Indoxyl sulfate, a gut microbiome-derived uremic toxin, is associated with psychic anxiety and its functional magnetic resonance imaging-based neurologic signature

Indoxyl sulfate, a gut microbiome-derived uremic toxin, is associated

with psychic anxiety and its functional magnetic resonance imaging-

based neurologic signature

Christopher R Brydges

, Oliver Fiehn

, Helen S Mayberg

, Henry Schreiber

, Siamak

Mahmoudian Dehkordi

, Sudeepa Bhattacharyya

, Jungho Cha

, Ki Sueng Choi

, W Edward

Craighead

, Ranga R Krishnan

, A John Rush

, Boadie W Dunlop

*, Rima Kaddurah-

Daouk

4,10,11

* for the Mood Disorders Precision Medicine Consortium

Affiliations

West Coast Metabolomics Center, University of California, Davis, United States.

Department of Neurology and Neurosurgery, Icahn School of Medicine at Mount

Sinai, New York, NY, USA., Department of Psychiatry and Behavioral Sciences,

Emory University School of Medicine, Atlanta, GA, USA.

Division of Biology & Biological Engineering, California Institute of Technology,

Pasadena, CA, USA.

Department of Psychiatry and Behavioral Sciences, Duke University School of

Medicine, Durham, NC, United States.

Department of Biomedical Informatics, University of Arkansas for Medical Sciences,

Little Rock, AR, United States.

Department of Psychiatry and Behavioral Sciences, Emory University School of

Medicine, Atlanta, GA 30329, USA and Department of Psychology, Emory

University, Atlanta, GA

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

Department of Psychiatry, Rush Medical College, Chicago, IL, United States.

Department of Psychiatry and Behavioral Sciences, Duke University School of

Medicine, Durham, NC, United States; Department of Psychiatry, Health Sciences

Center, Texas Tech University, Permian Basin, TX, United States; Duke-National

University of Singapore, Singapore.

Department of Psychiatry and Behavioral Sciences, Emory University School of

Medicine, Atlanta, GA, United States

10.

Department of Medicine, Duke University, Durham, NC, United States

11.

Duke Institute of Brain Sciences, Duke University, Durham, NC, United States

*Corresponding Authors

Rima Kaddurah-Daouk, PhD

Duke University Medical Center

DUMC 3903, Blue Zone South

Durham, NC, USA

Phone: 919- 684-2611

Email: kaddu001@mc.duke.edu

Boadie Dunlop, MD

Emory University College of Medicine

12 Executive Park Dr. NE, Room 347

Atlanta, GA 30329, USA.

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

Phone: 404-727-8474

Email: bdunlop@emory.edu

Running Title

Indoxyl sulfate associated with psychic anxiety

Key words

Metabolomics, gut microbiome, indoles, indoxyl sulfate, anxiety, depression, major

depressive disorder

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

ABSTRACT

Background

: It is unknown whether indoles,

metabolites of tryptophan that are derived

entirely from bacterial metabolism in the gut, are associated with symptoms of

depression and anxiety.

Methods

: Serum samples (baseline, 12 weeks) were drawn from participants (n=196)

randomized to treatment with cognitive behavioral therapy (CBT), escitalopram, or

duloxetine for major depressive disorder.

Results

: Baseline indoxyl sulfate abundance was positively correlated with severity of

psychic anxiety and total anxiety and with resting state functional connectivity to a

network that processes aversive stimuli (which includes the subcallosal cingulate cortex

(SCC-FC), bilateral anterior insula, right anterior midcingulate cortex, and the right

premotor areas). The relation between indoxyl sulfate and psychic anxiety was

mediated only through the metabolite’s effect on the SCC-FC with the premotor area.

Baseline indole abundances were unrelated to post-treatment outcome measures,

which suggests that CBT and antidepressant medications relieve anxiety via

mechanisms unrelated to gut microbiota.

Conclusions

: A peripheral gut microbiome-derived metabolite was associated with

altered neural processing and with psychiatric symptom (anxiety) in humans, which

provides further evidence that gut microbiome disruption can contribute to

neuropsychiatric disorders that may require different therapeutic approaches.

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

INTRODUCTION

The gut microbiota impacts numerous aspects of human health and disease (1),

including neuropsychiatric disorders. The “microbiota–gut–brain axis” refers to a

bidirectional communication pathway that connects the central nervous system (CNS),

the gut, and the microbial community that inhabits the gastrointestinal tract (2). Within

this axis, the gut microbiota modulates central processes through the activation of

neuronal pathways (e.g., the vagus nerve) as well as through the production of

microbial metabolites and immune mediators that can trigger changes in

neurotransmission, neuroinflammation, and behavior (3-6).

Disruptions to the gut microbiome have been correlated with several neurological

disorders, including Parkinson’s disease, autism spectrum disorder, schizophrenia, and

major depressive disorder (MDD) (7-10), though the specific mechanisms that underlie

the role of the gut microbiota in these diseases is not fully understood. However,

research in preclinical rodent models shows that the gut microbiota is sufficient to alter

host behavior, as shown by the increase in anxiety- and depressive-like behaviors in

rodents after fecal microbiota transfer from humans with depression relative to those

that received transfer of fecal microbiota from demographic controls (11-12). Further,

transferred microbes resulted in altered metabolic states in the recipient mice that

displayed depressive-like symptoms (12). These data implicate the gut microbiota as

direct contributors to behaviors associated with depression and anxiety through their

metabolic effects. In this study, we explore gut microbiota-associated tryptophan

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

metabolism and correlate levels of metabolites to clinical symptoms and severity of

depression and anxiety in humans.

