A young testicular microenvironment protects Leydig cells against age-related dysfunction in a mouse model of premature aging.

Testicular Leydig cells (LCs) are the primary source of circulating androgen in men. As men age, circulating androgen levels decline. However, whether reduced LC steroidogenesis results from specific effects of aging within LCs or reflects degenerative alterations to the wider supporting microenvironment is unclear; inability to separate intrinsic LC aging from that of the testicular microenvironment in vivo has made this question difficult to address. To resolve this, we generated novel mouse models of premature aging, driven by CDGSH iron sulfur domain 2 ( Cisd2) deletion, to separate the effects of cell intrinsic aging from extrinsic effects of aging on LC function. At 6 mo of age, constitutive Cisd2-deficient mice display signs of premature aging, including testicular atrophy, reduced LC and Sertoli cell (SC) number, decreased circulating testosterone, increased luteinizing hormone/testosterone ratio, and decreased expression of steroidogenic mRNAs, appropriately modeling primary testicular dysfunction observed in aging men. However, mice with Cisd2 deletion (and thus premature aging) restricted to either LCs or SCs were protected against testicular degeneration, demonstrating that age-related LCs dysfunction cannot be explained by intrinsic aging within either the LC or SC lineages alone. We conclude that age-related LC dysfunction is largely driven by aging of the supporting testicular microenvironment.-Curley, M., Milne, L., Smith, S., Jørgensen, A., Frederiksen, H., Hadoke, P., Potter, P., Smith, L. B. A Young testicular microenvironment protects Leydig cells against age-related dysfunction in a mouse model of premature aging.

The mammalian testis is divided into 2 distinct compartments that perform its principal functions. Spermatogenesis occurs within the seminiferous tubules, and androgen biosynthesis occurs in the interstitial space. Both of these processes are entirely dependent on the 2 major testicular somatic cell populations—the Sertoli cells (SCs) and Leydig cells (LCs), respectively. In human males, testicular function declines during the aging process [reviewed in Perheentupa and Huhtaniemi (1) and Gunes et al. (2)]. Of particular significance is the reported age-related decrease in LC androgen production (3–7) because androgens have been suggested to have a crucial role in supporting lifelong general health in men, with low circulating testosterone linked to an increased risk of developing chronic age-related cardiometabolic diseases (8–13). Therefore, elucidating the mechanisms governing the maintenance and function LCs is pertinent to our understanding of male health.

LC androgen production is under negative feedback control from the hypothalamic-pituitary-gonad (HPG) axis. Gonadotropin-releasing hormone, from the hypothalamus, stimulates the pituitary to secrete luteinizing hormone (LH), which stimulates LC testosterone production. Increased circulating testosterone then negatively regulates the hypothalamic–pituitary unit to control LH secretion. Reduced levels of circulating testosterone (hypogonadism) can be due to defective androgen biosynthesis at the level of the LC, usually accompanied by a compensatory increase in LH and subsequent distortion of the LH/testosterone ratio (primary hypogonadism) or as a result of diminished LH production by the hypothalamic pituitary unit (secondary hypogonadism). In aging males, the former is generally observed, although testosterone levels may remain within reference range because of a compensatory increase in LH (6, 14).

Current models for the study of age-related LC dysfunction are limited. The most widely used model has been the naturally aged brown Norway rat (Rattus norvegicus), which has provided significant insight into changes occurring in LCs during aging, including disruption of the pro/antioxidant balance, impaired LH-receptor signal transduction and cholesterol trafficking, and reduced expression and activity of enzymes involved in the conversion of cholesterol to testosterone (15–24). Interestingly, experimentally induced regeneration of LCs through ethylene dimethane sulfonate–mediated ablation in the aged testis was, for many years, thought to reinvigorate the regenerated LC population to function as young LCs (25). More recently, it has been shown that this capacity is not maintained long term (26), suggesting that either the aging of the LC stem cells themselves is a factor or that the aged microenvironment the regenerated LCs are in is unable to support continued testosterone production. A detailed dissection of this paradigm has been hampered by both the inability to separate individual components of the system (i.e., other testicular somatic cell populations that may influence LC function in a paracrine manner) and the prohibitive time and financial outlay required to study aged cohorts.

Rodent models of premature aging provide an attractive alternative for the expedited study of age-related processes. Indeed, a number of progeroid mouse lines have been generated to date, primarily through disruption of genes required for DNA repair/genomic stability or lamin processing/maintenance of nuclear architecture [reviewed in Kõks et al. (27) and Carrero et al. (28). However, most have been generated to model monogenic human progeroid syndromes and, thus, may be of limited utility for the study of age-related HPG axis dysfunction. Furthermore, in such models of constitutive gene ablation, it is difficult to separate the contribution of specific tissue/cell types to the observed phenotypes. As such, new models that provide a platform to study the mechanisms by which testis/LC function deteriorates during aging/disease are required. Ideally, a model permitting the separation of individual components of the system (i.e., other somatic cell populations that may influence LC function in a paracrine manner) would be valuable in assessing cause and effect.

To that end, we first validated a novel knockout (KO)-first conditional allele [Cisd2] of a previously reported premature aging model driven by CDGSH iron sulfur domain 2 (CISD2) deficiency (29). CISD2 is a redox active protein localized to the endoplasmic reticulum and is thought to be important for the maintenance of endoplasmic reticulum and mitochondrial structure and function (30). In mice, Cisd2 expression decreases with age, and its deletion results in a phenotype resembling premature aging, characterized by nerve and muscle degeneration, osteopenia, reduced body weight, and premature death (29). Conversely, Cisd2 overexpression has been reported to delay aging in mice (31). In the present studies, we sought to characterize the testicular phenotype of constitutive and cell-specific Cisd2-deficient mice to assess the utility of that premature aging model for the study of age-related LC dysfunction, and determine which cell types control and/or support LC function during the aging process.

MATERIALS AND METHODS

Ethics statement

Animals were bred and maintained in strict compliance with the Animals (Scientific Procedures) Act, 1986. All procedures were conducted in accordance with United Kingdom Home Office regulations under project licenses 60/4200 and 70/8804 held by L.B.S.

