Article

Metabolic rewiring promotes anti-

inflammatory effects of glucocorticoids

https://doi.org/10.1038/s41586-024-07282-7 Jean-Philippe Auger1,2, Max Zimmermann1,2, Maria Faas1,2, Ulrich Stifel3, David Chambers1,2,
Brenda Krishnacoumar4,5, R. Verena Taudte6,7, Charlotte Grund1,2, Gitta Erdmann8,
Received: 5 September 2022
Carina Scholtysek1,2, Stefan Uderhardt1,2,9, Oumaima Ben Brahim1,2,9, Mónica Pascual Maté1,2,
Accepted: 7 March 2024 Cornelia Stoll1,2, Martin Böttcher2,10, Katrin Palumbo-Zerr1,2, Matthew S. J. Mangan11,
Maria Dzamukova12, Markus Kieler13,14, Melanie Hofmann13,14, Stephan Blüml15,
Published online: xx xx xxxx
Gernot Schabbauer13,14, Dimitrios Mougiakakos2,10, Uwe Sonnewald16, Fabian Hartmann1,2,
Check for updates David Simon1,2,12, Arnd Kleyer1,2,12, Anika Grüneboom4, Susetta Finotto17, Eicke Latz11,18,
Jörg Hofmann16, Georg Schett1,2, Jan Tuckermann3 & Gerhard Krönke1,2,12,18 ✉

Glucocorticoids represent the mainstay of therapy for a broad spectrum of immune-

mediated inflammatory diseases. However, the molecular mechanisms underlying
their anti-inflammatory mode of action have remained incompletely understood1.
Here we show that the anti-inflammatory properties of glucocorticoids involve
reprogramming of the mitochondrial metabolism of macrophages, resulting in
increased and sustained production of the anti-inflammatory metabolite itaconate
and consequent inhibition of the inflammatory response. The glucocorticoid receptor
interacts with parts of the pyruvate dehydrogenase complex whereby glucocorticoids
provoke an increase in activity and enable an accelerated and paradoxical flux of
the tricarboxylic acid (TCA) cycle in otherwise pro-inflammatory macrophages.
This glucocorticoid-mediated rewiring of mitochondrial metabolism potentiates
TCA-cycle-dependent production of itaconate throughout the inflammatory
response, thereby interfering with the production of pro-inflammatory cytokines.
By contrast, artificial blocking of the TCA cycle or genetic deficiency in aconitate
decarboxylase 1, the rate-limiting enzyme of itaconate synthesis, interferes with the
anti-inflammatory effects of glucocorticoids and, accordingly, abrogates their
beneficial effects during a diverse range of preclinical models of immune-mediated
inflammatory diseases. Our findings provide important insights into the anti-
inflammatory properties of glucocorticoids and have substantial implications for
the design of new classes of anti-inflammatory drugs.

Glucocorticoids (GCs) represent the most potent class of anti-inflam- where the GR acts as ligand-activated transcription factor and actively
matory drugs and are an essential component of the therapy of vari- drives the expression of its target genes1,4. However, such a GR-mediated
ous immune-mediated inflammatory diseases such as rheumatoid transcriptional ‘transactivation’ does not account for the full spec-
arthritis, inflammatory bowel disease and asthma1. A major drawback trum of anti-inflammatory effects exerted by GCs in cells of the innate
of GC-based treatment regimens is their severe side effects, such as immune system. Here GCs are considered to also exert various, often
increased risk of infections, insulin resistance or osteoporosis1–3, incompletely defined, transrepressive effects that were reported to
a cir­cumstance that limits their long-term use and highlights the need involve GR-mediated tethering and blocking of other transcription
for a better understanding of their molecular mode of action. factors and signalling molecules, including different NF-κB-family
Binding of GCs to the cytoplasmic GC receptor (GR) results in its and AP-1-family members, resulting in suppression of the inflamma-
nuclear translocation and binding to specific responsive elements tory response1,5,6.
1
Department of Internal Medicine 3, University of Erlangen-Nuremberg and Universitätsklinikum Erlangen, Erlangen, Germany. 2Deutsches Zentrum für Immuntherapie (DZI), University of
Erlangen-Nuremberg and Universitätsklinikum Erlangen, Erlangen, Germany. 3Institute of Comparative Molecular Endocrinology (CME), Ulm University, Ulm, Germany. 4Leibniz-Institut für
Analytische Wissenschaften, ISAS, e.V, Dortmund, Germany. 5Institute of Physiology, University Hospital Essen, University of Duisburg-Essen, Essen, Germany. 6Institute of Experimental and
Clinical Pharmacology and Toxicology, University of Erlangen-Nuremberg, Erlangen, Germany. 7Institute of Laboratory Medicine and Pathobiochemistry, Molecular Diagnostics, Philipps
University Marburg, Marburg, Germany. 8Division of the Molecular Biology of the Cell I, German Cancer Research Centre (DKFZ), Heidelberg, Germany. 9Optical Imaging Competence Centre
(FAU OICE), Exploratory Research Unit, University of Erlangen-Nuremberg, Erlangen, Germany. 10Department of Hematology and Oncology, Otto-von-Guericke University Magdeburg,
Magdeburg, Germany. 11Institute of Innate Immunity, Medical Faculty, University of Bonn, Bonn, Germany. 12Department of Rheumatology and Clinical Immunology, Charité - Universitätsmedizin
Berlin, Berlin, Germany. 13Institute for Vascular Biology, Centre for Physiology and Pharmacology, Medical University Vienna, Vienna, Austria. 14Christian Doppler Laboratory for Arginine
Metabolism in Rheumatoid Arthritis and Multiple Sclerosis, Vienna, Austria. 15Division of Rheumatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria.
Division of Biochemistry, Department of Biology, University of Erlangen-Nuremberg, Erlangen, Germany. 17Department of Molecular Pneumology, University of Erlangen-Nuremberg and
16

Universitätsklinikum Erlangen, Erlangen, Germany. 18Deutsches Rheuma-Forschungszentrum Berlin, Berlin, Germany. ✉e-mail: [email protected]

Nature | www.nature.com | 1
Article
a b GC + LPS versus LPS, 24h c
4h 24 h 300 Oligo FCCP R/A
Suppressed by Induced by

OCR (pmol min–1)

Control
GC treatment GC treatment
Control 200
GC Il1b GC
300
LPS Cd163 LPS
GC + LPS Ccr5 Gpx3 100
GC + LPS
0
200 0 20 40 60 80
Ptx3
Time (min)

–log10[P]
z score 200 Glucose Oligo 2-DG

ECAR (mpH min–1)

2 Control
100 150
1 GC
100
0 Nos2 LPS
–1 50 GC + LPS
–2 0
0
–10 –5 0 5 10 0 20 40 60 80
log2[FC] Time (min)
d P = 0.002 P = 0.0001 e
80 Control 100
100

UK5099
per normalized unit)

per normalized unit)

Oxidat. Energ. Oxidat. Energ.

Vehicle

Vehicle
OCR (pmol per min

OCR (pmol per min

BPTES
MTRed+ cells of

GC
total cells (%)

60 80 80
LPS
GC + LPS 60 60
40 Control Control
40 GC 40 GC
20 LPS 20 LPS
20
Quiesc. Glycol. GC + LPS Quiesc. Glycol. GC + LPS
0 0 0
6h 24 h 0 20 40 60 80 0 20 40 60 80
ECAR (mpH per min ECAR (mpH per min
per normalized unit) per normalized unit)
f U-13C-glucose Control
GC
Relative ER (%)

80 Pyruvate LPS
60 GC + LPS
PDH

Relative ER (%)
40 Lactate 30
20 CO2
0 Acetyl-CoA 20
10
0

Relative ER (%)
α-Oxaloacetate Citrate 15
10
Itaconate
Relative ER (%)

5
8 CO2
6 0
4 Malate α-ketoglutarate
Relative ER (%)

2 10
0
5

0
Fumarate Succinate
Relative ER (%)
Relative ER (%)

6 10
4
5
2
0 0

g P = 0.001 P = 0.004
Acetyl-CoA (nmol g–1)

2,500 P = 0.004 200 P = 0.001 Control

Lactate (nmol g–1)

2,000 GC
150
LPS
1,500 GC + LPS
100
1,000
500 50

0 0
6h 24 h 6h 24 h

Fig. 1 | GCs promote mitochondrial reprogramming of inflammatory OCR and ECAR of BMDMs treated with vehicle, GC and LPS for 24 h in the
macrophages. a,b, Bulk mRNA-seq data for BMDMs treated with vehicle presence of vehicle, BPTES or UK5099 (10 µM each). Oxidat., oxidative; energ.,
(Control), dexamethasone (GC, 100 nM) and LPS (100 ng ml−1) for 4 or 24 h. energetic; quiesc., quiescent; glycol., glycolytic. f, Mass spectrometry (MS)-
Data are presented as a heat map illustrating differential gene expression (a) based analysis of U-[13C]glucose carbon tracing in BMDMs treated with vehicle,
and a volcano plot in which each dot represents a gene with adjusted GC and LPS. The graphs indicate the relative U-13C exchange rate (relative ER)
P < 0.05 (b). c, OCR data from a mitochondrial stress test and ECAR data from for highlighted metabolites. g, Measurement of acetyl-CoA and of lactate
a glycolysis stress test of BMDMs treated with vehicle, GC and LPS for 24 h. levels in BMDMs treated with vehicle, GC and LPS for 6 or 24 h. n = 4. Data are
2-DG, 2-deoxyglucose; FCCP, phenylhydrazone; oligo, oligomycin; R/A, mean + s.e.m. Statistical analysis was performed using Wald tests with
rotenone + antimycin A. d, Quantification of MTRed+ cells of BMDMs treated Bonferroni’s correction (a) and one-way analysis of variance (ANOVA) with
with vehicle, GC and LPS for 6 or 24 h. n = 6. e, Mito Fuel Flex test showing the Tukey’s multiple-comparison test (d and g).

genes (Fig. 1a,b). An in silico prediction of affected transcription factors

GCs fuel the TCA cycle of macrophages indicated that GC treatment resulted in a (direct or indirect) modulation
In an effort to revisit the anti-inflammatory mode of action of GCs in the activity of an extended spectrum of LPS-induced transcription
in cells of the innate immune system, we initially performed RNA- factors including different NF-κB, AP-1 and STAT family members, as
sequencing (RNA-seq) analysis of lipopolysaccharide (LPS)-treated well as of HIF-1α (Extended Data Fig. 1a). GCs only partially blocked the
bone-marrow-derived macrophages (BMDMs) in the absence and pres- initial transcriptional response to LPS, but foremost interfered with the
ence of GCs. The resulting datasets showed that GCs did not globally sustained activity of the transcription factors. Pathway enrichment
inhibit inflammatory gene expression, but exerted a differential and analyses also indicated that GCs modulated not only inflammatory
time-dependent effect on LPS-induced mRNA transcription. GCs sup- pathways, but also metabolic pathways in LPS-activated macrophages,
pressed the expression of distinct modules of pro-inflammatory genes which included mitochondrial oxidative phosphorylation, pyruvate
including Il1b (encoding interleukin-1β) and Nos2 (encoding inducible metabolism and TCA cycle activity (Extended Data Fig. 1b).
nitric oxide synthase), whereas we observed no GC-induced changes Recent data have highlighted the mutual influence of cellular meta­
or even an increase in the expression of other sets of LPS-responsive bolism and the inflammatory response7. GCs act as potent regulators of

2 | Nature | www.nature.com
systemic energy and glucose homeostasis in liver, muscle and adipose result from non-genomic actions of GCs9–11. Although the underlying
tissue8, but little is known about the GC-mediated regulation of the cel- molecular events are incompletely understood, a proteomic analysis
lular metabolism of different immune cell types or about a potential of GR-interacting proteins in hepatocarcinoma cells recently identi-
role of GC-induced immune–metabolic cross-talk. We subsequently fied parts of the PDH complex as direct binding partners of the GR12.
performed an in-depth characterization of the metabolic effects of GCs Co-immunoprecipitation as well as proximity ligation assays con-
in macrophages. Extracellular metabolic flux analyses showed that GC firmed an interaction between the GR and PDH in both resting and LPS-
treatment reversed the LPS-induced block of mitochondrial respira- activated macrophages (Fig. 2d,e). However, this interaction occurred
tion and simultaneously interfered with the process of LPS-induced primarily within the cytoplasm, and we did not detect an interaction
aerobic glycolysis as indicated by a GC-induced increase in oxygen- within the nucleus or the presence of the GR within the mitochondria
consumption rate (OCR) and a decrease in the extracellular acidifi­ (Fig. 2e and Extended Data Fig. 2c). Notably, we observed that GC
cation rate (ECAR), respectively (Fig. 1c and Extended Data Fig. 1c). treatment, which triggered GR nuclear translocation, also resulted in
A flow-cytometry-based measurement of the mitochondrial mem­ a simultaneous decrease in the interaction between the GR and PDH
brane potential using MitoTracker Red (MTRed) staining confirmed within the cytoplasm (Fig. 2e and Extended Data Fig. 2c). In turn, this
that these metabolic changes were paralleled by a GC-induced stabi- decrease in interaction was paralleled by a GC-induced translocation
lization of the membrane potential and an increase in the number of of the regulatory PDH component X (PDHX) subunit E3 binding protein
functional MTRed+ mitochondria that otherwise decreased in response (E3BP) into the mitochondria (Fig. 2f,g and Extended Data Fig. 2d) as
to LPS (Fig. 1d and Extended Data Fig. 1d). well as by a simultaneous GC-induced increase in PDH and PC activity
GC treatment of LPS-activated macrophages accordingly promoted in LPS-activated macrophages (Fig. 2h). GC treatment also resulted
a highly energetic phenotype. UK5099-mediated pharmacological in an increased mitochondrial volume of LPS-activated macrophages
inhibition of the mitochondrial import of pyruvate, but not BPTES- (Fig. 2i). However, total PDH protein levels remained stable in these
mediated inhibition of glutaminase, abrogated these GC-induced cells (Extended Data Fig. 2e).
effects on macrophage metabolism, suggesting that the observed Our data therefore indicated that GC treatment of LPS-activated mac-
GC-induced bioenergetic switch specifically required the import of rophages provoked a global increase in mitochondrial pyruvate con-
glucose-derived pyruvate into the mitochondria (Fig. 1e). sumption by increasing the enzymatic activity of mitochondrial PDH
To further narrow down the underlying molecular mechanisms, and PC. This resulted in a GC-mediated rescue and paradoxical increase
we performed carbon-tracing experiments using uniformly labelled of TCA cycle activity in otherwise pro-inflammatory macrophages. Such
(U)-[13C]glucose. These experiments confirmed that GCs substantially a GC-induced reprogramming of mitochondrial metabolism neither
redirected the glucose metabolism of pro-inflammatory macrophages. relied on genomic effects of GCs nor involved mitochondrial transloca-
In accordance with the onset of aerobic glycolysis in pro-inflammatory tion of the GR. Instead, we observed a GC-induced release of regulatory
macrophages, we detected an increased exchange rate of glucose- PDH subunits from preformed cytoplasmatic PDH–GR complexes into
derived carbon atoms within the pool of lactate in response to LPS the mitochondrial compartment.
(Fig. 1f and Extended Data Fig. 2a). By contrast, GCs reduced this glu-
cose-dependent lactate generation in LPS-activated macrophages and
shifted the use of glucose-derived carbon atoms into the mitochon- GC activity requires TCA cycle integrity
drial TCA cycle, in which we observed a substantial increase in carbon Next, we examined whether this GC-induced metabolic reprogram-
exchange rates within TCA cycle metabolites such as citrate, fumarate ming related to the anti-inflammatory effects of GCs in LPS-activated
and malate as well as within the TCA-cycle-derived metabolite itaconate macrophages. To globally interfere with the GC-mediated increase
(Fig. 1f and Extended Data Fig. 2a). Again, these findings therefore in mitochondrial respiration and TCA cycle activity, respectively, we
pointed towards a GC-mediated regulation of pyruvate metabolism initially exposed macrophages to a hypoxic environment and studied
in LPS-activated macrophages whereby GCs promoted the influx of the anti-inflammatory effects of GCs under 1% O2. Hypoxia-induced
pyruvate into the mitochondrial TCA cycle. By contrast, cytosolic blocking of TCA cycle activity indeed abrogated the GC-mediated inhi-
pyruvate reduction to lactate decreased in response to GCs. Direct bition of the LPS-induced production of pro-inflammatory cytokines
quantification of the two pyruvate products—lactate and acetyl-CoA, such as IL-1β, IL-6 and TNF in BMDMs (Fig. 3a). A hypoxic environment
the latter representing a direct product of pyruvate oxidation within also diminished the GC-induced suppression of pro-inflammatory
the mitochondria—confirmed a GC-induced decrease in lactate and a cytokines in human monocyte-derived macrophages, indicating that
parallel increase in acetyl-CoA in LPS-activated macrophages (Fig. 1g). the anti-inflammatory properties of GCs in both mouse and human
macrophages rely on a GC-induced regulation of mitochondrial activity
(Extended Data Fig. 3a). We also performed experiments in which we
GCs promote non-genomic reprogramming added the prolyl-hydroxylase inhibitor roxadustat (RXD) to GC- and
As expected, mutation of the GR nuclear-localization sequence LPS-treated macrophages, thereby blocking the prolyl-hydroxylase-
(NLS) abrogated GC-induced transcriptional effects, including the mediated degradation of HIF-1α (Fig. 3b) as well as the GC-mediated
GC-induced expression of the GR target genes C1qb, Cd163 and Fkbp5 rescue of mitochondrial respiration in LPS-activated macrophages
in fetal liver-derived macrophages (FLDMs) from Nlsmut embryos (Fig. 3c,d and Extended Data Fig. 3b). Both the forced HIF-1α-mediated
(Fig. 2a). By contrast, the observed GC-induced rescue of the mito- block of the TCA cycle as well as a UK5099-mediated inhibition of the
chondrial membrane potential and mitochondrial respiration did not mitochondrial pyruvate import resulted in significantly reduced anti-
require the GR NLS, indicating that the GC-induced metabolic rewiring inflammatory effects of GCs in LPS-activated macrophages (Fig. 3e
of LPS-activated macrophages relied on non-genomic events (Fig. 2b,c and Extended Data Fig. 3c). Collectively, these data confirmed that a
and Extended Data Fig. 2b). GC-induced increase in mitochondrial pyruvate metabolism and con-
In hepatocytes, GC treatment was previously shown to rapidly trigger sequent rescue of TCA cycle activity in pro-inflammatory macrophages
mitochondrial pyruvate uptake as well as an increase in the enzymatic were essentially involved in the GC-induced anti-inflammatory effects.
activity of pyruvate dehydrogenase (PDH) and pyruvate carboxylase
(PC), which represent the two enzymes that catalyse the mitochondrial
conversion of pyruvate to acetyl-CoA and oxalacetate, respectively9–11. Itaconate mediates the effects of GCs
This GC-induced increase in hepatic PDH and PC activity essentially TCA cycle metabolites such as α-ketoglutarate, fumarate and succinate
promotes both gluconeogenesis and TCA cycle activity and seems to can act as potent regulators of eukaryotic gene expression and the

