Immediate trauma to the vagina and perineum are serious and common sequelae of vaginal childbirth, affecting roughly 75% of women and causing significant short- and long-term morbidity. Indeed, maternal birth injury is by far the greatest risk factor for the development of pelvic floor dysfunction later in life. The cost to society of pelvic floor dysfunctions is not trivial, estimated at >$1 billion per year with approximately 338,000 patients requiring surgery in the United States annually for prolapse alone.

Despite the morbidity, damage to the vagina at the time of childbirth has received little attention. A major long-term motivation for understanding the tissue behavior following maternal birth injury is improvement of the healing response of the vagina, and ultimately tissue quality, when injury occurs. Enhancing recovery to the preinjured condition may in turn decrease susceptibility to pelvic floor disorders later in life. Current models of tissue behavior after maternal birth injury, which include the data acquired from magnetic resonance imaging before and after delivery, would be improved by the addition of biomechanical and biochemical (histomorphological) data characterizing the affected tissue. Such data are also necessary to evaluate the response to potential treatments that can be applied at the time of injury.

To explore possible treatment modalities for maternal birth injury, we turned to functional tissue engineering, which aims to enhance healing to improve the functional behavior of the injured tissue. Recently, extracellular matrix scaffolds have shown significant potential in a wide variety of applications including healing of the medial collateral ligament (MCL) and Achilles tendon after injuries. Small intestinal submucosa (SIS) is a naturally occurring acellular collagen matrix derived from porcine intestine that has also been used by gynecologists to augment prolapse repairs and as a tissue filler or replacement. In contrast to most graft materials, SIS is a bioinductive graft containing growth factors and cytokines that promote healing of damaged tissue. Some of the bioactive molecules that are contained within SIS include vascular endothelial cell growth factors, basic fibroblast growth factors, and transforming growth factor-β, all of which induce angiogenesis and cell infiltration into the wound site. In addition, the application of SIS has been shown to recruit bone marrow–derived cells (possibly of a pluripotent nature) that remain at the injury site and aid in the healing response helping the injured tissues recover. In animal models, it has been shown that immune-mediated inflammatory reactions evoked by SIS are limited due to the chemical composition of the scaffold. Lastly the scaffold is completely degraded and replaced by native tissue over a period of 8-12 weeks. These factors are all thought to be the reason why scar formation is minimized with the use of this scaffold and the functional behavior of healing tissues is improved. Consequently, SIS can be utilized as a bioactive agent to augment healing toward a more regenerative pathway. SIS has been successfully used in this way to aid in the repair of tissues from the human vascular, urogenital (bladder and urethra), and musculoskeletal systems. When used to treat orthopedic injuries, SIS enhanced the mechanical properties of the healing tissue (a reflection of tissue quality) with an increase in tangent modulus and tensile strength.

The principal determinants of strength in the vagina and its supportive tissues are thought to be fibrillar collagens I, III, and V. Changes in ratios and architecture of these collagens likely contribute to altered vaginal tissue behavior. Collagens III and V are increased in tissues following trauma and are associated with inferior mechanical properties: increased distensibility and lower tensile strength, compared to tissues with high concentration of collagen type I.

We hypothesized that untreated vaginal tissue heals incompletely following maternal birth injury, which is exhibited as inferior mechanical properties due to elevated levels of collagen types III and V relative to type I collagen when compared to the uninjured vagina. The application of SIS after a maternal birth injury would restore the vagina to its original mechanical behavior and biochemical composition. In this study, we therefore sought to characterize changes in vaginal collagen ratios and tissue microarchitecture as well as changes in the mechanical properties of the vagina that occur following simulated birth injury in the presence and absence of tissue augmentation using SIS in a well-established rat model.

Materials and Methods

Animals

Approval for this study was received from the Institutional Animal Care and Use Committee (IACUC) from the University of Pittsburgh, PA. A total of 52 Long-Evans 3-month-old virgin rats were used in this study. Animals were divided into the following groups: control (n = 18), injured untreated (n = 18), and injured treated with SIS (n = 16). Injured rats underwent a simulated birth injury via balloon distension and recovered for 4 weeks. Sixteen animals underwent treatment with either a sheet (n = 8) or a gel suspension (n = 8) of SIS, placed over the site of vaginal injury at the time of the simulated birth injury. Control and injured virgin animals were in similar phases of the menstrual cycle (estrus and metestrus) as determined by vaginal smears. Variables such as total weight, genital hiatus (GH) (diameter of vaginal opening), and total vaginal length (TVL) were obtained. Measurements of the GH and TVL were recorded before and after injury. After sacrifice, the weights and diameters of the excised vaginas were measured.

