Immunohistochemical staining of germ cell tumors and intratubular malignant germ cells of the testis using antibody to placental alkaline phosphatase and a monoclonal anti-seminoma antibody.

Antibody to placental alkaline phosphatase (PLAP) has been previously used in immunoperoxidase (IMP) staining studies of germ cell tumors and intratubular malignant germ cells (ITMGC) of the testis, the latter believed to be the precursor of these tumors. In this study, we compared staining by IMP using monoclonal antibody (mAb) and polyclonal antibody to PLAP with that seen using a mAb, M2A, which was previously shown to react with testicular seminomas and ITMGC. Antibody to PLAP and M2A reacted with different cellular components, as assessed by IMP staining of placenta and prepubertal testis and by Western blotting of seminoma lysates. Antibody to PLAP stained pure seminomas (seven of seven), pure embryonal carcinomas (four of four), and the seminoma (three of three) and embryonal carcinoma (six of six) components of mixed testicular germ cell tumors. M2A stained pure seminomas (26 of 26) and the seminoma component (three of three) of the mixed tumors, but failed to stain pure embryonal carcinomas (zero of four) or the embryonal carcinoma component (zero of five) of the mixed tumors. Both antibody to PLAP and M2A stained ITMGC of the testis. Since M2A stained seminomas and ITMGC but not embryonal carcinomas, seminomas would appear to be more closely related to ITMGC than embryonal carcinomas. This result has led us to speculate on the histogenesis of testicular germ cell tumors.     

MAGE-1 and MAGE-3 tumor rejection antigens in human germ cell tumors.

The MAGE-1 and MAGE-3 genes are members of the melanoma antigen-encoding gene family. These genes encode for HLA-restricted tumor-associated rejection antigens recognized by cytotoxic T lymphocytes. MAGE-1 and MAGE-3 gene expression has been identified in a number of human malignancies, and MAGE-1 and MAGE-3 antigenic peptides are potential targets for tumor-specific immunotherapy. The only normal tissues known to express these genes are testicular germ cells and placental tissue. The objective of this study was to examine MAGE-1 and MAGE-3 antigens immunohistochemically in testicular germ cell tumors, including seminoma, a germ cell tumor frequently associated with a lymphoid infiltrate. Forty-three germ cell tumors (24 seminomas, six embryonal carcinomas and 13 mixed germ cell tumors), and 10 Leydig cell tumors were selected for study, and standard immunohistochemical techniques were used on formalin-fixed paraffin-embedded tissues using mouse monoclonal antibodies to MAGE-1 (clone M454) and MAGE-3 (clone 57B) antigens. MAGE-1 antigen was identified in 16.6% of seminomas. No embryonal carcinomas, yolk sac tumors, or teratomas contained MAGE-1 protein. MAGE-3 antigen was identified in 41.8% of seminomas, and this protein was not identified in embryonal carcinomas, yolk sac tumors, or teratoma. Spermatogonia and primary spermatocytes contained MAGE-1 and MAGE-3 antigen, and more mature forms, including spermatids, were weakly positive to negative. Leydig cell tumors were negative for MAGE-1 and MAGE-3. In seminoma, the presence of MAGE-1 and MAGE-3 antigens did not correlate with tumor size, tumor stage, the presence of a lymphoid infiltrate, or patient outcome. The low frequency of MAGE-specific HLA alleles in the population, the loss of the HLA class I antigens in neoplastic germ cells, and the finding that the majority of seminomas and all non-seminomatous germ cell tumors lacked MAGE-1 and MAGE-3 antigenic peptides indicate that immunotherapy directed towards MAGE-1 and MAGE-3 antigen is not a likely treatment option for seminoma and nonseminomatous germ cell tumors. The significance of MAGE-1 and MAGE-3 proteins in normal spermatogonia and primary spermatocytes will require additional study.     

