The majority of adult population consumes alcohol in various quantities, and alcoholism is the leading cause of premature deaths in the world [1]. Long-term alcohol abuse exerts a spectrum of potent effects on different body systems, ranging from psychiatric symptoms, malnutrition and chronic pancreatitis to alcoholic liver disease, hepatocellular carcinoma, and coronary heart disease. Alcoholism increases the risk of oropharyngeal and esophageal cancer. Central and peripheral nervous systems are vulnerable and strongly affected by ethanol. As diet is one of the major factors affecting the community structure and functional potential of gut microbiota [2], alcohol and the products of its degradation by human organism can strongly modulate the human gut microbiota [3]. Moreover, gut dysbiosis contribute to neuroinflammation in the context of alcohol exposure and withdrawal, subsequently leading to the psychiatric symptoms of alcoholism—an understanding concordant with the concept of “gut-brain axis” [4]. However, most studies of alcohol dependence syndrome (ADS) are focused on examining the neurophysiologic effects as well as its action on the functions of liver and other organs, while the question of how alcohol dependence impacts the gut microbiota to a large extent remains unexplored. Recently, the application of cultivation-free approaches (qPCR and 16S rRNA metagenomic sequencing) demonstrated that the extent of alcohol impact on gut microbiota varies between individuals [5]. In a group of patients with ADS, it was shown to be associated with a decreased abundance of major commensal microbial taxa, including Roseburia, Faecalibacterium, Blautia, Bacteroides, and Lachnospiraceae, increased intestinal permeability, and inflammation [6–8].

Alcoholism is commonly accompanied by a number of pathologies each of which can individually shift the composition of the gut microbiota. Particularly, chronic alcohol abuse is a factor strongly contributing to the pathogenesis of various liver diseases [9]. Alcoholic liver cirrhosis (ALC) develops in 10–15% of alcoholics [10]. A growing number of recent studies support the hypothesis that the gut microbiota can play a significant role in the onset and progression of the alcoholic liver disease (ALD) [11, 12], nonalcoholic fat liver disease (NAFLD) [13], nonalcoholic steatohepatitis (NASH) [11], and other liver dysfunctions. One of the proposed mechanisms is via a direct impact of the microbial products like endotoxins on liver, as it interacts with the gut both directly via the hepatic artery as well as indirectly via the metabolism of bile acids [14]. However, the mechanisms of such interactions are yet to be elucidated. The gut microbial community of patients with ALC is known to be dysbiotic; its composition is characteristic by increased presence of Proteobacteria (particularly, Gammaproteobacteria) and Bacilli. These shifts are mirrored by a decrease in the commensal taxa Clostridia, Bacteroidetes, and Ruminococcaceae, as well as Lactobacillus and Bifidobacterium. A significant reduction of Lachnospiraceae, Roseburia, Faecalibacterium, and Blautia levels contributes to intestinal bacterial overgrowth, which, together with the acetaldehyde-caused increased intestinal permeability, leads to an increased blood concentrations of endotoxins and activation of inflammatory cascades likely inducing liver damage [7, 15–17].

Liver dysfunction is a major confounding factor in surveys of gut microbiota in patients with ADS. In order to identify the changes on microbiota composition and functions resulting from alcohol dependence and liver disease for the first time we conducted a “shotgun” metagenomic analysis of stool samples from a group of patients suffering from alcoholism represented by two cohorts—patients with ADS (without advanced liver disease) and patients with ALC (with advanced liver disease, i.e., liver cirrhosis). In order to disentangle the effects of these two pathologies, we carried out a multifactor statistical analysis of the metagenomic data. Moreover, we analyzed changes in the specific functional potential of microbiota linked to both virulence factors and alcohol metabolism.

