Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Interleukin (IL)-1 given peripherally has effects (Otterness et al., 1988; Saperas et al., 1990; Bluthe et al., 1991; Jacobs et al., 1991; Opp et al., 1991; Bianchi and Panerai, 1993; Opara et al., 1995) on the central nervous system (CNS). This demonstrates that some pathway exists by which blood-borne IL can offset the influence of the blood-brain barrier (BBB), which greatly limits the ability of circulating proteins to enter the CNS. Several pathways have been postulated by which various blood-borne cytokines can transmit information into the CNS (Dantzer, 1994; Banks et al., 1995; Watkins et al., 1995). Two pathways proposed for IL are binding to receptors and to transporters located on brain endothelial cells, the cells that comprise the BBB.

Binding to receptors capable of signal transduction would allow circulating IL to affect BBB function which, in turn, would affect CNS function. For example, cytokines have been shown or postulated to affect the integrity of the BBB (Rosenberg et al., 1995), cytoarchitecture of brain endothelial cells (Deli et al., 1995), release of substances from brain endothelial cells into the CNS (van Dam et al., 1996), and cell surface expression of adhesion molecules (Sharief et al., 1993; Barten and Ruddle, 1994). Type 1 but not type 2 IL receptor mRNA has been detected in brain endothelial cells and the choroid plexus (Cunningham et al., 1991; Cunningham and De Souza, 1993; Yabuuchi et al., 1994; Ericsson et al., 1995; van Dam et al., 1996).

Binding to transporters located at the BBB allows blood-borne IL direct, though limited, access to the CNS (Banks et al., 1989, 1991). The transporter is located on endothelial cells (Banks et al., 1994; Maness et al., 1998) and is particularly concentrated in the region of the posterior division of the septum (Maness et al., 1995). In vivo studies have shown that the transporter differs from the type 1 and type 2 IL receptors in its immunologic and binding-preference characteristics (Banks et al., 1991).

Binding sites for IL and IL-1 have been demonstrated on brain vasculature (Ban et al., 1991; Hashimoto et al., 1991; Cunningham and De Souza, 1993; van Dam et al., 1996). However, no study has indicated whether these binding sites participate in signal transduction or transport or has determined whether the characteristics of the binding sites resemble those of type 1 receptors or transporters. Herein, we characterize the binding of IL to murine brain endothelial cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Radioactive Labeling. Iodination of 5.0 µg of recombinant human (courtesy of Immunex, Seattle, WA) or murine (R&D Systems, Inc., Minneapolis, MN) IL with ¹²⁵I (Amersham Corp., Arlington Heights, IL) was performed by the enzymobead method (Bio-Rad Laboratories, Inc., Richmond, CA). The iodinated human (I-hIL) or murine (I-mIL) IL was purified by filtration on a G-10 column of Sephadex. Incorporation of ¹²⁵I, as determined by acid precipitation, was >90% and the specific activity was 700 Ci/mmol.

Microvessel Isolation. Murine cerebral microvessels were isolated by a modification of a method of Gerhart et al. (1988). All reagent volumes were proportionately adjusted for the quantity of tissue processed and, unless otherwise noted, all reagents were of cell culture quality and obtained from Sigma Chemical Company (St. Louis, MO). All glassware and plastics were precoated with PBS containing 1% BSA to minimize sticking and to maximize recovery of microvessels. Briefly, 8 to 10 cerebral cortexes from adult male ICR mice (Charles River Breeding Laboratories, Inc., Wilmington, MA) were pooled and homogenized in cold stock buffer [25 mM HEPES, 1% dextran in minimum essential medium (Gibco Laboratories, Grand Island, NY), pH 7.4] on ice. The homogenate was then filtered through a series of nylon mesh membranes (300 µm, followed by 2 × 100 µm; Spectrum Scientific Corp., Houston, TX), mixed with an equal volume of 40% dextran in stock buffer, and centrifuged at 5000g for 15 min at 4°C. The pellet was resuspended in stock buffer and filtered through a 25-µm nylon mesh membrane (Bio-Design, Carmel, NY). The microvessels were washed from the surface of the membrane with stock buffer four times, collected, and centrifuged at 2500g for 15 min at 4°C. They were resuspended in the designated assay buffer. For every new preparation, a small aliquot was taken for inspection to verify by light microscopy the purity of the microvessel preparation. Typically, >95% of the cells are brain microvessels (Koenig et al., 1992).

