Key words
urolithin A - anti-inflammatory - macrophages - M1 polarization - glycolysis
Introduction
Urolithin A (UroA) is one metabolite of ellagitannins produced by intestinal microbes
and proposed to be one of the active bioavailable principles of ellagitannins taken
up with fruits, berries, or nuts as part of a diet or herbal preparation. It is favorably
produced by a microbiome characterized by a higher Firmicutes to Bacteroidetes ratio
and a greater abundance of Clostridiales, Ruminococcaceae, and Akkermansia muciniphilia
[1]. Multiple strains possess the ability to metabolize ellagic acids to UroA, including
Bifidobacterium pseudocatenulatum INIA P815, Enterococcus faecium FUA027, and Lactococcus garvieae FUA009 (isolated from fecal samples) as well as Streptococcus thermophilus FUA329 (isolated from human milk) [2]. A recent study added Enterocloster bolteae CEBAS S4A9, E. bolteae DSM 29 485, E. bolteae DSM 15 670 T, Enterocloster
asparagiformis DSM 15 981 T, and Enterocloster citroniae DSM 19 261 T to the UroA producers (all strains were isolated from fecal samples)
[3]. UroA has been reported to confer several potential benefits, including a prolonged
health span [4], improved muscle function [5], or reduced inflammation. Concerning inflammation, UroA has been heavily investigated
in the context of colitis, where it could protect gut barrier function or improve
dysbiosis [6], [7], [8]. In vitro, UroA has been shown to reduce proinflammatory signaling in macrophages, osteoclasts,
neutrophils, and T cells [9], [10], [11]. Reported molecular modes of action of UroA include binding to cytochrome P450 (Cyp)
1A1, activation of the aryl
hydrocarbon receptor (AhR) or nuclear factor E2-related factor 2 (Nrf2), and inhibition
of nuclear factor (NF)-κB or mechanistic target of rapamycin (mTOR) signaling; the latter is closely linked
to induction of macroautophagy, in particular mitophagy, which is considered one prime
mechanism underlying the bioactivities of UroA [12], [13], [14], [15].
Macrophages are part of the innate immune system and have a variety of functions,
comprising control of infection and inflammation, phagocytic clearance of debris,
or building a bridge to the adaptive immune system by antigen processing and presentation
to T cells [16]. The distinct tasks demand that macrophages are highly plastic and can adopt distinctive
functional phenotypes (polarization states). They originally were divided into classically
activated proinflammatory M1 and the alternatively activated anti-inflammatory M2
type, which represent only the extremes of a continuum of possible distinct intermediates.
Proinflammatory macrophages are typically induced by interferon (IFN)-γ [referred to as M1(IFN)-γ, lipopolysaccharide (LPS) [M1(LPS)] or a combination of both [M1(IFN/LPS)] and express
typical M1 markers, such as interleukin (IL)1β, IL6, tumor necrosis factor (TNF)-α, or inducible nitric oxide (NO) synthase
(NOS2). M2(a) macrophages develop after stimulation with IL 4 [M2(IL4)] or 13 [M2(IL13)]
and are involved in tissue remodeling and wound repair, and express IL10, transforming
growth factor (TGF)-β arginase 1, mannose receptor (Mrc1; CD206), or C-type lectines (Mgl1/2) [17]. Notably, macrophage polarization coincides and is partly dependent on characteristic
shifts in cellular energy metabolism for optimal supply with ATP, electrons, building
blocks, and substrates for posttranslational modifications (immunometabolism). Murine
M1 macrophages usually show elevated activity of aerobic glycolysis, a broken tricarboxylic
acid (TCA) cycle, as well as use of mitochondria for reactive oxygen species (ROS)
rather than for ATP production. M2 macrophages, in contrast, tend to run coupled oxidative
phosphorylation (OXPHOS) with augmented fatty acid oxidation, an intact TCA cycle,
or an increased hexosamine pathway [18].
This study set out to investigate a potential immunometabolic facet of UroA, i.e.,
to address the question whether modulated bioenergetic programs in response to UroA
exposure may directly account for altered M1(LPS) macrophage polarization in vitro.
