DNA damage response induced by Etoposide promotes steroidogenesis via GADD45A in cultured adrenal cells

Glucocorticoid production is regulated by adrenocorticotropic hormone (ACTH) via the cyclic adenosine monophosphate (cAMP)/protein kinase A (PKA) pathway in the adrenal cortex, but the changes in steroidogenesis associated with aging are unknown. In this study, we show that cell-autonomous steroidogenesis is induced by non-ACTH- mediated genotoxic stress in human adrenocortical H295R cells. Low-dose etoposide (EP) was used to induce DNA damage as a genotoxic stress, leading to cellular senescence. We found that steroidogenesis was promoted in cells stained with γH2AX, a marker of DNA damaged cells. Among stress-associated and p53-inducible genes, the expression of GADD45A and steroidogenesis-related genes was significantly upregulated. Immunofluorescence analysis revealed that GADD45A accumulated in the nuclei. Metabolite assay using cultured media showed that EP-treated cells were induced to produce and secrete considerable amounts of glucocorticoid. Knockdown of GADD45A using small interfering RNA markedly inhibited the EP-induced upregulation of steroidogenesis-related gene expression, and glucocorticoid production. A p38MAPK inhibitor, but not a PKA inhibitor, suppressed EP-stimulated steroidogenesis. These results suggest that DNA damage itself promotes steroidogenesis via one or more unprecedented non-ACTH-mediated pathway. Specifically, GADD45A plays a crucial role in the steroidogenic processes triggered by EP-stimulated genotoxic stress. Our study sheds new light on an alternate mechanism of steroidogenesis in the adrenal cortex.


Induction of steroidogenesis by EP in H295R cells.
To determine whether the DDR induces steroidogenesis, we treated H295R cells with different concentrations of etoposide (EP) for 3 days (72 h), followed by culturing with normal growth medium. By treatment of 0.75 and 1.0 μM EP, the number of γH2AX-positive cells was significantly increased 4 days after treatment ( Fig. 1A and C). Until 2 days after EP treatment, the rate of γH2AX foci-positive cells remained to be approximately 10%. Three days after treatment, the number of γH2AX foci-positive cells began to increase and peaked after 4-5 days (Fig. 1B). Because γH2AX staining is established as a reliable quantitative indicator of the DDR, it was confirmed that DNA damage occurred in the cells 3 days after EP treatment. Interestingly, cells positive for the steroidogenic enzyme CYP21A2 also increased in the same manner as γH2AX foci-positive cells ( Fig. 1B and C). As shown in quantitative analyses of double staining of γH2AX and CYP21A2, a remarkable increase in the number of CYP21A2-positive cells by EP treatment was observed in γH2AX-positive cells, but not in γH2AX-negative cells ( Fig. 1C and D, and Supplementary Table S1). These findings indicate the strong relationship between EP-induced DDR and the induction of CYP21A2 in H295R cells.
We measured cortisol concentration in the medium of the cells to ensure cortisol production in response to treatment with EP. Cortisol synthesis increased more than 3-times 4 days after treatment (Fig. 1E).
Quantitative real-time PCR (RT-qPCR) experiments were performed using H295R cells treated with or without EP. The levels of HSD3β2, CYP21A2, CYP17A1, CYP11A1, CYP11B1, CYP11B2, Nurr1 and Nur77 mRNA transcript were significantly increased by EP in a time-and dose-dependent manner ( Fig. 1F and Supplementary Table S2 and Supplementary Fig. S1). Their expression was not increased until 2 days after the addition of EP. Conversely, the expression of steroidogenic factor 1 (SF1), which plays a major role in regulating steroidogenic enzymes 58 , and StAR was not changed (Supplementary Fig. S2). These data suggest that the DDR induced by EP promotes steroidogenesis in H295R cells and it is accompanied by the upregulation of certain steroidogenesis-related factors.
