1. Stem Cells and Regenerative Medicine
Download icon

Scleraxis-lineage cell depletion improves tendon healing and disrupts adult tendon homeostasis

  1. Katherine T Best
  2. Antonion Korcari
  3. Keshia E Mora
  4. Anne EC Nichols
  5. Samantha N Muscat
  6. Emma Knapp
  7. Mark R Buckley
  8. Alayna E Loiselle  Is a corresponding author
  1. Center for Musculoskeletal Research, University of Rochester Medical Center, United States
  2. Department of Biomedical Engineering, University of Rochester, United States
Research Article
  • Cited 1
  • Views 925
  • Annotations
Cite this article as: eLife 2021;10:e62203 doi: 10.7554/eLife.62203

Abstract

Despite the requirement for Scleraxis-lineage (ScxLin) cells during tendon development, the function of ScxLin cells during adult tendon repair, post-natal growth, and adult homeostasis have not been defined. Therefore, we inducibly depleted ScxLin cells (ScxLinDTR) prior to tendon injury and repair surgery and hypothesized that ScxLinDTR mice would exhibit functionally deficient healing compared to wild-type littermates. Surprisingly, depletion of ScxLin cells resulted in increased biomechanical properties without impairments in gliding function at 28 days post-repair, indicative of regeneration. RNA sequencing of day 28 post-repair tendons highlighted differences in matrix-related genes, cell motility, cytoskeletal organization, and metabolism. We also utilized ScxLinDTR mice to define the effects on post-natal tendon growth and adult tendon homeostasis and discovered that adult ScxLin cell depletion resulted in altered tendon collagen fibril diameter, density, and dispersion. Collectively, these findings enhance our fundamental understanding of tendon cell localization, function, and fate during healing, growth, and homeostasis.

Introduction

Despite the significant efforts toward improving tendon healing and regeneration, the specific cellular contributions during tendon healing have not been extensively characterized (Nichols et al., 2019). While many studies have examined the potential of using various stem cell populations to promote healing, originating from both tendon intrinsic (Walia and Huang, 2019) and extrinsic sources (Costa-Almeida et al., 2019), little focus has been directed toward defining the functions and therapeutic potential of tendon cells during tendon healing following an acute injury. Tendon cells are increasingly recognized as a heterogenous population of cells where many, but not all, express the gene Scleraxis (Scx) (Best and Loiselle, 2019; Kendal et al., 2020; De Micheli et al., 2020). Understanding the localization and function of tendon cell subpopulations during healing is likely to be instrumental in better defining the mechanisms that promote scar-mediated healing, which results in poor patient outcomes, and could therefore be used to develop pro-regenerative approaches to improve healing.

Scx, a basic helix-loop-helix transcription factor, is currently the most well-characterized marker that the field possesses to study tendon (Schweitzer et al., 2001). Scx has been utilized to examine tendon biology and development (Murchison et al., 2007; Pryce et al., 2009; Huang et al., 2019), healing and regeneration (Sakabe et al., 2018; Howell et al., 2017; Best and Loiselle, 2019; Dyment et al., 2014), differentiation (Bavin et al., 2017; Chen et al., 2012; Alberton et al., 2012; Nichols et al., 2018a), and mechano-transduction (Maeda et al., 2011; Nichols et al., 2018b). Functionally, Scx expression can drive matrix production and remodeling (Sakabe et al., 2018; Leéjard et al., 2007; Levay et al., 2008), epithelial-to-mesenchymal transition (Al-Hattab et al., 2018), development of force-transmitting tendons (Murchison et al., 2007), tendon growth (Gumucio et al., 2020), and effect focal adhesion morphology (Nichols et al., 2018b). Despite the effort to understand the functions of Scx as a transcription factor, the function and requirement of Scx-lineage (ScxLin) cells in tendon repair, post-natal growth, and homeostasis has not been defined.

Previous studies examining the localization and role of tendon cells during healing are limited. Scx-GFP mice (Pryce et al., 2007) have been used to visualize tendon cells in many studies; however, previous work has suggested that extrinsic, paratenon-derived Scx-GFP- cells can turn on Scx expression and become Scx-GFP+ by 14 days in a Patellar tendon injury model (Dyment et al., 2014). This makes interpretation using Scx-GFP mice complicated during in vivo studies as it becomes difficult to determine if a Scx-GFP+ cell is tendon-derived or simply activating Scx expression post-injury. Other studies have utilized the inducible Scx-CreERT2 mouse model to label Scx+ cells prior to injury to allow tracking of these cells post-injury (ScxAi9). Howell et al. determined that while Scx-GFP cells and ScxAi9 tendon cells localized to the regenerated tendon in neonatal mice, these cells were not recruited to the scar tissue/bridging tissue during adult healing in Achilles tendon (Howell et al., 2017). In contrast, we have previously demonstrated that ScxAi9 tendon cells organize into a linear, cellular bridge spanning the scar tissue between the tendon stubs following acute injury and repair of the adult flexor digitorum longus (FDL) tendon (Best and Loiselle, 2019). While these studies suggest that tendon type (ex. flexor, Achilles, etc.) and injury model-specific (ex. transection with no repair, transection with repair, etc.) differences may modulate the ScxLin cell contribution to injury, it also highlights the near complete lack of characterization of ScxLin cell function in the healing process.

In the present study, we hypothesized that ScxLin tendon cells would be required for successful healing in an adult model of acute flexor tendon repair by driving formation of a collagenous tissue bridge. We utilized a genetic mouse model of ScxLin cell depletion to directly assess the function of tendon cells during healing and surprisingly discovered that depletion of ScxLin tendon cells resulted in improved tendon biomechanical properties. We also examined alterations in wound healing-related cell populations, transcriptomics via RNA sequencing, and the effects of ScxLin cell depletion on tendon post-natal growth and homeostasis.

Results

Successful ablation of Scleraxis-lineage tendon cells using diphtheria toxin receptor mouse model

To determine the feasibility of depleting tendon cells using Scx-Cre; Rosa-DTRLSL (ScxLinDTR) mice (Figure 1A), diphtheria toxin (DT) was administered into the right hind paw for 5 consecutive days. Ten days after the final DT injection, both the injected and contralateral control hind paws were harvested (Figure 1B). This approach resulted in 57% depletion of tendon cells in uninjured ScxLinDTR FDL tendons relative to WT control littermates (p<0.0001) (Figure 1C and D). Tendon cell number was unaffected in the contralateral FDL, indicating that local DT injections did not induce cell death in locations other than the targeted hind paw (Figure 1C and D).

Figure 1 with 2 supplements see all
Efficiency of tendon cell and ScxLin cell depletion.

(A) To deplete ScxLin cells, Scx-Cre mice were crossed to the diphtheria toxin receptor mouse (ScxLinDTR). (B) Mice received five hind paw injections of DT and were harvested 10 days after the final injection. (C) Sections from injected and contralateral (CL) hind paws from WT and ScxLinDTR mice were stained with DAPI, and total DAPI+ cells within the tendon (white outline) were quantified (D). (E) To determine the depletion efficiency specifically of ScxLin cells, Scx-Cre; Rosa-DTRLSL; Rosa-Ai9 and Scx-Cre; Rosa-Ai9 reporter mice were given local, daily DT injections for 5 consecutive days and hind paws were harvested 24 and 38 days after the last injection (F) These are the contralateral control tendons from the mice in Figure 4 that underwent tendon injury and repair. (G) Hind paws from ScxLinAi9 and ScxLinAi9DTR were probed for Red Fluorescent Protein (RFP; Ai9) expression and counterstained with the nuclear dye DAPI. (H) Total tendon cell density (DAPI+), total ScxLinAi9+ cell density and the percentage of ScxLinAi9 cells (ScxLinAi9+ cells/ DAPI+ cells) were quantified in ScxLinAi9 and ScxLinAi9DTR tendons and demonstrate a significant reduction of ScxLinAi9 cells in ScxLinAi9DTR relative to ScxLinAi9 WT controls. N = 3–4 per genotype. Two-way ANOVA with Sidak’s multiple comparisons test used to assess statistical significance of tendon cell ablation between hind paw (injected with DT or contralateral) and genotype (ScxLinAi9 and ScxLinAi9DTR at 24 and 38 days). * indicates p<0.05 for the indicated comparison, ** indicates p<0.01 for indicated comparison, **** indicates p<0.0001 relative to all other groups.

To better understand how this depletion regimen specifically effects ScxLin cells we used Scx-Cre; Rosa-DTALSL; Rosa-Ai9 reporter mice (ScxLinAi9DTR) (Figure 1E). Samples were harvested at 24 and 38 days post-depletion (Figure 1F). At 24 days post-depletion, there was a 58.9% reduction in total tendon cells in uninjured ScxLinAi9DTR FDL tendons relative to WT control littermates (p<0.0105) and a 68.48% depletion of ScxLinAi9 cells in ScxLinAi9DTR FDL tendons relative to WT ScxLinAi9 controls (p<0.0068) (Figure 1G and H). At 38 days post-depletion, total tendon cells were reduced by 60.42% (p<0.0091), and ScxLinAi9 cells were reduced by 73.65% relative to ScxLinAi9 WT controls (p<0.0045) (Figure 1G and H). No significant differences in total tendon cell depletion efficiency (p<0.9935) and ScxLinAi9 cell depletion efficiency (p<0.9359) were found between D24 and D38 post-depletion timepoints.

To understand how ScxLinDTR affected previously established tendon cell sub-populations (Best and Loiselle, 2019), we evaluated active Scx and S100a4 expression in ScxLinDTR and WT uninjured flexor tendons 10 days following the final DT injection. The number of Scx+ cells was significantly reduced in the ScxLinDTR tendons relative to WT littermates, as expected (p=0.0448) (Figure 1—figure supplement 1A &B). Similarly, when Scx+ cells were normalized to total cell count, there was a trending reduction of Scx+ cells in ScxLinDTR tendons compared to WT (p=0.1000) (Figure 1—figure supplement 1C). ScxLinDTR tendons had a trending decrease in S100a4+ cells relative to WT littermates (p=0.0941) (Figure 1—figure supplement 1D &E). Interestingly, when S100a4+ cells were normalized to total cell count, there was no significant or trending difference between groups (p=0.3525), suggesting some remaining tendon cells may begin expressing S100a4 following ScxLincell depletion (Figure 1—figure supplement 1F). ScxLinDTR tendons exhibited a small number of apoptotic cells peripheral to, but not within, the tendon suggesting that a 10-day DT washout period was sufficient for all tendon-specific DT-induced cell death to occur (Figure 1—figure supplement 2A). Both ScxLinDTR and WT littermates exhibited PCNA+ cells within the muscle, but not within then tendon, and ScxLinDTR exhibited a few PCNA+ cells within the tendon that were not present in WT (Figure 1—figure supplement 2B). ScxLinDTR tendons exhibited more F4/80+ macrophages peripheral to the tendon, potentially due to macrophage recruitment to the tendon to clean up apoptotic tendon cell debris (Figure 1—figure supplement 2C).

Ablation of Scleraxis-lineage cells results in significantly increased biomechanical properties by day 28 post repair while not affecting gliding function

To define the functional effects of ScxLin cell depletion on tendon healing, mice received five local DT injections to deplete ScxLin cells followed by FDL repair 10 days following the final injection (Figure 2A). A trending improvement in MTP Range of Motion (ROM) in ScxLinDTR repairs was observed at day 14, relative to WT littermates (WT vs ScxLinDTR, p=0.0711), but this trend was absent by day 28 post-repair (Figure 2B). ScxLinDTR healing tendons did not significantly differ in gliding resistance at either day 14 or 28 post-repair (Figure 2C). While biomechanical properties were not altered between groups at day 14, both stiffness and maximum load at failure were significantly increased in ScxLinDTR healing tendons relative to wildtype littermates at day 28 post-repair (Stiffness: WT: 6.48 ± 0.75, ScxLinDTR: 11.22 ± 1.83, p=0.0237; Maximum load at failure: WT: 1.54 ± 0.17, ScxLinDTR: 2.44 ± 0.24, p=0.0061) (Figure 2D and E). Between days 14 and 28, WT tendon stiffness increased by 39.06% while ScxLinDTR stiffness increased by 109.33%, and WT tendon maximum load at failure increased by 52.48% while ScxLinDTR maximum load at failure increased by 162.37%, indicating that ScxLinDTR repairs heal at an accelerated rate relative to wild-type littermates.

