RESEARCH ARTICLE Open Access
Oxida i e s ess impai s ene gy me abolism
in p ima y cells and syno ial issue o
pa ien s wi h heuma oid a h i is
Emese Balogh
1
, Douglas J. Veale
2
, T udy McGa y
3
, Ca l O
2
, Zol an Szekanecz
1
, Chin-Teck Ng
4,5
, U sula Fea on
3†
and Monika Biniecka
2*†
Abs ac
Backg ound: In his s udy, we examined he e ec o oxida i e s ess on cellula ene gy me abolism and p o-
angiogenic/p o-in lamma o y mechanisms o p ima y heuma oid a h i is syno ial ib oblas cells (RASFC) and
human umbilical ein endo helial cells (HUVEC).
Me hods: P ima y RASFC and HUVEC we e cul u ed wi h he oxida i e s ess induce 4-hyd oxy-2-nonenal (4-HNE),
and ex acellula acidi ica ion a e, oxygen consump ion a e, mi ochond ial unc ion and p o-angiogenic/p o-
in lamma o y mechanisms we e assessed using he Seaho se analyse , complex I–V ac i i y assays, andom mu a ion
mi ochond ial cap u e assays, enzyme-linked immunoso ben assays and unc ional assays, including angiogenic
ube o ma ion, mig a ion and in asion. Exp ession o angiogenic g ow h ac o s in syno ial issue (ST) was assessed
by IHC in pa ien s wi h heuma oid a h i is (RA) unde going a h oscopy be o e and a e adminis a ion o umou
nec osis ac o inhibi o s (TNFi).
Resul s: In RASFC and HUVEC, 4-HNE-induced oxida i e s ess ep og ammed ene gy me abolism by inhibi ing
mi ochond ial basal, maximal and adenosine iphospha e-linked espi a ion and ese e capaci y, coupled wi h he
educed enzyma ic ac i i y o oxida i e phospho yla ion complexes III and IV. In con as , 4-HNE ele a ed basal
glycolysis, glycoly ic capaci y and glycoly ic ese e, pa alleled by an inc ease in mi ochond ial DNA mu a ions and
eac i e oxygen species. 4-HNE ac i a ed p o-angiogenic esponses o RASFC, which subsequen ly al e ed HUVEC
in asion and mig a ion, angiogenic ube o ma ion and he elease o p o-angiogenic media o s. In i o ma ke s o
angiogenesis ( ascula endo helial g ow h ac o , angiopoie in 2 [Ang2], y osine kinase ecep o [Tie2]) we e
signi ican ly associa ed wi h oxida i e damage and oxygen me abolism in he in lamed syno ium. Signi ican
educ ion in ST ascula i y and Ang2/Tie2 exp ession was demons a ed in pa ien s wi h RA be o e and a e
adminis a ion o TNFi.
Conclusions: Oxida i e s ess p omo es me abolism in a ou o glycolysis, an e ec ha may con ibu e o
accele a ion o in lamma o y mechanisms and subsequen dys unc ional angiogenesis in RA.
Keywo ds: Bioene ge ic me abolism, Oxida i e s ess, Angiogenesis, Rheuma oid a h i is
* Co espondence: [email p o ec ed]
†
Equal con ibu o s
2
Cen e o A h i is and Rheuma ic Diseases, Dublin Academic Medical
Cen e, S . Vincen ’s Uni e si y Hospi al, Dublin, I eland
Full lis o au ho in o ma ion is a ailable a he end o he a icle
© The Au ho (s). 2018 Open Access This a icle is dis ibu ed unde he e ms o he C ea i e Commons A ibu ion 4.0
In e na ional License (h p://c ea i ecommons.o g/licenses/by/4.0/), which pe mi s un es ic ed use, dis ibu ion, and
ep oduc ion in any medium, p o ided you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o
he C ea i e Commons license, and indica e i changes we e made. The C ea i e Commons Public Domain Dedica ion wai e
(h p://c ea i ecommons.o g/publicdomain/ze o/1.0/) applies o he da a made a ailable in his a icle, unless o he wise s a ed.
Balogh e al. A h i is Resea ch & The apy (2018) 20:95
h ps://doi.o g/10.1186/s13075-018-1592-1
Backg ound
Angiogenesis is one o he ea lies e en s in he de-
elopmen o heuma oid a h i is (RA). New blood
essels in ade he syno ial memb ane, esul ing in a
sel -pe pe ua ing and pe sis en in il a ion o immune
cells in o he join , ans o ming he syno ial issue
(ST) in o an agg essi e, umou -like ‘pannus’[1]. New
capilla ies also acili a e he deli e y o su icien oxy-
gen and nu ien s o suppo he p oli e a ing syno-
ium. Al hough angiogenesis is a p ominen ea u e
o RA, he neo ascula ne wo k is dys unc ional and
ails o es o e issue oxygen homeos asis, ende ing
he in lamed ST hypoxic. The inc ease in me abolic
u no e o he expanding syno ial pannus ou paces
he oxygen supply, esul ing in a demand o adeno-
sine iphospha e (ATP) and an al e ed egula ion o
cellula me abolic mechanisms [2,3].
