Adenosine mediated desensitization of cAMP signaling enhances T-cell responses

Abstract

Adenosine has long been regarded as a crucial anti-inflammatory agent that protects the host from excessive damage. It has been reported to play an important role in suppressing immune activation, particularly that of T cells. However, it is a general observation that induction of T-cell activation is an efficient event despite the high adenosine levels that are often present in the affected host due to injury or stress. We report here that prior to antigenic stimulation via TCR/CD3, exposure of T cells to adenosine desensitizes adeno-sine receptors, so as to create a window of time where the T cells are insensitive to this ubiquitous suppressor. T cells from mice that were pre-exposed to this manipulation showed stronger responses to antigenic stimulation; therefore, the P1 adenosine receptor desensitization demonstrated an adjuvant-like effect. Our results suggest that adenosine receptor desensitization may be a mechanism for T cells to escape the general suppression during early points of T-cell activation and may emerge as a potential alternative for vaccine adjuvants.

Introduction

As a general cell activation inhibitor, adenosine has long been known as an anti-inflammatory and immunosuppressive agent [1–3]. The strong effects are seen due to the engagement of P1 adenosine receptors (P1R) on immune cells that have been abundantly reported and are quickly becoming an important target for drug discoveries. The P1R can be divided into four subtypes, of which A1 and A2A are high affinity adenosine receptors and A2B and A3 are low affinity [4]. One common feature of the P1R is that they are all G-protein coupled transmembrane structures that regulate cAMP.

It was reported that A2AR deficiency rendered the host mice highly susceptible to septic shock, leading to massive tissue damage and mortality [5]. Subsequent work further demonstrated that A2AR signaling on T cells, likely in concert with adenosine deaminase (ADA, adenosine catabolic enzyme), controls adenosine levels and T-cell activation, as measured by CD25 expression [6, 7]. It is generally accepted that during infections and tissue damage [8], adenosine levels rise and signal via P1R, leading to the suppression of inflammation and T-cell activation [1, 9]. Two groups reported that adenosine production, via ecto-nucleotidase CD39 and CD73, is used by regulatory T cells as a means of controlling further T-cell activation [10, 11].

However, there are reasons to believe that the anti-activation dogma of adenosine in immune cells may not fully reflect the intricacies of its signaling. Some receptors in this group signal via inhibitory (Gi) instead of the conventional stimulatory (Gs), G proteins and therefore proteins and block cAMP production. It is also noted that the high affinity P1R, particularly A1 and A2A, can sense the nucleoside under physiological conditions where adenosine levels are thought to be low [4, 12]. Our general understanding of adenosine signaling on T cells also needs to be reconciled with the fact that T-cell activation is fully functional in situations where adenosine suppression is thought to be the highest. It is therefore prudent to search for mechanisms whereby T cells can escape from this suppression. On cultured T-cell lines, it was found that surface expression of ADA reduced adenosine availability [13–15]. Our previous work on DC similarly shows that the function of ADA bound to the cell surface is a prerequisite for DC activation in response to TLR ligation [16]. However, whether this mechanism involving ADA is sufficient to explain the insensitivity of T cells to adenosine during activation is unknown.

One of the mechanisms to reduce adenosine sensitivity, often observed in neurological tissues, is receptor desensitization. For instance, both A2AR and A1R are known to be functionally sequestered following their exposure to ligands [17–20]. The effect can be achieved via receptor phosphorylation or down-regulation [21–26]. Whether a similar event occurs during T-cell activation was unknown.

Here we report that pre-treatment of T cells in vivo with adenosine analogues significantly enhances the subsequent activation via TCR or CD3. This enhancement is achieved by cAMP mediated P1R desensitization. Following pre-exposure to adenosine analogues, T cells demonstrate stronger responses and are insensitive to adenosine signaling. The “priming” effect appears to be a result of induced functional dissociation of G-proteins to adenylate cyclase, rendering T cells insensitive to adenosine. Our work suggests a mechanism whereby T cells escape adenosine suppression during early activation. We found that this desensitization mechanism can be utilized to induce strong T-cell activities, indicating its potential value in human vaccine development.

Results

T cells are sensitive to the suppressive effect of multiple adenosine derivatives

