The trillions of microbes that inhabit the mammalian gut are a source of neuroactive compounds and mediate various interactions with the nervous system. Studies published just before the start of our project had shown that many pharmaceutical drugs, particularly those targeting the nervous system, can have off-target effects on human gut microbes^ref1. Such effects on microbes can lead to changes in the composition and function of the gut microbiome and communication with the host, which in turn can alter the therapeutic efficacy of drugs or mediate their side effects. But which microbes or microbial enzymes within the microbiome do these drugs primarily target? And how can we protect beneficial microbes and support positive microbiome-nervous system interactions, or reduce unwanted effects of drugs? These gaps in knowledge in the field of drug-microbiome interactions were the main drivers for our study.

Key to tackling these questions effectively was a highly collaborative team with complementary expertise, led by Michael Wagner (University of Vienna) and Ji-Xin Cheng (Boston University), and myself. Since 2019, our team has been working together on the development of a Raman spectroscopy-based platform to measure microbial activity at the single-cell level with high-throughputs in complex microbial communities^ref2. This platform, termed SRS-FISH, was an ideal tool to get our project off the ground and to study the effects of drugs on whole microbiome communities. This platform will surely be the basis for many more projects to come from this team – fortunately, Michi’s curiosity and questions never seem to end.

We started our project by recruiting a group of healthy people (mostly colleagues or friends) to provide a faecal sample, a good proxy for the distal gut microbiome and a non-invasive way to access these microbial communities. From the collected samples, we prepared a faecal slurry, distributed it into vials, supplemented each vial with the drug of interest and sampled it over a 24-hour period. Among drugs targeting the nervous system, we chose to focus on entacapone (Parkinson's disease adjuvant drug reducing the breakdown of the mainstay medication levodopa outside the brain) and loxapine succinate (to treat schizophrenia). These drugs were selected because of their established clinical use and documented strong inhibitory effects on microbial growth, with some selectivity for specific gut microbes. By the time we reached the second sampling point, we already suspected that entacapone affected the microbial community: cell pellets from the entacapone-treated samples were visibly smaller than those from the control or loxapine-treated samples. Microbial cell counts and sequencing results confirmed our suspicions: entacapone reduced total microbial loads and had a pronounced effect on microbiome composition. Loxapine also affected the microbiota, but to a much lesser extent. To determine how these drugs impact microbial activity, we supplemented our incubation vials with heavy water (D₂O), a tracer for metabolic activity^ref2. This allowed us to follow microbial activity at the single cell level using SRS-FISH - thank you Jannie and Klara - I remember that at one point we had more FISH labelling experiments running in parallel than timers available in the lab!

After many shipments to Boston - none of which resulted in lost or damaged samples, arguably one of our study’s greatest achievements - we observed changes in the activity of specific microbial community members even when these were supplemented with low drug concentrations. Another surprising finding from our stimulated Raman scattering (SRS) analyses was the detection of entacapone bioaccumulation in microbial cells, made possible by the intrinsic photothermal properties of entacapone, which we were able to image with single-cell resolution thanks to the Boston team's curiosity, skills and state-of-the-art optics instrumentation.

With these new findings about entacapone in mind, we turned to a chemist and collaborator, Thomas Boettcher, to help us track the fate of this drug in our samples using analytical chemistry techniques. Just one slide into the project presentation – the one slide featuring the entacapone chemical structure - Thomas interrupted with questions about iron concentrations and oxidation state in the vials. This was very confusing at first - why was he talking about metals? - but it all made sense when he explained that the presence of a catechol group probably enabled entacapone to chelate iron present in our incubation vials - potentially blocking microbes from accessing this micronutrient. We left the meeting with a new strategy: add an excess of iron to our incubations, more than what entacapone could chelate based on the stoichiometry of the entacapone:iron complex, and see if we could reverse the effect of entacapone. And indeed, we could! At the same initial meeting, Thomas also noted that the entacapone effect could be potentially circumvented by microbes capable of producing siderophores, molecules that bind iron with very high affinity, higher than entacapone. We then moved on to examine the genomes of microbiome members unaffected by entacapone and found out that these indeed possess the genomic ability to produce high-affinity iron-binding molecules. These observations were only possible due to the high degree of completion of the metagenome-assembled genomes retrieved, enabled by long-read sequencing and by a super skilled and efficient in-house sequencing facility (JMF Vienna) and nanopore wizard Rasmus! This study was indeed a big team effort - it was such a pleasure to work together with so many awesome scientists.

 Overall, our study underscores the broader concept that pharmaceuticals, even those not designed to target microbes, can influence the microbiome in unintended ways. Alarmingly, metagenomic analyses also revealed an increase in the abundance of virulence and antimicrobial resistance genes following entacapone treatment ex vivo.  Since oral iron supplementation reduces the absorption of entacapone in the small intestine, the microbiome-related effects of entacapone could potentially be mitigated by delivering iron directly to the colon in patients taking this drug. If administered in appropriate amounts, this new adjuvant therapy could be expected to aid in the preservation of microbiome homeostasis for Parkinson’s patients taking entacapone. However, further testing and investigation are needed before this can be implemented in patients. Additionally, as Parkinson’s patients are often on complex drug regimens, we plan to examine how these combined treatments collectively impact the gut microbiome.

You can read about all details of our study in the full paper here.