Inhibition of TNF-α–induced Sickle RBC Retention in Retina by a VLA-4 Antagonist

Abstract

purpose. Patients with sickle cell disease have elevated circulating levels of cytokines including tumor necrosis factor (TNF) α. TNF-α stimulates expression by endothelial cells of adhesion molecules, including vascular cell adhesion molecule (VCAM) 1. Others have demonstrated that VLA-4 (α₄β₁), a ligand for VCAM-1 or fibronectin, is present on a fraction of sickle reticulocytes. The intent of this study was to determine, using a rat model, if TNF-α increases retention of sickle erythrocytes in retina and if that retention can be inhibited.

methods. TNF-α was given intraperitoneally to rats 5 hours before IV administration of FITC-labeled, density-separated sickle erythrocytes. After 5 minutes, rats were exsanguinated, and retinas were excised and incubated for ADPase activity, permitting the determination of the number and location of retained cells.

results. TNF-α caused a three- to fourfold increase in retention of sickle erythrocytes in retinal capillaries (P < 0.05) but not of normal human erythrocytes. Preincubation of sickle erythrocytes with TBC772, a peptide that blocks the binding ofα ₄β₁ and α₄β₇, or a monoclonal antibody against VLA-4 (19H8), significantly inhibited the TNF-α–induced retention (P ≤ 0.02), whereas a control cyclic peptide and antibody had no effect. IV TBC772 also inhibited sickle erythrocyte retention (P = 0.01). Two intravenously administered anti-fibronectin antibodies inhibited sickle cell retention as well, but an anti-rat VCAM-1 antibody did not inhibit retention.

conclusions. The authors conclude that TNF-α stimulates retention of sickle erythrocytes in the retinal vasculature. This increased retention can be blocked by a VLA-4 antagonist, suggesting that the cells retained after cytokine stimulation are reticulocytes. The counter-receptor for VLA-4 in this rat retina model appears to be fibronectin and not VCAM-1, based on data obtained using antibodies against these molecules.

Erythrocyte-mediated vaso-occlusion is believed to be the major cause of organ damage in subjects with sickle cell disease. ^{1 2} Mechanisms for sickle erythrocyte initiation of vaso-occlusions have been proposed and summarized in reviews. ^{3 4 5 6} The occlusions in retina occur initially in the peripheral retinal microvasculature at a very early age. ⁷ With increased age, even large retinal blood vessels become occluded, and the entire peripheral retina becomes nonperfused. ^{7 8 9 10} Adjacent to the hypoxic peripheral retinal tissue, preretinal neovascularization forms, which is a hallmark of proliferative sickle cell retinopathy. ^{8 11 12 13}  

We have developed a rat model for sickle cell–mediated vaso-occlusion to study these events in the retina. ¹⁴ The initial study with this model demonstrated that dense erythrocytes (SS4 cells) from subjects with sickle cell anemia (SS genotype) were retained in normal rats during normoxic conditions, and the number increased as arterial Po ₂ decreased. In contrast, neither the reticulocyte-rich, normal-density fraction from SS subjects (SS2) nor density fractions from subjects with SC disease (SC2 and SC4) were retained in significant numbers under the conditions tested. ¹⁴ One possible explanation is that the rats used in our model were inflammation- and infection-free animals. Sickle cell disease subjects, on the other hand, have high circulating cytokine levels, high white cell counts, and frequent infections. ¹⁵  

Cytokines such as tumor necrosis factor (TNF) α and IL-1α upregulate leukocyte adhesion molecule expression by vascular endothelial cells. ^{16 17} TNF-α and IL-1α are elevated in steady state sickle cell subjects. ^{18 19} Sickle erythrocyte retention produces low-level tissue damage ¹⁵ and/or more severe tissue damage associated with painful crisis events. ² TNF-α is one of the most potent stimulators of endothelial cell upregulation of leukocyte/endothelial cell adhesion molecules, although the time course for upregulation of each adhesion molecule depends on the origin of the target endothelial cell. ^{20 21 22}  

Several groups have recently demonstrated that some reticulocytes from sickle cell subjects have the integrinα ₄β₁ (VLA-4) on their surface. ^{23 24 25} These cells could be stress reticulocytes that have been prematurely released from marrow because of rapid clearing of circulating RBCs in sickle cell subjects. The endothelial cell/leukocyte adhesion molecule VCAM-1 has been reported to mediate sickle reticulocyte binding to human endothelium in vitro via the VLA-4 counter-receptor present on some reticulocytes. ^{23 24 25} Setty and Stuart ²⁶ have found that VCAM-1/VLA-4 interaction is responsible for the in vitro adherence of dense cells and reticulocytes to both macrovascular and retinal microvascular endothelial cells. It remains to be determined if these findings can be reproduced in in vivo vasculatures. 