Tryptophan is an essential amino acid that can be metabolized in the gastrointestinal

tract via the serotonin, kynurenine, and indole metabolic pathways (Figure 1), which

have been associated with human maladies including autoimmunity, inflammatory

diseases, metabolic syndrome, and neurological diseases including depression and

anxiety disorders (13,14). Strikingly, the gut microbiota is exclusively responsible for the

conversion of tryptophan in the indole pathway, as there are no detectable levels of

indole or indole derivatives in gnotobiotic mice that lack a gut microbiome (15). Analysis

of biosynthetic pathways found that the genes necessary to make indole and indole

derivatives, such as indole-3-propionic acid (IPA), indole-3-acetic acid (IAA), and indole-

3-lactic acid (ILA), are found exclusively in the gut microbiome but not in mammalian

genomes (13) (Figure 1). These indoles can have important immunomodulatory effects

and are potent agonists for aryl hydrocarbon receptors (16) (AHRs), which regulate host

immunity and barrier function at mucosal sites (17).

Indole derivatives can also affect immune status in the brain, as some indole derivatives

(e.g., IPA and IAA) have anti-inflammatory effects on neurodegenerative diseases in the

experimental autoimmune encephalomyelitis (EAE) mouse model of multiple sclerosis

(18,19) as well as in a cell line model of Alzheimer’s disease (20). Other indole

derivatives can be further metabolized by host processes into molecules that may be

harmful to human health. Specifically, indole can be sulfonated in the liver into uremic

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

toxin indoxyl sulfate (IS), which crosses the blood-brain barrier (21) (Figure 1). IS, which

is normally cleared via the kidneys and excreted in the urine, is associated with

cardiovascular disease in patients who have chronic kidney disease via induction of

oxidative stress in endothelial cells (22), and peripheral IS concentrations are

associated with diminished cognitive function in renal dialysis patients (23).

IS is also associated with both neurodevelopmental and neurodegenerative diseases,

as levels of IS are increased in patients who have an autism spectrum disorder (24) or

Parkinson’s disease (25). Although the mechanistic role of IS in these diseases is

unknown, IS increases levels of oxidative stress and pro-inflammatory cytokine

signaling in astrocytes and mixed glial cells during

in vitro

administration (26), which

suggests that inflammation and reactive oxygen species may be involved. Further, IS

has been associated with behavioral defects in preclinical models of anxiety and

depression. The administration of IS into rodents’ drinking water results in increased

concentrations of IS in the brain and increased blood-brain barrier permeability in an

AHR-dependent manner, with accompanying increases in anxiety and cognitive deficits

(27,28). Monocolonization experiments with indole-producing

Escherichia coli

and

isogenic mutants have shown that indole production by gut bacteria is sufficient to drive

increases in anxiety- and depressive-like behavior in rats (29).

Taken together, the preclinical and clinical data indicate that indole derivatives provide

excellent models to study the microbiota-gut-brain axis given their connection to central

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

immune regulators (i.e., AHR), their link to human neurological diseases, and the

exclusivity of indole production to gut microbes.

To date, the effects of peripheral metabolic concentrations on neural functioning have

received little study, likely due to the paucity of datasets that contain concurrently

collected metabolomic and neuroimaging measures. Such research is crucial for

determining how changes in peripheral systems may yield alterations in brain function

that can produce clinically relevant symptoms such as depression, anxiety, or cognitive

impairment.

Using blood samples stored from the Prediction of Remission in Depression to

Individual and Combined Treatments (PReDICT) study, which was a large study of

treatment-naïve patients with MDD, we measured levels of four indole derivatives (IPA,

IAA, ILA, IS) to address the following questions:

1. Do levels of indoles and their ratios at baseline prior to treatment correlate with

depression and anxiety severity at baseline?

2. Do levels of indoles and their ratios at baseline correlate with specific individual

symptoms of depression?

3. Can symptom change after treatment with duloxetine, escitalopram, or cognitive

behavioral therapy (CBT) be predicted by baseline levels of indoles, and does

symptom change correlate with changes in levels of indoles after treatment?

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

4. Are there relationships between baseline peripheral metabolic concentrations of

indoles and brain resting state functional connectivity as determined using

functional Magnetic Resonance Imaging (fMRI).

MATERIALS AND METHODS

Study Design

The PReDICT study protocol (30), clinical results (31) and initial neuroimaging analyses

(32) have been published previously. The study was conducted through the Mood and

Anxiety Disorders Program of Emory University from 2007-2013. The study was

approved by Emory

s Institutional Review Board. All patients provided written informed

consent to participate.

PReDICT was designed to identify predictors and moderators of outcomes to three

randomly assigned first-line treatments for MDD: duloxetine, escitalopram, or CBT. The

study enrolled treatment-naïve adult outpatients, aged 18–65 years, who had current

MDD without psychotic symptoms. To be eligible for randomization, participants had to

score

≥

18 at screening and

≥

15 at baseline on the HAM-D. Key exclusion criteria

included the presence of any medically significant or unstable medication condition that

could impact study participation, safety, or data interpretation; any current eating

disorder, obsessive-compulsive disorder, or any current substance abuse or

dependence. Treatment was provided for 12 weeks with duloxetine (30-60 mg/day),

escitalopram (10-20 mg/day), or CBT (16 individual one-hour sessions).

