Abstract

Background. Hydatidiform moles (HM) are members of gestational trophoblastic diseases (GTD) and, in some cases, might progress to gestational trophoblastic neoplasia (GTN). HMs are either partial (PHM) or complete (CHM). Some HMs are challenging in arriving at a precise histopathological diagnosis. This study aims to investigate the expression of BCL-2 by immunohistochemistry (IHC) in HMs as well as in normal trophoblastic tissues “products of conception (POC) and placentas” using Tissue MicroArray (TMA) technique.
Methods. TMAs were constructed using the archival material of 237 HMs (95 PHM and 142 CHM) and 202 control normal trophoblastic tissues; POC and unremarkable placentas. Sections were immunohistochemically stained using antibodies against BCL-2. The staining was assessed semi-quantatively (intensity and percentage of the positive cells) in different cellular components (trophoblasts and stromal cells)
Results. BCL-2 showed cytoplasmic expression in more than 95% of trophoblasts of PHM, CHM and controls. The staining showed a significant reduction of the intensity from controls (73.7%), PHMs (76.3%) to CHM (26.9%). There was a statistically significant difference between PHM and CHM in the intensity (p-value 0.0005) and the overall scores (p-value 0.0005), but not the percentage score (p-value > 0.05). No significant difference was observed in the positivity of the villous stromal cells between the different groups. All cellular components were visible using the TMA model of two spots/case (3 mm diameter, each) in more than 90% of cases. Conclusions. Decreased BCL-2 expression in CHM compared to PHM and normal trophoblasts indicates increased apoptosis and uncontrolled trophoblastic proliferation. Construction of TMA in duplicates using cores of 3 mm diameter can overcome tissue heterogeneity of complex lesions.

Introduction

Gestational trophoblastic diseases (GTD) is a spectrum of lesions ranging from hydatidiform moles (HMs) to gestational trophoblastic neoplasia (GTN) ¹. Hydatidiform moles (HMs) are either partial (PHM) or complete (CNM) moles, whereas gestational trophoblastic neoplasia (GTN) includes invasive mole, choriocarcinoma, placental site trophoblastic tumor (PSTT) and epithelioid trophoblastic tumor ¹. Despite the well described histopathological criteria of HM lesions, some cases are challenging in arriving at a precise diagnosis ². Fortunately, most of HMs will regress spontaneously after curettage, although 8-30% of patients will progress to persistent GTD, especially following CHM, and will need special management, including chemotherapy ². To predict this aggressive behavior, current guidelines suggest weekly measurement of serum human chorionic gonadotropin (B-hCG) following the evacuation of pregnancy products ^2,3. Persistent GTD will be suspected if there is a plateau or increase in B-hCG. Therefore, making the initial correct diagnosis is crucial for the proper management of HM patients ^2,4,5. BCL-2 (B-cell lymphoma 2), encoded by the BCL-2 gene, is a member of the BCL -2 family of apoptotic regulatory proteins acting either by inhibiting (anti-apoptotic) or inducing (pro-apoptotic) apoptosis. It was the first apoptosis regulator identified in any organism ⁶. BCL-2 is proved to have a role in normal placental development and in some pregnancy disorders ^4,7. Several previous works explored the expression of BCL-2 in HMs. The data led to different conclusions and some opposing results. One early study showed significantly stronger expression of BCL-2 protein in CHMs and choriocarcinoma compared to both normal placentas and PHMs ⁸. In line with that observation, higher expression in molar pregnancies compared to non-molar ones was recently observed ⁵. Conversely, it was demonstrated that decreased BCL-2 expression might have a role in the pathogenesis of CHMs, but did not prove any association with disease progression ⁴. This decreased expression of BCL-2 was further noted in the trophoblasts of HMs, and even more in invasive mole and choriocarcinoma ³.

Tissue microArray (TMA) technology facilitates the simultaneous examination of multiple cases on the same glass slide, reducing time and cost for a given experiment ⁹. Many TMA models have been constructed in tissue-based research, including those of gynecological lesions. However, there are few attempts to use the technique in GTDs and placental tissues ^10-14.

This study aims to investigate the possible differences in the immunohistochemical expression patterns of BCL-2 in PHMs, CHMs and in normal placental tissues. In addition, we planned to construct a TMA model for these cases.

