A preliminary evaluation of in vivo response to a filament-wound macroporous collagen midurethral sling in an ovine model

Abstract

Stress urinary incontinence (SUI) impacts ~1/3 of women over age 50. Negative publicity around PP meshes used in pelvic prolapse repair drives the need for identifying alternative biomaterials for SUI repair. Our study evaluated in vivo response to collagen sling implanted in an ovine model. Electrocompacted collagen threads were filament wound as slings and crosslinked in genipin. Collagen slings were implanted suburethrally mimicking the transvaginal tape technique. Main study groups were: Collagen sling (n = 3, 6 months) and PP sling (n = 3, 6 months). Collagen sling was also tested at 3-weeks (n = 1) to observe early-stage tissue response and 1-year (n = 2) to assess biomaterial longevity in a preliminary capacity. Collagen slings healed to a fibrous ligament texture at 6 months and maintained such texture to 1 year. Histological scoring indicated biocompatible responses to collagen slings with no adverse events. All study groups exhibited complete tissue ingrowth and interstitial de novo collagen deposition at all time points. Collagen threads induced orderly de novo collagen deposition that was aligned along long axes of threads. Tissue infiltrated collagen slings that were explanted at 6 and 12 months presented similar structural strength with native tissues such as vagina and fascia, and PP (Lynx) slings (p > .05). With the limitation of low number of animals per time point in hindsight, this preliminary study justifies evaluation of collagen slings in a larger sample size of animals, particularly to assess persistence of ligamentous tissue response over longer durations than 1-year.

1 |. INTRODUCTION

Stress urinary incontinence (SUI) is estimated to affect 1 in 3 women over age 50.^1,2 SUI occurs predominantly due to parity related pelvic trauma to the connective tissue.³ For surgical management, synthetic midurethral slings (MUS) are implanted in ~400,000 patients/year in the United States.^4,5 Polypropylene (PP) mesh became the standard biomaterial for slings because of manufacturability, macroporosity, and ease of handling.⁶ PP meshes were also used for pelvic organ prolapse (POP) repair; however, a high erosion rate associated with PP in POP repair (~10%)⁷ prompted regulatory agency warnings, resulting in retraction of vaginal POP kits from the market. As a result, POP meshes were reclassified as high risk.^8,9 Past and ongoing legal actions for sling and POP mesh manufacturers^10,11 are impacting patient acceptance of PP slings negatively. Failure rates for MUS were reported to be 1%–19% within 2 years with no strong evidence for which intervention is best to utilize after failure.¹² PP slings are relatively contraindicated in patients undergoing revision of failed slings and during concurrent urethral surgery.¹³ As a result, the search for alternative biomaterials for sling repair is ongoing.

Repair of SUI by abdominal fascial autografts maybe associated with donor-site morbidity, and harvest of the autograft may prolong the surgery time.^14,15 Xenograft slings that were employed (e.g., submucosal intestine, dermis) in the past met limited success because of acute rejection,^16,17 and limited tissue integration due to poor interconnected macroporosity.^18,19 Furthermore, xenograft slings remodeled to inferior quality tissue, resulting in recurring incontinence.^20–22 These developments underline the need for developing innovative biomaterials for SUI treatment.²³

The basic building block of connective tissues, collagen, is biocompatible and available in medical grade formulations. Reconstituted collagen is generally available as weak gels or sponges which is not suitable for SUI repair. We overcame this limitation by the electrochemical compaction and alignment of collagen (ELAC) to manufacture strong monofilament threads.²⁴ ELAC threads can be woven or knitted in macroporous fabric forms that can hold sutures and deployed in a fully load-bearing context in tendon repair applications.²⁵ Baseline strength of woven ELAC scaffolds is 50% as strong as the native rabbit infraspinatus tendon.²⁶ We showed that woven collagen scaffolds that were implanted subcutaneously in rats demonstrated better biocompatibility, tissue integration rate, and structural property retention than dermal xenografts.¹⁸ This encouraging outlook on electrocompacted collagen threads motivated us to conduct a preliminary study to evaluate collagen slings in the dynamic suburethral environment of a large animal model in vivo.

