Vaginal host response to polycarbonate urethane, an alternative material for the repair of pelvic organ prolapse

Abstract

Complications following surgical repair of pelvic organ prolapse (POP) with polypropylene mesh (PPM) are common. Recent data attributes complications, in part, to stiffness mismatches between the vagina and PPM. We developed a 3D printed elastomeric membrane (EM) from a softer polymer, polycarbonate urethane (PCU). EMs were manufactured with more material given the low inherent material strength of PCU. We hypothesized that the EMs would be associated with an improved host response as compared to PPM. A secondary goal was to optimize the material distribution (fiber width and device thickness) within EMs, in regards to the host response. EM constructs (2×1cm²) with varied polymer stiffness, fiber width, and device thickness were implanted onto the vagina of New Zealand white rabbits for 12 weeks and compared to similarly sized PPMs. Sham implanted animals served as controls. Mixed effects generalized linear models were used to compare the effect of construct type accounting for differences in independent variables. EMs had an overall superior host response compared to PPM as evidenced by preservation of vaginal smooth muscle morphology (p-values<0.01), decreased total cellular response to construct fibers (p-values<0.001), and a reduced percent of macrophages (p-values<0.02) independent of how the material was distributed. Both PP and EMs negatively impacted vaginal contractility and glycosaminoglycan (GAG) content relative to Sham (all p-values<0.001) with EMs having less of an impact on GAGs (p-values<0.003). The results suggest that softer PCU EMs made with more material are well tolerated by the vagina and comprises a future material for POP repair devices.

Graphical Abstract

1. Introduction

Pelvic organ prolapse (POP) is characterized by the unnatural descent of the pelvic organs into the vagina. Approximately 13% of US women will undergo at least one surgical repair of POP by age 80 with costs estimated to exceed $1.5 billion annually[1, 2]. Surgical repairs utilizing a patient’s own tissue are associated with high recurrence rates with ~40% failing within 2 years, and ~70% failing by 5 years[3, 4]. To improve durability of POP repairs, synthetic meshes, most commonly polypropylene meshes, are implanted via a transabdominal route (e.g. sacrocolpopexy), which has been shown to be a safe and viable option[5]. However, mesh related complications after polypropylene mesh (PPM) procedures occur at ~10% over 5 years and increase with time[6–8]. The lack of an optimal POP repair option has prompted an urgent need for an alternative device that will improve patient outcomes particularly given the increased demand for POP repair surgery associated with the rapidly aging population[9].

Compelling mechanistic data in a nonhuman primate model show that high structural stiffness of a mesh negatively impacts vaginal morphology, structural integrity, and function, ultimately contributing to mesh complications[10–15]. Polypropylene has a polymer stiffness that is orders of magnitude greater than the human vagina (polypropylene ~1–2 GPa vs vagina ~6–34 MPa[16–18]). To reduce the stiffness, polypropylene is extruded into thin filaments or fibers which are then knitted into a product. Despite these design choices, a high discrepancy in polymer stiffness between the vagina and polypropylene persists. The stiffness mismatch predisposes to stress shielding, a phenomenon in which the stiffer material precludes (shields) the less stiff material (vagina) from experiencing normal physiologic loads[19, 20]. Indeed, evidence of stress shielding has been identified following implantation of PPM in both women and nonhuman primates[10–15]. In addition, these stiffer materials can cause chronic inflammation related to rigid micromotion of the device against the underlying soft tissue[21–24]. While we acknowledge that other mechanisms such as infection and the foreign body response contribute to complications, here, we focus on the impact of stress shielding and micromotion injury by manufacturing a POP repair device from a polymer with a stiffness that is similar to that of the vagina.

To this end, we fabricated devices from polycarbonate urethane (PCU) as a softer alternative polymer to polypropylene. PCU is a biocompatible, biodegradable resistant, thermoplastic elastomer with material properties that can be tuned to match various soft tissues, like the vagina[25–27]. Since softer materials have lower inherent material strength, we designed our initial iterations of the elastomeric membranes or EMs (i.e., a mesh-like device without knots) with more material than current polypropylene meshes. In the current study, we compared the host response to EMs manufactured from two grades of PCU with polymer stiffnesses that generally matched the stiffness of the human vagina - a lower polymer stiffness PCU (EM 1, ~8–65 MPa) and a higher polymer stiffness PCU (EM 2, ~28–149 MPa). Previously, membranes manufactured from the lower polymer stiffness of PCU were shown to have favorable performance in ex vivo mechanical tests[28]. To determine the optimal distribution of the material within the device, we compared the impact of varying fiber width and device thickness. Comparisons were made to the polypropylene meshes (PPMs) - Restorelle, a lightweight PPM (LWPP) and Gynemesh PS, a midweight PPM (MWPP, Ethicon, Somerville, NH, USA). Relative to PPMs, we hypothesized that the EMs would have an improved host response as characterized by integration into the vagina, preservation of vaginal structure and function, and a reduced host total immune cell and macrophage response. We further proposed that EM 1, because of the lower polymer stiffness, would have a more favorable response than EM 2. In this initial assessment, we sought to define the response to the device alone in the absence of tension which has an additional independent impact on the host response. We used a rabbit model to test multiple iterations of the device given the large number of implants and establishment of the rabbit as a large animal prolapse repair model[29]. We appreciate, however, that the rabbit vagina (~6 MPa) is substantially less stiff than the nonhuman primate vagina (~27 MPa)[30, 31] and stress mismatches will likely be observed.

2. Materials and Methods

2.1. Polycarbonate urethane

Medical grade pellets of polycarbonate-based thermoplastic polyurethanes (TPUs) with Shore Hardness 95A (Carbothane^™ AC-4095A - a low polymer stiffness grade of polycarbonate urethane, PCU) and Shore Hardness 75D (Carbothane^™ AC-4075D - a high polymer stiffness grade of PCU) were obtained from Lubrizol (Brecksville, OH). According to the manufacturer (Lubrizol), the polymer is composed of a soft segment (a long chain difunctional alcohol, polyol) and hard segment (consisting of a diisocyanate and a short chain diol which acts as a chain extender). Note: specific information regarding the composition of the polymers is Lubrizol proprietary information. Altering the hard and soft segments determines the hardness of the polymer in which a higher hard to soft segment ratio would result in a polymer with increased polymer hardness.

