Control of lymphocyte egress from lymph nodes through β₂-adrenergic receptors

Because β₂ARs are predominantly expressed in lymphocytes compared with other subtypes of adrenergic receptors (Sanders, 2012), we treated mice with selective β₂AR agonists, clenbuterol, or salbutamol, to mimic activation of adrenergic nerves and test the possible role of β₂ARs in lymphocyte dynamics. Administration of a single dose of either β₂AR agonist resulted in a rapid reduction of blood B cells and CD4⁺ or CD8⁺ T cells in a dose-dependent manner (Fig. 1 A; Fig. S1 A; and not depicted). Notably, the reduction of blood lymphocytes was accompanied by a sharp decline of lymphocyte numbers in lymph (Fig. 1 B and Fig. S1 B). Consistent with the reported pharmacological properties of β₂AR agonists (Smith, 1998), clenbuterol was more potent than salbutamol. The ED₅₀ values of clenbuterol and salbutamol for decreasing circulating B cell numbers were 0.01 and 0.3 mg/kg, respectively, which were comparable to those reported for rodents (McElroy and O’Donnell, 1988). Although the lymphocyte numbers after administration of either β₂AR agonist exhibited a trough at 2 h (Fig. 1, C and D), their recovery was faster in salbutamol-treated mice, reflecting a shorter half-life of salbutamol compared with clenbuterol (Smith, 1998). Lymphopenia produced by β₂AR stimulation was prominent in B cells compared with T cells, which was consistent with the predominant expression of transcripts for β₂ARs in B cells (Fig. 1 E and Fig. S1, C and D). Flow cytometric analysis of the surface expression of β₂ARs was hindered by lack of an antibody to produce specific labeling.

The effect of clenbuterol on the numbers of circulating B and T cells was abrogated in β₂AR-deficient mice (Fig. 1, F and G; and not depicted), confirming that clenbuterol specifically acts on β₂ARs in this experimental setting. To determine whether the action of the β₂AR agonist was mediated by β₂ARs expressed on the radiosensitive hematopoietic compartment or radioresistant nonhematopoietic compartment, we administered clenbuterol to BM chimeras generated using β₂AR-deficient mice as donors or recipients. Reduction of B cell numbers in blood and lymph was significantly blunted, but not completely abrogated, in WT mice reconstituted with β₂AR-deficient BM cells (Fig. 1, H and I). In contrast, B cell numbers in β₂AR-deficient mice receiving WT BM cells were reduced to the same extent as WT recipients (Fig. 1, H and I). The same was true of T cells in the BM chimeras (unpublished data). These observations suggest that lymphopenia induced by treatment with the β₂AR agonist largely depends on β₂ARs expressed on hematopoietic cells although the involvement of nonhematopoietic β₂ARs cannot be excluded.

Because the reduction of lymphocytes in lymph as well as blood was reminiscent of changes induced by FTY720 treatment (Mandala et al., 2002), we hypothesized that stimulation of β₂ARs might inhibit lymphocyte egress from LNs. To test this possibility, we implanted osmotic pumps in mice, allowing continuous administration of clenbuterol or salbutamol, and then blocked lymphocyte entry to LNs by treatment with neutralizing antibodies against α4 and αL integrins for 22 h (Fig. 2 A; Lo et al., 2005). Treatment with either β₂AR agonist increased the numbers of B cells and CD4⁺ or CD8⁺ T cells that were retained in the mesenteric LNs after entry blockade in a dose-dependent manner (Fig. 2 B; Fig. S2 A; and not depicted), suggesting that β₂AR stimulation retards lymphocyte egress from LNs. Consistent with the effect on circulating lymphocyte numbers, higher doses of salbutamol were required to achieve the levels of egress inhibition by clenbuterol. On the other hand, treatment with the β₂AR agonists did not affect lymphocyte entry to LNs (Fig. 2 C and Fig. S2 B), suggesting their selective action on lymphocyte egress. In WT mice reconstituted with β₂AR-deficient BM cells, LN retention of B and T cells was not promoted by clenbuterol treatment (Fig. 2 D and not depicted), whereas β₂AR stimulation resulted in enhanced lymphocyte retention in the LNs of β₂AR-deficient mice reconstituted with WT BM cells (Fig. 2 E and not depicted). These observations indicate that stimulation of β₂ARs inhibits egress of lymphocytes from LNs in a cell-intrinsic manner.

There was a trend for β₂AR-deficient mice to have higher lymphocyte numbers in both blood and lymph compared with WT mice (Fig. 1, F and G), suggesting that physiological adrenergic stimuli through β₂ARs are involved in lymphocyte recirculation. To test the contribution of β₂ARs expressed on hematopoietic cells, we generated mixed BM chimeras using a mixture of CD45.2⁺ β₂AR-deficient and CD45.1⁺ WT cells, and examined lymphocyte distribution in lymphoid tissues. Compared with their frequencies in mesenteric LNs, blood and lymph showed a significant accumulation of β₂AR-deficient B cells over WT B cells (Fig. 3 A and Fig. S3 A). This accumulation was not observed in control chimeras generated using a mixture of congenically distinct β₂AR-sufficient and WT BM cells. These observations support that β₂AR-mediated signals act to limit lymphocyte egress from LNs in a cell-intrinsic manner. However, such biased distribution was not observed for CD4⁺ or CD8⁺ T cells (unpublished data). Because of the lower expression of, and thereby weaker signals through β₂ARs in T cells compared with B cells (Fig. 1 E), the difference in LN egress between WT and mutant T cells might have been compensated for during long-term BM reconstitution.

