On the Role of the Translocator Protein (18-kDa) TSPO in Steroid Hormone Biosynthesis

Morohaku et al (1) report that the translocator protein (18-kDa) TSPO, previously known as peripheral-type benzodiazepine receptor, is not required for steroid hormone biosynthesis. The study was based on analysis of a conditional knockout of TSPO in Leydig and Sertoli cells. The findings were surprising and were coupled to a definitive statement implying that what was suggested for Leydig cells is the case for all steroid-synthesizing cells in the body (1). To understand the reported findings and place them within the context of the TSPO and steroidogenesis literature, I will try to address five questions: 1) What are the complexities of cholesterol mobilization and transport? 2) What are the data indicating that TSPO is integrally involved in steroid formation? 3) Are there limitations of the Morohaku paper? 4) Where do the discrepancies lie? and 5) Where do we go from here? The adrenal, gonads, placenta, and brain are the major tissues that are able to make steroids from the precursor cholesterol (2, 3). In the past 2 decades, thymus and skin also have been identified as having this ability (4, 5). Adrenal and gonadal steroidogenesis is under hormonal control; placental steroidogenesis is temporal and hormone independent. At yet, no hormone has been identified that induces brain steroid hormone formation. Steroidogenesis involves the metabolism of a series of substrates, the first being cholesterol, by enzymes distributed in the mitochondria and endoplasmic reticulum (2). The first step occurs in mitochondria where the enzyme cytochrome P450 (CYP)11A1 is localized in the matrix side of the inner mitochondrial membrane (IMM). Cholesterol is a highly lipophilic compound that cannot freely diffuse to IMM. Thus, there are mechanisms facilitating its movement from various intracellular stores to the outer mitochondrial membrane (OMM) where it stays segregated from the structural cholesterol, its translocation from OMM to IMM, and subsequent loading onto CYP11A1. In the adrenal and gonads, cholesterol translocation is induced by hormones and is limiting, controlling the availability of cholesterol to CYP11A1. For more than 50 years, there has been intense research to identify a single hormone-inducible protein that would be capable of binding and transferring cholesterol from intracellular stores to OMM, moving cholesterol from OMM to IMM across the aqueous intermembrane space, and loading cholesterol onto CYP11A1. No one protein has been identified that carries out all of these steps, but proteins have been identified which, functioning together in a complementary manner and within a large protein complex, are able to accomplish many of these functions. These include the steroidogenic acute regulatory protein, TSPO, and voltage-dependent anion channel (VDAC)1, among others (6–8). Snyder and colleagues (9) demonstrated the wide distribution of TSPO throughout the body, its abundance in steroid-producing cells of the adrenal and testis, and its localization at OMM. The findings that hypophysectomy, but not adrenalectomy or castration, controlled the expression of TSPO in the testis and adrenal, respectively, suggested the hormonal regulation of TSPO and its possible relationship to testicular and adrenal steroidogenesis (10). The identification of a natural polypeptide des-(Gly-Ile)-endozepine, also known as diazepam-binding inhibitor (DBI), which was able to stimulate cholesterol transport in isolated mitochondria and to bind TSPO, brought together the fields of steroidogenesis and TSPO (11, 12). These findings led to studies of the stimulatory effects of well-characterized TSPO-binding compounds on corticosteroid and androgen formation by adrenal cortical and testicular Leydig cells, respectively (13–15). Indeed, significant, 1.5- to 3.0-fold stimulation of basal steroid synthesis was found by all steroidogenic cells tested, including primary cells as well as cell lines, and also isolated mitochondria. Moreover, TSPO functional antagonists and partial agonists were able to block hormone- and TSPO ligand-induced steroid formation in both primary cells and cell lines (16–18). Consistent with this, DBI was shown to increase steroid formation when applied to isolated mitochondria, whereas its knockdown blocked steroidogenesis in Leydig cell lines (13). Interestingly, DBI levels in Leydig cells were reduced in hypophysectomized animals (19), and 2 knockout DBI (ACBD1/ACBP) mouse models resulted either in a lethal phenotype (20) or metabolic defects (21). TSPO was pharmacologically and biochemically characterized, its topography in OMM established, and its ability to move cholesterol through the OMM demonstrated in silico (22–24). TSPO was then shown to be a high-affinity cholesterol-binding protein, and mutations in its cholesterol recognition amino acid consensus domain were shown to eliminate its ability to bind cholesterol (25). Knocking down TSPO in various cell lines led to cell death, most likely because of TSPO's role in mitochondrial physiology. This also was true of the only Tspo-knockout cell line that we developed, which was unable to sustain steroid formation and could not be kept alive (13). The Tspo global knockout made in the mid-1990s was embryonic lethal. Unfortunately, the reasons for this lethality were not evaluated. TSPO drug ligands, administered in vivo, were shown to significantly affect circulating and tissue steroid levels (13) in a hormone-independent manner, as indicated by the following findings: First, TSPO ligands increased circulating corticosteroid levels in hypophysectomized animals (26). Second, in vitro and in vivo studies of the Brown Norway rat model of aging demonstrated that the pharmacologic activation of TSPO in aged cells and aged animals increased testosterone formation and circulating levels (27), suggesting