Published June 27, 2012, doi:10.4049/jimmunol.

1103033
The Journal of Immunology

CellCell Interaction with APC, not IL-23, Is Required for Naive CD4 Cells To Acquire Pathogenicity during Th17 Lineage Commitment
Guangpu Shi,* Jenna D. Lovaas,* Cuiyan Tan,* Barbara P. Vistica,* Eric F. Wawrousek, Mehak K. Aziz,* Rachael C. Rigden,* Rachel R. Caspi,* and Igal Gery*
Subpopulations of pathogenic or nonpathogenic Th17 cells were reported to develop when presensitized CD4 cells were activated with their target Ag during polarization by either IL-23 or IL-6 and TGF-b, respectively. In this study, we generated two Th17 subpopulations by using a system in which naive CD4 cells from TCR transgenic mice specic to hen egg lysozyme (HEL) are polarized with IL-6/TGF-b and, concurrently, are activated either with HEL presented by APCs, or with anti-CD3/CD28 Abs. Only the former cells were pathogenic, inducing inammation in eyes expressing HEL. Naive CD4 cells activated by the antiCD3/CD28 Abs acquired pathogenicity, however, when cocultured with HEL/APC. Importantly, the naive CD4 cells did not acquire pathogenicity when cocultured with APCs stimulated with LPS or when separated from the HEL-presenting cells by a semipermeable membrane. Unlike with presensitized Th17, soluble IL-23 does not participate in pathogenicity acquisition by naive CD4 cells; no pathogenicity was induced by adding IL-23 to cultures activated with anti-CD3/CD28 Abs. Furthermore, Abs against IL-23 or IL-23R did not inhibit acquisition of pathogenicity in cultures of naive CD4 cells activated by HEL/APC. Our data thus show that, unlike presensitized CD4 cells, naive CD4 cells polarized toward Th17 phenotype acquire pathogenicity only by direct interaction with APCs presenting the Ag, with no apparent involvement of soluble IL-23. We suggest that the Th17 lymphocytes derived from naive CD4 cells participate in pathogenic and other immune processes, along with the IL-23 dependent Th17 cells. The Journal of Immunology, 2012, 189: 000000. umerous studies in recent years have established the importance of the Th17 population in the immune system. Th17 cells are crucial in the protection against bacteria, parasites, and fungi, in particular in the gut and lungs, and they have a critical role in the majority of immune-mediated inammatory conditions in humans and experimental animals (15). The process of Th17 generation had been controversial in early studies, in particular with regard to the role of IL-23 in the process (2, 6). More recent studies have claried the issue, however, by providing evidence to show that Th17 cells are normally generated by the polarizing activity of two cytokines, IL-6 and TGF-b, whereas IL-23 is crucial for the maturity and pathogenic effectiveness of these cells (4, 7). In addition, IL-1 and IL-21 contribute to a vigorous generation of Th17 cells (8, 9). Unlike the phenotype stability of Th1, Th17 cells were found to exhibit a high level of plasticity, readily acquiring non-Th17 phenotypes when exposed to other cytokine environments (10 12). Thus, culturing Th17 cells in media containing polarizing cytokines specic for Th1 or Treg resulted in the majority of Th17
*Laboratory of Immunology, National Eye Institute, National Institutes of Health, Bethesda, MD 20892; and Laboratory of Molecular and Developmental Biology, National Eye Institute, National Institutes of Health, Bethesda, MD 20892 Received for publication October 21, 2011. Accepted for publication May 25, 2012. This work was supported by the Intramural Research Program of the National Eye Institute, National Institutes of Health. Address correspondence and reprint requests to Dr. Igal Gery, Laboratory of Immunology, National Eye Institute, National Institutes of Health, Building 10, Room 10N208, Bethesda, MD 20892-1857. E-mail address: geryi@nei.nih.gov The online version of this article contains supplemental material. Abbreviations used in this article: DC, dendritic cell; HEL, hen egg lysozyme; HEL/ APC, HEL presented by APC; PbAb, plate-bound anti-CD3/CD28 Ab; qPCR, quantitative PCR; Tg, transgenic. www.jimmunol.org/cgi/doi/10.4049/jimmunol.1103033

expressing the corresponding new phenotype within 23 d. Importantly, the acquisition of new phenotypes by Th17 cells was also observed in vivo (10, 13). Another unique feature of the Th17 population is its heterogeneity; Th17 lines generated by different procedures were found to exhibit different immunological features (14). Ghoreschi et al. (15) reported on a Th17 subpopulation generated by polarization of naive precursors by IL-6, IL-23, and IL-1b, but without TGF-b. These Th17 cells were pathogenic and capable of inducing EAE. More recently, Kim et al. (16) discovered the existence of a subpopulation of natural Th17, which are generated in the mouse thymus, before Ag exposure. Of particular interest to the current study is the early observation made by McGeachy et al. (17) of two Th17 subsets, one highly pathogenic and the other completely lacking this capacity. The two subsets originated from presensitized T cells, reactivated in culture with the Ag in the presence of either IL-23, or IL-6 and TGF-b. Only the previous subset developed pathogenicity, monitored by their capacity to induce immune-mediated inammation. We have previously reported on the activities of Th17 cells generated from naive CD4 cells with transgenic (Tg) TCR specic to hen egg lysozyme (HEL). When activated with HEL, presented by APCs, in the presence of IL-6 and TGF-b, these Th17 cells induced ocular inammation in recipient mice expressing HEL in their lens (10, 18, 19). In this study, we analyzed the phenotype of these pathogenic cells and compared it with that of another Th17 subpopulation, which is not pathogenic, that we generated by activating the naive CD4 cells with anti-CD3/CD28 Abs, along with the same polarizing cytokines, IL-6 and TGF-b. Our data indicate that the major factor determining pathogenicity of these Th17 cells, or lack thereof, is the availability of cell-cell contact with APCs presenting HEL, rather than soluble IL-23.

