Abstract
Pituitary follicle-stimulating hormone (FSH) synthesis is regulated by transforming growth factorβsuperfamily ligands, most notably the activins and inhibins. Bone morphogenetic proteins (BMPs) also regulate FSHβ subunit (Fshb) expression in immortalized murine gonadotrope-like LβT2 cells and in primary murine or ovine primary pituitary cultures. BMP2 signals preferentially via the BMP type I receptor, BMPR1A, to stimulate murine Fshb transcription in vitro. Here, we used a Cre–lox approach to assess BMPR1A’s role in FSH synthesis in mice in vivo. Gonadotrope-specific Bmpr1a knockout animals developed normally and had reproductive organ weights comparable with those of controls. Knockouts were fertile, with normal serum gonadotropins and pituitary gonadotropin subunit mRNA expression. Cre-mediated recombination of the floxed Bmpr1a allele was efficient and specific, as indicated by PCR analysis of diverse tissues and isolated gonadotrope cells. Furthermore, BMP2 stimulation of inhibitor of DNA binding 3 expression was impaired in gonadotropes isolated from Bmpr1a knockout mice, confirming the loss of functional receptor protein in these cells. Treatment of purified gonadotropes with small-molecule inhibitors of BMPR1A (and the related receptors BMPR1B and ACVR1) suppressed Fshb mRNA expression, suggesting that an autocrine BMP-like molecule might regulate FSH synthesis. However, deletion of Bmpr1a and Acvr1 in cultured pituitary cells did not alter Fshb expression, indicating that the inhibitors had off-target effects. In sum, BMPs or related ligands acting via BMPR1A or ACVR1 are unlikely to play direct physiological roles in FSH synthesis by murine gonadotrope cells.
Introduction
The pituitary glycoprotein follicle-stimulating hormone (FSH) is an essential regulator of ovarian function and testicular development in mammals. FSH synthesis by pituitary gonadotrope cells is primarily stimulated by gonadotropin-releasing hormone (GnRH) from the hypothalamus and intrapituitary autocrine/paracrine factors, the best studied of which are the activins (Bernard et al. 2010). Activins are the members of the transforming growth factor β(TGFβ) superfamily and signal via complexes of serine/threonine kinase receptors, homolog of Drosophila mothers against decapentaplegic (SMAD) proteins, and forkhead box L2 to regulate transcription of the FSHβ subunit gene (Fshb) (Weiss et al. 1995, Suszko et al. 2003, Bailey et al. 2004, Bernard 2004, Gregory et al. 2005, Safwat et al. 2005, Suszko et al. 2005, Bernard et al. 2006, Lamba et al. 2006, 2009, 2010, Tran et al. 2011, Wang & Bernard 2012, Bernard & Tran 2013). In response to FSH signaling, the gonads synthesize and secrete inhibins, which act in an endocrine fashion to suppress FSH synthesis by competitively binding to activin receptors (Lewis et al. 2000, Chapman et al. 2002).
Additional members of the TGFβ superfamily, the bone morphogenetic proteins (BMPs), have also been implicated in regulating Fshb expression both alone and in combination with other factors. For example, BMP6 and BMP7 (at 1µg/mL) stimulate ovine Fshb promoter-reporter activity in pituitaries of transgenic mice or in transiently transfected murine gonadotrope-like LβT2 cells (Huang et al. 2001). By contrast, antisera against BMP7 suppress FSH release from murine, rat, and ovine primary pituitary cultures, suggesting a role for endogenous intrapituitary BMP7 in FSH regulation. BMP15 similarly stimulates ovine Fshb promoter activity in LβT2 cells and FSH secretion from rat primary pituitary cell cultures, but at concentrations as low as 10–100ng/mL (Otsuka & Shimasaki 2002). BMP2 and BMP4 are ten-fold more potent in stimulating murine Fshb promoter-reporter activity than BMP6 or BMP7 in LβT2 cells (Lee et al. 2007). Moreover, BMP2 and activins synergistically stimulate murine, porcine, and ovine Fshb promoter-reporter activities as well as endogenous murine Fshb mRNA expression in this cell line. Similarly, BMP4 potentiates activin A and GnRH induction of Fshb mRNA expression and FSH release in LβT2 cells (Nicol et al. 2008), whereas BMP6 synergistically stimulates murine Fshb promoter-reporter activity with GnRH (Takeda et al. 2012). In contrast to these results, exogenous BMP4 and BMP6 suppress FSH secretion and Fshb mRNA expression in primary pituitary cultures from ewes and also counteract the stimulatory effects of activin A in this system (Faure et al. 2005, Young et al. 2008). Collectively, these data suggest that BMPs can regulate FSH at the level of the pituitary gland, though their effects may be ligand-, context-, and species-specific.
In LβT2 cells, BMP2 signals via the type I receptor, BMPR1A (also known as activin receptor-like kinase 3 or ALK3) to regulate murine Fshb transcription (Ho & Bernard 2009). Though a second type I receptor, BMPR1B (ALK6), is also present in these cells (Lee et al. 2007, Nicol et al. 2008), it is expressed at low levels and BMP2 signaling is intact when this receptor is knocked down using short interfering RNAs (Ho & Bernard 2009). Similarly, in sheep, BMPR1A, but not BMPR1B, is expressed in gonadotrope cells (Faure et al. 2005), suggesting that BMP2 and BMP4 likely signal via BMPR1A in these animals as well. Here, we tested the hypothesis that signaling via BMPR1A is required for FSH synthesis in vivo by selectively ablating the receptor in murine gonadotropes using a Cre-lox approach.
Materials and methods
Reagents
Human recombinant (rh-) BMP2 (355-BM) and activin A (338-AC) were from R&D Systems. RQ1 RNase-Free DNase (M6101), random primers (C1181), MMLV-reverse transcriptase (M1701), and RNasin (N2511) were from Promega. SB431542 (S4317), pancreatin (P3292), and collagenase (Type I-bib130) were from Sigma. Media 199 (M199; 31100-035), Hanks’ Balanced Salt Solution (HBSS) without calcium/magnesium (14170-112), TRIzol Reagent, and SYBRgreen Supermix for quantitative PCR (qPCR) were from Invitrogen. EvaGreen 2X qPCR MasterMix-S was from Applied Biological Materials, Inc. (Richmond, BC, Canada). Oligonucleotides were purchased from Integrated DNA Technologies, Inc. (Coralville, IA, USA). Gentamycin (450-135-XL), 100X antibiotic–antimycotic (450-115-EL), and deoxynucleotide triphosphates were from Wisent (St-Bruno, QC, Canada). LDN-193189 hydrochloride (1509) was purchased from Axon Medchem (Reston, VA, USA), and compound C (171260) was from Calbiochem (EMD Chemicals, Inc., Darmstadt, Germany).
