The translocator protein (TSPO) has been proposed to act as a key component in a complex important for mitochondrial cholesterol importation, which is the rate-limiting step in steroid hormone synthesis. However, TSPO function in steroidogenesis has recently been challenged by the development of TSPO knockout (TSPO-KO) mice, as they exhibit normal baseline gonadal testosterone and adrenal corticosteroid production. Here, we demonstrate that despite normal androgen levels in young male TSPO-KO mice, TSPO deficiency alters steroidogenic flux and results in reduced total steroidogenic output. Specific reductions in the levels of progesterone and corticosterone as well as age-dependent androgen deficiency were observed in both young and aged male TSPO-KO mice. Collectively, these findings indicate that while TSPO is not critical for achieving baseline testicular and adrenal steroidogenesis, either indirect effects of TSPO on steroidogenic processes, or compensatory mechanisms and functional redundancy, lead to subtle steroidogenic abnormalities which become exacerbated with aging.
The synthesis of pregnenolone (P5) from cholesterol, catalyzed by the P450 side-chain cleavage (P450scc), is the rate-limiting step in steroidogenesis . The rate of this reaction is determined by substrate availability, which is controlled by mitochondrial cholesterol import . Although essential for life, the molecular mechanisms mediating mitochondrial cholesterol transfer for steroidogenesis remain controversial. The translocator protein (TSPO), which binds cholesterol with high affinity and is enriched in the mitochondria of steroidogenic organs, has been proposed to mediate mitochondrial cholesterol importation for steroidogenesis [3,4]. A role for TSPO in steroidogenesis was first suggested pharmacologically, with a range of structurally and chemically distinct, high-affinity TSPO ligands found to stimulate steroidogenesis both in vitro and in vivo in various cell types and tissues [5–10]. More recently, a common TSPO polymorphism altering the cholesterol-binding domain has been associated with impaired cholesterol metabolism and reduced ability to synthesize P5 [11,12]. Structural studies have provided insights into the potential mechanism of TSPO ligand-mediated cholesterol import, finding that ligand binding stabilizes the tertiary structure of TSPO , hypothesized to facilitate cholesterol movement via a sliding mechanism along the external surface of TSPO .
However, the recent development of testis-specific and global TSPO knockout (TSPO-KO) mice has demonstrated that TSPO is not essential for testicular testosterone (T) or adrenal corticosteroid (CORT) synthesis [15–17], except perhaps under acute stress . Evidence of compensatory changes in the testis of TSPO-KO mice, including depleted cholesterol stores and increased expression of the luteinizing hormone receptor , has led to the suggestion that TSPO may be functionally redundant [19,20]. However, others have argued that these changes may be an indirect consequence of steroidogenesis-independent functions of TSPO leading to altered cholesterol availability, rather than evidence of compensation and redundancy [21,22].
Despite the controversy surrounding TSPO function in steroidogenesis, widespread interest in the potential therapeutic application of steroidogenic TSPO ligands remains. This includes for the treatment of androgen deficiency in the aging male [20,23,24], which is associated with dysfunction and disease in androgen-responsive tissues including bone, muscle, adipose and the brain . In aging rats, TSPO ligands can stimulate androgen production restoring T levels to those observed in young rats, even though the testes are insensitive to stimulation with the luteinizing hormone . Therefore, steroidogenic TSPO ligands may offer an alternative to T replacement therapy with reduced risks of adverse events such as prostate cancer and cardiovascular disease  through the stimulation of T production at physiological concentrations.
Here, we investigate the effect of TSPO deficiency on testicular and adrenal function in not only in healthy, young adult male mice but also in aging mice. The testis provides a good model for investigating the potential role of TSPO in gonadal steroidogenesis since the types of steroids produced do not change, unlike the ovary where levels of steroidogenic enzymes and types of secreted steroids fluctuate throughout the reproductive cycle . Using liquid chromatography tandem mass spectrometry (LC–MS/MS), we measure serum levels of androgens (T & 5α-dihydrotestosterone or DHT) and corticosteroids (CORT and aldosterone or ALDO), as well as their major bioactive precursors, P5 and progesterone (P4). As an indicator of steroidogenic flux, we examine precursor–metabolite relationships, an approach used to detect steroidogenic abnormalities both clinically and in animal models of disease [28–30]. The aim of these studies is to examine the potentially complex changes in the steroidogenic cascade resulting from loss of TSPO function, which may facilitate reconciliation of genetic and pharmacological studies into TSPO function in steroidogenesis.