To directly address the role of lymphocyte β₂ARs in LN egress under physiological conditions, we performed short-term transfer experiments. β₂AR-sufficient or -deficient lymphocytes were transferred into WT hosts, allowed to equilibrate for 2 d, and further lymphocyte entry to LNs was blocked by integrin-neutralizing antibodies for 14 h. β₂AR-deficient lymphocytes were retained in the mesenteric LNs to a lesser extent than β₂AR-sufficient lymphocytes (Fig. 3 B and Fig. S3 B). This result confirms that physiological inputs through β₂ARs expressed on lymphocytes inhibit their egress from LNs. Consistent with unaffected lymphocyte homing by agonist stimulation of β₂ARs (Fig. 2 C), transferred CD45.2⁺ β₂AR-deficient lymphocytes entered LNs as efficiently as CD45.1⁺ WT lymphocytes (Fig. 3 C and Fig. S3 C), further ruling out the involvement of β₂ARs in lymphocyte entry to LNs.

To evaluate the contribution of adrenergic nerves in controlling lymphocyte egress from LNs, we treated mice with 6-hydroxydopamine (6-OHDA) to deplete peripheral neuronal terminals containing noradrenaline, and then performed entry blockade with integrin-neutralizing antibodies for 14 h. Lymphocyte numbers in mesenteric LNs were underrepresented in 6-OHDA–treated mice compared with control mice (Fig. 3 D), suggesting that inputs from adrenergic nerves retard lymphocyte egress from LNs. However, the contribution of humoral adrenergic inputs provided by adrenal glands cannot be excluded. Collectively, these findings indicate that physiological levels of inputs through β₂ARs, a part of which is provided by adrenergic nerves, help retain lymphocytes within LNs by limiting their egress and maintain homeostasis of their distribution among lymphoid tissues.

A recent study demonstrated that CXCR4-deficient B cells exited LNs more rapidly than WT B cells (Schmidt et al., 2013), suggesting that CXCR4, in addition to CCR7, might promote lymphocyte retention in LNs by counteracting the egress-promoting receptor S1PR1. Thus, we anticipated that stimulation of lymphocyte β₂ARs might modulate responsiveness of these chemoattractant receptors and inhibit egress from LNs. To assess the effect of β₂AR stimulation on the signal strength of these and other lymphoid chemoattractant receptors, we analyzed the activation of a small GTPase Rac1 in a pull-down assay for GTP-bound Rac1. Stimulation with clenbuterol resulted in prolonged activation of Rac1 in response to the CCR7 ligand CCL21 and CXCR4 ligand CXCL12 in both B and T cells, whereas the effect of β₂AR stimulation was unremarkable in response to the CXCR5 ligand CXCL13 in B cells (Fig. 4 A). Treatment with S1P induced little activation of Rac1 in B and T cells, and the effect of β₂AR stimulation were not detectable (Fig. 4 A and not depicted). In β₂AR-deficient lymphocytes, the effect of clenbuterol treatment on CCR7- or CXCR4-mediated Rac1 activation were abrogated (Fig. 4 B), confirming specific action of clenbuterol on lymphocyte β₂ARs. Stimulation with the β₂AR agonist in the absence of chemoattractants did not induce activation of Rac1 in lymphocytes (Fig. 4 C). These observations suggest that stimulation of β₂ARs preferentially enhances signals mediated by the retention-promoting chemokine receptors CCR7 and CXCR4.

To examine whether the enhanced signals affect lymphocyte chemotaxis, we tested Transwell migration in the presence of the β₂AR agonist using β₂AR-deficient lymphocytes as a reference. We used suboptimal concentrations of chemoattractants to favor the likelihood that possible chemotaxis-enhancing effects of β₂AR stimulation, if any, could be detected. Consistent with the results in Rac1 activation assays, B cells exhibited enhanced chemotactic responses mediated by CCR7 or CXCR4, but not S1PR1 or CXCR5 (Fig. 4 D and Fig. S4). In CD4⁺ or CD8⁺ T cells, enhancement of Transwell migration by β₂AR stimulation was detectable only in response to the CXCR4 ligand (Fig. 4 E and not depicted). The β₂AR agonist alone did not induce Transwell migration of lymphocytes (unpublished data). Unaltered CCR7-mediated chemotaxis in T cells is consistent with a previous observation that transgenic overexpression of CCR7 did not enhance Transwell migration of mature T cells in response to CCR7 ligands (Kwan and Killeen, 2004). Because T cells show substantially higher chemotactic responses to CCR7 ligands compared with B cells (Ngo et al., 1998), we speculate that the effect of β₂AR stimulation might be negated by the robust CCR7 responsiveness in T cells. Surface levels of CCR7 and CXCR4 on lymphocytes were not affected by agonist stimulation or deficiency of β₂ARs (unpublished data). Overall, these findings suggest that β₂AR stimulation enhances responsiveness of CCR7 and CXCR4 in lymphocytes.

