Journal of Autoimmunity (1990) 3,237-246

Experimental Autoimmune Uveoretinitis in Mice. Induction by a Single Eliciting Event and Dependence on Quantitative Parameters of Immunization

Rachel R. Caspi, * Chi-Chao Ghan,* William C. Leake,* Makoto Higuchi ,* Barbara Wiggertt and Gerald J. Chaderl_* Laboratory

of Immunology and TLaboratory of Retinal Cell and Molecular National Eye Institute, NIH, Bethesda, MD 20892, USA

Biology,

Experimental autoimmune uveoretinitis (EAU) in the mouse is a recently developed model of ocular autoimmunity. Dependence of disease induction on qualitative and quantitative parameters of immunization was studied in BIO.A mice immunized with interphotoreceptor retindid-binding protein (IRBP). It was found that use of BordeteZZupertussis adjuvant as well as its mode of preparation was of critical importance for disease induction; no disease was induced if pertussis adjuvant was omitted. The minimal effective protocol for EAU induction when the vaccine form of B. pertussis adjuvant was used consisted of pretreatment with cyclophosphamide, two divided doses of IRBP in complete Freund’s adjuvant (CFA), and two divided doses of B. pertussis vaccine. Any reduction in the immunization schedule resulted in reduced incidence of disease. In contrast, substituting purified B. pertussis toxin (PTX) for the vaccine allowed reduction of the immunization schedule to a single dose of IRBP in CFA and omission of the cyclophosphamide pretreatment. Severity and incidence of disease could be quantitatively controlled by varying the respective doses of IRBP and PTX. In addition, a chronic or an acute clinical course of EAU could be obtained by using either a low-dose or a high-dose immunization, respectively. Establishment of a single dose induction protocol and the quantitation of theimmunopathogenic response as a function of the variables of immunization lay the foundation for the further development and utilization of this promising model of ocular autoimmunity.

Introduction Experimental autoimmune uveoretinitis (EAU), induced in rodents and in primates by immunization with retinal antigens, has characteristics similar to a variety of Correspondence to: RachelR. Caspi,PhD, Laboratory of Immunology, NationalEye Institute, NIH Building 10, Room lON216, Bethesda, MD 20892, USA. 237 0896~8411/90/030237+

10 $03.00/O

0 1990 Academic Press Limited

238

R. R. Caspi et al.

human ocular inflammatory diseases of a presumed autoimmune nature [l-3]. As with experimental autoimmune encephalitis (EAE) and adjuvant arthritis [4], EAU in the rat serves as a model system for the study of organ-specific autoimmunity mediated by T lymphocytes, yielding new insights into such basic questions as the role of immunodominance in the recognition of autopathogenic epitopes and the ability of organ-resident cells to regulate responses of autoimmune T lymphocytes [5-71. Development of an EAU model in the mouse, which is by far the bestcharacterized species immunologically as well as genetically, is important for furthering the understanding of immunogenetic and cellular mechanisms involved in ocular autoimmunity. Recently, we have reported the successful induction of EAU in the mouse, which had been considered to be a species refractory to this experimental disease [8]. The disease was induced in strains of the H-2k haplotype, and in related strains having the H-2k serological specificity in the Class II region, by using a high-intensity immunization protocol. The protocol consisted of pretreatment with cyclophosphamide (CY) on day - 2, immunization with antigen emulsified in complete Freund’s adjuvant (CFA) on days 0 and 7, and treatment with Bordetellu pertussis vaccine as an additional adjuvant on days 0 and 3. Murine EAU induced by this method with the interphotoreceptor retinoid-binding protein (IRBP) appears to approximate closely some types of human uveitis, for which no satisfactory model has existed before, because of its focal pathology and its relatively chronic course [8]. Although effective in inducing EAU, the need for a split-dose protocol of immunization severely limits the usefulness of this model. Some types of studies cannot be pursued in a model where the immune response is manipulated at several time points during the process of induction. Examples of this are studies involving anterior chamber-associated immune deviation (ACATD), which comprises a series of temporally-defmed induction events and involves participation of cyclophosphamidesensitive as well as cyclophosphamide-resistant suppressor cells [9]. We therefore set out to define a minimal immunization protocol, with the purpose of reducinginduction of disease to a single eliciting event. The present report describes the adaptation of murine EAU to a single-dose induction protocol, explores the influence of variables, such as the dose of antigen and the dose and type of pertussis-derived adjuvant, and identifies the characteristics of disease obtained under the different conditions of immunization. Materials and methods Mice

