- Email: [email protected]
Trials of new germ cell immunohistochemical stains in 93 extragonadal and metastatic germ cell tumors
Human Pathology (2008) 39, 275–281
www.elsevier.com/locate/humpath
Original contribution
Trials of new germ cell immunohistochemical stains in 93 extragonadal and metastatic germ cell tumors Kenneth A. Iczkowski MD a,b,⁎, Samantha L. Butler MD a , Jonathan H. Shanks MD c , Deloar Hossain MD d , Albrecht Schall MD d , Isabelle Meiers MD e , Ming Zhou MD f , Kathleen C. Torkko PhD g , Stacy J. Kim MD h , Gregory T. MacLennan MD h a
Department of Pathology, Immunology, and Laboratory Medicine, University of Florida, Gainesville, FL 32608, USA Department of Pathology and Laboratory Medicine, Malcom Randall Veterans Affairs Medical Center, Gainesville, FL 32610, USA c Christie Hospital, Manchester, M20 4BX UK d University of Manitoba Health Sciences Centre, Winnipeg, R3A1R9, Manitoba, Canada e Bostwick Laboratories, Glen Allen, VA 23060, USA f Cleveland Clinic, Cleveland, OH 44195, USA g Department of Pathology, University of Colorado, Aurora, CO 80045, USA h University Hospitals, Cleveland, OH 44106, USA b
Received 1 November 2006; revised 23 June 2007; accepted 3 July 2007
Keywords: Extragonadal; Germ cell tumor; Pineal; Mediastinum; c-kit; OCT3/4; D2-40; AP-2γ
Summary Organic cation transporter 3/4 (OCT3/4) is a transcription factor of embryonic stem cells; c-kit (CD117) is a tyrosine kinase receptor implicated in seminoma carcinogenesis. Their reactivity is well characterized in testicular, but not extragonadal and metastatic, germ cell tumors. A total of 93 germ cell tumors (41 seminoma, 22 embryonal carcinoma, 18 teratoma, and 12 yolk sac tumor) were obtained from the central nervous system (30), mediastinum (23), retroperitoneum/abdomen (31), and other locations (9). Immunohistochemical staining for c-kit, placental-like alkaline phosphatase (PLAP), OCT3/4, and new markers D2-40 and AP-2γ was performed on seminomas; CD30 and epithelial membrane antigen were added for nonseminomas. In embryonal carcinoma, c-kit reacted in 17 of 22 cases, OCT3/4 in 18 of 22, and PLAP in 13 of 22. OCT3/4 was superior to PLAP in intensity and percent cells staining. In seminoma, OCT3/4 and D2-40 were superior to PLAP in intensity and percent cells; c-kit and AP-2γ were superior in percent cells. D2-40 stained 23 of 24 seminomas strongly but had only weak focal reactivity in 6 of 17 embryonal carcinomas. Sensitivity and specificity were high for OCT3/4 discriminating seminoma and embryonal carcinoma, and c-kit discriminating seminoma, from other germ cell tumors. For embryonal carcinoma, OCT3/4 had higher specificity (0.94) than CD30 (0.786) owing to CD30 reactivity in 3 of 10 teratomas. Epithelial membrane antigen discriminated teratoma from other nonseminomas with a sensitivity of 1 but reacted occasionally in embryonal carcinoma (3/15) and yolk sac tumor (2/7). In conclusion, for extragonadal seminoma, OCT3/4, AP-2γ, D2-40, and c-kit were equivalently superior to PLAP. For embryonal carcinoma, OCT3/4 was superior to PLAP and more specific than CD30. D2-40 is recommended to discriminate between seminoma and embryonal carcinoma. © 2008 Elsevier Inc. All rights reserved.
⁎ Corresponding author. Pathology-P.O. Box 6511, UCHSC-Campus Mail Stop 8104, Aurora, CO 80045, USA. E-mail address: [email protected] (K. A. Iczkowski). 0046-8177/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.humpath.2007.07.002
276
1. Introduction Extragonadal germ cell tumors account for an estimated 2% to 5% of all germ cell tumors [1] and are usually encountered in the midline of 3 body sites: mediastinum, central nervous system (CNS), and retroperitoneum. Evidence suggests that such tumors in the mediastinum and in the CNS arise from primordial germ cells that failed to undergo apoptosis in fetal life [2]; germ cell tumors in the retroperitoneum are considered to represent metastases from a testicular primary site, whether or not a testicular tumor is demonstrable. The unique diagnostic dilemmas of tumors in these sites were reviewed [3] and result primarily from small sample size and frequent tumor necrosis. Accurate diagnosis is essential for treatment and prognosis. Seminomas, for instance, respond to radiotherapy and/or cisplatin-based chemotherapy, whereas nonseminomatous germ cell tumors do not respond to radiation therapy. New diagnostic immunohistochemical markers for germ cell tumors have been tested alongside established markers of metastatic seminoma [4] but not in a wide variety of extragonadal sites nor in a study that includes nonseminomatous germ cell tumors. Established germ cell immunohistochemical markers include placental-like alkaline phosphatase (PLAP), which reacts with all germ cell tumors; CD30 (Ki-1), which uniquely stains embryonal carcinoma; cytokeratin AE1/3, which reacts diffusely with embryonal carcinoma and yolk sac tumor and can give focal punctate perinuclear reactivity in seminoma; α-fetoprotein for yolk sac tumor; and epithelial membrane antigen (EMA) in teratoma and (with hCG) in choriocarcinoma. Four new markers for germ cell tumors have emerged recently. Organic cation transporter 3/4 (OCT3/4) is a nuclear transcription factor that regulates self-renewal and pluripotency of embryonic stem cells [5,6]. OCT3/4 protein is considered a specific and sensitive marker for intratubular germ cell neoplasia [7,8], seminoma and embryonal carcinoma [9], and ovarian dysgerminoma [10]. OCT3/4 was shown to give intense nuclear staining in 18 cases of seminoma and 29 cases of embryonal carcinoma from retroperitoneal metastases [11]. OCT3/4 is nonreactive in the other types of differentiated nonseminomatous tumors: yolk sac tumor, immature and mature teratoma, and choriocarcinoma [9]. Reactivity has been described in metastatic germ cell tumors [11] and CNS seminoma, commonly called germinoma [12]. c-kit (CD117), a 145-kD transmembrane glycoprotein, binds a stem cell factor that promotes dimerization and autophosphorylation of KIT receptors. KIT signaling is critical for normal development and survival of germ cells. Membranous c-kit gene expression is noted as a constant feature in intratubular germ cell neoplasia and primary seminoma [13], although its performance has not yet been compared with PLAP and conventional markers. The only c-kit expression reported in nonseminomatous tumors was in teratomas with sarcomatous histology, in a cytoplasmic pattern in 11 (48%) of 23 cases [14]; however, the status of c-kit in embryonal carcinoma was uncertain. D2-40, a membranous protein similar to podoplanin, has been proposed as a superior marker
K. A. Iczkowski et al. for seminoma, and AP-2γ, a transcription factor, for both seminoma and embryonal carcinoma [4]. We set out to study the reactivity of new markers in extragonadal germ cell tumors in primary mediastinal, CNS, and retroperitoneal and metastatic locations. We also aimed to compare the reactivity of new markers to established germ cell tumor immunohistochemical stains, including PLAP, EMA, and CD30.
