Vanderbilt University, Nashville, Tenn), which was diluted 1:500 in 0.01 mol/L phosphate-buffered saline solution that contained 2% normal goat serum and 0.1% bovine serum albumin. This antibody that was raised against mouse PDX-1 is known to recognize N-terminus of PDX-1 in both mice and humans. To exclude the possibility of background staining by a secondary antibody, adjacent sections from the same tumors were treated similarly with nonspecific mouse immunoglobulin G. The sections were washed 3
Fig 1. Expression of PDX-1 in pancreatic cancer tissue. Typical photomicrographs of pancreatic cancer tissues that were stained with anti-PDX-1 antibody. The cancer cells are surrounded with yellow dotted lines. PDX-1 was observed in pancreatic cancer cells (brown color), but not in the adjacent mesenchymal cells (original magnification, 200). The nucleus are counterstained with Mayer’s hematoxylin.
Table I. Clinical profiles of patients with pancreatic cancer PDX-1 staining Age <60 y ≥60 y Sex (n) Male Female
Negative*(n)
Positive†(n)
1 14
3 8
11 9
8 7
*n = 20 patients (57.1%). †n = 15 patients (42.9%).
times for 5 minutes in phosphate-buffered saline solution and incubated for 60 minutes with horseradish peroxidase–conjugated anti-rabbit immunoglobulin G (Envision Kit/horseradish peroxidase; DakoCytomation, Kyoto, Japan) as a secondary antibody. The serial sections were stained with hematoxylin and eosin, and the sections with confirmed diagnosis of adenocarcinoma were subjected to PDX-1 expression analysis.
262 Koizumi et al
Surgery August 2003
Table II. Comparison between the expression of PDX-1 and clinicopathologic features of pancreatic cancer PDX-1 Variables pT pN pM Union Internationale Contre le Cancer
Tumor size Grade
Category
Negative (n = 20)
Positive (n = 15)
P value
1-2* 3-4† Negative Positive Negative Positive Stage I Stage II Stage III Stage IVa Stage IVb <4 cm ≥4 cm 1-2§ 3-4
1 9 9 11 15 5 1 6 3 5 5 12 8 17 2
0 15 1 14 13 2 0 1 9 3 2 7 8 7 6
.29 .02‡ .38 .09
.43 .04‡
*Tumor is limited to the pancreas. †Tumor extends out of the pancreas. ‡Significant value. §Well to moderately differentiated adenocarcinoma. Poorly differentiated adenocarcinoma.
Fig 2. Survival curves of the patients with pancreatic cancer. The survival of patients with negative PDX-1 (n = 20) was significantly better than that of patients with positive PDX-1 (n = 15; P = .02, log-rank test).
The expression of PDX-1 was graded into 3 classes: 0 = undetectable; 1 = weak staining; 2 = strong staining. The proportion of positively stained cells among cancer cells was also graded into 3 classes: 0 = none; 1 = 1% to 49%; 2 = 50% to 100%. The staining intensity score was multiplied by the score of positively stained cells to obtain the overall score. The specimens with a score of no <2 were regarded as positive, and the specimens with a score of no >1 as negative. The expression of PDX-1 was evaluated
independently by 2 investigators (M.K.; R.D.) who had no knowledge of the patients’ clinicopathologic features. Cell lines. Seven human pancreatic cancer cell lines (AsPC-1, BxPC-3, Capan-2, CFPAC-1, HPAC, MIAPaCa-2, and PANC-1) and human embryonic kidney 293 cells were maintained in an appropriate medium that was supplemented with 10% fetal bovine serum, 100 U/mL penicillin, and 100 µg/mL streptomycin at 37°C in a humid atmosphere of 5% carbon dioxide/95% air. Western blot analysis. Pancreatic cancer cells were collected into microtubes with a cell scraper and lysed for 60 minutes in phosphorylation inhibitory radioimmunoprecipitation assay buffer that contained 50 mmol/L HEPES (pH 7.0), 250 mmol/L NaCl, 0.1% Nonidet P-40, 1 mmol/L phenylmethylsulfonyl fluoride, and 20 µg/mL gabexate mesilate. They were then sonicated for 20 seconds. Total extracts were centrifuged at 12,000 rpm for 10 minutes at 4°C. The supernatants were collected, were resuspended in the same volume of the gel loading buffer that contained 50 mmol/L Tris-HCl (pH 6.7), 4% sodium dodecylsulfate, 0.02% bromophenol blue, 20% glycerol, and 4% 2-mercaptoethanol, and were boiled at 95°C for 5 minutes. Ten-microgram aliquots that were taken from the total quantity of protein were size-fractionated to a single dimension by sodium dodecylsulfate–polyacrylamide gel electrophoresis (12% sodium dodecylsulfate gel) and
Koizumi et al 263
Surgery Volume 134, Number 2
Table III. Cox’s univariate and multivariate regression analysis Univariate analysis
Multivariate analysis
Variable
LRχ2
P value
Risk ratio (95% CI)
LRχ2
P value
pM (negative or positive) PDX-1 expression (negative or positive)
4.82 4.80
.03* .03*
0.47 (0.23-0.95) 0.53 (0.28—.95)
4.32 4.61
.04 .03*
LRχ2, Likelihood ratio χ2. *Significant value.
