The Prostate 28:372-378 (I996) There Are Multiple Forms of Glyceraldehyde-3-Phosphate Dehydrogenase in Prostate Cancer Cells and Normal Prostate Tissue Daniel E. Epner and Donald S. Coffey Departments of Oncology ( D E E , D.S.C.), Urology (D.S.C.), and Biochemistry, Cellular, and Molecular Biology (D.S.C.), The johns Hopkins University School of Medicine, Baltimore, Maryland ABSTRACT: We analyzed glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression in normal and malignant human prostate tissues, normal rat prostate, and Dunning R-3327rat prostate cancer cell lines. We detected multiple forms of GAPDH in Dunning cell lines by two-dimensional protein electrophoresis and Western analysis. Five forms of GAPDH that differed by isoelectric point were detected for each of the two metastatic Dunning cell lines, while four or fewer forms were detected for Dunning cell lines with low metastatic ability. We also detected multiple forms of GAPDH in normal and malignant human prostate specimens by two dimensional protein electrophoresis and immunohistochemical analysis. GAPDH was undetectable in normal human prostate secretory epithelium by immunohistochemistry, but was abundant in nuclei of normal basal cells and stromal cells. In human prostate cancer specimens, there was a rough correlation between cytoplasmic staining for GAPDH and tumor grade, but GAPDH staining was extremely heterogeneous. GAPDH was abundant in nuclei of some high-grade human prostate tumors. Both of the human prostate cancer bone metastases analyzed with immunohistochemistry had markedly elevated cytoplasmic GAPDH expression. We conclude that multiple forms of GAPDH may play diverse roles in normal prostate tissue and in prostate cancer. 0 1996 Wiley-Liss, Inc. KEY WORDS: prostatic neoplasms, glyceraldehydephosphatedehydrogenase, immunohistochemistry, two-dimensional gel electrophoresis INTRODUCTION GAPDH was initially identified several decades ago as a key regulatory glycolytic enzyme [1,2]. In recent years, there have been several studies that show that GAPDH expression is increased in a variety of human tumors and cancer cell lines [3-81. We previously showed that GAPDH RNA levels are increased in Dunning rat prostate tumors . Since increased glycolysis is one of the hallmarks of cancer [l], most authors have suggested that the role of GAPDH in cancer cells is exclusively that of a glycolytic enzyme [3-81. However, GAPDH is now known to have an astounding array of functions in both transformed and nontransformed cells that are seem0 1996 Wiley-Liss, Inc. ingly independent of its role in glycolysis. GAPDH is a DNA repair enzyme [10,11], microtubule associated protein [12-191, actin binding protein [20-231, protein kinase , and substrate for epidermal growth factor receptor kinase . GAPDH is also thought to play a role in detoxification of as-platinum and doxorubicin in cancer cells . We previously found a close cor- Received for publication November 28, 1994; accepted June 16, 1995. Address reprint requests to Dr. Daniel Epner, Baylor College of Medicine, VA Medical Center, Medical Service (111H), 2002 Holcombe Blvd., Houston, TX 77030. GAPDH and Prostate Cancer relation between GAPDH RNA levels and metastatic potential of Dunning rat prostate tumors, suggesting that GAPDH may play a role in metastasis . We undertook this study to determine whether there are multiple forms of GAPDH in normal and malignant prostate cells that may play diverse roles in normal prostate biology and prostate cancer progression. 