2.2. Elastomeric membrane (EM) and polypropylene mesh constructs

Pellets of Carbothane^™ AC-4095A and Carbothane^™ AC-4075D were dried and extruded into filaments. Extruded filaments were then used to manufacture the elastomeric membranes (EMs, AC-905A to manufacture EM 1 and AC-4075D to manufacture EM 2) using a Prusa MK3S (Prusa Research) fused deposition modeling three-dimensional (3D) printer (see Supplemental Material Section 2.1 for additional information regarding the EMs manufacturing and printing parameters). PCU EMs were designed to have a square pore geometry because of the favorable outcomes associated with a square pore mesh[10, 11, 13–15, 32]. To determine the best way to distribute the amount of material in the device, we investigated the following fiber width × device thickness configurations: 0.4 mm × 0.4 mm, 0.4 mm × 0.6 mm, and 0.6 mm × 0.4 mm. Hereafter, these fiber width × device thickness configurations have been represented as 4×4, 4×6, and 6×4, respectively. The pore size of the manufactured PCU EMs averaged between 2 mm and 2.2 mm as confirmed using a SkyScan 1272 microCT system (Bruker, Billerica, MA, USA, Supplemental Figure 1) and Blender Version 4.1 (blender.org) for pore size quantification. Commercially available knitted polypropylene meshes, Restorelle (LWPP, Coloplast, Minneapolis, MN, USA) and Gynemesh PS (MWPP, Ethicon, Somerville, NH, USA) were implanted for comparison. Restorelle was chosen because it is a square pore, lightweight (19 g/m²) and high porosity (78%) mesh that has been shown to have the least negative impact on the structure and function of the vagina among polypropylene meshes[10, 11, 13–15]. Gynemesh PS, the prototype prolapse mesh, has a hexagon-shaped pore geometry, and it is a midweight (42 g/m²) and low porosity (62%) mesh shown to induce a maladaptive remodeling response[10, 11, 13–15]. The PCU EMs were manufactured with square pores and considerably more material resulting in heavier weight products (EM 1 = 130 g/m² and EM 2= 169 g/m²) with wider fibers and thicker devices (Table 1). It is important for the reader to keep in mind that the polypropylene meshes are knitted, consisting of multiple fibers, and depending on the measurement technique, the fiber width (often reported as fiber diameter) and thickness can vary, hence we reported a range. The low aspect ratio (length to width) of the membranes and meshes implanted in this study precluded tensile testing. Thus, we calculated the structural stiffness of the EM constructs based on prior testing of larger constructs (9.0 cm × 1.5 cm) since structural properties are scalable. By this method, we found that the 2 cm × 1 cm EMs have a lower structural stiffness than the 2 cm × 1 cm polypropylene mesh constructs (Table 1).

Table 1.

Fiber width, device thickness, and structural stiffness of membrane and mesh constructs.

Construct	Fiber width (mm)	Device thickness (mm)	Structural stiffness (N/mm)^
EM 1 4×4	0.40	0.40	0.18 ± 0.01
EM 1 4×6	0.40	0.60	0.28 ± 0.02
EM 1 6×4	0.60	0.40	0.28 ± 0.02
EM 2 4×4	0.40	0.40	0.22 ± 0.08
EM 2 4×6	0.40	0.60	0.34 ± 0.13
EM 2 6×4	0.60	0.40	0.34 ± 0.13
LWPP	0.080	0.20 – 0.30*	0.47 ± 0.10
MWPP	0.080 – 0.10*	0.10 – 0.40*	0.81 ± 0.05

Data displayed as mean ± standard deviation

^*Pending the measurement technique, the fiber width and thickness can vary; therefore, a range is reported.

^{^}Structural stiffness reflects the stiffness of the 2 cm × 1 cm constructs implanted in this study. It was calculated based on prior testing of larger constructs (9.0 cm × 1.5 cm).

2.3. Animals

Twenty-two New Zealand white rabbit, retired breeders (multi-parous, median parity of 15) were utilized in this study in accordance with methods approved by the University of Pittsburgh Institutional Animal Care and Use Committee (IACUC #22030732). The experimental unit was defined as a single animal. Rabbits were fed a standard rabbit diet supplemented with hay and greens ad libitum and were housed in standard cages on a 12-hour alternating light/dark cycle.

2.4. Surgical procedures

All surgical procedures were performed with animals under anesthesia (isoflurane: 1–3%). Following intubation and a midline laparotomy, the bladder and rectum were gently dissected away from the anterior and posterior vagina, respectively. The anatomy was maintained with little perturbation beyond the creation of a bladder and rectal flap for placement of EM and mesh constructs along the entire length of the vagina. PCU EMs and polypropylene meshes with varied fiber width and device thickness were secured to the anterior and posterior vagina (n=3 per side, for a total of n=6 per animal) in construct units of 2 cm long × 1 cm wide (n=24 per EM 1 and EM 2) (Figure 1). For comparison, similarly sized constructs of the polypropylene meshes – LWPP (n=12) and MWPP (n=12) were implanted. Roughly 1 cm was maintained between adjacent constructs, and a random number generator was used to determine the order of placement. All constructs were placed in the absence of tension to isolate the host response to the device alone, as tension can independently modify the host response. A similar procedure was performed for Sham animals (n=10) without constructs implanted. The rectus fascia and skin were closed with delayed absorbable suture. After 12-weeks, the vagina-construct complexes and vagina alone (Sham) were excised and harvested for structural and functional analyses.

Figure 1:

Implantation scheme on the rabbit vagina with uterus and support structures left intact. Three constructs were implanted on both the anterior and posterior vagina in random order. At 12 weeks post implantation, differences in polypropylene mesh vs PCU elastomeric membranes are apparent with polypropylene meshes encapsulated vs PCU elastomeric membranes that have good tissue incorporation.

2.5. Histology staining

2.5.1. Masson’s trichrome

Tissue sections of vagina and vagina-constructs were stained with Masson’s Trichrome to assess the overall tissue morphology and thickness of the muscularis (i.e., smooth muscle layer). Previous studies by us have shown that this layer is the most susceptible to the effects of stress shielding[11, 13, 14]. Samples were fresh frozen in optimal cutting temperature (OCT) compound with 2-methylbutane, cooled with liquid nitrogen, and cryosectioned at 7 μm. Sections were stained with Trichrome Stain AB Solution (Sigma-Aldrich, St. Louis, MO, USA) and Hematoxylin Gill No. 2 (Sigma-Aldrich, St. Louis, MO, USA) solutions prior to imaging at 4X using Nikon Eclipse 90i and Ni-E microscopes (Melville, NY, USA). The inner and outer borders of the muscularis were outlined on cross sections, and the distance between these borders (i.e., the average thickness of the smooth muscle layer) was calculated using a custom script[29].

2.5.2. Hematoxylin & eosin staining

Cryosections of vagina with and without constructs were dried at room temperature, fixed in O-Fix Tissue Fixative (Leica Biosystems, Deer Park, IL), and stained with SelecTech Hematoxylin 560 MX (Leica Biosystems, Deer Park, IL). Slides were placed in SelecTech Blue Buffer 8 Ready-to-Use (Leica Biosystems, Deer Park, IL), dipped in 70% ethanol, and counterstained with Eosin Phloxine 515 (Leica Biosystems, Deer Park, IL). Sections were dehydrated with increasing concentrations of ethanol, cleared with Xylenes, and then mounted. Stained sections were imaged with a Nikon Ni-E microscope (Melville, NY, USA) at 10X. To assess the cellular response to construct fibers, images were imported in QuPath Version 0.4.3 (Queen’s University, Belfast, Northern Ireland)[33] for analysis. Fibers were outlined and the area quantified using QuPath. The cellular response to mesh fibers (i.e., within the fibroma) was identified and outlined according to the change in stain intensity and cell type/density immediately adjacent to the fiber as compared to normal adventitia at a slightly greater distance from each membrane or mesh fiber. The area of the cellular response was calculated and used to define the “total cellular response”. To take into account differences in fiber thickness and number of fibers within the cellular response, this value was reported as a ratio of the cellular response area to fiber area.

