Cerebrospinal fluid free kappa light chains versus IgG findings in neurological disorders: qualitative and quantitative measurements

Journal of Neuroimmunology ELSEVIER

Journal of Neuroimmunology

62

(1995)19-25

Cerebrospinal fluid free kappa light chains versus IgG findings in neurological disorders: qualitative and quantitative measurements K.J.B. Lamers a**, J.G.N. de Jong a, P.J.H. Jongen a, M.J.H. Kock-Jansen E.M.W. Prudon-Rosmulder a a University h University

Hnspitul

Hospital

Nijmegen,

Nijmegen.

Central

Institute

of Neurology,

Clinical

Chemical Laboratory,

PO Box 9101. 6500 HB Nijmegen.

b, M.A. Teunesen b,

The NetherlanA

PO Box 9101. 6500 HB Nijmegen.

The Netherlands

Received 21 February 1995; revised 18 April 1995; accepted 15 May 1995

Abstract In this study free kappa light chains in cerebrospinal fluid (CSF) were determined both by an affinity mediated capillary blotting technique after isoelectric fccusing (IEF) in agarose gel and by a quantitative enzyme linked immunosorbent assay (ELISA). The free kappa results were compared with the IgG findings in 4 neurological patient groups with a distinct CSF IgG pattern: (1) CSF without oligoclonal IgG bands, (2) CSF with setum derived IgG bands, (3) CSF restricted IgG bands and (4) CSF restricted and serum derived IgG bands. Oligoclonal free kappa bands are nearly absent in CSF of groups 1 + 2, and present in 88% of group 3 and 84% of group 4 patients. We could also establish free kappa indices from specimens in the 4 groups in analogy to IgG indices. Group I had a median free kappa index of 1.1, group 2: 1.0 and groups 3 + 4: 10.0. The correspondence between immunoblot and index findings for free kappa is better than for IgG. Free kappa index is more sensitive but somewhat less specific than IgG index for establishing intrathecal immune production. Keywords:

Cerebrospinal fluid; Free kappa light chain; IgG index; Free kappa index

1. Introduction Several diseases of the nervous system are associated with immunological abnormalities in cerebrospinal fluid (CSF). Selective increase in the CSF IgG level has been observed consistently in MS and it has become an important laboratory test for confirmation of the diagnosis (Poser et al., 1983). A selective CSF IgG increase can be established by using the IgG Index (Link and Tibbling, 1977), formulas for intrathecal produced IgG (Reiber and Felgenhauer, 1987; Ghman et a:., 1992) or by detection of CSF restricted oligoclonal IgG bands (Walker et al., 1983; Luxton et al., 1990). The IgG finding in MS is nonspecific, as also patients with central nervous system (CNS) infections or inflammatory diseases have regularly elevated CSF IgG levels. Besides IgG, also selective increase of free kappa and lambda light chains (free kappa, free lambda) in CSF has

* Corresponding

author.

Phcne

t 31 (80)

614567;

Fax

t 31 (80)

540297. 016S-5728/95/$09.50

0

SSDI 0165-5728(95)00086-O

1995 Elsevier Science B.V. All rights reserved

been found in MS and other neuro-immunological disorders (Fagnart et al., 1988; Rudick et al., 1986; Gallo et al., 1989; Decarli et al., 1987). CSF oligoclonal free kappa bands in MS have been observed with about the same frequency as that for CSF oligoclonal IgG bands (Sindic and Laterre, 1991a). Free light chain determination in CSF also has relevance for the diagnostics of patients with early MS (Lolli et al., 1991) and the presence of CSF free light chains correlates significantly with recent demyelinating activity in MS (Vakaet and Thompson, 1985; Bracco et al., 1987). From these studies it is clear that intrathecally produced CSF IgG is mostly accompanied by intrathecally produced CSF free light chains and free light chain detection is a complementary, although optional, test to establish a laboratory-supported diagnosis (Sindic and Laterre, 1991b). However, in all publications the detection limit for free light chain determination is too high to measure CSF levels in ‘normal’ patients. Furthermore, the coincidence of free light chain levels with oligoclonal free light chains in CSF and the relation between IgG- and free light chain findings in CSF are not systematically investigated. Therefore, we developed a sensitive ELISA test for

