Assessment of global DNA methylation in peripheral blood cell subpopulations of early rheumatoid arthritis before and after methotrexate

RESEARCH ARTICLE Open Access Assessmen o global DNA me hyla ion in pe iphe al blood cell subpopula ions o ea ly heuma oid a h i is be o e and a e me ho exa e Ma ía C. de And es 1 , E a Pe ez-Pampin 1 , Manuel Calaza 1 , F ancisco J. San acla a 1 , Ignacio O ea 1 , Juan J. Gomez-Reino 1,2 and An onio Gonzalez 1* Abs ac In oduc ion: DNA me hyla ion is an epigene ic mechanism egula ing gene exp ession ha has been insu icien ly s udied in he blood o heuma oid a h i is (RA) pa ien s, as only T cells and o al pe iphe al blood mononuclea cells (PBMCs) om pa ien s wi h es ablished RA ha e been s udied and wi h con lic ing esul s. Me hod: Fi e majo blood cell subpopula ions: T, B and NK cells, monocy es, and polymo phonuclea leukocy es, we e isola ed om 19 ea ly RA pa ien s and 17 heal hy con ols. Pa ien samples we e aken be o e and 1 mon h a e he s a o ea men wi h me ho exa e (MTX). Analysis included DNA me hyla ion wi h high-pe o mance liquid ch oma og aphy-elec osp ay ioniza ion- andem mass spec ome y-selec ed eac ion moni o ing (HPLC-ESI-MS/MS-SRM) and exp ession le els o se en me hyla ion-speci ic enzymes by quan i a i e polyme ase chain eac ion(qPCR). Resul s: Disease-modi ying an i- heuma ic d ug (DMARD)-naï e ea ly RA pa ien s showed global DNA hypome hyla ion in T cells and monocy es, oge he wi h a lowe exp ession o DNA me hyl asn e ase 1 (DNMT1), he main enance DNA me hyl ans e ase, which was also dec eased in B cells. Fu he mo e, signi ican ly inc eased exp ession o en-ele en ansloca ion1 (TET1), TET2 and TET3, enzymes in ol ed in deme hyla ion, was ound in monocy es and o TET2 in T cells. The e was also modes dec eased exp ession o DNMT3A in B cells and o g ow h a es and DNA-damage-inducible p o ein 45A (GADD45A) in T and B cells. T ea men wi h MTX e e ed hypome hyla ion in T cells and monocy es, which we e no longe di e en om con ols, and inc eased global me hyla ion in B cells. In addi ion, DNMT1 and DNMT3A showed a end o e e sion o hei dec eased exp ession. Conclusions: Ou esul s con i m global DNA hypome hyla ion in pa ien s wi h RA wi h speci ici y o some blood cell subpopula ions and hei e e sal wi h me ho exa e ea men . These changes a e accompanied by pa allel changes in he le els o enzymes in ol ed in me hyla ion, sugges ing he possibili y o egula ion a his le el. In oduc ion Epigene ics has become an a ea o in e es o he s udy o heuma oid a h i is (RA) [1]. I e e s o s able bu e e s- ible changes in gene exp ession ha a e he i able h ough cell di isions bu do no in ol e DNA a ian s. They main ain he cellula pheno ypes acqui ed du ing de elop- men and di e en ia ion, and e lec physiological changes and en i onmen al s ess. This so o cellula memo y is used o p omo e adap i e pheno ypic changes ha esul in inc eased i ness, bu when abe an hey could p o- mo e o pe pe ua e a disease s a us. Epigene ic mecha- nisms include DNA me hyla ion, his one modi ica ions, mic oRNA, o he non-coding RNA and nucleosome posi- ioning. Cu en in e es is based on indings indica ing ha epigene ic changes can become bioma ke s o di e - en ia e pa ien s om heal hy con ols and o sepa a e pa ien subg oups on p ognosis, o on esponse o ea - men , as well as o se e as new a ge s o ea men [2]. In addi ion, changes in DNA me hyla ion o blood cells ha e been cha ac e ized as media o s o gene ic isk in RA and a e o in e es o unde s and disease pa hogenesis [3]. * Co espondence: [email p o ec ed] 1 Labo a o io de In es igacion 10 and Rheuma ology Uni , Ins i u o de In es igación Sani a ia-Hospi al Clínico Uni e si a io de San iago, T a esia de Choupana, s/n, 15706 San iago de Compos ela, Spain Full lis o au ho in o ma ion is a ailable a he end o he a icle © 2015 de And es e al. Open Access This a icle is dis ibu ed unde he e ms o he C ea i e Commons A ibu ion 4.0 In e na ional License (h p://c ea i ecommons.o g/licenses/by/4.0/), which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o he C ea i e Commons license, and indica e i changes we e made. The C ea i e Commons Public Domain Dedica ion wai e (h p://c ea i ecommons.o g/publicdomain/ze o/1.0/) applies o he da a made a ailable in his a icle, unless o he wise s a ed. de And es e al. A h i is Resea ch & The apy (2015) 17:233 DOI 10.1186/s13075-015-0748-5 DNA me hyla ion is he mos widely s udied and well- cha ac e ized epigene ic change [4]. I happens mainly as 5-me hylycy osine (5mC) a C–phospha e–Gdinucleo- ides (CpG) by he enzyma ic ans e o a me hyl g oup om S-adenosyl-L-me hionine (SAM). In he bulk gen- ome CpG a e a e and highly me hyla ed, bu in clus e s o CpG dinucleo ides called CpG islands hey a e usually non-me hyla ed. When CpG islands a gene p omo e s a e me hyla ed hey a e associa ed wi h long- e m silen- cing o gene exp ession. In con as , a iable and issue- speci ic me hyla ion akes place wi h p e e ence ou side CpG islands [4]. Main enance o he me hyla ed s a us h ough mi osis equi es a speci ic DNA me hyl ans e - ase (DNMT), DNMT1, which ecognizes