ScxLinDTR tendons heal with significantly increased biomechanical properties.

Mice received five hind paw injections of DT on consecutive days, underwent flexor tendon repair surgery 10 days after the final DT injection, and were harvested at 14- and 28 days post-repair (A). Measurement of metatarsophalangeal (MTP) joint flexion angle (B), gliding resistance (C), stiffness (D), and maximum load at failure (E) of WT and ScxLinDTR repaired tendons. N = 9–12 per genotype per timepoint. Students t-test used to assess statistical significance between genotypes at a given timepoint. * indicative of p<0.05, ** indicative of p<0.01.

Scleraxis-lineage cells are not required for the formation of a bridging collagen matrix during tendon healing

We have previously demonstrated that a cellular bridge corresponds to a region of bridging collagen matrix in the scar tissue (Best and Loiselle, 2019). While there were no apparent differences in tissue morphology between groups at days 14 or 28 post-repair (Figure 3B), we wanted to determine if tendon cell depletion prevented formation of the collagen bridge. Masson’s trichrome staining revealed presence of bridging collagen through the scar in both groups at days 14 and 28, indicating that depletion of ScxLin tendon cells does not prevent formation of the bridging collagen matrix (Figure 3C).

ScxLin cell depletion does not disrupt formation of a bridging collagen matrix.

Mice received five hindpaw injections of DT on consecutive days, underwent flexor tendon repair surgery 10 days after the final DT injection, and were harvested at 14 and 28 days post-repair (A). Alcian blue/hematoxylin and Orange G stain utilized to assess overall morphology (B). Masson’s trichrome stain used to visualize collagen content and organization (C). Tendon is outlined by white dotted line and scar tissue by yellow dotted line. Black boxes indicate location of higher magnification images. Boundaries of bridging collagen indicated by black arrows. N = 4 genotype per timepoint. Suture indicated by *.

ScxLin cell depletion results in a significant lower Scxlin+ cells during tendon healing

To better understand how depletion of ScxLin cells prior to injury and repair affected ScxLin cell density during healing, we traced ScxLinAi9 cells at D14 and D28 post-injury (Figure 4A) No significant differences in ScxLinAi9 cells were detected between WT (ScxLinAi9) and ScxLinAi9DTR (p=0.3115) at D14 post-surgery (Figure 4B,C). In contrast, at D28 post-surgery, a significant decrease in ScxLinAi9 cells was observed in ScxLinAi9DTR repairs (p<0.0034) relative to WT repairs (Figure 4B,D). Collectively, these data suggest that depletion of ScxLinAi9 cells prior to injury does not alter the overall ScxLinAi9 content at D14, possibly due to additional labeling of cells that express Scx following injury. In contrast, by D28 the effects of depleting ScxLinAi9 cells prior to tendon injury is manifested in changes in both ScxLinAi9 content (Figure 4B,D,E) and phenotypic differences (Figure 2).

ScxLin cell depletion results in time-dependent changes in ScxLinAi9 cell presence during tendon healing.

(A) Mice received hind paw injections of DT on 5 consecutive days, underwent flexor tendon repair surgery 10 days after the final DT injection, and were harvested at 14 and 28 days post-repair. (B) Immunofluorescence for RFP (Ai9) in WT ScxLinAi9 and ScxLinAi9DTR tendon repairs at 14 and 28 days post-repair to define changes in ScxLinAi9 contribution following ScxLin cell depletion. Quantification of ScxLinAi9+ area in ScxLinAi9 WT repairs and ScxLinAi9DTR repairs at (C) D14 and (D) D28 post-surgery. Nuclei were stained with DAPI. N = 4 per genotype. Student’s t-test used to assess statistical significance between genotypes at a given timepoint. **indicates p<0.01. (E) Proposed model of the time-dependent contributions of ScxLinAi9 cells to the tendon healing process. During adult tendon homeostasis ScxLinAi9 cells are the predominant tenocyte population and ScxLinAi9DTR results in depletion of ~60% of these cells. Red cells indicate ScxLinAi9 cells that were present in the tendon when depletion was initiated. We hypothesize that no differences in the proportion of ScxLinAi9 cells is observed at D14 (concomitant with a lack of functional phenotypic differences) due to the predominance and functions of other cell populations, including those that express Scx in response to injury and are therefore labeled as ScxLin (blue cells). In contrast, we hypothesize that by D28 the contribution of ‘new’ ScxLin cells (blue cells) has waned, and that the ScxLinAi9 cells that were present in the tendon during adult tendon homeostasis (red cells) are now the predominant tenocyte population and exert their functions at this time as suggested by functional differences between WT and ScxLinDTR at this time. This schematic was made using http://www.biorender.com.

ScxLin depletion enhances myofibroblast content during tendon healing

We have previously demonstrated that elevated F4/80+ macrophages and αSMA+ myofibroblasts are associated with increased tendon maximum load at failure (Best et al., 2019). No significant differences in F4/80+ macrophages were detected between genotypes at either day 14 or 28 post-repair (Figure 5B). While αSMA+ myofibroblasts were not significantly altered at day 14 (D14: WT vs DTR, p=0.3790), ScxLinDTR healing tendons had significantly increased levels of αSMA+ myofibroblasts relative to wildtype littermates at D28 (p=0.0188) (Figure 5C). We have previously demonstrated that S100a4 is an important molecule that can influence tendon biomechanical properties and gliding function during healing (Ackerman et al., 2019). No differences in S100a4+ cells were observed in ScxLinDTR, relative to WT at either day 14 or 28 post-repair (Figure 5D). Thus, αSMA+ myofibroblasts are the most likely candidate driving the increased biomechanical properties seen in ScxLinDTR healing tendon, of the cell populations investigated, consistent with their roles in with matrix deposition, organization, and contraction. To demonstrate both the specificity of the αSMA+ myofibroblast response to injury and that this tendon injury and repair model does not induce degeneration of the proximal/ distal ends of the tendon due to altered loading, we examined αSMA expression adjacent to the repair site. No αSMA staining was observed in the proximal/distal native tendon away from the repair site at D14 or D28 (Figure 5—figure supplement 1).

Figure 5 with 1 supplement see all
ScxLinDTR repaired tendons heal with increased presence of αSMA+ myofibroblasts.

Mice received five hindpaw injections of DT on consecutive days, underwent flexor tendon repair surgery 10 days after the final DT injection, and were harvested at 14 and 28 days post-repair (A). Immunofluorescence of WT and ScxLinDTR repair tendons 14 and 28 days post-repair to assess F4/80+ macrophages (B), αSMA+ myofibroblasts (C), and S100a4+ cells (D). Tendon is outlined by white dotted line and scar tissue by yellow dotted line. Green boxes indicate location of higher magnification images. Examples of positive stain indicated by white arrows, while examples of auto-fluorescent blood cells and α-SMA+ blood vessels indicated by green arrows. Quantification of F4/80 (A’ and A’’), αSMA (B’ and B’’), and S100a4 (C’ and C’’) fluorescence. N = 4 per genotype per timepoint. Student’s t-test used to assess statistical significance between genotypes at a given timepoint, except for D28 F4/80 and S100a4 which required a Mann-Whitney test. * indicates p<0.05.

Identification of differentially expressed genes following ScxLin depletion

To further investigate the mechanisms driving altered biomechanical properties of healing ScxLinDTR tendons, bulk RNAseq analysis was conducted on days 14 and 28 post-repair samples from ScxLinDTR and WT. A total of three biological replicates per genotype per timepoint were submitted for analysis. At 14 days post-repair, 47 genes were up-regulated, and 313 genes were down-regulated in ScxLinDTR relative to WT (Figure 6A,B). At 28 days post-repair, 1237 genes were up-regulated, and 1296 genes were down-regulated in ScxLinDTR relative to WT (Figure 6C,D). Based on both the low number of differentially expressed genes (DEGs) and the lack of mechanical phenotype at D14 between WT and ScxLinDTR (Figure 2), our primary RNAseq analyses were focused on day 28.

Figure 6 with 1 supplement see all
Bulk RNA sequencing reveals differences between ScxLinDTR and wild-type healing flexor tendons at 14 and 28 days post-repair.

Representation of differentially expressed genes (DEGs) at 14 (A, B) and 28 (C, D) days post-repair. Volcano plots (A, C) depict significantly upregulated DEGs as red dots and significantly downregulated DEGs as green dots. DEGs are consider significant when the multiple test corrected (adjusted) p-value is < 0.05. The dotted line represents the unadjusted p-value of 0.05. Heat maps (B, D) depict all significant DEGs, with the data representing the regularized log transformation of the normalized count data.

ScxLin depletion drives differential expression of matrix-related gene expression

The day 28 dataset was analyzed using ingenuity pathway analysis (IPA) software. We hypothesized that a change in the matrix composition could be driving the altered biomechanical properties detected at day 28 post-repair in ScxLinDTR animals, possibly driven by the elevated myofibroblast content. To examine the biological effects of ScxLin cell depletion at the transcriptional level, downstream effects analysis was performed by utilizing the core analysis in IPA where activation states were assigned for biological processes with p-value<0.05 and z-score > 2 (Table 1). Included in the significantly increased disease and function annotations was ‘Fibrosis’ (p=3.37E-07, Z = 2.397, Table 1). As we have already demonstrated that depletion of ScxLin cells is not driving increased fibrotic healing (Figure 2B,C), we then examined if the significantly increased ‘Fibrotic’ annotation was indicative of altered expression of specific matrix-associated genes. Utilizing the comprehensive review of the matrisome by Hynes and Naba, 2012, genes coding for collagens, proteoglycans, basement membrane proteins, and glycoproteins were compiled and examined (Table 2). Many matrix-related genes were significantly increased in ScxLinDTR repairs relative to WT littermates at day 28 post-repair (ex. Col1a2, Col3a1, Col8a1, Epyc, Thbs4, complete list in Table 2), while other matrix components were significantly decreased (ex. Col6a4, Col9a1, Fras1, complete list in Table 2). To both validate the RNAseq data and to define the spatial localization of different matrix components, we performed immunofluorescence for Decorin (Dcn), Thbs4 and Mfap5. Both the staining intensity and staining extent of Dcn, Thbs4 and Mfap5 were substantially increased in D28 ScxLinDTR repairs, relative to WT (Figure 6—figure supplement 1), consistent with the increases in these matrix components identified by RNAseq (Table 2).

Table 1
Ingenuity pathway analysis downstream effects - Disease and Functions.

Table of all disease and function annotations marked as significant (p<0.05 and ABS(Z-score)>2) using IPA core analysis for ScxLinDTR vs WT at day 28 post-repair.

