Comparative in vitro and in vivo assessment of three experimental extracellular matrix meshes for soft tissue remodeling Authors: Wenting Ruan, Guowei Wang, Qiujin Qu, Xiaoxiao Gai, Changbin Wang, Jia Liu, Fuyu Zhu, Zhonghua Qu, Chenghu Liu Published: 2026 ABSTRACT The development of standardized cross-species evaluation frameworks is urgently needed for absorbable extracellular matrix (ECM) meshes, due to significant variations in safety and efficacy across tissue sources. This study establishes a multi-dimensional assessment system to quantify material-specific bio-functional correlations. Three ECM meshes—human acellular dermis (HAD), porcine acellular dermis (PAD), and porcine small intestinal submucosa (SIS)—were systematically characterized using integrated in vitro and in vivo analyses. Assessments included physicochemical properties (e.g., degradation kinetics via collagenase digestion), immunogenic residues (α-Gal epitopes and DNA quantification), and biocompatibility. In vivo performance was evaluated using rabbit subcutaneous (for degradation and immune response) and rat abdominal defect models (for macrophage polarization: CD206⁺/CD86⁺ ratio and collagen deposition). HAD exhibited superior structural stability (82.3% ± 1.2% residual mass after 7-day collagenase digestion), whereas SIS degraded most rapidly (10.5% ± 0.42%; p < 0.01 vs PAD) but provoked severe foreign body reactions (2.3-fold higher vs HAD at 13 weeks, p = 0.017). Notably, HAD uniquely promoted sustained M2-dominant macrophage polarization (CD206⁺/CD86⁺ ratio = 2.1 vs SIS = 0.7), which reduced inflammation and enhanced collagen remodeling (p < 0.05 vs SIS). Serum metabolomic analysis indicated that ECM mesh implantation induced a reprogramming of host amino acid metabolism, marked by a notable acceleration in L-methionine metabolism. The proposed integrated evaluation framework successfully identifies material-specific correlations among degradation behavior, immune response, and metabolic regulation, providing critical benchmarks for the rational design of immunomodulatory ECM meshes.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-47592-6. Introduction Absorbable biomaterials play an indispensable role in modern regenerative medicine, facilitating soft tissue repair in diverse applications such as hernia1, breast reconstruction2,3 and orthopedic applications4,5. Among these, extracellular matrix (ECM) materials-derived from decellularized tissues like skin, pericardium, or intestinal mucosa of xenogeneic (e.g., porcine, bovine) or allogeneic (human) origin6–10—occupy a prominent position. Following implantation, these meshes are not inert; they undergo complex host-mediated degradation and remodeling, a process orchestrated by immune responses and influenced by the local mechanobiological environment11–14. This process is modulated by numerous factors, including the biomechanical forces and the specific biochemical composition of the ECM15. Despite their promise, the clinical performance of ECM meshes varies significantly based on species origin, tissue source, and processing methods, leading to divergent inflammatory responses and remodeling outcomes10,11. Macrophages are central arbiters of this fate, with their polarization state (pro-inflammatory M1 vs pro-healing M2) critically influencing the balance between fibrosis and regeneration15–17. Both soluble and structural constituents of the ECM can modulate this macrophage polarization, thereby steering the host response18,19. However, a systematic, comparative framework to disentangle the individual and combined contributions of these variables is lacking. To address this gap, we established a multi-dimensional assessment platform focused on the intrinsic material properties. We selected three commercially relevant ECM meshes–human acellular dermis (HAD), porcine acellular dermis (PAD), and porcine small intestinal submucosa (SIS)–which differ deliberately in species origin (human vs porcine), tissue type (dermis vs intestine), and decellularization methodology. Our primary objective was not to simulate a specific clinical application with its unique functional requirements, but to perform a controlled, comparative characterization of their intrinsic physicochemical properties, biocompatibility, and elicited immune-metabolic responses in standardized models. To establish a meaningful comparative framework within a standardized context, this study evaluates ECM meshes using in vivo models that simulate implants intended for prolonged contact with muscle and connective tissues—a key application category defined by ISO 10993–6:201620. Our approach integrates detailed in vitro characterization with these defined in vivo models to establish correlative links between intrinsic material properties and the host response. This integrated framework aims to provide a predictive, multi-parametric protocol for the preclinical evaluation of ECM-based biomaterials for soft tissue repair applications. Results Structural characterization of ECM meshes To establish a baseline understanding of the physical properties that may influence degradation and host integration, we first characterized the surface topography and ultrastructure of the ECM meshes using high-resolution microscopy and scanning electron microscopy (SEM). A synthetic composite mesh (composed of non-degradable polyester fibers integrated with a collagen membrane) was included as a non-degradable control to provide a structural benchmark. The Composite mesh exhibited a tight, compact microstructure (Fig. 1A, E). In contrast, the acellular meshes showed distinct collagen architectures. HAD (Fig. 1B, F) and PAD (Fig. 1C, G) meshes displayed a fragmented and flaky collagen arrangement. The SIS mesh (Fig. 1D, H) featured a highly porous, three-dimensional network of interwoven collagen fibers with irregular pore sizes, which is characteristic of its submucosal origin. In summary, the ECM meshes exhibited markedly different ultrastructural organizations. SIS displayed a highly porous architecture, whereas the dermal-derived meshes (HAD and PAD) showed a more compact and fragmented morphology. These structural differences may predispose them to distinct degradation and host integration profiles. Fig. 1Surface and ultrastructural analysis of ECM meshes. Surface morphology was assessed using high-resolution optical microscopy (A–D), and ultrastructure was examined by scanning electron microscopy, SEM (E–H). The Composite