A mouse model of chondrocyte-specific somatic mutation reveals a role for Ext1 loss of heterozygosity in multiple hereditary exostoses

Abstract

Multiple hereditary exostoses (MHE) is one of the most common skeletal dysplasias, exhibiting the formation of multiple cartilage-capped bony protrusions (osteochondroma) and characteristic bone deformities. Individuals with MHE carry heterozygous loss-of-function mutations in Ext1 or Ext2, genes which together encode an enzyme essential for heparan sulfate synthesis. Despite the identification of causative genes, the pathogenesis of MHE remains unclear, especially with regard to whether osteochondroma results from loss of heterozygosity of the Ext genes. Hampering elucidation of the pathogenic mechanism of MHE, both Ext1(+/-) and Ext2(+/-) heterozygous mutant mice, which mimic the genetic status of human MHE, are highly resistant to osteochondroma formation, especially in long bones. To address these issues, we created a mouse model in which Ext1 is stochastically inactivated in a chondrocyte-specific manner. We show that these mice develop multiple osteochondromas and characteristic bone deformities in a pattern and a frequency that are almost identical to those of human MHE, suggesting a role for Ext1 LOH in MHE. Surprisingly, however, genotyping and fate mapping analyses reveal that chondrocytes constituting osteochondromas are mixtures of mutant and wild-type cells. Moreover, osteochondromas do not possess many typical neoplastic properties. Together, our results suggest that inactivation of Ext1 in a small fraction of chondrocytes is sufficient for the development of osteochondromas and other skeletal defects associated with MHE. Because the observed osteochondromas in our mouse model do not arise from clonal growth of chondrocytes, they cannot be considered true neoplasms.

Fig. 1.

Skeletal phenotype of Ext1-SKO mice. (A) X-ray images of P28 Ext1-SKO mice show multiple bony protrusions (arrows) in the long bones and the rib cage. (B) Macroscopic views of bony protrusions in the wrist joint and the rib cage of P28 Ext1-SKO mice (arrows). (C) Safranin O/Fast Green-stained sections of exostosis-like lesions developed in P28 Ext1-SKO mice. Note that these bony protrusions have a cartilage cap with the growth plate-like tissue structure, reminiscent to osteochondromas in human MHE. (D and E) Ext1-SKO mice mimic skeletal deformities seen in human MHE. (D) Bowing of the radius (arrow) and the subluxation/dislocation of its head (arrowheads). (E) Scoliosis of the thoracic vertebral column (arrowheads).

Fig. 2.

Col2-Cre^ERT transgene without induction with tamoxifen drives stochastic recombination of floxed alleles in a small fraction of chondrocytes. (A) Time course of Col2-Cre^ERT-mediated recombination under noninduced conditions. R26R;Col2-Cre^ERT mice were raised without tamoxifen treatment, and sections through the growth plate of the femur and tibia were stained with X-gal at indicated time points. Note that the frequency of lacZ⁺ cells is low, and that lacZ⁺ cells form clusters corresponding to the columnar organization of chondrocytes in the growth plate. See also Fig. S2 for a similar analysis in whole-mount skeletons. (B) Tissue specificity of Ext1 recombination driven by Col2-Cre^ERT without tamoxifen treatment. Genomic DNA from indicated tissues of P28 Ext1-SKO mice was analyzed by PCR specific for the recombined Ext1^ΔF allele. (C) Quantitative PCR analysis of recombination efficiency. The amounts of the recombined Ext1^ΔF allele and intact Ext1^F allele were determined by quantitative PCR and represented as the percentage of Ext1^ΔF alleles (black bars) to total Ext1 alleles. Genomic DNA samples from proximal tibial cartilage and rib cartilage of P28 Ext1-SKO mice were examined. The results from two quantitative controls (kidney DNA from Ext1-SKO mice, which should contain no Ext1^ΔF allele, and brain DNA from Ext1^+/− heterozygous mice, which is expected to give a 1:1 ratio) verify approximate linearity of the assay. All results are based on duplicate samples from three independent mice in each genotype.

Fig. 3.

Development of osteochondromas in Ext1-SKO mice. (A) Whole-mount skeletal preparations in the wrist joint area. (B) Sections through the wrist joint area. (C) Sections through the knee joint area. Note that at P7, no apparent abnormalities are observed in Ext1-SKO mice either by macroscopic (A) or microscopic (B and C) analysis. The first detectable abnormalities in the wrist joint area are the overgrowth of the growth plate cartilage (arrowheads in A and B) and the formation of microscopic exostoses (arrows in B) at P14. These abnormalities become more prominent at P21, and by P28, multiple bony protrusions with a cartilage cap are formed in the wrist joint area (A). Osteochondroma formation in the knee joint area takes a similar time course (C).

Fig. 4.

Histological characterization of osteochondromas in Ext1-SKO mice. (A) Analysis of the tissue organization of osteochondromas by in situ hybridization for Col2a1, Ihh, and Col10. Adjacent sections were stained with Safranin O and Fast Green (SafO/FG). (B) Analysis of cell proliferation. Sections of osteochondromas were double-stained with anti-phosphorylated histone H3 (pHH3) and DAPI. Quantitative analysis of pHH3⁺ cells (bar graph) shows that there is no significant difference in the frequency of replicating cells between the cartilage cap region of osteochondromas and the normal growth plate. GP, growth plate; OC, osteochondroma.

Fig. 5.

Chondrocytes comprising osteochondromas are mixtures of mutant and wild-type cells. (A) Analysis of the genotype of chondrocytes in osteochondromas by laser-capture microdissection (LCM). Cartilage cap regions of osteochondromas were isolated by LCM, and genomic DNA specimens from these tissue fragments were analyzed by allele-specific PCR. (Upper Left) Representative result of LCM. (Upper Right) Result of the PCR analysis of DNA from nine independent osteochondromas. As controls, DNA from cartilage of Ext1^+/− mice (Ext1^+/−) and that from cartilage of Ext1^F/F mice (Ext1^F/F) were also analyzed. Note that all osteochondromas tested contain both the recombined Ext1^ΔF and intact Ext1^F alleles. (Lower) Ratio of the Ext1^ΔF and Ext1^F alleles estimated by semiquantitative PCR in the 15 osteochondromas developed in the femur, radius, and ulna. Note that the percentage of the Ext1^ΔF allele is highly variable, ranging from 9.6% (#14) to 52.6% (#1). (B) Fate mapping analysis of chondrocytes comprising osteochondromas. To evaluate the contribution of Ext1 null chondrocytes to osteochondromas, Col2-Cre^ERT;Ext1^F/F mice carrying the R26R reporter gene were bred, and osteochondromas developed in these mice were stained with X-gal to identify cells that had undergone Cre-mediated recombination. A gallery of four representative results is shown at low and high magnifications. Osteochondromas indicated by arrowheads in Left are shown in Right at high magnification. Note that none of the osteochondromas examined were entirely composed of cells that undergone recombination. The contribution of recombined cells in osteochondromas is highly variable.