Epidermal growth factor regulates the development of stem and progenitor Leydig cells in rats

Abstract Epidermal growth factor (EGF) has many physiological roles. However, its effects on stem and progenitor Leydig cell development remain unclear. Rat stem and progenitor Leydig cells were cultured with different concentrations of EGF alone or in combination with EGF antagonist, erlotinib or cetuximab. EGF (1 and 10 ng/mL) stimulated the proliferation of stem Leydig cells on the surface of seminiferous tubules and isolated CD90+ stem Leydig cells and progenitor Leydig cells but it blocked their differentiation. EGF also exerted anti‐apoptotic effects of progenitor Leydig cells. Erlotinib and cetuximab are able to reverse EGF‐mediated action. Gene microarray and qPCR of EGF‐treated progenitor Leydig cells revealed that the down‐regulation of steroidogenesis‐related proteins (Star and Hsd3b1) and antioxidative genes. It was found that EGF acted as a proliferative agent via increasing phosphorylation of AKT1. In conclusion, EGF stimulates the proliferation of rat stem and progenitor Leydig cells but blocks their differentiation.


| Animals
Male Sprague-Dawley rats were purchased from Shanghai Laboratory Animal Co. Ltd. Forty male rats at age of 7 days were used for the isolation of CD90 + SLCs each time. Forty male rats at age of 21 days were used for the isolation of PLCs each time. Male rats at age of 90 days were used for ST isolation and culture of SLCs.
The animal procedure was approved by the Institutional Animal Care and Use Committee of Wenzhou Medical University and was performed in accordance with the Guide for the Care and Use of Laboratory Animals.

| Purification and culture of CD90 + SLCs
Testicular CD90 + cells were thought to be the SLCs. 30 Purification and culture of CD90 + SLCs from 7-day-old male rats were performed. Peritubular testicular cells were obtained from collagenase (0.1 mg/mL)-treated isolated STs and stained using CD90 antibody, and purified using BD IMag™ bead. Cells were incubated with CD90 antibody (1:100) in BD IMag™ Buffer for 20 minutes on ice. Beads were incubated with CD90-conjugated cells for 30 minutes. After washing, the cells were separated by BD IMagnet™ for 10 minutes.
Cells were suspended in M199 medium. The purity of CD90 + cells (SLCs) was over 99%. To study the proliferation of SLCs, CD90 + cells (1 × 10 4 cells/well) in LDM medium were seeded in 12-well plate and incubated with control (LDM), EGF (10 ng/mL), Cet (an EGF antagonist, 5 µg/mL) and EGF (10 ng/mL)+Cet (5 µg/mL) for 24 hours. Then, cells were washed using PBS, and EdU incubation was performed as following section. Our previous study has demonstrated that the SLCs are cultured during the first week of culture, and the number of SLCs is greatly amplified. 30 Then, SLCs were switched to LDM for additional 14 days, the amplified SLCs could be differentiated into ALCs, and thus, the increased number of ALCs could contribute into the robust increase of T level in the medium. 30 Using this approach, EGF (0, 10 ng/mL) with or without EGF antagonist (Cet, 5 μg/mL) was added to SLCs (1 × 10 4 cells/well) in M199 and cultured at 34°C and 5% CO 2 for 7 days, and then, SLCs were switched into LDM for additional 14 days. Media were collected for the measurement of T level.

| ST isolation and culture of SLCs on the surface of STs
The procedure for ST isolation and culture was performed as previously described. 27,30 Briefly, EDS was dissolved in a 1:3 solution mixture of DMSO and water. One 90-day-old rat was selected and injected intraperitoneally with a single dose (75 mg/kg body weight) of EDS, which can effectively eliminate all LCs in the testis without damaging SLCs. 12,13 Seven days after EDS, LCs were all eliminated. 32 The rat was euthanized under CO 2 . Two testes were taken out and placed in cold M199 medium and decapsulated. STs were mechanically separated using a fine forceps under a microscope. 27 STs were cut to about 3-cm-long fragments and distributed randomly into 12-well plates, with each well containing equal amount of ST fragments. The STs were cultured at 34°C and 5% CO 2 for up to 14 days in LDM for induction of SLCs into ALCs to produce T as previously described. 27,30 Twelve isolations were performed. To study the effects of EGF on the proliferation of SLCs on the surface of STs, EGF (0, 1, and 10 ng/mL) was added to M199 medium and cultured at 34°C and 5% CO 2 for 7 days, during which SLCs have the highest capacity of proliferation. 30