hemime hyla ed DNA sequences and me hyla es he new DNA s and. Two o he enzymes o he same amily, DNMT3A and DNMT3B, ha e been cha ac e ized as equi ed o de no o me hyla ion du ing de elopmen o in esponse o en i onmen al s imuli o o he s ess. Ac i e deme hyla- ion depends on he ac ion o he en-ele en ansloca ion (TET) (TET1, 2 and 3) enzymes [5]. These enzymes p o- mo e DNA deme hyla ion by con e ing 5mC o 5- hyd oxyme hylcy osine (5hmC), which hey u he oxidize in o 5- o mylcy osine and 5-ca boxylcy osine. The g ow h a es and DNA-damage-inducible p o ein 45A (GADD45A) enzyme could also con ibu e o deme hyla- ion [6], especially in au oimmune diseases [7, 8]. Global DNA hypome hyla ion has been ound in se - e al in lamma o y and au oimmune diseases whe e i en- ails abe an exp ession o genes and ibosomal RNAs p obably implica ed in hei pa hology [9, 10]. O he changes associa ed wi h DNA hypome hyla ion as genome ins abili y and mu a ions, o use o c yp ic p omo e s ha e no been desc ibed in he au oimmune diseases. In RA, DNA hypome hyla ion has been ho oughly demons a ed in ib oblas -like syno ial (FLS) cells [11–14]. Hypome hy- la ion a speci ic CpG si es in FLS is associa ed wi h o e - exp ession o genes ha a e keys o he disease p ocess. The consequences o DNA hypome hyla ion in FLS ha e been highligh ed by he ac i a ed pheno ype ha no mal FLS acqui e a e d ug (5-azacy idine)-induced deme hyla- ion [11]. Howe e , many aspec s o DNA me hyla ion in RA emain incomple ely explo ed o a e con o e sial. Fo example, he e a e con lic ing epo s ega ding changes in DNA me hyla ion o blood cells [15, 16], o he exp ession o me hyl ans e ase enzymes [11, 13, 16]. Also, we do no know he me hyla ion s a us o o he im- po an playe s in RA beyond FLS and blood T cells, he only cell popula ions s udied o da e [11–14, 17]. In addi ion, all he p e ious s udies ha e analyzed samples om es ablished RA pa ien s and could e lec e ec s o ea men o o disease e olu ion. In his espec , he e is a epo indica ing ha me ho exa e (MTX) e e s DNA hypome hyla ion in in lamma o y a h i is [18], in spi e o MTX inhibi ion o SAM syn hesis [19, 20]. Finally, some p e ious s udies ha e e alua ed global DNA me hyla ion wi h echniques ha a e insensi i e and unable o dis in- guish 5mC om 5hmC, which has di e en unc ional im- plica ions. These design and echnical issues could ha e con ibu ed o some o he p e ious disco dan esul s. The e o e, we aimed o add ess some o hese ques ions wi h disease-modi ying an i- heuma ic d ug (DMARD)- naï e ea ly RA pa ien s, be o e and a e ecei ing MTX and using sensi i e and accu a e echnology [21, 22]. Me hods Pa ien s and con ols New consecu i e pa ien s a i ing o he Rheuma ology Uni and ul illing he 2010 Ame ican College o Rheuma ology/Eu opean League Agains Rheuma ism (ACR/EULAR) classi ica ion c i e ia o RA we e included om Ap il 2011 o Feb ua y 2012 [23]. All had clinical symp oms o less han 2 yea s o e olu ion compa ible wi h hei classi ica ion as ea ly RA. Clinical da a and blood samples we e aken be o e and 1 mon h a e s a - ing ea men wi h MTX. Clinical ollow-up was done wi h comple e independence om his s udy. Gende and age-ma ched heal hy con ols we e ec ui ed a he same ime. All pa ien s and con ols we e o Caucasian Spanish ances y. The E hics Commi ee o Clinical Resea ch o Galicia app o ed his s udy, and w i en in o med consen was ob ained om all pa icipan s. Blood cell subpopula ions Double g adien sepa a ion by cen i uga ion was used o isola e g anulocy es and mononuclea cells wi h His opaque®-1077 and His opaque®-1119 (Sigma-Ald ich, S Louis, MO, USA) om 20 mL o EDTA an i- coagula ed blood as desc ibed [24]. Mononuclea cells o ming he bu y coa o e he 1077 laye we e ac ion- a ed by immune-magne ic posi i e selec ion in ou sub- popula ions CD56+ (na u al kille (NK) cells), CD14+ (monocy es), CD19+ (B lymphocy es) and CD3+ (T lym- phocy es). The MACS® sys em (Mil enyi Bio ek Be gisch Gladbach, Ge many) was used o NK cells and BD IMag™cell sepa a ion sys em (BD Biosciences, San Jose, CA, USA) o he o he subpopula ions. Pu i y o he isola ed cells was con olled by luo escence-ac i a ed cell so ing analysis on a FACScan™cy ome e wi h Cell- Ques P o So wa e (BD Biosciences) as double-s ained cells wi h an i-CD45 PE and subpopula ion-speci ic an ibodies labeled wi h FITC (an i-CD14, CD19, CD3 o he popula ions pu i ied wi h hese an ibodies and an i-CD15 o g anulocy es) and as CD56 + and CD3 − cells o he NK cells. All an ibodies we e om BD Biosciences. Pu i y o isola ed cell subpopula ions anged om 90 o 99 %. Pu i ied cells we e p ocessed wi h he illus a™ ipleP ep Ki (GE Heal hca e. Li le de And es e al. A h i is Resea ch & The apy (2015) 17:233 Page 2 o 9 Chal on , UK) acco ding o he manu ac u e ’sins uc- ions o ob ain genomic DNA and o al RNA. Global DNA me hyla ion 5mC ela i e le els we e quan i ied ollowing enzyma ic hyd olysis o genomic DNA as desc ibed [21]. Th ee s anda ds we e included in each analysis. They we e iden ical DNA sequences excep o he inclusion o 100 % unmodi ied cy osines, 100 % 5mC o 100 % 5hmC, e- spec i ely (Zymo