of 22
| Host-Microbial Interactions |
Research Article
Microbe-derived uremic solutes enhance thrombosis potential in
the host
Ina Nemet,
1,2
Masanori Funabashi,
3,4,5
Xinmin S. Li,
1,2
Mohammed Dwidar,
1,2
Naseer Sangwan,
1,2
Sarah M. Skye,
1,2
Kymberleigh A.
Romano,
1,2
Tomas Cajka,
6
Brittany D. Needham,
7
Sarkis K. Mazmanian,
7
Adeline M. Hajjar,
1,2
Federico E. Rey,
8
Oliver Fiehn,
6
W. H.
Wilson Tang,
1,2,9
Michael A. Fischbach,
3,4,5,10
Stanley L. Hazen
1,2,9
AUTHOR AFFILIATIONS
See
affiliation
list on p.
18
.
ABSTRACT
p
-Cresol sulfate (
p
CS) and indoxyl sulfate (IS), gut microbiome-derived
metabolites, are traditionally associated with cardiovascular disease (CVD) risks in the
setting of impaired kidney function. While pharmacologic provision of
p
CS or IS can
promote pro-thrombotic phenotypes, neither the microbial enzymes involved nor direct
gut microbial production have been linked to CVD. Untargeted metabolomics was
performed on a discovery cohort (
n
= 1,149) with relatively preserved kidney function,
followed by stable isotope-dilution mass spectrometry
quantification
of
p
CS and IS in an
independent validation cohort (
n
= 3,954). Genetic engineering of human commensals
to produce
p
-cresol and indole gain-of-function and loss-of-function mutants, followed
by colonization of germ-free mice, and studies on host thrombosis were performed.
Systemic
p
CS and IS levels were independently associated with all-cause mortality.
Both
in vitro
and within colonized germ-free mice
p
-cresol productions were recapitu­
lated by collaboration of two organisms: a
Bacteroides
strain that converts tyrosine
to 4-hydroxyphenylacetate, and a
Clostridium
strain that decarboxylates 4-hydroxyphe­
nylacetate to
p-
cresol. We then engineered a single organism,
Bacteroides thetaiotao­
micron
, to produce
p-
cresol, indole, or both metabolites. Colonizing germ-free mice
with engineered strains
,
we show the gut microbial genes for
p-
cresol (
hpdBCA
) and
indole (
tryptophanase
) are
sufficient
to confer a pro-thrombotic phenotype
in vivo
.
Moreover, human fecal metagenomics analyses show that abundances of
hpdBCA
and
tryptophanase
are associated with CVD. These studies show that
p
CS and IS, two
abundant microbiome-derived metabolites, play a broader potential role in CVD than
was previously known. They also suggest that therapeutic targeting of gut microbial
p-
cresol- and indole-producing pathways represent rational targets for CVD.
IMPORTANCE
Alterations in gut microbial composition and function have been
linked to numerous diseases. Identifying microbial pathways responsible for producing
molecules that adversely impact the host is an important
first
step in the development
of therapeutic interventions. Here, we
first
use large-scale clinical observations to link
blood levels of
defined
microbial products to cardiovascular disease risks. Notably, the
previously
identified
uremic toxins
p
-cresol sulfate and indoxyl sulfate were shown to
predict 5-year mortality risks. After identifying the microbes and microbial enzymes
involved in the generation of these uremic toxins, we used bioengineering technologies
coupled with colonization of germ-free mice to show that the gut microbial genes that
generate
p
-cresol and indole are
sufficient
to confer
p
-cresol sulfate and indoxyl sulfate
formation, and a pro-thrombotic phenotype
in vivo
. The
findings
and tools developed
serve as a critical step in both the study and targeting of these gut microbial pathways
in
vivo
.
November/December 2023 Volume 14
Issue 6
10.1128/mbio.01331-23
1
Invited Editor
Joseph Zackular, University of
Pennsylvania, Philadelphia, Pennsylvania, USA
Editor
Edward G. Ruby, University of Hawaii at
Manoa, Honolulu, Hawaii, USA
Address correspondence to Stanley L. Hazen,
hazens@ccf.org, or Michael A. Fischbach,
fischbach@fischbachgroup.org
.
