Evolution of human BCR–ABL1
lymphoblastic leukaemia-initiating cells
Faiyaz Notta1,2*, Charles G. Mullighan3*, Jean C. Y. Wang1,4, Armando Poeppl1, Sergei Doulatov1,2, Letha A. Phillips3, Jing Ma5,
Mark D. Minden4, James R. Downing3& John E. Dick1,2
Many tumours are composed of genetically diverse cells; however, little is known about how diversity evolves or the
impact that diversity has on functional properties. Here, using xenografting and DNA copy number alteration (CNA)
profiling of human BCR–ABL1 lymphoblastic leukaemia, we demonstrate that genetic diversity occurs in functionally
defined leukaemia-initiating cells and that many diagnostic patient samples contain multiple genetically distinct
leukaemia-initiating cell subclones. Reconstructing the subclonal genetic ancestry of several samples by CNA
profiling demonstrated a branching multi-clonal evolution model of leukaemogenesis, rather than linear succession.
For some patient samples, the predominant diagnostic clone repopulated xenografts, whereas in others it was
outcompeted by minor subclones. Reconstitution with the predominant diagnosis clone was associated with more
aggressive growth properties in xenografts, deletion of CDKN2A and CDKN2B, and a trend towards poorer patient
outcome. Our findings link clonal diversity with leukaemia-initiating-cell function and underscore the importance of
developing therapies that eradicate all intratumoral subclones.
A widely accepted tenet of cancer biology is that most tumours arise
fromsinglecells and that multiple genetic alterationsaccumulate over
time, resulting in transformation1. Historically, this process was con-
sidered to be a stepwise acquisition of new mutations, some of which
this model of tumour evolution, all clones are linearly related to each
other.However, new genomic technologies are revealing a more com-
plex clonal architecture in some cancers3–6. Analysis of chromosomal
translocation breakpoints and CNA profiling in twins with ETV6–
RUNX1-positive acute lymphoblastic leukaemia (ALL) showed that a
and then evolves with different kinetics and CNA acquisition in each
twin7,8. Genome-wide CNA profiling of paired diagnostic and relapse
ALL samples has been particularly informative9–11. In most cases, the
relapse clone shared only limited genetic identity with the predomi-
clone was either identical or a direct evolutionary product of the dia-
gnostic clone. These studies predicted the existence of an ancestral,
pre-diagnostic clone that gave rise to at least two clonal lineages that
evolved independently in many patients with ALL, with each clone
acquiring different genetic aberrations: one clone giving rise to the
clone at relapse with the acquisition of additional CNA. These results
indicate that tumour evolution may occur through a more complex
branching model that gives rise to genetically distinct subclones at
diagnosis that vary in aggressiveness and response to therapy12.
However, proof of this model requires studies directly examining the
functional properties of the cells in which genetic changes are found.
often present and thus incapable of contributing to long-term clonal
maintenance or to relapse. Therefore, functional studies of the cells
responsible for driving leukaemic growth in patients must be com-
and theirevolutionaryancestralprecursors,andtoshow whetherthey
possess biologically distinct growth properties.
Arguably the most important biological function a cancer cell can
by tumour initiation assays in primary and secondary recipients.
Indeed, some highly aggressive or metastatic tumours of mice and
humans seem to be functionally homogeneous because almost every
cell has tumour-initiating-cell capacity. However, most tumours
or leukaemia-initiating cells typically represent a minor fraction,
although their frequency varies widely in syngeneic14or xenograft15
recipients. It has been widely considered that intratumoral functional
heterogeneity results from stochastic processes that influence cell
growth but also from the variable behaviour of genetic subclones that
arise through clonal evolution. The cancer stem cell model proposes
an alternative explanation based on the hierarchical organization of
the tumour clone where cancer stem cells are solely responsible for
driving clonal growth and for therapeutic resistance. In the cancer
stem cell model, tumour-initiating cells and cancer stem cells are
synonymous and have the properties of self-renewal and maturation
that are canonical to all stem cells. Epigenetic or developmental pro-
grams contribute to functional differences between cancer stem cells
and non-cancer stem cells within a tumour clone that the model
assumes would be genetically identical13,16. The cancer stem cell and
clonal evolution models are the subject of intense debate and often
considered to be mutually exclusive16,17. The cancer stem cell model
focuses on the concept of functional heterogeneity but does not take
into account tumour evolution, intratumoral genetic variation, or the
evolution model focuses on genetic heterogeneity without considering
1Division of Stem Cell and Developmental Biology, Campbell Family Institute for Cancer Research/Ontario Cancer Institute, Toronto, Ontario M5G 1L7, Canada.2Department of Molecular Genetics,
University of Toronto, Toronto, Ontario M5G 1L7, Canada.3Department of Pathology, St Jude Children’s Research Hospital, Memphis, Tennessee 38105, USA.4Department of Medical Oncology and
Research Hospital, Memphis, Tennessee 38105, USA.
