α₄ Integrin-Dependent Leukocyte Recruitment Does Not Require VCAM-1 in a Chronic Model of Inflammation

Adjuvant immunization

Under light anesthetic (diethyl ether, BDH, Toronto, Canada), male Sprague Dawley rats (160–220 g) were injected s.c. at the base of the tail with a solution of heat-killed M. butyricum (Difco, Detroit, MI) in Freund’s mineral oil adjuvant (Difco; 0.75 mg of M. butyricum in 0.1 ml of adjuvant). This protocol has been used as a model of arthritis in other laboratories and was described in detail previously (14, 15, 26). Joint swelling became apparent 10–12 days after immunization, and the rats were used in this study at this time point. Previous experiments using intravital microscopy revealed a tremendous increase in leukocyte trafficking through mesenteric postcapillary venules 4–20 days after immunization (7, 8, 26).

Intravital microscopy

Rats were maintained on a purified laboratory diet and fasted for 18–24 h before surgery. Animals were anesthetized with an i.p. injection of sodium pentobarbitol (55 mg/kg body weight). The right jugular vein was cannulated to maintain anesthesia, and the right carotid artery was cannulated to measure systemic arterial blood pressure (model P23XL pressure transducer, Viggo-Spectramed, Oxnard, CA; model 7 physiologic recorder, Grass Instruments, Quincy, MA). Following laparotomy, rats were placed in a supine position on an adjustable Plexiglass microscope stage, and a segment of the midjejunum was exteriorized and prepared for intravital microscopy as previously described (7, 8, 9, 10, 11, 26).

The mesenteric preparation was observed through an intravital microscope (Optiphot-2, Nikon, Mississauga, Canada) with a ×25 objective lens (Wetzlar L25/0.35, Leitz, Munich, Germany) and a ×10 eyepiece. A video camera (model 5100 HS, Panasonic, Osaka, Japan) mounted on the microscope projected the image onto a color monitor (model PVM 2030, Sony, Tokyo, Japan), and the images were recorded using a videocassette recorder (model AG-1790, Panasonic) for subsequent playback analysis. The final magnification of the image on the monitor was ×1800. Single unbranched mesenteric venules (30–50 μm in diameter) were selected for study. The same section of venule was observed throughout the experiment to control for variations between different regions. Venular diameter was measured on-line using a video caliper (Microcirculation Research Institute, Texas A & M University, College Station, TX). Centerline RBC velocity was also measured on-line using an optical Doppler velocimeter (Microcirculation Research Institute, Texas A & M University).

The number of rolling and adherent leukocytes was determined off-line during video playback analysis. Leukocytes were considered adherent to the venular endothelium if they remained stationary for a period of time equal to or exceeding 30 s. Adhesion is expressed as cells per 100 μm of venule length. Rolling leukocytes were defined as those white blood cells that moved at a velocity less than that of erythrocytes within a given vessel. The flux of rolling leukocytes was determined as the number of white blood cells that rolled past a fixed point in the venule during a 1-min interval using frame-by-frame video analysis.