transblotted to 0.45-µm polyvinylidene difluoride membrane (Bio-Rad, Richmond, Calif) in a semi-dry electroblot apparatus. The blots were then incubated in the first antibody solution that contained antiPDX-1 (diluted 1:1000) and were applied with the horseradish peroxidase–conjugated anti-rabbit immunoglobulin G as a secondary antibody. Construction of Ad-pdx-1. We used a replicationdeficient recombinant adenovirus in which the E1A, E1B, and E3 regions of human adenovirus type 5 were deleted.9 The recombinant replicationdefective adenoviruses were constructed by the cosmid-adenoviral DNA terminal protein complex method.10,11 Briefly, the PDX-1 coding sequence was inserted into the pAdCAwt cassette cosmid that contains a CAG promoter (chicken β-actin promoter that is associated with cytomegalovirus enhancer), an artificial splice sequence and rabbit β-globulin poly A sequence. pAdCAG-PDX-1 was constructed by subcloning; pAdCAG-PDX-1 and adenoviral DNA-TPC. were co-transfected into 293 cells to produce recombinant adenovirus through homologous recombination. Ad-pdx-1, an adenovirus that contains the PDX-1 gene that is driven directly by the CAG promoter, strongly expresses the PDX-1 gene in all types of infected cells. Control viruses that express Escherichia coli-lacZ (AdlacZ) were also prepared. All of the adenoviral vectors were propagated in 293 cells, purified by 2 rounds of cesium chloride density centrifugation, dialyzed, and stored at –70°C. The titer, expressed as plaque-forming units per milliliter of each viral stock, was determined by plaque assay with 293 cells. All of the vector preparations were demonstrated to be free of replication-competent adenoviruses. The infection of Ad-pdx-1 and Ad-lacZ into PANC-1 was performed at various multiplicities of infection (MOIs) that ranged from 0 to 20. Cell proliferation assay. After the infection of the recombinant adenovirus, cells were maintained in 24-well plates (IWAKI, Tokyo, Japan), and the cell number was counted with a cell counter (Coulter Z1; Beckman-Coulter, Fullerton, Calif). Morphologic changes between Ad-pdx-1– and Ad-lacZ–infected cells were observed with the use
Fig 3. Expression of PDX-1 in pancreatic cancer cell lines. Positive control was COS7 cell infected with Ad-pdx-1 at 20 MOI. PDX-1 was not detected in any pancreatic cancer cell lines by Western blot analysis.
of phase-contrast microscope (Nikon, Tokyo, Japan). Cell migration assay. Cell migration assay was performed in triplicate by the use of 6.5-mm diameter chambers with 8-µm pore filters (Transwell 24-well cell cultures; Costar, Boston, Mass). The filters were dried subsequently with air that was blown into a clean ventilator. PANC-1 cells that were infected with Ad-pdx-1 and Ad-lacZ at various MOIs were suspended at 2 105 cells/mL in DMEM that contained 1% bovine serum albumin (serum-free media), and then 200 µL of the cell suspension was added to the upper chamber. Fetal bovine serum (0.1%) dissolved in 600 µL of media was placed in the lower well. The chambers were then incubated at 37°C for 12 hours in a humid atmosphere of 5% carbon dioxide/95% air. After the incubation, the filters were fixed in 10% acetic acid/90% methanol and stained with hematoxylin and eosin. The upper surfaces of the filters were scraped twice with cotton swabs to remove nonmigrating cells. The experiments were conducted in triplicate wells, and the number of migrating cells in 3 high-power fields per filter was counted microscopically at 400 magnification. Statistical analyses. Survival of the patients was calculated from the date of the surgical procedure to the date of death or latest follow-up. Clinicopathologic characteristics were compared with the expression of PDX-1 with the use of the chi-
264 Koizumi et al
Surgery August 2003
Fig 4. Effect of PDX-1 gene transfection on the proliferation of PANC-1 cells. After the transfection of PDX-1 gene or lacZ at 20 MOI, cell numbers were counted periodically. There was no significant difference in the proliferation rate between PANC-1 cells that were infected with Ad-pdx1 and those that were infected with Ad-lacZ. Values are expressed as mean ± SE.
squared test or the Fisher’s exact probability test. Overall survival curves were plotted according to Kaplan-Meier’s method,12 and the statistical differences were analyzed by log-rank test. To assess the prognostic significance of individual variables and to identify independent predictors of survival, Cox regression analysis and stepwise selection procedure were used. The statistical analyses were performed with JMP software (version 3.1.5; SAS Institute, Cary, NC) and Stat View software (version J-4.5; Abacus Concepts, Berkeley, Calif). Results of proliferation assay and migration assay were analyzed by 1-way analysis of variance with post hoc pairwise comparisons by Fisher’s least significant difference test. A probability value <.05 was considered statistically significant. RESULTS PDX-1 expression in pancreatic cancer tissues. PDX-1 was expressed in pancreatic cancer tissues. In pancreatic cancer, PDX-1 was observed in pancreatic cancer cells and not in the adjacent mesenchymal cells (Fig 1). The normal portion of human pancreas showed limited expression of PDX-1 mainly in β- and δ-cells of islets. According to the criteria described earlier, 15 of 35 pancreatic cancers (43%) were positive, and 20 pancreatic cancers were negative (57%). There were no significant differences between the 2 groups in terms of patient background, such as age and sex (Table I).
Fig 5. Effect of PDX-1 gene transfection on the morphologic condition of PANC-1 cells. The phase-contrast microscopic images of PANC-1 cells are shown. There was no apparent difference in microscopic morphologic condition between PANC-1 cells that were infected with Ad-lacZ (upper panel) and with Ad-pdx-1 (lower panel).
