J Cancer Res Clin Oncol (2004) 130: 521–526 DOI 10.1007/s00432-004-0563-x

O R I GI N A L P A P E R

Lirong Peng Æ Jinying Ning Æ Ling Meng Chengchao Shou

The association of the expression level of protein tyrosine phosphatase PRL-3 protein with liver metastasis and prognosis of patients with colorectal cancer Received: 20 September 2003 / Accepted: 22 April 2004 / Published online: 6 May 2004
Springer-Verlag 2004

Abstract Purpose: To investigate PRL-3 protein expression in normal colorectal epithelia and colorectal cancers with monoclonal antibody (MAb) against PRL-3. Methods: MAb against PRL-3 was prepared with the hybridoma technique, and its specificity was confirmed with ELISA and Western blotting assays. The expression of PRL-3 protein in normal colorectal epithelia and colorectal cancers was examined by immunohistochemistry assay. Logistic regression and survival analysis were performed to determine the clinical significance of PRL-3 expression. Results: MAb 3B6 against PRL-3 was obtained and showed high specificity. PRL-3 protein was expressed in two of 28 (7.1%) normal colorectal epithelia, 21 of 88 (23.9%) primary colorectal cancers, 22 of 41 (53.7%) metastatic lymph nodes and eight of 12 (66.7%) liver metastases, respectively. The PRL-3 expression rates of metastases were significantly higher than those of primary colorectal cancers and normal colorectal epithelia (P <0.05). PRL-3 expression was significantly associated with the liver metastasis of colorectal cancer (P = 0.004) and tended to shorten survival time (P = 0.0145). Conclusions: This is the first study demonstrating that PRL-3 is a potential marker for liver metastasis of colorectal cancer and negatively influences the prognosis of colorectal cancer patients.

This work was supported by the State Key Basic Research Program of China (G19980512). Key Project of Beijing Scientific Technology Committee (H020220020310), and Peking University Key Program of Oncology. L. Peng Æ J. Ning Æ L. Meng Æ C. Shou (&) Department of Biochemistry and Molecular Biology, Beijing Institute for Cancer Research, Peking University School of Oncology, 1 Da-Hong-Luo-Chang Street, Western District 100034 Beijing, China E-mail: [email protected] Tel.: +86-10-66160960 Fax: +86-10-66175832

Keywords PRL-3 Æ Monoclonal antibody Æ Colorectal cancer Æ Liver metastasis

Introduction In most colorectal cancer patients, the primary neoplasm has given rise to metastasis at the time of diagnosis. Metastasis is main cause of death in cancer patients, because of its resistant to chemotherapy or radiotherapy (Susan and Schumacher 2001). Therefore, studying the molecular basis of metastasis and identifying metastasis specific markers would provide information for the diagnosis and treatment of malignant diseases (Lindblom and Linder 1996). PRL-3 is a newly identified metastasis-related gene, which codes a 22 KDa nonclassical protein tyrosine phosphatase with a C-terminal prenylation motif. It is located at the cytoplasmic membrane when prenylated and in the nucleus when nonpreylated. PRL-3 has at least 75% amino-acid sequence similarity with PRL-1 and PRL-2, the other two members of the PRL family (Zeng et al. 1998). Overexpression of PRL-3 has been found to transform human embryonic kidney cell HEK293 and increase HEK293 cell growth (Matter et al. 2001). Saha et al. reported that PRL-3 mRNA expression was consistently increased in the liver metastases of colorectal cancers, suggesting that PRL-3 was associated with colorectal cancer metastasis (Saha et al. 2001). Recently, Zeng et al. showed its causal role in metastasis in Chinese hamster ovary cell CHO with PRL-3 stably transfected, which exhibited enhanced invasive activity and induced metastatic tumor formation in mice (Zeng et al. 2003). However, the PRL-3 protein level in normal colorectal epithelia and colorectal cancers has not been determined; and its clinical significance is unclear. In this study, we prepared MAb specific for PRL-3 with hybridoma technique first, because of the commercial antibody unavailable. Then PRL-3 protein expression in colorectal

522

cancers was examined by immunohistochemistry assay. Its prognostic potential for colorectal cancer patients was also investigated. This is the first report to examine PRL-3 protein expression in normal colorectal epithelia and colorectal cancers. The result suggests that PRL-3 is a predictor of liver metastasis and poor prognosis for patients with colorectal cancer.