Results and Discussion
We first investigated macrophage polarization from naïve M0 to M1(LPS) or M2(IL4)
in the absence and presence of UroA. A resazurin conversion assay showed that concentrations
up to 30 µM of UroA were safe for macrophage viability (Fig. 1SA, Supporting Information), whereas staurosporine, a known proapoptotic compound [19], strongly diminished cell vitality. Typical M1 marker expression based on il1β, il6, and nos2 mRNA expression as well as NO and TNF-α release were reduced by 10 and 30 µM UroA in M1(LPS) macrophages, again in line with
published reports [20], [21] (Fig. 1SB-F, Supporting Information). Sulforaphane was used as a reference compound due to its
reported inhibition of M1(LPS) polarization [22]. IL4-triggered induction of arg1, mrc1, mgl1, and 2 mRNA and polyamine or TGF-β release were not
markedly affected by the presence of UroA (Fig. 2SA-E, Supporting Information), although UroA was reported to elicit a modest increase
in M2 marker expression when macrophages were exposed to both IL4 and IL13 [23]. Overall, this data partly confirms previous findings or shows that UroA can impede
M1(LPS), but not M2(IL4) polarization in macrophages (prevention setting). A potential
influence on M1 → M2 or M2 → M1 repolarization (potential therapy setting) was beyond
the scope of this study, but clearly deserves further investigation in the future.
Given that macrophage polarization is driven by distinct metabolic programs, we next
examined the influence of UroA on glycolytic (needed for M1 polarization) and respiratory
activity (favored during M2 polarization). As expected from previous reports [24] in cultured murine macrophages, M1(LPS) cells showed higher glycolytic activity
than M0, as evident in an extracellular flux assay using the Seahorse technology.
Despite the observed anti-inflammatory action of UroA in macrophages and the presumably
proinflammatory characteristic of glycolysis, the presence of UroA further enhanced
basal and tendentially also compensatory glycolytic activity in M1(LPS) cells ([Fig. 1]). Looking at mitochondrial metabolism, UroA could not prevent the drop of respiratory
activity assessed as oxygen consumption rates (OCR) in M1(LPS) cells, at least at
the investigated time point and UroA concentrations, although UroA has already been
reported to protect mitochondrial function [23]. Basal, spare, and coupled respiration were not significantly different to control
M1(LPS). UroA did not prevent either the mitochondrial superoxide production during
M1(LPS) polarization ([Fig. 2]), presumed to occur due to breaks in the TCA cycle, succinate accumulation, and
reverse electron transport at complex I in the mitochondrial membrane [25].
Fig. 1 M1(LPS/UroA) display higher glycolytic activity than M1(LPS) macrophages. IBMDM macrophages
were pretreated with DMSO or with the indicated concentrations of UroA for 30 min
before they were stimulated with LPS (25 ng/mL) for 16 h. Then, 100 000 cells were
subjected to a glycolytic rate assay and extracellular flux analysis as described
in detail in the Methods section. Panel a depicts an overview of glycolytic activity over time before and after the addition
of the OXPHOS inhibitors rotenone (ROT) and antimycin A (A.A) (= driving cells from
basal to compensatory glycolytic activity) or the glycolysis inhibitor DOG (= showing
glycolysis-independent acidification), as evident in the extracellular acidification
rate (ECAR) in mPH/min. b Basal glycolytic activity. c Compensatory glycolytic activity. Bar graphs derive from compiled data of three different
biological replicates given as the mean ± SD (*p < 0.05; ANOVA, followed by multiple
comparisons test).
Fig. 2 Both M1(LPS) and M1(LPS/UroA) show impaired OXPHOS activity and higher mitochondrial
superoxide production than M0 macrophages. IBMDM macrophages were pretreated with
DMSO or with the indicated concentrations of UroA for 30 min before they were stimulated
with LPS (25 ng/mL) for 16 h. Then, 100 000 cells were subjected to a mitochondrial
stress test and extracellular flux analysis as described in detail in the Methods
section. Panel a depicts OXPHOS activity, as evident in the oxygen consumption rate (OCR) in pmol/min,
over time before and after the addition of the ATP synthase inhibitor oligomycin (oligo)
(= allowing deduction of OCR coupled to ATP production), the uncoupler FCCP (= eliciting
maximal OCR), and the complex I/III inhibitors rotenone (rot) and antimycin A (A.A)
(= showing non-mitochondrial OCR). b Basal respiration, c spare respiratory capacity, and d OCR coupled to ATP synthesis. e Mitochondrial superoxide
generation was measured by MitoSOX staining and the fluorescent signal was quantified
by flow cytometry. Bar graphs derive from compiled data of three different biological
replicates given as the mean ± SD (*p < 0.05; ANOVA, followed by multiple comparisons
test).
