Involvement of CREB-regulated transcription coactivators (CRTC) in transcriptional activation of steroidogenic acute regulatory protein (Star) by ACTH

Studies in vivo have suggested the involvement of CREB-regulated transcription coactivator (CRTC)2 on ACTH-induced transcription of the key steroidogenic protein, Steroidogenic Acute Regulatory (StAR). The present study uses two ACTH-responsive adrenocortical cell lines, to examine the role of CRTC on Star transcription. Here we show that ACTH-induced Star primary transcript, or heteronuclear RNA (hnRNA), parallels rapid increases in nuclear levels of the 3 isoforms of CRTC; CRTC1, CRTC2 and CRTC3. Furthermore, ACTH promotes recruitment of CRTC2 and CRTC3 by the Star promoter and siRNA knockdown of either CRTC3 or CRTC2 attenuates the increases in ACTH-induced Star hnRNA. Using pharmacological inhibitors of PKA, MAP kinase and calcineurin, we show that the effects of ACTH on Star transcription and CRTC nuclear translocation depend predominantly on the PKA pathway. The data provides evidence that CRTC2 and CRTC3, contribute to activation of Star transcription by ACTH, and that PKA/CRTC-dependent pathways are part of the multifactorial mechanisms regulating Star transcription.


Introduction
Glucocorticoid hormone release from the adrenal gland is essential for maintaining normal metabolic function and survival during severe stress (Cole et al., 1995;Whitehead et al., 2013). In basal (unstressed) conditions, glucocorticoids are secreted rhythmically, with both circadian (daily) and ultradian (pulsatile) variations (Park et al., 2013;Weitzman et al., 1971;Tapp et al., 1984). The episodic nature of glucocorticoid secretion is critical for homeostasis, and its alteration can result in changes in tissue steroid responsiveness, resulting in neuroendocrine, behavioural and metabolic dysfunction (Deuschle et al., 1997;van den Berg et al., 1995;Qian et al., 2012). Glucocorticoid circadian and ultradian rhythmicity follow preceding changes in circulating ACTH. As with other steroid hormones, stimulation of glucocorticoid secretion by ACTH requires de novo synthesis. These ACTHdependent glucocorticoid secretory episodes in vivo, are associated with rapid transcription of genes encoding steroidogenic proteins in the zona fasciculata of the adrenal cortex (Sewer and Waterman, 2002;Spiga et al., 2011a,b;Liu et al., 2013;Spiga et al., 2017), including the ratelimiting steroidogenic acute regulatory (StAR) protein (Stocco and Clark, 1996). Although initiation of steroidogenesis following ACTH stimulation depends on rapid post-translational modifications of StAR (Arakane et al., 1997), this protein is short-lived (Artemenko et al., 2001) and transcriptional episodes are essential for maintaining adequate mRNA and protein levels for subsequent secretory episodes (Ferguson, 1963;Garren et al., 1965;Clark et al., 1997).
The above evidence strongly suggests that CRTC2 is involved in the regulation of Star transcription. The adrenal cortex also expresses CRTC1 and CRTC3, though levels of CRTC1 are much lower than those of CRTC2 and CRTC3 both in rat; Fig. S1, and in humans (Conkright et al., 2003). Thus, it is reasonable to hypothesize that, in addition to CRTC2 Takemori et al., 2007;Spiga et al., 2011a,b), activation of the other isoforms, especially the highly expressed CRTC3, is also involved in the transcriptional regulation of Star. The objective of the present study was to elucidate the roles of different CRTC isoforms on ACTH-regulated Star transcription and the signalling pathways involved in this regulation. For this purpose, we investigated the dynamics of nuclear translocation of the three endogenous CRTC isoforms in response to ACTH, in relationship with the time course of Star transcription and CREB phosphorylation in two murine adrenal cell lines, Y1-BS1 (Watt and Schimmer, 1981), and ATC7-L (Ragazzon et al., 2006). In contrast with the original cell line, Y1 (Yasumura et al., 1966;Bloch and Cohen, 1960), widely used to study steroidogenesis (Whitehouse et al., 2002;Zaidi et al., 2014;Rainey et al., 2004), the sub-clone, Y1-BS1, is ACTH-responsive but as with the parent line, has the disadvantage of not producing corticosterone (Parker et al., 1985). The line ATC7-L, derived from an adrenal fasciculata tumour (induced by targeted expression of SV40 large T antigen (Sahut-Barnola et al., 2000)), produces corticosterone in response to physiological ACTH concentrations, as well as episodic secretion in response to ACTH pulses (Ragazzon et al., 2006;Hazell et al., 2019). Using these two ACTHresponsive cell lines, we examined the physiological role of CRTC2 and CRTC3 on Star transcription, employing siRNA knockdown, and chromatin immunoprecipitation assays to investigate the recruitment of CRTCs by the Star promoter.

