Cordelia Ziraldo

A Two-Compartment Mathematical Model of Endotoxin-induced Inflammatory and Physiologic Alterations in Swine

Objective:To gain insights into individual variations in acute inflammation and physiology. Design:Large-animal study combined with mathematical modeling. Setting:Academic large-animal and computational laboratories. Subjects:Outbred juvenile swine. Interventions:Four swine were instrumented and subjected to endotoxemia (100 µg/kg), followed by serial plasma sampling. Measurements and Main Results:Swine exhibited various degrees of inflammation and acute lung injury, including one death with severe acute lung injury (PaO2/FIO2 ratio &mgr;200 and static compliance &mgr;10 L/cm H2O). Plasma interleukin-1&bgr;, interleukin-4, interleukin-6, interleukin-8, interleukin-10, tumor necrosis factor-&agr;, high mobility group box-1, and NO2-/NO3- were significantly (p &mgr; .05) elevated over the course of the experiment. Principal component analysis was used to suggest principal drivers of inflammation. Based in part on principal component analysis, an ordinary differential equation model was constructed, consisting of the lung and the blood (as a surrogate for the rest of the body), in which endotoxin induces tumor necrosis factor-&agr; in monocytes in the blood, followed by the trafficking of these cells into the lung leading to the release of high mobility group box-1, which in turn stimulates the release of interleukin-1&bgr; from resident macrophages. The ordinary differential equation model also included blood pressure, PaO2, and FIO2, and a damage variable that summarizes the health of the animal. This ordinary differential equation model could be fit to both inflammatory and physiologic data in the individual swine. The predicted time course of damage could be matched to the oxygen index in three of the four swine. Conclusions:The approach described herein may aid in predicting inflammation and physiologic dysfunction in small cohorts of subjects with diverse phenotypes and outcomes.

A dynamic view of trauma/hemorrhage-induced inflammation in mice: principal drivers and networks.

Background Complex biological processes such as acute inflammation induced by trauma/hemorrhagic shock/ (T/HS) are dynamic and multi-dimensional. We utilized multiplexing cytokine analysis coupled with data-driven modeling to gain a systems perspective into T/HS. Methodology/Principal Findings Mice were subjected to surgical cannulation trauma (ST) ± hemorrhagic shock (HS; 25 mmHg), and followed for 1, 2, 3, or 4 h in each case. Serum was assayed for 20 cytokines and NO2 −/NO3 −. These data were analyzed using four data-driven methods (Hierarchical Clustering Analysis [HCA], multivariate analysis [MA], Principal Component Analysis [PCA], and Dynamic Network Analysis [DyNA]). Using HCA, animals subjected to ST vs. ST + HS could be partially segregated based on inflammatory mediator profiles, despite a large overlap. Based on MA, interleukin [IL]-12p40/p70 (IL-12.total), monokine induced by interferon-γ (CXCL-9) [MIG], and IP-10 were the best discriminators between ST and ST/HS. PCA suggested that the inflammatory mediators found in the three main principal components in animals subjected to ST were IL-6, IL-10, and IL-13, while the three principal components in ST + HS included a large number of cytokines including IL-6, IL-10, keratinocyte-derived cytokine (CXCL-1) [KC], and tumor necrosis factor-α [TNF-α]. DyNA suggested that the circulating mediators produced in response to ST were characterized by a high degree of interconnection/complexity at all time points; the response to ST + HS consisted of different central nodes, and exhibited zero network density over the first 2 h with lesser connectivity vs. ST at all time points. DyNA also helped link the conclusions from MA and PCA, in that central nodes consisting of IP-10 and IL-12 were seen in ST, while MIG and IL-6 were central nodes in ST + HS. Conclusions/Significance These studies help elucidate the dynamics of T/HS-induced inflammation, complementing other forms of dynamic mechanistic modeling. These methods should be applicable to the analysis of other complex biological processes.

Analysis of Serum Inflammatory Mediators Identifies Unique Dynamic Networks Associated with Death and Spontaneous Survival in Pediatric Acute Liver Failure

