Acid sphingomyelinase involvement in tumor necrosis factor alpha-regulated vascular and steroid disruption during luteolysis in vivo

Abstract

TNF is well known for its role in inflammation, including direct effects on the vasculature. TNF also is implicated in the regulation of reproduction by its actions to affect ovarian steroidogenic cells and to induce apoptosis of corpus luteum (CL)-derived endothelial cells in vitro. We hypothesized that the disruption of TNF signaling would postpone the regression of the highly vascularized CL in vivo, and this effect could be replicated in mutant mouse models lacking TNF receptor (TNFRI(-/-)) and/or a critical enzyme of TNF signaling, acid sphingomyelinase (ASMase(-/-)). In the current study, the treatment of pseudopregnant mice with the luteolytic mediator prostaglandin F2-alpha (PGF) significantly increased TNF in the ovaries when compared with saline-treated controls. Treatment with PGF also reduced serum progesterone (P4) concentrations and caused involution of the CL. However, pretreatment of pseudopregnant mice with Etanercept (ETA), a TNF-neutralizing antibody, inhibited the PGF-induced decrease in P4 and delayed luteal regression. A similar outcome was evident in pseudopregnant TNFRI(-/-) animals. Treatment of luteal microvascular endothelial cells (MVECs) with TNF provoked a significant increase in ASMase activity when compared with the corresponding controls. Furthermore, TNF-induced MVEC death was inhibited in the ASMase(-/-) mice. The ASMase(-/-) mice displayed no obvious evidence of luteal regression 24 h after treatment with PGF and were resistant to the PGF-induced decrease in P4. Together these data provide evidence that TNF plays an active role in luteolysis. Further studies are required to determine the deleterious effects of anti-inflammatory agents on basic ovarian processes.

Fig. 1.

Interruption of TNF activity impairs regression of the CL. (a–d) Morphology of pseudopregnant mouse ovaries 24 h after i.p. injection of saline (a), PGF (b), TNF-neutralizing antibody (ETA) (c), or PGF plus ETA (d). (e and f) Levels of circulating progesterone 24 h after treatment with saline or PGF either in the presence of absence of ETA (e) or in control mice (C567BL/6J) or mutant mice lacking the TNF type 1 receptor gene (Tnfrsf1a^−/−) (f).

Fig. 2.

TNF induces apoptosis of mouse CL MVECs. (a) Treatment with 50 ng/ml TNF induced apoptosis of isolated MVECs, which was prevented by treatment with 15 μg/ml ETA. (b) Treatment with 50 ng/ml TNF or an inactive ceramide dihydro ceramide (50 μM DHC) did not induce apoptosis of mouse luteinized granulosa cells, whereas treatment with active ceramide analogues (50 μM C2 and 200 nM C16) or 100 milliunits/ml rASMase increased apoptosis. (c) Treatment with 0.5 μg/ml filipin to block ceramide accumulation inhibited TNF-induced apoptosis of MVECs.

Fig. 3.

Mice lacking the acid sphingomyelinase gene, Smpd1 (ASMase^−/−), are resistant to PGF-induced CL regression. (a–d) Morphology of pseudopregnant mouse ovaries 24 h after injection of saline (a) or PGF (b) in WT mice or 24 h after injection of saline (c) or PGF (d) in ASMase^−/− mice and corresponding serum progesterone levels (e) 24 h after treatment.

Fig. 4.

MVD is disrupted during CL regression. MVD was assessed by staining with CD31 antibody. (a–c) Microvessel staining of pseudopregnant mouse ovaries 24 h after injection of saline (a) or PGF (b) in ETA-pretreated WT mice or 24 h after injection of PGF alone (c) and percentage of the MVD in total area counted (d).