A cross-sectional study observing the association of psychosocial stress and dietary intake with gut microbiota genera and alpha diversity among a young adult cohort of black and white women in Birmingham, Alabama

Background The relationships between psychosocial stress and diet with gut microbiota composition and diversity deserve ongoing investigation. The primary aim of this study was to examine the associations of psychosocial stress measures and dietary variables with gut microbiota genera abundance and alpha diversity among young adult, black and white females. The secondary aim was to explore mediators of psychosocial stress and gut microbiota diversity and abundance. Methods Data on 60 females who self-identified as African American (AA; n = 29) or European American (EA; n = 31) aged 21–45 years were included. Cortisol was measured in hair and saliva, and 16S analysis of stool samples were conducted. Discrimination experiences (recent and lifetime), perceived stress, and depression were evaluated based on validated instruments. Spearman correlations were performed to evaluate the influence of psychosocial stressors, cortisol measures, and dietary variables on gut microbiota genus abundance and alpha diversity measured by amplicon sequence variant(ASV) count. Mediation analyses assessed the mediating role of select dietary variables and cortisol measures on the associations between psychosocial stress, Alistipes and Blautia abundance, and ASV count. Results AA females were found to have significantly lower ASV count and Blautia abundance. Results for the spearman correlations assessing the influence of psychosocial stress and dietary variables on gut microbiota abundance and ASV count were varied. Finally, diet nor cortisol was found to partially or fully mediate the associations between subjective stress measures, ASV count, and Alistipes and Blautia abundance. Conclusion In this cross-sectional study, AA females had lower alpha diversity and Blautia abundance compared to EA females. Some psychosocial stressors and dietary variables were found to be correlated with ASV count and few gut microbiota genera. Larger scale studies are needed to explore the relationships among psychosocial stress, diet and the gut microbiome.


Introduction
For decades, research has shown how stress gets "under the skin" by activating neuroendocrine, cardiovascular, and metabolic systems, 1-3 ultimately contributing to the development of diseases and conditions such as hypertension, atherosclerosis, diabetes mellitus, nonalcoholic fatty liver disease, Alzheimer's, depression, and cancer. Stress, de ned as a state of threatened homeostasis, 4 is becoming more pervasive through psychosocial, environmental, and cultural means in the United States, 5 with discrimination recognized as a stressor likely contributing to ongoing health disparities 6 . Reported experience of discrimination has been associated with cardiometabolic risk factors disruption, and social defeat models in animals, with changes in gut microbiota diversity and abundance and gastrointestinal disorders 23 . Additionally, research observing associations between gut microbiota and cognitive functioning (emotions, memory, anxiety and depressive feelings) has involved interventions including probiotics, vitamin D, dairy products, and ber (inulin) 25 . The experience of discrimination in association with metrics of the gut microbiome has not been reported to date, but is paramount to health disparities research and future interventional strategies 18,26 .
The association between life stress and gut microbiota may be explained by physiological and behavioral responses. The experience of stress can have an immediate effect on physiology by activating the stress response system, resulting in increased production of cortisol and in ammatory cytokines 4 . Chronic, on-going stress impairs normal HPA-axis response and immune function, and has been associated with cardiometabolic and gastrointestinal disorders 27,28 . Limited research has examined whether chronic cortisol level, an objective measure of stress, mediates associations between subjective stressors and the gut microbiota. Second, diet is recognized as a major modi able factor in altering microbiota diversity, abundance, and function 29 . Habitual dietary patterns are associated with microbial clusters, mucosal protection, and anti/pro in ammatory features [30][31][32] . The intake of micro and macronutrients has also been associated with various gut microbiota taxa, with dietary ber often being a strong contributor to maintaining bacterial diversity 3334 . Diet quality has also been found to be highly correlated with various types of stressors and disease mortality 27,35 . This has led some to observe the mediating and moderating role of diet quality in stress-disease processes, 36,37 but these mediating and moderating inquiries have not yet included the gut microbiome. We propose that a diet score developed to assess the dietary in ammatory potential (dietary in ammation score, DIS) may mediate the association between the experience of discrimination, and other stressors and disease-related gut microbiota. We also recognize that individual nutrients may have speci c effects on the gut microbiota; thus we are interested in exploring other potential dietary mediators including the ratio of caloric intake over estimated energy expenditure (cal:EER), healthy eating index (HEI), ber, sugar, and other dietary variables. This research will add to the limited data related to subjective measures of stress, including the experience of discrimination, perceived stress, and depression, and various dietary variables and the gut microbiome of metabolically healthy young adult women. This study aims to explore the gut microbiota pro le of generally healthy young women in relation to diet and stress as well as extend upon the interest in uncovering gut microbiota differences by race. We also aim to explore the potential mediating role of diet and cortisol in the association between reported stress, alpha diversity, and speci c stress-related gut microbiota. First, we hypothesize there will be differences in gut genera, dietary, and stress variables between races. Second, we aim to explore associations between stressors and gut microbiota diversity and genus abundance. We hypothesize that stressors will be negatively associated with gut microbiota diversity, positively associated with potential pathogenic bacteria, and negatively associated with healthful bacteria identi ed through previous research. Third, we aim to uncover associations between diet scores (DIS and HEI), individual dietary variables, and gut microbiota diversity and genus abundance. Observance of similar correlations will lead us to explore the stress-gut relationship through mediation analyses. Last, we hypothesize that cortisol levels (physiological mediator) or diet (nutritional mediator) will mediate the associations found between stressors and gut microbiota diversity or abundance. We are primarily interested in recent (RED) and lifetime experience of discrimination (LED) as the stressors, DIS as the dietary mediator, and hair cortisol (HC) as the physiological mediator; thus all other stress and dietary variables included in the analyses are exploratory. The rationale for this study involves the use of stress measures (discrimination) and a dietary index (DIS) that has not been covered in this area of research, as well as the observance of gut microbiota data of young, generally healthy women, whereas the majority of research is among those with cardiometabolic disease or dysfunction (diabetes). Second, because diet is such a strong moderator of the gut microbiota, we are interested in the independent/dependent association of subjective stress with gut microbiota, and if dietary intake or cortisol levels explain the association between the two. Stress and changes in mood or emotion so often leads to reduced engagement in health promoting behaviors, including low quality diet. To our knowledge and review, this question of dietary mediation has not been observed in stress-gut research.

