Multiscale Model of Nutritional Immunity in Invasive Pulmonary Aspergillosis

Dr. Reinhard Laubenbacher, Dr. Borna Mehrad, Dr. Will Schroeder, Dr. Luis Sordo Viera, Dr. Yogesh Scindia, Brian Helba, Dr. Henrique de Assis Lopes Ribiero, Dr. Matt Wheeler, Dr. Adam Knapp, Spencer Whitten, Ganlin Qu, Yana Goddard, Marie Burdick, Marie D. Burdick, Dhruv Desai, Sadat Kassem, Annanya Agarwal, Dr. Ning Yang

The purpose of the work is to develop a whole-lung multiscale computational model to simulate the invasion of the human lungs by pathogens including spores of the opportunistic fungus Aspergillus fumigatus. In particular, we explore the role of nutritional immunity in clearing infection, with a focus on iron. This project is currently funded by The National Institutes of Health and The National Science Foundation of the United States of America. The team includes members from UF Health and Kitware.

Multiscale model of nutritional immunity in invasive pulmonary aspergillosis

News

June 8, 2021New preprint on the bioRxiv! – “The Innate Immune Response to Invasive Pulmonary Aspergillosis: A Systems Modeling Approach”

May 18, 2021Check out our new paper! – “A Modular Computational Framework for Medical Digital Twins”

Team Members

Laboratory for Systems Medicine

Dr. Reinhard Laubenbacher

Dr. Luis Sordo Viera

Dr. Henrique de Assis Lopes Ribiero

Dr. Adam Knapp

Dr. Mehrad’s Lab

Dr. Borna Mehrad

Dr. Yogesh Scindia

Dr. Matt Wheeler

Spencer Whitten

Ganlin Qu

Yana Goddard

Marie Burdick

Kitware

Dr. Will Schroeder

Brian Helba

Funding

Multiscale modeling of the battle over iron in invasive lung infection

Funding: National Institute of Allergy and Infectious Diseases 1R01AI135128-01

Abstract: Invasive aspergillosis is among the most common fungal infection in immunocompromised hosts and carries a poor outcome. The spores of the causative organism, Aspergillus fumigatus, are ubiquitously distributed in the environment. Healthy hosts clear the inhaled spores without developing disease, but individuals with impaired immunity are susceptible to a life-threatening respiratory infection that can then disseminate to other organs. The increasing use of immunosuppressive therapies in transplantation and cancer has dramatically increased suffering and death from this infection, and this trend is expected to continue. Current therapeutic approaches have been focused primarily on the pathogen, but a better understanding of the components of host defense in this infection may lead to the development of new treatments against this infection, possibly in combination with antifungal drugs. Iron is essential to all living organisms, and restricting iron availability is a critical mechanism of antimicrobial host defense against many microorganisms; conversely, successful pathogens have evolved potent mechanisms for scavenging iron from the host. These mechanisms have the potential to be harnessed therapeutically, for example with drugs that enhance the host’s iron sequestration mechanisms. The overarching goal of this project is to develop a multi-scale mathematical model that can serve as a simulation tool of the role of iron in invasive aspergillosis. The model will integrate mechanisms at the molecular scale with tissue-level events and a whole-body scale capturing the role of the liver. The project brings an innovative approach to the study of this infection, and introduces innovative features to multiscale modeling through a novel modular software design that improves flexibility, reproducibility, and model sharing.

Modular design of multiscale models, with an application to the innate immune response to fungal respiratory pathogens

Funding: National Institute of Biomedical Imaging and Bioengineering 1U01EB024501-01, National Science Foundation EAGER #1750183

