Role of oxygen in the regulation of Leydig tumor derived MA-10 cell steroid production: the effect of cobalt chloride.

Abstract We have earlier shown that cobalt chloride (CoCl2)-induced hypoxia and second messenger 8-bromoadenosine 3′, 5′-cyclic adenosine monophosphate (8-Br-cAMP) stimulates vascular endothelial growth factor (VEGF) production in Leydig tumor cell derived MA-10 cells. Both stimuli follow common signal transduction pathways including protein kinase A (PK-A), extracellular regulated kinase 1/2 (ERK1/2), and phosphatidyl inositol-3 kinase/akt (PI3-K/Akt) pathways in the stimulation of VEGF by MA-10 cells. In the present study we investigated the role of CoCl2 and 8-Br-cAMP on steroid production in MA-10 cells. The MA-10 cells were cultured in Waymouth MB 752/1 medium, supplemented with 15% heat inactivated horse serum. Progesterone was estimated by radioimmunoassay (RIA).We report that 8-Br-cAMP stimulated progesterone production by the MA-10 cells whereas CoCl2 inhibited the same. Also, 8-Br-cAMP stimulated steroidogenic acute regulatory protein (StAR) and cytochrome P450 side-chain cleavage enzyme (P450scc) mRNAs expression. However, CoCl2 had no effect on StAR mRNA. Cobalt chloride directly inhibited the expression of P450scc mRNA. The decrease in progesterone production could be attributed to three different mechanisms, (1) an increase in production of reactive oxygen species (ROS), (2) an increase in HIF-1α activity, and (3) ultimately a decrease in the level of cytochrome P450 side chain cleavage (CYT P450scc). Hypoxia has an action and mechanism of action similar to that of gonadotropins on VEGF production, whereas they have a contrasting effect on steroidogenesis. This study suggests that hypoxia could be as important as gonadotropins in regulating Leydig cell steroidogenesis.


Introduction
The mammalian testis works in a low oxygen environment [Carlsson et al. 2001;Cross and Silver, 1962;Free et al. 1976;Klotz et al. 1996;Massie et al. 1969;Yu and Cringle 2006] suggesting an important role of oxygen in the regulation of testicular function including steroidogenesis and spermatogenesis. Using a Leydig tumor cell derived MA-10 cell line, we have shown that cobalt chloride (CoCl 2 ), a hypoxia inducing agent [Maxwell and Salnikow 2004], like luteinizing hormone (LH) and its second messenger cAMP, stimulates vascular endothelial growth factor (VEGF) production [Kumar et al. 2012]. Cobalt chloride stimulates VEGF production sharing common signal transduction pathways with LH/cAMP, which include protein kinase A (PKA), extracellular regulated kinase 1/2 (ERK1/2), and phosphatidyl inositol-3 kinase/Akt (PI3-K/Akt) pathways. In addition, both CoCl 2 -induced hypoxia as well as cAMP partly mediates their stimulatory effect on VEGF synthesis, through the activation of hypoxia inducible factor-1a (HIF-1a), in MA-10 cells [Kumar et al. 2012].
However, hypoxia induced by a low oxygen environment has been shown to inhibit steroidogenesis in several steroidogenic tissues from different species including human adrenal cells [Panesar et al. 2003;Raff and Bruder 2006], monkey corpus luteum cells [Tesone et al. 2005], nonluteinized monkey granulosa cells [Martinez-Chequer et al. 2003], swine granulosa cells [Basini et al. 2004], and bovine luteal cells [Nishimura et al. 2006]. Cobalt chloride-induced hypoxia also inhibited steroidogenesis in non-luteinized monkey granulosa cells [Martinez-Chequer et al. 2003]. In contrast, Hwang et al. [2007] have reported that hypoxia enhanced the stimulatory effect of human chorionic gonadotropin (hCG) on testosterone release in the mouse Leydig cell derived TM3 cell line. The present study therefore was undertaken to study the effects of CoCl 2 induced hypoxia on steroidogenesis in MA-10 cells and to identify the associated mechanism of action.
Normal Leydig cells in vivo comprise a very low percentage of total cells of the testis. Therefore, to obtain large quantities of highly enriched and viable Leydig cells requires sacrificing a large number of mice. Also, a multi-step procedure that involves filtration with nylon mesh combined with centrifugation and Percoll density gradient sedimentation yields only 95% enriched Leydig cells [Salva et al. 2001]. To avoid such problems we used the MA-10 cell line, a well studied line, derived from mouse Leydig cells. The MA-10 cells have been shown to have LH receptors and behave physiologically like Leydig cells in terms of steroid and VEGF production in response to LH/ cAMP [Anand et al. 2003;Ascoli 1981;Schwarzenbach et al. 2004]. However, MA-10 cells lack 17a hydroxylase activity so the end product in steroidogenesis is progesterone and not androgens [Ascoli 1981].

