Altered TGFB1 regulated pathways promote accelerated tendon healing in the superhealer MRL/MpJ mouse

To better understand the molecular mechanisms of tendon healing, we investigated the Murphy Roth’s Large (MRL) mouse, which is considered a model of mammalian tissue regeneration. We show that compared to C57Bl/6J (C57) mice, injured MRL tendons have reduced fibrotic adhesions and cellular proliferation, with accelerated improvements in biomechanical properties. RNA-seq analysis revealed that differentially expressed genes in the C57 healing tendon at 7 days post injury were functionally linked to fibrosis, immune system signaling and extracellular matrix (ECM) organization, while the differentially expressed genes in the MRL injured tendon were dominated by cell cycle pathways. These gene expression changes were associated with increased α-SMA+ myofibroblast and F4/80+ macrophage activation and abundant BCL-2 expression in the C57 injured tendons. Transcriptional analysis of upstream regulators using Ingenuity Pathway Analysis showed positive enrichment of TGFB1 in both C57 and MRL healing tendons, but with different downstream transcriptional effects. MRL tendons exhibited of cell cycle regulatory genes, with negative enrichment of the cell senescence-related regulators, compared to the positively-enriched inflammatory and fibrotic (ECM organization) pathways in the C57 tendons. Serum cytokine analysis revealed decreased levels of circulating senescence-associated circulatory proteins in response to injury in the MRL mice compared to the C57 mice. These data collectively demonstrate altered TGFB1 regulated inflammatory, fibrosis, and cell cycle pathways in flexor tendon repair in MRL mice, and could give cues to improved tendon healing.


MRL tendons heal with reduced adhesions and improved biomechanical properties.
We first sought to compare the extent of adhesion formation and repair strength of healing flexor tendons in the C57 and MRL mice. The healing of partially lacerated deep digital flexor tendons (DDFT) was qualitatively assessed using hematoxylin-stained axial tendon sections. In general, the MRL tendons appeared to respond to the injury with a reduced cellular response compared to the C57 tendons. Specifically, there were no gross morphological differences in the uninjured (Fig. 1A,E), where the tendon bundles were sparsely populated by cells, and the tendons were encased with a very thin synovial sheath (black arrows in Fig. 1a,e). Fourteen days post injury (dpi), injured tendons in both mouse strains experienced substantial hyperplasia in the surrounding synovial sheath (black arrows, Fig. 1b,f), while no significant changes in cell number in the tendon proper were generally observed (yellow arrows Fig. 1B,F). At 28 dpi, substantial increases in cell density in both the tendon bundle and the peritendinous adhesions were observed in the C57 mice but not the MRL mice (yellow arrows, Fig. 1C,G). At this stage, the synovial hyperplasia began to resolve in the MRL mice but persisted in the C57 mice (black arrows, Fig. 1c,g). By 56 dpi, the injured MRL tendons appeared to have restored uninjured morphology where the tendon bundle was sparsely cellular and encased by a thin synovial membrane (Fig. 1H,h). In contrast, while the cellularity of the C57 tendon bundle decreased, the surrounding synovial sheath still demonstrated a hyperplasia morphology at 56 dpi (Fig. 1D,d). Quantification of the tendon perimeter adhered to the surrounding sheath and tissues (Fig. 1I), cellular density (Fig. 1J) in the tendon bundle demonstrated significant increases in adhesions and cellularity in the C57 tendons at 14 dpi and 28 dpi, respectively, compared to the MRL tendons. In addition, the synovial space was significantly increased in the MRL tendon at 14 and 28 dpi compared to the C57 tendons (Fig. 1K), while the cross sectional area of the MRL tendons was significantly larger than the C57 tendon at all time points (Fig. 1L).
To test whether the changes in cell density over time and between strains are attributable to cell proliferation, immunohistochemical staining for Proliferating Cell Nuclear Antigen (PCNA) was performed ( Fig. 2A) and demonstrated activation of cellular proliferation in the tendon bundle and the peritendinous tissues following injury, which peaked at 14 dpi. Furthermore, quantifying the percentage of PCNA + cells at all timepoints after injury revealed increased cellular proliferation in the healing C57 tendons relative to the MRL tendons (Fig. 2B).
