Rev. Bras. Reumatol. vol.56 número5

(1)

ww w . r e u m a t o l o g i a . c o m . b r

REVISTA

BRASILEIRA

DE

REUMATOLOGIA

Original

article

Clinical

characteristics

and

frequency

of

TLR4

polymorphisms

in

Brazilian

patients

with

ankylosing

spondylitis

Natalia

Pereira

Machado

a

_,

_Eliana

_Nogueira

b

_,

_Karen

_Oseki

a

_,

Pâmela

Carolina

Cruz

Ebbing

a

_,

_Clarice

_Silvia

_Taemi

_Origassa

b

_,

_Tatiane

_Mohovic

b

_,

Niels

Olsen

Saraiva

Câmara

b

_,

_Marcelo

_de

_Medeiros

_Pinheiro

a,∗

a_Divisão_de_{Reumatologia,}_Departamento_de_Medicina,_Universidade_Federal_de_São_Paulo_(UNIFESP),_São_Paulo,_SP,_Brazil b_Divisão_de_Nefrologia,_Departamento_de_Medicina,_Universidade_Federal_de_São_Paulo_(UNIFESP),_São_Paulo,_SP,_Brazil

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received10February2016

Accepted1May2016

Availableonline5August2016

Keywords:

Ankylosingspondylitis

TLR-4polymorphisms

HLA-B27

a

b

s

t

r

a

c

t

Objectives:Innateimmunityisinvolvedinthephysiopathologyofankylosingspondylitis

(AS),withtheparticipationofGram-negative bacteria,modulationofhumanleukocyte

antigen(HLA)B27andtheinvolvementofpatternrecognitionreceptors,suchasToll-like

receptors(TLRs).Theaimofthisstudywastoinvestigatetheclinicalcharacteristicsand

fre-quencyofTLR4polymorphisms(Asp299GlyandThr399Ile)inacohortofBrazilianpatients

withAS.

Methods:Across-sectionalstudywascarriedoutinvolving200patientswithadiagnosis

ofASandahealthycontrolgroupof200individuals.Diseaseactivity,severityand

func-tionalcapacityweremeasured.ThestudyofTLR4polymorphismswasperformedusingthe

restrictionfragmentlengthpolymorphismmethod.HLA-B27wasanalyzedbyconventional

polymerasechainreaction.TheIBMSPSSStatistics20programwasusedforthestatistical

analysis,withp-valueslessthan0.05consideredsignificant.

Results:Meanageanddiseasedurationwere43.1±12.7and16.6±9.2years,respectively.

Thesamplewaspredominantlymale(71%)andnon-Caucasian(52%).A totalof66%of

thegroupofpatientswerepositiveforHLA-B27.Thesampleofpatientswascharacterized

bymoderatefunctionalimpairmentandahighdegreeofdiseaseactivity.Nosignificant

associationwasfoundbetweenthetwoTLR4polymorphismsandsusceptibilitytoAS.

Conclusions:TLR4polymorphisms399and299werenotmorefrequentinpatientswithAS

incomparisontothehealthcontrolsandnoneoftheclinicalvariableswereassociatedwith

thesepolymorphisms.

license(http://creativecommons.org/licenses/by-nc-nd/4.0/).

∗ _{Corresponding}_author_.

E-mail:mpinheiro@uol.com.br(M.M.Pinheiro).

http://dx.doi.org/10.1016/j.rbre.2016.07.004

(2)

Características

clínicas

e

frequência

de

polimorfismos

em

TLR4

em

pacientes

brasileiros

com

espondilite

anquilosante

Palavras-chave:

Espondiliteanquilosante

PolimorfismosemTLR-4

HLA-B27

r

e

s

u

m

o

Objetivos: Aimunidadeinataestáenvolvidanafisiopatologiadaespondiliteanquilosante

(EA),comaparticipac¸ãodebactériasgram-negativas,modulac¸ãodoantígenoleucocitário

humano(HLA)B27eoenvolvimentodereceptoresdereconhecimentodepadrões,comoos

receptoresToll-like(TLR).Oobjetivodesteestudofoiinvestigarascaracterísticasclínicasea

frequênciadepolimorfismosemTLR4(Asp299GlyeThr399Ile)emumacoortedepacientes

brasileiroscomEA.

Métodos: Fez-seum estudotransversalqueenvolveu200pacientescomdiagnósticode

EAeumgrupocontrolesaudávelde200indivíduos.Mediram-seaatividadedadoenc¸a,a

gravidadeeacapacidadefuncional.OestudodospolimorfismosemTLR4foifeitocomo

métododepolimorfismodefragmentosderestric¸ão.OHLA-B27foianalisadoporreac¸ão

emcadeiadapolimeraseconvencional.Usou-seoprogramaSPSSStatistics20daIBMpara

aanáliseestatísticaeforamconsideradossignificativosvaloresdepinferioresa0,05.

Resultados: Amédiadeidadeeadurac¸ãodadoenc¸aforamde43,1±12,7e16,6±9,2anos,

respectivamente.Aamostrafoipredominantementedosexomasculino(71%)edenão

bran-cos(52%).Dogrupodepacientes66%eramHLA-B27positivos.Aamostradepacientesfoi

caracterizadaporumaalterac¸ãofuncionalmoderadaeumelevadograudeatividadeda

doenc¸a.Nãofoiencontradaassociac¸ãoestatisticamentesignificativaentreospolimorfismos

emTLR4easusceptibilidadeàEA.

Conclusões: OspolimorfismosemTLR4399e299nãoforammaisfrequentesempacientes

comEAemcomparac¸ãocomcontrolessaudáveisenenhumadasvariáveisclínicasesteve

associadaaessespolimorfismos.

