CLINICAL IMMUNOLOGY AND IMMUNOPATHOLOGY 49, 149-158 (1988)

BRIEF COMMUNICATION

HLA-DR2 Association with Excessive Somnolence in Narcolepsy Does Not Generalize to Sleep Apnea and Is Not

Accompanied by Systemic Autoimmune Abnormalities 1

ROBERT L. RUBIN, ROZA M. HAJDUKOVICH, AND MERRILL M. MITLER

Department of Basic and Clinical Research, Research Institute of Scripps Clinic, 10666 North Torrey Pines Road, La Jolla, California 92037

Recent reports that nearly all patients with narcolepsy have the HLA-DR2 phenotype suggest that autoimmunity may underly the etiology or pathogenesis of this disorder. Of 11 narcoleptic patients in the present study, 9 were HLA-DR2, confirming the strong association with this class II antigen but indicating that this is not an obligatory pheno­type. In contrast only 3/10 patients with sleep apnea were HLA-DR2, suggesting that this form of excessive somnolence has a different etiopathogenesis. Significant levels of rheumatoid factor, antinuclear antibodies or autoantibodies to native DNA, denatured DNA, histones, Sjogren's syndrome B antigen, or Smith antigen were undetectable in sera from narcoleptic patients. Antibodies to rodent brain, primate brain stem, and neurocytotoxic antibodies were also not found. These results along with the absence of laboratory signs and clinical features of a systemic inflammatory process indicate that if narcolepsy is an autoimmune disease, the underlying lesion or pathologic condition may be confined to the central nervous system. © 1988 Academic Press, Inc.

INTRODUCTION

Narcolepsy is a disease of unknown etiology characterized by spontaneous daytime somnolence often accompanied by pathologic rapid eye movement (REM)2 sleep periods. Honda and colleagues (1-3) and Langdon et al. (4) showed a universal association between narcolepsy and a class II product of the major histocompatibility complex (MHC), the HLA-DR2 phenotype which occurs in 22-34% ofnonnarcoleptic controls. The narcolepsy/DR2 association is the highest HLA antigen correlation with any disease described to date.

These findings have generated speculation that predisposition to narcolepsy or some forms of narcolepsy is inherited and that narcolepsy may be an autoimmune disorder (5) since HLA-DR2 is associated with systemic lupus erythematosus (SLE) (6, 7), especially in blacks (8), and with a high incidence of drug toxicity in rheumatoid arthritis patients (9). It may be particularly relevant that HLA-DR2 is

1 This work was supported by Grant NS 20459 from NINCDS and USPHS Grant RR00833. This is publication number 5099BCR from the Research Institute of Scripps Clinic, La Jolla, CA.

2 Abbreviations used: ANA, anti-nuclear antibody; ELISA, enzyme-linked immunosorbent assay; ESR, erythrocyte sedimentation rate; HLA, human (histocompatibility) leukocyte antigen; IF, immu­nofluorescence; MHC, major histocompatibility complex: MSLT, multiple sleep latency testing; PAGE, polyacrylamide gel electrophoresis; PBS, phosphate-buffered saline; REM, rapid eye move­ment; RF. rheumatoid factor; RFLP, restriction fragment length polymorphism; SDS, sodium dode­cylsulfate; SLE, systemic lupus erythematosus.

149 0090-1229/88 $1.50 Copyright© 1988 by Academic Press. Inc. All rights of reproduction in any form reserved.

150 BRIEF COMMUNICATION

also associated with the neurological diseases, multiple sclerosis (IO, 11) and optic neuritis (11). It is also of interest that circulating autoantibodies to antigens de­rived from nervous tissue are commonly found in the central nervous system disorders associated with SLE (12), Guillain-Barre syndrome (13), and sensory carcinomatous neuropathy (14). Evidence that narcolepsy has an autoimmune component could have significant impact on the understanding of its etiopatho­genesis and may open up new avenues of treatment for this disorder.

The present report replicates the HLA findings and extends HLA phenotyping data to patients with sleep apnea, the most common sleep disorder giving rise to the symptom of excessive somnolence (15). A comprehensive search for systemic autoimmune abnormalities of the humoral immune system in patients with nar­colepsy and sleep apnea was undertaken. Evaluation of the significance of the results was aided by comparison of laboratory data between patients with narco­lepsy and sleep apnea, a group which did not display the strong DR2/DQ l phe­notypic association characteristic of narcolepsy. The results appear to exclude systemic, antibody-mediated autoimmunity associated with narcolepsy.