Involvement of GADD45A in steroidogenesis in H295R cells. Having obtained evidence that EP induced steroidogenesis as well as DNA damage in H295R cells, we focused on GADD45A, whose mRNA expression is known to be increased in stressful growth arrest conditions and/or treatment with DNA-damaging agents [30][31][32][33][34] . Our RT-qPCR showed that GADD45A mRNA expression was elevated by 1.61-fold in 72 h EP-treated cells and 3.79-fold in 96 h EP-treated cells (Fig. 1F, and Supplementary Table S2). Further, GADD45A accumulated in the nuclei of approximately 30% of cells treated with EP, whereas control cells did not express GADD45A ( Fig. 2A). To demonstrate that GADD45A is involved in EP-induced steroidogenesis, we introduced small interfering RNA (siRNA) against GADD45A (siGADD45A) into H295R cells. As shown in Fig. 2A, siGADD45A decreased not only GADD45A expression but also CYP21A2 expression in EP-treated H295R cells. Moreover, cortisol synthesis was decreased by approximately 20% in the presence of siGADD45A (Figs 2B and S3A). The levels of HSD3B2, CYP21A2, CYP17A1, CYP11A1, CYP11B1, CYP11B2, Nurr1, and Nur77 mRNA transcripts were significantly decreased by siGADD45A (Fig. 2C, Supplementary Table S3, and Supplementary  Fig. S3B). Further, we performed transient introduction of a FLAG-tagged GADD45A expression vector into H295R cells and assayed steroidogenesis. As shown in Fig. 3A, HSD3B2, CYP21A2, CYP11B1, CYP11B2, and Nurr1 gene expression was significantly upregulated by approximately 1.7-, 1.9-, 1.4-1.9-, and 1.4-fold, respectively, in GADD45A-transfected H295R cells, in which a greater than 1000-fold increase in GADD45A was confirmed ( Fig. 3A and Supplementary Table S4). Immunofluorescence analysis also showed that CYP21A2 expression was induced in those cells (Fig. 3B). Approximately 60% of FLAG-positive cells expressed CYP21A2, whereas it was expressed in around 10% of empty vector-transfected, total FLAG-GADD45A-transfected, or FLAG-GADD45A-transfected FLAG (−) cells ( Fig. 3C and Supplementary Table S5), suggesting that CYP21A2 expression is induced predominantly in FLAG-GADD45A-expressing cells. These data suggest that GADD45A is involved in EP-induced steroidogenesis.

Inhibition of EP-induced steroidogenesis by SB203580 in H295R cells.
As p38MAPK is known as one of the main downstream molecules of GADD45A, the effect of siGADD45A on the phosphorylation level of p38MAPK was analyzed. Western blotting of p38MAPK showed that the level of phosphorylated p38MAPK, normalized by p38MAPK, was increased by approximately 2-fold in EP-treated cells, and decreased to control level by treatment with siGADD45A, suggesting that EP induced the phosphorylation of p38MAPK via GADD45A utilization ( Fig. 4A and B). RT-qPCR revealed that SB203580, an inhibitor of p38MAPK, significantly reduced the levels of HSD3B2, CYP21A2, CYP17A1, CYP11A1, CYP11B1, CYP11B2, Nurr1, and Nur77 mRNA levels ( Fig. 5A and Supplementary Table S6).