Bioene ge ics is undamen ally impo an o all cells
o enable p oli e a ion, di e en ia ion and ma u a ion,
wi h mi ochond ia being cen al o biosyn he ic and
bioene ge ic pa hways media ed by he ica boxylic acid
(TCA) cycle. Thus, al e a ions o mi ochond ial espi -
a ion can play a key ole in media ing pa hogenic mech-
anisms in ch onic in lamma o y diseases [4–6]. One
well-known example o mi ochond ial dys unc ion is he
bioene ge ic swi ch in cell me abolism om oxida i e
phospho yla ion (OXPHOS) owa ds ae obic glycolysis,
known as he Wa bu g e ec . Al hough he e iciency o
ATP p oduc ion pe molecule o glucose is much lowe
h ough glycolysis, he yield a e is much as e han ha
o OXPHOS, suppo ing apid cellula g ow h. I has
been demons a ed ha he Wa bu g e ec is p esen in
highly p oli e a ing and me abolically ac i e immune
cells in a manne simila o ha obse ed in umou
cells. In he in lamed join , an inc ease in he me abolic
s a e owa ds glycolysis has been shown in p ima y
heuma oid a h i is syno ial ib oblas s (RASFC), CD4
T cells, T-helpe ype 17 (T
H
17) cells, mac ophages and
dend i ic cells [7–10]. This is pa alleled by ele a ed lac-
a e le els and diminished glucose in RA syno ial luids
as well as by inc eased ac i i y o key glycoly ic enzymes
in he RA syno ium, indica ing ha anae obic glycolysis
is a ou ed in his hypoxic en i onmen [11–13]. Mo e
ecen ly, in i o s udies by ou g oup ha e shown ha
hypoxia and Toll-like ecep o 2 (TLR2)-induced in lam-
ma ion p omo ed mi ochond ial dys unc ion and oxida-
i e s ess and ep og ammed he na u e o cellula
espi a ion in RA syno ial cells [14,15].
Oxida i e damage occu s h ough he de imen al e ec
o hypoxia and is ecognised as an impo an sou ce o gen-
omic ins abili y ha leads o espi a o y al e a ions. Hyp-
oxia p omo es o e p oduc ion o eac i e oxygen species
(ROS) ha p o oke oxida ion o polyunsa u a ed a y acids
in plasma and mi ochond ial memb anes. This gene a es
an a ay o p ima y lipid pe oxida ion p oduc s, which sub-
sequen ly decompose and o m eac i e lipid elec ophiles,
among which 4-hyd oxy-2-nonenal (4-HNE) is he mos
impo an signalling molecule [16]. 4-HNE can o m co a-
len adduc s wi h DNA, phospholipids and nucleophilic
amino acids, impai ing hei s uc u e and biological p op-
e ies. In pa icula , mi ochond ia ha e been epo ed as a
p ominen a ge o 4-HNE ac i i y [17]. Mi ochond ial
p o eins ela ed o mi ochond ial ene gy me abolism, such
as adenosine iphospha e syn hase subuni β(ATP5B), suc-
cina e dehyd ogenase la op o ein subuni and educed
o m o nico inamide adenine dinucleo ide (NADH) de-
hyd ogenase i on–sulphu p o ein 2 in he elec on ans-
po chain (ETC), and i unc ional enzyme subuni αin
he TCA cycle, a e highly suscep ible o 4-HNE-induced in-
ac i a ion [18–20]. A ecen s udy has also demons a ed
4-HNE-induced inhibi ion o si uin 3, a majo mi ochon-
d ial nico inamide adenine dinucleo ide (NAD
+
)-dependen
deace ylase, wi h subsequen up- egula ion o ascula
endo helial g ow h ac o (VEGF) exp ession by b eas can-
ce cells [21], indica ing a close connec ion be ween oxida-
i e s ess, mi ochond ial unc ion and angiogenesis.
In p e ious s udies, ou g oup assessed le els o syn-
o ial lipid pe oxida ion in pa ien s wi h RA and demon-
s a ed a signi ican in e se co ela ion be ween 4-HNE
exp ession and oxygen ension in he in lamed join ,
e lec ing mi ochond ial damage [22]. Subsequen ly, we
ha e demons a ed ha high syno ial lipid pe oxida ion
posi i ely co ela ed wi h clinical disease ac i i y sco es,
and we ha e epo ed educed 4-HNE le els in pa ien s
wi h RA who esponded o umou nec osis ac o
(TNF) blocking he apy co esponding wi h a signi ican
inc ease in pa ial oxygen p essu e in syno ial issue, in-
dica ing a educ ion in syno ial oxida i e s ess as he
join issue becomes less hypoxic [23]. In addi ion, i was
obse ed ha inc eased syno ial in lamma ion and
angiogenesis was associa ed wi h highe oxida i e s ess
[22]. Gi en he impo an ole o mi ochond ial me ab-
olism in he egula ion o in lamma o y and angiogenic
esponses, in his s udy we in es iga ed he e ec o
oxida i e s ess on he mi ochond ial bioene ge ic p o ile
and he p o-angiogenic/p o-in lamma o y mechanisms
in RASFC and human umbilical ein endo helial cells
(HUVEC). Fu he mo e, we de e mined he e ec s o
umou nec osis ac o αinhibi o s (TNFi) on he
exp ession o angiogenic ma ke s in RA in ela ion o
syno ial oxida i e s ess in i o.
Me hods
Pa ien ec ui men , a h oscopy and sample collec ion
Fi een pa ien s wi h ac i e RA we e ec ui ed om he
Rheuma ology Depa men o S . Vincen ’s Uni e si y
Hospi al, Dublin, I eland. All pa ien s ga e ully in-
o med w i en consen app o ed by he ins i u ional
Balogh e al. A h i is Resea ch & The apy (2018) 20:95 Page 2 o 15
e hics commi ee, and he esea ch was pe o med in
acco dance wi h he Decla a ion o Helsinki. Clinical
disease ac i i y was assessed wi h he 28-join Disease
Ac i i y Sco e (DAS28) using he C- eac i e p o ein
le el. Unde local anaes hesia, all pa ien s wi h RA
unde wen a h oscopy o he in lamed knee join p io
o biologic ea men (T0) and a second a h oscopy 3
mon hs a e commencemen o TNFi (T3). ST biopsies
we e used o isola ion o p ima y syno ial ib oblas s
and his ological analyses.
