In recent years, the human microbiome has become an exciting field of study. Spurred by advances in sequencing technology, along with disciplines like epigenomics and metabolomics, we know more than we ever have done about the 100 trillion or so bacteria in the gut.

While much remains to be explored, it seems clear that each person’s microbial makeup can profoundly affect their physiology. In a healthy person, the gut plays host to a diverse bacterial community with an important range of functions. When the gut bacteria fall out of balance, various chronic conditions can result.

This in turn has implications for the way we think about antibiotics. While antibiotics have saved millions of lives, they are something of a blunt instrument in that they kill ‘good’ bacteria as well as ‘bad’ ones. To put it more scientifically, they disrupt the delicate intestinal ecosystem, wiping out useful species of bacteria while allowing other, opportunistic species to thrive.

A question that hasn’t been asked so widely is whether other types of drug might also affect the microbiome. Could certain non-antibiotic classes of drug (especially those with gastrointestinal side effects) inadvertently damage the gut bacteria, in much the same way as antibiotics? And if they work like antibiotics, might they contribute to antibiotic resistance too?

A team of researchers at the European Molecular Biology Laboratory (EMBL) have been attempting to find out. In their latest study, published in Nature in March, they screened more than 1,000 marketed drugs against 40 strains of gut bacteria. A surprisingly high proportion (24%) affected the growth of at least one bacterial species.

“How much damage drugs do to the microbiome is undecided, even for antibiotics, but more so for non-antibiotics,” says Nassos Typas, a group leader at EMBL and one of the study authors. “People expect that because these drugs are developed to target human proteins, they wouldn’t affect bacteria. But we found that the effect of non-antibiotic drugs on the gut microbiome is much more extensive than expected.”

Profiling drug actions

Typas and his colleagues are not the first to explore this topic. Over the past few years, a number of different drugs (anti-diabetics, proton pump inhibitors, NSAIDs and antipsychotics, to name a few) have been associated with changes in microbiome composition. One recent paper, for instance, pointed out that atypical antipsychotics are often associated with metabolic disease, and suggested that ‘gut dysbiosis’ (i.e. imbalances in the gut ecology) might be why.

However, the literature to date has left several questions hanging. It hasn’t explained, for instance, whether the drugs impact the microbiome directly, or whether there are additional factors at play. Nor has it determined whether this is a widespread phenomenon, or confined to just a few classes of drug.

Typas’s team decided to systematically profile the interactions between drugs and individual gut bacteria, with a view to generating a comprehensive resource of drug actions on the microbiome. This resource would add an extra dimension to clinical studies and ultimately improve drug design.

Their findings were striking – 203 of the drugs in their sample were active against certain bacteria (including 14 not previously reported to have antibacterial activity), with 40 of them affecting 10 strains of bacteria or more. What’s more, for reasons explored in the paper, the 24% figure is very likely an underestimate.

“Another very interesting finding is that the microbes that tend to be resistant to non-antibiotics tend to be resistant to antibiotics too,” says Typas. “This made us think there might be a common underlying resistance mechanism for the two classes of drug.”

Resistance and sensitivity

Keen to pursue this line of thinking, the team conducted a few preliminary experiments. They found there was indeed an overlap between the resistance mechanisms against antibiotics and against human-targeted drugs.

This raises a galling possibility – in principle, you might be able to acquire antibiotic resistance even without taking antibiotics.

“By getting resistance to non-antibiotics you might as a collateral damage be building resistance to antibiotics, and that’s something we definitely want to look into further,” says Typas.

While this finding may seem troubling, he points out that we don’t currently know enough to assess the degree of risk. On top of that, where cross-resistance exists, we might also see a related phenomenon – ‘collateral sensitivity’. In other words, when bacteria become resistant to one drug, they might simultaneously become more responsive to another.

“The general trend is that the resistance mechanism might be common for non-antibiotics and antibiotics, but for every pair of drugs you might also have the opposite situation,” says Typas. “In this case, developing resistance to non-antibiotic drugs – let’s say to a proton pump inhibitor – might make you sensitive to an antibiotic. This would give us room to exploit collateral sensitivity to drive off antibiotic resistance. So in a way it’s not only bad news – it also gives us a window of opportunity.”

Future directions

At present, the EMBL team is pursuing a number of different directions. Firstly, they are further exploring the risk that non-antibiotics might foster antibiotic resistance. Secondly, they want to look beyond individual microbial species, and see how drugs affect the complex communities actually found in the gut. Finally, they want to assess whether drug-microbe interactions might inform more than just the side effects.

“Antipsychotics for instance have an extensive effect on our gut microbes – we want to explore whether this effect on the gut microbiome extends only to side effects like weight gain, or whether it might also translate to the mode of action of the drug,” says Typas.

He feels that, sooner or later, drug developers will need to pay more attention to these kinds of interactions. This would give them more control over the drug’s side effects and even its primary mode of action, as well as better understanding the appropriate dosage.

“There’s a second aspect, which is how microbes might be changing the concentration of the drugs in the human body. So it’s not only drugs affecting microbes, it’s microbes affecting drugs,” he says.

Further down the line, he thinks this work might help researchers design optimal drug combinations, as well as leading to new possibilities within personalised medicine. Although the microbiomes of healthy individuals have a lot of similarities, in that they all carry a common set of species, they often carry very different strains.

This means one person might carry a strain of bacteria that is resistant to a drug, and the next person might carry a strain that is sensitive. If we know which is which, we might be able to pick the medications that best suit their microbiome.

“It might enable us to optimise drug choice for individuals depending on what microbes they have. It might also help us reduce side effects, or maybe even optimise the drug function,” Typas explains.