expression patientsexpression ofpublic idiotypes in patients with lymearthritis borreliaantigen....

Annals of the Rheumatic Diseases 1993; 52: 199-205

Expression of public idiotypes in patients withLyme arthritis

J S Axford, R A Watts, A A Long, D A Isenberg, A C Steere

AcademicRheumatology Unit,Division ofImmunology,St George's HospitalMedical School,London,United KingdomJ S Axford

Department ofRheumatologyResearch, UniversityCollege and MiddlesexSchool ofMedicine,University ofLondon,United KingdomR A WattsD A IsenbergDivisions ofRheumatology/Immunology, andAllergy, New EnglandMedical Center, TuftsUniversity School ofMedicine, Boston,Massachusetts, USAA Long

A C SteereCorrespondence to:Dr J S Axford,Academic RheumatologyUnit, Division ofImmunology,Department of Cellular andMolecular Sciences,St George's Hospital MedicalSchool, Cranmer Terrace,London SW17 ORE,United Kingdom.Accepted for publication29 October 1992

AbstractObjective Joints are often affected in

Lyme disease and in some instances thismay be due to immune autoreactivity. Tocharacterise further the immune responsein this disease investigations were carriedout to determine the expression of threepublic idiotypes on serum immuno-globulins in patients with Lyme diseaseduring the development of varyingdegrees of arthritis.Methods The expression of idiotypes

(Ids) 16/6, BEG2, and PR4, first identifiedon monoclonal antibodies to DNA, wasdetermined by an enzyme linked immuno-sorbent assay (ELISA) in serial bloodsamples from 12 patients with Lymedisease over a mean period of six yearsduring the development of a variety ofarthritic symptoms, and in serum samplesfrom healthy control subjects and controlsubjects with systemic lupus erythema-tosus.Results Expression of serum IgM or

IgG public Ids 16/6 and BEG2 wassignificantly increased in patients withLyme disease. IgA Id 16/6 expression, incontrast, was significantly increased onlyduring episodes of arthritis and was alsorelated to its severity. IgM and IgG Id 16/6expression was related to their respectivetotal immunoglobulin concentration and,in the case of IgM, to the level of IgMantibodies to Borrelia burgdorferi,whereas similar findings were notapparent with IgA antibodies. This mayindicate that the IgA response is related tothe pathogenesis of arthritis, especially astotal IgA and IgA Id 16/6 levels were foundto increase over the duration of disease.Sequential analysis of antibodies alsoshowed restriction in the expression of Id16/6 as it was never found on all immuno-globulin isotypes at the same time, and IdPR4 was never expressed. Ids 16/6 andBEG2 expression, however, may beassociated as seven patients expressedthese idiotypes simultaneously.

Conclusions These data indicate theuse of public idiotypes in the immuneresponse against B burgdorferi, whichmay be restricted in terms of idiotypeclass and isotype expression, and apossible association between IgAantibodies bearing Id 16/6 with arthritis.

(Ann Rheum Dis 1993; 52: 199-205)

Lyme disease is a multisystem disorder causedby a tick transmitted Gram negativespirochaete Borrelia burgdorferi. The initialmanifestations of the disease are thecharacteristic skin lesion, erythema migrans,and flu-like or meningitis-like symptoms,which may be followed by cardiac,neurological, or joint manifestations.' Jointsymptoms often begin with migratoryarthralgia followed months later by briefattacks of oligoarthritis. In about 10% ofpatients with joint disease chronic arthritisdevelops during the second or third year ofillness, sometimes leading to erosion of thecartilage and bone. The synovial lesion inLyme arthritis is histologically similar to thatfound in rheumatoid arthritis, and chronicLyme arthritis is thought to have anautoimmune basis because of the associationwith the HLA class II DR4 and DR2 genes.The cellular and humoral immune responses inLyme disease develop gradually over a periodofmonths to an increasing array of spirochaetalpolypeptides,2 which may be associated withthe production of a population of anticardio-lipin (aCL) antibodies, possibly as a directresponse to B burgdorfferi phospholipidantigen.3 Interestingly, in most patients Lymearthritis can be successfully treated withantibiotics, but a small percentage of patients,primarily those with HLA-DR4, do notrespond to this treatment.

