Embryonic and Postnatal Tendon Cells Respond Differently to IL-1beta

Abstract

Adult tendons heal as scar tissue, whereas embryonic tendons heal scarlessly via unknown mechanisms. Scarred tendon healing results from inflammation-driven imbalances in anabolic and catabolic functions. To test scarless vs. scarring age tendon cell responses to inflammatory conditions, we treated embryonic and postnatal tendon cells with interleukin (IL)-1β and characterized expression of collagens, matrix metalloproteinases (MMPs), inflammatory mediators, as well as phosphorylation of signaling molecules. At baseline, postnatal cells expressed significantly higher levels of inflammatory mediators. When treated with IL-1β, both postnatal and embryonic cells upregulated inflammatory mediators and MMPs. Notably, postnatal cells secreted inflammatory factors up to 12.5 times the concentration in embryonic cultures. IL-1β activated NF-kB p65 and p38 MAPK pathways in both cell types, but phosphorylated p38 MAPK levels were two-times higher in postnatal than embryonic cells. Our results suggest scarred healing tendon cells respond to pro-inflammatory cytokines by promoting an imbalance in anabolic and catabolic functions, and that the heightened response involves p38 MAPK signaling activity. In contrast, embryonic cell responses are smaller in magnitude. These intriguing findings support a potential role for tendon cells in determining scarless vs. scarred healing outcomes by regulating the balance between anabolic and catabolic functions during tendon healing.

INTRODUCTION

Tendon is a musculoskeletal tissue that transmits muscle-generated forces to bone, thereby enabling skeletal locomotion. It has been reported that musculoskeletal ailments and injuries lead to 130 million physician visits each year in the U.S., with half of these visits involving tendons and ligaments¹. Significant tendon ruptures require surgical repair, while other injuries are treated non-surgically. Unfortunately, in all cases, tendon heals as scar tissue, which is tissue with abnormal extracellular matrix (ECM) composition (e.g., altered collagen ratios), ECM organization, and mechanical properties. Consequently, tendon function is permanently compromised, leading to greater risk of re-injury and frequent association with pain. Thus, there is a critical need for more effective strategies to treat tendon injuries.

Postnatally, an acute inflammatory response occurs when tendon is injured, with pro-inflammatory cytokines appearing at high levels within minutes^2–4. Interleukin (IL)-1β is one of the earliest and most highly expressed pro-inflammatory cytokines at the injury site⁴. In response to pro-inflammatory cytokines, such as IL-1β, adult tendon cells upregulate expression of inflammatory mediators, including IL-1 receptor antagonist (IL-1RA), IL-1 receptors (IL-1R), tumor necrosis factor-α (TNF-α), IL-6, cyclooxygenase (COX)-2, and matrix metalloproteinases (MMP)-3 and −13^5–8. These molecules regulate the balance of anabolic functions (ECM synthesis) and catabolic functions (ECM degradation) during healing. Manipulating the levels of these anabolic and catabolic molecules during healing alters the quality of tendon formed^9–11, although scarless tendon healing has yet to be achieved.

Studies in sheep show that fetal tendons of 80–86 days of gestation (full term by 145 days) that are acutely injured in utero will heal regeneratively with restoration of native tissue properties (scarlessly), whereas adult tendons heal abnormally^{12, 13}. Furthermore, fetal tendons possesses fewer inflammatory cells and lower levels of inflammatory mediators during healing than adult tendons¹². When fetal and adult sheep tendon tissue were subcutaneously transplanted into severe combined immunodeficiency (SCID) adult mice (to avoid immune rejection of engrafted tendons) and then wounded, they retained their respective scarless and scarred healing responses¹³. Adult tendon grafts healed with significant disruption in collagen fiber alignment, formation of granulation tissue, and inferior mechanical properties. In contrast, fetal tendon grafts healed scarlessly and regained normal tissue properties. Notably, SCID mice mount inflammatory responses to injury, despite lower T-cell and B-cell levels¹⁴. Based on these studies, an immature immune system is not the primary reason for scarless tendon healing. Similar findings of fetal scarless healing vs. adult scarred healing have been reported for skin in human and sheep^15–18, whereas some fetal tissues, such as alimentary tract and diaphragm tissue, heal with scar regardless of developmental stage^{19, 20}. Taken together, an immature immune system is unlikely the major determinant of fetal scarless tendon healing.