2. Materials and methods With internal review board approval from each institution, surgical pathology files were searched for extragonadal and metastatic germ cell tumors at Shands Hospital, Gainesville, FL; University of Arkansas for Medical Sciences in Little Rock; University Hospitals, Cleveland; Cleveland Clinic; PCA Southeast, Columbia, TN; University of Manitoba Health Sciences Centre in Winnipeg; Christie Hospital, Manchester, UK; and the Institut de Pathologie et de Genetique, Gosselies, Belgium. We obtained and rereviewed 93 germ cell tumors (41 seminoma/germinoma, 23 embryonal carcinoma, 18 mature teratoma—2 with immature component—and 12 yolk sac tumor) from cerebrum (11), pineal (18), suprasellar (1), mediastinum (23), retroperitoneum/abdomen (31), and other sites (9). The other sites included lung (5), pericardium, neck lymph nodes, liver, and thigh. All CNS and mediastinal tumors were primary. Eleven cases were metastatic: known primary testicular tumors had been diagnosed in 1 case of CNS and 1 case of mediastinal embryonal carcinoma and in 8 (26%) retroperitoneal tumors. Chemotherapy is known to alter immunohistochemical expression of CD30 [15]; however, no specimens from patients who underwent it were included. The immunohistochemical panel used for seminomas was OCT3/ 4, PLAP, and c-kit (CD117). For nonseminomatous germ cell tumors, the panel included OCT3/4, PLAP, c-kit, EMA, and CD30 (Ki-1). Immunostains for AP-2γ and D2-40 were performed on a subset of 23 seminoma and 26 to 30 nonseminoma tumors, limited by availability. Heat-induced epitope retrieval was used for all antigens except EMA. The dilution of OCT3/4 (C20, sc8629; Santa Cruz Biotechnology, Santa Cruz, CA) goat antibody was 1:40 for 32 minutes at 42°C. D2-40 antibody (Abcam, Cambridge, MA) was used prediluted for 1 hour, and AP-2γ (H-77, Santa Cruz) was used at a 1:50 dilution for 1 hour. Rabbit antigoat secondary antibody was diluted at 1:100 for 32 minutes. The other 4 antibodies were detected using room temperature incubations with Ventana i-View DAB kit reagents (Ventana Medical Systems, Tucson, AZ). c-kit (IgG; CellMarque, Hot Springs, AR) came in prediluted form, as did PLAP (Ventana). The dilutions were 1:2 for CD30 (Cell Marque) and 1:50 for EMA (DakoCytomation, Carpinteria, CA). Reactivity was scored in terms of intensity and percent reactivity of the tumor. For OCT3/4 and AP-2γ, nuclear reactivity was evaluated; for all other stains, cytoplasmic or
Trials of new germ cell immunohistochemical stains
277
membranous reactivity was scored. Intensity of reactivity was recorded as 0 (no staining), 1+ (weak intensity), 2+ (moderate intensity), or 3+ (strong intensity). The percentage of reactive tumor cells (0%-100%) was estimated based on entire viable tumor present. Differences in intensity were assessed using Fisher exact test, and differences in percent using general linear mixed model. Matched pairs of immunostains were compared using Wilcoxon signed rank test. The sensitivity and specificity of markers for certain tumor types were calculated.
3. Results 3.1. Seminoma and embryonal carcinoma The reactivity of seminoma and embryonal carcinoma with the immunostaining panels is detailed (Table 1). Table 1 tumors Tumor
Protein expression by 93 extragonadal germ cell Embryonal carcinoma
PLAP, cytoplasm Cases + 13/22 Intensity 1.2 % cells 27 c-kit Cases + 17/22 Intensity 1.4, cytoplasmic % cells 32 † (1-90) OCT3/4, nucleus Cases + 18/22 Intensity 1.9 % cells 64 D2-40, cytoplasm Cases + 6/17 Intensity 0.7 % cells 5 AP-2γ, nucleus Cases + 10/14 Intensity 1.5 % cells 56 CD30, cytoplasm Cases + 16/22 Intensity 2.2 % cells 65 EMA, cytoplasm Cases + 6/18 Intensity 2.0 % cells 48
P⁎
Seminoma
P⁎
32/40 2.1 57 ‡
ns
40/41 2.4, membranous 84 (20-100)
.0001
.006 b.0001
41/41 2.6 84
.003 b.0001
ns ns
23/24 2.67 91
.008 b.0001
ns .006
22/23 2.5 82
ns .001
ns
ns
nd
nd
Abbreviations: ns, not significant; nd, not done. ⁎ P values represent superiority to PLAP based on Wilcoxon signed rank test. † Percent cells staining in embryonal carcinoma is less than in seminoma, P b .0001. ‡ Percent of seminoma staining with PLAP is less than with all 4 other antibodies, P b .0001.
Fig. 1 Pineal germinoma in a 15-year-old boy. All figures are on the same focus. A, OCT3/4 shows strong nuclear reactivity. B, c-kit with reactivity predominantly in membranous location. C, PLAP with very weak reactivity (×400).
For the purpose of this analysis, germinomas of the CNS were included in the group termed seminoma. OCT3/4 discriminated seminoma and embryonal carcinoma from all other germ cell tumors (sensitivity, 0.88; specificity, 0.94).
278
K. A. Iczkowski et al.
Fig. 2 Mediastinal embryonal carcinoma in a 25-year-old man. All figures are on the same focus. A, OCT3/4 is strongly reactive. B, c-kit was reactive in 11 of 15 extragonadal embryonal carcinomas, but the pattern was focal, and cytoplasmic reactivity predominated over membranous reactivity, unlike seminoma (Fig. 1B). C, PLAP shows no reactivity. D, CD30 gives focal reactivity (×400).
Compared with PLAP, the percentage of seminoma cells was higher for OCT3/4 (P b .001), c-kit (P b .0001), AP-2γ (P = .0009), and D2-40 (P b .0001); intensity of reactivity was superior for OCT3/4 (P = .0027) and D2-40 (P = .0078) only. For embryonal carcinoma, the percent of cells staining was higher than PLAP for OCT3/4 (P b .0001) and AP-2γ (P = .006) and D2-40; intensity was again superior to PLAP only for OCT3/4 (P b .006). Sensitivity was 0.96 for seminoma with AP-2γ and D2-40; it was 0.97 for c-kit, but specificity was low (0.57) owing to reactivity with 17 of 22 embryonal carcinomas. In all seminomas, the c-kit reactivity was predominantly crisply membranous. Reactivity of a pineal germinoma with PLAP, OCT3/4, and D2-40 is shown (Fig. 1). In this case, there is nuclear reactivity for OCT3/4, strong membranous reactivity for c-kit, and weak to absent reactivity for PLAP. OCT3/4 and CD30 both discriminated embryonal carcinoma from other nonseminomas. They were reactive in 18 of 22 and 16 of 22 cases and not significantly different in intensity or percent cells. However, OCT3/4 had higher specificity (0.94) than CD30 (0.786) for embryonal carcinoma owing to CD30 reactivity in 8 of 18 teratomas—
Fig. 3 Seminoma in retroperitoneal lymph node. D2-40 reacted with germinal center cells in lymph nodes (left), although weakly enough so as not to risk confusion with metastatic seminoma (right) (×200).
