Molecular Correlates of Aggressive Behavior and Biological Progression in Testicular Sertoli Cell Tumors.

Sertoli cell tumor (SCT) is the second most common type of sex cord-stromal tumor in men, and ∼10% exhibit malignant behavior. Although CTNNB1 variants have been described in SCTs, only a limited number of metastatic cases have been analyzed, and the molecular alterations associated with aggressive behavior remain largely unexplored. This study evaluated a series of nonmetastasizing and metastasizing SCTs using next-generation DNA sequencing to further characterize their genomic landscape. Twenty-two tumors from 21 patients were analyzed. Cases were divided into metastasizing SCTs and nonmetastasizing SCTs. Nonmetastasizing tumors were considered to have aggressive histopathologic features if they exhibited ≥1 of the following: size >2.4 cm, necrosis, lymphovascular invasion, ≥3 mitoses per 10 high-power fields, severe nuclear atypia, or invasive growth. Six patients had metastasizing SCTs, and the remaining 15 patients had nonmetastasizing SCTs; 5 nonmetastasizing tumors had ≥1 aggressive histopathologic feature(s). Gain-of-function CTNNB1 or inactivating APC variants were highly recurrent in nonmetastasizing SCTs (combined frequency >90%), with arm-level/chromosome-level copy number variants, loss of 1p, and CTNNB1 loss of heterozygosity occurring exclusively in CTNNB1-mutant tumors with aggressive histopathologic features or size >1.5 cm. Nonmetastasizing SCTs were almost invariably driven by WNT pathway activation. In contrast, only 50% of metastasizing SCTs harbored gain-of-function CTNNB1 variants. The remaining 50% of metastasizing SCTs were CTNNB1-wild-type and harbored alterations in the TP53, MDM2, CDKN2A/CDKN2B, and TERT pathways. These findings suggest that ∼50% of aggressive SCTs represent progression of CTNNB1-mutant benign SCTs, whereas the remaining ones are CTNNB1-wild-type neoplasms that exhibit alterations in genes of the TP53, cell cycle regulation, and telomere maintenance pathways.

[Clinicopathologic and molecular characterizations of Sertoli cell tumor, not otherwise specified of the testis].

Objective: To investigate the histomorpholgic spectrum, immunophenotypic, and molecular genetic features of Sertoli cell tumor, not otherwise specified (SCT, NOS) of the testis. Methods: Seven cases of SCT, NOS of the testis were analyzed(4 from Peking University Third Hospital and 3 from Zhejiang Provincial People's Hospital) between 2008 and 2017. The histopathologic features were examined based on HE staining, and EnVision method was used for immunohistochemistry staining of calretinin, inhibin, ß-catenin, cyclinD1, CD10, CKpan, neuroendocrine markers, WT1, Melan A, vimentin, SALL4, GATA3, PAX8, and S-100 protein. Mutational analysis of exon 3 of the CTNNB1 gene by polymerase chain reaction (PCR)-amplified sequences and direct sequencing was performed. Results: Patients ages ranged from 22 to 65 years (mean 43 years). The clinical manifestation in all was a slowly enlarging, painless testicular mass.The maximum diameter of the tumor ranged from 1.5 cm to 3.0 cm (mean 2.1 cm). Sectioning usually disclosed a tan-gray to white mass with vague lobular cut-surface. Microscopically, the tumors were well circumscribed and non-encapsulated; the tumor cells were rearranged in multiple growth patterns from diffuse solid sheets to trabeculae and cords, ribbon and solid or hollow tubules setting in variable amount of acellular fibrous stroma. Two cases showed acellular collagenous stroma constituted >50% of the tumor confirming to the diagnosis of sclerosing SCT. One case demonstrated a prominent myxoid stromal change. The tumor cells typically had moderate amounts of pale to lightly eosinophilic cytoplasm, 2 tumors had variable cells with abundant lipid-rich cytoplasm, and 1 other tumor showed scattered aggregates of multinucleated tumor cells. The tumor cells were bland-appearing without any evidence of atypia, mitoses were noted in 2 tumors (both were 1/50 HPF), but necrosis was absent. Immunohistochemical staining results as follows: vimentin (diffuse, 7/7), CD10 (diffuse membrane, 7/7); diffuse ß-catenin nuclear and cytoplasm staining in 5 of 7 cases, and all the 5 cases showed diffuse cyclin D1 nuclear staining, ß-catenin membrane staining in 2 of 7 cases, CKpan (5/7, focal or diffuse), calretinin (focal, 5/6), inhibin (focal, 3/7), synaptophysin (focal, 2/6), CD56 (focal or diffuse, 4/5), WT1 (diffuse nuclear, 4/5), and S-100 protein (diffuse, 3/7), and chromogranin A, Melan A, PAX8, GATA3 and SALL4 all were negative. Molecular genetic studies of PCR and direct sequencing showed CTNNB1 mutations in 4 of 7 (4/7) cases, 4 of the four mutation-carrying cases showed diffuse ß-catenin nuclear and cytoplasm immunoreactivity and diffuse cyclin D1 nuclear immunoreactivity in the tumor cells. Conclusions: SCT, NOS of the testis typically shows significant heterogeneities in both morphology and immunohistochemistry, thus causing differential diagnostic confusions. Molecular analyses showed mutations of exon 3 of CTNNB1 in more than half of these tumors, and nuclear accumulation of ß-catenin and over expression of cyclin D1 can be useful for the differential diagnosis of SCT, NOS.

