KAI1 gene is differently expressed in papillary and pancreatic cancer: Influence on metastasis

Abstract

AIM: To compare KAI1 in cancer of papilla of Vater and pancreas to evaluate whether there are differences in biologic behavior which might account for prognosis.

METHODS: We compared the expression in 24 papillay and 29 pancreatic cancers using Northern blot analysis, immunochemical assay and in situ hybridization, and investigated whether early diagnosis or molecular differences predict the outcome in these tumor entities.

RESULTS: By Northern blot analysis there is no statistical difference of KAI1 levels in normal and cancerous papilla. No association between KAI1 mRNA expression and tumor stage or tumor differentiation was found in the tumors. By immunohistochemical assay, KAI1 staining in cytoplasm of papillary cancer cells was similar to that of normal papillary cells. By in situ hybridization, the results of KAI1 mRNA expression in normal and cancerous papilla were similar to those with immunohistochemical assay. The normal and cancerous pancreas tissues were also analyzed by the methods used in papillary samples.

CONCLUSION: Although the biologic roles of KAI1 have not been clarified, our results suggest that KAI1 may restrict the progression of malignant papillary cancer, but its expression might not have any effect on the characteristics of papillary tumor, whereas by the analysis of KAI1 gene, its reduced expression is closely related to the progression and metastases of pancreatic cancer.

Key Words: pancreatic neoplasms; papillary neoplasms; KAI1 gene; immunohistochemistry; in situ hybridization; blotting, northern

INTRODUCTION

Metastasis is a complex process, involving local invasion, inward and outward infiltration of tumor cells, and decreased host immunological responses[1-4]. In carcinoma of papilla of Vater, lymph node metastases are present at the time of diagnosis in 31%-52% of the patients and by radical tumor resection, 5 year-survival rates of reached 21%-61%[3]. In contrast to cancers of the papilla of Vater, cancer of pancreas has a dismal prognosis. Most pancreatic cancers have already had local or distant metastasis which restricts palliative surgical procedures. Therefore, median survival period of 4-6 mo in most patients with pancreatic cancer remains in reality. The aggressive growth behaviour of pancreatic cancer results in a death/incidence ratio of approximately 0.99 in the United States and also in most European countries[4,5].

The reason why pancreatic cancer has a prognosis different from that of the papilla of Vater is not known. It was postulated that earlier diagnosis due to jaundice accounts mainly for the better prognosis of papilla of Vater cancer patients. However, it is not known whether differences in tumor biologic behavior play any role in the difference in prognosis.

KAI1 has been identified to influence the metastatic ability of a various gastrointestinal cancer cells or other tumors[6-22]. The gene is located on human chromosome 11p 11.2[23]. Recently, it was reported that decreased KAI1 mRNA expression correlated with the metastases of pancreatic cancer[24]. After transfer of the KAI1 gene into highly metastatic prostatic cancer cells, the metastatic ability was suppressed, whereas their primary tumor is not affected[23]. These results suggested that decreased KAI1 expression is involved in the progression to metastatic cancers. However, whether changes in expression of tumor metastases influencing gene account for the better prognosis of papilla of Vater cancers is not known. Currently, it is believed that the better prognosis of papilla of Vater cancer patients compared with pancreatic cancer patients caused no differences in tumor growth and metastasis formation but from the earlier establishment of the diagnosis. Therefore, in the present studies, we compared KAI1 in papilla of Vater and pancreatic cancer patients to evaluate whether there exist differences in tumor biological behavior which might account for the differences in prognosis.

MATERIALS AND METHODS

Patients

Nine normal human papilla of Vater tissue specimens (4 females, 5 males, mean age ± SD: 35.7 ± 6.5 years, ranging 23-43); and 16 normal human pancreatic tissue specimens (6 females, 10 males; mean age ± SD: 36.6 ± 10.7 years, ranging 10-47) were obtained through a multiorgan donor program in 9 cases. The whole pancreas was obtained with the duodenum and the papilla of Vater was completely resected. Tissue specimens from 24 patients with carcinoma of the papilla of Vater (9 females and 15 males, mean age ± SD: 58.4 ± 14.1 years, ranging 16-81) were obtained following a Whipple’s operation. The diagnosis of cancer of the papilla of Vater was confirmed by histopathological analysis. According to the TNM classification[25] there were 2 stage I, 9 stage II, 11 stage III and 2 stage IV tumors. Pancreatic cancer tissues were obtained from 14 female and 15 male patients after operation. The median age of the pancreatic cancer patients was 64 years, ranging 37-78. The partial duodeno pancreatectomy (Whipple’s operation) and distal pancreatectomy was performed in 26 and 3 patients, respectively. According to the TNM classification of the International Union Against Cancer[25], the patients were in stage I, 3 cases, stage II, 10 cases and stage III 16 cases; their gradings were well differentiated in 7, moderately in 17 and poorly differentiated in 5.

