Patient-derived xenografts (PDXs) as model systems for human cancer

Patient-derived

xenografts

(PDXs)

as

model

systems

for

human

cancer

Federica

Invrea

,

Roberta

Rovito

,

Erica

Torchiaro

,

Consalvo

Petti

,

Claudio

Isella

and

Enzo

Medico

Patient-derivedxenografts(PDXs)areobtainedby transplantingfragmentsofapatient’stumourinto immunodeficientmice.GrowthandpropagationofPDXs allowscorrelatingtherapeuticresponseinvivowithextensive, multi-dimensionalmolecularannotation,leadingto

identificationofpredictivebiomarkers.PDXsareincreasingly recognisedasclinicallyrelevantmodelsofcancerforseveral reasons,ofwhichthemainisthepossibilityofstudyingthe behaviourofcancercellsinanaturalmicroenvironment,where theyinteractwithstromalcomponentsaccruedfromthemouse host.PDXsmaintainclosesimilaritieswiththetumouroforigin, intermsoftissuearchitecture,molecularfeaturesand responsetotreatments.Indeed,preclinicaltrialsinPDXshave beenshowntomatchandalsoanticipatedataobtainedin patients.Explorationofmorecomplexprocesseslike metastaticevolutionandantitumourimmuneresponsesis activelypursuedwithPDXs,asnewgenerationsofhostmodels emergeonthehorizon.

Addresses

1_{CandioloCancerInstitute,FPO-IRCCS,stradaProv.142,km3,95,} 10060Candiolo(TO),Italy

2_{UniversityofTorino,DepartmentofOncology,stradaProv.142,km} 3,95,10060Candiolo(TO),Italy

Correspondingauthor:Medico,Enzo(enzo.medico@unito.it)

CurrentOpinioninBiotechnology2020,63:151–156

ThisreviewcomesfromathemedissueonSystemsbiology

EditedbyMarcoErcoleVanoniandPasqualePalumbo

https://doi.org/10.1016/j.copbio.2020.01.003

0958-1669/_ã2020TheAuthor(s).PublishedbyElsevierLtd.Thisisan openaccessarticleundertheCCBYlicense(http://creativecommons. org/licenses/by/4.0/).

Introduction

Developmentofexperimentalmodelsthatcorrectly

reca-pitulatetumourbiology,geneticheterogeneityanddrug

responsehasbeenakeyobjectiveofcancerresearchsince

itsinception.PDXsaregeneratedbyimplanting

patient-derivedtumourtissueintohost animalmodels,typically

immunocompromisedmice.Afterafirststepof

engraft-mentandadaptationtothenewhost,PDXtumourscan

begrown,explanted,aliquotedandimplantedinfurther

animals,aprocessknownas‘passage’,togeneratecohorts

ofanimalshostingthesametumour,forpreclinical

treat-ment efficacy studies (Figure 1). PDX-derived tumour

tissue canalsobecryopreserved,togenerateofa‘living

tissue biobank’for virtuallyunlimited tissueavailability

forexperimentsandmolecularprofiling.Themain

advan-tage of PDXs versus invitro cancer cell culturesis the

possibilityofstudyingcancercellsintheir

microenviron-mentandassessingtheinvolvementoffibroblasts,

endo-thelial cells and leukocytes in tumour biology and

response to treatments. Over many years, PDXs have

been used to study multiple aspects of the neoplastic

disease [1]. The PDX field isin continuousexpansion,

and large consortia like EurOPDX in Europe (URL:

http://www.europdx.eu/) and PDXNet in the USA

(URL: https://www.pdxnetwork.org/) have been

estab-lished to standardise proceduresandmake PDXs

avail-abletothecancerresearchcommunity.Inthisreview,we

summarisethemostrecentfindingsderivedfromtheuse

of PDXsascancermodels.

Modelling

drug

response

and

resistance

Patient-derived xenografts (PDXs) are widely used to

recapitulatehowthecomplexityoftumourbiologyaffects

drugresponse.Inthelastfewyears,severalstudieshave

been conducted to define pharmacological

vulnerabil-ities. In this context, a large-scale in vivo screen in

1000PDXmodelsassessedthevalueofsuchapproach

in term of reproducibility and clinical translatability, to

identify associations between genotypes and drug

responses [2]. As an alternative to using large PDX

cohorts to establish genotype-response correlations, a

moreempiricalstrategyisbasedonthe‘Avatar’approach:

an individual patient’s tumour tissue is used to derive

PDXs, onwhichmultipletreatments aretested,so that

themosteffectivecanbeadministeredtothepatientof

origin. PDXs can be used to investigate personalised

treatments in real-time (co-clinical trials), as well as to

findand validatenewtargets[3,4].PDX-basedstudies

allowed optimisation of clinical trial designs, providing

strong rationale for new combinatorial regimens in

patients refractory to conventional treatments [5,6].