Suppression of Sertoli cell tumour development during the first wave of spermatogenesis in inhibin α-deficient mice.

A dynamic partnership between follicle-stimulating hormone (FSH) and activin is required for normal Sertoli cell development and fertility. Disruptions to this partnership trigger Sertoli cells to deviate from their normal developmental pathway, as observed in inhibin α-knockout (Inha-KO) mice, which feature Sertoli cell tumours in adulthood. Here, we identified the developmental windows by which adult Sertoli cell tumourigenesis is most FSH sensitive. FSH was suppressed for 7 days in Inha-KO mice and wild-type littermates during the 1st, 2nd or 4th week after birth and culled in the 5th week to assess the effect on adult Sertoli cell development. Tumour growth was profoundly reduced in adult Inha-KO mice in response to FSH suppression during Weeks 1 and 2, but not Week 4. Proliferative Sertoli cells were markedly reduced in adult Inha-KO mice following FSH suppression during Weeks 1, 2 or 4, resulting in levels similar to those in wild-type mice, with greatest effect observed at the 2 week time point. Apoptotic Sertoli cells increased in adult Inha-KO mice after FSH suppression during Week 4. In conclusion, acute FSH suppression during the 1st or 2nd week after birth in Inha-KO mice profoundly suppresses Sertoli cell tumour progression, probably by inhibiting proliferation in the adult, with early postnatal Sertoli cells being most sensitive to FSH action.

A CASE OF AN AROMATASE-PRODUCING SERTOLI CELL TUMOR PRESENTING WITH GYNECOMASTIA.

A 64-year-old man had complained of a left scrotal mass and gynecomastia since June 2012. A left testicular tumor was suspected and the patient was referred to our department in December 2013. He presented with bilateral gynecomastia and a painless left scrotal mass that was firm, smooth surfaced, and the size of large hen's egg. Levels of markers of testicular germ cell tumors were all within normal range. Endocrinological examination revealed a marked elevation in serum estradiol (E2) level. The patient underwent high inguinal orchiectomy in December 2013.The pathological diagnosis was a Sertoli cell tumor of the left testis. Immunohistochemistry revealed the expression of aromatase synthesis; we speculated that this E2 production by the tumor caused the gynecomastia.Serum E2 level normalized after the orchiectomy. Owing to the diagnosis of malignancy, retroperitoneal lymph node dissection was performed in January 2014. No lymph node metastasis was found in the specimen. The gynecomastia improved gradually, and the patient has been free of disease since the surgery.

Immunohistochemical Characterization of Spontaneous Sertoli Cell Clusters in the Seminiferous Tubules of C57BL/6J Mice.

Cell clusters were observed in the seminiferous tubules of C57BL/6J mice as a spontaneous lesion in a 2-week toxicity study, and they were demonstrated to be basically composed of Sertoli cells by immunohistochemistry for claudin-11 and GATA-4 (GATA-binding protein 4), which are both Sertoli cell markers. The clusters were composed of about 5 to 50 cells, which had eosinophilic and occasionally vacuolated cytoplasm with an unclear cell boundary. The cell clusters involved some sperm. No mitotic figures were observed and no immunoreactivity for proliferating cell nuclear antigen (PCNA) was detected in the clusters. In most cases, the cell clusters were observed in seminiferous tubules that also showed degenerative changes. In rare instances, cell aggregates immunohistochemically positive for claudin-11 were observed in the lumen of the epididymis, suggesting that some of the Sertoli cell clusters were sloughed off from the seminiferous epithelium into the epididymal ducts. To our knowledge, this is the first report of Sertoli cell clusters in any animal species except for transgenic or surgically altered animals.

