RESEARCH ARTICLE
A e in i o and in silico app oaches used
app op ia ely o animal-based majo
dep essi e diso de esea ch?
Cons anc¸a Ca alhoID
1
*, Susana A. M. Va ela
2¤
, Tiago A. Ma ques
3,4
, And ew Knigh
5
,
Luı
´s Vicen e
1
1Faculdade de Ciências, Cen o de Filoso ia das Ciências, Uni e sidade de Lisboa, Lisboa, Po ugal,
2Faculdade de Ciências, cE3c – Cen e o Ecology, E olu ion and En i onmen al Changes, Uni e sidade
de Lisboa, Lisboa, Po ugal, 3Cen e o Resea ch in o Ecological and En i onmen al Modelling, Uni e si y
o S And ews, S And ews, Sco land, 4Depa amen o de Biologia Animal, Faculdade de Ciências, Cen o de
Es a ı
´s ica e Aplicac¸ões, Uni e sidade de Lisboa, Lisboa, Po ugal, 5Cen e o Animal Wel a e, Uni e si y o
Winches e , Winches e , Uni ed Kingdom
¤Cu en add ess: Ins i u o Gulbenkian de Ciência, Oei as, Po ugal
*cons anca.ca [email protected]
Abs ac
The cu en pa adigm o biomedical esea ch and d ug es ing pos ula es ha in i o and in
silico da a in o m animal s udies ha will subsequen ly in o m human s udies. Recen e i-
dence poin s ou ha animal s udies ha e made a poo con ibu ion o cu en knowledge o
Majo Dep essi e Diso de , whe eas he con ibu ion o in i o and in silico s udies o animal
s udies- wi hin his esea ch a ea- is ye o be p ope ly quan i ied. This quan i ica ion is
impo an since biomedical esea ch and d ug disco e y and de elopmen includes wo
s eps o knowledge ans e abili y and we need o e alua e he e ec i eness o bo h in o de
o p ope ly implemen 3R p inciples (Replacemen , Reduc ion and Re inemen ). He e, we
used he ci a ion acking acili y wi hin Web o Science o loca e ci a ions o o iginal
esea ch pape s on in i o and in silico ela ed o MDD published iden i ied in PubMed by
ele an sea ch e ms. 67 publica ions desc ibing a ge pape s we e loca ed. Bo h in i o
and in silico pape s a e mo e ci ed by human medical pape s han by animal pape s. The
esul s sugges ha , a leas conce ning MDD esea ch, he cu en wo s eps o knowledge
ans e abili y a e no being ollowed, indica ing a poo compliance wi h he 3R p inciples.
1. In oduc ion
Biomedical esea ch hea ily elies on animal s udies, despi e he e hical and clinical limi a ions
o hese [1].
The s anda d con empo a y pa adigm o biomedical esea ch, and d ug disco e y and
de elopmen , equi es scien is s o es pu a i e new clinical in e en ions, by p og essing
om simple o inc easingly complex models, p io o conduc ing human s udies and ials, as
shown in Fig 1.
PLOS ONE
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0233954 June 24, 2020 1 / 6
a1111111111
a1111111111
a1111111111
a1111111111
a1111111111
OPEN ACCESS
Ci a ion: Ca alho C, Va ela SAM, Ma ques TA,
Knigh A, Vicen e L (2020) A e in i o and in silico
app oaches used app op ia ely o animal-based
majo dep essi e diso de esea ch? PLoS ONE 15
(6): e0233954. h ps://doi.o g/10.1371/jou nal.
pone.0233954
Edi o : Jyo shna Kanungo, Na ional Cen e o
Toxicological Resea ch, UNITED STATES
Recei ed: Feb ua y 28, 2020
Accep ed: May 11, 2020
Published: June 24, 2020
Copy igh : ©2020 Ca alho e al. This is an open
access a icle dis ibu ed unde he e ms o he
C ea i e Commons A ibu ion License, which
pe mi s un es ic ed use, dis ibu ion, and
ep oduc ion in any medium, p o ided he o iginal
au ho and sou ce a e c edi ed.
