Malignant germ cell tumours of the testis express interferon-γ, but are resistant to endogenous interferon-γ

Cytokines possess discrepant effects on tumour cells varying from anti- to proapoptotic activities. We recently reported that testicular germ cell tumours (TGCT) express a functional form of the proinflammatory cytokine interferon-gamma (IFNγ). The present study asked whether TGCT-derived IFNγ influences survival or death of neoplastic germ cells. Analysis of TGCT cell lines demonstrated that they expressed and secreted IFNγ, but were resistant to the endogenous IFNγ since neutralisation of IFNγ by a specific blocking antibody had no influence on the proliferation and/or the degree of apoptosis of tumour cells. To study mechanisms providing tumour resistance to endogenous IFNγ, we analysed primary TGCT and two human TGCT cell lines (NTERA and NCCIT) for the expression of IFNγ receptor and for the level of phosphorylation of the signal transducer and activator of transcription (STAT)-1. In situ hybridisation, immunocytochemistry, Western blot analysis and flow cytometry indicated that primary TGCT as well as NCCIT and NTERA cell lines expressed the heterodimeric cell surface IFNγ receptor which consists of both 90-kDa α- and the 85-kDa β-chains. However, the downstream transcription factor STAT-1 was not phosphorylated constitutively, indicating that STAT-1 is not activated by the endogenous IFNγ. Upon application of recombinant human IFNγ in excess, however, STAT-1 was phosphorylated and the interferon regulatory factor-1 (IRF-1) was induced, suggesting that both IFNγR and STAT-1 are functionally intact in TGCT. Altogether our results suggest that despite secreting biologically active IFNγ, the concentration of the endogenous IFNγ is too low to stimulate the IFNγR/STAT signalling pathway in TGCT in an autocrine and/or paracrine manner.

Interferon-g (IFNg) is a pleiotropic cytokine mainly secreted by activated T lymphocytes and natural killer (NK) cells (Mosman and Sad, 1996). The cellular response to IFNg is mediated by a heterodimeric cell surface receptor (IFNgR), which consists of two subunits, the 90-kDa a chain (IFNgRa) and the 85-kDa b chain (IFNgRb) (Valente et al, 1992;Pestka et al, 1997). Ligation of IFNg to the a chain clusters the neighbouring b chain of the IFNgR followed by Janus kinase-mediated phosphorylation of the signal transducer and activator of transcription (STAT)-1, the homodimer of which migrates to the nucleus and stimulates gene transcription such as the interferon regulatory factor-1 (IRF-1) (Haque and Williams, 1998;Wagner et al, 2002).
After the first report of its ability to protect cells from viral infection, IFNg has been demonstrated to play a crucial role in host defence, inflammation and autoimmunity. For instance, IFNg has been shown to augment antigen presentation by upregulation of major histocompatibility complex class I and II molecules, to induce proinflammatory cytokines in effector cells and to orchestrate leukocyte -endothelium interaction by upregulation of adhesion molecules (Boehm et al, 1997). Regarding induction of apoptosis, however, IFNg seems to play a paradoxical role in different normal and neoplastic cell types. Whereas in normal macrophages and neoplastic myeloid and NK cells, IFNg prevents apoptosis (Lotem and Sachs, 1996;Mizuno et al, 1999;Xaus et al, 1999), it enhances apoptosis of tumour cells in malignancies such as pancreatic carcinoma, colon carcinoma and ovarian carcinoma (Adachi et al, 1999;Burke et al, 1999;Detjen et al, 2001).
Testicular germ cell tumours (TGCT) are the most common solid malignancy in young males from 20 to 40 years old and represent a heterogeneous group of different histological entities composed of seminomatous and nonseminomatous tumours (Ulbright, 1993) that are almost all infiltrated by various numbers of T lymphocytes (Torres et al, 1997). In an attempt to investigate mechanisms navigating T lymphocytes into the TGCT, we observed that neoplastic germ cells express IFNg (Schweyer et al, 2002). Based on this evidence and considering the dual role of IFNg in apoptosis of neoplastic cells, we asked whether TGCT-derived IFNg possesses any effect on survival or death of TGCT.