Effect of CoCl 2 on progesterone production
Time course studies of progesterone production for 48 hours were performed with hCG in combination with 3-isobutyl-1-methylxanthine (IBMX), a cyclic nucleotide phosphodiesterase inhibitor. The progesterone levels peaked at 24 hours (data not shown). Therefore subsequent experiments were performed for 24 hours. Progesterone production was stimulated 5-fold with hCG and 11.5 -fold with 8-bromoadenosine cAMP (8-Br-cAMP). Cobalt chloride decreased basal, hCG, and 8-Br-cAMP stimulated progesterone production by 15%, 77%, and 75%, respectively ( Figure 1). Cells treated with 25-hydroxy cholesterol (25-OHC) increased progesterone production in a dose (1.5 mg ml À1 to 15 mg ml À1 ) dependent manner. Doses higher than 15 mg ml À1 of 25-OHC progressively decreased cell viability and consequently progesterone production (data not shown). Therefore, 15 mg ml À1 of 25-OHC was selected as the preferred dose for subsequent experiments. Cells treated with 15 mg ml À1 of 25-OHC increased progesterone production by 3 fold. Cobalt chloride decreased the 25-OHC-induced progesterone production by 45% ( Figure 2A). Cells treated with pregnenolone resulted in a 13 fold increase in the level of progesterone. Cobalt chloride had no significant effect in regulating pregnenolone induced progesterone production ( Figure 2B).

Effects of CoCl 2 on StAR and P450scc mRNA expressions
Reverse transcription polymerase chain reaction (RT-PCR) analysis showed that the MA-10 cells have almost no expression of steroidogenic acute regulatory protein (StAR) mRNA under basal conditions. However, stimulation by 8-Br-cAMP significantly (p50.05) increased StAR mRNA expression, but CoCl 2 co-treatment failed to show any inhibitory response ( Figure 3A). In comparison, 8-Br-cAMP stimulated the mRNA levels of cytochrome P450 side chain cleavage (P450scc) enzyme by 30%. Cobalt chloride inhibited basal P450scc mRNA levels by 30% and 8-Br-cAMP stimulated P450scc levels by 28% ( Figure 3B).

Effect of inhibitors on progesterone productions
Cyclosporin (CSA) A hypoxia inducible factor-1a (HIF-1a) inhibitor alone did not produce any significant change either in basal progesterone production or hCG stimulated production of progesterone. However, in the presence of CoCl 2 , it partially reversed the inhibitory action of CoCl 2 on basal and hCG stimulated progesterone production ( Figure 4A). H-89 (protein kinase A (PKA) inhibitor), U0126 (MEK1/2 inhibitor), and LY294002 (PI3-K inhibitor) significantly decreased 8-Br-cAMP stimulated progesterone production by 60%, 72%, and 50%, respectively ( Figure 4B). These inhibitors had no significant effect on the basal and CoCl 2 treated progesterone production ( Figure 4C).

Effect of CoCl 2 on ROS and nitrogen reactive species production
Cobalt chloride treatment of MA-10 cells induced a rise in reactive oxygen species (ROS) by 44%. In comparison, 8-Br-cAMP did not alter the basal levels of ROS in these cells but supported the CoCl 2 stimulated rise ( Figure 5A). Pyrrolidine dithiocarbama (PDTC), a potent antioxidant agent decreased ROS production under basal, 8-Br-cAMP, and CoCl 2 induced conditions ( Figure 6A). PDTC had no significant effect on basal and hCG stimulated progesterone production in MA-10 cells. However, PDTC partially reversed the action of CoCl 2 on basal and hCG induced progesterone production ( Figure 6B). One method for indirect determination of NO involves the spectrophotometric measurement of its stable decomposition products, nitrate and nitrite radicals. MA-10 cells were treated with S-nitroso-N-acetylpenicilamine (SNAP, 0.01 mM), CoCl 2 , or 8-Br-cAMP for 24 hours. There was a robust increase in nitrite as well as nitrate in response to the NO releasing agent SNAP. However, both CoCl2 and 8-Br-cAMP Upper inset panel is a representative agarose gel picture. A, B, C, and D represent basal, CoCl 2 , 8-Br-cAMP, and CoCl 2 þ 8-Br -cAMP induced StAR mRNA expression in relation to internal control (L19). Bars represent mean AE SEM of ratios of the intensity of StAR mRNA bands with corresponding L19 bands from three different gels (n ¼ 3). *p50.05 between A and C; B and D. Panel (B) mRNA expression of P450 scc upon treatment with 8-Br-cAMP and CoCl 2 . Upper inset panel is a representative agarose gel picture. A, B, C, and D represent basal, CoCl 2 , 8-Br-cAMP, and CoCl 2 þ 8-Br -cAMP induced P450 scc mRNA expression in relation to internal control (S15). Bars represent mean AE SEM of ratios of the intensity of P450 scc mRNA bands with corresponding S15 bands from three different gels (n ¼ 3). p50.05 between A and B; C and D. failed to elicit any change in the production of nitrogen reactive species by MA-10 cells ( Figure 5B).

