STUDY QUESTIONDoes experimental manipulation of fibroblast growth factor 9 (FGF9)-signalling in human fetal gonads alter sex-specific gonadal differentiation?SUMMARY ANSWERInhibition of FGFR signalling following SU5402 treatment impaired germ cell survival in both sexes and severely altered the developing somatic niche in testes, while stimulation of FGF9 signalling promoted Sertoli cell proliferation in testes and inhibited meiotic entry of germ cells in ovaries.WHAT IS KNOWN ALREADYSex-specific differentiation of bipotential gonads involves a complex signalling cascade that includes a combination of factors promoting either testicular or ovarian differentiation and inhibition of the opposing pathway. In mice, FGF9/FGFR2 signalling has been shown to promote testicular differentiation and antagonize the female developmental pathway through inhibition of WNT4.STUDY DESIGN, SIZE, DURATIONFGF signalling was manipulated in human fetal gonads in an established ex vivo culture model by treatments with recombinant FGF9 (25 ng/ml) and the tyrosine kinase inhibitor SU5402 (10 μM) that was used to inhibit FGFR signalling. Human fetal testis and ovary tissues were cultured for 14 days and effects on gonadal development and expression of cell lineage markers were determined.PARTICIPANTS/MATERIALS, SETTING, METHODSGonadal tissues from 44 male and 33 female embryos/fetuses from first trimester were used for ex vivo culture experiments. Tissues were analyzed by evaluation of histology and immunohistochemical analysis of markers for germ cells, somatic cells, proliferation and apoptosis. Culture media were collected throughout the experimental period and production of steroid hormone metabolites was analyzed in media from fetal testis cultures by liquid chromatography–tandem mass spectrometry (LC-MS/MS).MAIN RESULTS AND THE ROLE OF CHANCETreatment with SU5402 resulted in near complete loss of gonocytes (224 vs. 14 OCT4+ cells per mm2, P < 0.05) and oogonia (1456 vs. 28 OCT4+ cells per mm2, P < 0.001) in human fetal testes and ovaries, respectively. This was a result of both increased apoptosis and reduced proliferation in the germ cells. Addition of exogenous FGF9 to the culture media resulted in a reduced number of germ cells entering meiosis in fetal ovaries (102 vs. 60 γH2AX+ germ cells per mm2, P < 0.05), while in fetal testes FGF9 stimulation resulted in an increased number of Sertoli cells (2503 vs. 3872 SOX9+ cells per mm2, P < 0.05). In fetal testes, inhibition of FGFR signalling by SU5402 treatment altered seminiferous cord morphology and reduced the AMH expression as well as the number of SOX9-positive Sertoli cells (2503 vs. 1561 SOX9+ cells per mm2, P < 0.05). In interstitial cells, reduced expression of COUP-TFII and increased expression of CYP11A1 and CYP17A1 in fetal Leydig cells was observed, although there were no subsequent changes in steroidogenesis.LARGE SCALE DATAN/ALIMITATIONS, REASONS FOR CAUTION Ex vivo culture may not replicate all aspects of fetal gonadal development and function in vivo. Although the effects of FGF9 were studied in ex vivo culture experiments, there is no direct evidence that FGF9 acts in vivo during human fetal gonadogenesis. The FGFR inhibitor (SU5402) used in this study is not specific to FGFR2 but inhibits all FGF receptors and off-target effects on unrelated tyrosine kinases should be considered.WIDER IMPLICATIONS OF THE FINDINGSThe findings of this study suggest that dysregulation of FGFR-mediated signalling may affect both testicular and ovarian development, in particular impacting the fetal germ cell populations in both sexes.STUDY FUNDING/COMPETING INTEREST(S)This work was supported in part by an ESPE Research Fellowship, sponsored by Novo Nordisk A/S to A.JØ. Additional funding was obtained from the Erichsen Family Fund (A.JØ.), the Aase and Ejnar Danielsens Fund (A.JØ.), the Danish Government’s support for the EDMaRC programme (A.JU.) and a Wellcome Trust Intermediate Clinical Fellowship (R.T.M., Grant no. 098522). The Medical Research Council (MRC) Centre for Reproductive Health (R.T.M.) is supported by an MRC Centre Grant (MR/N022556/1). The authors have no conflict of interest to disclose.