Disease or function annotationp-valueActivation stateZ-Score
Down-RegulatedContractility of skeletal muscle4.96E-16Decreased−3.595
Abnormal bone density0.000000572Decreased−3.299
Contractility of muscle8.91E-16Decreased−2.636
Intestinal cancer6.76E-46Decreased−2.561
Bleeding0.00000075Decreased−2.424
Malignant neoplasm of large intestine8.42E-46Decreased−2.343
Colorectal cancer6.67E-23Decreased−2.343
Large intestine neoplasm2.97E-46Decreased−2.256
Colorectal tumor1.28E-23Decreased−2.256
Function of muscle6.93E-14Decreased−2.245
Consumption of oxygen0.000000942Decreased−2.237
Function of skeletal muscle1.59E-08Decreased−2.186
Intestinal tumor9.72E-47Decreased−2.144
Development of lung carcinoma0.000000663Decreased−2.012
Up-regulatedCell movement4.98E-22Increased4.735
Migration of cells6.44E-17Increased4.733
Cell movement of tumor cell lines8.66E-10Increased4.343
Reorganization of cytoskeleton0.000000218Increased4.296
Migration of tumor cell lines0.00000032Increased4.162
Engulfment of cells1.05E-09Increased4.057
Endocytosis1.2E-11Increased3.821
Leukocyte migration4.37E-09Increased3.821
Cell movement of blood cells4.04E-09Increased3.818
Homing of cells0.00000055Increased3.792
Formation of cellular protrusions1.53E-14Increased3.669
Glucose metabolism disorder5.46E-10Increased3.516
Invasion of cells9.21E-08Increased3.396
Organization of cytoplasm3.25E-26Increased3.384
Organization of cytoskeleton2.94E-20Increased3.384
Cell movement of leukocytes7.74E-08Increased3.38
Endocytosis by eukaryotic cells1.16E-08Increased3.371
Engulfment of tumor cell lines0.00000031Increased3.348
Proliferation of neuronal cells3.67E-09Increased3.313
Metabolism of carbohydrate1.94E-13Increased3.285
Formation of lamellipodia0.000000932Increased3.121
Cell movement of breast cancer cell lines0.000000603Increased3.103
Cell movement of fibroblast cell lines3.42E-08Increased3.083
Growth of neurites6.25E-09Increased2.981
Microtubule dynamics5.21E-18Increased2.974
Cell spreading3.31E-11Increased2.875
Formation of filopodia0.000000177Increased2.873
Cell movement of connective tissue cells0.000000415Increased2.792
Concentration of lipid3.14E-08Increased2.779
Progressive neurological disorder6.74E-10Increased2.671
Outgrowth of neurites7.31E-08Increased2.662
Production of reactive oxygen species6.22E-12Increased2.625
Quantity of macropinosomes0.000000865Increased2.621
Neuromuscular disease1.47E-15Increased2.619
Progressive myopathy1.28E-11Increased2.611
Synthesis of carbohydrate0.000000499Increased2.553
Outgrowth of cells0.000000025Increased2.52
Advanced malignant tumor0.000000208Increased2.517
Differentiation of connective tissue cells1.48E-09Increased2.512
Secondary tumor0.000000643Increased2.461
Arrhythmia2.14E-08Increased2.4
Fibrosis0.000000337Increased2.397
Extension of cellular protrusions0.00000098Increased2.371
Invasive tumor2.47E-08Increased2.345
Synthesis of reactive oxygen species3.6E-14Increased2.312
Organization of actin cytoskeleton4.37E-08Increased2.298
Disassembly of filaments0.000000544Increased2.27
Metabolism of reactive oxygen species2.61E-15Increased2.269
Cancer of cells9.31E-14Increased2.254
Response of tumor cell lines2.31E-08Increased2.231
Morphogenesis of neurons1.44E-12Increased2.224
Neuritogenesis1.86E-12Increased2.224
Neoplasia of cells1.76E-16Increased2.221
Quantity of metal0.0000002Increased2.198
Ruffling0.000000297Increased2.157
Tubulation of cells0.00000126Increased2.132
Angiogenesis4.02E-15Increased2.109
Hereditary myopathy1.58E-23Increased2.104
Dystrophy of muscle1.85E-11Increased2.104
Development of vasculature2.67E-16Increased2.06
Growth of axons0.000000724Increased2.017
Migration of fibroblast cell lines0.000000461Increased2.002
Table 2
Regulation of matrix components in ScxLinDTR healing tendons at day 28.

Expression level, fold change, and adjusted p-value of key matrix-related genes in ScxLinDTR tendons vs WT at day 28 post-repair generated from bulk RNA-seq. Orange color indicative of increased expression and blue color indicative of decreased expression.

GeneBaseMeanFold change (Log2)p-adj
Collagens
Col1a1742982.8830.2750.117206
Col1a2720257.2330.3590.033355
Col2a1461.9650.1760.700754
Col3a1748380.280.6450.005567
Col4a135181.7730.170.559295
Col4a231627.0650.2650.206116
Col4a349.144−0.0090.988635
Col4a4143.520.2320.589161
Col4a5284.0590.3720.440755
Col4a652.693−0.0930.904823
Col5a1118821.8570.4950.031814
Col5a2105087.0420.460.092164
Col5a335303.6660.4180.06355
Col6a1122410.4230.4060.00366
Col6a2122396.3220.4550.003183
Col6a355660.6290.5450.021485
Col6a433.599−0.8790.0385
Col6a572.0460.5990.286724
Col6a651.234−0.070.901424
Col7a12215.503−0.5320.06829
Col8a16553.5620.9830.025451
Col8a24787.689−0.1280.727708
Col9a1390.971−1.3280.000857
Col9a250.18−0.0710.911757
Col9a352.9110.1770.713852
Col10a14.5850.254N/A*
Col11a113080.117−0.2390.496276
Col11a21974.589−0.6710.028981
Col12a134791.768−0.2040.545342
Col13a1308.833−0.0850.853468
Col14a17279.6530.6930.00124
Col15a18889.5170.3850.296359
Col16a124279.3470.2370.252556
Col17a1269.420.002N/A*
Col18a114754.8860.3910.169961
Col19a14.7380.133N/A*
Col20a1438.526−0.70.060808
Col22a12064−0.7970.022217
Col23a14457.185−0.6360.025351
Col24a1966.961−0.350.047102
Col25a1118.0330.0220.975523
Col26a178.90.3710.475453
Col27a14062.7840.4060.253217
Col28a1579.4410.2910.653743
ECM proteoglycans
Hspg247213.8670.2480.108795
Aspn19143.1910.8390.007556
Bgn151131.1530.2510.256881
Dcn95817.7180.6541.26E-05
Fmod132295.7480.1720.683103
Kera5207.2310.5220.154489
Lum43114.0990.3530.130133
Omd65.3240.560.151367
Prelp18853.1240.3810.057946
Epyc133.9351.1870.003793
Ogn5228.6560.6360.004079
Optc30.1930.0420.957221
Chad2558.426−0.1330.813626
Chadl146.3810.3770.295306
Nyx21.0480.4430.360179
Podn1763.3690.4320.131578
Podnl1216.612−0.0870.872241
Acan8738.435−0.4070.194754
Bcan125.867−0.7640.014902
Ncan1.472−0.047N/A*
Vcan5214.4630.4310.165678
Hapln1109.2010.0360.955062
Hapln28.84−0.024N/A*
Hapln331.920.0390.949531
Hapln4128.6990.0530.926101
Prg22.717−0.083N/A*
Spock112.3870.674N/A*
Spock2621.4890.0270.956371
Spock334.16−0.0340.963822
Prg426463.024−0.1910.759532
Srgn1500.611−0.080.843956
Impg249.211−0.1960.676547
Esm1120.466−0.020.971642
Basement membrane components
Lama13.29−0.445N/A*
Lama23430.8870.3610.201711
Lama3104.957−0.3030.484039
Lama47706.8190.340.081572
Lama53390.382−0.2370.430648
Lamb111272.5070.2360.389409
Lamb213730.50.3430.092701
Lamb377.394−0.060.924993
Lamc115292.130.3130.119095
Lamc2424.2660.3930.064032
Lamc321.5970.0930.886633
Nid112717.3060.5390.02301
Nid22799.8170.1790.586124
Colq358.747−0.3790.341156
Major ECM glycoproteins
Eln17607.4640.5180.270402
Emilin16960.5390.1240.684821
Emilin24405.4630.5120.061758
Emilin3595.3560.8570.005562
Emid1550.6850.2520.631487
Fbln12214.0230.230.320082
Fbln248333.4860.1270.615495
Fbln52725.6750.3260.023116
Fbln74562.4090.7840.00233
Efemp11308.9390.5430.077648
Efemp24858.9910.2220.11327
Fbn136959.1960.6680.005087
Fbn22856.248−0.0080.983466
Fn1510510.0530.3070.223595
Fras1430.729−0.8370.013797
Gldn464.820.4240.390168
Hmcn11034.8060.5890.036404
Hmcn26034.9220.2520.456036
Ibsp2.365−0.125N/A*
Matn10.6710.074N/A*
Matn26213.3810.5030.015668
Matn3201.369−0.5810.155595
Matn42796.498−0.2450.6134
Mfap1a536.8870.1010.639732
Mfap1b398.9580.030.902813
Mfap23608.7270.4650.013072
Mfap31213.9240.1190.519005
Mfap44217.930.40.210242
Mfap514548.0080.7535.34E-05
Mmrn1439.1830.4920.262418
Mmrn21557.776−0.2010.519886
Npnt1084.388−0.4530.209896
Papln16.561−0.1220.859483
Postn102294.8710.5910.048873
Sparc331616.1770.4790.000296
Sparcl112872.967−0.0270.959288
Spp129368.7240.2340.637831
Srpx25778.7940.3620.077229
Tnc28609.2970.3780.326135
Tnn5098.501−0.6230.126873
Tnr19.2710.4820.385953
Tnxa6.3230.944N/A*
Tnxb18888.4250.7810.015263
Thbs114477.7980.490.099587
Thbs236613.7870.390.149809
Thbs317595.8550.2670.193149
Thbs4203095.5420.8415.69E-05
Comp35501.7590.2930.10314

Effects on cell motility, cytoskeletal organization, metabolism, and oxidative stress identified in ScxLinDTR repair tendons

Utilizing the downstream effects analysis described above (Table 1), we next examined other significantly altered biological functions in ScxLinDTR tendon repairs relative to WT. The inhibited pathways indicated decreased contractility and function of muscle (ex. ‘Contractility of skeletal muscle,’ p=4.96E-16, Z = −3.595) and decreased oxygen consumption (‘Consumption of Oxygen,’ p=9.42E-07, Z = −2.237) in ScxLinDTR tendons relative to wildtype controls at day 28 post-repair (Table 1). While only 14 disease and function annotations were found to be inhibited, 62 annotations were found to be significantly activated. Interestingly, in addition to ‘Fibrosis,’ the activated biological pathways indicated increased cell movement and migration (ex. ‘Migration of cells,’ p=6.44E-17, Z = 4.733), reorganization of the cytoskeletal network (ex. ‘Organization of cytoskeleton,’ p=2.94E-20, Z = 3.384), metabolism disorders (ex. ‘Glucose metabolism disorder,’ p=5.46E-10, Z = 3.516), and production of reactive oxygen species (ex. ‘Production of reactive oxygen species,’ p=6.22E-12, Z = 2.625) (Table 1).

Differentially enriched canonical pathways following ScxLin cell depletion prior to flexor tendon repair

To better understand signaling cascades that could be driving alterations in ScxLinDTR healing, enriched canonical pathways were identified using IPA core analysis. Nineteen canonical pathways were identified, where 13 were positively enriched (activated) and 6 were negatively enriched (inhibited/suppressed) (Figure 7, Table 3). Consistent with the metabolism disorders identified from the downstream effects analysis (Table 2), canonical pathways related to metabolism were negatively enriched (ex. ‘Oxidative Phosphorylation,’ -log(p)=7.6, Z = −5.303; additional pathways provided in Figure 7 and Table 3). Additionally, calcium signaling was found to be negatively enriched (‘Calcium Signaling,’ -log(p)=5.72, Z = −2.335) (Figure 7 and Table 3).

Canonical pathways positively and negatively enriched in ScxLinDTR healing tendons at day 28 post-repair.

Ingenuity pathway analysis was utilized to determine positively and negatively significantly enriched pathways in ScxLinDTR healing tendons at day 28 post-repair. Canonical pathways were considered significant if p<0.05 and ABS(Z-score)>2. The orange color indicates pathways that are significantly, positively enriched (‘activated’), while the blue color indicates significantly, negatively enriched (‘inhibited’) pathways. The orange dotted line represents -log (1.3)=0.05, indicating the p-value cut-off.

