Hedgehog Activation for Enhanced Rotator Cuff Tendon-to-Bone Healing

Abstract

Background:

Rotator cuff repair is a common orthopaedic procedure, yet the rate of failure to heal after surgery is high. Repair site rupture is due to poor tendon-to-bone healing and lack of regeneration of the native fibrocartilaginous enthesis. During development, the enthesis is formed and mineralized by a pool of progenitors activated by hedgehog signaling. Furthermore, hedgehog signaling drives regenerative enthesis healing in young animals, in contrast to older animals, in which enthesis injuries heal via fibrovascular scar and without participation of hedgehog signaling.

Hypothesis:

Hedgehog activation improves tendon-to-bone healing in an animal model of rotator cuff repair.

Study Design:

Controlled laboratory study.

Methods:

A total of 78 adult Sprague-Dawley rats were used. Supraspinatus tendon injury and repair were completed bilaterally, with microsphere-encapsulated hedgehog agonist administered to right shoulders and control microspheres administered to left shoulders. Animals were sacrificed after 3, 14, 28, or 56 days. Gene expression and histological, biomechanical, and bone morphometric analyses were conducted.

Results:

03 days, hedgehog signaling pathway genes Gli1 (1.70; P = .029) and Smo (2.06; P = .0173), as well as Runx2 (1.69; P = .0386), a transcription factor of osteogenesis, were upregulated in treated relative to control repairs. At 14 days, transcription factors of tenogenesis, Scx (4.00; P = .041), and chondrogenesis, Sox9 (2.95; P = .010), and mineralized fibrocartilage genes Col2 (3.18; P = .031) and Colx (1.85; P = .006), were upregulated in treated relative to control repairs. Treatment promoted fibrocartilage formation at the healing interface by 28 days, with improvements in tendon-bone maturity, organization, and continuity. Treatment led to improved biomechanical properties. The material property strength (2.43 vs 1.89 N/m²; P = .046) and the structural property work to failure (29.01 vs 18.09 mJ; P = .030) were increased in treated relative to control repairs at 28 days and 56 days, respectively. Treatment had a marginal effect on bone morphometry underlying the repair. Trabecular thickness (0.08 vs 0.07 mm; P = .035) was increased at 28 days.

Conclusion:

Hedgehog agonist treatment activated hedgehog signaling at the tendon-to-bone repair site and prompted increased mineralized fibrocartilage production. This extracellular matrix production and mineralization resulted in improved biomechanical properties, demonstrating the therapeutic potential of hedgehog agonism for improving tendon-to-bone healing after rotator cuff repair.

Clinical Relevance:

This study demonstrates the therapeutic potential of hedgehog agonist treatment for improving tendon-to-bone healing after rotator cuff injury and repair.

Rotator cuff repair is one of the most commonly performed orthopaedic surgeries, with >450,000 surgeries performed every year in the United States.^6,23 Unfortunately, the rate of postoperative failure to heal remains unacceptably high despite advances in surgical techniques and rehabilitation strategies.^9,25 Clinical studies have shown that only about half of repaired rotator cuffs heal by 12 months after surgery, with the rate of recurrent tears ranging from 20% in the best case to as high as 94% in certain patient populations with massive tears.^9,25 Animal model studies indicate that the high rate of postoperative failure to heal after surgical repair is due to a lack of regeneration of the native tendon-to-bone transitional tissue.^13,17,30,34 This native transition zone, the enthesis, is a functionally graded, mineralized fibrocartilaginous structure that minimizes stress concentrations at the interface between the mechanically dissimilar tissues of tendon and bone.^10,12,17,33 Rather than re-formation of this fibrocartilaginous transition, which is necessary for effective load transfer between tendon and bone, an abrupt transition of fibrovascular scar forms at the healing interface and the underlying bone resorbs.^13,17,30,34

