Epigenetic landscape in blood leukocytes following ketosis and weight loss induced by a very low calorie ketogenic diet (VLCKD) in patients with obesity

O iginal a icle
Epigene ic landscape in blood leukocy es ollowing ke osis and weigh
loss induced by a e y low calo ie ke ogenic die (VLCKD) in pa ien s
wi h obesi y
Ana B. C ujei as
a
,
h
,
*
,
1
, And ea G. Izquie do
a
,
h
,
1
, Da id P imo
b
, Fe min I. Milag o
c
,
h
,
Ignacio Sajoux
d
, Amalia J
acome
e
, Al edo Fe nandez-Quin ela
,
h
, Ma ía P. Po illo
,
h
,
J.Al edo Ma ínez
c
,
h
, Miguel A. Ma inez-Olmos
a
,
h
, Daniel de Luis
b
,
Felipe F. Casanue a
g
,
h
a
Epigenomics in Endoc inology and Nu i ion G oup, Epigenomics Uni , Ins i u o de In es igacion Sani a ia de San iago de Compos ela (IDIS), Complejo
Hospi ala io Uni e si a io de San iago de Compos ela (CHUS/SERGAS), Spain
b
Cen e o In es iga ion o Endoc inology and Nu i ion, Medicine School and Depa men o Endoc inology and In es iga ion, Hospi al Clinico
Uni e si a io, Uni e si y o Valladolid, Valladolid, Spain
c
Depa men o Nu i ion, Food Science and Physiology, Cen e o Nu i ion Resea ch, Uni e si y o Na a a (UNAV) and IdiSNA, Na a a Ins i u e o
Heal h Resea ch, 31009, Pamplona, Spain
d
Medical Depa men P onokal G oup, P onokalG oup, Ba celona, Spain
e
Depa men o Ma hema ics, MODES G oup, CITIC, Uni e sidade da Co u~
na, Facul y o Science, A Co u~
na, Spain
Nu i ion and Obesi y G oup, Depa men o Nu i ion and Food Science, Uni e si y o he Basque Coun y (UPV/EHU), Lucio Lasca ay Resea ch Ins i u e
and Heal h Resea ch Ins i u e BIOARABA, Vi o ia, Spain
g
Molecula and Cellula Endoc inology G oup. Ins i u o de In es igacion Sani a ia de San iago de Compos ela (IDIS), Complejo Hospi ala io Uni e si a io
de San iago de Compos ela (CHUS) and San iago de Compos ela Uni e si y (USC), Spain
h
CIBER Fisiopa ologia de La Obesidad y Nu icion (CIBERobn), Spain
a icle in o
A icle his o y:
Recei ed 7 Janua y 2021
Accep ed 13 May 2021
Keywo ds:
Adiposi y
Me hyla ion
Nu i ional in e en ion
Ci cula ing blood cells
Bioma ke s
summa y
Backg ound: The molecula mechanisms unde lying he po en ial heal h benefi s o a ke ogenic die a e
unknown and could be media ed by epigene ic mechanisms.
Objec i e: To iden i y he changes in he obesi y- ela ed me hylome ha a e media ed by he induced
weigh loss o a e dependen on ke osis in subjec s wi h obesi y unde wen a e y-low calo ie ke ogenic
die (VLCKD).
Me hods: Twen y-one pa ien s wi h obesi y (n ¼12 women, 47.9 ±1.02 y , 33.0 ±0.2 kg/m
2
) a e 6
mon hs on a VLCKD and 12 no mal weigh olun ee s (n ¼6 women, 50.3 ±6.2 y s, 22.7 ±1.5 kg/m
2
)
we e s udied. Da a om he Infinium Me hyla ionEPIC BeadChip me hylomes o blood leukocy es we e
ob ained a ime poin s o ke o ic phases (basal, maximum ke osis, and ou o ke osis) du ing VLCKD
(n ¼10) and a baseline in olun ee s (n ¼12). Resul s we e u he alida ed by py osequencing in
ep esen a i e coho o pa ien s on a VLCKD (n ¼18) and co ela ed wi h gene exp ession.
Resul s: A e weigh educ ion by VLCKD, di e ences we e ound a 988 CpG si es (786 unique genes).
The VLCKD al e ed me hyla ion le els in pa ien s wi h obesi y had high esemblance wi h hose om
no mal weigh olun ee s and was concomi an wi h a down egula ion o DNA me hyl ans e ases
(DNMT)1, 3a and 3b. Mos o he encoded genes we e in ol ed in me abolic p ocesses, p o ein me a-
bolism, and muscle, o gan, and skele al sys em de elopmen . No el genes ep esen ing he op sco ing
associa ed e en s we e iden ified, including ZNF331,FGFRL1 (VLCKD-induced weigh loss) and CBFA2T3,
C3o 38,JSRP1, and LRFN4 (VLCKD-induced ke osis). In e es ingly, ZNF331 and FGFRL1 we e alida ed in
an independen coho and in e sely co ela ed wi h gene exp ession.
Conclusions: The beneficial e ec s o VLCKD he apy on obesi y in ol e a me hylome mo e sugges i e o
no mal weigh ha could be mainly media ed by he VLCKD-induced ke osis a he han weigh loss.
©2021 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND
license (h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
*Co esponding au ho . Ins i u o de In es igaci
on Sani a ia, Complejo Hospi ala io de San iago (CHUS), C/ Choupana, S/N, 15706, San iago de Compos ela, Spain.
E-mail add esses: [email p o ec ed],[email p o ec ed] (A.B. C ujei as).
1
Bo h au ho s equally con ibu ed o his wo k.
Con en s lis s a ailable a ScienceDi ec
Clinical Nu i ion
jou nal homepage: h p://www.else ie .com/loca e/clnu
h ps://doi.o g/10.1016/j.clnu.2021.05.010
0261-5614/©2021 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
Clinical Nu i ion 40 (2021) 3959e3972
1. In oduc ion
Ke osis has gained in e es o e ecen yea s due o i s induced
benefi s ha i impa s on se e al heal h condi ions [1,2]. Ke osis is
associa ed wi h a delay in he onse o diseases and inc eased
longe i y [3]. Simila ly, ke osis is sugges ed o ha e an ex ensi e
ange o heal h benefi s, om inc eased physical endu ance in
a hle es [4,5] o delayed aging [6]. Also o imp o e condi ions such
as neu odegene a i e disease [7e9] cance [10e12], ca dio ascula
disease [13], and obesi y [14]. Some o hese s udies in ol ed high
a ke ogenic die s and e en hough he main cha ac e is ic o
ke ogenic die s is he ca bohyd a es es ic ion, he specific
composi ion in mac onu ien s and calo ies should be aken in o
conside a ion o he impac in clinical p ac ice [15].