Experimental protocols

Leukocyte trafficking was examined in mesenteric postcapillary venules of naive and M. butyricum-immunized animals (12 days postimmunization). Anti-adhesion molecule therapies were administered following two baseline recordings (at the 20 min point) so that each animal could serve as its own control for effects of treatments on leukocyte rolling flux and adhesion. In some experiments animals were treated with a polyclonal rabbit serum against rat VCAM-1 (0.5 ml/rat i.v.), a mAb against rat VCAM-1 (3 mg/kg i.v.; 5F10) (22, 23), a mAb against the rat α₄ integrin (4 mg/kg i.v.; TA-2) (7), a mAb against the rat β₇ integrin (4 mg/kg i.v.; TA-6), or a mAb against the mouse β₁ integrin (2 mg/kg i.v.; HMβ1-1, PharMingen Canada, Mississauga, Canada). In preliminary experiments the β₁ integrin Ab was found to block the β₁ integrin, but induced β₂ integrin (CD18)-dependent adhesion. Therefore, experiments were performed using β₁ integrin Ab plus a blocking β₂ integrin mAb (2 mg/kg i.v.; WT-3). On its own, WT-3 blocks chemokine-induced adhesion in normal animals (27), but does not reduce adhesion in mesenteric venules of adjuvant-immunized rats (7, 8) or enhance the effect of the α₄ integrin or β₇ integrin mAbs (data not shown). In another experiment a polyclonal Ab directed against the connecting segment-1 (CS-1)³ portion of fibronectin (0.1 and 0.3 ml/rat) was given to immunized rats. All these Abs were used at concentrations that were previously reported to inhibit leukocyte recruitment. Nevertheless, to confirm that suprasaturating concentrations of Ab were used, and that the concentration of Ab was sufficient to block binding to the targeted ligand (VCAM-1 or CS-1), whole blood flow chamber assays were used as described below.

In other experiments rats were treated with a phosphorothioate antisense oligodeoxynucleotide (ISIS 18155, a gift from Dr. C. F. Bennett, ISIS Pharmaceuticals, Carlsbad, CA), which blocks translation of rat and mouse VCAM-1 mRNA (C. F. Bennett, unpublished observation). The sequence of ISIS 18155 is 5′-CGCGACCATCTTCACAGGCA-3′. ISIS 18155, or a scrambled control oligonucleotide, ISIS 18154, was given at 50 mg/kg i.v. 16–24 h before intravital microscopy. Antisense oligonucleotides given in this manner were found to be effective in murine and rat models of inflammation (28, 29). The dose of antisense in this study prevented significant VCAM-1 expression.

Whole blood flow chamber assay

CHO cells stably transfected with rat VCAM-1 (a gift from Dr. R. R. Lobb, Biogen, Cambridge, MA) were grown to confluence (2 days) on glass coverslips (Fisher Scientific, Ottawa, Canada) using Ham’s F-12 medium (Life Technologies, Grand island, NY) with 10% FBS (Life Technologies). In other experiments glass coverslips were coated with 5 μg/ml soluble recombinant human VCAM-1 (a gift from Dr. R. R. Lobb, Biogen) and incubated for 2 h at 37°C. It has been shown previously that rat α₄ integrins will bind to human VCAM-1 (30, 31). Coverslips were then incubated with 1% BSA (Sigma, St. Louis, MO) for an additional 2 h (37°C) to block nonspecific leukocyte interactions with glass. Coverslips were assembled into parallel plate flow chambers (2, 32) and washed with HBSS (Life Technologies) before experiments. Conditions were maintained at 37°C using a warm air cabinet surrounding the flow chamber.

Whole blood was isolated from normal and adjuvant-immunized rats by cardiac puncture, treated with heparin (30 U/ml) to prevent clotting, and maintained at 37°C in a water bath. Some animals had been treated with Abs against VCAM-1 or the α₄ integrin before blood isolation (see above). A syringe pump (model 22, Harvard Apparatus, South Natick, MA) was used to draw blood across VCAM-1-coated coverslips or over CHO cells expressing rat VCAM-1 at a defined shear force of 10 dyn/cm². Whole blood was perfused over coverslips for 5 min, followed by perfusion with HBSS as previously described (32). Within 20 s of perfusion with HBSS, the RBC cleared and leukocytes could be observed interacting with the surface of the coverslip. Leukocyte interactions were examined using an inverted phase contrast microscope (Axioskop, Carl Zeiss Canada, Don Mills, Canada) with a ×10 objective. Images were projected onto a video monitor using a CCD camera and were videotaped for subsequent analysis. Four different fields of view were each recorded for 15 s. The numbers of rolling and adherent leukocytes were determined. Rolling leukocytes could be distinguished from those in the bulk flow by phase contrast and their reduced velocity. In these experiments leukocytes were defined as adherent if they remained stationary for a period ≥10 s (33).