D2-40 immunohistochemistry in the differential diagnosis of seminoma and embryonal carcinoma: a comparative immunohistochemical study with KIT (CD117) and CD30.

The distinction between seminoma and embryonal carcinoma based on morphology alone can sometimes be problematic, requiring the use of immunohistochemistry to facilitate diagnosis. D2-40 is a monoclonal antibody that reacts with an oncofetal antigen expressed by fetal germ cells and testicular germ cell tumors. The diagnostic value of D2-40 immunohistochemistry in distinguishing seminoma from embryonal carcinoma has not been determined. D2-40 immunoreactivity was evaluated in a series of testicular germ cell tumors and compared with that of KIT (CD117) and CD30, to assess the relative utility of this marker in discriminating between seminoma and embryonal carcinoma. Forty testicular germ cell neoplasms were examined, which included 19 seminomas, three embryonal carcinomas, three teratomas, one yolk sac tumor, and 14 mixed germ cell tumors. The 14 cases of mixed germ cell tumors contained components of seminoma (n=7), embryonal carcinoma (n=11), teratoma (n=10), yolk sac tumor (n=2), and choriocarcinoma (n=1). All cases of pure seminoma and the seminomatous components of mixed germ cell tumors exhibited positive immunoreactivity for D2-40. Focal positivity for D2-40 was also observed in 29% of the embryonal carcinoma samples. D2-40 immunoreactivity in seminomas was characterized by diffuse membrane staining, whereas for embryonal carcinomas, staining was focal and distributed along the apical surfaces of the neoplastic cells. Immunohistochemical staining for KIT was observed in 92% of the seminoma samples and in none of the embryonal carcinomas. Conversely, CD30 expression was identified in 93% of the embryonal carcinoma samples and in none of the seminomas. Other germ cell components showed no immunoreactivity for D2-40, KIT, or CD30. KIT and CD30 are effective immunohistochemical markers in separating seminoma from embryonal carcinoma. Although a highly sensitive marker for seminomas, D2-40 positivity was also observed in a subset of embryonal carcinomas, thus limiting the utility of this antibody for discriminating between these two malignancies.     

Two monoclonal antibodies recognizing determinants on human embryonal carcinoma cells react specifically with the liver isozyme of human alkaline phosphatase

From a series of hybridomas that produced monoclonal antibodies reactive with the surface of human embryonal carcinoma cells, two that specifically recognized determinants of the liver/bone/kidney isozyme of alkaline phosphatase were isolated. They did not cross-react with the intestinal or placental isozymes. Phylogenetic studies revealed that both antibodies cross-reacted strongly with liver alkaline phosphatase from higher primates, but exhibited marked differences in their respective cross-reactions with liver alkaline phosphatase from other mammalian species.     

Immunohistochemical markers of carcinoma in situ of the testis also expressed in normal infantile germ cells

Carcinoma in situ of the testis is an intratubular, pre-invasive lesion preceding germ cell tumour. In adult men, carcinoma in situ cells differ in several aspects from normal germ cells. For example, placental-like alkaline phosphatase and/or the epitopes for the monoclonal antibodies M2A, 43-9F and TRA-1–60 are not seen in normal germ cells, whereas their presence is considered a specific sign of carcinoma in situ . As it is known that placental-like alkaline phosphatase and the epitope for TRA-1–60 are expressed in normal fetal germ cells it is possible that the markers could appear in normal infantile germ cells in a period after birth before they lose their expression. In children, carcinoma in situ cells may be difficult to identify morphologically and the use of the markers could be of great value. However, little information is available on the expression of the markers of adult carcinoma in situ in normal infantile germ cells. We investigated gonads from 66 boys less than 15 years old who died suddenly. Their deaths were unrelated to testicular disease. Immunohistochemical staining with anti-placental-like alkaline phosphatase antibody and monoclonal antibodies TRA-1–60 and 43–9F were performed. We found that these markers were expressed in some normal infantile germ cells until the age of 1 year. Therefore, these markers are not suitable for diagnosis of carcinoma in situ during the early postnatal period of life. Furthermore, our findings of placental-like alkaline phosphatase and the epitope for TRA-1–60 in normal germ cells before birth indicate that the markers of adult carcinoma in situ cells are of embryonic origin. This is in accordance with the hypothesis that carcinoma in situ cells are fetal germ cells malignantly transformed during fetal life, although re-expression of the antigens could provide an alternative explanation.     