Idiotypes (Ids) are variable region structureson immunoglobulins which are located at oradjacent to the binding site. Anti-idiotypicantibodies can be used to identify public (crossreactive) and private idiotypic determinants,and their analysis allows the study of thegenetic relation between antibodies. Ids 16/6,4PR4,5 and BEG26 are public Ids that were firstidentified on human monoclonal antibodies toDNA derived from the peripheral bloodlymphocytes of patients with haemolyticanaemia (16/6) and leprosy (PR4), and fromnormal fetal hepatocytes (BEG2). They havesubsequently been found to be expressed onserum and tissue bound immunoglobulins ofpatients with a variety of autoimmunerheumatic diseases. Id 16/6 has also beenidentified on antibodies to K30 in patients withklebsiella infections and on the serumimmunoglobulins of 60% of patients withMycobacterium tuberculosis infection. It has beenshown by mRNA sequencing that the heavychain of the monoclonal antibody bearing 16/6is identical to that of the germline gene VH 26.7It is thought that Id 16/6 is on the heavy chain.

199

on February 2, 2020 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.52.3.199 on 1 March 1993. D

ownloaded from

http://ard.bmj.com/

Axford, Watts, Long, Isenberg, Steere

Thus the public expression of Id 16/6 may wellbe the result of a conserved germline V gene,primarily involved in antibody productionagainst environmental pathogens. It mayalso be associated with pathological auto-antibodies, however, as it has been detected onimmunoglobulins deposited in the kidneys ofpatients with systemic lupus erythematosus(SLE).8We report here the longitudinal expression of

these three public idiotypes in serial serumsamples from 12 patients with erythemamigrans followed by brief or prolongedepisodes of Lyme arthritis.

Patients and methodsPATIENTS AND CONTROL SUBJECTS

We studied 68 serial serum samples in 12patients from New England, USA, who haderythema migrans followed by arthralgia or onebrief attack of arthritis lasting no longer thansix weeks (brief arthritis, four subjects), or byintermittent or chronic arthritis for three to sixyears (prolonged arthritis, eight subjects). Fivepatients had neurological symptoms early intheir illness (four meningitis, one Bell's palsy).The 12 patients had a mean age of 45 years(range 9-71 years), 10 were men and twowomen. They were followed up for a meanduration of six years (range one to nine). Lymedisease was initially diagnosed in these patientsby the presence of erythema migrans before thecausative agent or efficacy of antibiotictreatment was known. It was later shown thatall 12 patients had increased IgG antibodytitres to B burgdorferi. Patients were scored forthe severity of joint swelling as follows: O=noswelling, 1=mild, 2=moderate, and 3=severeswelling.For the investigation of Ids 16/6 and PR4,

serum samples from 23 healthy subjects withno personal or family history of autoimmunedisease were used as healthy controls andserum samples from 15 patients with activeSLE were used as disease controls. In the IdBEG2 assay a different panel of serum samplesfrom 20 healthy subjects was used, again withno personal or family history of autoimmunedisease.

IDIOTYPE ASSAYS

Antibodies against idiotypes 16/6, BEG2, andPR4 were prepared in rabbits, extensivelyadsorbed against human IgG/IgM, and used ata concentration at which there is negligiblebackground binding of irrelevant rabbitpolyclonal antibody, as detailed elsewhere.s6The identification of Ids 16/6 and PR4 onserum immunoglobulins (Ig) was carried outby an enzyme linked immunosorbent assay(ELISA) using the rabbit polyclonal anti-idiotype reagent. Negative and positive controlserum samples were used on each platetogether with the test serum samples. Serumdilution and capture antibody concentrationwere chosen to ensure saturation by serumimmunoglobulins. The following techniqueswere used:

Immulon 1 (Dynatech) microtitre plateswere coated with goat antihuman IgG Fc(Sigma Chemicals, 1 ,ug/ml in borate buffer),which had been previously adsorbed againstnormal rabbit serum for 18 hours at 4°C.Immulon 2 and Nunc plates were similarlycoated and incubated with goat antihumanIgM (Capell, 1 pug/ml in borate buffer) andgoat antihuman IgA (Sigma Chemicals, 0 5pug/ml in borate buffer) respectively. Each wellwas washed and then blocked with phosphatebuffered saline (PBS)-3% bovine serumalbumin (BSA) and then washed five timeswith PBS-0-05% Tween (PBS-T). A 75 [Lvolume of serum, diluted 1:200 in PBS, 3%goat serum, 1% BSA, and 0 05% Tween(Sigma Chemicals), was added in duplicate tothe wells and incubated for two hours at 37°C.Each plate was washed five times with PBS-T.Aliquots (75 Rl) of either rabbit anti-16/6(1/20 000) or rabbit anti-PR4 (1/2000),diluted in 3% goat serum, 1% BSA, and PBS-T, were then added to the wells and incubatedfor two hours at 37°C. Each plate was againwashed five times with PBS-T beforeincubating overnight with 75 RI aliquots ofgoat F(ab')2 antirabbit Ig alkaline phosphataseconjugate (Sigma Chemicals), which hadpreviously been adsorbed against humanserum and diluted in PBS-3% goat serum, 1%BSA, and 0.05% Tween. Each plate wasdeveloped using Sigma 104 alkaline phospha-tase substrate after washing five times withPBS-T, and absorbances were subsequentlymeasured at 405 nm after six hours at roomtemperature. Reference to healthy controlserum samples on each ELISA plate was madeto enable accurate comparison of absorbancereadings obtained from different experimentsand the presence of Ids 16/6 and PR4 wasexpressed as ratios of the average absorbanceobtained for the duplicate Lyme disease andSLE samples against the average absorbanceobtained from the healthy subjects on eachplate.A capture ELISA was used to detect the Id

BEG2 (Q) on Ig in human serum samples.6 IdBEG2 levels were expressed in arbitrary units(100 BEG2 units= 100% binding of standardserum) and this assay does not distinguishbetween IgG or IgM isotypes.

Quantification of the level of expression wasmade by comparison to the number ofstandard deviations (SD) greater than thehealthy control mean value. Antibody levels arereported as positive if levels were greater thanor equal to two SD above the mean.

MEASUREMENT OF SERUM ANTIBODIES TO

B burgdorferi AND TOTAL IMMUNOGLOBULINLEVELSAntibodies to B burgdorferi antigen weremeasured by ELISA as described previously.9Briefly, IgM and IgA antibodies to B burgdorferiwere detected using an immunoglobulincapture technique followed by exposure of theimmunoglobulin to B burgdorfieri antigen andthen detection of antibodies specific to Bburgdorferi with a rabbit antibody against the

200



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Expression ofpublic idiotypes in patients with Lyme arthritis

borrelia antigen. IgG antibodies to B burgdorferiwere detected by their ability to directly bindsonicated B burgdorferi antigen. Antibody to Bburgdorferi values are expressed either as directabsorbances or as titres, or, for IgA, as anabsorbance ratio of a positive standard. IgMtitres ¢ 1/200, IgG titres ¢ 1/400, and IgAratios ¢2-0 were considered increased. Totalimmunoglobulins were measured by standardnephelometry techniques (Beckman assayimmunochemistry system). Normal rangeswere IgM 0A45-1-50, IgG 8&0-15-0, and IgA0-98-3-25 g/l.

MEASUREMENT OF ANTICARDIOLIPINANTIBODIESAnticardiolipin antibodies were detected byELISA as reported previously.'0 Essentially,serum samples were diluted 1/50 in PBS-10%adult bovine serum and assayed in triplicate forIgM aCL and IgG aCL antibodies. Each plateincluded two positive and eight negativecontrol samples and positive results were thosewith absorbances above the mean +2 SD of thenegative controls.

MEASUREMENT OF ANTIBODIES TO B burgdorferiEXPRESSING ID 16/6To determine whether serum antibodies withB burgdorferi antigen binding specificity frompatients with Lyme arthritis expressed Id 16/6,a modification of the Id 16/6 ELISA assay wasused. Serum samples from three patients withprolonged arthritis who expressed Id 16/6 ontheir total serum Ig were selected, togetherwith serum samples from one patient withactive SLE and from one healthy subject. Nuncplates were coated with sonicated B burgdorferi(strain B31, 5 ,ug/ml in carbonate buffer),antibodies captured from the serum sampleswere doubly diluted four times (1/300-1/2400)and Id 16/6 expression determined asdescribed previously.

STATISTICS

Statistics used in analysing the data wereStudent's t test, the Pearson coefficient ofproduct-moment correlation, the standarderror of the mean (SEM), and the standardnormal deviant test was used to compareregression analyses.

