Effects of Ganghosides on the Expression of Autoimmune Demyelination in the Peripheral Nervous System Diego Ponzin, MS,*t Anna Maria Menegus, PhD,? Gunther Kirschner, PbD,? M. Grazia Nunzi, PhD,? Mario G. Fiori, MD,t and Cedric S. Raine, PhD, DSc, FRCPath*
TOtest whether gangliosides (GA) q i g h t exert neuritogenic effects in vivo, experimental allergic neuritis (EAN) was studied clinically, neuropathologically, and immunologically in Lewis rats immunized with bovine peripheral nerve, and GA alone, P, myelin pfotein, P2myelin protein plus two different doses of GA, P, with galactocerebroside (GC), each emulsified in adjuvant. A14 except the GA-treated group developed signs of EAN between days 11 and 14 after the injection. Rats immspized with P, alone were the most severely affected. Rats given Pz plus GA and those given P? plus GC displayed a significantly lower clinical score. Histological analysis revealed a comparable degree of inflammation of the peripheral nervous system and demyelination in the spinal nerve roots of bovine peripheral nerveand P,-immunized rats. The P, plus Glq and P, plus GC groups revealed similar degrees of pathology in the spinal nerve roots but the latter group stood apart from the rest in that it showgd widespread peripheral nervous system changes extending distally into the sciatic nerve. Serological analysis demonstrated that P, and GC, but not GA, elicited antibody (I&) responses, but there was no correlation between antibody titer and clinical or histological involvement. The present data fail to support an enhancing role for gangliosides in the expression of EAN and, by extrapolation, in the Guillain-Bard syndrome, for which EAN serves as the laboratory model, and in which suggestions haye been mads that antibodies to GA may have pathogenetic significance. Ponzin D, Menegus AM, lrschner G, Nunzi MG, Fiori MG, Raine CS. Effects of gangliosides on the expression of autoimmune demyelination in the peripheral nervous system. A n n Neurol 1991;30:678-685 In the central nervous system (CNS), autoimmune demyelination has been shown to be augmented by the incorporation of certain myelin glycolipids (e.g., galactocerebroside [GCI) into the inochting emulsion 11-31. This effect has been attributed to myelin glycolipids acting as haptens in the generation of myelinolytic antibodies operating in concert with the T cellsensitiziqg antigen, myelin basic protein (MBP) 121. In the peripheral nervous system (PNS), a similar pathological scpario involving the PNS myelin proteiu P,, and glycolipids has been implicated in experimental allergic neuritis (EAN), but the mechanisms underlying the inflammatory events remain to be elucidated {4}. Whether other myelin components may be operative in autoimmune demyelination in the PNS has yet to be assessed. In this regard, glycoconjugates, including neutral and acidic glycolipids, occurring naturally in both neuronal membranes and myelin 15-71, have been suggested to be causally related to specific monoclonal antibodies in a number of human autoimmune neuropathies [8-141. Polyclonal antiganglioside antibodies have been described in a variety of neurological
Materials and Methods AYZiTZdh Thirty-six adult male Lewis rats (Charles River, Wilmington, MA), were used. Animals weighed between 250 and 300
From the *Departments of Neuropathology and Neuroscience, Albert Einstein College of Medicine, The Bronx, NY, and tFidia Research Laboratories, Abano Terme, Itdy.
Address correspondence to Dr Raine, Dept. of Pathology (Neuropathofogy), Alberr Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461.
conditions with suspected immunological involvement (e.g., amyotrophic lateral sclerosis, multiple sclerosis, Guillain-Barre syndrome) [l5-20). While the specificity and significance of these antibodies remain unclear, recent communications to this journal raised the possibility that under certain conditions, antibodies to gangliosides (GAS) may have a pathogenetic role in some peripheral neuropathies, in particular, the Guillain-Barre syndrome 12 1-23]. To investigate the latter possibility, the present study was undertaken to determine whether purified GAS, delivered either alone or in combination with the major neuritogen in EAN, P, protein, may affect the expression of immune-mediated demyelination in the PNS. The relative effects of GC, which is known to augment immune demyelination [l, 27, were studied for comparison.
Received Oct 29, 1990, and in revised form Feb 27 and May 2, 1991. Accepted for publication May 2, 1991.
678
Copyright 0 1991 by fhe American Neurological Association
3 or 4 per cage, and were fed standard chow and water ad libitum.
gm, were housed in groups of
lnocula and Animal Groups Freshly dissected intradural spinal roots were homogenized 1: 1 in saline solution. Aliquots of the suspension were emulsified with equal volumes of complete Freund‘s adjuvant (CFA) containing 10 mg Mycobactm’urn tuberculosis per milliliter. Five rats were subcutaneously injected with 0.2 mi of the emulsion into the dorsum of each hindfoot. In this way, animals received about 200 pg of P, protein, a known neuritogenic dose 143. BOVINE PERIPHERAL NERVE (BPN GROUP).
Bovine P, was purified from spinal root myelin as previously described [24}. Five rats were given one inoculation containing 0.2 mg P, in 0.1 ml saline solution emulsified in CFA. BOVINE p2 PROTEIN (P, GROUP).
+
p2 PLUS GALACTOCEREBROSIDE (p2 GC GROUP). G C (Supelco, Bellefonte, PA) was prepared as a micelle preparation with lysolecithin and bovine serum albumin 111. Four animals were given a dosc of 0.2 ml, made up of 0.1 ml containing 1 rng GC plus 0.2 mg P,, and 0.1 ml CFA. The ratio between GC and P, was similar to that used previously for experimental allergic encephalomyelitis (EAE) {l], except that the concentration of P, in PNS myelin was estimated to be about half that of MBP in CNS myelin.
+
p2 PLUS GANGLIOSIDES (P, GA GROUP). A highly purified bovine brain G A mixture comprising 21% GM,, 40% GD,,, 16% GD,,, 19% GTIb, 2% GD,, and 2% G Q l b (FIDIA S.p.A, Abano Terme, Italy) was dissolved in saline solution. P, was added so that 0.1 ml contained either 0.2 mg P, plus 0.02 mg GA (P2 GA(a) group), where the P2/GA ratio was similar to that found in intact PNS myelin [6], or 0.2 mg P, plus 1 mg GA (P, GA(b) group), with GA at a concentration known to elicit a delayed-type hypersensitivity response 1251. The two GA solutions were emulsified with equal volumes of CFA and delivered to two groups of 4 rats, as described above.
+
+
Two groups of 3 and 7 rats received 0.02 mg (GA(a)) and 1 mg (GA(b)) of the G A mixture, respectively. This was emulsified in CFA and delivered as described above.
GANGLIOSIDE (GA GROUP).
GALACTOCEREBROSIDE (GC GROUP). G C was not tested alone in CFA because this compound has been shown in several studies not to elicit neuropathological changes after single inoculation [I, 2, 4}.
Four rats received 0.2 ml CFA emulsitied 1:1 in saline solution. COMPLETE FREUNII’S ADJUVANT GROUP.
Clinical Evaluation Animals were observed daily and weighed from day 8 after injection onward. Clinical signs were evaluated using a 5-point scale: 0 = normal, 1 = limp tail, 2 = slight paraparesis, 3 = one leg paralyzed, 4 = paresis of both hind and/or front limbs, 5 = moribund or dead. Scoring was carried out
by two independent, blinded observers. Rats were anesthetized and perfused with fixative 48 hours after disease onset. Animals failing to develop signs were sampled 15 days after injection. Comparisons between groups were made by analysis of variance with orthogonal contrasts. Animals were perfused with cold 2.5% glutaraldehyde in phosphate buffer [l, 2). Spinal cord, proximal segments of L5, L6, and S1 ventral and dorsal roots and corresponding dorsal root ganglia, plus both sciatic nerves were dissected, thinly sliced, postfixed in chrome osmium, dehydrated, and embedded in epoxy resin (Epon 812). Slides containing a total of 8 to 10 semithin (1-pm) sections of the entire root or nerve were stained with toluidine blue and evaluated in a blinded fashion by at least three independent observers. The index for inflammation was: 0 = none; 1 = a few scattered inflammatory cells; 2 = numerous scattered cells, an occasional perivascular cuff; 3 = many perivascular cuffs; 4 = confluent perivascular cuffs; 5 = extensive perivascular cuffing with involvement of the entire nerve. The index for demyelination was: 0 = none, 1 = a few scattered naked axons, 2 = small groups of naked axons, 3 = larger groups of demyelinated axons, 4 = confluent foci of demyelination, 5 = widespread demyelination. From each animal, mean values were determined for both ventral and dorsal roots and sciatic nerves, and were pooled for each experimental group. Comparisons between groups were made by analysis of variance with orthogonal contrast. MORPHOLOGY.