2.6. Immunofluorescence labeling

Cross sections of the vagina were embedded in optimal cutting temperature compound and cryosectioned at 7 μm. Tissue sections were incubated at 4°C overnight with the following primary antibodies (1:500 dilution): macrophage/monocyte monoclonal antibody, clone RbM2 (Abnova, Taipei City, Taiwan), goat polyclonal anti-alpha smooth muscle actin (αSMA, Abcam, Waltham, MA), and chicken polyclonal vimentin (Novus Biologicals, Centennial, CO). This triple labeling method allowed us to distinguish smooth muscle myocytes (aSMA⁺), fibroblasts (vimentin⁺), and macrophages (RbM2⁺). Following overnight incubation, tissue sections were washed, and the following secondary antibodies (1:200 dilution) were applied: Horse Anti-Mouse IgG Antibody (H+L), Biotinylated (Vector Laboratories, Newark, CA), Donkey anti-Goat conjugated with Alexa Fluor^™ Plus 594 (Thermo Fisher Scientific, Waltham, MA), and Donkey anti-Chicken conjugated with Alexa Fluor^™ 647 (Thermo Fisher Scientific, Waltham, MA). Tissue sections were then washed and incubated with Alexa Fluor^® 488 Streptavidin (1:200 dilution, Jackson ImmunoResearch Inc, West Grove, PA). Lastly, tissue sections were washed, cover slipped using VECTASHIELD^® PLUS Antifade Mounting Medium with DAPI (Vector Laboratories, Newark, CA), and imaged at 10X with a Nikon Ni-E microscope (Melville, NY).

Triple labeled images were imported into QuPath version 0.4.3 for analysis of the adventitia layer of the vagina, the site of device implantation, defined as the loose connective tissue region adjacent to the muscularis (positive αSMA labeling). All construct fibers were outlined and the area quantified. Areas within the adventitia containing suture which evokes an independent robust macrophage response were removed to isolate the macrophage reaction specific to the EMs or PPMs alone. The cell detection tool for the DAPI channel was used keeping all default settings except for the threshold which was set to 400 to obtain a total cell count within the adventitia. A single measurement classifier was created by filtering all DAPI detections through the FITC channel and then thresholding so that only cell nuclei surrounded by RbM2 labeling would be detected as positive. Utilizing these methods, the following parameters were quantified: total cells within the adventitia, total macrophages, percent macrophages per total cells, and ratio of mesh area per adventitia area – a quantitative measure of mesh burden.

2.7. Total collagen content

As previously described[13, 34], quantification of the total collagen content was conducted on lyophilized and papain-digested (final concentration: 125 μg/ml) tissue samples. Hydroxyproline solution (1 mg/ml) and purified collagen type I solution (3.6 mg/ml, Sigma, St. Louis, MO) served as standards. Collagen content was calculated by assuming hydroxyproline comprises 14% of the amino acid composition of collagen[13, 35]. The assay was performed in duplicate for each sample. Total collagen content was normalized to tissue dry weight.

2.8. Glycosaminoglycan content

The amount of sulfated glycosaminoglycan (sGAG) within papain-digested vagina samples (papain final concentration, 125 μg/ml) was quantified using the 1,9-dimethylmethylene blue (DMMB) assay as previously described[13, 34]. Samples and serial dilutions of Blyscan sGAG reference standard (100 μg/ml, Biocolor, United Kingdom) were diluted in ddH2O and freshly prepared papain digestion solution. After incubation with DMMB dye and centrifugation, supernatants were read on a SpectraMax iD3 microplate reader (Molecular Devices, San Jose, CA) at 595 nm. The assay was performed in duplicate for each sample and the amount of sGAG was normalized to tissue dry weight.

2.9. Vaginal contractile function

A Radnoti 8 Channel Tissue Bath System equipped with eight isometric force transducers (ADInstruments, Colorado Springs, CO, USA) was used to perform an organ bath assay to assess vaginal contractile function following the implantation of the constructs as previously described[29]. Briefly, vaginal strips (~7 mm × 2 mm) with and without constructs were cut along the circumferential direction, clamped on opposing ends, and submerged in an organ bath filled with oxygenated Krebs solution. Care was taken to ensure that only vagina and not constructs was clamped. Importantly, constructs were not separated from the vagina as this injures the vagina during the dissection. Following the application of a 1.0 g preload and an equilibrium incubation period, tissues were stimulated with 120 mM KCl to induce a smooth muscle contraction. The maximum force generated was recorded and normalized by the volume of the sample to account for any differences in sample dimensions.

2.10. Statistics

Sample size was determined based on differences in smooth muscle parameters in previously published data in which Sham and the polypropylene meshes (Restorelle and Gynemesh PS) were compared[11, 13, 14]. Using this data, we determined that eight constructs would be needed per group for a power of 80% using a one-way analysis of variance with a two-sided significance level of 0.05. Age and weight of the rabbits were compared between constructs using Kruskal-Wallis tests. Since 6 construct/position variations were implanted in each animal, mixed effects generalized linear models were used to approximate the effects of construct type, implantation position, age, and weight on all outcome measures. Estimates of the unadjusted mean difference and adjusted mean difference, when appropriate, for all outcome measures from the referent in each model are shown along with 95% confidence intervals and corresponding p-values. Pairwise comparisons were made using the ‘lincom’ command which calculated point estimates for the difference between coefficients estimated by the parent model, with statistical inference based on the z-statistic. Significance level was set to p<0.05 unless otherwise specified, and all analyses were performed using STATA Version 18 (Stata Corp, College Station, TX, USA).

3. Results

Twelve parous, New Zealand white rabbits underwent laparotomy; eight were implanted with elastomeric membrane constructs and four implanted with polypropylene mesh constructs. Three constructs (2 cm × 1 cm) were implanted onto both the anterior and posterior vagina for a total of six per animal (Figure 1). Ten animals in which no constructs were implanted served as controls. All animals had similar weight (p=0.530) but those implanted with PPMs were older than Sham (3.8 vs 2.7 years, p=0.002) but not different from EM groups (3.3 years, p>0.05) (Table 2). In total, 72 constructs were analyzed with the final sample size for each group: Sham (n=10), LWPP (n=12), MWPP (n=12), EM 1 4×4 (n=8), EM 1 4×6 (n=8), EM 1 6×4 (n=8), EM 2 4×4 (n=8), EM 2 4×6 (n=8), EM 2 6×4 (n=8). For some of the outcomes, the “implantation location” of the construct on the vagina (proximal, mid, distal) impacted the result. Therefore, in our mixed effects generalized linear model we provide data adjusted for age and location on the vagina where appropriate.

