Histochem Cell Biol (1999) 111:399–403

© Springer-Verlag 1999

O R I G I N A L PA P E R

Alexander Berndt · Xinmei Luo · Frank-D. Böhmer Hartwig Kosmehl

α Expression of the transmembrane protein tyrosine phosphatase RPTPα in human oral squamous cell carcinoma

Accepted: 2 February 1999

Abstract Little is known about the role of protein-tyrosine phosphatases (PTPs), the cellular counterparts of protein-tyrosine kinases, both for normal growth regulation and for its dysregulation in cancer. The receptor-like PTPα (RPTPα) may play a positive role in growth regulation and has been shown to be overexpressed in colon carcinoma. An RNA/RNA in situ hybridisation protocol for RPTPα as well as RPTPα immunohistochemistry was developed to evaluate RPTPα expression in oral squamous cell carcinomas (OSCCs) of different histological grade and to reveal the synthetically active cells and their tissue distribution. In well-differentiated OSCC (G1), RPTPα mRNA could be detected by in situ hybridisation exclusively in stroma cells (fibro/myofibroblasts and inflammatory cells). A higher histological grade (G2/G3) was associated with an increased number of RPTPα-synthesising carcinoma cells haphazardly distributed within invading tumour areas. Consistent results were obtained by immunocytochemistry. Thus, both carcinoma dedifferentiation and stroma recruitment and activation seem to be associated with an upregulation of RPTPα expression in OSCC. The results speak in favour of the important role of activation of stroma fibro/myofibroblasts influencing the biological behaviour of epithelial tumours and also suggest that elevated RPTPα expression may be a more general marker for proliferating or dedifferentiated cells.

A. Berndt and X. Luo contributed equally to this study A. Berndt (✉) · H. Kosmehl Institute of Pathology, Friedrich Schiller University, D-07740 Jena, Germany e-mail: [email protected] Tel.: +49-3641-633-624 Fax: +49-3641-633-111 X. Luo · F.-D. Böhmer Research Unit “Molecular Cell Biology”, Friedrich Schiller University, D-07740 Jena, Germany

Introduction Protein-tyrosine phosphorylation is an important cellular mechanism for the regulation of growth and differentiation, and aberrant tyrosine phosphorylation has been linked to various diseases including human malignancies. Whereas the significance of different protein-tyrosine kinases (PTKs) has been studied in great detail, much less is known about the role of their counterparts, protein-tyrosine phosphatases (PTPs), both for normal growth regulation and for dysregulation in cancer (Hunter 1995, 1998). Numerous PTPs have been identified over the last 10 years falling into two main classes, the cytosolic and the transmembrane, receptor-like PTPs (Neel and Tonks 1997; Schaapveld et al. 1997). The receptor-like PTPα (RPTPα) is a widely expressed enzyme with particularly high levels in brain and kidney (Matthews et al. 1990; Sap et al. 1990). It consists of a short, heavily glycosylated extracellular domain, a single transmembrane domain and a cytosolic domain with two intracellular catalytic PTP domains, D1 and D2. The membrane proximal domain D1 seems to harbour most of the PTP activity, although the domain D2 has been demonstrated to possess activity against some substrates in vitro (Wang and Pallen 1991). RPTPα can be activated by Ser-specific phosphorylation via PKC (Denhertog et al. 1995), and dimerisation of two RPTPα molecules may present an inactivation mechanism (Bilwes et al. 1996). Src-family kinases including pp60 src and fyn apparently present substrates of RPTPα in vitro and in certain cell types (Zheng et al. 1992; Denhertog et al. 1993; Bhandari et al. 1998). RPTPα dephosphorylates inhibitory C-terminal phosphotyrosines in these PTKs leading to PTK activation. This results in initiation of differentiation in neuronal cells (Denhertog et al. 1993) and can result in a transformed phenotype as demonstrated for rat embryo fibroblasts (Zheng et al. 1992). On the other hand, RPTPα has been implicated in negative regulation of insulin receptor signalling by dephoshorylation of the insulin receptor itself and/or by antagonising downstream signalling steps (Moller et al.

400

1995; Lammers et al. 1997; Jacob et al. 1998). In transient overexpression systems, RPTPα also potently dephosphorylates the receptor for platelet-derived growth factor (PDGF; S. Groß, X. Luo and F. D. Böhmer, unpublished data). Thus, depending on the cellular context, RPTPα expression and activity level may have positive or negative consequences for cell growth and differentiation. Due to this implication of RPTPα in cell growth regulation and possibly transformation, its tissue distribution in carcinoma may contribute to the understanding of carcinoma invasion and progression. Therefore, an RNA/ RNA in situ hybridisation protocol for RPTPα has been developed to reveal the synthetically active cells and their tissue distribution in tumours. Analysis of oral squamous cell carcinoma (OSCC) with this technique revealed RPTPα expression and cellular distribution in relation to the histological grade.