Materials and methods

SAMPLE SELECTION

In this cross-sectional study, data of female patients diagnosed with hydatidiform moles in Sultan Qaboos University Hospital and Khoula Hospital in Muscat, Oman, between 2010 and 2020 were retrieved from electronic hospital records. Data obtained from medical records included patients’ age and gestational age. A total of 237 cases of HMs (95 of PHM, 142 of CHM) and 202 cases of normal trophoblastic tissues (POCs and placentas) were collected. The archival routine hematoxylin and eosin (H&E) stained slides of all cases with corresponding tissue blocks were retrieved and the diagnosis of each case was re-checked by two independent pathologists. The criteria for the diagnosis of CHM and PHM were as described ¹. All the CHM cases were confirmed routinely at the time of the initial diagnosis by the absence of p57 immuno-reactivity in cytotrophoblasts and in villous stromal cells. For PHM, and due to the unavailability short tandem repeat DNA genotyping in our laboratories, all the revised PHM cases where those fulfilling all diagnostic morphological features (at least; the presence of dual “large distended and small fibrotic” villous population, irregular outlines of the distension chorionic villi, cistern formation, presence of pseudoinclusions and mild circumferential trophoblastic hyperplasia) and retained p57 staining. The study was approved by the research ethics committees of Sultan Qaboos University Hospital (SQU-EC/218/19) and the Ministry of Health (PRO0102019045).

TMA CONSTRUCTION

Two representative areas reflecting the diagnosis for each case were identified on archival slides and marked on the corresponding tissue blocks. The TMA construction was performed using tissue cores of 3 mm diameter, according to the manufacturers’ instructions (T-Sue array molds, Simport, Canada). Each TMA block contained 17 different cases in duplicates, 2 cores per case. A total of 14 TMA blocks of hydatidiform mole cases (PHM and CHM) and 12 TMA blocks of normal trophoblastic tissues (POCs and placentas) were prepared. Control tissues includes spots of normal kidney, liver, lung, intestine, tonsil, and lymph node.

IMMUNOHISTOCHEMISTRY

The TMA blocks were sectioned and immunohistochemically stained using antibodies against BCL-2 (Dako, Glostrup, Denmark), according to the manufacturer instructions. The expression of the marker was assessed in trophoblasts (cytotrophoblasts, syncytiotrophoblasts and non-villous trophoblasts) and in villous stromal cells independently. BCL-2 expression was analyzed semi-quantitatively as previously reported ^3,5. Evaluation of the staining intensity was scored as follows: score 0 “no staining”, score 1 “weak staining”, score 2 “moderate staining” and score 3 “strong staining”. Assessment of the percentage of the positive cells was scored as follows: score 0 “no stained cells”, score 1 “< 30% stained cells”, score 2 “30-60% stained cells” and score 3 “> 60% stained cells”. The sum of both the intensity and percentage scores (the overall score) for each cell type (trophoblasts and stromal cells) was obtained. The tissue cores of lymph nodes were chosen as positive controls for staining.

STATISTICAL ANALYSIS

The information collected was entered using Epi-data program and then transferred to SPSS-25 for analysis. Associations between different categorical variables were assessed by Chi Square test. In addition, the sum of the both the percentage and intensity scores for each type of cells was obtained.

Results

DEMOGRAPHIC FEATURES

The age of patients with HMs moles ranged from 16 years to 54 years. The mean age of patients with CHM was 30 ± 7.9 years with median of 29.0, minimum of 16 and maximum of 54. In contrast, the mean age of the patients with PHM was 32 ± 6.8 years with median of 33.0, minimum of 18 and maximum of 49. A statistically significant relation was identified between the age of the patient and the type of molar diseases (p-value 0.021). CHM was diagnosed in 88.9% of patients below the age of 20 years, while 65.7% of those who were above the age of 30 years were diagnosed with PHM. Likewise, the gestational age showed a statistically significant relationship with the type of molar diseases (p-value 0.012). The patients’ gestational age of pregnancy ranged from 4 to 25 weeks. CHM was diagnosed at earlier gestational age with mean of (10.20) weeks compared to (11.25) weeks in PHM. There was no significant relation between the number parity and the type of molar pregnancy.