MATERIALS AND METHODS

Subjects. The principal criterion for diagnosis of narcolepsy or obstructive sleep apnea syndrome was a history of excessive somnolence. In addition, nar­coleptics had objective evidence of REM sleep abnormalities including at least two abnormal REM onsets during multiple sleep latency testing (MSLT). Sleep apnea patients had obstructive respiratory events during sleep at a rate greater than 30 per hr and were free of major sleep abnormalities characteristic of nar­colepsy such as cataplexy. sleep paralysis. and hypnagogic hallucinations. Nar­coleptics had a REM latency of 8.7 ± 12.8 min after initiation of nocturnal sleep and achieved 3.9 ± 1.2 REM periods during daytime testing; in contrast patients with sleep apnea had a nocturnal REM latency of 112.9 ± 69.3 min and 0.7 ± 0.8 daytime REM periods (narcolepsy vs sleep apnea P < 0.01 for both REM param­eters). Daytime testing for the narcoleptics disclosed a mean sleep latency of 2. 74 (±2.03) min on the MSLT (instructions to fall asleep), which was 80% below control levels in our laboratory. For sleep apneics, mean sleep latency was 18.12 (±12.12) min on the four 40-min long trials of the maintenance of wakefulness testing (instructions to stay awake) which was 50% below normal controls. Thus both groups were excessively somnolent by objective criteria, but only the nar­coleptic patients displayed REM abnormalities. Consenting subjects were 11 pa­tients with narcolepsy (3 males, 8 females; 44 ± 12 years old) and 10 patients with sleep apnea (8 males, 2 females; 53 ± 10 years old).

HLA phenotype determination. Performance of HLA phenotyping was based on the microcytotoxicity method of Terasaki et al. (16). Blood was collected in heparin-containing tubes (Vacutainer) and lymphocytes were isolated by centrif­ugation into Ficoll-Hypaque. Washed lymphocytes were cryopreserved in liquid nitrogen at 2-5 x 106 cells/ml in complete medium containing 10% dimethyl sulf­oxide. Thawed cells were preincubated in complete medium for 1 day and dis­tributed into HLA-A, -B, -C typing trays (One Lambda Inc., Los Angeles, CA) and evaluated by Eosin-Y exclusion according to the manufacturer's instructions.

BRIEF COMMUNICATION 151

For determination of HLA-D antigens encoded by the DR and DQ loci, B lym­phocytes were enriched by lysing non-B cells with a cocktail of monoclonal an­tibodies plus complement (B Lympho-Kwik, One Lambda Inc.) and the B cell preparations, which were 60--90% pure based on sensitivity to anti-B cell mono­clonal antibody lysis, were assayed on HLA-DRw typing trays as described above.

Solid-phase immunoassays. A set of purified proteins, nucleoproteins, and nu­cleic acids which are common targets for autoantibodies in systemic rheumatic diseases ( 17) were tested for reactivity with the present sera using enzyme linked immunosorbent assays (ELISA). Sera were generally diluted 1 :200 and incubated in antigen-coated plates followed by detection of bound immunoglobulin with peroxidase-conjugated anti-human K plus >-.. as previously described (18). DNA (Calbiochem Behring, La Jolla, CA) was either heat denatured (dDNA) or treated with Sl nuclease to produce native DNA as previously described (19). Histones were from calf thymus and obtained from U.S. Biochemical Corp. (Cleveland, OH). Sm was purified from rabbit thymus acetone powder (Pel-Freez, Rogers, AR) and adsorbed to microtiter plates as previously described (20). SS-B was isolated from calf thymus (21). All solid-phase antigens were evaluated for reac­tivity with a panel of human sera that are prototypes of a wide variety of autoanti­body specificities. Positive and negative controls were run with each assay as noted in the results. Rheumatoid factor (RF) was detected by incubating sera with human lgG (Miles Laboratories, Elkhart, IN) coated plates at 1:500 dilution fol­lowed by detection of bound lgM antibody using peroxidase-conjugated goat anti­human lgM.

