Focal adhesion kinase regulates tendon cell mechanoresponse and physiological tendon development

Abstract

Tendons enable locomotion by transmitting high tensile mechanical forces between muscle and bone via their dense extracellular matrix (ECM). The application of extrinsic mechanical stimuli via muscle contraction is necessary to regulate healthy tendon function. Specifically, applied physiological levels of mechanical loading elicit an anabolic tendon cell response, while decreased mechanical loading evokes a degradative tendon state. Although the tendon response to mechanical stimuli has implications in disease pathogenesis and clinical treatment strategies, the cell signaling mechanisms by which tendon cells sense and respond to mechanical stimuli within the native tendon ECM remain largely unknown. Therefore, we explored the role of cell–ECM adhesions in regulating tendon cell mechanotransduction by perturbing the genetic expression and signaling activity of focal adhesion kinase (FAK) through both in vitro and in vivo approaches. We determined that FAK regulates tendon cell spreading behavior and focal adhesion morphology, nuclear deformation in response to applied mechanical strain, and mechanosensitive gene expression. In addition, our data reveal that FAK signaling plays an essential role in in vivo tendon development and postnatal growth, as FAK-knockout mouse tendons demonstrated reduced tendon size, altered mechanical properties, differences in cellular composition, and reduced maturity of the deposited ECM. These data provide a foundational understanding of the role of FAK signaling as a critical regulator of in situ tendon cell mechanotransduction. Importantly, an increased understanding of tendon cell mechanotransductive mechanisms may inform clinical practice as well as lead to the discovery of diagnostic and/or therapeutic molecular targets.

1 |. INTRODUCTION

Tendons enable locomotion by carrying high tensile forces via their dense, collagenous extracellular matrix (ECM),^1–4 which transmits mechanical stimuli to the resident cells. Tendons maintain a state of tensional homeostasis, in which both extracellular (e.g., applied loading) and intracellular (e.g., cell contractility) mechanical inputs regulate biological cell processes throughout physiological and pathological contexts.^5–8 Applied physiological levels of mechanical loading elicit an anabolic tendon cell response, including tendon hypertrophy, increased tenogenic and anabolic gene expression (e.g., Scx, Tnmd, and Col1a1), increased ECM synthesis, and cell proliferation.^9–15 Conversely, decreased mechanical loading evokes a rapid catabolic response, including cell apoptosis and increased synthesis and activation of proteolytic enzymes (e.g., MMP-1, MMP-3, MMP-13, and ADAMTS-5), which is followed by cell-mediated ECM contraction and remodeling leading to reduced tendon mechanical properties.^16–28 Taken together, the opposing anabolic and catabolic responses to physiological and de-tensioned mechanical loading environments, respectively, indicate that tensional homeostasis is critical for maintaining tendon structure and function.

Tendon pathologies, including acute ruptures as well as chronic overuse tendinopathy, are frequent, painful, and debilitating injuries.^29–32 Importantly, tendon tensional homeostasis has important implications in both disease pathogenesis and clinical treatment strategies. Aberrant mechanoresponse through pathological mechanical loading conditions can contribute to tendinopathy pathogenesis,^31–34 as it is hypothesized that overuse injuries generate tissue microdamage that disrupts the transmission of mechanical stimuli to the resident cells.³⁵ On the other hand, the anabolic response to applied loading at physiological levels is employed in physical therapy to treat tendon disease and injuries.^36,37 There exists a limited understanding of the in situ mechanotransductive signaling events by which mechanical stimuli regulate these clinically relevant processes, which represents a fundamental gap in knowledge that leads to uncertainty in prescribing and comparing physical therapy regimens.^38–42

Cells interpret mechanical stimuli with the surrounding three-dimensional ECM via integrin-mediated cell–ECM attachments called focal adhesions, which activate signaling cascades and transmit mechanical stimuli to the nucleus via the cytoskeleton to elicit a cell response.^43–46 Focal adhesion kinase (FAK, gene: Ptk2) is an intracellular protein kinase that regulates focal adhesion formation, maturation, and disassembly, with downstream effects on actin cytoskeletal organization and stress fiber formation.^{43,44,47–49} In addition, phosphorylation of FAK leads to downstream activation of other mechanotransductive molecules and signaling pathways (e.g., Src, ERK–MAPK, Rac1, and Rho/ROCK).^43–46 These roles for FAK are dependent on the cell’s mechanobiological environment, as intracellular phosphorylated FAK (pFAK) content is increased with cell-intrinsic and externally applied force at focal adhesions.⁵⁰ FAK is also required for promoting many mechanosensitive cell behaviors, including cell migration^51–54 and ECM synthesis, deposition, and remodeling.^55–57 While FAK signaling has not been studied in tendon cells within their native ECM, FAK phosphorylation is required for tenogenic gene expression (e.g., Scx, Tnmd, and Col1a1) and differentiation in response to mechanical and growth factor stimulation in mesenchymal and tendon-derived cells in monolayer culture.^58–64 Despite these well-described mechanotransductive roles for FAK in tendon, the degree to which FAK regulates mechanotransduction within the native tendon ECM as well as the regulatory role of FAK-dependent mechanotransduction in maintaining tendon homeostasis throughout postnatal growth remains unknown.

Therefore, the objectives of this study were to (1) elucidate the role of FAK-dependent mechanotransduction in regulating homeostatic tendon response to mechanical loading and (2) define the role of FAK in tenogenic cell differentiation and ECM synthesis during development. Our overall premise is that FAK-dependent tendon cell mechanotransduction regulates the cell response to altered mechanical loading throughout tendon development and homeostasis. Specifically, we hypothesized that FAK-dependent tendon cell mechanotransduction promotes the anabolic response to mechanical loading and mitigates the catabolic response to unloading. We further hypothesized that ablation of FAK in the tenogenic lineage in vivo results in impaired proliferation, differentiation, and ECM production that becomes more pronounced as locomotion increases during postnatal growth.

2 |. METHODS

2.1 |. Animal use

Animal use was approved by the University of Pennsylvania Institutional Animal Care and Use Committee (IACUC #805526). All animals were provided water and standard chow ad libitum and housed in a vivarium with 12 h of light exposure per day. We mated FAK^f/f mice, which contain loxP sites flanking the second kinase exon of Ptk2,⁶⁵ to ScxCre mice (MGI:5317938),⁶⁶ which will ablate FAK expression in the tenogenic lineage as Scx is expressed early in tendon development (Figure S1).⁶⁷ For the purposes of experimental breeding, male FAK^f/f mice were bred to female ScxCre;FAK^f/f mice to generate ScxCre+;FAK^f/f mice (FAK-KO) as well as ScxCre-;FAK^f/f (WT) littermate controls. All experimental animals were bred within an 18-month timespan and generated from breeding pairs made up of 16 males and 18 females. This study required a total of 234 mice (allocations to specific experiments are listed in Table S1), and no mice were excluded from this study. Mice were euthanized for ex vivo analyses via carbon dioxide inhalation per the American Veterinary Medical Association guidelines at postnatal days 10, 30, and 60 (P10, P30, and P60, respectively). P10 mice required confirmation of euthanasia via decapitation.