Resea ch, I ine, CA, USA). Samples and s anda ds, 0.1–1μg, we e hyd olyzed o hei com- ponen nucleosides be o e analysis by incuba ion wi h DNA Deg adase Plus (Zymo Resea ch) o 2 h a 37 °C. Sepa a ion and quan i ica ion o he nucleosides was done by high-pe o mance liquid ch oma og aphy- elec osp ay ioniza ion- andem mass spec ome y- selec ed eac ion moni o ing (HPLC-ESI-MS/MS-SRM) in he API 4000 LC/MS/MS Sys em (AB Sciex, F aming- ham, MA, USA) including a ZORBAX Eclipse XDB-C18 column (Agilen Technologies, San a Cla a, CA, USA) and a iple quad upole mass spec ome e . This sys em allows he accu a e and sensi i e di e en ia ion o 5mC om he ou deoxy ibonucleosides, he ou ibonucleo- sides and 5hmC [21]. Quan i ica ion was exp essed as he a io o 5mC o o al cy osine (5mC/5mC+C). DNA me hyl an e ase exp ession analysis To al RNA om each cell subpopula ion was immedia ely e e se- ansc ibed wi h a ian myeloblas osis i us e e se ansc ip ase (P omega, Madison, WI, USA) and andom p ime s. Gene exp ession o DNMT1,DNMT3A, DNMT3B,TET1,TET2,TET3 and GADD45A was quan i- ied wi h quan i a i e polyme ase chain eac ion (qPCR). P ime s we e designed wi h he p ime Exp ess 3.0 so - wa e (Applied Biosys ems, Fos e Ci y, CA, USA) (Table S1 in Addi ional ile 1). Fi e housekeeping genes we e ini- ially es ed (RPL13A,GAPDH,B2M,18S and TBP)as e- po ed [25]. Those showing s able exp ession (TBP and 18S o all cell subpopula ions excep g anulocy es, whe e only TBP showed s able le els) in qBase [26] we e used as e e ence. Quan i ica ions we e pe o med in iplica e 10 μl eac ions con aining 1 μlcDNA,5μlRT 2 SYBR G een qPCR Mas e mix (Qiagen, Venlo, The Ne he lands), and 250 nM o each p ime in a Ro o -Gene™6000 (Co be Li e Science, Venlo, The Ne he lands) he mocycle wi h ini ial ac i a ion a 95 °C o 10 minu es, ollowed by a wo-s ep p og am o 95 °C o 15 seconds and 60 °C o 60 seconds o 45 cycles (gain = 8). S anda ds we e included in each un o in e - un calib a ion. Speci ici y o he PCR eac ions was con i med by mel ing cu e analysis o he p oduc s as well as by size e i ica ion by DNA elec o- pho esis. T ans o med exp ession da a we e analyzed wi h qBase a e adjus ing o ampli ica ion e iciency o each ansc ip [26]. S a is ical analysis Da a analysis was pe o med wi h S a is ica 7.0 (S a So , Tulsa, OK, USA). Non-no mali y o he a iables was co ec ed wi h loga i hm o powe ans o ma ions. Di - e ences be ween pa ien s and con ols we e e alua ed using main e ec s analysis o a iance wi h co a ia es (ANCOVA). Co a ia es we e sex and age. Wi hin- pa ien compa isons, be o e and a e MTX ea men , we e done wi h pai ed-samples es s. Resul s Cha ac e is ics o pa ien s, con ols and he isola ed blood cell subpopula ions A o al o 19 consecu i e pa ien s wi h RA s a ing MTX ea men we e ec ui ed (Table 1). The ac ion o men (73.7 %) was highe han he ypical o pa ien s wi h RA. Symp oms had s a ed a median o 6.0 mon hs be- o e MTX ea men wi h only h ee pa ien s su passing he yea since he i s symp oms and none wi h mo e han 20 mon hs o e olu ion. These pa ien s had no e- cei ed DMARDs be o e s a ing MTX. All had al eady been ea ed wi h low-dose me hylp ednisolone (4–10 mg/day) o a median o 23 days be o e s a ing MTX (in e qua ile ange (IQR) 12–33.5 days). Ac i i y o RA was mode a e in mos pa ien s wi h mean disease ac i - i y sco e in 28 join s (DAS28) o 4.6. Only wo o he pa- ien s showed e osions on adiog aphic explo a ion in spi e o he p esence o an i-ci ullina ed pep ide an i- bodies (ACPA) o heuma oid ac o (RF) posi i i y in mo e han hal o hem (63.2 % se oposi i e). The ini ial dose o MTX was 10 mg/week excep o one pa ien who ecei ed 15 mg/week. A second sample om each pa ien was aken 1 mon h a e s a ing MTX ea men o assess he e ec on DNA me hyla ion and exp ession o DNMTs, TETsandGADD45A. All pa ien s we e in MTX mono he apy a ha ime and wi hou change in MTX o me hylp ednisolone doses. Al hough 1 mon h is oo ea ly o assessing esponse o MTX, a dec ease o RA ac i i y was al eady e iden in mos pa ien s (Table 1). Response o MTX was assessed a 6-mon h ollow-up. A ha ime, 15 pa ien s emained in mono- he apy wi h MTX, 12 o hem showed good esponse acco ding o he EULAR c i e ia [27], one showed a mode a e esponse and wo we e non- esponde s. O he ou pa ien s no emaining on MTX mono he apy a 6 mon hs, wo we e on a di e en DMARD due o in- e icacy o MTX and wo had in e up ed MTX due o ad e se e ec s. A o al o 17 heal hy con ols ma ched o age and sex we e ec ui ed and s udied. Th ee subpopula ions o blood cells monocy es, B and T lymphocy es we e isola ed om he 19 pa ien s and 17 heal hy con ols. G anulocy es and NK cells we e also isola ed in he i s eigh pa ien s and eigh con ols bu no pu sued u he because no di e ences we e de And es e al. A h i is Resea ch & The apy (2015) 17:233 Page 3 o 9 obse ed. The esul s om hese wo blood subpopula- ions will no be p esen ed in de ail. Pu i y o he iso- la ed subpopula ions anged om 90 o 99 %. Di e ences in global DNA me hyla ion be ween pa ien s wi h RA and con ols Le els o global 5mC we e simila o hose p e iously epo ed [21, 22]. Compa ison o he