NHLBI Special Emphasis Panel
New Approaches to Complex Biological Problems:
Inflammation and Risk Assessment

Minutes of Meeting

Bethesda, Maryland

September 19, 1997

Present:

Dr. Martin opened the meeting at 8:00 a.m. with introductions. Dr. Peavy explained the purpose of the meeting and reviewed the charge to the Panel. The panel members were informed about the requirement to read and sign the conflict of interest statement. All members signed the statement.

The panel was charged with the following tasks:

1. Identify the most important needs and opportunities and promising possibilities relating to mathematical modeling of complex biological processes that might advance lung research in the areas of TB/AIDS and ARDS. This process includes review of current scientific concepts and making recommendations concerning areas of scientific opportunity.
   
   
2. Formulate major questions and issues in the areas identified.
   
   
3. Identify steps needed to address the questions and issues. Indicating, if possible which are steps are most urgent, where obstacles to progress exist and what is needed to overcome these.
   
   
4. Discuss training issues when these are applicable.
   
   

Each panel member gave a brief presentation that included background information, new concepts and data on various aspects of inflammation in the lung and modeling. This was followed by delineation of gaps in our understanding of those areas as well as recommendations to help to fill those gaps. First, clinical and biological presentations outlined issues associated with TB and AIDS, the roles of lung macrophages, lymphocytes, chemokines and their receptors in HIV, disease, and acute respiratory distress syndrome (ARDS). This was followed by an overview of the role of mathematical modeling in medicine, examples of models in other organ systems, examples of statistical models and modeling applied to TB-HIV interactions. A short discussion followed each presentation. In the afternoon members of the panel prepared a list of specific recommendations.

Tuberculosis and Opportunistic Infections

Dr. Martin discussed mechanisms of disease transmission for opportunistic pulmonary Infections. He concluded that for respiratory pathogens, the pattern of transmission and the incidence of disease are not well linked. Better knowledge of how epidemics arise and why specific individuals become affected would greatly enhance our understanding of the disease and potentially its control.

We need to learn more about how epidemics arise and why specific individuals become affected. This would greatly enhance our understanding of the disease and potentially its control.

Lung Macrophages and Lymphocytes in HIV Disease

Dr. Twigg reviewed normal pulmonary immune responses, how pulmonary immune responses are thought to be altered in HIV infection, and how mathematical modeling might be applied to dissect specific abnormalities in these patients.

HIV progression is associated with a loss of diversity in potential immune responses to invading pathogens leaving the lung in a state of chronic cellular activation (macrophages and T cells) and resulting in chronic inflammation. Mathematical modeling of how these changes occur in the lung would provide important clues to the mechanisms involved.

HIV in the Lung: Chemokines and Receptors

Dr. Collman reviewed how HIV infects cells in the lung, how this affects local immune responses, interactions between chemokines, chemokine receptors and HIV, and what this may imply for progression of disease in the lung.

During HIV infection of the lung what determines whether pulmonary inflammation or immune incompetence predominates? How do local pulmonary and systemic immune changes interact to determine the level of pulmonary immune competence? What roles are played by chemokine receptors CCR5, CXCR3, CCR2 and others? How do viral, host genetic, immune, and environmental factors interact to determine systemic viral load and disease progression? How do these factors interact to regulate viral burden in the lung?

Acute Respiratory Distress Syndrome

Dr. Buchman discussed multiple organ dysfunction syndrome (MODS) as a consequence of inflammation and a cause of death. He proposed novel strategies for analysis and possibly for intervention.

Using MODS as an example, linear models are not adequate representations of human pathophysiology. The stability of biological systems appears to be a function of their interconnectedness. There is a pressing need to develop and test complex systems models as representations of human pathophysiology to identify therapeutic strategies which will promote transitions to basal, healthy physiologic states.

