PNAS PLUS

Adaptive immunity to murine skin commensals Wei Shena, Wenqing Lia, Julie A. Hixona, Nicolas Bouladouxb, Yasmine Belkaidb, Amiran Dzutzevc, and Scott K. Duruma,1 a Laboratory of Immunoregulation and cLaboratory of Experimental Immunology, Cancer and Inflammation Program, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Frederick, MD 21702; and bLaboratory of Parasitic Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD 20814

Edited by Robert L. Coffman, Dynavax Technologies, Berkeley, CA, and approved June 4, 2014 (received for review February 4, 2014)

T lymphocyte

| microbiome

M

ammalian epithelium is host to a myriad of microbial species, and this requires that the epithelium provides barrier functions. The intestinal epithelial barrier is especially complex because, whereas it must be permeable to nutrients, it must also block microbial invasion via mechanical and innate immune mechanisms. Therefore, the adaptive immune system of the gut contributes to blocking microbial invasion and yet has a delicate tolerance mechanism that attenuates reactions to enteric antigens. The skin epithelial barrier, in contrast to the intestinal barrier, would seem to require simpler immunological mechanisms for blocking entry of commensals and pathogens because the skin is relatively impermeable. The innate immune system has been ascribed a role in skin barrier function through release of antimicrobial peptides (1); defects in the innate immune system are implicated in atopic dermatitis (2). The adaptive immune system, in contrast, has received less attention regarding its contribution to barrier function of the skin. Although tolerance would seem less necessary than for the more permeable gut, there is some evidence for tolerance (3); excess T-cell reactions are implicated in several common skin diseases, including atopic dermatitis and psoriasis. In humans deficient in T and B cells, about a third show infectious cutaneous manifestations (4). T and B cell deficiency (TB−) that is specifically due to recombination activating gene (Rag) mutations similarly show a high frequency of cutaneous manifestations (5). These clinical findings implicate adaptive immunity in skin host defense in man. Susceptibility to disseminated mycobacterial infections is often seen in TB− severe combined immunodeficiency patients (SCID) with disseminated bacillus Calmette–Guérin (4). A Rag2-deficient SCID was reported with disseminated bacillus Calmette–Guérin (6) and a Rag1 hypomorphic patient was reported with disseminated nontuberculous www.pnas.org/cgi/doi/10.1073/pnas.1401820111

mycobacteria (7). These reports support a role for adaptive immunity in defense against mycobacteria in man. T cells within normal human dermis are predominantly of an effector memory (TEM) phenotype with a diverse repertoire (8) and have been suggested to be involved in immune surveillance (9). T-cell homing to skin is directed by E selectin and chemokines chemokine (C-C motif) ligand (CCL)20, -22, and -27 expressed by keratinocytes acting on cutaneous lymphocyte-associated antigen (CLA), chemokine receptor (CCR)6, -4, and -10, respectively, expressed by the subset of T cells that home to skin. There is implied interaction between the skin flora and the adaptive immune system, suggested by the defect in Th17 development in mice lacking skin flora (10). The adaptive immune system is implicated in skin barrier function by recent studies suggesting Th17 cells play a significant role in defense against cutaneous pathogens (11). Autoantibodies against Th17 cytokines, including IL-17A and F, IL-22 (12), and mutations in the receptor for IL-17 (13) have been implicated in skin infections. It remains to be determined whether skin “commensals,” i.e., bacteria that are normally not pathogenic, elicit T-cell immunity in a normal host and whether these T cells could contribute to skin barrier function. The present study derives from experiments showing that transfer of OT-1 cells, monoclonal CD8 T cells, into Rag1−/− mice required host IL-7 for survival and proliferation of the transferred cells (14), a phenomenon termed “homeostatic” proliferation. However, our group (15) and others (16) subsequently observed that, unexpectedly, normal polyclonal CD8 T cells transferred into Rag1−/− survived and proliferated independently of IL-7; moreover, their proliferation was much faster than homeostatic proliferation observed with OT-1 cells. This phenomenon, the rapid, IL-7–independent proliferation of normal T cells transferred into Rag1−/− mice, resembled the reaction of memory cells to antigens. Here we investigate this reaction to determine the nature of the Significance Barrier function of the skin in blocking microbial invasion has been attributed to the structural integrity of the epithelium, augmented by innate immune mechanisms. T cells and antigen-presenting cells have long been observed in the skin, but what is their role? Here we report, for the first time, that commensal skin bacteria are recognized by major populations of T cells in skin-draining lymph nodes of mice. We report a previously unrecognized role for T cells in preventing breach of the skin epithelial barrier by certain species of commensal bacteria, especially mycobacteria, and we examine the mechanism. Patients deficient in T cells frequently show infectious cutaneous manifestations and mycobacterial susceptibility, reflecting features of our study in mice. Author contributions: W.S., W.L., A.D., and S.K.D. designed research; W.S., W.L., J.A.H., and A.D. performed research; N.B. and Y.B. contributed new reagents/analytic tools; W.S., W.L., A.D., and S.K.D. analyzed data; and W.S. and S.K.D. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. 1

To whom correspondence should be addressed. Email: [email protected].

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1401820111/-/DCSupplemental.

PNAS | Published online July 7, 2014 | E2977–E2986

IMMUNOLOGY AND INFLAMMATION

The adaptive immune system provides critical defense against pathogenic bacteria. Commensal bacteria have begun to receive much attention in recent years, especially in the gut where there is growing evidence of complex interactions with the adaptive immune system. In the present study, we observed that commensal skin bacteria are recognized by major populations of T cells in skindraining lymph nodes of mice. Recombination activating gene 1 (Rag1)−/− mice, which lack adaptive immune cells, contained living skin-derived bacteria and bacterial sequences, especially mycobacteria, in their skin-draining lymph nodes. T cells from skin-draining lymph nodes of normal mice were shown, in vitro, to specifically recognize bacteria of several species that were grown from Rag1−/− lymph nodes. T cells from skin-draining lymph nodes, transferred into Rag1−/− mice proliferated in skin-draining lymph nodes, expressed a restricted T-cell receptor spectrotype and produced cytokines. Transfer of T cells into Rag1−/− mice had the effect of reducing bacterial sequences in skin-draining lymph nodes and in skin itself. Antibacterial effects of transferred T cells were dependent on IFNγ and IL-17A. These studies suggest a previously unrecognized role for T cells in controlling skin commensal bacteria and provide a mechanism to account for cutaneous infections and mycobacterial infections in T-cell–deficient patients.

postulated stimulatory antigens and to characterize the responding T cells and their physiological role. We will show that much of this T-cell reaction is triggered by skin commensal bacterial antigens and demonstrate a protective role of these T cells in barrier function that requires IFNγ and IL-17. Results T Cells Transferred into Rag1−/− Mice Undergo Antigen-Driven Proliferation. Transfer of T cells from C57BL/6 mice into T-cell–

deficient Rag1−/− mice revealed a prominent population of rapidly dividing cells that had diluted carboxyfluorescein succinimidyl ester to undetectable levels (Fig. 1 A and B). In contrast, hosts acutely depleted of T cells using radiation, did not support rapid T-cell division (15, 16). T cells isolated from skin-draining lymph nodes showed rapid proliferation in the recipients’ skindraining lymph nodes (Fig. 1A). Similarly, T cells isolated from mesenteric lymph nodes showed rapid proliferation in the recipients’ mesenteric nodes (Fig. 1A), consistent with a memory response to skin and enteric antigens. It should be noted that there was no regional specificity for proliferation, in that cells

A Donor LN:

B6 skin Recipient LN: Rag-/- skin

from skin-draining lymph nodes proliferated in mesenteric nodes as well as skin-draining nodes and vice versa; however, as will be shown, cells from skin-draining nodes displayed in vitro recognition of skin bacteria, whereas mesenteric lymph node cells did not. A comparison of the kinetics of proliferation of transferred cells (Fig. 1 A–C) shows the relative rate of homeostatic division of OT-1 cells is much slower than the rapid cell cycling described here. This suggests that rapidly proliferating T cells are undergoing a memory response to antigens, in contrast to OT-1 cells, which are not encountering their nominal ovalbumin antigen and are undergoing homeostatic proliferation in response to cytokines and self-peptides/MHC complexes. To evaluate the hypothesis that the proliferation could be antigen driven, analysis of the spectrotype of T-cell antigen receptor Vbeta of proliferating T cells was performed and compared with unfractionated T cells from lymph nodes before transfer (Fig. 2). Proliferating cells showed marked differences from unfractionated lymph node T cells or spleen T cells for a number of Vbeta families. The area under the curve (AUC) (Fig. 2A) showed highly significant differences in Vbeta 2, 3, 7, 8,

