Serum Thioredoxin-80 is associated with age, ApoE4, and neuropathological biomarkers in Alzheimer’s disease: a potential early sign of AD

Goikoleae al. Alzheime ’s Resea ch & The apy (2022) 14:37
h ps://doi.o g/10.1186/s13195-022-00979-9
RESEARCH
Se um Thio edoxin-80 isassocia ed wi hage,
ApoE4, andneu opa hological bioma ke s
inAlzheime ’s disease: apo en ial ea ly sign
o AD
Julen Goikolea1*, Go ka Ge enu2,3,4, Mak ina Daniilidou1,5, F ancesca Mangialasche5, Pa izia Mecocci6,
Tiia Ngandu5,7, Juha Rinne8, Alina Solomon5,9,10, Miia Ki ipel o5,10,11,12, Angel Cedazo‑Minguez1,
Anna Sandeb ing‑Ma on1,5,9† and Sil ia Maioli1†
Abs ac
Backg ound: Thio edoxin‑80 (T x80) is a clea age p oduc om he edox‑ac i e p o ein Thio edoxin‑1 and has
been p e iously desc ibed as a p o‑in lamma o y cy okine sec e ed by immune cells. P e ious s udies in ou g oup
epo ed ha T x80 le els a e deple ed in Alzheime ’s disease (AD) b ains. Howe e , no s udies so a ha e in es iga ed
pe iphe al T x80 le els in he con ex o AD pa hology and whe he could be associa ed wi h he main known AD isk
ac o s and bioma ke s.
Me hods: T x80 was measu ed in se um samples om pa icipan s om wo di e en coho s: he obse a ional
memo y clinic biobank (GEDOC) (N = 99) wi h AD CSF bioma ke da a was a ailable and he popula ion‑based
li es yle mul idomain in e en ion ial Finnish Ge ia ic In e en ion S udy o P e en Cogni i e Impai men and Dis‑
abili y (FINGER) (N = 47), wi h neu oimaging da a and blood ma ke s o in lamma ion a ailable. The GEDOC coho
consis s o pa icipan s diagnosed wi h subjec i e cogni i e impai men (SCI), mild cogni i e impai men (MCI), and
AD, whe eas he FINGER pa icipan s a e olde adul s a ‑ isk o demen ia, bu wi hou subs an ial cogni i e impai ‑
men . One‑way ANOVA and mul iple compa ison es s we e used o assess he le els o T x80 be ween g oups. Linea
eg ession models we e used o explo e associa ions o T x80 wi h cogni ion, AD CSF bioma ke s (Aβ42, ‑ au, p‑ au
and p‑ au/ ‑ au a io), in lamma o y cy okines, and neu oimaging ma ke s.
Resul s: In he GEDOC coho , T x80 was associa ed o p‑ au/ ‑ au a io in he MCI g oup. In he FINGER coho ,
se um T x80 le els co ela ed wi h lowe hippocampal olume and highe p o‑in lamma o y cy okine le els. In bo h
GEDOC and FINGER coho s, ApoE4 ca ie s had signi ican ly highe se um T x80 le els compa ed o non‑ApoE4 ca i‑
e s. Howe e , T x80 le els in he b ain we e u he dec eased in AD pa ien s wi h ApoE4 geno ype.
Conclusion: We epo ha se um T x80 le els a e associa ed o AD disease s age as well as o se e al isk ac o s
o AD such as age and ApoE4 geno ype, which sugges s ha T x80 could ha e po en ial as se um AD bioma ke .
© The Au ho (s) 2022. Open Access This a icle is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License, which
pe mi s use, sha ing, adap a ion, dis ibu ion and ep oduc ion in any medium o o ma , as long as 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 licence, and indica e i changes we e made. The images o
o he hi d pa y ma e ial in his a icle a e included in he a icle’s C ea i e Commons licence, unless indica ed o he wise in a c edi line
o he ma e ial. I ma e ial is no included in he a icle’s C ea i e Commons licence and you in ended use is no pe mi ed by s a u o y
egula ion o exceeds he pe mi ed use, you will need o ob ain pe mission di ec ly om he copy igh holde . To iew a copy o his
licence, isi h p:// c ea i eco mmons. o g/ licen ses/ by/4. 0/. The C ea i e Commons Public Domain Dedica ion wai e (h p:// c ea i eco
mmons. o g/ publi cdoma in/ ze o/1. 0/) applies o he da a made a ailable in his a icle, unless o he wise s a ed in a c edi line o he da a.
Open Access
*Co espondence: [email p o ec ed]
†Anna Sandeb ing‑Ma on and Sil ia Maioli con ibu ed equally o his
wo k.
1 Depa men o Neu obiology, Ca e Sciences and Socie y, Di ision
o Neu oge ia ics, Cen e o Alzheime Resea ch, Ka olinska Ins i u e ,
S ockholm, Sweden
Full lis o au ho in o ma ion is a ailable a he end o he a icle
Page 2 o 13
Goikoleae al. Alzheime ’s Resea ch & The apy (2022) 14:37
In oduc ion
Alzheime ’s disease (AD) is he leading cause o demen-
ia and i is conside ed a mul i ac o ial and complex dis-
o de [1]. The main pa hological hallma ks in AD b ains
a e neu o ib illa y angles (NFT) composed o hype -
phospho yla ed au p o ein (p- au) and amyloid plaques
consis ing o agg ega ed amyloid be a (Aβ). I has been
p oposed ha hese e en s, in u n, may lead o neu o-
in lamma o y p ocesses led by mic oglia, synap ic and
neu onal dys unc ion, and ul ima ely b ain a ophy [2].
The as majo i y o AD cases occu spo adically, known
as la e-onse AD (LOAD), and a e d i en by a complex
combina ion o gene ics and en i onmen al ac o s [1].