Table 2.

Demographics of rabbits implanted with no device, elastomeric membrane (EM), and polypropylene mesh (PPM) constructs.

	Age (yrs)	Weight (kg)
Sham rabbits	2.7 (2.3 – 2.8)	4.2 (3.9 – 4.5)
EM implanted rabbits	3.3 (3.2 – 3.4)	4.2 (4.0 – 4.3)
PPM implanted rabbits	3.8 (3.2 – 3.9)	4.1 (3.9 – 4.2)
P-value*	.002	.530

Data displayed as median (interquartile range)

Eight rabbits were implanted with EMs (six EM constructs implanted per rabbit).

Four rabbits were implanted with PPMs (six PPM constructs implanted per rabbit).

Devices were not implanted in ten rabbits (Sham).

^*P-value from Kruskal-Wallis test

For consistency, the results are presented by first discussing the impact of polymer stiffness which will consist of the results from 1) the overall group differences relative to a reference group (either Sham or LWPP) irrespective of fiber dimensions (i.e., Sham or LWPP vs MWPP, EM 1, and EM 2 were compared as appriopriate) and 2) the results of pairwise comparisons between the groups. Next, the results from comparing the fiber dimensions within EM 1 and EM 2.

3.1. Mesh integration and smooth muscle morphology

Grossly at the time of dissection, the EMs and PPMs appeared flat and maintained the implantation location. Trichrome stained micrographs of PCU EMs showed them to be well integrated into the vagina and to follow the vaginal contour whereas polypropylene was encapsulated above the adventitia. MWPPs showed the least amount of integration. Relative to Sham, smooth muscle architecture of most PCU EMs and all polypropylene constructs appeared less densely packed and had more longituindally oriented smooth muscle bundles than circular (Figure 2). Smooth muscle architecture was most preserved following implantation of PCU EM 1 thinner fiber (0.4 mm × 0.6 mm) constructs, closely resembling that of the Sham vagina.

Figure 2:

Representative tissue sections of Sham, lightweight (LWPP) and midweight (MWPP) polypropylene mesh constructs, and lower (EM 1) and higher stiffness (EM 2) heavyweight PCU elastomeric membrane constructs stained with Masson’s trichrome. Vaginal layers and normal architecture shown in Sham. Adventitia is the site of device implantation. Smooth muscle layer architecture showed the least amount of disruption following implantation of EM 1 constructs with more densely packed bundles that most closely resembled Sham compared to all other constructs. The EM 1 image is a 0.4 mm × 0.6 mm construct, and the EM 2 image is a 0.6 mm × 0.4 mm construct. Asterisks (*) correspond to fiber.

Impact of polymer stiffness

In previous studies, we have found that a decrease in vaginal muscularis thickness is a sensitive indicator of stress shielding[11, 13, 14]. Consistent with these studies, muscularis thickness was negatively impacted by polypropylene mesh with a roughly 20% decrease in thickness for both LWPP [mean difference −232.16 (−340.51, −123.80)] and MWPP [mean difference −259.42 (−343.50, −175.34)] relative to Sham (both p-values<0.001, Table 3 and Figure 3). In contrast, smooth muscle thickness was preserved at Sham levels following implantation of both the lower stiffness and higher stiffness EMs (p-values>0.05). Relative to the EMs, PPMs had a more negative impact on vaginal smooth muscle thickness. Specifically, LWPP and MWPP had decreased smooth muscle thickness relative to both EM 1 (both p-values<0.001) and EM 2 (p=0.006 and p=0.001, respectively). There was no difference in smooth muscle thickness between the PPM constructs (p=0.576).

Table 3.

Thickness of the muscularis (i.e., smooth muscle layer of the vagina).

	Thickness (μm) Mean (SD)	Unadjusted Mean Difference (95% CI)	P-value*
Sham (n=10)	1360.53 (129.35)	Referent	<.001†
LWPP (n=11)	1139.45 (251.59)	−232.16 (−340.51, −123.80)	<.001
MWPP (n=11)	1112.19 (84.93)	−259.42 (−343.50, −175.34)	<.001
EM 1 (n=24)	1404.17 (268.05)	44.95 (−93.63, 183.52)	.525
EM 2 (n=24)	1435.02 (375.56)	56.24 (−143.06, 255.55)	.580
Pairwise comparisons
LWPP vs MWPP		27.27 (−68.18, 122.72)	.576
LWPP vs EM 1		−277.10 (−422.27, −131.94)	<.001
LWPP vs EM 2		−288.40 (−492.39, −84.41)	.006
MWPP vs EM 1		−304.37 (−432.48, −176.25)	<.001
MWPP vs EM 2		−315.66 (−507.85, −123.48)	.001
EM 1 vs EM 2		−11.30 (−183.87, 161.27)	.898

Data presented reflects the overall group comparisons. EM 1 and EM 2 include the respective 4×4, 4×6, and 6×4 constructs.

^*P-value from mixed effects linear regression

^†Overall P-value from unadjusted mixed effects linear regression model. Position and age not associated after adjusting for construct (P>0.05).

Figure 3:

Smooth muscle thickness quantification. A significant difference from Sham is indicated by *, from LWPP is indicated by θ, and from MWPP is indicated by $.

Impact of fiber dimesions

Changing the fiber dimensions of the EMs, smooth muscle thickness was not impacted for the EM 2 constructs (p=0.335) whereas smooth muscle thickness was thinner for EM 1 6×4 compared to EM 1 4×6 [mean difference 167.57 (33.07, 302.07), p=0.015] (Supplemental Table 1).

3.2. Cellular response

To characterize the cellular response to the PCU EM and PPM constructs, the following were quantified: number of cells within the adventitita, total cellular response to construct fibers (i.e., fibroma cell area per fiber area), total macrophages, and percentage of mcrophages per total cells.

3.2.1. Total cells within the adventitia

Impact of polymer stiffness

Overall, the total number of cells within the adventitia trended to be higher for the PCU EMs relative to PPMs but did not reach statistical significance (Supplemental Material Table 2). A significant difference was only observed between LWPP and EM 2 (adjusted mean difference 4441 (124, 8758), p=0.044).

Impact of fiber dimensions

Changing the fiber width and thickness did not impact the total number of cells within the adventitia for EM 1 (p=0.362) nor EM 2 (p=0.977).

