Journal of the Neurological Sciences, 1982, 54:117-127
Elsevier Biomedical Press
AGAROSE ISOELECTRIC FOCUSING OF UNCONCENTRATED CSF AND R A D I O I M M U N O F I X A T I O N FOR DETECTION OF OLIGOCLONAL BANDS IN PATIENTS WITH MULTIPLE SCLEROSIS AND OTHER NEUROLOGICAL DISEASES
V.K. KOSTULAS and H. LINK
Department of Neurology, University Hospital, S-581 85 Link6ping (Sweden) (Received 8 September, 1981) (Accepted 14 September, 1981)
Agarose isoelectric focusing (AIEF) of concentrated CSF was compared with AIEF of unconcentrated CSF and subsequent immunofixation with radiolabeled antihuman IgG Fc fragment antiserum and autoradiography for the demonstration of oligoclonal bands in CSF from 287 neurological patients. Oligoclonal bands were demonstrated by AIEF in 98~o of 43 patients with multiple sclerosis, 72~ of 18 patients with infectious CNS diseases, and 23~ of 226 patients with other neurological diseases. The corresponding figures obtained with AIEF of unconcentrated CSF and radioimmunofixation were 98~o, 67~, and 21~o, respectively. In 15 of the patients, oligoclonal bands were demonstrated in CSF and serum by both techniques. They are both useful alternatives for the demonstration of oligoclonal bands in CSF, and the method for unconcentrated CSF can be safely applied when only small CSF volumes are available. The oligoclonal IgG pattern obtained by AIEF was not influenced by concentration of CSF by ultrafiltration and subsequent dilution to the original IgG concentration, nor by storage for 6 months.
Separation of immunoglobulin G (IgG) by electrophoresis or isoelectric focusing (IEF) for the demonstration of oligoclonal bands has become an important part of the routine examination of cerebrospinal fluid (CSF). The procedure involves
This study was supported by a grant from the Swedish Medical Research Council (Project No. 3381). 0022-510X/81/0000-0000/$02.75 © Elsevier Biomedical Press
118 the study of the patient's serum taken simultaneously. Oligoclonal IgG bands present in CSF but not in serum are considered to represent the product of a restricted number of lymphocyte clones transferred into the central nervous system (CNS) and CSF where they are transformed to antibody-producing plasma cells. Oligoclonal bands have been reported in the CSF from most patients with multiple sclerosis (MS), in a substantial proportion of patients with optic neuritis and chronic myelopathy, i.e. disorders which may be related to MS, in about a third of patients with acute aseptic meningitis, meningo-encephalitis and GuillainBarr6 syndrome (for review see Link 1978; Trotter and Brooks 1980), in neurosyphilis, tuberculous meningitis, and cryptococcal meningitis. Occasionally, oligoclonal bands in CSF may be observed in diseases which are not considered to be related to infection or immune responses, such as in cerebrovascular diseases (Rostr6m and Link 1981). Electrophoresis on agar or agarose is mostly used for the demonstration of oligoclonal bands in routine clinical work. Thin-layer polyacrylamide IEF of CSF and serum has been found to be a useful alternative (Delmotte 197l) and its high resolution power has yielded information about the complexity of the humoral immune response within the CNS in inter alia MS (Laurenzi 1981). A prerequisite for the use of any of these techniques is that the CSF is concentrated 100-200 times or, preferably, to a well-defined IgG level to enable the application of standardized amounts of IgG (Link 1973; Trotter and Brooks 1980). This means that about 10 ml of CSF are usually required, which is a limiting factor in e.g. pediatric