A look at mathematical models of pathogenesis with Josh Schiffer

Vaccine and Infectious Disease Division

A look at mathematical models of pathogenesis with Josh Schiffer

Dr. Josh Schiffer

VIDI senior fellow Dr. Josh Schiffer first became interested in infectious disease as a college student studying abroad and working in Vietnam, where he was involved with community service for those at high risk of HIV infection. He now lives in West Seattle with his wife Chihana and two boys, Kenzo, 4, and Kaito, 3.

Computational models in herpes simplex virus-2 (HSV-2) research are revealing new insights into how this virus maintains its long-term balance of latency and reactivation inside the human body.  By using math to model how the virus spreads in its host, VIDI senior fellow Dr. Josh Schiffer found some surprising results – including the finding that very small amounts of virus may be constantly released from neurons, where the virus normally lies dormant.  This tiny trickle of virus is enough to trigger HSV-2 reactivation episodes, which include genital lesions and viral shedding, where viruses break free of infected cells en masse.

“It’s like a continuum of tiny sparks that occasionally set off larger fires,” Schiffer said.

HSV-2 is a common virus, infecting approximately 16% of Americans. For many, HSV-2 poses little danger, causing either genital ulcers or no symptoms.  After initial infection, in most cases the virus is rapidly cleared from the genital skin and remains latent in neurons near the genitals.  Yet HSV-2 is considered a major public health threat, in part because HSV-2 infection greatly increases a person’s chances for acquiring HIV, or if they are already HIV positive, the likelihood of transmitting HIV through sexual activity.  Additionally, the virus can be dangerous for immunocompromised people, occasionally causing pneumonia and hepatitis.  Babies born to HSV-2 positive mothers are at risk of acquiring neonatal herpes – transmission at birth is rare but often fatal to the infant.

Modeling can be a useful tool for studying pathogens because of the nature of infectious disease, Schiffer said.  The spread of a virus from person to person or from cell to cell is a nonlinear process, meaning that the chances of a person or a cell becoming infected can change depending on how many infected people or cells they come into contact with.

“That fundamental concept can only be addressed with certain kinds of equations,” Schiffer said.  Epidemic modeling, which focuses on disease spread among multiple people, is a well-established field, but models of infection within a single person are lacking for many diseases, he said.

Schiffer and his colleagues created models based on clinical data gathered on HSV-2 over the last decade by groups led by VIDI member Dr. Larry Corey and affiliate investigator Dr. Anna Wald.  Schiffer created a stochastic model for HSV-2 infection in which the probability of infection of a given cell is calculated at very small intervals. He then ran multiple computational simulations of an HSV-2 infection event  to reveal general trends of infection.

Using known parameters based on experimental results, Schiffer asked how often and how much virus is released from neurons. He found that the clinical data could only be reproduced in his model when the parameters were set so that a tiny amount of virus was constantly released.  This finding has clear clinical implications, Schiffer said, as antiviral drugs such as acyclovir are known to reduce HSV-2 reactivation and transmission, but don’t knock them down to zero. 

“This finding tentatively points to one of the challenges of treating HSV-2, which is that you don’t need a lot of virus to come from the neurons to stimulate the system,” Schiffer said.  “Additionally, it’s not even clear if the drug enters the neuron, so if it’s only working peripherally, it would be a very tough system to shut down.”

The model also indicates that more severe lesions or shedding episodes with higher amounts of virus result in recruitment of a larger number of T cells.  When taken together with previous work by VIDI staff scientist Jia Zhu and others showing that CD4+ and CD8+ T cells persist in the genitalia for months after HSV-2 lesions are cleared, these results could explain why HSV-2 infection increases the risk of HIV acquisition.  Since HIV primarily infects CD4+ T cells, the increased number of these cells in the genitalia of HSV-2 infected individuals could make it easier for HIV to gain a toehold.  Clinical tests showed that acyclovir did not decrease HIV transmission or acquisition, even though it reduces HSV-2 reactivation rates.

“In order to truly test the hypothesis that controlling HSV can decrease HIV acquisition and transmission, you would need to completely eliminate shedding in the genital tract,” Schiffer said.  “The intervention that we used, while the best available, did not shut down herpes shedding.  So it really brings us back to the drawing board: How can we eliminate shedding?”

Schiffer, who just received a fundable score on a K23 grant that will allow him to continue his work at the Center for the next five years, now plans to incorporate drug activity into his models, and ask what happens to shedding at different drug levels.  He is working with postdoctoral fellows Abigail Waucher and Ramzi Alsallaq to figure out how much HSV-2 shedding is required for transmission, and how the virus spreads from cell to cell during a reactivation event.  In the future, Schiffer hopes to apply his models to such diseases as influenza and dengue by collaborating with experts on these infections.  He also hopes to establish collaborations in which experimental data could be generated specifically for the model.