The functional properties of injured tendons were evaluated using biomechanical testing, which involved displacement-controlled tensile stretching of the tendon to failure at 0.05 mm/s. In terms of the structural properties, the injury resulted in significant decreases in peak tensile force and stiffness at 14 dpi in both the C57 and the MRL tendons. In contrast to the C57 tendons, which exhibited sustained reductions in peak tensile force and stiffness up to 56 dpi, the MRL tendons recovered uninjured levels of peak tensile force by 28 dpi (Fig. 3A) and were significantly stiffer (Fig. 3B) than the C57 tendons at all time points up to 56 dpi. Since the crosssection area of the uninjured MRL tendons are approximately 20% larger than the C57 tendons, we computed the corresponding stress-strain tensile curves and determined the effects of injury on the size-independent material properties. At 14 dpi, the tensile strength (peak stress) of the MRL was significantly reduced compared www.nature.com/scientificreports/ to uninjured tendon and the C57 injured tendon. The tensile strength of the C57 continued to decline at 28 dpi but recovered at 56 dpi. In contrast, the MRL tendon recovered the tensile strength to levels comparable to uninjured tendon as early as 28 dpi (Fig. 3C). Injury resulted in significant reductions in the elastic modulus of the C57 tendon, which did not improve over 56 dpi. While the elastic modulus of the uninjured MRL tendon was significantly reduced compared to the uninjured C57, injury resulted in reductions in the elastic modulus on 14, 28, and 56 dpi, which were not significantly different than uninjured tendon (Fig. 3D). Collectively, these . High magnification (×20) regions of interest (a-h) are representative of five levels analyzed per tendon. Black arrows identify the thin synovial sheath in uninjured tendons, and the hyperplasia that ensues immediately after injury in both strains. The synovial sheath hyperplasia appears to persist in the C57 tendons up to 56 dpi, but resolves in MRL tendons by 28 dpi. Yellow arrows indicate regions of increased cellularity in the injured tendon and peritendinous tissues evident by the punctate, intense hematoxylin staining. Histomorphometric quantification of the healing tendon shows reduced adhesions (I) and tendon hyperplasia (J), as well as increased synovial space (K) and tendon area (L) in the MRL mouse. The length of the tendon perimeter adhered to the subcutaneous tissue and cell density are significantly reduced at 14-and 28-dpi in the MRL, respectively (N = 4 animals with n = 3-5 images (technical replicates) per animal).  www.nature.com/scientificreports/ data support the conclusion that MRL flexor tendons heal with reduced peritendinous adhesions and cellular proliferation and accelerated recovery of strength and stiffness.
Altered inflammatory, fibrotic, and cell cycle regulation in C57 and MRL tendon healing. To gain insights into the differences in the biological drivers of the tendon injury response between the C57 and MRL mice, we performed next generation sequencing of bulk tissue RNA (RNA-seq) at 7 days post-injury, which we reasoned to be a time point that defines the transition from the acute inflammatory phase to the reparative phases. Based on a twofold change (± 2FC) and adjusted p value < 0.05 differential expression criteria relative to healthy uninjured tendon, 891 genes were differentially expressed (594 upregulated and 297 downregulated) in the C57 tendons, and 1358 genes were differentially expressed (747 upregulated and 611 downregulated) in the MRL tendons. Of those, 406 genes (DEG) were differentially expressed, relative to respective uninjured controls, in both mice. We identified 22 significantly enriched Reactome pathways in the injured C57 tendon and 51 significantly enriched pathways in the injured MRL tendon based on False Discovery Rate (FDR) < 0.05. Functional annotation of these enriched gene sets identified remarkable differences in the transcriptional activity, wherein the gene sets enriched in the C57 injured tendon were functionally linked primarily to fibrosis, immune system signaling and extracellular matrix (ECM) organization (Fig. 4A), while the gene sets enriched in the MRL injured tendon were dominated by cell cycle pathways (Fig. 4B). Closer inspection of the top 15 Reactome pathways enriched in the C57 injured tendons (Fig. 5A) identified cytokine signaling, including interleukins 4, 10, and 13, and neutrophil degranulation as significantly enriched immune system pathways, and collagen formation (biosynthesis and assembly), ECM proteoglycans, ECM degradation, and integrin and non-integrin cell-ECM