Embed Size (px)
Citation preview
LETTERdoi:10.1038/nature11984
Induction of pathogenic TH17 cells by induciblesalt-sensing kinase SGK1Chuan Wu1*, Nir Yosef1,2*, Theresa Thalhamer1{*, Chen Zhu1, Sheng Xiao1, Yasuhiro Kishi1, Aviv Regev2,3 & Vijay K. Kuchroo1,2
TH17 cells (interleukin-17 (IL-17)-producing helper T cells) arehighly proinflammatory cells that are critical for clearing extracel-lular pathogens and for inducing multiple autoimmune diseases1.IL-23 has a critical role in stabilizing and reinforcing the TH17phenotype by increasing expression of IL-23 receptor (IL-23R)and endowing TH17 cells with pathogenic effector functions2,3.However, the precise molecular mechanism by which IL-23 sus-tains the TH17 response and induces pathogenic effector functionshas not been elucidated. Here we used transcriptional profiling ofdeveloping TH17 cells to construct a model of their signalling net-work and nominate major nodes that regulate TH17 development.We identified serum glucocorticoid kinase 1 (SGK1), a serine/threonine kinase4, as an essential node downstream of IL-23 sig-nalling. SGK1 is critical for regulating IL-23R expression and sta-bilizing the TH17 cell phenotype by deactivation of mouse Foxo1, adirect repressor of IL-23R expression. SGK1 has been shown togovern Na1 transport and salt (NaCl) homeostasis in other cells5–8.We show here that a modest increase in salt concentration inducesSGK1 expression, promotes IL-23R expression and enhances TH17cell differentiation in vitro and in vivo, accelerating the develop-ment of autoimmunity. Loss of SGK1 abrogated Na1-mediatedTH17 differentiation in an IL-23-dependent manner. These datademonstrate that SGK1 has a critical role in the induction of patho-genic TH17 cells and provide a molecular insight into a mechanismby which an environmental factor such as a high salt diet triggersTH17 development and promotes tissue inflammation.
To determine the molecular mechanisms by which naive T cellsdevelop into effector TH17 cells, we measured genome-wide messengerRNA expression profiles using microarrays along 18 time points over72 h, following the in vitro exposure of naive T cells to TH17 polarizingconditions (transforming growth factor b1 (TGF-b1) with IL-6). Toexamine the role of IL-23 in TH17 development, we added IL-23 at thelate time points (48–72 h) and monitored the transcriptional responsein both wild-type and Il23r–/– cells. We ranked the genes according totheir extent of induction in cells treated with TGF-b1 and IL-6 (relativeto non-polarized activated T cells) and repression in Il23r–/– cells(relative to wild-type cells) (Methods, Fig. 1a and SupplementaryTable 1). Murine Sgk1 was one of the top ranking genes, whose tran-scriptional regulation was strongly associated with both IL-23R sig-nalling and TH17 cell differentiation (Fig. 1a). Quantitative polymerasechain reaction (qPCR) analysis showed that Sgk1 is induced at lowlevels by TGF-b1 (induced regulatory T (iTreg) cells), and not inducedin other T cell subsets (TH0, TH1, TH2). As expected, it is most highlyexpressed under TH17 differentiation conditions (Fig. 1b). Sgk1expression is strongly induced during the first 2 h after stimulationof naive T cells under TH17-polarizing conditions. This is followed by asharp decline by 10 h to a steady expression level that is still substan-tially higher than in the control population (Fig. 1c and SupplementaryFig. 1a). Furthermore, Sgk1 expression is specifically induced and
maintained by exposure to IL-23 (Fig. 1c and Supplementary Fig.1b). Although Il23r–/– T cells initially produce Sgk1 mRNA, they can-not sustain this expression (Supplementary Fig. 1b, c). Finally, thekinase activity of SGK1 is also significantly higher in TH17 cells thanin other T cell subsets (Supplementary Fig. 1d), and restimulation ofdifferentiated TH17 cells with IL-23 elevates further SGK1 kinase activ-ity (Supplementary Fig. 1e). Thus, IL-23 signalling is critical for main-taining Sgk1 expression during TH17 cell differentiation.
Network analysis of the transcriptional changes in Il23r–/– T cellsusing the ANAT software9 singled out SGK1 as a potential nodal pointdownstream of IL-23R signalling. Based on a curated database ofprotein–protein interactions (PPIs), we constructed a network modelthat connects known proteins of the IL-23R signalling pathway(Methods) to the transcription factors whose function is dysregulatedin Il23r–/– cells (Methods, Fig. 1d and Supplementary Fig. 1f). Weranked the network’s nodes based on a centrality measure, definedas the fraction of IL-23R-affected transcription factors downstreamof that node in the network (Methods and Supplementary Table 1).SGK1 was the highest-ranking node (Supplementary Fig. 1g), suggest-ing that it acts both as a transcriptional target of IL-23R signalling andas a kinase that may mediate the transcriptional effects of the pathway.
Using Sgk1–/– mice, we studied the impact of loss of SGK1 on TH17differentiation in vitro. We observed no abnormality of SGK1-deficient T cells during primary differentiation into TH17 cells(Fig. 1e). However, Sgk1–/– TH17 cells restimulated with IL-23 showedimpaired IL-17 production (Fig. 1e and Supplementary Fig. 2b).Memory Sgk1–/– T cells also showed a defect in IL-17 production uponIL-23 stimulation, but not under stimulation with TGF-b1 and IL-6(Supplementary Fig. 2a). To study the function of SGK1 specifically inIL-17-producing T cells that carry the CD4 antigen (CD41 T cells),we generated Il17f CreSgk1fl/fl mice in which SGK1 was deleted in cellsproducing IL-17F, enabling us to analyse the function of SGK1 in themaintenance of TH17 phenotype. Il17f CreSgk1fl/fl T cells also showedno defect in primary TH17 differentiation, but displayed reduced IL-17production when restimulated with IL-23 (Fig. 1f and SupplementaryFig. 2c). IL-23R expression was also significantly reduced in Sgk1–/– Tcells (Fig. 1g). Thus, loss of SGK1 does not affect primary TH17 dif-ferentiation, but profoundly affects their stability and IL-23R expres-sion. One possible explanation for the dispensability of SGK1 duringprimary TH17 differentiation is redundancy with other kinases, such asits homologue AKT10. However, SGK1 seems to be indispensable forIL-23R-dependent stability and maintenance of TH17 cells.