BY-NC-ND(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Introduction

Pattern-recognition receptors (PRRs) are a set of receptors

involvedin recognitionofpathogens in multicellular

orga-nisms.Toll-like receptors(TLRs)1_function_as_PRRs_and_play

anessentialroleintherecognitionofmicrobialcomponents

and endogenous ligandsinduced duringthe inflammatory

response.2–4

AmongthegenepolymorphismsofTLR,someofthemost

widelystudiedaretwoco-segregatedfunctionalmutationsin

theextracellulardomainofhumanTLR4,whicharelocatedon

chromosome4andareassociatedwithhyporesponsiveness

tobacteriallipopolysaccharides(LPS).5_Based_on_its_evidence

ofassociation withanincreasedriskofinfectionby

Gram-negativebacteria,5_such_{polymorphisms}_have_been_evaluated

insomeinflammatorydiseasesinwhichtheparticipationof

thesemicroorganismshasbeenimplicatedinthe

etiopathol-ogy,suchasankylosingspondylitis(AS).

Basedonthepremiseofsubclinicalcolitisinpatientswith

ASandanimalmodelsthatdemonstratetheparticipationof

aninfectioustriggerbyGram-negativebacilli6,7_modulated_by

thepresentationofthe antigentoHLAB27,8 _the_aim_of_the

present study was to identifythe frequencyofTLR4

poly-morphisms (Asp299Glyand Thr399IIe) inBrazilianpatients

withASandinvestigatepossibleassociationsbetweenthese

polymorphismsand greater susceptibilitytothe disease as

wellasclinicalandlaboratoryaspectsofdiseaseactivityand

chronicity.

TheresultsoftheassociationbetweenTLRpolymorphisms

and AS have been controversial in some clinical trials,9

probably becauseof differentpopulation studied.This fact

motivates investigation ofthis association inmiscigenated

populations,likeBrazilianone.

Materials

and

methods

TwohundredpatientswithadiagnosisofASaccordingtothe

modifiedNewYorkcriteria10_or_axial_{spondyloarthritis}11_were

recruitedfromthespondyloarthritisclinicofthe

Rheumato-logySectoroftheFederalUniversityofSaoPaulo(Brazil)and

200healthyindividualswereselectedfromamongvolunteers

givingbloodatthehospitalofthesameinstitutionbetween

May2011andOctober2013.Allparticipantsagreedto

partici-pateinthestudybysigningastatementofinformedconsent

(ethicscommitteenumber1804/10).

Demographic and clinical data were collected and

spe-cificassessmenttoolswereemployedforthecharacterization

ofdisease activity andseverity, such asthe Bath

Ankylos-ingSpondylitisDiseaseActivityIndex(BASDAI),12_Ankylosing

SpondylitisDiseaseActivityScore(ASDAS),13 _Bath

Ankylos-ing SpondylitisFunctionalIndex (BASFI),14 _Bath_Ankylosing

Spondylitis Metrology Index (BASMI)15 _and _modified _Stoke

AnkylosingSpondylitisSpineScore(mSASSS).16

TheanalysisofAsp299GlyandThr399Ilepolymorphisms

of TLR4 was performed using polymerase chain reaction

(3)

(NCOIfor299andHinfIfor399).17,18_GenomicDNA_extraction

was performed using whole blood collected in

ethylenedi-aminetetraacetate (EDTA) usinga DNA extractionkit (DNA

NucleoSpin®, Macherey-Nagel). The PCRamplification was

performedwith50ngoftheDNAtobestudiedinatotal

vol-umeof20␮Lcontaining0.8␮Lofpotassiumchloride50mM,

2␮L ofTris(pH 8.4),0.6mMof magnesium chloride,0.4␮L

ofeachprimer(10nM),0.4␮Lofdeoxyribonucleotidemixture

(dATP,dCTP,dGTPanddTTP)and0.06␮LunitsofTaqplatinum

DNApolymerase(0.015U/␮L).

Theautomaticthermalcycler(MJResearchPTC-200)was

programmedforamplification:initialdenaturation(95◦_C_for

4min),followedby35cyclesof95◦_C_for₄₅_s,₅₅◦_C_for₃₀_s_and

72◦_C_for₁_min₃₀_s,_with_final_extension_at₇₂◦_C_for₁₀_min._The

forwardandreverseprimerswererespectively5′_-GAT_TAG_CAT

ACTTAGACTACTACCTCCATG-3′_and₅′_-GAT_CAA_CTT_CTG

AAAAAGCATTCCCAC-3′_for_Asp299Gly_and₅′_-GGT_TGC_TGT

TCTCAAAGTGATTTTGGGAGAAand5′_-CCT_GAA_GAC_TGG

AGAGTGAGTTAAATGCT-3′_for_Thr399Ile._{Electrophoresis}_in

2%agarosegelwasperformedtoconfirmtheDNA

amplifica-tion.

Analiquotof5␮Lwiththeappropriaterestrictionenzyme

wasusedfordigestionofthePCRproductat37◦_C_for₂_h.

Elec-trophoresiswasperformedin4%agarosegel(Agarose1000

Invitrogen, Eugene,OR, USA)for identification ofthe TLR4

alleles.ThegelwasstainedwithSybrGold(Nucleicacidgel

stain,Invitrogen,Eugene,OR,USA)andvisualizedusingthe

Storm849system(MolecularDynamics,USA).

Table1–Clinicalanddemographicdataofpatientswithankylosingspondylitisandhealthycontrols.

Clinicalcharacteristics Patients

(n=200)

Controls (n=200)

p

Age(years) 43.1±12.7 38.5±11.2 0.001

Malegender(n,%) 142(71%) 121(60.5%) 0.027

Caucasianethnicity(n,%) 96(48%)

Symptomduration(years) 16.6±9.2

Timesincediagnosis(years) 7.6±6.2

FamilyhistoryofAS(n,%) 46(23%)

PresenceofHLA-B27(n,%) 130(66%)

Peripheralinvolvement(n,%)

Arthritis(pastorcurrent) 99(49.5%) Enthesitis(pastorcurrent) 129(64.5%)

Isolatedaxialinvolvement(n,%) 42(21%)

Hipinvolvement(n,%) 26(13%)

Anterioruveitis(pastorcurrent)(n,%) 74(37%)

Currentlivinghabits(n,%)

Smoking 19(9.5%)

Regularphysicalactivity 44(22%)