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

Symptom Assessments

At the baseline visit, participants were assessed by trained interviewers using the HAM-

D and the HAM-A. The HAM-A is a 14-item measure that consists of two subscales,

“psychic anxiety” (items 1–6 and 14), and “somatic anxiety” (items 7–13) (33). Psychic

anxiety consists of the symptoms of anxious mood, tension, fears, depressed mood,

insomnia, impaired concentration, and restlessness. Somatic anxiety consists of

physical symptoms associated with the muscular, sensory, cardiovascular, respiratory,

gastrointestinal, genitourinary, and autonomic systems. Participants also completed the

QIDS-SR, which assesses the nine diagnostic symptom criteria for MDD (34). The

HAM-D, HAM-A, and QIDS-SR were repeated at the Week 12 visit.

Blood Sampling

Participants who met all eligibility criteria at the baseline visit underwent an antecubital

phlebotomy, without regard for time of day or fasting/fed status. Sampling was repeated

at the week 12 visit. Collected samples were allowed to clot for 20 minutes and then

centrifuged at 4

C to separate the serum, which was frozen at -80

C until being thawed

for the current analyses.

Neuroimaging

To explore associations between indole metabolites and brain function, we used the

resting state fMRI scans collected during the week prior to baseline, the details of which

have previously been published (32). Briefly, eyes-open scanning was performed for

7.4 minutes in a 3-T Siemens TIM Trio (Siemens Medical Systems, Erlangen,

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

Germany). Echo planar images were corrected for motion and slice-time acquisition.

Scans with head motion >2 mm in any direction were removed from the analysis. The

nuisance regressors, including head motion parameters, signal from the ventricle mask,

and signal from a region of local white matter, were cleaned. Subsequently, data were

applied a band-pass filter and smoothed using an isotropic Gaussian kernel of 8 mm full

width at half maximum. The imaging anatomical and functional data sets were co-

registered and normalized to standard Montreal Neurological Institute (MNI) 1-mm voxel

space. Image analysis was conducted using Analysis of Functional NeuroImages

(AFNI) (35,36) 3dvolreg. Consistent with our prior analyses (32), we used a region-of-

interest seed-based approach to assess the resting state functional connectivity (RSFC)

of the SCC. The SCC volume was defined using the Harvard-Oxford Atlas (37), and the

SCC was thresholded at 50% probability cent

ered on MNI coordinates 66, 24, –11. The

seeds comprised two 5-mm radius spheres, with a final volume of 485 mL each.

Utilizing 3dNetCorr (38), the mean time course of the bilateral seed was correlated

voxel-wise with the rest of the brain. The voxelwise correlation coefficients were then z-

scored by calculating the inverse hyperbolic tangent, yielding the seed-based RSFC

maps for analysis.

Metabolomics Data Acquisition

Metabolomics data focused on primary and polar metabolites using gas

chromatography – time of flight mass spectrometry (39). Briefly, 30

μ

l of plasma was

extracted at -20ºC with 1 mL degassed isopropanol/acetonitrile/water (3/3/2). Extracts

were dried down, cleaned from triacylglycerides using acetonitrile/water (1/1), and

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

derivatized with methoxyamine and trimethylsilylation. Samples (0.5

μ

L) were injected at

250ºC to a 30 m rtx5-SilMS column, ramped from 50-300ºC at 15ºC/min, and analyzed

by -70 eV electron ionization at 17 spectra/s. Raw data were deconvoluted and

processed using ChromaTOF vs. 4.1 and uploaded to the UC Davis BinBase database

(40) for data curation and compound identification (41). Result data were normalized by

SERRF software to correct for drift or batch effects (42).

Statistical Analyses

Indole abundance and ratios of each indole pair were included in all analyses. In order

to investigate the role of indoles in depression and anxiety symptomology at baseline,

partial Spearman rank correlations were conducted between the baseline

abundance/ratio of each indole and HAM-D 17-item total score, HAM-A total score,

HAM-A Psychic and Somatic subscores, QIDS-SR 16-item total score, and each

individual QIDS-SR item after accounting for age, sex, and body mass index (BMI).

Spearman correlations were also conducted between baseline indole abundance/ratio

and participant demographic factors (age, BMI, height, and weight). Additionally, sex

differences in baseline indole abundance/ratio were tested using Mann-Whitney

tests

and fold changes in median abundance/ratio between groups.

To investigate the potential effects of treatment on indoles, changes in indole

abundance from pre- to post-treatment were tested using Wilcoxon signed-rank tests

and fold changes. Partial Spearman rank correlations were conducted between post-

treatment indole abundance/ratio and post-treatment HAM-D 17-item total score, HAM-

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

A total score, and HAM-A Psychic and Somatic subscores, QIDS-SR 16-item total

score, and each individual QIDS-SR item, after accounting for age, sex, and baseline

BMI. The same analyses were also conducted with change from pre- to post-treatment

scores of all psychiatric variables, and also with fold changes from pre- to post-

treatment for each indole. Additionally, differences in indole post-treatment

abundance/ratio and fold change were investigated between each pair of treatment

response outcome groups (31) (treatment failure; partial response; response; remission)

by conducting Mann-Whitney

tests. This analysis was also repeated with baseline

indole abundances/ratios to investigate whether baseline levels of indoles may be

associated with treatment outcome. All reported

-values were adjusted for multiple

comparisons using the Holm method.