Tissue sampling

For RNA extraction and Northern blot analysis, normal and tumor specimens were frozen in liquid nitrogen immediately after surgical removal and stored at -80 °C until use. Additionally, freshly removed normal and cancerous tissue samples were immediately fixed in formaldehyde solution for 12 h-24 h and paraffin-embedded for in situ hybridization and immunohistochemical assay.

Northern blot analysis

Total RNA was extracted by the single-step guanidinium isothiocyanate method[26] size-fractionated on 1.2% agarose 1.8 mol/L formaldehyde gels[26], and stained with ethidium bromide for verification of RNA integrity and loading equivalency. The RNA was electro-transferred onto nylon membranes (Gene Screen, Du Pont International, Regensdorf, Switzerland) and cross-linked by UV irradiation. For hybridization a digoxigenin-(DIG) labeled KAI cRNA probe and ³²P-labeled 7S cDNA probe were used.

Prehybridization for KAI1 was performed for two hours at 65 °C in a buffer containing 50% formamide, 5 × SSC (sodium chloride/sodium citrate buffer), 2% blocking reagent (Boehringer Mannheim GmbH, Mannheim, Germany) and 0.1% Nauroylsarcosine. After adding the DIC-labeled KAI1 antisense probe, hybridization was carried out at 65 °C for 18 h. The filters were washed after wards for 5 min, in 2 × SSC, and 0.1% SDS. At room temperature, followed by two washes at 68 °C for 15 min each in 0.1 × SSC and 0.1% SDS. The filters were then incubated in 20 mL blocking buffer (1% blocking reagent in 100 mmol/L maleic-acid, 150 mmol/L sodium chloride and 175 mmol/L sodium hydroxide) containing 1 μL anti-DIG alkaline phosphatase antibodies (Boehringer Mannheim) for 30 min, washed with blocking buffer for 15 min, and incubated with 4 μL CDP-Star (25 mmol; Boehringer Mannheim). The membranes were then exposed to X-ray films for 15 s. At room temperature as previously reported[27,28]. In order to assess equivalent RNA loading, the membranes were rehybridized with the ³²P-labeled mouse 7 S cDNA probe that cross-hybridizes with human 7 S RNA[28-30] to verify equivalent RNA loading. The membranes were prehybridized for 4-8 h at 42 °C in a buffer that contained 50% formamide, 1% sodium dodecyl sulfate, 0.75 mol/L NaCl, 5 mmol/L EDTA, 5 × Denhardts solution, 100 mg/L salmon sperm DNA, 10% Dextran sulfate, and 50 mmol/L sodium phosphate (pH7.4). The hybridization was carried out at 42 °C for 18 h by adding the labeled cDNA probe 1 × 10⁵ cpm/mL. The blots were rinsed twice in 2 × SSC at room temperature and washed three times at 55 °C in 0.2 × SSC and 2% SDS.

The blots were then exposed at -80 °C to Fuji X-ray films with intensifying screens for 24-48 h. The intensity of the KAI1 and 7S signals was quantified by video densitometric analysis (Biorad 620, New York, USA) as previously reported[10,26]. The ratio between the KAI1 signal and the corresponding 7S signal was calculated for each sample.