The correlation between PDX-1 expression and clinicopathologic factors (Union Internationale Contre le Cancer pTNM system) was analyzed. Microscopic lymph node metastasis (P = .02) and histologic grade (P = .04) were correlated significantly with PDX-1 expression (Table II). Other variables such as tumor size, location, and distant metastasis were not correlated significantly with PDX-1 expression. Survival curves were produced for patients with PDX-1–positive and PDX-1–negative pancreatic cancers (Fig 2). Median survival time of patients with PDX-1-positive cancers was 14.0 months and that of patients with PDX-1-negative cancers was 21.4 months. The survival rate of the PDX-1 negative group was significantly better than the PDX-1 positive group (P = .02, log-rank test). To determine the independent prognostic value with respect to the overall survival of patients, PDX-1 expression status was included in the Cox
Surgery Volume 134, Number 2
proportional hazard model with all the clinicopathologic factors listed in Table II. By univariate or multivariate analysis, PDX-1 expression and distant metastasis were found to be independent variables that significantly effected overall survival. The result of multivariate analysis showed that the relative risk for death that was based on PDX-1 expression status was calculated as 0.53 (95% CI, 0.28-0.95; P = .03; Table III). PDX-1 expression in pancreatic cancer cell lines. Protein extracts from 7 pancreatic cancer cell lines (AsPC-1, BxPC-3, Capan-2, CFPAC-1, HPAC, MIA PaCa-2, and PANC-1) were examined by Western analysis. PDX-1 was not detected in these pancreatic cancer cell lines (Fig 3). The 293 cells that were infected with recombinant adenovirus carrying PDX-1 gene was used as a positive control that expressed corresponding 46-kd immunoreactivity. Effect of enforced expression of PDX-1 on proliferation and morphologic condition of pancreatic cancer cell lines. PANC-1 cells were infected with Ad-lacZ in an MOI-dependent manner when determined by X-gal staining. Expression of PDX-1 protein in Ad-pdx-1 infected PANC-1 cells was verified by Western blot. PANC-1 cells were infected with Ad-pdx-1 or Ad-lacZ at various MOIs. After being incubated for 2 days, an equal number of cells were seeded in a 24-well plate and further incubated. At day 6 of incubation, the cell number between Adpdx-1 and Ad-lacZ infected PANC-1 cells was not significantly different (Fig 4). Microscopically, there were no differences in morphologic findings between Ad-pdx-1 and Ad-lacZ infected PANC-1 cells (Fig 5). Effect of enforced expression of PDX-1 on migration of pancreatic cancer cell line. Transwell migration assay was performed to test the effects of enforced expression of PDX-1 on the motility of PANC-1 cells (Fig 6). PANC-1 cells that were infected with Ad-pdx-1 or Ad-lacZ at various MOIs were investigated. At an MOI of 5, there was no significant difference between the migration of PANC-1 with Ad-pdx-1 and Ad-lacZ. However, at MOIs of 10, 20, and 50, PANC-1 cells that were infected with Ad-pdx-1 showed significantly increased rate of migration compared to those cells with Ad-lacZ. DISCUSSION We first demonstrated that pancreatic cancer tissue expresses PDX-1. In our series, 43% of pancreatic cancer was positive for PDX-1 expression and 57% was negative. We also observed in the normal human pancreas that the expression of PDX-1 is seen only in some islet cells, mainly in β- and δ-cells. It has been reported that acinar cells
Koizumi et al 265
Fig 6. Effects of PDX-1 gene transfection on the mobility of PANC-1 cells. The migration of PANC-1 cells that were infected with Ad-pdx-1 or Ad-lacZ was tested by Transwell chamber. Transfection with PDX-1 gene induced an MOIdependent increase of migration of PANC-1 cancer cells. Values are expressed as mean ± SE. *P < .05 versus Ad-lacZ by 1-way analysis of variance with post hoc pairwise comparisons by Fisher least significant difference test.
express a very low level of PDX-1.13 Although normal ductal cells of the pancreas do not express PDX-1, several pancreatic regeneration experiments such as duct ligation model,14 partial pancreatectomy model,15 and cellophane wrapping model16 have suggested that ductal cells could be stimulated under certain conditions to differentiate into islet cells and acinar cells. In addition, it has been reported that PDX-1 is re-expressed in proliferating ductal cells during pancreatic regeneration.8 Although it is generally accepted that pancreatic cancer arises from ductal cells of the pancreas that do not express PDX-1, we thought that the re-expression of PDX-1 in pancreatic cancer tissue may represent a return to a more dedifferentiated state. As assumed, lymph node metastasis and histologic grade were correlated significantly with PDX-1 expression of pancreatic cancer. Patients with negative PDX-1 expression tended to show less lymph node metastasis and higher histologic grade. Further, patients with positive PDX-1 expression had a significantly worse survival when compared with patients with negative PDX-1. Importantly, PDX-1 is an independent variable that affects overall survival.
266 Koizumi et al
Contrary to pancreatic cancer tissue, all the human pancreatic cancer cell lines that were tested in this study showed no PDX-1 expression. During the passage of the cancer cell lines, it is possible that they lose PDX-1 expression in response to culture conditions that are considerably different from where they are isolated. Otherwise, cancer cells that express PDX-1 could not be established into cell lines for some reason because we observed that there is heterogeneity in the expression level of PDX-1, even in the same pancreatic cancer tissue. Because pancreatic cancer cell lines were negative for PDX-1 expression, the effects of the enforced expression of PDX-1 gene in pancreatic cancer cells were then investigated. The replication-deficient recombinant adenovirus-mediated gene transfer to PANC-1 cells was performed successfully in an MOIdependent manner. The transfection of PDX-1 gene and expression of PDX-1 in PANC-1 cells were verified by immunocytologic testing, Western blotting, and room temperature polymerase chain reaction. However, PDX-1 brought about no effects on proliferation and morphologic condition of PANC-1 cells. Further, immunocytologic testing of insulin, glucagon, and somatostatin was negative for Ad-pdx1–infected PANC-1 cells. Finally, we examined the effects of PDX-1 gene transfer on the mobility of PANC-1 cells because clinicopathologic analysis revealed metastasis of pancreatic cancer is related to the expression of PDX-1. We showed by migration assay that the transgene of PDX-1 increased the mobility of pancreatic cancer cells in culture. We found that expression of motility-related molecules such as E-cadherin, N-cadherin, neural cell adhesion molecule, and rho were not effected by Ad-pdx-1 infection when determined by Western analysis (data not shown). Increased motility by PDX-1 was observed in pancreatic cancer cells, but the downstream signals that were altered by gene transfer of PDX-1 remain to be elucidated. Many studies have revealed that the genetic profile of pancreatic cancer, that is, high frequent mutations of oncogene like K-ras and tumor suppressor genes (such as p16, p53 and DPC4) have been demonstrated in pancreatic cancer tissues and pancreatic cancer cell lines.17 The alteration of PDX-1 expression could represent one of those genetic changes. In conclusion, our results suggest that the re-expression of PDX-1 may represent a return to a more dedifferentiated state by more aggressive pancreatic cancers and may also represent an important new tumor marker for these aggressive cancers.