interactions as significantly enriched ECM organization pathways. In contrast, the top 15 Reactome pathways enriched in the MRL injured tendons (Fig. 5B) identified numerous cell cycle pathways related to mitotic phase, cell cycle checkpoints, and G1-S phase transition, as well as cell signaling by Rho GTPases and immune system signaling by interleukins and neutrophil degranulation. . Asterisks indicate significant mouse strain differences (p < 0.05). a and b indicate significant differences (p < 0.05) from uninjured within C57 and MRL mouse strains respectively. www.nature.com/scientificreports/ To investigate differences in activation of myofibroblasts and macrophages as drivers of fibrotic scarring that could explain these transcriptional differences, histology and immunofluorescent staining against α-SMA+ myofibroblasts, F4/80+ macrophages, and BCL-2 expressing cells were performed on C57 and MRL tendons at 7 dpi. Visual examination of Masson's trichrome stained injured tendon sections at this time point did not reveal any significant morphological differences in the injured tendons (Fig. 6A,B), except for increases in the vascularization of the scar tissue immediately surrounding the injured C57 tendon, which was notable in α-SMA+ smooth muscle cells surrounding the blood vessels (Fig. 6C). In addition to the increased vascularization in the C57 scars, activated (α-SMA+) myofibroblasts and F4/80+ macrophages could be observed in both the injured C57 tendon and the surrounding scar tissue (Fig. 6C and insets). Of note was the observation of a subset of cells coexpressing both SMA and F4/80 in the injured C57 tendon (yellow staining in Fig. 6c′), which could implicate the activation of macrophage-myofibroblast transition (MMT) as a feature of fibrotic tendon healing. In contrast, www.nature.com/scientificreports/ fewer α-SMA+ myofibroblasts and F4/80+ macrophages could be observed in the injured MRL tendon, but were abundant in the surrounding scar tissue ( Fig. 6D and insets). No localized expression of α-SMA and F4/80 could be observed in the injured MRL tendon. BCL-2 expression was notably abundant in the injured C57 tendon, but appeared to be reduced in the surrounding scar tissue ( Fig. 6E and insets) as well as in the MRL injured tendon and surrounding scar tissues ( Fig. 6F and insets).
Dichotomous roles for TGB1 in the C57 and MRL tendons. Ingenuity pathways analysis (IPA) of active or inhibited upstream transcriptional regulators were also examined to provide additional insights into the observed gene expression changes driving the altered mechanobiology of tendon healing. In the C57 injured tendons, the top 20 activated and inhibited upstream regulators (Fig. 7A) were associated with 184 and 62 unique genes, respectively, that enriched Gene Ontology (GO) terms related to chemotaxis and cell motility and migration, ECM organization, sprouting angiogenesis and blood vessel morphogenesis, positive regulation of cell growth and proliferation, inflammatory responses, collagen catabolism and regulation of phosphatidylinositol-and ERK1/2-signaling (Table S1). In the MRL injured tendon, the top 20 activated and inhibited upstream regulators ( Fig. 7B) were associated with 325 and 304 unique genes, respectively, that enriched GO terms related to regulation of cell cycle, including mitotic cell cycle, sister chromatid cohesion, G1/S transition of mitotic cell cycle, mitotic cytokinesis, DNA replication, chromosome segregation, regulation of cell cycle, microtubulebased movement, metaphase plate congression, mitotic spindle organization, G2/M transition of mitotic cell   www.nature.com/scientificreports/ cycle, negative regulation of cell cycle, and extracellular matrix organization (Table S2). Of particular note, was the inhibition of upstream regulators TP53, CDKN1A (p21), CDKN2A (p16), which are associated with DNA damage response and cell cycle arrest. TGFB1 was a significantly activated Upstream Regulator in both the C57 (Activation Z-score = 3.11 and p value = 6.4 × 10 -13 ) and MRL (Activation Z-score = 2.97 and p value = 1.1 × 10 -27 ) injured tendons. GO functional annotation revealed that activation of TGF-β1 drives different biological pathways in the C57 and MRL injured tendons. In the C57 injured tendons, the 76 target genes of TGFB1 (Table S3) enriched GO pathways related to cytokine and chemokine signaling, ECM organization, and cell motility and migration (Fig. 8A). In contrast, 160 target genes of TGFB1 in the MRL (Table S4) enriched GO pathways associated with not only ECM organization and cytokine signaling but mostly with regulation of cell cycle, mitosis, and DNA binding and transcription (Fig. 8B).