Materials and methods
For generation of TSPO-KO mice, the targeting construct was electroporated into C57BL/6 embryonic stem cells for homologous recombination. Correctly targeted cells were injected into C57BL/6 blastocysts generating chimeric mice. Female TSPOfl/fl mice were crossed with male CaMKIIa-Cre mice [RIKEN C57BL/6-TgN(a-CaMKII-nlCre)/10] expressing Cre in germ cells, generating heterozygous global TSPO-KO mice. To remove the Cre transgene, animals were bred to wild-type (WT) C57BL/6J mice (CLEAR, Japan), and the resulting heterozygous mice were used to breed homozygous offspring (TSPO-KO). Homozygous TSPO-KO mice on a C57BL/6J background were used to maintain the colony and generate mice used for the experiment. Mice were genotyped by PCR using DNA isolated from tail biopsies (forward primer probe: ATAGATGGAAACAGGATTGAAGTGA; reverse primer probe: ATTGTGTTTAAGTGCTTCAGTCCAT). WT C57BL/6J mice were purchased (CLEAR, Japan) and maintained at the National Institute of Radiological Sciences vivarium facilities with food and water available ad libitum.
For serum steroid hormone measurements, mice were anesthetized with isoflurane [3% (v/v) for induction and 1.5% (v/v) for maintenance], and blood was collected via cardiac puncture prior to euthanasia (4 months old, n = 10/group; 10 months old, n = 5/group; 22 months old, n = 5/group). For autoradiography and immunohistochemistry, anesthetized mice (4 months old, n = 3/group) were perfused with 0.9% saline, then organs immediately frozen and stored at −80°C until sectioning by cryostat (20 µm; HM560, Carl Zeiss).
All experimentation was approved by the National Institutes for Quantum and Radiological Science and Technology Institutional Animal Care and Use Committee.
18F-FEDAA1106, a selective imaging agent for TSPO, was radiosynthesized according to the previously described methods . Sections were incubated in 50 mM Tris–HCl with 5% ethanol (pH 7.4) containing 0.5 nM 18F-FEDAA1106 (specific activity 135.9 GBq/µM) in the presence or absence of 10 µM unlabeled PK-11195, a widely used TSPO ligand. Sections were then washed in 50 mM Tris–HCl (pH 7.4) containing 5% ethanol, rinsed in water and dried before contacting to an imaging plate (BAS-MS; Fuji Film), which was digitally scanned (BAS5000 system, Fuji Film).
Sections were fixed in 4% paraformaldehyde in PBS (pH 7.4) for 10 min and washed in PBS prior to Oil red O staining or immunolabeling. Oil red O (Sigma–Aldrich) staining was performed in 60% isopropyl alcohol for 30 min, prior to counterstaining with hematoxylin (Sigma–Aldrich) for 5 min. For immunolabeling, sections were incubated with blocking reagent (PerkinElmer) for 1 h at room temperature and then incubated overnight at 4°C with anti-TSPO (1 : 1000; Abcam ab109497). Sections were then washed three times in PBS and incubated for 1 h at room temperature with anti-rabbit IgG biotin (1 : 1000). Immunostaining was visualized using tetramethylrhodamine-labeled tyramide signal amplification (PerkinElmer). Sections were mounted in Vectashield antifade mounting medium with 4′,6-diamidino-2-phenylindole, dihydrochloride (DAPI, Vector Laboratories) and images were captured with a fluorescence microscope/digital camera (BZ-X700, Keyence).
LC–MS/MS assessment of steroid hormones
Samples and internal standards were diluted in 0.1 M KH2PO4 and then added to 4 ml of methyl tert-butyl ether (MTBE) for extraction. The organic phase was dried and purified with a solid-phase extraction cartridge (adande:1 TM PAX0301, Shiseido, Tokyo, Japan). The cartridge was washed with 1 ml of distilled water, 1 ml of methanol/distilled water/acetic acid (45 : 45 : 1), and 1 ml of 1% pyridine, before eluting the steroids with 1 ml of methanol/pyridine (100 : 1). The purified sample was dissolved in 0.5 ml of acetate/hexane/acetic acid (15 : 35 : 1) and applied to a preconditioned InertSep SI cartridge. The fraction containing P4 was eluted with 2.5 ml of MTBE solution, and the fraction containing P5, T and DHT was eluted with 2.5 ml of acetone–hexane (7 : 3). Fractions were dried, and P4 was converted to its O-ethyl oxime derivatives. The derivatives were then dissolved in acetonitrile/distilled water (2 : 3) and used for LC–MS/MS analysis.
