Characterization of Vaginal Immune Response to a Polypropylene Mesh: Diabetic vs. Normoglycemic Conditions

Abstract

Objective:

Urogynecology meshes, typically manufactured from polypropylene, are widely used in the surgical treatment of stress urinary incontinence and pelvic organ prolapse. However, mesh-associated complications such as mesh exposure can develop in women undergoing mesh implantations, for which diabetes is an independent risk factor. We aimed to define the impact of diabetes on the vaginal immune response to mesh by comparing diabetic vs. normoglycemic conditions longitudinally in a rat sacrocolpopexy model.

Methods:

Diabetes (blood glucose ≥ 300 mg/dL) was induced in middle-aged female Wistar rats with streptozotocin (STZ). A polypropylene mesh was implanted on the vagina via lumbo-sacrocolpopexy following bilateral ovariectomy and supracervical hysterectomy for 3-, 7- and 42-days. Sham-operated controls underwent the same procedures without mesh. Mesh-associated inflammation, immune cell populations and cytokine/chemokine profiles were examined in the excised vaginal tissues.

Results:

Diabetes was reliably induced starting on the 3^rd day following STZ injection. Under both normoglycemic and diabetic conditions, mesh caused a prolonged inflammatory response in the vagina with increased proinflammatory chemokines MCP-1 and MIP-1α as compared to Sham. Major differences between the two conditions were found at the later stage (42 days post-surgery), including an increased inflammation with larger foreign body granuloma and more giant cells at the mesh-tissue interface, increased fraction of macrophages in the immune cell population, and higher proinflammatory chemokine IP-10 in the diabetic group.

Conclusion:

Polypropylene mesh implanted on the vagina induces prolonged inflammation at mesh-tissue interface. Diabetes increases the mesh-associated inflammation in the long term, which is related to a dysregulated macrophage response.

Graphical Abstract

Introduction

Urogynecologic meshes, typically manufactured from polypropylene, have been used to augment surgical repairs of stress urinary incontinence (SUI) and pelvic organ prolapse (POP), which are common pelvic floor disorders occurring in aging women. For SUI, implantation of mid-urethral slings is the choice of surgery while for POP, prolapse meshes have been used to overcome the high failure rate of native tissue repair (~ 70% at 6 years) [1, 2]. Despite successful anatomical outcomes, mesh related complications, most commonly exposure of mesh through vaginal epithelium and pelvic pain, occur in ~2% of women receiving slings and 2–19% of women receiving POP meshes [3–7]. While the development of mesh complications can be attributed to multiple factors such as mesh properties, surgical procedures and surgeon experience, diabetes mellitus has been identified as an independent risk factor [5, 8–12].

Diabetes is a pandemic affecting over 1 in 10 adult women in the US. The prevalence is higher in aging women [13, 14] with ~87% cases diagnosed in women aged 45 years or older (National Diabetes Statistic Report 2017). Mesh-augmented surgical repairs are frequently performed in women with diabetes in this aging group to treat SUI and POP. It is estimated that women with diabetes comprise about 14% of the ~350,000 women receiving meshes annually in the US [8–12, 15–17]. Unfortunately, the incidence of mesh complications increases significantly in this cohort. Specifically, mesh exposure is developed in 8 – 11% of women with diabetes receiving slings [8, 11, 12] and 10 – 32% of those receiving prolapse meshes [5, 9, 10, 15–17], which is about 5 times higher than the rates in women without diabetes[3, 4]. In addition, mesh-associated pain is more severe in women with diabetes, and often not relieved by mesh removal [18]. These complications negatively impact women’s quality of life and significantly add to healthcare costs. To date, the mechanism by which diabetes increases the risk of mesh complications is unclear. As a result, no preventive or therapeutic treatment is available to improve the mesh outcomes in this population.

