Inhibition of Leukocyte Sticking and Infiltration, but Not Rolling, by Antibodies to ICAM-1 and LFA-1 in Murine Endotoxin–Induced Uveitis

Abstract

purpose. Cell-adhesion molecules are critical elements in intravascular rolling and sticking of leukocytes during acute inflammation. In this process, selectins are thought to be involved in initial adhesion and rolling, and integrin–Ig superfamily interactions are believed primarily to mediate stronger adhesion and transendothelial migration. This study clarifies the role of two adhesion molecules, intercellular adhesion molecule (ICAM)-1 and leukocyte functional antigen (LFA)-1, in endotoxin-induced uveitis (EIU).

methods. Intravital microscopy was used to record the movement and location of leukocytes in the irises of mice with uveitis induced by intravitreal injection of 250 ng Escherichia coli endotoxin. Each mouse concurrently received an intraperitoneal injection of monoclonal neutralizing antibodies for ICAM-1, LFA-1, or both or control irrelevant antibodies.

results. Mice treated with endotoxin and control antibodies had an inflammatory response that was clearly present at the 6- and 24-hour time points and was mostly resolved by 48 hours. Mice that received anti-ICAM-1 or anti-LFA-1 had significantly fewer cells infiltrating their irises at 6 and 24 hours. Detailed analysis of the 6-hour time point recordings revealed that neither anti-ICAM-1 nor anti-LFA-1 significantly reduced the number of leukocytes rolling on venule endothelial surfaces, but the treatments reduced the number of firmly adherent cells.

conclusions. These data confirm previous reports that ICAM-1 and LFA-1 are important mediators of EIU. The dynamic in vivo images clearly support the hypothesis that integrin-mediated cell adhesion is more critical for the firm adhesion of sticking cells than for leukocyte rolling.

Endotoxin-induced uveitis (EIU) is a model for acute inflammation with leukocytes, primarily neutrophils and monocytes, leaving iris venules and entering the surrounding tissues. ^{1 2} Leukocytes will roll and stick to the luminal surface of venule endothelial cells before extravasation. ³ These processes are controlled by adhesion molecules on the cells’ surfaces. Members of the carbohydrate-binding selectin family are primarily involved in leukocyte rolling and members of the integrin family, such as leukocyte functional antigen (LFA)-1, and their immunoglobulin superfamily counterreceptors, such as intercellular adhesion molecule (ICAM)-1, are believed to be primarily responsible for the firm adhesion of sticking cells. Some recent reports have suggested that integrin-mediated interactions may also have a role in leukocyte rolling. ^{4 5} We and others have documented the importance of these adhesion molecules in EIU in mice, rats, and rabbits. ^{6 7 8 9 10 11 12}  

In the present study, we used intravital microscopy to extend our knowledge about the role of cell-adhesion molecules, specifically ICAM-1 and one of its counterreceptors, LFA-1, in murine EIU. By visualizing leukocytes in and around iris venules of living mice that have been treated with neutralizing antibodies directed against a specific cell-adhesion molecule, we can clarify the role of these molecules in specific steps of the dynamic process of leukocyte migration. 

Materials and Methods

Animals

Female BALB/c mice (5–6 weeks old) were purchased from Jackson Laboratories (Bar Harbor, ME) and fed standard laboratory chow and water ad libitum. All procedures were in accordance with the ARVO Statement for Use of Animals in Ophthalmic and Visual Research. 

Induction of Uveitis and Experimental Design

Baseline intravital microscopy recordings were made of vessels on the day before induction of uveitis. Uveitis was induced by intravitreal injection of 2 μl containing 250 ng endotoxin (lipopolysaccharide; LPS) from Escherichia coli 055:B5 (List Biological Laboratories, Campbell, CA) and 0.25% human serum albumin (Baxter Healthcare, Glendale, CA) in saline. ¹³ At the same time, the mice received an intraperitoneal injection of experimental monoclonal or control polyclonal antibodies. One group of mice received 200 μl with 200 μg of clone YN-1 rat IgG2b antibody against mouse ICAM-1. ¹⁴ A second group received 200 μl with 200 μg of clone M17 rat IgG2a antibody against mouse LFA-1. ¹⁵ The third group received 200 μg of both anti-ICAM-1 and anti-LFA-1 antibodies. Similar doses of these antibodies have been reported to inhibit experimental autoimmune uveitis, systemic EIU, and experimental allergic conjunctivitis. ^{10 16 17} Control mice received 200μ l with 200 μg polyclonal rat IgG2b (PharMingen, San Diego, CA). Experimental intravital microscopy recordings were made 6, 24, and 48 hours after LPS injection. 