Nature | www.nature.com | 3
Article
a b

Normalized relative Fkbp5

Normalized relative Cd163
Normalized relative C1qb
P = 0.018
5 P = 0.0001 2,000 P = 0.0001 30 P = 0.0001 Control 60 P = 0.014 Control

mRNA expression
mRNA expression
mRNA expression

MTRed+ cells of
GC GC

total cells (%)

4 1,500 LPS
20 LPS 40
3 GC + LPS GC + LPS
1,000
2 10 20
1 500
0 0 0 0
Nls+/+ Nlsmut Nls+/+ Nlsmut Nls+/+ Nlsmut Nls+/+ Nlsmut

c Basal respiration Maximum respiration

250 Oligo FCCP R/A Nls+/+ Nlsmut 150 P = 0.0001 P = 0.0001 300 P = 0.001 P = 0.006 Control

OCR (pmol min–1)

OCR (pmol min–1)

OCR (pmol min–1)

200 Control 120 250 GC
200 LPS
150 90 GC + LPS
GC 150
100 60
LPS 100
50 30 50
GC + LPS
0 0 0
0 20 40 60 80 Nls+/+ Nlsmut Nls+/+ Nlsmut
Time (min)

d e P = 0.020

PLA signal per nucleus

50
IP P = 0.001
40 PLA DAPI PLA DAPI PLA DAPI PLA DAPI PLA DAPI
R

30
G
t

ti-
pu

WB
An
Ig
In

kDa 20

~100 GR 10
0
No Ab Control GC LPS GC + LPS
~55 C Ab

l
C
C PS

S
tro Anti-GR and Anti-GR and Anti-GR and Anti-GR and

LP
G
G L
on
o

anti-PDH anti-PDH anti-PDH anti-PDH

N

+
PDH
Anti-GR and anti-PDH
f 24 h
Control

GC
GC + LPS
LPS

PDHX Phalloidin PDHX/MitoTracker PDHX Phalloidin PDHX/MitoTracker

Merge merge Merge merge
MTRed DAPI PDHX/MitoTracker MTRed DAPI PDHX/MitoTracker
surface rendering surface rendering

g Control
GC
LPS
GC + LPS
h Control
GC
LPS
GC + LPS
i
Mitochondrial volume

120 P = 0.020 1.0 P = 0.001 2.5 P = 0.0058P = 0.001 300 P = 0.001

PDHX/MitoTracker

100
(U per 104 cells)

per cell (μm3)

0.8 2.0
PDH activity
overlap (%)

PC activity
(mU ml–1)

80 200
0.6 1.5
60
40 0.4 1.0
100
20 0.2 0.5
0 0 0 0
6h 24 h 6h 24 h 6h 24 h 6h 24 h

Fig. 2 | GCs control pyruvate metabolism through non-genomic mechanisms. vehicle, GC and LPS for 24 h. Fluorescence signal rendering for PDHX and
a, Quantification of C1qb, Cd163 and Fkbp5 mRNA expression in FLDMs from MitoTracker is shown at the bottom right. Scale bars, 10 μm. g, Quantification
Nls +/+ and Nls mut mice treated with vehicle, dexamethasone (GC, 100 nM) and of f. n = 5. h, Data from a PDH enzymatic activity assay (left) and PC enzymatic
LPS (100 ng ml−1) for 24 h. n = 5. b, Quantification of MTRed+ FLDMs from Nls+/+ assay (right) of lysates from BMDMs treated with vehicle, GC and LPS for 6
and Nls mut mice treated with vehicle, GC and LPS for 24 h. n = 4. c, The OCR of a or 24 h. n = 6 (PDH) and n = 4 (PC). i, Quantification of the mitochondrial
mitochondrial stress test of FLDMs from Nls+/+ and Nls mut mice treated with volume/cell of BMDMs treated with vehicle, GC and LPS for 6 or 24 h and stained
vehicle, GC and LPS for 24 h. n = 8. d, Co-immunoprecipitation of GR in BMDMs with MTRed, as determined by immunofluorescence microscopy. n = 8
with pyruvate dehydrogenase (PDH). WB, western blot. e, Proximity ligation (control), n = 5 (GC), n = 8 (LPS) and n = 7 (GC + LPS). Representative images are
assay in BMDMs treated with vehicle, GC and LPS for 24 h with antibodies (Ab) shown in Fig. 2f and Extended Data Fig. 2d. Data are mean + s.e.m. Statistical
against GR and PDH. Interactions (red) were quantified relative to nuclei. n = 5. analysis w.as performed using one-way ANOVA with Tukey’s multiple-comparison
Scale bars, 10 μm. f, The subcellular location of PDHX in BMDMs treated with test (a–c, e and g–i).

inflammatory response13,14. We therefore performed metabolomic an increase in intracellular and extracellular itaconate in LPS-treated
profiling to identify and quantify TCA-cycle-derived metabolites that macrophages (Fig. 4a–c).
would potentially modulate inflammatory gene expression in response The mitochondrial metabolite itaconate derives from the enzy-
to the observed GC-induced mitochondrial rewiring of LPS-activated matic conversion of the TCA cycle intermediate aconitate and can
macrophages. These analyses confirmed that the GC-mediated acceler- act as a potent anti-inflammatory molecule that interferes with the
ation of TCA cycle activity increasingly altered the metabolic profile of transcription of pro-inflammatory genes15. The generation of itaco-
LPS-activated macrophages. We observed that GC treatment reversed nate is catalysed by the enzyme aconitate decarboxylase (ACOD1),
the LPS-induced accumulation of multiple TCA cycle metabolites such which is encoded by Acod1, representing one of the most prominent
as citrate, malate or fumarate that otherwise piled up after the onset LPS-induced genes in macrophages16. We confirmed that LPS treat-
of aerobic glycolysis in pro-inflammatory macrophages 24 h after ment resulted in strong and sustained Acod1 mRNA and ACOD1 protein
LPS stimulation (Fig. 4a,b). One of the exceptions was a delayed, but expression as well as an early LPS-induced peak in itaconate levels at
substantial, GC-induced increase in the levels of succinate as well as 6 h (Fig. 4c–e). However, we also observed a substantial decrease in

4 | Nature | www.nature.com
 a P = 0.0001 P = 0.001 b
3.0 30 P = 0.009 6 P = 0.001
P = 0.001 Control

on S

S
2.5 P = 0.0001 25 5 GC

LP

LP
l

l
TNF (ng ml–1)

tro

tro
IL-1β (ng ml–1)

+
LPS

IL-6 (ng ml–1)

kDa

on

S
2.0 20

C
4

LP

LP
G

G
C

C
GC + LPS ~130
1.5 15 3 HIF1A
1.0 10 2 ~55
ACTB
0.5 5 1
0 0 0 Veh. RXD
GC – + – + – + – + GC – + – + – + – + GC – + – + – + – +
LPS – – + + – – + + LPS – – + + – – + + LPS – – + + – – + +
Normoxia 1% O2 Normoxia 1% O2 Normoxia 1% O2

c 150 Oligomycin FCCP

d 150
R/A Veh. RXD Glucose Oligo 2-DG Veh. RXD

ECAR (mpH min–1)

OCR (mpH min–1)

Control Control
100 100
GC GC

50 LPS 50 LPS

GC + LPS GC + LPS
0 0
0 20 40 60 80 0 20 40 60 80
Time (h) Time (h)

e P = 0.003 P = 0.003
P = 0.009
3.0 P = 0.009 15 P = 0.002
35 Control
30 P = 0.005
2.5 12 GC
IL-1β (ng ml–1)

25
TNF (ng ml–1)

LPS
IL-6 (ng ml–1)

2.0
20 9 GC + LPS
1.5
15 6
1.0
10
0.5 3
5
0 0 0
GC – + – + – + – + GC – + – + – + – + GC – + – + – + – +
LPS – – + + – – + + LPS – – + + – – + + LPS – – + + – – + +
Veh. RXD Veh. RXD Veh. RXD

Fig. 3 | Anti-inflammatory effects of GCs require TCA cycle integrity. glycolysis stress test (d) of BMDMs treated with vehicle, GC and LPS for 24 h in
a, Enzyme-linked immunosorbent assay (ELISA)-based quantification of the the presence of vehicle or RXD. e, ELISA-based quantification of the indicated
indicated cytokines in the supernatants of BMDMs under normoxia and cytokines in the supernatants of BMDMs treated with vehicle, GC and LPS
hypoxia (1% O2) and treated with vehicle, dexamethasone (GC, 100 nM) and LPS in the presence of vehicle or RXD. n = 6. Data are mean + s.e.m. Statistical
(100 ng ml−1). n = 6. b, Western blot analysis of HIF-1α protein levels in BMDMs analysis was performed using one-way ANOVA with Tukey’s multiple-comparison
after stimulation with vehicle, GC and LPS in the presence of a vehicle or RXD test (a and e).
(10 µM). c,d, OCR data of a mitochondrial stress test (c) and ECAR data of a

itaconate levels 24 h after LPS stimulation, coinciding with the delayed cytokine expression irrespective of ACOD1 expression (Extended Data
LPS-induced block of TCA cycle activity and the onset of aerobic glyco- Fig. 4a,b), strongly indicating that itaconate served as a central TCA-
lysis in LPS-activated macrophages (Fig. 4c). GCs, in turn, did not induce cycle-derived metabolite that mediated the anti-inflammatory effects of
ACOD1 expression nor did they alter the LPS-induced expression of this GCs in response to the observed TCA cycle acceleration in LPS-activated
enzyme, which was stably maintained over 24 h (Fig. 4d,e). We accord- macrophages. An RNA-seq analysis of BMDMs isolated from Acod1+/+
ingly observed that GC treatment did not directly increase itaconate, and Acod1−/− mice showed that ACOD1 expression was required for the
but substantially amplified itaconate generation in response to LPS at GC-mediated suppression of a large set of LPS-induced genes (Fig. 4h).
24 h (Fig. 4c). This observation was consistent with the GC-mediated Among others, this included the GC-mediated suppression of Il1b and
increase in TCA cycle activity in LPS-activated macrophages, which (in Nos2 mRNA expression (Fig. 4i). Moreover, the in silico prediction of
contrast to resting macrophages) displayed high ACOD1 expression affected transcription factors indicated that the direct and/or indi-
levels. These data suggested that GCs boosted itaconate production rect suppressive capacity of GCs on a broad spectrum of LPS-induced
not by affecting ACOD1 expression, but by sustaining TCA-cycle- transcriptional events was linked to ACOD1 expression (Extended Data
derived provision of the ACOD1 substrate aconitate in LPS-activated Fig. 4c). ERG344-mediated inhibition of ACOD1 enzymatic activity in
macrophages. Exposure to a hypoxic environment or RXD treat- human monocyte-derived macrophages and CRISPR–Cas9-mediated
ment accordingly abrogated the GC-induced increase in itaconate in disruption of ACOD1 in human THP-1 cells showed that ACOD1 was also
LPS-activated macrophages (Fig. 4f). involved in the GC-mediated suppression of LPS-induced cytokines
On the basis of the reported anti-inflammatory potential of itaconate, in human cells (Extended Data Fig. 4d,e). However, GCs retained a
we subsequently measured the cytokine expression in macrophages certain anti-inflammatory potential, indicating the involvement of
derived from wild-type and Acod1−/− mice to determine a potential role of additional metabolites that mediate the anti-inflammatory effects of
itaconate as a mediator of the anti-inflammatory effects of GCs. Acod1−/− GCs in response to the GC-induced reprogramming of the TCA cycle
macrophages produced regular levels of cytokines such as IL-1β, IL-6 in the human setting. The anti-inflammatory effects of itaconate were
and TNF in response to LPS (Fig. 4g). However, we observed that GCs previously shown to involve activation of the transcription factors NRF2
displayed a significantly reduced potential to block the production of and ATF317–19. GC treatment of LPS-activated macrophages accord-
these pro-inflammatory cytokines in the absence of ACOD1 (Fig. 4g). ingly resulted in an increased accumulation of NRF2, whereas genetic
The addition of the itaconate mimetics dimethyl itaconate or 4-octyl deletion of NRF in Nfe2l2−/− BMDMs significantly reduced the anti-
itaconate, in turn, was able to mimic the effects of GCs and blocked inflammatory potential of GCs (Extended Data Fig. 4f–h).