Germ Free skin Rag-/- skin

OT-1 Rag-/- skin

CD8 CFSE Donor LN: B6 skin Recipient LN: Rag-/- mesenteric

B6 mesenteric Rag-/- mesenteric

B6 mesenteric Rag-/- skin

CD8 17.51%

C

9.71%

CD8

B

B6 65.40%

4.16%

7.38%

42.84%

OT-1 49.33%

3.67%

GF

CFSE Fig. 1. Rapid proliferation of T cells transferred into Rag1−/− mice. Skin-draining (axillary, brachial, and inguinal) or mesenteric lymph node cells from C57BL/ 6 mice, GF mice, or OT-1 mice were labeled with CFSE and transferred into Rag1−/− recipients. After 5 d (A and B) or indicated days (C), skin-draining or mesenteric lymph node cells were analyzed. (A) Histograms comparing C57BL/6, GF, and OT-1 cells. (B) Flow cytometry showing CD8 staining vs. CFSE. (C) Frequency of CFSE-negative CD8 T cells. Graphs show mean ± SEM of four mice. **P < 0.01, ***P ≤ 0.001; ns, not significant. Results are representative of three experiments.

E2978 | www.pnas.org/cgi/doi/10.1073/pnas.1401820111

Shen et al.

PNAS PLUS Area under curve(AUC)/control

A

3.5

Proliferating T Cells Unfractionated T cells

3.0

***

2.5 2.0

***

1.5 1.0

***

*** *

*

4

5

***

0.5

*** ***

***

*

*** *** ***

9

10

*** *** ***

**

0.0 1

2

3

6

7

8.1

8.2

8.3

11

12

13

14

15

16

17

18

19

20

TCR Vbeta

B

*** Number of peaks

125

***

100 75

*** *** ***

***

***

***

Proliferating T Cells Unfractionated T cells

***

*** * ***

*** ***

***

*

50

*** ***

*

***

*

25 0 1

2

3

4

5

6

7

8.1 8.2 8.3

9

10

11

12

13

14

15

16

17

18

19

20 control

TCR Vbeta

Fig. 2. Diversity of TCR Vβ repertoire in proliferating T cells. Proliferating T cells were sorted from mouse Rag1−/− recipients 5 d after lymphocyte T-cell transfer, whereas unfractionated T cells were prepared from C57BL/6 skin-draining lymphocytes. Following isolation of total RNA from proliferating and unfractionated T cells, spectrotype of 22 mouse TCR Vβ gene families was determined by an Applied Biosystems 3730xl DNA Analyzer. (A) Total AUC for PCR fragments of individual Vβ gene families relative to positive control. (B) Number of peaks for PCR fragments of each TCR Vβ family members (n = 3, data represents mean ± SEM *P < 0.05, **P < 0.01, ***P ≤ 0.001, determined by t test). Results represent two independent experiments.

Rapidly Proliferating T Cells in Skin-Draining Lymph Nodes Are Effector and Effector Memory Cells. The rapidly proliferating T cells in skin-

draining lymph nodes were of heterogeneous phenotypes whether transferred cells were derived from skin-draining lymph nodes (Fig. 3), mesenteric lymph nodes (Fig. S1), or spleen (Fig. S2). CD4 and CD8 cells represented the majority of the rapidly dividing population, and most γδ T cells, although a minor population, also divided rapidly as did IgM+ B cells. Most rapidly dividing T cells expressed the memory marker CD44 and some expressed Sca-1. Cytokine receptors for IL-2 (CD25 and CD122) and IL-7 (CD127) were expressed on some rapidly proliferating cells, but few expressed the receptor for FasL (CD95). Chemokine receptors (CCR4, -6, -7, and -10) were expressed on some rapidly proliferating cells derived from skin-draining lymph nodes, but CCR7 was not expressed on cells derived from mesenteric lymph node or spleen. The selectin CD62L, which directs homing to lymph nodes was expressed on a large fraction of rapidly dividing cells, whereas the selectin CD62E, which directs homing to epidermis was not expressed. These markers are generally consistent with the phenotypes of effector and effector memory cells, although other populations may also be included as will be discussed. All of the markers observed on rapidly proliferating cells were expressed on subpopulations of lymph node T cells before transfer into the Rag1−/− recipient (Fig. S3), although there was enrichment for some markers, most notably chemokine receptors. Because some rapidly proliferating T cells in skin-draining lymph nodes had features of antigen-driven effectors with a skinhoming (CCR10) marker, we examined expression of several cytokines that could be involved in antimicrobial responses of the skin (Fig. 4). IL-17A, expressed from a reporter construct, was observed in the rapidly proliferating population, whereas IL-17F was not expressed. IFNγ and IL-5, visualized by intracellular staining, were highly expressed in some rapidly proliferating cells. IL-22, visualized from a reporter construct, showed a difShen et al.

ferent pattern, being expressed in both slowly dividing and rapidly dividing populations. Culturable Bacteria in Skin-Draining Lymph Nodes Are Recognized by T Cells. It therefore appeared that the T cells isolated from nor-

mal mice could be proliferating and producing cytokines in response to antigens present in Rag1−/− lymph nodes. Because immunodeficient HIV patients were shown to have bacterial translocation across the gut epithelial barrier (17), we considered the possibility that such barrier function might also fail in the gut of Rag1−/− mice. We observed that mesenteric lymph nodes of WT mice contained living bacteria that could be grown in culture (Fig. 5A); however, Rag1−/− mice did not contain more than WT in mesenteric nodes. To our surprise, skin-draining lymph nodes from Rag1−/− mice contained many more living, culturable bacteria than WT mice. The living bacteria did not appear to be associated with host cells, because they were not associated with a cell pellet following centrifugation of the lymph node cell suspension. The spleens from Rag 1−/− mice (Fig. 5A), and other organs including heart, lung, and kidney, did not contain culturable bacteria. Because only a small percentage of microbiome bacterial species can be grown in culture, we analyzed bacterial 16s ribosomal RNA sequences in these tissues (Fig. 5 B and C), which also showed a substantial difference between Rag1−/− versus WT skin-draining lymph nodes. These sequences could reflect both living bacteria as well as dead bacteria that might have been killed locally or at a distance and transported to the lymph node. The skin of Rag1−/− mice also contained substantially more bacterial sequences than WT (Fig. 5B). Rag1−/− mice are deficient in both T and B cells; however, B-cell–deficient mice (MuMT) did not contain culturable bacteria or 16s rRNA bacterial sequences (Fig. 5 A and B). Thus, T-cell deficiency accounted for observed differences between Rag1−/− and WT. The differences between Rag1−/− and WT mice was not attributable to different flora from different colonies because the two strains were housed together for 2 mo before these experiments were performed. Analysis of richness of bacterial species populating the skin of Rag1−/− and WT mice by sequencing of 16S rRNA amplicon we also found that the number of species populating the skin was higher in Rag1−/− mice. PNAS | Published online July 7, 2014 | E2979

IMMUNOLOGY AND INFLAMMATION

10–12, 15–17, 19, and 20. The number of peaks (Fig. 2B) showed highly significant differences in Vbeta 1–7, 8.1, 8.3, 10–12, 15–18, and 20. This is consistent with a selected repertoire for diverse antigens represented in the proliferating population.