Materials and methods Tissue material The RPTPα mRNA in situ hybridisation and the corresponding immunohistochemistry were performed on 12 OSCCs. Three carcinomas were well differentiated (G1), 6 were moderately differentiated (G2) and 3 were less differentiated (G3). Tissue specimens included areas of non-neoplastic, hyperplastic and dysplastic epithelia. Blocks of 4×4×4 mm of fresh, surgically obtained tissue were immediately shock frozen in isopropanol cooled by liquid nitrogen and stored at -75°C. The diagnosis was confirmed in the corresponding paraffin-embedded tissue according to the WHO classification criteria (Pindborg et al. 1997). Histological grading of malignancy was performed using the method reported by Bryne and coworkers (1992). In situ hybridisation A cDNA for murine RPTPαa (LRP) was kindly provided by Dr. M. Thomas (Washington). A 728-bp fragment (nucleotides 1–728 of LRP cDNA; Matthews et al. 1990) covering the coding sequence for the extracellular, transmembrane and juxtamembrane domain was obtained by EcoRI digestion and inserted into EcoRIdigested pBluescriptIIKS+ vector (Stratagene, La Jolla, USA). Sense or antisense RNA probes were generated by in vitro transcription of the HindIII or BamHI linearised plasmid using digoxigenin-labelled uridine triphosphate as substrate (DIG RNA labelling kit; Boehringer Mannheim, Mannheim, Germany) and T7 or T3 RNA polymerase. Unincorporated nucleotides were removed by ethanol precipitation. The precipitate was dissolved in 100 µl diethyl pyrocarbonate-treated and RNase inhibitor-containing water. The transcripts were analysed by agarose gel electrophoresis and dot blot assay. In situ hybridisation was carried as described previously (Berndt et al. 1998). Sections, 7–10 µm thick, of immediately snap-frozen tissue were heated for 2 min at 50°C to fix the RNA in the tissue. The sections were allowed to air dry for 30 min and were then fixed in 4% (w/v) paraformaldehyde in phosphate-buffered saline (PBS) for 20 min at 4°C. After washing in PBS (1× 5 min) and 2×SSC (2×5 min), the prehybridisation solution was applied on each slide for 60 min at 37°C. The prehybridisation solution contained 4×SSC, 10% dextran sulphate, 1×Denhardt’s solution, 2 mM EDTA, 50% deionised formamide, 100 µg/ml herring sperm DNA and 100 µg/ml tRNA. After removal of this solution, each section was covered with 100 µl prehybridisation solu-

Fig. 1 Specificity of anti-receptor-like protein-tyrosine phosphatase (anti-RPTPα) antibodies. Extracts of human U937 histiocytic lymphoma cells (lane 2), human embryonal kidney (HEK) 293 cells (lane 3; 50 mg protein per lane) or a lysate aliquot of E. coli expressing a fusion protein containing the first catalytic domain of RPTPα (lane 1) were subjected to gel electrophoresis and immunoblotting with anti-RPTPα antibody (clone 21). Marker protein positions are indicated tion containing 200–1000 ng/ml of DIG-labelled antisense cRNA probe and incubated at 37°C overnight. As hybridisation controls, the antisense cRNA probe was replaced by RPTPα sense cRNA or was completely omitted from the hybridisation to evaluate the quality of the colour-detection system. After hybridisation, unbound probe was washed from the sections as follows: 1×5 min with 2×SSC at 37°C, 3×5 min with 60% formamide in 0.2×SSC at 37°C and 2×5 min with 2×SSC at room temperature. Hybridised DIG-labelled cRNA probes were detected using the components of the DIG nucleic acid detection kit (Boehringer Mannheim) following the protocol recommended by the manufacturer. The colour reaction was carried out for up to 24 h at room temperature. Immunohistochemistry Cryostat sections of the respective frozen tissue samples were fixed in ice-cold acetone for 15 min and subjected to immunohistochemistry. As primary anti-RPTPα antibody, the clone 21 (Transduction Laboratories, Lexington, USA) was used. This antibody detects a single band of 140–150 kDa in immunoblotting with extracts of human U937 and HEK 293 cells (Fig. 1) or SKN human neuroblastoma cells (Transduction Laboratories, 1998 Catalogue). Immunohistochemical staining was performed using the alkaline phosphatase monoclonal anti-alkaline phosphatase (APAAP) method. The primary antibody was incubated with the tissue sections for 30 min at room temperature. After washing with TRIS buffer, sections were treated with rabbit anti-mouse immunoglobulin (diluted 1:70; Dako, Glostrup, Denmark), and then with the mouse APAAP complex (Dako). Both incubations were done for 30 min at room temperature. To increase the staining intensity, the incubation with the rabbit anti-mouse immunoglobulin and with the APAAP complex was repeated twice. Naphthol AS-BI phosphate (Sigma, St. Louis, USA.) and new fuchsin (Merck, Darmstadt, Germany) were used as substrate and developer, respectively. To inhibit endogenous tissue enzyme activity, the developing solution was supplemented with 0.25 mmol/l levamisole (Sigma). To evaluate the specificity of immunostaining, the primary antibody was replaced by non-immune serum as a negative control.