ResultsTOTAL SERUM ANTIBODIES, ANTIBODIES TOB burgdoferi, AND ANTICARDIOLIPINANTIBODIESIn patients with prolonged arthritis, the meantotal IgA concentration decreased over the firsttwo years of illness (r=-0-996; p<0 03), butincreased over the subsequent four years(r=0-947; p<002). The reverse was foundwith levels of IgA antibodies to B burgdorferiand a significant reduction in levels(mean(SEM)) was observed between years two(7-05(1-84)) and six (4 02(0 75)) of follow up(p<0.03). In contrast, the mean total serum

IgG and IgM levels decreased over the periodof investigation (IgG r=-0-939, p=0O002; IgMr=-0 831, p<0 02), and though a slightincrease in levels of IgG antibodies to Bburgdorferi was noted over the initial two years,levels of IgG and IgM antibodies to Bburgdorfferi also decreased (IgG r=-0-830,p<005; IgM r=-0-526, p<005).

In patients with brief arthritis, the mean totalIgM levels decreased over the period ofinvestigation (r=-0-987; p<005). Total IgGand A values were always within normal limitsand levels of antibodies to B burgdotfieri of allisotypes were increased and did not fluctuateover the period of investigation.

Anticardiolipin antibody levels wereincreased on one occasion in two asympto-matic patients with Lyme disease with noprevious history of neurological disease - onewith brief arthritis (IgG antibody greater thanthree SD of the negative control mean) and theother with prolonged arthritis (IgM antibodygreater than five SD of the negative controlmean).

ID 16/6 EXPRESSION ON TOTAL SERUMANTIBODIESThe table gives the total mean (range) (SEM)immunoglobulin (Ig) absorbance readingsafter six hours from all experiments .

Levels of IgM and IgG antibodies frompatients with Lyme disease and SLE weresignificantly increased compared with thosefrom healthy control subjects (Lyme diseaseIgM p<001_, IgG p<0 02; SLE IgM p<005,IgG p<0 001).

Expressing the data as a ratio of theabsorbance value obtained from healthysubjects, the levels of IgM and IgG isotypes

Total mean (range) (SEM) immunoglobulin absorbance readings after six hoursfrom experiments conducted to detectidiotype (Id) expression on total serum antibodies

Idiotype Isotype Patients with Lyme disease Patients with systemic Healthy subjectslupus erythematosus

Id 16/6 IgM 162-8(0722)(15-6)* 210-2(23-1200)(74-0)t 92-2(0-325)(16-6)IgG 44 8(0-136)(2 9)t 67-3(6-190)(18-9)§ 33 9(0-63)(3 4)IgA 19-3(3-47)(1-2) 19-1(2-44)(3-0) 14-9(0-45)(2-9)

Id BEG2 IgG/M 243-9(31-611)(17 4) ND 229-6(25-398)(17-6)Id PR4 IgM 333-5(44-1500)(16-8) 396-0(17-1500)(59-0) 302 8(79-704)(26 6)

IgG 143-1(13-1100)(9-8) 140-0(57474)(13-9) 131-6(17-263)(10-0)IgA 109 8(2-305)(6-8) 245-2(21-1000)(42-1)t 146 1(12-314)(16-1)

*p-O.O1.tp-0-02.tps0-05.SP-0.001.ND=Not done.

201



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


NS NS p < O*05p<005

0

0

... ....

... ....

NS NS

... .......................

............

Tr I I I I

Lyme SLE Lyme SLE SLE

IgM Id 16/6 IgG Id 16/6 IgA Id 16l6 IgA Id PR4

Figure 1 Expression of idiotypes (Id) 16/6 and PR4 on serum immunoglobulins in Lymedisease (n=68) and SLE (n=15). Idiotype ratio is the mean absorbance ratio of test/controlserum samples. The solid bars indicate mean values, the broken bars indicate standarddeviations above the control mean (X2=upper bar, X 1=lower bar), and the dots representeach serum sample tested. A significant increase in IgM and IgG Id 16/6 expression wasfound in patients with Lyme disease, but IgA Id 16/6 expression was only increased in thosepatients developing Lyme arthritis.

