Distinct molecular pathways mediate Mycn and Myc-regulated miR-17-92 microRNA action in Feingold syndrome mouse models

Abstract

Feingold syndrome is a skeletal dysplasia caused by loss-of-function mutations of either MYCN (type 1) or MIR17HG that encodes miR-17-92 microRNAs (type 2). Since miR-17-92 expression is transcriptionally regulated by MYC transcription factors, it has been postulated that Feingold syndrome type 1 and 2 may be caused by a common molecular mechanism. Here we show that Mir17-92 deficiency upregulates TGF-β signaling, whereas Mycn-deficiency downregulates PI3K signaling in limb mesenchymal cells. Genetic or pharmacological inhibition of TGF-β signaling efficiently rescues the skeletal defects caused by Mir17-92 deficiency, suggesting that upregulation of TGF-β signaling is responsible for the skeletal defect of Feingold syndrome type 2. By contrast, the skeletal phenotype of Mycn-deficiency is partially rescued by Pten heterozygosity, but not by TGF-β inhibition. These results strongly suggest that despite the phenotypical similarity, distinct molecular mechanisms underlie the pathoetiology for Feingold syndrome type 1 and 2.

Fig. 1

Generation of mouse models of Feingold syndrome. a Forelimbs of wildtype control (Ctrl), Mycn Knockout (Prx1-Cre:Mycn^fl/fl, Mycn cKO), Mir17-92 knockout (Prx1-Cre:Mir17-92^fl/fl, 17 KO), and Mir17-92 and Mir106b-25 doubly conditional knockout (Prx1-Cre:Mir17-92^fl/fl:Mir106b-25^−/−, 17 dKO) mice at postnatal day (P) 16. 17 KO mutants exhibit mild brachydactyly in the fifth digit (double-headed arrows). Cutaneous syndactyly (arrows) and severe brachydactyly are presented in 17 dKO and Mycn cKO mice. Scale bar, 0.5 cm. b Alizarin red and alcian blue staining of the forelimb and magnified views of the fifth digit of corresponding mice in a. 17 KO mutants show shortening of the middle phalanx (M) (brackets) of the fifth digit. 17 dKO and Mycn cKO mutants show severe brachymesophalangy and the absence of the middle phalanx of the fifth digit (arrows). c Alizarin red and alcian blue staining of the skull of corresponding mice in a. Mycn cKO, 17 KO, and 17 dKO mice show microcephaly and a frontal bone ossification defect (arrows). Scale bar, 1 cm. d Quantification of the length of the fifth M and fifth metacarpal bone (5th MC) and the ratio of 5th M to 5th MC. Values are expressed as mean ± SE (n = 5 in each group, *p < 0.001 vs. Ctrl). ND, not detected. e Relative expression of five miRNAs encoded in the Mir17-92 gene in limb bud mesenchymal cells isolated from Mycn cKO embryos at embryonic day 14.5. (relative to control littermate samples, n = 5 each group, *<0.001 vs. Ctrl at every time point). f Hindlimbs of wildtype Ctrl, Mycn cKO, and Mycn Knockout overexpressing Mir17-92 (Prx1-Cre:Mycn^fl/fl:Mir17-92^Tg, Mycn cKO:17 Tg) (upper panels) and alizarin red and alcian blue staining of the hindlimbs of corresponding genotypes (bottom panels). Mir17-92 overexpression partially rescues the skeletal defects of Mycn cKO mutants. Cutaneous syndactyly (white arrows) and shortening of the fifth digit in Mycn cKO mutants are rescued by overexpression of Mir17-92 whereas fused fourth and fifth metatarsal bones still remain (bottom panels). Scale bar, 0.5 cm