Trials of new germ cell immunohistochemical stains although this reactivity for CD30 would not be a diagnostic problem. c-kit was, as mentioned above, reactive in many embryonal carcinomas; the intensity and percent of cells were comparable to PLAP. Reactivity of a mediastinal embryonal carcinoma with these 4 immunostains is shown (Fig. 2A-D). The percent of cells in embryonal carcinoma reactive with PLAP, c-kit, AP-2γ, or D2-40 averaged significantly less than in seminoma, and the intensity was nonsignificantly lower. A constant observation was that patterns of reactivity were different from the membranous predominance in seminoma. Among 22 embryonal carcinomas, reactivity was cytoplasmic in 10 and cytoplasmic plus membranous in 12. The cytoplasmic pattern of reactivity is shown (Fig. 2B). Notably, both D2-40 and AP-2γ had consistent, significant reactivity for other cell types. D2-40 reacted with germinal center cells in lymph nodes, although weakly enough so as not to risk confusion with metastatic seminoma (Fig. 3). AP-2γ strongly reacted with the angulated nuclei of adipocytes, whose shape minimizes likelihood of confusion with the rounded nuclei of seminoma cells (Fig. 4).
3.2. Other germ cell tumors EMA discriminated teratoma from other nonseminomas with a sensitivity of 1 and a specificity of 0.76. The EMA was reactive in all nonneural components of the mature teratomas. The low specificity resulted from occasional reactivity in embryonal carcinoma (3/15) and yolk sac tumor (2/7). c-kit was reactive in 3 of 7 yolk sac tumors, with 1+, 2+, and 3+ intensity in 10% to 90% of cells and a membranous plus cytoplasmic pattern. None of the 7 tumors was reactive for PLAP or CD30, but two had some reactivity for EMA and one for OCT3/4. CD30 reactivity was not confined to embryonal carcinoma but was reactive in some (4/10) mature teratomas. In
Fig. 4 Seminoma in retroperitoneal lymph node. AP-2γ strongly reacted with the angulated nuclei of adipocytes (left), which would be unlikely to be confused with the rounded nuclei of seminoma cells (right) (×200).
279 these cases, reactivity was confined to the respiratory component in 3 cases, the squamous component in 2 cases, and gastrointestinal epithelium, nerve, or cartilage in 1 case each. Immature teratoma was not reactive with any stain.
3.3. Reactivity according to tumor site For each tumor type, all markers were compared for their reactivity according to site: CNS, mediastinum, retroperitoneum, or other. The only difference among markers by tumor site was more intense OCT3/4 reactivity in retroperitoneal/abdominal seminoma (3.0+) than at other sites (2.2-2.4+) (P = .0068). Furthermore, no difference was evident between the primary and the metastatic tumors for any tumor type or marker.
4. Discussion The current study is novel in 4 ways: (1) it shows that OCT3/4, AP-2γ, and D2-40 perform well as markers for mediastinal seminoma, extending the variety of extragonadal germ cell tumor locations in which these have been tested; (2) for nonseminomatous tumors, it tests these new markers in comparison with established ones, whereas all studies but two dealt only with seminomas [2,11]; (3) it is the first to test c-kit reactivity in nonseminomatous germ cell tumors, disclosing reactivity in 17 of 22 embryonal carcinomas; (4) it adds CD30 and EMA to the panel for the nonseminomatous tumors to evaluate their comparative diagnostic potential. With seminoma of extragonadal or metastatic sites, OCT3/ 4 and c-kit were superior to PLAP in the number of positive cases, intensity of staining, and percent cells. This finding reinforces a prior study in which c-kit was superior to PLAP in 25 CNS germinomas, staining 100% versus 28% of cases [16]. Recent studies involving 25 cases of CNS germinoma [12] and 30 biopsy specimens of retroperitoneal seminoma [17] found superior reactivity of OCT3/4 compared with PLAP. Interestingly, in the latter study, OCT3/4 was also superior to c-kit: OCT3/4 marked 100% of cases, whereas ckit marked 21 (70%) of 30 of cases, and PLAP marked 23 (77%) [17]. In contrast, we found essentially no difference in c-kit versus OCT3/4 sensitivity for extragonadal seminomas from a wider range of locations. Biermann et al [4] recently compared OCT3/4 and c-kit to seminoma markers D2-40 and AP-2γ. A total of 30 extragonadal seminomas were included, 28 metastatic and 2 pineal, using this panel. D2-40 and AP-2γ yielded strong reactivity compared with weak reactivity for PLAP in 63% of cases, OCT3/4 in 50%, and c-kit in 40%. However, our study showed all 4 seminoma markers essentially equivalent and superior to PLAP as shown previously for OCT3/4 [9,11,12] and c-kit [16,17]. Technical variations, particularly with antigen retrieval, may explain this disparity.
