Journal of the Neurological Sciences, 109 (1992)200-206 © 1992ElsevierSciencePublishersB.V. All rightsreserved0022-510X/92/$05.00
200
JNS 03751
Guillain-Barr6 syndrome associated with high titers of anti-GM1 antibodies Eduardo Nobile-Orazio a, Marineila Carpo a, Nicoletta Meucci a, Maria Pia Grassi b, Erminio Capitani c, Monica Sciacco a, Alfonso Mangoni b and Guglielmo Scarlato a a Institute of Clinical Neurology, Centro Dino Ferrari, Ospedale Maggiore Policlinico, and Departments of Neurology, b Ospedale L. Sacco, and c Ospedale San Paolo, University of Milan, Milan, Italy
(Received 10 September, 1991) (Revised, received21 November,1991) (Accepted 24 November,1991) Key words: GM1; Gangliosides; Guillain-Barr6 syndrome; Autoimmunity; Antibodies; IgG
Summary We found high titers of anti-GM1 antibodies (1/1280 or more) in 3 of 14 consecutive patients (21%) with Guillain-Barr6 syndrome (GBS) and in 2 additional patients who developed GBS, 10-11 days after starting parenteral treatment with gangliosides. Antibodies were IgG in 4 patients and lgM in one, and they all bound to asialo-GM1, and, in 3, to GDlb as well. Although the clinical features in all the patients with high anti-GM1 titers fulfilled the criteria for the diagnosis of GBS and in 4 of them, proteins but not cells were elevated in cerebrospinal fluid, electrodiagnostic studies in 3 patients showed prominent signs of axonal degeneration, that in one case were confirmed by morphological studies on sural nerve biopsy. No recent antecedent infection was reported by these patients, but in 3, including patients treated with gangliosides, anti-Campylobacter jejuni antibodies were elevated. In 3 patients a consistent decrease in anti-GM1 levels was observed after the acute phase of the disease suggesting that the frequent occurrence of these antibodies in patients with GBS and their frequent association with a prominent axonal impairment may have pathogenetic relevance.
Introduction
High titers of anti-GM1 lgM antibodies have been reported in some patients with motor neuron disease (MND), motor neuropathy or, less frequently, with sensorimotor neuropathy (llyas et al. 1988; Latov et al. 1988; Nardelli et al. 1988; Pestronk et al. 1988, 1990; Nobile-Orazio et ai. 1990a,b; Sadiq et al. 1990; Shy et al. 1990). In some of these patients therapeutic reduction of antibody levels was associated with clinical improvement supporting a pathogenetic role for these antibodies (Latov et al. 1988; Pestronk et al. 1988; Shy et al. 1990). Recently, high titers of anti-GM1 antibodies, mostly IgG, have been detected in some patients with the Guillain-Barr6 syndrome (GBS) (Ilyas et al. 1990; Oomes et al. 1990; Yuki et al. 1990; van den Berg
Correspondence to: E. Nobile-Orazio, M.D., Institute of Clinical Neurology,Universityof Milan, Via F. Sforza 35, 20122Milan, Italy. Tel.: 39/2/55184306; Fax.: 39/2/55190392.
et al. 1991; Walsh et al. 1991) and an association between these antibodies and Campylobacter infection has been reported (Oomes et al. 1990; Yuki et al. 1990; Walsh et al. 1991). We found high titers of anti-GM1 antibodies in 5 patients with GBS, 2 of whom developed symptoms after therapeutic injection of gangliosides. The clinical, electrophysiological and laboratory findings in these patients are reported.
Material and methods
Patients
Serum was obtained from 226 patients including 16 patients with a clinical diagnosis of GBS (Asbury and Cornblath 1990), 52 patients with neuropathy (PN) whose etiology was known in 25 (toxic in 5, vitamin B n deficiency in 1, diabetes in 2, inherited in 7, crioglobulinemic in 2, carcinomatous in 1, associated with Sj6gren's syndrome in 1, with benign lgG monoclonal
201 gammopathy in 6), unknown in 16, while 9 had chronic inflammatory demyelinating polyneuropathy (CIDP) and 2 had motor neuropathy with multifocal conduction blocks. Fifty patients had motor neuron disease (MND), 25 multiple sclerosis (MS), 23 other neurological diseases (OND) while 46 had other immune diseases (OID) (2 polymyositis; 4 myastenia gravis; 16 Sj6gren's syndrome and 24 SLE), and 14 were normal subjects. Samples were kept frozen at -80°C until use. Sera from GBS patients were obtained within 1 month from onset of symptoms and were in all but 2 cases from a consecutive series of GBS patients admitted to the Institute of Clinical Neurology of the University of Milan. None of these patients ever received gangliosides before onset of the disease. The two remaining sera were from patients admitted elsewhere who developed symptoms of GBS 10 and 11 days, respectively, after starting treatment with a mixture of commercially available bovine brain gangliosides containing GM1, GDla, G D l b and GTlb in approximately the same proportion (Cronassial, Fidia, Italy), at a dose of 20 mg i.m. on alternate days for 8 days (no. 3 in the tables) and 20 mg i.m. daily for 10 days (no. 4). This therapy was prescribed for low back pain in one (no. 3) and for a recurring ankle pain in one (no. 4). The latter patient also had diabetes but denied having antecedent neuropathy symptoms.
ELISA for anti-GM1 antibodies All sera were tested by ELISA for lgG and IgM reactivity with GM1 (Fidia, Italy) following a procedure reported previously (Nobile-Orazio et al. 1990b). Briefly, 1 /~g of purified ganglioside in 100 /tl of ethanol was added to each well of 96-well polystyrene microtitration plate (CeU-Cult, England) and ethanol was evaporated. Unreactive sites were saturated with 1% bovine serum albumin (BSA) in phoshate-buffered saline (PBS), pH 7.4 (saturating solution). After 3 washes with washing/antibody solution (saturating solution with 0.05% Tween 20), patient serum was added in triplicate to the wells at 1:20 dilution and incubated for 16 h at 4°C. After 6 washes, peroxidase-conjugated goat anti-human IgG or IgM (Cappel, Pa) was added to the wells for 1 h at room temperature, followed, after 6 washes, by 100/~1 substrate solution (0.04% o-phenylenediamine, 0.1 M citrate, 0.2 M Na2HPO4, pH 5 with 0.012% H202). The color reaction was stopped after 60 min with 50/tl/well 0.1 M H2SO 4 and the absorbance was measured at 490 rim. Results are expressed as the mean absorbance obtained from the wells coated with GM1 subtracted of absorbances obtained from wells coated with BSA only. Sera were considered positive when this difference exceeded the value of 0.05. Positive sera were titrated by serial 4-fold dilution until negative. To compare data from different experiments, serum from a patient with high antibody
titers was used as positive control in each assay. Sera with high anti-GM1 titers were also tested by ELISA for antibody reactivity with purified GDla, GDlb and asialo-GM1, all kindly provided by Prof. S. Sonnino.
lmmunostaining after high-performance thin-layer chromatography (HPTLC) In patients with high antibody titers by ELISA, reactivity with gangliosides was confirmed by immunostaining after HPTLC (Kasai and Yu 1983; NobileOrazio et al. 1987) using a mixture of bovine brain gangliosides containing GM1, GDla, GDlb and GTlb (Cronassial, Fidia, Italy) or asialo-GM1. Either 0.5 /Lg/lane of purified asialo-GM1 or 2/~g of Cronassial were separated by HPTLC on aluminium-backed silica gel 60 HPTLC plates (Merck, Germany) and the sera were tested at 1 : 100 dilution and counterstained using peroxidase-conjugated rabbit immunoglobulins to human IgG or IgM (Dakopatts, Denmark).
Immunoabsorption studies Sera from GBS patients reacting with both GM1 and Campylobacter jejuni (see later) were absorbed with GMI using an affinity column prepared by conjugation to Sepharose 4B (Pharmacia, Lurid, Sweden) of a GM1 derivative containing 12-aminododecanoic acid as acyi moiety (Sonnino et al. 1989). Absorbed sera were tested for reactivity with GM1 by ELISA and with C. jejuni by complement fixation assay. Sera from these patients were also tested for anti-GM1 reactivity by ELISA after incubation for 12 h at 4°C with an excess of the C. jejuni lysate used in the complement fixation assay (Virion, Zurich, Switzerland).
Results
We found high levels of anti-GM1 antibodies (1/1280 or more) in 5 patients with GBS, including 3 of the 14 consecutive GBS patients admitted to our institute (21%) and the 2 patients who developed GBS after treatment with gangliosides. In four of these patients antibodies were IgG (Fig. 1), while one had high levels of anti-GM1 IgM (1/2560) (open symbols in Fig. 1). None of the controls had high anti-GM1 IgG titers while one previously reported patient with MND (Nobile-Orazio et al. 1990a) had anti-GM1 IgM titers of 1/20480. Low anti-GM1 IgG levels (up to 1/320) were found in 4 additional patients with GBS (29%) but also in 7 patients with other PN (13%), 6 with MND (12%), 5 with OND (22%) (1 epilepsy, 1 dementia, 1 cerebellar degeneration, 1 stroke and 1 myopathy), and 13 with OID (28%) (Fig. 1), while the frequency of low anti-GM1 IgM in these groups ranged from 5 to 20% (Nobile-Orazio et al. 1990b).
