Protein kinase CK2 enables regulatory T cells to suppress ... - Nature

Articles

npg

© 2015 Nature America, Inc. All rights reserved.

Protein kinase CK2 enables regulatory T cells to suppress excessive TH2 responses in vivo Alexander Ulges1, Matthias Klein1, Sebastian Reuter2, Bastian Gerlitzki1, Markus Hoffmann1, Nadine Grebe1, Valérie Staudt1, Natascha Stergiou1, Toszka Bohn1, Till-Julius Brühl1, Sabine Muth1, Hajime Yurugi1, Krishnaraj Rajalingam1, Iris Bellinghausen3, Andrea Tuettenberg3, Susanne Hahn3, Sonja Reißig4, Irma Haben5, Frauke Zipp6, Ari Waisman4, Hans-Christian Probst1, Andreas Beilhack7, Thierry Buchou8, Odile Filhol-Cochet9, Brigitte Boldyreff10, Minka Breloer5, Helmut Jonuleit3, Hansjörg Schild1, Edgar Schmitt1,11 & Tobias Bopp1,11 The quality of the adaptive immune response depends on the differentiation of distinct CD4 + helper T cell subsets, and the magnitude of an immune response is controlled by CD4+Foxp3+ regulatory T cells (Treg cells). However, how a tissue- and cell type–specific suppressor program of Treg cells is mechanistically orchestrated has remained largely unexplored. Through the use of Treg cell–specific gene targeting, we found that the suppression of allergic immune responses in the lungs mediated by T helper type 2 (TH2) cells was dependent on the activity of the protein kinase CK2. Genetic ablation of the b-subunit of CK2 specifically in Treg cells resulted in the proliferation of a hitherto-unexplored ILT3+ Treg cell subpopulation that was unable to control the maturation of IRF4+PD-L2+ dendritic cells required for the development of TH2 responses in vivo. The maintenance of peripheral tolerance is based chiefly on the sup pressive properties of CD4+ regulatory T cells (Treg cells). The develop ment of Treg cells is orchestrated by the transcription factor Foxp3 (ref. 1), and mutations in the Foxp3 locus result in Treg cell deficiency associated with the development of fatal multiorgan autoimmune dis eases and allergies2,3. On the cellular level, deficiency in Treg cells results in diseases that are, among other features, characterized by polyclonal activation and proliferation of cells of the immune system belonging to the lymphoid lineage as well as the myeloid lineage, which indicates the vital importance of Treg cells for the immunological tolerance network4. Therefore, modulation of Treg cell activity in vivo is believed to be a valuable therapeutic strategy in many different diseases. Analyses of the molecular mechanisms that underlie the suppressive properties of Treg cells have revealed the involvement of various molecules5. Among these, kinases have emerged as valuable targets for the treatment of diseases caused by dysregulation of the immune system6. Therefore, exploration of the Treg cell ‘kinome’ (the kinases encoded by Treg cells) could provide important insights for the development of innovative therapeutic strategies. Here we performed a comparative analysis of the Treg cell kinome that revealed the protein kinase CK2 (‘casein kinase 2’; encoded by Csnk2) was ‘preferentially’ active in human and mouse Treg cells upon activation mediated by the T cell antigen receptor (TCR).

CK2 is a highly conserved serine-threonine kinase that is expressed in all eukaryotic organisms and is involved in many signal-transduction pathways, including the NF-κB, PI(3)K and Wnt pathways7. Due to its ability to phosphorylate a multitude of signaling proteins linked to tumorigenesis and tumor suppression, CK2 is believed to be a prom ising target for the treatment of malignant diseases8. It functions as a tetrameric complex of two catalytic α-subunits (α and α′) and two regulatory β-subunits. The β-subunits (CK2β; encoded by Csnk2b) seem to be important for the stability of CK2α and provide specificity to substrates9. In general, CK2 is thought to be constitutively active; however, its regulation and cell type–specific subcellular localiza tion and mode of action remain largely unknown. To investigate the function of CK2 in Treg cells, we studied mice carrying loxP-flanked Csnk2b alleles10 conditionally deleted in Treg cells by Cre recombinase expressed under control of the Foxp3 locus11. We found that Treg cell– specific deficiency in CK2β resulted in the spontaneous development of excessive T helper type 2 (TH2) responses in the lungs. RESULTS High CK2 activity in Treg cells To gain knowledge about kinases that specifically regulate the sup pressive ability of Treg cells, we used kinase profiling with peptide

1Institute for Immunology, University Medical Center of the Johannes Gutenberg-University Mainz, Mainz, Germany. 2Department of Pulmonary Medicine, III. Medical Clinic of the University Medical Center, Johannes Gutenberg-University Mainz, Mainz, Germany. 3Dermatology, University Medical Center of the Johannes Gutenberg-University Mainz, Mainz, Germany. 4Institute for Molecular Medicine, University Medical Center of the Johannes Gutenberg-University Mainz, Mainz, Germany. 5Department of Immunology and Virology, Bernhard Nocht Institute for Tropical Medicine, Hamburg, Germany. 6Department of Neurology, University Medical Center of the Johannes Gutenberg-University Mainz, Mainz, Germany. 7Experimentelle Stammzelltransplantation der Medizinische Klinik und Poliklinik II, Zentrum für Experimentelle Molekulare Medizin, Julius-Maximilians-Universität Würzburg, Würzburg, Germany. 8INSERM, U823, Université Joseph Fourier-Grenoble1, Institut Albert Boniot, Faculté de Médecine, Domaine de la Merci, La Tronche, France. 9INSERM U1036, Institute de Recherches en Technologies et Sciences pour le Vivant/Biologie du Cancer et de l’Infection, Commissariat à l’Énergie Atomique et aux Énerigies Alternatives Grenoble, Grenoble, France. 10KinaseDetect, Aarslev, Denmark. 11These authors jointly directed this work. Correspondence should be addressed to T. Bopp. ([email protected]).

