|Home | About | Journals | Submit | Contact Us | Français|
Irradiation with 6 Gy produces a complete block of spermatogonial differentiation in LBNF1 rats that would be permanent without treatment. Subsequent suppression of gonadotropins and testosterone (T) restores differentiation to the spermatocyte stage; however this process requires 6 weeks. We evaluated the role of Leydig cells (LCs) in maintenance of the block in spermatogonial differentiation after radiation exposure by specifically eliminating functional LCs with ethane dimethane sulfonate (EDS). EDS (but not another alkylating agent), given at 10 weeks after irradiation, induced spermatogonial differentiation in 24% of seminiferous tubules 2 weeks later. However differentiation became blocked again at 4 weeks as LCs recovered. When EDS was followed by treatment with GnRH antagonist and flutamide, sustained spermatogonial differentiation was induced in >70% of tubules within 2 weeks. When EDS was followed by GnRH antagonist plus exogenous T, which also inhibits LC recovery but restores follicle stimulating hormone (FSH) levels, the spermatogonial differentiation was again rapid but transient. These results confirm that the factors that block spermatogonial differentiation are indirectly regulated by T and probably FSH and that adult and possibly immature LCs contribute to the production of such inhibitory factors. We tested whether insulin-like 3 (INSL3), a LC-produced protein whose expression correlated with the block in spermatogonial differentiation, was indeed responsible for the block by injecting synthetic INSL3 into the testis and knocking down its expression in vivo with siRNA. Neither treatment had any effect on spermatogonial differentiation. The Leydig cell products that contribute to the inhibition of spermatogonial differentiation in irradiated rats remain to be elucidated.
Radiation, chemotherapy for cancer treatment, and other toxicants deplete germ cells, causing prolonged azoospermia in both rodents (Boekelheide & Hall, 1991; Delic et al., 1986; Kangasniemi et al., 1996; Sawhney et al., 2005) and humans (Meistrich & van Beek, 1990). In all rat strains tested, such agents can cause azoospermia by disrupting the differentiation of surviving type A spermatogonia (Abuelhija, 2012 #6176). In some strains, such as LBNF1 rats, the block of spermatogonial differentiation is complete even at moderate doses of cytotoxic agents (Meistrich & Shetty, 2003). In humans, histological evidence and the period of prolonged azoospermia before the recovery of spermatogenesis also shows that radiation therapy and chemotherapeutic drugs can block differentiation from surviving stem cells (Clifton & Bremner, 1983; Kreuser et al., 1989; Meistrich & van Beek, 1990). Thus, the block in spermatogonial differentiation in LBNF1 rats can be a model to study some aspects of how treatment for cancer often leads to male sterility.
In irradiated LBNF1 rats, the block does not appear to be a direct effect of irradiation as spermatogonial differentiation proceeds for 6 weeks after irradiation (Kangasniemi et al., 1996). Then, between 6 and 8 weeks post-irradiation (Porter et al., 2006), coinciding with the complete disappearance of differentiated germ cells from the seminiferous epithelium, spermatogonial differentiation becomes blocked. The surviving spermatogonia were fully capable of differentiation after transplantation to the permissive environment of irradiated nude mouse testes (Zhang et al., 2007), demonstrating that alterations in the somatic environment of the rat testis were the cause of the block. Although microarray analyses have revealed numerous changes in gene expression in the remaining cells, the factor(s) that are involved in producing the block (Zhou et al., 2010), the cellular origin of the factor(s), and whether the block is produced by stimulation of an inhibitory factor or repression of a supportive factor are not known.
Without additional treatment, the spermatogonial differentiation block is irreversible for at least one year. However spermatogenesis can be restored by the suppression of gonadotropins and testosterone (T) using GnRH analogs and androgen receptor antagonists (Blanchard et al., 1998; Meistrich & Kangasniemi, 1997; Meistrich et al., 2001; Setchell et al., 2001; Shetty et al., 2000; Udagawa et al., 2001). We have shown that both T and FSH independently contribute to producing the block in spermatogonial differentiation in toxin-treated rats, but T seems to be a more potent inhibitor (Shetty et al., 2000; Shetty et al., 2006). T and other androgens act directly though the androgen receptor to produce the inhibition and aromatization to estradiol is not involved (Shetty et al., 2002). These findings conform with the facts that alterations in the somatic environment are responsible for the spermatogonial differentiation block and that the receptors for androgen and FSH are localized in the somatic cells, and not the germ cells, of the testis (Bremner et al., 1994).
Although type A spermatogonial proliferation is initiated within 2 weeks after hormone suppression (Zhou et al., 2011), 6 weeks are required for a robust stimulation of differentiation of surviving type A spermatogonia to the B spermatogonial and spermatocyte stages (Porter et al., 2006; Shuttlesworth et al., 2000). The reasons for this delay are also unknown. Possibilities included the times required (i) for the complete suppression of T levels and action, (ii) for stem and progenitor spermatogonia to visibly differentiate, and (iii) for completion of the reversal of an indirect process by which T and/or FSH stimulates production of an inhibitory factor. Because primarily T but also FSH are required for the transformation of spermatocytes to mature sperm, no sperm are produced during hormone suppressive treatment; however, after the cessation of suppression complete spermatogenesis and fertility were restored (Meistrich et al., 2001).