202 TABLE 1 CLINICAL FEATURES IN PATIENTS WITH GBS AND HIGH TITERS OF ANTI-GM1 ANTIBODIES Patient No.
Presenting symptoms
1
LL weakness UL&LL pain LL weakness UL weakness and tingling LL pain and weakness UL&LL pain and weakness
2 3 4 5
Days of progression
Sensory loss
Cranial nerve impairment
Deep tendon reflexes
Autonomic signs
Maximal impairment
Impairment a at last control (time)
7
no
no
absent
no
2
0 (3 yrs)
12
no
no
absent
yes b
4
2 (6 yrs)
7
no
absent
no
4
2 (7 mos)
10
no
R& L VII no
absent
yes c
3
1 (6 mos)
12
yes
absent
no
1
0 (1 yr)
R&L VII & XII
UL = upper limbs; LL = lower limbs; R = right; L = left. a Impairment: 0 = walks normally; 1 = walks autonomously but with disturbances; 2 = needs help/support to walk; 3 = unable to walk; 4 = requires ventilatory assistance; 5 = dead. b Urinary and fecal incontinence; orthostatic hypotension. c Synus tachicardia.
In patients with GBS and high levels of anti-GM1 antibodies reactivity with GM1 was confirmed by immunostaining after HPTLC that also showed binding of Ig with the same isotype of anti-GM1 lg, to asialoGM1 and, in 3 patients, to GDlb (nos. 3, 4 and 5), but not to GDla or GTlb. Reactivity with asialo-GM1 and GDlb by ELISA in positive patients had approximately the same intensity than reactivity with GM1. The clinical features of the neuropathy in the 5 patients with GBS and high anti-GM1 titers are summarized in Table 1. All patients fulfilled the clinical criteria for the diagnosis of GBS (Asbury and Corn1/81920
•
~ 1/20480' E
1/5120 -
~ 1/12Bo-
$ ~
t~
1/320-
1/so1/20-
TABLE 2 COMPARISON OF SOME CLINICAL FEATURES IN PATIENTS WlTH GBS WITH AND WITHOUT HIGH ANT1-GM1 TITERS
Mean age (yrs) (range) Sex Days of progression (range) Limb pain at onset Requiring ventilatory assistance Sensory loss Cranial nerve impairment Autonomic impairment Mean maximal impairment a Complete recovery at 6 months
High titers (No. 5)
No high titers (No. 11)
48.4 (41-64) 2M/3F 9.4 (7-12) 3/5 2/5 1/5 2/5
40.3 (16-74) 6M/5F 17.7 (5-30) 2/11 3/11 (1 dead) 4/11 4/11
2/5
3/11
2.8
2.2
2/5
6/11
a See impairment scale in Table 1. n
am
~
GBS Neuropathy MND MS OND PATIENTS'GROUPS
OJD
Nor'real subjects
Fig. 1. Anti-GM1 lgG antibody titers by ELISA in patients with Guillain-Barr~ syndrome (GBS), neuropathy, motor neuron disease (MND), multiple sclerosis (MS), other neurological diseases (OND), other immunological diseases (OID), and in normal subjects. High titers (1/1280 or more) of anti-GM1 IgG antibodies were detected in 4 patients with GBS. One additional patient with GBS and a previously reported patient with MND (Nobile-Orazio et al. 1990a), both with low anti-GMl IgG titers (open dots), had high titers of anti-GM1 IgM antibodies (1/2560 and 1/20480 respectively).
blath 1990) with a rapidly progressive weakness and areflexia reaching the peak within 2 weeks from onset, followed by a more or less complete recovery. Although a mild sensory impairment was observed in only 1 patient (no. 5), pain in the lower limbs was an early and persistent symptom in 3 of them. When the clinical features of patients with and without anti-GM1 antibodies were compared (Table 2), the mean time to reach maximal disability was shorter in patients with high titers and their mean functional impairment was slightly more severe than in those without anti-GM1.
203 TABLE 3 ELECTROPHYSIOLOGIC FINDINGS IN PATIENTS WITH GBS AND HIGH TITERS OF ANTI-GMI ANTIBODIES Patient No.
Days of illness
Nerve conduction studies Median
1 2 3 4 5
5 13 22 3 42 12 22
EMG
Motor nerves
Sensory nerves Ulnar
Peroneal
amp
CV
DL
amp
CV
DL
amp
CV
8.0 5.5 NR nd nd nd 7.0
50.0 50.0
3.9 3.4
3.7 nd nd 3.0 NR 1.5 nd
51.0
2.4
3.7 4.0 0.5 1.3 NR 0.4 4.0
5 3 . 0 3.5 5 0 . 0 4.4 38.8 35.8 4.4
5.6
49.0
2.0
44.0
Normal limits: upper lower 4.0
47,0
4.7
62.1 4.0 33.8
4.4 49.0
2.6
Ulnar
amp
DL
amp
nd 17 10 11 9 3 6 . 6 6.4 b 10 4 2 . 7 5.3 8
2.5 2.7 2.6 2.4 4.2 3.6
nd nd 12 nd nd nd 7
6.5
3.4
3.5
DL
Median
8
Surai DL
2.1
2.9
amp
DL
nd nd nd nd nd 5 5
5.0 4.4
3.1 6
Fibrillation MUAPs (0--, + + + + ) (neurogenic) changes) a 0 0 +++ nd ++ 0/+ ++
+ + ++++ ++++ +++ ++
4.1 6
amp: amplitude (mY or/zV); CV: conduction velocity (m/sec); DL: distal latency (msec); EMG: electromyogram. a Increased MUAPs duration and reduced recruitment (graded 0 to + + + + ). nd: not done; NR: no response. b F-wave absent.
The electrophysiological findings in patients with GBS and high anti-GM1 titers are summarized in Table 3. In none of these patients the proposed electrodiagnostic criteria for demyelination (Cornblath 1990) were fulfilled, while in 3 (nos. 2, 3 and 4) there were prominent signs of axonal degeneration. In one of these patients (no. 4) sural nerve biopsy was performed (Nobile-Orazio et al. 1987) 29 days after onset of symptoms, showing a reduced number of myelinated
fibers (3210/mm 2) on semithin sections with several thinly myelinated axons and occasional clusters of regenerating fibers. No inflammatory cell infiltrate or blood vessel abnormality was detected and no immunoglobulin deposit was found by direct immunofluorescence (Takatsu et al. 1985). By teased fiber studies (Dyck et al. 1984), 43% of fibers were normal (conditions A and B), 7% had abnormalities consistent with axonai degeneration (conditions E and H) while 43%
TABLE 4 LABORATORY FINDINGS IN PATIENTS WITH GBS AND HIGH TITERS OF ANTI-GM1 ANTIBODIES
Patient No.
Anti-GMi
1
IgG 1/1280 a IgM 1/20 IgG 1/1280 a IgM neg. IgG 1/81920 a,b IgM 1/320 IgG 1/320 IgM 1/2560 a,b IgG ]/20480
2 3 4 5
Isotype titers
Antecedent events (time before onset of GBS)
Anti-
Cells (n/ram 3)
CSF
days (phase) of disease
Days of disease
Protein (mg/di)
C. jejuni Ab titer
6 (prog)
6
35
2
NR
n.d.
9 (prog)
31
138
3
NR
1/40
3 (prog)
2
84
0
1/160
29 (star)
10
70
0
16 (stat)
16
176
4
gangliosides (10 days) gangliosides (11 days) NR
1/160 1/80
IgM 1/320 Without high anti-GM1 titers Total Day of No. Ab measurement 11
17.6 (7-28)
a Ig also binding to asialo-GM1. b Ig also binding to GDlb.
CSF Protein
CSF Cells
Antecedent events
Anti-C. jejuni titer
< 100:7 > 100:4
< 10:10 > 10:1
Infection: 6 NR: 5
< 1/80:9 n.d.: 2
204 Patient 2
1181920-
Patient 3
1120480-
111280•
10G
o
1180-
1180• 1120-
o
•
lb
~.~
2'0 3'o~ys"
i :'~ 1/20480-
-
i~IM
6 ;ears Patient 4
0
10
2'0 3'0 4;)
1/81920-
5() S0days//71~lonths Patient S
1/5120111280-
1/128011320•
1/320.
c
1/80.
1/80. 0
IoM
1/20-
1/20480-
1/5120-
1/20.
