The role of miRNAs in regulating adrenal and gonadal steroidogenesis.

Currently very little information is available about the miRNA regulation of steroidogenesis in testicular Leydig cells. Studies mainly from our laboratories have shown that some miRNAs may be involved in the regulation of testicular steroidogenesis either directly or indirectly using model Leydig cell lines, MLTC-1 and R2C cells ( Hu et al . 2012 , 2013 ). It was reported by us that miRNA-125a is expressed at much higher levels than miRNA-125b in MLTC-1 cells ( Hu et al . 2012 ). In contrast, very little miRNA-455–5p or miRNA-145 expression was noted under basal conditions. Treatment of MLTC-1 cells with a cAMP analog, Bt 2 cAMP, significantly downregulated the expression of all four miRNAs (i.e., miRNA-125a, miRNA-125b, miRNA-455–5p and miRNA-145). We further demonstrated that miRNA-125a and miRNA-455 act as potent negative regulators of SR-BI and SR-BI-mediated selective HDL-CE transport function ( Hu et al. 2012 ). Overexpression of pre-miRNA-125a and pre-miRNA-455 significantly decreased the levels of progesterone produced in both R2C and MLTC-1 cells. Additional follow-up experiments demonstrated that treatment of MLTC-1 cells with Bt 2 cAMP for 6 h increased the expression of miRNA-212, miRNA-183, miRNA-132, miRNA-182 and miRNA-96 and inhibited the expression of miRNA-138 and miRNA-19a ( Hu et al . 2013 ). In mouse Leydig tumor MLTC-1 and murine Y1 adrenocortical cells, it has been demonstrated that inhibition of miRNA-200c is associated with increased vimentin expression and steroid production ( Hu et al. 2017 ). A recent study by Geng et al . (2017) demonstrated that miRNA-150 by targeting StAR protein negatively regulates testosterone production in mouse Leydig cells both in vivo and in vitro . Furthermore, Men et al . (2017) demonstrated that overexpression of H19 in MLTC-1 cells stimulated StAR expression by antagonizing let-7, which inhibits StAR protein expression at the posttranscriptional levels. Gao et al . (2018) provided evidence that dysregulation of mouse testicular Leydig cell miRNA-146a-5p/Mta3 signaling cascade contributes to the pathogenesis of impaired steroidogenesis, at least in part, in response to exposure of testis to the environmental toxin, bisphenol A (BPA). A similar type of study demonstrated that miRNA-6321/Map3k1-regulated the JNK/c-Jun/Nur77 signaling cascade, mediating the Triclosan (TCS)-induced inhibition of steroidogenesis in primary rat testicular Leydig cells and mouse Leydig tumor TM3 cells ( Ha et al. 2018 ).

This review summarizes the current understanding of the role of miRNAs in the regulation of adrenal and gonadal steroidogenesis and related processes. MiRNAs are endogenous single-stranded small RNAs of ~22 nucleotides in length that post-transcriptionally downregulate gene expression. A single RNA could suppress hundreds of genes, and a single gene could be modulated by multiple miRNAs. Although miRNA regulation of steroidogenesis is an emerging field, extensive literature exists about the important role played by miRNAs in the regulation of many other cellular processes such as inflammatory and immune responses, cell-cycle progression and proliferation, differentiation, tissue remodeling, apoptosis and disease pathology. The limited information that is currently available for three steroidogenic tissues, adrenal, ovary and testis, already suggests that miRNAs are putatively involved in virtually every aspect of the steroidogenic process, including receptor-mediated lipoprotein cholesterol delivery to the cell interior, intracellular cholesterol processing and transport to and within the mitochondria, other steroidogenesis-related accessory protein and transcription factors, and steroidogenic enzymes themselves. Table 2 depicts the current understanding of the roles of specific miRNAs in steroidogenesis. The large amount of data generated through genome-wide analyses of steroid producing adrenal and gonadal tissues, particularly ovarian granulosa cells, should provide ample opportunity for the identification and functional characterization of miRNAs that will likely lead to a much greater mechanistic understanding of the posttranscriptional regulation of steroidogenesis. Besides, with the introduction and availability of high-throughput, next-generation RNA sequencing tools, as well as technical advances in sequencing methodologies and newer bioinformatics, future studies should allow investigators to precisely identify novel, or low abundance miRNAs in steroidogenic tissues and their pathophysiological relevance to steroidogenesis. Furthermore, studies to determine the expression and role of miRNAs in pathological settings should lead to identification of miRNAs as biomarkers or validated targets and functions that are likely to greatly aid in the clinical management of steroid hormone-related diseases and/or reproductive diseases.