Tissue samples
Tumour specimens were obtained from 12 patients who underwent orchiectomy for testicular tumour. Patients had not received any chemotherapeutic or immunomodulatory treatment before operation. The mean age at the time of operation was 36.8 years, ranging Received 13 March 2003;revised 16 May 2003;accepted 30 June 2003 from 28 to 55 years. Tumour tissues were classified according to the classification system of the World Health Organization (six cases of pure seminoma, six cases of combined tumour containing nonseminomatous and seminomatous components) (Mostofi, 1980). Probes from normal testes were obtained from three patients who underwent bilateral orchiectomy because of prostatic cancer. Two blocks of each testis were immediately frozen in liquid nitrogen and stored at À801C until analysed by reverse transcription -polymerase chain reaction (RT -PCR). Samples of each tissue specimen were also fixed in neutral formalin and processed for histology, immunohistochemistry and in situ hybridisation (ISH).

Human monocytes
Blood monocytes were obtained from healthy volunteers as described previously (Schweyer et al, 2002). A total of 2.5 Â 10 5 monocytes were cultured in basis medium containing RPMI 1640, 2 mM L-glutamine, 1% penicillin and streptomycin, and 15% human serum for 1 day. The cells were then stimulated for 12 h by recombinant human IFNg (rhIFNg; R&D Systems) at a final concentration of 200 U ml À1 before being analysed for the expression of IFNgRa and IFNgRb and for the phosphorylation of STAT-1 (see below).

Human umbilical vein endothelial cells
Human umbilical vein endothelial cells (HUVEC) were isolated and cultured as described previously (Schweyer et al, 2002). The cells were then stimulated for 9 h by rhIFNg (1000 U ml À1 ) and rhTNFa (100 U ml À1 ) before being studied for the expression of IRF-1 as described elsewhere (Wagner et al, 2002).
To assess the proliferation activity of NCCIT and NTERA cells, plates were pulsed with 1 mCi per well of [ 3 H]thymidine. After 24 or 48 h, the cells were collected by an automated Inotech cell harvester (Dunn, Ansbach, Germany) and the radioactivity was measured with a b-counter (Hewlett-Packard, Meriden, CT, USA). Results were expressed as counts per minute (c.p.m.)7s.e.m. Each experiment was performed in triplicate and was repeated three times.
To determine the apoptotic rate of NCCIT and NTERA cells, they were harvested at 24 or 48 h after application of neutralising antibody or control IgG. Then the cells underwent May -Giemsa -Grunwald (MGG) staining, or immunocytochemistry for the active form of caspase-3, or in situ end labelling (ISEL) for DNA fragmentation. For control, NCCIT cells were incubated with 50 mM cisplatin (Holzkirchen, Germany) for 24 h and proved to be apoptotic as described previously (Burger et al, 1999). Each experiment was performed in triplicate and was repeated three times.