During acute stress, the activation of the HPA axis promotes steroidogenesis in adrenocortical cells by increasing the concentration of c-AMP, followed by PKA activation. Therefore, we explored whether EP-induced steroidogenesis is involved in PKA activation. H89, an inhibitor of PKA, significantly decreased the expression of Nurr1 mRNA by 22% in EP-treated cells ( Fig. 5B and Supplementary Table S7). The other factors examined were not changed. EP-induced cortisol synthesis was decreased to 40% or 20% by H89 or SB203580, respectively (Fig. 5C). SB203580 was significantly more effective than H89 at suppressing EP-induced cortisol synthesis. On the other hand, 8-bromo-cAMP (8-Br-cAMP), an analog of cAMP, remarkably increased the expression of steroidogenesis-related genes, but not GADD45A ( Supplementary Figs S4 and S5). Furthermore, H89 significantly decreased HSD3B2, CYP17A1, CYP11B1, CYP11B2, and Nurr1 expression ( Supplementary  Fig. S4). siGADD45A did not reduce the level of those steroidogenesis-related genes in 8-Br-cAMP-stimulated cells ( Supplementary Fig. S5). These data suggest that EP-induced steroidogenesis is mainly facilitated by GADD45A-p38 MAPK pathway, while 8-Br-cAMP-induced steroidogenesis is regulated by PKA, supporting our concept that EP-induced steroidogenesis occurs via a mechanism distinct from the HPA axis.

Discussion
In this study, we found that EP, which induces the DDR, promoted steroidogenesis in human adrenal H295R cells in a cell-autonomous manner, accompanied by the upregulation of steroidogenesis-related genes expression and the accumulation of CYP21A2 in particular, leading to significant cortisol production. The induction of steroidogenesis occurred through GADD45A upregulation and p38MAPK activation.
EP is a chemotherapeutic agent that induces DSBs by inhibiting DNA topoisomerase II (TopII) 59 . A high or low dose of EP induces apoptosis or cellular senescence, respectively 60 . The induction of DSBs triggers the phosphorylation of one of the variants of the nucleosome core histone H2A, namely, histone H2AX at Ser 139 (γH2AX). This phosphorylation, which is mediated by ATM, ATR, and/or DNA-dependent protein kinases, takes place in nucleosomes along a megabase of DNA flanking the DSB. As the immunocytochemical detection of γH2AX foci reveals that DNA damage with DSBs has occurred in cells, γH2AX is considered a biomarker of DNA damage. In this study, we treated H295R cells with a low dose of EP, which did not induce apoptosis. From 3 days after EP-treatment and thereafter, γH2AX foci-positive cells were increased significantly, confirming medium of the cells treated with EP (μM) 4 days after EP (0.75 μM) treatment was measured using an enzymelinked immunosorbent assay (ELISA). Data are presented as the mean ± SE of 3 independent experiments. *P < 0.05 vs. control. (F) Relative mRNA expression of the indicated genes was analyzed by qRT-PCR. RNA was extracted from H295R cells 3 days (72 h) or 4 days (96 h) after EP (0.75 μM) treatment. Data are presented as the mean ± SE of 3 independent experiments. *P < 0.05 vs. control.

Figure 2.
Inhibitory effect of GADD45A siRNA on steroidogenesis promoted by EP. H295R cells were transfected with the indicated siRNAs simultaneously with plating. After 24 h, the cells were treated with EP (0.75 μM) for 72 h, changed to the normal growth medium, and cultured for another 24 h. (A) Immunostaining for GADD45A and CYP21A2. H295R cells were fixed with 4% paraformaldehyde. Green staining shows the anti-GADD45A antibody, red staining shows the anti-CYP21A2 antibody, and blue staining shows DAPI (cell nuclei). Scale bars represent 100 μm. (B) Cortisol concentration in the medium was measured using ELISA, and corrected by the number of cells. Data are presented as the mean ± SE of 3 independent experiments. *P < 0.05 vs. control, † P < 0.05 vs. EP. (C) Relative mRNA expression of the indicated genes was analyzed by RT-qPCR. RNA was extracted at 12 h after changing of the medium (at 84 h after EP-treatment). Data are presented as the mean ± SE of 3 independent experiments. *P < 0.05 vs. control, † P < 0.05 vs. EP + siControl.  Table S1). These results indicate that steroidogenesis is activated in H295R cells undergoing the DDR without the aid of other stimuli. H295R cells are the only human adrenocortical cell line that has the potential to produce steroid hormones, mineralocorticoids, glucocorticoids, and adrenal androgens in a physiological manner 61 . There are two available and well-characterized adrenocortical cell lines that have the ability to produce steroid hormones, H295R cells and Y1 mouse adrenocortical tumor cells. We treated Y1 cells with EP and analyzed steroidogenesis. As Y1 cells do not produce corticosterone 56,62 , we assessed progesterone secretion in response to EP. As shown in Supplementary Fig. S6A and B, we observed a 2.5-fold increase in progesterone secretion, which was associated with the upregulation of StAR, CYP11A1, and CYP11B1 gene expression. As limited steroid hormone production is life-threatening, the induction of steroidogenesis in DNA-damaged adrenocortical cells might be essential for survival in vivo. We suspect that the activation of steroidogenesis in response to DNA damage might be a general mechanism in mammalian adrenocortical cells.