PDXmodelshavebeenemployedtorecapitulate

mech-anismsofprimaryandacquireddrugresistance[7,8].In

vivo studies highlighted that tumour resistance

mecha-nisms identified in PDXs are also found in theoriginal

patientsamplesresistanttotargettherapy.Thisprovided

potentialtreatmentoptionsinresistanttumours[9]and

acquired resistance [10]. A limitation of the PDX

approach is the considerable effort and time required

to testinvivo theresponseto multipledrugs. To

over-comethisissue,invivoimplantabledevicesallowtesting

simultaneouslydifferentdrugsinadjacentregionsofthe

same PDX tumour [11]. Moreover, loss-of-function

genetics screenings have been exploited to identify

newtherapeutictargets,employingpooledshorthairpin

RNA(shRNA)libraries,adaptedforinvivoscreensincell

linexenograftsandPDXs[12].

PDX-derived

in

vitro

models

In vivo experiments in PDXs are limited by cost,size,

timeandresources.Reachingtheadequatesamplesizeto

explore inter-patient heterogeneity is therefore quite

challenging. The use of in vitro tests on PDX-derived

cellsfollowedbyinvivovalidation isavalidalternative

and can be achieved in different ways. Short-term 2D

culturescanbeusedforrapid drugtests,thatdisplayed

goodconcordancewithinvivoexperiments[13].

Alter-natively,long-termculturesmaybeemployedto

estab-lishcelllinesretainingthepropertiesofthePDXoforigin

[14].Patient-derivedorPDX-derivedcellscanbegrown

as 3D organoids[15],that retaincertain architectural

featuresandcanbeusedfordrugefficacystudiesandalso

tofurthergeneratePDXswhendirectinvivopropagation

is limiting [16,17]. Cancer-initiating stem cells can be

derived from PDXs and maintained in suspension as

spheroids, a good model for studying the biology and

drugsensitivityofcancerstemcells[18].

Modelling

tumour-stroma

interactions

Itiswidelyrecognisedthatthebiologyofsolidtumours

depends on the ability of cancer cells to recruit and

educatevascular, mesenchymal and immune cells, that

collectively form the tumour microenvironment, or

tumour stroma. During engraftment in a PDX, cancer

cells retainthis ability, so thatwhile thetumour grows

humanstromalcellsarereplacedbymousecounterparts,

thusprovidinganidealexperimentalmodelto

character-isetumour-stromainteractions[19].Indeed,thefactthat

stromal cells in a PDX are of mouse origin has been

considereda limit,potentially affectingtumour biology

andevolution[20].However,thehistologicalarchitecture

ofthetumourtissueismaintainedduringPDX

propaga-tion,suggesting thatthekey mediatorsoftumor-stroma

interactions are functional [21]. Moreover, proteomic

analyses revealed high correspondence of matched

tumourand PDX stromalprofiles, as wellasadaptation

of the stromal proteome upon PDX engraftment, with

distinctprofiles fordifferentPDXsamples[22,23].

Figure1

F0 (Surgery)

“Xen

otr

ial

”

F1 (Engraftment) F2 (Expansion) Vehicle Drug X Drug Y Drug Z Time (days) % r o m u T e m ul o v F3 (Validation)

Vehic

he

Drug X

Drug Y

Drug Z

Biobanking,

molecular profiling,

cell isolation

Current Opinion in Biotechnology

Generation,expansionanduseofPDXs.Thepatient-derivedtumourisimplantedintoanimmunodeficientmouseand,aftertheengraftment phase,itisexplantedandexpandedinmultiplepassages,enablingthegenerationofcohortsofPDXs,suitableforpreclinical«xenotrials»to evaluatedrugefficacy.PDX-derivedtumoursamplescanbecollectedateverypassagetocreateafrozen,vitaltissuebiobankandtoperform molecularprofilesandexvivoexperiments.

Tumor-stroma interactions in PDX have been

exten-sively explored as potential therapeutic targets, we

report here the most recent findings. In colorectal

cancer (CRC)PDXs harbouringAPCmutations,

block-ade of the RSPO3 signal from cancer-associated

fibro-blasts (CAFs) to cancer cells enhanced the activity of

paclitaxel-based chemotherapy [24]. Prolonged

treat-ment oflung cancer xenografts with akinase inhibitor

induced ametabolicshift towardincreased lactate

pro-ductionincancercells, thatinturninducedaparacrine

supplyofHGF fromCAFs, promotingresistance. This

adaptive resistance mechanism was also observed in

patients, which showed clinical relevance [25]. In

CRC PDXs, non-canonical TGF-beta signalling from

cancer cells was found to activate CAFs.Inhibition of

this axis led to reduction of metastatic dissemination

and increased sensitivity to therapy [26]. Finally, in

breast cancer PDXs, microvesicle-mediated transfer of

miR-221 from CAFs to cancer cells was found to

pro-moteresistancetohormonaltherapy,highlightinganew

therapeutic avenue [27].