NAIVE CD4 NEED CELLCELL INTERACTION FOR Th17 PATHOGENICITY Measurement of cytokines released into Th17 cell culture supernatant
Cell culture supernatants were sampled on days 2, 3, or 4 of culture. The supernatants were tested for the levels of IL-10, IL-17, IL-22, and GM-CSF, using ELISA kits from R&D System, and for the levels of IL-21 and IL17F, using ELISA kits from BioLegend, according to the manufacturers instructions.

Materials and Methods

Mice
All studies were performed with (B10.BR 3 FVB/N) F1 mice, generated as detailed elsewhere (10, 18). The lines used in this study included Tg mice expressing HEL in their lens (HEL-Tg), HEL-specic TCR on the majority of their T lymphocytes (designated 3A9) (10, 18), or syngeneic wild type mice. HEL is expressed exclusively in eyes of the HEL-Tg mice; this neoself Ag cannot be detected in the circulation or any other tissue of these mice (20). All manipulations were performed in compliance with the National Institutes of Health Resolution on the Use of Animals in Research.

Flow cytometric analysis of surface, nuclear, and intracellular molecules

Conventional methods were used for analysis of surface molecule expression. For Foxp3 staining, Th17 cells were xed and permeabilized with the xation or permeabilization buffer for 1 h at 4C before intracellular staining with allophycocyanin-conjugated anti-Foxp3, following the procedure recommended by the manufacturer (eBioscience). For intracellular cytokine staining, polarized Th17 cell cultures were stimulated with 20 ng/ml PMA and 1 mM ionomycin (Sigma-Aldrich), plus Golgi-Stop (BD Biosciences), for 5 h. The stimulated cells were then stained with the corresponding Abs for surface and intracellular molecules. Data were acquired using a FACSCalibur (BD Biosciences) and analyzed using FlowJo (Tree Star).

Generation of Th17 lines

Naive T cells were enriched from splenocytes and lymph node cells of 3A9 mice, using T cell columns (R&D Systems). CD4 cells expressing the Tg TCR were then sorted by FACSAria II (BD Biosciences), using the clonotypic mAb 1G12 (10, 18). The naive CD4 cells were incubated for 4 d in 12-well plates, at 0.5 3 106 in 2 ml of Th17 polarizing medium (10, 18). The polarizing cytokines IL-6 and TGF-b were added at 100 and 10 ng/ml, respectively, similar to the study by McGeachy et al. (17). Activation of the naive CD4 cells, concurrently with the polarization, was performed by HEL (at 2 mg/ml, presented by irradiated wild type splenocytes as indicated by HEL/APC), or by spleen dendritic cells (as indicated by HEL/ DC), or by anti-CD3/CD28 Abs bound to the plate walls (PbAb) by incubation for 1 h (2 mg/well) (21). In certain experiments, additional agents and Abs were added to the culture medium, alone or in combination, as indicated. These included: IL23 (eBioscience), 20 ng/ml on day 1 or 2 and at 10 ng/ml on day 3 of culture; LPS (100 ng/ml; Difco); antiIL-23p40 Ab (50 mg/ml; a gift from Giorgio Trinchieri, National Cancer Institute, National Institutes of Health); or antiIL-23R Ab (10 mg/ml; R&D Systems). To generate Th1 lines, naive CD4 cells were incubated for 4 d in 12-well plates, at 0.5 3 106 in 2 ml of Th1 polarizing medium, as detailed elsewhere (10, 18).

Adoptive transfer of Th17 line cells

After 4 d of incubation, PbAb or HEL/APC Th17 cells were injected (at 4 3 106) via the tail vein, into groups of naive HEL-Tg mice. Recipient eyes were collected on day 5 after cell transfer and processed for histologic

DC isolation
DCs were isolated from naive syngeneic wild type mouse spleens, using the method described by Tang et al. (22). Spleens were minced into small fragments and digested with collagenase D (Roche) and DNase I (SigmaAldrich) for 45 min at 37C. Cells were collected after passing through a 70-mM nylon cell strainer (BD Falcon) and treated with EDTA for 5 min. Following red cell lysis and Fc receptor blocking, cells were incubated with antiCD11c-conjugated magnetic beads (Miltenyi Biotec) at 4C for 15 min. The positive population was puried using autoMACS (Miltenyi Biotec). The isolated preparation consisted of 90% cells positive for CD11c.

HEL/APC and PbAb Th17 cells generated in transwell cultures

For HEL/APC and PbAb Th17 generated in transwell cultures, the sorted CD4 cells (0.5 3 106) were cultured with APCs at the top compartment in a volume of 0.5 ml containing HEL and Th17 polarizing mixture (10, 18, 19), while the same number of sorted CD4 cells in a volume of 1.5 ml containing HEL and the polarizing mixture, but with no APCs, were cultured in the bottom compartments that were coated with anti-CD3/ CD28 Abs, as described above. Th1 cells were similarly cultured for comparison. The system is also demonstrated in Fig. 6A.