miRNAs are endogenous single-stranded small RNAs of ~22 nucleotides in length and generated from endogenous hairpin-shaped transcripts ( Bartel 2018 ). miRNAs post-transcriptionally downregulate gene expression by base-pairing to partially complementary sites in one or more target mRNAs, usually in the 3′ UTR, triggering their repression; mechanisms for miRNA-mediated downregulation of gene expression involve some combination of translational repression, mRNA deadenylation, decapping, 5′-to-3′ mRNA degradation and alteration of mRNA stability ( Fabian et al . 2010 , Jonas & Izaurralde 2015 ). miRNAs belong to a group of heterogeneous noncoding RNAs (ncRNAs) including long noncoding RNAs (lncRNAs) such as long intergenic ncRNAs, natural antisense transcripts, pseudogenes and circular RNAs (circRNAs), ribozymes, ribosomal RNAs (rRNAs), transfer RNAs (tRNAs), small nucleolar RNAs (snRNAs), Piwi-associated RNAs (piRNAs), small nuclear RNAs (snoRNAs), telomere-associated RNAs (TERC, TERRA), miRNAs and miRNAs that participate in RNA interference (RNAi or siRNA) ( Esteller 2011 ). Among these ncRNAs, miRNAs are one of the most extensively studied ncRNA species and have emerged as important regulators of gene expression ( Bartel 2018 ). Using computational and experimental approaches, it has been demonstrated that a single miRNA can bind to and regulate on an average, more than 100 mRNAs ( Di Leva et al . 2014 ). In contrast, multiple types of miRNAs can bind to the 3′-UTR of a single mRNA target and collectively control its expression ( Kim 2005 ). Furthermore, single miRNAs can have multiple target sites in the 3′-UTRs, thereby increasing repression efficiency ( Rottiers & Näär 2012 ). In addition, some miRNAs have also been shown to affect multiple targets in linear pathways or interconnected in regulatory networks, thereby causing a large additive effect ( Rottiers & Näär 2012 ). miRNAs are predicted to regulate more than 60% of human protein-coding genes ( Friedman et al . 2009 ). The biogenesis of miRNAs is a complex process involving multiple and distinct pathways ( Ha & Kim 2014 , Bartel 2018 , Gebert & MacRae 2019 ). miRNA sequences are usually located in intergenic or intronic or polycistronic regions of the genome. As illustrated in Fig. 1 , transcription of such sequences by RNA polymerase II generates a single miRNA hairpin precursor or clusters of multiple precursors, called primary miRNA (Pri-miRNAs) and consist of a 5′-cap ( 7 MGpppG) and polyadenylated (AAAAA) tail. Specific promoters facilitate the transcription of intergenic miRNA genes; intragenic miRNA is processed by their own promoters if antisense oriented or by their host-gene promoters if sense oriented ( Lima et al . 2017 ). In some cases, miRNAs are transcribed as a single long transcript called clusters, which may have analogous seed regions, and in this context, they are classified as a family. The biogenesis of miRNA is classified into canonical and non-canonical pathways ( O’Brien et al . 2018 ). The canonical pathway of miRNA biogenesis is the principal pathway by which miRNAs are processed. In this pathway, primary-miRNA (pri-miRNA, ~1–3 kb in length) are transcribed from their genes by RNA polymerase II in the nucleus, and they are typically spliced, capped and polyadenylated. Pri-miRNAs fold into hairpin structures and these pri-miRNAs are cleaved in the nucleus by the microprocessor complex containing a dsRNA-binding protein, DiGeorge Syndrome Critical Region 8 Protein (DGCR8) and an RNase III enzyme, Drosha, into ~70–100 nucleotide-long stem-loop structures, termed precursor-miRNAs (pre-miRNA). The pre-miRNAs are then transported from the nucleus