Impact of Clinical Characteristics on Proteomic Analysis of Mesh Complications

Objectives: We have previously presented proteomic analysis of patients undergoing vaginal polypropylene mesh excision demonstrating the importance of inflammation and epithelial and extracellular remodeling in mesh complications. In this analysis, we aimed to determine the impact of age, BMI, smoking, and systemic or topical estrogen on vaginal proteomic profiles between women with mesh exposure and mesh-related pain compared to controls. We hypothesized that many differences would remain between mesh explants and controls due to the foreign body response, that keratins would remain upregulated in exposure complications due to high epithelial turnover and pain would be associated with extracellular matrix (ECM) remodeling.

Methods: Samples were obtained by random selection from women 30-80 years old, BMI < 40 kg/m2 and parous. 10 well-integrated mesh samples, 10 mesh exposure samples, 10 mesh-pain samples and 10 control vaginal biopsies of patients without mesh underwent proteomic analysis. Samples were lysed, digested, and normalized to 50µg total protein. Nano-scale liquid chromatographic tandem mass spectrometry yielded 30,588 peptide sequence from 2,803 protein groups. One well-integrated sample and one exposure sample were excluded due to predominance of hemoglobin. Due to the collinearity of age and estrogen use, two models were considered: 1) group, body mass index (BMI), smoking, and systemic (premenopausal or oral hormone replacement therapy) or topical estrogen use. In 2) age replaced estrogen. Differentially expressed protein (adjusted p< 0.05) and gene ontology enrichment analyses were performed.

Results: Mean ages in years by groups were 51.9±6.9 mesh exposure, 49.1±13.6 pain, 62.1±12.5 well-integrated mesh, and 41.5±6.2 controls (p < 0.05 well-integrated vs control). Overall, mean BMI was 28.0±5.5 kg/m2, 24% of patients were smokers, and 76% were premenopausal or on systemic or topical estrogen (p>0.05, Figure 1A). Median duration of implantation was 4.5±5.7 years (p>0.05). Because replacing estrogen with age didn’t significantly impact the results, the final model consisted of group, BMI, smoking, and estrogen. In this model, group accounted for 1455 (18%) of differentially expressed peptides between exposure vs controls, 973 (18%) between pain vs controls and 728 (14%) between well-integrated vs controls. For exposure samples, keratinization proteins were enriched (Figure 1B 2A), while dysregulated cytoskeleton and ECM were seen in pain (Figure 1C, 2C-D). Calmodulin 3, keratin 17 and keratin 6 remained increased 8fold in exposure compared to pain (adjusted p< 0.05). Although univariate gene ontology analysis revealed “defense against bacterium” to be an important cluster in exposure vs well-integrated mesh, when controlling for known causes of inflammation, this was no longer significant.

Conclusions: In this proteomic analysis, clinical variables did not drive protein differences between mesh groups, suggesting that regardless of known risk factors, pain and exposure complications progress through epithelial remodeling, dysregulated ECM, and fibrosis in similar patterns.

Second abstract: Mesh burden drives host response in abdominal and vaginal polypropylene mesh

Objectives: We sought to compare the host response to low vs high polypropylene mesh burden after vaginal or abdominal implantation in Rhesus Macaques. We hypothesized that in both sites an increased mesh burden would result in increased malondialdehyde (MDA; a marker of oxidative damage), heightened macrophage response, and increased apoptosis, and that deformed mesh on tension would produce the strongest negative effect in the vagina due to increased burden and micromotion of mesh against the vagina.

Methods: In a nonhuman primate sacrocolpopexy model we used Restorelle, normally associated with a relatively low mesh burden and good outcomes, and intentionally and successively introduced deformations resulting in 3 mesh configurations: a. “Stable” pores to create a low mesh burden (n=10); b. “Deformed-1” with collapsed pores to create high mesh burden (n-10); c. “Deformed-2” with collapsed pores and suture-secured wrinkles to create highest mesh burden. Stable and Deformed-1 were implanted on tension; Deformed-2 was implanted with (n=10) and without tension (n=8). Vaginal sham operated animals served as controls (n=7). Gynemesh PS (n=8) and Restorelle (n=8) were implanted on the anterior abdominal wall with sham (n=17) and non-surgical (n=9) controls. Immunofluorescence on 12-week vaginal-mesh complexes with CD68 (macrophage) and TUNEL (apoptosis) was performed. MDA was determined by biochemical analysis. Kruskal Wallis with post-hoc ranksum tests and Pearson correlation tests were used as appropriate.

Results: There were no differences in cell types examined between abdominal and vaginal sham samples (p>0.05). Abdominal mesh samples showed more macrophages (0.52 [0.21-0.71] vs 0.04 [0.02-0.15] macrophages/mm2) and apoptotic cells (0.16 [0.08-0.24] vs 0.02 [0.00-0.04] cells/mm2, p< 0.001; Figure 2A). For macrophages, this was driven by a difference in abdominal Gynemesh compared to lower mesh burden groups (Stable and Deformed-1) and Deformed-2 without tension (p < 0.006 Figure 2B); there was no difference between abdominal Gynemesh and Deformed-2 on tension or between abdominal and vaginal Restorelle groups. In abdominal mesh, apoptosis was limited to areas immediately surrounding mesh, while in Deformed-2, increased apoptosis was seen in the epithelium and subeithlieum (p < 0.05). Abdominal mesh samples also had higher MDA than vaginal samples (0.054 [0.030-0.057] vs 0.002 [0.001-0.005] nmol/mg tissue, p< 0.001).

Regardless of implantation site, macrophages (rho 0.64 p< 0.001), apoptotic indices (rho 0.52, p< 0.001) and MDA levels (rho 0.37 p< 0.01) were strongly correlated with mesh burden (Figure 2C-E) with a linear 7.8% increase in macrophages for every 10% increase in percent fiber area (p < 0.001). In multivariable analysis, percent fiber area and abdominal Gynemesh were independent predictors of percent macrophages (p < 0.05).

Conclusions: Mesh burden is a main driver in both abdominal and vaginal mesh innate immune response and Restorelle on tension may behave more like Gynemesh if deformed. Apoptosis in the subepithelium likely indicates stress-shielding and plays a role in exposure.

Figure 1:
Vaginal Implantation Groups

Figure 2:
Comparison of abdominal and vaginal mesh and correlations with mesh burden