Prior to sampling, rats were bled from the heart and serum samples stored at - 20°C. For the detection of antibodies against P,, GA, and GC, an enzymelinked immunosorbent assay (ELISA) was employed 126, 271. For anti-GA antibody assays, 750 ng of GM,, GD,,, GD,,, or GT,, was diluted in ethanol and water in a 1 :1 ratio and added to each well of a flat-bottom microtiter plate (Flow Labs, McLean, VA). Incubation was carried out at room temperature for 90 minutes, and nonspecific binding was blocked for 90 minutes with 2% bovine serum albumin in phosphate-buffered saline (PBS) (Flow Labs). After washing with buffer, 50 pl of serum diluted 1: 100 to 1:6,400 in 2% bovine serum albumin in PBS was added and the plate incubated overnight at 4°C. After washing, 50 pl of horseradish peroxidase-conjugated goat anti-rat IgG (Sigma, St Louis, MO) or IgM (Cappel, West Chester, PA), diluted 1: 1,000, was added for 2 hours at room temperature and then washed. Antibodies were detected by adding 50 ~1 of orthophenylanine-diamine reagent (Sorin, Vercelli, Italy). Absorbance was read at 492 nm on a Titertek Multiskan ELISA reader (Flow Labs). For anti-P, and anti-GC antibody assays, 100 pl of a solution of P, in PBS (1 pg/ml) or GC in ethanol (10 pg/ml) was added to each well. Incubation was performed at 4°C overnight (P,) or at room temperature for 120 minutes (GC). Nonspecific binding was blocked with 1% polyvinylpyrrolidone for 20 minutes at room temperature and then 100 pl of serum diluted 1: 100 to 1:25,600 in PBS plus 0.05% Tween 20 (Sigma) was added and the plate incubated overnight at 4°C. Detection of antibodies was then performed as described above. IMMUNOLOGICAL ASSAYS.
Ponzin et al: Gangliosides and EAN 679
Table 1. Prodtlction of Experimental Allergic Neuritis in Lewis Rats with DIZfferentAntigensa
Dose (mgianimal)
Day of Onsetb (after Inoculation)
BPN
50
Complete Freund's adjuvant GA(a) GA(b)
-
11.0 0 0 0 11.2 k 13.2 k 12.5 ? 11.7 k
Antigen
P2 P2 P2 P2
0.02 1.00 0.20 0.20 0.20 0.20
+ GC
+ GA(a) + GA(b)
Clinical Disease
* 0.0
+
+ +
1.00 0.02 1.00
Incidence
Mean Scoreb
515 014 013
2.3 k 0.5 0 0 0 3.6 0.2 2.1 k 0.3 1.5 2 0.2 1.7 k 0.1
017 515
0.5 0.5
414 414 414
0.6 0.5
*
"P, and glycolipids were mixed together and emulsified with complete Freund's adjuvant containing 10 mgiml of Mycobarterium tabercalosis. Two-tenths milliliters of this mixture was injected subcutaneously in the dorsum of the posterior feet. bValues are mean 5 standard crror.
BPN
=
bovine peripheral nerve; P2
=
bovine P, protein; G C
=
galactocerebroside; GA
The results are presented as mean optical densities (OD) 1 standard deviation, and were obtained at 1: 100 (GA, P,, GC) and 1:3,200 (P2).Control values were determined from the sera of 15 normal Lewis rats from the same source. A positive result was 1 OD greater than 2 standard deviations above the mean OD for control animals analyzed in the same experiment. Data on all P,-treated groups were compared by Scheffe's multiple comparisons test. 2
Results Clinical Finding BPN-inoculated rats presented with neurological signs on day 11 after injection (Table 1). For neuropathology, 4 rats were sampled 48 hours after disease onset when they showed mild to severe paraparesis (clinical score, 1 to 4),and 1 was taken 5 days after onset, when the clinical course had stabilized at grade 3. The P, group developed significantly more severe signs than the BPN and the P, GA groups (p < 0.01). In these rats, signs began between days 11 and 13 after injection and, after 48 hours, reached a clinical level of 3 to 4. GC + P, animals became sick between days 12 and 14 after injection. At the time of death (48 hours later), neurological signs were less severe than those of the P,-injected group (grades 1 to 3, p < 0.01). GA-sensitized rats presented with different clinical pictures. When GA was given alone in CFA (both doses), no neurological signs developed. Between days l l and 14, all rats in the P, GA groups developed clinical signs similar to but less severe than those from P,-inoculated rats (clinical score 1 to 2 ) . In all groups, most rats developed an adjuvant arthritis with a time course overlapping that of the neurological syndrome. This was taken into account by not including the hindfeet stiffness and increased sensitivity in the evaluation of the clinical score. In addition, all animals displayed an abrupt overnight decrease in body weight that preceded or accompanied neurological
+
+
=
gangliosides.
signs. This reduction was even apparent in nonsymptomatic rats immunized with CFA alone or GA + CFA.
Neuropatbology PNS tissue from all animal groups was examined. Mean histological scores for ventral roots, dorsal roots, and sciatic nerves from all animals sampled (48 hours after onset) are detailed in Figure 1. Readings from all levels of either ventral or dorsal roots were pooled since there was no clear difference. Dorsal roots were usually more involved than were ventral roots. With regard to inflammation (see Fig 1, upper panel), there was no difference between groups except for the P, + GC group, in which significantly fewer inflammatory cells were found in ventral roots (p < O.O1), in comparison to rats treated with P, alone. The P, GC group was unusual in that animals showed inflammatory cells in the sciatic nerve as an early event (' < 0.001). One animal in the P, + GA group showed some inflammatory cells in the sciatic nerve. The one animal from the BPN group that was killed at a later time point, when the clinical score had stabilized at grade 3, demonstrated ventral and dorsal roots to be equally affected and some inflammatory cells in the sciatic nerve. With regard to demyelination (see Fig 1, lower panel), there was no difference between the groups, except for the P, + GC group in which demyelination was also found in the sciatic nerve (p < 0.001). In the single animal sampled later in the BPN group, dernyelination was also evident in the sciatic nerve. Comparison of the various groups confirmed the presence of relatively uniformly myelinated nerve fibers in spinal nerve roots and sciatic nerve of normal animals, the endoneurium containing no infiltrating cells. In GA-inoculated animals, the PNS (Figs 2A and 2B) was normal. CFA controls also showed no PNS changes. In BPN-sensitized rats, lesions typical of
680 Annals of Neurology Vol 30 No 5 November 1991
+
z
-
0 k
U
sciatic nerve was involved (i.e., within 48 hours of onset) and perivascular cuffing was more marked. Axonal loss occurred in all groups but was more apparent in animals treated with P, plus GC. Animals inoculated with P, plus GA ((a) and (b)) presented with milder lesions than did the BPN and P, groups (see Fig 2F). Spinal cord changes were not seen, but on occasion, infiltration of a few mononuclear cells into the leptomeningeal space was seen in neurologically afflicted animals. However, these cells did not invade CNS parenchyma.
“1 3
I 5
A U.
2
5 1
0
BPN
P2
PZ+GC
PZ+GA(a)
-F
BPN
P2
P2+GC
P2+GA(o)
P2+GA(b)
T
P2+GA(b)
Fig 1. Comparative histopathology of experimental allergic neuritis groups. Average infEammation (upper panel) and demyelination (lower panel) of ventral roots (white bars), dorsal roots (shaded bars), and sciatic nerves (black bars) 2 standard errors. See Materials and Methods fir a description of the groups and of the scoring system. Note the uniformity in disease expression in all groups except the one that was inoculated with P2 and galactocerebroside (Pz GC) where there is a distal shift in PNS involvement. BPN = bovine peripheral nerve; GA = galactoside.