Breeding of transgenic mice

Mice carrying Cisd2 were obtained from the International Mouse Phenotyping Consortium (IMPC; Project ). Male and female Cisd2 mice were intercrossed to generate Cisd2, Cisd2, and Cisd2 offspring [hereafter referred to as wild-type (WT), heterozygous (HET), and KO mice, respectively]. Mice carrying Cisd2 were also obtained from the IMPC Project and bred to red fluorescent protein (RFP) cyclic recombinase (Cre)-reporter mice [B6;129S6-Gt(ROSA)26Sor; referred to as tdTomato, hereafter] to generate a colony of mice carrying both transgenes (Cisd2;tdTomato). LC and SC–Cisd2 mice were generated, using Pdgfrb-Cre (32) mice (provided by Prof. Neil Henderson, University of Edinburgh, Edinburgh, United Kingdom) and anti-Müllerian hormone (Amh)-Cre (33) mice, respectively. First, Cisd2;tdTomato mice were mated with either Cre line to generate Cre;Cisd2;tdTomato and Cre;Cisd2;tdTomato offspring. Cre;Cisd2;tdTomato and Cre;Cisd2;tdTomato offspring were then intercrossed to obtain Cre;Cisd2;tdTomato and Cre;Cisd2;tdTomato offspring (the experimental animals; hereafter referred to as LC-WT or SC-WT and LC-KO or SC-KO, respectively). Genomic DNA isolated from tail or ear biopsies was used to genotype mice for the presence of the various transgenes by standard PCR techniques using either BioMix PCR Reaction Buffer (Bioline, London, United Kingdom) or Type-it Mutation Detect PCR Kit (Qiagen, Hilden, Germany), according to the manufacturer’s instructions. The details of the genotyping assays are listed in . For Pdgfrb-Cre, Amh-Cre, and tdTomato genotyping assays, primers against endogenous Il2 were included as a positive control for the PCR reaction. Representative gels showing PCR patterns for each assay are shown in Supplemental Fig. S1. In LC-KO and SC-KO mice, Cisd2 recombination (i.e., conversion of Cisd2 to Cisd2) was assessed in genomic DNA isolated from testis biopsies. PCR products were analyzed with the Qiaxcel Capillary Electrophoresis System (Qiagen).

TABLE 1

Primers used for genotyping assays

Assay	Primer sequence, 5′–3′
	Forward	Reverse
Cisd2tm1a
5′-HA WT	GAATATTCAATGTGTAAAGGTTTCAA
3′-HA WT		GAAAACATTTTCACTCCTTTCTTTTT
5′-HA MUT		GAACTTCGGAATAGGAACTTCG
Cisd2tm1c Tm1c	GAGTAAGTTGACTGATCAGG	CCTATCTCATTAGATCTGCT
Cisd2tm1d
5′Cas	AAGGCGCATAACGATACCAC
Cisd2 WT		GAAAACATTTTCACTCCTTTCTTTTT
3′ LoxP		ACTGATGGCGAGCTCAGACC
Pdgfrb-Cre	TGCCACGACCAAGTGACAGCA	AGAGACGGAAATCCATCGCTC
Amh-Cre	CACATCAGGCCCAGCTCTAT	GTGTACAGGATCGGCTCTGC
Td-tomato	CTGTTCCTGTACGGCATGG	GGCATTAAAGCAGCGTATCC
Il2	CTAGGCCACAGAATTGAAAGATCT	GTAGGTGGAAATTCTAGCATCATCC

Tissue collection

Animals were euthanized between the hours of 08:00 and 11:00 am, by methods conforming to Schedule 1 of the Animals (Scientific Procedures) Act, 1986. For adult animals [5.7 ± 0.13 mo old (means ± sem); referred to as 6 mo old, hereafter, unless otherwise stated], this was achieved by exposure to an increasing concentration of carbon dioxide, followed by confirmation of permanent cessation of the circulation by palpitation. Where human chorionic gonadotropin (hCG) stimulation of LCs was performed, mice were injected intraperitoneally with a single dose of hCG (Chorulon; Henry Schein, Melville, NY, USA) 16 h before euthanasia (20 IU in 100 μl saline). For embryonic ages, pregnant dams were dispatched as described above, and the fetuses rapidly removed and decapitated. Body weight, testis weight, and seminal vesicle weight were recorded, and tissues were either frozen on dry ice for storage at −80°C or immersed in Bouin’s fixative (Clin-Tech, Guildford, United Kingdom) for 6 h. Fixed tissues were processed, embedded in paraffin wax, and sectioned at 5 µm for histologic analysis. Tissues expressing the tdTomato-fluorescent Cre-reporter were imaged with an MZFLIII Fluorescence Stereomicroscope (Leica Biosystems, Wetzlar, Germany) fitted with a CoolSnap camera (Photometrics, Tucson, AZ, USA) and PMCapture Pro 6.0 software (Photometrics).

Hormone analyses

Blood was collected at euthanasia via cardiac puncture with a syringe and needle precoated with heparin. Plasma was harvested by centrifugation (17,136 g, 10 min, 4°C) and stored at −20°C. The quantification of LH and follicle-stimulating hormone in mouse plasma was performed with the Milliplex Map Pituitary Magnetic Bead Panel Kit (PPTMAG-86K; MilliporeSigma, Burlington, MA, USA) according to the manufacturer’s instructions. Samples were read on a Bio-Plex 200 suspension array system (Bio-Rad Laboratories, Hercules, CA, USA) with Bio-Plex Manager software (Bio-Rad Laboratories). To measure testosterone and corticosterone in mouse plasma, an isotope-dilution TurboFlow liquid chromatography–tandem mass spectrometry method was used, as previously described (34). Samples were analyzed in 7 batches. Each batch included standards for calibration curves, ∼60 unknown samples, 1 blank, and control material prepared from a pool of human serum; 3 unspiked controls; 3 controls spiked at low levels; and 3 controls spiked at high levels. The interday variation, expressed as the relative sd for testosterone was 4.7 and 3.0% for low and high spike levels, respectively, and for corticosterone, the relative sd was 16.1 and 6.1% for low and high spike levels, respectively. The recovery was >89% for all analytes in both spike levels. Supplemental Table S1 shows the limit of quantification (34), the linear range for each individual calibration curve, and the interday validation. All chemical analyses were performed at the Department of Growth and Reproduction, Rigshospitalet, Copenhagen University Hospital.

Histology and immunostaining

For general histologic analysis, tissue sections were stained with hematoxylin and eosin (H&E), following standard protocols. Immunolocalization of a single protein was performed by a chromogenic immunostaining procedure, as previously described (35) whereas immunolocalization of 2 proteins on the same section was performed with a fluorogenic immunostaining procedure, as previously described (36). Antibodies used are detailed in . Images were captured using either an Axio Scan Z1 slide scanner (Carl Zeiss, Oberkochen, Germany) or an LSM 780 confocal microscope (Carl Zeiss).