May -Giemsa-Grunwald staining
Centrifuged cells (2 Â 10 3 ) were dried for 24 h, fixed in 100% acetone for 10 min, stained with MGG and embedded in 'Super-Mount Medium'. Apoptotic cells were identified by cellular shrinkage and nuclear condensation and fragmentation.
Immunohistochemistry Immunohistochemical reactions for IFN-gRa and IFNgRb were performed on frozen sections. After incubation with the primary antibody, the sections were incubated with a horseradish peroxidase (HRP)-conjugated biotin -streptavidin amplified system (Dako) and the signals were visualised with 3,3 0 -diaminobenzidine (DAB; Dako) as described previously . Immunohistochemical reactions for PLAP and cytokeratin were performed on paraffin-embedded serial sections. After incubation with the primary antibody, the sections were incubated with an AP-conjugated or an HRP-conjugated biotin -streptavidin amplified system (Dako) and the signals were visualised with fast red or AEC (Dako) as described elsewhere . All samples were counterstained with Meyer's haematoxylin, mounted in Super Mount Medium and analysed by light microscopy. Controls were stained as above omitting primary or secondary antibodies.
Immunofluorescence and immunocytochemistry Fresh tumour tissue was mechanically dissociated in a suitable volume of RPMI 1640 supplemented with 100 IU ml À1 penicillin, 100 mg l À1 streptomycin and 2 mM L-glutamine. Cells were spun down at 1500 r.p. m. for 5 min and the pellet was resuspended in 500 ml ice-cold phosphate-buffered saline (PBS). Then, 5 Â 10 5 cells were cytocentrifugated on slides coated with 2% 3-aminopropyltriethoxy-silane, dried for 24 h and fixed with 100% acetone for 10 min at room temperature. For immunocytochemical staining, cells were incubated for 30 min with a primary antibody against IFNg, IFNgRa, IFNgRb, active form of caspase-3 or p-STAT-1. To visualise bound primary anti-IFNg-antibody, cells were incubated with an FITClabelled rabbit anti-goat IgG (working dilution 1 : 500) for 1 h (Dako), mounted with 'Fluorescent Mounting Medium' s and examined using fluorescence microscopy. To visualise bound primary anti-IFNgRa, anti-IFNgRb, anti-active caspase-3 and p-STAT-1 antibodies, cells were incubated with AP-conjugated or HRP-conjugated biotin -streptavidin amplified system (Dako) as described previously . Fast red or AEC (Dako) was applied as chromogen. After signal development, cells were counterstained with Meyer's haematoxylin, mounted in Super Mount Medium and analysed by light microscopy. IFNg-stimulated human monocytes served as positive control.
In situ end labelling (ISEL) Fixed centrifuged cells (2 Â 10 3 ) were incubated with TBS (50 mM Tris-HCl; 150 mM NaCl; pH 7.5) containing 10% FCS and 0.3% H 2 O 2 for 15 min. The cells were then incubated for 60 min at 371C with 50 ml of the labelling mix (250 U ml À1 terminal transferase, 20 ml ml À1 Digoxigen-DNA labelling mix at 10 Â concentration and 1 mmol l À1 CoCl 2 in reaction buffer for terminal transferase (Roche, Mannheim, Germany)). After rinsing in TBS, the cells were blocked with 10% FCS and incubated for further 60 min with a rabbit HRP-conjugated F(ab) 2 fragment against digoxigenin (working dilution: 1 : 500, Dako). 3,3 0 -Diaminobenzidine was next applied as chromogen. Cells with fragmented DNA revealed nuclear brown signals. DNA-fragmented cells with intact plasma membrane were considered to be apoptotic. Negative controls were stained as above but without terminal transferase.

Identification and quantification of apoptosis and statistical analysis
Apoptotic cells were identified by nuclear condensation and fragmentation in MGG staining, by positive cytoplasmic signals for the active form of caspase-3 in immunocytochemistry or by DNA fragmentation in ISEL. All experiments were performed in triplicate and were repeated three times with similar results.
Percentage of apoptotic cells was calculated as the ratio of apoptotic cells to 500 cells counted. Results are expressed as the mean 7s.e.m. The differences were analysed with a t-test and were considered significant at Po0.05.

RNA extraction and RT -PCR
Total RNAs were extracted from TGCT cell lines as well as from IFNg-stimulated human monocytes using the Qiagen RNA isolation kit (Qiagen, Hilden, Germany), digested with DNAse I and transcribed to cDNA using oligo-d(T) primers and SuperScript II reverse transcriptase (RT) (Life Technologies). In brief, 1 -5 mg of total cellular RNA was incubated for 50 min at 371C with 50 U of RT and 20 U placental RNase Inhibitor in a 20-ml volume containing 2.5 mmol l À1 oligo-d(T) primers, 5 mmol l À1 MgCl 2 , 50 mmol l À1 KCl, 10 mmol l À1 Tris-HCl and 1 mmol l À1 of each of the deoxynucleoside-triphosphate, heated to 701C and subsequently cooled to 51C.
Subcloning of IFNc, IFNcRa and IFNcRb cDNA, preparation of cRNA probes and nonradioactive ISH For preparation of riboprobes, fragments of the human IFNg, IFNgRa and IFNgRb cDNA were subcloned into pBluescript II KS þ phagemid (Stratagene, CA, USA). The subcloned fragment served as template for in vitro transcription of digoxigenin-11-dUTP labelled antisense and sense probes, which were generated by virtue of T3-and T7-polymerase according to the manufacturer's instructions (Roche).
In situ hybridisation for IFNg mRNA was performed according to the method described previously (Schweyer et al, 2002). After signal detection, specimens were subjected to double-staining by indirect immunofluorescence for TGCT markers PLAP and cytokeratin, as described above (Schweyer et al, 2002). To detect and visualise IFNgR transcripts, the Catalyzed Signal Amplification System was used as recommended by the manufacturer (Dako). DAB (Dako) was applied as chromogen. After signal detection, specimens were mounted in Super Mount Medium. For each tissue specimen, sense riboprobes were applied as controls and proved to be negative.