*These authors contributed equally to this work.
3 6 2 | N A T U R E | V O L 4 6 9 | 2 0 J A N U A R Y 2 0 1 1
Macmillan Publishers Limited. All rights reserved
the functional variation that might exist intratumorally within indi-
vidual genetic subclones. As a first step to resolve the basis for intra-
clones of tumour-initiating cells.
between intratumoral clonal diversity, genetic alterations and cellular
growth properties because it is considered a single clinical entity with
identifiable and recurrent genetic abnormalities. Detailed studies have
revealed a number of genetic alterations, notably deletions of the
lymphoid transcriptional regulator IKAROS (also called IKZF1),
PAX5, EBF1, as well as deletions involving CDKN2A/B that cooperate
with BCR–ABL1 in lymphoid leukaemogenesis18. Furthermore, the
association of IKZF1 deletion in Ph-negative ALL with poor patient
outcome predicts that it will be possible to link specific genetic altera-
xenograft system that was used to carry out a combined genetic and
initiating cells derived from diagnostic patient samples.
Modelling human Ph1ALL in xenografts
To determine whether a single genetic subtype of leukaemia like Ph1
ALL exhibits uniform growth properties, we used three xenograft
models of increasing immune deficiency: NOD.CB17-Prkdcscid/J
(NOD/SCID) mice; NOD/SCID mice treated with anti-CD122 to
deplete innate immune cells (NS122)20,21; or NOD/SCID mice with
deletion of the common gamma (c)-chain (NSG)22(Supplementary
Fig. 1). Diagnosis samples from 18 of 20 Ph1ALL patients efficiently
engrafted NS122 mice (Supplementary Fig. 2) and recapitulated
numerous aspectsof thehuman diseaseincluding tumour dissemina-
tion, immunophenotype (Supplementary Fig. 3) and morphology
(Supplementary Fig. 4). However, 10 of 20 patient samples caused
clinically manifest disease before 15weeks and were categorized as
aggressive group 1 samples, whereas the remaining xenograft mice
appeared healthy until they were killed and these were classified as
non-aggressive group 2 samples (Fig. 1a). Accordingly, the leukaemic
burden in bone marrow and systemic dissemination was significantly
higher in group 1 versus group 2 samples (Fig. 1b, c, Supplementary
Fig. 2b and Supplementary Table 1). Notably, group 1 samples
engrafted all recipient types (Fig. 1d and Supplementary Fig. 5) even
when transplanted at near-limiting dose (Supplementary Fig. 6). By
contrast, group 2 samples failed to engraft to NOD/SCID mice
(Fig. 1d and Supplementary Fig. 5) despite injection of 50-fold more
cells (Supplementary Fig. 7). Of twogroup 2 samples (patient 2-5and
2-6) unable to engraft NS122 recipients, one engrafted NSG mice
(Supplementary Fig. 8). The extent of leukaemic dissemination was
similar in NS122 or NSG mice for both group 1 and group 2 samples,
although NSG mice had higher peripheral engraftment levels (Sup-
plementary Fig. 9).
Genetic basis of functional heterogeneity
To examine a possible genetic basis for the distinct xenograft growth
properties between group 1 and group 2 samples, genome-wide CNA
profiling was undertaken. Overall, the frequency of genetic alterations
2 samples had focal and complete deletions of the IKZF1 locus.
