RESEARCH ARTICLE
Aging-associa ed pa e ns in he exp ession o
human endogenous e o i uses
Tapio Ne alainenID
1,2☯
*, A u Au io
1☯
, Binisha Hamal Mish a
1
, Saa a Ma ila
1,2¤
,
Ma ja Jylha
¨
2,3
, Mikko Hu me
1,2
1Facul y o Medicine and Li e Sciences, Uni e si y o Tampe e, Tampe e, Finland, 2Ge on ology Resea ch
Cen e (GEREC), Uni e si y o Tampe e, Tampe e, Finland, 3Facul y o Social Sciences, Uni e si y o
Tampe e, Tampe e, Finland
☯These au ho s con ibu ed equally o his wo k.
¤Cu en add ess: Ins i u e o Ageing and Ch onic Disease, Uni e si y o Li e pool, Li e pool, Uni ed
Kingdom
* apio.l.ne alainen@u a. i
Abs ac
Human endogenous e o i uses (HERV) a e elics o ancien e o i al in ec ions in ou
genome. Mos o hem ha e los hei coding capaci y, bu p o i al RNA o p o ein ha e
been obse ed in se e al disease s a es (e.g. in in lamma o y and au oimmune diseases
and malignancies). Howe e , hei clinical signi icance as well as hei mechanisms o ac ion
ha e s ill emained elusi e. As human aging is associa ed wi h se e al biological cha ac e -
is ics o hese diseases, we now analyzed he aging-associa ed exp ession o he indi idual
p o i uses o wo HERV amilies, HERV-K (91 p o i uses) and HERV-W (213 p o i uses)
using genome-wide RNA-sequencing (RNA-seq). RNA was pu i ied om blood cells de i ed
om heal hy young indi iduals (n = 7) and om nonagena ians (n = 7). The da a indica ed
ha in he case o HERV-K (HML-2) 33 p o i uses had a de ec able exp ession bu in only 3
o hose he exp ession le els we e signi ican ly di e en be ween he young and old indi id-
uals. In he HERV-W amily exp ession was obse ed in 45 loci and only in one case he
young/old di e ence was signi ican . Howe e , applying hie a chical clus e ing on he
HERV exp ession da a esul ed in he o ma ion o wo dis inc clus e s, one con aining he
young indi iduals and ano he he nonagena ians. This sugges s, ha e en hough he
aging-associa ed di e ences in he exp ession le els o indi idual p o i uses a e mino ,
he e seems o be some unde lying aging- ela ed pa e n. These da a indica e ha aging
does no ha e a s ong e ec on he exp ession o indi idual HERV p o i uses, bu ins ead
se e al p o i uses a e a ec ed mode a ely, leading o age-dependen exp ession p o iles.
In oduc ion
Du ing mammalian e olu ion, in eg a ion o e o i al RNA in o a ge m line cell may ha e
led in o a o ma ion o a p o i us ha is ansmi ed e ically and inhe i ed in a Mendelian
manne . In humans, hese endogenous e o i uses (HERV) comp ise ca. 8% o ou genome
[1]. While i is known ha some e oelemen s o he human genome a e s ill capable o
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0207407 Decembe 4, 2018 1 / 11
a1111111111
a1111111111
a1111111111
a1111111111
a1111111111
OPEN ACCESS
Ci a ion: Ne alainen T, Au io A, Mish a BH, Ma ila
S, Jylha¨M, Hu me M (2018) Aging-associa ed
pa e ns in he exp ession o human endogenous
e o i uses. PLoS ONE 13(12): e0207407. h ps://
doi.o g/10.1371/jou nal.pone.0207407
Edi o : Klaus Roeme , Uni e si a des Saa landes,
GERMANY
Recei ed: Augus 9, 2018
Accep ed: Oc obe 30, 2018
Published: Decembe 4, 2018
Copy igh : ©2018 Ne alainen e al. This is an open
access a icle dis ibu ed unde he e ms o he
C ea i e Commons A ibu ion License, which
pe mi s un es ic ed use, dis ibu ion, and
ep oduc ion in any medium, p o ided he o iginal
au ho and sou ce a e c edi ed.
Da a A ailabili y S a emen : Da a is a ailable a
GEO (h ps://www.ncbi.nlm.nih.go /geo/) wi h
accession numbe GSE122309.