The hedgehog (Hh) signaling pathway is integral to the growth and patterning of a number of musculoskeletal structures and is necessary for mineralizing tissues such as the growth plate.^4,8 The formation and mineralization of a functional rotator cuff enthesis are driven by a unique population of Hh-responsive cells.²⁹ One study showed that both the ablation of these cells and the conditional deletion of Smoothened (SMO), a transmembrane protein necessary for Hh signaling, impaired enthesis fibrocartilage formation.²⁹ Another study showed that, in contrast to adult animals, neonatal animals were able to regenerate the rotator cuff enthesis after injury, and this process was driven by Hh signaling.²⁸ Additional studies have shown that the administration of Hh agonist (HhAg) to a murine fracture model enhanced bone formation.¹⁸ This finding is of particular relevance given the lack of fibrocartilage formation and mineralization at the healing tendon-to-bone interface after rotator cuff injury and the promising effects of mineralization-promoting treatments for enhancing healing.^30,32

Motivated by the poor outcomes after rotator cuff repair and the potential to enhance healing by promoting fibrocartilage formation, the objective of the current study was to evaluate the effect of pharmacologic upregulation of the Hh signaling pathway on tendon-to-bone healing in a clinically relevant rat model of rotator cuff repair. We hypothesized that HhAg would promote rotator cuff enthesis regeneration, as evidenced by increased fibrocartilage formation and mineralization, leading to enhanced mechanical properties.

METHODS

A total of 78 adult Sprague-Dawley rats (~3–4 months old; 43 females and 35 males; Envigo) were used to test the effect of HhAg on rotator cuff tendon-to-bone healing (Figure 1). All procedures were approved by Columbia University’s Institutional Animal Care and Use Committee (protocol AC-AAAZ2456).

Figure 1.

In a paired study design, bilateral rotator cuff injury and repair were performed in rats to test the efficacy of hedgehog (Hh) agonism on tendon-to-bone healing. Outcome measures at 3, 14, 28, and 56 days included gene expression, histology, biomechanics, and micro–computed tomography (microCT).

Animal Care

Rats were given a standard irradiated diet (Purina Lab Diet #5053). They were exposed to standard 12:12 hour light cycles. Rats were typically housed in pairs but occasionally were housed singly.

Microsphere Synthesis

A small molecule HhAg (Hh-Ag1.5; Cellagen) was encapsulated in poly(D,L-lactide-co-glycolide) (PLGA; 75:25 lactideglycolide; Sigma P1941) as previously described.^26,27,31 A solution of 2 mg of HhAg dissolved in 0.5 mL of dimethyl sulfoxide (DMSO) was added to a PLGA solution consisting of 200 mg of PLGA dissolved in 4.5 mL of dichloromethane. The polymer-drug mixture was added dropwise to a 1% poly(vinyl alcohol) solution and stirred at 450 rpm for 4 hours. Unloaded microspheres were fabricated by adding 0.5 mL of DMSO to the PLGA solution. Microspheres were collected and purified by multiple rounds of washing in double-distilled water and centrifugation at 1500 RCF and subsequently were lyophilized for 48 hours and stored at 220°C until further use. As reported in previous publications, the described approach is highly efficient, with little to no drug lost during the process.^26,27,31 Microsphere delivery was used to sustain the release of HhAg through the remodeling phase of tendon-to-bone healing. In vitro HhAg release, as measured by spectrometry, showed that with a starting concentration of 0.167 mg/mL, the cumulative release of HhAg from microspheres increased over 12 days from ~0.02 mg/mL to ~0.08 mg/mL (ie, approximately half of the HhAg was released over the course of 12 days) (Figure 2). Based on this rate, all HhAg would be released by day 21.

Figure 2.

In vitro cumulative release of hedgehog agonist from poly(D,L-lactide-co-glycolide) microspheres increased over 12 days from ~0.02 mg/mL to ~0.08 mg/mL (starting concentration 0.167 mg/mL).

