Loss of MMP-2 disrupts skeletal and craniofacial development and results in decreased bone mineralization, joint erosion and defects in osteoblast and osteoclast growth

Abstract

The 'vanishing bone' or inherited osteolysis/arthritis syndromes represent a heterogeneous group of skeletal disorders characterized by mineralization defects of affected bones and joints. Differing in anatomical distribution, severity and associated syndromic features, gene identification in each 'vanishing bone' disorder should provide unique insights into genetic/molecular pathways contributing to the overall control of skeletal growth and development. We previously described and then demonstrated that the novel autosomal recessive osteolysis/arthritis syndrome, multicentric osteolysis with arthritis (MOA) (MIM #605156), was caused by inactivating mutations in the MMP2 gene [Al Aqeel, A., Al Sewairi, W., Edress, B., Gorlin, R.J., Desnick, R.J. and Martignetti, J.A. (2000) Inherited multicentric osteolysis with arthritis: A variant resembling Torg syndrome in a Saudi family. Am. J. Med. Genet., 93, 11-18.]. These in vivo results were counterintuitive and unexpected since previous in vitro studies suggested that MMP-2 overexpression and increased activity, not deficiency, would result in the bone and joint features of MOA. The apparent lack of a murine model [Itoh, T., Ikeda, T., Gomi, H., Nakao, S., Suzuki, T. and Itohara, S. (1997) Unaltered secretion of beta-amyloid precursor protein in gelatinase A (matrix metalloproteinase 2)-deficient mice. J. Biol. Chem., 272, 22389-22392.] has hindered studies on disease pathogenesis and, more fundamentally, in addressing the paradox of how functional loss of a single proteolytic enzyme results in an apparent increase in bone loss. Here, we report that Mmp2-/- mice display attenuated features of human MOA including progressive loss of bone mineral density, articular cartilage destruction and abnormal long bone and craniofacial development. Moreover, these changes are associated with markedly and developmentally restricted decreases in osteoblast and osteoclast numbers in vivo. Mmp2-/- mice have approximately 50% fewer osteoblasts and osteoclasts than control littermates at 4 days of life but these differences have nearly resolved by 4 weeks of age. In addition, despite normal cell numbers in vivo at 8 weeks of life, Mmp2-/- bone marrow cells are unable to effectively support osteoblast and osteoclast growth and differentiation in culture. Targeted inhibition of MMP-2 using siRNA in human SaOS2 and murine MC3T3 osteoblast cell lines resulted in decreased cell proliferation rates. Taken together, our findings suggest that MMP-2 plays a direct role in early skeletal development and bone cell growth and proliferation. Thus, Mmp2-/- mice provide a valuable biological resource for studying the pathophysiological mechanisms underlying the human disease and defining the in vivo physiological role of MMP-2.

Figure 1

Mmp2^-/- mice have abnormal craniofacial development. (A) Frontal and dorsal views of 12 week old Mmp2^-/-, Mmp2^+/-, and Mmp2^+/+ mice. (B-D) Lateral, superior and inferior views, respectively, of μCT images of 6 month old Mmp2^-/-, Mmp2^+/-, and Mmp2^+/+ mice. (E) Tabulated μCT measurements between paired Mmp2^-/- and Mmp2^+/+ mice showing those differences reaching statistical significance (p<0.05).

Figure 2

Photomicrograph of femoral condyles and proximal tibiae from 12 week old Mmp2^+/+ and Mmp2^-/- mice. Normal femoral condyle structure in Mmp2^+/+ is shown in A, with a smooth articular cartilage surface. The condylar surface in Mmp2^-/- mice (B) is arthritic in appearance, with a roughened and irregular cartilage surface and eroded areas of cartilage replaced by fibrous tissue (inset box). Similarly, cartilage destruction is evident on the proximal tibial surface (C), with large erosions of cartilage occupied by fibrous tissue and inflammatory cells (inset box).