3.2.2. Total cellular response to construct fibers

Impact of polymer stiffness

Qualitatively, on H&E stained micrographs, there was a greater total cellular response to polypropylene fibers as compared to PCU fibers (Figure 4). This was confirmed in the quantitative analysis in which the ratio of the cellular response area to fiber area (i.e., total cellular response) was calculated to take into account differences in fiber size (Table 4 and Figure 5). By this method, PCU EMs had a marked reduction in the cellular response as compared to LWPP (LWPP vs EM 1 – mean difference −0.82 (−1.07, −0.58), p<0.001 and LWPP vs EM 2 – mean difference −0.81 (−1.06, −0.56), p<0.001). Similar reductions in the cellular response to PCU EMs were observed relative to MWPP [MWPP vs EM 1 – mean difference 0.65 (0.44, 0.86), p<0.001 and MWPP vs EM 2 – mean difference 0.63 (0.41, 0.85), p<0.001]. The total cellular response was not significantly different between the PPM constructs (p=0.374) nor between the PCU EMs (p=0.755).

Figure 4:

H&E staining of constructs comparing cellular response to lightweight (LWPP) and midweight (MWPP) as compared to heavyweight PCU EMs −1 and −2. As shown in the representative images, the cellular response is limited to the perimeter immediately around the fiber and is markedly reduced in PCU elastomeric membranes as compared to polypropylene mesh constructs despite substantially increased weight. The EM 1 image is a 0.4 mm × 0.4 mm construct, and the EM 2 image is a 0.4 mm × 0.4 mm construct. Asterisks (*) correspond to fiber.

Table 4.

Area of the cellular response surrounding membrane or mesh fibers to fiber area.

	Total Cellular Response (mm2/mm2) Mean (SD)	Unadjusted Mean Difference (95% CI)	P-value*
LWPP (n=12)	1.19 (0.34)	Referent	<.001†
MWPP (n=12)	1.02 (0.36)	−0.18 (−0.57, 0.21)	.374
EM 1 (n=24)	0.37 (0.21)	−0.82 (−1.07, −0.58)	<.001
EM 2 (n=23)	0.39 (0.16)	−0.81 (−1.06, −0.56)	<.001
Pairwise comparisons
MWPP vs EM 1		0.65 (0.44, 0.86)	<.001
MWPP vs EM 2		0.63 (0.41, 0.85)	<.001
EM 1 vs EM 2		−0.01 (−0.10, 0.08)	.755

Data presented reflects the overall group comparisons. EM 1 and EM 2 include the respective 4×4, 4×6, and 6×4 constructs.

^*P-value from mixed effects linear regression

^†Overall P-value from unadjusted mixed effects linear regression model. Position and age not associated after adjusting for construct (P>0.05).

Figure 5:

Total cellular response (ratio of the cellular response area to fiber area). A significant difference in total cellular area from LWPP is indicated by θ, and from MWPP is indicated by $.

Impact of fiber dimensions

In regards to differences in fiber dimensions within the EMs, the total cellular response was significantly reduced for EM 1 4×6 [mean difference −0.15 (−0.29, −0.01), p=0.034) and EM 1 6×4 (mean difference −0.22 (−0.36, −0.08), p=0.002] as compared to EM 1 4×4 (Supplemental Table 3). No other significant differences were observed in the total cellular response for EM 1 4×6 vs 6×4 (p=0.066) nor among the different EM 2 constructs (p=0.272).

3.2.3. Macrophage quantification

Following the surgical implantation of a material, macrophages play an essential role in the host response. Qualitatively, a lower number of macrophages were associated wtih PCU fibers as compared to the polypropylene fibers (Figure 6). Additionally, macrophages responding to PCU fibers formed a monolayer around each fiber while those responding to the polypropylene fibers were present in multiple layers.

Figure 6:

Immunofluorescent micrographs depicting macrophage response to vaginally implanted constructs. Thin cross sections labeled with RbM2 (pan macrophage), α smooth muscle actin (blood vessel), vimentin (fibroblast), and DAPI (nucleus) demonstrating drastically reduced response to PCU elastomeric membranes as compared to polypropylene meshes. The EM 1 image is a 0.6 mm × 0.4 mm construct, and the EM 2 image is a 0.6 mm × 0.4 mm construct. Asterisks (*) correspond to fiber.

3.2.3.1. Total macrophages (Supplemental Table 4)

Impact of polymer stiffness

Overall, the total number of macrophages was highest for the MWPP constructs. Macrophages were 71% increased for MWPP constructs as compared to the LWPP constructs [adjusted mean difference 1010.92 (223.75, 1798.08), p=0.012]. Similarly, the number of macrophages was 88% [adjusted mean difference 1257.37 (732.86, 1781.88), p<0.001] and 67% [adjusted mean difference 1096.96 (404.15, 1789.76), p=0.002] times higher for the MWPP constructs relative to the EM 1 and EM 2 constructs, respectively. Unlike the response to MWPP, a significant difference in the number of macrophages was not observed between the LWPP and EM 1 (p=0.253) nor EM 2 (p=0.790) constructs.

Impact of fiber dimensions

The number of macrophages was not impacted by the changes in fiber width and thickness for EM 1 (p=0.386) nor EM 2 (p=0.594).

3.2.3.2. Percentage of macrophages (Table 5 and Figure 7-left)

Table 5.

Percentage of macrophages following device implantation.

	Macrophages (%) Mean (SD)	Unadjusted Mean Difference (95% CI)	P-value*	Adjusted‡ Mean Difference (95% CI)	P-value
LWPP (n=12)	10.08 (4.18)	Referent	<.001†	Referent
MWPP (n=12)	16.35 (5.84)	6.27 (4.08, 8.46)	<.001	6.27 (4.08, 8.46)	<.001
EM 1 (n=24)	7.35 (3.00)	−2.73 (−5.03, −0.44)	.020	−3.08 (−5.29, −0.86)	.006
EM 2 (n=24)	7.80 (4.98)	−2.27 (−5.07, 0.53)	.112	−3.11 (−5.61, −0.61)	.015
Pairwise comparisons
MWPP vs EM 1		9.01 (6.22, 11.79)	<.001	9.35 (7.18, 11.52)	<.001
MWPP vs EM 2		8.54 (5.32, 11.76)	<.001	9.38 (6.87, 11.90)	<.001
EM 1 vs EM 2		−0.46 (−1.75, 0.82)	.478	0.03 (−0.94, 1.00)	.948
Position
Proximal (n=24)	10.64 (6.48)	Referent	.111†	Referent
Mid (n=24)	9.40 (5.95)	−1.24 (−4.36, 1.89)	.438	−1.24 (−4.37, 1.88)	.435
Distal (n=24)	8.32 (3.27)	−2.31 (−4.60, −0.03)	.047	−2.32 (−4.60, −0.04)	.046
Mid vs Distal		1.08 (−1.31, 3.46)	.376	1.08 (−1.31, 3.46)	.376
Age, months		0.01 (−0.57, 0.59)	.981	−0.20 (−0.31, −0.08)	.001

Data presented reflects the overall group comparisons. EM 1 and EM 2 include the respective 4×4, 4×6, and 6×4 constructs.

^*P-value from mixed effects linear regression

^†Overall P-value from unadjusted mixed effects linear regression model

^‡Adjusted for construct, position, and age.