Modelling

metastasis

Thepossibilityofmodellingmetastaticprogressionusing

subcutaneousPDXimplantsisquitelimited.However,a

muchbetter mimicof thisprocessis obtained byimplanting

PDXsinthesametissuefromwhichthepatient’stumour

wasexplanted[28].Thistypeofimplant,calledorthotopic,

in mostcasesrequiresmicrosurgicalcompetencesandin

vivo imaging, for the implant [29] and for subsequent

evaluationof primary tumour and metastasesdevelopment.

Anotableexceptionisbreastcancer,forwhichthestandard

PDX implant in the mammary fat pad is intrinsically

orthotopicandallowsmodellingmetastasis[30].

Severalapproacheshavebeendevelopedandexploitedto

study the metastatic process and its therapeutically

actionabledependenciesinPDXs.Themost

straightfor-wardistheimplantof theprimarytumour inthe

corre-spondingsiteofthemouse.Metastaticbehaviourcanbe

furtherenhancedwhenstromalcellsareaddedtocancer

cells [31]. When the source material is limited (e.g.

metastasis biopsies), a first generation of PDXs can be

derived subcutaneously,and then implanted

orthotopi-cally, retaining the features of the tumour of origin,

includingmetastaticability[17].Humanmetastaseshave

beenimplantedorthotopicallyintheprimarytumoursite,

tomodelatthesametimetumourgrowthattheprimary

site and metastatic propagation [32]. Finally, human

metastases canbe implanted in the primary metastasis

site,asinthecaseofliver-metastaticCRC,torecapitulate

thebehaviourofthediseasewithinitsmetastatic

micro-environment[33].Itisworthnotingthatinallthesecases

morphological,molecularandpharmacologicalfeaturesof

theorthotopicPDXscorrectlyrecapitulatedthoseofthe

tumour oforigin.

Modelling

anticancer

immunotherapy

Toavoidrejectionofhumantumourimplants,PDXsare

invariablygeneratedinimmunocompromisedmice.

Var-ious mouse strains have been developed over several

decades with progressively increasing

immunosuppres-sion, from athymic nude mice in 1966 to non-obese

diabetic-severe combined immunodeficiency

(NOD-SCID) mice in 1995, to NOD-SCID-Gamma-null

(NSG)in2002.Higherlevelsofimmunodeficiencyallow

better engrafting[34],but alsoprogressively reducethe

possibilityto explore therole oftheimmune systemin

tumour biology and response to treatments [35]. This

drawback is particularlycritical for modelling

immuno-therapybasedoncheckpointinhibitors,asitrequiresan

intact human immune system. A wayto overcome this

limitationisthegeneration of‘humanized’mouse

mod-els, byengraftingNSG micewithhuman leukocytesor

hematopoieticstemcells.Patienttumourtissuesarethen

engraftedintothehumanizedmiceandusedforstudying

theefficacyofimmunomodulatorytreatments.However,

thisapproachhasahighriskofgraft-versus-hostdisease

andthereforeallowsonlyshorttermexperiments[36,37].

Mouse‘humanization’hasbeenimprovedbygenetically

insertinghumancytokinegenes(M-CSF,IL-3,GM-CSF

thrombopoietin) at their respective mouse loci, plus

transgenicexpressionofhumanSIRPa28whichenables

mouse phagocytesto toleratehuman cells[38].Indeed,

recent studies with humanizedmouse PDXs have

pro-videdsignificantadvancesin thecomprehensionof

can-cerimmunotherapy [39,40].

ConventionalPDXscanstillhaveagooduseformodelling

another type of immunotherapy, called ‘adoptive

immunotherapy’,inwhichcytotoxicT-lymphocytes,

nat-uralkillercellsorengineeredkillercellsareadministeredto

testtheirdirectanticanceractivity.Asanexample,

Jesper-senandcolleaguesinsertedmelanomatumourcellsandT

cells fromthesame patientintoNSG miceandshowed

tumour inhibition [41].Alternatively, cytokine-induced

killer(CIK)cellscanbegeneratedfromthepatientorfrom

donorsandtestedforanticancerefficacyinPDX,aloneorin

combination with cancer-targeting antibodies [42]. An

interesting researchfrontier in adoptiveimmunotherapy

isthe engineeringofkillercellswithachimeric antigen

receptor,todirectthemagainstaspecificantigenexpressed

bycancercells.Alsothisapproachhasbeensuccessfully

validatedinPDXmodels[43,44].

PDXs

for

biomarker

development

ThefactthatPDXsprovideavirtuallyunlimitedsource

oftumourtissueformulti-dimensionalmolecularprofiles

andallowtestingresponsetomultipledrugsinthesame

model opened a new avenue for biomarker discovery.

Moreover,thefactthatDNAandRNAfromstromalcells

are of mouse origin enables distinguishing cancer

cell-intrinsic (human) from microenvironmental (mouse)

orsingle-cellsequencing[45].Recentexamplesof

PDX-based biomarker discovery include: (i) identification of

lungcancercell-specificdownregulationof

histocompati-bilityantigens,furtherconfirmedinclinicalsamplesand

associatedtoresistancetoimmunotherapy[46];(ii)lackof

GATA4 expression as a predictor of sensitivity to

TGFBR1inhibition[47];(iii)identificationofp38-alpha

inhibition as sensitizer to taxanes in breast cancer[48].

