Differences in gut microbial patterns associated with salivary biomarkers in young Japanese adults

Abstract

Recent evidence suggests that psychological stress is associated with gut microbiota; however, there are no reports of its association with gut microbial structure. This cross-sectional study examined the relationship between psychological stress and gut microbial patterns in young Japanese adults. Analysis of fecal microbiota was performed using terminal restriction fragment length polymorphism (T-RFLP). Psychological stress was assessed using salivary biomarkers, including cortisol, alpha-amylase, and secretory IgA (S-IgA). Fecal microbial patterns were defined using principal component analysis of the T-RFLP profile and were classified into two enterotype-like clusters, which were defined by the B (microbiota dominated by Bacteroides) and BL patterns (microbiota dominated by Bifidobacterium and Lactobacillales), respectively. The Simpson index was significantly higher for the BL pattern than for the B pattern. The salivary cortisol level was significantly lower for the BL pattern than for the B pattern. Salivary alpha-amylase and S-IgA levels showed a negative correlation with the Simpson index. Our results raise the possibility that salivary biomarkers may be involved in the observed differences in microbial patterns.

INTRODUCTION

The human gut microbiota is a highly complex microbial community, comprising over 500 different species and consisting of approximately 10¹⁴ intestinal bacteria. The gut microbiota is known to change with geography, sex, and age of the individual [1,2,3]. Moreover, increasing evidence suggests that dietary habits influence the relative proportions of dominant bacteria [4, 5], as we recently reported [6].

Stress can impact the gut microbiota composition and has been associated with increased onset of mental illnesses. Mice exposed to chronic psychological stress have been shown to have an altered gut microbiota, including an increase in the relative abundance of Clostridium and a decrease in the relative abundance of Bacteroides [7]. A human study showed that patients with major depression have increased relative abundances of Bacteroidetes and Proteobacteria in the fecal microbiota compared with healthy control subjects, and additionally, the fecal bacterial diversity was different between the two groups [8]. Moreover, it is well known that irritable bowel syndrome (IBS) is affected by emotional disorders, including depression and anxiety. The gastrointestinal symptoms of IBS frequently accompany chronic constipation and/or diarrhea [9]. Several studies have also investigated differences in the composition of the gut microbiota between patients with IBS and healthy controls [10, 11]. Furthermore, human and animal model studies have revealed that stressful events are associated with changes in the gut microbiota composition [12,13,14].

The stress response system is modulated by the two cardinal stress axes, the sympathetic adrenomedullary (SAM) system and the hypothalamic-pituitary-adrenal (HPA) axis. When the SAM system is activated by stress stimulation, alpha-amylase is secreted from the salivary gland, which is usually measured as a salivary biomarker [15]. Salivary alpha-amylase has been widely used in psychiatric research and is increased by psychological stress [16, 17]. On the other hand, the activation of the HPA axis raises salivary cortisol, which is also influenced by chronic stress [18]. Salivary cortisol levels have been found to be highly correlated with free serum cortisol levels [19]. Several studies have shown that salivary cortisol is considerably elevated by chronic stress, such as stressful daily events and work-related stresses [20, 21]. In addition to these salivary biomarkers, salivary secretory immunoglobulin A (S-IgA) has been reported to be associated with mental stress [22].

In recent years, gut microbial patterns have been associated with several factors, such as dietary intake and lifestyle. For example, a metagenomic study indicated that fecal microbiota composition in humans can be clustered by three dominant bacterial enterotypes: Bacteroides, Prevotella, and Ruminococcus [23]. These enterotypes have been found to be associated with habitual dietary intake. A western diet that is rich in animal protein and fat reportedly favored the Bacteroides enterotype, whereas the Prevotella enterotype was associated with a carbohydrate-enriched diet [4]. Additionally, the fecal microbiota of Asian children could be classified into two enterotype-like clusters, which were dominated by the Prevotella and Bifidobacterium/Bacteroides types [24].

