The infantile myofibromatosis NOTCH3 L1519P mutation leads to hyperactivated ligand-independent Notch signaling and increased PDGFRB expression

Abstract

Infantile myofibromatosis (IMF) is a benign tumor form characterized by the development of nonmetastatic tumors in skin, bone, muscle and sometimes viscera. Autosomal dominant forms of IMF are caused by mutations in the PDGFRB gene, but a family carrying a L1519P mutation in the NOTCH3 gene has also recently been identified. In this report, we address the molecular consequences of the NOTCH3^L1519P mutation and the relationship between the NOTCH and PDGFRB signaling in IMF. The NOTCH3^L1519P receptor generates enhanced downstream signaling in a ligand-independent manner. Despite the enhanced signaling, the NOTCH3^L1519P receptor is absent from the cell surface and instead accumulates in the endoplasmic reticulum. Furthermore, the localization of the NOTCH3^L1519P receptor in the bipartite, heterodimeric state is altered, combined with avid secretion of the mutated extracellular domain from the cell. Chloroquine treatment strongly reduces the amount of secreted NOTCH3^L1519P extracellular domain and decreases signaling. Finally, NOTCH3^L1519P upregulates PDGFRB expression in fibroblasts, supporting a functional link between Notch and PDGF dysregulation in IMF. Collectively, our data define a NOTCH3-PDGFRB axis in IMF, where an IMF-mutated NOTCH3 receptor elevates PDGFRB expression. The functional characterization of a ligand-independent gain-of-function NOTCH3 mutation is important for Notch therapy considerations for IMF, including strategies aimed at altering lysosome function.

Keywords: Fibroblast; Infantile myofibromatosis; Notch; PDGF.

Fig. 1.

Aberrant processing of NOTCH3^L1519P. (A) Schematic overview of Notch proteolytic processing. (B) Expression of the full-length (FL), transmembrane intracellular domain (TMIC) and Notch extracellular truncated (NEXT)/Notch intracellular domain (NICD) forms was analyzed by western blotting of cell extracts from HEK 293T ΔN1-3 cells (Control) or HEK 293T ΔN1-3 cells expressing wild-type (WT) and NOTCH3^L1519P (L1519). β-actin levels were used as a loading control (n=3). Relative levels of TMIC/FL were quantified by ImageJ from three experiments and analyzed by paired Student's t-test, **P<0.01. ECD, extracellular domain; EGF, epidermal growth factor; NRR, negative regulatory region.

Fig. 2.

Lack of cell surface expression and increased retention of NOTCH3^L1519P in the endoplasmic reticulum (ER). (A,B) Confocal images of wild-type and NOTCH ^L1519P-expressing cells without (A) and with (B) permeabilization. Na,K-ATPase was used as a membrane marker. (C) Immunocytochemistry for NOTCH3 ECD (green) and the ER marker calnexin (red) from permeabilized cells; the plot shows the Pearson's correlation coefficients for the colocalization analysis. (D) Immunocytochemistry for NOTCH3 ECD (red) and the Golgi marker giantin (green); the graph shows the Pearson's correlation coefficient analysis (n=3). Significance was calculated using unpaired Student's t-test, ***P<0.001; ns, not significant. Scale bars: 20 µm.

Fig. 3.

NOTCH3^L1519P undergoes enhanced and ligand-independent S2 cleavage. (A) Western blot analysis of TMIC and NEXT/NICD fragments in the presence of the S2 inhibitor GI254023X or the γ-secretase inhibitor DAPT. (B) Treatment of the wild-type and NOTCH3^L1519P-expressing cells with cycloheximide at different time points, as indicated. The graph shows densitometric quantification of TMIC using ImageJ (n=3). Data are from three independent experiments.

Fig. 4.