These findings suggest scarless healing ability is intrinsic to the fetal (“embryonic” in other species, such as mouse) tissue. We propose that tendon cells are key regulators of tendon healing outcomes. We hypothesize that tendon cells of scarless and scarring healing ages possess intrinsic differences that lead to divergent responses to pro-inflammatory cytokines (e.g., IL-1β) and downstream regulation of molecules involved in ECM synthesis and degradation. In sheep, skin and tendon follow similar fetal scarless healing mechanisms, with fetal skin and tendon both healing scarlessly as late as 100 days of gestation^{16, 21–23}. Skin transitions from scarless to scarred healing in the sheep fetus at 120 days of gestation, at the beginning of the third trimester in human, and in mouse at 18 days of gestation (embryonic day (E) 18)^{16, 17, 23–25}. By E14.5 in mouse, the complex patterns of mature limb tendons are fully formed and marked by scleraxis (Scx)^26–28. Based on this, we chose E15 to represent a scarless healing stage for tendon. While the transition to scarred tissue healing occurs prenatally, injured early postnatal mouse limb tendons have been shown to heal more regeneratively than adult tendons²⁹. Thus, we chose postnatal day (P) 7 to represent a scarred tendon healing age that retains some regenerative capacity, with the idea that observed differences in P7 vs. E15 cells will identify major determinants that contribute to scarred vs. scarless healing outcomes. In the present study, following the skin healing paradigm, we characterized how P7 and E15 tendon cells regulate key molecules in response to IL-1β treatment. Identifying scarless tendon healing mechanisms will pave the path to developing cell-targeted strategies to redirect adult scarred tendon healing toward scarless outcomes.

MATERIALS AND METHODS

Experimental Overview.

Embryonic and postnatal mouse tendon cells were seeded in monolayer, cultured for 24 h in growth medium, followed by 24 h in reduced-serum medium, and then treated for 24 h with IL-1β or vehicle control. Samples were harvested after 15 min and 24 h to examine signaling pathway activation, and after 24 h to characterize mRNA and protein levels of tendon markers, inflammatory mediators, collagens, and MMPs. Results were statistically analyzed to identify significant changes. Materials were from Invitrogen (Carlsbad, CA) unless otherwise specified.

Tendon Cell Isolation and Culture.

Scx-(green fluorescent protein) GFP-expressing tendon cells were isolated from limbs as previously described^{30, 31}. Briefly, P7 and pregnant ScxGFP mice were sacrificed according to IACUC guidelines. E15 embryos were harvested from the pregnant mice and staged³², and limbs were harvested. ScxGFP-expressing cells were isolated from digested limbs of the litter via cell sorting by GFP signal (MoFlo Legacy, Beckman Coulter). Three independent P7 and E15 limb cell pools (litters) were harvested. Tendon cells were expanded to passages between 3 and 5 in growth medium (GM: high glucose Dulbecco’s Modified Eagle Medium (DMEM), 10% fetal bovine serum (FBS), 1% penicillin/streptomycin (P/S)) at 37°C and 5% CO₂ for experiments.

IL-1β Treatment.

Tendon cells were seeded at 30,000 cells/cm² on tissue culture plastic and cultured for 24 h in GM, followed by 24 h in reduced-serum medium (DMEM, 1% FBS, 1% P/S). At this time (0 h, designated as baseline condition), cells were washed with PBS and fed with reduced-serum medium supplemented with either 100 pM recombinant mouse IL-1β (R&D Systems) or phosphate buffered saline (PBS) without Ca²⁺ or Mg²⁺ as vehicle control.

Western Blot.

