Arrestin 3 Mediates Endocytosis of CCR7 following Ligation of CCL19 but Not CCL21¹

G protein-coupled receptors (GPCRs)³ are a diverse group of seven transmembrane receptors that are activated by multiple sensory and chemical stimuli including light, odors, tastes, chemokines, and neurotransmitters (as reviewed by Ref. 1). Activation of these receptors by their ligands induces a conformational change, which leads to activation of distinct downstream signaling events (2, 3, 4). Following activation by agonist binding, the GPCRs couple to heterotrimeric GTP-binding (G) proteins triggering an exchange of GDP for GTP on the Gα-subunit and the resultant release of the α subunit from the β/γ-dimer (5). These subunits can regulate numerous effectors including adenylyl cyclase and phospholipase Cβ (6, 7). The receptor is then phosphorylated at its C terminus, which provides a substrate for high affinity binding by arrestin 2 (also known as β-arrestin 1) and/or arrestin 3 (also known as β-arrestin 2), and the receptor is then internalized. Two of the four known arrestins, arrestin 2 and arrestin 3, are expressed ubiquitously in mammalian cells (8, 9, 10). Receptor internalization is an important immunological function that controls signaling, which controls immunological responses, receptor fate, and receptor desensitization (11). Yet it is unclear how distinct receptors are shuttled to the necessary internalization machinery to differentially control the signaling and fate of each internalized receptor.

Arrestins can regulate GPCR internalization, degradation and/or recycling. Arrestin binding can block association of G proteins with phosphorylated GPCRs and thus plays an important role in phosphorylation-dependent desensitization (9, 12). For instance, arrestins mediate internalization of the β2 adrenergic receptor (10, 13). Arrestins can also serve as adaptors between the receptors and the endocytic machinery. In such cases, arrestin binding targets activated receptors to clathrin-coated pits via interactions with clathrin, the clathrin adaptor AP-2, and the intracellular transport ATPase N-ethylmaleimide-sensitive factor (14, 15, 16). Arrestins can also regulate other signaling pathways including activating Src, ERK, and c-Jun N-terminal kinase (17, 18, 19). Alternatively, while the human N-formyl peptide receptor (FPR) does not require arrestins to internalize, arrestins are required for FPR recycling (20). In addition, arrestins can mediate trafficking of GPCRs to specific sites in the cell, where the receptors are degraded (8, 21, 22).

CCR7 is a GPCR that is expressed on B lymphocytes, NK cells, and naive and central memory T lymphocytes (23, 24, 25, 26, 27). CCR7 is activated by two known endogenous ligands, CCL19 (MIP-3β, ELC, CKβ-11, Exodus-3) and CCL21 (6Ckine, SLC, TCA4, Exodus-2) (26, 28, 29, 30, 31). CCL21 is expressed on the luminal side of high endothelial venules, while both ligands are found within the stromal region of the T cell-rich lymph node areas (32). The CCL19 ligand is expressed on dendritic cells, and is found at lower levels in the spleen and lymph nodes than CCL21 (33). The expression of these ligands in different regions of the immune system implies that each ligand might have a unique role in the regulation of this system and as a result differentially regulate signaling in T cells. Indeed, each ligand has been uniquely associated with different autoimmune diseases. Although CCL21 has been associated with Sjogren’s syndrome and lymphoproliferative disorders (34, 35, 36), CCL19 has been linked to Crohn’s (37, 38). It is unclear, however, how each of these ligands mediates distinct effects following association with CCR7.

CCR7 binds CCL19 and CCL21 with equal affinities (30), yet each ligand has distinct signaling pathways and modulation of cell migration. In one study, ectopic expression of CCL19 in pancreatic islets resulted in the development of small infiltrates of lymphocytes and dendritic cells, whereas ectopic expression of CCL21 resulted in larger, more organized infiltrates. From these observations it was inferred that CCL21 might be more effective than CCL19 in promoting the recruitment of T cells (33). It was surprising, therefore, when it was reported that only CCL19 mediates emigration of newly generated lymphocytes from the neonatal thymus (39). These observations provide further support that distinct signal cascades are initiated following ligand/receptor association, which control different cellular behaviors. The signal transduction events triggered by GPCRs are regulated both spatially and temporally (4).

To begin to unravel the differential regulation of the signaling events between CCL19 and CCL21, we examined the internalization of CCR7 when bound to CCL19 and CCL21. Receptor internalization can initiate signaling, processing and often desensitization of a GPCR. After we observed differences in receptor internalization following engagement of CCL19 and CCL21, we examined the role of arrestins in regulating the differential internalization of CCR7 following ligand binding. To this end, we examined CCR7 internalization following depletion of arrestin 2 or arrestin 3 using small interfering RNA (siRNA) in HuT 78 human T cells. In addition, we examined the effect of loss of each arrestin on migration of HuT 78 cells. Next, we examined internalization of CCR7 in murine embryonic fibroblasts (MEFs) that are null for arrestin expression. We report that arrestin 3 but not arrestin 2 is required for internalization of CCR7 when bound to CCL19. In contrast, arrestins are not required for internalization of CCR7/CCL21. In addition, both arrestins control migration of HuT 78 cells. Although signaling is typically associated with receptor clustering, we observed that arrestins do not appear to cluster at the membrane with CCR7 following ligand binding. Instead, CCR7 forms caps following ligand association. This is the first description of differential binding of arrestins that is regulated by two distinct endogenous chemokines, which bind to the same receptor.

