Antibody Detection in Gdain-Barr6 Syndrome Lars Svennerholm, MD, PhD, and Pam Fredman, PhD

Chromatograms with appropriate separated ganglioside fractions were overlaid with diluted patient sera, and antibody binding was detected with an alkaline phosphatase-labelled second antibody. Antiganglioside antibodies were present in the sera of 39 of 50 patients with Guillain-Barre syndrome (GBS), in 10 of 12 patients with Alzheimer’s disease, and in approximately 30% of the controls. The antibodies were directed against ganglioside 3’-LM1, shown to be the major ganglioside of peripheral nerve (femoral nerve and cauda equina) in most of the positive GBS sera but also in a high proportion in the controls. No correlation was found between the severity or the course of the disease and the antibody titer. Daily parenteral administration of purified bovine brain-derived GM1 ganglioside for three months to 12 patients with Alzheimer’s disease did not result in any antiganglioside GM1 antibodies. We have interpreted our findings in the following way. Human sera normally contain naturally occurring antibodies against gangliosides that in general do not cause any tissue damage. Thus, parenteral injection of gangliosides will not lead to any antibody formation. Svennerholm L, Fredman P. Antibody detection in Guillain-Barre syndrome. Ann Neurol 1990;27 (suppl):S36-S40

The etiology and pathogenesis of acute inflammatory demyelinating polyradiculoneuropathies such as Guillain-Barre syndrome (GBS) are unknown. Both humoral [l] and cellular [ 2 ) immune mechanisms are thought to be involved. In GBS, transient neurological symptoms are associated with an inflammatory demyelination of peripheral nerve in which myelin is the target of the immune attack [3, 41. Complement-fixing antibodies to peripheral nerve myelin have been demonstrated in the serum of patients with GBS and are highest when the neurological symptoms first occur 15, 61. The role of antibodies or other circulating factors has been further corroborated by the demonstration of shortening of the clinical course by plasmapheresis [7]. Lipid Composition of Peripheral Nerve No putative antigens have been elucidated, despite much effort over a long time to identify them. Recently serum antibodies to gangliosides were reported in 5 of 26 patients with GBS, but they showed no common epitope specificity IS]. Since the glycosphingolipids of peripheral nerve have aroused considerable interest, it has become important to know the ganglioside composition of peripheral nerve, particularly the ganglioside composition of its myelin. Perusal of the literature shows few studies of the lipid composition of peripheral nerve. MacMillan and Wherrett [9] described the occurrence of gangliosides in sciatic nerve

From the Department of Psychiatry and Neurochemistry, Gothenburg University, Hisings Backa, Sweden.

S36

and noted a ganglioside pattern similar to that for central nervous tissue. Svennerholm and colleagues [lo] found, however, that the major monosialoganglioside of femoral nerve contained N-acetylglucosamine instead of N-acetylgdactosamine, the only hexosamine detected in the major brain gangliosides. The structure of the N-acetylglucosamine-containing ganglioside was established the following year to be IV3 NeuAcnLcOse4Cer or 3’-LM1 ganglioside {ll). Fong and colleagues [12) showed later that the new ganglioside was prominent in myelin isolated from sciatic nerve. As no systematic study of human peripheral nerve lipids, particularly the ganglioside pattern, had ever been performed, we examined them in adult human femoral nerve and cauda equina (Svennerhoh et al, unpublished data). The ganglioside concentration was found to be 0.18 pmol and 0.29 pmol of lipid-bound sialic acid per gram of tissue in femoral nerve and cauda equina, respectively (Table 1). The concentrations of the characteristic myelin lipids, galactosylceramide (cerebroside) and galactosylceramide-3-0-sulfate (sulfatide), were two times higher in cauda equina than in femoral nerve, but their concentrations in cauda equina were only 40% of those in spinal cord or cerebral white matter. There were no changes in the lipid concentration with age. The ganglioside pattern (Table 2) was similar to the

Address correspondence to Dr Svennerholm, Department of Psychiatry and Neurochemistry, Gothenburg University, St. Jorgen Hospital, 422 03 Hisings Backa, Sweden.