that this approach might be used therapeutically for the gonadotropin-independent induction of testosterone in cases of both primary and secondary hypogonadism. Third, TSPO drug ligands provided the first pharmacologic means to regulate steroid formation in the brain (14, 15, 28) and was used to alleviate neurologic and psychiatric disease symptoms both at the central (29, 30) and peripheral nervous system in animal models and patients (14, 30). Indeed, the TSPO ligand XBD173 was found to exert anxiolytic properties in cholecystokinin-4-induced anxiety in healthy volunteers (29). XBD173 has reached advanced clinical trials in patients with generalized anxiety disorders and the TSPO ligand SSR180575 for the treatment of diabetic neuropathy (clinicaltrials.gov; NCT00108836 & NCT00502515). There are no published outcomes from these studies. However, the recent identification of the common rs6971 TSPO polymorphism, Ala147Thr amino acid substitution, affecting TSPO ligand binding in humans, would indicate a variable response by the patients treated with the compounds (31, 32). Interestingly, the rs6971 TSPO polymorphism has been shown also to influence pregnenolone production in circulating lymphomonocytes of healthy volunteers (33). The observations made of the Leydig/Sertoli-cell targeted knockout mice in this paper, although puzzling, may well be correct. That said, careful analysis of the reported results raises some critical questions. 1) Why were there no metabolic effects? Although it would not be surprising to see little or no biological effect of knocking out TSPO in cells that have little TSPO, such as those of kidney, liver, and spleen, absence of metabolic effect after knocking out TSPO in Leydig cells is surprising; TSPO has been estimated to make up 5%-10% of steroidogenic cell OMM protein (9). 2) The results presented are from one out of the 6 embryonic stem cell (ESC) clones generated; there is no information as to whether other mouse lines with germ line transmission were obtained and studied. 3) The data presented appear to indicate that exon 2 might still be present in the TspocΔ/Δ mice. Perhaps this is not surprising because the use of Amhr2 as a promoter also failed to completely knock out the androgen receptor in Leydig and Sertoli cells (34). This is relevant because the minimal level of TSPO needed for function is unknown. 4) Unfortunately, there is lack of wild-type, Cre transgene, and heterozygous knockout mouse controls. In our experience, the introduction of the loxP sites can alter Tspo mRNA levels in various tissue compared with true wild-type mice. 5) TspocΔ/Δ male mice were used to maintain the colonies, and this may have selected for unusual animals that can compensate for Tspo loss. 6) There is lack of Tspo mRNA detected by quantitative PCR in RNA extracts from whole testis, indicating complete absence of Tspo mRNA in the TspocΔ/Δ mouse testes. This is unexpected because TSPO is not only expressed in Leydig and Sertoli cells but also in germ cells and endothelial cells (35), where it should not have been knocked out. 7) Careful perusal of the data reveals great variability in circulating testosterone levels in the TspocΔ/Δ mice. Indeed, 3 of the generated TspocΔ/Δ mice were shown to have basal circulating levels of testosterone up to 30 ng/mL (or more). These values correspond to those seen with hormone treatment. It is likely that the results obtained by Morohaku et al (1) are correct and the limitations presented above will be resolved. Thus, taking the data at face value, it might be concluded that TSPO is not involved in Leydig cell hormone-induced steroid formation. However, the findings do not address the role of TSPO in controlling steroid hormone biosynthesis in other steroidogenic cell types, other species, or hormone-independent steroidogenesis. Certainly the paper does not refute or dismiss more than 30 years of work by numerous laboratories around the world, in academia and industry, that have shown that pharmacologic activation of TSPO can stimulate steroidogenesis in primary cells, cell lines, and organisms. Rather, this work does raise questions about whether all steroidogenesis is “monothetic,” ie, via one mechanism. TSPO genes are present in almost all organisms (36). Indeed, TSPO is a 3.5B year-old protein, possibly part of LUCA (last universal common ancestor), suggesting strong selection pressure to maintain this protein for basic cellular function. Thus, it is likely that deletion of this protein would have detrimental effects on mitochondrial function, suggesting that cells would adapt to compensate TSPO loss by transferring its function to another protein, in order to preserve cell function. Such a possibility is illustrated by the cases of Vdac1−/− and Vdac3−/− global null mice, and Vdac1−/−, Vdac2−/−, and Vdac3−/− cell lines, which failed to show changes in mitochondrial permeability, transition pore activity, and Bcl-2-driven cell death compared with wild-type mitochondria (37, 38). These observations led to the speculation that other β-barrel OMM proteins could functionally compensate for VDAC loss. Is it possible that TSPO could have been replaced by another α-helical protein or is it just a redundant remnant of evolution? Because cholesterol transport from OMM to IMM was shown only when TSPO was activated by specific drug ligands (39), TSPO might be an OMM bystander in the mitochondrial cholesterol transport process, not playing a role unless activated by high-affinity drug ligands. However, how could this explain the inhibitory effect seen in hormone-dependent steroidogenesis when the cholesterol recognition amino acid consensus domain of TSPO is blocked (17, 40)? Another attractive interpretation is that TSPO controls steroidogenic cholesterol segregation and availability in OMM for steroid formation. In this case, TSPO removal would result in the loss of