Simulated birth injury

We utilized an established model of simulated birth trauma. A 16F Foley catheter was custom fit with a balloon. The tip of the catheter was trimmed so that it was flush with the end of the balloon. Experimental animals were anesthetized and the catheter-balloon construct was placed in the vagina. Following filling of the balloon, the animal was placed supine on the edge of a table and the Foley catheter was allowed to hang with 130-g weight attached to its free-hanging end. The catheter with inflated balloon stayed in place for 2 hours. The balloon was then deflated and the catheter removed. Gross examination of vaginas was performed to assess the extent of injury. Balloon distention with 5-mL volume produced full-thickness tears in all animals ( Figure 1 ).

Representative photograph taken immediately after simulated birth injury

Urethra is located at 12 o’clock. Full-thickness vaginal tear is visible, extending from 7-10 o’clock.

Alperin. Collagen scaffold. Am J Obstet Gynecol 2010.

Application of the porcine SIS

Following balloon injury, the vaginal full-thickness tear was visualized. A strip of sheet SIS (Cook / Biotech Inc, Bloomington, IN), 10 × 3 × 0.2 mm, or the equivalent amount of the gel form of SIS was placed over at the site of vaginal injury underneath the vaginal epithelium. For the sheet form, the luminal side faced the tear. The gel form of the SIS was made following previously described technique. The edges of vaginal epithelium were reapproximated over the collagen scaffold with 5-0 polyglactin suture secured to a microclamp (Ethicon Inc, Somerville, NJ), thereby covering the SIS. For the animals that went untreated after birth injury, analogous sutures were placed in the vagina to control for the effects of suture placement. All animals were allowed to recover for 4 weeks after which they were euthanized according to IACUC guidelines.

Histological analysis

Thirty rats were utilized for the histological and immunofluorescence portions of the study: control (n = 10), injured untreated (n = 10), and injured treated with SIS (n = 10). Following sacrifice, the vagina was dissected away from its attachments to the pelvic sidewall, pubic symphysis, levator ani muscles, and sacrum. The full-thickness midportions of the vaginas were excised, embedded in Optimal Cutting Temperature media (Sakura Finetek Inc, Torrance, CA), cut into 5- to 7-μm sections with a cryostat, and stained with Masson trichrome for examination of gross morphologic features. Histopathologic examination was carried out to assess the extent of tissue injury. Our histologic endpoints included vaginal thickness on a cross section (epithelium to inferior margin of the muscularis), presence of a cellular infiltrate (absent, present), and disruption of tissue architecture (yes/no). The transverse diameter of the cross section of the midportion of the vagina was measured across from one to another antilumenal side of the epithelium.

Immunofluorescence

Frozen embedded tissues were cut into serial sections of 5-7 μm and stored at 20°C until they were ready for use. The sections were incubated with either collagen I and III or collagen I and V primary antibodies at room temperature for 1 hour. The primary antibody to collagen I is rabbit anticollagen I (1:100) (Abcam, Cambridge, MA), collagen III is mouse anticollagen III (1:1000) (Sigma, St Louis, MO), and collagen V is mouse anticollagen V (1:1000) (Chemicon, Billerica, MA). Optimal dilutions of primary antibody were determined by a series of previous titration experiments. The secondary antibodies for collagens I, III, and V are goat antimouse Cy3 (Jackson, West Grove, PA), goat antirabbit Alexa 488 (Molecular Probes, Eugene, OR), and goat antimouse Cy5 (Jackson). Smooth muscle F-actin was labeled with Alexa 647 Phalloidin (1:500) (Molecular Probes). Sections were incubated with secondary antibody for 60 minutes, followed by 5 washes of bovine serum albumin and 5 washes of phosphate-buffered saline. Slides were stained for 30 seconds with Hoechst stain, followed by 5 washes of phosphate-buffered saline. Sections were then mounted with gelvatol and a coverslip and dried overnight at 4°C in the dark. Samples simultaneously labeled with 3 different primary antibodies (collagens I, III, and smooth muscle actin or collagens I, V, and smooth muscle actin) were scanned with an Olympus Fluoview BX61 (Olympus Optical Corp, Tokyo, Japan) confocal scanning laser microscope with a ×60 objective. One of the authors (M.A.), blinded to the identity of the slides, performed all the analyses. Each specimen was analyzed at 10 random sites of the subepithelial and muscularis layers of the vagina. Fluorescence microscopy was interfaced to a quantitative computer program (Matlab; MathWorks, Inc, Natick, MA). Results are represented as mean pixel intensity ratios per square area. The ratio of collagen I/III or I/V was used as an indicator of remodeling and the quality of the healing response. Regions of interest were drawn around vessels, which were identified by their morphology and positive staining of vascular smooth muscle F-actin, labeled with Alexa 647 Phalloidin (1:500) (Molecular Probes). Initial studies showed no significant difference in mean pixel intensity ratios of collagens I/III and I/V with and without regions of interest containing vascular structures. Therefore, vascular structures were not excluded in the final analysis.