Recombinant human and murine CCL19 and CCL21, and PE-conjugated anti-human CCR7 Ab (FAB197P), were purchased from R&D Systems. M2 Abs (Catalog no. F1804) were purchased from Sigma-Aldrich. Anti-β-arrestin 1/2/3 was purchased from Santa Cruz Biotechnology. PE-conjugated anti-mouse was purchased from Jackson ImmunoResearch Laboratories. The cDNA encoding N-terminally tagged murine CCR7 was a generous gift from James J. Campbell (Harvard Medical School, Boston, MA). Plasmids encoding bovine arrestin2-GFP and arrestin3-GFP have been described (20) and were generously provided by Dr. Jeffrey Benovic (Thomas Jefferson University, Philadelphia, PA). The plasmid encoding the human FPR has been described (40) and was generously provided by Dr. Eric Prossnitz (University of New Mexico Medical School, Albuquerque, NM).

HuT 78 and Jurkat e.6 cell lines were purchased from the American Type Culture Collection, and were maintained in cRPMI 1640 (10% FBS/90% RPMI 1640/2 mM l-glutamine) (Invitrogen) in a humidified atmosphere of 5% CO₂/air at 37°C. The arrestin2^−/−/arrestin3^−/− MEFs (WT and arr2^−/−/arr3^−/−) cell lines were a generous gift from Dr. Robert Lefkowitz (Duke University, Durham, NC). Murine CCR7 was transiently transfected into the MEF arr2^−/−/arr3^−/− cell lines with Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions, with the following modification: we halved all recommended amounts. pCDNA3.1 (Invitrogen) vector was transfected as the vector control. MEF cell lines were maintained in 10% FBS (HyClone)/90% DMEM supplemented with 2 mM l-glutamine at 37°C/5% CO₂ in a humidified incubator. Cells were certified to be free of mycoplasma by Charles River Laboratories.

A total of 10⁶ HuT 78 cells were isolated from culture, rinsed twice with serum free (sf)RPMI 1640 and resuspended directly in 10 μl of PE directly conjugated anti-human CCR7 Ab or the PE directly conjugated isotype control (R&D Systems). Cells were allowed to incubate on ice for 30 min, rinsed three times with sfRPMI 1640, and analyzed immediately by flow cytometry on a LSRII (BD Biosciences).

A total of 10⁶ HuT 78 or Jurkat e.6 cells were isolated from culture, rinsed twice with sfRPMI 1640, and total RNA was isolated as per the manufacturer’s instructions (Qiagen) and quantified. Then, 5 μg of RNA was mixed with 50 ng of random hexamers (Invitrogen), 200 μM dNTPs, ddH₂O, Superscript II (Invitrogen) and 5× First Strand buffer, 10 mM DTT, RNasin (Promega). The reaction was mixed at room temperature and incubated at 42°C for 60 min. The Superscript II was inactivated by heating the mixture to 70°C for 15 min in the presence of NaOH. Of the resultant cDNA, 5 μl was mixed with 10× PCR buffer, 200 μM dNTPs, 1.5 μM MgCl2, ddH₂O, 0.5 μM of the forward primer 5′GGAAGCTTGGGATCGATGCCATGGACCGGGGAAACCAATGAAAAGCG-3′ and 0.5 μM of the reverse primer 5′GCGGCCGCATGGGGAGAAGGTGGTGGTGG-3′. The following conditions were used: a) 95°C for 5 min (1 cycle); b) 95°C for 30 s, 55°C annealing 30 s, and 72°C extension 1 min (30 cycles); c) 72°C terminal extension 10 min (1 cycle); and d) 4°C infinite hold on a PerkinElmer 9600 cycler. Products were separated by agarose gel electrophoresis; the gel was stained with ethidium bromide and imaged (Kodak Gel Logic 1000).

MEF Arr2^−/−/arr3^−/− cells were transiently cotransfected with arrestin-2 GFP, arrestin-3 GFP or pEGFP-N1 and muCCR7 using Lipofectamine 2000 per the manufacturer’s instructions using half of the recommended amounts of reagents and DNA. Twenty-four hours following transfection, MEF Arr2^−/−/arr3^−/− cells were harvested from dishes by trypsinization and re-plated in larger dishes. HuT 78 suspension cells were collected from flasks, counted, and washed in PBS. Cells were re-suspended in sfRPMI 1640 (HuT 78) or sfDMEM-(MEF arr2^−/−/arr3^−/−) at 5 × 10⁶ cells per ml, and 100 μl was used for each time point. Following a 10-min preincubation at 37°C, cells were treated with 200 nM CCL19 or CCL21 for the indicated time periods. Internalization was arrested by adding 10 volumes of ice-cold sfRPMI 1640 HuT 78 or sfDMEM MEF arr2^−/−/arr3^−/−. Human ligands were used with the HuT 78 cells, while the murine ligands were used with murine CCR7 in the MEF cell lines. Receptors on the surface of the HuT 78 cells were stained with PE-conjugated anti-human CCR7 Abs for 30 min on ice (as per the manufacturer’s instructions). FLAG-tagged murine CCR7 was detected by incubating the MEF arr2^−/−/arr3^−/− cells with anti-FLAG (M2) Ab on ice for 30 min, followed by three rinses with ice cold PBS and incubation with FITC anti-mouse secondary for 30 min on ice or with CCL-19 Ig. Cells were rinsed three times in serum free media and assayed immediately or fixed in 2% paraformaldehyde and stored at 4°C for up to 4 days. Receptor levels were assayed by flow cytometry on a Becton Dickinson LSRII with DIVA software. Ab binding was not inhibited by the presence of either CCL19 or CCL21 (data not shown). Level of receptor remaining of the surface was calculated as [mean channel fluorescence of PE-anti-CCR7 or FITC-anti-mouse at time-point]/[mean channel fluorescence of PE-anti-CCR7 at time = 0]. Statistical significance was determined using a paired t test. Curves or linear regressions were fit using Prism 4 software package (GraphPad).

