Ankylosing spondylitis and HLA-B27: polymorphism sequencing › content › annrheumdis › 51 › 7...

Annals ofthe Rheumatic Diseases 1992; 51: 855-862

Ankylosing spondylitis and HLA-B27: restrictionfragment length polymorphism and sequencingof an HLA-B27 allele from a patient with ankylosingspondylitis

Colleen M Higgins, Torben Lund, Michael E Shipley, Alan Ebringer,Maria Sadowska-Wroblewska,l Roger K Craig

AbstractTwo groups of patients with ankylosingspondylitis (AS) from England and Polandwere examined for restriction fragment lengthpolymorphisms (RFLPs) associated with thedisease. No preferential association wasfound between the 9-2 kb PvuII fragment inHLA-B27 positive patients with AS comparedwith HLA-B27 healthy subjects as had beenpreviously reported. In the English group,however, a 14 kb Pvull fragment was morecommon in HLA-B27 positive subjects withAS than in normal controls. Also 4-6 and3-7 kb PvuII fragments were more prevalentin subjects without AS than in the group withAS, but these results were confined to theEnglish group. Furthermore, the sequence ofan HLA-B*2705 gene isolated from a patientwith AS was examined, and no significantdifferences were found compared with thesequence isolated from a healthy subject.There do not seem to be significant geneticdifferences in the coding or in the regulatoryregion in HLA-B27 alleles, in subjects with orwithout AS.

(Ann Rheum Dis 1992; 51: 855-862)

HLA-B27 molecule.' Alternatively, it may notbe HLA-B27 itself which plays a part in thedisease development but a gene in linkagedisequilibrium with HLA-B27. There is reportedto be no difference between the amino acidsequence of the extracellular domains of HLA-B27 antigen in patients with AS and in non-diseased subjects.9

Restriction fragment length polymorphism(RFLP) analysis has been used to examine thepossibility of linking a particular allelic form ofHLA-B27 with AS. Using an HLA-B7 cDNAprobe, McDaniel et al showed that a poly-morphic 9-2 kb PvuII DNA fragment waspresent in 72-9% of HLA-B27 positive patientswith AS but in only 22% of HLA-B27 positivenormal white subjects from Alabama, USA.'°Segregation studies in families showed that the9-2 kb PvuII polymorphic DNA fragment is nota polymorphism of the HLA-B27 allele. Anumber of more recent RFLP studies havefailed to confirm an association between the9-2 kb PvuII fragment and AS."-'3We report here our studies of two groups of

European patients with AS, usually geneticallymore homogeneous than American patients,who were investigated to determine a possibleassociation between RFLPs and AS.

Departments ofImmunology andRheumatology Researchand The MedicalMolecular Biology Unit,University Collegeand Middlesex (UCHSH)School of Medicine,LondonC M HigginsT LundM E ShipleyR CraigDivision ofBiomolecular Sciences,King's College LondonA EbringerInstitute ofRheumatology,Warsaw, PolandM Sadowska-WroblewskaCorrespondence to:Dr T Lund,Department of Immunology,University College andMiddlesex Schoolof Medicine,40-50 Tottenham Street,London WIP 9PG,United Kingdom.Accepted for publication18 October 1991"Deceased.

Ankylosing spondylitis (AS) is an inflammatorydisease affecting the sacroiliac joints, spine, andperipheral joints. Patients with reactive arthritisassociated with HLA-B27 often have a historyof antecedent infection with microorganismssuch as salmonella, shigella, and yersinia,'whereas in patients with AS klebsiella has beenimplicated as a possible cause.2 3 A strong as-sociation exists between AS and the majorhistocompatibility complex (MHC) class Iantigen HLA-B27,4 5 with over 90% of patientswith AS carrying this HLA antigen comparedwith a prevalence of HLA-B27 of less than 8%in the white population. The relative risk forsubjects carrying HLA-B27 has been shown tobe greater than 70.6 HLA-B27 is present in 50%of American black subjects with AS, whereasonly 2% of the American black population havethe HLA-B27 allele.7

Despite the close association between AS andHLA-B27 the mechanism underlying it remainsunclear. It has been suggested that HLA-B27 islinked directly to the pathology of AS throughthe mechanism of molecular mimicry by meansof immunological cross reactivity betweenantigens on the associated bacteria and the

Materials and methodsSUBJECTSThirty four unrelated English patients with AS(seven women, 27 men), all with classical ASaccording to the New York criteria,'4 werestudied. Thirty two of the 34 patients werewhite. Thirty two patients had peripheral jointdisease and 14 had had uveitis. Only one HLA-B27 negative patient was included, who hadperipheral joint disease but not uveitis. Thirtytwo healthy English subjects with no previoushistory of AS in the family, including seven whowere positive for HLA-B27, served as controls.Twenty Polish white patients with AS were

also studied, of whom 10 had peripheral jointdisease. The incidence of uveitis among thesepatients was not known. Thirteen healthy,white Polish subjects, of whom seven wereHLA-B27 positive without any history of arth-ritis or spondylitis, served as controls.