A e y-low-calo ie-ke ogenic die (VLCKD) was demons a ed o
be an e ec i e s a egy in managing obesi y [16], including weigh
loss and main enance [17], inc eased p ese a ion o muscle mass
[18], and enhanced es ing me abolic a e [19]. Mo eo e , i is able
o imp o e me abolic pa ame e s in pa ien s wi h obesi y [20,21]
and ype 2 diabe es [22]. Addi ionally, i was demons a ed ha a
VLCKD is able o educe ood c a ing and imp o e psychobiological
pa ame e s o help imp o e quali y o li e in pa ien s wi h obesi y
[23]. Howe e , he molecula mechanisms unde lying hese bene-
fi s o ke ogenic die emain unknown.
The main molecula mechanism ha links he e ec o en i-
onmen al ac o s, such as nu i ion, wi h he egula ion o he
genes unc ion is epigene ics [24]. Indeed, die a y ac o s o die a y
pa e ns we e iden ified as modula o s o epigene ic mechanisms
[25e28].
Recen ly, i was sugges ed ha ke one bodies o ches a e gene
exp ession ia epigenomic mechanisms [29]. This molecula e ec
o nu i ional ke osis has been epo ed in neu ologic diso de s
such as epilepsy [30,31] and Kabuki synd ome and ela ed diso -
de s [32]. The e ec o ke one bodies on epigene ic egula ion was
also p oposed as a po en ial oppo uni y o an icance he apies
[33]. Mo eo e , ke ogenic die s, including die a y es ic ion, delay
aging h ough epigene ic e ec s [34].
On he o he hand, obesi y is associa ed wi h a specific
me hyla ion p ofile in se e al issues [35e38] and he obesi y-
ela ed me hylome is ela ed o complica ions such as insulin
esis ance [39] and cance [40,41]. Mo eo e , se e al me hyla ion
ma ks we e iden ified as po en ial bioma ke s o p edic ing he
success o weigh loss he apies du ing he ac i e phase o ea -
men [42e46] o du ing he pe iod o weigh loss main enance
[47,48]. The e o e, DNA me hyla ion was p oposed as a a ge o
p e en ing and managing obesi y [48e51].
In ligh o he abo e e idence, he aim o he cu en s udy was
o e alua e how a VLCKD migh a ec he obesi y me hylome.
Fu he mo e, his s udy aimed o iden i y he changes in he
obesi y- ela ed me hylome ha a e media ed by he induced
weigh loss o a e dependen on ke osis.
2. Ma e ials and me hods
2.1. Pa ien coho
The DNA and RNA o me hyla ion and gene exp ession assays
we e isola ed om blood samples o pa ien s om a 6-mon h
nu i ional in e en ion s udy pe o med a he Endoc inology
and Nu i ion Depa men o he Hospi al Clinico, Uni e si a io o
Valladolid; he pa ien s we e ecei ing ea men o obesi y. In
Abb e ia ions lis
ACACB Ace yl-CoA Ca boxylase Be a
AEBP1 AE binding p o ein 1
ANOVA analysis o a iance
BMI body mass index
C3o 38 ch omosome 3 open eading ame 38
CACNA1H Calcium Vol age-Ga ed Channel Subuni Alpha1 H
CAMKK1 Calcium/Calmodulin Dependen P o ein Kinase
Kinase 1
CBFA2T3 CBFA2/RUNX1 Pa ne T ansc ip ional Co-Rep esso 3
cDNA complemen a y DNA
CENPF Cen ome e P o ein F
CHAT choline O-ace yl ans e ase
CHUK Componen O Inhibi o O Nuclea Fac o Kappa B
Kinase Complex
COL1A1 collagen, ype I, alpha 1
COL5A1 collagen ype V alpha 1 chain
COL9A2 collagen, ype IX, alpha 2
CV coe ficien o a ia ion
DMCpGs di e en ially me hyla ed CpGs
DNMTs DNA me hyl ans e ases
DNA deoxy ibonucleic acid
EWAS epigenome-wide associa ion s udy
FDR alse disco e y a e
FF esh- ozen
FGFRL1 Fib oblas g ow h ac o ecep o (FGFR)-like
p o ein 1
GAPDH Glyce aldehyde-3-Phospha e Dehyd ogenase
GO gene on ology
HRAS HRas p o o-oncogene, GTPase
INSR insulin ecep o
JSRP1 Junc ional Sa coplasmic Re iculum P o ein 1
LAMA2 Laminin Subuni Alpha 2
LRFN4 Leucine Rich Repea And Fib onec in Type III Domain
Con aining 4
MKNK2 MAPK In e ac ing Se ine/Th eonine Kinase 2
PRKAG2 P o ein Kinase AMP-Ac i a ed Non-Ca aly ic Subuni
Gamma 2
PRKCZ P o ein Kinase C Ze a
qRT-PCR Real- ime quan i a i e polyme ase chain eac ion
RELA - el e iculoendo heliosis i al oncogene homolog A
RNA Ribonucleic acid
RPTOR Regula o y Associa ed P o ein O MTOR Complex 1
SD s anda d de ia ion
SEM s anda d e o o he mean
SGCB Be a-sa coglycan
SMTN Smoo helin
TRAF2 TNF Recep o Associa ed Fac o 2
TSC2 Tube ous Scle osis Complex 2
TSS 200 ansc ip ion s a si es 200
VLCKD e y-low-calo ie ke ogenic die
WC Wais ci cum e ence
ZFHX3 zinc finge Homeobox 3
ZNF331 zinc finge p o ein 331
b
-OHB be a-hyd oxy-bu y a e
A.B. C ujei as, A.G. Izquie do, D. P imo e al. Clinical Nu i ion 40 (2021) 3959e3972
3960
addi ion, samples om a g oup o heal hy olun ee s we e also
analyzed.
