Endothelial histamine H₁ receptor signaling reduces blood–brain barrier permeability and susceptibility to autoimmune encephalomyelitis

Abstract

Disruption of the blood–brain barrier (BBB) underlies the development of experimental autoimmune encephalomyelitis (EAE) and multiple sclerosis. Environmental factors, such as Bordetella pertussis, are thought to sensitize central endothelium to biogenic amines like histamine, thereby leading to increased BBB permeability. B. pertussis-induced histamine sensitization (Bphs) is a monogenic intermediate phenotype of EAE controlled by histamine H₁ receptor (Hrh1/H₁R). Here, we transgenically overexpressed H₁R in endothelial cells of Hrh1-KO (H₁RKO) mice to test the role of endothelial H₁R directly in Bphs and EAE. Unexpectedly, transgenic H₁RKO mice expressing endothelial H₁R under control of the von Willebrand factor promoter (H₁RKO-vWF^H1R Tg) were Bphs-resistant. Moreover, H₁RKO-vWF^H1R Tg mice exhibited decreased BBB permeability and enhanced protection from EAE compared with H₁RKO mice. Thus, contrary to prevailing assumptions, our results show that endothelial H₁R expression reduces BBB permeability, suggesting that endothelial H₁R signaling may be important in the maintenance of cerebrovascular integrity.

The blood–brain barrier (BBB) involves endothelial cells that line the blood vessels of the central nervous system (CNS) and the tight junction protein complexes between these endothelial cells. The BBB thereby acts as a physical and metabolic barrier by separating the vasculature from the parenchyma of the CNS (1). Breakdown of the BBB is associated with the onset and pathogenesis of multiple sclerosis (MS), a degenerative, demyelinating, inflammatory disease of the CNS (2). In MS and its autoimmune model, experimental autoimmune encephalomyelitis (EAE), activated T cells cross the BBB into the perivascular space of the CNS, causing damage to neurons (2). Surveillance of the CNS by T cells does occur (3), but the paucity of leukocytes found in the CNS under noninflammatory conditions suggests that extravasation of cells across the BBB is tightly regulated. Therefore, identifying factors that modulate BBB permeability may be of therapeutical value in treating inflammatory demyelinating diseases of the CNS.

Bordetella pertussis-induced hypersensitivity to histamine (Bphs/Bphs) is a genetically controlled intermediate phenotype associated with susceptibility to EAE (4). Bphs is a state wherein mice are rendered highly susceptible to histamine after a preceding injection of B. pertussis toxin (PTX) (5). Bphs-susceptible strains of mice die within 30 min after histamine challenge, presumably attributable to hypotensive and hypovolemic shock. The cells targeted during Bphs are not known. Histamine has also been implicated in the pathophysiology of MS and EAE. Increased tissue levels of histamine correlate with the onset of EAE (6–8). Mast cells, the major source of histamine (9), are present in MS lesions (10–12), and evidence of mast cell activation is found in the cerebrospinal fluid (CSF) of patients who have MS (13). In mice, mast cells (14) and their activation (15) are required for early EAE onset and maximal disease severity.

The effects of histamine are mediated by four surface histamine receptors: H₁R, H₂R, H₃R, and H₄R (16). H₁R protein and mRNA are highly expressed in MS lesions (17), and H₁ antihistamines reduce EAE in mice and rats (17, 18). In patients with MS, the use of sedating H₁ antihistamines is correlated with decreased disease incidence and amelioration of symptoms (19, 20). Our laboratory has shown that susceptibility to Bphs and EAE requires expression of Hrh1, the gene encoding H₁R (21).