While aging is s ill he highes isk ac o o AD, se -
e al suscep ibili y genes ha e been epo ed [3]. Among
hem, apolipop o ein E4 (ApoE4) is ega ded as he
s onges gene ic isk ac o o LOAD [4]. ApoE has
h ee iso o ms (ε2, ε3 and ε4) whe e he ApoE4 geno ype
con e s highe isk o AD as compa ed o ApoE3 and
ApoE2. Ca ying one copy o APOE4 inc eases AD isk
by 3–4 old and wo copies by 10–15 old compa ed o
hose ca ying wo copies o APOE3 [5–7]. The ApoE4
iso o m has a dec eased capaci y o bind and anspo
lipids including choles e ol a ec ing lipid homeos asis
[8], ye i s mechanisms o ac ion in ol e also o he dis-
ease pa hways. P e ious s udies show ha he p esence
o he ApoE4 iso o m accele a es Aβ accumula ion, glio-
sis, and au phospho yla ion compa ed wi h he o he
iso o ms [9, 10]. ApoE4 has also been shown o inc ease
in lamma ion in mouse b ain [11]. Indeed, mice exp ess-
ing humanized ApoE4 and adminis e ed wi h LPS
showed a signi ican inc ease in TNFα and IL-6 in he
b ain in compa ison o ApoE3 mice [4]. In he pe iph-
e y, APOE4 a ec s he oxida i e s a us in mac ophages,
p oducing mo e supe oxide anion adicals han ApoE3
mac ophages [12]. Al oge he , ApoE4 seems o ha e an
impac on edox and in lamma o y p ocesses, wo cen al
e en s in AD.
Thio edoxin-1 (T x1) is a highly conse ed endog-
enous di hiol wi h many di e en oles, including eac-
i e oxygen species (ROS) sca enging o chemokine
ac i i ies, and i is dec eased in AD neu ons [13]. T x1
can be clea ed by he α-sec e ase ac i i y o ADAM10/17
in o an 80 amino-acid long pep ide, known as Thio e-
doxin-80 (T x80) [14]. Mos o he s udies ega ding
T x80 unc ion ha e been pe o med in pe iphe al blood
mononuclea cells whe e i igge s inna e immuni y by
inducing he ac i a ion and di e en ia ion o human
monocy es [15], he up egula ion o cell su ace pa ho-
gen ecogni ion ecep o s, and he p oduc ion o se -
e al p o-in lamma o y cy okines [16]. P e ious s udies
epo ed ha T x80 le els inc ease signi ican ly in se um
wi h aging and unde ch onic in lamma o y condi ions
[17–19]. Ou lab showed ha T x80 le els a e deple ed in
b ains and ce eb ospinal luid (CSF) o MCI pa ien s who
la e con e ed o AD [14], sugges ing a po en ial use o
T x80 as a bioma ke o ele ance o AD p og ession.
Howe e , no s udies so a ha e in es iga ed pe iphe al
T x80 le els in he con ex o neu odegene a ion and AD.
In he cu en explo a o y s udy, T x80 was meas-
u ed in se um samples om subg oups o wo di e en
coho s: a obse a ional memo y clinic coho GEDOC
and he popula ion-based coho om he li es yle in e -
en ion ial Finnish Ge ia ic In e en ion S udy o P e-
en Cogni i e Impai men and Disabili y (FINGER). The
GEDOC g oup (n = 99) con ains pa ien s diagnosed wi h
subjec i e cogni i e impai men (SCI), mild cogni i e
impai men (MCI), and AD, whe eas he FINGER RCT
included pa icipan s om he gene al popula ion a isk
o demen ia bu wi hou subs an ial cogni i e impai -
men (n = 47).
The aim o his s udy was o in es iga e (a) whe he
se um T x80 le els a e al e ed a di e en s ages o
demen ia and (b) whe he se um T x80 le els a e asso-
cia ed wi h demog aphical o clinical AD isk ac o s.
Addi ionally, we aimed o in es iga e whe he any ound
associa ions o T x80 could also be ound in a subg oup
o a - isk pa icipan s o FINGER whe e we analyzed
T x80 associa ions wi h demen ia- ela ed ma ke s.
Finally, we measu ed T x80 le els in AD pos mo em
b ain issue o ApoE3 and ApoE4 pa ien s.
Ma e ials andme hods
GEDOC memo y clinic popula ion
This s udy included 99 pa ien s equally dis ibu ed
be ween he clinical diagnos ic g oups subjec i e cogni-
i e impai men (SCI) mild cogni i e impai men (MCI)
and AD om he Ka olinska Uni e si y Hospi al memo y
clinic in Huddinge, Sweden. The demog aphical cha ac-
e is ics o he GEDOC memo y clinic popula ion a e
Inc eased se um T x80 and dec eased b ain T x80 le els was pa icula ly seen in ApoE4 ca ie s. Whe he his could
con ibu e o he mechanism by which ApoE4 show inc eased ulne abili y o de elop AD would need o be u he
in es iga ed.
T ial egis a ion: Clini calT ials. go NCT01 041989. Regis e ed on 4 Janua y 2010— e ospec i ely egis e ed
Keywo ds: Thio edoxin‑80, In lamma ion, Demen ia, ApoE4, Alzheime ’s disease, Aging
Page 3 o 13
Goikoleae al. Alzheime ’s Resea ch & The apy (2022) 14:37
desc ibed in Table1. The clinical and demog aphical da a
ele an o he s udy (e.g., ApoE geno ype) was acqui ed
om he cen al GEDOC da abase (Ka olinska Uni e -
si y Hospi al, S ockholm, Sweden).
Diagnosis o MCI was done using he consensus c i-
e ia o MCI which equi e he p esence o bo h subjec-
i e and objec i e cogni i e impai men including one o
se e al cogni i e domains, bu no demen ia o impai -
men o daily li ing ac i i ies [20]. Demen ia diagnoses
we e ca ied ou ollowing he c i e ia o he Diagnos ic
and S a is ical Manual o Men al Diso de s, 4 h edi ion
(DSM-IV), as p e iously desc ibed [21].