Overview of math modeling in medicine

Dr. Kirschner provided a general overview on the role of mathematical modeling in medicine. Models are an abstraction, not a reflection of reality and must achieve a balance between simplicity and detail. They are based on available knowledge and may provide insight into interactions. However the model's results are based on the original assumptions. Mathematical models can be used to compare and contrast existing controversies, extend intuition about biological process, and can be done without data. Creating models generally requires input from many disciplines. They are useful in testing the understanding of a system, explaining phenomena, when experimental resolution is difficult, expensive or impossible (based on cost, physical, ethical constraints, etc.). Models may be able to provide details or understanding about an experimental system. Models have made an impact in many areas of science, e.g. neurophysiology, cardiology, epidemiology, health economics, immunology, pathogenesis of disease (cancer), physiology, radiology and other types of imaging. Differential equations are a major component of the models. Parameters in models represent rate constants of the processes. They can help identify key interactions, and allow for sensitivity and uncertaintly analyses as well as aid in obtaining bifurcation information. These methods can lead to a greater understanding of the important porcesses of the biological system. Because of the nature of these parameters, it is easy to perturb the system to explore changes in the biological assumptions. Applications which are already underway include models for control strategies of TB epidemics and models to understand the dynamics of HIV infection including, production, turnover, half-life, and possible reservoirs of infection. Modeling HIV changed concepts about viral growth and lead to the realization of the highly dynamic growth of the virus. Problems inherent in the use of models include validation, gaps in communication, difficulty in obtaining appropriate expertise for reviews, lack of funding. Research dealing with models is usually published in journals not read by basic scientists and clinicians.

Models can contribute to the understanding of complex biological processes and may offer new perspectives, clarify ideas, and raise new questions. They are constructed to test ideas and are process driven. Validation of the models is an important issue. A major obstacle is lack of communication between the clinical and basic scientists and theoretical scientists, who develop models.

Examples of models in organ systems other than lung

Dr. Mejia provided a framework for thought using examples from renal models.

Mathematical modeling in other organ systems has provided insights at the molecular, cellular and organ level. These include protein identification and information on structure and function, improved imaging techniques and understanding of blood flow. An essential factor in creating models is close collaboration between clinicians or bench scientists and biomathematicians.

Statistical Models

Dr. Albert discussed basic strategies used in developing statistical models for characterizing a disease process. He outlined the purposes of these models, the issues in making statistical inferences about model parameters, and the types of questions that can be addressed with these approaches. Statistical modeling has been used to model disease processes in all areas of medicine including AIDS, multiple sclerosis, epilepsy, bipolar disorder, and asthma.

Statistical modeling has been used to model disease processes in all areas of medicine including AIDS, multiple sclerosis, epilepsy, bipolar disorder, and asthma. Examples presented included a model of relapsing-remitting multiple sclerosis and a model of monotonic responses.

Mathematical Modeling of TB-HIV Interactions

Dr. Kirschner presented a model she developed of TB-HIV interaction.

A mathematical model of dual infection with tuberculosis and HIV predicts interactions. The predictions appear to fit well with clinical findings.

RECOMMENDATIONS:

• The panel explored the role of modeling techniques and concluded that these are a resource that can be used to enhance experimental design especially to define important variables and questions, but validation remains an issue.
  
  
• The chief priority is to improve communication and cooperation between modelers clinical and basic scientists.

Suggestions for achieving this included integration of pilot studies with larger grants, use of meetings, journal supplements and training to alert scientists to the existence and the potential of techniques that might be helpful in the study of a wide variety of lung diseases and processes with an inflammatory component.

These include HIV related lung disease, tuberculosis and other lung infections, interstitial lung disorders, obstructive airway diseases, trauma, lung injury and ARDS as well as inhalation injury secondary to environmental toxicities.

The meeting was adjourned at 3:00 p.m. on September 19, 1997

CERTIFICATION

I hereby certify that the foregoing minutes are accurate and complete.

William J. Martin, M.D.
Chairperson

Hannah H. Peavy, M.D.
Executive Secretary

Last Updated April 2011