Plasma and Fecal Metabolite Profiles in Autism Spectrum Disorder

Archival Report

Plasma and Fecal Metabolite Pro

fi

les in Autism

Spectrum Disorder

Brittany D. Needham, Mark D. Adame, Gloria Serena, Destanie R. Rose, Gregory M. Preston,

Mary C. Conrad, A. Stewart Campbell, David H. Donabedian, Alessio Fasano, Paul Ashwood,

and Sarkis K. Mazmanian

ABSTRACT

BACKGROUND:

Autism spectrum disorder (ASD) is a neurodevelopmental condition with hallmark behavioral

manifestations including impaired social communication and restricted repetitive behavior. In addition, many affected

individuals display metabolic imbalances, immune dysregulation, gastrointestinal dysfunction, and altered gut

microbiome compositions.

METHODS:

We sought to better understand nonbehavioral features of ASD by determining molecular signatures in

peripheral tissues through mass spectrometry methods (ultrahigh performance liquid chromatography

–

tandem mass

spectrometry) with broad panels of identi

fi

ed metabolites. Herein, we compared the global metabolome of 231

plasma and 97 fecal samples from a large cohort of children with ASD and typically developing control children.

RESULTS:

Differences in amino acid, lipid, and xenobiotic metabolism distinguished ASD and typically developing

samples. Our results implicated oxidative stress and mitochondrial dysfunction, hormone level elevations, lipid pro

fi

changes, and altered levels of phenolic microbial metabolites. We also revealed correlations between speci

fi

metabolite pro

fi

les and clinical behavior scores. Furthermore, a summary of metabolites modestly associated with

gastrointestinal dysfunction in ASD is provided, and a pilot study of metabolites that can be transferred via fecal

microbial transplant into mice is identi

fi

ed.

CONCLUSIONS:

These

fi

ndings support a connection between metabolism, gastrointestinal physiology, and com-

plex behavioral traits and may advance discovery and development of molecular biomarkers for ASD.

https://doi.org/10.1016/j.biopsych.2020.09.025

Many diseases are associated with informative metabolic

signatures, or biomarkers, that enable diagnoses, predict dis-

ease course, and guide treatment strategies. In contrast,

autism spectrum disorder (ASD) is diagnosed based on

observational evaluation of behavioral symptoms, including

reduced social interaction and repetitive/stereotyped behav-

iors (

). The average age of ASD diagnosis is between 3 and 4

years old (

), at which time children can receive behavioral

therapy, the gold standard treatment. Because earlier diag-

nosis improves ef

fi

cacy of behavioral therapies (

), molecular

biomarkers represent an attractive approach for identifying at-

risk populations and may aid development of personalized

therapies. This prospect is increasingly important given the

rising rate of ASD diagnoses, which currently stands at up to 1

in 59 children in the United States (

), with high variability of

worldwide estimates (approximately 1%

–

2%) (

) and no

Food and Drug Administration

–

approved drugs for core

behavioral symptoms.

Metabolic abnormalities have been reported in ASD (

though most studies have measured a small subset of

metabolites, and many outcomes have not been repro-

duced between cohorts. Mitochondrial dysfunction, which

heavily in

fl

uences systemic metabolism, is estimated to be

higher in ASD individuals than control subjects (5% vs.

approximately 0.01%) (

), and genes crucial for mitochon-

drial function are risk factors for ASD (

). The metabolic

abnormalities associated with mitochondrial dysfunction in

ASD affect cellular energy, oxidative stress, and neuro-

transmission in the gut and the brain (

–

). Other metabolic

pro

fi

les in ASD implicate aromatic and phenolic metabo-

lites, including derivatives of nicotinic acid, amino acid, and

hippurate metabolism (

–

). Various amino acids are

detected at differential levels across studies and across

sample types, but consistent patterns are dif

fi

cult to discern

(

–

Some discrepancies between studies are likely due to dif-

ferences in sample number, tissue analyzed, and methodol-

ogy. Other sources of variability include differences in

environmental factors, such as diet and gut bacteria, which

differ between ASD and typically developing (TD) populations

and can in

fl

uence each other (

–

). Diet is a major source of

circulating metabolites, impacting the metabolome directly or

indirectly through chemical transformation by the trillions of gut

microbes, the microbiome, which has been proposed to

modulate complex behaviors (

–

). Such proposed envi-

ronmental modulators of ASD may integrate with genetic risks

ª

2020 Society of Biological Psychiatry. This is an open access article under the

CC BY-NC-ND license (

http://creativecommons.org/licenses/by-nc-nd/4.0/

451

ISSN: 0006-3223

Biological Psychiatry March 1, 2021; 89:451

–

462

www.sobp.org/journal

Biological

Psychiatry

to impact behavioral end points through the actions of small

molecules produced in tissues outside the brain.