Aims of this study were: (1) to fabricate macroporous collagen slings sized comparably to clinically available slings, (2) assess long-term post-surgical structural strength and host response of implanted collagen slings suburethrally in an ovine model.

2 |. MATERIALS AND METHODS

2.1 |. Collagen thread production

Electrochemical compaction of collagen to produce aligned threads was developed by our group previously.^24,27–29 Briefly, dialyzed type I bovine collagen solution (#FS22008, Collagen Solutions, Glasgow, United Kingdom) was mixed with maltodextrin (collagen/maltodextrin, 1/2), stirred briefly, dispensed between two rotating electrodes, and subjected to an electric potential of 40 V DC between the electrodes to produce collagen threads. Continuous-length collagen threads were collected in isopropanol/water (4/1). Two threads were combined as a 2-ply yarn, dried, collected on a spool (Figure 1A), vacuum-sealed, and stored at 4°C. Supplemental material provides additional details on methods (Appendix S1).

FIGURE 1.

(A) Electrochemically compacted collagen threads on a spool (≈100 μm diameter). Two such threads were twisted as yarns. (B, C) Using a two axes controlled computer numeric controlled machine, yarns were wound as slings using filament winding in the pattern depicted by this schema (also see Figure S1). (D) Macrographs of filament wound slings demonstrated a diamond shape pore that was spaced about 1.5 mm apart at the equator (E) and made an angle of about 15° with the horizontal. (F) The full length of genipin-crosslinked collagen sling. The color turned dark blue after crosslinked with genipin solution for 3 days. (G) Collagen segment was grafted with polypropylene wings (refer to Figure S2 for fabrication of PP-wings). Wings were overlapped with a collagen segment by 5 mm at both ends and adhered to by polylactic acid dissolved in chloroform. The color of collagen sling turned yellow brownish after sterilizing in peracetic acid/ethanol solution. Polypropylene wings were used as a leader to place the collagen sling surgically, following which most wings were snipped at the abdomen. Otherwise, PP-wings were distantly positioned to the region of interest, that is, the mid-urethra.

2.2 |. Filament wound collagen sling

A computer numeric controlled (CNC) device (Sherline, CA) was programmed to wind collagen yarn around a mandrel to produce slings with diamond-shaped macroporosity (1.5 × 4 mm, Figure 1B–E). The dry yarn was passed through 1% maltodextrin solution to provide adherence. Slings remained at room temperature for an hour to dry. Samples were recovered from the mandrel by a longitudinal cut to obtain 11.5 cm long collagen slings (Figure 1F). PP wings were grafted serially onto both ends of collagen slings (Figure 1G). PP wings were overlapped with collagen segments over a length of 0.5 cm and adhered using poly-l-lactic acid (Goodfellow Corporation, Coraopolis, PA) dissolved in chloroform. These wings were positioned distantly to the midurethral collagen segment and simply acted as leaders to pull implants surgically toward the anterior abdomen. In this process, the collagen segment experienced the pull forces. Once positioned, excess PP length was cut and removed following implantation.

2.3 |. Crosslinking and Sterilization of Slings

Collagen slings were crosslinked with genipin (0.625%, 72 hours). Slings were sterilized using peracetic acid/ethanol solution (1/2.5, 4 h), rinsed in copious amounts of sterile deionized water (3 × 3 min, Figure 1G), dried, and placed separately in self-sealing sterile pouches. Commercially available polypropylene slings (n = 3, Lynx^™, Boston Scientific, Marlborough, MA) were tested as clinical standards in three sheep at 6 months.

2.4 |. Surgery and euthanasia

All animal experiments were conducted in accordance with rules and protocols approved by Institutional Animal Care and Use Committee at Case Western Reserve University (Protocol # 2017–0023) and Northeast Ohio Medical University (Protocol # 18–01-191). Primary study groups at 6 months included six ewes that were implanted either with collagen sling (N = 3) or PP sling (N = 3, Lynx TM, Boston Scientific Corporation, MA).

One ewe was included to assess early cell infiltration for collagen sling at 3 weeks. Two additional ewes implanted with collagen sling were sacrificed at 12 months to assess long-term structural robustness of remodeled collagen slings.

