Asymmetrical ligand-induced cross-regulation of chemokine (C-X-C motif) receptor 4 by α1-adrenergic receptors at the heteromeric receptor complex

Recently, we reported that chemokine (C-X-C motif) receptor (CXCR)4 and atypical chemokine receptor 3 regulate α1-adrenergic receptors (α1-AR) through the formation of hetero-oligomeric complexes. Whether α1-ARs also regulate chemokine receptor function within such heteromeric receptor complexes is unknown. We observed that activation of α1b-AR within the α1b-AR:CXCR4 heteromeric complex leads to cross-recruitment of β-arrestin2 to CXCR4, which could not be inhibited with AMD3100. Activation of CXCR4 did not cross-recruit β-arrestin2 to α1b-AR. A peptide analogue of transmembrane domain 2 of CXCR4 interfered with α1b-AR:CXCR4 heteromerization and inhibited α1b-AR-mediated β-arrestin2 cross-recruitment. Phenylephrine (PE) induced internalization of CXCR4 in HEK293 cells co-expressing CXCR4 and α1b-AR and of endogenous CXCR4 in human vascular smooth muscle cells (hVSMC). The latter was detectable despite blockade of CXCR4 with the neutralizing antibody 12G5. hVSMC migrated towards CXCL12 and PE, but not towards a combination of CXCL12 and PE. PE inhibited CXCL12-induced chemotaxis of hVSMC (IC50: 77 ± 30 nM). Phentolamine cross-inhibited CXCL12-induced chemotaxis of hVSMC, whereas AMD3100 did not cross-inhibit PE-induced chemotaxis. These data provide evidence for asymmetrical cross-regulation of CXCR4 by α1-adrenergic receptors within the heteromeric receptor complex. Our findings provide mechanistic insights into the function of α1-AR:CXCR4 heteromers and suggest alternative approaches to modulate CXCR4 in disease conditions.

To further consolidate that agonist stimulation of α 1b -AR leads to cross-recruitment of β-arrestin 2 to CXCR4, we then tested the effects of the selective α 1 -AR antagonist phentolamine and of the selective CXCR4 antagonist AMD3100 on PE-induced β-arrestin 2 cross-recruitment to CXCR4-Tango. AMD3100 inhibited CXCL12 induced β-arrestin 2 recruitment to CXCR4-Tango (Fig. 4a). While AMD3100 did not inhibit PE-induced β-arrestin 2 recruitment to CXCR4-Tango in cells co-expressing CXCR4-Tango and α 1b -AR, phentolamine inhibited PE-induced β-arrestin 2 recruitment under these conditions (Fig. 4b). Measurements of cell viability demonstrated that phentolamine was not cytotoxic at the concentration that was used in the β-arrestin 2 recruitment assays (Fig. 4c).
Recombinant CXCR4 internalizes upon agonist binding to α 1b -AR. Because β-arrestin recruitment to CXCR4 upon CXCL12 binding leads to internalization of the receptor, we then studied by immunofluorescence microscopy whether activation of α 1b -AR would also result in the internalization of CXCR4 in cells co-expressing  both recombinant receptors. After transfection with HA-CXCR4 plus pcDNA3 or α 1b -AR, HEK293T cells were labeled with FITC-conjugated anti-HA at 4 °C, followed by stimulation with vehicle, CXCL12 or PE at room temperature. In cells co-expressing HA-CXCR4 and pcDNA3 (Fig. 6a), the fluorescence signal for FITC-labeled anti-HA was localized to the cell membrane after vehicle and PE treatment (left and right). After treatment with CXCL12, punctate areas of fluorescence were detectable, indicating internalization of HA-CXCR4 (center). In cells co-expressing HA-CXCR4 and α 1b -AR (Fig. 6b), CXCL12 and PE treatment resulted in punctate areas of fluorescence, suggesting that HA-CXCR4 internalizes upon activation of α 1b -AR. Measurements of cell surface HA-CXCR4 expression levels by flow cytometry confirmed these observations and showed that PE did not affect HA-CXCR4 expression levels in cells co-expressing HA-CXCR4/pcDNA3 (Fig. 6c), but reduced HA-CXCR4 expression levels in cells co-expressing HA-CXCR4/α 1b -AR (Fig. 6d).

