Growth Hormone-induced STAT5B Regulates Star Gene Expression Through a Cooperation With cJUN in Mouse MA-10 Leydig Cells

Abstract

Leydig cells produce androgens that are essential for male sex differentiation and reproductive function. Leydig cell function is regulated by several hormones and signaling molecules, including growth hormone (GH). Although GH is known to upregulate Star gene expression in Leydig cells, its molecular mechanism of action remains unknown. The STAT5B transcription factor is a downstream effector of GH signaling in other systems. While STAT5B is present in both primary and Leydig cell lines, its function in these cells has yet to be ascertained. Here we report that treatment of MA-10 Leydig cells with GH or overexpression of STAT5B induces Star messenger RNA levels and increases steroid hormone output. The mouse Star promoter contains a consensus STAT5B element (TTCnnnGAA) at −756 bp to which STAT5B binds in vitro (electrophoretic mobility shift assay and supershift) and in vivo (chromatin immunoprecipitation) in a GH-induced manner. In functional promoter assays, STAT5B was found to activate a −980 bp mouse Star reporter. Mutating the −756 bp element prevented STAT5B binding but did not abrogate STAT5B-responsiveness. STAT5B was found to functionally cooperate with DNA-bound cJUN. The STAT5B/cJUN cooperation was only observed in Leydig cells and not in Sertoli or fibroblast cells, indicating that additional Leydig cell–enriched transcription factors are required. The STAT5B/cJUN cooperation was lost only when both STAT5B and cJUN elements were mutated. In addition to identifying the Star gene as a novel target for STAT5B in Leydig cells, our data provide important new insights into the mechanism of GH and STAT5B action in the regulation of Leydig cell function.

The production of testosterone by Leydig cells is a finely regulated process involving several hormones and signaling pathways [reviewed in (1-3)]. The pituitary luteinizing hormone (LH) is the main regulator of Leydig cell steroidogenesis and its mechanism of action has been intensely studied [reviewed in (4)]. However, several other molecules have been implicated in the regulation of steroidogenesis, including cytokines, growth factors, and many hormones including growth hormone (GH) [reviewed in (1-3,5,6)].

In addition to systemic GH originating from the pituitary gland, local production of GH occurs in many tissues. For instance, GH gene expression has been detected within the testis of several species including rat, human, chicken, and eel [reviewed in (7)]. GH acts via its receptor (GHR), which is expressed in many cell types of the testis (8,9). Proper functioning of the GH/GHR system is essential for male reproductive function. In humans, naturally occurring mutations in the GHR gene result in delayed sexual maturation associated with Laron syndrome (10). In animal models, GH-deficient male rats (dwarf rat model) (11) and male Ghr knockout mice (12,13) exhibit smaller testes. Furthermore, secondary sex organs are underdeveloped and puberty is delayed in Ghr^−/− male mice (12,13). The fertility rate of Ghr^−/− male mice is also reduced (14). Ghr-deficient male mice have reduced intratesticular testosterone levels and the increase in testosterone production in response to LH is also dampened (12-14). These data suggest that GH mediates its effects, at least in part, by acting directly on Leydig cells.

Consistent with a direct action of GH on Leydig cell steroidogenesis, Leydig cells express GHR (8,15) and treatment of Leydig cells with GH increases testosterone production (15-17). Treatment of Leydig cells with GH was found to increase the messenger RNA (mRNA) and protein levels of steroidogenic acute regulatory protein (STAR) (15). STAR protein catalyzes the shuttling of cholesterol, the substrate for the synthesis of all steroid hormones, from the outer to the inner mitochondrial membrane where steroidogenesis is initiated (18). The presence of a functional STAR protein is essential for steroidogenesis as revealed by the phenotype of adrenal and gonadal steroid hormone insufficiency in humans with naturally occurring mutations in the STAR gene and in Star^−/− mice [reviewed in (19)]. Although the regulation of Star gene expression in Leydig cell has been well studied, the mechanism of GH-induced Star expression have yet to be elucidated.

A classical downstream effector of GH action in many cell types is STAT5B, a member of the signal transducer and activator of transcription family of transcription factors [reviewed in (20)]. STAT5B is expressed in Leydig cells (21,22) and upon GH treatment, STAT5B becomes phosphorylated and translocates to the nucleus (22). Activated STAT5B is known to stimulate gene expression by binding as a dimer to the consensus sequence TTCN₃GAA, called the γ-interferon-activated sequence (GAS), found in the promoter of target genes [reviewed in (23)]. Despite being present in Leydig cells and its activation by GH treatment, no target gene has been identified for STAT5B in these cells.

In the present work, we demonstrate that STAT5B directly activates Star gene expression in Leydig cells. Overexpression of a constitutively active form of STAT5B in MA-10 Leydig cells is sufficient to activate the Star promoter, increase endogenous Star mRNA levels, and raise steroid hormone output. STAT5B is recruited to the Star promoter and binds to a perfect GAS element located at −756 bp. Furthermore, STAT5B functionally cooperates with cJUN on the Star promoter, and this cooperation is abrogated only when the GAS element at −756 bp and the AP-1 element at −78 bp are both mutated. Our data provide a direct link between GH-activated STAT5B and Star gene expression in Leydig cells.

Materials and Methods

Plasmids

The mouse Star luciferase reporter constructs (−980, −195, −144, −95, −70, and −43 bp to +16 bp) have been described previously (24-31). The −980 bp Star reporter with an inactivating mutation (shown in lowercase) in the AP-1 element at −78 bp (TGACTGATG to TGAgaGtTG) was previously described (26). The −980 bp Star reporter with a mutation (shown in lowercase) inactivating the GAS element at −756 bp (TTCCCAGAA to TTCCCActt), or a double mutation of the GAS and AP-1 elements, was generated by site-directed mutagenesis using the QuikChange XL mutagenesis kit (Stratagene, La Jolla, CA, USA) using either the wild-type or AP-1-mutated Star reporter as a template and the following pair of oligos (mutations are in lowercase): GAS element at −756: sense: 5’-GTC ATC TCA TTT CCA GAG AGa agC CAG AAT GAG AAG TTA GAG TG-3’, antisense: 5’- CAC TCT AAC TTC TCA TTC TGG ctt CTC TCT GGA AAT GAG ATG AC-3’. The rat Nur77 luciferase reporter constructs (−1007 and −59 bp to +53 bp) have been described previously (32). The GATA4 expression vector has been described previously (33). The expression vector for a constitutively active form of STAT5B (34) was kindly provided by Dr. Toshio Kitamura (University of Tokyo, Tokyo, Japan). The cJUN expression vector (35) was obtained from Dr. Dany Chalbos (Institut National de la Santé et de la Recherche Médicale, Endocrinologie Moléculaire et Cellulaire des Cancers, Montpellier, France). An empty expression vector served as control for all transfection experiments.