Materials and methods
All chemicals were purchased from Sigma Aldrich unless otherwise stated.

Rat adrenal cell isolation
Female Sprague Dawley rats were decapitated following CO 2 sedation, adrenal glands removed, decapsulated and quartered, and then digested with collagenase Type II, 2 mg/ml in DMEM/HEPES (Gibco), containing 1% pen/strep, 1% BSA fraction V, 0.02% Deoxyribonuclease I, for 20 min at 37°C, 95% air/5%CO 2 under agitation. Tissue was then sedimented, washed and resuspended in DMEM/HEPES containing 1% pen/strep, 1% BSA, 0.002% Deoxyribonuclease I and 0.01% Trypsin inhibitor, and mechanically dispersed by aspiration/release with a syringe attached to 3 mm tubing. The supernatant containing dispersed cells was filtered through a 100 μm nylon gauze, and the procedure repeated until the media became clear. A second collagenase incubation and dispersion was performed on undigested tissue. Pooled supernatant was centrifuged at 100×g, cell pellets resuspended and preincubated for 1 h in 20 ml DMEM/HEPES containing 0.1% BSA and 1% pen/strep, before resuspending in fresh medium at 250,000 cell/ml. Aliquots (1 ml) were incubated at 37°C for the time periods and conditions indicated. Incubations were terminated by placing vials on ice, then cells pelleted by centrifugation for RNA isolation.

RNA isolation and RT-qPCR
Cells were harvested in TRIzol reagent (Thermofisher Scientific) and RNA extracted using phase separation with chloroform. Total RNA was purified from the aqueous phase using RNeasy mini kit reagents and column DNase digestion (Qiagen, Valencia, CA, USA), per manufacturer instructions. Complementary (c)DNA was reverse transcribed from 1000 ng of RNA as previously described (Liu et al., 2008). qPCR primers for murine Star primary transcript (prior to splicing to mRNA), or heteronuclear RNA (hnRNA), were designed spanning an intronicexonic region: Forward 5′-TGTCTCGCTCGGGGTCACACA-3′, Reverse 5′-AGGCAGGGGCACCTCAAGCT-3′, Invitrogen). PCR reactions were performed using power SYBR green PCR mix (ThermoFisher Scientific), 166 nM of each primer and 2 μl cDNA, final volume 15 μl, in a spectrofluorometric thermal cycler 7900 HT Fast real-time PCR system (ThermoFisher Scientific) as previously described . Briefly, samples underwent denaturation at 50°C for 2 min, 95°C for 10 min, followed by 45 cycles at 95°C for 15 s and 60°C for 1 min hnRNA levels were calculated using relative quantification by standard curve, normalised to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA (murine Gapdh mRNA: Forward 5′-CCATCACTGCCAC CCAGAAGA-3′, Reverse 5′-GACACATTGGGGGTAGGAACA-3′, Invitrogen), as determined in separate qRT-PCR reactions. Absence of detection when omitting the reverse transcription enzyme Superscript III (Invitrogen) indicated a lack of genomic DNA contamination.

Statistical analyses
All data are expressed as mean ± SEM of values obtained from a minimum of three independent experiments. Normal distribution of the data and homogeneity of variance were verified using the Shapiro-Wilk test and Normal Q-Q plots, and the LEVENE test, respectively. Data were analysed using one-way ANOVA, Welch ANOVA or two-way ANOVA, as indicated in the figure legend. When appropriate, one-or two-way ANOVA was followed by Fisher Least Significant Difference (Fisher LSD) post-hoc test, Welch ANOVA was followed by the Dunnett T3 post-hoc test. Statistical significance was set at P ≤ 0.05, with a trend defined as P ≤ 0.1.

Time course of the effect of ACTH on Star transcription, CREB phosphorylation and CRTC translocation in Y1-BS1 and ATC7-L cells
The dynamics of Star transcription in response to ACTH were determined by measuring the levels of Star primary transcript (Star hnRNA). ACTH induced an overall increase in Star hnRNA levels both in Y1-BS1 and ATC7-L cells (P < 0.001; Fig. 1A and B), with no changes detected in the first 5-7 min, whereas a significant increase was observed at 15-60 min. Concomitant with the increase in Star hnRNA, ANOVA analysis revealed a significant effect of ACTH on pCREB in both cell lines (Y1-BS1: P = 0.02; Fig. 1C ATC7-L: P = 0.002; Fig. 1D). In Y1-BS1 cells, nuclear pCREB transiently increased 2.2-fold by 7 min, peaking at 15 min. In ATC7-L cells, pCREB increased by 7.5-fold at 15min and remained significantly elevated at 60 min (5.8-fold).