Microarray analysis of Sgk1–/– versus wild-type TH17 cells restimu-lated with IL-23 showed a significant overlap in differentially expres-sed genes with the Il23r–/– versus wild-type IL-23-restimulated TH17cell profiles, supporting further the functional relatedness of the SGK1and IL-23R pathways (Fisher exact test, P , 0.001) (Fig. 1h andSupplementary Fig. 2d). Consistent with this, genes downregulatedin Sgk1–/– cells are significantly enriched (Fisher exact test, P , 0.001)
*These authors contributed equally to this study.
1Center for Neurologic Diseases, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts 02115, USA. 2Broad Institute of MIT and Harvard, 7 Cambridge Center, Cambridge,Massachusetts 02142, USA. 3Howard Hughes Medical Institute, Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts 02140, USA. {Present address: Department ofMolecular Biology, University of Salzburg, Hellbrunnerstrasse 34, 5020 Salzburg, Austria.
2 5 A P R I L 2 0 1 3 | V O L 4 9 6 | N A T U R E | 5 1 3
Macmillan Publishers Limited. All rights reserved©2013
for genes that are upregulated in wild-type TH17 cells compared toother T cell subsets11 (Methods and Supplementary Fig. 2e). Selectedgenes were confirmed by qPCR analysis (Supplementary Fig. 2f). Genesfrom several other pathways are also enriched (over- or underex-pressed) (Supplementary Table 2), including cell cycle and prolifera-tion, which may be related to the known role of SGK1 as a regulator ofproliferation and apoptosis7,8,10. Although our analysis strongly associ-ates SGK1 with the TH17 program, genes important for developmentand function of other T cell subsets, such as Ifng, Tbx21 or Gata3 werealso dysregulated in Sgk1–/– cells, suggesting possible additional effectsof this kinase in other T cell subsets.
To determine the role of SGK1 in vivo, we immunized Cd4CreSgk1fl/fl
mice with myelin oligodendrocyte glycoprotein (MOG) peptide 35–55(MOG35-55) to induce experimental autoimmune encephalomyelitis(EAE). SGK1-deficient mice exhibited significantly reduced EAEincidence and severity. IL-17 production from infiltrated CD41 T cellsin different organs of SGK1-deficient mice was also reduced, whereasinterferon-c (IFN-c) levels were unaffected (Fig. 2a and Sup-plementary Fig. 3a). When we restimulated the isolated T cells fromimmunized mice with IL-23 in the presence of MOG35-55, the SGK1-deficient T cells also showed impaired IL-17 but normal IFN-c pro-duction (Supplementary Fig. 3b, c). Next, using Il23r gfp reporter mice,we observed reduced IL-23R–GFP (green fluorescent protein) expres-sion on infiltrating CD41 T cells in different organs of SGK1-deficientmice undergoing EAE (Supplementary Fig. 4a). Similar to the res-ponse of Cd4CreSgk1fl/fl mice, reduced TH17 differentiation and diseaseseverity were also observed in Il17f CreSgk1fl/fl mice during EAE(Fig. 2b). In addition, to exclude any effects of SGK1-deficientbystander cells, we transferred purified Il17f CreSgk1fl/fl CD41 T cellsinto Rag2–/– mice and induced EAE. Mice that received SGK1-deficient T cells developed less severe disease compared to mice thatreceived wild-type T cells (Supplementary Fig. 4b).
To determine the reason for fewer TH17 cells being found in SGK1-deficient mice, we transferred purified GFP1 cells from differentiatedCd4CreSgk1fl/flIl17agfp or control TH17 cells to congenic Ly5.1 mice andtraced the IL-17 GFP1 cells in different organs after immunizationwith MOG35-55 (Fig. 2c). Starting with the same number of CD41IL-171 T cells, we found that 7 and 12 days after transfer, SGK1-deficientTH17 cells failed to maintain IL-17 production, particularly in thecentral nervous system (CNS) (Fig. 2d and Supplementary Fig. 4c).Next, we crossed Il17f CreRosa26ReYFP mice onto the SGK1-deficientbackground, and analysed the expression of IL-17 in T cells that hadturned on the Il17f gene as determined by enhanced yellow fluorescentprotein (eYFP) expression. We induced EAE in these mice and analysedthe frequency of eYFP1 cells producing IL-17 in infiltrating CD41 Tcells in the lymph nodes and CNS. The Sgk1–/– reporter mice showed asmaller proportion of CD41eYFP1 T cells in both organs. Furthermore,there was a dramatic loss of IL-17 expression by eYFP1 T cells in theSGK1-deficient mice, indicating that TH17 cells could not stably retainIL-17 production during EAE (Fig. 2e and Supplementary Fig. 4d).
