Probiotics may impair post-antibiotic gut microbiome recovery

It has long been assumed that taking a probiotic supplement broad-spectrum antibiotic therapy will help to restore damage done to the gut microbiome. But a study published in the journal Cell offers evidence that the effect may in fact be the opposite--that probiotics markedly delayed reconstitution of the indigenous microbiome after antibiotic therapy, with potentially adverse consequences. This is among the surprises emerging as scientific investigation of the vast realm of the microbiome advances.

Antibiotic disruption of the commensal microbiome

Acknowledging the immense benefit of how antibiotics have transformed the treatment of life-threatening bacterial infections, the authors note the significant adverse effects that can ensue, including antibiotic-associated diarrhea (AAD), hypersensitivity, and the emergence of resistant strains. Importantly, it can take as long as years to recover the normal microbial ecology.

Human microbiome reconstitution from antibiotic treatment is often slow and incomplete and, in some cases, may take years to revert to naive configuration. Of note, studies in rodent models and humans suggest an association between antibiotic exposure, especially during early stages of life, and a host propensity for a variety of long-term disorders, including obesity, allergy, increased risk of autoimmunity, and inflammatory bowel disease.

Uncertainty about probiotics

Up until now, assumptions about the effectiveness of post-antibiotic recolonization of the gut by probiotics have been largely unproven. Additionally, adverse effects may have been under-reported in previous trials.

Probiotics have been proposed to constitute an effective preventive treatment for antibiotics-induced dysbiosis and associated adverse effects...in some but not all human studies. Importantly, adverse effects associated with probiotics consumption may be under-reported in clinical trials, further complicating the efficacy debate. The extent and pattern of probiotic gut mucosal colonization and impact on the indigenous gut microbiome following antibiotic use also remain unclear...the vast majority of studies extrapolate their conclusions from stool samples, resulting in inconclusive findings regarding probiotics capability to restore the pre-antibiotics microbiome configuration. Importantly, no in vivo studies have directly examined the global extent of human mucosal probiotic colonization following antibiotic treatment and their impact on reconstitution of the indigenous mucosal microbiome or the host gut transcriptome.

To resolve this uncertainty the authors studied the effect of probiotics after antibiotic exposure in the gut luminal, and fecal microbiome on supplementing with probiotics or autologous fecal microbiome transplantation (aFMT) in human volunteers treated with broad-spectrum antibiotics and in mice. Controls were allowed to recover spontaneously over time.

Probiotics markedly delayed reconstitution of the indigenous microbiome

After a standard oral broad-spectrum antibiotic treatment with ciprofloxacin and metronidazole for seven days, the human cohorts included seven who were followed by 'watchful waiting' to observe spontaneous microbiome reconstitution, eight received an 11-strain probiotic preparation given twice a day for four weeks, and six had an autologous fecal transplant. A colonoscopy and deep endoscopy were performed right after the antibiotics and again three weeks later. Multiple stool samples were collected at intervals for six months to assess reconstitution.

Expectedly, antibiotic treatment in humans triggered a profound fecal microbial depletion and disruption of microbial community composition as observed in stool, LGI [lower GI] mucosa, and UGI [upper GI] mucosa, with the latter region the least affected by antibiotics. Compositional changes were accompanied by alteration of microbiome function in the stool and LGI as assessed by shotgun metagenomic sequencing.

Interestingly, they showed that antibiotics diminished the native microbiome's resistance to the probiotic species introduced, especially in the large intestine.

Collectively in the antibiotics-perturbed human gut, reversal of colonization resistance to probiotics enabled incremental gut colonization by the tested exogenously administered probiotic strains, mainly in the large intestine, leading to long-term probiotic fecal shedding indicative of stable colonization and active proliferation.

But reconstitution of the pre-existing, indigenous microbiome was markedly impaired. This was in contrast to aFMT and doing nothing (watchful waiting). The autologous fecal transplant cohort recovered in a day, and the spontaneous recovery group normalized after 21 days.

In contrast, probiotics-consuming individuals did not return to their baseline stool microbiome configuration by the end of the intervention period (day 28), and dysbiosis was maintained even 5 months after probiotics cessation, with all stool samples collected through day 180 remaining significantly different from baseline.

And for the clinically important characteristic of microbial diversity they observed these dramatically adverse effects:

Importantly, following antibiotics treatment, the number of observed species in feces was halved but was restored in both the aFMT and the spontaneous recovery groups within 1 and 2 days, respectively. In contrast, the alpha diversity remained significantly low and did not return to baseline in the probiotics group throughout the intervention period, with the alpha diversity reconstitution curve remaining lower compared to its own baseline, as well as aFMT or spontaneous, up to 5 months post probiotics cessation. Likewise, fecal bacterial load failed to return to baseline after 3 weeks of probiotics supplementation, as compared to both aFMT and spontaneous recovery, and remained lower than baseline 1 month after probiotics supplementation ceased.