We first studied the presence of P1R on T cells. We purified C57BL/6 splenic CD4⁺ T cells with MACS beads, and performed real-time PCR to survey P1R messages. All the four messages were detected, although A1R was weaker (higher delta cycle threshold (ΔCT) values) (Fig. 1A). To analyze if adenosine level fluctuation impact on their expression, we also analyzed splenocytes from mice that had been previously injected with 5-(N-ethylcarboxamido) adenosine (NECA) overnight. It appears that only A3 was downregulated during the process (p-value 5 0.031) (Fig. 1A). We treated purified splenic CD4⁺ T cells from C57BL/6 mice with plate-bound anti-CD3 antibody and measured their IL-2 production as an indicator of T-cell activation. When high levels of exogenous adenosine were present in the cell culture, IL-2 production was reduced (Fig. 1B). We measured the effect of non-degradable adenosine derivatives on T-cell activation, because ADA in serum media rapidly converts adenosine to inosine [16]. We tested the OT-II splenocytes response to OVA, and their response in the presence of specific P1R agonists (Fig. 1C). Total splenocytes (containing antigen presenting cells) were cultured for 24 (data not shown) or 48 h and IL-2 production was measured. We confirmed in previous reports that CGS (A2A) and NECA (pan P1R) inhibited T-cell antigen-specific responses. 1-Deoxy-1-[6-[[(3-iodophenyl)methyl] amino]-9H-purine-9-y l]-N-methyl-b-d-ribofuranuronamide (IBMECA), a A3R agonist, enhanced T-cell activation, which was consistent with previous reports that A3R downregulates cAMP and increases cellular activation [1]. A1R agonists (N6-cyclopentyladenosine, CPA and 2-chloro-N-cyclopentyl-2'-methyladenosine, MeCCPA) also downregulated T-cell activation. To confirm that this response is not limited to this specific antigen reaction, we also cultured purified splenic CD4⁺ T cells from C57BL/6 mice and stimulated them with an anti-CD3 antibody coated plate (Fig. 1D). The suppressive effect of CPA and NECA was evident as well. We repeated the assay with purified CD8⁺ T cells from OT-I splenocytes. MACS purified CD8⁺ T cells from OT-I mice were stimulated with epitope peptide SIINFEKL in the presence of adenosine derivatives (Fig. 1E). Again CGS and NECA blocked the IFN-γ production by CD8⁺ T cells, as seen with the CD4⁺ T cells. CPA and MeCCPA showed identical inhibitory effects.

Figure 1.

Adenosine suppresses CD8⁺ and CD4⁺ T-cell activation. (A) Real-time PCR ΔCT analysis of P1R messages versus GAPDH. Filled bars: purified splenic CD4⁺ cells from untreated C57BL/6 mice; Open bars: from C57BL/6 mice pre-injected with NECA. (B) CD4⁺ T cells from C57BL/6 mice were purified and incubated with plate-bound anti-CD3 (activation) mAb in 96-well plates, with or without 5 μM adenosine as indicated. Supernatants were collected after different time points and IL-2 levels were determined by an ELISA kit. All data points shown (and henceforth) were performed in triplicates. UT: untreated. (C) Similar to (B), splenocytes (2 × 10⁶/mL) from OT-II mice were activated with a soluble peptide ISQAVHAAHAEINEAGR (OVA 323–339) (1 mg/mL) for IL-2 production in the presence of adenosine receptor agonists. CPA (10 μM): A1 agonist; CGS (10 μM): A2A agonist; IB-MECA (10 μM): A3 agonist, NECA (10 μM): a non-selective adenosine agonist, EHNA (10 μM): an ADA inhibitor. (D) CD4⁺ T cells from C57BL/6 mice were purified and incubated with anti-CD3 antibody as in (B), with or without adenosine agonist CPA (10 μM) or NECA (10 μM). Supernatants were collected at 24 h and IL-2 production was determined by ELISA. (E) Splenocytes (2 × 10⁶/mL) from OT-I mice were incubated with SIINFEKL (10^–7M) with or without adenosine agonists as indicated. Supernatants were collected after 48 h and IFN-γ levels were measured with an ELISA kit.

Pre-exposure to adenosine renders T cells hyper-reactive to antigen stimulation

Our central question is whether there are any balancing mechanisms to the seemingly dominant effect of adenosine suppression. One additional issue is whether adenosine can signal without concomitant T-cell activation [4]. We pre-injected C57BL/6 mice with adenosine derivatives overnight and studied their T-cell responses after purification. The design was to mimic T-cell responses in relation to adenosine regulation in vivo at the systemic level. Tissue stress and inflammation lead to higher adenosine levels, which happened before any antigenic encounter by T cells, as the latter event only follows antigen presenting cell migration to the draining LN, some point after 24–48 h. In other words, T cells are likely exposed to adenosine during injury or infection prior to their encounter with antigens presented by DC. Surprisingly, the in vivo pre-exposure to NECA or CPA upregulated the subsequent CD4 T-cell activation (IL-2 production), via anti-CD3 antibody stimulation. However, CGS (A2AR) pre-injection did not have the same effect (Fig. 2A). This result suggests that a positive regulatory signal was transmitted by CPA/NECA, and this pre-T-cell activation signal was not dependent on A2AR. To confirm the specificity, we performed experiments with 8-cyclopentyl-1,3-dipropylxanthine (DPCPX) to block the “A1R” activation. We repeated the assay as shown in Fig. 2A, and injected DPCPX to block CPA activity. It is clear from Fig. 2B that the upregulatory effect of CPA was indeed blocked by DPCPX (Fig. 2B).

