Primarily chronic progressive and relapsing/remitting multiple sclerosis: Two immunogenetically distinct disease entities

Vol. 86, pp. 7113-7117, September 1989 Immunology

Primarily chronic progressive and relapsing/remitting multiple

sclerosis: Two immunogenetically distinct disease entities

(HLA class IIgenes/HLA-DR/HLA-DQ/restrictionfragment length polymorphism analysis)

OLLE

OLERUP*tf,

JAN HILLERT*§,

STEN

FREDRIKSON§, TOMAS OLSSON§, SLAVENKA

KAM-HANSEN§,

ERNA

MOLLERt, BJORN CARLSSON*t,

AND

JOHAN WALLIN*t

*Center for Biotechnology,tDepartmentof ClinicalImmunology,and§DepartmentofNeurology, Karolinska Institute at Huddinge Hospital, S-141 86

Huddinge,Sweden

Communicated byBarujBenacerraf,June8, 1989

ABSTRACT HLA class IIgene polymorphism was inves-tigated in 100 patients with clinically definite multiple sclerosis (MS) by restriction fragmentlength polymorphism analysis of TaqI-digested DNA using DRB, DQA, and DQB cDNA probes.

Twenty-sixpatients had primarily chronic progressive MS and 74 had relapsing/remitting MS. The latter group included patients with a secondaryprogressiveevolution of symptoms. Both clinical forms of MS were found to be associated with the DRw15,DQw6 haplotpe. In addition, primarily chronic pro-gressive MS was positivelyassociated withtheDQBlrestriction fragmentpattern seeninDR4,DQw8, DR7,DQw9, and DRw8, DQw4haplotypes,as well as

negatively

associated with the Taq IDQBIallelic pattern corresponding to the serological speci-ficity DQw7. Relapsing/remitting MS was positively associ-ated with the DQBI allelic pattern observed in the

DRw17,DQw2 haplotype. These three DQB1 alleles are in strong negative linkage disequilibria with DRw15. The two susceptibility markersof each clinical form of MS act additively in determining the genetic susceptibility, as therelative risks for individualscarrying both markers roughly equal thesumof respective risks. Different alleles of theDQB1locusdefinedby restriction fragment lengthpolymorphisms contribute to

sus-ceptibilityand resistance toprimarily chronic progressive MS

as well as to

susceptibility

to

relapsing/remitting

MS. The observed immunogenetic heterogeneity between the different clinical forms of MS favors the hypothesis that primarily chronicprogressive MS andrelapsing/remittingMS are two distinctdiseaseentities.

Withthefinding ofastrongassociation betweenankylosing spondylitisand the HLAclassIantigenB27(1, 2),anew era

in the search forgeneticmarkers in humandiseasesbegan, and various HLA antigens have been demonstrated to be associated withsusceptibilityto alarge number of different diseases (for review,seeref. 3). Despite intense effortsthe mechanismsinvolved in theassociations betweenHLA

an-tigensanddiseases havenotbeenunraveled.

Multiple sclerosis (MS) is achronic inflammatory demy-elinating disease ofthecentralnervoussystemoften causing neurological deficits.Theetiologyaswellasmany aspectsof

the pathogenesis remain unclear. However, most current hypotheses postulate that the disease is caused by a T-cell-mediated autoimmune destruction ofthe myelinsheath

triggered byanenvironmentalinfluence, whichmaybeviral,

in thegenetically predisposed individual.

MostpatientswithMShaveadisease coursecharacterized by relapses andremissions, termed

relapsing/remitting

MS

(R/R

MS).Manyofthesepatients will, however, eventually enter aphasewithsecondaryprogressive evolution of

symp-toms (4). Approximately, 15-40%o of the patients with MS have aprimarily chronic progressive diseaseform(PCP MS)

(4, 5);i.e.,thediseasehas acontinuous progressive evolution

from onset. It has beenproposedthat PCP MS and R/R MS may be separate disease entities (6) since they differ with respect to epidemiology (6), age at onset (4, 5, 7), initial symptoms (4, 6),prognosis withregard to degreeofdisability (4, 7) andtomortality(8), and response to immunosuppres-sion(9). Effortstodifferentiate between PCP MSand R/R

MS by serological HLA class II typing have yieldeddivergent andinconclusiveresults(10-12).

Theextremely polymorphicHLAclassIIgenesbelongto theimmunoglobulingenesuperfamily andarelocatedonthe

short arm of chromosome 6 with three major subregions encoding expressed molecules-DR, DQ, and DP. HLA

class IIantigens aretransmembraneousglycoproteins com-posed ofa heavy achain anda light

A

chain. Mostof the polymorphism is found within discrete diversity regions of

themembrane-distal domain, encoded by thesecond exon,

especially of the 3 chains (for review, see ref. 13). HLA

antigens serve as markers for self during the ontogenetic development ofthe T-cellrepertoire.HLAclassIImolecules formacomplexwithimmunogenicdeterminants, whichcan be recognized by the T-cell receptor. Thus, the antigen-specificT-cell responseis restrictedbythemajor

histocom-patibility

complex.