Monoclonal antibody 43-9F: an immunohistochemical marker of embryonal carcinoma of the testis.

Discrimination between different types of germ cell tumours may be difficult in routine histological preparations. Additionally, none of the established immunohistochemical markers is completely reliable in diagnosis of embryonal carcinoma. A pilot study indicated that monoclonal antibody 43-9F--a marker of carcinoma-in-situ germ cells--may also react with embryonal carcinoma of the testis. In order to elucidate the applicability of 43-9F in diagnosis of embryonal carcinoma, 42 consecutive testicular germ cell tumours were tested immunohistochemically. Among the 42 tumours, 23 were seminomas and 19 were non-seminomas with seminomatous components in seven of them. Embryonal carcinomas were found in 15 tumours, two being of pure type and the remaining 13 a part of mixed tumours. Additionally, the material included 11 teratomas, nine yolk sac tumours and one choriocarcinoma. Immunohistochemical stainings were performed with 43-9F and additionally with antibodies against placental-like alkaline phosphatase, cytokeratins, alpha-foetoprotein and human chorionic gonadotropin. Using 43-9F a strong colour reaction was found in 13 of the embryonal carcinomas, whereas the reaction was moderate in the remaining two cases. A weak positive reaction was found in six seminomas and the remaining 24 did not react at all. 43-9F exhibited a positive reaction in four of 11 teratomas. The reactivity was generally weak with some focal areas with strong staining. In five cases the yolk sac tumour elements did not stain with this monoclonal antibody. The reaction was weak in three cases and in only one case was the staining intensity scored as moderate. Finally, no reaction was found in the choriocarcinoma element. Compared to the other antibodies tested, including the antibody against cytokeratins, in embryonal carcinoma immunohistochemical staining with 43-9F was more specific, stronger and more constantly expressed. The monoclonal antibody 43-9F may be of value in histological diagnosis of germ cell tumours. Additionally, the study confirmed the pathogenetical link between pre-invasive carcinoma in situ and embryonal carcinoma.     

Placental alkaline phosphatase immunohistochemistry of intratubular malignant germ cells and associated testicular germ cell tumors

Two hundred three testicular germ cell tumors were studied immunohistochemically for the presence of placental alkaline phosphatase (PLAP). Special emphasis was placed on the pattern and incidence of positive staining of intratubular malignant germ cells (ITMGCs) adjacent to tumors. 99% of cases with adjacent ITMGCs showed a positive staining reaction in some or all IT-MGCs present. Other germ cell elements showed at least a focal positive staining reaction in the following proportions: seminomas, 96%; embryonal carcinomas, 96%; yolk sac tumors, 25%; mature teratomas, 5%; immature teratomas, 4%; choriocarcinomas, 45%; and syncytiotrophoblasts, 43%. The staining pattern for seminomas tended to be diffuse, whereas for embryonal carcinomas the staining pattern was more focal. Yolk sac tumors stained inconsistently for PLAP and a positive reaction was limited to a small percentage of cells. Syncytiotrophoblasts, singly or in choriocarcinomas, also showed variable positivity. These results corroborate the fact that PLAP is a sensitive marker for ITMGC, seminoma, and embryonal carcinoma.     