mesh (A, E) displays a compact, integrated structure. Human (B, F) and porcine (C, G) acellular dermis meshes show a fragmented, flaky collagen morphology. Porcine SIS (D, H) possesses a characteristic porous, interwoven collagen fiber network. Scale bars: 500 μm (E), 50 μm (F–H). Magnification times: 20x (A), 280x (B), 280x (C), 820x (D). Surface and ultrastructural analysis of ECM meshes. Surface morphology was assessed using high-resolution optical microscopy (A–D), and ultrastructure was examined by scanning electron microscopy, SEM (E–H). The Composite mesh (A, E) displays a compact, integrated structure. Human (B, F) and porcine (C, G) acellular dermis meshes show a fragmented, flaky collagen morphology. Porcine SIS (D, H) possesses a characteristic porous, interwoven collagen fiber network. Scale bars: 500 μm (E), 50 μm (F–H). Magnification times: 20x (A), 280x (B), 280x (C), 820x (D). In vitro enzymatic degradation kinetics Given the critical role of collagenous structure in ECM scaffold integrity and remodeling, we evaluated the susceptibility of the meshes to enzymatic breakdown using type I collagenase, which primarily digests fibrillar collagens (I, II, III). We acknowledge that the relative abundance of these collagens, as suggested by our proteomic data (Fig. S1), may influence the degradation rate, and that different enzymatic profiles in vivo will contribute to the overall remodeling kinetics. The synthetic composite mesh served as a non-degradable control (Fig. 2). All acellular ECM meshes underwent significant and time-dependent degradation. HAD meshes exhibited the most rapid degradation during the initial 3-day period. The most pronounced mass loss occurred between days 3 and 7, with SIS meshes degrading most extensively. By day 10, the degradation of all ECM meshes had reached a plateau. In stark contrast, the synthetic composite mesh, as anticipated, showed minimal degradation throughout the study period (p < 0.05 for HAD/SIS vs Composite at all time points), confirming its structural stability under enzymatic challenge. Fig. 2In vitro enzymatic degradation kinetics of ECM meshes. Mesh strips (2 cm × 1 cm; initial dry weight W0) were incubated in 19.2 U/mL collagenase I (PBS, pH 7.4) at 37 °C with agitation (60 rpm). Enzyme solution was refreshed daily. Residual mass (W1) was measured after 3, 7, and 10 days. In vitro enzymatic degradation kinetics of ECM meshes. Mesh strips (2 cm × 1 cm; initial dry weight W0) were incubated in 19.2 U/mL collagenase I (PBS, pH 7.4) at 37 °C with agitation (60 rpm). Enzyme solution was refreshed daily. Residual mass (W1) was measured after 3, 7, and 10 days. Quantification of immunogenic residues Since the immunogenic potential of ECM meshes is a major concern for clinical translation, we quantified key immunogenic residues to assess decellularization efficacy. Residual DNA and the xenogeneic α-Gal epitope were measured as critical indicators of cellular clearance and potential immunogenicity. The effectiveness of decellularization was quantified by measuring residual DNA and the immunogenic α-Gal epitope (Fig. 3). The presence of the xenogeneic α-Gal epitope was detected in porcine-derived meshes (Fig. 3A). According to YY/T 1465.5 (Immunogenic evaluation method of medical devices–Part 5: Determination of α-Gal antigen clearance in medical devices utilizing animal tissues and their derivatives with M86 antibody), the α-Gal antigen is present in mammals other than humans and non-human primates. Therefore, materials derived from potentially antigenic animal tissues—specifically, PAD, SIS, and composite meshes—were selected for evaluation in this study to determine their α-Gal antigen clearance rates.Relative to native porcine dermal tissue, the α-Gal antigen clearance rate was 83.6% for PAD. For SIS, the clearance rate was 91.9% relative to its raw porcine small intestinal source material. Representative hematoxylin and eosin (H&E) stained sections confirmed effective cellular residue clearance across all dECM meshes (Fig. S7). The source data for Fig. 3 panels A-B, including all replicate values, are provided in Supplementary Data File Table S9 & S10. Quantification of residual DNA revealed significant inter-mesh differences (Fig. 3B). SIS contained 344.24 ± 6.06 ng/mg DNA, while HAD and PAD retained low DNA levels (10.40 ± 0.13 ng/mg and 3.47 ± 0.04 ng/mg, respectively). Fig. 3Quantification of immunogenic residues in decellularized ECM meshes. (A) α-Gal epitope content in PAD and SIS meshes. Values represent residual α-Gal relative to untreated porcine raw tissues (n = 3). Clearance efficiency was calculated via competitive ELISA inhibition assays. (B) Residual DNA content across ECM meshes. Linear standard curve (λ-DNA; y = 0.0172x - 0.9801, R2 = 0.9994) and PicoGreen quantification show DNA residues below ISO 10993–18 thresholds. Data are presented as mean ± SD (n = 3). Quantification of immunogenic residues in decellularized ECM meshes. (A) α-Gal epitope content in PAD and SIS meshes. Values represent residual α-Gal relative to untreated porcine raw tissues (n = 3). Clearance efficiency was calculated via competitive ELISA inhibition assays. (B) Residual DNA content across ECM meshes. Linear standard curve (λ-DNA; y = 0.0172x - 0.9801, R2 = 0.9994) and PicoGreen quantification show DNA residues below ISO 10993–18 thresholds. Data are presented as mean ± SD (n = 3). Proteomic analysis of mesh components To gain deeper insight into the compositional basis for the observed biophysical and immunological differences, we conducted a proteomic analysis of various mesh fractions following enzymatic digestion. As detailed in the Methods (“BCA protein assay for digested ECM fractions” Section), the digests were fractionated into gel (intact, hydrated ECM), supernatant (soluble peptides and proteins released during digestion), and precipitate (insoluble structural components) to mimic the physiological breakdown of ECM and separately analyze the bioactivity of released factors versus the remaining scaffold structure15. The total protein concentration was highest in HAD-derived fractions, followed by SIS and PAD (Fig. 4). SDS-PAGE combined with LC–MS/MS identification revealed that the predominant protein in all meshes was collagen (Fig. S1). Type I collagen α-chains were the major components identified in HAD fractions and PAD precipitate. Notably, keratin and intermediate filament proteins were detected in the supernatants and precipitates