to the cytoplasm by RAN-GTP complex and karyopherin exportin 5 (Exp5). Once in the cytoplasm, the pre-miRNAs undergo an additional processing step by the RNase III enzyme Dicer and transactivation-response RNA-binding protein (TRBP) RNase III enzyme complex to generate ~22 bp miRNA:miRNA* duplex ( Ha & Kim 2014 , Bartel 2018 , Gebert & MacRae 2019 ). This duplex is then loaded into the mRNA-associated multiprotein miRNA-induced silencing complex (miRISC), which includes Argonaute (AGO) 1–4 family of proteins. The mature single-stranded miRNA, termed guide strand, is preferentially retained in the complex. The mature miRNA/RISC complex then binds to complementary sites in the mRNA target to negatively regulate gene expression. The remaining strand, known as the passenger strand, is degraded as a RISC complex substrate. Although relatively less common, multiple non-canonical miRNA biogenesis pathways have been characterized ( O’Brien et al . 2018 ). These pathways have been shown to utilize various combinations of the key proteins involved in the canonical pathway, including Drosha, Dicer, Exp5, and AGO ( O’Brien et al . 2018 ). Recent advances have led to clear understanding that miRNAs contribute to the regulation of a variety of physiologic processes including embryonic development, differentiation and proliferation, apoptosis, metabolism, hemostasis and inflammation ( Bushati & Cohen 2007 , O’Connell et al . 2012 , Rottiers & Näär 2012 , Dong et al . 2013 , Alberti & Cochella 2017 , Bartel 2018 , Gebert & MacRae 2019 ). There is also ample evidence available that miRNAs play a key role in the pathogenesis of numerous diseases including cancer, hepatitis, cardiovascular diseases and metabolic diseases ( Esteller 2011 , Sayed & Abdellatif 2011 , Rottiers & Näär 2012 , Di Leva et al . 2014 , Price et al . 2014 , Adams et al . 2017 , Lin & He 2017 , Rupaimoole & Slack 2017 ). Moreover, because of their ability to contribute to the pathogenesis of various diseases, miRNAs have emerged as potential therapeutic targets to combat these diseases ( Rupaimoole & Slack 2017 ). Steroid hormones are involved in virtually every cellular process and are essential to maintain metabolic homeostasis ( Table 1 ). In steroidogenic cells of the adrenal gland (zona glomerulosa cells, zona fasciculata cells and zona reticularis cells), ovary (granulosa cells, theca cells and luteal cells) and testis (Leydig cells), the overall rate of steroid hormone production is controlled by trophic hormones (ACTH, LH, FSH, or angiotensin II (Ang II)/potassium (K+)) at two levels: (a) at the level of cellular cholesterol (substrate) mobilization and its transport to and within the mitochondria for side chain cleavage by CYP11A1 (P450scc) occurring within minutes, reflecting the acute regulation of steroidogenesis ( Miller & Bose 2011 , Hattangady et al . 2012 , Shen et al . 2016a , Selvaraj et al . 2018 ) and (b) at the level of gene transcription occurring between hours and days, reflecting chronic regulation of steroid hormone biosynthesis ( Simpson & Waterman 1988 , Simpson et al . 1992 , Payne & Youngblood 1995 , Romero et al . 2010 , Hattangady et al . 2012 , Shen et al . 2016a ). In addition, in recent years, it has been increasingly clear that steroidogenic pathways are also subject to post-transcriptional and post-translational regulations including processes such as phosphorylation/dephosphorylation and protein‒protein interactions, as well as the involvement of specific miRNAs ( Shen et al . 2016b ). This review presents a detailed description of miRNA regulation of steroid hormone biosynthesis, with emphasis on adrenal and gonadal steroidogenesis.