Received 8 October 2014; accepted 15 December 2014; published online 19 January 2015; doi:10.1038/ni.3083

nature immunology VOLUME 16 NUMBER 3 MARCH 2015

267

Articles

Spontaneous TH2 differentiation in Csnk2bfl/flFoxp3-Cre mice Flow cytometry of peripheral T cells demonstrated activation and population expansion of CD4+ T cells as well as CD8+ T cells in Csnk2bfl/flFoxp3-Cre mice. This was evident in the considerable increase in the frequency of CD62LloCD44hi (effector memory) T cells among peripheral CD4+ T cells solely in Csnk2bfl/flFoxp3-Cre mice 268

fl

Fo snk xp 2b f l/f 3C l re

a

fl/

b

Csnk2b 10

5

fl/fl

fl/fl

2.3

Csnk2b Foxp3-Cre

83.4

1.6

85.7

11.0

3.6

8.8

4

10

3

3.4 0 10 CD4

50 0

+

D C

fl/fl

10 8 6 4 2 0

Csnk2b fl/fl Csnk2b Foxp3-Cre

P

100

100 80 60 40 20 0

e

D 4S

150

fl/fl


P

200

d

D

Csnk2bfl/fl Csnk2bfl/fl Foxp3-Cre

4 5 10 10

8S P

c

3

C

β-actin

10 0

Foxp3 cells (%)

CD8

CK2β

D N

microarrays to comparatively analyze the kinomes of mouse as well as human Treg cells and CD4+Foxp3− effector T cells (Teff cells) stimu lated via the TCR. These analyses resulted in a list of kinases that were ‘preferentially’ active in Treg cells after activation. Among the kinases highly active mainly in mouse Treg cells, CK2 showed the greatest difference compared with its activity in Teff cells (Fig. 1a). Similarly, CK2 was highly active in human Treg cells compared with their activity in the respective Teff cell population (Fig. 1b). On the basis of the results obtained from the kinome analysis, we hypothesized that CK2 might crucially contribute to the suppressive properties of Treg cells. To test this, we generated a mouse model in which only Treg cells lacked CK2β by crossing Foxp3-IRES-Cre mice (which express Cre recombinase under control of the Foxp3 locus)11 with mice carrying loxP-flanked Csnk2b alleles (Csnk2bfl/fl)10, both on the C57BL/6 background. The Csnk2bfl/flFoxp3-Cre progeny were born at the expected Mendelian ratio and were phenotypically indistinguishable from their Cre-negative littermates (Csnk2bfl/fl mice) during the first 10 weeks of life (data not shown). Following Cre-mediated ablation of Csnk2b in Csnk2bfl/flFoxp3-Cre mice, CK2β was undetectable in Treg cells isolated from these mice (Fig. 2a). Comparison of Csnk2bfl/flFoxp3-Cre mice with Csnk2bfl/fl mice demonstrated unaltered thymic development of CD4+ T cells, CD8+ T cells (Fig. 2b–d) and Treg cells (Fig. 2e) and no impairment on Treg cell homeostasis in the periphery of Csnk2bfl/flFoxp3-Cre mice (Supplementary Fig. 1). However, histopathological analyses of these mice revealed increased cell numbers in the spleen and several lymph nodes (LNs), most prominently in the tracheal LNs (tLNs) (Fig. 3a,b). Accordingly, we observed massive infiltration of cells of the immune system into the lungs of Csnk2bfl/flFoxp3-Cre mice (Fig. 3c), whereas other organs (skin, liver, small intestine and colon) were affected far less (Fig. 3d). Histological staining revealed that the majority of lunginfiltrating cells were CD45+CD3+ T cells (data not shown). With increasing age, cell infiltration, splenomegaly and lymphadenopathy also became visible in other secondary lymphoid organs (mesenteric, inguinal and popliteal LNs; data not shown). However, immunopa thology was not comparable to that observed in Treg cell–deficient scurfy mice in any organ except the lungs (Fig. 3c,d), which indicated that CK2 controlled a suppressor program that was cell type–specific and was induced in a tissue-specific manner.

Cells (%)

npg


Figure 1 Mouse and human Treg cells display high activity of CK2. Peptide microarray analysis of the predicted kinase activity of the top six kinases deregulated in mouse Treg cells (a) and human Treg cells (b) upon polyclonal stimulation for 30 min by plate-bound anti-CD3 and anti-CD28; results are presented relative to those of Teff cells (Treg/Teff). Data are from one experiment.

b

–6

C

–4

k2

0 –2

sn

n Er k2 cSR C

Fy

–10

C K2 N EM O C D K5

–5

2

(Fig. 4a,b). Detailed flow cytometry of tLN-resident effector memory T cells from Csnk2bfl/flFoxp3-Cre mice and Csnk2bfl/fl mice demon strated a considerably enhanced frequency in GATA-3-expressing TH2 cells but a comparatively small increase in TH1 cells expressing the transcription factor T-bet and cells of the TH17 subset of helper T cells expressing the transcription factor RORγt in Csnk2bfl/flFoxp3Cre mice (Fig. 4c,d). Several independent experiments demonstrated no difference in the abundance of T-bet+ or RORγt+ T cells, while the proportion of GATA-3+ T cells was considerably enhanced in the Csnk2bfl/flFoxp3-Cre cohort (Fig. 4d). Concomitantly, ex vivo polyclonal stimulation of CD4+ effector memory T cells from both Csnk2bfl/flFoxp3-Cre mice and Csnk2bfl/fl mice revealed comparatively high production of interleukin 4 (IL-4) solely by effector memory T cells from Csnk2bfl/flFoxp3-Cre mice, whereas their comparatively low production of interferon (IFN-γ) and IL-17 was slightly reduced (IFN-γ) or marginally elevated (IL-17) (Fig. 4e). TH2 cell–derived IL-4 is important for class-switch recombination to immunoglobulin E (IgE)12. Therefore, we measured total serum IgE titers. Notably, under steady-state conditions, Csnk2bfl/flFoxp3Cre mice showed elevated titers of IgE in serum (Fig. 4f). In addition, we detected an increase in splenic B220 + germinal center B cells expressing the germinal center marker GL7 and the cell surface receptor Fas in Csnk2bfl/flFoxp3-Cre mice (Fig. 4g,h), which further confirmed our observation that Treg cell–specific deficiency in CK2β resulted in the spontaneous development of an uncontrolled TH2 cell–driven immune response. The aforementioned differences between Csnk2bfl/flFoxp3-Cre mice and Csnk2bfl/fl mice could potentially have been due to a decrease in the number of peripheral Treg cells or an altered homing capacity in absence of CK2β. However, the number of CK2β-deficient Treg cells

C

0

Human

Thymocytes (×106)

5

4

C K2 Ak t1 C D K2 PA K2 G SK 3β

b

Mouse

EG FR

Activity (Treg/ Teff)