Herein, we present a comprehensive comparison of an

extensive panel of identi

fi

ed metabolites in human plasma and

feces from a large cohort of matched children with ASD and TD

children. We identi

fi

ed differential levels of metabolites ranging

from hormones, amino acids, xenobiotics, and lipids, many of

which correlate with clinical behavior scores. To our knowl-

edge, this is the

fi

rst study to concurrently evaluate paired

intestinal and systemic metabolomes in a high-powered anal-

ysis with a large number of identi

fi

ed metabolites, allowing

direct associations between metabolites previously highlighted

in ASD samples and discovering new metabolites of interest.

These

fi

ndings support the emerging concept of evaluating

nonbehavioral features in the diagnosis of ASD.

METHODS AND MATERIALS

Samples for this study, from children 3 to 12 years of age, were

collected through the MIND Institute, University of California,

Davis (

). ASD diagnosis was con

fi

rmed by trained staff

using the Autism Diagnostic Observation Schedule (ADOS)

and the Autism Diagnostic Interview-Revised. Diagnosis made

in subjects before 2013 was based on DSM-IV. TD participants

were screened using the Social Communication Questionnaire

and scored within the typical range (

,

15). An additional eval-

uation to determine gastrointestinal (GI) symptoms was

completed by 97 participants who also provided stool sam-

ples. GI status was determined using the GI symptom scale,

based on Rome III Diagnostic Questionnaire on Pediatric

Functional Gastrointestinal Disorders (

) and described in

detail elsewhere (

). See

Supplemental Methods

Supplement 1

for further information.

RESULTS

Plasma and Fecal Metabolomes Differ Between

ASD and TD

Plasma samples from 130 children with ASD and 101 TD

children were analyzed along with fecal samples collected from

a subset of these same children with ASD (

= 57) and TD

children (

= 40) (

Figure S1A

–

Supplement 1

). Samples and

metrics of behavioral and GI scores were obtained from the

MIND institute (

). The ASD group was strati

fi

ed into

subsets of children with GI symptoms (ASD

1

GI) or without GI

symptoms (ASD

2

GI), to explore potential effects of comorbid

intestinal dysfunction in ASD (40 of 130 ASD samples were

ASD

1

GI). This strati

fi

cation was based on symptoms associ-

ated with ASD, including diarrhea, constipation, and irritable

bowel syndrome

–

like symptoms.

Samples were analyzed by the global metabolite panel

(plasma and fecal samples) and the complex lipid panel

(plasma samples) from Metabolon, Inc. (Morrisville, NC), which

identi

fi

ed a total panel of 1611 plasma and 814 fecal metab-

olites (

Tables S1

–

Supplement 2

). Overall, we discovered

that 194 metabolites were differentially abundant between the

ASD and TD groups in plasma and 19 metabolites were

differentially abundant in feces (

Figure 1A, B

;

Tables S1

and

Supplement 2

). Using a quantitative assay for a targeted

panel of metabolites, we observed a high correlation between

relative abundance and precise concentration (

= approxi-

mately .97) for all tested metabolites except

-acetylserine

(

= .63) (

Figure S1H

Supplement 1

). Overall, these data

expand on previous evidence that the metabolomic pro

fi

les of

ASD and TD populations display differences not only in the gut

compartment but also in circulation, which may affect the

levels of metabolites throughout the body, including the brain

(

We then used random forest machine learning analysis to

determine if metabolite pro

fi

les can predict without bias

whether the sample came from an ASD or a TD donor.

Overall, the modest predictive accuracy of this machine

learning approach was 69% for plasma and 67% for feces.

To test whether focusing on the most discriminating me-

tabolites would improve the prediction, we repeated the

random forest analysis using the top 30 metabolites and

found that the predictive accuracy for plasma improved

slightly to 70% when using all ASD samples and to 74%

when using only ASD

2

GI samples. For fecal samples, the

predictive accuracy improved to 75% using all ASD samples

and to 67% using only ASD

2

GI samples. The top 30 me-

tabolites, calculated by measuring the mean decrease in

accuracy of the machine learning algorithm, are useful to

describe the strongest drivers of overall metabolic differ-

ences between ASD and TD populations. These most

discriminatory metabolites we

re primarily from the lipid,

amino acid, xenobiotic, and cofactor/vitamin super pathways

(

Figure 1E, F

). Several of these metabolites have been pre-

viously linked to ASD, such as steroids, bile acids, acylcar-

nitines, and nicotinamide metabolites (

–

). Further,

multiple molecules known to be produced or manipulated by

the gut microbiota also featured prominently, including 4-

ethylphenyl sulfate (4-EPS), which is elevated in an ASD

mouse model, and indolelactate, a microbe-derived trypto-

phan metabolite (

Figure 1E

)(

). The two most discrimi-

natory molecules in plasma (

Figure 1G, H

) and feces (

Figure 1I,

) are depicted, which are detected in almost every sample

except for 9-HOTre, which has a slightly lower percent

fi

ll in ASD

compared with TD samples (84% vs. 91%). Metabolites corre-

lating most strongly with these discriminatory metabolites are

closely related on a structural and metabolic level (

Figure 1K, L

Metabolite Levels Correlate With Clinical

Behavioral Scores

Using clinical metadata for children with ASD, we correlated

the levels of individual metabolites to the verbal, social, and

nonverbal scores of the following standard diagnostic mea-

sures, all conducted by trained health professionals: the

Autism Diagnostic Interview-Revised (a parent questionnaire)