Da a A ailabili y S a emen : All analysed pape s
a e a ailable a h ps://jou nals.sagepub.com/doi/
pd /10.1177/0261192919885578.
Funding: This s udy was suppo ed by Animal ee
Resea ch–Swi ze land in he o m o unds
awa ded o CC. Cen o de Es a ı
´s ica e Aplicac¸ões
(CEAUL) p o ided pa ial suppo o TAM, h ough
p ojec unding om Fundac¸ão pa a a Ciência e a
Tecnologia, Po ugal (FCT) (UID/MAT/00006/
2019). The publica ion ee was inanced by
Po uguese na ional unds awa ded o LV wi hin
E en hough his pa adigm is mo e ocused owa d d ug disco e y, i is also encou aged o
b oade esea ch, by legisla ion and guidelines pe aining o animal esea ch, in a ious coun-
ies and egions (e.g.[3]).
Suppo e s o animal s udies wi hin biomedical esea ch claim ha 1) i is no possible o
discon inue hei use, as ha would jeopa dize human heal h, and ha 2) human-based me h-
ods (in silico and in i o) a e used in ea ly s eps o biomedical esea ch o in o m he animal
esea ch communi y, hence a oiding unnecessa y o excessi e use o animals. Fo example,
pu po edly, i a subs ance shows high le els o oxici y in i o i will no p og ess in o animal
es ing [4]. In he same way, a d ug ha shows high oxici y in animal es ing should no p o-
ceed o human ials. Howe e , i has been demons a ed ha human ials may some imes
occu simul aneously wi h animal ials, a he han sequen ially, as one would expec i animal
ials we e an essen ial s ep p io o human ials [5].
In ou p e ious s udy we compa ed he numbe o ci a ions in i o,in silico and non-
human p ima e-based (NHP) o iginal s udies ocused on Majo Dep essi e Diso de (MDD),
ha we e ecei ed (i) in o al, (ii) by unspeci ied human medical pape s, and (iii) by human
medical pape s ocused on MDD. We e i ied ha bo h in i o and in silico esea ch pape s
ecei ed mo e ci a ions by human medical pape s, han NHP pape s. This was unexpec ed,
conside ing ha mos coun ies es ic he use o NHPs, making i easonable o p esume ha
when hey we e used, hey should p o ide a signi ican con ibu ion o human heal h. How-
e e , his was no he case. Da a ob ained ia simple models (in i o and in silico) seemed o
be mo e isible o conside ed mo e impo an by he human medical esea ch communi y.
This called in o ques ion he con empo a y pa adigm o biomedical esea ch and d ug disco -
e y, in which knowledge is p esumed o ans e be ween animal and human models [6].
Conside ing ha his pa adigm p esumes wo s eps o knowledge ans e abili y: i) be ween
simple and complex models, and ii) be ween animals models and humans, we wonde ed i
he e could be knowledge ans e abili y p oblems in s ep (i), simila o hose we demons a ed
a s ep (ii).
Hence, he aim o he cu en s udy is o assess whe he in i o and in silico pape s desc ib-
ing o iginal da a on a human diso de (MDD) a e being app op ia ely ci ed by subsequen ani-
mal-based pape s. I is impo an o men ion ha animal models a e ex ensi ely used in MDD
esea ch. In ac , by he ime ou s udy was conduc ed he e we e abou wice as many o iginal
pape s using animal models in MDD esea ch han pape s using in i o and in silico
app oaches.
Du ing s udies ocused on MDD, animals equen ly unde go se e e p ocedu es such as
lea ned helplessness o o ced swim es p o ocols. Mos applicable legisla ions and guidelines
manda e ha such p ocedu es should be a oided whe e e possible. Hence i is easonable o
expec ha he MDD- ocused animal esea ch communi y should be pa icula ly ale o he
da a and insigh s p o ided by simple da a.
