1
Scien i ic RepoR S | (2020) 10:572 | h ps://doi.o g/10.1038/s41598-019-57348-0
www.na u e.com/scien i ic epo s
he mi ochond ial nega i e
egula o McJ modula es he
in e play be ween mic obio a and
he hos du ing ulce a i e coli is
Miguel Angel pascual-i oiz
1,8, Ainize peña-cea a1,2,8, I zia Ma ín-Ruiz1, José Luis La ín1,
ca olina Simó3, Héc o Rod íguez1, Es ibaliz A ondo1, Juana Ma ía lo es4, Ana ca e as-
González1, Julen omás-co áza 1, Diego Ba iales
1, Ainhoa palacios1, Vi ginia Ga cía-
cañas3, Aize pellón
1, Asie ullaondo2, Ana Mª A ansay
1,6, Ra ael p ados-Rosales1,
Rebeca Ma ín
5, Juan Angui a
1,7 & Le icia Abecia
1*
Recen e idences indica e ha mi ochond ial genes and unc ion a e dec eased in ac i e ulce a i e
coli is (UC) pa ien s, in pa icula , he ac i i y o Complex I o he elec on anspo chain is hea ily
comp omised. MCJ is a mi ochond ial inne memb ane p o ein iden i ied as a na u al inhibi o o
espi a o y chain Complex I. The induc ion o expe imen al coli is in MCJ-de icien mice leads o he
up egula ion o Timp3 exp ession esul ing in he inhibi ion o TACE ac i i y ha likely inhibi s Tn and
Tn 1 shedding om he cell memb ane in he colon. MCJ-de icien mice also show highe exp ession
o Myd88 and Tl 9, p oin lamma o y genes and disease se e i y. In e es ingly, he absence o MCJ
esul ed in dis inc mic obio a me abolism and composi ion, including a membe o he gu communi y
in UC pa ien s, Ruminococcus gna us. These changes p o oked an e ec on IgA le els. Gene exp ession
analyses in UC pa ien s showed dec eased le els o MCJ and highe exp ession o TIMP3, sugges ing
a ele an ole o mi ochond ial genes and unc ion among ac i e UC. The MCJ de iciency dis u bs he
egula o y ela ionship be ween he hos mi ochond ia and mic obio a a ec ing disease se e i y.
Ou esul s indica e ha mi ochond ia unc ion may be an impo an ac o in he pa hogenesis. All
oge he suppo he impo ance o MCJ egula ion du ing UC.
The pa hogenesis o in lamma o y bowel disease (IBD) is mul i ac o ial, in ol ing he in e play be ween he
mic obio a, he in es inal luminal en i onmen , he in es inal epi helial ba ie , and bo h he adap i e and inna e
immune esponses. Al hough he ole ha mi ochond ia dys unc ion plays in IBD is no well unde s ood, he e
is a signi ican connec ion be ween in es inal in lamma ion and mi ochond ial unc ion. Heal hy mi ochond ia
a e impo an o adequa e ene gy supply o all cellula p ocesses, p esen ing pa ien s wi h IBD educed ATP
le els wi hin he in es ine1–3. Con e sely, a ia ions in mi ochond ial DNA, which esul in inc eased concen-
a ion o in es inal ATP and augmen ed oxida i e phospho yla ion complex ac i i y, p o ec mice om coli is4.
Fu he mo e, mi ochond ia o IBD pa ien s show mo phological changes5–7 exhibi ing he colonic epi helial cells
o pa ien s wi h ulce a i e coli is (UC) mi ochond ial al e a ions be o e o he ul as uc u al abno mali ies a e
appa en in he epi helium and be o e he onse o mucosal in lamma ion8,9. Reduced mi ochond ial unc ion
o he espi a o y chain complexes II, III and IV (up o 60%) has been epo ed in UC pa ien s10 while pedia ic
C ohn´s disease (CD) pa ien s ha e unc ional de ec s a complexes III and IV11. Recen ly, Habe man e al.12,
obse ed a ma ked supp ession o mi ochond ial complex I ac i i y in ac i e UC pa ien s. Such mi ochond ial
1CIC bioGUNE. Bizkaia Science and Technology Pa k. bld 801 A, 48160, De io, Bizkaia, Spain. 2Facul y o Science
and Technology, Uni e si y o he Basque Coun y (UPV/EHU), Leioa, Bizkaia, Spain. 3Molecula Nu i ion and
Me abolism, Ins i u e o Food Science Resea ch (CIAL, CSIC), Mad id, Spain. 4Depa men o Animal Medicine
and Su ge y, Ve e ina y Facul y, Uni e sidad Complu ense de Mad id, Mad id, Spain. 5Commensal and P obio ics-
Hos In e ac ions Labo a o y, UMR1319 Micalis, INRA, Jouy-en-Josas, F ance. 6CIBERehd, ISCIII, Mad id, Spain.
7Ike basque, Basque Founda ion o Science, Bilbao, Bizkaia, Spain. 8These au ho s con ibu ed equally: Miguel
Angel Pascual-I oiz and Ainize Peña-Cea a. *email: [email p o ec ed]
open
2
Scien i ic RepoR S | (2020) 10:572 | h ps://doi.o g/10.1038/s41598-019-57348-0
www.na u e.com/scien i ic epo s
www.na u e.com/scien i ic epo s/
pe u ba ions esul in inc eased epi helial pe meabili y as a consequence o eac i e oxygen species (ROS)
p oduc ion and p omo e anscy osis o bac e ia ac oss he epi helial laye 13. Highe in es inal pe meabili y
has been epo ed in ac i e14 and UC in emission15 u he sugges ing a ole o mi ochond ial dys unc ion in
he pa hogenesis o IBD. In addi ion, esul s om Mo awea e al.16 indica ed a egula o y ela ionship be ween
mi ochond ia and mic obio a and a dis u bance o his ela ionship in CD pa ien s. Me hyla ion-con olled
J p o ein (MCJ), encoded by he Dnajc15 gene, is a small mi ochond ial p o ein ha nega i ely egula es he
elec on anspo chain (ETC)17,18. Endogenous MCJ associa es wi h complex I and ac s as a na u al inhibi o .