Studies showed that different types of G protein–coupled receptors (GPCRs) form heteromeric complexes on the cell surface and cross-regulate their signaling (González-Maeso et al., 2008; Fribourg et al., 2011). These findings prompted us to test whether there is a physical association between β₂ARs and CCR7 or CXCR4. Confocal microscopy for the surface localization of epitope-tagged receptors expressed in 2PK-3 B lymphoma cells demonstrated that β₂ARs co-localized more frequently with CCR7 or CXCR4 compared with S1PR1 (Fig. 4 F), suggesting that β₂ARs resides in the proximity of CCR7 or CXCR4. To examine whether the β₂ARs form complexes with CCR7 or CXCR4, we performed co-immunoprecipitation assays. Using lysates prepared from transduced 2PK-3 cells, we found that β₂ARs were co-precipitated with CCR7 or CXCR4, but not S1PR1 or CXCR5 (Fig. 4 G). Although we cannot exclude the involvement of additional molecules, this result provides evidence for physical interactions of β₂ARs with CCR7 or CXCR4. However, analysis of these receptor interactions in mouse primary lymphocytes was not successful because of the lack of available specific antibodies for immunoprecipitation or immunoblotting of β₂ARs. Collectively, these findings indicate that β₂ARs are physically associated with CCR7 or CXCR4, which might contribute to the preferential enhancement of their signals by β₂AR activation.

The aforementioned findings suggest that stimulation of β₂ARs inhibits lymphocyte egress from LN as a consequence of enhanced lymphocyte retention through CCR7 and CXCR4. To test the involvement of CCR7 in β₂AR-mediated promotion of lymphocyte retention, we transferred CCR7-sufficient or -deficient spleen cells into WT mice and performed entry blockade with integrin-neutralizing antibodies for 18 h during β₂AR stimulation with clenbuterol. Treatment with the β₂AR agonist promoted LN retention of CCR7-sufficient and -deficient CD4⁺ or CD8⁺ T cells to the same extent (Fig. 5 A; Fig. S5; and not depicted). In contrast, enhancement of B cell retention was blunted, but not completely abrogated, by CCR7 deficiency (Fig. 5 B), suggesting that B cells are at least in part dependent on CCR7 in β₂AR-mediated promotion of LN retention. A previous study showed that CCR7 deficiency resulted in more rapid LN egress of T cells, but not B cells (Pham et al., 2008), highlighting the retention-promoting role of CCR7 in T cells. However, our data suggest that when CCR7-mediated signals were increased from the basal level, they might also promote LN retention of B cells.

To test the involvement of CXCR4 in LN retention of T cells and confirm a role for B cell retention, we blocked CXCR4 function using a specific antagonist AMD3100 and analyzed the dynamics of lymphocytes. Treatment with AMD3100 increased the numbers of CD4⁺ T cells and B cells in blood and lymph (Fig. 5, C and D). Although part of the large increase of blood B cells might reflect the role of CXCR4 in their retention within BM (Ma et al., 1999), these data suggest that CXCR4 is involved in LN retention of both B and T cells. To test this, we performed entry blockade for 18 h during the continuous administration of AMD3100 through osmotic pumps. CD4⁺ T cells and B cells were underrepresented in the mesenteric LNs of mice treated with AMD3100 (40 mg/kg/d) compared with control mice (Fig. 5 E), confirming that CXCR4 functions as a retention-promoting receptor for both lymphocyte populations. Because a similar effect was observed with low (20 mg/kg/d) and high (80 mg/kg/d) doses of AMD3100 (unpublished data), we considered that complete blockade of CXCR4 was achieved with 40 mg/kg/d of the drug.

To assess the effect of CXCR4 blockade on β₂AR-mediated enhancement of lymphocyte retention in LNs, we performed entry blockade upon combined treatment with AMD3100 and clenbuterol. Enhancement of T cell retention in LNs was abrogated by blockade of CXCR4 (Fig. 5 F), whereas enhancement of B cell retention was unaffected (Fig. 5 G). These results suggest that enhanced LN retention of T cells by β₂AR stimulation is largely dependent on CXCR4. Thus, although we cannot exclude the involvement of other undefined retention-promoting receptors, β₂AR-mediated enhancement of LN retention of B and T cells appears to be dependent on CCR7 and CXCR4, respectively.

Analogous to the action of FTY720, we hypothesized that β₂AR-mediated control of lymphocyte egress might affect the pathogenesis of inflammatory diseases. Indeed, previous studies showed that treatment with β₂AR agonists suppressed experimental autoimmune encephalomyelitis (EAE), an animal model of multiple sclerosis (Chelmicka-Schorr et al., 1989; Wiegmann et al., 1995), although the mechanisms of their action were not addressed. To test the role of β₂ARs in T cell–mediated inflammation, we used mouse EAE and delayed-type hypersensitivity (DTH) responses in the skin. We treated mice with clenbuterol in the course of EAE induced by immunization with myelin oligodendrocyte glycoprotein (MOG) in adjuvant. To target the trafficking of MOG-primed encephalitogenic T cells from LNs to the central nervous system without affecting their generation, the β₂AR agonist was administered for 3 consecutive days at the onset of disease. Consistent with the previous studies, treatment inhibited the progression of paralytic symptoms characteristic of EAE (Fig. 6 A). Conversely, β₂AR-deficient mice developed more severe symptoms compared with β₂AR-sufficient mice (Fig. 6 B), suggesting that physiological levels of inputs through β₂ARs are involved in EAE pathology. The disease induced in β₂AR-deficient mice was not affected by clenbuterol treatment (Fig. 6 C), indicating β₂AR specificity of clenbuterol in an inflammatory condition.