Identity of the neoplastic alkaline phosphatase as revealed with monoclonal antibodies to the placental form of the enzyme

A panel of six monoclonal antibodies and a conventional polyclonal antibody raised against human placental alkaline phosphatase (ALP) were used to characterize the alkaline phosphatase detected by means of histochemistry on tumors of breast, ovary, lung, gastrointestinal tract, and kidney. Complete antigenic identity between the tumor ALP and the placental ALP was found only in one lung tumor. However, ten tumors reacted with the polyclonal antibody and some monoclonal antibodies, thus exhibiting partial identity with the placental ALP.     

Evaluation of cyclin expression in testicular germ cell tumors: cyclin E correlates with tumor type, advanced clinical stage, and pulmonary metastasis.

The measurement of proliferative index has yielded promising yet conflicting results in the evaluation of testicular tumors. We have examined the role of Ki-67, along with the cyclins A and E in testicular tumorigenesis. We compared the immunoreactivity of 20 pure seminomas with 20 mixed germ cell tumors composed predominantly of embryonal carcinoma with a variety of proliferation markers, including Ki-67, cyclin A, and cyclin E. All 40 tumors stained for Ki-67, and 19 of 20 (95%) seminomas and 18 of 20 (90%) embryonal carcinomas stained positively for cyclin A. Cyclin E stained 14 of 19 (74%) of the embryonal carcinomas and only 4 of 20 (20%) of the seminomas (Fisher's exact two-tailed test, P = .0012). There was a trend toward larger tumor size for cyclin E-positive seminomas (median, 5.92 cm versus 3.96 cm; P = .08), although the same correlation was not significant in embryonal carcinomas. For both seminomas and embryonal carcinomas, staining with cyclin E did not correlate with the presence of lymphovascular invasion or capsular invasion. However, patients who had cyclin E-positive tumors presented with higher clinical stage (P = .0015). In addition, pulmonary spread in embryonal carcinomas (four patients) and seminomas (one patient) occurred only in patients whose tumors were cyclin E positive (P = .014). Although Ki-67 and cyclin A offer little prognostic information in testicular germ cell tumors, cyclin E immunoreactivity correlates with tumor type and is strongly predictive of distant tumor spread.     

Monoclonal antibodies to human embryonal carcinoma cells: antigenic relationships of germ cell tumors

Fifteen monoclonal antibodies (mAb) that show specificity for human embryonal carcinoma cells are described. C57BL/6 mice were immunized with Tera-2 embryonal carcinoma cells, and hybridomas were isolated and tested versus a set of human developmental tumor cell lines. The antigens exhibit relatively restricted and unique distributions on normal tissues as shown by immunohistochemistry. The mAbs recognize a series of differentiation antigens since their expression changes when teratocarcinoma stem cells are induced to differentiate in vitro. The expression of these molecules in germ cell and related tumors is consistent with the data obtained from in vitro cell studies. Seven of these mAbs immunoprecipitate molecules from surface labeled Tera-2 cells that show distinct molecular weights on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The antigenic phenotypes defined by the set of mAbs can be used to investigate possible pathogenetic relationships of the various testicular tumors.     

Immunohistochemical localization of murine stage-specific embryonic antigens in human testicular germ cell tumors.

Monoclonal antibodies raised against and/or recognizing stage-specific antigens on preimplantation mouse embryos and stem cells of murine teratocarcinoma were used to localize these antigens immunohistochemically on human testicular germ cell tumors. SSEA-1, the antigen found on mouse embryonal carcinoma (EC) cells and embryonic cells from the 8-cell stage embryo onward, including the fetal primordial germ cells, was detected on yolk sac carcinoma components of human tumors, but not on EC cells. SSEA-3, the antigen found on follicular ova, fertilized eggs, early cleavage stage embryonic cells, and visceral endodermal cells of the mouse embryo, but not on mouse EC cells, was detected on human EC cells. Both antigens were found on the cell surface of fetal testicular germ cells but not in the seminiferous tubules of adult human testes. These data point out differences between human and murine EC cells suggesting that human EC cells correspond developmentally to a less mature embryonic cell than the murine EC cells. The possible histogenesis of human germ cell tumors from primordial and/or fetal germ cells is briefly discussed.     