and the cumulative ADOS severity score (SS) (

Tables S1, S4

Supplement 2

)(

). To contextualize the biological relevance of

different metabolomes between ASD and TD samples, indi-

vidual metabolites were integrated into biochemical pathways

for pathway enrichment analysis, revealing the degree of

change within each. We identi

fi

ed changes mostly in lipid,

xenobiotic, and nucleotide pathways associated with diverse

physiological processes (

Figure S2C

Supplement 1

). We

found that verbal and social scores primarily correlate with lipid

metabolism pathways and that nonverbal scores have the

Plasma and Fecal Metabolite Pro

fi

les in ASD

452

Biological Psychiatry March 1, 2021; 89:451

–

462

www.sobp.org/journal

Biological

Psychiatry

fewest correlations. Verbal scores show modest, negative

correlations with levels of individual metabolites, including

various plasma glycerolipids (i.e., diacylglycerols, glycer-

ophosphorylcholine, monoacylglycerols), and both plasma and

fecal levels of bile acids and sphingolipids. However, these

correlations are not particularly striking on an individual level

(

Figure S2D

Supplement 1

;

Tables S1, S4

Supplement 2

Social scores positively correlate with total free fatty acid levels

in plasma, most signi

fi

cantly with a C18 chain length

(

Figure S2E

Supplement 1

;

Table S1

Supplement 2

). After

false discovery rate correction, no individual metabolites

signi

fi

cantly correlated with nonverbal score.

nicotinamide

9-HOTrE

acetylcarnitine

mevalonate

hypotaurine

1,2-dilinoleoyl-GPC

pantothenate

3-OH-benzoate

sedoheptulose

carnitine

urea

succinate

trans-4-hydroxyproline

mevalonate 5-diP

1-linoleoylglycerol(18:2)

linolenoylglycerol(18:3)

betaine

2-linoleoylglycerol(18:2)

1-palmitoylglycerol(16:0)

androstenediol disulfate

xanthine

AMP

choline-phosphate

nicotinate

taurine

ribulose/xylulose

docosapentaenoate

1-oleoylglycerol(18:1)

xanthosine

valylleucine

mean-decrease-accuracy

Increasing Importance to Group Separation

Amino Acid

Carbohydrate

Cofactors and Vit.

Complex Lipids

Energy

Lipid

Nucleotide

Peptide

Xenobiotics

Super Pathway

Compared to TD

Total

All ASD

163

194

ASD-GI

120

ASD+GI

115

indolelactate

margaroylcarnitine

pregnenediol disulfate

beta-cryptoxanthin

4-allylphenol sulfate

androst. disulfate

(

)

ergothioneine

4-hydroxychlorothalonil

malate

androst. monosulfate

(

)

androst. disulfate

(

)

myristoylcarnitine

21-hydroxypreg. disulfate

threonate

pregnenediol sulfate

4-ethylphenylsulfate

PFOS

glutarylcarnitine

androst. monosulfate (2)

pregnenolone sulfate

stearoylcarnitine

sphinganine-1-P

arginine

palmitoylcarnitine

adrenoylcarnitine

hydroxy-CMPF

phytanate

DHEA-S

oxalate

α

-OH-isocaproate

lor b

Amino Acid

Carbohydrate

Cofactors and Vit.

Complex Lipids

Energy

Lipid

Nucleotide

Peptide

Xenobiotics

Super Pathway

mean-decrease-accuracy

ASD

0.0

0.5

1.0

1.5

3.0

ctat

****

ASD

nicotinamide

****

9-HOTr

ASD

012345

Acyl Carnitine (long)

Saturated FFA

CE Ester

LPC Ester

Acyl Carnitine (PUFA)

Androgenic Steroids

Pregnenolone Steroids

Enrichment Value

1.4E-3

7.5E-11

4.9E-5

7.0E-12

2.7E-7

0.02

0.05

Plasma

Feces

Lysophospholipid

Phospholipid Metabolism

Monoacylglycerol

Diacylglycerol

4.3E-9

7.2E-5

8.6E-3

3.0E-3

012345

Enrichment Value

Compared to TD

Total

All ASD

712 19

ASD-GI

35 8

ASD+GI

43 7

ASD

***

margaroylcarnitine

****

Correlated Metabolite

qval

Margaroylcarnitine

myristoylcarnitine (C14)

0.88

***

stearoylcarnitine (C18)

0.88

***

palmitoylcarnitine (C16)

0.87

***

palmitoleoylcarnitine (C16:1)