Figure 7:

Percent macrophages per total cells responding to each construct (left) and mesh burden - fiber area per tissue area (right). Mesh burden is increased in midweight polypropylene mesh (PPM) and PCU elastomeric membranes (EMs) relative to lightweight PPM. EMs have increased mesh burden relative to both midweight (MWPP) and lightweight (LWPP). Despite the increased mesh burden, the percent macrophages responding to the construct fibers per total cells in the adventitia is highest for the MWPP, and higher for both PPMs than the PCU EMs indicating a reduced inflammatory response to PCU as compared to polypropylene. A significant from LWPP is indicated by θ, and from MWPP is indicated by $.

Impact of polymer stiffness

Normalizing the number of macrophages by the total number of cells, the percentage of macrophages was 62% higher for the MWPP constructs relative to the LWPP constructs [adjusted mean difference 6.27 (4.08, 8.46), p<0.001]. Following the implantation of the EM constructs, the percentage of macrophages was significantly reduced. Compared to LWPP, the percentage of macrophages decreased by 27% [adjusted mean difference −3.08 (−5.29, −0.86), p=0.006] and 23% [adjusted mean difference −3.11 (−5.61, −0.61), p=0.015)] for EM 1 and EM 2 constructs, respectively. A greater reduction in the percentage of macrophages was observed in comparison to MWPP for both EMs, decreasing by 55% [EM 1 – adjusted mean difference 9.35 (7.18, 11.52), p<0.001] and 52% [EM 2 – adjusted mean difference 9.38 (6.87, 11.90), p<0.001]. The percentage of macrophages was not significantly different between EMs (p=0.948).

Impact of fiber dimensions

Interestingly, increasing EM fiber thickness while maintaining fiber width resulted in a 22% reduction in the percentage of macrophages when comparing EM 1 4×4 to EM 1 4×6 [mean difference −2.25 (−4.11, −0.39), p=0.018] (Supplemental Table 5). No other differences were observed.

3.3. Mesh burden quantification (Table 6 and Figure 7-right)

Table 6.

Mesh burden - mesh area per adventitia area.

	Mesh Burden (mm2/mm2) Mean (SD)	Unadjusted Mean Difference (95% CI)	P-value*	Adjusted‡ Mean Difference (95% CI)	P-value
LWPP (n=12)	0.03 (0.02)	Referent	<001†	Referent
MWPP (n=12)	0.07 (0.05)	0.05 (0.02, 0.07)	.001	0.05 (0.02, 0.07)	.001
EM 1 (n=24)	0.10 (0.06)	0.08 (0.06, 0.10)	<.001	0.07 (0.05, 0.10)	<.001
EM 2 (n=24)	0.12 (0.08)	0.10 (0.07, 0.12)	<.001	0.09 (0.06, 0.12)	<.001
Pairwise comparisons
MWPP vs EM 1		−0.03 (−0.06, −0.003)	.031	−0.02 (−0.05, −0.005)	.016
MWPP vs EM 2		−0.05 (−0.08, −0.02)	.004	−0.04 (−0.07, −0.02)	.002
EM 1 vs EM 2		−0.02 (−0.04, 0.004)	.102	−0.02 (−0.04, 0.003)	.095
Age		−0.004 (−0.01, 0.002)	.215	−0.002 (−0.003, −0.00004)	.044

Data presented reflects the overall group comparisons. EM 1 and EM 2 include the respective 4×4, 4×6, and 6×4 constructs.

^*P-value from mixed effects linear regression

^†Overall P-value from unadjusted mixed effects linear regression model

^‡Adjusted for membrane and age; position not associated (P>0.05).

Because our PCU EMs were substantially heavier than our PPMs, we sought to quantify this by calculating mesh burden which is represented as the area occupied by construct fibers relative to tissue area (i.e., area of the adventitia). By this method and perhaps not surprisingly, mesh burden progressively increased as the weight of the constructs increased with LWPP<MWPP<EMs (p<0.001). There was no difference in mesh burden between EM 1 and EM 2 (p=0.095). Thus, even though the EMs present a significantly larger material burden to the tissue, the percent macrophages responding to the PCU fibers was decreased.

3.4. Matrix quantification: collagen and sGAG assays

3.4.1. Collagen quantification (Figure 8-left and Supplemental Table 6 & 7)

Figure 8:

Biochemical analysis of key extracellular matrix components. A significant difference from Sham is indicated by *, from LWPP is indicated by θ, and from MWPP is indicated by $.

Impact of polymer stiffness

Total collagen content (measured via hydroxyproline assay) was not significantly different from Sham vagina following the implantation of either the EM or PPM constructs (p=0.300, Supplemental Table 6). Relative to the LWPP implanted vaginas, the amount of collagen was approximately 29% and 26% higher for the vaginas implanted with EM 1 [mean difference −5.08 (−10.02, −0.15), p=0.044] and EM 2 [mean difference −4.88 (−9.53, −0.23), p=0.040], respectively. Differences in collagen content were not observed between the MWPP and EM implanted vaginas (p-values all >0.05).

Impact of fiber dimensions

Adjusting for the dimensions of the EM constructs impacted EM 1 (p=0.015) but not EM 2 (p=0.656) (Supplemental Table 6). Specifically, there was significantly less collagen associated with the EM 1 6×4 implanted vaginas as compared to EM 1 4×4 implanted vaginas [mean difference −7.65 (−13.03, −2.28), p=0.005] meaning less collagen was associated with a thicker fiber in this construct (Supplemental Table 7).

3.4.2. Glycosaminoglycan (sGAG) quantification (Figure 8-right and Supplemental Table 6 & 7)

Impact of polymer stiffness

Overall, the amount of sGAG within the vagina was significantly reduced following the implantation of the PCU and polypropylene constructs as compared to Sham (p<0.001) with decreases in the PPM groups greater than those in the EM groups (Supplemental Table 6). Indeed, sGAG content in the LWPP groups was lower than EM 1 and EM 2 [mean difference −0.17 (−0.27, −0.07), p=0.001 and −0.16 (−0.25, −0.07), p<0.001, respectively]. Similarly, sGAGs were lower in MWPP constructs than EM 1 and EM 2 [mean difference −0.15 (−0.25, −0.05), p=0.002 and −0.15 (−0.24, −0.06), p=0.001, respectively].

Impact of fiber dimensions

Congruent with the trends observed for collagen content, changing the construct fiber dimensions impacted the EM 1 constructs (p<0.001) but not the EM 2 constructs (p=0.194) (Supplemental Table 7). Relative to EM 1 4×4, sGAG content decreased 15% for EM 1 4×6 [mean difference −0.19 (−0.31, −0.06), p=0.003] and an even greater reduction of 25% was observed for EM 1 6×4 [mean difference −0.31 (−0.45, −0.16), p<0.001]. There was no difference in sGAG content between EM 1 4×6 and EM 1 6×4 constructs (p=0.100).