PDXs havealso been usedto studycirculating tumour

cells (CTCs). In breast cancer, a molecular signature

associated with thepresence of CTC clusters in PDXs

wasassociatedwithworseoutcomeinpatients[49].The

possibilityof performingspecies-specificRNA

sequenc-ing of PDXs to distinguish cancer cell from stromal

transcriptomes hasbeen extensively exploited in CRC,

leadingto theidentificationononesideof apoor

prog-nosisstromalsignature[45],and ontheotherof

cancer-cell-intrinsic molecular subtypes of CRC (CRIS

sub-types) with unprecedented predictive and prognostic

power[50,51].Similarly,acancercell-intrinsic

microsat-elliteinstability transcriptionalsignaturedefinedin

gas-tric cancer PDXs was found to identify patients with

worseprognosis[52].

Conclusion

and

outlook

PDXsofferunprecedented opportunitiesfordeveloping

precisionmedicineapproachestocancertreatment.Vast

PDXcollectionshavebeengeneratedglobally,summing

up to thousands of cases and enabling population-scale

preclinicaltrialsthatcaptureinter-tumourheterogeneity.

Yet,eachindividualmodelrecapitulateswithinitselfthe

complexityofasinglehumantumour,allowingin-depth,

longitudinalexplorationofadaptiveandevolutive

dynam-icsthattypicallyconferresistancetotreatments.ThePDX

communityisfacingimportantchallengestoimprovethe

efficacyofthesemodels.Inparticular,thetime,effortand

costsnecessaryforinvivoexperimentsareurging

research-erstodevelopanduse exvivosystems,exploiting

PDX-derived cells for simpler in vitro experiments before

embarking in PDX tests. Improved ways of modelling

metastatic progression and interactions between cancer

anditsmicroenvironment arecontinuouslybeing

devel-oped,aswellasin-depthmolecularprofilingofalltumour

components.Thetimeinwhichnewlydiagnosedcancer

patientswillfindexistingPDXswithmatchingfeaturesfor

treatmentpersonalizationisnotfar.

Conflict

of

interest

statement

Nothingdeclared.

Acknowledgements

ThisworkwassupportedbyEUH2020ResearchandInnovation Programmeprojectn.731105‘EDIReX’toE.M.;AssociazioneItalianaper laRicercasulCancro5permille2018projectn.21091toE.M.;MyFirst AIRCGrantprojectn.19047toC.I.;andFondazionePiemonteseperla RicercasulCancro-ONLUS5permilleMinisterodellaSalute(year2015to E.M.).

References

and

recommended

reading

Papersofparticularinterest,publishedwithintheperiodofreview, havebeenhighlightedas:

 ofspecialinterest ofoutstandinginterest

1. ByrneAT,Alfe´rezDG,AmantF,AnnibaliD,ArribasJ,BiankinAV, BrunaA,Budinska´ E,CaldasC,ChangDKetal.:Interrogating openissuesincancermedicinewithpatient-derived xenografts.NatRevCancer2017,17:632.

2. GaoH,KornJM,FerrettiS,MonahanJE,WangY,SinghM, ZhangC,SchnellC,YangG,ZhangYetal.:High-throughput screeningusingpatient-derivedtumorxenograftstopredict clinicaltrialdrugresponse.NatMed2015,21:1318-1325.

3. SavaikarMA,WhiteheadT,RoyS,StrongL,FettigN,PrmeauT, LuoJ,LiS,WahlRL,ShoghiKI:SUV(25)andmPERCIST: precisionimagingofresponsetotherapyinco-clinical FDG-PETimagingoftriplenegativebreastcancer(TNBC) patient-derivedtumorxenografts(PDX).JNuclMed2019.pii: jnumed.119.234286.[Epubaheadofprint].

4.

 InoueCurrentA,DeemandfutureAK,KopetzhorizonsS,Heffernanofpatient-derivedTP,DraettaxenograftGF,CarugoA: modelsincolorectalcancertranslationalresearch.Cancers (Basel)2019,11pii:E1321

Thisreviewprovidesinformationonongoingco-clinicaltrialstoevaluate theutilityofPDXmodelsinpredictingdrugresponseinvariousprimary andmetastaticsolidtumoursbeyondCRC.

5. BertottiA,MigliardiG,GalimiF,SassiF,TortiD,IsellaC,Cora` D,Di NicolantonioF,BuscarinoM,PettiCetal.:Amolecularly annotatedplatformofpatient-derivedxenografts

(“xenopatients”)identifiesHER2asaneffectivetherapeutic targetincetuximab-resistantcolorectalcancer.CancerDiscov 2011,1:508-523.

6.