The present study evaluates the effect of the cytokine TNF-α on sickle erythrocyte retention in the retinal vasculature. The potential involvement of VLA-4 on sickle erythrocyte retention in retina was investigated using a cyclic peptide antagonist ofα ₄β₁ andα ₄β₇, TBC772, ²⁷ and IV antibodies againstα ₄β₁ and its receptors, VCAM-1 and fibronectin. The results demonstrate that increased sickle erythrocyte binding in cytokine-stimulated retina can be prevented by a VLA-4 antagonist. 

Methods

Preparation of Erythrocytes

Blood was obtained after informed consent from three normal (AA) subjects at the Wilmer Ophthalmological Institute and five subjects with sickle cell anemia (SS) at the Bronx Comprehensive Sickle Cell Center who had known homozygous genotype (SS) and had not been transfused for 3 months. On the same day that blood was drawn, it was passed through a Pall filter (Pall Biomedical, Inc., Glen Cove, NY) to remove microclots and most white blood cells. Erythrocytes from sickle cell patients (SS-RBCs) at the Bronx Comprehensive Sickle Cell Center were also separated on a Percoll-Larex continuous density gradient into normal-density (SS2) and high-density (SS4) fractions. ²⁸ The SS2 fraction is enriched in reticulocytes, and the SS4 is enriched in irreversibly sickled cells (ISCs). ²⁸ Immediately after isolation, the fractionated erythrocytes were shipped overnight at 4°C. 

Immediately on receipt, the cells were washed in phosphate-buffered saline and then labeled with fluorescein isothiocyanate (FITC; Research Organics Inc., Cleveland, OH): 0.1 ml of packed cells/2.0 ml of 0.1% FITC in PBS, for 1 hour at room temperature by the procedure of Butcher and Weissman. ²⁹ AA RBCs were stored overnight as well before labeling. Kaul et al. ³⁰ demonstrated that FITC-labeling of sickle cells has no effect on adhesion of SS2 cells or retention of SS4 cells in rat mesocecum by comparing results for pairs of FITC-labeled and unlabeled cells. All animal experimentation was completed within 24 hours after the cells were drawn. 

Preparation of Rats

Male Sprague–Dawley rats (Harlan, Frederick, MD) weighing 200 to 250 g were anesthetized with 0.1 ml of ketamine (50 mg/ml; Phoenix Scientific, Inc., St. Joseph, MO)/Rompun (5 mg/ml; Phoenix Scientific, Inc.) per 100 g body weight given IM. A tracheotomy was performed, and the animals were ventilated using a Harvard Rodent Ventilator (TV = 1.5 cm³/kg, RR = 100). Femoral arteries were catheterized with polyethylene tubing (PE-50) to draw samples for blood gas analysis (Acid Base Laboratory ABL-5; Radiometer, Copenhagen, Denmark). Arterial blood oxygen levels ranged between 100 and 150 mm Hg. A femoral vein in the contralateral leg was catheterized with polyethylene tubing (PE-50) tubing for delivery of FITC-labeled RBCs (300 μl, Hct 10). Cells were injected slowly and allowed to circulate for 5 minutes. Animals were treated in accordance with the ARVO Statement for Use of Animals in Ophthalmic and Vision Research, and the studies were approved by the Johns Hopkins University Institutional Review Board. 

