CXCR3 Internalization Following T Cell-Endothelial Cell Contact: Preferential Role of IFN-Inducible T Cell α Chemoattractant (CXCL11)

Reagents

Recombinant human IP-10, I-TAC, Mig, IFN-γ, and IL-2 were obtained from PeproTech (Rocky Hill, NJ). Neutralizing murine mAb to human IP-10 and Mig were purchased from R&D Systems (Minneapolis, MN). Neutralizing murine mAb to human I-TAC was a kind gift from K. Neote (Pfizer, Groton, CT). FITC-conjugated and nonlabeled mouse anti-human CXCR3 mAb (49801.111; IgG1) was purchased from R&D Systems. PE-conjugated mouse anti-CXCR4 mAb (12G5, IgG2a), irrelevant-purified and FITC-conjugated IgG1, and PE-conjugated IgG2a mouse Abs (isotype controls) were obtained from BD PharMingen (San Diego, CA). Texas Red-conjugated goat anti-mouse (GAM) Ab was obtained from Southern Biotechnology Associates (Birmingham, AL). Bispecific anti-CD3/CD8 was a kind gift from J. Wong (Massachusetts General Hospital, Boston, MA; Ref. 27). PTX and calcium ionophore were purchased from Calbiochem (La Jolla, CA). Genestein, wortmannin, PMA, staurosporin, and propidium iodine were obtained from Sigma-Aldrich (St. Louis, MO). T cell enrichment columns were purchased from R&D Systems.

Cell isolation and culture

Human vascular endothelial cells were isolated from saphenous veins from seven different donors by collagenase treatment as described elsewhere (5). Cells were maintained in M199 medium (BioWhittaker, Walkersville, MD) supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin (BioWhittaker), 5% FCS (Atlanta Biologicals, Norcross, GA), 100 μg/ml heparin (Sigma-Aldrich), and 50 mg/ml endothelial cell growth factor (Pel-Freez Biologicals, Rogers, AR). Human microvascular endothelial (HMEC)-1 cells, a human microvascular cell line, were cultured in complete endothelial growth medium (Clonetics, Walkersville, MD) supplemented with 5% FCS, gentamicin, amphotericin, bovine brain extract (12 μg/ml) and hydrocortisone (1 μg/ml), and epidermal growth factor (10 ng/ml).

CXCR3⁺ T cells were obtained from different sources. In some experiments, PBMC isolated by a Ficoll gradient were incubated with 200 ng/ml IL-2 and bispecific anti-CD3/CD8 or anti-CD3/CD4 for 4 days in RPMI 1640 (Cellgro and Mediatech, Herndon, VA) supplemented with 10% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin (Mediatech), and 2 mM of l-glutamine (Mediatech). Proliferating CXCR3⁺ CD4⁺ or CD8⁺ T cells were kept in IL-2 only for an additional 4–10 days. In other experiments, CD3⁺ T cells were derived from PBMC passed through a T cell enrichment column (R&D Systems) and incubated for 4 days with 5 μg/ml PHA (Sigma-Aldrich) and 200 ng/ml IL-2 and then in IL-2 only for 4–10 days. Human CXCR3⁺-transfected 300-19 cells were a gift from B. Moser (Theodor Kocher Institute, Bern, Switzerland; Ref. 6).

Coculture experiments

Human saphenous vein endothelial cells (HSVEC) were used from passage two to passage four and were cultured in 0.1% gelatin-precoated wells from 6-well plates (Costar, Cambridge, MA). When cells reached confluence, they were treated with or without 100 ng/ml IFN-γ for 16 h in fresh complete M199 medium. Following removal of the supernatant and two washes with prewarmed HBSS, CXCR3⁺ T cells (1 × 10⁶) in 1 ml of complete RPMI were then added per well. HSVEC-conditioned medium (250 μl) was used to resuspend a separate aliquot of 0.5 × 10⁶ CXCR3⁺ T cells pelleted in an Eppendorf. After various times, CXCR3⁺ cells were collected by being gently washed off of the endothelial monolayer in 6-well plates or by centrifugation of CXCR3⁺ cells incubated with HSVEC supernatants. Cells were analyzed for CXCR3 surface expression by flow cytometry. In control experiments, we found that IFN-γ treatment of T cells had no direct effect on CXCR3 expression.