MCJ de iciency esul s in inc eased complex I ac i i y and mi ochond ial memb ane po en ial wi hou a ec ing
mi ochond ial mass17. The ac i i y o complex I is enhanced by i s assembly in o “ espi asomes”, mi ochond ial
ETC supe complexes con aining complexes I, III, and IV19. Supe complexes acili a e he e icien ans e o
elec ons minimizing elec on “leak” ha esul s in ROS p oduc ion20. Loss o MCJ in mac ophages esul s in
inc eased mi ochond ial espi a ion and ele a ed basal le els o ROS. The ac i a ion o he JNK/c-Jun pa hway
also inc eased, leading o he up egula ion o he Tn con e ing enzyme (Tace) inhibi o Timp3, which e ec i ely
p e en s he shedding o Tn ( umo nec osis ac o ) om he memb ane. MCJ egula es he p oduc ion o Tn
by mac ophages in esponse o a a ie y o Toll-like ecep o (TLR) ligands and bac e ia21. MCJ was ini ially
iden i ied as a gene nega i ely egula ed by me hyla ion a CpG islands in o a ian cance 22, Wilms umo s23 and
melanoma24. La e , IFNγ was iden i ied as a ep esso o MCJ ansc ip ion in mac ophages25. Howe e , he ole
ha MCJ plays du ing in es inal in lamma ion is unknown.
In his s udy, we used a MCJ-de icien mu ine model o s udy he ole o he mi ochond ial dys unc ion
in expe imen al coli is. Loss o MCJ esul s in a mo e se e e disease ac i i y index h ough he egula ion o
cy okines. This is i s e lec ed in gu mic obio a composi ion and in es inal pe meabili y and hen impac ed
ia TLR in he p og ession o coli is. The e o e, MCJ plays a p o ec i e unc ion du ing in es inal in lamma ion.
Unde s anding he ole o mi ochond ial modula o MCJ in he pa hogenesis o UC may o e key insigh s in o
he ini ia ion and p opaga ion o he disease.
Ma e ials and Me hods
Animals and expe imen al design. Animal p o ocols we e app o ed by he Animal Resea ch E hics
Boa d o CIC bioGUNE in acco dance wi h Eu opean and Spanish guidelines and egula ions. MCJ-de icien
mice on a C57BL/6 backg ound and wild- ype B6 mice (8–10 wk) we e main ained unde speci ic pa hogen- ee
condi ions wi h con olled empe a u e (21–23 °C) and 12/12-hou ligh /da k cycles. Mice we e ed ad libi um on
s anda d mouse chow (Global die 2914, Ha lam, Madison, USA).
Dex an sodium sul a e (DSS) (36–50 kDa; TdB Consul ancy) was adminis e ed in d inking wa e (3%) o
6 days; hen, mice we e gi en au ocla ed wa e o 2 days. Animal body weigh , he p esence o g oss blood in
eces, and s ool consis ency we e indi idually e alua ed daily by a blind echnician. Each pa ame e was assigned
a sco e acco ding o he c i e ia p oposed p e iously26 and used o calcula e an a e age daily DAI (disease ac i i y
index).
ansepi helial pe meabili y assay. Mice we e ga aged wi h 600 mg kg−1 body weigh o FITC–dex-
an (4 kDa; TdB consul ancy) and whole blood was collec ed by ca diac punc u e 4 h a e ga age. Blood se um
was collec ed a e cen i uga ion a 6000 pm o 10 min. Se um luo escence in ensi y was measu ed using a
mul i-de ec ion mic opla e eade (Spec amax M2, Molecula de ices) wi h an exci a ion wa eleng h o 485 nm
and an emission wa eleng h o 528 nm. FITC concen a ion (mg ml−1) was calcula ed om a s anda d cu e
using se ial dilu ions o FITC–dex an.
Myelope oxidase ac i i y assay. One cen ime e leng h o he dis al colon was homogenized in 50 mM
phospha e bu e (pH 6.0) and 0.5% hexadecyl ime hylammonium b omide using a P ecellys 24 homogenize
(Be in Ins umen s). A e 4 cycles o 90 seconds a 6000 pm, 7 µl o supe na an was mixed wi h 200 µl o 0.02%
dianisidine (Sigma-Ald ich) in 50 mM phospha e bu e , pH 6.0, and 0.0005% H2O2 (Sigma-Ald ich). Human
myelope oxidase (MPO) (Me ck Millipo e, ca numbe 475911) was used as a s anda d o measu e samples’ ac i -
i y. All ac i i y assays we e pe o med in iplica es on 96 well mic o i e pla es and analyzed wi h a mic opla e
eade measu ing abso bance a 450 nm (Spec amax M2, Molecula de ices).
cell p epa a ion. Spleens and mesen e ic lymph nodes we e dissec ed pos -mo em and collec ed in PBS
(Gibco). Fo splenocy e and lymph node cell p epa a ion, o gans we e mashed h ough a 70-μm cell s aine
(Falcon), and e y h ocy es om spleens we e lysed using ACK Lysis Bu e .
Cell analysis by low cy ome y. Cells we e s ained wi h he ollowing luo och ome-conjuga ed an i-
bodies: CD11b APC (Mil enyi Bio ech, M1/70); CD11c PE Cy7 (Mil enyi Bio ech, N418); CD103 PE (Mil enyi
Bio ech, REA789); F4/80 FITC (Mil enyi bio ech, REA126); Fc ecep o s we e blocked wi h An i-mCD16/CD32
(BD). Only e en s ha appea ed single in o wa d-sca e wid h we e analyzed. A FACSCan o II and FACSDi a
so wa e (BD) we e used o low cy ome y and da a we e analyzed using FlowJo so wa e (T eeS a ).
His ology and immunohis ochemis y. Colon issue was ixed in 10% o malin o Ca noy´s ixa i e,
dehyd a ed, embedded in pa a in and cu in o 5 μm- hick sec ions. Fo his opa hology, sec ions we e depa a -
ined, hyd a ed and s ained wi h hema oxylin and eosin o pe iodic acid-Schi (PAS) acco ding o he s anda d
p o ocol. S ained sec ions we e analyzed by a pa hologis blinded o mouse geno ype and ea men . The numbe
o goble cells was de e mined on PAS s ained slides and exp essed as he pe cen age pe in es inal epi helial cells.