-+
acute EAN were encountered and were invariably centered on blood vessels. The endoneurial space was increased and contained infiltrating cells, some perivascular cuffing was apparent, and nerve fibers at various stages of demyelination were evident (see Fig 2C). Demyelination invariably occurred in association with infiltrating macrophages, some of which could be seen within the Schwann cell tube. Infiltration by polymorphonuclear leukocytes occurred and was common in all groups. P,-induced EAN was essentially identical to BPN-induced disease (see Fig 2D). In animals sensitized with P, plus GC, spinal nerve root lesions were qualitatively similar to those of the BPN and P, groups but sciatic nerves were extensively infiltrated (see Fig 2E). Therefore, EAN was more widespread when G C was combined with P,, and there was a shift in the pathology from radicular to more distal regions. The P, + GC group represented the only group in which
Serology No 1gM response was found against any of the antigens. Sera from the groups inoculated with CFA or GA contained no IgM or IgG response against P,, GA, or GC (Table 2). Antibody reactivity is shown for GM, only because it was representative of results from the other three GAS tested. P,- and BPN-sensitized rats developed antibodies against P,. The lowest P, response was seen in sera from BPN rats, whereas rats treated with P, or P2 plus GA(a) developed higher responses. For this reason, readings from the higher dilution tested are also shown, where the OD response was linear. Somewhat lower values, not statistically significant, were found in rats injected with P, plus G C or P, plus GA(b), even though some variation among individuals was found. Antibodies against the four main GAS in the mixture were not demonstrable in most animals. A borderline positive response was seen in 1 rat from the P, + GA(a) group and 2 from the P, + GA(b) group. Only the P, + GC group displayed a positive antibody response against GC at the time point tested (48 hours after onset). A borderline positive response was seen in the sera of 2 rats from the P, GA(b) group. Taken together, the ELISA findings showed no significant correlation between humoral response and disease findings.
+
Discussion On the basis of suggestions that GAS might serve as pathogens in PNS disease 121-231, the present study investigated whether GAS possess the ability to enhance immune-mediated tissue damage in a model of EAN in the Lewis rat. Such an augmentation effect had previously been shown with G C in EAE El-31. These latter studies on EAE concluded that tissue damage was the result of an initial T-cell response to MBP and a B-cell response to G C 111. In support of this, demyelination was observed in the rabbit retina after the introduction into the vitreous of anti-GC antiserum plus cytokines [2]. In contrast to the situation with MBP in EAE, sensitization to P, is sufficient to elicit full-blown EAN with infhnmation and demyelination [28, 291. However, the incorporation of G C into the protocol
Ponzin et al: Gangliosides and EAN 681
A
B
C
D
E
F
682
Annals of Neurology Vol 30 No 5
November 1991
Table 2.Antibody Reactivity by Enzyme-Linked lmmunosorbent Assay" IgG Anti-P,
Group
(1: 100)
Normal control
227 177 109 i 42 119 67 291 2 130 1,031 i 140 2,645 +- 63 2,102 i 142 2,113 80 2,273 232
*
Complete Freund's adjuvant GAb) GA(b) BPN
p2 P2 + GC P2 + GA(a) P, + GA(b)
(1 :3,200)
IgG Anti-GM, ( 1 : 100)
IgG Anti-GC (1:lOO)
-
54 2 17 41 t 18 43 2 16 53 2 32 65 2 12 85 20 50 & 15 109 2 54 125 ? 66
112 i 67 90 26 102 2 20 114 -t 37 216 & 71 146 & 65 420 2 173 221 58 220 105
-
*
-
52 ? 19 1,049 ? 302 452 2 176 1,570 ? 309 435 ? 346
* *
*
*
*
*
aVdues are mean optical densities ( x lo3)2 standard deviation. Numbers in parentheses indicate the dilution of the sera Each s e w sample was assayed in triplicate at the different dilutions. Mean values and standard deviations represent interanimd values. The intra-assay variatiqn was approximately 10%.
GA = gangliosides; BPN
=
bovine peripheral nerve; GC
=
galactocerebroside.
has been shown to augment tissue damage in the spinal nerve roots {4] and, in the present case, caused an early, distal involvement of the sciatic nerve. Sciatic nerve is usually not involved until later time points and the occurrence of pathology at this level in the present animals may be related to the low levels of circulating antibody to GC. Intraneural injection of anti-GC antiserum alone has been shown to be sufficient to produce demyelinated lesions of peripheral nerve {30].
2. (A) Ganglioside (GA(b))-sensitizedrat, epoxy 4 Fig resin section. Transverse section o f an LS ventral root from a rat 1-pm
displaying no clinical disease reveals normal nerve fibers and no injammatory changes. (B) Sciatic nerve of the same animal as A. Note the normal appearance of the nerve. (C) Bovine peripheral nerve (BPN)-inoculated rat, 1-pm epoxy wsin section. Detail from a transverse section o f an L6 ventral root from a rat displaying clinical signs. The entire section was scoved 3 for inflammation and 4 f i r LmyeLination. Note the injammuto y cells surrounding a x o m undergoing akmyelination. ( D ) Pzsensitized rat, 1-pm epoxy resin section. Detail from a transverse section of an S i dorsal root from an experimental allergic neuritis (EAN)-afjacted rat. This corresponded to grades 4 (inflammation) and 3 (demyelination).Note the presence of some polymorpbonuclear leukocytes and an axon undergoing demyelination (arrow}. (E) Pz and galactocerebroside (GC)-sensitized rat, 1-pm epoxy resin section. Transverse section of sciatic nerve from a rat with clinical EAN. Note the conjuent perivascular cufs and scattered demyelinated fibers. This corresponded t o graah 4 (inflammation) and 3 (&myelination). Note the perivascular cuf and some demyelinated axons (arrows). (F)Pz and GA(b)-sensitized rat, 1-pm epoxy resin section. Transverse section of an S1 ventral root of a rat displaying EAN. This root scored 3 for injammation and 4 for demyelination. Note the demyelinated axons (arrow) and an injammatory cell adhering t o the endothelium of the vessel (center). (A, C-F, toluidine blue; A, B, x 375; C-F, x 755.)
The present experimental protocol compared Lewis rats sensitized against a known neuritogenic dose of P, protein with animals given P, plus different doses of GAS and, for comparative purposes, GC.A mixture of GAS was used since it has been observed that the immune reaction to these compounds yields antibodies reacting with the Gal(P 1-3)GalNAc epitope of GM,. This is shared by other glycolipids and glycoproteins in the myelin and ponmyelin fractions {22]. Furthermore, when mixtures of total brain GAS are administered, the predominant antibody response is against GM, and asialo-GM1 1311. In agreement with previous observations, the combination of P, with G C in the inoculum increased the PNS lesion area but not its severity. While inflammation and demyelinarion in P, animals were limited to spinal roots supplying the sciatic nerve, the inclusion of GC into the protocol led to lesions in the distal nerve as well. In contrast to GC, GAS did not enhance lesion areas induced by P2.The data showed that when used alone, GAS produced no pathological changes, and when GASwere coupled with P, protein, signs that occurred were less severe than those seen in animals receiving P, alone (see Table 1). In the present study, the clinical score was lower when either GAS or G C were combined with P,, a finding discordant with the observation that there was essentially no difference in the degree of neuropathological involvement between the groups. The reduced clinical score for the P, + G C group, in spite of the increased pathology, might be related to the lesions being less radicular and more distal, thus rendering the effect less detectable clinically. A similar dissociation between clinical and structural changes has also been reported in previous studies of EAE and EAN C32, 331. The lack of change in the humoral response to P, protein after the addition of G C or GAS to the emulPonzin et al: Gangliosides and EAN
683
sion might be explained by the neuritogenic activity of the CN-1 peptide of P, correlating better with a T-cell, rather than a B-cell, response E343. Antibodies against GAS were not detected in most animals. This supports further the low immunogenicity of these glycolipids
c35, 361. The failure of exogenous GAS to enhance the response to P, might relate to a number of phenomena, particularly when one compares the effect with that of GC. The neuritogenicity of P, appears to be conformation dependent [ 3 7 ) and interactions between P, and GAS might lead to conformational changes that are ineffective in augmenting neuritogenicity, in contrast to the case with GC. Thus, it appears that GAS exert no augmenting role in autoimmune demyelination and it is possible that the presence of GASin the sensitizing mixture may have interfered with interactions between neuritogenic determinants and the immune system. Interestingly, the administration of exogenous GAS has been claimed to offer partial protection against EAN induced in myelin-sensitized rats [38). Other studies on EAE have documented a protective role for GM, and GM4 [39J and a delayed hypersensitivity response to GAS that was not encephalitogenic c251. Taken in concert, our observations have shown GAS to be nonantigenic in the PNS and suggest that when combined with P,, a strong neuritogen, exogenous GAS might lead to a lessened pathological effect. The possible immunological mechanisms underlying these discrepanices are currently under investigation. This study was supported in part by grants (to C.S.R.) from the National Instirutes of Health (NS 08952 and NS 11920) and National Multiple Sclerosis Society (RG 1001-G-7). The authors thank Drs Krzysztof W. Selma], Alessandro Bruni, Lanfranc0 Callegaro, Alberta Leon, and Gin0 Toffano for helpful discussion; Everett Swanson, Howard Finch, Miriam Pakingan, Marta Bassan, Paola Facco, and Anna Rosa Morandin for expert technical assistance; Dr George Hashim for P, protein; and Michele Brigs for careful preparation of the manuscript.