(mean(SEM)) bearing Id 16/6 were signifi-cantly increased in patients with erythemamigrans followed by brief or prolonged epi-sodes ofLyme arthritis (IgM 2 0(0 2), p<0Q01;IgG 1 3(0- 1), p<0003) compared with healthycontrol subjects, but the mean level of IgAimmunoglobulins bearing this idiotype was notincreased (fig 1). Positive levels of IgM andIgG Id 16/6 antibodies were detected in 5/12and 6/12 patients respectively, primarily duringthe first several years of illness. In contrast,significantly increased levels of IgA Id 16/6antibodies (1-4(0-2); p=005) were found onlyduring the period of joint disease whencompared with healthy control subjects, andpositive levels were detected in three of eightpatients who developed prolonged arthritis. Inthe 15 control patients with SLE, as in thepatients with Lyme disease, the mean level ofIgM and IgG Id 16/6 antibodies wassignificantly increased (IgM 2 6(0 9), p<003;IgG 2-2(0 7), p<0 005). These patients did nothave increased Id 16/6 expression on IgAantibodies, however.

In the eight patients with erythema migransfollowed by prolonged arthritis, IgA Id 16/6expression increased significantly during thefirst (from 1P1(0-1) to 1-7(0-3); p<005) andthird (from 1-2(0-3) to 1-8(0-2); p<0 05) yearsof illness and remained high thereafter (fig 2);levels ofIgG Id 16/6 remained high throughoutthe initial years of illness; and expression ofIgM Id 16/6 was highest during the first yearof illness and generally decreased in eachsubsequent year (r=-0-792; p<005). Incontrast, in the four patients with erythemamigrans followed by brief episodes of arthritis,Id 16/6 expression on all isotypes was highestwhen the skin lesion was present and decreased

thereafter. No patient had Id 16/6 expressionon all isotypes simultaneously, and only onepatient with prolonged arthritis had increasedexpression of this idiotype on each isotype atdifferent times.IgA Id 16/6 expression correlated directly

with the severity of arthritis (r=0-258; p<002),but not with the erythrocyte sedimentation rate(ESR); IgM and IgG Id 16/6 expressioncorrelated with neither the severity of arthritisnor the ESR. The levels of IgM and IgG Id16/6 correlated directly with their respectivetotal immunoglobulin concentrations (IgMr=0-538; IgG r=0-308; in both instancesp<0-001), whereas the levels ofIgA Id 16/6 didnot. Similarly, IgM Id 16/6 expression showeda positive correlation with levels of IgMantibodies to B burgdorferi (r=0-264; p<0 025),but the other antibody isotypes did not. Therewas a significant difference (p<0.05) betweenthe overall positive correlation of IgA Id 16/6expression (r=0-439; p=NS) and the negativecorrelation of IgM Id 16/6 expression withdisease duration.

ID 16/6 EXPRESSION ON ANTIBODIES TOB burgdorferiThe total mean (SEM) absorbance readingsfor the three patients with Lyme disease wereas follows: 1/300, 338-3(84 8); 1/600,253-3(103-6); 1/1200, 173-3(66.9), and1/2400, 135.0(60-7). Id 16/6 expression wasnot detected when serum samples, from thepatient with SLE nor from the healthy subjectwere used.

ID BEG2 EXPRESSION ON TOTAL SERUMANTIBODIESThe table gives the total mean (range) (SEM)absorbance readings from all experiments.When the data were expressed as the

percentage binding relative to a standardserum sample, patients with Lyme disease hadsignificantly increased levels of expression ofIgM and IgG BEG2 when compared withhealthy subjects (37(5) v 20(2); p<0 02).Positive levels of IgM and IgG BEG2expression were detected in 6/12 patients,which included some with brief or prolongedarthritis, and this was predominantly early inthe illness. There was no association betwen IdBEG2 expression and arthritis score, ESR,total immunoglobulin, or levels of antibodiesto B burgdorferi.

ID PR4 EXPRESSION ON TOTAL SERUMANTIBODIESThe table gives the total mean (range) (SEM)absorbance readings for all experiments.The IgA absorbance readings for subjects

with SLE were significantly increased (p<002)compared with healthy control subjects.When the data were expressed as a ratio of

the absorbance value obtained from healthysubjects, Id PR4 expression was again notsignificantly different in the patients with Lymedisease compared with the healthy control

16 7

14 -

12 -

p<0OO1 p<0.03 p<0005pp< 0003

0

coL.0

10 -

8-

6

0

0

0

O

.1-..11.......

t-@@@*@ *@

Lyme SLE

4 -

2 -

0 -

a *

.......-e*T......i..w - w - v w -202



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Expression ofpublic idiotypes in patients with Lyme arthitis