To understand better the molecular role of SGK1 in TH17 cells, weconducted another network analysis, using PPI data to connect SGK1to the transcription factors whose activity is dysregulated in Sgk1–/–
TH17 cells (Methods). The analysis suggested Foxo1 as one of thehighest-ranking nodes downstream of SGK1 (Fig. 3a, SupplementaryTable 1 and Supplementary Fig. 5a). Foxo1 phosphorylation by SGK1in adipocytes has been shown previously to lead to its deactivation andtranslocation from the nucleus to the cytoplasm12. Consistent with thisobservation, we found that SGK1 phosphorylates Foxo1 (Supplemen-tary Fig. 5b). Immunoblot analysis of Sgk1–/– TH17 cells restimulatedwith IL-23 confirmed that there is not only reduced phosphorylationof Foxo1 in the nucleus but increased mRNA and protein expression ofFoxo1 (Fig. 3b, c), suggesting that compromised phosphorylation ofFoxo1 can result in its own transcriptional upregulation. It has been
Gene
Il17a
Timp1Aqp3
TH1
7 (0
–2
h)
TH1
7 (0
–4
8 h
)T
H1
7 (5
0–7
2 h
)Il2
3r–
/– (<
60
h)
Il2
3r–
/– (>
60
h)
Netw
ork
sco
re
Igfbp4Sgk1Socs3SmoxGlipr2Frmd4bIer3Anxa2CpdPrnpCd24aGemMgllCdkn2dEts1RorcCd28
Cen
tralit
y s
co
re
300
200
100
0
(rela
tive e
xp
ressio
n)
T H0
T H1
TGF-
β + IL
-6T H2
iT reg
TGF-
β + IL
-21
TGF-
β + IL
-21
+ IL-2
3
TGF-
β + IL
-6 +
IL-2
3
TGF-
β + IL
-6 +
IL-2
3 + IL
-1β
TH17
600
400
200
0
Sg
k1
mR
NA
(rela
tive e
xp
ressio
n)
0 0.5 1 2 4 8 12 24 36 48 60 72
TGF-β + IL-6TGF-β + IL-6 + IL-23TH0
150
100
50
0
Time after stimulation (h)
Pdk1
SGK1Akt
Mapk1
Nedd4l Foxo3 Creb1
25% Centrality score
0
Transcription factor
IL-23R signalling
SGK1
No
de s
ize
IL-1
7A
1º TGF-β + IL-6 2º IL-23
CD
4
IL-1
7A
Il17f C
reSg
k1
+/+
Il17f C
reSg
k1
fl/fl
18.5 3.3637.6 0.13
0.21 4.94
35.3 0.09
0.16
9.61 1.06
2.31
IL-23R GFP
3.81 28.3
4.25 15.2
Sg
k1
+/+Il2
3r g
fpS
gk1
–/–Il2
3r g
fp
IFN-γ
WT
Sg
k1
–/–
38.4 0.01
0.15
41.2 0.01
0.25
33.8 0.73
14.3 0.63
1.93
3.48
Sp
p1
Il5Il3C
cl2
2Il4
i1T
nfs
f10
Il23
aIl2
rbIl2
raT
nfs
f8C
cl4
Ccl3
Il21
Csf2
Il13
ra1
Ccr1
Ccr6
Il1r2
Il17
aC
xcr3
Il7r
Il1r1
Il13
Ifng
Gen
e
Il2Il4Il9Lta
Pg
pS
lam
f7Ifi
tm3
Cd
38
Btla
Ab
cb
1a
Cd
83
Cd
48
Ifng
r1K
lrc1
Cd
80
Cd
68
Klrd
1P
lxn
c1
Sco
1N
t5e
Cd
93
Cd
86
Slc
44
a1
Pro
cr
Gen
eId
3Z
eb
1N
ab
1P
ja2
Etv
3F
oxp
1Jun
Eb
f1S
mad
4M
bn
l3R
un
x2
Nco
a1
Klf4
Pycr1
Ru
nx3
Bcl3
Atf4
Bach
2B
cl6
Fo
sD
dit3
Nfe
2l2
Cre
b3
l2A
hr
Po
u6
f1R
ora
Satb
1E
gr2
Eg
r3G
ata
3F
us
Tb
x2
1A
tf3M
sc
So
x4
Prm
t7T
rip1
3Ju
nd
Myb
bp
1a
Ph
f17
Ru
vb
l1M
llt3R
ep
in1
Pa2
g4
Mxd
1
Hiv
ep
3R
pl7
l1P
hb
Myc
Tra
db
pT
fdp
1
Cell-surface molecules Cytokines
Transcription factors
Il23r–/–
Sgk1–/–
–1 1
a b c d
e f g
h
0
15
104
103
102
101
100
104
103
102
101
100
100 101 102 103 104 100 101 102 103 104
104
103
102
101
100
104
103
102
101
100
100 101 102 103 104 100 101 102 103 104
1º TGF-β + IL-6 2º IL-23
IFN-γ
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
100 101 102 103 104 100 101 102 103 104
100 101 102 103 104100 101 102 103 104
1º TGF-β + IL-6 2º IL-23104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
100 101 102 103 104 100 101 102 103 104
100 101 102 103 104 100 101 102 103 104
Il23r–/–
Sgk1–/–
Sg
k1
mR
NA
–1
1
Fo
ld c
han
ge (lo
g2)
Fold change (log2)
Figure 1 | SGK1 is specifically induced in TH17 cells and is important fortheir maintenance. a, Top candidate genes ranked by their average of fold-increase in TH17 conditions (TGF-b1 with IL-6 compared to TH0) and folddecrease in Il23r–/– (knockout versus wild-type cells, TGF-b1, IL-6 and IL-23condition). The centrality score of a given protein is the percentage of IL-23R-affected transcription factors downstream of that protein in the network(Methods). b, Sgk1 mRNA expression in different T cell subsets. c, Kinetic analysisof Sgk1 gene expression in activated naive wild-type CD41 T cells differentiatedwith TGF-b, IL-6 and IL-23. d, IL-23R PPI network model (this is an enlargementof the SGK1 sub-network from the full network of Supplementary Fig. 1f). Nodes
are sized proportionally to their centrality score. e–g, Naive CD41 T cells fromSgk1–/– (e), Il17f CreSgk1fl/fl (f) or Sgk1–/–Il23rgfp (g) and control mice weredifferentiated into TH17 cells with TGF-b1 and IL-6 (left) or restimulated withIL-23 (right). IL-17 and IFN-c or IL-23R (GFP) expression were assessed.Numbers in the graphs indicate the percentage of cells in that quadrant. h, Heatmap displaying microarray data, fold change of selected gene subsets in the twoexperimental settings: Sgk1–/– versus wild-type, and Il23r–/– versus wild-typeTH17 cells (TGF-b1 and IL-6, restimulated with IL-23). Only genes with asignificant fold change in the Sgk1–/– TH17 cells are presented. Data arerepresentative of at least two independent experiments. Error bars, s.d.