References 1. Rame CS, Traugotr U, Farooq M, et al. Augmentation of immune-mediated demyelination by lipid haptens. Lab Invest 1981;45:174-182 2. Brosnan CF, Traugoa U, Raine CS. Analysis of humoral and cellular events and the role of lipid haptens during CNS demyelination. Acta Neuropathol (Berl) 1983;9(suppl):S59-S70 3. Hosein 22, Gilbert JJ, Strejan GH. The role of myelin lipids in experimental allergic encephalomyelitis. J Neuroimmunol 1986;10:219-233 4. Hughes RAC,Powell HC. Experimental allergic neuritis: demyelination induced by P2 alone and non-specific enhancement by cerebroside. J Neuropathol Exp Neurol 1984;43:154-161 5 . Yu RK, Saito M. Structure and localization of gangliosides. In: Margolis RU, Margolis F X , eds. Neurobiology of glycoconjugates. New York: Plenum Press, 1989:l-42 6. Fong JW, Ledeen RW, Kundu SK, Brostoff SW. Gangliosides of peripheral newe myelin. J Neurochem 1976;26:157-162
684 Annals of Neurology Vol 30 No 5
7. Ledeen RW, Cochran FB, Yu RK, et al. Gangliosidcs of the CNS myelin membrane. Adv Exp Med Biol1980;125:167-176 8. Quarles RR, Ilyas AA, Willison HJ. Antibodies to glycolipids in demyelinating diseases of the human peripheral nervous system. Chem Phys Lipids 1986;42:235-248 9. Latov N, Hays AP, Sherman WH. Peripheral neuropathy and anti-MAG antibodies. CRC Crit Rev Neurobiol 1988;3: 301-332 10. Ilyas AA, Wilhon HJ, Dalakas MO, et d. Identificacion and characterization of gangliosides reacting with 1gM paraproceins in three patients with neuropathy associated with biclonal gammopathy. J Neurochem 1988;5:85 1-858 11. Nardelli E, Steck AJ, Barkas T, et al. Motor neuron syndrome and monoclonal IgM with antibody activity against gangliosides GM1 and G D l b . Ann Neurol 1988;23:524-528 12. Larov N , Hays AP, Donofrio PD, et al. Monoclonal IgM with unique specificity to gangliosides GM1 and G D l b and to lactoN-tetraose associated with human motor neuron disease. Neurology 1988;38:763-768 3. Ito H, Latov N. Monoclonal IgM in two patients with motor neuron disease bind to the carbohydrate antigens Gal(B13)GalNAc and Gal(B1-3)GlcNAc. J Neuroimmunol 1988;19: 245-253 4. Marcus DM, Percy L, Gilbert S, et al. Human IgM monoclonal proteins that bind 3-fucosyllactosamine, asialo GM1 and GM1. J Immunol 1989;143:2929-2932 15. Endo T, Scott DD, Stewart SS, et al. Antibodies to glycosphingolipids in patients with mulciple sclerosis and SLE. J Immunol 1984;132:1793-1797 16. Pestronk A, Chaundhry V, Feldman EL, ec al. Lower motor neuron syndromes defined by parterns of weakness, nerve conduction abnormalities, and high titers of antiglycolipid antibodies. Ann Neurol 1990;27:316-326 17. Pestronk A, Adams RN, Clawson L,et al. Serum antibodies to GM1 ganglioside in amyotrophic lateral sclerosis. Neurology 1988;38:1457-1461 18. Ilyas AA, Willison HJ, Quarles RH, et al. Serum antibodies to gangliosides in Guillain-Barri. syndrome. Ann Neurol 1988; 23:440-447 19. Pestronk A, Cornbluth DR, Ilyas AA, et al. A treatable multifocal motor neuropathy with antibodies to GM1 ganglioside. Ann Neurol 1988;24:73-78 20. Shy ME, Evans VA, Lublin FD, et al. Antibodies to GMI and G D l b in patients with motor neuron disease without plasma cell dyscrasia. Ann Neurol 1989;25:511-513 21. Svennerholm L, Fredman P. Antibody detection in GuillainBarrC syndrome. Ann Neurol 1990;27(suppl):S36-S40 22. Latov N. Neuropathy and anti-GM1 antibodies. Ann Neurol 1990;27(~~ppl):S4 1443 23. Quarles RH, Ilyas AA, Willison HJ. Antibodies to gangliosides and myelin proteins in Guillain-Barre syndrome. Ann Neurol 1990;27(SUPPI):S~ 8-S 52 24. Brostoff SW, Sachs H, Dal Canto M, et al. The P2 protein of bovine root myelin: isolation and some chemical and immunological properties. J Neurochem 1974;23:1037-1043 25. Offner H , Standage BA, Burger DR, Vandenbark AA. Delayed-type hypersensitivity to gangliosides in the Lewis rat. J Neuroimmunol 1985;9:147-157 26. Quarles RH. Human monoclonal antibodies associated with neuropathy. Methods Enzymol 1989;179:291-299 27. Nakayasa H , Ota K, Tanaka H , er al. Suppression of actively induced and passively transferred experimental allergic neuritis hy cyclosporin A. J Neuroimmunol 1990;26:219-227 28. Kadlubowski M, Hughes RAC. Identification of the neuritogen of experimental allergic neuritis. Nature (Lond) 1979;277: 140-141 29. Kadlubowski M, Hughes RAC, Gregson IVA. Experimental al-
November 1991
30.
31.
32.
33.
34.
lergic neuritis in the Lewis rat: characterization of the activity of peripheral myelin and its major basic protein, P,. Brain Res 1980;184~439-454 Saida K, Sumner AJ, SaidaT, et al. Antiserum-mediated demyelinarion: relationship between remyelination and functional recovery. Ann Neurol 1980;8:12-24 Rapport MM, Graf L, Brunner W, Yu RK. Antibodies to total brain gangliosides: titer and specificity of antisera. In: Svennerholm L, Mandel P, Dreyfus H, Urban P-F, eds. Advances of experimental medicine and biology, vol 125. New York, Plenum Press, 1980:327-334 Raine CS, Snyder DH, Stone SH, Bornstein MB. Suppression of acute and chronic experimental allergic encephalomyelitis in strain 13 guinea pigs. J Neurol Sci 1977;31:355367 Rostami A, Brown MJ, ksak RP, et al. The role of myelin P, protein in the production of experimental allergic neuritis. Ann Neurol 1984;16:680-685 Milek DJ, Cunningham JM, Powers JM, Brostoff SW. Experi-
35.
36.
37.
38.