To further understand the hormone-mediated inhibition of spermatogonial differentiation after irradiation, it is important to identify the androgen-responsive cell types in the testis that mediate the spermatogenic inhibitory effect of T after irradiation. Analyses of microarray data indicated that most of the genes that were upregulated after irradiation were dependent on T were expressed in Leydig cells (LCs) (Zhou et al., 2011; Zhou et al., 2010). Hence in the present study, we studied the LCs as potential effectors of the T-induced inhibition of spermatogonial differentiation.
Rat LCs develop from 3β-hydroxysteroid dehydrogenase (HSD)-negative stem cells and become 3β-HSD-positive progenitor LCs, which remain as spindle-shaped cells at the periphery of the seminiferous tubule basal lamina during 14–21 days after birth (Ge et al., 2006). The progenitor LCs proliferate, and by day 28 most differentiate to round immature LCs, which are capable of androgen production and which finally transform into T-secreting adult LCs by day 56 (Shan et al., 1993).
Ethane dimethane sulfonate (EDS) can be used to specifically eliminate the adult LCs in the rat testis (Teerds, 1996). However, the stem LCs survive and produce LCs that, in the presence of luteinizing hormone (LH) (Molenaar et al., 1986; Sriraman et al., 2003), pass through differentiation stages similar to those that occur during development and reach control numbers at about 35 days after EDS administration in normal rats but recover more rapidly in germ cell–free busulfan-treated or irradiated testes (Molenaar et al., 1986; O’Shaughnessy et al., 2008). The effect of EDS-induced LC elimination on spermatogonial differentiation had been examined previously in hexanedione-treated rats, which also show a block in spermatogonial differentiation (Boekelheide et al., 2003; Richburg et al., 2002). There was a slight transient stimulation in differentiation when EDS alone was given, but the EDS treatment prevented the recovery of spermatogenesis with subsequent GnRH-agonist treatment.
In this study, we used EDS to eliminate the postprogenitor LCs in the irradiated rat testis to evaluate the role of LCs in mediating the regulation of spermatogonial differentiation after toxicant exposure and subsequent hormone suppression. We employed several techniques to modulate the recovery of LCs, including inhibition of their recovery by giving a GnRH antagonist (GnRH-ant) and stimulating the numbers of progenitor LCs with human chorionic gonadotropin (hCG). We modulated the levels or action of intratesticular T by the exogenous administration of T or flutamide, respectively. We also directly modulated the levels of a Leydig cell secreted candidate factor, insulin-like-3 (INSL3), by injecting synthetic peptides and by siRNA knockdown, to test whether it had inhibitory action on spermatogonial differentiation.
Adult LBNF1 (F1 hybrids of Lewis and Brown-Norway) male rats were obtained from Harlan Sprague-Dawley (Indianapolis, IN) and housed in animal facilities approved by the American Association for Accreditation of Laboratory Animal Care in accordance with the current regulations and standards of the Department of Agriculture and the National Institutes of Health. The rats were maintained on a 12-h light and 12-h dark cycle and were allowed food and water ad libitum. All rats were acclimatized for at least 3 days before the initiation of experiments, at which time they were 9-11 weeks old. All the animal procedures were approved by The University of Texas MD Anderson Cancer Center’s Institutional Animal Care and Use Committee.
Rats were anesthetized and their testes irradiated using procedures described in detail earlier (Shetty et al., 2002). Radiation was delivered to the lower part of the body, with the anterior edge of the field positioned about 6 cm above the base of the scrotum. A single dose of 6 Gy was administered at a dose rate of between 0.7-1.4 Gy/min.
The schedules of treatments administered after irradiation to modulate hormone levels are given in Fig. 1.
The GnRH antagonist acyline (kindly provided by Contraceptive Discovery and Development Branch, National Institute of Child Health and Human Development, Rockville, MD) was injected sc at a dose of 1.5 mg/kg per week to suppress gonadotropin levels in order to reduce intratesticular T levels (Porter et al., 2006) and to block the recovery of Leydig cells. The antiandrogen flutamide was administered to some of the GnRH-treated rats by implanting four 5-cm Silastic (Dow Corning, Midland, MI) capsules filled with flutamide sc, which was shown to effectively block androgenic action in irradiated rats (Porter et al., 2009), in order to further reduce the action of residual intratesticular T levels. T was administered to other GnRH-treated rats by implanting four 6-cm Silastic capsules filled with T sc in order to elevate and clamp interstitial fluid T (IFT) to levels that maintain normal spermatogenesis (Porter et al., 2009; Zirkin et al., 1989). hCG was administered sc in 4 daily injections at a dose of 100 IU/rat/day to stimulate proliferation of precursor Leydig cells (Teerds et al., 1988) just prior to the EDS injection. Flutamide, T and hCG were purchased from Sigma (St. Louis, MO).