10

Activity (Treg/ Teff)

a

Figure 2 Ablation of Csnk2b in Treg cells has no effect on their thymic development. (a) Immunoblot analysis of CK2β in ex vivo CD25+ Treg cells sorted by magnetic-activated cell sorting from Csnk2bfl/flFoxp3-Cre mice and their Csnk2bfl/fl littermates. (b) Distribution of CD4−CD8− thymocytes (bottom left), CD4+CD8+ thymocytes (top right), CD4+CD8− thymocytes (bottom right) and CD4−CD8+ thymocytes (top left) in Csnk2bfl/fl and Csnk2bfl/flFoxp3-Cre mice. Numbers in quadrants indicate percent cells in each throughout. (c) Quantification of thymocytes in Csnk2bfl/fl and Csnk2bfl/flFoxp3-Cre mice. Each symbol represents an individual mouse (n = 10 per genotype); small horizontal lines indicate the mean (± s.d.). (d,e) Frequency of CD4−CD8− (DN), CD4+CD8+ (DP), CD4+CD8− (CD4SP) and CD4−CD8+ (CD8SP) thymocytes (d) and of Foxp3+ Treg cells among CD4+CD8− thymocytes (e) in Csnk2bfl/fl and Csnk2bfl/flFoxp3-Cre mice. Data are representative of at least two independent experiments (a,b) or are pooled from three (c) or two (d,e) independent experiments (mean and s.d. of n = 6 mice per genotype in d,e).

VOLUME 16 NUMBER 3 MARCH 2015 nature immunology

npg

* Popliteal LN cells (×104)

Splenocytes (×107)

40 30 20 10 0

b

500

Csnk2bfl/fl

250

400 300 200 100

c tLN Spleen

* tLN cells (×105)

a

200 150 100 50

0 Colon

** Csnk2bfl/fl Csnk2bfl/fl Foxp3-Cre

0 Small intestine

Lung

Liver

Skin

Csnk2bfl/fl

Csnk2bfl/fl Foxp3-Cre

Csnk2bfl/fl Foxp3-Cre Scurfy

d 3 2 1 0

** 4 3 2 1 0

was unaltered in the blood, tLNs and lungs of Csnk2bfl/flFoxp3-Cre mice relative to their abundance in Csnk2bfl/fl mice (Fig. 5a,b and Supplementary Fig. 1), which thus excluded the possibility that the TH2 cell–driven pathologies resulted from a decreased number or altered homing ability of Treg cells. Csnk2bfl/flFoxp3-Cre mice develop asthmatic symptoms To further analyze the consequence of the uncontrolled TH2 response noted above in vivo, we established a mouse model of allergic asthma (Supplementary Fig. 2). For this, we either sensitized Csnk2bfl/flFoxp3Cre mice and Csnk2bfl/fl mice to ovalbumin (OVA) by intraperitoneal immunization of OVA emulsified in aluminum hydroxide (alum) or gave the mice injection of PBS instead. On days 15 and 16 following sensitization, we challenged all mice via the airways with OVA by ultrasonic nebulization. On day 18, histopathological analysis of lungs revealed massive infiltration of cells of the immune system (Fig. 6a,b), concomitant with considerable mucus production (Fig. 6c,d), in all mice sensitized to OVA. Of note, these asthma pathologies were already present in Csnk2bfl/flFoxp3-Cre mice without prior sensiti zation (i.e., those given injection of PBS) but were not detectable in the lungs of their PBS-treated Csnk2bfl/fl counterparts (Fig. 6a–d). After challenge with OVA, the substantial asthma pathologies that were detectable in nonsensitized Csnk2bfl/flFoxp3-Cre mice (i.e., those given injection of PBS) were only slightly increased in their sensitized Csnk2bfl/flFoxp3-Cre counterparts (i.e., those given injection of OVA emulsified in alum) (Fig. 6a–d). This suggested that Treg cell–specific deficiency in CK2β caused full-blown allergic inflammation in the lungs even in the absence of an additional disease-provoking treat ment. Accordingly, influx of eosinophils into the lungs was always nature immunology VOLUME 16 NUMBER 3 MARCH 2015

4

Lung inflammation score

Small intestine inflammation score

4

3 2 1 0

**

** Skin inflammation score

Colon inflammation score

Figure 3 Ablation of Csnk2b in Treg cells results in TH2-type lymphoproliferative disease in the lungs. (a) Cellularity of the spleen, popliteal LNs and tLNs of Csnk2bfl/fl and Csnk2bfl/flFoxp3-Cre mice. Each symbol represents an individual mouse (n = 20 (Csnk2bfl/fl) or 22 (Csnk2bfl/fl Foxp3-Cre) for spleen; n = 9 (Csnk2bfl/fl) or 10 (Csnk2bfl/flFoxp3-Cre) for popliteal LNs; n = 13 (Csnk2bfl/fl) or 14 (Csnk2bfl/flFoxp3-Cre mice) for tLNs); small horizontal lines indicate the mean (± s.d.). (b) Photographs of tLNs and spleens of Csnk2bfl/fl and Csnk2bfl/flFoxp3-Cre mice. (c) Histopathology of tissue sections from 10- to 15-week-old Csnk2bfl/fl and Csnk2bfl/flFoxp3-Cre mice, as well as from 3-week-old Foxp3-deficient (scurfy) male mice, stained with hematoxylin and eosin. Arrows indicate inflammatory cell infiltration. Original magnification, ×100. Scale bars, 100 µm. (d) Histopathology scores of hematoxylin and eosin–stained sections of tissue as in c. Each symbol represents an individual mouse (n = 16 (Csnk2bfl/fl) 17 (Csnk2bfl/flFoxp3-Cre) or 3 (scurfy) for colon; n = 17 (Csnk2bfl/fl) 17 (Csnk2bfl/flFoxp3-Cre) or 3 (scurfy) for small intestine; n = 22 (Csnk2bfl/fl), 24 (Csnk2bfl/flFoxp3-Cre) or 7 (scurfy) for lungs; n = 11 (Csnk2bfl/fl), 11 (Csnk2bfl/flFoxp3-Cre) or 5 (scurfy) for liver; n = 8 (Csnk2bfl/fl), 8 (Csnk2bfl/flFoxp3-Cre) or 3 (scurfy) for skin); small horizontal lines indicate the mean (± s.d.). *P < 0.01 and **P < 0.001 (unpaired t-test). Data are pooled from three independent experiments (a,d) or are representative of three experiments (b,c).