Quantitative PCR analysis

Total RNA was extracted from cultured cells with TRIzol (Invitrogen-Life Technologies). RNA (5 mg), SuperScript III Reverse Transcriptase (Invitrogen Life Technologies), and oligo(dT)1216 were used for rst-strand cDNA synthesis. Primer-probe sets for quantitative PCR (qPCR) quantifying the expression of mouse IL-23R, CD40L, ICOS, AHR, CCL2, CCL5, CCL20, CCL22, CXCL2, CXCL10, CCR2, CCR6, CXCR3 and GAPDH or b-actin (internal control) were purchased from Applied Biosystems. Primers used for ROR-gt and ROR-a were as follows: ROR-gt, 59-CCGCTGAGAGGGCTTCAC-39 and 59-TGCAGGAGTAGGCCACATTA CA-39; and ROR-a, 59-CGTGTCCATGGCAGAACTAGAA-39 and 59GCAAGTACTGGCAGGTTTCCA-39. Fluorescence-labeled probes used are: ROR-gt, 59-AAGGGCTTCTTCCGCCGCAGCCAGCAG-BHQ-139; and ROR-a, 59-CCTTGCCCAGAACATATCCAAATCCCA-BHQ-139. PCR parameters were as recommended for the TaqMan Universal PCR Master Mix kit (Applied Biosystems).

FIGURE 1. Unlike HEL/APC Th17, PbAb Th17 cells are nonpathogenic. (A) Four million cells of the two subpopulations were adoptively transferred into HEL-Tg mice; 5 d later, eyes of recipients were examined histologically for inammatory changes. Sections of representative eyes show no pathologic changes in the recipient of PbAb Th17 cells, but severe panuveitic changes in the recipient of HEL/APC Th17 cells. The changes mainly include edema and cellular inltration in the cornea, cellular, and proteinaceous exudates in the anterior chamber and the vitreous, and intense inltration in the retina. (B) Transferred PbAb Th17 cells are alive and proliferate in the recipient spleen when exposed to HEL. CSFE division analysis of transferred cells in recipients injected with HEL (PbAb Th17 + HEL) and their controls in recipients with no HEL injection (PbAb Th17). The data are shown by ow cytometry results of a representative experiment (upper panel) and as mean percentage 6 SEM of 1G12+ cells among total CD4 cells in recipient spleens, on day 3 after cell injection, in three independent experiments (lower panel). (C) Low proportions of Foxp3+ cells among both PbAb and HEL/APC Th17 subpopulations, measured by ow cytometry. A representative experiment; similar results were obtained in two additional experiments.

The Journal of Immunology

analysis by conventional H&E staining. Severity of the inammation was scored according to histologic changes, as described elsewhere (23).

3 into syngeneic recipients expressing HEL in their eyes (10, 18, 19). Fig. 1A depicts representative sections of eyes from recipients injected with cells of the two subpopulations. Severe ocular inammatory changes are seen in the recipient of HEL/APC Th17 cells, whereas no inammatory changes could be detected in the eyes of the recipient of PbAb Th17 cells. The lack of pathogenicity by PbAb-activated Th17 cells could be attributed to their low viability in vivo. To examine this possibility, we used the CFSE division method to test the capacity of the transferred cells to proliferate when exposed to their target Ag. We and others (25, 26) have reported that adoptively transferred autopathogenic T cells initially migrate to the host spleen, where they proliferate before invading the target tissue. Fig. 1B shows a representative experiment. Adoptively transferred CFSE-labeled PbAb Th17 cells proliferated poorly in the spleen of recipient mice, but vigorous proliferation was seen in recipient mice that were also injected with HEL. This observation thus provides evidence that the injected PbAb-activated cells are viable and fully capable of responding to their target Ag. Importantly, the PbAbactivated Th17 cells in this system retained their nonpathogenicity despite their reactivation by the Ag in vivo (data not shown). The possibility that the nonpathogenic activity of PbAb-stimulated Th17 cells was due to high levels of regulatory T cells was ruled out by the nding that the proportion of cells expressing Foxp3 in PbAb-activated cells was similar to that of the pathogenic HEL/APC Th17 cells (Fig. 1C).

Measuring cell division rate by CFSE dilution

PbAb or HEL/APC Th17 cells (107 cells/ml) were labeled with CFSE (10 mM; Molecular Probes) for adoptive transfer, as described elsewhere (19, 24). CFSE-labeled cells were washed, resuspended in RPMI 1640, and injected i.v. into recipient mice (4 3 106 cells). Recipient mice were euthanized on the indicated days after cell transfer, and splenocytes were collected and stained with antiCD4-PE. Data were acquired using a FACSCalibur. CFSE dye dilution was analyzed using FlowJo.

Statistical analysis
Unpaired, two-tail t test was performed for comparison of transcript level in Th17 or cytokine level in supernatant, as well as for comparison of the frequency of transferred cells recovered from recipient spleen. A p value # 0.05 was dened as statistically signicant.

Results
Th17 cells generated by activation with HEL/APC are pathogenic, whereas those generated by activation with anti-CD3/CD28 Abs are not We generated two subpopulations of Th17 cells by activation of sorted naive CD4 cells specic to HEL with either the target Ag, HEL, presented by APCs (HEL/APC), or with plate-bound antiCD3/CD28 Abs (PbAb), in the presence of the same polarizing cytokines, IL-6 and TGF-b. To test for the pathogenic capacity of the two Th17 subpopulations, we adoptively transferred the cells

FIGURE 2. HEL/APC and PbAb Th17 subpopulations differ in their cytokine production proles. (A) Expression of ROR-a and ROR-gt transcripts by the two Th17 subpopulations, after 4 d of incubation, determined by qPCR. (B) Mean levels 6 SEM of IL-17, IL-22, and IL-10 in supernatants collected at different time points of cultures in which the two subpopulations were generated. The data were collected in four independent experiments. (C) Immunostaining for intracellular expression of IL-17, IL-22, and IL-10 of cells from the two Th17 subpopulations. The cells were collected after 4 d in culture and stained with the specic mAbs, or isotype controls (IgG), as detailed in Shi et al. (10). Note that HEL/APC Th17, but not PbAb Th17, cells stained for intracellular IL-22 and that different subsets within this subpopulation stained for either IL-17 or IL-22, with just a minor portion of cells coexpressing both cytokines. Also of note is the coexpression of IL-17 and IL-10 by cells of the two Th17 subpopulations. A representative experiment; similar observations were made in another experiment. p , 0.05.