Restrictionfragment lengthpolymorphism (RFLP) analy-sis ofHLA-DR and -DQ genes has been developed intoa

powerful

toolfor genomic tissue typing (14-17). The poly-morphism detected byRFLPanalysis correlateswell with the

phenotypic polymorphism recognized by serologyand mixed

lymphocyte culture (ref. 16; for review, seeref. 18).

Addi-tional biologically relevant polymorphism is identified by

RFLPanalysiscompared toconventional serological tissue typing (ref. 16; for review, see ref. 18). Due to the strong

linkage disequilibrium between DR and DQ subregions, DR-DQ haplotypescanbeassigned by merelyinvestigating

theindividual withoutfamilydata(16-18).

Inapreliminary studywehavefound(19)thatPCP MSwas

associated with a specific heterozygous Taq I HLA-DQB restrictionfragmentpattern. Toidentifyimmunogenetic

het-erogenetiesbetween the twoclinicalformsofMS,PCP MS and

R/R

MS,

RFLPanalysis ofHLA-DRand-DQgeneswas performedinalargenumberofclinicallywellcharacterized MS patients.

Abbreviations: RR, relativerisk; MS, multiple sclerosis;PCPMS, primarilychronicprogressive MS;R/RMS, relapsing/remittingMS,

including patientswithasecondary progressive phase; RFLP,

re-strictionfragment length polymorphism; Pcorr58,etc., Pcorrected byafactor of58,etc.;EF, etiologic fraction.

tTowhomreprintrequestsshouldbe addressed.

7113

Thepublicationcostsof thisarticleweredefrayed in part by page charge payment.Thisarticlemustthereforebeherebymarked"advertisement"

MATERIALS AND METHODS

Patients. One hundred unrelated Swedish patients with clinically definite MS (20) were studied. Twenty-sixpatients had PCPMS-i.e., graduallyprogressive neurological symp-tomsfrom the onset of diseasewithout remissions.

Seventy-four patients had R/R MS. This latter group includedpatients who experienced secondaryprogressiveevolution of symp-tomsafter initialrelapses and remissions. Clinical and

cere-brospinal fluid data of thepatientsaregiveninTable 1. Controls. One hundred healthy unrelated Swedes,

HLA-DR and -DQ typed by Taq I RFLP analysis,were used as

controls.

Probes. The following almost full-length cDNA probes

were used: DRB, a790-base-pair(bp) HindIII-Sac Ifragment

of clone

pII-,3-3

(23); DQA,a584-bpRsaI-Stu Ifragmentof clone pII-a-5 (24); DQB, a627-bpAva I-Ava Ifragment of

clone

pll-,8-1

(25). Purified insertswere labeled byrandom hexanucleotidepriming (Pharmacia)to a specific activity of

1-2 x 109

cpm/hg.

RFLPAnalysis of HLA-DRB,-DQA, and -DQB Genes. DNA was extractedfrom peripheral blood leukocytes byphenol/ chloroform extraction ofproteinaseK-treated nuclei in mi-croscale. Samplesof

8,ug

of DNAweredigestedwith 2units of Taq I per

gg

of DNA (Boehringer Mannheim) in 20-/A

reaction mixtures for4 hr at 650C. Totalreaction mixtures were loaded on 0.7%agarose gels (20 x 20 cm). Gels were electrophoresed for 20 hrat40 Vin 0.5 x TBEbuffer (89mM

Trisborate/89mMboricacid/2mMEDTA, pH8.0). Dena-tured DNAfragmentswere transferredbycapillary blotting

with 20x SSC (3 M NaCl/0.3M sodiumcitrate, pH 7.0) to

nylon membranes (Pall) and cross-linkedby UV-irradiation for 3.5 min. After prehybridization for 2 hr at 420C

[50%

(vol/vol)formamide/10%

(wt/vol)

dextran

sulfate/lOx

Den-hardt's solution/50 mM Tris-HCl, pH

7.5/1

M

NaCl/1%

SDS/0.1% sodium

pyrophosphate/denatured

sheared salmon sperm DNA(150

,ug/ml)],

a32P-labeled cDNAprobe

wasadded(15-20 x106cpm permembrane). (1xDenhardt's solution = 0.02%

polyvinylpyrrolidone/0.02%

Ficoll/0.02%

bovineserumalbumin.)Filterswerehybridizedfor 12-18 hr

at 42°C. Two 5-min washes at room temperature in 2x

SSC/0.5%SDSwerefollowedbytwo30-minhigh-stringency

washes at 60°C in 0.lx SSC/0.1% SDS. Membranes were autoradiographed (X-Omat AR5, Kodak) with intensifying screens for 2-5 days at -70°C, and RFLP patterns were analyzed (16-18).