Reduced circulating cytokines and chemokines in the MRL injured mice. We have previously shown that the putative senescence marker plasminogen activator inhibitor 1 (PAI-1), downstream of TGF-β, is elevated in fibrotic tendon tissue in humans and mice 30 . In subsequent studies, we observed increases in the prosurvival signal BCL-2 in injured tendon, which appeared to be linked to inflammatory nuclear factor κB (NF-κB) signaling 31 . Therefore, in order to evaluate the roles of the immune response and systemic inflammatory signaling on the tendon healing, sera were extracted from whole blood draws from mice at different time points up to 14 dpi and analyzed for senescence associated secretory phenotype (SASP) proteins including inflammatory cytokines/chemokines, TGF-βs, and MMPs. In general, C57 and MRL mice exhibit similar circulatory concentrations of serum proteins in their uninjured state (Fig. 9A). Upon injury, the temporal changes in the circulating proteins in the C57 mice were in general more pronounced than in the MRL mice. Analyzed proteins were clustered in 4 major classes based on Pearson's correlations (Fig. 9B). In the first cluster, the chemokine CXCL9 (MIG) showed no significant difference over time in the C57 but increased significantly over time in the MRL injured tendons. Cluster 2 proteins showed no significant differences over time or between C57 and MRL mice, which included TGF-βs (1,2,3), MMP-3 and CCL 4 (MIP-1β). Cluster 3 proteins, which included IL-1α, -12 (P40), -13, MMP-2 and -8, CSF1 (M-CSF) and 3, CCL5 (RANTES) and 11 (Eotaxin), CXCL2 (MIP-2) and 5 (LIX), were significantly higher in the C57 mice compared to the MRL mice, but did not vary significantly over time. Cluster 4 proteins showed significant changes over time in the C57 mice only, peaking at either 1 dpi (IL-2 and -9, MMP-12, and CCL3 (MIP-1α)) or 5 dpi (IL-1β, -2, -4, -5, -6, -7, 10, -12 (P70), -17, TNF-α, LIF, MMP-9, CCL2 (MCP-1) and CXCL1(KC) and 10 (IP-10)). In contrast circulation levels of these proteins were significantly lower and remained mostly unchanged over time in the MRL mice. Protein association networks of cluster 3 (Fig. 9C) and cluster 4 (Fig. 9D) circulating proteins visualized in String DB identified strong functional associations, which involved cytokine signaling in immune system and more specifically signaling by interleukins, chemokine signaling and activation of matrix metalloproteinases.

Discussion
Development of fibrotic flexor tendon adhesions, particularly in zone II of the hand 32 , limits their natural gliding function, strength, and intricate joint flexion 33,34 . Recent biological advances in scarless repair of tendon injuries 24,26,35 have begun to elucidate the complex cellular and transcriptional interplay in restoring native tissue properties to injured tendons. For example, previous studies demonstrate that MRL tendon injuries heal to near normal biomechanical properties 25 , have reduced infiltrating immune cells 25 , attenuated inflammatory expression 24,36 , lower catabolic enzyme activity 28 , heightened MMP activity 24 , and improved ECM alignment 26 . These reports are consistent with our current findings in flexor tendons, but the biological drivers of this regenerative tendon healing phenotype have hitherto not been fully understood.