Similar results were found when they documented microbiome reconstitution or the lack of it in the gut mucosa itself, the long-term residence habitat of the gut microflora.

Deep functional significance

But the real question is what does this mean for the physiology of the host? Is our biology changed in practical ways by the difference in indigenous gut microbiome reconstitution? To answer this question, the authors conducted an analysis of the gut transcriptome according to the individual cohorts examined. (While the genome is the collection of DNA, the genes express their function through RNA. The 'transcriptome', the collection of RNA molecules derived from the genes, is a dynamic representation of the cellular functional state). Antibiotics markedly modified the transcriptome as expected, and the three post-antibiotic regimens effects on the transcriptome mirrored the results for the gut microbiome (gut microbial genome) for multiple metabolic and immune pathways.

Given the differential impact of probiotics and aFMT, as compared to watchful waiting, on the recovery of mucosal gut microbiome composition and function, we next sought to characterize the effect of the three post-antibiotics interventions on the host. To this aim, we performed a global gene expression analysis through RNA sequencing of transcripts collected from stomach, duodenum, jejunum, terminal ileum, cecum, and descending colon biopsies immediately after antibiotics treatment and 3 weeks later into the three post-antibiotic interventions. Of note, antibiotics affected the transcriptional landscape across the GI tract, with the majority of differences between naive and antibiotics state observed in the descending colon. Importantly, restoration of the antibiotics-naive host transcriptional landscape by the three post-antibiotics intervention arms mirrored our findings in the microbiome, as multiple genes across the GI tract that were significantly affected by antibiotics were reverted toward homeostatic expression levels by spontaneous recovery and aFMT, but not by probiotics.

Not only that, but the functional effect of the probiotics appeared to increase inflammatory mediators. Note the increase in expression for the proinflammatory cytokine IL1B (interleukin-1beta) for the probiotic cohort (green bar) in the graph below.

Interestingly, probiotics led to an elevation in the transcript levels of inflammatory mediators and regulators of anti-microbial peptide secretion, such as IL1B, and of some anti-microbial peptides, such as REG3G, potentially contributing to the inhibition of indigenous commensal such as Clostridiales.

The authors conclude:

Importantly, in both mice and humans, we demonstrate that enhanced post-antibiotic probiotic colonization comes at a tradeoff of delayed indigenous microbiome and host mucosal transcriptome reconstitution to a homeostatic configuration as compared to either watchful waiting or aFMT. In contrast, aFMT results in rapid and nearly complete reconstitution of the gut mucosal microbiome configuration and host gut transcriptome.

...our study highlights an important previously unappreciated tradeoff in which improved probiotic gut mucosal colonization under disruptive antibiotic conditions led to a markedly delayed indigenous gut mucosal reconstitution in terms of composition, function and bacterial load, and prolonged dysbiosis that lasted at least 5 months following the cessation of probiotic exposure. While our study is not aimed or powered to assess the effectiveness, or lack thereof, of probiotics in ameliorating post-antibiotics clinical symptoms, we demonstrate that their putative ‘‘placeholder’’ effect may come at a price of significant prolongation of dysbiosis and delayed recolonization of the indigenous microbiome, resulting in altered reversion of the host gut transcriptome toward homeostatic configuration. This probiotic-induced ‘‘adverse effect’’ may be important in light of multiple observations linking antibiotics-associated dysbiosis and lower microbial diversity with increased susceptibility to a myriad of chronic and infectious diseases. The duration, extent, and long-term health consequences of probiotics-induced delayed endogenous microbiome and host transcriptome reconstitution—and whether they occur with other probiotics not tested in our study—merit further studies.

What to do? While the aFMT (autologous fecal transplant) was highly successful at restoring indigenous microbiome colonization and gut transcriptome compared to probiotics, its widespread use presents serious technical and safety hurdles. The authors offer a direction for future clinical research that would be most welcome.

An alternative, scientifically sound aFMT replacement modality would necessitate characterization of a person-specific ‘‘core gut mucosal microbiome function’’ enabling the generation of individualized bio-active commensal probiotic consortia providing post-antibiotics mucosal protection and core microbiome function. We expect such a highly defined, individual-tailored modality to enable improved clinical efficacy and reproducibility of probiotic use while minimizing the related potential consequences of indiscriminate probiotics colonization. This ‘‘personalized probiotics’’ approach merits further research.