Measurement of CXCR3 surface expression

CXCR3⁺ cells (0.25–0.5 × 10⁶) were washed into an ice-cold FACS buffer (HBSS without Ca²⁺ and Mg²⁺ supplemented with 1% BSA, 2% FCS, and 0.1% sodium azide) and resuspended in 90 μl of FACS buffer and 10 μl of FITC-conjugated anti-CXCR3 mAb, PE-conjugated anti-CXCR4, or isotype controls. Cells were incubated for 30 min on ice in the dark. After two washings with the ice-cold FACS buffer, cells were resuspended in FACS buffer with propidium iodine (0.3 μg/ml, and only when CXCR3 was studied alone) to gate on live cells. FACS analysis was performed on 10⁴ cells using a FACSCaliber flow cytometer (BD Biosciences, San Jose, CA) using CellQuest software (San Jose, CA). Mean fluorescence intensity (MFI) values were obtained by subtracting the MFI of the isotype control from the MFI of the positively stained sample. CXCR3 surface expression was expressed as a percent of baseline expression using the formula: (MFI of tested cells/MFI of untreated cells) × 100.

Regulation of CXCR3 cell surface expression

CXCR3⁺ T cells or CXCR3⁺ 300-19 cells (0.25 × 10⁶) were incubated with various concentrations of rIP-10, I-TAC, or Mig for different times as indicated. In some experiments, T cells were preincubated for 2 h with 1 μM of wortmannin, 100 μM of genestein, 10 nM of staurosporin, or 0–1000 ng/ml PTX at 37°C, 5% CO₂ before the addition of I-TAC, IP-10, or Mig (250 ng/ml), or PMA (0.1 nM) for an additional 30 min. T cells were also incubated with different concentrations of PMA and calcium ionophore for 30 min. At the end of these experiments, ice-cold FACS buffer was added, then cells were studied for cell surface expression of CXCR3.

Indirect immunofluorescence

CXCR3⁺ T cells or 300-19 cells (10⁷ cells/ml) were incubated with anti-CXCR3 (10 μg/ml in PBS) for 30 min on ice. Cells were then washed twice in ice-cold PBS and then incubated for 30 min with 500 ng/ml I-TAC diluted in complete RPMI prewarmed at 37°C. Alternatively, they were exposed for 30 min to a monolayer of HSVEC treated or not with IFN-γ for 16 h and washed twice with prewarmed PBS. Cells were then resuspended in the ice-cold FACS buffer, centrifugated, and immediately fixed in 3.7% paraformaldehyde for 20 min on ice. After permeabilization with 0.1% saponin in FACS buffer, cells were incubated with Texas Red-conjugated GAM diluted in 0.1% saponin buffer for 30 min on ice and in complete darkness. CXCR3 expression was assessed by immunofluorescence microscopy (Axiophot microscope; Zeiss, Oberkochen, Germany).

Role of heparin

Confluent HSVEC were incubated with complete M199 supplemented with 0–500 μg/ml heparin with or without 100 ng/ml IFN-γ for 16 h. After removal of the supernatant, cells were washed twice with prewarmed HBSS (37°C). HSVEC and their supernatants were examined for their ability to induce CXCR3 internalization on T cells as mentioned above. In another set of experiments, 100 ng/ml IP-10 or I-TAC in either complete RPMI, complete M199 (100 μg/ml heparin), or complete M199 without heparin were used to resuspend CXCR3⁺ T cells. Following incubation for 2 h at 37°C, 5% CO₂, T cells were stained for CXCR3 surface expression and analyzed by FACS.

Effect of exogenous CXCR3 ligands added on HSVEC

Confluent HSVEC were incubated with complete M199 (including 100 μg/ml heparin) with or without 250 μg/ml I-TAC, IP-10, or Mig, or the three of them for 1 h at 37°C, 5% CO₂. After removal of the supernatant, cells were washed twice with prewarmed HBSS (37°C) to remove unbound chemokines. CXCR3⁺ T cells were then added onto HSVEC monolayers or resuspended into HSVEC supernatants and incubated for 30 min before CXCR3 surface expression was analyzed by FACS.