K.J.B. hmers

20

et al. /Journal

CSF W

free kappa boundkappa

w

serum CSF serum CSF serum CSF serum

free kappa

CSF serum

bound kappa

CSF serum

W free kappa boundkappa

CSF serum CSF serum CSF serum

w

CSF serum

free kappa

CSF serum

boundkappa

CSF serum

w

CSF serum

free kappa

CSF serum

boundkappa

CSF serum

Fig. 1. Examples of immunoblot findings. Always the CSF and the corresponding serum were tested against anti IgG, anti free kappa and anti bound kappa. From top to bottom you can find an example from groups 1, 2, 3, and 4 and from a patient with an IgG kappa paraproteinemia respectively. Oligoclonal free kappa is only present in the CSF specimen of the patient from group 3 and group 4.

quantification of free kappa and a method for detecting oligoclonal free kappa and investigated 4 groups of neurological patients with a distinct CSF IgG pattern according to the classification described in the consensus report (Andersson et al., 1994).

of Neuroimmunology 62 (1995) 19-25

2. Materials and methods 2.1. Patients We studied CSF and serum samples from 379 patients with various neurological disorders. All samples were routinely investigated for the presence of oligoclonal IgG bands with isoelectric focusing. Four patient groups were formed with respect to IgG oligoclonality: Group 1 (n = 183; ‘normal IgG pattern’) had a normal CSF and serum pattern without oligoclonal IgG bands: no local IgG synthesis; Group 2 (n = 93; ‘mirror IgG pattern’) had identical oligoclonal IgG bands in CSF and serum: no local IgG synthesis; Group 3 (n = 52; ‘CSF IgG-restricted’) had CSF-restricted oligoclonal IgG bands: local IgG synthesis; Group 4 (n = 51; ‘CSF IgG-restricted plus’) had CSF-restricted IgG oligoclonal bands in addition to identical IgG bands in CSF and serum: local IgG synthesis. 2.2. Methods All paired samples of CSF and serum were collected during the year 1993 and aliquots were stored at -20°C. IgG and albumin were determined by immunonephelometry on a Cobas FARA II apparatus. The upper limit for IgG index was 0.58 for a normal CSF/serum albumin ratio (Q albumin < 8.0 X 10e3) and 0.62 for Q albumin between 8.0 and 20 X 10e3. Intrathecally produced IgG was calculated according to the Reiber formula (Reiber and Felgenhauer, 1987). Isoelectric focusing (IEF) of oligoclonal IgG was performed with unconcentrated CSF exactly according to the technique described by Keir et al. (1990). 100 ng IgG from CSF and diluted serum was always applied. This method utilized IEF in agarose solution (0.3 g agarose in 2 ml Pharmalyte 3-10 and 0.5 ml Pharmalyte 8-10.5, both Pharmacia), followed by passive protein transfer onto nitrocellulose (NC) membrane with immune detection of IgG by double antibody, using horseradish peroxidase (HRP) as a visualizing agent. Goat antihuman IgG Fc (Sorin Biomedica, Amsterdam) and HRP labelled rabbit antigoat antiserum were used (Dakopatts code P160, Copenhagen).

Table 1 Numbers and percentages of free kappa positive immunoblots in CSF and serum Group

Oligoclonal IgG pattern

Number

1 2 3 4

CSF normal IgG CSF mirror IgG CSF &G-restricted CSF IgG-restricted (plus)

183 93 52 51

CSFfree kappa positive

Serum free kappa positive

Number

%

Number

46

6 7 46 43

3 8 88 84

0 6 0 1

0 6 0 3

K.J.B. Lumers et al/Journal

of Neuroimmunology 62 (1995) 19-25

21

Table 2 Median values and ranges of immunological parameters in CSF and serum in the 3 patient groups Parameter

Group 1: IgG normal (n = 37)

Unit

(Range) (9-51) (5.8-85) (6.0-21) (2.7-9.5) (2.5-10) (1.1-4.4) (1.2-11) (0.42-0.58) (0.5-2.9)

Median

CSFIgG

ma/l -.