PROBESFull length HLA-B27 cDNA probe pB l'5 waskindly provided by Dr E Weiss. HLA-B27 5'

855

on June 14, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.51.7.855 on 1 July 1992. D

ownloaded from

on June 14, 2020 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.51.7.855 on 1 July 1992. Dow

nloaded from


http://ard.bmj.com

/A

nn Rheum


ownloaded from

http://ard.bmj.com/

http://ard.bmj.com/

http://ard.bmj.com/

Higgins, Lund, Shipley, Ebringer, Sadowska-Wroblewska, Craig

and 3' cDNA probes were derived from pB 1.The 5' cDNA probe was a 415 bp PstI-PvuIIfragment containing the sequences encoded inthe leader sequence, the presequence, the ctldomain, and 194 bp of the a2 domain. The 3'cDNA probe consisted of the 550 bp PvuII-PstIfragment containing the 3' untranslated regionof pBI cDNA. The HLA-B27 3' flanking probewas a 1650 bp PstI-BamHI fragment from anHLA-B27 gene isolated from a subject withoutAS'6 and was also kindly provided by DrE Weiss. The probe contains sequencesimmediately downstream from the poly-Aaddition signal.

SOUTHERN BLOTTINGThe peripheral lymphocytes from 10 ml ofwhole EDTA blood were lysed in 90 ml of lysisbuffer containing 0-32 M sucrose (BDH),10 mM TRIS-HCI pH 7 5 (Sigma), 5 mMMgCI2 (BDH), and 1% (v/v) Triton-X100(Sigma). The nuclei were pelleted by centri-fugation at 2000 g for 20 minutes and thenresuspended in 5 ml 75 mM NaCl (BDH),24 mM EDTA (BDH) pH 8-0. The nuclearmembranes were lysed by addition of sodiumdodecyl sulphate (Sigma) to a final concentrationof0 5% and 0-2 mg/ml proteinase K (BoehringerMannheim) and incubated at 37°C for two to 12hours. The DNA was then purified by extractionwith an equal volume of phenol (BDH) followedby two extractions with chloroform (BDH).The DNA was further purified by addition of11 ml ethanol (BDH) after addition of sodiumacetate pH 5 -2 (BDH) to a final concentration of0-3 mol/l. The precipitated DNA was dissolvedin 0 5 ml TE buffer.DNA (10 ,tg) was digested to completion with

25-50 units of the restriction enzymes EcoRI,HindIII, orPvuII (BRL/Gibco) as recommendedby the manufacturers. The DNA fragmentswere separated through a 0-8% agarose (Sigma)gel at 60 V for 17 to 20 hours in 40 mM TRIS(Sigma), 20 mM acetic acid (BDH), 10 mMEDTA (BDH). The DNA fragments weredenatured, neutralised, and transferred toHybond-N membranes (Amersham, UK)followed by cross linking of the DNA tothe nylon membrane by ultraviolet irradiation(305 nm) for three minutes.

Blots were prehybridised for a minimum offour hours at 65°C in 6xSSC (900 mM NaCl,90 mM sodium citrate pH 6-15), 0-01% BSA(ICN Immunobiochemicals), 0-1% (w/v) poly-vinylpyrrolidone (Sigma), 0-1% Ficoll 600(Pharmacia), 10% dextran sulphate (Phar-macia), 0 1% sodium dodecyl sulphate, and0-1 mg/ml denatured sonicated herring spermDNA (Sigma). Fresh hybridisation buffercontaining 0 1 M EDTA was used for overnighthybridisation at 65°C. Denatured probe labelledwith [32P]dCTP (NEN) to a specific activity ofat least 108 cpm/tg according to Feinberg andVogelstein'7 was also added to the hybridisationbuffer at a final concentration of 1-2 x 10'cpm/ml. After hybridisation the membraneswere washed twice in 2xSSC/0-5% sodiumdodecyl sulphate at 65°C for 30 minutesfollowed by two 30 minute washes in 0 I x SSC/

0-5% sodium dodecyl sulphate at 65°C. Auto-radiography was carried out using screenedx ray cassettes at -70°C with Kodak XAR-5film.

CLONING OF THE HLA-B27 GENE FROM A PATIENTWITH ASLymphocytic DNA isolated from a patient withAS heterozygous for the HLA-B27 gene wasdigested to completion with EcoRI and selectedaccording to size by agarose gel electrophoresis.DNA fragments of about 6-5 kb containing theHLA-B27 gene were electroeluted and clonedinto k gtlO phages (Amersham, UK). Recom-binant phages (7x 104) were screened with the3' flanking probe, and one strongly crosshybridising clone was purified. The 6 5 kbgenomic insert was subcloned into pBluescriptSK+ (Stratagene) for further characterisation.The DNA sequence was determined withSequenase after subcloning of suitable DNAfragments into M13 mpl8 or M13 mpl9 bac-teriophages.

ResultsRFLP ANALYSISGenomic DNA from English and Polishpatients with AS and healthy normal subjectswas analysed for RFLPs by hybridisation with

HLA-B27 - AS B..B -

.w.*W

40:

4_ 4 4.

A

Figure 1 Autoradiograph ofPvuII digested genomic DNAfrom four HLA-B27 positive Polish subjects with AShybridised with full length HLA-cDNA probe. Afierhybridisation the blot was washed with 2xSSC, 0 1%/,sodium dodecyl sulphate (SDS) at 65°C and exposed againstx ray film for three days (A) after which the stringency of thewash was increased to 0-2 xSSC, 0 1% SDS and theSouthern blot was re-exposed to x ray films for sevendays (B).

856

..::o;w

:ft::W:

0. .10


http://ard.bmj.com

/A

nn Rheum


ownloaded from

http://ard.bmj.com/

Ankylosing spondylitis and HLA-B27

a number of probes derived from differentregions of the HLA-B27 gene or the corres-ponding cDNA. Hybridisation to Southernblots of EcoRI and HindIII digested DNAshowed that the HLA-B27cDNA probe detectedseveral DNA fragments. Some of these werepolymorphic, but none showed any associationwith the disease (data not shown).