Immunohistochemistry

From each normal and cancer tissue sample, three tissue sections were examined. After deparaffinizing and hydrating, tissue sections were submerged for 15 min in Tris-buffered saline (10 mmol/L Tris HCl, 0.85% NaCl, pH7.4) containing 0.1% (vol/vol) Triton X-100 and briefly rinsed 3 times for 1 min-2 min in TBS solution. Following incubation in methanol containing 0.6% hydrogen peroxide for 30 min to block endogenous peroxidase activity, the slides were covered with 10% normal goat serum at 23 °C for 30 min then incubated overnight with mouse monoclonal anti-human KAI1 antibody (antibody C33, kindly supplied by Dr. J.C. Barrett, Institute of Environmental Health Sciences, National Institutes of Health, Research Triangle Park, NC, USA). After washing with TBS buffer, biotinylated goat anti-mouse immunoglobulin and streptavidin-peroxidase complex (Kirkegaard & Perry Laboratories, Gaithersburg, MD) were added at 23 °C for 45 and 30 min, respectively, followed by incubation with a 3,3’-diaminobenzidine tetrahydrochloride and hydrogen peroxide mixture. The slides were counterstained with Mayer’s- hematoxylin.

In situ hybridization

In situ hybridization was performed as reported previously in detail using DIG-labeled cRNA probes[31]. Tissue sections of normal and cancerous sam ples were processed always simultaneously. In addition, the consecutive tissue slides were processed, one slide each was incubated with the sense probe, a antisense probe. The prehybridization, hybridization and washing conditions were the same for pancreatic and papilla of Vater tissue samples. Four μm tissue sections were deparaffinized, rehydrated, and incubated in 0.2 mol/L HCl for 20 min. After washed with 2 × SSC, the tissues were permeabilized with proteinase K at a concentration of 35 mg/L for 15 min at 37 °C. After post-fixation with 4% paraformaldehyde in saline phosphate buffer (5 min) and washing in 2 × SSC, the sections were prehybridized for 1 h at 60 °C, in a buffer containing 50% formamide (v/v), 4 × SSC, 2 × Denhardts reagent and 250 μg RNA/mL. Hybridization was performed overnight at the same temperature in 50% (v/v) formamide, 4 × SSC, 2 × Denhardt’s reagent, 500 μg RNA/mL and 1 0% dextran sulfate (v/v).The final concentration of the DIG-labeled KAI1 probes (antisense or sense) was approximately 0.5 ng/μL. After hybridization, excess probe was removed by washing in 2 × SSC, and by RNase treatment: 100 U/mL RNase TI and 0.2 U/mL RNase DNase-free (Boehringer Mannhein) at 37 °C for 30 min. Washings were performed at 60 °C for pancreatic tissue slides and 63 °C in 2 × SSC (10 min), and twice in 0.2 × SSC (10 min each). Afterward the sections were incubated with an anti-digoxigenin antibody conjugated with alkaline phosphatase (Boehringer Mannheim). For the color reaction 5-bromo-4 chlorl-3-indolyl phosphatase and nitro blue tetrazolium (Sigma, Buchs, Switzerland) were used.

Pretreatment of the slides with RNase abolished the hybridization signals, and hybridization with the sense probes corresponding to the antisense probes failed to produce an in situ hybridization signal.

The in situ hybridization signals were semiquantitatively evaluated by two independent observers blind to patient status followed by resolution of any differences by joint review and consultation with a third observer. The in situ hybridization results were scored as previously described[32]; (-) no detectable signal; (+) weak detectable signal; (++) moderate detectable signal; and (+++) strong detectable signal.

Preparation of KAI1 sense and antisense cRNA probes

To prepare digoxigenin-labeled KAI1 cRNA probes for Northern blot analysis and in situ hybridization, a 500 bp fragment of human KAI1 cDNA was subcloned into the pCR-II vector (Invitrogen, San Diego, USA), which contains promoters for DNA-dependent SP6 and T7 RNA polymerases. After linearization of the plasmid, the antisense KAI1 probes were transcribed using SP6 poly merase and the Ribomax System (Promega Biotechnology, Madison, W1, USA). A DIG-labeled KAI1 cRNA probe was used for Northern blot analysis and in situ hybridization. To evaluate the specificity of the in situ hybridization reaction, DIG-labeled sense probes of KAI1 were generated after linearization of the plasmid with Bam-HI and Hind-III, respectively and in vitro transcription with T7 polymerase and the Ribomax System (Promega Biotechnology, Madison, WI, USA)[33]. For the in situ hybridization experiments, the KAI1 antisense and sense probes were shortened to a length of approximately 150 bases[27].