Nature | www.nature.com | 5
Article
a GC + LPS versus LPS, 6 h b
–4 Citrate P = 0.043 P = 0.001
100 4 0.8 P = 0.001

Isocitrate (nmol g–1)

P = 0.007 P = 0.001

Citrate (nmol g–1)

α-Ketoglutarate
–3 Ctrl
80

–log10[P]
3 0.6

(nmol g–1)
Isocitrate GC
–2 60
0.4 LPS
2
–1 40 GC + LPS
1 0.2
0 20
–5 0 5 0 0 0
log2[FC] 6h 24 h 6h 24 h 6h 24 h

GC + LPS versus LPS, 24 h

80 5 P = 0.001 60 P = 0.001

Succinate (nmol g–1)

Fumarate (nmol g–1)

–6 P = 0.001

Malate (nmol g–1)

–5 Fum. Itaconate 4
60
–4 40
–log10[P]

Isocit. 3
–3 G6P 40
Cit. 2
–2 Lact. Succinate 20
20 1
–1
0 0 0 0
–5 0 5 6h 24 h 6h 24 h 6h 24 h
log2[FC]
c d
P = 0.026 P = 0.001

Normalized relative Acod1

50 P = 0.001 Extracellular itaconate P = 0.001 400 Ctrl
800
Intracellular itaconate

mRNA expression
40 Ctrl 300 GC
600 GC
(nmol g–1)

(nmol g–1)

30 LPS 200 LPS

400 GC + LPS
20 GC + LPS
200 100
10
0 0 0
GC – + – + –+ –+ GC – + – + 0 12 24
LPS – – + + – – ++ LPS – – + + Time (h)
6h 24 h
e f 800 P = 0.001 300 P = 0.014
Ctrl

Itaconate (nmol g–1)

Itaconate (nmol g–1)

S
GC
C LPS

LP
LP

600 LPS
+

200
+

kDa
S

trl
C

C
trl

trl
C

GC + LPS
LP
LP

LP

G
C
G

G
C

~55 400
ACOD1 100
200
~55 ACTB
0 0
GC – + – + – + – + GC – + – + –+ –+
6h 12 h 24 h LPS – – + + – – + + LPS – – + + – – ++
Normoxia 1% O2 Veh. RXD
g P = 0.0001 P = 0.0001
P = 0.0001 15 P = 0.0001 10 P = 0.0001
2.0
P = 0.0001 Ctrl
8
IL-1β (ng ml–1)

TNF (ng ml–1)

IL-6 (ng ml–1)

1.5 GC
10 6 LPS
1.0 GC + LPS
4
5
0.5 2
0 0 0
GC – + – + – + – + GC – + – + – + – + GC – + – + – + – +
LPS – – + + – – + + LPS – – + + – – + + LPS – – + + – – + +
Acod1+/+ Acod1–/– Acod1+/+ Acod1–/– Acod1+/+ Acod1–/–

Acod1–/– versus Acod1+/+, 24 h

h i Decreased Increased in
Acod1+/+ Acod1–/–
in Acod1–/– Acod1–/–
Ctrl
GC 300
Nos2 Il1b
LPS
GC + LPS
200
–log10[P]

z score 100 Ccr5

4
2 Ptx3
0 0
–2 –5 0 5 10 15
–4 log2[FC]

Fig. 4 | Amplification of itaconate production mediates the anti- quantification of the indicated cytokines in the supernatants of BMDMs from
inflammatory effects of GCs in vitro. a–c, MS-based quantification in cell Acod1+/+ and Acod1−/− mice treated with vehicle, GC and LPS. n = 6. h,i, Bulk
lysates (a–c) or supernatants (c) of BMDMs treated with vehicle, dexamethasone mRNA-seq data of BMDMs from Acod1+/+ and Acod1−/− mice treated with vehicle,
(GC, 100 nM) and LPS (100 ng ml−1) for 6 and 24 h. n = 4. a, Volcano plot of the GC and LPS for 24 h. Data are presented as a heat map illustrating differential
different intracellular metabolites in GC- and LPS-treated versus LPS-treated gene expression (h) or a volcano plot in which each dot represents a gene with
BMDMs. Cit., citrate; G6P, glucose-6-phosphate; isocit., isocitrate; lact., lactate. adjusted P < 0.05 (i). The volcano plot shows differentially expressed genes
b, Quantification of different intracellular TCA metabolites. c, Quantification in Acod1−/− versus Acod1+/+ BMDMs normalized for the treatment effect of GC
of intracellular and extracellular itaconate. d,e, RT–qPCR-based (d) and and LPS versus LPS alone. The blue dots represent downregulated genes and
western-blot-based (e) analysis of Acod1 mRNA and ACOD1 protein expression the green dots represent upregulated genes in GC- and LPS-treated versus
in BMDMs treated with vehicle, GC and LPS. f, MS-based quantification of LPS-treated Acod1−/− BMDMs, in comparison to Acod1+/+ BMDMs. Data are
itaconate in cell lysates of BMDMs treated with vehicle, GC and LPS under mean + s.e.m. Statistical analysis was performed using unpaired t-tests with
normoxia and hypoxia (1% O2; left), or in the presence of vehicle or RXD false-discovery rate correction (a), one-way ANOVA with Tukey’s multiple-
(10 µM; right). n = 3 (Ctrl and GC) and n = 4 (LPS and GC + LPS). g, ELISA-based comparison test (b, c, f and g) and Wald tests with Bonferroni correction (i).

Fig. 5a). GC treatment fully prevented the LPS-induced lung inflam-

GC therapy relies on ACOD1 activity mation in wild-type mice. This involved GC-mediated inhibition of the
To understand the in vivo relevance of the identified mechanism, production of the pro-inflammatory cytokines TNF and IL-6 as well as
we initially studied the effect of GC treatment in a mouse model of inhibition of the recruitment of CD45+CD11b+ myeloid cells, including
LPS-induced lung injury in which intranasal application of LPS results polymorphonuclear neutrophils, into the lungs (Fig. 5a,b and Extended
in severe lung inflammation and immune pathology (Extended Data Data Fig. 5b–d). By contrast, GC treatment of Acod1−/− mice showed no

6 | Nature | www.nature.com
 a Myeloid cells PMNs b
80 P = 0.0001 80 P = 0.0001 300 400 LPS
P = 0.004

Percentage of cells

Percentage of cells
P = 0.001 LPS P = 0.004
P = 0.002 GC + LPS GC + LPS
P = 0.014

IL-6 (pg ml–1)

60 60 P = 0.014 300 P = 0.022

TNF (pg ml–1)

P = 0.011 200
40 40 200
100
20 20 100
0 0 0 0
GC – + – + GC – + – + GC – + – + GC – + – +
LPS + + + + LPS + + + + LPS + + + + LPS + + + +
Acod1+/+ Acod1–/– Acod1+/+ Acod1–/– Acod1+/+ Acod1–/– Acod1+/+ Acod1–/–

c d e Vehicle GC

P = 0.030

Itaconate (nmol ml–1)

6 –GC 2.0 P = 0.024
Serum itaconate

+GC Acod1+/+
(nmol ml–1)

1.5
4
1.0
2
0.5
0 0
Ctrl Veh. GC
K/BxN Acod1–/–

f Start P = 0.008 g OVA GC + OVA

10 Acod1+/+ Ctrl 100 P = 0.011 Ctrl
GC
Cumulative score

Acod1+/+ GC GC
8 80

AUC score
Acod1–/– Ctrl
6 60
Acod1–/– GC Acod1+/+
4 40
2 20
0 0
0 2 4 6 8 10 12 14 Acod1+/+ Acod1–/–
Time (days)

Acod1–/–

h OVA GC + OVA i

Percentage of eosinophils
Percentage of myeloid
Percentage of CD4+ cells

5 P = 0.011 80 80 Ctrl
P = 0.003 P = 0.002
4 GC
P = 0.012 60 P = 0.007 60 P = 0.012 OVA
Acod1+/+ 3 GC + OVA
40 40
2
1 20 20
CD45 (FI)

0 0 0
GC – + – + – + – + GC – + – + – + – + GC – + – + – + – +
OVA – – + + – – + + OVA – – + + – – + + OVA – – + + – – + +
Acod1+/+ Acod1–/– Acod1+/+ Acod1–/– Acod1+/+ Acod1–/–
Acod1–/–
j
800 150 P = 0.0004 600 Ctrl
P = 0.006 P = 0.0001
P = 0.014 GC
P = 0.027
IL-13 (pg ml–1)

600
IL-4 (pg ml–1)

IL-5 (pg ml–1)

P = 0.021 100 400 OVA

CD11b (FI) GC + OVA
400
50 200
200
0 0 0
GC – + – + – + – + GC – + – + – + – + GC – + – + – + – +
OVA – – + + – – + + OVA – – + + – – + + OVA – – + + – – + +
Acod1+/+ Acod1–/– Acod1+/+ Acod1–/– Acod1+/+ Acod1–/–

Fig. 5 | Therapeutic properties of GC-based therapies rely on itaconate Acod1−/− mice with K/BxN serum transfer arthritis with or without 0.5 mg per kg
production. a,b, Flow cytometry analysis of the bronchoalveolar lavage fluid GC treatment starting on day 4. Acod1+/+ mice: n = 10 (Ctrl) and n = 9 (GC);
(BALF) of mice (n = 8) (a) and quantification of cytokines (b) (Acod1+/+ mice: Acod1−/− mice: n = 7 (Ctrl and GC). For e, scale bars, 900 µm. g–j, Representative
n = 8 (LPS and GC + LPS); Acod1−/− mice: n = 8 (LPS) and n = 10 (GC + LPS)) in lung H&E staining (g), flow cytometry analysis (h) and quantification (i) of the
Acod1+/+ and Acod1−/− mice that received intraperitoneal dexamethasone (GC; BALF (Acod1+/+ mice: n = 6 (Ctrl), n = 5 (GC) and n = 4 (LPS and GC + LPS); Acod1−/−
5 mg per kg) and intranasal LPS (40 µg). The experimental design and gating mice: n = 5 (Ctrl), n = 4 (GC) and n = 5 (LPS and GC + LPS)), and quantification of
strategy are shown in Extended Data Fig. 5. PMNs, polymorphonuclear cytokines ( j) in Acod1+/+ and Acod1−/− mice during OVA-induced allergic airway
neutrophils. c, Serum itaconate of patients with rheumatoid arthritis with and inflammation with or without GC (Acod1+/+ mice: n = 7 (Ctrl), n = 5 (GC) and n = 9
without GC treatment. n = 21 (without GC) and n = 14 (with GC). d, Serum (LPS and GC + LPS); Acod1−/− mice: n = 5 (Ctrl), n = 4 (GC) and n = 9 (LPS and
itaconate of control mice (Ctrl; n = 3) and mice with K/BxN serum transfer GC + LPS). For g, scale bars, 200 µm. The experimental design and gating
arthritis with or without GC (n = 10). e,f, Representative paw haematoxylin and strategy are shown in Extended Data Fig. 5. Data are mean + s.e.m. Statistical
eosin (H&E) staining (dashed outlines indicate inflammatory infiltrates) (e) and analysis was performed using two-tailed unpaired t-tests (c) and one-way
clinical disease course including area under the curve (AUC) (f) in Acod1+/+ and ANOVA with Tukey’s multiple-comparison test (a, b, d, f, i and j).

beneficial effect in this disease model and these mice did not display patients with rheumatoid arthritis receiving a therapy with or without
a GC-induced reduction in cytokines or in the number of recruited GCs showed that GC treatment was associated with significantly higher
polymorphonuclear neutrophils. ACOD1-deficient mice therefore itaconate serum levels in patients with rheumatoid arthritis (Fig. 5c). To
experienced a severe LPS-induced pulmonal inflammation despite GC determine the potential functional relevance of GC-induced itaconate
treatment, confirming that the production of itaconate was essential production for the beneficial effects exerted by GCs during rheuma-
for the GC-induced anti-inflammatory effects in vivo (Fig. 5a,b and toid arthritis, we subsequently used the mouse model of K/BxN serum
Extended Data Fig. 5b–d). transfer arthritis, a preclinical model of rheumatoid arthritis in which
A heterogenous spectrum of immune-mediated inflammatory dis- injection of arthritogenic autoantibodies against glucose-6-phosphate
eases, including autoimmune-mediated inflammatory diseases such isomerase results in severe inflammation of the peripheral joints20
as rheumatoid arthritis as well as allergic diseases such as asthma, (Extended Data Fig. 5e). Here we tested a therapeutic GC regimen and
successfully responds to GC treatment. A cross-sectional analysis of started daily GC treatment after the onset of arthritis, which again