Ina Nemet and Masanori Funabashi contributed
equally to this article. Author order was chosen at
random.
S.L.H. reports being named as co-inventor on
pending and issued patents held by the Cleveland
Clinic relating to cardiovascular diagnostics and
therapeutics, being a paid consultant formerly for
Procter & Gamble in the past, and currently being
with Zehna Therapeutics. He also reports having
received research funds from Procter & Gamble and
Zehna Therapeutics and being eligible to receive
royalty payments for inventions or discoveries
related to cardiovascular diagnostics or therapeutics
from Procter & Gamble, Zehna Therapeutics, and
Cleveland HeartLab, a wholly owned subsidiary of
Quest Diagnostics. M.A.F. is a co-founder and director
of Federation Bio and Kelonia and a co-founder
of Revolution Medicines. M.A.F. also reports the
following: Ownership Interest: Kelonia, NGM Bio;
Patents or Royalties: Federation Bio; and Advisory
or Leadership Role: Federation Bio, Kelonia, NGM
Bio, The Column Group, and Chan Zuckerberg
Science. W.H.W.T. reports being a consultant for
Sequana Medical A.G., Owkin Inc., Relypsa Inc.,
and PreCardiac Inc., having received honorarium
from Springer Nature for authorship/editorship, and
American Board of Internal Medicine for exam
writing committee participation—all unrelated to
the subject and contents of this paper. The
other authors have reported that they have no
relationships relevant to the contents of this paper
to disclose.
See the funding table on p.
19
.
Received
26 May 2023
Accepted
25 September 2023
Published
10 November 2023
Copyright © 2023 Nemet et al. This is an open-access
article distributed under the terms of the
Creative
Commons Attribution 4.0 International license
.
Downloaded from https://journals.asm.org/journal/mbio on 21 December 2023 by 131.215.249.156.
KEYWORDS
gut microbes, uremic toxins,
p
-cresol sulfate, indoxyl sulfate, cardiovascu­
lar disease, mortality
C
ardiovascular disease (CVD) remains the leading cause of mortality worldwide;
however, despite
significant
advances in preventive CVD treatments and preven­
tion
efforts,
there remains a
significant
“residual CVD risk,” with numerous atherothrom­
botic events and cardiovascular mortality risk occurring even among optimally treated
individuals (
1
–4
). The search for additional causative factors beyond traditional CVD risk
factors remains a
significant
area of investigation. One promising area in this regard is
the gut microbiome (
5,
6
).
Differences
in gut microbial composition and function have
been associated with CVD, including production of some bioactive metabolites (
5
,
7
–10
).
While most of studies on gut microbes and CVD are correlative and associative, multiple
mechanistic studies are identifying gut microbial metabolites as direct contributors to
CVD (
7
,
11
–16
).
p
-Cresol sulfate (
p
CS) and indoxyl sulfate (IS), products derived from
microbial fermentation of the aromatic amino acids tyrosine and tryptophan, respec­
tively, are associated with CVD mortality risk in individuals with chronic kidney disease
(CKD) and end-stage renal disease (
17
–19
).
Termed “uremic solutes” because levels of
p
CS and IS accumulate as renal function
declines, these compounds are also considered “uremic toxins” because they have
been implicated in mediating adverse phenotypes in the setting of renal disease (
20
).
The synthesis of both
p
CS and IS are also “metaorganismal,” since each is co-synthe­
sized by an initial (obligate) gut microbiota-dependent process, followed by the host
enzymatic transformations that facilitate metabolite excretion. Each molecule begins as
a byproduct of bacterial fermentation of aromatic amino acids in a protein-rich diet.