280 Interestingly, the presence of c-kit expression in all but one of our cases of extragonadal seminoma does not necessarily imply that all seminomas are sensitive to imatinib (Gleevec) treatment. Treatment sensitivity depends on c-kit activating mutations in either exon 11 or 17. In a recent study, these mutations were present in only 5 (23%) of 22 testicular seminomas, although 19 (86%) had membranous c-kit immunoreactivity [18]. In extratesticular embryonal carcinoma, OCT3/4 was again strikingly superior to PLAP in number of cases, intensity, and percent cells. We found that the most (18/22) of the embryonal carcinomas were reactive for OCT3/4, with a slight edge over the other nuclear stain, AP-2γ, whereas only 13 of 22 cases reacted with PLAP. The frequent reactivity for c-kit (17/22) was an unexpected finding, making D2-40 a better discriminator since the number of cases (6/17), intensity, and percent of cells reactive were less than for c-kit. The reactivity of c-kit in extragonadal embryonal carcinoma had been tested in just one case in a prior study [2], where it was negative. If c-kit is used, the likelihood of confusion with seminoma is reduced not only by the lower percent of cells showing reactivity in most cases but also by different localization. Embryonal carcinoma reactivity was cytoplasmic, not membranous as in seminoma. In a number of other tumor types studied, a variety of c-kit reactivity patterns have been reported. Predominant cytoplasmic reactivity of c-kit has been noted, for example, in sarcomatoid teratoma [14] and solid-pseudopapillary neoplasm of the pancreas [19]. In gastrointestinal stromal tumor, c-kit reactivity may be membranous, cytoplasmic, or punctate perinuclear (Golgi) [20]. CD30 and EMA stained predominantly embryonal carcinoma and teratoma, respectively, but had mild specificity problems. In embryonal carcinoma, OCT3/4 showed nearly identical reactivity to CD30 in number of cases, intensity, and percent of cells. This stands in contrast to a study of 25 metastatic embryonal carcinomas after chemotherapy, in which 76% had moderate to strong OCT3/4 staining, but only 28% had weak focal CD30 expression [21]. In all of our cases, patients had not undergone chemotherapy, so our observation supports the theory that chemotherapy markedly reduces CD30 [15]. The respiratory or squamous elements of some teratomas were noted to react with CD30. OCT3/4 was more specific for embryonal carcinoma and was not reactive in teratoma, in which columnar glands could be confused with embryonal carcinoma. The presence of c-kit reactivity in a membranous/ cytoplasmic pattern in yolk sac tumor was also an unexpected finding. None of the 7 extragonadal yolk sac tumors was PLAP-reactive, and other markers tested were usually negative. In a prior study of mediastinal versus testicular germ cell tumors, no difference in PLAP reactivity was found among yolk sac tumors [22]. In conclusion, the newer immunostains OCT3/4 (reviewed recently [23]), c-kit, AP-2γ, and D2-40 offer improved sensitivity and specificity for detecting extrago-
K. A. Iczkowski et al. nadal seminoma and embryonal carcinoma in various sites compared with PLAP. Both OCT3/4 and c-kit or D2-40 are recommended for establishing a diagnosis of extragonadal seminoma. An immunostaining panel to help with the differential diagnosis between seminoma and embryonal carcinoma should include c-kit and CD30, with the caveat that some c-kit reactivity may be expected in embryonal carcinoma, but it is usually focal and has a cytoplasmic pattern. For embryonal carcinoma in which the presence of other nonseminomatous elements is suspected, α-fetoprotein and EMA may be useful in addition to CD30. The diagnosis of mature teratoma was supported by EMA reactivity in the gastrointestinal and respiratory components. For extragonadal yolk sac tumor, PLAP did not show reactivity, although caution is needed with c-kit because 3 of 7 tumors were reactive. Thus, α-fetoprotein reactivity is probably still the best marker for yolk sac tumor.
Acknowledgment We thank Elaine Dooley (Univ. of Florida) for technical assistance. Jesse K. McKenney, MD (Little Rock, AR), and Cory Schlesigner, MD (Columbia, TN) contributed cases.
References [1] Bosl GJ, Motzer RJ. Testicular germ-cell cancer. N Engl J Med 1997; 337:242-53. [2] Hoei-Hansen CE, Sehested A, Juhler M, et al. New evidence for the origin of intracranial germ cell tumours from primordial germ cells: expression of pluripotency and cell differentiation markers. J Pathol 2006;209:25-33. [3] McKenney JK, Heerema-McKenney A, Rouse RV. Extragonadal germ cell tumors: a review with emphasis on pathologic features, clinical prognostic variables, and differential diagnostic considerations. Adv Anat Pathol 2007;14:69-92. [4] Biermann K, Klingmüller D, Koch A, et al. Diagnostic value of markers M2A, OCT3/4, AP-2γ, PLAP and c-kit in the detection of extragonadal seminomas. Histopathology 2006;49:290-7. [5] Looijenga LH, Stoop H, de Leeuw HP, et al. POU5F1 (OCT3/4) identifies cells with pluripotent potential in human germ cell tumors. Cancer Res 2003;63:2244-50. [6] Honecker F, Oosterhuis JW, Mayer F, et al. New insights into the pathology and molecular biology of human germ cell tumors. World J Urol 2004;22:15-24. [7] Jones TD, Ulbright TM, Eble JN, et al. OCT4: a sensitive and specific biomarker for intratubular germ cell neoplasia of the testis. Clin Cancer Res 2004;10:8544-7. [8] de Jong J, Stoop H, Dohle GR, et al. Diagnostic value of OCT3/4 for pre-invasive and invasive testicular germ cell tumors. J Pathol 2005; 206:242-9. [9] Jones TD, Ulbright TM, Eble JN, et al. OCT4 staining in testicular tumors: a sensitive and specific marker for seminoma and embryonal carcinoma. Am J Surg Pathol 2004;28:935-40. [10] Cheng L, Thomas A, Roth LM, et al. OCT4: a novel biomarker for dysgerminoma of the ovary. Am J Surg Pathol 2004;28:1341-6. [11] Cheng L. Establishing a germ cell origin for metastatic tumors using OCT4 immunohistochemistry. Cancer 2004;101:2006-10.
Trials of new germ cell immunohistochemical stains [12] Hattab EM, Tu PH, Wilson JD, et al. OCT4 immunohistochemistry is superior to placental alkaline phosphatase (PLAP) in the diagnosis of central nervous system germinoma. Am J Surg Pathol 2005;29: 368-71. [13] Kemmer K, Corless CL, Fletcher JA, et al. KIT mutations are common in testicular seminomas. Am J Pathol 2004;164:305-13. [14] Madani A, Kemmer K, Sweeney C, et al. Expression of KIT and epidermal growth factor receptor in chemotherapy refractory nonseminomatous germ-cell tumors. Ann Oncol 2003;14:873-80. [15] Berney DM, Shamash J, Pieroni K, et al. Loss of CD30 expression in metastatic embryonal carcinoma: the effects of chemotherapy? Histopathology 2001;39:382-5. [16] Nakamura H, Takeshima H, Makino K, et al. C-kit expression in germinoma: an immunohistochemistry-based study. J Neuroendocrinol 2005;75:163-7. [17] Sung M-T, MacLennan GT, Cheng L. Retroperitoneal seminoma in limited biopsies: morphologic criteria and immunohistochemical findings in 30 cases. Am J Surg Pathol 2006;30:766-73.
281 [18] Willmore-Payne C, Holden JA, et al. Detection of c-kit exons 11- and 17-activating mutations in testicular seminomas by high-resolution melting amplicon analysis. Mod Pathol 2006;19:1164-9. [19] Cao D, Antonescu C, Wong G, et al. Positive immunohistochemical staining of KIT in solid-pseudopapillary neoplasms is not associated with KIT/PDGFRA mutations. Mod Pathol 2006;19:1157-63. [20] Miettinen M, Lasota J. Gastrointestinal stromal tumors: review on morphology, molecular pathology, prognosis, and differential diagnosis. Mod Pathol 2006;130:1466-78. [21] Sung M-T, Jones TD, Beck SD, et al. OCT4 is superior to CD30 in the diagnosis of metastatic embryonal carcinomas after chemotherapy. HUM PATHOL 2006;37:662-7. [22] Suster S, Moran CA, Dominguez-Malagon H, et al. Germ cell tumors of the mediastinum and testis: a comparative immunohistochemical study of 120 cases. HUM PATHOL 1998;29:737-42. [23] Cheng L, Sung M-T, Cossu-Rocca P, et al. OCT4: biological functions and clinical applications as a marker of germ cell neoplasia. J Pathol 2007;211:1-9.