practice. The use of commercially available agarose with extremely low, or no demonstrable electroendosmosis was described in 1979 for the separation of macromolecular plasma proteins (Ros6n et al. 1979; Saravis and Zamcheck 1979). In the present paper, a modified AIEF technique was applied for the detection of oligoclonal bands in concentrated CSF. In order to demonstrate oligoclonal IgG bands in AIEF separated unconcentrated CSF, immunofixation was carried out with radiolabeled antiserum against human IgG Fc fragment, followed by autoradiography according to a modification of the technique described by Chazot et al. (1980) and Lasne et al. (1981). The two variants of AIEF were evaluated in the examination of CSF and serum from 287 patients. Patients CSF and serum was obtained by lumbar puncture from 287 patients examined at the Department of Neurology. The diagnoses are given in Tables 1 and 2. Of the 43 patients (29 females) with clinically definite MS, 10 had a duration of disease exceeding 10 years, and 8 had moderate or severe disability, i.e. the patients were dependent on a variable degree of assistance, and their professional and social lives were influenced to a variable and often high degree. None of the patients with MS was treated with immunosuppressive drugs for the last 6 months, and the majority of patients had never received such medication. Six patients had acute unilateral optic neuritis and their age varied between 19 and 61 years (mean 40); 14 had paraesthesia and their age was between 22 and 47 years (mean
119 TABLE 1 F R E Q U E N C I E S OF O L I G O C L O N A L BANDS IN CSF D E M O N S T R A T E D BY A G A R O S E ISOELECTRIC F O C U S I N G (AIEF) OF U N C O N C E N T R A T E D CSF A N D R A D I O I M M U N O F I X A T I O N , A N D BY AIEF OF C O N C E N T R A T E D CSF F R O M 287 PATIENTS WITH N E U R O L O G I C A L DISEASES Diagnosis
AIEF and radioimmunofixation
Any of the two techniques
Multiple sclerosis (n = 43)
Infectious CNS diseases (n = 18)
Other neurological diseases (n = 226)
TABLE 2 F R E Q U E N C I E S OF O L I G O C L O N A L BANDS IN CSF D E M O N S T R A B L E BY A G A R O S E ISOELECTRIC F O C U S I N G (AIEF) OF U N C O N C E N T R A T E D CSF A N D R A D I O I M M U N O F I X A T I O N , A N D BY AIEF OF C O N C E N T R A T E D CSF F R O M 226 PATIENTS WITH N E U R O L O G I C A L DISORDERS EXCEPT MS A N D INFECTIOUS CNS DISEASES Diagnosis
Cerebral infarction (n = 37) Headache (n = 28) Transient ischemic attacks (n = 19) Vertigo (n = 18) Paraesthesia (n = 14) Polyneuropathy (n = 14) Guillain-Barr~ syndrome (n = 7) Optic neuritis (n = 6) Chronic myelopathy of unknown cause (n = 6) Psychoneurosis (n = 6) Syncope (n = 5) Diplopia (n = 4) Cerebral concussion (n = 4) Brachialgia (n = 3) Systemic lupus erythematosus with CNS complications (n = 3) Myasthenia gravis (n = 2) Horner's syndrome (n = 1) Hemiplegia (n = 1) Liver encephalopathy (n = 1) Meningeal carcinomatosis (n = I) Lymphoma with mononeuropathy (n = 1) Myeloma (n _ 1) Other, see text (n = 44) Total
AIEF and radioimmunofixation 5 1 3 0 5 1 6 6 5 I 2 1 1 2
7 2 2
Any of the two techniques 7 2 4
1 1 1 1 1 1 1 1 0
2 1 1 1 1 1 1 1