Rotator Cuff Animal Model

Supraspinatus tendon tenotomy and repair were completed bilaterally, as previously described.^13,30,34 In brief, after the administration of inhaled isoflurane and subcutaneous buprenorphine and bupivacaine, a scalpel was used to detach the supraspinatus tendon from the humeral head at its attachment site. Using a paired study design in each animal, either 1.3 mg of HhAg encapsulated in PLGA microspheres or control PLGA microspheres were administered to the tendon-to-bone interface before surgical repair; right-sided repairs were assigned to the treated group and left-sided repairs were assigned to the control group. The dose of HhAg administered was determined using data from previous literature.¹⁸ Specifically, we calculated the locally applied in vivo dose as follows: 20 mg/kg (systemic mouse dose¹⁸) × 0.025 kg (mouse weight) = 0.5-mg dose per mouse; rat supraspinatus tendon length = 2.67 × mouse supraspinatus tendon length; to scale the dose, 0.5 mg × 2.67 = 1.3-mg local dose for the rat supraspinatus tendon repair. Direct visualization of microspheres at the time of delivery indicated some dispersion after application. However, the vast majority of microspheres remained adhered to the supraspinatus footprint after application, and no obvious off-target effects were seen at dissection and evaluation postmortem. After microsphere delivery, the supraspinatus tendon was repaired to its native humeral head attachment site with a No. 5–0 Vicryl 18-inch C-3 cutting suture, which was passed through the tendon in a modified Mason-Allen configuration and then passed through a bone tunnel in the humeral head, distal to the growth plate. After surgery, animals were allowed free cage activity, were checked daily for signs of pain or distress, and were sacrificed at 3, 14, 28, or 56 days after surgery. Supraspinatus tendon entheses were isolated from animals postmortem and used for gene expression analysis at 3, 14, and 28 days. Supraspinatus tendonhumeral head bone units were isolated from animals postmortem and used for histological analysis (all time points) or bone morphometry and biomechanical analysis (28 and 56 days).

Gene Expression

At 3, 14, and 28 days postoperatively, bilateral supraspinatus tendons were harvested and processed for reverse transcription quantitative real-time polymerase chain reaction (PCR) (n = 3–5 per time point). Immediately after animal sacrifice, the supraspinatus tendon entheses were microdissected (carefully avoiding inclusion of bone or muscle), snap frozen in liquid nitrogen, and homogenized via ball mill (Mikro-Dismembrator U; Sartorius). Next, 1 mL of TRIzol (Invitrogen) was added. Phase lock gel tubes (5PRIME phase lock gel Heavy; QuantaBio) were then used to separate nucleic acids from proteins, spin columns (PureLink RNA Mini Kit; Invitrogen) with on-column enzymatic DNA digestion (DNase I; Invitrogen) were used to purify RNA, and a spectrophotometer (Take3, Cytation 5; Biotek) was used to determine RNA yield and sample purity. A thermocycler (ProFlex; Applied Biosystems) was used to reverse transcribe 100 ng of RNA to complementary DNA (cDNA; High-Capacity RNA-to-cDNA kit; Invitrogen). A real-time PCR system (Quantstudio 6 Flex; Applied) and SYBR (PowerUp SYBR green master mix; Invitrogen) were used to perform quantitative real-time PCR. Primequest (Integrated DNA Technologies) was used to design primers. Analysis was conducted via the delta Ct method with Gapdh as the housekeeping gene. Genes involved in hedgehog signaling (Gli1, Smo), tenogenesis (Scx), osteogenesis (Runx2), chondrogenesis (Sox9), and fibrocartilage extracellular matrix (collagen type 2 [Col2], collagen X [Colx]) were assessed.

Histological Analysis

At 3, 14, 28, and 56 days postoperatively, tendon-to-bone units were isolated, fixed in 4% paraformaldehyde for 48 hours, decalcified in formic acid (Immunocal; StatLab), dehydrated in ethanol, and embedded in paraffin (n = 4 per time point). Next, 5-μm sections were collected and stained with hematoxylin and eosin and Toluidine blue. Blinded, semiquantitative evaluation was performed by 3 observers (S.T., X.F., A.J. Luzzi) for cellularity, fibroblast shape, matrix maturity, cell orientation, insertion organization, insertion continuity, bone quality, and overall maturity, according to the rubric defined previously (Table 1).¹³

TABLE 1.