Rou ine neu ological and physical examina ions we e
ca ied ou a he memo y clinic as desc ibed p e i-
ously [21] and included Mini-Men al S a e Examina ion
(MMSE), blood es s and CSF sampling. CSF samples
we e collec ed as p e iously desc ibed [21]. F esh sam-
ples we e used o measu e soluble Aβ42, - au, and p- au
concen a ions in CSF wi h comme cially a ailable sand-
wich enzyme-linked immunoso ben assays (ELISA)
(Innogene ics, Belgium) acco ding o s anda dized p o-
ocols in he memo y clinic.
FINGER s udy pa icipan s
This explo a o y sub-s udy included baseline da a om
47 FINGER ial pa icipan s (21 women and 26 men,
mean age 71 ± 5.1 yea s) who unde wen neu oimag-
ing in Tu ku (Finland) [22] wi h se um samples a ail-
able o T x80 measu emen s (Fo CONSORT lowcha ,
see Fig.1A). They we e selec ed a he ime when MRI
esou ces became a ailable, and i he e we e no con-
aindica ions. The demog aphical and clinical cha -
ac e is ics o hese pa icipan s ha e been p e iously
desc ibed [22, 23], and hey we e no di e en om he
es o he FINGER pa icipan s.
The FINGER popula ion cha ac e is ics [24], ial p o-
ocol de ails [25], main ele an indings [26], and neu-
oimaging sub-s udy ha e been p e iously published
in de ail [22, 27]. The demog aphical, clinical, and neu-
oimaging da a ele an o he s udy was acqui ed
om he cen al FINGER da abase (Finnish Ins i u e
o Heal h and Wel a e, Helsinki, Finland). In b ie , he
da a collec ion was ongoing be ween Sep embe 7, 2009,
and No embe 24, 2011, whe e 2654 indi iduals we e
sc eened, and 1260 pa icipan s o ages be ween 60 and
Table 1 Demog aphic, clinical, and bioma ke da a o GEDOC memo y clinic coho
Values a e means ± SD unless o he wise speci ied. Be ween-g oup di e ences we e analyzed wi h chi-squa e and ANOVA as app op ia e. Cogni i e sco es a e mean
alues, whe e highe sco es indica e be e pe o mance. ANCOVA es s adjus ing o age we e used o compa e CSF bioma ke and T x80 le els be ween g oups. p
alue was conside ed signi ican (ma ked in bold) i < 0.05
ApoE4 apolipop o ein E4, SCI subjec i e cogni i e impai men , MCI mild cogni i e impai men , AD Alzheime ’s disease, CSF ce eb ospinal luid, Aβ42 amyloid-be a 42
agmen , -Tau o al au p o ein, p-Tau phospho yla ed au p o ein, MMSE mini men al s a e examina ion
NSCI, mean (SD) NMCI, mean (SD) NAD, mean (SD) p
Demog aphic da a
Sex, % men/women 8/17 32.00/68.00 10/14 41.67/58.33 8/17 32.00/68.00 0.72
Age, yea s 25 59.00 (7.85) 24 69.21 (6.69) 25 73.40 (10.48) 0.0001
Yea s o educa ion 25 14.28 (2.79) 24 11.75 (3.65) 25 10.32 (3.08) 0.001
Cogni ion
MMSE es sco e 25 29.24 (1.16) 23 27.30 (1.84) 25 24.52 (3.63) 0.0001
ApoE geno ype
ApoE4 ca ie , % yes/no 6/9 40.00/60.00 10/12 45.45/54.55 6/8 42.86/57.14 0.95
ApoE allele equencies, %
e2/e2 1 2.70 0 0.00 0 0.00
e2/e3 2 5.41 2 6.45 2 6.45
e2/e4 1 2.70 2 6.45 0 0.00
e3/e3 6 16.22 10 32.26 6 19.35
e3/e4 4 10.81 6 19.35 5 16.13
e4/e4 1 2.70 2 6.45 1 3.23
Unknown 22 59.46 9 29.03 17 54.84
CSF measu emen s
Aβ42, ng/l 37 935.32 (229.86) 31 881.87 (350.37) 31 658.32 (302.59) 0.011
‑Tau , ng/l 37 224.97 (91.82) 31 325.58 (162.98) 31 499.71 (228.49) 0.001
p‑Tau, ng/l 37 46.46 (15.32) 31 59.13 (25.95) 31 81.03 (28.17) 0.01
Se um measu emen s
T x80, ng/ml 37 3.56 (4.87) 31 28.13 (30.39) 31 25.50 (28.58) 0.0001
Page 4 o 13
Goikoleae al. Alzheime ’s Resea ch & The apy (2022) 14:37
77 yea s om he gene al popula ion we e ec ui ed
based on cogni i e pe o mance a he mean le el o
sligh ly lowe han expec ed o hei age acco ding o
Finnish popula ion no ms o he Conso ium o Es ab-
lish a Regis y o Alzheime ’s Disease (CERAD) [28]
and 6 poin s o highe in he Ca dio ascula Risk Fac-
o s, Aging and Demen ia (CAIDE) isk sco e [29]. Diag-
nosis o demen ia o subs an ial cogni i e impai men
was used as eason o exclusion as well as any condi ion
a ec ing sa e pa icipa ion.
E hics app o al
The s udy was app o ed by he Regional E hical Re iew
Boa d in S ockholm, and w i en in o med consen was
ob ained om all pa icipan s.
Cogni i e ou comes
Mini-Men al S a e Examina ion (MMSE) was conduc ed
as s anda d cogni i e es ing o all GEDOC and FINGER
pa icipan s. Highe sco es indica ed be e pe o mance.