Several research groups recently investigated the relationship between gut microbial patterns and host conditions [25, 26]. However, to the best of our knowledge, no comprehensive study has been conducted describing the relationship between the typical bacterial clusters of gut microbiota and psychological stress in healthy humans. We hypothesized that gut microbial patterns would be associated with the salivary biomarkers. Moreover, our preliminary results showed that salivary alpha-amylase was negatively correlated with fecal microbial diversity in healthy young women (unpublished observations). The purpose of this cross-sectional study was to examine the enterotype-like clusters of healthy young adults and investigate their possible relationships with salivary biomarkers to understand the influence of psychological stress on gut microbial patterns. Previous studies have reported that levels of salivary cortisol and alpha-amylase differ depending on the age of the individual [27, 28]. We therefore conducted this study on young adults to avoid this limitation.

MATERIALS AND METHODS

Study subjects

A survey was performed involving 61 healthy young Japanese university students (22 males, 39 females; 18 to 22 years of age). Habitual drug takers and smokers were excluded from the study. The survey was carried out from August 2017 to July 2018. The study complied with the principles in the Declaration of Helsinki, and all subjects provided written informed consent before their participation. All aspects of the study were approved by the Ethics Committee of Aichi Shukutoku University (No. 2017-2). Samples and data are the same as those previously published elsewhere [29].

Measurement of body composition

Bodyweight, body mass index (BMI), skeletal muscle mass, and percent body fat were measured by the bioimpedance analysis method using an InBody 270 device (InBody, Tokyo, Japan).

DNA extraction and fecal microbiota analysis

Fecal samples were collected by subjects at home using fecal sampling tubes containing a guanidine thiocyanate solution (Techno Suruga Laboratory Co., Ltd., Shizuoka, Japan) and were stored at room temperature. The time of stool collection was not specified. Collection occurred at the time of defecation. Samples were immediately delivered to the university and stored at 4°C until analysis. DNA was extracted from fecal samples according to a previously described protocol [30]. Fecal microbiota were analyzed by Techno Suruga Laboratory Co., Ltd. using terminal restriction fragment length polymorphism (T-RFLP). Each sample was suspended in a solution containing 4 M guanidium thiocyanate, 100 mM Tris-HCl (pH 9.0), 40 mM EDTA, and 0.001% bromophenol blue. Each sample was then subjected to bead beating with zirconia beads. Thereafter, DNA was extracted from the bead-treated suspension using a Magtration System 12 GC and GC series Mag DEA DNA 200 (Precision System Science, Matsudo, Japan). The 16S rRNA gene was amplified from fecal DNA using a fluorescent-labeled 516f primer (5′-TGCCAGCAGCCGCGGTA-3′; Escherichia coli positions 516-532) and 1510r primer (5′-GGTTACCTTGTTACGACTT-3′; E. coli positions 1510-1492) [31, 32]. The resulting 16S rRNA amplicons were digested with 10 U of BSlI for 3 hours. The length of the terminal restriction fragment was determined using an ABI PRISM 3130xl Genetic Analyzer System (Applied Biosystems, Foster City, CA, USA) and analyzed using GeneMapper DNA analysis software (Applied Biosystems).

Analysis of salivary biomarkers

Participants were asked to collect whole mouth saliva between 10:00 and 11:00 to avoid daily fluctuations. Samples were collected by having participants spit into a 50 mL centrifuge tube for 5 minutes, according to the methods published by Michishige et al. [33] and Nagler and Hershkovich [34]. The collection of saliva samples was performed under resting conditions and within 2 days of collecting feces. Also, participants were not allowed to eat or drink 60 minutes before sampling. After saliva collection, samples were centrifuged at 1,500 × g (3,000 rpm) at 4°C for 15 minutes, and the supernatant of each sample was aliquoted and stored at −80°C until analysis. Salivary cortisol, alpha-amylase, and S-IgA concentrations were determined using a cortisol salivary immunoassay kit, salivary α-amylase kinetic enzyme assay kit, and salivary secretory IgA indirect enzyme immunoassay kit, respectively (Salimetrics LLC, Carlsbad, CA, USA).