Figure 2.

Pre-exposure to adenosine leads to enhanced subsequent T-cell activation. (A) C57BL/6 mice were injected i.p. with PBS, CPA (20 μM/kg), CGS (13 μM/kg) or NECA (1 μM/kg) (50 μL total volume). The same treatment was repeated after 18 h. After 24 h, mice were euthanized and splenocytes were harvested. CD4⁺ T cells were purified and activated with anti-CD3 antibody. IL-2 production was measured after 24 h. (B) same as A except CPA was co-injected with DPCPX, and supernatants were collected both at 24 and 48 h. (C) Same as (A), except CD8⁺ T cells were purified and tested. (D) Splenocytes from OT-I or OT-II were incubated with or without CPA (10 μM) or NECA (10 μM) for 24 h and were then stained with Annexin V and Propidium Iodide. Cells were analyzed by FACS to determine their viability. The percentage of apoptotic cells is indicated in each panel.

To see if this observation was limited to CD4⁺ T cells, we again purified CD8⁺ T cells from C57BL/6 splenocytes and performed an identical stimulation assay (Fig. 2C). The result was the same: CPA pre-injection upregulated the subsequent response to anti-CD3 stimulation. The pre-injection of CGS did not alter the IL-2 production at all, similar to the PBS control.

Since adenosine reduces cellular activation, it was very possible that T cells treated with adenosine might have an extended life span in comparison to regular T-cell activation. We checked the apoptotic state of T cells with annexin V and propidium iodine staining following their activation with plate-bound anti-CD3 antibody (Fig. 2D). The addition of CPA or NECA (data not shown) did not change the state of cell death/apoptosis in either OT-I or OT-II splenocytes at the time points tested (5 h, data not shown) (Fig. 2D).

Desensitization is not mediated exclusively by A1R or A2AR

The effect of CPA and DPCPX, displayed in the assays presented above, implicates A1R's involvement in desensitization. To confirm the results with a genetic approach, we tested A1R KO mice. These tested mice were genotyped both by PCR (Fig. 3A) and mRNA reverse transcription (Fig. 3B), and were found to be A1R deficient as previously reported [27]. They also display unique resistance to the tranquilizing effect of CPA and NECA, confirming the absence of A1R, as a neurological observation [28, 29].

Figure 3.

The enhanced T-cell response is mediated by NECA and CPA. (A) PCR genotyping of C57BL/6 and A1R KO mice for A1R deletion was performed as described in Materials and methods. Only WT and deletion mutant gene products were present in C57BL/6 and A1R KO mice, respectively. (B) cDNA, reverse transcribed from mRNA isolated from kidney cell lysate from C57BL/6 and A1R KO mice, were analyzed by PCR for the presence of A1R transcript as described in Materials and methods. Rodent GAPDH was used as the control. GAPDH transcript controls were present in both preparations. (C) A1R KO mice were injected i.p. with PBS or CPA (20 μM/kg) in a total volume of 50 μL. The same treatment was repeated after 18 h. After 24 h, mice were euthanized and splenocytes were harvested. CD4⁺ T cells were purified and activated with anti-CD3 antibody. IL-2 production was measured after 24 h incubation. (D) Similar to Fig. 2A except that A2AR KO mice were used in place. (E) CD4⁺ T cells from A1R KO mice were purified and incubated with anti-CD3 antibody, with or without adenosine agonist CPA (10 μM) or NECA (10 μM). IL-2 in the culture supernatants was measured after 24 h.

To our surprise, the priming effect in the A1R KO mice remaind intact with CPA treatment (Fig. 3C). The effect of NECA was still reversed by DPCPX. To ascertain the results were not a one time experimental outlier, we performed some of the experiments to be discussed below with A1R KO mice, and on no occasion did we see the A1R deficiency alter the experimental outcome. As a remote possibility, we tested the suggestion that CPA treatment may modulate A2AR signaling [30] and found that A2AR KO mice also showed heightened IL-2 production after the pre-injection of CPA or NECA (Fig. 3D).