Table 1. Lipid Composition of Adult Human Femoral Nerve and Cauda Equina Lipid Phospholipids Cholesterol Cerebroside Sulfatide Gangliosides

Femoral Nervea (n = 10)

OP”@€FO-CER Cauda Equina” (n = 9 )

*

~~~

86 68

12 49 35 -1- 9 9 + 4 3 + 0.8 0.18 2 0.03

+ +

17 15 18 2 3 7 2 1.4 0.29 0.05

+

n”@ow-CER

LA 1 3‘ -LM 1

*

~

~~~

“Measured in pmoVgm of fresh weight; values are mean t standard deviation.

Fag I . Gangliosides and allied neutral glycosphingolipids of the lactotetraose series found in peripheral nerve. (0= galactose; 0 = Glucose; = N-acetylglucosamine; = N-acetylneuraminic acid; CER = ceramide.)

0

Table 2. Ganglioside Pattern of Adult Human Femoral Nerve and Cauda Equina Ganglioside

Femoral Nervea (n = 10)

Cauda Equinaa (n = 10)

GM3 GM2 GM1 3’-LM1 3’-HexLM1 GD3 GD2 LD1 GDla GDlb GTlb GQlb

9 2 2 2 2 0.4 8 + 1 16 2 2 4 + l 6-1-2 2 2 1 3 2 1 16 -1- 2 14 2 2 16 2 1 4-11

6 2 1 1 2 0.3 9 2 2 17 2 2 4 ? 1 9 t 2 2 2 1 4 + 1 15 + 2 16 + 2 13 -1- 2 3 k 1

”Measured as percentage of sialic acid; values are mean t standard deviation.

one we [lo} described in 1972. Ganglioside 3’-LM1 was the major monosialoganglioside in femoral nerve and cauda equina, and constituted 15% to 20% of total ganglioside sialic acid. There was another ganglioside with the same terminal antigen determinant, 3’HexLM1, which reacted with the same affinity to our specific antiganglioside, 3 ’-LM1 antibody, SL-1. This ganglioside comprised 4% of total ganglioside sialic acid in femoral nerve and cauda equina. The ganglioside pattern did not change in individuals ranging from 22 to 91 years old. There was no sign of sialosylgalactosylceramide, the typical ganglioside of central nervous system myelin. Examination of isolated myelin confirmed our assumption that 3’-LM1 was the major ganglioside of peripheral nerve myelin, constituting approximately 30% of total ganglioside sialic acid. Figure 1 shows the schematic structure of 3’-LM1 and allied glycosphingolipids. Antiglycophingolipid Antibodies in GBS Serum The sensitivity of an antibody assay depends on three major factors: the amount of antigen, the concentra-

tion of the antibody solution, and the sensitivity of the detection system. We have previously used radiolabelled antihuman immunoglobulin antibodies, a method that can show great sensitivity if the specific activity is high. We have preferred, however, to use an enzyme-linked immunosorbent assay (ELISA) method, which is slightly less sensitive but obviates storage problems with radioactive waste. Our standard method for assaying antiganglioside antibodies is a thin-layer chromatographic (TLC) overlay technique. Because the details of the method are extremely important for evaluating results, the complete method and the amounts of antigens used are given in Tables 3 and 4, respectively. In all of our studies, we have used a serum dilution of 1:32, but other research groups have used a dilution of 1: 100 instead 18, 13, 141. The antigen amount we have used has in general been one-tenth or less of the amount used by the American researchers 18, 13, 141. We have also assayed circulating serum antibodies with an ELISA in microtiter wells. In a large number of sera, nonspecific binding was observed for serum dilutions above 1:400. Therefore, only results obtained by the TLC overlay technique have been used in the present study. Serum Samples from 50 Patients with GBS Sera from 42 patients with acute GBS and 8 patients in a chronic phase of the disease were examined. Because the results did not show any major difference between the two groups of patients, they were treated as a single cohort. The patients, 35 males and 15 females, ranging in age from 14 to 78 years, fulfilled the criteria of the National Institute of Neurological and Communicative Disorders and Stroke for GBS 1151. A detailed description is given in the report by Vedeler and colleagues 1161. Clinical symptoms of a preceding nonspecific respiratory or gastrointestinal infection with Svennerholm and Fredman: Antibody Detection in GBS S37