Discussion
In our earlier study we showed that both CoCl 2 and 8-Br-cAMP stimulate VEGF production in Leydig cell derived MA-10 tumor cell line via cAMP downstream pathways that include PK-A, mitogen-activated protein kinases (MAPK)/extracellular regulated kinase (ERK1/2), and PI3K/Akt pathways [Kumar et al. 2012]. In contrast, the current study reports that both CoCl 2 and 8-Br-cAMP have divergent actions on progesterone production. Cobalt chloride-induced hypoxia decreased basal hCG-and 8-Br-cAMP stimulated progesterone production. Inhibitors of the PK-A, MAPK/ERK1/2, and PI3K/ Akt pathways decreased 8-Br-cAMP stimulated progesterone production, but they had no effect on CoCl 2 induced inhibition of progesterone production. These observations suggest that CoCl 2 affects steroid synthesis by inhibiting a step(s) distal to the target of these inhibitors. The rate limiting step in steroidogenesis is the translocation of cholesterol from the outer mitochondrial membrane to the inner mitochondrial membrane by the StAR protein. However, 25-OHC, a lipid soluble side-chain oxygenated sterol, readily diffuses across mitochondrial membranes [Jefcoate et al. 1974;Meaney et al. 2002]. Presumably, the diffusion of polar cholesterol is independent of the StAR protein; however, 25-OHC has also been shown to stimulate the synthesis of the StAR protein to further promote steroidogenesis in MA10 cells [King et al. 2004]. In our studies, 25-OHC stimulated progesterone production by 3 fold in MA-10 cells. This stimulation of progesterone synthesis was inhibited by CoCl 2 .
Exogenous pregnenolone also stimulated progesterone production in the MA-10 cells. Stimulation of progesterone production was not abolished by CoCl 2 suggesting that CoCl 2 has possibly no effect on 3b-hydroxy steroid dehydrogenase (3b-HSD activity). Therefore, it seems that CoCl 2 probably inhibits progesterone production at a site proximal to the step catalyzed by 3b-HSD, namely either the StAR protein or cytochrome P450 side chain cleavage (CYT P450 scc ). The expression of StAR mRNA in the MA-10 cells was stimulated by 8-Br-cAMP. But CoCl 2 did not alter StAR mRNA expression in response to 8-Br-cAMP. The data suggests that the inhibitory action of CoCl 2 did not involve StAR mRNA.
Treatment with CoCl 2 resulted in a significant decrease in P450scc mRNA expression in both basal and 8-Br-cAMP treated cells. Nishimura et al. [2006] also demonstrated in bovine luteal cells a decline in the levels of P450scc mRNA in  response to hypoxia. This group, too, did not find any effect on StAR mRNA/protein, and 3b-HSD mRNA expression/ activity. The presence of cobalt along with heme leads to the formation of cobalt protophyrin owing to the competition of cobalt with ferrochelatase [Sinclair et al. 1979]. Sinclair et al. [1982] suggested that the formation of the cobalt protophyrin occurs at the expense of heme, leading to a decrease in CYT P450. Cobalt has also been shown to substitute for ferrous iron in the center of the porphyrin ring of heme of hemoglobin. The resultant cobalt hemoglobin has lower affinity to bind with oxygen [Yonetani et al. 1974]. These observations suggest that cobalt can also inactivate the presynthesized heme.
We showed that CoCl 2 -induced hypoxia and 8-Br-cAMP increased HIF-1a protein activity in MA-10 cells without any change in its mRNA levels. Also, the rise in HIF-1a protein activity was inhibited by CsA, a stimulant of prolyl hydroxylase activity [Kumar et al. 2012]. Similarly, D'Angelo et al. [2003] showed that CsA destabilized HIF-1a induced by hypoxia in C6 glioma cells. In our present study we report that CsA partially reversed the inhibitory action of CoCl 2 on basal and hCG stimulated progesterone production, suggesting a possible role of HIF-1a in the regulation of progesterone production by CoCl 2 .
Primary Leydig cells treated with H 2 O 2 resulted in a decrease in StAR and 3b-HSD protein levels [Allen et al. 2004]. Similarly, Diemer et al. [2003] showed that H 2 O 2 decreased progesterone production in MA-10 cells. The inhibition was due to a decrease in StAR protein levels without any change in their mRNA levels or P450 scc protein levels. In the present study, CoCl 2 induced hypoxia increased ROS levels which could subsequently lead to a reduction in steroid production. PDTC, a potent anti-oxidant, has been shown to enhance the degradation of ROS such as H 2 O 2 [Schreck et al. 1991]. It decreased ROS production under hypoxic conditions. Here, use of PDTC partially abolishes the inhibitory action of CoCl 2 on progesterone production suggesting that the inhibitory action of CoCl 2 is partly mediated by ROS. However, ROS induced inhibition of steroid production could be camouflaged because of direct inhibition of CYT P450scc by CoCl 2. Nitric oxide, a highly reactive inorganic free radical gas has been shown to inhibit steroidodgenesis in granulose -luteal cells and MA-10 cells [Punta et al. 1996;Van Vooris et al. 1994]. In the present study, there was no change in nitrogen species in response to CoCl 2 or 8-Br-cAMP.
The low level of oxygen in the testes seems to be as important as gonadotrophins in the regulation of Leydig cell function. Hypoxia has an action and mechanism of action similar to that of gonadotropins on VEGF production, whereas they have divergent effect on steroidogenesis. This inhibition of steroidogenesis by hypoxia seems to be necessary to regulate spermatogenesis and other androgen dependent effects of gonadotropins. In conclusion, hypoxia, particularly induced by CoCl 2 , lowers progesterone production in Leydig cell-derived MA-10 cells. This decrease in progesterone production could be attributed to three different mechanisms: (1) an increase in ROS production, (2) an increase in HIF-1a activity, and (3) ultimately a decrease in the level of CYT P450scc.

