Introduction

Top Abstract Introduction Methods Results Discussion References

Enhanced permeability of the blood-brain barrier (BBB) is a characteristic abnormality observed during the pathogenesis of the human demyelinating disease, multiple sclerosis (MS) (Adams et al., 1989). The constitutive endothelial cells of the neurovasculature operate selectively to maintain BBB homeostasis but malfunction during MS to allow the formation of vasogenic edema and infiltration of inflammatory cells into parenchymal tissues with subsequent demyelination of nerve fibers. The inducible autoimmune condition experimental allergic encephalomyelitis (EAE) has many pathological features in common with MS, of which cerebrovascular leakage in susceptible animals is a standard occurrence (Leibowitz and Kennedy, 1972; Hawkins et al., 1990). Original pharmacological studies by us in acute EAE showed that the glucocorticoid dexamethasone and the immunosuppressant cyclosporin A limit neurovascular breakdown, possibly through a reduction in the release of permeability-inducing factors (Paul and Bolton, 1994, 1995).

Although the mechanisms instigating loss of BBB integrity in MS and EAE remain unclear, investigations in nonimmune models of neurovascular damage have identified central nervous system (CNS)-derived polyamines and the nitrogen intermediate nitric oxide (NO) as primary mediators of cerebral vessel disruption (Koenig et al., 1983a; Trout et al., 1986; Faraci and Brian, 1994). Interestingly, studies in EAE and MS have indicated cerebral NO involvement (Koprowski et al., 1993; Bö et al., 1994), and preliminary observations by us have shown enhanced NO and polyamine levels in neurovascular isolates from EAE-diseased animals corresponding with initial BBB leakage (Bolton et al., 1994; Scott et al., 1996). Furthermore, dexamethasone and cyclosporin A have the potential to inhibit CNS vessel polyamine and NO production (Koenig et al., 1989; Ryffel, 1993), which supports our original hypothesis concerning the drugs' direct effects on the BBB during early EAE.

Both NO and the polyamines can be generated by activation of the N-methyl-D-aspartate (NMDA) receptor which has been linked to the neuronal injury seen in several neurodegenerative conditions including Parkinson's and Alzheimer's disease (Lee et al. 1988; Zeevalk et al., 1994). In addition to the well-demonstrated neuronal location, the NMDA subtype of glutamate receptor has been shown to reside in rat CNS-derived blood vessels (Garthwaite et al., 1988; Monaghan et al., 1989; Koenig et al., 1992), although species variation in receptor expression at the neuroendothelium is recognized (Beart et al., 1988; Faraci and Breese, 1993; Giese et al., 1995). Up-regulation of the neuroendothelial NMDA receptor and associated permeability-inducing factor release culminate in the loss of BBB integrity in nondisease models, which can be suppressed by the specific noncompetitive antagonist MK-801 (dizocilpine maleate) (Koenig et al., 1992). Interestingly, the potential for NMDA receptor activation exists in EAE as increased levels of excitatory amino acid agonists have been reported in the CNS of animals during disease (Honegger et al., 1989; Flanagan et al., 1995). Therefore, we were interested in examining the ability of MK-801 to suppress aberrant CNS vessel leakage during EAE and thereby provide clear evidence for NMDA receptor involvement in neurovascular breakdown in an immune-based model of human neurological disease.

	Methods

Top Abstract Introduction Methods Results Discussion References

Animals

Male Lewis rats, weighing 200 to 250 g, were used from stock bred on site and housed five rats per cage, with food (CRM diet) and water ad libitum.

Induction of EAE

EAE was induced in animals as described previously (Bolton and Flower, 1989). An emulsion comprising equal parts of guinea pig spinal cord, sterile PBS and incomplete Freund's adjuvant (Difco Laboratories, Detroit, MI) was prepared and supplemented with 10 mg ml¹ Mycobacterium tuberculosis H₃₇Ra (Difco Laboratories). Rats were inoculated with 0.1 ml of inoculum into each hind footpad. A minimum of 5 rats were used per treatment.

Evaluation of Neurological EAE

Animals were weighed daily beginning on day 0 PI and assessed for neurological EAE. Disease symptoms appeared after weight loss and were scored as follows: 1, flaccid tail; 2, hind limb hypotonia; 3, partial hind limb paralysis; 4, complete hind limb paralysis.

Assessment of Histological EAE

The cervical spinal cords of animals from vehicle and drug treatment groups were examined by light microscopy for inflammatory lesions sampling on day 12 PI. Spinal tissue was selected for analysis because a heavy lesion load can be guaranteed in this CNS area during the onset of acute EAE (Bolton et al., 1984). Furthermore, the efficacy of MK-801 to restrict lesion development during early EAE could be evaluated confidently. The upper 1.5 cm of tissue was dissected and snap frozen. Cervical cord sections were cut at 5-µm thickness, at one standard depth and stained with hematoxylin and eosin. Lesion number per section was quantitated "blind" and assessed for intensity of cellular infiltration.