RESEARCH LETTER
5 1 4 | N A T U R E | V O L 4 9 6 | 2 5 A P R I L 2 0 1 3
Macmillan Publishers Limited. All rights reserved©2013
shown previously that Foxo1 can regulate its own expression13 and wehave found that Foxo1 binds to a site located about 1 kilobase (kb)upstream of the first exon in the Foxo1 locus (Supplementary Fig. 6a).
Transfection of a Foxo1 luciferase reporter in the presence of Foxo1led to increased luciferase activity (Supplementary Fig. 6b), whereasincreasing expression of SGK1 in the presence of Foxo1 resulted in adose-dependent decrease in reporter activity, suggesting that SGK1inhibited Foxo1-mediated transactivation of its own promoter (Fig. 3d).
To decipher the consequences of Foxo1 expression on TH17 celldevelopment, we used Foxo1–/– CD41 memory T cells and observedhigher expression of Il23r and Il17a in these cells than in wild-type cells,indicating that Foxo1 may act as a repressor of TH17 cell developmentand of IL-23 signalling (Fig. 3e and Supplementary Fig. 6c). We alsofound potential binding sites of Foxo1 located about 1 kb upstream ofthe first exon of the Il23r locus by chromatin immunoprecipitation(ChIP)-PCR (Supplementary Fig. 6d). Moreover, there is significantlyenriched binding of Foxo1 on the Il23r promoter region in Sgk1–/– cellscompared to that in wild-type T cells, indicating enhanced suppressionof Il23r transcription in the absence of SGK1 (Fig. 3f). Retinoic-acid-receptor-related orphan receptor ct (RORct) has been suggested to bethe master transcription factor of TH17 development and ChIP-seq14
(ChIP coupled with high-throughput DNA sequencing) and ourChIP-PCR analysis confirmed that IL-23R is one of the targets ofRORct (Supplementary Fig. 6e). Indeed, we observed that the Il23rpromoter is transactivated by RORct in IL-23-restimulated TH17cells and it can be inhibited by Foxo1 in a dose-dependent manner(Supplementary Fig. 6f, g). While Foxo1 inhibited RORct-mediatedIl23r expression, co-expression of SGK1 together with RORct andFoxo1 abrogated the suppressive effects of Foxo1 and rescued Il23r
promoter transcriptional activity (Fig. 3g). Additionally, the inhibitionof Il23r transcription by a phosphorylation-insensitive triple alaninemutant of Foxo1, Foxo1 AAA, was not reduced in the presence ofSGK1 (Supplementary Fig. 6h). Furthermore, we observed an endogen-ous Foxo1–RORct interaction in primary TH17 cells (Fig. 3h). Thesedata support a model in which some of the effects of SGK1 are due tophosphorylation of Foxo1, which may be a key step in relieving RORctfrom Foxo1-mediated inhibition, enhancing the expression of IL-23R.
SGK1 has been reported to act as a mediator for sodium homeostasis.It can be induced by exogenous sodium chloride and is one of the majorkinases that regulates Na1 intake by phosphorylation of epithelialsodium channels (ENaCs)4,5. Considering the defects in TH17 develop-ment in Sgk1–/– mice, this raised the hypothesis that increasing sodiumconcentration may affect the TH17 cell phenotype through SGK1. To testthis, we first activated naive T cells in the presence of additional NaCl,but in the absence of any polarizing cytokines. Microarray analysis ofthese NaCl-treated cells showed a significant upregulation of Sgk1 and ofmultiple other genes associated with TH17 development (Fisher exacttest; P , 0.001; Supplementary Table 2 and Supplementary Fig. 7a),which we confirmed by qPCR analysis of selected genes (Sup-plementary Fig. 7b). We also found increased mRNA and protein levelsof IL-17 and IL-23R with additional NaCl under various TH17 polarizingconditions (Fig. 4a, b and Supplementary Fig. 7c). Furthermore, asodium-induced increase in TH17 development and IL-23R expressionwas not observed in SGK1-deficient T cells, specifically in the context ofIL-23–IL-23R signalling (Fig. 4c, d). Importantly, culturing cells withmannitol did not alter TH17 cell differentiation, excluding the possibilitythat the TH17 program is initiated simply by the alteration of osmoticpressure (Supplementary Fig. 7d).
