Genomics

A gut prediction

Article metrics

Characteristic profiles of gut microorganisms in people with type 2 diabetes could aid diagnostics and therapies, but differing signatures between ethnicities and genders highlight the need for further studies. See Letter p.99

Microbial cells make up the majority of cells in the human body, and most of these reside in the intestinal tract1,2. Researchers have long recognized that some intestinal microorganisms are associated with health, but the beneficial impact of most of the gut's microbes on human metabolism has been discovered only relatively recently2. It is of great medical and societal importance to pinpoint the associations of these intestinal microbes with health and with diseases such as obesity, metabolic syndrome and type 2 diabetes (T2D)3. A study by Karlsson et al.4 on page 99 of this issue is an important contribution to this growing body of evidenceFootnote 1.

Experiments in mice have revealed a causal relationship between certain intestinal microorganisms and obesity5, and evidence from work in humans suggests6 that intestinal microbes have a causal role in mitigating insulin resistance — the hallmark of T2D. However, the human intestinal microbiota is immensely complex and includes thousands of species that have a collective genome, termed the metagenome, of close to 5 million genes4,7,8. High-throughput sequencing of this metagenome, which is derived from stool samples, is now regularly used to analyse the intestinal microbiota and is replacing characterization of individual microbes. When metagenomic analysis is combined with clinical data, it is known as a metagenome-wide association study (MGWAS).

Karlsson and colleagues conducted a detailed MGWAS in 145 European women who had T2D or impaired glucose metabolism (an indicator of pre-T2D), or who were healthy. The authors' results complement those of a similar study, by Qin and colleagues8, in a group of Chinese men and women that comprised T2D and healthy cohorts (Table 1). The results of both studies are astonishing, showing highly significant correlations of specific intestinal microbes and their genes with T2D. These findings could take approaches for early diagnosis and treatment of T2D far beyond what is possible with existing methods. Moreover, the findings indicate that the predictive power of MGWASs surpasses that of genome-wide association studies, which include only human genes; this is testament to the fact that microbial genes, in addition to our genes, are intricately related to the pathogenesis of T2D, and almost certainly other diseases.

Table 1 Study comparison: A comparison of two metagenome-wide association studies of the gut microbiota of patients with type 2 diabetes.

Metagenome analysis is a rapidly emerging field and new sequencing methods and computational-analysis approaches are being developed. Karlsson et al. and Qin et al. used the same high-throughput sequencing platform, which generates sequences of about 100 nucleotides in size, but their subsequent methods differed: Qin and co-workers mapped their sequences to known metagenome data sets, whereas Karlsson and colleagues constructed their own sequence assembly that they then annotated using newly developed algorithms. Both teams then identified sequences that act as signatures of groups of correlated genes, which Qin et al. termed metagenomic linkage groups and Karlsson et al. called metagenomic clusters.

When Karlsson et al. compared the frequencies of these signatures in T2D patient and control groups, they found that the presence of fewer Clostridiales bacteria that produce the short-chain fatty acid butyrate (Roseburia species and Faecalibacterium prausnitzii) was highly discriminant of T2D — as had also been seen by Qin and colleagues. Thus, the two studies underscore the known role of butyrate-producing bacteria as regulators of human glucose and lipid metabolism3, and the function of butyrate (together with other short-chain fatty acids) in maintaining intestinal integrity. However, several associations differed between the two studies (Table 1). For example, Karlsson and colleagues identified an enrichment of Lactobacillus gasseri and Streptococcus mutans, usually found in the mouth and upper intestinal tract, in their T2D cohort, whereas Qin et al. saw an enrichment of Proteobacteria, which may produce inflammatory lipopolysaccharides that lead to endotoxaemia.

Microbial characteristics such as these might constitute early warning signals9 that indicate new ecological states in the intestinal microbiota that deviate from the resilient microbiota of healthy people1,2. The findings lend support to a model (Fig. 1) in which changes in the composition of gut microorganisms are followed by tissue destruction; in the case of T2D, this affects insulin sensitivity of the liver and muscles. These initial stages could progress to catastrophic shifts in the microbiota that are associated with chronic disease and that may be reversible only through therapeutic intervention to change the intestinal microbiota.