| EdU incorporation into SLCs
The proliferative capacity of SLCs after EGF treatment was measured by EdU kit as previously described. 33 Media containing EGF and its inhibitor were removed for both CD90 + SLCs and STs above, and 2 µL of 1:1000 diluted EdU was added in the well and incubated at 34°C and 5% CO 2 for 24 hours. STs were washed twice with 500 µL phosphate-buffered saline (PBS) containing 3% BSA. Fragments of STs were also brought down in 2% argarose gel after centrifugation at 750 g, and the cross sections were cut. CD90 + SLCs and ST cross sections were then fixed in 4% paraformaldehyde at room temperature for 30 minutes. The CD90 + SLCs and STs were washed and incubated with reaction solution in the dark for 45 minutes. CD90 + antibody was used for staining CD90 + SLCs in purified SLCs. DAPI served the counterstaining. Then, CD90 + SLCs and STs were washed again and mounted on slide for visualization under fluorescence microscope (Olympus) and images were captured. EdU-positive cells (green fluorescence in cell nucleus) were counted and calculated by the total surface area of STs using the ImageProPlus 7.0 software (Media Cybernetics).

| SLC differentiation
Our previous study has demonstrated that STs are cultured in LDM for 14 days, and SLCs on the surface of STs are capable of differentiating into ALCs, which robustly produce T. 30 To study the effects of EGF on the differentiation of SLCs, EGF (0, 1, and 10 ng/mL) with or without EGF antagonist (erlotinib, E, 100 nmol/L) was added to LDM and cultured at 34°C and 5% CO 2 for 14 days. Then, media containing EGF were removed for measurement of T levels. Fragments of STs were also brought down in 2% argarose gel after centrifugation at 750 g, and the cross sections were cut. ST cross sections were then fixed in 4% paraformaldehyde at room temperature for 30 minutes. The STs were stained immunohistochemically after incubating HSD11B1 antibody (for ALC biomarker) and smooth actin muscle (SMA for peritubular myoid cells) and then Alexa Fluor 488 (green colour for HSD11B1) or Fluor 594 (red colour for α-SMA) for 1 hour. The sections were stained with DAPI for the nucleus of the cells. The slides were covered with 50% glycerol. The fluorescence was visualized under a fluorescent microscope (Olympus).

| SLC proliferation
Our previous study has demonstrated that the STs are cultured during the first week of culture, and the number of SLCs is greatly amplified. 30 Then, STs were switched to LDM for additional 7 days, the amplified SLCs could be differentiated into ALCs, and thus, the increased number of ALCs could contribute into the robust increase of T level in the medium. 30 Using this approach, EGF (0, 1 and 10 ng/mL) with or without EGF antagonist (E, 100 nmol/L) was added to M199 and cultured at 34°C and 5% CO 2 for 7 days, and then, STs were switched into LDM for additional 7 days. Media were collected for the measurement of T level.