The surgical procedure was similar to human tension-free vaginal tape surgery as detailed in the Supplemental material S1. Following euthanasia, urogenital tissues (vagina, urethra, and sling in-between) were harvested in one piece (Figure S3). Rectus fascia and vaginal tissues were dissected as native tissue controls for mechanical testing. Vaginal tissue was collected about 5 cm distal to the region where the sling was implanted.

2.5 |. Histology and Immunohistochemistry

Tissues were stained with hematoxylin and eosin, Masson’s trichrome, and picrosirius red to visualize cells, collagen deposition, and collagen orientation, respectively. Slides were examined to assess material biocompatibility in accordance with ISO 10993; Biological evaluation of medical devices–Part 6: Tests for local effects after implantation³⁰ using a semi-quantitative scale rating from 0 to 4 (Scores represent 0: absent, 1: minimal, 2: mild, 3: moderate, 4: extensive). Acute (presence of macrophages and neutrophils) and chronic (presence of plasma cells and lymphocytes) inflammatory reactions were scored. Host reaction at implant surfaces was scored for foreign body reaction indicated by the presence of monocytes, macrophages and foreign body giant cells, and fibrous encapsulation indicated by the existence of acellular collagen around the perimeter of sling materials. Tissue integration was scored for cellular infiltration indicated by the density of nuclei within the bulk of the sling and new collagen deposition indicated by cellularized area with quasi aligned collagen fibers corresponding to the direction of the sling.

Masson’s trichrome-stained slides were assessed with stereological point counting^31,32 to determine area fractions of collagen threads and interstitial host collagen deposition. A grid (50 μm × 50 μm) was overlaid onto micrographs and points which fall within regions of interstitial collagen or collagen sling threads were counted for 6- and 12-month specimens. Measurements of fibrous capsule thickness surrounding the implanted material were recorded.

Immunohistochemistry (IHC) was carried out for collagen I (rabbit polyclonal to collagen I, #ab34710; Abcam, dilution 1:100) and collagen III (rabbit polyclonal to collagen III, #ab7778, Abcam, dilution 1:50) at 6- and 12-month timepoints on collagen sling slides. A conjugated secondary antibody (Goat Anti-Rabbit IgG horseradish peroxidase, #ab205718; Abcam) was utilized.

2.6 |. Mechanical testing

Uniaxial tension tests were performed on: (Group 1) blank collagen slings (referred to as control hereon) to measure baseline properties, (Group 2) blank PP slings at baseline, (Group 3) tissue-ingrown collagen sling complex retrieved at 6 months, (Group 4) tissue-ingrown PP sling complex explanted at 6 months, and (Group 5) tissue-ingrown collagen sling complex explanted at 1 year. Rectus fascia and vaginal tissues were tested as native tissue controls.

Mechanical testing was conducted using an MTS 858 Bionic Test System (MTS, Minnesota). Hydrated test samples were gripped via liquid-nitrogen cooled clamps, with gage length values ranging from 20 to 30 mm. Following position-controlled load preconditioning, samples were loaded to failure under monotonic tension (20 mm/min) to determine maximum load to failure. Width and thickness of retrieved samples along the gage length were recorded to calculate apparent stress, apparent strain and apparent modulus values^33–35 (Supplemental material S1). For tissue ingrown specimens, these values reflect an averaged value for the complex, rather than the individual implant or the ingrown tissue itself. Comparisons of groups 3 and 4 to groups 1 and 2, respectively, provided information on changes in the apparent values between baseline and 6 months. Comparisons of group 5 to 3 determined whether apparent properties of slings were maintained between 6 months and 1 year.

2.7 |. Statistics

At 6-months, biomechanical parameters, point count area fraction values, and fibrous capsule thickness measurements were compared between any two groups by using a nonparametric Mann–Whitney U tests with statistical significance defined as p < .05. Data for 3 weeks and 12 months were presented qualitatively.

