TLR7 Induces Anergy in Human CD4+ T Cells

A rt i c l e s

TLR7 induces anergy in human CD4+ T cells

Margarita Dominguez-Villar1, Anne-Sophie Gautron1, Marine de Marcken1, Marla J Keller2 & David A Hafler1

The recognition of microbial patterns by Toll-like receptors (TLRs) is critical for activation of the innate immune system. Although TLRs are expressed by human CD4+ T cells, their function is not well understood. Here we found that engagement of TLR7 in CD4+ T cells induced intracellular calcium flux with activation of an anergic gene-expression program dependent on the transcription factor NFATc2, as well as unresponsiveness of T cells. As chronic infection with RNA viruses such as human immunodeficiency virus type 1 (HIV-1) induces profound dysfunction of CD4+ T cells, we investigated the role of TLR7-induced anergy in HIV-1 infection. Silencing of TLR7 markedly decreased the frequency of HIV-1-infected CD4+ T cells and restored the responsiveness of those HIV-1+ CD4+ T cells. Our results elucidate a previously unknown function for microbial pattern– recognition receptors in the downregulation of immune responses.

Toll-like receptors (TLRs) represent the major pathway by which immunodeficiency virus type 1 (HIV-1), the immune responses medi- microorganisms interact with host cells. They are a family of highly ated by CD4+ helper T cells and CD8+ cytotoxic T cells determine the out- conserved pattern-recognition receptors that recognize distinct come of the infection, with chronic infections being correlated with late, pathogen-associated molecular patterns that are conserved in specific transient or narrowly focused responses by CD4+ or CD8+ T cells9–11. classes of microorganisms1. The human TLR family consists of at least Several studies have demonstrated impairment in the activation and/or ten members that can be classified into two different groups on the basis function of T cells during infection with HIV-1. Specifically, CD4+ of their cellular location. Intracellular TLRs (TLR3, TLR7, TLR8 and T cells from patients chronically infected with HIV-1 display an anergic TLR9) recognize nucleic acids; TLR7 and TLR8 recognize single- phenotype with defects in proliferation and the secretion of interleukin stranded RNA2,3, whereas TLR3 and TLR9 are receptors for double- 2 (IL-2) and interferon-γ (IFN–γ). The mechanisms by which RNA stranded RNA and double-stranded DNA, respectively. In contrast, viruses impair T cell function are not well understood. cell surface TLRs (TLR1, TLR2, TLR4, TLR5 and TLR6) recognize Here we describe a previously unrecognized pathway of TLR- various components of bacteria1. In mice, although TLR7 and TLR8 are mediated negative regulation of both the activation and the cytokine expressed at low levels in CD4+ T cells, there are species-specific dif- production of CD4+ T cells. Engaging TLR7 expressed on CD4+ Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature ferences in the recognition of ligands3 as well as in their functionality. T cells resulted in complete anergy by inducing intracellular calcium Specifically, mouse TLR7 and human TLR8 mediate species-specific flux, with activation of an anergic gene-expression program depend- recognition of GU-rich single-stranded RNA. It has been suggested ent on the transcription factor NFATc2 and with subsequent T cell npg that in contrast to its human TLR counterpart, mouse TLR8 is not unresponsiveness that was reversed by knockdown of TLR7 with functional and TLR7 is the only TLR that recognizes single-stranded short hairpin RNA (shRNA). In studies of the potential physiological RNA4. The expression and signaling pathways triggered by stimula- relevance of these findings, we found that knockdown of TLR7 via tion of TLRs have been described in antigen-presenting cells (APCs) shRNA decreased the frequency of HIV-1-infected CD4+ T cells in a process that leads to the activation of APCs with the secretion in vitro and restored the responsiveness of those HIV-1+ CD4+ of inflammatory and antiviral cytokines1,5. Although TLR expression T cells in vitro. Our results elucidate a previously unknown function has been studied mainly in APCs, several reports have described the for microbial pattern–recognition receptors in the downregulation expression of TLRs on lymphocytes6, and specifically on CD4+ T cells. of immune responses, inducing anergy by increasing intracellular As with APCs, such studies indicate that the engagement of TLRs acts calcium concentrations and interfering with secondary costimulation as a positive costimulatory signal that increases the secretion of proin- signals in the presence of signaling via TLRs7. flammatory cytokines, proliferation and cell survival7,8. While TLRs are central to the early host immune response to acute RESULTS viral infection, more-chronic infectious diseases are characterized by Inhibition of the activation of CD4+ T cells by TLR7 the inability of the host immune system to mount a strong, long-lasting While investigating a potential costimulatory role for TLRs in CD4+ response to the infectious agent. In particular, it has been shown that T cells, we observed that the entry of CD4+ T cells into the cell cycle during infection with RNA viruses such as hepatitis C virus and human after crosslinking of the T cell antigen receptor (TCR) with antibody

1Department of Neurology and Department of Immunobiology, Yale School of Medicine, New Haven, Connecticut, USA. 2Albert Einstein College of Medicine and Montefiore Medical Center, Bronx, New York, USA. Correspondence should be addressed to M.D.-V. ([email protected]) or D.A.H. ([email protected]).

Received 18 June; accepted 20 October; published online 17 November 2014; doi:10.1038/ni.3036

Figure 1 TLR7 signaling inhibits the proliferation and cytokine secretion of CD4+ T cells. (a) Proliferation of CD4+ T cells labeled with the cytosolic dye CFSE and stimulated for 3 d with various concentrations (above plots) of imiquimod (IMQ). Numbers above bracketed lines indicate percent viable proliferating CD4+ T cells. (b) Frequency of viable proliferating CD4+ T cells stimulated for 3 d with various concentrations (horizontal axis) of imiquimod. (c) ELISA of cytokine secretion by CD4+ T cells stimulated for 3 d with various concentrations of imiquimod. (d) Intracellular staining of IFN-γ and IL-2 (top) and of IL-17 and IL-4 (bottom) in CD4+ T cells stimulated for 4 d with various concentrations of imiquimod and then stimulated for with 4 h PMA and ionomycin. Numbers in quadrants indicate percent cells in each throughout. (e) Frequency of cytokine-producing CD4+ T cells (as in d). (f) Expression of CD25 (left), CD69 (middle) and CD137 (right) on CD4+ T cells stimulated with anti-CD3 and anti-CD28 in the presence (α-CD3 and α-CD28 + IMQ) or absence (α-CD3 and α-CD28) of imiquimod, and of cells stained with isotype-matched control antibody (Isotype). MFI, mean fluorescence intensity. *P < 0.05, **P < 0.005 and ***P < 0.0005 (paired t-test). Data are representative of eight (a–d) or six (e,f) independent experiments with one donor in each (mean and s.e.m. in b,c,e,f).

to the invariant signaling protein CD3 (anti-CD3) and crosslinking of of up to 15 µg/ml of imiquimod but found no effect on cell viability the coreceptor CD28 with anti-CD28 was blocked by coengagement (data not shown). The diminished proliferation correlated with less of TLR7 (Fig. 1a,b and Supplementary Fig. 1a,b). Treatment with the secretion of the cytokines IFN-γ, IL-17, IL-2 and IL-4, as measured by synthetic TLR7 agonist imiquimod resulted in considerably less pro- enzyme-linked immunosorbent assay (ELISA), at day 3 after stimu- liferation of CD4+ T cells than that of vehicle-treated control cells, as lation (Fig. 1c). We confirmed the diminished cytokine secretion at well as less secretion of IFN-γ and IL-17, in a dose-dependent fashion the single-cell level, as the frequency of cytokine-producing cells was (Fig. 1c–e and Supplementary Fig. 1c,d). We observed this inhibitory also diminished in a dose-dependent manner with increasing doses of effect as soon as 12 h after activation, with much less induction of imiquimod in culture (Fig. 1d,e). Furthermore, stimulation of CD4+ the expression of IL2, IFNG and IL4 after treatment with imiquimod T cells in the presence of imiquimod inhibited the expression of (Supplementary Fig. 1e). We assessed the effect of concentrations activation markers such as CD25, CD69 and CD137, measured 48 h

nature immunology VOLUME 16 NUMBER 1 JANUARY 2015 119 A rt i c l e s

a b Veh e Veh f GDQ IMQ 100

Veh GDQ Lox CL264 80 Lox 4 150 + 80

150 + CL264 60 3 60 100 ** 100 40 70.9 8.3 9.1 55.3 40 2

50 T cells (%) 50 20 IL-2 (U/ml)

T cells (%) *** Proliferating CD4 Cells *** IFN- γ ( ng / ml ) 2 3 4 5 20 1 0 10 10 10 10 *** *** Proliferating CD4 0 0 0.1 0.25 0.5 1 2.5 CFSE 0 0 0 NT TLR7(3) NT TLR7(3) ssRNA40 (µg/ml) Veh 150 4 1,000 2.0 400 8 c GDQ d NT g TLR7(3) * Lox 2.0 300 6 3 750 1.5 CL264 100 * 1.5 200 4 2 500 1.0 ** IL-2 (pg/ml)

100 IFN- γ ( ng / ml ) 2

IL-2 (U/ml) 1.0 50 IL-4 (pg/ml) IL-17 (ng/ml) IFN- γ ( ng / ml ) * *** 1 ** 250 0.5 *** * 0 0 *** * * * 0.5 0 0.1 0.25 0.5 1 2.5 0 0.1 0.25 0.5 1 2.5 * * TLR7 mRNA (AU) 0 0 0 0 *** ssRNA40 (µg/ml) ssRNA40 (µg/ml) 0 1,000 4 NT TLR7(3) 800 Figure 2 The inhibitory effect of imiquimod 3 150 Isotype Isotype * * is TLR7 specific. (a) Proliferation of CFSE- 600 * 2 labeled CD4+ T cells stimulated for 3 d with 100 400 * IL-4 (pg/ml)

anti-CD3 and anti-CD28 in the presence of * IL-17 (ng/ml) 1 50 200 * vehicle (Veh) or the TLR7 agonists Cells 2 3 4 5 0 0 gardiquimod (GDQ), loxoribine (Lox) or 0 10 10 10 10 0 0.1 0.25 0.5 1 2.5 0 0.1 0.25 0.5 1 2.5 TLR7 CL264 (numbers above bracketed lines as in ssRNA40 (µg/ml) ssRNA40 (µg/ml) Fig. 1a). (b) Frequency of viable proliferating CD4+ T cells as in a. (c) ELISA of cytokine secretion by CD4+ T cells as in a. (d) Expression of TLR7 mRNA (top) and TLR7 protein (bottom) in CD4+ T cells stimulated with anti-CD3 and anti-CD28 in the presence of shRNA specific for TLR7 (construct designation, TRCN0000056973; TLR7(3)) or nontargeting shRNA (NT), and also containing sequence encoding green fluorescent protein, then sorted at day 5 on the basis of the expression of green fluorescent protein; mRNA results are presented in arbitrary units (AU). Isotype (bottom), isotype-matched control antibody. (e) ELISA of IFN-γ and IL-2 secreted by CD4+ T cells transduced with shRNA as in d (horizontal axis) and stimulated for 3 d with anti-CD3 and anti-CD28 in the presence of vehicle or imiquimod (key). (f) Proliferation of CFSE-labeled CD4+ T cells stimulated with anti-CD3 and anti-CD28 in the presence of various doses of ssRNA40 (horizontal axis). (g) ELISA of IL-2, IFN-γ, IL-4 and IL-17 secreted by CD4+ T cells at day 3 after activation with anti-CD3 and anti-CD28 in the presence of various doses of ssRNA40. *P < 0.05, **P < 0.005 and ***P < 0.0005 (one-way analysis of variance (ANOVA) with Tukey’s post-test (b,c) or paired t-test (d,e,g)). Data are representative of seven independent experiments (a,b,d,e; mean and s.e.m. in d,e), five experiments (c; mean and s.e.m.) or four independent experiments (f,g; mean ± s.e.m. of n = 4 donors).