| Isolation of PLCs
The procedure for PLC isolation and cultured was performed as previously described. 9 Briefly, forty 21-day-old rats were euthanized under CO 2 for isolation of PLCs. All testes were taken out and decapsulated and put in sterile 50-mL tube (10 mL medium).
Testes were digested in medium 199 containing 0.25 mg/mL collagenase-D at 34°C in the shaking bath (75 rpm) for 10 minutes.
The testis fragments were gently shaken and filtered through two layers of nylon mesh (200 μm) and washed with medium 199. The filtered cells were centrifuged at 250 g for 10 minutes. Crude cell preparations were resuspended in the 55% isotonic Percoll.
Following density gradient centrifugation at 25 000 g at 4°C for 45 minutes, the PLC fraction was gently collected between densities of 1.064 and 1.070 g/mL. The cells were washed with HBSS and centrifuged at 250 g for 10 minutes. PLCs were resuspended in phenol red-free 1:1 DMEM: F12 supplemented with 1 mg/ mL BSA. Purity of PLCs was judged after histochemical staining of HSD3B1 activity with 0.4 mmol/L etiocholanolone and 0.4 mmol/L NAD + as previously described. 34 The purity of PLCs was typically more than 95%. The purifications of PLCs were repeated for four times.
After the incorporation, PLCs were washed twice with PBS and harvested. PLCs were lysed in 0.5 mL hyamine hydroxide, and radioactivity was measured in a liquid scintillation counter (PE, USA). Cpm per 10 6 PLCs was calculated for thymidine incorporation into PLCs.  To evaluate early and lately apoptotic activity, an Annexin V-FITC/ PI Apoptosis Detection Kit (Nanjing KeyGEN Biotech) was used as previously described. 35 Cells were harvested and washed with cold PBS and then were resuspended in 200 μL the annexin V-binding buffer. After cells were stained with FITC-labelled annexin V and PI, they were instantly measured using flow cytometer.

| PLC steroidogenesis after EGF treatment
Progenitor Leydig cells with a density of 0.5 × 10 6 cells per cell were cultured with DMEM: F12 (1:1) alone or in combination with 1 and 10 ng/mL EGF at 34°C 5% CO 2 for 24 hours. Media were collected for measurement of AO and T. PLCs were washed twice with PBS and harvested for isolation of RNAs and proteins.

| Medium T and androsterone analysis
Medium concentrations of T and AO were measured by the tritiumbased radioimmunoassay validated for the use of rat antiserum as using either anti-T antibody (Fitzgerald, MA) or anti-AO antibody. 9 Standards ranging between 10 and 2000 pg/mL T or AO were prepared in triplicate. Standards and samples were incubated with respective tracer and antibody at 4°C overnight, and charcoal-dextran suspension was used to separate the bound and free steroids. The bound steroids were mixed with a scintillation buffer and counted in a β-scintillation counter (PE, USA). The minimum detectable concentration of the assay for either T or AO was 5 pg/mL. The quality control had either 100 pg/mL T or 100 pg/mL AO dissolved in the same culture media. Interassay and intra-assay coefficients of variation for T and AO were within 10%.

| Microarray hybridization and scanning
Progenitor Leydig cells were treated with 0, 1 and 10 ng/mL EGF as well as 1 ng/mL LH. LH serves the positive control for induction of PLC proliferation and differentiation. Total RNAs were harvested from PLCs after EGF treatment using a Trizol kit (Invitrogen) for microarray analysis. The RatRef-12 Expression BeadChip containing 21 910 rat genes was used as previously described. 36 Genes are selected from the NCBI RefSeq database to cover the whole rat transcriptome. Four groups of samples were used: 0, 1 and 10 ng/mL EGF-treated as well as 1 ng/mL LH-treated PLCs. Four replicates per group were performed. Probe labelling, hybridization, washing and scanning were performed using the Illumina Total Prep Kit (Applied Biosystems) as previously described. 37 First strand of cDNA was synthesized in a total volume of 20 μL with the supplied reagents.
The first-strand product was used for the second-strand synthesis, F I G U R E 1 Effects of EGF on proliferation of stem Leydig cells (SLCs). Leydig cell (LC)-depleted seminiferous tubules (STs) were cultured without (Panel A) or with 1 ng/mL EGF (Panel B), or 10 ng/mL EGF (Panel C) for 7 d. CD90 + SLCs were cultured without (Panel G) or with 10 ng/mL EGF (Panel H), or 5 μg/mL Cet (cetuximab, Panel I), or 10 ng/mL EGF + 5 μg/mL Cet (Panel J) for 24 h. EdU incorporation into the nucleus of SLC on the STs (white arrow) or CD90 + SLC (yellow arrow) was observed. DAPI serves as the counterstaining. EdU-positive cells were located outside of STs. Bar = 50 μm. The EdU-positive SLCs per mm 2 STs were presented in Panel D (mean ± SEM, n = 5). In Panel E, STs were cultured without or with 1 or 10 ng/mL EGF for 7 d, and then, STs were switched to Leydig cell differentiation medium (LDM) for additional 7 d to produce testosterone (T), and the effects of EGF on SLC proliferation were indirectly analysed (mean ± SEM, n = 4). In Panel F, STs were treated with 0 ng/kg EGF (control, C) or 100 nmol/L EGF antagonist (erlotinib, E), or 10 ng/mL EGF (EGF) or 100 nmol/L E plus 10 ng/mL EGF (E + EGF) for 7 d, and then, STs were switched to LDM for 14 d to produce T (mean ± SEM, n = 4). The EdU-positive CD90 + SLCs were presented in Panel K (mean ± SEM, n = 6). In Panel L, CD90 + SLCs were treated with 0 ng/kg EGF (control, C) or 5 μg/mL EGF antagonist (Cet), or 10 ng/mL EGF (EGF) or 5 μg/mL Cet plus 10 ng/mL EGF (Cet + EGF) for 7 d, and then, cells were switched to LDM for 7 d to produce T (mean ± SEM, n = 4). Identical letters designate no significant difference between two groups at P < .05 followed by column purification. The purified product was then