The LC–MS/MS system consisted of a API 5000 triple-stage quadruple mass spectrometer (AB Sciex, Foster City, CA) equipped with a Shimadzu HPLC system (SCL-10Avp system controller, LC-20AD pump, SIL-HTc column oven, CTO-20A auto-sampler; Shimadzu Co. Ltd, Kyoto, Japan). The MS was operated with electrospray ionization in the positive ion mode. T and P4 standards were purchased from Tokyo Chemical Industry (Tokyo, Japan). P5 was purchased from Sigma–Aldrich Corp. (St. Louis, MO, U.S.A.). Data were analyzed using Analyst 1.6.1 (AB Sciex, MA, U.S.A.). The SRM transitions of the derivatives of P5, P4, T and DHT were m/z 422/281, 400/300, 394/253 and 396/203, respectively. The accuracy of all steroid measurement was between 92.5 and 109.0%.
Data were analyzed by Student's t-test (equal variance not assumed) or two-way ANOVA using the Statistical Package for Social Sciences (SPSS, version 11.5; SPSS, Inc., IL, U.S.A.). Levine's and Shapiro–Wilk tests were used to assess homoscedasticity and normality. Non-parametric data were transformed prior to ANOVA analysis. Spearman's correlation was used to examine linear and non-linear relationships between steroidogenic precursors and metabolites. Extreme values were identified using Grubb's test, but were included in statistical analyses. Significance was set at a threshold of P < 0.05 and pair-wise comparisons were made for significant interactions. Data are presented as mean ± SEM unless otherwise specified.
Generation of global TSPO-KO mice
Global TSPO-KO mice were generated through the targeted deletion of exons 2 and 3 (Figure 1A). Deletion of TSPO was confirmed by PCR and in vitro autoradiography in sections from the testis, ovary, adrenal, kidney, liver and heart using the TSPO radioligand 18F-FEDAA1106 (Figure 1). Immunofluorescence analysis confirmed the complete absence of TSPO in the testis and the adrenal TSPO-KO mice (Figure 2A,B, top panel). No differences in cellular morphology or neutral lipid accumulation assessed by Oil red O staining were observed between WT and TSPO-KO mice in either the testis or adrenal (Figure 2A,B, bottom panel).
Validation of TSPO-KO mice.
TSPO and steroidogenesis in the mouse testis and adrenal.
TSPO-KO perturbs testicular and adrenal steroidogenesis in young mice
To evaluate the effect of TSPO deficiency on steroidogenesis, concentrations of the key components of the testicular and adrenergic steroidogenic pathways (Figure 2C,D), including androgens (T and DHT), corticosteroids (CORT and ALDO) and their precursors (P5 and P4), were measured in the serum of young adult mice (4 months old, n = 10/group). As an indicator of steroidogenic capacity, we examined the sum of all measured steroidogenic metabolites, which are generated from cholesterol in a series of sequential reactions with an 1 : 1 reaction stoichiometry. Overall, a significant decrease in the total quantity of steroid hormones measured in TSPO-KO mice was observed (P5 + P4 + T + DHT + CORT + ALDO; t = 4.0, P = 0.001; Figure 3A). The most abundant hormone, CORT, which represented more than 90% of all measured steroidogenic metabolites, was depleted by more than 40% in TSPO-KO mice (t = 4.2, P = 0.001; Figure 3B). P4 was also markedly reduced by more than 65% in TSPO-KO mice (t = 4.4, P = 0.002; Figure 3B). No significant difference in P5, ALDO, T or DHT levels was observed between the genotypes, although a trend toward reduced P5 and elevated androgen levels in TSPO-KO mice approached significance (P5: t = 1.9, P = 0.08; T: t = −2.1, P = 0.07; DHT: t = −2.1, P = 0.07; Figure 3B).
TSPO-KO perturbs testicular and adrenal steroidogenesis in young mice.