HuT 78 suspension cells were collected from flasks. Cells were counted and washed in PBS. Cells were resuspended in sf RPMI 1640 at 5×10⁶ cells per ml and 100 μl was used for each time point. Following a 10-min preincubation at 37°C, cells were treated with 200 nM CCL19 or CCL21 for 20 min. Ligand was added again following 20 min of receptor internalization to prevent ligand depletion. At 30 min, cells were rinsed three times in ice-cold sfRPMI 1640. Receptor recycling was initiated by addition of prewarmed 37°C sfRPMI 1640. Cells were collected at the indicated time points. Recycling was arrested by addition of 10 volumes of ice-cold sfRPMI 1640. Receptors on the surface of the cells were stained with PE-conjugated anti-human CCR7 Abs for 30 min on ice (as per the manufacturer’s instructions). Cells were rinsed three times in sfRPMI 1640 and assayed immediately or fixed in 2% paraformaldehyde and stored at 4°C for up to 4 days. Receptor levels were assayed by flow cytometry on a Becton Dickinson LSRII with DIVA software. Level of receptor remaining of the surface was calculated as [mean channel fluorescence of PE-anti-CCR7 at time-point]/[mean channel fluorescence of PE-anti-CCR7 at time = 0]. Curves were fit using Prism 4 software package (Graphpad). Statistics were determined using a paired t test.

FITC-labeled, double-stranded siRNA, with 19-nucleotide duplex RNA and 2-nucleotide 3′-dTdT over-hangs were purchased from Invitrogen. The siRNA sequences targeting human β-arrestin1 and β-arrestin2 were 5′-AAAGCCUUCUGCGCGGAGAAU-3′ and 5′-AAGGACCGCAAAGUGUUUGUG-3′ corresponding to positions 439–459 and 148–168 relative to the start codon, respectively. A nonsilencing RNA duplex (5′-AAUUCUCCGAACGUGUCACGU-3′), as the manufacturer indicated, was used as a control. These siRNAs have been described (41). A total of 20 pmol of siRNAs were Nucleofected (AMAXA) using solution R (see below) or transfected with Oligofectamine (Invitrogen) to introduce the siRNA. Briefly, on the day of the transfection, cells were plated at ∼500,000/80 μl sfRPMI 1640 in a 24-well dish. A total of 1 μl of a 20 μM stock of each double-stranded siRNA was combined with 16 μl Opti-MEMI reduced serum medium. Oligofectamine was inverted several times to mix. In a separate Eppendorf tube 0.8 μl of Oligofectamine was diluted with 2.2 μl Opti-MEMI reduced serum medium. The mixtures in tubes were incubated separately at room temperature for 10 min, combined into one tube, and allowed to incubate at room temperature for an additional 20 min to form siRNA/Oligofectamine complexes. The 20 μl of complexes were added to the cells and incubated at 37°C in a humidified, CO₂ incubator for 4 h. A total of 50 μl of RPMI 1640/30%FBS/l-glutamine was added to each of the 24 wells. Following treatment with Oligofectamine, or nucleofection cells were allowed to incubate for 72 h and then assayed for internalization and arrestin expression. Changes in arrestin 2 and/or arrestin 3 expression levels were confirmed by Western blotting.

A total of 10⁶ HuT 78 cells were harvested for each condition (cells pretreated for 72 h with nonspecific, arrestin 2 or arrestin 3 siRNA) from flasks, counted and washed in PBS. Cells were pelleted, and the media was replaced with 200 μl of RIPA lysis buffer (50 mM HEPES (pH 7.5), 150 mM NaCl, 5 mM EDTA, 1 mM EGTA, 1 mM NaF, 2 mM Na₃VO₄, 1% NP40, 0.1% SDS, 1% deoxycholate, and 1 mM DTT with Protease Inhibitor mixture from Sigma-Aldrich). Lysates were separated by SDS-PAGE on 10% acrylamide gels, transferred to PVDF, and probed for arrestins with a rabbit anti-β-Arrestin-1/2/3 (H-290) Ab (Santa Cruz Biotechnology). Immune complexes were probed with a rabbit-HRP secondary Ab (Pierce) and visualized by Super Signal West Pico Substrate (Pierce) and exposed membranes to x-ray film or to the Fuji LAS-4000 and bands were quantified.