Cell Culture, Growth Hormone Treatment, and Transfections

The MA-10 mouse Leydig cell line (36) and the MSC-1 mouse Sertoli cell line (37) were provided by Dr. Mario Ascoli (University of Iowa, Iowa City, IA, USA) and Dr. Michael Griswold (Washington State University, Pullman, WA, USA), respectively. Mouse MLTC-1 Leydig cells (ATCC Cat# CRL-2065, RRID:CVCL_3544; https://web.expasy.org/cellosaurus/CVCL_3544) and African green monkey kidney CV-1 fibroblast cells (ATCC Cat# CRL-6305, RRID:CVCL_0229; https://web.expasy.org/cellosaurus/CVCL_0229) were obtained from ATCC. All cell lines were grown as previously described (25,28,30-32,38,39). Leydig cell lines were validated by morphology (as other cell lines) and by quantifying steroidogenic output (progesterone for MA-10 and testosterone for MLTC-1). To measure promoter activity, transient transfections were performed using the calcium phosphate coprecipitation method as described in Martin et al (25,32). Briefly, the cells were plated in 24-well plates and cotransfected with 500 ng of reporter vector along with 250 ng of expression vectors (empty expression vector pcDNA3 as control, STAT5B CA, cJUN, GATA4, or in combination). The following morning, the media was replaced, and 24 hours later the cells were lysed and luciferase assays performed as previously described (25,29,32). When indicated, cells were treated with 100 ng/mL of human GH (hGH; Sigma-Aldrich Canada, Oakville, Canada). For overexpression experiments, MA-10 Leydig cells (500 000 cells/60 mm plate) were transfected using polyethylenimine hydrochloride (Sigma-Aldrich Canada) as previously described (29) at a 3:1 ratio of polyethylenimine hydrochloride to DNA (w/w) with either 5 µg of an empty expression vector (pcDNA3), 2.5 µg of an expression vector for STAT5B CA plus 2.5 µg of pcDNA3, 2.5 µg of a cJUN expression vector plus 2.5 µg of pcDNA3, or 2.5 µg of both STAT5B CA and cJUN expression vector. The next morning, the media was replaced, and 24 hours later, a 500 µL-aliquot of media was harvested for steroid hormone quantification (described in the following text), and whole-cell extracts were prepared as described previously (25) for use in Western blots (described in the following text). For hGH treatment, MA-10 and MLTC-1 Leydig cells were seeded in 6-well plates at 500 000 cells/well. The next morning, the cells were rinsed twice with phosphate-buffered saline (PBS) and the media was replaced with serum-free media. After 2 hours, 100 ng/mL of hGH (Sigma-Aldrich Canada) was added for 0, 5, 15, 30, 60, 120, 240, and 360 minutes. At each time point, a 500 µL-aliquot of media was harvested for steroid hormone quantification, nuclear and cytoplasmic extracts were prepared as described in Schreiber et al (40), and total RNA was isolated as described in the following text.

Protein Purification and Western Blots

Nuclear and cytoplasmic extracts were prepared as described in Schreiber et al (40). For all Western blots, 10 µg-aliquots of protein were heated for 10 minutes in a denaturing loading buffer, separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and transferred onto polyvinylidene fluoride membrane (Sigma-Aldrich Canada). After blocking in 5% nonfat milk in 1× Tris-buffered saline with 0.1% Tween 20 detergent for 30 minutes, membranes were rinsed and incubated overnight at 4°C in 5% nonfat milk containing a rabbit polyclonal anti-STAT5B antiserum (1:200, Santa Cruz Biotechnology Cat# sc-835, RRID:AB_632446; http://antibodyregistry.org/AB_632446) or a mouse monoclonal anti-STAT5B antibody (1:200, Cat# sc-1656, RRID:AB_2197067; http://antibodyregistry.org/AB_2197067), a rabbit polyclonal anti-STAR antiserum (1:500, Santa Cruz Biotechnology Cat# sc-25806, RRID:AB_2115937; http://antibodyregistry.org/AB_2115937, or 1:5000, Cell Signaling Technology Cat# 8449, RRID:AB_10889737; http://antibodyregistry.org/AB_10889737), a rabbit polyclonal anti-cJUN antiserum (1:500, Santa Cruz Biotechnology Cat# sc-45, RRID:AB_2129862; http://antibodyregistry.org/AB_2129862), a mouse monoclonal anti-GAPDH antibody (1:5000, Santa Cruz Biotechnology Cat# sc-32233, RRID:AB_627679; http://antibodyregistry.org/AB_627679), a goat polyclonal anti-LMNB1 antiserum (1:500, Santa Cruz Biotechnology Cat# sc-6216, RRID:AB_648156; http://antibodyregistry.org/AB_648156), and a mouse monoclonal anti-αTUBULIN antibody (1:100000, Sigma-Aldrich Cat# T9026, RRID:AB_477593; http://antibodyregistry.org/AB_477593). The next morning, the membranes were washed in PBS and incubated for 1 hour at room temperature with the corresponding secondary biotinylated antibody [1:1000, rabbit antigoat antibody, Vector Laboratories Cat# BA-5000, RRID:AB_2336126; http://antibodyregistry.org/AB_2336126; goat antimouse antibody, Vector Laboratories Cat# BA-9200, RRID:AB_2336171; http://antibodyregistry.org/AB_2336171; or goat antirabbit immunoglobulin G (IgG) antibody, Vector Laboratories Cat# BA-1000, RRID:AB_2313606; http://antibodyregistry.org/AB_2313606] and revealed using the Vectastain ABC-AmP Reagent (Vector Laboratories Cat# AK-6000, RRID:AB_2336806; http://antibodyregistry.org/AB_2336806). Immunodetection was also performed using horseradish peroxidase-conjugated secondary antibodies at 1:5000, incubation 1 hour at room temperature (rabbit antigoat antibody, Jackson ImmunoResearch Labs Cat# 305-035-003, RRID:AB_2339400; http://antibodyregistry.org/AB_2339400; goat antimouse antibody, Jackson ImmunoResearch Labs Cat# 115-035-003, RRID:AB_10015289; http://antibodyregistry.org/AB_10015289; or goat antirabbit antibody, Jackson ImmunoResearch Labs Cat# 111-035-003, RRID:AB_2313567; http://antibodyregistry.org/AB_2313567) and either the Clarity Western ECL Substrate or Clarity Max Western ECL Substrate (Bio-Rad Laboratories, Québec, Canada).

