Female C57BL/6 (B6) mice and CD73^−/− and TCR-δ^−/− mice were purchased from Jackson Laboratory (Bar Harbor, ME); 12- to 16-week-old mice were used in all studies. All mice were housed and maintained in the animal facilities of the University of California Los Angeles. Institutional approval (Protocol number: ARC#2014-029-03A) was obtained from the Institutional Animal Care and Use Committee of the Doheny Eye Institute, University of California Los Angeles, and institutional guidelines regarding animal experimentation were followed. Veterinary care was provided by IACUC faculty. Immunized animals that displayed swelling joints were either humanely euthanatized or administered an analgesic (buprenorphine, 0.1 mg/kg sc. twice daily or ketoprofen, 2 mg/kg sc. daily) until the swelling resolved. By the end of the study, mice were euthanized by cervical dislocation after a lethal injection of ketamine and xylazine prior to tissue collection. Recombinant murine IL-1β, IL-7, IL-12 and IL-23 were purchased from R & D (Minneapolis, MN). Fluorescein isothiocyanate (FITC)-, phycoerythrin (PE)-, or allophycocyanin (APC)-conjugated antibodies (Abs) against mouse CD4 (GK1.5), αβ T cell receptor (TCR) (H57-597), or γδ TCR (GL3) and their isotype control antibodies were purchased from Biolegend (San Diego, CA). (PE)-conjugated anti-mouse IFN-γ (XMG1.2) and IL-17 (TC11-18H10.1) monoclonal antibody was purchased from Santa Cruz Biotechnology (Dallas, Texas). The non-selective AR agonist 5’-N-ethylcarboxamidoadenosine (NECA) (34) and AMP were purchased from Sigma-Aldrich (St. Louis, MO, USA). Toll-like receptors ligand LPS was purchased from Invivogen (San Diego, CA).

EAU was induced in B6 mice by subcutaneous injection of 200 μl of emulsion containing 200 μg of human IRBP_1-20 (Sigma-Aldrich, St. Louis, MO) in complete Freund’s adjuvant (CFA; Difco, Detroit) at six spots at the tail base and on the flank and intraperitoneal (i.p.) injection with 300 ng of pertussis toxin.

All αβ T cells used were purified from the spleen or draining lymph nodes of IRBP_1–20-immunized mice at day 13 post-immunization using an auto-MACS separator system, as described previously [29]. The purity of the purified cells was >95%, as determined by flow cytometric analysis using phycoerythrin-conjugated antibodies against αβ T cells. The cells were then cultured in RPMI 1640 medium containing 10% fetal calf serum (Corning).

Non-activated and activated γδ T cells were separated from either naïve B6 mice or IRBP_1–20-immunized B6 mice, respectively, by positive selection using a combination of FITC-conjugated anti-TCR-δ antibody and anti-FITC antibody-coated Microbeads, followed by separation using an auto-MACS.

Bone marrow dendritic cells were generated by incubating bone marrow cells for 5 days in the presence of 10 ng/ml of recombinant murine GM-CSF and IL-4 (R&D Systems), as described previously (35). Cytokine (IL-1β, IL-6, L-12 and IL-23) levels in the culture medium were measured by ELISA.

To determine the antigen-presenting function, BMDCs were incubated in a 24-well plate with responder T cells isolated from immunized B6 mice under Th1- or Th17-polarizing conditions. Forty-eight hours after stimulation, IFN-γ and IL-17 in the culture medium were measured by ELISA. The percentage of IFN-γ⁺ and IL-17⁺ T cells among the responder T cells was determined by intracellular staining after 5 days of culture as described above.

Adenosine in the medium of cultured cells was measured by an Adenosine Assay Kit (Fluorometric) from Biovision (CA). Briefly, 25 µl of cultured cell supernatant were mixed with assay buffer, adenosine convertor, adenosine detector, adenosine developer and adenosine probe from the kits to compose a 100 µl reaction system. Kept in room temperature for 15 min and protected from the light. Fluorescence was read in a SpectraMax iD5 multi-mode microplate reader (Molecular Devices, LLC. USA) at Ex/Em = 535/587 nm.