To induce skin DTH responses, we immunized mice with OVA in adjuvant, and 7 d later, ear skin was challenged with an intradermal injection of OVA that induced inflammation and ear swelling peaking ∼48 h later. Consistent with results in the EAE model, continuous administration of the β₂AR agonist started before OVA challenge in the ear attenuated the DTH responses (Fig. 6 D). Additionally, β₂AR-deficient mice exhibited an enhanced OVA-induced DTH response (Fig. 6 E) and unresponsiveness to clenbuterol treatment (Fig. 6 F). Despite increased inflammatory responses, β₂AR deficiency did not affect the generation of disease-related CD4⁺ T helper cells producing IL-17 (Th17) or IFN-γ (Th1) in the draining inguinal LNs of MOG- or OVA-immunized mice (Fig. 6, G and H; and Fig. S6). Collectively, these data establish that signals through β₂ARs act to attenuate T cell–mediated inflammation. Although stimulation of β₂ARs had some impacts on the differentiation of T helper cells in cultures (Sanders, 2012), our results obtained from β₂AR-deficient mice suggest that physiological inputs through β₂ARs might play a minimal role in antigen-induced T cell differentiation in vivo.

As an approach to reveal the role of β₂AR-mediated signals in T cell dynamics in an inflammatory condition, we used the skin DTH model in which trafficking of adoptively transferred antigen-specific T cells could be tracked. We first tested the effect of β₂AR stimulation on T cell recruitment into the ear in the initial phase of inflammation. Naive OT-II T cells expressing OVA-specific TCRs were transferred into recipient mice and then the mice were immunized with OVA in adjuvant. 7 d later, the immunized mice were subjected to continuous administration of clenbuterol through osmotic pumps, followed by OVA challenge in the ear. The number of OT-II T cells recruited into the inflamed ear 12 h after challenge was substantially reduced, but not completely blocked, by treatment with the β₂AR agonist (Fig. 7 A and Fig. S7 A). Concomitantly, reduction of OT-II T cells in blood and lymph was observed (Fig. 7 B and Fig. S7 B), suggesting that β₂AR-mediated inhibition of LN egress of OT-II T cells might be the cause of their impaired recruitment into the ear.

To assess the egress rate of antigen-primed OT-II T cells from LNs, we performed entry blockade with integrin-neutralizing antibodies upon β₂AR stimulation 7 d after immunization. At this time, all transferred OT-II T cells in the draining inguinal LNs exhibited an activated phenotype demonstrated by the up-regulation of CD44 (Fig. S7 C). Administration of clenbuterol dose-dependently enhanced the retention of CD44^hiCD62L^hi central memory OT-II T cells, but not CD44^hiCD62L^lo effector memory OT-II T cells, in the inguinal LNs (Fig. 7 C and Fig. S7C). Consistent with the established phenotypes of central memory T cells (Sallusto et al., 2004), CD44^hiCD62L^hi OT-II T cells had higher expression of CCR7, whereas CXCR4 was more abundantly expressed in CD44^hiCD62L^lo OT-II T cells (Fig. 7 D). The expression of β₂AR transcripts was higher in central memory OT-II T cells than in effector memory cells (Fig. 7 E). The expression pattern of these receptors suggests that higher expression of CCR7 and β₂ARs might contribute to the preferential retention of central memory T cells. These findings suggest that stimulation of β₂ARs in the course of T cell–mediated inflammation selectively inhibits egress of central memory T cells from LNs.

To examine the effect of β₂AR deficiency on the dynamics of antigen-primed T cells, we transferred β₂AR-sufficient or -deficient OT-II T cells (CD45.2⁺) together with WT OT-II T cells (CD45.1⁺CD45.2⁺) as an internal control and immunized the recipient mice (CD45.1⁺) with OVA in adjuvant. Consistent with the results obtained from MOG- or OVA-immunized β₂AR-deficient mice, we found that generation and differentiation of antigen-primed OT-II T cells were not affected by β₂AR deficiency (Fig. 7, F and G; and Fig. S7, D and E). Nevertheless, recruitment of β₂AR-deficient OT-II T cells to the inflamed ear was enhanced compared with β₂AR-sufficient cells (Fig. 7 H and Fig. S7 F). Additionally, β₂AR-deficient OT-II T cells were enriched in blood and lymph of the immunized mice (Fig. 7 I and Fig. S7 G), suggesting that LN egress of antigen-primed OT-II T cells is promoted by β₂AR deficiency. Indeed, β₂AR deficiency resulted in more rapid egress of CD44^hiCD62L^hi central memory OT-II T cells from the draining inguinal LNs relative to co-transferred WT control cells (Fig. 7 J). As expected from the low expression of β₂ARs in CD44^hiCD62L^lo effector memory OT-II T cells (Fig. 7 E), their LN egress was not affected by β₂AR deficiency (Fig. 7 J). Because the expression of CCR7 and CXCR4 was comparable between β₂AR-deficient and WT control OT-II T cells (unpublished data), the impaired LN retention was not attributable to reduced expression of these chemokine receptors. These findings indicate that physiological inputs through β₂ARs expressed on antigen-primed T cells control their dynamics in an inflammatory condition. However, we cannot exclude the contribution of β₂ARs expressed on other cell types, including endothelial cells, in which β₂AR-mediated signals are involved in the expression of adhesion molecules (Scheiermann et al., 2012).