4 Molecular analysis of the two Th17 subpopulations

NAIVE CD4 NEED CELLCELL INTERACTION FOR Th17 PATHOGENICITY populations were found in their expression proles of several chemokines and chemokine receptors (Fig. 3B). Whereas transcripts of CCL20 and CXCR3 were more elevated in PbAb cultures, those of CXCL10, CCL2, CCL22, and CXCL2 were higher in HEL/DC cultures. Th17 cells generated by stimulation with HEL/APC or with PbAb were also tested for production of other molecules: cells of the two subpopulations expressed similar levels of CD40L transcripts, but HEL/APC Th17 expressed higher levels of ICOS transcript, whereas moderately higher levels of AHR transcript were expressed by PbAb Th17 cells (Supplemental Fig. 2A). Supernatants of these cultures contained similar levels of GM-CSF and IL-21, but PbAb-stimulated cultures were superior in their production of IL-17F (Supplemental Fig. 2B), in line with their higher levels of IL-17A (Fig. 2B). Th17 generated by stimulation with either PbAb or HEL/APC proliferated vigorously, but a marked difference was noted in the kinetics of their response, with PbAb-stimulated Th17 cells responding earlier (Supplemental Fig. 3). This difference is in accord with the two mechanisms of stimulation process: the Abs directly stimulate the naive CD4 cells, whereas stimulation with HEL/APC requires prior processing of the Ag by the APC. The two Th17 subpopulations differ in their homing to and proliferation in the recipient mouse spleen As mentioned above, adoptively transferred activated T cells migrate to the recipient spleen, where they proliferate before invading the target organ (25, 26). To compare between the two Th17 subpopulations for this capacity, we adoptively transferred the tested cells into HEL-Tg mice and tracked them in the recipient spleen by the clonotypic Ab, 1G12. Fig. 4A depicts data of a representative experiment in which 1G12+ cells were identied by ow cytometry in recipient spleens 4 d after cell transfer; the gure shows that Th17 generated by activation with HEL/APC were superior to Th17 generated by PbAb activation in their capacity to accumulate and

To dene the two Th17 subpopulations molecularly, we compared their capacity to produce several molecules characteristic for the Th17 lineage. Both subpopulations expressed transcripts of the two Th17 specic transcription factors, ROR-a and ROR-gt, with higher levels seen in the PbAb cultures (Fig. 2A). Next, we analyzed the cytokine production by the two subpopulations, by measuring in the culture supernatants the levels of three Th17 productsIL-17, IL-10, and IL-22at different time points (Fig. 2B). PbAb cultures yielding nonpathogenic Th17 produced exceedingly high levels of IL-17 that were notably higher than those produced by HEL/APC cultures, which generate the pathogenic Th17. The two types of cultures produced similar or only moderately different levels of IL-10, but striking differences were seen in the pattern of IL-22 production between the cultures activated by PbAb or HEL/APC, with the latter cultures producing IL-22 levels higher by two orders of magnitude. The vigorous production of IL-22 by CD4 cells stimulated by HEL/ APC was not affected by the addition of Ab against IL-23R, or against 12p40, but was moderately inhibited by the addition of PbAb (Supplemental Fig. 1). Production of the three cytokines by Th17 cells was also directly demonstrated by ow cytometry. The data of a representative experiment are recorded in Fig. 2C and show that the intracellular staining patterns of the tested cytokines are in accord with our ELISA data. Of interest are the ndings that (1) IL-10 was expressed by the same subset of CD4 cells that produced IL-17, which is in line with our nding (Fig. 2B) that pathogenic Th17 cells in our study also secrete IL-10, and (2) in contrast to IL-10, IL22 production in cultures activated by HEL/APC was localized to a large extent in a subset of CD4 cells that did not coexpress IL-17. The subpopulations of pathogenic and nonpathogenic Th17 also differed in their staining for certain surface Ags. Differences were particularly clear in the expression of CCR5, CCR6, and a4b7 (Fig. 3A). More dramatic differences between the two sub-

FIGURE 3. The two Th17 subpopulations differ in their expression of surface molecules and certain chemokine and chemokine receptor transcripts. (A) Suspensions of PbAb and HEL/APC Th17 cells, harvested after 4 d in culture, were stained for the indicated surface molecules and examined by ow cytometry. (B) qPCR analysis of Th17 cells collected after 4 d of incubation at the HEL/DC or PbAb conditions. Transcript expression is presented relative to GAPDH. A representative experiment is shown; similar expression proles were observed in four other independent experiments.

