Response to psychotherapy in borderline personality disorder and methylation status of the BDNF gene.

0.39 MB·English

by  Perroud, N

Citation:TranslPsychiatry(2013)3,e207;doi:10.1038/tp.2012.140
&2013MacmillanPublishersLimited Allrightsreserved 2158-3188/13
www.nature.com/tp
Response to psychotherapy in borderline personality
disorder and methylation status of the BDNF gene
NPerroud1,2,ASalzmann1,PPrada2,RNicastro2,M-EHoeppli3,SFurrer3,SArdu4,IKrejci4,FKarege3andAMalafosse1,3
Downregulation of brain-derived neurotrophic factor (BDNF) gene expression with corresponding increased methylation at
specific promoters has been associated with stressful experiences in early life and may explain later adulthood
psychopathology. We measured the percentage ofmethylation at BDNF CpG exonsI andIV as well as plasma BDNF protein
levelsin115subjectswithborderlinepersonalitydisorder(BPD)and52controls.BPDsubjectsthenunderwenta4-weekcourse
of intensive dialectical behavior therapy (I-DBT). BDNF methylation status and protein levels were re-assessed at the end of
treatment.BPDsubjectshadsignificantlyhighermethylationstatusinbothCpGregionsthancontrols.Inaddition,thehigherthe
numberofchildhoodtrauma,thehigherwasthemethylationstatus.InBPDsubjects,BDNFmethylationsignificantlyincreased
afterI-DBT.Nonrespondersaccountedforthemajorityofthisincrease,whereasrespondersshowedadecreaseinmethylation
status over time. Accordingly, the changes in methylation status over time were significantly associated with changes in
depression scores, hopelessness scores and impulsivity. No association was found between protein levels and BDNF
methylation status. We here found a relationship between child maltreatment and higher DNA methylation of BDNF. These
resultsmoreoversupporttheideathattheseepigeneticmarksmaybechangedthroughpsychotherapeuticapproachesandthat
thesechangesunderlinechangesincognitivefunctions.
TranslationalPsychiatry(2013)3,e207;doi:10.1038/tp.2012.140;publishedonline8January2013
Introduction histoneH3lysine27methylation(H3K27)atthecorrespond-
ing promoter. Although posttranslational modifications of
Borderlinepersonalitydisorder(BPD)isadisordercharacter-
histones have been shown to be an important correlate of
ized by marked impulsivity, lack of emotional control,
environmental epigenetic modifications, DNA methylation
disturbed interpersonal relationships and frequent self-injur-
status at CpG sites within the promoter/exon IV has also
iousandsuicidalbehaviors.Notsurprisingly,itisassociated
been extensively studied and found to be an important
with high socioeconomic burden and high morbidity and
correlate of the impact of early developmental stress.9–11
mortality. One of the etiological factors believed to be
Rothetal.9,11foundthatpsychosocialstressinratsincreased
responsible for the development of BPD is childhood
BDNFDNAmethylationatpromoter/exonIV,suggestingthat
maltreatment.1 It is hypothesized that early-life adverse
theremaybeamechanismunderlyingcognitivedeficitsfound
events may favor the occurrence of core dimensions of this
in adult subjects maltreated in their childhood. Although
personality disorder through sustained epigenetic modifica-
epigeneticmodificationsoftheBDNFgenehaveneverbeen
tions of key developmental or stress-related genes such as
exploredinBPD,increasedDNAmethylationatspecificCpG
theonecodingfortheglucocorticoidreceptor(NR3C1).2One
sitesintheBDNFpromoter/exonIVandlowerBDNFmRNA
of the proteins that have recently been implicated in BPDis
levelsintheWernikeareahavebeenfoundinsuicidevictims,
brain-derived neurotrophic factor (BDNF).3–5 Because of its
aphenotypecloselyrelatedtoBPD.12
involvement in neurodevelopment, BDNF may indeed be
Several recent studies in depression have convincingly
seenasanaturalcandidateforabiologicalcorrelateofearly-
shown that the epigenetic processes involving BDNF are
life stress. The altered levels of this protein found in BPD
subjects5 may thus be the consequence of epigenetic responsibleforthealteredlevelsofBDNFassociatedwiththis
disorderandthattheseepigeneticprocessescanbechanged
modificationresultingfromchildmaltreatment.Higherhistone
through antidepressant treatment.13,14 In this perspective,
and DNA methylation of the BDNF gene and consecutive
Lopezetal.15werethefirsttoshowthat,inhumans,8weeks
downregulation of BDNF mRNA and decreased levels of
BDNF have been demonstrated in animal models of mal- of treatment with an antidepressant (citalopram) decreased
treatment.6,7 Tsankova et al.8 showed that chronic defeat H3K27 methylation levels at promoter/exon IV of the BDNF
stress in mice induced downregulation of the expression of gene. This reduction in BDNF H3K27 methylation was not
BDNFpromoter/exonIVandwasassociatedwithincreasesin only associated with increased BDNF levels but also with
1DepartmentofPsychiatry,UniversityofGeneva,Cheˆne-Bourg,Switzerland;2DepartmentofMentalHealthandPsychiatry,UniversityHospitalsofGeneva,Geneva,
Switzerland;3DepartmentofMedicalGeneticandLaboratories,UniversityHospitalsofGeneva,Geneva,Switzerlandand4DepartmentofCariologyandEndodontology,
TreatmentPlanUnitandDivisionofOperativeDentistry,DentalSchool,UniversityofGeneva,Geneva,Switzerland
Correspondence:DrNPerroud,DepartmentofMentalHealthandPsychiatry,UniversityHospitalsofGeneva,8ruedu31-De´cembre,Geneva,1207Switzerland.