of PAD and SIS meshes. While keratin is a common contaminant in proteomic studies, its consistent presence suggests potential residual cellular cytoskeletal components in these meshes, though this requires further validation. Fig. 4Protein concentration detection in components of ECM meshes using BCA kits. Protein concentration detection in components of ECM meshes using BCA kits. Extract protein from the prepared mesh sediment and supernatant samples using acetone precipitation method and protein contents in the prepared gel, precipitate, and supernatant samples were detected using a Pierce™ BCA protein kit. Data are presented as mean ± SD (*p = 0.026 for HAD hydrogel vs PAD hydrogel, *p = 0.026 for HAD hydrogel vs SIS hydrogel, **p = 0.00019 for HAD supernatant vs PAD supernatant, **p = 0.0075 for PAD supernatant vs SIS supernatant, **p = 0.0045 for HAD pellet vs PAD pellet, *p = 0.011 for PAD pellet vs SIS pellet, n = 3). Cytocompatibility and cell proliferation We then assessed the biocompatibility of the meshes at the cellular level. Cytocompatibility was assessed using L929 fibroblasts, a cell line recommended for cytotoxicity testing of medical devices according to ISO 10993–5:2009 protocol21. MTT assays revealed significant cytotoxicity in SIS extracts, reducing L929 fibroblast viability to 65.8% ± 2.3% after 72 h (Fig. 5A). In contrast, viability remained high for HAD (95.8% ± 2.3%) and PAD (86.7% ± 2.9%) extracts. Detailed data are provided in Supplementary Table S8. According to ISO 10993–5:2009, a reduction in cell viability to less than 70% of the control level is considered evidence of potential cytotoxicity.This finding is consistent with previous reports linking high residual DNA in SIS to inflammatory responses25. The RTCA profiles corroborated the MTT findings. The Cell Index (CI) for cells treated with HAD and PAD extracts recovered after an initial lag phase, indicating normal proliferation. The CI for the SIS group remained low, confirming its inhibitory effect on cell growth. The Composite and negative control groups showed steady proliferation. Fig. 5Cytotoxicity evaluation on L929 cell line by the extract of ECM meshes. (A) The average cell viability of L929 cells with different kinds of mesh extracts on 1×10^5/mL cell concentration during 72 h incubation in the MTT method (n = 6). (B) The dynamic cell index from different kinds of mesh extracts on 1 × 105/mL cell concentration during 72 h incubation in the RTCA method (n = 4). Record impedance signals every 20 min in the first 12 frames (4 h) and record every 15 min in the following 80 frames (20 h). The RTCA software provides the CI value based on the impedance signal. Data are presented as mean ± SD. Cytotoxicity evaluation on L929 cell line by the extract of ECM meshes. (A) The average cell viability of L929 cells with different kinds of mesh extracts on 1×10^5/mL cell concentration during 72 h incubation in the MTT method (n = 6). (B) The dynamic cell index from different kinds of mesh extracts on 1 × 105/mL cell concentration during 72 h incubation in the RTCA method (n = 4). Record impedance signals every 20 min in the first 12 frames (4 h) and record every 15 min in the following 80 frames (20 h). The RTCA software provides the CI value based on the impedance signal. Data are presented as mean ± SD. Modulation of cell migration (chemotaxis) As cell migration is a fundamental process in tissue repair, we investigated the influence of ECM components on this behavior. The effect of mesh fractions on the migration of BALB/c 3T3 cells (Fig. 6), a widely utilized model for in vitro cell migration and wound healing assays22. All ECM mesh components (gel, precipitate, supernatant) significantly inhibited cell migration compared to the DMEM control. However, the inhibitory potency varied. The gel components of HAD and PAD showed stronger anti-chemotactic effects than that of SIS. Furthermore, the supernatant of PAD was less inhibitory than that of HAD. Fig. 6Chemotaxis effect evaluation of ECM mesh components on BALB/c 3T3 cells. Load Bal/b/c 3T3 cells into the top well of the transwell chamber in a quantity of 3 × 104 cells per well. Add each group of ECM mesh gel, precipitation and supernatant samples at the bottom and then cultivate under the condition of 37 °C and 5% CO2 for 3 h and stain migrating cells with DAPI. Observe and count cells under a microscope. Results showed that the inhibitory effect of SIS mesh hydrogel on cell chemotaxis were significantly lower than that of HAD mesh hydrogel and PAD mesh hydrogel. The inhibitory effect of PAD mesh supernatant on cell chemotaxis was significantly lower than that of HAD mesh supernatant. Data are presented as mean ± SD (**p = 0.000036 for HAD hydrogel vs DMEM, **p = 0.000032 for HAD supernatant vs DMEM, **p = 0.000038 for HAD pellet vs DMEM, **p = 0.00045 for PAD hydrogel vs DMEM, **p = 0.000037 for PAD supernatant vs DMEM, **p = 0.000054 for PAD pellet vs DMEM, **p = 0.000054 for SIS hydrogel vs DMEM, **p = 0.00037 for SIS supernatant vs DMEM, **p = 0.000046 for SIS pellet vs DMEM., n = 3). Chemotaxis effect evaluation of ECM mesh components on BALB/c 3T3 cells. Load Bal/b/c 3T3 cells into the top well of the transwell chamber in a quantity of 3 × 104 cells per well. Add each group of ECM mesh gel, precipitation and supernatant samples at the bottom and then cultivate under the condition of 37 °C and 5% CO2 for 3 h and stain migrating cells with DAPI. Observe and count cells under a microscope. Results showed that the inhibitory effect of SIS mesh hydrogel on cell chemotaxis were significantly lower than that of HAD mesh hydrogel and PAD mesh hydrogel. The inhibitory effect of PAD mesh supernatant on cell chemotaxis was significantly lower than that of HAD mesh supernatant. Data are presented as mean ± SD (**p = 0.000036 for HAD hydrogel vs DMEM, **p = 0.000032 for HAD supernatant vs DMEM, **p = 0.000038 for HAD pellet vs DMEM, **p = 0.00045 for PAD hydrogel vs DMEM, **p = 0.000037 for PAD supernatant vs DMEM, **p = 0.000054 for PAD pellet vs DMEM, **p = 0.000054 for SIS hydrogel vs DMEM, **p = 0.00037 for SIS supernatant vs DMEM, **p = 0.000046 for SIS pellet vs DMEM., n = 3). In vivo degradation and local tissue response To transition from in vitro observations to in vivo performance, we implanted the meshes into rabbit paravertebral muscle to evaluate host response and degradation kinetics in a physiological environment. This model is widely used for assessing the