Nature | www.nature.com | 7
Article
resulted in increased itaconate serum levels as well as in a substantial the supernatant of cells and the sera of GC-treated mice and humans.
amelioration of arthritis severity (Fig. 5d–f). In turn, the absence of These observations show that GCs induce a systemic increase in ita-
ACOD1 abrogated the anti-inflammatory potential of GCs in this rheu- conate and strongly suggest that the levels of this metabolite reach
matoid arthritis mouse model, and GC-treated Acod1−/− mice displayed much higher concentrations in the inflamed tissue microenvironment
an unchanged disease course (Fig. 5e,f). in which myeloid-cell-derived itaconate might also act in a paracrine
To address the relevance of GC-induced itaconate production during manner thereby affecting other cell types such as stromal cells or lym-
the GC-mediated treatment of allergic diseases, we used the mouse phocytes. Metabolic reprogramming seems to essentially contribute
model of ovalbumin (OVA)-induced allergic airway inflammation as to the anti-inflammatory effects of GC in both mouse and human mac-
an established asthma model in which sensitization and exposure to rophages, although the relative contributions of itaconate and addi-
OVA promote eosinophil recruitment and an asthma-like pathology in tional, unidentified metabolites might differ between distinct species.
the lungs21,22 (Extended Data Fig. 5f). GC treatment of wild-type mice The GR was shown to exert various anti-inflammatory effects, includ-
resulted in a block of this OVA-triggered immune pathology, blunted ing both active induction of anti-inflammatory pathways27 and suppres-
eosinophil and T cell recruitment into the lung and inhibited the sion of inflammatory gene expression. Our current data argue against
expression of type 2 cytokines such as IL-4, IL-5 and IL-13 (Fig. 5g–j and the predominance of direct GR-mediated transcriptional transrepres-
Extended Data Fig. 5g). By contrast, GC-treated Acod1−/− mice showed sion of pro-inflammatory transcription factors. Instead, they suggest
full-blown allergic airway inflammation, continuous infiltration of that mitochondrial reprograming, which includes itaconate-mediated
immune cells and unchanged levels of type 2 cytokines despite GC suppression of a subset of inflammatory genes, accounts for a sub-
treatment (Fig. 5g–j). Together, these data suggest that the therapeutic stantial number of the anti-inflammatory effects observed after GC
anti-inflammatory effects exerted by GCs in diverse immune-mediated treatment in vitro and in vivo (Extended Data Fig. 6).
inflammatory diseases such as rheumatoid arthritis and asthma essen- The TCA-cycle-derived metabolite itaconate possesses potent anti-
tially depend on GC-induced mitochondrial reprogramming. inflammatory properties and exerts a range of molecular effects, includ-
ing the induction of an electrophilic stress response and activation of
anti-inflammatory transcription factors such as NRF2, which seems
Discussion to be involved in the anti-inflammatory effects of GCs as well. Another
Cellular metabolic pathways and metabolically active organelles such major consequence of itaconate production is the inhibition of succi-
as mitochondria have gained increased attention as key modulators nate dehydrogenase, which results in the reduction of mitochondrial
of the innate and adaptive immune response23. Both pro- and anti- ROS and an accumulation of succinate28, effects that we also observed
inflammatory stimuli promote differential metabolic reprogramming after GC treatment of LPS-activated macrophages.
that heavily shapes the cell’s polarization state24,25. GCs act as potent It will be of substantial interest to determine whether the non-
regulators of systemic energy homeostasis in the liver, fat and muscle, genomic GC-mediated modulation of mitochondrial metabolism and
suggesting that these compounds not only suppress the inflammatory TCA cycle acceleration also account for some of the systemic metabolic
response, but might also affect the cellular metabolism of immune side effects of long-term GC treatment. The side effects of continuous
cells. Recently, GCs were reported to modulate the metabolic state of long-lasting GC treatment include an increase in fat mass and devel-
macrophages, although the underlying molecular mechanisms and opment of type 2 diabetes, and have been considered to result from
potential consequences of this for the inflammatory response remain GR-mediated transcriptional activity within the nucleus of metaboli-
unclear26. Our current data reveal that GCs promote substantial mito- cally active cells such as adipocytes and hepatocytes1. More precise
chondrial reprogramming that involves an increase in PDH-mediated targeting of mitochondrial TCA cycle activity by GC analogues or small
pyruvate oxidation and PC-mediated pyruvate carboxylation, which molecules in immune cells might enable the disentangling of the benefi-
results in the acceleration and continuous fuelling of the TCA cycle cial and deleterious effects of GCs and therefore offer potential targets
as well as a rescue of mitochondrial functionality in LPS-activated for new classes of potent anti-inflammatory drugs.
macrophages.
Notably, similar GC-induced effects have been described in hepato-
cytes, in which GC-mediated activation of PDH and PC is essentially Online content
involved in hepatic gluconeogenesis9–11. These effects have been Any methods, additional references, Nature Portfolio reporting summa-
reported to occur within 30 min after GC treatment, indicating non- ries, source data, extended data, supplementary information, acknowl-
transcriptional mechanisms. Our data show that macrophages respond edgements, peer review information; details of author contributions
in a similar manner to hepatocytes and provide support that this and competing interests; and statements of data and code availability
GC-induced metabolic reprogramming is indeed independent of are available at https://doi.org/10.1038/s41586-024-07282-7.
GR-mediated transcriptional events, as it did not require nuclear trans-
location of the GR. Instead, the GR seems to be able to tether parts of 1. Hardy, R. S., Raza, K. & Cooper, M. S. Therapeutic glucocorticoids: mechanisms of actions
the PDH complex in the cytoplasm, where GC binding triggers the in rheumatic diseases. Nat. Rev. Rheumatol. 16, 133–144 (2020).
2. Buckley, L. & Humphrey, M. B. Glucocorticoid-induced osteoporosis. N. Engl. J. Med. 379,
release and mitochondrial translocation of regulatory PDH subunits, 2547–2556 (2018).
resulting in increased mitochondrial PDH activity and channelling of 3. Lillegraven, S. et al. Immunosuppressive treatment and the risk of diabetes in rheumatoid
pyruvate into the TCA cycle. arthritis. PLoS ONE 14, e0210459 (2019).
4. Vettorazzi, S., Nalbantoglu, D., Gebhardt, J. C. M. & Tuckermann, J. A guide to changing
The paradoxical persistence and acceleration of TCA cycle activ- paradigms of glucocorticoid receptor function-a model system for genome regulation
ity in GC-treated ACOD1-expressing pro-inflammatory macrophages and physiology. FEBS J. 289, 5718–5743 (2021).
then fosters potentiation of the production of the anti-inflammatory 5. Glass, C. K. & Saijo, K. Nuclear receptor transrepression pathways that regulate
inflammation in macrophages and T cells. Nat. Rev. Immunol. 10, 365–376 (2010).
metabolite itaconate. ACOD1-mediated itaconate production is other­ 6. Hübner, S., Dejager, L., Libert, C. & Tuckermann, J. P. The glucocorticoid receptor in
wise strictly limited owing to an early block of TCA cycle activity in inflammatory processes: transrepression is not enough. Biol. Chem. 396, 1223–1231 (2015).
pro-inflammatory cells (Fig. 4c and Extended Data Fig. 6). GC-mediated 7. O’Neill, L. A., Kishton, R. J. & Rathmell, J. A guide to immunometabolism for immunologists.
Nat. Rev. Immunol. 16, 553–565 (2016).
potentiation of itaconate production essentially contributes to the 8. Li, J. X. & Cummins, C. L. Fresh insights into glucocorticoid-induced diabetes mellitus and
anti-inflammatory effects of GCs in vitro as well as to the beneficial new therapeutic directions. Nat. Rev. Endocrinol. https://doi.org/10.1038/s41574-022-
effects of GC therapy in a range of mouse models of acute and chronic 00683-6 (2022).
9. Jones, C. G., Hothi, S. K. & Titheradge, M. A. Effect of dexamethasone on gluconeogenesis,
inflammatory diseases such as rheumatoid arthritis and asthma. The pyruvate kinase, pyruvate carboxylase and pyruvate dehydrogenase flux in isolated
GC-induced increase in itaconate was measurable in cellular extracts, hepatocytes. Biochem. J. 289, 821–828 (1993).

8 | Nature | www.nature.com
10. Martin, A. D., Allan, E. H. & Titheradge, M. A. The stimulation of mitochondrial pyruvate 22. Finotto, S. et al. Treatment of allergic airway inflammation and hyperresponsiveness by
carboxylation after dexamethasone treatment of rats. Biochem. J. 219, 107–115 (1984). antisense-induced local blockade of GATA-3 expression. J. Exp. Med. 193, 1247–1260 (2001).
11. Sistare, F. D. & Haynes, R. C. Jr. Acute stimulation by glucocorticoids of gluconeogenesis 23. Buck, M. D., Sowell, R. T., Kaech, S. M. & Pearce, E. L. Metabolic instruction of immunity.
from lactate/pyruvate in isolated hepatocytes from normal and adrenalectomized rats. Cell 169, 570–586 (2017).
J. Biol. Chem. 260, 12754–12760 (1985). 24. Ip, W. K. E., Hoshi, N., Shouval, D. S., Snapper, S. & Medzhitov, R. Anti-inflammatory effect
12. Karra, A. G. et al. Proteomic analysis of the mitochondrial glucocorticoid receptor of IL-10 mediated by metabolic reprogramming of macrophages. Science 356, 513–519
interacting proteins reveals pyruvate dehydrogenase and mitochondrial 60 kDa heat (2017).
shock protein as potent binding partners. J. Proteom. 257, 104509 (2022). 25. Cheng, S. C. et al. mTOR- and HIF-1α-mediated aerobic glycolysis as metabolic basis for
13. Santos, J. H. Mitochondria signaling to the epigenome: a novel role for an old organelle. trained immunity. Science 345, 1250684 (2014).
Free Radic. Biol. Med. 170, 59–69 (2021). 26. Stifel, U. et al. Glucocorticoids coordinate macrophage metabolism through the regulation
14. Ryan, D. G. & O’Neill, L. A. J. Krebs cycle reborn in macrophage immunometabolism. of the tricarboxylic acid cycle. Mol. Metab. 57, 101424 (2022).
Annu. Rev. Immunol. 38, 289–313 (2020). 27. Clark, A. R. Anti-inflammatory functions of glucocorticoid-induced genes. Mol. Cell.
15. Hooftman, A. & O’Neill, L. A. J. The immunomodulatory potential of the metabolite Endocrinol. 275, 79–97 (2007).
itaconate. Trends Immunol. 40, 687–698 (2019). 28. Lampropoulou, V. et al. Itaconate links inhibition of succinate dehydrogenase with
16. Peace, C. G. & O’Neill, L. A. The role of itaconate in host defense and inflammation. J. Clin. macrophage metabolic remodeling and regulation of inflammation. Cell Metab. 24,
Invest. https://doi.org/10.1172/jci148548 (2022). 158–166 (2016).
17. Mills, E. L. et al. Itaconate is an anti-inflammatory metabolite that activates Nrf2 via
alkylation of KEAP1. Nature 556, 113–117 (2018). Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
18. Bambouskova, M. et al. Electrophilic properties of itaconate and derivatives regulate the published maps and institutional affiliations.
IκBζ-ATF3 inflammatory axis. Nature 556, 501–504 (2018).
19. O’Neill, L. A. J. & Artyomov, M. N. Itaconate: the poster child of metabolic reprogramming Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this
in macrophage function. Nat. Rev. Immunol. 19, 273–281 (2019). article under a publishing agreement with the author(s) or other rightsholder(s); author
20. Christensen, A. D., Haase, C., Cook, A. D. & Hamilton, J. A. K/BxN serum-transfer arthritis self-archiving of the accepted manuscript version of this article is solely governed by the
as a model for human inflammatory arthritis. Front. Immunol. 7, 213 (2016). terms of such publishing agreement and applicable law.
21. Aun, M. V., Bonamichi-Santos, R., Arantes-Costa, F. M., Kalil, J. & Giavina-Bianchi, P. Animal
models of asthma: utility and limitations. J. Asthma Allergy 10, 293–301 (2017). © The Author(s), under exclusive licence to Springer Nature Limited 2024

Nature | www.nature.com | 9
Article
Methods Cloned oligonucleotides were as follows (5′ to 3′ orientation): Acod1
sgRNA 1F, CACCGACTACCCAACGTTCCTACCA; 1R, AAACTGGTAGGAA
Primary mouse cell cultures CGTTGGGTAGTC; Acod1 sgRNA 2F, CACCGTATGTGAAACACTTCCGTAG;
BMDMs. Bone marrow was isolated from the femurs and tibias of 8- to 2R, AAACCTACGGAAGTGTTTCACATAC; Acod1 sgRNA 3F, CACCGGT
12-week old C57BL/6J mice ( Janvier Labs), Acod1+/+ or Acod1−/− mice TTATACAGATTCACTCCA; 3R, AAACTGGAGTGAATCTGTATAAACC.
( Jackson Research Laboratory) or Nfe2le+/+ or Nfe2le−/− mice (provided by sgRNA oligos with overhangs were annealed using T4 DNA ligase
F. Gruber). After erythrocyte lysis, cells were incubated at 37 °C with 5% (Thermo Fisher Scientific) and ligated into LentiCRISPRv2 construct
CO2 in BMDM medium consisting of Dulbecco’s modified Eagle medium (Addgene plasmid, 52961) digested with BsmBI-v2 restriction enzyme
(DMEM) (GIBCO) supplemented with 10% fetal calf serum (Biochrom), (New England BioLabs) and amplified in chemically competent (Zymo
1% penicillin–streptomycin (GIBCO) and 10% M-CSF-containing L929- Research) Stbl3 Escherichia coli cultures (Thermo Fisher Scientific).
supernatant. L929 cells were a gift from R. Lang. The day after isolation, HEK293T/17 cells were transfected with polyethylenimine (linear, MW
non-adherent cells were plated in fresh BMDM medium for 5 days. 25,000; Polysciences Europe) using the respective LentiCRISPRv2
constructs and packaging plasmids psPAX2 (Addgene plasmid, 12260)
FLDMs. Livers were isolated from embryonic day 14.5 (E14.5) mouse and pMD2.G (provided by D. Trono). Medium was changed to DMEM
embryos and meshed using a 5 ml syringe. After erythrocyte lysis, cells alone 24 h and 48 h after transfection. Virus supernatants were col-
were incubated at 37 °C with 5% CO2 in BMDM medium supplemented lected 72 h and 96 h after transfection and concentrated with lentiviral
with 30% M-CSF-containing L929-supernatant. The day after isolation, concentrator solution containing 10% (w/v) PEG-8000 and 0.3 M NaCl.
non-adherent cells were plated in fresh BMDM medium for 5 days. THP-1 cells were infected by spinfection with 1.25 × 107 IU ml−1 con-
centrated virus supernatant (determined using the qPCR Lentivirus
In vitro treatments. Differentiated primary mouse macrophages were Titration Kit; Applied Biological Materials) in RPMI containing 8 µM
pretreated with 10 µM of RXD (Cayman Chemical), 10 µM of UK5099 cyclosporine-H (AdooQ Biosciences) and 1 mg ml−1 Lentiboost (Sirion
(Cayman Chemical), 62.5 µM dimethyl itaconate (DMI, Sigma-Aldrich) Biotech) at 800g for 45 min at 25 °C. The medium was replaced the next
or 62.5 µM of 4-octyl-itaconate (Sigma-Aldrich), where applicable, for day with RPMI alone. The next day, selection with 2 µg ml−1 puromycin
1 h before stimulation with 100 nM of dexamethasone (Sigma-Aldrich) (Sigma-Aldrich) was initiated and cells were maintained on selection for
and/or 100 ng ml−1 of LPS (Sigma-Aldrich). For stabilization of NRF2, 48 h. The knockout efficiency was determined using western blotting.
2 µM of the proteasome inhibitor MG-132 (Cell Signaling Technology)
was added 6 h before cell collection for western blotting. To evaluate Stimulation and treatments. THP-1 cells were seeded in 12-well plates
the effects of hypoxia, cells were incubated in normoxic conditions (1 × 106 cells per well in 1 ml RPMI) and treated with 25 ng ml−1 of PMA
(21% O2) or 1% O2 using the H25 Hypoxystation (Whitley Scientific). For (Sigma-Aldrich) for 48 h. The medium was then replaced with RPMI
measurement of secreted TNF and IL-6, cells were activated for 6 h and alone for an additional 24 h, after which cells were stimulated with
24 h, respectively. To induce inflammasome activation required for 100 nM of dexamethasone and/or 100 ng ml−1 of LPS in combination
IL-1β processing and secretion, cells were treated with 5 mM of ATP with 20 ng ml−1 human IFNγ (BioLegend) for 24 h. The levels of secreted
(Sigma-Aldrich) for 45 min, 6 h after activation. Data were pooled from TNF and IL-6 were then measured. To induce inflammasome activation
individual cultures over different experiments. required for IL-1β processing and secretion, cells were treated with 5 mM
of ATP (Sigma-Aldrich) for 45 min, 24 h after activation.
Human monocyte-derived macrophage cultures
Peripheral blood was collected from healthy donors and peripheral Patients
blood mononuclear cells were isolated using Ficoll-Paque gradient cen- Informed consent from all patients providing serum samples was
trifugation (Lymphoflot; Bio-Rad). Peripheral blood mononuclear cells obtained. Analysis of human material was approved by the ethics com-
were washed and CD14+ monocytes were isolated using the MojoSort mittee of the University Hospital Erlangen. The patients with rheuma-
human CD14+ monocyte isolation kit according to the manufacturer’s toid arthritis fulfilled the 2010 ACR/EULAR criteria for rheumatoid
instructions (BioLegend). Purified monocytes were cultured in RPMI arthritis30 and were selected on the basis of the absence or presence
(GIBCO) supplemented with 10% fetal calf serum, 10 mM HEPES (Gibco), of ongoing GC therapy (equivalent to 25 mg to 50 mg prednisolone
1% l-glutamine (Gibco) and 1% penicillin–streptomycin, containing per day).
50 ng ml−1 human recombinant M-CSF, for 6 days. Where needed, cells
were pretreated with 500 µM of the ACOD1 inhibitor ERG344 (pro- Mouse models
vided by A. E. Papatjamassiu) for 1 h before stimulation with 100 nM All animal experiments were conducted in accordance with German
of dexamethasone and/or 100 ng ml−1 of LPS for 24 h and the levels of guidelines and laws and were approved by the local animal ethics com-
secreted TNF and IL-6 were measured. To induce inflammasome acti- mittees of the Regierung von Mittelfranken and the Regierungsprä-
vation required for IL-1β processing and secretion, cells were treated sidium Karlsruhe and Tübingen. Acod1+/+ or Acod1−/− littermates were
with 5 mM of ATP (Sigma-Aldrich) for 45 min, 24 h after activation. To used for all of the in vivo experiments. Fetal liver cells were derived
evaluate the effects of hypoxia, cells were incubated under normoxic from mouse embryos with a mutation in the Nls1 sequence of the GR
conditions (21% O2) or 1% O2 using the H25 Hypoxystation (Whitley (NR3C1) at position 510–512, which results in the exchange of three
Scientific). lysine residues to three asparagine residues and is reported to disrupt
GR nuclear localization (referred to as Nlsmut), or from wild-type litter-
ACOD1-deficient THP-1 cell cultures mates31. Animals were housed and bred at the animal facilities of the
HEK293T/17 cells from ATCC (CRL-11268) were cultured in DMEM con- Friedrich-Alexander University of Erlangen-Nuremberg or Ulm Univer-
taining 1% penicillin–streptomycin and 10% fetal calf serum (Sigma- sity under specific-pathogen-free conditions. Mice were maintained
Aldrich). THP-1 cells from ATCC (TIB-202) were cultured in RPMI under a 12 h–12 h light–dark cycle at around 20–24 °C with 45–65%
containing 1% penicillin–streptomycin and 10% fetal calf serum. humidity with ad libitum access to food and water.