Tendon-to-Bone Histological Scores^a

	3 Days		14 Days		28 Days		56 Days
	HhAg	Control	HhAg	Control	HhAg	Control	HhAg	Control
Cellularity	4 (4)	3.9 (3.5–4)	2.9 (2–4)	3.6 (3–4)	2.7 (2–3.5)	2.8 (2–3.5)	2.5 (2–3)	2.8 (2–3)
Fibroblast shape	3.9 (3.5–4)	3.8 (3.5–4)	2.9 (2–4)	2.6 (2–3.5)	2.6 (2–3)	2.6 (2–4)	2.3 (2–3)	2.6 (2–3.5)
Matrix maturity	3.8 (3–4)	3.7 (3–4)	3.1 (2–4)	3 (2–4)	2.7 (2–3)	3.1 (3–3.5)	2.5 (2–3)	2.5 (2–3)
Cell orientation	3.8 (3–4)	3.8 (3–4)	2.9 (2–4)	2.8 (2–4)	2.7 (2–3)	2.5 (2–4)	2.3 (2–3)	2.6 (2–3)
Insertion organization	3.8 (3–4)	4 (4)	2.9 (1.5–4)	2.7 (2–4)	2.1 (2–3)	2.8 (2–3.5)	2.3 (2–3)	2.7 (2–3)
Insertion continuity	3.7 (3–4)	3.8 (3.5–4)	2.3 (2–3)	2.6 (1–3)	2.3 (2–3)	2.9 (2–3)	2.4 (2–3)	2.3 (2–3)
Bone quality	2.1 (2–3)	2.3 (2–3)	2.3 (2–3)	2.4 (2–3)	2.7 (2–4)	2.8 (2–4)	2.4 (1.3–3)	2.1 (1.5–3)
Overall maturity	25.1 (23–27)	25.3 (14–26)	19.9 (17–25)	19.7 (17–24)	17.7 (17–19)	19.5 (17–20)	16.6 (15–18)	17.4 (15–19)

^aResults are shown as the median (range).

Cellularity: 1 = low, 4 = high. Fibroblast shape: 1 = spindle-shaped, 4 = round. Matrix maturity: 1 = purple, 4 = light blue. Cell orientation: 1 = aligned, 4 = random. Insertion organization: 1 = fibrocartilage transition, 4 = fibrous/no transition. Insertion continuity: 1 = integrated, 4 = abrupt. Bone quality: 1 = enthesis lamellar and underlying trabecular bone similar to the healthy state, 4 = thin enthesis lamellar bone and loss of trabecular bone. Overall maturity (sum of individual scores): 7 = healthy/uninjured, 28 = disorganized scar tissue. HhAg, hedgehog agonist.

Immunohistochemistry

For immunohistochemical analysis, slides were first placed at 60°C for 35 minutes and then cooled to room temperature. The sections were deparaffined with xylenes for 5 minutes and then for an additional 2 minutes with fresh xylenes. Slides were rehydrated through graded ethanol, 2 minutes per wash: twice at 100%, twice at 95%, once at 70%, once at 50% ethanol, and once at 0% (ie, H₂O). Antigen retrieval was carried out in 100 μg/mL of hyaluronidase (H3506; Sigma-Aldridge) in phosphate-buffered saline (PBS) for 12 minutes at 37°C. Sections were then permeabilized in 1% Triton X-100 for 5 minutes and blocked in 5% goat serum in PBS plus Tween (PBST) for 1 hour at room temperature. Anti-mouse collagen X (COLX) (clone X53, 14–9771-95, 1:200; eBioscience) or anti-CD68 (ab213363, 1:200; Abcam) was incubated with the sections overnight at 4°C in blocking buffer containing 5% goat serum. After 3 PBST washes, Alexa Fluor-488 goat anti-mouse immunoglobulin G (IgG) (H 1 L; heavy and light chains) (A11001, 1:5000; ThermoFisher) or Alexa Fluor-555 goat anti-rabbit IgG (H 1 L) (A27039, 1:5000; Invitrogen) was incubated for 1 hour at room temperature. The sections were then washed 5 times and mounted with VectaShield Anti-fade Mounting Media with DAPI (H-1200; Vector Laboratories). Images were taken on a Zeiss Axio Observer microscope equipped with an Apotome 3 apparatus, and cell counts were made with Image J (National Institutes of Health).