Statistical analyses

All statistical analyses were done using IBM SPSS Statistics for Windows statistical software version 25.0 (IBM, Tokyo, Japan). We defined gut microbial patterns through a principal component analysis (PCA) of fecal microbiota that used the T-RFLP profile for 10 bacterial groups (Bifidobacterium, Lactobacillales, Bacteroides, Prevotella, Clostridium cluster IV, Clostridium subcluster XIVa, Clostridium cluster IX, Clostridium cluster XI, Clostridium cluster XVIII, and others). Differences between groups were assessed parametrically using an unpaired t-test and nonparametrically using the Mann-Whitney U test. The association of salivary biomarkers and fecal microbiota was examined through the Spearman correlation method. Microbial diversity of the fecal microbiota was calculated using the Simpson index [35]. The number of subjects was equally divided into three groups of salivary biomarker tertiles (low, medium, high). Simpson indices were compared using the Jonckheere-Terpstra test. A p-value <0.05 was considered significant. All data are expressed as the mean ± standard error (SE).

RESULTS

PCA based on T-RFLP data showed that the first principal component (PC1) explained 23.1% of the variance and that the second principal component (PC2) explained 19.4%. Thus, the fecal microbial pattern was classified based on PC1. PC1 was positively loaded (factor loading ≥ 0) with relative abundance of Bifidobacterium, Lactobacillales, Clostridium cluster IX, and Clostridium cluster XI. PC1 was negatively loaded (factor loading <0) with relative abundance of Clostridium cluster XIVa, Bacteroides, Clostridium cluster XVIII, Clostridium cluster IV, and Prevotella (Fig. 1). The principal components for fecal microbial patterns were divided into positive and negative regions. PC1 positive regions defined for Bifidobacterium and Lactobacillales were labeled the “BL pattern” group. PC1 negative regions defined for Bacteroides were labeled the “B pattern” group. In total, 28 of the 61 subjects were classified into the BL pattern, and the rest were classified into the B pattern. The two patterns of subjects were not significantly different in terms of body weight, BMI, skeletal muscle mass, and percent body fat (Table 1). Also, the BL/B pattern ratios were approximately equal for men and women (1.0 and 0.8, respectively; data not shown).

Fig. 1.

Enterotype classification (A) clustering using PC1 loading and (B) a PCA loading plot of 61 subjects based on terminal restriction fragment length polymorphism (T-RFLP) profiles. The similarity between samples was assessed by principal component analysis. (A) The factor loading represents the correlation between each bacterial group and the principal component and has a value in the range of −1 to 1. Fecal microbial patterns were named according to the bacterial group showing factor loading. Subjects were classified into the BL pattern when the value of the principal component loading was positive and into the B pattern when the value was negative. (B) In order to visualize the similarity relationship with the fecal microbial pattern, the first principal component (PC1) is shown on the horizontal axis, and the second principal component (PC2) is shown on the vertical axis.

Table 1. Characteristics of the subjects.

	All subjects (n=61)	BL pattern (n=28)	B pattern (n=33)
Proportion of men (%)	22 (36%)	11 (39%)	11 (33%)
Body weight (kg)	56.8 ± 1.3	58.1 ± 1.1	56.6 ± 1.4
BMI (kg/m2)	20.9 ± 0.3	21.2 ± 0.5	20.9 ± 0.3
Skeletal muscle mass (kg)	25.0 ± 0.8	25.4 ± 1.1	24.7 ± 1.0
Percent body fat (%)	20.9 ± 0.9	20.4 ± 1.4	19.6 ± 1.4

Values are mean ± SE. BL: microbiota dominated by Bifidobacterium and Lactobacillales, B: microbiota dominated by Bacteroides, BMI: body mass index.

Figure 2 shows the results for the fecal microbiota composition of the BL and B patterns based on the T-RFLP profiles. The relative abundances of Bifidobacterium (p<0.01), Lactobacillales (p<0.05), Clostridium cluster IX (p<0.01), and Clostridium cluster XI (p<0.05) were significantly higher for the BL pattern than for the B pattern. The abundances of Bacteroides (p<0.05), Clostridium cluster IV (p<0.01), Clostridium subcluster XIVa (p<0.01), and Clostridium cluster XVIII (p<0.01) were significantly lower for the BL pattern than for the B pattern.

Fig. 2.

Differences in bacterial profile based on terminal restriction fragment length polymorphism (T-RFLP) between the BL and B pattern groups. Statistical significance was tested by unpaired t-test or Mann-Whitney U test. *p<0.05, **p<0.01.