Nomenclature. Toavoid local Taq I RFLPnomenclature,

theassociated DRand DQ serological specificities (16, 18) are given whenever possible, using the nomenclature

sug-gested byTheNomenclature Committee onLeucocyte

An-tigens1987 (26). Allelic Taq I DRB, DQA, andDQBRFLP patterns frequently referred to in the text are described in

Table 1. Clinical andcerebrospinalfluid data of MSpatients

PCPMS R/RMS

Male/femaleratio 9:17 22:52

Ageatonset, years 40(20-56) 33(13-55) Duration ofdisease,years 8 (2-22) 9 (1-33) CSF

IgGindex -0.7* 23/26 54/73

Oligoclonal IgGbandst 24/26 68/73

Mononuclearpleocytosist 8/26 35/73

Twenty-six PCP MS patients and 74 R/R MS patients were

studied. Forage at onset andduration ofdisease,dataare meanand range(inparentheses).Cerebrospinalfluid(CSF)datawereavailable on99patientsand areexpressedas no.patients/no. totalpatients. *ReflectsIgG synthesiswithin the centralnervous system(21). tDemonstrated by agarose isoelectric focusing and avidin-biotin

peroxidaseimmunolabeling (22). t:5-x 106 cells per liter.

Table 2. TaqI DRBandDQBRFLPpatternspositivelyand

negativelyassociated with the different clinical forms ofMS areillustrated inFig. 1.

Statistics. Datawere

analyzed by

thex2test or

by

Fisher's

exacttest, when thecriteria of thex2test were notfulfilled. When

looking

for other HLA class IIassociationsthan the well-established DR2 association, gene

frequencies

of pa-tients and controls were

compared

after subtraction of T-DRB II,T-DQB II

(DRwl5,DQw6)-positive

haplotypes.

P

values of non-DR2 associationswerecorrected(byafactor of

29, Pcorr29)

for thetotalnumberofTaqIHLA-DRB

(n

= 16), -DQA (n=5),and-DQB(n=8) allelicpatternsidentified

inCaucasiansor werecorrected

(by

afactor of

19,

Pcorr19)

for the number ofTaqIDRB,DQA,andDQB

haplotypes (n

= 19) found in the healthy controls. P values were also corrected(byafactor of2,Pcorr2)for

dividing

the

patients

intotwogroups.Relative risks(RRs)and

etiologic

fractions (EF)werecalculatedbymethodsdescribedinrefs. 27 and 28.

RESULTS

AllMSPatients VersusHealthyControls.

Sixty

percentof

patients

(n = 100)

compared

to 30% of controls had the T-DRB II pattern, which

corresponds

to the

serological

DR2-split

DRw15[P<0.00005; RR,

3.5; EF,

0.43]

(Table 2;

Fig.

1, lane 1).

Sixty-nine

percentof

patients

compared

to

50% of controlswere T-DQB

II-positive,

which isassociated

withthe

serological DQwl-split DQw6

(P<

0.005;

RR,

2.2;

EF, 0.38)

(Table

2;

Fig.

1, lane 2). The T-DQB IIpattern

(DQw6)

is seen in

DRwl5,DQw6,Dw2

and

DRw13,DQw6,

Dw18 haplotypes as well as in some very rare

haplotypes.

The

primary

RFLP-defined associationin all MS

patients

was

withthe T-DRB II pattern

(DRw1S)

and notwith

T-DQB

II

(DQw6)

(P<

0.001).

PCP MS PatientsVersusHealthyControls. Theassociation

of PCP MS

(n

=

26)

withT-DRBII

(DRw15)

wasweaker than the association found in all MS

patients,

54% of

patients

compared

to 30% ofcontrols

(Pcorr2

<

0.05; RR,

2.7;

EF,

0.34)

(Fig. 1,

lane

1).

No

significant

association with

T-DQB

II

(DQw6)

wasobserved

(Fig.

1,lane 2).

Sixty-five

percentof PCP MS

patients

compared

to38%of controls had theT-DQBIVpattern

(Pcorr58

<

0.05; RR,

3.1;

EF, 0.44)

(Table

2;

Fig.

1, lane4). T-DQBIVinCaucasians

is found in

DR4,DQw8, DR7,DQw9,

and

DRw8,DQw4

hap-lotypes (Table 2). Most of T-DQB

IV-positive

patients

be-longed

to the

DR4,DQw8

haplotype-42%

of all PCP MS

patients compared

to 24% of controls

(Pcorr38

<

0.05; RR,

2.3; EF, 0.24). Noneof the 26

patients

compared

to29% of

controls were T-DQB

V-positive,

which is associated with

DQw7

(Pcorr58 <

0.05;

RR,

0.06)

(Table 2;

Fig. 1,

lane

5).