To investigate whether these changes in pCREB and Star transcription are associated with differential changes in CRTC subtype nuclear translocation, we measured the dynamics of cytosolic and nuclear levels of CRTC1, CRTC2 and CRTC3. ACTH had no significant effect on cytosolic CRTC1, CRTC2 or CRTC3 levels in Y1-BS1 ( Fig. 2A; data not shown) or ATC7-L cells ( Fig. 3A; data not shown). Y1-BS1 cells exhibited a significant overall increase in nuclear translocation of CRTC1 (P = 0.04), CRTC2 (P = 0.01) and CRTC3 (P = 0.04; Fig. 2A and B). Nuclear levels of CRTC3 were already significantly increased by 3 min, whilst nuclear levels of CRTC1 and CRTC2 were significantly increase at 7 min. A significant overall effect of ACTH on nuclear CRTC1 and CRTC3 was also detected in ATC7-L cells (CRTC1: P=0.05; CRTC3: P = 0.02; Fig. 3A and B), with nuclear levels of CRTC3 also significantly increased by 3 min. Despite an increase in nuclear levels of CRTC2 in each one of four experiments (range 1.5-5.3-fold at 3 min), the changes were not significant due to high variability in the magnitude of the effect across experiments (P = 0.22).

ACTH induces rapid association of pCREB, CRTC2 and CRTC3 at the proximal Star promoter in ATC7-L cells
The increases in nuclear CRTCs, preceding increases in Star hnRNA, suggest an involvement of the coactivators in the initiation of Star transcription. To determine whether CRTC2 and CRTC3 can interact with the Star promoter, we used the adrenocortical cell line, ATC7-L, and ChIP assays to measure recruitment of pCREB and CRTC proteins by the Star promoter. Recruitment of CRTC1 could not be measured because of the lack of a suitable antibody for ChIP. ACTH significantly increased the binding of pCREB (P = 0.04; Fig. 4A) and CRTC2 (P = 0.04; Fig. 4B) at the Star promoter. No increases in binding were detected at 7 min, but by 30 min, there was a significant increase above basal levels for both pCREB (P = 0.02) and CRTC2 (P = 0.02). Immunoprecipitation of the Star promoter by the CRTC3 antibody also revealed a significant effect of ACTH on CRTC3 binding (P < 0.001; Fig. 4C). Binding of CRTC3 to the Star promoter was detected by 15 min (P = 0.001), the earlier point measured, and remained at similar levels at 30 min (P < 0.001).

Knockdown of CRTC2 and CRTC3 attenuates Star transcription
The involvement of the different CRTC subtypes on ACTH-induced Star transcription was investigated further using siRNA oligonucleotides to inhibit the expression of CRTC1-3 in the adrenocortical cell line, Y1-BS1. While technical difficulties impaired CRTC1 knockdown, transfection with Crtc2 and Crtc3 siRNA oligonucleotides, alone or in combination effectively reduced the respective CRTC in whole cell protein extracts, for CRTC2 (P = 0.01; Fig. 5B) and CRTC3 (P = 0.002; Fig. 5C) protein levels. Transfection with Crtc2 siRNA decreased CRTC2 . Star hnRNA was normalised to Gapdh mRNA and is expressed as fold-changes of time 0 (n = 3-6/ time point). In C, pCREB was measured in the nuclear extract, was normalised to HDAC, and is expressed as fold-changes of time 0 (n = 4-5/time point). In D, pCREB was measured in the whole cell extract, was normalised to vinculin, and is expressed as fold-change of time 0 (n = 3/time point). Data in A were analysed using one-way ANOVA followed by Fisher LSD post-hoc test; data in B, C and D were analysed using Welch ANOVA followed by Dunnett T3 post-hoc test. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001 vs time 0. protein by 75.3% ± 8.4 (P = 0.02 vs non-coding siRNA sequence, siNC; Fig. 5B), whilst Crtc3 siRNA transfection decreased CRTC3 protein by 49.2% ± 14.3 (P = 0.03 vs siNC; Fig. 5C). Furthermore, there was no significant difference in whole cell CRTC2 and CRTC3 protein levels when cells were transfected with the heterologous siRNA, confirming specificity of the siRNA used. Knockdown of CRTC2 and CRTC3, alone or combined, did not affect CRTC1 protein levels (P = 0.49; Fig. 5A).
