Dynamics Modeling as a Weapon to Defend Ourselves Against Threats from Infectious Diseases and Bioterrorist Attacks

1
Dynamics Modeling as a Weapon to Defend Ourselves
Against Threats from Infectious Diseases and
Bioterrorist Attacks
Hulin Wu, Ph.D., Professor Director, Center for
Biodefense Immune Modeling Chief, Division of
Biomedical Modeling and Informatics Department
of Biostatistics Computational
Biology University of Rochester Medical Center

• SAMSI, February 25, 2011

2
Outline

• Introduction Impact of Infectious Diseases to
  Public Health
• Dynamic Modeling for HIV
• Dynamic Modeling for Influenza
• Conclusions and Discussions
• Acknowledgement

3
SARS Pandemic November 1, 2002-July 31, 2003

• Total Cases 8096
• Death 774
• Death rate 9.6
• 29 countries/regions
• USA 27 cases (no death)

4
Bird Flu (H5N1) Epidemics in Human

• Total Cases 285
• Death 170
• Death Rate 59.6
• 12 countries/regions

5
(No Transcript)
6
Flu Pandemics History

• 1918 Spanish flu (H1N1) pandemic kill 20-100
  million people worldwide
• 1957 Asian Flu (H2N2) 1-4 million infections
  worldwide, 69,800 deaths in the US
• 1968 Hong Kong Flu (H3N2) 500,000 infections
  worldwide, 33,000 deaths in the US

7
An Emergency Hospital for Influenza Patients
8
Two routes to a pandemic
H5N1
H5N1
Species barrier
H5N1
H3N2
9
Annual Influenza Epidemics around the World

• 5-15 of the population affected
• 3-5 million cases of severe illness
• 250,000-500,000 deaths around the world

10
Current Estimates of the Yearly Disease Burden of
Influenza in the US
40,000 100,000 40,000,000 4,000,000,000 8,000,000,
000
Deaths - Hospitalizations - Illnesses - Direct
costs () - Indirect costs () -
11
Global HIV/AIDS Epidemics 2006 Update
12
Global HIV/AIDS Epidemics 2006 Update
13
Global HIV/AIDS Epidemics 2006 Update
14
New HIV Infection Rate in 2006

• 8 infections per minute
• 458 infections per hour

15
Defend Ourselves Why and How to Use
Mathematics/Statistics as a Weapon?

• Understand pathogenesis of infection by
  infectious agents
• Identify therapeutic targets for intervention
• Design and evaluate the effects of treatments and
  other intervention/prevention strategies

16
Example HIV/AIDS Modeling

• 1st AIDS case reported in late 1970s
• AIDS virus discovered in 1983, named HTLV
• AIDS virus renamed as HIV in 1986
• HIV dynamics models in late 1980s Merrill 1987
  Mclean 1988 Anderson and May 1989 Perelson 1989
• HIV dynamics models for clinical studies David
  Ho and Alan Perelson (Nature 1995 Science 1996
  Nature 1997)
• My research in HIV dynamics modeling 1997-

17
Ho et al, Nature 1995
18
Ho et al., Nature 1995

• 20 HIV-1 infected patients
• A new antiviral drug a protease inhibitor,
  ABT-538 (Ritonavir)
• Observations Viral load declined exponentially
  in 2 weeks

19
Ho et al., Nature 1995
20
Ho et al., Nature 1995

• Tap-Tank Model
• Solution with perfect treatment P0
• Fit a linear regression model
• c viral clearance rate
• 1/c Mean life-span of HIV virus
• ln(2/c) Half-life of HIV virus

21
Ho et al., Nature 1995

• Estimate of c 0.34 (range 0.21 to 0.54)
• Half-life of HIV virus 2.1 (range1.3 to 3.3)
  days
• Daily production and clearance rate of HIV virus
  0.68x109 (range 0.05 to 2.07x109) virions

