Sprague-Dawley rats aged 9 months to 1 year were maintained according to protocols approved by the Institutional Animal Care and Use Committee of the University of Southern California (Protocol #20791). Mitochondrial fusion promoter M1 (4-Chloro-2-(1-(2- (2,4,6-trichlorophenyl) hydrazono)ethyl)phenol, Sigma-Aldrich) (Wang et al., 2012) was dissolved in DMSO and diluted to 2 mg/kg/day for each animal. Animals were separated into three groups (n = 4 per group): untreated (control), DMSO-only control, or 2 mg/kg/day M1 (M1) daily for 6 weeks. The concentration of M1 and treatment duration used was chosen based on previous studies (Ding et al., 2020). Rats were sedated under isoflurane during tail blood collection throughout the study. Briefly, anesthesia was induced with isoflurane in an induction chamber. The animal was placed in the induction chamber and the isoflurane levels were turned to 2%–3% with a flow rate of 0.8–1.0 L/min. Once the animal was under, the gas supply to the induction chamber was turned off, the animal was removed from the induction chamber and mounted with a nose cone. The gas flow was restored at levels of 1.5%–2% at a flow rate of 0.4–0.8 L/min while tail blood was collected. All rats were euthanized by CO2 inhalant followed by decapitation by guillotine. Rats were placed into the CO2 chamber with flow rate between 30% and 50%. Thereafter, the rat was decapitated by guillotine, trunk blood was collected, and plasma was separated by centrifugation at 2000 g for 15 min, stored at −80°C, and used for determination of circulating testosterone levels.

Leydig cells were extracted using magnet-activated cell sorting (MACS), as previously described (Ji et al., 2022). Briefly, testes were collected from rats after euthanasia, washed and placed in ice-cold PBS. Testes were then decapsulated, and the testicular milieu was digested in 1 mg/mL collagenase in DMEM/F12 medium containing 0.1% bovine serum albumin (BSA) for 30 min at 34°C in a slow shaking incubator at 90 cycles/min. The digested testicular milieu was then filtered through a 70 µM pore nylon mesh. Cells were pelleted, then washed with DMEM/F12 medium, and resuspended in Ca²⁺, Mg²⁺-free Hank’s balanced salt solution containing 0.1% BSA before sorting.

Briefly, following the manufacturer’s protocol, testicular cell pellets were suspended in ice-cold IMag buffer at 2 × 10⁷ cells/mL with prolactin receptor (PRLR) antibody (1:150) for 30 min at 4°C. The labeled cell suspension was washed twice with IMag buffer and labeled with anti-mouse IgG1 magnetic beads (1:20) at 4°C for 30 min. The labelled cell suspension was placed into the IMag Cell Separation Magnet holder (BD Biosciences, US) for 8 min. The supernatant was then removed, and the positive fraction was resuspended in IMag buffer and placed in the magnet holder for an additional 4 min. This resuspension was performed two more times to purify the cell population. Purification was verified by both staining for 3β-hydroxysteroid dehydrogenase and flow cytometry.

The purity of Leydig cells was determined by a 3β-hydroxysteroid dehydrogenase chemical reaction as described previously (Browning et al., 1981). In brief, Leydig cell fractions were incubated for 30 min at 37°C with the substrate dehydroepiandrosterone (100 g/mL; Sigma-Aldrich, St. Louis, MO, United States) in 0.07 M phosphate buffer (pH 7.2) containing 1 mg/mL nicotinamide, 6 mg/ml g-NAD, and 1.5 mg/mL nitro blue tetrazolium (Sigma-Aldrich). In parallel, flow cytometry was performed after the fractions were incubated with a PE-conjugated goat anti-mouse IgG fluorescent secondary antibody in darkness for 1 h. PRLR + cells were analyzed using a Fortessa X20 Flow Cytometer (BD Biosciences).

Proteins were extracted using RIPA buffer supplemented with protease inhibitors. Protein concentrations were measured using a Pierce BCA Protein Assay Kit (ThermoFisher Scientific, Waltham, MA, United States). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis was performed using 1 μg/μL of purified protein on a 4%–20% Tris-glycine gradient gel (Bio-Rad, Hercules, CA, United States). Protein bands were electro-transferred onto a polyvinylidene fluoride membrane and blocked with 5% BSA for 30 min. Anti-OPA1 mouse monoclonal antibody clone 1E8-1D9 that reacts with mouse, rat, and human protein was from ThermoFisher Scientific (UniProt ID: O60313). Rabbit monoclonal antibody against cleaved Caspase-3 (UniProt ID: P42574) from Cell Signaling Technology was used to visualize the protein. TSPO protein expression was assessed using an affinity purified rabbit anti-peptide antibody raised against the mouse TSPO C-terminal sequence as previously described (Payne and Hardy, 2007; Batarseh and Papadopoulos, 2010). This antiserum detects human, mouse, and rat TSPO. Other antibodies against CYP11A1 (UniProt ID: P05108), STAR (UniProt ID: P49675), VDAC1 (UniProt ID: P21796), RIPK1 (UniProt ID: T1503), and MFN2 (UniProt ID: O95140) were used as appropriate. Membranes were incubated with primary antibody overnight at 4°C at 1:1,000 dilution and secondary antibody at 1:5,000 for 1 h at room temperature. Membranes were then quenched with Radiance Peroxide and Radiance Plus (Azure Biosystems, Dublin, CA, United States) and subsequently imaged with an Azure c600 system (Azure Biosystems). Membranes were then stripped with Restore Western blot Stripping Buffer (ThermoFisher Scientific) and incubated with an antibody against the housekeeping protein GAPDH (1:1,000) (Proteintech, Rosemont, IL, United States) for normalization used as a control. Protein expression levels were measured using the ImageJ software, in which the density of the protein band of interest was compared with the GAPDH level for that particular band.