Evaluation of anti-Müllerian hormone in a dog with a Sertoli cell tumour.

BACKGROUND: Testicular tumours are common in elderly male dogs, and Sertoli cell tumours (SCTs) are among the most common. An increase in blood estradiol concentration is often seen in canine SCTs, but such measurements do not necessarily correlate with the clinical signs. CASE REPORT: A 6-year-old male Pembroke Welsh corgi was referred for nonpruritic alopecia. Clinical examination revealed cryptorchidism of the right testicle, and blood tests showed an increased estradiol concentration. The cryptorchid testis was removed by laparotomy, and SCT was diagnosed histologically. An enzyme-linked immunosorbent assay kit designed to measure human anti-Müllerian hormone (AMH) revealed a very high preoperative serum AMH concentration, which decreased after surgery. The serum AMH concentrations of two intact healthy control male dogs were lower than that of the dog with the SCT before treatment but higher than thoseof two healthy castrated male dogs. CONCLUSIONS AND CLINICAL IMPORTANCE: Canine serum AMH concentrations, as measured by a human AMH enzyme-linked immunosorbent assay, may be useful as a marker for canine SCT.

Gonadoblastoma: an immunohistochemical study and comparison to Sertoli cell nodule with intratubular germ cell neoplasia, with pathogenetic implications.

AIMS: To investigate the immunohistochemical properties of the sex cord cells of gonadoblastoma and Sertoli cell nodule with intratubular germ cell neoplasia unclassified (IGCNU) as a means of objective distinction and to provide insight into the pathogenesis. METHODS AND RESULTS: Immunohistochemical stains for SOX9, FoxL2 and SF-1 were performed on 10 gonadoblastomas (all phenotypical females) and 14 Sertoli cell nodules with IGCNU in normal phenotypical males with coexisting germ cell tumours. The sex cord cells of gonadoblastomas showed strong, diffuse FoxL2 and SF-1 positivity and focal weak to moderate SOX9 reactivity, whereas those of Sertoli cell nodules with IGCNU were uniformly, strongly positive for SOX9 and SF-1, while negative for FoxL2. CONCLUSIONS: Coexpression of SOX9 and FoxL2 in the sex cord cells of gonadoblastomas provides evidence that these morphologically ambiguous sex cord cells are incompletely differentiated. The strong, diffuse SOX9 and SF-1 positivity and absence of FoxL2 reactivity in the Sertoli cell nodules with IGNCU support full Sertoli cell differentiation of the sex cord cells and distinguish them from gonadoblastomas. Deficient SOX9 expression in gonadoblastoma supports a current model of pathogenesis where immature germ cells, in the absence of well-formed Sertoli cells, retain a fetal phenotype and susceptibility to malignant transformation.

Sclerosing Sertoli cell tumor of the testis: a clinicopathologic study of 20 cases.

Sclerosing Sertoli cell tumor (SSCT) of the testis is rare, with only 22 previously reported cases. Most have been small, circumscribed masses, and none has had a malignant clinical course; however, follow-up is limited. We have examined 20 new SSCTs to better characterize their features and report long-term follow-up. At least focal tubule formation by sex cord cells in a dense, hypocellular fibrous stroma occupying at least 50% of the lesion was required. The patient age ranged from 23 to 52 years (mean, 37; median, 39). All SSCTs were unilateral with 11 left sided and 9 right sided. The average size was 1.7 cm (range, 0.5 to 6 cm). In most cases, the stroma represented 50% to 70% of the mass but was at least 80% in 2. The Sertoli cells formed cords, trabeculae, small nests, focal tubules (sometimes with a vague pseudovascular or retiform appearance), and rarely single cells. Most tumor cells had small, round, oval to polygonal nuclei with finely granular chromatin, small nucleoli, and modest amounts of pale, eosinophilic cytoplasm. No mitotic figures, significant atypia, or necrosis was seen. All tumors but 1 were circumscribed and lacked lymphovascular invasion. Follow-up in 15 patients (3 mo to 16 y; mean, 6.1 y) showed 9 alive and free of disease, 5 alive with unknown disease status, and 1 patient, who presented with bone metastases, dead of disease at 27 months. The only features in the malignant case that differed from all others in our study were lymphovascular invasion and lack of circumscription. Combining our cases with previously reported ones shows that SSCTs are unilateral, usually small (80% <2 cm) tumors that occur in a wide age range (18 to 80 y old; mean, 35 y) and lack necrosis. Only 1 of 31 with follow-up (mean, 4.4 y) metastasized; this tumor was 3.8 cm and had lymphovascular invasion and invasive growth. We conclude that SSCTs<2 cm with the typical features and lacking those associated with malignancy in Sertoli cell tumors, not otherwise specified, have a negligible risk of metastasis and are adequately managed by orchiectomy alone.