This study demonstrates that activation of β₂ARs in lymphocytes enhances the responsiveness of retention-promoting chemokine receptors to inhibit lymphocyte egress from LNs. Under physiological conditions, the inputs to β₂ARs are provided at least in part by adrenergic nerves. Thus, this study has revealed a cell-intrinsic mechanism by which the nervous system controls lymphocyte recirculation through LNs. This marks a sharp contrast to the role of adrenergic nerves in controlling BM egress of hematopoietic progenitor cells and tissue recruitment of myeloid cells, in which adrenergic inputs affect the expression of adhesion molecules and chemokines in nonhematopoietic compartments of the microenvironments (Méndez-Ferrer et al., 2008; Scheiermann et al., 2012).

However, the sites of interplay between adrenergic nerves and lymphocytes remain to be determined. A previous study demonstrated that substantial levels of noradrenaline were detectable in T cell zones and interfollicular areas but not in B cell follicles in LNs (Felten et al., 1985). Thus, T cells might be constantly exposed to noradrenaline in LNs, whereas B cells might receive the inputs only when they travel to LN exit portals called cortical sinuses, which are located at the interface between T cell zones and B cell follicles (Grigorova et al., 2009; Sinha et al., 2009; Pham et al., 2010). Continuous exposure to noradrenaline would promote desensitization of β₂ARs in T cells, which might be an additional cause for the lower responsiveness of T cells to augmentation or depletion of β₂AR-mediated inputs compared with B cells. To understand how adrenergic nerves control lymphocyte behaviors, it will be important to visualize interactions between adrenergic nerves and lymphocytes using intravital microscopy.

Additionally, the physiological significance of β₂AR-mediated control of lymphocyte trafficking is currently unclear. Studies on circadian regulation of immunity suggested that the innate immune system prepares for higher risk of infection during the active phase, when the activity of adrenergic nerves is high, by up-regulating pathogen sensors (Silver et al., 2012) and recruiting myeloid cells into peripheral tissues (Scheiermann et al., 2012; Nguyen et al., 2013). Our data suggest that β₂AR-mediated signals sequester lymphocytes into lymphoid organs from the circulation during the period of high adrenergic activity. Considering that lymphocytes specific for a given antigen are very rare (20–200 cells/mouse) and the population size dictates magnitude of the response (Moon et al., 2007), sequestration of lymphocytes in lymphoid organs during the active phase may increase the chance of antigen encounter and potentiate adaptive immune responses.

Although stimulation of β₂ARs increased the responsiveness of both CCR7 and CXCR4 in vitro (Fig. 4, A and D), the enhanced LN retention of B and T cells was dependent on distinct chemokine receptors: CCR7 in B cells and CXCR4 in T cells (Fig. 5, A, B, F, and G). In Transwell migration assays, T cells exhibit higher chemotactic responses to the CCR7 ligand than to the CXCR4 ligand, whereas B cells show an opposite trend of responses (unpublished data). Therefore, the effect of β₂AR stimulation on LN retention may be masked by the strong signals evoked by CCR7 in T cells or CXCR4 in B cells, and β₂AR-promoted LN retention may rather reflect signal enhancement of CXCR4 in T cells or CCR7 in B cells. This might explain the observation that retention-promoting effects of β₂AR stimulation were abolished only when functions of the less-dominant chemokine receptors were blocked.

CCR7 deficiency (Förster et al., 1999) and CXCR4 deficiency in the absence of CCR7-mediated signals (Okada et al., 2002) resulted in significant reduction of lymphocyte homing to LNs, establishing the contribution of CCR7 and CXCR4 to lymphocyte entry to LNs. However, lymphocyte homing was unaffected by agonist activation (Fig. 2 C) or genetic ablation (Fig. 3 C) of β₂AR-mediated signals. During lymphocyte entry from blood to LNs, signals through chemokine receptors induce activation of α4 and αL integrins, which is essential for lymphocyte to firmly adhere to the lumen of high endothelial venules (von Andrian and Mempel, 2003). However, lymphocyte egress from LNs does not require the integrins (Lo et al., 2005) and the rate of egress appears to depend on the relative strength of egress-promoting signals and retention-promoting signals (Pham et al., 2008). We speculate that basal levels of CCR7- or CXCR4-mediated signals might be sufficient for optimal integrin activation, allowing efficient lymphocyte homing irrespective of inputs through β₂ARs. This might explain the fact that signal enhancement of CCR7 and CXCR4 by β₂AR activation is selectively reflected in lymphocyte egress from LNs.

Heterodimerization of GPCRs is an emerging concept that accounts for their functional and pharmacological properties (Milligan, 2009; Prezeau et al., 2010). Our data suggest that β₂ARs might selectively form heteromeric complexes with CCR7 or CXCR4, although a definitive demonstration of their interactions in primary lymphocytes has not been achieved. Because β₂AR stimulation preferentially enhances signals through CCR7 and CXCR4 (Fig. 4, A, D and E), the physical association could be key to functional crosstalk between β₂ARs and chemokine receptors. Heteromerization might have allosteric effects on the conformation of chemokine receptors that favor ligand binding or signal transduction, and help converge and synergize their signaling pathways with those of β₂ARs. These possibilities will be tested in future studies to reveal the molecular mechanism of the cross talk between the two classes of GPCRs.