373 Western Transfer MATERIALS AND METHODS Semidry transfer of proteins to Hybond'"-ECL nitrocellulose membranes (Amersham, Arlington Heights, IL) with the Milliblot"-SDE transfer system (Millipore) was done at 2 mA/cm2 for 30 min with a discontinuous buffer system (anode buffer 1:300 mM Tris, 20% methanol, pH 10.4; anode buffer 2:25 mM Tris, 20% methanol, pH 10.4; cathode buffer: 25 mM Tris, 40 mM glycine, 20% methanol, pH 9.4). Cell Culture lmmunoblotting Dunning R-3327 rat prostate cancer cell lines, which have been characterized previously [27-291, were maintained in WMI, 10% fetal bovine serum, dexamethasone 250 nM. Immunoblotting was performed at room temperature. Nitrocellulose membranes were blocked for 1hr in 10%nonfat dried milk/TBS-T (20 mM Tris, 137 mM sodium chloride, 0.1% Tween-20, pH 7.6), washed briefly with TBS-T alone, incubated for 1 hr with monoclonal antibody to GAPDH (Biogenesis, Franklin, MA) diluted 1:1,000 in 5% nonfat dried milk/ TBS-T, washed several times with TBS-T, incubated for 1 hr with horseradish peroxidase-labeled (Amersham) secondary antibody to mouse IgG, and washed again several times with TBS-T. Filters were then bathed in luminol and enhancer reagents (Amersham) according to the ECL'" Western blotting protocol for 1min and exposed to Kodak XAR-5 film for 10 sec-30 min. Sample Preparation for Two-Dimensional Protein Electrophoresis (2D PAGE) and Western Analysis Adherent cells were washed twice with ice cold phosphate buffered saline (PBS), incubated at 4°C for 45 min on a rocking platform in 1 ml of protein solubilization buffer (10 mM PIPES, 100 mM potassium chloride, 1 mM EGTA, 2 mM magnesium chloride, 300 mM sucrose, 0.5% Triton X-100, 1 mM phenylmethylsulfonylfluoride, 100 p,M leupeptin), scraped from the dish, and placed in 1.5 ml conical polypropylene tubes. Normal rat ventral prostate was rinsed in ice-cold PBS, snap frozen in liquid nitrogen, ground in liquid nitrogen with a mortar and pestle, and incubated for 45 min at 4°C in solubilization buffer. The fresh frozen human tissue sample was transferred directly to protein solubilization buffer and incubated at 4°C for 45 min. All samples were then centrifuged at 4"C, 10,000g for 10 min. The resultant supernatants, which contained >99% of total cellular GAPDH protein (data not shown), were analyzed as described below. 2D PAGE High-resolution two-dimensional gel electrophoresis was performed with the Investigator 2-D gel system (MilligenBiosearch, Bedford, MA) as described previously . One hundred and twenty micrograms of Carbamylated CPK isoelectric point markers (BDH Limited, Poole, UK) were included with each sample. Isoelectric focusing was performed for 18,000 V-h using 1-mm x 16-cm tube gels. For the second dimension, tube gels were then placed on 1-mm precast 10% Tris-acetate-SDS Duracryl polyacrylamide slab gels (Millipore Co., Bedford, MA) and eledrophoresis was camed out at 12"C, 16,000 mW until the bromophenol blue dye front was 17 cm from the top of the gel (approximately 5 hr). Patient Tissue Samples A total of 13 paraffin-embedded patient samples were analyzed for GAPDH expression by immunohistochemistry. Ten of the 13 were radical prostatectomy specimens removed from patients with clinical stage B adenocarcinoma. Post-operative Gleason scores for these 10 prostatectomy specimens ranged from 5-9. One sample was obtained from a 31-yearold man who underwent cystoprostatectomy for bladder disease and who was incidentally noted to have a small focus of adenocarcinoma of the prostate with Gleason score 1 2 = 3. The remaining two specimens were bone metastases resected from patients with stage D adenocarcinoma of the prostate. The fresh frozen radical prostatectomy specimen tested by 2D PAGE was from a patient with stage B adenocarcinoma, Gleason grade 7. This specimen was cut into serial 6 Fm sections and examined by hematoxylin and eosin staining at approximately 100 p,m intervals