local tissue reaction to implant materials. The host response and degradation kinetics of meshes were evaluated following implantation in rabbit paravertebral muscle (Fig. 7, Table S1). Histopathological analysis revealed that all ECM meshes elicited a characteristic foreign body response, involving macrophages, lymphocytes, fibroblasts, and neovascularization. HAD meshes induced a relatively mild and chronic inflammatory response, with gradual degradation over 26 weeks. PAD and SIS meshes provoked a more pronounced inflammatory reaction initially. SIS meshes, despite the robust cellular infiltration, showed rapid fragmentation and degradation. Composite meshes, as non-degradable controls, were encapsulated in dense fibrous tissue with persistent inflammation. Semi-quantitative scoring confirmed that the tissue reactivity decreased over time for ECM meshes but remained elevated for the Composite group. Fig. 7Local tissue reaction and degradation of ECM meshes (40 ×). Implant ECM meshes and negative controls into the back muscle tissue of rabbits, respectively. After 1 week, 4 week, 13 week and 26 week implantation, take the implanted samples to perform the mesh degradation and histopathological analysis. Local reaction scoring after implantation was perform according to the GB/T 16,886.6–2022 standards. Subtract the average value of the control from the average value of the test to determine the reactivity level. (A–D) 1 w, 4 w, 13 w and 26 w after PAD mesh implantation, (E–H) 1 w, 4 w, 13 w and 26 w after HAD mesh implantation, (I–L) 1 w, 4 w, 13 w and 26 w after SIS mesh implantation, (M–P) 1 w, 4 w, 13 w and 26 w after the composite mesh implantation, (Q–T) 1w, 4 w, 13 w and 26 w after the HDPE control implantation. (U) The average score for the PAD mesh, HAD mesh, SIS mesh, Composite mesh and HDPE negative control based on the semi-quantitative scoring criteria (n = 12). Scale bars: 20 μm (A–T). Local tissue reaction and degradation of ECM meshes (40 ×). Implant ECM meshes and negative controls into the back muscle tissue of rabbits, respectively. After 1 week, 4 week, 13 week and 26 week implantation, take the implanted samples to perform the mesh degradation and histopathological analysis. Local reaction scoring after implantation was perform according to the GB/T 16,886.6–2022 standards. Subtract the average value of the control from the average value of the test to determine the reactivity level. (A–D) 1 w, 4 w, 13 w and 26 w after PAD mesh implantation, (E–H) 1 w, 4 w, 13 w and 26 w after HAD mesh implantation, (I–L) 1 w, 4 w, 13 w and 26 w after SIS mesh implantation, (M–P) 1 w, 4 w, 13 w and 26 w after the composite mesh implantation, (Q–T) 1w, 4 w, 13 w and 26 w after the HDPE control implantation. (U) The average score for the PAD mesh, HAD mesh, SIS mesh, Composite mesh and HDPE negative control based on the semi-quantitative scoring criteria (n = 12). Scale bars: 20 μm (A–T). Functional performance in a rat abdominal wall defect model Building on the subcutaneous implantation study, we further evaluated the functional performance of the meshes in a rat abdominal wall defect model. This model provides a more clinically relevant scenario for assessing the mechanical integration and repair efficacy of the meshes under tension. All meshes were successfully implanted to repair 1 cm2 full-thickness abdominal wall defects in rats. No postoperative infections or complications occurred, supporting the use of this model for safety assessment. The extent of adhesions to the underlying viscera was moderate and comparable across all groups at the explant time points (4, 13, 26 weeks; Table S2). Serum hematology and biochemistry profiles showed no significant abnormalities, indicating no systemic toxicity. Inflammatory cytokine profile and macrophage polarization To elucidate the immune mechanisms underlying the observed in vivo responses, we characterized the local and systemic inflammatory profiles (Fig. 8). Serum levels of TNF-α were elevated in all mesh-implanted groups at 4 and 13 weeks compared to sham controls, indicating a persistent systemic inflammatory state.To assess the temporal dynamics of inflammation, we compared cytokine levels within each implant group across time points. One-way ANOVA of TNF-α levels for each mesh type across the 4-, 13-, and 26-week time points showed a significant decrease over time (p < 0.05 for HAD, PAD, and SIS groups; ). This indicates an attenuation of the inflammatory response with prolonged implantation. It is important to note that serum cytokine levels reflect systemic immune status, whereas local gene expression (Fig. 8D–F) more directly mirrors the inflammatory activity within the peri-implant microenvironment. Expression of pro-inflammatory genes (IFN-γ, TNF-α) was highest at early time points (4w, 13w), while the expression of the anti-inflammatory cytokine IL-4 increased later (26w), suggesting a temporal transition toward tissue remodeling. SIS-induced MMP2/9 upregulation was observed in THP-1 macrophages (Fig. S6). Consistent with its high residual DNA content, SIS was associated with a 2.3-fold stronger TNF-α response in vivo than HAD at 13 weeks (*p = 0.017; Fig. 8B), a factor that may contribute to its sustained local inflammation23. Immunohistochemistry revealed a temporal shift in macrophage polarization within the implant site (Fig. 8H, I). The ratio of pro-inflammatory M1 macrophages (CD86 +) was dominant at early stages, which gradually transitioned to a predominance of pro-healing M2 macrophages (CD206 +) by 26 weeks, coinciding with the resolution of inflammation and active remodeling. HAD uniquely promoted a sustained M2-dominant state (CD206⁺/CD86⁺ ratio = 2.1 at 26 weeks). Fig. 8ECM inflammation induced during mesh degradation and the polarization of macrophages. Protein expression levels of inflammatory cytokines in rat serum, (A) Rat abdominal wall injury repair model was established for 4 weeks, (B) Rat abdominal wall injury repair model for 13 weeks (C) Rat abdominal wall damage repair model for 26 weeks; Gene expression of inflammatory-associated cytokines, (D) A rat abdominal wall injury repair model was established for 4 weeks, (E) Establishment of the rat abdominal wall injury repair model for the first 13-week (F) rat abdominal wall injury repair model for the next 26 weeks; Immuno-histochemical analysis of the expression of macrophage surface molecules in rat abdominal wall tissue, (G & H) Expression of CD86 + M1 type