Several metagenomic studies have outlined the gut microbiota composition of patients with ALC [16, 28] and ADS [5, 29]—however, all of the projects describing microbiota in ADS were based on 16S rRNA amplicon sequencing. To the best of authors’ knowledge, the present study is the first one to describe the gut metagenome of ADS patients using shotgun (whole-genome) metagenomics and compare both its taxonomic as well as functional composition with ALC patients and healthy population. Our study of two groups of alcoholics with different degree of liver dysfunction demonstrated that the most expressed feature of each group is a substantial reduction of many commensal gut taxa, while the number of increased species is rather low. Such relative under representation of the “positive” markers in the pathology resembles a “burst of microbiota”—when shifts away from the balanced community structures have different directions that strongly vary between subjects. We found that patients with ALC manifest these “bursts” even to a greater extent than ADS patients—due to the combined impact of alcohol dependence and liver dysfunction. Visualization highlights extreme cases of this phenomenon as outliers. We confirmed the existence of such outlying samples using an alternative reference-free analysis of shotgun metagenomes.

Notably, the “drivers” of dysbiotic communities in ADS patients and ALC patients were found to be different. In an attempt to improve the precision of the taxonomic comparison, based on one common feature of our two groups of patients (i.e., alcohol dependence), we conducted a multifactor analysis using a state-of-art algorithm based on boosted general linear models. The results suggest that each of two factors, liver dysfunction and alcohol dependence, is associated with a significantly changed abundance of distinct bacterial taxa representing potential biomarkers of these two diseases. Interestingly, metagenomic signatures of two pathologies contain nonoverlapping set of markers; and few of the shared markers manifest opposite directions of change.

Liver cirrhosis is characterized by a higher level of gut dysbiosis than alcoholism. Massive depletion of major commensals from the Bacteroidales order is accompanied by a rise of taxa normally inhabiting the oral cavity, as identified during our analysis of relative abundances of respective genomes. This effect is likely linked to the abnormal bile secretion and composition and is in agreement with the results of an earlier study describing gut metagenome in nonalcoholic liver cirrhosis patients [16], thus confirming this hallmark of the liver dysfunction on another cohort. The presence of oral species Lactobacillus salivarius, Veillonella parvula, and Streptococcus salivarius is more pronounced in ALC patients in comparison with both healthy controls as well as the ADS group. Interestingly, we found that Bifidobacterium abundance was significantly associated with the cirrhosis. This observation might appear somewhat contradictory to a well-known enrichment of species exhibiting probiotic properties within this genus. While one cannot completely rule out the possibility that some of these taxa, along with the Streptococcus and Lactobacillus, could be transient food-associated microbes, there is a growing body of evidence suggesting that levels of Bifidobacterium can increase in pathologies like alcoholic hepatitis [30] and inflammatory bowel diseases [31]. Overall, it would be interesting to compare the shifts in microbiota composition with detailed data on liver function assessed using additional methods like liver biopsy (although the latter was not performed in our study as it is not a required standard part of examination of such patients and possesses certain risks).

Alcohol dependence syndrome was found to be the second most important factor contributing to changes in microbiota. Inter-individual variability of types and amount of consumed drinks could have affected the results—however, due to low reliability of self-reported data on alcohol consumption, this was not in the focus of the study. Considering the metagenomic signature, the range of decreased autochthonous taxa in patients suffering from alcoholic dependence was found to be quite different from the ones of liver cirrhosis and includes many butyrate-producing taxa from the Clostridiales order, among them—Faecalibacterium prausnitzii, Coprococcus eutactus, and Roseburia spp. Depletion of the gut community potential for producing butyrate, a vital “currency” metabolite providing anti-inflammatory protection and oxidative stress amelioration in the gut, is coupled with the uprise of opportunist species tolerant to the reactive oxygen species (ROS), beneficiaries of the inflammation: Enterococcus and members of Enterobacteriaceae family (including Escherichia and Klebsiella). The intestinal domination of each of these opportunist genera was previously found to be associated with a high risk of bacteremia [32]. Due to an increased intestinal permeability observed in alcoholics, the dominance of these taxa is likely to lead to an increased passage of microbial products into lamina propria and trigger a stronger immune response than in a healthy population with similar community structures. Noteworthy, a strong association of alcoholism with Clostridium sp. HGF2, a species able to use carbon monoxide as an electron donor, apparently reflects the selective advantage conveyed by this ability in inflammatory conditions [33].

