Comparative Assessment of Aspergillosis by Virtual Infection Modeling in Murine and Human Lung

Abstract

Aspergillus fumigatus is a ubiquitous opportunistic fungal pathogen that can cause severe infections in immunocompromised patients. Conidia that reach the lower respiratory tract are confronted with alveolar macrophages, which are the resident phagocytic cells, constituting the first line of defense. If not efficiently removed in time, A. fumigatus conidia can germinate causing severe infections associated with high mortality rates. Mice are the most extensively used model organism in research on A. fumigatus infections. However, in addition to structural differences in the lung physiology of mice and the human host, applied infection doses in animal experiments are typically orders of magnitude larger compared to the daily inhalation doses of humans. The influence of these factors, which must be taken into account in a quantitative comparison and knowledge transfer from mice to humans, is difficult to measure since in vivo live cell imaging of the infection dynamics under physiological conditions is currently not possible. In the present study, we compare A. fumigatus infection in mice and humans by virtual infection modeling using a hybrid agent-based model that accounts for the respective lung physiology and the impact of a wide range of infection doses on the spatial infection dynamics. Our computer simulations enable comparative quantification of A. fumigatus infection clearance in the two hosts to elucidate (i) the complex interplay between alveolar morphometry and the fungal burden and (ii) the dynamics of infection clearance, which for realistic fungal burdens is found to be more efficiently realized in mice compared to humans.

Keywords: Aspergillus fumigatus lung infection; human model; hybrid agent-based computer simulations; mouse model; virtual infection modeling.

Figure 1

Visualization of a to-scale alveolus in the hybrid agent-based model for mouse (A) and human (B). The alveolar entrance ring (left) and Pores of Kohn (black) represent entry/exit points for AM (green) and chemokine flow (white isolines) induced by conidium (red). Alveolar surface is covered with epithelial cells of type 1 (yellow) and 2 (blue).

Figure 2

Alveolar occupation number, the maximal expected number of present conidia per alveolus, as a function of the fungal burden in mouse (blue) and human (red). Black line represents the experimental range of fungal burden, which is reached in typical mice model experiment.

Figure 3

(A) Infection scores IS for random walk migration and selected examples of chemokine parameters in the limit of low fungal burden with AON = 1. Dashed-dotted black line indicates the threshold infection score at IS_t = 5%. Error bars represent 95%-confidence intervals received from bootstrapping. (B) Mean probability for directed AM migration p_directed following the underlying chemokine gradient as a function of the distance from the source AEC. Chemokine parameters are set to the corresponding optima $\overline{D_{opt}}$ and $\overline{s_{AE{C}_{opt}}}$ in mice and men.

Figure 4

Infection scores IS as a function of the AON (A,B) and the fungal burden (C,D) for selected secretion parameters with diffusion coefficient $D=200\mu m^2\stackrel{-1}{min}$ (A,C) and for optimal chemokine parameters $\overline{D_{opt}}$ and $\overline{s_{AE{C}_{opt}}}$ (B,D) in mice and men. Dashed-dotted black line indicates the threshold infection score at IS_t = 5%. Error bars represent 95% confidence intervals. Black line represents the experimental range of fungal burden, which is reached in typical mice model experiment.

Figure 5

Mean of the normalized alveolar chemokine concentration as a function of the AON in mice and men obtained from simulations. Error bars represent the standard deviation of this measurement.

Figure 6

Median of the relative infection score between human and mouse, ΔIS = 1 − IS^M/IS^H, depending on the scaling factor for chemotaxis parameters: $f_D=D^H/D^M$ for the diffusion coefficient (purple) and $f_{s_{AEC}}=s_{AEC}^H/s_{AEC}^M$ for the secretion rate (orange). Error bars represent the standard errors.