UBC Faculty Research and Publications

Single-cell sequencing reveals karyotype heterogeneity in murine and human malignancies Bakker, Bjorn; Taudt, Aaron; Belderbos, Mirjam E; Porubsky, David; Spierings, Diana C J; de Jong, Tristan V; Halsema, Nancy; Kazemier, Hinke G; Hoekstra-Wakker, Karina; Bradley, Allan; de Bont, Eveline S J M; van den Berg, Anke; Guryev, Victor; Lansdorp, Peter M; Colomé-Tatché, Maria; Foijer, Floris May 31, 2016

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-13059_2016_Article_971.pdf [ 3.73MB ]
JSON: 52383-1.0366931.json
JSON-LD: 52383-1.0366931-ld.json
RDF/XML (Pretty): 52383-1.0366931-rdf.xml
RDF/JSON: 52383-1.0366931-rdf.json
Turtle: 52383-1.0366931-turtle.txt
N-Triples: 52383-1.0366931-rdf-ntriples.txt
Original Record: 52383-1.0366931-source.json
Full Text

Full Text

RESEARCH Open AccessSingle-cell sequencing reveals karyotypeheterogeneity in murine and humanmalignanciesBjorn Bakker1†, Aaron Taudt1†, Mirjam E. Belderbos1,3, David Porubsky1, Diana C. J. Spierings1, Tristan V. de Jong1,Nancy Halsema1, Hinke G. Kazemier1, Karina Hoekstra-Wakker1, Allan Bradley2, Eveline S. J. M. de Bont3,Anke van den Berg4, Victor Guryev1, Peter M. Lansdorp1,5,6, Maria Colomé-Tatché1,7* and Floris Foijer1*AbstractBackground: Chromosome instability leads to aneuploidy, a state in which cells have abnormal numbers ofchromosomes, and is found in two out of three cancers. In a chromosomal instable p53 deficient mouse modelwith accelerated lymphomagenesis, we previously observed whole chromosome copy number changes affectingall lymphoma cells. This suggests that chromosome instability is somehow suppressed in the aneuploid lymphomasor that selection for frequently lost/gained chromosomes out-competes the CIN-imposed mis-segregation.Results: To distinguish between these explanations and to examine karyotype dynamics in chromosome instablelymphoma, we use a newly developed single-cell whole genome sequencing (scWGS) platform that provides acomplete and unbiased overview of copy number variations (CNV) in individual cells. To analyse these scWGS data,we develop AneuFinder, which allows annotation of copy number changes in a fully automated fashion andquantification of CNV heterogeneity between cells. Single-cell sequencing and AneuFinder analysis reveals highlevels of copy number heterogeneity in chromosome instability-driven murine T-cell lymphoma samples, indicatingongoing chromosome instability. Application of this technology to human B cell leukaemias reveals different levelsof karyotype heterogeneity in these cancers.Conclusion: Our data show that even though aneuploid tumours select for particular and recurring chromosomecombinations, single-cell analysis using AneuFinder reveals copy number heterogeneity. This suggests ongoingchromosome instability that other platforms fail to detect. As chromosome instability might drive tumour evolution,karyotype analysis using single-cell sequencing technology could become an essential tool for cancer treatmentstratification.Keywords: Aneuploidy, Karyotype heterogeneity, Single-cell sequencing, Copy number detection, Lymphoma,LeukaemiaBackgroundChromosomal instability (CIN) is a process leading tostructural and whole chromosome abnormalities andresults in cells with an abnormal DNA content, a statedefined as aneuploid. CIN and the resulting aneuploidycause physiological stress and growth defects in yeastand primary mouse embryonic fibroblasts [1, 2].Furthermore, some of the mouse models that were engi-neered to model CIN are characterised with a reducedlifespan, which can be rescued by reducing the levels ofCIN [3–5]. Although aneuploidy has detrimental conse-quences for untransformed cells, more than two out ofthree cancers are aneuploid, suggesting a fundamentalrelationship between aneuploidy and tumourigenesisthat so far remains poorly understood [6–8].Various mouse models have been engineered toinvestigate the relationship between aneuploidy and* Correspondence: maria.colome@helmholtz-muenchen.de; f.foijer@umcg.nl†Equal contributors1European Research Institute for the Biology of Ageing, University ofGroningen, University Medical Center Groningen, A. Deusinglaan 1,Groningen 9713 AV, The NetherlandsFull list of author information is available at the end of the article© 2016 Bakker et al. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Bakker et al. Genome Biology  (2016) 17:115 DOI 10.1186/s13059-016-0971-7tumourigenesis, typically by abrogating components ofthe spindle assembly checkpoint (SAC) [9–11]. The SACmonitors chromosome segregation by blocking mitoticcells in metaphase until proper kinetochore-microtubuleattachment and tension have been established [12–14],and its (partial) abrogation will therefore lead to CIN. Wehave previously shown that CIN as provoked by truncat-ing the SAC kinase Mps1 selectively in murine T-cells isnot sufficient to trigger malignant proliferation. However,when combined with loss of p53, CIN significantlyaccelerated lymphomagenesis. When we analysed theaverage chromosome content in the emerging T cell acutelymphoblastic lymphomas (T-ALLs) by array comparativegenomic hybridisation (aCGH), we found that all lymph-omas displayed highly similar karyotypes, suggestingclonal selection for the recurring chromosomal abnormal-ities [15]. This was surprising because we also showed thatMps1 truncation results in mitotic abnormalities in virtu-ally each cell division in tissue culture cells, which wouldcounteract clonal selection of any chromosomal abnor-mality in a tumour. One possible explanation is that aneu-ploid T-ALL cells can overcome the CIN phenotype andthus maintain a stable karyotype. Alternatively, selectionforces for specific chromosome alterations outcompetethe mis-segregation events. The resulting chromosomalimbalances are expected to have severe effects on geneexpression and thus cellular fitness. Although most formsof aneuploidy are expected to decrease fitness, karyotypeheterogeneity could result in selection of cells that haveaccumulated favourable copy numbers of chromosomesexpressing genes important for tumour evolution andoverall cellular fitness.Indeed, karyotype heterogeneity occurs in human can-cers [16, 17], has been associated with human tumourevolution, and might therefore impact therapeutic re-sponse to cancer therapy [18]. However, most currentcytogenetic and molecular techniques are limited in thenumber of karyotype alterations per cell they can detect,are biased towards dividing subpopulations, or can onlymeasure the population-average chromosome copy num-ber alterations [19, 20]. These shortcomings have pre-cluded thorough analysis of intratumour chromosomecopy number variations. Recent advances in single-cellgenomics allow researchers to dissect the heterogeneity ofthe cancer genome with greater resolution than ever be-fore [21, 22].In this study we describe the application of single-cellwhole genome sequencing (scWGS) to map and quantifykaryotype heterogeneity in primary mouse lymphomaand human leukaemia samples. For this purpose, wehave developed a new bioinformatics tool, AneuFinder,to identify chromosome copy numbers in scWGS data.Using these tools, we now show that besides recurrentchromosome copy number alterations, the aneuploidT-ALLs that arise in our mouse model display high-gradeongoing CIN as evidenced by severe intratumour karyo-type heterogeneity. Importantly, analyses of a number ofhuman paediatric B cell acute lymphoblastic leukaemia(B-ALL) samples using this platform revealed differentgrades of karyotype heterogeneity, demonstrating thatCIN rates differ between these malignancies. Therefore, asongoing CIN is an important hypothesised driver of can-cer evolution [17], our platform