Pineal Germinomas and Testicular Seminoma: A Comparative Ultrastructural Study with Special References to Early Carcinomatous Transformation

We have investigated the ultrastructural characteristics of 16 cases of pineal germinomas and compared them with those of 18 cases of testicular seminomas. Glandular differentiation of tumor cells was found in both though it was more consistently noted in pineal germinomas than in testicular seminomas. This feature was interpreted to represent early carcinomatous transformation of germinoma cells. It not only explains the difficulties occasionally encountered in distinguishing germinoma and its anaplastic variant from embryonal carcinoma, but also has implications for our understanding of germ cell neoplasia, particularly the place of germinoma/seminoma in the nosology of such tumors.     

Lectin histochemistry of embryonal carcinoma

Histological tissue sections of human testicular embryonal carcinoma from 13 patients and of a xenograft tumour in nude mice, as well as cell lines of human embryonal carcinoma, were investigated with eight different lectins to characterize the distribution of glycoconjugates in embryonal carcinoma. In all cases the malignant cells showed binding with Con A, WGA and RCA I conjugates, whereas other lectins were bound to some, but never to all, tumour cells in each group, revealing the heterogeneity of the malignant cells. A polarization of cancer cells was shown particularly with WGA and RCA I labelling, which was most intense on the luminal borders of the carcinoma cells, where pseudotubular structures were formed. The sugar staining properties were retained in cell culture and in the xenograft tumour. Regardless of the germ cell origin, embryonal carcinoma cells differed from normal germ cells. The distribution of glycoconjugates was also different from that of testicular carcinoma-in-situ germ cells, which share morphological features and the pattern of glycosylation with seminoma cells. However, the similarities in lectin binding pattern of seminomas and embryonal carcinomas suggest the close relationship between the two types of testicular malignancy, without excluding the possibility that embryonal carcinomas were derived from seminomas. Although lectins seem to be less important for differential diagnostic use in testicular cancer, our findings showed the usefulness of lectin histochemistry for characterization of embryonal carcinoma.     

Intermediate filament proteins in human testis and testicular germ-cell tumors

Normal testicular tissue and 76 testicular germ-cell tumors of various types were immunohistochemically evaluated for the expression of intermediate filament proteins of different types. In normal testes, the rete testis epithelium was positive to cytokeratin, and the Sertoli cells, stromal cells, and Leydig cells were positive for vimentin. Cytokeratin-positive cells were also found lining atrophic seminiferous tubules and were occasionally seen within nonatrophic seminiferous tubules. The classical seminomas showed vimentin positivity, but this was usually observed in a small number of tumor cells. In addition, nearly half the seminomas contained single cytokeratin-positive cells, some of which were multinucleated and appeared to represent syncytiotrophoblastic giant cells. The tumor cells in embryonal carcinomas, endodermal sinus tumors, and choriocarcinomas displayed cytokeratin positivity. In some embryonal carcinomas vimentin-positive tumor cells were also found, probably representing attempts at further differentiation of the tumor cells. In immature teratomas, both the immature and the mature epithelial structures showed cytokeratin positivity. The stromal components, including cartilage, contained vimentin, and the smooth-muscle elements, desmin. Neural tissue positive for neurofilaments and glial tissue positive for glial fibrillary acidic protein, were observed in 5 and 3 of 15 cases, respectively. It is considered that antibodies to intermediate filaments are suitable tools to characterize the differentiation patterns of testicular germ-cell tumors and have the potential to aid in the differential diagnosis especially between seminoma and embryonal carcinoma.     