Fo immuno luo escen analysis, issue sec ions we e depa a ined, hyd a ed and subjec ed o an igen e ie al
using p o einase K o 15 min a 37 °C o no hing. A e blocking (3% H2O2), sec ions we e incuba ed wi h p i-
ma y an ibodies o 1 h (F4/80, 1:50, Biolegend) o 2 h (Muc2, 1:100, San a C uz), ollowed by 30 min incuba ion
3
Scien i ic RepoR S | (2020) 10:572 | h ps://doi.o g/10.1038/s41598-019-57348-0
www.na u e.com/scien i ic epo s
www.na u e.com/scien i ic epo s/
wi h luo escen ly labelled seconda y an ibody. Finally DAPI was added o s ain nuclei. Pho og aphs we e aken
wi h a luo escence mic oscope (Axioimage .D1 Zeiss) and analyzed by wo people blinded o he ea men . A
leas 10 isual ields we e cap u ed andomly.
Colon se ial sec ions (5 µm) we e subjec ed o immunohis ochemis y (IHQ) wi h p ima y an ibody speci ic
o IgA, HRP conjuga ed (IgA; dilu ion 1:100). An igen e ie al was pe o med incuba ing wi h p o einase K o
15 minu es a 37 °C. A e incuba ion issue sec ions we e imme sed in diaminobenzidine solu ion o 2 min and
washed. Slides we e coun e s ained in Maye ’s hema oxylin o 30 seconds. Images we e cap u ed wi h a Zeiss
Axioimage A1 mic oscope and analyzed wi h F ida so wa e. A leas 10 isual ields we e cap u ed andomly.
ansmission elec on mic oscopy. Sec ions (2–3 mm) om dis al colon we e ixed in 2% glu a aldehyde
in 0.12 M phospha e bu e (PB, pH 7.4), o e nigh a 4 °C. Then, samples we e washed wi h 0.1 M PB (pH 7.4)
and imme sed in 1% OsO4 in 0.1 M PB o e nigh a 4 °C. A e a washing s ep wi h dis illed wa e , samples we e
s ained wi h 0.5% u anyl ace a e du ing 45 minu es a 4 °C. Colon slices we e washed again wi h dis illed wa e ,
dehyd a ed in a g aded e hanol se ies, and embedded in p opylene oxide. A e wa ds, samples we e incuba ed
45 minu es 1:1 p opylene oxide and esin, and embedded in esin o e nigh a oom empe a u e. Finally, samples
we e added o esin blocks and kep o e nigh a 60 °C. The 1 µm semi- hin and 60 nm ul a- hin sec ions we e
ob ained on a Leica Ul acu UCT ul amic o ome. Ul a- hin sec ions we e collec ed in 200-mesh coppe g ids
and obse ed in a JEOL JEM 1400 Plus ETM a 100 kV.
De e mina ion o ROS in colon issue sec ions. Samples we e sec ioned in a c yos a (8 µm) and incu-
ba ed wi h MnTBAP 150 µM o 1 h a RT. The samples we e hen incuba ed wi h DHE (5 µM) o 30 min a 37 °C.
Sec ions we e moun ed wi h moun ing media con aining DAPI. Pho og aphs we e aken wi h a luo escence
mic oscope (Axioimage .D1 Zeiss) and analyzed by wo people blinded o he ea men . A leas 10 isual ields
we e cap u ed andomly.
Mac ophage isola ion om colon. In es inal cells we e ex ac ed using se e al washing s eps and by
enzyma ic dis up ion media ed by collagenase ype IV, a ma ix deg ading enzyme ha b eaks down in e cellula
ma ices as desc ibed by Ha usa o e al.27. MACS LS magne ic column we e used o selec F4/80 posi i e cells.
RNA ex ac ion, cDNA syn hesis, and gene exp ession. Colon samples we e p ese ed in RNAla e
solu ion. To al RNA om colon issue was ex ac ed using RiboZol and Nucleospin RNA ki (Mache ey-Nagel)
acco ding o manu ac u e ’s p o ocol. The cDNA was syn hesized by using M-MLV e e se ansc ip ase
(The mo). Rela i e gene exp ession o a housekeeping gene (Rpl-19) was de e mined by using RT-PCR.
Ampli ica ion and de ec ion was pe o med on op ical g ade 384-well pla es in Quan S udio 6 Flex Real-Time
PCR sys em (The mo Fishe Scien i ic, Wal ham, MA, USA) wi h Pe eCTa qPCR Tough Mix (Quan abio,
Be e ly, MA, USA) and speci ic p ime s a hei annealing empe a u e (see Supplemen a y Table). To no malize
mRNA exp ession, he exp ession o 3 housekeeping genes was measu ed; Rpl-19 was anked as he bes can-
dida e. The mRNA ela i e quan i ica ion was calcula ed using he ΔΔC me hod. PCR e iciency was always
be ween 90 and 110%.
Human RNAseq samples analysis. Human samples we e ob ained om he GEO da ase ela ed o he
s udy o “UC Colon RNAseq subse analysis” eleased as pa o he GSE107593 se ies (h ps://www.ncbi.nlm.nih.
go /gds), a ailable since Ap 17, 2018. Raw ead coun able “GSE107593_ aw_ eads_BCHRNAseq” gene a ed
upon STAR alignmen s was downloaded and used as inpu o he Di e en ial Exp ession (DE) analysis, ca ied
ou by DESeq. 228, aimed o de ec di e en ially exp essed genes.
Ace ac i i y. Colon p o ein ex ac s we e incuba ed wi h 10 µM o TACE FRET Subs a e I (Anaspec,
F emon , CA) in black NUNC polys y ene 96-well mic o i e pla es (Fishe Scien i ic). Colon p o ein ex ac s
we e ea ed wi h he me allop o einase inhibi o TAPI-2 (50 µM; Enzo Li e Sciences, Fa mingdale, NY) o de e -
mine non-speci ic TACE ac i i y. Enzyme ac i i y was moni o ed using a BioTek Syne gy HT mic opla e lu-
o escence eade (BioTek, Winooski, VT) a an exci a ion wa eleng h o 355 nm and an emission wa eleng h
o 500 nm. Resul s a e exp essed as speci ic ac i i y esul ing om sub ac ing nonspeci ic ac i i y om o al
ac i i y.
TNF ELISA. The TNF le els we e de e mined by cap u e ELISA using he DuoSe II ki (R&D Sys ems,
Minneapolis, MN) acco ding o he manu ac u e ’s ecommenda ions.