1112801/320-
1/320-
"= ,r--
IgG
1/5120-
1/5120
t
10
~"~~.~/~....
toM
1/20. 2'0 3'0
40days Iff
5months
0
10 2'0 3'0 40days ~/"~
1'year
Time from onset of symptoms
Fig. 2. Longitudinal study of anti-GM1 IgG and IgM antibody titers during the disease in four patients with GBS. Arrow heads indicate the days when plasmapheresis (3 liters each) was performed in patients nos. 3 and 4.
showed multiple and 7% isolated segments of demyelination and remyelination (conditions C and D), indicating that demyelination at this level was probably secondary to axonal damage (Dyck et al. 1984). Prominent electrophysiological signs of axonal degeneration were also found early in the course of the disease, in 3 patients with GBS not associated with high levels of anti-GM1 antibodies. Some of the laboratory findings in patients with high anti-GM1 titers are summarized in Table 4. All but one patient (no. 1) had increased cerebrospinal fluid (CSF) proteins with normal CSF cell count at some point of the disease. Although none of these patients reported an antecedent history of febrile or diarrheal illness within a month before onset of symptoms, 3 of them, including the 2 patients injected with gangliosides, had increased levels of anti-Campylobacter jejuni antibodies by complement fixation assay (normal up to 1/40). A 4-fold decrease in anti-C, jejuni titers was observed during recovery in two of them (nos. 3 an 4) while in one (no. 5) antibodies decreased by 2-fold. In all these patients anti-C jejuni antibodies were of the same lg isotype than anti-GM1 antibodies and both anti-C jejuni and anti-GM1 activity were removed by previous absorption of patients' sera with GM1 by affinity chromatography. Similarly, antibody reactivity with GM1 was markedly decreased in the sera of these patients after preincubation with C. jejuni lysate. None of the patients without anti-GM1 antibodies had increased levels of anti-C jejuni antibodies. In Fig. 2 the temporal prophile of anti-GM1 antibodies in 4 patients during the disease is represented while no follow up data is available in patient no. 1. Although antibody titers were still elevated in all patients at the time of last follow up, a consistent decrease in antibody levels was observed in three of them
after the acute phase. This was most evident in patient no. 3, in whom antibody titers were extremely high 3 days after onset of symtoms, and rapidly decreased after 6 plasma exchanges, 3 1 each, then remained stable between 1/1280 and 1/2560 through the latest control, 7 months later, without further treatment. In the same patient serum IgG levels decreased from 1.7 to 0.3 g/dl after plasma exchange but progressively returned to pretreatment values within 6 months (1.8 g/dl). In the other 2 patients (nos. 4 and 5) a 20-30% decrease in Ig levels was observed during follow-up whereas anti-GM1 levels decreased 4-fold.
Discussion
We found high titers of anti-GM1 antibodies in 5 patients with GBS, two of whom developed symptoms after therapeutical injection of gangliosides. In all but one patient these antibodies were IgG and they all bound to asialo-GM1. In 3 patients these antibodies also bound to GDlb, indicating that they probably reacted with the GaI(/31-3)GalNAc epitope. The clinical features of the neuropathy in these patients fulfilled the criteria for the diagnosis of GBS (Asbury and Cornblath 1990) with a rapidly progressive weakness and areflexia, followed by partial or complete recovery. Although the number of patients was too small to detect consistent differences with respect to patients without anti-GM1 antibodies, patients with high antiGM1 titers tended to have a more acute and severe disease, frequently associated with pain in affected limbs at the onset of illness. In all but one patient CSF studies supported the clinical diagnosis but in none of them electrodiagnostic findings were consistent with a demyelinating process (Cornblath 1990) and 3 had
205 prominent signs of axonal degeneration, similarly to the other reported patients with GBS associated with anti-GM1 antibodies (Yuki et al. 1990; van den Berg et al. 1991; Walsh et al. 1991). Signs of axonal degeneration were not restricted to patients with high anti-GM1 titers, as they were also prominent in 3 patients without anti-GM1 antibodies, one of whom had all motor nerves electrically unexcitable, similarly to the patients described by Feasby and coll. (1986) under the name of "acute axonal form of GBS". However, since electrodiagnostic studies were not thoroughly performed in our patients it is not possible to establish whether the abnormalities reflected a primary axonal damage or were secondary to demyelination (Van der Mech~ et al. 1991), as in the case of experimental inflammatory polyneuropathy (Hartung et al. 1988). The pathogenetic significance of anti-GM1 antibodies is still unclear. High titers of anti-GM1 IgM antibodies have been associated with motor neuron syndromes (llyas et al. 1988; Latov et ai. 1988; Nardelli et al. 1988; Pestronk et al. 1988, 1990; Nobile-Orazio et al. 1990a,b; Sadiq et al. 1990; Shy et al. 1990), or, less frequently, with chronic sensorymotor neuropathy (Sadiq et al. 1990), and in some patients reduction of antibody levels was associated with clinical improvement supporting a pathogenetic role for these antibodies. On the other side there arestudies showing that anti-GM1 antibodies, mostly IgG, are increased in 1530% of patients with GBS (llyas et al. 1990; Oomes et al. 1990; Yuki et al. 1990; van den Berg et al. 1991; Walsh et al. 1991) (21% in our series), and tend to decrease after the acute phase of the disease. Although the significance of this reactivity in GBS is not known, it is unlikely that these antibodies merely represent a secondary response to neural damage since, at least in the few patients tested within a week from the onset of symptoms, antibody levels were most elevated very early in the disease (Yuki et al. 1990). Whether, bowever, this reactivity may contribute to nerve damage in these patients remains unclear as well as whether GM1 or other crossreactive glycoconjugates in nerve (Thomas et al. 1989) were the target for this antibody response. The prevalent localization of GM1 in peripheral neuronal membrane (Hansson et al. 1977; Wiegandt 1985) as compared to myelin, where the major ganglioside is LM1 (Li et al. 1973; Fong et al. 1976) might however explain the frequent association of these antibodies with a prominent axonal damage. A significant association between anti-GM1 antibodies in patients with GBS and Campylobacter infection has been recently reported (Oomes et al. 1990; Yuki et al. 1990; Walsh et al. 1991) and the possibility that this infection might trigger an immune response to GM1 has been raised (Yuki et al. 1990; Walsh et al. 1991). None of our patients with high anti-GM1 titers reported a recent diarrheal illness, but 3 of them, corre-
sponding to those reacting with the GaI(/31-3)GalNAc epitope, had increased levels of anti-C, jejuni antibodies early in the disease, suggesting a recent C. jejuni infection. However, since stool cultures were not available to confirm C. jejuni infection and a cross-reactivity between anti-C, jejuni and anti-GM1 antibodies was confirmed in our patients (Oomes et al. 1991), as suggested by the presence of the GaI(/31-3)GalNAc epitope on C. je]uni (O'Sullivan et al. 1990), it is difficult to determine whether these antibodies were originally directed against C. jejuni, GM1, GDlb or other cross-reacting antigens. This is particularly intriguing in the 2 patients who developed GBS 10-11 days after starting treatment with gangliosides. Gangliosides have been reported not to be immunogenic in humans (Svennerholm and Fredman 1990), and the increasing use of gangliosides in our country from 1985 through 1990 was not associated with an increasing incidence of GBS (Italian Consiglio Superiore di SanitY, unpublished data). However, a delayed type of hyperu sensitivity to gangliosides has been reported in Lewis rat after immunization with spinal cord (Offner et al. 1985) and in a patient with multiple sclerosis who developed an axonal form of GBS after therapeutic swine brain implantation (Knorr-Held et al. 1986), while a case of GBS associated with high anti-GM1 titers following treatment with gangliosides has been reported (Latov et al. 1991). Since an immune response to GM1 may be found in normal individuals (Latov et al. 1988; Nobile-Orazio et al 1990b; Pestronk et al. 1990; Sadiq et al. 1990), it can not be excluded that in some sensitized subjects exposure to one or more of the above mentioned antigens may occasionally trigger an antibody response to GM1. However, whether this may lead to an acute neuropathy presenting as GBS remains unclear. Acknowledgments This work was made possible by the financial support of"Associazione Amici Centro Dino Ferrari" and by a grant from "Associazione ltaliana Sclerosi Multipla". We would like to thank Prof. S. Sonnino and Dr. V. Chigorno (Dept. of Medical Chemistry and Biochemistry, University of Milan) for providing purified gangliosides and for absorbtion studies with GM1, and Dr. A. Stangalini (Laboratorio Fleming, Milan) for measuring antiCampylobacterjejuni antibodies. Our Colleagues at the Institute of Clinical Neurology of the University of Milan are acknowledged for helping in collecting th~ sera and for providing clinical informations on the patients.