Study Participants
Sixty-two African America (AA) and European American (EA) females were recruited from Birmingham, Alabama between August 2014 and April 2016. The same yer was distributed on all marketing platforms (the University of Alabama at Birmingham [UAB] website, word of mouth, and social media [Facebook]) and included messaging with interest in studying life stress, diet quality, and health status. A screening questionnaire was used to determine eligibility. Participants were eligible if they were under the age of 45 and had a body mass index (BMI) between 18.5 and 45 kg/m 2 . Participants were excluded if they had any medical diagnosis or medication known to affect body composition or metabolism (e.g., diuretics, beta-blockers, calcium channel blockers, angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, and other hypertension drugs). Participants were also excluded if they were taking monophasic or biphasic oral contraceptive (birth control) or reported an exercise frequency of greater than 2 hours per week, as these types of oral contraceptives 38 and exercise 39 alter hormone production, including cortisol, and would confound study results. Triphasic contraceptives are designed to mimic the natural rise and fall of estrogen and progesterone during a menstrual cycle.

Protocol
Data for this observational study were collected during two visits. Demographics, food frequency questionnaire (FFQ), discrimination and perceived stress questionnaires, and hair and saliva samples were collected at the rst visit. Stool samples were collected at home and returned at the second visit, within a week of the rst visit. If participants were taking triphasic oral contraceptive, they were scheduled for their visits during their luteal phase, when taking their placebo dose. Luteal phase was determined by patient report. All sample collections and analyses were conducted in the Core Laboratory of the UAB Center for Clinical and Translational Science (CCTS) Clinical Research Unit (CRU), the UAB Diabetes Research Center's Bio Analytical Redox Biology (BARB) Core, and the UAB Nutrition Obesity Research Center. Recruitment concluded once 31 participants from each race were reached, to ensure no overrepresentation or bias.

Demographics
Demographics obtained from participants included age, race, income, marital status, and education level.
Participants ages ranged from 18 to 45 years, which has been de ned as young adult. Income levels ranged from $0 to greater than $100,000, by increments of $10,000. Participants reported marital status as married, never married, separated, divorced, or widowed. Education levels included partial high school, high school graduate, partial college, college graduate, and graduate professional training.

Cortisol
Salivary cortisol (SC) and hair cortisol (HC) were assessed as acute and chronic objective measures of stress, respectively. Methodology for SC and HC can be found in previously published works 40 and from original sources [41][42][43][44] . Brie y, participants provided 5ml of saliva in a sterile collection tube over a period of 10 to 30 minutes and at least 60 minutes following their last meal. Hair samples of approximately 6mm in diameter and 3 cm in length were cut from the vertex posterior, as close to the scalp as possible, since hair growth is ~ 1cm/month.