120 35), and 6 had chronic myelopathy of unknown cause and their age was between 23 and 70 years (mean 46). METHODS Routine CSF and serum studies Cell counting and differentiation into polymorphonuclear and mononuclear cells, and erythrocytes was performed by phase contrast microscopy. Determinations of IgG and albumin were carried out on unconcentrated CSF and on serum by an automatic immunoprecipitation technique utilizing nephelometric analyses of antigen-antibody complexes in a continuous flow system. The CSF IgG was presented as the CSF IgG index equal to (CSF/serum IgG) : (CSF/serum albumin) (Tibbling et al. 1977). The CSF IgG index takes into account fluctuations in concentration of serum IgG as well as the status of the blood-brain barrier; an elevated CSF IgG index indicates synthesis of IgG within the CNS. The blood-brain barrier was determined by the CSF/serum albumin ratio. The age-dependent normal values previously reported were used (Tibbling et al. 1977). After cell counting, centrifugation at 200 x g for 10 min, and quantitation of proteins, CSF was concentrated by ultrafiltration at 4°C in collodion bags (Sartorius Membranfiiter, G6ttingen, F.R.G.). To allow the application for AIEF of well-defined amounts of IgG, the IgG level was also determined in concentrated CSF and adjusted with 0.15 M NaC1 to about 4 g/1. The corresponding serum was diluted with 0.15 M NaC1 to the same IgG concentration. The specimens were stored at 4°C up to 1 week after lumbar puncture before AIEF. Agarose isoelectric jocusing AIEF was carried out according to Saravis and Zamcheck (1979), with some modifications. Thirty-three ml of agarose aliquots, prepared from 0.8~o agarose (IsoGel; Marine Colloids Division of FMC Corporation, Rockland, ME, U.S.A.) in deionized water, were brought into solution in a boiling water bath, cooled to 60°C, mixed with 5.7~ ampholyte (Ampholine, pH 3.5-10; LKB, Stockholm, Sweden) and poured on hydrophilic polyester-based plastic plates (Gelbond; Marine Colloids) 110 x 205 mm. A cooled home-made slit former yielding 15 application slits with the dimensions of 9 x 0.5 mm, was immediately placed on the gel and removed 15 min later when the gel was solidified. The formed agarose gel plates were "aged" at 4 °C in a moist chamber for 1 h. Excess liquid was then removed by overlaying the gel with a Whatman 3 MM filter paper. Five/~1 of concentrated CSF and of the corresponding diluted serum were applied side by side; CSF and serum from 6 patients were run on each plate. Pooled serum from 200 blood donors diluted 1:3 with 0.15 M NaC1, carboxyhaemoglobin, and a reference to enable determination of isoelectric points in the pH range 3-10.5 (Pharmacia, Uppsala, Sweden) were run on every plate. Electrode strips were soaked with 0.5 M acetic acid for the anode, and 0.5 M NaOH for the cathode. The strips were slightly dried between filter papers before use. Using the
121 Multiphor equipment (LKB with power supply 2105) connected with a cooling system to allow separation at 5 °C, AIEF was started at 400 V, 50 mA and 15-20 W, and after 1 h, 800-1000 V and 15-20 mA were reached, when the carboxyhaemoglobin had formed several sharp bands. The pH gradient between the anode and the cathode was then measured by a surface pH-etectrode (type 403-30 M8; Ingold AG, Ziirich, Switzerland) at the temperature of the experiment. Thereafter, AIEF was continued for an additional 10 min. The plates were then placed in a fixative solution (containing 57.5 g trichloracetic acid and 17.3 g sulphosalicylic acid/500 ml deionized water) for 10 min, washed in 95~o ethanol for 20 min, pressed under several layers of chromatography paper (Whatman 3 MM, thickness 0.38 mm) soaked in 95~ ethanol, glass plate and 1 kg weight for 20 min, dried with warm air at 40 °C, stained with 0.5~o (weight/vol.) Coomassie Brilliant Blue R-250 (Sigma, St. Louis, MO, U.S.A.) in 35~ ethanol, 10~ acetic acid and deionized water for 10 min at 60 °C, destained with 35~ ethanol, 10~o acetic acid and deionized water for 2 x 15 min, and dried.