RASFC cul u e
RASFC biopsies ob ained a a h oscopy we e diges ed
wi h 1 mg/ml collagenase ype I (Wo hing on Biochem-
ical, Lakewood, NJ, USA) in Gibco RPMI 1640 medium
(The mo Fishe Scien i ic, Paisley, UK) o 4 hou s a
37 °C in humidi ied ai wi h 5% CO
2
. Dissocia ed cells
we e pla ed in RPMI 1640 medium supplemen ed wi h
10% Gibco FCS (The mo Fishe Scien i ic), 20 mM 4-(2-
hyd oxye hyl)-1-pipe azinee hanesul onic acid (The mo
Fishe Scien i ic), penicillin (100 U/ml), s ep omycin
(100 U/ml) and ampho e icin B (Fungizone 0.25 μg/ml;
(In i ogen, Plymou h, MN, USA). Cells we e g own o
con luence and used be ween passages 4 and 7. RASFC
we e seeded on o 96-well pla es and in o T25 lasks and
cul u ed in he p esence o 4-HNE (2.5 μM; Cayman
Chemical, Ann A bo , MI, USA), a highly eac i e end
p oduc o lipid pe oxida ion o ehicle basal medium
(0.1% e hanol). The concen a ion o 4-HNE used in he
expe imen s was based on a cell iabili y assay and p e-
iously published s udies [24]. Following s imula ion, he
e ec o ampli ied oxida i e s ess on mi ochond ial
unc ion, cellula me abolism and angiogenic esponses
was assessed as desc ibed below.
HUVEC cul u e
HUVEC (Lonza, Walke ille, MD, USA) we e incuba ed
in MCDB (The mo Fishe Scien i ic) supplemen ed wi h
L-glu amine (The mo Fishe Scien i ic), 0.5 ml epide mal
g ow h ac o (The mo Fishe Scien i ic), 50 ml FCS
(The mo Fishe Scien i ic), 0.5 ml o hyd oco isone, peni-
cillin (100 U/ml; Bioscience), s ep omycin (100 U/ml;
Bioscience) and Fungizone (0.25 μg/ml; Bioscience). Cells
we e cul u ed a 37 °C in humidi ied ai wi h 5% CO
2
and
ha es ed wi h ypsin-e hylenediamine e aace ic acid
(Lonza). Cells we e used be ween passages 20 and 30.
Oxygen consump ion a e and ex acellula acidi ica ion
a e measu ed using Seaho se echnology
Oxygen consump ion a e (OCR) and ex acellula acid-
i ica ion a e (ECAR), e lec ing OXPHOS and glycolysis,
espec i ely, we e measu ed be o e and a e ea men
wi h oligomycin (2 μg/ml), i luo oca bonylcyanide phe-
nylhyd azone (FCCP; 5 μM), an imycin A (2 μM) and 2-
deoxyglucose (2-DG; 25 mM) using he Seaho se XF24
analyse (Agilen Technologies, San a Cla a, CA, USA).
RASFC and HUVEC we e seeded a 30,000 cells pe well
in a Seaho se XF96 cell cul u e mic opla e (Agilen
Technologies) and allowed o adhe e o 24 hou s. Cells
we e insed wi h assay medium (unbu e ed DMEM sup-
plemen ed wi h 10 mM glucose, 1 mM sodium py u a e
and 2 mM L-glu amine, pH 7.4) be o e incuba ion wi h
assay medium o 30 minu es a 37 °C in a non-CO
2
in-
cuba o . Following incuba ion, cells we e s imula ed wi h
4-HNE (2.5 μM) and ehicle basal medium o 2 hou s.
Fou baseline OCR and ECAR measu emen s we e ob-
ained o e 28 minu es be o e injec ion o speci ic me a-
bolic inhibi o s. Mo eo e , o challenge he me abolic
capaci y o he RASFC and HUVEC, h ee OCR and
ECAR measu emen s we e ob ained o e 15 minu es ol-
lowing injec ion wi h oligomycin, FCCP, an imycin A
and 2-DG.
In i o mi ochond ial dys unc ion and mi ochond ial
DNA mu agenesis
ROS p oduc ion was assessed using he DCFDA Cellula
Reac i e Oxygen Species De ec ion Assay Ki (Abcam,
Camb idge, UK). RASFC we e seeded in o clea -bo omed,
da k-sided 96-well pla es a a densi y o 2.5 × 10
4
cells/well
and allowed o a ach o e nigh . Cells we e washed in 1×
bu e and s ained wi h 25 μM2′,7′-dichlo o luo escin
diace a e in 1× bu e o 45 minu es a 37 °C and 5% CO
2
.
A e s aining, cells we e washed, ea ed wi h 4-HNE and
incuba ed a 37 °C in 5% CO
2
. ROS luo escence signal was
measu ed using he Spec aMax Gemini sys em (Molecula
De ices, Sunny ale, CA, USA) wi h exci a ion and emission
wa eleng hs o 485 nm and 538 nm, espec i ely. Mean
luo escence alues om ou wells o each condi ion we e
ob ained. To cha ac e ise he equencies o andom mu a-
ionsinRASFCexposed o4-HNE o 24hou s,weuseda
mi ochond ial andom mu a ion cap u e assay.
Mi ochond ial DNA (m DNA) was ex ac ed using a
p e iously epo ed p o ocol [25]. Following ex ac ion,
10 μg o m DNA was diges ed wi h 100 U o Taq
α
I
es ic ion enzyme (New England Biolabs, Ipswich, MA,
USA), 1× bo ine se um albumin, and a Taq
α
I-speci ic
diges ion bu e (10 mM T is HCl, 10 mM MgCl
2,
100 mM NaCl, pH 8.4) o 10 hou s, wi h 100 U o Taq
α
I
added o he eac ion mix u e e e y hou . PCR
ampli ica ion was pe o med in 25-μl eac ion mix u es
con aining 12.5 μl o 2× SYBR G een B illian Mas e Mix
(S a agene, La Jolla, CA, USA), 0.1 μl o u acil DNA gly-
cosylase (New England Biolabs), 0.7 μl o o wa d and e-
e se p ime s (10 pM/μl; In eg a ed DNA Technologies,
Skokie, IL, USA), and 6.7 μlo H
2
O. The samples we e
ampli ied using a Roche Ligh Cycle 480 Ins umen
(Roche Diagnos ics, Indianapolis, IN, USA), acco ding o
he ollowing p o ocol; 37 °C o 10 minu es, 95 °C o 10
Balogh e al. A h i is Resea ch & The apy (2018) 20:95 Page 3 o 15
minu es, ollowed by 45 cycles o 95 °C o 15 seconds
and 60 °C o 1 minu e. Samples we e kep a 72 °C o 7
minu es and ollowing mel ing-cu e analysis we e imme-
dia ely s o ed a −80 °C. The p ime sequences used we e
as ollows: o m DNA copy numbe , 5′-ACAGTTTATG
TAGCTTACCTCC-3′( o wa d) and 5′-TTGCTGCG
TGCTTGATGCTTGT-3′( e e se); o andom mu a-
ions, 5′-CCTCAACAGTTAAATCAACAAAACTGC-3′
( o wa d) and 5′-GCGCTTACTTTGTAGCCTTCA-3′
( e e se).