after activation (Fig. 1f). Of note, the effect of imiquimod was not silenced that were stimulated with imiquimod produced concentra- related to the conversion of CD4+ T cells into a population of regu- tions of IFN-γ and IL-2 similar to those produced after treatment latory T cells (Treg cells), as neither expression of the transcription with vehicle (Fig. 2e). Both TLR7 and TLR8 are expressed in human Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature factor Foxp3 nor secretion of IL-10 was increased in the presence of CD4+ T cell subpopulations6,8; however, although reports suggest that imiquimod (data not shown). The unresponsive phenotype observed both TLR7 and TLR8 recognize single-stranded RNA as their natural 2,3 was not due to an indirect effect of TLR7 on Treg cells, as we obtained ligand in APCs , we observed no such effect when we used the TLR8 + + + npg the same results with a CD4 T cell population sorted as CD3 CD4 ligand ssRNA40 (a single-stranded RNA) in the transfection reagent + lo + or CD4 CD25 CD127 cells and depleted of Treg cells (data not LyoVec in these experiments in parallel with TLR7 ligands (Fig. 2f,g). shown). Only stimulation of TLR9 with the synthetic ligand oligode- Instead, stimulation of T cells in the presence of ssRNA40 significantly oxynucleotide ODN2006 also decreased the frequency of proliferating increased the production of IFN-γ and inhibited the secretion of IL-4 CD4+ T cells, but it did so to a lesser extent than did stimulation of by CD4+ T cells (Fig. 2g), with no effect on proliferation (Fig. 2f). TLR7 (Supplementary Fig. 1a,b). The effect observed was not due Vehicle alone (LyoVec) did not have any effect on cytokine secre- to an indirect effect of the triggering of TLR7 on contaminant APCs tion or proliferation under these experimental conditions (data not in the culture, as imiquimod exerted the same effect on T cell clones shown). We obtained the same results with CL075, another ligand grown from a single donor (Supplementary Fig. 2). predominantly of TLR8 (data not shown). These data were in agree- To confirm the specificity of our results, we stimulated CD4+ T cells ment with a published report8 and suggested that despite sharing in the presence of other synthetic ligands of TLR7, such as loxoribine, ligands, the signaling pathways that TLR7 and TLR8 triggered on CL264 or gardiquimod (Fig. 2). The three ligands tested induced a CD4+ T cells led to different phenotypic outcomes. significantly less CD4+ T cell proliferation than did a vehicle con- Ligation of TLR7 induces the activation and maturation of APCs, with trol, with gardiquimod and loxoribine inhibiting proliferation to a upregulation of activation markers and secretion of proinflammatory degree similar to that achieved by imiquimod treatment (Fig. 2a,b). cytokines12–14. In agreement with published data, CD14+ mono- The secretion of IL-2, IFN-γ, IL-4 and IL-17 was also inhibited after cytes isolated from the same donors studied above (Figs. 1 and 2) stimulation in the presence of each of the three ligands, as measured and stimulated in the presence or absence of imiquimod showed by ELISA after 3 d in culture (Fig. 2c). The specificity of the pheno- an activated phenotype, with upregulation of the expression of the type observed was further demonstrated by silencing of TLR7 in CD4+ human leukocyte antigen HLA-DR, the costimulatory molecule CD80 T cells (Fig. 2d). While control CD4+ T cells transduced with nontar- and the cytokine receptor CD25 and downregulation of the expression geting shRNA responded to treatment with imiquimod by decreasing of the costimulatory molecule CD86 (Supplementary Fig. 3a). Upon their secretion of IFN-γ and IL-2, CD4+ T cells in which TLR7 was being stimulated with imiquimod, CD14+ monocytes also secreted the

120 VOLUME 16 NUMBER 1 JANUARY 2015 nature immunology A rt i c l e s

a IMQ (5 µg/ml) IMQ (1 µg/ml) b c

2.5 1.0 IRS Ctrl + IMQ * IRS 661 + IMQ 2.0 0.8 * 1.5 0.6 * 1.0 0.4 * *

0.5 p-NFATc2 (AU) 0.2 1,000 Ionomycin ssRNA40

NFATc2/p-NFATc2 (AU) 0 0 800 0 45 90 0 10 20 30 40 600 Time (min) Time (min) Indo-1 AM (400/475) 400 200 400 600 800 p-NFATc2 NFATc2 97 kDa 200 Time (s)

Indo-1 AM (400/475) β-actin β-actin 0 200 400 600 Time (s) ) 3 ) ) ) ) ) ) 3 3 3 3 3 d 15 40 *** 3 6 ** 8 * 80 150 100 * ** ** 80 30 6 60 Veh 10 4 100 IMQ 60 IRS 661 + IMQ 20 4 40 40 5 2 50 10 2 20 20

0 0 0 0 0 0 0 FASL mRNA (relative × 10 EGR2 mRNA (relative × 10 CBLB mRNA (relative × 10 DGKA mRNA (relative × 10 GRG4 mRNA (relative × 10 CASP3 mRNA (relative × 10 CD98 mRNA (expression × 10 ) ) ) ) ) 3 ) 3 ) 3 3 3 3 50 120 60 10.0 * 4 3 50 50 ** *** ** * ** * 40 40 40 90 7.5 3 40 30 30 30 60 5.0 2 20 20 20 20 30 2.5 1 10 10 10

0 0 0 0 0 ITCH mRNA (relative × 10 0 0 EGR3 mRNA (relative × 10 LDHA mRNA (relative × 10 SIRT1 mRNA (relative × 10 IKZF1 mRNA (relative × 10 RAB10 mRNA (relative × 10 KMD6B mRNA (relative × 10 Veh

) NT IMQ 3 ) ) ) 3 3 3 6 NFATc2 100 IRS 661 + IMQ 50 5 15 6 e f 5 ** * 180 80 *** 40 4 4 135 10 4 4 ** 30 3 60 3 90 20 2 40 2 5 2 2 IL-2 (pg/ml) IFN- γ (ng/ml) IL-2 (pg/ml) IFN- γ (ng/ml) 45 10 1 20 *** 1 ***

Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature 0 0 0 0 0 0 0 0 RGS2 mRNA (relative × 10 TRAF5 mRNA (relative × 10 SOCS2 mRNA (relative × 10 Veh IMQ Veh IMQ TNFRSF9 mRNA (relative × 10

Figure 3 Mechanism of TLR7-induced anergy. (a) Calcium flux over time in CD4+ T cells stimulated (downward arrows) with 5 µg/ml or 1 µg/ml of imiquimod (top row), ionomycin (bottom left) or ssRNA40 in LyoVec (bottom right), assessed as the ratio of Indo-1 AM fluorescence at 400 nm to that npg at 475 nm (400/475). (b) Calcium flux in CD4+ T cells preincubated for 1 h with control sequence (IRS Ctrl + IMQ) or with the TLR7-specific inhibitory sequence IRS 661 (10 µM; IRS 661 + IMQ) and stimulated with imiquimod (5 µg/ml), presented as in a. (c) Immunoblot analysis of total NFATc2 (left) or phosphorylated (p-) NFATc2 (right) and β-actin (loading control) in CD4+ T cells at 0, 45 and 90 min (left) or 0, 10, 20, 30 and 40 min (right) after treatment with imiquimod (bottom), and ratio of dephosphorylated NFATc2 to phosphorylated NFATc2 (top left) or intensity of phosphorylated NFATc2, normalized to results obtained for β-actin (top right), assessed after stimulation as in blots below. (d) Expression of anergy-related genes by CD4+ T cells stimulated for 2 h with vehicle, imiquimod or IRS 661 plus imiquimod (key). (e) ELISA of IL-2 and IFN-γ secreted by CD4+ T cells transduced with nontargeting or NFATC2-specific shRNA and stimulated for 3 d with anti-CD3 and anti-CD28 in the presence or absence of imiquimod. (f) ELISA of IL-2 and IFN-γ secreted by CD4+ T cells incubated for 2 h with vehicle or imiquimod in the presence or absence or IRS 661, then washed and allowed to ‘rest’ for 12 h, then stimulated for 2 d with anti-CD3 and anti-CD28. *P < 0.05, **P < 0.005 and ***P < 0.0005 (paired t-test (c,e) or one-way ANOVA with Tukey’s post-test (d,f)). Data are representative of six independent experiments (a,b), five experiments (c,d), three experiments (e) or four experiments (f) with one donor in each (mean and s.e.m. in c–f).

proinflammatory cytokines IL-6, tumor-necrosis factor and IL-1β and and anti-CD28 in the presence or absence of imiquimod and meas- decreased their secretion of IL-10 (Supplementary Fig. 3b), whereas ured the secretion of IL-2 and IFN-γ. While CD4+ T cells transduced we observed no effect on monocyte proliferation (data not shown); with nontargeting shRNA showed a decrease in the production this suggested that stimulation of CD4+ T cells and CD14+ monocytes of IL-2 and IFN-γ secretion after treatment with imiquimod, this with imiquimod led to completely different outcomes. effect was abolished in cells expressing TLR7-specific shRNA To confirm the specificity of the unresponsive phenotype driven by (Fig. 2e), which confirmed that the unresponsive state of T cells TLR7 signaling, we stimulated CD4+ T cells with anti-CD3 and anti- observed in the presence of imiquimod was TLR7 specific. CD28 in the presence of shRNA specific for TLR7 or nontargeting shRNA, as a control. After 5 d of culture, we confirmed that the effi- Induction of calcium-NFATc2–driven CD4+ T cell anergy by TLR7 ciency with which protein and RNA was silenced was >80% (Fig. 2d). The inhibition of the proliferation, cytokine secretion and activation We stimulated shRNA-treated resting CD4+ T cells with with anti-CD3 of CD4+ T cells upon stimulation in the presence of TLR7 ligands

nature immunology VOLUME 16 NUMBER 1 JANUARY 2015 121 A rt i c l e s

a IMQ 200 IRS 661 + IMQ b IMQ 200 IRS 661 + IMQ Veh * IMQ Veh IMQ Veh * Veh 200 150 * 200 150 * 150 150 *** 100 100 ** 100 100 p-NF- κ B (MFI) p-IRAK4 (MFI) 50 50 50 50

