Journal ofNeurology, Neurosurgery, and Psychiatry 1990;53:662-666

Intrathecal synthesis of anti-mycobacterialantibodies in patients with tuberculousmeningitis. An immunoblotting study

C J M Sindic, D Boucquey, M P Van Antwerpen, M C Baelden, C Laterre, C Cocito

AbstractCerebrospinal fluid (CSF) and serumsamples from eight patients withbacteriologically proven (6) or clinicallysuspected (2) tuberculous meningitiswere tested for the presence ofanti-mycobacterial IgG antibodies by anaffinity-mediated immunoblot tech-nique. This technique is based onagarose gel isoelectric focusing ofpaired CSF and serum samples dilutedto the same IgG concentration, andtransfer of the specific IgG antibodiesonto mycobacterial antigen-loadednitrocellulose sheets. An intrathecal syn-thesis of anti-mycobacterial oligoclonalIgG antibodies, often superimposed ondiffuse polyclonal production was shownin all patients but not in patients withtension headache or other neurologicaldisorders. Similar results were obtainedwhen a purified mycobacterial antigen,A60, was used for coating thenitrocellulose sheets in place of a wholemycobacterial homogenate, indicatingthat A60 was a major immunogen. Thenumber of anti-mycobacterial oligo-clonal IgG bands increased with time,and persisted for years even in clinicallycured patients. Some IgG bands had nodetectable anti-mycobacterial activity,at least with the antigens preparationsused in this study. The demonstration ofsuch anti-mycobacterial IgG bands in heCSF could be a useful adjunct for thediagnosis of tuberculous meningitis,especially in the case of negative cul-tures.

Despite a decline in the incidence of tuber-culosis in western countries,' tuberculousmeningitis (TM) has not diminished inproportion and remains a serious illness withfatal consequences, if untreated.2 The rate andextent of recovery are related, in part, to therapidity with which an appropriate therapy isinitiated. However, a fast and specific test fordiagnosis is a prerequisite for early initiationof treatment. The demonstration of Mycobac-terium tuberculosis by culture of the cerebro-spinal fluid (CSF) is a lengthy process, whichdelays the diagnosis; moreover, cultures mayremain sterile even in cases proven at nec-ropsy.2 Direct demonstration of tubercularantigens in CSF by radioimmunoassays(RIA)," enzyme-linked immunosorbentassays (ELISA),56 and latex agglutination7 has

been reported, but the sensitivities andspecificities of these techniques have yet to beassessed. In addition, the presence of anti-mycobacterial antibodies and/or solubleimmune complexes could mask such tuber-cular antigens. TM is indeed characterised bythe intrathecal synthesis of oligoclonal IgG`9which, in one case,9 were proved to be direc-ted against mycobacterial antigens. The detec-tion of such anti-mycobacterial antibodies inthe CSF by RIA or ELISA would be a usefuladjunct for early diagnosis of TM.61"12The antigen composition of M tuberculosis

is very complex.'3 Some 30 antigens havebeen detected by a reference system based oncrossed immunoelectrophoresis." The lessmobile polymer in this system is antigen 60(A60), which is the main thermostableimmunogen of both "old tuberculin" and"purified protein derivative" (PPD), knownreagents for cutaneous testing in tubercu-losis.'5 A60 was purified from the cytoplasm ofM bovis.'6 It is a high molecular weight com-plex (106-107 daltons) containing threemoieties of free lipids, lipopolysaccharidesand lipoproteins.'7 This antigen is located inthe bacterial cytosol during the exponentialgrowth phase, accumulates in the walls ofstationary cells, and is later releasedextracellularly.'8 '9 A60 is a powerful immun-ogen inducing primary and secondaryhumoral responses and delayed hypersen-sitivity reactions when injected into mice.20An A60-based ELISA-type immunoassay hasbeen developed for serological analysis oftuberculosis in humans (Baelden et al,unpublished). The use of A60 as a new tuber-culin for cutaneous testing has beenproposed.2'

In this study, an immunoblotting methodwas applied to CSF and paired sera samples ofpatients with TM. Using this test the intra-thecal synthesis of oligoclonal anti-mycobac-terial IgG antibodies was followed. In addi-tion, immunoblotting experiments were per-formed in parallel with whole mycobacterialhomogenates and with the purified antigenA60, to explore the importance of this antigenin the case of intrathecal mycobacterialinfection.