The animals were killed by an overdose of sodium pentobarbital after perfusion at 5 ml/min with 60 ml heparinized saline via the left ventricle after severing the jugular veins, and then the eyes were removed. Retinas were removed and processed by our ADPase flat-mount technique. ³¹ The entire retinal vasculature was visualized under darkfield illumination (lead ADPase reaction product), and FITC-labeled cells were visualized by fluorescence microscopy. The number of RBCs in each retina and their location within the vasculature were noted for each set of experiments. AA cells were used as controls and yielded <10 cells retained per retina under all conditions. For some animals, after counting the number of FITC-labeled RBCs, granulocytes and monocytes were detected by incubating the retinas for enzyme histochemical demonstration of nonspecific esterase with napthol AS-D chloroacetate as substrate (kit 91; Sigma, St. Louis, MO), which labels granulocytes prominently red and monocytes pink. 

Ten or twenty micrograms of TNF-α (recombinant rat TNF-α; R&D Systems, Minneapolis, MN) per kg body weight was administered IP at 5 and 9 hours before delivery of FITC-labeled RBCs. The involvement of VLA-4 in sRBC retention was investigated with TBC772, a cyclic peptide that is an antagonist ofα ₄β₁ andα ₄β₇, and with TBC1194 as a control. TBC 772 is a cyclic hexapeptide with the sequence CWLDVC, and peptide TBC1194 is a cyclic hexapeptide with a scrambled sequence (CDLVWC). ²⁷ TBC772 blocks binding of VLA-4 to VCAM-1, fibronectin, and mucosal addressin cell adhesion molecule-1 (MadCAM-1) ²⁷ and can neutralize integrin coactivation systems as occur in activation of T cells. ³²  

Before administration to rats, SS-RBCs were also preincubated for 15 minutes with monoclonal antibody (mAb) 19H8, which recognizesα ₄β₁ but notα ₄β₇. ^{32 33} There are two vascular endothelial cell counter receptors that could be responsible for VLA-4–mediated adhesion, the CS-1 portion of fibronectin and VCAM-1. To determine whether SS-RBC retention in retina was modulated by either of these counter-receptors, we administered IV the following monoclonal antibodies, 30 or 60 minutes before administration of SS-RBCs: 5 and 7.5 mg/kg anti-VCAM-1 (clone 5-F-10; generously provided by Roy Lobb at Biogen, Boston, MA) ^{34 35} ; 7.5 mg/kg anti-human fibronectin (FN-15, Sigma); 7.5 mg/kg anti-CS-1 (clone 7E5, graciously provided by Tatiana Ugarova, Cleveland Clinic Foundation, OH); or an isotype matched mouse IgG (Jackson Laboratories, West Grove, PA). 

Data are presented as the mean number of cells in retina ± SD. The difference between cells retained with control versus experimental peptides and antibodies were analyzed by the Student’s t-test to determine statistical significance. P ≤ 0.05 was considered significant. 

Results

FITC-labeled cells were easily observed under fluorescein wavelength-filtered transmitted light (Fig. 1) , and the position of the cells in the retinal vascular hierarchy was determined by viewing the lead ADPase reaction product in the same area with darkfield illumination. Although many SS-RBCs were observed at bifurcations in the microvasculature (Fig. 1) , the majority of the cells were present in straight segments of capillaries. The total number of cells in each retina was counted. Eighty-four percent of all the SS-RBCs observed in retinas from this study were in the capillary networks, not in larger vessels. When normal, AA erythrocytes were administered, there were always 10 or fewer cells observed in a retina. 

Initially, we evaluated 10 μg of TNF-α/kg, administered 5 or 9 hours before administration of FITC-labeled SS-RBCs. Both times of administration resulted in a significant increase in SS-RBC retention in rat retinal vessels (P < 0.05), and there was no significant difference between the time periods (Fig. 2A ). Henninger et al. ³⁶ demonstrated that ICAM-1 (a neutrophil/endothelial cell adhesion molecule) is elevated on endothelium in mouse at 9 hours but not 5 hours after TNF-α. Therefore, TNF-α administration 5 hours before delivering SS-RBCs was used in all subsequent experiments to avoid any polymorphonuclear leukocyte (PMN)-mediated vaso-occlusion. We then evaluated 10 and 20μ g TNF-α/kg body weight and found that 20 μg of recombinant rat TNF-α yielded greater SS2 retention (P = 0.05) than 10 μg (Fig. 2B) . Furthermore, TNF-α stimulated the retention of both SS2 (P = 0.008) and SS4 erythrocytes (P = 0.001; Fig. 2C ). TNF-α at 20 μg/kg did not increase the number of AA cells retained in rat retina (results not shown). TNF-α at 20 μg/kg also did not change the number of granulocytes (mostly PMNs) retained in retina, and when nonspecific esterase positive cells were observed, they were not associated with retained FITC-labeled erythrocytes (results not shown). These data suggested that the increased number of FITC-labeled SS-RBCs observed in retina after TNF-α (20 μg/kg given 5 hours before FITC-erythrocytes) was not retained because of their association with PMNs. 