Male and female Bl0.A mice were purchased from Jackson Laboratory, Bar arbor, Maine, at 8 weeks of age. Animals were housed under specific pathogen-free conditions, were given water and laboratory chow (Purina) ad libitum and were used at 2-5 months of age. Antigens and adjuvants

IRBP was prepared from bovine retinas as reported previously, by Concanavalin A (Con-A) Sepharose afIinity chromatography and fast performance liquid

EAU in mice

239

chromatography (FPLC) [lo]. Antigen preparations were aliquoted and stored at - 20°C until use. B. pertussis vaccine preparations were purchased from Michigan Department of Public Health, Lansing, MI, Massachusetts Public Health Labs, Boston, MA, USA, and Pasteur Institute, Marmes La Coquette, France. The vaccine consisted of a formaldehyde-inactivated suspension of bacterial cells in phosphatebuffered saline; the Pasteur Institute vaccine additionally contained 1.25 mg of aluminum hydroxide for each 5 x lo9 bacteria. Purified B. pertussis toxin (PTX), bacterial lipopolysaccharide (LPS, S. typhimurium) and muramyl dipeptide (MDP) were from Sigma, St Louis, MO. AS-101, an immunoenhancing organic tellurate compound [ 111, was kindly donated by Dr B. Sredni of Bar-Ilan University, Israel. Immunization

For the original split-dose protocol [8], mice were pretreated intraperitoneally (i.p.) 2 days before immunization with 20 mg/kg of cyclophosphamide (CY), immunized on day 0 with 50 ug of antigen of CFA (0.1 ml of emulsion) in both hind footpads (fp), and given 0.5-2 x 10” B. pertussis organisms i.p. Three days later the mice were given a second i.p. dose of B. pertussis, and 7 days later a second identical dose of antigen in CFA intramuscularly (i.m.), divided between both hind legs. Modifications of the protocol (as described more fully in Results) consisted of: omitting CY pretreatment, injecting both doses of antigen/CFA at the same time (fp and i.m.), varying the batch of pertussis vaccine, supplementing the pertussis vaccine with additional immuno-enhancing agents or replacing it with PTX, and finally varying the respective doses of antigen and PTX. Histopathology

Freshly enucleated eyes were fixed for 1 h in 4% phosphate-buffered glutaraldehyde and transferred into 10% phosphate-buffered formaldehyde until processing. Fixed and dehydrated tissue was embedded in methacrylate or in paraffin, and 4-6 micron sections were stained by standard hematoxylin and eosin. EAUgrading

and presentation

of results

Six to eight sections cut at different planes were examined for each eye in a masked fashion by an independent observer, and the presence and extent of lesions were determined. Incidence and severity of EAU were scored on a scale of 0 to 4 in halfpoint increments, according to the semiquantitative grading system previously described [8]. Incidence is shown as the number of positive animals out of all animals in the group. Severity of disease is calculated as mean grading of all eyes from affected animals only. Results Feasibility

of reduction of the immunization

schedule depends on the potency of the

pertussis adjuvant

In a previous publication [8], we showed that EAU in mice could be induced by a high-intensity immunization schedule, involving pretreatment with cyclophosphamide on day - 2, immunization with antigen emulsified in CFA on days 0 and 7, and

240

R. R. Cad Table

et al.

1. Dependence

of pertussis

vaccine-supported EAU induction and dual immunization

on

cyclophosphamide pretreatment

Group Treatment‘

Cyclophosphamide 1st immunization 2nd immunization EAU incidence

1

(pg IRBP/mouse) (pg IRBP/mouse)

yes yes (50) yes (50) 418

2

3

yesnPs0)

yes yes (100)

yes (50) 018

l;“s

‘CY pretreatment was on day -2, first immunization on day 0 and second immunization on day 7. All animals were given 1O’Oorganisms of B. pertussis on day 0 and 3. EAU incidence is by histopathology.