Tendon cells were harvested after 15 min and 24 h treatment and lysed in RIPA buffer (Boston BioProducts) containing 1% (v/v) Halt protease inhibitor (ThermoScientific). Following sonication in sodium dodecyl sulfate (SDS), lysates were boiled and used for western blotting. Blots were probed with antibodies against p38 MAPK, phospho-p38 MAPK, p42/p44 ERK, phospho-p42/p44 ERK, NF-κB p65, phospho-NF-κB p65 (1:1000; Cell Signaling Technology), β-actin (1:1000; Abcam), and HRP-conjugated secondary antibodies (1:1000; Abcam). Protein content was quantified with densitometry using ImageJ (NIH), normalized to β-actin, and then normalized to total p38 MAPK, p42/p44 ERK, or NF-κB p65 content.

Quantitative Polymerase Chain Reaction (qPCR).

Cells harvested after 24 h treatment were homogenized in TRIzol reagent (Life Technologies). RNA was isolated and reverse-transcribed using SuperScript III First-Strand Synthesis System (Life Technologies). qPCR was performed with Brilliant II SYBR Green qPCR Master Mix (Agilent Technologies) on a Stratagene Mx3000P qPCR system (SAgilent Technologies). Mouse-specific primers for 18S, IL-6, TNFα, COX-2, IL-1R1, IL-1RA, MMP-3, MMP-13, collagen type (Col) I, Col III, Scx, and Tenomodulin (Tnmd) were used. Fold change values were calculated using the 2^−ΔΔCt method and normalized to 18S.

Enzyme-Linked ImmunoSorbent Assay (ELISA).

Conditioned medium was collected after 24 h treatment. IL-6, MMP-3, and IL-1RA protein content were determined using mouse-specific Quantikine ELISA kits (R&D Systems). DNA content was determined using the Quant-iT PicoGreen dsDNA Assay Kit (Life Technologies) and divided by 5.6 pg/cell, the approximate amount of DNA in a murine diploid cell^{33, 34}, to obtain total cell number. Reported protein content was normalized to respective cell counts.

Statistics.

Unpaired Student’s t-test was performed to compare IL-1β and vehicle control (PBS) treatments within each cell type, and to compare P7 cells and E15 cells under each condition tested, using p<0.05 to determine statistical significance. Statistical analyses were performed with Graphpad Prism (GraphPad Software Inc., San Diego, CA). All data are shown as mean ± standard deviation.

RESULTS

Baseline mRNA levels of IL-1RA, IL-6, and TNF-α were higher in P7 than E15 cells (p<0.05), whereas Scx, Tnmd, Col I, Col III, IL-1R1, COX-2, MMP-3, and MMP-13 did not differ (Fig 1a). Secreted protein levels of IL-1RA were higher in P7 than E15 cells (p<0.05), whereas IL-6 trended higher (p=0.07), and MMP-3 did not differ (Fig 1b).

Figure 1.

Baseline gene expression and secreted protein levels of IL-1RA, IL-6, and MMP-3 in P7 and E15 tendon cell cultures. (a) Gene expression levels of P7 and E15 cells were first normalized to 18S, and then to gene expression levels of P7 cells. (b) Secreted protein levels of P7 and E15 cells were first normalized to β-actin, and then to protein levels of P7 cells. (N=3, *p<0.05).

IL-1β treatment upregulated IL-1R1, IL-6, MMP-3, and MMP-13 expression levels in both P7 (Fig. 2a) and E15 cells (Fig. 2b) (p<0.05). COX-2 was also upregulated in P7 (p<0.05) (Fig. 2a) and appeared to increase in E15 cells but was not statistically different from the control (Fig. 2b). In contrast, TNF-α, IL-1RA, Col I, Col III, Scx, and Tnmd gene expression levels were not affected by IL-1β in either cell type. To directly compare P7 and E15 cell responses, we normalized mRNA levels (after normalization to 18S) to that of P7 tendon cells, revealing that IL-1β treatment induced P7 cells to express much higher gene expression levels of IL-6, TNF-α, COX-2, MMP-3, and MMP-13 than E15 cells (p<0.05) (Fig. 2c).