Star hnRNA levels were then measured either following 45 min treatment with 10 nM ACTH or vehicle (Fig. 5D). Two-way ANOVA  revealed an overall significant effect for both ACTH and siRNA on Star hnRNA (P < 0.001 and P = 0.03, respectively), but no interaction (P = 0.15). ACTH increased Star hnRNA levels in siNC cells (P < 0.001), an effect which was significantly attenuated in cells transfected with either Crtc2 or Crtc3 siRNA, or their combination. Although ACTH-stimulated values were still significantly higher than the respective basal (Crtc2 siRNA: P = 0.002; Crtc3 siRNA: P = 0.002; Crtc2+Crtc3 siRNA: P = 0.002), responses to ACTH were significantly lower than in siNC (Crtc2 siRNA: P = 0.05; Crtc3 siRNA: P = 0.004). No additivity was observed between the inhibitory effect of Crtc2 and Crtc3 siRNAs on ACTH-stimulated Star hnRNA, with the combined effect being similar to that of the individual oligonucleotides (P = 0.004, Fig. 4. ACTH-induced binding of pCREB, CRTC2 and CRTC3 to the Star promoter in ATC7-L cells. Cells were treated with 10 nM ACTH prior to chromatin immunoprecipitation by antibodies for pCREB (A), CRTC2 (B) and CRTC3 (C) proteins and normal rabbit IgG. Binding of each protein target to the Star promoter was calculated as percentage pulldown of the input. Data represent the mean ± SEM of data obtained in 3-4 independent experiments. pCREB, CRTC2 and CRTC3 data were analysed using one-way ANOVA followed by Fisher LSD post-hoc test. IgG data were analysed using Welch ANOVA (A-B) and one-way ANOVA (C). *P ≤ 0.05, ***P < 0.001 vs 0 min.

Fig. 5. Effect of CRTC2 and CRTC3 silencing on ACTH-induced increases in
Star hnRNA levels in Y1-BS1 cells. Y1-BS1 cells were transfected with non-coding siRNA (siNC) or siRNA oligonucleotides for CRTC2 (siCRTC2), CRTC3 (siCRTC3), or both (siCRTC2+3), 48 h prior to treatment with 10 nM ACTH for 45min. (A, B and C) Efficacy of silencing determined by Western immunoblot and semi-quantification of the levels of whole cell CRTC1, CRTC2 and CRTC3, normalised to actin, expressed as fold-change levels of the siNC transfected control group. No effect on CRTC1 levels was observed following transfection of siCRTC1 oligonucleotides (data not shown). (D) Effect of CRTC siRNAs on Star hnRNA levels, normalised to Gapdh mRNA and expressed as fold-change of siNC transfected controls. Data are the mean ± SEM of data obtained in 3 independent experiments. Data in A, B and C were analysed using one-way ANOVA followed by Fisher LSD post-hoc test; data in D were analysed using two-way ANOVA followed by Fisher LSD post-hoc test. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001 vs untreated control. # P ≤ 0.05 vs ACTHtreated siNC control. compared to siNC cells). No significant differences in basal Star hnRNA levels between siRNA treatment groups were found.

Stimulation of cAMP, but not of MAPK or PKC pathways, mimics the effect of ACTH on Star transcription
To test whether other signalling pathways, in addition to ACTH/ cAMP, can regulate Star transcription, Y1-BS1 and ATC7-L were treated for 30 min with either ACTH, 8-Br-cAMP, the PKC stimulator PMA, or the MAPK stimulator EGF (Fig. 6A and B). In Y1-BS1 cells (Fig. 6A), both ACTH (P = 0.001) and 8-Br-cAMP (P < 0.001) significantly increased the levels of Star hnRNA, whilst treatment with PMA or EGF had no effect on Star hnRNA levels. In ATC7-L cells (Fig. 6B), Star hnRNA levels also increased in cells treated with ACTH (P = 0.007). In each of the three experiments, 8-Br-cAMP increased in Star hnRNA levels, but due to variability in the magnitude of the increase (range 5.4-11.0-fold), the effect was not statistically significant (P = 0.23). Consistent with the findings in Y1-BS1 cells, neither PMA nor EGF had any effect on Star hnRNA levels in ATC7-L cells. These findings were confirmed in dispersed rat adrenal cells, with both ACTH and 8-Br-cAMP increasing Star hnRNA levels, while no effect of PMA treatment was seen (Fig. 6C).

ACTH-stimulated Star transcription is attenuated by inhibition of PKA and MAPK activity in a cell line specific manner
To further study the signalling pathways mediating Star transcription, murine adrenocortical cell lines were pre-treated with either vehicle, the PKA inhibitor H89, the MEK inhibitor UO126, or with a combination of both inhibitors for 15 min prior to incubation with 10 nM ACTH for 30 min. The inhibitors, alone or in combination, had no significant effect on basal Star hnRNA levels ( Fig. 6D-F). Both Y1-BS1 (Fig. 6D) and ATC7-L (Fig. 6E) cells showed a significant effect of ACTH (P < 0.001), inhibitors pre-treatment (Y1-BS1: P < 0.001; ATC7-L: P = 0.03), and a significant interaction (Y1-BS1: P = 0.01; ATC7-L: P = 0.03).
