Evolution of Drug resistance

Self-replicating (living) entities evolve

Interesting Evolutionary ID Topics

Immune escape (e.g. influenza, SARS-CoV-2, HIV)
Evolution of virulence (e.g. reduced mortality of myxomatosis virus in rabbits)
Evolution in response to vaccines (why do we need a new flu vaccine every year but a 30+ year old vaccine still works well against measles?)
Evolution of drug resistance (within-host drug resistance evolution of HIV, resistance to antibacterial drugs for a lot of bacteria – MRSA, XDR-TB,…)

Evolution of Drug Resistance

Resistance of pathogens to drugs is a major public health problem.
We need to understand how and why resistance arises and spreads, and what to do about it.

Evolutionary Dynamics 101

Evolution has several main “ingredients”

Mutation (including recombination, etc.)
Random changes (drift)
Selection

Mutations

When something (e.g. RNA, DNA, proteins) is copied/produced, changes (mistakes) happen.
Most of the time, the rate at which changes happen is constant.
- High for RNA-based organism
- Low for DNA-based organisms
Sometimes, rates of change/mutation can be influenced by the environment (e.g. increased mutation of “stressed” bacteria).

Selection

Different mutants are “selected” based on their fitness.
Fitness is being measured in different ways. Ultimately, what counts for pathogens is the ability to transmit from host to host and keep the chain going.
The mutant with the higher fitness (better able to transmit) “wins” and the one with the lower fitness tends to die out or remain at low levels.

Fitness of Mutants

Most organisms are already relatively fit.
Most mutations lead to mutants with reduced fitness compared to their ancestor (wild-type).
Those are often called deleterious mutations.

Fitness of Mutants

Occasionally, a mutation leads to a mutant with increased fitness.
There are more ways/mutations to increase fitness a bit, only a few increase fitness by a lot.

Mutation-selection balance

Mutants are constantly generated.
Most of the time, mutants are less fit (i.e. reproduce/transmit less well) and are out-competed by the wild-type strain.
Selection acts against mutants and purges them.
The number of mutants is determined by the rate of mutations (production) and the strength of (negative) selection.

Mutation-selection balance

Impact of Interventions on Fitness

Interventions generally do not change the mutation process.
Interventions can impact the selection process.
Certain interventions (e.g. drugs, vaccines) can lower the fitness of the wild-type. This gives the mutant a selective fitness and it can take over.

Fitness and Environment

Modeling drug resistance evolution

Drugs can (but don’t have to) lead to the evolution and emergence of resistance.
Models can help understand evolutionary dynamics.
Models can be used to explore different scenarios and try to predict best strategies to minimize resistance emergence, maximize overall benefit, etc.
Good references to get started & learn more:
- Temime et al. (2008) Epid Inf
- zur Wiesch et al. (2011) Lancet Inf Dis

Example: Influenza drug resistance

Assume a new pandemic flu strain emerges.
The strain is initially susceptible to antivirals.
Antivirals reduce the fitness of the wild-type strain, thus a resistant strain might emerge and spread.

Example: Influenza drug resistance

The question/problem:

No treatment = no reduction in wild-type (wt) infections but also no reduction in wt fitness, therefore no resistance emergence.
Lots of treatment = large reduction in wt infections but also strong reduction in wt fitness and potentially rapid emergence and spread of resistance.
What is the best population-level of antiviral treatment to minimize total cases?

Example: Flu drug resistance evolution

Answering the question:

Pick some population-level of treatment.
Simulate a number of pandemics. At end of each simulation, count total number of cases (both wt and resistant infected). Compute average over all simulations for a given treatment level.
Change value for treatment level, repeat.
Report distribution of cases as function of treatment level.

DSAIDE Exploration

R package to explore infectious disease models without having to write code:

https://ahgroup.github.io/DSAIDE/

Install or run online, then find the “Evolution of Drug Resistance” app.

Loosely based on Handel et al. 2009 JTB “Antiviral resistance and the control of pandemic influenza: The roles of stochasticity, evolution and model details”

Questions?

https://phdcomics.com/

Slides: https://www.andreashandel.com/presentations/