While the effect was not directly compared within all potential combinations, and some quantitative differences may exist between different knockouts and WT, it is still clear that the desensitization effect is not mediated exclusively by either A1R or A2AR. CPA treatment on CD4⁺ T cells appears to be targeting multiple P1R, behaving more similarly to NECA. Indeed, A1R KO CD4⁺ T-cell activation by plate-bound anti-CD3 antibody was blocked by CPA (Fig. 3E)

The “priming” effect is not due to inhibition of cAMP

There are several possible mechanisms that may explain the enhanced responses described in Fig. 2 and 3. One of which is via Gi, which could permit enhanced T-cell activities via down-regulation of cAMP. To this end, we conducted two sets of assays. First, we stimulated purified splenic CD4⁺ T cells with anti-CD3 antibody in the presence or absence of CPA and measured the levels of cAMP in the cell culture. We decided to measure several time points and determined that a 15–20 min period of exposure produced the highest readings (Fig. 4A). At this time point, CPA stimulated cAMP and showed a synergistic effect with forskolin, a Gi-independent direct adenylate cyclase stimulator (Fig. 4B). As a control, A3R engagement showed a decrease in cAMP levels. We also performed Gi versus Gs assay by using cholera toxin (CT) and pertussis toxin (PT). C57BL/6 CD4⁺ T cells were activated with anti-CD3 antibody and the activation was blocked with CGS and CPA (Fig. 4C). CGS mediated inhibition was reversed to some extent by PT, but not CT, as expected. CPA mediated suppression was identical to CGS in its response to the two inhibitors, confirming the sole involvement of Gs in A1R-mediated T-cell suppression.

Figure 4.

Treatments leading to increased T-cell responses upregulate cAMP production. (A) and (B) Purified C57BL/6 CD4⁺ T cells from C57BL/6 were suspended in the cell culture medium and stimulated with immobilized anti-CD3 antibody in the presence of adenosine agonists or Forskolin. At different time points, media were aspirated and lysed with 0.1 M HCL (20 μL in 200 μmL medium), and then incubated at room temperature for 20 min. Supernatants after centrifugation were acetylated with KOH and acetic anhydride. The concentration of cAMP was derived from the raw data following manufacturer's instructions. (A) The preliminary experiment that determined the optimal time point for (B). UT: untreated; For or F: Forskolin; C: CPA; A3ag or A3: IB-MECA. (C) Purified CD4⁺ T cells were suspended in the cell culture medium and incubated on immobilized anti-CD3 in the presence of adenosine agonist CPA (10 μM) or CGS (10 μM), with or without antibody CT (10 ng/mL) or PT (2 ng/mL). Supernatants were collected after 24 h incubation and IL-2 was determined by an ELISA kit. CON: anti-CD3 antibody only.

The “priming” effect is due to T-cell P1R desensitization

The signaling on T cells triggered by CPA quickly leads to increased cAMP, reaching a maximum in 15 min. However, the increased level of cAMP is not sustained. This observation presents the possibility that exposure to adenosine over a longer term may not be able to activate cAMP constitutively. In other words, adenosine receptors may become desensitized, removing this negative feedback in subsequent T-cell activation.

We performed both in vivo and in vitro assays to test this possibility. We injected C57BL/6 mice with NECA. Splenic CD4⁺ T cells were purified after one day. The cells were then stimulated with plate-bound anti-CD3 antibody. At various points, NECA was added to the culture to test if these T cells were still sensitive to the suppression. T cells from mice pre-injected with NECA produced several fold higher IL-2 than untreated T cells (Fig. 5A). The addition of NECA at the time of CD3 stimulation reduced IL-2 production. But the T cells from the treated mice showed a resistance to the suppression. Although the basal level of IL-2 production from T cells without the pretreatment was not completely eliminated, the change was not statistically significant. A near identical effect was seen with the same T cells that had been treated with NECA in vitro after the harvest (Fig. 5B). C57BL/6 splenocytes were harvested and treated with NECA for 24 h. After washing, CD4⁺ T cells were purified and activated with plate-bound anti-CD3 antibody. NECA was added at the time of activation. These T cells showed robust IL-2 production and a similar substantial resistance to the concomitantly added NECA (Fig. 5B).

Figure 5.

Pre-exposure to adenosine leads to P1R desensitization. (A) C57BL/6 mice were injected i.p. with PBS or NECA (1 μM/kg). 18 h later, the same treatment was repeated. After 24 h, mice were euthanized and splenocytes were harvested. CD4⁺ T cells were purified and activated with anti-CD3 antibody with or without the presence of NECA (10 μM). IL-2 levels in the overnight supernatant were determined by ELISA. (B) Splenocytes from C57BL/6 were incubated with or without NECA (10 μM) overnight. Splenocytes were washed, and CD4⁺ T cells from C57BL/6 were purified and activated with anti-CD3 antibody. The rest of the assay was identical to (A). (C) C57BL/6 mice were injected i.p. with PBS or NECA (1 μM/kg). The same treatment was repeated after 18 h. After 24 h, mice were euthanized and splenocytes were harvested. CD4⁺ T cells were purified and activated with anti-CD3 antibody in the absence or presence of NECA (10 μM). After a 15 min incubation, cAMP was determined as in Fig. 4A and B. The top X-axis labels are the pre-treatment in vivo and the bottom labels are the subsequent in vitro treatment.

To study cAMP change in this event, we harvested the same T cells, as shown in Fig. 5A, and stimulated them with NECA. As predicted, T cells from the pre-treated C57BL/6 mice completely failed to respond to NECA, in sharp contrast to the control T cells (Fig. 5C).

