Correction DEVELOPMENTAL BIOLOGY Correction for “Temporomandibular joint formation requires two distinct hedgehog-dependent steps,” by Patricia Purcell, Brian W. Joo, Jimmy K. Hu, Pamela V. Tran, Monica L. Calicchio, Daniel J. O’Connell, Richard L. Maas, and Clifford J. Tabin, which ap- peared in issue 43, October 27, 2009, of Proc Natl Acad Sci USA (106:18297–18302; first published October 8, 2009; 10.1073/ pnas.0908836106). The authors note the following statement should be added to the Acknowledgments: “Additional support was obtained from NIH/NIDCR Grant K12-DE14528, Dean’s Scholar, and Aina M. Auskaps Fellowships from the Harvard School of Dental Medi- cine (all to P.P.).” www.pnas.org/cgi/doi/10.1073/pnas.1000188107 3942 | PNAS | February 23, 2010 | vol. 107 | no. 8 www.pnas.org Downloaded by guest on September 28, 2021 Temporomandibular joint formation requires two distinct hedgehog-dependent steps Patricia Purcella,b, Brian W. Jooa,b, Jimmy K. Hua, Pamela V. Tranb, Monica L. Calicchioc, Daniel J. O’Connellb, Richard L. Maasb, and Clifford J. Tabina,1 aDepartment of Genetics, Harvard Medical School, Boston, MA 02115; bDivision of Genetics, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA 02115; and cDepartment of Pathology, Children’s Hospital Boston, Boston, MA 02115 Contributed by Clifford J. Tabin, August 10, 2009 (sent for review July 17, 2009) We conducted a genetic analysis of the developing temporo-man- develop by the cleavage or segmentation of a continuous skeletal dibular or temporomandi-bular joint (TMJ), a highly specialized sy- condensation (2–4). The first morphological sign of joint formation novial joint that permits movement and function of the mammalian is the appearance of a transverse stripe of cells, the interzone, a jaw. First, we used laser capture microdissection to perform a ge- three-layered region with reduced cell density in the center that nome-wide expression analysis of each of its developing components. marks the area destined to become the joint space (5, 6). This The expression patterns of genes identified in this screen were morphological change is presaged by molecular events, including examined in the TMJ and compared with those of other synovial the down-regulation of several genes expressed in the remainder of joints, including the shoulder and the hip joints. Striking differences the developing cartilage, such as Sox9, a member of the Sox family were noted, indicating that the TMJ forms via a distinct molecular of transcription factors present in all chondroprogenitor cells (7, 8). program. Several components of the hedgehog (Hh) signaling path- Conversely, a large number of genes are induced specifically in the way are among the genes identified in the screen, including Gli2, location of the future joint. Prominent among these are Wnt9a which is expressed specifically in the condyle and in the disk of the (formerly called Wnt14), a canonical Wnt ligand (9–11), and developing TMJ. We found that mice deficient in Gli2 display aberrant depending on the specific joint, Gdf5, Gdf6,orGdf7, members of TMJ development such that the condyle loses its growth-plate-like the BMP/TGF␤ superfamily (12–15). Strikingly, these genes are cellular organization and no disk is formed. In addition, we used a expressed during and act in the formation of joints that form conditional strategy to remove Smo, a positive effector of the Hh between the long bones by segmentation and in other classes of signaling pathway, from chondrocyte progenitors. This cell autono- joints such as those between vertebrae and those between calvarial mous loss of Hh signaling allows for disk formation, but the resulting membranous bones. Different members of the Gdf family are structure fails to separate from the condyle. Thus, these experiments expressed in diverse joints, and the loss of Gdf activity results in the establish that Hh signaling acts at two distinct steps in disk morpho- failure of joint formation (16). In addition, ␤-catenin activity, a key genesis, condyle initiation, and disk–condyle separation and provide effector of the canonical Wnt pathway, is required for joint forma- a molecular framework for future studies of the TMJ. tion, and ectopic Wnt9a is sufficient to initiate the formation of a joint interzone (9, 11, 17). Indian hedgehog ͉ Gli2 ͉ synovial joints ͉ microarray The condyle is an important growth site in the mandible with similarities to the growth plate of the long bones, and it displays four he temporomandibular joint (TMJ) is a complex structure that distinct zones: a fibrous cell layer, a progenitor cell layer, a zone of Tis essential for jaw movement and found only in mammals. Its flattened chondrocytes, and a zone of hypertrophic chondrocytes major components include the glenoid fossa of the temporal bone, (Fig. 1, upper right) (18, 19). One key gene previously noted to be the condylar head of the mandible, and a fibrocartilaginous disk that expressed during and function within