Intravital Microscopy

Details of the microscopy technique have been published previously. ¹³ Approximately 10 minutes before each time point, mice received an intravenous injection of 35 mg/kg rhodamine 6G (Sigma Chemical Co, St. Louis, MO) in phosphate-buffered saline to label circulating leukocytes. Mice were anesthetized with 1 l/min 1.7% isoflurane (Ohmeda, Liberty Corner, NJ) in oxygen and placed on the stage of an epifluorescence microscope. Pupils were constricted with topical pilocarpine. A viscous gel (Vidisic Gel, Dr. Mann Pharma, Berlin, Germany) was applied between the cornea and aqueous immersion objectives. Images were captured (NTSC video format) with a black-and-white camera coupled to an image intensifier (CF 8/4 NNIR; Kappa, Gleichen, Germany) and recorded by computer (Power PC; Apple, Cupertino, CA) with a video board (Ignitor; Aurora Video Systems, Shelby Township, MI). For each eye, five regions of interest containing one to three vessels were recorded for 20 seconds. To avoid operator bias, these regions were determined by a fixed grid pattern beginning at the first branching of an iris arteriole from the major iridal circle at 9 o’clock on the iris periphery. At the baseline time point, vessels were mapped to ensure that the operator returned to exactly the same segments for subsequent recordings. At least 10 vessels were recorded for each eye, but some were excluded from analysis if off-line analysis was compromised because of eye movement. 

Quantification of Visual Data

Details of the quantitative analysis have been published previously. ¹³ Measurements were made within 100- to 300-μm venule segments that did not branch or appreciably change diameter. The mean of five measurements equally spaced over the length of the vessel was used as the vessel’s internal diameter for surface area calculations. Rolling leukocytes were defined as cells whose contact with the vessel wall made the cell move considerably slower than the mean blood velocity and are reported as the number of rolling cells per square millimeter of endothelial surface per minute. Sticking leukocytes were defined as cells firmly attached to vessel walls and immobile for the 20-second observation period. Sticking cells are reported as the number of cells per square millimeter of endothelial surface. The categories of rolling and sticking cells are mutually exclusive. Unstained, freshly emigrated leukocytes were clearly seen as round, dark dots in the perivascular iris tissue and are reported as the number of infiltrating cells per square millimeter of iris tissue. 

Statistical Analysis

The groups of values for rolling and sticking leukocytes in each venule were subjected to the Mann-Whitney rank sum test, because they did not usually follow a normal distribution. The Mann-Whitney rank sum test was also used to evaluate the numbers of infiltrating cells and the χ² test was used for comparing the percentages of venules with rolling or sticking leukocytes. Differences with P < 0.05 were considered statistically significant. 

Results

Intravitreal injection of endotoxin causes an acute inflammatory response that can be visualized by intravital fluorescence microscopy of the iris. The importance of integrin-mediated cell adhesion was evaluated by treating mice with neutralizing antibodies for ICAM-1, LFA-1, or both at the time of endotoxin injection. Control mice received isotype-matched, nonspecific polyclonal antibodies. Large numbers of cells infiltrating the irises of control mice were evident at 6 and 24 hours after endotoxin injection (Figs. 1 and 2) . By 48 hours after injection, the numbers of infiltrating cells had nearly returned to baseline values. Mice receiving either neutralizing antibody had a blunted inflammatory response with only approximately one fifth the number of infiltrating cells as the control animals at both the 6- and 24-hour time points. Treatment with a combination of both antibodies was not statistically different from the effect of the separate antibodies. These results indicate that ICAM-1 and one of its counterreceptors, LFA-1, are both important in the inflammatory response to intravitreal endotoxin and are in accord with previous studies of EIU. ^{7 8 9 10 12}  