E en hough he e is a wide consensus ha he use o simple models such as in i o and in
silico me hods wi hin basic and applied biomedical esea ch helps animal esea che s o mee
he p inciples o Replacemen (o animals wi h al e na i es) and Reduc ion (o animal num-
be s), as desc ibed by Russell & Bu ch [7], o ou knowledge, he e has ne e been a sys ema ic
Fig 1. Cu en pa adigm o biomedical esea ch and d ug disco e y and de elopmen . Kindly p o ided by Taylo [2].
h ps://doi.o g/10.1371/jou nal.pone.0233954.g001
PLOS ONE
In i o and in silico use in animal-based majo dep essi e diso de esea ch
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0233954 June 24, 2020 2 / 6
he Cen o de Filoso ia das Ciência da Uni e sidade
de Lisboa (CFCUL) esea ch cen e s a egic
p ojec , unded by Fundac¸ão pa a a Ciência e a
Tecnologia (FCT) (UIDB/00678/2020). These
unde s played no ole in s udy concep ion, design,
da a collec ion and analysis, manusc ip au ho ship
o he decision o submi o publica ion.
Compe ing in e es s: No au ho s ha e compe ing
in e es s.
s udy ha empi ically e i ies whe he animal esea che s a e, indeed, applying his p inciples
o hei p ac ice i.e. i hey a e loca ing and using applicable da a ob ained ia such simple
models.
I in i o and in silico s udies a e indeed seen as an impo an s ep p io o conduc ing ani-
mal s udies in biomedical esea ch, and animal s udies a e in u n seen as impo an p io o
conduc ing human s udies, hen we would expec ha pape s desc ibing in i o o in silico
da a on a human diso de should be ci ed mo e equen ly by animal pape s, han by human
medical pape s. I , on he con a y, his is no he case, hen u he s udies on o he human
diso de s and d ug de elopmen should be conduc ed o con i m he ex en o which he con-
empo a y heo e ical pa adigm o biomedical esea ch is ac ually being ollowed in p ac ice.
I adhe ence is no as common as belie ed, hen his pa adigm should clea ly be e ised.
2. Me hods
We conduc ed a ci a ion analysis as de ined by Ga ield and Me on [8]. Concisely, in a ci a-
ion analysis, a ge pape s a e loca ed i s and hen a sea ch o all o he pape s ci ing he o -
me is pe o med.
The in o ma ion compiled comp ises he o al numbe o ci a ions, and he pa e ns o ci a-
ion. We used a o al o 67 a ge pape s o in i o o in silico s udies on MDD- u ilising only
human da a, selec ed om he ci a ion analysis da abase c ea ed in ou p e ious s udy [6]. The
ci a ion analysis was pe o med be ween Sep embe 2016 and June 2017. We conside ed all
published pape s using in i o o in silico me hods, ha aimed o gain knowledge abou
MDD, and we e published p io o 2011, o enable i e-yea ime o ci a ions–a equen ly
used imeline o ci a ion analysis [9]. To loca e a ge pape s we sea ched PubMed– he la ges
eely accessible bibliog aphic da abase, using he ollowing Medical Subjec Heading (MeSH)
sea ch e ms: ‘Dep essi e Diso de , Majo ’ AND (“in silico” OR ‘compu e model’ OR ‘ma he-
ma ical model’ OR ‘compu e simula ion’ OR ‘in i o’ OR ‘cell cul u e’ OR ‘cul u e echnique’
OR ‘cell line’ OR ‘o gan cul u e’ OR ‘ issue cul u e’. Ou goal was o selec o iginal publica ions
ha p esen ed new da a, so we used PubMed il e s o exclude e iew a icles (“ e iew”, “sys-
ema ic e iew”, “me a-analysis”, “bibliog aphy”) as well as opinion a icles (“biog aphy”,
“au obiog aphy”, “commen ”, “edi o ial”, “in e iew”). We also excluded by hand in i o
pape s ha used animal issue o cells. Using he ci a ion acking acili y wi hin Web o Sci-
ence, we coun ed he numbe o imes each a ge pape was ci ed by subsequen pape s in he
ollowing ca ego ies: ‘animal esea ch pape s‘, ‘human medical pape s‘, ‘in i o pape s‘, and
‘in silico pape s‘. Ci ing pape s may ha e been assigned o mo e han one ca ego y i hey
desc ibed di e en esea ch app oaches (e.g. human-based and in i o).