Microvessel Incubation Assay. Freshly isolated microvessels were resuspended in incubation buffer (129 mM NaCl, 2.5 mM KCl, 7.4 mM Na₂HPO₄, 1.3 mM KH₂PO₄, 0.63 mM CaCl₂, 0.74 mM MgSO₄, 5.3 mM glucose, 0.1 mM ascorbic acid, pH7.4) containing 1% BSA and divided into 180- to 190-µl aliquots containing ~300 µg of protein. Incubation buffer, 25 pM I-hIL or I-mIL, and any additives as indicated below were mixed with the suspensions to a final volume of 200 µl and incubated at 37°C unless otherwise specified.

(1)

Total and Specific Binding of I-hIL versus Time. Microvessels incubated as described above were harvested 15 min, 1 h, 4 h, and 24 h after addition of I-hIL (n = 3/time point). In a second set of microvessels, 3000 pM instead of 25 pM I-hIL was added at t = 0 and in a third set, 3000 pM I-hIL with 300 nM unlabeled hIL was added; these microvessels also were harvested at 15 min, 1 h, 4 h, and 24 h. The second and third sets of microvessels contained n = 4/time point. A group of tubes was incubated with 3000 pM I-hIL and buffer but no microvessels. Specific binding is reported as the %TB for the third set (3000 pM + 300 nM) of microvessels subtracted from the %TB for the first group (25 pM).

Inhibition of Specific Binding of I-IL. Self-inhibition was tested by either adding increasing amounts of I-IL, which maintains specific activity, or by adding varying amounts of unlabeled IL. Percent of specific binding was calculated by taking the %TB for microvessels incubated with 25 pM IL as 100% and with 8 nM (human) or 10 nM (murine) I-IL as 0%. An n of 3 was used per concentration.

Inhibition of I-IL Binding by Other Cytokines. I-hIL was incubated with mIL, IL-1 receptor antagonist (IL-1ra), murine IL-1, or murine tumor necrosis factor (TNF)  at concentrations ranging between 10 and 1000 pM. I-mIL was incubated with hIL or murine IL-1 in concentrations ranging from 10 to 300 pM. Results were expressed as percent of specific binding versus concentration of cytokine.

Effects of Antibodies on Specific Binding of I-hIL. The effect of four antibodies (courtesy of Immunex, Seattle, WA) directed at either the murine type 1 T cell receptor or the IL-1 molecule on specific binding was studied. The antibodies were MIL-1R M15, an antibody directed at the site on the receptor that binds IL (R-B); MIL-1R-M5, an antibody directed at the receptor that does not block IL binding to the T cell receptor (R-NB); M4, an antibody that binds to the site on IL that binds to the type 1 T cell receptor (I-B); and M4, which binds to IL at a site not associated with type 1 receptor binding (I-NB). The ability of the antibodies to inhibit the percent of specific binding of I-hIL was compared with 100 pM (1.7 ng/ml) of hIL. The antibodies were incubated at a concentration of 340 ng/ml, a 200-fold excess of the concentration of 100 pM I-hIL (100 pM) on a mass basis.

Effect of Buffer Components on Specific Binding of I-hIL. Percent of specific binding was determined in full buffer, PBS only (buffer with NaCl, KCl, Na₂HPO₄, KH₂PO₄, ascorbic acid, BSA, and adjusted to pH 7.4 but without CaCl₂, MgSO₄, or glucose), PBS with CaCl₂, PBS with MgSO₄, and PBS + glucose. These groups were compared by ANOVA followed by Duncan's range test. Each group had a n = 4.