3.5. Vaginal contractile function (Active mechanics)

Impact of polymer stiffness

As a functional outcome, we measured vaginal contractility in an organ bath assay. Contractility was negatively impacted by the polypropylene and EM constructs (Table 7 and Figure 9). Relative to Sham, vaginal contractility decreased by ~46–60% for LWPP, MWPP, EM 1 and EM 2 (all p-values<0.001). Comparing the polypropylene constructs, contractility was significantly higher for LWPP as compared to MWPP [adjusted mean difference 0.26 (0.04, 0.47), p=0.018]. Similarly, contractility was significantly reduced for MWPP as compared to EM 2 [adjusted mean difference −0.43 (−0.81, −0.04), p=0.029]. No other significant differences were observed in contractility between the polypropylene, EM 1, and EM 2 constructs (all p-values>.05) nor between EM 1 and EM 2 [adjusted mean difference −0.17 (−0.51, 0.16), p=0.314].

Table 7.

Contractile force in response to 120 mM KCl.

	Contractile Force, mN/mm3 Mean (SD)	Unadjusted Mean Difference (95% CI)	P-value*	Adjusted‡ Mean Difference (95% CI)	P-value
Sham (n=7)	2.91 (0.57)	Referent	<001†	Referent
LWPP (n=12)	1.42 (0.35)	−1.49 (−1.84, −1.14)	<.001	−1.55 (−1.89, −1.20)	<.001
MWPP (n=12)	1.17 (0.50)	−1.75 (−2.13, −1.36)	<.001	−1.80 (−2.19, −1.42)	<.001
EM 1 (n=24)	1.45 (0.74)	−1.44 (−1.96, −0.91)	<.001	−1.54 (−2.08, −1.01)	<.001
EM 2 (n=24)	1.57 (0.54)	−1.37 (−1.80, −0.93)	<.001	−1.38 (−1.86, −0.89)	<.001
Pairwise comparisons
LWPP vs MWPP		0.26 (0.04, 0.47)	.018	0.26 (0.04, 0.47)	.018
LWPP vs EM 1		−0.05 (−0.50, 0.40)	.828	0 (−0.43, 0.43)	.998
LWPP vs EM 2		−0.12 (−0.47, 0.22)	.485	−0.17 (−0.52, 0.17)	.330
MWPP vs EM 1		−0.31 (−0.79, 0.18)	.213	−0.26 (−0.72, 0.21)	.278
MWPP vs EM 2		−0.38 (−0.76, 0.004)	.053	−0.43 (−0.81, −0.04)	.029
EM 1 vs EM 2		−0.07 (−0.45, 0.30)	.703	−0.17 (−0.51, 0.16)	.314
Position
Proximal (n=24)	1.24 (0.59)	Referent	<001†	Referent
Mid (n=38)	1.94 (0.94)	0.21 (−0.01, 0.43)	.060	0.16 (−0.06, 0.38)	.149
Distal (n=24)	1.70 (0.53)	0.46 (0.28, 0.65)	<.001	0.50 (0.30, 0.70)	<.001
Mid vs Distal		−0.25 (−0.45, −0.04)	.020	−0.34 (−0.53, −0.14)	.001

Data presented reflects the overall group comparisons. EM 1 and EM 2 include the respective 4×4, 4×6, and 6×4 constructs.

^*P-value from mixed effects linear regression

^†Overall P-value from unadjusted mixed effects linear regression model

^‡Adjusted for construct and position; age not associated (P>0.05).

Figure 9:

Contractile force following a 120 mM KCl stimulant. A significant difference from Sham is indicated by *, from LWPP is indicated by θ, and from MWPP is indicated by $.

Impact of fiber dimensions

Adjusting the fiber dimensions, vaginal contractile function was significantly higher for EM 1 4×6 as compared to EM 1 6×4 [mean difference 0.77 (0.21, 1.33), p=0.007]. A similar trend was observed for the EM 2 4×6 and 6×4 constructs; however, changing the fiber width and thickness of the EM 2 constructs did not significantly impact the contractile function of the vagina (p=0.321).

4. Discussion

Stiffness mismatches between current polypropylene meshes and the vagina have been shown to result in stress shielding and to compromise vaginal structure and function[10, 13]. To overcome the limitations of current polypropylene meshes, we have developed 3D printed elastomeric membranes (EMs) from a softer material, PCU, that can be tuned to match the stiffness of the vagina. Given that PCU has a lower stiffness and strength than poypropylene, elastomeric membranes were made with substantially more material than the polypropylene meshes used in this study. PCU can be tuned to match mechanical needs, and as such, we fabricated EMs from a lower stiffness grade of PCU and a higher stiffness grade of PCU, both with varied amounts of material. The EMs manufactured in this study are considered to be heavyweight devices. We then compared the host response to these materials to lightweight (LWPP, Restorelle) and midweight (MWPP, Gynemesh PS) polypropylene meshes. We further asked how best to distribute the amount of material across an EM (increasing fiber width vs thickness of the EM) without negatively impacting the host response. To isolate the host response to the EM and PPM constructs alone, we implanted constructs onto the vagina in the absence of tension thereby minimizing the additional impact of deformation resulting from tension. The most important findings of this study were that despite the marked increase in weight, PCU EMs outperformed polypropylene mesh as evidenced by improved tissue integration, preservation of vaginal smooth muscle morphology, and a significantly reduced host response independent of how the material was distributed across the PCU devices. Additionally, the overall cellular (i.e., cell area within a fibroma/fiber area) and macrophage responses were significantly reduced to the PCU EMs as compared to the PPMs. In these study conditions, in which multiple contiguous EM subunits were implanted onto the rabbit vagina, all EM constructs negatively impacted vaginal contractility and sGAG content similar to PPM constructs. However, the PCU EMs did not negatively impact collagen content. Collectively, these results suggest that PCU is well tolerated by the vagina with an overall improved host response as compared to polypropylene.

The finding that PCU is well tolerated by the vagina and was better tolerated than polypropylene is a significant advancement for the field of urogynecology and pelvic reconstructive surgery, for other fields in which mesh is used in soft tissue applications, and for patients and surgeons seeking alternatives to polypropylene. Polypropylene has been the predominant polymer used in prolapse repair and incontinence surgeries for decades. Emerging evidence based on mechanistic data demonstrates that softer polymers like PCU may be more suitable for use in vaginal and urethral reconstruction than rigid polymers. PCU is already utilized in a diverse array of biomedical applications ranging from cardiac catheters and pacemaker leads to orthopedic applications (e.g., meniscus implant) [36–41] which will help to facilitate translation into other surgical subspecialties. Unlike most soft polymers, PCU has a unique toughness that makes it more resistant to fatigue fracture than conventional compliant materials[42], attributed to its viscoelastic nature and extensive hydrogen bonding between the polymer’s linear chains which dissipates energy by breaking and reforming as the chains move past one another in response to applied loading[43]. This makes it ideal for use in a device to repair POP that will be subject to continuous loading and unloading. As a thermoplastic, PCU can be 3D printed through processes such as fused deposition modeling as used in this study or selective laser sintering affording direct fabrication of complex geometries[25, 44]. This may be particularly relevant to the manufacturing and design of devices to repair vaginal prolapse which are subjected to much more complex loading conditions.

