Involvement of type XII collagen in bone formation. (A–F) Type XII collagen is localized to bone-forming regions. Reactivity for type XII collagen is localized to the periosteum (PO), endosteum (EO), and trabecular bone (TB). Immunofluorescence analysis of the mid-diaphysis in cross (A and B) and longitudinal (D) sections of the femur and calvaria (E and F) from P30 wild type (Col12a1^+/+) and the mid-diaphysis in cross sections of the femur in type XII collagen–null (Col12a1^−/−) mice (C). Green, anti–type XII collagen; blue, nuclei reactive with DAPI; CB, cortical bone; BM, bone marrow; OB, osteoblast. (G) A schematic representation of the wild-type Col12a1 exons 1–8 and target locus to create type XII collagen–null mice. After homologous recombination, the exon 2–5 locus was replaced by the selection cassette that includes the phosphoglycerolkinase (PGK) neomycin-polyadenylation (NEO polyA) signal. (H) Analysis of type XII collagen expression in tissues from wild-type and Col12a1^−/− mice by immunoblotting. FDL, flexor digitorum longus. (I) X-ray and skeletal staining with alizarin red of 5-mo-old Col12a1^+/+ and Col12a1^−/− mice. Col12a1^−/− mice demonstrated kyphosis (asterisk) in the x-ray image. Abnormal double spinous processes (arrows) were present in the vertebral column of Col12a1^−/− mice visualized by skeletal staining. L3, third lumbar. (J–L) Body weight, femur length, and percentage of tissue area/length were measured in P30 wild-type (control; n = 9) and type XII collagen–null (n = 10) mice. (J) Body weight is decreased in Col12a1^−/− mice. (K) Col12a1^−/− mice have shorter femurs compared with wild-type mice when normalized to body weight. (L) The percentage of tissue area/length normalized to body weight revealed that Col12a^−/− mice have more slender femurs compared with wild-type controls. **, P < 0.01. Error bars are mean ± SD. Bars: (A, D, and E) 25 µm; (B, C, and F) 250 µm; (I, left) 2 cm; (I, right) 1 mm.

Decreased bone formation in type XII collagen–null mice. (A) Decreased mineralized nodule formation was observed in bone marrow cells derived from P30 Col12a1^−/− mice (n = 6) compared with wild-type controls (n = 5). Cells were cultured in osteogenic medium for 14, 21, and 28 d. Mineralized nodules represented by alizarin red staining are detected in wild-type cells after 14 d in culture, whereas nodules are virtually absent from Col12a1^−/− cells under the same conditions. At day 21, although nodules are formed in the Col12a1^−/− bone marrow cells, the nodule area is smaller than that in wild-type cultures. Similar nodule areas are detected at 28 d in culture. The mineralized nodule area is significantly decreased in Col12a1^−/− bone marrow cells at both day 14 and 21 in culture compared with wild-type controls. At day 28, the areas are comparable, indicating a delay in the absence of type XII collagen. (B–E) Bone morphometry in the periosteum of cross sections of tibia from P30 Col12a1^−/− (n = 5) and wild-type control (n = 4) mice was conducted using calcein double labeling. (B) Calcein label is detected as double lines (arrows). The boxed areas on the left are shown at higher magnification on the right. The distance between labels is decreased in Col12a1^−/− bone compared with wild-type control. (C) Mineral apposition rate is reduced ∼30% in Col12a1^−/− mice compared with wild-type controls. (D and E) The type XII collagen–null bones have ∼40% of the percentage of mineralized bone surface/total bone surface (D) and ∼50% of the bone formation rate (E) in Col12a1^−/− mice compared with wild-type controls. All of the differences are significant. *, P < 0.05; **, P < 0.01. Error bars are mean ± SD. Bars, 100 µm.

Disorganized bone matrix collagen arrangement in type XII collagen–null mice. (A) Osteoblast morphology was analyzed in cross sections of the diaphysis from femurs of P30 Col12a1^+/+ and Col12a1^−/− mice with H&E staining. Osteoblasts in wild-type bone are localized on the surface of cortical bone (CB), where they are aligned and well organized in an epithelioid manner (arrows). In contrast, Col12a1^−/− osteoblasts (arrows) are poorly aligned and crowded in the periosteum (PO) with a disorganized arrangement associated with the surface of cortical bone. BM, bone marrow. (B) Collagen organization was analyzed in cross sections of the diaphysis from femurs of P14 Col12a1^+/+ and Col12a1^−/− mice. To analyze collagen organization, Sirius red staining with polarized light microscopy was performed. Collagen fibers (arrows) are arranged in parallel and are well organized in the wild-type cortical bone matrix, whereas birefringence is sparse in Col12a1^−/− mice, indicating smaller and/or less organized collagen fibers (asterisks) in the cortical bone. Bars, 50 µm.