Neuroimaging analyses were conducted using AFNI (35) and jamovi (www.jamovi.org).

Of 122 participants who had an adequate quality of resting-state fMRI data (32), 80 had

metabolomic measurements and clinical scores. Voxel-wise linear regression analyses

were performed to examine the relationship between SCC-FC and IS or psychic anxiety

scores (uncorrected

< 0.005 and > 250 voxels cluster size). A conjunction analysis

identified overlapping areas between the SCC-FC of IS and the SCC-FC of psychic

anxiety scores. Subsequently, mediation analyses were performed using the Medmod

module (43) in jamovi. Three regions identified by the whole brain linear regression

analysis between SCC-FC and IS were used for the mediation analysis. We explored

each region, and combinations of the three regions, in the mediation models. For each

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

model, the direct and indirect effects were estimated using bootstrapping with 5000

samples.

RESULTS

Of the 344 patients randomized in PReDICT, 196 had metabolomic measures available

for analysis at baseline and 127 were available at week 12. The demographic and

clinical characteristics of the 196 participants are presented in Table 1.

Baseline Associations

Associations of indole metabolites with demographic variables

Supplemental Figure 1 shows a heat map of correlations between baseline indole

abundance/ratio and participant demographic variables. Abundance of ILA was

positively associated with age, height, and weight (all

s > 0.18, all

s < 0.040). The

ratios of IAA/ILA (negative associations) and IS/ILA (positive associations) were also

significantly associated with height and weight (all

s < 0.034). For sex differences,

abundance of IAA (Fold Changes (FC) = 1.19,

= 0.039) and ILA (FC = 1.37,

< 10

-9

and ratios of ILA/IPA (FC = 1.40,

= 0.004) and ILA/IS (FC = 1.16,

= 0.013) were all

found to be significantly higher in men than in women.

Associations of indole metabolites with depression and anxiety

Figure 2 shows a heat map of correlations between baseline indole abundance/ratio

and baseline levels of the 17-item Hamilton Depression Rating Scale (HAM-D) (44) total

score, Hamilton Anxiety Rating Scale (HAM-A) (45) total score, and HAM-A psychic and

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

somatic subscores. Greater abundance of IS was associated with higher scores on the

HAM-D 17-item total score (

= 0.21,

= 0.018), HAM-A total score (

= 0.26,

= 0.002),

and HAM-A psychic subscore (

= 0.31,

= 0.0001), but not on the HAM-A somatic

subscore. Additionally, the ratios of ILA/IS and IPA/IS were negatively correlated with

HAM-A total and psychic scores (all

s > -0.20, all

s < 0.033), for which a negative

correlation indicates that increasingly severe symptoms are associated with a relative

increase in IS and/or a relative decrease in ILA or IPA. Additionally, IPA/IS was

negatively correlated with HAM-D total score (

= -0.24,

= 0.001).

Associations of indole metabolites with individual symptoms of depression

Correlations between Quick Inventory of Depressive Symptoms – Self-Report (QIDS-

SR) (34) items, and total scores and indole abundances/ratios are presented in Figure

3. Of note, IS positively correlated with items 4 (hypersomnia;

= 0.22,

= 0.016) and 6

(decreased appetite;

= 0.20,

= 0.034), and the IPA/IS ratio negatively correlated with

QIDS-SR total score (

= -0.21,

= 0.027).

Treatment Effects

Compound abundance significantly increased from pre- to post-treatment for ILA (FC =

1.05,

= 0.006), but not for any other compound/ratio (all

s > 0.12). This indicates that

the treatments had limited overall effect on indole composition and levels.

For post-treatment indole abundances and ratios, there were significant correlations

between IAA/IS and QIDS-SR item 15 (feeling slowed down;

= 0.30,

= 0.007).

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

Additionally, post-treatment ILA abundance was significantly higher in men than in

women (FC = 1.24,

= 0.00001), as was the ILA/IPA ratio (FC = 1.54,

= 0.002).

Conversely, the IPA/IS ratio was lower in men than in women (FC = 0.80,

= 0.048).

No other significant post-treatment effects were observed.

For fold changes, change in IAA/IS ratio correlated with post-treatment scores of QIDS-

SR item 5 (feeling sad;

= 0.27,

= 0.032). No significant associations were observed

when correlating indole fold changes with any other post-treatment scores, or with

change in depression/anxiety scores (all

s > 0.10). Additionally, no sex differences

were observed for fold changes (all

s > 0.07), and no differences in fold changes were

observed between response outcome groups (all

s > 0.12).

Baseline levels of indoles and their ratios did not significantly correlate with changes in

symptoms for any measure or item (all

s < 0.15,

s > 0.59). Comparison of categorical

response outcomes also showed no meaningful differences in baseline indole

abundances or ratios. These analyses indicate that pre-treatment indole compound

abundances are not predictive of eventual treatment outcomes.

Associations of indole metabolites with brain resting state functional connectivity

Relationships of Subcallosal Cingulate Cortex – Functional Connectivity (SCC-FC) with

IS and with psychic anxiety scores are shown in Figure 4. IS abundance was positively

correlated with SCC-FC with the bilateral anterior insula, anterior midcingulate cortex

(aMCC), supplementary motor area (SMA), and right premotor area (Figure 4A).