The objective of our study was to uncover the mechanisms of improved healing of MRL flexor tendon injuries to understand better the limitations of natural healing in common C57 mouse tendons. We first investigated the longitudinal histomorphometry and biomechanical properties of MRL and C57 mice flexor tendons after a zone II partial laceration injury. We demonstrated that the MRL flexor tendons heal with reduced tenocyte cellularity, reduced peritendinous adhesions, and accelerated recovery of their biomechanical properties of tensile strength, stiffness, and elastic moduli. We then performed bulk tissue transcriptomic analysis using next-generation RNA sequencing (RNA-Seq) to characterize enriched transcriptional pathways in the early reparative phase. These experiments uncovered dichotomous roles for TGFB1 in the healing of C57 and MRL tendons. Specifically, we identified TGFB1 as a major driver of positive enrichment of cytokine signaling, neutrophil degranulation, and ECM degradation pathways in injured C57 flexor tendons, and, by contrast, negative enrichment of cell cycle regulation and senescence pathways in the injured MRL flexor tendons. To validate these transcriptional observations, we investigated the systemic inflammatory response by analyzing the cytokines, chemokines, and growth factors in sera extracted from peripheral blood at different phases of healing. These assays revealed significantly lower circulatory concentrations of SASP proteins in the MRL relative to the C57 despite comparable baseline (day 0) serum concentrations. These findings provide insights into novel molecular pathways in tendon healing that could be therapeutic targets for regenerative tendon repair.
TGF-β is critical to the modulation of tendon healing and the effects of its isoforms are quite distinct. In flexor tendons, TGF-β1 expression is auto-inductive and is initially produced by inflammatory cells 37 . TGF-β1 stimulates fibroblast proliferation and migration, synthesis of collagen and fibronectin, inactivation of MMPs 38 and myofibroblast differentiation 12 . Chang et al. reported increased expression of TGF-β1 in both tenocytes and infiltrating fibroblasts and inflammatory cells from the tendon sheath in a rabbit zone II flexor tendon healing model 39 . The inflammatory cells stimulate synovial and epitenon fibroblasts, partly through TGF-β1, to produce fibronectin, a marker of scarring 37 . The role of TGF-β1 in scar formation is accentuated by the observation that www.nature.com/scientificreports/ while fetal cutaneous wound healing and tendon repair are scar-free 40,41 , the in utero administration of TGF-β1 to fetal rabbit wounds results in fibrotic scarring similar to that observed in adult rabbits 41 . In our study, we did not observe changes in circulating TGF-β1 levels over time or between the C57 and MRL mice. However, we  www.nature.com/scientificreports/ did observe dichotomous downstream effects of TGFB1 gene regulatory networks between the C57 and MRL healing tendons locally. In particular, genes regulated by TGFB1 in the C57 tendons positively enriched pathways relevant to cytokine signaling in chemotaxis and inflammation, ECM organization, angiogenesis, and epithelial to mesenchymal transition, whereas the TGFB1 regulated genes in the MRL tendons negatively enriched pathways predominately related to cell cycle regulation and senescence. These results corroborate reports of dual roles for TGF-β in different contexts, such as being protective and cytostatic against tumorigenesis but also a promoter of tumor progression, invasion, and metastasis in established tumors 42 . Dual roles of TGF-β have also been reported in autoimmune-induced tissue damage, wherein reduced TGF-β signaling in immune cells can lead to inflammatory dysregulation triggering the production of TGF-β in target organs, which subsequently leads to fibrogenesis and organ failure 43 . It is possible that the effects of TGF-β in the injured tendon are intertwined with the inflammatory cytokines and chemokines in the site of injury and systemically. Particularly intriguing was the detection of cells positive for both F4/80 and α-SMA in the C57 injured tendons but not the MRL tendons. This observation could be indicative of macrophage-myofibroblast transition (MMT), which is implicated as a mechanism of chronic inflammation and progressive fibrosis in kidney 44 but never reported in tendon repair. Previous studies have demonstrated the role of TGF-β/Smad3 signaling in regulating the transition of bone marrow-derived macrophages into myofibroblasts during tissue fibrosis 45 . It is also possible that intrinsic genetic differences in the tendon fibroblasts and infiltrating inflammatory cells could explain the dichotomous responses downstream of TGFB1. A third possibility is that the innate tissue