In the light of this a more detailed analysiswas performed on Southern blots of PvuIIdigested DNA. Hybridisation at low stringencyof a full length cDNA (fig 1) as well as a 5'cDNA probe (fig 2) showed about 20 DNAfragments. Most of these are derived from otherMHC class I genes which have DNA sequencehomology with the HLA-B27 gene. The twoprobes detect virtually the same fragments,except that the 5' probe detects no fragmentsbelow 2-6 kb. The table lists the frequencies ofall fragments detected with the 5' cDNA probe.A 14 kb PvuII fragment was found in 82%(28/34) of English HLA-B27 positive patientswith AS, whereas it appeared in only 290/o (2/7)of the English control subjects with HLA-B27.The 14 kb PvuII fragment, however, occurred

B27 - AS -

II

with the same frequency in the HLA-B27negative control group as seen among HLA-B27positive healthy subjects. The 14 kb PvuIIfragment was not an AS susceptibility poly-morphism among HLA-B27 positive Polishpatients.The prevalence of the 9-2 kb PvuII frag-

ment, described by McDaniel and coworkers,'0was low among HLA-B27 positive English andPolish patients with AS (28% and 15% respec-tively). In contrast, this 9-2 kb fragment wassignificantly more prevalent in both Englishand Polish HLA-B27 positive controls (86% and50% respectively). This suggests a negativeassociation between the disease and the 9-2 kbfragment among HLA-B27 positive subjects.The 4-6 and 3-7 kb fragments were more

prevalent in HLA-B27 negative healthy subjects(70% and 67% respectively) than in HLA-B27 positive English subjects (35% and 23%respectively). The 3-7 kb fragment was alsoseen more commonly in healthy subjects (64%)than in patients with AS (16%), irrespective ofthe presence of the HLA-B27 allele. Among theEnglish patients it was found that the 3-7 kb

B27 + AS - B27 - B27 + AS +AS +

II 11Size(kbp)14

9.2*

.....

0w : * : ; : ; ; w _

_m 4_ _P _I

_N ~~66

629W 60

5-65 34.6

UP 4_ 4.0_W _ _- 3.73 6

3.4VW _ 32

* 0 4_ _ _ -W 2 6

Figure 2 Autoradiograph ofPvuII digested genomic DNA from eight HLA-B27 negative healthy subjects,five HLA-B27positive healthy subjects, one HLA-B27 negative English patient with AS, andfour HLA-B27 positive English patientswith AS hybridised with the pB27-5' cDNA probe. The 9-2 kb PvuII fragment is indicated with an asterisk.

Prevalence of PvuII fragments hybridising to the pB27-5' cDNA probe in each subset of the English and Polish groups.The 9-2 kb PvuII fragment is indicated with an asterisk. The percentage of patients with each fragment is shown

Size(kbp)

149-2776

ow 6-6

6 0->_ ~~~~~5-6

5-_0 534-6

3.73-63.43-2

.~~~~~~~~~~ 6

Engli'sh Polish

B27+AS+ B27+AS- B27-AS- B27+AS+ B27+AS- B27-AS-(n=34) (n=7) (n=25) (n=20) (n=4) (n=9)

822810010012367630389416949

91100

2986100100432971014

10057861486100

7235100100

8687639707667761481100

7015

10010010'58530359555750

7550

100100

02510025507550100

095 100

10( 100

_ _w 0:

7844100100

0221004444783367089100

I r---l r- I

857

I

I.

I


http://ard.bmj.com

/A

nn Rheum


ownloaded from

http://ard.bmj.com/


1 1 11 4'II1

I 01111 u

413,111 1-

011111 LUI4

I ~~~~~~~~~~~LI)

ou 0

LI

I 011111 8UI 0

I . L

I 11 1I 011111~~~9

I 01 I~ g

01 0

IU011 11 I

u0111 0

I0 0

31110u

00j,N44-

tn.of.

:1M

-C

00

119

101101

III

111N

141

III

III

31110u

4N

0111

4'''

LI''U

LIII

0111L11

It,

4LII

LI'''

LI I'0' ILI I'

lu0'''

LIII'0''''

LI I'

0L11LI''

LI'U

LIII'

X gm

.; . 44 mC. -0 C. ar. f. f. r-ft eqM eq

I :1 10, IIC oc oc0 M 1=1

01 III.I 1111

1111

01111

11111

8 1111111111

41III

LII II

01 111141 111101

. 11111

01

41 1111

01

01 1111

1111141

I'll

111111141011111

01

01 II

.

01

01 II41III 1111

U 1111

01 III411111011111

01 11114101 II

4501 1111

01 II41I-I II II

01 II01 111101

0101 111101 1111

011111 1111LII 1111

011111b-I 1111011111

01 III0101

01 II01 111101

0141 III01 1111

0111101 1111LII 1111

0101 1111011111

011111b-I 111101 1111

LII 111141011111

LII 1111LIIl's

LII011111LII 1111

b-I 1111411111b-I 1111

011111

411111

('sI41 1111411111

011111

011.1 1111011111

01111141

01 III41 III01 III

I-sI01 IIIIII 511

01 111101

300

0111I

II 0'

01111 IILI'

00

U0I111 U-

U111 I

LI'

1-I

01I11 4

01111 ID0*4111 1

011

4'0111 U'U11 '

L,I0111 LI

0111U LU011 UIl

UI1

1-IU1144''UU'U11 1411U -01U1ILI

LI

Ii aIl'a.1

0,U 1 1

NI N N%020 20 00

UO

UH :::

LI'

-K8: :6: c

'''

IU 'U"H,

4u

o LI'

6::'IC'l 4

UII0UI' 4

ELIII LII

11 0'

uu4 U1

U:UILII0

UI1-l48:: U

U:

::8:II'I 1-

U.140'I LI

LI'' 0'UI~ ' 0'

LI' U-

4' LIt:: U:LII 41LI'U I

H 1: t:0:~E-8

0''ICiEK-

LII UI8::o::

01 IIICI bIC I LI

IILII cIILII I

'I O-l I'I U-II bI I

UI1IIII LII I'I 4' I

0''C .