Liver inflammation score


Articles

4 3 2

4

**

**

*

3 2 1 0

** Csnk2bfl/fl Csnk2bfl/flFoxp3-Cre Scurfy

1 0

considerably higher in Csnk2bfl/flFoxp3-Cre mice than in Csnk2bfl/fl mice (Supplementary Fig. 3). In agreement with the results presented above (Fig. 4e), analy sis of cytokine production by cells from the lungs of nonsensitized Csnk2bfl/flFoxp3-Cre mice and sensitized Csnk2bfl/fl mice demon strated similar amounts of the TH2 cell–associated cytokines IL-5 and IL-13 (Fig. 6e,f). Furthermore, cells from sensitized Csnk2bfl/fl Foxp3-Cre mice produced slightly higher concentrations of these cytokines than did those from sensitized Csnk2bfl/fl mice (Fig. 6e,f). In contrast, the TH1 cell–related cytokine IFN-γ was hardly detectable and was secreted in similar amounts by lung cells from both mouse strains regardless of the treatment regimen (Fig. 6g). This spontane ous TH2 differentiation could not be explained by an impaired ability of CK2β-deficient Treg cells to home to the lungs (Supplementary Fig. 4a,b), consistent with the results obtained with tLNs (Fig. 5). Hence, it can be assumed that the TH2 cell–associated lung inflam mation of Csnk2bfl/flFoxp3-Cre mice was based on a reduced ability of CK2β-deficient Treg cells to suppress TH2 responses in vivo. CK2b regulates expression of the receptor ILT3 In searching for the molecular mechanisms underlying the inability of CK2β-deficient Treg cells to control TH2 immune responses, we iso lated Treg cells from the tLNs of Csnk2bfl/fl mice and Csnk2bfl/flFoxp3Cre mice and comparatively analyzed the transcriptomes of these cells by next-generation sequencing–based RNA sequencing (RNA-seq). CK2β-deficient and CK2β-sufficient tLN-resident Treg cells exhibited similar expression of many molecules with a presumed role in the suppressive properties of Treg cells (Fig. 7a). We confirmed that result on the protein level by flow cytometry for the immunomodulatory 269

Articles

34.6

39.1

+

CD62L

CD8

5.3

4.7

e

0

**

40 20 0

20 10

20 10 0

0

fl/fl

50

*

40

0.4

30 20 10

7.9

0.8

0.3 0.2 0.1 0

0

2.5

30

5.8


f

0.6 0.4 0.2 0

fl/fl

11

21.9

GATA-3

T-bet

RORγt

CD62L

71.1 CD44

CD4

CD4

10 0

800 600 400 200

1.5

Csnk2b

0.1 fl/fl

*

1.0 0.5 0

fl/fl

10

5

10

4

0.8

Csnk2b 3 10 Foxp3-Cre 0 2

CD4

fl/fl


**

3

4

0 10 10 10 10 GL-7

5

Figure 4 Treg cell–specific deficiency in CK2β leads to the spontaneous development of TH2 cells. (a) Flow cytometry analyzing the expression of CD62L and CD44 in tLN-resident CD4+ and CD8+ T cells from Csnk2bfl/flFoxp3-Cre mice and their Csnk2bfl/fl littermates. (b) Frequency of CD4+CD62L−CD44+ and CD8+ CD62L−CD44+ effector memory T cells in Csnk2bfl/fl and Csnk2bfl/flFoxp3-Cre mice. Each symbol represents an individual mouse (n = 10 (CD4+ T cells) or 8 (CD8+ T cells) mice per genotype); small horizontal lines indicate the mean (± s.d.). (c) Flow cytometry analyzing the expression of RORγt, T-bet and GATA-3 in CD4+ CD62L−CD44+ T cells (sorting, far left) from the tLNs of Csnk2bfl/fl and Csnk2bfl/flFoxp3-Cre mice. Numbers in outlined areas (far left) indicate percent CD62L +CD44− cells (top left) or CD62L+CD44+ cells (top right) or CD4+CD62L−CD44+ T cells (bottom) analyzed further at right, or (right) percent transcription factor–expressing CD4 +CD62L−CD44+ T cells. (d) Frequency of GATA-3+, T-bet+ or RORγt+ T cells among CD4+Foxp3− T cells from the tLNs of Csnk2bfl/fl and Csnk2bfl/flFoxp3-Cre mice. Each symbol represents an individual mouse (n = 6 per genotype for GATA-3 and RORγt; n = 6 (Csnk2bfl/fl) or 7 (Csnk2bfl/flFoxp3-Cre) for T-bet); small horizontal lines indicate the mean (± s.d.). (e) Enzymelinked immunosorbent assay of IL-4, IFN-γ and IL-17A in culture supernatants of CD4+CD62L−CD44+ T cells sorted by flow cytometry from Csnk2bfl/fl and Csnk2bfl/flFoxp3-Cre mice and stimulated for 72 h with plate-bound anti-CD3 and anti-CD28. (f) Titer of total IgE in serum from Csnk2bfl/fl and Csnk2bfl/flFoxp3-Cre mice. (g) Frequency of GL-7+Fas+ germinal center (GC) B cells (B220+) in the spleen of Csnk2bfl/fl and Csnk2bfl/flFoxp3-Cre mice. Each symbol (f,g) represents an individual mouse (n = 8 per group); small horizontal lines indicate the mean (± s.d.). (h) Flow cytometry analyzing the expression of Fas and GL-7 in splenic B220+MHCII+ B cells from Csnk2bfl/fl and Csnk2bfl/flFoxp3-Cre mice. Numbers in outlined areas indicate percent GL-7+Fas+ cells among B220+MHCII+ B cells. *P < 0.01 and **P < 0.001 (unpaired t-test). Data are representative of three independent experiments (a,c,e; mean and s.d. of technical replicates in e) or two independent experiments (h) or are pooled from two independent experiments (b,d,f,g).

receptor CTLA-4 (CD152), the serine protease granzyme B, the ATP ectonucleotidase CD39, the extracellular AMP nucleotidase CD73 and the IL-2 receptor complex (Supplementary Fig. 5a–c). Accordingly, CK2β-deficient Treg cells showed an unaltered ability to suppress the proliferation of cocultured Teff cells (Supplementary Fig. 5d).

a

fl/fl

50

Csnk2b Csnk2bfl/flFoxp3-Cre

b 10

30 20 10 0

fl/fl

Csnk2b Foxp3-Cre

fl/fl

Csnk2b 105

40

29.5

29

4 3

Foxp3

Foxp3+ cells (%)

npg


fl/fl

Csnk2b Foxp3-Cre

+

3.6

GL7+Fas GC B cells (%)

