hH-Rev107, a class II tumor suppressor gene, is expressed by post-meiotic testicular germ cells and CIS cells but not by human testicular germ cell tumors.

Abstract

By systematic analysis of a human testis library, we have isolated the hH-Rev107-3 cDNA, identical to hH-Rev107-1 cDNA, which was previously described as a class II tumor suppressor gene (Husmann et al., 1998). In this study, two transcripts (1 and 0,8 kb) were detected by Northern blot in all human tissues, excepted in thymus. The strongest expression was found in testis, skeletal muscle and heart. These two mRNA are probably transcribed from only one gene that we mapped to the q12–q13 region of the chromosome 11. In human testis, hH-Rev107 gene expression was localized, by in situ hybridization, within the round spermatids. To investigate a possible role for hH-Rev107 protein in testicular malignant growth, we examined the expression of this gene in germ cell tumors. A strong hH-Rev107 gene expression was observed in normal testis as well as in samples with preinvasive carcinoma in situ but was completely absent in overt tumors, both seminomas and non-seminomas. By in situ hybridization, CIS was found hH-Rev107 positive and tumor negative. A semi-quantitative assessment of hH-Rev107 mRNA level in testicular germ cell tumors, by RT-PCR, exhibited a 9-fold decrease in the gene expression. No gross structural aberrations of hH-Rev107 gene were detected in these human primary tumors. The results suggest that down-regulation of hH-Rev107 may be associated with invasive progression of testicular germ cell tumors.

Introduction

Testicular tumors are rare, comprising approximately 2% of all cancers occurring in men. However, this neoplasia is the most common malignant disease in males aged 15–35 years (Richie, 1993) and represents the most frequent cause of death from solid tumors in this patient group. The incidence has steadily risen in the past 30–50 years, particularly in the western world (Peckham, 1988) (Schmidt et al., 1995).

The large majority of testicular tumors originate from germ cells, rarely from Sertoli or Leydig cells. Testicular germ cell tumors (TGCTs) are a heterogeneous group of different histological tumors which are divided into two major groups: seminomas, which proliferate along germ cell lineage, and non-seminomas, which display embryonal differentiation and exhibit embryonal (embryonal carcinoma), extra-embryonal (choriocarcinoma, yolk sac tumor), and somatic (teratoma) patterns of tissue differentiation. Most of TGCTs are derived from a common preinvasive precursor, carcinoma in situ (CIS) cells, which may represent gonocytes that have been delayed in their progression to spermatogonia (Skakkebaek, 1972) (Skakkebaek et al., 1987). Because the molecular pathogenesis of germ cell tumors is not known, identification of genes involved in testicular tumorigenesis will increase our understanding of development of testicular tumors and may provide the basis for new-targeted therapies.

By systematic sequence analysis of cDNAs expressed in human testis, we isolated new human transcripts (Pawlak et al., 1995), some of which correspond to genes that are potentially involved in cell division. Among them, one human testis cDNA sequence was found to be identical to the human H-Rev107-1 cDNA (Husmann et al., 1998), which was proposed to act as a class II tumor suppressor gene.

The rH-Rev107 cDNA was first isolated in rat cells from a phenotypic revertant (F9 cells) of H-Ras transformed rat fibroblasts (FE-8 cells). No rH-Rev107 transcript was detected in FE-8 cells. Apart from F9 cells, a strong expression of the rH-Rev107 gene was also found in a fibroblast line which is resistant to transformation by H-Ras, suggesting that the rH-Rev107 gene was directly involved in the resistance toward H-Ras mediated transformation (Hajnal et al., 1994). Further studies have shown that the rH-Rev107 gene was frequently down regulated in rat tumorigenic cell lines and in experimental tumors mediated by targeted expression of a H-Ras transgene. Additionally, forced expression of the rH-Rev107 cDNA, controlled by heterologous promoters in deficient rodent cell lines transformed by H-Ras, resulted in growth inhibition in vivo and in vitro (Sers et al., 1997). As repression of the H-Rev107 gene occurred within the tumor cell lines without any modification at the DNA level, the H-Rev107 gene was categorised as a class II tumor suppressor gene (Sers et al., 1997). This class of antioncogenes possesses a functionally intact gene, but its expression is repressed in tumors, and permits cell proliferation (Lee et al., 1991). The human orthologue of rH-Rev107 cDNA (hH-Rev107-1) was also shown to be down regulated in tumorigenic cell lines both at the level of transcription and translation (Husmann et al., 1998).