II111 -

3c touIn..1 ". I0 040 0 0r- r- r- r- r-

:1 :1 :1 N N

mooml mTI-C AC oc -Ca w w w m

39

4I.

31 CO U.11 .,.11

0 0 40 10 C)r- r- r- r- r-

:1 CV cli :07

0 I 'm 0 0,c

.4 :; ; 'M =

858

0'4'0'4'I-'LII

0'b-I0'0'0'

0'

8:0'

4'0'0'LI'0'4'LI'

8:b-I

4'LI'LI'8:0'8:LII4'

0'LI'

4'

6:LII4'8:

LI'

0'0'8'

8

8

LI'LI'0'

8:0'4'8:0'01

4'LI'

LI'

LII8:4'0'6:LIILII

LI'0'0'4'LIILI'0'0'LII4'LI'

88

8

10

a I

0 1

8'0iOg-

01

LI

4'

01OIOI

OIu141

CbP-I

LI'4'

01 4

b- 0

u: 0I- 4'

0' b-IC u:LI 0b- 0

LI LI' 0'

LII LI6:

LICI

I' 0'

6) 6

4 LI;LI 4u

4 0'

4 6:

0 LI

4 -I

L 0'C4 0'

1- LI

I- LILI LI

4 I'

1' 0

3c 3c

0 0

4 4C

01111101 1

II 1 1

LIlU11

II 0 11

II411

a$ U1

II fLI41 1

41

I

I C0

II LII

El-0 11

III LIIIl U~

3c U) y?V) N M to Ln .) C,4C3, 4* -0 -0 42 C2, C.f. r- r- r- r- r- r-44 C4 C4 C14 C4 04 :6

;% ; i I'm 4110 10 0SC oc oc -C IC oc ACid 0 m m m . m

3

4413C44't

It21

I1cIi

In in In tv fnC. 0 0S,0-, r- r. r,IN 14 14.7 .14I0 a so 1

dC .4 OC.ww m


http://ard.bmj.com

/A

nn Rheum


ownloaded from

http://ard.bmj.com/

Ankylosing spondylitis and HLA-B27

N -Id cm

Lo

LI~~~~~~~~11

0~~~~~~

11 lo0. 111 011 61~~~~~~~~~~I 1

0

0

0~~~~i

~ ~ ~

1~~~~~~o

LI,

LI Itu ~~~~~~~~~~~~IC

0~~~~L

LIC

1~~~~~~Is 11

lI I I

311to0

4ama.

~mma

IC

:1:: :~~~~~dO,::

U11

IC~ ~ ~ ~ ~ ~ S

U1 cc-U

ILI

LI1 U

I-'48::,

::~~~IO

0'''II)

i, ~~LI

UIl

0I

I I4IIIIII1~~~~~~~I

U:

EL

'~~~~I

fellIll 41ol

,

fIC1U.1U

I~~~~~~~~~~E-U U

'~~~~~U

OIl'

X 40 uI., ., ". m'n

Ca C3, 0 CDf. e- r- r- r-im :1 C-4 :0 Cld

c 0 0 a 1

.4I ; . ; l=C

x to u1; .11 C-4 tn0 c, c, -0 .!r- r- r- r-

:1 :1 :1 114 :1go to to I M

.3 .3I.c i =C =C

x w

6. an C4 enC. C3, C. 40r- r, r, r,

7:17:1go to 04 44

oc AC oc9 1=1

I

I

CI ',

0'

'C0

IC

x 0 U.1 .1 In. I0 40 0 0 =r- f- f- r- r,7 :1 t4 CM Cld

m m I 'm Iic NC ACr"'. :. m .i .

31110u

". .1

859

8-:8 :

ACI

6-:

LII

:

LII

ICI

01

4, .1 4

Ies

13,:

-0 4

.8qa-

t3TQ."

ce

LIII

LII

LI'

4

0

Llu-ILIII'"LII'

'IU

LIIIC

LIIII

Im

4'L'C0'.

:

fr-I

6e.

4-

4'LI

4'

LI4-

LI'

0'4ICA-

Eel

0'

it 0

tn . 914 f-i0 40 C. 0r- P. IZ r-14 " 44

. . :1;b a a a

I 1.; ; 5 L.

a

1 LC-0 -

.4

-1

I1

I


http://ard.bmj.com

/A

nn Rheum


ownloaded from

http://ard.bmj.com/


al Domain20 40 60 80

GSHSMRYFHTSVSRPGRGEPRFITVGYVDDTLFVRFDSDAASPREEPRAPWIEQEGPEYWDRETQICKALAQTDREDLRTLLRTTNQSEAB*2701 --------------------------------------------------------------------------Y--N---A---------B*2702 -----------------------------------------------------------------------------N--IA---------B*2703 ----------------------------------------------------------H-------------------------------B*2704 -----------------------------------------------------------------------------S-------------B*2705 ------------------------------------------------------------------------------------------