0.81

***

oleoylcarnitine (C18:1)

0.79

***

cysteinylglycine

0.78

***

LPC(17:0)

0.77

***

dihomo

linoleoylcarnitine (C20:2)

0.77

***

Indolelactate

phenyllactate (PLA)

0.51

3-(4-hydroxyphenyl)lactate

0.51

alpha-hydroxyisovalerate

0.48

2-hydroxy-3-methylvalerate

0.46

N-acetyltryptophan

0.41

kynurenate

0.40

imidazole lactate

0.38

LPC(14:0)

0.38

Correlated Metabolite

qval

Nicotinamide

hypotaurine

0.74

choline phosphate

0.67

glycerophosphorylcholine (GPC)

0.63

ergothioneine

0.57

acetylcarnitine (C2)

0.56

glycerol 3-phosphate

0.56

carnitine

0.53

adenosine 5'-monophosphate (AMP)

0.53

9-HOTrE

1-linolenoylglycerol (18:3)

0.68

diacylglycerol (16:1/18:2, 16:0/18:3)

0.67

13-HODE + 9-HODE

0.66

palmitoyl

-linolenoyl-

glycerol (16:0/18:3)

0.63

glycerol

0.57

oleoyl-linoleoyl-glycerol (18:1/18:2)

0.57

palmitoyl

-linoleoyl-

glycerol (16:0/18:2)

0.56

enterodiol

0.54

Figure 1.

Plasma and fecal metabolomes differ

between autism spectrum disorder (ASD) and typi-

cally developing (TD) groups.

(A, B)

The number of

signi

fi

cantly elevated and decreased metabolites (

,

.05,

,

.1) in ASD samples compared with the TD

control group by analysis of variance contrasts in

plasma and feces, respectively. Samples are strati-

fi

ed by all samples or samples without or with

gastrointestinal (GI) symptoms (

2

GI,

1

GI).

(C, D)

Pathway analysis results of human plasma and fecal

comparisons (all samples), indicating which metab-

olomic pathways are the most altered between

groups, with enrichment value plotted and

value to

the right of each bar. Metabolites within each

pathway could be observed at either higher or lower

levels, as this plot indicates only changes.

(E, F)

Top

30 most distinguishing metabolites between each

group in plasma and feces by random forest anal-

ysis, with mean decrease accuracy along the x-axis,

which is determined by randomly permuting a

variable, running the observed values through the

trees, and then reassessing the prediction accuracy.

If a variable is important to the classi

fi

cation, the

prediction accuracy will drop after such a permuta-

tion. Metabolites known to be produced by (aster-

isks) or in

fl

uenced by (triangles) the gut microbiota

are denoted. The super pathway to which each

metabolite belongs is indicated by color of sphere

and de

fi

ned in the legend.

(G, H)

Scaled intensity

values indicating relative levels of the most dis-

tinguishing molecules between ASD and TD

(all samples) in plasma. Asterisks indicate signi

fi

cance (

,

.05,

,

.1) in analysis of variance con-

trasts performed on total metabolomics

dataset. Data are represented as mean

6

SEM.

(I, J)

Scaled intensity values indicating relative

levels of the most distinguishing molecules

between ASD and TD (all samples) in feces. Aster-

isks indicate signi

fi

cance (

,

.05,

,

.1) in analysis

of variance contrasts performed on total

metabolomics dataset. Data are represented as

mean

6

SEM.

(K)

Top correlated plasma metabo-

lites that covary with margaroylcarnitine and

indolelactate.

(L)

Top correlated fecal

metabolites that covary with nicotinamide

and 9-HOTrE. AMP, adenosine monophosphate;

androst., androstane; CE, cholesterol ester;

DHEA-s, dehydroepiandrosterone sulfate; FFA,

free fatty acid; hydroxypreg., hydroxypregnenalone;

LPC, lysophosphatidylcholine; PFOS, per

fl

uoro-

octanesulfonic acid; PUFA, polyunsaturated fatty

acid; qval,

value.

Plasma and Fecal Metabolite Pro

fi

les in ASD

Biological Psychiatry March 1, 2021; 89:451

–

462

www.sobp.org/journal

453

Biological

Psychiatry

The ADOS SS correlated with diverse metabolite pathways,

including amino acids and food/plant component pathways

(

Figure S2C

Supplement 1

), which may be partly due to the

diverse array of symptoms integrated into the ADOS SS score.

On an individual metabolite level, we observed signi

fi

cant

positive correlations and trends between the ADOS SS and

metabolites in pathways of oxidative stress (cysteine, methi-

onine,

-adenosyl-

-methionine, and glutathione pathways)

(

Figure 2A

;

Table S1

Supplement 2

). Some of these mole-

cules were found in higher levels in the ASD fecal samples and

at lower levels in the ASD plasma samples, such as hypo-

taurine and, to a lesser degree, taurine (

Tables S1

and

ASD

1.5

2.0

2.5

Log ADOS SS

-4

-2

Log

g-AA (Log)

1.0

1.5

2.0

2.5

Log ADOS SS

Glutathione, Met,

Cys levels (Log)

-2

-4

1.0

1.5

2.0

2.5

Log ADOS SS

Energy Metabolites (Log)

-2

-4

1.0

ASD

Figure 2.