Expression of H₁R on T cells is required for their full encephalitogenic potential (22), but the contribution of H₁R signaling in other cell types to the pathogenesis of EAE has not been formally addressed. In this study, we focused on endothelial cells because they express H₁R, play a role in both Bphs and EAE, and are important in controlling vascular permeability. We generated mice overexpressing H₁R only on endothelial cells to test the hypothesis that signaling via endothelial H₁R promotes BBB permeability and susceptibility to both Bphs and EAE. Contrary to our hypothesis, endothelial-specific reexpression of H₁R did not restore Bphs susceptibility in Hrh1-KO (H₁RKO) mice. Moreover, we found that selective expression of H₁R on endothelial cells decreased BBB permeability and protected mice from EAE.

Results

Bphs Susceptibility Maps to the Nonhematopoietic Compartment.

In this study, we tested the hypothesis that PTX acts as an ancillary adjuvant in EAE by increasing BBB permeability via hypersensitization of endothelial cells to H₁R signaling. First, we determined whether bone marrow (BM)-derived or non–BM-derived cells were responsible for histamine hypersensitivity by assessing Bphs in reciprocal BM chimeras of Bphs-resistant (Bphs^R) and Bphs-susceptible (Bphs^S) strains of mice (C3H/HeJ and C3H.Bphs^SJL/J, respectively). Bphs^S BM-derived cells failed to rescue the phenotype of Bphs^R C3H/HeJ recipients (Table 1), indicating that the recipient Bphs genotype determined Bphs susceptibility. Both sedating (cross the BBB) and nonsedating (do not cross the BBB) H₁ antihistamines blocked Bphs in PTX-sensitized Bphs^S C57BL/6J mice (Table S1), indicating that histamine was not directly neurotoxic to the CNS. Thus, the H₁R-expressing target cells involved in Bphs are not located exclusively in the CNS or in the BM but, instead, reside in a nonhematopoietic compartment.

Table 1.

Non–BM-derived cells mediate Bphs

Donor	Recipient	No PTX	PTX	P value
C3H/HeJ	C3H/HeJ	0/2	0/4
BphsS	BphsS	0/2	7/8	0.004
BphsS	C3H/HeJ	0/2	2/10
C3H/HeJ	BphsS	0/2	9/10	0.002
Het	Het	0/2	6/7	0.006
Het	C3H/HeJ	0/2	1/6
C3H/HeJ	Het	0/2	6/6	0.007

Reciprocal BM chimeras using C3H/HeJ, C3H.Bphs^SJL/J (Bphs^S), or heterozygous C3H/HeJ × C3H.Bphs^SJL/J F₁ hybrid (Het) mice were generated as described in Materials and Methods. After reconstitution, mice were injected i.v. with 200 ng of PTX, and 3 d later, they were challenged i.v. with histamine (12.5, 6.25, 3.125, and 1.56 mg/kg of histamine). The data are the total number of animals that died within 30 min divided by the total number of animals studied at all doses. Control mice received carrier alone on day 0 and were challenged with 12.5 mg/kg of histamine 3 d later. A χ² test was used to detect the significance of differences.

Endothelial H₁R Overexpression Does Not Rescue Bphs in H₁R-Deficient Mice.

H₁R-expressing endothelial cells are thought to control Bphs and BBB permeability, phenomena that are both important in susceptibility to EAE. In addition, endothelial cells regulate immune cell entry into the CNS during EAE (23). To directly test the hypothesis that endothelial H₁R signaling is responsible for Bphs, increased BBB permeability, and EAE susceptibility, we generated transgenic H₁RKO mice expressing endothelial H₁R under control of the von Willebrand factor promoter (H₁RKO-vWF^H1R Tg). In these mice, endothelial-specific expression of H₁R is driven by the vWF promoter (24) on an H₁RKO background (Fig. S1A). We obtained two founder lines of H₁RKO-vWF^H1R Tg mice (Fig. S1B), but protein expression was only detectable in brain endothelial cells from the higher expressing line 4 (Fig. 1A), which we further studied. Vwf-HA-Hrh1 transgene expression was detectable in spleen, lymph node, and thymus that were depleted of lymphocytes (Fig. S1C). Activation of H₁R increased Ca²⁺ levels in cells from H₁RKO-vWF^H1R Tg mice but not in cells from H₁RKO mice (Fig. 1B). Taken together, these results demonstrate that H₁RKO-vWF^H1R Tg mice express H₁R in endothelial cells of the CNS and lymphoid organs and that CNS-expressed H₁Rs are functional.