B*2706 ----------------------------------------------------------------------------S-------------

20 40 60 80GSHTLQNMYGCDVGPDGRLLRGYHQDAYDGKDYITLNEDLSSWTAADTAAQITQRKWEAARVAEQLRAYLEGECVEWLRRYLENGKETLQRA-----------------------------------------------------------------------------___------------

--------------------------------------------------------------__-------------__-------------

--------------------------------------------------------------__-------------__-------------

----------------------------------------____ ---- -----E------- --------E--------------~~~~~ ~~

-- -- - --- --- -- - ------ --D-Y---- -- --- --- --- -- ---- --- --- --- --- ---E--- --- --- --- ------ --- --- -- --- -

Figure 4 Comparison of the deduced amino acid sequence of the (I and (52 domains of the HLA-B27 gene presented herewith those ofother HLA-B27 subtypes. The amino acid sequences were compiledfrom Rojo et al '' (B*2701), Seeman et al "'(B*2702), Choo et al,25 (B*2703), Rojo et al 27 (B*2704), Weiss et al "' (B*2705), and Vega et al 28(B*2706).

PvuII fragment was seen more commonlyamong patients with uveitis (36%) than in thosewithout (5%).

Hybridisation of the Southern blots of PvuIIdigested DNA with the 3' cDNA and the3' flanking probe showed a more simplehybridisation fragment pattern detecting fewpolymorphic fragments. None ofthese, however,showed any particular association with thedisease in either of the two groups studied.

DNA SEQUENCE OF AN HLA-B27 ALLELEAs the RFLP analysis did not show any strikingdisease association the HLA-B27 allele from a

patient with AS was cloned and sequenced toidentify possible differences between it and thatof a healthy subject. To ensure that the HLA-B27 allele isolated indeed would be associatedwith the disease, a partial genomic library ofDNA isolated from a patient with AS carryingonly one copy of the HLA-B27 allele was

prepared. As characterisation of cloned HLA-Bgenes had previously shown that the HLA-Bgene was contained on a 6 5 kb EcoRI frag-ment,'6 DNA fragments of about 6-5 kb were

cloned into k gtlO phages. Screening with 3'flanking probe identified one HLA-B27 clone.Figure 3 shows the DNA sequence of 4259 bpof this clone. DNA sequence comparisonbetween the HLA-B27 gene from normalsubjects and the one isolated here from a patientwith AS showed no striking differences either inthe promoter and regulatory region or in theregion containing the structural gene. In parti-cular exons V, VI, and VII were identical.Indeed, few nucleotide differences could bedetected in the sequence of the gene isolated byus and that isolated by Weiss et al 16 from a non-diseased subject. None of the nucleotide differ-ences changed the amino acid sequence. Thuswhen translated into an amino acid sequence theHLA-B27 gene we cloned encoded a proteinidentical to that encoded by the HLA-B*2705allele (fig 4).

DiscussionMore than 95% of white patients with AS are

positive for the MHC class I antigen HLA-B27,but only about 20% of all HLA-B27 positive

subjects are likely to develop some stigmata ofthe disease, though the incidence seems to behigher in families with a history of AS.'8 Thismight be due to the influence of other genetic as

well as environmental factors on the pathologyof the disease. As RFLP analysis has proveduseful in associating particular DNA poly-morphisms with diseases'9 an extensive searchfor RFLPs preferentially associated with ASalone or together with HLA-B27 was carriedout. McDaniel and coworkers described a 9-2kb PvuII DNA fragment which was detectedwith an HLA-B27 cDNA probe,'0 and in a

study of patients with AS from Alabama (USA)it was found to be present in 73% of HLA-B27positive patients with spondylitis comparedwith only 27% of the healthy subjects withHLA-B27. The 9-2 kb PvuII gene seems toincrease the relative risk from 90 (for HLA-B27alone) to a relative risk of 297 when bothmarkers are present.'0 The 9-2 kb PvuIIfragment was later identified as a polymorphismof the HLA-A3 or the HLA-A9 gene,'2 suggest-ing the existence of either a second MHC linkedAS susceptibility gene, a disease gene differentfrom HLA-B27 but in linkage disequilibriumwith the HLA-B27 gene, or a particular allelicform of the HLA-B27 gene associated with AS.RFLP studies of patients with AS in Balti-

more,'2 Minnesota and Texas," and in Ger-many,l3 however, have all failed to confirmthe preferential association of the 9-2 kb PvuIIfragment and AS in HLA-B27 positive subjectsas reported by McDaniel et al.'o Indeed none ofthese RFLPs identified by MHC class I poly-morphisms preferentially associated with ASin HLA-B27 positive subjects. We analysedpatients with ankylosing spondylitis from twogeographically isolated, but in themselveshomogeneous, groups from England andPoland. We found no increase in the prevalenceof the haplotype HLA-A3/A9, HLA-B27among patients with spondylitis compared withnormal subjects. We did find, however, a

preferential association between AS and a 14 kbPvuII fragment among HLA-B27 positivepatients compared with HLA-B27 healthysubjects in the English group (but not in thePolish group), though this failed to reachstatistical significance when corrected for thenumber of fragments examined. None of the