The mean diversity indices of the BL and B patterns are shown in Fig. 3. The Simpson index was significantly higher for the BL pattern than the B pattern.

Fig. 3.

The Simpson indices of terminal restriction fragment length polymorphism (T-RFLP) profiles in fecal microbiota for the BL and B pattern groups. Statistical significance was tested by Mann-Whitney U test. *p<0.05.

Figure 4 shows the results of the salivary biomarker analysis for the BL and B patterns. The salivary cortisol level was significantly lower for the BL pattern compared with the B pattern, whereas the alpha-amylase and S-IgA levels showed more modest reductions.

Fig. 4.

Differences in salivary biomarkers of cortisol (A), alpha-amylase (B), and s-IgA (C) between the BL and B pattern groups. Statistical significance was tested by unpaired t-test or Mann-Whitney U test. *p<0.05.

The results of the Spearman correlation analysis are shown in Table 2. Salivary cortisol was negatively correlated with the relative abundance of Clostridium cluster XI (R=−0.36; p<0.01) but positively correlated with the relative abundance of Clostridium cluster XVIII (R=0.27; p<0.05). Alpha-amylase was significantly negatively correlated with the relative abundance of Prevotella (R=−0.28; p<0.05).

Table 2. Correlation between fecal microbiota and salivary biomarkers.

	Cortisol	Alpha-amylase	S-IgA
	CC	CC	CC
Bifidobacterium	−0.17	−0.06	−0.01
Lactobacillales	0.02	0.00	0.15
Bacteroides	−0.03	0.17	0.08
Prevotella	0.08	−0.28*	−0.21
Clostridium cluster IV	0.04	−0.09	−0.08
Clostridium subcluster XIVa	0.17	0.04	0.10
Clostridium cluster IX	−0.16	−0.03	−0.11
Clostridium cluster XI	−0.36**	−0.01	−0.03
Clostridium cluster XVIII	0.27*	−0.08	0.09
Others	−0.18	0.10	0.06

CC: correlation coefficient. *p<0.05, **p<0.01 based on Spearman correlation analysis.

Salivary alpha-amylase and S-IgA levels showed a significant negative correlation with the Simpson index (p for trend <0.05, Fig. 5).

Fig. 5.

The trend of association between salivary biomarkers and Simpson index. P-value analyzed using the Jonckheere-Terpstra test.

DISCUSSION

The present study investigated the association between fecal microbial patterns and salivary biomarkers among healthy young Japanese adults. We identified two typical bacterial clusters. The BL pattern contained Bifidobacterium and Lactobacillales. The B pattern contained Bacteroides. A difference in salivary cortisol levels based on the pattern of bacterial clusters was evident. Moreover, salivary alpha-amylase and S-IgA showed a significant negative correlation with the Simpson index. Our results raise the possibility that salivary biomarkers may be involved in the difference in the microbial structures.

It is now clear that the relationship between the brain and the gut modulates mental health in the host. The bidirectional communication between the brain and gut is called the brain-gut interaction, which occurs through several pathways, including the autonomic nervous system and the HPA axis [36]. A recent study demonstrated that germ-free mice were more sensitive in their HPA response to restraint stress and had higher cortisol levels compared with specific pathogen-free mice [37]. Pregnant rats with high levels of cortisone gave birth to pups with lower total and gram-negative bacteria in the intestine [38]. There is increasing evidence that the composition of the gut microbiota is transmitted to the brain via the nervous system, causing emotional changes, such as depression and anxiety [39, 40].

In this study, the BL pattern showed reduced salivary cortisol and an increased Simpson index. Among the various microbial species, Bifidobacterium and Lactobacillus are recognized as probiotics, which modulate the gut environment and confer benefits to the host’s mental health. For example, the ingestion of a probiotic product containing Lactobacillus casei strain Shirota for 8 weeks increased alpha-diversity and reduced stress-induced anxiety and depression in healthy humans [41]. A previous study showed that Bifidobacterium longum R0175 and Lactobacillus helveticus T0052 consumption decreased urinary free cortisol levels and improved mood scores [42]. Although the mechanism by which Bifidobacterium and Lactobacillus relieve stress is not well understood, a previous study in rats showed that administration of an L. casei strain suppressed stress-induced activation of corticotropin-releasing factor–expressing cells in the paraventricular nucleus of the hypothalamus [43]. Another study demonstrated that the intake of probiotics containing Bifidobacterium attenuated the inflammatory cytokine interleukin-6, resulting in an improvement in depression [44]. The collective findings indicate that Bifidobacterium and Lactobacillales-rich enterobacterial patterns may maintain good mental health.