treatment with Bordetellapertussis vaccine as an additional adjuvant on days 0 and 3. Table 1 shows an attempt at reduction of this complex immunization protocol. Groups of mice were injected with the same total dose of antigen/CFA and pertussis vaccine, using a single-dose or a split-dose protocol, with or without pretreatment with CY. The results clearly showed that, under these conditions, any attempt at reducing the original immunization protocol resulted in lowered incidence of disease. An important limiting factor in the induction of EAU appeared to be the pertussis adjuvant. Without the pertussis vaccine, EAU could not be induced at all (Table 2). Moreover, only one vaccine preparation out of four that were tested was able to support EAU induction. Bacterial LPS, muramyl dipeptide (MDP) and the immunoenhancing compound AS- 101 [ 111, in a variety of doses and treatment schedules, were unable to supplement effectively inactive vaccine preparations (results not shown). In contrast, a purified preparation of the pertussis toxin (PTX) was highly efficient in supporting EAU induction (Table 2). Three different lots oftoxin were subsequently compared in parallel for supporting EAU induction, and no obvious variability in the efficacy of the different lots was seen (not shown). Reduction of the immunization protocol was attempted using PTX in place of the vaccine. Groups of animals were immunized with the same total dose of 100 pg antigen and 1,000 ng of PTX, by the original full schedule, by split-dose immunization without CY pretreatment, or by single-dose immunization without CY (Figure 1). It was found that use of PTX allowed omission of the CY pretreatment and administration of the antigen in CFA in a single dose, without sacrificing either the incidence or the severity of the resulting disease.

Severity

and incidence of disease can be controlled by varying PTX and antigen

the respective doses of

Having established that a single-dose immunization protocol can effectively elicit EAU in mice, it was important to establish the respective doses of antigen and PTX needed to induce disease. For this purpose a checkerboard titration was performed,

EAU in mice

241

Table 2. Dependence of EAU induction on pertussis aa’juvant EAU Pertussis (Dose/injection)

Group number

Score

Incidence

None

1

induction: (+ SD) 0

016 2.0f

1.3

2

Vaccine A’ (1 x 10” organisms)

617

3

Vaccine B (2 x 10” organisms)

015

0

4

Vaccine C (2 x 10” organisms)

O/5

0

5

Vaccine D (5 x lo9 organisms)

O/6

0

Toxin

6

2.7 f0.9

616

(500 ng) ‘Vaccine A: Michigan Dept. of Public Health, Lansing, MI, Lot # 91. Vaccine B: Michigan Dept. of Public Health, Lansing, MI, Lot # 93. Vaccine C: Massachusetts Public Health Labs., Boston, MA, Lot # WF262. Vaccine D: Pasteur Institute, Marmes La Coquette, France. Lot # 47. Animals were immunized by the split-dose protocol, after CY pretreatment. The specified number of B. pertussis organisms (groups 2-5) or the specified amount of pertussis toxin (group 6) were injected i.p. in 0.1 ml PBS on day 0 and on day 3 after primary immunization. EAU score.(4 weeks after primary immunization) was calculated as the average score of all eyes from all affected animals.

1st ”

imm. PTX

2nd PTX

2nd Imm.

d.0

d.3

6.7

1st

I

4.2

d.0

IMMUNIZATIONSCHEDULE

0

.

I

1

.

I

2

*

I

3

AVERAGE EAU SCORE @SD)

.

I

4 INCIDENCE (POSTTOTAL)

Figure 1. Use of PTX permits reduction of the immunization protocol to a single dose of antigen and toxin. All animals received the same total dose of IRBP (100 ug) and PTX (1,000 ng) either as one or as two injections. CY was at 20 mg/kg, where used. Disease incidence (positive/total) and severity (on a scale of 0 to 4) were determined by histopathological examination of sectioned eyes at 4 weeks after immunization. Values given are the mean k standard deviation from the mean of animals from five experiments.

the doses of PTX and IRBP were varied independently among the groups (Figure 2). The highest incidence and score of disease were obtained when both IRBP and PTX were at a high dose. Progressive reduction in the dose of antigen at a constant dose of PTX, or in the dose of PTX at a constant dose of antigen, resulted in a quantitative diminution of the mean incidence as well as the severity of EAU. EAU where

242

R. R. Caspi et al. PTX (ng/mouse)

Incid. 94% (16117) Incid. Score

2.4 + 1.1

Score

97%

(15/16)

1.6fl.O

InCid.

Score

Incid. 80% (1~12)

Incid. 62% (16122

hid.

Score 2.1 f 1.4

Score

Score

Incid. 30% (6/20)

kid.

Scare

Score

1.1 ?I0.9

Incid. 13% Score

(206)

0.9 f 0.9

1.8 f 1.1

23% (3113) 0.4i

1

Incid. 6% (1117) Score

0.4 f 0

hid.