3. Resul s
In o al, 464 (18%) o he 2,574 ci a ions ecei ed by he 38 in i o pape s we e by in asi e ani-
mal esea ch pape s, and 978 (40%) we e by human medical pape s. Fo he 29 in silico pape s,
44 (5%) o he 806 ci a ions we e by in asi e animal esea ch pape s, and 317 (39%) by human
medical pape s.
As shown in Fig 2, he majo i y o ci a ions ecei ed by bo h in i o o in silico a ge pape s
we e by pape s employing he same esea ch me hod, and by human medical pape s. The p o-
po ion o ci a ions by animal pape s and he o he esea ch me hod we e conside ably lowe .
Mo e impo an ly, he p opo ion o ci a ions by animal pape s was lowe han by human
medical pape s.
PLOS ONE
In i o and in silico use in animal-based majo dep essi e diso de esea ch
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0233954 June 24, 2020 3 / 6
4. Discussion
The esul s o ou ci a ion analysis sugges ha he s anda d app oach o es ing medical
hypo heses–which pos ula es ha in i o and in silico esea ch is an impo an s ep p io o
conduc ing animal es ing–is no suppo ed by ci a ion da a, a leas o MDD esea ch.
Clea ly, MDD biomedical esea ch u ilising in i o and in silico da a does no seem o be con-
side ed impo an by, o a leas mo e impo an o, he animal esea ch communi y, han i is
o he human medical communi y.
One can a gue ha i he animal esea ch communi y is no ci ing in i o and in silico
pape s on MDD, hese migh be o limi ed use. Howe e , ha is inconsis en wi h hei sub-
s an ial use by he human medical communi y, which ci es mo e his kind o esea ch han
esea ch based on animal s udies [6]. Addi ionally, his lack o ans e abili y o knowledge
be ween he animal and he human medical esea ch communi ies is u he e idenced by he
ac ha , in gene al, mos ci a ions ecei ed by animal esea ch pape s a e wi hin o he ani-
mal-based s udies, a he han wi hin human medical pape s [10].
MDD is a complex human men al diso de wi h mul i ac o ial ae iopa hogenesis [11], so
one canno ex apola e ha he ci a ion pa e ns ound he e will necessa ily be eplica ed in
o he diso de s ha ha e jus one cause (e.g. Down’s synd ome). Fu he mo e, a single disease
analysis is no enough o gene alize he esul s o he en i e ield o biomedical esea ch.
Hence, he nex s ep should be he use o a simila app oach a ge ing mono ac o ial diso -
de s and d ug ials. I , as whole, hese s udies p oduce simila esul s, hen i would be com-
pelling e idence ha he accep ed pa adigm o biomedical esea ch and d ug disco e y and
Fig 2. Boxplo s o he p opo ion o ci a ions ecei ed by esea ch ca ego y o in i o A) and in silico B) pape s on
MDD.
h ps://doi.o g/10.1371/jou nal.pone.0233954.g002
PLOS ONE
In i o and in silico use in animal-based majo dep essi e diso de esea ch
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0233954 June 24, 2020 4 / 6
de elopmen is no being su icien ly ollowed, which suppo s he claims made by se e al
au ho s [1] ha he 3Rs a e no being add essed as well as equi ed by applicable legisla ion
and good esea ch p ac ice. This sugges s ha animal s udies in biomedical esea ch a e mos ly
de ining hei esea ch p io i ies au onomously, a he han being pe ec ly amed in he bio-
medical esea ch pa adigm.