Animals
The mouse strains used here have been described previously: Bmpr1a +/−(Mishina et al. 1995), Bmpr1a fl/fl (Mishina et al. 2002), Acvr1 fl/fl (Dudas et al. 2004), and GnRH-receptor-IRES-Cre (GRIC) (Wen et al. 2008). In the latter model, Cre recombinase is expressed as part of a bicistronic mRNA with the endogenous Gnrhr mRNA. Bmpr1a +/ − males were crossed with Gnrhr GRIC/GRIC females. Resulting Bmpr1a +/ −;Gnrhr GRIC/+ females were then crossed with Bmpr1a fl/fl males to generate control (Bmpr1a fl/+;Gnrhr GRIC/+) and experimental animals (Bmpr1a fl/ −;Gnrhr GRIC/+). The GRIC allele is active in the male germ line (Wen et al. 2010); therefore, to avoid global recombination of the floxed allele, GRIC was always introduced from the female parent. The Bmpr1a +/ −;Gnrhr GRIC/+×Bmpr1a fl/fl cross generated animals of four genotypes at the expected frequencies (1:1:1:1). For genetic labeling of gonadotropes, Gnrhr GRIC/GRIC mice were crossed with Rosa26-loxSTOPlox-EYFP (hereafter R26-YFP) reporter mice (Srinivas et al. 2001) acquired from Jackson Laboratories. To label gonadotropes in control and experimental mice, we crossed Bmpr1a +/ −; Gnrhr GRIC/GRIC females with Bmpr1a fl/fl;R26-YFP/R26-YFP males. Acvr1 fl/fl and Bmpr1a fl/fl mice were crossed to generate Acvr1 fl/+;Bmpr1a fl/+ and eventually Acvr1 fl/fl;Bmpr1a fl/fl mice for in vitro recombination experiments. All animals were housed on a 12h light:12h darkness cycle and were given access to food and water ad libitum. All mouse work was conducted in accordance with federal and institutional guidelines and with the approval of the McGill Animal Care and Use Committee (animal use protocol #5204).
DNA extraction and genotyping
Genomic DNA was extracted from tail biopsies (~0.5cm) using 0.5mL of lysis buffer (100mmol/LTris–HCl (pH 8.5), 5mmol/L EDTA (pH 8.0), 200mmol/LNaCl, 0.2% (v/v) SDS, and 100µg/mL proteinase K). Tails were incubated at 55°C overnight in a water bath. Samples were then vortexed and centrifuged at 14,600g for 10 min. The supernatant was collected and mixed by inversion with 0.5mL isopropanol. Precipitated DNA was collected with a micropipette tip and dissolved in 40µL of 10mmol/L Tris (pH 8.0). For comparison of recombination across different tissues, DNA was extracted from approximately 5mg of the indicated tissues using the Gentra Puregene Blood Kit (Qiagen) following the manufacturer’s instructions. Wild-type, null, floxed, and recombined Bmpr1a alleles were detected by PCR using the protocols described in (Mishina et al. 1995, 2002). The GRIC allele was detected using the primer set indicated in Table 1.
Genotyping and quantitative RT-PCR primers.
Genotyping |
---|
Cre |
Fwd: GGACATGTTCAGGGATCGCCAGGC |
Rev: GCATAACCAGTGAAACAGCATTGCTG |
Acvr1 |
Fwd: CCCCCATTGAAGGTTTAGAGAGAC |
Rev: CTAAGAGCCATGACAGAGGTTG |
Bmpr1a |
Fwd1: CTCTGAATTTCTAGTCCACATCTGC |
Fwd2: AGACTGCCTTGGGAAAAGCGC |
Rev: GGACTATGGACACACAATGGC |
Quantitative RT-PCR |
Acvr1 |
Fwd: GCTGCATAGCAGATTTGGGC |
Rev: CACTTCCGGAGCCATGTAGC |
Bmpr1a |
Fwd: CATCAGATTACTGGGAGCCTG |
Rev: GGTATCCTCTGGTGCTAAAGTC |
Bmpr1b |
Fwd: GTCTCAGAGCTCGGGAAGTG |
Rev: ATAGCGGCCTTTTCCAATCT |
Fshb |
Fwd: GTGCGGGCTACTGCTACACT |
Rev: CAGGCAATCTTACGGTCTCG |
Id1 |
Fwd: GGTACTTGGTCTGTCGGAGC |
Rev: GCAGGTCCCTGATGTAGTCG |
Id3 |
Fwd: TTAGCCAGGTGGAAATCCTG |
Rev: TCAGTGGCAAAAGCTCCTCT |
Lhb |
Fwd: ACTGTGCCGGCCTGTCAACG |
Rev: AGCAGCCGGCAGTACTCGGA |
Rpl19 |
Fwd: CGGGAATCCAAGAAGATTGA |
Rev: TTCAGCTTGTGGATGTGCTC |
Hormone assays
Serum luteinizing hormone (LH) and FSH were measured by multiplex ELISA at the Ligand Assay and Analysis Core (LAAC) of the Center for Research in Reproduction at the University of Virginia (Charlottesville, VA, USA). Both hormones were measured in singlet from 10µL serum. The reportable ranges for LH and FSH were 0.24–30.0 and 2.40–300.0ng/mL, respectively.
Fluorescence-activated cell sorting of genetically labeled gonadotropes
Pituitaries from mice with yellow fluorescent protein (YFP)-labeled gonadotropes were enzymatically dispersed as described in (Ho et al. 2011) and single-cell suspensions were prepared in PBS. Cells were then passed through a 70µm nozzle at 70 psi into a Becton Dickinson FACSAria Sorter in the McGill University Flow Cytometry Core Facility. Sorting was performed using FACSDiva software (v. 6.0). Gating was established on a forward and side scatter plot (for relative cell size and granularity, respectively) to exclude debris and cell clusters. A control without YFP-labeled cells was run to establish a negative baseline profile using a single-parameter histogram. The pituitary cells were then run and gated to sort YFP+ and YFP− cells. YFP was excited with a 488nm argon ion laser and detected with a 530/30 band-pass filter and an LP 502 dichroic mirror. Depending on the preparation, we obtained roughly 0.8–15×103 YFP+ and 4.6–18×104 YFP− cells per pituitary.