EDS originally provided by Dr. Paul Juniewicz (Sterling-Winthrop Research Laboratories, Rensselaer, NY) was used except as noted. EDS was dissolved in DMSO:sterile water (1:3) and injected ip at doses of 75 or 100 mg/kg. The ability of EDS at these two dose levels to stimulate spermatogonial differentiation was almost identical (P = 0.88). That stock of EDS was depleted near the end of this study, and additional material was synthesized at the MD Anderson Translational Chemistry Core Facility under the direction of Dr. William Bornmann.
Cyclophosphamide (CP) (Baxter Healthcare, Deerfield, IL) was dissolved in sterile water and injected at doses of 170 or 200 mg/kg (Meistrich et al., 1995). The effect of CP at these 2 dose levels on spermatogonial differentiation was almost identical (P = 0.17). The rats were given post-treatment hydration and mesna (Mesnex, Baxter Healthcare) injections to minimize CP’s bladder toxicity while not affecting the response of the testis to CP-induced damage (Meistrich et al., 1995). Hydration treatment involved four 10-mL ip injections of saline containing 5% dextrose given every hour starting 15 min after CP injection. Mesna, at a dose of 200 mg/kg, was included in the saline-dextrose injections given 15 min and 3 h 15 min after CP (Meistrich et al., 1995).
To knock down the levels of Insl3, three In Vivo siRNAs (Life Technologies), modified for stability in vivo and designed by Ambion to target the 3′-region of rat Insl3 mRNA, s234924, s234923, and s137584, as well as Negative Control #1 siRNA were used. The three rat Insl3 siRNAs were mixed with Invivofectamine 2.0 Reagent (Life Technologies) and then dialyzed using a Float-A-Lyzer G2 cassette (Spectrum Laboratories, Rancho Dominguez, CA). Starting at 10 weeks after irradiation, rats were given daily treatments for 6 weeks. They were anesthetized with isoflurane and injected in the right testis with a mixture of 1.2 μg of each of the 3 siRNAs in 100 μl volume, using a 30-gauge needle. This dose was based on giving a tissue dose of 2 mg/kg to a 0.6 g testis (Ambion Technical Support, personal communication). Negative control siRNA was injected into the other testis. All rats were euthanized 1 day after the last siRNA injection and portions of each testis were prepared for histology or stored in RNAlater (Qiagen, Germantown, MD).
To directly modulate INSL levels in the testis, we performed intratesticular injections of synthetic INSL3 peptide. Rats were irradiated and, 9.4 weeks later, injected with EDS and given hormone suppression treatment with weekly injections of acyline to stimulate spermatogonial differentiation. Starting 3 days after EDS treatment, rats were anesthetized with isoflurane and given daily injections of 2 μg of INSL3 peptide in 100 μl to the right testis, using a 30-gauge needle for 2 weeks. Left testes were injected with PBS as control. In two rats, an INSL3 peptide based on the rat sequence (Smith et al., 2001) (provided by Dr. John Wade, Florey Institute, Melbourne, Australia) was used, but because of limited availably of this peptide, a synthetic mouse INSL3 peptide (Phoenix Pharmaceuticals, Burlingame CA, catalog #035-43) was used in another 2 rats. All rats were euthanized 1 day after the last INSL3 injection and a portion of each testis was prepared for histology.
When the rats were killed, blood was collected by cardiac puncture under ketamine-acepromazine anesthesia. The serum was separated and stored at −20°C. In all rats, the right testis was weighed, a portion was fixed in Bouin’s fluid, and the rest was weighed again without the tunica and either saved in RNAlater or frozen form later protein extraction. The left testis was used to collect the interstitial fluid as described previously (Porter et al., 2006). The interstitial fluid was weighed, mixed with 200 μL of PBS in a vortex mixer, spun, and stored at −20°C for the analysis of T levels.
Serum T and IFT were assayed using T-antiserum-coated tubes (DSL 4000, Diagnostics Systems Laboratories, Webster, TX; or TKTT1, Siemens Health Care Diagnostics, Deerfield, IL). Both kits gave equivalent values which have been validated for use with rat serum (Porter et al., 2006; Shetty et al., 2011; Shetty et al., 2000). The minimum T detection level was 0.04 ng/mL. The intra- and inter-assay coefficients of variation were 10.2 and 16.6%, respectively.
FSH and LH measurements were performed by the University of Virginia, Center for Research in Reproduction, Ligand Assay and Analysis Core. Rat serum FSH was measured by radioimmunoassay (Gay et al., 1970), and LH was measured by a sensitive two-site sandwich immunoassay (Haavisto, 1993). Intra- and inter-assay coefficients of variation for FSH assay were 4.9 and 9.6% respectively, and the respective values for LH assay were 7.6 and 9.3%.