CRISPR–Cas9-mediated gene deletion in THP-1 cells. sgRNAs were LPS-induced acute lung injury mouse model. Mice (aged 10–12
designed using the Broad Institute sgRNA design tool (https://portals. weeks) of both sexes were administered 40 µg of LPS intranasally (20 µl
broadinstitute.org/gppx/crispick/public). The control sgRNA tar­ per nostril) alongside an injection of 5 mg per kg of dexamethasone or
geting Renilla luciferase (sgRen) has been described previously29. vehicle (0.9% NaCl) intraperitoneally. Then, 24 h later, the mice were
euthanized, the BALF was recovered for flow cytometry and ELISA and anti-PDH (Cell Signaling Technology, 3205) antibodies. Rabbit anti-β-
the lungs were fixed in 4% buffered formalin (pH 7.4) and embedded in actin (Sigma-Aldrich, A2066) antibodies were used as a loading control.
paraffin before sectioning. Then, 2 mm sections were stained with H&E The original uncropped blots are provided in Supplementary Figs. 1–5.
using standard procedures. Photographs were taken using the Keyence
BZ-X710 Microscope using BZ-X View and BZ-X Analyzer software series Co-immunoprecipitation assay
2:22 (Keyence) at ×10 optical zoom (Nikon PlanApoλ10×0.45/4 mm BMDMs (5 × 106) were treated with 100 nM of dexamethasone for 24 h,
objective). lysed using 1× cell lysis buffer (Cell Signaling Technology) according
to the manufacturer’s instructions and 200 µg of whole-cell lysates
OVA-induced asthma mouse model. Mice (aged 6–8 weeks) of both was incubated with 2 µg of rabbit anti-GR (Cell Signaling Technology,
sexes received an intraperitoneal injection of 200 μg of OVA (Sigma- 12041) or normal rabbit IgG as a control (Cell Signaling Technology,
Aldrich) in 0.9% NaCl complexed with Alum (Sigma-Aldrich) on days 2729) at 4 °C with rotation overnight. Ligated proteins were precipitated
0 and 7. On days 13, 14 and 15, mice received 0.5 mg of dexamethasone using protein A agarose beads (Cell Signaling Technology), separated
intranasally 30 min before intranasal administration of 50 µg of OVA. on 12.5% SDS–polyacrylamide gel, transferred onto a polyvinylidene
Mice were euthanized on day 16, the BALF was recovered for flow cyto­ difluoride membrane and immunoblotted with rabbit anti-PDH (Cell
metry and ELISA, and the lungs were fixed in 4% buffered formalin Signaling Technology, 3205) antibodies.
(pH 7.4) and embedded in paraffin before sectioning. Then, 2 mm sec-
tions were stained with H&E using standard procedures. Photographs ELISA
were taken with the Keyence BZ-X710 Microscope using BZ-X View and ELISA-based measurements of BMDM supernatants or BALF were per-
BZ-X Analyzer software series 2:22 (Keyence) at ×10 optical zoom (Nikon formed according to the manufacturer’s protocol using the DuoSet
PlanApoλ10×0.45/4 mm objective). ELISA development System (R&D) for mouse TNF (DY410), mouse
IL-6 (DY406), mouse IL-1β (DY401), mouse IL-4 (DY404), mouse IL-5
K/BxN serum transfer arthritis mouse model. Mice (aged 8–12 weeks) (DY405) and mouse IL-13 (DY413), as well as for human TNF (DY210),
of both sexes were used. K/BxN serum transfer arthritis was induced human IL-6 (DY206) and human IL-1β (DY201). The optical density was
by intraperitoneal injection of 150 µl of K/BxN serum collected from measured at 450 nm with a wavelength correction at 540 nm using the
arthritic K/BxN mice. Clinical development of arthritis was evaluated SpectraMax190 system with the SoftMax Pro v.7.1 software (Molecular
using a clinical index ranging from 0 (minimum) to 16 (maximum), Devices).
which represents a cumulative score of 0 to 4 for each paw: 0, no signs
of inflammation; 1, minor swelling and reddening of a paw, or affect- RNA-seq analysis
ing only single digits; 2, moderate swelling and erythema, or affecting Total BMDM RNA was isolated using the RNeasy Mini Kit for RNA purifi-
multiple digits per paw; 3, severe swelling and erythema affecting the cation in according to the manufacturer’s protocol (Qiagen). Paired-end
whole paw; 4, maximum swelling and erythema. From day 4 onwards, 150 bp sequencing (PE150) was used to sequence the libraries on the
mice were injected daily with 0.5 mg per kg of dexamethasone or vehicle NovaSeq platform (Illumina). The reads were obtained in fastq format
(0.9% NaCl). On day 14, mice were euthanized and the hind paws were and quality controlled using fastqc v.0.11.7. Reads were then mapped
collected, fixed in Histofix 4% (Roth) overnight and bones were decal- against the mouse reference genome (Ensembl GRCm38, release 99)
cified for 10 days using Teitel buffer and embedded in paraffin before using the STAR aligner v.2.5.1b. Read counts per gene were then quanti-
sectioning. Then, 2 mm sections were stained with H&E using standard fied using featureCounts v.2.0.0. Subsequent analyses were performed
procedures. Photographs were taken with the Keyence BZ-X710 Micro- using the R programming language v.3.6.3. Specifically, differential
scope using BZ-X View and BZ-X Analyzer software series 2:22 (Keyence) gene expression analysis was completed using the DESeq2 package
at ×4 optical zoom (Nikon PlanApoλ4×0.2/20 mm objective). v.1.36.0, transcription factor activity prediction was performed using
the dorothea package v.1.8.0. and pathway enrichment analysis was
RT–qPCR performed using the pathfinder package v.1.6.3. Moreover, plots were
Total BMDM RNA was isolated using the NucleoSpin RNA Mini Kit for generated using the ggplot2 package v.3.3.6 and the ComplexHeatmap
RNA purification according to the manufacturer’s protocol (Machery- package v.2.12.0.
Nagel). A total of 500 ng of RNA was reverse-transcribed into cDNA.
mRNA expression was quantified using the SYBR Green I Kit (Roche) Extracellular flux assay
and the QuantStudio 6 Flex with QuantStudio Real Time Systems Real-time bioenergetic profiles of BMDMs and FLDMs were determined
LightCycler (Thermo Fisher Scientific). Acod1, C1qb, Cd163 and by measuring the OCR and ECAR using the XFe96 Extracellular Flux
Fkbp5 mRNA quantities are presented as ΔΔCt values normalized to Analyzer with Wave v.2.4.2 (Seahorse Bioscience, Agilent Technolo-
Actb expression. The following primers were used: Actb, 5′-TGTCCAC gies). BMDMs (80,000 cells) were pretreated with 10 µM RXD or vehicle
CTTCCAGCAGATGT-3′ (forward) and 5′-AGCTCAGTAACAGTCCGC for 1 h before treatment with the vehicle, dexamethasone (100 nM)
CTAGA-3′ (reverse); Acod1, 5′-GCGAACGCTGCCACTCA-3′ (forward) and/or LPS (100 ng ml−1) for 24 h. Assays were performed according
and 5′-ATCCCAGGCTTGGAAGGTC-3′ (reverse); C1qb, 5′-CAACCAG to the manufacturer’s protocol. In brief, the Cell Mito Stress Test was
GCACTCCAGGGATAA-3′ (forward) and 5′-CCAACTTTGCCTGGAGTC performed by sequential titration to 1 µM oligomycin, 1.5 µM carbonyl
CCAG-3′ (reverse); Cd163, 5′-TGCTGTCACTAACGCTCCTG-3′ (forward) cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP) and 0.5 µM
and 5′-TCATTCATGCTCCAGCCGTT-3′ (reverse); Fkbp5, 5′-AGCAACG rotenone/antimycin A. A cellular Glyco Stress Test was performed by
GTAAAAGTCCACCT-3′ (forward) and 5′-TTCCCCAACAACGAACACCA-3′ sequential titration to 10 mM glucose, 1 µM oligomycin and 50 mM
(reverse). 2-deoxyglucose. Mito Fuel Flex Tests were performed by adding 3 µM
BPTES or 2 µM UK5099 (provided in the test kit, Agilent) after measur-
Western blot ing the baseline respiration. The basal and maximal respiration as well
BMDMs were lysed with Laemmli-Buffer (Bio-Rad) containing as glycolysis and glycolytic capacity were subsequently calculated
β-mercaptoethanol. Protein extracts were separated on 4–20% pre- according to the manufacturer’s recommendations.
cast gels (Bio-Rad), transferred onto a polyvinylidene difluoride mem-
brane (Bio-Rad) and immunoblotted with rabbit anti-HIF-1α (Novus Metabolomics
Biologicals, NB100-449), rabbit anti-ACOD1 (Cell Signaling Technology, Perchloric acid was used to extract phosphorylated intermediates and
17805), rabbit anti-NRF2 (Cell Signaling Technology, 12721) and rabbit carboxylates as described previously32, applying ionchromatography
Article
with the ICS3000 HPLC-system (Dionex) and electrospray ionization–
tandem MS detection using the QTrap3200 Triple-Quadrupole mass Proximity ligation assay
spectrometer with turbo V ion source (Applied Biosystems) operated The proximity ligation assay (Duolink DUO92008, Sigma-Aldrich) was
in multiple reaction monitoring mode. Shock-frozen BMDMs (50 mg) performed with BMDMs treated with vehicle or 100 nM dexamethasone
or 100 µl of cultured supernatant, 100 µl of serum from mice injected for 24 h. After a PBS wash, cells were fixed with 4% PFA for 30 min at
or not with KBx/N serum and treated or not with 0.5 mg per kg dexa- room temperature. After fixation, cells were permeabilized with ice-
methasone after 14 days, or 100 µl of serum from patients with rheu- cold methanol for 30 min. The block with the provided blocking solu-
matoid arthritis who were treated or not with 25 to 50 mg per day of tion was performed for 1 h at room temperature, followed by incubation
prednisolone was used for extracts. with antibodies (GR mixture 1:200, Cell Signaling, 3660 and 12041) and
PDHE1a 1:100 (Santa Cruz, D-6) for 1 h at 37 °C. Enzymatic reactions
Carbon-tracing experiments (ligation and amplification) were performed according to the manu-
BMDMs were stimulated for 24 h with dexamethasone and/or LPS, after facturer’s instructions. Images were acquired using the ImageXpress
which the medium was exchanged with either complete medium or MicroConfocal Microscope at 40× magnification (Molecular Devices).
medium containing [13C6]glucose (4.5 g l−1, Sigma-Aldrich) and further The numbers of nuclei and interaction events were quantified using
incubated for 24 h. Cells were lysed with 1 ml of ice-cold methanol/ ImageJ v.1.52a (Fiji).
water (80/20, v/v). After centrifugation at 16,000 rpm and 4 °C for
5 min using the Eppendorf 5427R centrifuge (Eppendorf), 350 µl of Flow cytometry
the supernatant was pipetted into separate liquid chromatography MitoTracker. After stimulation with dexamethasone and/or LPS for
(LC) vials (VWR). A pool sample was prepared by mixing 230 µl of the 6 h or 24 h, BMDMs were stained with 50 nM MitoTracker Red CMXRos
supernatant from each sample. From this mixture, 350 µl was taken (Invitrogen) and 100 nM MitoTracker Green (Invitrogen) in BMDM
into LC vials for each quality-control sample. All of the samples were medium for 30 min at 37 °C with 5% CO2. Cells were then washed with
dried under a gentle stream of nitrogen at 30 °C. Subsequently, the PBS, detached using Cellstripper (Corning) and resuspended in 2%
dried samples were reconstituted in 200 µl eluent (99.5% water, 0.5% heat-inactivated FCS in PBS. Flow cytometry was performed using
136 methanol (VWR chemicals), both with 0.1% formic acid). the CytoFLEX S (Beckman Coulter) system. Data were analysed using
CytExpert (Beckman Coulter, v.2.4) and FlowJo (v.10.6.1).
LC–MS analysis. All analyses were performed on the Dionex Ultimate
3000 chromatography system hyphenated to a Q Exactive Focus mass BALF. BALF was recovered and cells were blocked with 10% rat serum
spectrometer (both from Thermo Fisher Scientific) with a heated and 1% Fc Block (TruStain FcX, BioLegend) in PBS for 10 min at room
electrospray source. The instrument was controlled using Trace- temperature and stained with BV510 anti-mouse CD45 (30-F11), PE
Finder 4.1 (Thermo Fisher Scientific). The installed column was an anti-mouse CD11b (M1/70), A700 anti-mouse Ly6G (HK1.4), A647 anti-
Acquity UPLC BEH C18, 1.7 µm, 2.1 ×100 mm column. A 2.1 ×5 mm guard mouse CD64 (X54-5/7.1), FITC anti-mouse SIGLECF (S1700SL), PE/Cy7
column was installed to prolong the column lifetime (all columns anti-mouse CD3 (17A2) or APC anti-mouse CD4 (GK1.5) fluorophore-
were from Waters). The column temperature was 40 °C, the flow rate conjugated antibodies for 30 min at 4 °C. After washing, cells were
was 0.35 ml min−1 and the injection volume was 2 µl. The eluent was resuspended in 2% heat-inactivated FCS in PBS. Flow cytometry was
as follows: A, water, 0.1% formic acid; and B, methanol, 0.1% formic performed using the CytoFLEX S (Beckman Coulter) system. Data were
acid. Chromatography separation was achieved using the follow- analysed using CytExpert (Beckman Coulter, v.2.4) and FlowJo (v.10.6.1).
ing gradient program: 10% B (0 min) increased to 98% B (9 min until
11 min), back to 10% B (11.1 min until 15 min). The Q Exactive Focus Immunofluorescence microscopy
mass spectrometer was operated with a capillary voltage of −3 kV in Sample preparation and stainings. BMDMs were seeded onto glass
negative mode. The capillary temperature was 380 °C and the auxil- coverslips at 70,000 cells per well in 24-well plates and stimulated
iary gas temperature was 400 °C. The sheath gas pressure, auxiliary with 100 nM of dexamethasone and/or 100 ng ml−1 of LPS for 6 h or
gas pressure and sweep gas flow rate were set to 60, 20 and 0 a.u., 24 h before fixation with 4% PFA in PBS for 5 min at room temperature.
respectively. Nitrogen 5.0 was used for these gases. The scanning Cells were then washed with PBS and permeabilized using a 0.1% Triton
range was 66.7 to 1,000 m/z in discovery mode. The resolution of the X-100 solution in PBS for 15 min. Subsequently, the blocking step was
analyser in full scan was set to 70,000 and the maximum inject time performed using a 5% FCS solution in PBS containing TruStain FcX (Bio-
was set at auto with an AGC target of 1 × 106. In ddMS2, the resolution Legend) for 1 h at room temperature. Staining with primary antibodies
was set to 35,000 with the maximum injection time set to auto and was performed in blocking buffer at 4 °C overnight using an unconju-
an AGC target of 5 × 104. gated anti-rabbit PDHX antibody (Proteintech), unconjugated anti-
For data analysis, Compound Discoverer v.3.1 was used (Thermo mouse GR antibody (Cell Signaling Technology, 12041), MitoTracker
Fisher Scientific). Compound annotations were made using mzcloud, Red CMXRos (Invitrogen) or Alexa Fluor 488 anti-mouse phalloidin
mzVault and ChemSpider databases. The annotations of TCA metabo- (Inivitrogen). Secondary staining was performed at room temperature
lites as well as itaconate were confirmed using an in-house library. for 2 h using Alexa Fluor plus 488 anti-rabbit IgG (Invitrogen) or Alexa
Fluor plus 647 anti-rabbit IgG (Invitrogen) antibodies. DAPI nuclear
PDH and PC enzymatic activity quantification staining (Carl Roth) was performed for 5 min at room temperature
BMDMs were stimulated for 6 h or 24 h with dexamethasone and/or before washing the unbound antibodies with PBS.
LPS, after which cells were lysed using the buffer provided in the kit
according to the manufacturer’s instructions, and PDH enzymatic Visualization of spatial relationships of PDHX or GR with mitochon-
activity (Sigma-Aldrich, MAK183) or PC enzymatic activity (Cohesion dria. PFA-fixed and immune-stained cells were imaged using a Leica
Biosciences CAK1262) was measured using colorimetric assays. TCS SP8 confocal laser scanning microscope with acousto-optic tune-
able filter and acousto-optic beam splitter, internal hybrid detectors
Acetyl CoA quantification (HyD SP) and las AF software v.2.7.3.9723 (Leica). Imaging of Fluoro-
BMDMs were stimulated for 6 h or 24 h with dexamethasone and/or mount G-embedded samples was performed using an HC PL APO CS2
LPS, after which cells were lysed using the buffer provided in the kit 63×/1.30 NA GLYC objective combined with digital zoom. Fluorescence
according to the manufacturer’s instructions (Sigma-Aldrich, MAK039), signals were measured with sequential scans: Phalloidin A488 was
and acetyl CoA was quantified using a fluorometric assay. excited with an argon laser at 488 nm and detected with an internal HyD
at 500–550 nm; MitoTracker Red was excited with a diode-pumped GSE250274. The following databases were used in this study: the Broad
solid-state laser at 561 nm and detected with an internal HyD at Institute sgRNA design tool (https://portals.broadinstitute.org/gppx/
571–621 nm. The third sequence to visualize PDHX or GCR counter- crispick/public), Ensembl mouse reference genome GRCm38 (http://
stained with Alexa Fluor plus 647 was performed with a helium-neon www.ensembl.org/Mus_musculus/Info/Index), mzcloud and mzVault
laser at 633 nm for excitation and an internal HyD at 650–700 nm for (https://www.mzcloud.org/) and ChemSpider (http://www.chemspider.
detection. DAPI was excited in the final sequence with a diode-pumped com/). Source data are provided with this paper.
solid-state laser at 405 nm and detected with an internal HyD at
430–580 nm. All confocal measurements for subcellular localization 29. Bigenzahn, J. W. et al. LZTR1 is a regulator of RAS ubiquitination and signaling. Science
362, 1171–1177 (2018).
of PDHX or GCR were performed according to Nyquist criteria. 30. Aletaha, D. et al. 2010 rheumatoid arthritis classification criteria: an American College of
The generated images were deconvoluted using Huygens Profes- Rheumatology/European League Against Rheumatism collaborative initiative. Ann. Rheum.
sional (v.22.10, Scientific Volume Imaging) and visualized using Imaris 69, 1580–1588 (2010).
31. Savory, J. G. et al. Discrimination between NL1- and NL2-mediated nuclear localization of
software (v.9.7.1, Bitplane). The amount of PDHX localized in mito- the glucocorticoid receptor. Mol. Cell. Biol. 19, 1025–1037 (1999).
chondria was determined by unbiased manual volume rendering of 32. Hofmann, J., Börnke, F., Schmiedl, A., Kleine, T. & Sonnewald, U. Detecting functional
groups of Arabidopsis mutants by metabolic profiling and evaluation of pleiotropic
the corresponding channels, Alexa Fluor plus 647 and MitoTracker,
responses. Front. Plant Sci. 2, 82 (2011).
respectively. All overlapping volumes were calculated as a percentage
for each condition. The mitochondrial content per cell was assessed
Acknowledgements We thank W. Baum who helped in generating the K/BxN serum;
by summing the rendered volumes of MitoTracker signal for each cell. A. Klej, R. Weinkam and R. Mancuso for technical assistance; and A. Papathanassiu (Ergon
Exemplar images for each conditions were extracted from deconvo- Pharmaceuticals) for providing the ACOD1 inhibitor ERG344. This work was supported by the
luted images using Imaris Bitplane software. The Phalloidin channel Deutsche Forschungsgemeinschaft (DFG, CRC1181-A03/A01/B08 to G.K., G. Schett and S.F.;
FOR2886 PANDORA, B01/A03/B05/Z01, to G.K., G. Schett, S.U. and A.K.; CRC1149, 251293561,
was adjusted separately for each condition; DAPI, MitoTracker and C02 and CRC 1506, 450627322, C05 to J.T.; Tu220/25-1, 505870049 and Major Research
PDHX were displayed with the same brightness and contrast within Instrumentation grant 441730715), the Emerging Field Initiative (EFI) of the Friedrich-Alexander
one dataset. University Erlangen-Nürnberg (FAU) (EFI_Verbund_Med_05_MIRACLE to G.K., U. Sonnewald and
J.H.), the Interdisziplinäres Zentrum für Klinische Forschung (IZKF) of the Friedrich-Alexander
University Erlangen-Nürnberg (IZKF J91 to J.-P.A.), the Bundesministerium für Bildung und
Statistics and reproducibility Forschung (BMBF) (MASCARA to G.K. and G.S and MelAutim to G.K.), the Christian Doppler
GraphPad Prism v.8.3 was used to calculate statistical significance. Laboratory Arginine Metabolism in Rheumatoid Arthritis and Multiple Sclerosis (to
G. Schabbauer, S.B., M.K. and M.H.), the FWF Sonderforschungsbereich SFB F83 (to G.S.)
Where statistics were derived, experiments were independently per- and the EU (Horizon 2020 ERC-2014-StG 640087, SOS and Horizon 2020 ERC-2020-CoG
formed three times for in vitro experiments, with the exception of the 101001866, INSPIRE to G.K.; Horizon Europe ERC-2021-StG 101039438, NEXUS, to S.U.; and
Horizon 2020 ERC-2018-SyG nanoSCOPE and RTCure to G.S.). J.-P.A. received additional
following experiments, which were performed twice (Figs. 1e, 2c–f
financial support from the Canadian Institutes for Health Research (CIHR, 201711MFE-395641-
and 3c and Extended Data Figs. 1d, 2b,c,e, 3b and 4f). For Extended 294537) and Fonds de recherche du Québec—Santé (FRQS, 259908). M.H. was awarded a
Data Fig. 4e, experiments were independently performed four times. DOC fellowship by the Austrian Academy of Sciences. S.U. was supported by the Hightech
Agenda Bavaria.
Experiments were independently performed twice for all of the in vivo
mouse experiments. Author contributions J.-P.A., M.Z., M.F. and U. Stifel designed the study, performed experiments,
interpreted results and wrote the manuscript. B.K., R.V.T., C.G., G.E., C. Stoll, O.B.B., M.P.M.,
Reporting summary C. Scholtysek, M.B., K.P.-Z., M.S.J.M., M.D., M.K. and M.H. performed experiments, collected data
and interpreted results. S.U., D.M., U. Sonnewald, D.S., A.K., S.F. and J.H. provided expertise,
Further information on research design is available in the Nature Port- essential material and input and wrote the manuscript. D.C. and F.H. performed bioinformatics
folio Reporting Summary linked to this article. analysis and interpreted the data. S.B., G. Schabbauer, A.G., E.L. and G. Schett wrote the
manuscript and provided valuable input. J.T. and G.K. designed the study and experiments and
wrote the manuscript. All of the authors read and commented on the manuscript.

