1、Over the past decades,our knowledge of the origin of cancer has increased immensely.Despite substantial progress in the treatment of(certain types of)cancer,it remains a major worldwide health problem1.The number of cancer deaths may be reduced by preven-tive measurements and early detection.In addi
2、tion,the development of new,more targeted therapies offers opportunities.One of the major hurdles for the devel-opment of novel treatment regimens is the challenge of translating scientific knowledge from bench to bedside,which is mainly due to the fact that many cancer models only poorly recapitula
3、te the patients tumour2,and as a consequence,many drugs that perform well in cancer models ultimately fail in clinical trials3.Although animal cancer models have provided important insights into the basics of cancer,their generation is time consuming,and it is argued that these models often do not f
4、aithfully reca-pitulate pathogenic processes in patients.For example,the histological complexity and genetic heterogeneity of human cancers are typically not reflected in genetically engineered mouse models of cancer4.Commonly used human cancer models include cancer cell lines and primary patient-de
5、rived tumour xenografts(PDTXs).Cancer cell lines are derived from primary patient material and have contributed tremen-dously to cancer research.However,they have several drawbacks.For instance,their generation from primary patient material is very inefficient and involves extensive adaptation and s
6、election to invitro 2D culture condi-tions.As only rare clones are able to expand and can be maintained over many passages,the derived cell lines may have undergone substantial genetic changes and no longer recapitulate the genetic heterogeneity of the original tumours.Other limitations of cell line
7、s include the absence of normal tissue-derived control cell lines as reference and the lack of stromal compartments(Table1).A recently developed method called condi-tional reprogramming facilitates the establishment of 2D cell cultures from normal and tumour epithelial cells with high efficiency5.Th
8、ese cultures can be main-tained long term and retain a stablekaryotype.The pro-cedure involves the presence of a RHO kinase inhibitor and fibroblast feeder cells5(Table1).PDTXs have the advantage of mimicking the biological characteristics of the human tumour much better than invitro culture models.
9、PDTXs are generated by transplanting freshly derived patient material subcutaneously or orthotopi-cally into immunodeficient mice.The ability to serially transplant tumour tissues into increasing numbers of animals allows for preclinical testing of novel therapies for cancer treatment(Table1).Limita
10、tions of PDTXs include the use of animals and limited engraftment effi-ciencies for subsets of patient tumours.Moreover,the approach is expensive,time consuming and resource consuming,and PDTXs may undergo mouse-specific tumour evolution6,7.Recently developed 3D culture technologies have led to the
11、development of novel and more physiological human healthy tissue and cancer models.Upon embed-ment into a 3D matrix,tissue-derived adult stem cells can be grown with high efficiencies into self-organizing organotypic structures,termed organoids.In 2009,Sato etal.8 demonstrated that 3D epithelial org
12、anoids can be established from a single leucine-rich repeat-containing G protein-coupled receptor 5(LGR5)+intes-tinal stem cell.Upon embedment into Matrigel,cells are cultured under serum-free conditions mimicking the invivo stem cell niche(involving R-spondin 1(a WNT KaryotypeThe number and appeara
13、nce of chromosomes in the nucleus of a cell.Feeder cellsa layer of cells that is used to support the growth of a cell culture(that is,stem cell cultures)by secretion of important growth factors into the culture medium.Matrigela mouse-derived ex vivo basement membrane substitute that is used to suppo
14、rt 3D growth of organoid cultures.Organoids in cancer researchJarnoDrost1*and HansClevers1,2,3Abstract|The recent advances in invitro 3D culture technologies,such as organoids,have opened new avenues for the development of novel,more physiological human cancer models.Such preclinical models are esse
15、ntial for more efficient translation of basic cancer research into novel treatment regimens for patients with cancer.Wild-type organoids can be grown from embryonic and adult stem cells and display self-organizing capacities,phenocopying essential aspects of the organs they are derived from.Genetic
16、modification of organoids allows disease modelling in a setting that approaches the physiological environment.Additionally,organoids can be grown with high efficiency from patient-derived healthy and tumour tissues,potentially enabling patient-specific drug testing and the development of individuali
17、zed treatment regimens.In this Review,we evaluate tumour organoid protocols and how they can be utilized as an alternative model for cancer research.1Princess Mxima Centre for Paediatric Oncology,Utrecht,Netherlands.2Hubrecht Institute,Royal Netherlands Academy of Arts and Sciences(KNAW)and UMC Utre