to identdy areas of nearly pure cancer or normal tissue. Cut sections were stored at -80°C until the time of analysis, when they were solubilized as described above. + lmmunohistochemistry Immunohistochemistry was performed with the automated Bio Tek Techmate 1,000 system (Santa Bar- 374 Epner and Coffey bara, CA). Six micron sections of paraffin-embedded specimens were mounted on Chem Mate Capillary Gap Plus microscope slides and microwave treated. Sections were deparaffinized by sequential treatment with xylene, absolute ethanol, 95% ethanol, and 80% ethanol; incubated with hydrogen peroxide to eliminate endogenous peroxidases and decrease background staining; incubated sequentially with monoclonal antibody to GAPDH diluted 1:2,000 (Biogenesis), followed by biotinylated secondary antibody, followed by avidin and peroxidase-complexed tertiary antibody; and exposed to diaminobenzene, a chromagen which yields a brown color. RESULTS There Are Multiple Forms of GAPDH Detectable in Dunning Prostate Cancer Cell Lines by Two-Dimensional Protein Electrophoresis (2D PAGE) and Western Analysis Results of 2D PAGE and Western analysis are shown in Figure 1. There were five forms of GAPDH detected in the two highly metastatic cell lines tested (MatLyLu and MatLu), while no more than four forms of GAPDH were detected in cell lines with low metastatic ability. Only one major form of GAPDH was detected in the G cell line, which is the most indolent and only androgen-sensitive Dunning cell line. GAPDH protein, which is ubiquitous, could not be detected in normal rat ventral prostate tissue under the conditions of the experiment. The absence of detectable GAPDH protein in samples from normal rat prostate illustrates the degree to which GAPDH expression is increased in cancer cells. Isoelectric standards were included in each sample to allow accurate assignment of PI values. The results shown in Figure 1are not quantitative, but instead show the relative abundance of various forms of GAPDH in the different cell lines. Twice as much total protein was loaded for G cells and normal prostate as for the other cell lines. Results of quantitative, one-dimensional Western analysis of four Dunning cell lines are shown in Figure 2. It is apparent from Figure 2 that GAPDH is much less abundant in nonmetastatic G cells than in other Dunning cell lines, and that total GAPDH protein levels do not correlate with metastatic potential. Human Prostate Cancer Tiswe I s Distinguishable From Normal Prostate Tissue by 2D PAGE and Western Analysis for GAPDH As shown in Figure 3, there were at least two distinct forms of GAPDH with PI of approximately 7.05 and 7.1 detected in normal human prostate tissue. In lsoelectric focusing pl6.8 I 37 kD - 37 kD - pl7.0 pl7.2 I I High Metastatic Potential 37 kD 37 kD - 37 kD - 37 kD - Fig. I. 2D PAGE and Western analysis of Dunning cell lines for GAPDH protein. The patterns were aligned based on the location of pl markers which were included in each sample. HIGH MET LOW MET 'G AT1 AT2' 'MLL1 37 kD - Optical Density 0.29 2.14 1.05 1.30 Fig. 2. I D PAGE and Western analysis of GAPDH levels in Dunning cell lines. Forty micrograms of total protein were loaded per lane. Corresponding optical density measurements are indicated directly beneath each band. Cell lines with low metastatic potential: G, ATI, and ATZ; cell line with high metastatic potential: Mat-LyLu (MLL). addition, there was a broad band of signal in the PI range of 6.75-6.9 which may either represent multiple forms of GAPDH of low abundance or one form which does not focus sharply. In contrast, the cancer specimen from the same patient did not contain the 375 GAPDH and Prostate Cancer lsoelectric focusing pl6.8 I PI 7.0 TABLE 1. Summary of ImmunohistochemicalAnalysis of GAPDH Expression in Human Prostate Tissue* pl7.2 I Normal Normal tissue Basal cells Secretory cells Stromal cells Adenocarcinoma Low grade High grade Metastasis (bone) 37 kD - 37 kD - Nuclear staining Cytoplasmic staining Heterogeneous pattern +++ - +++ - - - - + ++ +++ + +++ - ++ - - - ~ *-, none; Fig. 3. 