We demonstrated that signals through β₂ARs inhibited LN egress of antigen-primed T cells, reduced their numbers in circulation, and impaired their recruitment to inflamed peripheral tissues (Fig. 7). Although it is possible that these β₂AR-mediated effects on lymphocyte dynamics might contribute to the attenuated T cell–induced inflammation (Fig. 6), the direct causal connection has not been established. Previous studies showed that stimulation of β₂ARs in dendritic cells reduced their capacities for antigen presentation and production of proinflammatory cytokines, including IL-6, IL-12, and IL-23 (Grebe et al., 2009; Hu et al., 2012; Hervé et al., 2013; Nijhuis et al., 2014). Therefore, the alteration of dendritic cell functions might impair reactivation of central memory T cells at sites of inflammation. We speculate that combination of these β₂AR-mediated effects on immune functions might contribute to the suppression of inflammatory responses.

A recent study showed that inputs from adrenergic nerves induced production of a chemokine CCL20 from blood vessels to promote invasion of encephalitogenic T cells into the central nervous system during the onset of EAE (Arima et al., 2012). This effect of adrenergic stimulation appeared to be mediated by β₁ARs because treatment with a selective β₁AR antagonist diminished chemokine production by blood vessels and ameliorated disease. Combined with our results, these findings indicate that stimulation of distinct adrenergic receptors β₁ and β₂ might produce opposite outcomes in EAE. Thus, it would be worth exploring the therapeutic potential of β₂AR stimulation, possibly in combination with β₁AR blockade, in T cell–mediated inflammatory diseases.

C57BL/6 (B6) mice were obtained from Clea. CD45.1⁺ and Ccr7^−/− (Ccr7^tm1Rfor, MGI: 2180679) mice on a B6 background were purchased from The Jackson Laboratory. Adrb2^−/− mice on a B6 background were generated as described previously (Hanyu et al., 2012). OT-II TCR transgenic mice on a B6 background were provided by W.R. Heath (The University of Melbourne, Parkville, Victoria, Australia; Barnden et al., 1998) and crossed to Adrb2^−/− or CD45.1⁺ WT mice to generate Adrb2^−/− or CD45.1⁺CD45.2⁺ WT OT-II mice, respectively. 8–12-wk-old B6 mice were used for experiments. BM chimeras were generated by irradiating recipient mice with a single dose of 8 Gy, followed by i.v. transfer of BM cells. Chimeras were analyzed at least 8 wk after irradiation. Mice were housed in a specific pathogen–free facility, and all experiments were performed in accordance with protocols approved by the Osaka University Animal Care Committee.

Lymphoid tissues were disrupted by passage through a 40-µm cell strainer (BD). For blood samples, red blood cells were lysed with ammonium chloride potassium (ACK) buffer. Single cells were stained with the following fluorophore-labeled antibodies (BioLegend): CD3ε (145-2C11), CD4 (RM4-5), CD8α (53–6.7), CD19 (6D5), CD11b (M1/70), CD21/CD35 (7E9), CD23 (B3B4), CD44 (IM7), CD45 (30-F11), CD45.1 (A20), CD45.2 (104), CD62L (MEL-14), CD93 (AA4.1), CCR7 (4B12), IgD (11-26c.2a), ly 6G (1A8), TCRβ (H57-597), and TCR Vα2 (B20.1). Antibodies against CD95 (Jo2) and CXCR4 (2B11) were purchased from BD. For analysis of intracellular cytokines, cells were fixed and permeabilized with Cytofix/Cytoperm solution (BD) after surface staining and stained with antibodies against IFN-γ (XMG1.2) and IL-17A (TC11-18H10.1; both from BioLegend). Data were acquired on a FACSVerse cytometer (BD) and analyzed with FlowJo software (Tree Star). Cell sorting was performed on a FACSAria II (BD) to isolate the following populations: follicular B cells (CD19⁺CD93⁻CD23^hiCD21^int), CD4⁺ T cells (CD4⁺CD44^loCD62L^hi), and CD8⁺ T cells (CD8⁺CD44^loCD62L^hi) from the spleen; naive OT-II T cells (CD4⁺Vα2⁺ CD44^loCD62L^hi) from the spleen of OT-II mice; and OT-II T cells (CD4⁺Vα2⁺CD45.2⁺) with a central memory (CD44^hiCD62L^hi) or effector memory (CD44^hiCD62L^lo) phenotype from the draining LNs of OVA-immunized mice that had received OT-II T cells.

The MSCV2.2 retroviral vectors expressing Flag-tagged mouse CCR7, CXCR4, or S1PR1, upstream of an internal ribosomal entry site (IRES) and a cytoplasmic domain-truncated human CD4, were provided by J.G. Cyster (University of California, San Francisco, San Francisco, CA; Lo et al., 2005). Mouse CXCR5 was also cloned into this vector. Mouse β₂AR was cloned from splenic cDNA and inserted together with a Myc tag at the amino terminus into the MSCV2.2 vector containing Thy-1.1 reporter downstream of an IRES. Cultures of Phoenix-E packaging cell line were transfected with these constructs, supernatants containing retrovirus were collected and 2PK-3 cells (ATCC) were transduced as described (Lo et al., 2005).