We hypothesized at first that TNF-α had stimulated production of VCAM-1 by vascular endothelial cells and VLA-4/α₄β₁ on reticulocytes was modulating adhesion of some SS erythrocytes to vascular endothelium. Therefore, we investigated an antagonist ofα ₄β₁ andα ₄β₇, cyclic peptide TBC772. We evaluated the effect of TBC772 initially by preincubating the SS2 or SS4 cells with 200 μM TBC772 or control peptide TBC1194 and gently rocking the cell suspension for 15 minutes before administration of the SS-RBCs. Rats in these studies received 20 μg TNF-α, 5 hours before administration of cells. As in our previous study, more SS4 cells were retained in retina under normoxic conditions than SS2 cells (Figs. 2C and 3A) . TNF-α increased retention of both SS2 and SS4 (Figs. 2C and 3A) . Cells preincubated with TBC772 were retained in retina at the same level as in rats not receiving TNF-α, that is, TBC772 prevented TNF-α–stimulated retention of both cell types (P ≤ 0.02; Fig. 3A ). The inhibition of TNF-α–initiated retention by TBC772 was dose dependent in that 200 μM yielded complete inhibition of retention of SS2 cells (P < 0.001), whereas 50 μM inhibited retention to a lesser extent (NS; Fig. 3B ). TBC772 can block both α₄β₁ andα ₄β₇, so a monoclonal antibody that recognizesα ₄β₁ but notα ₄β₇ (mAb 19H8) ^{32 33} was evaluated. This antibody significantly inhibited SS-RBC retention in retina when FITC-labeled SS-RBCs were incubated with it for 15 minutes before administration (P < 0.002; Fig. 4 ). 

IV administration of TBC772 was then evaluated. Increased retention of SS2 cells in TNF-α–treated rats could also be inhibited by IV delivery of TBC 772 (25 mg/kg), 5 minutes before administering cells (P = 0.01; Fig. 3C ). If the peptide was administered IV 8 hours before administering cells, the inhibition was not statistically significant (P = 0.06). 

There are two well-established vascular endothelial cell counter-receptors that could be responsible for VLA-4–mediated SS-RBC retention, the CS-1 portion of fibronectin and VCAM-1. TBC772 and mAb 19H8 antagonize VLA-4 binding to either of these receptors. ²⁷ To determine whether either of these counter-receptors on endothelial cells was involved in increased retention of SS-RBCs in TNF-α–treated animals, we administered mAbs against them intravenously to rats before administering cells. The monoclonal antibody against VCAM-1 (clone 5-F-10), which had previously been shown to functionally block rat VCAM-1, ^{34 35} was administered 30 or 60 minutes before administration of SS-RBCs. Neither dose (5 or 7.5 mg/kg) nor time of administration prevented increased retention of SS-RBCs in TNF-α–treated rats (Fig. 4) . An mAb against human fibronectin (7.5 mg/kg, FN-15) or an isotype-matched mouse IgG was administered intravenously, 30 minutes before administration of SS-RBCs. Anti-fibronectin inhibited retention of SS-RBCs (P < 0.001; Fig. 5 ). Furthermore, a mAb against the CS-1 domain of human fibronectin ^{37 38} also significantly inhibited retention of SS-RBCs when administered IV 30 minutes before SS-RBCs (P = 0.01; Fig. 5 ). 