2D PAGE and Western analysis of human prostate specimens for GAPDH protein. PI 7.1 form, but instead had a prominent signal at PI 7-7.05 that appeared to represent two forms of GAPDH with slightly different isoeledric points. The broad band in the PI range of 6.75-6.9 was also present in the cancer specimen. lmmunohistochemical Analysis of Human Prostate Tissue Reveals Multiple Forms of GAPDH With Nuclear and Cytoplasmic Localization Results of immunohistochemical analysis of GAPDH expression in human prostate tissue are summarized in Table I. GAPDH was abundant in all basal cell nuclei and in many stromal cell nuclei of all normal tissues studied (Fig. 4a). In contrast, GAPDH was not detectable in normal secretory epithelium (Fig. 4a). While GAPDH staining of normal prostate epithelium was uniform for all patients, GAPDH staining of prostate cancer specimens was extremely heterogeneous. Most low-grade cancers stained faintly for GAPDH (Fig. 4b), but some stained with moderate intensity. All high-grade tumors had moderate to intense nuclear or cytoplasmic staining for GAPDH (Figs. 4c and d, respectively), and some tumors had both nuclear and cytoplasmic staining. While there was a rough correlation between staining intensity and histologic grade, there was no consistent relationship between GAPDH staining and long-term clinical outcome (data not shown). Both metastatic lesions had very intense cytoplasmic GAPDH staining (Fig. 4e). DISCUSSION We found that five forms of GAPDH with different isoelectric points can be detected in metastatic Dun- + , little; + +, moderate; + + +, much. ning rat prostate cancer cell lines, while one to four forms can be detected in nonmetastaticcell lines (Fig. 1).We also detected at least three forms of GAPDH by 2D PAGE of human prostate tissues (Fig. 3). All of the forms of GAPDH detected in these studies represent 37 kD monomers. Since GAPDH is a tetramer in vivo, there may be many different GAPDH isoenzymes present in prostate cancer cells [31,32], each of which could conceivably have a unique function. There may also be GAPDH isoenzymes that are specific for prostate cancer cells. Further experiments will be required to determine whether the various forms of GAPDH in prostate cancer cells result from transcriptional activation of genes that are normally silent, posttranslational modification of a single gene product, or both. The different forms of GAPDH may represent differently phosphorylated proteins with identical amino acid sequence, since GAPDH is known to undergo autophosphorylation  and to be phosphorylated by epidermal-growth-factor-receptor kinase  and Ca’+/calmodulin-dependent protein kinase I1 . Phosphorylation of GAPDH may lead to nuclear localization, as is the case for protein kinase C and other proteins [34,35]. The fact that GAPDH is abundant in nuclei of normal prostate basal cells suggests that it may play a role in DNA repair in prostate epithelium, as it does in other tissues. High levels of cytoplasmic GAPDH in prostate cancer cells are probably at least partially a reflection of elevated glycolysis, one of the hallmarks of cancer. However, cytoplasmic GAPDH may have nonglycolytic roles in prostate cancer cells, since it is known to have an astounding number of diverse functions in other tissues. Future work will aim to clarify the roles that GAPDH has in prostate tissue and to determine whether there are cancer-specific GAPDH isoenzymes. 