22
(No Transcript)
23
Perelson et al. and Ho, Science 1996

• A more complicated model
• Solution
• Clinical data 5 HIV patients

24
Perelson et al. and Ho, Science 1996

• Estimate of c 3.07
• Estimate of d 0.49
• Half-life of virus 0.24 (about 6 hours).
• Half-life of infected cells 1.55 days

25
(No Transcript)
26
Perelson et al. and Ho, Nature 1997

• Short-lived infected cells t1/21.1 days
• Long-lived inected cells t1/214.1 days
• Latently infected cells t1/28.5 days

27
(No Transcript)
28
My Research HIV and Influenza

• HIV/AIDS Use differential equation models to
  study antiretroviral treatment effects and
  treatment strategies in HIV/AIDS research
• Influenza Use differential equation models to
  study immune response to influenza infections and
  vaccinations

29
Dynamic Models for AIDS Treatment

• HIV Viral Dynamic Model in Vivo
• Viral fitness is related to antiviral drug
  efficacy
• Correlate the lab data to clinical data via the
  proposed model

30
(No Transcript)
31
(No Transcript)
32
(No Transcript)
33
Influenza Project

• Center for Biodefense Immune Modeling funded by
  NIH from 2005-2015 with 21.9 million in total
• To develop mathematical models and computer
  simulation tools to simulate immune response to
  influenza virus
• To design and conduct experiments to validate the
  mathematical models and simulation tools
• To expect that our modeling and simulation tools
  can help to rapidly design drugs or vaccines to
  fight against new and possibly engineered viruses

34
(No Transcript)
35
A Complex Dynamic System for Influenza Infection
Lee et al 2009 (J. of Virology)
6/2/10 Annual Meeting
36
Lung Compartment Sub-Model
37
Lung Compartment Sub-Model
Collected data
Fig 2. Cytokine secreting CD8 T cells per murine
lung

• Fig 1. HKX31 EID50/ml titers per murine lung

38
Lung Compartment Sub-Model
Collected data

• Fig 3. Smoothed data for IgG and IgM pg/ml murine
  serum

39
Model Fitting Results
40
Estimation Result Summary

• The CTL effect 6.4x10-5/day. Shorten the
  half-life of infected cells from 1.16 days to
  0.59 days in average.
• The death rate of infected cells due to effects
  other than CTL is 0.16/day which is 26 of the
  death rate during the first 5 days
• Antibody effect IgM dominates the clerance of
  viral particles with a rate about 4.4/day.
  Shorten the half-life from 4 hours to 1.8 minutes
  in average
• Antibody IgG not significant
• The clearance rate of viral particles due to
  factors other than antibody effect very small.

41
Immune Response Kinetics Useful

• Identify antiviral drug and vaccine targets
• Understand virulent viruses and their properties
• Prepareness

42
DEDiscover

• Software tool for developing, exploring, and
  applying differential equation models.
• Key Features
• ODE DDE Models
• Real-time interactive simulation
• Data fitting (Estimation)
• Clean, Cross-platform GUI
• High Quality Plots
• Ver 2.5b freely available
• https//cbim.urmc.rochester.edu/software/dedis
  cover

42
2010-06-02
CBIM DEDiscover Software
43
Conclusions and Discussions

• Efficiently fight against infectious diseases and
  bioterrorism
• Need global effort with efficient collaborations
  and communications
• Need efficient collaborations and communications
  among inter-disciplinary scientists
• Need long-term effort and huge resources
• Use any weapons available to defend ourselves
  including mathematics, computer and statistics
• Dynamics modeling an important weapon
• Can we defend ourselves?

44
(No Transcript)
45
Acknowledgments

• NIAID/NIH grant R01 AI 055290 AIDS Clinical
  Trial Modeling and Simulations
• NIAID/NIH grant N01 AI50020 Center for
  Biodefense Immune Modeling
• NIAID/NIH grant P30 AI078498 Developmental
  Center for AIDS Research
• NIAID/NIH grant R21 AI078842 Analysis of
  Differential Resistance Emergence Risk for
  Differential Treatment Applications
• NIAID/NIH grant RO1 AI087135 Estimation Methods
  for Nonlinear ODE Models in AIDS Research