3 |. RESULTS

3.1 |. Gross-morphological observation of retrieved constructs

All animals survived to target time-points with no complications. No signs of sling exposure or urinary obstruction were observed for any of the groups. At 6 and 12 months, explanted sling + tissue complex was observed to bridge the urethrovaginal complex bilaterally toward the back of pubic bones (Figure S3). None of the slings showed any evidence of migration (Figure S3A,B). Harvested collagen sling +tissue complex had smooth surfaces and were palpated as tough flexible bands, similar to native ligaments (Figure S3C). Explanted PP sling + tissue complex was thin, firm, and the edges felt sharp (Figure S3D). Ligamentous tissue thickness around implanted collagen slings 6 months (3.40 ± 1.19 mm) appeared to be thicker than that around implanted PP slings (1.20 ± 0.22 mm); however, the difference was not marginally significant (p = .081).

Partially degraded collagen threads were visible through portions of ligamentous tissue within and around implants at 12 months. Collagen threads appeared to resorb gradually from lateral to medial direction, where areas adjacent to the urethra/vaginal junction appeared to remodel with no obvious sling residuals (Figure S3E). Gross morphological pictures of explanted collagen sling + tissue complex for each sheep is shown in Figure S4.

3.2 |. Semiquantitative histology

At 3 weeks, the collagen sling + tissue complex showed evidence of early wound healing response with granulation tissue and fibrous capsule development. Host reaction was restricted to implanted collagen regions without extension to periurethral or paravaginal tissues. At all timepoints, new collagen deposition and focal foreign body reaction remained limited to implant regions (Figure 2). Masson’s trichrome stained sections indicated complete cellular infiltration (red nuclei) and collagen deposition (blue) in interstitial regions of collagen slings at all time points (Figure 2A).

FIGURE 2.

The host reaction to collagen sling over time. Typical fields of views taken from H&E (hematoxylin and eosin) and MT (Masson’s trichrome) stained sections demonstrating de novo tissue deposition in the interstitial regions of retrieved sling materials. Sections are taken in the frontal plane and fields of view are taken in the mid-urethral regions. (A) Longer axes of slings are oriented horizontally. C: Collagen sling residuals. The scale bar for H&E images is 1 mm. The scale bar for MT images is 200 μm. (B) Results for percentage of total area occupied by collagen sling threads using the point counting method in the suburethral area. (C) Percentage of total area occupied by interstitial collagen in the suburethral area was significantly greater at 12 months. (D) Thickness of the fibrous capsule around collagen slings declined significantly by 12 months.

Area fraction of collagen threads degraded significantly between 6 and 12 months (p < .001, Figure 2B) along with increasing area fraction of host-synthesized interstitial collagen (p < .05, Figure 2C), implying collagen slings gradually remodeled into connective tissue. Fibrous capsule thickness measurements showed a significant reduction at 12 months (p < 0.05, Figure 2D).

Picrosirius red staining revealed highly aligned de novo collagen around collagen threads as evidenced by strong birefringence (Figure 3). Degree of collagen alignment around polypropylene + tissue complex was qualitatively less pronounced at 6 months as reflected by discontinuous grainy birefringence relative to collagen sling + tissue complex where brightness of birefringence was pronounced and continuous.

FIGURE 3.

Picrosirius red-stained section under cross-polarized imaging conditions to visualize the alignment of collagen deposition in the interstitial regions of retrieved sling materials. C: Collagen thread, L: Lynx, P: Polypropylene thread in the wing section. Sections are taken in the frontal plane and fields of view are taken in the mid-urethral regions of slings except for 1 year and 3-weeks PP images that were taken from the proximal end-regions of slings. The longer axis of slings is indicated with a double-sided yellow arrow.

3.3 |. Histological scores for biocompatibility and healing response

Histological scores are reported for 6- and 12-month collagen and 6-month PP sling groups. Both materials were biocompatible and acute inflammation was absent in all groups. Chronic inflammation was absent in all treatment groups at all time points except for a few focal occurrences. Both PP and collagen slings demonstrated a high degree of cellular infiltration (Figure 4).

FIGURE 4.

Histological scoring for host response and tissue integration for study groups at 6 months (PP and Collagen) and 12 months (Collagen). Scores represent 0: absent, 1: minimal, 2: mild, 3: moderate, 4: extensive

Collagen deposition scored moderate to extensive for both sling types at 6 months. By 12 months, collagen deposition in collagen slings scored consistently as extensive. At 6 months, foreign body reaction for collagen sling was minimal whereas it was mild for PP sling. Foreign body reaction stayed minimal at 12 months for collagen slings. Median value for fibrous encapsulation around collagen slings declined from extensive to moderate between 6 and 12 months, respectively, in agreement with direct measurement of fibrous capsule thickness values (Figure 4).