The Journal of Immunology

5 PbAb cells, a nding in line with their higher proportion among the total spleen CD4 cells (Fig. 4A, 4B). We have reported previously (10) that Th17 cells generated by activation with HEL/APC exhibit plasticity and acquire the Th1 phenotype when exposed to other cytokine environments. To compare the two Th17 subpopulations for their plasticity, we collected recipient spleen cells on day 4 after cell injection and analyzed them for intracellular expression of IL-17 and IFN-g. As shown in Fig. 4D, the proportion of cells expressing IFN-g was higher among the HEL/APC recipients than among the PbAb recipients. Naive CD4 cells activated by plate-bound Abs acquire pathogenicity when cocultured with DC and HEL Soluble IL-23 was shown to be critical for acquisition of pathogenicity by CD4 cells preimmunized against the immunopathogenic Ag (7, 17). No pathogenicity was acquired, however, in our system by naive CD4 activated in culture by the plate-bound Abs when IL-23 was added to the medium during Th17 polarization, on day 1 (Fig. 5A, upper panel) or even on days 2 and 3 (not shown), when the expression of IL-23R was elevated (Fig. 5B). In contrast, PbAb-activated CD4 cells acquired pathogenic capacity when DCs and HEL were also added to the culture medium; these Th17 cells induced inammation levels similar to those achieved by Th17 cells of cultures stimulated by DCs and HEL (Fig. 5A, upper panel). Importantly, however, no pathogenicity was acquired in cultures in which DCs were added alone or with LPS, a potent APC stimulant (2729) (Fig. 5A, upper panel). All cultures in these experiments were polarized with IL-6 and TGF-b during the activation process and, as shown in the lower panel of Fig. 5A, Th17 cells generated in all these cultures released high levels of IL-17, despite the differences in their immunopathogenicity, but in agreement with data shown in Fig. 2B. It is of interest, however, that the lowest IL-17 levels were released by the pathogenic cultures activated by only DCs and HEL, again in agreement with the data recorded in Fig. 2B, underscoring the IL17stimulating capacity of the plate-bound Abs. It is also noteworthy that the addition of Abs against IL-23 or IL23R had no effect on acquisition of pathogenicity by CD4 cells activated by HEL/DC, or by plate-bound Abs and HEL/DC (data not shown). Cell-cell interaction between naive CD4 and APCs is essential for acquisition of pathogenicity by Th17 cells APCs affect T cells by the release of cytokines and by cell-cell interaction (30, 31). To analyze the mechanism whereby APCs induce pathogenicity in Th17 in our experimental system, we separated by semipermeable membranes the HEL-presenting cells and the naive CD4 cells activated with anti-CD3/CD28 Abs (Fig. 6A) and examined the development of pathogenicity by the latter cells. Naive CD4 cells were also added to the top compartment, along with the APCs. Th17 polarizing cytokines (IL-6 and TGF-b) and HEL were added to both compartments. Data collected in repeated experiments are summarized in Fig. 6B, whereas Fig. 6D shows representative eye sections of recipient mice injected with Th17 cells collected from the two compartments. As expected, the CD4 lymphocytes from the top compartment, cultured with HEL/APC, developed pathogenic capacity. In contrast, no pathogenicity was demonstrated by the CD4 cells from the bottom compartment, which were activated by anti-CD3/CD28 Abs. These observations thus provide evidence showing that acquisition of pathogenicity by naive CD4 cells during polarization toward Th17 phenotype requires cell-cell interaction with APCs, rather than soluble cytokines released by APCs.

FIGURE 4. HEL/APC Th17 are superior to PbAb Th17 cells in their capacity to migrate to and proliferate in the recipient spleen. Four million Th17 cells of the two subpopulations were adoptively transferred into HEL-Tg mice, and spleens of the recipients were collected at the indicated time points. (A) Flow cytometric analysis on day 4, identifying donor cells, positive for the clonotypic Ab, 1G12, among the CD4 population. The proportion of donor cells is higher in the recipients injected with HEL/ APC Th17 than in that of the PbAb Th17 recipients. A representative experiment is shown; similar data were obtained in two other experiments. (B) A summary of three experiments, depicting the mean percentage 6 SEM of PbAb or HEL/APC Th17 cells (1G12+) among CD4 cells in the recipient spleens, on days 2, 4, or 7 after cell injection. (C) Proliferation rate of the two subpopulations determined by the CFSE dilution method, at the indicated time points. Th17 cells of the HEL/APC subpopulation proliferated at a faster rate. (D) Th17 generated by HEL/APC activation switch to the Th1 phenotype in the recipient spleen more readily than PbAb-generated Th17. Intracellular ow cytometric analysis shows a higher proportion of the former subpopulation expressing IFN-g on day 4 after cell injection. p , 0.05.

proliferate in the recipient spleen. Repeated experiments in which recipient spleens were collected on days 2, 4, and 7 after cell transfer are summarized in Fig. 4B, showing the superiority of the HEL/APC Th17 at all tested time points. Next, we used the CFSE dilution procedure to compare the two subpopulations of Th17 cells for their proliferation rate in the recipient spleen. As seen in Fig. 4C, the HEL/APC Th17 proliferated considerably more vigorously than the nonpathogenic

NAIVE CD4 NEED CELLCELL INTERACTION FOR Th17 PATHOGENICITY

FIGURE 5. Naive CD4 cells activated by PbAb acquire pathogenicity when DCs and HEL are also added to the culture medium. (A) Top frame, severity levels of ocular inammation induced in recipient mice expressing HEL in their eyes, following adoptive transfer of 4 million Th17 cells, generated as indicated. The dots represent individual eyes, scored as detailed in Kim et al. (23). Bottom panel, Levels of IL-17 secreted by the different Th17 cultures, generated as indicated, following incubation for 4 d. (B) Expression levels of IL-23R transcripts by activated naive CD4 cells, at different time points. PbAb-stimulated naive CD4 cells, under Th17 polarizing conditions, were collected at the indicated time points and assayed by qPCR. Freshly isolated naive CD4 cells were used as control for the assay.