Preparation and Administration of MK-801

MK-801 (supplied by Dr. L.L. Iversen, Merck, Sharp and Dohme Research Laboratories, Harlow, England) was suspended in sterile PBS and administered prophylactically by i.p. injection, once daily at either 0.15 mg kg¹ bwt or 0.3 mg kg¹ bwt for 6 days, beginning day 7 PI. Animals were also treated therapeutically with MK-801 starting from initial weight loss, typically day 10 PI, for 3 days dosing i.p., once daily at either 0.3 mg kg¹ bwt or 0.6 mg kg¹ bwt. Control EAE-inoculated rats received vehicle alone.

Quantitation of BBB Integrity

BBB permeability in selected areas of the CNS was determined according to our previous methods (Paul and Bolton, 1995), which are briefly detailed below.

Labeling of red blood cells with ¹¹¹In-tropolonate. Cell-free plasma was prepared from pooled Wistar rat blood by repeated centrifugation. The remaining blood cells were resuspended in HEPES saline buffer (20 mM HEPES [Gibco, Ltd., Paisley, UK]; 0.8% NaCl), washed repeatedly to remove leukocytes and reconstituted to provide an erythrocyte concentration of 5 × 10⁸ cells ml¹. A preparation containing 20 µCi ¹¹¹In-tropolonate/5 × 10⁸ red blood cells was incubated at 37°C for 20 min followed by washing and resuspension at 5 × 10⁹ cells/0.5 ml cell-free plasma.

Determination of BBB permeability. Rats received 10 µCi ¹²⁵I-rat serum albumin i.v. under halothane/oxygen and, 24 hr later, 5 × 10^{9 111}In-red blood cells were injected as a blood volume marker. After a 4.5-min circulation time, cardiac blood was collected into heparin-coated tubes followed by a lethal injection of euthatal (RMB Animal Health Ltd., Dagenham, UK) at 5 min. Cerebella, medulla-pons and cervical spinal tissues were dissected out and the ¹¹¹In levels in samples and 100-µl blood aliquots from each animal were recorded by an LKB minigamma counter. Quantities of ¹²⁵I were measured in samples following ¹¹¹In decay after storage for 3 weeks at 20°C. BBB permeability, expressed as EVBE, was calculated from isotope levels in tissue and blood (equation 1) and is a measure of radiolabeled albumin that has crossed the neurovasculature and accumulated within CNS tissues.

(1)

Corticosterone Radioimmunoassay

Circulating corticosterone levels in rats were determined to exclude the possibility that treatment regimes enhanced endogenous steroid levels which are known to influence the course of EAE (MacPhee et al., 1989) and could therefore contribute to drug efficacy. Blood was collected at a standard time by cardiac puncture, dispensed into heparin-coated tubes for subsequent extraction of plasma by centrifugation at 300 × g with storage at 20°C before assay. Plasma corticosterone levels were measured in samples from all treatment groups with a Gamma-B ¹²⁵I-Corticosterone Radioimmunoassay Kit (IDS, Tyne and Wear, UK) according to the manufacturer's instructions.

Statistical Analysis

Plasma corticosterone data were analyzed by one-way analysis of variance. Significant differences between all other drug and vehicle treatment results were determined using the Mann-Whitney U test for nonparametric data with Bonferroni correction for multiple comparisons where required.

	Results

Top Abstract Introduction Methods Results Discussion References

The effects of prophylactic MK-801 administration on neurovascular permeability and disease development. The appearance of neurological EAE in vehicle-treated, sensitized rats, 12 days PI, was accompanied by increased vascular permeability in all CNS areas studied and was similar to previous results from untreated diseased animals detailed in our earlier studies (fig. 1) (Paul and Bolton, 1995). Repeated prophylactic doses of MK-801 at 0.15 mg kg¹ bwt suppressed BBB leakage and significantly reduced neurological deficits in EAE-diseased rats (P < .05) (table 1). However, a 2-fold increase in the administered dose of MK-801 maintained EVBE values within normal limits in the cerebella of treated rats (P < .01) and significantly inhibited BBB disruption in medulla-pons (P < .05) and cervical spinal tissue (P < .01). In addition, high-dose MK-801 markedly curtailed the development of paralytic disease in EAE-inoculated animals (P < .01). Body weight loss in drug-treated groups was not in excess of that shown by vehicle controls, indicating the dosing regimes used were well tolerated. Prophylactically treated animals receiving 0.3 mg kg¹ bwt MK-801 lost significantly less weight, from disease onset, than vehicle-dosed rats (P < .05) (table 1).