Immunohistochemical evaluation of the expression of anti-Müllerian hormone in mature, immature and neoplastic canine Sertoli cells.

In mammals, the earliest specific protein expressed by Sertoli cells (SCs) is the anti-Müllerian hormone (AMH), which induces the regression of Müllerian ducts and is produced by SCs until the functional maturation of the testes. The aim of this study was to evaluate the expression of AMH by canine SCs during testicular maturation and neoplastic transformation. Testes from two fetuses, 18 newborn puppies, five puppies aged 43-180 days and six adult dogs, and 24 canine Sertoli cell tumours (SCTs) were studied immunohistochemically for expression of AMH. Fifteen of the 24 SCTs were classified as typical, eight as lipid-rich and one was considered malignant based on evidence of lymph node metastasis. SCs from fetuses and neonatal puppies and puppies up to 45 days old expressed AMH, while SCs from older puppies and adults were negative. All SCTs expressed AMH, suggesting that AMH expression is a useful marker of immature and neoplastic canine SCs.

Identification of the estrogen receptor GPER in neoplastic and non-neoplastic human testes.

BACKGROUND: Estrogen signaling is mediated by estrogen receptor beta isoforms in normal and neoplastic human testes. Recently, a G-protein-coupled-receptor (GPER) has been suggested as being involved in rapid responses to estrogens in different normal and tumor cells. METHODS: This study investigated the GPER expression in paraffin-embedded samples from non neoplastic and neoplastic human testes (sex-cord stromal and germ cell tumors) by immunohistochemical and Western Blot analyses. RESULTS: In control testes, a positive GPER immunoreactivity was detected in Leydig and in Sertoli cells while all germ cells were immunonegative. Furthermore, neoplastic cells of the Sertoli cell tumor, Leydig cell tumor, seminoma and embryonal carcinoma samples were all immunopositive. The immunoblots of testis extracts confirmed the results. CONCLUSIONS: These findings suggest that GPER could mediate estrogen signaling in both normal and transformed somatic cells of human testis, but they reveal a differential expression of the novel estrogen receptor in non neoplastic and neoplastic germ cells.

Sertoliform cystadenoma: a case with overlapping features.

The 1st pediatric case of sertoliform cystadenoma with unique features is described herein. The patient is a 6-year-old boy who presented with gynecomastia and a left testicular cystic mass. Histopathologically the tumor was found to originate from the rete channels, filling and distending them with areas of mural Sertoli cell proliferations reminiscent of large cell Sertoli cell tumor (noncalcifying form) and showing widespread intratubular Sertoli cell proliferation islands in the vicinity. Histopathologic and immunohistochemical features are described in light of the relevant literature.

Macroscopic sertoli cell nodule: a study of 6 cases that presented as testicular masses.

Sertoli cell nodules are almost always incidental microscopic lesions found in both cryptorchid and normally descended testes. Sertoli cell nodules, when present as masses or ultrasonographic lesions, may create diagnostic confusion. Herein, we report 6 cases of macroscopic Sertoli cell nodules that were received in consultation. The referral diagnoses included Sertoli cell tumor (2 cases), sex cord tumor with annular tubules (1 case), and gonadoblastoma (1 case). The patients were 19 to 36 years old: 3 patients presented with palpable testicular masses and 3 with lesions that were worrisome for neoplasms in ultrasonographic examinations conducted for pain (2 cases) or infertility (1 case). All were phenotypically normal male patients who lacked endocrine symptoms. The Sertoli cell nodules ranged from 6 to 10 mm in diameter and on microscopic examination consisted of circumscribed proliferations of immature Sertoli cells, globules and trabeculae of basement membrane, and spermatogonia in varying proportions. In 2 cases the lesion was distinctly intratubular, consisting of closely packed tubules containing various components; in the other cases there was confluent growth of the tubules. Immunostains for α-inhibin highlighted the Sertoli cells (5 of 5 cases), with the germ cells appearing in negative relief. An antibody for testis-specific protein, Y-encoded (TSPY), stained the spermatogonia (2 of 2 cases), whereas OCT 3/4 was negative in all the cases (5 of 5 cases). We conclude that Sertoli cell nodules may present clinically as mass lesions, and that it is important to distinguish them from true neoplasms to avoid unnecessary procedures.