18、cht,Utrecht,Netherlands.3Oncode Institute,Utrecht,Netherlands.*e-mail:j.drostprinsesmaximacentrum.nlhttps:/doi.org/10.1038/s41568-018-0007-6Reviews 2018 Macmillan Publishers Limited,part of Springer Nature.All rights reserved.Nature reviews|CAnCER volume 18|JulY 2018|407agonist and ligand of LGR5(re
19、fs9,10),epidermal growth factor(EGF)and the bone morphogenetic protein(BMP)inhibitor noggin).Consequently,LGR5+intestinal stem cells grow out as organotypic,highly polarized epithelial structures with proliferative crypt and differentiated villus compartments8.This culture protocol formed the start-
20、ing point for other organoid culture protocols of multiple mouse and human epithelia,including colon11,12,liver13,pancreas14,prostate15,16,stomach17,fallopian tube18,taste buds19,salivary glands20,oesophagus21,lung22,endometrium23 and breast24(reviewed in ref.25).Organoids can be expanded long term,
21、can be cryopre-served and genetically modified and remain genetically and phenotypically stable.This allows for a wide range of applications in cancer research.Instead of culturing organoids in medium containing niche-recapitulating and tissue-specific growth factors,the Kuo laboratory26 pioneered a
22、n organoid culture system using an airliquid interface with stromal support cells as a source of essential growth factors.Organoids can also be derived from induced pluripotent stem cells(iPSCs)25.However,the efficiency of generating iPSC-based cancer mod-els from patients may depend on cancer type
23、and the presence or absence of specific oncogenic mutations,potentially selecting for outgrowth of tumour subclones and loss of the genetic heterogeneity of the tumour it is derived from27.Generally,it appears more practical to grow tumour organoids directly from cancers than to involve an intermedi
24、ate iPSC step.In this Review,we discuss the use of the adult stem cell-derived organoid technology in both basic and trans-lational cancer research.We highlight the approaches to exploit patient-derived tumour organoid biobanks for drug development and personalized medicine and eval-uate the applica
25、tion of organoid technology as an experi-mental tumour model.Finally,we discuss the limitations and potential of exploiting organoids for cancer research.Organoids for translational researchLiving organoid biobanks.The ability to grow orga-noids with high efficiency from healthy human adult stem cel
26、ls has paved the way to grow organoids from patient-derived tumour tissue.So far,we and others have shown that long-term organoid cultures can be established from primary colon11,28,29,oesopha-gus11,pancreas14,30,stomach17,liver31,endometrium23 and breast24 cancer tissues,as well as from meta-static
27、 colon32,prostate33,34 and breast24 cancer biopsy samples.Importantly,these studies have shown that tumour-derived organoids both phenotypically and genetically resemble the tumour epithelium they were derived from.Tumour organoids do not grow faster perse than their matching normal organoid counter
28、parts and,counterintuitively,in many cases even grow at slower rates,possibly owing to higher rates of mitotic failures and subsequent cell death35,36.Therefore,the overgrowth of tumour organoids by healthy epithelial organoids derived from remaining normal tissue present in tumour biopsy samples ne
29、eds to be avoided.Hence,it is essential to initiate tumour organoid cultures using either pure tumour material or to grow the samples under selective culture condi-tions.For example,in the vast majority of colorectal cancers(CRCs),activating mutations in the WNT signalling pathway are present37.In t
30、hese cases,pure tumour organoid cultures can be obtained by using Table 1|Comparison of the described preclinical cancer modelsFeatureCancer cell linesConditional reprogrammingAdult stem cell-derived organoidsPDTXSuccess rate of initiation+Ease of maintenance+Resource consumptionLowLowMediumHighExpa
31、nsion+3D growth+Retention of phenotypic features invitro+Retention of genetic features invitro+Representation of cancer spectrum+a+Amenable to genetic modification+NA+Matched normal controls+Tumourstroma interactions+Incorporation of an immune systembGenetic cancer modelling(initiation and progressi
32、on)NA+Low-throughput drug screens+High-throughput drug screens+Biobanking+Respective features were judged as best(+),suitable(+),possible(+),not very suitable()or unsuitable().NA,not available;PDTX,patient-derived tumour xenograft.aOnly in epithelial tumours.bThe immune system could be implemented b
33、y co-culturing organoids with haematopoietic cells66,68.Table adapted and updated with permission from ref.116,Elsevier.2018 Macmillan Publishers Limited,part of Springer Nature.All rights 2018|volume 18 culture medium lacking WNT and R-spondins11,which are essential growth factors for healthy tissu
