Engineered Decellularized Tendon Matrix Putty Preserves Native Tendon Bioactivity to Promote Cell Proliferation and Enthesis Repair

Abstract

Rotator cuff tears are a common soft tissue injury that can significantly decrease function of the shoulder and cause severe pain. Despite progress in surgical technique, rotator cuff repairs (RCRs) do not always heal efficiently. Many failures occur at the bone-tendon interface as a result of poor healing capacity of the tendon and failure to regenerate the native histological anatomy of the enthesis. While allografts are commercially available, clinical use is limited as they do not stimulate tissue regeneration and are associated with a structural failure of up to 40% in re-tear cases. Novel tissue engineering strategies are being developed with promise, but most involve addition of cells and/or growth factors which extends the timeline for clinical translation. Thus, there exists a significant unmet clinical need for easily translatable surgical augmentation approaches that can improve healing in RCR. Here we describe the development of a decellularized tendon matrix (DTM) putty that preserves native tendon bioactivity using a novel processing technique. In vitro, DTM promoted proliferation of tenocytes and adipose-derived stem cells with an increase in expression-specific transcription factors seen during enthesis development, Scleraxis and Sox9. When placed in a rabbit model of a chronic rotator cuff tear, DTM improved histological tissue repair by promoting calcification at the bone-tendon interface more similar to the normal fibrocartilaginous enthesis. Taken together, these data indicate that the engineered DTM putty retains a pro-regenerative bioactivity that presents a promising translational strategy for improving healing at the enthesis.

Figure 1

(a) DNA content was measured following decellularization using the using the DTM processing technique or using standard detergent methods, like 1% SDS and 0.1% EDTA. All decellularization methods effectively removed DNA content as compared to treatment with PBS (p < 0.0001; n = 2 donors) and no significant differences were found between decellularization treatments. Histological sections of (b) native tendon or (c) DTM decellularization stained with DAPI to identify cell nuclei, scale bar = 100 µm. An ordinary one-way ANOVA was conducted to determine statistical significance in DNA content between the treatment groups, F (5, 6) = 515.1, p < 0.0001, and all further significance gained from Tukey's multiple comparison testing is listed on graph (a).

Figure 2

Primary tenocytes and ADSCs were plated at 20,000 cells/well and proliferation was quantified at 48 hours and 7 days (a, b) after plating, generating significantly different proliferation rates (n = 3–5 donors tested/group). DTM showed significantly more cell proliferation at day 7 than the pepsin groups in both tenocytes (p = 0.0001) and ADSCs (p < 0.0001). (c–j) Collagen coating, tissue-culture (TC) treated, DTM, and pepsin-coated plates were seeded with tenocytes and still images from live cell imaging were taken at 0 (c–f) and 48 hours (g–j) where variance in tenocyte cell morphology can be viewed. A two-way ANOVA was conducted revealing a statistically significant interaction between days and treatment groups on cell number for ADSCs, F (4, 18) = 31.77, p < 0.0001, and for tenocytes, F (4, 16) = 5.150, p = 0.0073. Statistically significant results from Tukey's multiple comparison test are reported on graphs (a) and (b). Scale bars = 100 μm.

Figure 3

(a) Total protein content between native Achilles and patella tendons was analyzed and no significant difference was found between the tendon types (p = 0.7413; n = 10 donors). (b) Schematic diagram of Achilles and patella tendons dissected into equal thirds comprised of proximal (P), midcenter (M), and distal (D) regions. (c, d) The patellar and Achilles tendons had no significance in total protein content between proximal, mid, and distal regions (p = 0.9893 and p = 0.368, respectively; n = 6 donors). Ordinary one-way ANOVA was conducted to determine statistical significance in total protein content and treatment groups, F (2, 33) = 0.934, p = 0.403, in patella tendon protein content and treatment groups, F (2, 15) = 0.011, p = 0.989, and in Achilles tendon protein content and treatment groups, F (2, 15) = 1.069, p = 0.368.