50%

(W16)

1.3fl.l

25% (4/16) 12 f 0.7

17% (3/l 8)

hid.

24% (4/l 7)

Score

0.9 k 0.6

hid. Score

bid.

12% (2117; 1.1 f 0.2

11%

(~18)

0.9 f 1.2

Score

0.4 i 0.2

Incid. 0% (0117)

hcid.

0% (o/16)

Swre

Score

Swre

0

0

Figure 2. Checkerboard titration of the amounts of IRBP and PTX needed for single-dose induction of EAU. Animals were immunized on day 0 (without CY pretreatment). Disease incidence (positive/total) and severity (on a scale of 0 to 4) were determined by histopathological examination 4 weeks after immunization. (Values given are the mean f SEM of three separate experiments). Animals injected with a high dose of either IRBP + CFA alone or PTX + CFA alone, did not develop disease.

could not be induced by antigen + CFA alone, or by PTX + CFA alone, even at the highest doses (not shown). Course of disease and pathological

manifestations dose

are determined

by the immunization

Use of PTX at high-dose immunization changed the clinical course and the pathological manifestations of EAU in comparison to our previous results obtained in the presence of pertussis vaccine [8]. At highest doses of toxin and antigen (100 ug IRBP+ 1,000 ng of PTX, see Figure 2), most animals developed an early-onset, acute disease of short duration, with extensive serous retinal detachment and diffuse photoreceptor damage that was very similar in appearance to the disease typical of the Lewis rat [l-3]. At lower doses of immunization (50 ug IRBP+ 500 ng of PTX, or less), the chronic form of disease was predominant, with a later onset, focal photoreceptor damage, retinal vasculitis, some choroidal thickening and mild vitritis. Figure 3 shows the chronic form [panels (a) and (c)] and the acute form [panels (b) and (d)] of the disease at 2 weeks and at 4 weeks after immunization. In contrast to the extensive destruction of the photoreceptor cell layer of the retina seen in the late stage of the acute form of disease [panel (d)], most of the photoreceptors in the chronic form of EAU were preserved even at 8 weeks postimmunization, when most of the eyes were essentially quiescent (not shown). This form of disease was virtually identical to the EAU obtained previously with pertussis vacccine and split-dose immunization [8]. Discussion This report describes the development and quantitation of a single-dose induction protocol for the murine EAU model induced with IRBP in the BIO.A strain of mice.

Figure 3. EAU induced by high- compared to low-dose immunization. High-dose immunization was 100 ug of IRBP and 1,000 ng PTX. Low dose immunization was 50 ug of IRBP and 500 ng PTX. Paraffin-embedded sections stained with hematoxylin and eosin ( x 160): (a) Two weeks, low-dose. No pathology is apparent (V = vitreous, R = retina, C = choroid); (b) Two weeks, high-dose. Active-stage disease with extensive serous retinal detachment (asterisk) subretinal hemorrhage (arrow) and inflammatory cell infiltration (open arrow); (c) Four weeks, low dose. Active-stage disease with focal infiltration of the retina (arrows) and choroid (open arrow), and mild vitritis (arrowhead), (d) Four weeks, high dose. Late-stage disease, with diminished inflammatory infiltrate and complete destruction of the photoreceptor cell layer (arrow).