References Asbury, A.IL and D.R. Comblath (1990) Assessment of current diagnostic criteria for Guillain-Barr~ syndrome. Ann. Neurol., 27:
(Suppl.)S17-s20. Cornblath, D.R. (1990) Electrophysiology in Guillain-Barr~ Syndrome. Ann. Neurol., 27 (Suppl.): S17-$20. Dyck, P.J, J. Karnes, A. Lais, E.P. Lofgren and J.K. Stevens (1984) Pathologic alteration of the peripheral nervous system of humans. In: P.J. Dyck, P.K. Thomas, E.H. Lambert and R. Bunge (Eds.),
206 Peripheral Neuropathy, 2nd edn., WB Saunders, Philadelphia, PA, pp. 760-870. Feasby, T.E., J.J. Gilbert, W.F. Brown, C.F. Bolton, A.F. Hahn, W.F. Koopman and D.W. Zochodne (1986) An acute axonal form of Guillain-Barr6 polyneuropathy. Brain, 109: 1115-1126. Fong, J.W., R.W. Ledeen, S.K. Kundu and S. Brostoff (1976) Gangliosides of peripheral nerve myelin. J. Net.,.~chem., 26: 157-162. Hansson, H.A., J. Holmgren and L. Svennerholm (1977) Ultrastructural localization of cell membrane GM1 ganglioside by cholera toxin. Proc. Natl. Acad. Sci. USA, 74: 3782-3786. Hartung, H.P., K. Heininger, B. Shafer, W. Fierz and K.V. Toyka (1988) Ann. N.Y. Acad. Sci., 540: 122-161. llyas, A.A., H.J. Willison, M.C. Dalakas, J.N. Whitaker and R.H. Ouarles (1988) Identification and characterization of gangliosides reacting with IgM paraproteins in three patients with neuropathy associated with biclonal gammopathy. J. Neurochem., 51: 851858. llyas, A.A,. F.A. Mithen, Z.W. Chen and S.D. Cook (1990) Antibodies to neural glycolipids in sera of patients with Guillain-Barr~ syndrome. Peripheral Neuropathy Association meeting on Dysimmune neuropathies, Oxford, p. 80. Kasai, N. and R.K. Yu (1983) The monoclonai antibody A2B5 is specific to ganglioside GO1C. Brain. Res., 277: 155-158. Knorr-Held, S., W. Brendel, H. Kiefer, G. Paal and B.U. von Specht (1986) Sensitization against brain gangliosides after therapeutic swine brain implantation in a multiple sclerosis patient. J. Neurol., 233: 54-56. Latov, N., Hays, A.P., P.D. Donofrio, J. Liao, H. Ito, S. McGinnis, K. Manoussos, L. Freddo, M.E. Shy, W.H. Sherman, H.W. Chang, H.S. Greenberg, J.W. Albers, A.G. Alessi, D. Keren, R.K. Yu, L.P. Rowland, E.A. Kabat (1988) Monoclonal IgM with a unique specificity to gangliosides GM1 and GDlb and to lacto-N-tetraose associated with human motor neuron disease. Neurology, 33: 763-768, Latov, N., C.L. Koski, P.A. Walicke (1991) Guillain-Barr6 syndrome and parenteral gangliosides. Lancet, 338: 757. Li, Y.T., J.E. Mansson, M.T. Vanier and L. Svennerholm (1973) Structure of the glucose amine containing ganglioside of human tissues. J. Biol. Chem., 248: 2634-2636. Nardelli, E,, A.J. Steck, T. Barkas, M. Schluep and F. Jerusalem (1988) Motor neuron Syndrome and monoclonal IgM with antibody activity against gangliosides GM1 and GDIb. Ann. Neurol., 23: 524-528. Nobile-Orazio, E., P. Marmiroli, L. Baldini, G. Spagnol, S. Barbieri, M. Moggio, N. Polli, E. Polli and G. Scarlato (1987) Peripheral neuropathy in macroglobulinemia: incidence and antigen specificity of M-proteins. Neurology, 37: 1506-1514. Nobile-Orazio, E., G. Legname, R. Daverio, M. Carpo, A. Giuliani, S. Sonnino and G. Scarlato (1990a) Motor neuron disease in a patient with a monoclonal lgMk directed against GM1, GDlb and high-molecular-weight neural-sp~cific glycoproteins. Ann. Neurol., 28: 190-194. Nobile-Orazio, E., M. Carpo, G. Legname, N. Meucci, S. Sonnino and G. Scarlato (1990b) Anti-GM1 IgM antibodies in motor neuron disease and neuropathy. Neurology, 40: 1747-1750. Offner, H., B.A. Standage, D.R. Burger and A.A. Vandenbark (1985) Delayed type hypersensitivity to gangliosides in the Lewis rat. J. Neuroimmunol. 9: 145-157. Oomes, P.G., F.G.A. van der Mech,, K.V. Toyka and R.P. Kieyweg (1990) Antibodies to 8anglioside GM1 in GuiUain-Barr~ Syndrome (Abstract). Peripheral Neuropathy Association meeting on Dysimmune neuropathy, Oxford, p. 79.
Oomes, P.G., F.G.A. van der Mech,, B.C. Jacobs, M.P. Hazenberg and J.R.J. Banffer (1991) Antibodies to the ganglioside GM1 in sera of Guillain-Barr~ patients recognize epitopes on CampyIobacter bacteria. (Abstract) Peripheral Nerve Study Group. 10th Biennial Meeting, Arden House, New York. O'Sullivan, N., J. Benjamin, M.B. Skirrow (1990). Lectin typing of campylobacter isolates. J. Clin. Pathol., 43: 856-859. Pestronk, A., D.R. Cornblath, A.A. llyas, H. Baba, R.H Quarles, J.W. Griffin, K. Alderson and R.N. Adams (1988) A treatable multifocal neuropathy with antibodies to GM1 ganglioside. Ann. Neurol., 24: 73-78. Pestronk, A., V. Chaudhry, E.L. Feldman, J.W. Griffin, D.R. Cornblath, E.H. Denys, M. Glasberg, R.W. Kuncl, R.K. Olney and W.C. Yee (1990) Lower motor Neuron Syndromes defined by patterns of weakness, nerve conduction abnormalities, and high titers of antiglycolipid antibodies. Ann. Neurol., 27: 316-326. Sadiq, S.A., F.P. Thomas, K. Kilidireas, S. Protopsaltis, A.P. Hays, K.W. Lee, S.N. Romas, N. Kumar, L. van den Berg, M. Santoro, D.J. Lange, D.S. Younger, R.E. Lovelace, W. Trojaburg, W.H. Sherman, J.R. Sherman, J.R. Miller, J. Minuk, M.A. Fehr, R.I. Roelofs, D. Hollander, F.T. Nichols, III, H. Mitsumoto, J.J. Kelley, Jr., T.R. Swift, T.L. Munsat and N. Latov (1990) The spectrum of neurological disease associated with anti-GM1 antibodies. Neurology, 40: 1067-1072. Shy, M.E., T. Heiman-Patterson, G.J. Parry, AJ. Tahmoush, V.A. Evans and P.K. Schick (1990) Motor neuron disease in a patient with autoantibodies against Gai(/31-3)GalNAc in gangliosides GM1 and GDlb: improvement following immunotherapy. Neurology, 40: 842-844. Sonnino, S., V. Chigorno, D. Acquotti, M. Pitto, G. Kirschner and G. Tettamanti (1989) A photoreactive derivative of radiolabeled GM1 ganglioside: preparation and use to establish the involvement of specific proteins in GM1 uptake by human fetal fibroblasts in culture. Biochemistry, 28: 77-84. Svennerholm, L. and P. Fredman (1990) Antibody detection in Guillain.Barr6 Syndrome. Ann. Neurol., 27: (Suppl.) $36-40. Takatsu, M., A.P. Hays, N. Latov, G.M. Abrams, W.H. Sherman, R. Nemni, E. Nobile-Orazio, T. Saito and L. Freddo (1985) Immunofiuorescence study of patients with neuropathy and IgM M-proteins. Ann. Neurol., 18: 173-181. Thomas, F.P., A.M. Lee, S.N. Romas and N. Latov (1989) Mono. clonal IgMs with anti-Gal(/31-3)GalNAc activity in lower motor neuron disease; identification of glycoprotein antigens in neural tissue and cross-reactivity with serum immunoglobulins. J. Neuroimmunol., 23: 167-174. Van den Berg, L.H., J. Marrink, E.J. de Jager, N. Latov and S.A. Sadiq (1991) Guillain-Barr~ Syndrome with axonal degeneration and anti-GM1 antibodies (Abstract). Ann. Neurol., 30: 305. Van der Mech,, F.G.A., J. Meulstee and R.P. Kleyweg (1991) Axonal damage in Guillain-Ba~r~ syndrome. Muscle Nerve, 4: 997-1002. Waish, F.S., M. Cronin, S. Koblar, P. Doherty, J. Winer, A. Leon and R.A.C. Hughes (1991) Association between glycoconjugate antibodies and Campylobacter infection in patients with Guillain-Barr6 syndrome. J. Neuroimmunol., 34: 43-51. Wiegandt, H. (1985) Gangliosides. In: Wiegandt, H. (Ed.), Glycolipids, Elsevier, Amsterdam, pp. 199-260. Yuki, N., H. Yoshino, S. Sato, T. Miyatake (1990) Acute axonal polyneuropathy associated with anti-GM1 antibodies following Campylobacter enteritis. Neurology, 40: 1900-1902.