CXCR4 in human vascular smooth muscle cells internalizes upon agonist binding to α 1 -AR.
To assess whether PE stimulation also induces internalization of endogenous CXCR4, we then incubated hVSMC with rabbit anti-CXCR4 ACR-014 at 4 °C, followed by incubation with vehicle, CXCL12 or PE at 37 °C. Cells were then fixed, permeabilized and incubated with Alexa488-conjugated goat anti-rabbit IgG. Fluorescence microscopy showed a diffuse pattern of fluorescence in vehicle treated cells (Fig. 7a, left column, top). In cells treated with CXCL12 and PE (Fig. 7a, left column, center and bottom), bright fluorescent punctate areas were visible. These observations are consistent with CXCR4 internalization and with the staining patterns that have previously been observed in similar fluorescence microscopy internalization studies of other GPCRs 19 . When these experiments were performed with FITC-labeled anti-CXCR4 12G5, an antibody known to inhibit CXCL12 binding to CXCR4 20 , fluorescent punctate formation did not occur after CXCL12 treatment, but persisted after PE treatment (Fig. 7a, right column). These findings are in line with our observation that blockade of CXCR4 with AMD3100 did not prevent β-arrestin 2 cross-recruitment upon α 1b -AR activation (Fig. 4b). To confirm that PE treatment leads to internalization of CXCR4 in hVSMC, we then performed PLA to analyze expression levels of CXCR4, α 1B -AR and CXCR4:α 1B -AR heteromers after vehicle and PE treatment (Fig. 7b). The quantification of the PLA signals per cell from three independent experiments is shown in Fig. 7c. As compared with vehicle treated cells, PLA signals corresponding to CXCR4, α 1B -AR and heteromeric CXCR4:α 1B -AR complexes were significantly reduced after PE treatment. Together, our observations suggest that activation of α 1B -AR within the α 1B -AR:CXCR4 heteromer recruits β-arrestin to both receptor partners, leading to co-internalization of α 1B -AR and CXCR4. Furthermore, our findings that AMD3100 and anti-CXCR4 12G5 did not prevent β-arrestin 2 recruitment and internalization of CXCR4 upon PE stimulation indicate that this receptor cross-talk can occur when CXCR4 is in its ligand-free and in an antagonist-bound configuration. This assumption is supported by the recent observation that chemically-induced fusion of CXCR4 to β-arrestin 1/2 leads to internalization of the receptor, independent of agonist binding to and G protein coupling through CXCR4 21 .
Ligand binding to α 1 -AR inhibits CXCR4-mediated chemotaxis of human vascular smooth muscle cells. To assess consequences of ligand binding to α 1 -AR on CXCR4-mediated effects on cell function, we then studied chemotactic responses of hVSMC in trans-well migration assays. We reported previously that the dose-response profile of CXCL12 shows maximal chemotactic activity at concentrations between 10-100 nM in hVSMC under our experimental conditions 22 . As expected 23,24 , PE also dose-dependently induced chemotaxis in hVSMC under these conditions (Fig. 7d). Maximal chemotactic activity of PE was comparable with the chemotactic activity of CXCL12. When hVSMC were loaded onto the membrane in the presence of increasing concentrations of PE, migration towards 10 nM CXCL12 was significantly reduced. At a concentration of 10 nM PE, chemotaxis towards 10 nM CXCL12 was reduced by more than 50% and further increases in PE concentrations dose-dependently reduced the chemotactic index (CI) to 20% of the CI in the absence of PE (dark grey bars). When hVSMC migrated towards 10 nM CXCL12 plus increasing concentrations of PE, PE inhibited chemotaxis with an IC 50 of 77 ± 30 nM (Fig. 7d, dark grey bars and inlay). These findings are in agreement with previous observations on the pharmacological behaviors of the CXCR4:δ-opioid receptor and CXCR4:cannabinoid receptor 2 heteromers, in which agonist co-stimulation of both receptor partners abolished chemotaxis towards CXCL12 13,14 .