Consequently,

none of the

DR4-positive haplotypes

in PCP MS

compared

to31%of

DR4-positive haplotypes

incontrols

had the T-DQB V

(DQw7)

pattern

(P

<

0.00001).

The combined occurrence of T-DRB II

(DRwl5)

and

T-DQBIV

(DQw8, DQw9,

and

DQw4)

wasobserved in31%

of

patients

with PCP MS

compared

to 6% ofcontrols

(P

<

0.001; RR,

7.0).

Eighty-eight

percent of

patients

exhibited

one or both

susceptibility

markers

compared

to 62% of controls

(P

<

0.05; EF, 0.70).

R/R

MS Patients VersusHealthyControls.The

strength

of theassociationwith T-DRB II

(DRwlS)

in

R/R

MS

(n

=

74)

wasof thesame

magnitude

astheassociation found in all MS

patients-62%

of

patients compared

to 30% of controls

(Pcorr2

<

0.00005;

RR,

3.8; EF,

0.46)

(Fig.

1,

lane

1).

Furthermore,

68% of

R/R

MS

patients

were

T-DQB

II

(DQw6)-positive compared

to50%of controls

(Pcorr2

<

0.05;

RR,

2.2; EF, 0.37)

(Fig. 1,

lane

2).

In

R/R

MS,

asforall MS

patients,

the

primary

RFLP-definedassociationwaswith the

T-DRB II

(DRwlS)

allelic pattern and not with

T-DQB

II

(DQw6)

pattern (P< 0.0005).

Table 2. AllelicTaq I DRB, DQA,and DQB RFLP patterns referred to in the text and the associated serological specificities

Taq Irestrictionfragments, kb Assoc. ser. spec.*

DRB DQA DQB DRB DQA DQB DR DQ II H II 12, 2.0, 1.7, 1.4 6.5 2.5, 2.3, 1.8 DRw15 DQw6 IV IV III 11, 8.0, 4.1, 1.2 4.5 2.4, 1.5 DRw17 DQw2 VIII IV III 12, 8.0, 4.1, 1.2 4.5 2.4, 1.5 DRw17 DQw2 XII V IV 14, 5.7, 5.1, 2.7, 2.6 5.7 2.2, 2.0, 1.7 DR4 DQw8 XIII V IV 14, 5.7, 4.0, 2.7, 2.6 5.7 2.2, 2.0, 1.7 DR7 DQw9 V II IV 10,1.2 6.5 2.2, 2.0,1.7 DRw8 DQw4 - - V 2.1,1.4 DQw7

ForDRB, DQA, and DQB, local Taq Inomenclature is in Roman numerals (16). kb, Kilobases; assoc. ser. spec., associatedserological specificities.

*Data arefrom refs. 16 and 18 andO.O., unpublished data.

Thirty-two percent of R/R MS patients had the T-DQB III pattern,which isseenintheDRw17,DQw2haplotype, com-pared to15%of controls(Pcorr58<0.005; RR, 2.7;EF, 0.20)

go co 0 co cc CY Or a CY a az a a a l L ~ L P 12 2.3 2.0 1.4~ lane All MS n=100 PCP MS n=26 R/R MS n=74 Controls n- 100 2.4m 2.2 ,2.1> 2.0 1.5> 1.4> t 2 3 60%* 69% 54%* 73% 62% * 68%* 30% 50% 4 5 25% 37% 4% 65% 32%** 27% 15% 38% 14% 0%* 19% 29%

FIG. 1. TaqIDRBandDQB allelic restrictionfragmentpatterns

positivelyandnegativelyassociatedwith allMS,PCPMS, and R/R MS. Lanes: 1, Taq I-digested DNA hybridized withaDRBcDNA

probe; 2-5, Taq I-digested DNA hybridized with aDQB cDNA

probe,asindicated. To the left of eachlane, the restrictionfragments

characteristic fortheallelic patternare indicatedinkilobases. All

statisticalcomparisonsweremade withthehealthy controls.

Signif-icantdifferencesafter correctionformultiple comparisonsare

indi-cated:*,P< 0.05;**,P< 0.01;***,P<0.001.

(Table2; Fig. 1, lane 3). No negative (i.e., protective) HLA class II associationwasfound in this group of patients.

The combined occurrence of T-DRB II (DRw15) and T-DQB III(DRw17,DQw2) was found in19%of patients with

R/RMScompared to4%of controls (P<0.005; RR, 5.6).

Seventy-sevenpercentof patients had one or both

suscep-tibilitymarkers compared to 41%of controls (P< 0.00001;

EF,0.61).

PCP MS VersusR/RMS. PCP MS as well asR/RMSwere associated with T-DRB II (DRw15).