Chemotaxis assays performed as described (42, 43). In brief, siRNA-transfected cells were resuspended in assay buffer (Hank’s Buffered Saline Solution (Cellgro) plus 10 mM HEPES (pH 7.4) (Invitrogen) with 0.5 g BSA (A790 Sigma-Aldrich V)). Cells were diluted to 4 × 10⁶ cells/ml in complete RPMI 1640 and kept at room temperature during preparation of the chemotaxis chambers. Polycarbonate membranes (8-μm pore size) were pre-wet with serum-free RPMI 1640. Chemokines were diluted to 200 nM in 1%BSA/PBS and pipetted into lower chamber of a Neuroprobe chemotaxis chamber. Chambers were fitted with filters, gaskets, and lid. A total of 50 μl of cells (100,000 cells/well) were loaded and cells allowed to migrate for 2 h at 37°C in a 5% CO₂ humidified incubator. Following incubation the membranes were removed and cells that migrated to the lower chamber were counted.

A total of 2 × 10⁶ HuT 78 cells were transiently nucleofected with arrestin 2-GFP, arrestin 3-GFP, or pEGFP-N1 using the AMAXA R kit per the manufacturer’s instructions. In brief, cells were centrifuged for 10 min at 90 × g and resuspended in 100 μl Solution R. DNA was added to an Eppendorf and cells resuspended in SolutionR were layered on top. Before nucleofection, cells/DNA/SolutionR was transferred to an Amaxa cuvette and cells were nucleofected on program V-01. Immediately after nucleofection, 800 μl of cRPMI 1640 was added and cells were transferred back into the Eppendorf tube with the transfer pipette included in the Amaxa kit. Cells were then plated in complete RPMI 1640 and allowed to recover overnight. Twenty-four hours following transfection, HuT 78 cells were collected, stimulated with 200 nM ligand for 4 or 8 min, and fixed with 2% formaldehyde (30 min on ice). Cells were permeabilized with 1% Nonidet-P 40/1 × PBS for 15 min at room temperature and labeled with mouse anti-CCR7 (1/100) (Pharmingen no.550937) in 3% BSA/PBS for 1 h. The cells were rinsed three times and labeled with Texas red (1:250) directly conjugated to anti-mouse (Jackson ImmunoResearch Laboratories). Cells were rinsed three times in PBS and plated on glass slides with Prolong Gold anti-fade (Molecular Probes). Immunofluorescence was visualized on a Nikon 90i upright microscope via mercury fluorescent excitation then confirmed via confocal scanning (Nikon C1 series confocal scan head). Lasers used for emission detection were a 488 Multi line Argon (green IR) and 561 Diode laser (red IR) acquisition via the Nikon EZ-C1 series software (3.60).

To examine the expression of endogenous CCR7 in human T cells, we used RT-PCR (Fig. 1) to look for expression of CCR7 in two human cell lines. In line with previous observations, HuT 78 cells expressed CCR7 mRNA, in contrast with Jurkat cells, which failed to express CCR7 (44, 45) (Fig. 1,A). To verify surface localization of expressed CCR7 protein, we used FACS (Fig. 1 B). We confirmed that CCR7 was expressed on the surface of HuT 78 cells.

Receptor internalization is an important cellular function that mediates receptor desensitization (11, 46). To confirm that the HuT 78 cell line internalized and recycled in a manner similar to the CCR7 expressed in peripheral blood cells (44), we used FACS based internalization and receptor recycling assays. CCR7 is activated by two endogenous ligands, CCL19 and CCL21 (25, 31, 47). Initially, to identify an optimal concentration of ligand for our assays, we conducted a dose response (Fig. 2,A) and found that 200 nM CCL21 and 200 nM CCL19 provided maximal internalization for CCR7. We did not observe increased internalization levels at higher concentrations, possibly due to loss of ligand solubility. Therefore, we selected 200 nM ligand to carry out further studies. We compared the extent of internalization and rates of CCR7 internalization in the HuT 78 cells following binding to 200 nM CCL19 and CCL21 ligands (Fig. 2, B and C). Similar to published observations for peripheral blood leukocytes (44, 45), 76 ± 9% of CCR7 was internalized with a t_1/2 of 0.46 min following exposure of the HuT 78 cells to 200 nM CCL19. In contrast, only 22 ± 4% of CCR7 was internalized with a t_1/2 of 0.28 min following exposure of HuT 78 cells to 200 nM of CCL21. Thus, there was a more rapid internalization of CCR7/CCL21, although the total amount of CCR7/CCL19 internalization was greater.

When a receptor internalizes, the extent of this process is a balance between receptor internalization and recycling. Therefore, we hypothesized that the enhanced rate of CCR7/CCL19 internalization was due to a reduced rate of recycling. To test this hypothesis, we compared CCR7 recycling following receptor internalization with CCL19 or CCL21. The difference in the receptor recycling, between the two ligands, was significant (p = 0.02). We observed the CCL19 internalized CCR7 recycled slowly (Fig. 3), while the CCL21 internalized CCR7 recycled within 10 min (data not shown).