Cd4CreSgk1+/+
Cd4CreSgk1fl/fl3
2
1
00 10 15 20 25
Time after immunization (d)
Mean
clin
ical sco
re
LN Spleen Blood CNS
IFN-γ
IL-1
7
10.3 1.35
2.45
12.5 1.93
6.38 0.51
5.86
0.98
4.31 0.46 8.64 3.11
14.2 5.26
11.2
11.5 3.27
20.4 10.7
17.2
9.63 15.7
Cd
4C
reSg
k1
+/+
Cd
4C
reSg
k1
fl/fl
4.12
Il17fCreSgk1+/+
Il17fCreSgk1fl/fl3
2
1
00 10 15 20 25
Mean
clin
ical sco
re
Time after immunization (d)
IL-1
7
15.7 24.7
CD4
7.68 20.3
Il17
f CreS
gk1
fl/fl
Il17
f CreS
gk1
+/+
IFN
-γ
CD45.2 TGF-β + IL-6
In vitro cultureSort out IL-17A GFP+
CD4+ T cells
Cd4CreSgk1+/+Il17agfp
or
Cd4CreSgk1fl/flIl17agfp
CD45.1 MOG + CFA
immunization
Transfer
7–10 days Spleen, LN, blood,
CNS analysis on
CD45.2+ cellsEAE onset
dLN Spleen Blood CNS
Cd
4C
reSg
k1
+/+Il1
7a
gfp
Cd
4C
reSg
k1
fl/flIl1
7a
gfp
10.4 20.5
5.22 7.31
CD45.2
6.33
18.5 21.4
2.76
IL-1
7A
GF
P
CD
45.2
+ C
D4 +
IL-1
7+ c
ells
(%
)
Cd4CreSgk1+/+Il17agfp
Cd4CreSgk1fl/flIl17agfp
25
20
15
10
5
0
dLN
Spleen
CNS
Blo
od
LN
CNS
Il17fCreRosa26ReYFPSgk1+/+
Il17fCreRosa26ReYFPSgk1–/–
22.8
12.7
eY
FP
26.1
7.54Co
unts
CD4 IL-17
7.64 32.7
3.55
eY
FP
Co
unts
CD4 IL-17
3.18
a b
c
d
e
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104
100 101 102 103 104100 101 102 103 104100 101 102 103 104100 101 102 103 104
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
100 101 102 103 104 100 101 102 103 104
100 101 102 103 104 100 101 102 103 104
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104
100 101 102 103 104100 101 102 103 104100 101 102 103 104100 101 102 103 104
*
***
***
Il17fCreRosa26ReYFPSgk1+/+
Il17fCreRosa26ReYFPSgk1–/–
104
103
102
101
100
104
103
102
101
100
100 101 102 103 104 100 101 102 103 104
100 101 102 103 104 100 101 102 103 104
100 101 102 103 104100 101 102 103 104
100 101 102 103 104100 101 102 103 104
104
103
102
101
100
104
103
102
101
100
Day 5
Figure 2 | SGK1-deficient mice are resistant to EAE, owing to a defect inmaintaining the TH17 phenotype. a, EAE development in Cd4CreSgk11/1 andCd4CreSgk1fl/fl mice (left), and IL-17 and IFN-c secretion by CD41 T cellsisolated from indicated organs at the peak of disease (right) (n 5 12). b, EAEdevelopment in Il17f CreSgk11/1 and Il17f CreSgk1fl/fl mice (left), and IL-17 andIFN-c secretion by CD41 T cells within the CNS (right) (n 5 10). c, Schematicillustration of adoptive transfer experiments shown in d. d, IL-17 production
from the donor CD41 cells collected from the indicated organs 12 days aftertransfer; representative histograms (left) and quantification of the flowcytometry data (right; means and s.d. are shown in red, n 5 10). dLN, draininglymph nodes. e, IL-17A production by CD41eYFP1 T cells isolated from lymphnodes or CNS of wild-type and SGK1-deficient Il17f CreRosa26ReYFP fate-reporter mice 17 days after MOG35-55–CFA immunization (n 5 10). *P , 0.05,**P , 0.01 and ***P , 0.001 (Student’s t-test). Error bars, s.d.
LETTER RESEARCH
2 5 A P R I L 2 0 1 3 | V O L 4 9 6 | N A T U R E | 5 1 5
Macmillan Publishers Limited. All rights reserved©2013
Recent studies have shown that different components in the dailydiet and gut microbiota can strongly affect the frequency of effector Tcells in the gut15,16. Furthmore, previous data indicate that moleculesrelated to sodium homeostasis can influence TH17 cell responses17,18.To understand further the effect of NaCl on TH17 cell generation
in vivo, we fed a high salt diet (HSD) to wild-type or Cd4CreSgk1fl/fl
mice. After 3 weeks on HSD, we observed that un-immunized wild-type mice showed a marked increase in the frequency of TH17 cells inthe lamina propria, whereas no notable changes were observed in themesenteric lymph nodes or spleen. Conversely, SGK1-deficient mice
SGK1
Foxo1
Smad4
E2f4 Jun
Dyrk1a
Cry2
No
de s
ize
25% Centrality score
0
a WT Sgk1–/–b
α-CD3/28α-CD3/28
+ IL-23
0 10 30 60 10 30 60 0 10 30 60 10 30 60 min
pFoxo1
Foxo1
Histone H3
α-CD3/28α-CD3/28
+ IL-23
Foxo1
mR
NA
(rela
tive e
xp
ressio
n)
150
100
50
0
**
WT Sgk1–/–
c d
Foxo1 luciferase
2
1.5
1
0.5
0
Lucifera
se
activity (R
U)
Foxo1 – – + + + + +SGK1 – + –
e
Il17a Il23r IL-17A2.5
2
1.5
1
0.5
0
4
3
2
1
0
2
1.5
1
0.5
0
********
Cd4CreFoxo1+/+
Cd4CreFoxo1fl/fl
ng
ml–
1
mR
NA
(rela
tive e
xp
ressio
n)
f Foxo1 enrichment
0.5
0.4
0.3
0.2
0.1
0
Inp
ut
(%)
1 2 3 4 5 6
0.5 kb Il23r
1 2 3 4 5 6
g Il23r luciferase
1
0.8
0.6
0.4
0.2
0
Lucifera
se
activity (R
U)
RORγt – + – – + – + + + + +
Foxo1 – – + – – + + + + + +
SGK1 – – – + + + –
hIP
IgG α-R
ORγt
IB: α-RORγt
IB: α-Foxo1
Who
le ly
sate
IP
IgG
IB: α-Foxo1
α-Fox
o1
IB: α-RORγt
Who
le ly
sate
WTSgk1–/–
WT
Sgk1–/
–
Foxo1
Tubulin
Transcription factor
SGK1
Foxo1