Specificdiseaseindices

BASDAI 2.25±2.02

BASMI 4.28±2.30

BASFI 3.80±2.52

ASDAS-Sed.rate 2.20±1.06

ASDAS-CRP 2.07±1.08

HAQ-S 1.05±1.35

mSSASS 19.0±22.32

Testsforinflammatoryactivity

Sed.rate(mm) 21.92±20.47

CRP-us(mg/dL) 8.96±13.53

Currentconventionaltreatment(n,%)

NSAIDs 125(62.5%)

Glucocorticoids 12(6%)

Methotrexate 34(17%)

Sulfasalazine 18(9%)

TNF˛inhibitors(n,%) 81(40.5%)

Infliximab 35(17.5%)

Etanercept 24(12%)

Adalimumab 22(11%)

AS,ankylosingspondylitis;NSAID,non-steroidalanti-inflammatorydrug;BASDAI,Bathankylosingspondylitisdiseaseactivityindex;BASMI, Bathankylosingspondylitismetrologyindex;BASFI,Bathankylosingspondylitisfunctionindex;ASDAS,Ankylosingspondylitisdiseaseactivity score;HAQ-S,Healthassessmentquestionnaire-Spondylitis;mSSASS,modifiedStokesankylosingspondylitisspinescore;Sed.rate,blood sedimentationrate;CRP-us,ultrasensitiveC-reactiveprotein.

(4)

TheanalysisofHLA-B27wasperformedusingconventional

PCRinanautomaticthermalcycler (MJResearchPTC-200):

70ngofDNAfromeachsample,0.9␮mol/Lofeachforwardand

reverseprimer,1.1mmol/Lofmagnesiumchloride,200␮mol/L

ofdeoxyribonucleotidemixture(dATP,dCTP,dGTPanddTTP)

and2UofTaqDNApolymeraseforatotalvolumeof25␮L.The

forwardandreverseprimersforthereactionwererespectively

E91S(5′_-GGG_TCT_CAC_ACC_CTC_CAG_AAT-3′₎_and_136AS₍₅′_-CGG

CGGTCCAGGAGCT-3′_)._The_cycling_conditions_were₁₀₀_s_at

94◦_C,_followed_by₃₀ _cycles_of₁_min_at₉₄◦_C,₁_min_at₆₄◦_C

and2minat72◦_C,_with_final_extension_at₇₂◦_C_for₁₀_min_in

40consecutivecycles.

Toevaluatethereactionquality(internalcontrol),reactions

were performedusing primers for␤-globinfor all samples

underthesameconditionsusedforthe HLA-B27reactions.

TheprimersforthesereactionswerePCO4(5′_-CAA_CTT_CAT

CCACGTTCACC-3′₎_and_GH20₍₅′_-GAA_GAG_CCA_AGG_ACA_GGT

AC-3′_)._The_PCR_products_were_analyzed_through

electrophore-sisin1%agarosegelrunforonehourat100V.

Numerical data were expressed as mean and standard

deviation.TheKolmogorov–Smirnovtestwasusedto

deter-mine the distribution ofthe data (normal or non-normal).

EithertheMann–WhitneytestorKruskal–Wallistestwasused

forcomparisonsamongthecategoricalandnumericaldata.

Thechi-squaretestwasusedtodeterminethedistributionof

theTLR4polymorphisms betweengroupsaswellasforthe

comparisonofthecategoricalvariables.Spearman’s

correla-tion coefficientswere calculated todeterminethe strength

ofcorrelationsamongthecontinuousvariables.Bivariateand

multivariatelogisticregressionmodelswereconstructedwith

variables thatexhibited significant associationsin the

pre-vious tests. The IBM SPSS 20 program was used for the

statisticalanalysis, withp-valuesless than 0.05considered

significant.

Results

Table 1 displays the clinical and demographic data of the

200 patients. As expected, the male gender was

predomi-nant.Approximatelyonequarterofthepatientshadafamily

historyofASand halfthesamplereportedcurrent orpast

peripheralinvolvement.Nearly40%ofthepatientshad

ante-rioruveitis.Theprevalenceratesofcurrentsmokingandthe

regularpracticeofphysicalexercisewerelow.

Meandiseaseactivitywashigh,withsignificantfunctional

andmobilityimpairmentreflectingthelongdurationofthe

disease.Morethan60%ofthepatientsmaderegularuseof

anon-steroidalanti-inflammatorydrugand35%useda

syn-theticdisease-modifyinganti-rheumaticdrug.Nearlyhalfthe

patientsusedTNF␣inhibitors,withequaldistributionamong

etanercept,infliximabandadalimumab.

The study of HLA-B27 was performed on 197 (98.5%)

patientswithASand60(30%)healthyindividuals,66%(n=130)

and1.6%(n=1)ofwhomtestedpositive,respectively.

Extra-articularmanifestationswerefoundin80patients(40%),the

mostfrequentofwhichwasacuteanterioruveitis(n=74;37%),

followedbycircinatebalanitis(n=3;1.5%),nonspecificcolitis

(n=2;1%)andsterileurethritis(n=1;0.5%).

Table2–FrequencyofTLR4polymorphisms(299and 399)inpatientsandcontrols.

TLR-4polymorphisms Patients n=200

Controls n=200

p

Asp299Gly

Wild 182(91%) 178(89%) 0.505

Heterozygote 17(8.5%) 22(11%)

Homozygote 1(0.5%) 0

Thr399IIe

Wild 187(93.5%) 186(93%) 0.50 Heterozygote 13(6.5%) 14(7%)

Homozygote 0 0

Chi-squaretest.

Nostatisticallysignificantdifferencesbetweenthepatients and healthy controls were found regardingAsp299Gly and Thr399Ile polymorphisms. Due to the very low number of homozygotes,heterozygousandhomozygouspatientswere included in the same group for the tests. The 299 and 399 polymorphisms were in Hardy–Weinberg(HW) equilib-riuminthepatients(HW-2=0.73,p=0.39andHW-2=0.22, p=0.63,respectively)andcontrols(HW-2=0.68,p=0.41and

HW-2=0.26,p=0.61,respectively),demonstratingthe

conser-vationofgenotypefrequenciesacrossgenerations(Table2).A

tendency wasfoundtowardagreaterfrequencyof

non-co-segregationoftheallelesinthecontrolgroupincomparison

tothepatientswithAS(Table3).