Figure 2.

IL-1β treatment effects on gene expression of P7 cells and E15 cells after 24h of culture. (a) Gene expression levels of P7 cells were normalized to 18S, and then to gene expression levels of vehicle control-treated P7 cells. (b) Gene expression levels of E15 cells were normalized to 18S, and then to gene expression levels vehicle control-treated E15 cells. Dotted lines in (a) and (b) represent the respective vehicle control gene expression levels. (c) To directly compare P7 and E15, gene expression levels of P7 and E15 cells were normalized 18S, and then to gene expression levels of P7 cells. (N=3, *p<0.05).

IL-1β also affected P7 and E15 cells protein secretion. Compared to their respective vehicle control treatments, IL-1β-treated P7 cells secreted higher levels of IL-1RA, IL-6, and MMP-3 protein (p<0.05), whereas E15 cells only secreted higher levels of IL-6 protein (p<0.05) (Fig 3). We also observed that IL-1β-treated P7 cells secreted much higher IL-1RA and IL-6 protein levels than E15 cells (p<0.05) (Fig. 3).

Figure 3.

IL-1β treatment effects on secreted protein levels of IL-1RA, IL-6, and MMP-3 of P7 and E15 tendon cells after 24h of culture. IL-1RA, IL-6, and MMP-3 protein levels were first normalized to β-actin, and then normalized to 10,000 cells. (N=3, *p<0.05).

In both E15 and P7 cells, IL-1β treatment upregulated phosphorylated levels of p38 MAPK and NF-kB p65 (p<0.05) after 15 min (Fig. 4) but not after 24 h of treatment (data not shown). After 15 min, phosphorylated levels of p42/p44 ERK appeared to increase trend-wise compared to vehicle controls in both E15 and P7 cells, but not statistically significantly (Fig. 4), and no trends were observed after 24 h of treatment (data not shown). We also observed that 15 min of IL-1β treatment induced greater phosphorylated p38 MAPK levels in P7 than E15 cells (p<0.05), whereas phosphorylated levels of p42/p44 ERK and NF-κB p65 did not differ between P7 and E15 cells with either IL-1β or vehicle control treatments (Fig. 4).

Figure 4.

IL-1β treatment effects on phosphorylation of p38 MAPK, p42/p44 ERK, and NF-κB p65 proteins in P7 and E15 cells after 15 min of culture. (a) Representative western blot images of phosphorylated (p-) and total p38 MAPK, p42/p44 ERK, and NF-κB p65 in P7 and E15 cells. (C: control; I: IL-1β) (b) Densitometry analysis of western blots of phosphorylated and total protein levels, normalized to β-actin. N=3, *p<0.05.

DISCUSSION

In an injured adult tendon, it is believed that pro-inflammatory cytokines promote imbalances in anabolic and catabolic functions during healing, and that these imbalances lead to the formation of scar tissue with abnormal ECM content and organization. In contrast, injured fetal sheep tendon grafts retain their native scarless healing ability even when cultured in an animal with an adult immune system¹³. Based on these findings, we propose tendon cells are key regulators of tendon healing outcomes, and hypothesize that their responses to pro-inflammatory cytokines are primary determinants of whether tendon will heal scarlessly or with scar. To that end, the current study examined how representative scarless and scarred healing tendon cells (E15 and P7 cells, respectively) regulate molecules related to anabolic and catabolic functions in response to IL-1β, a cytokine that appears nearly immediately when tendon is injured^{5, 6, 8, 35, 36}. We chose a 24 h timepoint based on reports that adult tendon cells alter mRNA and protein levels as early as 24 h after treatment with IL-1β^{5, 8}. Similar gene expression levels of tendon markers Scx and Tnmd between P7 and E15 suggested the cells are similarly tenogenically committed, agreeing with our previous results that E15 and P7 mouse limb tendon cells express similar levels of Scx, Tnmd, Col I, and elastin³⁰. At baseline and also after IL-1β treatment, P7 and E15 cells expressed anabolic and catabolic molecules differently, revealing that tendon cells of scarless and scarred healing stages are intrinsically different. Notably, P7 responded with larger changes than E15 cells in upregulation of inflammatory mediators (IL-6, COX-2) as well as catabolic enzymes (MMP-3, MMP-13). Our novel findings establish a platform for future studies to investigate how tendon cells regulate age-specific scarred and scarless healing responses.