Flow cytometry
NCCIT-, NTERA-and IFNg-stimulated human monocytes were spun down at 1500 r.p. m. for 5 min and the pellet was resuspended and washed in 500 ml of ice-cold TBS. The cell suspension was then centrifuged again at 1500 r.p.m. and resuspended in 10 ml RPMI 1640 cell culture medium. Cells were then transferred to 96-well round-bottom microtitre plates (Nunc, Wiesbaden, Germany), which had been precoated by blocking buffer (10% heatinactivated rabbit serum and 0.1% NaN 3 in PBS). Thereafter, they were incubated for 30 min with a monoclonal antibody against human IFNgRa, a polyclonal antibody against IFNgRb or irrelevant mouse/rabbit IgG (Dako). After washing with PBS, the cells were incubated for further 30 min with an FITC-conjugated goat antimouse or a PE-conjugated goat anti-rabbit antibody (Dako). Cells were then washed again, fixed with 2% paraformaldehyde and subjected to flow cytometric analysis using FACStar plus (Becton Dickinson, San Jose, CA, USA).

ELISA
Extracellular IFNg level in culture supernatants of TGCT cell lines was measured via commercial ELISA (R&D) as recommended by the manufacturer. The lower limit for assay was IFNg o3 pg.

Testicular germ cell tumours express IFNc
To determine the expression and cellular localisation of IFNg in TGCT, in situ analyses were performed. Whereas no IFNg expression was noted within the normal testes (data not shown), nonradioactive ISH revealed that numerous tumour-infiltrating mononuclear cells and almost all neoplastic germ cells independent of their histological type expressed IFNg mRNA (Figure 1). Consequently, we asked whether IFNg mRNA is translated into IFNg protein. To answer this question, tumour cells were isolated from TGCT and subjected to immunofluorescence. Results demonstrated that primary TGCT not only produce IFNg mRNA but also IFNg protein (Figure 1).
For an autocrine effect, IFNg must be secreted by tumour cells. To prove this, we analysed two well-established human TGCT cell lines NCCIT and NTERA, for the expression and secretion of IFNg. Reverse transcription -polymerase chain reaction RT -PCR showed that both cell lines expressed IFNg mRNA (data not shown). Applying ELISA, significant amounts of secreted IFNg (343 -112 pg ml À1 ) were found in culture supernatants of NCCIT and NTERA, as described previously (Schweyer et al, 2002) (data not shown). Based on this background, we next asked whether TGCT-derived IFNg influences the proliferation and/or apoptosis of the testicular tumour cells in an autocrine manner.

Testicular germ cell tumour-derived IFNc does not affect proliferation or apoptosis of TGCT cell lines
To study the effect of the secretory IFNg on multiplication and/or death of TGCT cells, endogenous IFNg was blocked by adding different concentrations of a neutralising antibody to the cultured NCCIT and NTERA cells for 24 or 48 h. Results from the [ 3 H]thymidine assay demonstrated that IFNg has no effect on proliferation of the TGCT cell lines when compared to controls. In accordance, immunocytochemistry for the active form of caspase-3, ISEL for DNA fragmentation and MGG staining for the detection of nuclear condensation and fragmentation showed that although the apoptosis rate of tumour cells was increased following IFNg neutralisation when compared to controls, the differences among the groups remain, however, not significant (P40.05). Figure 2 demonstrates results from proliferation and apoptosis assays on TGCT cell lines 24 h after IFNg neutralisation. Similar results were also seen 48 h after beginning IFNg neutralisation (data not shown).

Testicular germ cell tumours express both IFNcR subunits a and b
Since it is known that IFNg mediates its effects by binding to a specific high-affinity receptor (Boehm et al, 1997), we asked whether the unresponsiveness of TGCT to the endogenous IFNg is due to the lack or dysfunction of IFNgR on neoplastic germ cells.
Nonradioactive ISH revealed that in addition to numerous tumour-infiltrating mononuclear cells, almost all tumour cells independent of their histological type expressed IFNgRa and IFNgRb mRNAs (Figure 3). Consequently, the expression of IFNgRa and IFNgRb proteins in primary tumours was examined by immunohistochemistry and immunocytochemistry. Results illustrated both IFNgR subunits on neoplastic germ cells in primary TGCT as well as on the surface membrane of tumour-infiltrating mononuclear cells (Figure 3). Within the normal testes, however, no expression of IFNgRa or IFNgRb was noted in germ cells (Figure 3). In situ hybridisation and immunohistochemistry, however, indicated that IFNgRb, but not IFNgRa, was expressed in Leydig cells (Figure 3). In addition to primary tumours, both human TGCT cell lines were also examined for the IFNgR expression. RT -PCR, Western blot and FACS analyses demonstrated that NCCIT and NTERA cell lines express IFNgRa and IFNgRb (Figure 4).