Fig. 1.

Endothelial-specific H₁R activation in vivo. (A) Brain endothelial cells from H₁RKO and H₁RKO-vWF^H1R Tg mice were stained with anti-HA mAb, and HA-H₁R expression (red) was visualized by confocal microscopy. Nuclei were stained with DAPI (blue). (B) Isolated brain endothelial cells were loaded with the Ca²⁺-sensing dye Fluo-4, and the change in Fluo-4 fractional fluorescence after stimulation with the H₁R agonist 2-[(3-trifluoromethyl)phenyl] histamine dimaleate was measured by real-time confocal microscopy.

We then assessed whether H₁RKO-vWF^H1R Tg mice were susceptible to Bphs. As expected (21), H₁RKO mice were resistant to Bphs, but, surprisingly, H₁RKO-vWF^H1R Tg mice were also completely resistant (Table 2). Sensitivity to histamine following PTX intoxication is prolonged (25), but mortality was not different between H₁RKO and H₁RKO-vWF^H1R Tg mice up to 1 wk after initial histamine challenge (Table S2). These results show that endothelial-specific H₁R activation does not influence acute or long-term histamine sensitivity following PTX intoxication.

Table 2.

H₁RKO-vWF^H1R Tg mice are resistant to Bphs

Strain	Histamine (mg/kg)	No. affected/ Total
C57BL/6J	100	4/4
	50	4/4
	25	2/2
	12.5	2/2
H1RKO	100	0/15
H1RKO-vWFH1R Tg	100	0/24

Mice were sensitized with 200 ng of purified PTX by i.v. injection on day 0. On day 3, animals were challenged with the indicated doses of histamine (mg/kg) by i.v. injection, and deaths were recorded at 30 min postchallenge. The results are expressed as the number of animals dead divided by the number of animals studied.

Endothelial-Specific H₁R Activation Reduces BBB Permeability.

We assessed the BBB permeability index (SI Materials and Methods and ref. 26) in PTX-treated WT, H₁RKO, and H₁RKO-vWF^H1R Tg mice. PTX-treated H₁RKO mice exhibited the greatest concentration of FITC-BSA in the CSF after systemic administration, followed by WT mice, with the lowest concentration in H₁RKO-vWF^H1R Tg mice (Fig. 2A). Bphs/Hrh1 does not regulate sensitivity to PTX (27), indicating that differences observed were not attributable to ineffective PTX intoxication. Thus, contrary to the current dogma, the results obtained in our model system show that H₁R overexpression in endothelial cells, in fact, negatively regulates BBB permeability, suggesting that H₁R signaling in endothelial cells may be important in the maintenance of CNS barrier integrity.

Fig. 2.

Decreased BBB permeability mediated by endothelial H₁R expression. (A) BBB permeability on day 10 after PTX injection of WT C57BL/6J (n = 9), H₁RKO (n = 9), and H₁RKO-vWF^H1R Tg (n = 6) mice was determined as described in Materials and Methods. Values were normalized to H₁RKO, and a one-sample t test was used to assess significance compared with H₁RKO mice (***P < 0.0001, H₁RKO vs. WT and H₁RKO vs. H₁RKO-vWF^H1R Tg). WT and H₁RKO-vWF^H1R Tg mice were also significantly different from each other (*P < 0.05) as assessed by the Student's t test. (B) Brain endothelial cells from WT, H₁RKO, and H₁RKO-vWF^H1R Tg mice (n = 4 per strain) were stimulated with PBS (no histamine) or with 10 μM histamine for 30 min, and intracellular cAMP was determined by enzyme immunoassay (*P < 0.05; ***P < 0.0001, as assessed by the Student's t test).