TABLE 2

Details of antibodies and detection methods used for chromogenic and fluorogenic immunostaining

Target	Antigen retrieval	Primary antibody			Secondary reagent			Detection method
		Source	No.	Dilution	Source	No.	Dilution
HSD3b	Citrate pH 6	Santa Cruz Biotechnology, Dallas, TX, USA	sc30820	1:2000	Vector Laboratories, Burlingame, CA, USA	MP-7405	NA	ImmPRESS HRP DAB
HSD3b	Citrate pH 6	Santa Cruz Biotechnology	sc30820	1:2000	Santa Cruz Biotechnology	sc-2961	1:200	ChαG HRP Tyramide
SOX9	Tris-EDTA pH 9	MilliporeSigma	AB5535	1:4000	Vector Laboratories	MP-7451	NA	ImmPRESS HRP DAB
SOX9	Tris-EDTA pH 9	MilliporeSigma	AB5535	1:4000	Vector Laboratories	PI-1000	1:200	GαR HRP Tyramide
RFP	n/a	Evrogen, Moscow, Russia	AB233	1:800	Vector Laboratories	BA-1000	1:500	Streptavidin HRP DAB
RFP	Citrate pH 6	Takara Bio, Kusatsu, Japan	632496	1:2000	Vector Laboratories	PI-1000	1:200	GαR HRP Tyramide
COUP-TFII	Citrate pH 6	Perseus Proteomics, Tokyo, Japan	PP-H7147-00	1:1000	Thermo Fisher Scientific	P0447	1:500	GαM HRP Tyramide
PDGFRb	Citrate pH 6	Abcam	ab32570	1:1500	Vector Laboratories	PI-1000	1:200	GαR HRP Tyramide

ChαG, chicken anti-goat; DAB, 3,3′-diaminobenzidine; GαM, goat anti-mouse; GαR, goat anti-rabbit; HRP, horseradish peroxidase; NA, not applicable.

Stereology

Leydig and Sertoli cells were identified in tissue sections by chromogenic immunostaining for HSD3B and SOX9, respectively. Quantification of cells was conducted with the stereology plugin for Image-Pro Plus 7.0 software (Media Cybernetics, Rockville, MD, USA) and a Zeiss Axio Imager A1 microscope fitted with a Qicam Fast 1394 digital camera (Qimaging, Surrey, BC, Canada) and a ProScan automated stage (Prior Scientific Instrument, Fulbourn, Cambridge, United Kingdom), as previously described (37). Briefly, the “count” function was used to score nuclei of immunopositive cells and to obtain the relative nuclear volume, which was then converted to absolute nuclear volume by testis weight. Mean nuclear volume was then measured with the “nucleator” function and used to calculate cell number/testis.

Seminiferous tubule diameter

Seminiferous tubule diameter was measured with the stereology plugin for Image-Pro plus 7.0 software and an Axio Imager A1 microscope fitted with a Qicam Fast 1394 digital camera and a ProScan automated stage, as previously described (38).

Epididymal sperm reserve

The epididymal sperm reserves were quantified from frozen epididymides. The cauda epididymis was homogenized on ice in homogenization buffer (0.9% NaCl, 0.5% Triton X-100). Homogenization-resistant elongate spermatids were quantified with a hemocytometer.

Estimation of the proportion of LCs targeted by Pdgfrb-Cre

To determine the proportion of LCs targeted by the Pdgfrb-Cre, dual-fluorogenic immunostaining for HSD3B and RFP was performed. RFP+;HSD3B+ double-positive cells were counted and expressed as a percentage of total HSD3B+ cells in a minimum of 10 images across 2 different testis sections from each of n = 4 animals.

Western blotting

Western blotting for CISD2 was performed, as previously described (39), with minor modifications. Briefly, 60 µg of protein was separated on an 8–16% Tris-glycine polyacrylamide gel (Thermo Fisher Scientific, Waltham, MA, USA) and transferred onto an Immobilon-Fl membrane (MilliporeSigma). Blots were probed with primary antibodies against CISD2 (13318-1-AP; 1:500; Proteintech, Rosemont, IL, USA) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH; ab9483; 1:4000; Abcam, Cambridge, United Kingdom). Primary antibodies were detected with donkey anti-rabbit IRDye 800CW (925-32213, 1:10,000; Li-Cor Biosciences, Lincoln, NE, USA) and donkey anti-goat IRDye 680RD (925-68074; 1:10,000; Li-Cor Biosciences) secondary antibodies. Blots were imaged with the Odyssey imaging system (Li-Cor Biosciences).

Quantitative RT-PCR

Total RNA was extracted from frozen tissue with the RNeasy Mini Kit (Qiagen), per the manufacturer’s instructions, including an on-column DNase (Qiagen) treatment step. An external control luciferase RNA (5 ng/testis; Promega, Madison, WI, USA) was added to each sample at the start of the extraction. RNA concentrations were determined with a NanoDrop 1000 v.3.8.1 spectrophotometer (Thermo Fisher Scientific). Random hexamer-primed cDNA was reverse transcribed from 2 μg total RNA with the SuperScriptVilo cDNA Synthesis Kit (Thermo Fisher Scientific), per the manufacturer’s instructions. Real-time PCR was performed in 384-well format with the ABI Prism 7900HT Real-Time PCR System (Thermo Fisher Scientific) and the Universal ProbeLibrary (F. Hoffmann-La Roche, Basel, Switzerland). Assays were designed with the online Universal ProbeLibrary Assay Design Center tool (). Details of each assay are listed in . The resulting data were analyzed using the cycle threshold ΔΔC method, with the expression of each gene related to the external luciferase control. The external luciferase control was assayed with forward, 5′‑GCACATATCGAGGTGAACATCAC‑3′ and reverse, 5′‑GCCAACCGAACGGACATTT‑3′, and, a 5′NED-labeled forward probe (5′‑TACGCGGAATACTTC‑3′).

TABLE 3

Details of quantitative PCR assays

Gene	Primer sequence, 5′–3′		UPL probe
	Forward	Reverse
Lhcgr	GATGCACAGTGGCACCTTC	CCTGCAATTTGGTGGAAGAG	107
Star	AAACTCACTTGGCTGCTCAGTA	TGCGATAGGACCTGGTTGAT	83
Cyp11a1	AAGTATGGCCCCATTTACAGG	TGGGGTCCACGATGTAAACT	104
Hsd3b1	GAACTGCAGGAGGTCAGAGC	GCACTGGGCATCCAGAAT	12
Hsd3b6	ACCATCCTTCCACAGTTCTAGC	ACAGTGACCCTGGAGATGGT	95
Cyp17a1	CATCCCACACAAGGCTAACA	CAGTGCCCAGAGATTGATGA	67
Hsd17b3	AATATGTCACGATCGGAGCTG	GAAGGGATCCGGTTCAGAAT	5
Fshr	AGCCTTACCTACCCCAGTCAC	CAAATTGGATGAAGTTCAGAGGT	98
Inha	GGAAGATGTCTCCCAGGCTA	TGGCTGGTCCTCACAGGT	33
Inhbb	GGCAACCAGAACCTATTCGT	CATAGGGGAGCAGTTTCAGG	5

UPL, Universal ProbeLibrary.