Altered correlations between steroid hormone precursors and metabolites
TSPO deficiency was associated with altered steady-state correlations between steroid precursors and downstream metabolites (Figure 4). While a positive correlation between P5 and P4 was observed irrespective of genotype (R2 = 0.50, P = 0.01; Figure 4A), a strong positive correlation was observed between P4 and CORT levels in TSPO-KO (R2 = 0.79, P = 0.01; Figure 4B), but not in WT mice (R2 = 0.14, P = 0.35; Figure 4B). Interestingly, a strong positive correlation between P5 and T (R2 = 0.76, P = 0.006; Figure 4D), but not P4 (R2 = 0.46, P = 0.09), was observed in TSPO-KO mice. In contrast, the opposite relationship between P5 and T was observed in WT mice (R2 = −0.68, P = 0.01; Figure 4D). No significant association was observed between CORT and ALDO concentrations in either WT or TSPO-KO mice (WT: R2 = 0.33, P = 0.18; TSPO-KO: R2 = 0.18, P = 0.6; Figure 4C), while a strong, positive correlation was observed between T and its metabolite, DHT, in both WT (R2 = 0.53, P = 0.05) and TSPO-KO mice (R2 = 0.97, P < 0.001; Figure 4E).
Altered correlations between serum levels of steroidogenic precursor metabolites in TSPO-KO mice.
TSPO-KO induces age-related androgen depletion
Next, we compared serum P5, P4, CORT, T and DHT levels with aging male mice (10 months and 22 months old, n = 5/group; Figure 5). As in young groups, a significant reduction in P4 and CORT levels was observed in TSPO-KO mice (P4 genotype effect: F = 27.89, P < 0.001; CORT genotype effect: F = 9.13, P = 0.005), although in aged TSPO-KO mice CORT levels were highly dispersed. Interestingly, a significant interaction between genotype and age was observed on T and DHT (genotype × age interaction: T, F = 13.89, P < 0.001; DHT, F = 33.98, P = 0.05), but not P5, P4 or CORT levels. We found TSPO to play an important role in maintaining androgen levels in aging, with serum T and DHT levels declining by greater than 90% in TSPO-KO mice by 22 months of age (P < 0.001 compared with all other groups; Figure 5A).
Age-dependent impairment of androgen production in TSPO-KO mice.
Steroid hormone precursor–metabolite relationships in aging TSPO-KO mice
To investigate if the effect of TSPO-KO on precursor–metabolite relationships was altered as a function of aging, we examined correlations between serum hormone concentrations in the aged mice (Figure 5B–E). While no significant correlation was observed between P5 and P4 in aged animals, when combined with data from young animals, a highly significant positive correlation was observed (R2 = 0.40, P = 0.008; Figure 5B). As in young adult mice, a strong positive correlation was observed between P4 and CORT in aged TSPO-KO (R2 = 0.79, P = 0.006) but not in WT mice (R2 = 0.14, P = 0.35; Figure 5C). While a strong positive correlation between P5 and T was observed in young TSPO-KO mice, no such association was observed in aged mice (R2 = 0.25, P = 0.27; Figure 5D). As in young mice, T concentrations positively correlated with its metabolite, DHT, in both WT (R2 = 0.69, P = 0.02) and TSPO-KO mice (R2 = 0.86, P < 0.003; Figure 5E).
Here, we demonstrate that although TSPO is not critical for maintaining steroid levels essential for life, TSPO deficiency is associated with reduced steroidogenic output, with specific reductions in levels of P4 and CORT in both young and aged male mice. Marked changes in the relationship between steady-state levels of steroidogenic precursors and metabolites suggest that loss of TSPO function was associated with abnormalities in steroidogenic flux. Furthermore, TSPO was critical for maintaining androgen production in aging, supporting the notion that TSPO may represent a therapeutic target for the treatment of hypogonadism and age-related androgen deficiency. These findings provide new insights into the complex interactions between TSPO and steroidogenesis, and may help resolve controversies surrounding apparent discrepancies in genetic versus pharmacological studies examining TSPO function.
