Omics Profiling of S2P Mutant Fibroblasts as a Mean to Unravel the Pathomechanism and Molecular Signatures of X-Linked MBTPS2 Osteogenesis Imperfecta

Abstract

Osteogenesis imperfecta (OI) is an inherited skeletal dysplasia characterized by low bone density, bone fragility and recurrent fractures. The characterization of its heterogeneous genetic basis has allowed the identification of novel players in bone development. In 2016, we described the first X-linked recessive form of OI caused by hemizygous MBTPS2 missense variants resulting in moderate to severe phenotypes. MBTPS2 encodes site-2 protease (S2P), which activates transcription factors involved in bone (OASIS) and cartilage development (BBF2H7), ER stress response (ATF6) and lipid metabolism (SREBP) via regulated intramembrane proteolysis. In times of ER stress or sterol deficiency, the aforementioned transcription factors are sequentially cleaved by site-1 protease (S1P) and S2P. Their N-terminal fragments shuttle to the nucleus to activate gene transcription. Intriguingly, missense mutations at other positions of MBTPS2 cause the dermatological spectrum condition Ichthyosis Follicularis, Atrichia and Photophobia (IFAP) and Keratosis Follicularis Spinulosa Decalvans (KFSD) without clinical overlap with OI despite the proximity of some of the pathogenic variants. To understand how single amino acid substitutions in S2P can lead to non-overlapping phenotypes, we aimed to compare the molecular features of MBTPS2-OI and MBTPS2-IFAP/KFSD, with the ultimate goal to unravel the pathomechanisms underlying MBTPS2-OI. RNA-sequencing-based transcriptome profiling of primary skin fibroblasts from healthy controls (n = 4), MBTPS2-OI (n = 3), and MBTPS2-IFAP/KFSD (n = 2) patients was performed to identify genes that are differentially expressed in MBTPS2-OI and MBTPS2-IFAP/KFSD individuals compared to controls. We observed that SREBP-dependent genes are more downregulated in OI than in IFAP/KFSD. This is coupled to alterations in the relative abundance of fatty acids in MBTPS2-OI fibroblasts in vitro, while no consistent alterations in the sterol profile were observed. Few OASIS-dependent genes are suppressed in MBTPS2-OI, while BBF2H7- and ATF6-dependent genes are comparable between OI and IFAP/KFSD patients and control fibroblasts. Importantly, we identified genes involved in cartilage physiology that are differentially expressed in MBTPS2-OI but not in MBTPS2-IFAP/KFSD fibroblasts. In conclusion, our data provide clues to how pathogenic MBTPS2 mutations cause skeletal deformities via altered fatty acid metabolism or cartilage development that may affect bone development, mineralization and endochondral ossification.

Keywords: MBTPS2; X-linked osteogenesis imperfecta; connective tissue; site 2 protease; transcriptomics.

FIGURE 1

Predicted secondary structure of S2P based on Uniprot topology annotations. The figure was generated using the TOPO2 display software for transmembrane proteins (Johns, 2020). The histidine residues of the His-Glu-(any amino acid)₂-His (H₁₇₁EXXH) motif and aspartate residue of the L₄₆₆DG sequence, depicted in green, coordinate the catalytic site zinc atom of S2P. Known IFAP/KFSD-causative variants are depicted in blue; OI-causative variants are depicted in magenta.

FIGURE 2

Volcano plots of transcriptome profiles of panels (A) MBTPS2-OI and (B) MBTPS2-IFAP/KFSD patient-derived fibroblasts compared to healthy control fibroblasts.

FIGURE 3

Heatmap depicting expression levels of genes that are transcriptionally regulated by ATF6 at basal conditions (nt: no treatment) and in the presence of thapsigargin (TG). The dendrogram on the left y-axis shows clustering of the genes.

FIGURE 4

In vitro electron microscopy analyses of fibroblasts derived from (A,B) healthy controls, (C–E) MBTPS2-OI and (F,G) MBTPS2-IFAP/KFSD patients. In each image a portion of the fibroblasts is depicted with abundant ER cisterns. Short and large shaped ER cisterns are shown in the MBTPS2-OI [(C–E), white arrows] and MBTPS2-IFAP/KFSD patients [(F,G), white arrows] in contrast to the long and slender shaped ER cisterns shown in the two controls [(A,B), white large arrows].

FIGURE 5

Heatmap depicting expression levels of genes that are transcriptionally regulated by OASIS at basal conditions (nt: no treatment) and in the presence of thapsigargin (TG). The dendrogram on the left y-axis shows clustering of the genes.

FIGURE 6

Heatmap depicting expression levels of genes that are transcriptionally regulated by BBF2H7 at basal conditions (nt: no treatment) and in the presence of thapsigargin (TG). The dendrogram on the left y-axis shows clustering of the genes.

FIGURE 7

Heatmap depicting expression levels of SREBP-dependent genes at basal conditions. The dendrogram on the left y-axis shows clustering of the genes.

FIGURE 8

Deposition of collagen type IV proteins in the extracellular matrix by (A,B) healthy control fibroblasts, (C,D) MBTPS2-OI patient fibroblasts, and (E,F) MBTPS2-IFAP/KFSD patient fibroblasts was determined by immunofluorescence staining. Collagen type IV proteins are depicted in red and nuclei are depicted in blue.

FIGURE 9

qRT-PCR was performed to validate the RNA-sequencing results for (A) SREBP-dependent genes involved in fatty acid metabolism, (B) SREBP-dependent genes involved in sterol metabolism, (C) OASIS-dependent genes involved in skeletal development, and (D) genes involved in bone or cartilage homeostasis that are differentially expressed in MBTPS2-OI compared to control. Gene expression was measured in four independent replicates per subject, and t-tests were performed for statistical analysis. Data are expressed as mean ± SEM.

FIGURE 10

A simplified diagram of mammalian fatty acid metabolism. The enzymes listed in this diagram are encoded by genes that are transcriptionally regulated by SREBP.

FIGURE 11

(A−C) The cellular content of several saturated, monounsaturated and polyunsaturated omega-6 and omega-3 fatty acids were determined by GC-MS/MS and expressed as ratios. Four independent measurements were performed per subject, and t-tests were performed for statistical analysis. Data are expressed as mean ± SEM.