Free kappa and lambda light chain levels in the cerebrospinal fluid of patients with multiple...

transcript

Journal of Neuroimmunology, 19 (1988) 119-132 Elsevier

JNI 00620

Free kappa and lambda light chain levels in the cerebrospinal fluid of patients with multiple

sclerosis and other neurological diseases

O.C. Fagnart, C.J.M. Sindic and C. Laterre Laboratory of Neurochemistry, Unwersitd Catholique de Louvain, 1200 Brussels, Belgium

(Received 28 October 1987) (Revised, received 4 March 1988)

(Accepted 21 March 1988)

Key words: Free fight chain; Multiple sclerosis; Immunoassay

Summary

Free kappa and lambda light chains were assayed by particle-counting immunoassay in cerebrospinal fluid (CSF) from patients with various neurological disorders. Detection limits were 25 and 50 ng/ml , respectively. Values of free kappa chain were higher than 50 ng /ml (upper reference limit) in 155 of 191 (81%) multiple sclerosis (MS) patients, in 100 of 168 (605[) patients with central nervous system (CNS) infections but in 41 of 217 (19%) patients with other neurological disorders. Free kappa chains were also assayed in 273 matched sera. The mean concentration in the control group (1.58 # g /m l ; SD: 0.41) did not differ significantly from those in MS sera (1.63 #g /ml ; SD: 0.43). The free kappa chain index was increased in 86% of MS patients and in 40% of patients with CNS infections. Regarding free lambda chains, CSF values were higher than 240 n g / m l (upper reference limit) in most neurological disorders (50-100%). However, the use of a lambda chain index increased the specificity of the assay as this index was higher than the upper reference value in 86% of MS patients and in only 23% of patients with infectious diseases. In MS, high levels of free kappa and lambda indices correlated significantly (P < 0.01) with either the presence of oligoclonal bands or a high IgG index. Local synthesis of free light chains is an additional marker of an ongoing immune response within the CNS, especially in MS.

Address for correspondence: C.J.M. Sindic, Laboratory of Neurochemistry, Universit6 Catholique de Louvain, 5359, Av. E. Mounier, 1200 Brussels, Belgium.

Introduction

Immunoglobulins G, A, and M are produced within the CNS in various neurological diseases, such as MS and CNS chronic infections (Kabat et al., 1948; Felgenhauer, 1982; Sindic et al., 1987a). In most cases, the local production of IgG is related to the appearance in electrophoresis of oligoclonal bands restricted to the CSF (Lowenthal et al., 1960; Laterre, 1965).

Free kappa chains were identified as a minor contaminant in preparations of beta-trace protein from normal CSF (Link, 1967). In addition, free kappa and lambda light chains were detected in CSF of either MS patients or patients with meningitis and encephalitis by double diffusion (Iwashita et al., 1974; Bollengier et al., 1975, 1978; Riberi et al., 1978), immunoelectrophoresis or crossed immunoelectrophoresis (Kolar, 1977; Vandvik, 1977; Fryden and Link, 1979; Perini et al., 1979) and immunofixation or immunoblotting after isoelectric focusing (Laurenzi et al., 1980; Mattson et al., 1982; Vakaet and Thompson, 1985; Bracco et al., 1987). These free light chains also appeared as homogeneous bands, more anodic, however, than the bulk of IgG. Only few studies dealt with their quantitative analysis (Rudick et al., 1986a, b; De Carli et al., 1987). Here we describe sensitive and fast assays of free kappa and lambda light chains in CSF and sera from patients with various neurological disorders and discuss their clinical relevance.

Materials and methods

Patients Sera (n = 273) and CSF samples (n = 626) from 626 patients were frozen at

- 2 0 ° C after centrifugation. Control patients (n=50) (group I) suffered from minor neurosis or tension headache but were devoid of clinical signs of neurological disorders. They were normal at electroencephalography, computed tomography and CSF analysis. Group II consisted of 217 samples from 217 patients with various, but non-infectious, neurological disorders: sciatica (n = 15), peripheral neuropathy (n = 50), pre-senile or senile dementia of the Alzheimer type (n = 18), stroke (n = 50), tumours (n = 20), optic neuritis (n = 27) and Guillain-Barr6 syndrome (n = 37). Group III included 168 CSF samples from 168 patients with infection of the CNS: 48 patients with aseptic meningitis, 24 with bacterial meningitis, 17 with acquired immunodeficiency syndrome (AIDS), ten with neurosyphilis, 24 with Borrelia burgdorferi meningoradiculitis or myelitis, and 45 with viral encephalitis. Group IV was the MS group and consisted of 191 CSF samples from 191 patients with either 'clinically definite' or 'laboratory-supported definite' MS (Poser et al., 1983); in all cases, CSF was collected for diagnosis and no serial samples were available for a given patient. About 15% of the MS patients were in apparent clinical remission at the moment of sampling.