global 5mC le el be ween pa ien s and con ols was done wi h he sam- ples o RA pa ien s be o e MTX ea men . Main e ec s ANCOVA wi h sex and age as co a ia es was used o hese compa isons. T lymphocy es showed signi ican DNA hypome hyla ion in ea ly RA pa ien s compa ed wi h heal hy subjec s (Fig. 1a; mean = 3.89 %, 95 % con- idence in e al (CI) = 3.80–3.99 s.4.15 %, 95 % CI = 3.96–4.38, espec i ely, P= 0.011). The mul i a ia e ana- lysis also showed signi ican DNA hypome hyla ion in he monocy es o RA pa ien s (Fig. 1c; mean = 3.96 %, 95 % CI = 3.87–4.07 s.4.13, 95 % CI = 3.98–4.33, espec i ely, P= 0.047). Al hough signi ican , hese di e ences we e small. No signi ican di e ences we e de ec ed in any o he o he blood cell subpopula ions: B lymphocy es, NK cells and polymo phonuclea (PMN) cells (Fig. 1b and no shown). No signi ican co ela ion be ween he o al me h- ylp ednisolone dose ecei ed un il blood d awing and le els o 5mC in any o he i e blood subpopula ions we e obse ed in RA pa ien s (no shown). In he same assay, 5hmC was also de e mined wi hou any signi ican di e - ence be ween pa ien s and heal hy con ols, showing e y low le els in all cell subpopula ions (a ound 0.02 % o o al cy osine) as is cha ac e is ic o mos adul issues [21, 28]. DNMT1,DNMT3A,DNMT3B,TET1,TET2,TET3 and GADD45A ela i e exp ession was de e mined by qPCR analysis in each o he blood cell subpopula ions. A signi i- can dec ease in pa ien s wi h RA was obse ed o some enzymes: DNMT1 exp ession was dec eased in T cells (Fig. 2a; mean = 5.6, 95 % CI = 4.3–7.2 s. 10.0, 95 % CI = 7.7–13.1; P= 0.0022), B cells (Fig. 2b; mean = 3.4, 95 % CI = 2.6–4.3 s. 6.0, 95 % CI = 4.6–7.4; P=0.00046) and monocy es (Fig. 2c; mean = 5.6, 95 % CI = 4.5–6.7 s. 8.4, 95 % CI = 7.4–9.4; P= 0.00020); DNMT3A was de- c eased, al hough less ma kedly, in B cells (Fig. 2d; mean = 27.5, 95 % CI = 17.9–42.2 s. 46.9, 95 % CI = 32.4–67.9; P= 0.044); no di e ence was ound in he ex- p ession o DNMT3B in any o he blood cell subpopula- ions (no shown); and GADD45A exp ession was also mode a ely educed in T cells (Fig. 2e; mean = 12.2, 95 % CI = 8.7–17.3 s. 20.0, 95 % CI = 13.9–28.8; P= 0.048) and in B cells (Fig. 2 ; mean = 9.6, 95 % CI = 5.8–45.9 s. 18.3, 95 % CI = 12.9–25.9; P= 0.021). On he con a y, he ela i e exp ession o TET enzymes was inc eased in some cell popula ions. Monocy es showed he la ges in- c eases, wi h he h ee TET enzyme genes showing highe ela i e le els o exp ession in RA pa ien s han in con- ols. TET1 and TET3 showed he mos ma ked di e ence be ween RA pa ien s and con ols (TET1, Fig. 3a; mean = 2.9, 95 % CI = 2.1–4.8 s. 1.9, 95 % CI = 1.6–2.3; P= 0.0044; and TET3, Fig. 3c; mean = 2.0, 95 % CI = 1.7–2.3 s. 1.35, 95 % CI = 1.15–1.6; P= 0.0014), bu also TET2 (Fig. 3b; mean = 2.5, 95 % CI = 1.9–3.2 s. 1.8, 95 % CI = 1.5–2.2; P= 0.019) was inc eased in monocy es o he pa- ien s wi h RA. TET2 was also modes ly inc eased in he T cells o RA pa ien s in ela ion wi h he heal hy con ols (Fig. 3d; mean = 3.4, 95 % CI = 2.4–5.2 s. 2.2, 95 % CI = 1.8–3.0; P= 0.045). No o he di e ences we e de ec ed. Inc ease o global DNA me hyla ion a e MTX ea men Global DNA me hyla ion was analyzed 1 mon h a e s a ing ea men wi h MTX. Compa ison o he pe cen - age o 5mC in each o he i e blood cell subpopula ions wi h es s o dependen a iables showed signi ican in- c eases a e ea men in h ee o hem: T cells ha showed he mos signi ican inc ease in global DNA 5mC (Fig. 4a; mean = 4.21 %, 95 % CI = 4.03–4.43 a e 1 mon h o ea men ; P= 0.0014), B cells (Fig. 4b; 4.09 %, 95 % CI = 3.94–4.28 a e ea men s. 3.94 %, 95 % Table 1 Cha ac e is ics o he pa ien s wi h ea ly RA and o he heal hy con ols included in he s udy RA pa ien s Heal hy con ols Numbe 19 17 Female, (%) 5 (26.3) 6 (35.3) Age a diagnosis, median (IQR) 61.6 (51–65) 58 (39–64) Symp oms o MTX, median mon hs (IQR) 6.0 (2.2–7.3) RF (%) 11 (57.9) ACPA (%) 11 (57.9) E osi e a h i is (%) 2 (10.5) Smoking (%) 8 (42.1) Me hylp ednisolone, median mg/day (IQR) 4(4–8) Me o h exa e, median mg/week (IQR) 10 (10–10) CRP (mg/L), median (IQR) Baseline 15.3 (3.9–23.2) 1 mon h 5.2 (1.2–9.4) ESR, median (IQR) Baseline 28.0 (6.0–38.0) 1 mon h 14.0 (6.5–24.0) DAS28, mean ± SD Baseline 4.6 ± 1.6 1 mon h 3.2 ± 1.5 RA heuma oid a h i is, IQR in e qua ile ange, MTX me ho exa e, RF heuma oid ac o , ACPA an i-ci ullina ed pep ide an ibody, CRP C- eac i e p o ein, ESR e y h ocy e sedimen a ion a e, DAS28 disease ac i i y ac o in 28 join s, SD s anda d de ia ion de And es e al. A h i is Resea ch & The apy (2015) 17:233 Page 4 o 9 CI = 3.82–4.09 be o e MTX; P= 0.018) and monocy es ha showedabo de lineinc ease(Fig.4c;4.09%;95% CI = 4.00–4.21 a e 1 mon h o ea men ; P=0.045). Howe e , some pa ien s did no ollow his end o e- co e y (Fig. 4). No di e ences in 5mC we e de ec ed in NK o PMN cells be ween baseline and a e 1 mon h wi h MTX. No de ec able a ia ion in global 5hmC le els was obse ed (da a no shown). We ha e also compa ed exp ession o DNMT1, DNMT3A,DNMT3B,GADD45A,TET1,TET2 and TET3 in he i e cell subpopula ions a e MTX ea men wi h hei baseline le els. No clea changes we e obse ed (Fig. 5). Only DNMT1 showed a end o inc eased le els a e ea men in T cells and monocy es ha was nea signi ican (P= 0.06 and 0.07, espec i ely). A no sig- ni ican end o an inc ease o DNMT3A in B cells (P= 0.06) was also obse ed. No simila end was ound o any o he o he enzymes. On he con a y, he di e - ences wi h heal hy con ols ha we e p esen be o e ea men we e s ill p esen a e 1 mon h on MTX ex- cep o TET2 in T cells ha was no longe signi ican ly di e en . In pa icula , he TET enzyme exp ession in monocy es showed he same le el o di e ence wi h heal hy con ols han be o e ea men (TET1 P be o e = Fig. 1 Di e ences in global DNA me hyla ion be ween heal hy con ols (HC) and ea ly heuma oid a h i is (RA) pa ien s. The pe cen age o 5-me hylcy osine (5mC) o e he o al con en o cy osine in o al DNA o aT cells, bB cells, and cmonocy es is shown. Each do ep esen s a subjec . Ho izon al ba s wi h an emp y squa e a e means. Di e ences in (a)and(c) we e signi ican , P<0.05 Fig. 2 Exp ession o me hyla ion enzymes in blood cells o heal hy con ols (HC) and ea ly heuma oid a h i is (RA) pa ien s. No malized ela i e exp ession o DNMT1 ob ained by quan i a i e polyme ase chain eac ion (qPCR) in aT cells, bB cells, and cmonocy es, DNMT3A in B cells (d) and o GADD45A in T cells (e) and B cells ( ) is shown. O he con en ions a e as in Fig. 1. All hese compa isons we e signi ican . No compa ison in o he cells subpopula ion and none o he DNMT3B analyses we e signi ican de And es e al. A h i is Resea ch & The apy (2015) 17:233 Page 5 o 9 0.037, P a e = 0.039; TET2 P be o e = 0.019, P a e = 0.013; and TET3 P be o e =0.0014,P a e =0.0033),aswellas he exp ession o GADD45A in T and B cells (T cells P be o e = 0.048, P a e =0.049;BcellsP be o e =0.021,P a e =0.041). The e o e, TET enzyme gene exp ession in monocy es and GADD45A in T and B cells we e no signi ican ly modi ied by his ea men . Discussion Ou esul s suppo he in ol emen o dec eased DNA me hyla ion in RA and i s e e sal in esponse o MTX. They also widen he spec um o changes in me hyla ion and in he enzymes egula ing his p ocess in RA, as well as p o ide cell-speci ic esul s. In pa icula , his is he i s ime ha changes in B cells and in monocy es ha e Fig. 3 Di e ences be ween heal hy con ols (HC) and ea ly heuma oid a h i is (RA) pa ien s in en-ele en ansloca ion (TET) enzymes exp ession. No malized ela i e exp ession o aTET1,bTET2, and cTET3 in monocy es; and o dTET2 in T cells. Con en ions a e as in Fig. 1. All di e ences we e signi ican a P< 0.05 Fig. 4 Inc eased global DNA me hyla ion in heuma oid a h i is (RA) pa ien s a e 1 mon h on me ho exa e (MTX). The pe cen age o 5-me hylcy osine (5mC) o e he o al con en o cy osine in o al DNA o aT cells, bB cells, and cmonocy es is shown. Each pai o do s joined by dashed lines ep esen s a subjec be o e and a e MTX. All changes we e signi ican wi h P<0.05 de And es e al. A h i is Resea ch & The apy (2015) 17:233 Page 6 o 9 been desc ibed and ha he main me hyla ion- ela ed enzymes ha e been s udied in he blood o RA pa ien s. The esul s suppo he implica ion o DNA hypome hy- la ion in RA and highligh he in e ac ion be ween MTX and DNA me hyla ion ha could con ibu e o i s he a- peu ic e icacy in RA. A couple o p e ious s udies ha e analyzed whole pe iphe al blood mononuclea cells (PBMCs) wi h dis- co dan esul s [15, 16]. One o hese s udies ound global DNA hypome hyla ion in PBMCs wi h an enzyme-linked immunoso ben assay (ELISA) ki o ques ionable accu acy gi en he low pe cen age o 5mC (1.22 % in con ols) [16] in compa ison wi h ou esul s, o he esul s in RA T cells [17], and in o he cells and issues [21, 22]. The second s udy did no ind di e ences in DNA me hyla ion be ween PBMCs o RA disco dan monozygo ic wins a 1505 CpG si es analyzed by bead a ay [15]. Howe e , hese si es we e no andomly dis ibu ed and hei ep e- sen a i eness o global DNA me hyla ion is unce ain, and could be insensi i e o de ec he small di e ence we ha e obse ed. The mos speci ic s udy o da e has analyzed PBMCs in wo subse s, T cells and non- T cells, wi h a HPLC me hod [17]. The T cells showed global DNA hypome hyla ion, whe eas he non-T cells did no . These esul s a e in ag eemen wi h ou obse a ions. Meaning o global DNA hypome hyla ion in au o- immune diseases has been bes s udied in sys emic lupus e y hema osus (SLE) CD4 T cells. These cells show global and si e-speci ic hypome hyla ion ha is associa ed wi h inc eased exp ession o immune esponse genes, many o hem in he in e e on (INF) signaling pa hway, o e ex- p ession o ibosomal RNA, modi ica ions o imp in ing, and eac i a ion o endogenous e o i us [8–10, 29]. All hese changes could con ibu e o he b eakdown o im- mune ole ance and o ch onic in lamma ion. Suppo o his causal ole has been p o ided by he d ug-induced SLE ha ollows ea men wi h hyd alazine o p ocainamide, which dec ease DNA me hyla ion, and by animal s udies [8–10]. Al hough some o he abo e- men ioned changes could be SLE-speci ic, i is likely ha DNA hypome hyla ion in RA T cells has a simila ole. In his ega d, he e is al eady e idence showing ha speci ic CpG si es a TNFSF5 