The inclusion c i e ia we e: age be ween 18 and 65 yea s, body
mass index (BMI) 30 kg/m
2
, s able body weigh o e he p e ious
3 mon hs, a desi e o lose weigh , and a his o y o ailed die a y
e o s. The main exclusion c i e ia we e hy oid al e a ion, diabe es
melli us, obesi y induced by o he endoc ine diso de s o d ugs,
and pa icipa ion in any ac i e weigh -loss p og am in he p e ious
3 mon hs. In addi ion, pa ien s wi h p e ious ba ia ic su ge y,
epo ed o suspec ed abuse o na co ics o alcohol, se e e
dep ession o any o he psychia ic disease, se e e hepa ic insu -
ficiency, any ype o enal insu ficiency o gou episodes, neph-
oli hiasis, neoplasia, p e ious ins ances o ca dio ascula o
ce eb o ascula disease, uncon olled hype ension, o hos a ic
hypo ension, and hyd oelec oly ic o elec oca diog aphic al e -
a ions we e excluded. Females who we e p egnan , b eas eeding,
o in ending o become p egnan and hose wi h child-bea ing
po en ial who we e no using adequa e con acep i e me hods
we e also excluded. Apa om obesi y and me abolic synd ome,
pa icipan s we e gene ally heal hy indi iduals. Unde hese
c i e ia, 21 pa ien s wi h obesi y and 12 olun ee s wi h no mal
weigh we e included in his s udy.
The s udy p o ocol was in acco dance wi h he Decla a ion o
Helsinki and was app o ed by he E hics Commi ee o Clinical
Resea ch o Hospi al Clinico Uni e si a io de Valladolid, Spain (C.I:
40/13, PNK-DHA2013-01). Pa icipan s p o ided w i en in o med
consen be o e any in e en ion ela ed o he s udy. Pa icipan s
ecei ed no mone a y incen i es.
2.2. Ve y-low-calo ie ke ogenic die p o ocol
Pa ien s included in his s udy de i ed om a andomized
clinical ial in es iga ing he e ec o docosahexaenoic acid (DHA)
supplemen a ion in a e y low-calo ie ke ogenic die . The clinical
ial consis ed in wo a ms: one a m whe e pa ien s ollow a VLCKD
and o he a ms whe e pa ien s ollowed a VLCKD þDHA [21].
Nu i ional in e en ion was based on a comme cial weigh -loss
p og am (PNK me hod ®), as desc ibed elsewhe e [18,21]. B iefly,
he in e en ion included an e alua ion by he specialis physician
conduc ing he s udy, an assessmen by an expe die ician, and
exe cise ecommenda ions. This me hod is based on high-
biological- alue p o ein p epa a ions ob ained om cow's milk,
soy, a ian eggs, g een peas, and ce eals. Each p o ein p epa a ion
con ained 15g p o ein, 4g ca bohyd a es, 3g a , and p o ided
90e100 kcal. The VLCKD þDHA a m was supplemen ed wi h
500 mg/day DHA [21].
The weigh -loss p og am has fi e s eps and adhe es o he mos
ecen guidelines o he EFSA (2015) on o al ca bohyd a e in ake
[52]. The fi s h ee s eps consis o a VLCKD (600e800 kcal/day)
ha is low in ca bohyd a es (<50g daily om ege ables) and lipids
(only 10g o oli e oil pe day). The amoun o high biological- alue
p o eins anged be ween 0.8 and 1.2g pe kg o ideal body weigh o
ensu e ha pa ien s we e mee ing hei minimum bodily e-
qui emen s and o p e en he loss o lean mass. In s ep 1, he
pa ien s a e high-biological- alue p o ein p epa a ions fi e imes a
day and ege ables wi h low glycemic indices. In s ep 2, one o he
p o ein se ings was subs i u ed wi h a na u al p o ein (e.g., mea
o fish) ei he a lunch o a dinne . In s ep 3, a second se ing o
low- a na u al p o ein was subs i u ed o he second se ing o
biological p o ein p epa a ion. Th oughou hese ke ogenic phases,
supplemen s o i amins and mine als, such as K, Na, Mg, Ca, and
omega-3 a y acids, we e p o ided in acco dance wi h in e na-
ional ecommenda ions [53]. These h ee s eps we e main ained
un il he pa ien los he a ge amoun o weigh , ideally 80%.
Because o his, he ke ogenic s eps a ied in ime depending on he
indi idual and he weigh -loss a ge . The o al ke osis s a e las ed
o a maximum o 60 days.
In ei he s ep 4 o 5, ke osis was ended by he physician in
cha ge o he pa ien based on he amoun o weigh los , and he
pa ien began a low-calo ie die (800e
1500 kcal/day). A his poin ,
he pa ien s unde wen a p og essi e inco po a ion o di e en
ood g oups and pa icipa ed in a p og am o alimen a y e-
educa ion o gua an ee long- e m main enance o he weigh
loss. The main enance die consis ed o an ea ing plan balanced o
ca bohyd a es, p o ein, and a . Depending on he indi idual, cal-
o ies consumed anged be ween 1500 and 2000 kcal/day, wi h he
goal o main aining he weigh loss and p omo ing a heal hy
li es yle.
Du ing his s udy, pa ien s ollowed he s eps o he me hod un il
hey eached he a ge weigh , o up o a maximum o 4 mon hs o
ollow-up, al hough pa ien s emained unde medical supe ision
o he ollowing mon hs. Pa ien s isi ed he esea ch eam e e y
15 ±2 days o e alua e adhe ence and po en ial side e ec s. Com-
ple e an h opome y, body composi ion, and biochemical assess-
men s we e pe o med a ou o he isi s, which we e de e mined
acco ding o he e olu ion o each pa ien h ough he s eps o
ke osis and weigh loss: Visi 1 (Baseline), isi 2 (Maximum
Ke osis), isi 3 (Reduced Ke osis) and isi 4 (Endpoin ). DNA
me hyla ionandgeneexp essionwe epe o med a isi s1, 2,and4.
In all isi s, pa ien s ecei ed die a y ins uc ions, indi idual
suppo i e counsel, and encou agemen o exe cise on a egula
basis using a o mal exe cise p og am. Addi ionally, a p og am o
ein o cemen elephone calls was ins i u ed, and a phone numbe
was p o ided o all pa icipan s o add ess any conce ns.
2.3. An h opome ic assessmen
All an h opome ic measu emen s we e pe o med a e an
o e nigh as (8e10 h) unde es ing condi ions in duplica e and
pe o med by well- ained heal h wo ke s. A each isi , pa ien s
we e weighed on he same calib a ed scale (Seca 200 scale, Medical
Resou ces, EPI Inc OH, USA). BMI was calcula ed as body weigh in
kg, di ided by he squa e o body heigh in me e s (BMI ¼weigh
(kg)/heigh
2
(m)). Wais ci cum e ence (WC) was measu ed using a
s anda d flexible non-elas ic me ic ape placed o e he midpoin
be ween he las ib and he iliac c es , wi h he pa ien s anding
and exhaling.