376 Epner and Coffer Fig. 4. lmmunohistochemical analysis of GAPDH expression in human prostate specimens. a: Normal gland, x 100; b low-grade adenucarcinoma. x 64 (bar = 40 km); c: nuclear staining in high-grade adenocarcinoma. x 64; d: cytoplasmic staining in high-grade adenocarcinoma X 64; e: bone metastasis, x 100. CONCLUSIONS 1. There are multiple forms of GAPDH in normal and malignant human prostate tissues and in rat prostate cancer cell lines. 2. Since GAF'DH is known to have many diverse functions, it may have multiple roles in prostate can- cer that are independent of its established role in glycolysis. ACKNOWLEDGMENTS Dr.Epner is the recipient of an American Cancer Society Physician's Research Training Award (PRTA- GAPDH and Prostate Cancer 14). This work was also supported by National Cancer Institute Grant CA15416; National Institute of Arthritis, Metabolism, and Digestive and Kidney Diseases Grants DK07552 and DK19300 (Department of Health and Human Services); and a SPORE grant for prostate cancer. The authors thank Dr.Jonathan Epstein for assistance in interpretation of immunohistochemical studies and Dr. Steven Bova for microdissecting the fresh frozen patient specimen analyzed by 2D PAGE and Western analysis. REFERENCES 1. Warburg OH: “The Metabolism of Tumors.” [Translated from the German edition by F. Dickens.] London: Constable, 1930. 2. Stryer, L: Biochemistry. New York: W.H. Freeman and Company, 1995. 3. Tokunaga K, Nakamura Y, Sakata K, Fujimori K, Ohkubo M, Sawada K, Sakiyama S: Enhanced expression of a glyceraldehyde-3-phosphate dehydrogenase gene in human lung cancers. Cancer Res 475616-1519, 1987. 4. Schek N, Hall BL, Finn OJ: Increased glyceraldehyde3-phosphate dehydrogenase gene expression in human pancreatic adenocarcinoma. Cancer Res 48:6354-6359, 1988. 5. Persons DA, Schek N, Hall BL, Finn OJ: Increased expression of glycolysis-associated genes in oncogenetransformed and growth-accelerated states. Mol Carcinog 288-94, 1989. 6. Perfetti V, Manenti G, Dragani TA: Expression of housekeeping genes in Hodgkin’s disease lymph nodes. Leukemia 5:lllO-1112, 1991. 7. Desprez PY, Poujol D, Saez S: Glyceraldehyde-3-phosphate dehydrogenase (GAPDH, E.C. 1.2.1. 12.) gene expression in two malignant human mammary epithelial cell lines: BT-20 and MCF-7. Regulation of gene expression by 1,25-dihydroxyvitaminD, (1,25-(OH),D,). Cancer Lett 64:219-224, 1992. 8. Ohkubo M, Nakamura Y, Tokunaga K, Sakiyam S: Similarity between glyceraldehyde-$phosphate dehydrogenase and a 37,000-dalton protein which is abundantly expressed in human lung cancers. Jpn J Cancer Res 77:554-559, 1986. 9. Epner DE, Partin AW, Schalken JA, Isaacs JT, Coffey DS: Association of glyceraldehyde-3-phosphate dehydrogenase expression with cell motility and metastatic potential of rat prostatic adenocarcinoma. Cancer Res 53:1995-1997, 1993. 10. Meyer-Siegler K, Mauro DJ, Seal G, Wurzer J, DeRiel JK, Sirover MA: A human nuclear uracil DNA glycosylase is the 37-kDa subunit of glyceraldehyde-3-phosphate dehydrogenase. Proc Natl Acad Sci USA 88: 8460-8464,1991. 11. Ronai Z: Glycolytic enzymes as DNA binding proteins. Int J Biochem 25:1073-1076, 1993. 12. Muronetz VI, Wang Z-X,Keith TJ, h u l l HR, Srivastava DK: Binding constants and stoichiometries of glyceraldehyde-3-phosphate dehyrogenase-tubulin complexes. Arch Biochem Biophys 313:253-260, 1994. 13. Somers M, EngelborghsY, Baert J: Analysis of the binding of glyceraldehyde-3-phosphate dehydrogenase to 377 microtubules, the mechanism of bundle formation and the linkage effect. Eur J Biochem 193:437-444, 1990. 