Radiolabeling of antiserum Rabbit antihuman IgG Fc fragment antibodies (code AA 089; Dako Immunoglobulins, Copenhagen, Denmark) were iodinated with 2 mCi 125I (code IMS 30; The Radiochemical Centre, Amersham, U.K.) per mg of protein using the chloramine-T method (Garvey et al. 1977). The radiolabeled antiserum was separated from free ~25I on a 0.7 x 15 cm column (Econo-column; BIO-RAD Laboratories, Richmond, CA 94804, U.S.A.) with Sephadex G 25 (Pharmacia) presaturated with 1 ml of 2 ~ bovine serum albumin. The estimated specific activity was 0.2 mCi/mg of antibodies. The iodinated antibodies were dialyzed three times against 2 1 of 0.05 M phosphate buffer, pH 7.0, and stored at 4°C until use. Before use, the solution was diluted with 0.1 M phosphate-buffered saline, pH 7.2-7.4, to an activity of 15 x 106 cpm/ml.
Agarose isoelectric Jocusing o["unconcentrated CSF and radioimmunofixation A modified version of the technique of Lasne et al. (1981) enabling the demonstration of oligoclonal IgG bands in CSF without previous concentration of the CSF was used. Five #1 of unconcentrated CSF with an IgG concentration of 20-30 mg/l after adjustment with 0.15 M NaC1, and serum diluted with 0.15 M NaC1 to the same IgG concentration were applied side by side. Carboxyhaemoglobin and pooled donor serum were run as references on every plate. AIEF was carried out as described above. Immediately after focusing, the plates were fixed for 20 min in fixative solution (14.25 g sulphosalicylic acid and 57.5 g trichloroacetic acid in 150 ml methanol and 350 ml deionized water), and then rinsed 5 times over 24 h in 1 1 of Tris buffer (0.1 M Tris-HC1, 0.85 M NaC1, pH 7.6). For immunofixation, cellulose acetate strips (Sepraphore III; Gelman Instrument Co., Ann Arbor, MI, U.S.A.) dipped in the radiolabeled antiserum were applied to the gel surface for 20 min. The strips were then removed, and incubation continued in a moist chamber at room temperature over night. After rinsing in 2 I of 0.15 M NaCI 5 times over
122 24 h, and final rinsing in deionized water to avoid salt crystal precipitation, the gel plate was dried under warm air at 40 °C. Antigen-antibody complexes precipitated on the agarose plates were visualized by autoradiography on the dried gel plate, using X-ray film (X-omat; Kodak) for 6-12 h at room temperature. RESULTS The results obtained by AIEF and radioimmunofixation of unconcentrated CSF, and with AIEF of concentrated CSF for the demonstration of oligoclonal bands in patients with MS, infectious CNS diseases, and other neurological disorders are summarized in Table 1, and exemplified in Figs. 1A and B.
Multiple sclerosis Among the 43 patients with clinically definite MS, 42 had oligoclonal bands in CSF demorlstrable by both techniques. Oligoclonal bands were demonstrable in serum from 8 of these patients. Infectious CNS diseases Eleven patients with aseptic meningo-encephalitis were included. AIEF revealed bands in CSF in l0 of them, while AIEF and radioimmunofixation revealed oligoclonal |gG bands in CSF in 9. Of 5 patients with acute aseptic meningitis, 2 displayed oligoclonal bands in CSF with both techniques. One of 2 patients with acute bacterial meningitis had oligoclonal bands, which were also seen in the corresponding serum by both techniques. Other neurological disorders Of the 226 patients with neurological diseases except MS and infectious CNS disorders, 54 (24~,,; Table 1) had oligoclonal bands in CSF demonstrable by any of the two techniques. Subgrouping of the 54 patients with oligoclonal bands, according to diagnosis (Table 2) revealed that altogether 16 of these patients had paraesthesia, acute unilateral optic neuritis or chronic myelopathy of unknown cause, i.e. disorders that might be related to MS. Furthermore, 6 patients with oligoclonal bands in CSF had Guillain-Barr6 syndrome and 2 had systemic lupus erythematosus, i.e. disorders which can be accompanied by an inflammatory response within the CSF compartment. Thirteen of the 54 patients with oligoclonal bands in CSF displayed one or more oligoclonal bands in serum as well. The following diagnoses were encountered: Transient ischemic attacks (2 patients), cerebral infarction (1), Guillain-Barr6 syndrome (1), syncope (1), liver encephalopathy (1), Horner's syndrome (1), polyneuropathy (1), meningeal carcinomatosis secondary to pulmonary carcinoma (1), lymphoma with mononeuropathy (1), myeloma (1), vertigo (1), and psychoneurosis (1). The age among these patients varied between 21 and 81 years (mean 65 years).