Receptor internalization is a key cellular function that controls the duration of signaling from a receptor. As such, the cellular mechanisms that are activated during receptor internalization play pivotal roles in controlling the fate of the internalized receptor and in T cells regulate the immune response (11). Arrestins regulate internalization of many GPCRs, such as the β2 adrenergic receptor, the V2 Vasopressin receptor, and the angiotensin II type 1A receptor, but are not required for internalization of the human FPR, or the PAR-1 receptor (8, 20, 48, 49, 50, 51). Thus, we questioned whether arrestin binding regulated the differential internalization of CCR7 following binding to either CCL19 or CCL21. For this purpose, we used a well-characterized combination of siRNAs (41, 52, 53) to knock down the expression level of arrestin 2 or arrestin 3 in the HuT 78 cells (Fig. 4). We used Western blots to confirm that arrestin 2 and arrestin 3 levels were knocked down (Fig. 4,C). When normalized to the β-actin loading control, arrestin 2 siRNA reduced levels of arrestin 2 by 58%, and arrestin 3 siRNA reduced levels of arrestin 3 by 75%. However, arrestin 3 siRNA also reduced arrestin 2 expression levels by 96%. Independent of a reduction in the expression levels of either arrestin 2 or arrestin 3, CCR7 was internalized following binding to CCL21 (Fig. 4). In contrast, the extent of internalization of CCR7/CCL19 was significantly reduced (p = 0.0013) in comparison to cells treated with nonspecific siRNA, following treatment with arrestin 3 siRNA.

Although these siRNA’s have been used to specifically knockdown arrestin 2 or arrestin 3 in human cells (41, 52, 53), in the HuT 78 cells used in this study, both arrestin 2 and arrestin 3 levels were reduced by the siRNA for arrestin 3. Therefore, it was unclear whether loss of internalization of CCR7/CCL19 in the arrestin 3 siRNA-treated cells was due to a reduction in the levels of arrestin 3 alone, or due to the loss of both arrestin 2 and arrestin 3 combined. An additional consideration was that since siRNA can have off target effects (54), which can knockdown the levels of nontargeted mRNAs in the cell, it was not clear whether the observed effects on CCR7/CCL19 internalization were due only to depletion of arrestins.

To examine internalization of CCR7/CCL19 and CCR7/CCL21 in the complete absence of arrestins, we examined internalization in the MEFs, which lack arrestin (MEF arr2^−/−/arr3^−/− cells). This MEF arr2^−/−/arr3^−/− cell line failed to internalize the β2 adrenergic receptor, which is an arrestin-dependent process (8). Since MEF cells lack endogenous CCR7, we transiently transfected MEF cells with a murine N-terminally FLAG-tagged murine CCR7. To facilitate detection with the M2 anti-FLAG Ab, the murine CCR7 was fused at the N terminus to the FLAG tag (28). The MEF transient transfectants were allowed to internalize in the presence of 200 nM ligand and assayed for receptor remaining on the surface of the cells by FACS (Fig. 5, A and B). Ligand binding did not affect Ab detection of the FLAG tag, as ≥100% of the receptor remained detectable on cells that were labeled with Ab, on ice, in the presence of ligand.

Since internalization of the CCR7/CCL19 was blocked by knocking down arrestin 3 expression (Fig. 4), we hypothesized that reconstitution of expression of arrestin 3 but not arrestin 2 would restore CCR7/CCL19 internalization in MEF arr2^−/−/arr3^−/− cells. Therefore, we cotransfected the MEF cells with muCCR7 and GFP, or muCCR7 and arrestin 2-GFP (Arr2-GFP), or muCCR7 and arrestin 3-GFP (Arr3-GFP). This allowed us to detect and gate on only the GFP or arrestin-GFP expressing cells during FACS for the internalization assay. MEF arr2^−/−/arr3^−/− cells internalized 23 ± 4% of CCR7/CCL21 within 30 min (Fig. 5 B) independent of arrestin expression, while CCR7/CCL19 failed to internalize in the absence of arrestin. When the MEF arr2^−/−/arr3^−/− were reconstituted with arrestin 3, the maximum level of CCR7/CCL19 receptor internalized was 32 ± 7%. When compared with internalization of CCR7/CCL19 in the presence of GFP alone, internalization of CCR7/CCL19 was significantly increased in the presence of arrestin 3, (p = 0.0051), but not in the presence of arrestin 2 (p = 0.1532). This observation suggests that the role of arrestin 3 is not redundant with the role of arrestin 2 during internalization of the CCR7/CCL19 receptor/ligand complex. In addition, this is the first observation of ligand dependent regulation of internalization of a chemokine receptor by arrestin.

Since we have shown that GPCRs can have lowered rates and extents of internalization when over-expressed in cells that do not endogenously express the receptor (20), we questioned whether the minimal internalization of CCR7 in the MEF cells was due to absence arrestin 2 or other as yet unidentified factors that are normally found in hematopoietic cells. To address this question, we over-expressed CCR7 in the wild-type MEF and examined internalization when CCR7 was bound to CCL19 (Fig. 5,C). There was no significant difference (p = 0.1093) between levels of CCR7 internalized in the wild-type MEF when compared with the Arr3-GFP reconstituted MEF arr2^−/−arr3^−/− cells (Fig. 5).