4.9

20

h Csnk2b fl/fl Csnk2b Foxp3-Cre

22.3

30

0

g

19.8

fl/fl


1,000

fl/fl

Csnk2b

40

+

30

50

RORγt cells (%)

+

20

60

24.7

–

c

40

80

CD44

74.7

60

40

Fas

22.3

71.4

71.1

80

50

40

Serum IgE titer

19.8

100

**

50

+

CD4

d

IL-4 (ng/mL)

+

+ CD62L–CD44+CD4 effector memory T cells (%)

4.9

+

22.3

+

4.7

74.7

**

T-bet cells (%)

fl/fl

Csnk2b Foxp3-Cre

IL-17 (ng/ml)

fl/fl

CD62L CD44 CD8 effector memory T cells (%)

Csnk2b

fl/fl


IFN-γ (pg/ml)

b

GATA-3 cells (%)

a

10 0 4

3 0 10 10 10 CD4

5

Figure 5 Ablation of Csnk2b in Treg cells has no effect on lung-homing ability. (a) Frequency of Foxp3+ Treg cells among in CD4+ T cells from the tLNs of Csnk2bfl/flFoxp3-Cre mice and their Csnk2bfl/fl littermates, assessed by flow cytometry. Each symbol represents an individual mouse (n = 6 (Csnk2bfl/fl) or 7 (Csnk2bfl/flFoxp3-Cre)); small horizontal lines indicate the mean (± s.d.). (b) Flow cytometry as in a. Numbers adjacent to outlined areas indicate percent Foxp3+ cells among CD4+ T cells. Data are representative of two independent experiments.

270

However, we observed completely different activity in CK2βdeficient Treg cells versus CK2β-sufficient Treg cells in their expression of Lilrb4, which encodes the inhibitory receptor ILT3 (also known as gp49B or CD85k): there was significantly higher expression of Lilrb4 (roughly 20-fold higher) in tLN-resident Treg cell from Csnk2bfl/ flFoxp3-Cre mice than in T cells from Csnk2bfl/fl mice (Fig. 7b). reg Flow cytometry with an antibody to gp49 to detect ILT3-expressing (ILT3+) Treg cells demonstrated a substantially elevated percentage of this Treg cell subpopulation especially in the spleen, mesenteric LNs (Supplementary Fig. 6a) and tLNs (Fig. 7c,d) of Csnk2bfl/flFoxp3-Cre mice. ILT3+ Treg cells were even detectable in thymus of both Csnk2bfl/fl Foxp3-Cre and Csnk2bfl/fl mice (Supplementary Fig. 6a), which suggested that the ILT3+ Treg cells represented an as-yet-undefined discrete subpopulation of Treg cells whose development was under the control of CK2. To exclude the possibility that the ILT3 expres sion was a consequence of TH2 cell–mediated inflammation and to demonstrate that it was instead a marker of a discrete subpopulation of Treg cells, we analyzed the ability of various helper T cell–skewing conditions to alter ILT3 expression on Treg cells. Neither TH1 conditions nor TH2 conditions were able to alter the expression of ILT3 on Treg cells upon 72 h of stimulation (data not shown). Pharmacological inhibition of CK2 with either of two different inhibitors VOLUME 16 NUMBER 3 MARCH 2015 nature immunology

npg

fl/fl Csnk2b (PBS) fl/fl Csnk2b Foxp3-Cre (PBS) Csnk2bfl/fl (OVA in alum) Csnk2bfl/flFoxp3-Cre (OVA/alum)

*** *** ***

4

c PAS+ cells per mm bronchius

a

3

1

fl/fl

fl/fl

(PBS)

Csnk2b Foxp3-Cre (PBS)

Csnk2bfl/fl (PBS)

80

Csnk2bfl/flFoxp3-Cre (PBS)

60

Csnk2bfl/fl (OVA in alum)

40

Csnk2bfl/flFoxp3-Cre (OVA in alum)

20 0

***

6

0

bCsnk2b

100

*** *** ***

e

2

IL-13 (ng/ml)

Lung inflammation score

** Csnk2bfl/fl (PBS) Csnk2bfl/flFoxp3-Cre (PBS)

4

Csnk2bfl/fl (OVA/alum) Csnk2bfl/flFoxp3-Cre (OVA in alum)

2 0

fl/fl

Csnk2b (OVA in alum)

d Csnk2b

fl/fl

(PBS)


Csnk2bfl/flFoxp3-Cre (PBS)

*

f

6 IL-5 (pg/ml)

Figure 6 Massive asthma-like airway pathology in Csnk2bfl/flFoxp3-Cre mice without provocation. (a) Histopathology scores for lung sections from Csnk2bfl/fl and Csnk2bfl/flFoxp3-Cre mice left unsensitized (PBS) or sensitized with OVA emulsified in aluminum hydroxide (OVA in alum) and then challenged with OVA. (b) Hematoxylin and eosin–stained lung sections from mice as in a. Arrows indicate inflammatory cell infiltration. Original magnification, ×100. (c) Mucus production in mice as in a, assessed by staining of bronchial cells with periodic acid–Schiff (PAS). (d) Periodic acid–Schiff–stained lung sections from mice as in a. Arrows indicate mucus-producing cells. Original magnification, ×100. (e–g) Total IL-13 (e), IL-5 (f) and IFN-γ (g) in culture supernatants of lung and tLN cells isolated from mice as in a, then pooled and cultured for 72 h without further stimulation. ND, not detectable. Each symbol (a,c,e–g) represents an individual mouse; small horizontal lines indicate the mean (± s.d.). *P < 0.05, **P < 0.01 and ***P < 0.001 (unpaired t-test). Data are representative of two independent experiments with five mice per group.

