BoneKEy-Osteovision | Commentary

And then they were two… progressing along the osteoblast differentiation pathway



DOI:10.1138/2002029

The hallmark and the strength of the fields of adipogenesis, myogenesis or hematopoiesis has been the existence of a known cascade of transcription factors acting as successive differentiation triggers to progressively specify cell lineages. Notwithstanding the importance of delineating molecular interactions between transcription factors or between transcription factors and cell-specific structural genes, the knowledge of such cascades has allowed researchers to pinpoint accurately discrete stages along the differentiation process of a cell lineage. With the identification of Osx as a second transcription factor that is required for osteoblast differentiation during development and acts downstream of Cbfa1 (OMIM 600211) (), bone biologists can better characterize genetically the origin, expression particularities and fate of cells progressing toward a mature osteoblastic phenotype (Fig. 1). Of course, two factors is still a short list and drawing the genetic map of skeletogenesis still involves many question marks and hypotheses, but on the positive side of things, we can now count on two, instead of one, “entry points” on which to base our future investigations.

Fig. 1: A model of osteoblast and chondrocyte specification during skeletogenesis. The mesenchymal cells grouped in skeletal condensations first differentiate into type II collagen (Col II) producing cells. Cbfa1 then promotes their differentiation into Ihh-secreting hypertrophic chondrocytes or into type I collagen (Col I) producing pre-osteoblasts. The absence of Cbfa1 therefore blocks differentiation in both the osteoblastic and chondrocytic lineages (orange crosses) and maintains the cells present in the skeletal condensations in a chondrocyte-like stage (orange arrows). Osx acts downstream of Cbfa1 to further differentiate osteoblast progenitors into mature cells. Osx, therefore promotes the expression of early osteoblast marker genes such as bone sialoprotein and osteonectin. Osx does not seem to play a role during chondrogenesis, as its absence does not alter cell differentiation along this pathway. In contrast, Osx deficiency (green cross) arrests osteoblast differentiation, forcing progenitor cells to redirect their fate toward a chondrocyte phenotype (green arrow).

Using differential screening between C2C12 myoblastic cells treated or not by BMP-2 (OMIM 112261), Nakashima et al. isolated a cDNA clone encoding a zinc-finger domain containing protein they termed Osterix (Osx) (OMIM 606633). An expression pattern exquisitely restricted to skeletal cells during development as well as the ability of Osx to induce type I collagen and osteocalcin expression when overexpressed in two non-osteoblastic cell lines prompted them to delete the Osx gene in mice. This identification of Osx and the characterization of the Osx-deficient phenotype was reported in the January 11th issue of Cell (). As occurred in the absence of Cbfa1, Osx-null mice die at birth from an inability of their rib cage to sustain breathing. As was also the case for the Cbfa1-deficient mice, Osx-null mice harbor a perfectly patterned skeleton, although entirely composed of cartilage, and present a general lack of osteoblasts. This is demonstrated by the absence of expression of osteoblast-specific genes such as bone sialoprotein, osteonectin and osteocalcin (). These striking similarities between the Osx- and the Cbfa1-deficient mice raised the question of the relative genetic positioning of the two transcription factors. To address this critical point, expression studies were performed in the respective deficient mouse models. Whereas Cbfa1 was expressed normally in Osx-deficient mice, the expression of Osx was abolished in Cbfa1-deficient mice, indicating that Osx is genetically located downstream of Cbfa1. Does this mean that Cbfa1 directly controls Osx expression? Most likely but not necessarily. Intermediate factor(s) still to be identified may create the link between the two regulators.

Two major differences exist, however, between the consequences of Cbfa1-deficiency and Osx-deficiency. First, although Osx is expressed weakly and transiently in chondrocytes, its absence does not cause any obvious cartilaginous defect. This result is essential as it indicates that Osx functions strictly within the osteoblast differentiation pathway. The second difference resides in the possibility for Osx-null cells present in the intramembranous bones to evolve toward a chondrocyte fate since they cannot differentiate into osteoblasts. This result reinforces the hypothesis that osteoblasts and chondrocytes share a common progenitor () but also enlarges it to the possibility that Cbfa1 could be a moving force toward either lineage, while the presence of Osx would be decisive for an osteoblastic fate. As such, differentiation along the chondrocyte lineage could be interpreted as a default pathway caused physiologically by the absence of activation or specific inhibition of Osx expression.

A point that the analysis of the Osx-deficient mice also strengthens is the possibility for TRAP-positive (OMIM 190445) multinucleated resorbing cells to differentiate and to function in the complete absence of osteoblasts. Indeed, the long bones of the Osx-deficient mice are bent, as were the long bones of the Cbfa1-deficient mice partially rescued by restoration of Cbfa1 expression in chondrocytes (), due to the existence of a resorptive process eroding the cartilaginous template. Considering that the resorbing cells present in Osx-null long bones exhibit MMP-9 (OMIM 120361) and cathepsin K (OMIM 601105) in addition to TRAP expression, Nakashima et al. propose that they are fully functional osteoclasts. Such observations suggest that the expression of the factors regulating osteoclast differentiation is not restricted to the cells of the osteoblastic lineage.

The absence of a known skeletal disorder mapping at the Osx locus (Sp7 locus, human chromosome 12q13), as well as the absence of phenotypic abnormalities in heterozygous Osx-deficient mice, contrasts with the Cleidocranial dysplasia characterizing Cbfa1 haploinsufficiency in mice and patients (). Based on the uniqueness of its amino acid sequence outside the Zinc finger region, Nakashima et al. dismiss the hypothesis of a compensatory mechanism by a structurally related factor. However, one could hypothesize that other factors, functionally if not structurally related, could compensate for a partial loss of Osx. Another possibility would be that regulatory mechanisms could upregulate the expression of Osx and thereby re-establish its normal level in the heterozygous deficient mice. Lastly, Osx could be such a potent factor that only minor amounts of it would be sufficient to fulfill its function. Additional analyses of Osx biology should answer these questions. One can also hope that such analyses will shortly delineate an Osx-specific binding site and identify its target genes.


Creative Commons License This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 United States License.