Next, we were prompted to examine whether the observed increased glycolytic activity
may be implicated in the blunted M1(LPS) polarization by UroA, especially as the simplistic
view of elevated glycolysis as an essential M1 prerequisite and invariably proinflammatory
feature had already been questioned [22], [26], [27]. For this, the inhibitory effect of UroA on M1 marker expression (il1β, il6, and nos2 mRNA as well as NO release) was compared under normal and conditions of partially
inhibited glycolysis. Diminishing glycolytic activity by the pharmacological inhibitor
deoxyglucose (DOG) at a concentration of 2 mM (which achieved submaximal inhibition
but still allowed moderate M1 marker gene expression as confirmed in our previous
study [22]) dampened the relative inhibitory activity of UroA towards all selected readouts
([Fig. 3]). Therefore, the observed increased glycolytic activity may truly take part in the
UroA-mediated inhibition of M1(LPS) polarization. The increased glycolytic activity
was not associated with an elevated expression of enzymes, characteristic for proinflammatory
macrophages like hexokinase 2, pyruvate kinase M2, or lactate dehydrogenase on mRNA
(Fig. 3S, Supporting Information) or protein (not shown) level. However, this does not yet
allow a reliable deduction about possibly altered enzymatic activities due to posttranslational
or allosteric modification, which would require further examination. Notably, M1(LPS/UroA)
did not take up more glucose than M1(LPS) either (Fig. 3S, Supporting Information), suggesting that M1(LPS/Uro) preferentially fuels the incorporated
glucose into aerobic glycolysis rather than into mitochondrial respiration.
Fig. 3 Interference with glycolytic activity blunts the inhibition of M1(LPS) marker expression
by UroA. iBMDMs were left untreated or treated with DOG (2 mM) and exposed to Uro
A (10 and 30 µM) as indicated for 30 min prior to stimulation with LPS (25 ng/mL)
for 6 h. Then, mRNA expression [relative to the respective M1(LPS) control w/o UroA]
was assessed for (a) il1β, (b) il6, and (c) nos 2 by qPCR (using ppia as the reference gene). d Cells were left untreated or treated with DOG (2 mM) and exposed to Uro A (0 – 30 µM)
as indicated for 30 min prior to stimulation with LPS (25 ng/mL) for 24 h. Then, NO
release was quantified by the Griess assay and referred to the respective M1(LPS)
control without UroA treatment. Bar graphs depict compiled data from three biological
replicates (mean ± SD, n = 3, *p < 0.05, ANOVA, compared to respective solvent control).
The question of how to reconcile the observed elevated glycolysis and impaired mitochondrial
OXPHOS with the generally accepted pro-mitophagic, mitoprotective, and herewith connected
anti-inflammatory activity of UroA still remains [28], [29], [30]. The process of mitophagy involves mitochondrial fission, i.e., the segregation
of one big mitochondrium into smaller ones, engulfment by phagosomes, and temporary
higher glycolytic activity to replenish ATP during the phase of reduced mitochondrial
ATP production [31]. Thus, it may be conceivable that UroA – in the context of its pro-mitophagic activity
– leads to (transient) mitochondrial fission in macrophages, which entails reduced
respiration and increased glycolysis. Using a mitochondria-selective dye and confocal
microscopy, M1(LPS/UroA) macrophages indeed showed transient fission of mitochondria,
which occurred
much earlier (8 h) than in control M1(LPS) (at 24 h; as part of their full polarization
to M1 cells [32] ([Fig. 4]). To see whether this early fission accounts for glycolytic and anti-inflammatory
activity of UroA, we made use of mdivi, an accepted inhibitor of mitochondrial fission
[33], [34]. The presence of mdivi blunted LPS-triggered induction of investigated M1 markers
(based on published reports [35], [36]) but also diminished the extent of UroA-mediated inhibition of their expression
([Fig. 5 a], from approx. 80 – 100% down to ~ 20 – 30% inhibition only). Moreover, mdivi overcame
the increased glycolytic activity in M1(LPS/UroA) when compared to M1(LPS) macrophages
([Fig. 5 b]).
Fig. 4 UroA leads to comparably early and transient mitochondrial fission in M1(LPS) macrophages.
iBMDMs were left as is (M0) or treated with vehicle (0.1% DMSO) or UroA (30 µM) for
30 min prior to stimulation with LPS (25 ng/mL) for the indicated periods of time.
A MitoTracker Deep Red probe was used to detect changes in mitochondrial morphology
by maximum intensity projection (MIP) imaged by confocal microscopy. The scale bar
indicates 10 µm. Representative images are shown.