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

Psychic anxiety scores showed a significant positive correlation with SCC-FC with the

left aMCC, right precuneus, and right premotor area; there was a negative correlation

with SCC-FC with the ventromedial prefrontal cortex, right orbitofrontal, and left

Brodmann Area 47 (Figure 4B). The conjunction analysis identified one overlapping

area: the right premotor region (Figure 4C).

The mediation analyses explored whether the association of IS with psychic anxiety was

mediated through its effects on SCC-FC. Figure 5A shows the overall association

between IS and psychic anxiety (z = 1.976,

= 0.048). Figure 5B shows that the

identified overlapping area in the SCC-FC analyses

–

the right premotor region

–

mediated the association between IS and psychic anxiety (indirect pathway: z = 2.138,

= 0.033). Because our whole brain SCC-FC analyses had also found IS concentrations

to be significantly associated with two other regions previously identified in

neuroimaging studies of anxiety (the right anterior insula and the aMCC, Figure 4A), we

conducted further mediation analyses incorporating these two regions along with the

right premotor region. Even though the three regions were highly correlated with each

other in their functional connectivity to SCC (Figure 5C), only the right premotor region

mediated the relationship between IS and psychic anxiety scores when all three regions

were included in the model (Figure 5D, indirect pathway: z = 1.991,

= 0.046).

DISCUSSION

Increasing evidence suggests that gut bacteria can complement human metabolism,

and that together they define the metabolome comprised of the collection of small

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

molecules in blood and in different organs. Bacteria can further metabolize compounds

available through human metabolism, food-intake, and/or human ingestion of chemicals.

Also, humans can further metabolize compounds produced by bacteria, which results in

human-bacteria co-metabolism and the production of a large number of chemicals that

can impact human health, including brain function. Examples include the metabolism of

cholesterol and its clearance mediated by bacteria, which can produce secondary bile

acids that we recently implicated in the pathogenesis of Alzheimer’s disease (46,47).

Several compounds produced from the metabolism of phospholipids and choline by gut

bacteria lead to compounds like trimethylamine N-oxide, which have been implicated in

cardiovascular diabetes and CNS disease (48,49).

Indoles represent a class of gut bacterially-derived compounds that are produced from

tryptophan, an essential amino acid that can also be converted (through separate

pathways) into tryptamine, serotonin, skatol, and melatonin, among other metabolites

involved in CNS functioning and diseases. Mounting evidence suggests that indoles

derived from gut bacterial metabolism exert significant biological effects and may

contribute to the etiology of cardiovascular, metabolic, and psychiatric diseases. To

date, research in this area has been mainly limited to experimental studies in model

systems.

In this investigation, we interrogated levels of four indoles produced by gut bacteria and

their relationship to anxiety and depression severity and response to treatment. At

baseline, IS abundance was found to positively correlate with severity of Psychic

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

Anxiety and total anxiety. IPA seems protective, as noted earlier, indicating that indoles,

as a class, can have mixed effects on neuropsychiatric health. Different strains of

bacteria can lead to the production of different indoles; for example, tryptophanase-

producing bacteria produce toxic IS while

Clostridium sporogenes

and other bacterial

strains produce protective IPA (50,51). Notably, IS levels did not meaningfully change

with treatment, and changes in IS were not correlated with improvement in depression

or anxiety measures. This suggests that gut microbiome composition and activity might

be modulated as an approach for developing additional classes of therapies effective in

the treatment of anxiety and depression.

This association between IS and anxiety seems to be mediated by the impact of IS on

the functional connectivity between the SCC and the right premotor region. IS

abundances were also associated with activation of a well-established network for the

processing and control of emotionally salient, particularly aversive, stimuli, comprising

the anterior insula and aMCC. Taken together, these results suggest that the co-

metabolism of tryptophan by certain gut microbiota that result in the production of IS,

which can induce anxiety through the activation of established brain networks, and that

existing treatments do not specifically resolve this pathogenetic process when they lead

to clinical improvement.

The neuroimaging analyses indicate that the effect of IS on psychic anxiety symptoms is

mediated through the functional connectivity of the SCC with the premotor cortex, as

part of a network involved in processing emotionally salient stimuli. The anterior insula,

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

aMCC, and supplementary motor area form a network that is involved in the attention

to, interpretation of, and control of emotional responses (52-56). The premotor cortex is

functionally and structurally connected to the SMA and the aMCC, which act together in

the preparation and readiness for voluntary movement in response to internal and

external stimuli (57), and the aMCC is a site of integration for the processing of pain and

motor control (58). Outputs from this network include projections to the spinal cord and

adrenal medulla (59), which may contribute to the sympathetic arousal and heightened

cortisol release under situations of psychic stress.

Although the activity of the premotor cortex has not been a major focus in studies of

anxiety and depression, Pierson and colleagues (60), using the electroencephalography

measure of contingent negative variation (which localizes to the premotor cortex (61)),

demonstrated abnormal activation of this region in anxious MDD patients compared to

MDD patients with psychomotor retardation. Ma and colleagues (62) found that patients

with generalized anxiety disorder have increased resting state functional connectivity

between the habenula and right premotor cortex. Others have found abnormal premotor

function in social anxiety disorder (63).