Metabolite levels correlate with Autism

Diagnostic Observation Schedule (ADOS) severity

score (SS).

(A)

Correlation of ADOS SS with me-

tabolites from the cysteine, methionine, and gluta-

thione pathways. Signi

fi

cant metabolites

corresponding to the linear regression in the graph

are listed along with Spearman coef

fi

cients and

values. Refer to color legend at bottom.

(B, C)

Scaled intensity values indicating relative levels of

hypotaurine in feces

(B)

and plasma

(C)

(all sam-

ples). Data are represented as mean

6

SEM. As-

terisks indicate signi

fi

cance in analysis of variance

contrasts performed on total metabolomics dataset

with a false discovery rate cutoff of

,

.1 (**

,

.01,

***

,

.001).

(D)

Correlations of gamma-glutamyl

amino acids with ADOS SS, with Spearman co-

fi

cients and

values to the right. Refer to color

legend at bottom.

(E)

The top 5 most positively

correlated plasma metabolites with ADOS SS, with

Spearman coef

fi

cients and

values to the right.

Refer to color legend at bottom. AA, amino acid;

Corr, correlation; Cys, cysteine; Cys-gly, cys-

teinylglycine; GSSG, glutathione disul

fi

de; Met,

methionine; pval,

value; qval,

value.

Plasma and Fecal Metabolite Pro

fi

les in ASD

454

Biological Psychiatry March 1, 2021; 89:451

–

462

www.sobp.org/journal

Biological

Psychiatry

Supplement 2

;

Figure 2B, C

;

Figure S3A, B

Supplement 1

which might indicate altered fecal production, excretion, or

differential uptake into the plasma potentially through varied

intestinal permeability. Taurine plays many roles throughout

the host and has previously been measured at altered levels in

ASD, although with little consensus (

–

Hypotaurine and taurine de

fi

ciency has been shown to lead to

defects in cell differentiation in the brain (

), and their dysre-

gulation could alter neuronal signaling (

Dysregulated amino acid degradation, homeostasis, and

import into the brain have been implicated as a cause of

neuronal stress in ASD, and supporting metabolomic data

have shown perturbations of various amino acid pathways,

such as glutamate, methionine, glutathione, and gamma-

glutamyl metabolites (

), some of

which show associations with the ADOS SS (

Table S1

Supplement 2

). Oxidative stress

–

related glutathione pathway

precursors, gamma-glutamyl amino acids, can in

fl

uence levels

of neurotransmitters such as GABA (gamma-aminobutyric

acid) (

), which is widely thought to play a role in ASD. In a

recent ASD study, an experimental treatment that led to

increased gamma-glutamyl amino acids and other redox

pathway metabolites correlated with improved behavior met-

rics in children (

). Here, we observed signi

fi

cant negative

correlations and trends between many gamma-glutamyl amino

acids and the ADOS SS (

Figure 2D

;

Table S1

Supplement 2

We also observed some perturbations in the urea cycle, which

processes the amino group of amino acids for excretion in

urine (

Figure S3C, D

Supplement 1

). Abnormalities in this

pathway can be indicative of altered amino acid degradation

observed in ASD and can lead to neurotoxic accumulation of

nitrogen-containing compounds in the blood (

Additionally, the 5 plasma metabolites most positively

correlated with ADOS SS all are involved in cellular energy

pathways (

Figure 2E

;

Table S1

Supplement 2

). Differences in

energy markers could indicate a neurodevelopmental pheno-

type during periods when the high lipid and energy require-

ment in the brain is crucial (

–

). Overall, these correlations

support the involvement of lipid, amino acid, and xenobiotic

metabolism in the etiology of ASD, as previously described

(

), and reveal new candidates for ASD biomarkers that

correlate with symptom severity.

Altered Levels of Cellular Energy and Oxidative

Stress Metabolites in ASD

In line with the observed correlations between oxidative stress

and cellular energy with the ADOS-SS, we further observed

altered levels of related metabolites between ASD and TD

samples. Metabolites that are markers of mitochondrial and

oxidative stress can offer a snapshot into cellular metabolic

states. These markers include acylcarnitines, which have

been highlighted in various ASD studies and are

established indicators of mitochondrial dysfunction

(

–

). Acylcarnitines are formed

to allow transport of lipids across the mitochondrial mem-

branes for beta-oxidation, and abnormal levels of these con-

jugated lipids accumulate to higher levels with a decrease in

beta-oxidation. Interestingly, high levels of short acylcarni-

tines are found in rodent models where ASD-like behaviors are

induced with the short-chain fatty acids valproic and propionic

acids (

We found differential levels of acylcarnitines in ASD,

creating a pattern of more abundant short-chain acylcarnitines

and less abundant long-chain acylcarnitines in the ASD

2

samples compared with TD samples (

Figure 3A

). Acylcarnitines

trend toward positive correlations with more severe social

behavior, an effect driven by structures with shorter moieties

(C2

–

C14) (

Figure S2C

Supplement 1

;