macrophages in the extract, (I & J) Expression of CD206 + M2 type macrophages in the extract. Scale bars: 50 μm (H&G). ECM inflammation induced during mesh degradation and the polarization of macrophages. Protein expression levels of inflammatory cytokines in rat serum, (A) Rat abdominal wall injury repair model was established for 4 weeks, (B) Rat abdominal wall injury repair model for 13 weeks (C) Rat abdominal wall damage repair model for 26 weeks; Gene expression of inflammatory-associated cytokines, (D) A rat abdominal wall injury repair model was established for 4 weeks, (E) Establishment of the rat abdominal wall injury repair model for the first 13-week (F) rat abdominal wall injury repair model for the next 26 weeks; Immuno-histochemical analysis of the expression of macrophage surface molecules in rat abdominal wall tissue, (G & H) Expression of CD86 + M1 type macrophages in the extract, (I & J) Expression of CD206 + M2 type macrophages in the extract. Scale bars: 50 μm (H&G). Note: Compared with the composite mesh group, data are presented as mean ± SD(n = 6, **p = 0.000 for Control vs PAD/ HAD/ SIS in Fig. 8A , 8B, 8C. For IL-1β expression , **p = 0.00 for Control vs Composite/ PAD/ HAD/ SIS in Fig. 8D. For IL-4 expression, **p = 0.000 for Control vs PAD/ HAD, and **p = 0.006 for Control vs SIS in Fig. 8D. For IFN-γ expression, **p = 0.006 for Control vs PAD/ HAD, and **p = 0.002 for Control vs SIS in Fig. 8D. For TNF-α expression, *p = 0.021 for Control vs composite, *p = 0.031 for Control vs PAD, *p = 0.011 for Control vs HAD, *p = 0.028 for Control vs SIS in Fig. 8D. For IL-1β expression , **p = 0.003 for Control vs PAD, p = 0.009 for Control vs HAD, p = 0.002 for Control vs SIS in Fig. 8E. For IL-4 expression, **p = 0.001 for Control vs PAD, *p = 0.012 for Control vs HAD, and **p = 0.000 for Control vs SIS in Fig. 8E. For IFN-γ expression, **p = 0.001 for Control vs composite, and **p = 0.000 for Control vs SIS in Fig. 8E. For TNF-α expression, **p = 0.003 for Control vs PAD **p = 0.000 for Control vs HAD, **p = 0.005 for Control vs SIS in Fig. 8E. For IL-1β expression , **p = 0.002 for Control vs SIS in Fig. 8F. For IL-4 expression, **p = 0.002 for Control vs SIS in Fig. 8F. For IFN-γ expression, *p = 0.049 for Control vs PAD, and **p = 0.002 for Control vs HAD in Fig. 8F. For TNF-α expression, *p = 0.044 for Control vs PAD *p = 0.045 for Control vs HAD, **p = 0.003 for Control vs SIS in Fig. 8F. For 13w, **p = 0.000 for Composite vs PAD/ SIS, and *p = 0.020 for Composite vs HAD in Fig. 8G. For 26w, *p = 0.027 for Composite vs PAD, *p = 0.020 for Composite vs HAD, *p = 0.041 for Composite vs SIS in Fig. 8G). Collagen deposition and tissue remodeling The ultimate goal of ECM scaffold implantation is constructive tissue remodeling. Therefore, we analyzed the quality of neotissue formation and collagen deposition at the implant site. The quality of neotissue formation and collagen deposition was analyzed using Sirius Red and Masson’s Trichrome staining (Fig. 9). All acellular ECM meshes (HAD, PAD, SIS) promoted robust deposition of new, densely packed collagen fibers by 26 weeks, facilitating strong integration with the native tissue. The Composite mesh, while effective in repair, was associated with a looser collagen architecture and encapsulation, typical of a foreign body response to non-degradable materials. H&E staining confirmed that macrophage infiltration decreased over time as the ECM meshes degraded and were replaced by host tissue. Fig. 9Collagen remodeling in rat abdominal wall defects. (A) Sirius Red staining of collagen types, (B) Masson’s trichrome staining of collagen fibers, (C) H&E staining showing macrophage infiltration. Scale: 20 × /40 × as indicated. Scale bars: 100 μm (A), 50 μm (B–C). Collagen remodeling in rat abdominal wall defects. (A) Sirius Red staining of collagen types, (B) Masson’s trichrome staining of collagen fibers, (C) H&E staining showing macrophage infiltration. Scale: 20 × /40 × as indicated. Scale bars: 100 μm (A), 50 μm (B–C). Systemic metabolic alterations Beyond local tissue responses, we sought to determine if ECM mesh implantation induced systemic effects. Serum metabolomics analysis was performed to identify potential metabolic reprogramming in host animals. Serum metabolomics analysis indicated that implantation of ECM meshes influenced host metabolism. Levels of several amino acids (L-methionine, L-threonine, L-serine, etc.) were upregulated, indicating activated amino acid biosynthesis and protein translation pathways. Distinct metabolic pathways, including starch/sucrose metabolism, were specifically altered in the ECM mesh groups compared to the Composite group. Modulation of macrophage function in vitro To directly probe the interaction between ECM components and immune cells, we investigated the modulation of human macrophage (THP-1 cell) function in vitro (Fig. S6). THP-1 cells are a well-established human monocytic cell line that can be differentiated into macrophages, allowing for the study of human-specific immune responses. ECM hydrogels altered the phagocytic capacity of THP-1 cells. Furthermore, RT-qPCR and ELISA analyses demonstrated that ECM components significantly downregulated the expression of key signaling molecules (STAT-1, STAT-3) and inflammatory cytokines, while upregulating the expression of remodeling enzymes (uPA, MMP2, MMP9), suggesting that the meshes actively promote a pro-remodeling phenotype in macrophages. Discussion This study provides a comparative analysis of three ECM meshes within a defined preclinical context relevant to soft tissue implantation. We aimed to elucidate how intrinsic material properties drive biological responses in this setting, thereby establishing a foundational framework to inform future application-specific optimization. Our multi-faceted analysis began with structural and in vitro degradation studies. The rapid in vitro degradation of SIS, attributable to its porous, loosely organized collagen ultrastructure (Fig. 1H) and potentially its collagen composition, aligns with its potential for fast turnover. However, a critical divergence emerged in vivo. Its propensity for rapid fragmentation and sustained inflammation (Fig. 7I–L, 8A–C) highlights a key limitation of relying solely on enzymatic degradation assays and underscores the paramount role of the host immune system in driving remodeling outcomes. This is supported by our observation of SIS-induced MMP upregulation in