Major Subsets

24.28%

21.88%

51.55%

14.83%

4.10%

0.56%

23.70%

85.40%

9.94%

0.66%

0.08%

18.72%

3.43%

89.88%

9.38%

70.69%

7.17%

15.94%

7.96%

22.77%

1.57%

55.55%

20.56%

69.88%

5.78%

1.84%

74.26%

0.20%

IgM

29.92%

11.74%

TCRγδ γδ

12.70%

CD8

CD4

33.10%

Cytokine Receptors

7.23%

31.35%

3.14%

CD127

10.12%

58.28%

CD95

81.20%

0.38%

CD122

CD25

8.30%

Chemokine Receptors

5.72%

17.02%

50.24%

10.21%

CCR6

CCR10

55.81%

22.53%

CCR7

2.71%

CCR4

35.76%

Selectins

3.26%

0.91%

89.04%

7.51%

7.20%

9.05%

1.34%

CD44

82.41%

21.77%

17.98%

Sca-1

13.76%

2.55% CD62e

20.31%

CD62L

62.68%

Memory Markers

1.02%

59.24% 2

0

2

3

4

5

CFSE Fig. 3. Heterogeneous phenotypes of rapidly proliferating T cells in skin-draining lymph nodes. Five days after transfer of T cells from C57BL/6 skin-draining lymph nodes into Rag1−/− recipients, cells from skin-draining lymph nodes were analyzed by flow cytometry. CD4, CD8, CD25, CD44, CD62, CD95, CD122, Tcrγδ, Sca-1, CCR7, and CCR10 are shown. Data are representative of three or more experiments.

To evaluate further if skin bacteria drive the observed T-cell proliferation, T cells from germ-free (GF) mice were transferred into Rag1−/− mice and no significant rapid proliferation was observed (Fig. 1 A and B). Antibiotics were used to deplete gut commensals, but sparing skin commensals, in both donor and recipient; this did not impair the proliferative response in skindraining lymph nodes (Fig. S4). The living bacterial colonies grown from Rag1−/− lymph nodes were analyzed and compared with living colonies grown from other sites (Table 1 and Fig. 6). Mycobacteria were the most frequent colonies grown from lymph nodes, and although found in skin of both C57BL/6 and Rag1−/−, they were not the most frequent. This suggested that loss of adaptive immunity resulted in a selective defect in preventing penetration of mycobacteria through the skin barrier, whereas defense against staphylococci, the most common skin flora in these mice, was relatively intact. We next evaluated whether the T-cell proliferation observed in lymph nodes in vivo (Figs. 1–4) could be accounted for by bacterial antigens in the lymph nodes. Individual bacteria colonies were grown from Rag1−/− lymph nodes or skin and identified by 16s rRNA typing. Bacteria were heat-killed and added to cultures of lymph node cells supplemented with dendritic cells (DCs). Vigorous proliferative responses of CD8 cells were observed against corynebacteria, staphylococci, mycobacteria, and enterococci (Fig. 7A). T cells from germ-free mice did not proliferate in response to bacteria (Fig. 7B). Moreover, T cells from mesenteric lymph nodes and spleens did not recognize skin bacteria in vitro (Fig. S5 A and B). CD4 cells mounted less vigorous responses to bacteria as well as to the polyclonal stimE2980 | www.pnas.org/cgi/doi/10.1073/pnas.1401820111

ulus phytohemagglutinin (PHA). Therefore, antigens from these bacteria could account for much of the T-cell proliferative response observed in vivo, at least for the CD8 subset. IFNγ Rescues the Elimination of the Bacteria in Skin or Skin-Draining Lymph Nodes. Rag1−/− mice were then reconstituted with normal

lymph node T cells from WT mice to determine whether the bacterial sequences in skin-draining lymph nodes could be reduced to the levels found in WT lymph nodes. Following T-cell transfer, a steady decline in bacterial sequences was observed in skin-draining lymph nodes, including axillary (Fig. 8, Top Left), brachial, and inguinal (Fig. 8, Top Right) nodes. There was also a modest decline in bacterial sequences from the skin (Fig. 8, Top Left). No decline was observed in mesenteric lymph node (Fig. 8, Top Right). Because Rag1−/− mice lack both T and B cells, we tested whether deleting B cells from the transferred lymph node population had an effect, but there was no impairment observed in reduction of bacterial sequences (Fig. 8, Middle and Bottom Left and Right). The mechanism by which T cells control bacteria could involve cytokines that are known to activate innate immune functions. Several candidate cytokines, IFNγ, IL-17A, and IL-22, were produced by T cells following transfer into Rag1−/− mice (Fig. 4). IFNγ activates microbicidal functions in macrophages (18), IL-17 can activate neutrophils and induce β-defensin 2 production from keratinocytes (19), and IL-22 can induce production of antimicrobial peptides by keratinocytes (20). IFNγ−/− T cells were unable to reduce bacterial sequences in skin-draining lymph nodes (Fig. 8, Middle and Bottom Left) or in skin (Fig. 8, Shen et al.

PNAS PLUS

A

Donor from skin-draining lymph nodes

64.87%

9.78%

4.14%

0.22%

68.55%

27.09%

0.50%

1.68%

75.42%

22.40%

IL-17F

1.86%

1.54%

IL-22-tdtomato

80.94%

23.81%

IL-5

0.26%

IFNγγ

IL-17A-tdtomato

16.94%

0.09%

1.08%

44.95%

53.88%

CFSE

B

Donor from mesenteric lymph nodes

43.06%

23.52%

73.38%

0.32%

0.28%

0.49%

66.19%

33.04%

1.03%

63.22%

35.44%

IL-17F

53.01%

0.45%

2.65%

IL-22-tdtomato

37.89%

1.17%

IL-22-tdtomato

58.20%

2.76%

IL-5

1.04%

IFNγγ

IL-17A-tdtomato

2.87%

0.22%

0.51%

68.76%

30.51%

CFSE

C

Donor from spleen

52.57%

1.15%

43.81%

2.38%

0.42%

73.13%

24.07%

IL-17F

18.95%

2.47%

IL-5

77.08%

0.15%

IFNγγ

IL-17A-tdtomato

3.81%

0.08%

0.50%

41.36%

58.07%

CFSE

Fig. 4. Cytokine expression in rapidly proliferating T cells in skin-draining lymph nodes. Five days after transfer of T cells from C57BL/6 into Rag1−/− recipients, intracellular cytokines were examined by flow cytometry. IL-17A and IL-22 were detected from reporter T cells, whereas IFNγ, IL-5, and IL-17F were detected by intracellular staining. Donor T cells were prepared from skin-draining lymph nodes (A), mesenteric lymph nodes (B), and spleens (C). Results are representative of two experiments.

Shen et al.

B ***

Mu MT C57Bl/6 Rag1-/-

*** 10

***

p=0.25 ns

ns

5

0

ALN

ns

ns

BLN

ILN

MLN spleen

C

9 8 7 6 5 4 3 2 1 0

*

**

Mu MT C57Bl/6 Rag1-/-

** ** ns

ALN

ns

ns

BLN

ILN

ns ns

ns

MLN Skin

* 70 60 50 40 30 20 10 0

Rag1-/- skin

C57Bl/6 skin

Fig. 5. Detection of live bacteria and bacterial sequences in skin-draining lymph nodes of Rag1−/− mice. (A) Skin-draining lymph node (axillary, ALN; brachial, BLN; and inguinal, ILN), mesenteric lymph node (MLN), and spleen suspensions were prepared from Rag1−/− mice, MuMT, or normal C57BL/6 mice (WT). Live bacteria were cultured on media plates and total colony numbers per organ from individual mice were counted. Data are expressed as mean ± SEM (n = 15 in each group). Data show two experiments. (B) 16s rRNA quantification of bacterial sequences in preparations of skin, skindraining, or mesenteric lymph nodes. Geometric means of relative quantification of total microbiome are shown (n = 8 in each group). Results are representative of two experiments. (C) Chao analysis of bacterial species richness in skin of Rag1−/− and C57BL/6 mice. Data were obtained by 454 sequencing; mean values and SD bars are shown. *P < 0.05, determined by t test. Data show two experiments.