the growth plate of the is located between these bones, dividing the joint cavity into two condylar cartilage is Indian hedgehog (Ihh) (20–22). Ihh has been compartments. Both the condyle and the glenoid fossa are endo- studied extensively during endochondral ossification of the long chondral in origin. The first evidence of TMJ formation during bones, where it plays several distinct roles. Secreted by prehyper- development is the appearance of distinct mesenchymal conden- trophic condrocytes that are just entering the differentiation path- sations, the temporal and condylar blastemas. The condylar blast- way, Ihh is critical for maintaining the growth of adjacent prolif- ema rapidly grows toward the temporal blastema, closing the gap erating chondrocytes. In addition, Ihh plays an indirect role in between them while a distinct articular disk forms within the joint regulating the rate of chondrocyte differentiation by acting in a negative feedback loop with a second secreted protein, parathyroid- as a separate condensation (1). hormone-related protein (PTHrP), in the periarticular perichon- The TMJ differs from most synovial joints in several ways. First, drium. Chondrocytes within the range of PTHrP signaling are in the TMJ forms by appositional growth, as opposed to segmentation turn blocked from entering the differentiation pathway. Thus, Ihh, of a continuous skeletal condensation. Second, in the TMJ, the in conjunction with PTHrP, plays a crucial role in organizing the articular surfaces of the condyle and glenoid fossa are covered by growth plate (23–25). In potentially analogous fashion, in the a layer of fibrous rather than hyaline cartilage. Last, the two bones absence of Ihh, the organization of the growth-plate-like zone in the are in contact with an intervening fibrocartilaginous disk rather TMJ condyle is disrupted, and the TMJ disk does not form (21). Ihh than articulating with each other directly. The development of the signals through its receptor Ptc1, itself a transcriptional target of TMJ during prenatal life also lags behind other joints in both the Ihh. Acting through a second transmembrane protein, Smo, Ihh BIOLOGY time of its initiation and its development. In the mouse, all of the activity serves to regulate the processing and activity of the Gli DEVELOPMENTAL major anatomical features of the TMJ, including the disk, are family of transcription factors (26). Gli1 itself is transcriptionally present by E16.5, although the condyle and glenoid fossa continue up-regulated by Ihh signaling and is a transcriptional activator of to increase in size and density into adulthood. Although the structural features of the TMJ are well docu- mented, little information is available with respect to the genetic, Author contributions: P.P. and C.J.T. designed research; P.P., B.W.J., and M.L.C. performed cellular, and molecular mechanisms involved in TMJ morphogen- research; J.K.H. and P.V.T. contributed new reagents/analytic tools; P.P., D.J.O., R.L.M., and esis. In contrast, studies of other skeletal elements, most notably of C.J.T. analyzed data; and P.P. and C.J.T. wrote the paper. the developing limb, have provided a wealth of information about The authors declare no conflict of interest. signals involved in synovial joint formation. Most synovial joints 1To whom correspondence should be addressed. E-mail: [email protected]. www.pnas.org͞cgi͞doi͞10.1073͞pnas.0908836106 PNAS ͉ October 27, 2009 ͉ vol. 106 ͉ no. 43 ͉ 18297–18302 Fig. 1. RNA expression of Fgfr1, Fgfr2, Fgfr3, Sox5, Sox6, Sox9, Zfp445, Wnt6, and BMP7 in mouse embryonic tem- poromandibular joint (TMJ). (A–I) In situ hybridization was performed on serial coronal cryosections of E16.5 TMJ. At this developmental stage, the mandibular condyle, tempo- ral fossa, and TMJ disk are morphologically well defined. (Upper right) Different cell zones in the condyle. (Lower right) An in situ hybridization with Collagen X (ColX), a marker for hypertrophic chondrocytes. f, glenoid fossa of the temporal bone; c, condylar head of the mandible; d, joint disk; pe, perichondrium; fl, fibrous cell layer; pl, pro- genitor cell layer; fc, flattened chondrocytes; hc, hypertro- phic chondrocytes. Ihh targets. Gli3 acts predominantly as a transcriptional repressor and spatially restricted patterns. Fgfr1 was expressed in the perios- and is down-regulated transcriptionally by Ihh signaling. Moreover, teum of the condyle and fossa, Fgfr2 in the perichondrium of the Ihh activity alters the processing of full-length Gli3 protein such that condyle and fossa, and Fgfr3 in the immature chondrocytes of the the repressor form is not produced. Finally, Gli2 functions mostly condyle [Fig. 1 A–C (32)]. The transcription factor Sox9 was highly as a transcriptional activator, although in the absence of Ihh activity expressed in proliferating chondrocytes in the condyle at E16.5 and it retains some repressor function (27, 28). Functional Gli2 is was down-regulated in hypertrophic chondrocytes (Fig.