2PK-3 cells were stably transduced with Myc-tagged β₂AR and Flag-tagged CCR7, CXCR4, CXCR5, or S1PR1. The transduced 2PK-3 cells (8 × 10⁶ cells) were washed with PBS and lysed in 0.8 ml lysis buffer (0.875% Brij 97, 0.125% Nonidet P-40, 150 mM NaCl, and 10 mM Tris-HCl, pH 7.2) containing protease inhibitors (Nacalai Tesque). Supernatants were precleared before incubation with Flag M2 beads (Sigma-Aldrich) for 12 h at 4°C. Beads were washed with lysis buffer containing 0.5 M NaCl. Samples were eluted and reduced before separation in NuPAGE 4–12% Bis-Tris gels (Life Technologies) and transferred to Immobilon-P^SQ membranes (EMD Millipore). Membranes were blocked with 4% skim milk in TBS containing 0.1% Tween 20 and blotted with mouse monoclonal anti-Flag (M2; Sigma-Aldrich) or rabbit polyclonal anti-Myc (Cell Signaling Technology), followed by goat anti–mouse IgG or donkey anti–rabbit IgG conjugated with horseradish peroxidase (HRP; both from Jackson ImmunoResearch Laboratories). Activation of Rac1 was analyzed using an assay kit from Millipore. In brief, 10⁷ B or T cells, isolated from mouse spleens by negative selection using AutoMACS (Miltenyi Biotec), were stimulated with chemoattractants in the presence or absence of 10 µM clenbuterol (Sigma-Aldrich). Cell extracts were incubated with a Rac-binding domain of PAK1 bound to agarose beads for 1 h at 4°C. Reduced eluates were resolved in NuPAGE 10% Bis-Tris gels (Life Technologies) and blotted with mouse monoclonal anti-Rac1 (23A8), followed by HRP-conjugated goat anti–mouse IgG (Jackson ImmunoResearch Laboratories). Images were acquired with a luminescent image analyzer (ImageQuant LAS4000 mini; GE Healthcare), and band density was analyzed with ImageJ software (National Institutes of Health).

For single-cell analysis of 2PK-3 cells expressing Myc-tagged β₂ARs together with Flag-tagged CCR7, CXCR4, or S1PR1, cells were stained with Alexa Fluor 647–conjugated anti-Myc (9B11; Cell Signaling Technology) and biotin-labeled antibodies against CCR7 (4B12) or CXCR4 (2B11), followed by PE-labeled streptavidin (BioLegend). S1PR1 was detected with a purified rat antibody (713412; R&D Systems), followed by biotin-labeled anti–rat IgG (Jackson ImmunoResearch Laboratories) and PE-labeled streptavidin (Arnon et al., 2011). Labeled cells were seeded onto slides and fixed with 4% paraformaldehyde. Slides were mounted with FluoSave Reagent (EMD Millipore). Images were acquired at room temperature with Fluoview FV10-ASW version 3.00 using a Fluoview FV1000 inverted confocal microscope (Olympus), equipped with plan apochromat UPLSAPO objectives (100× oil immersion with 1.40 NA). Co-localization of fluorescent signals was analyzed with a plug-in ‘RG2B Colocalization’ for ImageJ software.

Total RNA was extracted with a Nucleospin RNA kit (Macherey-Nagel) from FACS-sorted cells. Complementary DNA was synthesized with a ReverTra Ace qPCR RT kit (Toyobo). Quantitative PCR was performed with SYBR Premix Ex Taq II on a Thermal Cycler Dice Real Time System II (Takara). The following primers were used to detect Adrb2 transcripts: 5′-GTACCGTGCCACCCACAAGA-3′ (forward) and 5′-CCCGGGAATAGACAAAGACCATC-3′ (reverse).

Spleen cells from WT and Adrb2^−/− mice were treated with ACK buffer to lyse red blood cells, washed five times, and incubated for 30 min at 37°C in RPMI containing 0.5% fatty acid-free BSA (EMD Millipore). Cells were put into the upper chambers of Transwells (Costar) and tested for transmigration across filters with 5-µm pores for 3 h at 37°C in response to chemoattractants in the lower chambers: 100 ng/ml CCL21 (R&D Systems), 10 ng/ml CXCL12 (R&D Systems), 300 ng/ml CXCL13 (R&D Systems), or 100 nM S1P (Sigma-Aldrich). Clenbuterol was added to the lower chambers with the chemoattractants. CD4⁺ or CD8⁺ T cells and B cells that migrated to the lower chambers were enumerated by flow cytometry and chemotactic responses were determined as percentages of their numbers relative to those of input cells.

After mice were euthanized with CO₂, lymph was drawn from the cysterna chyli using a fine microcapillary pipette (Matloubian et al., 2004), and then blood was collected from the inferior vena cava using a syringe with a 26-gauge needle. To determine the cell concentration in lymph, cell numbers counted by flow cytometry were divided by the volume of collected lymph. Except for the time course analysis of lymphocyte numbers after treatment with β₂AR agonists, blood and lymph were collected between 2 and 4 h after the onset of light to avoid circadian variation of lymphocyte counts.

Lymphocyte entry to LNs was blocked by i.v. injection of neutralizing antibodies against integrin αL (M17/4) and α4 (PS/2, both from Bio X Cell) at 100 µg each per mouse and, 14–22 h later, lymphocytes remaining in the LNs were enumerated (Lo et al., 2005). Mini osmotic pumps (2001D; Alzet) containing saline, AMD3100 (40 mg/kg/d; Sigma-Aldrich) and/or various doses of clenbuterol or salbutamol were surgically implanted s.c. in the back of mice to continuously administer the drugs up to 24 h. Entry blockade was performed 1 h after implantation of the pumps. For depletion of adrenergic nerves, mice were injected i.p. with 100 mg/kg 6-OHDA (Sigma-Aldrich) dissolved in saline containing 0.01% antioxidant ascorbate on days −7 and −5 and 200 mg/kg on day −3 (Grebe et al., 2010). For short-term transfers, mutant or matched control spleen cells were labeled with 0.5 µM CFSE in DMEM containing 1% FBS for 10 min at 37°C. Each recipient mouse received 2 × 10⁷ WT or Adrb2^−/− cells, or 4 × 10⁷ Ccr7^−/− cells to overcome a defect in LN entry caused by CCR7 deficiency (Förster et al., 1999). After 2 d of equilibration, entry blockade was performed.