To distinguish the types of leukocytes that were interacting with VCAM-1, the coverslips were gently removed from the flow chamber and stained using a Wright-Giemsa staining kit. This procedure resulted in minimal loss of interacting leukocytes as determined by comparing the number of leukocytes on the coverslip before removal from the chamber with that after the staining procedure (32).

Flow cytometry

Leukocytes were isolated from citric acid, sodium citrate, and dextrose (ACD) anticoagulated rat blood by a 45-min sedimentation over 4% dextran (250,000 m.w.; Spectrum, Gardena, CA) followed by hypotonic lysis of contaminating RBC. Cells were diluted to 1 × 10⁶/tube and incubated with an Ab against the α₄ integrin (TA-2; 1 μg/tube) at room temperature for 30 min. Cells were washed and stained with a FITC-labeled goat anti-mouse IgG (Becton Dickinson, Mississauga, Canada) for 30 min. After washing, the level of α₄ integrin expression was measured using a FACScan flow cytometer (Becton Dickinson). Neutrophils and lymphocytes were differentiated based on forward and side scatter characteristics.

Radioiodination of mAbs for VCAM-1 expression

The binding (5F10) and nonbinding (P-32) mAbs were labeled with ¹²⁵I and ¹³¹I (DuPont-NEN, Boston, MA), respectively, using the Iodogen method previously described (34). Untreated M. butyricum-immunized rats, and M. butyricum-immunized rats treated with ISIS 18155 as previously described were anesthetized, and the carotid artery and jugular vein were cannulated. A mixture of ¹²⁵I-labeled 5F10 and a fixed dose of ¹³¹I-labeled nonbinding mAb was administered through the jugular vein catheter. Exactly 5 min after injection of the mAb mixture a blood sample was obtained from the carotid artery. Immediately thereafter, the animals were heparinized (40 U of sodium heparin) and rapidly exsanguinated by perfusion of bicarbonate-buffered saline through the jugular venous catheter with simultaneous blood withdrawal through the carotid arterial catheter. This was followed by perfusion of 60 ml of buffered saline through the carotid artery after severing the inferior vena cava at the thoracic level. The mesentery was harvested and weighed.

The method of calculating VCAM-1 expression has been described previously (34). Briefly, the ¹²⁵I binding and ¹³¹I nonbinding mAb activities in the mesentery and in 100-μl samples of cell-free plasma were measured. A 3-μl aliquot of the radiolabeled mAb mixture was assayed to determine the total injected activity of each labeled mAb. The accumulated activity of each mAb in the mesentery was expressed as a percentage of the injected activity per gram of tissue. VCAM-1 expression was calculated by subtracting the accumulated activity per gram of tissue of the nonbinding mAb (¹³¹I-labeled P-23) from the activity of the binding VCAM-1 (¹²⁵I-labeled 5F10). This value, expressed as a percentage of the injected dose per gram of tissue, was converted to nanograms of mAb per gram of tissue by multiplying the above value by the total injected binding mAb.

Statistical analysis

Data are presented as the mean ± SEM. Means were compared using Student’s t test with Bonferroni’s correction for multiple comparisons where appropriate. Statistical significance was set at p < 0.05.