Patients (table 1)We studied eight patients (six females, twomales; seven white, one oriental; age: 141-53years), six of them with bacteriologicallyproven and two with clinically suspected TM.Three patients were also affected by lung

Laboratoire deNeurochimie etService de Neurologie,Universit6 Catholiquede Louvain, Brussels,BelgiumC J M SindicD BoucqueyM P Van AntwerpenC LaterreMicrobiology andGenetics Unit,Institute of CellularPathology, Brussels,BelgiumM C BaeldenC CocitoCorrespondence to:C J M Sindic, Laboratoire deNeurochimie, UCI, 53-59,Avenue Mounier, 1200Brussels, Belgium.Received 11 September 1989and in revised form1 1 December 1989.Accepted 3 January 1990

662

on January 11, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.53.8.662 on 1 August 1990. D

ownloaded from

Intrathecal synthesis of anti-mycobacterial antibodies in patients with tuberculous meningitis. An immunoblotting study

Table 1 Clinical and biological data ofpatients with tuberculous meningitis

LungFirst CSF analysis Microbiological analysis tuberculosis

Sex of CSFPatient Malel Age Protein Glucose (Radiographicnumber Female) (years) (mg/dl) Cells/mm3 (mg/dl) (serum) Staining Culture analysis) Outcome

1 F 1 312 106 (10)* 60 (104) - + - Death2 M 17 122 890 (61) 29 (82) - + + Complete recovery3 F 53 50 800 (15) 44 (89) - + + Complete recovery4 F 14 23 78 (79) 67 (98) - + - Major sequelaet5 F 14 262 900 (35) 11 (72) - + - Slight sequelae6 M 13 5000 1000 (39) 41 (88) + + - Major sequelae7 F 26 220 232 (96) 27 (74) - - - Complete recovery8 F 50 281 455 (96) 15 (78) - - - Complete recovery

*Percentage of lymphocytes in total CSF cells.tMajor sequelae included hydrocephalus, bradypsychia, spastic paraparesis, and tardive syringomyelia.

tuberculosis. The youngest (patient 1) wasinfected by her grandmother and her sisteralso had lung tuberculosis. All patients hadHIV seronegativity. Initial CSF analysisrevealed pleocytosis (mostly lymphocytes) inall cases with high protein content in six andglucose levels below 40 mg/dl in 4. The Zielh-Neelsen staining was positive in one case andthe cultures were positive for M tuberculosis insix. Patient 1 died after being in a comatosestate for six months. In the other cases thestandard anti-tuberculous therapy led to areturn to normal CSF values in cell numberand glucose levels, although an alteration ofthe blood-brain barrier persisted in patients 4and 6. Slight or major sequelae persisted inpatients 4, 5 and 6; clinical recovery wascomplete in four patients (2, 3, 7, 8).

Materials and methodsSamplesTwenty two CSF and serum paired sampleswere collected at various intervals from TMpatients, up to sixty nine months after onset

(table 2). After centrifugation, aliquots of 0 5-2 ml were kept frozen at - 20°C. The totalprotein content of CSF was determined bynephelometry and albumin and IgG in bothCSF and serum by immunonephelometry.

Immunoblotting techniqueThe technique used to identify anti-mycobac-terial antibodies was similar to the immuno-blotting procedure first described by Dorriesand Ter Meulen,22 with slight modi-fications.2324

Agarose gel plates were prepared with0-36 g Isogel Agarose-EF for electrofocusingand 4-3 g sorbitol containing 2 ml ampholine(pH range 3 5-9-5).The nitrocellulose sheet was dipped over-

night into either soluble mycobacterialantigens at a concentration of 300 jug/mlproteins (Lowry method) or A60 at the sameprotein concentration, and washed in Trisbuffer saline (TBS, Tris 20 mM, NaCl500 mM, pH 7 5) containing 0 1% Tween 20(Technicon Diagnostics) for 60 minutes withthree changes.