MRI imaging
All pa icipan s unde wen a b ain 3T MRI (3D Tu bo
Field Echo sequence, oxel size 1.0 × 1.0 × 1.0 mm, o al
slices 160, ield o iew 240 × 240 mm, epe i ion ime
8.1 ms, echo ime 3.7 ms, Philips Ingenui y TF PET/MR,
Ams e dam, he Ne he lands). The speci ic p o ocol o
MRI imaging has been p e iously desc ibed [22]. T1WI
and FLAIR images we e quali y checked and isually
inspec ed o any abno mali ies by an expe ienced neu-
o adiologis . Images we e excluded i he e we e b ain
lesions po en ially a ec ing olume y and/o scanning
issues. Regula phan om scans we e pe o med, and
quan i a i e measu es o signal- o-noise a io, uni o m-
i y, and geome ic dis o ion we e ca ied ou .
F eesu e image analysis ( e sion 5.3.0) was used
o measu e co ical hickness and egional b ain ol-
umes. Manual edi ing was pe o med in cases whe e
au oma ed whi e ma e (WM) segmen a ion p esen ed
geome ic inaccu acies in bounda ies be ween CSF,
g ay, and whi e ma e . B ain olumes we e adjus ed o
head size o accoun o be ween-pa ien a ia ions in
head size [30]. Co ical hickness measu emen in AD
signa u e egions was calcula ed as he mean a e age o
co ical hickness in he en o hinal, in e io empo al,
middle empo al, and usi o m egions [31].
Whi e ma e lesions (WML) olume was meas-
u ed h ough he segmen a ion o WM hype in ensi-
ies on T1 and FLAIR images acco ding o a p e iously
desc ibed me hod [32]. The segmen a ion was done in
3 s eps based on he expec a ion-maximiza ion algo-
i hm: [1] om he T1 images, WM was segmen ed in o
wo classes, no mal-b igh WM egions, and hypoin-
ense WM egions [2]; FLAIR images we e segmen ed
o h ee classes: CSF, no mal b ain issue and hype in-
ense oxels [3]; he WM and subco ical egions we e
segmen ed in o wo classes om he FLAIR images. The
segmen a ion o WM hype in ensi ies was ega ded as
he class wi h highe in ensi ies [32, 33].
Thio edoxin‑80 ELISA measu emen s
De e mina ion o T x80 le els in se um was pe o med
wi h a sandwich ELISA as p e iously desc ibed [34]
bu wi h a ew modi ica ions. In b ie , s anda d sam-
ples o ecombinan human T x80 (#11522), coa ing
an i-T x80 monoclonal mouse an ibody (clone 7D11
#11543), and de ec ion an ibody (bio inyla ed goa
polyclonal an i-T x1 #11541) we e pu chased om
Cayman chemicals (USA). S anda d dilu ions o T x80
(0.2–125 ng/ml) and se um samples we e p epa ed in
blocking bu e (0.5% BSA, 0.05% Tween-20); 50 μl o
s anda ds o samples we e added in duplica es and
incuba ed o e nigh (O/N). Bio inyla ed goa an i-T x1
an ibody was p epa ed a a concen a ion o 2 μg/ml in
blocking bu e (50 μl/well) and incuba ed 2 h a oom
empe a u e. S ep a idin-HRP (#OR03L, Me ck,
USA) was dilu ed o 10 ng/ml in PBS and incuba ed
45 min a oom empe a u e. TMB subs a es (#34021,
The mo Fishe Scien i ic, USA) we e mixed 1:1, added
100 μl/well, and incuba ed 25 min in he da k; 50 μl/
well o s op solu ion (2N sul u ic acid, DY994, R&D
sys ems, USA) was added and abso bance was imme-
dia ely measu ed a 450 nm in he spec opho om-
e e (Tecan Sa i e 2 Mul i-De ec ion Pla e Reade ,
Swi ze land).
Fig. 1 A CONSORT diag am o he FINGER explo a o y Thio edoxin‑80 sub‑s udy. CERAD, Conso ium o Es ablish a Regis y o Alzheime ’s Disease.
B Se um T x80 le els by disease diagnosis. Pa icipan s we e di ided in o ou g oups acco ding o hei disease s a e. FINGER coho pa icipan s
and SCI, MCI, and AD pa icipan s om he GEDOC coho . The g aph shows se um T x80 le els (ng/ml) be ween g oups. p alues a e calcula ed
om one‑way ANCOVA adjus ed o age. C Hippocampal olume o pa ien s wi h highes and lowes se um T x80 le els in FINGER coho . The
g aph shows hippocampal olume (ml) be ween g oups. ‑ es was used o analyze he di e ences be ween g oups. D Se um T x80 le els in ApoE4
ca ie s and non‑ca ie s om me ged GEDOC and FINGER coho s. The g aph shows se um T x80 le els (ng/ml) be ween g oups. p alues a e
calcula ed om one‑way ANCOVA adjus ed o age. E Wes e n‑bo analysis o T x80 le els in pos ‑mo em AD and non‑AD b ain samples. Samples
a e so ed by APOE geno ype ApoE3/ApoE3 (E3/E3) and E4/E4. S uden ‑ es was used o analyze he di e ences be ween g oups. *p < 0.05;
**p < 0.01; ****p < 0.0001
(See igu e on nex page.)
Page 5 o 13
Goikoleae al. Alzheime ’s Resea ch & The apy (2022) 14:37
Fig. 1 (See legend on p e ious page.)

Page 6 o 13
Goikoleae al. Alzheime ’s Resea ch & The apy (2022) 14:37
Plasma cy okines andchemokines measu emen s
inFINGER
In he FINGER s udy, enous blood samples we e aken a
baseline in as ing s a us and using EDTA ubes. Plasma
aliquo s we e s o ed a –80°C un il analysis. A panel o
cy okines, chemokines, and g ow h ac o s (IL-1α, IL-1β,
IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-8, IL-10, IL-12(p40),
IL-12(p70), IL-13, IL-15, IL-17, IFN-α2, IFN-γ, MIP-1α,
TNF-β) we e analyzed wi h a mul iplex suspension a ay
sys em using Bioplex Luminex 200 ins umen (Bio-Rad
Labo a o ies, He cules, CA, USA) and he MILLIPLEX®
MAP Human Cy okine/Chemokine panel (Me ck Milli-
po e, Da ms ad , Ge many).