Activation of H₁R through Gα_q/11 proteins leads to an elevation of inositol 1,4,5-trisphosphate, which induces Ca²⁺ release from the endoplasmic reticulum of endothelial cells (28, 29). H₁R can directly induce cAMP in cell lines (30) or indirectly increase cAMP levels by enhancing the effect of other cAMP-inducing stimuli (28, 31). With regard to BBB function, increased intracellular Ca²⁺ enhances tight junctions (32) and elevated cAMP increases BBB tightness (33). H₁R triggering increased intracellular Ca²⁺ in H₁RKO-vWF^H1R Tg brain endothelial cells (Fig. 1B); thus, we next asked whether H₁R signaling in endothelial cells could modulate cAMP levels. Endothelial cells from H₁RKO mice did not increase cAMP in response to histamine, but cAMP was induced by WT endothelial cells and was strongly induced by histamine in endothelial cells from H₁RKO-vWF^H1R Tg mice (Fig. 2B). These results suggest that functional overexpression of H₁R in endothelial cells directly contributes to maintenance of BBB integrity through a mechanism that may involve increased cAMP and intracellular Ca²⁺.

H₁RKO-vWF^H1R Tg Mice Are Highly Protected from Active EAE.

Given that breakdown of BBB integrity is implicated in the onset and pathogenesis of EAE/MS (34), we also determined whether H₁RKO-vWF^H1R Tg mice were susceptible to myelin oligodendrocyte glycoprotein peptides 35–55 (MOG_35–55)-induced EAE. Age-matched cohorts of WT, H₁RKO, and H₁RKO-vWF^H1R Tg mice were immunized with MOG_35–55 + complete Freund’s adjuvant (CFA) + PTX, and clinical scores over a 30-d period were recorded. As expected (21, 22), H₁RKO mice developed less severe disease than WT mice (Fig. 3A). Consistent with our findings on Bphs, H₁RKO-vWF^H1R Tg mice also did not develop severe EAE (Fig. 3A); surprisingly, however, EAE symptoms were even less severe in these mice than in H₁RKO mice (Fig. 3A). Similar results were observed when using the 2× MOG_35–55 + CFA protocol, which does not include PTX (Fig. 3B). EAE-associated clinical quantitative trait variables, including cumulative disease score, number of days affected, overall severity index, and peak score, were significantly different between WT, H₁RKO, and H₁RKO-vWF^H1R Tg mice (Table S3); only day of onset was the same among genotypes. Compared with WT mice, the severity of disease parameters were decreased in H₁RKO mice and further reduced in H₁RKO-vWF^H1R Tg mice. Similar disease trait results were obtained for cohorts immunized with the 2× MOG_35–55 + CFA protocol for EAE (Table S4). At 30 d postimmunization with MOG_35–55 + CFA + PTX, we also evaluated brain and spinal cord pathology, including demyelination, monocyte/lymphocyte infiltration, lesion score, and total disease score. Overall spinal cord pathology was significantly reduced in both H₁RKO and H₁RKO-vWF^H1R Tg mice compared with WT mice, but the extent of this reduction was greater in H₁RKO-vWF^H1R Tg mice (Fig. S2). Thus, although global H₁R expression is necessary for full development of EAE, selective overexpression of H₁R on endothelial cells is protective.

Fig. 3.

Endothelial H₁R activation is protective in EAE. EAE was elicited in WT, H₁RKO, and H₁RKO-vWF^H1R Tg mice using a 1× MOG_35–55 + CFA + PTX (A) or 2× MOG_35–55 + CFA (B) immunization protocol. Number in parentheses indicates number of animals per group. Regression analysis revealed that the disease course in mouse strains fit a variable slope sigmoidal curve (shown as solid lines) that was significantly different (P < 0.0001) in all strains as assessed by the extra sum-of-squares F test.