RNA Isolation and Quantitative Reverse Transcription Polymerase Chain Reaction

RNA isolation, complimentary DNA (cDNA) synthesis, and quantitative reverse transcription polymerase chain reaction (PCR) were performed as previously described (25). Briefly, total RNA from MA-10 and MLTC-1 Leydig cells grown and treated as previously described was isolated using TRIZOL (Life Technologies, Burlington, Canada) and reverse transcribed using either the Transcriptor Reverse Transcriptase kit (Roche Diagnostics, Laval, Canada) or the iScript Advanced cDNA Synthesis Kit (Bio-Rad Laboratories). Quantitative real-time PCR was performed using a LightCycler 1.5 instrument (Roche Diagnostics) or a C1000 Thermal Cycler (Bio-Rad Laboratories) along with the LightCycler FastStart DNA Master SYBR Green I kit (Roche Diagnostics) or the SsoAdvanced Universal SYBR Green Supermix kit (Bio-Rad Laboratories) according to the manufacturer’s protocols. PCRs to detect Star were performed using the following specific primers: forward, 5’-TTG GGC ATA CTC AAC AAC CA-3’, and reverse, 5’-CCT TGA CAT TTG GGT TCC AC-3’. PCRs to detect Stat5b were performed using the following specific primers: forward, 5’-GAG CCC GCA ACT GCG AAA G-3’, and reverse, 5’-TCA TGA CTG TGC GTG AGG GA-3’. PCRs to detect Cjun were performed using the following specific primers: forward, 5’- TGG GCA CAT CAC CAC TAC AC-3’, and reverse, 5’- TCT GGC TAT GCA GTT CAG CC-3’. As an internal control, PCRs were performed using Rpl19-specific primers (25). The PCRs were done using the following conditions: 10 minutes at 95°C, followed by 35 cycles of denaturation (5 seconds at 95°C), annealing (5 seconds at 62°C), and extension (20 seconds at 72°C) with single acquisition of fluorescence at the end of each extension step. PCR product specificity was confirmed by melting curve analysis and agarose gel electrophoresis. Quantification of gene expression was performed using the Relative Quantification Software (Roche Diagnostics) and is expressed as a ratio of Star to Rpl19 mRNA levels. Each amplification was performed in duplicate using 3 different preparations of first strand cDNAs for each of the 3 different RNA extractions.

Immunohistochemistry

Immunohistochemistry was performed as described in (25,28,39,41) using sections of CD-1 mouse testis at postnatal day 32. Testes were fixed with ice cold 4% paraformaldehyde (w/v) for 24 hours, dehydrated with ethanol, cleared with xylene, embedded in paraffin, and cut into 5-µm sections. Following paraffin removal in xylene, the tissue was incubated for 30 minutes in 0.3% H₂O₂ (Sigma-Aldrich, Oakville, Canada)/methanol, rehydrated in graded alcohols (95%, 70%, and 50%), and treated for antigen retrieval by heating in a microwave oven for 15 minutes (heat at 800 watts for 10 seconds, no heat for 20 seconds) in a 0.01 M citrate buffer pH 6.0. After cooling down to room temperature, sections were blocked in 0.1% to 0.5% bovine serum albumin in PBS for 1 h at 25°C and immunolocalization of STAT5B was performed using an anti-STAT5B antiserum (1:500 Santa Cruz Biotechnology Cat# sc-835, RRID:AB_632446) overnight at 4°C. The next morning, slides were washed in PBS and incubated for 1 hour at room temperature with a biotinylated anti-rabbit IgG (1:1000, Vector Laboratories Cat# BA-1000, RRID:AB_2313606). After washing in PBS, sections were submitted to an avidin-biotin complex solution for 30 minutes at room temperature (Vector Laboratories Cat# PK-6100, RRID:AB_2336819). The negative control corresponds to the same procedure with the omission of anti-STAT5B antiserum (data not shown). Final revelation was done using 3-amino-9-ethylcarbazole as substrate (Sigma-Aldrich Canada), and the sections were counterstained with hematoxylin Gill 1 (VWR International, Mount-Royal, PQ). All experiments were conducted according to the Canadian Council for Animal Care and have been approved by the Animal Care and Ethics Committee of Laval University (protocol # 2009011).

Electromobility Shift Assay

Electromobility shift assays were performed as previously described (25,38,39) using 5 µg of nuclear extracts from untreated or hGH-treated (100 ng/mL for 2h) MA-10 Leydig cells in 20 µL of 4 mM Tris-HCl (pH 8.0), 24 mM KCl, 0.5 mM EDTA (pH 8.0), 0.4 mM dithiothreitol, 5 mM MgCl₂, 10% glycerol, and 1 µg poly(dI-dC) for 1 hour on ice. A  ³²P-labeled double-stranded oligonucleotide corresponding to the −756 bp GAS element in the mouse Star promoter was used as probe (the GAS element is underlined): sense 5’-AGA GAG TTC CCA GAA TGA GA-3’ and antisense 5’-TCT CAT TCT GGG AAC TCT CT-3’. Competition experiments were performed using 5× and 25× (molar excess) of unlabeled double-stranded oligonucleotides corresponding to the probe (WT GAS sequence) or harboring a mutation (shown in lowercase) known to prevent STAT5 binding (sense 5’-AGA GAG aag CCA GAA TGA GA-3’). For supershift experiments, 4 µg of an anti-STAT5B (Santa Cruz Biotechnology Cat# sc-835, RRID:AB_632446) antiserum or a nonimmune IgG were also added to the binding reaction.

Progesterone and Testosterone Quantification

Enzyme-linked immunosorbent assays for progesterone (Cayman Chemical Cat# 582601, RRID:AB_2811273; http://antibodyregistry.org/AB_2811273) and testosterone (Cayman Chemical, Cat# 582701, RRID:AB_2895148; http://antibodyregistry.org/AB_2895148) quantification were performed as previously described (28,30,31,42) according to the manufacturer’s recommendations (Cayman Chemical Company, Ann Arbor, MI, USA). To quantify progesterone produced by MA-10 cells, a dilution of 1:20 [50 µL of media in 950 µL of enzyme immunoassay (EIA) buffer] was used for media from overexpressing cells while a dilution of 1:15 (30 µL of media in 420 µL of EIA buffer) was used for media from hGH-treated cells. To quantify testosterone produced by MLTC-1 cells, a dilution of 1:10 (50 µL of media in 450 µL of EIA buffer) was used for media from hGH-treated cells. Each experiment was repeated 3 times in duplicate.