Table 3. Thin-Layer Chromatography Overlay Technique

1. Ganglioside antigens are applied to Polygram Sil G-precoated plastic sheets (Marchery-Nagel). 2. The plate is developed in chlorofordmethanol/2.5 mol/L NH3 (50:40: 10, by volume). 3. After drying, the plate is dipped twice in 0.1% polyisobutylmethacrylate in hexane and nonspecific binding blocked with 3% dried skim milk in Tris buffer. 4. The plate is overlaid with patient's serum, diluted 1 : 32 or 1: 100 in Tris buffer (pH8) containing 3% dried skim milk, followed by alkaline phosphate-conjugated goat antihuman IgG and IgM (H L) in a 1.5% bovine serum album-Tris buffer (pH 8) for 3 hours at room temperature. 5. Bound antibody is detected by incubating the plate with substrate solution of 0.01% 5-bromo-4-chloro-3-indolyl phosphate in 0.1 moVL glycine buffer (pH 10.4) at 37°C for 1 hour. 6. A serum is considered positive when specific staining of one or more ganglioside bands is obtained with a 1:32 or greater dilution of the patient's serum.

+

Table 4. Ganglioside Antigens Used with Thin-Layer Chromatography Overlay Technique

Ganglioside(s)"

Amount of Antigen

Total human brain gangliosides Total peripheral nerve gangliosides isolated from femoral nerve and cauda equina Monosialoganglioside fraction of peripheral nerve Neutral glycolipid fraction from PFAL brains { 171 Individual gangliosides GM2, G M I , GD3, GD2, G D l a , G D l b , G T l b , G Q l b , 3'-LM1, and 3'-isoLM1

0.3-3 nmol NeuAc 0.3-3 nmol NeuAc 0.1- 1 nmol NeuAc 10-1 5 nmol glycosphingolipid 10-500 pmol NeuAc

"All individual gangliosides are tested for carbohydrate homogeneity with fast atom bombardment mass spectrometry and specific monoclonal antibodies or cholera toxin B subunit or both before and after sialidase hydrolysis. Neu-Ac = N-acetyheuraminic acid; PFAL

=

polyunsaturated fatty acid lipidosis.

an interval of two weeks before paralysis were observed in approximately half the patients. Only patients with a cerebrospinal fluid (CSF) cell count less than 50 cells per mm3 were included (mean cell count, 2.3 ? 1.7 cells per mm3). CSF albumin and CSF IgG were increased two standard deviations above the control mean in 45 patients. Serum IgM and IgA were increased in 12 patients, while serum IgG was at normal levels in all. The sera were examined with the TLC overlay technique (see Table 3). A serum was considered positive when specific staining was obtained with one or more ganglioside bands with a 1: 32 dilution of the patient's serum. When the serum reacted only with gangliosides from peripheral nerve but not from brain, it was also tested with 100 pmol of 3'-LM1. All 50 sera reacted with the three major neutral glycolipid fractions of brain from a patient who died from polyunsaturated fatty acid lipidosis [17]: globotriaosylceramide, globotetraosylceramide, and fucose LA1 (Fig 2). Antibodies to gangliosides were detected in 39 of 50 patients. Most of the sera reacted with more than one ganglioside (Fig 3). Twenty-nine of the sera reacted with 3'-LMl, the major monosialoganglioside of peripheral nerve. With the lower antigen amount, which is only one one-hundredth of the

glob- tri glob-tetra

GM 1 GDla

Fuc-LA 1

-1

2

GDlb GTlb

- -- - -_

origin

3

4

5

6

Fig 2. Binding of serum l g (G + Mi from a patient with GBS to glycolipids as examined with thin-layer chromatography and enzyme-linked immunosorbent assay. Lane 1 , 100 pmol sulfatide; lane 2, 100 pmol galactosylceramide;lane 3, 10 nmol glycosphingolipids of brain from a patient who died from polyunsaturatedfatty acid lipidosis { I 7); lane 4,3 nmol human brain gangliosides; lane 5 , 1 nmol human brain gangliosides; lane 6, 0.2 nmol peripheral nerve gangliosides.