Given the evidence for the expression of hH-Rev107 gene in human testis and the antiproliferative function of the encoded protein, we wished to explore the potential role of H-Rev107 protein in human testicular germ cell pathogenesis. In this study, we analyzed the expression of the hH-Rev107 gene in human tissues; we localized the transcript in human testis within the germ cells, in CIS but not in TGCTs. We quantified gene expression by RT-PCR in a series of human TGCTs compared with normal testis and with the preinvasive CIS. Our data identify hH-Rev107 as candidate novel marker associated with progression of human TGCTs from preinvasive to invasive lesions.

Results

Cloning and sequencing of hH-Rev107-3 cDNA

We cloned a human testis cDNA (hH-Rev107-3) (Accession number: AF317086) which exhibited 99,4% identity with the hH-Rev107-1 cDNA (Husmann et al., 1998) with two insertions in the 5’ and 3’ non coding sequences and two mutations in the coding sequence leading to serine versus threonine. Sequences of 354 bp and 42 bp were missing in 5’ and 3’, respectively. The presence of a poly(A)+ tail at the 3’ end of hH-Rev107- 3 cDNA indicates that the transcript cloned from the testis has used an alternative polyadenylation site (Figure 1).

Figure 1. Sequence of the human hH-Rev107-3 cDNA.

Nucleotide sequence of the human hH-Rev107-3 cDNA (Accession number: AF317086) and the predicted amino acid sequence. As compared to hH-Rev107-1 sequence (Accession number: X92814), the two inserted nucleotides are represented in bold and the two mutated nucleotides are boxed.

Expression of hH-Rev107 gene in human tissues

The distribution of hH-Rev107 mRNA in human tissues revealed one or two transcripts of 1 and 0,8 kb from all tissues analyzed, apart from thymus where no signal was detected (Figure 2). The strongest expression was found in testis, skeletal muscle and heart. The same tissue distribution was reported by Husmann (Husmann et al., 1998) but the presence of two messengers were previously unnoticed, probably due to an overexposure of the blots. The relative distribution of these two mRNA was quantified from a weak exposure (data not shown) and was found tissue-specific: the 0,8 kb mRNA was only detected in leucocytes and spleen, and was preferentially expressed in pancreas, skeletal muscle and heart; the 1 kb mRNA was mainly expressed in testis, liver and brain. In mouse, only one messenger of 1 kb was detected by Northern blot of mouse tissues probed with hH-Rev107-3 cDNA, except in testis where two messengers were present (1 and 0,8 kb) (data not shown).

Figure 2. Expression of hH-Rev107 gene in human tissues.

Northern blot analysis was performed using human multiple-tissue blots containing 2 μg of poly(A)+ RNA per lane, and hybridized with ³²P labeled hH-Rev107-3 cDNA. Arrows on the right indicate the size of the mRNA.

Analysis of human genomic DNA

Human genomic DNA was digested using restriction enzymes cutting only once (Pst I) or none (BamH I, Kpn I, EcoR I and Hind III) in the hH-Rev107-3 cDNA sequence. Southern blot, hybridized with hH-Rev107-3 probe, revealed a simple pattern of 3 to 5 bands, leading to a total of 16 to 40 kb for the hH-Rev107 gene, that led us to conclusion that only one gene for hH-Rev107 exists in the human genome (Figure 3).

Figure 3. Analysis of human genomic DNA.