CXCR3 recycling

CXCR3⁺ T cells (0.75 × 10⁶) were treated with I-TAC (500 ng/ml) in complete RPMI for 30 min (37°C, 5% CO₂). Following two washes with prewarmed (37°C) HBSS to remove unbound chemokines, they were resuspended in 750 μl of complete RPMI for up to 3 h. CXCR3 surface expression was determined at baseline in untreated cells (−30 min) and at times 0, 1, and 3 h following I-TAC stimulation.

Neutralizing HSVEC chemokine activity

When tested for their ability to inhibit CXCR3 internalization induced by recombinant chemokines, all neutralizing mAb were effective in a dose-dependent manner. Fifty micrograms per milliliter neutralizing mAb significantly reduced CXCR3 surface disappearance induced by 250 ng/ml Mig (43.4 vs 18%), IP-10 (47 vs 18%), and I-TAC (65 vs 19%). IFN-γ-treated HSVEC supernatants were pretreated with 50 μg/ml neutralizing anti-Mig, anti-IP-10, or anti-I-TAC mAb or control mAb for 1 h at 37°C. CXCR3⁺ T cells were then incubated with these supernatants for 30 min and CXCR3 surface expression examined as described above. The ability of neutralizing mAb to inhibit the activity of HSVEC monolayers was also studied. After removal of their supernatant and two washes, HSVEC monolayers (in 24-well plates) were treated with 50 μg/ml neutralizing mAb or control mAb for 1 h at 37°C before the addition of CXCR3⁺ T cells.

Transendothelial migration

HMEC-1 cells were trypsinized and added (5 × 10⁵ in 50 μl of complete endothelial cell growth medium) on the top of the filter (polycarbonate, pore size: 5 μm) of a 96-well ChemoTx plate (NeuroProbe, Gaithersburg, MD). After 96 h, cell adherence and confluence were confirmed by staining with Diff-Quick and acetyl/low density lipoprotein-DiI (Biomedical Technologies, Stoughton, MA), 10 μg/ml in DMEM 1% BSA. Indicated concentrations of chemokines were added in the bottom chamber (in 31 μl of RPMI/1% BSA). After gentle washing of the HMEC-1 cells, 2.5 × 10⁴ of human CXCR3⁺ 300-19 cells in 25 μl were added on top of HMEC-1 cells and incubated for 5 h at 37°C, 5% CO₂. The filter was then removed and the migrated cells were counted. Each experiment was performed in triplicate.

Data analysis

Experiments were performed using at least five different donors for HSVEC and CXCR3⁺ T cells. Results are shown as the mean of at least three independent experiments ± SE. Statistical analysis was calculated using a paired Student’s t test and statistical significance was considered with p < 0.05.

CXCR3 down-regulation induced by IFN-γ-activated HSVEC

Initial experiments showed that overnight coincubation of freshly isolated CD3⁺ T cells with IFN-γ-stimulated, but not resting, HSVEC led to the disappearance of CXCR3 surface expression on T cells (Fig. 1⇓A). We found this for both CD8⁺ and CD4⁺ CXCR3⁺ T cells (data not shown). Time-course experiments showed that CXCR3 surface expression decreased from 100 to 30–35% within the first 15 min of coincubation (Fig. 1⇓B). To assess whether this process was specific for CXCR3, we studied CXCR4 expression during these kinetic experiments and found no significant decrease in CXCR4 expression (Fig. 1⇓B). Thus, IFN-γ activated, but not resting, HSVEC quickly and selectively down-regulates CXCR3 expression on T cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Cell surface expression of CXCR3 on T cells cocultured with HSVEC. A, Freshly isolated peripheral blood CD3⁺ T cells (1 × 10⁶) were incubated with resting or IFN-γ-activated (100 ng/ml for 24 h) confluent HSVEC. After 16 h, T cells were removed from the HSVEC monolayers and analyzed by FACS for CXCR3 cell surface expression. Percentages in the upper right squares represent the percentage of CXCR3⁺ cells (37.4 vs 2%). B, Kinetics of CXCR3 and CXCR4 expression on T cells cocultured with HSVEC. CXCR3⁺ T cells were incubated with resting or IFN-γ-stimulated HSVEC (100 ng/ml overnight in complete medium). At indicated times, cell surface expression of CXCR3 and CXCR4 was assessed by FACS analysis. Data represent the mean ± SE of three independent experiments. C, Efficacy of HSVEC monolayers and HSVEC supernatants at inducing CXCR3 internalization on T cells. CXCR3⁺ T cells (1 × 10⁶) were cocultured with resting or IFN-γ-treated HSVEC (100 ng/ml, 16 h). Alternatively, 250 μl of supernatant collected from these same HSVEC cultures were used to resuspend CXCR3⁺ T cells (0.25 × 10⁶). After 2 h at 37°C and 5% CO₂, T cells were removed and analyzed for CXCR3 surface expression by FACS. Compared with respective controls, CXCR3 internalization was greater with HSVEC monolayers (72.3 ± 4.6%) compared with their supernatants (48.3 ± 12.8%), p < 0.015. Data represent the mean ± SE of four independent experiments.