Funding: MH ecei ed Compe i i e Resea ch Fund
o Pi kanmaa Hospi al Dis ic (h ps://www. ays. i/
en-US/Resea ch_and_ aining/Resea ch). MH
ecei ed g an om Tampe e Tube culosis
Founda ion (h p://www. ube kuloosisaa io. i). The
unde s had no ole in s udy design, da a collec ion
and analysis, decision o publish, o p epa a ion o
he manusc ip .
e o ansposi ion, DNA sequences o he HERVs ha e accumula ed mu a ions o he poin
whe e e o ansposi ion o o ma ion o i al pa icles is no aking place anymo e [2].
Despi e his mu a ion-d i en unc ional inac i a ion, he e a e hund eds o publica ions
demons a ing associa ions be ween HERV exp ession and a ious disease s a es (malig-
nancies, in ec ions, neu ological and au oimmune diseases), howe e , he causal ela ion-
ship has emained enigma ic [3–6].
Since he mechanism o ac ion canno be explained by de no o inse ional mu agenesis no
wi h he o ma ion o i al pa icles, i has been p oposed, ha po en ial pa hogenici y o he
HERVs could simply unde lie in he p esence o p o i al DNA, ac ing as a ansc ip ional eg-
ula o y sequence, modi ying he exp ession o neighbo ing and e en mo e dis an genes.
HERVs can do his o example by ac ing as ansc ip ion ac o binding si es. F om his
hypo hesis i na u ally ollows, ha po en ial e ec s o he HERVs would be es ic ed in some
genomic window a ound he p ima y p o i al inse ion si e. Howe e , he e is also e idence
suppo ing mo e global mode o ac ion as HERVs ha e been shown o ac i a e immune and
in lamma o y esponses o he body di ec ly. Fo example, hei RNA could be ecognized as a
pa hogen-associa ed molecula pa e n (PAMP) by Toll-like ecep o s and his would induce
ype I in e e on p oduc ion con ibu ing o he pa hogenesis o au oin lamma o y diseases
[7]. Some HERVs a e s ill able o encode an in ac en elope p o ein (En ) and i s p esence has
been obse ed in some i al in ec ions o in au oimmune diseases [3–6]. I has been p oposed
ha he mechanism o ac ion o En is based on he an igenici y o he molecule, possibly caus-
ing a polyclonal ac i a ion o lymphocy es, i.e. unc ioning as a “supe an igen” [8].
As he diseases, whe e HERV-associa ions ha e been obse ed, demons a e some o he
undamen al and cha ac e is ic aspec s o aging, e.g. inc eased le el o in lamma ion and
changes in he p opo ions o he a ious lymphocy e subse s [9,10], we now quan i a ed he
RNA le els o all p e iously cha ac e ized p o i uses o HERV-K (HML-2) and HERV-W
amilies in pe iphe al blood mononuclea cells (PBMC) de i ed om young and 90-yea old
indi iduals. Aging-associa ed inc ease in he exp ession o se e al HERV amilies has been
epo ed p e iously using quan i a i e PCR [11]. Howe e , qPCR app oach u ilizes degene a e
p ime s o each HERV amily, hus missing he in o ma ion ega ding indi idual p o i uses.
RNA-sequencing possesses he capabili y o ob ain his c ucial da a and hence i was he
me hod o choice.
The mos ecen en an s o ou genome a e ep esen ed by HERV-K (HML-2) amily (ca.
0.2–2 million yea s ago), o which Sub amanian e al. ha e iden i ied 91 ull-leng h p o i al
sequences [12]. HERV-W ep esen s an olde g oup o HERVs (p ima y in ec ion ca. 40 mil-
lion yea s ago) and i con ains 213 ull-leng h o nea ull-leng h elemen s [13].
Me hods
S udy popula ions
Two popula ions, ep esen ing young and elde ly indi iduals, we e used. The young ones con-
sis ed o heal hy labo a o y pe sonnel, all emale, aged 26 o 32 yea s (n = 7, median age 28)
who did no ha e any medically diagnosed ch onic illnesses, we e non-smoke s and had no
had any in ec ions o ecei ed any accina ions wi hin he wo weeks p io o blood sample
collec ion. The elde ly indi iduals (n = 7) we e selec ed among ela i ely heal hy, communi y
li ing, non- ail, nonagena ian emales, wi hou any se e e aging-associa ed diseases, ha
we e pa icipan s in The Vi ali y 90+ s udy. The nonagena ians we e bo n in 1920 and he
samples we e collec ed in 2014. The ec ui men and cha ac e iza ion o pa icipan s we e pe -
o med as has been epo ed p e iously [14]. The s udy pa icipan s p o ided hei w i en
in o med consen . This s udy was conduc ed acco ding o he p inciples exp essed in he
Aging-associa ed pa e ns in he exp ession o human endogenous e o i uses
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0207407 Decembe 4, 2018 2 / 11
Compe ing in e es s: The au ho s ha e decla ed
ha no compe ing in e es s exis .