When compared with the respective basal, ACTH treatment in Y1-BS1 cells increased Star hnRNA levels in cells pre-treated with vehicle (P < 0.001), UO126 (P = 0.03) or H89 (P = 0.01), but the effect of ACTH was significantly attenuated in cells pre-treated with H89 (P = 0.02) and UO126 (P = 0.01) compared vehicle pre-treatment. Furthermore, pre-treatment with the combination of H89 and UO126 had a significant additive inhibitory effect (P < 0.001), preventing a significant stimulatory effect of ACTH (P = 0.26 vs basal). Similarly, ACTH increased Star hnRNA levels in ATC7-L cells pre-treated with vehicle (P = 0.004), UO126 (P < 0.001) and H89 + UO126 (P = 0.01) compared with respective basal values, but pre-treatment with H89 alone only tended to reduce the stimulatory effect of ACTH (P=0.15). Compared with ACTH treatment in vehicle pre-treated controls, ACTH-stimulated Star hnRNA levels also tended to be lower in H89 pre-treated cells (P = 0.09). The lack of a significant effect of H89 in the overall analysis was likely due to the wide range in the magnitude of the ACTH responses in the UO126 group, since there was a significant reduction when comparing ACTH-stimulated values in H89 and vehicle pre-treated groups by t-test analysis (P = 0.04). In contrast to Y1-BS1 cells, preincubation of ATC7-L cells with the MAPK inhibitor, UO126, significantly augmented the stimulatory effect of ACTH on Star hnRNA levels compared with ACTH in vehicle pre-treated cells (P = 0.02).
Similar to the effects in the cell lines, in vehicle pre-treated collagenase-dispersed rat adrenal cells (Fig. 6F) there was a significant effect of ACTH (P = 0.002), while the overall effect of the inhibitors or Fig. 6. Effect of mimicry and inhibition of ACTH signalling on Star hnRNA levels in Y1-BS1, ATC7-L and dispersed rat adrenals cells. Star hnRNA levels were measured either prior to or 30 min after treatment with 10 nM ACTH, 1 mM 8-Br-cAMP, 100 nM PMA or 3 nM EGF in Y1-BS1 (A) and ATC7-L (B), or 60 min after treatment in dispersed rat adrenal cells (C). For inhibition studies, Y1-BS1 cells (D), ATC7-L (E) and dispersed rat adrenal cells (F) were pre-incubated for 15 min with either vehicle (0.5% DMSO), 10 μM H89, 1 μM UO126, or 10 μM H89+1 μM UO126 before addition of ACTH at a final concentration of 10 nM. Bars represent the mean ± SEM of 3-6 experiments, normalised to Gapdh mRNA and expressed as fold-change from basal vehicle values. PMA treated cells in C represent the mean of data from 2 independent experiments. Data in A were analysed using one-way ANOVA followed by Fisher LSD post-hoc test; data in B were analysed using Welch ANOVA followed by Dunnett T3 posthoc test. Data in D, E and F were analysed using Two-way ANOVA followed by Fisher LSD post-hoc test. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001 vs untreated vehicle control. #P ≤ 0.05, ###P ≤ 0.001 vs ACTH-treated vehicle control. their interaction was not significant (P = 0.213 and P = 0.412, for inhibitors and interaction, respectively). As in ATC7-L cells, Star hnRNA increases in the ACTH-treated vehicle (P = 0.035) and UO126 (P = 0.008) groups were significant compared to basal, whilst pre-incubation with H89, but not with UO126, blunted Star hnRNA response to ACTH (P = 0.242). Also similar to ATC7-L cells, pre-incubation with UO126 did not inhibit but tended to increase the effect of ACTH compared with vehicle pre-treated cells.
3.6. Inhibition of PKA and calcineurin decreases ACTH-mediated nuclear translocation of CRTC1, CRTC2 and CRTC3 in ATC7-L cells but not ACTHmediated Star transcription To investigate the role of PKA, MEK and calcineurin on CRTCmediated regulation of ACTH-induced Star transcription, the effect of 10 min treatment with ACTH on Star hnRNA levels and nuclear levels of CRTC1, CRTC2 and CRTC3 were measured in ATC7-L cells pre-treated with UO126, H89 or the calcineurin inhibitor CsA. There was a significant effect of ACTH (P < 0.001) and inhibitors treatment (P < 0.001), as well as interaction (P = 0.006), on Star hnRNA levels (Fig. 7A). ACTH significantly increased Star hnRNA in cells pre-incubated with vehicle (P < 0.001), UO126 (P = 0.003) and CsA (P < 0.001) but not in cells pre-incubated with H89 alone (P = 0.17) or in combination with the other antagonists (H89 + UO126 + CsA: P = 0.37). Interestingly, the effect of ACTH on Star hnRNA was potentiated by CsA (P = 0.01 vs ACTH alone), whereas the potentiation by UO126 observed after 30 min ACTH treatment (Fig. 6D) was not present after 10 min treatment.