*

Csnk2bfl/fl (PBS) Csnk2bfl/flFoxp3-Cre (PBS)

4

Csnk2bfl/fl (OVA/alum)

2


0

ND

g of CK2 resulted in substantial upregula 6 Csnk2bfl/fl (PBS) fl/fl tion of ILT3 expression upon stimulation of Csnk2b Foxp3-Cre (PBS) fl/fl fl/fl Csnk2b Csnk2b Foxp3-Cre 4 ILT3− Treg cells isolated from the spleen of Csnk2bfl/fl (OVA/alum) (OVA in alum) (OVA in alum) Csnk2bfl/flFoxp3-Cre C57BL/6 wild-type mice (Supplementary 2 (OVA in alum) Fig. 6b). This result supported the proposal of an intrinsic role for CK2 in the regulation 0 of ILT3 expression. To study fluctuation of the ILT3-expressing Treg cell subpopulation of allergic patients (Fig. 7e,f) but not that of patients with malignant noted above during typical TH2 immune responses, we analyzed the stage IV melanoma (data not shown), which suggested a correlation frequency of ILT3+ Treg cells in the lungs of asthmatic C57BL/6 wild- between the presence of a comparatively high percentage of ILT3+ Treg type mice as well as in the periphery of BALB/c wild-type mice upon cells and an inability to control TH2 cell–driven immune responses infection with the helminth Litomosoides sigmodontis. We detected in mice and humans. Exactly how CK2 might regulate ILT3 expression needs further substantially more of ILT3+ Treg cells in both experimental mouse models (Supplementary Fig. 6c,d), which demonstrated a positive analysis. Many reports have demonstrated that CK2 is a positive correlation between this Treg cell subpopulation and TH2 cell–driven regulator of canonical Wnt signaling that leads to the stabilization of immune responses. Nonetheless, we made use of a mixed–bone mar β-catenin and its translocation to the nucleus9,13–16. Notably, 1,25row (BM) chimera system to investigate the possibility that the exces dihydroxyvitamin D3 (vitamin D3) has been described as having a sive TH2 cell–mediated lung inflammation observed in Csnk2bfl/fl strong inhibitory effect on canonical Wnt signaling17 and as positively Foxp3-Cre mice might have caused the expression of ILT3 or the affecting ILT3 expression18. Hence, we investigated the relationship proliferation of ILT3+ Treg cells. For this, we transferred CD90.1+ between canonical Wnt signaling and CK2β deficiency that led to Csnk2bfl/fl and CD90.2+ Csnk2bfl/flFoxp3-Cre BM cells together ILT3 expression in Treg cells. Ex vivo analysis of canonical Wnt sig into irradiated C57BL/6 recipient mice deficient in recombination- naling in Treg cells revealed reduced expression of β-catenin in the activating gene 1. While the resulting chimeras did not show any signs absence of CK2β (Fig. 8a). To further analyze the effect of Wnt sig of TH2 lung inflammation (Supplementary Fig. 6e,f) or activation naling on ILT3 expression, we stimulated CK2β-deficient Treg cells and proliferation of peripheral CD62LloCD4+ T cells (Supplementary for 48 h in the presence of the Wnt ligand Wnt1. Flow cytometry Fig. 6g), we observed a greater frequency of ILT3-expressing Treg cells revealed a negative correlation between the Wnt1-mediated induc solely among the CD90.2+ Treg cells originating from Csnk2bfl/flFoxp3- tion of β-catenin and the expression of ILT3 (Fig. 8b). Although Treg Cre BM cells (Supplementary Fig. 6h). These results challenged the cells with high β-catenin expression had low ILT3 expression, Treg proposal of induction of ILT3 expression in Treg cells solely under cells with comparatively low expression of β-catenin had the highest TH2 cell–driven inflammation and instead demonstrated that dis expression of ILT3 (Fig. 8b), indicative of CK2β-steered regulation equilibrium between ILT3+ Treg cells and ILT3− Treg cells resulted in of ILT3 expression by β-catenin. excessive TH2 cell–driven immune responses. To translate our find With its cytoplasmic immunoreceptor tyrosine-based inhibitory ings into human TH2 cell–driven diseases, we comparatively analyzed motifs, ILT3 most probably transduces negative signals that attenuate the frequency of ILT3+ Treg cells in the peripheral blood of healthy TCR signaling via dephosphorylation of the signaling kinase Zap70 donors and allergic patients classified by a blood test for the detection by the tyrosine phosphatases SHP-1 and SHP-2 (refs. 19,20). To test of allergen-specific IgE. Regardless of classification, we detected a this hypothesis, we made use of the Nur77GFP reporter mouse strain, comparatively high frequency of ILT3+ Treg cells in peripheral blood in which expression of green fluorescent protein (GFP) is under IFN-γ (ng/ml)


Articles

nature immunology VOLUME 16 NUMBER 3 MARCH 2015

271

Articles

150

Csnk2b Csnk2bfI/fI Foxp3-Cre

100 50

II2 ra C tla En 4 tp d1 N t5 e Eb i3 II1 0

0

10

3.8

17.1

4

10

103

1

0 3

–4

–3

–2 –1 0 1 2 Expression (log2 fold)

3

4

4

5

0 10 10 10 Foxp3

d

Csnk2bfI/fI Csnk2bfI/fI Foxp3-Cre

*

50 40 30 20 10 0

f

*

control of the promoter of the gene encoding the nuclear hormone receptor Nur77 (Nr4a1; called ‘Nur77’ here)21. In these mice, the amount of GFP expressed during activation reflects the strength of TCR stimulation. Flow cytometry–based isolation of ILT3+ and ILT3− Treg cells from Nur77GFP mice and subsequent polyclonal stimulation with antibody to the invariant signaling protein CD3 (anti-CD3) and antibody to the costimulatory molecule CD28 (anti-CD28) revealed considerably attenuated TCR signaling in ILT3+ Treg cells, as evident by reduced GFP expression (Fig. 8c,d). The receptors Nr4a1, Nr4a2 and Nr4a3 are involved in the develop ment and function of Treg cells of thymic origin22. Genetic ablation of all receptors of the Nr4a family leads to the development of fatal T H2 cell–driven inflammation. Hence, we analyzed the expression of all three Nr4a receptors in ILT3+ and ILT3− Treg cells. Consistent with the results obtained with Nur77GFP mice, upon in vitro stimulation, the expression of all Nr4a receptors was lower in ILT3+ Treg cells than in ILT3− Treg cells (Fig. 8e). Physiologically, such a dampening of Treg cell activation is required when a fulminant immune response is essential—for example, in the early phase of an antimicrobial immune response. Thus, these data suggested that ILT3+ Treg cells might repre sent an as-yet-undescribed subpopulation with unique properties. ILT3+ Treg cells promote PD-L2+IRF4+ dendritic cells Two independent groups have shown that T H2 differentiation in vivo is regulated by a specialized dendritic cell (DC) subpopula tion that expresses the ligand PD-L2 and the transcription factor IRF4 (refs. 23,24). Therefore, we comparatively analyzed the expres sion of IRF4 in DCs from Csnk2bfl/flFoxp3-Cre and Csnk2bfl/fl mice as well as the percentage of DCs that coexpressed PD-L2 and IRF4 in these mice. IRF4 expression in DCs, as well as the frequency of 272