DECREASED TROPHOBLASTIC EXPRESSION OF BCL-2 FROM POC AND PHM TO CHM

BCL-2 was expressed in the cytoplasm of almost all trophoblasts (> 95%), namely villous types, in hydatidiform moles and control groups. BCL-2 staining showed a significant reduction in its intensity from strong staining in 73.7% of normal trophoblasts and 76.3% of PHM cases to only 26.9% of cases of CHM. Among CHMs, moderate trophoblastic cytoplasmic staining was the predominant pattern, present in 67.7% of cases. There was a significant difference between PHM and CHM in the intensity of the staining score (p- value 0.0005) and the overall total score (p- value 0.0005) (Fig. 1a) but not the percentage score (p-value > 0.05). The overall score in the trophoblasts showed predominance of score 6 in both the control trophoblasts and those of PHMs, while most (66.7%) CHM cases had a score of 5 (Fig. 1a). BCL-2 expression was noted in the villous stromal cells of the different categories as weak staining. However, there was no significant relation between the overall villous stromal cells staining scoring between the different diagnostic entities (p-value > 0.05) (Fig. 1b). Examples of the different staining patterns are illustrated in Figure 2.

TMA REPRESENTED THE LESION HETEROGENEITY

The two TMA cores (each 3 mm in diameter) were retained in all slides, during sectioning, in 91.6% of cases, while 4.6% of cases had one core only. Both cores were lost in 3.8% of cases. In 78% of the retained cases, more than half of the arrayed area was occupied by the target tissue including the chorionic villi with their different components (villous cytotrophoblasts, syncytiotrophoblasts and stromal cells) and non-villous intermediate trophoblasts (Fig. 3a). On the other hand, less than half of the area of the core showed the target tissue in 18% of cases (Fig. 3b). Only 4% of the cores showed a complete loss of tissue components (Fig. 3c). In 49.8% of cores, both the chorionic villi and non-villous trophoblasts were present in the same TMA spot (Fig. 3d), whereas 48.6% of the cores showed only villi (Fig. 3e), 1.6% of the cores showed only extravillous trophoblasts (Fig. 3f).

Discussion

The histopathological criteria of HMs are well described and the diagnosis can be made with confidence in the majority of patients ¹. However, some cases continue to be problematic in arriving at a definite diagnosis, especially in first trimester abortions. With the use of p57 immunohistochemistry, many of these obstacles were resolved to make a diagnosis of CHM ². On the other hand, a precise histological distinction of PHMs is based largely on morphological features in addition to DNA genotyping ¹. Therefore, we investigated the immunohistochemical expression patterns of BCL-2 in molar pregnancies with a special attention to a possible diagnostic value, especially for PHMs. In addition, a TMA using duplicates, each of 3 mm diameter from each case, was constructed aiming at providing a tissue based model for current and future studies. Regarding the patients’ demographic data, we observed significant differences in the gestational age of patients at the time of diagnosis between PHM and CHM, as CHM was discovered during earlier weeks of pregnancy. This finding is inconsistent with the data of Moussa and colleagues, who, also in contrast with our data, found CHMs to affect older patients compared to PHMs ¹⁵.

BCL-2, discovered decades ago, is an anti-apoptotic protein that regulates cell death and survival without directly affecting cell proliferation ¹⁶. Our data revealed that the overall total IHC score (intensity + percentage) of BCL-2 expression in the trophoblasts of CHM was significantly lower than that of PHM and control specimens (unremarkable placentas and normal products of conception). This finding is in line with the previous works ^3,4,17-19. However, these data are in contrast to another study that found the inverse; increased BCL-2 expression in CHM ⁵. In one earlier study, there was prominent BCL-2 expression in molar trophoblasts suggesting its role in the development of CHM. However, the intensity was moderate compared to the more intense staining pattern observed in the first trimester chorionic villi ⁷. Most of the aforementioned works reported the positivity of BCL-2 in syncytiotrophoblasts only. In addition, we found the cytotrophoblasts and the non-villous intermediate trophoblasts to express the marker as well, despite being to a less extent compared to syncytiotrophoblasts ⁷. Moreover, the villous stromal cells in the current study showed an overall weak staining in some samples. These expression patterns could be explained by matters related to fixation, processing, antibody used for IHC and antigen retrieval protocols. The decreased expression of BCL-2 in CHM compared to PHM and normal control tissues indicates an increase in apoptosis and favors the excessive proliferation of trophoblastic cells that characterize CHM ^3,4. Some studies identified that, beside its anti-apoptotic role, BCL-2 can acts as anti-proliferative protein by preventing the cell cycle transmission to S-phase. This indicates that, in CHM, decrease BCL-2 expression leads indirectly to trophoblastic proliferation as they become responsive to mitotic stimuli ^4,7.