Prolonged arthritis

Control 0 1 2 3 4 5 6

IgG

Ii

Control 0 1 Control 0 1 2 3 4 5 6

Control 0 1 2 3 4 5 6

Years YearsFigure 2 Serum immunoglobulin idiotype (Id) 1616 expression (mean ratio (SEM)), in patients with Lyme disease whodeveloped briefperiods of arthritis (n=4) and prolonged arthritis (n=8), and in healthy controls (n=23). In patients withprolonged arthritis, IgA Id 16/6 expression increased significantly during the first and thirdyear (p<0-05), whereas therewas a significant reduction (p<0-05) in expression ofIgM Id 16/6 with time. There was a significant difference between theincrease and decrease in expression ofIgA and IgM Id 16/6 respectively (p<0 05). In those patients with brief arthritis,Id 16/6 expression on all isotypes tended to reduce.

subjects, but the control patients with SLE hadincreased expression of IgA PR4 (1 7(0-4);p<005) compared with healthy subjects.

COEXPRESSION OF IDS

Seven patients with Lyme disease hadsimultaneously increased expression of Ids16/6 and BEG2. Only one patient withprolonged arthritis had no expression on any ofthe isotypes.

Case reportA 58 year old man had erythema migransfollowed by several brief episodes of arthritisaffecting his wrists, knees, or interphalangealjoints. During the third year of illness hedeveloped chronic arthritis of the right

shoulder and hip, which was treatedsuccessfully with antibiotics during the fifthyear of illness. IgM Id 16/6 expression was

greatest during the first year of illness; itdecreased rapidly and increased before a flareof chronic arthritis in the second year and againduring a period of arthritis in the fifth year

(fig 3). IgG Id 16/6 expression was greatestduring the initial three years of illness. Incontrast, IgA Id 16/6 expression increasedduring the initial episode of arthritis and againduring the fifth year. No change in IgA Id 16/6expression was detected during a flare ofarthritis in the second year of illness. IgM Id16/6 levels correlated with total and Bburgdorfeni specific IgM levels, but the amountsof IgG and IgA Id 16/6 did not correlate withtheir respective total or specific B burgdorferiimmunoglobulin levels. Id 16/6 expression on

Brief arthritis2-3

2-1

1-9'

1*7

1-5

1-3

1-1

0-9

0.7

2-3-

2-1 -

19 -

1-7

15-

13-

1 1

0-9

0-7

I

c

0._aCO)

a,x

0.

0

:5

1-8

1-6

1-4.

1-2

1-0-

0.8

0-6

0.4'

0-2-

0-0

IgM4-5

4-0

3-5

3-0

2-5

2-0

1-5

1-0

0-5

2-8-2-6-

2-4-22-

2-01-8-1-61-4

1-21-0

I

Control 0 10-8fX I-

--- --r 1-

203



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


4-

c

0

0.

cn

x0

Q0.

0

V6

2

0

3

%%IA %

S! %.Jo '. %

IgGIgA

IgM

l\ A0 2-

0,~1I\

0 20 40 60 80 100 120Months

Figure 3 Serum idiotype (Id) 16/6 expression andarthritis score in a 58year old man with chronic Lymearthritis. IgM Id 16/6 expression (A- - -A) reducedfrominitially increased values on presentation with erythemamigrans. IgG Id 16/6 expression (U U) increased to apeak at 30 months of disease and then decreased to those ofthe healthy controls. IgA Id 16/6 expression (0. - -0)peaked at 10 and 50 months, before decreasing to those ofthe healthy controls. Antibiotic treatment, indicated by thearrow, was associated with a simultaneous decrease inId 16/6 expression on all isotypes.

all isotypes decreased simultaneously aftertreatment with antibiotics.

Similar findings were also found in the otherseven patients investigated, although thispatient was one of the more extensivelyinvestigated and had one of the longer followup times.

DiscussionData are presented that characterise theexpression of public Ids in a cohort of patientswith Lyme disease who developed varyingdegrees of arthritis, and this offers a uniqueopportunity to observe the humoral responseto B burgdorferi in patients not treated withantibiotics.