ILT3+ cells (%)

Figure 7 CK2 regulates the * Healthy Healthy CAP3 CAP4 CAP5 + CAP6 development of ILT3+ Treg cells. CAP3 5 * 10 (a) RNA-seq analysis of the 30 1.0 12.7 14.8 9.08 CAP4 expression of genes encoding 104 CAP5 + CAP6 3 molecules with a presumed role in 10 20 2 Treg cell–mediated suppression, 10 0 assessed in Foxp3-GFP+ Treg cells 10 0102 103 104 105 sorted by flow cytometry from the Foxp3 fl/fl tLNs of Csnk2b Foxp3-Cre mice and 0 their Csnk2bfl/fl littermates. (b) Geneexpression profile of Treg cells isolated from the tLNs of Csnk2bfl/fl and Csnk2bfl/flFoxp3-Cre mice (plotted as difference in expression versus P value) for genes downregulated (negative expression values) or upregulated (positive expression values) in Csnk2b-deficient Treg cells relative to their expression in Csnk2b-sufficient Treg cells; red indicates genes with difference in expression of at least twofold and a P value of 90% Foxp3+ T cells were used. For the use in in vitro coculture assays or in vivo experiments, ILT3+, ILT3−, CK2β-sufficient or CK2β-deficient CD4+CD25hi Treg cells were purified from the pre-enriched CD4+CD25+ cell fraction with a FACSAria II cell sorter (BD) to a purity of >95%. CD4+CD62L− effector memory T cells, CD4+CD62L+ naive T cells and CD4+CD25− T cells were isolated from splenocytes by positive selection with MACS separation (Milteny Biotech) as described42. CD4+CD62L− effector memory T cells were isolated for the analysis of cytokine production. CD4+CD62L+CD44− naive T cell were used for the dif ferentiation of TH1 or TH2 cells in vitro. Therefore, CD4+CD25− T cell frac tions underwent enrichment as described above. Subsequently, CD4+CD62L− effector memory T cells or CD4+CD62L+CD44− naive T cells were purified from the pre-enriched fraction with a FACSAria II (BD) to a purity of >98%. To obtain single-cell suspensions, lungs were isolated, and minced and then were exposed in a water bath at 37 °C to enzymatic digestion by 0.5 mg/ml collagenase type IA (C9891; Sigma-Aldrich) in PBS. After 1 h of incubation, single-cell suspensions were achieved by pushing of the digested lung tissue fragments through a 0.9- × 40-mm canula (BD Microlance) and a 70-µm nylon cell strainer. Residual erythrocytes were removed with Gey’s lysis buffer. Human CD4+ T cells and CD25+ Treg cells were isolated from leukapheresis products (up to 1.5 × 1010 whole cells) of healthy volunteers as described before43; this resulted in a purity of ≥95% CD4+CD25hi T cells. In vitro stimulation. T cells or Treg cell populations were cultured in Iscove’s modified Dulbecco’s medium (Sigma), supplemented with 5% FCS, 1% penicillin-streptomycin, 1% l-glutamine, sodium pyruvate and 50 µM β-mercaptoethanol. Cells were stimulated for various times with Dynabeads mouse CD3/CD28 T Cell Expander beads (Invitrogen) or plate-bound antiCD3 and anti-CD28 (4 µg/ml each; Supplementary Table 1). Where indi cated Wnt1 (400 ng/ml; Biovision) was added. The pharmacological inhibitors DMAT (2-dimethylamino-4,5,6,7-tetrabromo-1H-benzimidazole; Merck Millipore) and CX-4945 (Selleckchem) were used at the appropriate concen trations. Staining with CFSE (carboxyfluorescein diacetate succinimidyl ester) and proliferation assays were performed as described42. For measurement of the ability of various Treg cell subsets to suppress the differentiation of TH1 or TH2 cells, the following assays were performed. Naive CD4+CD62L+ T cells were isolated from Ly5.1+ C57BL/6 mice as described above (Cell purification) and were stimulated with splenic CD11c+ DCs and soluble anti-CD3 (4 µg/ml; Supplementary Table 1) at a ratio of 1:5 in the pres ence or absence of various numbers of Treg cell populations. TH1 polarization was achieved by the addition of mouse IL-12 (10 ng/ml), human recombinant IL-2 (10 ng/ml) and anti-IL-4 (10 µg/ml; 11B11; Supplementary Table 1). TH2 polarization was achieved by the addition of anti-IFN-γ (10 µg/ml; XMG1.2; Supplementary Table 1). The efficiency of T cell differentiation was analyzed after 4 d of culture in Iscove’s modified Dulbecco’s medium (Sigma). Medium was supplemented with 5% FCS, 1% penicillin-streptomycin, 1% l-glutamine, sodium pyruvate and 50 µM β-mercaptoethanol. For analysis of ex vivo cytokine production after the induction of mouse allergic asthma, cell suspensions from tLNs were prepared as described 44. Cells were cultured for 3 d in Iscove’s modified Dulbecco’s medium (Sigma) supplemented with 10% FCS, 1% penicillin-streptomycin, 1% l-glutamine and sodium pyruvate and 50 µM β-mercaptoethanol without further stimulation. Culture supernatants were collected and cytokine production was measured by specific ELISA. For analysis of the ability of ILT3− and ILT3+ Treg cells to inhibit (suppress) splenic DC maturation, the following suppression assay was performed. CD11c+ DCs were cultured for 16 h together with either ILT3+ ILT3− Treg cells

doi:10.1038/ni.3083

at a ratio of 1:3. To activate DCs and Treg cells, 100 ng/ml lipopolysaccharide and 4 µg/ml anti-CD3 (Supplementary Table 1) was added. For polyclonal activation of human T cells, 1 µg/ml monoclonal antibody (mAb) to CD3 (OKT-3; Supplementary Table 1) and 2 µg/ml mAb to CD28 (CD28.2; BD Pharmingen) were used.