For a more detailed analysis, the patients were

sepa-ratedintosubgroupsbasedonclinicalcharacteristics.Female

patientshadahigherbodymassindex(BMI)aswellashigher

HAQ-SandASDAS-Sed.ratescores.Femalesalsohadshorter

diseaseduration,lesserseverityofsacroiliacinvolvementas

well aslower mSASSS and BASMI scoresincomparisonto

males (Table3).Aftercontrolling forconfoundingvariables

inthelogisticregressionmodel,onlyBMI(p=0.014)andthe

BASMIscore(p=0.02)remainedsignificantlyassociatedwith

gender.

InthelogisticregressionanalysisusingpositivityforB27as

the dependentvariable,significant associationswerefound

forperipheral arthritis(p=0.039),uveitis (p=0.033)and the

useofTNF␣inhibitors(p=0.003)(Table4).Amongthepatients

positiveforHLA-B27,atendencywasfoundtowarda

predom-inanceofCaucasians(p=0.058).Moreover,greaterprevalence

rateswerefoundinthisgroupregardinganterioruveitis,more

severesacroiliacinvolvement(GradeIV),longerdisease

dura-tion and a greater frequency of biological agents (p<0.05)

(Table5).

Patients with the adult formof the disease had higher

scoresontheBASFI,BASDAIandHAQ-S,wereolderandhad

longerdiseaseduration.Thosewiththejuvenileformofthe

diseasemorefrequentreportsofafamilyhistoryofAS,made

moreuseofsulfasalazineandreportedmoresideeffectsfrom

TNF␣inhibitors.However,noneofthesevariablesremained

statistically significant in the final multiple regression

model.

Non-CaucasianpatientshadahigherBMIandBASMIscore

incomparisontoCaucasianpatients(Table6).However,only

(5)

Table3–Characteristicsofpatientswithankylosing spondylitiesaccordingtogender.

Variables Male

(n=142)

Female (n=58)

p

Ethnicity

Caucasian 67(47.2%) 29(50%)

Non-Caucasian 75(52.8%) 29(50%) 0.717

Onset

Adult 118(83.1%) 50(86.2%)

Juvenile 24(16.9%) 8(13.8%) 0.586

HLA-B27+ 96(67.6%) 34(58.6%) 0.158

Arthritis 68(47.9%) 31(53.4%) 0.475

Entesitis 93(65.5%) 36(62%) 0.646

Anterioruveitis 53(37.3%) 21(36.2%) 0.882 Familyhistoryof

AS

36(25.3%) 10(17.2%) 0.216

Pureaxial involvement

29(20.4%) 13(22.4%) 0.754

BMI(kg/m2₎ _24.81_±_3.72 _25.80_±_3.87 _0.02

BASMI 4.63±2.22 3.43±2.27 0.001

mSASSS 22.54±23.34 10.64±17.2 <0.0001 ASDAS-Sed.rate 2.11±1.05 2.42±1.04 0.036

HAQ-S 1.03±0.61 1.11±2.33 0.03

Sed.rate(mm) 19.3±19.6 28.35±21.31 0.001 Diseaseduration

(years)

17.5±9.43 14.43±8.5 0.02

Radiographicsacroilitis

II 11(7.7%) 4(7.0%) 0.044

III 66(46.5%) 38(65.5%)

IV 65(44.0%) 16(27.6%)

AS, ankylosing spondylitis; BMI, body mass index; BASMI, Bath ankylosing spondylitismetrology index; ASDAS, Ankylos-ingspondylitisdiseaseactivityscore;HAQ-S,Healthassessment questionnaire-Spondylitis; mSASSS,modified Stokes ankylosing spondylitisspinescore;Sed.rate,bloodsedimentationrate; chi-squareandMann–Whitneytests.

Thepvalueinboldmeansstatisticallysignificant.

analysis(p=0.004).Themostfrequentdegreeofradiographic sacroiliitiswasGradeIII(n=103;51.8%),followedbyGradesIV (40.5%)andII(7.5%).

AmongthepatientsonTNF␣inhibitors,20(24.7%)needed tochangeagents:five(6.2%)duetoprimaryfailure(6.2%),six (7.4%)duetosecondaryfailureand 13(16%)duetoadverse events,especiallyinfusionreactionsandinfection.Moreover, four(2%)ofthesepatientschangedTNF␣inhibitorsformore thantwoagents.

Anterioruveitiswasassociatedwithlongerdisease dura-tion(p=0.002),butlostitsstatisticalsignificanceinthefinal model.Hipinvolvementwasassociatedwithchronicityand lower disease activity scores(datanotshown). In thefinal logisticregressionmodel,longerdiseaseduration(p=0.039), a higher BASFI score (p=0.027) and a lower BASDAI score (p=0.024)remainedstatisticallysignificant.

PatientswithlongerdiseasedurationhadhighermSASSS andBASMIscoresaswellasalowerASDAS-Sed.ratescore. In the final model, only ASDAS-ESH (p=0.015) remained significant. Longer symptom duration was correlated with higherBASMI,BASFI,HAQ-SandmSSASS scoresaswell as alower BASDAIscore.Inthe finalmodel, BASMI(p<0.0001)

Table4–Characteristicsofpatientswithankylosing spondylitisaccordingtopositivityforHLA-B27.

Variables HLA-B27 positive (n=130)

HLA-B27 negative (n=67)

p

Ethnicity

Caucasian 69(53.1%) 26(39%) 0.058 Non-Caucasian 61(46.9%) 41(61.2%)

Onset

Adult 103(82.4%) 55(87.3%) 0.386 Juvenile 22(17.6%) 8(12.7%)

Arthritis 58(44.6%) 40(59.7%) 0.045

Enthesite 85(65.4%) 43(64.2%) 0.867

Uveitis 56(43.1%) 17(25.4%) 0.015

FamilyhistoryofAS 31(23.8%) 14(20.9%) 0.640

Radiographicsacroiliitis

II 9(6.9%) 6(8.9%)

III 59(45.4%) 44(65.7%)

IV 62(47.7%) 17(25.4%) 0.01

28(21.5%) 12(17.9%) 0.549

Diseaseduration (years)

8.28±6.57 6.39±5.07 0.039

TNF˛inhibitors 59(45.4%) 16(23.9%) 0.003

Chi-squareandMann–Whitneytests. AS,ankylosingspondylitis.

andBASDAI(p=0.016)remainedsignificantafterthemultiple adjustments.