Abstract: Increased availability of biomedical data sets across spatial and temporal scales makes it possible to calibrate complex models that capture integrated processes from the molecular to the whole organism level. This complexity poses multiple challenges related to mathematical modeling, software design, validation, reproducibility, and extensibility. Visualization of model features and dynamics is a key factor in the usability of models by domain experts, such as experimental biologists and clinicians. The proposed project addresses these challenges in the context of the immune response to an important respiratory fungal infection. Its goal is to develop a novel modular approach to model architecture, using a recently introduced technology of lightweight virtual machines and our user-friendly open-source platform for the construction and linking of these so-called “Docker containers” to create complex modular models in a transparent fashion. A key benefit of software containers is that they can encompass the entire computational environment of a model, enabling unprecedented reproducibility of computational results. The overarching computational goal is to develop a novel approach to the modular design of multiscale models. While broadly applicable, this novel computational modeling approach will be focused on the development of a multiscale model capturing the early stages of invasive aspergillosis, an important health problem. Invasive aspergillosis is one of the most common fungal infections in immunocompromised hosts and carries a poor prognosis. The spores of the causative organism, Aspergillus fumigatus, are ubiquitously distributed in the environment. Healthy hosts clear the inhaled spores without developing disease, but individuals with impaired immunity are susceptible to a life-threatening respiratory infection that can then disseminate to other organs. The increasing use of immunosuppressive therapies in transplantation and cancer has dramatically increased suffering and death from this infection, and this trend is expected to continue. Current therapeutic approaches have been focused primarily on the pathogen, but a better understanding of the components of host defense in this infection may lead to the development of new treatments. In particular, restricting iron availability is a critical mechanism of antimicrobial host defense; conversely, successful pathogens have evolved potent mechanisms for scavenging iron from the host. These mechanisms have the potential to be harnessed therapeutically. The biological focus of the proposed project is the battle over iron between the fungus and the host. The overarching biomedical goal is to develop a simulation tool to explore the role of iron in invasive aspergillosis across biochemical and biophysical conditions.

Publications

Masison, J., J. Beezley, Y. Mei, H. a. L. Ribeiro, A. C. Knapp, L. Sordo Vieira, B. Adhikari, et al. “A Modular Computational Framework for Medical Digital Twins.” Proceedings of the National Academy of Sciences 118, no. 20 (May 18, 2021). https://doi.org/10.1073/pnas.2024287118.

Cagnina, R. Elaine, Kathryn R. Michels, Alexandra M. Bettina, Marie D. Burdick, Yogesh Scindia, Zhimin Zhang, Thomas J. Braciale, and Borna Mehrad. “Neutrophil-Derived TNF Drives Fungal Acute Lung Injury in Chronic Granulomatous Disease.” The Journal of Infectious Diseases, April 5, 2021. https://doi.org/10.1093/infdis/jiab188.

Wooten, David J., Jorge Gómez Tejeda Zañudo, David Murrugarra, Austin M. Perry, Anna Dongari-Bagtzoglou, Reinhard Laubenbacher, Clarissa J. Nobile, and Réka Albert. “Mathematical Modeling of the Candida Albicans Yeast to Hyphal Transition Reveals Novel Control Strategies.” Preprint. Systems Biology, January 20, 2021. https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1008690

L. Sordo Vieira, R. Laubenbacher, and D. Murrugarra, Control of Intracellular Molecular Networks Using Algebraic Methods, Bull. Math. Biol., 82(2), 2019.

Aguilar, Boris, Pan Fang, Reinhard Laubenbacher, and David Murrugarra. 2020. “A Near-Optimal Control Method for Stochastic Boolean Networks”. Letters in Biomathematics 7 (1), 67–80.

E. Paul, G. Pogudin, W. Qin and R. Laubenbacher, The dynamics of canalizing Boolean networks, Complexity, https://doi.org/10.1155/2020/3687961, 2020.

Michels KR, Solomon AL, Scindia Y, Vaulont S, Burdick MD, Laubenbacher R, Mehrad B. Aspergillus utilizes heme as an iron source during invasive pneumonia. Submitted to mBio (revising manuscript to resubmit).

Solomon A, Michels K, Laubenbacher R, Scindia Y, Mehrad B. The role of heme uptake and hemopexin in invasive pulmonary aspergillosis. American Thoracic Society Virtual Conference, 2020.

Qu G, Solomon A, Yang N, Lin C, Scindia Y, Mehrad B. Alveolar macrophages protect against acute lung injury. American Thoracic Society Virtual Conference, 2020.

Masison, J., J. Beezley, Y. Mei, H. a. L. Ribeiro, A. C. Knapp, L. Sordo Vieira, B. Adhikari, et al. “A Modular Computational Framework for Medical Digital Twins.” Proceedings of the National Academy of Sciences 118, no. 20 (May 18, 2021). https://doi.org/10.1073/pnas.2024287118.