Background Tools to predict death or spontaneous survival are necessary to inform liver transplantation (LTx) decisions in pediatric acute liver failure (PALF), but such tools are not available. Recent data suggest that immune/inflammatory dysregulation occurs in the setting of acute liver failure. We hypothesized that specific, dynamic, and measurable patterns of immune/inflammatory dysregulation will correlate with outcomes in PALF. Methods We assayed 26 inflammatory mediators on stored serum samples obtained from a convenience sample of 49 children in the PALF study group (PALFSG) collected within 7 days after enrollment. Outcomes were assessed within 21 days of enrollment consisting of spontaneous survivors, non-survivors, and LTx recipients. Data were subjected to statistical analysis, patient-specific Principal Component Analysis (PCA), and Dynamic Bayesian Network (DBN) inference. Findings Raw inflammatory mediator levels assessed over time did not distinguish among PALF outcomes. However, DBN analysis did reveal distinct interferon-gamma-related networks that distinguished spontaneous survivors from those who died. The network identified in LTx patients pre-transplant was more like that seen in spontaneous survivors than in those who died, a finding supported by PCA. Interpretation The application of DBN analysis of inflammatory mediators in this small patient sample appears to differentiate survivors from non-survivors in PALF. Patterns associated with LTx pre-transplant were more like those seen in spontaneous survivors than in those who died. DBN-based analyses might lead to a better prediction of outcome in PALF, and could also have more general utility in other complex diseases with an inflammatory etiology.

Central Role for MCP-1/CCL2 in Injury-Induced Inflammation Revealed by In Vitro , In Silico , and Clinical Studies

The translation of in vitro findings to clinical outcomes is often elusive. Trauma/hemorrhagic shock (T/HS) results in hepatic hypoxia that drives inflammation. We hypothesize that in silico methods would help bridge in vitro hepatocyte data and clinical T/HS, in which the liver is a primary site of inflammation. Primary mouse hepatocytes were cultured under hypoxia (1% O2) or normoxia (21% O2) for 1–72 h, and both the cell supernatants and protein lysates were assayed for 18 inflammatory mediators by Luminex™ technology. Statistical analysis and data-driven modeling were employed to characterize the main components of the cellular response. Statistical analyses, hierarchical and k-means clustering, Principal Component Analysis, and Dynamic Network Analysis suggested MCP-1/CCL2 and IL-1α as central coordinators of hepatocyte-mediated inflammation in C57BL/6 mouse hepatocytes. Hepatocytes from MCP-1-null mice had altered dynamic inflammatory networks. Circulating MCP-1 levels segregated human T/HS survivors from non-survivors. Furthermore, T/HS survivors with elevated early levels of plasma MCP-1 post-injury had longer total lengths of stay, longer intensive care unit lengths of stay, and prolonged requirement for mechanical ventilation vs. those with low plasma MCP-1. This study identifies MCP-1 as a main driver of the response of hepatocytes in vitro and as a biomarker for clinical outcomes in T/HS, and suggests an experimental and computational framework for discovery of novel clinical biomarkers in inflammatory diseases.

SPARK: A Framework for Multi-Scale Agent-Based Biomedical Modeling

Multi-scale modeling of complex biological systems remains a central challenge in the systems biology community. A method of dynamic knowledge representation known as agent-based modeling enables the study of higher level behavior emerging from discrete events performed by individual components. With the advancement of computer technology, agent-based modeling has emerged as an innovative technique to model the complexities of systems biology. In this work, the authors describe SPARK Simple Platform for Agent-based Representation of Knowledge, a framework for agent-based modeling specifically designed for systems-level biomedical model development. SPARK is a stand-alone application written in Java. It provides a user-friendly interface, and a simple programming language for developing Agent-Based Models ABMs. SPARK has the following features specialized for modeling biomedical systems: 1 continuous space that can simulate real physical space; 2 flexible agent size and shape that can represent the relative proportions of various cell types; 3 multiple spaces that can concurrently simulate and visualize multiple scales in biomedical models; 4 a convenient graphical user interface. Existing ABMs of diabetic foot ulcers and acute inflammation were implemented in SPARK. Models of identical complexity were run in both NetLogo and SPARK; the SPARK-based models ran two to three times faster.

A Computational, Tissue-Realistic Model of Pressure Ulcer Formation in Individuals with Spinal Cord Injury.

People with spinal cord injury (SCI) are predisposed to pressure ulcers (PU). PU remain a significant burden in cost of care and quality of life despite improved mechanistic understanding and advanced interventions. An agent-based model (ABM) of ischemia/reperfusion-induced inflammation and PU (the PUABM) was created, calibrated to serial images of post-SCI PU, and used to investigate potential treatments in silico. Tissue-level features of the PUABM recapitulated visual patterns of ulcer formation in individuals with SCI. These morphological features, along with simulated cell counts and mediator concentrations, suggested that the influence of inflammatory dynamics caused simulations to be committed to “better” vs. “worse” outcomes by 4 days of simulated time and prior to ulcer formation. Sensitivity analysis of model parameters suggested that increasing oxygen availability would reduce PU incidence. Using the PUABM, in silico trials of anti-inflammatory treatments such as corticosteroids and a neutralizing antibody targeted at Damage-Associated Molecular Pattern molecules (DAMPs) suggested that, at best, early application at a sufficiently high dose could attenuate local inflammation and reduce pressure-associated tissue damage, but could not reduce PU incidence. The PUABM thus shows promise as an adjunct for mechanistic understanding, diagnosis, and design of therapies in the setting of PU.