a2 Domali

B*2701B*2702B*2703B*2704B*2705B*2706

860


http://ard.bmj.com

/A

nn Rheum


ownloaded from

http://ard.bmj.com/

Ankylosing spondylitis andHLA-B27

other studies'0-'3 reported a similar association.Only seven healthy subjects in our Englishcontrol group had the HLA-B27 class I alleleso this preferential association might not persistif a larger control group were to be studied.There is insufficient tissue typing data of ourpatient group to identify the MHC class I genecontaining the 14 kb polymorphism. In addition,4-6 kb and 3-7 kb polymorphic fragments werefound which, in the English control group, weremore prevalent in HLA-B27 negative subjectsthan in those who were HLA-B27 positive, andthe 3-7 kb fragment was more prevalent inhealthy subjects than in patients with ASirrespective of whether the subjects were posi-tive for HLA-B27 or not. This seems to bea negative correlation among English patientswith AS, but this association was not seen in thePolish group.The conflicting results might be due to

genetic differences between the groups studied.The HLA-A3/A9 allele might be linked with thedisease allele in the Alabama group studied byMcDaniel et al, 10 whereas such a linkage mightnot exist in the other groups with AS. Likewise,the 14 kb PvuII fragment, which appears to bepreferentially linked together with HLA-B27 toAS in the English group as reported here, butabsent in the other group studied, might reflectthe genetic diversity between different popu-lations. This implies that no additional MHCclass I gene is involved in the pathology of AS.If the HLA-B27 gene is not the disease geneitselfor ifadditionalMHC encoded susceptibilitygenes exist then they must be very close to theHLA-B27 allele.The discrepancy between the number of

people carrying HLA-B27 and those developingAS could be explained if there existed particularsusceptibility alleles of HLA-B27 which cannotbe identified by restriction fragment analysis.We sequenced the entire structural HLA-B*2705gene from a patient with AS. Our results showthat the translated amino acid sequence isidentical with the sequence of an HLA-B*2705antigen isolated from a healthy person. 16 20 Thisis in complete agreement with Coppin andMcDevitt,9 who found no differences betweenthe translated sequence of exon II, III, and IVencoding the extracellular portion of the MHCclass I antigen in HLA-B27 proteins frompatients with AS and healthy subjects. Further-more, our DNA sequence shows that thepromoter and regulatory region of the HLA-B27 gene isolated by us is virtually identical tothe non-diseased gene characterised by Weiss etal.'6 This eliminates the possibility that thepathology of AS might be due to alteredexpression of the HLA-B27 antigen. We foundvery few nucleotide differences between theHLA-B*2705 gene isolated by us and thenucleotide sequence of a similar allele from anormal subject cloned and sequenced by Weisset al 16 and Seeman et al.20 We noted somenucleotide differences in exon II and IV butnone which would result in amino acid differ-ences. We also found sequence differenceswithin intron I and II when compared withpreviously published sequences for non-diseasedHLA-B*2705 alleles. 16 20 Interestingly, some of

these differences create or destroy potentialbinding sequences for the transcription factorsSPI or AP2. It has not been shown that suchbinding sites are used in regulation of thetranscription, however.The absence of any differences between the

coding sequence of the HLA-B27 genes frompatients and healthy subjects does not eliminatea direct role for HLA-B27 in the pathogenesis ofthe disease through its function as restrictionelement for antigen presentation to cytotoxic Tcells.21 Many organ specific autoimmune di-seases, like insulin dependent diabetes mellitusand Hashimoto's disease, are characterised by acellular as well as a humoral immune response.Ankylosing spondylitis might well be initiatedby an infection of klebsiella or related bac-

22teria. The fact that the concordance ratefor AS in identical twins is well below 50%clearly suggests the involvement of an environ-mental factor in a genetically susceptible host23who may have more than one susceptibilitygene. Family studies suggest a possible role fora second susceptibility gene, not necessarilylocated in the MHC region. Such a secondsusceptibility gene could code for a self proteinsharing a T cell epitope with a bacterial protein,which might have a restricted tissue distribution,and this might account for the tissue specificityof the disease. A shared T cell epitope wouldthen only be presented by HLA-B27 or a relatedMHC class I antigen with a nearly identicalantigen binding groove.24

Clearly, further studies are required toresolve the problem of the association of HLA-B27 with AS.

We thank Drs G. Melmer and P Delves for critical comments inpreparation of this manuscript. Supported by grants from thetrustees of the Middlesex Hospital, London.

1 Toivanen A, Granfors K, Lahesmaa-Rantala R, Leino R,Stahlberg T, Vuento R. Pathogenesis of yersinia-triggeredreactive arthritis: immunological, microbiological andclinical aspects. Immunol Rev 1985; 86: 47-70.

2 Ebringer R, Cooke D, Cawdeli 0 R, Cowling P, Ebringer A.Ankylosingspondylitis: klebsiellaand HLA-B27. RheumatolRehabil 1977; 16: 190-6.

3 Ebringer A. The relationship between klebsielia infection andankylosing spondylitis. In: Rooney P J, ed. The gut andrheumatic disease. Balliere's Clin Rheumatol 1989; 3: 321-38.

4 Brewerton D A, Caffrey M, Hart F D, James D C 0,Nicholls A, Sturrock R D. Ankylosing spondylitis andHLA-B27. Lancet 1973; i: 904-7.

5 Schlosstein L, Terasaki P I, Bluestone R, Pearson C M. Highassociation of an HL-A antigen, W27, with ankylosingspondylitis. N Engl J Med 1973; 288: 704-6.