E-mail:[email protected]
Keywords:BDNF;borderlinepersonalitydisorder;psychotherapy;DBT;epigenetic
Received12September2012;revised19November2012;accepted24November2012
MethylationofBDNFandborderlinepersonalitydisorder
NPerroudetal
2
depression response. The same group had already ThestudywasapprovedbytheethicscommitteeofGeneva
demonstratedthatincreasedexpressionofBDNFpromoter/ University Hospital. Informed written consent was obtained
exonIVanddecreasedmethylationofH3K27intheprefrontal fromallparticipants.
cortex was associated with a history of antidepressant
treatment.16ThesestudiesthussuggestthatBDNFmethyla- Treatment. ParticipantswerereferredtoI-DBTfor4weeks
tion,atleastinhistones,isadynamicprocessunderlyingthe of treatment as previously described.17,25 I-DBT is an
cognitivechangesobservedduringtreatment.Theseresults adaptationofstandardDBTthataimstoprovideanoverview
clearly need to be extended to CpG DNA methylation as of the traditional DBT behavioral skills (emotion regulation,
this has been shown to be critical for regulation of interpersonal effectiveness, distress tolerance and mind-
BDNF expression and thus for the development of mental fulness).I-DBTconsistsofdailygroupandindividualtherapy
healthdiseases. the aim of which is to decrease the behavioral targets that
Followingtheseobservations,thisstudyaimstoanswerthe werechosenwithinaDBTframework,withsuicidalbehaviors
following questions: is the DNA methylation status of BDNF treatedasapriority,followedbybehaviorsthatinterferewith
exonsIandIVhigherinBPDcomparedwithcontrolsubjects? therapy and then by behaviors that interfere with quality of
IsthishigherDNABDNFmethylationinBPDsubjectslinked life.EachparticipantwasfollowedbyatrainedDBTtherapist
to child maltreatment? And finally can these epigenetic (nurseorpsychologist).25
processesbechangedthroughaspecificpsychotherapeutic
approach to subjects with BPD and are they linked to DNAextraction. FortheBPDsubjects,bloodsampleswere
responsetotreatment? systematically collected 1 week before the commencement
of the program and on the last day of I-DBT. For control
subjects, blood samples were collected just after the inter-
views. DNA was extracted fromperipheral blood leukocytes
Patientsandmethods
by using the Nucleon kit (Bioscience Amersham, GE
A total of 115 outpatients with BPD were admitted to a Healthcare, Glattbrugg, Switzerland). After extraction, 2mg
specialized center proving intensive dialectical behavior ofgenomicDNAwasbisulfite-modifiedbyusingtheEpiTect
therapy (I-DBT) as the main treatment for this disorder.17,18 Bisulfite Kit (Qiagen, Germantown, MD, USA) according to
Briefly,patientsdisplayingsuicidalorpara-suicidalbehaviors, themanufacturer’sinstructions.
severe impulse control disorders and anger problems were
referredtothe specializedcenterbytheirphysicianorother Commercial controls of methylation. Methylation status
medicalservices.Allparticipantsreceivedpsychopharmaco- wasidentifiedbyhigh-resolutionmeltassayonaRotor-Gene
logical treatment, which was refined by a psychiatrist if 6000 instrument (Corbett Life Science, Chadstone, Victoria,
necessary before and during follow-up. However, pharma- Australia). This technique proved to be accurate, rapid and
ceuticaltreatmentduringI-DBTremainedunmodifiedinmost sensitive.26,27DetailsontheDNAmethylationbyhigh-resolution
of the subjects (N¼89). Only subjects fulfilling DSM-IV meltanalysisaregivenintheSupplementaryMaterial.
(Diagnostic and Statistical Manual of Mental Disorders, 4th
Edition)criteriaforBPDwereacceptedforintensivetherapy. SelectionofBDNFregionsformethylationanalysisand
Inthisperspective,eachparticipantwasinterviewedbyeither PCR design. We identified by UCSC (University of Califor-
atrainedpsychiatristorapsychologistbeforeacceptanceinto nia, Santa Cruz) genome browser (http://genome.ucsc.edu/
treatment using the Screening Interview for Axis II Disorder index.html?org=Human&db=hg19&hgsid=242691861) a
(SCID-II) BPD part.19 In addition, the French version of the CpG island region located at chr11:27743473–27744564
DiagnosticInterviewforGeneticStudies(DIGS)wasusedto (CpG count: 81, % GC¼60.5 and Obs/Exp¼0.83) in the
assessAxisIdisorders.20At1weekbeforethestartofI-DBT, BDNFpromoterregion,upstreamoftheATGstartinexonI.
eachparticipantcompletedtheBeckDepressionInventoryII PCRdesignincludes9CpGsandislocatedfrom–1096to–
(BDI-II)21 to assess the current severity of depression, the 900(numbersrelatetothetranslationstartsiteconsideredas
BeckHopelessnessScale(BHS)22toestimatethedegreeof þ1inexonI).WecalledthisCpGislandBDNFCpGexonI.
pessimism and negativity about the future and the Barrat ThesecondCpGregion(chr11:27723060–27723294),which
Impulsiveness Scale (BIS-10) to measure the three compo- isnotaCpGislandbyUCSCgenomebrowsercriteria(%GC
nents of impulsivity: motor (behavior), attentional (cognitive) 450,length4200bpandObs/Exp40.6),waspositionedin
and non-planning.23 On the last day of I-DBT, participants exonIV,upstreamoftheATGstartinexonVII.Thepresent
were asked to complete these questionnaires again. The ampliconencompasses17CpGsandislocatedfrom–2341
ChildhoodTraumaQuestionnaire(CTQ),whichexaminesfive to –2107 (numbers relate to the translation start site
types of trauma (sexual abuse, physical abuse, physical considered as þ1 in exon VII). We called this CpG region
neglect,emotionalabuseandemotionalneglect),24wasused BDNFCpGexon IV(SupplementaryFigureS1).
toassesschildhoodtraumaticexperience.Withtheexception Allsamplesweretestedinduplicate.Intheseassays,the
oftheBIS-10,thesebaselinemeasurementswerepreviously methylationpercentageofsampleswasdeterminedbyhigh-
used in our previous study examining the effect of child resolution melt profile by comparing them with the scale of
maltreatmentonNR3C1methylationstatus.2 commercialcontrolsdescribedintheSupplementaryMaterial.