might become an import-ant tool to predict treatment outcome or even treatmentstratification.ResultsChromosomal instable T-ALLs show recurrentchromosome copy number changesWe previously reported that CIN in T-cells induced by atruncation of the SAC kinase Mps1 synergises with p53loss in lymphomagenesis [15]. When we re-examinedour earlier aCGH data [15], we again identified recurrentchromosome copy number changes in a large cohort of an-euploid lymphomas, most notably gains of chromosomes4, 9, 14 and 15 (Fig. 1a, Additional file 1: Figure S1). Thefact that these recurrent chromosomal abnormalities weredetectable by bulk measurement aCGH (i.e. measuring theaverage copy number changes in a piece of tumour andtherefore millions of cells) indicates that the majority ofthe T-ALL cells in the individual lymphomas displayedthese aneuploidies [19]. Indeed, when we determinedchromosome 15 aneuploidy in individual cells using inter-phase FISH, we confirmed that >70 % of the cells had threeor more copies [15]. As Mps1 truncation is expected tocause ongoing chromosome instability, these clonal karyo-types were unexpected. Two possible explanations for thisare: (1) the malignancies somehow compensate for Mps1truncation, thus alleviating the CIN phenotype; or (2) theongoing CIN is outcompeted by a selection that ultimatelydrives lymphoma cells to converge towards favourablechromosome-specific copy number states. If the latterexplanation is true, T-ALLs should display cell-to-cellvariability for chromosome numbers, i.e. karyotype hetero-geneity [15].Traditional methods to examine karyotypes depend ondividing cells (in case of normal and spectral karyotyping[SKY]), or are limited in the number of chromosomes thatcan be quantified per cell (in case of interphase FISH). Analternative to measure copy number alterations in atumour is to measure the average DNA content (e.g. byaCGH) [19, 23], but this obscures intratumour heterogen-eity. We therefore moved to single-cell sequencing as amethod for karyotyping, making use of a modified scWGSprotocol, described in more detail in van den Bos et al.[24]. Briefly, this scWGS platform involves single-cellsorting of primary tumour cells as nuclei by flow cytome-try, followed by automated DNA fragmentation, barcodedBakker et al. Genome Biology  (2016) 17:115 Page 2 of 15next generation sequencing library preparation andshallow multiplexed sequencing [24]. To validate ourplatform, we first sequenced the genomes of 25 primaryT-ALL cells isolated from an Mps1f/f; p53f/f; Lck-Cre+lymphoma that we had previously assessed [15] by aCGH-analysis (T-ALL 1, aCGH data in Fig. 1a). We first com-pared the single-cell sequencing data to the existing aCGHdata by creating an artificial ‘bulk sequencing file’ that hasthe cumulative data of all individual single-cell sequencinglibraries (Fig. 1b) to determine how representative thesampled cells are for the bulk tumour. Indeed, we foundthat the copy number changes in the bulk sequencing ana-lysis were identical to those observed in the aCGH data(compare Fig. 1a and b). However, when we plotted the in-dividual single-cell libraries, we detected many additionalcopy number changes: while some of the single cellsshowed the exact karyotype as found by aCGH analysis(Ts2, Ts4, Del7, Ts9, Ts14, Ts15; Fig. 1c; cell 1, more exam-ples in Additional file 2: Figure S2), most cells displayedadditional chromosomal aberrations (Fig. 1c; compareFig. 1 Chromosomal instable T-ALL display recurring chromosome copy numbers, as assessed by array CGH. a Two representative T-ALLs analysedusing array CGH, compared to a euploid reference, showing recurrent gains of chromosomes 4, 9, 14 and 15, and other tumour-specific alterations.The purple bars indicate the mean log-value of the respective chromosome. b Cumulative single-cell sequencing libraries to simulate bulk data,showing a comparable karyotype as found by aCGH. c Single-cell sequencing analysis of four representative cells from T-ALL 1, showing identicalchromosome copy numbers to the aCGH profile (cell 1), or cell-unique copy numbers (cells 2, 3 and 4; red arrows)Bakker et al. Genome Biology  (2016) 17:115 Page 3 of 15cells 2, 3 and 4 to cell 1, more single-cell libraries inAdditional file 2: Figure S2), reflecting karyotype hetero-geneity and suggestive of ongoing CIN. Indeed, when wemanually annotated the individual karyotypes of all 25cells, we found that 56 % of the cells had a unique karyo-type (Additional file 3: Figure S3), further emphasising theheterogeneity that our earlier aCGH analysis hadfailed to detect.AneuFinder: a tool to analyse high throughput single-cellsequencing dataWhile the scWGS data provided greater insight into thediversity of karyotypes than ‘bulk’ aCGH analysis, anno-tating the individual karyotypes is labour-intensive, and,more importantly error prone and possibly biased askaryotypes are annotated by visual inspection. Further-more, a minority (~11 % of the libraries, Additional file 4:Table S1) of the single-cell sequencing libraries are of poorquality, making unbiased quality control essential forfaithful copy number annotation. To automate qualitycontrol, and to facilitate copy number calling of single-cellsequencing data, we developed an automated analysispipeline called AneuFinder with the following key fea-tures: (1) independence of an external reference for copynumber analysis; (2) automated quantification of CNVsusing a Hidden Markov model [25]; (3) stringent semi-automated quality control of individual sequencing librar-ies; and (4) generation of BED files for an external genomebrowser to zoom into small amplified/lost regions. Aftersample sorting and sequencing (Fig. 2a), AneuFinder ana-lyses the aligned sequencing reads (e.g. BAM files). Thereads are counted in non-overlapping bins of variable sizebased on mappability, averaging at 1 Mb in size, followedby a GC correction (Fig. 2b, Additional file 5: Supplemen-tary Materials & Methods). A Hidden Markov model isthen applied to the binned sequencing data, assuming sev-eral possible (i.e. biologically relevant) states, from nullis-omy up to decasomy (10 copies). All states are modelledby a negative binomial distribution, except for the nullis-omy that is modelled by a delta distribution (see Add-itional file 5: Supplementary Materials & Methods). TheBaum-Welch algorithm [25] is used to obtain the best fitfor the distribution parameters, transition probabilitiesand posterior probabilities (Fig. 2c). Subsequently, eachbin is assigned the copy number state with the highestposterior probability (Fig. 2d). Since single-cell sequencingcan be inherently noisy, we included a stringent qualitycontrol using a multivariate clustering approach of indi-vidual libraries based on several quality measures calcu-lated by AneuFinder. These include, for example, the bin-to-bin variation in read count (spikiness), the number ofcontiguous stretches of bins with the same copy numberstate and the Bhattacharyya distance [26] between thenegative binomial distributions (also see Additional file 5:Supplementary Materials & Methods). We then select thebest scoring cluster for downstream analysis, resulting in~89 % of libraries being used in downstream analyses(Fig. 2e, Additional files 4 and 5: Table S1 and Materials &Methods). To assess the extent of karyotypic heterogen-eity in a set of cells, we developed a heterogeneity score,as well as an aneuploidy score (divergence from euploidy;Fig. 2f, Additional file 5: Supplementary Materials &Methods, Additional file 6: Table S2). Finally, the copynumber states are plotted in a genome-wide fashion, withcells being clustered based on the similarity of their copynumber profile. We extensively compared the AneuFinderpipeline to another tool used for identification of CNVs insingle-cell sequencing data, Gingko [27], and found thatcopy number calls were in general concordant betweenboth methods, although Gingko was less sensitive thanAneuFinder for the detection of small CNVs, while beingmore robust to sequencing noise (see Additional file 5:Supplementary Materials & Methods, and Additional file 7:Figure S4). The AneuFinder software was imple-mented as R-package AneuFinder and is freely avail-able through Bioconductor: http://bioconductor.org/packages/AneuFinder/.scWGS and AneuFinder reveal karyotype heterogeneity inmurine CIN T-ALLsTo further investigate karyotype heterogeneity in CINlymphomas, we set up cohorts of Mps1f/f; Lck-Cre+,Mps1f/f; p53f/f; Lck-Cre+ andMps1f/f; p53f/+; Lck-Cre+ mice,and Lck-Cre− mice as controls. Mice were sacrificed whenexhibiting signs of lymphoma (typically dyspnoea due toan enlarged thymus), thymuses were harvested andprimary T-ALL single cell suspensions were frozen forsubsequent single cell sequencing analysis. None of theMps1f/f Lck-Cre+ (n = 71) or Lck-Cre− animals (n = 63)succumbed to lymphoma within the first year (Fig. 3a,green and purple lines, respectively), while both theMps1f/f; p53f/f; Lck-Cre+ (n = 31) and Mps1f/f; p53+/f;Lck-Cre+ (n = 58) developed T-ALLs with median latenciesof 3.4 and 4.0 months, respectively (Fig. 3a, blue and redlines, respectively), identical as in our previous cohorts[15]. Thymic weights were in the range of 400–1420 mg(median: 1300 mg) and 590–1780 mg (median: 1560 mg)for Mps1f/f p53f/f Lck-Cre+ and Mps1f/f p53+/f Lck-Cre+animals, respectively, while Mps1f/f Lck-Cre+ thymuses were20–88 mg (median: 67 mg) and Mps1f/f p53f/f Lck-Cre−control thymuses were 20–60 mg (median: 40 mg; Fig. 3b).To determine aneuploidy at the single-cell level, we se-lected four Mps1f/f; p53f/f; Lck-Cre+ T-ALLs, one Mps1f/f;p53f/+; Lck-Cre+ T-ALL (Fig. 3a, arrows) and one controlthymus (Mps1f/f; p53f/f; Lck-Cre−) for single-cell se-quencing. For each sample, we sequenced 48 single-celllibraries. While AneuFinder did not detect any chromo-some copy number alterations in the male Lck-Cre−Bakker et al. Genome Biology  (2016) 17:115 Page 4 of 15abcde fgFig. 2 (See legend on next page.)Bakker et al. Genome Biology  (2016) 17:115 Page 5 of 15control thymus (Fig. 3c), all five T-ALL samples revealedclonal copy number gains of chromosomes 4, 9, 14 and 15(Fig. 3d–h), highly similar as identified in Mps1; p53; Lck-Cre T-ALLs assessed by aCGH analysis (Fig. 1a and [15]).However, more importantly, AneuFinder also detectedmany other chromosome copy number variations thatwere present in a minority of the lymphoma cells that wepredict aCGH analysis would have failed to detect. Indeed,these additional chromosome copy number alterationsdisappeared when we analysed our single - cell sequencingin bulk using the ‘bulk analysis sequencing files’ similar toFig. 1b (Additional file 8: Figure S5). Furthermore, Aneu-Finder also revealed heterogeneity in the copy numberstates for the frequently gained chromosomes that wefailed to detect previously by aCGH [15]. For instance, intumours T158, T170 and T257 half of the cells had threecopies of chromosome 14, while the other half had four oreven more copies (compare Fig. 3e–g). All together, thesedata clearly show that Mps1f/f; p53f/f; Lck-Cre+ T-ALLs arenot only highly aneuploid, but also display high-gradekaryotype heterogeneity, suggestive of ongoing CIN.Quantifying karyotype heterogeneity in CIN T-ALLsOur analyses so far revealed that the endpoint tumoursin the Mps1 mouse model are aneuploid with recurrentkaryotypes for some chromosomes, but also displayhigh-grade karyotype heterogeneity for most of the otherchromosomes, suggesting ongoing CIN. To furtherquantify the extent of CIN (i.e. karyotype heterogeneity)during tumourigenesis, we next harvested Mps1; p53;Lck-Cre thymuses from mice at ages 10, 13 and 14 weekswhen mice did not show any external evidence forT-ALL yet. Indeed, thymic weights were lower than end-point lymphomas (40, 590 and 650 mg, respectively).While AneuFinder analysis revealed that few cells wereaneuploid at 10 weeks, the karyotypes observed in the en-larged thymuses harvested from 13- and 14-week-oldMps1; p53; Lck-Cre mice were already very similar tothose observed in endpoint T-ALLs (compare Additionalfile 9: Figure S6b, c to Fig. 3d–g) indicating that selectionfor the frequently gained chromosomes 4, 9, 14 and 15 oc-curs early in lymphoma development. We then comparedthe levels of karyotype heterogeneity between developingand endpoint T-ALLs. ‘Baseline’ heterogeneity in perfectlydiploid Lck-Cre− control T-cells was virtually zero(0.0009, Fig. 4a, black diamond). Indeed, the 10-week-oldthymus showed an increase in the heterogeneity score(0.0226) compared to the control thymus as a result ofcopy number changes in the minority of the cells (Fig. 4a,karyotypes in Additional file 9: Figure S6a). Furthermore,five endpoint T-ALLs (Fig. 4a, blue circles, scWGS T-ALLkaryotypes in Fig. 3d–g) showed a dramatic increase inthe heterogeneity score (0.1032 – 0.2550), further em-phasising the large variation in karyotypes present in thoseT-ALLs. Interestingly, the developing lymphomas har-vested from 13- and 14-week-old Mps1; p53; Lck-Cre miceshowed comparable heterogeneity and aneuploidy scoresto the five assessed endpoint lymphomas (Fig. 4a, compareorange circle and square to blue circles, lymphoma karyo-types in Additional file 9: Figure S6b, c), suggesting thatCIN rates remain constant during tumourigenesis in ourmouse model.Since different chromosomes showed varying degreesof gain or loss in the T-ALLs (Fig. 3d–h and Additionalfile 9: Figure S6b, c), we wondered whether calculatingthe aneuploidy and heterogeneity scores for individualchromosomes would reveal whether specific chromo-somes more often showed changes in copy number thanothers. To this end, we plotted both scores per chromo-some for all samples that were analysed by single-cellsequencing. For the control thymus, all chromosomesclustered together in the bottom left, indicating thatnone of the cells displayed chromosome copy numberalterations (Fig. 4b, control thymus). In contrast, in thetumours we identified three ‘types’ of chromosomes:(1) chromosomes that were (virtually) never lost or gained(Fig. 4b, green chromosomes in T260 and T158), presum-ably due to lethality associated with such gain/loss events;(2) chromosomes that show a high heterogeneity rate, butlow aneuploidy rate, for which copy number changes arepresumably not selected for but that are tolerated rela-tively well (Fig 4b, blue chromosomes); and (3) chromo-somes with high aneuploidy, but low(er) heterogeneityscores, for which the observed copy number changes are(See figure on previous page.)Fig. 2 AneuFinder – automated copy number analysis of single-cell sequencing data. a Samples are homogenised, single-cell sorted and sequenced.b Aligned sequencing reads are counted in non-overlapping bins of variable size based on mappability. c A Hidden Markov Model with multiplehidden states is applied to the binned read counts in order to predict copy number state of every single bin. Emission distributions are modelled asnegative binomial distributions (NB (r,p,x)). d The model parameters are estimated using the Baum Welch algorithm and every binned read count isassigned to the copy number state that maximises the posterior probability. e Quality of each single-cell library is assessed based on the followingmeasures: spikiness, loglikelihood of the model determined by the Baum-Welch algorithm, number of separate copy number segments and Bhattacharyyadistance. Libraries are clustered based on these measures: the highest scoring cluster is selected for further analysis. f The extent of aneuploidy ismeasured as the divergence of a given chromosome from the normal euploid state. At the cell population level, heterogeneity is measured as thenumber of cells with a distinct copy number profile within the population. g Example of a genome-wide copy number profile of a population ofT-ALL cells. Each row represents a single cell with chromosomes plotted as columns. Copy number states are depicted in different colours. Cells areclustered based on the similarity of their copy number profileBakker et al. Genome Biology  (2016) 17:115 Page 6 of 15a bcdefghFig. 3 (See legend on next page.)Bakker et al. Genome Biology  (2016) 17:115 Page 7 of 15selected for in the tumours (Fig 4b, red chromosomes).These results (additional plots in Additional file 10:Figure S7) indicate that while some chromosomes copynumber changes are favoured and others are not toler-ated in these tumours, there is a third group of chro-mosomes for which copy number changes contribute toheterogeneity, but that are not (yet) selected for.Even though we observed high levels of karyotype het-erogeneity, the single-cell sequencing experiments areendpoint measurements and we therefore formally couldnot rule out that the T-ALLs consisted of a large numberof different aneuploid but chromosomal stable clones.To determine whether chromosome mis-segregationevents still occurred in the primary malignancies, wederived primary lymphoma cell lines from three end-point T-ALLs. This yielded three viable primary murineT-ALL lines (T302, T392 and T397) that displayedchromosome numbers similar as observed in primarylymphomas as assessed by metaphase spread-basedchromosome counting (median 47.5–51 for the T-ALLlines [passage 8], compared to 46–50 for the primary T-ALLs [compare Fig. 4c to Fig. 3d–h]). Furthermore, thelines displayed variability in chromosome numbers,indicating karyotype heterogeneity (Fig. 4c). We thenlabelled primary T-ALL cell line 1 (T302) with anH2B-GFP fusion protein to monitor the actual rates ofchromosome mis-segregation by time-lapse microscopy.Indeed, we found that half of the mitoses (n = 32) showedclear mitotic abnormalities, mostly lagging chromosomesand failed alignments, similar to those observed previouslyin Mps1f/f; p53f/f mouse embryonic fibroblasts (Fig. 4d, e,Additional files 11, 12 and 13: Movies 1–3 [15]). Wetherefore conclude that the karyotype heterogeneity thatwe observed in developing and endpoint Mps1; p53;Lck-Cre lymphomas is the result of ongoing CIN.However, as we observed roughly 10 % of the mitosesin our primary T-ALL lines to result in tetraploid cells(Fig. 4d, e) and as for single-cell sequencing purposes wetypically sort near diploid (2n) cells, we next investigatedwhether this resulted in an underestimation of the het-erogeneity in endpoint tumours. To test this, we first ex-amined the fate of T-ALL cells undergoing binucleationor polyploidisation events by time-lapse imaging in pri-mary murine T-ALL cultures. While all cells displayingpolyploidisation or binucleation events died between 4and 11 h after mitotic exit (n = 7, e.g. Additional file 14:Movie 4), only 20 % of the cells that showed no clearmitotic abnormalities died in this experiment (n = 10).Even though this indicates that tetraploidisation eventsare selected against in a tissue culture setting, this stillcould be different in vivo, given that whole genome dupli-cation is known to occur in human cancers. However, aswith our existing strains we could not distinguish G2 cellsfrom G1 tetraploid cells by FACS, we therefore analysedaneuploid murine T-ALLs for cells with a DNA contentlarger than 4n as an indication for proliferating tetraploidcells. Indeed, two out of the four analysed tumours (T158and T257) had cells with a DNA content larger than 4n(see arrows Additional file 15: Figure S8a) indicating thatsome of the cells with a 4n DNA content had to be tetra-ploid. We then sorted individual G1 and G2-diploid/G1-(near-)tetraploid cells followed by single-cell sequencingfrom T158 (Fig. 4f, g) and T257 (Fig. 4h, i), respectively.Even though in this experiment we could not distinguishG2 cells from (near) tetraploid cells, the data revealed thatnear tetraploid cells were present as some of the cells hadodd copy numbers of chromosomes and G2 cells by defin-ition must have an even number of chromosome copies.Importantly, this allowed us to confirm that AneuFindercan detect such events: when we forced AneuFinder to fitthe 4n population as diploid, the resulting fit was poor forcells with odd numbers (and thus true near-tetraploidcells; compare left and right hand panels in Additionalfile 15: Figure S8b, c) indicating that AneuFinder candetect tetraploid events even without a priori know-ledge about the most prevalent state (diploid/tetra-ploid), similar to Ginkgo.Further analysis of the 4n population of T158 revealedthat 12 out of 37 analysed cells had an odd copy numberfor at least one of the chromosomes, indicating that atleast 32 % of the 4n cells were true near-tetraploid andnot G2 cells. The heterogeneity score for these near-tetraploid cells (0.6503) was much larger than for the 2npopulation (0.3099), indicating that the near-tetraploidcells can generate greater karyotype diversity than near-diploid cells, presumably because of the larger numberof copy number states available. For the other 4n cells(68 %) we could not discriminate whether these were G2cells or G1 near-tetraploid cells in this experimentalsetup. However, the 4n population only represented4.1 % of all the live tumour cells, and therefore only be-tween 1.3 % (32 % of 4.1 % near-tetraploid cells withodd numbers of chromosomes) and 4.1 % (all cells withnear 4n DNA content) in tumour T158 were near-(See figure on previous page.)Fig. 3 Single-cell sequencing analysis confirms both recurrent chromosome copy numbers and karyotype heterogeneity in CIN T-ALLs. a Kaplan-Meiersurvival curve of the listed genotypes. T-ALLs that were single-cell sequenced in this study are indicated with a black arrow and tumour ID. b Thymic/T-ALL weights of the listed genotypes. T-ALLs indicated in red were analysed using single-cell sequencing. c–h Genome-wide copy number plots asgenerated by the AneuFinder algorithm for six samples: one Lck-Cre− control thymus and five CIN-driven T-ALLs. Individual cells are represented inrows, with the copy number state for ~1 Mb bins indicated in colours (see legend)Bakker et al. Genome Biology  (2016) 17:115 Page 8 of 15ac edbfghiFig. 4 (See legend on next page.)Bakker et al. Genome Biology  (2016) 17:115 Page 9 of 15tetraploid cells. The contribution of these cells tointratumour heterogeneity is therefore limited and inagreement with the observed cell death events in thetime-lapse experiments in primary aneuploid T-ALL cul-tures (Additional file 14: Movie 4). In the T257 4n popu-lation (5.3 % of all tumour cells) we found two out of 34analysed cells to have odd chromosome numbers, re-presenting ~0.3 % of the total tumour. Therefore, thecontribution of genuine near-tetraploid cells to tumourT257 lies between 0.3 % and 5.3 %. We conclude thatwhile the near-tetraploid cells contribute to intratumourheterogeneity, this contribution is limited, presumablydue to cell death events after polyploidisation leading tolow fractions of near-tetraploid cells in the primary an-euploid T-ALLs.Karyotype heterogeneity in human B cell acutelymphoblastic leukaemiaOur results so far indicate that the lymphomas that arisein our Mps1; p53; Lck-Cre T-ALL model are aneuploid,with recurring chromosomes affected, and that chromo-some numbers vary between cells for all chromosomes,resulting in intratumour karyotype heterogeneity. To in-vestigate the relevance of our findings for human cancer,we next assessed to what extent human aneuploidtumours display karyotype heterogeneity. For this, we se-lected three B-ALL samples, one near-euploid (sample A),one intermediate aneuploid (sample B) and one highlyaneuploid (sample C) as quantified by ‘traditional’ cyto-genetics at the time of diagnosis (Fig. 5a). Indeed, whenwe analysed all single-cell sequencing libraries per tumouras bulk, we found the average karyotypes of the tumoursto be similar to the reported cytogenetic reports (compareFig. 5a to Additional file 16: Figure S9a). When assessed atthe single-cell level, the near-euploid B-ALL A (Fig. 5b,top panel) did not show any whole chromosome copynumber alterations, except for a deletion of 9p from p12and p13 towards the telomeres. Interestingly, scWGSanalysis also revealed a previously unidentified amplifica-tion on chromosome 8 in 20 out of 35 examined cells.Single-cell sequencing analysis of the intermediate aneu-ploid B-ALL B revealed that while the bulk of the cellshad the karyotype as determined by traditional cytogeneticassessment, a small number of cells had a deviating karyo-type. For instance, four cells lacked the extra copy ofchromosome 9, and one cell did not have an extra copy ofchromosome 18 (Fig. 5b, middle panel: B-ALL B). Inaddition, we identified a tumour-specific ~7.5 Mb CNVon chromosome 2 harbouring genes as ITGA4, CERKL,PPP1R1C and PDE1A (Additional file 16: Figure S9b),underscoring the potential of our platform for identifyingstructural abnormalities. Finally, the highly aneuploidB-ALL C showed a near-triploid karyotype with numer-ous whole chromosome and local copy number alterationsand aneuploidy rates similar as observed in our Mps1;p53; Lck-Cre mouse model (Fig. 5b, bottom panel: B-ALLC). Indeed, when we quantified the karyotype heterogen-eity using AneuFinder, we found that, while heterogeneityscores were relatively low for B-ALL A and B (0.0147 and0.0176, respectively), heterogeneity in B-ALL C (0.1314)was nearing the heterogeneity scores observed in our an-euploid murine T-ALL samples (ranging: 0.1032–0.3010,compare Fig. 5d to Fig. 4a), emphasising the physiologicalrelevance of our mouse model.To assess whether the observed karyotype heterogen-eity in the primary human B-ALL samples could be indi-cative for ongoing chromosomal instability, we nextengrafted B-ALL B into sub-lethally irradiated immune-deficient mice. Mice were sampled every four weeks toassess the levels of chimerism (the fraction of human-specific CD45+ cells in the peripheral blood; Additionalfile 17: Figure S10), and were sacrificed when these levelsbecame larger than 90 %. Interestingly, when we analysedthe bone marrow of these mice (B-ALL B-1 and B-2)using scWGS and AneuFinder, we found that leukaemiccell karyotypes had changed after transplantation. In bothtransplanted samples, a fraction of cells gained chromo-some 2 while none of the primary B-ALL B cells displayedchromosome 2 trisomy, and one cell in transplant B2showed unique gain of chromosome 10 (Fig. 5c). Inaddition, the cells that had gained chromosome 2 had lostthe extra copy of chromosome 9. This could be the resultof a copy neutral loss of heterozygosity (cnLOH) event.To address this, we extracted all single nucleotide poly-morphisms (SNPs) on chromosome 9 that were identifiedby our single-cell sequencing analysis in the cells(See figure on previous page.)Fig. 4 Early time point T-ALLs show similar levels of karyotype heterogeneity as endpoint lymphomas. a Aneuploidy and heterogeneity scores forthe listed samples. The black diamond indicates the ‘baseline’ aneuploidy and heterogeneity based on the Mps1f/f p53f/f Lck-Cre− control thymus.b Aneuploidy and heterogeneity scores for a control thymus, T260, and T158 plotted per chromosome. Colours of the labels indicate clustersof chromosomes that favour a euploid copy number (green), show random copy number changes (blue) or favour copy number changes (red).c Chromosome counts acquired by metaphase spreads of three independently derived T-ALL cell lines (line 1, n = 35; line 2, n = 48; line 3, n = 30). Barsare median number of chromosomes (49, 47.5 and 51 for lines 1, 2 and 3, respectively). The black dotted line indicates the euploid chromosome countof 40 for mice. d Still frame of a mitotic cell from line 1 (T302) labelled with H2B-GFP, showing a lagging chromosome (white arrowhead). Frame isdeconvolved and maximally projected. e Frequency of mitotic errors as analysed using live-cell time-lapse imaging of the H2B-GFP labelled lymphomacell line 1 (n = 32). f, g Genome-wide copy number plots for G1 (f) cells and G2/near-tetraploid cells (g) for tumour T158. h, i Genome-wide copynumber plots for G1 (h) cells and G2/near-tetraploid cells (i) for tumour T257Bakker et al. Genome Biology  (2016) 17:115 Page 10 of 15bcadFig. 5 (See legend on next page.)Bakker et al. Genome Biology  (2016) 17:115 Page 11 of 15harbouring two copies of chromosome 9. Unfortunately,we found that the coverage of the single-cell libraries wastoo low to sufficiently sample alternative alleles at definedSNP positions. To detect such events deeper sequencingis required (which increases cost price).Heterogeneity scores for both transplants revealed anincreased heterogeneity score for both transplanted leu-kaemias B1 and B2 (Fig. 5d, an increase from 0.0176[black diamond] to 0.0565 and 0.0398 [blue triangles],respectively). Even though we formally cannot rule outthat the cells with ‘new’ karyotype patterns were pre-existing in the primary leukaemia as low-abundantclones, these data strongly suggest that human tumoursalso exhibit ongoing CIN that can facilitate tumour evo-lution when tumour cells are exposed to stress, in thiscase survival and propagation in another niche (i.e. themouse hematopoietic system).Discussion and conclusionsAneuploidy is a hallmark of cancer cells and results fromchromosome missegregation events during mitosis, aprocess called chromosomal instability (CIN). Indeed,most cancer cell lines exhibit CIN with ~10 % to 60 % ofthe cells displaying lagging chromosomes in mitosis,resulting in a distribution of cells with various aneu-ploidies [28–30]. However, while measuring CIN ratesand the resulting karyotype variation in cancer cell linesis straight-forward, assessing karyotype heterogeneity inprimary tumours is technically challenging [19]. Onlyrecently, the first platforms to measure full karyotypesof single non-dividing primary cells have been re-ported [22, 23, 31].In this study we describe a whole genome single-cellsequencing pipeline to quantify aneuploidy in primary(tumour) cells in an unbiased and high throughput fash-ion. We use this pipeline to measure karyotype hetero-geneity in aneuploid murine T-ALLs that arise in ourearlier published Mps1; p53; Lck-Cre mouse model as ameasure for ongoing CIN [15]. We showed that ourpipeline confirms our earlier observations that Mps1;p53; Lck-Cre lymphomas display recurrent chromosomecopy number alterations for some chromosomes. Inaddition, we observed groups of chromosomes to showdistinct degrees of aneuploidy and heterogeneity, withspecific chromosomes favouring either aneuploidy oreuploidy, and a third group of chromosomes showingapparent random copy number changes. This lattergroup of chromosomes could be important for tumourevolution as CIN events for these chromosomes are notselected against in the primary tumours. However, thesecopy number changes might become beneficial when thetumour environment changes, for instance during therapyand metastasis and thus lead to tumour recurrence or dis-ease progression. It will therefore be extremely interestingto measure chromosome dynamics in primary tumours,metastases and recurring tumours in future experiments.Using AneuFinder, we were able to detect high-gradekaryotype heterogeneity that would likely have been ob-scured using other platforms such as array CGH. This isimportant for two reasons: (1) it confirms that chromo-some mis-segregation is ongoing in the T-ALLs thatarise in our Mps1; p53; Lck-Cre mice despite clonal se-lection for some chromosome number alterations andtherefore that the selection forces outcompete mis-segregation rates. Indeed, we confirm ongoing CIN byshowing that 50 % of the primary T-ALL cell divisionsdisplay mitotic abnormalities; and (2) it shows thatkaryotype heterogeneity as a result from high grade CINwill go unnoticed when using one of the most com-monly employed platforms to assess (aCGH). Therefore,karyotype heterogeneity might be a gravely under-estimated feature of cancer. Indeed, when we assessedkaryotype heterogeneity in three paediatric B-ALLsamples we found that heterogeneity greatly variedbetween the three assessed samples. The heterogeneitywas not directly predictive of tumour outcome, as thetwo tumours with the lowest and highest heterogeneityrate had a favourable outcome while the tumour withmedium CIN rates had a poor outcome. However, eventhough our sample size is too small to draw any conclu-sions, it is tempting to hypothesise that tumours benefitmost from medium CIN rates as low CIN rates do notallow for any karyotype evolution and too high CINrates prevent clonal selection of chromosomes when theenvironment has changed. This hypothesis appears tohold true in the mouse [32], but further studies comparinghuman tumours with known outcome are required to testthis. Importantly, when we xenografted the humananeuploid B-ALL with a medium CIN rate into mice,we found that its propagation led to additional karyo-typic changes, presumably the result of propagation inanother niche in line with our hypothesis. Therefore,we argue that ongoing CIN and the resulting karyo-typic heterogeneity are important factors to consider(See figure on previous page.)Fig. 5 Karyotype heterogeneity of human B-ALLs increases upon engraftment into recipient mice. a Overview of B-ALL patient material used inthis study. b Genome-wide copy number plots using ~1 Mb bins for bone marrow cells of three B-ALL patients. For B-ALL A and B, respectively,8 and 3 euploid cells are present (non-cancer cells). c Genome-wide copy number plots using ~1 Mb bins for bone marrow cells of two mice,28 weeks after engraftment with B-ALL B. d Aneuploidy and heterogeneity scores for the analysed B-ALL patient material and engraftments B-1and B-2. The orange triangle indicates the baseline level of aneuploidy and heterogeneity of the near-euploid B-ALL ABakker et al. Genome Biology  (2016) 17:115 Page 12 of 15when predicting disease progression/outcome andpotentially treatment response.Genomic heterogeneity is a common feature of manycancers, but because of technical limitations, it is rarelyproperly quantified. Instead, primary tumour cell popu-lation averages are used for research and diagnosticspurposes that may not entirely represent the diversity inthe primary tumour [33]. The use of single-cell sequen-cing technology allows us to better estimate this level oftumour heterogeneity [22, 33, 34]. High resolutionquantification of this heterogeneity is of the utmost im-portance as karyotype heterogeneity will drive tumour evo-lution and ultimately affect response to treatment [34–37].While the mutational landscape of many tumour types hasbeen characterised in detail, we have only begun to under-stand the role of karyotype heterogeneity in tumour evolu-tion [18, 33, 38], a role that can be further examined usingthe scWGS-AneuFinder pipeline.MethodsAnimal housing and experimentsAnimals used in this study had mixed C57BL/6 geneticbackgrounds. All mice were bred in the Central AnimalFacility (University Medical Centre Groningen [UMCG],Groningen, The Netherlands). The conditional deletionof Mps1 was described before [15]. For survival studies,mice were monitored for tumour development startingat the age of 2.5 months by looking for signs of dys-pnoea (a consequence of thymic hypertrophy) and gen-eral animal wellbeing.EthicsAnimal protocols were approved by the UMCG Committeeon Animal Care (DEC) (RUG-DEC-6369). Human sam-ples (bone marrow isolated from patients with acutelymphoblastic leukaemia) were collected as part of routinediagnostic procedures and leftover cells were used in thisstudy (METc 2013.281). The use of these leftover sampleswas approved by the Medical Ethical Committee of theUniversity Medical Center Groningen when patients (orguardians) had signed an informed consent letter. Allpatients and their guardians provided this written in-formed consent. All experimental methods comply withthe Helsinki declaration for medical research involvinghuman subjects.Array CGH data analysisFor the aCGH data analysis, probe signals were dividedby corresponding signals of a euploid control referenceand log2-transformed. The data were binned into 1 Mbsegments in steps of 500 Kb and plotted. Array CGHdata have previously been deposited in the NationalCenter for Biotechnology Information Gene ExpressionOmnibus (NCBI GEO) database under accession no.GSE57334 [15].Single-cell whole genome sequencing and dataprocessingSingle cells for the purpose of single-cell sequencingwere isolated from thymuses by dissecting the organ andmincing the tissue through a 100 μm cell strainer. Cellswere washed in PBS twice before stored in FBS with10 % DMSO at −80 °C until sorted. Prior to sorting, cellswere thawed, and incubated for 10 min on ice in thedark in a DNA-staining buffer containing 10 μg/mL pro-pidium iodide and 10 μg/mL Hoechst, as well as 0.1 %Nonidet P-40 to dissociate the cytoplasm. The G1 popu-lation was sorted as individual nuclei into a 96-well plateformat. Per sample, 10 nuclei were sorted into a singlewell to serve as a positive control, as well as an emptywell acting as a negative control. Library preparationwas performed using a modified scWGS protocol [24, 31]using a Bravo Automated Liquid Handling Platform(Agilent Technologies, Santa Clara, CA, USA). Clustersfor sequencing were generated on the cBot (Illumina).Single-end 50 bp sequencing was performed on anIllumina HiSeq 2500 at ERIBA (Illumina, San Diego, CA,USA). Raw sequencing data were demultiplexed based onlibrary-specific barcodes and converted to fastq formatusing standard Illumina software (bcl2fastq version 1.8.4).The resulting reads were mapped to mouse (mm10)or human (hg19) reference genome using Bowtie2(version 2.2.4). Duplicate reads were marked usingBamUtil (version 1.0.3). All single-cell sequencing datahave been deposited at ArrayExpress under accessionno. E-MTAB-4183.Copy number analysis using AneuFinderAnalysis of single-cell sequencing data was performedwith the AneuFinder pipeline, using a blacklist strategyto exclude reads from artefact-prone regions. For ana-lyses, we divided genomes into non-overlapping bins ofvariable size based on mappability with a mean of 1 Mb(for more details, see Additional file 5: Methods).Quality control was performed using a clustering ap-proach (minimum of 1 and maximum of 3 clusters) basedon the quality parameters described in the main text.T-ALL cell culture and live-cell time-lapse imagingMouse T-ALL cell lines were derived from primary mouselymphomas and cultured as described by Jinadasa et al.