Table S1

Supplement 2

). In fecal samples, acetylcarnitine (C2) and free

carnitine were elevated in ASD (

Figure 3B, C

) and were highly

discriminatory (

Figure 1F

). Other mitochondrial markers in both

plasma and feces were also differentially abundant in ASD

(

Figure 3D

) along with markers of phospholipid metabolism,

which occurs largely in the mitochondria and was an enriched

pathway in fecal comparisons (

Figures 3D

and

). Such de-

fects in cellular metabolism support the theory that mito-

chondrial dysfunction may not only be comorbid with ASD but

also may be a potential contributing factor, as suggested by

numerous previous reports (

), and alterations to

levels of tricarboxylic acid cycle intermediates have been

observed in human ASD prefrontal cortex samples (

Together with the observed correlations between metabolites

and the ADOS SS, these results corroborate and extend a

growing body of research into altered mitochondrial meta-

bolism and oxidative stress in ASD.

Transfer of ASD Fecal Microbiota Into Mice Is

Accompanied by Metabolic Signatures

As microbial metabolites ranked highly in the random forest

machine learning analysis, we tested if any of the observed

metabolite differences in humans could be transferred to mice

via fecal microbial transplant. We selected 4 male donor

samples from each of the ASD and TD groups and colonized 2

or 3 male germ-free mice per donor for 3 weeks before col-

lecting plasma and fecal samples for metabolite pro

fi

ling and

bacterial DNA sequencing, respectively (

Figure S4A

Supplement 1

). Global metabolomic analysis revealed that

colonized mice modestly cluster by donor and group when

differential metabolites are considered in principal component

analysis (

Figure S4B

Supplement 1

). We selected the human

donors based on 4-EPS levels (

Figure S4C

Supplement 1

owing to the involvement of 4-EPS in an ASD mouse model

(

) and dysregulation of similar phenolic compounds in hu-

man ASD (

). 4-EPS is not produced by the host

and is strictly a bacterial metabolite (

). Surprisingly, we

observed 4-EPS levels in a bimodal distribution in mouse

samples (

Figure S4D, E

Supplement 1

). Despite the sur-

prising results with 4-EPS, many of the metabolites with the

highest fold change and lowest

value are indeed other

phenolic molecules, such as metabolites of hippurate, tyrosine,

and diet-derived phenols, some of which have been similarly

measured in a previous study transferring fecal microbes from

individuals with ASD into mice (

Figure S4F

Supplement 1

;

Table S10

Supplement 2

)(

). While preliminary, these

studies reveal dysregulated microbiome-mediated effects on

xenobiotic pathways and phenolic metabolites in ASD.

Plasma and Fecal Metabolite Pro

fi

les in ASD

Biological Psychiatry March 1, 2021; 89:451

–

462

www.sobp.org/journal

455

Biological

Psychiatry

Steroid Hormone Levels Are Elevated in ASD

Multiple human ASD studies have examined the levels of

speci

fi

c steroid metabolites within androgenic, pregnenolone,

and progesterone metabolism, with some

fi

nding aberrant

levels, positive correlations with ASD severity, and behavioral

improvement following treatments that lower levels of certain

hormones (

–

). In contrast, in a recent clinical trial of

children with ASD given an antioxidant treatment, levels of

pregnenolones and androgens increased and correlated with

improved behavior (

). In our dataset, we found increases of

many detected metabolites within the pregnenolone and

androgen pathways (

Figures 1C, 4A

;

Table S1

Supplement

). This is an indicator that the physiological pathways asso-

ciated with the downstream metabolism of cholesterol are

signi

fi

cantly altered between ASD and TD populations. There

does not appear to be a global change in steroid metabolism,

as most primary bile acid and sterol metabolites were unaf-

fected (

Tables S1

and

Supplement 2

). We observed some

elevation of these hormone levels independent of sex, which is

notable considering the male bias in ASD, re

fl

ected in our

primarily male sample set (7%

–

15% female) (

Figure S6A, B

Supplement 1

)(

). Because a cluster of our samples are from

older individuals in the ASD group, and to account for age-

dependent increases in androgens (

Table S7

Supplement

), we strati

fi

ed by age and still observed heightened andro-

genic and pregnenolone metabolite levels in ASD sub-

populations (

Figure S6C

Supplement 1

). Taken together,

these data indicate that steroidal hormone metabolism may be

altered in the ASD population relative to TD samples and that

these differences are not driven solely by sex or age differ-

ences in our cohort.