Purified CD3⁺ T cells from IRBP_1–20-immunized TCR-δ^−/− mice were stained with CFSE (Sigma-Aldrich) as described previously (36). Briefly, the cells were washed and suspended at 5 × 10⁶ cells/ml in serum-free RPMI 1640 medium; cells were then incubated at 37°C for 10 min with gentle shaking with a final concentration of 5 μM CFSE before being washed twice with, and suspended in, complete medium, stimulated with anti-CD3 antibodies which were pre-coated on 24 well plate. Some 48 h later, the T cells were harvested and analyzed by flow cytometry.

In a 6-well plate, 2 × 10⁶/well BMDCs were co-cultured with 1 × 10⁵/well γδ or αβ T cells for 24 h. After DCs were separated from T cells, the BMDCs were irradiated (5,000 Rad) and seeded in a 24-well plate at 3 × 10⁴/well with CD3⁺ cells isolated from immunized B6 mice. Five days later, the percentage of IFN-γ⁺ and IL-17⁺ T cells among the responder T cells was determined by intracellular staining, followed by FACS analysis, as described previously (37).

αβ T cells (1.8 × 10⁶) were collected from IRBP_1-20-immunized B6 mice on day 13 post-immunization. To obtain enough cells, the cells obtained from all six mice in the same group are routinely pooled before the T cells are further enriched. The cells were co-cultured for 48 h with irradiated spleen cells (1.5 × 10⁶/well) as APCs and IRBP_1–20 (10 μg/ml) in a 24-well plate under either Th1 (culture medium supplemented with 10 ng/ml of IL-12) or Th17 polarized conditions (culture medium supplemented with 10 ng/ml of IL-23) (37, 38). Cytokine (IFN-γ and IL-17) levels in the serum and 48 h of culture supernatants were measured by ELISA (R & D Systems).

ELISA kits (E-Bioscience) were used to measure serum IFN-γ and IL-17 levels on day 13 post-immunization and in the 48 h culture supernatants of responder T cells isolated on day 13 post-immunization from IRBP_1–20-immunized B6 or TCR-δ^−/− mice.

The results in the figures are from a representative experiment, which was repeated 3–5 times. We used 2-way Students t-tests unless otherwise specified. Data were presented as the means with error bars for standard error of mean (SEM). Statistical analysis and graphing were performed in Excel software (Microsoft Corp). Asterisks (**) in graphs indicated p ≤0.05 representing statistical significance.

Our recent studies showed that adenosine has a distinct effect on Th17 versus Th1 responses; it tips the balance between Th1 and Th17 responses towards the latter (38, 39). To determine the contribution of DCs to such contradictory effects of adenosine on antigen-specific Th1 and Th17 responses, we examined the antigen presenting (AP) function of BMDCs, before and after exposure to NECA—a non-selective adenosine receptor agonist (34). The αβTCR⁺ responder T cells obtained from immunized B6 mice were stimulated in vitro with the immunizing antigen and the treated BMDCs at ratio of DC:T = 1:20 under Th1 (culture medium added with IL-12) or Th-17 (culture medium added with IL-23) polarizing conditions (39, 40). Th17 responses were assessed by evaluating numbers of IL-17⁺ T cells activated after a 5-day in vitro stimulation under Th17 polarizing conditions. After a 5-day in vitro stimulation by the immunizing antigen, under Th1 or Th17 polarizing conditions (27, 38), a significantly increased number of the CD3⁺ responder T cells became IFN-γ and IL-17 positive after stimulation by lipopolysaccharide (LPS)-treated BMDCs ( Figure 1A , upper panel). The IL-17⁺, but not the IFN-γ⁺, cells further increased if BMDCs were dually treated with LPS and NECA ( Figure 1A , lower panel). Cytokine tests ( Figure 1B ) showed that IL-17 production by responder T cells was also enhanced after stimulation by BMDCs dually treated with LPS and NECA, whereas the IFN-γ production was inhibited.