Activation of Endothelial H₁R Does Not Affect Encephalitogenic T-Cell Responses.

T cells and endothelial cells interact extensively at the BBB during EAE (35, 36). To ask whether endothelial H₁R expression would have an impact on the encephalitogenic T-cell response, we examined the ex vivo MOG-specific response in spleen and draining lymph node cells from MOG_35–55 + CFA + PTX-immunized WT, H₁RKO, and H₁RKO-vWF^H1R Tg mice. MOG_35–55-specific proliferative responses did not differ significantly in any of the strains (Fig. 4A). In agreement with our previous results (22), loss of H₁R led to lower IFN-γ levels, higher IL-4 levels, and no change in IL-17 production by restimulated H₁RKO cells compared with WT cells. However, cells from H₁RKO-vWF^H1R Tg mice responded like cells from H₁RKO mice (Fig. 4 B–D). Similarly, none of the 20 other cytokines/chemokines we analyzed by multiplex assay were significantly affected by loss of H₁R or by its overexpression in endothelial cells (Fig. S3). Furthermore, IFN-γ, IL-4, or IL-17 production by MOG_35–55-restimulated cells from 2× MOG_35–55 + CFA-immunized H₁RKO mice was not different compared with cells from H₁RKO-vWF^H1R Tg mice (Fig. S4). In addition, expression of the endothelial activation marker intercellular adhesion molecule (ICAM)-1 (37) was not different in the CNS of 1× MOG_35–55 + CFA + PTX-immunized WT, H₁RKO, or H₁RKO-vWF^H1R Tg mice (Fig. S5). Likewise, endothelial degranulation, as assessed by serum vWF levels, in these mice was not different (Fig. S6). Thus, modulation of H₁R expression did not affect the inflammatory capacity of endothelial cells.

Fig. 4.

Endothelial H₁R expression did not alter the antigen-specific cytokine response in 1× MOG_35–55 + CFA immunized mice. Spleen and draining lymph node cells were isolated from WT, H₁RKO, and H₁RKO-vWF^H1R Tg (n = 15–17 per strain) mice that were immunized 10 d previously with the 1× MOG_35–55 + CFA + PTX protocol. (A) Cells were restimulated ex vivo with the indicated doses of MOG_35–55 for 72 h, and proliferation was determined by [H³]-thymidine incorporation. The mean cpm ± SD were calculated from triplicate wells, and the results shown are representative of two experiments. (B–D) Cells were isolated as in A and restimulated with 50 μg/mL MOG_35–55 for 72 h. Supernatants were harvested and IFN-γ (B), IL-4 (C), and IL-17 (D) levels were determined by ELISA (*P < 0.05; **P < 0.01, as determined using one-way ANOVA with Kruskal–Wallis and Dunn's multiple comparison tests).

Discussion

Susceptibility to EAE and MS is determined, in part, through gene-by-environment interactions. PTX is an example of an environmental factor derived from an infectious agent that influences susceptibility to EAE; as such, it is widely used as an ancillary adjuvant in EAE. Exposure to PTX leads to increased BBB permeability that is thought to be controlled, in part, by sensitization of endothelial cells to vasoactive amines, such as histamine (38). In this study, we provide data demonstrating the opposite: that H₁R signaling in endothelial cells decreases BBB permeability and strongly reduces susceptibility to EAE (Fig. S7).