Cells were treated with 50 ng/mL human chorionic gonadotropin (hCG; National Hormone and Peptide Program, Harbor-UCLA Medical Center Torrance, CA, United States) or control media and incubated for various time points at 37°C. Media were collected for testosterone measurement using the Testosterone ELISA Kit (Cayman Chemical, Ann Arbor, MI, United States). Similarly, blood serum was collected and evaluated using the same Testosterone ELISA Kit.

Cultured cells (1 × 10⁴) and isolated primary Leydig cells (3 × 10⁴) were plated onto Seahorse XF Cell Culture Microplates overnight (Agilent Technologies, Santa Clara, CA, United States). Cell media were replaced with Agilent Seahorse XF DMEM Medium supplemented with 1 mM glucose, 1 mM pyruvate, and 2 mM glutamine and incubated at 37°C in a non-CO₂ incubator. Cells were evaluated using a Seahorse XF Cell Mito Stress Test Kit or a Seahorse XF Real-Time ATP Rate Assay Kit according to the manufacturer’s specifications. Briefly, working solutions of oligomycin (2.5 µM), FCCP (2.0 µM) and Rot/AA (0.5 µM) were prepared and loaded into a sensor cartridge that had been hydrated in Seahorse XF Calibrant at 37°C in a non-CO₂ incubator overnight. The assay was performed using a Seahorse XFe96 Analyzer with templates designed in Wave 2.6.1.

Cells were imaged by TEM at the USC Core Center of Excellence in Nano Imaging. Cells were primarily fixed using 2.5% glutaraldehyde, 2% paraformaldehyde, 0.1 M HEPES, and 0.115 M Sucrose. After washing with 0.1 M cacodylate, cells were placed in a secondary fixative containing 1% osmium tetroxide. Cells were then stained with uranyl acetate and dehydrated with a series of 30%–100% EtOH washes. The dehydrated cells were transitioned to a microfuge tube using propylene oxide and infiltrated using increasing concentrations of Poly/Bed 812 epoxy resin. The cell-containing-block was sectioned using a Leica EM UC6 Ultramicrotome (Leica Biosystems, Nussloch, Germany) and examined on a FEI Talos F200C G2 Biological Transmission Electron Microscope (ThermoFisher Scientific). The images produced were quantified using ImageJ software, where the number of observable mitochondria and their size were measured to quantify mitochondrial biogenesis and mitochondrial size.

The Simplified Molecular Input Line Entry System format structure of M1 was obtained from the PubChem database and uploaded into the GastroPlus software 9.8 version (Simulations Plus Inc., United States). The Absorption, Distribution, Metabolism, Excretion and Toxicity (ADMET) Predictor software 11 version and PKPlus modules of GastroPlus v9.8 were used to simulate physicochemical and pharmacokinetic properties, respectively, in normal rats. Formulation and dosing parameters for the model were defined as following: 2 mg/kg/day injected into the rat model over 6 weeks (i.e., subsequent dosing of 2 mg/kg at an interval of every 24 h for the duration of 1,008 h, to mimic in vivo study conditions).

Data from experiments performed in technical triplicate are expressed as the mean ± standard error of the mean. GraphPad Prism (v.7; GraphPad Software, San Diego, CA, United States) was used for graphical presentation and statistical analysis. Analysis was performed using a Student’s t-test or ANOVA with multiple comparisons where appropriate. Results were considered statistically significant at p < 0.05.

Our previous studies isolated primary Leydig cells using isosmotic continuous Percoll (Gibco Inc.) gradients generated by centrifugation and subsequent BSA gradients (Browning et al., 1981). It was recently demonstrated that Leydig cells could be isolated from adult rats using magnetic-activated cell sorting (Ji et al., 2022), saving a considerable amount of time in experimental procedures. Evidence indicates that the prolactin receptor is a unique Leydig cell surface marker (Green et al., 2018; Guan et al., 2022), which was targeted in our MACS protocol. This cell sorting technique was used to obtain a purified Leydig cell population that produced testosterone under both basal and hormone stimulated conditions (Figure 1A). To further confirm that the sorted cell population was Leydig cells, the recovered cells were incubated with a PE-conjugated antibody and subsequent flow cytometry analysis revealed >90% of isolated cells expressed PRLR (Figure 1B). Additionally, the Leydig cell population was reaffirmed after staining for 3β-hydroxysteroid dehydrogenase activity (Figure 1C), an enzyme specific to steroidogenic cells.