In Lyme disease a significant increase in theexpression of serum IgM or IgG public Ids16/6 and BEG2 was shown and positive levelsdetected (greater than two SD above the meanfor healthy control subjects) in halfthe patientsinvestigated at some stage of illness. Incontrast, IgA Id 16/6 expression was increasedsignificantly only during episodes of arthritisand positive levels were detected in three of theeight patients with prolonged arthritis. IgA Id16/6 expression was also related to the severityof arthritis, but not the acute phase response,whereas IgM and IgG expression were not.Interestingly, IgM and IgG Id 16/6 expressioncorrelated with their respective total immuno-globulin concentrations and, for IgM, to thelevel of IgM antibodies to B burgdorferi,

whereas similar findings were not found withIgA antibodies. Id 16/6 bearing antibodies bindto the B burgdorferi antigen, as wassubsequently confirmed using B burgdorferisonicate in a capture ELISA. These dataindicate that the IgM and IgG production isprobably the major response against Bburgdorferi and perhaps IgA Id 16/6 expressionis involved with more subtle factors relating tothe pathogenesis of arthritis; perhapsassociated with specific antibodies to Bburgdorferi, for example, against the outersurface protein A.

Sequential antibody studies showed thatIgM and IgG Id 16/6 expression tended todecrease during prolonged illness, whereas IgAId 16/6 expression tended to increase.Furthermore, the difference between IgM andIgA expression was significant, which mayindicate that isotype switching is occurring. Inthis respect it is pertinent that the humoralresponse against outer surface protein A inLyme arthritis occurs late in the disease'1 andin the patients reported here total IgA levelsincreased as the disease progressed. Noevidence was found to suggest an associationbetween the production of aCL antibodies andarthritis, indicating that the Id 16/6 bearingantibodies do not also have cardiolipinspecificity. Comparable findings were observedin the patient reported, though the effects ofantibiotic treatment seemed to correct thesechanges, giving further weight to the dataindicating that Id 16/6 is expressed onantibodies to B burgdorferi. These changes werenot apparent if the illness was brief, though thepatients were not investigated for the sameperiod of time.

Restriction in the expression of Id 16/6 wasalso observed. Simultaneous expression of Id16/6 on all isotypes was never found and onlyone patient had increased expression of thisidiotype on each isotype at different diseasestages. Restriction was also found when Id PR4was examined, as this was never expressed inLyme disease; however, Ids 16/6 and BEG2seem to be associated with a similar repertoireof antibodies to B burgdorferi as seven patientsexpressed Id 16/6 simultaneously with IdBEG2. These Ids are not associated with thesame V gene family, however; Id 16/6 and IdBEG2 are VH3 and VH4 associatedrespectively. By using Ids 16/6 and BEG2 asmarkers ofVH gene usage, an impression of thecomplexity of the humoral response to the Bburgdorferi bacterium can be seen. It is evidentthat the response is far from static as these Idswere not expressed continuously, but as aseemingly well orchestrated response with Idsappearing on different isotypes at differenttimes. It is interesting that there is a similarexpression of Ids 16/6 and BEG2 in thepatients with SLE studied, but with the impor-tant exception that Id PR4 rather than Id 16/6was expressed on IgA. Perhaps these subtle Idfluctuations have clinical associations.So does Id 16/6 play a part in the patho-

genesis of the disease? These data suggest thatit does. It may simply be a marker of infectionrelated antibodies or it may be a marker of a

204

1



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Expression ofpublic idiotypes in patients with Lyme arthritis

component of an idiotype network in which acascade of interrelating antibodies is generated,any one of which may be pathogenic."2 Forexample, IgA bearing 16/6 may be antibody 3(anti-Ab2) in a parallel set of Id 16/6 positiveand DNA binding negative autoantibodies,13which may have antigen binding specificity tothe B burgdorferi antigen due to the presence ofan internal image of the B burgdorfieri antigenoccurring at the V region of antibody 2 (anti-Abl - that is, the B burgdorferi antibody)."4After processing the 16/6 idiotope may be seenby T cells, which, in the presence of certainHIA class II gene products (DR4, forexample), cause them to become autoreactiveand arthritogenic.'5 Subsequent B cellactivation may follow, which could result inpertubation of the antibody response.

In summary, these data indicate restrictionin the expression of some public idiotypes inLyme disease and a possible associationbetween IgA antibodies bearing Id 16/6 andthe development of arthritis. The associationswith infection shown here further strengthenthe suggestion that the main reason for theconservation of the public idiotypes is not toform part of an autoantibody response, but tohelp protect the species against environmentalpathogens.