Examina ion o mi ochond ial complexes I–V ac i i y
Mi ochond ial complexes I–V OXPHOS ac i i y assay
ki s (Abcam) we e used o sc een he di ec e ec o
4-HNE on all complexes o he mi ochond ial espi a-
o y chain. These assays a e pe o med using whole
bo ine hea mi ochond ia, a ich sou ce o OXPHOS
complexes. The ac i i y o mi ochond ial complexes
I–V was measu ed as pe he manu ac u e ’sins uc-
ions. B ie ly, OXPHOS complex I (NADH ubiquin-
one oxido educ ase) ca alyses elec on ans e om
NADH o he elec on ca ie , ubiquinone, concomi-
an ly pumping p o ons ac oss he inne mi ochon-
d ial memb ane. The p og ession o his eac ion was
moni o ed ollowing he oxida ion as a dec ease in
abso bance a op ical densi y (OD) 340 nm. OXPHOS
complex II (succina e-coenzyme Q educ ase) ca aly-
ses elec on ans e om succina e o he elec on
ca ie , ubiquinone. The p oduc , ubiquinol, is used
by complex III in he espi a o y chain, and uma a e
is necessa y o main ain he TCA cycle. The p oduc-
ion o ubiquinol in he p esence o 4-HNE was mon-
i o ed a OD 600 nm. To examine OXPHOS complex
III ac i i y, succina e (elec on dono o complex II)
and oxidised cy och ome c (elec on accep o o com-
plex III) we e added o he mi ochond ia o s a he
elec on ans e eac ion ha akes place du ing
OXPHOS.
The a e o coupled complex II + III eac ion was mea-
su ed by moni o ing he con e sion o oxidised cy o-
ch ome c in o educed o m, obse ed as an inc ease in
abso bance a OD 550 nm. OXPHOS complex IV (cy o-
ch ome c oxidase) ans e s elec ons om educed
cy och ome c o molecula oxygen and concomi an ly
pumps p o ons ac oss he inne mi ochond ial mem-
b ane. The p og ession o his eac ion was moni o ed
ollowing he oxida ion as a dec ease in abso bance a
OD 550 nm. OXPHOS complex V makes abou 95% o a
cell’s ATP using ene gy gene a ed by he p o on-mo i e
o ce and can also unc ion in he e e se di ec ion in
he absence o a p o on-mo i e o ce, hyd olysing ATP
o gene a e adenosine diphospha e (ADP) and ino ganic
phospha e. The p oduc ion o ADP by ATP syn hase can
be coupled o he oxida ion o NADH o NAD
+
, and he
p og ess o he coupled eac ion in he p esence o 4-
HNE was moni o ed as a dec ease in abso bance a OD
340 nm. Resul s we e calcula ed using So Max P o 5.3
mic opla e analysis so wa e (Molecula De ices). The
ac i i y o complexes I, II, IV and V is p opo ional o
he dec ease in abso bance, and he linea a e o educ-
ion in abso bance o e ime was calcula ed. The ac i i y
o complex III is p opo ional o he inc ease in abso b-
ance, and he linea a e o inc ease in abso bance o e
ime was calcula ed. Fo each complex, esul s a e
g aphically demons a ed as he pe cen age o enzyma ic
ac i i y in he p esence o 4-HNE ela i e o he pe cen -
age o basal ac i i y.
Quan i ica ion o p o-angiogenic media o s in RASFC
To assess he e ec s o oxida i e s ess on sec e ion o
VEGF, angiopoie in 2 (Ang2), pla ele -de i ed g ow h
ac o subuni B (PDGF-B), basic ib oblas g ow h ac-
o (bFGF), in e leukin (IL)-8, egula ed on ac i a ion,
no mal T cell exp essed and sec e ed (RANTES) and
in e cellula adhesion molecule (ICAM), RASFC we e
seeded in o 96-well pla es. Con luen RASFC we e
se um-s a ed o 24 hou s and hen cul u ed wi h 4-
HNE o 24 hou s. Supe na an s we e ha es ed, and
p o ein sec e ion le els we e quan i ied using MSD as-
says (Meso Scale Disco e y, Rock ille, MD, USA) o spe-
ci ic enzyme-linked immunoso ben assays (ELISAs)
(R&D Sys ems, Minneapolis, MN, USA).
Induc ion o p o-angiogenic mechanisms o HUVEC in
esponse o oxida i e s ess-ac i a ed RASFC
To examine i oxida i ely ac i a ed RASFC could u he
a ec p o-angiogenic mechanisms o HUVEC, RA ib o-
blas cells we e s imula ed wi h 4-HNE o 24 hou s,
and condi ioned media (CM) we e ha es ed. As a basal
medium, we used ib oblas -condi ioned media om
RASFC cul u ed in he absence o 4-HNE. Nex , he cul-
u e o HUVEC was supplemen ed wi h 10% ib oblas -
condi ioned media. To ensu e ha he e ec s on
HUVEC unc ion we e no due o esidual 4-HNE in he
10% ib oblas -condi ioned media, HUVEC we e also
cul u ed wi h RPMI 1640 medium con aining 4-HNE a
he same concen a ion (0.25 μM), which is he same
concen a ion as ha in he 10% RASFC CM. Following
24-hou exposu e o HUVEC o ib oblas -condi ioned
media, p o-angiogenic esponses o endo helial cells
we e assessed as desc ibed in he subsec ions ha
ollow.