Altered cell–cell interaction/communication in type XII collagen–null bone in vivo. Analysis of cross sections from the diaphysis of femurs in P14 Col12a1^+/+ and Col12a1^−/− mice using transmission EM. (A) Wild-type osteoblasts (OB) attached to the surface of cortical bone (CB) are elongated and polarized relative to the bone matrix. In contrast, Col12a1^−/− osteoblasts (arrows) are disorganized and less polarized. PO, periosteum. (B and C) High magnification of adjacent osteoblasts in the periosteum. In wild-type bone, osteoblasts have numerous processes (double arrows) that connect osteoblast to osteoblast. Col12a1^−/− osteoblasts have large intracellular spaces (asterisks) with fewer processes (double arrows). (D) Osteocyte within cortical bone matrix. Similar to the osteoblast phenotype, an osteocyte has processes (double arrows) that extend toward the cortical bone surface in wild-type bone connecting to other osteocytes or osteoblasts; however, Col12a1^−/− osteocytes have fewer and less developed processes (double arrows). Bars: (A) 5 µm; (B) 1 µm; (C) 500 nm; (D) 2 µm.

Model for role of type XII collagen in osteoblast development and bone formation. Type XII collagen surrounding immature osteoblasts responds to mechanical stimuli that initiate osteoblast maturation. This is associated with increased osteoblast organization, cell–cell interaction, and polarization along bone deposition interfaces. Further localized production of type XII collagen by well-organized polar osteoblasts reinforces the osteoblast organization and maturation. Cell–cell interaction drives Cx43 expression, thereby establishing cell communication via gap junctions, resulting in terminal differentiation, increased bone matrix protein production, and organized deposition of a functional bone matrix. In the absence of type XII collagen, there is a dysfunctional response to the mechanical stimuli and an alteration in the osteoblast maturation sequence. This defect delays/alters osteoblast organization, polarity, and cell–cell contacts, and, therefore, the downstream pathways, including up-regulation of Cx43 and bone matrix proteins, are compromised. Together, this results in decreased bone formation and quality.

Reduction of cortical bone mass and mechanical properties in Col12a1^−/− long bones. (A) Col12a1^−/− mice have smaller diameters and less/thinner cortical bone compared with wild-type controls. 3D micro-CT images from the midshaft of a femur from P30 Col12a1^+/+ and Col12a1^−/− mice. Bars, 0.5 mm. (B) Relative cortical area is significantly decreased in Col12a1^−/− femurs. The relative cortical area was measured by 3D micro-CT analysis as the percentage of mineralized cortical bone area/mineralized tissue bone area normalized to body weight in P30 Col12a1^+/+ (n = 9) and Col12a1^−/− (n = 10) mice. **, P < 0.01. Error bars are mean ± SD. (C–E) The biomechanical properties of the Col12a1^−/− femurs were significantly reduced when compared with wild-type controls. Maximum to load failure (C), bone robustness (D), and bone stiffness (E) were measured in femurs of both Col12a1^+/+ (n = 9) and Col12a1^−/− (n = 10) P30 mice by loading to failure in four-point bending at 0.05 mm/s and were normalized to body weight. Maximum load to failure, bone robustness, and bone stiffness are significantly decreased in Col12a1^−/− mice. TtAr/Le, total cross-sectional area/bone length; N/mm, Newtons per millimeter. P < 0.01.