Materials and methods Patients and tissue specimens Tissues of 88 colorectal cancers, 28 adjacent normal colorectal epithelia (at least 5 cm distant from the tumor edge), 41 metastatic lymph nodes, and 12 liver metastases were obtained from the Department of Pathology, Peking University School of Oncology. Each sample had been fixed in formalin, routinely processed, and embedded in paraffin. Specimens were diagnosed histopathologically and staged according to the TNM-International Union Against Cancer classification system (Sobin and Witterind 1997).

trations of MAb were added (50 ll/well) and incubated for 2 h at 37 C. Following washing the wells, bound antibodies was detected by ELISA using horseradish peroxidase-conjugated goat anti-mouse antibody (DAKO, Carpinteria, Calif., USA). For the Western blotting assay, the total proteins of COS-7 cells transfected with pcDNA3-PRL-3 by lipofectin (Invitrogen, Life Technologies Pacific, Beijing) and nontransfected were extracted. Equal amounts of proteins were subjected to 12% sodium dodecyl sulfatepolyacrylamide electrophoresis (SDS-PAGE) and electroblotted to a nitrocellulose membrane (Amersham Pharmacia Biotech, Uppsala, Sweden). Nonspecific binding was blocked with 5% nonfat milk in phosphatebuffered saline overnight at 4 C. Then the nitrocellulose membrane was incubated with MAb 3B6 for 1 h at room temperature, followed with horseradish peroxidase-conjugated goat anti-mouse IgG for 1 h at room temperature. After being washed, it was processed for enhanced chemoluminescence according to the kit instructions (Amersham Pharmacia Biotech). The result was documented on an X-ray film. Immunohistochemistry

Immunization and cell fusion Fusion protein GST-PRL-1, -2, and -3 was obtained by gene cloning, recombinating and expressing in Escherichia coli (data not shown). The purified fusion proteins were used as immunogen or screening antigen in enzyme-linked immunosorbent assay (ELISA) for the identification of monoclonal antibody against PRL-3. BALB/c mice (purchased from the Animal Center of the National Medical Academe of China) were subcutaneously immunized with purified fusion protein GST-PRL-3. Spleen cells from BALB/c mice immunized with GST-PRL-3 were fused with myeloma cell SP2/0; and hybridomas secreting antibodies against PRL-3 were selected, prepared in ascites, and purified according to standard procedure and kit instructions (Faseks de et al. 1980; Fuller et al. 2002).

Paraffin sections (4-lm thick) were deparaffinized, rehydrated, and placed in 3% hydrogen peroxide solution to inhibit endogenous peroxidase activity. Then they were blocked with 1% bovine serum albumin for 1 h and subsequently incubated with MAb 3B6 at a concentration of 2.5 lg/ml overnight at 4 C. The reaction product was visualized by incubation with secondary antibody EnVision+ (DAKO) for 20 min and diaminobenzidine (DAB, Sigma Chemical, St. Lous, Mo., USA) for 1 min at room temperature. Counterstaining was performed with hematoxylin. For the negative control, primary antibody was changed for preimmune mouse serum. The number of tumor cells or normal colorectal glandular epithelial cells was counted under ten randomly chosen microscopic fields (Olympus Optical, Tokyo, Japan) by two independent pathologists, and any specimen with more than 5% positivestaining cells was classified as positive.