Statistical analyses

Statistical analyses were performed with Prism 7.02 software (GraphPad Software, La Jolla, CA, USA). Data distribution was assessed by either the D’Agostino & Pearson or Shapiro-Wilk normality tests. Data were compared with either an unpaired, 2 tailed Student’s t test or a Mann-Whitney U test as appropriate. Where required, data were normalized with the Box-Cox transformation using the optimal λ obtained from the online Free Statistics Software (v.1.2.1; ). Inheritance of transgenes was assessed with a χ2 test. In each case, a value of P ≤ 0.05 was considered statistically significant.

RESULTS

KO of CISD2 in homozygous Cisd2 mice

To determine whether CISD2-deficient mice would provide a suitable platform to study age-related testicular dysfunction, we first confirmed the absence of CISD2 protein in Cisd2 mice. Heterozygous Cisd2 mice were intercrossed as described in Materials and Methods and genotyping primers designed around the synthetic cassette () were used to identify WT, HET, and KO animals (Fig. 1). The χ2 analysis of offspring derived from heterozygous crossings confirmed that inheritance of WT and KO alleles conformed to predicted mendelian ratios (Fig. 1). Western blotting on a panel of tissues, including liver and testis, confirmed absence of CISD2 protein in KO mice (Fig. 1). Furthermore, growth curves from WT, HET, and HOM mice (Fig. 1) revealed CISD2-deficient animals were significantly lighter from 8 wk of age compared with WT mice, consistent with previous reports of Cisd2 disruption (29). We, therefore, concluded that the KO-first Cisd2 allele functioned as expected (a null allele) and that CISD2 loss associated with that allele was functionally significant.

Figure 1

Validation of the knockout-first Cisd2 allele. A) Schematic of the WT and KO-first alleles detailing the location of genotyping primers and expected size of PCR products. B) PCR analysis of genomic DNA isolated from ear clips of WT (257 bp), HET (257 and 178 bp), and homozygous (KO; 178 bp) mice. C) Inheritance of WT and KO alleles in offspring derived from heterozygous crossings conformed to predicted Mendelian ratios (χ2; P = 0.4224), with no gender bias. D) Western blot analysis of liver and testis confirmed CISD2 protein was absent in KO mice. The lower band is predicted to be CISD1 based on the size and sequence homology in the region of the protein that the antibody was raised against. E) Growth curves of WT, HET, and KO mice revealed CISD2-deficient animals were significantly lighter from 8 wk of age compared with WT animals [P = 0.003691 at 8 wk, P = 0.000419 at 26 wk (WT vs. KO individual Student’s t test; n = ≥5 animals/genotype/time point)]. Values are means ± sem.

Testicular atrophy in CISD2-deficient mice

Having validated ablation of CISD2 in KO animals, we next sought to assess the effect of CISD2 deficiency (and, by inference, premature aging) on testicular architecture. No overt dysmorphology of the reproductive tract was noted, although testes and seminal vesicles appeared slightly smaller in KO animals (). In addition, a striking reduction in epididymal fat pad mass was apparent in the KO mice, in line with the previously reported (40) requirement of CISD2 for normal adipogenesis. Indeed, testis weight was significantly reduced in KO mice compared with WT controls (Fig. 2). Although testicular histology was largely normal in KO mice, with abundant interstitial LCs and full spermatogenesis occurring within seminiferous tubules, degenerating tubules were occasionally noted (Fig. 2).

Figure 2

Testis weight is reduced in CISD2-deficient mice. A) Representative reproductive tracts from WT and KO mice at ∼6 mo of age. Open arrowheads, seminal vesicle; closed arrows, testis. Note the striking reduction in epididymal fat-pad mass (asterisks). Scale bar, 10 mm. B) A significant reduction in testis weight was observed in KO mice compared with WT controls. ****P = <0.0001, unpaired Student’s t test; n = 9–10). Error bars = sem. C) H&E–stained testis sections from WT and KO mice. In each case, abundant interstitial LCs and seminiferous tubules (ST) with full spermatogenesis were present. However, occasional degenerating tubules (asterisks) were noted in the KO testis. Scale bars, 100 µm.

Given that the structure and function of the seminiferous tubules is entirely dependent on the SC population (35, 41), we next asked whether the tubular degeneration in the KO testis could be explained by a deficit in this cell population. Immunoreactivity of the SC marker SOX9 was observed, as expected, around the periphery of the seminiferous tubules in WT and KO testes (). However, stereologic analysis revealed a significant reduction in SC number in the KO testis compared with WT controls (Fig. 3). This decrease in the SC population was associated with a significant reduction in the diameter of the seminiferous tubules (Fig. 3) and a reduction in the epididymal sperm reserve (Fig. 3) in KO mice. Circulating follicle-stimulating hormone (Fig. 3) and testicular mRNA expression of Fshr, Inha, and Inhbb (Fig. 3) were similar between WT and KO animals. Taken together, these observations suggest CISD2 loss results in premature focal testicular atrophy consistent with age-related testicular atrophy observed in both rodents and humans (42–50).

Figure 3

SC number, seminiferous tubule diameter, and epididymal sperm reserves are reduced in CISD2-deficient mice up to 6 mo of age. A) Representative chromogenic immunostaining of the SC marker SOX9 (SRY-Box 9) in WT and KO testes. B) SC number was reduced in KO testes compared with WT controls. **P = 0.0016 (unpaired Student’s t test; n = 8). C) Seminiferous tubule diameter was significantly reduced in KO mice compared with WT controls. **P = 0.0033, (Mann Whitney U test; n = 8). D) Epididymal sperm reserves were significantly reduced in KO mice compared with WT controls. **P = 0.0091 (unpaired Student’s t test; n = 8). E) Circulating follicle-stimulating hormone was similar between WT and KO mice (P = 0.3700, Student’s t test; n = 11–12). F–H) Testicular mRNA expression of follicle stimulating hormone receptor (Fshr) (F), inhibin subunit α (Inha) (G) and inhibin subunit β (Inhbb) (H) was similar between WT and KO mice [P = 0.2801, P = 0.0816, and P = 0.1931, respectively (Student’s t test; n = 7–8)]. Error bars = sem.