2.2 |. Tendon cell culture and morphology quantification

Tendon cells were isolated from the tails of WT male and female mice at P30. Prior to cell isolation, sterilized 18-mm coverslips were coated with fibronectin (Sigma F1141; 400 μL at 20 μg/mL) in a 12 well-plate. To isolate cells, tail tendons were digested in 0.4% (w/v) Collagenase IV (Worthington LS004188) and 0.3% (w/v) Dispase II (Sigma–Aldrich D4693) in high glucose DMEM with 1% antibiotic-antimycotic (Thermo 15240096) and 25 mM HEPES (Thermo 15630080) at 37°C. We have previously validated that this cell isolation protocol generates overwhelmingly Col1a1 and Scx-positive tenocytes.^25,68 Tendon cells were seeded at 2.5 × 10³ cells/well and cultured in high glucose DMEM with 5% fetal bovine serum (FBS; R&D Systems S11150) and 1% antibiotic-antimycotic at 37°C and 5% CO₂. At 36 h post-isolation and seeding, cells were treated with 10 μM FAK inhibitor (FAK-I) treatment (PF-573228; Tocris; Minneapolis, MN) in dimethyl sulfoxide (DMSO) solvent, consistent with previous publications demonstrating that 10 μM FAK-I treatment robustly attenuates FAK phosphorylation and catalytic activity.^69–71 Cells in the control group were treated with an equivalent DMSO media concentration. This pharmacological inhibitor study was performed on cells following isolation without passage (i.e., passage zero) to prevent phenotypic drift that would occur with higher passage numbers on tissue culture plastic.⁷² Previous studies in fibroblast monolayers have demonstrated that attenuated FAK signaling begins to generate a morphologically distinct cell phenotype between 2 and 12 h post-seeding.⁵¹ Therefore, we harvested cells following 6 h of inhibitor treatment by fixing them with 4% paraformaldehyde (PFA; Santa Cruz Biotechnology SC-281692) and storing them in phosphate-buffered saline (PBS) at 4°C until immunofluorescence staining was performed (within 48 h of fixation).

For immunofluorescence staining, blocking and cell permeabilization were performed with a solution of 5% normal goat serum and 0.1% Triton X-100 in PBS for 30 min at 25°C. Coverslips were stained with primary antibodies overnight at 4°C, rinsed with PBS, and stained with secondary antibodies for 1 h at 25°C. Coverslips were counterstained with a nuclear stain and mounted on glass slides with a ProLong Diamond Antifade Mountant (Invitrogen P36965). Antibody and nuclear stain product numbers and working concentrations are listed below (Table S2). Coverslips were imaged with a ZEISS Axioscan (ZEISS; Oberkochen, Baden-Württemberg, Germany) at 20× magnification. Images were analyzed via a custom CellProfiler⁷³ pipeline (RRID:SCR_007358) to quantify cell spreading behavior, focal adhesion morphology, and phosphorylated FAK localization.

2.3 |. Tendon explants: Mechanical stimulation under confocal microscopy

To visualize and quantify real-time tendon cell nuclear deformation in response to applied strain, we developed a custom mechanical loading device to apply mechanical tension to a tendon sample while imaging on an inverted ZEISS LSM 710 confocal microscope (Figure S2). This device was inspired by similar devices in the literature for applying tensile strain to a sample while imaging with confocal microscopy.^74,75 This device consists of 2 electric linear actuators (Zaber X-LHM050A) to apply mechanical strain as well as a 25 lb. load cell (Honeywell 060–1430-04) to monitor force in the tendon. The Zaber linear stages each have a 50 mm travel distance, combining to enable the mechanical loading device to achieve a range of motion of 100 mm. The load cell is calibrated by the manufacturer, which was confirmed by manually pulling on it with known weights. Manual stages (MT-XY; Newport Corporation; Irvine, CA) center the tendon in the x- and z-directions over the objective and place it within the objective working distance. All other device components were designed in Solidworks (Dassault Systèmes SolidWorks Corp; Waltham, MA) and fabricated out of Aluminum 6061 (McMaster-Carr; Douglasville, GA) using a vertical milling machine (Bridgeport Machine Company; Kalamazoo, MI). Custom LabView software (RRID:SCR_014325; National Instruments; Austin, TX) was developed to operate the device.

For the tendon explant study evaluating nuclear mechanoresponse, flexor digitorum longus (FDL) tendons from adult male (P120–220) WT mice were freshly dissected and maintained in high glucose, phenol-free DMEM supplemented with 25 mM HEPES, and 5% FBS. Tendons were randomized to control and FAK-I groups (n = 5–6 tendons per group). For FAK-I tendons, media was supplemented with 10 μM FAK-I for 1 h at 37°C. Following treatment, cell nuclei were stained with DRAQ5 (1:1000; Thermo 62251) for 30 min, mounted within a custom mechanical loading device, and imaged with confocal microscopy at 0%, 5%, and 10% applied strain. To control for differential effects of tendon viscoelasticity, we applied mechanical strains at ~1% strain/s (or 0.06 mm/s displacement) and waited for the force values to stabilize prior to imaging. Imaging was performed with a Zeiss LSM 710 confocal microscope using a 633 nm excitation laser and 10× objective by imaging through the maximum light penetration depth (~50–60 μm) at a z-stack interval of 5 μm. Nuclei were segmented with FIJI, and nuclear aspect ratios (nARs) were computed by fitting nuclei to ellipses and calculating the ratio of their major and minor axes. Finally, nuclei were manually tracked between strain levels using custom MATLAB software (RRID:SCR_001622). Live/dead staining was performed with calcein-AM and ethidium homodimer-1 (Thermo L3224) to confirm tissue viability. Notably, we selected calcein-AM and ethidium homodimer-1 staining to confirm tissue viability as this method specifically confirms the viability of the cells that we would be otherwise imaging for our mechanical loading experiment.

2.4 |. Tendon explants: Gene expression following stress deprivation

To evaluate FAK’s regulatory role on the mechanosensitive gene expression response to acute de-tensioning, FDL tendons from male and female P30 WT mice were freshly dissected and randomized to receive 10 μM FAK-I or control (DMSO) media (high glucose DMEM supplemented with 5% FBS and 25 mM HEPES at 37°C). Previous studies have indicated that the tendon response to de-tensioning begins as early as 30 min following loss of mechanical load.²⁶ Similarly, our previous work documents a substantial increase in proteolytic enzyme production in tenocytes within 6–24 h after matrix de-tensioning and/or modulation of actomyosin contractility.²⁵ Therefore, we maintained tendons in explant culture conditions for 1.5, 4, or 12 h prior to isolating RNA for gene expression quantification (n = 5/treatment group/timepoint).

2.5 |. Gene expression analysis

Gene expression analysis was performed on free-floating control and FAK-I treated FDL explants to evaluate mechanotransductive gene expression. In addition, gene expression analysis was performed on P30 WT and FAK-KO Achilles tendons (ATs), FDLs, and patellar tendons (PTs) to evaluate knockdown efficiency. For FDL explants, tendons were immediately homogenized in TRIzol (Thermo 15596018) following the experiment. For WT and FAK-KO ATs, FDLs, and PTs, samples were harvested immediately post-euthanasia and stored at −20°C in RNAlater^™ (Thermo AM7024) until being homogenized in TRIzol. Following TRIzol homogenization, all samples underwent RNA extraction (Zymo Research R2062) following manufacturer’s instructions. RNA purity was assessed using a Nanodrop spectrophotometer (Thermo) by measuring 260 nm/280 nm and 260 nm/230 nm absorbance ratios (reported here as mean ± standard deviation). Specifically, for the FDL explant experiment, we measured 260/280 and 260/230 ratios of 1.995 ± 0.074 and 1.973 ± 0.211, respectively. For the analysis of P30 WT and FAK-KO tendons, we measured 260/280 and 260/230 ratios of 2.040 ± 0.232 and 1.874 ± 0.282, respectively. Following cDNA reverse transcription (Thermo 4368814), samples underwent pre-amplification for specific gene targets using TaqMan probes as listed below (Table S3). Quantitative polymerase chain reactions were run using a Thermo QuantStudio6. ΔCt was calculated by subtracting the gene of interest Ct value from the Ct value for the housekeeping gene Abl1. Abl1 is an established housekeeping gene for many cell and tissue types,^76,77 which we have previously validated within the tendon and other fibrous musculoskeletal tissues.^68,78–80