 LonardiSartore-BianchiS,BergamoA,TrusolinoF,ZagonelL,MartinoV,LeoneC,F,BencardinoDepetrisI,MartinelliK, E etal.:Dual-targetedtherapywithtrastuzumabandlapatinibin treatment-refractory,KRAScodon12/13wild-type, HER2-positivemetastaticcolorectalcancer(HERACLES):a proof-of-concept,multicentre,open-label,phase2trial.LancetOncol 2016,17:738-746

SuccessfulclinicaltrialonmetastaticCRCbasedonacombinedtargeted treatmentpreviouslydevelopedandvalidatedinPDXs,seeBertottietal. [5].

7. MarangoniE,LaurentC,CoussyF,El-BottyR,Chaˆteau-JoubertS, ServelyJL,dePlaterL,AssayagF,DahmaniA,MontaudonEetal.: CapecitabineefficacyiscorrelatedwithTYMPandRB1 expressioninPDXestablishedfromtriple-negativebreast cancers.ClinCancerRes2018,24:2605-2615.

8.

 SoleGo´mez-MiragayaA,Pare´ L,Go´mezJ,Mora´nA,SerraS,Calleja-CervantesV,DobroleckiLE,ME,LewisCollado-MT, Diaz-LagaresAetal.:ThealteredtranscriptomeandDNA methylationprofilesofdocetaxelresistanceinbreastcancer PDXmodels.MolCancerRes2019,17:2063-2076

AnexampleofinductionofsecondaryresistanceinvivoinPDXmodels. TheauthorsrenderedPDXsoftriple-negativebreastcancerresistantin vivo to docetaxel, and identified molecular markers of acquired resistance.

9.

 YaoForestYM,A,DonohoNovosiadlyGP,RD,IversenWebsterPW,YW,ZhangEbertY,VanP,BrayHornSRD,etal.: MousePDXtrialsuggestssynergyofconcurrentinhibitionof RAFandEGFRincolorectalcancerwithBRAForKRAS mutations.ClinCancerRes2017,23:5547-5560

ThisPDX-basedtrialrecapitulatedclinicalresultsofcetuximabon color-ectalpatientsanddemonstratedthatconcurrentEGFRandRAFinhibition displayssynergisticantitumoractivityforKRAS-mutantorBRAF-mutant PDXs.

10. KemperK,KrijgsmanO,Cornelissen-SteijgerP,ShahrabiA, WeeberF,SongJY,KuilmanT,VisDJ,WesselsLF,VoestEEetal.: Intra-andinter-tumorheterogeneityina vemurafenib-resistantmelanomapatientandderivedxenografts.EMBOMol Med2015,7:1104-1118.

11. JonasO,LandryHM,FullerJE,SantiniJTJr,BaselgaJ,TepperRI, CimaMJ,LangerR:Animplantablemicrodevicetoperform

high-throughputinvivodrugsensitivitytestingintumors.Sci TranslMed2015,7:284ra.

12. CarugoA,GenoveseG,SethS,NeziL,RoseJL,BossiD, CicaleseA,ShahPK,VialeA,PettazzoniPFetal.:Invivo functionalplatformtargetingpatient-derivedxenografts identifiesWDR5-Mycassociationasacriticaldeterminantof pancreaticcancer.CellRep2016,16:133-147.

13.

 CriscitielloPunziS,MeliksetianC,OrsiF,SpaggiariM,RivaL,L,MarocchiCasiraghiF,M,PruneriDellaVignaG, Petal.: Developmentofpersonalizedtherapeuticstrategiesby targetingactionablevulnerabilitiesinmetastaticand chemotherapy-resistantbreastcancerPDXs.Cells2019,8pii: E605

Theauthorsderivedshort-terminvitroculturesfrommetastaticbreast cancerPDXsandverified concordantinvitroandinvivoresponseto drugs.

14.

 DurinikovaLazzariL,CortiE,ZanellaG,PiccoER,G,NovaraIsellaL,C,BarbosaMontoneFM,etArcellaal.:Patient-P, derivedxenograftsandmatchedcelllinesidentify pharmacogenomicvulnerabilitiesincolorectalcancer.Clin CancerRes2019,25:6243-6259

Thisworkshowsthatlong-termculturedcelllinesderivedfromPDXs retainthegeneticfeaturesanddrugsensitivityprofileofthePDXsof origin.

15.

 MateosVlachogiannisJ,KhanG,K,HedayatLampisA,S,EasonVatsiouK,HuntingfordA,JaminY,I,Ferna´ndez-BurkeRetal.: Patient-derivedorganoidsmodeltreatmentresponseof metastaticgastrointestinalcancers.Science2018, 359:920-926

Organoidsderivedfromcancerpatientswereusedtogenerate xeno-grafts.Thedrugresponseoforganoidsandderivedxenografts recapi-tulatedthoseobservedinpatients.

16.

 ShiMartins-FilhoR,RadulovichS,RaghavanN,NgC,V,LiuLiN,Q,NotsudaMerASetH,al.:CabaneroOrganoidMN, culturesaspreclinicalmodelsofnon-smallcelllungcancer. ClinCancerRes2019.pii:clincanres.1376.2019.[Epubaheadof print]

Theauthorsderivedorganoidsfrompatients’tumoursorPDXofearly stagenon-smalllungcancer.Theorganoidsrecapitulatedthegenetic, transcriptional andhistologicfeatures oftheparentaltumour,that is, patient tumours or PDX, and tumorigenicityin vivo when implanted subcutaneously.