Cell contact is more efficient at inducing CXCR3 down-regulation than is conditioned medium

The most likely explanation for selective chemokine receptor internalization is selective chemokine receptor activation (26). IFN-γ-activated endothelial cells secrete the three known CXCR3 ligands: IP-10, Mig, and I-TAC (5, 28). Following overnight culture, we have found ∼100 ng/ml IP-10, 10 ng/ml Mig, and 1 ng/ml I-TAC in the conditioned medium of IFN-γ activatedHSVEC (28). Although chemokines are secreted from cells, they are basic proteins that under physiological conditions have an affinity for cell surface negatively charged heparan sulfate proteoglycans (HSPG) and accumulate on the surface of endothelial cells (29). Therefore, we compared the ability of HSVEC-conditioned medium with that of direct cell contact at inducing CXCR3 down-regulation. Coincubation of CXCR3⁺ T cells with IFN-γ-activated HSVEC for 2 h induced an 80% decrease in CXCR3 expression compared with a 25% decrease following coincubation with unstimulated HSVEC controls (Fig. 1⇑C). In contrast, supernatants collected from these same IFN-γ-activated HSVEC decreased CXCR3 expression on T cells by only 50%, while supernatants collected from unstimulated HSVEC cultures had no effect on CXCR3 expression. Thus, IFN-γ-activated and control HSVEC were significantly more potent at inducing CXCR3 down-regulation than their respective conditioned media (p < 0.015 and p < 0.03, respectively).

CXCR3 down-regulation induced by CXCR3 ligands: IP-10, Mig, and I-TAC

Because we recently showed that IFN-γ-stimulated HSVEC produced IP-10, Mig, and I-TAC, the three known ligands for CXCR3 (5, 28), we sought to determine whether these chemokines induced CXCR3 internalization on T cell surfaces. CXCR3⁺ T cells and CXCR3⁺ 300-19 cells were treated with 500 ng/ml IP-10, Mig, or I-TAC for various times and then analyzed by FACS for CXCR3 surface expression (Fig. 2⇓A). We found that maximum receptor internalization occurred within the first 15 min with all three ligands, as shown for HSVEC-mediated internalization. For both cell types, I-TAC was more potent than IP-10 or Mig, but this difference was much greater for T cells compared with CXCR3-transfected 300-19 B cell line. Indeed, CXCR3 expression on T cells diminished by ∼75–85% with I-TAC, by ∼35–40% with IP-10, and by ∼30% with Mig. In addition, we observed that in presence of the ligand, there was no significant CXCR3 reappearance on T cells for up to 120 min. In contrast, CXCR3 reappeared on 300-19 cells by 120 min despite the continuous presence of the ligands (Fig. 2⇓A). The differences in the magnitude and kinetics of CXCR3 internalization and recycling in T cells compared with transfected 300-19 B cells is not clear. However, we found that simply washing CXCR3⁺ 300-19 cells for flow cytometric analysis or into fresh medium increased CXCR3 expression ∼1.5-fold and ∼2.5-fold above baseline, respectively. This effect was not affected by cycloheximide, indicating that these treatments induced the mobilization of an intracellular pool of receptors in these transfected cells (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Time- and dose-dependent internalization of CXCR3 on T cells and CXCR3⁺ 300-19 cells. A, CXCR3⁺ T cells and CXCR3-transfected 300-19 cells (2 × 10⁶ cells/ml) were incubated in complete RPMI supplemented with 500 ng/ml IP-10, Mig, or I-TAC for the indicated times. B, top panels, CXCR3⁺ T cells and CXCR3⁺ 300-19 cells (2 × 10⁶ cells/ml) were treated with the indicated concentrations of chemokine for 30 min. Propidium iodine (PI) was used to gate on viable cells. Bottom panels, Representative FACS data obtained with T cells treated with varying concentrations of I-TAC (0–1000 ng/ml). Values are the mean ± SE of three to four independent experiments.