Six y yea s ago Russell and Bu ch [7] es ablished he ounda ions o much cu en legisla-
ion ega ding animal expe imen a ion, wi h he o mula ion o he 3R p inciples. E en
hough he esea ch communi y unanimously welcomes hem, he ocus o hei applica ion
has p edominan ly been e inemen , and no always in an e ec i e way [1].
Nowadays he e is an inc easing numbe o da abases on human and animal p o ein
exp ession di e ences ( o a e iew see [12]) which, on he one hand, makes i easie o
esea che s o loca e and ci e exis ing da a; bu , on he o he hand, migh s imula e animal
esea ch o be conduc ed independen ly o in i o and in silico da a o popula e such
da abases.
In heo y, he educ ion p inciple depends upon he s anda d use o in silico and in i o
echniques p io o animal s udies. I o iginal da a on human diso de s om in i o and in sil-
ico app oaches a e no being used by he animal esea ch communi y, hen he educ ion p in-
ciple is no being p ope ly ul illed. The easons behind his mus su ely be mul iple.
One o he possible easons is he inadequacy o sys ema ic e iews ha animal esea che s
some imes pe o m on hei esea ch opic, p io o conduc ing animal expe imen s. These
should p e en unnecessa y animal use [13], bu by excluding om he sea ch in i o and in
silico s udies, esea che s can exclude an impo an sou ce o knowledge.
Based on ou esul s we ecommend ha changes a e made in cu en sys ema ic e iew
p o ocols in o de o include in i o and in silico da a.
Ano he eason ha became salien wi h ou s udy and dese es a en ion, is ha in i o
and in silico app oaches a e, by de ini ion, human-based me hods, no animal-based me hods.
Concei ably human da a is no ele an enough o animal pape s, in he same way animal
s udies do no seem o be ele an o subsequen human s udies [6,10].
This highligh s ha he cu en pa adigm o biomedical esea ch and d ug disco e y and
de elopmen includes wo s eps o knowledge ans e abili y be ween he animal and he
human models, nei he o which appea o wo k well. I simila esul s a e ound in o he dis-
o de s and mo e impo an ly, in d ug disco e y, han he cu en pa adigm mus be changed.
Speci ically, animal es ing mus be dep io i ized, wi h g ea e in es men in human-based in
i o and in silico esea ch app oaches.
Suppo ing in o ma ion
S1 File.
(ODS)
Au ho Con ibu ions
Concep ualiza ion: Cons anc¸a Ca alho, Susana A. M. Va ela.
Da a cu a ion: Cons anc¸a Ca alho, Tiago A. Ma ques.
Fo mal analysis: Cons anc¸a Ca alho, Tiago A. Ma ques.
In es iga ion: Cons anc¸a Ca alho.
Me hodology: Cons anc¸a Ca alho, And ew Knigh .
PLOS ONE
In i o and in silico use in animal-based majo dep essi e diso de esea ch
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0233954 June 24, 2020 5 / 6

Supe ision: Tiago A. Ma ques, And ew Knigh , Luı
´s Vicen e.
Valida ion: And ew Knigh .
W i ing – o iginal d a : Cons anc¸a Ca alho, Susana A. M. Va ela.
W i ing – e iew & edi ing: Cons anc¸a Ca alho, Susana A. M. Va ela, Tiago A. Ma ques,
And ew Knigh .