Surgery August 2003
REFERENCES 1. Ahlgren U, Jonsson J, Edlund H. The morphogenesis of the pancreatic mesenchyme is uncoupled from that of the pancreatic epithelium in IPF1/PDX1-deficient mice. Development 1996;122:1409-16. 2. Jonsson J, Carlsson L, Edlund T, Edlund H. Insulinpromoter-factor 1 is required for pancreas development in mice. Nature 1994;371:606-9. 3. Stoffers DA, Zinkin NT, Stanojevic V, Clarke WL, Habener JF. Pancreatic agenesis attributable to a single nucleotide deletion in the human IPF1 gene coding sequence. Nat Genet 1997;15:106-10. 4. Miller CP, McGehee RE Jr, Habener JF. IDX-1: a new homeodomain transcription factor expressed in rat pancreatic islets and duodenum that transactivates the somatostatin gene. Embo J 1994;13:1145-56. 5. McKinnon CM, Docherty K. Pancreatic duodenal homeobox-1, PDX-1, a major regulator of beta cell identity and function. Diabetologia 2001;44:1203-14. 6. Watada H, Kajimoto Y, Kaneto H, Matsuoka T, Fujitani Y, Miyazaki J, et al. Involvement of the homeodomaincontaining transcription factor PDX-1 in islet amyloid polypeptide gene transcription. Biochem Biophys Res Commun 1996;229:746-51. 7. Itkin-Ansari P, Demeterco C, Bossie S, de la Tour DD, Beattie GM, Movassat J, et al. PDX-1 and cell-cell contact act in synergy to promote delta-cell development in a human pancreatic endocrine precursor cell line. Mol Endocrinol 2000;14:814-22. 8. Fernandes A, King LC, Guz Y, Stein R, Wright CV, Teitelman G. Differentiation of new insulin-producing cells is induced by injury in adult pancreatic islets. Endocrinology 1997;138:1750-62. 9. Ishii A, Hagiwara Y, Saito Y, Yamamoto K, Yuasa K, Sato Y, et al. Effective adenovirus-mediated gene expression in adult murine skeletal muscle. Muscle Nerve 1999;22:592-9. 10. Saito I, Oya Y, Yamamoto K, Yuasa T, Shimojo H. Construction of nondefective adenovirus type 5 bearing a 2.8kilobase hepatitis B virus DNA near the right end of its genome. J Virol 1985;54:711-9. 11. Miyake S, Makimura M, Kanegae Y, Harada S, Sato Y, Takamori K, et al. Efficient generation of recombinant adenoviruses using adenovirus DNA-terminal protein complex and a cosmid bearing the full-length virus genome. Proc Natl Acad Sci U S A 1996;93:1320-4. 12. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457-81. 13. Wu KL, Gannon M, Peshavaria M, Offield MF, Henderson E, Ray M, et al. Hepatocyte nuclear factor 3beta is involved in pancreatic beta-cell-specific transcription of the PDX-1 gene. Mol Cell Biol 1997;17:6002-13. 14. Wang RN, Kloppel G, Bouwens L. Duct- to islet-cell differentiation and islet growth in the pancreas of duct-ligated adult rats. Diabetologia 1995;38:1405-11. 15. Bonner-Weir S, Baxter LA, Schuppin GT, Smith FE. A second pathway for regeneration of adult exocrine and endocrine pancreas: a possible recapitulation of embryonic development. Diabetes 1993;42:1715-20. 16. Rafaeloff R, Barlow SW, Rosenberg L, Vinik AI. Expression of Reg gene in the Syrian golden hamster pancreatic islet regeneration model. Diabetologia 1995;38:906-13. 17. Hilgers W, Kern SE. Molecular genetic basis of pancreatic adenocarcinoma. Genes Chromosomes Cancer 1999;26:1-12.
Background. During pancreatic development, pancreatic duodenal homeobox gene-1 (PDX-1) is expressed in pancreatic duct cells that have the potential to differentiate into islets. Therefore, PDX-1 is thought to be a marker of de-differentiated cells with the capacity to redifferentiate into several pancreatic cell types. We analyzed PDX-1 expression in human pancreatic cancer specimens, pancreatic cancer cell lines, and the effects of forced expression of PDX-1 in pancreatic cancer cells. Methods. Thirty-five pancreatic adenocarcinomas were immunohistochemically stained with a polyclonal rabbit antibody against mouse PDX-1. Correlations with tumor characteristics were made with chisquared analysis. The influence of clinicopathologic factors on survival was assessed. The expression of PDX-1 in pancreatic cancer cells was examined. Replication-deficient recombinant adenoviruses were constructed by the cosmid-adenoviral DNA terminal protein complex method. PANC-1 cells were infected with Ad-pdx-1 or Ad-LacZ. PANC-1 cells that were infected with adenovirus were used in a cell growth assay and a migration assay and for morphologic analysis. Results. Interestingly, 43% of pancreatic cancers were positive for PDX-1 expression, and 57% of pancreatic cancers were negative (normal pancreatic exocrine tissue shows little or no staining for PDX-1). Lymph node metastasis (P = .02) and histologic grade (P = .04) were correlated significantly with PDX1 expression. Patients with positive PDX-1 had a significantly worse prognosis than those patients with negative PDX-1 (P = .02). Importantly, PDX-1 was an independent variable that effected overall survival (P = .03). Pancreatic cancer cell lines showed no PDX-1 expression. There were no significant differences in cell proliferation or morphologic condition between Ad-pdx-1– and Ad-lacZ–infected PANC-1 cells. However, Ad-pdx-1–infected PANC-1 cells did show a significantly higher migration rate than Ad-lacZ–infected PANC-1 cells. Conclusion. Re-expression of PDX-1 may represent a return to a more de-differentiated state by more aggressive pancreatic cancers and may also represent an important new tumor marker for these aggressive cancers. (Surgery 2003;134:260-6.) From the Department of Surgery and Surgical Basic Science, Kyoto University, Kyoto, Japan; and the Department of Surgical Research, Children’s Mercy Hospital, Kansas City, Mo
