There are currently six glypican family members (GPC1 to GPC6) that have been identified in mammals, and two in Drosophila melanogaster
]. Research of glypicans has increased due to the discovery that mutations of the gene encoding GPC3 results in a rare, X-linked overgrowth and dysmorphic syndrome called Simpson-Golabi-Behmel Syndrome (SGBS) [23
]. The GPC3 gene is localized to Xq26, and its product is thought to regulate cellular growth and embryonic cell differentiation through interactions with morphogenic or growth factors such as Wnt5a, fibroblast growth factor 2, bone morphogenetic protein 7, tissue factor pathway inhibitor and hedgehog proteins [11
]. A mutation in the gene speculated to render GPC3 nonfunctional results in the clinical phenotype of SGBS patients, which includes numerous craniofacial, skeletal, and genitourinary abnormalities [28
]. In addition, patients with SGBS are at increased risk for certain tumors, most frequently hepatocellular carcinoma, hepatoblastoma and Wilms tumor [28
]. GPC3 has also been linked to various sporadic tumors, particularly hepatocellular carcinoma, for which it has been shown to be a useful diagnostic marker helpful in differentiating hepatocellular carcinoma from non-neoplastic liver disease [31
GPC3 is expressed by normal human placental tissue, notably by the syncytiotrophoblast at term, and may play an important role in placental growth and development [18
]. In addition, a prior study has proposed that GPC3 may act as an anchor for placental protein 5 (PP5)/tissue factor pathway inhibitor-2 (TFPI-2) on placental villi, and a decrease in GPC3 expression in patients with preeclampsia may explain the increased levels of PP5/TFPI-2 in the serum of these patients [34
]. In testicular germ cell tumors, choriocarcinoma was consistently positive for GPC3 with strong staining in the malignant syncytiotrophoblast and weaker staining in the cytotrophoblast [19
]. Within ovarian mixed germ cell tumors, choriocarcinoma was found to be immunoreactive for GPC3, showing strong membranous reactivity with an additional case of endometrial choriocarcinoma in the same study demonstrating similar immunoreactivity [35
]. Although choriocarcinoma of the ovary is histologically identical to primary gestational choriocarcinomas, previous studies have not yet investigated the expression of GPC3 in GTN (including gestational choriocarcinoma, PSTT and ETT) or other trophoblastic lesions, such as exaggerated placental site (EPS) and PSN.
Although PSTT is very rare (about 1-2% of trophoblastic tumors overall), differentiation of PSTT from other types of GTN, non-neoplastic gestational trophoblastic disease and non-trophoblastic tumors is important clinically due to differences in their therapeutic approaches [36
]. Specifically, PSTT is refractory to chemotherapy and usually requires treatment that includes surgical resection or hysterectomy [16
]. Because PSTT commonly presents with a low and variable concentration of hCG in serum, tumors that secrete low to negligible hCG are clinical considerations. The histologic differential diagnosis of PSTT includes gestational choriocarcinoma, ETT, PSN, EPS, and non-trophoblastic tumors such as epithelioid leiomyosarcoma and poorly differentiated carcinoma, such as squamous cell carcinoma (SCC) of the cervix or endometrial adenocarcinoma [16
]. The histopathology of PSTT has been distinctly described as a proliferation of implantation site intermediate trophoblast cells, mostly mononuclear with occasional multinucleated tumor cells. The cells have marked nuclear atypia, and tend to form confluent sheets while extensively invading the uterus in nests that separate the myometrial fibers reminiscent of normal implantation [16
]. Unlike choriocarcinoma, which is usually dimorphic, PSTT is monophasic [36
]. However, there are overlapping morphologic features and immunoprofiles between subtypes of GTN, as well as mixed lesions, which can make a definitive diagnosis difficult [37
]. In addition, the small amount of tissue obtained from biopsy or endometrial curettage may limit the accuracy of the initial pathologic examination.