Identification of Radioactivity Bound to Endothelial Cells. I-hIL was incubated with endothelial cells for 24 h at 37°C (incubated cells). Cells were centrifuged at 10,000g for 1 min at 4°C and the pellets collected. The pellets were washed with 400 µl of incubation buffer, centrifuged at 10,000g for 1 min at 4°C, and this second pellet collected. A solution of 10% NH₄OAc with 1% BSA (0.5 ml) was added to the second pellet at the end of 24 h and the cells disrupted by sonication. This mixture was lyophilized until submitted for HPLC. This experiment was replicated once. As a control, cells were incubated for 24 h, washed as described above, and I-hIL added with the solution of 10% NH₄OAc/1% BSA, sonicated, and lyophilized (nonincubated cells). As an additional control, I-hIL was added to a solution of 10% NH₄OAc with 1% BSA without cells and lyophilized. The lyophilized material was resuspended in distilled water, filtered, applied to a C4 column, and eluted with a gradient that started with 70% of solution A (NH₄OAc) and increased in 30 min to 90% of solution B (CH₃CN).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Total and Specific Binding of I-hIL versus Time. Fig. 1 (top) shows the %TB for the three groups of microvessels, reaching a plateau beginning at ~4 h. Sticking to incubation tubes was insignificant as shown by the group of tubes that did not contain microvessels having 0.30 (15 min), 0.29 (1 h), 0.34 (4 h), and 0.28% (24 h) of the total counts retained in the tube that would have contained the microvessel fraction. Specific binding reached a maximum at 4 h.

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Total and specific binding of I-hIL to brain microvessels. Cells were incubated from 15 min to 24 h with 25 pM I-hIL, 3000 pM I-hIL, or 3000 pM I-IL + 300 nM unlabeled hIL.

Inhibition of Specific Binding of I-IL. Figure 2 shows the specific binding for I-hIL (top) and I-mIL (bottom). For this analysis, the results are expressed with the specific binding at 25 pM set to 100% and nonspecific binding at 0%. Results differed little between determinations that relied on increasing amounts of I-IL and those that relied on increasing amounts of unlabeled IL. Therefore, for the subsequent analysis that determined the pM specifically bound, the results for both these methods were used.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Self-inhibition of specific binding of I-hIL (top) or of I-mIL (bottom). Specific binding at 25 pM was set as 100% and nonspecific binding as 0%. The amount of IL in the incubation medium (x-axis) was increased by adding increasing amounts of I-IL () or by adding increasing amounts of unlabeled IL to a constant concentration of 25 pM I-IL.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Bmax and Km values for hIL (top) and for mIL (bottom). The values for hIL were 0.955 fmol for Bmax and 292 pM for Km; for mIL the values were 0.796 fmol for Bmax and 41.3 pmol for Km.

Inhibition of I-IL Binding by Other Cytokines. Figure 4 shows that the specific binding of I-hIL was inhibited in a dose-dependent manner by mIL, IL-1ra, and murine IL-1 but not by TNF- (top). The bottom graph of Fig. 4 shows that I-mIL was inhibited by hIL and by murine IL-1.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Inhibition of I-hIL (top) and I-mIL (bottom) by other cytokines. The percentage of specific binding of I-hIL was inhibited by mIL, IL-1ra, and murine IL-1 but not by TNF-. The percentage of specific binding of I-mIL was inhibited by hIL and by murine IL-1.

Effects of Antibodies on Specific Binding of I-hIL. ANOVA showed a statistically significant effect among the groups shown in Fig. 5 (F_6,14 = 19.8, p < .01). The range test showed that all five treatments decreased specific binding (p < .05). No statistically significant difference occurred among the effects for 100 pM hIL, R-NB, R-B, or I-B. The nonblocking antibody directed at IL, I-NB, was less effective at blocking specific binding than any of the other three antibodies or 100 pM IL (p < .05).

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Effects of antibodies on the percentage of specific binding of I-IL. Antibodies were blocking (B) and nonblocking (NB) antibodies directed at the murine T cell receptor (R) or the IL molecule (I). All antibodies were effective at inhibiting percentage of specific binding, with I-B being the least effective. The results are compared with inhibition of binding by 100 pM hIL.

Effect of Buffer Components on Specific Binding of I-hIL. ANOVA found a significant effect among the buffers (F_4,15 = 3.68, p < .01) (Fig. 6). The range test showed that binding in the PBS-only buffer reduced specific binding by 50% (p < .05). Binding was restored by glucose and MgSO₄ but not PBS with CaCl₂, which still had a significantly lower binding than full buffer (p < .05). These results show that glucose and magnesium are important components in I-hIL binding.

View larger version (35K): [in this window] [in a new window]   Fig. 6.   Effects of buffer components on specific binding of I-hIL. PBS only and PBS with CaCl2 had significantly less specific binding than full buffer. Glucose or MgSO4 restored binding.