properties in the MRL mouse tendon 26 can modulate the effects of TGF-β1 resulting in accelerated repair. These possibilities warrant further investigation in future studies. RNA-Seq analyses revealed negative enrichment of cell cycle regulation in the MRL tendons compared to the C57 healing flexor tendons who exhibited positive enrichment of inflammatory and ECM organization pathways, driven differentially by TGFB1 regulated gene sets. The increased responsiveness of injured C57 tendons to circulating inflammatory cytokines, chemokines, and growth factors may be associated with its positive enrichment of cell cycle evidenced by cyclin D1 (ccnd1) pathways. In contrast, injured MRL tendons negatively enriched pathways associated with cell cycle regulators TP53, CDKN1A (p21), CDKN2A (p16), RB1, FOXO3, and PTEN. The enrichment of these cell cycle regulators is associated with cellular senescence 46 . Cell cycle regulation and cellular senescence have recently become therapeutic targets in a number of conditions including fibrosis. Specifically, the cell cycle regulator p53 is induced in multiple fibrotic conditions such as idiopathic pulmonary fibrosis (IPF) and kidney fibrosis. In the absence of p53, bleomycin induced pulmonary fibrosis is prevented 47 and its absence reduces chronic renal injury 48 . Furthermore, the antifibrotic agent used to treat IPF, Pirfenidone, has been shown to induce cell cycle arrest 49,50 . Reduced cell apoptosis, cell cycle regulation through p21 (cdkn1a) 51 , p53 (trp53), and p16 (cdkn2a), increased cell proliferation 52 , and enhanced stem cell function 53 may command augmented regenerative capacity.
While the prolonged systemic response in the C57 is surprising given the focal nature of the injury, the diminished response in the MRL is likely associated with decreased recruitment of inflammatory cells to the injury site similar to the scar-free healing of fetal wounds 54 . Interestingly, chemokine (C-X-C motif) ligand 9, CXCL9 is the single cytokine in cluster 1 and is transiently upregulated in MRL healing. In other fibrotic conditions increased CXCL9 has been beneficial to patient function and reduced fibrosis 55,56 . A study of serum cytokine www.nature.com/scientificreports/ levels of patients with hypersensitivity pneumonitis pulmonary fibrosis demonstrated a worsening lung function with reduced levels of CXCL9 55 . Additionally, systemic administration of CXCL9 prevented the development of CCL4 induced liver fibrosis in a murine model 56 . Altogether, the observed elevation of CXCL9 in the MRL sera after injury points to a potential therapeutic benefit of CXCL9 administration, which merits testing in tendon injury. In addition, proteins in clusters 3 and 4 can potentially be viewed as serum biomarkers or potentially candidates for therapeutics inhibition to mitigate fibrovascular scar healing in tendon. This study has yielded novel insights into the transcriptional upstream regulators that drive enhanced tendon healing and fibrosis. However, there are limitations that qualify the conclusions of our study. First, our study was limited in assessing transcriptional regulation in bulk tendon tissue in native tendon and 7 days after injury. Whole-genome transcriptional profiling from the injured tendon tissue site represents a heterogeneous cell population of tenocytes, fibroblasts, myofibroblasts and immune cells, which could be resolved by single-cell RNA-sequencing in the tendon healing as recently described 57 . Future analysis will benefit from the utilization of single cell RNA-sequencing or spatial transcriptomics to elucidate signaling mechanisms at a cellular level given the involvement of the heterogeneous cellular populations at the tendon injury site and at later time-points that would represent the fibro-proliferative and remodeling phases of wound healing. Another limitation is that only male mice were included in this study. There are mixed reports on sex differences in the mechanical properties and biochemical composition of male and female tendons 58 and in their response to injury 59 . With respect to incidence of flexor tendon lacerations in humans, males are 5:1 more likely to experience flexor tendon lacerations compared to females 60 , which justifies the focus on male animals in this study. However, future studies should evaluate potential sex differences flexor tendon healing in the C57 and MRL mice.
In summary, this study identified critical cell cycle regulation and differentiation roles for TGFB1 signaling that contribute to improved tendon repair following injury in MRL flexor tendons but not in the C57 tendons. Uncovering the unique cellular and molecular pathways in which TGFB1 induces regenerative tendon healing in the MRL model could help identify therapeutic targets for regenerative tendon healing.