Biomechanical Testing

After micro–computed tomography (microCT) scanning, humeri with attached supraspinatus tendons were biomechanically tested (n = 11 or 12 per time point). After removal of the supraspinatus muscle from the tendon, the humerus with associated supraspinatus tendon was placed in a custom 3 dimensional–printed grip and tested in uniaxial testing at 37°C in a saline bath (Electroforce 3230; TA Instruments).¹⁴ Sutures were cut before testing. Uniaxial tension testing consisted of 5 cycles of preconditioning (2% strain; 0.2% per second), followed by 300 seconds of rest, followed by extension to failure at a strain rate of 0.2% per second. Grip-to-grip displacement relative to initial gauge length was used to determine strain. MicroCT scans were used to measure the supraspinatus tendon cross-sectional area at the attachment site for stress calculation. Load-deformation curves were used to determine structural properties (stiffness, failure load, and work to failure [area under the curve through failure load]). Stress-strain curves were used to determine material properties (modulus, strength, and toughness [area under the curve through failure]).

Bone Morphometry

At 28 and 56 days postoperatively, humeri with attached supraspinatus tendons were harvested for bone morphometric analysis (n = 11 or 12 per time point). Samples were scanned using microCT at an intensity of 145 μA with an aluminum 0.25-mm filter, an energy of 55 kVp, and a standard resolution of 12.3 μm (Skyscan 1272; Bruker). A custom segmentation algorithm (CTAn; Bruker) was used to evaluate microCT images to separate cortical and trabecular bone. Bone volume fraction, bone and tissue mineral density, trabecular thickness, trabecular spacing, and trabecular number were calculated, and the values were compared between the 2 groups and across time points.

Statistical Analysis

For quantitative measures, the treated group was compared with the contralateral control at each time point using paired t tests to eliminate interspecimen variability. The effects of time (3, 14, 28, and 56 days) and treatment (HhAg vs control) were evaluated using analysis of variance (ANOVA). Parametric testing assumptions were assessed and confirmed. The Robust regression and Outlier removal (ROUT) method was used to identify outliers, which were excluded before analysis. Significance was set to P < .05. All statistical analyses were performed using GraphPad Prism 8. All data are presented as mean ± SD. The study was not powered to test for the effect of sex, but approximately equal numbers of male and female animals were used to ensure generalizability of the results.

RESULTS

HhAg Activated Hh Signaling and Promoted Expression of Fibrocartilage Genes at the Healing Tendon-to-Bone Interface

At 3 days, Hh signaling pathway genes Gli1 (1.70 ± 0.40; P = .029) and Smo (2.06 ± 0.53; P = .017), as well as Rumx2 (1.69 ± 0.44; P = .039), a transcription factor involved in osteogenesis, were significantly upregulated in HhAg-treated tissues relative to nontreated tissues (Figure 3). At 14 days, transcription factors for tenogenesis (Scx, 4.00 ± 2.48; P = .041) and chondrogenesis (Sox9, 2.95 ± 1.08; P = .010), as well as genes related to the extracellular matrix (ECM) of mineralized fibrocartilage (Col2, 3.18 ± 0.82, P = .031; Colx, 1.85 ± 0.29, P = .006), were upregulated in HhAg-treated tissues relative to control tissues. Col1, which codes for the tendon ECM protein type 1 collagen, approached significance at 14 days (7.28 ± 5.91; P = .061). At 28 days, none of the genes examined were significantly upregulated in treated shoulders relative to nontreated shoulders.

Figure 3.

Expression of genes related to hedgehog signaling, differentiation, and extracellular matrix in hedgehog agonist–treated tendon-to-bone attachments relative to controls (**P < .01, *P < .05, ^#P < .10). Error bars represent the SD. The y-axes for 2 and 28 days are in log2 scale to better visualize fold changes.