Our finding of an association between IS and activation of the insula bilaterally is

consistent with the insula’s known involvement in processes relevant to anxiety,

including emotional salience (64), empathy for others’ pain, and processing of

uncertainty (55,65-67). In contrast, we did not find an association between IS

abundances and somatic anxiety scores, nor was there an association of IS

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

abundances with functional connectivity of the SCC-posterior insula, the insular region

involved in sensorimotor integration. This reveals the specificity of the IS-anterior insula

association for psychic anxiety.

Conceptualizing psychic anxiety as a chronic aversive stimulus akin to long-term pain

may explain the positive correlation between higher IS levels and the QIDS-SR loss of

appetite item. In mice, inflammatory pain is inhibited in the presence of hunger,

mediated by neuropeptide Y signaling in the parabrachial nucleus (68). The association

of higher IS concentrations with both reduction in appetite and increased connectivity

between brain regions involved in pain processing (anterior insula and aMCC) may

indicate that the symptom of low appetite reflects a compensatory response to this

chronic anxiety-type pain.

Limitations of this study include the absence of a healthy control comparison group. We

lacked fecal samples we could analyze which would allow for a more direct correlation

between specific gut microbiome species and the IS measures. We could not determine

whether IS is the etiological agent of the anxiety because IS also acts to reduce the

integrity of the blood brain barrier (28), thereby creating the possibility that CNS

penetration by an alternative molecule in the periphery is responsible for the observed

association between anxiety and IS.

Taken together, our results indicate that increases in IS lead to the activation of an

established network that is involved in the processing and control of aversive stimuli, but

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

that the conscious experience of anxiety depends upon the degree of IS-related

activation of SCC-right premotor cortex functional connectivity. The absence of an

association between psychic anxiety scores and anterior insula/aMCC SCC-FC (Figure

4B) may indicate that although IS activates this control network in all patients, it is only

when network function is inadequate that psychic anxiety ensues in conjunction with

premotor activation in preparation for action (69). These analyses reveal the potential of

integrated peripheral metabolomic-neuroimaging analyses to reveal mechanistic

pathways that are associated with neuropsychiatric symptoms, especially for

characterizing the pathological impact of specific gut microbiome-derived metabolites.

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

Acknowledgements

We acknowledge the editorial services of Mr. Jon Kilner, MS, MA (Pittsburgh) and the

assistance of Ms. Lisa Howerton (Duke). This work was funded by grant support to Dr.

Rima Kaddurah-Daouk (PI) through NIH grants R01MH108348, R01AG046171 and

U01AG061359. Dr. Boadie Dunlop has support from NIH grants P50-MH077083 (PI

Mayberg), R01-MH080880 (PI Craighead), UL1-RR025008 (PI Stevens), M01-RR0039

(PI Stevens) and the Fuqua Family Foundations.

Disclosures

Dr. Dunlop has received research support from Acadia, Compass, Aptinyx, NIMH,

Sage, and Takeda, and has served as a consultant to Greenwich Biosciences, Myriad

Neuroscience, Otsuka, Sage, and Sophren Therapeutics.

Dr. Rush has received consulting fees from Compass Inc., Curbstone Consultant LLC,

Emmes Corp., Holmusk, Johnson and Johnson (Janssen), Liva-Nova, Neurocrine

Biosciences Inc., Otsuka-US, Sunovion; speaking fees from Liva-Nova, Johnson and

Johnson (Janssen); and royalties from Guilford Press and the University of Texas

Southwestern Medical Center, Dallas, TX (for the Inventory of Depressive Symptoms

and its derivatives). He is also named co-inventor on two patents: U.S. Patent No.

7,795,033: Methods to Predict the Outcome of Treatment with Antidepressant

Medication, Inventors: McMahon FJ, Laje G, Manji H, Rush AJ, Paddock S, Wilson AS;

and U.S. Patent No. 7,906,283: Methods to Identify Patients at Risk of Developing

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

Adverse Events During Treatment with Antidepressant Medication, Inventors: McMahon

FJ, Laje G, Manji H, Rush AJ, Paddock S.

Dr. Kaddurah-Daouk in an inventor on a series of patents on use of metabolomics for

the diagnosis and treatment of CNS diseases and holds equity in Metabolon Inc.

Author Contributions

CRB did analysis of data and helped write the manuscript; OF and his team generated

biochemical data and wrote its methods and helped with interpretation of findings; HSM,

WEC, JC, KSC and BWD did analysis connecting metabolomics data to imaging data

and helped with writing of manuscript; SMD, SB and HS helped with background

literature searches and with interpretation of findings; RRK, BWD and AJR helped with

interpretation of findings and clinical relevance; RKD is PI for project helped with

concept development, study design, data interpretation and connecting biochemical and

clinical data, and with writing of manuscript.

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

REFERENCES

1. Frank DN, Pace NR. (2008): Gastrointestinal microbiology enters the

metagenomics era.

Curr Opin Gastroenterol

24, 4-10.

2. Foster JA, Neufeld KAM. (2013): Gut–brain axis: how the microbiome influences

anxiety and depression.

Trends Neurosci

36, 315-312.

3. Sherwin E, Sandhu KV, Dinan TG, Cryan JF. (2016): May the force be with you: the

light and dark sides of the microbiota–gut–brain axis in neuropsychiatry.

CNS Drugs

30, 1019-1041.