Figure 1: Microbiota in health and disease.
figure1

Studies such as that by Karlsson et al.4 contribute to a model of how the composition of gut microorganisms can influence the health of an individual. The model proposes that external factors such as infection or diet alter the healthy, resilient microbial composition to form one that shows early warning signals, such as a reduced number of butyrate-producing bacteria. This altered microbial activity can cause tissue destruction. Further progression can lead to catastrophic composition shifts and chronic disease. If this occurs, it is possible that healthy microbial diversity can be restored, and damaged tissue repaired, only by delivery of specific 'therapeutic' bacteria.

However, further work is needed before our understanding of altered intestinal microbiota can be used in a diagnostic or therapeutic setting. Differences in study design and some confounding factors between the two existing studies (Table 1) could explain some of the differences in outcomes, but long-term, prospective and multi-ethnic cohort studies are needed to determine whether there are in fact substantive differences in predictors of gut microbiota between ethnic groups. Such studies will also need to take diet and gender into account, because both can influence the composition of the gut microbiota10,11. Furthermore, for early warning signs (Fig. 1) to be used in a diagnostic capacity, they will need to be shown to have consistent associations with other biomarkers, such as fasting glucose or cholesterol levels, and to be detectable at an early stage in otherwise healthy individuals who are at risk of developing T2D10.

Despite the need for corroboration of the findings, the potential value of such approaches is underscored by Karlsson and colleagues' demonstration of predictive capacity — the model based on the metagenomic characteristics of their T2D cohort was able to identify women in the pre-T2D cohort who also had high levels of blood-plasma markers associated with T2D. Moreover, such studies might lay the foundation for designing therapeutic bacteria that can be used to 'reset' the intestinal microbiota to the composition that is characteristic of healthy individuals1,2. Support for this therapeutic potential comes from microbiota 'transplantations' in patients with T2D and infections of the bacterium Clostridium difficile that were able to (temporarily) reshape the composition of the gut microbiota with concomitant beneficial metabolic effects6,12. Future strategies might include oral administration of certain intestinal strains as a personalized therapy to postpone or even cure T2D by improving metabolic control in patients. Many of these strains have already been cultured and characterized, so the goal of manipulating our microbiota to keep disease at bay could be closer than we think.

Notes

  1. 1.

    *This article and the paper under discussion4 were published online on 29 May 2013.

References

  1. 1

    Lozupone, C. A. et al. Nature 489, 220–230 (2012).

  2. 2

    de Vos, W. M. & de Vos, E. A. J. Nutr. Rev. 70, S45–S56 (2012).

  3. 3

    Vrieze, A. et al. Diabetologia 53, 606–613 (2010).

  4. 4

    Karlsson, F. et al. Nature 498, 99–103 (2013).

  5. 5

    Turnbaugh, P. et al. Nature 444, 1027–1031 (2006).

  6. 6

    Vrieze, A. et al. Gastroenterology 143, 913–916 (2012).

  7. 7

    Qin, J. et al. Nature 464, 59–65 (2010).

  8. 8

    Qin, J. et al. Nature 490, 55–60 (2012).

  9. 9

    Scheffer, M. et al. Nature 461, 53–59 (2009).

  10. 10

    Morrow, D. A. et al. Circulation 115, 949–952 (2007).

  11. 11

    Markle, J. G. et al. Science 339, 1084–1088 (2013).

  12. 12

    Van Nood, E. et al. N. Engl. J. Med. 368, 407–415 (2013).

Download references

Author information

Correspondence to Willem M. de Vos or Max Nieuwdorp.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

de Vos, W., Nieuwdorp, M. A gut prediction. Nature 498, 48–49 (2013) doi:10.1038/nature12251

Download citation

Further reading

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.