Immunofluorescence. Binding of antibodies to HEp-2 cells (Bion, Ridge, IL) and unfixed rat brain sections was evaluated at 1 :20 dilution as previously de­scribed (22). Fluorescein-conjugated goat anti-human polyvalent detecting re­agent (Tago, Inc., Burlingame, CA) was subsequently added and immunofluores­cence examined using a Leitz Ortholux II fluorescence microscope at 625 x mag­nification.

Antigenicity of primate brain stem. Extracts of a baboon (Papio anubios) brain stem were analyzed by Ouchterlony and Western blot for antigenicity. The brain stem including all tissue between the corpus callosum and the caudal aspect of the medulla was homogenized in PBS, and the soluble fraction was concentrated to 14 mg protein/ml and tested against undiluted sera by Ouchterlony double diffusion as previously described (22). The insoluble fraction was dissolved by heating in 1 % sodium dodecyl sulfate (SDS). The protein composition of the aqueous and SDS extracts was evaluated by SDS polyacrylamide gel electrophoresis (PAGE) (23) as shown in Fig. 1. Protein bands were electrophoretically transferred to nitrocellulose (24) and strips were used to probe sera diluted 1: 100 for bound antibody as detected by 1251-protein A from Staphylococcus aureus (ICN, Irvine, CA) (25).

Neurocytotoxic antibodies. Anti-neuronal antibodies were measured by a com­plement dependent 51Cr-release assay as described by Bluestein (12) with the inclusion of 0.02% NaN3 • Target cells were a neuroblastoma cell line (SK-N-SH) and a glioblastoma cell line (U-87-MG) obtained from American Type Culture

152

200-

116-92.5-

66.2-

BRIEF COMMUNICATION

A B

10 20 40 80 10 20 40 80

FIG . l . Protein composition of baboon brain stem extrac ts. Increasing amounts (µg) of an aqueous extract (A) and the residual fraction solubilized by SDS <Bl were separated by 15% SOS-PAGE and stained with Coomassie blue . Molecular weight markers are shown on the outside lanes .

Collection (Rockville, MD). Positive control samples were obtained from patients with SLE with central nervous system involvement (26) . The percentage of cy­totoxicity was calculated by subtracting the cpm of 51 Cr released by normal hu­man sera from that released by the sample and dividing by the total releasable cpm (determined by incubating cells with 0.4% NP40). Samples were assayed in trip­licate. The upper limit of normal sera induced release (two standard deviations above the mean) was 38% for the glial cells and 26% for the neuronal cells. Total releasable cpm were 5100 for the neuronal cells and 21 ,500 for the glial cells.

Other assays. Complement activation in vivo was evaluated by determination of the C4 split product, C4d, in heparinized plasma stored at - 70°C. Separation of C4d from C4 was performed by rocket immunoelectrophoresis and the ratio of the quantity of C4d to C4 was determined by planimetry according to the method of Nitsche et al. (27). Erythrocyte sedimentation rate (ESR) was determined in Westergren graduated tubes on blood anticoagulated with ethylenediamirielet­raacetic acid .

RESULTS

Table l shows the results from HLA phenotyping of the serologically defined antigens. Of the 11 narcoleptic patients examined, 9 were heterozygous for the HLA-DR2 phenotype. Using a x square test based on 25 .2% as the expected frequency, narcoleptics had a significantly higher prevalence of HLA-DR2 (X square = 18.8, P < .001). In contrast, only 3 out of 10 patients with sleep apnea

BRIEF COMMUNICATION 153

TABLE 1 HLA PHENOTYPES OF PATIENTS WITH NARCOLEPSY (N) AND SLEEP APNEA (A)