PANCREATIC DUODENAL HOMEOBOX GENE-1 (PDX-1) is a transcription factor that is important for the regulation of pancreatic endocrine development and for adult islet β-cell function. PDX-1 expression is detected at e8.5 in endoderm cells in the dorsal gut and at e9.5 in dorsal and ventral pancreatic bud and in the duodenal wall between them.1 Recent Presented at the 64th Annual Meeting of the Society of University Surgeons, Houston, Texas, February 12-15, 2003. Supported by a Grant-in-Aid (#15390395) from the Ministry of Education of Japan. Reprint requests: Ryuichiro Doi, MD, Department of Surgery and Surgical Basic Science, Kyoto University, 54 Shogoinkawaracho, Sakyo, Kyoto 606-8507, Japan. © 2003 Mosby, Inc. All rights reserved. 0039-6060/2003/$30.00 + 0 doi:10.1067/msy.2003.231
260 SURGERY
studies have shown that PDX-1 is an essential factor for pancreatic islet cell differentiation, especially in cells of β-cell lineage.2,3 In adults, PDX-1 is found in both pancreatic islet β-cells and duodenal epithelial cells.4,5 PDX-1 regulates a number of genes that are involved in maintaining β-cell identity and function, including insulin, glucose transporter gene-2, glucokinase, and islet amyloid polypeptide; it also regulates somatostatin in δ-cells.5-7 On the other hand, it has been demonstrated that PDX-1 is re-expressed significantly in proliferating ductal cells during pancreatic regeneration.8 Therefore, PDX-1 is thought to be a marker of redifferentiated cells to regain their pluripotency to differentiate into any pancreatic cell type. Based on these collective findings, we hypothesized that PDX-1 could be reexpressed in pancreatic cancer cells, although the
Koizumi et al 261
Surgery Volume 134, Number 2
histogenesis of pancreatic cancer cell has not been clarified. Here, we first analyzed PDX-1 expression and its correlation with clinicopathologic factors in human pancreatic cancer and investigated the effects of enforced expression of PDX-1 in pancreatic cancer cells that lack PDX-1 expression. MATERIAL AND METHODS Patients and tumor samples. Thirty-five patients with pancreatic cancer (ductal adenocarcinoma) that was confirmed by histologic analysis who underwent pancreatic resection at the Department of Surgery and Surgical Basic Science, Kyoto University, between January 1994 and December 2000 were included in this study. The follow-up data of the patients were updated in August 2002. The patients included in this study were 19 men and 16 women, with ages that ranged from 46 to 76 years at the time of the operation (64.9 ± 8.2 years [mean ± SD]). Patients with other pancreatic malignancies such as intraductal papillary mucinous carcinoma, acinar cell carcinoma, or endocrine tumors were excluded. Tumor specimens were collected after the patients provided informed consent, in accordance with the institutional guidelines. Samples were fixed in 4% paraformaldehyde, embedded in paraffin, and cut into consecutive 4-µm–thick sections. Immunohistochemical analysis. The serial sections were deparaffinized in 3 changes of xylene, rehydrated in descending concentrations of ethanol, and washed 3 times for 5 minutes each with doubledistilled water. After rehydration, the sections were placed in 0.01 mol/L sodium citrate buffer (pH 6.0) for 20 minutes at 95°C and then incubated for 30 minutes at room temperature in 0.3% hydrogen peroxide in methanol to block endogenous peroxidase activity. The sections were then incubated for 30 minutes at room temperature with 0.01 mol/L phosphate-buffered saline solution (pH 7.4) that contained 5% normal goat serum and 1% bovine serum albumin (fraction V; Sigma Chemical Co, St. Louis, Mo) followed by overnight incubation at 4°C with polyclonal rabbit antibody against mouse PDX-1 (a gift from Dr C. V. Wright, Vanderbilt University, Nashville, Tenn), which was diluted 1:500 in 0.01 mol/L phosphate-buffered saline solution that contained 2% normal goat serum and 0.1% bovine serum albumin. This antibody that was raised against mouse PDX-1 is known to recognize N-terminus of PDX-1 in both mice and humans. To exclude the possibility of background staining by a secondary antibody, adjacent sections from the same tumors were treated similarly with nonspecific mouse immunoglobulin G. The sections were washed 3
Fig 1. Expression of PDX-1 in pancreatic cancer tissue. Typical photomicrographs of pancreatic cancer tissues that were stained with anti-PDX-1 antibody. The cancer cells are surrounded with yellow dotted lines. PDX-1 was observed in pancreatic cancer cells (brown color), but not in the adjacent mesenchymal cells (original magnification, 200). The nucleus are counterstained with Mayer’s hematoxylin.
Table I. Clinical profiles of patients with pancreatic cancer PDX-1 staining Age <60 y ≥60 y Sex (n) Male Female
Negative*(n)
Positive†(n)
1 14
3 8
11 9
8 7
*n = 20 patients (57.1%). †n = 15 patients (42.9%).
times for 5 minutes in phosphate-buffered saline solution and incubated for 60 minutes with horseradish peroxidase–conjugated anti-rabbit immunoglobulin G (Envision Kit/horseradish peroxidase; DakoCytomation, Kyoto, Japan) as a secondary antibody. The serial sections were stained with hematoxylin and eosin, and the sections with confirmed diagnosis of adenocarcinoma were subjected to PDX-1 expression analysis.
262 Koizumi et al
Surgery August 2003
Table II. Comparison between the expression of PDX-1 and clinicopathologic features of pancreatic cancer PDX-1 Variables pT pN pM Union Internationale Contre le Cancer
Tumor size Grade
Category
Negative (n = 20)
Positive (n = 15)
P value
1-2* 3-4† Negative Positive Negative Positive Stage I Stage II Stage III Stage IVa Stage IVb <4 cm ≥4 cm 1-2§ 3-4
1 9 9 11 15 5 1 6 3 5 5 12 8 17 2
0 15 1 14 13 2 0 1 9 3 2 7 8 7 6
.29 .02‡ .38 .09
.43 .04‡
*Tumor is limited to the pancreas. †Tumor extends out of the pancreas. ‡Significant value. §Well to moderately differentiated adenocarcinoma. Poorly differentiated adenocarcinoma.
Fig 2. Survival curves of the patients with pancreatic cancer. The survival of patients with negative PDX-1 (n = 20) was significantly better than that of patients with positive PDX-1 (n = 15; P = .02, log-rank test).