CSF free kappa Serum IgG Serum free kappa Ratio albumin Ratio IgG Ratio free kappa Index IgG Index free kappa

8/l m8/1 x 10-3 x10--3 x lo--3 -

21 25 11 4.2 4.0 1.7 5.0 0.46 1.1

fig/l

Group 2: IgG mirror (n = 31)

42 50 11 6.1 7.7 3.6 9.6 0.46 1.0

0

Median

35 330 10 5.0 5.2 3.9 68 0.67 10

(Range) (12-130) (20-2100) (4.9-23) (1.2-18) (2.4-19) (1.4-11) (5.6-330) (0.44-2.9) (1.6-89)

The NC membrane was washed in PBS and blocked during 1 h in 3% bovine serum albumin/PBS and again washed in PBS. The focused proteins in agarose were blotted onto the NC membrane for 30 min. After blocking of the NC membrane for 60 min in 2% milk powder in 0.15 M NaCl the membrane was incubated with (1:500) diluted rabbit antihuman total kappa, HRP labelled, in 0.2% milk powder/NaCl for 2 h (code P129 Dakopatts, Copenhagen), followed by color development.

kappa

. 0

(Range) (10-150) (8.8-410) (4.1-20) (1.6-23) (3.0-49) (1.3-35) (3.3-55) (0.36-0.71) (0.5-4.4)

Median

2.2.1. Free kappa immunobbt For oligoclonal free kappa the same method as for oligoclonal IgG was used with minor modifications. Always 10 ~1 unconcentrated (or diluted) CSF and 7.5 ~1 (1: 100) diluted serum were applied upon the agarose gel. Instead of passive protein transfer affinity-mediated capillary blotting was performed. The NC membrane was incubated overnight with rabbit antihuman free kappa, diluted (1:500) with PBS (code Al00 Dakopatts, Copenhagen).

Q-free

Group 3 + 4: IgG restricted (plus) (n = 38)

1

I

I

5

10

15

20

Q-albumin

Fig. 2. Relation between Q albumin (X lo-‘) and Q free kappa (X 10m3) in patients of group (3 + 4) (top) and in patients of groups 1 and 2 (bottom).

22

K.J.B. L.umers et al. /Journnl

2.2.2. Free kappa ELISA Free kappa light chains were measured by a solid phase ‘sandwich’ ELISA. Microtiter plates were coated overnight with 200 ~1 (1: 1000) diluted rabbit antihum~ free kappa (Dakopatts, code AlOO, Copenhagen). Reagents are the same as in an earlier described method (Haraldson et al., 1991). After washing, blocking with 2% bovine albumin and washing 100 ~1 native CSF and (1:lO and 1:lOO) diluted CSF and corresponding serum (1:2000) diluted were incubated 90 min at 37°C. After washing, 100 ~1 HRP conjugated rabbit antihuman total kappa (1:lOOO) diluted (Dakopatts, code P129, Copenhagen) was incubated 90 min at 37°C. Color development was performed with OPD-reagent. Standard curves were performed with dilutions of an oligoclonal free kappa pool CSF. The pooled CSF was left 2 days at 4°C centrifuged and filtered through a 0.2 pm filter and divided in aliquots of.0.5 ml and stored at -70°C. The CSF pool was calibrated on a commercial purified free kappa standard (2.7 mg/ml, Tago, Burlingame, CA, USA). CSF pool is 100%. We used always 9 standard dilutions (from 0.1% to 2% of the CSF pool). 1% CSF pool solution corresponds with 14.6 pg/l Tago free kappa light chain concen~ation. With help of the CSF pool standards the reproducibility of the assay was much better than with Tago standards. The lower limit of sensitivity is 2 pg/l. To evaluate intra-assay precision, we studied two difuted CSF samples containing 3 and 7 pg/l free kappa, repeating the assays 20 times on the same day. The coefficient of variation (CV) was respectively 6.1% and 2.9%. The inter assay precision was determined by daily assay of 1 CSF sample, containing 3 pg/l free kappa for 9 days. The CV was 8.1%. Concentrations remained stable during storage at - 70°C and - 20°C for 3 months. To test the specificity of the assay we used CSF specimens with a high free kappa level in our ELISA test. In a modified ELISA test we used in stead of HRP conjugated rabbit antihuman total kappa as second anti-

CSF-free

62 (19953 19-2.5

of Neuroimmunology Index

- free

kappa

._-

Qll Groupnr.

/

1

I

2

Fig. 4. Distribution of index free kappa values measured in patients of groups 1, 2 and (3 +4). The tine represents the upper limit (+ 2 SD) of the control group.

body, HRP conjugated goat antihuman IgG (Fc), gamma chain specific. The free kappa positive specimens all gave a negative reaction with this second antibody. This negative reaction excludes a cross reaction of intact IgG molecules with rabbit antihuman free kappa.