However, there were marked differences in the proportion with
deletions of CDKN2A/B (group 1, 90%; group 2, 0%; P50.0001)
and PAX5 (group 1, 60%; group 2, 10%; P50.057) genes (Fig. 2e
and Supplementary Table 2). Genomic quantitative polymerase
chain reaction (qPCR) confirmed that the CDKN2A/B locus was not
hypermethylated in an independent cohort of Ph1ALL patients
(Supplementary Fig. 10).
Because previous studies have reported a positive association
between the efficiency of xenograft engraftment and clinical outcome
of group 1 and group 2 samples. We found a trend towards poorer
outcome of group 1 patients with increased early relapse, although
significance was not reached owing to the small sample number
(Fig. 2f, P50.08). Limiting dilution analyses (LDA) of 11 patient
samples showed that the leukaemia-initiating-cell frequency in group
1 samples was 80-fold higher than group 2 samples (Fig. 2g). The
leukaemia-initiating-cell frequency in one group 1 patient was 11%,
similar to the leukaemia-initiating-cell frequency previously observed
in comparable murine models24. Although comparison of absolute
ents is subject to some uncertainty25, the relative difference between
group 1 and group 2 samples is consistent with their distinct growth
properties. Moreover, preliminary evidence indicates that leukaemic
evolution, defined by increasingly aggressive and less restrictive xeno-
graft growth upon serial passage, also correlates with reduced func-
tional heterogeneity as reflected by increased leukaemia-initiating-cell
frequency (Supplementary Fig. 11 and Supplementary Table 4).
Collectively, we provide the first data linking engraftment properties
and leukaemia-initiating-cell frequency to both specific genetic events
in cancer and clinical outcome of patients.
Clonal dynamics of Ph1ALL pathogenesis
Despite widespread use of tumour xenografts, there are few studies
comparing genetic alterations in primary samples versus xenografts6.
To determine whether genetic abnormalities of the diagnostic sample
are propagated upon transplantation, we tracked the clonal dynamics
of ALL growth in xenografts by comparing CNA profiles of 12 dia-
gnosis samples (eight group 1 and four group 2 tumours) with paired
primary and secondary xenografts (Fig. 2a). Overall, xenografts did
tumours (Supplementary Fig. 12 and Supplementary Table 5). In six
samples, five of which were group 1, xenografts exhibited the same
distribution of CNA as the diagnostic patient sample, and detailed
analysis of the antigen receptor (AgR) loci confirmed that the pre-
dominant clone present at diagnosis was propagated in xenografts
(Supplementary Fig. 13).
By contrast, multiple xenografts derived from the six other patient
samples (three group 1 and three group 2) harboured distinct genetic
changes compared to the predominant diagnostic clone (Fig. 2b, c and
Supplementary Fig. 14), while also sharing major CNA such as IKZF1
and CDKN2A/B (data not shown). The presence of multiple xenograft
recipients from the same patient sample with both identical and new
in the diagnostic sample, that harbour additional genetic alterations.
and relapse ALL patient samples where the majority of relapsed cases
represent the evolution of a new clone that is related to, but distinct
in xenografts (CNA concordant) was different from patients where
xenografts were engrafted with a minor subclone that outcompeted
the predominant clone (CNA discordant). CNA concordant samples
To determine whether the detection of minor subclones might be
hindered by out-competition in xenografts of dominant or aggressive
clones, we analysed the CNA profiles of two patient samples trans-
planted at limiting (to engraft single leukaemia-initiating cells) and at
non-limiting cell doses. In patient 1-8, non-limiting (bulk) cell doses
2 0 J A N U A R Y 2 0 1 1 | V O L 4 6 9 | N A T U R E | 3 6 3
Macmillan Publishers Limited. All rights reserved
generated xenografts that differed at three major CNA compared to
the patient sample (chromosome (Chr) 1 gain, Chr 8p deletion and