The assays we e pe o med in one ba ch, and samples
p epa a ion and se ing o he sys em unning p o ocol
we e done ollowing he manu ac u e ’s ins uc ions. All
samples and s anda ds we e un in duplica e and we e
measu ed as pg/ml. Quali y con ols we e pe o med
acco ding o he manu ac u e guidelines o ensu e accu-
acy o measu emen s. A e he pla e eading, he esul s
iles we e gene a ed using Bio-Plex Manage so wa e 4
(Bio-Rad Labo a o ies, He cules, CA, USA).
Human b ain samples
Human b ain issue samples we e ob ained om CIEN
ounda ion issue bank o neu ological esea ch (BT-
CIIN). In o ma ion ega ding age, sex, ApoE4 geno ype,
pos mo em in e al, and demen ia diagnosis om he
b ain issue is displayed in Supplemen a y Table1.
Sample p epa a ion andwes e n blo
Immunoblo ing was pe o med as desc ibed [35].
B ie ly, human b ain issue was homogenized in lysis
bu e (50 mM T is-HCL, 150 mM NaCl, 1% T i on-X)
con aining phospha ase and p o ease inhibi o s (Sigma-
Ald ich, USA). Homogena es we e hen cen i uged
a 15000×g o 15 min 4 °C, and samples we e mixed
wi h equal olumes o loading bu e (160 mM T is-HCl
pH 6.8, 4% SDS, 20% glyce ol, 0.01% b omophenol blue,
100 mM DTT). P ocessed samples we e hen un in 12.5%
polyac ylamide gels (Bio-Rad, USA) and ans e ed o
BioT aceTM ni ocellulose memb anes (GE heal hca e,
USA) o 2 h oom empe a u e a 50 V. Memb anes we e
hen blocked o 1 h in 5% skim milk in TBS-Tween-20
(TBS-T) p io o o e nigh incuba ion wi h dilu ed p i-
ma y an ibody a 4 °C. P ima y an ibodies: 1:1000 an i-
T x80 monoclonal mouse an ibody (clone 7D11 #11543,
Cayman chemicals, USA), 1:1000 an i-T x1(human) Poly-
clonal Goa An ibody (#11538 Cayman chemicals, USA),
and 1:10000 monoclonal An i-β-Ac in mouse an ibody,
clone AC-15 (Sigma-Ald ich, USA). Fluo escen sec-
onda y an ibodies (1:10000, LI-COR Biosciences, USA)
we e used o 2 h a oom empe a u e, and bands we e
isualized using ODYSSEY In a ed Imaging Sys em (LI-
COR Biosciences). Band in ensi y signal was quan i ied
by ImageJ so wa e. Each band signal alue was no mal-
ized agains loading con ol (ac in signal alue).
S a is ical analysis
Demog aphic da a we e compa ed ac oss clinical diag-
nos ic g oups wi h ANOVA o con inuous and χ2 o
ca ego ical a iables. Simple eg ession analysis was
applied when explo ing he ela ionship be ween age, sex
and ApoE4 ca ie ship and T x80 le els. Fu he associa-
ions we e analyzed by mul iple linea eg ession models
adjus ed o age and clinical diagnosis when app op ia e.
Fo a iables ha we e no no mally dis ibu ed (CSF
AD bioma ke s (Ab42, - au, p- au and p- au/ - au a io),
in lamma o y cy okines and T x80), ze o-skewness log
ans o ma ion was applied. ANCOVA es s adjus ing
o age we e used o compa e CSF bioma ke and T x80
le els be ween g oups and T x80 le els be ween ApoE4
ca ie s and non-ca ie s. S uden - es was used o
compa e hippocampal olumes be ween g oups. Co -
ec ion o mul iple es ing has no been applied, and all
he analyses pe o med a e pos hoc. Le el o signi icance
was se o p < 0.05 in all analyses. Analysis was pe o med
using he S a a so wa e, e sion 14 (S a aCo p), and
G aphPad P ism, e sion 9 (G aphPad So wa e, CA).
Resul s
Popula ion cha ac e is ics o GEDOC andFINGER
sub‑s udies
Demog aphical and clinical da a as well as CSF and
se um measu emen s om a subse o he Ka olinska
Uni e si y Hospi al memo y clinic coho GEDOC a e
lis ed in Table1. The MCI and AD g oups we e signi i-
can ly olde han he SCI g oup (69 ± 7 and 73 ± 10 e -
sus 59 ± 8 yea s, p < 0.0001), had lowe le el o educa ion
(12 ± 4 , 10 ± 3 e sus 14 ± 3 yea s, p < 0.001), and had
lowe MMSE sco es compa ed o SCI (27.30 ± 1.84,
24.52 ± 3.63 e sus 29.24 ± 1.16 poin s, p < 0.0001).
As expec ed om p e ious epo ed measu emen s
[2], bo h MCI and AD g oups had signi ican ly lowe
Aβ42 le els compa ed o SCI (881.87 ± 350.37 ng/l and
658.32 ± 302.59 e sus 935.32 ± 229.86 ng/l, p < 0.001),
highe - au le els han he SCI g oup (324.58 ± 162.98,
499.71 ± 228.49, e sus 224.97 ± 91.82 ng/l, p < 0.0001),
and highe p- au le els in CSF han SCI g oup
(59.13 ± 25.95, 81.03 ± 28.17 e sus 46.46 ± 15.32 ng/l,
p < 0.0001).
The baseline cha ac e is ics o he FINGER sub-s udy
(n = 47) a e lis ed in Table2. In addi ion o se um T x80
le els, he he ein analyzed da a con ains demog aphics,
cogni i e measu emen s, dis ibu ion o ApoE geno ype,
neu oimaging, and in lamma o y ma ke da a. Since only
Page 7 o 13
Goikoleae al. Alzheime ’s Resea ch & The apy (2022) 14:37
baseline da a we e included in his explo a o y s udy,
bo h con ol and in e en ion g oups a e analyzed as one
“a isk o demen ia” ca ego y.