200
JNS 03751
Guillain-Barr6 syndrome associated with high titers of anti-GM1 antibodies Eduardo Nobile-Orazio a, Marineila Carpo a, Nicoletta Meucci a, Maria Pia Grassi b, Erminio Capitani c, Monica Sciacco a, Alfonso Mangoni b and Guglielmo Scarlato a a Institute of Clinical Neurology, Centro Dino Ferrari, Ospedale Maggiore Policlinico, and Departments of Neurology, b Ospedale L. Sacco, and c Ospedale San Paolo, University of Milan, Milan, Italy
(Received 10 September, 1991) (Revised, received21 November,1991) (Accepted 24 November,1991) Key words: GM1; Gangliosides; Guillain-Barr6 syndrome; Autoimmunity; Antibodies; IgG
Summary We found high titers of anti-GM1 antibodies (1/1280 or more) in 3 of 14 consecutive patients (21%) with Guillain-Barr6 syndrome (GBS) and in 2 additional patients who developed GBS, 10-11 days after starting parenteral treatment with gangliosides. Antibodies were IgG in 4 patients and lgM in one, and they all bound to asialo-GM1, and, in 3, to GDlb as well. Although the clinical features in all the patients with high anti-GM1 titers fulfilled the criteria for the diagnosis of GBS and in 4 of them, proteins but not cells were elevated in cerebrospinal fluid, electrodiagnostic studies in 3 patients showed prominent signs of axonal degeneration, that in one case were confirmed by morphological studies on sural nerve biopsy. No recent antecedent infection was reported by these patients, but in 3, including patients treated with gangliosides, anti-Campylobacter jejuni antibodies were elevated. In 3 patients a consistent decrease in anti-GM1 levels was observed after the acute phase of the disease suggesting that the frequent occurrence of these antibodies in patients with GBS and their frequent association with a prominent axonal impairment may have pathogenetic relevance.
Introduction
High titers of anti-GM1 lgM antibodies have been reported in some patients with motor neuron disease (MND), motor neuropathy or, less frequently, with sensorimotor neuropathy (llyas et al. 1988; Latov et al. 1988; Nardelli et al. 1988; Pestronk et al. 1988, 1990; Nobile-Orazio et ai. 1990a,b; Sadiq et al. 1990; Shy et al. 1990). In some of these patients therapeutic reduction of antibody levels was associated with clinical improvement supporting a pathogenetic role for these antibodies (Latov et al. 1988; Pestronk et al. 1988; Shy et al. 1990). Recently, high titers of anti-GM1 antibodies, mostly IgG, have been detected in some patients with the Guillain-Barr6 syndrome (GBS) (Ilyas et al. 1990; Oomes et al. 1990; Yuki et al. 1990; van den Berg
Correspondence to: E. Nobile-Orazio, M.D., Institute of Clinical Neurology,Universityof Milan, Via F. Sforza 35, 20122Milan, Italy. Tel.: 39/2/55184306; Fax.: 39/2/55190392.
et al. 1991; Walsh et al. 1991) and an association between these antibodies and Campylobacter infection has been reported (Oomes et al. 1990; Yuki et al. 1990; Walsh et al. 1991). We found high titers of anti-GM1 antibodies in 5 patients with GBS, 2 of whom developed symptoms after therapeutic injection of gangliosides. The clinical, electrophysiological and laboratory findings in these patients are reported.
Material and methods
Patients
Serum was obtained from 226 patients including 16 patients with a clinical diagnosis of GBS (Asbury and Cornblath 1990), 52 patients with neuropathy (PN) whose etiology was known in 25 (toxic in 5, vitamin B n deficiency in 1, diabetes in 2, inherited in 7, crioglobulinemic in 2, carcinomatous in 1, associated with Sj6gren's syndrome in 1, with benign lgG monoclonal
201 gammopathy in 6), unknown in 16, while 9 had chronic inflammatory demyelinating polyneuropathy (CIDP) and 2 had motor neuropathy with multifocal conduction blocks. Fifty patients had motor neuron disease (MND), 25 multiple sclerosis (MS), 23 other neurological diseases (OND) while 46 had other immune diseases (OID) (2 polymyositis; 4 myastenia gravis; 16 Sj6gren's syndrome and 24 SLE), and 14 were normal subjects. Samples were kept frozen at -80°C until use. Sera from GBS patients were obtained within 1 month from onset of symptoms and were in all but 2 cases from a consecutive series of GBS patients admitted to the Institute of Clinical Neurology of the University of Milan. None of these patients ever received gangliosides before onset of the disease. The two remaining sera were from patients admitted elsewhere who developed symptoms of GBS 10 and 11 days, respectively, after starting treatment with a mixture of commercially available bovine brain gangliosides containing GM1, GDla, G D l b and GTlb in approximately the same proportion (Cronassial, Fidia, Italy), at a dose of 20 mg i.m. on alternate days for 8 days (no. 3 in the tables) and 20 mg i.m. daily for 10 days (no. 4). This therapy was prescribed for low back pain in one (no. 3) and for a recurring ankle pain in one (no. 4). The latter patient also had diabetes but denied having antecedent neuropathy symptoms.
ELISA for anti-GM1 antibodies All sera were tested by ELISA for lgG and IgM reactivity with GM1 (Fidia, Italy) following a procedure reported previously (Nobile-Orazio et al. 1990b). Briefly, 1 /~g of purified ganglioside in 100 /tl of ethanol was added to each well of 96-well polystyrene microtitration plate (CeU-Cult, England) and ethanol was evaporated. Unreactive sites were saturated with 1% bovine serum albumin (BSA) in phoshate-buffered saline (PBS), pH 7.4 (saturating solution). After 3 washes with washing/antibody solution (saturating solution with 0.05% Tween 20), patient serum was added in triplicate to the wells at 1:20 dilution and incubated for 16 h at 4°C. After 6 washes, peroxidase-conjugated goat anti-human IgG or IgM (Cappel, Pa) was added to the wells for 1 h at room temperature, followed, after 6 washes, by 100/~1 substrate solution (0.04% o-phenylenediamine, 0.1 M citrate, 0.2 M Na2HPO4, pH 5 with 0.012% H202). The color reaction was stopped after 60 min with 50/tl/well 0.1 M H2SO 4 and the absorbance was measured at 490 rim. Results are expressed as the mean absorbance obtained from the wells coated with GM1 subtracted of absorbances obtained from wells coated with BSA only. Sera were considered positive when this difference exceeded the value of 0.05. Positive sera were titrated by serial 4-fold dilution until negative. To compare data from different experiments, serum from a patient with high antibody
titers was used as positive control in each assay. Sera with high anti-GM1 titers were also tested by ELISA for antibody reactivity with purified GDla, GDlb and asialo-GM1, all kindly provided by Prof. S. Sonnino.
lmmunostaining after high-performance thin-layer chromatography (HPTLC) In patients with high antibody titers by ELISA, reactivity with gangliosides was confirmed by immunostaining after HPTLC (Kasai and Yu 1983; NobileOrazio et al. 1987) using a mixture of bovine brain gangliosides containing GM1, GDla, GDlb and GTlb (Cronassial, Fidia, Italy) or asialo-GM1. Either 0.5 /Lg/lane of purified asialo-GM1 or 2/~g of Cronassial were separated by HPTLC on aluminium-backed silica gel 60 HPTLC plates (Merck, Germany) and the sera were tested at 1 : 100 dilution and counterstained using peroxidase-conjugated rabbit immunoglobulins to human IgG or IgM (Dakopatts, Denmark).
Immunoabsorption studies Sera from GBS patients reacting with both GM1 and Campylobacter jejuni (see later) were absorbed with GMI using an affinity column prepared by conjugation to Sepharose 4B (Pharmacia, Lurid, Sweden) of a GM1 derivative containing 12-aminododecanoic acid as acyi moiety (Sonnino et al. 1989). Absorbed sera were tested for reactivity with GM1 by ELISA and with C. jejuni by complement fixation assay. Sera from these patients were also tested for anti-GM1 reactivity by ELISA after incubation for 12 h at 4°C with an excess of the C. jejuni lysate used in the complement fixation assay (Virion, Zurich, Switzerland).
Results
We found high levels of anti-GM1 antibodies (1/1280 or more) in 5 patients with GBS, including 3 of the 14 consecutive GBS patients admitted to our institute (21%) and the 2 patients who developed GBS after treatment with gangliosides. In four of these patients antibodies were IgG (Fig. 1), while one had high levels of anti-GM1 IgM (1/2560) (open symbols in Fig. 1). None of the controls had high anti-GM1 IgG titers while one previously reported patient with MND (Nobile-Orazio et al. 1990a) had anti-GM1 IgM titers of 1/20480. Low anti-GM1 IgG levels (up to 1/320) were found in 4 additional patients with GBS (29%) but also in 7 patients with other PN (13%), 6 with MND (12%), 5 with OND (22%) (1 epilepsy, 1 dementia, 1 cerebellar degeneration, 1 stroke and 1 myopathy), and 13 with OID (28%) (Fig. 1), while the frequency of low anti-GM1 IgM in these groups ranged from 5 to 20% (Nobile-Orazio et al. 1990b).