17. MizukoshiK,OkazawaY,HaenoH,KoyamaY,SulidanK, KomiyamaH,SaekiH,OhtsujiN,ItoY,KojimaYetal.:Metastatic seedingofhumancoloncancercellclustersexpressingthe hybridepithelial/mesenchymalstate.IntJCancer2019.Sep 10.[Epubaheadofprint].

18.

 LuraghiBaccoF,P,IsellaBigattoC,BellomoV,CiprianoSE,E,MedicoReatoEG,etOrzanal.:AF,molecularlySassiF,De annotatedmodelofpatient-derivedcoloncancerstem-like cellstoassessgeneticandnongeneticmechanismsof resistancetoanti-EGFRtherapy.ClinCancerRes2018, 24:807-820

Colon-cancerinitiatingcellsderivedfromPDXsusingastem cell-selec-tivemediumgrowinsuspensionasspheroidcoloniesandsuccessfully generatexenografts.Thislong-termplatformallowsthestudyofthemore resistantsubpopulationofcancerstemcellsresponsibleforresistance andrelapse.

19.

 ChaoGomezC,G,WidenQiuS,SG,ThompsonWoodTG,J,ZatarainSprattHJR,etal.:JohnsonPatient-derivedP,GajjarA, xenograftsfromcolorectalcarcinoma:atemporaland hierarchicalstudyofmurinestromalcellreplacement. AnticancerRes2017,37:3405-3412

Accurateworkdepictingtemporalkineticsofhumanstromalcell replace-mentwithmousecounterpartsduringPDXengraftmentandpropagation.

20. Ben-DavidU,HaG,TsengYY,GreenwaldNF,OhC,ShihJ, McFarlandJM,WongB,BoehmJS,BeroukhimRetal.: Patient-derivedxenograftsundergomouse-specifictumorevolution. NatGenet2017,49:1567-1575.

21. DelittoD,PhamK,VladaAC,SarosiGA,ThomasRM,BehrnsKE, LiuC,HughesSJ,WalletSM,TrevinoJG:Patient-derived xenograftmodelsforpancreaticadenocarcinoma demonstrateretentionoftumormorphologythrough incorporationofmurinestromalelements.AmJPathol2015, 185:1297-1303.

22.

 BlommeOtakaY,A,ShererVanSimaeysF,LovinfosseG,DoumontP,BoutryG,S,CostanzaPalaciosB,APBellieretal.:J, Murinestromaadoptsahuman-likemetabolicphenotypein thePDXmodelofcolorectalcancerandlivermetastases. Oncogene2018,37:1237-1250

Metabolicprofileswereanalysedtoshowthatcancercell-driven repro-grammingofmousestromalcellsinPDXsrecapitulateswhatfoundinthe tumoursoforigin.

23. WangX,MooradianAD,Erdmann-GilmoreP,ZhangQ,VinerR, DaviesSR,HuangKL,BomgardenR,VanTineBA,ShaoJetal.: Breasttumorseducatetheproteomeofstromaltissueinan individualizedbutcoordinatedmanner.SciSignal2017,10pii: eaam8065.

24.

 FischerMurrielC,MM,EvansYeungJW,VP,O’YoungCattaruzzaG,BrunnerF,HusseinAL,WangR,YenMWC,etal.: RSPO3antagonisminhibitsgrowthandtumorigenicityin colorectaltumorsharboringcommonWntpathway mutations.SciRep2017,7:15270

NiceexampleoftheuseofPDXstoidentifyatherapeuticallyactionable paracrinesignalfromCAFstocancercells.

25. ApicellaM,GiannoniE,FioreS,FerrariKJ,Ferna´ndez-Pe´rezD, IsellaC,GranchiC,MinutoloF,SottileA,ComoglioPMetal.: Increasedlactatesecretionbycancercellssustains non-cell-autonomousadaptiveresistancetoMETandEGFRtargeted therapies.CellMetab2018,28:848-865.e6.

26.

 Guille´nCimasFJ,Dı´az-MarotoGarcı´a E,N,Gonc¸Sanz-Pamplonaalves-RibeiroS,R,AlbertBerdiel-AcerN,Garcia-M, Vicie´nG,CapellaG,MorenoVetal.:NoncanonicalTGFb pathwayrelievestheblockadeofIL1b/TGFb-mediated crosstalkbetweentumorandstroma:TGFBR1andTAK1 inhibitionincolorectalcancer.ClinCancerRes2019, 25:4466-4479

Inhibitionofthetumour-stromacross-talkreducedstromalaccrualand sensitizedtochemotherapyinCRCPDXmodels.

