Title: Mutation analysis of thePTEN gene in uveal melanoma cell lines
Abstract: Uveal melanoma is the most common primary intra-ocular malignancy in adults, with an annual incidence of 0.7 per 100,000 people. Half of the patients die within 15 years after diagnosis due to metastatic disease, predominantly located in the liver. Until now, no predisposing factors have been identified; however, some tumour parameters, e.g., epithelioid cell type, large tumour diameter, and anterior location, have been correlated with a high risk of metastases (Mooy and De Jong, 1996). Cytogenetic studies have revealed that non-random abnormalities do occur. Deletions of chromosome 1p, loss of chromosome 3, gain of chromosome 8q, or alterations of chromosome 6 are found in most tumours; but specific chromosomal regions or genes involved in uveal tumorigenesis have yet to be identified (Prescher et al., 1996; Sisley et al., 1997; White et al., 1998). In up to 40% of cutaneous melanoma cell lines, somatic mutations of the PTEN gene have been detected (Guldberg et al., 1997). This prompted us to investigate the role of the PTEN gene in uveal melanoma cell lines since both cutaneous and uveal melanomas arise from neural crest–derived melanocytes. PTEN is a known tumour-suppressor gene, and somatic mutations of the PTEN gene or deletions of the PTEN locus at chromosome 10q23 have also been found in primary cutaneous melanoma (Tsao et al., 1998), glioblastomas (Wang et al., 1997), prostate carcinomas (Cairns et al., 1997), and breast carcinomas (Li et al., 1997). Germline mutations of PTEN are responsible for Cowden disease and Bannayan-Zonana syndrome (Marsh et al., 1998). The PTEN gene encodes a dual specific phosphatase, and this protein plays a major role in the inhibition of cell migration and the formation of focal adhesions (Tamura et al., 1998). In the present study, we used cytogenetic analysis and FISH to examine 9 uveal melanoma cell lines for deletions or translocations affecting the PTEN locus at 10q23. Expression of the gene was investigated by RT-PCR, and SSCP analysis of all 9 coding regions was carried out to screen for intragenic mutations. Nine uveal melanoma cell lines, isolated from different primary (EOM3, EOM29, OCM1, 92.1, Mel202, and Mel270) or metastatic (OMM1, OMM2, and OMM3) uveal melanomas, were analysed. Cytogenetic data of cell lines EOM3, EOM29, OCM1, OMM1, OMM2, and OMM3 have been published previously (Luyten et al., 1996). The karyotype of 92.1 (De Waard-Siebinga et al., 1995) was re-evaluated and showed, besides the already described diploid clone, a tetraploid clone with no additional abnormalities of chromosome 10. Cytogenetic studies on Mel202 and Mel270 (Chen et al., 1997) were carried out in our laboratory using standard cytogenetic techniques (Mel202: 50-53,XX[3],–X[13], +dic(1;9)(p11;p11),+5[3],add(6)(q?15),+add(6)(q?15),+7,+8,+8,+8,del(9)(q2?1)[2], del(9)(q3?2)[2],i(9)(p10)[3],add(11)(q22),add(16)(q11),add(18)(q21),der(20)t(8;20) (q12;q13),add(22)(p11)[cp16]; Mel270: 43∼48,XY,add(2)(p2?4)[4],?add(3)(q2)[5],psu i dic(6)(q11),der(7)del(7)(q22q22)add(7)(q32)[3],+der(7)del(7)(q22q22)add(7)(q32)[3], ?der(8)(p)[3],–9[5],?add(9)(p)[4],add(9)(p)[2],+der(9)add(9)(p)del(9)(q)[2],–10[3], add(10)(q2?)[3],add(12)(p11)[2],–13[6],add(13)(p11),add(16)(q11),der(6;17)t(6;17) (p2?5;q25)del(6)(q?),–19,add(21)(p1)[2],+1∼3mar[cp7]/81∼87,psu i dic(6)(q11)x2, der(7)del(7)(q22q22)add(7)(q32)x2,add(10)(q2?),der(6;17)t(6;17)(p2?5;q25)del(6)(q?), +marx2[cp2]). To search for smaller deletions, FISH analysis was performed using a PAC probe, 190D6, containing the entire coding sequence of the PTEN locus. This PAC probe was isolated by screening a genomic human PAC library on gridded filters (Genome System, St. Louis, MO) with an RT-PCR probe of PTEN. Single-colour FISH using 75 ng of biotin-labeled PAC190D6 was performed on chromosome preparations as described (Hagemeijer et al., 1998). For each cell line, 300 interphase nuclei were scored and the cut-off value for deletions (10%, mean of aberrant signals + 3 standard deviation) was calculated from hybridisation on normal lymphocytes. The cytogenetic and FISH results are summarised in Table I. In 6 cell lines, there was agreement between the results from the cytogenetic studies and FISH analysis. Two cell lines (Mel270 and OMM1) had structural abnormalities [add(10)(q2?) and t(2;10)(q32;q25), respectively] involving the long arm of chromosome 10, though no disruption or loss of the PTEN gene due to these translocations was found with FISH analysis. Cytogenetic analysis showed under-representation of chromosome 10 in 3 of 7 metaphases of Mel270. Interphase FISH studies on Mel270 revealed one signal for 190D6 in only 12% of the nuclei. The discrepancy between the FISH findings of 12% monosomy for this