It is of interest to note that the synthesis of both P4 and CORT, which were impaired in TSPO-KO mice, occurs predominantly in the mitochondria . Meanwhile, androgens, which are synthesized predominantly in the endoplasmic reticulum (ER) , were unaltered by TSPO-KO, with a tendency toward elevated concentrations. A trend toward elevated T levels has also been observed in the testis-specific TSPO-KO model . One possible explanation is that TSPO-KO may alter the global mitochondrial environment, impairing steroidogenesis in the mitochondrial compartment. For example, TSPO may alter mitochondrial redox states  leading to impaired electron transfer from catalytic redox partners, essential for P450 and hydroxysteroid dehydrogenase (HSD) enzyme activity . However, this explanation seems unlikely since no effect of TSPO-KO was observed on ALDO, which is also exclusively synthesized in the mitochondria. Previous studies have examined the effect of conditional and global TSPO-KO on transcriptional expression of the steroidogenic enzymes in the testes or adrenal, including StAR, P450scc and 3β-HSD, finding no difference between WT and TSPO-KO mice [15–18]. However, this does not address potential differences in protein level, subcellular compartment distribution or enzymatic activity. Analysis of precursor–metabolite relationships has proved useful in the diagnosis of disorders associated with steroidogenic enzyme deficiencies [28–30] and may provide mechanistic insights into the observed reduction in P4 and CORT in TSPO-KO mice. Assuming the steroidogenic system is operating at steady state, precursor–metabolite relationships can be used as an indicator of steroidogenic flux [35,36]. The validity of steady-state analysis is supported by the evidence that most metabolic systems operate close to steady state even in disease  and mathematical models of steroidogenesis have used steady-state assumptions to reliability predict the effects of endocrine disruptors on in vitro steroidogenic systems .
In TSPO-KO mice, despite a marked reduction in P4 levels, the relationship between P4 and its precursor, P5, was unaltered, suggesting normal 3β-HSD functionality. A linear relationship was observed between P5 and P4, consistent with first-order reaction kinetics where substrate concentration is directly proportional to the reaction rate. Reduced efficiency of P5 conversion to P4 resulting from a modest, non-significant reduction in substrate availability could lead to a marked reduction in P4 concentrations, as seen in the TSPO-KO mice. Interestingly, a marked difference in the relationship between P4 and CORT was observed between WT and TSPO-KO mice. In TSPO-KO mice, P4 was positively associated with CORT by a power function, consistent with second-order reaction kinetics. In contrast, no association was observed between P4 and CORT in WT mice, suggesting zero-order reaction kinetics whereby substrate concentration does not affect reaction efficiency. These differences in the P4–CORT relationship are consistent with a typical saturation curve, where P4 concentrations in WT mice exceed the maximal reaction efficiency. The relationship (or lack of) between CORT and ALDO in both WT and TSPO-KO was also consistent with zero-order reaction kinetics. Given the high absolute concentrations of CORT, we hypothesize that excess CORT substrate was available even in TSPO-KO mice, leading to saturation of aldosterone synthase (P450c18) activity.
Perhaps, the most striking difference between WT and TSPO-KO mice was the association between T and P5, with a strong positive correlation observed in TPSO-KO mice versus a negative relationship observed in WT mice. This could potentially be explained by compensatory steroidogenic activity in the ER, or a switch to the synthesis of the androgens via P5 → dehydroepiandrosterone (DHEA), rather than the normally predominant P4-substrate pathway  (Figure 2C). In normal mouse testis, P5 is competitively metabolized by 3β-HSD and cytochrome P450 17 (P540c17) to form P4 and DHEA, respectively . Enzyme–substrate proximity is thought to result in the preferential conversion of P5 to P4 by mitochondrial 3β-HSD, which is found to be closely associated with P450scc in the mitochondria . However, P5 may also diffuse into the ER where it becomes available for conversion to the intermediate, 17α-hydroxypregnenolone (17-P5), by P450c17, which is then converted into DHEA → androstenedione (A4) → T, promoting androgen synthesis via an alternative pathway. Modeling of the competitive P4 → A4 versus P5 → DHEA synthesis pathway has demonstrated how reductions in P5 supply can paradoxically lead to increased 17-P5 → DHEA production accompanied by a reduction in P4 production . Future studies could evaluate this hypothesis in pulse-labeled experiments in TSPO-KO models using a radiolabeled cholesterol substrate.