Chr 8q gain) (Fig. 2e). Xenografts derived from limiting cell doses
(clonal) differed from both non-limiting xenografts (lacked Chr 1
gain and 8p deletion) and the patient sample (Chr 8q gain) (Fig. 2e,
m6). Patient sample 1-1 was highly aggressive in xenografts. All reci-
contained the major diagnostic clone distinguished with a bi-allelic
dose group did not have this CNA but retained the larger flanking
CNA(Fig. 2fandSupplementaryFig.15, m2).Thetopographyof this
lesion from each clone, together with the similarity of AgR regions in
Each clone remained stable after secondary transplantation (Sup-
plementary Fig. 16). Thus, our data provide formal evidence that
ALL is composed of genetically distinct subclones that are present in
events that probably occurred in these patients and provides unique
Multi-clonal model of Ph1ALL pathogenesis
To gain an insight into the evolutionary processes that underlie the
0 100 200 300 400 500 600
P = 0.083
Group (no. patients)
7 × 103
4 × 105
05 10 15 20 25 30
Survival (NS122) (%)
Group 1 Group 2
1-1*1-2* 1-3* 1-4 1-5*
2-12-2 2-3 2-42-5 2-6* 2-7*2-10
Group 1 Group 2
Leukaemia blast (%)
Leukaemia blast (%)
Figure 1 | Functional and genetic analysis of Ph1ALL. a, Survival ofNS122
mice transplanted with 20 diagnostic Ph1ALL samples. Xenografts moribund
samples, 1-2 and 2-8, determined by flow cytometry of the injected femur (IF),
bone marrow (BM, left femur/tibiae), spleen (SP) and peripheral blood (PB)
(n54 mice per sample; error bars, mean6s.e.m.). c, Cumulative leukaemia
engraftment in xenografts from panel a transplanted with group 1 and group 2
samples (error bars, mean6s.e.m.; ***P,0.0001). d, Comparison of
SCID or NS122/NSG recipients. Human CD45 or CD44 (Supplementary Fig.
group 2 patient samples using Affymetrix 6.0 SNP arrays. Regions containing
IKZF1 (top panel), CDKN2A/B (middle panel) and PAX5 (bottom panel)
3–6, consistent with formation of dominant-negative IKAROS isoform IK6.
group 2 patients (P50.083). g, Minimum cell dose required for leukaemia
initiation in NS122 and NSG recipients from group 1 (n55) versus group 2
(n56) patient samples.
3 6 4 | N A T U R E | V O L 4 6 9 | 2 0 J A N U A R Y 2 0 1 1
Macmillan Publishers Limited. All rights reserved
analysis obtained from xenografts with clonal analysis carried out by
CNA profiling on multiple xenografts derived from a group 1 and
group 2 sample. For both patient samples, the dominant diagnostic
was not detected in xenografts, rather they were repopulated with
several related but distinct genetic subclones. Detailed tracking of
CNA provided an unprecedented opportunity to gain an insight into
the sequence of lesion acquisition in independent subclones. For
example, in patient 1-6, deletion of the AgR region of Chr 11 in all
genesis that was shared by all subclones that outgrew after transplant,
Patient 2-9 displayed a CNA (gain) in Chr Xp (Fig. 3b, Chr X) at
Only two recipients contained a subsequent Chr 9q deletion (m37 and
Chr 8q, but lacked a deletion in chromosome 9. Therefore, it remains
unclear which CNA (Chr 9 deletion or Chr 8q duplication) was
acquired first in the patient. Data are summarized pictorially for both
patient samples (Fig. 3, right panel). These data indicate that multiple
tumour clones coexist in the diagnostic patient sample, and that these
ing the branching model of tumour progression12.
Intratumoral heterogeneity may promote clonal evolution by
increasing the number of selectable traits under any given stress.
Therefore, genetic diversification is probably important for tumour
to clinical aggressiveness of breast tumours26, and is associated with
metastasis in pancreatic cancer4,5. To determine whether the various
primary xenografts were transplanted into secondary recipients. All
CNA from primary xenografts were detected in secondary recipients,
indicating overall stability of each subclone (Supplementary Fig. 17a).
However, new CNA were detected in three of nine recipients, indi-
cating that ongoing evolution and further progression of disease can
occur, although it seems to be largely stochastic (Supplementary Fig.
