HSD3B (3β-hydroxysteroid dehydrogenase) is required for the biosynthesis of steroid hormones.1 Among the isoforms, hydroxyl-C-5-steroid dehydrogenase (HSD3B1) is abundantly expressed in the human placenta and is involved in the biosynthesis of progesterone for the maintenance of pregnancy.2,3 Polymorphism of the HSD3B1 gene has been associated with prostate cancer susceptibility,4 an increased risk of benign prostate hyperplasia,5 and essential hypertension.6 Immunohistochemically, it has been found to be expressed in normal trophoblastic tissue and benign and neoplastic trophoblastic lesions, but not in carcinomas from the lung, breast, uterine endometrium, or cervix,7 thus serving as a sensitive and specific marker for the diagnosis of trophoblastic lesions and tumors. However, the expression of HSD3B1 in carcinomas from sites other than the lung, breast, and uterus is largely unknown. To verify the specificity of HSD3B1, we examined its expression in carcinomas from various sites including the lung, breast, ovary, uterine cervix, liver, pancreas, stomach, colon, pancreas, kidney, and prostate and in tumors with potential synthesis of steroid hormones, including ovarian granulosa cell tumors and adrenocortical adenomas.
MATERIALS AND METHODS
Paraffin tissues of carcinomas from various sites were constructed onto tissue microarrays (TMAs). A total of 10 TMAs containing 473 carcinomas from the lung (n = 54), breast (n = 39), ovary (n = 52), uterine cervix (n = 38), liver (n = 85), pancreas (n = 29), stomach (n = 47), colon (n = 52), kidney (n= 45), and prostate (n = 32) and 1 TMA containing 32 ovarian granulosa cell tumors and 12 adrenocortical adenomas were included in this study (Table 1). One TMA containing normal tissues from the cerebrum, oral tonsil, major salivary gland, thyroid, gastrointestinal tract, pancreas, spleen, adrenal gland, testis, ovary, endometrium, and placenta was also included. All cases were anonymized, and the use of archival material was approved by the institutional review board.
Immunohistochemistry was carried out using a mouse monoclonal antibody against HSD3B1 (clone 3C11-D4; Abnova Corporation, Taipei, Taiwan). The specificity of the antibody was proved by the presence of a single band on a Western blot in our previous study.7 Briefly, the deparaffinized sections were heated in 0.01 M citric buffer (pH 6.0) in an autoclave at 120°C for 10 minutes. After incubation with the primary antibody at 4°C overnight, a positive reaction was detected by the EnVision+System (DAKO, Carpinteria, CA) and developed with 3,3-diaminobenidine.
The immunostains were examined according to the cytoplasmic intensity, and the percentages of stained cells were determined independently by 2 pathologists. The staining intensity was scored based on the cytoplasmic staining ranging from 0 to 4+, and the percentage of stained cells from 0% to 100% was recorded. An H score of (percentage score) × (intensity score + 1) was used as in our previous study.7 Because most tumors are negative for HSD3B1, for practical analysis, tissues with weak, moderate, or strong staining were lumped together as the HSD3B1-positive group and compared with HSD3B1-negative tissues. Normal placenta served as a positive control.
Expression of HSD3B1 in Normal Tissues
HSD3B1 was strongly expressed in the syncytiotrophoblasts and intermediate trophoblasts in the implantation site and chorion laeve of normal placentas (Fig. 1). Positive staining was noted in the cytoplasm of the trophoblasts but not in the surrounding stromal cells. The picture is the same as in our previous study.7 All the normal tissues from the cerebrum, oral tonsil, major salivary gland, thyroid, gastrointestinal tract, pancreas, spleen, adrenal gland, testis, endometrium, and ovary were negative for HSD3B1.
