Tissue remodeling macrophages morphologically dominate at the interface of polypropylene surgical meshes in the human abdomen

Currently, textile mesh structures made of polypropylene are widely used to reinforce a hernia repair in the abdominal wall. The tissue reaction to the implants leads to the formation of a foreign body granuloma. Here, the mesh fibers are surrounded by dense layers of inflammatory cells with accompanying fibrotic encapsulation that shields the mesh from the surrounding tissue. After early recruitment of neutrophils, macrophages become the predominant cell mediator and are suspected to regulate the subsequent scar formation [1].

In 1992, Stein et al. demonstrated that IL-4 is an alternative activator for macrophages [2]. His findings later led to the separation of activated macrophages into two main populations, designated classically activated (M1) and alternatively activated (M2) [3]. This nomenclature is too simplistic and does not accurately describe the heterogeneity of macrophage phenotype in complex environments in situ; however, it remains useful for clinical assessment. Typically, in humans, ‘M1’ macrophages express the cell surface marker CD86 and produce high levels of proinflammatory cytokines. A prolonged or persistent M1 response can damage the tissue. In contrast, ‘M2’ macrophages encompass several different macrophage populations, which have overlapping marker expression and are relatively anti-inflammatory [4]. Though, two key markers are the mannose receptor CD206 and the scavenger receptor CD163. These cells produce lower amounts of proinflammatory cytokines and are considered to play a primary physiological role in promoting constructive healing and tissue remodeling [5]; hence, being termed by some as wound healing macrophages. However, it is recognized that a long-lasting strong ‘M2’ response can lead to excessive collagen formation, resulting in fibrosis [1, 6].

Experimental studies in animals with meshes have demonstrated that at least within 12 weeks after implantation, mainly M1 (CD86⁺) macrophages surrounded the mesh fibers and formed the foreign body granuloma [7, 8]. The predominance of these cells was confirmed by Nolfi et al. at human mesh explants from the pelvic floor, which were explanted because of inflammatory or fibrotic complications [9].

However, due to several conflicting studies, it still is not clear whether adequate fibrous scar formation with collagen deposition depends mainly on the presence of M1 or M2 macrophages [10]. In a mouse model, a shift from the M1 towards M2 phenotype led to a reduced fibrotic response [11], whereas in another mouse model the M2 phenotype was responsible for the development of fibrosis in the lung, likely by promoting myofibroblast differentiation of mesenchymal stem cells [12]. This highlights the need for further studies to determine the role of macrophages in both collagen deposition and degradation.

Therefore, we investigated whether M1 or M2 macrophages are related to local deposition of collagen-I and -III at seven polypropylene (PP) meshes, which were explanted from the abdominal wall due to recurrences. We examined the foreign body reaction to the meshes using cell densities, defined as cells per mm², and the percentages of macrophage populations in six regional zones with increasing distance from the mesh up to 350 µm. In addition, we quantified the collagen deposition in each regional zone and the expression of two different matrix metalloproteinases: gelatinase A (MMP-2) and neutrophil collagenase (MMP-8) as indicators of collagen degradation.

We analyzed tissue sections of seven polypropylene (PP) monofilament meshes (used for the repair of abdominal wall hernias), which were explanted due to recurrences. After removal, samples were embedded in aqueous formalin solution to preserve the tissue before sectioning and further processing to FFPE blocks. Operative reports from the initial mesh surgery were reviewed, and the following meshes were recorded: 3 large pore Ultrapro®, 2 Ventralex® with a layer of PTFE, and 2 small pore plugs (Table 1). To optimize antibody dilutions and to check specificity of labeling, we used human spleen and tonsil tissues that did not display gross pathology as control tissues (Online Resource 1).

Prior to immunofluorescence staining, mesh samples were checked for the presence of characteristic cells and morphology by hematoxylin and eosin (Fig. 1). All specimens showed a typical, highly localized foreign body reaction around the mesh fibers with an inner infiltrate of predominantly mononuclear cells and an outer fibrotic capsule. Multinucleated giant cells were observed at the surface of some mesh fibers in all samples.

The examination of the spatial morphology of the macrophages at the foreign body granuloma around meshes by combining a custom Euclidean distance map algorithm with immunofluorescence staining of entire tissue sections of polypropylene (PP) mesh explants demonstrated that cell density and macrophage ratio decreased with distance to the PP fibers, and that these were predominantly M2 macrophages. In addition, more than 70% of the macrophages showed co-expression with collagen-I or -III and more than 50% with MMP-2, which demonstrates these cells are important participants in mesh-tissue remodeling.

When examined, the cell densities and the percentages of macrophages decreased with increasing distance from the mesh fibers. Interestingly, the proportion of M2 macrophages in relation to the total number of macrophages was not significantly different between regional zones, highlighting the consistent M2 environment. The highest percentages of both M1 and M2 macrophages in situ were observed within the first 50 µm from the mesh surface, which is in agreement with other studies on PP meshes [8, 11, 14]. On the contrary to our findings, several animal studies on PP meshes have shown mainly M1 macrophages to surround the mesh fibers [7, 8, 11]. The predominance of M1 macrophages was confirmed by Nolfi et al. at human PP mesh explants from the pelvic floor, which were explanted due to mesh exposure and chronic pain [9]. Any quantification strongly depends on marker, cutoff, the distance to the mesh fibers, and the location of the investigated region, however the difference we record here is striking (‘M1’ ≈ 6% vs ‘M2’ ≈ 60%). The mesh placement at the pelvic floor may favor proinflammatory morphology of macrophages, whereas locations within the abdominal wall tends to show less inflammation [9, 15]. In addition, the medical reason for mesh removal may be decisive for macrophage morphology. Our explants represent the chronic foreign body reaction, but they have not been removed because of acute clinically apparent inflammation or erosion. Finally, M1/M2 morphology in humans may differ to the results in animals, especially considering many of these markers are not shared and that meshes are not adapted to the physiology of animals but rather to that of humans.