Lowered in ASD:

p<0.05, q<0.1

p<0.05, q>0.1

Elevated in ASD:

p<0.05, q<0.1

p<0.05, q>0.1

No change:

k

TAG (

SAT

)

DAG (

PUFA

)

FFA (

SAT

PUFA

)

acyl CoA

carnitine

acyl carnitine

(

short

long

)

β-Oxidation

glycerol

Glycolysis

Mitochondria

Plasma Mitochondrial

Markers

TCA

succinate

aconitate

malate

choline

lysoPC

GPC

lysoPE

GPE

Dimethylgly

ergothioneine

betaine

Phospholipid Metabolism

betaine

carnitine

acetyl carnitine

DAG

Fecal Mitochondrial

Markers

ace

nitin

****

(C2)

ASD

-1.0

-0.5

0.0

0.5

1.0

Log2 Fold Change

3OH-C4

3OH-C4(2

)

C6-DC

C2C

8-DC

C10:1

16:1

18-

22:

C18:1

18:3

C20

C20:2

20:1

C20:3or6

C26:1

C24

C26

C22

(All)

(Al

(All)

(

All)

(All)

Figure 3.

Autism spectrum disorder (ASD) abnormalities within cellular energy and oxidative stress metabolites.

(A)

Log

fold change of acylcarnitines in the

plasma of ASD samples without gastrointestinal symptoms compared with typically developing (TD) control samples. Signi

fi

cance indicated by color according

to legend below, determined by analysis of variance contrasts. Star indicates that a trend or signi

fi

cance was observed but only in the comparison between all

samples.

(B, C)

Scaled intensity values indicating relative levels of acetylcarnitine (C2) and carnitine in ASD fecal samples compared with TD control samples

(all samples). Data are represented as mean

6

SEM. Asterisks indicate signi

fi

cance in analysis of variance contrasts (false discovery rate cutoff

,

.1)

performed on total metabolomics dataset (**

,

.01, ****

,

.0001).

(D)

Schematic of mitochondrial markers and other metabolites closely associated with

cellular energy in plasma (within center box) and feces (boxed to left). *Signi

fi

cant only in ASD samples without gastrointestinal symptoms. Color of text in-

dicates direction and signi

fi

cance of change according to legend above. DAG, diacylglycerol; FFA, free fatty acid; GPC, glycerophosphocholine; GPE,

glycerophosphoethanolamine; GPG, glycerophosphoglycerol; GPI, glycerophosphoinositol; GPS, glycerophosphoserine; PC, phosphatidylcholin

e; PE,

phosphatidylethanolamine; PUFA, polyunsaturated fatty acid; SAT, saturated fatty acid; TAG, triacylglycerol.

Plasma and Fecal Metabolite Pro

fi

les in ASD

456

Biological Psychiatry March 1, 2021; 89:451

–

462

www.sobp.org/journal

Biological

Psychiatry

Lipid Metabolite Levels Differ in ASD

Lipids are crucial for energy storage, cellular membrane

integrity, and cell signaling. They play a variety of roles in the

central nervous system, and their dysregulation has been

linked to ASD (

). Phospholipids have been measured at

lower levels, while long-chain fatty acids are reportedly

elevated, but polyunsaturated fatty acids (PUFAs) have been

measured at both higher and lower levels, depending on the

cohort (

). Here, we performed a comprehensive,

quantitative metabolite analysis on complex lipids. The con-

centrations of 999 lipids were quanti

fi

ed, and 13.7% of these

were signi

fi

cantly different in the ASD samples, with 4.6%

increased and 9.1% decreased compared with TD samples.

Many of these differentially abundant lipids included phos-

pholipids, cholesterol esters, and glycerolipids (

Figure S7A

Supplement 1

). In general, shorter (14

–

18 chain length) satu-

rated fatty acids were less abundant in ASD throughout the

lipid classes (

Figure 4C

;

Figure S7B

Supplement 1

PUFA lipid levels were also different, with elevations in

diacylglycerols and free fatty acids and an enrichment of the

18:2 (linolenic) chain length in most lipid classes (

Figure 4C

Fecal samples generally trended in the opposite direction from

plasma in PUFA lipids (

Figure S7C

Supplement 1

;

Table S4

Supplement 2

). PUFA lipids containing linolenic acids are

precursors to other PUFAs (arachidonic and docosahexaenoic

acids) important for brain development, function, and structural

integrity (

). Future studies are needed to determine if these

changes in lipid levels contribute to ASD.