39.
mental allergic neuritis: humoral and cellular response to the cyanogen bromide peptides of the P, protein. J Neuroimmunol 1983;4:105-117 Marcus DM. A review of the immunogenic and immunomodulatory properties of glycosphingolipids. Mol Immunol 1984; 21:1083-1091 Lvingston PO, Ritter G, Calves MJ. Antibody response after immunization with the gangliosides GM,, GM?, GM,, GD, and GD3 in the mouse. Cancer Immunol Immunother 1989;29: 179-184 Brostoff SW. Immunological responses to myelin and myelin components. In: Morel1 P, ed. Myelin. New York, Plenum Press, 1984:405-439 Ledeen RW, Oderfeld-Now& B, Brosnan CP, Cervone A. Gangliosides offer partial protection in experimental allergic neuritis. Ann Neurol 14c)0;27(suppl):S6c)-S74 Mullin RB, Patrick DH, Poore CMB, et al. Prevention of experimental allergic encephalomyelitis by ganghoside GMS. A follow-up study. J Neurol Sci 1986;73:55-60
Ponzin et al: Gangliosides and EAN
685
TOtest whether gangliosides (GA) q i g h t exert neuritogenic effects in vivo, experimental allergic neuritis (EAN) was studied clinically, neuropathologically, and immunologically in Lewis rats immunized with bovine peripheral nerve, and GA alone, P, myelin pfotein, P2myelin protein plus two different doses of GA, P, with galactocerebroside (GC), each emulsified in adjuvant. A14 except the GA-treated group developed signs of EAN between days 11 and 14 after the injection. Rats immspized with P, alone were the most severely affected. Rats given Pz plus GA and those given P? plus GC displayed a significantly lower clinical score. Histological analysis revealed a comparable degree of inflammation of the peripheral nervous system and demyelination in the spinal nerve roots of bovine peripheral nerveand P,-immunized rats. The P, plus Glq and P, plus GC groups revealed similar degrees of pathology in the spinal nerve roots but the latter group stood apart from the rest in that it showgd widespread peripheral nervous system changes extending distally into the sciatic nerve. Serological analysis demonstrated that P, and GC, but not GA, elicited antibody (I&) responses, but there was no correlation between antibody titer and clinical or histological involvement. The present data fail to support an enhancing role for gangliosides in the expression of EAN and, by extrapolation, in the Guillain-Bard syndrome, for which EAN serves as the laboratory model, and in which suggestions haye been mads that antibodies to GA may have pathogenetic significance. Ponzin D, Menegus AM, lrschner G, Nunzi MG, Fiori MG, Raine CS. Effects of gangliosides on the expression of autoimmune demyelination in the peripheral nervous system. A n n Neurol 1991;30:678-685 In the central nervous system (CNS), autoimmune demyelination has been shown to be augmented by the incorporation of certain myelin glycolipids (e.g., galactocerebroside [GCI) into the inochting emulsion 11-31. This effect has been attributed to myelin glycolipids acting as haptens in the generation of myelinolytic antibodies operating in concert with the T cellsensitiziqg antigen, myelin basic protein (MBP) 121. In the peripheral nervous system (PNS), a similar pathological scpario involving the PNS myelin proteiu P,, and glycolipids has been implicated in experimental allergic neuritis (EAN), but the mechanisms underlying the inflammatory events remain to be elucidated {4}. Whether other myelin components may be operative in autoimmune demyelination in the PNS has yet to be assessed. In this regard, glycoconjugates, including neutral and acidic glycolipids, occurring naturally in both neuronal membranes and myelin 15-71, have been suggested to be causally related to specific monoclonal antibodies in a number of human autoimmune neuropathies [8-141. Polyclonal antiganglioside antibodies have been described in a variety of neurological
Materials and Methods AYZiTZdh Thirty-six adult male Lewis rats (Charles River, Wilmington, MA), were used. Animals weighed between 250 and 300
From the *Departments of Neuropathology and Neuroscience, Albert Einstein College of Medicine, The Bronx, NY, and tFidia Research Laboratories, Abano Terme, Itdy.
Address correspondence to Dr Raine, Dept. of Pathology (Neuropathofogy), Alberr Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461.
conditions with suspected immunological involvement (e.g., amyotrophic lateral sclerosis, multiple sclerosis, Guillain-Barre syndrome) [l5-20). While the specificity and significance of these antibodies remain unclear, recent communications to this journal raised the possibility that under certain conditions, antibodies to gangliosides (GAS) may have a pathogenetic role in some peripheral neuropathies, in particular, the Guillain-Barre syndrome 12 1-23]. To investigate the latter possibility, the present study was undertaken to determine whether purified GAS, delivered either alone or in combination with the major neuritogen in EAN, P, protein, may affect the expression of immune-mediated demyelination in the PNS. The relative effects of GC, which is known to augment immune demyelination [l, 27, were studied for comparison.
Received Oct 29, 1990, and in revised form Feb 27 and May 2, 1991. Accepted for publication May 2, 1991.
678
Copyright 0 1991 by fhe American Neurological Association
3 or 4 per cage, and were fed standard chow and water ad libitum.
gm, were housed in groups of
lnocula and Animal Groups Freshly dissected intradural spinal roots were homogenized 1: 1 in saline solution. Aliquots of the suspension were emulsified with equal volumes of complete Freund‘s adjuvant (CFA) containing 10 mg Mycobactm’urn tuberculosis per milliliter. Five rats were subcutaneously injected with 0.2 mi of the emulsion into the dorsum of each hindfoot. In this way, animals received about 200 pg of P, protein, a known neuritogenic dose 143. BOVINE PERIPHERAL NERVE (BPN GROUP).
Bovine P, was purified from spinal root myelin as previously described [24}. Five rats were given one inoculation containing 0.2 mg P, in 0.1 ml saline solution emulsified in CFA. BOVINE p2 PROTEIN (P, GROUP).
+
p2 PLUS GALACTOCEREBROSIDE (p2 GC GROUP). G C (Supelco, Bellefonte, PA) was prepared as a micelle preparation with lysolecithin and bovine serum albumin 111. Four animals were given a dosc of 0.2 ml, made up of 0.1 ml containing 1 rng GC plus 0.2 mg P,, and 0.1 ml CFA. The ratio between GC and P, was similar to that used previously for experimental allergic encephalomyelitis (EAE) {l], except that the concentration of P, in PNS myelin was estimated to be about half that of MBP in CNS myelin.
+
p2 PLUS GANGLIOSIDES (P, GA GROUP). A highly purified bovine brain G A mixture comprising 21% GM,, 40% GD,,, 16% GD,,, 19% GTIb, 2% GD,, and 2% G Q l b (FIDIA S.p.A, Abano Terme, Italy) was dissolved in saline solution. P, was added so that 0.1 ml contained either 0.2 mg P, plus 0.02 mg GA (P2 GA(a) group), where the P2/GA ratio was similar to that found in intact PNS myelin [6], or 0.2 mg P, plus 1 mg GA (P, GA(b) group), with GA at a concentration known to elicit a delayed-type hypersensitivity response 1251. The two GA solutions were emulsified with equal volumes of CFA and delivered to two groups of 4 rats, as described above.
+
+
Two groups of 3 and 7 rats received 0.02 mg (GA(a)) and 1 mg (GA(b)) of the G A mixture, respectively. This was emulsified in CFA and delivered as described above.
GANGLIOSIDE (GA GROUP).
GALACTOCEREBROSIDE (GC GROUP). G C was not tested alone in CFA because this compound has been shown in several studies not to elicit neuropathological changes after single inoculation [I, 2, 4}.
Four rats received 0.2 ml CFA emulsitied 1:1 in saline solution. COMPLETE FREUNII’S ADJUVANT GROUP.
Clinical Evaluation Animals were observed daily and weighed from day 8 after injection onward. Clinical signs were evaluated using a 5-point scale: 0 = normal, 1 = limp tail, 2 = slight paraparesis, 3 = one leg paralyzed, 4 = paresis of both hind and/or front limbs, 5 = moribund or dead. Scoring was carried out
by two independent, blinded observers. Rats were anesthetized and perfused with fixative 48 hours after disease onset. Animals failing to develop signs were sampled 15 days after injection. Comparisons between groups were made by analysis of variance with orthogonal contrasts. Animals were perfused with cold 2.5% glutaraldehyde in phosphate buffer [l, 2). Spinal cord, proximal segments of L5, L6, and S1 ventral and dorsal roots and corresponding dorsal root ganglia, plus both sciatic nerves were dissected, thinly sliced, postfixed in chrome osmium, dehydrated, and embedded in epoxy resin (Epon 812). Slides containing a total of 8 to 10 semithin (1-pm) sections of the entire root or nerve were stained with toluidine blue and evaluated in a blinded fashion by at least three independent observers. The index for inflammation was: 0 = none; 1 = a few scattered inflammatory cells; 2 = numerous scattered cells, an occasional perivascular cuff; 3 = many perivascular cuffs; 4 = confluent perivascular cuffs; 5 = extensive perivascular cuffing with involvement of the entire nerve. The index for demyelination was: 0 = none, 1 = a few scattered naked axons, 2 = small groups of naked axons, 3 = larger groups of demyelinated axons, 4 = confluent foci of demyelination, 5 = widespread demyelination. From each animal, mean values were determined for both ventral and dorsal roots and sciatic nerves, and were pooled for each experimental group. Comparisons between groups were made by analysis of variance with orthogonal contrast. MORPHOLOGY.