2.6 |. Viscoelastic mechanical testing

ATs, FDLs, and PTs were isolated from P10, P30, and P60 WT and FAK-KO mice. Tendon cross-sectional areas (CSAs) were measured with a custom laser device as described.⁸¹ Tendons were then subjected to a viscoelastic mechanical testing protocol within a 37°C 1× phosphate buffered saline bath mounted within an Instron 5848 mechanical testing device (Norwood, MA). During mechanical testing, gripping the samples was performed with methods specific to the tendon being tested. For the proximal end of the Achilles tendon as well as the distal and proximal ends of the FDL tendon, sandpaper was superglued to the tendon tissue, which was then mounted between serrated screw-tightened grips. For the proximal end of the patellar tendon and the distal end of the Achilles tendon, mineralized tissue of the patella and calcaneus, respectively, were gripped within a screw-tightened toothed grip. For the distal end of the patellar tendon, the tibia was secured with polymethylmethacrylate within a custom 3D-printed container. The testing protocol included preconditioning (10 cycles of 0.5% strain amplitude oscillations at 1 Hz centered at 1% strain), a viscoelastic stress relaxation (5% strain for ATs and PTs or 3% strain for FDL rapidly applied and maintained for 5 min) and dynamic frequency sweep (10 cycles of 0.0125% strain amplitude oscillations at 0.1, 1, 5, and 10 Hz),⁸² and a quasi-static ramp to failure (0.1% strain/s). Tendon structural and viscoelastic material properties were then calculated from CSA measurements and load–displacement data using a custom MATLAB (Mathworks; Natick, MA) script.⁸² During analysis, CSA measurements, load–displacement data, and mechanical testing videos (captured with a CCD camera as the mechanical strain was applied) were manually inspected by a blinded investigator to ensure successful mechanical tests. Samples were excluded if there were technical errors during dissection, laser malfunction during CSA measurement, slipping within the sand-paper tissue grips, or premature mechanical failure at the growth plate. Overall, 218 out of 243 (~90.5%) tendons designated for mechanical testing were successfully tested and included in this study.

2.7 |. Tendon histology

All histological imaging for both paraffin-embedded and cryo-embedded samples was performed on a ZEISS Axioscan at 20× magnification. For paraffin histology of whole ankle and knee joints, limbs were dissected following sacrifice. Samples were fixed in 4% PFA in 1× PBS for 24 h, decalcified in 1% PFA in 0.5 M Sodium EDTA (Santa Cruz Biotechnology SC-286969), and embedded in paraffin. Decalcification timing depended on sample age (24 h for P10 joints and 10 days for P30 and P60 joints). Samples were sectioned in the sagittal plane at 7 μm or transverse plane at 10 μm. Cell density was visualized through Hoechst staining (1:1000 Hoechst 33342 in 50% (w/v) glycerol in PBS) and quantified with custom CellProfiler and MATLAB scripts. Tissue morphology was visualized via toluidine blue staining.

Cryosectioning was performed on FDL explants for analysis of phosphorylated FAK content as well as P10 knee joints for analysis of cell proliferation within the patellar tendon. For FDL explants, tendons were fixed at 0% strain while mounted within the loading device (Figure S2) for 5 min at 25°C, embedded in optimal cutting temperature medium (OCT; VWR 25608–930), and cryosectioned at 7 μm onto charged slides. Sections were stored at −80°C until staining. The staining protocol included fixing in 4% PFA for 10 min at 4°C, blocking with 5% Goat serum in PBS with 0.1% Triton for 20 min at 25°C, primary antibody incubation overnight at 4°C, and second antibody incubation for 1 h at 25°C. Primary and secondary antibody working concentrations are listed below (Table S4). Following staining, coverslips were mounted with 1:1000 Hoechst 33342 in 50% (w/v) glycerol in PBS.

For cell proliferation analysis of P10 patellar tendons, EdU (5-ethynyl-2′-deoxyuridine; Invitrogen A10044) in PBS solvent was injected at P0 and P2 at 3 μg/g mouse body weight. Following euthanasia at P10, knees were dissected and fixed in 4% PFA in PBS for 24 h and incubated in a 30% (w/v) sucrose solution for 8 h prior to embedding in OCT. Sagittal knee cryosections were taken at 7 μm using an established tape stabilization procedure and adhered to glass slides using a 0.75% (w/v) chitosan (Sigma–Aldrich 419419) in 0.25% (v/v) acetic acid (Sigma–Aldrich 695092).⁸³ Sections were stored at −20°C until staining, which was performed with Click-iT^™ Cell Reaction Buffer Kit (Invitrogen C10269) and Alexa Fluor 647 Azide (Invitrogen A10277) per manufacturer’s instructions. Coverslips were mounted on slides with 1:1000 Hoechst 33342 in 50% (w/v) glycerol in PBS. Nuclei segmentation and quantification of proliferating cells (i.e., the percentage of EdU-positive stained nuclei) was performed with custom CellProfiler and MATLAB scripts.

2.8 |. Collagen fibril morphology

PTs from P10, P30, and P60 mice were dissected, fixed in Karnovsky’s fixative (4% paraformaldehyde, 2.5% glutaraldehyde, 0.1 M sodium cacodylate, and 8.0 mM calcium chloride), post-fixed with 1% osmium tetroxide, dehydrated in ethanol, and embedded in Epon as described.^84,85 Transverse ultrathin sections (~85 nm) were acquired with an ultramicrotome and post-stained with UranyLess (EMS 22409) and 1% phosphotungstic acid. Imaging of the collagen fibril cross-sections was performed with a JEOL 1010 transmission electron microscope (TEM) at 60,000× magnification and an accelerating voltage of 80 kV. Collagen fibrils were segmented using custom MATLAB scripts to compare fibril diameter distributions and densities between genotypes.

2.9 |. Statistics

All ex vivo experiments and subsequent data analyses were performed by blinded investigators. Statistical analyses were performed in Prism and MATLAB. Analyses were generally two-tailed t-tests (statistical significance was set at p < .05) comparing control and FAK-I groups or WT and FAK-KO groups. Data distributions were checked for exclusion of statistical outliers (2.2×IQR as decided a priori) and tested for normality (Shapiro–Wilks normality tests) prior to subsequent statistical analysis. For comparing fibril diameter distributions between WT and FAK-KO tendons from TEM images, we utilized Kolmogorov–Smirnov tests. Relevant statistical analyses and p-value indicators are described in the figure captions. Initial sample sizes were estimated based on prior studies,^25,78 and mid-study power analyses were performed in MATLAB using “sampsizepwr” to calculate the final necessary sample sizes to achieve 80% power (i.e., the probability of correctly rejecting the null hypothesis). For these calculations, effect sizes (i.e., the difference between group means divided by the pooled standard deviation) were generated using the collected mid-study data.