To determine whether AMP has a similar effect on BMDCs or whether BMDCs can convert AMP to adenosine and thus augment Th17 responses, the AP function of BMDCs was examined, before and after treatment with AMP, with or without a prior LPS treatment. Results in Figure 2A (upper panels) showed that AMP enhances Th17 promoting activity in LPS-treated, but not untreated, BMDCs. The number of IL-17⁺ T cells among the responder T cells was higher when BMDCs were pretreated with LPS. The number was further increased when the LPS-treated BMDCs were additionally treated with AMP ( Figure 2B ). However, BMDCs that were not treated with LPS did not respond to AMP, suggesting that LPS-treatment enabled BMDCs to respond to AMP. The effective doses of AMP range are from 1 to 8 µM and dose dependent effect was not apparent. Therefore, in subsequent studies we used the concentration of 1 µM.

CD73 is the main ecto-enzyme converting AMP to adenosine (12, 15). To test the prediction that LPS treatment enables BMDCs to express CD73, which allows BMDCs to convert AMP to adenosine, leading to augmented Th17 responses, we prepared CD73^+/+ (isolated from CD73^+/+ B6 mouse) and CD73^−/− BMDCs (isolated from CD73^−/− mouse) and compared the AP function between CD73^+/+ and CD73^−/− BMDCs, before and after treatment with LPS (100 ng/ml) and/or AMP (1 μM) ( Figure 2A ). The results show that treatment with LPS enhanced the Th17-promoting effect of both CD73^+/+ and CD73^−/− BMDCs; however, additional treatment of BMDCs with AMP augmented the Th17 responses presented by CD73^+/+, but not by CD73^−/−, BMDCs ( Figure 2A , lower panels). These results indicate that CD73^−/− BMDCs were unresponsive to AMP and that CD73 expressed by BMDCs may convert AMP to adenosine, which augments Th17 responses (41). We also measured IL-12/IL-23 production of BMDCs, after treatment with LPS and/or AMP. As demonstrated in Figures 3A, B , LPS exposure induced both CD73^+/+ and CD73^−/− BMDCs to produce increased amounts of IL-12 and IL-23. However, double exposure to LPS plus AMP enabled CD73^+/+, but not CD73^−/−, BMDCs to further increase IL-23 production and decrease IL-12 production.

Testing of AMP function by assessing cytokine production of responder T cells showed that AMP treatment further enhanced Th17 responses presented by CD73^+/+ but not CD73^−/− BMDCs ( Figure 3C ). Phenotypic examination of BMDCs, before and after LPS treatment, showed ( Figure 3D ). LPS-treated BMDCs expressed significant levels of CD73, whereas the CD73 level had been undetectable before treatment. We also measured adenosine amounts in the cultured supernatants of CD73^+/+ and CD73^−/− BMDCs after stimulation with LPS with or without AMP treatment. The results show that the adenosine concentration was significantly increased in AMP-treated CD73^+/+ but not CD73^−/− BMDCs ( Figure 3E ).

To determine whether CD73 expressed on splenic DCs is also functionally important, we examined the AP function of CD73^+/+ (isolated from B6 mouse) and CD73^−/− (isolated from CD73^−/− mouse) splenic DCs. As demonstrated in Figure 4A , approximately one third of the splenic CD11c⁺ cells in wt-B6, but not the CD73^−/− mouse are CD73⁺. αβTCR⁺ responder T cells separated from immunized B6 mice were stimulated for 5 days in vitro with the immunizing antigen and irradiated splenic DCs, under Th1- or Th17-polarizing conditions. Evaluating numbers of IFN-γ⁺ and IL-17⁺ T cells among responder T cells shows that AMP enhances Th17 responses ( Figure 4B , left panels) stimulated by CD73^+/+, but not CD73^−/−, splenic DCs and it inhibits Th1 responses ( Figure 4B , right panels) stimulated by CD73^+/+, but not CD73^−/− as well. Cytokine test ( Figure 4C ) showed that responder T cells produced comparable amounts of IFN-γ and IL-17 when stimulated by either CD73^+/+ or CD73^−/−, splenic DCs in the absence of AMP. In the presence of AMP, however, increased IL-17 production was seen in stimulation of responder T cells by CD73^+/+, but not CD73^−/−, splenic DCs. Also, AMP inhibits IFN-γ production of responder T cells stimulated by CD73^+/+, but not CD73^−/−, splenic DCs.

