HIV: New math model may help to predict strains for vaccine development
A major challenge in the development of an HIV vaccine is the fact that the virus is continually evolving, and it is very difficult to predict which direction it will take. Now, a new study shows how ideas from two disparate fields - stock price prediction and particle diffusion in liquids - can help. Inspired by the financial model, the researchers develop a diffusion-based tool that accurately predicted how an HIV surface protein evolved in a patient population over 30 years.
The research is the work of a team from the Carver College of Medicine at the University of Iowa in Iowa City and is published in the journal PLOS Biology.
HIV spreads through the exchange of certain body fluids such as blood, semen, and vaginal secretions. It can also spread through breast milk. Once it enters the body, HIV attacks the immune system, targeting T cells in particular.
If untreated, HIV carries on destroying T cells, gradually weakening the body's ability to defend itself against infection and disease. Eventually, it progresses to an advanced stage of infection known as AIDS.
According to the World Health Organization (WHO), HIV has infected more than 70 million and claimed the lives of around 35 million people worldwide.
Since the start of the AIDS epidemic in the 1980s, more than 1.2 million people have been diagnosed with the disease in the United States.
There is currently no effective cure or vaccine for HIV or AIDS, but with proper medical care in the form of antiretroviral therapy, the infection can be controlled.
A person diagnosed with HIV today who receives proper medical care before the disease has advanced too far can expect to live nearly as long and productive a life as a person who does not have the virus.
Virus evolution challenges vaccine development
In developing an HIV vaccine, researchers are essentially trying to make a safe tool that mimics the virus so that the patient's immune system learns how to recognize and attack the real one.
However, as Hillel Haim, senior author of the new study and assistant professor of microbiology, explains: "HIV is a highly dynamic virus. It continuously changes, both in an infected individual and, as a consequence of that, in the greater population."
So, he adds, the challenge for vaccine developers is, "how can you design a vaccine to hit a moving and continuously changing target?"
The "moving target" that Prof. Haim and colleagues investigated in their study is a protein that sits on the surface of HIV called the envelope glycoprotein (Env).
The rapid mutation of Env is the main stumbling block for attempts to develop an HIV vaccine. To make progress, vaccine developers need to find a way to anticipate and match the virus over time. They need to know which Env variants are current in the population and then predict how they will change.
For their study, Prof. Haim and colleagues used blood samples collected from hundreds of patients attending an HIV clinic in Iowa City that opened in the 1980s. The clinic has collected samples for more than 30 years.
The team carefully analyzed hundreds of Env proteins in the samples, focusing in particular on the integrity of specific structures in the protein. Prof. Haim says that the changes reminded him of patterns that occur when viruses randomly diffuse through liquids - something he had studied years earlier.
Diffusion model inspired by idea of volatility
Holding on to this idea of particle diffusion, the researchers then looked more closely at the Env structure changes to see whether they could pick up any other clues.
They looked at the properties of Envs in different viruses in the same blood sample. The noticed that while some changed quite a lot, others stayed largely similar.
They called this variation in Env properties "volatility" and further examination revealed that the volatility of each feature of Env structure was very similar among samples from different patients.
Financial models of the stock market also use the concept of diffusion and volatility. The researchers found that the volatility characteristic of an Env property was similar to the small fluctuations or volatility in price observed in a particular stock.
Inspired by the idea of volatility in modeling stock prices, the researchers then developed a new diffusion-based model and applied it to the blood samples. They note how it efficiently described the evolution of HIV Env proteins:
"Using volatilities measured from a few patient samples from the 1980s, we accurately predict properties of viruses that evolved in the population over the course of 30 years."
The researchers suggest that this means that just testing a few HIV patients may potentially give enough information to predict virus changes well enough to allow vaccines to match the particular HIV strains in particular populations.
Just in case people are wondering how the unpredictability of financial markets can inform such research, Prof. Haims concludes:
"Fortunately, relative [to] the financial market models that inspired this work, our predictions of changes in HIV are remarkably accurate, due to the highly conserved nature of randomness in this virus."