Expression of HSD3B1 in Tumors
All 473 carcinomas from the lung, breast, ovary, uterine cervix, liver, pancreas, stomach, colon, kidney, and prostate showed negative reactivity for HSD3B1 (Figs. 2A–D). Although ovarian granulosa cell tumors and adrenocortical adenomas are capable of producing steroid hormones, all 32 granulosa cell tumors were negative. Most of the adrenocortical adenomas were also negative except for two that showed weak focal immunoreactivity, with much lower H scores as compared with those of normal placentas and trophoblastic lesions from our previous study.7
HSD3B encoded by the HSD3B gene family at 1p13 is required for the biosynthesis of all active steroid hormones. Several isoforms have been identified in humans. HSD3B1 is the only isoform expressed in the placenta. It plays a major role in the biosynthesis of progesterone and is necessary for the maintenance of pregnancy.2,3 Our previous investigation revealed that HSD3B1 is a specific trophoblastic marker that is highly expressed in syncytiotrophoblasts and intermediate trophoblasts in the implantation site and the chorion laeve, variably expressed in the intermediate trophoblasts in the trophoblastic column, and undetectable in cytotrophoblasts. Only 3 of 319 nontrophoblastic carcinomas from the uterus, lung, and breast showed focal and weak immunoreactivity.7 In this study, we collected a much wider variety of nontrophoblastic tumors and studied the expression of HSD3B1. We found all 473 nontrophoblastic carcinomas and 12 normal tissues did not react with anti-HSD3B1 antibody. Although ovarian granulosa cell tumors and adrenocortical carcinomas are capable of synthesizing steroid hormones, only 2 adrenocortical adenomas showed weak staining. Among the HSD isoforms, HSD3B2 but not HSD3B1 is expressed in normal ovaries and adrenal glands. Hence, the negative immunoreactivity of ovaries, adrenal glands, granulosa cell tumors, and adrenocortical adenomas to HSD3B1 in this study suggests that the anti-HSD3B1 antibody we applied is highly specific and does not cross-react with HSD3B2.
In addition to molar gestation, trophoblast-associated lesions include tumorlike lesions, such as exaggerated placental sites and placental site nodules (PSNs), and tumors such as placental site trophoblastic tumors (PSTTs), epithelioid trophoblastic tumors (ETTs), and choriocarcinomas.8 These trophoblastic tumorlike lesions and trophoblastic tumors can be morphologically similar to other nonneoplastic lesions and carcinomas. Furthermore, because of their relatively low incidence, recognizing them as trophoblastic lesions and making a correct diagnosis can sometimes be difficult. Thus, a specific marker that can demonstrate most types of trophoblastic cells and is not expressed in other types of tissues and tumors is important. As an extension of our previous study, we showed that HSD3B1 is not expressed in a variety of normal and neoplastic lesions, confirming its high specificity. HSD3B1 shows restricted expression in trophoblasts, unlike other commonly used trophoblastic markers such as β-human chorionic gonadotropin, human placental lactogen, placental alkaline phosphatase, melanoma cell adhesion molecule (CD146), human leukocyte antigen G,9 and p63, which either are expressed only in a subtype of trophoblasts or are not specific for trophoblasts and can be expressed in other various nontrophoblastic lesions. Thus, our study further supports that HSD3B1 is superior to other commercially available trophoblast antibodies in distinguishing trophoblastic tumorlike lesions or tumors from nontrophoblastic lesions or carcinomas because it is expressed in virtually all types of trophoblastic tumorlike lesions and trophoblastic tumors. Moreover, HSD3B1 is highly specific and not expressed or only weakly expressed in other nontrophoblastic lesions and carcinomas.
Primary nongestational choriocarcinomas have been reported in the lung,10 uterine cervix,11 stomach,12 and liver.13 Besides choriocarcinoma, other trophoblastic tumors including ETT and PSTT may metastasize to sites such as the lung, liver, brain, and spine.14,15 On rare occasions, they may present with metastatic disease before the primary trophoblastic tumor is diagnosed. Furthermore, ETT has been reported primarily in extrauterine sites.16,17 In the uterus, ETT and PSTT may present in the lower uterine segment or cervix, simulating cervical cancer. The characteristic dimorphic trophoblastic population with marked hemorrhage in choriocarcinomas is not hard to identify as long as the clinician is familiar with this tumor. However, choriocarcinomas may occasionally contain attenuated syncytiotrophoblasts, whereas mononucleate trophoblasts are the predominant population. Furthermore, the monomorphic population of tumor cells in sheets in ETT and PSTT can be very similar to poorly differentiated carcinomas. These trophoblastic tumors are epithelioid in morphology and positive for epithelial markers. Without awareness of their trophoblastic nature, a correct diagnosis and proper treatment cannot be rendered. Immunohistochemical study using a specific HSD3B1 antibody is of great help in these conditions.
A stepwise immunohistochemical algorithm called a trophogram using a panel of commercially available antibodies in the diagnosis of lesions suspected to be of trophoblastic origin is suggested.18 In this 3-tiered approach, HSD3B1 appears to be the first-line antibody to be used because of its specificity and sensitivity to all types of trophoblastic tumorlike lesions and tumors. Once a lesion is confirmed to be trophoblastic in nature, β-human chorionic gonadotropin, p63, and human placental lactogen can be applied to classify the lesion into a choriocarcinoma, a lesion with implantation site intermediate trophoblasts, or a lesion with chorionic-type intermediate trophoblasts. Ki-67 is further utilized to differentiate an exaggerated placental site from a PSTT and Ki-67 plus cyclin E to differentiate a PSN from an ETT. The sequence of application of these antibodies is important because p63, cyclin E, and Ki-67 are useful only in the context of a lesion confirmed to be trophoblastic but cannot be applied in distinguishing nontrophoblastic lesions. The importance of HSD3B1 in the first setting in this algorithm cannot be overemphasized.