27. SansoneP,BerishajM,RajasekharVK,CeccarelliC,ChangQ, StrillacciA,SaviniC,ShapiroL,BowmanRL,MastroleoCetal.: Evolutionofcancerstem-likecellsinendocrine-resistant metastaticbreastcancersismediatedbystromal microvesicles.CancerRes2017,77:1927-1941.

28. HoffmanRM:Patient-derivedorthotopicxenografts:better mimicofmetastasisthansubcutaneousxenografts.NatRev Cancer2015,15:451-452.

29. HayCA,SorR,FlowersAJ,ClendeninC,ByrneKT: Ultrasound-guidedorthotopicimplantationofmurinepancreaticductal adenocarcinoma.JVisExp2019,153http://dx.doi.org/10.3791/ 60497.

30.

 FatimaKuenzingerI,El-AyachiWL,WendI,PlayaP,PenceHC,JC,Alva-OrnelasBrakefieldJA,L,KrutilinaKhalidAB,RIetal.: Simultaneousmulti-organmetastasesfromchemo-resistant triple-negativebreastcancerarepreventedbyinterferingwith WNT-signaling.Cancers(Basel)2019,11

Niceexampleoftherapeutic interferencewiththemetastaticprocess usingbreastcancerPDXsimplantedinthemammaryfatpad.

31. GillsJ,MoretR,ZhangX,NelsonJ,MareshG,HellmersL, CanterD,HudsonM,HalatS,MatranaMetal.:Apatient-derived orthotopicxenograftmodelenablinghumanhigh-grade urothelialcellcarcinomaofthebladdertumorimplantation, growth,angiogenesis,andmetastasis.Oncotarget2018, 9:32718-32729.

32.

 KongChoiSI,SY,JeonYoonAR,KA,KimKohMK,YHLeeetal.:YS,DevelopmentImJE,KohJW,ofpatient-HanSS, derivedpreclinicalplatformformetastaticpancreaticcancer: PDOXandasubsequentorganoidmodelsystemusing percutaneousbiopsysamples.FrontOncol2019,9:875

Theauthorsderivedametastaticmodelofpancreaticcancerbymeansof orthotopic implantation of human liver metastases into the primary tumour site, that is, pancreas. The model recapitulated the genetic and histopathological characteristicsof theprimarytumours, aswell asdistantmetastasesanddrug-responseobservedinpatients.

33. Roque-LimaB,RoqueCCTA,BegnamiMD,PeresiP,LimaENP, MelloCAL,CoimbraFJ,ChojniakR,GossSantosT:Development ofpatient-derivedorthotopicxenograftsfrommetastatic colorectalcancerinnudemice.JDrugTarget2019,27:943-949.

34. OkadaS,VaeteewoottacharnK,KariyaR:Applicationofhighly immunocompromisedmicefortheestablishmentof patient-derivedxenograft(PDX)models.Cells2019,8.

35.

 LinnebacherMaletzkiC,BockM:NSGS,FruhmiceP,asMaciushostsK,forWittoncologicalA,PrallF,precision medicine.LabInvest2019.[Epubaheadofprint]

Thisstudyisaclearexampleofhowhighimmunodeficiencyresultsina good engraftment but reduced drug tolerance, highlighting how the choiceofthemousestrainaffectssubsequentresults.

36. AliN,FlutterB,SanchezRodriguezR,Sharif-PaghalehE, BarberLD,LombardiG,NestleFO:Xenogeneic graft-versus-host-diseaseinNOD-scidIL-2RgnullmicedisplayaT-effector memoryphenotype.PLoSOne2012,7:e44219.

37. TanaskovicO,VergaFalzacappaMV,PelicciPG:Humancord blood(hCB)-CD34+humanizedmicefailtorejecthumanacute myeloidleukemiacells.PLoSOne2019,14:e0217345.

38. RongvauxA,WillingerT,MartinekJ,StrowigT,GeartySV, TeichmannLL,SaitoY,MarchesF,HaleneS,PaluckaAKetal.: Developmentandfunctionofhumaninnateimmunecellsina humanizedmousemodel.NatBiotechnol2014,32:364-372.

39.

 CapassoDiamondA,JR,LangKopetzJ,PittsS,BarbeeTM,JordanJ,PetersonKR,LieuJetCH,al.:DavisSL, Characterizationofimmuneresponsestoanti-PD-1monoand combinationimmunotherapyinhematopoietichumanized miceimplantedwithtumorxenografts.JImmunotherCancer 2019,7:37

HumanizedmousePDXswereusedtostudyimmunecheckpoint block-adeinbreastandcoloncancer.Tumourgrowthinhibitioncorrelatedwith increasedtumour-infiltratingTcells.

40.

 WangMaAH,M,DeYaoVereLC,WhiteChengRW,M,AirhartCaiD,SMartineketal.:HumanizedJ,PanCX,miceShiW,in studyingefficacyandmechanismsofPD-1-targetedcancer immunotherapy.FASEBJ2018,32:1537-1549

InthisworkhumanizedmousePDXsallowedtodemonstratethat inhibi-tionoftumorgrowthbypembrolizumab(PD-L1inhibitor)insolidtumours isdependentonhCD8+Tcells,asdemonstratedbyantibody-mediated depletion.