Fig. 1. A: Patterns from immunofixation with radiolabeled antiserum against IgG Fc fragment, of diluted serum (S) and unconcentrated cerebrospinal fluid (CSF) separated by agarose isoelectric focusing. (1) and (2) correspond to serum and CSF from a patient with benign monoclonal gammopathy; (3) and (4) paraesthesia; (5) and (6) Guillain-Barr6 syndrome; (7) and (8) multiple sclerosis; (9) and (10) chronic aseptic meningo-encephalitis of unknown cause; (11) and (12) psychoneurosis; (13) pooled normal serum. B: Patterns from agarose isoelectric focusing of serum (S) and concentrated CSF from the same subjects as in A. (14) corresponds to the reference used for determinations of isoelectric points, and (15) to carboxyhaemoglobin.
Fig. 2. Serum (S), unconcentrated CSF (CSF A) and CSF that had been concentrated and diluted to the original IgG concentration (CSF B) from a patient with benign monoclonal gammopathy (1), (2) and (3); cerebral concussion (4), (5) and (6); and multiple sclerosis (7), (8) and (9). Normal serum (NS) was run as reference (10).
Effect of concentration procedure and storage on oligoclonal band pattern of CSF CSF from. 20 patients, 6 of them with MS, was first concentrated by ultrafiltration to about 4 g of IgG/l and then diluted to the original IgG concentration and examined by AIEF and radioimmunofixation together with the same patient's unconcentrated CSF, and serum. The concentration procedure used had no influence on the number or intensity of oligoclonal IgG bands (Fig. 2). Patterns of oligoclonal bands obtained with both variants of AIEF applied within a week after lumbar puncture and up to one year later were compared in order to analyse the effect of storage. The oligoclonal bands were not significantly influenced after 6 months, but lower in number and less distinct after longer times of storage. DISCUSSION
The principle of radioimmunofixation after AIEF as described by Chazot et al. (1980) and Lasne et al. (1981) has been adopted in the present study. These authors described the presence of a regular, "oligoclonal" IgG band pattern in the pH region 7-9 for normal CSF, and they claimed that the difference between normal and MS CSF regarding this regularly occurring oligoclonal IgG band pattern is mainly quantitative. With our modification of the procedure, we were also able to identify a pattern of IgG common to CSF and serum from patients with various neurological disorders, including patients with psychoneurosis who may be
125 considered as "normal" controls. This pattern corresponds according to our opinion to polyclonal IgG. However, bands of intrathecally produced oligoclonal IgG were found in patients with MS, infectious CNS diseases, and in a minority of patients with other neurological disorders at frequencies similar to those obtained when concentrated CSF was examined by AIEF. AIEF of unconcentrated CSF and subsequent immunofixation with radiolabeled antiserum against IgG Fc fragment and subsequent autoradiography, seems therefore to be a new tool for the identification of locally produced oligoclonal IgG in neurological disorders. The main advantages of this technique - in addition to enable the examination of CSF without previous concentration - are that (1) oligoclonal IgG bands can be separated even when only small CSF quantities are available, such as in pediatric practice, and (2) the occurrence of artefacts such as the appearance of bands consisting of proteins other than IgG is minimized. The technique can also be adapted for the analysis of the IgG band pattern in e.g. AIEF-separated proteins from directly applied frozen nervous and muscular tissue sections (Olsson and Link, manuscript in preparation), and for analysis of patterns of proteins other than that of IgG (Laurenzi and Link, manuscript in preparation). IEF on thin-layer polyacrylamide gel (PAG) of concentrated CSF has been used in recent years for the demonstration of oligoclonal bands in CSF and has been claimed to be more sensitive regarding detection of oligoclonal bands in MS CSF than electrophoresis on agar (Delmotte 1971; Delmotte and Gonsette 1977) or agarose (Laurenzi and Link 1978). The same is true for AIEF (Link and Kostulas, manuscript in preparation). This difference is less important from a diagnostic point of view. However, drawbacks are inherent to the method of agarose electrophoresis commonly used: bands containing other proteins than IgG may migrate into the region where the majority of oligoclonal IgG bands are found (Link 1973; Hershey