Examination of the recordings obtained by intravital fluorescence microscopy allows an analysis of leukocyte rolling and sticking, two critical steps in inflammation. Examples of dynamic images are posted on our web site (http://www.ohsuhealth.com/cei/research/inflammatory.asp). Neither rolling nor sticking can be quantified by conventional histology. Based on the general model that leukocyte rolling is primarily mediated by selectins and that firm attachment before extravasation is primarily mediated by integrins, it might be predicted that neutralizing antibodies against ICAM-1 and LFA-1 would affect leukocyte sticking more than rolling. This is indeed what we found. The data in Table 1 show that all treatment groups had more rolling and sticking cells at 6 hours after endotoxin injection than at baseline. When all venules were included in the statistical analysis, there was no significant difference in the number of rolling cells between the control antibody group and any of the neutralizing-antibody–treated groups. In contrast, all three treated groups had significantly fewer sticking cells than the control animals (Table 2) . 

Our previous studies have shown that only a fraction of the venules in an iris appear to participate actively in the inflammatory response to intravitreal endotoxin. Therefore, a reduction in the number of participating venules may also contribute to a reduced number of infiltrating cells. In the four treatment groups, the percentage of venules with rolling cells ranged from 1% to 9% at baseline (Table 1) . By 6 hours, the percentage of venules with rollers had significantly (P < 0.001) increased to 36% to 47%. The differences among the treated groups were not significant. Therefore, administration of the neutralizing antibodies does not appear to interfere with the initial activation of venule endothelial cells. 

Applying the same analysis to venules with sticking cells again demonstrates the effectiveness of the anti-ICAM-1 and anti-LFA-1 antibodies in blocking leukocyte sticking. At baseline, only 2% to 5% of venules had at least one sticking cell (Table 2) . By 6 hours after endotoxin, 66% of venules in the control group had sticking cells. The experimental groups all had significantly fewer venules with sticking cells (P ≤ 0.005). 

Discussion

Our findings are in accord with previous studies indicating that ICAM-1 and LFA-1 are key effector molecules in EIU. ^{7 8 9 10 12} These data also add support to the argument that Ig superfamily-integrin–mediated cell adhesion is more critical for the firm adhesion of sticking cells than for leukocyte rolling. The incomplete blockage of cell sticking and infiltration by the neutralizing antibodies is qualitatively in accord with our previous report that mice genetically deficient in either ICAM-1 orβ ₂-integrin (one subunit of LFA-1) remain susceptible to uveitis from intravitreal endotoxin, although the resultant disease is less severe than in control animals. These data indicate that additional adhesion molecules are involved or able to compensate for the loss of LFA-1 and ICAM-1 function. Other possibilities include the integrin Mac-1, which binds both ICAM-1 and ICAM-2, the latter of which we recently found to be expressed by iris vascular endothelial cells (Silverman et al., manuscript in preparation). 

The anti-ICAM-1 and anti-LFA-1 antibodies may block leukocyte infiltration to a greater extent than leukocyte sticking. All the eyes from mice receiving either of these antibodies had fewer infiltrating cells per square millimeter of iris than the median count in the control animals, and most were less than the 25th percentile in the control animals. In contrast, 11% to 24% of the venules in the experimental groups had more sticking leukocytes per square millimeter of venule surface than the control median. This discrepancy raises the question of whether these two adhesion molecules have critical roles in addition to trapping circulating cells on the endothelium in areas of active inflammation. Some possibilities are participation in the extravasation process itself and reduction of leukocyte migration from the perivascular tissues. ICAM-1 is expressed on leukocytes as well as the vascular endothelium. ^{18 19} Direct binding of anti-ICAM-1 to the leukocytes may have an effect on cell migration. 