The durations of desensitization

The desensitization mechanism is a clear advantage for immune activation. However, it is also evident that such an event must be short-lived to permit the protective effect of adenosine. We therefore measured how quickly T cells regained sensitivity to adenosine suppression. We injected C57BL/6 mice with CPA overnight and the CD4⁺ T cells were purified and stimulated with plate-coated anti-CD3 antibody. At various points, NECA was added into the culture. As shown in Fig. 6, it appears that the desensitization was quickly reversed upon the removal of CPA. Within the first 1.5 h, T cells retained full activation even in the presence of the adenosine analog introduced concomitantly with anti-CD3 antibody (data not shown). However, after a rest period of 2 h, they became fully sensitive to adenosine inhibition. Similarly, tested CD8⁺ T cells produced a nearly identical result (data not shown). This information is important in that it provides a logical explanation that may bridge various reports in T-cell responses to adenosine, and points out that the desensitization is a transient effect. Physiologically, since high adenosine levels persist, T cells in fact may be constantly desensitized until convalescence. Paradoxically, as long as adenosine is present at high levels, T cells would be insensitive to its effects.

Figure 6.

Desensitization of P1R is rapidly reversible. C57BL/6 mice were pre-injected i.p. with CPA as in Fig. 2A. After 24 h, mice were sacrificed and CD4⁺ T cells were purified from the splenocytes and stimulated with plate-bound anti-CD3 antibody. NECA (1 μM) was added at the beginning of or at various points into the culture. The IL-2 production in the supernatant was measured after 24 h.

P1R desensitization exerts a strong adjuvant effect on CD8⁺ T cells

Several folds higher than the standard T-cell activation conditions, the robust T-cell response following adeno-sine desensitization (Fig. 5) seemed to implicate a potential adjuvant effect. We followed up on this observation and tested whether the pre-treatment of adenosine could be used akin to a long-sought adjuvant to enhance T-cell antigen-specific responses.

We injected C57BL/6 mice with CPA or PBS overnight. The next day, the mice were immunized i.p. with 1 mg of soluble OVA. A same amount of OVA mixed with CFA was used as the positive control. One week later, the splenocytes were harvested and activated with soluble OVA and the IL-2 production after 48 h was measured. The pre-injection of CPA increased cytokine production, although the intensity was lower than the effect of CFA (Fig. 7A), and the overall response was low as well. To analyze if the desensitization had a stronger adjuvant effect on CD8⁺ T-cell response, we then injected OT-I mice with NECA, CPA and PBS overnight. The next day, the mice were immunized s.c. with 5 ug of OVA-coated latex beads. The same beads were mixed with CFA as the positive control. Two days later, splenocytes from treated mice were stimulated with C57BL/6 BM DC pulsed with SIINFEKL peptide and the cytokine production was measured after 48 h. At several settings of splenocytes density (Fig. 7B and C), both CPA and NECA exerted strong adjuvant activities, very similar to that of CFA.

Figure 7.

The transient desensitization of P1R exerts an adjuvant effect on T-cell activation. (A) C57BL/6 mice were injected i.p. with PBS or CPA (20 μM/kg). The same treatment was repeated after 18 h. After 24 h, soluble OVA (200 mg/kg) was used to immunize mice as described in Materials and methods. As a control, a mixture of CFA+OVA was injected into control mice. After 7 days, mice were euthanized and splenocytes were harvested and activated with soluble OVA. IL-2 concentration in the supernatant was determined by ELISA. (B) and (C) OT-I mice were injected i.p. with PBS, CPA or NECA as in (A). After 24 h, latex beads (5 μg) were injected s.c. into mice. As a control, a mixture of CFA+latex beads was injected into control mice. In total 24 h later, mice were euthanized and splenocytes (10 × 10⁶/mL for (B) or 5 × 10⁶/mL for (C)) were harvested and activated with SIINFEKL. After a 24-h incubation, IL-2 in the supernatant was determined by ELISA. This figure was independently repeated four times.

Therefore, the pre-treatment of adenosine behaved similarly to the use of CFA and provided a strong immune enhancing effect to the antigen-specific T-cell activation, particularly that of CD8⁺ T cells.

Discussion

The negative feedback caused by adenosine signaling has been reported by various groups and described under various settings. A2AR, A2BR and A3R are all known to play some roles in this suppressive regulation. Although the protective effect is important, it should be reconciled with the robust T-cell activation under high levels of adenosine during infection and injury. Prior to this report, this issue had not been directly assessed. Our results point out that T cells use a common method of impediment: the ligand-induced receptor desensitization, a prevalent feature of P1R in neurologic and other tissues [18, 20–23, 26].