Additionally, we demonstrated a decrease in the relative abundance of Bacteroides and Clostridium subcluster XIVa in the BL pattern group. Also, salivary cortisol levels were negatively associated with Clostridium cluster XI, whereas they were positively associated with Clostridium cluster XVIII. A recent study using a rat model of stress reported a marked increase in fecal Bacteroides and Clostridium subcluster XIVa [45]. Moreover, Clostridium cluster XI was reduced in patients suffering from depression compared with healthy subjects [46]. Human and animal studies have been consistent with the conclusion that psychological stress changes the gut microbiota composition. While most studies are based on comparative investigations, by ingestion of probiotics and a survey of patients with mental illness, our cross-sectional study is the first to show that chronic stress may affect the Clostridium clusters, and salivary stress markers closely reflect the host’s enterotype in healthy subjects.

The gut microbiota composition plays an important role in dietary habits and lifestyle-related diseases, and higher gut microbial diversity has been linked to health benefits. Low dietary fiber consumption can alter the bacterial community and significantly decrease microbiota diversity [47]. A quantitative metagenomic study that compared obese and lean twin subjects revealed the association of obesity with decreased gut microbial diversity [48]. Psychological stress also influences the gut microbiota composition and can decrease species richness and diversity [49]. In the present study, we observed that the B pattern, which was associated with high cortisol levels, had a significantly lower Simpson index. Moreover, high salivary alpha-amylase and S-IgA levels were negatively correlated with the Simpson index. We hypothesize that psychological stress might decrease microbial diversity and be influenced by the gut microbial structure. Although the results obtained in this study indicate that salivary biomarkers can influence not only gut microbiota composition but also microbiota diversity in healthy subjects, further research with a larger sample size is necessary.

Maternal depressive symptoms during pregnancy have been associated with reduced fecal S-IgA in infants [50]. A similar association was detected in an animal study, in which restraint stress applied to mice reduced the intestinal levels of S-IgA [51]. Only S-IgA in saliva was analyzed in our study. A future task will be to elucidate the effect of gut microbial patterns on fecal S-IgA.

In the present study, we used the T-RFLP method to investigate fecal microbiota in healthy subjects and classified them into two separate clusters based on PCA. T-RFLP is a PCR-based tool, commonly used to study the microbial community structure and dynamics, that can exhaustively analyze the microbial community and is suitable for the examination of many samples [52]. Moreover, T-RFLP profiles enable easy classification of bacterial groups based on PCA and cluster analysis. However, in recent years, human microbiome research has mainly relied on the comprehensive analysis of fecal microbiota using high-throughput DNA sequencing technology, such as next-generation sequencing, and clustering by multidimensional cluster analysis. Using 16S rDNA sequencing, fecal microbiota have been classified into three enterotype clusters, each characterized by high levels of Bacteroides, Prevotella, and Ruminococcus, respectively [23]. Another metagenomic study described the gut microbiota of Asian children from five countries, including Japan, China, Thailand, Taiwan, and Indonesia, and classified then into Prevotella and Bifidobacterium/Bacteroides enterotype-like clusters [24]. However, the Japanese population has a unique gut microbial community, with Prevotella as a minor component [53]. Additionally, research using the T-RFLP method failed to detect Prevotella in the fecal microbiota of Japanese subjects [6, 54]. Therefore, the enterotypes of Japanese gut microbiota may be different from other countries. From these standpoints, it is necessary to determine Japanese-specific enterotypes, and we are currently investigating this using high-throughput DNA sequencing technology.

In conclusion, our results demonstrate that salivary biomarkers may be associated with differences in gut microbial patterns, and these differences may be influenced by psychological stress. Further work in this area to reveal the effect of salivary biomarkers on enterotype clusters is ongoing.

CONFLICT OF INTEREST

We have no conflicts of interest to declare.