| Microarray data analysis
Microarray data analysis was performed as previously described. 36 Briefly, after scanning, the microarray data were imported into the BeadStudio software (Illumina) for normalization, preliminary analysis and filtering. The background subtraction was performed, and the Illumina custom error model was used to generate present/absent calls for each probe ("present" defined as P < .01 for signal detection) and to call differentially expressed genes (defined as P < .05 after false discovery rate correction). For each array, all probe sets were normalized to a mean signal intensity value of 100. Normalized data from BeadStudio were filtered to exclude genes not expressed in PLCs (ie data from probes that were classed as "absent" in all samples). All of the 21 910 genes were present in the data based on which further analyses were carried out. The data were further imported into Microsoft Access  2010, and queries to find the increased and decreased genes after EGF treatment when compared to the control were generated to find the expression levels.

| Biological pathway analysis
Biological pathway analysis was performed as previously described. 36 The Gene MicroArray Pathway Profiler 2.1 (GenMAPP2.1) software was used to find the biological pathway, and GO pathway was generated according to the software developer's instruction.
The GenMAPP2.1 was used to create a map of signal pathways for the potential pathways. We imported our statistical results into the program and illustrated biological pathways containing differentially expressed genes. The results of the differential gene expression profile were validated by RT-qPCR.

| Quantitative real-time PCR (RT-qPCR)
Briefly, first-strand synthesis of DNA and RT-qPCR were performed as previously described. 38 qPCR was carried out in a 20 µl volume in a 96-well plate using the SYBR Green PCR Kit from Applied Biosystems. Primer titration was performed with the concentration of 300 nmol/L. Fluorescence was detected using the ABI 7700 system (PE Applied Biosystems). Each sample was run in duplicate and in parallel with no template controls.
The relative mRNA levels of targeted genes were adjusted to housekeeping gene, ribosomal protein S16 (Rps16), as the internal control. Rsp16 in LCs has been used as the internal control in many studies because it showed consistent expression. 32,39 The Ct value was read, and the levels of the target mRNAs were calculated using the standard curve method as previously described. 36 All primers in the present study were designed by Primer  Table S1.

| Western blotting
Progenitor Leydig cells were homogenized and lysed. Protein concentrations in samples were measured using the Bio-Rad Protein Assay Kit (Bio-Rad) as previously described. 33     The density was scanned by ImageJ software.

| Statistics
Data were subjected to analysis by Student's t test to identify significant differences whenever two groups (a single concentration of EGF vs. control) were compared. Data were subjected to analysis by one-way ANOVA test followed by ad hoc Tukey multiple comparisons to identify significant differences between the tested group and the controls whenever there were three or more groups (multiple concentrations of etomidate vs control) were compared. All data are expressed as means ± SEM. Differences were regarded as significant at P < .05.