An immunohistochemical study of normal and neoplastic canine sertoli cells.

Immunohistochemical studies of human fetal Sertoli cells (SCs) have shown transient expression of cytokeratin (CK) and desmin (DES) that is replaced after birth by expression of vimentin (VIM) and inhibin-α (INH-α). Human Sertoli cell tumours (SCTs) are characterized by re-expression of CK and DES. The aim of the present study was to evaluate immunohistochemically the expression of VIM, INH-α, CK and DES in normal and neoplastic canine SCs. Normal testicular tissue from three adult dogs, one 6-month-old puppy and two neonatal pups was examined in addition to samples from 21 canine SCTs. VIM was not expressed by neonatal SCs, but was present in SCs from the puppy, the adult dogs and in all SCTs. Conversely, INH-α was expressed by neonatal SCs and most SCTs, but not by normal SCs of adult dogs and the puppy. DES and CK were expressed only by some SCTs. These results show that, contrary to findings in man, canine SCs do not express VIM at the time of birth. SCs from neonatal dogs do express INH-α, but such expression was lost in the puppy and the adult dogs. Canine SCs therefore differ from human SCs, as expression of INH-α characterizes immature SCs, whereas the expression of VIM characterizes mature SCs. Canine SCTs may express CK and DES, suggesting that the neoplastic cells undergo de-differentiation during transformation.

Large cell calcifying Sertoli cell tumor: a clinicopathologic study of 1 malignant and 3 benign tumors using histomorphology, immunohistochemistry, ultrastructure, comparative genomic hybridization, and polymerase chain reaction analysis of the PRKAR1A gene.

Four cases of large cell calcifying Sertoli cell tumor, 3 benign and 1 malignant, with no clinical signs of Carney complex or Peutz-Jeghers syndrome are reported with results of histologic, immunohistochemical, ultrastructural, and comparative genomic hybridization studies. Analysis of PRKAR1A gene was performed on 2 cases. The age range of the patients was 19 to 54 years. The patient with a malignant large cell calcifying Sertoli cell tumor died of disease 4 years after surgery. Patients with benign tumors have had an uneventful follow-up for 1 and 3 years. All tumors were well circumscribed, unencapsulated, and composed of solid sheets, irregular cords, tubular structures, and nests in a fibrous and/or myxoid stroma with cellular atypia in the malignant case. All tumors showed diffuse immunoreactivity for inhibin, vimentin, calretinin, and S100 protein. Focal positivity for cytokeratin (AE1/AE3) was noticed in 1 case. Tumors were negative for CAM 5.2, Mic-2, Melan-A laminin, placental alkaline phosphatase, and alpha-fetoprotein. The proliferation index was 5% and 10% for 2 of the benign tumors and 30% for the malignant tumor. Comparative genomic hybridization was performed in 2 cases. There was no evidence of any major chromosomal changes. In one case, no PRKAR1A gene mutation was found. In the other case, a heterozygous shift mutation c.65_84dup was found, despite the absence of other clinical signs of Carney complex or Peutz-Jeghers syndrome. Although the combination of large cell calcifying Sertoli cell tumor and PRKAR1A mutation fulfills the criteria for establishing a diagnosis of Carney complex, the clinical relevance of finding a PRKAR1A gene mutation in a patient without any clinical signs of Carney complex or Peutz-Jeghers syndrome remains to be established.

Immunohistochemical expression of the KIT protein (CD117) in normal and neoplastic canine testes.

The aim of the present study was to characterize the expression of the KIT protein (CD117) in normal and neoplastic canine testes. Archival samples of normal testis (n=5), interstitial cell tumours (ICTs; n=10), Sertoli cell tumours (SCTs; n=10) and seminomas (n=10) were selected. Seminomas were subclassified on the basis of expression of placental alkaline phosphatase (PLAP) as classical seminoma (SE; PLAP positive; n=5) or spermatocytic seminoma (SS; PLAP negative; n=5). In normal testes, KIT expression was observed in Leydig cells and in spermatogonia. All ICTs expressed KIT, but no SCT was positively labelled. Seven of 10 seminomas expressed KIT and these tumours were reclassified on this basis as SS (KIT negative) or SE (KIT positive). These findings are consistent with observations of SE in man where many of the neoplastic cells reach the stage of spermatogonia where PLAP expression is lost and that of KIT is maintained. It would therefore appear that immunolabelling for KIT expression is a more appropriate means of distinguishing between canine SE and SS.