Thefrequency ofT-DQB IV(DQw8, DQw9, andDQw4), positively associated with PCPMS, is reduced in R/RMS

compared toPCP MS (P <0.001)(Table 3). Likewise, the occurrenceofT-DQB III(DRw17,DQw2) positively

associ-atedwithR/R MS is reduced inPCPMS comparedtoR/R MS (P< 0.005)(Table 3).

DISCUSSION

The familial occurrenceofMS with a tendencyofincreasing prevalence withincreasingdegreeof kinship (forreview,see

ref.29) suggests thatgeneticfactorsmaybeof importancein theetiology ofMS. In aCanadianpopulation-basedstudyof MS in twins(30),asignificantly higherconcordancerate was found in monozygotic twins compared to dizygotic twins, indicatingamajor genetic etiologiccomponent.

Theinheritance ofMSmight bepolygenic, with lociof the HLAclass IIregion contributingamajorpartofthegenetic predisposition. Linkage tothe HLAregion issupported by

many (e.g., refs. 29, 31), but not all (32, 33), studies of haplotype sharinginaffected sibling pairs. In formal

segre-gationandlinkage analysis(34), anexcellent fit wasobtained foramodelwithadominantMS determinanttightlylinkedto

the DR2 haplotype and a second unlinked determinant.

However, asonly limiteddata areavailableonthe concor-dancerateofMS inHLA-identicalsiblings,theinfluenceof non-HLA-linkedgenes isdifficulttoascertain.

The penetrance of MSin individuals carrying the associ-ated Taq I HLA-DRB and -DQB patterns islow. Thisdoesnot necessarily implythat thedisease ispolygenicorprecipitated by an environmental influence. The T-cell-receptor

a/,8

heterodimersareformedby random geneticrearrangements.

During maturationthe T cells are submittedtopositiveand

negative thymic selection controlled by theHLA antigens. Thus,the T-cellrepertoire ofanindividual isinfluencedby

stablegenetic markers,the HLA genes,andarandomgenetic

process. Thelikelihood for formation ofspecificT-cellclones

that might contribute to disease will, therefore, be

HLA-dependent. The penetrance ofdisease will depend on the

stochastic event offorming T-cellclonalrepertoires witha capacityto cause disease.

InthepresentstudyanEFof0.70is found for thepresence

ofthe RFLP-definedalleles T-DRBII(DRw15), T-DQB IV

-i MIN ',I; ."AL, AM 0 0

Table 3. Frequencies of Taq I HLA-DQB allelicrestrictionfragmentpatternscontributingto susceptibility and resistance toPCP MS compared toR/RMS

P

Frequency,t PCP MS PCP MS R/R MS

Pattern PCP MS R/RMS Controls* vs. R/RMS vs.controlst vs.controlst

T-DQB IV(DQw8,DQw9,DQw4)t 65 27 38 <0.001 <0.05 NS

T-DQB V(DQw7)§ 0 19 29 <0.01 <0.05 NS

T-DQB I1(DRw17,DQw2)1 4 32 15 <0.005 NS <0.005

NS, notsignificant. nvalues are asfollows: for PCP MS, 26; forR/RMS, 74; for controls, 100. *FrequenciesoftheMS-associatedDQB patternsin the healthy controls are given for

comparison.

tpcorrected by a factor of 58.

WPositivelyassociatedwith PCP MS.

§Negatively associated withPCP MS. Positivelyassociated withR/RMS.

(DQw8, DQw9, DQw4), or bothin PCP MS, and an EF of 0.61 is found in R/R MS for the occurrence of T-DRB HI (DRw15), T-DQB III (DRw17,DQw2), or both. Thesefigures

are in good agreement with the concept that most ofthe

genetic susceptibilitytoMS isconferred bygenesintheHLA class II region.

In population studies ofMS, associations were first

ob-served with the HLAclassIantigensA3 and B7. However,

subsequentlystrongerassociations werefoundwith the class II specificities Dw2 and DR2. In most population studies of Caucasian MSpatients consistently strongassociationswith the Dw2 and DR2specificitieshavebeenobtained (ref.35;for

review, see ref. 36). A Dw2 association has also been

observed in American black MS patients. A weak DR3 association has been observed in a few studies (37). In

Arabian (38) as well as in Sardinian (39) MS patients an

association with the DR4antigen has been found. Aweak

association with the DQwl antigen (DR2 associated) was

foundinScottish MS patients (11),but this finding has not been confirmed (12, 40). An association with the DPw4

specificity has been reported inSwedish MS patients (41). However, in aNorwegian study (42), where a DPw4 asso-ciation was also observed, this was interpreted to be

sec-ondaryto aprimaryDR2association. Itis worthnotingthat inpopulation studies ofDPantigen frequencies in

Norwe-gians (43), the Dutch (44), and Danes (45), weak linkage disequilibriumbetween the DR2 and DPw4specificitieshas beenobserved.