Testicular endocrine dysfunction resembling compensated LC failure in CISD2-deficient mice

In addition to atrophy of the seminiferous tubules, aging is associated with a reduction in LC androgen biosynthesis. As such, we next asked whether LC number or function was altered in prematurely aging CISD2-deficient mice. As expected, immunoreactivity of the LC marker HSD3B was observed in the interstitial compartment in WT and KO testes (). However, when quantified, a significant reduction in LC number was observed in KO mice compared with WT controls (Fig. 4). Next, we assessed whether LC function was compromised in the KO testis. Total circulating plasma testosterone and seminal vesicle weight (as a biomarker of peripheral androgen action) were both reduced in KO mice compared with WT controls (Fig. 4, respectively). Importantly, the reduction in circulating testosterone could not be attributed simply to reduced LC numbers because circulating plasma testosterone relative to LC number was also reduced in KO mice (Fig. 4). No significant difference in circulating LH was noted between WT and KO animals (Fig. 4). However, the LH, luteinizing hormone to testosterone (LH/T) ratio was significantly distorted (Fig. 4), indicative of compensated LC failure. Additionally, mRNAs of key LC-expressed genes required for the conversion of cholesterol to testosterone were significantly decreased in KO testes (), consistent with age-related LC dysfunction (17, 18, 20). Together, those observations point toward primary LC dysfunction in KO mice and suggest that the CISD2-KO mouse has utility as a model of the premature aging of the male reproductive system.

Figure 4

LC number and circulating testosterone are reduced in CISD2-deficient mice up to 6 mo of age. A) Representative chromogenic immunostaining of the LC marker hydroxysteroid dehydrogenase 3-β (HSD3B) in WT and KO testes. Scale bars, 100 µm. B) LC number was significantly reduced in KO mice compared with WT controls (unpaired Student’s t test; n = 8). **P = 0.0085. C, D) Both total circulating testosterone (C) and seminal vesicle weight (D), as biomarker of peripheral androgen action, were significantly reduced in KO mice. **P = 0.0085 (Student’s t test; n = 12), ***P = 0.0010 (Mann Whitney U test; n = 9–10). E) Circulating testosterone relative to LC number was also reduced in KO mice compared with WT controls. *P = 0.0280 (unpaired Student’s t test; n = 12). F) No difference in circulating luteinizing hormone was observed between WT and KO mice (P = 0.2215, unpaired Student’s t test; n = 10–12). G) The luteinizing hormone/testosterone ratio was significantly increased in KO mice compared with WT controls (P = 0.0232, unpaired Student’s t test; n = 10–12). Error bars = sem.

Figure 5

Steroidogenic gene expression is altered in CISD2-deficient testes. Testicular mRNA expression of genes involved in LC testosterone biosynthesis was reduced in KO mice compared with WT controls at ∼6 mo of age. Luteinizing hormone/chorionic gonadotropin receptor (Lhcgr), ****P = 0.0001; steroidogenic acute regulatory protein (Star), *P = 0.0102; P450 cholesterol side-chain cleavage enzyme (Cyp11a1), ****P < 0.0001; hydroxysteroid dehydrogenase 3-β type 1 (Hsd3b1), P = 0.2309; hydroxysteroid dehydrogenase 3-β type 6 (Hsd3b6), ***P = 0.0002; 17α-hydroxylase, 17,20-lyase (Cyp17a1), ****P = <0.0001; and hydroxysteroid dehydrogenase 17-β type 3 (Hsd17b3), **P = 0.0012 (unpaired Student’s t tests; n = 8). Error bars = sem.

Generation of LC- and SC-specific Cisd2-KO mice

To better understand the cause and effect relationship between aging and declining testosterone levels, we next sought to separate the effects of aging in the testis from systemic aging. First, to determine whether LC dysfunction in constitutive CISD2-KO mice was due to specific effects of premature aging intrinsic to LCs, we used Cisd2 to disrupt Cisd2 in LCs. Additionally, because SCs have an important paracrine role in the retention of the adult LC population (35, 51), we also disrupted SC Cisd2 to determine whether premature aging in SCs affects LC function. We first assessed the utility of Pdgfrb-Cre mice in targeting the adult LC population. Pdgfrb-Cre mice were bred to tdTomato Cre-reporter mice as described in Materials and Methods. Dual fluorescent immunohistochemistry for RFP and HSD3B revealed colocalization in ∼55% of LCs in the adult testis (). RFP expression was also observed in peritubular and perivascular cells as well as in a subpopulation of SCs. As the population of LCs in the adult testis developed from a pool of peritubular and/or perivascular stem/progenitor cells during pubertal development (52–59), we further demonstrated that the Pdgfrb-Cre targets adult LCs from the stem/progenitor stage in fetal life (Supplemental Fig. S2). We also confirmed SC-specific activity of Amh-Cre in RFP reporter mice by dual-fluorescent immunohistochemistry for RFP and SOX9. As expected, Cre expression was confined to testicular SCs as evidenced by the pattern of RFP-positive staining, characteristic of SC cytoplasm, in cells positive for the SC nuclear marker SOX9 (Fig. 6). Having demonstrated LC and SC targeting with Pdgfrb-Cre and Amh-Cre lines, respectively, we next generated LC and SC Cisd2-KO mice, as described in the Materials and Methods (referred to as LC-KO and SC-KO, respectively, hereafter). PCR interrogation of testicular genomic DNA isolated from testes of LC-KO, SC-KO, and their respective WT controls confirmed recombination of the conditional Cisd2 allele in the presence of either Cre recombinase (Fig. 6).

Figure 6

Generation of LC- and SC-specific Cisd2-KO mice. A) Analysis of Pdgfrb-Cre activity in the testis young adult tdTomato Cre-reporter mice. Coimmunostaining of RFP with the LC marker HSD3B revealed that Cre-recombinase is active in ∼55 ± 3.9% (mean ± sem) of adult LCs. Expression was also observed in the peritubular cells (arrows) as well as in a proportion of SCs (closed arrowheads). Scale bar, 20 µm. B) Confirmation of SC-specific targeting with Amh-Cre. Coimmunostaining of RFP with the SC marker SOX9 showed RFP expression was restricted to SCs. C) i) Schematic of the conditional Tm1c allele detailing the location of genotyping primers and expected size of PCR products following Cre-mediated recombination. ii, iii) PCR analysis of testicular genomic DNA confirmed the conditional Cisd2 allele was recombined upon exposure to either Pdgfrb-Cre (LC-KO) (ii) or Amh-Cre (SC-KO) (iii). WT = 268 bp; KO (recombined) = 174 bp.