3.4 |. Immunohistochemical characterization of collagen I and III

Qualitative assessment of immunostained sections of interstitial tissue within collagen sling suggested a more prominent presence of collagen I than collagen III, whereas tissue within interstitial domain of PP slings appeared to have greater presence of collagen III than collagen I (Figure 5). At 12 months, collagen I was still present around the collagen sling fibers and collagen type III was subjectively diminished in comparison to collagen sling specimens at 6 months.

FIGURE 5.

Immunohistochemistry results of staining of type III and type I collagen. “*” indicates collagen and polypropylene threads of the corresponding slings. PP, polypropylene Lynx sling. Scale bar is 100 μm.

3.5 |. Mechanical testing

Maximum load of explanted collagen sling + tissue complex showed an apparent increase from baseline to 6 months and baseline to 12 months, but no statistically significant differences were noted (p = .052 and .105, respectively, Figure 6). Maximum load values for explanted collagen sling + tissue complex versus explanted PP slings + tissue complex was not significantly different at 6 months (p = .190). At 6 months and 12 months, maximum load for all groups was comparable to that of the rectus fascia (p > .05).

FIGURE 6.

Structural strength, apparent modulus and apparent strain values for baseline slings (collagen, or PP), explanted collagen sling + tissue complex, and explanted PP sling + tissue complex at different time points. Maximum load indicates the ultimate load bearing capacity of test samples whereas “strain at maximum stress” reflects the extensibility of specimens. An apparent strain value of 1 pertains to 100% extension from the original length of the sample. Material properties for abdominal fascial strips and vagina are also provided as reference points of the apparent modulus and apparent strain at maximum stress for native tissues. For mechanical testing (n = 3), averages and standard deviation are presented for 6-month samples as well as the individual data points.

Collagen sling apparent modulus at baseline was comparable to that of rectus fascial (p = .86) and vaginal samples (p = .105) (Figure 6). Apparent modulus obtained from the explanted PP sling + tissue complex was comparable to that of vaginal sample at baseline (p = .773) and at 6 months (p = .149). Pairwise comparisons between apparent moduli of explanted collagen sling +tissue complex, vagina (p = .149), or fascia (p = 1.000) were not significantly different at 6 months.

The apparent strain at maximum stress values for explanted collagen sling + tissue complex was comparable to that of rectus fascia tissue at 6 and 12 months (p > .1 for both comparisons).

4 |. DISCUSSION

In this study, functionally implantable and fully load-bearing pure collagen slings were biofabricated with dimensions matching those that are used clinically. Slings were implanted successfully in the anatomically relevant mid-urethral domain in an ovine model. Type and extent of de novo tissue deposition and biocompatibility were assessed at 6 and 12 months in vivo. Tensile properties of tissue-integrated collagen slings were characterized and compared to clinical standard polypropylene slings with integrated tissue at 6 months. Results indicated rapid cellular infiltration between collagen threads with induction of de novo host-collagen deposition oriented to follow the collagen thread pattern, implying collagen threads imparted a tissue guiding effect. Importantly, synthesis of new ligament-like tissue compensated for degradation of the original implant, and tensile properties were maintained over the study duration, despite notable degradation of the original collagen sling’s threads. More prominent presence of collagen type I in comparison to collagen type III in the interstitial matrix suggests formation of a ligament-like structure as opposed to a scar-like tissue.

The original goal of mid-urethral slings was to act as an artificial ligament to support the damaged pubourethral ligament.³⁶ Accordingly, failure of such support results in recurrence of incontinence, as observed for decellularized xenografts.²¹ A major disadvantage of commercial xenografts is poor interconnected porosity which prevents cellular and vascular infiltration.¹⁸ Success of PP in treating SUI is attributed, in part, to a macroporous structure accommodating rapid tissue integration. Furthermore, fibrous tissue encapsulation and foreign body reaction elicits collagen-rich scar formation around PP fibers (Figure S5), ensuring continuity of the critical reinforcement function. Collagen slings demonstrated several of the favorable properties of PP in vivo: rapid cellular integration, and ligament-like tissue induction descending from pubic bone to urethrovaginal complex. Importantly, collagen slings maintained structural integrity over 12 months in vivo, suggesting that collagen slings likely have a greater longevity than xenografts that lost their functional integrity over time.