THP-1 macrophages (Fig. S6). We propose that the rapid, uncontrolled physical disintegration of SIS generates a high burden of particulate debris, perpetuating a foreign body response, a phenomenon documented for other biomaterials. This perspective complements the discussion on immunogenic residues and suggests that optimal scaffold design must balance degradation rate with structural integrity to avoid adverse fragmentation. A key finding was the incomplete decellularization of SIS, which contained significant residual DNA (344.24 ± 6.06 ng/mg). This likely contributed to its significant cytotoxicity in vitro (65.8% viability; Fig. 5A) and the elevated TNF-α response in vivo (Fig. 8A–C), potentially via TLR9 activation. While we focused on DNA, we acknowledge the critique that DNA itself may not be the primary immunogen, and other factors like residual detergents or endotoxin could play a significant role. Future work will explicitly quantify endotoxin levels and employ targeted assays to dissect the relative contributions of these factors. Furthermore, proteomic identification of residual keratin in PAD and SIS, while a common contaminant, suggests that DNA quantification alone is an insufficient metric for decellularization efficacy, reinforcing the need for complementary histological assessment. Functionally, all ECM meshes supported a temporal shift from M1 to M2 macrophage polarization (Fig. 8H–I). However, HAD uniquely promoted a sustained M2-dominant state (CD206⁺/CD86⁺ ratio = 2.1). We hypothesize that the superior structural integrity of HAD provides a more stable mechanical microenvironment that promotes a pro-regenerative phenotype, whereas the fragmented morphology of SIS perpetuates a pro-inflammatory state through continuous phagocytic challenge. Although SIS exhibited compensatory mechanisms, such as IL-4 upregulation and MMP secretion, its significantly stronger initial inflammatory response may compromise long-term functional integration. The use of different pig breeds is acknowledged as a potential confounder affecting decellularization efficiency and warrants investigation in future breed-standardized studies. Beyond the local implant site, our serum metabolomics analysis revealed that ECM scaffolds can induce systemic metabolic reprogramming, notably accelerating L-methionine metabolism. As a key methyl donor, L-methionine is integral to the epigenetic regulation of immune cell plasticity. These findings suggest a novel paradigm in which ECM scaffolds systemically recruit metabolic resources to support localized repair processes, opening avenues for non-invasive monitoring of remodeling status via metabolic biomarkers. In conclusion, our data suggest that HAD, with its optimal biocompatibility and controlled remodeling, may be suited for applications requiring long-term mechanical support. In contrast, the rapid degradation of SIS, while a liability in some contexts, might be leveraged in time-critical repairs where swift scaffold turnover is desired. The practical implications of our framework are threefold: First, decellularization protocols should be standardized and assessed using duplex assays for both DNA and α-Gal, complemented by histology. Second, ultrastructural preservation must be prioritized to guide favorable immune kinetics and prevent adverse fragmentation. Finally, metabolic signatures hold promise as biomarkers for clinical monitoring. Future work should focus on refining tissue-specific decellularization to maximize the retention of beneficial bioactive motifs while eliminating immunogenic residues, and on incorporating application-specific functional models to validate these material-specific insights in clinically relevant scenarios. This study has several limitations. Histological evaluations were primarily descriptive; future quantitative digital histomorphometry across entire implant sections would provide more robust metrics for tissue remodeling. Finally, the in vivo models, while simulating soft tissue contact, do not fully recapitulate the biomechanical stresses of specific clinical applications such as hernia repair under load. Method Ethics statement Human tissue Donor skin was obtained from a licensed tissue bank (The Yinfeng Institute of Cryomedicine, Jinan, Shandong Province, China; donor’s information is provided in Supplementary Fig. S8). The use of human tissue for research was approved by the Research Ethics Committee of Shandong Institute of Medical Device and Pharmaceutical Packaging Inspection (Approval No. KY2022001), and informed consent was obtained from all donors. All methods involving human tissue were carried out in accordance with relevant guidelines and regulations. Animal experiments All animal experiments were carried out in accordance with relevant guidelines and regulations, including the ARRIVE guidelines and the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The experimental protocols were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of Shandong Institute of Medical Device and Pharmaceutical Packaging Inspection (Approval Nos. P20200506001 & M20221011002). The study was designed and conducted in strict adherence to the ‘3Rs’ principles (Replacement, Reduction, and Refinement) to minimize animal use and discomfort. ECM mesh preparation and sourcing Human acellular dermis (HAD) Healthy, full-thickness human skin was obtained from a licensed tissue bank (The Yinfeng Institute of Cryomedicine, Jinan, Shandong Province, China). Subcutaneous fat was surgically removed. The dermis was sterilized in 0.05% (v/v) peracetic acid (Sigma-Aldrich, #Z112402) for 2 h, followed by extensive rinsing with sterile phosphate-buffered saline (PBS, Gibco). Lipids were extracted using isopropanol (Sigma-Aldrich) for 18 h. Decellularization was performed by agitation (60 rpm) in a solution containing 0.5% (w/v) TrypLE Express (Thermo Fisher Scientific) and 6% (w/v) NaCl in 10 mM HEPES buffer (pH 8.0) for 24 h at room temperature. The supernatant was discarded, and the tissue was further treated with 0.5% (w/v) sodium deoxycholate (Sigma-Aldrich) in 10 mM HEPES (pH 8.0) for 20 h under agitation. Finally, the dermis was thoroughly washed in PBS and lyophilized. Porcine acellular dermis (PAD) Full-thickness skin was harvested from six-month-old female Yorkshire pigs (n = 6 per group; a different breed from the SIS source, acknowledged as a study limitation) raised in a closed-loop, specific-pathogen-free facility (Tianjin Bainong