Fig. 5 Reduced M1 marker expression and boosted glycolytic activity in M1(LPS/UroA) is blunted
by mdivi, an inhibitor of mitochondrial fission. a Murine macrophages (RAW 264.7) were pretreated with DMSO or UroA (30 µM) for 30 min
before they were stimulated with LPS (25 ng/mL) in the presence and absence of mdivi
(25 µM) for 16 h. Then, mRNA expression (relative to the respective M0 control w/o
LPS) was assessed for (a) il1β, (b) il6, and (c) nos 2 by qPCR (using ppia as the reference gene). b Murine macrophages (RAW 264.7) were pretreated with DMSO or UroA (30 µM) for 30 min
before they were stimulated with LPS (25 ng/mL) in the presence and absence of mdivi
(25 µM) for 16 h. Then, 100 000 cells were subjected to a glycolytic rate assay and
extracellular flux analysis as described in detail in the Methods section. An overview
of glycolytic activity over time before (= basal glycolytic activity) and
after the addition of the OXPHOS inhibitors rotenone (ROT) and antimycin A (A.A) (= driving
cells from basal to compensatory glycolytic activity) or the glycolysis inhibitor
DOG (= showing glycolysis-independent acidification) is depicted. The graphs are representative
for two biological replicates [in technical triplicate or quadruplicate in case of
b] with consistent results.
Overall, this study shows that UroA reduces the expression of proinflammatory genes
and elicits higher glycolytic activity in M1(LPS) macrophages. The increased glycolytic
activity is needed for maximal inhibitory action of UroA and might be explained as
a compensatory response to mitochondrial fission, occurring as a prerequisite for
boosted mitophagy by UroA. These findings disclose a so far unappreciated facet in
the anti-inflammatory activity of UroA and support the notion that increased glycolytic
does not always need to result in a proinflammatory outcome but can also contribute
to anti-inflammatory signal relays. What still remains to be resolved in more detail,
though, is whether the increased glycolytic activity is only required for proper mitophagy,
which then mediates the anti-inflammatory action in UroA-treated cells [30], or whether it also directly modulates inflammatory signaling, e.g., by providing
metabolites as precursors for
epigenetic posttranslational modifications or lactate as an anti-inflammatory metabolite
and signal [37], [38]. Moreover, it needs to be noted that experiments were performed in cultivated murine
macrophages with LPS as the sole trigger for M1 polarization and on only few selected
time points. Therefore, further studies are needed, as, e.g., possibly altered behavior
of M1(IFN/LPS) or M1(IFN) macrophages, fuel competition and crosstalk between different
(immune) cells, dynamic cellular responses during inflammation, or differences between
human and murine macrophage biology were not considered [39], [40], [41], [42], [43]. In addition, for application of UroA in an in vivo setting, one should keep in mind that its metabolization to (less active) phase II
metabolites, such
as glucuronides or sulfates, may hamper bioactivity [9], and that in vivo achievable levels of UroA may well be below 10 or 30 µM, however, with likely local
peaks in the colon or sites of inflammation [44], [45].
Materials and Methods
Reagents, chemicals, and cells
Stimuli for macrophage polarization, i.e., LPS from Escherichia coli O55:B5 and mouse IL4, as well as 2-deoxy-D-glucose (2-DOG), sodium pyruvate, α-D-glucose, sulfanilamide, naphthylendiamine, and mdivi were obtained from Sigma-Aldrich.
UroA came from Tocris. DMSO served as the vehicle control and solvent for all stock
solutions and was obtained from Sigma. The DMSO concentration was even throughout
the different conditions of one experiment and never exceeded a final concentration
of 0.2%. Media and supplements for cell culture were purchased from Invitrogen, Lonza,
or Sigma. Immortalized bone marrow-derived macrophages (iBMDMs) were kindly provided
by Laszlo Nagy (Debrecen University, Hungary) and the RAW 264.7 cell line came from
ATCC.
Cultivation and treatment of murine macrophages
The detailed protocol was previously published in [22]. Briefly, iBMDMs were cultured in phenol red-free DMEM high glucose supplemented
with 10% filtered L-929 cell-conditioned medium containing macrophage colony-stimulating
factor (M-CSF), 10% heat-inactivated FBS, 2 mM L-glutamine, 100 IU/mL penicillin,
and 100 mg/mL streptomycin (iBMDM medium) at 37 °C with 5% CO2. RAW 264.7 were cultivated in DMEM high glucose supplemented with 10% heat-inactivated
FBS, 2 mM L-glutamine, 100 IU/mL penicillin, and 100 mg/mL streptomycin. Cells were
seeded into appropriate plates and left overnight prior to polarization and treatment
as indicated. Both cell lines gave consistent results in the assessed readouts.