HLA phenotype

Patient Diagnosis A B c DR DQ

PC N All, A25 BIS, B52 DR2, DRw6 DQwl,-cs N A I, A31 B 7, B35 Cw3, Cw4 DR2, DR7 DQwl, DQw3 DG N A 3, All B 7, B35 Cw4,- DR!, DR2 DQwl,-KB N A l, A 3 B44, B57 -,- DR4, DR7 DQw2, DQw3 cs N A28, A30 849,- DR2, DR7 DQwl, DQw2 RC N A 2, A30 827, 838 Cwl, Cw7 DR2, DRw8 DQwl, PE N A 2,All B 7, B18 Cw7,- DR2, DR4 DQwl, DQw3 GJ N A 3,- B 7, B51 Cw7,- DR!, DR2 DQwl,-CM N All, A29 844, B51 Cw3,Cw5 DR2, DR4 DQwl, DQw3 RH N N.D.a N.D. N.D. DR!, DR3 DQwl, DQw2 BS N A 2, A28 855,- Cw3, Cw6 DR2,- DQwl,-RH A A 2, A29 844, 862 Cw!, Cw5 DR7, DRw6 DQw2, DQw3 FH A A28, 844, 858 DR3, DQw3,-RH A A 1, A30 B 8, B55 CwL Cw7 DR3, DR5 DQw2, DQw3 GR A A 2, A24 835, B51 Cw4,- DR!, DQw3,-PC A N.D. N.D. N.D. DR4, DR9 DQw3,-EG A A 2, A23 844, 851 Cw5,- DR4,- DQw3,-SC A A 3, A24 849. 862 Cw3,- DR3, DR4 DQw3,-GS A A30, A33 8 7. 851 Cw4,- DR2, DR5 DQwl,-HC A A24, 8 7, 862 Cw3, Cw7 DR2,- DQwl.-VT A A 3, A23 861,- Cw2,- DR2, DR5 DQwl, DQw3

"Not done.

displayed the HLA-DR2 phenotype, a normal population prevalence. All patients with HLA-DR2 also displayed the HLA-DQwl phenotype.

Two narcoleptics did not have the HLA-DR2 phenotype. One of these patients was a male DRl, DR3; the other a female DR4, DR7. On diagnostic polysomnog­raphy both had two or more REM sleep periods on daytime testing. However, the male was DQwl and had a history of head trauma prior to onset of symptoms. The female had a mildly abnormal neuropsychiatric evaluation and cognitive com­plaints.

Autoantibodies. Antibodies to predominant nuclear antigens commonly en­countered in systemic autoimmune diseases were examined by solid-phase im­munoassays using purified antigens. Results are expressed in Table 2 as raw OD units of antigen-bound immunoglobulin assessed by ELISA using an antihuman IgM reagent for detecting rheumatoid factor (RF) or anti-human K plus A. reagent for the other specificities. Compared to the assay background (>2 SD above the mean of normal serum binding) slightly elevated reactivity was detected in the narcoleptic group to Sm (one patient), dDNA (three patients), and IgG (two patients). Two sera showed weak antinuclear antibody (ANA) activity. However, as a group, narcoleptic sera were not elevated for any autoantibody specificity compared to the normal controls. Furthermore, scattered reactivity was also ob­served for the sleep apneic group (one dDNA, one SS-B, one histone, three RF,

TABLE 2 AUTOANTIBODILS IN NARCOLEPSY (N) AND SLEEP APNEA (A) PATIENTS -·---- ---------- --------~----- Vl

IF on HEp-2 % Toxicity to """ ELISA" determined autoantibody to cells" target cellsc

--·----·-----

Patient Diagnosis Sm dDNA nDNA SS-B Histone lgG Cytoplasmic Nuclear Neuronal Giial ---------·-·

Assay bkg" 0.20 1.20 0.10 0.20 0.40 1.30 - - 0 0

PC N 0.12 1.44 0.04 0.13 0.26 l.44 - - l 0 cs N 0.06 0.81 0.05 0.15 0.29 0.21 - - I 0 DG N 0.12 0.63 0.03 0.15 0.25 0.21 - - 2 0 KB N 0.15 0.39 0.04 0.11 0.20 0.18 9 0 cs N O.o7 1.50 0.03 0.13 0.26 0.45 0 0 RC N 0.05 0.54 0.03 0.11 0.24 0.65 - - 0 4 PE N 0.31 2.99 0.06 0.18 0.40 2.97 1+ 5 0

Ill c:

GJ N 0.12 0.65 0.04 0.12 0.26 1.26 2+ 12 0 tll

CM N 0.05 0.34 0.02 0.10 0.16 0.11 1-2 2+ 0 0 'l:l

RH N 0.10 0.77 0.03 0.12 0.24 0.13 0 0 ("')