Cells 0 Cells 0 0 102 103 104 105 15 30 45 60 90 120 0 102 103 104 105 15 30 45 60 90 120 p-IRAK4 Time (min) p-NF-κB Time (min) c IMQ 250 IRS 661+ IMQ d IMQ 400 IRS 661 + IMQ Veh IMQ Veh IMQ Veh Veh 200 200 200 * 300 ** 150 * 150 150 ** * 200 * 100 100 100 p-Jnk (MFI) p-p38 (MFI) 100 50 50 50 Cells 0 Cells 0 0 102 103 104 105 15 30 45 60 90 120 0 102 103 104 105 15 30 45 60 90 120 p-Jnk Time (min) p-p38 Time (min)

Figure 4 Imiquimod inhibits the phosphorylation of Jnk. Phosphorylated IRAK4 (a), NF-κB (p65) (b), Jnk (c) and p38 (d) in CD4+ T cells stimulated for 30 min with vehicle or imiquimod (left) or for 15–120 min with vehicle alone or with imiquimod in the presence or absence of IRS 661 (right). *P < 0.05, **P < 0.005 and ***P < 0.0005 (paired t-test). Data are representative of eight independent experiments with one donor each (mean ± s.e.m. at right).

resembled the unresponsive phenotype that characterizes clonal stimulated for 40 min with imiquimod and probed with a phospho- anergy15. Various model systems have been used to induce clonal rylation-specific antibody to NFATc2 (Fig. 3c). anergy, including treatment with the calcium ionophore ionomy- After being dephosphorylated, NFATc2 translocates to the nucleus, cin16,17, and the activation of calcium signaling in the absence of where it becomes transcriptionally active22. We used nuclear and activating signals in costimulatory signaling pathways is common cytoplasmic protein extracts to further confirm translocation of in all these models. Thus, the main characteristic of an stimulus that NFATc2 from the cytoplasm to the nucleus upon stimulation with elicits anergy is its ability to induce an unopposed increase in intra- imiquimod (data not shown). This translocation of NFATc2 to the cellular calcium concentrations17. To test the hypothesis that TLR7 nucleus in the absence of a concomitant costimulatory signal leads signaling on CD4+ T cells was inducing clonal anergy, we stained to the transcription of a set of NFATc2-dependent, anergy-related sorted CD4+ T cells with the ratiometric calcium indicator Indo-1 genes that are different from those upregulated with full activation AM and treated them with various doses of imiquimod (Fig. 3a). in the presence of a costimulatory signal17. To investigate whether Imiquimod induced a significant and sustained (maintained for at stimulation of TLR7 and subsequent dephosphorylation of NFATc2 Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature least 20 min) increase in intracellular calcium concentration in a induce the expression of anergy-related genes17, we incubated CD4+ dose-dependent manner that was TLR7 specific, as blocking TLR7 T cells for 2–16 h with imiquimod in the presence or absence of the with IRS 661, a specific inhibitory oligonucleotide sequence18, TLR7-inhibitory sequence IRS 661 and isolated RNA for analysis. As

npg impaired the increase in calcium concentration upon stimulation a control for the expression of anergy-related genes, we incubated with imiquimod (Fig. 3b and Supplementary Fig. 4a). We did not CD4+ T cells with either ionomycin plus the phorbol ester PMA (a observe that increase in intracellular calcium concentration in cells non-anergic stimulus) or ionomycin alone (an anergic stimulus)17 stimulated with the TLR8 agonist ssRNA40 or with ligands of other (Supplementary Fig. 5). Ten of the twelve anergy-related genes exam- intracellular TLRs, such as the synthetic RNA duplex poly(I:C) ined were significantly upregulated in cells stimulated with imiqui- (for TLR3) or ODN2006 (for TLR9) (Fig. 3a and Supplementary mod relative to their expression in the control cells stimulated with Fig. 4b). As a positive control, we treated cells with the calcium vehicle (Fig. 3d and Supplementary Fig. 5). The effect observed in ionophore ionomycin, which has been used as an anergy-inducing the regulation of these genes was TLR7 specific, as preincubation of agent in in vitro experiments17,19, and obtained concentrations of CD4+ T cells with IRS 661 before treatment with imiquimod abro- intracellular calcium similar to those induced by 10 µg/ml of imiq- gated the increase in the expression of anergy-related genes (Fig. 3d). uimod (Fig. 3a and Supplementary Fig. 4a). Furthermore, we also assessed the expression of other genes encod- An immediate consequence of an increase in the concentration ing molecules that have been functionally related to establishment of intracellular calcium is the activation of NFATc2, a transcription and maintenance of the anergic phenotype and are targets of NFAT, factor that is highly phosphorylated in resting cells and becomes such as SIRT1 (ref. 23), ITCH24, CBLB24,25, DGKA26, EGR2 and EGR3 dephosphorylated by the calcium-calmodulin–dependent phos- (ref. 27), and found that all of these but EGR2 showed significant phatase calcineurin when the concentration of intracellular calcium upregulation upon treatment with imiquimod (Fig. 3d). Our data is increased20,21. To determine whether stimulation with imiquimod suggested that imiquimod induced clonal anergy in CD4+ T cells via induced dephosphorylation of NFATc2, we stimulated CD4+ T cells an increase in intracellular calcium concentration and activation of an with imiquimod and purified total protein extracts 0, 45 and 90 min NFATc2-dependent anergy-related gene-expression program. To fur- later. Immunoblot analysis with anti-NFATc2 showed that treat- ther investigate the role of NFATc2 in TLR7-induced T cell anergy, we ment with imiquimod resulted in dephosphorylation of NFATc2; we used two shRNAs to silence NFATC2 and used imiquimod to induce confirmed these results by analysis of total extracts of CD4+ T cells anergy in resting CD4+ T cells after knockdown of NFATc2. After 3 d

122 VOLUME 16 NUMBER 1 JANUARY 2015 nature immunology A rt i c l e s

Figure 5 Imiquimod inhibits the activation Veh a PMA + iono + IMQ of Jnk and Jun after full stimulation of CD4+ PMA + iono PMA + iono + IMQ PMA + iono T cells through signaling via the TCR and 500 300 costimulation. (a) Phosphorylated NF-κB (p65), 200 400 200 + 200 Jnk, p38 and Jun in CD4 T cells stimulated 150 300 150 * * * for 60 min (left of each pair) or 15–120 min 100 200 100 100 (right of each pair) with vehicle (left) or with 50 100 50 p-Jnk (MFI) PMA and ionomycin (iono) in the presence or p-NF- κ B (MFI) Cells Cells 0 0 absence of imiquimod (right). (b) Expression of 0102 103 104 105 15 30 45 60 90 120 0102 103 104 105 15 30 45 60 90 120 + FOS and JUN mRNA in CD4 T cells stimulated p-NF-κB Time (min) p-Jnk Time (min) for 12–96 h with anti-CD3 and anti-CD28 500 600 of in the presence or absence of imiquimod. 200 * 400 ** 200 *P < 0.05, **P < 0.005 and ***P < 0.0005 ** 400 * 150 300 150 (paired t-test). Data are representative of * * * 100 200 100 eight independent experiments with one 200 p-Jun (MFI) donor each (a) or four independent experiments 50 p-p38 (MFI) 100 50 (b; mean ± s.e.m. of n = 4 donors). Cells 0 Cells 0 0102 103 104 105 15 30 45 60 90 120 0102 103 104 105 15 30 45 60 90 120 p-p38 Time (min) p-Jun Time (min) ) in culture, cells transduced with nontarget- ) 3 b 3 α-CD3 + α-CD28 ing shRNA and stimulated with imiquimod 500 *** 6 α-CD3 + α-CD28 + IMQ significantly reduced their production of 400 ** 4 IL-2 and IFN-γ, while cells transduced with 300 200 * NFATC2-specific shRNA were not affected * 2 by treatment with imiquimod (Fig. 3e 100 ** and Supplementary Fig. 6); this suggested 0 0

JUN mRNA (relative × 10 12 24 36 48 72 96 12 24 36 48 72 96 FOS mRNA (relative × 10 that NFATc2 was necessary for the anergic Time (h) Time (h) phenotype driven by TLR7 signaling in CD4+ T cells. Treatment with imiquimod did not have an effect on the imiquimod, then assessed the phosphorylation status of IRAK4, expression of NFATC2 (Supplementary Fig. 6a). In agreement with Jnk, the p65 subunit of NF-κB and the mitogen-activated protein those data and published observations17, treatment with imiquimod kinase p38 by flow cytometry. In agreement with published reports failed to upregulate the expression of anergy-related genes (KMD6B, of other cell types2,3,5, stimulation of TLR7 with imiquimod induced IKZF1, GRG4 and RAB10) in CD4+ T cells in which NFATc2 was phosphorylation of IRAK4, NF-κB and p38 at different time points knocked down relative to their expression in cells transduced with than did treatment with vehicle (Fig. 4). Preincubation with IRS 661 nontargeting shRNA (Supplementary Fig. 6b), which confirmed that before stimulation with imiquimod inhibited the phosphorylation this gene-expression program driven by treatment of CD4+ T cells of these molecules (Fig. 4), which suggested that the effect observed with imiquimod was largely NFATc2 dependent. was TLR7 specific. Incubation with IRS 661 by itself did not have any As reported in published studies of an ionomycin-induced effect on protein phosphorylation (data not shown). Engagement of anergy model17, we hypothesized that pretreatment of CD4+ T cells TLR7 decreased the basal levels of phosphorylated Jnk (Fig. 4). One Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature with imiquimod would be sufficient to diminish their subsequent of the targets phosphorylated by activated Jnk is Jun, a component cytokine secretion and proliferative response to stimulation with both of AP-1, which is an essential transcription factor involved in cos- crosslinking of the TCR and costimulatory signaling through CD28. timulatory signal transduction17. We hypothesized that inhibition of + npg We pretreated memory CD4 T cells for 2 h with imiquimod in the Jnk activity by imiquimod might explain, at least in part, the anergic presence or absence of the inhibitory sequence IRS 661 and allowed phenotype observed in CD4+ T cells after stimulation with imiqui- them to ‘rest’ for 12 h after washout to remove all traces of imiquimod in the presence of full stimulation of the TCR and costimula- mod and then stimulated the cells for 3 d with anti-CD3 and anti- tory signaling (Figs. 1 and 2). To test this hypothesis, we stimulated CD28. Cells pretreated with imiquimod showed significantly lower CD4+ T cells with PMA plus ionomycin in the presence or absence production of both IL-2 and IFN-γ than that of cells pretreated with of imiquimod and assessed the phosphorylation status of Jnk and Jun vehicle (Fig. 3f). We observed no significant change in cell viability at various time points. We used NF-κB and p38 as positive controls. upon pretreatment wtih imiquimod (data not shown). Again, the pro- Although treatment with imiquimod did not produce an additive anergic effect of imiquimod was prevented by pretreatment with the effect on the phosphorylation of NF-κB with stimulation, imiqui- TLR7 antagonist IRS 661 (Fig. 3f). These data suggested that imiq- mod further increased the phosphorylation of p38 (Fig. 5a). Of note, uimod induced clonal anergy in CD4+ T cells by inducing calcium- upon stimulation with PMA plus ionomycin, the phosphorylation dependent activation of NFATc2 and subsequent expression of anergy- of Jnk was inhibited by imiquimod, and this decrease in Jnk activity related genes. was accompanied by a decrease in the phosphorylation of Jun (Fig. 5a). Moreover, imiquimod decreased the activity of both Jnk and Effect of TLR7 signaling on costimulatory signals Jun, as measured by phosphorylation after activation (Fig. 5a), TLR7 signaling has been studied mainly in APCs, in which it is and it decreased JUN expression (Fig. 5b). These data supported linked to activation of the transcription factors IRF7 and NF-κB the hypothesis that imiquimod treatment both induced anergy in and the kinase Jnk through pathways dependent on or independ- CD4+ T cells and interfered with costimulatory signaling during T ent of the adaptor MyD88 and the kinase IRAK4 (ref. 1). To assess cell stimulation. The decrease in the expression of CD69 and CD137 the consequences of TLR7 signaling on CD4+ T cells relative to the (both transcriptional targets of AP-1 (refs. 28,29)) observed after consequences of ‘conventional’ TLR signaling, we isolated CD4+ stimulation in the presence of imiquimod (Fig. 1f) further supported T cells from healthy donors and stimulated the cells ex vivo with this hypothesis.