Materials and methods
Animal care and flexor tendon surgery. The study is in accordance with relevant guidelines and regulations. All methods are reported in accordance with ARRIVE guidelines (https:// arriv eguid elines. org) 61 . All animal experimental protocols were approved by University of Rochester Committee on Animal Research (UCAR). Breeding pairs for the C57BL/6J (C57; stock number: 000664) and MRL/MpJ (MRL; stock number: 000486) were obtained from the Jackson Laboratory (Bar Harbor, ME) and bred in-house. Nine to eleven-week-old male mice from each strain were randomized into experimental groups. The mouse surgery protocol involves a partial laceration of the deep digital flexor (DDF) tendon of the 3rd digit in the hind paw, as previously described 30 . Surgeries were performed under a stereomicroscope on the hind paws using sterile aseptic techniques. Mice were anesthetized with 60 mg/kg ketamine and 4 mg/kg xylazine. A transverse incision was made on the 3rd digit between the metatarsophalangeal (MTP) and the proximal interphalangeal (PIP) joints. A transverse, mediolateral cut across roughly 50% of the DDF tendon width was completed. The skin incision was closed using interrupted sutures (Ethicon Suture, V950G, 9-0). To minimize pain, all mice were subcutaneously, Buprenorphine was administered immediately after closure of the skin incision (0.05 mg/kg) and postoperatively every 24 h up to 72 h. Blood was collected from mice whose hind paws were used for gene expression analysis.
Biomechanical testing. Tensile elastic biomechanical testing was completed as previously described 30 .
The DDFT was released at the myotendinous junction, precisely dissected along the tibia and tarsal bones, and released from the tarsal tunnel pulleys. The middle (3rd) tendon was isolated, with the tendon-to-bone attachment (enthesis) left intact for mounting and gripping. Tendons were loaded in a custom uniaxial microtester (eXpert 4000 MicroTester, ADMET, Inc., Norwood, MA) with the 3rd digit tip gripped between the two stainless steel plates. After preloading to 0.05 N, tendon width, thickness, and gauge length were measured while the tissue was immersed in a 1× PBS bath and imaged on the microscope. Tendon gauge length was measured as the length of the tendon between the Microtester system clamps. Width measurements were taken first, then the tendon was rotated 90 degrees in the clamps to measure thickness. Width and thickness measurements were taken at 3-5 locations along the length of the tendon and averaged. Cross sectional area was assumed to be an ellipse. Based on each sample's total gauge length, each sample underwent uniaxial displacement-controlled biomechanical testing as follows; 1% strain preconditioning for 10-cycles, returned to initial gauge length to rest for 5 min, stretched to 5% strain (0.05 strain/8 s) and held for 10 min to allow for stress-relaxation, returned to initial gauge length to rest for 5 min, and concluded with a displacement-controlled ramp to failure at 0.05 mm/ sec. Structural and material properties were determined from the ramp to failure force-displacement and stressstrain curves, respectively. All mechanical testing analysis was completed with a custom Matlab (MathWorks, Natick, MA) code. Sample size of N = 4-6 for each mouse strain and each experimental timepoint.
Histology and immunohistochemistry. The hind paw was transected near the tarsal bones and prepared for histology by scoring the skin superficial to the footpad to enhance 10% Neutral Buffered Formalin (NBF) tissue infiltration. Tissues were fixed in 10% NBF for 48 h at room temperature (RT), followed by decalcification in calis-EDTA (pH 7.2-7.4) for 7 days at RT. The tissues were then dehydrated in an ethanol gradient and embedded in paraffin. Paraffin-embedded tissues were cut axially into 5 μm thick sections between the metatarsophalangeal (MTP) joint and the A3 pulley of each sample, then stained for 15 s with Mayer's Hematoxylin. Sections for histomorphometric analysis were immediately cover slipped with a water-based mounting medium (Faramount Aqueous Mounting Medium, Dako, #S3025), whereas sections for immunohistochemistry (IHC) were cover slipped after IHC. www.nature.com/scientificreports/ For immunohistochemistry, paraffin-embedded and sectioned tissues were collected at 0, 14, 28, and 56 days post injury (dpi). Slides were incubated in citrate buffer (10 mM Na citrate, pH 7.2) for 3 h at 70 °C for antigen retrieval, circled with a hydrophobic barrier pen (Vector Laboratories, Burlingame, CA), and blocked with an endogenous peroxidase/ alkaline phosphatase inhibitor (Bloxall, Vector Laboratories). After washing with PBST (PBS+ 0.1% Tween-20), they were incubated for 1 h at room temperature in blocking solution (5% normal horse serum+ PBST). Sections were incubated for 1 h at RT with the primary antibody mouse monoclonal PCNA (ab29, Abcam) at 1:1000 (1 μg/mL) dilution. Followed by Vectastain Elite Kit with DAB chromogen (Vector Laboratories). Immunohistochemical sections were rinsed with deionized water, counterstained with Mayer's Hematoxylin, cover slipped with a water-based mounting medium (Faramount Aqueous Mounting Medium, Dako, #S3025), and imaged with a VS120 Virtual Slide Microscope (Olympus, Waltham, MA).