Testicular germ cell tumours lack STAT-1 activation
After demonstration of the IFNgR expression on TGCT, the functional activity of the IFNgR was analysed. For this, we examined whether the transcription factor STAT-1 is phosphorylated in IFNg-expressing neoplastic germ cells because it is known that a sufficient stimulation of IFNgR results in a STAT-1 activation through phosphorylation. Immunocytochemistry on tumour cells isolated from primary TGCT and Western blot analysis of NCCIT and NTERA cell lines revealed that STAT-1 is not constitutively phosphorylated in neoplastic germ cells ( Figure 5). To examine whether the IFNgR/STAT signaling pathway is intact in TGCT cells, we stimulated NCCIT and NTERA cell lines with different doses of rhIFNg. Western blot analysis demonstrated that upon stimulation STAT-1 was phosphorylated and IRF-1 was induced in both cell lines (Figure 5), suggesting that both IFNgR and STAT-1 are biologically intact.

DISCUSSION
In the present study, we investigated the effect of IFNg on proliferation and apoptosis of TGCT. Analyses showed IFNg in almost all neoplastic germ cells of primary TGCT. The data presented in this report extended our previous findings (Schweyer et al, 2002) because they demonstrated that primary TGCT not only expressed IFNg mRNA but also IFNg protein. These findings were surprising, because IFNg is normally expressed and secreted by inflammatory leucocytes (Elgert et al, 1998) but not by tumour cells. Moreover, IFNg is primarily known as a cytokine with several antitumour properties. For instance, it has been shown that IFNg possesses direct cytotoxic effects on ovarian carcinoma cell lines (Kim et al, 2002), augments apoptosis-inducing capacity of TNFa in cervical carcinoma cells (Suk et al, 2001), reduces the proliferation activity of colon carcinoma cells and melanoma cells (Raitano and Korc, 1993;Krasagakis et al, 1995), and is able to upregulate MHC molecules on renal cell carcinomas, thus leading to a better recognition of neoplastic cells by cytotoxic T cells (Totpal and Aggarwal, 1991;Hillman et al, 1997). In addition to these antitumour activities, however, IFNg seems to have also some powerful protumour effects. For instance, it is well known that IFNg is a potent inhibitor of apoptosis in some haematological malignancies (Lotem and Sachs, 1996;Mizuno et al, 1999).
Based on this background, we hypothesised that endogenous IFNg affects the proliferation and/or apoptosis of neoplastic germ cells. To prove this hypothesis, we first analysed two human TGCT cell lines for the expression and secretion of IFNg. Results demonstrated that both NCCIT and NTERA cell lines produce and release significant amounts of IFNg. Next, we neutralised the IFNg in culture supernatants of the TGCT cell lines by applying a specific antibody to study the role of secretory IFNg on the neoplastic germ cells. Using independent proliferation and apoptosis assays, we did not however note any evidence showing that the endogenous IFNg influences the multiplication and/or the NCCIT and NTERA cells were incubated with different concentrations of a neutralising antibody (Ab) against human IFNg (100, 500 or 1000 mg ml À1 ) or with control isotype-matched IgG (100, 500 or 1000 mg ml À1 ). After 24 h, the proliferation activity and the apoptosis rate of the cells were assessed by [ 3 H]thymidine assay and morphological methods, respectively. The [ 3 H]thymidine assay does not reveal any significant difference in the proliferation activity of NCCIT or NTERA cells following IFNg neutralisation (white columns) when compared with controls (black columns).
Results are expressed as counts per minute (c.p.m.)7s.e.m. (A). To prove the effect of IFNg neutralisation on apoptosis of NCCIT and NTERA cells, they were stained with MGG for the detection of nuclear condensation and fragmentation, with immunocytochemistry for the active form of caspase-3 and with ISEL for DNA fragmentation. Representative photomicrographs of active caspase-3 (black cytoplasmic signals) and of DNA fragmentation (black nuclear signals) illustrate no significant difference in apoptosis rate of cells incubated with the anti-IFNg Ab (1000 mg ml À1 ) or incubated with control IgG (1000 mg ml À1 ) (B). The diagrams show quantification of tumour cell apoptosis with MGG staining following