HUVEC answell in asion chambe s
BD BioCoa Ma igel in asion chambe s (BD Biosciences,
Wokingham, UK) we e used o examine HUVEC in asion.
Cells we e seeded a a densi y o 2.5 × 10
4
pe well in he
mig a ion chambe on 8-μm memb anes p e-coa ed wi h
Balogh e al. A h i is Resea ch & The apy (2018) 20:95 Page 4 o 15
Ma igel. HUVEC media con aining 10% ib oblas -
condi ioned media was placed in he lowe well o he
chambe , and cells we e allowed o mig a e o 48 hou s.
Non-mig a ing HUVEC we e emo ed om he uppe
su ace by gen le sc ubbing. Cells ha had in aded we e
a ached o he lowe memb ane and ixed wi h 4% pa a-
o maldehyde (PFA) and s ained wi h 0.1% c ys al iole .
To assess he a e age numbe o in ading HUVEC, cells
we e coun ed in i e andom high-powe ields.
HUVEC ube o ma ion
Ma igel (50 μl; BD Biosciences, San Jose, CA, USA) was
pla ed in 96-well cul u e pla es a e hawing on ice and
allowed o polyme ise o 30 minu es a 37 °C in humidi-
ied ai wi h 5% CO
2
. HUVEC we e emo ed om cul u e,
ypsinised and esuspended a a concen a ion o 4 × 10
4
cells/ml in endo helial cell g ow h medium. Fi e hund ed
mic oli e s o cell suspension was added o each chambe
in he p esence o 10% ib oblas -condi ioned media and
cul u ed o 8 hou s. The ube analysis was de e mined
om i e sequen ial ields (magni ica ion × 10) wi h a
ocus on he su ace o he Ma igel by wo blinded ob-
se e s and a connec ing b anch be ween wo disc e e
endo helial cells was coun ed as 1 ube.
HUVEC wound epai assay
HUVEC we e seeded on o 24-well pla es and g own o
con luence. A single sc a ch wound was induced h ough
he middle o each well wi h a s e ile pipe e ip. Cells we e
subsequen ly s imula ed o 24 hou s wi h 10% ib oblas -
condi ioned media. HUVEC mig a ion ac oss he wound
ma gins om 8 hou s was assessed and pho og aphed
using a phase-con as mic oscope. Semi-quan i a i e ana-
lysis o cell epopula ion o he wound was assessed. B ie ly,
images o he sc a ch wound assays we e aken a × 10
magni ica ion. The mean closu e o he wound was manu-
ally calcula ed om he a e age o h ee indi idual mea-
su emen s om each wound. This p ocess was epea ed o
all echnical eplica es. Measu emen o sc a ches a ime 0
we e designa ed as 100% open. F om his, he pe cen age o
closu e o all sc a ches was calcula ed.
HUVEC p oli e a ion
A c ys al iole cell p oli e a ion assay was used o assess
HUVEC p oli e a ion in he p esence o RASFC-
condi ioned media. HUVEC we e seeded in o 96-well
cul u e pla es a a densi y o 5000 cells/well and le
o e nigh a 37 °C and 5% CO
2
. Nex , cells we e s imu-
la ed wi h 10% ib oblas -condi ioned media o 24 hou s.
Following cell cul u e, cells we e washed wi h PBS, ixed
in 4% PFA and s ained wi h 1% c ys al iole solu ion.
Pla es we e washed wi h ap wa e and hen d ied
o e nigh . Cells we e esuspended in 1% T i on X-100
solu ion (Sigma-Ald ich, S . Louis, MO, USA), and cell
numbe was measu ed wi h a mic opla e eade a a
wa eleng h o 550 nm.
Quan i ica ion o p o-angiogenic media o s in HUVEC
HUVEC we e seeded in o 96-well pla es and le o e -
nigh a 37 °C and 5% CO
2
. The ollowing day, cells we e
s imula ed wi h 10% ib oblas -condi ioned media o
24 hou s. Nex , supe na an s we e ha es ed, and p o-
ein sec e ion le els o Ang2 and PDGF-B we e quan i-
ied by using a speci ic ELISA (R&D Sys ems).
Immuno luo escence s aining o RASFC and syno ial
issue
Single-immuno luo escence s aining was pe o med on
RASFC ollowing 24-hou cell s imula ions wi h 4-HNE.
To isualise immunoexp ession o VEGF, cells we e ixed
in 4% PFA and s ained wi h p ima y abbi an ibody agains
VEGF (Abcam). To demons a e ST co-exp ession o
ma ke s o angiogenesis, oxida i e s ess and bioene ge ics,
dual-immuno luo escence s aining was pe o med on c yo-
s a syno ial sec ions. ST sec ions we e ixed wi h ace one
o 10 minu es and co-incuba ed wi h p ima y mouse an i-
body agains human 4-HNE (GENTAUR, Kampenhou ,
Belgium) and wi h p ima y abbi an ibodies agains VEGF,
Ang2, Tie2, ATP5B and glucose anspo e 1 (GLUT1) (all
om Abcam), glyce aldehyde 3-phospha e dehyd ogenase
(GAPDH) (T e igen, Gai he sbu g, MD, USA) and py u-
a e kinase isozyme 2 (PKM2) (Abgen , San Diego, CA,
USA). Following o e nigh incuba ion in a humidi ied
chambe , RASFC and ST samples we e incuba ed wi h
In i ogen Alexa Fluo 488-conjuga ed goa In i ogen
Supe clonal™an i-mouse seconda y an ibody (The mo
Fishe Scien i ic) and Cy™3–conjuga ed goa an i-
abbi seconda y an ibody (Jackson ImmunoResea ch,
Wes G o e, PA, USA) o 60 minu es and coun e s ained
wi h 4′,6-diamidino-2-phenylindole (DAPI) nuclea s ain
(Sigma-Ald ich) o 10 minu es. Samples we e moun ed
wi h Molecula P obes an i ade moun ing medium
(The mo Fishe Scien i ic) and assessed by immuno luo -
escence mic oscopy (Olympus BX51; Olympus, Hambu g,
Ge many).