Previous studies on non-immune cells have revealed some basic mechanisms of receptor desensitization [31]. For P1R, such a functional reduction appears to be the result of receptor phosphorylation, which can occur among either homologous or heterologous receptors [32–34]. Such a change leads to reduced adenylate cyclase activities [35, 36]. Several aspects of this desensitization effect in the immune cells are interesting. First, the desensitization effect is bypassed by forskolin which directly modulates cAMP production. This suggests that adenylate cyclase activation remains intact and functional, and that the desensitization effect is caused by the functional dissociation between the Gi and the cAMP synthesis, as reported previously [37]. Second, following desensitization, the overall ability to sense adenosine (NECA) is reduced. Since immune cells are in general sensitive to adenosine signaling via A2AR [38] (Fig. 1), which is targeted by NECA, our results suggest that the overall adenosine feedback is blocked. This cross-inhibition may have important biological implications in terms of total adenosine signaling via all P1R on T cells under high adenosine levels. Third, the desensitization effect tapers off rapidly. Within 2 h, T cells completely regain their sensitivity to adenosine inhibition. This observation suggests that in vivo, once adenosine levels are falling back to normal, T cells are under the protection of adenosine. Conversely, it appears that as long as adenosine remains high, T cells are unlikely to be suppressed by adenosine. It may indicate that during infection and tissue stress, T cells remain insensitive to the adenosine suppression. It should be noted, however, that the injection of adenosine derivatives is in general a systemic effect. However, how local adenosine levels fluctuate in tissue stress, in relationship with the desensitization effect, needs to be studied in the future.

Adenosine suppression on immune cells can be potent. In DC and in some settings with T cells, the suppression of cellular activation is nearly absolute [16]. It is therefore reasonable to assume that the immune system uses multiple methods to escape such an inhibition. We recently reported that the availability of surface ADA is crucial to ensure mouse DC activation. The same mechanism is likely used on the T-cell surface [13]. The question is whether these mechanisms are enough. In addition, the ADA-related adenosine removal is likely constitutive, unlike the dynamic nature of P1R desensitization, as it may come into effect only upon inflammation. We propose that T cells combine several mechanisms to offset the rapid surge of adenosine at the earliest stages of inflammation and injury.

It was a surprise that in our settings, the desensitization effect is mediated by CPA and blocked by DPCPX, yet not via A1AR. Both reagents trace their specificity origins to adenosine signaling studies in non-immune cells [39–42]. The A1R deficient mice used in our assays were known to have A1 signaling defects in other physiological functions, such as urine retention in the kidneys and cardiovascular outputs [27, 43]. Our work seems to indicate that in the immune cells, CPA lacks its proclaimed specificity in other tissues and targets multiple P1R, similar to NECA. This is not surprising as P1R agonists demonstrate vastly different targeting specificities and efficiencies in different tissues [4], and the specificity issues of A1 agonists and antagonists in the immune system have surfaced in some earlier reports [30, 44, 45]. With the realization that A2AR does not mediate this desensitization effect, we propose that multiple P1R participate in the process.

Our work points out an additional layer of shielding from the suppressive effect by adenosine in T-cell activation. This is in addition to adenosine deamination by ADA on T cells or other immune cells. We would like to argue this short-lived desensitization effect is necessary for T-cell activation to overcome the initial hurdle of activation induction, while not interfering with the general protective effect of adenosine at resting times.

Materials and methods

Mice, cells and reagents

C57BL/6 (WT), OT-I (C57BL/6-Tg (TcraTcrb) 1100Mjb/J) and OT-II (C57BL/6-Tg (TcraTcrb) 425Cbn/J) mice were purchased from Jackson Laboratories. T-cell receptor-specific recognitions utilized in this report are as follows: OT-I mice are C57BL/6 mice that express a transgenic TCR that recognizes a peptide derived from residues of 257–264 OVA. OT-II mice are C57BL/6 mice that express a transgenic TCR that is specific for OVA 323–339. Animal protocols were approved by the University of Calgary Animal Care Committee and met the Canadian Guidelines for Animal Research. Splenic CD4⁺ T cells and CD8⁺ T cells were isolated with MACS CD4 or CD8 Microbeads (anti-mouse) from Miltenyi Biotec. In all spleen T-cell assays, mice were euthanized by CO₂, and spleens were harvested and ground in cell culture media (10% FBS plus 1 mM HEPES, 25 μM 2-ME and pen/strep antibiotics). All secondary antibodies were purchased from Jackson ImmunoResearch. All other antibodies and ELISA kits were from eBioscience except for the cAMP EIA kit, which was from Cayman Chemical, and annexin V apoptosis staining kit, from R&D systems. The Real Time QuantiTect SYBR Green PCR kit and RNeasy Mini Kit were from QIAGEN and the reverse transcription kit was from Invitrogen.