Figure 3 | SGK1 signalling promotes IL-23R expression throughphosphorylation of Foxo1. a, SGK1 PPI network model. Left, networkcomposed of all the protein nodes with a P value of under 0.0001 (Methods); right,enlargement of the Foxo1 sub-network. Nodes are sized relative to their centralityscore. In the subnetwork, directed edges (arrows) from one protein to anothercorrespond to post-translational modifications of the second protein by the first.Non-directed edges (lines without arrowheads) correspond to PPIs between oneprotein and another with no known directionality. b, Phosphorylated Foxo1(p-Foxo1; phosphorylation site Ser 256) and total Foxo1 levels were assessed innuclear extracts after restimulation of wild-type (WT) and Sgk1–/– TH17 cells.c, Levels of mRNA (left) and protein (right) of Foxo1 were analysed 3 daysafter IL-23 restimulation of differentiated TH17 cells. d, HEK293T cells weretransfected with a Foxo1 promoter-driven luciferase reporter along with theindicated plasmids, and promoter activity was assessed. e, Memory CD41 T cells
from Foxo1–/– mice were stimulated for 24 h with PMA (phorbol myristateacetate) and ionomycin, and levels of IL-17A and IL-23R expression weredetermined by qPCR or ELISA (enzyme-linked immunosorbent assay) (IL-17A).f, The binding of Foxo1 to the Il23r promoter in wild-type and Sgk1–/– IL-23-restimulated TH17 cells was assayed by ChIP-PCR. Six horizontal bars representthe locations of the Foxo1-binding sites on the Il23r locus detected by real-timePCR. Two thick vertical bars represent exons 1 and 2 of the Il23r locus. g, Il23rpromoter activity was measured in HEK293T cells transfected with an Il23rpromoter-driven luciferase reporter along with the indicated plasmids.h, Immunoprecipitation (IP) (control IgG (immunoglobulin-c), anti-RORct oranti-Foxo1) of lysates of wild-type IL-23-restimulated TH17 cells, followed byimmunoblot (IB) analysis with indicated antibodies. **P , 0.01 and***P , 0.001 (Student’s t-test). Data are representative of three independentexperiments. RU, relative units. Error bars, s.d.
None TGF-β + IL-6 IL-23 IL-6 + IL-23
0.01 0.01 39.2 0.01 0.76 0.02 0.39 0.01
2.48 0.01 3.65 4.74
2.12 0.01 59.4 0.01 12.9 0.01 22.8 0.01
0.25 0.01 0.46 0.21
Control
IFN-γ
IL-1
7
NaCl
a b
CD
4
None None TGF-β + IL-6 IL-23 IL-6 + IL-23
0.23 0.17 1.75
0.88 0.43 12.6
5.42
56.7
Control
NaCl
IL-23R GFP
NaCld
0.33
CD
4
1.54 2.19
0.46 0.42 1.78 24.4
55.3 77.6
40.8
Sg
k1
+/+Il2
3r g
fpS
gk1
–/–Il2
3r g
fp
IL-23R GFP
c NaCl
1.33 0.12 44.2 0.01
IL-1
7
4.55 0.74 25.3 1.52
WT
12.5 0.12
0.67 0.23 47.1 0.01 2.23 0.59
7.64 1.26
1.138.450.0111.8
13.2 1.21
Sgk1–/–
IFN-γ
e f3.5
3
2.5
2
1.5
1
0.5
0
Mean c
linic
al sco
re
Cd4CreSgk1+/+
Cd4CreSgk1fl/fl
7 10 13 16 19 22 25
Time after immunization (d)
Cd4CreSgk1+/+
Cd4CreSgk1fl/fl
60
40
20
0
80
60
40
20
0
* *
*
**
Spleen mLN CNS Spleen mLN CNS
CD
4+IL
-17
+ c
ells
(%
)
Ctrl
HSD
Ctrl
HSD
Ctrl
HSD
Ctrl
HSD
Ctrl
HSD
Ctrl
HSD
CD
4+IF
N-γ
+ c
ells
(%
)
*
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
None TGF-β + IL-6 IL-23 IL-6 + IL-23104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104
100 101 102 103 104100 101 102 103 104100 101 102 103 104 100 101 102 103 104
100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104
100 101 102 103 104100 101 102 103 104100 101 102 103 104100 101 102 103 104
None TGF-β + IL-6 IL-23 IL-6 + IL-23104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104
100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
104
103
102
101
100
100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104
100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 100 101 102 103 104
Cd4CreSgk1fl/fl HSD
Cd4CreSgk1+/+ HSD
Figure 4 | NaCl potentiates TH17 celldifferentiation in vitro and in vivo, enhancingEAE induction. a, IL-17 and IFN-c production bynaive wild-type CD41 T cells stimulated withindicated cytokines for 3 days with or without40 mM NaCl. Ctrl, control. b, IL-23R (GFP)expression in Il23rgfp CD41 T cells stimulatedunder conditions described in a. c, IL-17 and IFN-cproduction by wild-type (WT) and Sgk1–/– CD41 Tcells stimulated with the indicated cytokines in thepresence of 40 mM NaCl. d, IL-23R (GFP)expression in Sgk1–/–Il23rgfp and control CD41 Tcells stimulated under conditions described inc. e, Clinical scores of EAE in Cd4CreSgk11/1 andCd4CreSgk1fl/fl mice fed with high salt diet (HSD) orcontrol diet (n 5 15). f, Quantification of CD41IL-171 or CD41IFN-c1 T cells from the indicatedorgans of Cd4CreSgk11/1 and Cd4CreSgk1fl/fl micefed HSD or control diet on day 17 afterimmunization with MOG35-55–CFA. *P , 0.05(Student’s t-test). Data are representative of threeindependent experiments. mLN, mesenteric lymphnode. Error bars, s.d.