The expression of PDX-1 was graded into 3 classes: 0 = undetectable; 1 = weak staining; 2 = strong staining. The proportion of positively stained cells among cancer cells was also graded into 3 classes: 0 = none; 1 = 1% to 49%; 2 = 50% to 100%. The staining intensity score was multiplied by the score of positively stained cells to obtain the overall score. The specimens with a score of no <2 were regarded as positive, and the specimens with a score of no >1 as negative. The expression of PDX-1 was evaluated
independently by 2 investigators (M.K.; R.D.) who had no knowledge of the patients’ clinicopathologic features. Cell lines. Seven human pancreatic cancer cell lines (AsPC-1, BxPC-3, Capan-2, CFPAC-1, HPAC, MIAPaCa-2, and PANC-1) and human embryonic kidney 293 cells were maintained in an appropriate medium that was supplemented with 10% fetal bovine serum, 100 U/mL penicillin, and 100 µg/mL streptomycin at 37°C in a humid atmosphere of 5% carbon dioxide/95% air. Western blot analysis. Pancreatic cancer cells were collected into microtubes with a cell scraper and lysed for 60 minutes in phosphorylation inhibitory radioimmunoprecipitation assay buffer that contained 50 mmol/L HEPES (pH 7.0), 250 mmol/L NaCl, 0.1% Nonidet P-40, 1 mmol/L phenylmethylsulfonyl fluoride, and 20 µg/mL gabexate mesilate. They were then sonicated for 20 seconds. Total extracts were centrifuged at 12,000 rpm for 10 minutes at 4°C. The supernatants were collected, were resuspended in the same volume of the gel loading buffer that contained 50 mmol/L Tris-HCl (pH 6.7), 4% sodium dodecylsulfate, 0.02% bromophenol blue, 20% glycerol, and 4% 2-mercaptoethanol, and were boiled at 95°C for 5 minutes. Ten-microgram aliquots that were taken from the total quantity of protein were size-fractionated to a single dimension by sodium dodecylsulfate–polyacrylamide gel electrophoresis (12% sodium dodecylsulfate gel) and
Koizumi et al 263
Surgery Volume 134, Number 2
Table III. Cox’s univariate and multivariate regression analysis Univariate analysis
Multivariate analysis
Variable
LRχ2
P value
Risk ratio (95% CI)
LRχ2
P value
pM (negative or positive) PDX-1 expression (negative or positive)
4.82 4.80
.03* .03*
0.47 (0.23-0.95) 0.53 (0.28—.95)
4.32 4.61
.04 .03*
LRχ2, Likelihood ratio χ2. *Significant value.
transblotted to 0.45-µm polyvinylidene difluoride membrane (Bio-Rad, Richmond, Calif) in a semi-dry electroblot apparatus. The blots were then incubated in the first antibody solution that contained antiPDX-1 (diluted 1:1000) and were applied with the horseradish peroxidase–conjugated anti-rabbit immunoglobulin G as a secondary antibody. Construction of Ad-pdx-1. We used a replicationdeficient recombinant adenovirus in which the E1A, E1B, and E3 regions of human adenovirus type 5 were deleted.9 The recombinant replicationdefective adenoviruses were constructed by the cosmid-adenoviral DNA terminal protein complex method.10,11 Briefly, the PDX-1 coding sequence was inserted into the pAdCAwt cassette cosmid that contains a CAG promoter (chicken β-actin promoter that is associated with cytomegalovirus enhancer), an artificial splice sequence and rabbit β-globulin poly A sequence. pAdCAG-PDX-1 was constructed by subcloning; pAdCAG-PDX-1 and adenoviral DNA-TPC. were co-transfected into 293 cells to produce recombinant adenovirus through homologous recombination. Ad-pdx-1, an adenovirus that contains the PDX-1 gene that is driven directly by the CAG promoter, strongly expresses the PDX-1 gene in all types of infected cells. Control viruses that express Escherichia coli-lacZ (AdlacZ) were also prepared. All of the adenoviral vectors were propagated in 293 cells, purified by 2 rounds of cesium chloride density centrifugation, dialyzed, and stored at –70°C. The titer, expressed as plaque-forming units per milliliter of each viral stock, was determined by plaque assay with 293 cells. All of the vector preparations were demonstrated to be free of replication-competent adenoviruses. The infection of Ad-pdx-1 and Ad-lacZ into PANC-1 was performed at various multiplicities of infection (MOIs) that ranged from 0 to 20. Cell proliferation assay. After the infection of the recombinant adenovirus, cells were maintained in 24-well plates (IWAKI, Tokyo, Japan), and the cell number was counted with a cell counter (Coulter Z1; Beckman-Coulter, Fullerton, Calif). Morphologic changes between Ad-pdx-1– and Ad-lacZ–infected cells were observed with the use
Fig 3. Expression of PDX-1 in pancreatic cancer cell lines. Positive control was COS7 cell infected with Ad-pdx-1 at 20 MOI. PDX-1 was not detected in any pancreatic cancer cell lines by Western blot analysis.
of phase-contrast microscope (Nikon, Tokyo, Japan). Cell migration assay. Cell migration assay was performed in triplicate by the use of 6.5-mm diameter chambers with 8-µm pore filters (Transwell 24-well cell cultures; Costar, Boston, Mass). The filters were dried subsequently with air that was blown into a clean ventilator. PANC-1 cells that were infected with Ad-pdx-1 and Ad-lacZ at various MOIs were suspended at 2 105 cells/mL in DMEM that contained 1% bovine serum albumin (serum-free media), and then 200 µL of the cell suspension was added to the upper chamber. Fetal bovine serum (0.1%) dissolved in 600 µL of media was placed in the lower well. The chambers were then incubated at 37°C for 12 hours in a humid atmosphere of 5% carbon dioxide/95% air. After the incubation, the filters were fixed in 10% acetic acid/90% methanol and stained with hematoxylin and eosin. The upper surfaces of the filters were scraped twice with cotton swabs to remove nonmigrating cells. The experiments were conducted in triplicate wells, and the number of migrating cells in 3 high-power fields per filter was counted microscopically at 400 magnification. Statistical analyses. Survival of the patients was calculated from the date of the surgical procedure to the date of death or latest follow-up. Clinicopathologic characteristics were compared with the expression of PDX-1 with the use of the chi-
264 Koizumi et al
Surgery August 2003
Fig 4. Effect of PDX-1 gene transfection on the proliferation of PANC-1 cells. After the transfection of PDX-1 gene or lacZ at 20 MOI, cell numbers were counted periodically. There was no significant difference in the proliferation rate between PANC-1 cells that were infected with Ad-pdx1 and those that were infected with Ad-lacZ. Values are expressed as mean ± SE.