chromosome and the cytogenetic results (43% monosomy) may be due to the fact that this particular subclone actively proliferates. Lines OMM1 and OMM2 had small populations of interphase nuclei with 4 FISH signals for the PTEN gene. In both instances, these subclones were not detected by chromosome analysis or by FISH analysis on metaphases. However, they could have been in vitro culture artefacts since the FISH studies were carried out at a later passage than the original cytogenetic analyses. Mutation analysis on genomic DNA was performed using exon-specific SSCP. All 9 exons and flanking sequences were amplified using 11 primer pairs, and PCR products were analysed as described by Vlietstra et al. (1998). We separated 2 single-stranded bands for all primer pairs in all cell lines. No abnormal banding patterns, indicative of PTEN mutations, were observed. Examples of amplification of exons 3 and 5 are shown in Figure 1a. Expression of the PTEN gene was investigated using RT-PCR. cDNA, synthesised from 1 to 3 μg of total RNA (RNeasy kit; Qiagen, Santa Clarita, CA), and 2 primer pairs (977F, CCACCAGCAGCTTCT GCCATCTCT, and 1736R, CCAATTCAGGACCCACACGACGG; 1649F, GTTCAGTGG CGGAACTTGCAATCCTCA, and 2423R, CCCTATACATCCACAGGGTTTTGACACTT G) were used to amplify the partially overlapping PTEN cDNAs. In all cell lines, the 2 PCR products of 759 and 774 bp amplified normally (Fig. 1b) and no abnormally sized bands were observed. PTEN expression levels corresponded with expression of a housekeeping gene (data not shown) and do not reflect differences between cell lines. PTEN mRNA expression and SSCP analysis of uveal melanoma cell lines. (a) Exon-specific SSCP for exons 3 and 5 of the PTEN gene in uveal melanoma cell lines. The 2 single-stranded bands are clearly visible (arrowheads). (b) Agarose gel analysis of RT-PCR–amplified fragments. Two primer pairs were used to amplify partially overlapping products of 759 (977F/1736R) and 774 (1649F/2423R) bp. The present study demonstrates that loss, structural alterations, or somatic mutations of the PTEN gene are absent in uveal melanoma cell lines. These findings are in contrast with data obtained in cutaneous malignant melanoma cell lines. Other genetic differences between cutaneous and uveal melanomas have been reported. Whereas in up to 100% of cutaneous melanomas the involvement of the CDKN2 gene or its downstream target genes has been postulated (Walker et al., 1998), CDKN2 mutations or losses are rarely observed in uveal melanoma cell lines and primary tumours (Singh et al., 1996; Naus et al. unpublished results). Also, cytogenetically, there is a clear difference between uveal and cutaneous melanomas. The typical abnormalities found in uveal melanoma, such as loss of chromosome 3 or gain of 8q, are rarely observed in cutaneous melanoma. However, abnormalities of chromosomes 1 and 6 are common to both melanoma types. Biological differences also exist between these 2 tumours. Integrin expression, which is important for the growth and metastatic capacity of melanoma cells, as well as the expression of melanoma-associated antigens, differ markedly between uveal and cutaneous melanomas (Marshall et al., 1998; Mooy and De Jong, 1996). These data suggest that, despite the common embryonic origin, these tumours follow different paths towards tumorigenesis and have different biological behaviours. The question remains, however, whether we can extrapolate our findings on uveal melanoma cell lines to fresh tumour material. The cytogenetic findings in primary uveal melanoma show, in general, a simple karyotype with only a few recurrent abnormalities, such as loss of chromosome 1p and chromosome 3, gain of chromosome 8q, and structural abnormalities of chromosome 6 (Prescher et al., 1996; Sisley et al., 1997; White et al., 1998). Some of the uveal melanoma cell lines in our study harboured these genetic changes, but these were accompanied by additional chromosomal aberrations. Whether these complex abnormalities are the result of prolonged cell culture or the fact that tumours with a more complex karyotype are more likely to grow in vitro is unclear. However, in our study, no PTEN mutations were detected in uveal melanoma cell lines and cytogenetic abnormalities involving chromosome 10q23 were not observed. Furthermore, in contrast to primary cutaneous melanoma, chromosome 10q23 abnormalities are rarely found in primary uveal melanoma. This suggests that PTEN mutations or deletions probably do not play a role in the pathogenesis of uveal melanoma and that the tumorigenesis of uveal melanoma involves a different set of genes from those involved in cutaneous melanoma. Yours sincerely, Nicole C. Naus, Wieke Zuidervaart, Nazik Rayman, Rosalyn Slater, Ellen van Drunen, Bruce Ksander, Gregorius P.M. Luyten, Annelies Klein