Whenthepatientswereclassifiedbyremission(<1.3), mod-erate(≥1.3and<2.1),high(≥2.1and<3.5)orveryhigh(≥3.5)

Table5–Characteristicsofpatientswithankylosing spondylitisaccordingtoethnicity.

Variables Caucasian (n=96)

Non-Caucasian

(n=104)

p

Onset

Adult 78(81.3%) 90(86.5%)

Juvenile 18(18.7%) 14(13.5%) 0.308

Arthritis 47(49%) 52(0.5%) 0.883

Enthesitis 61(63.5%) 68(65.4%) 0.786

Uveitis 34(35.4%) 40(38.5%) 0.656

FamilyhistoryofAS 23(24%) 23(22.1%) 0.757

Radiographicsacroilitis

II 6(6.2%) 9(8.6%)

III 59(61.4%) 45(43.3%)

IV 31(32.3%) 50(48.1%) 0.037

21(21.8%) 21(20.2%) 0.770

BMI(kg/m2₎ _24.81_±_3.72 _25.8_±_3.87 _0.048

BASMI 3.89±2.18 4.65±2.36 0.023

AS,ankylosing spondylitis;BMI,bodymass index;BASMI,Bath ankylosingspondylitismetrologyindex.

Chi-squareandMann–Whitneytests.

(6)

Table6–Characteristicsofpatientswithankylosing spondylitisaccordingtohipinvolvement.

Variables Hipinvolvement p

Yes(n=26) No(n=174)

Age(years) 47.7±11.3 42.4±12.7 0.035 Timesince

diagnosis (years)

10.3±7.9 7.2±5.9 0.04

Diseaseduration (years)

23.4±9.6 15.6±8.8 <0.0001

BASDAI 1.20±1.43 2.42±2.06 0.001 BASMI 5.55±1.52 4.10±2.33 0.002 BASFI 5.03±2.52 3.61±2.48 0.011 mSASSS 27.33±24.46 17.77±21.80 0.035 ASDAS-CRP 1.55±0.88 2.14±1.09 0.009 HAQ-S 1.86±3.72 0.93±0.61 0.017

BASDAI,Bathankylosingspondylitisdiseaseactivityindex;BASMI, Bathankylosingspondylitismetrologyindex;BASFI,Bath anky-losingspondylitisfunctionindex;ASDAS,Ankylosingspondylitis disease activity score; mSASSS, modified Stokes ankylosing spondylitisspinescore.

Chi-squareandMann–Whitneytests.

diseaseactivitybasedontheASDASscore,thefollowing fre-quencieswerefound:14%(n=28),31.5%(n=63),37%(n=74) and17.5%(n=35),respectively.

TheBASDAI score waspositively correlated withBASFI, ASDAS-Sed. rate, ASDAS-CRP, Sed. rate, CRP and HAQ-S and negatively correlated with mSSASS. Following the lin-ear multivariate regression, BASDAI remained significantly correlated with BASFI (p=0.001), HAQ-S (p=0.047), ASDAS-Sed.rate,ASDAS-CRPandSed.rate(p<0.0001),butnotCRP (p=0.247).

Discussion

TLR4polymorphisms399and299werenotmorefrequentin BrazilianpatientswithASincomparisontothehealthcontrols becausetherewerenostatisticaldifferencesbetweenpatient and control groups. Theinfluence ofTLR4 polymorphisms in the etiopathogenesis of infections, especially by Gram-negativebacteria,is wellknown,19 _as_such_{polymorphisms}

resultinaphenotypethatislittleresponsivetoendotoxins

stemmingfrom the infectious process that cause aberrant

transduction of the signal in the presence of

microorgan-isms.However,anumberofauthorshaverecentlyquestioned

theinfluenceofpolymorphismsofthisreceptoronthe

pro-gression,severityandoutcomeofinfections,suggestingthat

factorsrelatedtothehostaremoreimportantthan

polymor-phismsperse.20,21

Bacterial or intracellular components can initiate the

inflammatoryprocessandtriggersensitizationtoan

endoge-nousantigenthroughmolecularmimicry,persistently

acti-vating adaptive innate immunity and perpetuating the

inflammatory process. The trigger may not necessarily be

pathogenic,butrathermakesofpartofthenormalresident

microbiotaandcanculminateinthedevelopmentofdiseases,

suchasAS,ingeneticallysusceptibleindividuals.2

ThereissomeevidenceofanassociationbetweenTLR

poly-morphisms (299and399)and AS,6,22,23 _but_this _association

has notbeen confirmed in other studies.24–26 _The_same _is

true forAsp896Gly.27 _The_S180L_polymorphism _of_an

adap-torproteinofTLR2and4(TIRAP),whichhasdemonstrated

toplayaprotectiveroleagainsttheoccurrenceofsystemic

lupuserythematosus,hasalsodemonstratednoassociation

withAS.28

Kyoetal.foundthatmutantmice(C3H/Hej)forTLR4did

notdeveloparthritisaftertheintra-jointinjectionofLPSfrom

E.coli,unlikethewildgroup.29_This_has_raised_the_hypothesis

thatthemutationinTLR4maydiminishtheintensityofthe

innateimmuneresponse,playingaprotectiveroleratherthan

promotingautoimmunity.