Data availability Competing interests The authors declare no competing interests.

The data supporting the findings of this study are available on rea-
Additional information
sonable request from the corresponding author or J.-P.A. The data Supplementary information The online version contains supplementary material available at
are not publicly available as they comprise information that could https://doi.org/10.1038/s41586-024-07282-7.
compromise research participant privacy, although source data are Correspondence and requests for materials should be addressed to Gerhard Krönke.
Peer review information Nature thanks Luke O’Neill, David Ray and the other, anonymous,
provided with this paper. Bulk RNA-seq datasets generated in this study reviewer(s) for their contribution to the peer review of this work.
are available at the GEO under accession numbers GSE250273 and Reprints and permissions information is available at http://www.nature.com/reprints.
Article
a 4h 24 h b Diabetic cardiomyopathy
Oxidative phosphorylation
E2f4 Ctrl Parkinson disease
Jun GC

1
Prion disease
Nfkb1 LPS Non−alcoholic fatty liver disease
Rela GC + LPS
Cardiac muscle contraction
Spi1 SNARE interactions in vesicular transport

2
Stat3
Irf9 Base excision repair
Rel Mismatch repair − log 10 ( p)
Fos Fanconi anemia pathway
12.5

3
Foxl2 Nucleotide excision repair
Hif1a DNA replication 10.0
Colorectal cancer
Stat2 7.5
Irf1 NF−kappa B signaling pathway
Stat1 Th17 cell differentiation 5.0
Nfya Human T−cell leukemia virus 1 infection
2.5
Rest Insulin resistance
Egr1 Th1 and Th2 cell differentiation

4
Esr1 Adipocytokine signaling pathway
# genes
Fosl2 PD−L1 expression and PD−1 checkpoint pathway in cancer
Ets1 mTOR signaling pathway 3
Sp1 T cell receptor signaling pathway
Toxoplasmosis 6
Srebf1
Trp53 Spliceosome 9

5
Hnf1a 12
Thermogenesis
Hnf4a
Tight junction 15
Nfkb2 Apoptosis
Sox10 Pathogenic Escherichia coli infection
Stat5a Shigellosis
Stat5b

6
Viral myocarditis
Erg Proteoglycans in cancer
Klf4 Neutrophil extracellular trap formation
Etv4 Leukocyte transendothelial migration
Twist1 Salmonella infection
Nfatc2
Rxra RNA transport

7
mRNA surveillance pathway
Arnt
Nr5a1 1 2 3 4 5
Rfx5 Fold Enrichment
Sox2 Neurotrophin signaling pathway
Nfe2l2 Relative Kaposi sarcoma−associated herpesvirus infection
Pax6 TF activity Human cytomegalovirus infection
Pou2f1 Gastric acid secretion
Creb1 10 Rap1 signaling pathway
Ets2

8
Human immunodeficiency virus 1 infection
Atf4 5 Fluid shear stress and atherosclerosis
Epas1 0 Glucagon signaling pathway
Jund Oxytocin signaling pathway
Smad3 -5 Dopaminergic synapse
Bach1 JAK−STAT signaling pathway
Fosl1 -10 Autophagy − animal − log 10 ( p)
Cellular senescence

9
12.5
Spinocerebellar ataxia
Autophagy − other 10.0
c Viral carcinogenesis 7.5
Basal respiration Maximal respiration Longevity regulating pathway

10
150 250 P = 0.001
Circadian rhythm 5.0
P = 0.001
OCR (pmol/min)
OCR (pmol/min)

200 Cell cycle 2.5

100 Chronic myeloid leukemia

11
150 ErbB signaling pathway
# genes
50 100 Ribosome

12
Coronavirus disease − COVID−19 3
50
Citrate cycle (TCA cycle) 6
0 0 HIF−1 signaling pathway
9

13
Pyruvate metabolism
Glycolysis Glycolytic capacity Central carbon metabolism in cancer 12
150 200 Ctrl Glyoxylate and dicarboxylate metabolism 15

14
P = 0.001 P = 0.005 One carbon pool by folate
ECAR (mpH/min)

GC
120
ECAR (mpH/min)

150 LPS Proteasome

15 16
90 GC + LPS
100 Transcriptional misregulation in cancer
60 Ubiquitin mediated proteolysis
50
17
30 Protein processing in endoplasmic reticulum

0 0 Histidine metabolism
18 19

TGF−beta signaling pathway

1 2 3 4 5
Fold Enrichment
d Cells Single cells
MitoTracker Red
SSC-A
SSC

FSC SSC-H MitoTracker Green

Ctrl GC LPS GC + LPS
MitoTracker Red(FI)

MitoTracker Green (FI)

Extended Data Fig. 1 | Glucocorticoid-mediated control of transcription stress test (n = 4) and a glycolysis stress test (n = 8) of BMDMs treated with Ctrl,
and metabolism. (a) Heatmap illustrating the predicted activity of indicated GC and LPS for 24 h. Oxygen consumption rate (OCR) and extracellular
transcription factors (TF) from the bulk mRNA sequencing data of bone marrow- acidification rate (ECAR) were measured. (d) Gating strategy of the flow
derived macrophages (BMDMs) treated with vehicle (Ctrl), dexamethasone cytometry analysis of BMDMs treated with Ctrl, GC and LPS for 6 or 24 h and
(GC, 100 nM) and LPS (100 ng/ml) for 4 or 24 h. (b) Pathway analysis derived stained with MitoTracker Green and Mitotracker Red (corresponding to
from the bulk mRNA sequencing data illustrating pathways enriched in BMDMs data in Fig. 1d and Fig. 2b). Data are presented as mean + SEM. One-sided
treated with a combination of GC and LPS in comparison to BMDMs treated Hypergeometric test with Bonferroni’s correction (b); one-way ANOVA with
with LPS only. (c) Quantification of indicated parameters derived from a mito Tukey’s multiple comparison test (c).
 a Lactate Citrate Itaconate -Ketoglutarate
100 100 100

Rel. exchange rate (%)

100
Rel. exchange rate (%)

Rel. exchange rate (%)

Rel. exchange rate (%)

M3 M6 M5 M5
M2 M5 M4 M4
M1 M4 M3 M3
50 M0 50 M3 50 M2 50 M2
M2 M1 M1
M1 M0 M0
M0
0 0 0 0
C

C
trl

trl

trl

C
S

trl

S
S
S

S
G LP

G LP

G LP

G LP
LP

LP

LP

LP
G

G
C

C
+
+

+
C

C
Succinate Fumarate Malate
100 100

Rel. exchange rate (%)

Rel. exchange rate (%)

100 M4 M4
Rel. exchange rate (%)

M4
M3 M3 M3
M2 M2 M2
M1 M1 M1
50 50 M0 50 M0
M0

0 0 0
C

C
trl

trl
S

S
S

S
C
trl

G LP

G LP
LP

LP
G

G
C

C
G LP
LP
G
C

+
+

C
C

b Glycolysis Glycolytic Capacity

150 Glucose Oligo 2-DG +/+ mut 60 P = 0.040 P = 0.035 80 P = 0.003 P = 0.004
Nls Nls Ctrl
GC
ECAR (mpH/min)

ECAR (mpH/min)
ECAR (mpH/min)
Ctrl 60
100 40 LPS
GC GC + LPS
40
50 LPS 20
20
GC + LPS
0 0 0
0 20 40 60 80 Nls+/+ Nlsmut Nls+/+ Nlsmut
Time (minutes)
c e
MitoTracker Red GC receptor Merge MitoTracker Red GC receptor Merge

S
C LPS

LP
LP
(kDa)

+
+

S
S

trl
C

C
trl

trl
C

LP
~55

LP

LP

G
C
G

G
C
PDH
Ctrl GC ~55
ACTB

6h 12 h 24 h

LPS GC + LPS

d
6h
Ctrl

GC
GC + LPS
LPS

PDH-X Phalloidin PDH-X/Mitotracker PDH-X Phalloidin PDH-X/Mitotracker

Merge Merge
Merge MTRed DAPI Merge MTRed DAPI
PDH-X/Mitotracker PDH-X/Mitotracker
Surface rendering Surface rendering

