br STAR Methods br Acknowledgments

STAR★Methods
Acknowledgments We thank the TCGA PanCanAtlas Analysis Working Group. This study was supported in part by grants from the NIH (CA175486 and CA209851 to H.L., CA217842 to G.B.M., and CCSG grant CA016672); a grant from the Cancer Prevention and Research Institute of Texas (RP140462 to H.L.); a University of Texas System STARS award (to H.L.); a Swiss National Science Foundation fellowship (P2BSP3_161941 to M.T.D.) and a fellowship from the Gulf Coast Consortia on the Computational Cancer Biology Training Program (CPRIT Grant No. RP170593 to Y.W.). We thank the MD Anderson high-performance computing core facility for computing and LeeAnn Chastain for editorial assistance.
Introduction With an incidence of 0.5–1 per 1000 births, congenital scoliosis (CS) is characterized by a lateral curvature of the spine exceeding 10° that is due to the presence of one or more abnormally formed vertebrae [1,2]. Genetic inheritance is responsible for some congenital vertebral anomalies, such as spondylocostal dysostosis and spondylothoracic dysostosis [3]. Wu et al. reported that 11% of CS cases harbored a rare null mutation and/or a hypomorphic allele of TBX6 [4]. Therefore, some CS cases may develop due to a TBX6 abnormality, although the genetic etiology of most CS patients remains unclear. Olfactomedin-like 1 (OLFML1) is a member of the olfactomedin (OLF) protein family containing an olfactomedin-like domain in the C-terminus [5]. OLF family expression is restricted to certain tissues and has been implicated in both normal development and pathology [[6], [7], [8], [9], [10], [11], [12], [13], [14]]. OLF proteins contribute to the formation of extracellular matrix and play a role in developmental processes [[13], [14], [15], [16]]. OLFML1 is considered a secreted glycoprotein that enhances trans-isomer progression in human cancer cell lines in vitro [17]. However, its physiological roles require clarification [18]. This study examined a Japanese patient with CS and his parents by whole-exome sequencing (WES) to identify a de novo missense mutation in the OLFML1 gene (Met250Leu; Supplemental Figure S1A‒D). We evaluated a transcript abundance of Olfml1 in newborn mice and found it to be predominantly expressed in osteoblasts. Experimentally induced gain or loss of Olfml1 function revealed that the protein suppressed the Hippo signaling pathway and inhibited mineralization in osteoblasts. Our data provide evidence of a novel cell-autonomous mechanism by which osteoblasts avoid excess mineralization through Olfml1.
Material and methods
Results
Discussion As WES of a patient with congenital scoliosis had identified a de novo missense mutation in the OLFML1 gene, this study sought to determine whether OLFML1 played a role in skeletal development. Our results in mice provide evidence that Olfml1 inhibits the Hippo signaling pathway in osteoblasts, thereby suppressing osteoblast mineralization. We therefore hypothesize the impact of an OLFML1 mutation as osteoblast differentiation impairment leading to abnormal development of bone tissue. YAP belongs to the group I of WW domain proteins, which are known to interact with various PPxY-, PPRxxP-, and PPPPP-motif-containing proteins such as p73 [21], Wbp2 [22], p73 [23], Runx2 [24,25], Smads [26], and Angiomotin family proteins [27,28]. In addition, YAP contains an SH3-binding motif as well as a PDZ-binding motif and binds to several molecules, including the porphyrin molecule verteporfin [29], the scaffold protein IQGAP1 [30], and the PDZ-scaffold protein NHERF1 [31]. In the current study, Olfml1 co-localized with YAP in the cytoplasm and suppressed its nuclear translocation. Motif analysis of murine Olfml1 showed that it lacked any known domains binding to YAP (Supplementary Fig. S1C). Thus, Olfml1 might bind to YAP by an unknown sequence or interact with one or more adapter proteins for YAP recruitment. Despite its pivotal roles in bone metabolism, the function of Hippo signaling in osteoblast differentiation is controversial. Knockdown of YAP and TAZ in human mesenchymal progenitors suppressed the expression of osteoblastic markers in vitro [32], and YAP and TAZ haploinsufficiency impaired osteoblast differentiation of murine osteoblast progenitors in vitro [33]. The deletion of YAP and TAZ in Osx-cre mice also decreased osteoblast number and bone mass [34]. Mechanistically, TAZ promoted osteoblast progenitor proliferation and differentiation by binding to TEAD and Runx2 [35]. Those data indicated a role of Hippo signaling as a positive regulator of osteoblast differentiation; however, several reports have concluded the opposite. Knockdown of YAP in murine bone marrow-derived mesenchymal cells [36] or deletion of YAP and TAZ in osteoprogenitor cells [37] increased osteoblast differentiation in vitro. Xiong et al. described that postnatal deletion of YAP and TAZ in Osx-cre mice resulted in increased osteoblast number [37]. YAP presumably inhibited Runx2 transcriptional activity [25], and cytoplasmic TAZ bound Axin and facilitated sequestering of β-catenin in the destruction complex, thereby promoting its proteosomal degradation and inhibiting Wnt signaling [38].