Gut Microbiome
In the comfort of their homes, participants used testing kits by Zymo Research to obtain fecal samples for isolation of microbial genomic DNA. Participants were instructed to collect an early morning sample or their rst bowel movement of the day. Polymerase chain reaction was used on the prepared DNA samples with unique bar-coded primers to amplify the V4 hypervariable region of the 16S ribosomal DNA gene to create an amplicon library from individual samples. Base paired-end reads were sequenced using Illumina MiSeq. Analysis of the sequence data utilized the QIIME-based pipeline21 45 to produce sample alpha diversity (amplicon sequence variant, ASV) tables.
Analysis included quality control, merging of paired-end reads, ASV grouping, and taxonomic assignment. ASVs with average abundance > 0.005% were further processed and grouped by taxonomy. Each sample had at least 20,350 sequences per sample. Alpha diversity, measured by ASV count, and genus abundance were used in analyses for this manuscript.

Recent and Lifetime Experiences of Discrimination
Details of the discrimination scale developed by Shariff Marco et al is described in previous studies 40,46,47 . The rst section of the questionnaire which assesses recent experience of discrimination (RED; within the last year) and lifetime experience of discrimination (LED), was used in this secondary analysis. Brie y, participants reported whether they had experienced examples of unfair treatment, what they felt was attributed to the unfair treatment, and how stressful they perceived these experiences. Recent discrimination sections 1, 2, and 3 include 8, 6, and 1 questions for each section, respectively. Lifetime discrimination sections 1, 2, and 3 include 5, 6, and 1 questions for each section, respectively. Section 1 questions score from 0 to 3 for frequency, section 2 questions are scored by 0 or 1 (No/Yes) for attribution, and section 3 questions score from 0 to 3 for severity. Total scores for RED and LED range from 0 to 33 and 0 to 24, respectively.

Perceived Stress Score (PSS)
Cohen's PSS is composed of 10 questions assessing the experience or appraisal of stress over the last 30 days. An example question asks, In the last month, how often have you been angered because of things that happened that were outside of your control? with responses ranging from 0 (never) to 4 (very often). Reverse scoring is applied to questions 4, 5, 7, and 8 as they were worded in a positive direction and a response such as very often would indicate low perceived stress. Scores ranging from 0 to 13 indicate low stress, 14 to 26 indicate moderate stress, and 27 to 40 indicate high stress. In ongoing observational research from years 1983, 2006, and 2009, higher perceived stress scores remain more prevalent among women, minorities, and younger age groups 48 .
Patient Health Questionnaire-8 (PHQ-8) The eight-item patient health questionnaire has been validated and established as a depression diagnostic tool, measuring the severity of depression disorders 49 . Scores range from 0 to 24 with scores of 0 to 4, 5 to 9, and 10 or greater indicating none-minimal, mild, or moderate/current depression, respectively. Questions assess feelings of interest or pleasure, hopelessness, sleep hygiene, energy, appetite, self-worth, concentration, and communication di culties over the past 2 weeks, with responses ranging from 0 (not at all) to 3 (nearly every day). A depression score of 10 or greater, or current depression, has been found to be prevalent among women, but surprisingly, more prevalent among those who are married, employed, college educated, and white race.