As pharmacological cross-inhibition has previously been described for other GPCR heteromers, we then tested the effects of AMD3100 and phentolamine on the chemotactic responses of hVSMC to CXCL12, PE and a combination of both agonists. As in Fig. 7d, hVSMC migrated towards PE and CXCL12 and this response was inhibited when cells migrated towards both agonists (Fig. 7e). Phentolamine inhibited PE-and CXCL12-induced chemotaxis, but did not alter the chemotactic response to both agonists. In contrast, AMD3100 inhibited CXCL12-induced chemotaxis, but did not significantly affect PE-induced chemotaxis. AMD3100, however, restored a chemotactic response of hVSMC to the combination of PE and CXCL12. In combination with the effects of PE on β-arrestin 2 cross-recruitment to and co-internalization of CXCR4, these findings indicate asymmetrical cross-activation and cross-inhibition of CXCR4 within the heteromeric receptor complex by α 1 -AR ligands. Such a pharmacological behavior of the CXCR4:α 1 -AR heteromeric complex is similar to the signaling behavior of other GPCR heteromers, for which both ligand-induced symmetrical and asymmetrical cross-activation and cross-inhibition of various signaling read-outs have previously been described [25][26][27][28] . For example, symmetrical ligand induced cross-inhibition and cross-sensitization was observed for the angiotensin II receptor type 1:prostaglandin F2α receptor heteromer when aortic ring contraction was used as a functional read-out, whereas the effects of agonist stimulation on [ 3 H]thymidine incorporation in vascular smooth muscle cells showed asymmetrical cross-inhibition, i.e. a prostaglandin F2α receptor antagonist inhibited angiotensin II-induced effects but an angiotensin II receptor type 1 antagonist did not affect prostaglandin F2α induced-effects. Similarly, agonist-induced cross-phosphorylation and cross-internalization was observed at the NK1 receptor:μ-opioid receptor heteromer 25 It has been proposed previously that simultaneous agonist binding to both receptor partners can inhibit signaling of heteromeric receptor complexes, such as the CXCR4:cannabinoid receptor 2, CXCR4:δ-opioid receptor or μ-opioid receptor:CCR5 heteromers 13,14,29 . As the results from the present study point towards β-arrestin 2 cross-recruitment to AMD3100-bound CXCR4 upon α 1 -AR activation, a possible explanation for the observation that AMD3100 restored chemotaxis of hVSMC towards both agonists is that coupling of the heteromeric receptor complex to both Gα i and Gα q upon ligand binding is required to prevent PE-induced chemotaxis. The finding that phentolamine did not restore chemotaxis of hVSMC towards both agonists is not contradictive because phentolamine cross-inhibited chemotactic responses of hVSMC towards CXCL12 alone.
In conclusion, our findings provide new mechanistic insights into molecular signaling events at the CXCR4:α 1b -AR hetero-oligomeric complex and evidence that α 1 -ARs regulate CXCR4-mediated functions in hVSMC. Our findings of asymmetrical agonist-and antagonist-induced cross-regulation of CXCR4 by α 1 -AR within the heteromeric receptor complex have several implications. While these data provide a mechanistic basis for interactions between the neuroendocrine and immune system, our findings also point towards α 1 -AR as new pharmacological target to modulate CXCR4 expression levels and CXCR4-mediated effects on cell migration in various disease processes, such as cancer metastasis or autoimmune diseases. These data further support the notion that the development of drugs which selectively target heteromeric receptor complexes could have novel and unique pharmacological properties and also imply that drugs with high selectivity for a specific GPCR can have unforeseen pharmacological effects through its actions on heteromerization receptor partners. Thus, the detailed elucidation of the molecular properties of heteromeric receptor complexes will provide new opportunities for drug development and may also help to better anticipate unwanted side-effect profiles of drugs targeting GPCRs.