Table 2 Biological data ofCSF and serum samples studied by immunoblotting

CSF analysisImmunoblottingtPatient Samples Protein Glucoset

number number Time* (mgwdl) (mg/dl) Cells/mm3 CSF Serum1 11 12d 26 52 0 - -

2 15 1900 20 176 (95)§ + -311 26 46 76 4 + _411 41 95 39 1 ++ _511 3m 100 66 10 +++ ++

2 6 60d 143 14 405 (53) + -7 72 185 36 323 (76) + + + +8 85 102 52 176 (87) + -9 108 109 45 125 (63) + -10 12m 42 62 7

3 11 22d 136 14 269 (66) + + +12 31 96 69 1110 (18) + + + +

4 13 69m 158 55 0 +

5 14 47d 270 28 1197 (30) ++ -

6 15 7m 4200 57 83 (99) + + + +16 20 2400 56 10 (96) + + +17 34 2600 63 7 (97) ++ -

7 18 14d 220 29 544 (50) - -19 27 170 20 232 (99) ++ -20 76 195 43 216 (94) + + + ++

8 21 4m 83 50 26 (88) + +22 7 44 53 10 (99) ++ -

*Time after the appearance of clinical symptoms, expressed in days (d) or months (m).tPaired samples of CSF and serum were tested at the same IgG concentration.$Serum glucose levels varied between 68 and 125 mg/dl.§Percentage of lymphocytes in total CSF cells.IlVentricular CSF.

663

on January 11, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.53.8.662 on 1 August 1990. D

ownloaded from

Sindic, Boucquey, Van Antwerpen, Baelden, Laterre, Cocito

Figure I Immunoblots ofanti-mycobacterial specificIgGfrom unconcentratedCSF and thecorresponding serum(SER) diluted to the sameIgG concentration, afterisoelectric focusing inagarose gel and transfer byimmunoaffinity to sheetscoated with totalmycobacterial antigens.Samples 15, 16 and 17from patient 6 (table 2)were collected seven, 20and 34 months after onsetof the disease, respectively.Staining of anti-mycobacterial oligoclonalantibodies is restricted toor more clearly marked inCSF than in thecorresponding serum. Thetotal amounts of IgGapplied on the gel were forthe three paired samples,26, 21 and 31 pgrespectively.

Figure 2 Immunoblots ofanti-mycobacterial specificIgGfrom unconcentratedCSF and thecorresponding serum(SER) diluted to the sameIgG concentration. Thenitrocellulose sheets werecoated with a crudepreparation of cytoplasmicantigens ofM tuberculosisH37Ra (A),or withpurified A60 (B). Lane I:samples 20from patient 7;lane II: samples 17frompatient 6 (table 2); laneIII: control samplesfroma patient with herpeticmeninigitis. The totalamounts of IgTG applied onthe gel were 2-3, 31 and1 2 ,sg respectively.

immunoblotting procedure on nitrocellulosesheet coated with an anti-human IgG anti-serum (Dako).

Each serum was diluted with distito obtain an IgG concentration equof the paired CSF sample. Ten MI o

CSF were then applied side byisoelectrically focused for 60 minutiin a LKB Multiphor Unit. Unprproteins were removed by applyirpaper moistened in PBS on the gtPrecipitated proteins were blottedmycobacterial antigen-loaded nitrsheet under a uniform weight of 1

minutes at 10GC. The immunoblotwashed in TBS-Tween and incubaminutes at 20°C with biotinylated ar

IgG antiserum (Dako Lot 016, CoDenmark) diluted 20-fold in TBS (

0 3% bovine serum albumin (BSIchem). After three 15 minute was}TBS-Tween, the immunoblot was

with the streptavidin-biotin-peroxi(plex (Amersham, United Kingdon400-fold in TBS-BSA. Staininjantigen-antibody bands by 4 chlorthol (Biorad, United Kingdom) w

med by standard technique.In addition the total IgG patte

focused samples was revealed by

Mycobacterial antigensA preparation of unfractionated cytoplasmantigens was obtained by ultrasonication (typeB12 Branson Sonifier) ofM tuberculosis strainH 37 Ra (Difco Laboratories, Detroit, UnitedStates) suspended in Na phosphate buffer 0 1M, containing 0-15 M NaCl, pH 7-6 (PBS):the homogenate was centrifuged at 6,000 g for15 minutes at 4°C, the pellet was discardedand the supernatant stored at - 20°C.The mycobacterial antigen A60 was

prepared from the cytoplasm ofM bovis BCGby exclusion gel chromatography on columnsof Sepharose 6B.'6 A60 was identified by cros-sed immunoelectrophoresis according to areference system'4 on 10% agarose gel slab