decla a ion o Helsinki, and he s udy p o ocol was app o ed by he e hics commi ee o he
ci y o Tampe e (1592/403/1996).
Sample collec ion
Blood samples we e collec ed by a ained labo a o y echnician in he labo a o y acili ies. All
blood samples we e d awn be ween 8 am and 12 am and collec ed in o EDTA con aining
ubes. Samples we e di ec ly subjec ed o leucocy e sepa a ion on a Ficoll-Paque densi y g adi-
en (Ficoll-Paque P emium, ca . no. 17-5442-03, GE Heal hca e Bio-Sciences AB, Uppsala,
Sweden). The PBMC laye was collec ed and cells used o RNA ex ac ion we e suspended in
150 μl o RNAla e solu ion (Ambion Inc., Aus in, TX, USA). Nonagena ian and con ol sam-
ples we e collec ed a he same ime.
RNA ex ac ion
RNA used o RNA sequencing was pu i ied using a miRNeasy mini ki (Qiagen, CA, USA)
and he RNA used o PCR analysis using RNeasy mini ki (Qiagen, CA, USA) acco ding o
manu ac u e ’s p o ocol wi h on-column DNA diges ion (Qiagen). The concen a ion and
quali y o he RNA was assessed wi h a NanoD op ND-1000 spec opho ome e (NanoD op
Technologies, Wilming on, DE, USA).
RNA sequencing
Agilen Bioanalyze RNA nano chips (Agilen ) we e used o e alua e he in eg i y o o al
RNA and Qubi RNA–ki (Li e Technologies) o quan i a e RNA in samples. 1 μg o o al
RNA was used o Sc ip Seq Comple e Gold Sys em (Epicen e) o ibodeple e RNA and
u he o RNA-seq lib a y p epa a ion. SPRI beads (Agencou AMPu e XP, Beckman
Coul e ) we e used o pu i ica ion o RNAseq lib a ies. The lib a y QC was e alua ed on
High Sensi i i y chips by Agilen Bioanalyze (Agilen ). Pai ed-end sequencing o RNAseq
lib a ies was done using Illumina HiSeq echnology wi h a minimum o 60 million 2x100bp
pai ed-end eads pe sample.
Da a p ep ocessing and analysis
Raw eads we e aligned o human genome e e ence build hg19 using TopHa 2.0.13 [15]
wi h he de aul pa ame e s. Only uniquely mapped eads we e conside ed in he ansc ip
abundance es ima ion and o his end SAM ools [16] was used o il e ou eads mapping o
mul iple egions o he genome. The downs eam analyses we e all conduc ed using he ools
in cu links2 . 2.2.1 [17,18]. The aw exp ession es ima es we e calcula ed using cu quan
and he exp ession we e no malized using cu no m, which gi es he no malized ead coun s
and he agmen s pe kilobase pe million alues (FPKM) o each gene as an ou pu . The geo-
me ic no maliza ion me hod was used which scales he ead coun s as well as he FPKM al-
ues acco ding o p ocedu e desc ibed in [19].
The anno a ion da a o HERV-K (HML-2) was om Sub amanian e al. [12] and ha o
HERV-W om G andi e al [13]. To ensu e he obus ness o he no maliza ion he exp es-
sions o HERV elemen s we e quan i ied and no malized oge he wi h ENSEMBL . 82 gene
e e ence se [20,21]. Fo each indi idual s udy subjec , a gi en HERV elemen was conside ed
signi ican ly exp essed when he indi idual exp ession le el exceeded no malized ead coun
o 16 [22].