Gonadotrope cell culture
Purified gonadotropes were plated at a density of 4−5×104cells per well in 96-well plates in Medium 199 with 10% fetal bovine serum (FBS), 1X antibiotic–antimycotic, and 120µg/mL gentamicin. Cells were incubated at 37°C/5% CO2 for at least 16h before treatment. Cells were treated in medium M199 plus 2% (v/v) FBS overnight with the indicated ligands and/or inhibitors. RNA and DNA were then collected with AllPrep DNA/RNA mini kits (cat. 80204; Qiagen) following the manufacturer’s instructions. DNA was analyzed as described previously. RNA was analyzed by quantitative RT-PCR (discussed later).
Primary pituitary cultures
Primary cultures were performed as described previously (Fortin et al. 2013). Briefly, pituitaries were collected from 10-week-old Bmpr1a f l/f l;Acvr1 f l/f l male and female mice in M199 medium supplemented with 10% (v/v) FBS. Pituitaries were washed 3 times in HBSS with 150µmol/L of CaCl2, cut several times with a scalpel, and digested in collagenase (1.5mg/mL) (Sigma #C-0130; diluted in HBSS with 30mg/mL BSA, pH 7.4, 40µL/pituitary) at 37°C for 2h. The suspension was then washed with 5mL calcium-free HBSS, centrifuged for 5min at 1000 g , and resuspended in pancreatin solution (Sigma P3292; 4.5mg/mL in calcium-free HBSS; 40µL/pituitary). Pancreatin digestion was performed in a 37°C water bath with manual agitation for 8–10min. Finally, the cell suspension was washed 3times in 5mL M199 media containing 10% FBS and cells were seeded at density of 5×105/well in 48-well plates.
Adenoviral transduction and treatment of primary pituitary cultures
Primary cultures were transduced with adenoviruses as described previously (Fortin et al. 2014). Pituitary cultures were prepared as described previously. After 24h, viral transductions were performed using adenoviruses that express enhanced green fluorescent protein (eGFP) or Cre-IRES-eGFP (Vector Development Laboratory, Baylor College of Medicine, Houston, TX, USA) at a multiplicity of infection of 60 in M199 medium containing 10% (v/v) FBS. The following day, virus-containing medium was removed and replaced with medium containing 2% (v/v) FBS with 25ng/mL activin A, 25ng/mL BMP2, 1µmol/L SB431542 or 1µmol/L LDN-193189, or DMSO as vehicle. All treatments were performed for 24h in duplicate, and the experiment was repeated 5–7 times. Cells were harvested, and RNA and DNA were extracted using the Allprep DNA/RNA kit (Qiagen). RNA was eluted in 23µL RNase-free water and reverse transcribed. The resulting cDNA was analyzed by qPCR.
Quantitative RT-PCR
Fshb and Lhb mRNA levels (normalized to Rpl19) in whole pituitary RNA were measured using the relative standard curve method with LβT2 cell RNA as standard, as described previously (Lamba et al. 2009). Fshb, Lhb, Id3, Bmpr1a, Bmpr1b, Acvr1a, and Id mRNA levels (normalized to Rpl19) in isolated gonadotropes or mixed pituitary cultures were measured using the 2− ΔΔ/fl Ct method as described in (Ho & Bernard 2010). Primer sequences are indicated in Table 1.
Statistics
Data from control and experimental mice and/or cell cultures were compared with t-tests or analyses of variance as indicated using Systat (Systat Software Inc, San Jose, CA, USA) . Post hoc pair-wise comparisons were made with Bonferroni corrections. Data were log transformed when variances were unequal. Significance was assessed relative to P<0.05.
Results
Gonadotrope-specific Bmpr1a knockout mice
We generated gonadotrope-specific Bmpr1a knockout mice by crossing ‘floxed’ Bmpr1a animals with mice expressing Cre recombinase from the endogenous Gnrhr locus (so-called GRIC mice, see the ‘Materials and methods’ section). Because Bmpr1a +/ − mice are viable and fertile, we elected to cross in a single null allele such that only one floxed Bmpr1a allele required recombination in our model. As a result, the experimental mice (hereafter conditional knockouts or cKOs) had the following genotype: Bmpr1a f l/ −;Gnrhr GRIC/+. Control littermates from the Bmpr1a +/ −;Gnrhr GRIC/+×Bmpr1a f l/f l cross used to generate cKO mice had the following genotype: Bmpr1a fl /+;Gnrhr GRIC/+. These animals allowed us to control for potential effects of the GRIC allele. Recombination of a single floxed allele in gonadotropes of these animals was not a concern because, as mentioned previously, Bmpr1a +/ − mice are fertile. Mice of the two other genotypes derived from this cross (Bmpr1a fl/+;Gnrhr +/+ or Bmpr1a fl/ −;Gnrhr +/+) were not analyzed. The GRIC allele is active in pituitary gonadotropes and in male germ cells. Therefore, as anticipated, we observed the recombination of the floxed Bmpr1a allele in pituitaries, testes, and epididymides, but not in other tissues, including ovaries and uteri of adult cKOs (Fig. 1A).
(A) PCR analysis of genomic DNA from the indicated tissues in cKO male and female mice. The presence of a band indicates recombination of the ‘floxed’ Bmpr1a allele. (B) PCR analysis of genomic DNA from purified gonadotrope (YFP+, right) and non-gonadotrope (YFP−, left) cells from pituitaries of adult male control and cKO mice. Top: Primer set that detects the floxed (top band) or wild-type/null (bottom band) Bmpr1a alleles. Bottom: Primer set that detects the recombined Bmpr1a allele. (C) Gonadotropes (YFP+) purified from control and cKO mice were plated in primary culture and then treated with vehicle or 25ng/mL BMP2 for 24h. Id3 mRNA expression was measured by quantitative RT-PCR. Data are means (+s.e.m.) of five independent experiments. Data from male and female cultures did not differ and were combined for analysis. ** indicates different from all other groups as determined by two-way ANOVA of log transformed data (P<0.01).