Cell culture
MA-10 cells were cultured in T-75 flask in Waymouth MB 752/1 medium, supplemented with 15% heat inactivated horse serum. Cells were incubated at 37 C in 5% CO 2 in a humidified chamber and grown until 75% confluency. The cells were trypsinised using a 0.25% trypsin-EDTA solution.
A total of 10 5 cells ml À1 per well were plated in medium containing serum in 24/48-well plates for 17 h. Thereafter, the medium was replaced with fresh serum-free medium (SFM). The cells were pre-incubated without or with 100 mM CoCl 2 for 6 h. The dose of CoCl 2 was selected based on previous studies [Kumar et al. 2012]. Cells were further incubated with one of the following reagents: 1 mM 8-Br-cAMP (a cAMP analogue), 0.5 IU hCG þ 0.5 mM IBMX, 25-OHC and 20 mM pregnenolone.
In experiments with CsA, cells were pre-treated with CsA for 2 h, before incubation without or with 100 mM CoCl 2 . After the specified incubation period, serum-free medium (SFM) was removed and stored at À20 C until assayed for progesterone. All the treatments were done at least in three replicates. Each experiment was performed twice or thrice to check the reproducibility.

Progesterone production
Progesterone was estimated by Radioimmunoassay (RIA, Panesar et al. 2003). Progesterone concentration in each sample was measured in duplicate. The antibody used for RIA was highly specific and showed no cross reactivity for pregnenolone as well as testosterone.

Semi quantitative reverse transcriptase PCR
Total RNA from MA-10 cells was isolated according to the manufacturer's instructions using RNeasy mini kit from Qiagen (Hilden, Germany). RNA (2 mg) from each isolate was reverse transcribed in oligo (dT)-primed reverse transcription reactions with reverse transcriptase RETROSCRIPT kit provided by Ambion according to the manufacturer's instructions. A 1 ml aliquot of the reverse-transcribed cDNAs was then used to amplify mRNA sequences coding for StAR and P450scc. L19 and S15 were taken as internal controls. The sequence of the primers used is provided in Table 1. The polymerase chain reaction was carried out for 31 cycles with annealing temperatures 60 C and 58.5 C for StAR and P450scc, respectively. The amplified fragments were sizefractionated in a 1.5% (wt/v) agarose gel and visualized by staining with ethidium bromide (0.2 ng/ml). The amplified bands were quantified by Life Science software from UVP (Upland, CA, USA).

Determination of ROS and nitrogen reactive species
A total of 2 Â 10 5 cells ml À1 per well were plated in medium containing serum in 24-well plates for 17 h. Thereafter, the medium was replaced with fresh SFM. The cells were preincubated without or with pyrrolidine dithiocarbamate (PDTC) for 1.5 h. Fluorescence intensity was determined with the help of Hitachi F 4500 spectrofluoremeter with excitation wavelength 488 nm and emission wavelength 522 nm. The values were expressed in terms of arbitrary units (AU). Nitrogen reactive species were determined with the Total NO determination kit obtained from R&D systems (Minneapolis, USA).