Thepresentdatashowthatthe299and399polymorphisms

ofTLR4arenotgeneticfactorsofgreatersusceptibilitytoAS,

whichisinagreementwitharecentmeta-analysisinvolving

data compiledfrom nine studieson this topic.30 _Moreover,

it isimportantto pointout thatone ofthelargest genetic

studies(TheAustralo-Anglo-AmericanSpondyloarthritis

Con-sortium), involving more than two thousand patients of

EuropeandescentwithAS,alsofoundnoassociationbetween

TLRs and susceptibility to the disease.9 _According _to _the

authorscited,themainassociationsoccurredwithtwodesert

genes(2p15and21q22)aswellaswithIL-23R,IL-1R2,ANTXR2

andERAP-1.

TheexpressionofTLR4isrelatedtotheinterfacebetween

the immunesystemand the environmentaswell asacute

andchronicinflammatoryresponsesinpatientswithAS,31–33

but polymorphisms doesn’t seem to be associated with a

greater risk of developing the disease.30 _{Population-based}

studies involving healthy individuals indicate

heterogene-ity in the geographic distributionof thesepolymorphisms.

Thefrequencyrangesfrom4to10%inCaucasians34_in_the

presentcohort,whileafrequencyof16%isreportedamong

Africans.35 _There _are _no _reports _of _this _polymorphism _in

Asians.25

ConsideringthatBrazilianpopulationishighlymixedwith

multipleethnic groups, wecannotassure similar resultsif

patientsfromotherpartsofthecountry,particularlyno

Euro-peanbackground,hadbeenincluded.

Onecannotdiscardthepossibilityofaweakassociation

thatmayberepresentativeinalargersamplesizethanthat

employedinthepresentstudy.However,astudycarriedoutin

theUnitedKingdominvolvingmorethan500patientswithAS

alsofoundnoassociation.6_The_same_has_been_true_for_other

ethnicities,suchasHungarian,Finnish,Korean,Canadianand

Dutchpopulations.Anotherpointtoconsideristhelow

preva-lenceofhomozygousgenotypesfortheTLR4polymorphisms

inthepresentcohort,whichissimilartofindingsdescribedin

otherpopulations.

IthasbeendemonstratedthattransgenicratsforHLA-B27

donotdevelopinflammationofthe enthesesor intestines,

althoughtheseanimalsdevelopgenitalandskinlesionsunder

sterileconditions,demonstratingtheroleofinfectiousagents

asaninflammatorytrigger.36_Moreover,_the_{participation}_of

the environmentisadeterminantintheemergence ofthe

disease,asonlytwotofivepercentofthepopulationpositive

forHLA-B27developsAS.Basedonthisfact,theinteraction

(7)

inthepresentsample,butwasnotdemonstrated,asno

signif-icantdifferencesinthepolymorphismswerefoundbetween

thepatientsandhealthycontrols.Moreover,thenumberof

patientswithpolymorphismswasrelativelysmall.

Regarding clinical aspects, males exhibited moresevere

axialradiographicdamageandlongerdiseasedurationthan

females, which is similar to findings reported in another

Braziliancohort37_as_well_as_in_other _populations,_although

with no significant gender difference regarding peripheral

involvement.38_However,_these_findings_should_be_interpreted

withcaution,sincetheonlysignificantdifferencesafter

con-trollingforconfoundingvariables,especiallydiseaseduration

andage,wereBASMIscore,whichwashigheramongmales,

andBMI,whichwashigheramongfemales.

Miscegenationlikelycontributedtothetwopeculiarities

found inthe present sample. Thefirst was the lesser

fre-quency of positivity for HLA-B27 (66%) in comparison to

anotherBraziliancohort(78.2%)37_and_studies_involving

Cau-casianpopulations (more than 90%).39 _Fifty-two _percent_of

thepresentcohortwasnon-Caucasian,whichbetterreflects

the Brazilianpopulationincomparisonto previousstudies

(75.5%Caucasianpatients).37_The_second_peculiarity_was_the

greaterperipheralormixedinvolvement,whichhasalsobeen

reported inother mixed-race populations.39,40 _Among_only

theCaucasianpatientsinthepresentstudy,thefrequencyof

positivityforHLA-B27was75%,whichisveryclosetothe

fre-quencyreportedinapreviousBraziliancohort39_as_well_as_that

reportedinarecentFrenchstudy.41

Thepatientswithhighchronicityindexscoresgenerally

hadlowerdiseaseactivity indexscores,suchasthe

associ-ationbetweenhipinvolvementandbothaworseBASFIand

lowerBASDAIscore.Likewise,diseasedurationwascorrelated

withalowerBASDAIandhigherBASMIscore.Interestingly,the

BASDAIscorewascorrelatedwiththeotheractivitymarkers

(ASDAS-Sed.rateandCRP)aswellasworsefunctionality,but

thecorrelationwithCRP,whichisoneofthemost

standard-izedparametersforstudyinginflammatoryactivityinpatients

withAS,42_was_{non-significant.}

Asexpected, the presenceofHLA-B27 was significantly

associatedwithextra-articularmanifestations(uveitis)43_and

lesserperipheralinvolvement,butnoassociationswerefound

withtheradiographicscoreoranyspecificASassessmenttool,

includingactivity,functionandmobility.Thisfindingreflects

themorerecentknowledgethatthisgeneticaspectisnotan

associatedfactorofapoorerprognosisregardingthe

forma-tionofnewbone.44

The prevalence of the juvenile form (16%) was similar

to that found in Caucasians (8.6–21%) Turks (13.4%) and

other Brazilian cohort (16%)45 _as _well _as _lower _than _the

ratesreportedforMexicans(28–54%)andKoreans(41.3%).46

Theinclusionofthesepatientsdidnotexert animportant

impact on the clinical, laboratory and imaging outcomes

analyzed, whichis inagreement withthe notionthat this

subgroup is part of the same spectrum of the disease.In

contrast,someauthors reportlesseraxialinvolvementand

agreaterproportionofperipheralinvolvement,especiallythe

knees,incomparisontoonsetofthediseaseaftertheageof

16years.37,45,46

The present study has limitations that should be

addressed, such as the sample size for studies on gene

polymorphismsandthehigherproportionofCaucasian

indi-viduals in this cohort, what may not reflect Brazilian AS

population.OurdatasuggestthattheseTLR4variationsare

unlikelytoplayaroleintheetiopathogenesisofAS.This

find-ing,however,doesnotexcludethepossibilitythatfunctional

abnormalitiesoftheTLRsorothermoleculesclosely

associ-atedwiththeTLRssignalingareimportantinthepathogenesis

ofspondylarthropathies.