Bouin’s-fixed testis tissue was embedded in paraffin or methacrylate, and 5-μm sections were cut. In methacrylate sections stained with PAS-hematoxylin, the numbers of spermatogonia were counted relative to the number of Sertoli cells in a minimum of 100 seminiferous tubules. To evaluate the recovery of spermatogenesis following irradiation and hormone modulation, we scored a minimum of 200 tubules in one section from each animal. A tubule was scored as differentiating if it contained 3 or more cells that had reached the type B spermatogonia stage or later (Meistrich & van Beek, 1993). The tubule differentiation index (TDI), which is the percentage of tubules showing differentiation, was then computed.
LCs were counted in paraffin sections immunostained for 3β-HSD as described previously (Shetty et al., 2008). A rabbit polyclonal antibody against human placental 3β-HSD (1:2000), kindly provided by Dr. C. Richard Parker, Jr. (University of Alabama, Birmingham, AL), was used (Parker et al., 1995). Immunoreactivity was detected using biotinylated goat anti-rabbit followed by an avidin-biotinylated horseradish peroxidase complex visualized with diaminobenzidine according to the manufacturer’s instructions (Vectastain ABC kit; Vector Laboratories, Burlingame, CA). Slides were counterstained with hematoxylin. Specificity of the immunohistochemical reaction was confirmed using sections processed identically except for the omission of the primary antibody and no staining was observed (data not shown). About 100-150 60×60-μm frames per section were selected by systematic random sampling using the Stereo Investigator (MicroBrightField, Colchester, VT) software package (Shetty et al., 2006). All LCs and Sertoli cells, identified by the presence of their nucleus with a characteristic nucleolus in the section, were counted in the chosen frames. Data are presented as the numbers of Leydig cells normalized to the numbers of Sertoli cells. In some experiments, progenitor LCs with spindle-shaped nuclei at the peritubular region of the tubules were distinguished from the postprogenitor, immature, and adult LCs in the interstitium.
Tissue stored in RNAlater was processed for RNA extraction using a Qiagen RNeasy Midi Kit (Qiagen, Valencia, CA). The A260/280 ratios of the RNA preparations were between 2.0-2.1 and A260/230 ratos were between 2.1-2.3. Total RNA (3 μg) was used to generate the template cDNA using the Transcriptor First Strand cDNA Synthesis Kit (Roche Applied Sciences, Indianapolis, IN). Quantitative real-time PCR was performed using the Rotor-Gene 3000 thermocycler (Corbett Research, Sydney, Australia) (Zhou et al., 2010). Hsd3b1 was used since it is the most highly expressed gene for the LC marker 3β-HSD in irradiated rat testes; its mRNA levels are not affected by hormone levels (Zhou et al., 2011; Zhou et al., 2010). Insl3 was evaluated, since the expression of this LC gene Insl3 is largely dependent on T levels and in previous studies correlated with the inhibition of spermatogonial differentiation. The primer sequences used were: Hsd3b1, forward 5′-CACTGCTGCTGTCATTGATGT-3′ and reverse 5′-TTGAACACAGGCCTCCAATAG-3′ Insl3, forward 5′-CACTGCTGCTGTCATTGATGT-3′ and reverse 5′-TTGAACACAGGCCTCCAATAG-3′. Relative levels of mRNA concentration were calculated from the Ct value and amplification efficiency by the Rotor-Gene 6.0 software and were normalized to levels of ribosomal protein S2 (Rps2) mRNA. All samples were run in triplicate.
CR15 (rabbit anti-INSL3) antiserum was kindly provided by Dr. Stefan Hartung (University Hospital Hamburg-Eppendorf, Hamburg, Germany) and used at a dilution of 1:100-1000. Anti-actin (pan) polyclonal antibody (Catalog #AAN01) was obtained from Cytoskeleton Inc. (Denver, CO) and used at a dilution of 1:1000. Total testis protein was obtained from frozen tissue by homogenization in RIPA buffer. Protein concentrations were determined by a BCA protein assay kit (Pierce Chemical, Rockford, IL), and then 30 μg of protein were loaded on Ready gel Tris-HCL (10%–20% gradient) or Any kD Mini-PROTEAN TGX Precast Gel (Bio-Rad Laboratories, Hercules, CA) and subjected to electrophoresis, membrane transfer, and immunostaining with CR15 antiserum and SuperSignal West Dura Chemiluminescent Substrate (Pierce). Western blot images were collected using an Alpha Innotech imaging system (ProteinSimple, San Jose CA).
Organ weights, TDI, LC counts, and mRNA levels were represented as arithmetic mean ± SEM. For serum T and IFT measurements, the means and SEMs were calculated using log-transformed data. When only two groups were being compared, the significance of differences between different treatments was evaluated by t-tests. When multiple time or treatment groups were being compared, a one-way analysis of variance (ANOVA) test was performed to test whether there were significant differences between the groups (ANOVA P < 0.05) and then individual groups were compared using a post-hoc-Tukey test. All analyses were performed with the IBM SPSS (version 19) statistical package.
In all the studies described here, rats were irradiated with 6 Gy and starting 10 weeks later, at which time the only germ cells in the tubules are undifferentiated spermatogonia, they were given various treatments, as indicated in Figure 1.