Immunoassay of free kappa light chains Free kappa chains were assayed by particle counting in the IMPACT instrument

(Acade Diagnostic Systems, Brussels, Belgium). The reliability of this technique has

been shown elsewhere (Masson et al., 1981). Briefly, latex particles (Estapor K 150, Rh6ne Poulenc, Courbevoie, France), covalently coated with F(ab')2 fragments of rabbit antibodies specific for free kappa light chains (Dakopatts, Copenhagen, Denmark) were agglutinated by the antigen, the agglutination being measured by optical counting of the unagglutinated particles. For the assay in CSF, samples were diluted 5 times automatically in glycine-buffered isotonic saline (GBS) containing 50 mM ethylenediaminetetraacetic acid (EDTA), aggregated F(ab')2 fragments (0.5 mg/ml) from normal rabbit serum (Collet-Cassart et al., 1981) and Tween 20 (polyoxyethylene 20 sorbitan monolaurate) surfactant (40 ml/1).

The serum samples were manually prediluted 8 times in GBS and incubated for 40 min with the same volume (30 #1) of latex and of a third reactive (50 mM GBS-EDTA containing 10 g of Dextran T500 (Pharmacia, Uppsala, Sweden) and 1 ml of Tween 20 per live). The reaction was stopped by addition of 1 ml of GBS and the suspension was sent to the flow cell for measurement.

A pool of Bence Jones kappa chains (Tago, Burlingame, CA, U.S.A.) was used as standard (concentration estimated to be 3.4 mg/ml by the Lowry method). The purity of the preparation was confirmed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The standard was serially diluted in GBS containing 0.2~ bovine serum albumin (BSA, Calbiochem, Hoechst, F.R.G.). The number of non-ag- giutinated particles expressed on the recorder as peak height plotted vs. the log concentration of kappa chains gave a sigmoidal curve extending from 25 to 7200 ng/ml (Fig. 1, left panel). Excess antigen decreased the agglutination for concentrations exceeding 25 000 ng/ml, which would be read as 3600 ng/ml on the calibration curve. The minimum detectable concentration defined as the concentration giving a peak height 3 SD lower than the mean of ten peak heights obtained in the absence of free kappa, was 25 ng/ml. To evaluate intra-assay precision, we studied three CSF samples containing 120, 450, and 1000 ng/ml of free kappa, repeating the assays 15 times on the same say. The coefficients of variation (CV) were respectively 3.5, 4.2, and 3.6%. Interassay precision was assessed by daily assay of

~o ~6o ~ I~ ,.6oo Free light chains [ ng/ml]

~ 0 Whole IgG (mg/ml)

Fig. 1. Left panel: Standard curves for determination of free kappa (0) and free lambda (o) light chain levels by IMPACT: peak height is proportional to the concentration of non-agglutinated particles flowing through the particle counter. Right panel: Displacement curves for the determination of cross-reacting

whole IgG in the kappa (e) and lambda (o) assay system.

three samples containing 50, 650, and 1600 ng/ml for 15 days. The CVs were, respectively, 13.8, 9.8, and 2.9%. Analytical recovery, evaluated on ten CSF and ten sera supplemented with different concentrations of free kappa, was 91.4% (CV = 8.6) for 50 ng/ml, 96.5% (CV = 3.5) for 100 ng/ml, 95.7% (CV = 4.2) for 450 ng/ml in CSF, and 107% (CV = 11) for 200 ng/ml in sera. To test the specificity of the assay, IgG isolated from normal human serum by double precipitation with ammonium sulfate and Rivanol followed by gel chromatography on Ultrogel (LKB, Bromrna, Sweden) was added at increasing concentrations. No agglutination occurred even at a concentration of 0.5 mg/ml (Fig. 1, right panel).

lmmunoassay of free lambda light chains The immunoassay of free lambda chains was similar to that of free kappa chains.

Latex coated with F(ab')2 fragments of rabbit anti-human free lambda chains (Dakopatts, Copenhagen, Denmark) was agglutinated by the antigen to be assayed. No third reactive was used and the reaction time was 15 min. The 20-fold automated dilution was in GBS containing 1 g of BSA (United Biochemical Corporation, U.S.A.) per litre and 10 ml of Tween 20 per litre. Serum samples were manually prediluted 10 times in GBS. A pool of Bence Jones lambda type chains (Tago, Burlingame, CA, U.S.A.) serially diluted in GBS-BSA was used for the calibration curve, which ranged from 50 to 3750 ng/ml (Fig. 1, left panel). The detection limit was 50 ng/ml. Analytical recoveries were 98% (CV = 9.3) for 350 ng/ml, 99% (CV = 6.8) for 650 ng/ml and 105% (CV = 5.2) for 2000 ng/ml in CSF, and 108% (CV = 12.5) for 500 ng/ml in sera. The intra-assay CVs were 4.4, 3.7, and 3.8% for 120, 1500, and 2350 ng/ml, respectively. The interassay CVs were 12.4, 8.2, and 6.4% for 300, 600, and 1200 ng/ml, respectively. Regarding the specificity of the assay, agglutination was already observed with an excess of whole IgG of 0.05 mg/ml. Therefore, the commercially available antiserum had to be first absorbed on pooled human IgG coupled to CNBr-activated Sepharose 4B (Pharmacia, Uppsala, Sweden) before preparation of the F(ab')2 fragments. After such absorption, cross-reaction appeared with an excess of whole IgG reaching 1 mg/ml (Fig. 1, right panel). These conditions were suitable to assay free lambda chains in serum.