We found that both alcohol dependence and liver cirrhosis have certain commonalities in their metagenomic signatures. The central point is an increase of multiple species belonging to Lactobacillus and Bifidobacterium genera (although the precise set of differentially abundant species varies between these two pathologies). ALC patients tend to have a larger number of Lactobacillus species increased in abundance (among them oral species), while ADS patients are uniquely characterized by a statistically significant increase of the whole genus. We suggest that this could be associated with genus’ ability to metabolize alcohol and its products [34]. Noteworthy, the detected increase of Bifidobacterium is contrary to previous observations for the Japanese cohort of alcoholics. We speculate that two important factors could play role in contributing to these differences. The first factor is genetic difference between the cohorts from the two studies. Compared to the population of Russia, the East Asian people often carry genetic variants of alcohol metabolism genes that are associated with increased acetaldehyde levels after ethanol consumption [35]. Besides worse physiological responses to alcohol, this specific trait in the dynamics of ethanol degradation by the host is likely to modulate the gut microbiota community structure in the Japanese.

The second suggested factor is hazardous drinking with consumption of manufactured ethanol-based liquids not intended for consumption specific for Russia. Nonbeverage alcohol is a potentially major contributor to the mortality [36]. Its consumption results in higher frequency of episodes of very high blood concentrations of ethanol and its metabolites—comparing to heavy drinkers who consume equal amounts of ethanol in the form of beverage alcohol. Moreover, such surrogate products contain toxic aliphatic alcohols [37]. Apparently, these distinctive features also affect the way gut microbiota is changed in ADS patients—possibly, leading to increased fraction of Bifidobacterium in ADS patients from Russia. Members of the genus are potentially important acetaldehyde accumulators—species able to accumulate acetaldehyde in the colon in mutagenic concentrations [25]. Therefore, the observed enrichment of this genus coupled with a significantly increased abundance levels of microbial ADH and other genes related to alcohol metabolism in gut metagenomes suggests that lumen microbiota of Russian ADS patients is characterized by a high potential of acetaldehyde synthesis thus contributing to the risk of colon carcinogenesis.

Furthermore, the detected dominance of the Bifidobacterium and Lactobacillus in patients with ADS and ALC points out that application of probiotic products based on the species from these two genera should be reconsidered. These bacteria have a reputation of “beneficial” agents in the human gut capable of anti-inflammatory activities, protecting against pathogens. Thus, they are being actively used for treatment of various intestinal disorders including alcohol liver disease [38]. However, our results showing enrichment of Lactobacillus and Bifidobacterium in patients suffering from alcohol dependence or liver disease suggest that these microbes could be potentially detrimental. Thus, prescription of the species to these types of patients should be approached with caution.

An additional cause for the enrichment of these two genera in guts of ADS and ALC patients could be due to involvement of these microbes in the “gut-microbiota-brain” axis. There is a frequent rate of psychological and psychiatric problems among patients suffering from alcohol abuse. One of the neuromediators playing a central role in the effect of the alcohol on the brain is serotonin [39]. Serotonin is linked to alcohol effects on the brain, and its altered levels were shown to be associated with the alcohol-related depression. It is estimated that around 80% of this neuromediator in a human organism is synthesized in the gut by enterochromaffin cells, and its synthesis has been recently shown to be regulated by the gut microbiota [40]. On the other hand, certain species of Lactobacillus and Bifidobacterium have been shown to have an effect on higher nervous functions including the behavior, with some of the species referred to as “psychobiotics” [41]. The shotgun metagenomic approach used in our study provides species-level resolution allowing one to distinguish between individual species Bifidobacterium and Lactobacillus species (the species-resolving powers of shotgun metagenomics are superior to the resolution of 16S rRNA amplicon sequencing). One of the Bifidobacterium species we found to be associated with liver cirrhosis is B. dentium. This species has been recently shown to be capable of producing large amounts of gamma-aminobutyric acid (GABA), an important neurotransmitter [42]. Based on these facts, we speculate that an increased presence of Bifidobacterium and Lactobacillus genera in the gut of alcoholics could reflect a compensatory feedback of the microbiota to altered metabolism of serotonin and other neuromediators in organisms of alcoholics. Our hypothesis could be evaluated in further experimental studies involving inoculation of germ-free animal models with species isolated from the ADS and ALC patients.