Food Frequency Questionnaire (FFQ) and Dietary In ammation Score (DIS)
Dietary intakes were assessed with a validated, self-administered graphical FFQ through VioScreen 50 , which utilizes the food and nutrient information from the Nutrition Coordinating Center (NCC) Food and Nutrient Database. Data from the FFQ were used to calculate a DIS for each participant. The DIS was developed by grouping FFQ foods into 19 food groups, a priori, based on biological plausibility and prior literature 51 . Using multivariable linear regression, the authors determined each DIS food groups weight based on its association with an in ammation biomarker score (hs-CRP, IL-6, IL-8, IL-10). Brie y, scores were calculated by 1) grouping foods of the FFQ by the amount (grams) into DIS food groups, 2) disaggregating mixed dishes into DIS food groups, 3) calculating the supplement score (if data are available), 4) standardizing each food group, 5) multiplying each DIS component by weight, and 6) summing weight values to equal DIS score (more negative values indicate a more anti-in ammatory score).
Vitamin and mineral supplementation was not included in the calculation of DIS in our cohort due to inadequate questioning of supplementation. Refer to the supplemental le of the DIS validation study for a step-by-step process in calculating the DIS 51 . Following protocol established by Byrd et al, participant data were excluded from analyses if the FFQ estimated implausible caloric intake (< 500 and > 4500 kilocalories). Dietary variables of interest gathered from Vioscreen FFQ include: HEI, healthy eating index; cal:EER, the ratio of estimated daily caloric intake to estimated energy requirements (EER); pcarb, percentage carbohydrate intake; pfat, percentage fat intake; pprot, percentage protein intake; bcar, beta carotene (mcg); acar, alpha carotene (mcg); vitc, vitamin c (mg); ber (g); vitdiu, vitamin D (IU); viteiu, vitamin E (IU); n-3, omega 3; n-6, omega 6; 6:3 ratio, the ratio of omega 6 to omega 3; mfa, monounsaturated fat; pfa, polyunsaturated fat; sfa, saturated fat; tfa, trans fats; sugar, added sugar intake (tsp); fruit, servings of fruit; veg, servings of vegetables; sweet, servings of sweets; and FF, servings of fast food.

Statistical Analyses
Descriptive statistics (mean, standard deviation [SD], and frequencies) were calculated to summarize demographic characteristics of the cohort. ANOVA was used to determine differences in gut microbiota measures, RED score, dietary variables, DIS by race. Parametric and non-parametric p-values were reported due to violation of the assumptions of homogeneity of variances and normal distribution. Spearman correlation was used to rst determine the correlations among stress variables, gut microbiota diversity and the abundance of the top 25 genera of this cohort. Spearman correlation was also used to assess the correlation of various dietary variables and gut microbiota diversity and the abundance of the top 25 genera of this cohort. Mediation analyses were then run to assess the potential mediating role of dietary variables or HC in the signi cant correlations found between life stressors and gut diversity and genus abundance. Independent, or predictor variables, included RED, LED, and PSS. Dependent, or outcome variables, included ASV count, Alistipes abundance, and Blautia abundance. Mediator variables included HC, DIS, HEI, calories to EER ratio, ber, and added sugar. Signi cance was set at α = 0.05 for all statistical analyses. All analyses were performed with SAS statistical software (version 9.4, 2002-2012 by SAS Institute Inc., Cary, NC).

Results
Participant Characteristics Table 1 describes study participants residing in the greater Birmingham area. Women were between the ages of 18 and 45 years with an average of 29 years, and the majority reported partial college education or higher, annual income of greater than $20,000, and never married (Table 1). These characteristics did not signi cantly differ by race. In Table 2, gut microbiota, dietary, and stress variables are reported by race. ASV count and abundance of genera Blautia were lower among AA vs EA women. HEI scores, intakes of ber, alpha and beta carotene, vitamin E, fruit servings and vegetable servings were lower among AA vs EA, and DIS, sweet servings, and fried food servings were higher among AA vs EA. HC levels were signi cantly higher in AA vs EA, and AA had signi cantly higher scores for lifetime discrimination (LED) and PSS compared to EA.  P values for unequal variances were determined using Welch's ANOVA.

Correlations between Life Stress Variables and Top 25 Gut Microbiota Genera (Heatmap)
Life stress variables were found to be positively and negatively correlated with many of the top 25 genera (Fig. 1

Mediating Role of Diet or Cortisol in the Association between Life Stress and Gut Microbiota Diversity and Abundance
There was no mediating effect of dietary variables or cortisol variables in the associations between subjective life stress variables (PSS, RED, LED) and ASV count, Alistipes or Blautia abundance (Fig. 3, Table 3). We tested the signi cance of this indirect effect using bootstrapping procedures. Unstandardized indirect effects were computed for each of 5,000 bootstrapped samples, and the 95% con dence interval was computed by determining the indirect effects at the 2.5th and 97.5th percentiles. RED and LED were not signi cantly associated with any mediator variables. Perceived stress was a better stress predictor of dietary intake (HEI, p = 0.000; DIS, p = 0.001; ber, p = 0.035), but did not signi cantly associate with ASV count.