For IS production, tryptophan from the diet can be degraded into indole by microbial
tryptophanase genes encoded in the genome of a variety of gut commensals, as
we recently showed (
21
). The gut microbiota-derived indole, once absorbed via the
portal circulation, can then be oxidized by CYP2E1 and sulfated by SULT1A1, both
endogenous host enzymes. Synthesis of
p
CS follows a similar overall trajectory. Dietary
protein-derived tyrosine is metabolized by gut microbes via various mechanisms to
yield
p
-cresol, which, following absorption into the host, can then undergo sulfation
by SULT1A1 to yield
p
CS (
22,
23
). While the host enzymes in these metaorganismal
biochemical transformations are well-known and characterized, the gut microbes and
the microbial genes utilized to produce the precursor molecules in these metaorganis­
mal pathways remain poorly understood.
Our interest in
p
CS and IS in the present studies began as a result of untargeted
metabolomics investigations, which
identified
microbial precursor of
p
CS as candidate
molecule whose circulating (blood) levels predict incident CVD risks among subjects
with normal renal function. Through subsequent targeted mass spectrometry analysis
on an independent cohort, we now show that
p
CS and IS are associated with overall
mortality in individuals with preserved kidney function. We then identify a missing link
in the microbial biosynthetic pathway of
p
-cresol and take advantage of this
finding
to engineer strains of
Bacteroides thetaiotaomicron
that produce
p
-cresol, indole, or
both. By colonizing germ-free mice with these strains, we
confirm
that microbial genes
responsible for production of either
p
-cresol or indole lead to formation of
p
CS and
IS, respectively. We then further demonstrate that this is
sufficient
to confer within
the host, formation of either
p
CS or IS, and induction of a pro-thrombotic phenotype.
Finally, in human metagenomics analyses, we show the fecal abundances of microbial
hpdBCA
and
tryptophanase
genes are independently associated with CVD. Our work,
thus, demonstrates that microbiota targeting therapies aimed at reducing levels of these
uremic solutes/toxins should be explored, and suggests potential microbial enzymatic
transformations as rational therapeutic targets for treatment of residual CVD risk.
Research Article
mBio
November/December 2023 Volume 14
Issue 6
10.1128/mbio.01331-23
2
Downloaded from https://journals.asm.org/journal/mbio on 21 December 2023 by 131.215.249.156.
RESULTS
Uremic solutes are associated with overall mortality in individuals with
preserved kidney function
Plasma samples from individuals undergoing elective diagnostic cardiac evaluation
with longitudinal follow-up (
n
= 1,149; Table S1) were analyzed using untargeted gas
chromatography
time-of-flight
mass spectrometry (GC-MS-TOF) as described under
Materials and Methods. A derivatized analyte consistent with
p
-cresol (
Fig. 1A
), the
microbiome-derived metabolite and precursor to
p
CS, was noted to be
significantly
higher among subjects who died over the ensuing period of follow-up (5 years)
compared to those that did not (
Fig. 1B
;
P
< 0.0001). Kaplan-Meier survival analyses
also suggested that levels of the candidate ion
identified
as
p
-cresol were associated
with incident mortality risk in this cohort (
Fig. 1C
). Subsequent Cox proportional hazards
regression with time-to-event analysis showed subjects with elevated (fourth quartile
[Q4]) levels had higher mortality risk compared to those with low
(first
quartile [Q1])
levels, even after adjustment for traditional risk factors (hazard ratio [HR] 95%
confidence
interval [CI] for incident [5 years] death risk of HR = 1.86 [1.15–3.03],
P
= 0.01) (
Fig. 1D
).
We were somewhat surprised by the
p
-cresol
finding
because our GC-TOF methods
employed were not optimized for detection of a volatile compound like
p
-cresol.