Citation: Journal of Endocrinology 229, 3; 10.1530/JOE-16-0053
To further demonstrate the efficacy of recombination, we genetically labeled and purified gonadotropes from control and cKO mice. Here, an enhanced YFP (eYFP) reporter allele, which is activated by Cre recombinase-mediated removal of a transcriptional stop cassette, was introduced into the genetic background of the control and cKO strains. Their pituitaries were extracted and enzymatically dissociated. eYFP-labeled (YFP+) cells were then separated from eYFP-negative (YFP–) cells by FACS. DNA was extracted from the purified cell populations and analyzed by PCR. The data confirm the efficient recombination of the floxed Bmpr1a allele in YFP+, but not YFP− cells in males of both genotypes (Fig. 1B; recall that both control and cKO mice harbor a single floxed Bmpr1a allele; therefore, recombination is predicted in both genotypes). Similar results were obtained in females (data not shown).
We previously showed that BMP2 signals via BMPR1A to regulate Id3 expression in LβT2 cells and in primary pituitary cell cultures (Ho et al. 2011). Therefore, to confirm functional deletion of BMPR1A, we treated purified gonadotropes from control and cKO mice in primary culture with 25ng/mL BMP2 for 24h. BMP2-stimulated Id3 mRNA expression was significantly reduced in cKOs relative to control gonadotropes (Fig. 1C).
Fertility is normal in Bmpr1a cKO mice
Eight- to ten-week-old control and cKO males and females were paired with wild-type C57BL/6 (Charles River) opposite sex partners for a period of 6 months to assess their fertility. As shown in Fig. 2, there were no significant differences between genotypes in the number of days from pairing to delivery of the first litter, the average numbers of litters per animal, the average number of pups per litter, or the inter-litter interval.
Female and male control and cKO mice were paired with wild-type C57BL/6 mice for a period of 6 months (N=6–8 per sex per genotype). (A) Number of days since pairing until delivery of the first litter, (B) average number of litters per pair, (C) average number of pups per litter, and (D) average number of days between litters. Data are means (+s.d.) (Data were analyzed with unpaired t-tests; NS, not significantly different (P>0.05)).
Citation: Journal of Endocrinology 229, 3; 10.1530/JOE-16-0053
Reproductive tissues and hormone levels are normal in Bmpr1a cKO mice
We next measured reproductive organ weights (Table 2), serum gonadotropin levels (Fig. 3A and B), and pituitary Lhb and Fshb mRNA levels (Fig. 3C and D) in postpubertal (6-week-old) control and cKO mice of both sexes. Again, we did not observe any differences between genotypes in any of these parameters, with the exception of testicular weight, which was slightly elevated in cKO mice.
Serum and pituitaries were collected from control and cKO mice at 6weeks of age. (A) LH and (B) FSH were measured by ELISA. RNA was extracted from whole pituitaries. (C) Lhb and D) Fshb mRNA levels were assessed by quantitative RT-PCR. Hormone data reflect means (+s.e.m.) of N=8 per sex per genotype. RNA data are means (+s.e.m.) of N=5 per sex per genotype. Data were analyzed with unpaired t-tests; NS, not significantly different (P>0.05).
Citation: Journal of Endocrinology 229, 3; 10.1530/JOE-16-0053
Body and reproductive organ weights.
Sex | Genotype | N | Body weight (g) | Ovarian weight (g)* | Uterine weight (g) | |||
---|---|---|---|---|---|---|---|---|
F | Control | 10 | 17.65±0.26 | P>0.1 | 0.0084±0.001 | P>0.1 | 0.110±0.013 | P>0.1 |
F | cKO | 10 | 18.27±0.45 | 0.011±0.001 | 0.092±0.013 | |||
Sex | Genotype | N | Body weight (g) | Testicular weight (g)* | ||||
M | Control | 11 | 21.36±0.7 | P>0.1 | 0.168±0.006 | P<0.01 | ||
M | cKO | 14 | 22.68±0.8 | 0.206±0.008 |
Data are means ± s.e.m.
*Paired organ weights. Data were analyzed with unpaired t-tests. P values are presented for each comparison.
F, female; M, male.
BMP2 does not stimulate Fshb mRNA expression in purified gonadotropes
Previous analyses of BMP2-regulated Fshb expression were conducted almost entirely in the LβT2 cell model (Lee et al. 2007, Ho & Bernard 2009, 2010, Ho et al. 2011). We therefore asked whether BMP2 similarly regulates Fshb expression in isolated murine gonadotropes of wild-type mice. Whereas activin A robustly stimulated Fshb mRNA in these cells, BMP2 had no effect (Fig. 4A). By contrast, BMP2 stimulated and activin A inhibited Id3 mRNA expression in these cultures (Fig. 4B). Thus, BMP2 was active in primary gonadotropes but did not regulate the Fshb gene therein.
Gonadotropes were purified from wild-type (GRIC-YFP) mice and plated in primary culture. Cells were treated with 25ng/mL activin A or BMP2 overnight. RNA was extracted and (A) Fshb or (B) Id3 mRNA levels were measured by quantitative RT-PCR. The data reflect the means (+s.e.m.) of five (Id3) or six (Fshb) independent experiments (two or three cultures were from males, two from females, and one from mixed sex cultures). ***indicates different from control as determined by one-way ANOVA of log transformed data (P<0.001). (C) Cells were treated with 1µmol/L SB431542 or LDN-193189 (or DMSO vehicle) overnight. RNA was extracted and Fshb mRNA levels were measured by quantitative RT-PCR. Data are the means (+s.e.m.) of three (SB) or five (LDN) independent experiments (three cultures were from females and two from males). ***indicates different from control as determined by one-way ANOVA (P<0.001).
Citation: Journal of Endocrinology 229, 3; 10.1530/JOE-16-0053
Inhibitors of BMP type I receptors suppress Fshb expression in primary gonadotropes
Activin B is generally regarded as the primary autocrine/paracrine stimulator of Fshb transcription (Corrigan et al. 1991, Sallon et al. 2010). Consistent with this idea, treatment of isolated gonadotropes from wild-type mice with a small-molecule inhibitor (SB431542) (Inman et al. 2002) of the known activin B type I receptors (ACVR1B and ACVR1C) (Tsuchida et al. 2004, Bernard et al. 2006) dramatically reduced Fshb expression (Fig. 4C). To determine whether autocrine/paracrine BMP-like molecules (other than BMP2) might also play a role in Fshb expression, we also treated gonadotropes with LDN-193189 (Cuny et al. 2008). This molecule blocks the activities of ACVR1, BMPR1A, and BMPR1B, the three best-characterized BMP type I receptors (for convenience, these receptors will be referred to as ALK2/3/6). LDN-193189 also suppressed Fshb mRNA levels (Fig. 4C). We obtained similar results with a second ALK2/3/6 inhibitor, compound C (also known as dorsomorphin; Yu et al. 2008) (data not shown).