In our MS patients, the well-known DR2 association is

confirmed(for review, seeref. 36). Inaddition, the associ-ationisfound tobe withthe RFLP pattern (T-DRB

If)

that

correspondstothe serological DR2-split DRw15. The three

cellularly defined specificities associated with DR2 (Dw2,

Dw12, and Dw2l) can readily be distinguished by Taq I RFLPanalysis andDRB, DQA, and DQB linkage analysis (18, 46). AllT-DRB II(DRw15)-positive patientsin our study have the Taq I RFLP patternscorrespondingtoDw2,which is in accordancewithprevious findings(35).

An association with T-DQB II, corresponding to the DQwl-split DQw6, is also observed. Based on RFLP data

this association is statistically secondary to the T-DRB II (DRw15)association (P < 0.001). However, as theprotein

sequencesof thepolymorphic amino-terminaldomain of the DRBI (for review, see ref. 13) and the DQB1 (47) gene

productsmost likely are uniquein the DRwl5,DQw6,Dw2 haplotype, it is notpossible by Taq I RFLP analysis or by

serologicalDRandDQ typingtodetermine thebiologically primary association.

Asit has beensuggestedthatPCPMS and

R/R

MSmaybe

distinct diseases(6), it is ofgreatinterestto notethefinding ofpositive and negative associations with different

RFLP-definedDQBI alleles in thetwoclinical forms ofMS. InPCPMS anassociationwith T-DQBIVisfound. This RFLP pattern is seen in DR4,DQw8, DR7,DQw9, and

DRw8,DQw4 haplotypes. Theamino acid sequences ofthe

DQB chain probably differ between these three haplotypes (47). However, asignificantproportionof T-DQBIV-positive PCP MS patients exhibitTaq I DRB,DQA,andDQB RFLP patternscorrespondingtotheDR4,DQw8 haplotype (Pcorr38 <0.05). The five Dw types (Dw4, DwlO,Dw13, Dw14, and Dw15)associatedwith the serologicalspecificityDR4 cannot be separated by RFLP analysis. The distribution ofthese

cellularlydefinedspecificities is not known inPCPMS.If the

distributiondoes not markedlydiffer betweenDR4-positive

PCP MS patients and controls, it would be tempting to

assume that the observed DR4,DQw8 association might primarilybewiththe DQBI locus. However, thiswouldhave tobeconfirmed by specific sequencing.

A negative (i.e., protective) association is observed with T-DQBV(DQw7) in PCPMS.ThisTaqIDQBJallelicpattern is mainly seen in the DR4,DQw7, DRwll,DQw7, DRw12,DQw7, and DRw8,DQw7 haplotypes. The amino acid sequences ofthe first domain of the DQB chains are identical in all these haplotypes (47). However, the amino acidsequencesoftheDQA(47) and DRB(for review,seeref.

13) chains differ. Thus, one might speculate whether the protective effectmaybe conferred by theDQBJgene

prod-uct.

InR/R MSapositive associationisfoundwith T-DQB III

(DRw17,DQw2). Itisof interestto note thatnoassociation

is found with the Taq I DQB pattern seenintheDR7,DQw2 haplotype. Althoughthe RFLP patternsofthesetwo

haplo-types are different, the sequences of the membrane-distal domain oftheDQB chainsare identical(47). However,the

amino acidsequencesofthefirst domainoftheDQA chains differ (47) between the two DQw2 haplotypes, which may

explain why an association is only found with the DQB

pattern of the DRw17,DQw2 haplotype and not with the

DR7,DQw2 haplotype. No association isobservedwith the Taq I DQA RFLP pattern ofthe DRw17,DQw2 haplotype.

This DQA pattern is also found in DRwll,DQw7 and

DRw12,DQw7 haplotypes. The sequences ofthe polymor-phic domain oftheDQA chainsinthehaplotypesexhibiting

this Taq I DQA pattern areidentical(47). Furthermore, the

two RFLP patterns (T-DRB IV and T-DRB VIII; Table 2)

associated with DRw17 have identical sequences of the

secondexonof the DRBI locus (48).They differonlyin the DRB3 locus, which encodes the DRw52 specificities (49). Thus,RFLPdata obtained in the presentstudysupported by

available sequence information confirms that

R/R

MS is

associated with the specific DQA-DQB heterodimer ofthe

DRw17,DQw2 haplotype, whereneitherthesequenceofthe

DQAortheDQB chain isunique. However, it isnotpossible toexcludethat the

biologically

primary association

might

be

with the

DRBJ

locusofDRw17,DQw2 haplotype.