6 Tiwari J L, Terasaki P I. Ankylosing spondylitis. HLA anddisease association. Berlin: Springer, 1985: 85-100.

7 Keat A. Is spondylitis caused by klebsiella? Immunol Today1986; 7: 144-9.

8 Ebringer A. The cross-tolerance hypothesis, HLA B27 andankylosing spondylitis. BrJ Rheumatol 1983; 22 (suppl 2):53-66.

9 Coppin H L, McDevitt H 0. Absence of polymorphismbetween HLA-B27 genomic exon sequences isolated fromnormal donors and ankylosing spondylitis patients.J Immunol 1986; 137: 2168-72.

10 McDaniel D 0, Acton R T, Barger B 0, Koopman W J,Reveille J D. Association of a 9-2 kilobase Pvu II class Imajor histocompatibility complex restriction fragmentlength polymorphism with ankylosing spondylitis. ArthritisRheum 1987; 30: 894-900.

11 Durant J P, Taurog J D. Association between ankylosingspondylitis and a 9-2 kb Pvu II class I HLA DNArestriction fragment: a reassessment. J Rheumatol 1988; 15:1119-22.

12 Ahearn J M, Calomiris J J, Wigley F M, Jabs D A, BiasW B,Hochberg M C. Characterization of the class I HLA 9-2 kbPvu II restriction fragment length polymorphism. Linkageto HLA-A and lack of disease association. Arthritis Rheum1989; 32: 870-6.

13 Weiss E H, Bloemer K, Doerner C, et al. Molecular biology

861


http://ard.bmj.com

/A

nn Rheum


ownloaded from

http://ard.bmj.com/


of the HLA-B27 locus. BrJ Rheumatol 1988; 27 (suppl 2):12-18.

14 Bennett P H, Burch T A. New York symposium onpopulation studies in the rheumatic diseases: new diagnosticcriteria. BullRheum Dis 1967; 32: 354-63.

15 Szoets H, Riethmuller G, Weiss E H, Meo T. Completesequence of HLA-B27 cDNA identified through thecharacterization of structural markers unique to the HLA-A, -B, and -C allelic series. Proc Nail Acad Sci USA 1986;83: 1428-32.

16 Weiss E H, Kuon W, Doerner C, Lang M, Riethmulier G.Organization, sequence of expression of the HLA-B27gene: a molecular approach to analyze HLA and diseaseassociation. Immunobiology 1985; 170: 367-80.

17 Feinberg A P, Vogelstein B. A technique for radiolabellingDNA restriction endonuclease fragments to high specificactivity. Anal Biochem 1983; 132: 6-13.

18 Woodrow J C, Nichol F E, Whitehouse G H. Genetic studiesin ankylosing spondylitis. Br Rheumatol 1983; 22 (suppl2): 12-17.

19 Tsui L C, Bughwald M, Barker D, et al. Cystic fibrosis locusdefined by genetically linked polymorphic DNA markers.Science 1985; 230: 1054-7.

20 Seeman G H A, Rein R S, Brown C S, Ploegh H L. Geneconversion-like mechanisms may generate polymorphismin human class I genes. EMBOJ7 1986; 5: 547--52.

21 Benjamin R, Parham P. Guilt by association: HLA-B27 andankylosing spondylitis. Immunol Today 1990; 11: 137-42.

22 Trull A, Ebringer A, Panayi G, Ebringer R, James D C 0.

HLA-B27 and the immune response to enterobacterialantigens in ankylosing spondylitis. Clin Exp Immunol 1984;55: 74-80.

23 Eastmond C J, Woodrow J C. Discordance for ankylosingspondylitis in monozygotic twins. Ann Rheum Dis 1977; 36:360-4.

24 Ewing C, Ebringer R, Tribbick G, Geysen H M. Antibodyactivity in ankylosing spondylitis sera to sites on HLA-B27.1 at the MHC groove region (within sequence 65-85),and to a Klebsiella pneumoniae nitrogenase reductasepeptide (within sequence 181-199). Exp Med 1990; 171:1635-47.

25 Choo S Y, St John T, Orr H T, Hansen J A. Molecular analy-sis of the variant alloantigen HLA-B27d (HLA-B*2703)identifies a unique single amino acid substitution. HumImmunol 1988; 21: 209-19.

26 Rojo S, Aparicio P, Choo S Y, Hansen J, Lopez de CastroJ A. Structural analysis of an HLA-B27 population variant,B27f. J Immunol 1987; 139: 831-6.

27 Rojo S, Aparicio P, Hansen J A, Choo S Y, Lopez de CastroJ A. Structural analysis of an HLA-B27 functional variant,B27d, detected in American blacks. Immunol 1987; 139:3396-401.

28 Vega M A, Bragado R, Ivanyi P, Palaez J L, Lopez de CastroJ A. Molecular analysis of a functional subtype ofHLA-B7.A possible evolutionary pathway for HLA-B27 poly-morphism. J Immunol 1986; 137: 3557-65.

862


http://ard.bmj.com

/A

nn Rheum


ownloaded from

http://ard.bmj.com/

Letters to the editor

Food intolerance in

rheumatoid arthritis

Sir: We read with interest the two papers byDrs van de Laar and van der Korst on foodintolerance in rheumatoid arthritis (RA).1 2 Itappears from this extensive but short study ofpatients with seropositive RA that rheumatoiddisease in a small subgroup is sensitive todietary manipulation. Identification of thesepatients is, however, a problem.We have previously shown that significant

gut abnormalities are associated with raisedlevels of IgA rheumatoid factor (RF) anddietary protein specific IgG in patients withRA and suggested that these changes mightidentify patients who would respond to dietarymanipulation.3Gendre and colleagues have described villous

atrophy and other histological abnormalities ofthe small intestine in patients with RA.4 Thesechanges were more significant among IgM RFnegative patients (who are often IgA RFpositive).