nature immunology VOLUME 16 NUMBER 1 JANUARY 2015 123 A rt i c l e s

Inhibition of HIV-1 in vitro via TLR7-Ca2+ signaling blockade RNA virus2,3 could mediate CD4+ T cell anergy. The responses of To investigate the potential biological and clinical relevance of the CD4+ T cells from patients chronically infected with HIV-1 are observations noted above in ex vivo and in vitro model systems, impaired and are insufficient to clear the virus, while the cells display we determined whether engagement of TLR7 by a single-stranded features of anergy30–33. Several viral proteins have been shown to

NT TLR7(3) TLR7(4) IRS 661 (µM) 0 1 20 a b NT c 30 *** 12.1 2.06 3.1 + TLR7(3) 15.8 2.3 0.3 TLR7(4) CD4

+ 20 **

NL-D * T cells (%) NL-D

10 NL-D HIV-1 HIV-1 0 HIV-1 Viability 0 1 2 3 4 5 6 7 8 9 10 11 Viability Time (d) NT TLR7(3) TLR7(4) d 20 e NT + Mock f + HIV-1NL-D + HIV-1NL-D + HIV-1NL-D 80 20 4 6.4 8.3 2.9 4.8 3.8 12.9 3.2 6.4 + 10 15 60 15 3

CD4 10 + 40 10 10 2 * *

10 cells (%) NL-D cells (%) + + 20 5

T cells (%) ** 1 5 10 IL-2 HIV-1 0 IFN- γ 0 *** 0 55.7 19.4 59.4 41.3 *** *** IFN- γ D D *** 1 2 3 4 0 0 10 10 10 10 NL-D NL-D NL- NL-D NL-D NL- 0 0.5 1 5 10 20 IL-2 HIV-1 NT + Mock HIV-1 NT + Mock IRS 661 (µM) NT + HIV-1 NT +

TLR7(3)TLR7(4) + HIV-1 + TLR7(3)TLR7(4) + HIV-1 + HIV-1 – + + g HIV-1 HIV-1 (IRS Ctrl) HIV-1 (IRS 661) h NL-D NL-D NL-D 25 Veh ) ) ) 3 ) Veh Quin-2 AM 3 ** 3 3 Quin-2 AM 300 1,500 300 1,000 20 104 13.5 5.0

** * T cells (%)

* 800 3 + 15 200 1,000 200 10 * 600 *

2 CD4 10

10 +

400 NL-D 100 500 100

1 NL-D 5 200 10

HIV-1 0 0

0 0 0 0 HIV-1 1 2 3 4

FASL mRNA (relative ×10 0 10 10 10 10

CD98 mRNA (relative ×10 5 7 9 11 GRG4 mRNA (relative ×10 CASP3 mRNA (relative ×10 Viability Time (d) )

) i

3 NT NFATc2(1) NFATc2(2) ) ) 3 20 3 3 NT ** 1,500 400 1,500 250 104 10.6 0.6 0.4 NFATc2(1) * 15 NFATc2(2) 300 * ** 200 103 T cells (%)

1,000 1,000 + 150 *** 200 102 10 **

100 CD4 NL-D 500 500 101 + 100 50 5 0 NL-D

LDHA mRNA (relative ×10 0 IKZF1 mRNA (relative ×10 RAB10 mRNA (relative ×10 HIV-1 KMD6B mRNA (relative ×10 Viability 5 7 9 11 Time (d) ) 3 ) ) 3 ) 3 3 j 400 300 400 400 Veh VIVIT 25 Veh 4 10 VIVIT

npg ** 300 ** 300 ** 300 11.3 1.4 20 200 3 10 T cells (%) 200 200 200 + 15 2 100 10 *** **

CD4 10 NL-D 100 100 100 1 + 10

0 0 0 0 0 NL-D 5 HIV-1

RGS2 mRNA (relative ×10 * TRAF5 mRNA (relative ×10 SOCS2 mRNA (relative ×10 1 2 3 4

TNFRSF9 mRNA (relative ×10 0 10 10 10 10 0 HIV-1 Viability 5 7 9 11 Time (d)

Figure 6 Knockdown of TLR7 or NFAT abolishes infection with HIV-1. (a) Viability of CD4+ T cells transduced with nontargeting shRNA or TLR7-specific shRNA (clone 3 or 4) (above plots), then infected with HIV-1NL-D 2 d later, and assessed 7 d after infection. Numbers adjacent to outlined areas indicate + + + percent viable HIV-1 cells. (b) Kinetics of the infection shRNA-transduced CD4 T cells (as in a) with HIV-1NL-D. (c) Viability of CD4 T cells stimulated for 2 d with anti-CD3 and anti-CD28 in the presence of various doses of IRS 661 (above plots), then infected with HIV-1NL-D and assessed at day 7 after + + infection (numbers in plots as in a). (d) Frequency of HIV-1NL-D cells among CD4 T cells incubated with various doses of IRS 661 (horizontal axis) and + infected with HIV-1NL-D. (e) Expression of IFN-γ and IL-2 by CD4 T cells stimulated for 2 d with anti-CD3 and anti-CD28 in the presence of shRNA as in a, then mock infected (Mock) or infected with HIV-1NL-D and then, 7 d later, stimulated for 4 h with PMA and ionomycin. (f) Frequency of IL-2- or IFN-γ- producing CD4+ T cells among cells treated as in e. (g) Expression of anergy-related genes by CD4+ T cells left uninfected with no further treatment − – + (HIV-1NL-D ) or infected in vitro with HIV-1NL-D in the presence of IRS 661 or a control sequence, and sorted (as HIV-1NL-D (uninfected) or HIV-1NL-D + (infected)) at day 7 after infection. (h) Viability of CD4 T cells preincubated with vehicle or Quin-2 AM (above plots) and then infected with HIV-1NL-D, assessed at day 7 after infection (left) or on days 5–11 after infection (right). (i) Viability of CD4+ T cells stimulated in the presence of nontargeting shRNA or NFATC2-specific shRNA (clone 1 or 2) and infected with HIV-1NL-D 2 d later, assessed at day 7 after infection (left) or on days 5–11 after infection (right). + (j) Viability of CD4 T cells preincubated with vehicle or VIVIT peptide and infected with HIV-1NL-D, assessed at day 7 after infection (left) or on days 5–11 after infection (right). Numbers in plots (left, h–j) as in a. *P < 0.05, **P < 0.005 and ***P < 0.0005 (one-way ANOVA with Tukey’s post-test (b,f,g,i,j) or paired t-test (d,h)). Data are from one experiment representative of five (a,c) or four (e) independent experiments with one donor each (a,c,e) or are representative of five independent experiments (b,d; mean ± s.e.m. of n = 5 donors), four independent experiments (f; mean ± s.e.m. of n = 4 donors), six independent experiments (g; mean and s.e.m. of n = 6 donors) or six independent experiments with one donor each (h–j; mean ± s.e.m. at right).

124 VOLUME 16 NUMBER 1 JANUARY 2015 nature immunology A rt i c l e s

a b c d Iono IMQ ( g/ml) 0 1 10 Veh Iono Veh µ 50 ** 40 4 4 10 17.8 + ** 10 + 8.3 41.0 40 11.5 24.1 * 3 3 30 10 10 CD4 CD4 ** + 2 + 30 * 10 102 20 * NL-D NL-D NL-D 1 NL-D 20 10 101 T cells (%) T cells (%) 10 * 10 0 HIV-1 0 HIV-1 HIV-1 HIV-1 0 101 102 103 104 0 0 101 102 103 104 0 Viability 0 0.1 0.5 1 5 10 Viability 3 5 7 9 11 Time (d) IMQ (µg/ml) Figure 7 Calcium-induced anergy favors + e Veh α-CD3 f g HIV-1 replication. (a) Viability of CD4 40 α-CD3 25 CsA 4 + 10 12.3 20.1 + Veh * Veh T cells stimulated for 2 d with anti-CD3 20 ** 3 30 10 CD4 and anti-CD28 in the presence of various CD4 + + 15 102 20 * ***

doses of imiquimod (above plots) and NL-D * NL-D NL-D 10 * 101 T cells (%) infected with HIV-1NL-D, assessed at day 7 T cells (%) 10 0 5 HIV-1 HIV-1 after infection. (b) Frequency of HIV-1 * 1 2 3 4 HIV-1 + CD4+ T cells among cells 0 10 10 10 10 0 0 NL-D Viability 3 5 7 9 11 5 7 9 11 + treated as in a. (c) Viability of CD4 T cells Time (d) Time (d) preincubated for 12 h with vehicle or ionomycin + + (above plots), then infected with HIV-1NL-D and assessed at day 9 after infection. (d) Frequency of HIV-1NL-D CD4 T cells among cells treated as in c, + assessed on days 3–11 after infection. (e) Viability of CD4 T cells preincubated for 12 h with vehicle or anti-CD3 (above plots), then infected with HIV-1NL-D + + and assessed at day 9 after infection. (f) Frequency of HIV-1NL-D CD4 T cells among cells treated as in e, assessed on days 3–11 after infection. + + (g) Frequency of HIV-1NL-D CD4 T cells among cells stimulated for 2 d with anti-CD3 and anti-CD28, with vehicle or cyclosporine A (CsA) added to the cultures 6 h before infection with HIV-1NL-D, assessed on days 5–11. Numbers in plots (a,c,e) as in Figure 6a. *P < 0.05, **P < 0.005 and ***P < 0.0005 (paired t-test). Data are from one experiment representative of three (a) or six (c,e) independent experiments with one donor each (a,c,e) or are representative of three independent experiments with one donor each (b; mean ± s.e.m.) or six independent experiments (d,f,g; mean ± s.e.m. of n = 6 donors).