Biomechanics

To characterize the biomechanical properties of vaginal tissue, a uniaxial tensile test in the longitudinal direction was performed. In all, 22 rats were utilized for this portion of the study: control (n = 8), injured untreated (n = 8), and injured treated with SIS (n = 6). SIS-treated animals utilized for biomechanical testing included 3 rats treated with the gel form of SIS and 3 treated with the sheet form. After sacrifice, the vaginas were carefully dissected from the surrounding connective tissue, wrapped in saline-soaked gauze, placed in a plastic bag, and immediately stored at –20°C. On the day of testing, the tissue was thawed immediately before performance of the uniaxial tensile test. Custom-designed soft-tissue clamps were utilized to grip each vaginal sample at the proximal and distal ends, thus forming the clamp-specimen-clamp complex. To provide a uniform stress and strain distribution within the midsubstance of the vaginal sample, for each uniaxial tensile test an aspect ratio (length/width) of the sample was insured to be a minimum of 5. The cross-sectional area of each sample was measured using a laser micrometer system in 3 areas along the length of the sample to calculate an average cross-sectional area for the tissue, as previously described. Throughout the experimental protocol, the sample was kept moist using 0.9% saline. Black contrast markers were placed on the tissue near the midline of the sample for strain measurements. These markers were captured and tracked using a camera system (Keyence CV-2600; Keyence Corp, Osaka, Japan) and motion analysis software (Spicatek Inc, Maui, HI) to calculate the strain in the midsubstance of the tissue. The clamp-specimen-clamp complex was placed into a 37°C saline bath and attached to a uniaxial tensile testing machine (Instron 5565; Instron, Grove City, PA). The proximal vagina was attached to a 50-lb load cell (Honeywell model 31, resolution 0.1 N; Sensotec, Columbus, OH), and the distal end was secured to the base of the material testing machine. The specimen was aligned with the loading axis of the machine and allowed to equilibrate in the saline bath for 30 minutes prior to testing. A small preload (0.1 N) was applied to the tissue and 10 cycles of preconditioning to 7% of the clamp-to-clamp distance was performed at a rate of 10 mm/min. A load to failure test was performed immediately after the preconditioning regimen, as previously described. The load (force, N) and elongation (distension, mm) of the tissue were recorded and used to generate load-elongation curves, which were then converted to a stress-strain relationship to calculate the mechanical properties of the vagina in the longitudinal direction. Here, stress is defined as the load (a measure of the force applied to the tissue) divided by cross-sectional area (measured by the laser micrometer), and strain was defined as the change in the marker distance divided by the original distance between the markers (Δl/l _o or change in length of the specimen relative to its initial length). The slope of the linear region of the stress-strain curve was defined as the tangent modulus (a measurement of stiffness), while the tensile strength and maximum strain were recorded at failure (point at which the specimen breaks apart). The strain energy density was calculated by taking the area underneath the stress-strain curve until the point of failure.

Data analysis and statistics

All statistical analyses were performed using statistical software (SPSS v. 17.0; SPSS Inc, Chicago, IL). Our sample size was calculated based on preliminary data from virgin rats. Based on previously determined collagen ratios, we calculated that 10 rats per group would have 80% power at the 2-sided .05 significance level to detect at least a 35% difference in the collagen I/V ratio and 50% difference in the ratio of collagen I/III between injured and uninjured animals. For mechanical data, 6 rats per group would have 80% power at the 2-sided .05 significance level to detect at least a 45% difference in the tensile strength and a 30% difference in the tangent modulus. Since the values of the skewness and kurtosis statistics did not indicate a departure from symmetry in the data distribution, Student t tests and 1-way analysis of variance with a Sidak post hoc were used to evaluate differences in the mean collagen ratios (I/III and I/V), mechanical properties (tensile strength, maximum strain, tangent modulus, and strain-energy density), as well as the continuous variables of the anatomical and demographic data between the injured and uninjured and treated and untreated rats. Since the variances for vaginal diameter were not equal based on Levene test, post hoc P values were determined using Student t test for unequal variances and adjusted using the Bonferroni correction. Statistical tests were evaluated at the 2-sided significance level of .05.