www.elsevier.com/locate/humpath
Original contribution
Trials of new germ cell immunohistochemical stains in 93 extragonadal and metastatic germ cell tumors Kenneth A. Iczkowski MD a,b,⁎, Samantha L. Butler MD a , Jonathan H. Shanks MD c , Deloar Hossain MD d , Albrecht Schall MD d , Isabelle Meiers MD e , Ming Zhou MD f , Kathleen C. Torkko PhD g , Stacy J. Kim MD h , Gregory T. MacLennan MD h a
Department of Pathology, Immunology, and Laboratory Medicine, University of Florida, Gainesville, FL 32608, USA Department of Pathology and Laboratory Medicine, Malcom Randall Veterans Affairs Medical Center, Gainesville, FL 32610, USA c Christie Hospital, Manchester, M20 4BX UK d University of Manitoba Health Sciences Centre, Winnipeg, R3A1R9, Manitoba, Canada e Bostwick Laboratories, Glen Allen, VA 23060, USA f Cleveland Clinic, Cleveland, OH 44195, USA g Department of Pathology, University of Colorado, Aurora, CO 80045, USA h University Hospitals, Cleveland, OH 44106, USA b
Received 1 November 2006; revised 23 June 2007; accepted 3 July 2007
Keywords: Extragonadal; Germ cell tumor; Pineal; Mediastinum; c-kit; OCT3/4; D2-40; AP-2γ
Summary Organic cation transporter 3/4 (OCT3/4) is a transcription factor of embryonic stem cells; c-kit (CD117) is a tyrosine kinase receptor implicated in seminoma carcinogenesis. Their reactivity is well characterized in testicular, but not extragonadal and metastatic, germ cell tumors. A total of 93 germ cell tumors (41 seminoma, 22 embryonal carcinoma, 18 teratoma, and 12 yolk sac tumor) were obtained from the central nervous system (30), mediastinum (23), retroperitoneum/abdomen (31), and other locations (9). Immunohistochemical staining for c-kit, placental-like alkaline phosphatase (PLAP), OCT3/4, and new markers D2-40 and AP-2γ was performed on seminomas; CD30 and epithelial membrane antigen were added for nonseminomas. In embryonal carcinoma, c-kit reacted in 17 of 22 cases, OCT3/4 in 18 of 22, and PLAP in 13 of 22. OCT3/4 was superior to PLAP in intensity and percent cells staining. In seminoma, OCT3/4 and D2-40 were superior to PLAP in intensity and percent cells; c-kit and AP-2γ were superior in percent cells. D2-40 stained 23 of 24 seminomas strongly but had only weak focal reactivity in 6 of 17 embryonal carcinomas. Sensitivity and specificity were high for OCT3/4 discriminating seminoma and embryonal carcinoma, and c-kit discriminating seminoma, from other germ cell tumors. For embryonal carcinoma, OCT3/4 had higher specificity (0.94) than CD30 (0.786) owing to CD30 reactivity in 3 of 10 teratomas. Epithelial membrane antigen discriminated teratoma from other nonseminomas with a sensitivity of 1 but reacted occasionally in embryonal carcinoma (3/15) and yolk sac tumor (2/7). In conclusion, for extragonadal seminoma, OCT3/4, AP-2γ, D2-40, and c-kit were equivalently superior to PLAP. For embryonal carcinoma, OCT3/4 was superior to PLAP and more specific than CD30. D2-40 is recommended to discriminate between seminoma and embryonal carcinoma. © 2008 Elsevier Inc. All rights reserved.
⁎ Corresponding author. Pathology-P.O. Box 6511, UCHSC-Campus Mail Stop 8104, Aurora, CO 80045, USA. E-mail address: [email protected] (K. A. Iczkowski). 0046-8177/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.humpath.2007.07.002
276
1. Introduction Extragonadal germ cell tumors account for an estimated 2% to 5% of all germ cell tumors [1] and are usually encountered in the midline of 3 body sites: mediastinum, central nervous system (CNS), and retroperitoneum. Evidence suggests that such tumors in the mediastinum and in the CNS arise from primordial germ cells that failed to undergo apoptosis in fetal life [2]; germ cell tumors in the retroperitoneum are considered to represent metastases from a testicular primary site, whether or not a testicular tumor is demonstrable. The unique diagnostic dilemmas of tumors in these sites were reviewed [3] and result primarily from small sample size and frequent tumor necrosis. Accurate diagnosis is essential for treatment and prognosis. Seminomas, for instance, respond to radiotherapy and/or cisplatin-based chemotherapy, whereas nonseminomatous germ cell tumors do not respond to radiation therapy. New diagnostic immunohistochemical markers for germ cell tumors have been tested alongside established markers of metastatic seminoma [4] but not in a wide variety of extragonadal sites nor in a study that includes nonseminomatous germ cell tumors. Established germ cell immunohistochemical markers include placental-like alkaline phosphatase (PLAP), which reacts with all germ cell tumors; CD30 (Ki-1), which uniquely stains embryonal carcinoma; cytokeratin AE1/3, which reacts diffusely with embryonal carcinoma and yolk sac tumor and can give focal punctate perinuclear reactivity in seminoma; α-fetoprotein for yolk sac tumor; and epithelial membrane antigen (EMA) in teratoma and (with hCG) in choriocarcinoma. Four new markers for germ cell tumors have emerged recently. Organic cation transporter 3/4 (OCT3/4) is a nuclear transcription factor that regulates self-renewal and pluripotency of embryonic stem cells [5,6]. OCT3/4 protein is considered a specific and sensitive marker for intratubular germ cell neoplasia [7,8], seminoma and embryonal carcinoma [9], and ovarian dysgerminoma [10]. OCT3/4 was shown to give intense nuclear staining in 18 cases of seminoma and 29 cases of embryonal carcinoma from retroperitoneal metastases [11]. OCT3/4 is nonreactive in the other types of differentiated nonseminomatous tumors: yolk sac tumor, immature and mature teratoma, and choriocarcinoma [9]. Reactivity has been described in metastatic germ cell tumors [11] and CNS seminoma, commonly called germinoma [12]. c-kit (CD117), a 145-kD transmembrane glycoprotein, binds a stem cell factor that promotes dimerization and autophosphorylation of KIT receptors. KIT signaling is critical for normal development and survival of germ cells. Membranous c-kit gene expression is noted as a constant feature in intratubular germ cell neoplasia and primary seminoma [13], although its performance has not yet been compared with PLAP and conventional markers. The only c-kit expression reported in nonseminomatous tumors was in teratomas with sarcomatous histology, in a cytoplasmic pattern in 11 (48%) of 23 cases [14]; however, the status of c-kit in embryonal carcinoma was uncertain. D2-40, a membranous protein similar to podoplanin, has been proposed as a superior marker
K. A. Iczkowski et al. for seminoma, and AP-2γ, a transcription factor, for both seminoma and embryonal carcinoma [4]. We set out to study the reactivity of new markers in extragonadal germ cell tumors in primary mediastinal, CNS, and retroperitoneal and metastatic locations. We also aimed to compare the reactivity of new markers to established germ cell tumor immunohistochemical stains, including PLAP, EMA, and CD30.