The current data help shed light on the possible diagnostic role of BCL-2 in molar pregnancies to provide more evidence for the diagnosis of CHM. However, with the great success of p57 to make this distinction, the potential role of BCL-2 could be reserved only for supporting the diagnosis in a few cases, for example those rare cases with discordant p57 expression ^19,20.

TMA technology is being heavily used for tissue based research, especially for neoplasms, including gynecological malignancies and pre-neoplastic lesions ^9,21.

In contrast, a few studies utilized the technique for placental and gestational diseases ^3,4,17-19. The initial step in constructing a successful TMA is to understand the nature of the lesion and to identify the target tissue appropriately. The heterogeneity of placental tissues and, even more with HMs, add more challenges for the proper construction of TMA ^13,22. There are different approaches for constructing TMAs for placentas and HMs. For a placental TMA, using three cores, each of 0.6 mm diameter, provided adequate representation of the trophoblastic components, especially villous trophoblasts ¹¹. Compared to placenta, HMs exhibit more complex features including villous edema, trophoblastic hyperplasia, hemorrhage, and more widely spaced tissue components with empty spaces in-between. These specific features make it more challenging to construct an optimum TMA for HMs ¹³. In order to represent the different components, such as villous trophoblasts, villous stromal cells and implantation site, which may be apart from each other in many cases, we have chosen cores of wider diameters (3 mm) in duplicates. The yield of this model assure the presence of the variable target tissues of HMs, especially those cases with complex natures. This model can save the use of reagents by 17 times, as 17 cases can be arrayed in duplicates in one TMA block. The model is kept to continue serving future studies.

In conclusion, BCL-2 expression in lower in CHM compared to PHM and non-molar tissues, a finding reflecting increased apoptosis and enhanced proliferation of the trophoblasts. Construction of TMA with wider core diameter can overcome heterogeneity of complex tissues.

CONFLICTS OF INTEREST

The authors declare no conflict of interest.

FUNDING

The study was funded by a grant of the research council of Oman to OMSB residents.

ETHICAL CONSIDERATION

The study was approved by the research ethics committees of Sultan Qaboos University Hospital (SQU-EC/218/19) and the Ministry of Health (PRO0102019045) in Oman.

AUTHORS’ CONTRIBUTIONS

MAJ: collecting material, data, performing experiment and writing manuscript. SAB: TMA construction. HAK: collecting material and revising diagnosis. MA: creating the idea, revising the diagnosis, analysing IHC, writing and revising the manuscript. All the authors reviewed and approved the final manuscript.

Figures and tables

Figure 1.BCL2 immunohistochemistry combined score (intensity and percentage) in the trophoblasts (A) villous stromal cells (B) of control tissues, partial (PHM) and complete (CHM) hydatidiform moles.

Figure 2.Examples of the variable intensity of the IHC expression of BCL-2 in the normal trophoblasts (strong) (A), trophoblasts of PHM (moderate to strong) (B), and CHM (weak) (C). Positive staining of villous stromal cells is noted, asterisk (A). (Immunoperoxidase, 200x)

Figure 3.Examples for cases having more (A) or less (B) than half of the areas of cores occupied by the target tissues (chorionic villi with or without non-villous intermediate trophoblasts). Cores showing complete absence of the target tissue components, presented only with decidua or blood (C). Cases having both chorionic villi and non-villous intermediate trophoblasts in the same TMA spot (D). Cores with only chorionic villi (E) and only non-villous trophoblasts (F). (H&E 40x)