Typical of any implants, a urogynecologic mesh induces a robust foreign body response at the mesh-tissue interface. Ideally, the proinflammatory response initiated at the early stage of implantation retreats over time and is gradually replaced by tissue remodeling with minimal signs of immune activity. Under adverse circumstances, the proinflammatory response persists, resulting in chronic inflammation and triggering a maladaptive remodeling response characterized by tissue degradation and fibrosis, which increases the risk of complications and implant removal [19]. Evidence suggests that diabetes compromises the function of immune cells, leading to sustained inflammation in wounds and contributing to diabetes-associated chronic diseases [20–23]. Patients with diabetes also experience higher-than-normal rates of implant-related complications [24]. While the mechanism is attributable to an altered foreign body reaction [25], findings have been inconstant due to the variations of implant type, material, and implantation site. To date, most research has focused on dental and orthopedic implants which are different from soft tissue implants, and subcutaneous and percutaneous devices in which the implantation environment is physiologically and mechanically different from vagina. Thus, to understand the mechanism accounting for the increased risk of mesh complications in women with diabetes and improve the outcomes of mesh implantation in this population, it is imperative to study the vaginal host response to mesh under the diabetic condition.

In this study, we aimed to define the impact of diabetes on the vaginal immune response to mesh by comparing diabetic vs. normoglycemic conditions longitudinally in a rat sacrocolpopexy model. This model simulated mesh placement on the vagina through an abdominal incision with supracervical hysterectomy in clinical practice. To mimic post-menopausal status in women who receive mesh implantations, equalize estrogen level, and minimize impact of estrogen variability, we performed bilateral ovariectomy (OVX) at the time of mesh sacrocolpopexy. Streptozotocin (STZ) was used to selectively destruct pancreatic β cells, thereby inducing diabetes with hyperglycemia as the major attribute [26–29]. We hypothesized that diabetes dysregulates the vaginal immune response to mesh, leading to an increased and prolonged inflammation at the mesh-tissue interface. To capture the dynamic change of the response overtime, we examined the outcomes at very early (3 days), early (7 days) and late stages (42 days) following the mesh implantation. The vaginal immune response to mesh was characterized by semi-quantitative histomorphology, differentiation of immune cell populations and cytokine/chemokine profiling.

Materials and Methods

Animals

All animal studies were approved by the University of Pittsburgh Institutional Animal Care and Use Committee (IACUC 18113612). Animals were housed at the animal facility in the Magee-Womens Research Institute under conditions of 12-hour light/dark cycles, 72°F and food/water ad libitum. All investigations conformed to the regulations of the National Institute of Health. Female Wistar rats (9 – 12 months, 300 – 500 g) purchased from the Hilltop Lab Animals (Scottdale, PA) were used for diabetes induction (n=43) or as normoglycemic groups (n=43). Animals were further divided into groups with or without mesh implantation (Sham-operated groups) (Figure 1).

Figure 1.

Animal distribution in different groups.

Induction of diabetes

Diabetes was induced by a single dose streptozotocin (STZ) injection via tail vein (45 mg/kg of body weight) without fasting. In a pilot study, we tested the dosage of STZ ranging from 40 – 65 mg/kg body weight and found that STZ at 45 ± 5mg/kg led to the least death loss in this aging strain of female rats. In this study, STZ at 45mg/kg was used to induce reliable diabetic status, defined as blood glucose level ≥ 300mg/dL. Blood glucose was measured by sampling ~5 μl of blood from tail tip before and 72 hours after STZ injection with a glucometer (Accutrend GCT, Roche). Blood glucose was then monitored every 3 – 4 days. Rats with confirmed diabetes for 2 weeks were used for further studies.

Mesh

An ultralight (20.9g/m²) large pore (~1.8mm) polypropylene mesh (Restorelle, Coloplast, Minneapolis, MN) was used (Figure 2A). This ultralight mesh features high porosity (78%) and low stiffness (0.16 ± 0.07 N/mm) [30], which were associated with favorable host response.[31] Thus, it is appropriate to use for a comparison of host response between physiological and pathological conditions. To minimize any adverse impact of mesh tensioning and pore collapse/deformation incurred by mechanical loading of the mesh following implantation, we implanted Restorelle flat on the vagina with its square pores oriented along the longitudinal axis of vagina with minimal tensioning.

Figure 2.

A) Gross morphology of a polypropylene mesh (Restorelle) used in this study. B) Schematic drawings demonstrating anatomy of rat reproductive organs and procedures of mesh implantation on the rat vagina following bilateral ovariectomy and supracervical hysterectomy. C) Restorelle (width ~0.7 cm × length ~5.0 cm) was implanted upon the anterior (left) and posterior vagina (right) of rat via lumbo-sacrocolpopexy. D) Representative images of mesh-grafted posterior vaginas at 42 days post-mesh implantation in normoglycemic and diabetic rats.