In summary, our study demonstrated that HSD3B1 is a highly specific trophoblast-associated marker that can be used in the distinction of trophoblastic tumorlike lesions and tumors from nontrophoblastic lesions and tumors. The correct diagnosis of trophoblastic tumors especially at sites outside the uterus is of paramount importance for proper treatment and management of patients.
The authors thank Dr Wilber Huang (Abnova Corporation) for technical support of tissue microarray construction.
1. Simard J, Ricketts ML, Gingras S, et al.. Molecular biology of the 3beta-hydroxysteroid dehydrogenase/delta5-delta4 isomerase gene family. Endocr Rev
. 2005; 26: 525–582.
2. Simard J, Durocher F, Mebarki F, et al.. Molecular biology and genetics of the 3β-hydroxysteroid dehydrogenase/delta5-delta4 isomerase gene family. J Endocrinol
. 1996; 150: S189–S207.
3. Peng L, Huang Y, Jun F, et al.. Transcription enhancer factor-5 and a GATA-like protein determine placental-specific expression of the type I human 3β-hydroxysteroid dehydrogenase gene, HSD3B1
. Mol Endocrinol
. 2004; 18: 2049–2060.
4. Chang BL, Zheng SL, Hawkins GA, et al.. Joint effect of HSD3B1
and HSD3B2 genes is associated with hereditary and sporadic prostate cancer susceptibility. Cancer Res
. 2002; 15: 1748–1749.
5. Roberts RO, Bergstralh EJ, Farmer SA, et al.. Polymorphisms in genes involved in sex hormone metabolism may increase risk of benign prostate hyperplasia. Prostate
. 2006; 66: 392–404.
6. Shimodaira M, Nakayama T, Sato N, et al.. Association of HSD3B1
and HDS3B2 gene polymorphisms with essential hypertension, aldosterone level, and left ventricular structure. Eur J Endocrinol
. 2010; 163: 671–680.
7. Mao TL, Kurman RJ, Jeng YM, et al.. HSD3B1
as a novel trophoblast
-associated marker that assists in the differential diagnosis of trophoblastic tumors and tumorlike lesions. Am J Surg Pathol
. 2008; 32: 236–242.
8. Shih IM. Gestational trophoblastic neoplasia—pathogenesis and potential therapeutic targets. Lancet Oncol
. 2007; 8: 642–650.
9. Snger G, Kurmain RJ, McMaster MT, et al.. HLA-G immunoreactivity is specific for intermediate trophoblast
in gestational trophoblastic disease and can serve as a useful marker in differential diagnosis. Am J Surg Pathol
. 2002; 26: 914–920.
10. Pushchak MJ, Farhi DC. Primary choriocarcinoma of the lung. Arch Pathol Lab Med
. 1987; 111: 477–479.
11. Brunner J, Högberg T, Malmström H, et al.. Postmenopausal extragenital choriocarcinoma. A case report and review of the literature. Eur J Gynaecol Oncol
. 1991; 12: 395–398.
12. Krulewski T, Cohen LB. Choriocarcinoma of the stomach: pathogenesis and clinical characteristics. Am J Gastroenterol
. 1988; 83: 1172–1175.
13. Arai M, Oka K, Nihei T, et al.. Primary hepatic choriocarcinoma—a case report. Hepatogastroenterol
. 2001; 48: 424–426.
14. Guvendag Guven ES, Guven S, Esinler I, et al.. Placental site trophoblastic tumor in a patient with brain and lung metastases. Int J Gynecol Cancer
. 2004; 14: 558–563.
15. Chohan MO, Rehman T, Cerilli LA, et al.. Metastatic epithelioid trophoblastic tumor involving the spine. Spine
. 2010; 35: E1072–E1075.
16. Kuo KT, Chen MJ, Lin MC. Epithelioid trophoblastic tumor of the broad ligament: a case report and review of the literature. Am J Surg Pathol
. 2004; 28: 405–409.
17. Shih IM, Kurman RJ. Epithelioid trophoblastic tumor: a neoplasm distinct from choriocarcinoma and placental site trophoblastic tumor simulating carcinoma. Am J Surg Pathol
. 1998; 22: 1393–1403.
18. Shih IM. Trophogram, an immunohistochemistry
-based algorithmic approach, in the differential diagnosis of trophoblastic tumors and tumorlike lesion. Ann Diagn Pathol
. 2007; 11: 228–234.