Cagnina, R. Elaine, Kathryn R. Michels, Alexandra M. Bettina, Marie D. Burdick, Yogesh Scindia, Zhimin Zhang, Thomas J. Braciale, and Borna Mehrad. “Neutrophil-Derived TNF Drives Fungal Acute Lung Injury in Chronic Granulomatous Disease.” The Journal of Infectious Diseases, April 5, 2021. https://doi.org/10.1093/infdis/jiab188.

Wooten, David J., Jorge Gómez Tejeda Zañudo, David Murrugarra, Austin M. Perry, Anna Dongari-Bagtzoglou, Reinhard Laubenbacher, Clarissa J. Nobile, and Réka Albert. “Mathematical Modeling of the Candida Albicans Yeast to Hyphal Transition Reveals Novel Control Strategies.” Preprint. Systems Biology, January 20, 2021. https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1008690

E. Paul, G. Pogudin, W. Qin and R. Laubenbacher, The dynamics of canalizing Boolean networks, Complexity, https://doi.org/10.1155/2020/3687961, 2020.

L. Sordo Vieira, R. Laubenbacher, and D. Murrugarra, Control of Intracellular Molecular Networks Using Algebraic Methods, Bull. Math. Biol., 82(2), 2019.

Aguilar, Boris, Pan Fang, Reinhard Laubenbacher, and David Murrugarra. 2020. “A Near-Optimal Control Method for Stochastic Boolean Networks”. Letters in Biomathematics 7 (1), 67–80.

E. Paul, G. Pogudin, W. Qin and R. Laubenbacher, The dynamics of canalizing Boolean networks, Complexity, https://doi.org/10.1155/2020/3687961, 2020.

Michels KR, Solomon AL, Scindia Y, Vaulont S, Burdick MD, Laubenbacher R, Mehrad B. Aspergillus utilizes heme as an iron source during invasive pneumonia. Submitted to mBio (revising manuscript to resubmit).

Solomon A, Michels K, Laubenbacher R, Scindia Y, Mehrad B. The role of heme uptake and hemopexin in invasive pulmonary aspergillosis. American Thoracic Society Virtual Conference, 2020.

Qu G, Solomon A, Yang N, Lin C, Scindia Y, Mehrad B. Alveolar macrophages protect against acute lung injury. American Thoracic Society Virtual Conference, 2020.

Grants

Multiscale modeling of the battle over iron in invasive lung infection, National Institute of Allergy and Infectious Diseases 1R01AI135128-01

Modular design of multiscale models, with an application to the innate immune response to fungal respiratory pathogens, National Institute of Biomedical Imaging and Bioengineering 1U01EB024501-01, National Science Foundation EAGER #1750183

Models

A description of the model, the workflow of model building, and associated data.

An overview of invasive pulmonary aspergillosis.

The average human inhales hundreds of spores of the A. fumigatus fungus on a daily basis. Despite of this, the vast majority of immunocompetent patients are able to clear the fungus without any seeming effects. A small number of spores are capable of reaching the alveolar sac, where resident macrophages and epithelial cells quickly mount an immune response recruiting various other cells from the circulation such as dendritic cells (DCs), monocyte derived macrophages, and neutrophils. A weakened immune system (e.g. neutropenia) presents a window of opportunity for the fungus to start germinating and breach the epithelial layer, getting into the circulation and causing a systemic infection. Thus, the immune system must act quickly and efficiently to clear the fungus. See e.g. (missing reference).

Nutritional Lung Immunity

Focusing on nutritional immunity

Although the response of the innate immune system to A. fumigatus is multifaceted, the sequestration of iron (in turn starving the fungus from its needed iron) by the host’s immune cells is a crucial response to reduce fungal burden. The A. fumigatus mould is equipped with a sideorophore system that attempts to scavenge iron from its environment. The host immune system is equipped with various mechanisms of iron retention and scavenging. We thus focus on this battle over iron that occurs in the alveolar space between the host and the fungus.