RNA extraction.
Tendon tissue was collected from the injury site for RNA extraction and subsequent gene expression analysis by the URMC Genomics Core. Briefly, partially lacerated, or uninjured tendon tissue was flash frozen in liquid nitrogen and stored at − 80 °C until extraction. Peritendinous tissue and surrounding sheath tissue were not removed due to size limitations. Total RNA was isolated from single tendons utilizing Trizol (ThermoFisher, #15596026) extraction methods and a bullet blender for tissue homogenization. RNA concentration was determined with the NanoDrop 1000 spectrophotometer (NanoDrop, Wilmington, DE) and RNA quality assessed with the Agilent Bioanalyzer 2100 (Agilent, Santa Clara, CA).
Next-generation sequencing (NGS) data processing and alignment. Samples of 1 ng of total RNA were pre-amplified with the SMARTer Ultra Low Input kit v4 (Clontech, Mountain View, CA) per manufacturer's recommendations. The quantity and quality of the subsequent cDNA was determined using the Qubit Flourometer (Life Technnologies, Carlsbad, CA) and the Agilent Bioanalyzer 2100 (Agilent, Santa Clara, CA). 150 pg of cDNA was used to generate Illumina compatible sequencing libraries with the NexteraXT library preparation kit (Illumina, San Diego, CA) per manufacturer's protocols. The amplified libraries were hybridized to the Illumina single end flow cell and amplified using the cBot (Illumina, San Diego, CA). Single end reads of 100nt were generated for each sample. Raw reads generated from the Illumina HiSeq. 2500 sequencer were demultiplexed using bcl2fastq version 2.19.0. Quality filtering and adapter removal are performed using Trimmomatic version 0.36 with the following parameters: "TRAILING:13 LEADING:13 ILLUMINACLIP:adapters. fasta:2:30:10 SLIDINGWINDOW:4:20 MINLEN:15". Quality reads were then mapped to the Mus musculus reference sequence (GRCm38.p5) with STAR_2.5.2b with the following parameters: "--twopassMode Basic --run-Mode alignReads --genomeDir ${GENOME} --readFilesIn ${SAMPLE} --outSAMtype BAM SortedByCoordinate --outSAMstrandField intronMotif --outFilterIntronMotifs RemoveNoncanonical". Differential expression analysis and data normalization were performed using default methods in DESeq. 2-1.14.1 with an adjusted p value threshold of 0.05.
Bioinformatics and statistical analyses of DEG. Differential expression analysis and data normalization were performed using DESeq. 2-1.14.1 on RNA extracted from injured tendons at 7 days post injury (dpi) relative to uninjured expression levels within each strain independently (N = 4 animals per experimental group). DEG was identified by filtering DeSeq. 2 pairwise comparisons for biological (ABS(Log2FC) > 1) and statistical significance (adjusted p < 0.05). Functional enrichment of upstream regulators was completed in pathway enrichment analysis using Ingenuity Pathway Analysis (IPA; http:// www. ingen uity. com). Default settings were used to determine upstream regulators using the global molecular network contained in the IPA knowledge base. Downstream targets of the upstream regulator TGF-β1 identified in IPA were exported for further functional analysis using Enrichr (https:// maaya nlab. cloud/ Enric hr/) and Reactome (https:// react ome. org).