Wes e n blo . 20 µg o p o ein we e un on SDS-PAGE, ans e ed o ni ocellulose memb anes and es ed
wi h an ibodies speci ic o TIMP3, TNF bound o memb ane and TNFR1. Equal loading was de e mined using
an ibodies agains GAPDH om San a C uz Bio echnology, Dallas, TX).
DNA ex ac ion and mic obiome analysis. Colon con en was collec ed a sac i ice. DNA was isola ed
om eeze-d ied colon samples (40 mg) using he Fa o P ep S ool DNA Isola ion Mini ki (Vienna, Aus ia) by
ollowing manu ac u e ’s ins uc ions. In addi ion, he lysis empe a u e was inc eased o 95 °C. Elu ed DNA was
ea ed wi h RNase and he DNA concen a ion assessed spec opho ome ically by using a NanoD op ND-100
Spec opho ome e (NanoD op Technologies, Wilming on, DE, USA). Pu i ied DNA samples we e s o ed a
−20 °C un il use.
Using high h oughpu sequencing pla o ms and ba coded p ime se s, phylogene ic-based me hods a ge -
ing he 16S RNA gene we e used o deeply cha ac e ize he mic obial popula ions p esen in he colon o expe -
imen al mice. DNA ex ac s we e used as he empla e o PCR-based ampli ica ion o he bac e ial V4 egion o
4
Scien i ic RepoR S | (2020) 10:572 | h ps://doi.o g/10.1038/s41598-019-57348-0
www.na u e.com/scien i ic epo s
www.na u e.com/scien i ic epo s/
he 16S RNA gene using a ba coded py o agging app oach a BGI (Beijing, China) using Illumina Miseq wi h
2 × 250 bp pai ed-end eads based on a s anda d p o ocol om he manu ac u e .
Da a p ocessing was pe o med using QIIME ( .1.9.0): Quan i a i e Insigh s In o Mic obial Ecology so wa e
package29. Sequences we e clus e ed as ope a ional axonomic uni s (OTUs) o 97% simila i y using UCLUST30.
OTU we e checked o chime as using RDP gold da abase and assigned axonomy using he G eengenes da a-
base ( e sion 4 eb2011)31. Richness (numbe o obse ed species) and alpha and be a di e si y me ics (Chao1,
Shannon index, and phylogene ic Di e si y whole ee) we e calcula ed using he QIIME pipeline. The signi -
ican old changes o OTU’s we e pe o med in DESeq. 228. Me aCoMET was used o isualize co e mic obi-
ome32. The signi icances o g ouping in he PCoA plo s we e es ed and analysis o simila i y (ANOSIM) wi h
999 pe mu a ions. Raw sequences we e deposi ed in he Eu opean Nucleo ide A chi e (ENA) unde he p ojec
numbe PRJEB19385.
Me abolomic analysis. Colon con en was collec ed a sac i ice and o each 20 mg o eeze-d ied issue,
300 µL cold me hanol-wa e (2:1, / ) and 200 µL chlo o o m wi h in e nal s anda ds we e sequen ially added
and mixed o 10 s wi h a o ex. Th ee cycles o eezing/ hawing and mechanical homogeniza ion we e pe -
o med. Samples we e cen i uged o 10 min a 3000 g a 10 °C, and he me hanol-wa e uppe phase was s o ed
a −80 °C un il LC-MS analysis. LC-MS was used o b oad me aboli e p o iling33. Me abolomic analyses we e
pe o med in a Wa e s Acqui y UPLC sys em hyphena ed o a B uke maXis II UHR-QTOF mass spec ome e .
Ch oma og aphic sepa a ion was pe o med using an ACQUITY UPLC BEH C18 (2.1 mm × 50 mm, 1.7 μm)
column. Mobile phase A was 0.1% o mic acid in wa e , and he mobile phase was B 0.1% o mic acid in ace oni-
ile. Mass de ec ion was un in he MS scan mode om m/z 20 o 2000 in ESI (+). Samples we e analyzed in
andomized o de and in duplica e. Resul s we e p ocessed wi h MZMine .2.334 and uni a ia e and mul i a ia e
s a is ical analysis was pe o med wi h Me aboAnalys 4.035. Me aboli e en a i e iden i ica ion was pe o med by
he que y o he exac mass o he de ec ed ea u es agains online da abases (HMDB, and Me lin) wi hin a ± 10
ppm mass ange.
S a is ical analysis. G aphPad so wa e was used o s a is ical analysis (G aphPad So wa e, San Diego,
CA, USA). Resul s we e g aphed as box and whiske plo s wi h median, qua iles, and ange. The signi icance was
assessed by wo ways analysis o a iance (ANOVA) ollowed by alse disco e y a e (FDR) pos - es co ec ion.
The signi icance o a esul is shown by an as e isk “*” in he box o indica e DSS ea men (DSS+) e sus con ol
(DSS−). A P alue o less han 0.05 was conside ed signi ican . Spea man´s co ela ion es was used o assess
he ela ionships among gene exp ession and bac e ial composi ion om he colon wall. Only co ela ions wi h
a alue o p < 0.05 we e conside ed signi ican and only co ela ions wi h a alue o p < 0.001 we e ep esen ed.
Resul s
MCJ a enua es he disease ac i i y index o DSS-induced coli is and dec eases colonic issue
damage. To assess he ole o MCJ in he egula ion o in es inal in lamma o y esponses, we s udied a model
o DSS-induced coli is. MCJ-de icien mice showed inc eased disease se e i y han WT mice, as e lec ed by
highe disease ac i i y index (DAI; Fig.1A). In ag eemen wi h p e ious s udies, he en i e colon o DSS- ea ed
mice showed his opa hological changes, wi h loss o c yp egions. S ikingly, MCJ-de icien mice showed highe
his opa hology sco e (18.3 ± 1.29), compa ed o DSS- ea ed WT (14 ± 1.33) mice (p ≤ 0.01) (Fig.1B).