Fig. 2

Proliferation defect in cells missing Mir17-92 and Mir106b-25. a–d Loss of Mir17-92 and Mir106b-25 (17 dKO) impairs skeletal mesenchymal cell proliferation. Since Mir106b-25-null mice show no skeletal abnormalities, Mir106b-null mice were used as the control. a BrdU labeling on limb bud sections of Mir17-92^fl/fl:Mir106b-25^−/−(Ctrl) and Prx1-Cre:Mir17-92^fl/fl:Mir106b-25^−/− (17 dKO) embryos at embryonic day (E) 10.5. b BrdU labeling of coronal skull sections of Ctrl and 17 dKO embryos at E13.5. c The BrdU-positive cells were significantly reduced in the 17 dKO limb bud and skull mesenchyme (n = 6, *p < 0.05). d Cell proliferation assay on limb bud and skull mesenchymal cells (MC) in vitro. Primary mesenchymal fibroblasts isolated from Mir17-92^fl/fl:Mir106b-25^−/− embryos were transduced with adenoviruses expressing a Cre recombinase (17 dKO) or a yellow fluorescent protein (Ctrl) to delete Mir17-92 in vitro. The efficient reduction (85 ± 5%) in miR-17-92 miRNAs was confirmed by qRT-PCR. Mir17-92-deficient limb bud and calvarial mesenchymal cells showed significantly impaired proliferation over time compared with control, as assessed by an index of the increase in cell mass (n = 5, *p < 0.05). e Whole-mount in situ hybridization for Sox9 shows a reduction in size of the Sox9-positive domain at E9.5 and E11.5. FS frontal suture. Scale bar, 100 μm

Fig. 3

TGF-β and STAT3 upregulation in miR-17-92 miRNA-deficient cells. a Immunoblot analysis on indicated proteins using protein lysates from cells cultured in DMEM containing 10% FBS. Levels of p-Smad2, Tgfbr2, and p-Stat3 (arrows) are increased in Mir17-92:Mir106b-25-deficient limb bud cells (17 dKO). b Quantification of the p-Smad2 to total Smad2 (t-Smad2) and p-Stat3 to total Stat3 (t-Stat3) protein ratios based on western blot data (n = 3, *p < 0.05). c Increased TGF-β signaling in 17 dKO cells upon stimulation. Serum-starved cells were treated with 50 ng/ml TGF-β1. Smad2 phosphorylation was analyzed at the indicated time points. d Relative mRNA expression of known targets of miR-17-92 miRNAs in 17 dKO limb bud cells (n = 3, *p < 0.05). e Luciferase reporter assay for miR-17 regulation on Tgfbr2 binding site. Primary limb bud cells were co-transfected with control miRNA mimic (Ctrl miR) or mmu-miR-17-5p (miR-17) and a luciferase reporter construct carrying a wildtype (Wt 3′UTR) or mutated 3′UTR (Mut 3′UTR) sequence of mouse Tgfbr2. The predicted binding sequence of Tgfbr2 3′UTR and relative luciferase units (RLU) are shown (n = 6, *p < 0.001)

Fig. 4

Rescue of cell proliferation defects with TGF-β receptor inhibitors. a Cell proliferation assay on control (Ctrl) and Mir17-92:Mir106b-25-deficient (17 dKO) limb bud mesenchymal cells treated with vehicle (DMSO) or a STAT3 inhibitor (STAT3 inh, S31-201, 100 µM). Ctrl and 17 dKO limb bud cells were prepared by transducing YPF (Ctrl) or Cre (17 dKO) in vitro. STAT3 inhibitor treatment shows no significant improvement in 17 dKO cell proliferation. b Treatment of limb bud cells with S31-201 downregulates the p-Stat3 level; *non-specific band. c Cell proliferation assay on Ctrl and 17 dKO (dKO) cells treated with DMSO or a TGF-β receptor inhibitor (Tgfbr inh, Ly364947, 0.2 µM). Treatment with Ly364947 has no significant effect on control cells whereas it significantly ameliorates the proliferation defect of 17 dKO cells (n = 6, *p < 0.05 vs. 17 dKO + DMSO). d Treatment of limb bud cells with Ly364947 downregulates the p-Smad2 level. e–g Ly364947 treatment efficiently rescues the skeletal defects of 17 dKO mutants. Ly364947 (1 mg/kg/day, i.p.) was injected into pregnant and nursing mothers from E9.5 through P7.5, and then injected directly into individual mice. The shortening of the fifth digit (double arrows) syndactyly (white arrows) (e), missing middle phalanx (M) (f), and microcephaly and frontal bone ossification defect (black arrows) (g) in 17 dKO mutants were ameliorated by Ly364947 treatment. h Quantification of the length of the fifth mesophalanx (5th M), fifth metacarpal bone (5th MC), and the ratio of 5th M to 5th MC. Values are expressed as mean ± SE (n = 6 each group, p < 0.001 vs. dKO). Scale bars: 0.5 cm in e, 1.0 cm in g. D distal phalanx; P proximal phalanx