Southern blot was performed with human genomic DNA digested with the indicated restriction enzymes, and hybridized with ³²P labeled hH-Rev107-3 cDNA. The blot was exposed 7 nights with two screens. The size of the markers is indicated on the left.

hH-Rev 107 chromosomal localization

In the 100 metaphase cells examined after in situ hybridization, 26.6% of silver grains were located on chromosome 11; 80% of them mapped to the q12–q13.2 region chromosome 11 long arm. These results allow us to localize the gene hH-Rev 107 to the 11 q12–q13.2 region of the human genome (Figure 4). A similar result was obtained by Husmann (Husmann et al., 1998), on chromosome 11 q11–12, by FISH.

Figure 4. Chromosomal localization of the human gene hH-Rev107 .

Idiogram of the human G-banded chromosome 11 illustrates the distribution of labelled sites with the hH-Rev107–3 probe

Human EST database analysis

The hH-Rev107-3 cDNA nucleotide sequences were searched for homology in the dbEST database using the BLAST program. We only detected EST sequences that were 100% identical to hH-Rev107-3 cDNA, no mutated sequence; this validates the sequence of hH-Rev107-3. From 3’ end EST analysis, we found an EST population produced from different human tissues, similar to the end of hH-Rev107-3 cDNA. No spliced mRNA was found from EST analysis. Taken together, these results indicated the presence of only one gene for hH-Rev107 in the genome and the two mRNA are likely transcribed from two different promoters with the use of two polyadenylation sites.

Localization of hH-Rev107 gene expression in human testis

Detection of hH-Rev107 transcripts was further characterized in human testis by in situ hybridization of DIG-labelled hH-Rev107 antisense oligonucleotide (Figure 5, B and D). The presence of the hH-Rev107 mRNA was confined to the seminiferous tubules in association with germ cells and more precisely in round spermatids. In situ hybridization of DIG-labelled hH-Rev107 sense oligonucleotide to human testis showed no staining (Figure 5, A and C).

Figure 5. Localization of hH-Rev107 gene expression in normal human testis.

Paraffin-embedded testis sections (tissue number c3, Table 1) were hybridized with: A and C, hH-Rev107-3 sense DIG-oligonucleotide, B and D, hH-Rev107-3 antisense DIGoligonucleotide. Sections were counterstained with methyl green. Original magnifications 200 X (A and B), 630 X (C and D). Arrows indicate signal.

Expression of hH-Rev107 gene in CIS and testis tumors

We analyzed, by Northern blot, the hH-Rev107 mRNA level in a panel of 4 human TGCTs (2 seminomas and 2 non-seminomas), 2 pure CIS and 4 normal testes (Figure 6) (Table 1). A strong expression of 1 and 0,8 kb mRNA was detected in the four normal testis samples as well as in the two CIS. In contrast, there was a lack of expression of these two hH-Rev107 transcripts in all the TGCTs tested. No significant variation in the amount of loaded mRNA was detected with the ß-actin control probe (data not shown).

Figure 6. Expression of hH-Rev107 gene in testis tumors.

Northern blot analysis was performed with 3 μg of poly(A)+ RNA per lane isolated from 4 normal testes (N), 2 carcinoma in situ (CIS), 2 non-seminoma (NSE) and 2 seminoma (SE). 4789-IIA to 4808-I corresponds to the tissue numbers described in Table 1. The blot was hybridized with ³²P labeled hH-Rev107-3 cDNA. Arrows on the left indicate the size of the mRNA.

Table 1. Description of the testis samples.

The samples are identified by their tissue numbers, (I-macroscopically tumor; IImacroscopically surrounding tissues), the diagnosis and the histology.