3 |. RESULTS

3.1 |. FAK regulates tendon cell–ECM engagement in monolayer culture

While FAK signaling is known to regulate tenogenic gene expression in tendon-derived cells in monolayer culture,^58–64 FAK’s role in regulating cell–ECM engagement via focal adhesions in tendon cells remains unexplored. Therefore, we isolated tail tendon cells from mature wild-type (WT) male and female mice and cultured them on fibronectin-coated coverslips (Figure 1A). Cultured cells were treated with an FAK-I (10 μM PF-573228; FAK-I) or DMSO control. Following 6 h of inhibitor treatment, the cells were fixed and used for immunofluorescence staining to visualize and quantify cell and focal adhesion morphology. FAK-I treated cells developed pronounced cell protrusions compared to the spreading behavior observed in control cells (Figure 1B). In FAK-I treated cells, focal adhesions (i.e., visualized via Paxillin immunofluorescence staining) colocalized less with pFAK staining (Figure 1C), indicating that FAK-I treatment successfully attenuated phosphorylated FAK content at the focal adhesions. To evaluate the effect of FAK signaling on cell spreading behavior, we quantified cell morphology and focal adhesion localization. Although cell area was not different between groups (Figure 1D), FAK-I treated cells had higher cell compactness values relative to control cells (Figure 1E), which is indicative of the increased cell protrusion phenotype. Finally, focal adhesions localized closer to the cell periphery (Figure 1F) and were more elongated (Figure 1G) in FAK-I cells compared to control cells. These data indicate that FAK signaling is essential for regulating tendon cell spreading behavior as well as cell–ECM engagement.

FIGURE 1.

FAK regulates tendon cell spreading behavior and focal adhesion morphology in monolayer culture. (A) Study design and morphology quantification diagram. (B) Representative images of control and FAK-I treated tail tendon cells. Scale: 50 μm. Bottom row: Zoomed in images of focal adhesions labeled by the white boxes. Scale: 10 μm. (C) pFAK-Paxillin correlation coefficient, (D) cell area, (E) cell compactness, (F) focal adhesion distance to the cell periphery, and (G) focal adhesion eccentricity quantifications of control and FAK-I treated tail tendon cells. Data are pooled from cells derived from three mice (two males and one female) in two independent experiments (n = 30 cells/group/mouse). Bars indicate significant differences between groups (t-test: ***p < .001).

3.2 |. FAK regulates the tendon cell nuclear mechanoresponse within the in situ ECM microenvironment

To evaluate if FAK-dependent cell–ECM engagement regulates the in situ tendon cell response to mechanical stimuli, we applied mechanical load to freshly explanted FDL tendons under confocal microscopy using a custom mechanical loading device with or without FAK-I treatment (Figure 2A). Specifically, we asked whether FAK-dependent cell–ECM engagement was necessary for the transmission of mechanical strain from the tendon ECM to the cell nucleus. When tendons were mounted within the mechanical loading system, there were no discernable differences in cell viability between control and FAK-I treated tendons (Figure 2B), validating our system and indicating that FAK-I treatment was not acutely cytotoxic. To validate that FAK-I treatment successfully reduced FAK phosphorylation in situ, we performed pFAK immunofluorescence staining on cryosections taken of tendons treated in this manner. pFAK staining was drastically reduced in FAK-I treated tendons compared to controls (Figure 2C), suggesting successful inhibition of FAK signaling. As expected, nuclei tracked across strain levels in control tendons became increasingly elongated with applied strain (Figure 2D,E). Strikingly, nuclei from FAK-I treated tendons did not demonstrate the same nuclear elongation response across strain levels (Figure 2D,F). While there was no difference in nAR at 0% strain between control and FAK-I groups (Figure 2G), the attenuated nuclear elongation response effect of FAK-I treatment was consistent across all tendons imaged as normalized nAR was decreased at both 5% and 10% strain in FAK-I tendons relative to control tendons (Figure 2H). Ultimately, these findings indicate that FAK signaling is required for the mechanical tethering of the cell nucleus to the native 3-D ECM.

FIGURE 2.

FAK regulates the tendon cell nuclear mechanoresponse to extrinsic mechanical strain. (A) Study design. Nuclei within freshly excised FDL tendons were imaged at 0%, 5%, and 10% applied mechanical strain. (B) Live/dead staining within our custom mechanical stimulation system indicated that tendon resident tendon cells were overwhelmingly viable in both the control (left) and FAK-I treated (right) groups, as approximately 98% of cells within either group stained positively for calcein-AM and negatively for ethidium homodimer. n = 1–2 tendons per treatment group; n = 123–160 nuclei per tendon. Scale: 50 μm. (C) Phalloidin and pFAK staining of control (left) and FAK-I treated (right) FDL tendons demonstrate attenuated pFAK-staining in FAK-I treated tendons. Quantitatively, ~92.9% of tendon cells in control tendons stained positively for pFAK compared to ~7.1% of tendon cells in FAK-I treated tendons. n = 1–2 tendons per treatment group; n = 70–85 nuclei per tendon. Scale: 50 μm. (D) Representative images with overlayed nuclear segmentation of control and FAK-I treated nuclei at indicated applied mechanical strains. The displayed nuclei for the control and FAK-I groups exhibited increases in nAR of 12.3% and 0.7%, respectively (indicated on the nuclear segmentation overlay panel). Scale: 5 μm. (E) Control and (F) FAK-I treated representative plots for individual nuclei tracked across applied mechanical strain values. Non-normalized data were analyzed with repeated-measures one-way ANOVAs. (a) Significant difference from 0% strain; (b) Significant difference from 5% strain. (G) Average nAR for all explant samples at 0% applied macroscale mechanical strain. (H) Normalized nAR for all explant samples across increasing applied macroscale mechanical strain. Data represented as mean ± standard deviation for G and H. Non-normalized data were analyzed with t-tests comparing treatment groups at 0, 5, and 10% strain. Astericks indicate significant differences between treatment groups (*p < .05; **p < .01). n = 5–6 tendons per treatment group; n = 20–30 nuclei per tendon.

3.3 |. FAK regulates tendon cell mechanotransductive gene expression

While we have described that FAK is essential for the transmission of mechanical strain within the native tendon cell environment (Figure 2), which is likely due to FAK’s role in regulating focal adhesion morphology (Figure 1), whether FAK regulates downstream tendon cell mechanotransduction, including mechanosensitive gene expression, in response to altered mechanical stimuli within the native tendon ECM remains unknown. Following matrix detensioning, tendons rapidly respond by producing proteolytic enzymes, followed by actomyosin-mediated contraction of the tendon ECM.^16–24 In addition, we have demonstrated that both cell-intrinsic and cell-extrinsic de-tensioning leads to a rapid reduction in chromatin accessibility for mechanosensitive genes, such as alpha-smooth muscle actin (Acta2) and the Yap/Taz target gene cysteine-rich angiogenic inducer 61 (Cyr61), as well as increased catabolic enzyme production leading to degraded tissue mechanical properties.²⁵

Therefore, to determine the role of FAK in early mechanoresponse to de-tensioning, we evaluated the mechanotransductive and catabolic gene expression responses of explanted FDLs to stress-deprived (i.e., free-floating) culture conditions with or without FAK-I treatment (Figure 3A). In control tendons, we observed the expected substantial increases in Mmp3 (~90-fold) and Mmp13 (~15-fold) gene expression with time in culture (Figure 3B,C). Interestingly, FAK-I treated tendons demonstrated reduced Mmp3 expression at all timepoints and reduced Mmp13 expression at the 12 h timepoint (Figure 3B,C). In addition, FAK-I treatment modulated the expression levels of Acta2 and the Yap/Taz target gene Cyr61 (Figure 3D,E). Taken together, the profound effect of FAK-I treatment on catabolic and mechanosensitive gene expression suggests that FAK signaling is an integral component of cell sensation and response to changes in extrinsic mechanical stimuli.