One such pulse-labeled study has been carried out in isolated mitochondria from TSPO-KO mouse testis, reporting no difference in P5 production between WT and TSPO-KO mice . However, there are several key limitations in the interpretation of the present study. Firstly, the authors quantified radioactivity in fractions separated by chromatography to determine the cholesterol : P5 ratio, but no control experiments were carried out to confirm the recovery, specificity and reproducibility of cholesterol : P5 fractionation and purification . Secondly, no quality control experiments were reported to confirm the quality and steroidogenic functionality of isolated mitochondria. For example, the ability to stimulate or disrupt steroidogenesis (e.g. using specific enzymatic inhibitors such as aminoglutethimide) or time-dependent accumulation of P5 would be appropriate positive controls. Furthermore, while the study also evaluated serum P5 in WT and TSPO-KO mice, the concentrations reported were 1000-fold higher than the normal expected range (e.g. ), adding to difficulty interpreting the findings. Finally, since P5 is rapidly converted to downstream steroid hormones due to the much higher reaction efficiency of subsequent reactions relative to P450scc (Km = 63 µM), steady-state serum P5 levels are a poor indicator of steroidogenic output. This is reflected in the very low steady-state serum concentrations of P5, representing less than 0.3% of total steroid hormones measured in the current study. Here, we consider the sum of all measured hormones as an indicator of steroidogenic capacity to circumvent this problem, since P5 is converted to downstream products with a 1 : 1 stoichiometry.
In contrast with the current findings, previous studies have reported no effect of TSPO-KO on CORT levels , or impaired CORT production in response to adrenocorticotropic hormone stimulation . Reasons for the discrepancies between these studies are not clear, but differences in sex, blood collection procedure, anesthesia and hormone measurement methods may be contributing factors. For example, different classes of anesthesia differentially promote CORT production [42,43], which could potentially mask or reveal TSPO-dependent differences in adrenal function.
Importantly, our findings indicate a critical role of TSPO in maintaining androgen levels during aging. Of note, in aged mice, no relationship was observed between P5 and T, in contrast with the strong positive correlation observed in young–adult TSPO-deficient mice. Meanwhile, TSPO deficiency was associated with reduced P4 and CORT levels irrespective of age, and in aged mice, the relationship between steady-state levels of P5 and CORT and T and DHT mirrored observations in young mice. These findings suggest an interaction between TSPO function and age-related changes in the testis. Such an interaction could involve age-induced changes in the supply of cholesterol , steroidogenic enzyme expression  or antioxidant capacity and ER stress . Interestingly, although previous studies have found that TSPO mRNA, protein and binding sites decrease by up to 50% in the aging rat testes [47,48], the steroidogenic capacity of synthetic TSPO ligands is preserved in the testis of aged rats with hypogonadism . Taken together with the current findings, the therapeutic potential of TSPO-targeted treatments for the management of age-related androgen deficiency warrants further investigation.
Steroid hormones are essential for life, and understanding the mechanisms mediating their synthesis has broad implications for our understanding of not only basic biology but also the treatment of disease. The present study provides new evidence that TSPO modulates steroidogenic pathways, adding to the growing literature elucidating complex interactions between TSPO and steroidogenic pathways.
liquid chromatography–tandem mass spectrometry
methyl tert-butyl ether
cytochrome P450 17
cytochrome P450 side-chain cleavage
translocator protein knockout
A.M.B., T.S., M.H. and B.J. designed the research; A.M.B., B.J. and S.K. performed the research; A.M.B., M.H. and B.J. analyzed the data and wrote the paper.
This work is supported by Grants-in-Aid for Core Research for Evolutional Science and Technology to T.S. and M.H. , the Brain Mapping by Integrated Neurotechnologies for Disease Studies , Research and Development Grants for Dementia  to M.H. and the Strategic Research Program for Brain Sciences to T.S. from the Japan Agency for Medical Research and Development and the Japan Advanced Molecular Imaging Program and Scientific Research on Innovative Areas [23111009 to M.H.] from the Ministry of Education, Culture, Sports, Science and Technology, Japan.
The authors thank T. Minamihisamatsu, S. Sasaki and S. Uchida for the technical support and assistance, M.-R. Zhang for radiosynthesis (National Institutes for Quantum and Radiological Science and Technology) and P. Matthews (Imperial College London) for the insightful discussion on the manuscript.
The Authors declare that there are no competing interests associated with the manuscript.