[39]. Cells were incubated at 37 °C, 5 % CO2 and lowoxygen (3 % O2). Cells were retrovirally transduced with aconstitutive H2B-GFP construct (pSF91.2 H2B-GFP) byspinfection for 90 min at 1000 × g and 32 °C in retroviralsupernatant containing polybrene 4 μg/mL (Sigma). Trans-duced cells were harvested, washed and seeded onto GlassBakker et al. Genome Biology  (2016) 17:115 Page 13 of 15Bottom Dishes (Greiner) 1 h prior to imaging. Mitoses werecaptured at 2-min intervals over the course of 16 h on aDeltaVision microscope (GE healthcare) at 100× mag-nification using SoftWorx software (GE Healthcare).In vivo leukaemia transplantationDiagnostic bone marrow cells of patients with acutelymphoblastic leukaemia were thawed and transplantedby tail vein injection into sub-lethally irradiated immunedeficient NOD/SCID/IL2Ry−/− (NSG) mice, at a dose of2.5 × 106 cells/mouse. At 4-weekly intervals, blood wasdrawn from the retro-orbital plexus and leukaemia devel-opment was followed by flow cytometric analysis using thefollowing markers against human antigens: CD19-PE,CD45-PECy7, CD34-APC, CD10-BV410 and CD20-BV605.Propidium Iodide was used to exclude dead cells. Humanchimerism was defined as >1 % hCD45+ cells among totalperipheral blood mononuclear cells (PBMC).Additional filesAdditional file 1: Figure S1. Additional comparisons of aCGH andexome-sequencing analyses of T-ALLs driven by Mps1 and p53 mutation.Six additional T-ALLs analysed using array CGH, compared to a euploidreference, showing recurrent gains of predominantly chromosomes 4, 9,14 and 15, and other lymphoma-specific alterations. (PDF 23139 kb)Additional file 2: Figure S2. Additional single-cell sequencing analysesof Mps1 T-ALL 1. Single-cell sequencing plots for Mps1 T-ALL cells 1(continuation of Fig. 1c). (PDF 63616 kb)Additional file 3: Figure S3. Quantification of single-cell sequencing dataof Mps1 T-ALL 1. a Quantification of the gains and losses of chromosomesfor 25 Mps1 T-ALL 1 cells as analysed using single-cell sequencing (Additionalfile 2: Figure S2). Losses and gains were scored as 1 s and 3 s, respectively;2 s indicates no change or disomies. The focal loss on chromosomes 7 wasscored as a loss to discriminate cells that did not show this CNV. Fourteenout of 25 cells (56 %) displayed a unique karyotype. Cells with identicalkaryotypes are clustered together, resulting in 18 groups. b Frequencypercentages of the gain, no change and loss events for chromosomes ofMps1 T-ALL 1. Gain of chromosome 4, 9, 14 and 15 are the most frequentevents, occurring in >90 % of the cells, with gain of chromosomes 2 and thefocal loss on chromosome 7 occurring in ~50 % of the cells. (PDF 548 kb)Additional file 4: Table S1. Overview of single-cell sequencing samplesand sequencing statistics. An overview of general sequencing and analysisstatistics of the single-cell sequencing data. The number of analysed readscorresponds to the number of uniquely mappable reads used in the copynumber annotation pipeline. (XLSX 38 kb)Additional file 5: Supplementary Materials and Methods. (DOCX 35 kb)Additional file 6: Table S2. Simulating the effects of differentaneuploidies on the aneuploidy and heterogeneity score. Table showingthe effect of modelling various aneuploidies on the aneuploidy andheterogeneity scores. (XLSX 42 kb)Additional file 7: Figure S4. Examples of discordant copy numbercalls between AneuFinder and Ginkgo. Top panels show the AneuFinderprofiles, bottom panels show the Ginkgo profiles, respectively. a Lowquality library showing a highly segmented fit with AneuFinder. bWrongly chosen ploidy state with Ginkgo. c Red boxes indicatechromosomes with unusually high read count dispersion whereAneuFinder fails to assign a clear copy number state. d Small copynumber change that is detected with AneuFinder but not with Ginkgo.(PDF 2236 kb)Additional file 8: Figure S5. Cumulative single-cell sequencing data ofcontrol thymus and aneuploid T-ALLs. Copy number plots showing thereads per 1 Mb of cumulative single-cell sequencing data analysed assimulated bulk data, showing an obscuring effect on the karyotypeheterogeneity. (PDF 3267 kb)Additional file 9: Figure S6. Single-cell sequencing of early time pointT-ALLs. Genome-wide copy number plots using ~1 Mb bins for threethymuses harvested from 10-, 13- and 14-week-old mice, showing highlevels of karyotype heterogeneity at 13 and 14 weeks. (PDF 451 kb)Additional file 10: Figure S7. Aneuploidy and heterogeneity perchromosome observed in a control thymus and T-ALLs. Aneuploidy andheterogeneity scores plotted per chromosomes of all T-ALLs examined inthe study. Chromosomes indicated in green do not favour copy numberchange and show minimal heterogeneity. Chromosomes in blue showapparent random copy number changes. Red chromosomes favour copynumber changes. (PDF 440 kb)Additional file 11: Time-lapse imaging of a T cell labelled with H2B-GFP,showing a lagging chromosome. (MOV 3783 kb)Additional file 12: Time-lapse imaging of a T cell labelled with H2B-GFP,showing tetraploidisation. (MOV 5675 kb)Additional file 13: Time-lapse imaging of a T cell labelled with H2B-GFP,showing failed alignment. (MOV 2525 kb)Additional file 14: Time-lapse imaging of a T cell labelled with H2B-GFP,showing tetraploidisation followed by cell death. (MOV 20714 kb)Additional file 15: Figure S8. Single-cell sequencing of (near)-4n cellsin T158 and T257. a PI/Hoechst FACS plots showing for four tumours,showing apparent cycling tetraploid cells in T158 and T257. b Comparisonof AneuFinder copy number calling of T158; comparing the fit when forcingAneuFinder to call the majority state tetrasomy (left) or disomy (right).c Comparison of AneuFinder copy number calling of T257; comparing thefit when forcing AneuFinder to call the majority state tetrasomy (left) ordisomy (right). (PDF 4660 kb)Additional file 16: Figure S9. Additional scWGS data for human B-ALLs. a Copy number plots showing the reads per 1 Mb of cumulativesingle-cell sequencing data analysed as simulated bulk, showing anobscuring effect on the karyotype heterogeneity. b Genomic context ofthe CNV on chromosome 2 in B-ALL B. (PDF 1964 kb)Additional file 17: Figure S10. Chimerism levels for B-ALL B over time.Chimerism is defined as >1 % hCD45+ peripheral blood mononuclear cells(PBMCs). Plotted is the percentage of hCD45+ PBMCs at 4-week intervalsB-ALL B (n= 2). Mice engrafted with B-ALL B showed ~80 % chimerism after28 weeks at which time they were sacrificed. (PDF 324 kb)AcknowledgementsWe are grateful to the other members of the Foijer and Lansdorp groups fortheir technical support and fruitful discussions. We thank Maria Sarkis Azkanaz forher assistance with analysing time-lapse microscopy data. Furthermore, we thankthe operators at the Central Flow Cytometry Unit (UMCG), Geert Mesander, HenkMoes and Roelof Jan van der Lei, for their assistance with sorting, and the animalcaretakers at the Central Animal Facility (UMCG). The authors are indebted to allpatients who provided consent for the use of their material in this study.FundingThis work was supported by a Dutch Cancer Society grant (KWF grantRUG-2012-5549) and a Groningen Foundation for Paediatric Oncology (SKOG)grant to FF and PML, and by an NWO (The Netherlands Organisation forScientific Research) MEERVOUD grant and a University of Groningen RosalindFranklin Fellowship grant to MCT.Availability of data and materialsAll array CGH and sequencing data have been deposited at public databases.aCGH data can be accessed from the Gene Expression Omnibus (NCBI GEO)database (http://www.ncbi.nlm.nih.gov/geo/) under accession no. GSE57334[15]. Sequencing data can be downloaded from ArrayExpress (https://www.ebi.ac.uk/arrayexpress/) under accession no. E-MTAB-4183. AneuFinderis available through BioConductor at: http://bioconductor.org/packages/AneuFinder/ and is licensed under Artistic-2.0Bakker et al. Genome Biology  (2016) 17:115 Page 14 of 15Authors’ contributionsBB, AT, MCT and FF conceived the study. BB and MEB performed experimentsand ESJMB provided human B-ALL samples. AT, BB, MCT and DP designed andtrained the AneuFinder pipeline. BB, AT, MCT and FF analysed data.DCJS, NH, HGK and KHW performed single-cell library preparation and providedsequencing support under the supervision of DCJS and PML. AT, DP, TVJ, VGand MCT provided bioinformatics support. PML, MCT and FF provided financialsupport. BB, AT, MCT and FF prepared the manuscript with input from MEB, DP,DCJS, PML and TVJ. All authors read and approved the manuscript.Competing interestsThe authors declare that they have no competing interests.Author details1European Research Institute for the Biology of Ageing, University ofGroningen, University Medical Center Groningen, A. Deusinglaan 1,Groningen 9713 AV, The Netherlands. 