2. Materials and methods With internal review board approval from each institution, surgical pathology files were searched for extragonadal and metastatic germ cell tumors at Shands Hospital, Gainesville, FL; University of Arkansas for Medical Sciences in Little Rock; University Hospitals, Cleveland; Cleveland Clinic; PCA Southeast, Columbia, TN; University of Manitoba Health Sciences Centre in Winnipeg; Christie Hospital, Manchester, UK; and the Institut de Pathologie et de Genetique, Gosselies, Belgium. We obtained and rereviewed 93 germ cell tumors (41 seminoma/germinoma, 23 embryonal carcinoma, 18 mature teratoma—2 with immature component—and 12 yolk sac tumor) from cerebrum (11), pineal (18), suprasellar (1), mediastinum (23), retroperitoneum/abdomen (31), and other sites (9). The other sites included lung (5), pericardium, neck lymph nodes, liver, and thigh. All CNS and mediastinal tumors were primary. Eleven cases were metastatic: known primary testicular tumors had been diagnosed in 1 case of CNS and 1 case of mediastinal embryonal carcinoma and in 8 (26%) retroperitoneal tumors. Chemotherapy is known to alter immunohistochemical expression of CD30 [15]; however, no specimens from patients who underwent it were included. The immunohistochemical panel used for seminomas was OCT3/ 4, PLAP, and c-kit (CD117). For nonseminomatous germ cell tumors, the panel included OCT3/4, PLAP, c-kit, EMA, and CD30 (Ki-1). Immunostains for AP-2γ and D2-40 were performed on a subset of 23 seminoma and 26 to 30 nonseminoma tumors, limited by availability. Heat-induced epitope retrieval was used for all antigens except EMA. The dilution of OCT3/4 (C20, sc8629; Santa Cruz Biotechnology, Santa Cruz, CA) goat antibody was 1:40 for 32 minutes at 42°C. D2-40 antibody (Abcam, Cambridge, MA) was used prediluted for 1 hour, and AP-2γ (H-77, Santa Cruz) was used at a 1:50 dilution for 1 hour. Rabbit antigoat secondary antibody was diluted at 1:100 for 32 minutes. The other 4 antibodies were detected using room temperature incubations with Ventana i-View DAB kit reagents (Ventana Medical Systems, Tucson, AZ). c-kit (IgG; CellMarque, Hot Springs, AR) came in prediluted form, as did PLAP (Ventana). The dilutions were 1:2 for CD30 (Cell Marque) and 1:50 for EMA (DakoCytomation, Carpinteria, CA). Reactivity was scored in terms of intensity and percent reactivity of the tumor. For OCT3/4 and AP-2γ, nuclear reactivity was evaluated; for all other stains, cytoplasmic or
Trials of new germ cell immunohistochemical stains
277
membranous reactivity was scored. Intensity of reactivity was recorded as 0 (no staining), 1+ (weak intensity), 2+ (moderate intensity), or 3+ (strong intensity). The percentage of reactive tumor cells (0%-100%) was estimated based on entire viable tumor present. Differences in intensity were assessed using Fisher exact test, and differences in percent using general linear mixed model. Matched pairs of immunostains were compared using Wilcoxon signed rank test. The sensitivity and specificity of markers for certain tumor types were calculated.
3. Results 3.1. Seminoma and embryonal carcinoma The reactivity of seminoma and embryonal carcinoma with the immunostaining panels is detailed (Table 1). Table 1 tumors Tumor
Protein expression by 93 extragonadal germ cell Embryonal carcinoma
PLAP, cytoplasm Cases + 13/22 Intensity 1.2 % cells 27 c-kit Cases + 17/22 Intensity 1.4, cytoplasmic % cells 32 † (1-90) OCT3/4, nucleus Cases + 18/22 Intensity 1.9 % cells 64 D2-40, cytoplasm Cases + 6/17 Intensity 0.7 % cells 5 AP-2γ, nucleus Cases + 10/14 Intensity 1.5 % cells 56 CD30, cytoplasm Cases + 16/22 Intensity 2.2 % cells 65 EMA, cytoplasm Cases + 6/18 Intensity 2.0 % cells 48
P⁎
Seminoma
P⁎
32/40 2.1 57 ‡
ns
40/41 2.4, membranous 84 (20-100)
.0001
.006 b.0001
41/41 2.6 84
.003 b.0001
ns ns
23/24 2.67 91
.008 b.0001
ns .006
22/23 2.5 82
ns .001
ns
ns
nd
nd
Abbreviations: ns, not significant; nd, not done. ⁎ P values represent superiority to PLAP based on Wilcoxon signed rank test. † Percent cells staining in embryonal carcinoma is less than in seminoma, P b .0001. ‡ Percent of seminoma staining with PLAP is less than with all 4 other antibodies, P b .0001.
Fig. 1 Pineal germinoma in a 15-year-old boy. All figures are on the same focus. A, OCT3/4 shows strong nuclear reactivity. B, c-kit with reactivity predominantly in membranous location. C, PLAP with very weak reactivity (×400).
For the purpose of this analysis, germinomas of the CNS were included in the group termed seminoma. OCT3/4 discriminated seminoma and embryonal carcinoma from all other germ cell tumors (sensitivity, 0.88; specificity, 0.94).
278
K. A. Iczkowski et al.
Fig. 2 Mediastinal embryonal carcinoma in a 25-year-old man. All figures are on the same focus. A, OCT3/4 is strongly reactive. B, c-kit was reactive in 11 of 15 extragonadal embryonal carcinomas, but the pattern was focal, and cytoplasmic reactivity predominated over membranous reactivity, unlike seminoma (Fig. 1B). C, PLAP shows no reactivity. D, CD30 gives focal reactivity (×400).
Compared with PLAP, the percentage of seminoma cells was higher for OCT3/4 (P b .001), c-kit (P b .0001), AP-2γ (P = .0009), and D2-40 (P b .0001); intensity of reactivity was superior for OCT3/4 (P = .0027) and D2-40 (P = .0078) only. For embryonal carcinoma, the percent of cells staining was higher than PLAP for OCT3/4 (P b .0001) and AP-2γ (P = .006) and D2-40; intensity was again superior to PLAP only for OCT3/4 (P b .006). Sensitivity was 0.96 for seminoma with AP-2γ and D2-40; it was 0.97 for c-kit, but specificity was low (0.57) owing to reactivity with 17 of 22 embryonal carcinomas. In all seminomas, the c-kit reactivity was predominantly crisply membranous. Reactivity of a pineal germinoma with PLAP, OCT3/4, and D2-40 is shown (Fig. 1). In this case, there is nuclear reactivity for OCT3/4, strong membranous reactivity for c-kit, and weak to absent reactivity for PLAP. OCT3/4 and CD30 both discriminated embryonal carcinoma from other nonseminomas. They were reactive in 18 of 22 and 16 of 22 cases and not significantly different in intensity or percent cells. However, OCT3/4 had higher specificity (0.94) than CD30 (0.786) for embryonal carcinoma owing to CD30 reactivity in 8 of 18 teratomas—
Fig. 3 Seminoma in retroperitoneal lymph node. D2-40 reacted with germinal center cells in lymph nodes (left), although weakly enough so as not to risk confusion with metastatic seminoma (right) (×200).