induce the state of T cell unresponsiveness that precedes the loss of of early or late apoptotic CD4+ T cells after infection (Supplementary CD4+ T cells in HIV-1-infected patients30. We hypothesized that a Fig. 8). We further confirmed the role of TLR7 in decreasing the + + virus–CD4 T cell interaction through TLR7 would be responsible, HIV-1NL-D infection rate by blockade of TLR7 on CD4 T cells at least in part, for the anergic phenotype observed in HIV-1-infected from healthy donors with various doses of IRS 661 before in vitro CD4+ T cells from patients. First, we isolated CD4+ T cells from four infection (Fig. 6c,d). Moreover, the secretion of cytokines by healthy donors, infected the cells in vitro with physiological concen- HIV-1NL-D-infected cells was significantly different for cells transduced trations (multiplicity of infection, 0.001) of a replication-competent with nontargeting shRNA and their counterparts transduced with strain of HIV-1 derived from the prototype HIV-1NL432 virus and TLR7-specific shRNA (Fig. 6e,f). Although HIV-1NL-D-infected cells 34,35 + tagged with the red fluorescent reporter DsRed (HIV-1NL-D) and in cultures of CD4 T cells transduced with nontargeting shRNA assessed their ability to produce IL-2 and IFN-γ 7 d after infection. produced less IL-2 and IFN-γ protein than did mock-infected CD4+ 30–33 + In agreement with published reports , infection with HIV-1NL-D T cells, the low frequency of HIV-1NL-D-infected CD4 T cells trans- Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature markedly diminished the ability of viable CD4+ T cells to produce duced with TLR7-specific shRNA secreted significantly more IL-2 IL-2 and IFN-γ after 4 h of stimulation with PMA and ionomycin and IFN-γ than did their counterparts transduced with nontargeting (Supplementary Fig. 7). We did not observe this decrease in cytokine shRNA (Fig. 6e,f), which suggested that the anergic phenotype did

npg secretion only in the bulk T cell population infected with the virus but not develop after infection in cells transduced with TLR7-specific specifically in HIV-1NL-D-infected cells, which suggested that direct shRNA. To confirm that observation, we assessed the expression of interaction of the virus with infected CD4+ T cells rendered the CD4+ anergy-related genes in CD4+ T cells obtained from healthy donors, T cells unresponsive. infected with HIV-1NL-D in vitro in the presence of IRS 661 (to inhibit To test the hypothesis that TLR7 signaling in CD4+ T cells accounted TLR7) or a control sequence and assessed at day 7 after infection. + + in part for the anergic phenotype after infection with HIV-1, we iso- Sorted HIV-1NL-D CD4 T cells showed higher expression of eight of + – + lated CD4 T cells from healthy donors and treated the cells ex vivo the twelve anergy-related genes examined than did HIV-1NL-D CD4 with either of two shRNAs specific for TLR7 (clone 3 or 4), to silence T cells, while HIV-1+ CD4+ T cells sorted from IRS 661–treated cul- TLR7, or with nontargeting shRNA, as a control (Fig. 6). After 2 d, tures did not upregulate any of these genes at the time point analyzed we infected the shRNA-treated CD4+ T cells with concentrations of (Fig. 6g). The expression of other genes encoding molecules that have 34 HIV-1NL-D within the physiological range (multiplicity of infection, been functionally linked to anergy, such as ITCH, DGKA, CBLB and + + 0.001) and measured the frequency of infected cells as well as their SIRT1, was also upregulated in HIV-1NL-D CD4 T cells but not in + ability to produce proinflammatory cytokines every 48 h for a total IRS 661–treated HIV-1NL-D T cells (data not shown). These data sug- + of 11 d. Although CD4 T cells transduced with nontargeting shRNA gested that an interaction of HIV-1NL-D with TLR7 was responsible, were infected with HIV-1NL-D to an extent similar to that of untrans- at least in part, for the anergic phenotype observed in HIV-1NL-D- duced cells, cell populations transduced with TLR7-specific shRNA infected CD4+ T cells. + had a much lower frequency of HIV-1NL-D-infected CD4 T cells at We investigated the role of the intracellular calcium– and NFATc2- all time points examined (Fig. 6a,b). This lower frequency of HIV-1+ activated gene-expression signaling events we observed by engage- cells was not due to a specific increase in the death of cells in which ment of TLR7 with imiquimod, after infection with HIV-1NL-D TLR7 was silenced, as there was no difference between cell popula- in vitro. We hypothesized that blockade of intracellular calcium and tions transduced with TLR7-specific shRNA and control cell popula- silencing of NFATC2 would lead to a decrease in the frequency of tions transduced with nontargeting shRNA in terms of the frequency HIV-1+ T cells even in the presence of a functional TLR7. To induce

nature immunology VOLUME 16 NUMBER 1 JANUARY 2015 125 A rt i c l e s

a IRS Ctrl b 2.0 IRS Ctrl c NT d 2.5 NT e 100 IRS Ctrl f 100 NT 1.5 IRS 661 IRS 661 1.5 TLR7(4) TLR7(3) IRS 661 TLR7(3) TLR7(3) 2.0 TLR7(4) TLR7(4) 1.5 80 80

1.0 1.0 1.5 60 60 1.0 1.0 40 40 p24 (ng/ml) 0.5 0.5 p24 (ng/ml) p24 (ng/ml) p24 (ng/ml) 0.5 ** ** 0.5 * * 20 * 20 ** DNA proviral load (%) DNA proviral load (%)

0 0 0 0 0 0 4 7 9 11 14 4 7 9 11 14 Time (d) Time (d)

Figure 8 Inhibition of TLR7 decreases infection in HIV-1+ patients. (a) Concentration of p24 in CD4+ T cells isolated from an HIV-1-infected patient and stimulated for 4–14 d (horizontal axis) in vitro with anti-CD3 and anti-CD28 in the presence of control sequence (IRS Ctrl) or IRS 661 (key). (b) Concentration of p24 in CD4+ T cells treated as in a, assessed on day 14. (c) Concentration of p24 in CD4+ T cells isolated from an HIV-1-infected patient and stimulated for 4–14 d (horizontal axis) in vitro with anti-CD3 and anti-CD28 in the presence of nontargeting shRNA or TLR7-specific shRNA (clone 3 or 4). (d) Concentration of p24 in CD4+ T cells treated as in c, assessed on day 14. (e) DNA proviral load in CD4+ T cells obtained from HIV-1-infected patients and stimulated for 11 d in vitro as in a. (f) DNA proviral load in CD4+ T cells from five HIV-1-infected patients and stimulated for 11 d in vitro in the presence shRNA as in d. *P < 0.005 and **P < 0.0005 (paired t-test (b,e) or one-way ANOVA with Tukey’s post-test (d,f)). Data are from one experiment representative of seven independent experiments with nine donors (a,c) or are representative of seven independent experiments (b,d–f; mean and s.e.m. of n = 9 donors).

calcium blockade, we preincubated CD4+ T cells with the chelation increase in intracellular calcium concentration were prerequisites for 36 agent Quin-2 AM before infection with HIV-1NL-D. Calcium chela- productive infection with HIV-1. tion significantly decreased the frequency of viable cells, even at the To further investigate whether inhibiting TLR7-induced anergy lowest concentration of the chelating agent (Fig. 6h), perhaps due to would affect the frequency of HIV-1-infected cells, we added either the essential role of calcium in many cellular processes. Nevertheless, the calcineurin inhibitor cyclosporine A or concentrations of IL-2 + 17 there was a lower frequency of HIV-1NL-D T cells among the remain- that have been shown to reverse anergy in several in vitro settings + ing viable CD4 T cells in cultures treated with Quin-2 AM than before in vitro infection of cells with HIV-1NL-D. Although the addi- + among CD4 T cells treated with vehicle alone (Fig. 6h), in agreement tion of IL-2 6 h before infection with HIV-1NL-D did not affect the with published investigations suggesting a role for calcium in the frequency of HIV-1NL-D-infected cells (data not shown), perhaps HIV-1 life cycle37. We then silenced NFATC2 with either of two differ- due to response kinetics, the broad spectrum of IL-2 functions or ent shRNAs or blocked NFATc2 with VIVIT peptide, which interferes the inability of IL-2 to ‘rescue’ CD4+ T cells from TLR7-induced anergy, with the calcineurin-NFATc2 interaction and inhibits dephosphor- blocking anergy with cyclosporine A significantly decreased the 38 + ylation of NFAT . In both cases, the absence of functional NFATc2 frequency of HIV-1NL-D T cells (Fig. 7g). This further supported led to a significantly lower frequency of HIV-1NL-D-infected T cells the hypothesis that HIV-1-induced anergy was necessary for than among CD4+ T cells left untreated or treated with vehicle alone productive infection with HIV-1. (Fig. 6i,j), which indicated a role for NFATc2 in HIV-1 infection, as previously suggested39. These data suggested that calcium-dependent Inhibition of TLR7 diminishes HIV-1 in T cells ex vivo Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature activation of NFATc2 upon triggering of TLR7 was involved in We next sought to directly investigate the role of TLR7 in CD4+ HIV-1NL-D infection in vitro. T cells from HIV-1-infected patients. Specifically, on the basis of the in vitro model system with HIV-1NL-D, we hypothesized that inhi- + npg Calcium-induced anergy favors HIV-1 replication bition of the TLR7 pathway in CD4 T cells from HIV-1-infected Given the decreased infection rate of cells in which TLR7 was silenced patients would decrease the infection rate. We isolated CD4+ T cells and the non-anergic phenotype of infected CD4+ T cells in which from HIV-1 infected patients (Supplementary Table 1) and stimu- TLR7 was silenced, we hypothesized that the anergic state induced by lated the cells in the presence of IRS 661 or transduced them with stimulation via TLR7 during infection with HIV-1 would be a neces- either of two TLR7-specific shRNAs, then collected supernatants sary step for HIV-1 to replicate in the host. In the absence of signal- every 3 d for a total of 14 d to measure virus concentration by ELISA ing via TLR7, the virus would not be able to render the infecting cell of the HIV-1 core antigen p24. The inhibition of TLR7 by IRS 661 anergic and long-term infection would not occur. To test this hypoth- substantially decreased the concentration of p24 in culture (Fig. 8a,b), esis, we obtained CD4+ T cells from healthy donors, induced anergy and we obtained similar results for cells transduced with TLR7- in the cells via the TLR7 pathway with various doses of imiquimod specific shRNA (Fig. 8c,d). We assayed CD4+ T cells from healthy and subsequently infected them with HIV-1NL-D and analyzed them donors in parallel as a negative control but found no detectable p24 + 7 d later. Of note, the frequency of HIV-1NL-D-infected CD4 T cells at any time point (data not shown). Furthermore, we measured the directly correlated with the concentration of imiquimod used and the load of proviral integrated DNA at day 7 after stimulation with anti- increase in intracellular calcium (Fig. 7a,b), and thus with the degree CD3 and anti-CD28 in the presence of IRS 661 or a control sequence of anergy in the culture, which suggested that calcium-induced anergy and after transduction with TLR7-specific or nontargeting shRNA favored infection with HIV-1. Moreover, the induction of anergy via to monitor the cellular viral reservoir. Inhibition of TLR7 by either other well-established in vitro methods, such as treatment with iono- IRS 661 (Fig. 8e) or knockdown of TLR7 (Fig. 8f) resulted in a mycin (Fig. 7c,d) or stimulation with anti-CD3 without costimulatory significantly lower proviral DNA load than that in control cells. signals (Fig. 7e,f), before infection with HIV-1NL-D also increased Together these results showed that inhibition of TLR7 signaling was + the frequency of HIV-1NL-D-infected cells. These data supported the able to diminish the HIV-1 load in infected CD4 T cells derived hypothesis that HIV-1-induced anergy via ligation of TLR7 and an from infected patients.