Surgical procedures

Restorelle was implanted on the anterior and posterior vagina via a modified sacrocolpopexy. Briefly, the animals were anesthetized with isoflurane. A lower abdominal incision at mid-line was made followed by bilateral ovariectomy (OVX) and supracervical hysterectomy. The vagina was bluntly dissected from bladder anteriorly and rectum posteriorly. Mesh strips (~0.7 cm in width and ~5 cm in length) were attached to the anterior and posterior vagina wall with 6–0 Polysorb (Ethicon) sutures (Figure 2B and 2C). The mesh stems were fixed to ligamentous tissue lateral to sacro-lumbar vertebrae without tensioning. The abdominal fascia and skin were subsequently closed, and skin sutures were removed at 7 days post-surgery. The animals were euthanized and vaginas with/without meshes were collected for analysis at 3-, 7- and 42-days post-surgery.

Histomorphology

Cross-sectional tissue cryosections (perpendicular to longitudinal axis of vagina) at 7μm thickness were fixed and stained following procedures for hematoxylin & eosin (H&E) and Masson’s trichrome staining (Millipore-Sigma, St Louis, MO). Large images for the whole tissue sections were taken using a 90i NIKON microscope at magnification of 100. Images were analyzed using NIS-Elements imaging software (Nikon, Melville, NY). Specifically, mesh fibers were identified at the vaginal adventitia. Mesh load was defined as the area of single mesh fiber or the sum area of fiber clusters (mesh knots). Mesh-associated inflammation was quantified with the following inflammatory indices: 1) overall inflammation (OI), defined as area of inflammatory foreign body granuloma plus fibrous capsule minus mesh area; 2) area of granuloma, defined as granulomatous area minus mesh area; 3) number of foreign body giant cells (FBGCs), defined as large multi-nucleated cells at mesh-tissue interface (Figure 3A) [32, 33]. In the circumstances that the mesh fibers were cut obliquely (oblong shape), miniferets (the smallest diameter) were used to calculate the cross-sectional areas of mesh fibers and inflammation.

Figure 3.

A) Quantification of overall inflammation and foreign body giant cells (arrows) in H&E staining images, and granuloma in trichrome staining images. Images were taken from the same sample at 10 × magnification. B) Positive correlations between inflammatory indices and mesh load in both normoglycemic and diabetic groups at 7- and 42-days post mesh implantation with Pearson’s r and regression lines. * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001.

Cell isolation and flow cytometry

Cells were isolated from mesh-grafted vaginal tissues with enzymic digestion and prepared for flow cytometry to differentiate and quantify immune cell populations [34, 35]. Specifically, tissue was minced and digested in enzyme solution composed of Liberase^™ DL (Millipore-Sigma) at 1:30 and DNAase I (Millipore-Sigma) at 1:100 in a 37°C iso-temperature incubator for 30 minutes with shaking. A single cell suspension was prepared by sequentially filtering through 70μm and 50μm cell strainers. Following blocking with mouse serum and anti-CD32 (clone 93), a cocktail of anti-rat antibodies identifying surface markers on different types of immune cells including CD45 (pan-marker of immune cells), CD3 (T cells), CD45R (B cells), CD161a (NK cells) and macrophage markers (CD11b, CD86, CD163, CD172a), and isotype antibodies were applied at 1 μg per 10⁶ cells (Table 1). Cell populations including CD45+ immune cells, lymphocytes (T, B and NK cells), and macrophages were identified and quantified with a gating strategy sequentially differentiating fluorescence signals at different emission spectrum (Figure 4A) [35].

Table 1:

Antibodies used in the flow cytometry. The antibodies were all generated in mouse and had reactivity to the specific antigens in rat.

Application	Antigens	Conjugates	Source
Blocking	CD32	none	BD
Pan-immune cells	CD45	V450	BD
Lymphocytes	CD3	APC	BD
	CD45R	APC	BD
	CD161a	APC	BD
Macrophages	CD11b	FITC	BioRad
	CD86	FITC	BioRad
	CD163	FITC	BioRad
	CD172a	FITC	BioRad
Controls	Isotype IgG	V450	BD
	Isotype IgG	APC	BD
	Isotype IgG	FITC	BioRad

BD: BD Biosciences, Franklin Lakes, NJ; BioRad: Bio-Rad Laboratories, Hercules, CA

Figure 4.