Funding

Supported byaresearch grant from theBrazilian fostering

agencyFundac¸ãodeAmparoàPesquisadoEstadodeSãoPaulo

(FAPESP2011/05517-6).

Conflicts

of

interest

Theauthorsdeclarenoconflictsofinterest.

Acknowledgment

We thank Rheumatology and Nephrology Division, Federal

UniversityofSãoPaulo(UNIFESP/EPM).

r

e

f

e

r

e

n

c

e

s

1.MedzhitovR,Preston-HurlburtP,JanewayCAJr.Ahuman homologueoftheDrosophilaTollproteinsignalsactivationof adaptiveimmunity.Nature.1997;388(6640):394–7.

2.PollanenR,SillatT,PajarinenJ,LevonJ,KaivosojaE, KonttinenYT.MicrobialantigensmediateHLA-B27diseases viaTLRs.JAutoimmun.2009;32(3-4):172–7.

3.Pacheco-TenaC,ZhangX,StoneM,Burgos-VargasR,Inman RD.Innateimmunityinhost–microbialinteractions:beyond B27inthespondyloarthropathies.CurrOpinRheumatol. 2002;14(4):373–82.

4.TakedaK,KaishoT,AkiraS.Toll-likereceptors.AnnuRev Immunol.2003;21:335–76.

5.LeaverSK,FinneySJ,Burke-GaffneyA,EvansTW.Sepsissince thediscoveryofToll-likereceptors:diseaseconceptsand therapeuticopportunities.CritCareMed.2007;35(5):1404–10.

6.LamarqueD,NhieuJT,BrebanM,BernardeauC,

Martin-GarciaN,SzepesZ,etal.Lymphocyticinfiltrationand expressionofinduciblenitricoxidesynthaseinhuman duodenalandcolonicmucosaisacharacteristicfeatureof ankylosingspondylitis.JRheumatol.2003;30(11):2428–36.

7.DeRyckeL,KruithofE,VandoorenB,TakPP,BaetenD. Pathogenesisofspondyloarthritis:insightsfromsynovial membranestudies.CurrRheumatolRep.2006;8(4):275–82.

8.SparksJA,CostenbaderKH.Geneticsenvironment,and gene–environmentinteractionsinthedevelopmentof systemicrheumaticdiseases.RheumDisClinNorthAm. 2014;40(4):637–57.

9.ReveilleJD,SimsAM,DanoyP,EvansDM,LeoP,PointonJJ, etal.Genome-wideassociationstudyofankylosing

spondylitisidentifiesnon-MHCsusceptibilityloci.NatGenet. 2010;42(2):123–7.

(8)

11.RudwaleitM,vanderHeijdeD,LandewéR,ListingJ,AkkocN, BrandtJ,etal.ThedevelopmentofAssessmentof

SpondyloArthritisinternationalSocietyclassificationcriteria foraxialspondyloarthritis(partII):validationandfinal selection.AnnRheumDis.2009;68(6):777–83.

12.GarrettS,JenkinsonT,KennedyLG,WhitelockH,GaisfordP, CalinA.Anewapproachtodefiningdiseasestatusin ankylosingspondylitis:theBathAnkylosingSpondylitis DiseaseActivityIndex.JRheumatol.1994;21(12):2286–91.

13.MachadoP,LandeweR,LieE,KvienTK,BraunJ,BakerD,etal. AnkylosingSpondylitisDiseaseActivityScore(ASDAS): definingcut-offvaluesfordiseaseactivitystatesand improvementscores.AnnRheumDis.2011;70(1):47–53.

14.CalinA,GarrettS,WhitelockH,KennedyLG,O’HeaJ,Mallorie P,etal.Anewapproachtodefiningfunctionalabilityin ankylosingspondylitis:thedevelopmentoftheBath AnkylosingSpondylitisFunctionalIndex.JRheumatol. 1994;21(12):2281–5.

15.vanderHeijdeD,LandeweR,FeldtkellerE.Proposalofa lineardefinitionoftheBathAnkylosingSpondylitisMetrology Index(BASMI)andcomparisonwiththe2-stepand10-step definitions.AnnRheumDis.2008;67(4):489–93.

16.CreemersMC,FranssenMJ,van’tHofMA,GribnauFW,vande PutteLB,vanRielPL.Assessmentofoutcomeinankylosing spondylitis:anextendedradiographicscoringsystem.Ann RheumDis.2005;64(1):127–9.

17.JeffreysAJ,WilsonV,TheinSL.Individual-specific ‘fingerprints’ofhumanDNA.Nature.1985;316(6023):76–9.

18.LorenzE,FreesKL,SchwartzDA.DeterminationoftheTLR4 genotypeusingallele-specificPCR.Biotechniques.

2001;31(1):22–4.

19.MedzhitovR.Toll-likereceptorsandinnateimmunity.Nat RevImmunol.2001;1(2):135–45.

20.DraismaA,DorresteijnM,PickkersP,vanderHoevenH.The effectofsystemiciNOSinhibitionduringhuman

endotoxemiaonthedevelopmentoftolerancetodifferent TLR-stimuli.InnateImmun.2008;14(3):153–9.

21.JessenKM,LindboeSB,PetersenAL,Eugen-OlsenJ,BenfieldT. CommonT.N.F-alpha,IL-1beta,PAI-1,uPA,CD14andTLR4 polymorphismsarenotassociatedwithdiseaseseverityor outcomefromGramnegativesepsis.BMCInfectDis. 2007;7:108.

22.McCormackWJ,ParkerAE,O’NeillLA.Toll-likereceptorsand NOD-likereceptorsinrheumaticdiseases.ArthritisResTher. 2009;11(5):243.

23.SnelgroveT,LimS,GreenwoodC,PeddleL,HamiltonS, InmanR,etal.Associationoftoll-likereceptor4variantsand ankylosingspondylitis:acase-controlstudy.JRheumatol. 2007;34(2):368–70.