Endothelial cells express H₁R (28) and regulate vascular tone, BBB permeability, and migration and extravasation of leukocytes into tissues (39, 40). Because H₁R is important for both Bphs and EAE, and because both Bphs and EAE involve some or all of these physiological processes, we asked whether endothelial H₁R signaling could similarly affect EAE or Bphs susceptibility. BM chimera experiments suggested a nonhematopoietic compartment mediating Bphs. Endothelium-specific expression of H₁R did not rescue the Bphs^R phenotype of H₁RKO mice. Our results argue against a failure of H₁R signaling in H₁RKO-vWF^H1R Tg endothelial cells because these cells responded to H₁R stimulation with elevations in both intracellular Ca²⁺ and cAMP. It is also possible that concomitant H₁R expression on another cell type, such as vascular smooth muscle cells or perivascular neurons, is required for Bphs. H₁R signaling has been shown to modulate arterial tone via vascular endothelium and adrenergic and nonadrenergic/noncholinergic perivascular nerves (41). Our results do not support a role for endothelial H₁R signaling as a physiological mechanism for Bphs. However, it is possible that H₁R activation on perivascular nerves stimulates release of vasoactive neurotransmitters to elicit the characteristic hypovolemic and hypotensive shock associated with Bphs.

We have previously reported decreased EAE susceptibility in H₁R-deficient mice (21) and that expression of H₁R in T cells complements EAE in H₁RKO mice (22). In contrast, in this study, we show that overexpression of H₁R on endothelial cells further suppressed the residual disease symptoms present in H₁RKO mice. Mechanistically, we propose that overexpression of H₁R in endothelial cells enhanced histamine-induced Ca²⁺ flux and cAMP production to augment BBB function. The increased BBB function presumably impaired the entrance of encephalitogenic T cells into the CNS, resulting in nearly complete protection from EAE in H₁RKO-vWF^H1R Tg mice. Thus, although H₁R activation in T cells is disease-promoting, its expression on endothelial cells impairs disease development, emphasizing cell-specific roles for susceptibility genes in complex traits, such as EAE.

In addition to their barrier function, endothelial cells store and release inflammatory mediators from Weibel–Palade bodies (WPBs) (42) and cytoplasmic granules (43). Although histamine is known to be a secretagogue for the release of WPB contents (44), our data do not support a role for H₁R in triggering the inflammatory capacity of endothelial cells in the CNS. Modulation of H₁R expression did not affect CNS endothelial activation status during EAE as assessed by ICAM-1 staining. Furthermore, endothelial H₁R overexpression did not affect the MOG_35–55-specific T-cell response compared with that of H₁RKO mice. However, compared with WT mice, both H₁RKO and H₁RKO-vWF^H1R Tg T cells produced less IFN-γ in response to MOG_35–55 in vitro. We believe that the reduced encephalitogenicity of T cells in H₁RKO and H₁RKO-vWF^H1R Tg mice likely reflects a T-cell–inherent defect, presumably attributable to the requirement for H₁R for the full encephalitogenic capacity of T cells (22) rather than an effect on endothelial cells.

Histamine signaling appears to play differing and complex roles in the cerebral microvasculature during EAE. For leukocytes, H₁R signaling promotes rolling in the cerebral microvasculature (45). Histamine also induces release of P-selectin from WPBs (44), which, within the CNS, is required for lymphocyte rolling but not adhesion (46). To our knowledge, it is not clear whether the secretagogue function of histamine for release of endothelial contents is H₁R-dependent. Our data here do not support a role for H₁R in the induction of ICAM-1 on CNS endothelial cells during disease. Nonetheless, regarding the barrier function of endothelial cells, our data indicate an important role for H₁R in this process. H₁RKO mice showed increased BBB permeability compared with WT mice, whereas overexpression of this receptor in the endothelium strongly reduced BBB permeability. We conclude that H₁R activation primarily affects the barrier functions and not the inflammatory function of CNS endothelium during disease. Collectively, the enhanced barrier effect mediated by H₁R in endothelial cells may increase the threshold for CNS immune surveillance by inhibiting activated T cells from entering the CNS without affecting their encephalitogenic potential.