illed water using anti-BCG antiserum in the secondial to that dimension. A quantitative determination of)f sera and A60 was made by spectrophotometricside and measurements of its polysaccharide and

es at 10°C protein moieties.'6recipitatedig a filterel surface. Resultsonto the Oligoclonal anti-mycobacterial IgG antibodies

rocellulose restricted to, or more marked in the CSF thankg for 60 in the corresponding serum were detected inwas then samples from the six patients with bacterio-ted for 90 logically provenTM (table 2; figs 1, 2), but alsonti-human from the two patients with clinically suspectedIpenhagen, TM (fig 3, 4, lanes II A and B). In some cases,containing the oligoclonal bands were barely detectable onk, Calbio- a background of polyclonal antibodies but iniings with most cases, the oligoclonal pattern of theseincubated antibodies was very pronounced. As CSF anddase com- paired sera were tested at the same IgG concen-n) diluted tration after appropriate dilutions, the farg of the deeper staining seen in the CSF is an indication0o-1-naph- of an intrathecal synthesis of these antibodies.,as perfor- The occurrence of such a local synthesis was

graded from 0 to + + + by visual inspection ofrn of the the immunoblots (table 2). Uniformly negativea similar results were obtained with samples from

664

.M-mm I, P.7,--

I II..-'t

,

MI 1--':::.,.:..z 4.4.

on January 11, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.53.8.662 on 1 August 1990. D

ownloaded from

Intrathecal synthesis of anti-mycobacterial antibodies in patients with tuberculous meningitis. An immunoblotting study

Figure 3 Immunoblots ofanti-mycobacterial specificIG from unconcentratedCSF and thecorresponding serum(SER) diluted to the sameIgG concentration afterisoelectric focusing inagarose gel and transfer byimmunoaffinity to sheetscoated with totalmycobacterial antigens.Samples 18, 19 and 20from patient 7 (table 2)were collected at days 14,27 and 76 after clinicalonset of the disease and thetotal amounts of IgGapplied on the gel were 35,2-9 and 2-3 pAg,respectively. No reactionwas observed at day 14,whereas two oligoclonalbands of specific anti-mycobacterial IgG werepresent in CSF, but not inthe serum, at day 27. Asimilar oligoclonal patternof specific antibodies, butwith a staining intensityslightly differentfrom onehand to another, waspresent in CSF and inserum at day 76.

Figure 4 Comparison ofall IgG (A: upper panel)and anti-mycobacterialspecific IgG (B: lowerpanel) in CSF and serum(SER) of TM patients.In lanes I (sample 20,patient 7), the asteriskindicates a CSF-restrictedIgG band with nodetectable anti-mycobacterial antibodyspecificity. In contrast,numerous oligoclonal anti-mycobacterial IgG bandswere visualisedfrom thepolyclonal background ofIgG. In lanes IIA andIIB (samples 21 and 22,patient 8) arrows indicateCSF oligoclonal IgGbands with anti-mycobacterial activity (forone band, lane IIb, thisactivity was onlydetectable by visualinspection on theimmunoblot). The totalamounts of IgG applied tothe gel were 0-05 ugfor allsamples in the upperpanel; in the lower panelthese amounts were 2-3 pgin lane I, 0-69 pg in laneIIA and 0 3 pg in laneIIB.

85.

3 0

7.5.

710 -

G 5 -

patient(10), n(10) or(5) (figMos

late insampleafter tlpatientHowevpatient7, fig 31

-..... severe permanent sequelae (therapy was stop-2S- j--; .;S:--o=.:s ;s= SF- ped after 18 months). Although an alteration of

the permeability of the blood-CSF barrier toplasma proteins was evident in both cases andwas likely due to persistent foci of arach-noiditis, glucose level and cell number in theCSF were normal. Intra-thecal synthesis ofanti-mycobacterial antibodies, however, was

..> X no longer detectable in sample 10 which wastaken from patient 2 (table 2), after 12 monthsof successful treatment.