Discussion

This study has demonstrated that cytokines such as TNF-α increase retention of sRBCs in retina, especially fractions rich in reticulocytes. It has already been demonstrated that sickle cell subjects, even in steady state, have increased circulating levels of TNF-α and IL1β. ^{18 19} The high levels of these cytokines may be due to frequent infections, low-level inflammation from a continuum of vaso-occlusions in peripheral tissue, increased white cell count, or more severe tissue damage that is incurred during painful crisis. ¹⁵ It is interesting that infections often are precipitating factors for painful crisis in sickle cell disease. ^{15 39} Administration of lipopolysaccharide from Gram-negative bacteria 20 hours before SS-RBC delivery increases retention of SS-RBCs, especially reticulocyte rich fractions, in retina. ⁴⁰  

TNF-α upregulates expression of many leukocyte/endothelial cell adhesion molecules like ICAM-1 and VCAM-1 and activates leukocytes such as PMNs. ^{16 17 36} Administration of TNF-α 5 hours before injection of FITC-labeled RBCs was chosen for most experiments to minimize the expression of ICAM-1, based on the observations of Henninger et al. ³⁶ in mouse. PMNs, which bind to ICAM-1 and can easily obstruct capillary lumens, were visualized in some retinas by nonspecific esterase activity and were not increased in retina nor associated with retained FITC-labeled-RBCs. VLA-4 is expressed not only on sickle reticulocytes but also on mononuclear leukocytes and prominently on T cells. ⁴¹ Because we completely blocked retention of FITC-RBCs by preincubating these cells with peptide TBC772, it is unlikely that rat monocyte or T-cell adherence to retinal endothelium via VLA-4 was involved in increased SS-RBC retention. Finally, AA cell retention did not increase in TNF-α–treated rats, suggesting further that adhesion of rat leukocytes did not cause human RBC retention. 

It seemed surprising that both SS2 and SS4 cell retention was increased in TNF-α–treated rats. The SS4 fraction is the densest and contains the most irreversibly sickled cells, whereas the SS2 fraction is enriched for reticulocytes. Heterogeneity in the composition of density-derived fractions is well established, as is heterogeneity in erythrocyte characteristics between sickle cell patients. ⁴² Retention of both fractions may be due to the presence of VLA-4–positive reticulocytes in both fractions. FACS analysis of SS2 and SS4 fractions using antibody 19H8 demonstrated VLA-4–positive cells in the SS4 fraction of some sickle cell subjects (McIntyre B, Lutty G, unpublished results, 1998). Additionally, we observed variation in cytokine-stimulated SS4 cell retention in retina, suggesting that there may be interindividual variation in the number of VLA-4–positive cells in the SS4 fraction (cf. Figs. 2C and 3A ). Increased retention of both fractions was observed in the in vitro study of Setty and Stuart, ²⁶ in which they suggested that both fractions bind to VCAM-1. 

The inhibition of SS-RBC retention by a peptide (TBC772) and a monoclonal antibody (19H8) that block VLA-4 and not by a control peptide or antibody suggests that the retention may be modulated by VLA-4 on reticulocytes. ^{23 24 25} There are two well characterized vascular endothelial cell counter-receptors that could be responsible for VLA-4–mediated SS-RBC retention, the CS-1 portion of fibronectin and VCAM-1. In vitro studies have demonstrated that either VCAM-1 or fibronectin on endothelial cells could be the counter-receptor for reticulocyte VLA-4. ^{26 43 44 45} TBC772 antagonizes VLA-4 binding to either of these receptors. ²⁷ Attempts to block the retention of cells by administering anti-rat VCAM-1 to the rats before administration of SS-RBCs were unsuccessful. The antibody used in those experiments has been previously reported to block rat T-cell entry into thymus at the same in vivo dose used in the present study, to block adherence of T cells to rat high endothelial cells in vitro, and to recognize VCAM-1 in histologic sections of rat lymph nodes. ^{34 35} This suggests either that VCAM-1 is not elevated as predicted in the rat retinal vasculature 5 hours after administration of 20 μg TNF-α/kg ³⁶ or that the level of antibody administered was not sufficient to neutralize the VCAM-1 expressed on retinal vascular endothelial cells in this experiment. However, anti-fibronectin (FN-15) completely inhibited increased retention so it is probable that TBC772 is blocking the binding of VLA-4 to the CS-1 segment of fibronectin, although the epitope on fibronectin that this antibody recognizes is unknown. Additionally, IV administration of a mAb against CS-1 in our model significantly inhibited retention of SS-RBCs in TNF-α–treated rats, further suggesting that CS-1 is the receptor on retinal endothelial cells. Kasschau et al. ⁴⁶ previously demonstrated in vitro adherence of sickle RBCs to fibronectin, whereas normal RBCs do not adhere. Kumar et al. ⁴⁵ have demonstrated that the same fibronectin antibody used in the present study blocked adherence of phorbol ester–treated sickle erythrocytes to endothelial cells in a flow chamber. They demonstrated further that CS-1 peptides could also block this adherence. 