Chromatin Immunoprecipitation Assays

Chromatin immunoprecipitation (ChIP) assays were performed as previously described with some modifications (29,43). MA-10 Leydig cells were cultured in 150-mm dishes until they reached ~80 % confluency and treated for 2 hours with 100 ng/mL of hGH. Next, the cells were crosslinked using 1% formaldehyde for 10 minutes. The crosslinking reaction was stopped by addition of glycine for 2 to 5 minutes. Then the cells were rinsed with PBS, scraped, centrifuged, and stored at −80°C. Next, ~12.5 million cells were resuspended in 0.4 mL lysis buffer supplemented with 1 mM phenylmethylsulfonyl fluoride (PMSF), 1 µg/mL pepstatin, 1 µg/mL leupeptin, and 1 µg/mL aprotinin; incubated on ice for 10 minutes; and sonicated using Misonix Ultra sonic processor part S-4000. The supernatants were diluted in 20 mM Tris pH 8.0, 2 mM EDTA, 0.1% sodium dodecyl sulfate (SDS), 150 mM NaCl, 1% Triton X-100, 1 mM PMSF, 1 µg/mL pepstatin, 1 µg/mL leupeptin, and 1 µg/mL Aprotinin. The amount of the chromatin was quantified by Nanodrop, and the shearing efficiency was determined by running an aliquot on an agarose gel (1× TAE). Immunoprecipitation was performed by using 100 µg of sheared chromatin, 50 µL magnetic Dynabeads protein A beads (Invitrogen, Ontario, Canada), 5 µg of a STAT5B antibody (Santa Cruz Biotechnology Cat# sc-835, RRID:AB_632446) or a rabbit IgG as control (Thermo Fisher Scientific Cat# 02-6102, RRID:AB_2532938; http://antibodyregistry.org/AB_2532938) in Tris/saline/EDTA (TSE) I supplemented with 1 mM PMSF, 1 µg/mL pepstatin, 1 µg/mL leupeptin, 1 µg/mL aprotinin at 4°C overnight with constant rotation. The beads were separated by a magnet and washed 1 time with TSE I, 2 times with TSE II, 3 times with buffer III, and 3 times with Tris/EDTA. The beads were resuspended in 200 µL of 10 mM Tris-HCl pH 8.0, 0.3 M NaCl, 5 mM EDTA, 0.5% SDS, 50 µg/mL RNase A and incubated for 4 h at 65°C. The supernatant was collected using a magnet. To prepare the reference input control sample, 200 µL of sheared chromatin aliquot was diluted with 275 µL of Tris/EDTA, 30 µL 5 M NaCl, and 25 µL of 20% SDS; RNAse A was added; and the sample was incubated for 4 hours at 65°C. Proteinase K was added to the supernatant and reference input control sample, and incubation was continued overnight to complete decrosslinking. The chromatin was purified by phenol/chloroform, precipitated by ethanol, and resuspended in 10 mM Tris-Cl, pH 8.5. ChIP DNA fragments and reference input samples were diluted 10-fold and quantified by quantitative PCR (qPCR) using specific primers targeting the Star proximal promoter region (−299/−41 bp; forward: 5’-TGA TGC ACC TCA GTT ACT GG-3’, reverse: 5’-GCT GTG CAT CAT CAC TTG AG-3’). The qPCR experiments were performed using SsoAdvanced Universal SYBR Green Supermix (Bio-Rad Laboratories) and analyzed as previously described (29). The analyzed and plotted data come from 3 independent experiments.

Statistical Analyses

Single comparisons between 2 experimental groups were done using a Mann-Whitney U (Wilcoxon rank-sum) nonparametric test. To identify significant differences between multiple groups, statistical analyses were done using a nonparametric 1-way analysis of variance on ranks Kruskal-Wallis test. For all statistical analyses, P < 0.05 was considered significant.

Results

Induction of STAT5B by Growth Hormone Precedes That of STAR in MA-10 Leydig Cells

Stimulation of Leydig cells with GH has been shown to increase Star mRNA levels (15) and to induce phosphorylation of STAT5B (22). Whether STAT5B is involved in the regulation of Star gene expression remains to be demonstrated. To address this question, we first tested whether STAT5B is expressed in the MA-10 Leydig cell line and whether this cell line is responsive to GH stimulation. As shown in Figure 1A, a band corresponding to STAT5B is detected in control MA-10 Leydig cells (CTL), and this band is markedly increased in cells overexpressing a constitutively active form of STAT5B (STAT5B CA). This demonstrates that STAT5B is expressed in MA-10 Leydig cells and that the anti-STAT5B antiserum recognizes the STAT5B protein. Immunohistochemistry on adult mouse testis sections was performed and confirmed the presence of STAT5B in Leydig cells (Fig. 1B). We next determined the kinetics of STAT5B and STAR protein upregulation in MA-10 Leydig cells in response to GH. MA-10 Leydig cells were treated with GH for different durations (0-120 minutes), and protein extracts were prepared and used to assess STAT5B and STAR protein levels. As shown in Figure 1C, within 15 minutes of GH treatment STAT5B nuclear levels were increased concomitant with a reduction in cytoplasmic levels, which consistent with the fact that activated STAT5B translocates to the nucleus as reported previously (22). Treatment with GH increased STAR protein levels but at a later time point (30-60 minutes). These data indicate that MA-10 Leydig cells respond to GH and that GH-dependent induction of STAT5B precedes that of STAR.

Figure 1.

STAT5B induction in response to human growth hormone (hGH) precedes that of STAR in MA-10 Leydig cells. (A) MA-10 Leydig cells were transfected with an empty expression vector (CTL) or an expression vector for a constitutively active STAT5B (STAT5B CA) and nuclear extracts were prepared for immunodetection of STAT5B. LAMIN B1 (LMNB1) was used as a loading control. This experiment was repeated at least 5 times and produced similar results. A representative image is shown. (B) STAT5B is present in interstitial Leydig cells of the mouse testis. Immunohistochemistry was performed on P32 mouse testis section using an anti-STAT5B antiserum. Brownish staining is observed in interstitial Leydig cells (arrows). (C) MA-10 Leydig cells were treated with hGH for different durations as indicated after which nuclear and cytoplasmic extracts were prepared for immunodetection of STAT5B and STAR. LAMIN B1 (LMNB1) αTUBULIN, and GAPDH were used as loading control.