Tissue number	Diagnosis	Histology of testis
ISH
c3	normal
5423-II	CIS	CIS (80 %) + tubules with spermatogenesis (20 %)
5294-II	CIS	CIS (50 %) + tubules with spermatogenesis (50 %)
5327-II	Seminoma	Seminoma (30 %) + tubules with spermatogenesis (70 %)
Northern blot
4789-IIA	normal
4800	normal
4749	normal	atrophic testis, few tubules with spermatogenesis
4789-IIB	normal
4825-II	CIS	CIS (100 %)
4818	CIS	CIS (100 %)
4846	Non-Seminoma	Embryonal Carcinoma
4825-I	Non-Seminoma	Teratoma + Endodermal Sinus Tumor
4464-I	Seminoma	sample may contain a few CIS tubules
4808-I	Seminoma
RT-PCR
1575	normal
3522-II	normal
4084-II	normal	partly atrophic testis
4232-II	CIS	CIS (20 %) + normal or atrophic tubules (80%)
3769-II	CIS	CIS (50 %) + normal (50 %)
4227-II	CIS	CIS (50 %) + normal spermatogenesis (50%) ; may also contain tumor
4237-II	CIS	CIS (50 %) + normal (50 %) ; may also contain tumor
4296-II	CIS	CIS (80–90 %) + normal (10–20 %)
5423-II	CIS	CIS (90–100 %) + normal (10 %)
4134-II	CIS	CIS ; no spermatogenesis seen, may also contain Seminoma
4217-II	CIS	CIS (20–30%) ; may also contain Seminoma, the rest is partly atrophic
4037-II	CIS	CIS (30 %) ; may contain tumor, no spermatogenesis seen
5220-II	CIS	CIS (50–70 %) ; may contain foci of Seminoma, no spermatogenesis seen
5304-II	CIS	CIS (100 %) ; no spermatogenesis seen
4332-II	CIS	CIS (100 %) ; no spermatogenesis seen, but may also contain some tumor
4084-I	Non-Seminoma	Teratoma
4296-IA	Non-Seminoma	Teratoma
4227-I	Non-Seminoma	Teratoma + Embryonal Carcinoma
4296-IB	Non-Seminoma	Teratoma + Embryonal Carcinoma
5239-I	Non-Seminoma	Teratoma + Embryonal Carcinoma
5303-I	Non-Seminoma	Embryonal Carcinoma
4134-I	Seminoma
4332-I	Seminoma
5220-I	Seminoma
5225-I	Seminoma
5423-I	Seminoma
3522-I	combined tumor	Seminoma + Teratoma + Endodermal Sinus Tumor
3769-I	combined tumor	mainly Seminoma + very little Embryonal Carcinoma
4037-I	combined tumor	Seminoma + Teratoma + Endodermal Sinus Tumor + Embryonal Carcinoma
4217-I	combined tumor	Seminoma + Embryonal Carcinoma
5158-I	combined tumor	Teratocarcinoma + Seminoma

To assess the localization of hH-Rev107 expression in CIS and to verify the lack of signal in tumor, we performed an in situ hybridization of a DIG-labeled hH-Rev107 antisense oligonucleotide on two CIS and one seminoma (Table 1). The presence of the hH-Rev107 mRNA was detected within tubules in association with germ cells of the CIS (Figure 7, E, F and G). No signal was detected in seminoma (Figure 7, H). In situ hybridization of DIG-labelled hH-Rev107 sense oligonucleotide to CIS and tumor showed no staining (Figure 7, A–D).

Figure 7. Localization of hH-Rev107 gene expression in human CIS and testis tumor.

Paraffin-embedded testis sections were hybridized with the hH-Rev107-3 sense DIGoligonucleotide (A, B, C and D) or the antisense hH-Rev107-3 DIG-oligonucleotide. (E, F, G and H). Sections were counterstained with methyl green. Original magnifications 200 X. A, E: CIS, tissue number 5294-II, B-G: CIS, tissue number 5423-II, D, H: seminoma, tissue number 5327-II (Table 1).