Table 3
Ingenuity pathway analysis canonical pathways.

All enriched pathways marked as significant (-log(p-value)>1.3 and ABS(Z-score)>2) using IPA core analysis for ScxLinDTR vs WT at day 28 post-repair.

Canonical pathway-log(p)Z-Score
Negatively enrichedOxidative phosphorylation7.6−5.303
TCA Cycle II (Eukaryotic)5.96−3.464
White Adipose Tissue Browning Pathway3.23−2.353
Calcium Signaling5.72−2.335
Glycolysis I4.63−2.111
Gluconeogenesis I4.63−2.111
Positively enrichedEphrin Receptor Signaling6.133.888
Rac Signaling3.853.674
Actin Nucleation by ARP-WASP Complex5.163.441
Fcγ Receptor-mediated Phagocytosis in Macrophages and Monocytes3.773.411
Remodeling of Epithelial Adherens Junctions6.233.162
Regulation of Actin-based Motility by Rho5.753.128
Sphingosine-1-phosphate Signaling3.942.711
Ephrin B Signaling3.092.53
RhoA Signaling3.972.502
Glioma Invasiveness Signaling2.132.496
fMLP Signaling in Neutrophils4.442.449
Actin Cytoskeleton Signaling7.862.214
Production of Nitric Oxide and Reactive Oxygen Species in Macrophages2.012.117

Consistent with reorganization of the cytoskeletal network identified from the downstream effects analysis (Table 2), numerous canonical pathways related to actin nucleation and polymerization were positively enriched (Figure 7 and Table 3). These include ‘Rac Signaling’ (-log(p)=3.85, Z = 3.674), ‘RhoA Signaling’ (-log (p) = 3.97, Z = 2.502), and ‘Ephrin B Signaling’ (-log(p)=3.09, Z = 2.53). Other pathways that are significantly enriched in ScxLinDTR day 28 post-repair tendons include, but are not limited to, ‘sphingosine-1-phosphate signaling’ (-log(p)=3.94, Z = 2.711) and ‘production of nitric oxide and reactive oxygen species in macrophages’ (-log(p)=2.01, Z = 2.117) (Figure 7 and Table 3).

Identification of possible upstream regulators driving altered ScxLinDTR flexor tendon healing

To identify key molecules that may be driving ScxLinDTR tendon healing at day 28 post-repair, predicted upstream regulators were identified using IPA core analysis. Eight possible activated upstream regulators and four inhibited upstream regulators were identified (Table 4). The eight activated regulators included the calcium-binding protein S100a4, peptidase F2, receptors BTNL2 and F2R, transcription factors EBF2 and SOX2, the kinase NTRK2, and growth factor FGF2. The four inhibited regulators included enzyme LDHB and transcription factors FOXO4, MEF2C, and SMYD1.

Table 4
Ingenuity pathway analysis upstream regulators.

All possible upstream regulators where expression log ratio > 0.5, ABS(Z-score)>2, p-value<0.05, and agreement between predicted activation state and directionality of regulator’s gene expression, compiled using IPA core analysis for ScxLinDTR vs WT at day 28 post-repair.

Upstream regulatorExpression log ratioPredicted activation stateActivation Z-scorep-value of overlap
Activated in ScxLinDTR
S100A40.518Activated2.9460.00456
F20.533Activated2.6060.000134
BTNL20.805Activated2.3240.012
EBF20.657Activated2.2230.000738
F2R0.599Activated2.220.0243
NTRK20.608Activated2.1370.012
SOX21.138Activated2.0710.0000143
FGF20.544Activated2.0170.000393
Inhibited in ScxLinDTR
FOXO4−0.501Inhibited−2.6970.0236
MEF2C−0.905Inhibited−2.5773.06E-08
SMYD1−0.836Inhibited−2.2192.63E-12
LDHB−0.609Inhibited−2.2190.000241

Ablation of ScxLin tendon cells does not significantly affect tendon post-natal growth 3 months post-ablation

In addition to roles in tendon healing, Scx expression is required for appropriate tendon development and growth processes (Murchison et al., 2007; Gumucio et al., 2020). However, the role of Scx+ cells during post-natal growth and adult homeostasis have not been evaluated. Local injection of DT into pre-pubescent mice (3–4 weeks old) resulted in 55% depletion of tendon cells in uninjured ScxLinDTR FDL tendons relative to WT littermates (p=0.0007) (Figure 8A–C). To assess the requirement for tendon cells in post-natal growth, pre-pubescent mice who were still undergoing periods of rapid growth were injected with DT and harvested at the 3 month timepoint (ScxLinDTR,3weeks) (Figure 8D). ScxLinDTR,3weeks tendons did not exhibit an influx of F4/80+ macrophages or tendon cell differentiation into αSMA+ myofibroblasts in either genotype (Figure 8E); however, ScxLinDTR mice had a 42.5% decreased tendon cell number relative to WT littermates (p=0.0454) demonstrating that the tendon cell environment was not repopulated following depletion (Figure 8E and F). There were no significant changes in MTP range of motion, gliding resistance, stiffness, or maximum load at failure between groups (Figure 8G–J). Second harmonic generation revealed no significant differences in collagen fibril dispersion between genotypes (Figure 8K and L). Taken together, these data suggest that ScxLin cells are not required for early post-natal tendon growth.

Tendon cell ablation does not negatively affect post-natal tendon growth 3 months post-ablation.

Pre-pubescent mice (3–4 weeks old) received five hindpaw injections of DT and were harvested 10 days after the final injection to assess tendon cell depletion (ScxLinDTR) (A). Hindpaw sections from both WT and ScxLinDTR hindpaws (B). Quantification of WT and ScxLinDTR,3weeks tendon cell number in pre-pubescent mice (C). To assess effects of tendon cell depletion on post-natal tendon growth, mice received five hindpaw injections of DT on consecutive days at 3–4 weeks of age and were harvested uninjured 3 months later for biomechanical, gliding, and histological evaluation (ScxLinDTR,3weeks) (D). Co-immunofluorescence of F4/80 (macrophages) and αSMA (myofibroblasts) in uninjured WT and ScxLinDTR,3weeks tendons (E). Quantification of WT and ScxLinDTR,3weeks tendon cell number (F). Measurement of metatarsophalangeal (MTP) joint flexion angle (G), gliding resistance (H), stiffness (I), and maximum load at failure (J) of WT and ScxLinDTR,3weeks uninjured tendons. N = 7–10 per genotype. Second harmonic generation (K) and quantification (L) of collagen fibril dispersion of WT and ScxLinDTR,3weeks. N = 3 per genotype. Nuclei stained with DAPI. Tendon is outlined by white dotted lines. Student’s t-test used to assess statistical significance between genotypes. * indicates p<0.05, *** indicates p<0.001.

Ablation of ScxLin tendon cells significantly affected tendon homeostasis 3 months post-depletion

To assess the requirement for ScxLin tendon cells in maintaining adult tendon homeostasis, 10–12 week-old WT and ScxLinDTR mice were injected with DT to induce cell death and harvested after 3 months (ScxLinDTR,10weeks) (Figure 9A). ScxLinDTR,10weeks mice had a 62.3% decrease in tendon cell number relative to WT littermates (p<0.0001)(Figure 9B and C), demonstrating that tendon cell number had not rebounded in the three months since initial depletion (Figure 1). Interestingly, we consistently observed a significant accumulation of unidentified cells on the top and bottom regions of ScxLinDTR,10weeks tendons (Figure 9B). Quantification of the cellular density revealed a significant increase on the top (p<0.0001) and bottom (p<0.01) regions of the ScxLinDTR,10 weeks compared to WT littermates (Figure 9D). ScxLinDTR,10weeks tendons did not exhibit an influx of F4/80+ macrophages or tendon cell differentiation into αSMA+ myofibroblasts in either genotype (Figure 9E). Functionally, there were no significant changes in tendon gliding function or biomechanical properties between genotypes after 3 months (Figure 9F–I). However, second harmonic generation imaging revealed a significant increase in overall collagen fibril dispersion in ScxLinDTR,10weeks relative to WT littermates (WT: 8.327 degrees ± 0.39, ScxLinDTR,10weeks: 9.815 degrees ± 0.53, p=0.0393) (Figure 9J and K). Based on the spatial changes in cellularity (Figure 9D), we also quantified fibril dispersion in a more spatially specific manner (top, middle and bottom thirds of the tendon) (Figure 9—figure supplement 1A–D). Interestingly, the top third of ScxLinDTR,10weeks tendons showed a significant increase in dispersion compared to all regions of WT (Figure 9—figure supplement 1E). These data suggest a potential relationship between the increased cellularity and fibril dispersion levels on the top region of the ScxLinDTR FTs.

Figure 9 with 2 supplements see all
Tendon cell ablation negatively affected tendon homeostasis 3 months post-ablation.

Mice received five hindpaw injections of DT on consecutive days at 10–12 weeks of age and were harvested uninjured 3 months later for biomechanical, gliding, and histological evaluation (ScxLinDTR,10weeks) (A). Cellularity was assessed using DAPI (B) and quantified (C) 3 months after ScxLin cell depletion. H and E staining was used to better define the hypercellular regions near the tendon epitenon. Cell density was quantified at the top and bottom boundaries of the tendon (D). N = 7 per genotype. Co-immunofluorescence of F4/80 (macrophages) and αSMA (myofibroblasts) in uninjured WT and ScxLinDTR tendons (E). N = 3 per genotype. Measurement of metatarsophalangeal (MTP) joint flexion angle (F), gliding resistance (G), stiffness (H), and maximum load at failure (I) of WT and ScxLinDTR,10weeks uninjured tendons. N = 9–12 per genotype. Second harmonic generation (SHG) (J) and quantification (K) of collagen fibril dispersion of WT and ScxLinDTR,10weeks. N = 7–8 per genotype. Nuclei stained with DAPI. Tendon is outlined by white dotted lines. Student’s t-test used to assess statistical significance between genotypes. * indicates p<0.05.

Given these changes in collagen alignment, we further assessed the ECM structure using TEM. (Figure 9—figure supplement 2A and B). Collagen fibril diameter distribution was substantially altered between WT (median = 147.95, Q1 = 118.73, Q3 = 179.45) and ScxLinDTR,10weeks (median = 213.41, Q1 = 167.49, Q3 = 261.91) (Figure 9—figure supplement 2C). Collagen fibril diameter of ScxLinDTR,10weeks increased by 30.67% compared to WT (p<0.0001) (Figure 9—figure supplement 2D). The collagen fibril density of ScxLinDTR,10weeks decreased by 48.72% compared to WT (p<0.0001) (Figure 9—figure supplement 2E). Finally, both the WT and the ScxLinDTR,10weeks groups exhibited similar levels of fibril irregularity (p=0.9023) (Figure 9—figure supplement 1F).

Discussion

Previous work has established the importance of Scx expression in tendon development (Murchison et al., 2007), growth (Gumucio et al., 2020), and healing (Sakabe et al., 2018), but few studies have considered the direct contributions of ScxLin tendon cells to these processes. In the present study, we examined the function of ScxLin tendon cells during adult flexor tendon healing and made the surprising discovery that ScxLin cell depletion prior to tendon injury and repair significantly enhanced biomechanical properties by day 28 post-repair. We characterized key cell populations known to be important in healing and regeneration and utilized RNA-Seq to investigate the mechanism driving these biomechanical changes. Lastly, we examined the effects of ScxLin cell depletion on post-natal tendon growth and adult tendon homeostasis and found that ScxLin cells are required for maintenance of collagen ECM alignment and organization in adult, but not early post-natal flexor tendon growth.