References

1. Cheung AN, Hui P, Shih I. Female genital tumours. 5th Edition. WHO Classification of Tumours. IARC Press: Lyon; 2020.
2. Khashaba M, Arafa M, Elsalkh E. Morphological Features and Immunohistochemical Expression of p57Kip2 in Early Molar Pregnancies and Their Relations to the Progression to Persistent Trophoblastic Disease. J Pathol Transl Med. 2017; 51:381-387. DOI
3. Wargasetia TL, Shahib N, Martaadisoebrata D. Characterization of apoptosis and autophagy through Bcl-2 and Beclin-1 immunoexpression in gestational trophoblastic disease. Iran J Reprod Med. 2015; 13:413.
4. Rath G, Soni S, Prasad CP. Bcl-2 and p53 expressions in Indian women with complete hydatidiform mole. Singapore Med J. 2011; 52:502-507.
5. Missaoui N, Landolsi H, Mestiri S. Immunohistochemical analysis of c-erbB-2, Bcl-2, p53, p21WAF1/Cip1, p63 and Ki-67 expression in hydatidiform moles. Pathol Res Pract. 2019; 215:446-452.
6. Hardwick JM, Soane L. Multiple functions of BCL-2 family proteins. Cold Spring Harb Perspect Biol. 2013; 5:a008722.
7. Hussein MR. Analysis of p53, BCL-2 and epidermal growth factor receptor protein expression in the partial and complete hydatidiform moles. Exp Mol Pathol. 2009; 87:63-69. DOI | PubMed
8. Fulop V, Mok SC, Genest DR. c-myc, c-erbB-2, c-fms and bcl-2 oncoproteins. Expression in normal placenta, partial and complete mole, and choriocarcinoma. J Reprod Med. 1998; 43:101-110.
9. Arafa M, Boniver J, Delvenne P. Progression model tissue microarray (TMA) for the study of uterine carcinomas. Dis Markers. 2010; 28:267-272. DOI
10. Richani K, Romero R, Kim YM. Tissue microarray: an effective high-throughput method to study the placenta for clinical and research purposes. J Matern Fetal Neonatal Med. 2006; 19:509-515. DOI | PubMed
11. Zhang Z, Zhang L, Yang X. Construction and validation of a placental tissue microarray from specimens of well-documented preeclampsia patients. Placenta. 2013; 34:187-192. DOI | PubMed
12. Abbas RK, Al-Khafaji KR. Expression of P57 immunohistochemical marker in complete and partial hydatidiform mole by using tissue microarray technique. IOSR Journal of Applied Chemistry (IOSR-JAC). 2014; 7:90-95.
13. King JR, Wilson ML, Hetey S. Dysregulation of placental functions and immune pathways in complete hydatidiform moles. Int J Mol Sci. 2019; 20:4999. DOI
14. López CL, Figueira Gouvêa AL, Rodrigues FR. Human epidermal growth factor receptor 2 fluorescence in situ hybridization and P57KIP2 immunohistochemistry using tissue microarray: improving histopathological subtyping of hydatidiform mole. Placenta. 2020; 99:166-172. DOI | PubMed
15. Moussa RA, Eesa AN, Abdallah ZF. Diagnostic Utility of Twist1, Ki-67, and E-Cadherin in Diagnosing Molar Gestations and Hydropic Abortions. Am J Clin Pathol. 2018; 149:442-455. DOI
16. Marzioni D, Mühlhauser J, Crescimanno C. BCL-2 expression in the human placenta and its correlation with fibrin deposits. Hum Reprod. 1998; 13:1717-1722. DOI | PubMed
17. Kim Y T, Huh J R, Kim J H. The Expression of bcl-2 protein in Gestational Trophoblastic Disease and Chorionic Villi. Obstet Gynecol Sci. 1998; 41:2386-2393.
18. Candelier JJ, Frappart L, Yadaden T. Altered p16 and Bcl-2 expression reflects pathologic development in hydatidiform moles and choriocarcinoma. Pathol Oncol Res. 2013; 19:217-227. DOI
19. Erol O, Suren D, Tutus B. Comparison of p57, c-erbB-2, CD117, and Bcl-2 expression in the differential diagnosis of hydatidiform mole and hydropic abortion. Eur J Gynaecol Oncol. 2016; 37:522-529. PubMed
20. Murphy KM, Carrick K, Gwin K. Rare Complete Hydatidiform Mole With p57 Expression in Villous Mesenchyme: Case Report and Review of Discordant p57 Expression in Hydatidiform Moles. Int J Gynecol Pathol. 2022; 41:45-50. DOI
21. Visser NCM, van der Wurff AAM, Pijnenborg JMA. Tissue microarray is suitable for scientific biomarkers studies in endometrial cancer. Virchows Arch. 2018; 472:407-413. DOI
22. Horn LC, Purz S, Leo C, Stepan H. Application of tissue microarrays in placental research. Ann Diagn Pathol. 2008; 12:48-49. DOI