Dose-response studies were also performed using CXCR3⁺ T cells and CXCR3⁺ 300-19 cells which were treated with 0, 10, 100, and 1000 ng/ml ligand for 30 min. At ≤10 ng/ml, only a minor decrease in CXCR3 expression was seen for all three ligands on T cells (Fig. 2⇑B). At higher concentrations, I-TAC was much more potent than IP-10 or Mig, and IP-10 was more potent than Mig. Representative results from FACS analysis performed on T cells treated with different doses of I-TAC are shown in the bottom panel of Fig. 2⇑B. A similar attenuated pattern was also found with CXCR3⁺ 300-19 cells (Fig. 2⇑B). These results demonstrate that IP-10, Mig, and I-TAC have differential abilities to down-regulate CXCR3 internalization on T cells, and that among these three IFN-γ-inducible CXCR3 agonists, I-TAC is the most potent inducer of CXCR3 down-regulation.

Internalization of CXCR3 following ligand binding

To study CXCR3 internalization, cells were treated with an anti-CXCR3 mAb and the fate of CXCR3 was followed after exposure to HSVEC cells and rI-TAC. Immunofluorescence microscopy using a Texas Red-conjugated GAM revealed CXCR3 at the cell surface and in a perinuclear vesicular distribution in untreated CXCR3⁺ T cells and 300-19 cells (Fig. 3⇓). Following I-TAC exposure as well as exposure to IFN-γ-treated HSVEC, an increase in CXCR3 internalization occurred, discernible as a decrease of cell surface expression and an increase in intracellular accumulation. In contrast, the cellular distribution of CD4 was unaffected by CXCR3 ligand exposure, demonstrating the specificity of I-TAC-induced CXCR3 internalization (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

IFN-γ-treated HSVEC- and I-TAC-induced internalization of CXCR3 on T cells and CXCR3⁺ 300-19 cells. CXCR3⁺ T cells and CXCR3⁺ 300-19 cells were first incubated on ice with 10 μg/ml anti-CXCR3 mAb and then exposed to resting or IFN-γ-treated HSVEC or to I-TAC (500 ng/ml) for 30 min. Cells where then fixed, permeabilized in 0.1% saponin, and incubated with GAM-Texas Red. In control cells, CXCR3 was found on the cell surface and accumulated into perinuclear vesicles (spontaneous internalization). After exposition to IFN-γ-treated HSVEC and I-TAC, CXCR3 surface expression strongly decreased and increased into perinuclear vesicles. Bottom panels, Anti-CXCR3 mAb was omitted and cells were only incubated with GAM-Texas Red.

CXCR3 recycling following agonist exposure

To determine whether CXCR3 is recycled back to the cell surface after agonist-induced internalization, CXCR3⁺ T cells were treated for 30 min with I-TAC (250 ng/ml) and then washed into fresh medium without I-TAC (Fig. 4⇓). CXCR3 surface expression was assessed at times 0, 1, and 3 h. Following agonist stimulation, receptor recycling occurred rapidly and by 3 h, CXCR3 cell surface expression reached 79 ± 4% of its prestimulation levels.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

CXCR3 recycling on T cells. CXCR3⁺ T cells (2 × 10⁶ cells/ml) were incubated with or without I-TAC (250 ng/ml) in complete RPMI for 30 min. After two washes with HBSS prewarmed at 37°C to remove unbound chemokines, cells were resuspended in fresh prewarmed complete RPMI. CXCR3 expression was determined at baseline and at times 0, 1, and 3 h after chemokine removal by washing. Data represent the mean ± SE of four independent experiments.