FIGURE 3.

FAK regulates tendon cell mechanotransductive gene expression. (A) Study design. Gene expression levels (represented as 2^ΔΔCt relative to expression in the control group at 1.5 h of explant treatment) for (B) Mmp3, (C) Mmp13, (D) Acta2, and (E) Cyr61 for control and FAK-I treated tendons following 1.5, 4, and 12 hours of free-floating explant conditions. Datapoints represent individual FDL explants, with connecting lines indicating contralateral tendons (n = 5/treatment group/timepoint). Bars indicate significant differences between groups (t-test: *p < .05; **p < .01; ***p < .001).

3.4 |. Tendon-targeted FAK conditional knockout reduces tendon growth

Thus far, we have demonstrated that FAK is essential for cell sensation and response to mechanical loading. Next, we asked whether FAK would regulate tendon physiological development in vivo. We generated a Cre/LoxP knockout mouse line to attenuate FAK expression in the tendon lineage cells using scleraxis (Scx) expression to regulate Cre expression (ScxCre;FAK^fl/fl, referred to herein as FAK-KO), which attenuated the expression of FAK in the tenogenic lineage early on in development (Figure 4A).⁶⁷ In this mouse model, we evaluated the role of FAK in the development of three tendons, Achilles tendons (ATs), FDL, and patellar tendons (PTs), which provide a spectrum of loading levels. FAK (gene: Ptk2) expression was reduced in all three tendons (Figure 4B), validating our novel model system. In the mouse, the tendon reaches an inflection point at P10 where growth shifts from cell proliferation to matrix production and assembly, which generates a fully established tendon by P30.^86–92 Tendons continue to mature after P30, which is driven in part by continued extrinsic mechanical stimuli to resident tendon cells through locomotion. Therefore, we assessed the effect of FAK knockout throughout these crucial stages of postnatal tendon maturation. Upon dissection, FAK-KO ATs, FDLs, and PTs were strikingly smaller than WT tendons at all timepoints (Figure 4C–E), despite no difference in animal bodyweight (Figure 4F). To visualize this morphological difference, we performed sagittal paraffin histology with toluidine blue staining on WT and FAK-KO ATs, FDLs, and PTs (Figure S3A–C), which demonstrated morphologically intact yet thinner tendons with subtle changes in cell composition and organization. Focusing specifically on the PT, which demonstrated the largest size effect of FAK-KO, the difference in tendon size between genotypes was most easily visible in transverse sections (Figure 4G), in which we observed reduced CSA in FAK-KO PTs at P10 and P60 (Figure 4H,I). In summary, these results demonstrate that FAK-KO reduces tendon size in multiple tendons. Moreover, this FAK-KO tendon phenotype is present early postnatally and is maintained throughout continued tendon growth and maturation.

FIGURE 4.

Tendon-Targeted FAK Conditional Knockout Reduces Tendon Size. (A) ScxCre;FAK^f/f (FAK-KO) mouse model and study design. (B) FAK (gene: Ptk2) expression (represented as 2^ΔΔCt) was reduced in FAK-KO ATs, FDLs, and PTs compared to WT tendons at P30. Cross-sectional areas (CSAs) of dissected FAK-KO ATs, FDLs, and PTs were smaller than WT tendons at (C) P10, (D) P30, and (E) P60 timepoints. (F) Body weight measurements across timepoints revealed that FAK-KO mice were not significantly smaller compared to WT mice at any age evaluated despite reduced tendon size. Interestingly, FAK-KO mice had increased body weight at P60 relative to WT mice. (G) Representative images from WT and FAK-KO patellar tendons sectioned in the transverse plane. Dotted lines encircle the tendon CSA and are overlayed in the rightmost panel to illustrate reduced CSA in FAK-KO tendons. Scale: 500 μm. CSA quantifications of histological sections for (H) P10 and (I) P60 PTs. Data represented as mean ± standard deviation. Bars indicate significant differences between groups (t-test: *p < .05; **p < .01; ***p < .001; dotted lines represent trending (p < .1) differences).

3.5 |. FAK regulates postnatal maturation of material properties

Given the role of FAK signaling in promoting the development of tissue size throughout development, we next explored the role of FAK signaling in regulating tendon biomechanical properties, which we evaluated through tensile mechanical testing. We first perform this assay on tendons from male and female P30 mice. We noted expected decreases in stiffness in FAK-KO PTs (Figure 5A), and maximum load in FAK-KO ATs and PTs (Figure 5B). Interestingly, the linear modulus was increased in FAK-KO ATs and FDLs (Figure 5C), but maximum stress was increased only in FAK-KO ATs (Figure 5D). Finally, while there were no differences in viscoelastic stress relaxation between groups (Figure 5E), dynamic moduli were increased in all FAK-KO tendons at all frequencies evaluated (Figure 5F). Collectively, these results indicate that despite the impact of FAK-KO on tendon size, FAK-KO tendons still generate an ECM with material properties that are superior to those of WT tendons in some tendons.

FIGURE 5.

Quasi-static and viscoelastic material properties are increased in FAK-KO tendons at P30. (A) stiffness, (B) max load, (C) modulus (D) maximum stress, (E) stress relaxation, and (F) dynamic moduli measurements for male and female WT and FAK-KO P30 tendons. Dynamic modulus was measured at 0.1, 1, 5, and 10 Hz. Data represented as mean ± standard deviation. Bars indicate significant differences between groups (t-test: *p < .05; **p < .01; ***p < .001).

Importantly, sexual dimorphisms exist in tendon physiology and regulate homeostatic tendon laxity and rupture risk as well as collagen synthesis and inflammatory gene expression during healing.^93–103 Therefore, we sought to define potential sex differences in the role of FAK signaling in tendon development in our novel mouse model. FAK-KO ATs, FDLs, and PTs were smaller than the corresponding WT tendons for both male and female mice (Figure S4A). However, there were few differences in the FAK-KO biomechanical phenotype between sexes at P30 (Figure S4A–E), suggesting that sexual dimorphisms in our mouse model are minimal.

Next, we sought to define how FAK-KO affected tendon biomechanical properties throughout postnatal development by performing additional viscoelastic mechanical testing experiments at the P10 and P60 timepoints (Figures S5A and S6A). Consistent with the previously noted reduced CSA (Figure 4C,H), we observed a reduced stiffness in FAK-KO FDL and PTs (Figure S5B), as well as a decreased maximum load in FAK-KO PTs (Figure S5C). Modulus and maximum stress parameters were not different between groups (Figure S5D,E). Finally, viscoelastic stress relaxation values were increased in FAK-KO FDLs (Figure S5F), while dynamic moduli were modulated by FAK-KO in the ATs and FDLs (Figure S5G). As before, we observed decreases in CSA in all FAK-KO tendons relative to WT tendons, with the most severe effects in the ATs and PTs (Figure 4E,I). Structurally, FAK-KO resulted in reduced stiffness in the ATs (Figure S6B) and reduced maximum load in the FDLs (Figure S6C). Interestingly, the material properties modulus and maximum stress were increased in FAK-KO ATs and PTs relative to WT tendons (Figure S6D,E). While there were no differences in viscoelastic stress relaxation (Figure S6F), dynamic moduli measurements were increased in the FAK-KO ATs and PTs at all frequencies evaluated (Figure S6G). In summary (Figure 6), this data indicates that reduced tissue size is present in FAK-KO tendons as early as P10 and is maintained at older timepoints, suggesting that FAK is a critical regulator of tendon growth. FAK-KO tendons exhibited some expected decreases in the structural properties (stiffness and maximum load), although these effects appeared to largely dissipate with increasing age. Interestingly, the robustly increased tendon material properties observed in FAK-KO tendons are only present in the P30 and P60 datasets, indicating that FAK’s role in regulating tendon material properties becomes more pronounced during the ECM deposition phase of postnatal maturation. Finally, the reduced size and increased material properties of FAK-KO tendons increased in severity and became more specific to the AT and PTs at P30 and P60 timepoints, which may indicate that FAK’s role in regulating tendon physiology is dependent upon mechanotransduction as the energy-storing AT and PT experience higher mechanical loads compared to the positional FDL.^104,105

FIGURE 6.