To test whether there was a factor that might be inhibiting spermatogonial differentiation after irradiation that was derived from the LCs, adult LCs were specifically eliminated with EDS. As expected, EDS caused depletion of 95% of total LCs within 1 week as determined by immunohistological analysis (Figs. 2A and and3D,3D, Supplemental Fig. 4B), which was confirmed by analysis of Hsd3b1 mRNA (Fig. 2B). The loss of LCs resulted in a 70% reduction in intratesticular fluid T (IFT) at 1 week (Fig. 2C). Insl3, another LC-expressed gene, showed an even greater decline than did Hsd3b1 (Fig. 2D), since it is dependent on T levels. There was an increase in LH after EDS treatment (Supplemental Fig. 1) as expected since Leydig cell elimination causes a reduction in T; however the increase in FSH was not expected since there is no decrease in germ cells. There also was a marked reversal of the block in spermatogonial differentiation at 2 weeks, as reflected by the 26% TDI (Fig. 2E). This increase in TDI was much more rapid than that produced by GnRH-ant plus flutamide treatment (P < 0.001) (Fig. 2E). As Leydig stem cells remain after EDS treatment and the rate of recovery of LCs is accelerated in germ cell-depleted testes, the total LC numbers (P < 0.05) and IFT already showed appreciable restorations at 2 weeks. Likely as a consequence of the Leydig cell recovery, which reached 47% of control levels by 4 weeks, spermatogonial differentiation again became blocked at 4 weeks after EDS treatment. These results indicated that an inhibitory factor produced by LCs is involved; however, the possible effect of reduced IFT levels needed to be also considered.
We noted that the suppression of IFT levels was actually greater with the GnRH-ant or GnRH-ant–flutamide treatment than with EDS (P < 0.001) (Fig. 2C). Also the decline in T levels with GnRH-ant treatment was quite rapid, as shown by the reduction of intratesticular T levels to 10% of the control levels within 1 day and to 5% within 1 week (Supplemental Fig. 2). Furthermore, the flutamide treatment blocks the action of residual T in the testis during the GnRH-ant treatment (Porter et al., 2009). Thus, the delay in differentiation with GnRH-ant–flutamide treatment cannot be attributed to a slow reduction in T action, and the stimulation of spermatogonial differentiation after EDS must be primarily related to the LC loss and not the reduction of T.
Next we examined whether inhibiting the recovery of LCs and suppressing hormones could produce a rapid and sustained recovery of spermatogenesis after EDS treatment. Since the regeneration of LCs after EDS is primarily dependent on LH, we combined EDS treatment with GnRH-ant to suppress LH and thereby prolong the reduction in LC numbers. Flutamide was also given with the GnRH-ant to produce total androgen ablation. Indeed, this treatment produced a more sustained suppression of LC numbers (P < 0.001) (Fig. 4A) and Hsd3b1 mRNA levels (P < 0.002) (Fig. 4B) than did EDS alone. There was, however, some increase in LC numbers and Hsd3b1 between weeks 2 and 4 but the recovery was incomplete. Immunohistochemistry showed that some newly formed LCs were past the progenitor LC stage (Fig. 3F).
Combined treatment with EDS and GnRH-ant-flutamide did result in a rapid and sustained stimulation of spermatogonial differentiation in irradiated rats (P < 0.001) (Fig. 4E). The TDI reached 72% at week 2 and increased to 84% at week 4 (Figs. 3B and and4E),4E), and was much greater than that observed with EDS alone or with GnRH-ant-flutamide alone; the latter resulted in a TDI of only 1% at week 2 and 15% at week 4. The more rapid initiation of differentiation with EDS given before the GnRH-ant-flutamide alone supports the conclusion that an inhibitory factor is derived from the LCs.
Next we examined whether the stimulation of differentiation produced by elimination of LCs in the absence of T and FSH could also be sustained in the presence of these hormones in the testis. In this case we combined EDS treatment with GnRH-ant and high-dose T to maintain moderate levels of intratesticular T (~40 mg/ml). It should be noted that the administration of T to rats along with GnRH-ant partially reverses the suppressive effect of GnRH-ant on FSH (Supplemental Fig. 1), and produced supraphysiological (>20 ng/ml) levels of T in the serum (Supplemental Fig. 3).
EDS injection followed by GnRH-ant plus T treatment produced a rapid increase in TDI to 40% at 2 weeks (Fig. 5F). This also must be a result of the elimination of LCs, since GnRH-ant+T treatment alone produced minimal differentiation in irradiated rats (TDI ~1%) after 2 weeks of hormone treatment (Porter, 2006; Zhou, 2010). However, despite the continued reduction of post-progenitor LCs at 4 weeks after EDS (Fig. 5B), the TDI was reduced to nearly zero at week 4 (Fig. 5F). It is possible that the modestly increased numbers of progenitor LCs (P < 0.001) (Fig. 5A) and/or immature LCs (P < 0.05) (Figs. 3F, ,5B)5B) that are formed by week 4, may produce sufficient levels of the factor that inhibits spermatogonial differentiation. In any case, the reappearance of the block in differentiation at 4 weeks after EDS with GnRH-ant plus T treatment depends on the action of T and/or FSH, since subsequent suppression of these hormones by continuing the GnRH-ant treatment but replacing the T with flutamide, induces differentiation in 33 ± 7% of tubules within 4 weeks (data not shown).