A) A gating strategy used in the flow cytometry to identify immune cell populations. Cells were isolated following the steps: (1) single cells with SSC and FSC; (2) viable cells with propidium iodide negative labeling; (3) CD45+ immune cells with V450 positive labeling; (4) lymphocytes with APC positive labeling; (5) macrophages with FITC positive labeling and low SSC-A, and granulocytes with FITC positive labeling and high SSC-A in the APC negative population. (B) Representative histograms for CD45+ immune cells, lymphocytes, macrophages, and granulocytes in the normoglycemic and diabetic groups at 3-, 7- and 42- days post-mesh implantation. Quantitative data were shown by box-whiskers plots with medians, minimal and maximal values. Points represent individual samples. * p<0.05, ** p<0.01, *** p<0.001.

Profile of cytokines and chemokines in the vagina with Luminex multiplex assays

Frozen tissue was ground, and protein was extracted in 50mM Tris buffer (pH 7.4 – 7.5) containing 150mM NaCl, 1% Triton X-100, 10mM EDTA, 0.1% SDS, 0.5% sodium deoxycholate with Halt^™ proteinase inhibitor cocktail (Thermo Fisher, Waltham, MA). Protein concentration was determined in duplicate (DC Protein Assay, Bio-Rad Laboratories, Hercules, CA). A multiplex Luminex assay kit (Rat Th1/Th2 Cytokine & Chemokine 22-plex, ThermoFisher) was then used to determine the contents of cytokines including G-CSF/CSF-3, GM-CSF, IFNϒ, IL-1α, IL-1β, IL-2, IL-4, IL-5, IL-6, IL-10, IL-12p70, IL-13, IL-17A/CTLA-8, TNFα, and chemokines including Eotaxin/CCL11, GROα/CXCL1, IP-10/CXCL10, MCP-1/CCL2, MCP-3/CCL7, MIP-1α/CCL3, MIP-2, and RANTES/CCL5. Procedures were followed as instructed by the manufacture and 40μg of protein from each extract was used. Results were presented as pg/μg total protein. This panel was chosen as it determines the major cytokines and chemokines involved in innate and adaptive immune responses in a foreign body reaction.

Power analysis, animal use and statistical analysis

Based on power analysis using data obtained from literature comparing cytokine production in diabetic vs. normoglycemic wounds, at least 4 animals in each group will be needed to achieve 80% power to detect meaningful differences with p < 0.05 [36]. The distribution of rats in experimental and sham-operated control groups and the use of rats for different time points are shown in Figure 1. The number of animals may not be evenly distributed across the groups due to loss-to-death from diabetes and surgeries.

Data distribution patterns were determined by Shapiro-Wilk tests. Quantitative measurements were expressed as mean ± standard deviation or median (1^st quartile, 3^rd quartile). Student t-test or Mann-Whitney U or Kruskal-Wallis test with pairwise comparisons was used. The impact of diabetes on mesh-associated inflammation was analyzed with mixed-effects generalized linear models (GLM) considering the involvement of multiple variables (mesh load and time points). The associations between the mesh load and inflammation were determined by Pearson’s r. The infection rates were compared between groups with Chi-square test. Statistical difference was accepted as p<0.05. SPSS software (14.0 student version for windows, Inc. SPSS) was used.

Results

Establishment of the diabetic rat model with mesh implanted on the vagina

The average blood glucose in normoglycemic rats was 182 ± 37mg/dL. Hyperglycemia was reliably induced in rats starting on the 3^rd day following STZ injection and lasted until the end of this study (8 weeks in total). The level of glucose was increased by ~ 2.5-fold at 72 hours following STZ induction (p<0.001), stabilized at 7 days ranging from 300 – 750 mg/dL (Figure 5A). The rats developed polydipsia and polyuria. Body weights of rats gradually decreased by ~15% at 8 weeks after the STZ induction when compared to the starting point (all p<0.001, Figure 5B).

Figure 5.