Aging-associa ed pa e ns in he exp ession o human endogenous e o i uses
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0207407 Decembe 4, 2018 3 / 11
Clus e analysis
Hie a chical clus e ing o he samples based on no malized ead coun s was done sepa a ely
o bo h HERV-K (HML-2) and HERV-W. Spea man co ela ion was used as he dis ance
me ic, which is obus agains ou lie s and non-Gaussian dis ibu ions, and can cap u e non-
linea ela ionships [23,24]. Wa d’s minimum inc ease o sum-o -squa es was used as he
linkage me hod, which has been epo ed o pe o m be e wi h gene exp ession da a han he
mo e adi ional me hods o a e age and comple e linkage [23]. Mul is ep-mul iscale boo -
s ap esampling was done o e alua e he unce ain y in ol ed in he clus e ing [25]. Thou-
sands o samples o a ying sizes a e andomly c ea ed om he da a and hen clus e ed. An
app oxima ely unbiased (AU) p- alue is ob ained, which indica es he bias co ec ed pe cen -
age o dend og am a ian s whe e he speci ic clus e was obse ed.
Resul s
The esul s o he RNA-seq analysis indica ed ha 33 HERV-K (HML-2) loci ou o 91 had a
de ec able exp ession, bu o en a low le el and no in all indi iduals. The exp ession le els o
PBMCs de i ed om young and elde ly indi iduals we e gene ally simila . Only a h ee loci
(1q22, 10p14 and 12q24.33) he di e ence was s a is ically signi ican as shown in Table 1.
In he case o HERV-W, he esul s we e simila , in 45 p o i uses ou o 213 he ead coun
was >16 a leas in one indi idual, and in he case o Xp11.21 he di e ence in exp ession le -
els be ween he young and old was signi ican as shown in Table 2.
Hie a chical clus e ing o he samples based on no malized p o i us ead coun s was done o
in es iga e exp ession pa e ns. Clus e ing o samples based on HERV-K (HML-2) exp ession
esul ed in wo g oups sepa a ed along he age g oup lines (Fig 1A). The e we e wo de ia ions
om his, wi h one nonagena ian in he p edominan ly young sample clus e and one young
sample in he nonagena ian clus e . Hea maps o he clus e ing o he HERV-K (HML-2) and
HERV-W p o i us exp ession le els a e shown in Figs 2and 3, espec i ely.
Boo s ap esampling o he clus e ing was done o quan i y he ce ain y o he clus e ing.
Bo h clus e s ha e an app oxima ely unbiased (AU) p- alue o 97, which is he bias co ec ed
pe cen age o esampling dend og am a ian s whe e he speci ic clus e was obse ed. AU p-
alue o 97 is equi alen o a p- alue o 0.03, indica ing s a is ical signi icance. The same signi -
ican clus e s esul ed e en i he signi ican ly di e en ially exp essed 1q22, 10p14 and
12q24.33 we e excluded om clus e ing.
HERV-W exp ession based clus e ing o samples esul ed in one s a is ically signi ican
clus e (p- alue o 0.04), which con ains he same six nonagena ian samples ha a e g ouped
oge he in he HERV-K (HML-2) based clus e ing (Fig 1B).
Conclusions
The RNA le els o indi idual p o i uses a ied conside ably be ween samples. I was no he
case ha some indi iduals would ha e been mo e ac i e p oduce s han o he s, bu ins ead di -
e en p o i uses seemed o be exp essing non-sys ema ically wi hin and be ween indi iduals.
A o al o eigh HERV-K p o i uses and nine HERV-W p o i uses we e ound o be exp essed
in all 14 samples and consequen ly hese p o i uses we e exp essed wi h highes RNA le els.
This sugges s ha some indi idual p o i uses could be less es ic ed in e ms o hei exp es-
sion po en ial, ha is b ough by he egula ion machine y o he cell. Se e al p o i uses we e
exp essed only in small pa o indi iduals and i is emp ing o hink ha hese could be he
ones behind po en ial ad e se e ec s, especially i hey would be mainly exp essed in nonage-
na ians, as hey a e p obably silenced o a eason. The e we e no p o i uses ha we e
exp essed exclusi ely in nonagena ians, bu o example HERV-K 8p23.1a was exp essed in 6
Aging-associa ed pa e ns in he exp ession o human endogenous e o i uses
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0207407 Decembe 4, 2018 4 / 11
nonagena ians and only in 1 young indi idual. Fu he mo e, only 4 aging-associa ed di e en-
ially exp essed p o i uses we e iden i ied (in HERV-K (HML-2) 1q22 and 10p14 ha ing a
highe and 12q24.33 a lowe exp ession in he elde ly and in HERV-W Xp11.21 a lowe
exp ession). Pu ing all his oge he , i seems o be he case, ha aging has only mode a e
e ec on he exp ession le els o indi idual p o i uses.