NUCLEAR LAMIN EXPRESSION IN NORMAL TESTIS AND TESTICULAR GERM CELL TUMOURS OF ADOLESCENTS AND ADULTS

Nuclear A- and B-type lamins are differentially expressed in tissues, depending on the degree of cellular differentiation and proliferative status. By studying lamin expression in testis parenchyma and testicular germ cell tumours, further insight may be gained into the degree of cellular differentiation in normal testis and into the whole spectrum of differentiation lineages found in testicular germ cell tumours. Frozen tissue sections of normal testis and the different types of testicular germ cell tumours were immunostained with monoclonal antibodies to distinct lamin subtypes. Lamin reactivity was evaluated in relation to the lineage and degree of cellular differentiation and the reactivity patterns were compared with each other and with those in normal testis. In normal testis, both A- and B-type lamins were expressed in Sertoli, Leydig, and peritubular cells, while in spermatogonia only B-type lamins were found and spermatocytes showed weak reactivity with the A-type lamin antibodies. Carcinoma in situ was most often positive for both of the B-type lamins and negative for the A-type lamins (lamins A and C). In testicular germ cell tumours, B-type lamins were always expressed, while A-type lamins were differentially expressed. Differentiated non-seminomas were positive for both of the A-type lamins, whereas embryonal carcinomas were positive for lamin C and negative for lamin A. Seminomas were negative for both of the A-type lamins, with the exception of seminomas containing a Ras mutation. Spermatogonia and seminoma cells, which follow a differentiation pathway along the spermatogenic lineage and show characteristics of germ cells, do not express A-type lamins. Non-seminomas, showing embryonal or extraembryonal differentiation, express A-type lamins to varying degrees, distinguishing embryonal carcinoma cells from other non-seminomatous components. This may aid in the evaluation of the percentage of embryonal carcinoma in non-seminomatous testicular germ cell tumours as a prognostic parameter. © 1997 John Wiley & Sons, Ltd.     

Immunohistochemical differences between intracranial germinomas and their gonadal equivalents. An immunoperoxidase study of germ cell tumors with epithelial membrane antigen, cytokeratin, and vimentin

Twenty-six intracranial germ cell tumours (11 germinomas, 10 teratomas, 2 endodermal sinus tumours, 1 teratocarcinoma, and 2 undifferentiated embryonal carcinomas) and 26 gonadal germ cell tumours (13 testicular seminomas, 2 ovarian dysgerminomas, 9 ovarian teratomas, and 2 myometrial choriocarcinomas) were studied by immunoperoxidase with monoclonal antibodies (MAbs) against epithelial membrane antigen (EMA), cytokeratin, and vimentin. Typical tumour cells in three of the 11 germinomas (two of the latter being situated in the posterior fossa) expressed both EMA and cytokeratin, whereas those in the seminomas and dysgerminomas did not. In one seminoma, a few multinucleated giant cells expressed cytokeratin. In three of seven germinomas, vimentin-positive tumour cells were found, but all seminomas and dysgerminomas were negative. In the other forms of intracranial and gonadal germ cell tumours, epithelial and mesenchymal elements displayed the expected patterns of immunoreactivity to the respective determinants. The immunoperoxidase differences between the intracranial germinomas and their gonadal equivalents indicate that, in the former, early epithelial or mesenchymal differentiation of the primordial germ cells may be present. The findings draw attention to the heterogeneous cellular composition of these otherwise morphologically homogeneous-appearing tumours and, especially in the posterior fossa, to their transitional links to the immature teratomas.     

Demonstration using monoclonal antibodies of inter-locus heteromeric isozymes of human alkaline phosphatase

Monoclonal antibodies were used to demonstrate hybrid forms of human alkaline phosphatase (ALP) composed of subunits from both placental and intestinal loci. Four main isozymes with alkaline phosphatase activity appear on polyacrylamide gels of the HeLa cell line, Hep 2/5 after electrophoresis. The mobilities of all 4 isozymes are retarded after incubating cellular extracts with a monoclonal antibody specific for placental ALP, while the mobilities of 2 isozymes are also affected by a monoclonal antibody which reacts specifically with intestinal ALP. These 2 isozymes, therefore represent interlocus heteromers (placental/intestinal ALP).