Extended Data Fig. 2 | Regulation of TCA cycle metabolism by glucocorticoids. for 24 h and stained with MitoTracker Red and an antibody directed against
(a) Mass spectrometry-based analysis of U-13C glucose carbon tracing in the glucocorticoid receptor. (d) Immunofluorescence microscopy of PDH-X
bone marrow-derived macrophages (BMDMs) treated with vehicle (Ctrl), protein in BMDMs upon stimulation with Ctrl, GC and LPS for 6 h. Fluorescence
dexamethasone (GC, 100 nM) and LPS (100 ng/ml). Relative U-13C exchange rate signal rendering for PDH-X and MitoTracker is represented in the lower right
(RelER) for the indicated metabolites and the number of exchanged carbon panel, with quantification in Fig. 2g. (e) Western blot analysis of pyruvate
atoms (M0 to M6). (b) Extracellular acidification rate (ECAR) of a glycolysis stress dehydrogenase (PDH) protein levels in BMDMs treated with Ctrl, GC, 100 nM and
test, with quantified glycolysis and glycolytic capacity of fetal liver-derived LPS for 6, 12 and 24 h. Data are presented as mean + SEM. One-way ANOVA with
macrophages from Nls+/+ and Nlsmut mice treated with vehicle, GC and LPS for 24 h Tukey’s multiple comparison test (b).
(n = 8). (c) Immunofluorescence imaging of BMDMs treated with Ctrl, GC and LPS
Article
a
P = 0.0001 P = 0.003 P = 0.044 P = 0.034
0.25 P = 0.0001 3 4 Ctrl
P = 0.0001 P = 0.007 P = 0.042
0.20 GC
3
IL-1β (ng/ml)

TNF (ng/ml)
LPS

IL-6 (ng/ml)
2
0.15 GC + LPS
2
0.10
1
1
0.05
0.00 0 0
GC - + - + - + - + GC - + - + - + - + GC - + - + - + - +
LPS - - + + - - + + LPS - - + + - - + + LPS - - + + - - ++
Normoxia 1% O2 Normoxia 1% O2 Normoxia 1% O2

b Basal respiration Maximal respiration Glycolysis Glycolytic capacity

P = 0.0001
60 100 80 P = 0.0001 150 Ctrl
P = 0.005
OCR (pmol/min)

ECAR (mpH/min)

ECAR (mpH/min)
OCR (pmol/min)

80 P = 0.0001 GC
60 LPS
40 100
60 GC + LPS
40
40
20 50
20 20

0 0 0 0
Veh RXD Veh RXD Veh RXD Veh RXD

c
5 P = 0.002 50 P = 0.0001 P = 0.036 10 P = 0.001
Ctrl
4
P = 0.004
40 P = 0.025 8 GC
IL-1β (ng/ml)

LPS
TNF (ng/ml)
IL-6 (ng/ml)

3 30 6 GC + LPS
2 20 4
1 10 2
0 0 0
Veh UK5099 Veh UK5099 Veh UK5099

Extended Data Fig. 3 | Parallel regulation of metabolism and inflammation derived macrophages (BMDMs) treated with Ctrl, GC and LPS in the presence of
by glucocorticoids. (a) ELISA-based quantification of the indicated cytokines a vehicle or roxadustat (RXD) (10 µM) for 24 h (n = 8). Oxygen consumption rate
in the supernatants human monocytes-derived macrophages from healthy (OCR) and extracellular acidification rate (ECAR) were measured. (c) ELISA-
donors under normoxia and hypoxia (1% O2) and treated as indicated with based quantification of the indicated cytokines in the supernatants of BMDMs
vehicle (Ctrl), dexamethasone (GC, 100 nM) and LPS and IFN-γ (100 ng/ml and treated with Ctrl, GC and LPS in the presence of a vehicle or UK5099 (10 µM) as
20 ng/ml, respectively) (n = 12). (b) Quantification of indicated parameters indicated (n = 6). Data are presented as mean + SEM. One-way ANOVA with
derived from a mito stress test and a glycolysis stress test of bone marrow- Tukey’s multiple comparison test (a-c).
 a b c
LPS LPS
GC + LPS GC + LPS Acod1+/+ Acod1-/-
DMI + LPS 4OI + LPS Ctrl
E2f4 GC
P = 0.002 P = 0.012 Myc LPS
5 4 Hif1a
P = 0.0001 P = 0.0001 P = 0.0001 Nfkb1 GC + LPS
4 P = 0.0001 Irf9
3

IL-1β (ng/ml)
IL-1β (ng/ml)

Stat2
3 Stat1
2 Irf1
2 Spi1
Rel
1 1 Epas1
Foxl2
Fos
0 0 Jun
Acod1+/+ Acod1-/- Acod1+/+ Acod1-/- Rela
Srebf1
Srebf2
Zfp263
P = 0.037 P = 0.003 Atf6
P = 0.006 Nr3c1
P = 0.005 Foxo3
16 P = 0.0001 20 P = 0.011 P = 0.031 Gata3
Atf1
P = 0.035 15 Sp3
12 Egr1
IL-6 (ng/ml)
IL-6 (ng/ml)

Nfya
8 10 Tcf7l2
Esr1
Prdm14
4 5 Vdr
Elk1
Klf4
0 0 Creb1
Acod1+/+ Acod1-/- Acod1+/+ Acod1-/- Etv4
Trp63
Rxra
Smad4
P = 0.025 Erg Relative
8 P = 0.046 15 P = 0.003
Ets1 TF activity
Sp1
P = 0.0001 P = 0.015 Ets2 10
P = 0.011 P = 0.004 Pax6
6
TNF (ng/ml)
TNF (ng/ml)

10 Fosl2
Smad3 5
4 Trp53
Nfe2l2 0
5 Stat3
2 Jund -5
Atf4
0 0
Fosl1 -10
Acod1+/+ Acod1-/- Acod1+/+ Acod1-/-

d P = 0.0001 P = 0.002 f
1.5 P = 0.0001 P = 0.016 15
P = 0.004 5 P = 0.043 Ctrl

+
2 -/-

2 -/-
2 +/

2 +/
P = 0.012 P = 0.014
4 GC

2l

2l

2l

2l
IL-1β (ng/ml)

fe

fe

fe

fe
TNF (ng/ml)
IL-6 (ng/ml)

1.0 10 LPS (kDa)

N
3 GC + LPS
2 ~100 NRF2
0.5 5
~55
1 ACTB
0.0 0 0
GC - + - + - + - + GC - + - + - + - + GC - + - + - + - +
LPS - - ++ - - ++ LPS - - ++ - - ++ LPS - - ++ - - ++
Veh ERG344 Veh ERG344 Veh ERG344

e g
P = 0.0001 P = 0.001 P = 0.0001 P = 0.001 P = 0.0001 P = 0.0001
P = 0.036
2.0 8 15 P = 0.025 Ctrl

PS
P = 0.002
GC

+L
S
1.5 6
trl

C
LPS
LP

(kDa)
IL-1β (ng/ml)

TNF (ng/ml)
IL-6 (ng/ml)

G
C

10
GC + LPS
1.0 4 ~100 NRF2
5 ~55
0.5 2 ACTB
0.0 0 0
sgRen sgAcod1 sgRen sgAcod1 sgRen sgAcod1

h P = 0.0001 P = 0.028
2.5 P = 0.0001 8 P = 0.0001 4 P = 0.0001 P = 0.002
Ctrl
P = 0.0001 P = 0.0001
2.0
6 GC
3
IL-1 (ng/ml)

TNF (ng/ml)
IL-6 (ng/ml)

LPS
1.5
4 2
GC + LPS
1.0
0.5 2 1

0.0 0 0
+/+ Nfe2l2-/- +/+ +/+
Nfe2l2 Nfe2l2 Nfe2l2-/- Nfe2l2 Nfe2l2-/-

Extended Data Fig. 4 | See next page for caption.

Article
Extended Data Fig. 4 | Itaconate as mediator of the action of glucocorticoids. (e) ELISA-based quantification of the indicated cytokines in the supernatants
(a,b) ELISA-based quantification of the indicated cytokines in the supernatants of human THP-1 monocyte-derived sgRen and sgAcod1 macrophages treated as
of bone marrow-derived macrophages (BMDMs) from Acod1+/+ and Acod1−/− indicated with Ctrl, GC and LPS and IFN-γ (n = 6). Similar results were observed
mice treated as indicated with vehicle (Ctrl), dexamethasone (GC, 100 nM), LPS with two other sgRNAs targeting Acod1. (f) Western blot analysis of NRF2 protein
(100 ng/ml) and (a) dimethyl itaconate (DMI; 62.5 µM) or (b) 4-octyl itaconate level in Nfe2l2+/+ and Nfe2l2−/− BMDMs. Six hours prior to collection, cells were
(4OI; 62.5 µM) (n = 7 for a, n = 6 for b). (c) Heatmap illustrating the predicted treated with 2 µM MG-132 to promote NRF2 stabilization. (g) Western blot
activity of indicated transcription factors (TF) derived from bulk mRNA analysis of NRF2 protein level in BMDMs treated with Ctrl, GC and LPS for 24 h.
sequencing data of BMDMs from Acod1+/+ and Acod1−/− mice treated with Ctrl, Six hours prior to collection, cells were treated with 2 µM MG-132 to promote
GC and LPS for 24 h. (d) ELISA-based quantification of the indicated cytokines NRF2 stabilization. (h) ELISA-based quantification of the indicated cytokines in
in the supernatants of human monocytes-derived macrophages from healthy the supernatants of BMDMs from Nfe2l2+/+ and Nfe2l2−/− mice treated as indicated
donors treated as indicated with Ctrl, GC and LPS and IFN-γ (20 ng/ml) in the with Ctrl, GC and LPS (n = 6). Data are presented as mean + SEM. One-way ANOVA
presence of a vehicle or the ACOD1 inhibitor ERG344 (500 µM) (n = 12). with Tukey’s multiple comparison test (a-b, d-e and h).
 a i.p. i.n. Histology Flow Cytometry
GC LPS BALF
ELISA

0h 1h 24 h

b Cells Single cells CD45+ CD11b+

Myeloid cells PMN

SSC-A

CD45

Ly6G
SSC

FSC SSC-H CD11b FSC

c LPS GC + LPS d LPS GC + LPS

Acod1+/+ Acod1+/+

100 µm 100 µm
CD45 (FI)

Acod1-/- Acod1-/-

100 µm 100 µm

CD11b (FI)

e i.p. GC treatment
i.p.
Histology
K/BxN serum

d0 d4 d5 d6 d7 d8 d9 d10 d11 d12 d13 d14

Scoring

f i.n. GC at t=0 min

i.p. i.p. i.n. OVA at t=30 min
OVA OVA Histology Flow Cytometry
Alum Alum BALF
ELISA

d0 d7 d13 d14 d15 d16

g Cells Single cells CD45+ CD11b+

Myeloid
cells
SSC-A

CD64
CD45
SSC

FSC SSC-H CD11b FSC

CD45+ CD64-

Eosinophils
CD4+
SiglecF
CD4

CD3 FSC

Extended Data Fig. 5 | Murine models of immune-mediated diseases. mice during LPS-induced lung injury (scale bars = 100 µm). (e) Timeline of the
(a) Timeline of the treatment protocol used for the mouse model of LPS-induced used treatment protocol for the mouse model of K/BxN serum transfer arthritis.
lung injury. (b-c) Gating strategy and representative plots of the flow cytometry (f) Timeline of the treatment protocol used for the mouse model of ovalbumin
analysis of cells isolated from the bronchoalveolar lavage fluid (BALF) of Acod1+/+ (OVA)-induced allergic airway inflammation. (g) Gating strategy of the flow
and Acod1−/− mice during LPS-induced lung injury (corresponding to data in cytometry analysis of cells derived from the BALF of mice during ovalbumin-
Fig. 5a). (d) Representative H&E stainings of lung sections of Acod1+/+ and Acod1−/− induced allergic airway inflammation (corresponding to data in Fig. 5i).
Article
a
LPS response Glucocorticoid treatment

GC

LPS LPS
TLR4 TLR4
PDH
NF-B GR TCA NF-B PDH TCA
Itaconate Itaconate

pro-inflammatory pro-inflammatory
genes genes GR
TCA cycle activity

TCA cycle activity

GC
+
Itaconate

Itaconate
LPS LPS

Time Time

Acod1 Expression Acod1 Expression

Extended Data Fig. 6 | Metabolic rewiring promotes anti-inflammatory effects of glucocorticoids. (a) Graphical summary of the results presented in the
manuscript.
 nature portfolio | reporting summary
Corresponding author(s): Gerhard Krönke
Last updated by author(s): 2024/02/16

Reporting Summary
Nature Portfolio wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency
in reporting. For further information on Nature Portfolio policies, see our Editorial Policies and the Editorial Policy Checklist.

Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section.
n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement
A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly
The statistical test(s) used AND whether they are one- or two-sided
Only common tests should be described solely by name; describe more complex techniques in the Methods section.

A description of all covariates tested

A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons
A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient)
AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals)

For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted
Give P values as exact values whenever suitable.

For Bayesian analysis, information on the choice of priors and Markov chain Monte Carlo settings
For hierarchical and complex designs, identification of the appropriate level for tests and full reporting of outcomes
Estimates of effect sizes (e.g. Cohen's d, Pearson's r), indicating how they were calculated
Our web collection on statistics for biologists contains articles on many of the points above.

Software and code

Policy information about availability of computer code
Data collection Data were collected using the following instruments, as detailed in the methods: Leica TCS SP8 confocal laser scanning microscope with Las AF
software v2.7.3.9723 (Leica); ImageXpress MicroConfocal microscope (Molecular Devices); Keyence BZ-X710 microscope using BZ-X View
series 2:22 (Keyence); ICS3000 HPLC-system (Dionex) and ESI/MS/MS detection using a QTrap3200 Triple-Quadrupole mass spectrometer
with turbo V ion source (Applied Biosystems); Dionex Ultimate 3000 chromatographic system hyphenated to a Q Exactive Focus mass
spectrometer (Thermo Fisher Scientific); CytoFLex S system (Beckman Coulter); XFe96 Extracellular Flux Analyzer with Wave v2.4.2 (Seahorse
Bioscience, Agilent Technology); Quantstudio 6 Flex with QuantStudio Real Time Systems LightCycler (ThermoFisher); SpectraMax190 with
SoftMax Pro 7.1 software (Molecular Devices).

Data analysis Data were analyzed using the following softwares, as detailed in the methods: Huygens Profession v22.10 (Scientific Volume Imaging B.V.) and
Imaris software v9.7.1. (Bitplane); BZ-X Analyzer software series 2:22 (Keyence); ImageJ 1.52a (Fiji); CytExpert v2.4 (Beckman Coulter) and
FlowJo v10.6.1; Compound Discoverer 3.1 (Thermo Fisher Scientific); GraphPad Prism v8.3. Differential expression analysis was performed
with the DESeq2 package v1.36.0, transcription factor activity prediction was performed using the dorothea package v1.8.0. and pathway
enrichment analysis was carried out using the pathfinder package v1.6.3. Additionally, plots were generated with the ggplot2 package v3.3.6
and the ComplexHeatmap package v2.12.0. in R software; STAR aligner v2.5.1b, fastqc v0.11.7 and featureCounts v2.0.0, R programming
language v.3.6.3. Data analysis strategies are described in the methods and extended figures.
April 2023

For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors and
reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Portfolio guidelines for submitting code & software for further information.