Collagen ratios (e.g., Col I:III), fiber structure, and organization are typically abnormal in tendons after healing³⁷. Consequently, scar tissue forms. In our study, Col I and III gene expression levels were unaffected by IL-1β treatment for both P7 and E15 cultures, which differ from reports that IL-1β downregulates Col I levels in adult human and mouse tendon cells^{7, 38}. These differences between studies could be due to variations in timepoints, IL-1β concentration, and culture conditions. Additionally, changes in collagen protein levels due to post-transcriptional mechanisms can occur without corresponding changes in gene expression levels³⁹. Future studies should consider the effects of these variables on gene expression, synthesis, and remodeling of collagen and other ECM proteins.

IL-1β-induced upregulation of MMPs is thought to transition adult tendon from a homeostatic to catabolic state^{5, 6, 8, 35, 36}. MMP-3, which degrades tendon components Col III, proteoglycans, fibronectin, and elastin^{40, 41}, is lower in both mRNA and protein forms in ruptured and tendinopathic human tendons^{4, 42} but higher in injured canine and rat tendons^{4, 43}. While the reasons for these inconsistent findings are unknown, these studies point to a role for MMP-3 in tendon healing. MMP-13, which degrades Col XIV and proteoglycans^{44, 45}, is higher in both mRNA and protein forms in injured canine, rabbit, and rat tendons^{2, 4, 43}. IL-1β treatment upregulates MMP-3 and −13 mRNA and protein levels in postnatal tendon cells of various species^{5, 6, 8, 35, 36}, and treatment of postnatal rat tendon cells with siIL-1β abrogates increases in MMP-13 mRNA expression in vitro⁴⁶. Interestingly, MMP-3 gene expression is higher in wild-type mouse skin wounds, which heal with scar, than in nude mouse skin wounds, which heal scarlessly⁴⁷. In our study, IL-1β upregulated MMP-3 to higher mRNA and protein levels in P7 cultures than in E15 cultures (Fig. 2c, Fig. 3). Based on these results, perhaps scarred healing tendon cells (e.g., P7 cells) respond to pro-inflammatory cytokines by inducing MMP-mediated degradation of ECM and activation of growth factors and other MMPs that promote scar tissue formation. In contrast, if the changes in MMP activity regulated by scarless healing age embryonic tendon cells (e.g., E15 cells) are too small to create imbalances between anabolic and catabolic functions, regeneration of normal (scarless) tissue may still proceed. Future studies should perturb MMP activity and assess scarless and scarred healing outcomes. Additional MMP roles should also be examined, including activation of cytokines, growth factors, and other MMPs^{48, 49}.