PNAS | Published online July 7, 2014 | E2981

IMMUNOLOGY AND INFLAMMATION

15

16srRNA quantification relative to C57BL6 mice

A

Chao Index

Discussion Whereas much recent interest has focused on gut commensal interaction with the adaptive immune system, skin flora have received less attention. We show for the first time, to our knowledge, that a major proportion of T cells in skin-draining lymph nodes of normal mice are reactive to commensal skin bacteria. These T cells were dramatically reactivated following transfer into Rag1−/− mice or upon stimulation with commensal bacteria in vitro. T cells were shown to contribute to the barrier function of the skin, at least in part, via IFNγ and IL-17A. A balance of immunity and tolerance is observed in the gut immune system: although tolerance mechanisms are dominant, antibodies to gut commensals are detected in healthy humans (21). In skin, a similar balance may exist between antigen recognition and tolerance. Epidermal Langerhans cells may induce T-cell tolerance due to inefficient presentation of antigens from skin flora, whereas dermal DCs are stimulatory (3, 22). Thus, commensals restricted to the surface of the skin may maintain tolerance, whereas bacteria penetrating the epithelial basement membrane to the dermis may induce an immune response. According to this model, our detection of commensal-reactive T cells in lymph nodes suggests that T cells encountered antigens presented by dermal DCs, perhaps after minor wounding. We show that IFNγ and IL-17A are involved in excluding live bacteria from lymph nodes, but it remains to be determined how

much of this protection occurs at the level of the epithelium, or in the node itself. Because the skin of Rag1−/− mice showed increased live bacteria, bacterial sequences and diversity (by Chao index) compared with WT, it suggests that the adaptive immune system exerts some protection on, or in, the epithelial

Colony number

Middle and Bottom Right). Deletion of IL-17 showed a partial defect in lymph nodes (Fig. 8, Middle and Bottom Left) and a substantial defect in skin (Fig. 8, Middle and Bottom Right). Deletion of IL-22 showed no defect in lymph nodes (Fig. 8, Middle and Lower Left) or skin (Fig. 8, Middle and Bottom Right). Therefore, T-cell production of IFNγ and IL-17, but not IL-22, were required to prevent bacterial translocation of these commensals into skin-draining lymph nodes, and their absence likely contributed to the barrier defect observed in Rag1−/− mice lacking adaptive immunity.

Table 1. Bacteria cultured from mouse tissues and identified by 16s rRNA PCR fragment Tissues Most frequent bacteria species

Skin

Axillary LNs

Brachial LNs

Staphylococcus Corynebacterium Mycobacterium Rhodopseudomonas Gordonia Bacillus Streptococcus

Mycobacterium Rhodopseudomonas Rothia Methylobacterium Arthrobacter

Mycobacterium Corynebacterium Methylobacterium Rhodopseudomonas Arthrobacter

Inguinal LNs

Small intestine

Large intestine

Mesenteric LNs

Mycobacterium Lactobacillus Enterococcus Lactobacillus Rhodopseudomonas Corynebacterium Lactobacillus Corynebacterium Corynebacterium Arthrobacter Staphylococcus Rhodopseudomonas Streptococcus Bacillus

LNs, lymph nodes.

surface; this was verified by transferring T cells into Rag1−/− mice resulting in bacterial reduction. T cells from normal donor mice used in these studies did not become activated until transfer into Rag1−/− mice (Figs. 1–4) or in vitro upon exposure to commensal bacterial antigens (Fig. 7). The unactivated state in the normal lymph node may be explained by a relatively low level of commensal antigens in the normal lymph node, compared with the “leaky” Rag1−/− epithelium. Microbial translocation in the gut has been well characterized in CD4-depleted humans with HIV (17) and monkeys with simian immunodeficiency virus infections (23). Our study suggests that the adaptive immune system also contributes to the skin barrier. Rag1−/− mice, in addition to harboring more bacteria, also lack Tregs, which may attenuate the T-cell responses in situ in normal mice. TGFβ and Langerhans cells have been suggested to promote regulatory T cells in skin (22) and may contribute to maintaining T-cell tolerance in situ. It has also been reported that migratory dendritic cells from dermis can induce regulatory T cells in skin-draining lymph nodes (24). The heterogeneous T-cell populations we observed reacting to commensals did not express the markers coinciding with most well-characterized T-cell subsets. Some markers resembled those of TEM cells, which have been identified for successive infections (25–30), rather than central memory (TCM) cells, which are favored by antigens that are cleared (31). Nevertheless, unlike

Rag1-/- skin-lymph nodes

TEM, the reactive populations we detected were found in lymph nodes, where TCM cells preferentially reside, and most of these proliferating cells expressed the selectin CD62L that is also expressed on TCM cells and directs migration to lymph nodes. Expression of chemokine receptors on the rapidly proliferating cells may provide a clue to their migration patterns. CCR7 is expressed on TCM and not TEM cells. CCR7 is the receptor for CCL21 expressed on lymph node endothelium (32) and was expressed on about a fifth of rapidly proliferating cells if the transferred cells derived from skin-draining lymph nodes. It is possible that a higher frequency of these cells expressed CCR7 before activation, although it has been reported that activation transiently increases, rather than decreases CCR7 expression (33). If the transferred cells derived from mesenteric lymph nodes or spleen, no CCR7 was observed on rapidly proliferating cells. CCR10 was expressed on about a fourth of proliferating cells derived from skin-draining or mesenteric lymph nodes, and 40% derived from spleen. Because CCR10 directs skin migration of T cells via its ligand CCL27 (reviewed in ref. 34) perhaps these T cells first migrate to skin, then to lymph nodes via draining lymphatics, along with DCs. How CCR10 is induced on peripheral T cells has not been examined. Some markers of the rapidly proliferating T cells we observed in lymph nodes resemble resident memory (TRM) cells, a population that was identified in epithelial tissues (35, 36). However,

Rag1-/- skin

C57BL6 skin

Fig. 6. Species of bacteria cultured from Rag1−/− skin-draining lymph nodes or skin. At least three single colonies per different color/shape in all four plates were picked up and identified by 16s rRNA PCR fragment. Total bacteria colonies of each species were counted, and the percentage of each species in total cultured bacteria was plotted. (Left) Distribution of cultured bacteria in whole skin-draining lymph nodes. (Right) Distribution of cultured bacteria in Rag1−/− skin (1 cm2 dorsal skin homogenate diluted in 1:100). (Lower) Distribution of cultured bacteria in C57BL/6 skin. Data represent three independent experiments.

E2982 | www.pnas.org/cgi/doi/10.1073/pnas.1401820111

Shen et al.

PNAS PLUS

A

donor T cells from C57BL6 mice Corynebacterium Staphylococcus

PHA

-Bacteria

Mycobacterium

Enterococcus

CD8

38.69% 43.39%

22.39%

2.98%

40.45% 7.26%

35.97% 5.51%

35.80% 2.24%

39.26%

0.71%

59.80% 2.71%

31.51%

1.64%

54.93% 2.58%

54.19% 5.76%

52.93% 1.37%

57.13%

0.36%

55.91% 18.44%

56.59% 1.35%

56.38% 2.69%

56.06% 0.70%

51.08% 0.69%

56.03%

10.43% 2.51%

39.76% 5.66%

35.58% 6.18%

42.04% 2.69%

40.58%

CD4

0.81%

5

4

3

2

1.25%

42.48% 14.54%

CFSE

B

donor T cells from germ free mice

CD8 CD4

0.11%

32.29%

0.25%

33.68% 0.12%

29.91% 0.12%

31.20% 0.07%

22.22%

0.07%

31.48%

4.02%

63.58% 6.12%

59.95% 5.44%

64.54% 2.83%

65.85% 30.13%

47.57% 2.49%

65.96%

0.15%

56.81% 0.19%

74.34% 0.37%

66.81% 0.20%

62.03% 0.52%

61.13% 0.13%

58.96%

0.10%

42.94% 0.09%

25.38% 0.09%

32.73% 0.06%

37.71% 0.06%

38.28% 0.08%

40.83%

CFSE

TRM cells were found to reside in skin, did not circulate following clearance of vaccinia or herpes viruses, and were found to express E-selectin. Because the T cells we detect are found in lymph nodes draining skin, and are reactive to skin commensals, they may represent a population that recirculates, but is related to TRM cells. They may be continually sensitized in draining lymph nodes, then circulate to skin and produce IFNγ, providing antimicrobial defense. It is clear that these T cells represent a prominent, perhaps major, T-cell population in lymph nodes draining skin. It was recently shown that lack of skin commensals results in a deficiency in IL-17A–producing T cells (10) and that this T-cell population could be restored by reconstituting mice with staphylococci. This is reminiscent of the capacity of segmented filamentous bacteria to promote development of IFNγ-producing T cells in gut-associated lymphoid tissues (37, 38). These reports, like the present study, demonstrate the importance of communication between T cells and microflora on epithelial surfaces. Human skin bacterial flora has been extensively characterized (39–41) and varies widely between individuals and anatomical sites. The skin flora of laboratory mice differs from colony to colony and is reported to be somewhat less diverse than that of humans. The skin of SCID mice was found to be dominated by staphylococci (more than 90% of sequences) compared with cohoused normal mice (5.2% of sequences) (42). We also observed an abundance of Staphylococcus sequences from skin in our colony of Rag1−/− mice. We have not determined the route by which live bacteria reach lymph nodes from skin in these mice. Entry could occur through small wounds or perhaps from hair follicles. The living bacteria we detected in lymph nodes did not appear to be cell associated (for example within DCs or macrophages), because they did not sediment with cells following centrifugation. Shen et al.