Specificity analysis of MAb against PRL-3 Statistical analysis Two methods were used to analyze the specificity of MAb 3B6. In ELISA analysis, GST-PRL-1 and GST-PRL-2 were used to eliminate antibodies cross reacting with PRL1 and PRL-2. Unrelated fusion protein GST-VEGF was used to eliminate antibodies against GST protein. At first, Polystyrene microplates (Maxisorb/NUNC, Roskilde, Denmark) were coated with GST-PRL-1, -2, -3 and GSTVEGF (5 lg/ml, respectively) in 0.05 M bicarbonate buffer (pH 9.6) overnight at 4 C, respectively. After washing with phosphate-buffered saline containing 0.05% Tween 20, the plates were blocked with 1% bovine serum albumin at room temperature for 2 h. Various concen-

Student’s t-test and Pearson’s v2 test were performed to evaluate the possible differences of PRL-3 expression in different groups of patients, and the associations of PRL-3 expression with clinicopathologic factors. The correlation of PRL-3 with liver metastasis was studied by logistic regression. Survival analyses were calculated by the Kaplan-Meier method and Cox proportional hazard regression model. All statistical tests were twosided and carried out with the SPSS statistical software package (SPSS 10.0, SPSS, Chicago, Ill., USA). P values less than 0.05 were considered statistically significant.

523

Results Preparation and characterization of MAb against PRL-3 MAb 3B6, highly specific for PRL-3, was obtained after cell fusion and subcloning. The concentrations of 0.1 lg/ml, 0.5 lg/ml, 1.0 lg/ml, 5.0 lg/ml, and 10.0 lg/ml of 3B6 were used to react with 5 lg/ml of GST-PRL-1, -2, -3, and GST-VEGF in ELISA. The result showed that the binding of MAb 3B6 with PRL-3 was dose-dependent and reached a saturation at 5 lg/ml of 3B6. There were no cross reactions between 3B6 and GST-PRL-1, -2, and -VEGF, respectively, even at 10 lg/ml of 3B6 in this assay, indicating that MAb 3B6 is highly specific to PRL-3 (Fig. 1). Because PRL-3 used in immunization was obtained form prokaryotic cells, it may be somewhat different from the actual structure of PRL-3 in eukaryotic cells. In order to test the binding ability of MAb 3B6 with PRL-3 in eukaryotic cells, we constructed the eukaryotic expression vector pcDNA3-PRL-3 and transfected it into COS-7 cells. The total proteins of COS-7 cells transfected with pcDNA3-PRL-3 and nontransfected were extracted and subjected to Western blotting with 3B6 as primary antibody. The result showed a reaction band at about 22 KDa in the transfected COS-7 cell lane, but not in the nontransfected COS-7 cell lane, which suggested that MAb 3B6 has specific binding with PRL-3 protein in eukaryotic cells (Fig. 2). Expression of PRL-3 protein in normal colorectal epithelia, colorectal cancers, and metastases PRL-3 expression in normal colorectal epithelia, colorectal cancers, and metastases was detected by immunohistochemistry assay. Cytoplasmic staining in

Fig. 2 Western blotting analysis of the binding ability of MAb 3B6 with PRL-3 protein expressed in eukaryotic cells. Lane 1, transfected pcDNA3-PRL-3 COS-7 cells show a band at about 22 KDa; Lane 2, nontransfected COS-7 cells show no reaction band; Lane 3, protein weight marker (reduced IgG)

colorectal cancer cells and individual normal glandular epithelia was evidenced by the presence of granular immunoreaction products (Fig. 3). PRL-3 was expressed in 2 of 28 (7.1%) normal colorectal epithelia, 21 of 88 (23.9%) primary colorectal cancers, 22 of 41 (53.7%) metastatic lymph nodes, and 8 of 12 (66.7%) liver metastases, respectively. PRL-3 expression intensity and positive rate were not significantly different between normal colorectal epithelia and colorectal cancers, or between metastatic lymph nodes and liver metastases (P>0.05). However, PRL-3 expression rate was significantly higher in metastases when compared with normal colorectal epithelia and primary colorectal cancers (P<0.05). To more precisely determine the differences of PRL-3 protein expression in colorectal cancers, we compared the PRL-3 expression in 36 primary colorectal cancers with that in their corresponding metastatic lymph nodes. The positive rate of metastatic lymph nodes was significantly higher than that of primary colorectal cancers (Table 1). Nine of the ten PRL-3immunoreactivity-positive primary cancers had positive corresponding metastatic lymph nodes. Of the 26 negative primary cancers, 11 had positive corresponding metastatic lymph nodes. The above data indicated that PRL-3 was associated with the metastasis of colorectal cancer. Prognostic significance of PRL-3 expression