Our approach

Multiscale modeling

The progression of invasive pulmonary aspergillosis is dependent on processes that occur at various spatiotemporal scales. For example, after the spore is inhaled into the alveolar space (tissue scale), receptors on immune cells detect the pathogen and initiate intracellular signaling processes (intracelullar scales) that result in the translation of various cytokines and chemokines. Some of these cytokines then get into the circulation and recruit other immune cells to the site of infection. For example, the cytokine interleukin-6 signals the liver (whole-body scale) to produce the hormone hepcidin which works to reduce iron export from host cells, arresting iron from pathogens. These biophysical processes take effect in ranges from milliseconds to hours. As such, the area of multiscale modeling is becoming increasingly popular for the modeling of diseases that affect multiple scales. Multiscale modeling seeks to model system behavior by modeling the events occurring at various scales and bridging them appropriately.

Agent-based modeling

We are building a 3D agent-based model of the innate immune response to invasive pulmonary aspergillosis with a focus on the “battle over iron” that occurs in the alveolar space between the host and the fungus. Agent-based modeling aims to model how autonomous individuals interact with each other and their computational environment. For us, the individual agents are:

  1. Alveolar macrophages.
  2. Monocyte derived macrophages.
  3. Monocyte derived dendritic cells (DCs).
  4. Neutrophils.
  5. Epithelial cells.
  6. A. fumigatus
  7. Liver cells

Agents behave according to a set of rules that are derived from literature and de novo experiments. For example, the monocyte derived macrophages respond to the sensing of A. fumigatus and respond accordingly to the output of an intracellular Generalized Boolean Network (GBN). The GBN is built from a combination of literature results and analysis of a transcriptomics dataset generated as part of this project.

Calibrating the model with experimental data

In order to build the 3D agent based model, we have performed several in vivo and in vitro experiments.

For example, we co-cultured monocyte derived macrophages with A. fumigatus in a time-series and extracted mRNA for RNA sequencing. We used the transcriptomic data for the calibration of the intracellular model of macrophages. We also performed other protein-level assays such as ELISA and blot assays to calibrate parameters related to extracellular behavior. For the purpose of transparency, all experimental data that was generated with the goal of calibrating the 3D model can be found on this website.

How we build features of the model

We provide an example from our own project to describe how the research process proceeds within our interdisciplinary research group.

Example Goal: Implement a model of the iron-handling behavior of an immune cell in response to A. fumigatus

To answer this question, the following steps could be taken:

  1. An experimentalist might perform a co-cultured experiment of an immune cell type with A. fumigatus.
  2. Various experiments including transcriptomics and ELISA are used to measure transcriptional and protein-level changes.
  3. After sequencing, the data is preprocessed and analyzed by a computational biologist.
  4. Such analysis (such as differential expression analysis, pathway analysis, etc.),together with the protein-level data, is then used to enrich a preliminary intracellular mathematical model built from literature knowledge.
  5. The implementation of this individual-cell model is then incorporated into the tissue-level 3D model.

From this example alone, we see the various data and artifacts that were required and/or generated:

  1. Experimental notebooks describing the protocol of the experiment.
  2. Raw data (e.g. RNA seq fastq files).
  3. Computational scripts analyzing the generated data.
  4. Versions of different software dependencies used in the analysis.
  5. Literature results.
  6. Implementation of mathematical model.

Bibliography

Overview of innate immunity to invasive pulmonary aspergillosis

Acknowledgements of figures.

Many of the individual figures were generated by the scientific illustration team at The Jackson Laboratory.

Software Design

Digital twins in biomedicine need to evolve continuously to represent the current state of knowledge and data. A large-scale implementation of the digital twin paradigm for human health requires the construction and execution of highly complex models, composed of several component models which span multiple spatial and temporal scales. A flexible software development platform is needed that enables multidisciplinary and distributed teams to work together, supports reproducibility, and facilitates the integration of data and component models. Design patterns common to traditional model implementations impair the development of integrative digital twins. Some of these patterns include:

  1. Lack of transparency in the implementation of computational models;
  2. Intertwined component models and simulation processes dependent on each other;
  3. Use of incompatible data structures and computer languages;
  4. Brittle architectures that do not easily accommodate extensions of a model;
  5. Software environments that do not easily support distributed collaboration.