Immunohistochemical expression of dog TERT in canine testicular tumours in relation to PCNA, ki67 and p53 expression.

The objective of the study was to determine the immunohistochemical expression of canine TERT in canine testicular tumours comparing two different antibodies for TERT, and to correlate them with well established markers specific to dividing cells such as PCNA and ki67, and with expression of the p53 tumour suppressor gene. The study included 36 cases of canine testicular tumours, which were categorized as 12 Sertoli Cell Tumours (SCT), 20 seminomas, 3 interstitial cell tumours and 1 mixed germ cell-sex cord stromal tumour (MT). Two antibodies for hTERT were examined; a highly specific TERT antibody, RCK-hTERT, was evaluated for the first time. Immunodetection of RCK-hTERT was observed in 31% of tumours examined (6/20 Seminomas, 4/12 SCT, 1/3 interstitial cell tumour and 0/1 mixed germ cell-sex cord stromal tumour), while the NCL-hTERT in 67% of them (15/20 Seminomas, 6/12 SCT, 3/3 interstitial cell tumour and 0/1 ΜΤ). PCNA immunoreactivity was detected in all cases. Regarding ki67, 3 SCT, 12 seminomas and all interstitial cell tumours showed clear immunoreaction. p53 immunoreactivity was detected in 6 SCT, 15 seminomas and all interstitial cell tumours. The immunohistochemical expression of both TERT antibodies are discussed and compared in order to clarify their potential usefulness in canine testicular malignancies in relation to the expression of well known cell cycle markers. Our results indicate that TERT and PCNA are useful proliferation markers but not helpful to evaluate prognosis. Instead of that ki67 and p53 could be used for predicting aggressiveness in this group of tumours.

Overactive beta-catenin signaling causes testicular sertoli cell tumor development in the mouse.

Overactive WNT/beta-catenin signaling has been found in many forms of cancer in human patients. Mouse models with mutations in different components of the WNT/beta-catenin signaling pathway have been generated to mimic tumorigenesis in humans. Mice with mutations that result in overactive WNT/beta-catenin signaling developed tumors in some tissues, such as digestive tract, skin, and ovary, but they failed to develop tumors in other tissues, such as mammary gland, liver, kidney, and primordial germ cells. To investigate whether overactive beta-catenin signaling is capable of inducing Sertoli cell tumorigenesis in testes, we generated Ctnnb1(tm1Mmt/+);Tg(AMH-cre)1Flor male mice that express a constitutively active form of beta-catenin specifically in Sertoli cells. No tumors were observed at 4 mo of age, but 70% of the mutant males developed Sertoli cell tumors at 8 mo of age. At 1 yr of age, more than 90% of the mutant males developed tumors. No instances of extratesticular spread of the tumors were found in the mutant mice. These studies show a causal link between overactive WNT/beta-catenin signaling and Sertoli cell tumor development and provide a novel mouse model for the study of Sertoli cell tumor biology.

Immunohistochemical evaluation of GATA-4 in canine testicular tumors.

GATA-4 is a transcription factor expressed in Sertoli cells and less commonly in Leydig (interstitial) cells but not germ cells in adult human beings, cattle, pigs, and mice. We examined GATA-4 in 76 formalin-fixed, paraffin-embedded canine testicular tumors, including 21 Sertoli cell tumors (SCT), 28 Leydig (interstitial) cell tumors (LCT), 24 seminomas (GCT), and 3 mixed germ cell sex cord-stromal tumors (MGSCT). Our hypothesis was that immunohistochemistry for GATA-4 could discriminate between germ cell and sex cord-stromal tumors of the canine testis. SCTs (21/21) had strong diffuse nuclear and weak and inconsistent cytoplasmic staining. LCTs (27/28) also had strong diffuse nuclear staining and much weaker and granular cytoplasmic staining. GCTs were negative for this marker. Sex cord-stromal cells of MGSCT were also positive. These results indicate that GATA-4 is mainly expressed in sex cord-stromal tumors and not in germ cell tumors of the canine testis.