Se um T x80 le els a e signi ican ly inc eased inAD
i espec i ely o age
As epo ed in Table1, he SCI g oup in GEDOC was
signi ican ly younge han MCI and AD g oups. Acco d-
ing o wha p e iously has been shown [17], age (β = 0.33,
p= 0.00) bu no sex (men, 13.33 ± 3.40 ng/ml; women,
19.41 ± 3.44 ng/ml; p = 0.26) had a signi ican in lu-
ence on T x80 in he GEDOC coho . In con as o he
GEDOC coho , whe e he e was a signi ican di e -
ence in age among pa icipan s, in he FINGER coho ,
he e was no associa ion be ween se um T x80 le -
els and age (β = − 0.04, p= 0.77) no wi h sex (men,
21.26 ± 8.95 ng/ml; women, 18.24 ± 5.41 ng/ml; p = 0.26).
Based on his, we decided o include FINGER sub-s udy
pa icipan s in he compa ison since hey ep esen
non-demen ed subjec s a isk o AD, whose a e age
age (70 ± 5 yea s) is compa able o MCI and AD g oups.
When se um T x80 le els we e analyzed be ween hese
4 g oups (Fig.1B), we obse ed ha he SCI g oup had
signi ican ly lowe T x80 le els (3.56 ± 4.87 ng/ml) han
he FINGER sub-s udy (19.46 ± 38.37 ng/ml; p < 0.01),
MCI g oup (28.13 ± 30.39 ng/ml; p < 0.001), and AD
g oup (25.50 ± 28.58 ng/ml; p < 0.001). When compa -
ing he age ma ched FINGER sub-s udy pa icipan s o
AD, signi ican ly highe T x80 le els we e seen in AD
(p < 0.02), and a simila end could be obse ed o MCI
g oup (p < 0.10). Based on his esul , we combined he
da a om bo h sub-s udies (GEDOC and FINGER) o
pe o m u he analysis on se um T x80 associa ions
(n = 145).
T x80 isassocia ed wi hage andApoE4 inGEDOC
andFINGER combined da ase
In ag eemen wi h he abo e esul s, T x80 le els signi i-
can ly co ela ed wi h age (β = 0.23, p = 0.01) bu no wi h
sex (β = 0.14, p = 0.12) in he me ged g oup (Table 3).
When adjus ing o age, he e was a signi ican associa-
ion be ween T x80 le els and ApoE4 ca ie s (β = 0.34,
p = 0.00).
Se um T x80 le els a e nega i ely associa ed wi hp‑ au
in heSCI g oup andposi i ely wi hp‑ au/ ‑ au a io inMCI
g oup o  heGEDOC coho
Da a on clinical CSF AD bioma ke s we e a ailable
om he GEDOC memo y clinic coho . To explo e
how T x80 a ies wi h known AD pa hology, we in es-
iga ed he associa ion be ween se um T x80 le els and
CSF Aβ42, - au, p- au le els, and p- au/ - au a io in he
GEDOC coho (Table4). No associa ions be ween T x80
Table 2 Demog aphic, clinical, and bioma ke da a o he
FINGER neu oimaging coho a baseline
Values a e means ± s anda d de ia ion (SD) unless o he wise speci ied.
Cogni i e sco es a e mean alues o cogni i e es s, whe e highe sco es indica e
be e pe o mance. *B ain olumes a e shown unadjus ed; AD signa u e
co ical hickness was calcula ed as he a e age o co ical hickness in he
en o hinal, in e io empo al, middle empo al, and usi o m egions. N, numbe
o pa icipan s wi h a ailable da a o each analysis. MRI magne ic esonance
imaging; IL-1 in e leukin-1, Mip-1a mac ophage in lamma o y p o ein 1a, TNF-b
umo nec osis ac o -be a
Numbe Mean (±SD)
Demog aphical da a
Sex, % male/ emale 26/21 55.32/44.68
Age, yea s 47 70.66 (5.06)
Cogni i e assessmen
MMSE 47 27.02 (1.78)
ApoE geno ype
ApoE4 ca ie s, % (yes/no) 14/32 30.43/69.57
ApoE allele equencies (%)
2.3 2 4.35
2.4 2 4.35
3.3 30 65.22
3.4 11 23.91
4.4 1 2.17
MRI
*To al g ay ma e olume, ml 47 575.76 (55.50)
*To al hippocampal olume, ml 47 7.47 (0.99)
AD signa u e co ical hickness, mm 47 2.73 (0.13)
*Whi e ma e lesion olume, ml 46 13.27 (16.18)
To al in ac anial olume, ml 47 1548.82 (252.08)
Se um measu emen s
T x80, ng/ml 47 19.46 (36.37)
In lamma o y ma ke s (pg/ml)
IL‑1α 41 1465.39 (2558.60)
IL‑1β 41 123.47 (379.81)
IL‑2 41 65.23 (159.62)
IL‑3 32 15.93 (9.48)
IL‑4 41 169.00 (239.26)
IL‑5 41 57.16 (78.22)
IL‑6 41 58.17 (57.09)
IL‑7 41 44.73 (22.54)
IL‑8 41 58.17 (57.09)
IL‑10 41 151.23 (165.27)
IL‑12 p40 41 435.83 (639.53)
IL‑12 p70 41 130.07 (149.32)
IL‑13 41 112.00 (157.59)
IL‑15 41 81.81 (85.99)
IL‑17 41 77.17 (104.78)
In e e on‑α2 41 204.32 (149.36)
In e e on‑γ 41 132.48 (178.42)
Mip‑1a 36 43.22 (37.77)
TNF‑β 41 109.24 (146.66)
B adykinin 47 6610000 (3870000)
Page 8 o 13
Goikoleae al. Alzheime ’s Resea ch & The apy (2022) 14:37
and Aβ42 o T x80 and au we e ound when adjus ing
o age. As shown in Table1, se um T x80 le els a e sig-
ni ican ly highe in MCI and AD g oups in compa ison
o SCI. Since se um T x80 is inc eased wi h demen ia
and aking in o accoun ha Aβ42, - au, and p- au le -
els a e de e minan in he de elopmen o AD, we nex
analyzed whe he hese ma ke s co ela ed wi h se um
T x80 le els a di e en disease s ages when adjus ed
o age (SCI, MCI and AD, Table4). Highe Tx80 le -
els we e signi ican ly associa ed wi h lowe p- au le -
els (β = − 0.48, p = 0.02) in he SCI g oup, and a simila
end could be obse ed o - au (β = − 0.36, p = 0.08).