To address these problems we have developed a novel approach based on an open source, highly modularized, computational representation of a digital twin architecture. While the concept of modular design of models and software is well-established, the way modules are assembled generally suffers from the shortcomings listed above. The central principle of the architecture we have developed is the separation of computational algorithms for the different dynamic processes from each other, eliminating dependencies that make model modifications and extensions cumbersome or impossible in complex models. It also features the separation of computational algorithms from data, in the sense that all data describing the global model state, including model parameters, are separate from the individual computational modules, in a ‘hub-and-spoke’ transparent architecture designed to facilitate extension and modification. This approach differs fundamentally from the conventional approach to building and simulating such models in biomedicine.

The core principle underlying the highly modularized architecture we propose here is to treat each dynamic biological process in the model, or related collections of processes, as a separate module of the digital twin. In a biological context a molecular module might contain the algorithms for diffusion and transport of that molecule, while a cellular model could contain sub-models related to that cell’s function. The individual modules are only indirectly connected by communicating through a central data structure, the global model state, rather than passing data to each other directly.

This prevents any direct dependencies between the computational portion of modules, a key feature that enables the model to be readily extended or modified. The global model state is the repository for all data describing the state of the simulated model at a given point in time, including any information about the underlying physical structure, if included, and variable states of all computational models in the modules. All computational algorithms, on the other hand, are contained in the modules, providing a clear separation between the model and the data on which it operates during model simulation. The resulting computational structure naturally separates model components so that they may be validated by the distinct modalities natural for each of the dynamic processes in the model, facilitating continued model refinement and personalization. Our implementation contains four components:

  1. a runtime configuration file,
  2. a global model state,
  3. modules, and
  4. a simulation framework that controls simulation runtime and provides data structures and algorithms useful for the development of new modules.

These four components and their relationship are represented in Figure 1, providing a structure for the model components and their dependencies.

The modular design implementation contains five components:

  1. The runtime configuration file, that contains all configuration and parameter settings for a given simulation run (config.ini).
  2. A simulation solver, that reads the configuration (config.ini) and constructs, initializes, and advances the simulation in time by executing each module according to its inherent time scale.
  3. The model state contains all data describing the state of the model at a given point in time, including any physical space geometry, and states of model objects. In this example, the model includes a spatial component. The model state is a contiguous block of memory as shown by the partitioned rectangle, with the hierarchical Python referencing syntax shown to the left of the representation.
  4. Each module consists of a computational model that takes all input data from the model state and stores none itself.
  5. These modules extend classes provided as part of the simulation framework, ModuleState and ModuleModel which handle the connection to the simulation solver and model state access so the developer only needs to consider the biological additions to the model. Extending the ModuleState results in the fields defined in the extending class being appended to the model state. The initialize and advance functions in the ModuleModel extension will be called by the simulation solver so the module can participate in the simulation.

The source code for the simulator is licensed under the Apache 2.0 License and is available on Github.

Data

As part of the model building, calibration, and validation effort, various heterogeneous data are generated. As new publications appear, a brief description with appropriate links will be included here.

  • Time series characterization of Aspergillus infected mice.
    • To characterize the dynamics of cytokine production, as well as cell recruitment into the lung, we measured CXCL2, IL6 protein levels (ELISAs), and performed flow cytometry to characterize neutrophil and monocyte recruitment into the lung 12 hours post infection, 1dpi, 2dpi, 3dpi. This data was utilized to validate the main computational model presented in  https://doi.org/10.1101/2021.06.08.447590, and the processed data can be found here. Details of methodology can be found in https://doi.org/10.1101/2021.06.08.447590.
  • RNA sequencing data of macrophages exposed to A. fumigatus
    • To characterize the transcriptional response to A.fumigatus, we performed RNA sequencing of human monocyte derived macrophages exposed to A. fumigatus conidia in a time series fashion after 0 hours, 2 hours, 4 hours, 6 hours, and 8 hours of co-culture (See https://www.biorxiv.org/content/10.1101/2022.01.24.477648v1 for details). This data was utilized to create an intracellular network model of the altered iron-handling phenotype of macrophages in response to A. fumigatus. The gene count matrix, as well as all scripts utilized to preprocess sequencing data and mathematical model can be found here.

The workflow to generate the model is summarized below.