View larger version (K): [in this window] [in a new window]   Fig. 1.   The ability of prophylactically administered MK-801 to suppress neurovascular permeability in EAE-inoculated Lewis rats. MK-801 treatment was initiated on day 7 PI and continued for 6 days dosing i.p. at either 0.15 mg kg1 bwt (hatched columns, n = 6) or 0.3 mg kg1 bwt (grey columns, n = 6). Vehicle control EAE-sensitized animals (solid column, n = 10) received sterile PBS. Normal neurovascular permeability was demonstrated in uninoculated Lewis rats (open column, n = 6). Columns represent mean EVBE values ± S.E. * P < .05 and ** P < .01 compared with vehicle treatment.

The effects of therapeutic MK-801 administration on neurovascular permeability and disease development. Vascular leakage in defined areas of the CNS from inoculated, vehicle-treated rats were similar to values recorded for prophylactically dosed controls (fig. 2). Therapeutic doses of MK-801 at 0.3 mg kg¹ bwt completely counteracted neuroendothelial disruption in the cerebella of treated rats (P < .05) and markedly reduced BBB dysfunction in remaining isolated tissues. However, increasing the dose of drug to 0.6 mg kg¹ bwt did not further enhance suppression of neurovascular breakdown to a significant level compared with the inhibition achieved with the lower-dose therapy. The mean neurological scores for animals receiving either 0.3 mg kg¹ bwt or 0.6 mg kg¹ bwt were not significantly different from vehicle controls (table 1).

View larger version (K): [in this window] [in a new window]   Fig. 2.   The ability of therapeutically administered MK-801 to suppress neurovascular permeability in EAE-inoculated Lewis rats. Treatment with MK-801 at either 0.3 mg kg1 bwt (hatched columns, n = 6) or 0.6 mg kg1 bwt (grey columns, n = 5) commenced at weight loss and continued for 3 days dosing i.p. once daily. Vehicle control EAE-sensitized animals (solid column, n = 6) received sterile PBS. Normal neurovascular permeability was demonstrated in uninoculated Lewis rats (open column, n = 6). Columns represent mean EVBE values ± S.E. * P < .05 compared with vehicle treatment.

Evaluation of CNS lesions after MK-801 treatment. Prophylactic administration of MK-801 at 0.15 mg kg¹ bwt from day 7 PI did not reduce the number of inflammatory lesions in cervical spinal tissues or limit the intensity of infiltration (table 2). However, increasing the dosage significantly reduced the density of perivascular lesions (P < .05). Furthermore, the extent of inflammatory cell infiltration was markedly restricted compared with vehicle control tissues (P < .02). Therapeutic treatments of 0.3 mg kg¹ bwt and 0.6 mg kg¹ bwt MK-801 did not alter the appearance or severity of inflammatory infiltrates.

The effect of vehicle and MK-801 treatment regimes on circulating corticosterone levels. Prophylactic vehicle treatment resulted in elevated endogenous corticosterone levels (table 1) similar to previous values recorded in untreated EAE-inoculated rats during the early onset of disease (Mackenzie et al., 1989; Elderfield et al., 1993). Both prophylactic MK-801 treatments had circulating glucocorticoid levels above normal limits (32.9 ± 20.6; n = 5), but below vehicle control values, this corresponding with the reduced severity of symptoms.

	Discussion

Top Abstract Introduction Methods Results Discussion References

EVBE values recorded in target tissues from EAE-affected, vehicle-dosed rats were equivalent to estimates previously made by us in similar CNS isolates, reconfirming the occurrence of BBB breakdown during early disease and emphasizing the reproducibility of the radioisotope technique for quantitating neuroendothelial leakage (Paul and Bolton, 1995). The present study demonstrates that prophylactic and therapeutic doses of the NMDA receptor antagonist and ion channel blocker MK-801 significantly suppresse neuroantigen-induced cerebrovascular disruption. The reduced symptom expression observed after long-term administration of the drug is in agreement with recent studies by Wallström et al. (1996) using the glutamate receptor antagonist memantine to inhibit the expression of EAE. However, although our investigation clearly demonstrates limited perivascular lesions in the spinal cord 12 days PI after prophylactic treatment with high-dose MK-801, immunohistochemical assessment of cellular infiltration into the CNS suggested no improvement after memantine treatment. Furthermore, our studies confirm MK-801 effects on the course of EAE are not mediated through excess circulating glucocorticoids, which have previously been shown by us and others to influence BBB permeability and determine the course of EAE (Bolton and Flower, 1989; MacPhee et al., 1989; Paul and Bolton, 1995).