Arrestin-3 mediates chemotaxis through the CXCR4 receptor in HeLa and human embryonic kidney cells (55). To determine whether arrestins regulate migration of T cells via CCR7, we examined migration of HuT 78 cells following knockdown of arrestin 2 or arrestin 3. Since our maximal internalization of ligand was at 200 nM, we selected this ligand concentration for our studies. In the presence of a full complement of arrestins, the nonspecific siRNA-treated cells were capable of migration to CCL19 and CCL21 (Fig. 6). Similar to the effects on internalization, loss of arrestin 2 or arrestin 3 had no effect on the migration of HuT 78 cells to 200 nM CCL21. In contrast, depletion of either arrestin 2 or arrestin 3 blocked migration to CCL19, when compared with cells treated with nonspecific siRNA (Fig. 6).

Receptor internalization has multiple important functions within the cell. Initially, internalization moves a receptor from an area of high ligand concentration on the surface of the cell to inside the cell. Changes in spatial temporal localization of a GPCR can control signaling from these receptors. Finally, during internalization, a receptor is trafficked to an area within the cell where the receptor releases its ligands and is recycled or degraded. Generally, when a receptor binds its cognate ligand, this binding leads to a conformational change and receptor clustering. These clusters form signaling domains (56) and terminate further receptor stimulation by coordinating internalization. Arrestins serve as scaffolds that recruit signaling proteins to the C-terminal domains of activated GPCRs. To determine whether activation of CCR7 could induce coclustering with arrestins, we transiently transfected the HuT 78 cells with GFP, arrestin 3-GFP, or arrestin 2-GFP and examined the clustering of CCR7 by immunofluorescence microscopy (Fig. 7). The FPR was used as a control for clustering of arrestins. The cells were plated on glass coverslips, stimulated for the indicated periods of time with either 200 nM CCL19 or CCL21. Fixed/stained cells were examined for clustering of human CCR7 and arrestin. In the absence of ligand, at t = 0 min, we found CCR7 distributed evenly over the cell surface and in the cytoplasm.

Following treatment of HuT 78 cells with 200 nM CCL19 or CCL21, in both live (data not shown) and fixed cells (Fig. 7), we failed to detect clustering of arrestins at the cell surface during receptor internalization. In addition, unlike the FPR, which we have shown forms multiple distinct clusters at the cell membrane following ligand binding (20) (Fig. 7,C), we found that CCR7 moved into capped structures with CCL19 or CCL21. In these structures, we observed that CCR7 moved to the same region of the cell as arrestin 3 following receptor internalization with CCL19 at 4 and 8 min. In addition, CCR7/CCL21 and CCR7/CCL19 colocalized, weakly, with arrestin 2 at 4 min, but this colocalization was lost at 8 min. In the absence of arrestin over expression (Fig. 7 B), CCR7/CCL21 appeared more dispersed. We did not observe colocalization of either CCR7 or FPR with GFP. In ∼10% of the cells we observed movement of CCR7/CCL19 to the nucleus in large clusters with arrestin 3-GFP (data not shown).

This report is the first to describe differences in the mechanisms that are used by T cells to internalize CCR7 following CCL19 vs CCL21 ligand binding. Since mice lacking either CCL19 or CCL21 have not yet been described, it is important to distinguish the physiologically distinct roles that each ligand plays to better understand the regulation of CCR7 and downstream signaling pathways. We observed that while both CCL19 and CCL21 are capable of mediating internalization of CCR7, the mechanisms controlling the internalization lead to different outcomes. In this report, we used a human T cell line, HuT 78, which like naive T cells expresses an endogenous CCR7. Although HuT 78 internalized 76% of CCR7/CCL19, only 22% of CCR7/CCL21 was internalized (Fig. 2). In addition, we found that while CCL21 internalized receptor is rapidly recycled, CCL19 internalized receptor recycled at a slow rate in HuT 78 (Fig. 3 and data not shown). These observations are in line with previously published results describing CCR7 internalization and recycling in human peripheral blood lymphocytes (44, 45). The HuT 78 cell line, however, provided a model that allowed us to manipulate arrestin levels and examine the role of arrestins in the internalization of CCR7 and in the chemotaxis of T cells. Loss of arrestin 3, but not arrestin 2 expression blocked internalization of CCL19/CCR7 but had no effect on internalization of CCL21/CCR7. Therefore, we were surprised to find that the association of CCR7 with either arrestin 2 or arrestin 3 mediated chemotaxis of HuT 78 cells to CCL19 but had no effect on chemotaxis to CCL21 (Fig. 6). Equally important, using confocal immunofluorescence microscopy, we found that both arrestin 2 and arrestin 3 polarized or capped with CCR7 at some point following internalization (Fig. 7).

The CCR7 GPCR, like the nociceptin GPCR demonstrates ligand-dependent differences in receptor internalization and desensitization (57, 58, 59, 60). Both GPCRs differentially regulate receptor fate following ligand binding (45). The differential regulation of receptor fate suggests that ligand specific mechanisms are activated at the surface of the cell when each ligand binds that rapidly control ligand/receptor behavior. Unlike our studies in which both CCR7 ligands that are full agonists (61) were maximally internalized within 8 min (Fig. 2), the nociceptin receptor required different ligand-dependent incubation times to mediate receptor internalization. This may reflect the differential functions of the described ligands, since agonists and antagonists were internalized at different rates (57, 58, 59, 60).