In order to quantify the decrease of the hH-Rev107 mRNA level occurring in testicular tumors, we performed a semi-quantitative RT-PCR on 16 TGCTs: 6 non-seminomas, 5 seminomas and 5 combined tumors (mixed seminoma and non-seminoma in the same sample) and compared the result with surrounding tissues (Table 1). A unique PCR product, with the expected size of 571 bp, was obtained suggesting that the two hH-Rev107 mRNA have a common sequence between the positions 319 and 890 (referred to accession number X92814). This result is in agreement with the hypothesis of two mRNA transcribed from two different promoters with the use of two polyadenylation sites. RT-PCR of ß2 microglobulin was performed in parallel as a control of the amount of starting mRNA. The relative hH-Rev107/ß2 microglobulin ratio has been estimated on the mean of three independent experiments. In all tumors tested, we observed a decrease of the hH-Rev107 mRNA level, as compared to the surrounding tissues (Figure 8). It is noteworthy that surrounding tissues exhibit a great heterogeneity based on histology results and can be classified in three groups: group 1, corresponding to normal tissues (N); group 2, corresponding to normal tissues containing CIS (N +%CIS) and group 3, corresponding to CIS with variable amount of tumor (CIS +%T) (Table 1). We observed a 9-fold decrease of the hH-Rev107 mRNA level between the group 1 (mean = 58; SE = 21) and the group of TGCTs (mean= 6,18; SE = 7,74), which is statistically significant (P value = 0,0021) (Mann-Whitney test). It is worth mentioning that there is no difference (P = 1) between the group 1 (mean = 58; SE = 21) and the group 2 (mean = 53,4; SE = 8,8) whatever the percentage of CIS present in the parenchyma tissue. This result is in good agreement with the hH-Rev107 strong signal observed in pure CIS by Northern blot and by in situ hybridization. On the contrary, we observed a statistical difference (P = 0,038) between the group 1 (mean = 58; SE = 21) and 3 (mean = 14,16; SE = 3,3) that most likely reflects the presence of tumor tissues in these CIS.

Figure 8. Semi-quantitative analysis of hH-Rev107 gene expression in testis tumors.

The quantitative hH-Rev107 and ß2 microglobulin gene expression was performed by RT-PCR on total RNA extracted from surrounding testicular tissues: 3 normal testicular tissues ( ), 6 normal tissues with increasing amounts of CIS ( ) and 6 CIS with variable amounts of tumors ( ) and from testicular tumors: 5 seminomas ( ), 6 non-seminomas ( ) and 5 combined tumor ( ), samples are identified by their number, I-macroscopically tumor; II-macroscopically surrounding tissues, (Table 1). The PCR products were revealed after blotting and hybridization with specific ³²P oligonucleotide. The hH-Rev107 / ß2 microglobulin gene expression ratio is referred to the normal sample 1575 (in percentage). The results are based on the mean of three independent experiments.

Discussion

In this study, we have cloned a human testis hH-Rev107-3 cDNA and provided evidences indicating that only one hH-Rev107 gene is present in the human genome and encodes two mRNA. From the results of sequencing, EST database analysis and RT-PCR, it is inferred that the two mRNA are likely transcribed from two promoters with the use of two polyadenylation sites. As ESTs identical to hH-Rev107-3 cDNA 3’ end were obtained from different tissues, the use of the 3’ polyadenylation site found in hH-Rev107-3 cDNA is not testis-specific. The deduced protein sequence did not exhibit a particular motif, which would be relevant to a specific biochemical function.

The physiological antiproliferative and tumor-suppressing activities of the Hrev-107 protein were demonstrated by rH-Rev107 over expression in H-Ras transformed cells in vitro and in vivo (Sers et al., 1997). In accordance with these observations, we have attempted to analyze the expression of hH-Rev107 gene in normal human testis and diverse types of testicular tumors. Given the unique features of this natural model system for simultaneous analyses of human spermatogenesis, cell division and differentiation and tumorigenesis (Skakkebaek et al., 1987), we shall consider our present findings in relation to the diverse patterns of differentiation, and then discuss their potential implications for the pathogenesis of testicular cancer.