14. Aithal NH, Walsh-Reitz MM, Kartha S, Janulis MP, Martin TE, Toback FG: Glyceraldehyde-3-phosphate dehydrogenase modifier protein is associated with microtubules in kidney epithelial cells. Am J Physiol266: F612-619, 1994. 15. Dumeu C, Bernier-Valentin F, Rousset B: Microtubules bind glyceraldehyde 3-phosphate dehydrogenase and modulate its enzyme activity and quaternary structure. Arch Biochem Biophys 252:32-40, 1987. 16. Balaban N, Goldman R The association of glycosomal enzymes and microtubules: a physiological phenomenon or an experimental artifact? Exp Cell Res 191:219226, 1990. 17. Walsh JL, Keith TJ, Knull HR: Glycolytic enzyme interactions with tubulin and microtubules. Biochim Biophys Acta 99954-70, 1989. 18. Launay JF, Jellali A, Vanier MT: Glyceraldehyde-3phosphate dehydrogenase is a microtubule binding protein in a human colon tumor cell line. Biochim Biophys Acta 996:103-109, 1989. 19. Huitorel P, Pantaloni D: Bundling of microtubules by glyceraldehyde-3-phosphate dehydrogenase and its modulation by ATP. Eur J Biochem 150:265-269, 1985. 20. Masters CJ, Reid S, Don M: Glycolysis-new concepts in an old pathway. Mol Cell Biochem 76:3-14, 1987. 21. Beitner R Regulation of carbohydrate metabolism. Boca Raton, FL: CRC Press, Inc., 1985. 22. K n d HR, Walsh JL: Association of glycolytic enzymes with the cytoskeleton. Curr Top Cell Regul 33:15-30, 1992. 23. Masters C Interactions between glycolytic enzymes and components of the cytomatrix. J Cell Biol99:222s 225.5, 1984. 24. Kawamoto RM, Caswell AH: Autophosphorylation of glyceraldehydephosphate dehydrogenase and phosphorylation of protein from skeletal muscle microsomes. Biochemistry 25656-661, 1986. 25. Reiss N, Kanety H, Schlessinger J: Five enzymes of the glycolytic pathway serve as substrates for purified epidermal-growth-factor receptor kinase. Biochem J 239: 691-697, 1986. 26. Hao XY, Bergh J, Brodin 0, Hellman U, Mannervik B: Acquired resistance to cisplatin and doxorubicin in a small cell lung cancer cell line is correlated to elevated expression of glutathione-linked detoxification enzymes. Carcinogenesis 15:1167-1173, 1994. 27. Isaacs JT, Isaacs WB, Feitz WFJ, Scheres J: Establishment and characterization of seven Dunning rat prostatic cancer cell lines and their use in developing methods for predicting metastatic abilities of prostatic cancer. Prostate 9:261-281, 1986. 28. Isaacs JT, Heston WDW, Weissman RM, Coffey DS: Animal models of the hormone-sensitive and -insensitive prostatic adenocarcinomas, Dunning R-3327-H, R-3327-HI, and R-3327-AT. Cancer Res 384353-4359, 1978. 29. Isaacs JT, Wake N, Coffey DS, Sandberg AA: Genetic instability coupled to clonal selection as a mechanism for tumor progression in the Dunning R-3327 rat prostatic adenocarcinoma system. Cancer Res 42:23532361, 1982. 30. Patton WF, Pluskal MG, Skea WM, Buecker JL, Lopez MF, Zimmerman R, Belanger LM, Hatch PD: Develop- 378 Epner and Coffey ment of a dedicated two-dimensional gel electrophoresis system that provides optimal pattern reproducibility and polypeptide resolution. Bio Techniques 8518-527, 1990. 31. Edwards YH, Clark P, Hams H: Isozymes of glyceraldehyde-%phosphate dehydrogenase in man and other animals. Ann Hum Genet 40:67-77, 1976. 32. Weinhouse S, Shatton JB, Criss WE, Farina FA, Morris Hp: Isozymes in relation to differentiation in transplantable rat hepatomas. In Weinhouse S, Ono T (eds): ”Isozymes and Enzyme Regulation in Cancer.” Tokyo: University of Tokyo Press, 1972, pp 1-17. 33. Ashmarina LI, Louzenko SE, Severin SE Jr, Muronetz VI, Nagradova NK Phosphorylation of D-glyceraldehyde-3-phosphate dehydrogenase by Ca2+/calmodulin-dependent protein kinase II. FEBS Lett 231:413-416, 1988. 34. Dingwall C, Laskey R The nuclear membrane. Science 258:942-947, 1992. 35. Silver PA: How proteins enter the nucleus. Cell 64:489497, 1991.