Our findings show that an IFNγ mechanism is important in preventing bacterial translocation to lymph nodes. Mycobacterial sequences were the most common bacterial sequences found in lymph nodes from Rag1−/− mice. IFNγ arms macrophages to kill ingested mycobacteria, at least in part through phagosome acidification (43) and autophagy mechanisms (44) (reviewed in ref. 45). IFNγ−/− mice were previously shown to be especially susceptible to mycobacterial infection (46). Humans with defects in the IFNγ pathway are susceptible to infection with nontuberculous mycobacteria, which is a signature of that defect. Infection of lymph nodes with Mycobacterium avium or bacillus Calmette–Guérin was found in all patients with complete IFNγ receptor deficiency (18). These patients are also susceptible to infection with histoplasma, Salmonella, and some viruses. The types of defects in the IFNγ pathway found in humans include autoantibodies against IFNγ itself (47) and genetic defects in IFNγ receptor and the signaing molecule Stat1 (48). IL-17 has also been implicated in host defense against mycobacteria (49, 50) and we observed a partial failure to clear bacterial sequences by IL-17A−/− T cells. A major role in Staphylococcus immunity by Th17 cells is revealed in hyperimmunoglobin E syndrome (HIES) or Job’s syndrome, in which patients have abnormal susceptibility to Staphylococcus in skin epithelial surfaces. In the case of HIES, heterozygous mutations in signal transducer and activator of transcription 3 (STAT3) DNA binding protein cause the systemic deficiency of IL-17 production due to a failure to express sufficient levels of Th17-specific transcription regulator retinoid-related orphan receptor gamma t (RORγt) (51–53). We did not evaluate TNF, which would be another candidate because TNF−/− mice and humans treated with TNF antagonists are susceptible to mycobacterial infection. It has been shown that both Tcell and myeloid-cell TNF are required for mycobacterial control in mice (54). IL-22 deficiency did not impair the effect of transferred T cells in reducing bacterial transcripts in Rag−/− lymph nodes. PNAS | Published online July 7, 2014 | E2983

IMMUNOLOGY AND INFLAMMATION

Fig. 7. Proliferation of T cells stimulated by bacteria from Rag1−/− lymph nodes and skin. Bacteria from skin-draining lymph nodes and skin of Rag1−/− were grown on media plates. Individual bacterial colonies were selected and typed by 16s rRNA. Enterococcus are cultured from small intestine homogenate. Heatinactivated bacteria were added to cultures of skin-draining lymph node cells (CFSE labeled) from C57BL/6 mice or GF mice, and cultures were supplemented with dendritic cells. (A and B) Flow cytometric analysis of CD4 and CD8 populations in T cells derived from C57BL/6 mice (A) and GF mice (B) at day 3. Data are representative of two or three experiments.

Rag 1-/-

0

6

5

4

Weeks

Weeks

IL-17A-/IL-22-/IFNγ-/Mu MT

4 3 2 1

Skin

5 4

IL-17A-/IL-22-/-

3

IFNγ-/Mu MT

2 1 0

7

6

5

4

2

weeks

1

0

7

6

5

4

2

3

0 1

Rag 1-/16sRNA quantification relative to WT mice

lymph nodes

5

0

16sRNA quantification relative to WT mice

KO

3

3

1

2

0

5

1

6

2

BLN ILN MLN

4

3

3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0

3

4

1

16sRNA quantification relative to WT mice

ALN Skin

5

2

6

0

16sRNA quantification relative to WT mice

WT

week

weeks

7

6

5

7

6

5

4

3

2

1

0

4

1

IL-17A-/WT IL-22-/IFNg-/Mu MT

3

2

Skin

2

3

7 6 5 4 3 2 1 0 1

4

0

IL-17A-/WT IL-22-/IFNg-/Mu MT

ALN

5

16sRNA quantification relative to WT mice

6

0

16sRNA quantification relative to WT mice

WT versus KO

weeks

Fig. 8. Role of IFNγ, IL-17, and IL-22 in skin barrier function. T cells were prepared from skin-draining lymph nodes of normal C57BL/6, IFNγ−/−, IL-17−/−, IL22−/−, or B-cell deficient (MuMT) mice. T cells were transferred into Rag1−/− mice. At the indicated times after reconstitution, tissues were examined for bacterial components by 16s rRNA sequencing. (Top Left) Axillary lymph node (ALN) vs. skin. (Top Right) Brachial (BLN), inguinal (ILN), and mesenteric (MLN) lymph nodes. Data are expressed as mean ± SEM (n = 5 in each group). Results are from two independent experiments. (Middle Left) Skin-draining lymph node cells from mice receiving indicated KO T cells. (Middle Right) Skin from mice receiving indicated KO T cells. Data are expressed as mean ± SEM (n = 3 in each group). Results are from two experiments. (Bottom Left) Comparison of skin-draining lymph node cells from WT mice and indicated KO mice. (Bottom Right) Comparison of skin from WT mice and indicated KO mice. Data are expressed as mean ± SEM (n = 3 in each group). Results are from two experiments.

Although IL-22 has been reported to inhibit growth of mycobacteria in macrophages (55), there was no effect on mycobacterial infection in IL-22−/− mice (56). Several common skin conditions are thought to relate to excess reactions of the adaptive immune system in response to commensals. Atopic dermatitis is proposed to result, in part from loss of mechanical barrier function (57) followed by excessive Th-2–driven recognition of commensals. It is suggested that T-cell recognition of commensals partly underlies atopic dermatitis in humans (2), perhaps reflecting an insufficiency of Th17 cells and an excess of IL-22 producers (reviewed in ref. 58). Psoriasis is thought to be initially triggered by skin microbial antigens and excess IL-23 and later develops a sterile lesion due to production of antimicrobial peptides, which is propagated by self-antigens (reviewed in ref. 58). Bacterial commensals are the focus of this study, which indicates a role in their regulation by the adaptive immune system. Adaptive mechanisms may also regulate commensal viruses and fungi. Better understanding of this important system could alleviate human conditions resulting from its defects and excesses. Materials and Methods Mice. Mice were maintained in a specific pathogen-free barrier facility at the National Cancer Institute (NCI, Frederick, MD) in accordance with the procedures outlined in the 2011 Guide for Care and Use of Laboratory Animals (National Institutes of Health, Bethesda). C57BL/6Ncr mice were obtained from the Animal Production Program of NCI/Charles River Laboratories.