Disruption of Cisd2 in either LCs or SCs does not result in premature testicular dysfunction

In contrast to the phenotype observed in constitutive CISD2-KO animals in which body weight was significantly reduced from 8 wk of age (Fig. 1), body weight was maintained in LC-KO and SC-KO animals (). In fact, an increase in body weight was observed in LC-KO animals. Because testicular atrophy accompanied by decreased SC number, seminiferous tubule diameter, and epididymal sperm reserves was observed in constitutive-KO testes (Figs. 2 and 3), we next asked whether that phenotype was recapitulated when Cisd2 was disrupted in either LCs or SCs alone. Testis weight was within reference range in both LC-KO and SC-KO animals (Fig. 7), with no histologic evidence of tubular degeneration (Fig. 7) as noted in constitutive-KO animals (Fig. 2). Additionally, SC number (), tubule diameter (Fig. 8), and epididymal sperm reserves (Fig. 8), as well as circulating follicle-stimulating hormone (Fig. 8) and testicular Fshr, Inha, and Inhbb mRNA expression (Fig. 8) were similar between LC-KO, SC-KO, and their respective controls. Taken together, these observations suggest that neither LCs nor SCs alone are responsible for the degenerative testicular phenotype observed in constitutive-KO animals.

Figure 7

Testis weight is unaltered when Cisd2 is disrupted either in LCs or in SCs. A) Contrary to the phenotype observed in constitutive CISD2-KO mice, body weight was not reduced either in LC-KO (left) or in SC-KO (right) animals at 6 mo of age. In fact, an increase in body weight was noted in LC-KO animals compared with LC-WT controls. *P = 0.0166 [SC-WT vs. SC-KO, P = 0.3688 (unpaired Student’s t test; n = 7–8)]. B) No difference in testis weight was observed between LC-KO or SC-KO mice and their respective WT controls [LC-WT vs. LC-KO, P = 0.5126l SC-WT vs. SC-KO, P = 0.5382 (unpaired Student’s t test; n = 7–8)]. Error bars = sem. C) Representative H&E stained testis sections of LC-KO (top) and SC-KO (bottom) mice and their respective WT controls. Testicular architecture remained normal when Cisd2 was disrupted in either LCs or SCs. Scale bars, 100 µm.

Figure 8

SC number, seminiferous tubule diameter, and epididymal sperm reserves are maintained when Cisd2 is disrupted in LCs or SCs. A) Representative chromogenic immunostaining of the SC marker SOX9 in 6-mo-old LC-KO and SC-KO testes. B) SC numbers were maintained in LC-KO and SC-KO testes when compared with their respective WT controls [LC-WT vs. LC-KO, P = 0.1604; SC-WT vs. SC-KO, P = 0.9264 (unpaired Student’s t test; n = 7–9)]. C, D) No difference in seminiferous tubule diameter (C) [LC-WT vs. LC-KO, P = 0.1736; SC-WT vs. SC-KO, P = 0.6188 (unpaired Student’s t test; n = 7–8)] or epididymal sperm reserve (D) [LC-WT vs. LC-KO, P = 0.5625; SC-WT vs. SC-KO, P = 0.6429 (unpaired Student’s t test; n = 6–8)] was noted in LC-KO or SC-KO mice compared with their respective WT controls. E) Circulating follicle-stimulating hormone was unaltered both in LC-KO and in SC-KO mice [LC-WT vs. LC-KO, P = 0.2303; SC-WT vs. SC-KO, P = 0.2024 (unpaired Student’s t test; n = 8–11)]. F–H) No difference in testicular mRNA expression of follicle-stimulating hormone receptor (Fshr; F) [LC-WT vs. LC-KO, P = 0.7971; SC-WT vs. SC-KO, P = 0.0507 (Student’s t test; n = 8)], inhibin subunit α (Inhα; G) [LC-WT vs. LC-KO, P = 0.0736; SC-WT vs. SC-KO, P = 0.6970 (Student’s t test; n = 8)], or inhibin subunit β (Inhbb; H) 9 LC-WT vs. LC-KO, ns P = 0.9582; SC-WT vs. SC-KO, P = 0.4391 (unpaired Student’s t test; n = 8)], and LC-KO, SC-KO, and their respective controls were similar. Error bars = sem.

We next asked whether LC dysfunction observed in constitutive CISD2-KO animals (Figs. 4 and 5) was a direct consequence of CISD2 loss from the LC population or was due to indirect paracrine effects mediated by SC dysfunction. LC number, circulating testosterone, and seminal vesicle weight were normal in both in LC-KO and in SC-KO mice compared with their respective WT controls (). To determine whether circulating testosterone levels are maintained through functional compensation by the remaining 45% WT LCs in the LC-KO testis (i.e., those not targeted by the Pdgfrb-Cre; Fig. 6), mice were injected with hCG (an LHCGR agonist) to stimulate maximal LC testosterone production. However, no difference in hCG-stimulated testosterone was observed (Fig. 9), effectively ruling that possibility out. In addition, no difference in circulating LH (Fig. 9), the LH/T ratio (Fig. 9), or testicular expression of steroidogenic mRNAs () was detected in LC-KO or SC-KO mice compared with their respective controls, suggesting that LC steroidogenic function is maintained when Cisd2 is disrupted either in LCs or SCs alone. Taken together, these data demonstrate that the mechanisms underlying the primary testicular failure observed in prematurely aged constitutive CISD2-KO animals cannot be assigned to specific dysfunction of the LC or SC populations alone.

Figure 9

Circulating testosterone is maintained when Cisd2 is disrupted either in LCs or in SCs. A) Representative chromogenic immunostaining of the LC marker HSD3B in 6-mo-old LC-KO, SC-KO, and control testes. B) Stereologic analysis revealed that LC number was normal in LC-KO and SC-KO testes compared with their respective controls (unpaired t-tests, LC-WT vs. LC-KO, P = 0.5158; SC-WT vs. SC-KO, P = 0.0736; n = 7–9). C) Circulating testosterone was normal in LC-KO and SC-KO mice compared with their respective controls [LC-WT vs. LC-KO, P = 0.9325; SC-WT vs. SC-KO, P = 0.2827 (unpaired Student’s t test; n = 8–11)]. D) Weight of the androgen-dependent seminal vesicle was maintained in LC-KO and SC-KO mice [LC-WT vs. LC-KO, P = 0.3823; SC-WT vs. SC-KO, P = 0.3843 Mann Whitney U test and unpaired Student’s t test, respectively; n = 8–11)]. E) LC response to maximally stimulating hCG was normal in both LC-KO and SC-KO mice [LC-WT vs. LC-KO, P = 0.1869; SC-WT vs. SC-KO, P = 0.5177 (unpaired Student’s t test; n = 4–6)]. F) Circulating luteinizing hormone was normal in both LC-KO and SC-KO mice [LC-WT vs. LC-KO, P = 0.3817; SC-WT vs. SC-KO, P = 0.8225 (unpaired Student’s t test; n = 8–11)]. G) The luteinizing hormone/testosterone ratio was unaltered when Cisd2 was disrupted in either LCs or SCs [LC-WT vs. LC-KO, P = 0.6385; SC-WT vs. SC-KO, P = 0.5115 (unpaired Stuedent’s t test; n = 8–11)]. Error bars = sem.