The thickness of the ligament-like tissue that formed around collagen slings was notably greater than that observed around PP, indicating more exuberant de novo tissue induction. Also, interstitial host collagen followed collagen sling thread orientation along longer axes of the sling, the direction along which functional support is needed. The woven structure of PP sling is oriented isotropically which likely contributed to some disorder in the collagen deposition and inconsistency of the birefringence visualized under polarized imaging (Figure 3). As an alternative to collagen thread induced alignment of host collage, it may be postulated that the tension that is imparted on the sling may have defined the alignment of de novo collagen. In the current study, slings were placed under the urethra mimicking the tension-free vaginal sling surgery. In the past when aligned collagen threads were implanted subcutaneously, in the absence of physiological tension, we observed de novo collagen alignment along the collagen threads.¹⁸ Thus, we believe that physiological tension is less likely to be the main contributor to de novo collagen alignment.

Host reaction to biomaterials starts with acute inflammation which may eventually advance to either rejection or resolution.¹⁸ Collagen is a long-standing material with an excellent record of biocompatibility for many clinical applications. Histological scoring performed in this study confirmed biocompatibility and safety of genipin cross-linked collagen slings in sheep, consistent with our previous study using dimensionally scaled-down woven collagen scaffolds implanted subcutaneously in rat model.²¹

It is noteworthy to mention that host response to collagen slings attained a steady-state healing process that resolved toward a ligament-like tissue formation by 6 months and was persistent at 12 months. Such resolution is evidenced by a number of findings as follows. The fibrous capsule became denser between 6 and 12 months, indicating fibrous capsule maturation. Second, inflammation at the collagen thread interface was largely absent at 6 and 12 months as evidenced by scarce macrophage presence, and absence of polymorphonuclear lymphocytes. Third, ligament-like end-stage healing was supported by increased collagen type I presence over time.

Vaginal exposure is a clinically reported drawback of PP-based mesh biomaterials.³⁷ In our study, host reaction was microscopically limited to implant regions. One hypothesized cause for vaginal exposure is stress shielding,^38,39 as an implant with a substantially greater apparent modulus than native tissue absorbs functional loads, depriving native tissue of mechanical stimulation, resulting in atrophy. Our data indicate apparent moduli of tissue integrated collagen and PP slings were comparable to reference native soft tissues, rectus fascia and vagina. Therefore, compatibility between the moduli of native tissue and sling-tissue complex may potentially avert atrophy that is driven by stress-shielding.

An important limitation for this study was the low number of animals per time point; therefore, all the conclusions should be interpreted reservedly. Another limitation was that incontinence was not induced in the sheep to test continence restoration because our focus was to assess biocompatibility, tissue integration, and long-term mechanical robustness at the current stage. Future research is needed to test functionality against incontinence. We strived to collect the longest possible segment for mechanical testing of retrieved implants. Being a biological sample, attainment of ideal test conditions is not always possible. Therefore, gage length of mechanical test samples was only 2.5-fold greater than the width. In this regard, the tensile tests deviate from idealized tensile test of monolithic materials. Similarly, the cross-sectional area of specimens varied along the length of the specimen. Therefore, we had to use post-failure cross sectional area which is likely lower than the original area. As a result, the modulus values we obtained may be overestimated by an unknown amount.

Despite such limitations, results of this preliminary study are consistent in that a ligament-like tissue results from collagen slings whose structural properties persist over time which could suggest a promising for SUI repair and possibly for POP repair in the future.

5 |. CONCLUSIONS

Biocompatibility, structural robustness, and physical conformation to the complex, dynamic suburethral environment in a human-sized animal model highlights collagen slings as a potentially useful biomaterial for pelvic floor and reconstructive surgeries.

DATA AVAILABILITY STATEMENT

The data supporting this study’s findings are available from the corresponding author upon reasonable request.