3. Results 3.1. Free kappa investigation with immunoblot

All CSF and serum samples were tested for the presence of oligoclonal free kappa bands. Fig. 1 shows examples of immunoblot findings in patients out of the 4 groups and also in one patient with an IgG kappa p~aproteinemia. There is no cross reaction of antibodies to oligoclonal free kappa with oligoclonal IgG molecules. Table 1 summarizes the results of the free kappa immunoblots in the 4

kappa (ug/L)

lntrathecal

CSF-IgG

’

3+6

Img/L)

Fig. 3. Relation between intrathecal CSF IgG (mg/l) and CSF free kappa ( pg/l) in patients of group (3 + 4, CSF restricted plus).

K.J.B. hmers

et al. /Journal

of Neuroimmunology 62 (1995) 19-25

groups. Generally, oligoclonal free kappa bands are very scarce in serum: the 6 specimens from group 2 with positive serum free kappa had an identical free kappa pattern in serum and CSF. In patients with local IgG synthesis (groups 3 + 4), CSF oligoclonal free kappa is frequent (88% in group 3 and 84% in group 4). The free kappa oligoclonal pattern was always different from the IgG oligoclonal pattern in all positive specimens.

3.2. Free kappa investigation with ELISA

23

Table 4 IgG index and free kappa index results in the three patient groups Group

Oligoclonal IgG pattern

Number

IgG index increased

Free kappa index increased

1 2 3+4

Normal Mirror Restricted (plus)

37 31 38

0 2 26

2 3 34

3.3. Correspondence between immunoblot and index find-

ings In a number of patients in each group we measured in CSF and serum also the quantitative levels of free kappa, IgG and albumin, and calculated ratios and indices. We selected 37 patients from group 1, 31 patients from group 2 and 38 patients from groups 3 + 4. The results of the free kappa immunoblots in groups 3 + 4 were not different. Therefore we handled groups 3 + 4 as one group. In Table 2 we present the median values and the ranges of the immunological parameters for the 3 groups. The median serum levels of free kappa in the 3 groups are comparable but the median CSF levels are different: 25, 50 and 330 pg/l for groups 1, 2 and (3 + 4) respectively. Furthermore, the median level of index free kappa in group (3 + 4) (local IgG synthesis) is lo-times higher than the median level of index free kappa of groups 1 and 2 (no local IgG synthesis). All 37 patients in group 1 had normal white cell number in CSF, an albumin ratio I 10 X 10T3, no oligoclonal IgG bands and normal IgG index. The mean f SD of the free kappa index for this ‘IgG-normal’ group is 1.3 + 0.5. Therefore, we considered patients of group 1 as a nonimmunological ‘control group’ for free kappa index with 2 SD limits: 0.3-2.3. In Fig. 2, the relation between ratio serum/CSF (Q) albumin and Q free kappa is presented for the 3 groups. The correlation coefficient between Q albumin and Q free kappa is 0.64 for group 1, 0.49 for group 2 and - 0.064 for group (3 + 4). The correlation coefficient between intrathecal CSF IgG and CSF free kappa is 0.91 (Fig. 3). In Fig. 4 the free kappa indices for the 3 patient groups are presented.

In Table 3 we present the results of immunoblots and indices of free kappa and IgG in the 3 patient groups. In the total group of 106 patients the index and immunoblot correspond in 102 patients for free kappa (96%) and in 92 patients for IgG (87%). In Table 4 we present the results of IgG- and free kappa indices in the 3 patient groups. Based on the acceptance that IgG isoelectric focusing is the ‘gold standard’ for indication of an intrathecal IgG production, it can be concluded that for free kappa index 4 false negatives and 5 false positives were found and for IgG index 12 false negatives and 2 false positives.

4. Discussion A selective increase in CSF of free kappa has been found in patients with MS and in patients with infections of the CNS. The incidence of CSF oligoclonal free kappa in neuroimmunological disorders is not systematically studied. We investigated the coincidence of free kappa and IgG oligoclonal bands in CSF and serum in 4 neurological patient groups. Generally, free kappa bands are very scarce in serum. In 88% of patients with CSF IgG restricted bands (group 3) and in 84% patients with CSF IgG restricted and additional bands (group 4), CSF free kappa bands are present. These findings can confirm observations from other studies that CSF oligoclonal free kappa bands can be ,demonstrated in a high percentage in patients with local IgG synthesis (Sindic and Laterre, 1991b; Gallo et

Table 3 Correspondence of immunoblot and index results of free kappa and IgG in the three patient groups Group

Oligoclonal IgG pattern

Number

Free kappa Immunoblot ‘: Index b:

1 2 3+4 Total

Normal Mirror Restricted (plus1

a + , positive; - , negative. ’ + , increased; - , normal.