Discussion
This in vitro study shows that nuclear translocation of the 3 isoforms of endogenous CRTC (CRTC1, CRTC2 and CRTC3) parallels or precedes the increases in Star hnRNA induced by ACTH, in agreement with previous reports of ex vivo work in rats. Using chromatin immunoprecipitation and siRNA knockout we now demonstrate that CRTC2 and CRTC3 are involved in the regulation of Star transcription by ACTH in a PKA signalling dependent manner. The use of ACTHresponsive adrenocortical Y1-BS1 and ATC7-L cell lines allowed us to study the effect of physiological stimulator, ACTH, on endogenous proteins, rather than cAMP analogues, as with most previous studies (Takemori et al., 2007;Whitehouse et al., 2002;Zaidi et al., 2014;Jefcoate et al., 2011;Manna et al., 2002). The effects of ACTH on Star transcription and on nuclear pCREB and CRTC were similar in both cell Fig. 7. Effect of inhibition of ACTH signalling on Star hnRNA levels and CRTC activity in ATC7-L cells. Cells were pre-treated for 15 min with either vehicle (0.5%DMSO), protein kinase A inhibitor H89 (H, 10 μM), the MAP kinase inhibitor UO126 (U, 1 μM) or the calcineurin inhibitor CsA (5 μM), either alone or in combination (H89 + UO126 + CsA), prior to treatment with 10 nM ACTH for 10 min. (A) RT-qPCR quantification of Star hnRNA, normalised to Gapdh mRNA and expressed as fold-change levels of the basal vehicle control. Bars represent the mean ± SEM of 3-5 independent experiments.
(B) Representative Western immunoblot of nuclear CRTC1, CRTC2 and CRTC3 levels. Quantification of Western immunoblot data of CRTC1 (C), CRTC2 (D), and CRTC3 (E); data are normalised to HDAC and expressed as foldchange of basal. Bars represent the mean ± SEM of 3-5 independent experiments. Data were analysed using two-way ANOVA followed by Fisher LSD post-hoc test. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001 vs untreated vehicle control. # P ≤ 0.05, ## P ≤ 0.01, ### P ≤ 0.001 vs ACTH-treated vehicle control. L.I.F. Smith, et al. Molecular and Cellular Endocrinology 499 (2020) 110612 lines, however, the actions of signalling inhibitors differed, with the MEK inhibitor, UO126, having an inhibitory effect in Y1-BS1 cells and a potentiating effect in ATC7-L cells. Interestingly, the effect of signalling inhibitors in ATC-7 cells resembles that observed in collagenase dispersed rat adrenal cells, suggesting that, at least in rodents, they are more representative of normal adrenal fasciculata cells. However, resistance to transfection made ATC7-L cells unsuitable for use in CRTC knock down experiments. Whilst in earlier studies ACTH increases Star mRNA levels within 2 h in Y1-BS1 cells (Lin et al., 2001) and by 30 min in ATC7-L cells (Ragazzon et al., 2006), measuring Star hnRNA made it possible to show significant transcriptional activation in response to ACTH by 15 min in both cell lines. These rapid increases are consistent with findings in rat adrenal tissue 15 min after ACTH injection (Spiga et al., 2011a(Spiga et al., ,b, 2017, or cAMP incubation in Y1 cells (Lee et al., 2015;Jefcoate et al., 2011). Consistent with the view that CREB phosphorylation is critical for Star transcription (Spiga et al., 2011a(Spiga et al., ,b, 2017Manna et al., 2003;Lefrancois-Martinez et al., 2011), ACTH treatment induced rapid and transient increases in nuclear pCREB in both cell lines.