a e simila ly hypome hyla ed in T cells o SLE and RA [30]. In addi ion, a ecen la ge s udy in RA has ound e idence o me hyla ion changes in PBMCs as possible media o s in gene ic suscep ibili y [3]. As al eady men ioned, global DNA me hyla ion has no been p e iously analyzed in monocy es o RA pa- ien s. In SLE, monocy es sha e se e al hypome hyla ed si es wi h CD4 T cells, al hough hey we e mo e nume - ous and p ominen ly hypome hyla ed in T cells [29]. In RA monocy es, he me hyla ion le el a a CpG si e in he IL6 p omo e ha was hypome hyla ed in RA PBMCs was in e sely co ela ed wi h in e leukin (IL)-6 lipopolysaccha ide (LPS)-induced exp ession [31]. I can only be p esumed ha changes in me hyla ion as his one could also con ibu e o he disease p ocess. DNMT1 was he mos ma kedly and consis en ly de- c eased DNA me hyl ans e ase in ou s udy. This esul is in con as wi h a p e ious epo ha ound inc eased DNMT1 exp ession in PBMCs o RA pa ien s [16], bu i is in ag eemen wi h mul iple obse a ions in o he au oimmune diseases [8, 9, 15]. In addi ion, impai ed DNMT1 unc ion has a causal ole in global DNA hypo- me hyla ion and au oimmuni y as demons a ed by d ug- induced SLE [9]. In RA, s udies o DNMT1 ha e been much mo e limi ed and es ic ed o FLS in cul u e. They show compa able le els o he obse ed in os eoa h i is (OA) FLS, bu DNMT1 exp ession is ma kedly down egu- la ed by incuba ion wi h small amoun s o in lamma o y cy okines [11, 13]. These p e ious epo s sugges possible mechanisms o he dec eased DNMT1 exp ession and in- dica e ha down egula ion o DNMT1 is e y likely a majo ac o in DNA hypome hyla ion. Howe e , co el- a ion be ween DNMT1 le els and DNA hypome hyla ion was no p esen in B cells in ou s udy. The lack o Fig. 5 La ges changes in exp ession o me hyla ion enzymes a e 1 mon h on me ho exa e (MTX). No malized ela i e exp ession o aDNMT1 in T cells, and bin monocy es, and cDNMT3A in B cells, be o e and a e 1 mon h on MTX is shown. Figu e con en ions a e as in Fig. 4. The h ee changes showed P< 0.1, bu none was signi ican de And es e al. A h i is Resea ch & The apy (2015) 17:233 Page 7 o 9 hypome hyla ion in B cells was especially ema kable be- cause hey we e he only cell ype showing dec eased ex- p ession o DNMT1 and DNMT3A.No educ iono DNMT3A le els has been p e iously desc ibed in RA FLS o in SLE T cells [13, 32–34], which a e he wo cell ypes ha ha e been s udied. In u n, no di e ences in DNMT3B ha e been desc ibed in ag eemen wi h ou nega i e esul s ega ding his enzyme [13, 32–34]. In addi ion o he changes in he enzymes ha me hy- la e DNA, ou s udy is he i s add essing he DNA deme hylases in an au oimmune disease. Exp ession o he genes o he h ee TET enzymes we e inc eased in he monocy es o RA pa ien s, and TET2 was also in- c eased in T cells. These esul s could con ibu e o- ge he wi h he dec ease in DNMT1 o he global hypome hyla ion obse ed in he pa ien s. O possible ele ance o RA is ha he TET3 locus has been associ- a ed wi h SLE suscep ibili y in Asians [35], and ha TET1 egula es ansc ip ion and p ocessing o IL1βand o he p o-in lamma o y genes in expe imen s wi h cell lines [36]. P e iously, GADD45A was he main known ac i e DNA deme hylase [6], howe e his ole seems o co espond o he TET enzymes [5]. Howe e , in e es in GADD45A has emained in he au oimmune diseases because i is o e exp essed in CD4 T cells o SLE pa ien s, co ela ing wi h global and si e-speci ic DNA hypome hyla ion in hese cells [7, 8]. In con as , GADD45A exp ession was diminished in ou analysis o RA pa ien s, bo h in T and B cells. The obse ed dec ease is no cong uen wi h global DNA hypome hyla ion in T cells gi en i s p oposed deme hyla ing unc ion, bu GADD45A has o he unc ions and one o hem could ha e a ole he e: i s ole as inhibi o o T cell ecep o (TCR) signaling [37]. Se e al o he di e ences we obse ed in RA pa ien s we e e e ed a e 1 mon h o MTX ea men . Fi s , glo- bal DNA hypome hyla ion in T cells, B cells and mono- cy es was e e ed in mos pa ien s. These esul s a e eminiscen o a smalle s udy including pa ien s wi h in- lamma o y a h i is (RA and pso ia ic a h i is (PsA)) on long- e m ea men wi h MTX [18]. This ou come is con- a y o he expec ed gi en he MTX supp ession o SAM [38], which is he majo sou ce o me hyl g oups o DNA me hyla ion. Howe e , con ol o in lamma ion by MTX could lead o educed cell p oli e a ion, which is a sou ce o SAM consump ion h ough inc eased ecycling o poly- amines [39], and o e e sal o DNMT1 and DNMT3A exp ession gi en hei sensi i i y o p o-in lamma o y cy okines [11, 13]. The wo mechanisms could con ibu e o he eco e y o DNA me hyla ion. Conclusions Ou esul s suppo he implica ion o DNA me hyla ion in RA and in i s esponse o MTX ea men . They also widen he spec um o changes and he ypes o cells ha a e a ec ed by hem: global DNA hypome hyla ion in T cells and monocy es associa ed wi h a lowe exp es- sion o DNMT1 and inc eased exp ession o he h ee TET enzymes in monocy es and TET2 in T cells, o- ge he wi h dec eased DNMT1 and DNMT3A exp ession in B cells. Se e al o hese changes we e e e ed a e MTX ea men , mos ema kably