Prior to sampling, rats were bled from the heart and serum samples stored at - 20°C. For the detection of antibodies against P,, GA, and GC, an enzymelinked immunosorbent assay (ELISA) was employed 126, 271. For anti-GA antibody assays, 750 ng of GM,, GD,,, GD,,, or GT,, was diluted in ethanol and water in a 1 :1 ratio and added to each well of a flat-bottom microtiter plate (Flow Labs, McLean, VA). Incubation was carried out at room temperature for 90 minutes, and nonspecific binding was blocked for 90 minutes with 2% bovine serum albumin in phosphate-buffered saline (PBS) (Flow Labs). After washing with buffer, 50 pl of serum diluted 1: 100 to 1:6,400 in 2% bovine serum albumin in PBS was added and the plate incubated overnight at 4°C. After washing, 50 pl of horseradish peroxidase-conjugated goat anti-rat IgG (Sigma, St Louis, MO) or IgM (Cappel, West Chester, PA), diluted 1: 1,000, was added for 2 hours at room temperature and then washed. Antibodies were detected by adding 50 ~1 of orthophenylanine-diamine reagent (Sorin, Vercelli, Italy). Absorbance was read at 492 nm on a Titertek Multiskan ELISA reader (Flow Labs). For anti-P, and anti-GC antibody assays, 100 pl of a solution of P, in PBS (1 pg/ml) or GC in ethanol (10 pg/ml) was added to each well. Incubation was performed at 4°C overnight (P,) or at room temperature for 120 minutes (GC). Nonspecific binding was blocked with 1% polyvinylpyrrolidone for 20 minutes at room temperature and then 100 pl of serum diluted 1: 100 to 1:25,600 in PBS plus 0.05% Tween 20 (Sigma) was added and the plate incubated overnight at 4°C. Detection of antibodies was then performed as described above. IMMUNOLOGICAL ASSAYS.
Ponzin et al: Gangliosides and EAN 679
Table 1. Prodtlction of Experimental Allergic Neuritis in Lewis Rats with DIZfferentAntigensa
Dose (mgianimal)
Day of Onsetb (after Inoculation)
BPN
50
Complete Freund's adjuvant GA(a) GA(b)
-
11.0 0 0 0 11.2 k 13.2 k 12.5 ? 11.7 k
Antigen
P2 P2 P2 P2
0.02 1.00 0.20 0.20 0.20 0.20
+ GC
+ GA(a) + GA(b)
Clinical Disease
* 0.0
+
+ +
1.00 0.02 1.00
Incidence
Mean Scoreb
515 014 013
2.3 k 0.5 0 0 0 3.6 0.2 2.1 k 0.3 1.5 2 0.2 1.7 k 0.1
017 515
0.5 0.5
414 414 414
0.6 0.5
*
"P, and glycolipids were mixed together and emulsified with complete Freund's adjuvant containing 10 mgiml of Mycobarterium tabercalosis. Two-tenths milliliters of this mixture was injected subcutaneously in the dorsum of the posterior feet. bValues are mean 5 standard crror.
BPN
=
bovine peripheral nerve; P2
=
bovine P, protein; G C
=
galactocerebroside; GA
The results are presented as mean optical densities (OD) 1 standard deviation, and were obtained at 1: 100 (GA, P,, GC) and 1:3,200 (P2).Control values were determined from the sera of 15 normal Lewis rats from the same source. A positive result was 1 OD greater than 2 standard deviations above the mean OD for control animals analyzed in the same experiment. Data on all P,-treated groups were compared by Scheffe's multiple comparisons test. 2
Results Clinical Finding BPN-inoculated rats presented with neurological signs on day 11 after injection (Table 1). For neuropathology, 4 rats were sampled 48 hours after disease onset when they showed mild to severe paraparesis (clinical score, 1 to 4),and 1 was taken 5 days after onset, when the clinical course had stabilized at grade 3. The P, group developed significantly more severe signs than the BPN and the P, GA groups (p < 0.01). In these rats, signs began between days 11 and 13 after injection and, after 48 hours, reached a clinical level of 3 to 4. GC + P, animals became sick between days 12 and 14 after injection. At the time of death (48 hours later), neurological signs were less severe than those of the P,-injected group (grades 1 to 3, p < 0.01). GA-sensitized rats presented with different clinical pictures. When GA was given alone in CFA (both doses), no neurological signs developed. Between days l l and 14, all rats in the P, GA groups developed clinical signs similar to but less severe than those from P,-inoculated rats (clinical score 1 to 2 ) . In all groups, most rats developed an adjuvant arthritis with a time course overlapping that of the neurological syndrome. This was taken into account by not including the hindfeet stiffness and increased sensitivity in the evaluation of the clinical score. In addition, all animals displayed an abrupt overnight decrease in body weight that preceded or accompanied neurological
+
+
=
gangliosides.
signs. This reduction was even apparent in nonsymptomatic rats immunized with CFA alone or GA + CFA.
Neuropatbology PNS tissue from all animal groups was examined. Mean histological scores for ventral roots, dorsal roots, and sciatic nerves from all animals sampled (48 hours after onset) are detailed in Figure 1. Readings from all levels of either ventral or dorsal roots were pooled since there was no clear difference. Dorsal roots were usually more involved than were ventral roots. With regard to inflammation (see Fig 1, upper panel), there was no difference between groups except for the P, + GC group, in which significantly fewer inflammatory cells were found in ventral roots (p < O.O1), in comparison to rats treated with P, alone. The P, GC group was unusual in that animals showed inflammatory cells in the sciatic nerve as an early event (' < 0.001). One animal in the P, + GA group showed some inflammatory cells in the sciatic nerve. The one animal from the BPN group that was killed at a later time point, when the clinical score had stabilized at grade 3, demonstrated ventral and dorsal roots to be equally affected and some inflammatory cells in the sciatic nerve. With regard to demyelination (see Fig 1, lower panel), there was no difference between the groups, except for the P, + GC group in which demyelination was also found in the sciatic nerve (p < 0.001). In the single animal sampled later in the BPN group, dernyelination was also evident in the sciatic nerve. Comparison of the various groups confirmed the presence of relatively uniformly myelinated nerve fibers in spinal nerve roots and sciatic nerve of normal animals, the endoneurium containing no infiltrating cells. In GA-inoculated animals, the PNS (Figs 2A and 2B) was normal. CFA controls also showed no PNS changes. In BPN-sensitized rats, lesions typical of
680 Annals of Neurology Vol 30 No 5 November 1991
+
z
-
0 k
U
sciatic nerve was involved (i.e., within 48 hours of onset) and perivascular cuffing was more marked. Axonal loss occurred in all groups but was more apparent in animals treated with P, plus GC. Animals inoculated with P, plus GA ((a) and (b)) presented with milder lesions than did the BPN and P, groups (see Fig 2F). Spinal cord changes were not seen, but on occasion, infiltration of a few mononuclear cells into the leptomeningeal space was seen in neurologically afflicted animals. However, these cells did not invade CNS parenchyma.
“1 3
I 5
A U.
2
5 1
0
BPN
P2
PZ+GC
PZ+GA(a)
-F
BPN
P2
P2+GC
P2+GA(o)
P2+GA(b)
T
P2+GA(b)
Fig 1. Comparative histopathology of experimental allergic neuritis groups. Average infEammation (upper panel) and demyelination (lower panel) of ventral roots (white bars), dorsal roots (shaded bars), and sciatic nerves (black bars) 2 standard errors. See Materials and Methods fir a description of the groups and of the scoring system. Note the uniformity in disease expression in all groups except the one that was inoculated with P2 and galactocerebroside (Pz GC) where there is a distal shift in PNS involvement. BPN = bovine peripheral nerve; GA = galactoside.