2.4. De e mining le els o ke one bodies
Ke osis was de e mined by measu ing ke one bodies, speci -
ically
b
-hyd oxy-bu y a e (
b
-OHB), in capilla y blood using a
po able me e (GlucoMen LX Senso , A. Mena ini Diagnos ics,
Neuss, Ge many; sensi i i y <0.2 mmol/l). As wi h an h opome ic
assessmen s, all de e mina ions o capilla y ke onemia we e made
a e an o e nigh as o 8e10 h. These measu emen s we e pe -
o med daily by each pa ien du ing he en i e VLCKD, and he
co esponding alues we e e iewed using machine memo y by
he esea ch eam o con ol adhe ence. Addi ionally,
b
-OHB le els
we e de e mined a each isi by he physician in cha ge o he
pa ien .
2.5. DNA me hyla ion analysis
2.5.1. DNA p epa a ion and bisulphi e con e sion
DNA om esh- ozen (FF) blood samples was isola ed using a
s anda d phenol-chlo o o m/p o einase-k p o ocol acco ding o
he manu ac u e 's ins uc ions, wi h sligh modifica ions.
Genomic DNA was isola ed om leukocy es using he Mas-
u Pu e™DNA pu ifica ion ki (Epicen e Bio echnologies,
A.B. C ujei as, A.G. Izquie do, D. P imo e al. Clinical Nu i ion 40 (2021) 3959e3972
3961
Madison, WI, USA). The isola ed DNA was ea ed wi h RNase A o
1 h a 45

C. All DNA samples we e quan ified using he fluo o-
me ic me hod (Quan-iT PicoG een DsDNA Assay, Li e Technolo-
gies) and we e assessed o pu i y using a NanoD op (The mo
Scien ific) o de e mine 260/280 and 260/230 a io measu emen s.
The in eg i y o he FF DNA was e ified by elec opho esis in 1.3%
aga ose gel. DNA (500 ng) was bisulfi e con e ed using he EZ DNA
me hyla ion ki Me hyla ion-Gold (Zymo Resea ch, CA, USA) ac-
co ding o he manu ac u e 's ins uc ions, which con e s non-
me hyla ed cy osine in o u acil.
2.5.2. Infinium Me hyla ionEPIC BeadChip
High-quali y DNA samples (500 ng) ob ained om blood leu-
kocy es o pa ien s included in he VLCKD þDHA a m o he clinical
ial (disco e y coho ; n ¼10 pa ien s, 3 pai ed samples/pa ien )
we e selec ed o bisulfi e con e sion (Zymo Resea ch; EZ-96 DNA
Me hyla ion™Ki ) and hyb idiza ion o Infinium Me hyla ionEPIC
BeadChip (Illumina) ollowing he Illumina Infinium HD me hyl-
a ion p o ocol. DNA quali y checks, bisulfi e modifica ion, hyb idi-
za ion, da a no maliza ion, s a is ical fil e ing, and alue
calcula ions we e pe o med as p e iously desc ibed [54,55].
Whole-genome amplifica ion and hyb idiza ion we e hen
pe o med on a BeadChip ollowed by single-base ex ension and
analysis on a HiScan SQ module (Illumina) o assess cy osine
me hyla ion s a es. The anno a ion o CG islands (CGIs) used he
ollowing ca ego iza ion: 1) sho e, o each o he 2-kb sequences
flanking a CGI; 2) shel , o each o he 2-kb sequences nex o a
sho e; and 3) open sea, o DNA no included in any o he p e ious
sequences o in CGIs [54,55]. The ansc ip ion s a si e 200 and
he ansc ip ion s a si e 1500 indica e egions ei he 200 o 1500
bp om he ansc ip ion s a si e, espec i ely.
2.5.3. Py osequencing analysis
Py osequencing was used o assess selec ed ma ke s in 18 pa-
ien s ( alida ion coho : (n ¼7 de i ed om he disco e y coho
and n ¼11 om an independen coho o pa ien s; 3 pai ed
samples/pa ien )). DNA samples analyzed in he alida ion coho
we e de i ed om pa ien s included in he wo a ms o he clinical
ial and me ged o he s a is ical analysis (VLCKD: n ¼10;
VLCKD þDHA: n ¼8). The p ime sequences used in his analysis
we e designed using Qiagen's Py oMa k Assay Design 2.0 so wa e
o hyb idize o CpG- ee si es o ensu e me hyla ion-independen
amplifica ion (de ails and p ime sequences a e a ailable in
Supplemen a y Table S1). Genomic DNA was isola ed om FF blood
leukocy es using he Mas u Pu e™DNA pu ifica ion ki (Epicen e
Bio echnologies, Madison, WI, USA), acco ding o he manu ac-
u e 's ins uc ions. DNA me hyla ion analyses we e pe o med
using bisulfi e- ea ed DNA (Zymo Resea ch; EZ-96 DNA Me hyl-
a ion™Ki ) ollowed by a highly quan i a i e analysis based on
PCR-based py osequencing using he Py oMa k Q24 Sys em e sion
2.0.7 (Qiagen). Me hyla ion le el was exp essed as he pe cen age
o me hyla ed cy osine o e he sum o me hyla ed and unme hy-
la ed cy osines. Non-CpG cy osine esidues we e used as buil -in
con ols o e i y bisulfi e con e sion. The alues a e exp essed as
he mean o all si es. We also included human non-me hyla ed and
me hyla ed DNA se as con ols in each un (Zymo Resea ch). The
in e -assay p ecision (%CV) was <2.5%, in a-assay (%CV) was <1.0%.
2.6. Exp ession assay by qRT-PCR
RNA om blood leukocy es (n ¼18 pa ien s) was ex ac ed
using T izol (In i ogen) acco ding o he manu ac u e 's ecom-
menda ions. The RNA concen a ions we e measu ed wi h a
Nanod op 2000 spec opho ome e (The mo Scien ific). F om o al
ex ac ed RNA, 2
m
g we e DNase ea ed using a DNA- ee ki as a
empla e (Ambion) o gene a e fi s -s and cDNA syn hesis using
he High-Capaci y cDNA Re e se T ansc ip ion Ki (Applied Bio-
sys ems). Real- ime quan i a i e polyme ase chain eac ion (qRT-
PCR) was pe o med using TaqMan Uni e sal PCR Mas e Mix,
TaqMan P obes (Applied Biosys ems) (de ails and p ime sequences
a e a ailable in Supplemen a y Table S1), and he S ep OnePlus
Real-Time PCR Sys em (Applied Biosys ems). All expe imen s we e
pe o med in duplica e, and gene exp ession le els we e no mal-
ized o he le els o housekeeping gene GAPDH. The old change in
gene exp ession was calcula ed using he 2

DD
C
ela i e quan i a-
ion me hod acco ding o he manu ac u e 's guidelines (Applied
Biosys ems), and da a a e epo ed as he geome ic mean (SEM).
qRT-PCR expe imen s we e pe o med in compliance wi h he
MIQE (Minimum In o ma ion o Publica ion o Quan i a i e Real-
Time PCR Expe imen s) guidelines (h p://www. dml.o g/miqe).