A total of 52 controls completed the same self-report
questionnaires and were recruited from the School of BDNF protein levels. BDNF protein levels were measured
DentistryattheUniversityofGeneva. inplasmaaspreviouslydescribed.28Bloodobtainedfromthe
TranslationalPsychiatry
MethylationofBDNFandborderlinepersonalitydisorder
NPerroudetal
3
antecubital vein was collected in a suitable tube containing p10%improvement;410top20%improvementand420%
an anticoagulant solution and centrifuged at 2000g for improvement.
10min at41C.Plasma wasaliquotedand storedat (cid:2)201C Finally, to ensure that the observed effect was not better
until use. BDNF levels were measured with an enzyme- explainedbychangesinmedicationduringfollow-up(26BPD
linked immunosorbent assay kit (Promega, Wallisellen, subjectshadachange/adjustmentoftreatmentduringfollow-
Switzerland) after appropriate dilution with the Block and up) or by antidepressant treatment (55 BPD subjects were
Sample solution (provided with the kit). A microplate reader taking antidepressants), these two variables (changes in
(AnthosLabtecInstrument,ChatelSt-Denis,Switzerland)set medication yes/no and taking antidepressant yes/no) were
at4501CwasusedtodetermineplasmaBDNFvalues(intra- addedintothemodelsascovariates.Theresultsofregression
assay and interassay variations were o6 and 7%, modelsarepresentedasstandardizedregressioncoefficients
respectively). (b)with 95%confidenceintervalsthatcanbeinterpretedas
effect size. The threshold of significance was P¼0.05. All
Statistical analyses. Linear regression with adjustment for analyses were performed using STATA release 10 (Stata-
age and gender was used to assess the effect of disease CorpLP,CollegeStation,TX,USA).
status (BPD vs controls) and childhood maltreatment status
(number of childhood maltreatments as previously
described).2 Results
MeanmethylationpercentageofbothBDNFCpGexonsI The clinical and demographic characteristics of control and
andIVwasusedasthedependantvariable.Inasecondary BPDsubjectsaredescribedinTable1.Afteradjustmentfor
analysis,thesetwoexonswereanalyzedseparately. ageandgender,BPDsubjectshadsignificantlyhighermethy-
Linear mixed models with maximum likelihood estimation lation status in both CpG regions compared with controls
wheretreatmenttimewasafixedeffectandtheindividuala (32.84%vs14.86%,b¼0.61,P¼0.002,95%CI0.22–0.99;
random effect, as described elsewhere,29,30 were initially and11.27%vs3.42%,b¼0.66,P¼0.001,95%CI0.24–0.99
used to analyze the effect of I-DBT on percentage of forCpGexonIVandCpGexonI,respectively;Figure1).
methylation.Secondarily,thesemodelswereusedtoassess These results may reflect the association between the
theassociationbetweenchangesinmethylationandchanges number of childhood maltreatments (sum of abuses and
indepressionseverity,hopelessnessandimpulsivityrespec- neglects) and the total percentage of methylation (mean
tively.Inaddition,todefinebetterthelinkbetweenchangesin percentage at both CpG sites). The greater the number of
methylation and treatment response, changes in BDNF childhood trauma, the higher was the methylation status
methylationduringI-DBTwereassessedwithinthefollowing (b¼0.12,P¼0.005,95%CI0.04–0.21;Figure2).
categories according to treatment response: for depression
and hopelessness: p25% improvement; 425 to p50% EffectofI-DBTonBDNFmethylationstatus. Therewasa
improvement; 450 to p75% improvement and 475% significantpositiveassociationbetweendepressionseverity,
improvement; and for impulsivity: 0% improvement; 40 to hopelessness and impulsivity and BDNF methylation status
Table1 DemographicandclinicalcharacteristicsoftheBPDandcontrolsamples
Controls(N¼52) BPD(N¼115) t(d.f.) P-value
Mean s.d. Mean s.d.
Age(years) 40.65 12.04 30.36 9.19 6.06(165) 8.69(cid:3)10–9
BDI(baseline) 6.04 5.38 34.10 11.80 (cid:2)14.94(148) 2.42(cid:3)10–31
Hopelessness(baseline) 4.03 2.5 11.30 4.82 (cid:2)9.39(149) 9.33(cid:3)10–17
BIS-10(baseline) 42.26 11.53 69.30 17.75 (cid:2)9.81(152) 6.45(cid:3)10–18
SCID-IIBPDpart — — 6.52 1.75
N % N % X2(d.f.) P-value
Female 24 46.15 108 93.91 38.03(1) 6.98(cid:3)10–10
AxisI
Bipolardisorder 0 0 22 19.13
Majordepressivedisorder 0 0 84 73.04
Schizoaffectivedisorder 0 0 9 7.83
CTQ
Physicalneglect 6 14.29 52 48.60 14.94(1) 1.11(cid:3)10–4
Physicalabuse 4 9.52 56 52.34 22.98(1) 1.63(cid:3)10–6
Emotionalabuse 12 28.57 91 85.05 45.07(1) 1.89(cid:3)10–11
Emotionalneglect 18 42.86 90 84.11 25.73(1) 3.91(cid:3)10–7
Sexualabuse 8 19.05 57 53.27 14.36(1) 1.51(cid:3)10–4
Abbreviations:BDI,BeckDepressionInventory;BIS-10,BarratImpulsivenessScale;BPD,borderlinepersonalitydisorder;CTQ,ChildhoodTraumaQuestionnaire;
SCID-II,ScreeningInterviewforAxisIIDisorder.
