Skip to main content
Premium Trial:

Request an Annual Quote

Predictive Model Helps Explain Why Protease Inhibitor-Based HIV Therapies Fail

Premium

A computational model developed by researchers at Johns Hopkins and Harvard Universities has provided a possible explanation for why HIV treatment regimens that include protease inhibitors fail even though there are no observed resistance mutations in the viral genome.

According to a recent Nature Medicine paper describing the method, the simulation showed that wild-type virologic failure occurs soon after a decline from "perfect adherence" with protease inhibitors. These drugs have short half-lives, which means they progress rapidly into the so-called wild-type growth window, in which "drug concentrations are so low that wild-type virus is not adequately suppressed and failure can occur even without resistance."

As a result, even a short interruption in treatment with protease inhibitors can lead to failure, the authors concluded.

Besides explaining why protease-based therapies fail, the method can also “determine when a resistance mutation will be selected” as well as predict the outcomes of different drug therapies, the paper states.

"With the help of our simulation, we can now tell with a fair degree of certainty what level of viral suppression is being achieved – how hard it is for the virus to grow and replicate – for a particular drug combination, at a specific dosage and drug concentration in the blood, even when a dose is missed," Robert Siliciano, a professor of medicine at Johns Hopkins and a co-author on the paper, said in a statement.

It can also be used to predict how well patients are likely to do on a specific regimen based on their adherence to the treatment, he said.

The model uses a series of mathematical algorithms that factor in information on drug properties, fitness differences between susceptible and resistant strains, viral mutations, and patients’ adherence to treatment, the paper explains.

Once that information is put into the model, it then runs simulations using “hypothetical patients that have different levels of adherence,” for example, and looks for cases where the treatment failed, Siliciano explained to BioInform.

In the case of protease inhibitors, the model arrived at a "straightforward" reason why they fail despite a lack of resistance mutations. The sharp slope of PI dose-response curves means that resistance mutations can only develop within a narrow range of drug concentrations, but these concentrations decay rapidly, leading to wild-type resistance if there is any gap in adherence.

"We predict that patients who fail PI therapy with wild-type virus should be able to re-suppress virus if the same drug is taken with improved adherence," the authors said.

Furthermore, the model predicted that PIs would be more effective if they were provided as “single-pill combination therapies” as opposed to the current formulation where they are taken separately. According to the paper, a single pill would “prevent resistance regardless of patient adherence.”

Based on the model's predictions, a single pill — combining, for example, Tibotec’s darunavir, a protease inhibitor, with Merck’s raltegravir, an integrase inhibitor — should make it “theoretically impossible to develop resistance,” Siliciano said.

For one thing, patients would have to take both drugs simultaneously so there isn’t a chance that they would take one without the other and run the risk of developing resistance, he explained.

Those conclusions are supported by a simulation that showed that when the two drugs were administered as a combination pill, their “concentrations would rise and fall roughly together, reducing the chance that they reach the discordant levels that select for resistance” — in some cases where the drugs are taken separately, resistance mutations in the integrase gene are selected when darunavir concentrations are low and raltegravir concentrations are moderate to high, according to the Nature Medicine paper

If the model’s predictions are true and a pharmaceutical company developed such a pill, it “would have a lot of implications for the global [HIV] epidemic … particularly as treatment is expanded to … populations of people who live in resource-limited settings where they don’t get frequent monitoring” to ensure adherence and where the development of drug resistance is a concern, Siliciano said.

That’s an aspect of drug development “that I hope [pharma] would look into now that they know about this,” he added.

Although protease inhibitors were the focus of this study, the model can be used to study patients’ responses to other HIV treatment regimens, Siliciano said.

“This [study] was done to help everybody in the HIV field understand the evolution of drug resistance and why some drug combinations are better than others,” Siliciano said.

More generally, the developers believe their model can help scientists streamline development and clinical trials of future combination therapies by ruling out combinations that are unlikely to work.

The Scan

Booster Push

New data shows a decline in SARS-CoV-2 vaccine efficacy over time, which the New York Times says Pfizer is using to argue its case for a booster, even as the lower efficacy remains high.

With Help from Mr. Fluffington, PurrhD

Cats could make good study animals for genetic research, the University of Missouri's Leslie Lyons tells the Atlantic.

Man Charged With Threatening to Harm Fauci, Collins

The Hill reports that Thomas Patrick Connally, Jr., was charged with making threats against federal officials.

Nature Papers Present Approach to Find Natural Products, Method to ID Cancer Driver Mutations, More

In Nature this week: combination of cryogenic electron microscopy with genome mining helps uncover natural products, driver mutations in cancer, and more.