In he MCI g oup, highe p- au/ - au a io was associa ed
wi h highe T x80 le els (β = 0.41, p = 0.05). The e we e
no u he signi ican associa ions o Aβ42, - au, and
p- au o he a io in MCI o AD g oups.
Se um T x80 isassocia ed wi hlowe hippocampal olume
andhighe p o‑in lamma o y cy okine le els in heFINGER
coho
We nex sough o de e mine i T x80 is ela ed o b ain
olume measu es. We ound ha he e was a nega i e
associa ion be ween hippocampal olume and se um
T x80 le els (β = − 0.32, p = 0.01, Table 5). No o he
associa ion was ound signi ican be ween g ay ma e
olume, co ical hickness, o whi e ma e lesions and
se um T x80 le els.
No clinical cu o o T x80 has been epo ed; how-
e e , we wan ed o explo e i he e we e di e ences in
neu oimaging da a among FINGER pa icipan s when
hey a e di ided in o wo g oups acco ding o hei
se um T x80 le els. Pa icipan s wi h he highes se um
T x80 le els had signi ican ly lowe hippocampal olume
han he pa icipan s wi h he lowes se um T x80 le -
els (7.07 ± 1.05 ml; 7.82 ± 0.78 ml, espec i ely, p < 0.01;
Fig.1C). No di e ences we e ound when compa ing g ay
ma e olume, co ical hickness o whi e ma e lesions
be ween pa icipan s wi h he highes and lowes T x80
le els.
Since p e ious s udies epo ed ha T x80 is impli-
ca ed in he in lamma o y esponse p esen in a he oscle-
o ic lesions [17, 19], and i induces a p o-in lamma o y
esponse in monocy es and mac ophages [15, 36, 37], we
in es iga ed possible associa ions be ween se um T x80
le els and se um in lamma o y cy okine le els o he
pa icipan s. Highe T x80 le els we e associa ed wi h
highe le els o IL-8 (β = 0.33, p = 0.05), IL-13 (β = 0.33,
p= 0.05), Mip-1a (β = 0.39, p = 0.03), and b adykinin
(β = 0.32, p= 0.04), and a simila end, howe e no sig-
ni ican , was obse ed o IL-4 (β = 0.32, p = 0.06), IL-5
(β = 0.31, p = 0.07), IL-6 (β = 0.30, p = 0.07), in e e on-α
(β = 0.29, p = 0.08), in e e on-γ (β = 0.28, p = 0.10), and
TNF-β (β = 0.30, p = 0.08).
ApoE4 geno ype impac s se um T x80 le els andi s
associa ion oCSF Aβ42
As shown in Fig. 1D, ApoE4 ca ie s om bo h
GEDOC and FINGER coho s had signi ican ly highe
(app oxima ely wo old) T x80 le els han non-ca ie s
(30.09 ± 41.00 ng/ml; 13.39 ± 18.88 ng/ml, espec i ely,
p < 0.01). Due o he signi ican ole o ApoE4 in AD
onse and p og ession, we in es iga ed he associa ions
be ween T x80 and CSF AD bioma ke s in ApoE4 ca i-
e s and non-ca ie s (Table6). Highe se um T x80 le -
els we e associa ed wi h lowe Aβ42 in CSF (β = − 0.46,
p = 0.04) in non-ApoE4 ca ie s when adjus ed o age.
A simila end could be obse ed when adjus ing o
disease diagnosis (β = − 0.36, p = 0.07). This associa-
ion could no be obse ed in ApoE4 ca ie s. To al au
Table 3 Se um T x80 associa ions wi h age, sex, and ApoE4
geno ype in a combined da ase
Values a e s anda dized β coe icien s (p alues) om linea eg ession models
wi h se um T x80 le els as a dependen a iable. Linea eg ession models
a e non-adjus ed o adjus ed o age. p alues conside ed signi ican (bold) i
p < 0.05
Non‑adjus ed
β (p)Adjus ed o age
β (p)
Demog aphic cha ac e is ics
Age 0.23 (0.01)
Sex 0.14 (0.12) 0.13 (0.14)
ApoE4 ca ie 0.17 (0.10) 0.34 (0.00)
Table 4 Associa ions be ween se um T x80 and CSF AD bioma ke s in he GEDOC coho
Values a e s anda dized β coe icien s (p alues) om linea eg ession models wi h se um T x80 le els as a dependen a iable. Linea eg essions a e adjus ed o
age. p alues conside ed signi ican (bold) i p < 0.05; #, p < 0.10. N, numbe o pa icipan s wi h a ailable da a o each analysis
To al SCI MCI 6AD
N β (p)N β (p)N β (p)N β (p)
Aβ42 99 − 0.14 (0.23) 37 0.21 (0.38) 31 − 0.15 (0.50) 31 − 0.12 (0.56)
‑ au 99 0.12 (0.31) 37 − 0.36 (0.08)#31 − 0.23 (0.29) 31 0.29 (0.17)
p‑ au 99 0.13 (0.27) 37 − 0.48 (0.02) 31 − 0.03 (0.89) 31 0.27 (0.20)
p‑ au/ ‑ au 99 − 0.07 (0.55) 37 − 0.07 (0.73) 31 0.41 (0.05) 31 − 0.21 (0.34)
Page 9 o 13
Goikoleae al. Alzheime ’s Resea ch & The apy (2022) 14:37
and p- au did no show any signi ican associa ions wi h
T x80 in hese g oups. The a io p- au/ - au was signi i-
can ly associa ed o se um T x80 le els in ApoE4 non-
ca ie s (β = 0.49, p = 0.00).