- - 0 BS N O.Q7 0.50 0.04 0.12 0.20 0.62 - - 0 0 ~

RH A 0.10 0.55 0.04 0.17 2.19 0.14 9 0 ~ c:

FH A 0.18 0.53 0.08 0.14 0.17 0.40 - 2 0 z -RH A 0.05 0.89 0.02 0.92 0.21 0.20 -- I+ 4 0 ("')

> GR A 0.11 0.42 0.05 0.15 0.20 0.44 -- 12 0 -I PC A O.Q7 0.87 0.05 0.12 0.24 3.19 I+ 4 0 0 EG A 0.07 0.36 0.03 0.10 0.23 3.78 0 0 z - -SC A 0.11 1.33 0.05 0.14 0.27 1.32 - 1+ I 0 GS A 0.10 0.26 0.03 0.12 0.16 0.16 - 0 0 HC A 0.08 0.47 O.Q3 0.12 0.19 1.06 0 ()

VT A 0.05 0.81 0.03 0.10 0.16 0.97 2 15 Positive controls

I 3.61 4.26 1.72 1.92 0.95 5.58 4+ 4+ 69 30 2 2.64 4.01 0.91 2,()() 1.74 9.01 3 4.22 0.84 4 J.78

"A polyvalenl Jeteding reagent wa~ used lanti-K t anti·A) except for the rheumat01d factor in which anti-lgM was used. 1

' A tluoresceinateJ anti-heavy chain (G +- A t MJ reagent was used. · % cytotoxicity = (sample cpm - NHS cpm)/(total release - NHS cpm) x. 100% '1 Backcround determined bv 2 Sp > mean of normal sera binding.

BRIEF COMMUNICATION 155

and three weak ANA). No significant differences in the mean reactivity of the narcoleptic and sleep apneic groups were observed. Immunofluorescence studies using whole rat brain as the antigenic substrate were also negative.

Table 2 also shows the results of screening for neurocytotoxic antibodies using a complement-dependent 51Cr-release assay. The percentage cytotoxicity was calculated after subtracting 2 SD above the mean of normal serum binding. Six out of 11 sera from the narcoleptic group and 7/10 from the sleep apnea patients displayed slightly elevated capacity to promote cytolysis of the neuronal cells but only one serum from each group had greater than 10% neurocytotoxicity. Even less reactivity was directed against the glioblastoma cell line target. No significant differences between the narcoleptic and sleep apneic groups were observed.

Narcoleptic and sleep apnea patients were also evaluated for indices of general inflammation. No evidence for in vivo complement activation based on increase in C4d/C4 ratio was detected in any patient, and the ESR was not elevated in patients in either group.

Reactivity with extracts from primate brain stem. Baboon brain stem was frac­tionated into an aqueous extract and an aqueous insoluble, detergent solubilized fraction, the latter containing 20% of the total solubilized protein. The two frac­tions displayed markedly different protein compositions as shown in Fig. 1. The aqueous extract showed over 40 protein bands differing in molecular mass from 15 ,000 to greater than 200,000 Da, whereas the residual material requiring SDS for solubilization had a more limited array of proteins, including predominant bands at 21,000 and 51,000 Da. The antigenicity of the aqueous fraction was determined by Ouchterlony double diffusion but, because of interference by detergent, the SDS solubilized fraction was assayed by Western blot technique.

No precipitin lines were observed between the aqueous extract and sera from patients of either category of sleep disorder. Using the detergent solubilized frac­tion in a Western blot format weak reactivity was observed with all sera for the two predominant proteins at 21 and 51 kDA (data not shown). Three narcoleptic and two sleep apnea sera each reacted with one additional band whose size varied from 46 to 70 kDa, and antibody binding to a 95 kDa protein was observed with one other sleep apnea patient. However, other than the nonspecific reactivity displayed by all sera, no bands common to a significant number of narcoleptic patients were detected.