202 TABLE 1 CLINICAL FEATURES IN PATIENTS WITH GBS AND HIGH TITERS OF ANTI-GM1 ANTIBODIES Patient No.
Presenting symptoms
1
LL weakness UL&LL pain LL weakness UL weakness and tingling LL pain and weakness UL&LL pain and weakness
2 3 4 5
Days of progression
Sensory loss
Cranial nerve impairment
Deep tendon reflexes
Autonomic signs
Maximal impairment
Impairment a at last control (time)
7
no
no
absent
no
2
0 (3 yrs)
12
no
no
absent
yes b
4
2 (6 yrs)
7
no
absent
no
4
2 (7 mos)
10
no
R& L VII no
absent
yes c
3
1 (6 mos)
12
yes
absent
no
1
0 (1 yr)
R&L VII & XII
UL = upper limbs; LL = lower limbs; R = right; L = left. a Impairment: 0 = walks normally; 1 = walks autonomously but with disturbances; 2 = needs help/support to walk; 3 = unable to walk; 4 = requires ventilatory assistance; 5 = dead. b Urinary and fecal incontinence; orthostatic hypotension. c Synus tachicardia.
In patients with GBS and high levels of anti-GM1 antibodies reactivity with GM1 was confirmed by immunostaining after HPTLC that also showed binding of Ig with the same isotype of anti-GM1 lg, to asialoGM1 and, in 3 patients, to GDlb (nos. 3, 4 and 5), but not to GDla or GTlb. Reactivity with asialo-GM1 and GDlb by ELISA in positive patients had approximately the same intensity than reactivity with GM1. The clinical features of the neuropathy in the 5 patients with GBS and high anti-GM1 titers are summarized in Table 1. All patients fulfilled the clinical criteria for the diagnosis of GBS (Asbury and Corn1/81920
•
~ 1/20480' E
1/5120 -
~ 1/12Bo-
$ ~
t~
1/320-
1/so1/20-
TABLE 2 COMPARISON OF SOME CLINICAL FEATURES IN PATIENTS WlTH GBS WITH AND WITHOUT HIGH ANT1-GM1 TITERS
Mean age (yrs) (range) Sex Days of progression (range) Limb pain at onset Requiring ventilatory assistance Sensory loss Cranial nerve impairment Autonomic impairment Mean maximal impairment a Complete recovery at 6 months
High titers (No. 5)
No high titers (No. 11)
48.4 (41-64) 2M/3F 9.4 (7-12) 3/5 2/5 1/5 2/5
40.3 (16-74) 6M/5F 17.7 (5-30) 2/11 3/11 (1 dead) 4/11 4/11
2/5
3/11
2.8
2.2
2/5
6/11
a See impairment scale in Table 1. n
am
~
GBS Neuropathy MND MS OND PATIENTS'GROUPS
OJD
Nor'real subjects
Fig. 1. Anti-GM1 lgG antibody titers by ELISA in patients with Guillain-Barr~ syndrome (GBS), neuropathy, motor neuron disease (MND), multiple sclerosis (MS), other neurological diseases (OND), other immunological diseases (OID), and in normal subjects. High titers (1/1280 or more) of anti-GM1 IgG antibodies were detected in 4 patients with GBS. One additional patient with GBS and a previously reported patient with MND (Nobile-Orazio et al. 1990a), both with low anti-GMl IgG titers (open dots), had high titers of anti-GM1 IgM antibodies (1/2560 and 1/20480 respectively).
blath 1990) with a rapidly progressive weakness and areflexia reaching the peak within 2 weeks from onset, followed by a more or less complete recovery. Although a mild sensory impairment was observed in only 1 patient (no. 5), pain in the lower limbs was an early and persistent symptom in 3 of them. When the clinical features of patients with and without anti-GM1 antibodies were compared (Table 2), the mean time to reach maximal disability was shorter in patients with high titers and their mean functional impairment was slightly more severe than in those without anti-GM1.
203 TABLE 3 ELECTROPHYSIOLOGIC FINDINGS IN PATIENTS WITH GBS AND HIGH TITERS OF ANTI-GMI ANTIBODIES Patient No.
Days of illness
Nerve conduction studies Median
1 2 3 4 5
5 13 22 3 42 12 22
EMG
Motor nerves
Sensory nerves Ulnar
Peroneal
amp
CV
DL
amp
CV
DL
amp
CV
8.0 5.5 NR nd nd nd 7.0
50.0 50.0
3.9 3.4
3.7 nd nd 3.0 NR 1.5 nd
51.0
2.4
3.7 4.0 0.5 1.3 NR 0.4 4.0
5 3 . 0 3.5 5 0 . 0 4.4 38.8 35.8 4.4
5.6
49.0
2.0
44.0
Normal limits: upper lower 4.0
47,0
4.7
62.1 4.0 33.8
4.4 49.0
2.6
Ulnar
amp
DL
amp
nd 17 10 11 9 3 6 . 6 6.4 b 10 4 2 . 7 5.3 8
2.5 2.7 2.6 2.4 4.2 3.6
nd nd 12 nd nd nd 7
6.5
3.4
3.5
DL
Median
8
Surai DL
2.1
2.9
amp
DL
nd nd nd nd nd 5 5
5.0 4.4
3.1 6
Fibrillation MUAPs (0--, + + + + ) (neurogenic) changes) a 0 0 +++ nd ++ 0/+ ++
+ + ++++ ++++ +++ ++
4.1 6
amp: amplitude (mY or/zV); CV: conduction velocity (m/sec); DL: distal latency (msec); EMG: electromyogram. a Increased MUAPs duration and reduced recruitment (graded 0 to + + + + ). nd: not done; NR: no response. b F-wave absent.
The electrophysiological findings in patients with GBS and high anti-GM1 titers are summarized in Table 3. In none of these patients the proposed electrodiagnostic criteria for demyelination (Cornblath 1990) were fulfilled, while in 3 (nos. 2, 3 and 4) there were prominent signs of axonal degeneration. In one of these patients (no. 4) sural nerve biopsy was performed (Nobile-Orazio et al. 1987) 29 days after onset of symptoms, showing a reduced number of myelinated
fibers (3210/mm 2) on semithin sections with several thinly myelinated axons and occasional clusters of regenerating fibers. No inflammatory cell infiltrate or blood vessel abnormality was detected and no immunoglobulin deposit was found by direct immunofluorescence (Takatsu et al. 1985). By teased fiber studies (Dyck et al. 1984), 43% of fibers were normal (conditions A and B), 7% had abnormalities consistent with axonai degeneration (conditions E and H) while 43%
TABLE 4 LABORATORY FINDINGS IN PATIENTS WITH GBS AND HIGH TITERS OF ANTI-GM1 ANTIBODIES
Patient No.
Anti-GMi
1
IgG 1/1280 a IgM 1/20 IgG 1/1280 a IgM neg. IgG 1/81920 a,b IgM 1/320 IgG 1/320 IgM 1/2560 a,b IgG ]/20480
2 3 4 5
Isotype titers
Antecedent events (time before onset of GBS)
Anti-
Cells (n/ram 3)
CSF
days (phase) of disease
Days of disease
Protein (mg/di)
C. jejuni Ab titer
6 (prog)
6
35
2
NR
n.d.
9 (prog)
31
138
3
NR
1/40
3 (prog)
2
84
0
1/160
29 (star)
10
70
0
16 (stat)
16
176
4
gangliosides (10 days) gangliosides (11 days) NR
1/160 1/80
IgM 1/320 Without high anti-GM1 titers Total Day of No. Ab measurement 11
17.6 (7-28)
a Ig also binding to asialo-GM1. b Ig also binding to GDlb.
CSF Protein
CSF Cells
Antecedent events
Anti-C. jejuni titer
< 100:7 > 100:4
< 10:10 > 10:1
Infection: 6 NR: 5
< 1/80:9 n.d.: 2
204 Patient 2
1181920-
Patient 3
1120480-
111280•
10G
o
1180-
1180• 1120-
o
•
lb
~.~
2'0 3'o~ys"
i :'~ 1/20480-
-
i~IM
6 ;ears Patient 4
0
10
2'0 3'0 4;)
1/81920-
5() S0days//71~lonths Patient S
1/5120111280-
1/128011320•
1/320.
c
1/80.
1/80. 0
IoM
1/20-
1/20480-
1/5120-
1/20.