he DNA global hypo- me hyla ion. The di e ences and changes ound he e in pe iphe al blood could e lec impo an mechanisms bo h in disease e olu ion and in i s con ol by MTX and sugges mul iple new a eas o u u e esea ch. Addi ional ile Addi ional ile 1: Table S1. P ime s used o assessing he exp ession le el o me hyla ion- ele an enzymes by qPCR. (DOC 36 kb) Abb e ia ions 5hmC: 5-hyd oxyme hylcy osine; 5mC: 5-me hylcy osine; ACPA: an i-ci ullina ed pep ide an ibody; ACR: Ame ican College o Rheuma ology; ANCOVA: analysis o a iance wi h co a ia es; CI: con idence in e al; CpG: cy osine-phospha e- guanine dinucleo ide; DAS28: disease ac i i y sco e in 28 join s; DMARD: disease-modi ying an i- heuma ic d ug; DNMT: DNA me hyl asn e ase; EULAR: The Eu opean League Agains Rheuma ism; FLS: ib oblas -like syno iocy e; GADD45A: g ow h a es and DNA-damage-inducible p o ein 45 alpha; HPLC-ESI-MS/MS-SRM: high-pe o mance liquid ch oma og aphy- elec osp ay ioniza ion- andem mass spec ome y-selec ed eac ion moni o ing; IQR: in e qua ile ange; MTX: me ho exa e; NK: na u al kille ; PBMCs: pe iphe al blood mononuclea cells; PMN: polymo phonuclea ; qPCR: quan i a i e polyme ase chain eac ion; RA: heuma oid a h i is; RF: heuma oid ac o ; SAM: S-adenosyl-L-me hionine; SLE: sys emic lupus e y hema osus; TET: en-ele en ansloca ion. Compe ing in e es s The au ho s decla e ha hey ha e no compe ing in e es s Au ho s’con ibu ions MCA designed he s udy, pe o med expe imen s, analyzed esul s and w o e he i s d a o he manusc ip . EP-P ollowed he pa ien s and ec ui ed pa ien s and con ols. MC con ibu ed o qPCR expe imen s and pe o med s a is ical analyses. FJS pe o med TET enzyme expe imen s. IO con ibu ed o DNA me hyla ion quan i ica ion. JJG-R con ibu ed o ec ui men and ob ained unding. AG designed he s udy, ob ained unding, analyzed esul s and supe ised he s udy. All au ho s con ibu ed o in e p e a ion o he esul s and o he inal e sion o he manusc ip ha all ha e app o ed a e e ising i c i ically o impo an in ellec ual con en . Acknowledgemen s We hank Ma ia del Sol Po o-Sil a o he help in ec ui ing he pa ien s and Ca men Pena-Pena o echnical assis ance. Funding The p esen wo k was suppo ed by Fondo de In es igacion Sani a ia o he Ins i u o de Salud Ca los III (Spain), g an s PI11/01048, PI12/01909 and RD12/0009/0008 ha a e pa ially inanced by he Eu opean Regional De elopmen Fund o he Eu opean Union. Au ho de ails 1 Labo a o io de In es igacion 10 and Rheuma ology Uni , Ins i u o de In es igación Sani a ia-Hospi al Clínico Uni e si a io de San iago, T a esia de Choupana, s/n, 15706 San iago de Compos ela, Spain. 2 Depa men o Medicine, Uni e si y o San iago de Compos ela, Rúa de San F ancisco, s/n, 15782 San iago de Compos ela, Spain. de And es e al. A h i is Resea ch & The apy (2015) 17:233 Page 8 o 9 Recei ed: 31 Janua y 2015 Accep ed: 10 Augus 2015 Re e ences 1. Bo ini N, Fi es ein GS. Epigene ics in heuma oid a h i is: a p ime o heuma ologis s. Cu Rheuma ol Rep. 2013;15:372. 2. Heyn H, Es elle M. DNA me hyla ion p o iling in he clinic: applica ions and challenges. Na Re Gene . 2012;13:679–92. 3. Liu Y, A yee MJ, Padyuko L, Fallin MD, Hesselbe g E, Runa sson A, e al. Epigenome-wide associa ion da a implica e DNA me hyla ion as an in e media y o gene ic isk in heuma oid a h i is. Na Bio echnol. 2013;31:142–7. 4. Jones PA. Func ions o DNA me hyla ion: islands, s a si es, gene bodies and beyond. Na Re Gene . 2012;13:484–92. 5. Kohli RM, Zhang Y. TET enzymes, TDG and he dynamics o DNA deme hyla ion. Na u e. 2013;502:472–9. 6. Ba e o G, Scha e A, Ma hold J, S ach D, Swamina han SK, Handa V, e al. Gadd45a p omo es epigene ic gene ac i a ion by epai -media ed DNA deme hyla ion. Na u e. 2007;445:671–5. 7. Li Y, Zhao M, Yin H, Gao F, Wu X, Luo Y, e al. O e exp ession o he g ow h a es and DNA damage-induced 45alpha gene con ibu es o au oimmuni y by p omo ing DNA deme hyla ion in lupus T cells. A h i is Rheum. 2010;62:1438–47. 8. Zhang Y, Zhao M, Sawalha AH, Richa dson B, Lu Q. Impai ed DNA me hyla ion and i s mechanisms in CD4(+)T cells o sys emic lupus e y hema osus. J Au oimmun. 2013;41:92–9. 9. Meda F, Folci M, Bacca elli A, Selmi C. The epigene ics o au oimmuni y. Cell Mol Immunol. 2011;8:226–36. 10. Ja ie e BM, He nando H, Balles a E. En i onmen al igge s and epigene ic de egula ion in au oimmune disease. Disco Med. 2011;12:535–45. 11. Ka ouzakis E, Gay RE, Michel BA, Gay S, Neidha M. DNA hypome hyla ion in heuma oid a h i is syno ial ib oblas s. A h i is Rheum. 2009;60:3613–22. 12. Ka ouzakis E, Rengel Y, Jungel A, Kolling C, Gay RE, Michel BA, e al. DNA me hyla ion egula es he exp ession o CXCL12 in heuma oid a h i is syno ial ib oblas s. Genes Immun. 2011;12:643–52. 13. Nakano K, Boyle DL, Fi es ein GS. Regula ion o DNA me hyla ion in heuma oid a h i is syno iocy es. J Immunol. 2013;190:1297–303. 14. Nakano K, Whi ake JW, Boyle DL, Wang W, Fi es ein GS. DNA me hylome signa u e in heuma oid a h i is. Ann Rheum Dis. 2013;72:110–7. 15. Ja ie e BM, Fe nandez AF, Rich e J, Al-Shah ou F, Ma in-Sube o JI, Rod iguez-Ub e a J, e al. Changes in he pa e n o DNA me hyla ion associa e wi h win disco dance in sys emic lupus e y hema osus. Genome Res. 2010;20:170–9. 