Regulation of CXCR3 internalization

To begin to dissect the mechanisms responsible for chemokine-induced CXCR3 internalization, we investigated the role of different intracellular signaling pathways using selective inhibitors and activators. Activation of chemokine receptors is known to activate multiple intracellular signaling pathways, including heterotrimeric G proteins, PI3K, tyrosine kinases, and PKC. To investigate the role of PI3K and tyrosine kinases, T cells were pretreated for 2 h with wortmannin to inhibit PI3K or genestein to inhibit tyrosine kinases. The expression of CXCR3 was then determined by FACS analysis following 30 min incubation with 250 ng/ml I-TAC or IP-10 (Fig. 5⇓A). Neither wortmannin nor genestein inhibited the ability of I-TAC or IP-10 to induce CXCR3 internalization, suggesting that PI3K and tyrosine kinases are not mediating the internalization signal.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Regulation of chemokine-induced internalization of CXCR3 in T cells. CXCR3⁺ T cells (2 × 10⁶ cells/ml) were pretreated for 2 h at 37°C with (A) 1 μM wortmannin or 100 μM genestein or (B) PTX (250–1000 ng/ml). I-TAC or IP-10 (250 ng/ml) was then added for 30 min at which point CXCR3 surface expression was assessed by FACS analysis. Data represent mean ± SE of three independent experiments. C, CXCR3⁺ T cells (2 × 10⁶ cells/ml) were stimulated with Ca ionophore or PMA at indicated concentrations, then examined for CXCR3 surface expression by FACS analysis. Results are expressed as the mean ± SE of three independent experiments. D, CXCR3⁺ T cells (2 × 10⁶ cells/ml) were pretreated with 10 nM of staurosporin for 2 h prior to PMA (0.1 nM) or IP-10, Mig, or I-TAC (250 ng/ml) were added. CXCR3 surface expression was then determined. Results are expressed as the mean ± SE of four independent experiments.

In addition, because heterotrimeric G proteins of the G_i subclass have been shown to mediate CXCR3-induced chemotaxis and calcium flux responses, we studied the effects of PTX, a selective G_α1 inhibitor, on I-TAC-induced CXCR3 internalization. T cells were therefore pretreated with PTX (250, 500, or 1000 ng/ml) for 2 h at 37°C and then stimulated with I-TAC (250 ng/ml; Fig. 5⇑B). PTX pretreatment at concentrations that inhibited I-TAC- and IP-10-induced chemotaxis and calcium flux did not inhibit I-TAC- or IP-10-induced CXCR3 internalization. This suggests that agonist-induced CXCR3 internalization is not dependent on G_α1 activation.

Because other chemokine receptors, such as CXCR4 (30) and CCR3 (31) have been shown to be internalized following activation of PKC, we determined whether direct activation of PKC by PMA or calcium ionophore (A12387) would induce CXCR3 internalization. We found that both PKC activators induced a dose-dependent CXCR3 down-regulation (Fig. 5⇑C), suggesting that global activation of PKC can induce CXCR3 internalization in T cells and might be involved in agonist-induced CXCR3 internalization. Therefore, we pretreated CXCR3⁺ T cells for 2 h with 10 μM of staurosporin, a PKC inhibitor, prior to exposure to 250 ng/ml I-TAC, Mig, or IP-10 (Fig. 5⇑D). Although staurosporin completely inhibited PMA (nonagonist)-induced CXCR3 internalization, it did not affect the effect of IP-10 and Mig and very mildly affected the effect of I-TAC. Therefore, PKC activation does not seem to be involved in agonist-induced CXCR3 internalization.