FIGURE 6. Cell-cell interaction is required for acquisition of pathogenicity by Th17 cells, but not by Th1 cells. Transwell cultures (A) were established with HEL-specic naive CD4 cells activated with either APCs in the top compartment or anti-CD3/CD28 Abs in the bottom compartment, along with HEL and polarizing cytokines for either Th17 or Th1 in both compartments. Following 4 d of incubation, the cells collected from each compartment were injected into HEL-Tg recipients. (B and C) Summaries of three repeated experiments, recording the severity levels of individual eyes of recipient mice injected with Th17 or Th1 cells, collected from the top or bottom compartments and scored as detailed in Kim et al. (23). (D and E) Sections of the anterior and posterior eye segments of representative recipient eyes collected 5 d after cell transfer. Severe changes were induced by Th1 collected from both culture compartments (E), but only from the top compartment of Th17 cultures (D). C, Cornea; I, iris; L, lens; O, optic nerve; R, retina; V, vitreous.

To further examine the observations we made with Th17 subpopulations using the transwell system, we performed similar experiments in which Th1 cell subpopulations were generated and tested for pathogenicity (Fig. 6C, 6E). Th1 cells were generated by polarization with IL-12 (10, 18). Unlike with Th17 cells, Th1 cells generated by activation with PbAb during polarization acquired pathogenicity. As shown in Fig. 6C and 6E, Th1 collected from the bottom compartment (activated by PbAb) induced inammation in recipient eyes, similar to that induce by Th1 cells collected from the top compartment (activated by HEL/APC). These observations thus underscore the profound difference between Th17 and Th1 populations in their acquisition of pathogenicity. Activation of naive CD4 cells with PbAb during polarization elicits pathogenic capacity in Th1 but not in Th17 cells.

Discussion
As cited above, recent publications revealed the heterogenicity of the Th17 population, with different subpopulations being generated

by different procedures (14, 15, 17). Data collected in the current study further extend the information concerning the variety of Th17 subpopulations. These data show that activating naive CD4 cells with either the Ag presented by APCs, or with Abs against CD3/CD28, during polarization with IL-6 and TGF-b, generate two Th17 subpopulations that differ by their immunopathogenic capacity: the former subpopulation is pathogenic, whereas the latter is not. It is of note that the pathogenic and nonpathogenic subpopulations in our study differed from the analogous two Th17 subpopulations described by McGeachy et al. (17) in their cell source (i.e., naive versus preimmunized cells) as well as the activation and polarization procedures used for their generation. In addition, the pathogenic Th17 lines of the two studies also differed in their expression of IL-10; pathogenic cells in the study by McGeachy et al. (17) did not produce this cytokine, whereas the pathogenic cells in our study did, as shown by both their release and intracellular production of this cytokine (Fig. 2B, 2C).

The Journal of Immunology The two subpopulations in our study also differed in their production of IL-22, a characteristic product of Th17 cells (32): only cells of the pathogenic Th17 subpopulation produced this cytokine. Interestingly, ow cytometric analysis revealed that the majority of cells that express IL-22 do not coproduce IL-17 (Fig. 2C). More studies are needed to further analyze the biological function of the subset of IL-22producing Th17 cells and their possible relationship to the Th22 subpopulation recently identied in humans (33, 34). The capacity to produce IL-22 is not related to pathogenicity of Th17 cells, because treatment of recipients of pathogenic Th17 with antiIL-22 Ab had no effect on the development of ocular inammation in these mice (data not shown). The two subpopulations in our study differed in their proles of cell surface Ags, chemokines, and chemokine receptors, and it is conceivable that these differences determined their pathogenic capacity to a large extent. This notion is also supported by the nding that the proles of chemokines and chemokine receptors of the pathogenic and nonpathogenic Th17 subpopulations generated in our study (Fig. 3B) resembled those of the analogous subpopulations generated by McGeachy et al. (17), despite the remarkable differences in their cell origin, mode of generation, and lymphokine proles. Our study shows that pathogenic Th17 cells are superior to the nonpathogenic cells in their capacity to migrate into and proliferate in the recipient spleen. The migration was monitored by tracking the injected cells with the clonotypic Ab 1G12, and their proliferation rate was measured by the CFSE dilution method (Fig. 4A C). It is conceivable that these cellular capacities are essential for the pathogenic Th17 cells in their induction of inammation in the target organ, the eye in the current study. Our study thus provides a new parameter for the capacity of Th populations to migrate into and proliferate in the recipient organs. In accord with our previous study (10), a portion of Th17 cells generated by HEL/APC activation acquire the Th1 phenotype in the recipient spleen (Fig. 4D). Lower proportions of Th17 cells generated by PbAb activation also exhibited phenotype switching, but it seems unlikely this lower activity of the PbAb Th17 cells contributed signicantly to their lack of pathogenicity. Importantly, data collected in the current study show that, unlike presensitized CD4 lymphocytes (17), naive CD4 cells are not affected by soluble IL-23 during their polarization toward the pathogenic Th17 phenotype. Addition of soluble IL-23 to cultures stimulated with anti-CD3/CD28 Abs had no effect on the generation of nonpathogenic phenotype, whereas addition of APCs and HEL to these cultures yielded pathogenic Th17 (Fig. 5A). Furthermore, the generation of pathogenic Th17 from naive cells could not be blocked in culture by Abs against IL-23 or IL-23R (data not shown). It is noteworthy that the APC effect of generating pathogenicity in naive CD4 cells in our experimental system was achieved only when the APC presented the target AgHEL. No effect was seen when the APCs were added with no addition of HEL, or even when LPS, a potent APC stimulator (2729), was added to the culture system (Fig. 5A). These data suggest that naive CD4 cells acquire pathogenicity only by cell-cell interaction with HELpresenting APCs, a notion that was further examined by the transwell method. This method allowed us to test in the same culture system activation of naive CD4 cells by either cell-cell interaction with APC-presenting HEL, or by cytokines released from these APCs. Pathogenicity was acquired only by the former mode of activation (Fig. 6B). In conclusion, the data in this study provide evidence to show that naive CD4 cells polarized into the Th17 phenotype acquire pathogenic capacity only by directly interacting with APCs pre-

7 senting the specic Ag. Our data also show that IL-23, which was found to be crucial for development of pathogenicity by preimmunized CD4 cells (4, 7, 12, 17), is not essential for the process in naive CD4 cells (Fig. 5A). These data thus show for the rst time, to our knowledge, that the process that generates Th17 from naive CD4 is remarkably different from the process that generates Th17 from mature preimmunized Th cells. More investigation is needed to analyze the functions of Th17 originated from naive cells in the total Th17-induced immune response against microbial invasion, or the initiation of pathogenic inammation. We suggest that the Th17 lymphocytes derived from naive CD4 cells participate in these immune processes along with, or ahead of, the IL23dependent Th17 cells.