Other procedures Total protein in CSF was determined by the trichloroacetic acid method (Sindic

and Cambiaso, 1982). IgG and albumin in serum and CSF were assayed by immunonephelometry (AlP System, Technicon, Tarrytown, NY, U.S.A.). Upper reference limits for CSF IgG and IgG index (Delpech and Lichtblau, 1972; Tibbling et al., 1977) were 5.5/zg/ml and 0.7, respectively. Agar gel electrophoresis of CSF and matched serum was run as described elsewhere (Laterre, 1965).

Free kappa chains The concentration of free kappa light chains in CSF was below 25 ng/ml (limit

of detection) in all but two samples of the control group (group I; n = 50). The

highest value, 50 ng/rnl, was considered as the upper reference limit. Values exceeding 50 ng /ml were found in 41 samples out of 217 (19%) of group II (non-infectious diseases), in 100 samples out of 168 (60%) of group III (infectious diseases), and in 155 samples out of 191 (81%) of group IV (MS) (Fig. 2). In MS, free kappa chains did not correlate with IgG concentration ( r = 0.11). In contrast, a highly significant correlation was found in this group between high levels of free kappa chains and either high IgG index or oligoclonal patterns in agar gel electrophoresis (P < 0.001; Table 1). It should be noted that out of 26 CSF samples from MS patients without oligoclonal bands, seven (26%) had increased free kappa chains. In optic neuritis (n = 27), high levels of free kappa chains and presence of ohgoclonal bands correlated also significantly ( P < 0 . 0 0 1 ; Table 2), as already described (Rudick et al., 1986b). In infectious diseases, kappa chain levels did not

20000 2000

E soot

~ 2o0 a .

~ IOO U -

0/5 0 0

2/15 1/50 7/50 15/37 2/18 7/20 7!271 20124 28/45 12117 7/10 17/48 16/2/. 13 2 14 40 11 35 26 / 83 62 70 70 35 67

• , . .

Q Q I Q

7 .....

155/191

° •

t.- v.

" ; J ; • , , • •

I ,¥•

• J J , ,

. ' . ' .

. l l . . .

• • •

. . . .

:- . . . . . . . . . (I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ',--; . . . . . . . . . . . . . . . . . . . . . . . : . : : : . ; . . . . . . . . ', ............. I ! . . , , . . , . , . . . . , . . , . .

• , . , . . . .

I . , . . , . ; ; | ; , . .

25111111111111 ...ii.. lllill!! LW.I.t " ' " ' " " " " ,mu, ..-;;: . . . . " . . . . . .

i fi 111 iv

Fig. 2. Distribution of free kappa chains in CSF measured in the control group and in patients with various neurological disorders. The top row gives the number of patients in each group whose free kappa

chain levels exceoded the upper reference limit (broken line), i.e., 50 ng/ml.

20000 "2o0o

. . . . . . . . . . 100

. . . . . . . . . . 1 2 5

TABLE 1

CHI-SQUARE ANALYSIS OF THE ASSOCIATION OF O L I G O C L O N A L BANDS A N D IgG

INDEX WITH THE CONCENTRATION OF F R E E L I G H T CHAINS A N D THE KAPPA A N D LAMBDA INDICES IN MS

Oligoelonal bands

Free kappa chains

< 50 n g / m l 19 17 > 50 n g / m l 7 148 X 2 = 40.0; P < 0.001

Free lambda chains

< 240 n g / m l 8 7 > 240 n g / m l 18 158 X 2 = 21.5; P < 0.001

Kappa index

< 5.5 8 12 > 5.5 12 111 X 2 = 13; P < 0.001

Lambda index

<17 6 9 > 17 10 76 X z = 7.7; P < 0.01

I g G l n d e x

< 0.7 > 0.7

Kappa index < 5.5 18 2 > 5.5 29 94 X 2 = 36.9; P < 0.001

Lambda index

<17 11 4 > 17 28 58 X 2 = 8.9; P < 0.01

correlate with IgG concentrations (r = 0.19; non-significant), and high levels of flee kappa chains were not associated with a high IgG index (X 2 = 0.8; non-significant).