Assessment of nitric oxide/nitrite
Cells in 96-well plates (3 × 104 cells per well in 200 µL) were treated as indicated. Then, an aliquot (100 µL) of
the cell culture supernatant was mixed with an even volume of Griess reagent (0.5%
sulfanilamide, 0.05% naphthylendiamine) before absorbance at 550 nm was assessed using
a microplate spectrophotometer (TECAN Sunrise Austria).
Ribonucleic acid extraction and real time quantitative polymerase chain reaction analysis
The detailed protocol has already been published in [22]. Briefly, total ribonucleic acid (RNA) from 0.5 – 1 million cells was extracted
using an RNA isolation kit (IST Innuscreen GmbH), quantified, and checked for purity
using a NanoDrop 2000 (Thermo Fisher Scientific). cDNA was synthesized from 1 µg of
RNA using a High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher), and then
subjected to quantitative polymerase chain reaction (qPCR) using the Luna Universal
qPCR Master Mix (New England Biolab) and the LightCycler 480 Real-Time PCR System
(Roche Diagnostics GmbH). Data analysis was performed using the (2–ΔΔCt) method. Primer sequences are listed in [Table 1].
Table 1 Primer sequences for qPCR experiments.
nos2
|
Fwd: CAGAGGACCCAGAGACAAGC Rev: TGCTGAAACATTTCCTGTGC
|
il6
|
Fwd: GAGGATACCACTCCCAACAGACC Rev: AAGTGCATCATCGTTGTTCATACA
|
il1β
|
Fwd: CAACCAACAAGTGATATTCTCCATG Rev: GATCCACACTCTCCAGCTGCA
|
ppia
|
Fwd: CCAAGACTGAATGGCTGGATG Rev: TGTCCACAGTCGGAAATGGTG
|
Extracellular flux analysis (Seahorse technology)
The detailed procedure can be found in [22]. In short, cells were treated, scraped, and plated in XF24e-cell culture plates
precoated with Corning Cell-Tak adhesive (Sigma-Aldrich) at a density of 1 × 105 cells/well. After 1 h at 37 °C in a non-CO2 incubator in XF assay medium (pH 7.4/37 °C) (Agilent Technologies) supplemented with
2 mM glutamine, 1 mM pyruvate and 25 mM glucose real-time extracellular acidification
(ECAR) and OCR rates were assessed in an XF24e Flux Analyzer (Seahorse Bioscience,
now Agilent Technologies). ECAR and OCR were monitored under basal conditions, as
well as after rotenone (Rot, 0.5 µM)/antimycin A (AA; 0.5 µM), FCCP (1 µM), and 2-DOG
injections (50 mM), respectively, using the proprietary glycolytic rate assay and
mitochondrial stress test protocols from Agilent. Data were analyzed using the Wave
software package and report generators were provided by Agilent Technologies.
Assessment of mitochondrial superoxide
Cells were seeded on 12-well plates (5 × 105 cells per well) and treated as indicated. The following day, cells were washed with
PBS and incubated with 5 µM MitoSOX (Thermo Fisher Scientific) in Hankʼs balanced
salt solution with calcium and magnesium (HBSS/Ca/Mg) (Gibco) for 10 min at 37 °C.
After washing with PBS, the red fluorescence intensity was detected and quantified
by flow cytometry [FACSCalibur (BD Biosciences)].
Confocal imaging of mitochondrial morphology
Cells were seeded at a density of 5 × 105 per well in 12-well plates containing coverslips and treated as indicated. Then,
they were washed with PBS and incubated with DMEM medium supplemented with 250 nM
MitoTracker Deep Red FM (Invitrogen) for 20 min at 37 °C according to the manufacturerʼs
instructions. The medium was removed, and cells were fixed with 4% MeOH-free paraformaldehyde
for 10 min at 37 °C. Following the fixation step, cells were rinsed three times in
PBS and mounted in Fluoromount Aqueous Mounting Medium (Sigma). MitoTracker Deep Red
fluorescence with excitation/emission maxima of 644/665 nm was detected using a confocal
microscope (Olympus), and images were collected and analyzed by Leica Application
Suite X (LAS X) image analysis software.
Statistics
Unless stated otherwise, at least three independent biological replicates were performed
per experiment. The bar graphs depict the mean ± SD (standard deviation). Groups were
compared via ANOVA and multiple comparisons test by using GraphPad Prism 9 software.
Differences were considered as significant if p < 0.05.
Contributorsʼ Statement
S. B. designed and performed the experiments, analyzed and interpreted data, prepared
figures, and contributed to writing and finalizing the manuscript. B. B. performed
and analyzed the experiments. E. H. H. conceived and supervised the study, provided
funding, performed experiments, and drafted the manuscript. All authors have read
and agreed to the submitted version of the manuscript.