The role of IL-6 in tendon healing is unclear, as IL-6 can be “anti-inflammatory” or “pro-inflammatory”⁵⁰. Ruptured and tendinopathic human tendons both possess higher IL-6 levels than healthy tendons^{51, 52}. Injured tendons of IL-6−/− mice have inferior mechanical properties and higher IL-1β, TNF-α, and Col I and III protein levels¹¹. While IL-6 treatment of human tendon cells in vitro has no detectable effects on TNF-α, IL-1β, IL-10, MMP-1, or elastin mRNA levels or Col I protein levels⁵³, IL-6 promotes collagen synthesis in human Achilles tendon in vivo⁵⁴. IL-6 also enhances IL-1β upregulation of MMP-3 in human nucleus pulposis cells⁵⁵. In our study, IL-1β treatment upregulated IL-6 mRNA and IL-6 protein levels by 12.5 and 3.1 times higher, respectively, in P7 cultures than in E15 cultures (p<0.05) (Fig. 2c, Fig. 3). Taken together, we hypothesize that tendon cells of scarred healing ages (e.g., P7) promote scar formation during tendon healing by dramatically increasing IL-6 production and altering IL-6-regulated levels of collagens and MMPs in response to pro-inflammatory cytokines. In contrast, perhaps a lower upregulation of IL-6 by scarless healing embryonic tendon cells (e.g., E15 cells) has insignificant effect on the balance of matrix synthesis and MMP activities, and thus allows for regeneration of normal (scarless) tissue during healing. Consistent with our hypothesis, adult human skin fibroblasts express higher IL-6 mRNA levels than fetal skin fibroblasts of a scarless healing stage, and adult human skin wounds that heal with scar produce IL-6 more persistently than fetal wounds that heal scarlessly¹⁵. Furthermore, fetal skin wounds treated with exogenous IL-6 heal with scar¹⁵. Based on these data, IL-6 may play an important role in the extent of scarring during tendon healing, and should be investigated further.

COX-2 regulates synthesis of prostaglandin E2 (PGE2), a pro-inflammatory mediator that impairs collagen production in vitro and during tendon injury and healing⁵⁶. IL-1β treatment has been reported to upregulate COX-2 gene expression and PGE2 levels in tendon cells in vitro^5–8. Interestingly, COX-2 mRNA and protein as well as PGE2 levels are higher in E18 fetal mouse skin wounds during scarred healing than in E15 fetal mouse skin wounds during scarless healing²⁵. Here, IL-1β treatment significantly upregulated COX-2 gene expression in P7 but not E15 cells (Fig. 2a). Perhaps scarred healing tendon cells (e.g., P7 cells) promote scarring by upregulating COX-2 to decrease collagen production in response to the pro-inflammatory cytokines that appear at injury.

TNF-α is another pro-inflammatory factor expressed during adult tendon healing^{4, 51}. Interestingly, in our study TNF-α mRNA levels were unaffected by IL-1β treatment, but were consistently expressed at lower levels in E15 than P7. Despite these interesting differences in mRNA levels between cell types, TNF-α protein was undetectable in both cultures. Although TNF-α is a pro-inflammatory factor expressed during adult tendon healing, others have also reported TNF-α protein is undetectable in tendon cells before and after cytokine treatment⁶.

IL-1β binds IL-1R1 to transduce signaling intracellularly⁵⁷, whereas IL-1β binding to IL-1R2 inhibits signaling⁵⁸. Adult tendon cells express IL-1R1, but IL-1R2 is barely detectable⁸, and thus we focused on IL-1R1. IL-1R1 gene expression was upregulated by IL-1β treatment in both P7 and E15 cells, but did not differ between P7 and E15 cells either at baseline (Fig. 1) or after IL-1β treatment (Fig. 2). These results suggest the different P7 and E15 responses to IL-1β were not due to a difference in IL-1R1 levels. In contrast to IL-1β, IL-1RA binds to IL-1R1 to inhibit IL-1 signaling^{59, 60}. IL-1RA treatment of healthy human patellar tendon cells inhibits IL-1β-induced upregulation of COX-2 and PGE2⁷. Treatment of tendinopathic rat tendons with IL-1RA prevents downstream production of MMP-3 and PGE2 that typically occur with disease progression⁶¹. Interestingly, IL-1RA knockout mice have delayed skin wound closure, decreased collagen production, and higher IL-1β levels at the wound site after skin injury, compared to wildtype mice¹⁰. Taken together, IL-1RA is considered an anti-inflammatory mediator that at high enough concentration may be expected to reduce the scarring effects of inflammation. However, IL-1β treatment induced P7 cells, but not E15 cells, to secrete higher levels of IL-1RA than control samples (Fig. 3), and IL-1RA levels were higher in P7 than in E15 IL-1β-treated cultures (Fig. 3). One potential explanation for why P7 tendon would heal with scar despite much higher secretion of IL-1RA than E15 cells is that the upregulation of inflammatory mediators, such as IL-6 and COX-2, outcompeted the effects of IL-1RA. It is also possible that IL-1β induced P7 cells to produce more IL-1β, which then outcompeted IL-1RA in binding IL-1R1. Future studies could perturb IL-1β and IL-1RA levels to test these hypotheses.