Preparation of 7S cDNA probe

To verify equivalent RNA loading on Northern blot membranes, all filters were rehybridized with a murine 190 bp BamH I fragment or 7S cDNA which cross -hybridizes with human 7S RNA as previously reported[28,29,34]. The 7S cDNA probe was radiolabeled with [alpha ³²P]dCTP (3000Ci/mmol; DuPont, Boston, USA) using a random primer labeling system (Pharmacia Biotech AG, Dubendorf, Switzerland)[28,29].

Statistical analysis

Results were expressed as median and range or as mean ± SD. For statistical analysis the Mann-Whitney U test and the Chi-square test were used. Significance was defined as P < 0.05.

RESULTS

KAI1 mRNA expression by Northern blot analysis

Qapillary samples Measureable KAI1 mRNA signal was detected in 67% of the normal papillary tissue samples. In papilla of Vater cancer samples KAI1 mRNA expression was present in 46%. Densitometric analysis of the expression signals revealed a 1.07 fold (not significant) increase in KAI1 mRNA levels in papillary cancer compared with the normal controls when all cancerous tissue samples were included. When only cancer samples with increased KAI1 mRNA expression level were statistically analyzed, the increase was 1.60-fold (not significant). The papillary cancer samples with lymph node metastases present at the time of tumor resection were compared with papillary tumor samples in lymph node-free metastases, no difference (P > 0.05, Table 1) in KAI1 mRNA levels was found.

Table 1 KAI1 gene expression in papillary and pancreatic cancer.

Carcinoma of papilla of Vater			Pancreatic cancer
Clinical data	Total No.	P	Clinical data	Total No.	P
Sex
Femal	9		Femal	14
Male	15	NS	Male	15	NS
Age (years)
< 60	10		< 60	8
> 60	14	NS	> 60	21	NS
Tumor-staging
T1-T2	11		T1-T2	13
T3-T4	13	NS	T3-T4	16	0.001
Differentiation
			G1-G2	24
			G3-G4	5	NS

NS: no significance

Pancreas The expression signals by densitometric analysis exhibited 2.2-fold increase (P < 0.05) in KAI1 levels in pancreatic cancer compared with the normal controls when all cancerous tissue samples were included. When only-cancer samples with increased KAI1 expression levels were statistically analyzed, the increase was 2.8-fold (P < 0.05). There was a significant negative correlation between KAI1 (r = 0.59) mRNA levels and the tumor staging (P < 0.0007, Table 1). Primary pancreatic cancer samples in which lymph node metastasis were present at the time of tumor resection (stage III) exhibited significantly lower KAI1 mRNA levels compared with primary tumor samples without lymph node metastasis at the time of tumor resection (stage I/II; Table 1, P < 0.005).

Comparison between papillary and pancreatic cancer samples The mRNA levels of KAI1 gene in pancreatic cancers were higher than that in papillary cancers (P = 0.03). Statistical analysis of the densitometric data revealed that these differences were statistically significant.

Immunohistochemistry of KAI1

Papilla of Vater In the normal samples of papilla of Vater, moderate to strong KAI1 immunoreactivity was present in the cytoplasm of epithelial cells. In addition, strong membranous immunostaining was found for KAI1 in the normal samples of papilla of Vater. Papillary cancer cells exhibited a similar staining like normal samples of papilla of Vater. However, only a few cancer cells showed membraneus KAI1 immunoreactivity in tumors with or without metastases.

Pancreas KAI1 immunoreactivity was weakly detectable in the normal pancreas in a few acinar and ductal cells and strong KAI1 immunostaining was present in all pancreatic islets. Other results were similar to that seen with in situ hybridization.

KAI1 mRNA expression by in situ hybridization

To localize the exact site and cellular distribution of KAI1 mRNA expression, in situ hybridization was performed in normal and cancerous tissue samples.

Papilla of Vater In the normal papilla of Vater, moderate to strong KAI1 mRNA staining was present in the cytoplasm of most epithelial cells. Lymphocytes in the submucosal areas of the normal papilla of Vater exhibited weak or moderate expression of KAI1 mRNA. In the papilla of Vater cancer samples, a similar intensity of KAI1 mRNA staining was present in the cytoplasm of cancer cells compared with normal epithelial cells. Fibroblasts of the connective tissue surrounding papillary cancer cells showed weak to moderate KAI1 mRNA staining. The intensity of KAI1 mRNA signals in samples of primary papilla of Vater cancer with or without metastases was not different.