Discussion
Our results of these analyses reveal interesting ndings of the gut microbiota in relation to race, psychosocial stress, and dietary intake. First, ASV count and Blautia abundance were signi cantly lower among AA vs EA. Racial differences of the gut microbiota have been reported in previous yet limited research 26,52,53 , but our ndings regarding differences in Blautia abundance are new. Blautia was recently found to be inversely associated with visceral fat accumulation 54 and children with obesity and diabetes. A recent literature review of Blautia discusses it's use as a potential bene cial probiotic as it has been found to be involved in avonoid conversion, free radical scavenging, bacteriocin production thus inhibition of pathogenic bacteria colonization, and maintenance of environmental balance through upregulating T regulatory cells and short-chain fatty acid production 55 .
Several dietary variables were signi cantly different between AA and EA women. AA women had signi cantly higher DIS and lower HEI scores, which coincides with higher carbohydrate percentage, fried food servings, and sweet servings, and signi cantly lower intakes of ber, beta and alpha carotene, vitamin E, fruit servings, and vegetable servings. Regarding stress variables, AA women also had signi cantly higher HC levels, and higher reports of lifetime discrimination (LED), and perceived stress (PSS).
Of the top 25 genera of our cohort, Bacteroides and Ruminococcus 1, Ruminococcus 2, and Ruminococcus torques were among the most abundant genera and Bi dobacterium, Lachnospiraceae, and Lactobacillus were among the least abundant genera, which are similar gut microbiota characteristics of those following a western, animal-based dietary pattern, high-sugar/high-fat diet, and of individuals who have undergone antibiotic treatments 56,57 . Blautia and Faecalibacterium, however, were the second and third most abundant genera of the cohort, genera that are reduced in individuals with cardiometabolic risk and disorders and increased with plant intake, respectively. Overall, it appears the cohort has a mixed abundance of bacteria important in various nutrient metabolism and host health.
Next, we observed some signi cant correlations between stress, dietary variables and gut microbiota diversity and genus abundance. Stress variables were not consistently or similarly correlated with gut diversity or genus abundance. PSS was negatively correlated with ASV count and Lachnospiraceae NK4A136 group. Depression did not correlate with any metrics. RED scores were correlated with more genera than any other stressor, with stronger correlations found with Blautia (negative), Alistipes (positive), Ruminococcus UCG 002 (positive), and Clostridium sensu (positive). LED scores were correlated with Alistipes (positive) and Ruminococcus UCG 002 (positive). Hair cortisol was positively correlated with Agathobacter, Ruminococcus 1, Bi dobacterium, and Eubacterium coprostanoli. Last, salivary cortisol was positively correlated with Subdoligranulum and Ruminococcus torques.
Regarding diet quality and gut microbiota diversity and abundance, trans fat intake was negatively correlated to ASV, which is similar to recent studies observing dietary fat intake and reduced alpha diversity 58-60 . Adversely, highglucose and high-fructose diets administered in mice have shown to have this effect on the alpha diversity and increases in proteobacteria phylum, one of the best sources of lipopolysaccharide (LPS), which trigger activation of the innate immune system and in ammatory conditions 61 . Variables thought to effect gut bacteria diversity were not correlated with ASV including high ber food groups (fruits, vegetables, ber), cal:EER ratio (energy balance), and fried food servings. Fruit servings, alpha carotene, vitamin C, and ber were all positively correlated with the genus Anaerostipes, a genus that has been found to increase in abundance following the consumption of prebiotic inulin, and improve reports of constipation and stool consistency 62 , possibly through its role in producing butyrate from lactate, contributing to colonic and GI health 63 . Alpha and beta carotene, vitamin C, and ber were also positively correlated with Lachnospiraceae NK4A136, however, this family of bacteria has been found to be controversial in its role in health and disease 64 . Lachnospiraceae NK4A136 was recently found to be restored after completion of a high-fat diet protocol that induced dysbiosis in mice 65 , was found to be diminished in a small pilot study including individuals with dementia 66 , and was among other short-chain fatty acid producing bacteria that were increased following an in ammation-reducing prebiotic trial in mice 67,68 . More recently, a 4-week tannin supplementation trial in humans was found to increase the abundance of healthy gut bacteria, including Lachnospiraceae NK4A136, and increase short chain fatty acid production 69 .
Variables of fat intake were inversely correlated with Akkermansia, which is consistent with studies reporting bodyweight and high-fat diet (HFD) in children and pregnant women [70][71][72] . Akkermansia supplementation in mice with high-fat induced obesity led to bene cial effects on weight, blood glucose control, and memory decay 70 .