C

244

R. R. Caspi et al.

The original induction schedule [8] depended on two injections of antigen in CFA given one week apart, after pretreatment of the animals with CY, and using pertussis vaccine as an additional adjuvant. A similar high-intensity protocol has previously been shown to be effective in inducing experimental autoimmune encephalitis in some resistant strains of mice [ 121. In order to develop a minimal effective protocol of disease induction, it was important to dissect the relative importance of the various components in the immunization process. The relative importance of the CY pretreatment and the number of immunizations for the induction of EAU pathology were assessed under conditions where pertussis adjuvant treatment remained identical between groups (Table 1). Omission of the CY pretreatment resulted in lack of EAU induction, suggesting that endogenous suppressor mechanisms may play a role in this model. Similarly, injecting the total amount of antigen in a single rather than a split dose, resulted in reduced EAU incidence, suggesting that prolonging the duration of the immunological stimulus is also important. Conversely, the important of pertussis adjuvant was studied while maintaining the other variables constant (Table 2). The critical importance of the pertussis adjuvant was indicated by several observations. Pertussis treatment was an absolute prerequisite for disease induction, with the purified toxin being superior to the vaccine form. In fact, several vaccine preparations were unsuitable for support of EAU induction, and a number of stimulants with a known adjuvant effect (MDP, LPS and AS-101) were unable to reconstitute the activity. This suggests that the ‘active ingredient’ in the pertussis vaccine for promoting EAU induction in the mouse is the toxin component, which may be lower in some vaccine preparations than in others. Notably, use of the toxin in place of the vaccine overcame the need for CY pretreatment and allowed a reduction of the number of immunizations to a single dose, without sacrificing either the incidence or the severity of the resulting disease (Figure 1). The importance of this finding for the further development of the murine EAU model cannot be overemphasized, since the ability to elicit disease by a single induction event puts the mouse on a par with the rat for usefulness as a ‘general’purpose’ model of ocular autoimmunity. The mode of action of the pertussis adjuvant in promoting the development of organ-specific autoimmunity is still unresolved, and may involve several steps in the induction and the effector phases of the immunopathogenic process. PTX has mitogenie activity on T-lymphocytes [13], and in viva PTX treatment increases the proportion of the T-helper lymphocyte subset in comparison to the suppressor subset [ 141. These properties of PTX might facilitate development of autoaggressive clones of CD4+ lymphocytes, and may help to explain its ability in the murine EAU system to overcome the need for pretreatment with CY (presumed to downregulate suppressor mechanisms) and for prolonging the immune stimulus by double immunization. In the murine EAE model, treatment of primed donor animals with pertussis was found necessary for the detection and recovery of antigen-reactive cells capable of adoptively transferring disease, while treatment of recipient animals was not essential for EAE induction [ 151. Similarly, in the rat EAU model, lymphocytes from PTX-treated donors had enhanced ability to transfer disease, while treatment of recipient animals actually delayed disease induction [ 161. These studies suggest a primary effect of PTX on the lymphoid cells involved in the pathogenesis of disease.

EAU in mice

245

A possible contribution of PTX in the effector phase of disease induction may involve enhancement of vascular permeability, facilitating extravasation of lymphocytes into the target organ. This mechanism was suggested by Linthicum et al. on the basis of their findings in the mouse model of EAE, where PTX was found to sensitize the microvasculature to vasoactive amines released from mast cells [ 17,181. Involvement of mast cells in ocular autoimmunity was suggested in the rat EAU model, where the number of choroidal mast cells was found to correlate with susceptibility to EAU induction [ 191. In the present study, the incidence and the severity of disease were quantitatively dependent on the respective doses of PTX and antigen; reduction in either stimulus was followed by a concomitant reduction in EAU intensity. In contrast to the pattern of disease previously observed in the presence of pertussis vaccine (where EAU was always of the chronic type), it is now possible to control the course of the disease and the type of pathology induced, by varying the respective doses of antigen and PTX. The mild, chronic form of the disease is obtained at low-dose immunization (50 ug IRBP and 500 ng PTX, and lower), while high-dose immunization (100 ug IRBP and 1,000 ng PTX) induces an acute form of EAU characterized by a rapid onset, short duration and massive photoreceptor destruction, similar to the disease obtained in the Lewis rat [l-3]. It has been proposed that EAU (which, in contrast to human uveitis, is a nonrelapsing disease in the acutely responding Lewis rat) is a self-limiting condition, relapses being prevented by elimination of the source of autologous antigen. We have evidence that the chronic type of EAU in the mouse, where damage to the photoreceptors is moderate, can take the form of a relapsing disease, with lesions reappearing at a later time (Chan et al., in press, and Caspi et al., in press). Thus, the ability to control the type of disease by adjusting the intensity of the immunological stimulus, makes the mouse EAU a versatile model that can be adapted to the needs of the particular experimental system. In summary, the present study shows that it is possible in the murine model to elicit EAU by a single immunization event, and to control the disease intensity and its clinical course by means of adjusting the immunization dose. The quantitation of the immunopathogenic response as a function of the variables of immunization lays the foundation for the further development and utilization of this promising model of ocular autoimmunity.

