Article Text


Changing the narrative on antibiotics
  1. Fergus Shanahan
  1. Correspondence to Professor Fergus Shanahan, Department of Medicine, Alimentary Pharmabiotic Centre, Clinical Science Building, Cork University Hospital, Cork, Ireland; f.shanahan{at}

Statistics from

Campaigning for antibiotic conservation is a bit like the struggle against global warming. The polar ice melts while politicians, not known for long-range thinking, debate deadlines for others to meet. Similarly, new drug discovery diminishes as resistance to existing antibiotics increases. Public health officials and policy makers predict a bleak future and plead for reform of prescribing practices, but clinicians deal with individual patients, not populations, in the present. Perhaps, it is time to direct the educational narrative on antibiotics towards the consumer. Enthusiasm for antibiotic treatment, particularly for soft indications, might be tempered if the potential adverse effects on the microbiota were better known. Antibiotic-associated diarrhoea and Clostridium difficile-associated disease are well recognised but more subtle effects are less well appreciated. Few clinicians may know that antibiotics can induce an immune deficiency in experimental animals. Furthermore, increasing evidence suggests that disturbance of the indigenous microbiota by antibiotic exposure, particularly in early life, is a risk factor for later development of obesity-related metabolic disorders and several immunologically mediated disorders, such as IBD and asthma.1– ,3 Concern about the adverse effects of antibiotic exposure will escalate with better understanding of host–microbe interactions in host development and homeostasis, especially when the full spectrum of antibiotic effects on both host and microbiota is appreciated.

In this issue of GUT, Morgun et al report a comprehensive assessment of the impact of a cocktail of antibiotics (ampicillin, vancomycin, neomycin and metronidazole) on both the microbiota and on host gene expression in intestinal tissue using germ-free and conventionally colonised mice.4 As anticipated, the total bacterial mass was depleted with most taxa decreased and some increased, and the associated changes in host gene expression attributable to this reduced microbial stimulus were determined by comparing germ-free mice with antibiotic-treated and conventional mice. Less than half of the gene expression changes were due to antibiotic-induced depletion of the normal microbiota. Most of the antibiotic-induced gene expression changes in the gut were due to either residual (antibiotic-resistant) organism or direct effects on the host. The potential effects of the residual or antibiotic-resistant microbes were shown by comparing germ-free mice that were colonised with microbes from antibiotic treated and from conventional mice. Direct or non-antimicrobial effects on host gene expression were determined by the administration of the drugs to germ-free mice and this accounted for about a third of the antibiotic-induced changes in gene expression seen in conventional mice. There were, as expected, overlapping effects, particularly between genes directly influenced by antibiotics and those regulated by antibiotic-resistant microbes.

Laser microdissection of the small intestine was used to show that changes in gene expression due to depletion of the normal microbiota occurred mainly in the lamina propria. These were associated with alterations in mucosal innate and acquired immunity, including marked reductions in T lymphocytes and immunoglobulin A-producing plasma cells evident on immunofluorescent confocal microscopy. In contrast, the genes influenced directly by antibiotics or by antibiotic-resistant bacteria were predominantly found within the villus and crypt epithelium and included mainly reduced mitochondrial gene expression associated with reduced epithelial viability. A clever computational strategy integrating the abundance of microbial genes and intestinal gene expression was used for transkingdom network analysis and suggested that the epithelial toxicity was mediated by virulence/quorum sensing in the antibiotic resistant bacteria which was confirmed in vitro.

Some of the effects of antibiotics might have been anticipated from earlier work by others, but this is the most precise demonstration of the spectrum of antibiotic effects with clear identification of mechanisms. Of course, the antibiotic cocktail studied is not used in humans, and the microbiota of experimental mice is far less complex than that of humans. However, the antibiotic constituents are often prescribed and the broad spectrum effects of antibiotics alone or in combination are likely to have similar clinical implications. In addition, the investigators studied adult mice, but it seems likely that the adverse effects of antibiotics, particularly those on the immune system, might be more pronounced in early postnatal life, when the microbiota is becoming established and immune function is maturing.

The direct and indirect effects of antibiotics on the host immune and intestinal function demonstrated by Morgun et al should inform the assessment of new drugs. Although the discovery pipeline for antibiotics has almost collapsed,5 ,6 a deeper understanding of the mechanisms of action of some of the old drugs, such as the β-lactams, has revealed new opportunities for drug design.7 There is also a place for novel screening strategies, particularly for drugs targeting lipid molecules which should have are relatively low rate of development of resistance.8 Another promising strategy is to ‘mine’ the indigenous microbiota for naturally occurring antibiotics. Whether these would have less toxicity is uncertain but the principle has already been confirmed with the discovery of a new peptide/bacteriocin agent with marked specificity against C. difficile.9

The inescapable irony of the efficacy of antibiotics, regardless of whether they are prescribed appropriately or inappropriately, is their capacity to drive resistance by selection. Thus, the pessimistic view holds that conservation strategies will ultimately fail, which implies that non-antibiotic infection control strategies must be developed.10 Meanwhile, the work of Morgun et al has broadened the narrative on antibiotics to bolster the case for slowing down the emergence of resistance with more judicious use of broad-spectrum agents. Their data will also be of value to other investigators, as a compendium of transkingdom gene networks influencing epithelial function, and will inform the design of experiments on the functional effects of the microbiota.


View Abstract


  • Funding The author's work has been supported in part by grants from Science Foundation Ireland in the form of a centre grant (Alimentary Pharmabiotic Centre; Grant Numbers SFI/12/RC/2273 and 12/RC/2273).

  • Competing interests The author is a shareholder in a university campus company Alimentary Health and directs a research centre which holds collaborative grants with Janssen pharmaceutical, Trino Therapeutics, General Mills, the Kerry Group, Mead Johnson Nutrition, Friesland, Cremo, Sigmoid Pharma, Second Genome and Nutricia.

  • Provenance and peer review Commissioned; internally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles