1、OpinionTumor Functional Heterogeneity Unraveled byscRNA-seq TechnologiesLaura Gonza lez-Silva,1Laura Quevedo,1and Ignacio Varela1,*Effective cancer treatment has been precluded by the presence of various forms of intratumoralcomplexity that drive treatment resistance and metastasis.Recent single-cel
2、l sequencing tech-nologies are significantly facilitating the characterization of tumor internal architecture duringdiseaseprogression.Newapplicationsandadvancesoccurringatafastpacepredictan imminentbroad application of these technologies in many research areas.As occurred with next-genera-tion sequ
3、encing(NGS)technologies,once applied to clinical samples across tumor types,sin-gle-cell sequencing technologies could trigger an exponential increase in knowledge of themolecularpathways involved in cancer progression andcontribute to the improvement of cancertreatment.Multifaceted Heterogeneity an
4、d Its Impact on Cancer ProgressionTumors comprise various cell populations in constant evolution.Some of this complex heterogeneityderives from genetic diversification and Darwinian selection of tumor cells as they adapt to variableenvironments.Next-generation sequencing(NGS;see Glossary)used for th
5、e past decade hadenough sensitivity to detect mutations present in minor cell populations and,combined with multi-sampling of human tumors(multisampling sequencing),fostered many studies that characterized in-tratumor heterogeneity in various cancers 1.The level of intratumor heterogeneity is consid
6、ered amain driver of therapy resistance and metastasis and is associated with poor prognosis 2.Inaddition,humancancersfrequentlyhavetumorcellpopulationswithdifferenttranscriptionalprograms.Thisfunctionaldiversityislikelyassociatedwiththegeneticheterogeneitydescribedabovebutisalsotheresult of many ot
7、her factors.First,the presence of a hierarchical structure,where a group of quiescentstem-likecellsfostersthegrowthofatumorcomprisingcellsindifferentdifferentiationstates,wasdemon-stratedinvarioustumortypes3.Additionally,differenttranscriptionalprogramscanbeactivatedintumorcells as a responseto stoc
8、hastic factorsor to a variable tumormicroenvironment.This functionaldiversityprovides tumors with a plasticity that grants a high capacity for adaptation 4.Finally,human tumors comprise not only malignant/transformed cells but also a plethora of differentcell types recruited from the surrounding tis
9、sue and the immune system.The tumor microenviron-ment shows also genetic and transcriptional diversity and plays important roles in tumor progression,metastasis,and treatment resistance 1,5.Fine characterization of these levels of tumor heterogeneity is essential to the successful treatment ofcancer
10、 patients.The recent development of technologies based on sequencing individual cells(sin-gle-cell sequencing technologies)opens new ways for the characterization of tumor heterogeneity.At the genetic level,single-cell DNA-seq technologies offer higher sensitivity in the detection of mi-nority clone
11、s,the reconstruction of clone structure,and the identification of concurrent or exclusivealterations in the same cells.However,it is in the study of functional heterogeneity that single-cellRNA-seq(scRNA-seq)significantly improves on previous technologies,increasing our molecularcomprehension of can
12、cer progression.A precise cell-type annotation of complex cellular samplesfrom primary tumors is possible thanks to the recent generation of single-cell transcriptome atlases.These comprise normal and pathological samples from human and mouse 6,7.The Emergence of scRNA-seq TechnologiesIn just a few
13、years,the ability to perform single-cell expression profiles increased from a handfulof cells to thousands of cells in a single experiment 8.After the first scRNA-seq experiment in a1Instituto de Biomedicina y Biotecnolog ade Cantabria,Universidad de Cantabria CSIC,Santander,Spain*Correspondence:ign
14、acio.varelaunican.esHighlightsTumors are highly complex entitiescomprising cell populations withvarious transcriptional programs.Single-cell sequencing technolo-gies are evolving fast and have thecapacity to finely characterize thehuge heterogeneity inside tumors.New single-cell sequencing pro-tocol
15、s do not need special infra-structure and can be applied to ahuge multitude of cancer sampletypes in many research areas.A fine characterization of liquid bi-opsies,tumor functional heteroge-neity,and the tumor microenviron-mentwillbefollowedbyanexponential increase in our knowl-edge on tumor progre
16、ssion and willsignificantlyimprovecancertreatment.Trends in Cancer,January 2020,Vol.6,No.1https:/doi.org/10.1016/j.trecan.2019.11.010 2019 The Authors.Published by Elsevier Inc.This is an open access article under the CC BY-NC-ND license(http:/creativecommons.org/licenses/by-nc-nd/4.0/).13Trends in
17、Cancerfour-cell-stage blastomere 9,several studies were published based on cell isolation and individualgenomic library preparation.These initial protocols were laborious and expensive,required RNAamplification steps that introduced bias in the data,and were characterized by reduced throughput10,11.
18、The subsequent introduction of unique molecular identifiers(UMIs),which are random se-quences that label individual molecules,significantly removed cDNA amplification bias 12.Furtherdevelopments in STRT-seq and CEL-seq protocols included the introduction of an individual bar-coding step on isolated
19、cells before a single retrotranscription reaction reducing batch artifacts13,14.In 2015,the introduction of microfluidic devices(Drop-seq 15 and InDrop 16)enabledthe processing of thousands of cells at once.Following this strategy,10 x Genomics automatedequipment recently characterized 1.3 million c
20、ells at the single-cell level 8.Unfortunately,microfluidics-based methods are not efficient in the removal of the abundant rRNA.Consequently,they use poly-T oligonucleotides to sequence the end of poly-A-tailed RNAs.This is useful ingenerating expression profiles in this group of RNAs but does not p
21、rovide complete transcriptomicinformation.Split-seq and Sci-seq strategies avoid physical cell isolation,taking advantage of acombinatorial barcoding strategy that permits the individual labeling of more than 100 000 sin-gle-cell transcriptomes 17,18.These techniques do not require expensive microfl
22、uidics infrastruc-ture and permit greater control over the number of analyzed cells.Finally,single-cell multiomicsapproaches that allow the study of genetic,epigenetic,and transcriptomic profiles in the samecell have been developed 19,20.This opens a window of opportunity for comprehensive cellchara
23、cterization.Single-cell sequencing data analysis is a great challenge,similar to the early years of the use of NGStechnologies.Due to a great variety of sequencing strategies and biological questions,there aremany different reported analysis workflows.Analysis tools for subpopulation identification,
24、differen-tial expression,functional signatures,pseudotiming modeling,and network reconstruction are pub-licly available for researchers with limited bioinformatics resources 21,22.Dissecting the Tumor Ecosystem with scRNA-seqFunctional Diversity of Tumor CellsTranscriptional heterogeneity among tumo
25、r cells has clear and direct clinical implications.First,molecular classification according to transcriptional signatures is commonly used for clinical man-agement in many tumor types.Regarding this,the presence of different transcriptional programsinside the same tumor might prevent,or at least bia
26、s,molecular classification from a single biopsy.In this context,scRNA-seq experiments have demonstrated the presence within the tumor ofmultiple cell populations belonging to different molecular groups according to standard classifica-tions 2325.Second,the presence of functional diversity within tum
27、ors likely improves their adaptation to hostileenvironments.Functionally diverse cell populations with symbiotic,mutually beneficial relationshipshave been reported in tumors 26.This diversity can also be hierarchical,as described in several tu-mor types in which a minority of highly specialized cel
28、ls,termed cancer stem cells(CSCs),might havespecialcapacitiestomaintaintumorgrowth,metastasize,andresistantitumortreatments27.Never-theless,the lack of universally accepted CSC markers and properties has generated controversy inthese studies.scRNA-seq technologies offer an opportunity for the unbias
29、ed identification and studyof those populations that supposedly are present in very low numbers and in a quiescent or dormantstate,and to design more specific antitumor treatments 28.scRNA-seq experiments recentlydemonstrated the presence of populations with stem-like and treatment-resistance proper
30、ties in oli-godendroglioma and melanoma 29,30.Finally,single-cell technologies can detect minor treatment-resistant cell populations inside com-plex tumors,which can be used to select appropriate therapies.For instance,the presence of amelanomacellpopulationexpressinghighlevelsofAXLanticipatedtheocc
31、urrenceofclonal selection after treatment with RAF or MEK inhibitors and the eventual development ofdrug resistance29.GlossaryCirculating tumor cells(CTCs):cancer cells that have escapedfrom the primary tumor and trav-elled through the blood vessels.Functional heterogeneity:pres-ence of cells with d
32、ifferent tran-scriptional programs insidetumors.Genetic heterogeneity:existenceof cell clones with different ge-netic somatic mutations insidehuman tumors.Genomic library:collection ofDNA fragments with commonadapters ready to be analyzed bynext-generation sequencingtechnologies.Intratumor heterogen
33、eity:thepresence of cell diversity insidehuman tumors.Microfluidics:group of tech-niques that allow the manipula-tion of fluids in the range of mi-croliters to picoliters.Multisampling sequencing:comprehensive analysis ofregionally distant samples fromthe same tumor by next-genera-tion technologies.