The process of germ cell differentiation begins in embryo when primordial germ cells differentiate into gonocytes, which subsequently mature into spermatogonia during infancy and early childhood. At the onset of puberty, gonadotropic stimulation induces real spermatogenesis. Using in situ hybridization, we localized hH-Rev107 mRNA in the round spermatids in human testis. These cells belong to the differentiation phase just before the condensation of the nuclei occurring in elongating spermatids. The proteins translated just after meiosis correspond to a limited number of genes that are involved in the final differentiation of spermatozoon (Erickson, 1990) (Schmidt & Schibler, 1995) (Hecht, 1998) (Braun, 1998). The post-meiotic expression of hH-Rev107 gene is still unclear, but the restricted expression pattern of hH-Rev107 mRNA in round spermatids is consistent with a specialized role during spermiogenesis. This hypothesis is in good agreement with the putative function of H-Rev107 protein, which is to suppress cell proliferation and/or to maintain the cell in a differentiated state (Husmann et al., 1998).

Several studies have provided pieces of evidence in line with the hypothesis that inhibition of gonocyte differentiation into spermatogonia during fetal life or infantile period may be crucial for the development of CIS (Skakkebaek et al., 1987) (Rajpert-De Meyts et al., 1998). CIS cells may differentiate, in the young adulthood along pathways resulting in several histological tumor types such as seminoma and non-seminoma. Using in situ hybridization and semi-quantitative RT-PCR, we demonstrated the presence of hH-Rev107 mRNA in preinvasive CIS but not in overt tumors (seminoma, non-seminoma and combined tumor). This result is in favor of the putative role for H-Rev107 protein in control of cell division, thus its absence may permit a proliferative process. There are very few reliable molecular markers that correlate closely with progression of the TGCTs. Among the cell cycle regulators studied in testicular cancer, p18^INK4C, a CDK inhibitor and candidate tumor suppressor, has the same pattern of expression in TGCTs and CIS as hH-Rev107 gene. But its expression, within normal tubules, is not restricted to a specific phase of spermatogenesis; (Bartkova et al., 2000). Due to its specific expression in the normal testis, H-Rev107 protein could be considered as a candidate regulator potentially involved in the process of post-meiotic maturation stages of human spermatogenesis. The relative down-regulation of hH-Rev107 gene in TGCTs in comparison to CIS might serve as indicator of progression of tumorigenesis.

All studies of H-Rev107 in rat or in human demonstrated an inverse correlation between H-Rev107 and the H-Ras gene expressions (Hajnal et al., 1994; Husmann et al., 1998). It is well documented that Ras plays a role in the normal cell growth differentiation and activated Ras oncogenes have been identified in a wide variety of human malignancies (Bos, 1989). In testis tumors, no mutation was observed in the coding sequence of H-Ras and no significant variation was observed in the mRNA level of H-Ras in seminoma as well as embryonal carcinoma or teratoma (Mulder et al., 1989) (Moul et al., 1992) (Shuin et al., 1994). We performed semi-quantative RT-PCR for H-Ras on a small series of tumors with their surrounding tissues. No evident correlation could be demonstrated between hH-Rev107 and H-Ras gene expression (data not shown).

To explain the loss of the hH-Rev107 mRNA in TGCTs, we examined if there were genomic rearrangements in DNA from human primary tumors. No difference was observed when comparing Southern blot profiles of DNA obtained from human testis tumor tissues and from human blood (data not shown). As hH-Rev107 mRNA was down-regulated in human testis tumors without gross alteration at the DNA level, the loss of expression may be due to: i) a limited genomic modification (mutation or small deletion) in the promoter region; ii) the absence of regulatory protein which is upstream in the cascade of cell proliferation pathway. Because repression of the rH-Rev107 gene occurred within the tumor cell lines without any modification at the DNA level, Sers considered rH-Rev107 as belonging to the class II tumor suppressor gene (Sers et al., 1997). The concept of class II tumor suppressor gene was first developed by Lee (Lee et al., 1991) when loss of function is caused by a regulatory block to expression. The study of this type of gene is of clinical interest as such a gene may theoretically be up regulated by drugs or special treatments.