Similar immunoblotting patterns wereobtained when antigen A60 was used for coat-ing the nitrocellulose sheet (fig 2 B) in place ofawhole mycobacterial homogenate (fig 2 A).Indeed, the same oligoclonal IgG bands weredetected at the same isoelectric points withsimilar intensity of staining.

ts with. tension headache (10), sciatica When the complete IgG pattern was com-nultiple sclerosis (10), viral meningitis pared with that of anti-mycobacterialBorrelia burgdorferi meningoradiculitis antibodies (fig 4), it was noted that some IgG2, lane III). bands were not directed against mycobacterial

it samples under study were collected antigens (fig 4, lane I). In addition, it wasthe course of the disease. The earliest possible to detect oligoclonal anti-mycobac-

:s which were collected 12 and 14 days terial IgG antibodies from the polyclonal back-he appearance of clinical symptoms (1, ground of IgG in CSF as well as in thet 1 and 18, patient 7) were negative. corresponding serum. The failure therefore to!er, samples collected at days 15 (2, detect oligoclonal IgG bands did not excludet 1), 22 (11, patient 3) and 27 (19, patient the presence of oligoclonal anti-mycobacterial), that is, at days 6,7 and 14 ofadmission antibodies.

to hospital, were already clearly positive. Whenserial samples were available, the stainingintensity for specific antibodies generallyincreased with time and reached a peak one tothree months after clinical onset. High levelsof CSF anti-mycobacterial antibodies wereobserved in samples 13 (table 2) and 17 (fig 1)from patients 4 and 6: these samples were taken69 and 34 months respectively, after onset. Atthis time these patients were considered curedof the tuberculous infection, but presented

H.

8-5-

8.0-

7.5-

70-

65-

8 5 -

80-

7-5

IA lIB

SE .- S iC. - ri

7:0 -

6-5 -

DiscussionIn normal conditions, CSF IgG originate fromblood as demonstrated by studies in humanswith radiolabelled IgG.25 Consequently, therelative amounts of IgG antibodies directedagainst a given antigen should be identical inboth fluids even in the case of an altered blood-CSF barrier. In contrast, when immune reac-tions occur within the central nervous system,the proportion of locally produced antibodieswill be higher in CSF than in serum. In ourstudy since paired CSF and serum sampleswere tested at the same IgG concentration, theresults ofthe immunoblots ofboth fluids can bedirectly compared.With this technique, we were able to detect a

local synthesis of antimycobacterial antibodiesin all of the patients with bacteriologicallyproven or clinically suspected TM. In the twocases with negative cultures positive immun-oblotting tests were a valid and the solebiological confirmation ofa clinical diagnosis ofTM.

Polyclonal and/or oligoclonal anti-mycobac-terial specific IgG were detected in samplescollected 15 to 27 days after the clinical onset ofmeningitis. Since cultures of M tuberculosisfrom CSF samples may yield visible coloniesonly after six weeks and may also remain sterileeven in cases that have been proved at necrop-sy,2 the demonstration of the intrathecal synth-esis ofanti-mycobacterial antibodies appears tobe a useful tool for diagnosis of TM. Our dataindicate that the synthesis of these antibodiesmay persist for several years (see late samplesfrom patients 4 and 6, who were consideredcured after a standard 18 months treatment).

It should be noted that these two patients

665

AK, W-1

'MI* 14:,.: Z-1.7-

on January 11, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.53.8.662 on 1 August 1990. D

ownloaded from

Sindic, Boucquey, Van Antwerpen, Baelden, Laterre, Cocito

presented severe meningoencephalitis, withmajor sequelae (paraplegia, dysarthria,intellectuel impairment), whereas patient 2(who has completely recovered), had no detec-table intrathecal synthesis of IgG antibodiesone year after the clinical onset of the disease.The long-lasting intrathecal synthesis of anti-mycobacterial antibodies might be due to thepersistence of possibly unviable M tuberculosiswithin the central nervous system. Anotherpossibility is that immunoregulatory mechan-isms for antibody synthesis fail to suppressproliferation and differentiation of B-cellclones after their activation within the brain.Persistence of intrathecal production ofantibodies over long periods of time, as long aseight years, has also been reported in otherinfectious diseases of the nervous system, suchas herpetic encephalitis,26 mumps meningitis27and varicella-zoster meningoencephalitis.' Incontrast, we have observed the spontaneousdisappearance of anti-B burgdorferi oligoclonalIgG in some patients recovering from menin-goradiculitis.29The presence of some oligoclonal IgG bands

restricted to the CSF but devoid of detectableanti-mycobacterial antibody specificity is pos-sibly due to non-specific polyclonal B-cellactivation. The intrathecal production ofantibodies to unrelated agents was observed inpatients with herpetic encephalitis or mumpsmeningitis.26 2730 Another explanation could bethat our antigen preparations do not containsufficient amounts of a minor, but highlyimmunogenic epitope to obtain a transfer byimmunoaffinity on the nitrocellulose sheet.The fact that the immunoblotting patterns

with A60 were similar to those with wholemycobacterial homogenates suggests that most,of the intrathecally synthesised IgG in the caseof TM were specifically directed against theformer antigen.