Howe e , he hie a chical clus e ing o he exp ession da a indica ed ha he exp ession
p o iles o he young and elde ly subpopula ions we e di e en . The simples way o achie e
his kind o di e ence would be i , o example, all he p o i uses we e exp essed sys ema ically
Table 1. Median exp ession le els (no malized ead coun s) o HERV-K (HML-2) p o i uses. P o i uses we e deemed exp essed i exhibi ing a ead coun o 16 o
mo e [22]. Known aliases a e de i ed om Sub amanian e al. [12].
HERV-K locus Aliases Median exp ession le el
(no malized ead coun ) in
nonagena ians/young
con ols
Numbe o nonagena ians/young con ols exp essing he p o i us
1p31.1a K4, K116, ERVK-1 6.99 6.30 2 / 1
1q21.3 - 7.76 18.72 1 / 5
1q22 K102, K(C1b),K50a,ERVK-7 339.70 261.01�7 / 7
1q23.3 K110, K18,K(C1a), ERVK-18 95.31 78.79 7 / 7
1q32.2 - 39.31 42.30 7 / 7
3q12.3 K(II), ERVK-5 774.61 916.86 7 / 7
3q13.2 K106, K(C3),K68, ERVK-3 19.77 11.47 4 / 3
3q21.2 K(I), ERVK-4 10.81 19.72 2 / 7
4p16.1a K17b 24.46 26.10 6 / 6
4p16.1b - 15.61 9.44 3 / 1
4p16.3a - 15.45 16.20 3 / 4
7q34 K(OLDAC004979),ERVK-15 66.42 74.16 7 / 7
8p23.1a K115, ERVK-8 28.81 11.39 7 / 1
8p23.1b K27 14.32 17.09 3 / 4
8p23.1c - 13.13 22.86 3 / 6
9q34.11 K31 40.53 36.90 6 / 7
9q34.3 K30 1.01 3.81 0 / 1
10p14 K(C11a), K33,ERVK-16 70.77 18.00�7 / 4
10q24.2 ERVK-17, c10_B 7.16 8.95 0 / 1
11p15.4 K7 7.29 12.35 0 / 3
11q12.1 - 8.19 14.65 3 / 3
11q12.3 K(OLDAC004127) 13.53 8.93 2 / 3
12p11.1 K50e 0.00 0.00 1 / 0
12q24.11 - 12.49 4.88 3 / 2
12q24.33 - 87.55 97.25�7 / 7
14q11.2 - 54.12 27.67 7 / 7
16p13.3 - 2.62 4.41 0 / 1
19q11 K(C19), ERVK-19 2.31 4.91 0 / 1
19q13.12a - 7.76 18.10 1 / 4
19q13.12b K(OLDAC012309),KOLD12309 122.78 146.49 7 / 7
19q13.41 - 12.23 9.71 2 / 2
20q11.22 K(OLDAL136419),K59 13.17 8.57 2 / 2
22q11.21 K101, K(C22),ERVK-24 5.12 4.76 0 / 1
�
S a is ically signi ican (Mann-Whi ney U- es ) di e en ial exp ession and exp essed in majo i y o samples.
h ps://doi.o g/10.1371/jou nal.pone.0207407. 001
Aging-associa ed pa e ns in he exp ession o human endogenous e o i uses
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0207407 Decembe 4, 2018 5 / 11

Table 2. Median exp ession le els (no malized ead coun s) o HERV-W p o i uses. P o i uses we e deemed
exp essed i exhibi ing a ead coun o 16 o mo e [22].