RESEARCH LETTER
5 1 6 | N A T U R E | V O L 4 9 6 | 2 5 A P R I L 2 0 1 3
Macmillan Publishers Limited. All rights reserved©2013
showed a much milder enhancement of TH17 cell frequency in the gut,whereas there was no increase in IFN-c production in any of the micefed with HSD (Supplementary Fig. 8a, b).
Finally, we studied whether HSD would affect the development ofTH17 and EAE in vivo. Wild-type mice fed HSD had more severe EAEthan mice fed a normal diet, and this increased severity was dramat-ically reduced in SGK1-deficient mice (Fig. 4e and Supplementary Fig.8c,d). We also observed a significantly higher frequency of TH17 cellsin mesenteric lymph nodes and CNS of wild-type mice fed with HSDthan in those of SGK1-deficient mice fed with HSD. The percentage ofIFN-c producing T cells in the CNS, but not in the peripheral immunecompartments, of wild-type mice was increased in mice fed HSD,suggesting that HSD may increase infiltration but not expansion ofIFN-c1 effector T cells in the target organs (Fig. 4f and SupplementaryFig. 8e). Consistent with our in vitro data, we observed elevated IL-17but not IFN-c production from CD41 T cells isolated from EAE-immunized wild-type mice fed with HSD and restimulated in vitrowith MOG35-55, compared to production from cells from EAE micefed a normal diet (Supplementary Fig. 8f). The data presented hereindicate that high sodium intake potentiates TH17 cell generation invivo in an SGK1-dependent manner and therefore has the potential toincrease the risk of promoting autoimmunity.
In conclusion, we used a combination of microarray data analysis,large-scale PPI network analysis and experimental data from severaldifferent knockout mice to establish IL-23R-SGK1-Foxo1 as a criticalaxis in TH17 stabilization. We show that Foxo1 acts as a repressor of IL-23R expression by binding directly to the Il23r promoter and inhibi-ting RORct-mediated Il23r transactivation. Phosphorylation of Foxo1,mediated by SGK1, leads to its deactivation and promotes unopposedRORct-mediated Il23r transcription. SGK1 has been studied exten-sively in the context of NaCl transport19,20. Modest increase of theNaCl concentration induces SGK1 expression in T cells with increasedIL-23R expression and TH17 cell generation in vitro. Interestingly, evenin un-immunized mice fed with HSD, enhancement of TH17 differ-entiation was observed in vivo in the gut and gut-associated lymphoidtissue, and this increase in TH17 cells can be recalled at other peripheralsites after immunization. Although our data suggest an essential rolefor SGK1 in this process, it is likely that other immune cells and path-ways are also influenced by increased salt intake. Furthermore, ourresults do not exclude additional alternative mechanisms by whichan increase in NaCl affects TH17 cells. Nevertheless, the elevated invivo TH17 differentiation resulting from HSD raises the importantissue of whether increased salt in westernized diets and in processedfoods contributes to an increased generation of pathogenic TH17 cellsand for an unprecedented increase in autoimmune diseases.
METHODS SUMMARYMicroarrays and network analysis. For gene-expression analysis Affymetrixmicroarray chips were used. Data were processed using the GenePattern suite21. Dif-ferentially expressed genes were detected using fold-change and t-test analysis (forSgk1–/– and NaCl-treated T cells) or a consensus of fold-change, the EDGE software22
and a novel sigmoid-based method23 (for the Il23r–/– TH17 cell time-course data). Acommand-line version of the ANAT software9 was used for network analysis.In vitro T cell differentiation. Naive T cells were FACS-sorted, stimulated withplate-bound anti-CD3 and anti-CD28 antibodies and the indicated cytokines orNaCl, and cells were analysed by qPCR or flow cytometry at different time points.Experimental autoimmune encephalomyelitis model. Mice were immunizedsubcutaneously with MOG35-55 in complete Freud’s adjuvant (CFA), and heat-inactivated Mycobacterium tuberculosis and with intraperitoneal injection ofBordatella pertussis toxin.In vivo cell transfer. Naive T cells were differentiated towards TH17 cells, thentransferred into MOG35-55–CFA-immunized hosts and T cells isolated from vari-ous organs were analysed by flow cytometry at 7–12 days after onset of EAE.Western blot and immunoprecipitation. Differentiated T cells or transfectedHEK293T cells were lysed in whole cell extract (WCE) buffer (containing 50 mMTris buffer, pH 7.5, 100 mM NaCl, 0.1% Triton X-100, 10% v/v glycerol, 1 mMDTT, 1 mM PMSF and protease inhibitors (Sigma)), and lysates were subjected towestern blot or immunoprecipitation analysis.
Promoter-activity reporter assay. HEK293T cells were transfected with lucifer-ase reporter constructs and expression vectors, and luciferase expression wasdetermined after 48 h.
Received 30 May 2012; accepted 6 February 2013.
Published online 6 March 2013.
1. Korn, T., Bettelli, E., Oukka, M. & Kuchroo, V. K. IL-17 and Th17 Cells. Annu. Rev.Immunol. 27, 485–517 (2009).
2. Aggarwal, S., Ghilardi, N., Xie, M. H., de Sauvage, F. J. & Gurney, A. L. Interleukin-23promotes a distinct CD4 T cell activation state characterized by the production ofinterleukin-17. J. Biol. Chem. 278, 1910–1914 (2003).