To u he assess he se e i y o coli is, colonic leng h, mucus laye hickness and he numbe o goble cells
we e measu ed in all expe imen al g oups. Bo h DSS- ea ed g oups showed a ma ked educ ion in colon leng h
(Fig.1C), mucus laye hickness (Figs.1D; S1A) and numbe o goble cells (Figs.1E; S1B), and highe pe mea-
bili y o he pa acellula ace FITC-dex an (Figs.1F; S1C), wi h no di e ences be ween WT and MCJ-de icien
mice. Fu he mo e, analysis o ROS measu ed by dihyd oe hidium (DHE) s aining in colon sec ions showed
inc eased le els in bo h geno ypes (Fig.1G). Mo eo e , mi ochond ia shape obse ed by TEM showed ha in
MCJ-de icien mice changed om ubula o ci cula o m wi h in lamma ion (Figs.1H; S1D). These esul s
sugges ha MCJ unc ion helps o empe he issue damage upon colon inju y, suppo ing he hypo hesis
ha an inhibi ion in mi ochond ial espi a o y con ibu es o his opa hological changes du ing coli is disease
agg essi eness.
MCJ inc eases colonic MPO ac i i y and CD11c+CD103+ cells in mesen e ic lymph nodes. We
hen analyzed he cellula composi ion and cell ac i a ion s a us in he p esence o absence o MCJ. Fi s , we
e alua ed he in il a ion and ac i a ion o neu ophils in he local in lamma o y colonic si es by measu ing mye-
lope oxidase (MPO) le els. MPO le els we e signi ican ly inc eased (p ≤ 0.0001) in bo h WT and MCJ KO mice
ea ed wi h DSS. Howe e , DSS- ea ed MCJ KO mice showed (p ≤ 0.05) lowe MPO ac i i y han WT mice
(Fig.2A), sugges ing a educed in il a ion o neu ophils. Mac ophages in il a ed in lamina p op ia and submu-
cosa we e also examined by exp ession o he cell su ace ma ke F4/80 in se ial colon sec ions. The numbe o
F4/80+ cells was signi ican ly highe in he in lamed mucosa o DSS ed mice, compa ed o un ea ed mice and
no ob ious di e ence was obse ed be ween geno ypes (Fig.2B). Then, colon mac ophages we e isola ed om
issue and TNF was de e mined by low cy ome y (Fig.2C). Resul s showed ha TNF bound o memb ane le els
in bo h DSS+ g oups and MCJ de icien DSS- g oup we e highe han WT DSS- one.
We also de e mined cellula popula ions in he spleen and mesen e ic lymph nodes (MLN) by low cy ome y.
We iden i ied an inc eased popula ion o CD11bhigh and F4/80 double posi i e cells (monocy e-de i ed dend i ic
cells: MDDC) in spleen om mice de icien in MCJ ea ed wi h DSS (Fig.2D). In con as , he popula ion o cells
ha we e posi i e o CD11b and CD11c (TNF and ni ic oxide p oducing DC, Tip-DCs) dec eased. In MLN,
he CD11c and CD103 (dend i ic cells) posi i e popula ion did no change as a esul o he ea men wi h DSS,
al hough we no ed a small, albei signi ican dec ease in MCJ KO DSS- ea ed mice compa ed o DSS- ea ed WT
5
Scien i ic RepoR S | (2020) 10:572 | h ps://doi.o g/10.1038/s41598-019-57348-0
www.na u e.com/scien i ic epo s
www.na u e.com/scien i ic epo s/
mice (Fig.2E). O e all, hese esul s showed ha inna e immune popula ions a e no d ama ically changed in he
absence o MCJ du ing coli is.
MCJ a ec s gene exp ession in mu ine colonic issue. In o de o measu e he po en ial e ec o
MCJ on he in lamma o y ou pu in he colonic issue du ing DSS-induced coli is, we quan i ied mRNA le els o
se e al genes by quan i a i e RT-PCR. The exp ession le els o Tl 2, 4, 5 and 9, he p ima y mucosal ecep o s o
bac e ial componen s, we e analyzed. In un ea ed mice, he exp ession le els o Tl 5 and Tl 9 we e no di e en
in WT and MCJ-de icien mice, while Tl 2 and Tl 4 exp ession le els we e signi ican ly lowe in he absence o
MCJ (Fig.3A). The ea men wi h DSS did no esul in signi ican changes in he exp ession le el o Tl 2 o Tl 5
ega dless o he geno ype. On he o he hand, he ea men esul ed in dec eased exp ession o Tl 4 ha was
no a ec ed by he lack o MCJ. Howe e , Tl 9 exp ession le els inc eased in WT mice ha had been ea ed wi h
DSS, and he absence o MCJ esul ed in a u he signi ican inc ease o i s exp ession le els (Fig.3A). Al hough
he exp ession le els o T i did no change along he di e en condi ions and geno ypes, we ound a signi ican
inc ease in he exp ession le els o Myd88 bo h as a consequence o he ea men wi h DSS and he geno ype
o he mice (Fig.3B). O e all, hese da a sugges ha he absence o MCJ implies an inc eased exp ession o key
genes egula ing he in lamma o y ou pu in he colon, including Tl 9-MyD88.
We hen measu ed he cy okine and p oin lamma o y gene exp ession p o ile o colonic issue. The exp ession
le els o Il6, Il1b, P gs2, Nos2, Tn , Tg b and Il10 we e signi ican ly inc eased as a esul o he ea men wi h DSS
bo h in WT and MCJ-de icien mice (Fig.3C–E). The absence o MCJ caused signi ican ly highe exp ession o
genes such as Nos2, Il1b, Tg b, Il10, and Tn . In his ega d, in lamma ion was con i med by he a io IL10/TNF
measu ed by ELISA. The up egula ion o Tn and Il1b genes exp ession in he absence o MCJ we e pa icula ly
in iguing due o i s ole in epi helial ba ie dis up ion du ing IBD. In pa icula , memb ane-bound Tn is asso-
cia ed wi h mo e se e e in lamma o y condi ions when in e ac ing wi h Tn 136. The e o e, we in es iga ed he
exp ession o bo h Adam17 and Timp3. As epo ed in bone ma ow-de i ed mac ophages21, he absence o MCJ
esul ed in signi ican ly inc eased exp ession le els o Timp3, while he le els o Adam17 emained in a iable
Figu e 1. Expe imen al design: WT and MCJ-de icien mice ecei ed 3% DSS s e ile wa e o 6 days,
ollowed by 2 days o egula d inking wa e . (A) DAI alues o e he expe imen al pe iod; da a a e exp essed
as means ± SEM (n = 20); *p < 0.05 e sus DSS con ol g oup. (B) His ological sco es. E o ba s indica e
SEM. Rep esen a i e images. (C) Colon leng h (cm). (D) Mucus laye hickness (µm). (E) Goble cells: % o
posi i e cells s ained wi h PAS (Pe iodic Acid-Schi ). (F) Pe meabili y o he ace FITC-dex an (ng/ml). (G)
Reac i e oxygen species (ROS) measu ed by dihyd oe hidium (DHE) s aining in colon sec ions. (H) Elec on
mic oscopy showing mi ochond ial mo phology in DSS induced coli is g oups (Scale ba : 500 nm). Whi e:
wild- ype; g ey: MCJ-de icien . Box and whiske plo s o median, qua iles and ange, n = 10 mice pe g oup a
a minimum. S a is ical analysis: wo-way ANOVA. “*” in boxes e sus con ol, “*” e sus WT DSS ea ed.
