FAK-Mediated mechanotransduction in skeletal regeneration

Abstract

The majority of cells are equipped to detect and decipher physical stimuli, and then react to these stimuli in a cell type-specific manner. Ultimately, these cellular behaviors are synchronized to produce a tissue response, but how this is achieved remains enigmatic. Here, we investigated the genetic basis for mechanotransduction using the bone marrow as a model system. We found that physical stimuli produced a pattern of principal strain that precisely corresponded to the site-specific expression of sox9 and runx2, two transcription factors required for the commitment of stem cells to a skeletogenic lineage, and the arrangement and orientation of newly deposited type I collagen fibrils. To gain insights into the genetic basis for skeletal mechanotransduction we conditionally inactivated focal adhesion kinase (FAK), an intracellular component of the integrin signaling pathway. By doing so we abolished the mechanically induced osteogenic response and thus identified a critical genetic component of the molecular machinery required for mechanotransduction. Our data provide a new framework in which to consider how physical forces and molecular signals are synchronized during the program of skeletal regeneration.

Figure 1

In vivo implant device permits defined stimulation of the bone marrow tissue. (A) A motion device, consisting of an intra-osseous, pin-shaped implant (im), held in place by a subcutaneous fixation plate is secured to the mouse tibia by two screws (dotted line is approximate skin level). An O-ring placed between the head of the implant and the center column of the fixation plate acts as a spring to return the implant to its starting position after axial displacement. (B) In vivo setting of micromotion device on murine tibia. (C) A linear variable differential transducer (LVDT) and load cell connected to the implant head and fixation plate allows the application and recording of displacement (∼150 µm) and the force (∼1N) required to produce motion.

Figure 2

Mechanical stimulation expands the pool of osteoprogenitor cells and accelerates their differentiation into osteoblasts. (A) On post-surgical d3, cells in the stimulated peri-implant space are densely packed within a proteoglycan-rich extracellular matrix (blue). (B) In the stationary environment, cells are loosely organized with no evidence of a mineralized extracellular matrix. (C) By d7, a thick (250 µm), fully mineralized bony sheath encapsulates the stimulated implant. (D) The tissue surrounding the stationary implant is absent of any bone matrix. (E) By d14, the bony encasement is more organized and still retains its original thickness. (F) The stationary tissue exhibits first sign of mineralization (90 µm thick) after 14 days. Scale bar: 100 µm.

Figure 3

Molecular and cellular response mirrors strain pattern. (A,B) PCNA staining reveals no differences in cell proliferation between unloaded and loaded samples. (C) In stimulated and (D) stationary implants sox9 is diffusely expressed throughout the surrounding bone marrow cavity. (E) Runx2 is broadly and strongly expressed in the peri-implant region in unstimulated samples, (F) whereas physical stimulation induces restriction of the runx2 transcripts to the cells adjacent to the implant. (G) Finite element modeling shows strain concentrations (tensile strain (t), compressive strain (c)) at the circumferential ridges and at the bottom of the implant (for illustration purposes, tensile strains were plotted on the right and compressive strains on the left). (H,I) µCT was used to record displacement of Tantalum particles, and principal strains were calculated by digital image correlation. Implant displacement generated a range of strain fields concentrated around circumferential ridges (cr)(*). (J,K) Picrosirius red staining in conjunction with polarized light microscopy reveals that in loaded samples, the peri-implant collagen fibrils (yellow-red) are abundant, tightly packed, and aligned parallel to the displacement trajectory, (L) whereas in unloaded samples, the collagen fibrils are unorganized. Scale bar: 100 µm.

Figure 4

FAK inactivation specifically blocks mechanically induced osteogenesis in vivo. (A) Col I expression marks peri-implant cells, (B) including those juxtaposed to the implant (im). (C) The schematic indicates the genomic structure of floxed FAK mice; crossing these mice with Cre mice carrying a 2.3Kb osteoblast-specific Col1a1 promoter resulted in Col1Cre^+/+;FAK^fl/fl (FAK mutant) mice. PCR was used to identify deletion of the fak allele in the animal. (D) In wildtype animals, seven days of stimulation result in abundant bone formation. (E) High magnification (Aniline blue) shows newly deposited bone matrix (blue) interlaced with blood vessels. (F) In FAK mutant mice, mechanical stimulation failed to induce osteogenesis. Note that FAK mutants were able to regenerate bone in unstimulated regions, as seen on the right periosteal surface. (G) Aniline blue staining shows complete absence of mineralized tissue in the peri-implant site. (H) Vascular ingrowth is not impeded by the deletion of FAK. (I,J,K,L) FAK mutant cells express sox9, runx2, col I and osteocalcin indicating that loss of FAK does not hamper the recruitment of osteochondroprogenitor cells to the peri-implant site. (M) Quantitative histomorphometric assessment of newly deposited bone matrix in unstimulated wild type (wt) bone marrow cavities (white), in stimulated wt bone marrow cavities (light gray), stationary FAK mutant bone marrow (gray), and in stimulated FAK mutant bone marrow cavities (black). * (P<0.1), # (p<0.001) indicates significant difference. Scale bar in A,D and F: 300 µm; in B,E and G–L: 100 µm.