Sertoli cell androgen receptor signalling in adulthood is essential for post‐meiotic germ cell development

Androgens are key drivers of spermatogenesis, and germ cells in mice lacking androgen receptor (AR), specifically from Sertoli cells, arrest in meiosis (reviewed in Smith and Walker, 2014). When Sertoli-cell AR is ablated during fetal life (De Gendt and Verhoeven, 2012), however, it is impossible to determine whether the meiotic-arrest phenotype observed in adults results from perturbed Sertoli cell development or perturbed function in adulthood. We used a lentiviral approach to determine if Sertoli-cell AR is essential for supporting spermatogenesis specifically in adult testes an organ where tamoxifen-inducible knockout may present off-target effects. Specifically, we introduced Cre recombinase into the Sertoli cells of adult male AR mice (De Gendt et al., 2004) to generate adult Sertoli-Cell AR Knockout (aSCARKO) mice. Lentiviral particles contained both CMV-Cre recombinase and tRFP635 (red fluorescent protein) transgenes separated by an IRES, or CMV-tRFP635 alone. Shuttle vectors were packaged using a third-generation lentiviral vector pseudotyped for VSV-G, produced at a viral titer of >1 10. Virus was introduced into the seminiferous tubules of adult male AR via injection into the efferent ducts, using 10ml of Cre virus, tRFP635 control virus, or optiMEM (vehicle); an additional shamoperated, but not injected, control was also evaluted. To control for systemic effects, combinations of Cre/control Cre/optiMEM, Cre/sham, control/optiMEM, control/sham, were generated in testes from individual mice (one treatment per testis; n1⁄4 10 per group). Tissueswerecollected40daysaftersurgery(onecomplete cycleofspermatogenesis).Bodyandseminalvesicleweight(a biomarkerofcirculatingandrogenconcentrations)didnotdiffer betweenanytreatmentgroup(datanotshown),buttheweightof testes injected with Cre recombinase virus was significantly reduced (sham, 92.52 4.29; OptiMEM, 100.12 4.96; tRFP635 control, 70.22 17.31; Cre, 35.48 3.45mg) to a finalweightconsistentwithdevelopmental-SCARKOmice(De Gendt et al., 2004). tRFP635 was specifically detected in the cytoplasmofSertoli cells (Fig. 1A,B), but not in other testicular cell types. AR expression was observed in all somatic cells in both sham-operated and control tRFP635 lentivirus-injected testes.Incontrast,injectionoftesteswithCrerecombinasevirus resultedinSertolicell-specific localisationoftRFPwithalossof AR expression only in Sertoli cells; AR was retained in other testicular somatic cell types. Thus, AR had been selectively ablated in adult Sertoli cellswhilst leaving the remainder of the testis untouched. Forty days post-injection, seminiferous tubules from control testes retainednormalspermatogenesis,withnoobvious defects (Fig. 1A,C). Testes injected with Cre recombinase virus,however,displayedevidenceofgerm-cell arrestduring meiosis (Fig. 1B,D), similar to that observed in SCARKO mice (De Gendt et al., 2004). Furthermore, epididymides continuous with the Cre-encoding virus-injected testes contained nomature spermatozoa (Fig. 1F). Therefore, loss of Sert9oli-cell AR in adulthood recapitulates the spermatogenic-blockphenotypeobserved indevelopmentally induced SCARKO models, unequivocally demonstrating that Sertoli cellARisessential for continuousspermatogenesis inadults.