2.7. S a is ical analysis
The sample size o he cu en s udy was calcula ed o de ec
di e ences o me hyla ion le els aking in o accoun published
alues o epigenome-wide analysis in he field o obesi y [35,40,41].
The in e en ional di e ences we e examined in wo independen
coho s. Mic oa ay-based DNA me hyla ion analysis was pe -
o med in he disco e y coho (n ¼10 pa ien s; 3 pai ed samples/
pa ien ), and hen he iden ified genes we e alida ed in an inde-
penden coho o pa ien s ( alida ion coho ; n ¼18 pa ien s; 3
pai ed samples/pa ien ). Finally, he associa ion be ween DNA
me hyla ion le el o he iden ified CpG si es and he an h opo-
me ic o biochemical pa ame e s was assessed in he global coho
o pa ien s included in his s udy (n ¼28).
The me hyla ion le el o each cy osine was exp essed as a
b
alue, which was calcula ed as he fluo escence in ensi y a io o
he me hyla ed o he unme hyla ed e sion o he p obe.
b
alues
anged be ween 0 (unme hyla ed) and 1 (comple ely me hyla ed)
acco ding o he combina ion o he Cy3 and Cy5 fluo escence in-
ensi ies. Colo balance adjus men and no maliza ion we e pe -
o med o no malize he samples be ween he wo-colo channels
using Genome S udio Illumina so wa e (V2010.3). Genome S udio
no malizes da a using di e en in e nal con ols ha a e p esen
on he Infinium Me hyla ionEPIC BeadChip. This so wa e also
no malized da a depending on in e nal backg ound p obes [55].
b
alues wi h de ec ed p- alues >0.01 we e conside ed o all below
he minimum in ensi y and h eshold, and hese CpGs we e
consequen ly emo ed om u he analysis. Addi ionally, p obes
ha con ained single nucleo ide polymo phisms (SNPs) a he 10
bp 3
0
end o he in e oga ing p obe we e fil e ed ou . To iden i y
consis en pa e ns o DMCpGs due o he nu i ional in e en ion,
a linea model was fi ed using a B-spline app oxima ion [56].
Th ee linea models we e fi ed: Model 1 was fi ed by including
he h ee poin s o he nu i ional in e en ion o e alua e he
gene al e ec o VLCKD; Model 2 including baseline and maximum
ke osis o e alua e he e ec o ke osis and weigh loss; Model 3
including me hyla ion le els a maximum ke osis and endpoin o
e alua e he e ec o only weigh loss, wi hou ke osis. P alues
we e adjus ed o mul iple compa isons using he alse disco e y
a e (FDR) p ocedu e o Benjamini and Hochbe g, and esul s we e
conside ed s a is ically significan when FDR <0.10. Addi ionally,
we applied a h eshold o he significan si es based on he mean
di e ence be ween isi s wi h a minimum
b
alue change o ±0.02.
Euclidean clus e analysis o significan CpGs was pe o med using
a hea map unc ion. The global me hyla ion le el was compa ed
be ween he nu i ional in e en ion isi s by uni a ian ANOVA
and a Bon e oni pos -hoc analysis. All o he a o emen ioned s a-
is ical analyses we e pe o med using R so wa e ( e sion 3.2.0).
A.B. C ujei as, A.G. Izquie do, D. P imo e al. Clinical Nu i ion 40 (2021) 3959e3972
3962
To es ima e en ichmen in biological p ocesses, a hype geo-
me ic es was pe o med using he GOs a s package on he bio-
logical p ocesses defined by gene on ology (GO) [57]. This analysis
de ec ed significan o e - ep esen a ion o GO e ms in one se (i.e.,
lis o iden ified genes) wi h espec o he en i e genome. GO e ms
wi h an adjus ed p- alue <0.05 we e conside ed significan .
Wi h SPSS e sion 21.0 so wa e (SPSS Inc., Chicago, IL) o Win-
dows XP (Mic oso , Redmond, WA), he genomic dis ibu ion o he
di e en ially me hyla ed CpGs was compa ed wi h he dis ibu ion o
he CpGs om all analyzed si es on he Infinium Me hyla ionEPIC
BeadChip. P alues we e compu ed using he chi-squa e es o
de e mine o e - o unde - ep esen a ion o he CpGs. The po en ial
associa ion be ween an h opome ic o biochemical pa ame e s and
DNA me hyla ion le els (
b
- alues) was e alua ed using he Spea man
coe ficien es . Di e ences in DNA me hyla ion le els and exp ession
o he iden ified genes du ing he ime-cou se o he in e en ion and
be ween he nu i ional in e en ion isi s we e assessed by he non-
pa ame ic es s, K uskal Wallis and ManneWhi ney U, espec i ely.
P0.05 was conside ed s a is ically significan .
3. Resul s
3.1. Pa ien cha ac e is ics
Samples om a o al o 21 pa ien s who ollowed he nu i ional
in e en ion based on a VLCKD we e compa ed wi h samples om
12 heal hy olun ee s wi h no mal weigh and e alua ed in his
s udy (Supplemen al Fig.1). We fi s e alua ed he disco e y coho
(n ¼10 pa icipan s wi h obesi y who ollowed a VLCKD þDHA
(n ¼5 women) and n ¼12 (6 women) olun ee s wi h no mal
weigh ). An ex ended alida ion coho composed o 11 pa ien s (7
women) wi h obesi y ha ollowed a VLCKD þDHA o a VLCKD-
DHA was also analyzed. S a is ically significan di e ences we e
no obse ed be ween ei he coho in age, gende , heigh , body
weigh , BMI, wais ci cum e ence, ke osis, o he esponse o
weigh loss ea men (Table 1). Di e ences s a is ically significan
we e only de ec ed in body weigh , BMI and wais ci cum e ence
be ween pa ien s wi h obesi y and subjec s wi h no mal weigh
(Table 1). All pa ien s los weigh a e nu i ional in e en ion
(21.8 ±4.9%), oge he wi h educ ions in BMI (21.9 ±5.1%) and
wais ci cum e ence (19.3 ±4.4%).