IHC and sco ing o syno ial issue
IHC was pe o med using 7-μm c yos a ST sec ions
and he DAKO ChemMa e EnVision ki (Dako/Agi-
len Technologies, Glos up, Denma k). Sec ions we e
de os ed a oom empe a u e o 20 minu es, ixed
in ace one o 10 minu es and washed in PBS o 5 mi-
nu es. Non-speci ic binding was blocked using 1% casein
in PBS o 20 minu es. The sec ions we e incuba ed wi h
abbi monoclonal p ima y an ibodies agains human
VEGF, Ang2, Tie2, ATP5B (all om Abcam), GAPDH
(T e igen) and mouse monoclonal an ibodies agains hu-
man 4-HNE (GENTAUR). Immunoglobulin G con ol
Balogh e al. A h i is Resea ch & The apy (2018) 20:95 Page 5 o 15

an ibodies we e used as nega i e con ols. Following 1-
hou incuba ion wi h p ima y an ibody, endogenous pe -
oxidase ac i i y was blocked using 0.3% hyd ogen pe ox-
ide o 5 minu es. Slides we e incuba ed o 30 minu es
wi h seconda y an ibody/ho se adish pe oxidase (Dako/
Agilen Technologies). 3,3'-Diaminobenzidine (1:50) was
used o isualise s aining, and Maye ’shaema oxylin
(BDH Labo a o ies, Poole, UK) was incuba ed o 30 sec-
onds as a coun e s ain p io o moun ing in DPX moun -
ing media. Slides we e sco ed sepa a ely o lining laye
(LL), sublining laye (SL) and ascula egion (BV) using a
well-es ablished and alida ed semi-quan i a i e sco ing
me hod [26], whe e he pe cen age o cells ha we e posi-
i e o a speci ic ma ke was compa ed wi h he pe cen -
age o cells ha we e nega i e. Pe cen age posi i i y was
g aded using a 0–4 scale, whe e 0 = no s ained cells, 1 =
1–25%, 2 = 25–50%, 3 = 50–75 and 4 = 75–100% s ained
cells. Images we e cap u ed using an Olympus DP50 ligh
mic oscope and AnalySIS so wa e (Olympus So Imaging
Solu ions, Lakewood, CO, USA).
S a is ical analysis
IBM SPSS S a is ics e sion 20 o Windows so wa e
(IBM, A monk, NY, USA) was used o s a is ical ana-
lysis. Wilcoxon’s signed- ank es , Spea man’s ank-
co ela ion coe icien and he Mann-Whi ney U es
we e used o analysis o non-pa ame ic da a. Pa ame -
ic da a we e analysed using one-way analysis o a i-
ance. All p alues we e wo-sided, and p alues less han
0.05 we e conside ed s a is ically signi ican .
Resul s
Oxida i e s ess al e s cellula bioene ge ics in RASFC and
HUVEC in i o
P e ious s udies by ou g oup demons a ed al e ed cel-
lula bioene ge ics in RASFC in he p esence o hypoxia
[14], and we ha e also demons a ed high oxida i e
s ess in he in lamed syno ium [22]. The e o e, in his
s udy, we u he in es iga ed whe he oxida i e s ess in
he in lamed join is in ol ed in me abolic ep og am-
ming o RASFC and HUVEC. Figu e 1a demons a es
ep esen a i e OCR and ECAR p o iles be o e and a e
injec ions o oligomycin, FCCP, an imycin A and 2-DG
in basal and 4-HNE-s imula ed RASFC. We show, o
he i s ime o ou knowledge, ha inhibi ion o OCR
ollowing 4-HNE-induced oxida i e s ess was associa ed
wi h a shi in RASFC me abolism owa ds glycolysis. 4-
HNE educed basal mi ochond ial espi a ion (p<0.05),pa -
alleled by a educ ion in maximal mi ochond ial espi a ion
(p< 0.001), ATP syn hesis (p= 0.1) and ese e capaci y (p<
0.01) (Fig. 1b). This me abolic ep og amming was u he
accompanied by inc eased le els o basal glycolysis (p<0.01),
glycoly ic capaci y (p< 0.01) and glycoly ic ese e (p=0.
2) in RASFC subjec ed o oxida i e s ess (Fig. 1b).
Rep esen a i e HUVEC OCR and ECAR p o iles be-
o e and a e injec ions o oligomycin, FCCP, an imy-
cin A and 2-DG a e shown on Fig. 2a. Simila ly o
RASFC, 4-HNE inhibi ed basal mi ochond ial espi -
a ion, maximal mi ochond ial espi a ion, ATP syn hesis
and ese e capaci y (all p< 0.01) wi h concomi an ele a-
ion o basal glycolysis (p< 0.01) and glycoly ic ese e (p
< 0.05) in HUVEC exposed o oxida i e s ess (Fig. 2b).