Results Analysis of synthesis and distribution of RPTPα: RPTPα immunohistochemistry In normal adult and hyperplastic squamous epithelium, RPTPα was not detectable using the antibody 21. In the G1 and G2 OSCCs, RPTPα distribution could be demonstrated by immunohistochemistry to be restricted to

401

Fig. 2A–F RPTPα protein and mRNA demonstration in oral squamous cell carcinoma (OSCC). In G1 OSCC, RPTPα distribution is demonstrated by immunohistochemistry to be restricted to the tumour stroma. A Carcinoma cells show only a slight intracellular immunostaining (monoclonal antibody 21, alkaline phosphatase anti-alkaline phosphatase (APAAP) technique red, haematoxylin counterstaining, ×150). B In G3 tumours, in addition to a similar stroma cell and vascular staining, tumour cells themselves reveal a clear RPTPα immunopositivity (arrowheads, monoclonal antibody 21, APAAP technique red, without counterstaining,

×150). C, D In comparison to the sense control (C), the whole tumour compartment showed a strong in situ hybridisation signal for RPTPα mRNA in less-differentiated carcinomas (G3) (D, brown in situ hybridisation signal, without counterstaining, ×150). E, F In well-differentiated carcinomas (G1/G2), RPTPα mRNA could be clearly detected preferentially in stromal cells (F, brown in situ hybridisation signal, without counterstaining, ×150) displaying the fibro/myofibroblastic or lymphocytic phenotype (E, haematoxylin/eosin staining, ×150). Scale bar 70 µm

402

the tumour stroma. The most pronounced staining was observed in stromal fibro/myofibroblasts and inflammatory cells immediately surrounding the tumour areas as well as in endothelial cells of stromal vessels. In these tumours, carcinoma cells showed a slight intracellular immunostaining (Fig. 2A). Conversely, in G3 tumours we observed a similar stroma cell and vascular staining but, in these less-differentiated tumours, the tumour cells themselves showed an intracellular and membranous RPTPα immunostaining (Fig. 2B). Analysis of synthesis and distribution of RPTPα: RNA/RNA in situ hybridisation for RPTPα The omission of the probe from the hybridisation solution showed no background staining in all cases. The use of sense cRNA as hybridisation probe led to a weak or moderate background staining affecting all cells (Fig. 2C). True positively stained cells could be identified in comparison to the negative controls by their strong cytoplasmic hybridisation signal using the RPTPα antisense probe. Hybridisation signals were interpreted as specific for RPTPα if they were absent in all negative controls. In normal and hyperplastic epithelia, RPTPα mRNA was below detection threshold. In G1 and G2 carcinomas, RPTPα mRNA could be clearly detected in tumour stroma fibroblasts and inflammatory cells (Fig. 2E,F). RPTPα mRNA positive stromal cells displaying the fibro/myofibroblast phenotype were preferentially visible in areas with desmoplastic stroma reaction. In moderately differentiated tumours (G2), only some single carcinoma cells displayed a clear hybridisation signal, whilst well-differentiated OSCC (G1) tumour cells were negative. In less-differentiated carcinomas (G3), the whole tumour cell compartment showed a strong in situ hybridisation signal (Fig. 2D), in addition to the signals detected in the stroma.