Next, STAT5B CA (constitutively active form of STAT5B) was overexpressed in MA-10 Leydig cells and Star mRNA levels were quantified by qPCR. As shown in Figure 2A, endogenous Star mRNA levels were increased by 2-fold in STAT5B CA–transfected cells. This is similar to the increase in Star mRNA levels observed following a stimulation with GH, whereas Stat5b mRNA levels remained unchanged (Fig. 2B). Taken together, these data are consistent with a role for STAT5B in the regulation of Star gene expression in Leydig cells.

Figure 2.

Human growth hormone (hGH) and STAT5B increase endogenous Star mRNA levels. MA-10 Leydig cells were transfected with an empty expression vector (CTL) or an expression vector coding for a constitutively active form of STAT5B (STAT5B CA) (A) or treated with hGH for different durations as indicated (B). Total RNA were prepared and Star (grey line) and Stat5b (black line) messenger RNA levels were quantified by quantitative polymerase chain reaction. Data were corrected with the Rpl19 housekeeping gene. The number of experiments is indicated. An asterisk indicates a statistically significant difference (P < 0.05) from the control (time 0).

Since STAR is involved in steroid hormone biosynthesis and since GH and STAT5B increase Star mRNA and STAR protein levels, we tested whether treatment with GH increases the steroidogenic output of Leydig cell lines. As shown in Figure 3, treatment of 2 Leydig cell lines, MA-10 (Fig. 3A) and MLTC-1 (Fig. 3B), with GH led to a rapid and progressive increase in progesterone (MA-10) and testosterone (MLTC-1) production. These data are in agreement with a previous study reporting an increase in androgen production by primary Leydig cells stimulated with GH (15).

Figure 3.

Treatment with hGH increases steroid hormone production in Leydig cells. MA-10 (A) and MLTC-1 (B) Leydig cells treated with human growth hormone for different durations as indicated, and progesterone (A) and testosterone (B) secreted in the media was quantified by enzyme-linked immunosorbent assay. The number of experiments is indicated. For a given cell line, an asterisk indicates a statistically significant difference (P < 0.05) from the control (time 0).

STAT5B Binds to and Activates the Mouse Star Promoter

Sequence analysis of the first ~1000 bp upstream of the transcription start site of the mouse Star promoter revealed the presence of 5 potential STAT binding motifs (GAS elements). The GAS element at −756 bp is a perfect match with the consensus sequence TTCnnnGAA [reviewed in (23)], while the other 4 GAS elements contain a single nucleotide mismatch (Fig. 4A). To test whether STAT5B could bind to the consensus GAS element at −756 bp, electromobility shift assays were performed using as a probe the sequence of the −756 bp GAS element. As shown in Figure 4B, no binding was detected when nuclear extracts from untreated MA-10 cells were used (Fig. 4B, lane 2). However, a band was detected in nuclear extracts from GH-treated MA-10 cells (Fig. 4B, lane 3). This binding was efficiently competed by an increasing concentration of unlabeled wild-type oligonucleotides (Fig. 4B, lanes 4 and 5) but not by oligonucleotides harboring a mutation in the GAS sequence (TTCCCAGAA to aagCCAGAA) known to prevent STAT5 binding (Fig. 4B, lanes 6 and 7). To determine whether the protein binding to the GAS element was STAT5B, supershift assays were performed. As shown in Figure 4C, strong binding is only detected after treatment with GH (Fig. 4C, compare lanes 6-8 with lanes 2-5). If the protein responsible for this binding is STAT5B, addition of an anti-STAT5B antiserum should result in a supershift while incubation with a nonimmune control IgG should have no effect. The strong binding detected in nuclear extracts from GH-treated MA-10 Leydig cells (Fig. 4C, lane 6) was completely supershifted in the presence of an anti-STAT5B antiserum (Fig. 4C, lane 8), while the control IgG had no effect (Fig. 4C, lane 7). These data confirm that STAT5B present in nuclear extracts from GH-treated MA-10 Leydig cells can bind specifically to the GAS element at −756 bp in the Star promoter.

Figure 4.

STAT5B binds to the −756 bp γ-interferon-activated sequence (GAS) element in the mouse Star promoter. (A) The mouse Star promoter contains 5 potential STAT5 binding sites (GAS elements), 1 perfect at −756 bp and 4 containing a 1-nucleotide mismatch (shown in lowercase). The GAS elements are represented by the black diamonds. The sequence of the consensus GAS element is shown. (B) Electromobility shift assays was used to determine the binding of STAT5B to the −756 bp consensus GAS element in the mouse Star promoter. Nuclear extracts were isolated from MA-10 Leydig cells either unstimulated or stimulated with 100 ng/mL of hGH for 2 hours. Protein binding was challenged by increasing doses (black triangles; molar excesses of 5× and 25×) of unlabeled oligonucleotides corresponding to the probe (WT GAS sequence) or harboring a mutation known to prevent STAT5 binding (TTCCCAGAA to aagCCAGAA). (C) Supershift (Ss) experiments were performed by adding 4 µg of an anti-STAT5B antiserum or a nonimmune control immunoglobulin G to nuclear extracts isolated from unstimulated or GH-stimulated MA-10 Leydig cells. Abbreviation: n.s., nonspecific.

Next, to test whether STAT5B can activate the Star promoter, transient transfections were first performed in MA-10 Leydig cells with a −980 bp mouse Star reporter construct along with increasing amounts of a STAT5B CA expression plasmid. As shown in Figure 5A, the mouse Star promoter was activated about 3-fold in the presence of STAT5B CA. Since the mouse Star contains 1 perfect GAS element at −756 bp to which STAT5B can bind (Fig. 4B and 4C), we tested whether the integrity of this GAS element was essential for STAT5B responsiveness. As shown in Figure 5B, a −980 bp Star reporter harboring a mutation that eliminates the GAS element (TTCCCAGAA to aagCCAGAA) preventing the binding of STAT5B (Fig. 4B) was still activated by STAT5B in transient transfection assays. This result confirmed that STAT5B can activate the mouse Star promoter but that the perfect GAS element at −756 bp is not essential for STAT5B-dependent activation.

Figure 5.