24.GergelyPJr,BlazsekA,WeiszharZ,PazarB,PoorG.Lackof geneticassociationoftheToll-likereceptor4(TLR4) Asp299GlyandThr399Ilepolymorphismswith spondylarthropathiesinaHungarianpopulation. Rheumatology(Oxford).2006;45(10):1194–6.

25.NaKS,KimTH,RahmanP,PeddleL,ChoiCB,InmanRD. AnalysisofsinglenucleotidepolymorphismsinToll-like receptor4showsnoassociationwithankylosingspondylitis inaKoreanpopulation.RheumatolInt.2008;28(7):627–30.

26.AdamR,SturrockRD,GracieJA.TLR4mutations(Asp299Gly andThr399Ile)arenotassociatedwithankylosing

spondylitis.AnnRheumDis.2006;65(8):1099–101.

27.vanderPaardtM,CrusiusJB,deKoningMH,MorreSA,vande StadtRJ,DijkmansBA,etal.Noevidenceforinvolvementof theToll-likereceptor4(TLR4)A896GandCD14-C260T polymorphismsinsusceptibilitytoankylosingspondylitis. AnnRheumDis.2005;64(2):235–8.

28.CantaertT,StoneMA,terBorgM,MoggR,DeVriesN,Wilson AG,etal.Afunctionalpolymorphismof

TIR-domain-containingadaptorproteinisnotassociated withaxialspondyloarthritis.AnnRheumDis.2008;67(5): 720–2.

29.KyoF,FutaniH,MatsuiK,TeradaM,AdachiK,NagataK,etal. Endogenousinterleukin-6,butnottumornecrosisfactor alpha,contributestothedevelopmentoftoll-likereceptor 4/myeloiddifferentiationfactor88-mediatedacutearthritis inmice.ArthritisRheum.2005;52(8):2530–40.

30.XuWD,LiuSS,PanHF,YeDQ.LackofassociationofTLR4 polymorphismswithsusceptibilitytorheumatoidarthritis andankylosingspondylitis:ameta-analysis.JointBoneSpine. 2012;79(6):566–9.

31.DeRyckeL,VandoorenB,KruithofE,DeKeyserF,VeysEM, BaetenD.Tumornecrosisfactoralphablockadetreatment down-modulatestheincreasedsystemicandlocalexpression ofToll-likereceptor2andToll-likereceptor4in

spondylarthropathy.ArthritisRheum.2005;52(7):2146–58.

32.AssassiS,ReveilleJD,ArnettFC,WeismanMH,WardMM, AgarwalSK,etal.Whole-bloodgeneexpressionprofilingin ankylosingspondylitisshowsupregulationoftoll-like receptor4and5.JRheumatol.2011;38(1):87–98.

33.YangZX,LiangY,ZhuY,LiC,ZhangLZ,ZengXM,etal. IncreasedexpressionofToll-likereceptor4inperipheral bloodleucocytesandserumlevelsofsomecytokinesin patientswithankylosingspondylitis.ClinExpImmunol. 2007;149(1):48–55.

34.CarvalhoA,MarquesA,MacielP,RodriguesF.Studyof disease-relevantpolymorphismsintheTLR4andTLR9genes: anovelmethodappliedtotheanalysisofthePortuguese population.MolCellProbe.2007;21(4):316–20.

35.MockenhauptFP,CramerJP,HamannL,StegemannMS, EckertJ,OhNR,etal.Toll-likereceptor(TLR)polymorphisms inAfricanchildren:commonTLR-4variantspredisposeto severemalaria.ProcNatlAcadSciUSA.2006;103(1): 177–82.

36.TaurogJD,RichardsonJA,CroftJT,SimmonsWA,ZhouM, Fernandez-SueiroJL,etal.Thegermfreestateprevents developmentofgutandjointinflammatorydiseasein HLA-B27transgenicrats.JExpMed.1994;180(6):2359–64.

37.Sampaio-BarrosPD,BertoloMB,KraemerMH,NetoJF,Samara AM.Primaryankylosingspondylitis:patternsofdiseaseina Brazilianpopulationof147patients.JRheumatol.

2001;28(3):560–5.

38.LeeW,ReveilleJD,DavisJCJr,LearchTJ,WardMM,Weisman MH.Aretheregenderdifferencesinseverityofankylosing spondylitis?ResultsfromthePSOAScohort.AnnRheumDis. 2007;66(5):633–8.

39.GallinaroAL,VenturaC,SampaioBarrosPD,GoncalvesCR. Spondyloarthritis:analysisofaBrazilianseriescompared withalargeIbero-Americanregistry(RESPONDIAgroup).Rev BrasReumatol.2010;50(5):581–9.

40.KohemCL,BortoluzzoAB,Gonc¸alvesCR,BragadaSilvaJA, XimenesAC,BértoloMB,etal.Profileoftheuseofdisease modifyingdrugsintheBrazilianRegistryof

Spondyloarthritides.RevBrasReumatol.2015;55(1):48–54.

41.CostantinoF,TalpinA,Said-NahalR,GoldbergM,HennyJ, ChiocchiaG,etal.Prevalenceofspondyloarthritisin referencetoHLA-B27intheFrenchpopulation:resultsofthe GAZELcohort.AnnRheumDis.2015;74:689–93.

42.MachadoP,LandeweR.Spondyloarthritis.Isittimetoreplace BasdaiwithAsdas?NatRevRheumatol.2013;9(7):388–90.

43.KhanMA.Update:thetwentysubtypesofHLA-B27.CurrOpin Rheumatol.2000;12(4):235–8.

(9)

45.DuarteAP,MarquesCDL,BortoluzzoAB,Gonc¸alvesCR,da SilvaJA,XimenesAC,etal.Epidemiologicprofileof

juvenile-onsetcomparedtoadultonsetspondyloarthitisina largeBraziliancohort.RevBrasReumatol.2014;54(6): 424–30.

46.BaekHJ,ShinKC,LeeYJ,KangSW,LeeEB,YooCD,etal.