cholesterol segregation and thus in the direct passage of cholesterol through OMM to CYP11A1 at IMM, resulting in unregulated steroid formation. Indeed, this could explain the elevated and nonphysiological levels of testosterone observed by the authors in some of the TspocΔ/Δ mice. Results obtained with TSPO drug ligands in vitro and in vivo suggest that TSPO may be important in controlling steroid formation in a hormone-independent manner. Thus, TSPO provides a physical means to induce steroidogenesis via its ability to bind cholesterol, location in OMM, and physical association (directly or indirectly) with the steroidogenic machinery. In this way, TSPO is a valid pharmacologic target to induce, reduce, or restore normal rates of steroid formation in cases of abnormal androgen synthesis, adrenal function, and neurosteroid formation. It is also possible that TSPO plays a more general role in cholesterol movement from OMM to IMM needed for membrane biogenesis. This has been proposed for rapidly proliferating cells, such as tumor cells in which TSPO is overexpressed (40–42), or activated microglia in brain (43). In cancer, blocking TSPO or reducing TSPO expression leads to changes in the rates of cell proliferation and tumor growth in vivo (40, 42). If this is the case, it would be true that results obtained in steroidogenic cell lines, all of which are of tumor origin, could be misleading. However, the effects of TSPO drug ligands on steroid formation also have been shown with primary cells and in vivo. The results presented in the manuscript by Morokahu et al suggest our ignorance of the physical mechanism(s) regulating hormone-dependent androgen formation in the testis. Previous studies indicated that cholesterol transport into mitochondria is not a single-protein event but rather relies on a multiprotein complex (7, 8). However, the chain of events regulating cholesterol movement from intracellular stores to and then into the mitochondria is still unknown. In recent studies, we were able to physically and functionally trap at least some of the involved proteins together in a complex comprised of steroidogenic acute regulatory protein and OMM and IMM proteins, the latter including TSPO, VDAC1, and the AAA+ ATPase ATAD3 (45, 46). The details of how these and perhaps other proteins interact to mediate hormone-dependent and hormone-independent effects involved in steroid formation in given steroidogenic cell types remain uncertain. Moreover, proteomic analysis of steroidogenic cell mitochondria make it evident that there are numerous additional proteins that also could play roles both in cholesterol coming to, and then translocating into, the mitochondria. Additionally, we know little about how cholesterol is coming to mitochondria, how the flow is maintained, how the cell senses the lack or excess of steroid synthesis, and how hormones regulate these processes. There is evidence that organelle interactions, the cytoskeleton, membrane lipid reorganization, ion exchanges, and/or second messengers might be involved. However, the mechanisms involved are unknown. It is premature, at best, to attribute all of this to any one protein, or to rule out the role of any one protein. Answers to biological questions typically are obtained by combining a multitude of positive and negative experiments and by using various methodologies, cell types, and animal models. Of course, the ideal model is the human. There is no single experiment that can provide the “definitive” answer to any research question. However, such experiments can stimulate and guide us toward the next steps. With this in mind, the paper by Morohaku et al provides clear evidence that further studies are needed to identify the cellular and molecular mechanisms driving cholesterol import into mitochondria for gonadal steroidogenesis. I could not agree more. There is much left to learn. Results from our publication clearly show that mice deficient of TSPO in testicular Leydig cells do not have any deficits in their ability to produce testosterone (46). In our short response to this counterpoint, we will focus only on a few specific points that Papadapoulos has indicated are “limitations of the Morohaku paper,” because other comments and speculations will eventually be sorted out by future research in this field. Our data definitely do not indicate that a TSPO transcript containing exon 2 is still present in TSPOcΔ/Δ mice testes. In our genetic design, loxP sites flank a region that includes both exon 2 and exon 3. Thus, a cre recombination event would remove both exon 2 and exon 3. Given the complete absence of exon 3 in RT-PCR, the extremely weak signal observed in the reaction for exon 2 is, in all probability, nonspecific. Our quantitative PCR data support this probability, because we did not detect amplification with a primer pair that annealed to exon 2 and exon 3 (specific amplicon was detected using a hydrolysis probe). The failure to detect TSPO transcripts in TSPOcΔ/Δ mice testes provides further proof that TSPO is not expressed in testicular germ cells. In a previous study, we examined the developmental expression of TSPO in murine reproductive tissues and concluded that TSPO is not expressed in germ cells of the adult testis (35). That Papadopoulos has cited this same study to claim the direct opposite to what we published is highly inaccurate. In our current study we are disproving a long-standing dogma surrounding TSPO function. We hope that our results will lead to critical reevaluation of existing models and advance understanding of mechanisms underlying regulation of steroid hormone production. Disclosure Summary: The author has nothing to disclose. cytochrome P450 diazepam-binding inhibitor inner mitochondrial membrane outer mitochondrial membrane translocator protein (18-kDa) voltage-dependent anion channel.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.