Fshb production by primary pituitary cultures is ACVR1/BMPR1A-independent
Though these data suggested a role for endogenous BMP-like molecules (other than BMP2) in FSH synthesis, LDN-193189 and compound C are known to antagonize other kinases in cells (Vogt et al. 2011). To assess potential ‘off-target’ effects, we examined the inhibitor activity in pituitary cells expressing or lacking BMPR1A and ACVR1. It should be noted that BMPR1B/ALK6 is expressed at very low levels (if at all; Yi et al. 2001) in the pituitary and in these cultures. Though we did not perform absolute quantification, the Ct values for Bmpr1b were 3–5 cycles higher than for Acvr1 or Bmpr1a, a difference of 8- to 32-fold. Similarly, a recent RNA-seq analysis indicates that Bmpr1a and Acvr1 mRNA levels are ~32- and ~16-fold greater than Bmpr1b in purified murine gonadotropes (Qiao et al. 2016). We prepared pituitary cultures from mice harboring floxed alleles for both BMPR1A (ALK3) and ACVR1 (ALK2) (Bmpr1a f l/f l; Acvr1 f l/f l). In these experiments, the cultures were prepared from whole pituitaries rather than from purified gonadotropes (it was not feasible to perform these experiments in purified gonadotropes). Half of the cells were infected with a control adenovirus expressing GFP. The other half was infected with a Cre-expressing adenovirus to recombine the floxed alleles. As shown in Fig. 5A and B, Acvr1 and Bmpr1a mRNA levels were significantly depleted in cells transduced with the Cre adenovirus. There was no compensatory increase in Bmpr1b mRNA expression (Fig. 5C). Stimulation of Id1 expression by BMP2 (Hollnagel et al. 1999) was significantly impaired in Cre adenovirus-treated cells, demonstrating the loss of functional type I receptor proteins (Fig. 5D). The loss of the receptors, however, did not alter (reduce) basal Fshb mRNA expression (Fig. 5E). Thus, endogenous TGFβ superfamily ligands regulate Fshb expression independently of these receptors. Importantly, both LDN-193189 and compound C impaired Fshb mRNA expression in mixed pituitary cultures (though to a lesser extent than in purified gonadotropes; compare Fig. 5E with Fig. 4C) whether the Bmpr1a and Acvr1 alleles were recombined or not. Therefore, these inhibitors appear to suppress Fshb expression in a BMPR1A/ACVR1-independent fashion. Again, activin A, but not BMP2, stimulated Fshb mRNA levels in these cultures (Fig. 5E). None of the treatments affected Lhb expression (Fig. 5F).
Primary pituitary cultures were prepared from Bmpr1a fl/fl;Acvr1 fl/fl mice. Cells were then transduced with GFP or Cre-encoding adenoviruses before being treated 24h later with 25ng/mL activin A, 25ng/mL BMP2, 1µmol/L SB431542, or 1µmol/L LDN-193189 in 2% (v/v) serum-containing medium. Twenty-four hours later, RNA was collected and Acvr1a (A), Bmpr1a (B), Bmpr1b (C), Id1 (D), Fshb (E), or Lhb (F) mRNA levels were measured by quantitative RT-PCR. The data presented reflect the means (+s.e.m.) of four to seven independent experiments. Cultures for females and males were run separately, but then combined for analyses when no sex differences were noted. All data were analyzed by two-way ANOVA (ligand × virus). Post hoc pair-wise comparisons (Bonferroni) were made of interactions (panels A, B, and D) or significant main effects (panel E). ***P<0.001 for GFP vs Cre (panels A and B), compared with all other treatments (panel D), or compared with no ligand treatment.
Citation: Journal of Endocrinology 229, 3; 10.1530/JOE-16-0053
Discussion
The data presented here indicate that BMPR1A is dispensable for normal gonadotropin synthesis and fertility in mice in vivo. Though these observations do not definitively rule out a role for BMP signaling in FSH regulation, they do suggest that any biologically relevant BMP-like ligands signal via an alternative type I receptor in the absence of BMPR1A (i.e., there is some form of compensation) and/or that these ligands actually prefer a receptor other than BMPR1A.
We focused on BMPR1A for in vivo analyses based on earlier results with BMP2 and BMP4 in LβT2 cells (Lee et al. 2007, Ho & Bernard 2009, 2010, Ho et al. 2011) and with BMP4 in ovine pituitary cultures (Faure et al. 2005, Young et al. 2008). Both of these ligands preferentially signal via BMPR1A and BMPR1B compared with the other type I receptors in the family (ten Dijke et al. 1994, Liu et al. 1995). BMPR1B is expressed at low levels, if at all, in gonadotropes (Lee et al. 2007, Nicol et al. 2008, Qiao et al. 2016) and BMP2 preferentially signals via BMPR1A in LβT2 cells to regulate both Fshb and Id3 transcription (Ho & Bernard 2009, Ho et al. 2011). The latter data were collected in the context of transient transfection experiments, which may have precluded compensatory mechanisms from developing. In the in vivo knockout model presented here, Cre expression occurs as early as embryonic day 12.75 (Wen et al. 2010), and we only examined mice in adulthood. As a result, it is possible that another receptor(s) could have compensated for the absence of BMPR1A during this protracted time frame. That said, Bmpr1b mRNA levels are low in isolated murine gonadotropes under normal conditions (Qiao et al. 2016) and do not increase in Bmpr1a cKO mice (data not shown).
BMP2 and BMP4 can also signal via the type I receptor ACVR1 (ALK2) (Liu et al. 1995, Macias-Silva et al. 1998), which is expressed in murine gonadotropes (Qiao et al. 2016). Nonetheless, BMP2 signaling in isolated gonadotropes from Bmpr1a cKO mice was greatly impaired (Fig. 1C), arguably ruling out a compensatory role for ACVR1 or another type I receptor, at least for this particular BMP ligand.