6
Scien i ic RepoR S | (2020) 10:572 | h ps://doi.o g/10.1038/s41598-019-57348-0
www.na u e.com/scien i ic epo s
www.na u e.com/scien i ic epo s/
(Fig.3E). Howe e , TACE ac i i y was signi ican ly educed in MCJ de icien g oup (Fig.3F) indica ing a unc-
ional con ol o he enzyme by MCJ. In acco dance wi h sligh ly lowe le el o soluble TNF measu ed by ELISA
(Fig.3F). In addi ion, he le els o Tn 1 exp ession emained cons an in WT mice when ea ed wi h DSS, as
opposed o he obse ed inc ease in Tn 2 le els. O no e, he absence o MCJ speci ically esul ed in an aug-
men ed exp ession o Tn 1 (Fig.3E). On he o he hand, mRNA abundance o I ng and S a 1 was inc eased only
in he absence o MCJ in DSS ea ed mice (Fig.3G). These da a suppo he augmen ed p oduc ion o Tn and he
signaling h ough Tn 1 in he absence o MCJ. In addi ion, MCJ seems o be implica ed in he egula ion o I ng
and S a 1 exp ession le els du ing in lamma o y p ocesses. Rep esen a i e wes e n blo e lec ing TIMP3, TNFR1
and TNF bound o memb ane p o ein le els in colon we e p esen ed in Fig.S2.
S udies in oden s indica e ha inna e ecogni ion o bac e ia o bac e ial componen s igge s epi helial
exp ession o sec e ed C- ype lec ins Reg3g and Reg3b. As expec ed, he exp ession le el o Reg3b and Reg3g
aised wi h DSS ea men (Fig.3G). In e es ingly he absence o MCJ in coli is-induced mice esul ed in lowe
exp ession o bo h genes, indica ing lowe p o ec i e ole agains in es inal ansloca ion which migh be associ-
a ed wi h he se e i y o he disease.
We also es ed he exp ession o genes ela ed o he in eg i y o he epi helial cell ba ie , by analyzing he
genes Muc2, Muc3, Tjp1, Cldn2, Cldn5, Ocln and Pig . As an icipa ed, he exp ession o Muc2 and Muc3 we e
signi ican ly ep essed upon DSS ea men , bu no e ec was obse ed by he lack o MCJ (Fig.3H,I). No di -
e ences we e obse ed in he exp ession o Tjp1 (encoding he p o ein Zonula Occludens 1). Tigh junc ion
associa ed genes such as Ocln, Cldn2 and Cldn5 (encoding he p o eins Occludin, Claudin 2 and Claudin 5,
espec i ely) did no show di e ences due o DSS ea men in WT mice (Fig.3J). Howe e , he MCJ-de iciency
in DSS- ea ed mice esul ed in he dec eased exp ession o Cldn2 and inc eased le els o Cldn5 ansc ip s in he
colon, sugges ing an e ec on igh junc ion pe meabili y, al hough esul s should be in e p e ed wi h cau ion as
only di e ences in concen a ion and no in localiza ion ha e been de e mined.
In o de o de e mine whe he ou indings in he mu ine model esemble human disease condi ions, we
analyzed a public da ase o UC pa ien s ( o al o 48 samples) o he exp ession le els o MCJ. DNAJC15 exp es-
sion le els we e ound o be signi ican ly lowe in in lamed issue compa ed o non in lamed issue (Fig.3K).
Impo an ly, he exp ession o he genes TIMP3 and TNF was inc eased in in lamed compa ed o non in lamed
issue om UC pa ien s, s ongly sugges ing ha MCJ exp ession and he subsequen egula ion o downs eam
genes a e in ima ely ela ed o he in lamma o y ou pu du ing colonic in lamma ion.
Figu e 2. (A) MPO ac i i y (U/mg p o ein). (B) Mac ophage in il a ion in he colon o DSS- ea ed mice.
Rep esen a i e immuno luo escence sec ions o colonic mac ophages by F4/80 ma ke . Ba ep esen s
50 µm. (C) The a e age o he mean luo escence in ensi y (MFI) de e mined by low cy ome y showing
he exp ession o TNF bound o memb ane p esen in mac ophages isola ed om colon issue. The alues
p esen ed co espond o one expe imen o wo wi h simila esul s. (D,E) Flow cy ome y analysis o lymphoid
popula ions in mesen e ic lymphoid nodes. Whi e: wild- ype; g ey: MCJ-de icien . Box and whiske plo s o
median, qua iles and ange, n = 7 mice pe g oup a a minimum. S a is ical analysis: wo-way ANOVA. “*” in
boxes e sus con ol, “*” e sus WT DSS ea ed.