17b). Although limiting dilution studies were not performed to com-
pletely rule out the possibility that additional minor subclones,
undetected in the primary xenograft, contributed to these new CNA,
these data indicate that genetic diversification can continue in the
Here we establish that individual Ph1ALL samples at diagnosis are
composed of genetically diverse subclones that are related through a
complex evolutionary process. These subclones vary in their xenograft
Ph1ALL patient samples can be segregated into two subgroups on the
basis of functional xenograft growth properties and specific genetic
0201,000 2,000 3,000 4,000
P = 0.04
m1 m2 m3
Figure 2 | Clonal dynamics of Ph1ALL upon transplant into xenografts.
a, Schema of experimental design for tracking primary leukaemia clones in
xenografts using limiting and bulk cell doses. b, Representative patient sample
displaying a new CNA in each of three xenografts in the HBS1L locus that was
below SNP array detection limit in the patient sample (additional examples on
thatgeneratedCNAconcordantanddiscordantxenografts.e, f, CNAprofiling
of xenografts transplanted with limiting and bulk doses of patient cells. In
patient 1-8 (e), multiple CNA present on Chr 1 (middle) and Chr 8 (bottom)
were detected in all recipients (m1–m4) at bulk cell doses (1 3106per mouse)
that were not detected in the diagnostic sample. At limiting cell dose, one of
panel) and the diagnostic sample (Chr 7, AgR deletion). In patient 1-1 (f), two
of nine mice were engrafted at limiting cell doses (50 cells per mouse; m1, m2)
and a single recipient at non-limiting cell dose (,13106cells; m3). The
m2 (bottom panel). All samples share common CNA at the AgR locus (top
panel). Data are log2ratio, median smoothing format (blue, deletion; white,
normal; red, gain).
2 0 J A N U A R Y 2 0 1 1 | V O L 4 6 9 | N A T U R E | 3 6 5
Macmillan Publishers Limited. All rights reserved
survival correlated with aggressive dissemination in xenografts and
higher leukaemia-initiating-cell frequency. These results are consistent
with the aggressive, tyrosine-kinase-inhibitor-resistant Ph1ALL seen
genetically distinct samples and subclones already possess variably
aggressive growth properties points to the need to develop effective
therapies to eradicate all intratumoral genetic subclones to prevent
further evolution and recurrence. The ability to segregate even minor
subclones in xenografts is a powerful tool for the preclinical develop-
ment of new therapeutic strategies.
an opportunity to reconstruct the functional genetic ancestry of sub-
clones present in diagnostic samples. Our data illustrate that leukaemic
progression can occur in either a linear or branching fashion, with mul-
tively. The xenograft growth characteristics of minor subclones were
distinct: sometimes they out-competed the dominant clone, whereas
the dominant clone also appeared to associate with poorer clinical
outcome, although the size of the cohort was small. We speculate that
some genetic events, such as loss of CDKN2A/B, contribute to clonal
trast, the reduced competitive advantage of minor subclones indicates
However, if minor subclones survive therapy, further evolution and
expansion could occur, leading to future relapse, consistent with pre-
gene silencing and other epigenetic events contribute to tumour pro-
subclonal complexity and underscores the importance of gaining a
better molecular understanding of each subclone within a tumour.
Outgrowth of subclones in serial xenografts can only be sustained
by leukaemia-initiating cells, and our findings establish that genetic
discovery that specific genetic events influence leukaemia-initiating-
cell frequency and that genetically distinct leukaemia-initiating cells
evolve through a complex evolutionary process indicates that a close
connection must exist between genetic and functional heterogeneity.
our findings in the absence of prospective isolation proving the exist-
ence of leukaemia stem cells in Ph1ALL; however, strong evidence is
accumulating for the existence and relevance of leukaemia stem cells
in other forms of leukaemia27and it is likely that genetically diverse
leukaemia stem cells will eventually be found. If the leukaemia-
and represent units of selection in tumour evolution. As tumours
evolve, the frequency of leukaemia stem cells increases, eventually
progressing to a highly advanced state that might no longer adhere
to a cancer stem cell model. The high leukaemia-initiating-cell fre-
quency that we observed in some samples and from several murine
models16,24supports this idea. Finally, as tumours are composed of
genetically diverse subclones, prospective isolation of leukaemia stem
could segregate genetically distinct subclones with variable tumour-
initiating-cell capacity as opposed to genetically identical cells with
differing epigenetic/developmental programs. Because the hierarchy
model posits that cancer stem cells give rise to non-cancer stem cells,
future studies must account for subclonal diversity and establish the
our findings indicate that there may be more commonalities between
thought and that future studies may lead to a unification of these
Diagnostic Ph1ALL patient samples were intrafemorallytransplanted intofemale
NOD.CB17-Prkdcscid/J (NOD/SCID) mice, NOD/SCID mice treated with mouse
mice. Xenograft recipients were monitored for disease sickness, and chimaerism
was evaluated in various haematopoietic tissues using flow cytometry. DNA copy
Subclone phylogeny (diagnosis)
m37 m38 m39 m40 m41
Figure 3 | Detection of genetically diverse leukaemia-initiating cells in Ph1
ALL. a, CNA profiling shows that three engrafted recipients from sample 1-6
share a common CNA on Chr 11 (top) that is not detected in the diagnostic
sample, but each is distinct for CNA on Chr 6 (middle) and Chr 14 (bottom).
b, In sample 2-9, all five engrafted recipients shared a deletion (at varying
degrees) in a region of Chr Xp (top), two recipients displayed a common CNA
regionofChr9q(middle;m37, m39), andthreerecipientshaddeletions(m37)
andduplications(m39,41)onregionsofChr8p (bottom). Aphylogenetictree
depicting the relationship between major and minor genetic subclones present
shown to the right of each sample. The dashed line represents clonal evolution
from disease origin (Ph, Philadelphia chromosome).
3 6 6 | N A T U R E | V O L 4 6 9 | 2 0 J A N U A R Y 2 0 1 1
Macmillan Publishers Limited. All rights reserved
number alteration (CNA) was carried out with Affymetrix 6.0 SNP arrays on the
diagnostic patient sample and corresponding xenografts.
Full Methods and any associated references are available in the online version of
the paper at www.nature.com/nature.
Received 10 June; accepted 3 December 2010.
Hanahan, D. & Weinberg, R. A. The hallmarks of cancer. Cell 100, 57–70 (2000).
Nowell, P. C. The clonal evolution of tumor cell populations. Science 194, 23–28
Barrett, M. T. et al. Evolution of neoplastic cell lineages in Barrett oesophagus.
Nature Genet. 22, 106–109 (1999).
Campbell, P. J. et al. The patterns and dynamics of genomic instability in
metastatic pancreatic cancer. Nature 467, 1109–1113 (2010).
Yachida, S. et al. Distant metastasis occurs late during the genetic evolution of
pancreatic cancer. Nature 467, 1114–1117 (2010).
Ding, L. et al. Genome remodelling in a basal-like breast cancer metastasis and
xenograft. Nature 464, 999–1005 (2010).
Bateman, C. M. et al. Acquisition of genome-wide copy number alterations in
monozygotic twins with acute lymphoblastic leukemia. Blood 115, 3553–3558
Hong, D. et al. Initiating and cancer-propagating cells in TEL-AML1-associated
childhood leukemia. Science 319, 336–339 (2008).
Li, A. et al. Sequence analysis of clonal immunoglobulin and T-cell receptor
gene rearrangements in children with acute lymphoblastic leukemia at
diagnosis and at relapse: implications for pathogenesis and for the clinical
utility of PCR-based methods of minimal residual disease detection. Blood 102,
10. Zuna, J. et al. TEL deletion analysis supports a novel view of relapse in childhood
acute lymphoblastic leukemia. Clin. Cancer Res. 10, 5355–5360 (2004).
11. Mullighan, C. G. et al. Genomic analysis of the clonal origins of relapsed acute
lymphoblastic leukemia. Science 322, 1377–1380 (2008).
12. Greaves, M. Cancer stem cells: back to Darwin? Semin. Cancer Biol. 20, 65–70
13. Dick, J. E. Stem cell concepts renew cancer research. Blood 112, 4793–4807
14. Bruce, W. R. & Van Der Gaag, H. A quantitative assay for the number of murine
lymphoma cells capable of proliferation in vivo. Nature 199, 79–80 (1963).
15. Diehn, M., Cho, R. W. & Clarke, M. F. Therapeutic implications of the cancer stem
cell hypothesis. Semin. Radiat. Oncol. 19, 78–86 (2009).
cancer stem cells versus clonal evolution. Cell 138, 822–829 (2009).