To determine whether or not ACVR1 can compensate for the loss of BMPR1A or is perhaps the preferred type I receptor for the biologically relevant BMP ligands in vivo, one could conditionally ablate the receptor (Kaartinen & Nagy 2001, Dudas et al. 2004) in gonadotropes either alone or in combination with BMPR1A (Yoon et al. 2005, Orvis et al. 2008, Edson et al. 2010). We have not systematically analyzed such mice to the extent that we have with Bmpr1a cKOs. However, we did generate a small number of Gnrhr GRIC/+; Bmpr1a fl/fl; Acvr1 fl/fl mice, which lack both receptors in gonadotropes. Thus far, double knockout females produce litters of normal size. This would suggest that FSH (and LH) synthesis is unimpaired, though we have not assessed this directly. Indeed, in light of these preliminary observations and the results of the in vitro recombination experiments presented in Fig. 5, a thorough analysis of the double knockouts appears to be unjustified.
We pursued BMP2 and BMPR1A (by extension) because BMP2 and BMP4 were ten-fold more potent than BMP6 or BMP7 in stimulating Fshb transcription in LβT2 cells (Lee et al. 2007). As BMP2 and BMP4 are expressed at relatively low levels in gonadotropes (Lee et al. 2007, Qiao et al. 2016), we argued that they would most likely regulate Fshb in a paracrine fashion. However, in light of our observations here with conditional Bmpr1a knockout mice and previous data implicating endogenous BMP7 in FSH regulation (Huang et al. 2001), it is fair to question whether a singular focus on BMP2 and its canonical receptor BMPR1A might have been limiting. That is, perhaps autocrine BMP7 is the primary BMP ligand regulating FSH in vivo. Indeed, Bmp7 is among the most highly expressed BMP ligands in isolated gonadotropes (Qiao et al. 2016). However, ACVR1 is the preferred BMP7 type I receptor (Macias-Silva et al. 1998), and the recombination experiments in Fig. 5 rule out a role for an endogenous ligand that signals via this receptor in Fshb synthesis, at least in primary culture.
In sum, though in vitro analyses in the gonadotrope-like LβT2 cell line suggested a role for BMP2 as a regulator of Fshb transcription both alone and in synergy with activins, the data presented here demonstrate that the ligand fails to stimulate Fshb expression in isolated primary gonadotropes and that its primary type I receptor, BMPR1A, is not required for normal FSH production or fertility in mice. Small-molecule BMPR1A, BMPR1B, and ACVR1 inhibitors suppressed Fshb expression in primary cells, suggesting a role for endogenous BMPs or related ligands in FSH synthesis. However, it now appears that the inhibitors’ effects were likely independent of BMP signaling as the genetic deletion of BMPR1A and ACVR1 together did not alter Fshb production in the same cells. In light of these data, we suggest that studies on the direct regulation of FSH synthesis by TGFβ superfamily ligands should focus on activins and related proteins that signal via ACVR1B (ALK4), TGFBR1 (ALK5), and/or ACVR1C (ALK7).
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
This work was supported by the Canadian Institutes of Health Research Operating Grant MOP-133394 and Natural Sciences and Engineering Research Council of Canada Discovery Grant 2015-05178 to DJB, and the National Institutes of Health (R01DE020843) to YM.
Author contribution statement
X Z performed all of the in vivo experiments; prepared, treated, and analyzed the pituitary cultures; and edited the manuscript. Y W performed several of the quantitative RT-PCR analyses and edited the manuscript. L O wrote sections of the manuscript and edited the final version. U B, V K, and Y M produced the floxed and Cre strains, provided advice regarding experimental design, and edited the manuscript. D J B designed the experiments, analyzed the data, generated the final figures and tables, wrote the original draft of the manuscript, and edited and formatted the final version of the manuscript.
Acknowledgments
The authors thank Ken McDonald from the McGill University Flow Cytometry Core Facility for technical assistance with cell sorting. The LAAC of the Center for Research in Reproduction at the University of Virginia is supported by the Eunice Kennedy Shriver NICHD/NIH (SCCPIR) Grant U54-HD28934.
References
Bailey JS, Rave-Harel N, McGillivray SM, Coss D & Mellon PL 2004 Activin regulation of the follicle-stimulating hormone beta-subunit gene involves Smads and the TALE homeodomain proteins Pbx1 and Prep1. Molecular Endocrinology 18 1158–1170. (doi:10.1210/me.2003-0442)
Bernard DJ 2004 Both SMAD2 and SMAD3 mediate activin-stimulated expression of the follicle-stimulating hormone beta subunit in mouse gonadotrope cells. Molecular Endocrinology 18 606–623. (doi:10.1210/me.2003-0264)
Bernard DJ & Tran S 2013 Mechanisms of activin-stimulated FSH synthesis: the story of a pig and a FOX. Biology of Reproduction 88 78. (doi:10.1095/biolreprod.113.107797)
Bernard DJ, Lee KB & Santos MM 2006 Activin B can signal through both ALK4 and ALK7 in gonadotrope cells. Reproductive Biology and Endocrinology 4 52. (doi:10.1186/1477-7827-4-52)
Bernard DJ, Fortin J, Wang Y & Lamba P 2010 Mechanisms of FSH synthesis: what we know, what we don't, and why you should care. Fertility and Sterility 93 2465–2485. (doi:10.1016/j.fertnstert.2010.03.034)
Chapman SC, Bernard DJ, Jelen J & Woodruff TK 2002 Properties of inhibin binding to betaglycan, InhBP/p120 and the activin type II receptors. Molecular and Cellular Endocrinology 196 79–93. (doi:10.1016/S0303-7207(02)00227-7)