| EGF increases EdU incorporation into SLCs
To investigate the effects of EGF on SLC proliferation, we treated STs for 7 days with EGF and/or EGF inhibitor in vitro. SLCs reside on the surface of the ST (Figure 1A-C). 30 EGF concentration-dependently increased EdU incorporation into SLCs ( Figure 1D). This indicates that EGF stimulates SLC proliferation. Our previous study has shown that during the first week SLCs had the highest proliferative capacity and LDM was capable of inducing the differentiation of SLCs into ALCs, which robustly produced T after additional 7 days in LDM. 30 Using this indirect approach, we treated STs with 0-10 ng/ mL EGF for 7 days and then switched STs in LDM for additional 7 days to induce the formation of ALCs, which produced T. As shown in Figure 1E, EGF concentration-dependently increased medium T levels after 7-day culture, indicating that EGF is capable of increasing the number of SLCs that are differentiated into ALCs, which produce more T. We also tested EGF action using EGF antagonist (Erlotinib, E).
Indeed, E (100 nmol/L) alone did not affect T level but it can reverse EGF (10 ng/mL)-induced action ( Figure 1F). This indicates that EGF acts via EGF receptor. In order to test the effects of EGF on SLC proliferation, purified CD90 + SLCs were treated with EGF and/or its antagonist (Cetuximab, Cet) for 24 hours. EGF (10 ng/mL) significantly increased EdU incorporation into CD90 + SLCs ( Figure 1H,K) when compared to the control ( Figure 1G). Cet (5 µg/mL, Figure 1I) alone did not affect EdU incorporation but it can reverse EGF (10 ng/mL)induced action ( Figure 1J,K). Using the indirect approach, we treated CD90 + SLCs with 0-10 ng/mL EGF for 7 days and then switched CD90 + SLCs in LDM for additional 7 days to induce the formation of ALCs. As shown in Figure 1L, EGF increased medium T levels after 7-d culture, indicating that EGF is capable of increasing the number of SLCs that are differentiated into ALCs. This further confirms that EGF acts via EGF receptor to stimulate SLC proliferation.

| EGF inhibits differentiation of SLCs into the LC lineage
To examine whether EGF is capable of affecting SLC differentiation, STs were cultured in M199 for 7 days and then were switched to EGF (0, 1, and 10 ng/mL)-containing LDM for additional 7 days. We used HSD11B1 as the biomarker of ALCs.
After 7-day culture in LDM, many HSD11B1 positive ALCs were formed on the surface of STs, designating ALCs (Figure 2A). EGF antagonist (E, 100 nmol/L) did not affect the number of LCs ( Figure 2B). However, EGF (10 ng/mL) significantly reduced the number of HSD11B1 positive LCs, while EGF antagonist E reversed EGF-mediated inhibition ( Figure 2D). EGF concentrationdependently lowered LC number per ST ( Figure 2E) and medium T levels ( Figure 2F). EGF antagonist E did not affect medium T level but reversed EGF (10 ng/mL)-induced suppression of T synthesis ( Figure 2G). We further cultured CD90 + SLCs in the presence of EGF (10 ng/mL) and/or Cet (5 μg/mL) for 14 days. EGF antagonist Cet did not affect medium T output but reversed EGF (10 ng/ mL)-induced suppression of T synthesis ( Figure 2H). These data indicate that EGF inhibits SLC differentiation into the LC lineage via EGF receptor.

| EGF inhibits differentiation of PLCs
To examine whether EGF can affect PLC differentiation, PLCs were cultured with EGF (0, 1, and 10 ng/mL) in DMEM: F12 medium for 48 hours. EGF (10 ng/mL) significantly lowered AO and total androgen (T plus AO) production by PLCs ( Figure 3A). The data indicate that EGF inhibits PLC differentiation.

| EGF increases thymidine incorporation into PLCs
To investigate the effects of EGF on PLC proliferation, we treated PLCs for 48 hours in vitro. Both concentrations of EGF significantly increased thymidine incorporation into PLCs ( Figure 3B). This indicates that EGF stimulates PLC proliferation.