41.

 JespersenKellerU,NyH,L,LindbergSvaneIM,MF,NilssonDoniaM,LM,So¨derbergNilssonEMV,JA:ClinicalAndersenR, responsestoadoptiveT-celltransfercanbemodeledinan autologousimmune-humanizedmousemodel.NatCommun 2017,8:707

Inthisworkmelanomatumourcellsandtumour-infiltratingTcellsfrom the same patient were transplantedsequentially in NSG mice. This systemprovidesthepotential frameworktofurthermodel genetically diversehumancancersforassessingtheefficacyofimmunotherapiesas wellascombinationtherapies.

42. GolayJ,MartinelliS,AlzaniR,CribioliS,AlbaneseC,GottiE, PasiniB,MazzantiB,SaccardiR,RambaldiA,IntronaM:Cord blood-derivedcytokine-inducedkillercellscombinedwith blinatumomabasatherapeuticstrategyforCD19(+)tumors. Cytotherapy2018,20:1077-1088.

43.

 TengYangR,WZhaoetal.:J,ChimericZhaoY,GaoantigenJ,LiH,receptor-modifiedZhouS,WangY,SunTcellsQ,LinZ,

repressedsolidtumorsandtheirrelapseinanestablished patient-derivedcoloncarcinomaxenograftmodel.J Immunother2019,42:33-42

This study showed that CAR-T-cell treatment may be a promising approachforsolidtumourclearanceandthatthePDXmodelmaybe usefultoevaluatetheeffectsofCAR-Tcells.

44. MagnaniCF,MezzanotteC,CappuzzelloC,BardiniM, TettamantiS,FazioG,CooperLJN,DastoliG,CazzanigaG, BiondiA,BiagiE:PreclinicalefficacyandsafetyofCD19CAR cytokine-inducedkillercellstransfectedwithsleepingbeauty transposonforthetreatmentofacutelymphoblastic leukemia.HumGeneTher2018,29:602-613.

45. IsellaC,TerrasiA,BellomoSE,PettiC,GalatolaG,MuratoreA, MellanoA,SenettaR,CassentiA,SonettoCetal.:Stromal contributiontothecolorectalcancertranscriptome.NatGenet 2015,47:312-319.

46. PereiraC,Gimenez-XavierP,ProsE,PajaresMJ,MoroM, GomezA,NavarroA,CondomE,MoranS,Gomez-LopezGetal.: Genomicprofilingofpatient-derivedxenograftsforlung canceridentifiesB2Minactivationimpairing

immunorecognition.ClinCancerRes2017,23:3203-3213.

47. GaoL,HuY,TianY,FanZ,WangK,LiH,ZhouQ,ZengG,HuX, YuLetal.:Lungcancerdeficientinthetumorsuppressor GATA4issensitivetoTGFBR1inhibition.NatCommun2019, 10:1665.

48. Ca´novasB,IgeaA,SartoriAA,GomisRR,PaullTT,IsodaM, Pe´rez-MontoyoH,SerraV,Gonza´lez-Sua´rezE,StrackerTHetal.: Targetingp38aincreasesDNAdamage,chromosome instability,andtheanti-tumoralresponsetotaxanesinbreast cancercells.CancerCell2018,33:1094-1110.e8.

49. ThangavelH,AngelisC,VasaikarS,BhatR,JollyMK,NagiC, CreightonCJ,ChenF,DobroleckiLE,GeorgeJTetal.:A CTC-cluster-specificsignaturederivedfromOMICSanalysisof patient-derivedxenografttumorspredictsoutcomesin basal-likebreastcancer.JClinMed2019,8pii:E1772.

50. IsellaC,BrunduF,BellomoSE,GalimiF,ZanellaE,PorporatoR, PettiC,FioriA,OrzanF,SenettaRetal.:Selectiveanalysisof cancer-cellintrinsictranscriptionaltraitsdefinesnovel clinicallyrelevantsubtypesofcolorectalcancer.NatCommun 2017,8:15107.

51.

 AlderdiceMcCorryAM,M,RichmanRoddyAC,SD,VimalachandranGollinsS,StewartD,JP,IsellaHurtCC,etAdamsal.: R, Prospectivepatientstratificationintorobustcancer-cell intrinsicsubtypesfromcolorectalcancerbiopsies.JPathol 2018,245:19-28

ThispapershowsthattranscriptionalclassifiersdevelopedinPDXsand based on cancer cell-intrinsic transcripts have higher concordance acrossmultiplesamplingsitesofthesametumour.

52. CorsoS,IsellaC,BellomoSE,ApicellaM,DurandoS,MiglioreC, UghettoS,D’ErricoL,MenegonS,Moya-RullDetal.:A comprehensivePDXgastriccancercollectioncaptures cancercell-intrinsictranscriptionalMSItraits.CancerRes 2019,79:5884-5896.