DISCUSSION

Our results confirm the strong association between narcolepsy and the MHC class II antigens HLA-DR2 and DQwl. Over 80% of our sample of narcoleptics were DR2 positive and greater than 90% were DQwl positive compared to a normal prevalence of these phenotypes of 25 and 55%, respectively. In contrast a normal HLA-DR/DQ phenotype distribution was displayed by the sleep apnea patients, consistent with the report of Langdon et al. (5), that only 1/6 sleep apnea patient was DR2 positive. These results, along with clinical features and poly­somnography data, further support the view that these two types of excessive somnolence have different etiological bases. However, the present study differs from the absolute association of most (1-4) but not all (5, 28, 29) previous reports

156 BRIEF COMMUNICATION

in that 2/11 patients in our narcoleptic group did not have the HLA-DR2 pheno­type. Both patients participated in other research projects and repeatedly dem­onstrated excessive somnolence and abnormal predilection for REM sleep periods on daytime testing. The non-DR2 narcoleptic female also suffered from cataplexy, and both patients complained of cognitive impairment. Taken together, the HLA phenotype data supports the notion of genetic predisposition to the clinical and polysomnographic signs of narcolepsy marked by the HLA-DR2 phenotype, but also suggests that narcolepsy may be acquired in patients who lack this pheno­type. It may be relevant in this context that restriction fragment length polymor­phism (RFLP) analysis using HLA-DQ 13 chain and HLA-DR 13 chain cDNA probes confirmed the strong association between DR2 and narcolepsy in man but failed to detect the existence of an analogous class II allele in canine narcolepsy (30). Overall these results may suggest that a single MHC gene product is not necessary for the development of narcolepsy in humans or dogs, but people with the DR2 phenotype are at greater risk.

MHC-disease associations, especially with class II antigens, usually involve autoimmune disorders (31), and the DR2 phenotype has been particularly asso­ciated with neurological disorders with autoimmune components (6-8, 10-12). The association of a class II antigen with a disease involving the central nervous system could be explained by unusual expression of DR antigens in a neurologic tissue as has been reported for murine astroglial cells in tissue culture (32), and murine brain cells exposed to interferon--y in vitro or in vivo (33). If such a phenomenon created a functional immune unit by association of Ia antigens with a cell component, an autoimmune response to that component may be triggered. With this concept, the DR2-Ia antigen would have enhanced capacity to present the putative self-component. The demonstration that all DR2 narcoleptics belong to the Dw2 subtype and showed the DQR-2,6 DNA fragment as determined by RFLP analysis using a DQ 13-chain cDNA probe lends credence to this model since this subtype/allogenotype was found in only 19% of normal individuals with the DR2/DQwl phenotype (34). Although narcoleptic DR2-Dw2 genes have a RFLP pattern indistinguishable from that of the normal Dw2 prototype (29), au­toimmunization may require both the predisposing la subtype and another, un­known event. Autoantibodies could be directly pathogenic to those parts of the brain in which the target antigen occurs, or interleukin-I (IL-1) released as part of the autoimmune response involving the target antigen could trigger one or more pathologic sleep processes (35). IL- I in the presence of other mediators can in­duce slow wave sleep when administered intraventricularly (36).

The autoimmunity scenario as an underlying cause or at least a sign of narco­lepsy prompted us to test blood samples from narcoleptic patients in a broad array of laboratory assays commonly used to confirm various autoimmune diseases. Patients with sleep apnea were used as controls. Results of assays for specific autoantibodies characteristic of diseases such as SLE (anti-DNA, anti-Sm, anti­histone), Sjogren's syndrome (anti-SS-B), rheumatoid arthritis (RF), and unde­fined antinuclear or anticytoplasmic antibodies were largely negative. The few sera with slightly elevated autoantibodies were not restricted to patients with narcolepsy so their significance is unclear. Since narcolepsy is clearly a neuro-

BRIEF COMMUNICATION 157

logic problem, and autoantibodies to brain cells have been reported in central nervous system diseases (10-12, 26), we tested sera for the capacity to react with the standard targets of neurocytotoxic antibodies. No significant reactivity was detected. It is also possible that narcolepsy sera may contain antibodies to brain tissue or to specific brain cell types. The narcoleptic lesion is thought to lie within the pontomedullary region of the brain stem (37). We failed to detect antibodies reacting with either soluble or insoluble fractions of primate brain stem which contain the pontine and medullary regions thought to be important for the sleep/ wake cycle. Signs of systemic inflammatory processes (elevated ESR and in vivo complement activation) were not observed, and narcolepsy patients do not dis­play the constitutional symptoms of fever, malaise, or arthralgias characteristic of patients with systemic autoimmune disease. These observations suggest that sys­temic autoimmunity is unlikely to underlie the etiopathogenesis of narcolepsy.