Fig. 1 Affinity and specificity analysis of MAb 3B6 with ELISA assay. Fusion protein GST-PRL-1, -2, -3, and GST-VEGF of 5 lg/ml were used to react with MAb 3B6 at the concentrations of 0.1 lg/ml, 0.5 lg/ml, 1.0 lg/ml, 5.0 lg/ml, and 10.0 lg/ml

The correlations of PRL-3 immunostaining with clinicopathologic factors, which were obtained from the retrospective records of 88 patients with colorectal cancer, are summarized in Table 2. PRL-3 expression was positively correlated with liver metastasis (logistic regression; P = 0.004). Although there were significant differences in PRL-3 expression between primary cancers and metastatic lymph nodes, we could not predict whether lymph node metastasis was present or not just according to the

524

Fig. 3A–F Immunohistochemical staining of PRL-3 protein in normal colorectal epithelia, primary cancers and metastases. Positive staining tissues with granular immunoreaction products in cytoplasm (original magnification ·400). Negative tissues (original magnification ·250). A positive normal colorectal epithelium; B positive primary colorectal cancer; C positive metastatic lymph node; D positive liver metastasis. Scale bar = 10 lm; E negative normal colorectal epithelium; F preimmune serum negative control. Scale bar = 10 lm

PRL-3 expression in primary cancer (Table 2; P>0.05), but PRL-3 positive in primary cancer could indicate liver metastasis (Table 2; P = 0.004). In addition, the correlations of liver metastasis with the clinical variables of age, sex, histological differentiation, depth of invasion, lymph node metastasis, and PRL-3 expression in primary colorectal cancer were analyzed by logistic regression to identify predictors for liver metastasis. The results demonstrated that PRL-3 expression was the only independent predictor for liver metastasis (Table 3). Moreover, we used the Kaplan-Meier survival curve and Cox proportional hazard model to examine the prognosis significance of PRL-3 expression for outcome. The Kaplan-Meier survival curve showed that PRL-3 expression tended to shorten the survival time of colorectal cancer patients (Fig. 4; P = 0.0145). Univariate analyses identified invasion depth, lymph node metastaTable 1 Expression of PRL-3 protein in 36 primary colorectal cancers and their corresponding metastatic lymph nodes. Statistical analysis was performed by the chi-square test. P<0.05 was considered significant Tissue type

PRL-3 expression

Positive v2 rate

P

Positive (n) Negative (n) Colorectal 10 cancer Metastatic 20 lymph node

26

27.8%

16

55.6%

5.714 0.031

sis, liver metastasis, and PRL-3 positive expression as predictors for poor outcome. Multivariate analyses of these variables indicated that liver metastasis was an independent prognostic factor, but PRL-3 expression was not an independent factor (Table 3). Table 2 Correlations of PRL-3 protein expression with the clinicopathologic factors in 88 patients with primary colorectal cancer. Statistical analysis was performed using the chi-square test. P <0.05 was considered significant Factor

PRL-3 expression

Positive P rate

Positive (n) Negative (n) Gender Male Female

11 10

33 34

25.0% 22.7%

Age £ 60 years >60 years

6 15

34 33

15.0% 31.3%

Histological differentiation Good Moderate Poor

>0.999

0.085

0.901 10 5 6

35 16 16

22.2% 23.8% 27.2%

Depth of invasion Muscularis propria 5 Adventitia 13 Subserosa 3

17 46 4

22.7% 22.0% 42.9%

TNM stage £ III IV

63 4

14.8% 63.6%

0.489

<0.0001 13 8

Lymph node metastasis Present Absent

0.317 14 7

35 32

28.6% 18.0%

Liver metastasis Present Absent

8 13

4 63

66.7% 16.5%

<0.0001

525 Table 3 Univariate and multivariate analyses of 88 patients with colorectal cancer. Binary logistic regression was performed to identify factors for liver metastasis and the Cox proportional hazard model was performed to identify factors for outcome. P<0.05 was considered significant Variable