Biophys. Acta 1805, 105–117 (2010).
18. Mullighan,C. G. et al. BCR-ABL1 lymphoblastic leukaemia is characterized by the
deletion of Ikaros. Nature 453, 110–114 (2008).
19. Mullighan, C. G. et al. Deletion of IKZF1 and prognosis in acute lymphoblastic
leukemia. N. Engl. J. Med. 360, 470–480 (2009).
stem cells are efficiently detected following intrafemoral transplantation into
NOD/SCID recipients depleted of CD1221cells. Blood 106, 1259–1261 (2005).
21. Taussig, D. C. et al. Anti-CD38 antibody-mediated clearance of human
22. Shultz, L. D. et al. Human lymphoid and myeloid cell development in NOD/LtSz-
scid IL2Rcnullmice engrafted with mobilized human hemopoietic stem cells. J.
Immunol. 174, 6477–6489 (2005).
AML: implications for our understanding of the heterogeneity of AML. Blood 107,
24. Williams, R. T., den Besten, W. & Sherr, C. J. Cytokine-dependent imatinib
resistance in mouse BCR-ABL1, Arf-null lymphoblastic leukemia. Genes Dev. 21,
Clin. Invest. 120, 636–644 (2010).
27. Tehranchi, R. et al. Persistent malignant stem cells in del(5q) myelodysplasia in
remission. N. Engl. J. Med. 363, 1025–1037 (2010).
Supplementary Information is linked to the online version of the paper at
Acknowledgements We would like to thank S. Minkin for statistical analysis of patient
outcome, the Dick Laboratory and B. Neel for critical review of this manuscript,
The Stem Cell Network of Canadian National Centres of Excellence, the Canadian
Cancer Society and the Terry Fox Foundation, Genome Canada through the Ontario
of Ontario, the Leukemia and Lymphoma Society, the Canadian Institutes for Health
Research, a Canada Research Chair, and the American and Lebanese Syrian
in part by the Ontario Ministry of Health and Long Term Care (OMOHLTC). The views
expressed do not necessarily reflect those of the OMOHLTC.
Author Contributions F.N. designed study, analysed data and prepared figures. F.N.,
C.G.M., J.C.Y.W., A.P., S.D. and L.A.P. performed experiments. M.D.M. provided patient
samples. J.C.Y.W. and M.D.M. provided patient outcome data. J.M. performed paired
and unpaired segmentation analysis of SNP array data. F.N. and C.G.M analysed and
interpretedSNPdata. C.G.M., J.C.Y.W.,S.D.and J.R.D.criticallyreviewedand edited the
manuscript. F.N. and J.E.D. wrote the manuscript. J.E.D. supervised the study.
Author Information Reprints and permissions information is available at
www.nature.com/reprints. The authors declare no competing financial interests.
Readers are welcome to comment on the online version of this article at
www.nature.com/nature. Correspondence and requests for materials should be
addressed to J.E.D. (firstname.lastname@example.org).
2 0 J A N U A R Y 2 0 1 1 | V O L 4 6 9 | N A T U R E | 3 6 7
Macmillan Publishers Limited. All rights reserved
METHODS Download full-text
Patient samples. Patient samples (primarily peripheral blood) were obtained
from newly diagnosed Philadelphia-positive acute lymphoblastic leukaemia
patients according to pre-established guidelines approved by the Research
Ethics Board of University Health Network. Three out of twenty patient samples
Cell viability, as assessed immediately after thawing, was greater then 90% for all
cases. Detailed patient data are provided in Supplementary Tables 6 and 7.
Xenotransplantation assay and analysis. NOD.CB17-Prkdcscid/J (NOD/SCID)
established and approved by the Animal Care Committee at University Health
Network. Ten-to-twelve-week-old old mice were sublethally irradiated at
225cGy 24h before transplant. NOD/SCID mice were also treated with mouse
anti-CD122 monoclonal antibody (NS122) after sublethal conditioning as previ-
ously described20. Only female mice were used in these studies28. NSG mice were
TGGACAACAAAT-39; mutated reverse, 59-GCCAGAGGCCACTTGTGTAG-39).