HhAg Promoted an Integrated Fibrocartilage Attachment at the Healing Tendon-to-Bone Interface

Tendon-to-bone healing was characterized by a substantial influx of inflammatory cells at the earlier time points followed by a fibrovascular scar response at the later time points (Figure 4; Table 1). At 3 and 14 days, tendon-to-bone histological scores were similar between treated and untreated groups, with a substantial improvement in scores over time (Table 1). Cellularity decreased, fibroblasts aligned with the tendon axis, and fibroblast shape shifted from rounded to spindle shape during the first 2 weeks of healing. The histology score improved further at the later time points, with an effect of HhAg treatment becoming clear by 28 days and continuing through 56 days. The effects of HhAg treatment were most apparent in the organization and continuity of the attachments: HhAg-treated repairs had fibrocartilaginous transitions with improved continuity and columnar arrangement of fibrochondrocytes between the tendon and bone compared with untreated repairs. Expression of CD68, a marker of macrophages, was elevated at all time points of healing, with few apparent differences between groups (Figure 5; Table 2). Expression of COLX was low at all time points examined, with similar levels between groups (Figure 5; Table 2).

Figure 4.

Representative histological results at 14, 28, and 56 days for control and hedgehog agonist (HhAg)–treated tendon-to-bone repairs. HhAg treatment led to improved fibrocartilage formation (purple staining) by 28 days and columnar alignment of fibrochondrocytes by 56 days (Toluidine blue stain; scale bar, 100 μm). Tendon (t) is on the left and bone (b) is on the right; 14 days, healing interface is indicated by dotted line; 28 days, healing interface is indicated between dotted lines; 56 days, disorganized fibrochondrocytes are indicated in control by dotted oval, and columnar arrangement of fibrochondrocytes is indicated in HhAg by dotted rectangles.

Figure 5.

Immunohistochemistry results for CD68 (red, top rows) and collagen X (COLX) (green, bottom rows). Cell nuclei are stained with DAPI (blue). Dotted lines delineate the bone and tendon borders. Hh, hedgehog; ND, not detected.

TABLE 2.

Expression of CD68 and COLX Based on Immunohistochemical Staining^a

	3 Days		14 Days		28 Days		56 Days
	HhAg	Control	HhAg	Control	HhAg	Control	HhAg	Control
CD68	2 (1–3)	2 (1–4)	1 (1)	3 (2–4)	1.7 (1–3)	ND	2.7 (1–4)	3 (2–4)
COLX	1.7 (1–2)	1.7 (1–3)	1 (1)	2 (1–4)	1.3 (1–2)	ND	1 (1)	1.7 (1–3)

^aResults are shown as the median (range).

1, <10% positive cells. 2, 10%−20% positive cells. 3, 20%−30% positive cells. 4, >30% positive cells. HhAg, hedgehog agonist; ND, not detected.

HhAg Enhanced the Biomechanical Properties of the Healing Tendon-to-Bone Interface

All samples failed at the tendon-to-bone repair site; no artifactual failures occurred at the growth plate or at the tendon grip. We found significant effects of treatment and time on tendon-to-bone biomechanical properties. At 28 days, we noted a significant increase in strength (2.43 vs 1.89 N/m²; P = .046) in the HhAg-treated samples compared with the paired control samples (Figure 6). At 56 days, a significant increase in work to failure (29.01 vs 18.09 mJ; P = .030) was found in the HhAg-treated samples compared with the paired control samples. Additionally, a significant effect of HhAg on stiffness (P = .022) was revealed in the ANOVA analysis. Also of note, maximum force (P < .0001), work to failure (P < .0001), and toughness (P < .0001) were significantly increased and modulus (P = .013) was significantly decreased between 28 and 56 days.

Figure 6.

Biomechanical properties of hedgehog agonist–treated samples and control tendon-to-bone samples (solid line over bars, P < .05; dashed line over bars, P < .10 based on paired t test). Error bars represent the SD. Effects of treatment and time revealed by analysis of variance are indicated below each plot.