| EGF and H 2 O 2 induce ROS generation in PLCs
To investigate the effects of EGF on ROS generation by PLCs,

| EGF-mediated gene expression of PLCs
Luteinizing hormone is an important hormone for PLC development. 2 We used LH (1 ng/mL) as the positive control. PLCs were cultured with EGF (0, 1 and 10 ng/mL) in DMEM: F12 medium for 24 hours. Microarray of gene profiles of PLCs after EGF treatment were compared to LH (1 ng/mL). A genome-wide expression containing 21 910 probes was analysed. Of these probes, 8667 probes were detected in the control group. Of them, 33 genes were upregulated and 53 genes down-regulated more than twofold in 1 ng/ mL LH control group. Eighteen genes were up-regulated (Table 1) and 81 genes down-regulated (Table 2) more than twofold in 1 ng/ mL EGF group, while 66 genes were up-regulated (Table 1) and 198 genes down-regulated (Table 2) more than twofold in 10 ng/ mL EGF group. GO analysis showed that up-regulation of genes by EGF (10 ng/mL) includes the following 10 major categories: lipid metabolism, response to stimulus, steroid metabolism, hormone metabolic process, complement activation, activation of plasma proteins, humoral immune response, lipid transport, positive regulation of Wnt, axon guidance, response to nutrient, and immune response ( Figure 4A). The down-regulation of genes by EGF (10 ng/ mL) includes the following 10 major categories: motor axon guidance, collagen catabolism, peptidoglycan metabolism, neuron migration, regulation of apoptosis, cell adhesion, brain development, G-protein, amino acid dephosphorylation and organismal development ( Figure 4C). When compared to LH (1 ng/mL), most of categories up-regulated by EGF (10 ng/mL) were overlapped with those of LH ( Figure 4B), except one (humoral immune response) for EGF and genes were also up-regulated by EGF ( Figure 4E). When compared to LH, the categories down-regulated by EGF were more diversified ( Figure 4D).

| PCR analysis of gene expression after EGF treatment
We used PCR to analyse gene expression in EGF-treated PLCs.
We compared EGF with LH ( Figure 5). The following patterns were shown: (a) no change in EGF but up-regulation in LH (Scarb1,

| Major pathways after EGF treatment
Using GenMAPP2, we discovered several pathways that were specific to the regulation of EGF. The expression of matrix

F I G U R E 4 Effects of EGF and LH on gene expression in progenitor Leydig cells: GO analysis. After rat progenitor Leydig cells (PLCs)
were cultured with 0, 1 and 10 ng/mL EGF as well as 1 ng/mL LH for 24 h, cells were subjected to gene microarray analysis. Panels A and B: up-regulation by EGF and LH, respectively; Panels C and D: down-regulation by EGF and LH, respectively; and Panel E: overlapping genes in EGF or LH up-regulated genes  Figure 6). Interestingly, the antioxidant genes (Ugt1a6, Idh1, Gsst1, Gstm2 and Sod3) were down-regulated by EGF (Figure 6), indicating that EGF might lead to ROS accumulation. Of these 5 genes, two (Gsst1 and Gstm2) were also down-regulated by LH ( Figure 6).
F I G U R E 6 EGF up-regulates some matrix metalloproteinase expression and down-regulates some antioxidative protein expression in progenitor Leydig cells. After rat progenitor Leydig cells (PLCs) were cultured with 0, 1 and 10 ng/mL EGF as well as 1 ng/mL LH for 24 h, cells were subjected to gene microarray analysis. Expression of matrix metalloproteinases was analysed. Expression of antioxidative proteins was analysed

| EGF lowers steroidogenesis-related protein levels via EGFR signalling
We performed Western blot to identify changes of steroidogenesisrelated proteins (including StAR, HSD3B1 and CYP11A1) in PLCs after treatment with EGF and/or EGFR antagonist (E) for 24 hours.
As shown in Figure 7A, EGF concentration-dependently lowered StAR and HSD3B1 protein levels without affecting CYP11A1 and E can reverse the action of EGF (10 ng/mL). This confirms the EGFinduced mRNA changes.