These negative results certainly do not exclude the possibility of an organ­specific autoimmune basis for narcolepsy. Autoantibodies may react with a minor component not present in sufficient quantities or not detectable by the methods employed. Alternatively, local production of autoantibodies in the cerebral spinal fluid as observed in multiple sclerosis (38) would have gone undetected. We also did not search for cell-mediated autoimmune reactions such as the cytotoxic T cells involved in the neurological disease experimental allergic encephalomyelitis (39) in animals. If narcolepsy has an autoimmune component, pathogenic anti­bodies or cytotoxic lymphocytes could be limited in distribution to areas of the brain controlling the sleep/wake cycle or to their neurological interconnections. Therefore, analysis of cerebral spinal fluid may be warranted. It may also be possible that a transient autoantibody-mediated reaction caused a nonrepairable lesion manifested as narcoleptic symptoms. Finally narcolepsy may be an autoim­mune disease without detectable autoantibodies as in systemic necrotizing vas­culitis (40).

If narcolepsy is not an organ-specific autoimmune disease, its association with HLA-DR2 may be due to the existence of a narcolepsy predisposing gene in close genetic linkage to and in linkage disequilibrium with the gene encoding the Dw2 13 chain. This view is supported by our observation and those of others (5, 28, 29) on the occurrence of narcolepsy in a few HLA-DR2 negative individuals.

ACKNOWLEDGMENTS We thank Sandra Chambers and Kendis Cox for the C4d/C4 and ESR determinations, respectively,

and Dr. Stephen R. Hanson and John Dietrich for making available the baboon brain. We thank William Ornelis for help with the HLA phenotype determinations, Peter Tan for the Western blot studies, and Dr. Steven J. Henriksen for assistance in the rat brain irnrnunofluorescence studies. Assistance in the preparation of the manuscript by the BCR Word Processing Center and its critical review by Dr. Eng M. Tan and Dr. John H. Vaughan are greatly appreciated.

REFERENCES I. Honda, Y .. Asaka, A., Tanaka, Y., and Juji, T., Sleep Res. 12, 254, 1983. 2. Juji, T., Satake, M., Honda, Y., and Doi, Y., Tissue Antigens 24, 316, 1984. 3. Matsuki, K., Juji, T., Tokunaga, K., Naohara, T., Satake, M., and Honda, Y., J. Clin. lnvest. 76,

2078, 1985. 4. Langdon, N., Welsh, K. I., van Darn, M., Vaughan, R. W., and Parkes, D., Lancet 2, 1178, 1984.

158 BRIEF COMMUNICATION

5. Langdon, N., Lock, C., Welsh, K., Vergani, D., Dorow, R., Wachtel, H., Palenschat, D., and Parkes, J. D., Sleep 9, 143, 1986.

6. Reinertsen, J. L., Klippel, J. H., Johnson, A. H .. Steinberg, A. D., Decker, J. L., and Mann, D. L., N. Engl. J. Med. 299, 515, 1978.

7. Reinerstein, J. L., Klippel, J. H., Johnson, A.H., Steinberg, A. D., Decker, J. L., and Mann, D. L., J. Rheumatol. 9, 253, 1982.

8. Kachru, R. B., Sequeira, W., Mittal, K. K., Siegel. M. E .. and Telischi, M .. J. Rheumatol. 11, 471, 1984.

9. Panayi, G. S., Wooley, P., and Batchelor, J. R., Brit. Med. J. 2, 1326, 1978. 10. Opelz, G., Terasaki, P., Myers, L., Ellison, G., Ebers, G., Zabriskie, J., Weiner, H., Kempe. H.,

and Sibley, W., Tissue Antigens 9, 54, 1977. 11. Jersild, C., Fog, T., Hansen, G. S., Thomsen, M., Svejgaard, A., and DuPont, B., Lancet 2, 1221,

1973. 12. Bluestein, H. G., Clin. Exp. Immunol. 35, 210, 1979. 13. Melnick, S. C., Brit. Med. J. 1, 368, 1963. 14. Wilkinson, P. C., Lancet 1, 1301, 1964. 15. Coleman, R. M., Roffward, H.P., Kennedy, S. J., Guilleminault, C., Cinque, J .. Cohn, M.A.,

Karacan, I.. Kupfer, D. J., Lemi, H., Miles, LE .. Orr, W. C .. Phillips, E. R., Roth. T .. Sassin, J. F .. Schmidt, H. S., Weitzman, E. D., and Dement. W. C., JAMA 247, 997, 1982.