HERV-W locus Median exp ession
le el (no malized
ead coun ) in
nonagena ians/
young con ols
Numbe o nonagena ians/young con ols exp essing he p o i us
1p12 5.22 10.59 0 / 1
1p22.2a 19.16 26.36 6 / 7
1p34.2 57.92 48.72 7 / 7
1q22 18.79 11.82 6 / 3
1q32.1 8.35 11.82 1 / 3
1q42.13 36.67 33.35 7 / 7
2p16.2 86.88 89.99 7 / 7
2p23.1a 11.07 16.17 2 / 4
2q11.2 33.96 27.27 7 / 6
2q22.2 108.33 113.64 7 / 7
2q24.3 3.71 14.38 0 / 2
2q31.2a 52.18 31.76 6 / 5
2q32.3 7.24 13.64 1 / 3
3q11.2 5.31 6.67 1 / 1
3q13.31 169.19 182.88 7 / 7
3q13.32 53.11 58.60 7 / 7
3q23b 86.98 73.15 7 / 7
3q26.32 8.35 12.71 2 / 2
4p16.3 13.66 11.72 3 / 3
4q21.22 9.31 15.45 1 / 3
5q22.2 3.03 4.49 0 / 1
6p22.3 25.05 30.80 4 / 5
6q21a 132.84 111.17 7 / 7
6q21c 8.92 38.13 3 / 7
6q24.2a 8.71 10.00 0 / 1
6q27b 35.17 36.58 7 / 6
7p14.2 4.16 0.48 2 / 0
7q21.2 18.18 17.77 4 / 5
7q31.1a 0.00 1.93 0 / 1
8q21.11 10.34 9.18 0 / 1
9p13.3 17.74 16.31 4 / 4
10q24.1 26.89 25.16 7 / 5
11q14.1 34.83 36.36 7 / 7
11q14.2 15.86 4.69 3 / 0
12q24.31 37.57 37.27 6 / 7
13q13.3 16.55 13.64 4 / 1
14q21.2 29.71 31.63 6 / 7
14q32.11 9.39 7.06 1 / 0
15q21.3 14.61 8.89 2 / 2
17q12a 11.32 10.00 1 / 2
17q12b 8.35 9.09 1 / 2
17q22 29.22 25.42 6 / 6
18p11.31 9.90 9.63 2 / 0
(Con inued)
Aging-associa ed pa e ns in he exp ession o human endogenous e o i uses
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0207407 Decembe 4, 2018 6 / 11
sligh ly up o down in one o he g oups. This kind o beha io could be a ibu ed o some
kind o common egula o ha has only one simple mode o ac ion. Howe e , his was no he
case, as di e en p o i uses we e up- and down egula ed equally in he nonagena ians. This
equi es mo e complex egula ion and is possibly e lec ing mul ilaye ed epigene ic egula ion
machine y in ol ing, among o he hings, DNA me hyla ion and his one modi ica ions, and
inducing dis inguishable aging-associa ed exp ession p o ile. Due o Spea man co ela ion
based dis ance me ic in he clus e ing, each p o i us has iden ical weigh in he clus e ing
Table 2. (Con inued)
HERV-W locus Median exp ession
le el (no malized
ead coun ) in
nonagena ians/
young con ols
Numbe o nonagena ians/young con ols exp essing he p o i us
19q13.2a 9.90 17.05 0 / 5
Xp11.21 16.54 34.60�4 / 7
�S a is ically signi ican (Mann-Whi ney U- es ) di e en ial exp ession and exp essed in majo i y o samples.
h ps://doi.o g/10.1371/jou nal.pone.0207407. 002
Fig 1. Hie a chical clus e ing o HERV-K(HML-2) and HERV-W p o i uses. Hie a chical clus e ing o he samples was ca ied ou wi h no malized (A) HERV-K
(HML-2) and (B) HERV-W ead coun s, using Spea man co ela ion dis ance me ic. Nonagena ian samples a e indica ed by an iden i ie s a ing wi h he numbe 2,
while con ol sample iden i ie s s a wi h 4. The heigh sepa a ing clus e s has been calcula ed wi h Wa d’s minimum inc ease o sum-o -squa es linkage me hod and
indica es p opo ional dissimila i y be ween clus e s. The ed squa es indica e clus e s ha we e deemed s a is ically signi ican h ough boo s ap esampling. AU p- alue,
in ed on , indica es he bias co ec ed pe cen age o dend og am a ian s whe e he clus e was p esen .
h ps://doi.o g/10.1371/jou nal.pone.0207407.g001
Aging-associa ed pa e ns in he exp ession o human endogenous e o i uses
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0207407 Decembe 4, 2018 7 / 11
esul , ega dless o le el o exp ession. The e o e his esul would indica e ha he e a e di -
e ences be ween he age g oups ha a e e ealed when he p o i al exp ession p o iles a e
examined as a whole. The unde lying cause behind obse ed exp ession p o ile di e ence hus
has o a ec he exp ession o many di e en p o i uses. Unde s anding wha causes his di -
e ence could inc ease knowledge o HERV exp ession associa ed disease s a es and o age-
ela ed decline. Since he same nonagena ian samples a e clus e ed by bo h HERV-K (HML-2)
and HERV-W exp ession, his phenomenon may no be limi ed o hese amilies, and could be
p esen in o he HERV amilies as well. I is no ewo hy, ha ou analysis is only limi ed o
HERV-K (HML-2) and HERV-W amilies. P e ious s udies ha e indica ed ha up egula ion
o some o he HERV subclasses migh also ha e implica ions in umo immuni y [26–28].
