Distinct microRNA profiles for complete hydatidiform moles at risk of malignant progression.

BACKGROUND: MicroRNAs are small noncoding RNAs with important regulatory functions. Although well-studied in cancer, little is known about the role of microRNAs in premalignant disease. Complete hydatidiform moles are benign forms of gestational trophoblastic disease that progress to gestational trophoblastic neoplasia in up to 20% of cases; however, there is no well-established biomarker that can predict the development of gestational trophoblastic neoplasia. OBJECTIVE: This study aimed to investigate possible differences in microRNA expression between complete moles progressing to gestational trophoblastic neoplasia and those regressing after surgical evacuation. STUDY DESIGN: Total RNA was extracted from fresh frozen tissues from 39 complete moles collected at the time of uterine evacuation in Brazil. In the study, 39 cases achieved human chorionic gonadotropin normalization without further therapy, and 9 cases developed gestational trophoblastic neoplasia requiring chemotherapy. Total RNA was also extracted from 2 choriocarcinoma cell lines, JEG-3 and JAR, and an immortalized normal placenta cell line, 3A-subE. MicroRNA expression in all samples was quantified using microRNA sequencing. Hits from the sequencing data were validated using a quantitative probe-based assay. Significantly altered microRNAs were then subjected to target prediction and gene ontology analyses to search for alterations in key signaling pathways. Expression of potential microRNA targets was assessed by quantitative real-time polymerase chain reaction and western blot. Finally, potential prognostic protein biomarkers were validated in an independent set of formalin-fixed paraffin-embedded patient samples from the United States (15 complete moles progressing to gestational trophoblastic neoplasia and 12 that spontaneously regressed) using quantitative immunohistochemistry. RESULTS: In total, 462 microRNAs were identified in all samples at a threshold of <1 tag per million. MicroRNA sequencing revealed a distinct set of microRNAs associated with gestational trophoblastic neoplasia. Gene ontology analysis of the most altered transcripts showed that the leading pathway was related to response to ischemia (P<.001). Here, 2 of the top 3 most significantly altered microRNAs were mir-181b-5p (1.65-fold; adjusted P=.014) and mir-181d-5p (1.85-fold; adjusted P=.014), both of which have been shown to regulate expression of BCL2. By quantitative real-time polymerase chain reaction, BCL2 messenger RNA expression was significantly lower in the complete moles progressing to gestational trophoblastic neoplasia than the regressing complete moles (-4.69-fold; P=.018). Reduced expression of BCL2 was confirmed in tissue samples by western blot. Immunohistochemistry in the independent patient samples revealed significantly lower cytoplasmic expression of BCL2 in the villous trophoblasts from cases destined for progression to gestational trophoblastic neoplasia compared with those that regressed, both with respect to staining intensity (optic density 0.110±0.102 vs 0.212±0.036; P<.001) and to the percentage of positive cells (16%±28% vs 49.4%±28.05%; P=.003). CONCLUSION: Complete moles progressing to gestational trophoblastic neoplasia are associated with a distinct microRNA profile. miR-181 family members and BCL2 may be prognostic biomarkers for predicting gestational trophoblastic neoplasia risk.