7
Scien i ic RepoR S | (2020) 10:572 | h ps://doi.o g/10.1038/s41598-019-57348-0
www.na u e.com/scien i ic epo s
www.na u e.com/scien i ic epo s/
MCJ a ec s he composi ion o he hos mic obiome. Sec e o y IgA plays a ole in he homeos a ic
main enance o he in es inal mic obio a, p ima ily by p e en ing mucosal in lamma ion h ough immune exclu-
sion, emo al o an igen-an ibody complexes in he lamina p op ia and neu aliza ion o in lamma o y media-
o s. The e o e, we measu ed IgA le els in he colon wall by immunohis ochemis y. A ma ked inc ease o IgA
was ound in coli is-induced MCJ-de icien mice compa ed o WT animals (Fig.4A), sugges ing ha he absence
o MCJ could shape in es inal mic obio a in he in lamed colon. The e o e, we e alua ed mic obio a composi ion
in animals ha we e induced coli is compa ed o heal hy con ol mice. Illumina sequencing o he V4 egion o
he 16S RNA gene o colonic bac e ial communi ies yielded a o al o 3,161,263 pai ed and me ged sequences
a e quali y il e ing anging om 56,302 o 162,730 sequences pe sample. A Good’s co e age a e age o 96.7%
( ange 95.3–98.2%) indica ed ha he e was su icien communi y co e age using his da ase so ha he e ec s
on he communi y s uc u e o he mic obio a could be assessed. Di e si y indices dec eased a e DSS ea -
men p esen ing wo indices (Obse ed species and Shanon) signi ican ly lowe di e si y in he MCJ-de icien
g oup (Fig.4B). Di e ences we e de ec ed using p incipal componen analysis (PCoA) o non-phylogene ic (B ay
Cu is) and phylogene ic dis ances (weigh ed and unweigh ed Uni ac), showing changes in pa icula OTUs
abundances. The heal h s a us o he mice p o ided he s onges e ec on mic obio a composi ion, ollowed by
he p esence o absence o MCJ (e.g., B ay Cu is dis ances; Fig.4C). ANOSIM de ec ed ha di e ences be ween
coli is-induced g oups we e signi ican (p = 0.01). Co e mic obio a was ep esen ed in a Venn diag am (Fig.4D).
The compa ison o axa p opo ions among expe imen al g oups iden i ied a numbe o hem ha could be
con ibu ing o he di e ences obse ed in he mic obial communi y. In homeos asis, he ela i e abundance o
Figu e 3. Gene exp ession le els in colon issue in a DSS-induced coli is model. mRNA old change no malized
o Rpl19 gene, and o WT (whi e) o MCJ KO (g ey) con ol, espec i ely. (A) Tl 2, Tl 4, Tl 5, and Tl 9. (B)
Myd88 and T i . (C) Il6, Il1b, P gs2, and Nos2. (D) Tg b, and Il10. (E) Tn , Adam17, Timp3, Tn 1, and Tn 2.
(F) Tace ac i i y and TNF concen a ion (G) I ng, and S a 1. (H) Reg3b, and Reg3g we e de e mined by eal-
ime qPCR. Resul s a e old change no malized o Rpl19 gene, and o WT (whi e) o MCJ KO (g ey) con ol,
espec i ely. (I) Muc2, Muc3, Tjp1, and Pig . (J) Cldn2, Cldn5, and Ocln. Box and whiske plo s o median,
qua iles, and ange, n = 7 mice pe g oup a a minimum. S a is ical analysis: wo-way ANOVA. “*” in boxes
e sus con ol, “*” e sus WT DSS ea ed. (K) Gene exp ession le els o MCJ, TIMP3,and TNF in a public
human da ase (H = non in lamed; UC = in lamed issue). Sample size: H = 24; UC = 24.
8
Scien i ic RepoR S | (2020) 10:572 | h ps://doi.o g/10.1038/s41598-019-57348-0
www.na u e.com/scien i ic epo s
www.na u e.com/scien i ic epo s/
he phylum Ve ucomic obia was signi ican ly highe (p = 0.008) in MCJ-de icien mice (7.1 ± 0.09%) compa ed
o WT animals (0.05 ± 0.0001%) (n = 8 pe geno ype) (Fig.4E). The examina ion a he genus le el showed ha
wi hin he phylum Ve ucomic obia only Akke mansia muciniphila was ep esen ed. In addi ion, Ruminococcus
and a small g oup belonging o he Lachnospi aceae amily ha has ye o be assigned o a genus, we e signi i-
can ly inc eased in heal hy MCJ-de icien mice (Fig.4F). In MCJ-de icien coli is-induced mice, he analysis indi-
ca ed a educ ion in Bac e oide es phylum based on he dec eased le els o unclassi ied genus om Bac e oide es
S24–7 amily, Bac e oides and P e o ella, and he inc ease in he p esence o membe s o he Fi micu es phy-
lum such as Lac obacillus and Ruminococcus gna us (Fig.4G). The En e obac e iaceae amily enla ged i s pe -
cen age om 10 in WT o 20% in MCJ-de icien mice. The disp opo iona e inc ease in mucoly ic bac e ia in
MCJ-de icien mice could con ibu e o bac e ial ansloca ion and se e i y o he disease.
We hen analyzed he unc ionali y o he mic obiome p esen in he colon o all expe imen al g oups by
a me abolomics app oach using LC-MS. PCA analysis o he da a showed a di e ence be ween samples om
mice as a esul o he ea men wi h DSS (Fig.5A) wi h wo clea ly di e en ia ed clus e s: non- ea ed and
DSS- ea ed mice. Supe ised pa ial leas -squa es disc imina e analysis (PLS-DA) was applied o u he di -
e en ia e he con ibu ions o pa icula me aboli es o he sepa a ions be ween me aboli e le els om WT
and MCJ-de icien mice du ing DSS-induced coli is. Those me aboli es wi h a iable impo ance in he p o-
jec ion (VIP) > 1 we e conside ed as po en ial bioma ke candida es o g oup disc imina ion. The lis o iden-
i ied me aboli es ( old changes o e 1.7) was en e ed in o he pa hway analysis module om Me aboAnalys .
Figu e 4. (A) De e mina ion o IgA le els by immunohis ochemis y in colon issue om WT and MCJ-
de icien mice du ing DSS-induced coli is. E o ba s indica e SEM. Box and whiske plo s o median, qua iles
and ange, n = 7 mice pe g oup a a minimum. S a is ical analysis: wo-way ANOVA. “*” in boxes e sus
con ol, “*” e sus WT DSS ea ed. (B) Alpha di e si y measu es o colon mic obiomes ac oss di e en
expe imen al g oups: To al obse ed axonomic uni s, Chao1 es ima es, and Shannon di e si y index. (C)
PCoA o β-di e si y compa ison e ealed a signi ican sepa a ion o mic obial communi y based on geno ype
in DSS expe imen al coli is. P = 0.01 using ANOSIM. (D) Venn diag am wi h OTUs in he colon mic obiome
wi hin each expe imen al g oup. (E) Phylum le el mic obial composi ion. (F) OTUs signi ican ly di e ence
be ween WT and MCJ-de icien con ol g oups a he genus le el. (G) Ope a ional axonomic uni s signi ican ly
di e en (q < 0.05 FDR) be ween he colon con en om WT and MCJ-de icien DSS-induced coli is g oups.