Previous studies on polypropylene meshes provided good evidence that weight was a critical parameter in downstream outcomes. A large prospective cohort study of women undergoing sacrocolpopexy comparing women who had been implanted with a heavier 44 g/m² polypropylene mesh with those implanted with a lighter 28 g/m² mesh demonstrated that the heavier devices were associated with worse long-term outcomes[45]. Specifically, at ~8 years follow up women implanted with heavier meshes had more mesh related complications that were more frequently symptomatic and required more frequent intervention than women implanted with the lighter weight mesh. Similarly, studies on the same meshes utilized in the current study in a nonhuman primate sacrocolpopexy model showed that LWPP (e.g., Restorelle 19 g/m²) was associated with improved outcomes that included passive mechanics, contractility, and tissue remodeling as compared to MWPP (Gynemesh 42 g/m²)[10, 11, 13, 14]. Here, we further corroborate the previous findings in a nonhuman primate model as we found that MWPP was associated with a heightened cellular and macrophage response and had a more negative impact on vaginal structure and function as compared to LWPP. The current study, however, provides compelling evidence that the stiffness of the polymer from which a device is manufactured has more of an impact on the host response than the weight of the device.

If a softer polymer such as PCU is used, a broad range of solutions can be manufactured with substantially more material to distribute across the device (i.e., varied fiber width and device thickness) and with varied polymer stiffness. This means that devices like PCU EMs can be tailored to match a patient’s needs such as stage of prolapse, BMI, and amount of physical activity. PCU devices used in this study were heavy (EM 1 = 130 g/m² and EM 2 = 169 g/m²) and had an increased mesh burden relative to polypropylene and yet, because they were manufactured from a substantially softer polymer, had an overall improved host response as compared to polypropylene. Evidence for this included improved tissue integration, absence of stress shielding, and a reduced inflammatory cellular response. Our finding that the macrophage response to the EMs was lower than that to even the polypropylene mesh with the least amount of material on the market, suggests that we have yet to define a “tipping point” or maximum amount of material allowable in a PCU EM before negatively impacting the host response.

In this study, we used a pan macrophage marker as a measure of a key cell in the foreign body response and found a reduced response of these cells to PCU as compared to polypropylene. We appreciate that the M1/M2 paradigm plays a key role in the host response to biomaterials. As one of the earliest immune cells to encounter antigens at the site of a surgical bed, the response of macrophages is critical to the remainder of the immune response. Macrophages are highly plastic and can be polarized into a variety of context specific phenotypes that exist along a continuum. M1 macrophages, at one end of the spectrum, are highly proinflammatory and involved in clearance of surgical debris and foreign material from the wound. M2 macrophages are at the opposite end and are associated with promotion of an anti-inflammatory constructive remodeling response[46]. Our previous studies in women have demonstrated that polypropylene meshes elicit a highly predominant M1 macrophage response with few if any M2 cells present[47]. While we appreciate the importance of macrophage phenotyping in assessing the host response to PCU, the lack of antibodies to M1 and M2 cell surface markers specific for the rabbit limited our ability to phenotype macrophages in this study. Thus, here we used the overall macrophage response as an indicator of the magnitude of the foreign body reaction. Additionally, given the limited number of cytokine/chemokines in the rabbit, future studies will use whole transcriptomic analyses in the rabbit or other animal models with improved antibody availability like the nonhuman primate to characterize the M1 to M2 macrophage response to PCU EMs.

Vaginal contractility plays a central role in maintaining vaginal tone and function during sexual intercourse and childbirth. While evidence is still emerging regarding the precise mechanisms involved in mediating vaginal contractility and the changes that occur over the lifespan, definitive evidence has established that polypropylene prolapse meshes negatively impact vaginal contractility[10, 11, 14]. Expanding on this evidence, here we found that PCU EMs also negatively impacted vaginal contractility in the rabbit model. Possible mechanisms include a direct or indirect toxic (nerve or receptor mediated) effect on vaginal smooth muscle. Evidence supporting this would be the change in smooth muscle architecture observed after both polypropylene and the EMs were implanted. A more likely explanation for the PCU EMs, is that the experimental setup and weight imposed by the presence of the heavyweight EMs “artificially reduces” the contractile force of the smooth muscle myocytes to a stimulus like KCl. For the polypropylene meshes, the matrix “stiffening” imposed by the polypropylene constructs may have hindered contractility. It is important to note that the rabbit vagina has more smooth muscle than the human vagina and may be more susceptible to this stiffening effect. Additionally, the rabbit vagina is thin and markedly less stiff than the human vagina. For these reasons, additional follow up studies are needed in which vaginal contractility is assessed in a more relevant preclinical model where the stiffness of the EMs and the vagina are truly matched, like the nonhuman primate.

Strengths of this study include use of a device that is designed to match the material properties of the human vagina. In addition, we used the rabbit as our animal model, that afforded implantation of multiple iterations of the device in a single animal that reduces experimental variability and is less resource expensive than a nonhuman primate model. A limitation is that the precise number of pregnancies and deliveries of the rabbits is not known. However, all animals in this study were retired breeders, purchased from the same company, with a median parity of 15. We used multiparous animals as 1) POP predominantly impacts parous individuals and 2) previous studies have demonstrated (across multiple species) that parity significantly impacts the vagina compared to a nulliparous vagina[17, 48–50]. In this study, none of the rabbits demonstrated signs of pelvic organ prolapse and the surgical implantation methods utilized were not intened to mimic POP repairs. Instead, all membranes and meshes were implanted in the absence of tension affording us the ability to assess the host response to the material alone. Studies are ongoing to assess the host response to full-sized PCU membranes (12 cm × 3 cm) implanted on tension. Lastly, very subtle differences were observed between low and high stiffness PCU and between the different fiber width and thickness configurations. It is possible that the small size of the constructs and the absence of tension essentially negated any differences that existed, that the differences between the PCU EMs were not drastic enough to be observed, or indeed there are no differences. It is anticipated that full-length studies of the PCU EMs will elicit a clearer picture regarding the impact of polymer stiffness and the distribution of PCU within the device fiber width and thickness on the host response.

5. Conclusion

In summary, this study examined the impact of elastomeric membranes developed from the soft polymer, PCU, on ECM organization and components, tissue architecture, and inflammation. Overall, the PCU EMs were associated with a more favorable host response as compared to polypropylene meshes. These results demonstrate the potential of PCU being used as a promising future material for POP repair and incontinence devices.

Data Availability

The raw/processed data required to reproduce these findings cannot be shared at this time as the data also forms part of an ongoing study.