In the present study the well-established

DRwl5,Dw2

association in MS is confirmed. Furthermore, different

con-tributetosusceptibility and resistancetoPCP MSaswellas to susceptibility to R/R MS. Due to the strong linkage

disequilibrium between DR and DQ subregions, it is, as

previously mentioned, not possible by RFLP analysis to

exactly determine which DR loci, DQ loci,orboth thatconfer susceptibilitytothedifferentclinical forms of MS. This issue

hastobeaddressedbysequencing of the polymorphicsecond

exon of DRBI and DQBJ genes in haplotypes positively associated with MS. Theprotective association observed in PCPMS might primarily be with the DQBJ locus.

It is of great interest to note that the differences in frequencies of the DQB RFLPpatternspositively associated with PCP MS and R/R MS are more pronounced when

comparingthe twoforms ofMS with each other than when comparing PCPandR/RMSpatients with controls(Table 3). The observed immunogenetic heterogeneity gives strong support tothehypothesis that PCP MS and R/R MSare two distinctdisease entities. This maywellprove tobe ofgreat importancewhen studying the etiology andpathogenesis of

MS andalsowhenthe effectof differenttherapeutic regimes

areevaluated.

Thisstudywassupported by grants from the Karolinska Institute; the SwedishMedical Research Council (Project 00793); 'Forenade Liv' MutualGroupLife InsuranceCompany, Sweden;theMedical Research Council of the Swedish Life Insurance Companies; the Multiple Sclerosis Foundation, Sweden; the Swedish Society for Medical Research;theSwedish SocietyofMedicine, and the Lars HiertaMemorial Foundation.

1. Schlosstein, L., Terasaki, P.T., Bluestone, R. & Pearson, G. M.(1973) N. Engl. J. Med.288,704-706.

2. Brewerton, D. A.,Caffrey,M., Hart, F. D., James, D. C. O., Nicholls,A. &Sturrock,R. D. (1973) Lancet i, 904-907. 3. Tiwari, J. L. & Terasaki, P.I. (1985) in HLA and Disease

Associations(Springer, NewYork), pp. 28-48.

4. Confavreux, C., Aimard, G. & Devic, M. (1980) Brain 103, 281-300.

5. Leibowitz, U. & Alter, M. (1970) Acta Neurol. Scand. 46, 53-70.

6. Larsen,J.P.,Kvaale,G., Riise,T.,Nyland,H. &Aarli, J. A. (1985)ActaNeurol. Scand.72, 145-150.

7. Vegans, E., Theys, P., Delmotte, P. & Carton, H. (1983) J. Neurol. 229, 155-165.

8. Poser, S.,Poser, W., Schlaf, G., Firnhaber, W., Lauer, K., Wolter, M.& Evers,P. (1986)ActaNeurol. Scand. 74, 387-392.

9. Aimard, G., Confavreux,C. &Devic,M.(1980) inProgress in Multiple Sclerosis Research,ed. Bauer,H.(Springer, Berlin), pp. 376-380.

10. Madigand, M., Oger,J. J.-F.,Fauchet,R., Sabouraud,0. & Genetet,B.(1982)J. Neurol. Sci. 53, 519-529.

11. Francis, D.A., Batchelor, J. R., McDonald, W.I., Hing, S.N.,Dodi,I.A., Fielder,H.L., Herm,J. E.C. &Downie, A. W.(1987)Brain 110,181-196.

12. Lambalgen,R. V.,Sanders,E. A. C. M. &D'Amaro,J. (1986) J.Neurol. Sci. 76, 13-21.

13. Kappes, D.& Strominger,J. L.(1988) Annu. Rev. Biochem. 57,991-1028.

14. Wake,C. T., Long,E. 0.&Mach,B.(1982) Nature (London) 300,372-374.

15. Cohen, D., Le Gall, I., Marcadet,A., Font, M.-P., Lalouel, J.-M. & Dausset, J. (1984) Proc. NatI.Acad. Sci. USA 81, 7870-7874.

16. Carlsson, B., Wallin, J., Bohme,J.&Moller,E.(1987) Hum. Immunol.20,95-113.

17. Bidwell,J. L.,Bidwell,E. A.,Laundy,G. J.,Klouda,P. T.&

Bradley,B. A. (1987) Mol. Immunol. 24,513-522. 18. Bidwell,J. L. (1988)Immunol. Today9, 18-23.

19. Olerup, O., Fredrikson, S., Olsson, T. & Kam-Hansen, S. (1987) Lancetii,327.

20. Schumacher, G., Beebe, B., Kibler, R., Kurland, L.,Kurtzke, J. F., McDowell, F., Nagler, B., Sibley, W. A.,Tourtellotte, W. W. & Willmon, T. L. (1965) Ann. N.Y. Acad. Sci. 122, 552-568.