Corrigan AZ, Bilezikjian LM, Carroll RS, Bald LN, Schmelzer CH, Fendly BM, Mason AJ, Chin WW, Schwall RH & Vale W 1991 Evidence for an autocrine role of activin B within rat anterior pituitary cultures. Endocrinology 128 1682–1684. (doi:10.1210/endo-128-3-1682)
Cuny GD, Yu PB, Laha JK, Xing X, Liu JF, Lai CS, Deng DY, Sachidanandan C, Bloch KD & Peterson RT 2008 Structure-activity relationship study of bone morphogenetic protein (BMP) signaling inhibitors. Bioorganic & Medicinal Chemistry Letters 18 4388–4392. (doi:10.1016/j.bmcl.2008.06.052)
Dudas M, Sridurongrit S, Nagy A, Okazaki K & Kaartinen V 2004 Craniofacial defects in mice lacking BMP type I receptor Alk2 in neural crest cells. Mechanisms of Development 121 173–182. (doi:10.1016/j.mod.2003.12.003)
Edson MA, Nalam RL, Clementi C, Franco HL, Demayo FJ, Lyons KM, Pangas SA & Matzuk MM 2010 Granulosa cell-expressed BMPR1A and BMPR1B have unique functions in regulating fertility but act redundantly to suppress ovarian tumor development. Molecular Endocrinology 24 1251–1266. (doi:10.1210/me.2009-0461)
Faure MO, Nicol L, Fabre S, Fontaine J, Mohoric N, McNeilly A & Taragnat C 2005 BMP-4 inhibits follicle-stimulating hormone secretion in ewe pituitary. Journal of Endocrinology 186 109–121. (doi:10.1677/joe.1.05988)
Fortin J, Kumar V, Zhou X, Wang Y, Auwerx J, Schoonjans K, Boehm U, Boerboom D & Bernard DJ 2013 NR5A2 regulates Lhb and Fshb transcription in gonadotrope-like cells in vitro, but is dispensable for gonadotropin synthesis and fertility in vivo. PLoS ONE 8 e59058. (doi:10.1371/journal.pone.0059058)
Fortin J, Boehm U, Deng CX, Treier M & Bernard DJ 2014 Follicle-stimulating hormone synthesis and fertility depend on SMAD4 and FOXL2. FASEB Journal 28 3396–3410. (doi:10.1096/fj.14-249532)
Gregory SJ, Lacza CT, Detz AA, Xu S, Petrillo LA & Kaiser UB 2005 Synergy between activin A and gonadotropin-releasing hormone in transcriptional activation of the rat follicle-stimulating hormone-beta gene. Molecular Endocrinology 19 237–254. (doi:10.1210/me.2003-0473)
Ho CC & Bernard DJ 2009 Bone morphogenetic protein 2 signals via BMPR1A to regulate murine follicle-stimulating hormone beta subunit transcription. Biology of Reproduction 81 133–141. (doi:10.1095/biolreprod.108.074211)
Ho CC & Bernard DJ 2010 Bone morphogenetic protein 2 acts via inhibitor of DNA binding proteins to synergistically regulate follicle-stimulating hormone beta transcription with activin A. Endocrinology 151 3445–3453. (doi:10.1210/en.2010-0071)
Ho CC, Zhou X, Mishina Y & Bernard DJ 2011 Mechanisms of bone morphogenetic protein 2 (BMP2) stimulated inhibitor of DNA binding 3 (Id3) transcription. Molecular and Cellular Endocrinology 332 242–252. (doi:10.1016/j.mce.2010.10.019)
Hollnagel A, Oehlmann V, Heymer J, Ruther U & Nordheim A 1999 Id genes are direct targets of bone morphogenetic protein induction in embryonic stem cells. Journal of Biological Chemistry 274 19838–19845. (doi:10.1074/jbc.274.28.19838)
Huang HJ, Wu JC, Su P, Zhirnov O & Miller WL 2001 A novel role for bone morphogenetic proteins in the synthesis of follicle-stimulating hormone. Endocrinology 142 2275–2283. (doi:10.1210/endo.142.6.8159)
Inman GJ, Nicolas FJ, Callahan JF, Harling JD, Gaster LM, Reith AD, Laping NJ & Hill CS 2002 SB-431542 is a potent and specific inhibitor of transforming growth factor-beta superfamily type I activin receptor-like kinase (ALK) receptors ALK4, ALK5, and ALK7. Molecular Pharmacology 62 65–74. (doi:10.1124/mol.62.1.65)
Kaartinen V & Nagy A 2001 Removal of the floxed neo gene from a conditional knockout allele by the adenoviral Cre recombinase in vivo. Genesis 31 126–129. (doi:10.1002/gene.10015)
Lamba P, Santos MM, Philips DP & Bernard DJ 2006 Acute regulation of murine follicle-stimulating hormone beta subunit transcription by activin A. Journal of Molecular Endocrinology 36 201–220. (doi:10.1677/jme.1.01961)
Lamba P, Fortin J, Tran S, Wang Y & Bernard DJ 2009 A novel role for the forkhead transcription factor FOXL2 in activin A-regulated follicle-stimulating hormone beta subunit transcription. Molecular Endocrinology 23 1001–1013. (doi:10.1210/me.2008-0324)
Lamba P, Wang Y, Tran S, Ouspenskaia T, Libasci V, Hebert TE, Miller GJ & Bernard DJ 2010 Activin A regulates porcine follicle-stimulating hormone beta-subunit transcription via cooperative actions of SMADs and FOXL2. Endocrinology 151 5456–5467. (doi:10.1210/en.2010-0605)
Lee KB, Khivansara V, Santos MM, Lamba P, Yuen T, Sealfon SC & Bernard DJ 2007 Bone morphogenetic protein 2 and activin A synergistically stimulate follicle-stimulating hormone beta subunit transcription. Journal of Molecular Endocrinology 38 315–330. (doi:10.1677/jme.1.02196)
Lewis KA, Gray PC, Blount AL, MacConell LA, Wiater E, Bilezikjian LM & Vale W 2000 Betaglycan binds inhibin and can mediate functional antagonism of activin signalling. Nature 404 411–414. (doi:10.1038/35006129)
Liu F, Ventura F, Doody J & Massague J 1995 Human type II receptor for bone morphogenic proteins (BMPs): extension of the two-kinase receptor model to the BMPs. Molecular and Cellular Biology 15 3479–3486. (doi:10.1128/MCB.15.7.3479)
Macias-Silva M, Hoodless PA, Tang SJ, Buchwald M & Wrana JL 1998 Specific activation of Smad1 signaling pathways by the BMP7 type I receptor, ALK2. Journal of Biological Chemistry 273 25628–25636. (doi:10.1074/jbc.273.40.25628)
Mishina Y, Suzuki A, Ueno N & Behringer RR 1995 Bmpr encodes a type I bone morphogenetic protein receptor that is essential for gastrulation during mouse embryogenesis. Genes & Development 9 3027–3037.