squared test or the Fisher’s exact probability test. Overall survival curves were plotted according to Kaplan-Meier’s method,12 and the statistical differences were analyzed by log-rank test. To assess the prognostic significance of individual variables and to identify independent predictors of survival, Cox regression analysis and stepwise selection procedure were used. The statistical analyses were performed with JMP software (version 3.1.5; SAS Institute, Cary, NC) and Stat View software (version J-4.5; Abacus Concepts, Berkeley, Calif). Results of proliferation assay and migration assay were analyzed by 1-way analysis of variance with post hoc pairwise comparisons by Fisher’s least significant difference test. A probability value <.05 was considered statistically significant. RESULTS PDX-1 expression in pancreatic cancer tissues. PDX-1 was expressed in pancreatic cancer tissues. In pancreatic cancer, PDX-1 was observed in pancreatic cancer cells and not in the adjacent mesenchymal cells (Fig 1). The normal portion of human pancreas showed limited expression of PDX-1 mainly in β- and δ-cells of islets. According to the criteria described earlier, 15 of 35 pancreatic cancers (43%) were positive, and 20 pancreatic cancers were negative (57%). There were no significant differences between the 2 groups in terms of patient background, such as age and sex (Table I).
Fig 5. Effect of PDX-1 gene transfection on the morphologic condition of PANC-1 cells. The phase-contrast microscopic images of PANC-1 cells are shown. There was no apparent difference in microscopic morphologic condition between PANC-1 cells that were infected with Ad-lacZ (upper panel) and with Ad-pdx-1 (lower panel).
The correlation between PDX-1 expression and clinicopathologic factors (Union Internationale Contre le Cancer pTNM system) was analyzed. Microscopic lymph node metastasis (P = .02) and histologic grade (P = .04) were correlated significantly with PDX-1 expression (Table II). Other variables such as tumor size, location, and distant metastasis were not correlated significantly with PDX-1 expression. Survival curves were produced for patients with PDX-1–positive and PDX-1–negative pancreatic cancers (Fig 2). Median survival time of patients with PDX-1-positive cancers was 14.0 months and that of patients with PDX-1-negative cancers was 21.4 months. The survival rate of the PDX-1 negative group was significantly better than the PDX-1 positive group (P = .02, log-rank test). To determine the independent prognostic value with respect to the overall survival of patients, PDX-1 expression status was included in the Cox
Surgery Volume 134, Number 2
proportional hazard model with all the clinicopathologic factors listed in Table II. By univariate or multivariate analysis, PDX-1 expression and distant metastasis were found to be independent variables that significantly effected overall survival. The result of multivariate analysis showed that the relative risk for death that was based on PDX-1 expression status was calculated as 0.53 (95% CI, 0.28-0.95; P = .03; Table III). PDX-1 expression in pancreatic cancer cell lines. Protein extracts from 7 pancreatic cancer cell lines (AsPC-1, BxPC-3, Capan-2, CFPAC-1, HPAC, MIA PaCa-2, and PANC-1) were examined by Western analysis. PDX-1 was not detected in these pancreatic cancer cell lines (Fig 3). The 293 cells that were infected with recombinant adenovirus carrying PDX-1 gene was used as a positive control that expressed corresponding 46-kd immunoreactivity. Effect of enforced expression of PDX-1 on proliferation and morphologic condition of pancreatic cancer cell lines. PANC-1 cells were infected with Ad-lacZ in an MOI-dependent manner when determined by X-gal staining. Expression of PDX-1 protein in Ad-pdx-1 infected PANC-1 cells was verified by Western blot. PANC-1 cells were infected with Ad-pdx-1 or Ad-lacZ at various MOIs. After being incubated for 2 days, an equal number of cells were seeded in a 24-well plate and further incubated. At day 6 of incubation, the cell number between Adpdx-1 and Ad-lacZ infected PANC-1 cells was not significantly different (Fig 4). Microscopically, there were no differences in morphologic findings between Ad-pdx-1 and Ad-lacZ infected PANC-1 cells (Fig 5). Effect of enforced expression of PDX-1 on migration of pancreatic cancer cell line. Transwell migration assay was performed to test the effects of enforced expression of PDX-1 on the motility of PANC-1 cells (Fig 6). PANC-1 cells that were infected with Ad-pdx-1 or Ad-lacZ at various MOIs were investigated. At an MOI of 5, there was no significant difference between the migration of PANC-1 with Ad-pdx-1 and Ad-lacZ. However, at MOIs of 10, 20, and 50, PANC-1 cells that were infected with Ad-pdx-1 showed significantly increased rate of migration compared to those cells with Ad-lacZ. DISCUSSION We first demonstrated that pancreatic cancer tissue expresses PDX-1. In our series, 43% of pancreatic cancer was positive for PDX-1 expression and 57% was negative. We also observed in the normal human pancreas that the expression of PDX-1 is seen only in some islet cells, mainly in β- and δ-cells. It has been reported that acinar cells
Koizumi et al 265
Fig 6. Effects of PDX-1 gene transfection on the mobility of PANC-1 cells. The migration of PANC-1 cells that were infected with Ad-pdx-1 or Ad-lacZ was tested by Transwell chamber. Transfection with PDX-1 gene induced an MOIdependent increase of migration of PANC-1 cancer cells. Values are expressed as mean ± SE. *P < .05 versus Ad-lacZ by 1-way analysis of variance with post hoc pairwise comparisons by Fisher least significant difference test.
express a very low level of PDX-1.13 Although normal ductal cells of the pancreas do not express PDX-1, several pancreatic regeneration experiments such as duct ligation model,14 partial pancreatectomy model,15 and cellophane wrapping model16 have suggested that ductal cells could be stimulated under certain conditions to differentiate into islet cells and acinar cells. In addition, it has been reported that PDX-1 is re-expressed in proliferating ductal cells during pancreatic regeneration.8 Although it is generally accepted that pancreatic cancer arises from ductal cells of the pancreas that do not express PDX-1, we thought that the re-expression of PDX-1 in pancreatic cancer tissue may represent a return to a more dedifferentiated state. As assumed, lymph node metastasis and histologic grade were correlated significantly with PDX-1 expression of pancreatic cancer. Patients with negative PDX-1 expression tended to show less lymph node metastasis and higher histologic grade. Further, patients with positive PDX-1 expression had a significantly worse survival when compared with patients with negative PDX-1. Importantly, PDX-1 is an independent variable that affects overall survival.