All primer sets were obtained from Jackson Laboratories website.
l21DNase (Roche AppliedScience). After centrifugation, cells were resuspended
in IMDM and counted using ViCell XR (Beckman coulter) or by trypan blue
exclusion. For transplantation, cells were placed in microfuge tubes and spun
down to remove excess media. Cells were resuspended in the correct volume
was performed as previously described29. Briefly, mice were anaesthetized using
isoflurane. The right knee of mice was bent and drilled with a 27.5-g needle and
followed by injection of cells with a 28.5-g insulin syringe (BD Biosciences).
After transplant, animals were monitored for the appearance of disease symp-
toms, such as weight loss, hunch-back, decreased activity, and killed soon after.
Micetransplanted with samplesthatdid notinduce diseasesymptoms(in NS122
or NSG recipients) were killed by 16–24weeks. Upon death, injected right femur
(IF), non-injected bones (left femur, right and left tibiae, bone marrow (BM)),
spleen and peripheral blood were removed and analysed for the presence of
human leukaemia blasts using flow cytometry. Aliquots of cells were stained in
96-well round-bottomed plates (BD Falcon). Human cells were distinguished
from mouse cells using human-specific CD45PC7 and CD44PE (Supplemen-
tary Fig. 18). B-cell-specific markers (various combinations of IgM FITC,
CD19 PC5 (Beckman Coulter), CD20 APC7, CD10 APC, CD34 APC7) were
used to evaluate blast phenotype after transplant. Detection of normal haemato-
poietic stem cell activity was monitored using CD33 PC5 (Beckman Coulter) or
APC. All fluorochromes were obtained from BD Biosciences unless otherwise
indicated. Mice were considered to be engrafted when multiple human leukaemia
0.5% threshold. Virtually all cases of engraftment were well above this threshold.
All flow cytometry analysis was performed on the LSRII (BD Biosciences). The
remaining marrow (IF1BM) was frozen viably in FCS 110% DMSO.
DNA SNP microarray analysis. DNA isolated from patients at diagnosis and
xenograft samples was analysed using Affymetrix 6.0 SNP arrays. SNP array data
binary segmentation as previously described11,18,19,30–32. To distinguish inherited
from somatic DNA copy number alterations for primary patient samples lacking
filtered using public copy number polymorphism databases33,34, and an in-house
database of SNP array data from several hundred samples. SNP array data are
available at dbGaP (phs000329.v1.p1) and is also hosted through St Jude
Children’s Research hospital (http://hospital.stjude.org/forms/genome-down-
load/request/). Sample information is shown in Supplementary Table 7.
Quantitative PCR of the CDKN2A/B locus. Primers for genomic quantitative
Supplementary Table 3. Taqman RNase P primers (Applied Biosystems) were used
(Applied Biosystems), using the 7500 universal cycling conditions: 50uC for 2min,
followed by 95uC for 10min, then 40 cycles of 95uC for 1min and 60uC for 1min.
Standard curves for each CDKN2A/B exon and RNase P were generated using
copy number values were normalized by dividing the value obtained for each test
Statistical analysis. All data were analysed using GraphPad Prism version 5.00 for
Mac OS X (http://www.graphpad.com). The Mann–Whitney U-test was used to
assess statistically significant differences in chimaerism in xenografts. Clinical out-
come data was analysed using the Gehan–Wilcoxon method based on random
10,000 permutations. Because the normal distribution of data can only be assumed
in larger cohorts, this modified test more accurately computes P-values for small
is more efficient in female NOD/SCID/IL-2Rgc-null recipients. Blood 115,
stem cells. Nature Med. 9, 959–963 (2003).
of-heterozygosity data. Bioinformatics 20, 1233–1240 (2004).
31. Mullighan, C. G. et al. Genome-wide analysis of genetic alterations in acute
lymphoblastic leukaemia. Nature 446, 758–764 (2007).
32. Pounds, S. et al. Reference alignment of SNP microarray signalsfor copy number
analysis of tumors. Bioinformatics 25, 315–321 (2009).
33. Iafrate, A. J. et al. Detection of large-scale variation in the human genome. Nature
Genet. 36, 949–951 (2004).
and copy number variation. Nature Genet. 40, 1166–1174 (2008).
Macmillan Publishers Limited. All rights reserved