4. Kelly JR, Kennedy PJ, Cryan JF, Dinan TG, Clarke G, Hyland NP. (2015): Breaking

down the barriers: the gut microbiome, intestinal permeability and stress-related

psychiatric disorders.

Front Cell Neurosci

9, 392.

5. Sampson TR, Mazmanian SK. (2015): Control of brain development, function, and

behavior by the microbiome.

Cell Host Microbe

17, 565-576.

6. Forsythe P, Kunze WA. (2013): Voices from within: gut microbes and the CNS.

Cell

Mol Life Sci

70, 55-69.

7. Adams JB, Johansen LJ, Powell LD, Quig D, Rubin RA. (2011): Gastrointestinal

flora and gastrointestinal status in children with autism–comparisons to typical

children and correlation with autism severity.

BMC Gastroenterol

11, 22.

8. Li C, Cui L, Yang Y, Miao J, Zhao X, Zhang J,

et al

. (2019): Gut microbiota differs

between Parkinson’s disease patients and healthy controls in northeast China.

Front Mol Neurosci

12, 171.

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

9. Aizawa E, Tsuji H, Asahara T, Takahashi T, Teraishi T, Yoshida S,

et al

. (2016):

Possible association of Bifidobacterium and Lactobacillus in the gut microbiota of

patients with major depressive disorder.

J Affect Disord

202, 254-257.

10.

Xu R, Wu B, Liang J, He F, Gu W, Li K,

et al

. (2020): Altered gut microbiota and

mucosal immunity in patients with schizophrenia.

Brain Behav Immun

85, 120-127.

11.

Kelly JR, Borre Y, O’Brian C, Patterson E, El Aidy S, Deane J,

et al

. (2016):

Transferring the blues: depression-associated gut microbiota induces

neurobehavioural changes in the rat.

J Psychiatr Res

82, 109-118.

12. Zheng P, Zeng B, Zhou C, Liu M, Fang Z, Xu X,

et al

. (2016): Gut microbiome

remodeling induces depressive-like behaviors through a pathway mediated by the

host’s metabolism.

Mol Psychiatry

21, 786-796.

13. Agus A, Planchais J, Sokol H. (2018): Gut microbiota regulation of tryptophan

metabolism in health and disease.

Cell Host Microbe

23, 716-724.

14. Gao J, Xu K, Liu H, Liu G, Bai M, Peng C,

et al

. (2018): Impact of the gut microbiota

on intestinal immunity mediated by tryptophan metabolism.

Front Cell Infect

Microbiol

8, 13.

15. Wikoff WR, Anfora AT, Liu J, Schultz PG, Lesley SA, Peters EC, Siuzdak G. (2009):

Metabolomics analysis reveals large effects of gut microflora on mammalian blood

metabolites.

Proc Natl Acad Sci U S A

106, 3698-3703.

16. Schroeder JC, Dinatale BC, Murray IA, Flaveny CA, Liu Q, Laurenzana EM,

et al

(2010): The uremic toxin 3-indoxyl sulfate is a potent endogenous agonist for the

human aryl hydrocarbon receptor.

Biochemistry

49, 393-400.

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

17. Lamas B, Natividad JM, Sokol H. (2018): Aryl hydrocarbon receptor and intestinal

immunity.

Mucosal Immunol

11, 1024-1038.

18. Rothhammer V, Mascanfroni ID, Bunse L, Takenaka MC, Kenison JE, Mayo L,

. (2016): Type I interferons and microbial metabolites of tryptophan modulate

astrocyte activity and central nervous system inflammation via the aryl hydrocarbon

receptor.

Nat Med

22, 586-597.

19. Rothhammer V, Borucki DM, Tjon EC, Takenaka MC, Chao C-C, Ardura-Fabregat

et al

. (2018): Microglial control of astrocytes in response to microbial metabolites.

Nature

557, 724.

20. Chyan YJ, Poeggeler B, Omar RA, Chain DG, Frangione B, Ghiso J, Pappolla MA.

(1999): Potent neuroprotective properties against the Alzheimer

-amyloid by an

endogenous melatonin-related indole structure, indole-3-propionic acid.

J Biol

Chem

274, 21937-21942.

21. Banoglu E, King RS. (2002): Sulfation of indoxyl by human and rat aryl (phenol)

sulfotransferases to form indoxyl sulfate.

Eur J Drug Metab Pharmacokinet

27, 135-

140.

22. Yu M, Kim YJ, Kang DH. (2011): Indoxyl sulfate–induced endothelial dysfunction in

patients with chronic kidney disease via an induction of oxidative stress.

Clin J Am

Soc Nephrol

6, 30-39.

23. Lin YT, Wu PH, Liang SS, Mubanga M, Yang YH, Hsu YL,

et al

. (2019): Protein-

bound uremic toxins are associated with cognitive function among patients

undergoing maintenance hemodialysis.

Sci Rep

9, 20388.

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

24. Gevi F, Zolla L, Gabriele S, Persico AM. (2016): Urinary metabolomics of young

Italian autistic children supports abnormal tryptophan and purine metabolism.

Mol

Autism

7, 47.

25. Cassani E, Barichella M, Cancello R, Cavanna F, Iorio L, Cereda E,

et al

. (2015):

Increased urinary indoxyl sulfate (indican): new insights into gut dysbiosis in

Parkinson's disease.

Parkinsonism Relat Disord

21, 389-393.