To investigate the relationship between mitochondrial fusion and function in aged Leydig cells, we induced mitochondrial fusion ex vivo by treating isolated Leydig cells with cell-permeable M1, which has been shown to promote mitochondrial tubular network formation, increase OPA1 expression, and increase ATP levels (Wang et al., 2012; Ding et al., 2020). Treatment with M1 resulted in increased mitochondrial function and increased steroid biosynthesis (Figure 2A). The oxygen consumption rate and hormone production of the M1 treated cells showed improved cellular function. Based on our previous findings (Garza et al., 2022), we had expected that mitochondrial structure would improve after M1 treatment. Dysfunctional mitochondria accumulate with aging and present abnormal shape and structure. Using TEM to visualize mitochondrial morphology, we found that untreated aged Leydig cells had fewer healthy mitochondria when compared to the M1-treated Leydig cells (Figure 2B). The treated aged Leydig cells also contained more mitochondria with normal cristae structure (Figure 2C).

Throughout the 6-week treatment period, serum and weight measurements were taken to assess the impact of a daily intraperitoneal injection of M1. M1 administered in vivo daily gradually decreased testosterone levels and weight over the 6-week treatment period when compared to untreated animals (Figure 3A), while isolated primary aged Leydig cells treated with M1 ex vivo had produced increased levels of testosterone (Figure 2A). Furthermore, analysis of primary Leydig cells isolated after the end of the in vivo treatment showed increased expression of apoptosis marker cleaved Caspase-3 and decreased expression of steroidogenic proteins (CYP11A1, TSPO, STAR, VDAC) (Figure 3B). Collectively, these data show declining Leydig cell health when M1 is administered in vivo.

Deterioration of mitochondria is associated with aging and observed in numerous diseases and conditions, so mitochondrial degradation was expected in aged Leydig cells here. Because the promotion of mitochondrial fusion enhanced mitochondrial phenotypes in previous studies, we were interested in visualizing the effects of M1 treatment on mitochondrial structure. After treatment was complete, Leydig cells were isolated, fixed, and observed using transmission electron microscopy, revealing abnormal phenotypic characteristics when compared with isolated cells from untreated animals (Figure 4A). Untreated Leydig cells displayed the expected normal degradation that occurs with aging. Weak mitochondrial integrity was observed in the untreated group, while mitochondria from treated animals displayed mitochondrial damage, including abnormal cristae structure and deterioration of the mitochondrial matrix. Additionally, the treated Leydig cells produced significantly decreased mitochondrial biogenesis and size (Figure 4B).

Previous research using M1 have showed beneficial effects in a cardiovascular disease model and their healthy counterparts after injection with M1³³. Using the same animals as above, adrenal, liver, and heart samples were evaluated. Immunoblots of extracted proteins from these tissues showed decreased markers for apoptosis and increased expression of some steroidogenic proteins (Figure 5A). Although the adrenal gland produces steroids, this production does not affect circulating testosterone levels. The data for adrenal cells suggest slightly increased health and increased steroidogenic capacity. Analysis of protein expression levels in liver samples showed significantly increased levels of apoptotic markers and decreased expression of mitochondrial proteins (Figure 5B), suggesting declining liver health and cell integrity after treatment. The expression of apoptotic markers decreased in heart samples (Figure 5C), confirming the beneficial effects found previously by Ding et al. (2020). Collectively, the effect of systemic M1 treatment assessed by protein expression varies at the tissue level, with some tissues showing beneficial health effects (adrenal, heart) and deleterious effects in others (testis, liver).

ADMET Predictor and GastroPlus software were used to predict ADME properties and simulate the in vivo rat experiment. Predicted physicochemical, metabolism, and pharmacokinetic properties of M1 were generated and used in physiologically based pharmacokinetic modeling. Quantitative structure activity relationship analysis through ADMET Predictor suggested that M1 is a substrate for numerous glucuronosyltransferases and cytochrome P450 enzymes (Figure 6A). Additionally, M1 was predicted to act as an inhibitor of several cytochrome P450 enzymes and could display androgen receptor-related toxicity. Next, plasma concentrations were predicted using the actual experimental conditions for dose and route of administration as input settings in the pharmacokinetic model. The results suggest that there should be no accumulation of M1 after multiple dosing and it should be cleared prior to subsequent doses. To better understand the varying effects amongst tissues that were observed, as shown in Figure 4, a physiologically based pharmacokinetic model of a rat was developed to estimate specific organ exposure to M1. Total exposure to the native M1 compound was highest in the liver, followed by the reproductive organs and kidney. Furthermore, since M1 is highly hydrophobic it was predicted to undergo glucuronidation as a possible phase 2 metabolic clearance mechanism to facilitate excretion (Figure 6B).