3. Zhou, L. et al. IL-6 programs TH-17 cell differentiation by promotingsequential engagement of the IL-21 and IL-23 pathways. Nature Immunol. 8,967–974 (2007).
4. Lang, F. et al. (Patho)physiological significance of the serum- and glucocorticoid-inducible kinase isoforms. Physiol. Rev. 86, 1151–1178 (2006).
5. Wulff, P. et al. Impaired renal Na1 retention in the sgk1-knockout mouse. J. Clin.Invest. 110, 1263–1268 (2002).
6. Salker, M. S. et al. Deregulation of the serum- and glucocorticoid-inducible kinaseSGK1 in the endometrium causes reproductive failure. Nature Med. 17,1509–1513 (2011).
7. Zhang, L., Cui, R., Cheng, X. & Du, J. Antiapoptotic effect of serum andglucocorticoid-inducible protein kinase is mediated by novel mechanismactivating IkB kinase. Cancer Res. 65, 457–464 (2005).
8. Shelly, C.& Herrera, R. Activation ofSGK1 by HGF,Rac1and integrin-mediatedcelladhesion in MDCK cells: PI-3K-dependent and -independent pathways. J. Cell Sci.115, 1985–1993 (2002).
9. Yosef, N. et al. ANAT: a tool for constructing and analyzing functional proteinnetworks. Sci. Signal. 4, pl1 (2011).
10. Brunet, A. et al. Protein kinase SGK mediates survival signals by phosphorylating theforkhead transcription factor FKHRL1 (FOXO3a). Mol. Cell. Biol. 21, 952–965 (2001).
11. Wei, G. et al. Global mapping of H3K4me3 and H3K27me3 reveals specificity andplasticity in lineage fatedeterminationofdifferentiatingCD41 Tcells. Immunity30,155–167 (2009).
12. Di Pietro, N. et al. Serum- and glucocorticoid-inducible kinase 1 (SGK1) regulatesadipocytedifferentiationvia forkheadboxO1.Mol.Endocrinol.24,370–380 (2010).
13. Essaghir, A.,Dif, N., Marbehant, C. Y., Coffer, P. J.& Demoulin, J. B. The transcriptionof FOXO genes is stimulated by FOXO3 and repressed by growth factors. J. Biol.Chem. 284, 10334–10342 (2009).
14. Ciofani,M.et al. A validated regulatory network for Th17cell specification.Cell 151,289–303 (2012).
15. Berer, K. et al. Commensal microbiota and myelin autoantigen cooperate to triggerautoimmune demyelination. Nature 479, 538–541 (2011).
16. Sczesnak, A. et al. The genome of th17 cell-inducing segmented filamentousbacteria reveals extensive auxotrophy and adaptations to the intestinalenvironment. Cell Host Microbe 10, 260–272 (2011).
17. Stegbauer, J. et al. Role of the renin-angiotensin system in autoimmuneinflammation of the central nervous system. Proc. Natl Acad. Sci. USA 106,14942–14947 (2009).
18. Herrada, A. A. et al. Aldosterone promotes autoimmune damage by enhancingTh17-mediated immunity. J. Immunol. 184, 191–202 (2010).
19. Diakov, A. & Korbmacher, C. A novel pathway of epithelial sodium channelactivation involves a serum- and glucocorticoid-inducible kinase consensus motifin the C terminus of the channel’s alpha-subunit. J. Biol. Chem. 279,38134–38142 (2004).
20. Loffing, J. et al. Aldosterone induces rapid apical translocation of ENaC in earlyportion of renal collecting system: possible role of SGK. Am. J. Physiol. RenalPhysiol. 280, F675–F682 (2001).
21. Reich, M. et al. GenePattern 2.0. Nature Genet. 38, 500–501 (2006).22. Storey, J. D., Xiao, W., Leek, J. T., Tompkins, R. G. & Davis, R. W. Significance analysis
of time course microarray experiments. Proc. Natl Acad. Sci. USA 102,12837–12842 (2005).
23. Chechik, G. & Koller, D. Timing of gene expression responses to environmentalchanges. J. Comput. Biol. 16, 279–290 (2009).
Supplementary Information is available in the online version of the paper.
Acknowledgements We thank D. Kozoriz for cell sorting and A. Waisman for providingIl17fCre mice. L. Zhou, D. Accili, J. Demoulin and K. Sato provided reagents. This work wassupportedby theUSNational InstitutesofHealth (NS030843,NS045937,AI073748andAI045757 to V.K.K.; 1P01HG005062-01, 1P50HG006193-01 and DP1-OD003958-01to A.R.; and K01DK090105 to S.X.); the National MS Society (RG2571 to V.K.K.); theHoward Hughes Medical Institute (A.R.); the Klarman Cell Observatory; Guthy JacksonFoundation; and the Austrian Science Fund (FWF, J 3091-B12 to T.T.).
Author Contributions C.W, N.Y. and T.T. carried out experiments and wrote themanuscript. C.Z., S.X. and Y.K. carried out experiments. N.Y. analysed the data. A.R. andV.K.K. supervised the study and edited the manuscript.
Author Information The microarray data sets have been deposited in the GeneExpression Omnibus database under accession numbers GSE43956, GSE43957 andGSE43969. Reprints and permissions information is available at www.nature.com/reprints. The authors declare no competing financial interests. Readers are welcome tocomment on the online version of the paper. Correspondence and requests formaterials should be addressed to A.R. ([email protected]) or V.K.K.([email protected]).
LETTER RESEARCH
2 5 A P R I L 2 0 1 3 | V O L 4 9 6 | N A T U R E | 5 1 7
Macmillan Publishers Limited. All rights reserved©2013