A) Blood glucose levels and B) body weight changes relative to the original weights following the injection of streptozotocin (STZ). Surgeries were perfomed 2 weeks after the STZ injection. Data were presented by mean ± standard deviation. * p<0.001.

In the diabetic groups, the rate of lose-to-death was 8.3% (5 rats) following STZ injection and 8.3% (5 rats) following surgery. The deaths following STZ injection might be related to individual sensitivity to the drug while the deaths following surgery might be caused by respiratory failure during anesthesia (1 rat), abdominal internal bleeding (2 rats), intestine volvulus (1 rat), or bladder injury (1 rat). In the normoglycemic groups, the rate was 4% (2 rats) following surgery, both caused by self-opening of abdominal incisions. Abdominal incisions were completely healed at 7-days post-surgery without differences between diabetic and normoglycemic rats. However, diabetic rats incurred higher frequency of incisional infection at abdominal fascia and/or skin relative to normoglycemic rats (14.5% vs. 1.9%, p<0.001). In addition, abscess was found on the cervical stump of some diabetic rats (5.5%) but not normoglycemic rats. These observations correlate with literature showing higher risk of infections in patients with diabetes [37–39]. These rats were not included in subsequent analyses to avoid interference from infection-related changes in inflammatory response.

Gross morphology

We did not observe any measurable changes in mesh size and morphology at different time points after surgery. Inflammation was present in all mesh-implanted groups at all time points while no obvious inflammation was observed in the Sham groups (Figure 6). At 3-day post-implantation, the mesh was easily detached from the vaginal wall in both normoglycemic and diabetic groups. Histomorphologically, the mesh-associated inflammation appeared diffused. At 7- and 42-days, the mesh was grossly incorporated into the vagina in both groups with discernible contour (Figure 2D, showing the gross morphology of mesh-grafted posterior vagina at 42 days). Typical mesh-associated inflammation was present at the mesh-tissue interface with the formation of encapsulated foreign body granuloma around each mesh fiber and FBGCs. In the normoglycemic group, the granuloma at 7 days was prominent surrounding the mesh fiber, and the fibrotic capsule was thin with some spindle-shaped cells scattered along fibrotic bands. Relative to 7 days, the granuloma in the normoglycemic group at 42 days became smaller with the cellular components typically present on one side of the mesh fiber, and the fibrotic capsule appeared thicker with more aligned spindle-shaped cells. No differences were observed between the normoglycemic and diabetic groups at 7 days. However, at 42 days, the diabetic group demonstrated larger granulomas with poorly defined capsule (Figure 6).

Figure 6.

Hematoxylin & eosin images of cross-sectional vaginal wall with (mesh) or without mesh implantation (Sham) in normoglycemic and diabetic rats at 3-, 7- and 42-days post-surgery. Images were taken at 10 × magnification. Arrows indicate the mesh fiber-associated inflammation, showing larger granuloma and poorly defined fibrotic capsule in the diabetic vs. normoglycemic groups at 42 days. For each image from top to bottom, E = epithelium, S = sub-epithelium, M = muscularis, A = adventitia, * = mesh fiber.

Quantification of mesh-associated inflammation under normoglycemic or diabetic conditions

The confined mesh-associated inflammation at 7- and 42-days post-implantation allowed us to perform a quantitative analysis. To compare the inflammation between glucose conditions as well as time points, we first investigated the association between mesh load and the scale of inflammation. Under both normoglycemic and diabetic conditions, the inflammation indices including OI, granuloma and FBGC counts were all positively correlated to mesh load independently (all p < 0.001) at 7- and 42-days post-implantation (Figure 3B).

Next, we adjusted the impact of mesh load using mixed-effect GLM analysis to compare the inflammation between groups. When evaluating the time course of inflammation by comparing the results at 42 days to 7 days, we found that the extent of OI and granuloma were significantly decreased with time in the normoglycemic (p=0.005, 0.006) but not the diabetic group (p=0.62, 0.56). The number of FBGCs, in contrast, was not different between the two time points in the normoglycemic group (p=0.22) but increased with time in the diabetic group (p=0.026).