P value on logistic regression Univariate

Age (years) Gender Histological differentiation Depth of invasion Lymph node metastasis PRL-3 expression Variable Age (years) Gender Histological differentiation Depth of invasion Lymph node metastasis PRL-3 expression Liver metastasis

Multivariate

0.369 0.713 >0.999 0.568 0.131 0.305 0.027 0.198 0.730 0.739 0.001 0.004 P value on survival analyses Univariate Multivariate 0.351 0.599 0.649 0.809 0.198 0.931 0.002 0.205 0.005 0.132 0.021 0.024 <0.0001 0.646

Fig. 4 Clinical outcome of patients with colorectal cancer after surgery with a 34-month follow-up period. Patients with PRL-3 protein expression have significantly shorter survival time than those without (log rank test; P = 0.0145)

Discussion Protein tyrosine phosphatases (PTPs) and protein tyrosine kinases play key roles in the regulation of cellular growth, differentiation, cell cycle, cell-cell communication, and other activities by the dephosphorylation and phosphorylation of tyrosine (Hunter 2000; Neel and Tonks 1997; Tan et al. 1993). Because many protein tyrosine kinases are proto-oncogenes, PTPs, which reverse the actions of protein tyrosine kinases, were initially thought to have the potential of tumor suppression (Di Cristofano et al. 2000; Brown-Shimer et al.

1992). However, more and more studies show that some PTPs might have positive roles in carcinogenesis. For example, PTPa activates Src kinases by dephosphorylating inhibitory C-terminal pTyr, and thereby promotes cell transformation (Su et al. 1999; Ponniah et al. 1999; Zheng et al. 1992). There is evidence that PTPs Cdc25A and Cdc25B may be oncogenes, which play important roles in cell-cycle regulation by removing the inhibitory phosphates from tyrosine and threonine residues of cyclin-dependent kinases (Galaktionov et al. 1995; Cangi et al. 2000). Recent studies show that PRL-3 may also be an oncogene that plays a causal role in cancer metastasis (Matter et al. 2001; Saha et al. 2001; Zeng et al. 2003), but the differences of PRL-3 protein expression in colorectal cancers have not been established at the protein level. Because there is no commercial monoclonal antibody available, the antibodies used in the previous studies on PRLs were polyclonal antibodies. The polyclonal antibodies might be not specific because of the high homology between PRL-1, -2, and -3. So we prepared MAb 3B6, highly specific against PRL-3, first, which provided a good tool for the following research. Next, PRL-3 protein expression in normal colorectal epithelia, primary colorectal cancers, and metastases was examined by immunohistochemistry assay. In this study, we demonstrated that PRL-3 expression rates were higher in metastatic lymph nodes and liver metastases than in primary colorectal cancers and normal epithelia, which accorded with the mRNA expression level of PRL-3 reported before (Saha et al. 2001). Statistical analyses indicated that PRL-3 expression was a predictor of liver metastasis (logistic regression; P = 0.004); and patients expressing PRL-3 showed a significantly shorter survival time. In this study, we did not find significant correlation of PRL-3 expression with lymph node metastasis (P >0.05), which may be due to an insufficient number of lymph nodes excised in surgical operation or the omission of micro-lymph-node metastasis by routine histopathologic examination. In conclusion, for the first time, we determined PRL3 expression in normal colorectal epithelia, primary colorectal cancers, and metastases at the protein level and demonstrated a close association between PRL-3 expression and liver metastasis of colorectal cancers. This study suggests that PRL-3 immunohistochemical assessment can be used to predict liver metastasis of patients with colorectal cancer. In the future, this may provide us with new therapeutic strategies for treatment. However, the process of metastasis is highly complex. It is the result of the accumulation of gene mutations and expression change in different places and times, rather than the alteration of a single gene (Ma 2001). Therefore, further experiments will be required to clarify the exact function of PRL-3 in metastasis. Acknowledgements The authors thank Dr. Jiyou Li, Dr. Jing Gu, Dr. Jiafu Ji, and Dr. Jingsheng He of Beijing Cancer Hospital for providing the tissue specimens