3.2. DNA me hyla ion changes du ing he global VLCKD
in e en ion
DNA me hyla ion p ofiles o blood leukocy es in ol ing
app oxima ely 850 housand CpGs we e analyzed a e VLCKD
in e en ion. This analysis e ealed s a is ically significan di e -
ences (cu -o poin
D
0.02; FDR 0.10) a 988 CpG si es, om a
o al o 739,222 alid CpGs (see de ailed lis in Supplemen a y
Table S2). The di e en ially me hyla ed CpGs we e mos ly cha ac-
e ized as changes owa ds CpG hypome hyla ion occu ing a e
nu i ional in e en ion, in bo h o al DMCpGs (Fig. 1A) and in he
DMCpGs loca ed in p omo e s, and islands o sho es (Fig. 1B). This
esul o global hypome hyla ion was co ela ed wi h he down-
egula ion in he exp ession o DNA me hyl ans e ase (DNMT) 1
and he de no o me hyl ans e ases DNMT3A and DNMT3B (Fig. 1C
and Supplemen al Fig. 2).
The iden ified CpG si es mapped o 786 unique genes and we
we e able o sepa a e he samples acco ding o nu i ional in e -
en ion isi s using a hie a chical clus e app oach (Fig. 1D). I
should be no ed ha he me hyla ion le els o he 988 CpG si es
di e en ially me hyla ed a e nu i ional in e en ion we e
al e ed o esemble me hyla ion le els obse ed in samples om
subjec s wi h no mal weigh (Fig. 1C). The 20 DMCpGs wi h he
highes di e ence wi h espec o baseline among genes ha a e
ep esen ed by mo e han 2 CpGs a e ep esen ed in Table 2.
Rega ding he unc ional dis ibu ion (Fig. 2), he di e ences in
DNA me hyla ion we e mainly obse ed in open sea egions
(Fig. 2A), wi h 43.2% o he DMCpG si es loca ed in p omo e e-
gions and he majo i y o DMCpGs in he body (Fig. 2B). Mo eo e ,
he DMCpGs we e mainly ound in ch omosomes 2, 9,16,17,19, and
22 when compa ed wi h all CpGs analyzed (Fig. 2C).
Among he 988 DMCpGs, we ound 886 (89.7%) wi h
dec eased and 102 (10.32%) wi h inc eased le els o DNA
me hyla ion a e nu i ional in e en ion. Mo eo e , he
DMCpGs ha los me hyla ion wi h espec o baseline we e
loca ed mainly in he open sea (Fig. 2D) and body (Fig. 2E). In
con as , he DMCpGs ha gained me hyla ion a e he nu i-
ional in e en ion we e mos ly loca ed in p omo e s (TSS 200
and 1s exon) (Fig. 2D) and in CpG islands (Fig. 2E). Rega ding
ch omosome dis ibu ion, he DMCpGs wi h dec eased me hyl-
a ion le els a e nu i ional ea men we e ound on ch omo-
somes1,4,67,8,9,14,and16,whe easDMCpGswi hinc eased
Table 1
Clinical cha ac e is ics o pa ien s a baseline and du ing he in e en ion wi h a VLCKD.
Disco e y coho Valida ion coho P alue
No mal
weigh
Obesi y Obesi y
Baseline
(day 0)
Maximum
ke osis
(day 30)
Endpoin
(day 180)
Baseline
(day 0)
Maximum
ke osis
(day 30)
Endpoin
(day 180)
Adiposi y Time Coho Time x
Coho
N1210101018*18 18 eeee
Age (yea s) 50.3 ±6.2 48.8 ±9.20 ee 47.1 ±9.8 ee 0.654 e0.678 e
Gende (men/women) 6/6 5/5 ee 7/11 ee eeee
Heigh (m) 1.67 ±0.08 1.68 ±0.08 ee 1.65 ±0.10 ee 0.773 e0.446 e
Body weigh (Kg) 63.8 ±8.7 93.4 ±9.9
a
84.3 ±8.4 73.9 ±6.7
b
90.8 ±11.6 81.9 ±10.4 70.4 ±10.4
b
<0.0001 <0.001 0.488 0.649
BMI (Kg/m
2
) 22.7 ±1.49 32.9 ±1.4
a
29.8 ±1.7 26.1 ±1.8
b
33.2 ±1.7 30.06 ±1.7 25.7 ±1.7
b
<0.0001 <0.001 0.986 0.334
Wais ci cum e ence (cm) 77.4 ±7.2 111.1 ±6.6
a
102.3 ±7.4 90.2 ±3.1
b
108.3 ±8.5 100.1 ±7.3 86.9 ±7.4
b
<0.0001 <0.001 0.325 0.836
Ke onemia (mM) —0.15 ±0.07 2.11 ±1.11
b
0.15 ±0.07 0.18 ±0.08 1.71 ±1.08
b
0.18 ±0.08 e<0.0001 0.445 0.374
Weigh loss (%) ee 9.69 ±2.51 20.58 ±5.17
c
e9.80 ±1.99 22.52 ±4.75
c
e<0.0001 0.336 0.455
Da a shown a e mean ±SD (s anda d de ia ion).*The sample size o he alida ion coho was comple ed wi h n ¼7 pa ien s om he disco e y coho and 11 pa ien s om
an independen coho .
Abb e ia ions: VLCKD, e y-low-calo ie ke ogenic die .
a
S a is ically significan di e ences compa ed wi h con ol No mal weigh in disco e y coho .
b
S a is ically significan di e ences compa ed wi h Baseline in bo h coho s.
c
S a is ically significan di e ences compa ed wi h Maximum Ke osis in bo h coho s.
A.B. C ujei as, A.G. Izquie do, D. P imo e al. Clinical Nu i ion 40 (2021) 3959e3972
3963

me hyla ion le els we e mainly ound in ch omosomes 3, 5, 10,
11, 12, 17, 21, and 22 (Fig. 2F).
3.3. Biological significance o he die a y in e en ion- ela ed
DMCpG si es and associa ed genes
In e es ingly, mos di e en ially me hyla ed genes belonged o
a ne wo k significan ly en iched in p o ein in e ac ions (p <0.001)
acco ding o STRING analysis (Fig. 3A). GO analysis was pe o med
o es whe he ce ain molecula unc ions o biological p ocesses
we e significan ly en iched wi hin he 786 genes associa ed wi h
he 988 DMCpGs disco e ed h ough VLCKD in e en ion (Fig. 3B).