-+
acute EAN were encountered and were invariably centered on blood vessels. The endoneurial space was increased and contained infiltrating cells, some perivascular cuffing was apparent, and nerve fibers at various stages of demyelination were evident (see Fig 2C). Demyelination invariably occurred in association with infiltrating macrophages, some of which could be seen within the Schwann cell tube. Infiltration by polymorphonuclear leukocytes occurred and was common in all groups. P,-induced EAN was essentially identical to BPN-induced disease (see Fig 2D). In animals sensitized with P, plus GC, spinal nerve root lesions were qualitatively similar to those of the BPN and P, groups but sciatic nerves were extensively infiltrated (see Fig 2E). Therefore, EAN was more widespread when G C was combined with P,, and there was a shift in the pathology from radicular to more distal regions. The P, + GC group represented the only group in which
Serology No 1gM response was found against any of the antigens. Sera from the groups inoculated with CFA or GA contained no IgM or IgG response against P,, GA, or GC (Table 2). Antibody reactivity is shown for GM, only because it was representative of results from the other three GAS tested. P,- and BPN-sensitized rats developed antibodies against P,. The lowest P, response was seen in sera from BPN rats, whereas rats treated with P, or P2 plus GA(a) developed higher responses. For this reason, readings from the higher dilution tested are also shown, where the OD response was linear. Somewhat lower values, not statistically significant, were found in rats injected with P, plus G C or P, plus GA(b), even though some variation among individuals was found. Antibodies against the four main GAS in the mixture were not demonstrable in most animals. A borderline positive response was seen in 1 rat from the P, + GA(a) group and 2 from the P, + GA(b) group. Only the P, + GC group displayed a positive antibody response against GC at the time point tested (48 hours after onset). A borderline positive response was seen in the sera of 2 rats from the P, GA(b) group. Taken together, the ELISA findings showed no significant correlation between humoral response and disease findings.
+
Discussion On the basis of suggestions that GAS might serve as pathogens in PNS disease 121-231, the present study investigated whether GAS possess the ability to enhance immune-mediated tissue damage in a model of EAN in the Lewis rat. Such an augmentation effect had previously been shown with G C in EAE El-31. These latter studies on EAE concluded that tissue damage was the result of an initial T-cell response to MBP and a B-cell response to G C 111. In support of this, demyelination was observed in the rabbit retina after the introduction into the vitreous of anti-GC antiserum plus cytokines [2]. In contrast to the situation with MBP in EAE, sensitization to P, is sufficient to elicit full-blown EAN with infhnmation and demyelination [28, 291. However, the incorporation of G C into the protocol
Ponzin et al: Gangliosides and EAN 681
A
B
C
D
E
F
682
Annals of Neurology Vol 30 No 5
November 1991
Table 2.Antibody Reactivity by Enzyme-Linked lmmunosorbent Assay" IgG Anti-P,
Group
(1: 100)
Normal control
227 177 109 i 42 119 67 291 2 130 1,031 i 140 2,645 +- 63 2,102 i 142 2,113 80 2,273 232
*
Complete Freund's adjuvant GAb) GA(b) BPN
p2 P2 + GC P2 + GA(a) P, + GA(b)
(1 :3,200)
IgG Anti-GM, ( 1 : 100)
IgG Anti-GC (1:lOO)
-
54 2 17 41 t 18 43 2 16 53 2 32 65 2 12 85 20 50 & 15 109 2 54 125 ? 66
112 i 67 90 26 102 2 20 114 -t 37 216 & 71 146 & 65 420 2 173 221 58 220 105
-
*
-
52 ? 19 1,049 ? 302 452 2 176 1,570 ? 309 435 ? 346
* *
*
*
*
*
aVdues are mean optical densities ( x lo3)2 standard deviation. Numbers in parentheses indicate the dilution of the sera Each s e w sample was assayed in triplicate at the different dilutions. Mean values and standard deviations represent interanimd values. The intra-assay variatiqn was approximately 10%.
GA = gangliosides; BPN
=
bovine peripheral nerve; GC
=
galactocerebroside.
has been shown to augment tissue damage in the spinal nerve roots {4] and, in the present case, caused an early, distal involvement of the sciatic nerve. Sciatic nerve is usually not involved until later time points and the occurrence of pathology at this level in the present animals may be related to the low levels of circulating antibody to GC. Intraneural injection of anti-GC antiserum alone has been shown to be sufficient to produce demyelinated lesions of peripheral nerve {30].
2. (A) Ganglioside (GA(b))-sensitizedrat, epoxy 4 Fig resin section. Transverse section o f an LS ventral root from a rat 1-pm
displaying no clinical disease reveals normal nerve fibers and no injammatory changes. (B) Sciatic nerve of the same animal as A. Note the normal appearance of the nerve. (C) Bovine peripheral nerve (BPN)-inoculated rat, 1-pm epoxy wsin section. Detail from a transverse section o f an L6 ventral root from a rat displaying clinical signs. The entire section was scoved 3 for inflammation and 4 f i r LmyeLination. Note the injammuto y cells surrounding a x o m undergoing akmyelination. ( D ) Pzsensitized rat, 1-pm epoxy resin section. Detail from a transverse section of an S i dorsal root from an experimental allergic neuritis (EAN)-afjacted rat. This corresponded to grades 4 (inflammation) and 3 (demyelination).Note the presence of some polymorpbonuclear leukocytes and an axon undergoing demyelination (arrow}. (E) Pz and galactocerebroside (GC)-sensitized rat, 1-pm epoxy resin section. Transverse section of sciatic nerve from a rat with clinical EAN. Note the conjuent perivascular cufs and scattered demyelinated fibers. This corresponded t o graah 4 (inflammation) and 3 (&myelination). Note the perivascular cuf and some demyelinated axons (arrows). (F)Pz and GA(b)-sensitized rat, 1-pm epoxy resin section. Transverse section of an S1 ventral root of a rat displaying EAN. This root scored 3 for injammation and 4 for demyelination. Note the demyelinated axons (arrow) and an injammatory cell adhering t o the endothelium of the vessel (center). (A, C-F, toluidine blue; A, B, x 375; C-F, x 755.)
The present experimental protocol compared Lewis rats sensitized against a known neuritogenic dose of P, protein with animals given P, plus different doses of GAS and, for comparative purposes, GC.A mixture of GAS was used since it has been observed that the immune reaction to these compounds yields antibodies reacting with the Gal(P 1-3)GalNAc epitope of GM,. This is shared by other glycolipids and glycoproteins in the myelin and ponmyelin fractions {22]. Furthermore, when mixtures of total brain GAS are administered, the predominant antibody response is against GM, and asialo-GM1 1311. In agreement with previous observations, the combination of P, with G C in the inoculum increased the PNS lesion area but not its severity. While inflammation and demyelinarion in P, animals were limited to spinal roots supplying the sciatic nerve, the inclusion of GC into the protocol led to lesions in the distal nerve as well. In contrast to GC, GAS did not enhance lesion areas induced by P2.The data showed that when used alone, GAS produced no pathological changes, and when GASwere coupled with P, protein, signs that occurred were less severe than those seen in animals receiving P, alone (see Table 1). In the present study, the clinical score was lower when either GAS or G C were combined with P,, a finding discordant with the observation that there was essentially no difference in the degree of neuropathological involvement between the groups. The reduced clinical score for the P, + G C group, in spite of the increased pathology, might be related to the lesions being less radicular and more distal, thus rendering the effect less detectable clinically. A similar dissociation between clinical and structural changes has also been reported in previous studies of EAE and EAN C32, 331. The lack of change in the humoral response to P, protein after the addition of G C or GAS to the emulPonzin et al: Gangliosides and EAN
683
sion might be explained by the neuritogenic activity of the CN-1 peptide of P, correlating better with a T-cell, rather than a B-cell, response E343. Antibodies against GAS were not detected in most animals. This supports further the low immunogenicity of these glycolipids
c35, 361. The failure of exogenous GAS to enhance the response to P, might relate to a number of phenomena, particularly when one compares the effect with that of GC. The neuritogenicity of P, appears to be conformation dependent [ 3 7 ) and interactions between P, and GAS might lead to conformational changes that are ineffective in augmenting neuritogenicity, in contrast to the case with GC. Thus, it appears that GAS exert no augmenting role in autoimmune demyelination and it is possible that the presence of GASin the sensitizing mixture may have interfered with interactions between neuritogenic determinants and the immune system. Interestingly, the administration of exogenous GAS has been claimed to offer partial protection against EAN induced in myelin-sensitized rats [38). Other studies on EAE have documented a protective role for GM, and GM4 [39J and a delayed hypersensitivity response to GAS that was not encephalitogenic c251. Taken in concert, our observations have shown GAS to be nonantigenic in the PNS and suggest that when combined with P,, a strong neuritogen, exogenous GAS might lead to a lessened pathological effect. The possible immunological mechanisms underlying these discrepanices are currently under investigation. This study was supported in part by grants (to C.S.R.) from the National Instirutes of Health (NS 08952 and NS 11920) and National Multiple Sclerosis Society (RG 1001-G-7). The authors thank Drs Krzysztof W. Selma], Alessandro Bruni, Lanfranc0 Callegaro, Alberta Leon, and Gin0 Toffano for helpful discussion; Everett Swanson, Howard Finch, Miriam Pakingan, Marta Bassan, Paola Facco, and Anna Rosa Morandin for expert technical assistance; Dr George Hashim for P, protein; and Michele Brigs for careful preparation of the manuscript.