When evaluating the effect of hyperglycemic condition on the inflammation by comparing diabetic to normoglycemic groups, we found that the size of granuloma and the count of FBGC were significantly higher in the former at the later (42 days, p=0.015, 0.012, respectively) but not the early stage (7 days, p=0.98, 0.27, respectively) post-implantation. While the OI did not differ significantly between the groups at both time points (at 7 days, p=0.24; at 42 days, p=0.14, respectively), it had a tendance of increase at 42 days in the diabetic group (Figure 3B).

Immune cell response to mesh under normoglycemic or hyperglycemic conditions

In our pilot experiments, the CD45+ immune cells isolated from the Sham tissues were in small amount (<3% of viable cells), which is consistent with the result of a previous study investigating hernia mesh [40]. It is likely that the sham surgery (blunt separation of vagina from bladder and rectum, etc.) did not induce a significant increase of immune cells in the vagina at either short or longer terms. Therefore, only mesh-implanted groups were included in this experiment. In the normoglycemic rats, the median % of CD45+ immune cells in total viable cells isolated from the mesh-implanted vagina peaked at 7-day post-implantation (53%) and decreased at 42 days (25%), which remained higher than 3 days (11%) (all p<0.05, Figure 4B), demonstrating a long-lasting immune response to mesh. A similar pattern was observed in the diabetic rats with the median % of CD45+ cells being 7%, 36% and 23% at 3-, 7- and 42-days albeit the significance was not found between 7 and 42 days (p=0.10). There were no differences between the normoglycemic and diabetic groups except that the value was slightly lower in the latter at 7 days (p=0.042).

Major immune cell types involved in the foreign body response to mesh including macrophages, lymphocytes and granulocytes were identified in the CD45+ cell population (Figure 4B). In the normoglycemic groups, the fraction of macrophages tended to decrease with time (median % being 60%, 42% and 34% at 3-, 7- and 42-days, respectively; p=0.032 between 3- and 42-days, p=0.06 between 7- and 42-days) while the fraction of lymphocytes tended to increase (median % being 12%, 31%, and 44% at 3-, 7- and 42-days, respectively; p=0.031 between 3- and 7-days, p=0.016 between 3- and 42-days, p=0.05 between 7- and 42-days). The fraction of granulocytes remained relatively stable over time (median % being 14%, 15%, and 10% at 3-, 7- and 42-days, respectively; all p>0.05).

In contrast, in the diabetic groups, the fractions of macrophages and lymphocytes both peaked at 7 days and remained at parallel levels at 42 days. For the macrophages, the median % values were 32%, 57% and 49% (p=0.018 between 3- and 7- days) while those for the lymphocytes were 8%, 26% and 29% (p=0.018 between 3- and 7-days, p=0.030 between 3- and 42-days) at 3-, 7- and 42-days, respectively. Meanwhile, the fraction of granulocytes peaked at 7 days and decreased at 42 days (median % being 10%, 14% and 8%, p=0.012 between 3- and 7-days, p=0.003 between 7-and 42-days). When comparing the diabetic to the normoglycemic groups, we found that the fractions of these immune cells were not significantly different except that the % of macrophages was increased in the former by 44% at 42 days (p=0.010).

Cytokine/chemokine profiles in the mesh-implanted vagina under normoglycemic or diabetic conditions

Among the 22 cytokines and chemokines, 12 were detected in the mesh-grafted or Sham vaginal tissues, including G-CSF, IFNϒ, IL-1α, IL-1β, GROα, MIP-2, IP-10, MCP-1, MIP-1α, RANTES, MCP-3, and eotaxin. Relative to the Sham groups, the mesh implantation induced a significant increase of IFNϒ at the very early stage (3 days), and increases of eotaxin, IP-10, MCP-1, and MCP-3 at the early stage (7 days) in both normoglycemic and diabetic groups (all p<0.05). There was also an increase of MIP-1α at 7 days in the normoglycemic group. The increase of these factors was mostly diminished at the late stage (42 days) except MCP-1, which remained elevated by 8 folds in the normoglycemic (p<0.001) and 5.3 folds in the diabetic rats (p=0.036). In addition, MIP-1α was increased by 13.4 folds in the normoglycemic (p<0.001) and 15 folds in the diabetic rats (p<0.001) at the 42 days post-surgery (Figure 7A &B).

Figure 7.