In this study, we found that GADD45A expression was upregulated, and knockdown of GADD45A using siRNA inhibited EP-induced steroidogenesis, which was accompanied by blocking the expression of steroidogenesis-related genes, such as HSD3B2 (Figs 2 and S3). Further, the transient overexpression of GADD45A in H295R cells upregulated the expression of HSD3B2 and CYP21A2 mRNA, which was accompanied with CYP21A2 protein expression (Fig. 3). These findings indicate that GADD45A plays an important role in EP-induced steroidogenesis. By contrast, 8-Br-cAMP-induced steroidogenesis was not inhibited by siGADD45A (Supplementary Fig. S5). These data indicate that GADD45A is associated specifically with EP-induced steroidogenesis. GADD45A protein lacks enzymatic activity and exerts its function by interacting with effectors of the cell cycle, apoptosis, DNA repair, and cellular senescence, and decides cell fates. In this study, an inhibitor of p38MAPK inhibited EP-induced steroidogenesis ( Fig. 5A and C), suggesting that p38MAPK is involved in this process. Several papers reported that GADD45A binds to and activates p38MAPK [40][41][42] , or MAP3K4, which is known as an upstream kinase of p38MAPK 55,63-65 . Burllard et al. reported that GADD45A activates MEKK4 by forming a complex, which is followed by the induction of skeletal muscle fiber atrophy 65 . GADD45A might induce steroidogenesis by interacting with MEKK4. On the other hand, GADD45A expression was not increased in EP-treated Y1 cells (Supplementary Fig. S6B). We revealed that StAR gene expression was upregulated by EP ( Supplementary Fig. S6B). It is reported that KRAS is expressed at a high level in Y1 cells 56 . Thus, a mechanism other than the induction of GADD45A might underlie EP-induced steroidogenesis in Y1 cells.
We demonstrated that siGADD45A reduced p38MAPK phosphorylation (Fig. 4). An inhibitor of p38MAPK, SB203580, reduced the EP-induced upregulation of steroidogenesis-related gene expression and glucocorticoid production ( Fig. 5A and C). These data suggest that p38MAPK is located downstream of GADD45A in EP-induced steroidogenesis. It is important to identify the target(s) of p38MAPK to clarify the mechanism of  Supplementary Fig. S9. (B) The level of phosphorylated p38MAPK was normalized by p38MAPK expression. Quantification of the expression level was performed by ImageJ software. Data are presented as the mean ± SE of 3 independent experiments. *P < 0.05 vs. control. † P < 0.05 vs. EP + siControl.
EP-induced steroidogenesis. p38MAPK phosphorylates several major transcriptional factors, such as CREB and GATA. There are GATA binding sites in the promoter of HSD3B2 66 , while CREB binding sites are present in the promoters of Nurr1, Nur77, CYP11B1, and CYP11B2 67-69 . Thus, it seems likely that p38MAPK regulates steroidogenesis-related gene expression and glucocorticoid production. The identification of factors connecting p38MAPK to steroidogenesis is an important aim of a future study.