E2984 | www.pnas.org/cgi/doi/10.1073/pnas.1401820111

Rag1−/− (C57BL/6 background) mice were originally purchased from The Jackson Laboratory. GF mice were bred as previously described (10). B-cell– deficient mice (MuMT) were generously provided by Giorgio Trinchieri (NCI). IFNγ−/− mice were kindly provided by Robert Wiltrout (NCI). OT-1 × RAG−/− (C57BL/6-Tg (OT-I)-RAG1tm1Mom) mice were kindly provided by Thomas Sayers (NCI). For the generation of IL-17 reporter mice (C57BL/6 background), murine bacterial artificial chromosome (BAC RP23-4E16) (Invitrogen) was modified to introduce a tdTomato reporter gene into the Il17 locus using recombineering technology as described previously (59). By homologous recombination, the sequence of the signal peptide of Il17a in the BAC was disrupted and the tdTomato gene with polyA was inserted immediately after the ATG start site of Il-17a. For the generation of IL-22-tdtomato reporter mice (C57BL/6 background), murine bacterial artificial chromosome (BAC RP23-401E11) (Invitrogen) was modified to introduce a tdTomato reporter gene into the Il22 locus using the same strategy as described above. Animal care was provided in accordance with the National Institutes of Health Animal Use and Care guidelines. All mice used were 8–12 wk old. Flow Cytometric Analysis. To perform surface staining, 1 × 106 cells were stained for 30 min at room temperature in PBS containing 1% FBS with the following antibodies: APC-conjugated anti-CD3 [Becton Dickinson (BD) Biosciences], APC-conjugated anti-CD4 (BD Biosciences), PerCPCy5.5- conjugated anti-CD8 (BD Biosciences), APC-conjugated anti-TCRγδ (BioLegend), PE-conjugated anti-CCR10 (R&D Systems), PerCPCy5.5-conjugated anti-Sca-1 (BD Biosciences), PerCPCy5.5-conjugated anti-CD44 (BD Biosciences), APC-conjugated anti-CD62L (BD Biosciences). The cell preparations were analyzed on FACSCalibur and LSRII. For intracellular cytokine staining, cells were stimulated with phorbol 12-myristate 13-acetate (PMA)/ionomycin (eBioscience) in the presence of GolgiPlug (BD Biosciences) for 5 h and then stained with APC-conjugated IL-17A (BioLegend), APC-conjugated IL-17F (BioLegend), perCPefluo710conjugated anti-IL-22 (eBioscience), APC-conjugated IL-5 (eBioscience), APCconjugated IFNγ (BioLegend) using Cytofix/Cytoperm Fixation/Permeabilization Solution kit with BD GolgiStop (BD Biosciences) according to the manufacturer’s instructions. Data were analyzed with FACS Express or FlowJo software (BD Biosciences). Cells were sorted using FACS Aria (BD Biosciences). Differentiation of CD8 TCM Cells from OT-I Mice. Spleens and lymph nodes from OT-1 mice were harvested and ground individually between two frosted slides into a single-cell suspension in HBSS buffer on ice. Red blood cells from splenic suspensions were removed using ACK lysis buffer (Gibco) and the resulting leukocytes were washed by HBSS to remove lysis solution. The suspension was then passed through a 100-μm cell strainer and pelleted. Cells were resuspended in RPMI media with 10 μg/mL of SIINFEKL peptide (Peprotech) and incubated at 37 °C for 1 h, followed by washing with HBSS. For in vitro differentiation, cells were cultured at 4–5 × 106 cells/mL in complete RPMI medium (with 100 IU/mL penicillin, 100 mg/mL streptomycin, and 50 μM 2-mercaptoethanol) supplemented with IL-15 at 20 ng/mL for 6 d. Cells were placed in fresh medium every 2 d. CFSE Labeling and Adoptive Transfer of T Cells. Skin-draining lymph nodes (axillary, brachial, and inguinal) and mesenteric lymph nodes from IL-17A– tdtomato reporter mice, IL-22–tdtomato reporter mice, IFNγ−/− mice, germfree mice, or C57BL/6 mice were homogenized in RPMI containing 5% (vol/vol) FBS and filtered through a 100-μm mesh nylon screen (BD Falcon). Cells were resuspended in PBS containing 5% FBS and warmed to 37 °C and then incubated for 10 min with 5 μM CFSE (Invitrogen) followed by three washes with PBS (15). A total of 2–5 × 106 CFSE-labeled cells were suspended in PBS and adoptively transferred into Rag1−/− recipient mice by i.v. injection. Dorsal skin and lymph nodes from the host mice were harvested at various times after transfer. Lymphocytes were analyzed by flow cytometry, whereas skin genomic DNA was isolated for 16s RNA quantification. Spectrotype of TCR Vβ Diversity. CFSE-labeled mouse skin-draining lymphocyte T cells were transferred into Rag1−/− mice followed by harvest of lymph nodes at day 5. Proliferating (non-CFSE) T cells were sorted by flow cytometry using a FACSAria (BD Biosciences). Following isolation of RNA from both proliferating T cells and unfractionated T cells, TCR Vβ diversity was tested by SuperTCRExpressTM Mouse TCR Vb Repertoire CDR3 Diversity Determination kit (BioMed Immunotech) based upon manufacturer’s manual. Briefly, 22 individual Vβ gene families and controls (both positive and negative) were amplified by cDNA with prelabeled primers and analyzed by Applied Biosystems 3730xl DNA sequencer. Total area under curve of each peak and number of peaks for every Vβ gene were measured by GeneMapper software (version 4.00; Applied Biosystems).

Shen et al.

Bacterial Identification by 16S rRNA Gene Sequencing. Bacterial identification to the species level was performed using 16S rRNA gene sequencing. A single colony of bacteria was lysed in 100 μL of “PrepMan Ultra” Sample Preparation Reagent (Applied Biosystems) to prepare a DNA template for PCR amplification. 16S rDNA fragments (900 base pairs, bp) were amplified with universal bacterial primers (Integrated DNA Technology) 8f (5′-AGAGTTTGATCCTGGCTCAG-3′) and 926r (5′-CCGTCAATTCCTTTRAGTTT-3′, R = A/G) (61), purified with a Gel Extraction kit (Qiagen), and sequenced with primer 8f and 926r. Resulting sequences were subjected to nucleotide–nucleotide BLAST (blastn) in comparison with known 16S rRNA genes in the public databases to identify the bacteria species. A species was assigned to an isolate when it shared 95% or higher identity to known genes. High-Throughput Bacterial Sequencing. For some of the samples we also performed high throughput sequencing using a 454 Roche machine. Before the 16S sequencing, we performed PCR amplification step of the V1–V3 region using 27F 5′-AGAGTTTGATCCTGGCTCAG-3′ and 534R 5′-ATTACCGCGGCTGCTGG-3′ primers, which also included barcodes and 454 sequencing primers. PCR was done with the following conditions: 95 °C for 2 min, 35 cycles of 30 s at 95 °C, 30 s at 56 °C, and 5 min at 72 °C. Resulting PCR products were purified using AMPure (Beckman Coulter), quantified and pooled in equimolar concentrations. Emulsion PCRs were made using a Roche 454 emPCR Lib-A kit. Sequencing was done using the Roche 454 FLX-Ti protocol. Sequence analysis was done using Mothur v1.22.0. Briefly, sequences were deconvoluted and low-quality sequences and chimeras were removed (UChime algorithm, Mothur plugin). To obtain operational taxonomic units (OTU), sequences were aligned to the SILVA bacterial reference dataset and clustered; sequences with more than 97% similarity were binned into the same OTUs. Bacterial diversity indexes (Chao and Inverse Simpson Index) were also calculated using Mothur software.