Rapid ACTH-induced nuclear translocation of all 3 CRTC isoforms (CRTC1, 2 and 3) by 10 min in both Y1-BS1 cells and ATC7-L cells, coinciding with the earliest detected increases in Star hnRNA following ACTH exposure, supports a role for CRTC in the initiation of Star transcription. These kinetics differ from previous findings in transfected Y1 cells, showing nuclear translocation of CRTC2, but not CRTC1 or CRTC3, in response to cAMP stimulation (Lee et al., 2015). In keeping with the rapid nuclear translocation of the three CRTC isoforms shown here, using ATC7-L cells we demonstrate, for the first time, an early interaction of CRTC2 (by 30 min) and CRTC3 (by 15 min) with DNA fragments including the −104 bp to −5 bp Star promoter region containing three cAMP-response element (CRE) half-sites (Manna et al., 2002). The present finding of pCREB binding to the Star promoter at 30 min ACTH incubation is consistent with previous observations in Y1 cells stimulated with cAMP analogues (Jefcoate et al., 2011). The latter study also showed delayed (60 min) CRTC2 recruitment, from which authors concluded that CRTC2 is involved in maintaining, rather than initiating, high rates of Star transcription. The difference between the former (Lee et al., 2015;Jefcoate et al., 2011) and present findings could reflect the use of different cell lines, or studying the effect of ACTH on endogenous CRTC2 as opposed to the effect of cAMP on transfected protein (which may exhibit altered intracellular localisation and bioactivity). Interestingly, in contrast to the parallel nuclear translocation patterns of the three CRTC isoforms, extrapolation of the ChIP time courses suggests that ACTH-induced recruitment of CRTC3 by the Star promoter, which is already maximal at 15 min, could potentially precede that of CRTC2. Although CRTC2 recruitment was not measured at 15 min, the lack of any promoter association at 7 min renders it unlikely that binding is maximal at 15 min. Since CRTC lacks DNA-binding activity, its recruitment by the Star promoter CRE requires association with pCREB through its CREB binding domain (CBD) (Luo et al., 2012). Amino acid substitution studies have shown that small changes in key motifs of the CBD region of CRTC2 can impact its binding affinity with pCREB (Conkright et al., 2003;Luo et al., 2012). Thus, while the similar nuclear translocation time courses of CRTC1-3 could predict parallel dynamics of recruitment to the Star promoter, differences in pCREB affinity between the three CRTC isoforms could affect the recruitment time by the Star promoter. Testing this possibility will require further detailed analyses of the early kinetics of the three CRTC isoform pCREB association and recruitment by the Star promoter.
Previous reports demonstrate that CRTC2 overexpression increases Star transcription (Takemori et al., 2007), while activation of salt inducible kinase (SIK) 1, which prevents activation and nuclear localisation of CRTC2, inhibits Star transcription (Lin et al., 2001;Katoh et al., 2004). We demonstrate that siRNA knockdown of CRTC2 and CRTC3 in Y1-BS1 cells only partially inhibits ACTH stimulation of Star hnRNA levels, directly establishing a role for CRTC mediating ACTHinduced Star transcription. However, in contrast to the full inhibition of cAMP-dependent Crh transcription by simultaneous knockdown of CRTC2 and CRTC3 seen in rat hypothalamic 4B cells , knocking down both CRTC isoforms in Y1-BS1 cells had no additive effect on Star transcription after 45 min ACTH exposure. This suggests CRTC2 and CRTC3 both act through a similar mechanism of action, and that at the time of maximal Star transcription either CRTC2 of CRTC3 are capable of inducing full activation. However, it is possible that differential recruitment of CRTC isoforms by the atypical Star half-CREs (Luo et al., 2012;Song et al., 2018), due to differential affinity with pCREB or other mechanism (Conkright et al., 2003;Luo et al., 2012), could have different functional implications at earlier time points of ACTH stimulation.
It is also evident from these experiments that other factors are capable of mediating considerable ACTH-stimulation of Star transcription in the absence of both CRTC2 and CRTC3. Although CRTC1 expression is far lower than that of CRTC2 and CRTC3 in rat (Fig. S1) and human adrenals (Conkright et al., 2003), we show clear nuclear translocation of CRTC1 following ACTH treatment, suggesting a possible role for CRTC1 in Star transcription. This could not be examined in the present study because of the inability to knockdown CRTC1 protein in Y1-BS1 cells, and future studies will be needed for elucidating the potential role of CRTC1 on Star transcription. In addition, SF-1, which is upregulated by ACTH (Ragazzon et al., 2006;Hazell et al., 2019), binds the Star promoter and is necessary for cAMP-induced Star transcription (Sandhoff et al., 1998;Sugawara et al., 1997). Furthermore, subsets of CREB-inducible genes can be alternatively regulated by CREB coactivator CREB binding protein (CBP) (Kasper et al., 2010). The role of these factors and their interaction with the transcriptional complex during Star transcriptional initiation will require further investigation.