application of anti-IFNg Ab (white columns) or control IgG (black columns) (B). Each experiment was performed in triplicate and was repeated three times. The values given (*P40.05, Student's t-test) are for the statistical significance of the difference between the two groups. Immunostaining shows the expression of IFNgRa (E and G, brown signals) and IFNgRb proteins (F and H, brown signals) in the same tumour (E and F) and in the same normal testis (G and H). Note that neoplastic but not normal germ cells in tubuli seminiferi (asterisks) express both IFNgRa and IFNgRb. Also note that, in the normal testis, Leydig cells (arrow) express IFNgRb, but not IFNgRa.
death rate of TGCT cells. These findings, however, must be strengthened by in vitro experiments with neoplastic cells isolated from primary tumours. Taking into account that the TGCT-derived IFNg is biologically active, because it induces the c-x-c chemokine IP-10 in cultured endothelial cells, as shown in our previous report (Schweyer et al, 2002), and considering the fact that IFNg mediates its effects through a high-affinity receptor consisting of a ligand-binding polypeptide chain a and a signal-transducing chain b (Pestka et al, 1997), and many tumours (e.g. hepatocellular carcinoma, prostatic carcinoma, basal cell carcinoma) do not express both receptor chains, thus providing tumour resistance to IFNg (Kooy et al, 1998;Nagao et al, 2000;Royuela et al, 2000), we proved whether the TGCT unresponsiveness to endogenous IFNg is due to the absence of IFNgR. Applying nonradioactive ISH and immunohistochemistry, however, IFNgR mRNA and protein for both a and b chains were detected in primary TGCT. Immunoblots and flow cytometry revealed that not only primary tumours but also TGCT cell lines express both receptor chains on their cell surface. Based on the fact that TGCT express IFNgR and stimulation of the IFNgR results in activation of downstream transcription factor STAT-1, we studied the level of STAT-1 phosphorylation in neoplastic germ cells. Analysis of primary tumours and cell lines indicated that the transcription factor STAT-1 is not constitutively phosphorylated/ activated in TGCT. For this phenomenon, we considered three possibilities. Firstly, the concentration of the endogenous IFNg may be too low to stimulate the IFNgR; secondly, IFNgR is functionally inactive, as demonstrated in renal cell carcinomas (Dovhey and Ghosh, 2000); and, finally, STAT-1 lacks TGCT, as shown in pul-monary carcinoma and malignant melanoma cells (Kaplan et al, 1998;Lee et al, 1999). Taking these possibilities into account, we stimulated the cell lines with rhIFNg and examined whether STAT-1 was phosphorylated and, if yes, whether pSTAT-1 acts as a functional transcription factor eliciting the expression of  Production of phosphorylated STAT-1 (pSTAT-1) protein was studied in primary TGCT as well as in TGCT cell lines. Immunocytochemistry (A and B) and Western blot (C) demonstrate the lack of pSTAT-1 in tumour cells from a representative TGCT with seminomatous differentiation (A) and in NCCIT and NTERA cell lines (C). At 24 h after application of rhIFNg (10, 50 or 100 U ml À1 ), however, STAT-1 was phosphorylated (D) and IRF-1 was induced in both TGCT cell lines (E). IFNg-stimulated monocytes (B, red signals), commercially purchased lysates of stimulated HeLa cells (C) or HUVECs costimulated with rhIFNg and rhTNFa (E) served as positive controls for immunocytochemistry and Western blots, respectively.
interferon-regulated proteins such as IRF-1. Results demonstrated that upon application of rhIFNg in excess (on an average five times the concentration of endogenous IFNg measured in supernatants of the TGCT cell lines), STAT-1 was phosphorylated and IRF-1 was induced. Thus, we suggest that IFNgR and STAT-1 are biologically intact in TGCT, but the level of the endogenous IFNg is not able to activate the IFNgR/STAT signalling pathway in an autocrine and/ or paracrine manner. Despite the lack of direct influence on neoplastic germ cells, an outstanding question may be whether endogenous IFNg alters the stromal microenvironment in TGCT, including enhancement of angiogenesis, modification of extracellular matrix composition, recruitment of inflammatory cells and dysbalance of protease activity and thereby the tumour development and progression.