Free kappa chains were assayed in 273 matched sera collected on the same day as CSF samples. The mean serum concentration found in the control group (1.5 #g /ml , SD: 0.41) did not differ significantly from those found in MS sera (1.63 #g /ml , SD: 0.43). To discriminate between local production and passive transuda-

TABLE 2

CHI-SQUARE ANALYSIS OF THE ASSOCIATION OF OLICKX~LONAL BANDS WITH THE CONCENTRATION OF FREE KAPPA CHAINS IN THE CSF OF PATIENTS WITH OPTIC

NEURITIS

Oligoelonal bands

Free kappa chains < 50 n g / m l 17 3 > 50 n g / m l 0 7 X 2 - 30.9; P < 0.001

1000O 7/501t,

1000" E

200" ---~ .... ~ .

1.1...

100: ".,,1~1.. ~.

,oO~ ,,o,~

2115 2~7 1~/50 2~37 l l /p 12~0 I&/15

13 5 28 73 64 57 93

: : 3. .8. ___. ~ ..~..

: @ ;~ .... 2"i

~°° | e.~*° °°

| ° ~ ° o

I i I i i i m .~, ,~.~, ~, _@ ~ ,, ,..4 ..-" ~%" .'° ~7,~" -~ o° o.°"

+, %_ |

18124 43/45 14/14 I0/I0 27//,3 24/24 176119 l

7s 96 ,oo ,oo s3 1oo ~z

• o I o o . . • o ° e

• . . ; . . .

°, , .o - ,

o, : I

,o • --%

I "%" i o~e

I .-_It ~t_-

• ; •

I 1 i i i I " I

.,.'?.o~; .'-'~Lo ~' ° e

Fig. 3. Distribution of free lambda chains in CSF measured in the control group and in patients with various neurological disorders. The top row gives the number of patients in each group whose free

lambda chain levels exceeded the upper reference limit (broken line), i.e., 240 ng/ml.

tion of free kappa chains, we calculated a 'free kappa chain index' by comparison with the Ig index ((CSF kappa/serum kappa) : (CSF albumin/serum albumin)). The highest index observed in controls (5.5) was considered as the upper reference limit. Indices higher than 5.5 were found in few samples (4/25, 16%) from group II (non-infectious diseases), in 30/75 samples (40%) from group III (infectious diseases) and in 123/143 (86%) MS patients (Fig. 4, left panel). In the infectious group, an increased kappa index was not correlated with either a high IgG index or the presence of oligoclonal bands (X 2= 0.99 and 0.32, respectively) and was mainly observed in viral encephalitis (seven out of 11 patients) and in meningoradiculitis due to B. burgdorferi (14 out of 20 patients) but rarely in viral meningitis (one out of 26 patients) (Fig. 4). In contrast, in MS, a high kappa index was significantly associated with a high IgG index and presence of oligoclonal bands (Table 1). It should be noted that, in MS, the kappa index was more often increased than the IgG index (86% vs. 67%), with a frequency exactly similar to the presence of oligoclonal bands (86%).

I(3(30 800

~, Io(

0•30 4/25 30/75 1231143

°iX:If"

o•°

.. -..~" = .:.:- ... "-

10001 0•30 2/25 15164 861101

go :~:

• : ..'(-

20 20 . . . . . . . . . . . . . . . . . • . . . . . . . . . 3L . . . . . . . . ~k'_ . . . . . .

• =. ,~ .~. ," "..~. ~-" • 6" ..L'L

. . . . . ~'~"~:'- . . . . . - ; ~ - . . . . . . ~ . . . . . "- '"~" . . . . o --- Group n ° I II Ili IV I 11 III 1V

Fig. 4. Distribution of kappa (left panel) and lambda (right panel) chain indices in controls (group I) and in various neurological disorders (group II: non-infectious diseases; group III: infectious diseases; group IV: MS). The broken lines represent the upper reference limits. In the infectious group, high kappa indices were observed in four out of six patients with bacterial meningitis, one out of 26 with viral meningitis, two out of eight with AIDS, two out of four with neurosyphilis, seven out of 11 with viral encephalitis and 14 out of 20 with meningoradiculitis due to B. burgdorferi. In the same group, high lambda indices were observed in two out of 14 patients with bacterial meningitis, two out of 20 with viral meningitis, three out of 12 with viral encephalitis and eight out of 18 with meningoradiculitis due to B.

burgdorferi.

Free lambda chains F o r the c o n t r o l g r o u p (n --- 50), the m e a n c o n c e n t r a t i o n o f free l a m b d a c h a i n s in

C S F was 120 ( S D = 60) n g / m l . E x c e p t in sc ia t i ca a n d p e r i p h e r a l n e u r o p a t h y (Fig .