References 1. Rame CS, Traugotr U, Farooq M, et al. Augmentation of immune-mediated demyelination by lipid haptens. Lab Invest 1981;45:174-182 2. Brosnan CF, Traugoa U, Raine CS. Analysis of humoral and cellular events and the role of lipid haptens during CNS demyelination. Acta Neuropathol (Berl) 1983;9(suppl):S59-S70 3. Hosein 22, Gilbert JJ, Strejan GH. The role of myelin lipids in experimental allergic encephalomyelitis. J Neuroimmunol 1986;10:219-233 4. Hughes RAC,Powell HC. Experimental allergic neuritis: demyelination induced by P2 alone and non-specific enhancement by cerebroside. J Neuropathol Exp Neurol 1984;43:154-161 5 . Yu RK, Saito M. Structure and localization of gangliosides. In: Margolis RU, Margolis F X , eds. Neurobiology of glycoconjugates. New York: Plenum Press, 1989:l-42 6. Fong JW, Ledeen RW, Kundu SK, Brostoff SW. Gangliosides of peripheral newe myelin. J Neurochem 1976;26:157-162
684 Annals of Neurology Vol 30 No 5
7. Ledeen RW, Cochran FB, Yu RK, et al. Gangliosidcs of the CNS myelin membrane. Adv Exp Med Biol1980;125:167-176 8. Quarles RR, Ilyas AA, Willison HJ. Antibodies to glycolipids in demyelinating diseases of the human peripheral nervous system. Chem Phys Lipids 1986;42:235-248 9. Latov N, Hays AP, Sherman WH. Peripheral neuropathy and anti-MAG antibodies. CRC Crit Rev Neurobiol 1988;3: 301-332 10. Ilyas AA, Wilhon HJ, Dalakas MO, et d. Identificacion and characterization of gangliosides reacting with 1gM paraproceins in three patients with neuropathy associated with biclonal gammopathy. J Neurochem 1988;5:85 1-858 11. Nardelli E, Steck AJ, Barkas T, et al. Motor neuron syndrome and monoclonal IgM with antibody activity against gangliosides GM1 and G D l b . Ann Neurol 1988;23:524-528 12. Larov N , Hays AP, Donofrio PD, et al. Monoclonal IgM with unique specificity to gangliosides GM1 and G D l b and to lactoN-tetraose associated with human motor neuron disease. Neurology 1988;38:763-768 3. Ito H, Latov N. Monoclonal IgM in two patients with motor neuron disease bind to the carbohydrate antigens Gal(B13)GalNAc and Gal(B1-3)GlcNAc. J Neuroimmunol 1988;19: 245-253 4. Marcus DM, Percy L, Gilbert S, et al. Human IgM monoclonal proteins that bind 3-fucosyllactosamine, asialo GM1 and GM1. J Immunol 1989;143:2929-2932 15. Endo T, Scott DD, Stewart SS, et al. Antibodies to glycosphingolipids in patients with mulciple sclerosis and SLE. J Immunol 1984;132:1793-1797 16. Pestronk A, Chaundhry V, Feldman EL, ec al. Lower motor neuron syndromes defined by parterns of weakness, nerve conduction abnormalities, and high titers of antiglycolipid antibodies. Ann Neurol 1990;27:316-326 17. Pestronk A, Adams RN, Clawson L,et al. Serum antibodies to GM1 ganglioside in amyotrophic lateral sclerosis. Neurology 1988;38:1457-1461 18. Ilyas AA, Willison HJ, Quarles RH, et al. Serum antibodies to gangliosides in Guillain-Barri. syndrome. Ann Neurol 1988; 23:440-447 19. Pestronk A, Cornbluth DR, Ilyas AA, et al. A treatable multifocal motor neuropathy with antibodies to GM1 ganglioside. Ann Neurol 1988;24:73-78 20. Shy ME, Evans VA, Lublin FD, et al. Antibodies to GMI and G D l b in patients with motor neuron disease without plasma cell dyscrasia. Ann Neurol 1989;25:511-513 21. Svennerholm L, Fredman P. Antibody detection in GuillainBarrC syndrome. Ann Neurol 1990;27(suppl):S36-S40 22. Latov N. Neuropathy and anti-GM1 antibodies. Ann Neurol 1990;27(~~ppl):S4 1443 23. Quarles RH, Ilyas AA, Willison HJ. Antibodies to gangliosides and myelin proteins in Guillain-Barre syndrome. Ann Neurol 1990;27(SUPPI):S~ 8-S 52 24. Brostoff SW, Sachs H, Dal Canto M, et al. The P2 protein of bovine root myelin: isolation and some chemical and immunological properties. J Neurochem 1974;23:1037-1043 25. Offner H , Standage BA, Burger DR, Vandenbark AA. Delayed-type hypersensitivity to gangliosides in the Lewis rat. J Neuroimmunol 1985;9:147-157 26. Quarles RH. Human monoclonal antibodies associated with neuropathy. Methods Enzymol 1989;179:291-299 27. Nakayasa H , Ota K, Tanaka H , er al. Suppression of actively induced and passively transferred experimental allergic neuritis hy cyclosporin A. J Neuroimmunol 1990;26:219-227 28. Kadlubowski M, Hughes RAC. Identification of the neuritogen of experimental allergic neuritis. Nature (Lond) 1979;277: 140-141 29. Kadlubowski M, Hughes RAC, Gregson IVA. Experimental al-
November 1991
30.
31.
32.
33.
34.
lergic neuritis in the Lewis rat: characterization of the activity of peripheral myelin and its major basic protein, P,. Brain Res 1980;184~439-454 Saida K, Sumner AJ, SaidaT, et al. Antiserum-mediated demyelinarion: relationship between remyelination and functional recovery. Ann Neurol 1980;8:12-24 Rapport MM, Graf L, Brunner W, Yu RK. Antibodies to total brain gangliosides: titer and specificity of antisera. In: Svennerholm L, Mandel P, Dreyfus H, Urban P-F, eds. Advances of experimental medicine and biology, vol 125. New York, Plenum Press, 1980:327-334 Raine CS, Snyder DH, Stone SH, Bornstein MB. Suppression of acute and chronic experimental allergic encephalomyelitis in strain 13 guinea pigs. J Neurol Sci 1977;31:355367 Rostami A, Brown MJ, ksak RP, et al. The role of myelin P, protein in the production of experimental allergic neuritis. Ann Neurol 1984;16:680-685 Milek DJ, Cunningham JM, Powers JM, Brostoff SW. Experi-
35.
36.
37.
38.
39.
mental allergic neuritis: humoral and cellular response to the cyanogen bromide peptides of the P, protein. J Neuroimmunol 1983;4:105-117 Marcus DM. A review of the immunogenic and immunomodulatory properties of glycosphingolipids. Mol Immunol 1984; 21:1083-1091 Lvingston PO, Ritter G, Calves MJ. Antibody response after immunization with the gangliosides GM,, GM?, GM,, GD, and GD3 in the mouse. Cancer Immunol Immunother 1989;29: 179-184 Brostoff SW. Immunological responses to myelin and myelin components. In: Morel1 P, ed. Myelin. New York, Plenum Press, 1984:405-439 Ledeen RW, Oderfeld-Now& B, Brosnan CP, Cervone A. Gangliosides offer partial protection in experimental allergic neuritis. Ann Neurol 14c)0;27(suppl):S6c)-S74 Mullin RB, Patrick DH, Poore CMB, et al. Prevention of experimental allergic encephalomyelitis by ganghoside GMS. A follow-up study. J Neurol Sci 1986;73:55-60
Ponzin et al: Gangliosides and EAN
685