Viscoelastic mechanical testing indicates that FAK-KO tendons are smaller throughout postnatal maturation and develop increased material with increased age, specifically in the AT and PT. Color and numbers within the cells indicate the ratio of the FAK-KO group mean relative to the WT group mean for that parameter. Asterisks represent significant differences between WT and FAK-KO groups (t-test: *p < .05; **p < .01; ***p < .001).

3.6 |. FAK-KO tendons demonstrate altered cell composition and matrix deposition

Given the interesting temporal roles of FAK in regulating tendon development and postnatal maturation, we sought to investigate the mechanistic origin of the FAK-KO tendon phenotype. Specifically, we asked whether attenuated FAK signaling in FAK-KO tendons leads to a defect in the cell proliferation or matrix deposition phases of tendon growth. To investigate this question, we quantified cell density and overall cell content in transverse histological sections of the PT (Figure 4G). Specifically, we observed a trending increase in cell density (Figure 7A) and reduced overall cell content in FAK-KO PTs relative to WT PTs (Figure 7B). These data led us to hypothesize that the decreased cell composition in FAK-KO PTs may be due to reduced cell proliferation during embryonic or early postnatal development. In addition, the decreased cell density in FAK-KO PTs indicates less deposited ECM between neighboring cells, suggesting that FAK signaling regulates matrix deposition during postnatal tendon growth and maturation. To investigate whether the reduced size of FAK-KO tendons was due to reduced cell proliferation, we assessed the histological tendon phenotype at P10, prior to the transition to ECM-deposition-based growth.^86–92 While we did not observe a difference in cell density between groups at this age (Figure 7C), there were fewer cells in the tendon cross-section in FAK-KO PTs relative to WT PTs (Figure 7D). We injected EdU to label and quantify the proliferating cells in WT and FAK-KO tendons during early postnatal development. Specifically, we performed injections at P0 and P2 and harvested at P10 to label proliferating cells during this early postnatal developmental period (Figure 7E). We did not detect a significant difference in EdU labeling in FAK-KO tendons relative to WT tendons (Figure 7F). Collectively, these data suggest that FAK-KO does affect cellularity at neonatal ages, although additional experiments profiling cell proliferation and apoptosis behavior in FAK-KO tendons are necessary to determine the extent to which FAK regulates cell proliferation-driven growth.

FIGURE 7.

FAK-KO PTs have increased cell density at P60 and decreased overall cell content at P10 and P60. Quantifications of histological transverse sections, including (A, C) cell density and (B, D) total cell content per cross-section (CS) for P60 and P10 WT and FAK-KO patellar tendons. (E) Study design depicting the timeline of EdU injections and tissue harvesting. (F) The percentage of nuclei positively labeled with EdU did not indicate an increased number of proliferating cells in FAK-KO P10 PTs. Data represented as mean ± standard deviation. Bars indicate significant differences between groups (t-test: *p < .05; dotted lines indicate trending (p < .1) differences).

To investigate altered ECM deposition in FAK-KO tendons, we performed transmission electron microscopy on transverse sections of P10, P30, and P60 PTs and segmented deposited collagen fibrils from the surrounding interfibrillar matrix. In the nascent ECM of P10 PTs, there was not a robust difference in fibril diameter distributions between WT and FAK-KO tendons (Figure 8A,B). Interestingly, in the more established ECM of P30 PTs, the distribution of collagen fibrils in FAK-KO tendons was shifted to the left, resulting in a smaller average fibril diameter (Figure 8A,C). Additionally, this difference in deposited ECM persists at P60, with a subtle leftward shift in the FAK-KO fibril diameter distributions and a smaller average fibril diameter (Figure 8A,D). Finally, there was a trending increase in fibril density in the FAK-KO P60 PTs compared to WT PTs (Figure S7). Collectively, these data demonstrate that FAK-KO regulates cellular composition and collagenous matrix deposition during postnatal tendon development.

FIGURE 8.

FAK-KO tendons demonstrate generally smaller collagen fibril diameters at P30 and P60. (A) Representative TEM images of collagen fibrils and collagen fibril diameter distributions of (B) P10, (C) P30, and (D) P60 PTs. Fibril diameter distributions were compared between groups using Kolmogorov–Smirnov tests, which demonstrated statistical significance (p < .001) between groups at all experimental timepoints. (B–D) Insets compare the average fibril diameter in their respective age group. Data represented as mean ± standard deviation. Bars indicate significant differences between groups (t-test: *p < .05; **p < .01).

4 |. DISCUSSION

The objectives of this study were to (1) elucidate the role of FAK in regulating tendon cell sensation of mechanical stimuli within the native tendon ECM and (2) define the role of FAK-dependent mechanotransduction in regulating cell phenotype and ECM deposition during physiological tendon development. Consistent with our hypotheses, we found that FAK signaling is a critical regulator of cell mechanotransduction within the native tendon ECM, and that FAK signaling is necessary for tendon development and postnatal growth. Specifically, we determined that FAK regulates tendon cell spreading behavior and focal adhesion morphology, nuclear deformation in response to applied mechanical strain, and longer-term changes in mechanosensitive gene expression in response to altered mechanical stimuli. In FAK-KO tendons, we observed reduced tendon size throughout postnatal growth as well as altered mechanical properties at later stages of postnatal maturation. In addition, we found that FAK-KO tendons had altered cell content, cell density, and ECM deposition. Taken together, these findings suggest that FAK plays a critical role in in situ tendon cell mechanotransduction through which FAK regulates the physiological development of tendon cellular and mechanical properties.

4.1 |. FAK regulates in situ tendon cell mechanotransduction

To evaluate the role of FAK in regulating tendon cell mechanotransduction, we performed in vitro cell culture and ex vivo tendon explant experiments. Previous experiments in other fibroblast cell types with attenuated FAK signaling via genetic knockdown or pharmacological inhibition have demonstrated that while FAK signaling is essential to focal adhesion assembly, it has a more substantial role in regulating the timing of focal adhesion disassembly.^{43,51,106,107} As a result, attenuated FAK signaling leads to increased persistence of focal adhesions, which become increasingly mature (i.e., larger and more elongated), migrate toward the cell periphery, and generate pronounced cell protrusions with robust actin stress fibers.^{43,51,106,107} We observed similar results in isolated tail tendon cells that were treated with an FAK-I, which likely indicates a conserved role for FAK signaling in regulating focal adhesion assembly and disassembly dynamics in tendon cells. While future studies are necessary to confirm that this finding is conserved in cells from energy-storing tendons, this result supports our hypothesis that FAK is an essential regulator of tendon cell sensation of their local mechanical microenvironment.