A) Profiles of cytokines and chemokines in vaginal tissues with (mesh) or without (Sham) mesh implantation in normoglycemic and diabetic rats at 3-, 7- and 42-days post-surgery. Median values were used to construct the heat map. Arrows point to factors depicted in detail in B) and C). NSO = normoglycemic Sham; NMO = normoglycemic with mesh; DSO = diabetic Sham; DMO = diabetic with mesh. B) Levels of MCP-1 and MIP-1α in Sham and mesh-implanted groups in normoglycemic and diabetic rats at 3-, 7- and 42-days post-surgery. C) Level of IP-10 in mesh-implanted groups in normoglycemic and diabetic rats at 3-, 7- and 42-days post-surgery. For B) and C), quantitative data were shown by box-whiskers plots with medians, minimal and maximal values. Points represent individual samples. * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001.

When compared to the normoglycemic rats, the diabetic rats had a 1.6-fold increase of IP-10 at 42 days post-surgery (p=0.041, Figure 7C), corresponding to the increased fraction of macrophages in the diabetic rats at this time point. For the Shams, no differences were found between the diabetic and normoglycemic groups throughout the time course.

Discussion

With an in-depth and longitudinal analysis of the mesh-associated inflammation, our work characterized the vaginal immune response to a polypropylene mesh under normoglycemic and diabetic conditions, thereby providing insights into the mechanisms that account for the increased risk of mesh complications in women with diabetes. We demonstrated that mesh induced prolonged inflammation with persistent increase of CD45+ immune cells in the vagina under both glycemic conditions, which was concurrent with an increase in proinflammatory chemokines MCP-1 and MIP-1α. Diabetes increased the inflammation with bigger granuloma and more FBGCs in the longer term (42 days following mesh implantation), supporting our hypothesis that diabetes dysregulates the vaginal immune response to a mesh implant. The increased mesh-associated inflammation in the diabetic rats corresponded to an increase of macrophage population in the immune cells and an up-regulation of IP-10 at the later stage.

While larger animals such as rhesus macaque, sheep, and rabbits are preferred models in the mesh research to facilitate mesh implantation on the vagina, they are usually costly and not good models for diabetes. Rats, on the other hand, has always been used to generate a diabetes model with low death rate and pathologic outcomes similar to human [27, 29, 41]. It has been recently used as a sacrocolpopexy model to investigate host responses to biomaterials [42]. Our study further tested the feasibility of using the rat model in the research of urogynecologic meshes. In addition, we demonstrated that this model was appropriate to investigate the pathogenesis of mesh complications under diabetic condition.

In line with literature [31], our findings demonstrated positive correlations between mesh load and mesh-associated inflammation under both normoglycemic and diabetic conditions, indicating that mesh load is a critical factor defining the scope of inflammation at the mesh-tissue interface. Restorelle, which is a knitted polypropylene mesh with the lightest weight, largest pore and lowest stiffness on the market, is associated with an overall favorable host response and mesh-tissue incorporation as compared to other widely used meshes [31, 43, 44]. In this study, we minimized the impact of variability introduced by mesh tensioning, which allowed us to evaluate the effect of mesh load on the host immune response. Our results provide further evidence supporting that meshes with lighter weight (less mesh load) and stable pore geometry (less deformation-caused increase of mesh load) are beneficial in the regard of foreign body response to mesh, which in turn promotes mesh outcomes.

Under normoglycemic conditions, our findings from both histomorphology and flow cytometry showed that the vaginal inflammatory response to mesh decreased gradually but was present for a sustained period. This phenomenon was also observed in studies of hernia meshes [33, 40]. We further found that despite a temporary increase of cytokines and chemokines in IFNϒ, eotaxin, IP-10, MCP-3, MCP-1 and MIP-1α at the early stages following mesh implantation, only the levels of proinflammatory cytokines MCP-1 and MIP-1α remained high at the later stage. It is likely that the short-term increase of immune mediators signifies wound healing response due to tissue injuries incurred with mesh implantation while the longer-term increase of MCP-1 and MIP-1α characterizes immune response to mesh. These chemokines have strong chemotactic potency for monocytes and macrophages, indicating their roles in the maintenance of the prolonged mesh-associated inflammation.