The mammalian stress response involves the activation of the HPA axis, followed by an increase of cAMP and PKA activation, resulting in steroidogenesis in adrenocortical cells through the upregulation of SF1 transcripts. In this study, the steroidogenesis induced by EP via the DDR was partially inhibited by the PKA inhibitor H89 (Fig. 5B and C), while the overall steroidogenesis induced by 8-Br-cAMP, a cAMP analog and PKA activator, was effectively inhibited by H89 (Supplementary Fig. S4). Of particular interest, SF1 and StAR gene expression was not affected in EP-treated cells (Supplementary Fig. S2). These data strongly indicate that the DDR may surely promote steroidogenesis via a pathway other than the HPA axis.
The HPA axis is self-regulated by a negative feedback exerted by serum cortisol levels in both the hypothalamus and the pituitary gland. As EP-induced steroidogenesis is independent of the HPA axis, the production of too much glucocorticoid cannot be self-regulated by the normal negative feedback mechanism. For example, various degrees of DNA damage occur in aged organs, and it has been reported that glucocorticoid concentration is elevated in the serum of aged mice 22 . Age-related increases in cortisol production have been documented in human using blood and salivary samples [23][24][25][26] . Several papers reported that a decrease in the sensitivity of the HPA axis to glucocorticoid feedback suppression occurs with aging [70][71][72][73][74][75] . The increases in cortisol observed with aging have been attributed to an impairment of feedback inhibition of HPA activity due to neuronal loss in the hippocampal area 23,25 . Thus, our present study might suggest that such an elevation of glucocorticoid may originate from DNA damage in aged adrenal glands through the upregulation of the stress-associated protein GADD45A and the activation of p38MAPK. The elucidation of the mechanism of DNA damage-induced glucocorticoid production, leading to prohibition of glucocorticoid excess caused by DNA damage, may provide a new approach toward the prevention of some of the complications associated with aging and age-related diseases.
In this study, we treated H295R cells with a low dose of EP, which induced DNA damage, to mimic cellular senescence. We examined the expression of certain factors of the senescence-associated secretary phenotype (SASP) and senescence marker p16 Ink4 (p16). As shown in Supplementary Fig. S7A, at 96 h, a significant upregulation of IL8 and MMP10 mRNA was observed (IL8: 7.0-fold, MMP10: 3-fold). Thereafter, at 7 days and 20 days after EP treatment, the expression decreased to the control level. On the other hand, p16 remained unchanged throughout. The cell cycle analysis by flow cytometry revealed that the population of cells in the G2/M phase was significantly increased following EP treatement, and was decreased to the control level at 20 days after EP treatment ( Supplementary Fig. S7B). These findings indicate that the upregulation of IL8 and MMP10 and cell cycle inhibition in the G2/M transition occurred temporally and thereafter recovered. These data might suggest that the cells were immortalized and not induced to be fully senescent by EP treatment. Indeed, it may be difficult to induce cellular senescence in H295R cells. It has been reported that H295R cells lack telomerase activity, but maintains their telomeres by the alternative lengthening of telomeres (ALT) mechanism 76 . Further, Sampaoli C, et al. reported that H295R cells have p53 mutation, the lack of exons 8 and 9, including the part of the DNA binding domain and the entire C-terminal domain 77,78 , suggesting that p53 activity is completely lost. Ragazzon B et al. reported that H295R cells carry a homozygous deletion of the RB transcriptional corepressor 1 (RB1) gene 79 , and Hadjadj et al. reported that RB is not detected in H295R cells 80 . Indeed, we did not detect exon 8-9 of p53 or whole RB1 in these cells as shown in Supplementary Fig. S8. Thus cellular senescence might barely occur in H295R cells, due to the lack of p53 activity and RB pathways as well as the presence of the ALT pathway. Although the present study may not represent a typical model that uses an aged adrenal gland, an increase in the glucocorticoid production was one of the features of adrenal glands in aging mammals [22][23][24][25][26] .