Tendon cell depletion had surprisingly beneficial effects on healing, with biomechanical properties significantly increased at day 28 post-repair relative to WT, with no impact on gliding function. These results indicate that the improved biomechanical properties are likely not due to increased levels of disorganized matrix/scar within the healing tendon, but that the healing process may be shifted toward a regenerative phenotype. Equally striking is that the significant improvements in biomechanical properties did not emerge until 28 days post-repair, which is firmly into the remodeling phase of healing and is consistent with the low DEG number at day 14 post-repair relative to day 28. This suggests that ScxLin cells are important in the late proliferative-remodeling phases of healing and possibly enact their greatest effects by modulating the remodeling process. Consistent with this is the lack of difference in proportion of ScxLinAi9 cells at D14 between WT and ScxLinAi9DTR, while a significant decrease in ScxLinAi9 cells was observed in ScxLinAi9DTR tendon repairs, relative to WT, at D28. While further studies are needed to completely define this process, these data suggest that ‘new’ ScxLin cells (e.g. those that express Scx in response to tendon injury and are therefore labeled as ScxLinAi9, but that were not ScxLinAi9 at the time of depletion) may predominate during early healing, so the effects of depleting ScxLin prior to injury are minimal. However, by D28 these ‘new’ ScxLinAi9 cells may undergo apoptosis and be cleared during progressive tissue remodeling, such that the cells that remain at D28 are primarily derived from the ScxLinAi9 cells present in the adult tendon prior to injury. Therefore, the effects of ScxLinAi9DTR become more apparent and allow interrogation of the functional effects on healing of depleting adult tendon resident ScxLin cells by D28. Finally, it is also possible that there is a compensatory response by other tenocyte subpopulations. While we do not see any changes in S100a4+ cells, it is possible that there may be an increase in S100a4-lineage cells, which can express Scx during tendon healing (Best and Loiselle, 2019), or other cells.

Our RNA-Seq results revealed significant changes in matrix-related gene expression at day 28 post-repair, suggesting that the matrix composition of the healing tendon may be substantially altered in the absence of ScxLin cells, with a shift toward either a more regenerative or mature matrix composition. Future proteomics analysis and additional experimentation will be required to further examine the matrix differences between ScxLinDTR and WT healing flexor tendons. It has previously been reported that tendon strength remains significantly decreased relative to uninjured controls 63 days post-repair using this flexor tendon repair model in C57Bl/6J mice (Loiselle et al., 2009). Therefore, it could be beneficial to examine time points beyond day 28 to see if ScxLinDTR tendons reach equivalency with uninjured controls. The improved biomechanical properties contrasted with our original hypothesis which predicted that ScxLin cells would be necessary for proper healing and would contribute to formation of the bridging collagen tissue. Both ScxLinDTR and WT mice exhibited a collagen bridge at 14- and 28 days post-repair suggesting that the ScxLin cells targeted for depletion prior to injury are not the predominant cell population laying down this collagenous matrix. This is surprising due to studies demonstrating that Scx promotes collagen formation (Sakabe et al., 2018; Leéjard et al., 2007), highlighting the important distinction between ScxLin cells and active Scx expression.

While myofibroblast persistence is considered a primary driver of fibrosis (Hinz and Lagares, 2020), recent work has revealed that myofibroblasts are a highly heterogenous population with differences in pro-fibrotic markers, gene expression, and cross-linking ability, suggesting that myofibroblasts contribute to both fibrotic and regenerative processes (Shook et al., 2018). Despite the elevated myofibroblast presence in ScxLinDTR repairs at day 28, these tendons healed with increased biomechanical properties while experiencing no deficits in tendon range of motion, suggesting a regenerative rather than fibrotic healing process. Future work to comprehensively define the myofibroblast landscape in both WT and ScxLinDTR mice is necessary to determine if ScxLinDTR myofibroblasts are more ‘pro-regenerative’ than those present in WT repairs. Similarly, understanding how myofibroblast subtypes affect matrix deposition at the repair site represents an important area of future study. Related to this, previous work has shown that regeneration is mediated in large part by the immune response, including macrophages (Vagnozzi et al., 2020). However, we did not see any differences in overall macrophage content, or in pathways related to immune response in the RNAseq. Future work, including single-cell RNA sequencing studies will be important to provide a comprehensive understanding of how ScxLin cell depletion alters the overall cellular environment and how this dictates the functional changes that are observed.

The RNA-seq data suggests that healing ScxLinDTR tendons exhibit less overall metabolic activity compared to WT repairs at day 28. This is highlighted in IPA’s canonical pathways core analysis, where ‘Oxidative Phosphorylation,’ ‘TCA Cycle II,’ ‘Glycolysis I,’ and ‘Gluconeogenesis I’ were all predicted to be negatively enriched (‘suppressed’) (Figure 7, Table 3). Additionally, IPA’s disease and function core analysis predicts an inhibitory effect on ‘Consumption of oxygen,’ while simultaneously predicting ‘Metabolism of carbohydrate’ activation, increased ‘Synthesis of carbohydrate,’ and ‘Glucose metabolism disorder (Table 1).” It has previously been demonstrated in a variety of tissues that metabolic reprogramming is important for cellular differentiation and regeneration (Chen et al., 2019; De Santa et al., 2019; Osuma et al., 2018; Lai et al., 2019). In addition to the dysregulation of metabolism, our RNA-seq data also revealed an increase in 'Production of reactive oxygen species', 'Synthesis of reactive oxygen species', and 'Metabolism of reactive oxygen species (Table 1)'. Despite reactive oxygen species (ROS) being implicated in various pathologies (Schieber and Chandel, 2014), recent studies have also identified functional roles for ROS in regeneration (Santabárbara-Ruiz et al., 2019; Youm et al., 2019; Labit et al., 2018; Santabárbara-Ruiz et al., 2015). While little is currently understood about the role of metabolism and ROS in tendon healing, these data clearly identify how changes in the cellular composition of the healing tendon can shift both the metabolic profile and the healing program. As such, investigating metabolic reprogramming and the dynamic roles of ROS in acute tendon healing represents an exciting area of future work.

In addition to investigating the role of tendon cells during healing, we also utilized ScxLinDTR mice to assess how tendon cell ablation affected post-natal tendon growth (ScxLinDTR,3weeks) and adult tendon homeostasis (ScxLinDTR,10weeks). Although there were no significant differences in gliding function or biomechanical properties between ScxLinDTR and WT genotypes in either age group, ScxLinDTR,10weeks animals exhibited significantly increased overall collagen fibril dispersion, as well as significant changes in dispersion in specific regions of the tendon. While it is unclear what is driving these spatially-specific changes in fibril organization, it is notable that the most profound changes occur in the top epitenon region rather than endotenon, and that this area is consistent with altered and increased cellularity in ScxLinDTR,10weeks tendons, though the identity and function of these cells are as yet unknown. In addition to matrix alignment, ScxLinDTR, 10 weeks also increased collagen fibril diameter as measured by TEM. Interestingly, one potential interpretation of the TEM data relates to the loss of proteoglycans (PGs) on the surface of the collagen fibrils, which normally prevents lateral fusion of collagen fibrils. Numerous in vitro and in vivo studies have shown the importance of the PGs to maintain the normal collagen fibril structure. Using proteinase treatment to remove surface bound PGs, Graham et al., demonstrated lateral collagen fibril aggregation, resulting in fused fibrils with increased diameter (Graham et al., 2000). In addition, Decorin-/- tail tendons also demonstrated lateral fusion and increased fibril diameter, relative to WT (Danielson et al., 1997). While there is no direct evidence of PG production by Scx+ cells, Scx-/- hearts have decreased PG content (Barnette et al., 2013; Barnette et al., 2014), while Sakabe et al., found that Scx-/- tendons lacked fibromodulin production during tendon healing (Sakabe et al., 2018). Moreover, recent work has shown that Scx knockdown in adult equine tendon cells alters the PG environment (Paterson et al., 2020). Collectively, these data suggest that the increase in fibril diameter observed in ScxLinDTR,10weeks could be due to diminution of PG production following depletion of ScxLin cells. However, we have not directly tested this, and future proteomic studies will be required to determine if and which specific PGs may be altered following ScxLin cell depletion. Collectively, these data indicate that loss of ScxLin cells may be detrimental to tendon tissue maintenance, but that negative biomechanical effects are not yet manifested 3 months post-depletion. Future studies looking beyond the 3-month time-point will be informative to understand the role of tendon cells on maintenance of tendon tissue. It is also possible that negative effects on tendon biomechanical properties could be the result of ScxLin cell death in non-tendon tissues, such as muscle in the hindpaw (Mendias et al., 2012). Despite nearly identical depletion efficiencies between ScxLinDTR,3weeks and ScxLinDTR,10weeks animals, only ScxLinDTR,10weeks tendons exhibited collagen disorganization and differences in fibril size. This suggests possible differences in tendon cell sub-populations present during growth and homeostasis, as well as the potential contribution of extrinsic progenitor populations to tendon growth (Dyment et al., 2014), such that depletion could differentially impact post-natal growth and adult tendon homeostasis. Moreover, it is possible that ScxLinDTR,3weeks did not disrupt tendon growth due to more rapid compensation by non-depleted ScxLin cells, or the addition of ‘new’ ScxLin cells during this period of rapid tissue growth.

We had initially planned to deplete tendon cells using the inducible Scx-CreERT2 crossed to the diphtheria toxin A mouse (Voehringer et al., 2008), but insufficient recombination occurred, and targeted cell death was not achieved. Therefore, to successfully deplete tendon cells, Scx-Cre mice were crossed to a diphtheria toxin receptor mouse model (ScxLinDTR) (Buch et al., 2005). The initial attempt to deplete cells using this model employed a series of intraperitoneal injections (200 ng/day) and resulted in all ScxLinDTR mice dying within four days of the initial injection while WT animals were unaffected. ScxLinDTR mice likely died due to apoptosis of non-tendon/ligament associated ScxLin cells. For example, it has previously been shown that Scx is expressed in the lung (Pryce et al., 2007; Perez et al., 2003), kidney (Pryce et al., 2007), muscle (Mendias et al., 2012), and brain, (Perez et al., 2003) among others. Therefore, to successfully utilize this model of cell depletion while simultaneously preventing ScxLinDTR-associated death, a series of low-dose (20 ng/day) local hind paw injections were administered. We successfully utilized this model of in vivo tendon cell depletion and reached a depletion level of 57%. Therefore, the ScxLinDTR data should be viewed as a model of partial depletion as ~40% of tendon cells remained using this approach. Future work utilizing increased or prolonged DT treatment could be attempted to obtain more complete tendon cell depletion, and to determine if the beneficial effects of tendon cell depletion may be reversed with complete ablation of ScxLin cells from the tendon. Additionally, our current ScxLinDTR ablation model targets cells prior to injury and thus does not target any cells that turn on Scx after injury. It has previously been shown that Scx knockout during Achilles tendon healing drives incomplete remodeling of type III collagen to type I collagen (Sakabe et al., 2018). As such, it could be interesting to assess if depletion of cells actively expressing Scx during healing resulted in deleterious effects on healing. However, local DT injection into the hind paw following repair surgery represents a technical challenge as repeated injections to the healing tendon is likely to disrupt the normal healing process. Moreover, the overlap between those ScxLin cells in the adult tendon prior to injury, and those that express Scx after injury is not yet defined. Finally, while there are likely to be some differences between tendons in terms of the functional contribution of ScxLin cells to healing process, the type of injury model used is also likely to impact data interpretation. For example, adult ScxLin cells do not contribute to tissue bridging in a model of Achilles tendon transection without repair (Howell et al., 2017), but are found in the bridging tissue following flexor tendon injury and repair (Best and Loiselle, 2019). As such, future studies will be needed to clarify if, or how, the functional contribution of ScxLin cells differ between tendons and injury models.