P1R agonists and antagonists

Adenosine, NECA and CPA were obtained from Sigma-Aldrich. Erythro-9-(2-hydroxy-3-nonyl) adenine (EHNA), CGS, MeCCPA, and DPCPX were from Tocris Cookson. Most of these reagents were dissolved in DMSO per manufacture's instructions. All preparations were warmed to 37°C and immediately added to assay wells to avoid crystallization. Unless indicated otherwise, the reagents were used at the following concentrations: Forskolin 25 μM, Adenosine 5 μM, CPA 10 μM, CGS 1 μM, DPCPX 10 μM, EHNA 10 μM, IB-MECA 100 nM and NECA 10 μM. The presumed specificities of the reagents are as follows: CPA, A1R agonist; CGS, A2AR agonist; IB-MECA, A3R agonist; NECA, pan-P1R agonist; DPCPX, A1R antagonist; and EHNA, ADA inhibitor.

T-cell stimulation assays

In some experiments, splenocytes were harvested from C57BL/6 mice and CD4⁺ or CD8⁺ T cells were purified with CD4 or CD8 magnet microbeads according to the manufacture's protocols. Anti-CD3 epsilon chain mouse antibody was diluted with PBS (5 μg/mL) and loaded into 96-well plates. Plates were coated at 37°C, in a 5% CO₂ humidified incubator for 6 h and washed twice with PBS. Purified CD4⁺ or CD8⁺ T cells were then added into the anti-CD3 antibody coated plates and incubated at 37°C, in a 5% CO₂ humidified incubator for various hours. Supernatants were collected and cytokine levels were measured by ELISA.

In OT-I mouse experiments, splenocytes were harvested from OT-I mice and control mice, and suspensions (2 × 10⁶ cells/mL) were transferred into 24-well plates. SIINFEKL peptide (10^–7M) was added to the corresponding wells. Cells were incubated at 37°C, in a 5% CO₂ humidified incubator for various hours. Supernatant was collected and cytokine levels were measured by ELISA. In OT-II mouse experiments, splenocytes were similarly prepared, except that soluble OVA (50 μg/mL, or as indicated) was added to the corresponding wells in place of the SIINFEKL peptide.

Pre-treatment with adenosine derivatives

C57BL/6, A1R KO or other mice received i.p. injections with PBS, CPA (2 × 10^–5 M/kg), NECA (10^–6 M/kg) or DPCPX (10^–6 M/kg). In total, 18 h after the first injection, the same injection was repeated to boost the effect. Unless otherwise indicated, after 24 h, mice were euthanized and splenocytes were harvested. The apoptosis assay was as previously reported [16].

cAMP measurement

CD4⁺ T cells were isolated from splenocytes from C57BL/6 with magnet microbeads. Cells were activated with anti-CD3 antibody, in the presence or absence of adenosine agonists. In some assays, Forskolin was added to the plate. At different time points (10, 20 and 30 min; 1, 4 and 24 h), media was aspirated and lysed with 0.1 M HCL (20 μL in 200 μL medium). The lysate was incubated at room temperature for 20 min, and then centrifuged at 1000 × g for 10 min. The supernatant was transferred into a clean tube. To maximize the reading intensity, the samples were acetylated with 4 M KOH (100 μL into 500 μL sample) and acetic anhydride (25 μL into 500 μL sample) in quick succession. The mixture was vortexed for 15 s, and an additional 25 μL KOH was added followed by vortexing. The mixture was then incubated at 4°C overnight. The developing and the reading of the plate were carried out according to the manufacturer's protocol. The plate was read with a visible light 96-well plate reader at a wavelength between 405 and 420 nm. The raw data points were converted into cAMP units using a spreadsheet program provided by the manufacturer.

RNA isolation, RT-PCR and real-time PCR

Total RNA was isolated from kidney cells from both A1 adenosine receptor knockout mice and C57BL/6 mice using the RNeasy Mini kit by Qiagen and its corresponding protocol. cDNA from C57BL/6 and A1R^–/– mice was synthesized from the total isolated RNA by reverse transcription using a cDNA synthesis kit by Invitrogen. The freshly synthesized cDNA was then used in the PCR reaction with a pair of A1R specific primers (forward: GTGATTTGGGCTGTGAAGGT, and reverse: CAAGGGAGAGAATCCAGCAG) with an expected product size of 321 bp. Standard rodent GAPDH primers (Applied Biosystems) were used as a quantitative control. Each PCR reaction was made with 5 μL 10 × PCR Buffer with 4 mM MgCl2, 5 μL 10 × CoralLoad, 0.4 μL 25 mM dNTPs, 2 μL (10 uM) of each primer, 0.25 μL Taq, 30.35 μL ddH20 and 5 μL cDNA.