266 Koizumi et al
Contrary to pancreatic cancer tissue, all the human pancreatic cancer cell lines that were tested in this study showed no PDX-1 expression. During the passage of the cancer cell lines, it is possible that they lose PDX-1 expression in response to culture conditions that are considerably different from where they are isolated. Otherwise, cancer cells that express PDX-1 could not be established into cell lines for some reason because we observed that there is heterogeneity in the expression level of PDX-1, even in the same pancreatic cancer tissue. Because pancreatic cancer cell lines were negative for PDX-1 expression, the effects of the enforced expression of PDX-1 gene in pancreatic cancer cells were then investigated. The replication-deficient recombinant adenovirus-mediated gene transfer to PANC-1 cells was performed successfully in an MOIdependent manner. The transfection of PDX-1 gene and expression of PDX-1 in PANC-1 cells were verified by immunocytologic testing, Western blotting, and room temperature polymerase chain reaction. However, PDX-1 brought about no effects on proliferation and morphologic condition of PANC-1 cells. Further, immunocytologic testing of insulin, glucagon, and somatostatin was negative for Ad-pdx1–infected PANC-1 cells. Finally, we examined the effects of PDX-1 gene transfer on the mobility of PANC-1 cells because clinicopathologic analysis revealed metastasis of pancreatic cancer is related to the expression of PDX-1. We showed by migration assay that the transgene of PDX-1 increased the mobility of pancreatic cancer cells in culture. We found that expression of motility-related molecules such as E-cadherin, N-cadherin, neural cell adhesion molecule, and rho were not effected by Ad-pdx-1 infection when determined by Western analysis (data not shown). Increased motility by PDX-1 was observed in pancreatic cancer cells, but the downstream signals that were altered by gene transfer of PDX-1 remain to be elucidated. Many studies have revealed that the genetic profile of pancreatic cancer, that is, high frequent mutations of oncogene like K-ras and tumor suppressor genes (such as p16, p53 and DPC4) have been demonstrated in pancreatic cancer tissues and pancreatic cancer cell lines.17 The alteration of PDX-1 expression could represent one of those genetic changes. In conclusion, our results suggest that the re-expression of PDX-1 may represent a return to a more dedifferentiated state by more aggressive pancreatic cancers and may also represent an important new tumor marker for these aggressive cancers.
Surgery August 2003
REFERENCES 1. Ahlgren U, Jonsson J, Edlund H. The morphogenesis of the pancreatic mesenchyme is uncoupled from that of the pancreatic epithelium in IPF1/PDX1-deficient mice. Development 1996;122:1409-16. 2. Jonsson J, Carlsson L, Edlund T, Edlund H. Insulinpromoter-factor 1 is required for pancreas development in mice. Nature 1994;371:606-9. 3. Stoffers DA, Zinkin NT, Stanojevic V, Clarke WL, Habener JF. Pancreatic agenesis attributable to a single nucleotide deletion in the human IPF1 gene coding sequence. Nat Genet 1997;15:106-10. 4. Miller CP, McGehee RE Jr, Habener JF. IDX-1: a new homeodomain transcription factor expressed in rat pancreatic islets and duodenum that transactivates the somatostatin gene. Embo J 1994;13:1145-56. 5. McKinnon CM, Docherty K. Pancreatic duodenal homeobox-1, PDX-1, a major regulator of beta cell identity and function. Diabetologia 2001;44:1203-14. 6. Watada H, Kajimoto Y, Kaneto H, Matsuoka T, Fujitani Y, Miyazaki J, et al. Involvement of the homeodomaincontaining transcription factor PDX-1 in islet amyloid polypeptide gene transcription. Biochem Biophys Res Commun 1996;229:746-51. 7. Itkin-Ansari P, Demeterco C, Bossie S, de la Tour DD, Beattie GM, Movassat J, et al. PDX-1 and cell-cell contact act in synergy to promote delta-cell development in a human pancreatic endocrine precursor cell line. Mol Endocrinol 2000;14:814-22. 8. Fernandes A, King LC, Guz Y, Stein R, Wright CV, Teitelman G. Differentiation of new insulin-producing cells is induced by injury in adult pancreatic islets. Endocrinology 1997;138:1750-62. 9. Ishii A, Hagiwara Y, Saito Y, Yamamoto K, Yuasa K, Sato Y, et al. Effective adenovirus-mediated gene expression in adult murine skeletal muscle. Muscle Nerve 1999;22:592-9. 10. Saito I, Oya Y, Yamamoto K, Yuasa T, Shimojo H. Construction of nondefective adenovirus type 5 bearing a 2.8kilobase hepatitis B virus DNA near the right end of its genome. J Virol 1985;54:711-9. 11. Miyake S, Makimura M, Kanegae Y, Harada S, Sato Y, Takamori K, et al. Efficient generation of recombinant adenoviruses using adenovirus DNA-terminal protein complex and a cosmid bearing the full-length virus genome. Proc Natl Acad Sci U S A 1996;93:1320-4. 12. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457-81. 13. Wu KL, Gannon M, Peshavaria M, Offield MF, Henderson E, Ray M, et al. Hepatocyte nuclear factor 3beta is involved in pancreatic beta-cell-specific transcription of the PDX-1 gene. Mol Cell Biol 1997;17:6002-13. 14. Wang RN, Kloppel G, Bouwens L. Duct- to islet-cell differentiation and islet growth in the pancreas of duct-ligated adult rats. Diabetologia 1995;38:1405-11. 15. Bonner-Weir S, Baxter LA, Schuppin GT, Smith FE. A second pathway for regeneration of adult exocrine and endocrine pancreas: a possible recapitulation of embryonic development. Diabetes 1993;42:1715-20. 16. Rafaeloff R, Barlow SW, Rosenberg L, Vinik AI. Expression of Reg gene in the Syrian golden hamster pancreatic islet regeneration model. Diabetologia 1995;38:906-13. 17. Hilgers W, Kern SE. Molecular genetic basis of pancreatic adenocarcinoma. Genes Chromosomes Cancer 1999;26:1-12.