To test whether progenitor LCs were producing an inhibitory factor, we treated rats with hCG for 4 days to increase the proliferation of precursor LCs prior to injection of EDS and the initiation of hormone modulation with GnRH-ant plus T (Figs. 3G and H). Pretreatment with hCG resulted in at least 10-fold more precursor LCs (P < 0.02) and 2-fold more (P < 0.05) later-stage LCs being present 2 weeks later than in rats just treated with EDS, GnRH-ant, and T (Figs. 5A and B). The higher levels of Hsd3b1 and Insl3 mRNA (Fig. 5C and D) resulting from hCG pretreatment are consistent with the presence of high levels of these messages in precursor LCs (Stanley et al., 2011; Teerds et al., 1999). The hCG pretreatment reduced the TDI produced by EDS and GnRH-ant plus T administration for 2 weeks only slightly, from 42% to 30% (P = 0.12) (Fig. 5F). These results indicated that progenitor LCs were not a major source of the inhibitory factor blocking spermatogonial differentiation in the presence of T and FSH. The modest increase in more mature LCs, likely immature LCs, could be responsible for a slightly reduced peak in TDI levels.
Since the stimulatory effects of EDS on spermatogonial differentiation were transient despite more prolonged effects on LC numbers and gene expression, we questioned whether this stimulation might be due to some other nonspecific effect of EDS, such as direct action on spermatogonia. Since EDS is a strong bifunctional alkylating agent, we tested this possibility by treating irradiated rats with another strong bifunctional alkylating agent, cyclophosphamide (CP), which is known to act on spermatogonia (Cai et al., 1997; Drumond et al., 2011). CP did not appear to kill LCs (Supplemental Fig. 4C), although these cells appeared to be smaller and more irregular and IFT (P = 0.052) and seminal vesicle weight (P < 0.001), which is an indicator of peripheral levels of T, were reduced (Fig. 6A and B). However, CP, given alone or with GnRH-ant-flutamide, did not stimulate spermatogonial differentiation in 2 weeks (Fig. 6C). In the same experiment, other rats treated with EDS alone or with GnRH-ant-flutamide, showed differentiation in 14% or 50% of tubules, respectively at 2 weeks after injection (Fig. 6C), further supporting the role of Leydig cell killing in stimulation of spermatogonial differentiation. Note that the EDS used in this experiment was from a second batch and the stimulation appeared lower than the corresponding TDI values, without and with additional GnRH-ant-flutamide, of 26% or 72%, respectively, that was obtained with the initial preparation (Figs. 2E, ,4E).4E). However, the differences in the levels of spermatogonial differentiation between the batches were not significantly different (P = 0.11), and the second batch also effectively eliminated LCs (Supplemental Fig. 4B) and reduced T levels (Figs. 6A & B). We next checked whether the inability of CP to stimulate the differentiation could have been due to killing or functional impairment of spermatogonia. At 1 week after CP treatment (at 200 mg/kg), there was a reduction in numbers of type A spermatogonia (per 100 Sertoli cells) to 3.0 ± 0.2, from the value of 5.3 ± 0.3 in rats treated with radiation only (P < 0.001). However, more importantly, the numbers of functional stem spermatogonia were not affected by CP treatment since the stimulation of differentiation by GnRH-ant–flutamide treatment for 6 weeks after the CP injection was as effective as with GnRH-ant–flutamide treatment without CP (P = 0.23) (Fig. 6D).
Insl3 gene expression is strongly upregulated after irradiation (Zhou et al., 2010), which causes inhibition of spermatogonial differentiation, and suppressed by GnRH-ant and estrogen treatments, which stimulate spermatogonial differentiation (Zhou et al., 2011). To further investigate its relationship to the regulation of spermatogonial differentiation, we compared its expression levels to the differentiation observed after various EDS-containing treatment regimens. After EDS treatment, the Insl3 mRNA levels were inversely related to subsequent spermatogonial differentiation (Fig. 2D and E). However when other treatments were given, spermatogonial differentiation was unrelated to Insl3 mRNA levels. When the rats were treated with GnRH-ant-flutamide, Insl3 gene expression and protein levels were downregulated within 3 days (Fig. 2D, Supplemental Fig. 5), but robust differentiation did not occur until 6 weeks. When the EDS-treated rats were given GnRH-ant plus T, the Insl3 levels were much lower than in the group treated with EDS only, but spermatogonial differentiation was only slightly higher at 2 weeks and was also blocked at 4 weeks (compare Figs. 2D and E with 5D and F). Also, when rats were pretreated with hCG before giving EDS and GnRH-ant plus T, there were large increases in Insl3 mRNA levels compared with the same treatment without hCG, but this resulted in only a marginal decrease in spermatogonial differentiation (Figs. 5D and F).