34、e-derived colon organoids.Similarly,tumours harbouring muta-tions in the EGF receptor(EGFR)signalling pathway can be selected by EGF withdrawal29,35,38.The small molecule nutlin-3,which stabilizes p53 by disrupting the binding of p53 to its negative regulator E3 ubiquitin ligase MDM2,has been used a
35、s a strategy to remove TP53-wild-type(WT)healthy cells from TP53-mutant cancer organoid cultures24,35.When such selection methods are not available,using pure tumour cells as starting material is a prerequisite.Currently,large collections of patient-derived tumour and matching healthy organoids are
36、generated and biobanked.These resources can be used to deter-mine whether organoids have predictive value for drug responses for individual patients(fig.1).One of the ear-liest efforts to store large collections of primary tumour cultures was undertaken by Inoue and colleagues39.A collection of cult
37、ures derived from patients with CRC Drug screeningGenomicsLiving organoid biobanksDrug BDrug B used totreat the patientDrug ANormal tissueTumour tissueDrug CDrug BDrug CHealthy organoidHepatocyte organoidApoptotic cellTumour organoidC G A G C TDevelop tumour-specific therapiesLiver toxicity testingP
38、ersonalized medicineFig.1|Organoid cultures for personalized cancer treatment and drug development.Organoids can be established from patient-derived healthy and tumour tissue samples(upper panel).The organoid cultures can be genetically characterized and used for drug screening,which makes it possib
39、le to correlate the genetic background of a tumour with drug response.Organoids can be cryopreserved and stored in living organoid biobanks.The establishment of organoids from healthy tissue of the same patient(middle panel)gives the opportunity to develop less toxic drugs by screening for compounds
40、 that selectively kill tumour cells while leaving healthy cells unharmed.Self-renewing hepatocyte organoid cultures may be used to test for hepatotoxicity one of the causes of drug failure in clinical trials of a potential new drug(lower panel).In this example,drug B seems most suitable for treating
41、 the patient as it specifically kills tumour organoids and does not induce hepatotoxicity.2018 Macmillan Publishers Limited,part of Springer Nature.All rights reserved.Nature reviews|CAnCERReviews volume 18|JulY 2018|409was assembled using a method to expand tumour tissues as tumour spheroids invitr
42、o.This culture method did not sustain the growth of healthy colon epithelium39.More recently,a tumour organoid biobank derived from patients with CRC was generated that consists of a set of 20 genetically diverse tumour organoid cultures and their matching normal tissue-derived organoids28.Integrati
43、ng genomic and monotherapy drug-response data demonstrated that only one of the tumour orga-noid cultures was sensitive to the inhibitor of WNT secretion LGK974(ref.40)owing to a mutation in the WNT antagonist E3 ubiquitin ligase RNF43(refs28,41,42).Besides several other known correlations between d
44、rug responses and the presence of mutations,several com-pounds were identified with differential cytotoxicities among patient-derived tumour organoids without an apparent genetic marker28.Increasing the number of biobanked organoids will be necessary to increase the statistical power to that require
45、d to correlate genetic markers with differences in drug sensitivity.In a separate study,a subset of CRC organoids from this same bio-bank were used for proteomic analyses and comparative transcriptomic analysis43.Substantially different protein profiles were observed between tumour organoids and mat
46、ching healthy organoids as well as among tumour organoids from individual patients displaying distinct personalized profiles43.This finding points towards an important role for proteomic profiling in personalized medicine.However,a direct comparison should be made between the proteomic profiles of t
47、umour organoids and the primary tumour tissue to ensure preservation of protein expression profiles following expansion invitro.In line with the previous study28,Sato and colleagues29 generated an organoid biobank derived from 55 patients with CRC.For certain CRC subtypes,organoids could be establis
48、hed only when cultured under specific conditions.For instance,some organoid lines were maintained in medium lacking an inhibitor of p38 MAPK,a component of the original human colon organoid culture medium,or under hypoxic conditions29.These dependencies may reflect differ-ences in the mutational bac
49、kgrounds of tumours.As the genetic background of tumours is in many cases not determined when establishing organoid cultures from fresh patient-derived tissues,high efficiencies may result from culturing under various media conditions that dif-fer in their combinations of growth factors.Importantly,
50、it was shown that tumour organoids preserve the histo-pathological features of the original tumours not only invitro but also following xenotransplantation under the kidney capsule of immunodeficient mice29.This finding suggests that organoid transplantations can be used to validate invitro drug res