16. Liu CC, Fang TJ, Ou TT, Wu CC, Li RN, Lin YC, e al. Global DNA me hyla ion, DNMT1, and MBD2 in pa ien s wi h heuma oid a h i is. Immunol Le . 2011;135:96–9. 17. Richa dson B, Scheinba L, S ahle J, G oss L, Hanash S, Johnson M. E idence o impai ed T cell DNA me hyla ion in sys emic lupus e y hema osus and heuma oid a h i is. A h i is Rheum. 1990;33:1665–73. 18. Kim YI, Logan JW, Mason JB, Roubeno R. DNA hypome hyla ion in in lamma o y a h i is: e e sal wi h me ho exa e. J Lab Clin Med. 1996;128:165–72. 19. Neshe G, Moo e TL. The in i o e ec s o me ho exa e on pe iphe al blood mononuclea cells. Modula ion by me hyl dono s and spe midine. A h i is Rheum. 1990;33:954–9. 20. Wang YC, Chiang EP. Low-dose me ho exa e inhibi s me hionine S-adenosyl ans e ase in i o and in i o. Mol Med. 2012;18:423–32. 21. Le T, Kim KP, Fan G, Faull KF. A sensi i e mass spec ome y me hod o simul aneous quan i ica ion o DNA me hyla ion and hyd oxyme hyla ion le els in biological samples. Anal Biochem. 2011;412:203–9. 22. Rocha MS, Cas o R, Ri e a I, Kok RM, Smulde s YM, Jakobs C, e al. Global DNA me hyla ion: compa ison o enzyma ic- and non-enzyma ic-based me hods. Clin Chem Lab Med. 2010;48:1793–8. 23. Ale aha D, Neogi T, Silman AJ, Funo i s J, Felson DT, Bingham 3 d CO, e al. 2010 Rheuma oid a h i is classi ica ion c i e ia: an Ame ican College o Rheuma ology/Eu opean League Agains Rheuma ism collabo a i e ini ia i e. A h i is Rheum. 2010;62:2569–81. 24. English D, Ande sen BR. Single-s ep sepa a ion o ed blood cells. G anulocy es and mononuclea leukocy es on discon inuous densi y g adien s o Ficoll-Hypaque. J Immunol Me hods. 1974;5:249–52. 25. Pombo-Sua ez M, Calaza M, Gomez-Reino JJ, Gonzalez A. Re e ence genes o no maliza ion o gene exp ession s udies in human os eoa h i ic a icula ca ilage. BMC Mol Biol. 2008;9:17. 26. Hellemans J, Mo ie G, De Paepe A, Speleman F, Vandesompele J. qBase ela i e quan i ica ion amewo k and so wa e o managemen and au oma ed analysis o eal- ime quan i a i e PCR da a. Genome Biol. 2007;8:R19. 27. an Ges el AM, Haagsma CJ, an Riel PL. Valida ion o heuma oid a h i is imp o emen c i e ia ha include simpli ied join coun s. A h i is Rheum. 1998;41:1845–50. 28. Globisch D, Munzel M, Mulle M, Michalakis S, Wagne M, Koch S, e al. Tissue dis ibu ion o 5-hyd oxyme hylcy osine and sea ch o ac i e deme hyla ion in e media es. PLoS One. 2010;5:e15367. 29. Abshe DM, Li X, Wai e LL, Gibson A, Robe s K, Edbe g J, e al. Genome-wide DNA me hyla ion analysis o sys emic lupus e y hema osus e eals pe sis en hypome hyla ion o in e e on genes and composi ional changes o CD4+ T-cell popula ions. PLoS Gene . 2013;9:e1003678. 30. Liao J, Liang G, Xie S, Zhao H, Zuo X, Li F, e al. CD40L deme hyla ion in CD4(+) T cells om women wi h heuma oid a h i is. Clin Immunol. 2012;145:13–8. 31. Nile CJ, Read RC, Akil M, Du GW, Wilson AG. Me hyla ion s a us o a single CpG si e in he IL6 p omo e is ela ed o IL6 messenge RNA le els and heuma oid a h i is. A h i is Rheum. 2008;58:2686–93. 32. Zhu X, Liang J, Li F, Yang Y, Xiang L, Xu J. Analysis o associa ions be ween he pa e ns o global DNA hypome hyla ion and exp ession o DNA me hyl ans e ase in pa ien s wi h sys emic lupus e y hema osus. In J De ma ol. 2011;50:697–704. 33. Balada E, O di-Ros J, Se ano-Acedo S, Ma inez-Los ao L, Rosa-Ley a M, Vila dell-Ta es M. T ansc ip le els o DNA me hyl ans e ases DNMT1, DNMT3A and DNMT3B in CD4+ T cells om pa ien s wi h sys emic lupus e y hema osus. Immunology. 2008;124:339–47. 34. Liu Y, Chen Y, Richa dson B. Dec eased DNA me hyl ans e ase le els con ibu e o abno mal gene exp ession in “senescen ”CD4(+)CD28(−) T cells. Clin Immunol. 2009;132:257–65. 35. Yang W, Tang H, Zhang Y, Tang X, Zhang J, Sun L, e al. Me a-analysis ollowed by eplica ion iden i ies loci in o nea CDKN1B, TET3, CD80, DRAM1, and ARID5B as associa ed wi h sys emic lupus e y hema osus in Asians. Am J Hum Gene . 2013;92:41–51. 36. Ne es-Cos a A, Moi a LF. TET1 is a nega i e ansc ip ional egula o o IL-1be a in he THP-1 cell line. Mol Immunol. 2013;54:264–70. 37. Sal ado JM, Mi els ad PR, Belo a GI, Fo nace J AJ, Ashwell JD. The au oimmune supp esso Gadd45alpha inhibi s he T cell al e na i e p38 ac i a ion pa hway. Na Immunol. 2005;6:396–402. 38. Ne adil J, Pa laso a G, Veselska R. New mechanisms o an old d ug; DHFR- and non-DHFR-media ed e ec s o me ho exa e in cance cells. Klin Onkol. 2012;25:2S87–92. 39. Ka ouzakis E, Gay RE, Gay S, Neidha M. Inc eased ecycling o polyamines is associa ed wi h global DNA hypome hyla ion in heuma oid a h i is syno ial ib oblas s. A h i is Rheum. 2012;64:1809–17. Submi you nex manusc ip o BioMed Cen al and ake ull ad an age o : • Con enien online submission • Tho ough pee e iew • No space cons ain s o colo figu e cha ges • Immedia e publica ion on accep ance • Inclusion in PubMed, CAS, Scopus and Google Schola • Resea ch which is eely a ailable o edis ibu ion Submi you manusc ip a www.biomedcen al.com/submi de And es e al. A h i is Resea ch & The apy (2015) 17:233 Page 9 o 9