The present study suggests that, in TNF-α–treated rats, the counter-receptor for VLA4 in retinal blood vessels is fibronectin. TNF-α can modulate production of matrix components like fibronectin. ⁴⁷ Perhaps TNF-α, while increasing vascular permeability, exposes the CS-1 portion of fibronectin in the process. Conformational changes in fibronectin have been demonstrated to alter its cellular adhesive properties. ⁴⁸ Alternatively, changes in endothelial cell membrane fluidity may cause CS-1 on endothelial cell fibronectin to be exposed. Increased luminal exposure of CS-1 has been observed in synovia during arthritis, a condition where TNF-α is prominent. ⁴⁹ Manodori et al. have observed increased adhesion of sickle RBCs after induction of interendothelial cell gaps by thrombin. ⁵⁰ Inflammatory cytokines such as TNF-α and IL-1b induce increased permeability in retinal blood vessels by producing gaps between endothelial cells. ⁵¹  

Although fibronectin appears to be the receptor in our rat model, in vitro studies have suggested that either the VCAM-1 or CS-1 portion of fibronectin can serve as the counter-receptor for sickle reticulocyte VLA-4. ^{26 43 44 45} It has been suggested by in vitro experiments that the counter-receptor in human sickle cell retina may be VCAM-1. ²⁶ Stuart and Setty ⁵² have recently observed increased serum levels of soluble VCAM-1 during sickle cell acute chest syndrome. We observed, in a recent immunohistochemical study, increased expression of VCAM-1 in retinal blood vessels and preretinal neovascularization of sickle cell subjects. ⁵³ In ex vivo rat mesocecum after stimulation with platelet activating factor, antibodies against αVβ3-inhibited sickle cell retention. ⁵⁴ TBC772, however, shows no inhibition of αVβ3 at doses as high as 1 mM. ⁵⁵ Therefore, it appears that there are many pathways involved in retention of sickle RBCs in vivo and adherence to endothelial cells in vitro. The pathway involved is probably dependent on the vasculature studied and the stimulus for retention. 

There are probably two mechanisms for sickle RBC retention in retina. We have previously demonstrated that the number of dense cells from SS subjects retained in retina increases after 5 minutes exposure to decreasing arterial Po ₂, ¹⁴ which is compatible with mechanical obstruction, i.e., obstruction due to nondeformability. The present study demonstrates a second mechanism that occurs when the retinal vasculature is stimulated with TNF-α, adherence of VLA-4–positive sRBCs to endothelium. Both of these mechanisms could co-exist in the sickle cell subjects retina. If elevated cytokines are present, VLA-4–positive cells may adhere to activated endothelium. Although the reticulocytes tend to be more pliable, their presence in the microvasculature would block the dense, rigid irreversibly sickled cells from passing through the microvasculature. This would create hypoxia that, as we have already demonstrated, will increase the dense cells retained in these vascular segments. Fabry et al. ⁵⁶ have documented a similar scenario in hind leg of rats. Kaul et al. ^{30 57 58} have also observed retention of SS2 sRBCs in the mesocecum and mesoappendix and retention of dense cells after lighter density cells have adhered, or during hypoxia. 

The effects of TBC772 and mAb 19H8 on sickle RBC retention suggests VLA-4 as a target for prevention of cytokine-stimulated sickle RBC retention. TBC772 or similar molecules have potential therapeutic application in sickle cell disease because it interferes with VLA-4 binding to both VCAM-1 and CS-1 on fibronectin, so either VLA-4 binding site would be blocked. This has implications well beyond retinal vascular occlusion and retinopathy. If this approach can disrupt and/or prevent occlusion in other organ systems, it might be used in shortening the duration of sickle cell painful crisis and improve the quality of life and survival of these patients. 

