Smarter antibiotic use beats resistant bacteria

Laura Blackburn

10 April 2015

The consequences of rapidly emerging antimicrobial resistance (AMR), in particular bacterial resistance to antibiotics, are generally acknowledged as being extremely serious. Multiple issues such as over-prescribing, incorrect prescribing and lack of development of new drugs, need urgent attention.

The concern at government level is such that antimicrobial resistance has been added to the UK National Risk Register of Civil Emergencies for 2015, which estimates that if a widespread outbreak of bacterial blood infection occurred, around 200,000 people would probably be infected, of whom around 80,000 would die. In addition, the UK government launched the Review on Antimicrobial Resistance in 2014, which will publish its final package of actions and recommendations in summer 2016 on how the threat of AMR can be tackled nationally and internationally.

Strategies to combat AMR

Any strategy for dealing with AMR must combine many aspects, including not only new drug development but also optimising the use of drugs we currently have. A paper published in PLoS Biology this week outlines a strategy that uses two drugs – doxycycline and erythromycin – in a sequential manner to treat a strain of E. coli bacteria already partially resistant to these drugs, due to the presence of a genetic region coding a drug efflux pump.

Using drugs in combination is a common strategy and can work due to a number of effects. Sometimes two drugs given together have a bigger effect than their combined individual effects, known as synergy. Another effect is known as collateral sensitivity, where a bacteria's response to one antibiotic effectively sensitises it to the second drug. The question then becomes, how can this process be exploited to optimise the use of these two drugs, and at lower doses?

Combination therapy vs. sequential therapy

The researchers explored this question in the paper, by alternating the use of doxycycline and erythromycin over eight treatment rounds on E. coli in laboratory culture, in every possible combination where each drug was given four times, comparing the results to the two drugs being given at the same time as a combination therapy.

They found that five of the sequential treatment combinations cleared the bacteria completely, and that the low concentrations of the drugs used sequentially failed to clear the infection when given in combination. The researchers also used whole genome sequencing to study the genetic changes occurring due to the sequential treatment, finding that the genomic region coding the drug efflux pump (known as the acrRAB operon) was persisting or being duplicated in response to drug treatment. So in situations where drug treatment completely cleared the bacteria, the bacteria were still developing resistance to the drugs – but could not compensate for the collateral sensitivity caused by the sequential treatment.

Clinical application?

The authors are keen to point out that this study was carried out in the laboratory using cell culture, and as such is not currently applicable for use in a clinical setting. For example, it is not clear how treatment would work in the complex physiological environment of the human body. There is also currently little data available on how treatments would work, and which drug doses should be used and when. However, the principle of their approach could be explored and developed to see if it could work for the treatment of bacterial infections in patients.

The benefits of the approach would be that it uses drugs that are already available, at lower doses than combination therapy, and in the right sequential combination drugs can clear bacteria even in the presence of continued genetic resistance. This process could also extend the lifetime of current drugs.

However, there is much to be done before sequential treatment could be used in a clinical setting. The use of whole genome sequencing and other genomic technologies is important to fully understand the biology of the bacteria and how they develop resistance. One finding from the paper was that the collateral sensitivity was non-reciprocal in certain situations, in that many cycles of erythromycin followed by one of doxycycline led to bacterial resistance, whereas many cycles of doxycycline followed by one of erythromycin led to collateral sensitivity. Understanding processes such as this will be vital in helping to develop protocols where sequential treatment works effectively.

It is also clear from the paper that while there are many different potential sequential combinations available, only a few are effective. Should the approach be effective in patients, a clear and consistent approach will need to be developed for treatment planning and management of the associated protocols and data.

We discuss antimicrobial resistance and other aspects of pathogen genomics as part of our infectious disease genomics project; our final extensive report, which also provides a national policy roadmap for actions to make the most of genomic technologies for infectious disease surveillance and control, will be released soon.  

More from us