126 VOLUME 16 NUMBER 1 JANUARY 2015 nature immunology A rt i c l e s

DISCUSSION the induction of the anergic phenotype observed in HIV-1-infected Microbial pattern–recognition receptors are critical for the early sens- cells and in increases in the degree of infection of CD4+ T cells, which ing of diverse bacterial and viral infections for activation of the innate suggests that TLR7-induced upregulation of calcium is involved in the immune system. Here we have described a previously unknown func- viral life cycle. In this context, TLR8-dependent activation of NF-κB tion of these pattern-recognition receptors in shutting off the immune has been shown to be critical for HIV-1 replication in dendritic cells43. responses of human CD4+ T cells. Engagement of TLR7 by its ligands Although we did not investigate the role of TLR8 in the infection in CD4+ T cells prevented entry into the cell cycle and the secretion of CD4+ T cells by HIV-1, it is possible that contributions of both of proinflammatory cytokines after stimulation. Mechanistically, TLR7-driven activation of NF-κB and TLR7-induced anergy have stimulation of TLR7 increased intracellular calcium concentrations, a role in the increased viral load. The observation that anergy which led to dephosphorylation of NFATc2 and its translocation favored HIV-1 infection is somewhat paradoxical given existing lit- to the cell nucleus; this activated an anergic gene-expression pro- erature showing that HIV-1 infection is favored in activated CD4+ gram. Furthermore, TLR7 signaling interfered with costimulatory T cells11. We speculate that HIV-1 might trigger TLR7 to induce signals, as activation of Jnk and Jun was inhibited in the presence of an increase in intracellular calcium concentrations, which has TLR7 ligands and full stimulation through the TCR and costimulation been suggested to be important for the viral life cycle37. Whether via CD28. Our results have potential clinical implications, as silenc- the induction of anergy in host cells by HIV-1 infection via TLR7 ing of TLR7 inhibited the frequency of cells infected in vitro with is a bystander consequence of the increase in intracellular calcium HIV-1NL-D, and those infected cells did not display an anergic phe- driven by TLR7, and whether this status is beneficial for the host notype, which suggested that the calcium-induced TLR7-dependent or not, are currently unknown. However, these observations open anergic state of HIV-1NL-D-infected cells might have a role in up a new field of investigation related to the mechanism, kinetics HIV-1 persistence. and consequences of the interaction of HIV-1 with the host cell TLR signaling has been studied predominantly in APCs, in which through TLR7. engagement of the ligand induces the secretion of proinflammatory Although infections are generally regarded as illnesses, mammals cytokines and upregulation of activation molecules12–14. Although are colonized with bacteria and viruses. The ‘prime’ example of this TLR signaling in CD4+ T cells has not been studied in depth, the is bacteria adopted for digestion, although humans are also colonized few reports published have demonstrated a positive costimulatory with common DNA and RNA viruses, including endogenous retrovi- role for TLR signaling7,8,40. Our results have demonstrated a previ- ruses. It has been suggested that these endogenous retroviruses pro- ously undescribed TLR signaling pathway with an inhibitory effect on vide an adoptive selective advantage in generating genetic diversity44. T cell proliferation and cytokine secretion. The differences between On the basis of our data here, we speculate that TLR7 might have been the phenotypes of monocytes and those of CD4+ T cells upon TLR7 co-opted in human CD4+ T cells to co-evolve so that the cells do not stimulation could be due to the intrinsic differences in activation enter into the cell cycle in response to endogenous retroviruses, with requirements for each cell population. Although CD4+ T cells need potential consequences such as leukemia (HTLV-1) and autoimmune engagement of the TCR and a costimulatory signal to enter cell cycle, diseases. There are examples in the literature showing that certain CD14+ cells require one signal. Thus, engagement of TLR7 by its lig- endogenous retrovirus sequences are upregulated in human autoim- and is sufficient to induce activation and secretion of proinflammatory mune diseases45. Although it will be of interest in future investiga- cytokines in monocytes, as expected for a cell of the innate immune tions to elucidate precisely when TLR7 is engaged in the life cycle of system. In contrast, stimulation of TLR7 in CD4+ T cells leads to a HIV-1 during infection, these data demonstrate a mechanism by Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature significant increase in intracellular calcium that in the absence of a which HIV-1 may avoid elimination by co-opting NFAT-dependent costimulatory signal triggers the TLR7-driven anergic program17. In TLR7-induced T cell anergy. this context, our results also showed a decrease in CD4+ T cell prolif- In summary, we have demonstrated a previously unknown role for + npg eration upon stimulation of TLR9, with no effect on IFN-γ secretion TLR7 in CD4 T cell function that is in direct opposition to its role and a trend toward increased IL-17 secretion, although this result was in cells of the innate immune system. Moreover, TLR7 ligands may not significant. Furthermore, there was no increase in the intracellular be used as a means of inducing ‘tolerance’ on CD4+ T cells in human calcium concentration when CD4+ T cells were stimulated with CpG autoimmune diseases. Finally, our data have demonstrated a novel B, which suggested that the mechanism by which CpG B modifies function for microbial pattern–recognition receptors in the inhibition CD4+ T cell functionality is not common to ligation of TLR7. of immune responses. We note that all the ligands we used here were pharmacological agonists of TLR7. We have not been able to find any immunostimulatory Methods RNA specific for human TLR7, which would be a more physiological Methods and any associated references are available in the online ligand, as these immunostimulatory RNA sequences are specific for version of the paper. mouse TLR7 but recognize human TLR8 instead. We have overcome this issue through the use of HIV-1 as a potential ‘natural’ ligand of Note: Any Supplementary Information and Source Data files are available in the TLR7. The physiological relevance of our data is suggested by the online version of the paper. effect of TLR7 signaling on HIV-1-infected cells. Several reports have highlighted the role of innate sensors in infection with HIV-1, Acknowledgments We thank L. Devine and Z. Wang for technical assistance; Y. Tsunetsugu-Yokota focusing on the function of these molecules in CD4+ T cells that had (Tokyo University of Technology) for HIV-1NL-D proviral DNA; D. Bruce, not been productively infected41,42. Our results add a new layer of H. Zapata and B.C. Herold and the laboratory of B.C. Herold for the recruitment complexity to the understanding of the HIV viral life cycle, which of patients; and R. Medzhitov, A. Iwasaki and members of the Hafler laboratory needs further investigation, and are in contrast with published data for comments and suggestions. Supported by the National MS Society (CA1061- A-18), the US National Institutes of Health (P01 AI045757, U19 AI046130, U19 describing the cytoplasmic sensors of viral DNA that trigger cell death AI070352 and P01 AI039671 to D.A.H., and R01 AI065309 to M.J.K.), the Penates + 41,42 by pyroptosis of unproductively infected CD4 T cells . Moreover, Foundation (D.A.H.), the Nancy Taylor Foundation for Chronic Diseases (D.A.H.) the observations reported here suggest involvement of TLR7 both in and the Race to Erase MS Foundation (M.D.-V.).