It should be noted, however, that thestructure of A60 is very complex:'7 besides thelipoglucan, lipopeptidoglycan and lipidmoieties, A60 preparations also contain proteincomponents. After hydrolysis of the proteinmoiety, the lipoglycan part of A60 is stillreactive with anti-BCG antiserum.'7 Thisindicates that both protein and polysaccharidemoieties are responsible for the immunogen-icity of the antigen. A recent ELISA study3'showed that IgG antibody responses to both a14-kilodalton protein and lipoarabinomannanwere immnunodominant in the CSF of TMpatients.

We thank Professor P Evrard, Dr F Zech and Dr E Urbain whoreferred patients 1, 6, 7 and 8. This work was supported by grantnumber 3-4529-79 from the "Fonds de la Recherche Scien-tifique Medicale', Belgium. DB was a research fellow from the"Ministere de la Communaute Francaise-RechercheAppliquee.

1 Clark WC, MetcalfJC, MuhlbauerMS, Dohan FC, Robert-son JH. Mycobacterium tuberculosis Meningitis: a report oftwelve cases and a literature review. Neurosurgery 1986;18:604-10.

2 Ogawa SK, Smith MA, Brennessel DJ, Lowy FD. Tuber-culous meningitis in an urban medical center. Medicine1987;66:317-26.

3 Samuel AM, Ashtekar MD, Ganatra RD. Significance ofcirculating immune complexes in pulmonary tuberculosis.Clin Exp Immunol 1984;58:317-24.

4 Kadival GV, Samuel AM, Mazarelo TBMS, Chaparas SD,Radioimmunoassay for detecting Mycobacterium tuber-culosis antigen in cerebrospinal fluids of patients withtuberculous meningitis. J Inf Dis 1987;155:608-1 1.

5 Sada E, Ruiz-Palacios GM, Lopez-Vidal Y, Ponce de LeonS. Detection of mycobacterial antigens in cerebrospinalfluid of patients with tuberculous meningitis by enzyme-linked immunosorbent assay. Lancet 1983;ii:651-2.

6 Watt G, Zaraspe G, Bautista S, Laughlin LW. Rapiddiagnosis of tuberculous meningitis by using an enzyme-linked immunosorbent assay to detect mycobacterialantigen and antibody in cerebrospinal fluid. J Inf Dis1988;158:681-6.

7 Krambovitis E, Mclllmurray MB, Lock PE, Hendrickse W,Holzel H. Rapid diagnosis of tuberculous meningitis bylatex particle agglutination. Lancet 1984;ii: 1229-31.

8 Laterre EC. Lesproteines du liquide ciphalo-rachidiend P'etatnormal et pathologique. (Thesis) Arscia-Maloine, Paris,1965.

9 Kinnman J, Link H, Fryden A. Characterization ofantibodyactivity in oligoclonal Immunoglobulin G synthesizedwithin the central nervous system in a patient withtuberculous meningitis. J Clin Microbiol 1981;13:30-5.

10 Kalish SB, Radin RC, Levitz D, Zeiss CR, Phair JP. Theenzyme-linked immunosorbent assay method for IgGantibody to Purified Protein derivative in Cerebrospinalfluid ofpatients with tuberculous meningitis. Ann Int Med1983;99:630-3.

11 Chandramuki A, Allen PRJ, Keen M, Ivanyi J. Detection ofmycobacterial antigen and antibodies in the cerebrospinalfluid of patients with tuberculous meningitis. J MedMicrobiol 1985;20:239-47.

12 Ashtekar MD, Dhalla AS, Mazarello TBMS, Samuel AM.A study of Mycobacterium tuberculosis Antigen andAntibody in Cerebrospinal Fluid and Blood in Tuber-culous Meningitis. Clin Immunol Immunopath 1987;45:29-34.