Pancreas In normal pancreatic tissue samples only very faint KAI1 mRNA staining was found in a focal pattern in a few acinar and ductal cells of the normal pancreas. In contrast, pancreatic cancer cells demonstrated moderate to strong cytoplasmic KAI1 mRNA staining. However, the staining intensity for KAI1 mRNA in the pancreatic cancer cells was dependent on the tumor staging. Cancer cells of tumors without lymph node or distant metastases (Stage I/II) exhibited stronger KAI1 mRNA staining than primary tumor samples in which lymph node metastases were present (stage III). The stroma cells surrounding the pancreatic cancer lesions exhibited low levels of KAI1 mRNA expression. Similarly, only low KAI1 mRNA level was found in lymphocytes infiltrated in the cancer samples.

DISCUSSION

Patients with carcinoma of papilla of Vater have the best prognosis in all patients with periampullary carcinomas[35]. The most possible reason for it is the aggressive surgery carried out in early tumor stage or the tumor possessed a different biological local growth behavior and spreading characteristic. In contrast to cancer of papilla of Vater, pancreatic head carcinoma has a dismal prognosis and it was believed previously to be due to late diagnosis[36]. It is very difficult to distinguish patient with jaundice between papillary cancer and pancreatic cancer. Hence, the search for sensitive and reliable prognostic factor is of primary importance. The level of KAI1 mRNA expression were associated with clinical parameters influencing metastasis in gastrointestinal cancers in different ways[24,28,29]. Tumor invasion and metastasis might be contributed by the down-regulation of KAI1[24], whereas no significant KAI1 mRNA expression was found in gastric cancer and esophageal cancer with poor prognosis and metastases[29]. These findings suggest that the effects of KAI1 on metastasis depend on the underlying malignancy.

In papillary cancer, no information about KAI1 has been obtained recently. By Northern blot analysis our results showed that the KAI1 mRNA level was not different in primary papillary carcinoma with or without lymph node metastases.

KAI1 immunoreactivity was mainly located in the membrane and/or cytoplasm of normal and/or cancerous papillary epithelia cells. By in situ hybridization, the results of KAI1 mRNA expression in normal and cancerous papilla were similar to those with immunohistochemical assay. Our findings indicate that KAI1 in papillary cancers exhibits different KAI1 mRNA expression patterns from prostate and pancreatic cancers[23,29]. This suggests that KAI1 may limit the progression only in malignant papillary cancers, and this expression might have no effect on the characteristics of papillary tumors. Furthermore, the divergent expression patterns of KAI1 in the investigated cancer tissuess how that it plays a role in the formation and metastases of these malignancies, that are different from those in previously analyzed tumors of the prostate, pancreas, breast, or the lung[23-25,35]. However, by immunohistochemical assay and in situ hybridization, there was some heterogeneity of KAI1 mRNA expression levels in the cancers of papilla.

Our results showed that reduced expression of KAI1 mRNA might be regarded as the potential candidates for predicting tumor metastasis and invasion in pancreatic cancer. A few prospective randomized trials were reported that decreased expression of KAI1 gene was associated with metastasis in pancreatic cancer[24]. By contrast to Northern blot analysis, similar expressions of KAI1 mRNA in immunohistochemistry and in situ hybridization were found. Down-regulation expression was only present in the patients with lymph nodes and distant metastases. We suggested that the expression of KAI1 gene could predict the prognosis of patients with pancreatic cancer. It is well known that the accumulation of genetic alterations causes the progression of tumors[37]. But there are very few reports on its relationship with the mechanism of metastases in pancreatic cancer. K-ras, P16, p53, DPC4 and BRCA2 gene influenced the tumor-suppressive pathways in pancreatic cancer. Underscoring the multigenic nature of cancer, and tumor PX101, having alterations identified in the five genes examplified the extent of accumulation of genetic alterations[38].

Cancer is a disease of gene alterations accumulated in several genes resulting in the development of the tumor[39-50]. Multiple genetic lesions with either activating dominant oncogenes or inactivating tumor-suppressor genes have been recognized in human pancreatic cancer[38]. Although our results suggest that expression of KAI1 gene is not associated with papillary cancer, it is closely related to metastases and prognosis of pancreatic cancer, and serves as a biological marker beneficial to diagnosis and treatment of pancreatic cancer. By further study on the pathogenesis of metastasis, the problem of how to prevent the early metastasis of pancreatic cancer will be solved.