The le side ep esen s OTU’s wi h a log2 old posi i e di e ence o WT colon con en s ela i e o he MCJ-
de icien while he igh side is he nega i e old change o he WT colon ela i e o he MCJ-de icien con en s.
Each poin ep esen s a single OTU colo ed by phylum and g ouped by axonomic amily o genus le el, size o
poin e lec s he log2 mean abundance o he sequence da a.
9
Scien i ic RepoR S | (2020) 10:572 | h ps://doi.o g/10.1038/s41598-019-57348-0
www.na u e.com/scien i ic epo s
www.na u e.com/scien i ic epo s/
The esul s e lec ed signi ican al e a ion in bile acid me abolism in he diges i e ac (Fig.5B). In e es ingly,
MCJ-de icien mice we e cha ac e ized by he highe p esence o bile acids in he in lamed mucosa. Mo e speci -
ically, au ocholic acid and au ohyochola e (Fig.5C) we e inc eased in MCJ-de icien mice unde in lamma o y
condi ions. High bile acid concen a ion in he colon leads o aqueous s ool and e a ic bowel unc ion, exace -
ba ing dia hea (s ool consis ency) which con ibu es o DAI. Mo eo e , highe le els o ecal bile acids ha e been
iden i ied o wo sen he se e i y o coli is in DSS- ea ed mice37.
In summa y, hese da a show signi ican shi s in he mic obio a o WT and MCJ-de icien mice unde in lam-
ma o y condi ions, sugges ing a egula o y ela ionship be ween mi ochond ial unc ion and mic obio a, and a
dis u bance o his ela ionship in coli is. Mo eo e , ou da a suppo an associa ion be ween he me abolism o
bile acids, mic obio a and in lamma ion.
Co ela ion be ween bac e ial communi y and gene exp ession in he colon unde
expe imen ally-induced coli is. In o de o unde s and whe he he obse ed changes in mic obio a com-
posi ion we e ela ed o he esponse o colonic damage o he absence o MCJ, we pe o med Pea son co ela ion
analyses. The a ia ion in bac e ial communi y composi ion and gene exp ession om colon wall in e ac ion
showed a signi ican co ela ion (Fig.6). The p esence o Akke mansia co ela ed (p ≤ 0.001) wi h I nγ exp ession
while Pa abac e oides co ela ed wi h he genes di e en ially exp essed in he absence o MCJ (Adam17, Timp3
and Tn ecep o s). E en mo e impo an was he posi i e co ela ion obse ed be ween En e ococcus and he
En e obac e iaceae bac e ial g oup wi h Myd88 and Tl 9 exp ession. Those bac e ial g oups we e also co ela ed
wi h he exp ession o cy okines such as Il1b, Il6, Tg b, and Il10, as well as o P gs2 and Nos2 exp ession indi-
ca ing a c i ical ole du ing in lamma ion especially in MCJ-de icien mice. On he o he hand, he p esence o
one unclassi ied genus belonging o he Rikenellaceae and S24-7 amilies showed a nega i e co ela ion wi h he
exp ession o genes associa ed wi h an igen p esen ing cells, poin ing o a po en ial bene icial ole o bo h g oups
o bac e ia du ing in lamma o y condi ions.
Discussion
Mi ochond ial dys unc ion may play an impo an ole in he pa hogenesis associa ed wi h UC. Mi ochond ia a e
he main sou ce o ATP (adenosine iphospha e) gene a ed h ough oxida i e phospho yla ion in he mi ochon-
d ial espi a o y chain. MCJ is an endogenous nega i e egula o o he ETC ha egula es mi ochond ial unc-
ion in esponse o al e ed me abolic condi ions. He e, we show ha he loss o MCJ esul s in mo e se e e coli is,
sugges ing ha ETC unc ion needs o be igh ly con olled o egula e he pa hological consequences upon he
ini ia ion o in lamma ion. We also show ha he absence o MCJ esul s in changes upon he ini ia ion o colonic
damage, which may be ela ed o a ia ions in in es inal mic obio a-hos mi ochond ia c oss alk and he ensuing
exp ession le els o key genes associa ed wi h he ini ia ion and p og ession o he disease.
We ha e iden i ied a pa hway coo dina ed by MCJ a gu le el. MCJ absence du ing coli is esul s in he up eg-
ula ion o he Tace inhibi o Timp3, which inhibi s TACE ac i i y p obably a ec ing Tn and Tn 1 shedding om
he cell memb ane. Indeed, we ha e epo ed ha he loss o MCJ in mac ophages inhibi s Tn shedding om
he plasma memb ane21. Simila ly, bo h TNF ecep o s a e TACE subs a es38. In addi ion, ansmemb ane TNF
induces he o ma ion o STAT1-dependen dea h-inducing signaling complexes (DISC) and apop osis h ough
i s in e ac ion wi h TNFR139. These da a sugges ha MCJ de iciency con ibu es o disease se e i y a leas pa ly
due o he egula ion o memb ane TNF and i s abili y o signal h ough TNFR1. Mo e impo an ly was ha ou
Figu e 5. (A) Sco e plo o PCoA model wi h PC1 plo ed agains PC2 wi h 95% con idence ellipses a ound
he expe imen al g oups (KO DSS-, KO DSS+, WT DSS- and WT DSS+). (B) Me abolic pa hway analysis. The
me abolic pa hways a e ep esen ed as ci cles acco ding o hei p alues om en ichmen analysis (Y-axis) and
pa hway impac alues (X-axis) using Me aboAnalys 4.0. Da ke ci cle colo s indica e mo e signi ican changes
o me aboli es in he co esponding pa hway (p alues). The size o he ci cle co esponds o he pa hway
impac sco e and is co ela ed wi h he cen ali y o he in ol ed me aboli es. (C) Boxplo s o he abundances o
impo an bile acids signi ican ly di e ed among he g oups. The cen e line o each box ep esen s he median,
and he op and bo om o he box ep esen he 75 h and 25 h pe cen iles o he da a, espec i ely. The op and
bo om o he e o ba s ep esen he 95 h and 5 h pe cen iles o he da a, espec i ely. n = 8 mice pe g oup.