S38 Annals of Neurology Supplement to Volume 27, 1990

38-LMl

OM1

GDlblGTlb

PNS

CNS.PNS

3’-LM1

Glycolipid antigens

OM1 GDlblGTlb GA1

PNS

CNSePNS

Glycolipid antigens ~~

Fig 3. Antigangliosideantibodies in serum from 50 patients with GBS. The experimental conditions are given in Tables 3 and 4. (PNS = peripheral nervous system; CNS = central nervous system.)

Fig 4. Antigangliosideantibodies in serum from 12 patients with Alzheimer‘s disease. The experimental conditions are given in Tables 3 and 4. (PNS = peripheral nervous system; CNS = central nervous system.)

amount used by the American research groups [8, 131, only 8 sera had distinctly positive reactions, suggesting a relatively low titer or avidity. Except for the reaction with 3’-LM1, only 2 sera reacted with single gangliosides, 1 serum with GM1 and 1 serum with GD1b/GTlb. Control sera from 20 age-matched healthy blood donors were collected at the same time in Bergen. Six of the 20 controls had antibodies directed against 3’-LM1 ganglioside. The results from the examination of the sera from the patients and a larger control group with a 1: 100 serum dilution are in preparation. As a control group, sera were examined from 197 age-matched patients who were in our psychiatric hospital for endogenous depression, psychoneurosis, and cerebrovascular or senile dementia. Antibodies to gangliosides were detected in 47 of the 197 patients. Twenty-eight of the patients had antibodies against 3’LM1; this means that 15% of the controls with no signs of peripheral neuropathy also had antibodies against the ganglioside shown to be particularly involved in GBS. Therefore, we assumed that the occurrence of low-affinity antibodies against neutral glycosphingolipids and gangliosides is a common phenomenon and not related to any specific disease process.

from that reported by Ilyas and colleagues 181, who found a fall of the antibody titer concurrent with clinical improvement.

Antiganglioside Antibodies in Patients with GBS Undergoing Immunoglobulin Removal To test our hypothesis that antiganghoside antibodies did not have any causal connection with the pathogenesis for GBS, 19 patients who were treated with plasma exchange or protein A immunoadsorption (Freiburghaus et al, unpublished data) were studied. Antiganghoside antibodies were measured with the TLC overlay technique before the treatment was initiated and directly after completion of treatment. All patients improved, but the titer of antiganglioside antibodies did not change significantly. This finding differs

Antiganglioside Antibodies in AIzbeimerIr Disease Treated with G M l Twelve patients with presenile Alzheimer’s disease were given bovine brain-derived ganglioside GM 1 (100 mg per day) intramuscularly or subcutaneously for 90 days 1181. The patients were examined for antiganglioside antibodies before the treatment was initiated and after 30, 60,90, and 180 days. Ten of the 12 patients had antiganglioside antibodies before GM 1 treatment (Fig 4). These titers fluctuated during the period of ganglioside administration, but at completion the antibody titers were slightly lower. A particularly pertinent finding in this study is that no patient produced any antiganglioside GM1 antibody during the entire treatment period, although each patient received 9 gm of GM1 ganghoside. Discussion Our present studies of serum antibodies to gangliosides in patients with GBS and psychiatric diseases have shown that a large proportion of these patients have antibodies that are largely directed against ganglioside 3’-LM1. We are aware that Ilyas and associates [8] found antiganglioside antibodies in only 5 of 26 patients and none in the sera from healthy control subjects. To a certain extent, these differences can be explained by differences in the experimental procedure. We have used a threefold stronger serum dilution, but our antigen amount has been only one-tenth of that used by Ilyas and co-workers E81. The antigen amount is critical for the result. When we tested for antibodies against neutral glycolipids, we used a fraction with unusually high amounts of more complex glycosphingolipids {171 and found all 50 GBS sera to

Svennerholm and Fredman: Antibody Detection in GBS S39

be positive, while Ilyas and colleagues 181 did not find any positive sera for neutral glycolipids. We believe our present data show the following: 1. Human sera contain naturally occurring antibodies against a number of antigenic carbohydrate determinants on the glycoconjugates. 2. Their serum titer may increase in disorders displaying increased levels of immunoglobulins. 3. The high frequency of antibodies against 3’-LM1 might depend on an increased shedding of this ganglioside from various blood cells as a result of the inflammatory process. 4. Parented (intravenous, intramuscular, or subcutaneous) injection of pure gangliosides will not lead to any antibody formation. The costs of this study were defrayed by grants from the Swedish Medical Research Council (03X-627), the Bank of Sweden Tercentenary Foundation (86/326), and the Eiwind and Elsa K:son Sylvan’s Foundation.