Proximity Ligation Assays (PLA). PLA were performed as described in detail previously 15,16 . To visualize individual receptors, slides were incubated with mouse anti-FLAG M2 (Sigma-Aldrich, F1804), rabbit anti-HA (Abcam, Ab9110), rabbit anti-α 1B -AR (Abcam Ab169523) or goat anti-CXCR4 (Abcam Ab1670) at 37 °C for 105 min in a humidifying chamber at 37 °C. Anti-α 1B -AR and anti-CXCR4 have been validated for sufficient selectivity previously 16,31 . Slides were then washed and incubated (1 h, 37°C) with secondary species specific (rabbit/mouse/goat) antibodies conjugated with plus and minus Duolink II PLA probes (1:5), as appropriate. To visualize receptor-receptor interactions, slides were incubated with a combination of rabbit anti-α 1B -AR (Abcam Ab169523) and goat anti-CXCR4 (Abcam Ab1670) or a combination of mouse anti-FLAG M2 and rabbit anti-HA. All antibodies, except anti-HA (dilution 1:1000), were used in dilutions of 1:500. Slides were then washed and incubated with secondary species specific antibodies conjugated with plus and minus Duolink II PLA probes (1:5) corresponding to the two primary antibodies. PLA signals were quantified using the Duolink Image Tool software (Sigma Aldrich). Comparisons and statistical analyses were performed only when PLA assays were performed on the same day in parallel experiments and fluorescence microscopy was performed with the identical settings. For each experiment and condition, 10 randomly selected non-overlapping vision fields were analyzed. Plasmids and transfections. FLAG-tagged Tango plasmids (CXCR4-TANGO, #66262; α 1b -AR-TANGO, #66214; α 1d -TANGO, #66215) were from Addgene deposited by the laboratory of Dr. Bryan Roth. Myc-tagged TANGO plasmids were generated by replacing the original FLAG-tag plasmids with Myc tag after PCR amplification with primers carrying the Myc tag sequence. HA-or FLAG-tagged CXCR4, α 1b -AR or α 1d -AR were generated by PCR amplification using corresponding TANGO plasmids as cDNAs with primers carrying Xho I and Xba I sites and inserted in pcDNA3 with an N-terminal FLAG or HA tag. All plasmids were verified by sequencing. HEK293T cells were transiently transfected with a single DNA or two DNAs that encode HA-or FLAG-tagged CXCR4, α 1b -AR or α 1d -AR or TANGO plasmids at the amounts indicated, using Lipofectamine 3000 (Thermo Scientific) as per manufacturer's protocol. For single GPCR-encoding DNA transfection, empty vector pcDNA3 was added to maintain the total DNA amount per transfection constant.