That CCL19 and CCL21 regulate CCR7 internalization differently may reflect their unique functions in the immune system. Although CCL19 is expressed primarily by dendritic cells and within the T cell zone of the lymph nodes, CCL21 is constitutively expressed by the high endothelial venules of lymph nodes and is found within the spleen, Peyer’s patches, the thymic medulla, and the lymphatic endothelium of many tissues (62, 63). Normally, CCR7 functions to induce migration of cells into the lymph nodes (28, 64, 65, 66, 67). Initially, CCR7 expressing cells are exposed to CCL21 in the high endothelial venules of the lymph nodes. During this exposure, if a large number of CCR7 receptors were internalized, this could prevent the cell from reaching the CCL19 expressing dendritic cells. Recently, systemic application of CCL19-Ig, which likely internalized most available CCR7, prevented T cells from forming conjugates with dendritic cells in the lymph nodes, in vivo (68). In addition, whereas in the presence of a CCL19 antagonist cells were able to migrate to the lymph nodes, they failed to exit (69). These studies suggested that CCL19 likely contributes to pathways involved in T cells exiting the lymph nodes. The signaling pathway regulated by CCL19 is currently under investigation in our laboratory (L. Shannon, M. Byers, and C. Vines, manuscript in preparation). In contrast, if the cell fails to encounter a CCL19-expressing dendritic cell, then the low level of internalization and the high rate of CCR7 recycling promoted by the CCR7/CCL21 receptor/ligand complex (Fig. 2 and data not shown) would allow the cell to continue to traffic through the lymph node. Since stimulation with CCL19 removes ∼80% of the CCR7 from the cell surface, it is not surprising the T cell becomes unable to respond to the CCL21 stimulus of CCR7 following encounter with CCL19 (Fig. 2 and Ref. 61). By reducing the numbers of CCR7 receptors on the surface and desensitizing the cell to further stimulation with CCR7 ligands (61), CCL19 may help to retain the T cell in the lymph node once it has formed a T cell/dendritic cell conjugate.

Further evidence of the differential regulation of CCR7/CCL19 and CCR7/CCL21 has been revealed in studies of receptor phosphorylation in human embryonic kidney (HEK293) cells transiently transfected with human CCR7. In these studies, stimulation of CCR7 with CCL19 led to an 83% increase in phosphorylation of the C terminus of CCR7 over basal levels in contrast with only an 18% increase in phosphorylation levels when activated by CCL21 (61). Phosphorylation of GPCRs is a prerequisite for recruitment of arrestins to the C-terminal domains of the receptors. In our study, we find that in T cells and in MEFs, arrestins are required for internalization of the CCR7/CCL19 ligand receptor combination but not of CCR7/CCL21 (Figs. 4 and 5).

For ligand internalization assays and isotype controls, 10⁴ to 2 × 10⁴ cells were assayed per sample. The flow cytometry data was reported as percent of total cells to be able to directly compare the peaks. There were equal maximal downshifts for both CCL19 and CCL21 ligands; however, a subpopulation of CCL21-treated cells showed no internalization after 8 min. These results suggest that HuT 78 cells are heterogeneous in regards to CCL21 internalization, which is not surprising, since HuT 78 cells, similar to primary human peripheral blood cells, are not homogeneous but derived from an heterogeneous human lymphoma (70). In addition, these cells are at different phases of the cell cycle and may respond differently to stimulation with CCL21 depending upon their state in the cell cycle. Like the HuT 78 cells, primary cells, which are heterogeneous and found at different stages of the cell cycle, internalize ∼25% of CCL21 and 60% of CCL19 in flow cytometry (45). While it is beyond the scope of this study to try to identify different subpopulations of primary or HuT 78 cells that behave differently in response to CCL21, it is important to note that HuT 78 cells have been used to isolate several subclones that display distinct phenotypic differences from others cells in the HuT 78 human T cell lymphoma line (71, 72).

In our studies, we used well-characterized siRNAs against arrestin 2 and arrestin 3 in HuT 78 T cells. These siRNAs have been used successfully to knockdown arrestin 2 or arrestin 3 in human embryonic kidney cells (41). Although the transfection efficiency was ∼90%, as measured by FACS, we were unable to specifically knock down levels of arrestin 2 or arrestin 3. The inability to knockdown only arrestin 2 or arrestin 3 in the HuT 78 cells may be due to the high level of conservation between arrestin 2 and arrestin 3 in the regions against which the siRNAs were generated. Unlike previous studies that were able to specifically deplete arrestin 3 with arrestin 3 siRNA, we failed to see specific depletion of arrestin 3 but also observed reduced levels of arrestin 2. This made it difficult to determine whether the effects of the siRNA were due to depletion of arrestin 3 alone or due to the concomitant loss of arrestin 2. Loss of arrestin 2 and 3 did not effect CCR7/CCL21 internalization when compared with cells treated with nonspecific siRNA (Fig. 4,A). Therefore, although arrestin 3 mediated CCR7/CCL19 internalization, neither arrestin 2 nor arrestin 3 regulated CCR7/CCL21 internalization in T cells. In support of this observation, we were able to rescue CCR7/CCL19 internalization in MEF arr2^−/−arr3^−/− cells by transient transfection of arrestin 2 but not of arrestin 3 (Fig. 5).