1
 nature portfolio | reporting summary
Data
Policy information about availability of data
All manuscripts must include a data availability statement. This statement should provide the following information, where applicable:
- Accession codes, unique identifiers, or web links for publicly available datasets
- A description of any restrictions on data availability
- For clinical datasets or third party data, please ensure that the statement adheres to our policy

The data supporting the findings of this study are available upon reasonable request from G.K. or J-P.A.. The data are not publicly available due to comprised
information that could compromise research participant privacy, though source data are provided with this paper. Bulk RNA sequencing datasets generated in this
study are available on the GEO Platform under the accession numbers GSE250273 and GSE250274. The following databases were used in this study: the Broad
Institute sgRNA design tool (https://portals.broadinstitute.org/gppx/crispick/public), Ensembl mouse reference genome GRCm38 (http://www.ensembl.org/
Mus_musculus/Info/Index), mzcloudTM and mzVault (https://www.mzcloud.org/), and ChemSpider (http://www.chemspider.com/).

Research involving human participants, their data, or biological material

Policy information about studies with human participants or human data. See also policy information about sex, gender (identity/presentation),
and sexual orientation and race, ethnicity and racism.
Reporting on sex and gender Findings equally apply to both sexes. Therefore sex was not considered in the study design and samples from both sexes
were analyzed together. The individual-level data could compromise research participant privacy and cannot be detailed.
Samples from 35 patients were included in this study.

Reporting on race, ethnicity, or Reporting on race, ethnicity or other socially relevant groupings does not apply to this study.
other socially relevant
groupings

Population characteristics No healthy donors were excluded from this study. Rheumatoid arthritis patients fulfilled the 2010 ACR/EULAR criteria for
rheumatoid arthritis and were selected based on the absence or presence of an ongoing glucocorticoid therapy. All donors
and patients were ≥ 18 years of age.

Recruitment Healthy donors and rheumatoid arthritis patients were recruited by the Department of Internal Medicine 3 - Rheumatology
and Immunology, Universitätsklinikum Erlangen. Informed consent was obtained from all participants. Healthy donors were
selected based on their willingness to provide blood and the absence of immunomodulatory drug intake. Rheumatoid
arthritis patients were selected based on the absence or presence of an ongoing glucocorticoid therapy (equivalent to 25 mg
to 50 mg prednisolone per day). Patients were receiving treatment at the Universitätsklinikum Erlangen and were
approached for participation by the acting physician.

Ethics oversight Approval was given by the ethics committee of the University Hospital Erlangen

Note that full information on the approval of the study protocol must also be provided in the manuscript.

Field-specific reporting
Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection.

Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences
For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf

Life sciences study design

All studies must disclose on these points even when the disclosure is negative.
Sample size Sample size was chosen based on prior experience in our working group and published studies with similar layouts (Faas et al., 2021;
Culemann & Grüneboom et al., 2019; Rothe, Ipseiz & Faas et al., 2017).

Data exclusions No data were excluded from the analysis.

Replication The number of replicates for each experiment is indicated in figure legends. Within each experimental group, the reproducibility was
successful, although a degree of variability was detected due to inter- individual diversity.

Randomization Experimental groups were randomly allocated. Treated groups were housed together with control groups.
April 2023

Blinding The investigators were blinded to group allocation during data collection and data analysis was performed blindly to experimental conditions.

2
 nature portfolio | reporting summary
Reporting for specific materials, systems and methods
We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material,
system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response.

Materials & experimental systems Methods

n/a Involved in the study n/a Involved in the study
Antibodies ChIP-seq
Eukaryotic cell lines Flow cytometry
Palaeontology and archaeology MRI-based neuroimaging
Animals and other organisms
Clinical data
Dual use research of concern
Plants

Antibodies
Antibodies used Antibodies used in this study are:
- CD45-BrilliantViolet 510 (clone 30-F11, BioLegend #103133)
- CD11b-PE (clone M1/70, BioLegend #101207)
- Ly6G-Alexa Fluor 700 (clone HK1.4A8, Biolegend 127621)
- CD64-Alexa Fluor A647 (clone X54-5/7.1, BioLegend 139321)
- SiglecF- FITC (clone S1700SL, BioLegend 155504)
- CD3-PE/Cy7 (clone 17A2, BioLegend 100219)
- CD4-APC (clone GK1.5, BioLegend 100411)
- TruStain FcX PLUS (anti-CD16/32) (clone S17011E, BioLegend 156603)
- Rabbit anti-HIF-1α (Novus Biologicals NB100-449)
- Rabbit anti-ACOD1 (Cell Signaling Technology 17805)
- Rabbit anti-NRF2 (Cell Signaling Technology 12721)
- Rabbit anti-PDH (Cell Signaling Technology 3205)
- Rabbit anti-glucocorticoid receptor (Cell Signaling Technology 12041)
- Rabbit anti-glucocorticoid receptor (Cell Signaling Technology 3660)
- Normal rabbit IgG (Cell Signaling Technology 2729)
- Rabbit anti-β-actin (Sigma A2066)
- Mouse anti-PDHE1a (clone D-6, Santa Cruz sc-377092)
- Rabbit anti-PDHX (Proteintech 10951-1-AP)
- Phalloidin-Alexa Fluor 488 (Invitrogen A12379)
- Anti-rabbit IgG Alexa Fluor 488 (Invitrogen A-11008)
- Anti-rabbit IgG Alexa Fluor 647 (Invitrogen A-21245)

Validation All antibody used are commercially available. Antibody specificity and quality validation were performed by the manufacturers,
including for the species and application. Validation statements of the manufacturers can be found on their webpages (direct testing
or by providing adequate references).
- CD45-BrilliantViolet 510 (clone 30-F11, BioLegend #103133): Verified reactivity against mouse; valided for flow cytometry
https://www.biolegend.com/fr-lu/products/brilliant-violet-510-anti-mouse-cd45-antibody-7995?GroupID=BLG1932
- CD11b-PE (clone M1/70, BioLegend #101207): Verified reactivity against mouse; valided for flow cytometry
https://www.biolegend.com/fr-lu/products/pe-anti-mouse-human-cd11b-antibody-349
- Ly6G-Alexa Fluor 700 (clone HK1.4A8, Biolegend 127621): Verified reactivity against mouse; valided for flow cytometry
https://www.biolegend.com/fr-lu/products/alexa-fluor-700-anti-mouse-ly-6g-antibody-6754
- CD64-Alexa Fluor A647 (clone X54-5/7.1, BioLegend 139321): Verified reactivity against mouse; valided for flow cytometry
https://www.biolegend.com/fr-lu/products/alexa-fluor-647-anti-mouse-cd64-fcgammari-antibody-12425
- SiglecF- FITC (clone S1700SL, BioLegend 155504): Verified reactivity against mouse; valided for flow cytometry
https://www.biolegend.com/fr-lu/products/fitc-anti-mouse-cd170-siglec-f-antibody-16371
- CD3-PE/Cy7 (clone 17A2, BioLegend 100219): Verified reactivity against mouse; valided for flow cytometry
https://www.biolegend.com/fr-lu/products/pe-cyanine7-anti-mouse-cd3-antibody-6060
- CD4-APC (clone GK1.5, BioLegend 100411): Verified reactivity against mouse; valided for flow cytometry
https://www.biolegend.com/fr-lu/products/apc-anti-mouse-cd4-antibody-245
- TruStain FcX PLUS (anti-CD16/32) (clone S17011E, BioLegend 156603): Verified reactivity against mouse; valided for flow cytometry
https://www.biolegend.com/fr-lu/products/trustain-fcx-plus-anti-mouse-cd16-32-antibody-17085
- Rabbit anti-HIF-1α (Novus Biologicals NB100-449): Verified reactivity against mouse; valided for Western blotting
https://www.novusbio.com/products/hif-1-alpha-antibody_nb100-449
- Rabbit anti-ACOD1 (Cell Signaling Technology 17805): Verified reactivity against mouse; valided for Western blotting
April 2023

https://www.cellsignal.com/products/primary-antibodies/irg1-antibody/17805
- Rabbit anti-NRF2 (Cell Signaling Technology 12721): Verified reactivity against mouse; valided for Western blotting
https://www.cellsignal.com/products/primary-antibodies/nrf2-d1z9c-xp-rabbit-mab/12721
- Rabbit anti-PDH (Cell Signaling Technology 3205); Verified reactivity against mouse; valided for Western blotting
https://www.cellsignal.com/products/primary-antibodies/pyruvate-dehydrogenase-c54g1-rabbit-mab/3205

3
 - Rabbit anti-glucocorticoid receptor (Cell Signaling Technology 12041) : Verified reactivity against mouse; valided for Western
blotting, co-immunoprecipitation and immunofluorescence

nature portfolio | reporting summary

https://www.cellsignal.com/products/primary-antibodies/glucocorticoid-receptor-d6h2l-xp-rabbit-mab/12041
- Rabbit anti-glucocorticoid receptor (Cell Signaling Technology 3660): Verified reactivity against mouse; valided for
immunofluorescence
https://www.cellsignal.com/products/primary-antibodies/glucocorticoid-receptor-d8h2-xp-rabbit-mab/3660
- Normal rabbit IgG (Cell Signaling Technology 2729): Verified reactivity against mouse; valided for co-immunoprecipitation
https://www.cellsignal.com/products/primary-antibodies/normal-rabbit-igg/2729
- Rabbit anti-β-actin (Sigma A2066): Verified reactivity against mouse; valided for Western blotting
https://www.sigmaaldrich.com/DE/en/product/sigma/a2066
- Mouse anti-PDHE1a (clone D-6, Santa Cruz sc-377092): Verified reactivity against mouse; valided for immunofluorescence
https://www.scbt.com/p/pdh-e1alpha-antibody-d-6
- Rabbit anti-PDHX (Proteintech 10951-1-AP): Verified reactivity against mouse; valided for immunofluorescence
https://www.ptglab.com/products/PDHX-Antibody-10951-1-AP.htm
- Phalloidin-Alexa Fluor 488 (Invitrogen A12379): Verified reactivity against mouse; valided for immunofluorescence
https://www.thermofisher.com/order/catalog/product/de/de/A12379

Eukaryotic cell lines

Policy information about cell lines and Sex and Gender in Research
Cell line source(s) HEK293T/17 from ATCC (#CRL-11268)
THP-1 from ATCC (#TIB-202)
L929 from ATCC (#CCL-1)

Authentication None of the cell lines were authenticated.

Mycoplasma contamination Cell lines were not tested for mycoplasma contamination.

Commonly misidentified lines No commonly misidentified lines were used in this study.
(See ICLAC register)

Animals and other research organisms

Policy information about studies involving animals; ARRIVE guidelines recommended for reporting animal research, and Sex and Gender in
Research

Laboratory animals In this study the following mouse lines were used:

- C57BL/6 (strain C57BL/6JRj; Janvier Labs)

- Acod1 (strain C57BL/6NJ-Acod1em1(IMPC)J/J; The Jackson Laboratory #029340)
- Nfe2le (strain B6.129X1-Nfe2l2tm1Ywk/J; The Jackson Laboratory #017009)
- Nls1mut (mutantion in the sequence of the glucocorticoid receptor (NR3C1) at position 510-512, which results in the exchange of
three lysins to three asparagines, generated by J. Tuckermann; 10.1128/MCB.19.2.1025)

Animals were housed and bred at the animal facilities of the Friedrich-Alexander University of Erlangen-Nuremberg under specific-
pathogen-free conditions. Mice were maintained on a 12 h/12 h light/dark cycle, at around 20-24 °C with 45-65% humidity with ad
libitum access to food and water. Male and female mice aged 6-12 weeks were used.

Wild animals No wild animals were used in this study.

Reporting on sex We used age and sex-matched males and females.

Field-collected samples No field-collected samples were used in this study.

Ethics oversight Animal experiments were approved by the local animal ethics committees of the Regierung von Mittelfranken and the
Regierungspräsidium Karlsruhe and Tübingen.
Note that full information on the approval of the study protocol must also be provided in the manuscript.
April 2023

4
 nature portfolio | reporting summary
Plants
Seed stocks Report on the source of all seed stocks or other plant material used. If applicable, state the seed stock centre and catalogue number. If
plant specimens were collected from the field, describe the collection location, date and sampling procedures.

Novel plant genotypes Describe the methods by which all novel plant genotypes were produced. This includes those generated by transgenic approaches,
gene editing, chemical/radiation-based mutagenesis and hybridization. For transgenic lines, describe the transformation method, the
number of independent lines analyzed and the generation upon which experiments were performed. For gene-edited lines, describe
the editor used, the endogenous sequence targeted for editing, the targeting guide RNA sequence (if applicable) and how the editor
was applied.
Authentication Describe any authentication procedures for each seed stock used or novel genotype generated. Describe any experiments used to
assess the effect of a mutation and, where applicable, how potential secondary effects (e.g. second site T-DNA insertions, mosiacism,
off-target gene editing) were examined.

Flow Cytometry
Plots
Confirm that:
The axis labels state the marker and fluorochrome used (e.g. CD4-FITC).
The axis scales are clearly visible. Include numbers along axes only for bottom left plot of group (a 'group' is an analysis of identical markers).
All plots are contour plots with outliers or pseudocolor plots.
A numerical value for number of cells or percentage (with statistics) is provided.

Methodology
Sample preparation Murine bone marrow-derived macrophages were treated in culture plates for the required times before being stained with
50 nM MitoTracker Red CMXRos and 100 nM MitoTracker Green in medium for 30 min at 37°C with 5% CO2. Cells were then
washed with PBS, detached using Cellstripper® and resuspended in 2% heat-inactivated FCS in PBS.

Bronchoalveolar lavage fluid was recovered from treated mice, centrifuged and the cells were blocked with 10% rat serum
and 1% Fc Block in PBS for 10 min at room temperature and stained with BV510 anti-mouse CD45 clone 30-F11, PE anti-
mouse CD11b clone M1/70, A700 anti-mouse Ly6G clone HK1.4, A647 anti-mouse CD64 clone X54-5/7.1, FITC anti-mouse
SiglecF clone S1700SL, PE/Cy7 anti-mouse CD3 clone 17A2 or APC anti-mouse CD4 clone GK1.5 fluorophore-conjugated
antibodies for 30 min at 4°C. After washing, cells were resuspended in 2% heat-inactivated FCS in PBS.

Instrument Flow cytometry was performed with a CytoFLex S (Beckman Coulter).

Software Flow cytometry data and cell sorting data were analyzed using the CytExpert and FlowJo softwares.

Cell population abundance This study does not include sorted cells.

Gating strategy Gate boundaries for positive and negative populations were defined based on single color controls and fluorescence minus
one controls. Specific gating strategies are provided in detail in the extended data figures.
- Gating strategy of MitoTrackerRed and MitoTrackerGreen BMDMs (Extended Data Fig. 1d): Singlets (SSC-H vs. SSC-A were
gated on cells to exclude debris (FCS vs. SSC). MitoTrackerRed and MitoTrackerGreen+ populations were then defined and
gated.
- Gating strategy of myeloid cells and PMNs in LPS-induced lung injury model (Extended Data Fig. 5b):Singlets (SSC-H vs. SSC-
A) were gated on cells to exclude debris (FCS vs. SSC). CD45+CD11b+ cells (defined as myeloid cells) were gated upon and
Ly6G+ cells were defined as PMNs.
- Gating strategy of CD4+ T cells, myeloid cells and eosinophils in ovalbumin-induced asthma model (Extended Data Fig. 5g):
Singlets (SSC-H vs. SSC-A) were gated on cells to exclude debris (FCS vs. SSC). CD45+CD11b+ cells (defined as myeloid cells)
were gated on and CD64-SiglecF+ cells were defined as eosinophils. CD4+ T cells were defined as being CD45+CD11b- cells
but CD3+CD4+ (gated).

Tick this box to confirm that a figure exemplifying the gating strategy is provided in the Supplementary Information.
April 2023