In the recent study the number of patientsexamined histologically was small, therefore a

firm conclusion cannot be reached. Moreover,it is likely that by excluding IgM RF negativepatients with RA some food intolerant patientsmight have been missed. Furthermore,investigation of intestinal structure andfunction in RA is severely hampered by thedamaging effects of non-steroidal anti-inflam-matory drugs, which are so often used bypatients with RA. Thus a fundamental questionremains extremely difficult to answer-that is,whether gut abnormalities and any of themanifestations associated with these abnor-malities in patients with RA are secondary tothe disease process and its treatment or are ofprimary pathogenic significance.

MOHAMED ABUZAKOUKCLIONA O'FARRELLYDepartment of Immunology

St _'ames's HospitalDublin 8Ireland

Correspondence to: Dr O'Farrelly.

1 van de Laar M A F J, van der Korst J K. Foodintolerance in rheumatoid arthritis. I. A doubleblind, controlled trial of the clinical effects ofelimination of milk allergens and azo dyes. AnnRheum Dis 1992; 51: 298-302.

2 van de Laar M A F J, Aalbers M, Bruins F G,van Dinther-Janssen A C H M, van der KorstJ K, Meijer C J L M. Food intolerance inrheumatoid arthritis. II. Clinical and histo-logical aspects. Ann Rheum Dis 1992; 51:303-6.

3 O'Farrelly C, Marten D, Melcher D, et al.Association between villous atrophy in rheu-

matoid arthritis, rheumatoid factor andgliadin-specific IgG. Lancet 1988; ii: 819-23.

4 Gendre J P, Luboinski J, Prier A, Camus J P, leQuintrec Y. Anomalies de la muqueusejejunale et polyarthrite rhumatoide: 30 cas.Gastroenterol Clin Biol 1982; 6: 772-5.

AUTHORS' REPLY We appreciate the interestof Drs Abuzakouk and O'Farrelly in ourpapers on food intolerance in rheumatoidarthritis (RA).' 2 In the first paper theexistence of intolerance for food in a minorityof patients was suggested and in the secondpaper the patients with food intolerance weredescribed in more detail. We hoped thatthe clinical, immunological, and histologicalinvestigations in these patients might providesome clues to the problem of how to identifyfood intolerant patients. However, possiblybecause of the limited number of patients,such predicting parameters were not found.As suggested byDrsAbuzakoukand O'FarrellyIgA rheumatoid factor (RF) might by such aparameter-this possibility was not consideredin our paper. We studied several dietaryprotein specific IgGs, however, but a dis-criminatory high level of these antibodies wasnot found.3

Histological abnormalities of the smallintestine are of particular interest in con-nection with food intolerance in RA.4However, as this needs an invasive tech-nique the availability of material for studyis limited. Moreover, several other factors,such as the effect of non-steroidal anti-inflam-matory drugs, do influence intestinal histology.The existence of food intolerance in RF

negative RA might indeed exceed that inpositive disease. In these studies, how-ever, we preferred to study the more homo-geneous type of RA-namely, RF positivecases.We should mention that we are at present

studying the identification of food intolerantpatients with RA. In this study many of thesuggestions made by Drs Abuzakouk andO'Farrelly are included.

MARTIN A F J VAN DE LAARJAN K VAN DER KORST

Vakgroep Reumatologie TwenteZiekenhuizen: Twenteborg,

Almelo en Medtsch Spectrum Twente,Enschede

PO Box 76007600 SZ AlmeloThe Netherlands

I van de Laar M A F J, van der Korst J K. Foodintolerance in rheumatoid arthritis. I. A doubleblind, controlled trial of the clinical effects ofelimination of milk allergens and azo dyes. AnnRheum Dis 1992; 51: 298-302.

2 van de Laar M A F J, Aalbers M, Bruins F G,

van Dinther-Janssen A C H M, van der KorstJ K, Meijer C J L M. Food intolerance inrheumatoid arthritis. II. Clinical and histo-logical aspects. Ann Rheum Dis 1992; 51:303-6.

3 van de Laar M A F J, Aalbers M, van der KorstJ K. Radioallergosorbent tests to foods inrheumatoid arthritis. In: van de Laar M A F J,ed. Rheumatoid arthritis and food allergy?Amsterdam: Thesis Publishers, 1991: 81-91.

4 Marcolongo R, Bageli P F, Montagnami M.Gastrointestinal involvement in rheumatoidarthritis: a biopsy study. Jf Rheumatol 1979;6: 426-40.

5 Doube A, Collins A J. Is the gut intrinsicallyabnormal in rheumatoid arthritis. Ann RheumDis 1988; 47: 617-9.

6 van de Laar M A F J, van der Korst J K.Rheumatoid arthritis, food and allergy. SeminArthritis Rheum 1991; 21: 12-23.

CorrectionAnkylosing spondylitis and HLA-B27:restriction fragment length polymorphismand sequencing of an HLA-B27 allele from apatient with ankylosing spondylitisIn the paper by C M Higgins, T Lund, M EShipley, et al (Ann Rheum Dis 1992; 51:855-62) we regret that fig 1 was incorrectlyreproduced. The correct version of this figureis shown below.

I, 4W :51

.4'* A* 0*

88

Ankylosing spondylitis and HLA-B27: polymorphism sequencing › content › annrheumdis › 51 › 7...

Documents

Transcript of Ankylosing spondylitis and HLA-B27: polymorphism sequencing › content › annrheumdis › 51 › 7...