STAT5B activates the mouse Star promoter. (A) MA-10 Leydig cells were cotransfected with a −980 bp mouse Star reporter and increasing amount (25, 50, 125, 250, 500 ng) of an expression vector encoding a constitutively active (CA) STAT5B (black bars). The number of experiments is indicated. Results are shown as fold activation over control (cells transfected with an empty expression vector, white bar). Different letters (a, b, c, or d) indicate a statistically significant difference between groups (P < 0.05). (B) MA-10 Leydig cells were cotransfected with a −980/+16 bp Star reporter either wild-type or harboring a mutation (TTCCCAGAA to aagCCAGAA, depicted by a large X) in the γ-interferon-activated sequence (GAS) element at −756 bp, along with either an empty expression vector (open bars) or an expression vector for a constitutively active STAT5B (STAT5B CA, black bars). The position of the 5 potential GAS elements is shown (diamonds). The number of experiments is indicated. Results are shown as fold activation over control (± SE of the mean). For a given reporter, the asterisk indicates a statistically significant difference (***P < 0.001) from the control (empty expression vector).

To locate the STAT5B responsive element, a series of 5’ progressive deletion constructs of the mouse Star promoter were transfected in MA-10 Leydig cells with or without an expression vector for STAT5B CA. As shown in Figure 6A, a deletion to −95 bp that removes all 5 GAS elements was still activated 2.7- to 3-fold in the presence of STAT5B, as was the −980 bp construct. Further deletion to −70 bp resulted in a significant reduction in STAT5B responsiveness, down to 1.4-fold, while a deletion to −43 bp was no longer activated by STAT5B (Fig. 6A). These data indicate that a 53 bp region, from −95 to −43 bp, is essential for STAT5B-mediated activation of the Star promoter in Leydig cells. Since this region is known to contain a binding site for the nuclear receptor NUR77 (NR4A1, NGFI-B), which is a key regulator of hormone-induced Star expression (25,26,44,45), we tested whether STAT5B could also activate the Nur77 promoter. As shown in Figure 6B, a −1007 to +53 bp Nur77 reporter, which contains putative GAS elements, was activated 3-fold by STAT5B CA while a minimal −59 bp reporter was unresponsive. So in addition to directly activating the Star promoter, STAT5B may also act by stimulating the expression of Nur77, an important activator of Star transcription.

Figure 6.

The STAT5B-responsive element maps to the proximal region of the Star promoter. (A) MA-10 Leydig cells were cotransfected with various 5’ deletion constructs of the mouse Star promoter (the 5’-end point of each construct is indicated on the left of the graph) along with either an empty expression vector (open bars) or an expression vector for a constitutively active STAT5B (STAT5B CA, black bars). The position of the 5 potential γ-interferon-activated sequence elements is shown (diamonds). (B) The Nur77/Nr4a1 promoter is activated by STAT5B CA. MA-10 Leydig cells were cotransfected with Nur77 reporters as indicated (−1007/+53 bp and −59/+53 bp) along with either an empty expression vector (open bars) or an expression vector for a constitutively active STAT5B (STAT5B CA, black bars). The number of experiments is indicated. Results are shown as fold activation over control (± SE of the mean). For a given reporter, the asterisk indicates a statistically significant difference (*P < 0.05; ***P < 0.001) from the control (empty expression vector).

STAT5B Cooperates With cJUN on the Mouse Star Promoter

In addition to the NUR77 binding site at −91 bp, the 53 bp region (−95 to −43 bp) of the Star promoter contains 2 previously characterized elements for the binding of and activation by AP-1 (cJUN) at −78 bp (24,46) and by GATA4 at −64 bp (24,27) in Leydig cells, elements that are conserved across species (Fig. 7A). We therefore tested whether STAT5B can cooperate with either GATA4 or cJUN to activate the Star promoter in MA-10 Leydig cells. As shown in Figure 7B, both STAT5B and GATA4 were found to activate the −980 bp Star promoter individually, as expected (24,27), but no cooperation was observed when both factors were combined. As previously reported (24), cJUN was found to activate the −980 bp Star reporter by about 10-fold (Fig. 7C). Combination of cJUN and STAT5B CA led to a 17-fold cooperative activation of the Star promoter (Fig. 7C). Furthermore, treatment of MA-10 Leydig cells with GH revealed that cJun mRNA (Fig. 8A) and to a lesser extent cJUN protein (Fig. 8B) levels were increased in response to GH treatment, supporting a role for this transcription factor in GH responsiveness. We next tested whether the STAT5B/cJUN cooperation was also observed in other cell types. As shown in Figure 9, the cooperation between STAT5B and cJUN was observed in 2 Leydig cell lines (MA-10 and MLTC-1) but not in MSC-1 Sertoli cells or CV-1 fibroblast cells. These results indicate that other factors present in Leydig cells but absent from Sertoli or fibroblast cells are required for the STAT5B/cJUN cooperation on the Star promoter. We next tested whether transient overexpression of STAT5B CA and cJUN in Leydig cells was sufficient to increase progesterone production by MA-10 Leydig cells. As shown in Figure 10A, overexpression of either STAT5B CA or cJUN resulted in a small but statistically significant increase in progesterone production by MA-10 cells. A tendency to an increase progesterone production was observed when both factors were combined, but it did not reach statistical significance (Fig. 10A). Similarly, there was a tendency of increased STAR protein levels in MA-10 Leydig cells transiently overexpressing either transcription factor, but it did not reach statistical significance (Fig. 10B).

Figure 7.

STAT5B cooperates with cJUN but not GATA4 to activate the mouse Star promoter. (A) DNA sequence alignment of the −95 to −43 bp region of the Star promoter from different species highlighting the AP-1 and GATA elements (grey shaded box). The sequence of the consensus AP-1 and GATA elements is also shown. S = G or C, W = A or T, R = G or A. (B, C) MA-10 Leydig cells were cotransfected with an empty expression vector (CTL) or expression vectors for a constitutively active STAT5B (STAT5B CA), GATA4 (B) or cJUN (C) as indicated, along with a mouse Star −980/+16 bp reporter construct. The number of experiments is indicated. Results are shown as fold activation over control (± SE of the mean). Different letters (a, b, c, or d) indicate a statistically significant difference between groups (P < 0.05).

Figure 8.