TranslationalPsychiatry
MethylationofBDNFandborderlinepersonalitydisorder
NPerroudetal
4
0.4 0.4 and changes in medication during follow-up (yes or no) as
covariates in the model did not modify the observed
associations.
xon IV 0.3 xon 1 0.3 weTreheanreaslyuzltesdwseerpeasraimteillya.rwhenBDNFCpGexons IVandI
E E
F F
N N
D D BDNF protein levels. At baseline, BPD subjects had
n B 0.2 ** n B 0.2 significantly higher BDNF protein levels than controls
o o
ati ati (Supplementary Figure S2; b¼1.13, P¼2.51(cid:3)10(cid:2)6, 95%
yl yl
h h CI0.69–1.58).
Met 0.1 Met 0.1 There was a significant decrease in BDNF protein levels
% % over time in BPD subjects (b¼ (cid:2)0.50, P¼0.008, 95% CI
**
(cid:2)0.87 to (cid:2)0.13), which was inversely associated with
0 0 treatment response (b¼(cid:2)0.02, P¼0.013, 95% CI (cid:2)0.04
Controls BPD Controls BPD to (cid:2)0.01fordepression;andb¼(cid:2)0.049,P¼0.020,95%CI
(cid:2)0.09to (cid:2)0.01forhopelessness).Poorresponders(p50%
Figure1 Percentagemethylationofbrain-derivedneurotrophicfactor(BDNF)
CpGexonIV(a) andBDNFCpG exonI(b) regionsincontrolsandborderline in BDI) showed a nonsignificant decrease in BDNF protein
personalitydisorder(BPD)subjectsatbaseline.**Po0.01.Barsrepresents.d. levels (b¼(cid:2)0.58, P¼0.062, 95% CI (cid:2)1.18 to 0.03),
whereas responders (450% in BDI) had a nonsignificant
increase in protein levels (b¼0.44, P¼0.78, 95% CI 0.05–
0.3 0.95). However, these changes were not significantly asso-
ciatedwithchangesinBDNFmethylationstatus(b¼ (cid:2)0.57,
on P¼0.387,95%CI (cid:2)1.88to0.73).Furthermore,noassocia-
ati tionwasfoundbetweenproteinlevelsandBDNFmethylation
yl
h 0.2 status in the whole sample (controls and BPD; b¼0.35,
et
m P¼0.634,95%CI (cid:2)1.10to1.79).
F
N
D
B Discussion
n 0.1
a
e ThepresentstudysupportstheideathatDNAmethylationisa
M
% dynamic process that underlines cognitive fluctuations in
humans. Recently, Lopez et al.15 showed in a convincing
0 study that changes in BDNF H3K27 methylation were
0 1 2 3 4 5 associated with antidepressant response. Although it might
Number of childhood maltreatment bethoughtthatthesechangesarerestrictedtoantidepressant
treatment and to histones, we found in this case that a
Figure2 Percentageofmethylation(meanofbrain-derivedneurotrophicfactor
psychotherapeutic approach can have the same impact on
(BDNF)CpGexonIVandBDNFCpGexonIregions)accordingtothenumberof
childhoodmaltreatments.Barsrepresents.d. DNACpGislands.Thus,ourresultsunderscorethefactthat
methylationchangesinCpGsareassociatedwithtreatment
(b¼0.02,P¼0.0001,95%CI0.01–0.03;b¼0.04;P¼0.002, responseindependentlyofthemethodusedtotreatpatients.
95% CI 0.016–0.07; and b¼0.02, P¼2.32(cid:3)10(cid:2)6, 95% CI Ourresultsshowedthatnonrespondersshowedanincrease
0.01–0.025,respectively;Figure3). in BDNF methylation during treatment, whereas responders
There was a significant decrease in depression severity showedadecreaseinBDNFmethylationstatus.Subjectswith
(from34.10(s.d.¼11.80)to20.37(s.d.¼12.42),b¼(cid:2)0.94, the best response achieved the level of BDNF methylation
P¼5.68(cid:3)10(cid:2)35, 95% CI (cid:2)1.09 to (cid:2)0.79), hopelessness foundincontrolsubjects.Itisworthnotinginthelightofour
(from 11.30 (s.d.¼4.82) to 7.98 (s.d.¼5.04), b¼(cid:2)0.63, findingsthatD’Addarioetal.31foundhigherDNAmethylation
P¼1.51(cid:3)10(cid:2)13,95%CI (cid:2)0.80to (cid:2)0.47)and impulsivity inBDNFexonIpromoterofbipolardisordersubjectsduring
(from69.30(s.d.¼17.75)to65.99(s.d.¼17.31),b¼(cid:2)0.16, pharmacological treatment with mood stabilizers plus anti-
P¼0.006, 95% CI (cid:2)0.28 to (cid:2)0.05) over time in BPD depressants compared with those exclusively on mood-
subjectsfollowingI-DBT. stabilizing agents, suggesting that antidepressants may
Surprisingly,BDNFmethylationsignificantlyincreasedover increaseDNAmethylation.Weproposethatthisoccurswhen
time(b¼0.50,P¼0.0001,95%CI0.25–0.74).Poorrespon- subjectsarenonresponderstotreatment.