Altogether, these data demonstrate that ScxLin cell depletion is beneficial for tendon healing as it increases biomechanical properties several weeks after repair, possibly due to alterations in matrix composition, deposition, and/ or remodeling by αSMA+ myofibroblasts. Moreover, we have identified several molecular pathways that are altered following ScxLin cell depletion, such as changes in metabolism, cell motility, and the cytoskeleton, which may represent important targets for successfully modulating the healing process. Finally, ScxLin depletion does not significantly disrupt the biomechanical properties of tendons during early post-natal growth or adult tendon homeostasis within three months of ScxLin cell depletion. However, given the changes in matrix alignment and organization in adult ScxLinDTR tendons, it is possible that mechanical changes may occur following longer periods of depletion. Understanding the complex nature of tendon cells during healing, growth, and homeostasis can provide us with insights for driving regenerative healing and healthy tendon maintenance in the future.

Materials and methods

Key resources table
Reagent type (species)
or resource
DesignationSource or referenceIdentifiersAdditional information
Genetic reagent (Mus musculus)Scx-CreDr. Ronen SchweitzerMGI:5317938Referred to as ScxLinin manuscript
Genetic reagent (Mus musculus)C57BL/6-Gt(ROSA)26Sortm1(HBEGF)Awai/J
(Rosa-DTRLSL)
Jackson LaboratoryStock #: 007900
RRID:IMSR_JAX:007900
Referred to as ScxLinDTRin manuscript
Genetic reagent (Mus musculus)B6.Cg-Gt(ROSA)26Sortm9(CAG-tdTomato)Hze/J (ROSA-Ai9)Jackson LaboratoryStock #: 007909
RRID:IMSR_JAX:007909
Referred to as ScxLinAi9or ScxLinAi9DTR in manuscript
AntibodyAnti-SCXA
(rabbit polyclonal)
AbcamCatalog #: ab58655
RRID:AB_882467
(1:500)
AntibodyAnti-S100a4 (rabbit monoclonal)AbcamCatalog #: ab197896
RRID:AB_2728774
(1:2000)
AntibodyAnti-cleaved caspase 3
(rabbit polyclonal)
Cell Signalling TechnologyCatalog #: 9661
RRID:AB_2341188
(1:100)
AntibodyAnti-PCNA
(mouse monoclonal)
AbcamCatalog #: ab29
RRID:AB_303394
(1:100)
AntibodyAnti-F4/80
(rabbit polyclonal)
Santa Cruz BiotechnologyCatalog #: sc-26643
RRID:AB_2098331
(1:500)
AntibodyAnti-THBS4 (rabbit monoclonal)AbcamCatalog #: ab263898(1:250)
AntibodyAnti-MFAP5 (rabbit monoclonal)AbcamCatalog #: ab203828(1:2000)
AntibodyAnti-Decorin
(Rabbit polyclonal)
AbcamCatalog #: ab175404(1:250)
AntibodyAnti-alpha-SMA-Cy3
(mouse monoclonal)
Sigma-AldrichCatalog #: C6198
RRID:AB_476856
(1:200)
AntibodyAnti-alpha-SMA-FITC
(mouse monoclonal)
Sigma-AldrichCatalog #: F3777
RRID:AB_476977
(1:500)
AntibodyRhodamine Red-X (RRX) AffiniPure F(ab')₂ Fragment Donkey Anti-Rabbit IgG (H+L)
(Donkey polyclonal)
Jackson ImmunoResearchCatalog #: 711-296-152(1:200)
AntibodyAlexa Fluor 488 AffiniPure F(ab')₂ Fragment Donkey Anti-Goat IgG (H+L)
(Donkey polyclonal)
Jackson ImmunoResearchCatalog #: 705-546-147
RRID:AB_2340430
(1:200)
AntibodyRhodamine Red-X (RRX) AffiniPure F(ab')₂ Fragment Donkey Anti-Mouse IgG (H+L)
(Donkey polyclonal)
Jackson ImmunoResearchCatalog #: 715-296-150
RRID:AB_2340834
(1:200)
Chemical Compound, DrugDiphtheria Toxin (DT)Millipore SigmaD0564-1MG20 ng DT / injection
Software, algorithmsGraphPad Prism softwareGraphPad Prism (https://graphpad.com)Version 7.02
Software, algorithmsOlyVIA softwareOlympus (https://www.olympus-lifescience.com/en/support/downloads/)Version 2.9
Software, algorithmsImageJ softwareImageJ (http://imagej.nih.gov/ij/)

Mice

Scx-Cre mice were generously provided by Dr. Ronen Schweitzer. ROSA-Ai9F/F (#007909), and Rosa-DTRLSL (#007900) mice were obtained from the Jackson Laboratory (Bar Harbor, ME, USA). ROSA-Ai9LSL mice express Tomato red fluorescence in the presence of Cre-mediated recombination (Madisen et al., 2010). Scx-Cre mice were crossed to ROSA-Ai9F/F mice to label Scx lineage cells (ScxLin). Diphtheria toxin receptor (DTRLSL) mice can be utilized to temporally ablate cell populations driven by a non-inducible Cre driver (Buch et al., 2005). In short, expression of the diphtheria toxin receptor is inhibited prior to Cre-mediated recombination due to the presence of a STOP cassette flanked by loxp site (Loxp-STOP-Loxp; LSL). Following Cre-mediated recombination the STOP cassette is deleted, resulting in expression of DT receptor, in this case specifically on ScxLin cells. As such, administration of diphtheria toxin (DT) to these mice results in targeted cell death of ScxLin cells. Scx-Cre mice were crossed to DTRLSL animals to generate a model of ScxLin tendon cell depletion (ScxLinDTR) and Scx-Cre-; DTRLSL littermates were used as wild-type (WT) controls. To simultaneously deplete and visualize ScxLin tendon cells, Scx-Cre; Ai9F/F mice were crossed to the DTRLSL to generate a model of Scx-Cre+; Ai9F/+; DTRF/+ littermates (ScxLinAi9DTR) and Scx-Cre+; Ai9F/+; DTR+/+ (ScxLinAi9)were used as WT to visualize ScxLinAi9 cells without depletion. All mouse studies were performed with 10–12 week-old male and female mice except where otherwise noted (Figure 7 3–4-week-old male and female mice). All mouse work (injections, surgeries, harvests) were performed in the morning. Mice were kept in a 12 hr light/dark cycle.

Flexor tendon repair

Request a detailed protocol

Complete transection and repair of the murine flexor digitorum longus (FDL) tendon was performed as previously described (Ackerman and Loiselle, 2016). Mice received a 15–20 μg injection of sustained-release buprenorphine. Mice were anesthetized with Ketamine (60 mg/kg) and Xylazine (4 mg/kg). To reduce chances of rupture at the repair site, the FDL tendon was first transected at the myotendinous junction and the skin was closed with a 5–0 suture. This MTJ transection results in a transient decrease in tendon loading, with progressive reintegration of the MTJ observed by D7-10 post-surgery. As shown by the absence of αSMA staining in the uninjured tendon adjacent to the repair (Figure 5—figure supplement 1), this transient alteration in loading does not induce a widespread tendon response such as degeneration, remodeling, or cellular activation. Next, a small incision was made to the posterior surface of the right hind paw, the FDL tendon was isolated from surrounding tissue and completely transected. The tendon was repaired using 8–0 suture and the skin was closed with a 5–0 suture. Animals resumed prior cage activity, food intake, and water consumption immediately following surgery.

Quantification of tendon cell depletion and ScxLincells

Request a detailed protocol

ScxLinDTR mice were injected with 20 ng of diphtheria toxin (DT) for five consecutive days (100 ng total DT). Uninjured hind paws were harvested 10 days after the final injection for frozen sectioning. ScxLin hind paws were harvested uninjured hind paws were fixed in 10% NBF for 24 hr, decalcified in 14% EDTA for four days, and processed in 30% sucrose to cryo-protect the tissue. Samples were embedded using Cryomatrix (Thermo Fisher Scientific, Waltham, MA, USA) and sectioned into 8 mm sagittal sections using an established cryotape-transfer method (Dyment et al., 2016). Sections were stained with DAPI to visualize nuclei and imaged using a VS120 Virtual Slide Microscope (Olympus, Waltham, MA). Using Image J, a region of interest (ROI) was drawn around the tendon and an area was obtained. For ScxLinDTR and WT littermate mice, nuclei within the ROI were manually counted and total nuclei number was normalized to area. For ScxLin mice, fluorescent cells in uninjured sections were manually counted and ScxLin+ cells reported as a percentage of total cells counted in each section. An n = 3–4 (adult mice) or n = 3–5 (young mice) was used for quantification.

Paraffin histology and immunofluorescence

Request a detailed protocol

ScxLinDTR hind paws were harvested 10 days after the final DT injection for homeostasis studies, and at 14- and 28 days post-repair. Additionally, uninjured ScxLinDTR hind paws from adult (10–12 weeks) and pre-pubescent (3–4 weeks) mice were harvested 3 months following the final DT injection to assess effects of tendon cell depletion on tendon growth and homeostasis. Hind paws were fixed in 10% neutral buffered formalin (NBF) at room temperature for 72 hr and were subsequently decalcified in Webb Jee EDTA (pH 7.2–7.4) for 7 days at room temperature, processed, and embedded in paraffin. Three-micron sagittal sections were utilized for analysis, except for the 3 month study which were cut at 5-µm to facilitate SHG imaging. ScxLinDTR repair sections were stained with Alcian blue/hematoxylin and Orange G (ABHOG) or Hematoxylin and eosin (H and E) to assess tissue morphology and cellularity, and Masson’s Trichrome to assess collagen deposition. For immunofluorescent staining, sections were stained with Cleaved Caspase 3 (1:100, Cat#: 9661, Cell Signaling, Danvers, MA), PCNA (1:100, Cat#:ab29, Abcam, Cambridge, MA), F4/80 (1:500, Cat#: sc-26643, Santa Cruz, Dallas, TX), α-SMA-CY3 (1:200, Cat#: C6198, Sigma Life Sciences, St. Louis, MO), α-SMA-FITC (1:500, Cat#: F3777, Sigma Life Sciences, St. Louis, MO), S100a4 (1:2000, Cat#: ab197896, Abcam, Cambridge, MA), SCXA (1:500, ab58655, Abcam, Cambridge, MA), Decorin (1:250, Cat # ab175404, Abcam), Thbs4 (1:250, Cat # ab263898, Abcam), Mfap5 (1: 2000, Cat # ab203828, Abcam), or tdTomato (1:500, Cat#: AB8181, SICGEN, Cantanhede, Portugal). Sections were counterstained with the nuclear DAPI stain and imaged with a VS120 Virtual Slide Microscope (Olympus, Waltham, MA).

Quantification of fluorescence

Request a detailed protocol

Fluorescent images scanned by the virtual slide scanner were quantified using Visiopharm image analysis software v.6.7.0.2590 (Visiopharm, Hørsholm, Denmark). Automatic segmentation via a threshold classifier was utilized to define and quantify specific cell populations based on fluorescence. An ROI was drawn to encapsulate both the scar tissue and tendon stubs. The area of fluorescent signal was determined and normalized to the total ROI area to determine percentages of each cell type. An n = 4 was used for quantification. Three samples were utilized for quantification of Scx+ and S100a4+ cells in uninjured ScxLinDTR hind paws 10 days after the final DT injection. Fluorescent cells were manually counted within an ROI and normalized to either the ROI area or to both the ROI area and total cell number.

Quantitative assessment of gliding function and biomechanical properties

Request a detailed protocol

Gliding function of uninjured and repaired ScxLinDTR tendons was assessed as previously described (Hasslund et al., 2008). Hindlimbs were harvested at the knee-joint and the FDL tendon was disconnected at the myotendinous junction. The FDL tendon was secured between two pieces of tape using superglue and the tendon was loaded incrementally with small weights ranging from 0 to 19 g. Images were captured unloaded and after each load and measurements of the flexion angle of the metatarsophalangeal (MTP) joint were made using Image J. Gliding resistance was derived from the changes in MTP flexion angle over the range of applied loads. An increase in Gliding Resistance and decrease in MTP Flexion Angle is associated with restricted range of motion and increased scar tissue. After conclusion of gliding testing, the FDL tendon was released from the tarsal tunnel while the proximal end of the tendon and the toes of the hind paw were secured into an Instron 8841 uniaxial testing system (Instron Corporation, Norwood, MA). The tendon was loaded until failure at a rate of 30 mm/minute. Seven to 12 samples per genotype per time-point were assessed.