The authors thank Dr R S Schwartz for critical comment andfor allowing us to use idiotype reagents prepared in hislaboratory. We also thank Dr A Rabson for measuring immuno-globulin levels and for reviewing the manuscript, Dr L Tuckerfor developing the Id 16/6 ELISA protocol, Dr A Puccetti forhelpful advice while carrying out this investigation, Ambrin Darfor technical assistance, and Professor I M Roitt and Mrs AAlavi for constructive criticism of the manuscript. JSA was aFulbright Scholar and thanks the Fulbright Commission andthe Peel Medical Research Trust for financial assistance. RAWwas supported by the Wellcome Foundation and the Emily LeRossignol Fund. AAL was the recipient of a fellowship in basicsciences from the Canadian Arthritis Society and is a Sandosscholar in medicine. ACS was supported by public healthservice grants AR-20358 and AR-40576. The anti-PR4 reagentwas kindly donated by DrW Williams.

1 Steere A C. Lyme disease. N Engl J Med 1989; 321:586-96.

2 Grodzicki R L, Steere A C. Comparison of immunoblottingand indirect enzyme-linked immunosorbent assay usingdifferent antigen preparations for diagnosing early Lymedisease. Infect Dis 1988; 157: 790-7.

3 Mackworth-Young C G, Harris E N, Steere A C, et al.Anticardiolipin antibodies in Lyme disease. ArthritisRheum 1988; 31: 1052-6.

4 Shoenfeld Y, Isenberg D A, Rauch J, Madaio M P, StollarB D, Schwartz R S. Idiotypic cross-reactions ofmonoclonal human lupus autoantibodies. J Exp Med1983; 158: 718-30.

5 Williams W, Zumla A, Behrens R, et al. Studies of acommon idiotype PR4 in autoimmune rheumatic disease.Arthritis Rheum 1988; 31: 1097-104.

6 Watts R A, Ravirajan C T, Staines N A, Isenberg D A. Ahuman fetal monoclonal DNA-binding antibody sharesidiotypes with fetal and adult murine monoclonal DNA-binding antibodies. Immunology 1990; 69: 348-54.

7 Chen P P, Liu M R, Sinha S, Carson D A. A 16/6 idiotype-positive anti-DNA antibody is encoded by a conserved VHgene with no somatic mutation. Arthritis Rheum 1988; 31:1429-31.

8 Isenberg D A, Collins C. Detection of cross-reactive anti-DNA antibody idiotypes on renal tissue bound immuno-globulins from lupus patients. J Clin Invest 1985; 67:287-94.

9 Beradi V P, Weeks K E, Steere A C. Serodiagnosis of earlyLyme disease: analysis of IgM and IgG antibodyresponses by using an antibody-capture enzyme immuno-assay.7InfectDis 1988; 158: 754-60.

10 Long A A, Ginsberg J S, Brill-Edwards P, et al. Therelationship of antiphospholipid antibodies tothromboembolic disease in systemic lupus erthematosus:a cross sectional study. Thromb Haemost 1991; 66:520-4.

11 Kalish R, Leong J, Steere A C. Delay in the immuneresponse to outer-surface proteins (Osp) A and B ofBorrelia burgdorferi: correlation with arthritis andtreatment failure in susceptible individuals. ArthritisRheum. 1991; 34: S43.

12 Watts R A, Isenberg D A. Idiotypes and anti-idiotypes: whatare they trying to tell us? Ann Rheum Dis 1988; 47:705-7.

13 Naparstek Y, Mackworth-Young C, Breitrnan L, SchwartzR. Sequential anti-idiotypes define reciprocal idiotopes onthe same anti-DNA antibody. Clin Immun Immunopathol1989; 50: S106-16.

14 Puccetti A, Koizumi T, Migliorini P, Andre-Schwartz J,Barrett K J, Schwartz R S. An immunoglobulin light chainfrom a lupus-prone mouse induces autoantibodies innormal mice.J ExpMed 1990; 171: 1919-30.

15 Van-Eden W, Holoshitz J, Nevo Z, Frankel A, Klajman A,Cohen I R. Arthritis induced by a T-lymphocyte clonethat responds to mycobacterium tuberculosis and tocartilage proteoglycan. Proc Natl Acad Sci USA 1985; 82:5117-20.

205



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

expression patientsexpression ofpublic idiotypes in patients with lymearthritis borreliaantigen....

Documents