Acknowledgments
We thank Dr. Ronald H. Schwartz for helpful suggestions, Lindsey Nugent for expert assistance, R. Steven Lee for tail DNA analysis, the National Eye Institute Flow Cytometry Core for technical support, and the Histology Core for tissue section preparations.

Disclosures
The authors have no nancial conicts of interest.

References
1. Bettelli, E., M. Oukka, and V. K. Kuchroo. 2007. T(H)-17 cells in the circle of immunity and autoimmunity. Nat. Immunol. 8: 345350. 2. Weaver, C. T., R. D. Hatton, P. R. Mangan, and L. E. Harrington. 2007. IL-17 family cytokines and the expanding diversity of effector T cell lineages. Annu. Rev. Immunol. 25: 821852. 3. Park, H., Z. Li, X. O. Yang, S. H. Chang, R. Nurieva, Y. H. Wang, Y. Wang, L. Hood, Z. Zhu, Q. Tian, and C. Dong. 2005. A distinct lineage of CD4 T cells regulates tissue inammation by producing interleukin 17. Nat. Immunol. 6: 11331141. 4. Morrison, P. J., S. J. Ballantyne, and M. C. Kullberg. 2011. Interleukin-23 and T helper 17-type responses in intestinal inammation: from cytokines to T-cell plasticity. Immunology 133: 397408. 5. Shainheit, M. G., K. W. Lasocki, E. Finger, B. M. Larkin, P. M. Smith, A. H. Sharpe, C. A. Dinarello, L. I. Rutitzky, and M. J. Stadecker. 2011. The pathogenic Th17 cell response to major schistosome egg antigen is sequentially dependent on IL-23 and IL-1b. J. Immunol. 187: 53285335. 6. McKenzie, B. S., R. A. Kastelein, and D. J. Cua. 2006. Understanding the IL-23IL-17 immune pathway. Trends Immunol. 27: 1723. 7. McGeachy, M. J., Y. Chen, C. M. Tato, A. Laurence, B. Joyce-Shaikh, W. M. Blumenschein, T. K. McClanahan, J. J. OShea, and D. J. Cua. 2009. The interleukin 23 receptor is essential for the terminal differentiation of interleukin 17-producing effector T helper cells in vivo. Nat. Immunol. 10: 314324. 8. Geri, G., B. Terrier, M. Rosenzwajg, B. Wechsler, M. Touzot, D. Seilhean, T. A. Tran, B. Bodaghi, L. Musset, V. Soumelis, et al. 2011. Critical role of IL-21 in modulating TH17 and regulatory T cells in Behcet disease. J. Allergy Clin. Immunol. 128: 655664. 9. Sutton, C., C. Brereton, B. Keogh, K. H. Mills, and E. C. Lavelle. 2006. A crucial role for interleukin (IL)-1 in the induction of IL-17-producing T cells that mediate autoimmune encephalomyelitis. J. Exp. Med. 203: 16851691. 10. Shi, G., C. A. Cox, B. P. Vistica, C. Tan, E. F. Wawrousek, and I. Gery. 2008. Phenotype switching by inammation-inducing polarized Th17 cells, but not by Th1 cells. J. Immunol. 181: 72057213. 11. Lee, Y. K., H. Turner, C. L. Maynard, J. R. Oliver, D. Chen, C. O. Elson, and C. T. Weaver. 2009. Late developmental plasticity in the T helper 17 lineage. Immunity 30: 92107. 12. Hirota, K., J. H. Duarte, M. Veldhoen, E. Hornsby, Y. Li, D. J. Cua, H. Ahlfors, C. Wilhelm, M. Tolaini, U. Menzel, et al. 2011. Fate mapping of IL-17producing T cells in inammatory responses. Nat. Immunol. 12: 255263. 13. Martin-Orozco, N., Y. Chung, S. H. Chang, Y. H. Wang, and C. Dong. 2009. Th17 cells promote pancreatic inammation but only induce diabetes efciently in lymphopenic hosts after conversion into Th1 cells. Eur. J. Immunol. 39: 216224. 14. Peters, A., Y. Lee, and V. K. Kuchroo. 2011. The many faces of Th17 cells. Curr. Opin. Immunol. 23: 702706. 15. Ghoreschi, K., A. Laurence, X. P. Yang, C. M. Tato, M. J. McGeachy, J. E. Konkel, H. L. Ramos, L. Wei, T. S. Davidson, N. Bouladoux, et al. 2010. Generation of pathogenic T(H)17 cells in the absence of TGF-b signalling. Nature 467: 967971. 16. Kim, J. S., J. E. Smith-Garvin, G. A. Koretzky, and M. S. Jordan. 2011. The requirements for natural Th17 cell development are distinct from those of conventional Th17 cells. J. Exp. Med. 208: 22012207. 17. McGeachy, M. J., K. S. Bak-Jensen, Y. Chen, C. M. Tato, W. Blumenschein, T. McClanahan, and D. J. Cua. 2007. TGF-b and IL-6 drive the production of IL17 and IL-10 by T cells and restrain T(H)-17 cell-mediated pathology. Nat. Immunol. 8: 13901397.