Mishina Y, Hanks MC, Miura S, Tallquist MD & Behringer RR 2002 Generation of Bmpr/Alk3 conditional knockout mice. Genesis 32 69–72. (doi:10.1002/gene.10038)
Nicol L, Faure MO, McNeilly JR, Fontaine J, Taragnat C & McNeilly AS 2008 Bone morphogenetic protein-4 interacts with activin and GnRH to modulate gonadotrophin secretion in LbetaT2 gonadotrophs. Journal of Endocrinology 196 497–507. (doi:10.1677/JOE-07-0542)
Orvis GD, Jamin SP, Kwan KM, Mishina Y, Kaartinen VM, Huang S, Roberts AB, Umans L, Huylebroeck D & Zwijsen A et al. 2008 Functional redundancy of TGF-beta family type I receptors and receptor-Smads in mediating anti-Mullerian hormone-induced Mullerian duct regression in the mouse. Biology of Reproduction 78 994–1001. (doi:10.1095/biolreprod.107.066605)
Otsuka F & Shimasaki S 2002 A novel function of bone morphogenetic protein-15 in the pituitary: selective synthesis and secretion of FSH by gonadotropes. Endocrinology 143 4938–4941. (doi:10.1210/en.2002-220929)
Qiao S, Nordstrom K, Muijs L, Gasparoni G, Tierling S, Krause E, Walter J & Boehm U 2016 Molecular plasticity of male and female murine gonadotropes revealed by mRNA sequencing. Endocrinology 157 1082–1093. (doi:10.1210/en.2015-1836)
Safwat N, Ninomiya-Tsuji J, Gore AJ & Miller WL 2005 Transforming growth factor beta-activated kinase 1 is a key mediator of ovine follicle-stimulating hormone beta-subunit expression. Endocrinology 146 4814–4824. (doi:10.1210/en.2005-0457)
Sallon C, Faure MO, Fontaine J & Taragnat C 2010 Dynamic regulation of pituitary mRNAs for bone morphogenetic protein (BMP) 4, BMP receptors, and activin/inhibin subunits in the ewe during the estrous cycle and in cultured pituitary cells. Journal of Endocrinology 207 55–65. (doi:10.1677/JOE-10-0035)
Srinivas S, Watanabe T, Lin CS, William CM, Tanabe Y, Jessell TM & Costantini F 2001 Cre reporter strains produced by targeted insertion of EYFP and ECFP into the ROSA26 locus. BMC Developmental Biology 1 4. (doi:10.1186/1471-213X-1-4)
Suszko MI, Lo DJ, Suh H, Camper SA & Woodruff TK 2003 Regulation of the rat follicle-stimulating hormone beta-subunit promoter by activin. Molecular Endocrinology 17 318–332. (doi:10.1210/me.2002-0081)
Suszko MI, Balkin DM, Chen Y & Woodruff TK 2005 Smad3 mediates activin-induced transcription of follicle-stimulating hormone beta-subunit gene. Molecular Endocrinology 19 1849–1858. (doi:10.1210/me.2004-0475)
Takeda M, Otsuka F, Takahashi H, Inagaki K, Miyoshi T, Tsukamoto N, Makino H & Lawson MA 2012 Interaction between gonadotropin-releasing hormone and bone morphogenetic protein-6 and -7 signaling in LbetaT2 gonadotrope cells. Molecular and Cellular Endocrinology 348 147–154. (doi:10.1016/j.mce.2011.08.001)
ten Dijke P, Yamashita H, Sampath TK, Reddi AH, Estevez M, Riddle DL, Ichijo H, Heldin CH & Miyazono K 1994 Identification of type I receptors for osteogenic protein-1 and bone morphogenetic protein-4. Journal of Biological Chemistry 269 16985–16988.
Tran S, Lamba P, Wang Y & Bernard DJ 2011 SMADs and FOXL2 synergistically regulate murine FSHbeta transcription via a conserved proximal promoter element. Molecular Endocrinology 25 1170–1183. (doi:10.1210/me.2010-0480)
Tsuchida K, Nakatani M, Yamakawa N, Hashimoto O, Hasegawa Y & Sugino H 2004 Activin isoforms signal through type I receptor serine/threonine kinase ALK7. Molecular and Cellular Endocrinology 220 59–65. (doi:10.1016/j.mce.2004.03.009)
Vogt J, Traynor R & Sapkota GP 2011 The specificities of small molecule inhibitors of the TGFss and BMP pathways. Cell Signaling 23 1831–1842. (doi:10.1016/j.cellsig.2011.06.019)
Wang Y & Bernard DJ 2012 Activin A induction of murine and ovine follicle-stimulating hormone beta transcription is SMAD-dependent and TAK1 (MAP3K7)/p38 MAPK-independent in gonadotrope-like cells. Cell Signaling 24 1632–1640. (doi:10.1016/j.cellsig.2012.04.006)
Weiss J, Guendner MJ, Halvorson LM & Jameson JL 1995 Transcriptional activation of the follicle-stimulating hormone beta-subunit gene by activin. Endocrinology 136 1885–1891. (doi:10.1210/endo.136.5.7720634)
Wen S, Schwarz JR, Niculescu D, Dinu C, Bauer CK, Hirdes W & Boehm U 2008 Functional characterization of genetically labeled gonadotropes. Endocrinology 149 2701–2711. (doi:10.1210/en.2007-1502)
Wen S, Ai W, Alim Z & Boehm U 2010 Embryonic gonadotropin-releasing hormone signaling is necessary for maturation of the male reproductive axis. PNAS 107 16372–16377. (doi:10.1073/pnas.1000423107)
Yi SE, LaPolt PS, Yoon BS, Chen JY, Lu JK & Lyons KM 2001 The type I BMP receptor BmprIB is essential for female reproductive function. PNAS 98 7994–7999. (doi:10.1073/pnas.141002798)
Yoon BS, Ovchinnikov DA, Yoshii I, Mishina Y, Behringer RR & Lyons KM 2005 Bmpr1a and Bmpr1b have overlapping functions and are essential for chondrogenesis in vivo. PNAS 102 5062–5067. (doi:10.1073/pnas.0500031102)
Young JM, Juengel JL, Dodds KG, Laird M, Dearden PK, McNeilly AS, McNatty KP & Wilson T 2008 The activin receptor-like kinase 6 Booroola mutation enhances suppressive effects of bone morphogenetic protein 2 (BMP2), BMP4, BMP6 and growth and differentiation factor-9 on FSH release from ovine primary pituitary cell cultures. Journal of Endocrinology 196 251–261. (doi:10.1677/JOE-07-0148)
Yu PB, Hong CC, Sachidanandan C, Babitt JL, Deng DY, Hoyng SA, Lin HY, Bloch KD & Peterson RT 2008 Dorsomorphin inhibits BMP signals required for embryogenesis and iron metabolism. Nature Chemical Biology 4 33–41. (doi:10.1038/nchembio.2007.54)