37 31 38 106

IgG

+ +

+ -

+

-

+ +

+ -

+

-

0 3 34 37

1 1 0 2

2 0 0 2

34. 27 4 65

0 0 26 26

0 0 12 12

0 2 0 2

37 29 0 66

24

K.I.B. Lumers ef al. /Journal

of Neuroimmunology

al., 1989; Lolli et al., 1991). In patients with identical IgG bands in CSF and serum (group 21, CSF free kappa bands are only present in 8%. From these data it is obvious that in contrast to in~athecally produced oligoclonal IgG, systemically produced oligoclonal IgG is seldom accompanied with oligoclonal free kappa. This phenomenon can probably be explained by the rapid removal of these small free kappa proteins from blood to urine as is found for Bence Jones proteinuria in patients with myeloma. In case a laboratory receives only CSF and no serum of a patient it is more conclusive to perform both oligoclonal IgG and oligoclonal free kappa isoelectric focusing. When both investigations are positive, this is a strong indication for intrathecal IgG production. When oligoclonal IgG is positive and olig~lon~ free kappa is negative, this is a strong indication for systemic IgG production. With respect to the quantitative levels, in patients without immunological CNS disturbances, the levels in CSF of free kappa are mostly below the detection limit of the assay in literature (Fagnart et al., 1988; Lolli et al., 1991; Rudick et al., 1986). We have developed a sensitive method in order to determine ‘normal’ CSF free kappa levels. Our median free kappa serum level (4.2 mg/l) is nearly 3-times higher than the value of 1.58 mg/l reported by Fagnart et al. (19881, comparable with the value of 3.36 mg/l reported by Rudick et al. (19861, and 4-times lower than the value reported by Sollink (1975). Absolute levels of free light chains in body fluids are difficult to compare owing to both the differences in specific antisera and the absence of a valid free kappa light chain standard. The use of an index mitigates the problem of an appropriate standard. As free light chains are smaller molecules than albumin, theoretically the index must be higher than 1.0 since the filtration process of the blood CSF barrier is size dependent and proportional for free light chains. The median index of 1.l for free kappa and the relationship between Q albumin and Q free kappa for the control group can confirm the validity of the use of a free+kappa index for establishing intrathecal free kappa production. Based on the acceptance that IgG isoelectric focusing is the ‘gold standard’ for indication of intrathecal IgG production, we can establish for free kappa index a sensitivity of 90% and a specificity of 93%, and for IgG index a sensitivity of 68% and a specificity of 97%. The pathophysiological significance of intrathecal production of free light chains remains obscure. Their presence is not due to local degradation of immunoglobulins but excess of free light chains is likely a marker of the ongoing intrathecal humoral immune response. In conclusion: CSF oligoclonal free kappa is observed with about the same frequency as that for CSF restricted oligoclonal IgG (groups 3 and 4 in Table 1). In case a laboratory receives only CSF and no serum of a patient oligoclonal free kappa investigation is more conclusive than oligoclonal IgG.

l

l

62 (I 995) 19-25

The correspondence between immunoblot and index findings for free kappa is better than for IgG (Table 3). Free kappa index is more sensitive but somewhat less specific than IgG index for establishing intrathecal immune production (Table 4).