A search for homology in the GenBank database revealed that H-Rev107 might be a member of a gene family. The second member would be the retinoic acid receptor responder 3 gene (RARRE3, also called TIG3) (DiSepio et al., 1998), which was also cloned as the retinoid-inducible gene (RIG1) (Huang et al., 2000) and as hHRev107-2 (Husmann et al., 1998). The nucleotide sequence of this second member is 54% identical to hH-Rev107 gene. The tissue expression pattern and the gene regulation of TIG3/RIG1 gene are different from the one of hH-Rev107 gene (DiSepio et al., 1998) (Huang et al., 2000) But TIG3 would also have an anti-proliferative function and may function as a tumor suppressor in both normal and malignant epidermal differentiation (DiSepio et al., 1998) (Duvic et al., 2000).

The gene family is localized on the same chromosome 11 in 11q12 region for hH-Rev107 gene (this work, (Husmann et al., 1998)) and 11q23 region for RARRE gene (DiSepio et al., 1998). The distribution of genes along 11q is non-uniform. Three regions of long arm of this chromosome, 11q13, 11q22–23 and 11q24 have a higher density of genes. The reasons for this clustering are not completely explained, though an ancestral pericentric inversion juxtapositioning the gene-rich telomeric part of the centromere to its current location at 11q13 has been invoked (Saccone et al., 1992). Since hH-Rev107 and RARRE genes are localized on the same region, they are likely to be related genes coming from the same ancestral sequence duplicated on this chromosome.

H-Rev107 gene belongs to a gene family, which plays a role in the control of cell proliferation. It would be important to determine exactly the regulatory pathways in which these genes are involved. This could improve our understanding of cell proliferation in tumorigenesis. In addition, analysis of their promotor regions may contribute in the development of new drugs that can induce hH-Rev107 expression and thereby exert an antiproliferative effect.

Materials and methods

Tissues, genomic DNA and RNA preparation

Small specimens (0.1 to 2 mg) of human testicular tumors and surrounding tissues were obtained at the time of surgery from patients undergoing orchidectomy. Immediately after macroscopic evaluations, tissues were frozen at −80°C. Histological evaluation was based on routine pathology reports and included seminomas, non-seminomas, mixed germ cell tumors (seminoma and non-seminoma) and CIS. Normal testicular tissues were obtained from patients undergoing bilateral orchidectomy for treatment of prostatic carcinoma (Table 1).

Genomic DNA was obtained from peripheral blood cells and prepared as previously described (Bulle et al., 1996).

Total RNA was isolated from frozen tissue specimens using the method of Chomczynski (Chomczynski & Sacchi, 1987).

cDNA cloning and sequencing

The hH-Rev107-3 cDNA (Accession number: AF317086) was isolated from a human testis library and the complete sequence of the cDNA (676 bp) was determined as previously described (Pawlak et al., 1995).

Northern and Southern blot analysis

Human multiple-tissue Northern blots were purchased from Clontech (Palo Alto, Calif.). Poly(A)+ RNA from different human tumor tissues were analyzed by Northern blot as previously described (Levy et al., 1996). 20 μg of genomic DNA were digested by BamH I, Kpn I, Pst I, EcoR I and Hind III. The fragments were separated on 0.8% agarose gel and blotted onto Hybond N membrane (Amersham) as previously described (Bulle et al., 1996).

The blots were hybridized with the hH-Rev107-3 or β-actin cDNA (Accession number: NP 001092) sequences labelled by random priming, with a ³²P dCTP, according to the Megaprime protocol (Amersham).