Effect of heparin on ligand-induced CXCR3 internalization

Chemokines are basic proteins that bind negatively charged glycosaminoglycan molecules, such as heparin and HSPG, at physiological salt concentrations. This interaction can significantly modulate the activity of chemokines and influence chemokine binding to their receptors and subsequent signaling. In particular, IP-10 has a relatively high affinity for heparin (25 nM) and can bind to HSPG on endothelial cells. Heparin has been shown to displace an IP-10 alkaline-phosphatase fusion protein bound to endothelial cells (29). Therefore, we sought to determine whether heparin could affect the down-regulation of CXCR3 induced by incubation of T cells with HSVEC monolayers or their supernatants. To study this, HSVEC were incubated overnight with or without IFN-γ (100 ng/ml) in medium containing various concentrations of heparin (0–500 μg/ml). HSVEC were then washed to remove the heparin, and CXCR3⁺ T cells were added to theHSVEC monolayers or resuspended in HSVEC supernatants for 30 min. We found that heparin, in a dose-dependent manner, was able to very modestly decrease CXCR3 internalization on T cells incubated with HSVEC monolayers or their supernatants (13 ± 3% vs 24 ± 3%, p < 0.007, and 33 ± 5% vs 47 ± 10%, NS, for monolayers and supernatants, respectively, at 500 ng/ml heparin) (Fig. 6⇓A). These results suggest that heparin, at least at these concentrations, only moderately displaced CXCR3 ligands fromHSVEC surfaces.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Modulation of CXCR3 internalization in T cells by heparin. A, Confluent HSVEC were cultured overnight with or without IFN-γ (100 ng/ml) in complete M199 medium supplemented with 0–500 μg/ml heparin. CXCR3⁺ T cells were resuspended in HSVEC supernatants or added onto HSVEC monolayers and incubated for 30 min at 37°C, 5% CO₂. CXCR3 expression was then assessed by FACS analysis. Data represent mean ± SE of three independent experiments. ∗, p < 0.03 compared with control without heparin. B, rI-TAC and IP-10 (100 ng/ml final) were diluted in either complete RPMI medium or complete M199 medium with or without heparin (100 μg/ml). These media were then used to incubate CXCR3⁺ T cells for 30 min at 37°C and 5% CO₂. CXCR3 expression was then assessed by FACS analysis. Data represent the mean ± SE of three independent experiments. ∗, p < 0.003; ∗∗, p < 0.001 compared with control.

To further characterize the role of glycosaminoglycans, we preincubated IP-10 or I-TAC (100 ng/ml) in medium with or without heparin (100 μg/ml). Interestingly, we found that heparin significantly decreased IP-10- or I-TAC-induced internalization of CXCR3 (Fig. 6⇑B). This is in contrast to experiments using HSVEC supernatants, and suggests that IP-10 or I-TAC secreted fromHSVEC or sequestered on HSVEC surfaces might be already bound to HSPGs in such a way that does not inhibit their activity on CXCR3.

To examine whether exogenous chemokines would efficiently bind to resting HSVEC and induce CXCR3 internalization on T cells, IP-10, Mig, and I-TAC were individually added (250 ng/ml) to the medium of HSVEC for 1 h and then unbound chemokines were removed by washing. We found that I-TAC, IP-10, and Mig remained bound to HSVEC surfaces after washing because they induced CXCR3 sequestration in T cells, with a potency order (I-TAC > IP-10 > Mig) identical to soluble recombinant proteins (Fig. 7⇓). It is worth noting that the simultaneous addition of all three ligands was not significantly different from I-TAC alone and that I-TAC (250 ng/ml) was similar to the effect of IFN-γ-stimulated HSVEC monolayers. These data suggest that the level of I-TAC found in HSVEC supernatants might not be an accurate reflection of the amount of biologically active I-TAC bound to HSVEC surfaces following IFN-γ treatment.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

Exogenous chemokines bind HSVEC and induce CXCR3 internalization on T cells. rI-TAC, IP-10, or Mig (250 ng/ml), or all three were added to the medium of resting HSVEC monolayers for 1 h. After washes, CXCR3⁺ T cells were incubated with chemokine-treated HSVEC for 30 min. CXCR3 expression was then assessed by FACS analysis. Data represent the mean ± SE of four independent experiments.