References

Andersson,M., Alvarez-Cermeao, J., Bernardi, G., Cogato, I., Fredman, P., Frederiksen, J., Frederikson, S., GaIlo, P., GrimaIdi, L.M., Gunning, M., Keir, G., Lamers, K., Link, H., Magalh5es. A., Massaro, A.R., Ghman, S., Reibcr, H., RSnnback, L., Schluep, M., Schuller, E., Sindic, C.J.M., Thompson, E.J., Trojano, M. and Wurster, U. (1994) The role of cerebrospinal fluid analysis in the diagnosis of multiple sclerosis: a consensus report. J. Neurot. Neurosurg. Psych. 57,897-902. Bracco, F., Gallo, P., Menna, R., Battistin, L. and Tavolato, B. (1987) Free light chains in the CSF in multiple sclerosis. J. Neurol. 234, 303-307. Decadi, C., Menegus, M.A. and Rudick, R.A. (1987) Free fight chains in multiple sclerosis and infections of the CNS. Neurology 37, 13341338. Fagnart, O.C., Sindic, C.J.M. and Latene, C. (1988) Free kappa and Iambda hght chain levels in the cerebrospinal fluid of patients with multiple sclerosis and other neurological diseases. J. Neuroimmunol. 19, 119-132. Gallo, P., Tavolato, B., Bergenbrant, S. and Siden, A. (1989) A study with special reference to the occurrence of free light chains in cerebrospinal fluid with and without oligocfonal i~unoglobuIin G. J. Neural. Sci. 94, 241-253. Handdson, A., Kock-Jansen, M.J.H., Jaminon, M., van Eck-Atts, P.B.J.M., de Boo, T., Weemacs, C.M.R. and Bakkeren, J.A.J.M. (1991) ~te~nation of kappa and lambda light chains in serum i~unoglobuIins G, A and M. Ann. Clin. B&hem. 28.461-466. Keir, G., Luxton, R.W. and Thompson, E.J. (1990) Isoelectric focusing of cembrospinal fluid immunoglobulin G: an annotated update. Ann. Clin. Biochem. 27,436-443. Link, H. and Tibb~ng, G. (1977) Principles of ~bu~n and IgG disorders. Evaluation of IgG synthesis within the central nervous system in multiple sclerosis. Stand. J. Clin. Lab. Invest. 37, 397-401. Lolli, F., Siracusa, G., Amato, M.P., Fratiglioni, L., Dal Pozzo, G., Galli, E. and Amaducci, L. (1991) Intrathecal synthesis of free immunoglobulin light chains and IgM in initial multiple sclerosis. Acta Neurol. Stand. 83, 239-243. Luxton, R.W., McLean, B.N. and Thompson, E.J. (1990) Isoelectric focusing versus quantitative measurements in the detection of intrathecal local synthesis of IgG. Clin, Chim. Acta 187, 297-308. Ghman, S., Emerudh, H., Forsbcrg, P., Henriksson, A., von Schenk, H. and Vrethem, M. (1992) Comparison of seven formulae and isoelectrofocusing for determination of intrathecally produced IgG in neurological disease. Ann. Clin. Biochem. 29, 405-410. Poser, CM., Paty, D.W., Scheinberg, L., Mc~n~d, W.I., Davis, EA., Ebers, G.C., Johnson, K.P., Sibley, W.A., Silberberg, D.H. and Tourtellotte, W.W. (1983) New diagnostic criteria for multiple sclerosis: Guidelines for research protocols. Ann. Neural. 13, 227-231. Reiber, H. and Felgenhauer, K. (1987) Protein transfer at the biood cercbrospinal fluid barrier and the quantitation of the humoral immune response. within the central nervous system. CIin. Chim. Acta 163.319-328. Rudick, R.A., Pallant, A., Bidlack, J.M. and Hemdon R.M. (1986) Free. kappa light chains in mul~ple scterosis spinal Ruid. Ann. Nemo]. 20, 63-69. Sindic, C.J.M. and Laterre, C. (1991a) Free light chains in CSF from MS

K.J.B. Lamers et al. / Journal of Neuroimmunology 62 (1995) 19-25

patients: an affinity-mediated blot study. In Wicthdlter et al. (Eds.), Current Concepts in MS, Elsevier Science Publishers, pp. 129-133. Sindic, C.J.M. and Lateme, C. (1991b) Oligoclonal free kappa and lambda bands in the cerebrospinal fluid of patients with MS and CNS infections. An immunoaftinity-mediated capillary blot study. J. Neuroimmunol. 33, 62-73. Sollink, K. (1975) Free light chains of immunoglobulins in normal serum and urine determined by radioimmunoassay. Stand. J. Clin. Lab. Invest. 35, 407-412.

25

Vakaet, A. and Thompson, E.J. (1985) Free light chains in the cerebtospinal fluid: an indicator of recent immunological stimulation. J. Neurol. Neurosurg. Psych. 48, 995-998. Walker, R.W.H., Keir, G., Johnson, M.H. and Thompson, E.J. (1983) A rapid method for detecting oligoclonal IgG in unconcentrated CSF, by Agarose isoelectric focusing, transfer to cellulose nitrate and immunoperoxidase staining. J. Neuroimmunol. 4, 14I - 148.