 Supported by National Institutes of Health Grants HL45922–07 (GAL), HL38655 (RLN), CA62596 (BWM) and EY01765 (Wilmer) and in part by a grant from Texas Biotechnology Corp., Houston, Texas. GAL is an American Heart Association Established Investigator and a recipient of an Research to Prevent Blindness Lew Wasserman Merit Award.

 Submitted for publication November 20, 2000; revised January 17, 2001; accepted February 7, 2001.

 Commercial relationships policy: E (PV); F (all).

 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.

 

FITC-labeled SS2 cells in retinal vasculature of a rat treated with 20μ g of TNF-α per kg body weight, 5 hours before administration of sRBCs. Cells in both micrographs are present at bifurcations in the capillaries.

Figure 2.

 

Effect of TNF-α on sickle RBC retention in retina. Mean number of fluorescently labeled sickle RBCs per retina ± SD is used in (A) through (C). (A) TNF-α, 10μ g/kg body weight, was administered IP either 5 or 9 hours before SS2 cells. There was a significant increase in retention of sickle RBCs when TNF-α was given either 5 (P < 0.03) or 9 hours (P < 0.01) before administration of cells compared with retention in rats not receiving TNF-α (n = 4 in each group). (B) Ten or 20 μg TNF-α/kg were administered IP 5 hours before SS2 cells were given. There was a dose-dependent increase in mean number of cells retained in retina but only the 20-μg/kg dose yielded a statistically significant increase compared with rats not receiving TNF-α (P < 0.01; n = 3 in each group). (C) Effect of 20μ g/kg TNF-α given IP 5 hours before administration of SS2 or SS4 cells. Increased retention of both cell types after administration of TNF-α was statistically significant (*P < 0.01; n = 4 for each group).

Figure 3.

 

Effect of experimental cyclic peptide TBC772 (EP) or control, scrambled cyclic peptide 1194 (CP) on sickle RBC retention. Mean number of fluorescently labeled sickle RBCs per retina ± SD is expressed in (A) through (C). (A) SS2 or SS4 cells were preincubated with 200 μM TBC 772 (EP) or TBC1194 (CP), 15 minutes before administration to rats. TBC772 significantly inhibited TNF-α–induced retention of both SS2 and SS4 cells (*P ≤ 0.02; n = 4 in each group). (B) Preincubation with 200 μM TBC772 yielded greater inhibition in TNF-α–induced retention of SS2 cells than 50 μM (P < 0.05). Inhibition of cell retention in retina by 200 μM TBC772 was highly significant (P < 0.001; n = 3 in each group). (C) Effect of administering 25 mg/kg TBC772 IV to rats either 5 minutes or 8 hours before administration of SS2 cells (*P = 0.01; n = 4).

Figure 4.

 

Effects of monoclonal antibodies against VCAM-1 (5F10) and VLA-4 (19H8) on SS-RBC retention in retina. TNF-α, 20 μg, significantly increased retention of SS-RBCs when 7.5 mg/kg nonimmune IgG (NIM) was administered IV 30 minutes before (n = 10). IV administration of 7.5 mg/kg of mAb 5F10 30 minutes before administration of SS-RBCs did not affect retention of cells in retina (n = 4). However, preincubating the SS-RBCs for 15 minutes with mAb 19H8 (VLA-4) significantly inhibited retention of SS-RBCs (*P < 0.05; n = 3).

Figure 5.

 

Effect of administering 7.5 mg/kg monoclonal antibodies against fibronectin. Anti-fibronectin (FN15; Sigma) and anti-CS-1 domain of fibronectin (7E5) or control IgG (NIM) were administered IV 30 minutes before administering SS2 cells to rats receiving 20 μg/kg TNF-α. Both anti-fibronectin (P = 0.0002) and anti-CS-1 (P = 0.001) significantly inhibited SS2 retention (n = 4 in each group).

The authors acknowledge the excellent technical assistance of Carol Merges and Sandra Suzuka, Roy Lobb from Biogen for graciously providing mAb 5F10, and Tatiana Ugarova from Cleveland Clinic Foundation, Ohio, for graciously providing mAb 7E5.