The differential phosphorylation of CCR7 following engagement of CCL19 vs CCL21 positions the receptors for binding to unique effectors, as we observed with arrestins in this study (Figs. 3–6). How the phosphorylation of CCR7/CCL19 and CCR7/CCL21 are regulated, leading to homologous desensitization, is unknown. Basal levels of phosphorylation are mediated by the second messenger kinase, protein kinase C (73). Other kinases, however, such as the GPCR kinases, which regulate homologous desensitization of GPCRs, have not yet been identified for CCR7. Studies are currently underway in our laboratory to begin to define the GPCR kinases that differentially regulate phosphorylation of CCR7.

Arrestin 2 and arrestin 3 controlled chemotaxis to CCL19 (Fig. 6). This was unexpected, since CCR7/CCL19 is internalized via an arrestin 3-dependent, but arrestin 2-independent, pathway. However, this was in line with previous studies that showed that internalization of a GPCR is not required for chemotaxis (74). Signaling through chemokine receptors such as CCR7 can control T cell chemotaxis (75, 76, 77) since migration of lymphocytes can be regulated through integrins, heterodimeric adhesion molecules that respond to signals originating at activated GPCRs (78). Therefore, it is curious that although only arrestin 3 regulates receptor internalization, loss of arrestin 2 or arrestin 3 affects migration only to CCL19 and not to CCL21. This observation could suggest some redundancy of arrestin function in mediating activation of adhesion proteins that are used for migration following activation of CCR7. Although we and others (78, 79) have shown that GPCRs can regulate integrin adhesion, the roles of arrestins in the regulation of integrin adhesion/migration has not yet been defined. In this study, we have observed distinct roles for arrestin 3 in the regulation of CCR7/CCL19 internalization but not CCR7/CCL21. Arrestin 2 did not regulate the internalization of CCR7/CCL19 or CCR7/CCL21 but along with arrestin 3 controlled migration to CCR7/CCL19. Since arrestins can serve as scaffolds for many different effectors proteins, we anticipate that this early difference in binding to intracellular signaling proteins can be better defined by further dissecting the proteins that are recruited by arrestins following engagement of CCR7 by its ligands, CCL19 and CCL21.

As we and others have discussed, arrestin recruitment to GPCRs can have different consequences. With the β2 adrenergic receptor, arrestins bind to the phosphorylated form to mediate recruitment of receptors to clathrin-coated pits (13, 20). With the V2 Vasopressin receptor, arrestins prevent recycling by trafficking the receptor to the lysosome (21). In contrast, while the FPR does not use arrestins to mediate receptor internalization (80), we have shown that the FPR uses arrestins to regulate receptor recycling (20) and prevent GPCR-mediated apoptosis (81). In these studies, arrestins form distinct small clusters with the GPCRs during internalization and trafficking. Therefore, it was surprising to us that, although we observed large areas of colocalization that resembled capping of proteins (82), we failed to detect ligand-bound CCR7 in small clusters with arrestin 2-GFP or arrestin 3-GFP either in video microscopy studies (data not shown) or in fixed cells (Fig. 7). We observed that CCR7 recruits arrestin 3 only when it is in association with CCL19 but not with CCL21. (Fig. 7 A). CCR7 bound to CCL21 fails to cluster with arrestin 3 at 4 min and only weakly at 8 min. This in contrast to CCR7 bound to CCL19, which clusters with arrestin 3 at both time points. Arrestin 2 forms weak clusters with CCR7 at 4 min. These clusters are gone by 8 min. The HuT 78 cells are capable of clustering arrestins with the FPR, demonstrating that these cells have the machinery to tightly cluster arrestins with GPCRs. In contrast, we were not able to observe the same strong clusters of arrestins with CCR7 in the HuT 78 cells, suggesting that the cellular mechanisms used by CCR7 that control internalization are distinct from the mechanisms used by the FPR. Interestingly, we did observe that a few of the cells stimulated with CCL19 showed aggregates of receptor/arrestin 3 in the nucleus (data not shown). Developing a more complete understanding of the role of arrestins in the regulation of these internalization and trafficking events is the focus of ongoing investigations in our laboratory.

These are the first studies that define a role for arrestin 3 in the internalization of a chemokine receptor following binding of one but not both endogenous ligands. We have found that internalization of CCR7 is regulated by two distinct mechanisms; an arrestin-dependent and an arrestin-independent mechanism. In response to binding either CCL19 or CCL21, CCR7 uses either an arrestin 3-dependent or an arrestin 3-independent mechanism, respectively, to be internalized. This study provides insight into how differential recruitment of effectors may control immunological events that are initiated by CCR7 when it is bound to its two ligands CCL19 or CCL21.

We thank Colin A. Bill and Brant Wagener for careful reading of the manuscript and Steve Benedict, Thomas Yankee, Kelly Cool, and Abby Dodson for helpful discussions.

The authors have no financial conflict of interest.