The α₄β₁ integrin mediates leukocyte recruitment in vivo

In Fig. 1⇓ (A and B) the data reveal that on day 12 ∼250–300 cells rolled/min and 30–40 cells adhered/100 μm of venule length in rats immunized with M. butyricum 12 days earlier. These values are at least 10-fold higher than those in control untreated animals (data not shown). Treatment with an α₄ integrin mAb (4 mg/kg i.v.; TA-2) significantly reduced both the leukocyte rolling flux (∼55%) and leukocyte firm adhesion (∼75%) observed in mesenteric postcapillary venules of adjuvant-immunized rats. Previous work has demonstrated that the remaining 50% of leukocyte rolling is L-selectin dependent (7). Treatment with an Ab against the β₁ integrin (2 mg/kg i.v.; HMβ1-1) caused a similar reduction in leukocyte rolling flux (Fig. 1⇓). The inset in Fig. 1⇓ (top) demonstrates the temporal kinetics of β₁ integrin Ab inhibition; leukocyte rolling was ∼300 cells/min in adjuvant-treated rats and decreased to <150 cells/min following β₁ integrin Ab administration. As we were unable to find a β₁ integrin Ab that did not activate CD18 on rat leukocytes, we inhibited the CD18-dependent adhesion when testing the role of β₁ integrin. It should be noted that the mAb to CD18 did not block adhesion on its own (Fig. 1⇓). In the presence of a CD18 mAb (2 mg/kg i.v.; WT-3), the β₁ integrin mAb reduced adhesion by about 55%. In direct contrast, an Ab against the β₇ integrin (4 mg/kg i.v.; TA-6) did not reduce leukocyte rolling or adhesion alone (Fig. 1⇓) or in the presence of a CD18 Ab (data not shown). These data suggest that the α₄ integrin-dependent leukocyte recruitment in adjuvant-immunized rats is dependent on very late activation Ag-4 (α₄β₁) rather than α₄β₇ integrin.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Role of the α₄β₁ and α₄β₇ integrins in the leukocyte rolling (A) and adhesion (B) in mesenteric postcapillary venules of M. butyricum-immunized (day 12) rats. After a baseline control period of 20 min, rats were treated with Abs against α₄ integrin (TA-2; 4 mg/kg i.v.; n = 5), β₁ integrin (HMβ1-1; 2 mg/kg i.v.; n = 3), β₇ integrin (TA-6; 4 mg/kg i.v.; n = 4), and/or CD18 (WT-3; 2 mg/kg i.v.; n = 5). The administration protocol is illustrated in the inset; the administration of HMβ1-1 at 20 min significantly reduced leukocyte rolling flux at later time points. Time points after Ab administration were compared with the initial baseline time points within each animal. ∗, p < 0.05 compared with time zero; †, p < 0.05 compared with untreated day 12 immunized animals.

VCAM-1 does not mediate α₄ integrin-dependent leukocyte recruitment in vivo

As VCAM-1 is the major ligand for very late activation Ag-4 in vitro (4, 5, 6, 17, 18, 19), animals were treated with a function-blocking mAb against rat VCAM-1 (3 mg/kg i.v.; 5F10). Surprisingly, this treatment did not reduce leukocyte rolling or adhesion (Fig. 2⇓). As multiple VCAM-1 epitopes have been shown to mediate α₄ integrin/VCAM-1 interactions (17, 18, 19, 35), a polyclonal serum against rat VCAM-1 (0.5 ml/rat) was also tested. As with the mAb, this treatment had no effect on leukocyte rolling or adhesion (Fig. 2⇓). It should be noted that in the same animals in which inhibition of VCAM-1 had no effect on leukocyte rolling in the microvasculature of an adjuvant-immunized rat, the Ab directed against α₄ integrin was again able to inhibit the rolling of >150 cells/min (not shown). In additional experiments, VCAM-1 inhibition with an antisense oligonucleotide failed to reduce leukocyte trafficking in vivo (Fig. 2⇓).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

The role of VCAM-1 in leukocyte rolling (A) and adhesion (B) in mesenteric postcapillary venules of M. butyricum-immunized (day 12) rats. After a baseline control period of 20 min, rats were treated with an mAb against VCAM-1 (5F10; 3 mg/kg i.v.; n = 4), or a polyclonal serum against rat VCAM-1 (0.5 ml/rat i.v.; n = 4). Other rats were pretreated 16–24 h before the experiment with a phosphorothioate antisense oligonucleotide against VCAM-1 (ISIS 18155; 50 mg/kg i.v.; n = 4) or a scrambled control oligonucleotide (ISIS 18154; 50 mg/kg i.v.; n = 4). The administration protocol is illustrated in the inset. The administration of 5F10 at 20 min did not alter leukocyte rolling flux at later time points. Time points after Ab administration were compared with the initial baseline time points within each animal. ∗, p < 0.05 compared with time zero; †, p < 0.05 compared with untreated day 12 immunized animals.