Interestingly, these findings in the cell culture monolayer may help to explain the attenuated mechanical strain transmission from the ECM to the nucleus in FAK-I treated tendon explants (Figure 3). The rapid effect of FAK-I treatment (i.e., 1 h of FAK-I media treatment) in our experiment suggests that in situ tendon cells dynamically remodel their actin cytoskeletons by reestablishing focal adhesions to maintain ECM–nucleus mechanical tethering. Moreover, the severely attenuated nuclear elongation response in FAK-I treated tendons is particularly surprising within the densely crowded ECM of a mature adult tendon, in which mechanical strain might be expected to be transmitted to the resident cells regardless of cell–ECM adhesions. Future experiments exploring nuclear mechanoresponse throughout tendon development (i.e., at varying stages of ECM maturity) are necessary to define the role of FAK in regulating the dynamic nature of the tendon cell cytoskeleton.

We have two primary hypotheses to explain FAK’s regulatory role on tendon cell nuclear mechanosensitivity. First, attenuated FAK signaling may lead the resident tendon cells to mechanically release from the ECM due to the inability to remodel their cytoskeleton and establish new focal adhesions. In this way, FAK-I treatment would directly disrupt the transmission of mechanical strain from the ECM to the resident cells. An alternative hypothesis is that FAK-I treatment may lead to aberrantly robust focal adhesions as supported by the literature^{43,51,106,107} and our cell monolayer experiments (Figure 1), and the increased rigidity of the cytoskeletal architecture surrounding may limit nuclear deformation. Importantly, this potential role for FAK in regulating intracellular tension may explain how FAK signaling promotes tendon differentiation and growth, as reduced intracellular tension via actin depolymerization in developing tendons significantly reduces the structural maturity of the ECM.⁹⁰ Similarly, we have recently described how manipulated intracellular tension alters tenocyte epigenetic gene regulatory mechanisms via Yap/Taz.²⁵ Therefore, deciphering between these competing hypotheses is of critical importance to establishing the mechanism by which FAK regulates tendon mechanobiology, and additional experiments exploring the role of FAK signaling in regulating in situ tendon cell cytoskeletal architecture and nuclear mechanobiology are necessary. Specifically, we plan to assess the role of FAK in regulating the actin cytoskeleton or lamin-based nucleoskeleton^108–110 by visualizing and quantifying these structures via immunofluorescence staining. In addition, we will aim to enhance the image quality obtained with our mechanical loading device, which is currently limited by the 10× objective required to provide the appropriate working distance when imaging tendon samples. Notably, by improving our imaging methodologies, we may elucidate more intricate details of nuclear deformation within the tendon ECM, thereby allowing for deeper insights into the hypotheses outlined above.

Previous studies in embryonic tendons and 3-D tendon constructs have demonstrated that the actin cytoskeleton is required for the maintenance of tendon cell fate, collagen fibril deposition, collagen crosslinking, and ECM tensioning.^22,90,91 Therefore, given FAK’s role in regulating the actin cytoskeleton,⁴³ it seems appropriate that it would play an essential role in regulating in situ tendon cell behavior. FAK is known to regulate tenogenic gene expression in monolayer culture,^58–64 which is also regulated by extrinsic mechanical stimuli.²⁶ In addition, the actin cytoskeleton and nuclear morphology are implicated in regulating chromatin accessibility,¹¹¹ and our lab has recently shown that cell-extrinsic and cell-intrinsic mechanical tension regulates chromatin accessibility specifically in mechanosensitive loci in tendon cells.²⁵ Taken together, the results from the present study suggest that the dependence on FAK for tenogenic gene expression may be due to its role in regulating nuclear mechanosensitivity.

In line with this hypothesis, we have demonstrated that FAK-I treatment severely attenuated the catabolic gene expression response to tendon de-tensioning, in addition to modulating expression of the mechanosensitive genes Acta2 and the Yap/Taz-target gene Cyr61 (Figure 3). Interestingly, the suppressed catabolic gene expression in FAK-I treated explants conflicts with some studies in the literature, in which FAK^−/− mouse embryonic fibroblasts exhibit increased collagen degradation and MMP expression.⁵⁵ Although we measured acute changes to matrix de-tensioning, these findings may differ from those in longer-term culture studies. Nevertheless, other studies have demonstrated a positive correlation between levels of MMP and FAK expression.^112–114 Importantly, previous experiments using tendon-derived cells in monolayer culture and 3-D constructs revealed that blebbistatin treatment, which inhibits actomyosin contraction, increases MMP expression.^25,115 Given the overlapping but distinct roles for FAK signaling and actomyosin contractility, these conflicting results may require additional experiments to resolve these potentially context-dependent discrepancies. For example, with loss of ECM tension, tendons first demonstrate an immediate decrease in tendon-specific gene expression (Scx) and a concomitant increase in catabolic (Mmp3, Mmp13, etc.) gene expression,^25,26 which is followed by a longer-term cell contractility and ECM degradation response.^{22,23,116,117}

Taken together, our results demonstrate that FAK signaling is essential for tendon cell mechanotransduction, with specific regulatory roles on tendon cell spreading behavior and focal adhesion morphology as well as nuclear deformation and mechanosensitive gene expression in response to extrinsic mechanical stimuli. A primary limitation of this work involves the use of the free-floating explant model, which we selected as it is the most well-established experimental system for exploring the tendon mechanobiological response to de-tensioning.^16–25 Nevertheless, this experimental design has significant limitations, including the non-physiological zero-stress state. While we can conclude from our experiments that FAK plays a critical role in regulating the gene expression response within this specific experimental system at these acute timepoints, future studies will need to address the role of FAK in regulating the physiologically relevant anabolic gene expression response to increased mechanical stimuli. In addition, we have demonstrated that FAK signaling regulates the short-term gene expression response to loss of ECM tension. Future work evaluating the longer-term effect of FAK-I treatment may increase our understanding of the regulatory role of FAK signaling on proteolytic enzyme synthesis and downstream ECM remodeling. Finally, while we found similar effects of FAK-I treatment in male and female tendon cells and explanted tendons (albeit our study is underpowered to statistically compare between sexes), our future work will investigate sexual dimorphisms as we continue to define the role of FAK signaling in regulating tendon cell mechanotransduction.

4.2 |. FAK regulates physiological tendon development and postnatal growth

To evaluate the role of FAK signaling in regulating tendon development and postnatal growth in vivo, we generated a novel tendon-targeted FAK knockout mouse model (ScxCre;FAK^f/f; FAK-KO). Our primary assay for defining tendon phenotype in this mouse model was viscoelastic mechanical testing, in which we found that FAK-KO tendons demonstrate reduced size at all ages with differences in ECM material properties appearing at the later postnatal timepoints. Notably, we found that the tendon size was significantly reduced at all timepoints, yet the overall animal body weights were not reduced (Figure 4F). This finding reinforces the importance of the tendon-specific role for FAK signaling in development and validates that our tendon-targeted conditional knockout mouse has limited off-target effects. Due to FAK’s role in regulating cell–ECM adhesions, this phenotype likely arises due to reduced tendon cell mechanotransduction throughout embryonic development and postnatal maturation. This hypothesis is supported by the increasing severity of the tendon phenotype throughout postnatal growth as mechanical stimuli to the resident cells increase during this period via increased locomotion and body weight. In addition, the increased material and viscoelastic properties in FAK-KO tendons demonstrated a larger effect size in the ATs and PTs compared to the FDLs, which may further indicate that differences in mechanotransduction drive the phenotype in FAK-KO tendons as the “energy-storing” AT and PT experience higher mechanical forces compared to the “positional” FDL. Despite this supporting evidence, additional experiments are necessary to confirm whether attenuated mechanotransduction is the mechanistic driving factor of the FAK-KO tendon phenotype. Specifically, future in vivo experiments in which tendon mechanical loading is increased or decreased would allow assessment of the role of FAK in regulating cell sensation and overall tendon response to changing extrinsic mechanical stimuli.