nature immunology VOLUME 16 NUMBER 1 JANUARY 2015 127 A rt i c l e s

AUTHOR CONTRIBUTIONS 22. Okamura, H. et al. Concerted dephosphorylation of the transcription factor NFAT1 M.D.-V. designed and performed the experiments, analyzed data and wrote the induces a conformational switch that regulates transcriptional activity. Mol. Cell 6, manuscript; A.-S.G. and M.d.M. performed experiments; M.J.K. provided HIV-1 539–550 (2000). 23. Gao, B., Kong, Q., Kemp, K., Zhao, Y.S. & Fang, D. Analysis of sirtuin 1 expression samples; and D.A.H. assisted with the design of experiments, supervised the project reveals a molecular explanation of IL-2-mediated reversal of T-cell tolerance. and wrote the manuscript. Proc. Natl. Acad. Sci. USA 109, 899–904 (2012). 24. Heissmeyer, V. et al. Calcineurin imposes T cell unresponsiveness through targeted COMPETING FINANCIAL INTERESTS proteolysis of signaling proteins. Nat. Immunol. 5, 255–265 (2004). The authors declare no competing financial interests. 25. Jeon, M.S. et al. Essential role of the E3 ubiquitin ligase Cbl-b in T cell anergy induction. Immunity 21, 167–177 (2004). Reprints and permissions information is available online at http://www.nature.com/ 26. Zha, Y. et al. T cell anergy is reversed by active Ras and is regulated by diacylglycerol reprints/index.html kinase-alpha. Nat. Immunol. 7, 1166–1173 (2006). 27. Safford, M. et al. Egr-2 and Egr-3 are negative regulators of T cell activation. 1. Song, D.H. & Lee, J.O. Sensing of microbial molecular patterns by Toll-like receptors. Nat. Immunol. 6, 472–480 (2005). Immunol. Rev. 250, 216–229 (2012). 28. Castellanos, M.C. et al. Expression of the leukocyte early activation antigen CD69 2. Diebold, S.S., Kaisho, T., Hemmi, H., Akira, S. & Reis e Sousa, C. Innate antiviral is regulated by the transcription factor AP-1. J. Immunol. 159, 5463–5473 responses by means of TLR7-mediated recognition of single-stranded RNA. Science (1997). 303, 1529–1531 (2004). 29. Kim, J.O., Kim, H.W., Baek, K.M. & Kang, C.Y. NF-kappaB and AP-1 regulate 3. Heil, F. et al. Species-specific recognition of single-stranded RNA via toll-like activation-dependent CD137 (4–1BB) expression in T cells. FEBS Lett. 541, receptor 7 and 8. Science 303, 1526–1529 (2004). 163–170 (2003). 4. Kugelberg, E. Innate immunity: Making mice more human the TLR8 way. Nat. Rev. 30. Miedema, F. Immunological abnormalities in the natural history of HIV infection: Immunol. 14, 6 (2014). mechanisms and clinical relevance. Immunodefic. Rev. 3, 173–193 (1992). 5. Kawai, T. & Akira, S. Innate immune recognition of viral infection. Nat. Immunol. 31. Faith, A., O’Hehir, R.E., Malkovsky, M. & Lamb, J.R. Analysis of the basis of 7, 131–137 (2006). resistance and susceptibility of CD4+ T cells to human immunodeficiency virus 6. Kabelitz, D. Expression and function of Toll-like receptors in T lymphocytes. (HIV)-gp120 induced anergy. Immunology 76, 177–184 (1992). Curr. Opin. Immunol. 19, 39–45 (2007). 32. Cayota, A., Vuillier, F., Gonzalez, G. & Dighiero, G. In vitro antioxidant treatment 7. Gelman, A.E., Zhang, J., Choi, Y. & Turka, L.A. Toll-like receptor ligands directly recovers proliferative responses of anergic CD4+ lymphocytes from human promote activated CD4+ T cell survival. J. Immunol. 172, 6065–6073 (2004). immunodeficiency virus-infected individuals. Blood 87, 4746–4753 (1996). 8. Caron, G. et al. Direct stimulation of human T cells via TLR5 and TLR7/8: flagellin 33. Maggi, E. et al. Reduced production of interleukin 2 and interferon-γ and enhanced and R-848 up-regulate proliferation and IFN-gamma production by memory CD4+ helper activity for IgG synthesis by cloned CD4+ T cells from patients with AIDS. T cells. J. Immunol. 175, 1551–1557 (2005). Eur. J. Immunol. 17, 1685–1690 (1987). 9. Grakoui, A. et al. HCV persistence and immune evasion in the absence of memory 34. Chen, H. et al. CD4+ T cells from elite controllers resist HIV-1 infection by selective T cell help. Science 302, 659–662 (2003). upregulation of p21. J. Clin. Invest. 121, 1549–1560 (2011). 10. Lechner, F. et al. Analysis of successful immune responses in persons infected with 35. Yamamoto, T. et al. Selective transmission of R5 HIV-1 over X4 HIV-1 at the dendritic hepatitis C virus. J. Exp. Med. 191, 1499–1512 (2000). cell-T cell infectious synapse is determined by the T cell activation state. 11. Walker, B. & McMichael, A. The T-cell response to HIV. Cold Spring Harb. Perspect. PLoS Pathog. 5, e1000279 (2009). Med. 2, a007054 (2012). 36. Bigby, M., Wang, P., Fierro, J.F. & Sy, M.S. Phorbol myristate acetate-induced 12. Chamberlain, N.D. et al. Ligation of TLR7 by rheumatoid arthritis synovial fluid down-modulation of CD4 is dependent on calmodulin and intracellular calcium. single strand RNA induces transcription of TNFα in monocytes. Ann. Rheum. Dis. J. Immunol. 144, 3111–3116 (1990). 72, 418–426 (2013). 37. Cloyd, M.W., Lynn, W.S., Ramsey, K. & Baron, S. Inhibition of human 13. Cros, J. et al. Human CD14dim monocytes patrol and sense nucleic acids and immunodeficiency virus (HIV-1) infection by diphenylhydantoin (dilantin) implicates viruses via TLR7 and TLR8 receptors. Immunity 33, 375–386 (2010). role of cellular calcium in virus life cycle. Virology 173, 581–590 (1989). 14. Dzopalic, T. et al. Loxoribine, a selective Toll-like receptor 7 agonist, induces 38. Aramburu, J. et al. Affinity-driven peptide selection of an NFAT inhibitor more maturation of human monocyte-derived dendritic cells and stimulates their Th-1- selective than cyclosporin A. Science 285, 2129–2133 (1999). and Th-17-polarizing capability. Int. Immunopharmacol. 10, 1428–1433 (2010). 39. Kinoshita, S. et al. The T cell activation factor NF-ATc positively regulates HIV-1 15. Chappert, P. & Schwartz, R.H. Induction of T cell anergy: integration of environmental replication and gene expression in T cells. Immunity 6, 235–244 (1997). cues and infectious tolerance. Curr. Opin. Immunol. 22, 552–559 (2010). 40. Crellin, N.K. et al. Human CD4+ T cells express TLR5 and its ligand flagellin 16. Quill, H. & Schwartz, R.H. Stimulation of normal inducer T cell clones with antigen enhances the suppressive capacity and expression of FOXP3 in CD4+CD25+ presented by purified Ia molecules in planar lipid membranes: specific induction T regulatory cells. J. Immunol. 175, 8051–8059 (2005). of a long-lived state of proliferative nonresponsiveness. J. Immunol. 138, 41. Monroe, K.M. et al. IFI16 DNA sensor is required for death of lymphoid CD4 Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature 3704–3712 (1987). T cells abortively infected with HIV. Science 343, 428–432 (2014). 17. Macián, F. et al. Transcriptional mechanisms underlying lymphocyte tolerance. 42. Yan, N., Regalado-Magdos, A.D., Stiggelbout, B., Lee-Kirsch, M.A. & Cell 109, 719–731 (2002). Lieberman, J. The cytosolic exonuclease TREX1 inhibits the innate immune 18. Lenert, P.S. Classification, mechanisms of action, and therapeutic applications of response to human immunodeficiency virus type 1. Nat. Immunol. 11, 1005–1013 npg inhibitory oligonucleotides for Toll-like receptors (TLR) 7 and 9. Mediators Inflamm. (2010). 2010, 986596 (2010). 43. Gringhuis, S.I. et al. HIV-1 exploits innate signaling by TLR8 and DC-SIGN 19. Schwartz, R.H. Models of T cell anergy: is there a common molecular mechanism? for productive infection of dendritic cells. Nat. Immunol. 11, 419–426 J. Exp. Med. 184, 1–8 (1996). (2010). 20. Luo, C. et al. Recombinant NFAT1 (NFATp) is regulated by calcineurin in T cells 44. de la Casa-Esperón, E. Horizontal transfer and the evolution of host-pathogen and mediates transcription of several cytokine genes. Mol. Cell. Biol. 16, interactions. Int. J. Evol. Biol. 2012, 679045 (2012). 3955–3966 (1996). 45. Balada, E., Vilardell-Tarres, M. & Ordi-Ros, J. Implication of human endogenous 21. Luo, C. et al. Interaction of calcineurin with a domain of the transcription factor NFAT1 retroviruses in the development of autoimmune diseases. Int. Rev. Immunol. 29, that controls nuclear import. Proc. Natl. Acad. Sci. USA 93, 8907–8912 (1996). 351–370 (2010).

128 VOLUME 16 NUMBER 1 JANUARY 2015 nature immunology ONLINE METHODS ELISA. IFN-γ, IL-2, IL-17, IL-4 and IL-10 in supernatants of stimulated CD4+ Cell-culture reagents and antibodies. Cells were cultured in RPMI-1640 T cell culture and of tumor-necrosis factor, IL-1β and IL-6 in CD14+ cultures media supplemented with 2 nM l-glutamine, 5 mM HEPES, and 100 U/µg/ml were measured with an ELISA kit according to manufacturer’s recommenda- penicillin-streptomycin (Biowhittaker), 0.5 mM sodium pyruvate, 0.05 mM tions (BD Biosciences). Antibody pairs for IL-2, IL-17 and IL-10 were from nonessential amino acids (Life Technologies) and 5% human AB serum R&D Systems (Quantikine ELISA kits); biotin-labeled monoclonal antibody (Gemini Bio-Products). The antibodies used for stimulation were anti–human to IFN-γ ( 2G1) and monoclonal antibody to human IFN-γ (PA1-84955) CD3 (UCHT1 and Hit3a; BD Biosciences) and anti–human CD28 (28.2; BD were from Thermo Scientific; and antibody pairs for IL-10 (JES3-19F1 and Biosciences) at 1 µg/ml. The antibody to TLR7 (533707) used for staining JES3-12G8), IL-6 (MQ2-13A5 and MQ2-39C3), tumor-necrosis factor (MAb11 was from R&D Systems. IL-2 was obtained through the AIDS Research and MAb1) and IL-1β (CRM56 and CRM57) were from BD Biosciences. and Reference Reagent Program (Division of AIDS, National Institute of Allergy and Infectious Diseases, US National Institutes of Health), and was Quantification of mRNA expression by RT-PCR. RNA was isolated with an used at 10 U/ml at the initiation of cultures. All TLR ligands (<0.001 EU/µg; RNeasy Micro Plus Kit according to the manufacturer’s guidelines (Qiagen) EndoFit) were resuspended in endotoxin-free water according to the man- and was converted to cDNA by reverse transcription with random hexam- ufacturer’s recommendations (Invivogen). PMA (phorbol 12-myristate ers and Multiscribe RT (TQMN, Reverse Transcription Reagents; Applied 13-acetate) and Ionomycin were from Sigma-Aldrich. IRS 661, the TLR-specific Biosystems). For mRNA expression assays, the following probes were used (all inhibitory sequence46, was synthesized by Sigma-Aldrich on a phosphorothio- from Applied Biosystems): CBLB Hs00180288_m1; DGKA, Hs00176278_m1; ate backbone. EGR2, Hs00166165_m1; EGR3, Hs00231780_m1; NFATC2, Hs00234855_m1; SIRT1, Hs01009003_m1; ITCH, Hs00395201_m1; IL17A, Hs00936345_m1; Study subjects. Peripheral blood was drawn from healthy donors after IL4, Hs00929862_m1; IFNG, Hs00989291_m1; IL10, Hs00961622_m1; informed consent was provided and approval was obtained from the Foxp3, Hs01085834_m1; TBX21, Hs00203436_m1; RORC2, Hs01076112_m1; Institutional Review Board at Yale University. GATA3, Hs00231122_m1; IL2, Hs00914135_m1; TLR7, Hs00152971_m1; TLR8, Hs00152972_m1; CASP3, Hs00991554_m1; CD98, Hs00374243_m1; HIV-1+ patients. Peripheral blood was obtained from nine HIV-1-infected FASL, Hs00181225_m1; GRG4, Hs00419101_m1; IKZF1, Hs00958474_m1; patients after informed consent was provided and approval was obtained KMD6B, Hs00996325_g1; LDHA, Hs00855332_g1; RAB10, Hs00211643_m1; from the Institutional Review Board at Yale University. Patient information is RGS2, Hs01009070_g1; SOCS2, Hs00919620_m1; TNFRSF9, Hs00155512_m1; provided in Supplementary Table 1. TRAF5, Hs00182979_m1; and B2M, 4326319E). The reactions were set up following manufacturer’s guidelines and were run on a StepOne Real-Time Cell isolation and sorting of T cell populations by flow cytometry. PCR System (Applied Biosystems). Values are presented as the difference in Peripheral blood mononuclear cells were isolated from healthy donors by cyclineg threshold (Ct) values normalized to thos eof mRNA encoding β2- Ficoll Hypaque gradient centrifugation. Total CD4+ T cells were isolated by microglobulin for each sample, calculated by the following formula: relative negative selection with CD4+ T cell isolation kit (StemCell Technologies) RNA expression = (2–∆Ct) × 103. and were stained for fluorescence-activated cell sorting with the following antibodies: anti-CD45RO (UCHL1), CD45RA (Hl100), CD25 (M-A251; Immunoblot analysis. Total protein extracts were isolated with an all from BD Biosciences); and CD127 (eBioRDR5; eBioscience). M-PER protein extraction kit (Thermo Scientific) and protein was quanti- + hi lo–neg Populations of Treg cells (CD4 CD25 CD127 ), memory T cells fied with a BCA kit (Thermo Scientific). 20 µg of protein extract was loaded (CD4+CD45RA−CD45RO+CD25low/neg) and naive T cells (CD4+CD45RA+ in each lane, followed by separation by 10% SDS-PAGE and transfer to a nitro- CD45RO−CD25lo–neg) were sorted on a FACSAria (BD Biosciences). Unless cellulose membrane. Anti-NFAT1 (D43B1) and anti-β-actin (13E5) were from specified otherwise, CD4+ T cell populations used in the experiments were Cell Signaling Technology. Antibody to NFAT phosphorylated at Ser213, Ser217 + pos–lo + + depleted of Treg cells and were sorted as CD4 CD25 CD127 . CD14 cells and Ser229 (sc-32991) was from Santa Cruz Biotechnology. Primary antibodies were isolated by positive selection with an EasySep Human CD14 Positive were detected by the secondary antibody horseradish perixidase–conjugated Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature Selection Kit (StemCell Technologies). anti-rabbit (Cell Signaling Technology), and images were obtained with a charge-coupled device camera instrument. Bands were quantified with Cell activation and intracellular staining. T cell populations were stimulated QuantityOne software.