Under diabetic condition, the extent of mesh-associated inflammation did not decrease over time. As a result, the inflammation relatively increased when compared to the normoglycemic rats at the later stage (42 days). Such prolonged inflammation at higher levels may result in tissue pathologies such as tissue degradation and/or fibrosis, which underly the pathogenesis of mesh complications [19]. Overall, our results suggest that diabetes has a long-term negative impact on the vaginal immune response to mesh, which may account for the diabetes-associated risk in the development of mesh complications.

Macrophage is a key innate immune cell orchestrating the progress of foreign body response. It is a main component of the granuloma and forms FBGCs at the interface of tissue and foreign bodies. Consistent with the literature [31, 32, 40, 45], we found that macrophages, along with lymphocytes, were the primary cells recruited to the mesh fibers at early and later stages. Interestingly, the fraction of macrophages in the immune cells was higher at the later stage (42 days) in the diabetic vs. normoglycemic rats, which parallels the findings of histomorphology showing unrepressed granuloma and increased FBGC formation. Research showed that macrophages have high plasticity in their energy metabolism to accommodate phenotypical and functional transitions in response to stimuli in the microenvironment [46–49]. Under hyperglycemic condition, the normal energy metabolism in macrophages can be disturbed, resulting in macrophage dysfunction. Our findings are consistent with the scenario, indicating that diabetes negatively impacts the behavior of macrophages, which may contribute to the increased mesh-associated inflammation.

IP-10 is a pro-inflammatory chemokine secreted by multiple cell types in response to IFN-γ, mainly macrophages [50, 51]. By binding to a G protein-coupled receptor (CXCR3) on cell surface, IP-10 promotes “homing” of CXCR3-positive cells, such as macrophages, dendritic cells, NK cells and activated T cells, toward inflamed tissue [52]. The increase of IP-10 in the diabetic vs. normoglycemic rats at 42 days post-implantation corresponds to the increased mesh-associated inflammation with higher fraction of macrophages in the immune cells, indicating its involvement in the diabetes-altered immune response to mesh. On the other hand, the absence of a change in IFN-γ in the diabetic rats at the longer term suggests that the increase of IP-10 might be caused by other mechanisms such as epigenetic modulations induced by hyperglycemia, which will be investigated in future studies.

Our study provides the first-hand preclinical data elucidating the mechanism accounting for the increased risk of mesh complications in women with diabetes. The longitudinal study design allowed us to demonstrate the time course of mesh-associated inflammation under both normoglycemic and diabetic conditions, which was facilitated by a cost-efficient rat model. The limitation of this study includes: 1) host response to mesh in rats might not be the same as in human. Particularly, the profile change of cytokines and chemokines under diabetic condition might be subjected to environment-sensitive regulatory mechanisms in different species. However, rat models have been widely used to test host response to materials in preclinical studies. We believe our model is feasible to provide mechanistic information for future clinical studies; 2) STZ-induced hyperglycemic model simulated type 1 diabetes while type 2 diabetes, which has other metabolic disturbance, e.g., hyperlipidemia and insulin resistance, is more prevalent in women receiving urogynecologic meshes. Yet, this model allowed us to isolate out the impact of hyperglycemia on the mesh-elicited immune response, which is the primary harmful attribute in both types of diabetes.

In summary, polypropylene mesh implanted on the vagina induces prolonged inflammation at mesh-tissue interface while diabetes escalates the inflammation in the long term, which is possibly mediated by macrophage dysregulation. As such, women with uncontrolled diabetes are prone to developing mesh complications following mesh implantation. While polypropylene mesh with ultralight weight and stable pores has prospective advantage in eliciting more favorable foreign body response, novel biomaterials with less immunogenicity are desired due to the prolonged mesh-associated inflammation. For women with diabetes, a better understanding of the mechanism underlying the dysregulated macrophage response to mesh will inform future cell-based therapies to improve the outcomes of vaginal implants in this population.

Significance-revised.

This study investigated the underlying mechanism accounting for the increased risk in women with diabetes for developing mesh complications such as mesh exposure. The significance includes: 1) It is the first study investigating vaginal host response to a prosthesis under the influence of diabetes; 2) the longitudinal study design elucidated dynamic changes of vaginal host response to mesh from early to late stages; 3) our findings may inform future mechanistic studies and studies investigating preventive/therapeutic strategies to improve the outcomes of women with diabetes receiving vaginal implants.