Relocalization of the ECD-TMIC heterodimer and exacerbated export of NOTCH3^L1519P ECD into the cell medium. (A) Schematic representation of the proximity ligation assay (PLA) identifying the ECD-TMIC heterodimer. (B) PLA (red) was performed on wild-type or NOTCH3^L1519P-expressing cells, using mouse anti-ECD 1E4 and rabbit anti-NOTCH3 ICD antibodies. The cell membrane (green) was directly labeled with anti-Na/K ATPase-Plasma Membrane Marker (Alexa Fluor^® 488). (C) 3D surface rendering of the PLA and cell membrane staining. (D) Immunoprecipitation of supernatants from cells expressing wild-type NOTCH3 or NOTCH3^L1519P mutation using V5-agarose beads followed by western blot analysis using anti-V5 antibody. β-actin levels were used as a loading control. (E) Western blot analysis of the conditioned media from cells expressing NOTCH3^L1519P treated with dimethyl sulfoxide (DMSO), GM6001, GI254023X, chloroquine, DAPT and MitMAB, as indicated. β-actin levels were used as a loading control for whole-cell extracts. (F) Immunocytochemistry for NOTCH3 ECD (green) and the endosomal marker EEA1 (red). (G) Immunocytochemistry for NOTCH3 ECD (green) and LAMP1 (red). (H) Immunocytochemistry for NOTCH3 ECD (green) and LAMP1 in cells treated with chloroquine (CQ). The plots show the Pearson's correlation coefficiency for the colocalization analysis (n=3). Significance was calculated using unpaired Student's t-test, **P<10⁻², ***P<10⁻³. Scale bars: 10 µm.

Fig. 5.

NOTCH3^L1519P produces ICD in a ligand-independent manner. (A) Western blot analysis of FL, TMIC and NEXT and ICD fragments from cell extracts using an anti-NOTCH3 ICD antibody, following activation of the cells by immobilized jagged2 (Jag2-Fc) and treatment with the γ-secretase inhibitor DAPT or MG132, as indicated. (B) NIH3T3 cells were transfected with wild-type, NOTCH3^L1519P or control plasmid together with 12XCSL-luc reporter and β-gal plasmids and cultured on immobilized jagged2 (Jag2) in combination with treatment by DMSO or DAPT, as indicated (n=3). Statistical analysis was performed from three experiments using paired Student's t-test, *P<0.1, **P<10⁻², ***P<10⁻³, RLU, relative luminescence units.

Fig. 6.

NOTCH3^L1519P hyperactivates expression of Notch downstream genes. (A) Western blot analysis of PDGFRB expression in HMFΔN2 cells expressing wild-type NOTCH3 (WT) or NOTCH3^L1519P (L1519P) compared to the parental HMFΔN2 cells (Ctrl). The graph shows PDGFRB expression normalized to β-actin using ImageJ. (B) Quantitative real-time PCR analysis of the Notch downstream target genes NOTCH3, HES1, HEY1, NRARP and PDGFRB from wild-type or NOTCH3^L1519P-expressing HMFΔN2 cells. (C) Western blot analysis of PDGFRB, AKT and p42/44 MAPK phosphorylation upon PDGF-BB stimulation as indicated. The blot with p42/44 MAPK was stripped and re-probed with the β-actin antibody for a loading control. Relative levels were normalized to β-actin using ImageJ (n=3). Statistical significant was analyzed from three experiments using paired Student's t-test, *P<0.1, **P<0.01, ***P<0.001; ns, not significant.

Fig. 7.

PDGFRB IMF mutations are gain-of-function mutations. (A) Analysis of phosphorylation levels from different PDGRFB mutants after stimulation with PDGF-BB, as indicated. (B) Quantification of phosphorylation of specific tyrosine residues (Y751, Y771, Y1009, Y1021) in the various PDGFRB mutants, as indicated (n=3). (C) Analysis of phosphorylation levels of PDGFRB, SHP2, Akt, p42/44 MAPK and eIF4E (as control) at different time points after PDGF-BB stimulation (0, 5, 15 and 60 min), as indicated. ***P<0.001 compared to the stimulated WT control (black bars); ^#P<0.05 and ^###P<0.001 compared to unstimulated WT control (unpaired two-tailed Student's t-test). KD, knockdown.