DNA was extracted from mouse-tail samples using Qiagen DNeasy Blood and Tissue Kit. DNA was tested using PCR for the presence of WT and mutant genes, using A1R (WT) (forward: TTGGCTGGAACAACCTGAGT, reverse: GTGGTATCGGAAGGCATAGA) and LacZ specific (forward: TTCACTGGCCGTCGTTTTACAACGTCGTGA, reverse: ATGTGAGCGAGTAACAACCCGTCGGATTCT) primers with products of expected sizes of 449 and 364 bp, respectively. Cycling conditions were 95°C for 10 min, followed by 40 repeats of 95°C for 1 min, 60 or 65°C (WT and LacZ, respectively) for 1 min and 72°C for 1 min 30 s, ending with 10 min at 72°C.

For real-time PCR, total RNA was isolated from purified CD4⁺ T cells from both the PBS and NECA injected C57BL/6 mice using the RNeasy Mini Kit supplied by QIAGEN. Five to ten million cells were homogenized using QIAshredder spin columns. The RNA quality and harvest efficiency were determined using the Nano-drop 1000. cDNA synthesis was performed using reagents supplied from Invitrogen. The reaction mixture, consisting of 1 μL random primers, 9 μL RNA, 1 μL 10 mM dNTP, and 1 μL sterile H₂O, was heated at 65°C for 5 min. The reaction mixtures were then placed on ice and centrifuged to ensure all the contents were on the bottom. Briefly, 4 μL 5 × first strand buffer, 2 μL 0.1 M DTT, and 1 μL RNase OUT were added to the reaction mixture and incubated at 37°C on a heating block for 2 min. In total 1 μL of M-MLVRT was added and mixed by pipetting up and down and left at room temperature for 10 min. The reaction mixture was then incubated at 37°C for 50 min and then the reaction was inactivated by heating the mixture to 70°C for 15 min. The cDNA was then stored at –20°C until real-time PCR. For the QuantiTect SYBR Green PCR Kit (AB), primers covering exon–exon junctions were designed to produce 100–150 bp PCR fragments. Mouse GAPDH was used as the internal control while the targets were mouse A1, A2a, A2b and A3 receptors. The primer sequences were as follows: mouse GAPDH forward-5'TTCACCACCATGGAGAAGGC3', reverse 5'GGCATGGACTGTGGTCATGA3'; mouse A1 receptor forward-5'TCCCTCTCCGGTACAAGACAGT3', reverse 5'CAGGTTGTTCCAGCCAAACA3'; mouse A2a receptor forward-5'GCTATTGCCATCGACAGATACATC3', reverse 5'AATGACAGCACCCAGCAAATC3'; mouse A2b receptor 5'TGGCGCTGGAGCTGGTTA3', reverse 5' GCAAAGGGGATGGCGAAG3'; and mouse A3 receptor forward 5'CAGTCAGATATAGAACGGTTACCACTCA3', reverse 5'GTTGCTTTTCTATTCCAGCCAAA3'. Real-time PCR was performed in the ABI PRISM 7000 using QuantiTect SYBR Green PCR protocol per manufacturer's instructions. Using the MicroAmp optical 96-well reaction plate, 20 μL of master mix and 5 μL of cDNA were added to the appropriate wells. Non-template controls, along with RNA controls, were run in order to determine the background noise. The total reaction volumes consisted of 12.5 μL of 2XQuantiTect SYBR Green PCR Master Mix, 6 μL of RNase-free water, 0.75 μL of Primer A and Primer B (final concentration 0.3 μM), and 5 μL of a 1 in 5 dilution of cDNA. The optical 96-well reaction plate was then sealed with an optical adhesive cover and placed in the centrifuge for 5 min. The ABI PRISM 7000 was setup to have 2 min at 50°C, 15 min at 95°C for the PCR initial activation step, and 40 cycles of three steps consisting of 15 s at 94°C, 30 s at 60°C and 30 s at 72°C. A dissociation step was also added after the 40 cycles were completed. Real-time PCR results were obtained using the ΔCT method as follows: ΔCT = Target CT –Normalizer CT (GAPDH).

Statistical analysis

All experiments were independently repeated at least three times, and all error bars are standard errors from each particular assay. All statistical analyses were performed with two-tailed Student t-test.

Abbreviations

ADA

adenosine deaminase

CGS

CGS 21680

CPA

N6-cyclopentyladenosine

CT

cholera toxin

ΔCT

delta cycle threshold

DPCPX

8-cyclopentyl-1,3-dipropylxanthine

EHNA

erythro-9-(2-hydroxy-3-nonyl) adenine

Gi

G-protein

Gs

G simulatory

IBMECA

1-deoxy-1-[6-[[(3-iodophenyl)methyl]amino]-9H-purine-9-y l]-N-methyl-b-d-ribofuranuronamide

MeCCPA

2-chloro-N-cyclopentyl-2'-methyladenosine

NECA

5-(N-ethylcarboxamido) adenosin

P1R

P1 adenosine receptor

PT

pertussis toxin