ders were mainly responsible for the increase (Table 2 and ThesecondissueisthatBDNFmethylationatCpGexonsI
Figure3). and IV was significantly higher in BPD patients than in
Afteradjustmentforbaselineclinicalscores,thechangesin controls, an association that may be the consequence of
methylationstatusovertimeweresignificantlyassociatedwith adverse events occurring early in life. These results are
changes in depression scores (b¼1.05, P¼4.60(cid:3)10(cid:2)11, consistentwithpreviousfindingsshowingthatBPDsubjects
95%CI0.75–1.36),hopelessnessscores(b¼0.64,P¼0.001, exhibited higher methylation status in several candidate
95%CI0.27–1.01)andimpulsivity(b¼0.51,P¼0.0002,95% genes including NR3C1, HTR2A, MAOA, MAOB and
CI 0.24–0.77). Adding antidepressant treatment (yes or no) COMT.2,32 They are also concordant with findings showing
TranslationalPsychiatry
MethylationofBDNFandborderlinepersonalitydisorder
NPerroudetal
5
***
0.8
0.5
**
n n
o o
ati 0.6 ati0.4
yl yl
h h
et et
m m0.3
F 0.4 F
N N
D D
B B0.2
n n **
a a
e 0.2 e
M M0.1
% %
0 0
0 20 40 60 Controls A B C D
Depression severity % Depression response
n 0.8 0.5
o ***
ati
ethyl 0.6 ation0.4 *
m yl
NF 0.4 meth0.3
D
B F
an 0.2 DN0.2
e B
M n
% 0 Mea0.1
0 5 10 15 20 %
Hopelessness
0
Controls BPD T0
Controls A B C D
BPD T4 RES BPD T4 nonRES
% Hopelessness response
0.8 0.5
***
n
o n
hylati 0.6 ylatio0.4 *
DF met 0.4 F Meth0.3
B N
ean 0.2 n BD0.2
M a
% Me0.1
0 %
20 40 60 80 100 120
0
BIS−10 total score
Controls a. b. c. d.
Controls BPD T0 BPD T4
% BIS−10 response
Figure3 Percentageofmethylation(meanofbrain-derivedneurotrophicfactor(BDNF)CpGexonIVandBDNFCpGexonIregions)accordingtodepressionseverity(a),
hopelessness(b)andimpulsivity(c).Therightpanelsin(a–c)indicatetreatmentresponseandpercentageofmethylationaccordingtotreatmentresponse:fordepressionand
hopelessness: A, r25% improvement; B, 425% to r50% improvement; C, 450% to r75% improvement; D, 475% improvement; and for impulsivity: a, 0%
improvement;b,40%tor10%improvement;c,410%tor20%improvement;d,420%improvement(graybarsindicateborderlinepersonalitydisorder(BPD)at
baselineandblackbarsindicateBPDaftertreatment).Barsrepresents.d.*Po0.05;**Po0.01;***Po0.001.Theleftpanelsin(a–c)indicatescatterplotsofpercentageof
methylationandseverityofdepression,hopelessnessandimpulsivity.BPDT0,borderlinepersonalitydisordersubjectsatbaseline;BPDT4,borderlinepersonalitydisorder
subjectsaftertreatment;non-RES,nonresponders(r50%improvement);RES,responders(450%improvement).
highermethylationofBDNFCpGexonIVinsuicidevictims,a humansshowingincreasedDNAandhistonemethylationof
phenotypecloselyrelatedtoBPD.12Finally,thecurrentdata BDNFaswellasofseveralothercandidategenesinsubjects
corroboratetheincreasingnumberofstudiesinanimalsand who have endured early-life events such as childhood
TranslationalPsychiatry
MethylationofBDNFandborderlinepersonalitydisorder
NPerroudetal
6
Table2 Percentageofmethylation(meanofBDNFCpGexonIandBDNFCpGexonIVregions)beforeandafterI-DBTaccordingtotreatmentresponse
Mean(baseline) s.d. Mean(end) s.d. N b P-value 95%CI
BDI
r25% 32.77 13.67 30.63 10.64 29 1.43 4.07(cid:3)10–13 1.04 1.81
425tor50% 37.67 10.27 24.24 7.29 21 0.51 0.01 0.15 0.87
450tor75% 35.68 7.50 13.59 4.38 22 (cid:2)0.36 0.06 (cid:2)0.74 0.02
475% 32.27 8.99 4.73 3.17 15 (cid:2)0.67 0.01 (cid:2)1.18 (cid:2)0.16
Hopelessness
r25% 11.49 4.71 7.98 5.04 40 0.87 0.00 0.48 1.27
425tor50% 11.05 4.84 4.42 2.87 14 0.63 0.01 0.14 1.11
450tor75% 11.11 4.73 3.46 2.13 21 (cid:2)0.38 0.14 (cid:2)0.88 0.13
475% 10.70 4.86 1.57 0.89 9 (cid:2)0.36 0.20 (cid:2)0.92 0.19
BIS-10
0 64.06 17.52 71.26 16.37 36 0.85 0.0004 0.38 1.31
40tor10% 71.54 19.27 68.21 18.32 18 0.52 0.04 0.03 1.01
410tor20% 74.69 18.20 64.48 15.56 13 (cid:2)0.32 0.19 (cid:2)0.81 0.16
420% 70.07 14.01 50.94 11.36 16 (cid:2)0.36 0.14 (cid:2)0.83 0.12
Abbreviations:BDI,BeckDepressionInventory;BDNF,brain-derivedneurotrophicfactor;BIS-10,BarratImpulsivenessScale;CI,confidenceinterval;I-DBT,
intensivedialecticalbehaviortherapy.