Examina ion o mi ochond ial mu agenesis and ac i i y o
enzymes o mi ochond ial OXPHOS complexes unde 4-
HNE-induced oxida i e s ess
We ha e p e iously shown ha inc eased m DNA mu a-
ion equency and mi ochond ial dys unc ion in he RA
join we e s ongly associa ed wi h syno ial in lamma ion
and hypoxia [27,28]. We ha e also epo ed, a a unc-
ional le el, induc ion o p o-angiogenic esponses o
endo helial cells in he p esence o oxida i e s ess [29]. In
he p esen s udy, we assessed he equency o m DNA
mu a ions and mi ochond ial dys unc ion in RASFC sub-
jec ed o 4-HNE. We obse ed inc eases in ROS p oduc-
ion and m DNA poin mu a ions in RASFC in he
p esence o 4-HNE compa ed wi h basal cells (p<0.001
and p= 0.06, espec i ely) (Fig. 3a). 4-HNE p o ein
adduc ion may al e p o ein ac i i y; he e o e, we nex ex-
amined he ac i i y o he indi idual p o eins o mi ochon-
d ial OXPHOS complexes I–V. 4-HNE signi ican ly
educed he ac i i y o complex III by 8% and complex IV
by 70% compa ed wi h basal alues (bo h p<0.01).Lowe
enzyma ic ac i i y ollowing 4-HNE s imula ion was also
de ec ed o complex I by 9%, complex II by 22% and
complex V by 12% (all p=0.2)(Fig.3b).
In i o sec e ion o p o-angiogenic and p o-in lamma o y
media o s unde oxida i e s ess condi ions
Because we ound a close associa ion o edox s a e wi h
ene gy me abolism in RASFC,we nex examined he e -
ec o oxida i e s ess on angiogenic and in lamma o y
media o s om RASFC. Figu e 4demons a es inc eased
VEGF immuno luo escence s aining in RASFC cul u ed
in he p esence o 4-HNE compa ed wi h he basal cells.
In addi ion, 4-HNE signi ican ly inc eased sec e ion o
key p o-in lamma o y and p o-angiogenic media o s
compa ed wi h basal RASFC (VEGF, Ang2, bFGF, IL-8
[all p< 0.05], PDGF-B, RANTES, ICAM [all p< 0.01]).
These indings, along wi h ou p e iously published in
i o s udy showing TNF-α-induced mi ochond ial dys-
unc ion [28], u he suppo he concep o he com-
plex in e play be ween oxida i e damage, oxygen
me abolism and angiogenesis in RA. The e o e, we nex
de e mined angiogenic in i o esponses ollowing TNFi
in 15 pa ien s wi h RA a baseline (T0) and 3 mon hs
a e he commencemen o biologic ea men (T3).
Addi ional ile 1: Figu e S1A shows changes o
Balogh e al. A h i is Resea ch & The apy (2018) 20:95 Page 6 o 15
mac oscopic ascula i y and ST exp ession o VEGF,
Ang2 and Tie2 om T0 o T3. Addi ional ile 1: Figu e
S1B g aphically illus a es dec eases in ST VEGF (p=0.
1), Ang2 (p< 0.005) and Tie2 (p< 0.005) a e TNFi
he apy.
Oxida i e s ess-ac i a ed RASFC p omo e p o-angiogenic
mechanisms in HUVEC
RASFC a e known o be s ongly in ol ed in egula ing
pa hological angiogenesis in he in lamed join [30].
The e o e, we nex examined i he obse ed al e a ions in
cellula bioene ge ics and p o-in lamma o y p ocesses in
RASFC in esponse o oxida i e s ess could subsequen ly
in luencep o-angiogenicmechanismsinHUVEC.Wes im-
ula ed RASFC in he p esence o absence o 4-HNE and ha -
es ed he supe na an s, e med condi ioned media.Fig.5a
demons a es he e ec o basal o 4-HNE RASFC-CM on
in asion, he o ma ion o ube-like s uc u es and mig a ion
o HUVEC. Figu e 5b g aphically illus a es ma kedly in-
duced in asion (p< 0.001), p oli e a ion (p< 0.05),
numbe o o med ube-like s uc u es (p< 0.001), cell
mig a ion ac oss he wound (p< 0.001) and sec e ion
o Ang2 and PDGF-B (bo h p alues < 0.05) in
HUVEC in esponse o basal o 4-HNE RASFC-CM.
To con i m ha he inc ease in p o-angiogenic e-
sponses o HUVEC was due o oxida i ely ac i a ed
RASFC and no o esidual 4-HNE p esen in he CM,
addi ional expe imen s we e pe o med, consis ing o
RPMI 1640 media supplemen ed wi h 4-HNE (0.25 μM;
4-HNE RPMI 1640 con ol), which would be a he same
concen a ion o 4-HNE in he 10% RASFC-CM. A sig-
ni ican inc ease in in asion (p< 0.001), numbe o o med
Fig. 1 Bioene ge ic me abolism in p ima y heuma oid a h i is syno ial ib oblas cells (RASFC) subjec ed o 4-hyd oxy-2-nonenal (4-HNE)-induced
oxida i e s ess. aRep esen a i e oxygen consump ion a e (OCR) and ex acellula acidi ica ion a e (ECAR) Seaho se analyse p o iles be o e and
a e injec ions o oligomycin, i luo oca bonylcyanide phenylhyd azone (FCCP), an imycin A and 2-deoxyglucose (2-DG) in RASFC in he p esence
and absence o 4-HNE. bBa g aphs demons a e quan i ica ion o basal mi ochond ial (M ) espi a ion, maximal M espi a ion, adenosine
iphospha e (ATP) syn hesis, ese e capaci y, basal glycolysis, glycoly ic capaci y and glycoly ic ese e in RASFC (n= 5) subjec ed o oxida i e
s ess. Da a a e p esen ed as mean ± SEM. *p< 0.05, **p< 0.01, and ***p< 0.001, signi ican di e ences om basal le el
Balogh e al. A h i is Resea ch & The apy (2018) 20:95 Page 7 o 15
ube-like s uc u es (p< 0.01) and cell mig a ion ac oss
he wound (p < 0.001) in HUVEC in esponse o 4-HNE
RASFC-CM compa ed wi h 4-HNE RPMI 1640 con ol
media u he suppo s he di ec e ec o 4-HNE on
RASFC-induced angiogenesis in he in lamed join (Add-
i ional ile 2: Figu e S2).