npg


Kinome array. For mouse assays, 5 × 106 mouse CD4+CD25−Foxp3− or CD4+CD25+Foxp3+ T cells were isolated from the spleen of C57BL/6 mice and were stimulated in the presence of anti-CD3 and anti-CD28 (Supplementary Table 1). After 30 min of stimulation, cells were washed twice with ice-cold PBS and cell pellets were frozen at −80 °C. For human kinome array samples, 1 × 107 PBMCs were suspended in 5 ml of IMDM. Stimulations were terminated by a wash in ice-cold phosphatebuffered saline. PBMCs were lysed in 200 ml of cell lysis buffer (20 mm TrisHCl, pH 7.5, 150 mm NaCl, 1 mm Na2EDTA, 1 mm EGTA, 1% Triton X-100, 2.5 mm sodium pyrophosphate, 1 mm MgCl2, 1 mm β-glycerophosphate, 1 mm Na3VO4, 1 mm NaF, 1 µg/ml leupeptin, 1 µg/ml aprotinin and 1 mm PMSF) and the volume of the cell lysate was ‘equalized’ with distilled H2O. PepChip kinase profiling of human and mouse material and subsequent in silico analysis was performed by qualified employees of Pepscope as described45. Immunoblot analysis. After isolation of Treg cells by magnetic-activated cell sorting, 1 × 106 Treg cells were lysed in 50 µl RIPA buffer. Aliquots correspond ing to 3 × 105 cells were fractionated by SDS-PAGE. Subsequently, proteins were transferred on a 0.45-µm PVDF membrane. Afterward, CK2β was analyzed by immunoblot with rabbit anti-CK2β (EP1995Y; Merck Millipore) and horse radish peroxidase–conjugated polyclonal goat anti-rabbit (Supplementary Table 1). Detection with horseradish peroxidase–conjugated anti-β-actin (AC-16; Sigma) served as a loading control. Flow cytometry. Flow cytometry data were acquired on a BD LSR II flow cytometer and were analyzed with BD FACSDiva software (version 6.0) and FlowJo software (TreeStar). Samples were excluded from analysis if 98% for Foxp3+ CD4+ T cells. RNA from Treg cells was isolated with TRIzol (spleen samples) or with an RNeasy Micro Kit (tLN samples; Qiagen). RNA was quantified with a Qubit 2.0 fluorometer (Invitrogen) and RNA quality was assessed on a Bioanalyzer 2100 (Agilent) with an RNA nano chip (spleen) or pico chip (tLNs) (Agilent). Samples with an RNA integrity number of >8 were used for library preparation. 600 ng of total RNA derived from splenic Treg cells was used for library preparation with a TruSeq RNA Prep Kit v2 (Illumina). After the synthesis of cDNA with a SMARTer Kit V1 (Clontech), 1 ng of the resulting cDNA from Treg cells isolated from tLNs as used for library prepara tion with a Nextera XT DNA Sample Prep Kit according to the manufacturer’s instructions (Illumina). Library size distribution was assessed with a High Sensitivity DNA chip (Agilent) on a Bioanalyzer 2100. Sequencing (50–base pair single ‘reads’) was performed on a MiSeq (Illumina) with MiSeq Reagent Kit v2, which resulted in at least 7 × 106 ‘reads’ per sample. Sequencing data were analyzed with CLC Genomic Workbench 6.5 (Qiagen). Statistical analysis. For statistical analysis, Student’s t-test was performed with the software Graph Pad Prism for calculation of the statistical significance

37. Lahl, K. et al. Selective depletion of Foxp3+ regulatory T cells induces a scurfy-like disease. J. Exp. Med. 204, 57–63 (2007). 38. Mittrücker, H.W. et al. Requirement for the transcription factor LSIRF/IRF4 for mature B and T lymphocyte function. Science 275, 540–543 (1997). 39. Tuettenberg, A. et al. Induction of strong and persistent MelanA/MART-1-specific immune responses by adjuvant dendritic cell-based vaccination of stage II melanoma patients. Int. J. Cancer 118, 2617–2627 (2006). 40. Muth, S., Schütze, K., Schild, H. & Probst, H.-C. Release of dendritic cells from cognate CD4+ T-cell recognition results in impaired peripheral tolerance and fatal cytotoxic T-cell mediated autoimmunity. Proc. Natl. Acad. Sci. USA 109, 9059–9064 (2012). 41. Ruedl, C., Rieser, C., Böck, G., Wick, G. & Wolf, H. Phenotypic and functional characterization of CD11c+ dendritic cell population in mouse Peyer’s patches. Eur. J. Immunol. 26, 1801–1806 (1996). 42. Bopp, T. et al. NFATc2 and NFATc3 transcription factors play a crucial role in suppression of CD4+ T lymphocytes by CD4+CD25+ regulatory T cells. J. Exp. Med. 201, 181–187 (2005). 43. Jonuleit, H. et al. Infectious tolerance: human CD25+ regulatory T cells convey suppressor activity to conventional CD4+ T helper cells. J. Exp. Med. 196, 255–260 (2002). 44. Dehzad, N. et al. Regulatory T cells more effectively suppress Th1-induced airway inflammation compared with Th2. J. Immunol. 186, 2238–2244 (2011). 45. Diks, S.H. et al. Kinome profiling for studying lipopolysaccharide signal transduction in human peripheral blood mononuclear cells. J. Biol. Chem. 279, 49206–49213 (2004). 46. Staudt, V. et al. Interferon-regulatory factor 4 is essential for the developmental program of T helper 9 cells. Immunity 33, 192–202 (2010). 47. Powrie, F., Leach, M.W., Mauze, S., Caddle, L.B. & Coffman, R.L. Phenotypically distinct subsets of CD4+ T cells induce or protect from chronic intestinal inflammation in C. B-17 scid mice. Int. Immunol. 5, 1461–1471 (1993). 48. Webster, K.E. et al. In vivo expansion of T reg cells with IL-2-mAb complexes: induction of resistance to EAE and long-term acceptance of islet allografts without immunosuppression. J. Exp. Med. 206, 751–760 (2009). 49. Hartmann, W., Haben, I., Fleischer, B. & Breloer, M. Pathogenic nematodes suppress humoral responses to third-party antigens in vivo by IL-10-mediated interference with Th cell function. J. Immunol. 187, 4088–4099 (2011).

npg


Helminth infection model. BALB/c mice were naturally infected by exposure to L. sigmodontis–infected mites (Ornithonyssus bacoti) as described49. 76 d after infection, ILT3 expression in splenic CD4+Foxp3+ Treg cells was assessed by flow cytometry.

of the mean values. For non-Gaussian distributions, the Mann-Whitney U-test was used for the calculation of statistical significance. Gaussian distri bution was analyzed with the Kolmogorov-Smirnov test. Variance was similar between groups being compared statistically.

nature immunology

doi:10.1038/ni.3083