In our study, GPC3 was expressed in the majority (80%) of PSTT cases with an overall weak mean staining intensity. This expression is consistent with the expression of GPC3 in cells with extra-embryonic differentiation including strong expression in normal placenta as well as tumors which recapitulate the phenotype of extra-embryonic membranes (amnion, yolk sac, chorion, and allantois), such as gonadal and extra-gonadal choriocarcinoma and yolk sac tumor [18
]. The weak staining intensity of GPC3 in PSTT is concordant with previous conclusions that PSTT consists of implantation site intermediate trophoblastic cells that originate from transformed cytotrophoblasts, which have been described to stain weakly for GPC3 compared to the differentiated syncytiotrophoblast [18
]. On examination, the pattern of GPC3 in PSTT appears similar but weaker than that seen in gonadal and extra-gonadal choriocarcinoma, with stronger expression mostly limited to larger/multinucleate cells and weaker expression in smaller, mononucleate cells. This pattern is also described for hCG staining in PSTT and could be due to the greater level of differentiation of these larger cells [38
]. Other immunohistochemical markers have been reported for PSTT to help discriminate PSTT and other trophoblastic/non-trophoblastic diseases including hPL, β-hCG, Mel-CAM, cytokeratin 18, HLA-G and p63 [39
]. Each of these markers each have their own limitations, due to the lack of specificity of some of these markers for the trophoblast alone, as well as their variable expression in the different subtypes of trophoblast [37
]. GPC3 represents an additional marker with good sensitivity to corroborate the placental lineage of PSTT.
There have been no studies of GPC3 in other types of GTN or non-neoplastic placental lesions. Although previous studies have not looked specifically at GPC3 expression in gestational choriocarcinoma, one may predict that--like the results obtained from choriocarcinoma of testicular and ovarian germ cell tumors--gestational choriocarcinoma is expected to have similar strong reactivity in malignant syncytiotrophoblastic cells [21
]. Like PSTT, ETT consists of intermediate trophoblastic cells derived from a common trophoblastic stem cell. One could predict that ETT will stain for GPC3 in a similar fashion to PSTT, and so GPC3 might not be useful in distinguishing between the two. However, because both tumors are treated in similar ways, differentiation between the two may not be clinically crucial [16
]. The few cases of PSN tested in this study showed weak staining for GPC3, as expected for a lesion composed of intermediate trophoblastic cells. GPC3 does not appear to discriminate between PSTT and PSN.
Tumors with non-trophoblastic differentiation common to the uterus did not express GPC3. We did not find expression of GPC3 in endometrial adenocarcinoma similar to a prior report [33
]. In our study, a minority of cases of invasive cervical SCC had expression of GPC3. The pattern of basal staining was quite different than for PSTT and PSN and should allow for discrimination between these entities. A previous study also found that the majority of cervical SCC did not express GPC3 (15%) [42
]. GPC3 expression was also low in non-cervical SCC (anus, 20%; esophagus, 8%; oral cavity, 10%; skin, 2%; urinary bladder, 13%; vulva, 12%) [42
]. GPC3 was not expressed in uterine leiomyosarcoma and leiomyoma in the few cases we examined. Similarly Baumhoer et al tested a large number of leiomyosarcomas and found low GPC3 expression [42
]. GPC3 may be useful as an additional immunohistochemical marker to distinguish PSTT from non-trophoblastic tumors common to the uterus.
Recent studies have shown that GPC3 may act not only as a histochemical marker, but also as a serum marker for early detection of HCC with potential identification of patients who have high levels of GPC3 for possible targeted therapy [43
]. As serum levels of hCG are variable in PSTT and GPC3 is found in the majority of primary and metastatic cases of PSTT, GPC3 could represent a new serum marker to aid in the clinical diagnosis of primary lesions and in patient follow-up for the detection of metastases. Additionally, it has been shown that anti-GPC3 monoclonal antibodies targeting the C-terminal 30-kDa fragment of GPC3 in serum induces antibody-dependent cellular cytotoxicity (ADCC) and/or complement-dependent cytotoxicity (CDC) against GPC3-positive human HCC cells in culture [45
]. With the increasing detection of GPC3 in various tumors other than HCC, now including PSTT, extending this targeted therapy approach to other GPC3 positive tumors may be possible.