RNA extraction, next-generation sequencing, and data analysis for RNA-Seq

Request a detailed protocol

Tendons (three samples per genotype per time point) were harvested at 14- and 28 days post-repair and flash frozen in liquid nitrogen. Total RNA was isolated using the Bullet Blender (Next Advance) to homogenize the tissue. The RNA was isolated from the resulting extract using Trizol (Life Technologies, Carlsbad, CA) and the RNeasy Plus Micro Kit (Qiagen, Valencia, CA) per manufacturer’s recommendations. The total RNA concentration was determined with the NanoDrop 1000 spectrophotometer (NanoDrop, Wilmington, DE) and RNA quality assessed with the Agilent Bioanalyzer (Agilent, Santa Clara, CA). The RNA integrity number (RIN) for all harvested samples was 8.4 ± 0.85 (mean ± standard deviation). The TruSeq Stranded mRNA Sample Preparation Kit (Illumina, San Diego, CA) was used for next-generation sequencing library construction per manufacturer’s protocols. Briefly, mRNA was purified from 200 ng total RNA with oligo-dT magnetic beads and fragmented. First-strand cDNA synthesis was performed with random hexamer priming followed by second-strand cDNA synthesis using dUTP incorporation for strand marking. End repair and 3` adenylation was then performed on the double-stranded cDNA. Illumina adaptors were ligated to both ends of the cDNA and amplified with PCR primers specific to the adaptor sequences to generate cDNA amplicons of approximately 200–500 bp in size. The amplified libraries were hybridized to the Illumina flow cell and single end reads were generated for each sample using Illumina NovaSeq6000. The generated reads were demultiplexed using bcl2fastq version 2.19.0. Data cleaning and quality control was accomplished using FastP version 0.20.0. Read quantification was accomplished using subread-1.6.4 package (featureCounts). Data normalization and differential expression analysis of ScxLinDTR relative to WT at a given time point was performed using DESeq2-1.22.1 with an adjusted p-value threshold of 0.05 on each set of raw expression measures. The ‘lfcShrink’ method was applied, which moderates log2 fold-changes for lowly expressed genes. DeSeq2 data was uploaded to Qiagen’s ingenuity pathway analysis (IPA, http://www.ingenuity.com) and submitted to core analysis. Canonical pathways, upstream regulators, and downstream disease and functions were deemed biologically and statistically significant when the calculated ABS(z-score)>2 and p-value<0.05. Additionally, imposed upon the upstream regulator data was an expression log ratio cut-off of 0.5, and agreement between predicted activation state and directionality of expression log ratio. The data consists of three biological replicates per genotype per timepoint. The data generated in this study have been uploaded to the Gene Expression Omnibus under accession number GSE156157.

Second harmonic generation two-photon confocal imaging

Request a detailed protocol

Five-micron paraffin sections of WT and ScxLinDTR hind paws were utilized for second harmonic generation (SHG) imaging. Sections were scanned with a Spectra-Physics MaiTai HP DeepSee Ti:Sapphire Laser, tuned to 1000 nm, under 25x magnification, with a 2.5X optical zoom, with a step size of 0.25 μm. 3D projections of image stacks were generated using the 3D-Project macro in ImageJ and analyzed for collagen fibril uniformity using the Directionality macro. The Directionality macro utilizes Fourier transform analysis to derive spatial orientation of image stacks. Three to 10 samples per genotype were used to quantify overall collagen fibril dispersion for pre-pubescent studies, while sections were analyzed from 7 to 8 mice per genotype for the adult homeostasis studies. To quantify spatial dispersion in adult samples, each image stack was divided into equal thirds (top, middle, and bottom), and dispersion was calculated within each region using Directionality macro as above. N = 6 per genotype per age group were used to quantify the spatial collagen fibril dispersion.

Transmission electron microscopy imaging and analysis

Request a detailed protocol

FDL tendons were isolated (N = 4 for WT; N = 3 for ScxLinDTR,10weeks) and fixed in Glutaraldehyde Sodium Cacodylate fixative. One-micron axial sections were cut and stained with Toluidine blue. One-micron sections were then trimmed to 70 nm and stained with uranyl acetate and lead citrate. Sections were placed on grids for imaging on a Hitachi 7650 Analytical TEM. Three non-overlapping images were taken from mid-substance of each tendon at ×40,000 magnification. For measurement of fibril diameter, a region of interest (ROI) was determined within each image so that a minimum of 80 fibrils could be measured. Diameters were measured along the y-axis. The perimeter and the area of the collagen fibrils were quantified. The radii based on the calculated perimeter and area were quantified. The ratio of these two radii represent a measure of fibril roundness (fibril irregularity factor; FIF). An FIF different than one suggest that the fibril is not a perfect circle.

Statistical analysis and animal stratification

Request a detailed protocol

Experimental N determined based on previously published work (Best and Loiselle, 2019; Best et al., 2019). Quantitative data was analyzed via GraphPad Prism and is presented as mean ± standard error of the mean (SEM). Either a student’s t-test or two-way analysis of variance (ANOVA) with Sidak’s multiple comparisons test was used to analyze data when data was normal. A Mann-Whitney test was utilized when data was not distributed normally [Scx+ cells normalized to area and total cell number (Figure 1—figure supplement 1C), D28 F4/80 and S100a4 immunofluorescence (Figure 4)]. GraphPad Prism was used to detect outlier data points (ROUT method, Q-value = 1%) and no outliers were found. Mice were randomly selected for specific experimental outcome metrics prior to surgery and quantitative data (ex. fluorescence quantification, gliding, and biomechanical properties) were analyzed in a blinded manner. For all experiments, an N = 1 represents one mouse. p values ≤ 0.05 were considered significant. * indicates p<0.05, ** indicates p<0.01, *** indicates p<0.001, **** indicates p<0.0001.

Study approval

Request a detailed protocol

This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All animal procedures described were approved by the University Committee on Animal Research (UCAR) at the University of Rochester Medical Center.

Data availability

Sequencing data have been deposited in GEO under accession code GSE156157. All other data generated during this study are included in the manuscript and supporting files.

The following data sets were generated
    1. Best KT
    2. Loiselle AE
    (2021) NCBI Gene Expression Omnibus
    ID GSE156157. RNA-seq analysis of Scx-lineage cell depletion to investigate tendon cell functions during flexor tendon healing.
The following previously published data sets were used
    1. Lincoln J
    2. Barnette DN
    (2014) NCBI Gene Expression Omnibus
    ID GSE57423. Expression data from embryonic day 15.5 atrioventricular canal regions were isolated from Scx-/- and Scx+/+ mice.
    1. Mendias C
    2. Swanson J
    (2019) NCBI Gene Expression Omnibus
    ID GSE138515. Single cell transcriptional atlas of mouse Achilles tendons.

References

    1. Schweitzer R
    2. Chyung JH
    3. Murtaugh LC
    4. Brent AE
    5. Rosen V
    6. Olson EN
    7. Lassar A
    8. Tabin CJ
    (2001)
    Analysis of the tendon cell fate using scleraxis, a specific marker for tendons and ligaments
    Development 128:3855–3866.

Decision letter

  1. Clifford J Rosen
    Senior Editor; Maine Medical Center Research Institute, United States
  2. Mone Zaidi
    Reviewing Editor; Icahn School of Medicine at Mount Sinai, United States
  3. Nathaniel Dyment
    Reviewer
  4. Stavros Thomopoulos
    Reviewer

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Acceptance summary:

The study utilized a mouse model of ScxLin cell depletion to directly assess the function of tendon cells during tendon healing. The finding that ScxLin-depleted tendon cells result in improved biomechanical properties is a noteworthy observation. This is buttressed by RNA sequencing data that provides useful mechanistic information. Issues raised by the reviewers have been adequately addressed.

Decision letter after peer review:

Thank you for submitting your article "Scleraxis-Lineage Cell Depletion Improves Tendon Healing and Disrupts Adult Tendon Homeostasis" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by Mone Zaidi as the Reviewing Editor and Clifford Rosen as the Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Nathaniel Dyment (Reviewer #1); Stavros Thomopoulos (Reviewer #3).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

As the editors have judged that your manuscript is of interest, but as described below that additional experiments are required before it is published, we would like to draw your attention to changes in our revision policy that we have made in response to COVID-19 (https://elifesciences.org/articles/57162). First, because many researchers have temporarily lost access to the labs, we will give authors as much time as they need to submit revised manuscripts. We are also offering, if you choose, to post the manuscript to bioRxiv (if it is not already there) along with this decision letter and a formal designation that the manuscript is "in revision at eLife". Please let us know if you would like to pursue this option. (If your work is more suitable for medRxiv, you will need to post the preprint yourself, as the mechanisms for us to do so are still in development.)

This is an interesting study exploring the role of ScxLin cells on tendon healing, post-natal development, and adult homeostasis. The findings are considered worthy and potentially publishable, particularly due to the unexpected result of an absent phenotype in haploinsufficiency or Scx lineage cells. With that said, several aspects need to be strengthened with new experiments. While point 4 is considered important as it would enhance the conclusions, the Editors are cognizant of the time that may be required. Hence, the latter aspect could be discussed.

Essential revisions

1) In addition to tenocytes, ScxCre will also label epitenon cells and many other cells adjacent to tendons including neighboring sheath cells (see Guak Kim Tan's eLife paper, Jan 2020). Since these cells likely also contribute to the healing response, ablation of these cells may be one cause of improved healing. It is critical that Scxlin cells are assessed and quantified during the healing process. The authors should use the RosaT reporter and/or ScxGFP reporter to perform lineage tracing of ScxCre cells after ablation and/or visualization of ScxGFP cells.

2) Since the authors reported grip-to-grip structural properties and not local strains or material properties, the reported structural properties will be impacted by changes to the adjacent tendon stubs in addition to the healing tissue. Since the MTJ is cut, presumably there will be an unloading-induced remodeling response that may be impacted in the DTA mice because 57% of the cells were ablated. Instead of redoing the biomechanics with local strain measurements and material properties, the authors should analyze the adjacent tendon stubs (i.e., tendon midsubstance adjacent to injury site) using using aSMA staining, SHG and/or TEM (with fibril diameter/distribution quantification) to determine if that tissue was protected from unloading-induced degradation.

3) Since the injury site is a heterogeneous mix of cells, the authors should validate and localize matrix markers identified from RNASeq using immunohistochemistry and/or in situ hybridization. These assays would give insight into the mechanism by which the ablated tendons show improved healing.

4) The reviewers prefer a delayed ablation experiment. They also suggest that the authors inject dead cells prior to injury to see if the effects are not due to Scx cell ablation per se, but instead to induction of an immune response that leads to better healing. While these experiments would exceed the allowable 6 month time for revision, the authors should discuss these points. For instance, Vagnozzi et al., 2019 found that in a cardiac wound injury-repair model, improved healing associated with cell therapies was due to the inflammatory response and recruitment of macrophages (not injection of the cells itself).

https://doi.org/10.7554/eLife.62203.sa1

Author response

Essential revisions

1) In addition to tenocytes, ScxCre will also label epitenon cells and many other cells adjacent to tendons including neighboring sheath cells (see Guak Kim Tan's eLife paper, Jan 2020). Since these cells likely also contribute to the healing response, ablation of these cells may be one cause of improved healing. It is critical that Scxlin cells are assessed and quantified during the healing process. The authors should use the RosaT reporter and/or ScxGFP reporter to perform lineage tracing of ScxCre cells after ablation and/or visualization of ScxGFP cells.