Fig. 5

Distinct mechanisms in Feingold syndrome mouse models. a Forelimbs of mice with indicated genotypes at postnatal day 16. Wildtype control (Ctrl), Mycn conditional Knockout (Prx1-Cre:Mycn^fl/fl, Mycn cKO), Mycn cKO in which one allele of Tgfbr2 is deleted (Prx1-Cre:Mycn^fl/fl:Tgfbr2^fl/+, Mycn cKO:Tgfbr2 Het), Mir17-92 and Mir106b-25 doubly conditional knockout (Prx1-Cre:Mir17-92^fl/fl:Mir106b-25^−/−, 17 dKO), and 17 dKO in which one allele of Tgfbr2 is deleted (Prx1-Cre:Mir17-92^fl/fl:Mir106b-25^−/−:Tgfbr2^fl/+, 17 dKO:Tgfbr2 Het). Unlike 17 dKO mice, the digit abnormalities of Mycn cKO mutants are not rescued by Tgfbr2 heterozygous deletion. b Alizarin red and alcian blue staining of forelimbs of corresponding mice in a. Scale bars, 0.5 cm. More than three rescued mice per each model were analyzed to confirm the reproducibility. c Immunoblot analysis for indicated proteins upon treatment with 10% fetal bovine serum (FBS) at indicated time points (top panel). Primary mesenchymal fibroblasts from limb buds of E10.5 Mycn^fl/fl embryos were transduced with adenoviruses expressing a Cre recombinase (KO) or a yellow fluorescent protein (Ctrl) to delete Mycn in vitro. Cells were serum starved for 1 h before stimulation. The efficient reduction in the Mycn protein level was confirmed. Levels of p-Akt (Thr308) and p-S6k are decreased in Mycn-deficient limb bud cells (KO). Quantification of p-Akt relative to total Akt (t-Akt) and p-Smad2 to total Smad2 (t-Smad2) at time 0 based on the western blot data (bottom panels) (n = 3, *p < 0.05). d Relative expression of previously reported targets of miR-17-92 miRNAs in Mycn cKO limb bud mesenchymal cells isolated from embryos at age E10.5.  Y axis, arbitrary units

Fig. 6

PI3K downregulation contributes to the Mycn cKO phenotype. a Hindlimbs of mice with indicated genotypes at postnatal day 16. Wildtype control (Ctrl), heterozygous Pten mutant (Prx1-Cre:Pten^fl/+, Pten Het), Mycn Knockout (Prx1-Cre:Mycn^fl/fl, Mycn cKO), Mycn cKO in which one allele of Pten is deleted (Prx1-Cre:Mycn^fl/fl:Pten^fl/+, Mycn cKO:Pten Het), and Mycn Knockout overexpressing Mir17-92 (Prx1-Cre:Mycn^fl/fl:Mir17-92^Tg, Mycn cKO:17 Tg). Pten heterozygous deletion partially rescues the skeletal abnormalities of Mycn cKO mutants. Cutaneous syndactyly and brachydactyly are partially rescued, while the fused metatarsals and shortening of mid phalanxes remain. The fifth metacarpal bones, subjected to measurement in b, are indicated by brackets. Scale bars, 0.5 cm. b Quantification of the length (µm) of fifth metacarpal bones of mice with indicated genotypes (n = 6 for Ctrl and Mycn cKO, n = 5 for Pten Het and Mycn cKO:Pten Het, n = 4 for Mycn cKO:17 Tg; *p < 0.01 vs. Ctrl, **p < 0.01 vs. Mycn cKO). c PI3K signaling was assessed by p-Akt (Thr308) in limb bud cells isolated from E 10.5 mice with indicated genotype. d Immunoblot analysis for phospho-Pten (p-Pten) and total Pten (t-Pten) isolated from Mycn cKO and wildtype littermate limb bud cells. The p-Pten level is decreased in Mycn-deficient skeletal progenitor cells whereas the t-Pten level is unchanged. e Immunoblot analysis for casein kinase 2 beta phosphorylation on S209 (p-CK2) in limb bud mesenchymal cells isolated from E10.5 Mycn cKO and wildtype littermates. p-CK2 is decreased in Mycn-deficient cells; *non-specific band. f Proposed model. Upregulation of TGF-β signaling plays a causal role in the skeletal defect of Feingold syndrome type 2 (Mir17-92 mutation), whereas downregulation of the PI3K signaling plays a major pathophysiologic role in Feingold syndrome type 1 (Mycn mutation). Mycn partially regulates miR-17-92 miRNA levels, but the contribution of this regulation in limb development is limited (dotted arrow). Overexpression of miR-17-92 miRNAs can suppress Pten, although miR-17-92 miRNAs have limited regulatory effects on Pten expression at the physiological level (dotted inhibitory line)