References 1. Faure, J. P. 1980. Autoimmunity and the retina. Curr. Top. Eye Res. 2: 215-301 2. Gery, I., M. Mochizuki, and R. B. Nussenblatt. 1986. Retinal specific antigens and immunopathogenic processes they provoke. Prog. Retinal Res. 5: 75-109 3. Caspi, R. R. 1989.Basic mechanisms in immune-mediated uveitic disease. In ZmmunoZogy ofEye Disease. S. L. Lightman, ed. MTP Press Ltd, London. pp. 61-86 4. Cohen, I. R. 1986. Regulation of autoimmune disease: physiological and therapeutic. Immunol. Rev. 94: 5-21 5. Sanui, H., M. T. Redmond,

S. Kotake, B. Wiggert, L.-H. Hu, H. Margalit, J. A. Berzofsky, G. J. Chader, and I. Gery. 1989. Identification of an immunodominant and highly immunopathogenic determinant in the retinal interphotoreceptor retinoidbinding protein (IRBP). J. Exp. Med. 169: 1947-1960 6. Caspi, R. R., F. G. Roberge, and R. B. Nussenblatt. 1987. Organ-resident, nonlymphoid cells suppress proliferation of autoimmune T lymphocytes. Science 237: 1029-1032

246 R. R. Caspi et al. 7. Roberge,

a.

9.

10.

11.

12. 13.

14.

15.

16.

17.

18.

19.

F. G., R. R. Caspi, and R. B. Nussenblatt. 1988. Glial retinal Miiller cells produce IL1 activity and have a dual role in regulation of autoimmune T-lymphocytes: antigen presentation manifested after removal of suppressive activity. J. Zmmunol. 140: 2193-2196 Caspi, R. R., F. G. Roberge, C.-C. Chan, B. Wiggert, C. J. Chader, L. A. Rozenszajn, Z. Lando, and R. B. Nussenblatt. 1988. A new model of autoimmune disease: experimental autoimmune uveoretinitis induced in mice with two different retinal antigens. J, Immunol. 140: 1490-1495 Streilen, J. W. 1987. Immune regulation and the eye: a dangerous compromise. FASEB J. 1: 199-208 Redmond, T. M., B. Wiggert, F. A. Robey, N. Y. Nguyen, M. S. Lewis, L. Lee, and G. J. Chader. 1985. Isolation and characterization of monkey interphotoreceptor retinoidbinding protein, a unique extracellular matrix component of the retina. Biochemistry 24: 787-793 Sredni, B., R. R. Caspi, A. Klein, Y. Kalechman, Y. Danziger, M. BenYa’akov, T. Tamari, F. Shalit, and M. Albeck. 1987. A new immunomodulating compound (AS-101) with potential therapeutic application. Nature 330: 173-176 Lando, Z., D. Teitelbaum, and R. Arnon. 1980. Induction of experimental allergic encephalomyelitis in genetically resistant strains of mice. Nature (Lond.) 257: 551-552 Fish, F., J. L. Cowell, and C. R. Manclark. 1984. Proliferative response of immune mouse T-lymphocytes to the lymphocytosis-promoting factor of Bordetella pertussis. Infect. Zmmun. 44: l-6 Sewell, W. A., J. J. Munoz, R. Scollay, and M. A. Vadas. 1984. Studies on the mechanism of the enhancement of delayed-type hypersensitivity by pertussigen. J. Zmmunol. 133: 1716-1722 Lando, Z. and A. Ben-Nun, 1984. Experimental autoimmune encephalomyelitis mediated by T-cell line. II specific requirements and the role of pertussis vaccine for the in vitro activation of the cells and induction of disease. Clin. Zmmunol. Zmmunopathol. 30: 290-303 McAllister, C. G., B. I’. Vistica, R. Sekura, T. Kuwabara, and I. Gery. 1986. The effects of pertussis toxin on the induction and transfer of experimental autoimmune uveoretinitis. Clin. Zmmunol. Immunopathol. 39: 329-336 Linthicum, D. S., J. J. Munoz, and A. Blaskett. 1982. Acute experimental encephalomyelitis in mice. I. Adjuvant action of Bordetella pertussis is due to vasoactive amine sensitization and increased vascular permeability of the central nervous system. Cell. Immunol. 73,229-310 Linthicum, D. S. and J. A. Freilinger. 1982. Acute autoimmune encephalomyelitis in mice. II. Susceptibility is controlled by the combination of H-2 and histamine sensitization genes.Y. Exp. Med. 155: 31-40. Mochizuki, M., T. Kuwabara, C. C. Chan, R. B. Nussenblatt, D. D. Metcalfe, and I. Gery. 1984. An association between susceptibility to experimental autoimmune uveitis and choroidal mast cell numbers. J. Immunol. 133: 1699-1701