SPARK: a framework for multi-scale agent-based biomedical modeling

Multi-scale modeling of complex biological systems remains a central challenge in the systems biology community. A method of dynamic knowledge representation known as agent-based modeling enables the study of higher level behavior emerging from discrete events performed by individual components. In this work, we describe SPARK (Simple Platform for Agent-based Representation of Knowledge), a framework for agent-based modeling specifically designed for systems-level biomedical model development. SPARK is a standalone application written in Java. It provides a user-friendly interface, and a simple programming language for developing Agent-Based Models (ABMs). SPARK has the following features specialized for modeling biomedical systems: 1) continuous space that can simulate real physical space; 2) flexible agent size and shape that can represent the relative proportions of various cell types; 3) multiple spaces that can concurrently simulate and visualize multiple scales in biomedical models; 4) a convenient graphical user interface. Existing ABMs of diabetic foot ulcers and acute inflammation were implemented in SPARK. Models of identical complexity were run in both NetLogo and SPARK; the SPARK-based models ran two to three times faster. We are currently utilizing SPARK to develop multi-scale inflammation models in diverse settings such as cancer, viral infection, and spinal cord injury.

Integrating Data-Driven and Mechanistic Models of the Inflammatory Response in Sepsis and Trauma

Inflammation can drive both homeostasis and disease via dynamic, multiscale processes. The inflammatory response can be studied using multiplexed platforms, but there is no straightforward means by which to deal with the consequent “data deluge” in order to glean basic insights and clinically useful applications. Systems approaches, including data-driven and mechanistic computational modeling, have been employed in order to study the acute inflammatory response in the settings of trauma/hemorrhage and sepsis. Through combined data-driven and mechanistic modeling based on such “meso-dimensional” datasets, computational models of acute inflammation applicable to multiple preclinical species as well as humans were generated. A key hypothesis derived from these studies is that inflammation may be regulated via positive feedback loops that control switching between beneficial and detrimental inflammatory responses. Self-resolving inflammation may occur when specific signals feedback in a positive fashion to drive anti-inflammatory responses, while proinflammatory signals remain below certain thresholds. In contrast, self-amplifying, detrimental inflammation may occur when different signals feedback in a positive fashion to drive proinflammatory responses, setting in motion the positive feedback loop of inflammation → tissue damage/dysfunction → inflammation driven by damage-associated molecular pattern molecules. These insights may drive a future generation of targeted, personalized therapies for acute inflammation.

Abstract 102: inflammatory mediators modulate alloreactive T cell susceptibility to immune-regulation in reconstructive transplantation.

Saami Khalifian, BA1; Yawah Nicholson, BS2; Cordelia Ziraldo, PhD2; Ravi Starzl, PhD3; Stefan Schneeberger, MD1; Angus Thomson, PhD, DSc2; Gerald Brandacher, MD1; Yoram Vodovotz, PhD2; Giorgio Raimondi, PhD1 1Johns Hopkins University School of Medicine, Dept of Plastic & Reconstructive Surgery, Baltimore, MD, University of Pittsburgh Medical Center, Starzl Transplantation Institute, Pittsburgh, PA, Language Technologies Institute, Carnegie Mellon University, Pittsburgh, PA

Agent-Based Models of Wound Healing

Wounds can arise either from a disease process or are surgically created as part of therapy. The healing of damaged tissue is a fundamental biological process, involving a complex set of cellular and molecular components acting within a specific spatial context. Impairment or aberration of the wound healing process is a considerable source of morbidity and mortality, likely only to increase in clinical significance given an aging population. Despite considerable advances in the mechanistic knowledge of wound healing, as with all complex biological processes converting that knowledge into effective therapeutics is a substantial translational challenge. As result, wound healing has been a major focus in the field of Translational Systems Biology, and, in particular, been a subject of agent-based mechanistic computational modeling. The intuitive mapping between biological knowledge and the rules in an agent-based model (ABM), the ability of an ABM to readily represent stochastic processes, and the inherent spatial representation capability of an ABM all facilitate the utilization of this method in the practice of Translational Systems Biology. Presented herein are a series of ABMs of wound healing that demonstrate the translational potential and utility of this methodology in advancing the rational development of wound healing therapeutics.