Northern blots were washed in 0.05 X SCC, 0.1% SDS at 65°C and Southern blots in 0.5 X SCC, 0.1% SDS at 68°C.

Chromosomal localization

In situ hybridization was carried out on chromosome preparations obtained from phytohemagglutini-stimulated lymphocytes and hybridized with the tritium labelled hH-Rev107 -3 clone (Mattei et al., 1985). The slides were exposed for 20 days at chromosome spreads were first stained with buffered giemsa solution and metaphases photographed. R-banding was then performed by the fluorescencephotolysis-giemsa (F.P.G.) method and metaphases rephotographed before analysis.

In situ hybridization

In situ hybridization was carried out on sections of paraffin-embedded testis. After deparaffinization in xylene and rehydratation, sections were incubated in 50% acetic acid for 30 sec, and then rinsed in sterile water. Sections were digested with 10 μg/mL proteinase K in 100 mM Tris-HCl pH 7.5, 1 mM CaCl₂ for 15 min at 37°C. Slides were then washed in sterile water, and postfixed in 2% paraformaldehyde (PFA) for 20 min at room temperature. After one wash in sterile water, sections were dehydrated in gradual ethanol baths. Following air-dry treatment, sections were prehybridized for 2 h at 37 °C in 100 μL of hybridization buffer (4% deionized formamide, 1 X Denhardt’s solution, 1.5 X SSC and 330 μg/mL yeast tRNA) in a humid chamber. In situ hybridization was performed for 16 h at 37 °C in 20 μL of the hybridization buffer containing 10% dextran sulfate and 25 ng DIG-labeled oligonucleotide antisense to the hH-Rev107 mRNA (position 571–598, accession number: X92814) in a humid chamber. After hybridization, the sections were briefly rinsed in 2 X SSC at room temperature, then in 2 X SSC for 1 h at 37°C, and finally in 1 X SSC for 30 min at room temperature. Hybridization signals were detected using DIG Nucleic Acid Detection Kit (Boehringer Mannheim) with alkaline phosphatase-conjugated anti-DIG antibody, nitroblue tetrazolium salt/bromo-chloro-indolyl-phosphate (NBT/BCIP) as chromogen, and high molecular weight polyvinyl alcohol (10%). Sections were then counterstained using methyl green. Specificity experiments included hybridization with the DIG-labelled sense oligonucleotide.

Semi-quantitative RT-PCR analysis

Amplification of hH-Rev107 mRNA was performed from primers, which map at position 320–345 (5’-TTAATGACAATAAATCCCTGCTCCCC-3’) and 869–890 (5’- CTCAAGAAACAAGCGACAAAAG-3’) (Accession number: X92814) and generates a 571 bp fragment. Primers used for the amplification of ß2 microglobulin mRNA (Accession number: XM_007650) map at position 11–36 (5’-GAGATGTCTCGCTCCGTGGCCTTAGC-3’) and 358–385 (5’-GGATCGAGACATGTAAGCAGCATCATGG-3’) and generate a 375 bp fragment. All primers were synthesized commercially (Genset, France). RT-PCR reactions were carried out with 5 or 1 ng of total RNA for hH-Rev107 or ß2 microglobulin amplification, respectively using the RT-PCR access kit (Promega). RT-PCR was performed by using 30 or 40 amplification cycles in a thermal cycler (Perkin Elmer) at an annealing temperature of 54°C or 58°C, respectively. The PCR products were separated on a 2% agarose gel and transferred onto Hybond N membrane (Amersham). The specific products were revealed by hybridization to a ³²P oligonucleotide probe specific of hH-Rev107 mRNA (5’-CTGACTGACAAGGCCATCGTG-3’) (nt 567–587) or ß2 microglobulin (5’-GAAGTTGACTTACTGAATGGAGAG-3’) (nt 179–203). The signal was quantified in an Instant Imager using the Imager program. The amount of RNA, the number of amplification cycles and the conditions of exposure were selected in order to obtain a signal in a linear concentration range.