We also in es iga ed whe he he associa ions be ween
ApoE4 and MMSE sco es wi h T x80 we e a ec ed o e
he disease spec um (Table7). To his end, we analyzed
hese associa ions wi hin each coho (FINGER pa ici-
pan s and memo y clinic pa ien s (SCI, MCI and AD)).
A posi i e associa ion was ound be ween se um T x80
le els and ApoE4 in FINGER pa icipan s (β = 0.30,
p = 0.05) and in he MCI g oup om GEDOC coho
(β = 0.59, p = 0.02). This associa ion was no ound in he
SCI nei he in he AD g oups. Rega ding cogni i e sco e
da a, he e was a signi ican associa ion be ween high
T x80 le els and low MMSE sco es only he MCI g oup
(β = − 0.41, p = 0.05).
T x80 le els a e dec eased inApoE4 AD b ains
Finally, we measu ed T x80 p o ein le els in pos -mo -
em human b ain homogena es o AD and non-AD
pa ien s. Demog aphical and clinical da a om he b ain
samples has been collec ed in Supplemen a y Table1.
AD samples we e g ouped by ApoE geno ype in ApoE3/
ApoE3 (E3/E3) o E4/E4 (Fig.1E). Despi e he low num-
be sample size (n = 3 pe g oup), only E4/E4 AD pa ien s
had signi ican ly lowe T x80 le els han E3/E3 non-
demen ed con ols (p < 0.01). As p e iously epo ed by
ou g oup [14], T x80 le els in E3/E3 AD pa ien s we e
also lowe han he non-AD con ols, al hough his di -
e ence did no each s a is ical signi icance (p = 0.08)
in his sho sample numbe s udy. The e we e no sig-
ni ican di e ences in he p ecu so T x1 p o ein le els
be ween g oups.
Discussion
T x80 has been p e iously desc ibed as a p o-in lamma-
o y cy okine sec e ed by immune cells in he pe iphe y
[15, 34, 36]. P e ious s udies showed ha T x80 is able
o p e en Aβ agg ega ion and o inhibi Aβ oxic e ec s
Table 5 Associa ions be ween se um T x80 and neu oimaging
and in lamma o y ma ke s in he FINGER coho
Values a e s anda dized β coe icien s (p alues) om linea eg ession models
wi h se um T x80 le els as a dependen a iable. Linea eg essions a e adjus ed
o age. *B ain olumes a e shown adjus ed o head-size and ime be ween
blood sample collec ion and b ain scan; AD signa u e co ical hickness
was calcula ed as he a e age o co ical hickness in he en o hinal, in e io
empo al, middle empo al, and usi o m egions. p alues conside ed signi ican
(bold) i p < 0.05. MRI magne ic esonance imaging, IL-1 in e leukin-1, Mip-1a
mac ophage in lamma o y p o ein 1a, TNF-α umo nec osis ac o -be a
Adjus ed o age
β (p)
MRI*
To al g ay ma e olume, ml − 0.07 (0.67)
Hippocampal olume, ml − 0.32 (0.01)
AD signa u e co ical hickness, mm − 0.08 (0.61)
Whi e ma e lesions, ml − 0.11 (0.50)
In lamma o y ma ke s
IL‑1α 0.09 (0.59)
IL‑1β 0.12 (0.47)
IL‑2 0.10 (0.55)
IL‑3 0.19 (0.36)
IL‑4 0.32 (0.07)
IL‑5 0.31 (0.08)
IL‑6 0.30 (0.08)
IL‑7 0.12 (0.49)
IL‑8 0.33 (0.05)
IL‑10 0.16 (0.33)
IL‑12 p40 0.10 (0.56)
IL‑12 p70 0.05 (0.77)
IL‑13 0.33 (0.05)
IL‑15 0.14 (0.41)
IL‑17 0.17 (0.32)
In e e on‑α2 0.29 (0.10)
In e e on‑γ 0.28 (0.10)
Mip‑1a 0.40 (0.03)
TNF‑β 0.30 (0.08)
B adykinin 0.32 (0.04)
Table 6 Se um T x80 associa ions wi h CSF Aβ42, p‑ au, and ‑ au in ApoE4 ca ie s and non‑ca ie s
Values a e s anda dized β coe icien s (p alues) om linea eg ession models wi h se um T x80 le els as a dependen a iable in GEDOC coho . Linea eg essions
a e adjus ed o age and disease diagnosis. Co esponding p alues a e shown om analyses wi h alues adjus ed ze o-skewness log- ans o med as app op ia e. p
alues conside ed signi ican (bold) i p < 0.05; #, p < 0.10. N, numbe o pa icipan s wi h a ailable da a o each analysis
Adjus ed o age Adjus ed o diagnosis
ApoE4 non‑ca ie s ApoE4 ca ie s ApoE4 non‑ca ie s ApoE4 ca ie s
N β (p)N β (p)N β (p)N β (p)
Aβ42 29 − 0.48 (0.04) 22 0.51 (0.11) 29 − 0.36 (0.07)#22 0.27 (0.25)
‑ au 29 − 0.13 (0.60) 22 − 0.39 (0.18) 29 − 0.38 (0.06)#22 − 0.04 (0.81)
p‑ au 29 0.04 (0.87) 22 − 0.32 (0.27) 29 − 0.19 (0.37) 22 − 0.09 (0.62)
p‑ au/ ‑ au 29 0.34 (0.16) 22 0.23 (0.50) 29 0.49 (0.00) 22 0.01 (0.98)