This is a very good point. To address this, we have generated Scx-Cre+; Rosa-Ai9; Rosa-DTRLSL mice to track changes in ScxLinAi9 content before injury and during healing, relative to Scx-Cre+; Rosa-Ai9; DTRWT controls. In new Figure 1 we show that there is a significant decrease in ScxLinAi9 cells in the depleted tendons through 38 days post-depletion. Interestingly, no changes in ScxLinAi9 cells were observed in depleted repairs at D14 post-surgery. However, at D28 a significant decrease in ScxLinAi9 cells were observed, relative to WT repairs (new Figure 4). We have also added a schematic to this figure to help with interpretation of these surprising data and have discussed this at length in the Discussion. Briefly, we think that these data suggest that the ScxLin cells that are present in the adult tendon during homeostasis make their predominant contribution to healing at or around D28. Thus, the effects of depletion are manifested at this time as evidenced by changes in ScxLinAi9 content, function, and transcriptional profile. In contrast, we hypothesize that the lack of differences in ScxLinAi9 content, function and transcriptional profile at D14 are due to a potential increase in cells that express Scx in response to injury. These cells would still be labelled as ScxLinAi9 due to the non-inducible nature of the Scx-Cre driver used, but would not have been targeted for depletion due to their lack of Scx expression prior to injury. The induction of Scx expression after injury in the tendon is well supported by prior studies (Dyment et al., 2014, Dyment et al., PlosOne 2013, Howell et al., 2017).

2) Since the authors reported grip-to-grip structural properties and not local strains or material properties, the reported structural properties will be impacted by changes to the adjacent tendon stubs in addition to the healing tissue. Since the MTJ is cut, presumably there will be an unloading-induced remodeling response that may be impacted in the DTA mice because 57% of the cells were ablated. Instead of redoing the biomechanics with local strain measurements and material properties, the authors should analyze the adjacent tendon stubs (i.e., tendon midsubstance adjacent to injury site) using using aSMA staining, SHG and/or TEM (with fibril diameter/distribution quantification) to determine if that tissue was protected from unloading-induced degradation.

Thank you for this question. The MTJ transection to protect the initial repair from rupturing results in only a transient decrease in tendon loading. By 7-10 days post-surgery, there is substantial re-integration of the MTJ, however, we do acknowledge that this altered loading environment could impact the un-injured tendon adjacent to the repair site. To address this, we examined aSMA staining in the uninjured sections of the tendon proximal/ distal to the repair site (Figure 5—figure supplement 1). A complete lack of aSMA staining was observed in these sections of the tendon, with aSMA staining concentrated at the repair site, suggesting that the transient decrease in tendon loading due to MTJ transection does not result in a widespread tendon response unloading-induced degeneration, remodeling or cellular response. We have also edited the Materials and methods section to clarify the transient nature of the MTJ-transection induced alterations in loading.

3) Since the injury site is a heterogeneous mix of cells, the authors should validate and localize matrix markers identified from RNASeq using immunohistochemistry and/or in situ hybridization. These assays would give insight into the mechanism by which the ablated tendons show improved healing.

To validate and provide spatial information related to the RNAseq study, we performed immunofluorescence for Decorin (Dcn), Thrombospondin 4 (Thbs4) and Microfibril Associated Protein 5 (Mfap5) ECM components that were differentially regulated between WT and ScxLinDTR repairs at D28 post-surgery (Figure 7—figure supplement 1). Consistent with the RNAseq data, we observe substantial increases in the intensity and extent of staining of these ECM components in ScxLinDTR vs. WT repairs, further supporting an enhanced matrix response in ScxLinDTR repairs that differs from that of WT. These data are consistent with an enrichment for “Fibrosis” related genes in ScxLinDTR repairs, though it is clear that these alterations in matrix composition and quantity are actually more reflective of enhanced healing rather than fibrosis.

4) The reviewers prefer a delayed ablation experiment. They also suggest that the authors inject dead cells prior to injury to see if the effects are not due to Scx cell ablation per se, but instead to induction of an immune response that leads to better healing. While these experiments would exceed the allowable 6 month time for revision, the authors should discuss these points. For instance, Vagnozzi et al., 2019 found that in a cardiac wound injury-repair model, improved healing associated with cell therapies was due to the inflammatory response and recruitment of macrophages (not injection of the cells itself).

This is a very good suggestion and is the focus of on-going studies in our laboratory. However, we have confined the current studies to depletion of ScxLin cells prior to injury due in large part to the potentially confounding effects of depleting ScxLin during healing. Since systemic treatment with DT is lethal to Scx-Cre+; DTRLSL animals, likely due to Scx expression in the heart and other organs, we have utilized local injection of DT directly into the hindpaw and tendon. We have conducted pilot experiments and determined that local injections, even of saline, during the healing process results in some alterations in healing. As such, alternative approaches to locally deliver DT during healing are being developed. Therefore, it is not feasible to conduct the proposed experiments at this time. Moreover, as supported by the new data in Figure 4, there are very likely cells that turn-on Scx expression in response to injury, and the overlap between these new ScxLin cells and those from the adult tendon prior to injury is unknown. As such, it is likely that a delayed ablation approach would target different subpopulations of ScxLin cells and our goal with this study was to understand the functional contribution of ScxLin cells that reside in the adult tendon prior to injury. We have added text on this to the Discussion.

Regarding the potential effects of cell death on the immune response, as shown in Figure 1—figure supplement 3, we did not detect any inflammatory infiltration following ScxLin cell depletion. Moreover, as shown in Figure 5, no changes in overall macrophage content were observed between depleted and WT controls at 14 or 28 days post-surgery. Finally, alterations in the immune response was not noted as a primary pathway/ process that was altered in the RNAseq experiment. We have expanded on this in the Discussion.

https://doi.org/10.7554/eLife.62203.sa2

Article and author information

Author details

  1. Katherine T Best

    Center for Musculoskeletal Research, University of Rochester Medical Center, Rochester, United States
    Contribution
    Conceptualization, Formal analysis, Funding acquisition, Methodology, Writing - original draft, Writing - review and editing
    Competing interests
    No competing interests declared
  2. Antonion Korcari

    1. Center for Musculoskeletal Research, University of Rochester Medical Center, Rochester, United States
    2. Department of Biomedical Engineering, University of Rochester, New York, United States
    Contribution
    Conceptualization, Formal analysis, Methodology, Writing - original draft, Writing - review and editing
    Competing interests
    No competing interests declared
  3. Keshia E Mora

    1. Center for Musculoskeletal Research, University of Rochester Medical Center, Rochester, United States
    2. Department of Biomedical Engineering, University of Rochester, New York, United States
    Contribution
    Formal analysis, Methodology, Writing - review and editing
    Competing interests
    No competing interests declared
  4. Anne EC Nichols

    Center for Musculoskeletal Research, University of Rochester Medical Center, Rochester, United States
    Contribution
    Formal analysis, Methodology, Writing - review and editing
    Competing interests
    No competing interests declared
  5. Samantha N Muscat

    Center for Musculoskeletal Research, University of Rochester Medical Center, Rochester, United States
    Contribution
    Formal analysis, Writing - review and editing
    Competing interests
    No competing interests declared
  6. Emma Knapp

    Center for Musculoskeletal Research, University of Rochester Medical Center, Rochester, United States
    Contribution
    Data curation, Writing - review and editing
    Competing interests
    No competing interests declared
  7. Mark R Buckley

    1. Center for Musculoskeletal Research, University of Rochester Medical Center, Rochester, United States
    2. Department of Biomedical Engineering, University of Rochester, New York, United States
    Contribution
    Supervision, Methodology, Writing - review and editing
    Competing interests
    No competing interests declared
  8. Alayna E Loiselle

    1. Center for Musculoskeletal Research, University of Rochester Medical Center, Rochester, United States
    2. Department of Biomedical Engineering, University of Rochester, New York, United States
    Contribution
    Conceptualization, Supervision, Funding acquisition, Writing - original draft, Writing - review and editing
    For correspondence
    alayna_loiselle@urmc.rochester.edu
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-7548-6653

Funding

National Institute of Arthritis and Musculoskeletal and Skin Diseases (F31 AR074815)

  • Katherine T Best

National Institute of Arthritis and Musculoskeletal and Skin Diseases (K01AR068386)

  • Alayna E Loiselle

National Institute of Arthritis and Musculoskeletal and Skin Diseases (R01AR073169)

  • Alayna E Loiselle

National Institute of Arthritis and Musculoskeletal and Skin Diseases (R01AR070765)

  • Mark R Buckley

National Institute of Arthritis and Musculoskeletal and Skin Diseases (T32 AR076950)

  • Anne EC Nichols

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank the Histology, Biochemistry and Molecular Imaging (HBMI) and the Biomechanics, Biomaterials and Multimodal Tissue Imaging (BBMTI) for technical assistance with the histology and biomechanical testing, respectively. We would also like to thank the URMC Multiphoton and Analytical Imaging Center (MAGIC) for assistance with Second Harmonic Generation Imaging, the UR Genomics Research Core for assistance with RNA sequencing experiment and the Electron Microscopy Core for assistance with the transmission electron microscopy data.

Ethics

Animal experimentation: This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All of the animals were handled according to approval by the University Committee on Animal Resources (UCAR) for protocols #2014-004E and 2017-030 at the University of Rochester. All surgery was performed under ketamine anesthesia, and every effort was made to minimize suffering.

Senior Editor

  1. Clifford J Rosen, Maine Medical Center Research Institute, United States

Reviewing Editor

  1. Mone Zaidi, Icahn School of Medicine at Mount Sinai, United States

Reviewers

  1. Nathaniel Dyment
  2. Stavros Thomopoulos

Publication history

  1. Received: August 17, 2020
  2. Accepted: January 21, 2021
  3. Accepted Manuscript published: January 22, 2021 (version 1)
  4. Version of Record published: February 1, 2021 (version 2)

Copyright

© 2021, Best et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 925
    Page views
  • 166
    Downloads
  • 1
    Citations

Article citation count generated by polling the highest count across the following sources: PubMed Central, Crossref, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Cell Biology
    2. Stem Cells and Regenerative Medicine
    Shintaro Ide et al.
    Research Article Updated

    Overwhelming lipid peroxidation induces ferroptotic stress and ferroptosis, a non-apoptotic form of regulated cell death that has been implicated in maladaptive renal repair in mice and humans. Using single-cell transcriptomic and mouse genetic approaches, we show that proximal tubular (PT) cells develop a molecularly distinct, pro-inflammatory state following injury. While these inflammatory PT cells transiently appear after mild injury and return to their original state without inducing fibrosis, after severe injury they accumulate and contribute to persistent inflammation. This transient inflammatory PT state significantly downregulates glutathione metabolism genes, making the cells vulnerable to ferroptotic stress. Genetic induction of high ferroptotic stress in these cells after mild injury leads to the accumulation of the inflammatory PT cells, enhancing inflammation and fibrosis. Our study broadens the roles of ferroptotic stress from being a trigger of regulated cell death to include the promotion and accumulation of proinflammatory cells that underlie maladaptive repair.

    1. Stem Cells and Regenerative Medicine
    Thomas H Ambrosi et al.
    Research Article

    Skeletal stem and progenitor cell populations are crucial for bone physiology. Characterization of these cell types remains restricted to heterogenous bulk populations with limited information on whether they are unique or overlap with previously characterized cell types. Here we show, through comprehensive functional and single-cell transcriptomic analyses, that postnatal long bones of mice contain at least two types of bone progenitors with bona fide skeletal stem cell (SSC) characteristics. An early osteochondral SSC (ocSSC) facilitates long bone growth and repair, while a second type, a perivascular SSC (pvSSC), co-emerges with long bone marrow and contributes to shape the hematopoietic stem cell niche and regenerative demand. We establish that pvSSCs, but not ocSSCs, are the origin of bone marrow adipose tissue. Lastly, we also provide insight into residual SSC heterogeneity as well as potential crosstalk between the two spatially distinct cell populations. These findings comprehensively address previously unappreciated shortcomings of SSC research.