34、Next-generation sequencing(NGS)technologies:family of ap-plications that allow the afford-able parallel sequencing of hun-dreds of millions of smallfragments in a single reaction.Single-cell multiomics:technolo-gies that allow the simultaneousanalysis of different cell molecularcharacteristics such
35、as genomics,transcriptomics,epigenomics,orproteomics.Single-cell RNA-sequencing(scRNA-seq):analysis of the RNAcontent of single cells by next-generation sequencingtechnologies.Transcriptional signature:a spe-cific set of genes expressed by acell in a given moment underparticular circumstances.14Tren
36、ds in Cancer,January 2020,Vol.6,No.1Trends in CancerTumor MicroenvironmentCancer-associated fibroblasts(CAFs)are present in many if not all solid tumors and participateactively in tumor development 31.The molecular mechanisms behind CAFs role remain largely un-known and the lack of reliable cell mar
37、kers to identify CAFs prevents a clear statement of their abun-dance and importance in solid tumors 32.The origin of CAFs is also under debate.They can be theresult of the transformation of resident fibroblasts previously present in the normal tissue or new cellsgenerated from special cell precursor
38、s recruited to the tumor 33.scRNA-seq reports in thepast years have provided useful information in this respect.Different types of CAFs have been re-ported in breast and colorectal tumors,which is likely to be associated with different cell origins3436.Additionally,each group of CAFs has special fun
39、ctions in the recruitment of immune cellsand in the induction of the epithelialmesenchymal transition(EMT)in tumor cells 24,29,34,36.Tumors are also frequently infiltrated by immune cells.The activation of the immune system to attack tu-mor cells is attractive as an antitumoral therapy 37.Consequent
40、ly,the so-called immunotherapies havebecome a promising toolinfightingcancer,althoughvariableresponseshavebeenobservedwhentheyare applied to cancer patients 38.Thereis a great diversityof immune cells with differing,andprobablyopposite,functions in tumor development.This complexity requires a correc
41、t transcriptional character-ization of all the different cell types present in the tumor 39.Here,scRNA-seq studies offer anunprecedented opportunity.A recent study demonstrated that a high proportion of active versus ex-hausted CD8+T lymphocytes is associated with a better outcome in non-small cell
42、lung cancer 40.Bycontrast,tumors presenting large proportions of regulatory T lymphocytes or myeloid-derived suppres-sor cells have a poor prognosis 4143.The complex relationship between the different immune cellspresent in the tumor will determinean overall tolerant or nontolerant environment.Final
43、ly,some studiessuccessfullyidentifiedtumorneoantigensbysingle-cellcharacterizationoftheTcellreceptor(TCR)reper-toire,which might be useful in the diagnosis and treatment of cancer 40(Figure 1,Key Figure).Circulating Tumor CellsThe characterization of cells that extravasate into the blood circulation
44、,circulating tumor cells(CTCs),constitute a good and low-invasive alternative for the diagnosis and,more importantly,moni-toring of tumors 44.The utility of this strategy has been widely shown in many tumor types and thequantification ofCTCs can be used asa prognostic factor 45.Whereas many authors
45、claimthat CTCsrecapitulate intratumor diversity perfectly,others have reported that they resemble metastasis morethan primary tumors 46.The low number of CTCs present in the blood circulation has forced many studies to purify CTCs ac-cording to specific epithelial surface markers.Debate about the sp
46、ecificity of these markers calls intoquestion some of the reported observations 45.Some current platforms for CTC isolation are basedon physicochemical properties,but it remains unclear whether this constitutes a less biased isolationmethod 47.The high throughput of modern single-cell sequencing tec
47、hnologies offers withoutdoubt an opportunity to reduce the need for extensive purification,which will help to clarify the na-ture and the source of CTCs(Figure 1).Relevant observations came recently from CTC single-cell sequencing studies.The presence of hetero-geneous CTC populations with both epit
48、helial and mesenchymal markers was identified,stressing thatisolation methods based on epithelial markers are likely to be inadequate to capture all CTCs 48,49.Additionally,a recent study on prostate cancer CTCs identified the activation of the noncanonicalWnt signaling pathway,anticipating the appe
49、arance of drug resistance 49.Finally,the presenceof pla-koglobin in breast cancer CTCs was associated with earlier metastasis appearance 50.This suggeststhat we need to include the study of CTCs in therapeutic decision-making in oncological practice.Limitations of Single-Cell Technologies in Human C
50、ancerA major limitation in the application of scRNA-seq technologiesto solid tumor samples is the require-ment for complex dissociation protocols to obtain viable,individualized fresh cells.This limitation isespecially important as several studies have raised caution on potential transcriptional cha