Re e ences
1. He mann K. Re inemen on he way owa ds eplacemen : A e we doing wha we can? In He man K,
Jane K, edi o s. Animal Expe imen a ion: Wo king Towa ds a Pa adigm Change. Bos on: B ill;
2019. pp. 1–64. h ps://doi.o g/10.1163/9789004391192
2. Taylo K. Recen de elopmen s in al e na i es o animal es ing. In He man K, Jane K, edi o s, Animal
Expe imen a ion: Wo king Towa ds a Pa adigm Change. Bos on: B ill; 2019. pp. 583–609. h ps://doi.
o g/10.1163/9789004391192_002
3. Wo kman P, Aboagye EO, Balkwill F, Balmain A, B ude G, Chaplin DJ, e al. Guidelines o he wel a e
and use o animals in cance esea ch. B i ish Jou nal o Cance . 2010; 102(11): 1555–1577. h ps://
doi.o g/10.1038/sj.bjc.6605642 PMID: 20502460
4. Choudhu i S, Pa on GW, Chande bhan RF, Ma ia A, Klaassen CD. F om classical oxicology o
Tox21: Some c i ical concep ual and echnological ad ances in he molecula unde s anding o he
oxic esponse beginning om he las qua e o he 20 h cen u y. Toxicological Sciences. 2017;
161(1): 5–22. h ps://doi.o g/10.1093/ oxsci/k x186 PMID: 28973688
5. Pound P, Eb ahim S, Sande cock S, B acken MB, Robe s I. Whe e is he e idence ha animal
esea ch bene i s humans? BMJ. 2004; 328(7438): 514–517. h ps://doi.o g/10.1136/bmj.328.7438.
514 PMID: 14988196
6. Ca alho C, Va ela SAM, Bas os LF, O
´ ão I, Beja V, Sapage M, e al. The Rele ance o In silico,In
i o and Non-human P ima e based App oaches o Clinical Resea ch on Majo Dep essi e Diso de .
ATLA. 2019a; 47 (3–4): 128–139. h ps://doi.o g/10.1177/0261192919885578 PMID: 31838868
7. Russell WMS, Bu ch RL. The p inciples o humane expe imen al echnique. London: Me huen; 1959.
8. Ga ield E, Me on RK. Ci a ion indexing: I s heo y and applica ion in science, echnology, and humani-
ies (Vol. 8). New Yo k: Wiley; 1979.
9. Wooding S, Polli A, Cas le-Cla ke S, Coch ane G, Diepe een S, e al. Men al Heal h Re osigh :
Unde s anding he e u ns om esea ch (lessons om schizoph enia): policy epo . Rand heal h qua -
e ly 2014. 4(1). A ailable e h ps://www. and.o g/pubs/ esea ch_ epo s/RR325.h ml.
10. Ca alho C, Al es D, Knigh A, Vicen e L. Is animal-based biomedical esea ch being used in i s o iginal
con ex ? In: He man K, Jane K, edi o s. Animal Expe imen a ion: Wo king Towa ds a Pa adigm
Change. Bos on: B ill; 2019b. pp. 376–390. h ps://doi.o g/10.1163/9789004391192_017
11. Chi iţă AL, Gheo man V, Bonda i D, Rogo eanu I. Cu en unde s anding o he neu obiology o majo
dep essi e diso de . Rom J Mo phol Emb yol. 2015;. 56(2 Suppl): 651–658. PMID: 26429155
12. Yin JY, Sun W, Li FC, Hong JJ, Li XX, e al. VARIDT 1.0: Va iabili y o D ug T anspo e Da abase.
Nucleic Acids Resea ch. 2020; 48(D1): D1042–D1050. h ps://doi.o g/10.1093/na /gkz779 PMID:
31495872
13. Leenaa s M, Hooijmans CR, an Veggel N, Te Rie G, Lee lang M, Hoo L, e al. A s ep-by-s ep guide
o sys ema ically iden i y all ele an animal s udies. Labo a o y Animals. 2012; 46(1): 24–31. h ps://
doi.o g/10.1258/la.2011.011087 PMID: 22037056
PLOS ONE
In i o and in silico use in animal-based majo dep essi e diso de esea ch
PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0233954 June 24, 2020 6 / 6