Among he ca ego ies o unc ional p ocesses ha exhibi ed s a-
is ical significance (FDR <0.05), we ound p ocesses ela ed o
egula ion o ansc ip ion, signal ansduc ion, cell di e en ia ion,
p oli e a ion and apop osis, me abolic p ocesses, esponse o hyp-
oxia, p o ein p ocessing, muscle o gan and skele al sys em de el-
opmen , ne ous sys em de elopmen and axon guidance (Fig. 3B).
Among hese, we highligh ed ele an pa hways in he field o
obesi y physiopa hology such as insulin signaling, p o ein diges ion
and abso p ion, adipocy okine signaling and muscle de elopmen
(Fig. 4AeD). To in es iga e biological ele ance, he CpG si es
ep esen ing p omo e egions (TSS1500, TSS200, 5
0
UTR and 1s
Exon) a CpG islands/sho es we e selec ed. This selec ion yielded
141 CpGs ep esen ing 150 unique genes. Based on his fil e we
iden ified CpG si es ha could be gene ic a ge s whose me hyl-
a ion is associa ed wi h VLCKD esponses. Among hese, he mos
ep esen a i e gene was ZNF331, which was ep esen ed by 2 CpG
si es loca ed in he p omo e and island wi h he highes di e ence
wi h espec o baseline. Fu he mo e, FGFRL1 was also selec ed
om among he DMCpG si es because o i s biological ele ance in
me abolic pa hways and obesi y pa hogenesis and because he
me hyla ion le el o his gene in leukocy es was p e iously p o-
posed as an episigna u e ha mi o s me hyla ion le els o
dys unc ional adipose issue in obesi y [35].
3.4. DNA me hyla ion changes associa ed wi h he e ec s o
die a y-induced ke osis
An analysis compa ing baseline (day 0) wi h maximum ke osis
(day 30) yielded 1365 DMCpGs. Wi h espec o all CpGs analyzed,
hese CpGs we e ound mainly in he open sea (Fig. 5A) and body
(Fig. 5B) and in ch omosomes 1, 10, 11, 17, 19, and 22 (Fig. 5C). The
DMCpGs ha gained me hyla ion we e ound in he island and
Fig. 1. P ofile o DNA me hyla ion ollowing he VLCKD (n ¼10; Disco e y coho ). (A). Global di e ences in me hyla ion le els o 988 DMCpGs iden ified by Infinium
Me hyla ionEPIC BeadChip analysis. (B). Global di e ences in me hyla ion le els o DMCpGs loca ed in he p omo e and island/sho e. (C). Exp ession le els o DNA me hyl-
ans e ase (DNMT) 1 and he de no o me hyl ans e ases DNMT3A and DNMT3B (D). Supe ised clus e ing o he 988 CpGs ha we e ound o be di e en ially me hyla ed
be ween pa ien s wi h obesi y and heal hy olun ee s wi h no mal weigh (n ¼12) g oups. The iden ified CpG si es mapped o 786 unique genes. Di e ences s a is ically significan
de ec ed (P <0.0001). DMCpGs, di e en ially me hyla ed CpGs; VLCKD, e y-low calo ie ke ogenic die .
A.B. C ujei as, A.G. Izquie do, D. P imo e al. Clinical Nu i ion 40 (2021) 3959e3972
3964
Table 2
Lis o DMCpGs wi h he highes di e ences o Baseline - Maximum Ke osis - Endpoin be ween genes ha a e ep esen ed by mo e han 2 CpGs.
Ta ge ID CHR Posi ion Gene Name Gene egion CpG con ex Me hyla ion le els (mean) Di e ences p alue FDR
BMKEMK-BEB
cg04254103 19 1794380 ATP8B3; ATP8B3;
ATP8B3
Body; Body; Body N_Sho e 0.826 0.786 0.796 0.039 0.030 0.001 0.091
cg06643002 14 105735797 BRF1; BRF1; BRF1;
BRF1;
Body; Body; Body;
Body; Body
0.782 0.741 0.748 0.040 0.034 0.001 0.099
cg12063937 1 7731375 CAMTA1 Body 0.837 0.785 0.802 0.051 0.034 0.000 0.054
cg00035197 16 88962986 CBFA2T3; CBFA2T3 Body; Body Island 0.803 0.748 0.763 0.056 0.040 0.001 0.082
cg09962824 21 44479417 CBS Body N_Sho e 0.754 0.707 0.714 0.047 0.040 0.000 0.051
cg23790296 10 73233854 CDH23; CDH23;
CDH23; CDH23;
CDH23
Body; Body; Body;
Body; Body
0.846 0.814 0.820 0.032 0.026 0.001 0.095
cg26493726 19 36508614 CLIP3 Body S_Shel 0.808 0.771 0.773 0.037 0.035 0.000 0.069
cg01818220 9 23821773 ELAVL2; ELAVL2;
ELAVL2;
ELAVL2
5
0
UTR; TSS1500 Island 0.092 0.148 0.134 0.056 0.042 0.000 0.008
cg01933710 5 132596864 FSTL4 Body 0.828 0.778 0.799 0.049 0.029 0.001 0.076
cg07390459 2 121582002 GLI2 Body 0.768 0.720 0.727 0.048 0.041 0.001 0.097
cg14777822 2 121728297 GLI2 Body 0.770 0.728 0.726 0.042 0.044 0.001 0.093
cg16120742 7 50345049 IKZF1 5
0
UTR S_Sho e 0.114 0.058 0.041 0.057 0.073 0.000 0.026
cg02999309 10 134502626 INPP5A Body N_Sho e 0.836 0.799 0.807 0.037 0.029 0.000 0.024
cg14010696 19 5119250 KDM4B Body Island 0.853 0.812 0.819 0.041 0.034 0.000 0.008
cg16333587 14 101369826 MEG8 Body 0.857 0.817 0.828 0.040 0.029 0.000 0.051
cg20963002 14 101360088 MEG8 TSS1500 0.862 0.821 0.828 0.041 0.034 0.001 0.098
cg18026309 2 26683651 OTOF; OTOF; OTOF;
OTOF; OTOF
Body; Body; Body;
Body; Body
0.884 0.851 0.857 0.033 0.027 0.000 0.051
cg01764953 11 70814657 SHANK2 Body 0.914 0.887 0.887 0.027 0.026 0.000 0.067
cg19696891 19 54057705 ZNF331; ZNF331;
ZNF331
5
0
UTR; 5
0
UTR;
TSS1500
Island 0.379 0.431 0.447 0.052 0.068 0.000 0.015
cg03643149 19 54041519 ZNF331; ZNF331;
ZNF331; ZNF331;
ZNF331; ZNF331
TSS1500; 1s Exon;
TSS200; TSS200;
5
0
UTR; 5
0
UTR
Island 0.397 0.445 0.458 0.048 0.061 0.000 0.019
Da a shown a e mean.
