Blood Microbiome in CKD

A diverse and complicated population of microbes lives in the human body (1). Accumulating evidence supports an important role for these microbes in health and disease (1,2). The term “human microbiome” broadly refers to either the entire population of microbes inhabiting an individual or to the total microbial genetic material within that individual (1). The latter forms the basis of modern microbial classification. It has become increasingly clear that each individual hosts many distinct microbial ecosystems and that microbial genetic material can be detected in essentially any human tissue. Recent studies have detected microbial DNA in the blood of healthy humans, a space traditionally considered microbe-free (3). It is not clear, however, to what extent this “blood microbiome” represents microbes living in the blood versus genetic material derived from microbes that is simply present in blood cells.

In this issue of CJASN, Shah et al. (4) describe microbial genetic material in the blood of patients with CKD and of individuals with normal kidney function. They compared blood from 20 patients with moderate CKD (median eGFR of 42 ml/min per 1.73 m²) and 20 controls (median eGFR of 86 ml/min per 1.73 m²). They isolated DNA from the buffy coat layer, which consists of mostly white blood cells and platelets and in what over 90% of microbial DNA in the blood has previously been located (3). It is not known, however, which specific white cells contain the microbial genetic material. Shah et al. (4) compared patients with CKD and controls in two ways: (1) by measuring the total concentration of microbial DNA, and (2) by profiling the microbial DNA’s composition. The technique commonly used to profile microbial DNA is to isolate and amplify a region of the 16S ribosomal DNA (rDNA) gene (1). The advantage of the 16S rDNA gene is that it contains a conserved or slowly evolving region. Sequences of this conserved region in microbial DNA samples can thus be compared with reference sequences in databases to delineate the composition and diversity of the microbial population from which the DNA was derived.

The first finding of Shah et al. (4) was that blood from patients with CKD contained the same total levels of microbial DNA as the control group (117 versus 122 copies/ng DNA). Previous studies have shown higher blood levels of bacterial products such as endotoxin in patients with CKD and patients on hemodialysis compared with controls (5). The increased bacterial product levels in CKD has been attributed to impaired function of the intestinal epithelial barrier and is considered to cause inflammation and cardiovascular disease. Other studies have shown higher levels of bacterial DNA in the whole blood of patients on hemodialysis or peritoneal dialysis, and in patients with advanced CKD (5). The absence of elevated microbial DNA in the blood of patients with CKD observed by Shah et al. (4) is therefore a notable “negative” finding. A possible reason for this difference from the results of prior studies is that Shah et al. (4) enrolled patients with less advanced CKD. Further work is needed to test whether levels of blood microbial DNA rise in proportion to declining kidney function.

Profiling 16S rDNA in the blood of patients with CKD presents a more complex problem. After adjusting for the large number of comparisons and confounding clinical characteristics, Shah et al. (4) found two differences in the blood microbial DNA profile of patients with CKD and controls: one at the phylum level (Proteobacteria: 61% in CKD versus 54% in control) and one at the class level (Gammaproteobacteria: 45% in CKD versus 38% in control). Although these differences reached statistical significance, their absolute magnitudes were small. More importantly, the extent to which differences in microbial DNA profiles represent differences in microbial function is uncertain. The use of blood 16S rDNA profiling may not provide information about microbial metabolism. There is limited correlation between 16S rDNA classifications and the presence of genes coding for specific microbial metabolic pathways (6). Metagenomic sequencing of the entirety of microbial genetic material may provide more insight into the functional capacity of a microbial population. Sequence analysis of 16S rDNA, despite being sometimes referred to as metagenomics, provides limited information on microbial behavior (1,6).

The broad definitions of a “microbiome” and of “metagenomics” highlight the challenge of interpreting data in the “-omics” era. Advanced analytical techniques can detect proteins, small metabolites, and microbial genetic material in various biologic samples (i.e., blood, urine, stool). These techniques generate large amounts of data, which are analyzed using sophisticated bioinformatic methods. There has been a rise in kidney medicine of these “-omics” analyses. The advantage of “-omics” analyses is the ability to obtain large amounts of data from relatively few samples and low sample volumes. A major challenge, however, is determining the relation of “-omics” data to disease mechanisms. In microbiome studies, it is important to recognize that the microbial profiling, despite providing a description of the microbial population, may provide very limited information on the function of the microbes.

An important question raised by the findings of Shah et al. (4) is whether microbial DNA in the blood is derived from the colon microbiome. The colon microbiome is the most thoroughly studied microbial population in humans (1). Interest in the colon microbiome in CKD has been motivated by the finding that several uremic solutes are products of colon microbes (7). The uremic solutes p-cresol sulfate, indoxyl sulfate, trimethylamine oxide, and phenylacetylglutamine are derived from colon microbial metabolism and have all been associated with poor outcomes in patients with CKD and ESKD(7). Efforts have been made to reduce the levels of such solutes (8). More intensive hemodialysis thus far has not provided substantial reductions in the plasma levels of colon-derived uremic solutes (9,10). There is, therefore, much interest in suppressing solute production using prebiotics, probiotics, and oral adsorbents (7). Future interventions may involve “reprogramming” of the colon microbial population such that the microbes that produce toxic solutes are eliminated (2,7). Indeed, studies have shown that the colon microbiome is altered in advanced CKD, which may contribute to the increased production and systemic absorption of uremic solutes (5). Alterations in the colon microbiome may also be responsible for CKD progression, cardiovascular disease, and inflammation (7). In theory, the blood microbiome could be a convenient method for profiling and monitoring the colon microbiome in CKD.

Some important points must be addressed, however, before blood microbial 16S rDNA profiling can be related to the colon microbiome. First, Shah et al. (4) found that more than half of the 16S rDNA sequences in the blood of the control participants was derived from microbes in the Proteobacteria phylum. Proteobacteria, however, represents a small minority of the normal colon microbial population, suggesting the blood microbiome was not derived from the colon microbiome (1). Second, other sites throughout the human body contain microbes. The microbial DNA in the blood buffy coat may be derived from microbes at other sites, such as from microbes residing in the oral pharynx.

In conclusion, Shah et al. (4) carefully profiled blood microbial DNA in CKD. The findings contribute to the understanding of the relationship between kidney disease and the microbiome and motivate future studies. Interpreting the function of microbial populations, however, remains a challenge. Only time will tell whether microbial DNA in the blood, a material whose origin and relationship to live microbes is unclear, will contribute to the understanding of the health of patients with kidney disease.

Disclosures

None.