The lack of effect with VCAM-1 Ab treatments in vivo was not due to insufficient doses of Ab, as blood isolated from these animals had sufficient Ab to inhibit leukocytes from interacting with rat VCAM-1-transfected CHO cells under flow conditions in vitro. Fig. 3⇓ shows that ∼450 leukocytes from adjuvant-immunized rats interacted with rat VCAM-1-transfected CHO cells in the flow chamber. Leukocytes from adjuvant-immunized rats treated with an anti-α₄ integrin Ab (Ab given in vivo) were not able to tether to and roll on VCAM-1, suggesting that VCAM-1 can indeed function as a ligand for α₄ integrin in the rat system. Moreover, animals that had been treated with the function-blocking mAb against rat VCAM-1 had sufficient plasma Ab titers that leukocytes from these animals were not able to interact with rat VCAM-1 in vitro (Fig. 3⇓). Similarly, sufficient quantities of a polyclonal serum against rat VCAM-1 were used to prevent any subsequent interactions of leukocytes with VCAM-1 from antiserum-treated animals in vitro. Clearly, saturating concentrations of functionally blocking VCAM-1 Abs were used in the in vivo experiments.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Leukocyte recruitment onto rat VCAM-1 under flow conditions in vitro. Whole blood from adjuvant-immunized rats was perfused over monolayers of CHO cells transfected with rat VCAM-1. Some animals had been pretreated with mAbs against the α₄ integrin (TA-2; 4 mg/kg i.v.; n = 2) or VCAM-1 (5F10; 3 mg/kg i.v.; n = 3), or a polyclonal serum against rat VCAM-1 (0.5 ml/rat i.v.; n = 3). Total leukocyte interactions (rolling and adhesion) were quantitated after 5 min of perfusion. ∗, p < 0.05 compared with untreated day 12 immunized animals.

VCAM-1 is expressed in adjuvant-immunized microvasculature

Fig. 4⇓ demonstrates that there was a very small constitutive amount of VCAM-1 in the mesentery of animals that were not immunized with M. butyricum. By contrast, VCAM-1 expression was increased >10-fold in this vascular bed in rats immunized 12 days earlier with M. butyricum. This increase was ∼50% of the response seen with optimal levels of LPS administration (1.88 ng of mAb/g), suggesting a significant amount of VCAM-1 expression in the mesentery of adjuvant-immunized rats. Fig. 4⇓ also demonstrates that ISIS 18155 antisense inhibited VCAM-1 synthesis, as VCAM-1 expression was reduced by >60% in the mesentery of adjuvant-immunized rats.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

VCAM-1 expression in mesenteric vessels of healthy nonimmunized rats (n = 4) and rats immunized with M. butyricum 12 days (n = 5) before measurement of adhesion molecule expression. Also shown is VCAM-1 expression in rats immunized with M. butyricum receiving ISIS 18155 (n = 6) antisense 16–24 h before measurement of VCAM-1 expression. ∗, p < 0.05 compared with untreated healthy control rats.