and Trotter 1980), which might necessitate that the procedure is repeated and followed by immunofixation with monospecific antisera to demonstrate the oligoclonal IgG bands (Laurenzi 1981). Furthermore, the variability in quality of the commercial agarose (Panagel ®) plates has been pointed out by others (Hershey and Trotter 1980) and has also resulted in problems in our laboratory. Therefore, IEF will probably be more commonly used in the future for the routine demonstration of oligoclonal bands in CSF. In this context, agarose has advantages over polyacrylamide, inter alia easiness in handling, non-toxicity, and separation of high molecular weight proteins such as IgM. The frequency of oligoclonal bands in neurological diseases - excluding MS and infectious CNS diseases - is related to (1) the separation technique applied, those with higher resolution capacity such as IEF yielding higher frequencies, and (2) the proportion of patients with certain other disorders where oligoclonal bands have been reported, such as Guillain-Barr6 syndrome (Link 1975), myasthenia gravis (Adornato et al. 1978), and acute cerebrovascular diseases (Rostr6m and Link 1981). The present patient material with disorders other than MS and CNS infections, included substantial numbers of such patients. In addition, there were 6 patients with acute unilateral optic neuritis and 6 patients with chronic myelopathy
126 of unknown cause, i.e. MS-related disorders. These are the explanations of the high numbers of positive findings obtained with both techniques in the present study of patients with diseases other than MS and infectious CNS diseases. O f special interest is the demonstration of oligoclonal bands in CSF from 5 of 14 patients with paraesthesia. The follow-up of clinical and CSF variables in such patients might throw new light on the relation between early MS and CSF immunoglobulin abnormalities. It has been debated since the introduction of techniques for the demonstration of oligoclonal bands in CSF, whether the procedure used for concentration of the fluid had any influence on the oligoclonal band pattern but, hitherto, no study has been reported to elucidate this question. Our data obtained with A I E F indicate that concentration of CSF by ultrafiltration has no influence on the oligoclonal IgG band pattern. Comparisons of oligoclonal bands over the course of disease in individual patients are hampered by variations induced by the technique, yielding inter alia differences in pattern when lots of the same specimen are examined simultaneously by A I E F on different plates. Previously, the simultaneous examination by agarose electrophoresis of specimens obtained at different occasions from the same patient has revealed a band pattern which was constant in the individual patient with MS (Olsson and Link 1973). On the other hand, storage of CSF as well as repeated freezing and thawing of the fluid resulted in loss of bands when the CSF was examined by electrophoresis on commercial agarose (Panagel ®) plates (Johnson et al. 1977). The present study indicates that storage has a significant influence on the A I E F pattern of oligoclonal bands, since they were unaffected up to 6 months of storage at - 2 0 ° C but then become fainter and could disappear after one year of storage. This makes A I E F less suitable for the analysis of banding patterns over the course of disease in various long-standing inflammatory nervous system diseases. ACKNOWLEDGEMENTS We thank Birgitta FranzOn, AnneChatrine Axelsson and Annalena Bj6rklinger for skilful secreterial work. REFERENCES Adornato, B.T., S.A. Houff, M. Dalakas, D.L. Madden and J.L. Sever (1978) Abnormal immunoglobulin bands in cerebrospinal fluid in myasthenia gravis, Lancet, ii: 367-368. Chazot, G., Y. Lasne, O. Benzerara, C. Confavreux and B. Schott (1980) Radio-immunofixation .... Une nouvelle technique de caract~risation des immunoglobulines darts le liquide c6phalo-rachidien non concentr& Rev. neurol. (Paris), 136: 783-786. Delmotte, P. (1971) Gel isoelectric focusing of cerebrospinal fluid proteins -- A potential diagnostic tool, Z. klin. Chem. klin. Biochem., 9: 334-336. Delmotte, P. and R. Gonsette (1977) Biochemical findings in multiple sclerosis, Part 4 (Isoelectric focusing of the CSF gamma globulins in multiple sclerosis (262 cases) and other neurological diseases (272 cases), J. Neurol., 215: 27-37. Garvey, J.S., N.E. Cremer and D.H. Sussdorf (1977) Methods in Immunology, 3rd edition. W.A. Benjamin Inc., Reading. MA. pp. 175 178.