Discussion In this study, evidence was obtained for expression of RPTPα in OSCC cells which seems to correlate to tumour cell dedifferentiation. Consistent results were obtained both with immunocytochemistry and with in situ hybridisation, supporting the specificity of the observed signals and suggesting that regulation of RPTPα expression in these tissues is at the level of transcription. Overexpression of RPTPα has been observed before in late stage colon carcinomas (Tabiti et al. 1995) and may thus be a more general phenomenon in epithelial tumours. In fact, we also observed strong RPTPα positivity in a small number of inspected mammary carcinomas (own unpublished data). The observation that RPTPα expression is elevated in tumours is consistent with the earlier finding that overexpression of RPTPα can have transforming effects in rat embryo fibroblasts (Zheng et al. 1992) and that

RPTPα has the capacitiy to activate Src-family kinases by dephosphorylating the inhibitory C-terminal phosphotyrosine (Zheng et al. 1992; Denhertog et al. 1993). However, elevated RPTPα expression may be a cellular reaction to elevated tyrosine phosphorylation in transformed or stimulated cells and be considered as an “attempt” to counteract aberrantly activated signalling pathways. Indeed, upregulation of RPTPα mRNA in response to growth stimulation has been observed in fibroblasts stimulated for prolonged periods of time with PDGF and also in fibroblasts transformed with the sisoncogene (Celler et al. 1995). Further understanding of the cellular roles of RPTPα is necessary to interpret the observed elevated expression level in tumours. Given the fact that RPTPα is a membrane protein, it is striking that the immunostaining pattern of tumour cells is mainly intracellular. This may be due to an active synthesis and intracellular accumulation of the receptor protein. As shown by Daum et al. (1994), RPTPα is a heavily glycosylated molecule (130 kDa of which 45 kDa represents carbohydrate) which undergoes a lengthy posttranslational maturation. Upon transient expression in different cell types, a 100-kDa precursor is detectable in significant amounts which by virtue of its incomplete glycosylation is located in the ER/Golgi compartment. Incompletely glycosylated forms of RPTPαa are likely to correspond to the intracellularly detectable immunostaining. Interestingly, not only the tumour cells but also stromal fibro/myofibroblasts as well as inflammatory cells account for RPTPα expression in OSCC. In fact, expression in these cells is already detectable in well-differentiated tumours (G1) with largely negative tumour cells as well as in less-differentiated carcinomas. This is in line with the important role of activation of stromal fibro/myofibroblasts influencing the biological behaviour of epithelial tumours (Schürch et al. 1992; Gregoire and Lieubeau 1995; Iozzo 1995). Stromal myofibroblasts in tumours may deliver numerous growth factors and are involved in remodelling the extracellular matrix towards an embryonic microenvironment (Brouty-Boyé and Magnien 1994; Berndt et al. 1995, 1998; Kosmehl et al. 1996). Therefore, one could speculate that tumour-derived or autocrine growth factors, such as PDGF, mediate stroma activation leading to elevated RPTPα expression and to raised growth factor susceptibility. PDGF has been shown to elicit at least phenotypic conversion of resting fibroblasts into spindle-type myofibroblasts (Oh et al. 1998). Moreover, elevated RPTPα expression seems not to be a phenomenon specifically confined to malignantly transformed cells but may rather be a more general marker for actively proliferating or dedifferentiated cells which exist in the tumour tissue itself as well as in the stroma. Acknowledgements The authors are grateful to Prof. Dr. Dr. Peter Hyckel, Clinic of Maxillofacial Surgery of the Friedrich Schiller University, Jena, for providing native tumour tissue. The authors would like to thank Dr. Gareth M. Thomas for careful reading of the manuscript and to Mrs. Christiane Rudolph and Mrs. Carola König for skillful technical assistance. The investigation was supported by the grant ThMWK no. 973 117.