To directly test whether Insl3 gene expression was related to inhibition of spermatogonial differentiation, we knocked down those levels with siRNA for Insl3. Two weeks of daily injection of siRNA to the irradiated rat testes reduced the levels of Insl3 mRNA expression in vivo to about 25% of control levels, and this knockdown was maintained for the 6-week injection period (Fig. 7). Despite the reduction in Insl3 levels, the percentage of tubules showing differentiation remained at 0.1%, which is typical of 6-Gy irradiated rats not given any treatment. Thus reduction of Insl3 levels did not result in any significant change in the block in spermatogonial differentiation (P=0.28).
To directly test whether INSL3 protein inhibits spermatogonial differentiation, spermatogenic recovery was simulated in irradiated rats by giving a combination of EDS and GnRH-ant. Starting 3 days after the EDS, one of the testes of these rats was injected daily with either rat or mouse synthetic INSL3 for two weeks, to test whether exogenous INSL3 inhibits stimulated spermatogenic recovery, as compared to the contralateral sham-injected testis. As expected, sham-treated testes that were only exposed to EDS and GnRH-ant treatment showed TDIs ranging from 18 to 42%. The TDIs for the contralateral testes treated with INSL3, although marginally lower than those of the respective sham-treated testes, were not significantly different (Fig. 7C).
We have shown here that in irradiated rats, in which there was a total and potentially permanent block in spermatogonial differentiation, elimination of functional LCs by treatment with EDS resulted in a transient stimulation of spermatogonial differentiation at 2 weeks, but spermatogonial differentiation became blocked again at 4 weeks as the Leydig cells recover. However EDS treatment followed by GnRH-ant plus flutamide, resulted in high levels of stimulation of spermatogonial differentiation at 2 weeks, which were sustained. This stimulation was much more rapid than that observed after GnRH-ant plus flutamide alone, which required 6 weeks to produce robust spermatogonial differentiation. When the EDS treatment was followed by GnRH-ant plus T, spermatogonial differentiation was observed at 2 weeks, but was not sustained. Nevertheless the stimulation at 2 weeks was due to the EDS as it was not observed with GnRH-ant plus T alone. The rapid differentiation observed when LCs are eliminated indicates that the block in spermatogonial differentiation in irradiated rats involves an inhibitory factor that originates from LCs and furthermore is regulated at least in part by autocrine action of androgen, as had been proposed in a different model (O’Hara et al., 2015). We also tested whether this factor might be INSL3, but obtained negative results.
We were concerned about possible direct effects of EDS on germ cells. Treatment with a different alkylating agent, which does not kill Leydig cells, did not produce any stimulation of spermatogonial differentiation, arguing against such a direct effect (Fig. 6). Multiple studies demonstrate that the only effects of EDS on spermatogenesis in rats are indirect and a consequence of decreased T resulting from Leydig cell killing as they were prevented when T supplementation was given (De Kretser et al., 1989; Sharpe et al., 1988a; Sharpe et al., 1988b). Suggestions from other studies that T supplementation did not prevent the effects of EDS on spermatogenesis were based on fertility data (Jackson & Jackson, 1984) and did not account for epididymal damage from EDS; or on spermatid counts in rats treated with EDS and T-supplementation was compared to controls (Sprando et al., 1990) but the results were not significantly different from the T-supplemented group not treated with EDS. Another study suggesting a direct effect of EDS on spermatogenesis (Kerr et al., 1987) involved multiple injections of high doses to mice; however multiple EDS injections of 75 mg/kg to rats din not have any direct effect on germ cells (Morris, 1985). Although we cannot rule out minor effects on germ cells, the data generally support the selectivity of the doses of EDS used for Leydig cell toxicity in rats.
Our results showing a transient increase in tubule cross-sections containing differentiated germ cells in irradiated rats after EDS-alone treatment was similar to that previously observed in rats with a hexanedione-induced block in spermatogonial differentiation (Boekelheide et al., 2003; Richburg et al., 2002). However, in contrast to our study, when they followed the EDS treatment with a GnRH-agonist, the EDS treatment prevented the induction of spermatogonial differentiation by the GnRH agonist (Richburg et al., 2002). The differences may stem from the flare in LH and T levels right after the agonist is given, the greater suppression of T action we have obtained by GnRH-ant than by a GnRH agonist, and also addition of flutamide.
Our observation that, when EDS is followed by treatment with GnRH-ant plus T, the restoration of the blockage of spermatogonial differentiation occurs at 4 weeks after EDS despite the fact that Leydig cell numbers were only about 10% of control at this time, raises a question as to whether the blockage is solely due to a Leydig cell factor. It is still possible that the LCs that are formed at this time produce high levels of the inhibitory factor. The recovering LCs must be immature LCs because LH suppression, even with exogenous T administration, prevents their maturation to adult LC stage (Sriraman et al., 2003; Teerds et al., 1989). It is possible that the immature LCs produce higher levels of the inhibitory factor so that even if total LCs are reduced, they are still effective enough to completely block spermatogonial differentiation. The fact that hCG stimulation of progenitor LC numbers before EDS treatment only decreased the transient recovery of spermatogonial differentiation slightly indicates that progenitor LCs are not a major source of a inhibitory factor, leaving only the immature LC as a possible source.