Consistent with this notion, leukocyte infiltration into the spinal cord during EAE was lowest in H₁RKO-vWF^H1R Tg mice. However, compared with H₁RKO mice, the Ag-specific immune response was not affected by restoration of endothelial H₁R. However, if adhesion signals, such as the interaction between α4 integrin or lymphocyte function-associated antigen (LFA)-1 on activated T cells and vascular cell adhesion molecule (VCAM)-1 or intercellular adhesion molecule (ICAM)-1, respectively, on endothelial cells (37), are sufficiently high, T cells could still enter the CNS. Indeed, targeting α4 integrin is a viable but not comprehensive therapeutical approach for MS (47). Thus, we propose that H₁R does not affect the elicitation of inflammatory cues produced by endothelial cells during the induction of disease but that an important role of endothelial H₁R signaling is to increase the stringency for entry of immune cells into the CNS across the BBB. Finally, the results obtained here encourage the speculation that selectively activating H₁R in endothelial cells may be an effective preventative or therapeutical strategy for EAE/MS.

Materials and Methods

Mice.

C57BL/6J (B6) mice and C3H/HeJ mice were purchased from the Jackson Laboratory. B6.129P-Hrh1^tm1Wat (H₁RKO) mice (48) and C3H.Bphs^SJL/J (Bphs^S) congenic mice (21) were maintained in-house at the University of Vermont. All animals were housed at the vivarium of the Given Medical Building at the University of Vermont. All animal studies were approved by the Institutional Animal Care and Use Committee of the University of Vermont.

H₁RKO-vWF^H1R Tg mice were generated by injection of a DNA fragment containing the murine Vwf promoter (24), the HA-tagged Bphs^S Hrh1 allele (22), and the human growth hormone (hGH) intron/polyadenylation signal directly into fertilized C57BL/6J eggs at the University of Vermont Transgenic/Knockout Mouse Facility. Transgene-positive founders were identified by DNA slot-blot using a BamHI-SacI 0.5-kb fragment from the hGH gene as a probe. Two founder lines were generated and were crossed to H₁RKO mice to establish H₁RKO-vWF^H1R Tg mice.

Primary Mouse Brain Endothelial Cell Cultures.

Microvascular endothelial cells were isolated and cultured as previously described (49), except that dialyzed “histamine-free” FBS (22) was used in the culture medium.

Confocal Microscopy.

Brain endothelial cells were fixed with 4% paraformaldehyde, permeabilized with 0.1% (vol/vol) Triton X-100 in PBS, and stained using an anti-HA mAb (Cell Signaling Technologies), followed by incubation with AlexaFluor 568-conjugated goat anti-mouse IgG Ab (Molecular Probes/Invitrogen). DAPI (Sigma–Aldrich) was used as a nuclear marker. Cells were examined by confocal microscopy using a Zeiss LSM 510 META Confocal Laser Scanning Imaging System (Carl Zeiss USA Microimaging, Inc.).

Ca²⁺ Imaging in Endothelial Cells.

Mouse brain endothelial cells were loaded with Fluo-4 (10 μM) for 7 min at 35 °C in the presence of pluronic acid (2.5 μg/mL). Ca²⁺ was imaged using a Revolution confocal system with an electron multiplying charge-coupled device camera (Andor Technology) mounted on an upright Nikon microscope with a 60× water-immersion objective (1.0 N.A.). After adding the H₁R agonist 2-[(3-trifluoromethyl)phenyl] histamine dimaleate, images were acquired at 15 frames per second with Revolution TL acquisition software (Andor Technology). Fluo-4 fluorescence was excited by a krypton/argon laser (488 nm), and emitted fluorescence was collected at 495 nm. The images were processed using custom-designed software (supplied by A. Bonev, University of Vermont), and the fractional fluorescence was evaluated by dividing the fluorescence of a region of interest (ROI; 5 × 5 pixels per box) by the average fluorescence of 10 images from the same ROI.

Determination of cAMP Production.

Mouse brain endothelial cells were cultured in collagen-coated six-well plates and treated with PBS or 10 μM histamine for 30 min. Cells were then washed twice with ice-cold PBS and immediately lysed with 1 mL of 0.1 M HCl in PBS. The protein concentration in lysates was adjusted to 1.6 mg/mL with PBS, and cAMP was measured using an enzyme immunoassay kit (Cayman Chemical) according to the manufacturer's instructions.