Physiologically, Akkermansia administration has been shown to improve insulin sensitivity, attenuate adaptive changes related to caloric intake following cold exposure (negative energy balance), increase fat browning, induce anti-in ammatory effects through Treg cell induction in adipose tissue, and provide protective effect against atherosclerosis 72 . Interestingly, HFD has been shown to decrease Akkermansia abundance while sh oil consumption has been shown to increase Akkermansia 72,73 . Vegetable servings were inversely correlated with Alistipes, another controversial genus 74 that has been shown to increase in individuals following a calorie restricted high-fat diet that induced weight loss.
Some of our correlational ndings between dietary variables and gut microbiota genera were unexpected. For example, ber was inversely correlated with Eubacterium coprostanoli, a genus that has been found aid in the conversion of cholesterol to coprostanol, which is important in cholesterol excretion 75  Finally, variables we hypothesized to mediate the association between subjective stress and gut microbiota parameters did not hold true. In the current study, AA women reported higher perceived stress and lifetime discrimination, greater intakes of sweet and fried food servings, lower diet quality (increased DIS and reduced HEI scores), increased hair cortisol, and reduced ASV count. We hypothesized that a worse dietary intake or elevated cortisol in response to stress may be a way that AA have reduced alpha diversity, increased Alistipes, and reduced Blautia abundance. Although diet did not mediate the association between PSSUM and ASV count, the simple regressions between PSSUM and dietary variables ( ber, DIS, HEI) were signi cant (Table 3). Additionally, RED and LED were independently signi cantly associated with Alistipes and Blautia abundance. These separate associations are signi cant and reveal the independent associations between stress, diet quality, and speci c gut bacteria abundance. Previous research has similarly explored associations between added sugar intake, the abundance of various pathogenic gut bacteria, and cognitive function. Added sugar intake was found to increase in ammatoryrelated bacteria such as Proteobacteria, reduce bene cial bacteria, Bacteroidetes 72 and Lachnobacterium 73 , and increase species of Parabacteroides, which were found to impair memory performance 74 . The alterations in these bacteria abundances can lead to lipopolysaccharide-induced in ammation and impaired gut integrity through modi ed tight junctions and increased intestinal permeability.
Still, subjective stress questionnaires may not be the best predictor of dietary, physiologic, or gut microbiota metrics because of the nature of the variable accounting for past and cumulative experience, not current or acute response.
Additionally, RED, LED, and other discrimination questionnaires do not capture one's resilience or emotional response to these experiences. Induced stress interventions, such as the Trier social stress test 79 , and observation of immediate food choice, stress response, and gut microbiota changes will better capture these relationships and mediations. Further research in this area is important to explore effective protocols in reducing the burden of stress.
This research is not without limitations. First, the cross-sectional, observational design of this study cannot describe causal relationships. Additionally, the sample size and metabolically healthy state of our cohort may limit the strength of our ndings. Prospective studies observing the gut microbiota across the young adult period would be bene cial in learning about the resilience of one's gut microbiome and any signi cant changes that may be associated with declining or stable health, chronic experiences of stress, and diet quality. Future research should continue to observe the relationships among stress, diet, and the gut microbiome in early adolescence and young adulthood and conduct interventions involving mental health treatment and lifestyle modi cation, with hopes to halt or reverse concerning trends in obesity, prehypertension and hypertension, and pre-diabetes and diabetes rates of this population 14, 80-82 .

Conclusion
In conclusion, we observed several disparities between races in diet and stress but few differences in microbiota. We also found correlations between the top 25 gut microbiota genera and diet quality (animal, plant, fat, ber, etc) to be mixed. To date, it appears that greater microbial diversity is indicative of resilience and good health 83 , and reduced diversity is associated with autoimmune disorders and cardiometabolic disease 20 . There was not partial or full mediation by dietary or cortisol variables in the associations between subjective stress and alpha diversity or gut bacteria abundance.
The associations between perceived stress and ber, DIS, and HEI is important to note as young adults, especially those between the ages of 18 and 25, were shown to have experienced the greatest increase in symptoms consistent with major depression, suicidal thoughts, and serious psychological distress over the last decade (2008-2017). The prevalence of mental illness symptoms is cause for concern as individuals are less likely to engage in health-promoting behaviors. Choosing low-nutrient, high-calorie, sugary and fatty foods as a response to stress may alter the health and diversity of a young individuals gut microbiome, potentially leading to changes in metabolic health. Availability of data and materials

Abbreviations
The datasets generated and/or analyzed during the current study are not publicly available due to privacy protection of the participants but are available from the corresponding author on reasonable request.
Ethics approval and consent to participate