PRESTO-Tango β-arrestin recruitment assay. The PRESTO-Tango (parallel receptorome expression and screening via transcriptional output, with transcriptional activation following arrestin translocation) assay was performed as recently described 18 . HTLA cells (2.5 × 10 5 /well) were seeded in a 6-well plate and transfected with 750 ng of each of the Tango plasmids using Lipofectamine 3000 (Thermo Scientific). The following day, transfected HTLA cells (75,000 cells/well) were plated onto Poly-L-Lysine pre-coated 96-well microplates and allowed to attach to the plate surface for at least 4 hours prior to treatment. Cells were treated with receptor agonists for 2 h and then replaced with fresh full medium and incubated overnight at 37 °C, 5% CO 2 in a humidified environment. To test the effects of TM peptides (50 μM), AMD3100 (10 μM) or phentolamine (100 μM), cells were pre-incubated with these reagents for 15 min at 37 °C before adding agonists. The following morning, medium was removed from cell culture plates and replaced with a 100 µL 1:5 mixture of Bright-Glo (Promega) and 1× HBSS, 20 mM HEPES solution. Plates were then incubated at room temperature for 20 min before measuring luminescence on a Biotek Synergy II plate reader. Relative luminescence (RLU) is expressed as % of luminescence upon stimulation with the highest concentration of the corresponding receptor agonist when the Tango receptor was co-expressed with pcDNA3 (=100%). When HTLA cells were not transfected with a Tango plasmid, no change in luminescence was detectable upon agonist treatment ( Supplementary Fig. S3).
Immunofluorescence receptor internalization assays. Receptor internalization assays were performed with HEK293T cells and hVSMC. HEK293T cells were transfected with HA-CXCR4 plus pcDNA3 or FLAGα 1b -AR at 375 ng each DNA/well of a 12-well plate. After 24 hours, cells were re-plated in poly D-lysine coated 8-well chamber slides. After overnight incubation, cells were chilled on ice and labeled with FITC-conjugated anti-HA (Sigma-Aldrich, H7411, 2 μg/ml in Opti-MEM) for 1 h at 4 °C. Subsequently, labeled cells were washed with Opti-MEM twice and then incubated with CXCL12 (200 nM) or PE (200 μM) for 1 h at room temperature. Thereafter, cells were washed with PBS and fixed with 4% paraformaldehyde. hVSMC were plated in 8-well chamber slides. The next day, slides were cooled on ice, washed once with Opti-MEM and then incubated with FITC-conjugated anti-CXCR4 antibody (R&D, FAB170F, 1:150 dilution in Opti-MEM) or rabbit anti-CXCR4 antibody (Alomone, ACR-014, 1:150 in Opti-MEM) at 4 °C for 1 h. Subsequently, slides were washed twice and then treated with CXCL12 (100 nM) or PE (10 μM) in Opti-MEM at 37 °C for 20 min. Cells were fixed with 4% paraformaldehyde/PBS at room temperature for 10 min. After three washes with PBS, cells incubated with non-conjugated anti-CXCR4 were permeabilized with 0.2% Triton X-100/PBS at room temperature for 10 min and then blocked with 1% BSA/PBS for 30 min, followed by incubation with Alexa 488-conjugated goat anti-rabbit antibody (1:250) at room temperature for 1 h. HEK293T cells and hVSMC Cells were mounted with mounting medium for fluorescence microscopy. Images were obtained under a florescence microscope (Carl Zeiss Axiovert 200 M) with a 40x objective lens equipped with Axio CamMRc5. For each well, 10 vision fields were imaged. Chemotaxis assays. Cell migration was assessed using the ChemoTx 96-well cell migration system, as described 17,22,32 . The chemotactic index (CI) was calculated as the ratio of cells that transmigrated through the filter in the presence versus the absence (=PBS/control) of the test solutions.
Cell viability assays. Cell viability was assessed using the Trypan blue exclusion method. HTLA cells were incubated with various concentrations of phentolamine under the same conditions used in Presto-Tango assays. After staining cells with Trypan blue, at least 100 cells per condition and experiment were examined in a hemocytometer and the percentage of viable cells quantified.
Data analyses. Data are expressed as mean ± standard error of the mean (SEM) from n independent experiments that were performed on different days. Data were analyzed using the GraphPad-Prism 7 software. Unpaired Student's t test or one-way analyses of variance (ANOVA) with Dunnett's multiple comparison post-hoc test for multiple comparisons were used, as appropriate. Dose response curves were generated using non-linear regression analyses. A two-tailed p < 0.05 was considered significant. Data availability. The datasets generated during the current study are available from the corresponding author on reasonable request.