IL-1β signaling via IL-1R1 can activate p38 MAPK, p42/p44 ERK, and NF-κB p65 pathways⁶². We first characterized phosphorylated levels of the signaling proteins at 24 h after treatment, but did not observe significant differences (data not shown). We then focused on 15 min after treatment, based on reports that phosphorylated p38 MAPK levels peak in chondrocytes at 15 min after IL-1β treatment^{35, 63}. IL-1β treatment increased phosphorylated levels of p38 MAPK and NF-kB p65 in both cultures, with significantly higher phosphorylated p38 levels in P7 than E15 cells (Fig. 4). Phosphorylated p42/p44 ERK levels appeared to increase in both E15 and P7 cells, but were not significantly different between cell types or compared to control treatment (Fig. 4). Little is known about these pathways in scarless wound healing, however IL-1β treatment of adult human tendon cells upregulates COX-2 and MMPs via activation of the p38 MAPK and NF-κB pathways and does not activate the p42/p44 ERK pathway^{7, 35}. Treatment of human patellar tendon fibroblasts with p38 MAPK inhibitor SB203580 abrogates IL-1β-induced increases in COX-2 and PGE2⁷. Here, IL-1β treatment led to lower phosphorylated p38 MAPK, IL-6, COX-2, and MMP-3 and −13 levels in E15 than P7 cultures (Fig. 2, 3, 4). Perhaps lower upregulation of inflammatory mediators in E15 cultures was due to less p38 MAPK signaling. While the consistent trend that we observed across three biological samples is compelling, we present these findings with caution, as densitometry can be subject to imaging artifacts and technical errors. Furthermore, some of the observed trends, such as with p42/p44 ERK phosphorylation, could be significant at different IL-1β concentrations and timepoints. Future studies that perturb these pathways would strengthen these results.

Based on our findings, we hypothesize that tendon cells play a significant role in determining scarless vs. scarred healing outcomes (Fig. 5). Our study showed that scarred healing tendon cells respond to pro-inflammatory cytokines by dramatically upregulating inflammatory mediators and MMPs, whereas scarless healing tendon cells respond with smaller changes. Future studies should consider additional cytokines present in tendon injuries. We used 100pM of IL-1β, which induces adult mouse tendon cell responses in vitro^{5, 8}, whereas 10 pM IL-1β has no effect^{5, 8}. However, IL-1β concentration of full-thickness torn rabbit supraspinatus tendon tissue one day after injury is approximately 10 pM³. It may be worthwhile to perform a dose-response study considering cytokine levels fluctuate with injury severity and healing stage. Future studies should also focus on protein and enzyme activity levels and cell membrane-bound receptor distribution since gene expression and protein levels may not reliably coincide^{39, 64–66}. Cell migration and traction-mediated wound closure should be considered as well. As mouse embryo tendons are too small to access in utero and sheep fetus surgery is prohibitively expensive, an economical embryo tendon injury model is needed. The transition stage from scarless to scarred tendon healing will need characterization. P7 represents a scarred tendon healing age, but future studies will include tendons at older ages that commonly sustain injuries. Future research based on these findings may elucidate key mechanisms of scarless tendon healing to inform the development of novel therapeutics for adult tendon regeneration.

Figure 5.

Schematic to highlight significance of findings. We propose IL-1β activation of p38 MAPK signaling in scarred healing age (e.g., P7) tendon cells is sufficient to create an imbalance in regulation of anabolic vs. catabolic functions, whereas IL-1β effects on scarless healing age (e.g., E15) tendon cells are too small to offset this critical balance.