Trials of new germ cell immunohistochemical stains although this reactivity for CD30 would not be a diagnostic problem. c-kit was, as mentioned above, reactive in many embryonal carcinomas; the intensity and percent of cells were comparable to PLAP. Reactivity of a mediastinal embryonal carcinoma with these 4 immunostains is shown (Fig. 2A-D). The percent of cells in embryonal carcinoma reactive with PLAP, c-kit, AP-2γ, or D2-40 averaged significantly less than in seminoma, and the intensity was nonsignificantly lower. A constant observation was that patterns of reactivity were different from the membranous predominance in seminoma. Among 22 embryonal carcinomas, reactivity was cytoplasmic in 10 and cytoplasmic plus membranous in 12. The cytoplasmic pattern of reactivity is shown (Fig. 2B). Notably, both D2-40 and AP-2γ had consistent, significant reactivity for other cell types. D2-40 reacted with germinal center cells in lymph nodes, although weakly enough so as not to risk confusion with metastatic seminoma (Fig. 3). AP-2γ strongly reacted with the angulated nuclei of adipocytes, whose shape minimizes likelihood of confusion with the rounded nuclei of seminoma cells (Fig. 4).
3.2. Other germ cell tumors EMA discriminated teratoma from other nonseminomas with a sensitivity of 1 and a specificity of 0.76. The EMA was reactive in all nonneural components of the mature teratomas. The low specificity resulted from occasional reactivity in embryonal carcinoma (3/15) and yolk sac tumor (2/7). c-kit was reactive in 3 of 7 yolk sac tumors, with 1+, 2+, and 3+ intensity in 10% to 90% of cells and a membranous plus cytoplasmic pattern. None of the 7 tumors was reactive for PLAP or CD30, but two had some reactivity for EMA and one for OCT3/4. CD30 reactivity was not confined to embryonal carcinoma but was reactive in some (4/10) mature teratomas. In
Fig. 4 Seminoma in retroperitoneal lymph node. AP-2γ strongly reacted with the angulated nuclei of adipocytes (left), which would be unlikely to be confused with the rounded nuclei of seminoma cells (right) (×200).
279 these cases, reactivity was confined to the respiratory component in 3 cases, the squamous component in 2 cases, and gastrointestinal epithelium, nerve, or cartilage in 1 case each. Immature teratoma was not reactive with any stain.
3.3. Reactivity according to tumor site For each tumor type, all markers were compared for their reactivity according to site: CNS, mediastinum, retroperitoneum, or other. The only difference among markers by tumor site was more intense OCT3/4 reactivity in retroperitoneal/abdominal seminoma (3.0+) than at other sites (2.2-2.4+) (P = .0068). Furthermore, no difference was evident between the primary and the metastatic tumors for any tumor type or marker.
4. Discussion The current study is novel in 4 ways: (1) it shows that OCT3/4, AP-2γ, and D2-40 perform well as markers for mediastinal seminoma, extending the variety of extragonadal germ cell tumor locations in which these have been tested; (2) for nonseminomatous tumors, it tests these new markers in comparison with established ones, whereas all studies but two dealt only with seminomas [2,11]; (3) it is the first to test c-kit reactivity in nonseminomatous germ cell tumors, disclosing reactivity in 17 of 22 embryonal carcinomas; (4) it adds CD30 and EMA to the panel for the nonseminomatous tumors to evaluate their comparative diagnostic potential. With seminoma of extragonadal or metastatic sites, OCT3/ 4 and c-kit were superior to PLAP in the number of positive cases, intensity of staining, and percent cells. This finding reinforces a prior study in which c-kit was superior to PLAP in 25 CNS germinomas, staining 100% versus 28% of cases [16]. Recent studies involving 25 cases of CNS germinoma [12] and 30 biopsy specimens of retroperitoneal seminoma [17] found superior reactivity of OCT3/4 compared with PLAP. Interestingly, in the latter study, OCT3/4 was also superior to c-kit: OCT3/4 marked 100% of cases, whereas ckit marked 21 (70%) of 30 of cases, and PLAP marked 23 (77%) [17]. In contrast, we found essentially no difference in c-kit versus OCT3/4 sensitivity for extragonadal seminomas from a wider range of locations. Biermann et al [4] recently compared OCT3/4 and c-kit to seminoma markers D2-40 and AP-2γ. A total of 30 extragonadal seminomas were included, 28 metastatic and 2 pineal, using this panel. D2-40 and AP-2γ yielded strong reactivity compared with weak reactivity for PLAP in 63% of cases, OCT3/4 in 50%, and c-kit in 40%. However, our study showed all 4 seminoma markers essentially equivalent and superior to PLAP as shown previously for OCT3/4 [9,11,12] and c-kit [16,17]. Technical variations, particularly with antigen retrieval, may explain this disparity.