Testicular Leydig cells are the principal source of testosterone; in fact, over 95% of the testosterone is secreted by the Leydig cells. In the ‘classic’ pathway, LH binds to its receptor on the Leydig cell surface, which is also coupled to the cAMP/PKA signaling cascade and initiates testosterone production ( Matsumoto & Bremmer 2016 ). In this classic pathway, pregnenolone is converted through the Δ 5 pathway by P450c17 to 17α-hydroxypregnenolone and then to DHEA with the first reaction catalyzed by its 17α-hydroxylase activity using P450 oxidoreductase (POR) as a coenzyme and the second reaction by the 17,20 lyase activity in the presence of POR and cytochrome b5 (CYB5). DHEA is then converted to testosterone through androstenedione or androstenediol catalyzed by 3βHSD2 and 17βHSD3/5, respectively ( Fig. 4 ). In the Δ 4 pathway, pregnenolone is converted sequentially to 17α-hydroxyprogesterone, androstenedione, and testosterone. It should be noted that biosynthesis of testosterone in human testis proceeds predominantly through the Δ 5 pathway and very little testosterone is generated via the Δ 4 pathway ( Flück & Pandey 2014 , Matsumoto & Bremmer 2016 ). The most potent endogenous androgen 5α-dihydrotestosterone (DHT) is formed from testosterone catalyzed by steroid 5α-reductase type 2 (SRD5A2). DHT formation is low in testis, but high levels are produced in genital skin and prostate. There is an alternative pathway or ‘backdoor’ pathway that was first described for androgen production in the tammar wallaby pouch young testis and subsequently in the immature mouse testis ( Auchus 2004 , Miller & Auchus 2019 ). In this pathway, androgen synthesis (DHT) occurs through conversion of 17α-hydroxyprogesterone to DHT without going through androstenedione or testosterone as intermediates ( Fig. 4 ). Evidence is accumulating that this pathway plays an important role in hyperandrogenic disorders ( Miller & Auchus 2019 ) such as 21-hydroxylase deficiency-linked congenital adrenal hyperplasia ( Kamrath et al . 2012 ), polycystic ovary syndrome (PCOS) ( Martin et al . 2017 ), some virilized female newborns with P450 oxidoreductase deficiency ( Shackleton et al . 2004 , Homma et al . 2006 , Krone et al . 2012 ) and physiologic ‘minipuberty of infancy’ ( Dhayat et al . 2017 ). Furthermore, AKR1C2 and AKR1C4 gene mutation studies ( Flück & Pandey 2014 ) led to the conclusion that both ‘classic’ pathways of androgen synthesis and ‘backdoor’ are required for normal male genital development ( Biason-Lauber et al . 2013 ). A very recent study demonstrated that the major human ‘backdoor’ androgen is androsterone and is primarily derived from placental progesterone ( O’Shaughnessy et al . 2019 ).

Few studies have directly attempted to identify steroidogenic enzymes, protein factors that mediate cholesterol delivery to cells and its transport to and within the mitochondria for steroidogenesis and gonadotropin receptors, as potential targets for miRNAs. Using human adrenal tissue samples and human H295R adrenocortical cells, CYP11A1, CYP11B1, CYP17A1 and CYP19A1, have been identified as potential targets for various miRNAs ( Table 2 ). Likewise, in mouse Y1 adrenocortical tumor cells, miRNA-132 was shown to directly target cholesterol transport protein, StAR and methyl CpG-binding protein with a direct relevance to steroidogenesis ( Hu et al. 2017 ). Relatively more information is currently available about the ovarian steroidogenic targets for several miRNAs ( Table 2 ). These include gonadotropin receptors (FSHR, LHCGR), steroidogenic enzymes (CYP19A1), transcription factors (Creb1, Foxl2, LRH-1, SF-1, E2F1, RUNX2, SREBP-1a, SREBP-2), cholesterol transport proteins (StAR) and signaling proteins (HRas, EFNA3, IGF-1, MAP3K8) ( Table 2 ). Likewise, in testicular Leydig cells, miRNA-150 and let-7b target cholesterol transport protein, StAR ( Geng et al. 2017 , Men et al. 2017 ), whereas miRNA-200c targets vimentin, a protein involved in intracellular cholesterol transport ( Hu et al. 2017 ). Two miRNAs, miRNA-125a and miRNA-455, directly target testicular Leydig cell SR-B1 ( Hu et al. 2012 ), and miRNA-6321 targets Map3k1 ( Ha et al. 2018 ) ( Table 2 ).