maltreatment.8,9,11,33–35 Overall, these studies and ours CpG exon I, as it has been described as the brain-specific
stronglysupporttheideathatmaltreatmentisassociatedwith induciblepromoter,42andonCpGexonIV,asithaspreviously
thebroadepigeneticsignatureofgenescruciallyinvolvedin been shown to be sensitive to environmental factors.35 In
brainfunctionanddevelopment. addition,themethyl-CpG-bindingprotein2(MeCP2),whichis
WealsofoundasignificanthigherlevelofBDNFproteinin knowntorepressBDNFgenetranscription,bindsselectively
patients than in controls. This is a striking result as most tomethylatedDNAatthepromoterofexonIV,43andisthusof
studiesindepressiontodatehavefoundlowerBDNFlevelsin particular interest for epigenetic regulation. Moreover, the
patientsthanincontrols.36BPDsubjectsareusuallytreated latterpromoterregioncontainsaspecificbindingsiteforthe
for years with antidepressants and other psychotropic cyclic AMP-responsive element-binding protein (CREB),38
medications. This chronic treatment may have biased the which is of particular interest in suicidal behavior and
levelofBDNFmeasurements.Indeed,althoughitisgenerally consequently for BPD. Moreover, another gene called
accepted that depressed patients have decreased serum antiBDNF (or BDNFOS) overlaps with the BDNF gene and
BDNF levels, treatment with antidepressants is also widely is transcribed in reverse orientation.44 Recently, Pruunsild
known to promote an increase in these levels.37 To our et al.39 showed that BDNF and antiBDNF transcripts form
knowledge, only one study has examined BDNF proteins dsRNA duplexes in brain in vivo. These findings reveal that
levelsandfoundlowerlevelsofthisproteininBPDsubjects.5 antiBDNFmRNAsseemtohaveasignificantroleinregulating
These discrepant results definitely need to be explored in BDNF expression in humans. Its regulatory functions could
futurestudies.AnotherpointrelatingtoBDNFproteinlevelsis actdirectlytoinhibitBDNFtranslationintoprotein.39Conse-
thefactthat,contrarytotheresultsobtainedbyLopezetal.15 quently, the discrepancy between BDNF protein levels and
showingthatBDNFH3K27methylationcorrelatedwithBDNF BDNFmethylationlevels couldbebecauseof alowlevelof
proteinlevels,wewereunabletofindanycorrelationbetween antiBDNF mRNAs, which leads to an increase in BDNF
DNA methylation at CpG exons I and IV and BDNF protein expressionwithoutanycorrelationwithmethylationlevelsof
levels. The first obvious explanation is that it may be more BDNF. Finally, the origin of plasma or serum BDNF is still
difficult to correlate CpG DNA methylation at two specific unknown.IthasbeenshownthatbloodBDNFisessentially
exonsonlywithchangesinproteinlevelsinperipheralblood. storedinbloodplatelets,fromwhichitcanbereleasedintothe
Methylationathistonesmaybeamoresuitablecandidatefor plasma through activation or clotting processes.45 As DNA
this. BDNF shows a complex gene architecture described methylationwasassessedonlymphocyteDNAinthecurrent
extensivelyinseveralpreviouspapers38withdifferentsplice study, its effect on BDNF expression (probably a decrease)
variants of BDNF mRNA differing by the 5’ extremity, each failstoovercomethemassiveeffectofplateletrelease.
regulated by specific promoter regions.39 These different Thisstudyhasseverallimitationsthatmay,inpart,explain
BDNF mRNA splice variants underline the tissue- and time- the striking results showing no correlation between BDNF
specific regulation of BDNF expression.40 In addition to this protein levels and BDNF DNA methylation. One is that we
complex control at the promoter level, BDNF expression is used peripheral blood that may only be a proxy for what is
alsoregulatedattheposttranscriptionallevelduringcleavage happeninginthebrain.Secondarily,mostofoursubjectswere
of pro-BDNF into mature BDNF by means of several under psychotropic treatment, especially antidepressants.
mechanisms.41 This complexity may thus explain the pro- Our findings may not actually be attributable to psychother-
blems experienced in linking BDNF protein levels to CpG apy.Nevertheless,ourresultswereunchangedafteradjust-
methylation. We decided to focus on the promoter of BDNF mentforantidepressantstatus(takingornot),whichindicates
TranslationalPsychiatry
MethylationofBDNFandborderlinepersonalitydisorder
NPerroudetal
7
at least that the observed associationis independent of this 5. KoenigsbergHW,YuanP,DiazGA,GuerreriS,DorantesC,MaysonSetal.Platelet
medication.Wemayalsobedealingwithnaturalfluctuations proteinkinaseCandbrain-derivedneurotrophicfactorlevelsinborderlinepersonality
disorderpatients.PsychiatryRes2012;199:92–97.
inmethylationovertime.Acontrolsamplefollowedupforthe
6. ElzingaBM,MolendijkML,OudeVoshaarRC,BusBA,PrickaertsJ,SpinhovenPetal.
same period would be required to exclude this hypothesis Theimpactofchildhoodabuseandrecentstressonserumbrain-derivedneurotrophic
definitively.Thedurationoffollow-up(4weeksoftreatment) factorandthemoderatingroleofBDNFVal66Met.Psychopharmacology(Berl)2011;214:
wasrelativelysmallinlength.Alongerdurationwouldideally 319–328.
7. KozlovskyN,MatarMA,KaplanZ,KotlerM,ZoharJ,CohenH.Long-termdown-regulation
beneededinthefuturetoestablishafirmcorrelationbetween
ofBDNFmRNAinrathippocampalCA1subregioncorrelateswithPTSD-likebehavioural
responsetotreatmentandchangesinmethylationstatus.An stressresponse.IntJNeuropsychopharmacol2007;10:741–758.
additionallimitationpertainstothespecificityofourfindings. 8. Tsankova NM, Berton O, Renthal W, Kumar A, Neve RL, Nestler EJ. Sustained
hippocampalchromatinregulationinamousemodelofdepressionandantidepressant
In other words, are the changes in methylation status
action.NatNeurosci2006;9:519–525.
observedin ourstudy specificto BDNF?In order to answer 9. RothTL,ZoladzPR,SweattJD,DiamondDM.Epigeneticmodificationofhippocampal
thisquestion,wetooktheopportunitytoassessthemethyla- BdnfDNAinadultratsinananimalmodelofpost-traumaticstressdisorder.JPsychiatr
tionstatusoftheexon1 NR3C1promoterregionin61ofthe Res2011;45:919–926.