3), c o n c e n t r a t i o n s h ighe r t h a n 240 n g / m l c o n s i d e r e d as t he u p p e r r e f e r e n c e l imi t

w e r e f o u n d in a p r o p o r t i o n v a r y i n g f r o m 50 to 100% o f C S F samp le s in m o s t

n e u r o l o g i c a l d i so rde r s w i t h o u t g rea t speci f ic i ty . T h e r e was no s ign i f i can t r e l a t i on

b e t w e e n k a p p a a n d l a m b d a C S F c o n c e n t r a t i o n s in the M S g r o u p ( r = 0.15), w h e r e a s

in in fec t ious d i seases these two p a r a m e t e r s c o r r e l a t e d s ign i f i can t ly ( r = 0.64; P <

0.001). F r e e l a m b d a cha ins d id n o t c o r r e l a t e w i th to ta l I g G n e i t h e r in the M S g r o u p

( r = 0.23) n o r in t he g r o u p of i n f ec t i ous d i seases ( r = 0.10). W e also a s sayed f ree

l a m b d a cha ins in 220 sera co l l ec t ed o n the s a m e days as the C S F samples• T h e m e a n

c o n c e n t r a t i o n f o u n d in the con t ro l g r o u p (3.9 # g / m l ; S D : 1.6) d i d n o t d i f f e r f r o m

those found in MS sera (4 /xg/ml; SD: 1.3) but was significantly (P < 0.02) increased in sera from patients with infectious (5.9 # g /m l ; SD: 2) or non-infectious (6.3 ~g /ml ; SD: 1.6) neurological disorders. The difference for this latter group was mainly due to the high levels of free lambda chains observed in sera from patients with Guillain-Barr~ syndrome (8.0 ~g /ml , SD: 2.8; n = 10). We also calculated a free lambda chain index using the formula (CSF lambda/se rum lambda) : (CSF albumin/serum albumin) (Fig. 4, right panel). The mean index for the control group was 9.0 (SD: 4.0). By this method of calculation, the MS group was strikingly differentiated from other groups as 86% of the MS patients had values higher than the upper reference limit, i.e., 17 (mean + 2 SD); in contrast, only 8% and 23% of patients from groups II and III, respectively, had increased indices (Fig. 4). In the infectious group, a high lambda index was more frequently observed in meningoradiculitis due to B. burgdorferi (Fig. 4) and was significantly correlated with an increased IgG index (X 2 = 6.15; P < 0.02); in MS, a significant correlation was also found with the IgG index and the presence of oligoclonal bands (Table 1). Among 101 MS patients, 78 had an increase of both kappa and lambda indices, eight an increase of the lambda index, seven an increase of the kappa index, and eight normal indices for both types of light chains. A high kappa (but not lambda) index was significantly associated with CSF pleocytosis (> 5 ce l ls /mm 3) (X 2 = 5.01; P < 0.05). It was possible to compare the kappa and lambda indices in 60 MS patients with active disease and in ten patients in apparent clinical remission: no clear-cut differences were observed (data not shown). All patients with a disease duration shorter than 6 months (n = 16) or longer than 15 years (n = 9) displayed a high kappa index, and all but one, in both groups, a high lambda index.

Discussion

The specificity of our immunoassay for free light chains was tested in presence of an IgG excess and a starting cross-reaction was observed for IgG concentrations of about 1 mg/ml , i.e., a 4-log excess over the specific reaction. It should be noted, however, that the commercially available antiserum against free lambda chains has to be further absorbed on pooled IgG coupled to a Sepharose column. The high specificity of this immunoassay by particle counting may also be explained by the observation that, on a same weight basis, small molecules have greater agglutinating activity than molecules of higher molecular weight, as it was previously shown for monomeric and dimeric IgA (Sindic et al., 1984). In addition, the mean reference value we observed in serum for free kappa chains (1.5/~g/ml + 0.41) is very close to the value of 1.2 btg/ml reported by Axiak et al. (1987) who used a murine monoclonal antibody in an enzyme-linked immunoassay. These values are about 10 times lower than those reported in previous studies (Soiling, 1975; Soiling et al., 1982; Brouwer et al., 1985). Regarding free lambda chain concentrations in control sera, our mean value (3.9 ~tg/ml + 1.6) is 2 and 10 times lower than the value reported by Soiling et al. (1982) and Brouwer et al. (1985), respectively.

By this immunoassay, free lambda chains were detected even in normal CSF (mean + SD: 120 + 60 ng/ml), whereas free kappa chains were undetectable in

most controls (detection limit: 25 ng/ml); in sera, levels of free lambda chains (approx. 4 #g/ml) were 2.5-3 times higher than the levels of free kappa chains (approx. 1.5 #g/ml), although the normal kappa/lambda ratio of bound chains to immunoglobulins is about 2. The CSF/serum ratio for free lambda chains was therefore 1/31, whereas the corresponding ratio for free kappa chains was around 1/80. This difference was unexpected since kappa and lambda light chains have similar molecular weights. However, complexes between free kappa (but not lambda) chains and plasma proteins such as al-antitrypsin and prealbumin have been described (Laurell and Thullin, 1974); the occurrence of such complexes could modify the diffusion rate of the light chains across the blood-brain barrier or their immunoreactivity. No definite conclusions may be drawn from the present data.