403

References Berndt A, Kosmehl H, Mandel U, Gabler U, Luo X, Celeda D, Zardi L, Katenkamp D (1995) TGFß and bFGF synthesis and localization in Dupuytren’s disease (nodular palmar fibromatosis) relative to cellular activity, myofibroblast phenotype and oncofetal variants of fibronectin. Histochem J 27:1014–1020 Berndt A, Kosmehl H, Borsi L, Luo XM, Zardi L, Katenkamp D (1998) Evidence of ED-B+ fibronectin synthesis in human tissues by non radioactive RNA in situ hybridization. Investigations on carcinoma (oral squamous cell and breast carcinoma), chronic inflammation (rheumatoid synovitis) and fibromatosis (Morbus Dupuytren). Histochem Cell Biol 109:249–255 Bhandari V, Lim KL, Pallen CJ (1998) Physical and functional interactions between receptor-like protein-tyrosine-phosphatasealpha and p59(fyn). J Biol Chem 273:8691–8698 Bilwes AM, Den HJ, Hunter T, Noel JP (1996) Structural basis for inhibition of receptor protein-tyrosine phosphatase-alpha by dimerization. Nature 382:555–559 Brouty-Boyé D, Magnien V (1994) Myofibroblast and concurrent ED-B fibronectin phenotype in human stromal cells cultured from non-malignant and malignant breast tissue. Eur J Cancer 30A:66–73 Bryne M, Koppang HS, Lilleng R, Kjaerheim A (1992) Malignancy grading of the deep invasive margins of oral squamous cell carcinoma has high prognostic value. J Pathol 166:375–381 Celler JW, Luo XM, Gonez LJ, Böhmer FD (1995) MessengerRNA expression of 2 transmembrane protein-tyrosine phosphatases is modulated by growth factors and growth arrest in 3T3 fibroblasts. Biochem Biophys Res Commun 209:614– 621 Daum G, Regenass S, Sap J, Schlessinger J, Fischer EH (1994) Multiple forms of the human tyrosine phosphatase RPTPα. Isozymes and differences in glycosylation. J Biol Chem 14:10524–10528 DenHertog J, Pals CEGM, Peppelenbosch MP, Tertoolen LGJ, Delaat SW, Kruijer W (1993) Receptor protein tyrosine phosphatase-alpha activates pp60(c-src) and is involved in neuronal differentiation. EMBO J 12:3789–3798 DenHertog J, Sap J, Pals CEGM, Schlessinger J, Kruijer W (1995) Stimulation of receptor protein tyrosine phosphatase alpha activity and phosphorylation by phorbol ester. Cell Growth Differ 6:303–307 Gregoire M, Lieubeau B (1995) The role of fibroblasts in tumour behaviour. Cancer Metastasis Rev 14:339–350 Hunter T (1995) Protein kinases and phosphatases: The yin and yang of protein phosphorylation and signaling. Cell 80:225– 236

Hunter T (1998) Anti-phosphatases take the stage. Nat Genet 18:303–305 Iozzo RV (1995) Tumor stroma as a regulator of neoplastic behavior. Lab Invest 73:157–160 Jacob KK, Sap J, Stanley FM (1998) Receptor like protein tyrosine phosphatase alpha specifically inhibits insulin increased prolactin gene expression. J Biol Chem 273:4800–4809 Kosmehl H, Berndt A, Katenkamp D (1996) Molecular variants of fibronectin and laminin: structure, physiological occurrence and pathohistological aspects. Virchows Arch 429:311–322 Lammers R, Møller NPH, Ullrich A (1997) The transmembrane protein tyrosine phosphatase alpha dephosphorylates the insulin receptor in intact cells. FEBS Lett 404(N1):37–40 Matthews RJ, Cahir ED, Thomas ML (1990) Identification of an additional member of the protein tyrosine phosphatase family: evidence for alternative splicing in the tyrosine phosphatase domain. Proc Natl Acad Sci USA 87:4444–4448 Møller NPH, Møller K, Lammers R, Kharitonenkov A, Hoppe E, Wiberg F, Sures I, Ullrich A (1995) Selective down-regulation of the insulin receptor signal by protein tyrosine phosphatases alpha and epsilon. J Biol Chem 270:23126–23131 Neel BG, Tonks NK (1997) Protein-tyrosine phosphatases in signal transduction. Curr Opin Cell Biol 9:193–204 Oh SJ, Kurz H, Christ B, Wilting J (1998) Platelet derived growth factor ß induces transformation of fibrocytes into spindle shaped myofibroblasts in vivo. Histochem Cell Biol 109:349–357 Pindborg JJ, Reichart PA, Smith CJ, Wal I van der (1997) Histological typing of cancer and precancer of the oral mucosa/World Health Organization, 2nd edn. Springer, Berlin Heidelberg New York Sap J, D’Eustachio P, Givol D, Schlessinger J (1990) Cloning and expression of a widely expressed receptor tyrosine phosphatase. Proc Natl Acad Sci USA 87:6112–6116 Schaapveld R, Wieringa B, Hendriks W (1997) Receptor like protein tyrosine phosphatases – alike and yet so different. Mol Biol Rep 24:247–262 Schürch W, Seemayer TA, Gabbiani G (1992) Myofibroblast. In: Sternberg SS (ed) Histology for pathologists. Raven Press, New York, pp | Tabiti K, Smith DR, Goh HS, Pallen CJ (1995) Increased mRNA expression of the receptor like protein tyrosine phosphatase alpha in late stage colon carcinomas. Cancer Lett 93:239–248 Wang Y, Pallen CJ (1991) The receptor like protein tyrosine phosphatase RPTP alpha has two active catalytic domains with distinct substrate specificities. EMBO J 10:3231–3237 Zheng XM, Wang Y, Pallen CJ (1992) Cell transformation and activation of pp60 (c-src) by overexpression of a protein tyrosine phosphatase. Nature 359:336–339