Associa ion be ween ST angiogenesis, oxida i e s ess
and bioene ge ics
Finally, he co ela ion o angiogenic ac o s wi h p e i-
ously assessed ma ke s o oxida i e s ess and me abol-
ism in his pa ien coho was examined [14]. ST 4-HNE
exp ession was associa ed wi h inc eased exp ession o
VEGF ( = 0.63; p= 0.015) and Tie2 ( = 0.56; p= 0.029),
GAPDH ( = 0.60; p= 0.03) and wi h educed le els o
ATP5B (p=−0.52, p= 0.017). Fu he mo e, ep esen a i e
immuno luo escence images demons a ing co-localisa ion
o 4-HNE wi h angiogenic ac o s (VEGF, Ang2, Tie2)
, as well as wi h mi ochond ial (ATP5B) and glycoly ic
(GAPDH, PKM2, GLUT1) p o eins, is demons a ed in
Fig. 6. Addi ional ile 3: Figu e S3 and Addi ional ile 4:
Figu e S4 show single images o VEGF, Ang2, Tie2,
GAPDH, PKM2, GLUT1 and ATP5B (all in ed), single
images o 4-HNE immuno luo escence (in g een), as well
as single images o DAPI (in blue), along wi h hei con-
ols wi h iso ype-ma ched an ibodies.
Discussion
In his s udy, we demons a e, o he i s ime o ou
knowledge, ha oxida i e s ess ep og ams cellula bio-
ene ge ics o RASFC and HUVEC by down egula ing
OXPHOS and p omo ing glycolysis. This change was
Fig. 2 Bioene ge ic me abolism in human umbilical ein endo helial cells (HUVEC) subjec ed o 4-hyd oxy-2-nonenal (4-HNE)-induced oxida i e
s ess. aRep esen a i e oxygen consump ion a e (OCR) and ex acellula acidi ica ion a e (ECAR) Seaho se analyse p o iles be o e and a e
injec ions o oligomycin, i luo oca bonylcyanide phenylhyd azone (FCCP), an imycin A and 2-deoxyglucose (2-DG) in HUVEC in he p esence and
absence o 4-HNE. bBa g aphs demons a e quan i ica ion o basal mi ochond ial (M ) espi a ion, maximal M espi a ion, adenosine
iphospha e (ATP) syn hesis, ese e capaci y, basal glycolysis, glycoly ic capaci y and glycoly ic ese e in HUVEC (n= 3) subjec ed o oxida i e
s ess. Da a a e p esen ed as mean ± SEM. *p< 0.05 and **p< 0.01, signi ican di e ences om basal le el
Balogh e al. A h i is Resea ch & The apy (2018) 20:95 Page 8 o 15
e lec ed by a dec ease in mi ochond ial maximal and
ATP-linked espi a ion and ese e capaci y, whe eas
glycoly ic capaci y and glycoly ic ese e we e ele a ed
in he p esence o 4-HNE. A bioene ge ic swi ch was
coupled wi h highe ROS p oduc ion and m DNA
mu a ions, in addi ion o he educed enzyma ic ac i i y
o mi ochond ial complexes III and IV. Oxida i e s ess
also induced sec e ion o p o-angiogenic and p o-
in lamma o y media o s by RASFC. CM om 4-HNE-
ac i a ed RASFC po en ia ed p o-angiogenic mecha-
nisms in HUVEC, as e lec ed by ele a ed cell in asion,
p oli e a ion, mig a ion, he o ma ion o ube-like s uc-
u es and sec e ion o p o-angiogenic media o s. In i o
co-exp ession o angiogenic ma ke s, oxida i e damage
and oxygen me abolism was demons a ed in ST. Finally,
a dec ease in ST angiogenesis was obse ed in pa ien s
wi h RA ollowing TNFi he apy.
Hypoxia is a undamen al me abolic change in ST o RA
associa ed wi h ele a ed mi ochond ial ROS p oduc ion
and lipid pe oxida ion. Co alen modi ica ions o m DNA,
lipids and p o eins by 4-HNE ha e been epo ed o com-
p omise mi ochond ial in eg i y and unc ion, including e-
spi a o y me abolism, p o ein anspo a ion, mi ochond ial
dynamics and quali y con ol h ough ission, usion and
mi ophagy [16]. We ha e p e iously shown ha inc eased
m DNA mu a ion equency and mi ochond ial dys unc-
ion in he RA join co ela ed wi h g ea e hypoxia, oxida-
i e s ess, ascula i y and p o-in lamma o y cy okines [27,
28]. Ou p esen in i o indings using RASFC u he
demons a e high suscep ibili y o he mi ochond ial gen-
ome o oxida i e damage. A mi ochond ial andom mu a-
ion cap u e assay was used o quan i y he equency o
andom mi ochond ial poin mu a ions in RASFC ollow-
ing 4-HNE s imula ions. This me hodology elies on single-
Fig. 3 Mi ochond ial mu agenesis and ac i i y o enzymes o mi ochond ial oxida i e phospho yla ion (OXPHOS) complexes unde 4-hyd oxy-2-nonenal
(4-HNE)-induced oxida i e s ess. aBa g aphs demons a e inc eased p oduc ion o eac i e oxygen species (n= 7), pa alleled by he g ea e equency o
mi ochond ial DNA mu a ion (n= 5) in p ima y heuma oid a h i is syno ial ib oblas cells (RASFC) in esponse o 4-HNE. bAc i i y o mi ochond ial
OXPHOS complexes I–V in he p esence o 4-HNE. 4-HNE educes he ac i i y o complex I by 9%, complex II by 22%, complex III by 8%, complex IV by
70% and complex V by 12% (all complexes measu ed in iplica e). Fo each complex, esul s a e g aphically demons a ed as he pe cen age o enzyma ic
ac i i y in he p esence o 4-HNE ela i e o he pe cen age o basal ac i i y. Da a is ep esen ed as Mean ± SEM, **p<0.01; ***p<0.001 signi ican ly di e en
o basal
Balogh e al. A h i is Resea ch & The apy (2018) 20:95 Page 9 o 15