High levels of free kappa and lambda chains were observed in the CSF from patients with MS but also with other neurological, especially infectious, disorders. Our results disagree with those of Rudick and his coworkers (Rudick et al., 1986a; De Carli et al., 1987) who proposed that high CSF levels of free kappa and lambda chains were specific to MS and acute CNS infections, respectively. On the one hand, this assumption was based on the study of few patients with infectious disorders, on the other, the serum levels of these free light chains were not taken into account for the assessment of a transudation process. As for other proteins, especially immunoglobulins, an increase of the CSF free light chain levels may be the result of either a transudation from blood through an impaired blood-brain barrier or a local synthesis within the CNS. The use of an index for the discrimination between passive transudation and local production rests on the assumption that albumin is a quantitative marker of the barrier permeability and that albumin and the protein under study permeate the disturbed barrier in the same way as the normal barrier, i.e., that the filtration process remained size-dependent and proportional. This assumption is certainly wrong for proteins of high molecular weight, like IgM, in the case of severe blood-brain barrier impairment (Reiber et al., 1987). However, as free light chains are small molecules, we used such indices and compared the results with other well-established markers of a local humoral immune response, as the presence of oligoclonal bands restricted to the CSF.

In MS, high free kappa and lambda chain levels and high indices were significantly associated with either the presence of oligoclonal bands, or an high IgG index. These results clearly support the hypothesis of a local production of both free light chains in up to 86% of MS patients. Oligoclonal bands were observed in the same proportion of MS patients, whereas the IgG index was increased in only 66% of cases. In our hands, the proportion of CSF samples with oligoclonal bands, as determined by agar gel electrophoresis, is about 85% in MS and is not higher when isoelectric focusing is used. Olsson and Nilson (1979) reported similar results in comparing both methods.

Absolute levels of free kappa and lambda chains correlated significantly in CNS infections, but not in MS: their simultaneous increases were likely due to a transudation process from the blood through an impaired blood-brain barrier. We found only a significant correlation between the free lambda index and the IgG index in this group of patients. A local production of free lambda chains has already

been found in viral meningitis (Fryden and Link, 1979) and in subacute sclerosing panencephalitis (Riberi et al., 1975). We found high free kappa levels and indices in 60% and 40%, respectively, of patients with infectious diseases, in discordance with others (Rudick et al., 1986a; De Carli et al., 1987). We considered in this group 45 patients with viral encephalitis and 24 with meningoradiculitis due to B. burgdorferi infections, which are both characterized by a strong immune response within the CNS (Sindic et al., 1985, 1987b). These conditions were not studied by Rudick and coworkers; however, in viral meningitis, no discordance was observed, as only one out of 26 patients had a high kappa index.

In MS, the local production of free kappa chains was initially supported by the observation of double ring formation in single radial immunodiffusion for kappa chain (Iwashita et al., 1974) which was associated with the known increased kappa/lambda chain ratio of CSF IgG (Link and Zetterwall, 1970). Local production of free lambda chains has already been observed by immunofixation and immunoblotting techniques under the form of discrete bands restricted to the CSF (Laurenzi et al., 1980; Vakaet and Thompson, 1985; Bracco et al., 1987). In the study of Bracco et al. (1987), out of 32 MS patients, 15 had both free kappa and lambda chains, ten only lambda chains and three only kappa chains. These electrophoretic studies, like our quantitative data but in discordance with the results of Rudick et al. (1986a), suggest a local production of free lambda chains also in MS.

The diagnostic value of the assays of free light chains in MS is clearly shown by the presence of high levels and indices in CSF samples from MS patients devoid of detectable oligoclonal bands (Table 1). High levels of kappa and lambda chains, and high kappa and lambda indices were observed in 27, 69, 60, and 62% of samples without oligoclonal bands, respectively. However, we have shown that high levels of free lambda but also of free kappa chains are not specific for MS, and that the calculation of an index is required to take in account the blood levels and the transudation process through the blood-brain barrier. Therefore, the detection of CSF oligoclonal bands remains the single best laboratory procedure for the diagnosis of MS.

The physiopathological significance of this local production of free light chains remains obscure. Their presence is not due to local degradation of immunoglobulins, as bands of free gamma heavy chains were never detected by immunoblotting procedures (Bracco et al., 1987), but is more likely a marker of the ongoing intrathecal humoral immune response in MS. Stimulated B cells and plasma cells have been shown to synthesize and release free light chains in vitro (Shapiro et al., 1966). Excess of free light chains has been detected in sera from patients with increased synthesis of immunoglobulins: chronic lymphatic leukaemia and multiple myeloma on the one hand (Soiling et al., 1982), immune disorders as Sj~gren's syndrome on the other (Moutsopoulos et al., 1983). The significant increase of the levels of free lambda chains in sera from patients with Guillain-Barr6 syndrome should be considered in this context. It should be noted that the serum concentrations of free chains were similar in MS and in controls, even if the presence of clonally restricted B cells expressing predominantly kappa light chain determinants

has b e e n r epor t ed in the pe r iphe ra l b l o o d of MS p a t i e n t s ( H a u s e r et al., 1982). In a d d i t i o n to a s t rong ac t iva t ion of cells i n v o l v e d in the h u m o r a l i m m u n e response , o n e m u s t keep also in m i n d the hypo thes i s o f a ' p s e u d o n e o p l a s t i c ' c o n d i t i o n in which B cells local ized w i th in the C N S w o u l d escape n o r m a l r egu la t ion .