Delayed terminal differentiation in type XII collagen–null osteoblasts. (A–E) Primary osteoblasts were derived from neonatal calvaria of Col12a1^+/+ and Col12a1^−/− mice and cultured in osteogenic medium. The cells are harvested after 0, 7, 14, and 21 d in culture and used for quantitative real-time PCR analysis. Each group represents the mean of triplicate determinations. mRNA was normalized to Gapdh expression. (A) Type XII collagen mRNA expression is stable during osteoblast development. (B and C) Similar expression patterns are detected in mRNA expressions of Runx2 (B) and type I collagen (C) in Col12a1^+/+ and Col12a1^−/− osteoblasts. (D and E) The expression of osteocalcin (Ocn; D) and osteopontin (Opn; E), which are terminal osteoblast differentiation markers, is significantly decreased in Col12a1^−/− osteoblasts as compared with wild-type controls. **, P < 0.01. Error bars are mean ± SD. (F) Immunoblot analysis of P30 femurs from Col12a1^+/+ (n = 2) and Col12a1^−/− (n = 3) mice was performed to validate the in vitro data. Opn and Ocn expression is decreased, whereas type I expression is stable in Col12a1^−/− femurs compared with the mRNA data.

Defective organization and polarization of osteoblasts in type XII collagen–null bone. Osteoblast polarization was analyzed by the distribution of nucleus, Golgi apparatus, and F-actin staining with confocal fluorescence microscopy. (A) Wild-type osteoblasts (OB) in an epithelioid arrangement are well organized into a layer between the periosteum (PO) and cortical bone (CB), and F-actin fibers (green) are highly enriched in regions of the osteoblasts, which face the periosteum and cortical bone (white arrows). (B) In contrast, Col12a1^−/− osteoblasts are not well organized between the periosteum and cortical bone, and F-actin fibers appear to form aggregates distributed randomly within osteoblasts (arrows). (C) Wild-type osteoblasts exhibit a polarized distribution of the Golgi apparatus (red) and the nucleus (blue). Most Golgi reactivity is localized toward the bone matrix (arrows), whereas the nuclei were positioned to the opposite side. (D) Col12a1^−/− osteoblasts also fail to show a polarized distribution of Golgi and nucleus. It is noticeable that the F-actin fibers (yellow arrows) extend to the cortical bone matrix, establishing a network throughout the bone matrix in wild-type bone. This network is not as extensive in Col12a1^−/− bone. Bars: (A and B) 20 µm; (C and D) 10 µm.

Impaired cell–cell communications via gap junction associated with altered Cx43 expression in type XII collagen–null osteoblasts. (A–C) Analysis of Cx43 (Cx43; A), Cad2 (Cad2; B), and Cad11 (Cad11; C) mRNA expression in cultures of primary osteoblasts by quantitative real-time PCR normalized to Gapdh expression. Each group represents the mean of triplicate determinations. (A) Cx43 expression is significantly decreased at every time point in Col12a1^−/− osteoblast cultures. (B) No significant difference between genotypes is detected in Cad2 expression. (C) Cad11 expression is decreased in early stages in Col12a1^−/− osteoblast cultures but not in late stages. **, P < 0.01. (D) Immunoblot analysis of Cx43 in cultures of primary osteoblasts. β-Actin is used as an internal control. (E) The relative band density of Cx43 normalized to β-actin. Each group represents the mean of triplicate determinations. Cx43 expression is decreased in Col12a1^−/− osteoblasts. (F) Immunoblot analysis of Cx43, Cad2, and Cad11 in P30 femurs from Col12a1^+/+ (n = 2) and Col12a1^−/− (n = 3) mice. Cx43 and Cad11 expression is decreased in Col12a1^−/− femurs. Molecular mass is indicated in kilodaltons. (G) Dye coupling study in primary osteoblasts. Calcein AM– and Cell Tracker–labeled osteoblasts (DCC, dye-carrying cells) were seeded onto confluent osteoblasts (host cells) and analyzed by confocal microscopy. The dashed lines indicate the outline of host cells. High magnification of the site of dye-carrying wild-type osteoblasts (+/+) attaching to the wild-type host osteoblasts demonstrates transfer, i.e., the green dots (arrows) to the host cell. Although the Col12a1^−/− osteoblast is attaching to the Col12a1^−/− host cell, calcein dye stays within the donor cell (i.e., no transfer is observed). Bars: (top) 25 µm; (bottom) 10 µm. (H) The number of dye-transferred cells per total parachuted cells was determined in 10 random images acquired from each well (24-well plate), and 6 wells were used per experimental group. As compared with dye-carrying wild-type cells (+/+) on wild-type host cells, the percentage of dye-transferred cells is significantly decreased when either dye-carrying or host cells are Col12a1^−/−. A reduction of ∼40% is detected when compared with wild-type versus wild-type and Col12a1^−/− versus Col12a1^−/− cells. *, P < 0.05; **, P < 0.01. Error bars are mean ± SD.