Neutrophils from adjuvant-immunized rats bind VCAM-1 in vitro

In vitro whole blood experiments were also used to examine the populations of leukocytes attaching to VCAM-1. In Fig. 5⇓A it can be seen that equivalent numbers of leukocytes from naive and adjuvant-immunized rats were recruited onto VCAM-1. However, when leukocytes on the coverslips were stained with a Wright-Giemsa staining kit for identification of cell types, differences in recruited cell types were observed between control and adjuvant-immunized animals (Fig. 5⇓B). Neutrophil and monocyte recruitment from blood of adjuvant-immunized rats were significantly increased compared with those in naive control rats. In contrast, the number of lymphocytes recruited from adjuvant-immunized rats was significantly reduced compared with that in control animals. The recruitment of neutrophils was confirmed using an esterase staining kit (data not shown). To determine whether these differences in recruitment were due to the large increase in the number of neutrophils and monocytes in the blood of adjuvant-immunized rats (Table I⇓), the recruited cell populations were normalized for differences in circulating numbers of leukocytes. Fig. 5⇓C shows that the increased neutrophil recruitment and reduced lymphocyte recruitment in blood from adjuvant-immunized rats were still apparent despite normalization for circulating numbers of cells, while the increased recruitment of monocytes was entirely accounted for by the increase in circulating monocyte number. These results suggest that there may be differences in the activation state and/or adhesion molecule expression of circulating neutrophils and lymphocytes in adjuvant-immunized rats.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Differential leukocyte recruitment onto purified VCAM-1 under flow conditions in vitro. Whole blood from naive and adjuvant-immunized rats was perfused over coverslips coated with purified recombinant VCAM-1. Total leukocyte interactions (rolling and adhesion) were quantitated after 5 min of perfusion (A). Flow chambers were dismantled, and coverslips were stained with a Wright-Giemsa staining kit to determine leukocyte differentials (B). Leukocyte differentials were corrected for differences in leukocyte distribution (C) as shown in Table I⇓. ∗, p < 0.05 compared with naive control rats.

α₄ integrin expression on neutrophils is not altered in adjuvant-immunized rats

Flow cytometry was used to examine expression of the α₄ integrin on leukocytes from control and adjuvant-immunized rats. As reported previously (31, 36), α₄ integrin was detected on the surface of circulating rat neutrophils isolated from untreated control animals (Fig. 6⇓). However, neutrophils from adjuvant-immunized rats did not exhibit an increase in the expression of α₄ integrin compared with that in control animals (Fig. 6⇓). The lymphocyte results were more equivocal, as a significant proportion of lymphocytes in adjuvant-immunized animals expressed a similar level of α₄ integrin as nonimmunized rats, but a significant proportion of lymphocytes expressed less α₄ integrin. This may account for the reduced lymphocyte recruitment onto VCAM-1 in vitro.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Expression of α₄ integrin on leukocytes from naive and adjuvant-immunized rats. Mean fluorescence with α₄ integrin mAb (TA-2) was increased in naive (dotted line) and adjuvant-immunized (thick line) rats. There was no difference in the binding of an α₄ integrin Ab to neutrophils and only a small reduction of binding to lymphocytes from naive or adjuvant-immunized rats. Plots are representative of five separate experiments.

CS-1 is not the ligand for α₄ integrin

In a final series of experiments, we examined the possibility that the CS-1 portion of fibronectin might be the ligand for the α₄ integrin. This peptide has been detected on the luminal surface of activated endothelium and mediates binding of monocytes in static adhesion assays (37). However, an Ab directed against the CS-1 portion of fibronectin had no effect on leukocyte recruitment when used in vivo at 0.1 or 0.3 ml/rat (Fig. 7⇓, A and B). When whole blood from adjuvant-immunized animals was perfused over a fibronectin-40 fragment containing the CS-1 region, no interactions were noted, suggesting that leukocytes cannot interact with CS-1 under flow conditions (data not shown). However, many leukocytes adhered to the CS-1-containing fragment when the perfusion was stopped, and leukocytes were allowed to settle (Fig. 7⇓C). Leukocytes from an animal treated with the anti-CS-1 Ab displayed no interactions with fibronectin-40 under these static conditions (Fig. 7⇓C), suggesting that ample Ab was used in vivo.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

The effects of an Ab against the CS-1 portion of fibronectin on leukocyte rolling (A) and adhesion (B) in vivo and adhesion of leukocytes to fibronectin-40 fragment in vitro (C). Whole blood from treated and untreated animals was perfused into a flow chamber containing the CS-1 portion of fibronectin and then stopped for 3 min before flow was resumed. Due to limitations of the amount of Ab and identical results for two different concentrations, the data are grouped.