npg with 1 µg/ml anti-CD3, 1 µg/ml anti-CD28 and 10 U/ml IL-2 (antibodies identified above) in the presence or absence of the following TLR ligands (all Measurement of intracellular calcium. Isolated CD4+ T cells were labeled 2+ from Invivogen): Pam3CSK4 (0.5 µg/ml; tripalmitoyl cysteinyl seryl tetralysine ex vivo for 1 h at 37 °C with 5 µM Indo-1 AM in PBS and 1 mM Ca . Cells lipopeptide), heat-killed Listeria monocytogenes (5 × 107 cells per ml), poly were washed to remove traces of Indo-1 AM, then were resuspended in buffer (I:C) (high molecular weight; 10 µg/ml), ultrapure lipopolysaccharide from at 1 × 106 cells per ml and were incubated for 30 min at 37 °C before analysis. Escherichia coli K12 strain (10 µg/ml), recombinant flagellin from Salmonella Where necessary, IRS 661 was added at the appropriate concentration during enterica (0.2 µg/ml), synthetic diacylated lipoprotein (FSL-1; 0.2 µg/ml), the 30 min of incubation. For acquisition, samples were acquired for approxi- imiquimod (0.1–10 µg/ml), gardiquimod (5 µg/ml), loxoribine (1 mM), mately 3 min to obtain basal levels of calcium, and subsequently imiquimod CL264 (9-benzyl-8 hydroxyadenine; 20 µg/ml), ssRNA40 in LyoVec (0.5 µg/ml) or other reagents were added at the appropriate concentrations. Samples were and ODN2006 (2.5 µM). At day 4, cells were stimulated for 4 h with 50 ng/ml acquired for a minimum of 10 min on a BD Fortessa and were analyzed with phorbol-12-myristate-13-acetate (PMA) and 250 ng/ml ionomycin in the pres- FlowJo software. ence of GolgiStop (BD Biosciences) and intracellular staining of cytokines (IFN-γ, IL-10, IL-17, IL-4 and IL-2) and Foxp3 was performed with Foxp3 Gene silencing by lentiviral transduction. Lentiviral particles encod- staining buffers per the manufacturer’s recommendations (eBioscience) and ing shRNA were from Sigma-Aldrich (TRCN0000016144 (NFAT1), the following antibodies: anti-IFN-γ (4S.B3; BioLegend), anti-IL-17 (BL168; TRCN0000056973 (TLR7 clone 3) and TRCN0000056974 (TLR7 clone 4)). BioLegend), anti-IL-10 (JES3-19F1; BD Biosciences), anti-IFN-γ (B27; BD CD4+ T cells (5 × 104 cells per well) were stimulated for 12 h with plate-bound Biosciences), anti-IL-4 (3010.211; BD Biosciences) and anti-Foxp3 (PCH101; anti-CD3 (1 µg/ml) and soluble anti-CD28 (1 µg/ml) (antibodies identified eBioscience). For both extracellular staining and intracellular staining, above) before transduction. Cells were then transduced with viral particles LIVE/DEAD reagent (Molecular Probes) was used to exclude dead cells containing a vector expressing shRNA specific for NFAT1 or TLR7, or with before surface staining. CD14+ monocytes were cultured for 24 h with various nontargeting shRNA as a control, at a multiplicity of infection (MOI) of 4. All TLR ligands and cell surface staining was performed as described above constructions contained sequence encoding green fluorescent protein. After with an initial FcR-blocking step (FcR blocking reagent, human; Miltenyi 5 d in culture, transduced cells were sorted on the basis of expression of green Biotech). Samples were run on a BD Fortessa instrument and were analyzed fluorescent protein, and the efficiency of gene silencing was determined by with FlowJo software (TreeStar). TaqMan real-time PCR and protein staining.

doi:10.1038/ni.3036 nature immunology + Preparation of HIV-1 viral stocks. HIV-1NL-D (HIV-1 proviral DsRed-tagged Staining to assess apoptosis. CD4 T cells were stimulated for 48 h with anti- DNA plasmid pNL-D, derived from the prototype HIV-1 proviral DNA plas- CD3 (1 µg/ml) and anti-CD28 (1 µg/ml) (antibodies identified above) and mid pNL432) was provided by Y. Tsunetsugu-Yokota. 293T human embryonic were subsequently infected with HIV-1NL-D at an MOI of 0.001. Annexin V kidney cells were used to prepare viral stocks following published protocols35 (AnnV) and 7-amino-actinomycin D (7-AAD) were used to stain cells every with minor modifications. 293T cells were plated at a density of 0.5 × 106 24 h to assess early apoptosis (AnnV+7-AAD—) and late apoptosis (7-AAD+), cells per well in 12-well plates and were transfected for 48 h with 2 µg pNL-D for a total of 11 d. through the use of Lipofectamine 2000 (Invitrogen). Culture supernatants were treated for 30 min at 37 °C with benzonase (1 U/ml), were cleared by filtration Analysis of phosphorylation by flow cytometry. CD4+ T cells were stimulated and were ‘titrated’ on 293T cells. Stocks were stored at –80 °C. with 10 µg/ml imiquimod in the presence or absence of 50 ng/ml PMA and 250 ng/ml ionomycin, and the cells were fixed at various time points with Fixation In vitro infection with HIV-1. CD4+ T cells were stimulated for 48 h with buffer (BD Biosciences), permeabilized with Perm buffer III according to the anti-CD3 (1 µg/ml) and anti-CD28 (1 µg/ml) (antibodies identified above) manufacturer’s recommendations (BD Biosciences), and were stained with and were subsequently infected with HIV-1 at an MOI of 0.001 (a concentra- phycoerythrin-labeled mouse antibody to p38 phosphorylated at Thr180 and tion within the physiological range34). Viability and cytokine secretion was Tyr182 (36/p38; BD Biosciences), Alexa Fluor 488–labeled antibody to NF-κB measured every 48 h for a total of 11 d after infection. p65 phosphorylated at Ser529 (K10-895.12.50; BD Biosciences), Alexa Fluor For infection of TLR7-deficient cells with HIV-1, CD4+ T cells were stim- 647-labeled antibody to mouse Jnk phosphorylated at Thr183 and Tyr185 ulated anti-CD3 (1 µg/ml) and anti-CD28 (1 µg/ml) (antibodies identified (9257S; Cell Signaling), phycoerythrin-labeled rabbit antibody to IRAK4 phos- above) and were transduced with the corresponding lentiviral particles. 2 d phorylated at Thr345 and Ser346 (D6D7; Cell Signaling) and phycoerythrin- after transduction, the cells were infected with HIV-1NL-D at an MOI of 0.001. labeled antibody to c-Jun phosphorylated at Ser73 (D47G9; Cell Signaling). The viability and frequency of HIV-1-infected cells was measured at various time points by staining with LIVE/DEAD cell dye and fixation with 1% PFA. In vitro anergy induction. CD4+ T cells were incubated for 12 h with either For intracellular staining after infection, the cells were restimulated for 4 h 1 µM ionomycin or 0.5 µg/ml plate-bound anti-CD3 (identified above) (or with with PMA and ionomycin in the presence of GolgiStop and were stained as vehicle as a control), then were washed and then restimulated with anti-CD3 recommended with a Foxp3 staining kit (eBioscience) with a pre-fixation step and anti-CD28 (antibodies identified above) after a 6-hour resting period. with 1% PFA to avoid loss of DsRed signal. 2 d later, cells were infected with HIV-1, and the frequency of infected cells was assessed every other day for a total of 11 d. DNA proviral load. 50 × 103 CD4+ T cells were stimulated as described above anti-CD3 (1 µg/ml) and anti-CD28 (1 µg/ml) (antibodies identified above) Statistics. GraphPad Prism software was used for statistical analysis. A stand- and DNA was isolated after 7 d for measurement of the DNA proviral load ard paired two-tailed t-test was used for statistical analysis and a one-way as published47. analysis of variance (ANOVA) with a Tukey’s post-test was for comparisons of more than two groups; P values of 0.05 or less were considered significant. T cell clones. CD4+ T cells were sorted at a density of one cell per well in 96-well plates and were stimulated for 18 d with anti-CD3 (1 µg/ml), anti- CD28 (1 µg/ml), IL-2 (25 U/ml) (antibodies identified above) and irradiated 46. Pawar, R.D. et al. Inhibition of Toll-like receptor-7 (TLR-7) or TLR-7 plus TLR-9 peripheral blood mononuclear cell samples (5 × 104 cells) depleted of T cells. attenuates glomerulonephritis and lung injury in experimental lupus. J. Am. Soc. The medium was replaced every 3 d. After 18 d, clones were stimulated for 3 d Nephrol. 18, 1721–1731 (2007). 47. Gibellini, D., Vitone, F., Schiavone, P., Ponti, C., La Placa, M. & Re, M.C. Quantitative with anti-CD3 and anti-CD28 (antibodies identified above) in the presence or detection of human immunodeficiency virus type 1 (HIV-1) proviral DNA in peripheral absence of imiquimod, and supernatants were collected for ELISA and cells blood mononuclear cells by SYBR green real-time PCR technique. J. Clin. Virol. were lysed for isolation of RNA. 29, 282–289 (2004). Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature npg

nature immunology doi:10.1038/ni.3036