Figure 3

Bone changes resulting from MMP-2 deficiency. Long bone radiographs from 6 month old mice showing (A) slight shortening of long bones and lower density in Mmp2^-/- bone compared to Mmp2^+/+ bone; this lower apparent density was confirmed and shown to be progressive by DEXA-based bone mineral density (BMD) measurements (B). Photomicrograph of the long bones (tibiae) reveal poorly developed cancellous bone in the Mmp2^-/- metaphyses at 4 days (C, F). Cortical bone at both at the metaphyseal (D, G, enlargement from areas shown in C, F) and diaphyseal (E, H) regions was poorly consolidated in Mmp2^-/- bones; the knockout bone was also more woven in appearance and showed decreased numbers of osteocytes compared to Mmp2^+/+ cortical bone. The major cortical bone defect resolved by 4 and 12 weeks of age; however, the Mmp2^-/- bone continued to demonstrate large areas of woven bone (polarized light photomicrographs, I, J) compared to the normal lamellar tissue in seen Mmp2^+/+ bone.

Figure 4

Osteoblast, osteoclast, and proliferating cell numbers are decreased in Mmp2^-/- early postnatal bones. (A) Low power micrograph showing distal femur head (fh), growth plate (gp), and metaphysis (met) of Mmp2+/+ 4 day old mouse. Box indicates approximate region of high power micrographs (B-G). (B-G) High power micrographs of Mmp2^+/+ (B,D,F) and Mmp2^-/-(C,E,G) 4 day old mouse femurs stained with procollagen I as a marker of osteoblasts (B,C), cathepsin K for osteoclasts (D,E), and PCNA (F,G) show reduced number of positive cells in the Mmp2^-/- bone. (H) Quantification of procollagen I, cathepsin K, or PCNA positive cells per bone surface area; (*p<0.01; **p<0.005).

Figure 5

MMP-2 deficiency results in decreased osteoblast- and osteoclastogenesis ex vivo. (A-B) Mmp2^-/- bone marrow derived alkaline phosphatase positive colonies are fewer, smaller, and sparser than those derived from Mmp2^+/+ BMC under either (A) normal or (B) reduced stringency washing conditions. (C) Quantification of low stringency colony formation assays. (D,E) Mmp2^+/+ and Mmp2^-/- calvarial cells six days after plating equal cell numbers show decreased proliferation of Mmp2^-/- calvarial cells. (F,G) Representative low power micrograph of Mmp2^+/+ (F) and Mmp2^-/- (G) splenic multinucleated, TRAP positive osteoclasts demonstrates reduction of osteoclastogenic efficiency from Mmp2^-/- cells. Inset high power micrograph illustrates multi-nucleation. (H) Mmp2^-/- BMC and splenic cells produce ∼65% fewer osteoclasts. Osteoclast numbers reflect averages of the total numbers of multi-nucleated, TRAP positive cells present in six P35 dishes. * p<0.005.

Figure 6

MMP-2 inhibition by siRNA decreases proliferation of osteoblast cell lines. (A & C) MMP-2 mRNA expression levels in (A) SaOS2 and (D) MC3T3 cells transfected with pools of siRNA targeting MMP-2 (siMMP-2) are reduced by up to 80% compared to cells transfected with an unrelated siRNA control (siNTC). (B&E) Conditioned media MMP-2 activity levels are reduced by greater then 50% in (B) SaOS2 and (E) MC3T3 cells transfected with siMMP-2. (C & F) Proliferation rates as measured by ³H thymidine incorporation are reduced in (C) SaOS2 and (F) MC3T3 cells transfected with siMMP-2 compared to those transfected with siNTC. *p<0.05. (G) Proliferation rates are significantly reduced in SaOS2 clones stably expressing siMMP2 (A, B) compared to control clones infected with empty-vector-containing virus (E), regardless of the addition of exogenous MMP2 in the culture media. *p<0.05.