Figure 10

Steroidogenic gene is unaltered when Cisd2 is disrupted in either LCs or SCs. No difference in testicular expression of mRNAs involved in androgen biosynthesis was observed in either 6-mo-old LC-KO or SC-KO mice compared with their respective WT controls. Luteinizing hormone/chorionic gonadotropin receptor (Lhcgr), LC-WT vs. LC-KO, P = 0.1450; SC-WT vs. SC-KO, P = 0.2838; steroidogenic acute regulatory protein (Star), LC-WT vs. LC-KO, P = 0.4109; SC-WT vs. SC-KO, P = 6698; P450 cholesterol side-chain cleavage enzyme (Cyp11a1), LC-WT vs. LC-KO, (NS) P = 0.2404; SC-WT vs. SC-KO, P = 0.4037; hydroxysteroid dehydrogenase 3-β type 1 (Hsd3b1), LC-WT vs. LC-KO, P = 0.3852; SC-WT vs. SC-KO, P = 0.4390; hydroxysteroid dehydrogenase 3-β type 6 (Hsd3b6), LC-WT vs. LC-KO, (NS) P = 0.7086; SC-WT vs. SC-KO P = 0.5602; 17α-hydroxylase, 17,20-lyase (Cyp17a1), LC-WT vs. LC-KO, P = 0.2627; SC-WT vs. SC-KO, P = 0.2924; hydroxysteroid dehydrogenase 17-β type 3 (Hsd17b3), LC-WT vs. LC-KO, (NS) P = 0.5221; SC-WT vs. SC-KO, P = 0.8524; n = 6–8. Error bars = sem.

Alterations in the wider endocrine milieu during aging may affect LC function. Circulating glucocorticoids are reported to increase in aging rats and humans (60–62), and the suppressive effects of glucocorticoids on testosterone production are well documented (63–69). With this in mind, we next assessed circulating corticosterone levels and found a significant increase in constitutive CISD2-KO mice, whereas levels were unchanged in both LC-KO and SC-KO mice (). Furthermore, testicular mRNA expression of Nr3c1, encoding the glucocorticoid receptor, which was previously reported to be under autoregulation (70, 71), was significantly decreased in constitutive-KO mice (Fig. 11). Additionally, although testicular mRNA expression Hsd11b1 was similar between constitutive and conditional KO mice and their respective controls, expression of Hsd11b2 was decreased in the constitutive KOs, potentially indicating a perturbation in intercellular glucocorticoid metabolism (Fig. 11).

Figure 11

Increased circulating corticosterone in CISD2-deficient mice. A) Circulating corticosterone was increased in constitutive CISD2-KO mice at ∼6 mo of age. **P = 0.0054 (unpaired Student’s t test; n = 12), but not in 6-mo-old LC-KO or SC-KO mice [LC-WT vs. LC-KO, P = 0.2806; n = 8; SC-WT vs. SC-KO, (NS) P = 0.6807; n = 10–11 (unpaired Student’s t test)]. B) Testicular mRNA expression of the glucocorticoid receptor was decreased in constitutive CISD2-KO mice (Gr). ****P < 0.0001 (unpaired Student’s t test; n = 8), whereas expression was normal in LC-KO and SC-KO mice compared with their respective controls [LC-WT vs. LC-KO, (NS) P = 0.7052; n = 8; SC-WT vs. SC-KO, (NS) P = 0.8109; n = 8 (unpaired Student’s t test)]. C) Testicular mRNA expression of Hsd11b1 was normal in constitutive CISD2-KO, LC-KO, and SC-KO mice [P = 0.2869, P = 0.0733, and P = 0.7123, respectively, n = 8 (unpaired Student’s t test)]. D) Expression of Hsd11b2 was reduced in constitutive CISD2-KO testes (*P = 0.0322; n = 8), whereas expression was maintained both in LC-KO and SC-KO testes [P = 0.0806 and P = 0.4049, respectively (unpaired Student’s t test; n = 8)].

DISCUSSION

Aging in men is accompanied by a decline in testicular function. Of particular importance is the reported age-related decrease in the endocrine function of the testis (e.g., androgen production) (3–7, 14) because an inverse relationship between circulating testosterone levels and cardiometabolic disease risk has been suggested (8–12). However, the cause–consequence relationship between androgens, aging, and disease is unclear. Specifically, the precise mechanisms by which LC testosterone production becomes compromised, leading to age-related primary hypogonadism, remains to be established. Using a series of novel mouse models of premature aging, we now demonstrate that age-related testicular atrophy and LC dysfunction are not entirely explained by intrinsic aging in either LCs or SCs per se. Instead, we suggest that aging-induced disruption to the testicular microenvironment and/or wider systemic effects of aging may be significant contributing factors to age-related testicular dysfunction.

Age-related testicular atrophy is well documented both in humans and in rodents. Testicular volume declines with advancing age and is linked to a reduction in SC number and/or function, involution of the seminiferous tubules, and diminished spermatogenesis (42–49, 72–74). Consistent with that, our histomorphometric analyses of constitutive CISD2-KO mice revealed a testicular phenotype characterized by reduced testis weight, seminiferous tubule diameter, SC number, and epididymal sperm reserves. Significantly, in comparison with previous reports of aging in the rodent testis, signs of age-related atrophy in constitutive CISD2-KO mice are observed ∼18 mo earlier than in naturally aged animals (43–45). As such, prematurely aging CISD2-KO mice provide a valuable platform for the expedited study of aging in the testis.

Testicular endocrine function (i.e., testosterone production) also deteriorates during the aging process, which may be attributed to either a defect in LC androgen biosynthesis (primary hypogonadism) or to reduced stimulation of LCs because of decreased hypothalamic–pituitary LH secretion (secondary hypogonadism). The latter has been associated with obesity in men (75), and its prevalence is not reported to increase with advancing age (76). Instead, progressive, age-related reduction in testosterone level is thought to be predominantly caused by primary testicular dysfunction (14, 76, 77). It has been reported that testosterone levels decrease in aging mice because of reduced frequency and amplitude of LH pulses (i.e., secondary hypogonadism) (78, 79). Conversely, others have reported that LH and testosterone levels remain stable up to 31 mo of age in mice (80, 81). As such, naturally aged mice may be refractory to age-related LC dysfunction, potentially limiting their utility as a model of human primary hypogonadism. The ratio of luteinizing hormone to testosterone is a well-established indicator of LC function, with an increased ratio indicative of primary LC dysfunction. In the present studies, we noted a reduction in total circulating testosterone accompanied by an increase in the LH/testosterone ratio in constitutive CISD2-KO mice, consistent with age-related primary hypogonadism in aging men. This phenotype was not recapitulated when Cisd2 disruption (and by inference premature aging) was restricted to LCs, suggesting that the otherwise young testicular microenvironment in this conditional model is conducive to testosterone production by prematurely aged LCs.