Acknowledgements

W e are gra tefu l to Prof. P.L. M a s s o n for r ev iewing this p a p e r a n d to Ms. P. L e m o i n e for her ski l l ful t echn ica l ass is tance . Th i s w o r k was s u p p o r t e d b y g ran t s f rom the ' F o n d s de la Reche rche Sc ien t i f ique M&iica le ' (No. 3.4529.79) a n d f rom the ' G r o u p e beige d ' E t u d e de la Sc l&ose en P laques ' .

References

Axiak, S.A., Krishnamoorthy, L., Guman, J. and Raison, R.L. (1987) Quantitation of free kappa light chains in serum and urine using a monoclonal antibody based inhibition enzyme-linked immunoassay. J. Immunol. Methods 99, 141-147.

Bollengier, F., Lowenthal, A. and Henrotin, W. (1975) Bound and free light chains in subacute sclerosing panencephalitis and multiple sclerosis serum and cerebrospinal fluid. Z. Klin. Chem. Klin. Biochem. 13, 305-310.

Bollengier, F., Rabinovitch, N. and Lowenthal, A. (1978) Oligoclonal immunoglobulms, light chain ratios and free light chains in cerebrospinal fluid and serum from patients affected with various neurological diseases. J. Clin. Chem. Clin. Biochem. 16, 165-173.

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.

Brouwer, J., Otting-van de Ruit, M. and Busking-van der Lely, H. (1985) Estimation of free light chains of immunoglobulins by enzyme immunoassay. Clin. Claim. Acta 150, 267-274.

Collet-Cassart, D., Magnusson, C.G.M. and Ratcliffe, J.G. (1981) Automated particle counting immunoassay for alpha-fetoprotein. Clin. Chem. 27, 64-68.

De Carli, C., Menegus, M.A. and Rudick, R.A. (1987) Free light chains in multiple sclerosis and infections of the CNS. Neurology 37, 1334-1338.

Delpech, B. and Lichtblau, E. (1972) Etude quantitative des immunoglobulins G et de l'albumine du liquide c, tphalo-rachidien. Clin. Claim. Acta 37, 15-23.

Felgenhauer, K. (1982) Differentiation of the humoral immune response in inflammatory diseases of the central nervous system. J. Neurol. 228, 223-237.

Fryden, A. and Link, H. (1979) Predominance of oligoclonal IgG type lambda in CSF in aseptic meningitis. Arch. Neurol. 36, 478-480.

Hauser, S.L., Weiner, H.L. and Ault, K.A. (1982) Clonally restricted B cells in peripheral blood of multiple sclerosis patients: kappa/lambda staining patterns. Ann. Neurol. 11,408-412.

lwashita, H., Grunwald, F. and Bauer, H. (1974) Double ring formation in single radial immunodiffusion for kappa chains in multiple sclerosis cerebrospinal fluid. J. Neurol. 207, 45-52.

Kabat, E.A., Glusman, M. and Knaub, V. (1948) Quantitative estimation of the albumin and gamma globulin in normal and pathologic cerebrospinal fluid by immunochemical methods. Am. J. Med. 4, 653-662.

Kolar, O.J. (1977) Light chains in cerebrospinal fluid in multiple sclerosis. Lancet ii, 1030. Laterre, E.C. (1965) Les prottines du liquide ctphalo-rachidien ~ rttat normal et pathologique. Thesis,

Arsia, Maloine, Paris. LaureU, C.B. and Thulin, E. (1974) Complexes in plasma between light chains, kappa immunoglobulins

and al-antitrypsin respectively prealbumin. Immunochemistry 11,703-709.

Laurenzi, M.A., Mavra, M., Kam-Hansen, S. and Link, H. (1980) Oligoclonal IgG and free light chains in multiple sclerosis demonstrated by thin-layer polyacrylamide gel isoelectric focusing and immunofixation. Ann. Neurol. 8, 241-247.

Link, H. (1987) Immunoglobulin G and low molecular weight proteins in human cerebrospinal fluid. Chemical and immunological characterization with special reference to multiple sclerosis. Acta Neurol. Scand. 43 (Suppl. 28), 1-136.

Link, H. and Zettervall, O. (1970) Multiple sclerosis: disturbed kappa : lambda chain ratio of immunoglobulin G in cerebrospinal fluid. Clin. Exp. Immunol. 6, 435-438.

Lowenthal, A., Vansande, M. and Karcher, D. (1960) The differential diagnosis of neurological diseases by fractionating electrophoretically the CSF gamma-giobulin. J. Neurochem. 6, 51-56.