Investigation of signalling pathways showed that whilst MAPK activation by EGF had no effect on Star transcription in either Y1-BS1 or ATC7-L cells, MEK inhibition by UO126 attenuated ACTH-induced Star hnRNA in Y1-BS1 cells, but not ATC7-L cells. Furthermore, combination of H89 and UO126 abolished Star transcriptional responses to ACTH in Y1-BS1 cells, suggesting both pathways are required for full cAMP-dependent stimulation of Star transcription in this cell line. Similar results were found when using the PKA inhibitor PKI or MAPK inhibitor SL327 (Smith, Huang and Aguilera, unpublished observations). Previous studies also suggest cell-specific differences in the MAPK involvement in Star transcriptional regulation; Lefrancois-Martinez et al. (2011) found no effect of MAPK inhibition on ACTH-stimulated StAR protein levels in ATC1 cells, whilst, Gyles et al. (2001) showed inhibition of forskolinstimulated Star mRNA and protein by MEK inhibitors in Y1 cells. Transactivation of the MAPK pathway by ACTH and cAMP has been implicated in steroidogenesis (Gyles et al., 2001;Winnay and Hammer, 2006). Furthermore, consistent with previous reports (Le and Schimmer, 2001), ACTH increased pERK levels in Y1-BS1 cells and, to a lesser extent, ATC7-L cells (Smith, Olah and Aguilera, unpublished observations).
H89 blunted ACTH-induced increases in nuclear CRTC, consistent with the established cAMP dependence of CRTC activation and translocation to the nucleus (Bittinger et al., 2004;Takemori et al., 2007;Liu et al., 2010). This was associated with complete inhibition of ACTHinduced Star hnRNA at 10 min, while there was a partial recovery by 30 min. This is consistent with the effects of siRNA CRTC knockdown, in which there was only a partial reduction of Star hnRNA by 45 min ACTH stimulation, despite reduced CRTC2 and CRTC3 levels. Although earlier Star hnRNA responses to ACTH were not measured in the siRNA experiments, the overall findings strongly suggest that PKA/CRTC mechanisms are essential for early transcriptional activation of Star, and that additional signalling mechanisms are important to sustain the activation. The fact that MAPK appears to play a role in CRTC regulation only in Y1-BS1 cells, raises a note a caution when extrapolating data obtained from cell lines to in vivo regulation in various species.
No effect of PMA indicates PKC activation alone is insufficient for initiating Star transcription. PKC has previously been implicated in the regulation of StAR protein levels in bovine adrenal cells (Nishikawa et al., 1996), whilst low levels of steroidogenesis were previously induced by 2 h incubation with PMA in Y1 cells (Frigeri and Armelin, 1996). Furthermore, Leydig M-10 cell studies have shown phorbol esters increase Star mRNA levels (Manna and Stocco, 2005;Manna et al., 2011). Ca 2+ -sensitive calcineurin can be activated by cAMP (Antoni, 1996), and it is known to activate CRTC (Bittinger et al., 2004;Lee et al., 2015;Rahnert et al., 2016;Screaton et al., 2004). Thus, ACTH signalling could lead to calcineurin activation and play a role in CRTC activation and nuclear translocation in the adrenal fasciculata. However, CsA potentiated the effect of ACTH on Star transcription, despite attenuating ACTH-induced nuclear translocation of CRTC2. It is noteworthy that Y1-BS1 cells lack part of the steroidogenic enzyme expression profile, which characterises zona fasciculata cells (Bloch and Cohen, 1960;Parker et al., 1985). The lost differentiation of these cells may be responsible for the apparent dependence of Star transcription on calcium (Lee et al., 2015) and MAPK signalling (Gyles et al., 2001), that was not observed in ATC7-L cells in this study or elsewhere (Lefrancois-Martinez et al., 2011;Schiebinger et al., 1985). Furthermore, CsA only partially inhibits CRTC1 and CRTC2 translocation, suggesting PKA may be the main mechanism regulating CRTC1 and CRTC2 activity. Although less sensitive to calcium signalling than zona glomerulosa cells (Braley et al., 1986;Omura et al., 2007;Spat et al., 2016), intracellular calcium depletion in rat zona fasciculata cells blunts cAMP-induced corticosterone release, suggesting that cAMP activation of adrenal steroidogenesis in the adrenal fasciculata requires calcium release from intracellular stores (Schiebinger et al., 1985).
This study provides evidence that of the 3 isoforms of CRTC present in the adrenal fasciculata, at least CRTC2 and CRTC3 are involved in the initiation of Star transcription. The rapid nuclear translocation and recruitment by the Star promoter of CRTC2 and CRTC3 in response to ACTH, preceding the maximal increases in Star primary transcript, and the ability of siRNA knockdown of these CRTC isoforms to attenuate ACTH-induced Star transcription, strongly suggest that PKA-mediated activation of CRTC2 and CRTC3 plays a role in the initiation of Star transcription. However, the fact that knockdown of both subtypes (or preventing their activation by PKA inhibitors) reduces but does not prevent ACTH-induced Star hnRNA increases, indicates the participation of additional factors and emphasizes the complexity of the mechanisms regulating StAR expression.