Interestingly, when healthy and diseased aged mice are considered separately, a reduction in circulating testosterone is reported in diseased mice (82). However, as circulating gonadotropins were not measured in that study, it is not known whether the hypogonadism in the aged, diseased animals was primary or secondary in nature. Additionally, the cross-sectional nature of the study prevents any interpretation of causality (i.e., whether low testosterone levels drive disease progression or vice versa). In line with previous studies by Chen et al. (29) and Wang et al. (40), the CISD2-KO mice described herein had a marked reduction in adipose tissue. Although resection of the epididymal white adipose tissue depot has been shown to have a negative effect on spermatogenesis, circulating testosterone and LH are unaffected (83). As such, we suggest that prematurely aging CISD2-KO mice represent a novel model for the study of age-related primary hypogonadism, without confounding issues associated with increased adiposity and associated hypothalamic/pituitary dysfunction observed in naturally aged animals.

In humans, circulating testosterone is either specifically bound to sex hormone binding globulin (SHBG), weakly bound to albumin, or is not associated with binding proteins (i.e., is free) (84, 85). SHBG is primarily produced by the human liver, whereas in rodents, the equivalent, androgen binding protein (ABP), is produced by testicular SCs (86–88). In aging men, SHBG levels increase, contributing to reduced free circulating testosterone (6), which represents an additional layer of complexity in the progression of age-related hypogonadism. Whether ABP levels are altered in aging rodents is unclear, but it is possible that species differences in steroid binding proteins, particularly very low levels/absence of ABP in the mouse compared with the rat (87), may partially explain why naturally aged mice do not fully model human age-related primary hypogonadism. This requires further study.

Testosterone deficiency may be attributed to reduced number and/or function of LCs. Early studies reported up to a 44% reduction in LC numbers in the aging human testis, accompanied by a 2-fold increase in LH to maintain circulating testosterone levels (89–91). However, a more-recent study reported decreased numbers of SCs, but not LCs in the aging human testis (92). In the brown Norway rat, LC number is maintained in aged animals, whereas steroidogenic function is reduced, in part, because of reduced expression of genes required for testosterone biosynthesis (93, 94). In the present studies, we report a reduction in the number of LCs in prematurely aging CISD2-KO mice. Additionally, mRNA expression of genes required for LC androgen biosynthesis was reduced in constitutive CISD2-KO animals. This could potentially limit the capacity for increased steroidogenic flux to maintain normal testosterone levels in the face of reduced LC number. For example, in other rodent models, with up to a 75% reduction in LC number, testosterone levels are maintained because of functional compensation by the remaining LC population (35, 41, 95). Indeed, reduced circulating testosterone in constitutive CISD2-KO animals was not entirely explained by correction for the reduction in LC number, confirming a primary LC dysfunction in this model. Surprisingly, LC number and function were maintained in conditional LC-KO and in SC-KO mice, suggesting that impaired LC function in constitutive KO mice is not explained by aging specifically in either LCs or SCs alone.

In addition to reduced expression and activity of proteins required for the conversion of cholesterol to testosterone, damage to steroidogenic machinery as a result of perturbation in the balance between the generation of reactive oxygen species (ROS), and their neutralization by antioxidants, is hypothesized to underlie the decreased testosterone production by aged LCs (21, 23, 24, 94, 96). In other rodent models of premature aging, including the senescence accelerated SAMP8 mouse line (97–99) and premature aging induced by d-galactose administration (100, 101), primary testicular dysfunction is associated with increased testicular ROS production. Studies by Wiley et al. (30) demonstrated that CISD2-loss results in a pro-oxidative intracellular environment, which may explain the reduction in testosterone observed in CISD2-KO mice. However, in prematurely aging mitochondrial DNA mutator mice, despite a 7-fold increase in LC superoxide production, testosterone biosynthesis was unaffected; therefore, increased ROS may not be directly toxic to LCs (95). As such, maintenance of LC function in our LC-KO mice supports the hypothesis of Chen et al. (26), stating that alterations in factors extrinsic to LCs may be responsible for impaired steroidogenic function of the aged testis. This would also be in agreement with the reported rejuvenation of aging tissues by exposure to a young systemic milieu in models of heterochronic parabiosis (102–106) and, with the observation that activity of aged stem cells is restored after transplantation into young hosts (107).

Drawing comparisons between prematurely aging, constitutive CISD2-KO mice and our conditional LC-KO and SC-KO models effectively enables the separation of cell intrinsic aging and systemic aging effects on LC function. In the present studies, we noted that circulating corticosterone was significantly elevated in constitutive CISD2-KO mice but not in LC- or SC-conditional KO mice. Because glucocorticoids are known to suppress testicular androgen biosynthesis (63–69), this may offer a partial explanation as to why LC function is impaired in constitutive-KO, but not in LC-KO or SC-KO, animals. The balance of intercellular, bioactive glucocorticoid is regulated by HSD11B enzymes, which are responsible for interconversion between active (corticosterone) and inactive (11-dehydrocorticosterone) forms. Although we did not assess enzymatic activity specifically within LCs we did note a difference in testicular expression of Hsd11b2, but not Hsd11b1, in constitutive-KO mice. This may represent an inadequate response to elevated corticosterone, leaving LCs in constitutive-KO mice unprotected from the suppressive effects of glucocorticoid on steroidogenesis.

Although old animals may be considered the simplest rodent models of aging, significant time, cost, and welfare implications, concomitant with the generation of naturally aged animals, limit their practicality. Furthermore, aging represents a multifactorial series of complex changes across multiple systems, making it difficult to assign cell/tissue specific contributions to age-related dysfunction. Although mouse models of accelerated aging may not fully model the natural aging process, they have been employed as alternatives to shed light on the mechanisms underpinning degenerative processes associated with aging (27). The testicular phenotype in constitutive CISD2-KO mice closely resembles that of the aging human testis, rendering this line a useful tool for the expedited study of testicular aging. Furthermore, in contrast to naturally aged mice, the hormonal profile of CISD2-KO animals may better reflect the status of the HPG axis in aging men, thus providing a useful resource to test novel therapeutic interventions aimed at reversing age-related primary hypogonadism. Based on our observations of both LC and SC conditional Cisd2-KO models, we suggest that, in addition to intrinsic LC aging, age-related disruption to the testicular microenvironment and/or wider endocrine milieu, likely has a significant role in age-associated LC dysfunction.

Supplementary Material

This article includes supplemental data. Please visit to obtain this information.

Click here for additional data file.

Click here for additional data file.

Click here for additional data file.