The gut dysbiosis observed in ALC and ADS patients at the taxonomic level was manifested by pronounced shifts at the level of the functional potential. Firstly, we observed significant increase of the abundance levels of the genes related to alcohol metabolism. Although this result reflects the high potential of the gut community to perform these reactions, the extent to which it manifests could be validated by measuring the expression levels using metatranscriptomic or metaproteomic techniques. Our results extend the previous observations on Japanese population showing that that feces of ADS patients display lower ethanol- and acetaldehyde-metabolizing activities than control group [29]. The difference from that study might be linked to the fact that the mentioned assay of fecal ethanol metabolism was performed under aerobic conditions, while the intestinal environment is anaerobic. The in vivo metabolic activity of gut microbiota in ADS patients should be elucidated using multi-omics experiments on larger cohorts.

Further, among functions enriched in ALC patients were glutathione metabolism (reflecting the functional adaptation of microbiota to oxidative stress associated with the inflammatory environment [43]), porphyrin metabolism and biosynthesis of siderophore group nonribosomal peptides (likely related to the iron-scavenging potential of gut pathobionts [44]), as well as ABC transporters and PTS responsible for nutrient uptake. We observed an increase in simple sugars metabolism (fructose/mannose and galactose metabolism pathways) previously reported to be enriched in subjects with inflammatory bowel disease in comparison with healthy controls [24, 45]. The observed shifts in metabolic potential of microbiota in our cohort of ADS patients were different: the list of decreased functions included phenylalanine, tyrosine, and tryptophan biosynthesis likely related to a decreased ability of the microbiota to synthesize these amino acids characteristic of auxotrophic and pathobiont species [46]. As a common feature, the functions depleted in microbiota of both ALC and ADS patients included bacterial chemotaxis and flagellar assembly. Together with a strong increase of specific gene groups related to virulence factors and metabolism of alcohol and its toxic metabolites, these results signify that gut microbiota in alcoholics might take active role in pathogenesis of associated comorbidities—to a different extent depending on the degree of the liver damage. In the future, the identified gene groups could be used as potential biomarkers in diagnostic kits for detecting gut microbiota alterations associated with pathologies linked to alcoholism and alcohol-related liver disease.

This metagenomic study disentangled the impact of two main clinical factors—alcohol dependence and liver dysfunction—on the gut microbial community and its functional potential in patients suffering from alcoholic dependence. These two factors were associated with profoundly different changes in the species-level composition, suggesting that the mechanisms of microbiota involvement in the pathogenesis vary in patients with different degrees of liver damage. While specific changes associated with alcoholic dependence suggest the onset of inflammatory environment in the gut, the hallmarks of the liver cirrhosis are likely linked to the impaired bile secretion. The observation that each of these pathologies is associated with a different pool of Bifidobacterium and Lactobacillus species suggests alterations of the gut-microbiota-brain signaling in alcoholics involving these clades. Overall, our associative findings related to isolated species pave the way to understanding the altered community-level ecology within the gut community in alcohol-related diseases. Particularly, they lay the foundation for a personalized design of treatment of alcoholism in patients with different degree of liver damage. Based on patients’ microbiota composition, personalized treatment would identify the most promising pro- or/and prebiotics, as well as enable rational selection of individual donor for fecal mass transplantation (FMT)—the method whose therapeutic potential has already been demonstrated for a large number of microbiota-linked pathologies [47]. Finally, the expressed enrichment of functions related to inflammation, virulence factors, and alcohol metabolism in ALC and ADS patients suggest active contribution of gut microbiota to formation of psychiatric symptoms of alcoholic dependence, as well as to the increased risk of serious diseases associated with alcohol abuse.