Masson, P.L., Cambiaso, C.L., Collet-Cassart, D., Magnusson, C.G.M., Richards, C.B. and Sindic, C.J.M. (1981) Particle-counting immunoassay (PACIA). Methods Enzymol. 74(B), 106-139.

Mattson, D.H., Roos, R.P., Hopper, J.E. and Amason, B.G.W. (1982) Light chain composition of CSF oligoclonal IgG bands in multiple sclerosis and subacute sclerosing panenceplialitis. J. Neuroim- munol. 3, 63-76.

Moutsopoulos, H.M., Steinberg, A.D., Fauci, A.S., Lane, H.C. and Papadopoulos, N.M. (1983) High incidence of free monoclonal lambda light chains in the sera of patients with SjSgren's syndrome. J. Immunol. 130, 2663-2665.

Olsson, J.E. and Nilsson, K. (1979) Gamma globulins of CSF and serum in multiple sclerosis: isoelectric focusing on polyacrylamide gel and agar gel electrophoresis. Neurology 29, 1383-1391.

Perini, J.M., Lebas, J., Roussel, P. and Biserte, G. (1979) Evidence for heterogeneous or incomplete immunoglobulins in oligoclonal CSF studied by electroimmunofixation. Clin. Chim. Acta 96, 205-214.

Poser, C.M., Paty, D.N., Scheinberg, L. et al. (1983) New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann. Neuroi. 13, 227-231.

Reiber, H. and Felgenhauer, K. (1987) Protein transfer at the blood cerebrospinal fluid barrier and the quantitation of the humoral immune response within the central nervous system. Clin. Chim. Acta 163, 319-328.

Riberi, M., Bernard, D. and Depieds, R. (1975) Evidence for the presence of lambda chain dimers in cerebrospinal fluid of patients suffering from subacute sclerosing panencephalitis. Clin. Exp. Im- munol. 19, 45-53.

Rudick, R.A., Pallant, A., Bidlack, J.M. and Herndon, R.M. (1986a) Free kappa light chains in multiple sclerosis spinal fluid. Ann. Neurol. 20, 63-69.

Rudick, R.A., Jacobs, L., Kinkel, P.R. and Kinkel, W.R. (1986b) Isolated idiopathic optic neuritis. Analysis of free kappa light chains in cerebrospinal fluid and correlation with nuclear magnetic resonance findings. Arch. Neurol. 43, 456-458.

Shapiro, A.L., Scharff, W.R., Maizel, J.V. and Uhr, J.W. (1966) Synthesis of excess light chains of gamma globulin by rabbit lymph node cells. Nature 211,243-245.

Sindic, C.J.M., Delacroix, D.L., Vaerman, J.P., Laterre, E.C. and Masson, P.L. (1984) Study of lgA in the cerebrospinal fluid of neurological patients with special reference to size, subclass and local production. J. Neuroimmunol. 7, 65-75.

Sindic, C.J.M., Kevers, L., Chalon, M.P., Laterre, E.C. and Masson, P.L. (1985) Monitoring and tentative diagnosis of herpetic encephalitis by protein analysis of cerebrospinal fluid. Particular relevance of the assay of ferritin and S-100. J. Neurol. Sci. 67, 359-369.

Sindic, C.J.M., Boon, L., Chalon, M.P. and Laterre, E.C. (1987a) Intrathecal synthesis of immunoglobulins in MS patients. In: A. Lowenthal and J. Raus (Eds.), Cellular and Humoral Immunological Components of Cerebrospinai Fluid in Multiple Sclerosis, Plenum Publishing, New York, pp. 47-59.

Sindic, C.J.M., Depre, A., Bigaignon, G., Goubau, P.F., Hella, P. and Laterre, C. (1987b) Lymphocytic meningoradiculitis and encephalomyelitis due to Borrelia burgdorferi: a clinical and serological study of 18 cases. J. Neurol. Neurosurg. Psychiatry 50, 1565-1571.

Soiling, K. (1975) Free light chains of immunoglobulins in normal serum and urine determined by radioimmunoassay. Scand. J. Clin. Lab. Invest. 35, 407-412.

Soiling, K., Nielsen, J.L., Soiling, J. and Ellegaard, J. (1982) Free light chains of immunoglobulins in serum from patients with leukaemias and multiple myeloma. Scand. J. Haematol. 28, 309-318.

Tibbling, G., Link, H. and Ohman, S. (1977) Principles of albumin and IgG analyses in neurological disorders. I. Establishment of reference values. Scand. J. Clin. Lab. Invest. 37, 385-390.

Vakaet, A. and Thompson, E.J. (1985) Free light chains in the cerebrospinal fluid: an indicator of recent immunological stimulation. J. Neurol. Neurosurg. Psychiatry 48, 995-998.

Vandvik, B. (1977) Oligocional IgG and free light chains in the cerebrospinal fluid of padents with multiple sclerosis and infectious diseases of the central nervous system. Scand. J. Immunol. 6, 913-922.

Free kappa and lambda light chain levels in the cerebrospinal fluid of patients with multiple...

Documents