CD73 is constantly expressed in ~80% of αβ T cells ( Figure 5A ). To determine whether CD73 molecules expressed on αβ T cells contribute to AMP-mediated enhancement of Th17 responses, αβTCR⁺ responder T cells enriched by MACS column were CFSE labeled before stimulating with plated bound anti-CD3 antibodies, in the absence or presence of AMP (1 µM), under Th17 polarizing conditions. The results show that AMP treatment does not significantly affect the Th17 responses ( Figure 5B ), which agreed with our previous finding that CD73 expressing αβ T cells are incapable of degrading AMP to adenosine (26). To further exclude the possibility that CD73 expressed on BMDCs but not on αβ T cells accounted for the augmented Th17 responses, we conducted a crisscross test in which responder T cells isolated from immunized CD73^+/+ (B6) and CD73^−/− (CD73^−/− mouse) mice were stimulated by CD73^+/+ and CD73^−/− splenic DCs. The results showed that AMP enhancement of Th17 responses is seen only in the presence of CD73^+/+ splenic DCs ( Figure 5C ). AMP suppressed Th1 responses, which is also seen only in the presence of CD73^+/+, but not CD73^−/−, splenic DCs, suggesting that CD73 expressing DCs are required in degradation of AMP to adenosine leading to inhibited Th1 and enhanced Th17 responses.

We have observed that mouse BMDCs acquired an increased ability to enhance Th17 responses after exposure to γδ T cells, and a significant portion of the BMDCs expressed CD73 after exposure to γδ T cells. To determine whether CD73 expression by BMDCs is required for DC-γδ interaction, we compared Th17 responses presented by CD73^+/+ and CD73^−/− BMDCs, after exposure to γδ T cells. Three BMDC preparations were compared for their interactive ability with γδ T cells—the CD73^+/+ BMDCs with ( Figure 6A , mid panels) or without ( Figure 6A top panels) prior exposure to LPS and the CD73^-/- BMDCs treated with LPS ( Figure 6A , lower panels). The γδ T cells were separated from immunized B6 mice using a MACS column (38). BMDCs were exposed to γδ T cells at a pre-determined optimal ratio of T: DC = 1:20 for 24 h. Then, the γδ T cells were removed and the BMDCs were collected, irradiated, and seeded onto 24-well plates (5 × 10⁵/well), followed by co-culture with responder T cells. The AP function of BMDCs was assessed by measuring IFN-γ⁺ and IL-17⁺ T cells among responder T cells and cytokine production, under Th1-, or Th17-polarizing conditions (27, 38). Results show that the LPS-treated ( Figure 6A , mid panels), but not the untreated ( Figure 6A , Top panels), CD73^+/+ BMDCs stimulated a significantly greater number of IL-17⁺ cells among responder T cells, after exposure to γδ T cells. However, the CD73^-/- BMDCs were unable to augment Th17 response by exposure to γδ T cells ( Figure 6A , lower panels). Cytokine tests showed that the LPS-treated, but not the untreated, CD73^+/+ BMDCs induced greater amounts of IL-17 production of responder T cells after interacting with γδ T cells ( Figure 6B ). IFN-γ production of responder T cells remained unchanged after the BMDCs were exposed to γδ T cells, further supporting the prediction that expression of CD73 by DCs crucially triggers a strong Th17 response; whereas disabling CD73 function on DCs prevents higher Th17 responses.