280 Interestingly, the presence of c-kit expression in all but one of our cases of extragonadal seminoma does not necessarily imply that all seminomas are sensitive to imatinib (Gleevec) treatment. Treatment sensitivity depends on c-kit activating mutations in either exon 11 or 17. In a recent study, these mutations were present in only 5 (23%) of 22 testicular seminomas, although 19 (86%) had membranous c-kit immunoreactivity [18]. In extratesticular embryonal carcinoma, OCT3/4 was again strikingly superior to PLAP in number of cases, intensity, and percent cells. We found that the most (18/22) of the embryonal carcinomas were reactive for OCT3/4, with a slight edge over the other nuclear stain, AP-2γ, whereas only 13 of 22 cases reacted with PLAP. The frequent reactivity for c-kit (17/22) was an unexpected finding, making D2-40 a better discriminator since the number of cases (6/17), intensity, and percent of cells reactive were less than for c-kit. The reactivity of c-kit in extragonadal embryonal carcinoma had been tested in just one case in a prior study [2], where it was negative. If c-kit is used, the likelihood of confusion with seminoma is reduced not only by the lower percent of cells showing reactivity in most cases but also by different localization. Embryonal carcinoma reactivity was cytoplasmic, not membranous as in seminoma. In a number of other tumor types studied, a variety of c-kit reactivity patterns have been reported. Predominant cytoplasmic reactivity of c-kit has been noted, for example, in sarcomatoid teratoma [14] and solid-pseudopapillary neoplasm of the pancreas [19]. In gastrointestinal stromal tumor, c-kit reactivity may be membranous, cytoplasmic, or punctate perinuclear (Golgi) [20]. CD30 and EMA stained predominantly embryonal carcinoma and teratoma, respectively, but had mild specificity problems. In embryonal carcinoma, OCT3/4 showed nearly identical reactivity to CD30 in number of cases, intensity, and percent of cells. This stands in contrast to a study of 25 metastatic embryonal carcinomas after chemotherapy, in which 76% had moderate to strong OCT3/4 staining, but only 28% had weak focal CD30 expression [21]. In all of our cases, patients had not undergone chemotherapy, so our observation supports the theory that chemotherapy markedly reduces CD30 [15]. The respiratory or squamous elements of some teratomas were noted to react with CD30. OCT3/4 was more specific for embryonal carcinoma and was not reactive in teratoma, in which columnar glands could be confused with embryonal carcinoma. The presence of c-kit reactivity in a membranous/ cytoplasmic pattern in yolk sac tumor was also an unexpected finding. None of the 7 extragonadal yolk sac tumors was PLAP-reactive, and other markers tested were usually negative. In a prior study of mediastinal versus testicular germ cell tumors, no difference in PLAP reactivity was found among yolk sac tumors [22]. In conclusion, the newer immunostains OCT3/4 (reviewed recently [23]), c-kit, AP-2γ, and D2-40 offer improved sensitivity and specificity for detecting extrago-
K. A. Iczkowski et al. nadal seminoma and embryonal carcinoma in various sites compared with PLAP. Both OCT3/4 and c-kit or D2-40 are recommended for establishing a diagnosis of extragonadal seminoma. An immunostaining panel to help with the differential diagnosis between seminoma and embryonal carcinoma should include c-kit and CD30, with the caveat that some c-kit reactivity may be expected in embryonal carcinoma, but it is usually focal and has a cytoplasmic pattern. For embryonal carcinoma in which the presence of other nonseminomatous elements is suspected, α-fetoprotein and EMA may be useful in addition to CD30. The diagnosis of mature teratoma was supported by EMA reactivity in the gastrointestinal and respiratory components. For extragonadal yolk sac tumor, PLAP did not show reactivity, although caution is needed with c-kit because 3 of 7 tumors were reactive. Thus, α-fetoprotein reactivity is probably still the best marker for yolk sac tumor.
Acknowledgment We thank Elaine Dooley (Univ. of Florida) for technical assistance. Jesse K. McKenney, MD (Little Rock, AR), and Cory Schlesigner, MD (Columbia, TN) contributed cases.
References [1] Bosl GJ, Motzer RJ. Testicular germ-cell cancer. N Engl J Med 1997; 337:242-53. [2] Hoei-Hansen CE, Sehested A, Juhler M, et al. New evidence for the origin of intracranial germ cell tumours from primordial germ cells: expression of pluripotency and cell differentiation markers. J Pathol 2006;209:25-33. [3] McKenney JK, Heerema-McKenney A, Rouse RV. Extragonadal germ cell tumors: a review with emphasis on pathologic features, clinical prognostic variables, and differential diagnostic considerations. Adv Anat Pathol 2007;14:69-92. [4] Biermann K, Klingmüller D, Koch A, et al. Diagnostic value of markers M2A, OCT3/4, AP-2γ, PLAP and c-kit in the detection of extragonadal seminomas. Histopathology 2006;49:290-7. [5] Looijenga LH, Stoop H, de Leeuw HP, et al. POU5F1 (OCT3/4) identifies cells with pluripotent potential in human germ cell tumors. Cancer Res 2003;63:2244-50. [6] Honecker F, Oosterhuis JW, Mayer F, et al. New insights into the pathology and molecular biology of human germ cell tumors. World J Urol 2004;22:15-24. [7] Jones TD, Ulbright TM, Eble JN, et al. OCT4: a sensitive and specific biomarker for intratubular germ cell neoplasia of the testis. Clin Cancer Res 2004;10:8544-7. [8] de Jong J, Stoop H, Dohle GR, et al. Diagnostic value of OCT3/4 for pre-invasive and invasive testicular germ cell tumors. J Pathol 2005; 206:242-9. [9] Jones TD, Ulbright TM, Eble JN, et al. OCT4 staining in testicular tumors: a sensitive and specific marker for seminoma and embryonal carcinoma. Am J Surg Pathol 2004;28:935-40. [10] Cheng L, Thomas A, Roth LM, et al. OCT4: a novel biomarker for dysgerminoma of the ovary. Am J Surg Pathol 2004;28:1341-6. [11] Cheng L. Establishing a germ cell origin for metastatic tumors using OCT4 immunohistochemistry. Cancer 2004;101:2006-10.
Trials of new germ cell immunohistochemical stains [12] Hattab EM, Tu PH, Wilson JD, et al. OCT4 immunohistochemistry is superior to placental alkaline phosphatase (PLAP) in the diagnosis of central nervous system germinoma. Am J Surg Pathol 2005;29: 368-71. [13] Kemmer K, Corless CL, Fletcher JA, et al. KIT mutations are common in testicular seminomas. Am J Pathol 2004;164:305-13. [14] Madani A, Kemmer K, Sweeney C, et al. Expression of KIT and epidermal growth factor receptor in chemotherapy refractory nonseminomatous germ-cell tumors. Ann Oncol 2003;14:873-80. [15] Berney DM, Shamash J, Pieroni K, et al. Loss of CD30 expression in metastatic embryonal carcinoma: the effects of chemotherapy? Histopathology 2001;39:382-5. [16] Nakamura H, Takeshima H, Makino K, et al. C-kit expression in germinoma: an immunohistochemistry-based study. J Neuroendocrinol 2005;75:163-7. [17] Sung M-T, MacLennan GT, Cheng L. Retroperitoneal seminoma in limited biopsies: morphologic criteria and immunohistochemical findings in 30 cases. Am J Surg Pathol 2006;30:766-73.
281 [18] Willmore-Payne C, Holden JA, et al. Detection of c-kit exons 11- and 17-activating mutations in testicular seminomas by high-resolution melting amplicon analysis. Mod Pathol 2006;19:1164-9. [19] Cao D, Antonescu C, Wong G, et al. Positive immunohistochemical staining of KIT in solid-pseudopapillary neoplasms is not associated with KIT/PDGFRA mutations. Mod Pathol 2006;19:1157-63. [20] Miettinen M, Lasota J. Gastrointestinal stromal tumors: review on morphology, molecular pathology, prognosis, and differential diagnosis. Mod Pathol 2006;130:1466-78. [21] Sung M-T, Jones TD, Beck SD, et al. OCT4 is superior to CD30 in the diagnosis of metastatic embryonal carcinomas after chemotherapy. HUM PATHOL 2006;37:662-7. [22] Suster S, Moran CA, Dominguez-Malagon H, et al. Germ cell tumors of the mediastinum and testis: a comparative immunohistochemical study of 120 cases. HUM PATHOL 1998;29:737-42. [23] Cheng L, Sung M-T, Cossu-Rocca P, et al. OCT4: biological functions and clinical applications as a marker of germ cell neoplasia. J Pathol 2007;211:1-9.