Human growth hormone (hGH) increases cJun expression in Leydig cells. MA-10 Leydig cells treated with hGH for different durations as indicated. (A) Total RNA were prepared and cJun messenger RNA levels were quantified by quantitative polymerase chain reaction. Data were normalized with the Rpl19 housekeeping gene. The number of experiments is indicated. An asterisk indicates a statistically significant difference (P < 0.05) from the control (time 0). (B) Nuclear extracts were prepared for immunodetection of cJUN. GAPDH was used as loading control.

Figure 9.

The STAT5B/cJUN cooperation on the Star promoter occurs in Leydig but not in Sertoli or fibroblast cells. MA-10 (top left panel, black bars) and MLTC-1 (top right panel, grey bars), MSC-1 Sertoli cells (lower left panel, hatched bars) and CV-1 fibroblast cells (lower right panel, open bars) were cotransfected with an empty expression vector (CTL) or expression vectors for a constitutively active STAT5B (STAT5B CA), GATA4, or cJUN individually or in combination as indicated, along with a mouse Star −980/+16 bp reporter construct. The positions of the 5 potential γ-interferon-activated sequence elements (diamonds), AP-1 element (hatched oval), and GATA element (grey rectangle) are indicated. The number of experiments is indicated. Results are shown as fold activation over control (± SE of the mean). For a given cell line, different letters (a, b, c, or d) indicate a statistically significant difference between groups (P < 0.05).

Figure 10.

Overexpression of STAT5B and cJUN in MA-10 Leydig cells increases steroid hormone production. (A) MA-10 Leydig cells were transfected with an empty expression vector (CTL) or expression vectors for a constitutively active STAT5B (STAT5B CA) or for cJUN individually or in combination as indicated. The next day, cells were rinsed and grown in media without serum. After 24 hours, the media was harvested, and progesterone was quantified by enzyme-linked immunosorbent assay. The number of experiments is indicated. Different letters (a or b) indicate a statistically significant difference between groups (P < 0.05). (B) MA-10 Leydig cells were transfected as previously described, and total protein extracts were prepared for immunodetection of STAT5B, cJUN, and STAR. GAPDH was used as loading control. The experiment was repeated 4 times and the results from each experiment were quantified and are plotted as mean ±SE of the mean.

To determine the site requirement(s) for the STAT5B/cJUN cooperation, MA-10 Leydig cells were transfected with different −980 bp Star reporter constructs, either wild-type or harboring mutations in the GAS element at −756 bp, the AP-1 element at −78 bp, or in both elements. Mutation of the GAS or the AP-1 element individually had no effect and the STAT5B/cJUN cooperation was similar to that observed with the wild-type Star reporter (Fig. 11). However, simultaneous mutation of the GAS and AP-1 elements abrogated the STAT5B/cJUN cooperation (Fig. 11). Taken together, these results indicate that either the GAS element at −756 bp or the AP-1 element at −78 bp is sufficient for STAT5B/cJUN cooperation on the Star promoter.

Figure 11.

The STAT5B/cJUN cooperation requires either the γ-interferon-activated sequence (GAS) element at −756 bp or the AP-1 element at −78 bp. MA-10 Leydig cells were cotransfected with an empty expression vector (control) or expression vectors for constitutively active STAT5B (STAT5B CA) and cJUN as indicated, along with −980/+16 bp Star reporter constructs either wild type, harboring a mutation in the GAS element at −756 bp (TTCCCAGAA to aagCCAGAA), containing a mutation in the AP-1 element at −78 bp (TGACTGATG to TGAgaGtTG), or with both elements mutated (−756 bp and −78 bp). The mutated elements are depicted by large X. The number of experiments is indicated. Results are shown as fold activation over control (± SE of the mean). For a given reporter, different letters (a, b, c, or d) indicate a statistically significant difference between groups (P < 0.05).

Although STAT5B binds directly to the GAS element at −756 bp (Fig. 4B and 4C), we wanted to determine whether STAT5B is also recruited to the proximal region of the Star promoter in a native chromatin environment in MA-10 Leydig cells by performing ChIP-qPCR. As shown in Figure 12, STAT5B was found to be associated with the proximal Star promoter, between −299 and −41 bp, a region that contains the AP-1 element at −78 bp. Furthermore, recruitment of STAT5B to the proximal Star promoter was dramatically increased in MA-10 Leydig cells treated with GH (Fig. 12), a condition known to induce and activate STAT5B in these cells (Figs. 1C and 4B and 4C) (22). No recruitment was observed when an IgG was used instead of an anti-STAT5B antiserum (Fig. 12). These data confirm that STAT5B is recruited in a GH-stimulated manner to the proximal Star promoter in Leydig cells.

Figure 12.

STAT5B is recruited to the mouse Star proximal promoter region in Leydig cells. MA-10 Leydig cells were treated with either vehicle (CTL, open bars) or 100 ng/mL human growth hormone for 2 hours (black bars) after which chromatin immunoprecipitation (ChIP) assays were performed using an immunoglobulin G (IgG; negative control) or an anti-STAT5B antiserum (αSTAT5B). A 258 bp fragment of the proximal Star promoter (−299/−41 bp) containing an imperfect γ-interferon-activated sequence element at −236 bp (black diamond) and the previously characterized AP-1 element at −78 bp (hatched oval) was amplified by quantitative polymerase chain reaction immediately following the ChIP assay. The amplified region of the Star gene is shown. Results are represented as fold enrichment of ChIP DNA αSTAT5B over ChIP DNA IgG from 3 independent experiments (± SE of the mean). Different letters (a, b, or c) indicates a statistically significant difference between groups (P < 0.05).

Discussion

GH is an important regulator of Leydig cell steroidogenesis [reviewed in (2,3)] but our understanding of the molecular mechanism of GH action in this process remained incomplete. Here we show that in Leydig cell lines, treatment with GH activates the STAT5B transcription factor, which then stimulates the activity of the Star promoter alone and in cooperation with the transcription factor cJUN, leading to increased steroid hormone production (Fig. 13).

Figure 13.

Proposed model for growth hormone (GH)-induced STAT5B action in Leydig cell steroidogenesis. GH binds to its receptor (GHR) present on the surface of Leydig cells. This activates a signaling cascade involving JAK kinases ultimately leading to phosphorylation and dimerization of STAT5B. STAT5B activates the expression of several genes including the gene encoding the NUR77/NR4A1 nuclear receptor (arrow 1), a well-known regulator of several steroidogenic genes, including Star (25,26,31,32,44,45,53,58-64). STAT5B also contributes to GH-induced steroidogenesis by activating the promoter of the Star gene (arrow 2). Action of STAT5B on Star promoter activity involves direct binding to a γ-interferon-activated sequence element and cooperation/interaction with DNA-bound cJUN.