In summary, intravital microscopy is useful in defining the role of adhesion molecules in endotoxin-induced eye inflammation. Intravital microscopy is superior to other techniques, such as conventional histology. For example, in a previous publication we showed thatβ ₂-integrin and perhaps ICAM-1 gene knockout mice had reduced cell infiltration in intravitreal EIU. ¹² The present study goes further and allows the stepwise process of leukocyte migration to be dissected by demonstrating a major effect on sticking and not on rolling. In designing optimal pharmacotherapy for anterior uveitis, combining an inhibitor of rolling with an inhibitor of sticking may be synergistic, whereas two drugs that each blocked sticking may not show an additive response. Dynamic visualization of the inflammatory response is an invaluable tool in clarifying the process of inflammation. 

²  Present affiliation: University Augenklinik, Heidelberg, Germany.

 Supported by National Institutes of Health Grants EY06484, EY06477, and EY13093; Research to Prevent Blindness (SRP, JTR); the Casey Eye Institute; and Deutsche Forschungsgemeinschaft (MDB).

 Submitted for publication February 23, 2001; revised May 31, 2001; accepted June 11, 2001.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked“ advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Intravital microscopy of iris tissues. The line in each panel represents 100 μm. (A) Before treatment only a few refractile cells were seen in the perivascular tissue (arrows). These cells were not fluorescent and were visualized by transillumination of the iris by light reflecting from the retina. (B, C) Six hours after induction of uveitis by endotoxin, an influx of infiltrating leukocytes was apparent by the many perivascular dots. (C, arrows) Fluorescent intravascular rolling or sticking cells. The dynamics of rolling and sticking can be appreciated only on video, examples of which can be seen on our web site (http://www.ohsucasey.com/research/inflammatory.asp). (D) An iris 6 hours after injection of endotoxin with antibodies to both ICAM-1 and LFA-1 showed fewer intravascular adherent cells and fewer perivascular infiltrating cells than with control antibodies.

Figure 2.

 

Antibodies to ICAM-1 and LFA-1 inhibit cell infiltration. The density of infiltrating cells is shown in irises before treatment (Baseline) and at 6, 24, and 48 hours after injection of endotoxin with control, anti-ICAM-1, anti-LFA-1, or both anti-ICAM-1 and anti-LFA-1 antibodies. The boxes indicate the 25th and 75th percentiles, the whiskers indicate the 10th and 90th percentiles, and the central lines show the medians for each time point.* P < 0.05, **P < 0.01,*** P ≤ 0.001 compared with the control antibody group.

Table 1.

 

Rolling Cells

Time Point	Antibody Specificity	Eyes (n)	Venules (n)	Rolling Cells*					Venules with Rolling Cells
								Mean	75th Percentile	P , †	Percent	P , †
Baseline	Control	7	59	17	0		1.7
	ICAM-1	9	87	16	0		1.1
	LFA-1	9	73	30	0		5.5
	ICAM-1 and LFA-1	7	57	50	0		8.8
Six hours	Control	7	58	588	863		47
	ICAM-1	8	75	293	361	0.11	36	0.29
	LFA-1	9	87	543	756	0.48	38	0.39
	ICAM-1 and LFA-1	7	54	414	844	0.59	43	0.82

*  Data are number of rolling cells per square millimeter of vessel surface.

†  Compared with control antibody group.

Table 2.

 

Sticking Cells

Time Point	Antibody Specificity	Eyes (n)	Venules (n)	Sticking Cells*					Venules with Sticking Cells
								Mean	75th Percentile	P , †	Percent	P , †
Baseline	Control	7	59	6.3	0		5.1
	ICAM-1	9	87	1.5	0		2.3
	LFA-1	9	73	4.9	0		2.7
	ICAM-1 and LFA-1	7	57	4.2	0		3.5
Six hours	Control	7	58	401	625		66
	ICAM-1	8	75	104	75	<0.001	27	<0.001
	LFA-1	9	87	277	248	0.004	39	0.003
	ICAM-1 and LFA-1	7	54	108	154	<0.001	37	0.005

*  Data are number of sticking cells per square millimeter of vessel surface.

†  Compared with control antibody group.

The authors thank Son Nguyen, My Cao, and Thao Bui for their assistance in the quantitative measurements and leukocyte counts.