The M2 response observed in this study was accompanied by abundant collagen deposition, leading to fibrosis. This is in agreement with several studies on fibrotic pathogenesis like pulmonary fibrosis or hepatic fibrosis [12, 16‐19]. Most collagen-I was found directly at the interface to the meshes where the cell density and percentage of M2 macrophages was highest, while in the more distant zones there was more collagen-III, but significantly less M2 macrophages. These observations are consistent with the knowledge that profibrotic M2 macrophages produce various mediators that directly activate fibroblasts known to be mainly located in the outer fibrotic capsule, which in turn control ECM deposition [19]. The observed co-localization of macrophages with collagen-I, especially directly at the mesh–tissue interface (> 50%), could indicate that M2 macrophages are responsible for collagen-I degradation via a mannose receptor-mediated (CD206) pathway after MMP cleavage, as discovered by Madsen et al. [20]. Alternatively or additionally, macrophages may be involved in the synthesis of collagen-I, since the synthesis of other collagen types has been demonstrated in animals and humans [21, 22].

Profibrotic M2 macrophages are also known to produce their own MMPs, which regulate inflammatory cell recruitment and ECM turnover [19]. MMP-8 was rarely observed, whereas abundant MMP-2 expression was present in all regional zones, which is consistent with previous studies that reported an up-regulated MMP-2 protein synthesis and enzymatic activity following mesh implantation [9, 23, 24]. About half of the macrophages were found to produce MMP-2. MMP-2 is a gelatinase that degrades collagens, its production by macrophages here may be an attempt to remodel the foreign body. However, the side effects may cause instability of the granuloma, as observed in atherosclerotic plaque rupture [25], and likely inhibit normal scar formation and maintain persistent cell turnover. Future studies will have to verify functionality of MMPs, e.g., by ELISA, zymography techniques, or RT-PCR gene expression, although the overall proteinase activity is altered by tissue inhibitors of metalloproteinases (TIMPs), as well, but in this study, we just aimed to demonstrate the local expression without going into the functionality discourse.

Several limitations of our study have to be considered. Our collection of explanted meshes represents a selection from patients who apparently had a recurrence of an abdominal wall hernia with different intervals after implantation in nonhealthy tissues. Furthermore, only a limited sample size was available, covering three different mesh types with different textile properties (e.g., textile structures, pore sizes). In addition to the many confounders given by the complex staining protocol, it is the location and size of the investigated region that affect the results. Therefore, we used a Euclidean distance map algorithm that created six regional zones adjacent to the mesh fibers. This allowed us to spatially quantify cells and collagen deposits. Our choice of six regional zones, which covered 350 µm from the mesh interface ensured that most of the inflammatory infiltrate around each mesh fiber was included. By relating the total number of cells or positive signal area to the total area of the regional zone, we were able to create a robust, normalized measurement. For cells, the analysis of the staining intensity only in the area of the nuclei cannot exclude positive staining based on some overlapping cytoplasm membrane, but the probability is considerably reduced by sectioning at a thickness of about 2 µm. In addition, we applied a cutoff value of 100 for the mean staining intensity of cells, as described [13]. This allowed us to determine positive cells in an objective and reliable way and provided us with the most confidence for true positive detection.

In addition, as mentioned, macrophage polarization nomenclature is under scrutiny because of overlapping markers and the broad spectrum of macrophage origin and functional distinctions in tissues [26]. Hence, even though CD86 is considered an M1-specific marker, it can be identified in the M2 macrophage subset: M2b. Therefore, although we label these cells M1 for simplicity, the reality is more complex, and these could also be M2b or other macrophages in the multidimensional spectrum of polarization. Regardless of nomenclature, our findings do not implicate M1 or M2b cells as major components of the mesh implants we examined. In addition, even though we report on two M2 macrophage groups (CD163⁺ or CD206⁺), we acknowledge that these may indicate the same cells, as M2a and M2c cells can express both markers [27], while tumor macrophage studies show approximately half of CD206⁺ macrophages expressed CD163 [28]. Although it may be useful to fully categorize these cells with MALDI proteomics approaches for example, this was beyond the scope of our study—which aimed to address which broad category of macrophage (inflamed or wound healing) was related to collagen deposition at the foreign body granuloma.

In conclusion, the findings of the present study suggest that the foreign body reaction to PP meshes is associated with an M2 response that persists even years after implantation and is related to collagen deposition and turnover. Therefore, our findings challenge the idea that a stimulated transformation of macrophages towards the M2 phenotype is a solution to ameliorate the long-term foreign body reaction [11, 14, 29]. Future studies are required to assess whether these M2 macrophages are responsible for excessive fibrosis with all its clinical side effects, and whether attenuating this response is able to reduce complications after abdominal wall hernia repair.