526

References Brown-Shimer S, Johnson KA, Hill DE, Bruskin AM (1992) Effect of protein tyrosine phosphatase 1B expression on transformation by the human neu oncogene. Cancer Res 52:478–482 Cangi MG, Cukor B, Soung P, Signoretti S, Moreira G Jr, Ranashinge M, Cady B, Pagano M, Loda M (2000) Role of Cdc25A phosphatase in human breast cancer. J Clin Invest 106:753–761 Di Cristofano A, Pandolfi PP (2000) The multiple roles of PTEN in tumor suppression. Cell 100:387–390 de St. Groth SF, Scheidegger D (1980) Production of monoclonal antibodies strategy and tactics. J Immunol Methods 35:1–21 Fuller SA, Takahaski M, Hurrell JG (2002) Immunology. In: Ausubel FM, Brent R, Kingston RE, Moore D, Seidman JG, Smith J, Struhl K (eds) Current protocols in molecular biology, vol. 2. Wiley, New York, pp 11.1.1–11.11.4 Galaktionov K, Lee AK, Eckstein J, Draetta G, Meckler J, Loda M, Beach D (1995) Cdc25 phosphatases as potential human oncogenes. Science 269:1575–1577 Hunter T (2000) Signaling-2000 and beyond. Cell 100:113–127 Lindblom A, Linder S (1996) The metastatic phenotype—prognostic implications. Crit Rev Oncol Hematol 24:71–96 Ma D (2001) Tumor metastsis. In: Zeng YX (eds) Oncology. People Health, Beijing, pp 183–185 Matter WF, Estridge T, Zhang C, Belagaje R, Stancato L, Dixon J, Johnson B, Bloem L, Pickard T, Donaghue M, Acton S, Jeyaseelan R, Kadambi V, Vlahos CJ (2001) Role of PRL-3, a human muscle-specific tyrosine phosphatase, in angiotensin-II signaling. Biochem Biophys Res Commun 283:1061–1068

Neel BG, Tonks NK (1997) Protein tyrosine phosphatase in signal transduction. Curr Opin Cell Biol 9:193–204 Ponniah S, Wang DZ, Lim KL, Pallen CJ (1999) Targeted disruption of the tyrosine phosphatase PTPalpha leads to constitutive downregulation of the kinases Src and Fyn. Curr Biol 9:535–538 Saha S, Bardelli A, Buckhaults P, Velculescu VE, Rago C, St Croix B, Romans KE, Choti MA, Lengauer C, Kinzler KW, Vogelstein B (2001) A phosphatase associated with metastasis of colorectal cancer. Science 294:1343–1346 Sobin LH, Witterind C (1997) International Union Against Cancer. TNM classification of maglignant tumors (5th edn). Wiley, New York, p 53 Su J, Muranjian M, Sap J (1999) Receptor protein tyrosine phosphatase alpha activiates Src-family kinases and controls integrin-mediated responses in fibroblasts. Curr Biol 9:505–511 Susan A, Schumacher BU (2001) Metastasis research protocols, vol. 1. Humana, New Jersey, p 10 Tan YH (1993) Yin and yang of phosphorylation in cytokine signaling. Science 262:376–377 Zeng Q, Hong W, Tan YH (1998) Mouse PRL-2 and PRL-3, two potentially prenylated protein tyrosine phosphatases homologous to PRL-1. Biochem Biophys Res Commun 244:421–427 Zeng Q, Dong JM, Guo K, Li J, Tan HX, Koh V, Pallen CJ, Manser E, Hong W (2003) PRL-3 and PRL-1 Promote cell migration, invasion, and metastasis. Cancer Res 63:2716–2722 Zheng XM, Wang Y, Pallen CJ (1992) Cell transformation and activation of pp60-src by overexpression of a protein tyrosine phosphatase. Nature 359:336–339