Considering the prevalence of intestinal issues in ASD, GI-

dependent metabolite analysis between individuals may pro-

vide useful insights. Accordingly, we found a total of 87 and 24

metabolites in plasma and feces, respectively, that were

potentially differential within ASD

1

GI compared with ASD

2

individuals by analysis of variance contrasts, but none passed

the false discovery rate cutoff (

Figure S5A, B

Supplement 1

At the pathway level, we found enrichments in free fatty acids

between ASD

1

GI and ASD

2

GI plasma samples (

Figure S5C

Supplement 1

), but none in fecal samples (

Table S4

Supplement 2

). In plasma, free fatty acids of multiple chain

lengths trended lower in ASD

1

GI compared with ASD

2

samples (

Figure S5D

Supplement 1

). Some fatty acids, such

as PUFAs, are anti-in

fl

ammatory, and their lower levels may

contribute to GI dysfunction directly (

). We did not observe

signi

fi

cant correlations between lipid (or other) metabolite

patterns and GI symptoms or disease, including diarrhea,

constipation, or irritable bowel syndrome. While these results

are interesting, the implications of correlating an altered

metabolome and GI symptoms in ASD remain to be

determined.

Differential Phenolic Xenobiotic Metabolite Levels

in ASD

Phenolic metabolites, a diverse structural class of thousands

of molecules containing a phenol moiety, come from dietary

ingredients or biotransformation of aromatic amino acids by

the gut microbiota. In their free phenolic forms, they can be

readily absorbed through intestinal tissues; however, microbial

modi

fi

cation of these molecules can signi

fi

cantly alter their

absorption, bioavailability, and bioactivity (

), leading to

various bene

fi

ts or harm to the host (

). In fact, altered

levels of phenolics have been highlighted in many ASD

metabolomic studies (

–

). How-

ever, a consensus of these metabolite changes in ASD has yet

FATTY ACI DS

Composi ti on Matr i x

SAT

MUFA

PUFA

12:0

14:0

5:0

6:0

17:0

0:0

2:0

24:0

4:1

16:1

0:1

2:1

24:1

0:2

20:3

0:4

2:4

18:3

0:5

2:5

22:6

LIPID CLASSES

Phosphatidylcholine(

PC)

0.79

0.86

0.85

0.83

1.03

0.93

Phosphatidylethan olam in e(

PE)

1.02

0.98

0.92

1.03

1.43

1.07

0.96

Phosphatidylinositol(

PI)

1.09

0.93

0.97

LysoPC(

LPC)

0.85

0.83

0.97

0.82

1.19

1.14

1.05

LysoPE(

LPE)

0.94

0.86

0.92

1.08

C hol ster yl ester (

CE)

0.85

0.82

0.84

0.95

Triacylglycerol(

TAG)

0.83

0.81

0.83

0.90

1.1

0.90

Diacylglycerol(

DAG)

0.93

1.08

1.11

Free fatty acid(

FFA)

0.95

0.96

0.72

1.11

0.93

1.08

1.09

1.06

1.11

cholest erol

nen

nes

pregnenolone-S

1.61

-hydroxypregnenolone-SS

1.51

pregnenediol

-SS

1.52

pregnenediol

-S

1.51

dehydroisoandrosterone

-S

1.78

2.69 (

)

16a

-hydroxy DHEA 3-S

1.88

androsterone glucuronide

1.82

et iocholanolone glucuronide

2.2

androstenediol (3β,17β)

-S (1)

2.49

androstenediol (3β,17β)

-S (2)

1.82

androstenediol (3β,17β)

SS ( 1)

2.17

3.22

androstenediol (3β,17β)

SS ( 2)

1.7

2.37 (

)

androstenediol (3α, 17α)

S(2)

1.8

androstenediol (3α, 17α)

S(3)

1.75

5α

-androstan-3β,17β-diol-SS

2.21

3.39

andro steroid-S (1)

1.81

+ / - = Sig. in only + or - G I group

No sig. change:

Not det ect ed:

+/- = Only ASD + or - GI group

Lowered in ASD:

Elevat ed in ASD:

p< 0.05; q< 0.1

p< 0.05; q> 0.1

detected i n < 70% of sam pl es

Figure 4.

Steroid hormone levels are elevated in autism spectrum disorder (ASD), and other lipid metabolite levels differ in ASD.

(A)

Signi

fi

cant alterations to

levels of all metabolites are detected in the pregnenolone, progestin, and androgen steroid pathways in plasma (P) and feces (F), with colors indicat

ing

signi

fi

cance and fold change direction according to legend at bottom right and numerical fold change in text within the box.

(B)

Complex lipid panel results for

all ASD plasma samples compared with typically developing control samples with acyl chain length of lipids across the top, described by chain length a

degree of unsaturation and categorized by saturated (SAT), monounsaturated (MUFA), and polyunsaturated (PUFA) fatty acids. Lipid classes are list

ed along

the left. Direction of change and signi

fi

cance are indicated by the legend. Signi

fi

cance determined by analysis of variance contrasts; false discovery rate cutoff

,

.1. DHEA, dehydroepiandrosterone; GI, gastrointestinal (symptoms); sig., signi

fi

cant.

Plasma and Fecal Metabolite Pro

fi

les in ASD

Biological Psychiatry March 1, 2021; 89:451

–

462

www.sobp.org/journal

457

Biological

Psychiatry