Laboratory Animal Breeding Technology Co., Ltd., License No.: SCXK (Jin) 2020–0002). The decellularization protocol was identical to that used for HAD to allow for comparative analysis, as detailed in “Human Acellular Dermis (HAD)” Section. Porcine small intestinal submucosa (SIS) Pigs were euthanized by intravenous injection of pentobarbital sodium (100 mg/kg) followed by exsanguination. Jejunum was harvested from conventionally raised adult female miniature pigs (Tianjin Bainong Laboratory Animal Breeding Technology Co., Ltd.) slaughtered on the day of processing. Tissues were pooled from 6 animals to obtain the required quantity of SIS. After cleaning and disinfecting, the mucosa and submucosa were separated and decellularized by agitation in 1 M NaCl containing 0.3% (w/v) Triton X-100 (Sigma-Aldrich #X100-5 mL) for 18 h. Residual mucosa and muscle layers were mechanically scraped off. The tissue was then soaked in purified water for 90–120 min, sterilized in 0.10% (v/v) peracetic acid (PBS, pH 7.4) for 60 min, treated with isopropanol for lipid removal, and finally lyophilized. Note: The use of different pig breeds (Yorkshire for PAD, Miniature for SIS) and tissue sources is acknowledged as a potential confounder. Animals Rabbits Twenty-four (24) female New Zealand White rabbits (2.5–3.5 kg; Jinan Jinfeng Laboratory Animal Co., Ltd., Certificate No.: SCXK (Lu) 20180006) were used for muscle implantation studies.Rabbits were euthanized at each time point (1, 4, 13, 26 weeks) via intravenous injection of pentobarbital sodium (100 mg/kg). Rats Sixty (60) female Sprague Dawley rats (180–200 g; SPF grade; Beijing Vital River Laboratory Animal Technology Co., Ltd., License No.: SCXK (Jing) 20,210,011) were used for the abdominal wall defect model. Animals were randomly assigned to experimental groups (n = 10 per group/time point).Rats were euthanized at 4, 13, and 26 weeks via intraperitoneal injection of pentobarbital sodium (100 mg/kg). Mice BALB/c 3T3 cells (ATCC, #CCL-163) were used for chemotaxis assays. In vitro degradation assay Mesh samples (n = 5 per group) were cut into 2 cm × 1 cm strips, weighed (initial dry weight, W0), and incubated in 10 ml of PBS (pH 7.4) containing 60 μg/ml of Type I collagenase from Clostridium histolyticum (Sigma-Aldrich, #C0130) at 37 °C with constant agitation at 60 rpm. The enzyme solution was replaced every 24 h to maintain activity. Samples were retrieved at pre-determined time points (3, 7, 10 days), rinsed extensively with deionized water, lyophilized, and weighed (degraded dry weight, W1). Mass loss was calculated as degradation rate (%) = 1-\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\frac{{W}_{1}}{{W}_{0}}$$\end{document}. BCA protein assay for digested ECM fractions Rationale and fraction preparation To evaluate the protein content released from and remaining within the ECM meshes during enzymatic digestion—a process critical for understanding their in vivo degradation and remodeling—we fractionated the digests into soluble and structural components. This approach is supported by Slivka et al.12, who demonstrated that soluble and structural fractions of ECM hydrogels differentially regulate macrophage behavior and immune responses. Gel This term refers to the intact, hydrated ECM mesh prior to digestion, representing the initial implantable material. Supernatant Following collagenase digestion (as detailed in “Proteomic analysis of mesh components” Section), the digest was centrifuged at 12,000×g for 20 min at 4 °C. The resulting supernatant contains the soluble fraction of the ECM, encompassing peptides, growth factors, and other bioactive molecules liberated during degradation. Precipitate The insoluble pellet after centrifugation represents the structural fraction, comprising primarily undigested or partially digested structural proteins like collagen fibrils and elastin. These divisions were employed to mimic the physiological breakdown of ECM in vivo and to separately analyze the bioactivity of released soluble factors versus the remaining scaffold structure . LC–MS/MS proteomic analysis Protein samples from gel, precipitate, and supernatant fractions were reduced with 10 mM dithiothreitol (DTT) at 56 °C for 30 min, alkylated with 55 mM iodoacetamide (IAA) in the dark for 30 min, and digested with trypsin (Promega) at a 1:50 (w/w) enzyme-to-protein ratio overnight at 37 °C. The resulting peptides were desalted using C18 StageTips. LC–MS/MS analysis was performed on a Q-Exactive HF-X mass spectrometer (Thermo Fisher Scientific) coupled to an EASY-nLC 1200 system (Thermo Fisher Scientific). Peptides were separated on a 75 μm × 25 cm PepMap RSLC C18 column (Thermo Fisher Scientific) using a 60-min gradient from 4 to 28% acetonitrile in 0.1% formic acid at a flow rate of 300 nL/min. The mass spectrometer was operated in data-dependent acquisition (DDA) mode. Full MS scans were acquired at a resolution of 60,000 with an AGC target of 3e6. The top 20 most intense ions were selected for HCD fragmentation with a normalized collision energy of 28%. MS/MS scans were acquired at a resolution of 15,000 with an AGC target of 1e5. Raw data were processed using MaxQuant software (version 2.0.3.0) against the Swiss-Prot human or porcine database (as appropriate) with carbamidomethylation of cysteine set as a fixed modification and oxidation of methionine and hydroxylation of proline and lysine set as variable modifications. For label-free quantification (LFQ), peptide intensity values were normalized to the total protein intensity per sample to account for loading variations. The LFQ was performed using the MaxLFQ algorithm. Pathway analysis of the proteomic data is presented in Fig. S2. Details regarding serum metabolomics are provided in the Supplementary Information. Quantification of immunogenic residues α-Gal epitope content The α-Gal content was quantified using a modified ELISA. The α-Gal epitope content was quantified using a standardized ELISA inhibition assay established by Lu et al. 201824, which employs artificial Gal-BSA antigen as a calibration standard and optimizes antigen exposure through lysate buffer homogenization. This method has been validated for high reproducibility (RSD = 12.48%) and sensitivity (LDL ≈ 7.1 × 1011 epitopes/reaction), providing a critical QC metric for assessing xenogeneic immunogenicity in human applications. Briefly, mesh samples (n = 5) were incubated with a mouse monoclonal anti-Gal IgM antibody (M86; Alexis, #01,031,931) at a 1:100 dilution overnight at 4 °C. The remaining antibody in the supernatant was measured by ELISA using α-gal-bovine serum albumin (α-gal-BSA;Dextra, #ATDX276-139) as a capture antigen and an HRP-conjugated goat anti-mouse IgM (1:5000; Elabscience, #E-AB-1008) as the secondary antibody. Absorbance was read at 450 nm. A standard curve was generated using α-gal-BSA, and results are expressed as ng of α-gal epitope per mg of dry mesh. Residual DNA content Residual DNA was quantified using the PrepSEQ™ Sample Preparation Kit and the Quant-iT™ PicoGreen™ dsDNA Assay Kit (Thermo Fisher Scientific, Waltham, MA, #P11496) according to the manufacturer’s instructions. Briefly, 10 mg of each mesh (n = 5) was digested. DNA was purified using magnetic particles, eluted in 50 µl of elution buffer, and quantified fluorometrically with PicoGreen reagent against a λ-DNA standard curve. Results are expressed as ng DNA per mg dry weight. Real-time cell analysis (RTCA) for cytocompatibility Cytocompatibility of mesh extracts was assessed using the xCELLigence RTCA DP system (ACEA Biosciences). L929 fibroblasts (ATCC, #CCL-1) were seeded at a density of 3 × 104 cells/well into E-Plates L8 PET (Agilent , #300,600,870). After 24 h, the culture medium was replaced with extracts prepared by incubating 1.25 cm2/ml of each mesh in MEM for 24 h at 37 °C. The negative control was prepared using high-density polyethylene at a ratio of 1 mL of cell culture medium per 6 cm2. The positive control was MEM containing 0.1% zinc diethyldithiocarbamate, and the blank control was fresh MEM alone. Data were normalized to the CI value at the time of extract addition. Chemotaxis assay Chemotaxis was evaluated using 24-well Transwell plates with 8.0 μm pore polycarbonate membranes (Corning, #3422). BALB/c 3T3 cells were serum-starved overnight in DMEM with 0.5% heat-inactivated FBS. Cells were trypsinized, resuspended in serum-free DMEM, and 3 × 104 cells in 200 µl were added to the upper chamber. The gel, precipitate, and supernatant fractions (500 µl each, prepared as described in Fig. 4 legend) or a positive control (DMEM with 10% FBS) were added to the lower chamber. After incubation for 3 h at 37 °C and 5% CO2, non-migrated cells on the upper membrane surface were removed with a cotton swab. Migrated cells on the bottom surface were fixed, stained with DAPI (1 µg/ml; Sigma-Aldrich, #D9542), and imaged using a fluorescence microscope (Olympus IX73; 10 × objective). Cells were counted in five random fields per well (n = 3 wells per group). Images were captured at 1360 × 1024 resolution (0.65 μm/pixel) using identical exposure settings (200 ms) across all groups. Cell counting was performed by two independent investigators blinded to group assignments. In vivo implantation models Rabbit muscle implantation Rabbits (n = 6 per time point) were anesthetized with intravenous Zoletil 50 (6 mg/kg; Virbac). A 1 cm × 1 cm sample of each ECM mesh or a negative control material (High-density polyethylene, HDPE) was aseptically implanted into the paravertebral muscle. Implants and surrounding tissue were explanted at 1, 4, 13 and 26 weeks post-implantation for histopathological analysis. Rat abdominal wall defect model Rats (n = 10 per group) were anesthetized with intraperitoneal Zoletil 50 (25 mg/kg). A 1 cm × 1 cm full-thickness defect was created in the abdominal wall. Defects were repaired by implanting and suturing in place a 1.5 cm × 1.5 cm mesh of the respective ECM mesh. Sham-operated animals served as controls. Animals were euthanized at 4 and 26 weeks for evaluation. Histological and immunohistochemical staining Explanted tissues were fixed in 10% neutral buffered formalin, paraffin-embedded, and sectioned (5 µm thickness). Sections were stained with: H&E Collection and Processing of Implanted Meshes for Histopathological Evaluation: The entire implant area with ≥ 5 mm of adjacent tissue was collected. Tissues were bisected longitudinally along the midline parallel to the animal’s body axis. One half was allocated for biomechanical testing, and the other for histopathological assessment. Histopathological processing Tissues designated for histopathology were immediately immersed in 10% neutral buffered formalin. After complete fixation, samples were processed for microscopic evaluation. Tissue specimens ≥ 1 cm × 1 cm were sectioned using the diagonal continuous slicing method. Samples were fixed in 10% neutral formalin, paraffin-embedded, sectioned (5 μm thickness), and stained with Hematoxylin and Eosin (H&E) for histopathological examination. Scoring methodology Parameters recorded: Inflammatory cell types and tissue responses within the repair zone. Semi-quantitative scoring was performed according to predefined criteria. Microscopy: Tissues were examined at 40 × , 100 × , and 200 × magnifications. Quantification: A minimum of 5 fields per sample were analyzed at 400 × magnification, and averaged. Score calculation: Average test group score = (Σ scores of test samples) / (number of test samples) Average control group score = (Σ scores of control samples) / (number of control samples) Final score = Average test group score—Average control group score (Negative values recorded as 0). Scoring criteria were as follows: 0.0–2.9, none to minimal stimulation; 3.0–8.9, mild stimulation; 9.0–15.0, moderate stimulation; ≥ 15.1, severe stimulation. Masson’s Trichrome Using a commercial kit (Solarbio, #G1340) to evaluate collagen deposition and fibrosis. IHC For macrophage polarization, sections were stained with monoclonal antibodies against CD86 (M1 marker;CD86-PE, Genetex, #GTX43009) and CD206 (M2 marker; CD206-PE, Biolegend, #141,708). Appropriate isotype controls were included. Staining was visualized using a DAB kit (Boster, #AR1000). All histological slides were reviewed by a board-certified veterinary pathologist blinded to the experimental groups. Statistical analysis All quantitative data are presented as mean ± standard deviation (SD). Statistical analysis was performed using GraphPad Prism software (version 9.0). Normality was assessed using the Shapiro–Wilk test. For comparisons across multiple groups, a two-way analysis of variance (ANOVA) was performed, followed by Tukey’s post-hoc test for multiple comparisons. A p-value of < 0.05 was considered statistically significant. The sample size (n) for each experiment is specified in the respective method section. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1