1112801/320-
1/320-
"= ,r--
IgG
1/5120-
1/5120
t
10
~"~~.~/~....
toM
1/20. 2'0 3'0
40days Iff
5months
0
10 2'0 3'0 40days ~/"~
1'year
Time from onset of symptoms
Fig. 2. Longitudinal study of anti-GM1 IgG and IgM antibody titers during the disease in four patients with GBS. Arrow heads indicate the days when plasmapheresis (3 liters each) was performed in patients nos. 3 and 4.
showed multiple and 7% isolated segments of demyelination and remyelination (conditions C and D), indicating that demyelination at this level was probably secondary to axonal damage (Dyck et al. 1984). Prominent electrophysiological signs of axonal degeneration were also found early in the course of the disease, in 3 patients with GBS not associated with high levels of anti-GM1 antibodies. Some of the laboratory findings in patients with high anti-GM1 titers are summarized in Table 4. All but one patient (no. 1) had increased cerebrospinal fluid (CSF) proteins with normal CSF cell count at some point of the disease. Although none of these patients reported an antecedent history of febrile or diarrheal illness within a month before onset of symptoms, 3 of them, including the 2 patients injected with gangliosides, had increased levels of anti-Campylobacter jejuni antibodies by complement fixation assay (normal up to 1/40). A 4-fold decrease in anti-C, jejuni titers was observed during recovery in two of them (nos. 3 an 4) while in one (no. 5) antibodies decreased by 2-fold. In all these patients anti-C jejuni antibodies were of the same lg isotype than anti-GM1 antibodies and both anti-C jejuni and anti-GM1 activity were removed by previous absorption of patients' sera with GM1 by affinity chromatography. Similarly, antibody reactivity with GM1 was markedly decreased in the sera of these patients after preincubation with C. jejuni lysate. None of the patients without anti-GM1 antibodies had increased levels of anti-C jejuni antibodies. In Fig. 2 the temporal prophile of anti-GM1 antibodies in 4 patients during the disease is represented while no follow up data is available in patient no. 1. Although antibody titers were still elevated in all patients at the time of last follow up, a consistent decrease in antibody levels was observed in three of them
after the acute phase. This was most evident in patient no. 3, in whom antibody titers were extremely high 3 days after onset of symtoms, and rapidly decreased after 6 plasma exchanges, 3 1 each, then remained stable between 1/1280 and 1/2560 through the latest control, 7 months later, without further treatment. In the same patient serum IgG levels decreased from 1.7 to 0.3 g/dl after plasma exchange but progressively returned to pretreatment values within 6 months (1.8 g/dl). In the other 2 patients (nos. 4 and 5) a 20-30% decrease in Ig levels was observed during follow-up whereas anti-GM1 levels decreased 4-fold.
Discussion
We found high titers of anti-GM1 antibodies in 5 patients with GBS, two of whom developed symptoms after therapeutical injection of gangliosides. In all but one patient these antibodies were IgG and they all bound to asialo-GM1. In 3 patients these antibodies also bound to GDlb, indicating that they probably reacted with the GaI(/31-3)GalNAc epitope. The clinical features of the neuropathy in these patients fulfilled the criteria for the diagnosis of GBS (Asbury and Cornblath 1990) with a rapidly progressive weakness and areflexia, followed by partial or complete recovery. Although the number of patients was too small to detect consistent differences with respect to patients without anti-GM1 antibodies, patients with high antiGM1 titers tended to have a more acute and severe disease, frequently associated with pain in affected limbs at the onset of illness. In all but one patient CSF studies supported the clinical diagnosis but in none of them electrodiagnostic findings were consistent with a demyelinating process (Cornblath 1990) and 3 had
205 prominent signs of axonal degeneration, similarly to the other reported patients with GBS associated with anti-GM1 antibodies (Yuki et al. 1990; van den Berg et al. 1991; Walsh et al. 1991). Signs of axonal degeneration were not restricted to patients with high anti-GM1 titers, as they were also prominent in 3 patients without anti-GM1 antibodies, one of whom had all motor nerves electrically unexcitable, similarly to the patients described by Feasby and coll. (1986) under the name of "acute axonal form of GBS". However, since electrodiagnostic studies were not thoroughly performed in our patients it is not possible to establish whether the abnormalities reflected a primary axonal damage or were secondary to demyelination (Van der Mech~ et al. 1991), as in the case of experimental inflammatory polyneuropathy (Hartung et al. 1988). The pathogenetic significance of anti-GM1 antibodies is still unclear. High titers of anti-GM1 IgM antibodies have been associated with motor neuron syndromes (llyas et al. 1988; Latov et ai. 1988; Nardelli et al. 1988; Pestronk et al. 1988, 1990; Nobile-Orazio et al. 1990a,b; Sadiq et al. 1990; Shy et al. 1990), or, less frequently, with chronic sensorymotor neuropathy (Sadiq et al. 1990), and in some patients reduction of antibody levels was associated with clinical improvement supporting a pathogenetic role for these antibodies. On the other side there arestudies showing that anti-GM1 antibodies, mostly IgG, are increased in 1530% of patients with GBS (llyas et al. 1990; Oomes et al. 1990; Yuki et al. 1990; van den Berg et al. 1991; Walsh et al. 1991) (21% in our series), and tend to decrease after the acute phase of the disease. Although the significance of this reactivity in GBS is not known, it is unlikely that these antibodies merely represent a secondary response to neural damage since, at least in the few patients tested within a week from the onset of symptoms, antibody levels were most elevated very early in the disease (Yuki et al. 1990). Whether, bowever, this reactivity may contribute to nerve damage in these patients remains unclear as well as whether GM1 or other crossreactive glycoconjugates in nerve (Thomas et al. 1989) were the target for this antibody response. The prevalent localization of GM1 in peripheral neuronal membrane (Hansson et al. 1977; Wiegandt 1985) as compared to myelin, where the major ganglioside is LM1 (Li et al. 1973; Fong et al. 1976) might however explain the frequent association of these antibodies with a prominent axonal damage. A significant association between anti-GM1 antibodies in patients with GBS and Campylobacter infection has been recently reported (Oomes et al. 1990; Yuki et al. 1990; Walsh et al. 1991) and the possibility that this infection might trigger an immune response to GM1 has been raised (Yuki et al. 1990; Walsh et al. 1991). None of our patients with high anti-GM1 titers reported a recent diarrheal illness, but 3 of them, corre-
sponding to those reacting with the GaI(/31-3)GalNAc epitope, had increased levels of anti-C, jejuni antibodies early in the disease, suggesting a recent C. jejuni infection. However, since stool cultures were not available to confirm C. jejuni infection and a cross-reactivity between anti-C, jejuni and anti-GM1 antibodies was confirmed in our patients (Oomes et al. 1991), as suggested by the presence of the GaI(/31-3)GalNAc epitope on C. je]uni (O'Sullivan et al. 1990), it is difficult to determine whether these antibodies were originally directed against C. jejuni, GM1, GDlb or other cross-reacting antigens. This is particularly intriguing in the 2 patients who developed GBS 10-11 days after starting treatment with gangliosides. Gangliosides have been reported not to be immunogenic in humans (Svennerholm and Fredman 1990), and the increasing use of gangliosides in our country from 1985 through 1990 was not associated with an increasing incidence of GBS (Italian Consiglio Superiore di SanitY, unpublished data). However, a delayed type of hyperu sensitivity to gangliosides has been reported in Lewis rat after immunization with spinal cord (Offner et al. 1985) and in a patient with multiple sclerosis who developed an axonal form of GBS after therapeutic swine brain implantation (Knorr-Held et al. 1986), while a case of GBS associated with high anti-GM1 titers following treatment with gangliosides has been reported (Latov et al. 1991). Since an immune response to GM1 may be found in normal individuals (Latov et al. 1988; Nobile-Orazio et al 1990b; Pestronk et al. 1990; Sadiq et al. 1990), it can not be excluded that in some sensitized subjects exposure to one or more of the above mentioned antigens may occasionally trigger an antibody response to GM1. However, whether this may lead to an acute neuropathy presenting as GBS remains unclear. Acknowledgments This work was made possible by the financial support of"Associazione Amici Centro Dino Ferrari" and by a grant from "Associazione ltaliana Sclerosi Multipla". We would like to thank Prof. S. Sonnino and Dr. V. Chigorno (Dept. of Medical Chemistry and Biochemistry, University of Milan) for providing purified gangliosides and for absorbtion studies with GM1, and Dr. A. Stangalini (Laboratorio Fleming, Milan) for measuring antiCampylobacterjejuni antibodies. Our Colleagues at the Institute of Clinical Neurology of the University of Milan are acknowledged for helping in collecting th~ sera and for providing clinical informations on the patients.