The e o e, i is possible ha hese HERVs could con ibu e o aging mo e han HERV-K
(HML-2) and HERV-W. This emains o be explo ed in u u e s udies.
The e is a gene al ag eemen ha he exp ession o HERVs should be unde a s ic con ol,
i.e. allowing hei exp ession in he ge m line bu silencing in mos soma ic cells, whe e hei
ac i i y could dis up no mal gene exp ession o ansc ip p ocessing. Se e al o hese con ol
mechanisms ha e been cha ac e ized in de ail [29,30]. As human aging is associa ed wi h d a-
ma ic epigene ic changes, e.g. DNA me hyla ion [31], i is maybe su p ising ha exp ession
le els be ween he young and old indi iduals we e no s ikingly di e en . Howe e , i is possi-
ble ha his epigene ic egula ion is esponsible o he obse ed di e ences and he exp ession
Fig 2. Hea map o clus e ed samples and p o i uses based on HERV-K (HML-2) no malized ead coun s. The
colo g een indica es ela i ely high exp ession, while ed indica es ela i ely low exp ession. The g ey blocks benea h
sample dend og am indica e age g oup membe ship, wi h ligh g ey o young con ols and da k g ey o
nonagena ians.
h ps://doi.o g/10.1371/jou nal.pone.0207407.g002
Aging-associa ed pa e ns in he exp ession o human endogenous e o i uses
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0207407 Decembe 4, 2018 8 / 11
p o iles would be due o di e en ial sensi i i y o he indi idual p o i uses o hese aging-asso-
cia ed epigene ic changes.
The gene al exp ession p o ile o HERV-K (HML-2) in he es ing blood cells used he e,
was domina ed by a ew loci, i.e. 3q12.3, 19q13.12b and 1q22., esembling he si ua ion in in
i o p e-ac i a ed lymphocy es [32], sugges ing ha he p oli e a i e s a e o he cells has
p obably only a mino e ec . This a , no simila da a in he case o HERV-W is a ailable.
In conclusion, ansc ip ional egula ion o he p o i uses belonging o HERV-K (HML-2)
and HERV-W amilies appea s o be wo-dimensional in he PBMCs; a subse o HERVs a e
exp essed cons an ly in age-independen manne ha ing only sligh aging-associa ed di e ences
in he exp ession le els. These di e ences migh be explained by a ine- uning o ansc ip ional
egula ion ha is b ough by DNA me hyla ion and is known o be hea ily al e ed in aging. On
he o he hand, p o i uses in ano he subse o HERVs we e cha ac e ized by o al lack o
exp ession in some indi iduals. This could be he esul o some mo e d as ic mode o egula-
ion such as ha o H3K4me3, ha is also known o be al e ed in aging [33]. Aging-dependen
HERV p o ile ound wi h clus e ing migh e lec his aspec o egula ions and i is also possible
ha ad e se e ec s o HERVs a e d i en by hose p o i uses ha unde go mo e adical an-
sc ip ional elaxa ion o es ic ion, ha is no necessa ily seen in he median RNA le els ( o
example HERV-K 8p23.1a in Fig 2 and HERV-W 11q14.2 in Fig 3).
Fig 3. Hea map o clus e ed samples and p o i uses based on HERV-W no malized ead coun s. The colo g een
indica es ela i ely high exp ession, while ed indica es ela i ely low exp ession. The g ey blocks benea h sample
dend og am indica e age g oup membe ship, wi h ligh g ey o young con ols and da k g ey o nonagena ians.
HERV-W has mo e p o i uses lis ed, which causes changes o he appea ance o he hea map, in addi ion o he
di e ences in exp ession.
h ps://doi.o g/10.1371/jou nal.pone.0207407.g003
Aging-associa ed pa e ns in he exp ession o human endogenous e o i uses
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0207407 Decembe 4, 2018 9 / 11