13 DanielTM, Janicki BW. Mycobacterial antigens: a review oftheir isolation, chemistry and immunological properties.Microbiol Revs 1978;42:84-113.

14 Closs 0, Harboe M, Axelsen NH, Bunch-Christensen K,Magnusson M. The antigens of Mycobacterium bovisstrain BCG studied by crossed immunoelectrophoresis: areference system. Scand JImmunol 1980;12:249-63.

15 Harboe M. Antigens ofPPD, old tuberculin and autoclavedMycobacterium bovis BCG studied by crossed immuno-electrophoresis. Am Rev Respir Dis 1981;124:80-7.

16 Cocito C, Vanlinden F. Preparation and properties ofantigen 60 from Mycobacterium bovis BCG. Clin ExpImmunol 1986;66:262-72.

17 Fabre I, L'Homme 0, Bruneteau M, Michel G, Cocito C.Chemical composition of antigen 60 from Mycobacteriumbovis BCG. Scand J Immunol 1986;24:591-602.

18 Cocito C, Vanlinden F. Subcellular localisation andsedimentation behaviour of antigen 60 from Mycobact-erium bovis BCG. Med Microb Immunol 1988a;177:15-25.

19 Cocito C, Vanlinden F. Metabolism of the TMA group ofantigens during the growth cycle of mycobacteria. MedMicrob Immunol 1988b;177:357-67.

20 Cocito C, Baelden MC, Benoit C. Immunological propertiesof antigen 60 of BCG: Induction of humoral and cellularimmunity reactions. Scand J Immunol 1987;25:579-85.

21 Benoit CH, Beschin A, Desmecht M, Dekeyser P, Cocito C.Delayed hypersensitivity reactions by the mycobacterialantigen A60 and cutaneous testing in tuberculosis. MedMicrob Immunol 1989;178:105-12.

22 Dorries R, Ter Meulen V. Detection and identification ofvirus-specific IgG in unconcentrated cerebrospinal fluidby immunoblot technique. JNeuroimmunol 1984;7:77-89.

23 Bukasa KSS, Sindic CJM, Limet JN, Laterre C. Activiteanticorps des IgG oligoclonales du LCR dans les path-ologies neurologiques infectieuses. Detection par"Immunoblotting". Acta Neurol Belg 1988;88:203-20.

24 Bukasa KSS, Sindic CJM, Bodeus M, Burtonboy G, LaterreC, Sonnet J. Anti-HIV antibodies in the CSF of AIDSpatients: a serological and immunoblotting study. JNeurolNeurosurg Psychiatry 1988;51:1063-8.

25 Tourtellotte WW, Potvin AR, Fleming JO, et al. Multiplesclerosis: measurement and validation of central nervoussystem IgG synthesis rate. Neurology 1980;30:240-4.

26 Vandvik B, Skoldenberg B, Forsgren M, Stiernstedt G,Jeansson S, Norrby E. Long-term persistence of intra-thecal virus-specific antibody responses after herpessimplex virus encephalitis. J Neurol 1985;231:307-12.

27 Vandvik B, Nilsen RE, Vartdal F, Norrby E. Mumpsmeningitis: specific and non-specific antibody responsesin the central nervous system. Acta Neurol Scand1982;65:468-84.

28 Vartdal F, Vandvik B, Norrby E. Intrathecal synthesis ofoligoclonal IgG, IgA and IgM virus-specific antibodies ina case of varicella-zoster meningoencephalitis. JNeurolSci 1982;57:121-32.

29 Sindic CJM, DepreA, Bigaignon G, Goubau PF, Hella P,Laterre C. Lymphocytic meningoradiculitis and encep-halomyelitis due to Borrelia burgdorferi: a clinical andserological study of 18 cases. JNeurol Neurosurg Psy-chiatry 1987;50:1565-71.

30 Link H, Laurenzi MA, Fryden A. Viral antibodies inoligoclonal and polyclonal IgG synthesized within thecentral nervous system over the course ofmumps menin-gitis. JNeuroimmunol 1981;1:287-98.

31 Chandramuki A, Bothamley GH, Brennan PJ, Ivanyi J.Levels of antibody to defined antigens of Mycobacteriumtuberculosis in tuberculous meningitis J Clin Microbiol1989;821-5.

666

on January 11, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.53.8.662 on 1 August 1990. D

ownloaded from