I-TAC is the physiological inducer of CXCR3 internalization

Because IP-10, Mig, and I-TAC are produced to varying levels from IFN-γ-activated HSVEC (IP-10 > Mig > I-TAC) and exogenously added recombinant ligands can, to varying degrees, induce CXCR3 internalization (I-TAC > IP-10 > Mig), we sought to determine the contribution of these ligands to the ability of IFN-γ-activated HSVEC monolayers and their supernatants to induce CXCR3 internalization on T cells. This was achieved using neutralizing Abs. Anti-IP-10 and anti-Mig mAbs (50 μg/ml), which were able to completely inhibit IP-10- and Mig- (250 ng/ml) induced CXCR3 down-regulation, had no effect on HSVEC monolayer- or supernatant-induced CXCR3 down-regulation (Fig. 8⇓). In contrast, anti-I-TAC mAb (50 μg/ml), which was able to inhibit I-TAC (250 ng/ml)-induced CXCR3 down-regulation, was able to partially inhibit the ability of HSVEC supernatants from down-regulating CXCR3 expression on T cells (Fig. 8⇓). The addition of all three mAbs was no better than the I-TAC mAb alone (Fig. 8⇓). These results suggest that I-TAC is the physiological inducer of CXCR3 internalization produced by IFN-γ-activated HSVEC. The fact that anti-I-TAC did not completely inhibit HSVEC-induced CXCR3 down-regulation suggests that I-TAC bound to HSPG on HSVEC may not be completely accessible to mAb neutralization, or that the surface of HSVEC has an I-TAC independent mechanism for inducing CXCR3 internalization.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 8.

Neutralization of CXCR3 ligands reveals a preferential role for I-TAC in HSVEC-mediated CXCR3 internalization. Confluent HSVEC monolayers (A) or HSVEC supernatants (B) were incubated with anti-I-TAC, anti-IP-10, anti-Mig, or isotype control (50 μg/ml) for 1 h at 37°C. CXCR3⁺ T cells were resuspended in HSVEC supernatant (B) or added onto HSVEC monolayers (A) and incubated for 1 h at 37°C, 5% CO₂. CXCR3 expression on T cells was then assessed by FACS analysis. Data represent the mean ± SE of three independent experiments. Values of p for anti-I-TAC compound to no Ab control.

Role of CXCR3 ligands in transendothelial migration

The ability of IP-10, Mig, and I-TAC to induce transendothelial cell migration of CXCR3⁺ 300-19 cells was compared with the ability of these ligands to induce chemotaxis across bare filters. Both assays used 5 μM-pore size filters with transendothelial cell migration being performed with filters coated with HMEC-1 endothelial cells (Fig. 9⇓, A and B). Chemotaxis across uncoated filters essentially confirmed published studies, demonstrating comparable chemotaxis for all three ligands. However, in contrast, I-TAC was markedly more potent and more efficacious at inducing migration across endothelial cells. Although the presence of HMEC-1 increased the number of cells that migrated across the filter for all ligands, this effect was more pronounced for I-TAC, especially at low concentrations. For example, at 10 ng/ml I-TAC there was an ∼33-fold increase in the number of cells that migrated across HMEC-1 coated filters compared with bare filters, while HMEC-1 cells only caused an ∼5-fold increase to IP-10 and an ∼6-fold increase to Mig.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 9.

CXCR3 ligands induced transendothelial migration. Migration of CXCR3–300-19 cells through bare filters (5 μm) or through these same filters covered with a confluent monolayer of HMEC-1 was analyzed in ChemoTx 96-well chambers (NeuroProbe). After 5 h of incubation at 37°C in 5% CO₂, migration was determined by counting migrating cells in the bottom chambers containing CXCR3 ligands at indicated concentrations. Top panels, Migration through bare filters (left) or through HMEC-1 monolayers (right). Comparison between Mig, IP-10, and I-TAC. Bottom panels, Individual comparison for Mig, IP-10, and I-TAC. ⋄, +HMEC-1 cells; •, −HMEC-1 cells. Data is representative of four independent experiments and is presented as the mean ± SE of one representative experiment performed in triplicate.