In summary, we showed that EP activates cell-autonomous steroidogenesis in H295R cells undergoing DNA damage, which may constitute a part of senescence, through the expression of genes for steroidogenic factors and enzymes. The EP-induced upregulation of the stress-associated protein GADD45A and activation of p38MAPK were followed by steroidogenesis. This study provides novel information on one of the diverse mechanisms of steroidogenesis associated with aging. EP treatment. At  anti-phospho-p38MAPK antibody, 1:1000 in 5% milk/TTBS). The membranes were incubated with the secondary antibodies for 1 h at room temperature. After the final washes, the membranes were scanned using Typhoon Trio (Amersham), and the signals of reactive bands were quantified using ImageJ software.

Materials. EP and 8-Br
Cortisol assays. Cortisol production from H295R cells was analyzed by measuring its concentration in culture medium using a Chemiluminescent Enzyme Immunoassay (LSI Medience Corporation, Tokyo, Japan), and normalized by cell number. Three independent experiments were performed in triplicate, and followed by statistical analyses.

Reverse-transcription and real-time quantitative PCR (RT-qPCR).
Total RNA was extracted with RNA iso-Plus (TaKaRa Bio, Inc., Shiga, Japan) according to the manufacturer's instructions. Total RNA (1 μg) was reverse-transcribed to cDNA using a PrimeScript RT Reagent Kit with gDNA Eraser (RR047A, TaKaRa) according to the manufacturer's instructions. qRT-PCR was performed using a One Step SYBR PrimeScript RT-PCR Kit Ver. 1 (RR066A; TaKaRa) and a Thermal Cycler Dice Real-Time System (TP800; TaKaRa) according to the manufacturer's instructions. The thermal cycling conditions consisted of an initial denaturation step at 95 °C for 30 s followed by 40 cycles of PCR under the following conditions: 95 °C for 5 s and 60 °C for 60 s. GAPDH was used as an internal control because this gene, along with the cyclophilin gene, is employed widely as an internal control for the changes of mRNA levels of steroidogenic enzyme genes 57,81,82 . Our choice of appropriately distant primer sets and experiments with or without reverse transcriptase excluded the possibility that our real-time RNA quantification counted genomic DNA. The relative amount of each transcript was calculated with the 2 −∆∆Ct method 83 using the cycle threshold value, which was determined automatically by the real-time PCR system by means of the second derivative maximum method 84 . Primer pairs were subsequently tested for performance: absence of primer dimers and efficiency of amplification >95%, <105%. The primer sets are described in Table S8. Three independent experiments were performed in triplicate, and followed by statistical analyses.
Immunofluorescence analysis. The cells were cultured on coverslips in 6-well plates. The cells were fixed with 4% paraformaldehyde and 4% sucrose in PBS for 20 min at room temperature. Permeabilization was carried out with 0.2% Triton X-100 in PBS for 20 min at room temperature. Nonspecific binding was blocked by incubation in 3% bovine serum albumin and 0.1% Triton X-100 in PBS for 30 min at 37 °C. The antibodies were diluted in the above blocking solution at the indicated concentrations and incubated for O/N at 4 °C. Secondary antibodies were also diluted in the blocking solution and incubated for 1 h at room temperature. Nuclei were stained with 2 μg/mL 4′6-diamidine-2′-phenylindole dihydrochloride (DAPI) (Roche, Mannheim, Germany) in PBS for 15 min at room temperature. Images were acquired using a laser-scanning confocal image system (A1R-A1 Confocal Microscope System; Nikon, Tokyo, Japan). The primary anti-flag, anti-γH2AX, anti-GADD45A, and anti-CYP21A2 antibodies were used at concentrations of 1:1000, 1:50, 1:250, and 1:50, respectively, and the secondary antibodies were used at a concentration of 1:500. Statistical analysis. Quantitative data are expressed as the mean ± standard error (SE). Statistical analysis was performed by Student's t test or one-way analysis of variance followed by a post-hoc Turkey's test using JMP9 software.