Although LCs are involved in producing a factor that contributes to the block in spermatogonial differentiation, the roles of other cells should also be considered and there are three observations indicating involvement of Sertoli cells. First the initiation of the block in spermatogonial differentiation after irradiation is a delayed process first observed about 8 weeks later (Kangasniemi et al., 1996; Porter et al., 2006), which is the time required for differentiated germ cell loss by maturation depletion. Very similar blocks are observed after treatment with procarbazine (Meistrich et al., 1999), busulfan (Udagawa et al., 2001), or hexanedione (Allard et al., 1995), which have different mechanisms of action. From these observations, we conclude that differentiated germ cell loss, rather than specific toxicant-induced somatic cell damage, is the trigger for initiating the spermatogonial block. Since Sertoli cells interact most closely with germ cells and display changes in gene expression when germ cells are lost (Cheng & Bardin, 1987; Guitton et al., 2000) they are most likely involved in signaling the loss of germ cells to lead to the induction of the block. A second observation is that transplantation of normal Sertoli cells from young rats to the interstitium of irradiated rat testes stimulates recovery of spermatogonial differentiation in adjacent tubules (Zhang et al., 2009). And the third is our previous observations that FSH enhances the spermatogonial differentiation block (Shetty et al., 2006), which clearly shows that Sertoli cells also have some role in the inhibition of spermatogonial differentiation. The reestablishment of spermatogonial block with the partial restoration of FSH levels when exogenous T was given to EDS and GnRH-ant treated rats in the current study further supports this argument.
There are numerous possible mechanisms by which Sertoli and Leydig cells might interact to inhibit spermatogonial differentiation. Whereas the EDS experiments have indicated that Leydig cells produce an inhibitory factor that is dependent on T, it is possible that in the irradiated rat the Sertoli cells produce either a stimulatory factor that is only expressed when T and FSH are suppressed or an inhibitory factor that is dependent on T and FSH. Both cells could produce factors that directly impact the spermatogonia. It is interesting that the genes for the subunits of inhibin B, which can affect spermatogonial differentiation by inhibiting activin action (Mithraprabhu et al., 2010) are expressed in both Sertoli and Leydig cells. Inhbb is exclusively expressed in the Sertoli cells (Johnston et al., 2008) and is upregulated by suppression of T (Zhou et al., 2010), but this could contribute both to production of activin B and inhibin B. Inha is expressed equally in both Sertoli and Leydig cells (Johnston et al., 2008; Stanley et al., 2011). Inha is downregulated by suppression of FSH and should be reduced by loss of Leydig cells, either of which would reduce inhibin B dimer levels and could enhance spermatogonial differentiation. When T supplementation is given along with the GnRH-ant it reverses of the GnRH-ant-induced suppression of FSH (Supplemental Fig. 1B) which may account for the restoration of the block in spermatogonial differentiation (Fig. 5F). Future experiments would be necessary to test these hypotheses.
It is also possible that only one of these cells produces the factor that regulates spermatogonial differentiation but the other cell interacts with it in a paracrine manner. The gradual nature of the reversal of the block in spermatogonial differentiation, requiring 4 to 6 weeks (Porter et al., 2006; Shuttlesworth et al., 2000) when only hormone suppression with GnRH-ant plus flutamide was given to irradiated rats also suggests that such indirect interactions are involved. The present study shows that this cannot be a result of the time required for the germ cells to progress from undifferentiated type A spermatogonia (Shuttlesworth et al., 2000) to the B spermatogonia or spermatocyte stage, since differentiation is observed within 2 weeks after EDS-treatments. One possible model is that after germ cell depletion the Sertoli cell produce a hormonally-regulated paracrine factor that regulates the LC-derived factor that blocks spermatogonial differentiation.
Although our study indicates that a factor from the LCs contributes to the inhibition spermatogonial differentiation in the irradiated rat, there might be other explanations, such as EDS action on gene expression in other cells. To prove this one way or another, it will be necessary to identify a specific factor or factors. The identification of factor(s) that regulate spermatogonial differentiation and non-hormonal methods of modulating such factors could lead to approaches of accelerating recovery of spermatogenesis in male cancer patients or even a male contraceptive.
This work was supported by grants ES 008075 from NIH/NIEHS to MLM, Cancer Center Support Grant CA P30 16672 from the NIH to MD Anderson Cancer Center, and the Florence M. Thomas Professorship in Cancer Research to MLM. We sincerely thank Drs. R.P. Blye, Hyun K. Kim, June Lee, and Min S. Lee of the National Institute for Child Health and Human Development for providing the acyline. We also thank Bryan Tutt in MD Anderson’s Department of Scientific Publications for the scientific editing of the manuscript.
The authors have no conflicting financial interests.
AUTHOR CONTRIBUTIONSGS & MLM, conception, design, analysis of results and preparation of the manuscript; WZ collection of data and analysis of results; GS, CCW and SHS collection of data.