Identification of Radioactivity Bound to Endothelial Cells. For both incubated cell experiments, the nonincubated cell control and the lyophilized I-hIL, 80 to 90% of the radioactivity eluted in the position of I-hIL. Representative results for incubated and nonincubated cells are shown in Fig. 7.

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Characterization by HPLC of radioactivity recovered from endothelial cells incubated with I-hIL. Incubated cells were incubated with I-hIL for 24 h at 37°C before extraction of radioactivity and nonincubated cells had I-hIL added at the end of the incubation period.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The results herein show that I-IL binds to brain endothelial cells by a saturable mechanism. Such binding could represent either the transporter responsible for the passage of I-IL across the BBB, receptors involved in signal transduction, or a combination. The characteristics of the binding found herein more closely resemble those of the transporter than signal transduction receptors. This is consistent with findings that have shown that the site of transport for IL occurs at brain endothelial cells in addition to or instead of at the epithelial cells of the choroid plexus and circumventricular organs.

The binding of I-IL was time-dependent and saturable with maximal specific binding occurring at ~4 h (Figs. 1 and 2). Subsequent studies were conducted at 4 h. I-IL was resistant to enzymatic degradation and remained largely intact even at 24 h of incubation. The radioactivity taken up by endothelial cells also was shown to be intact I-IL (Fig. 7). The resistance of I-IL to enzymatic degradation is consistent with in vivo studies that have shown that all radioactivity recovered from blood 30 min after i.v. injection elutes as intact I-IL by HPLC (Banks et al., 1991). Binding also was found to be dependent on glucose and magnesium but not on calcium (Fig. 6).

Saturation was demonstrated by both increasing the concentration of I-IL and by adding increasing amounts of unlabeled IL (Fig. 2). The results were similar for these two methods, which suggests that the iodine attached to IL does not interfere with uptake and transport. This is consistent with previous findings that iodination does not affect the biological activity of IL (Dower et al., 1986).

The results for B_max and K_m show that IL binds at a single high-affinity, low-capacity site. The K_m was ~7 times higher for hIL than for mIL and demonstrates a species specificity of the binding site.

Binding was inhibited by IL-1 and by IL-1ra (Fig. 4). Both of these are ligands for the type 1 IL receptor and for the IL transporter located at the BBB. When tested as inhibitors of I-hIL binding, mIL and murine IL-1 were equally effective. When tested as inhibitors of I-mIL binding, murine IL- was a more effective inhibitor than hIL. Both of these relations also occur for the in vivo inhibition of I-IL transport across the BBB (Banks et al., 1991). TNF did not affect binding.

Blocking antibodies inhibited the binding of I-IL to brain endothelial cells (Fig. 5). Two antibodies directed at the IL-1 molecule were used. One was directed at the site on the IL molecule that binds to the type 1 receptor on the T cell (blocking, I-B) and the other that binds to another site (nonblocking, I-NB). Both antibodies reduced binding, but the nonblocking antibody was much less effective. This indicates that the same region of the IL molecule that binds to the type 1 receptor also is binding to the endothelial cell binding site. Two antibodies directed at the type 1 T cell receptor also were used. One was directed at the site on the receptor that binds the IL molecule (blocking, R-B) and the other to another site (nonblocking, R-NB). Blocking antibody can totally inhibit the binding of IL to the type 1 IL-1 receptor expressed by aortic endothelial cells (Akeson et al., 1992). Both antibodies were highly effective at inhibiting transport. These results show that the binding site on the endothelial cell is immunologically similar but not identical with the type 1 receptor. Therefore, the binding site on brain endothelial cells is distinct from that of the type 1 receptor. This is the same pattern that was found for in vivo IL transport across the BBB (Banks et al., 1991) and suggests that the binding sites found herein on the endothelial cell represent the transporter.

The lack of classical type 1 receptor binding despite consistent findings of type 1 receptor mRNA (Cunningham et al., 1991; Cunningham and De Souza, 1993; Yabuuchi et al., 1994; Ericsson et al., 1995; van Dam et al., 1996) seems contradictory. One possibility is that the mRNA for the transporter is homologous to the mRNA for the type 1 receptor. Alternatively, a single gene may code for both the transporter and the receptor with post-translational processing accounting for the immunologic distinction.

These results characterize the binding of I-IL to brain endothelial cells. They show that binding is most similar to that previously described for I-IL transport across the BBB and, therefore, suggest that most of the binding sites represent transporters.