Towards predicting implant-induced fibrosis: A standardized network model of macrophage-fibroblast interactions

Abstract

The foreign body response (FBR) is a complex and multifaceted process that remains incompletely understood, often leading to complications in medical device integration. In this study, we constructed a literature-based network of the FBR and developed it into a semi-quantitative predictive model to better understand its dynamics. The in silico FBR model incorporates key material-related factors, including immunogenic properties and mechanical mismatch, which influence immune cell activation and extracellular matrix (ECM) deposition. Predictions align with existing knowledge, showing that material stiffness and tissue progressive stiffening due to increased ECM deposition can exacerbate the FBR and that feedback interactions can protect the system from pathological outcome by gradually reducing the initial inflammatory input. The model also successfully replicated six out of eight experimental cases of anti-fibrotic interventions, demonstrating its potential as a predictive tool. Assessing implant safety in the early pre-clinical stages of device development is critical for ensuring long-term functionality and reducing adverse reactions. By systematically analyzing and integrating all interacting aspects of the FBR, in silico modeling can provide valuable insights and complement in vitro and in vivo studies for improved implant safety assessment.

Keywords: Fibrotic tissue; Foreign body response; Immune response; Mechanotransduction; Standardized ordinary differential equations.

Graphical abstract

Fig. 1

a) Schematic overview of the different phases of the foreign body response (FBR) highlighting main cell types involved and material characteristics. Abbreviations: M1 (macrophages of type 1), M2 (macrophages of type 2), FBGC (foreign body giant cells). b) Literature-based network of processes involved in the FBR. The network represents all variables included in our in silico model and their interactions as described in the literature. Sharp blue arrows indicate interactions leading to activation or an increase in the target variable, while blunt orange arrows represent inhibitory effects or decreases in the target variable. Abbreviations: TNF-α (tumor necrosis factor-α), IFN-γ (interferon-γ), IL- (interleukin-), M1 (macrophages of type 1), M2 (macrophages of type 2), MMPs (matrix metalloproteinases), TIMPs (tissue inhibitors of metalloproteinases), TGF-β (transforming growth factor-β), PDGF (platelet derived growth factor), ECM (extracellular matrix). Ant is used to indicate other antagonists of TGF-β that were not explicitly included in the model. A indicates the input variable activating immune cells, B the input of immunogenic properties of the material and C the input variable for the mechanical mismatch in the peri-implant area (see Table 1 for a more detailed description).

Fig. 2

Effect of increasing the starting value of input A on the FBR network variables. Blue: A(0)= 0.3, Yellow A(0)= 0.9. In both cases BC= 0. The normalized value at steady state between 0 and 1 represents the normalized concentration level (see Methods Section 2.2). M1 (macrophages of type 1), M2 (macrophages of type 2), TGF-β (transforming growth factor-β), PDGF (platelet derived growth factor), MMPs (matrix metalloproteinases), TIMPs (tissue inhibitors of metalloproteinases), F (fibroblasts), mF (myofibroblasts), ECM (extracellular matrix).

Fig. 3

Effect of initial values of variables B and C on ECM outcomes, with varying initial values of A. The figure illustrates the influence of initial conditions for variables B and C on the normalized extracellular matrix (ECM) levels at steady state, under different initial values of variable A. Each subplot corresponds to a specific initial value of A, from the left: A = 0.1, A = 0.4, A = 0.7, and A = 0.9. The normalized ECM levels are categorized into two outcomes: low (below 0.35, shown in green) and high (above 0.35, shown in blue).

Fig. 4

The effect of feedback loops involving input variables on the high ECM outcome in the in silico FBR model. a) steady state ECM values as a function of the initial value of the input variable A, highlighting the value of A (A*) for which the threshold of 0.35 is reached. Solid line = B= 1, C= 1, with feedback; Dotted line: B= 1, C= 1, without feedback. b) A* for different combinations of B and C input values. Solid line, solid marker: baseline model with feedback loops to input variables. Dotted line, hollow marker: model without feedback loops to input variables.