Ratio of Klebsiella/Bifidobacterium in early life correlates with later development of paediatric allergy

Abstract

Several studies have reported that intestinal microbial colonisation patterns differ between non-allergic and allergic infants. However, the microbial signature underlying the pathogenesis of allergies remains unclear. We aim to gain insight into the development of the intestinal microbiota of healthy infants and infants who develop allergy in early life, and identify potential microbiota biomarkers of later allergic disease. Using a case-control design in a Chinese sub-cohort of a Singaporean birth cohort (GUSTO), we utilised 16S rRNA gene sequencing to assess intestinal microbial composition and diversity of 21 allergic and 18 healthy infants at 3 weeks, 3 months and 6 months of age, and correlated the microbiota with allergy at ages 18 and 36 months. Pronounced differences in intestinal microbiota composition between allergic and healthy infants were observed at 3 months of age. The intestine of healthy infants was colonised with higher abundance of commensal Bifidobacterium. Conversely, Klebsiella, an opportunistic pathogen, was significantly enriched in the allergic infants. Interestingly, infants with a high Klebsiella/Bifidobacterium (K/B) ratio (above the population median K/B ratio) at age 3 months had an odds ratio of developing allergy by 3 years of age of 9.00 (95% confidence interval 1.46-55.50) compared to those with low K/B ratio. This study demonstrated a relationship between the ratio of genera Klebsiella and Bifidobacterium during early infancy and development of paediatric allergy in childhood. Our study postulates that an elevated K/B ratio in early infancy could be a potential indicator of an increased risk of allergy development. This line of research might enable future intervention strategies in early life to prevent or treat allergy. Our study provides new insights into microbial signatures associated with childhood allergy, in particular, suggests that an elevated K/B ratio could be a potential early-life microbiota biomarker of allergic disease.

Keywords: 16S rRNA; Bifidobacterium spp.; Enterobacteriaceae; atopic dermatitis; eczema.

Figure 1

(A) Relative abundance (%) of bacterial genus between allergic and healthy infants at 3 weeks, 3 months and 6 months of age as determined by 16S rRNA sequencing. Only genera with relative abundance above 1% were represented in the graph above (B) Scatter plots show the log₁₀ relative abundance (%) of genus Klebsiella between allergic infants and healthy controls at 3 weeks, 3 months and 6 months of age as determined by 16S rRNA sequencing. Each dot represents the log₁₀-transformed relative abundance of Klebsiella harbored by each infant, and lines indicate medians.

Figure 2

Boxplots show comparison of microbial richness of allergic and healthy infants at 3 weeks, 3 months and 6 months of age based on (A) Observed species (B) Chao1 (C) ACE. The boxplots are shown as median, with whiskers represent minimum and maximum values. Single asterisk (*) indicates P ≤ .05 as evaluated by Mann-Whitney U test. The differences in microbial richness are regarded as significant if P ≤ .05.

Figure 3

Principle coordinates analysis based on (A) Unweighted and (B) Weighted Unifrac metrics to compare intestinal microbiota profiles of allergic and healthy infants at 3 months of age. (Red – Healthy controls, Blue – Allergic Infants). PC1 and PC2 explain the percentage variation observed within the data.

Figure 4

PCA biplot based on the compositional structure of the intestinal microbiota of allergic infants and healthy controls at 3 months of age. The PCA biplot is explained by two principal components (PC) i.e. PC1 and PC2 which could explain 27.56% variation observed within the data. Each circle represents the composition of intestinal microbiota of a subject (Red – Healthy controls, Blue – Allergic Infants). The allergic infants who had levels above median K/B ratio (K/B ratio > 0.09) are represented by blue triangle. The eight unique genera explaining the differences are represented by the arrows. There were 2 boxes with “Bac” and “Vei” as there are two loading vectors representing similar OTUs belonging to the Bacteroides and Veillonella genera, respectively.