21. Link, H. & Tibbling, G. (1977) Scand. J. Clin. Lab. Invest. 37, 397-401.

22. Olsson, T., Kostulas, V. & Link, H. (1984) Clin. Chem. 30, 1246-1248.

23. Gustafsson, K., Wiman, K., Emmoth, E., Larhammar, D., Bohme, J., Hyldig-Nielsen, J. J., Ronne, H., Peterson, P. A.& Rask, L. (1984) EMBO J. 3, 1655-1661.

24. Schenning, L., Larhammar, D., Bill, P., Wiman, K., Jonsson, A.-K., Rask, L. & Peterson, P. A. (1984)EMBOJ. 3,447-452. 25. Larhammar, D., Schenning, L., Gustafsson, K., Wiman, K., Claesson, L., Rask, L. & Peterson, P. A. (1982) Proc. NatI. Acad. Sci. USA 79, 3687-3691.

26. WHO-HLA Nomenclature Committee (1988) Vox. Sang. 55, 119-126.

27. Woolf, B. (1955) Ann. Hum. Genet. 19, 251-253.

28. Bengtsson, B. 0. & Thomson, G. (1981) Tissue Antigens 17, 356-363.

29. Suarez, B., O'Rourke, D. & Van Eerdewegh, P. (1982) Am. J. Med. Genet. 12, 309-326.

30. Ebers, G. C., Bulman, D. E., Sadovnick, A. D., Paty, D. W., Warren, S., Hader, W., Murray, T. J., Seland, T. P., Duquette, P., Grey, T., Nelson, R., Nicolle, M. & Brunet, D. (1986) N. Engl. J. Med. 315, 1638-1642.

31. Stewart, G. J., McLeod, J. G., Basten, A. & Bashir, H. V. (1981) Hum. Immunol. 3, 13-29.

32. Ebers, G. C., Paty, D. W., Stiller, C. R., Nelson, R. F., Se-land, T. P. &Larsen, B. (1982) Lancet i, 88-90.

33. Govaerts, A., Gony, J., Martin-Mondiere, C., Poirier, J. C., Schmid, M., Schuller, E., Degos, J. D. & Dausset, J. (1985) Tissue Antigens 25, 187-190.

34. Ho, H.-Z.,Tiwari, J. L., Haile, R. W., Tersaki, P. I. & Mor-ton, N. E. (1982)Immunogenetics 15, 509-517.

35. Jersild, C., Fog, T., Hansen, G. S., Thomsen, M., Svejgaard, A. & Dupont, B. (1973)LancetH,1221-1225.

36. Tiwari, J. L. & Terasaki, P. I. (1985) in HLA and Disease Associations (Springer, New York), pp. 152-167.

37. de Moerloose, P., Jeannet,M., Martins-da-Silva, B., Werner-Favre, C., Rohr, J. & Gauthier, G. (1979) Tissue Antigens 13, 357-360.

38. Kurdi, A., Ayesh, I., Abdallat, A., Maayta, U., McDonald, W.I.,Compston, D. A. S. & Batchelor, J. R. (1977) Lancet i, 1123-1125.

39. Marrosu, M. G., Muntoni, F., Murru, M. R., Spinicci, G., Pischedda, M. P., Goddi, F., Cossu, P. & Pirastu, M. (1988) Neurology 38, 1749-1753.

40. Swingler, R. J., Kirk, P. F., Darke, C. & Compston, D. A. S. (1987) J. Neurol. Neurosurg. Psychiatry 50, 1153-1155. 41. 0dum, N.,Hyldig-Nielsen, J. J., Morling, N.,

Sandberg-Woll-heim, M., Platz, P. & Svejgaard, A. (1988) Tissue Antigens 31, 235-237.

42. Moen, T., Stien, R.,Bratlie,A. &Bondevik, E. (1984) Tissue Antigens 24, 126-127.

43. Moen, T., Bratlie,A., Kiss, E., Bruserud,0. & Thorsby, E. (1983) Tissue Antigens 22, 298-309.

44. Termijtelen,A., Naipal-vanden Berge,S.,Suwandi-Thung, L. & van Rood, J. J. (1985) Tissue Antigens 26, 234-242. 45. 0dum, N.,Hartzman,R.,Jakobsen, B.K.,Morling,N.,Platz,

P., Robbins,F.-M., Ryder, L. P. & Svejgaard, A. (1986) Tissue Antigens28, 105-118.

46. Lee, B.S.M., Bell, J. I., Rust, N. A. & McDevitt, H. O. (1987) Immunogenetics 26, 85-91.

47. Horn,G. T., Bugawan, T. L., Long, C. M. &Erlich, H. A. (1988) Proc.Natl. Acad. Sci. USA 85, 6012-6016.

48. Todd, J. A., Bell, J. I. & McDevitt, H.0. (1987) Nature (London) 329, 599-604.

49. Tiercy, J.-M., Gorski, J., Jeannet, M. & Mach, B. (1988) Proc. Natl.Acad. Sci. USA 85, 198-202.