Untargeted analyses as performed are not quantitative, and given use of a sample drying
step where the analyte could be lost, we reasoned that the
p
-cresol detected using our
analytical method most probably was a product of
p
CS hydrolysis during sample
preparation (see Materials and Methods). Given the novelty of these
findings
(
p
CS
previously has been linked to CVD and mortality risks, but only among those with
significantly
impaired renal function like end-stage renal disease and hemodialysis), we
decided to validate our observations using quantitative methods in an independent
(non-overlapping) cohort (
n
= 3,954 subjects; Table S2) using a complementary techni­
que: targeted stable isotope-dilution LC-MS/MS (see Materials and Methods). We also
expanded our targeted LC-MS/MS analyses to also include IS, since both
p
CS and IS are
commonly referred to as “uremic toxins” and studied concurrently in clinical observatio­
nal studies (
17
,
24
–27
). As shown in
Fig. 2
, both
p
CS and IS each were
significantly
higher
in subjects who died over the course of 5 years (
P
< 0.001 each, Wilcoxon rank-sum test,
Fig. 2A
and D
). Kaplan-Meier survival analyses revealed that subjects with either high
p
CS
or IS levels had overall poorer survival over 5 years of follow-up (
Fig. 2B
and E
).
Specifi-
cally,
individuals with
p
CS and IS levels in the highest vs lowest quartile (Q4 vs Q1) for
each compound demonstrated a
significantly
increased risk of incident (5 years) death
(HR = 2.83 [2.19–3.64];
P
< 0.0001 and HR = 2.81 [2.16–3.65];
P
< 0.0001, respectively). The
associations between
p
CS and IS and death each remained
significant
after adjusting for
traditional risk factors, as well as following additional adjustments for renal function (HR
= 1.52 [1.16–2.00];
P
= 0.002 and HR = 1.68 [1.27–2.21];
P
= 0.0002 for
p
CS and IS,
respectively) (
Fig. 2C
and F
). Furthermore, the association between
p
CS or IS and risk of
incident (5 years) death holds true when subjects were divided into subgroups of
subjects with relatively preserved kidney function (estimated glomerular
filtration
rate
[eGFR]
60 mL/min/1.73 m
2
) and subjects with impairment in kidney function (eGFR <
60 mL/min/1.73 m
2
) (Table S3). In further sensitivity analyses, we observed the associa­
tions between
p
CS and IS and mortality risk remained
significant
in both males and
females alike, younger versus older, as well as within multiple
different
subgroups
including subjects with or without hyperlipidemia or hypertension (
Fig. 3
). Collectively,
these data demonstrate that
p
CS and IS, which are generally considered to be relevant
only in the setting of renal disease, are associated with mortality in the broader popula­
tion of individuals with predominantly preserved renal function, as well as in the absence
of traditional cardiovascular risk factors.
Research Article
mBio
November/December 2023 Volume 14
Issue 6
10.1128/mbio.01331-23
3
Downloaded from https://journals.asm.org/journal/mbio on 21 December 2023 by 131.215.249.156.
FIG
1
Untargeted metabolomics reveals that
p
-cresol is associated with overall mortality. (
A
) Comparison of electron
ionization spectra of the compound detected in plasma and an authentic standard of the trimethylsilyl derivative of
p
-cresol.
(
B
) Relative plasma levels of
p
-cresol in sequential stable subjects undergoing elective diagnostic cardiac evaluation. Subjects
(
n
= 1,149; discovery cohort) were divided into groups as indicated based on whether or not they died during the 5-year
follow-up. In the box-whisker plot, the upper and lower boundaries of the box represent the 25th and 75th percentiles,
the median is marked by a horizontal line inside the box, and the whiskers represent 10% and 90% measured values.
(
C
) Kaplan-Meier estimates of the risk of incident death by quartile of relative amounts of
p
-cresol from the untargeted
analysis. (
D
) Forest plots showing overall mortality within 5 years among test subjects according to the quartiles for the
relative level of
p
-cresol (black), or a multivariable Cox model for hazard ratio that includes adjustments for age, sex, current
smoking, high-density lipoprotein, low-density lipoprotein, triglyceride level, systolic blood pressure, diabetes mellitus, and
high-sensitivity C-reactive protein (adjusted, red). Symbols represent hazard ratios and the 95% CIs are indicated by the line
length.
Research Article
mBio
November/December 2023 Volume 14
Issue 6
10.1128/mbio.01331-23
4
Downloaded from https://journals.asm.org/journal/mbio on 21 December 2023 by 131.215.249.156.