16. Terasaki, P. I., Bemoco, D., Park, M. S., et al., Amer. J. Clin. Pathol. 69, 103, 1978. 17. Tan, E. M., Adv. Immunol. 33, 167, 1982. 18. Rubin, R. L., In "Manual of Clinical Laboratory Immunology" (N. R. Rose, H. Friedman, and

S. L. Fahey, Eds.), p. 744, ASM, Washington, 1983. 19. Rubin. R. L., Joslin, F. G., and Tan, E. M., J. lmmunol. Methods 63, 359, 1983. 20. Pollard, K. M .. and Tan, E. M., Clin. Exp. Immunol. 60, 586, 1985. 21. Chan, E. K. L., Francoeur, A.-M., and Tan, E. M., J. Immunol. 136, 3744. 1986. 22. Nakamura, R. M., Peebles, C. L., Rubin, R. L., Molden, D. P .. and Tan, E. M .. "Autoanti-

bodies to Nuclear Antigens (ANA)," 2nd ed., ASCP Press. Chicago, 1985. 23. Laemmli, U. K., Nature (London) 227, 680, 1970. 24. Towbin. H., Staehelin, T., and Gordon, J., Proc. Natl. Acad. Sci. USA 76, 435, 1979. 25. Johnson, D. A .. Gaustsch, J. W., Sportsman, J. R .. and Elder, J. H .. Gene. Anal. Tech. l, 3.

1984. 26. Bluestein, H. G., Williams, G. W., and Steinberg, A. D .. Amer. J. Med. 70, 240, 1981. 27. Nitsche, J. F., Tucker, E. S., Sugimoto, S .. Vaughan, J. H., and Curd. J. G., Amer. J. Clin.

Pathol. 76, 679. 1981. 28. Andreas-Zeitz, A .. Keller, E., Sholz, A. E. D., Roth, B .. Nevsimalova, S., Sonka. K., Docekal.

P., Ivaskova, E., Schulz, H., and Giesler, P., Lancet 2, 684, 1986. 29. Holloman, J. D .. Bell, J. I., Kilduff, T. S., Dement, W. C., Guilleminault, C., and McDevitt.

H. 0., J. Neurosci. Res. 18, 239, 1987. 30. Kilduff, T. S .. Dean, R., Motoyama, M., Burtch, M .. Radde, L., Steinman. L.. McDevitt, H. 0 ..

Lee, B. S. M., Grumet, C., Guilleminault, C., and Dement, W. C .. Sleep Res. 16, 369, 1987. 31. Svejgaard, A., Friis, J., Morling, N., and Ryder, L. P., J. Rheumatol. 8, 541. 1981. 32. Fontana, A., Kristensen, F., Dubs, R., Gemsa, D .. and Weber, E., J. Immuno/. 129, 2413, 1982. 33. Wong, G. H. W., Bartlett, P. F .. Clark-Lewis, I., Battye, F., and Schneider, J. W., Nature

(London) 310, 688, 1984. 34. Marcadet, A .. Gebuhrer, L., Betuel, H., Seignalet, J .. Freidel, A. C., Confavreux, C .. Billiard,

M., Dausset, J., and Cohen, D., Immunogenetics 22, 679, 1985. 35. Kamovsky, M. L., N. Engl. J. Med. 315, 1026, 1986. 36. Tobler, I., Borbely, A. A., Schwyzer, M., and Fontana. A., Eur. J. Pharmaco/. 104, 191, 1984. 37. Lai, Y. Y., and Siegel, J.M., Sleep Res. 16, 19, 1987. 38. Tourtellotte, W. W., and Ma, B. I., Neurology 28, 76, 1978. 39. Weigle, W. D., Adv. Immunol. 30, 159, 1980. 40. Fauci, A. S., Haynes, B. F., and Katz, P .. Ann. Intern. Med. 89, 660, 1978.

Received January 25, 1988; accepted with revision May 5, 1988