HhAg Had Little Effect on the Morphology of the Bone Underlying the Healing Tendon-to-Bone Interface

We noted few significant effects of HhAg treatment on the morphometry of the humeral head bone underlying the repaired supraspinatus tendon. A significant increase in trabecular thickness was seen at 28 days (0.08 vs 0.07 mm; P = .035) (Figure 7). Cortical bone mineral density (P < .0001), trabecular bone mineral density (P < .0001), bone volume over total volume (P = .0001), and trabecular thickness (P = .0034) increased, and trabecular spacing (P = .0002) decreased between 28 and 56 days.

Figure 7.

Bone morphometry of hedgehog agonist (HhAg)–treated and control humeral head bone (solid line over bars, P < .05; dashed line over bars, .05 < P < .10 based on paired t test). Error bars represent the SD. Effects of treatment and time revealed by analysis of variance are indicated below each plot; arrow indicates supraspinatus footprint. BMD, bone mineral density; BV/TV, bone volume fraction; Tb.N, trabecular number; Tb.Sp, trabecular spacing; Tb.Th, trabecular thickness.

DISCUSSION

Failure to heal after rotator cuff repair can be largely attributed to insufficient regeneration of the mineralized fibrocartilage at the interface between tendon and bone.^13,17,30,34 The Hh signaling pathway is integral to the development and homeostasis of numerous tissues, including bone and enthesis.^4,8 Previous studies showed that the formation of the rotator cuff enthesis is driven by Hh signaling.^28,29 Specifically, this pathway is activated in a pool of progenitor cells at the interface of developing tendon and bone, leading to deposition and mineralization of fibrocartilage. It was further demonstrated that Hh signaling is active during regenerative neonatal enthesis healing and absent in scar-mediated adult enthesis healing. Consistent with this result, a recent enthesis healing study reported that mesenchymal stem cell treatment was associated with increased Hh activity, leading to more organized staining for COL2 and an increase in SOX9+ cells.³⁶ These previous studies motivated our current work investigating the therapeutic potential of Hh upregulation on tendon-to-bone healing. Because Hh signaling acts on enthesis progenitor cells to drive enthesis formation during postnatal development, we hypothesized that increasing Hh signaling in progenitor cells during enthesis healing should drive regeneration of the enthesis. We targeted Hh signaling during the proliferative through early remodeling phases of healing, when resident and migrating mesenchymal stem cells are present.

HhAg treatment activated Hh signaling at the healing tendon-to-bone attachment, as demonstrated by the upregulation of genes related to the Hh signaling pathway in the period immediately after repair. The activation of Hh signaling resulted in increased expression of transcription factors related to tenogenesis, chondrogenesis, and osteogenesis. Ultimately, the upregulation of these transcription factors led to the upregulation of genes involved in the production of fibrocartilage ECM. Notably, the lack of significant gene expression changes at the 28-day time point was consistent with the drug release profile from the microspheres, as in vitro testing suggested complete drug release by 21 days. This timeline is appropriate for in vivo healing, because increased Hh signaling should be maintained from the proliferative phase through the early remodeling phase of healing.

Histological analysis showed some effects of HhAg treatment on ECM production and mineralization, with improvements in tendon-bone continuity, organization, and maturity. These qualitative changes toward a more mature fibrocartilaginous interface led to a biomechanically superior attachment in HhAg-treated repairs. Improvements were noted at both 28 days and 56 days for a subset of structural and material properties. Of note, there was a decrease in modulus with time. This is not necessarily a negative outcome, as there are typically tradeoffs between strength, modulus, and toughness when optimizing materials, and modifications that increase modulus may come at the expense of toughness. In the current study, improvements in strength and work to failure may have come at the cost of modulus. The changes indicate both improved structural integrity of the repair to prevent rupture as well as improved quality of the material, consistent with the histological evidence of increased fibrocartilage formation. Because retears typically occur within the first several months after surgery, it is critical for treatments to exert their effect within the early postsurgical period, particularly at the healing interface.¹⁹ The early time frame of biomechanical improvements is notable for potentially reducing surgical repair failures due to insufficient reattachment of tendon to bone.