Quantitative Real-Time PCR Amplification of 16S rRNA Gene Sequences. The quantity of bacteria in skin and lymph nodes was measured by quantitative real-time PCR (qPCR) using a 7300 Real-Time PCR system (Applied Biosystems) with 16S rRNA gene primers (Integrated DNA Technology). A short segment of the 16S rRNA gene (200–300 bp) was specifically amplified by real-time PCR, using the conserved 16S rRNA-specific primer pair UniF340 (5′ ACTCCTACGGGAGGCAGCAGT 3′) and UniR514 (5′ ATTACCGCGGCTGCTGGC 3′) to determine the total amount of commensal bacteria in skin or lymph nodes (62). The real-time PCR program started with an initial step at 95 °C for 2 min, followed by 40 cycles of 10 s at 95 °C and 45 s at 63 °C. Data were acquired in the final step at 63 °C. The real-time PCRs were done using SYBR Green Supermix (Roche). Bacterial numbers were determined based on the relative level to GAPDH internal controls. In Vitro Bacteria Stimulation of Lymph Node Cells. Bacteria single colonies were picked from four plates and separately grown in related liquid culture medium overnight at 35 °C. Cultures were then treated as reported in detail (63). Briefly, cells were centrifuged and washed three times in PBS. The recovered bacteria were resuspended in complete RPMI and incubated at 80 °C for 2 h to generate a heat-killed preparation. Bacterial killing was confirmed by lack of growth on blood agar plates. Bacterial concentration was enumerated by comparing the absorbance of a serial dilution of bacteria at 590 nm with McFarland nephelometer standards. The lymph node single-cell suspensions were prepared in PBS, labeled with CFSE, and adjusted to 1 × 106 cells/mL in complete RPMI medium. One hundred microliters of diluted cell suspensions was dispensed into 96-well round-bottom culture plates, and added with or without 1 × 104 bacteria or 10 μg/mL phytohemagglutinin-A (Sigma). Splenic cells purified from irradiated Rag−/− mouse were added as feeder cell at the concentration of 2.5:1 or 0.5:1 to host cells. Cells were harvested after 3 d, stained with anti-CD4 or anti-CD8, and analyzed by flow cytometry. Treatment of Mice with Oral Antibiotic Mixture. Seven-week-old C57BL/6 and Rag1−/− mice were treated with freshly prepared antibiotic mixture [neomycin (1 g/L), vancomycin (0.5 g/L), and imipenem/cilastatin (Primaxin) 0.5 g/L] in drinking water for 1 mo (64). The water bottle was changed every other day. Statistics. Statistical analysis was performed using GraphPad Prism 5.0 software. Data are expressed as mean ± SEM. The Student two-tailed unpaired, parametric t test was used to assess statistical differences between two groups. Asterisks indicate statistical differences, *P < 0.05, **P < 0.01, ***P < 0.001.

Bacterial Genomic DNA Extraction. The dorsal skin harvested from C57BL6 or Rag1−/− mice were homogenized using a Polytron PT 10–35 homogenizer (Kinematica) in 1 mL sterile PBS. The lymph nodes were minced with the

ACKNOWLEDGMENTS. We thank Drs. N. Boudaloux and Y. Belkaid for the generous gift of germ-free mice, and Drs. Mark Udey and Joost Oppenheim for helpful suggestions on the study and manuscript. The study was funded by the intramural program of the National Cancer Institute and a grant (to S.K.D.) from the Ely and Edythe Broad Foundation.

1. Afshar M, Gallo RL (2013) Innate immune defense system of the skin. Vet Dermatol 24(1):32–8.e8-9. 2. Kuo IH, Yoshida T, De Benedetto A, Beck LA (2013) The cutaneous innate immune response in patients with atopic dermatitis. J Allergy Clin Immunol 131(2):266–278. 3. Shklovskaya E, et al. (2011) Langerhans cells are precommitted to immune tolerance induction. Proc Natl Acad Sci USA 108(44):18049–18054. 4. Rezaei N, et al. (2006) Frequency and clinical manifestations of patients with primary immunodeficiency disorders in Iran: Update from the Iranian Primary Immunodeficiency Registry. J Clin Immunol 26(6):519–532. 5. Villa A, et al. (2001) V(D)J recombination defects in lymphocytes due to RAG mutations: Severe immunodeficiency with a spectrum of clinical presentations. Blood 97(1): 81–88. 6. Sadeghi-Shabestari M, et al. (2009) Novel RAG2 mutation in a patient with T- B- severe combined immunodeficiency and disseminated BCG disease. J Investig Allergol Clin Immunol 19(6):494–496. 7. Avila EM, et al. (2010) Highly variable clinical phenotypes of hypomorphic RAG1 mutations. Pediatrics 126(5):e1248–e1252. 8. Clark RA, et al. (2006) The vast majority of CLA+ T cells are resident in normal skin. J Immunol 176(7):4431–4439. 9. Kupper TS, Fuhlbrigge RC (2004) Immune surveillance in the skin: Mechanisms and clinical consequences. Nat Rev Immunol 4(3):211–222. 10. Naik S, et al. (2012) Compartmentalized control of skin immunity by resident commensals. Science 337(6098):1115–1119. 11. Milner JD, et al. (2008) Impaired T(H)17 cell differentiation in subjects with autosomal dominant hyper-IgE syndrome. Nature 452(7188):773–776.

12. Kisand K, et al. (2010) Chronic mucocutaneous candidiasis in APECED or thymoma patients correlates with autoimmunity to Th17-associated cytokines. J Exp Med 207(2):299–308. 13. Puel A, et al. (2011) Chronic mucocutaneous candidiasis in humans with inborn errors of interleukin-17 immunity. Science 332(6025):65–68. 14. Schluns KS, Kieper WC, Jameson SC, Lefrançois L (2000) Interleukin-7 mediates the homeostasis of naïve and memory CD8 T cells in vivo. Nat Immunol 1(5):426–432. 15. Li WQ, et al. (2006) IL-7 promotes T cell proliferation through destabilization of p27Kip1. J Exp Med 203(3):573–582. 16. Kieper WC, et al. (2005) Recent immune status determines the source of antigens that drive homeostatic T cell expansion. J Immunol 174(6):3158–3163. 17. Brenchley JM, et al. (2006) Microbial translocation is a cause of systemic immune activation in chronic HIV infection. Nat Med 12(12):1365–1371. 18. Dorman SE, et al. (2004) Clinical features of dominant and recessive interferon gamma receptor 1 deficiencies. Lancet 364(9451):2113–2121. 19. Eyerich K, et al. (2009) IL-17 in atopic eczema: Linking allergen-specific adaptive and microbial-triggered innate immune response. J Allergy and Clin Immunol 123(1): 59–66.e4. 20. Liang SC, et al. (2006) Interleukin (IL)-22 and IL-17 are coexpressed by Th17 cells and cooperatively enhance expression of antimicrobial peptides. J Exp Med 203(10): 2271–2279. 21. Zimmermann K, Haas A, Oxenius A (2012) Systemic antibody responses to gut microbes in health and disease. Gut Microbes 3(1):42–47. 22. van der Aar AM, et al. (2013) Langerhans cells favor skin flora tolerance through limited presentation of bacterial antigens and induction of regulatory T cells. J Invest Dermatol 133(5):1240–1249.

Shen et al.

PNAS PLUS

rough side of the slides in 1 mL sterile PBS. Bacterial genomic DNA was extracted from the skin and lymph nodes by use of a Qiagen DNA mini kit according to the kit directions.

PNAS | Published online July 7, 2014 | E2985

IMMUNOLOGY AND INFLAMMATION

In Vitro Culture of Bacteria. C57BL6 and Rag1−/− mice were housed in the same cage for 2 mo to equilibrate the microbiome of the two different strains of mice. Axillary (ALN), brachial (BLN), inguinal (ILN), mesenteric lymph nodes (MLN), spleen (as control), skin, small intestine, and large intestine of individual mice were removed aseptically. Lymph nodes were homogenized using Kontes RNase-Free Pellet Pestle Grinders (Kimble ChaseBTX) in 400 μL PBS buffer, whereas spleen, small intestine (1 cm of the middle part), large intestine (1 cm of the distal part), and dorsal skin (1 cm2) tissues were homogenized using a tissue homogenizer (Omni International) in 1 mL sterile PBS buffer. All lymph node and spleen homogenized samples were plated into four different plates including chocolate agar (Teknova) plate, sheep blood plate (Mediatech), MacConkey agar (Teknova) plate, and thioglycolate plate (Teknova), at 100 μL/plate. Intestine and skin homogenized solution was diluted 100-fold in PBS buffer and then plated onto four plates as described above. The bacteria were grown at 35 °C for 3 d and colonies were counted (60).