F 10. LubinFD,RothTL,SweattJD.EpigeneticregulationofBDNFgenetranscriptioninthe
BPDsubjectsparticipatinginourpreviousreportforwhichwe
consolidationoffearmemory.JNeurosci2008;28:10576–10586.
hadavailableDNAatbaselineandonthelastdayofI-DBT.2 11. RothTL,LubinFD,FunkAJ,SweattJD.Lastingepigeneticinfluenceofearly-lifeadversity
Analyzingthesedata,weobservedthattherewasabsolutely ontheBDNFgene.BiolPsychiatry2009;65:760–769.
12. KellerS,SarchiaponeM,ZarrilliF,VideticA,FerraroA,CarliVetal.IncreasedBDNF
nochangesintheNR3C1methylationstatusbetweenthetwo
promotermethylationintheWernickeareaofsuicidesubjects.ArchGenPsychiatry2010;
measures(meanmethylationstatusatbaselineandattheend 67:258–267.
ofI-DBT¼0.134(s.d.¼0.02)and0.134(s.d.¼0.02),respec- 13. SenS,DumanR,SanacoraG.Serumbrain-derivedneurotrophicfactor,depression,and
tively; P¼0.758). This result strongly suggests that our antidepressantmedications:meta-analysesandimplications.BiolPsychiatry2008;64:
527–532.
currentfindingsarespecifictoBDNF.Finally,otherenviron-
14. MolendijkML,BusBA,SpinhovenP,PenninxBW,KenisG,PrickaertsJetal.Serumlevels
mentalfactorsthatwerenotmeasuredinthisstudymayhave ofbrain-derivedneurotrophicfactorinmajordepressivedisorder:state-traitissues,clinical
modifiedDNAmethylationduringfollow-up. featuresandpharmacologicaltreatment.MolPsychiatry2011;16:1088–1095.
15. LopezJP,MamdaniF,LabonteB,BeaulieuMM,YangJP,BerlimMTetal.Epigenetic
Inconclusion,theseresultsprovidefurthersupportforthe
regulationofBDNFexpressionaccordingtoantidepressantresponse. MolPsychiatry
impact of environmental factors on BDNF epigenetic pro- advanceonlinepublication,1May2012;doi:10.1038/mp.2012.38.
cesses.First,theyprovidefurtherevidenceoftherelationship 16. ChenES,ErnstC,TureckiG.Theepigeneticeffectsofantidepressanttreatmentonhuman
between child maltreatment, higher DNA methylation of prefrontalcortexBDNFexpression.IntJNeuropsychopharmacol2011;14:427–429.
17. PerroudN,UherR,DiebenK,NicastroR,HugueletP.Predictorsofresponseanddrop-out
BDNF and the development of adulthood psychopathology.
duringintensivedialecticalbehaviortherapy.JPersDisord2010;24:634–650.
Therefore,itmaybepossibletouseBDNFmethylationstatus 18. PerroudN,NicastroR,JermannF,HugueletP.Mindfulnessskillsinborderlinepersonality
asabiologicalmarkerofthispathway.Second,theysupport disorderpatientsduringdialecticalbehaviortherapy:preliminaryresults.IntJPsychiatry
ClinPract2012;16:189–196.
the idea that these epigenetic marks are susceptible to
19. First M, Gibbon M, Spitzer R, Williams JBW, Smith Benjamin L. Structured Clinical
change through targeted therapeutic approaches and may InterviewforDSM-IVPersonalityDisorders(SCID-II).AmericanPsychiatricAssociation:
underlietheameliorationincognitivefunction.Thus,evalua- WashingtonDC,1994.
tion of BDNF methylation levels may have relevance in 20. PreisigM,FentonBT,MattheyML,BerneyA,FerreroF.Diagnosticinterviewforgenetic
studies (DIGS): inter-rater and test-retest reliability of the French version. Eur Arch
predicting response to treatment. This hypothesis clearly
PsychiatryClinNeurosci1999;249:174–179.
needstobeextendedinfuturestudiestodeterminewhether 21. BeckAT,SteerRA,BallR,RanieriW.ComparisonofBeckDepressionInventories-IAand
thisistrulythecase. -IIinpsychiatricoutpatients.JPersAssess1996;67:588–597.
22. Beck AT, Weissman A, Lester D, Trexler L. The measurement of pessimism: the
hopelessnessscale.JConsultClinPsychol1974;42:861–865.
23. BayleFJ,BourdelMC,CaciH,GorwoodP,ChignonJM,AdesJetal.[Factoranalysisof
FrenchtranslationoftheBarrattimpulsivityscale(BIS-10)].CanJPsychiatry2000;45:
Conflictofinterest 156–165.
24. BernsteinDP,FinkL.ChildhoodTraumaQuestionnaire.ARetrospectiveSelf-Report.The
Theauthorsdeclarenoconflictofinterest.
PsychologicalCorporation:SanAntonio,USA,1998.
25. McQuillanA,NicastroR,GuenotF,GirardM,LissnerC,FerreroF.Intensivedialectical
behaviortherapyforoutpatientswithborderlinepersonalitydisorderwhoareincrisis.
Acknowledgements. WethankDominiqueMouthonandMoniqueVessaz PsychiatrServ2005;56:193–197.
fordatamanagementandDNAextractionandfortechnicalsupport.Wealsothank 26. WojdaczTK,DobrovicA,AlgarEM.RapiddetectionofmethylationchangeatH19in
Ge´raldBouillault,Jean-JaquesKunckler,BrigitteBlanchon,VenusKaby,andKaren humanimprintingdisordersusingmethylation-sensitivehigh-resolutionmelting.HumMutat
2008;29:1255–1260.