A marginal osteogenic effect of HhAg treatment was evident in the bone morphometric analysis. Previous studies have demonstrated the importance of bone quality on the strength of the tendon-to-bone attachment. Particularly, a correlation has been shown between mineral content and enthesis mechanical properties.^7,11 Furthermore, a decrease in mineralized tissue at the enthesis has been shown to correlate with increased retear rates.⁹ As such, we examined the effect of HhAg treatment on attachment site mineralization. Although most bone morphometric parameters were unchanged because of treatment, we noted small increases in trabecular bone number and thickness due to HhAg. Previous studies have demonstrated that HhAg treatment can enhance bone formation and improve fracture healing.¹⁸ In the current study, the lack of dramatic bone-specific effects from HhAg treatment is likely because of the local delivery approach, which focused the application of the agonist on the tendon-to-bone interface and not on the underlying bone. Consistent with a local effect, gene expression and histological analyses support the conclusion that mineralization was enhanced at the interface, despite only marginal mineralization in the underlying trabecular bone. Specifically, expression of COLX, a marker of chondrocyte hypertrophy and cartilage mineralization, was increased in HhAg-treated interfaces, as was expression of RUNX2, a marker of osteogenesis. Furthermore, histological evidence of enhanced enthesis fibrocartilage formation and organization was evident in HhAg-treated interfaces by 28 days compared with control interfaces.

There remain hurdles for translation of the current results to patients. Despite the therapeutic potential of Hh agonism, no Hh agonists have been approved by the Food and Drug Administraiton (FDA). This is in contrast to Hh inhibitors, of which several have been approved by the FDA as cancer therapeutics.⁵ Although the application of Hh agonists is not as developed as that of Hh inhibitors, numerous Hh agonists have been studied in animal and clinical trials. Several synthetic and endogenous oxysterols, as well as several synthetic glucocorticoids, are capable of Hh signal upregulation via action on SMO (a necessary cell surface receptor for Hh activation).^20,21 Furthermore, several small molecule agonists of Smo have been developed.^2,3 In addition to the therapeutic potential of Hh agonism in the context of tendon enthesis pathology and tendon-to-bone repair, preclinical evidence supports the therapeutic potential of Hh agonism in other systems, including skeletal muscle and bone.^{1,15,16,18,22,24,35} The opportunity for the development of HhAg therapeutics for tendon-to-bone repair is therefore high, particularly if delivered locally to avoid any off-target effects.⁸

The current study had several limitations. Most rotator cuff tears in patients occur after chronic tendon degeneration in an elderly population; the current study used an acute model of rotator cuff injury and repair in relatively young adult animals. The responsiveness to HhAg may be decreased in elderly patients with reduced numbers of responding progenitor cells. Nonetheless, the current study demonstrates exciting validation of the approach for the younger population with acute tears, and further studies in chronically degenerated rabbit models, which better re-create the chronicity of tears seen and the surgical repair methods used in the typical patient, will examine how applicable the approach will be to the older population. It is also worth noting that the present analysis did not examine the results of HhAg treatment beyond a 56-day time point. This timeframe is equivalent to the critical early period after surgery in patients, when the repair is most likely to fail. Future studies will examine longer term effects of HhAg treatment on tendon-to-bone healing. The current work and its potential to predict outcomes in humans are inherently limited by the absence of an outcome analogous to functional outcomes in patients. Additionally, the chosen delivery method, although arguably safer and more translatable than systemic delivery, still carries a risk of off-target effects, which were not explicitly assessed, and is subject to the imperfections that accompany microsphere fabrication and delivery. Going forward, FDA approval of the approach will require testing in large animals followed by clinical trials in patients. These future studies will define the specifics of drug delivery kinetics and dosage and the application in an arthroscopic setting, as well as the safety profile of HhAg treatment.

CONCLUSION

HhAg treatment promoted the formation of mineralized fibrocartilage at the interface between the supraspinatus tendon and its humeral head attachment. This ECM deposition resulted in improved biomechanical properties at 28 and 56 days after repair. This preclinical study demonstrated the therapeutic potential of HhAg treatment to improve tendon-to-bone healing after rotator cuff injury and repair.