STAT5B Activates Star Transcription in Leydig Cells

Because of its vital role in steroid hormone production, expression of the Star gene has been the focus of intense studies, which led to the identification of several hormones, signaling pathways, and transcription factors involved in its regulation [reviewed in (47, 48)]. Although it is well established that treatment of Leydig cells with GH increases testosterone secretion (15,16) and Star mRNA and STAR protein levels (15), the transcription factor(s) downstream of the GH-induced pathway responsible for increased Star transcription remained uncharacterized. In other tissues, the STAT5B transcription factor is a classical mediator of GH action [reviewed in (20)]. The testis is no exception as we detected STAT5B in Leydig cells of the adult mouse testis as well as in the MA-10 Leydig cell line where its expression is induced in response to GH. Our data are in agreement with previous findings from Kanzaki and Morris who reported the presence of STAT5B in MA-10 Leydig cells and in primary rat Leydig cells and showed that STAT5B becomes phosphorylated and translocates to the nucleus upon GH treatment (22). In addition, our present data revealed that STAT5B is recruited to the Star promoter in Leydig cells in a GH-dependent manner in a native chromatin environment, further supporting a role for STAT5B in Star gene expression.

In our present work, overexpression of a constitutively active form of STAT5B in MA-10 Leydig cells is sufficient to increase steroid hormone production and endogenous Star mRNA levels, consistent with a role for STAT5B in Star gene transcription. Indeed, we found that STAT5B activates the mouse Star promoter in Leydig cells. The mouse Star promoter contains 5 potential GAS elements for the binding of STAT5B. However, only 1 located at −756 bp (TTCCCAGAA) perfectly matches the consensus sequence TTCnnnGAA [reviewed in (23)]. STAT5B was found to directly bind to the element at −756 bp and mutation of the sequence from TTCCCAGAA to aagCCAGAA abolished STAT5B binding (Fig. 13, arrow 2). Surprisingly, this mutation did not abrogate STAT5B-dependent activation of the Star promoter, which indicates alternate mechanisms of STAT5B action in Star transcription. There are several possibilities that are not mutually exclusive that can be put forward to explain the mechanism of STAT5B action in Star transcription. For instance, STAT5B could function via 1 of the imperfect GAS elements since some flexibility in STAT5 binding specificity has been described (49). However, functional analysis of the Star promoter performed in this work revealed that a deletion to −95 bp, which removes all the potential GAS elements, was still activated by STAT5B. These data suggest that the GAS elements are dispensable and that other mechanisms are involved. Indeed, we found that STAT5 can functionally cooperate with DNA-bound cJUN even in the absence of a functional GAS element. However, we cannot exclude the possibility that the various GAS elements (perfect and imperfect) still contribute to the STAT5-dependent activation of the Star promoter.

Another potential mechanism to explain STAT5B action on the Star promoter is that STAT5B acts indirectly. In this scenario, STAT5B could act by upregulating the expression of another transcription factor known to activate the Star promoter. For instance, the −95 bp deletion construct retains the binding site for the nuclear receptor NUR77 (NR4A1, NGFI-B) at −91 bp. NUR77 is a well-characterized immediate response factor known to be induced by various stimuli [reviewed in (50, 51)]. In addition, NUR77 is an important regulator of Star gene expression (24-26,31,44,45,52-54). It is therefore possible that STAT5B could induce transcription of the Nur77 gene, and NUR77 would in turn activate the Star promoter. In support of this, we found that the Nur77 promoter is activated by STAT5B in Leydig cells (Fig. 13, arrow 1).

STAT5B Functionally Cooperates With cJUN on the Star Promoter

Another possibility to explain the STAT5B-dependent activation of the Star promoter in the absence of any GAS element involves the recruitment of STAT5B to the Star promoter via another DNA-bound transcription factor. The −95 to −43 bp region of the Star promoter contains several binding sites for various transcription factors [reviewed in (55)], including GATA4 and cJUN, 2 key activator of Star gene expression (24,27,46). Interestingly, both GATA4 and cJUN were found to physically interact and functionally cooperate with STAT family members on the Angiotensin II promoter in vascular smooth muscle cells. GATA4 was found to interact with STAT1 leading to a synergistic activation of the Angiotensin II promoter (56) while cJUN interacts and cooperates with STAT5B on the same promoter (57). In MA-10 Leydig cells, we found that STAT5B did not cooperate with GATA4 on the Star promoter. However, a functional cooperation was observed between STAT5B and cJUN. This STAT5B/cJUN cooperation was abrogated only when the binding sites for both factors (GAS at −756 bp and AP-1 at −78 bp) were mutated as mutation of a single element (GAS or AP-1) did not affect the cooperation. This indicates that the 2 transcription factors interact in a common regulatory complex (Fig. 13, arrow 2) as previously demonstrated by Han et al using co-immunoprecipitation assay (57). In Leydig cells, cJUN has been shown to regulate expression of the Star gene on its own and in cooperation with other transcription factors including SF1/NR4A1 and GATA4 (24,26,46). Consistent with a role in GH-mediated signaling, we also found that cJUN expression is induced by GH in MA-10 Leydig cells.

It is interesting to note that the STAT5B/cJUN cooperation on the Star promoter described in this work was specific to Leydig cells. When assayed in the MSC-1 Sertoli cell line, another somatic cell lineage of the testis that shares several common transcription factors with Leydig cells, or in the CV-1 fibroblast cell line, STAT5B did not cooperate with cJUN on the Star promoter. These data indicate that the STAT5B/cJUN cooperation requires additional Leydig cell–enriched transcription factor(s) that have yet to be identified, which warrants further investigations.

Funding

This work was funded by a grant from the Canadian Institutes of Health Research (funding reference number MOP-81387) to J.J.T. S.M. and K.D.M. were the recipient of a studentship from the Fondation du CHU de Québec-Université Laval. K.D.M. is the recipient of a studentship from the Fonds de recherche du Québec-Santé.

Author Contributions

P-O.H-M. performed the majority of the experiments along with F.B., N.M.R., S.M., K.J.P., R.E.M-V., K.D.M., and C.B. J.J.T. conceived the original idea and supervised the project. J.J.T. wrote the manuscript. All authors provided critical feedback and helped shape the research, analysis, and manuscript.

Disclosures

The authors have nothing to disclose.

Data Availability

All data generated or analyzed during this study are included in this published article or in data repositories.