Diebenforthedatacollection.ThisworkwassupportedbyAXAResearchFund
27. WojdaczTK,DobrovicA,HansenLL.Methylation-sensitivehigh-resolutionmelting.Nat
‘Vulnerabilitytoriskybehaviorsasanomaliesinneuro-developmentaltrajectories:
Protoc2008;3:1903–1908.
anindividualizedapproach’grantedtoAlainMalafosse. 28. KaregeF,BondolfiG,GervasoniN,SchwaldM,AubryJM,BertschyG.Lowbrain-derived
neurotrophicfactor(BDNF)levelsinserumofdepressedpatientsprobablyresultsfrom
loweredplateletBDNFreleaseunrelatedtoplateletreactivity.BiolPsychiatry2005;57:
1. LeichsenringF,LeibingE,KruseJ,NewAS,LewekeF.Borderlinepersonalitydisorder. 1068–1072.
Lancet2011;377:74–84. 29. UherR,MaierW,HauserJ,MarusicA,SchmaelC,MorsOetal.Differentialefficacyof
2. PerroudN,Paoloni-GiacobinoA,PradaP,Olie´E,SalzmannA,NicastroRetal.Increased escitalopramandnortriptylineondimensionalmeasuresofdepression.BrJPsychiatry
methylationofglucocorticoidreceptorgene(NR3C1)inadultswithahistoryofchildhood 2009;194:251–258.
maltreatment:alinkwiththeseverityandtypeoftrauma.TranslPsychiatry2011;1:e59. 30. Rabe-HeskethS,SkrondalA.MultilevelandLongitudinalModelingUsingStata.Stata
3. WagnerS,BaskayaO,DahmenN,LiebK,TadicA.Modulatoryroleofthebrain-derived Press:CollegeStation,TX,2005.
neurotrophic factor Val66Met polymorphism on the effects of serious life events on 31. D’AddarioC,Dell’OssoB,PalazzoMC,BenattiB,LiettiL,CattaneoEetal.SelectiveDNA
impulsive aggression inborderline personality disorder. GenesBrain Behav 2010; 9: methylationofBDNFpromoterinbipolardisorder:differencesamongpatientswithBDIand
97–102. BDII.Neuropsychopharmacology2012;37:1647–1655.
4. TadicA,ElsasserA,VictorA,vonCubeR,BaskayaO,WagnerSetal.Association 32. DammannG,TeschlerS,HaagT,AltmullerF,TuczekF,DammannRH.IncreasedDNA
analysisofserotoninreceptor1B(HTR1B)andbrain-derivedneurotrophicfactorgene methylation of neuropsychiatric genes occurs in borderline personality disorder.
polymorphismsinBorderlinepersonalitydisorder.JNeuralTransm2009;116:1185–1188. Epigenetics2011;6:1454–1462.
TranslationalPsychiatry
MethylationofBDNFandborderlinepersonalitydisorder
NPerroudetal
8
33. McGowanPO,SudermanM,SasakiA,HuangTC,HallettM,MeaneyMJetal.Broad 42. Timmusk T, Palm K, Metsis M, Reintam T, Paalme V, Saarma M et al. Multiple
epigeneticsignatureofmaternalcareinthebrainofadultrats.PLoSOne2011;6:e14739. promoters direct tissue-specific expression of the rat BDNF gene. Neuron 1993; 10:
34. McGowanPO,SzyfM.Theepigeneticsofsocialadversityinearlylife:implicationsfor 475–489.
mentalhealthoutcomes.NeurobiolDis2010;39:66–72. 43. ImHI,HollanderJA,BaliP,KennyPJ.MeCP2controlsBDNFexpressionandcocaine
35. FuchikamiM,YamamotoS,MorinobuS,TakeiS,YamawakiS.Epigeneticregulationof intake through homeostatic interactions with microRNA-212. Nat Neurosci 2010; 13:
BDNFgeneinresponsetostress.PsychiatryInvestig2010;7:251–256. 1120–1127.
36. KaregeF,PerretG,BondolfiG,SchwaldM,BertschyG,AubryJM.Decreasedserum 44. LiuQR,WaltherD,DrgonT,PolesskayaO,LesnickTG,StrainKJetal.Humanbrain
brain-derivedneurotrophicfactorlevelsinmajordepressedpatients.PsychiatryRes2002; derived neurotrophic factor (BDNF) genes, splicing patterns, and assessments of
109:143–148. associations with substance abuse and Parkinson’s Disease. Am J Med Genet B
37. ShimizuE,HashimotoK,OkamuraN,KoikeK,KomatsuN,KumakiriCetal.Alterationsof NeuropsychiatrGenet2005;134B:93–103.
serumlevelsofbrain-derivedneurotrophicfactor(BDNF)indepressedpatientswithor 45. FujimuraH,AltarCA,ChenR,NakamuraT,NakahashiT,KambayashiJetal.Brain-
withoutantidepressants.BiolPsychiatry2003;54:70–75. derived neurotrophic factor is stored in human platelets and released by agonist
38. BoulleF,vandenHoveDL,JakobSB,RuttenBP,HamonM,vanOsJetal.Epigenetic stimulation.ThrombHaemost2002;87:728–734.
regulationoftheBDNFgene:implicationsforpsychiatricdisorders.MolPsychiatry2012;
17:584–596.
Translational Psychiatry is an open-access journal
39. PruunsildP,KazantsevaA,AidT,PalmK,TimmuskT.DissectingthehumanBDNFlocus:
bidirectionaltranscription,complexsplicing,andmultiplepromoters.Genomics2007;90:397–406. published by Nature Publishing Group. This work is
40. AnJJ,GharamiK,LiaoGY,WooNH,LauAG,VanevskiFetal.Distinctroleoflong3’UTR licensedundertheCreativeCommonsAttribution-NonCommercial-No
BDNFmRNAinspinemorphologyandsynapticplasticityinhippocampalneurons.Cell
DerivativeWorks3.0UnportedLicense.Toviewacopyofthislicense,
2008;134:175–187.
41. BarkerPA.WhitherproBDNF?NatNeurosci2009;12:105–106. visithttp://creativecommons.org/licenses/by-nc-nd/3.0/
SupplementaryInformationaccompaniesthepaperontheTranslationalPsychiatrywebsite(http://www.nature.com/tp)
TranslationalPsychiatry