127 Hershey, L.A. and J.L. Trotter (1980) The use and abuse of cerebrospinal fluid IgG profile in the adult - - A practical evaluation, Ann. Neurol., 8 : 426~34. Johnson, K.P., S. C. Arrigo, B.J. Nelson and A. Ginsberg (1977) Agarose electrophoresis of cerebrospinal fluid in multiple sclerosis, Neurology (Minneap.), 27: 273-277. Lasne, Y., I. Benzerara, G. Chazot and R. Creyssel (1981) A sensitive method for characterization of oligoclonal immunoglobulins in unconcentrated cerebrospinal fluid, J. Neurochem., 36 : 1872-1874. Laurenzi, M.A. (1981) Immunochemical characterization of immunoglobulins and viral antibodies synthesized within the central nervous system in patients with multiple sclerosis and controls, Acta neurol, scand., 63 (Suppl. 84): 1-84. Laurenzi, M.A. and H. Link (1978) Comparison between agarose gel electrophoresis and isoelectric focusing of CSF for demonstration of oligoclonal immunoglobulin bands in neurological disorders, Acta neurol, scand., 58: 148-156. Link, H. (1973) Comparison of electrophoresis on agar and agarose gel in the evaluation of gammaglobulin abnormalities in cerebrospinal fluid and serum in multiple sclerosis, Clin. chim. Acta, 46 : 383-389. Link, H. (1975) Demonstration of oligoclonal immunoglobulin G in Guillain-Barr6 syndrome, Acta neurol, scand., 52:111-120. Link, H. (1978) Characteristics of the immune response within the CNS in neurological disorders, Acta neurol, scand., 57 (Suppl. 67): 177-190. Olsson, J.-E. and H. Link (1973) Immunoglobulin abnormalities in multiple sclerosis - - Relation to clinical parameters: exacerbations and remissions, Arch. Neurol. (Chic.), 28: 392-399. Ros6n, A., K. Ek and P. ,~man (1979) Agarose isoelectric focusing of native human immunoglobulin M and ~2-macroglobulin, J. immunol. Methods, 28: 1-11. Rostr6m, B. and H. Link (1981) Oligoclonal immunoglobulins in cerebrospinal fluid in acute cerebrovascular diseases, Neurology (Minneap.), 31: 590-596. Saravis, C. A. and N. Zamcheck (1979) Isoelectric focusing in agarose, J. irnmunol. Methods, 29: 91-96. Tibbling, G., H. Link and S. Ohman (1977) Principles of albumin and IgG analyses in neurological disorders, Part 1 (Establishment of reference values), Scand. J. clin. Lab. Invest., 37: 385-390. Trotter, J.L. and B.R. Brooks (1980) Pathophysiology of cerebrospinal fluid immunoglobulins. In: J. H. Wood (Ed.), Neurobiology of Cerebrospinal FluM, Vol. 1, Plenum Press, New York, London, pp. 465~,86.