Tendon development occurs through subsequent cell proliferation and matrix deposition phases of growth that reach a critical transition point at around P10.^86–92 To explain the reduced size of FAK-KO tendons, we hypothesized that attenuated FAK signaling leads to defects in these phases of postnatal tendon growth. In the literature, attenuated FAK signaling through pharmacological inhibition or genetic knockout restricts cell proliferative capacity within in vitro and in vivo contexts.^{54,69,118–120} In our model, we observed conflicting evidence for reduced cell proliferation in FAK-KO tendons, including reduced cellularity (Figure 7B,D) and reduced tendon size immediately following the proliferative growth phase at P10 (Figure 4C,H). However, EdU labeling at P0 and P2 did not demonstrate a difference in cell proliferative behavior between WT and FAK-KO tendons (Figure 7E,F). Therefore, we cannot confirm that cell proliferative behavior is altered in FAK-KO tendons. An alternate possibility is that the reduced cell content at P60 is indicative of increased apoptosis due to attenuated FAK signaling in FAK-KO tendons throughout postnatal growth and maturation. In the literature, FAK signaling is known to suppress apoptosis through the regulation of mitotic cell cycle markers (such as the cyclin D1) and the apoptotic markers caspase-3 and p53.¹¹⁸ Interestingly, this hypothesis is consistent with FAK’s regulatory role in cell mechanotransduction, as decreased extrinsic mechanical stimuli leads to increases in tendon cell apoptosis.^21,26

In addition to a reduced proliferative phase of tendon growth, we hypothesized that attenuated FAK signaling leads to a defect in the ECM deposition phase of postnatal tendon development,^86,87 which is supported by the increasing severity of the reduced CSA observed in FAK-KO tendons with increased postnatal age (Figure 6) as well as increased cell density in P60 transverse sections (Figure 7A). Finally, collagen fibril diameter distributions in both P30 and P60 PTs demonstrated immature matrix deposition (Figure 8), which may contribute to reduced macroscale tendon size. Consistent with this hypothesis, fibroblast-targeted genetic FAK knockout is known to regulate collagen expression, synthesis, deposition, and tractional remodeling within in vitro and in vivo contexts.^55,57 Importantly, the deposition and remodeling of collagen ECM structure in tendon cells is known to require mechanical stimuli and cellular actomyosin machinery.^90,121,122

One potential hypothesis to explain the reduced size in FAK-KO tendons could be altered ECM production and composition compared to WT tendons. For example, in P60 PTs, we found that the reduced size of FAK-KO tendons compared to WT tendons was more pronounced in the viscoelastic mechanical dataset compared to the histological data. This finding indicates that the tissue size difference between genotypes is likely dependent upon the tissue hydration state, which may suggest a difference in the tendon ECM composition, particularly in the non-collagenous molecules that play an essential role in water retention.¹²³

Consistent with this hypothesis, our data demonstrate that the ECM of FAK-KO tendons is indeed different compared to WT tendons, specifically including increased tissue material properties and reduced collagen fibril diameters. Interestingly, these differences in material properties and collagen fibril diameters are not present at P10 but are pronounced at the P30 and P60 timepoints. These interesting data led us to ask whether the phenotype in FAK-KO tendons is due to altered ECM deposition or differences in the matrix remodeling response. Specifically, the secondary developmental stage of ECM deposition occurs primarily through lateral growth of the collagen fibrils rather than the deposition of new fibrils.^88,124 This process of fibril remodeling is known to require actomyosin-dependent structures called fibripositors.^122,125,126 Therefore, regulation of the actomyosin machinery within resident tendon cells may be the mechanism by which FAK regulates the phenotype of the deposited ECM at the P30 and P60 timepoints. Alternatively, we have previously demonstrated that smaller collagen fibrils are representative of the tissue remodeling response to injury.⁷⁸ Therefore, it may be that the smaller fibrils in FAK-KO tendons are representative of cell response to tissue injury, rather than developmental differences in matrix deposition. To address this potential hypothesis in the future, we plan to evaluate potential differences in the tendon response to injury in FAK-KO tendons using a tamoxifen-inducible CreERT2 knockout mouse model.

In summary, we observed that FAK signaling is essential for tendon development and postnatal maturation, as FAK-KO tendons demonstrated reduced size and altered mechanical properties due to reduced tendon cell content and immature ECM deposition. A primary limitation of this work is that while we hypothesize that the defects in FAK-KO tendon development are due to decreased mechanotransduction, we cannot confirm that this effect is dependent upon mechanotransduction based on our present data. Future experiments could be undertaken to utilize a spectrum of in vivo mechanical loading paradigms, such as rodent treadmill running or botox injection, which could further substantiate this hypothesis. In addition to mechanical loading models, transcriptomic analysis of the expression of mechanotransductive pathways in FAK-KO tendons would also further elucidate the in vivo mechanotransductive roles of FAK.

5 |. CONCLUSION

This work was inspired by the lack of mechanistic understanding of the mechanotransductive cellular mechanisms that regulate the sensation of mechanical stimuli within the native tendon ECM. To address this gap in knowledge, we specifically targeted cell–ECM adhesions by perturbing FAK signaling in tendon-derived cell culture monolayer experiments, ex vivo tendon explant experiments, and in a novel in vivo mouse model. This multi-dimensional approach is a significant strength of this work as it allowed us to explore the mechanism of FAK-dependent cell mechanotransduction while confirming physiological relevance. Results from the experiments presented herein have established a foundation on which future experiments will continue to explore the regulatory role of FAK on tendon cell mechanotransduction within the native tendon ECM. Importantly, an increased understanding of these mechanotransductive mechanisms will inform how tendon cells maintain tensional homeostasis and has ultimate implications for the pathogenesis and future treatment of overuse tendinopathy. In addition, the investigation of downstream mechanotransductive pathways that underlie tendon response to mechanical stimuli may lead to the discovery of diagnostic and/or therapeutic molecular targets.

Abbreviations:

Acta2

Alpha-actin-2

ADAMTS

A Disintegrin And Metalloproteinase with Thrombospondin motifs

AT

Achilles Tendon

cDNA

Complimentary DNA

Col1a1

Collagen Type I Alpha 1

CSA

Cross-Sectional Area

Cyr61

Cysteine-rich angiogenic inducer 61

DMSO

Dimethyl Sulfoxide

DMEM

Dulbecco’s Modified Eagle Medium

ECM

Extracellular Matrix

EdU

5-ethynyl-2’-deoxyuridine

FAK

Focal Adhesion Kinase

FAK-I

Focal Adhesion Kinase Inhibitor

FDL

Flexor Digitorum Longus

FBS

Fetal Bovine Serum

FIJI

ImageJ software distribution for biological image analysis

HEPES

4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

IACUC

Institutional Animal Care and Use Committee

KO

Knockout

MMP

Matrix Metalloproteinase

nAR

Nuclear Aspect Ratio

PBS

Phosphate Buffered Saline

PFA

Paraformaldehyde

pFAK

Phosphorylated Focal Adhesion Kinase

PT

Patellar Tendon

Ptk2

Protein Tyrosine Kinase 2 (gene for FAK)

RNA

Ribonucleic Acid

RNAse

Ribonuclease

RT-qPCR

Reverse Transcription Quantitative Polymerase Chain Reaction

Scx

Scleraxis

WT

Wild Type

Yap/Taz

Yes-associated Protein/Transcriptional Co-Activator with PDZ-Binding Motif.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author.