To test the effect of β₂AR stimulation, 3 × 10⁷ CD45.1⁺ WT spleen cells were transferred to CD45.2⁺ mice that had received i.v. injection of saline, clenbuterol (0.4 mg/kg), or salbutamol (12 mg/kg) immediately before cell transfer. To assess the effect of β₂AR deficiency, CD45.2⁺ Adrb2^+/− or Adrb2^−/− and CD45.1⁺ WT spleen cells were mixed at a ratio of 50:50, labeled with 0.1 µM CFSE in DMEM containing 1% FBS for 15 min at 37°C and transferred to CD45.2⁺ mice (4 × 10⁷ cells per mouse). 90 min after cell transfer, LNs were collected and subjected to flow cytometry.

On day 0, mice were immunized s.c. with 100 µg MOG_35-55 peptide (BEX) emulsified in CFA (Sigma-Aldrich) at the base of the tail. The mice received i.v. 100 ng pertussis toxin (List Laboratories) on days 0 and 2. The clinical symptoms were scored daily as follows: 1, limp tail or abnormal gait; 2, partial hind limb paralysis; 3, complete hind limb paralysis; 4, forelimb weakness; 5, death. For analysis of effector T cell generation, the inguinal LNs were collected on day 14 and disrupted by passage through a 40-µm cell strainer (BD). Cells were stimulated with 50 ng/ml phorbol 12-myristate 13-acetate (Sigma-Aldrich) plus 1 µM ionomycin (Sigma-Aldrich) in the presence of brefeldin A (Sigma-Aldrich) for 6 h and stained for intracellular cytokines as described above. Because initial signs of EAE were reproducibly observed on day 7 or 8 in our particular set of experiments, treatment with clenbuterol (0.4 mg/kg/d, i.v.) was started on day 6, aiming at inhibiting entry of pathogenic T cells into the central nervous system in the early phase of disease.

Mice were immunized s.c. with 100 µg OVA emulsified in CFA at the base of the tail. 7 d after immunization, ears were injected i.d. with saline or 20 µg OVA. Mini osmotic pumps (1003D, Alzet) containing saline or clenbuterol (40 mg/kg/day) were implanted 1 h before OVA challenge into the ear and left for 3 d. Thickness of ear was measured with a spring-loaded caliper (Mitutoyo). Net ear swelling was calculated by subtracting the increased thickness of the saline-injected ear from that of the OVA-injected ear. For analysis of effector T cell generation, single cells prepared from the inguinal LNs 7 d after immunization were stimulated and stained for intracellular cytokines as described above.

To analyze the dynamics of OT-II T cells in the DTH model, 2 × 10⁵ CD45.2⁺ WT OT-II T cells or mixtures of CD45.2⁺ Adrb2^+/− or Adrb2^−/− plus CD45.1⁺CD45.2⁺ WT OT-II cells (2 × 10⁵ cells of each) were transferred i.v. into CD45.1⁺ mice and then the mice were immunized with 100 µg OVA in CFA the next day. In some experiments, clenbuterol was administered through mini osmotic pumps (2001D) 7 d after immunization. For analysis of LN egress, entry blockade was performed 7 d after immunization. The draining inguinal LNs were collected after the indicated time and subjected to flow cytometry for detection of central memory and effector memory OTII-T cells. For analysis of T cell recruitment to inflammatory sites, ears were injected i.d. with saline or 20 µg OVA 7 d after immunization, and collected 12 h later. The dorsal half of the ear was treated with 1 U/ml Dispase I (Roche) in RPMI containing 1% FBS for 1 h at 37°C to remove the epidermis. The dermis was cut into pieces and digested with 0.3 mg/ml Liberase TL (Roche) and 0.05% DNase I (Roche) in RPMI containing 1% FBS for 1.5 h at 37°C. The tissue was disrupted by passage through a 70-µm cell strainer (BD) and subjected to flow cytometry for detection of OT-II T cells.

GraphPad Prism (GraphPad Software) was used for all statistical analyses. For comparison of two nonparametric datasets, Mann-Whitney U test was used. Means of two groups were compared with unpaired Student’s t test. For multigroup comparisons, we applied one-way ANOVA with post hoc testing using Bonferroni’s multiple comparison test. The ED₅₀ values were obtained from nonlinear regression analysis of the curves.

Figs. S1–S7 show representative flow cytometry plots and gating strategies for the data shown in Figs. 1–7, respectively.

We thank J.G. Cyster for reagents, T. Okada for technical assistance and discussion, H. Kikutani and A. Kumanogoh for discussion, Y. Ezura and S. Sakakibara for supporting mouse experiments, T. Sugimoto for technical support, and RIMD core facilities for cell sorting.

K. Suzuki is supported by the Japan Science and Technology Agency (PRESTO), the Japan Society for the Promotion of Science (23890097 and 24790470), the Naito Foundation, the Senri Life Science Foundation, the Daiichi-Sankyo Foundation of Life Science, the Sumitomo Foundation, and the Kowa Life Science Foundation.

The authors have no conflicting financial interests.