References 1. Cook SD, Dowling PC. The role of autoantibody and immune complexes in the pathogenesis of Guillain-Barre syndrome. Ann Neurol 1981;9(suppl):S7O-S79 2. Iqbal A, Oger J J-F, Arnason BGW. Cell-mediated immunopathy in idiopathic polyneuritis. Ann Neurol 1981;9:65-69 3. Arnason BGW. Acute inflammatory demyelinating polyradiculoneuropathies. In: Dyck PJ, Thomas PK, Larnbert EH, Bunge R, eds. Peripheral neuropathy. Philadelphia: Saunders, 1984:2050-2100 4. Koski CL. Guillain-Barre syndrome. Neurol Clin 1984;2:355366 5. Latov N, Gross RB, Kastelman J, et al. Complement fixing antiperipheral nerve myelin antibodies in patients with inflamma-

tory polyneuritis and with polyneuropathy and paraproteinemia.

Neurology 1981;31 :1530- 1534

6. Koski CL, Gratz E, Sutherland J, et al. Clinical correlation with anti-peripheral myelin antibodies in Guillain-Barre syndrome. Ann Neurol 1986;19:573-577 7. Guillain-Barre Study Group. Plasmapheresis of acute GuillainBarre syndrome. Neurology 1985;35:1096-1104 8. Ilyas AA, Willison HJ, Quarles RH, et al. Serum antibodies to ganglioside in Guillain-Barre syndrome. Ann Neurol 1988; 23~440-447 9. MacMillan VH, Wherren JR. A modified procedure for the analysis of mixture of tissue gangliosides. J Neurochem 1969; 16:162 1-1624 10. Svennerholm L, Bruce A, Mksson J-E, et al. Sphingolipids of human skeletal muscle. Biochim Biophys Acta 1972;280:626636 11. Li Y-T, Mksson J-E, Vanier MT, Svennerholm L. Structure of the major glucosamine-containingganglioside of human tissues. J Biol Chem 1973;248:2634-2636 12. FongJW, Ledeen RW, Kundu SK, et al. Gangliosides of peripheral nerve myelin. J Neurochem 1976;26:157-162 13. Ilyas AA, Quarles RH, Dalakas MC, et al. Polyneuropathy with monoclonal gammopathy: glycolipids are frequently antigens for IgM paraproteins. Proc Natl Acad Sci USA 1985;82:66976700 14. Pestronk A, Adams RN, Clawson L, et al. Serum antibodies to GM1 ganghoside in amyotrophic lateral sclerosis. Neurology 1988;38:1457-1461 15. Asbury AK, Arnason BG, Karp HR, McFarlin DE. Criteria for diagnosis of Guillain-Barre syndrome. Ann Neurol 1978;3: 565-566 16. Vedeler CA, Matre R, Nyland H. Immunoglobulins in serum and cerebrospinal fluid from patients with acute Guillain-Barre syndrome. Acta Neurol Scand 1986;73:388-393 17. Svennerholm L, Fredman P, Jungbjer B, et al. Large alterations in gangliosides and neutral glycolipid patterns in brains from cases with infantile neuronal ceroid lipofuscinoses, polyunsaturated fatty acid lipidosis. J Neurochem 1987;49:1772-1783 18. Svennerholm L, Gotdries CG, Blennow K, et al. Parented administration of GM 1 ganglioside to presenile Altheimer patients. Acta Neurol Scand 1990;81:48-53

S40 Annals of Neurology Supplement to Volume 27, 1990

Antibody detection in Guillain-Barré syndrome.

Chromatograms with appropriate separated ganglioside fractions were overlaid with diluted patient sera, and antibody binding was detected with an alka...
530KB Sizes 0 Downloads 0 Views