| EGF lowers the phosphorylation of AKT1 signalling
We performed Western blot to identify the changes of AKT1 as well as its phosphorylated protein in PLCs after treatment with EGF and/or E for 24 hours. As shown in Figure 7B, EGF concentrationdependently increased phosphorylation of AKT1 without affecting total AKT1 level ( Figure 7B). This suggests that the phosphorylation of AKT1 is involved in EGFR pathway.

| D ISCUSS I ON
In the present study, we found that EGF stimulated SLC/PLC proliferation but blocked their differentiation.
There is a growing evidence to indicate that EGF is a critical proliferative growth factor for SLCs or PLCs during pubertal development. 13,[40][41][42] Apparently, EGF stimulated SLC and PLC proliferation in two-type LC precursor cells: SLCs on STs and CD90 + SLCs ( Figure 1) and isolated PLCs (Figure 3). Previous studies have demonstrated that CD90 + cells on the surface of STs were SLCs and can be induced into ALCs in LDM in vitro. 30  and Cd82) were also up-regulated by EGF (Table 1). Apparently, SLCs did not possess LHCGR, 13 and LH was not able to increase EdU incorporation into SLCs. 30   Previous studies indicated that EGF activated phosphatidylinositol 3-kinase (PI3K). 50 PI3K is able to phosphorylate AKT1, thus in turn activating cyclin D1 pathway. 50,51 The data presented here show that EGF indeed increased AKT1 phosphorylation without affecting total AKT1 protein levels ( Figure 7). This regulation of phosphorylation of AKT1 in PLCs is dependent on EGFR, because the EGF antagonist (E) can reverse EGF action (Figure 7).
The new data presented here also showed that activation of EGF inhibited H 2 O 2 -induced apoptosis in PLCs ( Figure 3). Indeed, EGF inhibited apoptosis in ILCs. 46 The present data confirmed what was found for EGF action in ILCs. As AKT1 signalling mediates both proliferative and anti-apoptotic effects, EGF-mediated anti-apoptotic action might be also mediated by AKT1.
Effects of EGF on T production in LCs are still controversial.
Several in vitro studies found that EGF inhibited LH or cAMP-induced steroid production in LCs. 52,53 Other studies indicated that EGF stimulated T production in LCs. 45,46,[54][55][56] The effects of EGF on androgen production depend on cell maturity, cell types of LCs, and EGF treatment duration.
In the current study, we clearly demonstrated that EGF Interestingly, the expression levels of many genes in antioxidative proteins (Ugt1a6, Idh1, Gsst1, Gstm2 and Sod3) were significantly down-regulated more than twofold after EGF treatment. This could lead to ROS accumulation in PLCs ( Figure 6). Indeed, EGF significantly led to ROS accumulation ( Figure 3C). Several studies demonstrated that the accumulation of ROS is capable of damaging testis function 60 and inhibiting T production by LCs. 61 Mitochondrion is an important organelle for regulation of T synthesis and is sensitive to ROS attack. [62][63][64] ROS is capable of disturbing StAR level. 65 Although ROS accumulation might lead to cell apoptosis, 66 the concentration of EGF (10 ng/mL) tested in the current study was not enough to induce PLC apoptosis ( Figure 3). Actually, EGF can antagonize H 2 O 2induced apoptosis (Figure 3).
TSP2 (encoded by Tsp2) was significantly down-regulated by EGF ( Figure 5). TSP2 functions the differentiation of mesenchymal stem cells into bones. 67 The down-regulation of Tsp2 by EGF could lead to stem/progenitor Leydig cells stay in the Leydig cell lineage not into other cell lineage as SLCs were multipotent and SLCs can also differentiate into bone cells. 8 In conclusion, the present study demonstrated that EGF induced the proliferation of SLCs/PLCs but blocks their differentiation into LCs.

ACK N OWLED G EM ENT
The present study was supported by NSFC (81730042, 81601266, 81871150, 81871155) and Department of Science and Technology of Zhejiang Province (2019C03035).

CO N FLI C T O F I NTE R E S T
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.