Figure 4

(a) DTM reconstituted and stretched to demonstrate the viscoelasticity of the DTM. (b) DTM reconstituted. (c) Oscillation stress sweep of human DTM with concentrations of 0, 1, 3, 5, and 7 g of lyophilized DTM to 1 mL of PBS resulted in all reconstitution concentrations tested to maintain an elastic modulus (n = 1 donor, 2-3 runs each). (d) Collagenase activity was measured to confirm successful removal of the added collagenase solution during enzymatic digestion. No significant differences were observed in collagenase activity between native tendon and enzymatically digested and filtered samples (p = 0.922; n = 3–10 donors). (e) The total protein of the samples was tested to confirm bioactivity was maintained (n = 4 donors) and no significant differences were found between native, DTM, and pepsin samples. An ordinary one-way ANOVA was conducted to determine statistical significance in collagenase activity and treatment groups, F (4, 22) = 18.06, p < 0.0001, and in total protein content between groups, F (2, 9) = 5.282, p = 0.0304. Additional significance from Tukey's multiple comparison is listed on graphs (d) and (e).

Figure 5

RNA isolated from ADSCs cultured on tissue-culture (TC) treated, DTM, and pepsin-coated plates was measured for level of tenocyte differentiation by probing for tenocyte markers (n = 3–5 donors) (a) Tenomodulin (Tnmd) and (b) Scleraxis (Scx), (c) tenocyte recruiter Sox9, and (d) TGF-β marker Smad3. DTM had significantly more Sox9 (p = 0.023) and Smad3 (p = 0.0395) than the control at day 7. Performing a two-way ANOVA revealed a statistically significant interaction between days and treatment groups on Sox9 values, F (2, 16) = 6.365, p = 0.0093, and on Smad3 values, F (2, 17) = 3.907, p = 0.0402. Statistically significant results from Tukey's multiple comparison test are reported for Sox9 and for Smad3 on each respective graph above. Two-way ANOVA for Scx, F (2, 23) = 0.4200, p = 0.6620, and Tnmd, F (2, 19) = 0.9607, p = 0.4005, was found to have no statistical significance.

Figure 6

The TGFβ profile was then measured to determine bioactivity of the processed tendons compared to the native tendons (n = 10 donors). No statistical significance was found in (a) TGFβ1 and in (b) TGFβ2 between combined native tendons, DTM, and pepsin. However, statistical significance was found in (c) TGFβ3 between the combined native tendons and DTM (p < 0.0001) and combined native tendon and pepsin (p < 0.0001), with no statistical difference between DTM and pepsin. One-way ANOVA was conducted to determine significance between treatment groups on TGF-β1, F (2, 24) = 4.054, p = 0.6712, on TGF-β2, F (2, 24) = 0.8386, p = 0.4446, and on TGF-β3, F (2, 21) = 35.55, p < 0.0001. Statistically significant results from Tukey's multiple comparison test are reported on graph (c).

Figure 7

Rotator cuff repair (RCR) surgery was performed in a rabbit model with a chronic rotator cuff tear (n = 4/group). (a) Initial rotator cuff tear of the supraspinatus; (b) DTM placed under the supraspinatus 6 weeks after the initial tear; (c) repair of the initial tear, suture-anchoring the supraspinatus to the greater tuberosity. 8 weeks following the repair, shoulders were harvested, processed through histological analysis, and stained with (d–f) HBQ, (g–i) H&E, and (j–l) Picrosirius Red. Images display the bone, fibrocartilage, and tendon tissues at the tendon enthesis in (d, g, j) contralateral shoulders, (e, h, k) repair only, and (f, i, l) DTM treatment with repair. T = tidemark; FC = fibrocartilaginous zone; B = bone; arrow in (h) shows collagen fiber orientation. Scale bars represent 100 μm.