PTX-Induced Sensitization to Histamine.

As described previously (21), mice were injected i.v. with 200 ng of purified PTX (List Biological Laboratories) dissolved in buffer containing 25 mM Tris, 0.5 M NaCl, and 0.017% (vol/vol) Triton X-100. Control mice received carrier. Three days later, sensitivity to histamine was determined by i.v. injection of indicated doses (mg/kg of dry base) of histamine dihydrochloride (Sigma–Aldrich) in PBS. H₁ antihistamines were injected i.v. 20 min before histamine challenge. Deaths were recorded 30 min after histamine challenge or as indicated. Control mice received PBS. The results are expressed as the number of deaths divided by the number of animals studied.

BM Chimeras.

Femurs from 6- to 8-wk-old C3H/HeJ, C3H.Bphs^SJL/J (Bphs^S), or heterozygous C3H/HeJ × C3H.Bphs^SJL/J F₁ hybrid (Het) mice were aspirated with RPMI through an 18-gauge needle to obtain BM cells. BM cells were washed and resuspended in RPMI and kept on ice until injection. Recipient mice (also 6–8 wk old) were irradiated with 700 rad divided over two doses given 4 h apart. Immediately following the second irradiation, mice received 2 × 10⁶ BM cells of the indicated donor genotype i.v. and were allowed to reconstitute for 8 wk. Mice were then injected with 200 ng of PTX by i.v. injection, and 3 d later, they were challenged i.v. with histamine (12.5, 6.25, 3.125, and 1.56 mg/kg). Control mice received carrier alone on day 0 and were challenged with 12.5 mg/kg of histamine 3 d later.

Determination of BBB Permeability.

Determination of BBB permeability was performed essentially as described (26). Briefly, mice were injected i.v. with 200 ng of PTX, and 10 d later, they were given 50 μg/g of FITC-labeled BSA (Sigma–Aldrich) by i.v. injection. Four hours later, CSF and plasma were isolated (SI Materials and Methods), diluted in PBS, and centrifuged at 835 × g for 15 min. The fluorescence intensity (FI; excitation wavelength of 485 nm, emission wavelength of 528 nm) of FITC in the CSF and serum samples was determined with a microplate fluorescence reader (Flx-800-I; Bio-Tek Instruments, Inc.). The BBB permeability index is expressed as the ratio of the CSF FI divided by the plasma FI.

Induction and Evaluation of EAE.

Mice were immunized for MOG_35–55-induced EAE and scored for clinical quantitative trait variables and quantitative histopathology as described previously (22). Non-linear regression analyses (50) were performed to examine the effect of strain on measures of disease severity (SI Materials and Methods).

Ex Vivo MOG_35–55 Responses.

Mice were immunized for the induction of EAE, and 10 d later, spleens and draining lymph nodes were isolated. For proliferation assays, 5 × 10⁵ cells per well were cultured in 96-well plates for 72 h at 37 °C in the presence of 0.05, 0.5, 2, 10, or 20 μg/mL MOG_35–55. Cells were pulsed for the last 18 h of culture with 1.0 μCi of ³H-thymidine (PerkinElmer) and harvested onto glass fiber filters, and thymidine uptake was determined by liquid scintillation (Tomtec, Inc.). Counts per minute were measured, and data are presented as the mean ± SEM of triplicate wells. For cytokine assays, 1 × 10⁶ cells per well were cultured in 24-well plates for 72 h in the presence of 50 μg/mL MOG_35–55. Supernatants were harvested, and cytokines were analyzed by ELISA as described previously (22).

Statistical Analysis.

Statistical analyses, as indicated in the figure legends, were performed using GraphPad Prism 4 software (GraphPad Software, Inc.).