2Wellcome Trust Sanger Institute,Wellcome Trust Genome Campus, Hinxton CB10 1SA, UK. 3Department ofPaediatrics, University of Groningen, University Medical Center Groningen, A.Deusinglaan 1, Groningen 9713 AV, The Netherlands. 4Department ofPathology & Medical Biology, University of Groningen, University MedicalCenter Groningen, A. Deusinglaan 1, Groningen 9713 AV, The Netherlands.5Terry Fox Laboratory, BC Cancer Agency, Vancouver, BC V5Z 1L3, Canada.6Division of Hematology, Department of Medicine, University of BritishColumbia, Vancouver, BC V6T 1Z4, Canada. 7Institute of ComputationalBiology, Helmholtz Zentrum München, Ingolstädter Landstr. 1, Neuherberg85764, Germany.Received: 29 December 2015 Accepted: 29 April 2016References1. Williams BR, Prabhu VR, Hunter KE, Glazier CM, Whittaker CA, Housman DE,et al. Aneuploidy affects proliferation and spontaneous immortalization inmammalian cells. Science. 2008;322:703–9.2. Torres EM, Sokolsky T, Tucker CM, Chan LY, Boselli M, Dunham MJ, et al.Effects of aneuploidy on cellular physiology and cell division in haploidyeast. Science. 2007;317:916–24.3. Baker DJ, Jeganathan KB, Cameron JD, Thompson M, Juneja S, Kopecka A,et al. BubR1 insufficiency causes early onset of aging-associated phenotypesand infertility in mice. Nat Genet. 2004;36:744–9.4. Baker DJ, Jeganathan KB, Malureanu L, Perez-Terzic C, Terzic A, van Deursen JM.Early aging-associated phenotypes in Bub3/Rae1 haploinsufficient mice. J CellBiol. 2006;172:529–40.5. Baker DJ, Dawlaty MM, Wijshake T, Jeganathan KB, Malureanu L, van Ree JH,et al. Increased expression of BubR1 protects against aneuploidy and cancerand extends healthy lifespan. Nat Cell Biol. 2013;15:96–102.6. Sheltzer JM, Amon A. The aneuploidy paradox: costs and benefits of anincorrect karyotype. Trends Genet. 2011;27:446–53.7. Giam M, Rancati G. Aneuploidy and chromosomal instability in cancer: ajackpot to chaos. Cell Div. 2015;10:3.8. Siegel JJ, Amon A. New insights into the troubles of aneuploidy. Annu RevCell Dev Biol. 2012;28:189–214.9. Foijer F. Mouse models for chromosomal instability. In: Storchova Z, editor.Aneuploidy in health and disease. Rijeka: Intech; 2012. p. 59–78.10. Foijer F. CINister thoughts. Biochim Biophys Acta. 2010;38:1–7.11. Simon J, Bakker B, Foijer F. CINcere modelling: What have mousemodels for chromosome instability taught us? Results Cancer Res. 2015;200:39–60.12. Musacchio A, Salmon ED. The spindle-assembly checkpoint in space andtime. Mol Cell. 2007;8:379–93.13. Kops GJPL, Shah JV. Connecting up and clearing out: how kinetochoreattachment silences the spindle assembly checkpoint. Chromosoma. 2012;121:509–25.14. Vleugel M, Hoogendoorn E, Snel B, Kops GJPL. Evolution and function ofthe mitotic checkpoint. Dev Cell. 2012;23:239–50.15. Foijer F, Xie SZ, Simon JE, Bakker PL, Conte N, Davis SH, et al. Chromosomeinstability induced by Mps1 and p53 mutation generates aggressivelymphomas exhibiting aneuploidy-induced stress. Proc Natl Acad Sci.2014;111:13427–32.16. Nicholson JM, Cimini D. Cancer karyotypes: survival of the fittest. FrontOncol. 2013;3:148.17. Gerlinger M, Rowan AJ, Horswell S, Larkin J, Endesfelder D, Gronroos E,et al. Intratumor heterogeneity and branched evolution revealed bymultiregion sequencing. N Engl J Med. 2012;366:883–92.18. Heng HHQ, Liu G, Stevens JB, Abdallah BY, Horne SD, Ye KJ, et al. Karyotypeheterogeneity and unclassified chromosomal abnormalities. CytogenetGenome Res. 2013;139:144–57.19. Bakker B, van den Bos H, Lansdorp PM, Foijer F. How to countchromosomes in a cell: An overview of current and novel technologies.BioEssays. 2015;37:570–7.20. McGranahan N, Burrell RA, Endesfelder D, Novelli MR, Swanton C. Cancerchromosomal instability: therapeutic and diagnostic challenges. EMBO Rep.2012;13:528–38.21. Macaulay IC, Voet T. Single cell genomics: advances and future perspectives.PLoS Genetics. 2014;10:e1004126.22. Navin N, Kendall J, Troge J, Andrews P, Rodgers L, McIndoo J, et al. Tumourevolution inferred by single-cell sequencing. Nature. 2011;472:90–4.23. Knouse K, Wu J, Whittaker C, Amon A. Single cell sequencing revealslow levels of aneuploidy across mammalian tissues. Proc Natl Acad Sci.2014;111:13409–14.24. van den Bos H, Spierings DCJ, Taudt AS, Bakker B, Porubský D, Falconer E, etal. Single-cell whole genome sequencing reveals no evidence for commonaneuploidy in normal and Alzheimer's disease neurons. Genome Biol. 2016.doi:10.1186/s13059-016-0976-2.25. Baum L, Petrie T, Soules G, Weiss N. A Maximization technique occurring inthe statistical analysis of probabilistic functions of Markov chains. Ann MathStat. 1970;41:164–71.26. Bhattacharyya A. On a measure of divergence between two statisticalpopulations defined by their probability distributions. Bull Calcutta MathSoc. 1943;35:239–58.27. Garvin T, Aboukhalil R, Kendall J, Baslan T, Atwal GS, Hicks J, et al.Interactive analysis and assessment of single-cell copy-numbervariations. Nat Methods. 2015;12:1058–60.28. Thompson SL, Compton DA. Examining the link betweenchromosomal instability and aneuploidy in human cells. J Cell Biol.2008;180:665–72.29. Ganem NJ, Godinho SA, Pellman D. A mechanism linking extracentrosomes to chromosomal instability. Nature. 2009;460:278–82.30. Silkworth WT, Nardi IK, Scholl LM, Cimini D. Multipolar spindlepole coalescence is a major source of kinetochore mis-attachmentand chromosome mis-segregation in cancer cells. PLoS One.2009;4:e6564.31. Falconer E, Hills M, Naumann U, Poon SSS, Chavez EA, Sanders AD, etal. DNA template strand sequencing of single-cells mapsgenomic rearrangements at high resolution. Nat Methods.2012;9:1107–12.32. Silk AD, Zasadil LM, Holland AJ, Vitre B, Cleveland DW, Weaver BA.Chromosome missegregation rate predicts whether aneuploidywill promote or suppress tumors. Proc Natl Acad Sci U S A.2013;110:E4134–41.33. Abdallah BY, Horne SD, Stevens JB, Liu G, Ying AY, Vanderhyden B, et al.Single cell heterogeneity: Why unstable genomes are incompatible withaverage profiles. Cell Cycle. 2013;12:3640–9.34. Navin N, Hicks J. Future medical applications of single-cell sequencing incancer. Genome Med. 2011;3:31.35. Alizadeh AA, Aranda V, Bardelli A, Blanpain C, Bock C, Borowski C, et al.Toward understanding and exploiting tumor heterogeneity. Nat Med.2015;21:1–8.36. Sotillo R, Hernando E, Díaz-Rodríguez E, Teruya-Feldstein J, Cordón-Cardo C,Lowe SW, et al. Mad2 overexpression promotes aneuploidy andtumorigenesis in mice. Cancer Cell. 2007;11:9–23.37. Laughney AM, Elizalde S, Genovese G, Bakhoum SF. Dynamics of tumorheterogeneity derived from clonal karyotypic evolution. Cell Rep.2015;12:1–12.38. Saadatpour A, Lai S, Guo G, Yuan G-C. Single-cell analysis in cancergenomics. Trends Genet. 2015;31:576–86.39. Jinadasa R, Balmus G, Gerwitz L, Roden J, Weiss R, Duhamel G. Derivation ofthymic lymphoma T-cell lines from Atm(−/−) and p53(−/−) mice. J Vis Exp.2011;50:6–8.Bakker et al. Genome Biology  (2016) 17:115 Page 15 of 15


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items