NAIVE CD4 NEED CELLCELL INTERACTION FOR Th17 PATHOGENICITY

up- and down-regulation of chemokine receptor CXCR3 on inammationinducing Th1 cells. Eur. J. Immunol. 34: 28852894. Flugel, A., T. Berkowicz, T. Ritter, M. Labeur, D. E. Jenne, Z. Li, J. W. Ellwart, M. Willem, H. Lassmann, and H. Wekerle. 2001. Migratory activity and functional changes of green uorescent effector cells before and during experimental autoimmune encephalomyelitis. Immunity 14: 547560. Gery, I., and B. H. Waksman. 1972. Potentiation of the T-lymphocyte response to mitogens. II. The cellular source of potentiating mediator(s). J. Exp. Med. 136: 143155. Piani, A., J. P. Hossle, T. Birchler, C. A. Siegrist, D. Heumann, G. Davies, S. Loeliger, R. Seger, and R. P. Lauener. 2000. Expression of MHC class II molecules contributes to lipopolysaccharide responsiveness. Eur. J. Immunol. 30: 31403146. Ulevitch, R. J., and P. S. Tobias. 1995. Receptor-dependent mechanisms of cell stimulation by bacterial endotoxin. Annu. Rev. Immunol. 13: 437457. Gutcher, I., and B. Becher. 2007. APC-derived cytokines and T cell polarization in autoimmune inammation. J. Clin. Invest. 117: 11191127. Yokosuka, T., and T. Saito. 2010. The immunological synapse, TCR microclusters, and T cell activation. Curr. Top. Microbiol. Immunol. 340: 81107. Liang, S. C., X. Y. Tan, D. P. Luxenberg, R. Karim, K. Dunussi-Joannopoulos, M. Collins, and L. A. Fouser. 2006. Interleukin (IL)-22 and IL-17 are coexpressed by Th17 cells and cooperatively enhance expression of antimicrobial peptides. J. Exp. Med. 203: 22712279. Eyerich, S., K. Eyerich, D. Pennino, T. Carbone, F. Nasorri, S. Pallotta, F. Cianfarani, T. Odorisio, C. Traidl-Hoffmann, H. Behrendt, et al. 2009. Th22 cells represent a distinct human T cell subset involved in epidermal immunity and remodeling. J. Clin. Invest. 119: 35733585. Wan, Q., L. Kozhaya, A. ElHed, R. Ramesh, T. J. Carlson, I. M. Djuretic, M. S. Sundrud, and D. Unutmaz. 2011. Cytokine signals through PI-3 kinase pathway modulate Th17 cytokine production by CCR6+ human memory T cells. J. Exp. Med. 208: 18751887.

18. Cox, C. A., G. Shi, H. Yin, B. P. Vistica, E. F. Wawrousek, C. C. Chan, and I. Gery. 2008. Both Th1 and Th17 are immunopathogenic but differ in other key biological activities. J. Immunol. 180: 74147422. 19. Shi, G., M. Ramaswamy, B. P. Vistica, C. A. Cox, C. Tan, E. F. Wawrousek, R. M. Siegel, and I. Gery. 2009. Unlike Th1, Th17 cells mediate sustained autoimmune inammation and are highly resistant to restimulation-induced cell death. J. Immunol. 183: 75477556. 20. Zhang, M., M. S. Vacchio, B. P. Vistica, S. Lesage, C. E. Egwuagu, C. R. Yu, M. P. Gelderman, M. C. Kennedy, E. F. Wawrousek, and I. Gery. 2003. T cell tolerance to a neo-self antigen expressed by thymic epithelial cells: the soluble form is more effective than the membrane-bound form. J. Immunol. 170: 3954 3962. 21. Huter, E. N., G. H. Stummvoll, R. J. DiPaolo, D. D. Glass, and E. M. Shevach. 2009. Pre-differentiated Th1 and Th17 effector T cells in autoimmune gastritis: Ag-specic regulatory T cells are more potent suppressors than polyclonal regulatory T cells. Int. Immunopharmacol. 9: 540545. 22. Tang, J., W. Zhu, P. B. Silver, S. B. Su, C. C. Chan, and R. R. Caspi. 2007. Autoimmune uveitis elicited with antigen-pulsed dendritic cells has a distinct clinical signature and is driven by unique effector mechanisms: initial encounter with autoantigen denes disease phenotype. J. Immunol. 178: 55785587. 23. Kim, S. J., M. Zhang, B. P. Vistica, C. C. Chan, D. F. Shen, E. F. Wawrousek, and I. Gery. 2002. Induction of ocular inammation by T-helper lymphocytes type 2. Invest. Ophthalmol. Vis. Sci. 43: 758765. 24. Fujimoto, C., C. R. Yu, G. Shi, B. P. Vistica, E. F. Wawrousek, D. M. Klinman, C. C. Chan, C. E. Egwuagu, and I. Gery. 2006. Pertussis toxin is superior to TLR ligands in enhancing pathogenic autoimmunity, targeted at a neo-self antigen, by triggering robust expansion of Th1 cells and their cytokine production. J. Immunol. 177: 68966903. 25. Chen, J., B. P. Vistica, H. Takase, D. I. Ham, R. N. Fariss, E. F. Wawrousek, C. C. Chan, J. A. DeMartino, J. M. Farber, and I. Gery. 2004. A unique pattern of

26.

27.

28.

29. 30. 31. 32.

33.

34.