References Asbury, A.IL and D.R. Comblath (1990) Assessment of current diagnostic criteria for Guillain-Barr~ syndrome. Ann. Neurol., 27:
(Suppl.)S17-s20. Cornblath, D.R. (1990) Electrophysiology in Guillain-Barr~ Syndrome. Ann. Neurol., 27 (Suppl.): S17-$20. Dyck, P.J, J. Karnes, A. Lais, E.P. Lofgren and J.K. Stevens (1984) Pathologic alteration of the peripheral nervous system of humans. In: P.J. Dyck, P.K. Thomas, E.H. Lambert and R. Bunge (Eds.),
206 Peripheral Neuropathy, 2nd edn., WB Saunders, Philadelphia, PA, pp. 760-870. Feasby, T.E., J.J. Gilbert, W.F. Brown, C.F. Bolton, A.F. Hahn, W.F. Koopman and D.W. Zochodne (1986) An acute axonal form of Guillain-Barr6 polyneuropathy. Brain, 109: 1115-1126. Fong, J.W., R.W. Ledeen, S.K. Kundu and S. Brostoff (1976) Gangliosides of peripheral nerve myelin. J. Net.,.~chem., 26: 157-162. Hansson, H.A., J. Holmgren and L. Svennerholm (1977) Ultrastructural localization of cell membrane GM1 ganglioside by cholera toxin. Proc. Natl. Acad. Sci. USA, 74: 3782-3786. Hartung, H.P., K. Heininger, B. Shafer, W. Fierz and K.V. Toyka (1988) Ann. N.Y. Acad. Sci., 540: 122-161. llyas, A.A., H.J. Willison, M.C. Dalakas, J.N. Whitaker and R.H. Ouarles (1988) Identification and characterization of gangliosides reacting with IgM paraproteins in three patients with neuropathy associated with biclonal gammopathy. J. Neurochem., 51: 851858. llyas, A.A,. F.A. Mithen, Z.W. Chen and S.D. Cook (1990) Antibodies to neural glycolipids in sera of patients with Guillain-Barr~ syndrome. Peripheral Neuropathy Association meeting on Dysimmune neuropathies, Oxford, p. 80. Kasai, N. and R.K. Yu (1983) The monoclonai antibody A2B5 is specific to ganglioside GO1C. Brain. Res., 277: 155-158. Knorr-Held, S., W. Brendel, H. Kiefer, G. Paal and B.U. von Specht (1986) Sensitization against brain gangliosides after therapeutic swine brain implantation in a multiple sclerosis patient. J. Neurol., 233: 54-56. Latov, N., Hays, A.P., P.D. Donofrio, J. Liao, H. Ito, S. McGinnis, K. Manoussos, L. Freddo, M.E. Shy, W.H. Sherman, H.W. Chang, H.S. Greenberg, J.W. Albers, A.G. Alessi, D. Keren, R.K. Yu, L.P. Rowland, E.A. Kabat (1988) Monoclonal IgM with a unique specificity to gangliosides GM1 and GDlb and to lacto-N-tetraose associated with human motor neuron disease. Neurology, 33: 763-768, Latov, N., C.L. Koski, P.A. Walicke (1991) Guillain-Barr6 syndrome and parenteral gangliosides. Lancet, 338: 757. Li, Y.T., J.E. Mansson, M.T. Vanier and L. Svennerholm (1973) Structure of the glucose amine containing ganglioside of human tissues. J. Biol. Chem., 248: 2634-2636. Nardelli, E,, A.J. Steck, T. Barkas, M. Schluep and F. Jerusalem (1988) Motor neuron Syndrome and monoclonal IgM with antibody activity against gangliosides GM1 and GDIb. Ann. Neurol., 23: 524-528. Nobile-Orazio, E., P. Marmiroli, L. Baldini, G. Spagnol, S. Barbieri, M. Moggio, N. Polli, E. Polli and G. Scarlato (1987) Peripheral neuropathy in macroglobulinemia: incidence and antigen specificity of M-proteins. Neurology, 37: 1506-1514. Nobile-Orazio, E., G. Legname, R. Daverio, M. Carpo, A. Giuliani, S. Sonnino and G. Scarlato (1990a) Motor neuron disease in a patient with a monoclonal lgMk directed against GM1, GDlb and high-molecular-weight neural-sp~cific glycoproteins. Ann. Neurol., 28: 190-194. Nobile-Orazio, E., M. Carpo, G. Legname, N. Meucci, S. Sonnino and G. Scarlato (1990b) Anti-GM1 IgM antibodies in motor neuron disease and neuropathy. Neurology, 40: 1747-1750. Offner, H., B.A. Standage, D.R. Burger and A.A. Vandenbark (1985) Delayed type hypersensitivity to gangliosides in the Lewis rat. J. Neuroimmunol. 9: 145-157. Oomes, P.G., F.G.A. van der Mech,, K.V. Toyka and R.P. Kieyweg (1990) Antibodies to 8anglioside GM1 in GuiUain-Barr~ Syndrome (Abstract). Peripheral Neuropathy Association meeting on Dysimmune neuropathy, Oxford, p. 79.
Oomes, P.G., F.G.A. van der Mech,, B.C. Jacobs, M.P. Hazenberg and J.R.J. Banffer (1991) Antibodies to the ganglioside GM1 in sera of Guillain-Barr~ patients recognize epitopes on CampyIobacter bacteria. (Abstract) Peripheral Nerve Study Group. 10th Biennial Meeting, Arden House, New York. O'Sullivan, N., J. Benjamin, M.B. Skirrow (1990). Lectin typing of campylobacter isolates. J. Clin. Pathol., 43: 856-859. Pestronk, A., D.R. Cornblath, A.A. llyas, H. Baba, R.H Quarles, J.W. Griffin, K. Alderson and R.N. Adams (1988) A treatable multifocal neuropathy with antibodies to GM1 ganglioside. Ann. Neurol., 24: 73-78. Pestronk, A., V. Chaudhry, E.L. Feldman, J.W. Griffin, D.R. Cornblath, E.H. Denys, M. Glasberg, R.W. Kuncl, R.K. Olney and W.C. Yee (1990) Lower motor Neuron Syndromes defined by patterns of weakness, nerve conduction abnormalities, and high titers of antiglycolipid antibodies. Ann. Neurol., 27: 316-326. Sadiq, S.A., F.P. Thomas, K. Kilidireas, S. Protopsaltis, A.P. Hays, K.W. Lee, S.N. Romas, N. Kumar, L. van den Berg, M. Santoro, D.J. Lange, D.S. Younger, R.E. Lovelace, W. Trojaburg, W.H. Sherman, J.R. Sherman, J.R. Miller, J. Minuk, M.A. Fehr, R.I. Roelofs, D. Hollander, F.T. Nichols, III, H. Mitsumoto, J.J. Kelley, Jr., T.R. Swift, T.L. Munsat and N. Latov (1990) The spectrum of neurological disease associated with anti-GM1 antibodies. Neurology, 40: 1067-1072. Shy, M.E., T. Heiman-Patterson, G.J. Parry, AJ. Tahmoush, V.A. Evans and P.K. Schick (1990) Motor neuron disease in a patient with autoantibodies against Gai(/31-3)GalNAc in gangliosides GM1 and GDlb: improvement following immunotherapy. Neurology, 40: 842-844. Sonnino, S., V. Chigorno, D. Acquotti, M. Pitto, G. Kirschner and G. Tettamanti (1989) A photoreactive derivative of radiolabeled GM1 ganglioside: preparation and use to establish the involvement of specific proteins in GM1 uptake by human fetal fibroblasts in culture. Biochemistry, 28: 77-84. Svennerholm, L. and P. Fredman (1990) Antibody detection in Guillain.Barr6 Syndrome. Ann. Neurol., 27: (Suppl.) $36-40. Takatsu, M., A.P. Hays, N. Latov, G.M. Abrams, W.H. Sherman, R. Nemni, E. Nobile-Orazio, T. Saito and L. Freddo (1985) Immunofiuorescence study of patients with neuropathy and IgM M-proteins. Ann. Neurol., 18: 173-181. Thomas, F.P., A.M. Lee, S.N. Romas and N. Latov (1989) Mono. clonal IgMs with anti-Gal(/31-3)GalNAc activity in lower motor neuron disease; identification of glycoprotein antigens in neural tissue and cross-reactivity with serum immunoglobulins. J. Neuroimmunol., 23: 167-174. Van den Berg, L.H., J. Marrink, E.J. de Jager, N. Latov and S.A. Sadiq (1991) Guillain-Barr~ Syndrome with axonal degeneration and anti-GM1 antibodies (Abstract). Ann. Neurol., 30: 305. Van der Mech,, F.G.A., J. Meulstee and R.P. Kleyweg (1991) Axonal damage in Guillain-Ba~r~ syndrome. Muscle Nerve, 4: 997-1002. Waish, F.S., M. Cronin, S. Koblar, P. Doherty, J. Winer, A. Leon and R.A.C. Hughes (1991) Association between glycoconjugate antibodies and Campylobacter infection in patients with Guillain-Barr6 syndrome. J. Neuroimmunol., 34: 43-51. Wiegandt, H. (1985) Gangliosides. In: Wiegandt, H. (Ed.), Glycolipids, Elsevier, Amsterdam, pp. 199-260. Yuki, N., H. Yoshino, S. Sato, T. Miyatake (1990) Acute axonal polyneuropathy associated with anti-GM1 antibodies following Campylobacter enteritis. Neurology, 40: 1900-1902.