23. Estes JD, et al. (2010) Damaged intestinal epithelial integrity linked to microbial translocation in pathogenic simian immunodeficiency virus infections. PLoS Pathog 6(8):e1001052. 24. Guilliams M, et al. (2010) Skin-draining lymph nodes contain dermis-derived CD103(-) dendritic cells that constitutively produce retinoic acid and induce Foxp3(+) regulatory T cells. Blood 115(10):1958–1968. 25. Appay V, et al. (2002) Memory CD8+ T cells vary in differentiation phenotype in different persistent virus infections. Nat Med 8(4):379–385. 26. Gillespie GM, et al. (2000) Functional heterogeneity and high frequencies of cytomegalovirus-specific CD8(+) T lymphocytes in healthy seropositive donors. J Virol 74(17):8140–8150. 27. Jabbari A, Harty JT (2006) Secondary memory CD8+ T cells are more protective but slower to acquire a central-memory phenotype. J Exp Med 203(4):919–932. 28. Klenerman P, Hill A (2005) T cells and viral persistence: Lessons from diverse infections. Nat Immunol 6(9):873–879. 29. Masopust D, Ha SJ, Vezys V, Ahmed R (2006) Stimulation history dictates memory CD8 T cell phenotype: Implications for prime-boost vaccination. J Immunol 177(2):831–839. 30. Unsoeld H, Pircher H (2005) Complex memory T-cell phenotypes revealed by coexpression of CD62L and CCR7. J Virol 79(7):4510–4513. 31. Badovinac VP, Porter BB, Harty JT (2004) CD8+ T cell contraction is controlled by early inflammation. Nat Immunol 5(8):809–817. 32. Stein JV, et al. (2000) The CC chemokine thymus-derived chemotactic agent 4 (TCA-4, secondary lymphoid tissue chemokine, 6Ckine, exodus-2) triggers lymphocyte function-associated antigen 1-mediated arrest of rolling T lymphocytes in peripheral lymph node high endothelial venules. J Exp Med 191(1):61–76. 33. Sallusto F, et al. (1999) Switch in chemokine receptor expression upon TCR stimulation reveals novel homing potential for recently activated T cells. Eur J Immunol 29(6): 2037–2045. 34. Xiong N, Fu Y, Hu S, Xia M, Yang J (2012) CCR10 and its ligands in regulation of epithelial immunity and diseases. Protein Cell 3(8):571–580. 35. Jiang X, et al. (2012) Skin infection generates non-migratory memory CD8+ T(RM) cells providing global skin immunity. Nature 483(7388):227–231. 36. Mueller SN, Gebhardt T, Carbone FR, Heath WR (2013) Memory T cell subsets, migration patterns, and tissue residence. Annu Rev Immunol 31:137–161. 37. Ivanov II, et al. (2008) Specific microbiota direct the differentiation of IL-17-producing T-helper cells in the mucosa of the small intestine. Cell Host Microbe 4(4):337–349. 38. Umesaki Y, Okada Y, Matsumoto S, Imaoka A, Setoyama H (1995) Segmented filamentous bacteria are indigenous intestinal bacteria that activate intraepithelial lymphocytes and induce MHC class II molecules and fucosyl asialo GM1 glycolipids on the small intestinal epithelial cells in the ex-germ-free mouse. Microbiol Immunol 39(8):555–562. 39. Costello EK, et al. (2009) Bacterial community variation in human body habitats across space and time. Science 326(5960):1694–1697. 40. Gao Z, Tseng CH, Pei Z, Blaser MJ (2007) Molecular analysis of human forearm superficial skin bacterial biota. Proc Natl Acad Sci USA 104(8):2927–2932. 41. Grice EA, et al.; NISC Comparative Sequencing Program (2009) Topographical and temporal diversity of the human skin microbiome. Science 324(5931):1190–1192. 42. Garcia-Garcerà M, et al. (2012) Staphylococcus prevails in the skin microbiota of longterm immunodeficient mice. Environ Microbiol 14(8):2087–2098. 43. Schaible UE, Sturgill-Koszycki S, Schlesinger PH, Russell DG (1998) Cytokine activation leads to acidification and increases maturation of Mycobacterium avium-containing phagosomes in murine macrophages. J Immunol 160(3):1290–1296.

E2986 | www.pnas.org/cgi/doi/10.1073/pnas.1401820111

44. Gutierrez MG, et al. (2004) Autophagy is a defense mechanism inhibiting BCG and Mycobacterium tuberculosis survival in infected macrophages. Cell 119(6):753–766. 45. Ní Cheallaigh C, Keane J, Lavelle EC, Hope JC, Harris J (2011) Autophagy in the immune response to tuberculosis: Clinical perspectives. Clin Exp Immunol 164(3):291–300. 46. Flynn JL, et al. (1993) An essential role for interferon gamma in resistance to Mycobacterium tuberculosis infection. J Exp Med 178(6):2249–2254. 47. Turano A, Caruso A (1993) The role of human autoantibodies against gamma-interferon. J Antimicrob Chemother 32(Suppl A):99–105. 48. Chapgier A, et al. (2006) Human complete Stat-1 deficiency is associated with defective type I and II IFN responses in vitro but immunity to some low virulence viruses in vivo. J Immunol 176(8):5078–5083. 49. Umemura M, et al. (2007) IL-17-mediated regulation of innate and acquired immune response against pulmonary Mycobacterium bovis bacille Calmette-Guerin infection. J Immunol 178(6):3786–3796. 50. Khader SA, et al. (2007) IL-23 and IL-17 in the establishment of protective pulmonary CD4+ T cell responses after vaccination and during Mycobacterium tuberculosis challenge. Nat Immunol 8(4):369–377. 51. Holland SM, et al. (2007) STAT3 mutations in the hyper-IgE syndrome. N Engl J Med 357(16):1608–1619. 52. Ma CS, et al. (2008) Deficiency of Th17 cells in hyper IgE syndrome due to mutations in STAT3. J Exp Med 205(7):1551–1557. 53. Minegishi Y, et al. (2007) Dominant-negative mutations in the DNA-binding domain of STAT3 cause hyper-IgE syndrome. Nature 448(7157):1058–1062. 54. Allie N, et al. (2013) Prominent role for T cell-derived tumour necrosis factor for sustained control of Mycobacterium tuberculosis infection. Sci Rep 3:1809. 55. Dhiman R, et al. (2009) IL-22 produced by human NK cells inhibits growth of Mycobacterium tuberculosis by enhancing phagolysosomal fusion. J Immunol 183(10): 6639–6645. 56. Behrends J, Renauld JC, Ehlers S, Hölscher C (2013) IL-22 is mainly produced by IFNγsecreting cells but is dispensable for host protection against Mycobacterium tuberculosis infection. PLoS ONE 8(2):e57379. 57. Palmer CN, et al. (2006) Common loss-of-function variants of the epidermal barrier protein filaggrin are a major predisposing factor for atopic dermatitis. Nat Genet 38(4):441–446. 58. Weaver CT, Elson CO, Fouser LA, Kolls JK (2013) The Th17 pathway and inflammatory diseases of the intestines, lungs, and skin. Annu Rev Pathol 8:477–512. 59. Mazzucchelli RI, et al. (2009) Visualization and identification of IL-7 producing cells in reporter mice. PLoS ONE 4(11):e7637. 60. Fukata M, et al. (2005) Toll-like receptor-4 is required for intestinal response to epithelial injury and limiting bacterial translocation in a murine model of acute colitis. Am J Physiol Gastrointest Liver Physiol 288(5):G1055–G1065. 61. Yu RR, et al. (2009) A Chinese rhesus macaque (Macaca mulatta) model for vaginal Lactobacillus colonization and live microbicide development. J Med Primatol 38(2): 125–136. 62. Barman M, et al. (2008) Enteric salmonellosis disrupts the microbial ecology of the murine gastrointestinal tract. Infect Immun 76(3):907–915. 63. Jazani NH, et al. (2010) Evaluation of the adjuvant activity of naloxone, an opioid receptor antagonist, in combination with heat-killed Listeria monocytogenes vaccine. Microbes Infect 12(5):382–388. 64. Iida N, et al. (2013) Commensal bacteria control cancer response to therapy by modulating the tumor microenvironment. Science 342(6161):967–970.

Shen et al.

Adaptive immunity to murine skin commensals.

The adaptive immune system provides critical defense against pathogenic bacteria. Commensal bacteria have begun to receive much attention in recent ye...
1MB Sizes 1 Downloads 5 Views