Abb e ia ions: B, baseline; CHR, ch omosome; E, endpoin ; FDR, alse disco e y a e; MK, maximum ke osis.
Fig. 2. Cha ac e iza ion o he DMCpGs a e VLCKD in he disco e y coho (n ¼10). (A) Genomic dis ibu ion o he DMCpGs and hei espec i e loca ions ega ding he
b oade CpG con ex , (B) gene egion and (C) ch omosome. (D) Genomic dis ibu ion o DMCpGs compa ing hose ha exhibi ed an inc ease wi h hose ha exhibi ed a dec ease in
me hyla ion le els ollowing VLCKD, and hei espec i e loca ions in he b oade CpG con ex , (E) he gene egion and (F) ch omosome. DMCpGs, di e en ially me hyla ed CpGs;
VLCKD, e y-low calo ie ke ogenic die .
A.B. C ujei as, A.G. Izquie do, D. P imo e al. Clinical Nu i ion 40 (2021) 3959e3972
3965
Fig. 3. Biological implica ions o he DMCpGs ollowing a VLCKD. (A) Summa y o he GO analysis o he biological p ocess ca ego ies ep esen ing he di e en ially me hyla ed
genes associa ed wi h DMCpG si es. (B) Gene-p o ein in e ac ion ne wo k-STRING analysis. Mos o he genes egula ed by me hyla ion belonged o a ne wo k significan ly en iched
in p o ein in e ac ions (p <0.001) acco ding o STRING analysis. DMCpGs, di e en ially me hyla ed CpGs; GO, gene on ology; VLCKD, e y-low calo ie ke ogenic die .
Fig. 4. Genes wi h known unc ions ha p esen DMCpGs ollowing a VLCKD. (A) No el genes epigene ically egula ed ollowing VLCKD belonged o he insulin signaling
pa hway, (B) adipocy okine signaling pa hway, (C) p o ein diges ion and abso p ion unc ions, and (D) muscle o gan de elopmen unc ions. DNA me hyla ion alues a e exp essed
as b- alues om he Infinium Me hyla ionEPIC BeadChip. *Deno es di e ences s a is ically significan (P <0.05). DMCpGs, di e en ially me hyla ed CpGs; VLCKD, e y-low calo ie
ke ogenic die .
A.B. C ujei as, A.G. Izquie do, D. P imo e al. Clinical Nu i ion 40 (2021) 3959e3972
3966
p omo e and in ch omosomes 1, 2, 3, 4 and 7. The DMCpGs ha
los me hyla ion we e ound in he open sea and body and in
ch omosomes 9, 10, 11, 12, 17, 21, 22 (Fig. 5DeF).
Wi h he aim o isola ing he specific e ec s o ke osis on
me hyla ion p ofile, addi ional analysis was pe o med. A Venn
diag am was c ea ed by including: DMCpGs Baseline eMaximum
Ke osis ¼1365, DMCpGs Baseline eEndpoin ¼405, and DMCpGs
Maximum Ke osis eEndpoin ¼21. Using hese condi ions, we
iden ified 280 CpGs whose me hyla ion was a ec ed by he
induced weigh loss pe se (see de ailed lis in Supplemen a y
Table S3) and 1239 CpGs we e iden ified as VLCKD-induced
ke osis- ela ed DMCpGs (Fig. 6A). These VLCKD-induced ke osis-
ela ed DMCpGs co esponded o 966 anno a ed genes, and 161 o
hese we e ep esen ed by 137 VLCKD-induced ke osis- ela ed
DMCpGs loca ed in he p omo e and island o sho e (see de ailed
lis in Supplemen a y Table S4). Among he ca ego ies o
unc ional p ocesses ha exhibi ed s a is ical significance
(FDR <0.05), we ound p ocesses ela ed o egula ion o an-
sc ip ion, signal ansduc ion, cell adhesion, cell di e en ia ion,
p oli e a ion and apop osis, as well as ne ous sys em de elop-
men and axon guidance. Mo eo e , hese ke osis- ela ed di e -
en ially me hyla ed genes belonged o pa hways in ol ed in ocal
adhesion, insulin and adipocy okine signaling pa hways, MAPK
and P53 signaling pa hways and in cance and ype II diabe es
melli us- ela ed pa hways (Fig. 6B). O e all pa hways associa ed
wi h obesi y physiopa hology.
Among hese, we highligh ed ele an pa hways in he field o
obesi y physiopa hology such as insulin signaling, p o ein diges-
ion and abso p ion, adipocy okine signaling and muscle
de elopmen .
To iden i y po en ially no el signa u es o DNA me hyla ion
associa ed wi h VLCKD-induced ke osis, hose genes wi h mo e
Fig. 5. Cha ac e iza ion o he DMCpGs du ing ke osis induced by a VLCKD om he disco e y coho (n ¼10). (A) Genomic dis ibu ion o he DMCpGs and hei espec i e
loca ions ega ding he b oade CpG con ex , (B) gene egion and (C) ch omosome. (D) Genomic dis ibu ion o DMCpGs compa ing hose ha exhibi ed an inc ease wi h hose ha
exhibi ed a dec ease in me hyla ion le els du ing he VLCKD-induced ke osis and hei espec i e loca ions in he b oade CpG con ex , (E) he gene egion and (F) ch omosome.
DMCpGs, di e en ially me hyla ed CpGs; VLCKD, e y-low calo ie ke ogenic die .
Fig. 6. Biological implica ions o he DMCpGs ela ed o VLCKD-induced ke osis. (A) Venn diag am o he DMCpGs de ec ed be ween baseline and maximum ke osis, be ween
baseline and endpoin , and be ween maximum ke osis and endpoin . F om his analysis, 1239 CpGs we e iden ified as nu i ional ke osis- ela ed DMCpGs. (B) Summa y o he GO
analysis o he biological p ocess ca ego ies ep esen ing he di e en ially me hyla ed genes associa ed wi h nu i ional ke osis- ela ed DMCpG si es. DMCpGs, di e en ially
me hyla ed CpGs; GO, gene on ology; VLCKD, e y-low calo ie ke ogenic die .
A.B. C ujei as, A.G. Izquie do, D. P imo e al. Clinical Nu i ion 40 (2021) 3959e3972
3967