Increased permeability of cerebral vessels during the course of MS or with the onset and development of EAE has been well recognized and acknowledged as an intrinsic feature of both conditions (Barlow, 1956; Broman, 1964; Leibowitz and Kennedy, 1972; Gay and Esiri, 1991). However, and despite subsequent studies describing physiological and biochemical abnormalities associated with BBB dysfunction (Kristensson and Wisniewski, 1977; Claudio et al. 1989; Hawkins et al., 1990, 1992), the events which precipitate neuroendothelial leakage remain unclear. Pharmacological studies by us and others have offered an additional approach to elucidating the mechanisms which trigger abnormal leakage at CNS vascular sites (Claudio and Brosnan, 1992; O'Neill et al., 1992; Bolton et al., 1994; Paul and Bolton, 1995). In particular, the use of specific antagonists to characterize receptor-mediated events in BBB breakdown is exemplified through sequential investigations with the quinazoline derivative prazosin to identify the intimate involvement of the cerebroendothelial alpha-1 adrenoceptor in neurovascular dysfunction during EAE (Brosnan, 1985; Brosnan et al., 1986; Goldmuntz et al., 1986; Claudio and Brosnan, 1992).

The usefulness of drug-receptor studies to determine mechanisms inducing BBB disruption is well illustrated by the related work of Koenig et al. (1992) and the recent studies of Miller et al. (1996) who used nonimmune-mediated models of barrier leakage and specific antagonists to identify the cerebrovascular-located NMDA receptor as pivotal in the control of CNS vessel permeability. In particular, Koenig and co-workers showed that neuroendothelial NMDA receptor activation, after binding of the amino acid agonist glutamate, could be prevented by the antagonistic actions of MK-801. Moreover, the study demonstrated that in vivo treatment with the drug inhibited polyamine synthesis which these workers had previously shown to cause BBB breakdown via cytotoxic mechanisms (Koenig et al., 1983a, b, 1989). Interestingly, increased levels of NMDA receptor ligands such as glycine and quinolinic acid have been found in EAE CNS tissue (Honegger et al., 1989; Flanagan et al., 1995) providing the potential to enhance receptor function. Therefore, increased excitatory amino acid availability during EAE may indirectly account for our findings of excess polyamines in neurovascular isolates coinciding with enhanced permeability described in the current study and earlier investigations (Bolton et al., 1994; Paul and Bolton, 1994). Finally, more recent work by us has shown that elevated polyamine levels in CNS tissues from EAE-sensitized rats can be markedly reduced by MK-801 treatment reinforcing a role for the compounds in BBB disruption (Paul et al., 1996).

An additional consequence of NMDA receptor activation is the generation, via a constitutive synthase enzyme, of the vasoactive nitrogen intermediate, NO (Garthwaite et al., 1988, 1989; Southam et al., 1991). NMDA receptor-dependent NO production at cerebroendothelial sites has been closely associated with changes in local blood flow and vascular tone, which could ultimately affect BBB permeability (Faraci and Breese, 1993). In addition, more recent studies have proposed that NO, produced by target tissues during the onset and progression of EAE (Lin et al., 1993; Bolton et al., 1994; Scott et al., 1994; Hooper et al., 1995), may give rise to the generation of longer acting peroxynitrite products with increased cytotoxic potential (Hooper et al., 1995). Indeed, pharmacological studies by us, using selective inhibitors of NO synthase enzymes, strongly support a secondary role for the molecule in the pathogenesis of EAE (Scott et al., 1995; 1996). Therefore, MK-801 may correct abnormal BBB permeability by acting at neurovascular sites to limit NO generation and polyamine production by down-regulating vessel NMDA receptor activation after stimulation by excess glutamate.

In addition to an action on neuroendothelial-associated NMDA receptors MK-801 has well-documented effects on receptors at central neuronal sites (Wong et al., 1986). Hence, the drug could act indirectly to prevent BBB leakage by inhibiting the production of neuronal-derived, permeability-inducing factors disruptive at cerebroendothelial locations. A recent study by Purcell et al. (1996) also provides evidence, through the use of MK-801, for the presence of NMDA receptors on mast cells which appear to be involved in the pathogenesis of EAE (Bö et al., 1991; Levi-Schaffer et al., 1991). Finally, the possibility that MK-801 may have, as yet, unknown peripheral effects which influence the development of EAE cannot be ignored. However, we have excluded the possibility that MK-801 is acting via the consequence of up-regulated endogenous glucocorticoids and preliminary data indicate the drug has no effects on in vitro lymphocyte proliferation or macrophage function.

In conclusion, we have described the potent suppressive actions of the NMDA receptor antagonist, MK-801, on BBB breakdown, lesion formation and symptom onset during early neurological EAE and thus identified a novel mechanism through which neurovascular disruption may occur. NMDA receptor involvement in the pathogenesis of MS has, to our knowledge, not been documented but clearly requires investigation. Studies are ongoing to determine the precise mode of action through which MK-801 corrects abnormal BBB leakage during EAE and thereby indicate possible targets for the therapeutic control of MS.