Although this review focuses on miRNA regulation of adrenal and gonadal steroid hormone biosynthesis, to familiarize the readers, a brief overview regarding critical events involved in steroidogenesis is warranted. The steroid hormones are primarily synthesized by endocrine glands, adrenal cortex, and gonads (ovary and testis) and by placenta during pregnancy. There are five major classes of steroid hormones: (i) cortisol is the major glucocorticoid in humans and most mammals (a similar analog of this hormone, designated as corticosterone, is synthesized by rodents, birds and reptiles); (ii) aldosterone is the major mineralocorticoid in humans; (iii) androgens, such as testosterone and dihydrotestosterone; (iv) estrogens, mainly estradiol and estrone; and (v) progestins, such as progesterone ( Miller & Auchus 2011 ). The outermost zona glomerulosa layer of the adrenal cortex produces aldosterone; there are two major regulators of aldosterone biosynthesis: angiotensin II (AngII) and high extracellular potassium levels and a secondary regulator, adrenocorticotrophic hormone (ACTH) ( Williams 2005 , Hattangady et al . 2012 , Bollag 2014 ). The zona fasciculata, the middle and largest layer of the adrenal cortex is the major site of cortisol (glucocorticoid) production ( Turcu et al . 2014 , Stewart & Newell-Price 2016 ). The innermost zone of the adrenal cortex, zona reticularis, predominantly secretes DHEA and DHEAS and to a lesser extent androstenedione ( Turcu et al . 2014 , Stewart & Newell-Price 2016 ). These zones are primarily regulated by ACTH. Follicle-stimulating hormone (FSH) controls progesterone and estrogen synthesis in ovarian granulosa cells, whereas luteinizing hormone (LH) regulates progesterone synthesis in luteinized granulosa cells and luteal cells, androgen production in ovarian theca-interstitial cells ( Hoffman et al . 2012 ), and testosterone synthesis in testicular Leydig cells ( Matsumoto & Bremmer 2016 ). Cholesterol is the principal precursor for the biosynthesis of all steroid hormones in the adrenal gland, ovaries and testis ( Azhar & Reaven 2002 , Miller & Bose 2011 ). In these glands, there are multiple sources of cholesterol, which could potentially supply cholesterol substrate needed for steroidogenesis; the cholesterol can be synthesized de novo from acetate, directly obtained from circulating lipoproteins (low-density lipoprotein (LDL) and high-density lipoprotein (HDL)) or acquired via the hydrolysis of stored cytoplasmic cholesteryl ester lipid droplets ( Azhar & Reaven 2002 ). In addition, cholesterol for steroidogenesis can also be derived from the plasma membrane ( Deng et al . 2019 ). The current evidence suggests that plasma lipoproteins are the major source of cholesterol for steroidogenesis in adrenal gland, ovary and under certain conditions in testicular Leydig cells ( Azhar & Reaven 2002 , Miller & Bose 2011 , Shen et al . 2018 ). Human steroidogenic cells obtain cholesterol from circulating LDL via the LDL-receptor/endocytic pathway ( Miller & Bose 2011 ), whereas HDL particles supply cholesterol to mostly rodent steroidogenic cells via the SR-B1/selective pathway ( Shen et al . 2018 ). The first common step in the production of all steroid hormones is the enzymatic conversion of cholesterol (a carbon27 (C27) steroid) to pregnenolone (the primary C21 product) catalyzed by cholesterol side-chain cleavage (P450scc) enzyme, encoded by the CYP11A1 gene and which is under the control of trophic hormones (ACTH, LH, FSH or AngII/high K + depending on the tissue) using cAMP or calcium as second messenger. The P450scc is localized in the inner mitochondrial membrane (IMM) along with its cofactor protein ferredoxin reductase. The key steps in this reaction are the mobilization and transport of lipid droplet-associated cholesterol ester (cholesterol is esterified by acyl-coenzyme A:cholesteryl acyltransferase (ACAT), stored in lipid droplets, and released by trophic hormone stimulation of hormone-sensitive lipase (HSL)) to the outer mitochondrial membrane (OMM) mediated by HSL, cytoskeleton, SNAREs and possibly other factors ( Shen et al . 2003 , 2012 , 2016 b , Kraemer et al . 2004 , 2017 , Lin et al . 2016 ), and subsequent cholesterol transport from the OMM to the IMM and eventual loading of the precursor into the active site of P450scc. The movement of cholesterol from OMM to IMM is considered the rate-limiting step in all steroid hormone production and is primarily mediated by the StAR protein (encoded by the STARD1 gene) ( Miller 2017 , Selvaraj et al . 2018 ). Trophic hormones rapidly induce the expression of StAR protein, which then facilitates inter mitochondrial membrane cholesterol transfer and initiate steroidogenesis. Although a role for translocator protein (TSPO)/peripheral benzodiazepine receptor (PBR) has also been suggested in the movement of cholesterol from OMM to IMM ( Papadopoulos et al . 2018 ), however, based on the current evidence, there is no consensus whether or not TSPO/PBR is critically involved in intermembrane cholesterol transfer ( Costa et al . 2018 , Selvaraj et al . 2018 ). Once P450scc-catalyzed pregnenolone is formed, it can then be transformed to progesterone, androgens, estrogens, and corticosteroids in a tissue-specific manner.