26. Adesso S, Magnus T, Cuzzocrea S, Campolo M, Rissiek B, Paciello O,

et al

(2017): Indoxyl sulfate affects glial function increasing oxidative stress and

neuroinflammation in chronic kidney disease: interaction between astrocytes and

microglia.

Front Pharmacol

8, 370.

27. Karbowska M, Hermanowicz JM, Tankiewicz-Kwedlo A, Kalaska B, Kaminski TW,

Nosek K,

et al

. (2020): Neurobehavioral effects of uremic toxin–indoxyl sulfate in

the rat model.

Sci Rep

10, 9483.

28. Bobot M, Thomas L, Moyon A, Fernandez S, McKay N, Balasse L,

et al

. (2020):

Uremic toxic blood-brain barrier disruption mediated by AhR activation leads to

cognitive impairment during experimental renal dysfunction.

J Am Soc Nephrol

31,

1509-1521.

29. Jaglin M, Rhimi M, Philippe C, Pons N, Bruneau A, Goustard B,

et al

. (2018):

Indole, a signaling molecule produced by the gut microbiota, negatively impacts

emotional behaviors in rats.

Front Neurosci

12, 216.

30.

Dunlop BW, Binder EB, Cubells JF, Goodman MM, Kelley ME, Kinkead B,

et al

(2012): Predictors of remission in depression to individual and combined treatments

(PReDICT): study protocol for a randomized controlled trial.

Trials

13, 106.

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

31.

Dunlop BW, Kelley ME, Aponte-Rivera V, Mletzko-Crowe T, Kinkead B, Ritchie JC,

et al

. (2017): Effects of patient preferences on outcomes in the Predictors of

Remission in Depression to Individual and Combined Treatments (PReDICT) study.

Am J Psychiatry

174, 546-556.

32. Dunlop BW, Rajendra JK, Craighead WE, Kelley ME, McGrath CL, Choi KS,

et al

(2017): Functional connectivity of the subcallosal cingulate cortex and differential

outcomes to treatment with cognitive-behavioral therapy or antidepressant

medication for major depressive disorder.

Am J Psychiatry

174, 533-545.

33. Maier W, Buller R, Philipp M, Heuser I. (1988): The Hamilton Anxiety Scale:

reliability, validity and sensitivity to change in anxiety and depressive disorders.

Affect Disord

14, 61-68.

34. Rush AJ, Giles DE, Schlesser MA, Fulton CL, Weissenburger J, Burns C. (1986):

The inventory for depressive symptomatology (IDS): preliminary findings.

Psychiatry

Res.

18, 65-87.

35. Cox RW. (1996): AFNI: software for analysis and visualization of functional

magnetic resonance neuroimages.

Comput Biomed Res

29, 162-173.

36.

Cox RW, Hyde JS. (1997): Software tools for analysis and visualization of fMRI

data.

NMR in Biomed.

10, 171-178.

37. Desikan RS, Ségonne F, Fischl B, Quinn BT, Dickerson BC, Blacker D,

et al

(2006): An automated labeling system for subdividing the human cerebral cortex on

MRI scans into gyral based regions of interest.

Neuroimage

31, 968-980.

38. Taylor PA, Saad ZS. (2013): FATCAT: (an efficient) functional and tractographic

connectivity analysis toolbox.

Brain Connect

3, 523-535.

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint

39. Fiehn O. (2016): Metabolomics by gas chromatography–mass spectrometry:

Combined targeted and untargeted profiling.

Curr Protoc Mol Biol

114, 30-34.

40. Fiehn O, et al. (2008): Quality control for plant metabolomics: reporting

MSI

‐

compliant studies.

Plant J

53, 691-704.

41. Lai Z, Tsugawa H, Wohlgemuth G, Mehta S, Mueller M, Zheng Y,

et al

. (2018):

Identifying metabolites by integrating metabolome databases with mass

spectrometry cheminformatics.

Nat Methods

15, 53-56.

42. Fan S, Kind T, Cajka T, Hazen SL, Wilson Tang WH, Kaddurah-Daouk R,

et al

(2019): Systematic error removal using random forest for normalizing large-scale

untargeted lipidomics data.

Anal Chem

91, 3590-3596.

43. Rosseel Y. (2012): Iavaan: An R package for structural equation modeling and

more.

J Stat Softw

48, 1-36 (2012).

44. Hamilton M. (1960): A rating scale for depression.

J Neurol Neurosurg Psychiatry

23, 56-62.

45. Hamilton M. (1959): The assessment of anxiety states by rating.

Br J Med Psychol

32, 50-55.

46. Mahmoudian Dehkordi S, Arnold M, Nho K, Ahmad S, Jia W, Xie G,

et al

. (2019):

Altered bile acid profile associates with cognitive impairment in Alzheimer’s

disease–An emerging role for gut microbiome.

Alzheimers Dement.

15, 76-92.

47.

Nho K, Kueider-Paisley A, MahmoudianDehkordi S, Arnold M, Risacher SL, Louie

et al

. (2019): Altered bile acid profile in mild cognitive impairment and

Alzheimer's disease: Relationship to neuroimaging and CSF

biomarkers.

Alzheimers Dement.

15, 232-244.

CC-BY-NC-ND 4.0 International license

available under a

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint

this version posted December 9, 2020.

;

https://doi.org/10.1101/2020.12.08.388942

doi:

bioRxiv preprint