論文の公開元へ

書き出し

Refer/BibIX

RIS

BibTeX

TSV

Combined landscape of single-nucleotide variants and copy number alterations in clonal hematopoiesis

Saiki, Ryunosuke 京都大学 DOI:10.14989/doctor.k24507

2023.03.23

概要

Articles
https://doi.org/10.1038/s41591-021-01411-9

Combined landscape of single-nucleotide
variants and copy number alterations in clonal
hematopoiesis
Ryunosuke Saiki 1, Yukihide Momozawa2, Yasuhito Nannya1, Masahiro M. Nakagawa1,3,
Yotaro Ochi 1, Tetsuichi Yoshizato 1, Chikashi Terao 4, Yutaka Kuroda5, Yuichi Shiraishi6,
Kenichi Chiba6, Hiroko Tanaka 7, Atsushi Niida8, Seiya Imoto9, Koichi Matsuda10, Takayuki Morisaki11,
Yoshinori Murakami11, Yoichiro Kamatani4,10, Shuichi Matsuda5, Michiaki Kubo12, Satoru Miyano7,
Hideki Makishima 1 and Seishi Ogawa 1,3,13 ✉
Clonal hematopoiesis (CH) in apparently healthy individuals is implicated in the development of hematological malignancies
(HM) and cardiovascular diseases. Previous studies of CH analyzed either single-nucleotide variants and indels (SNVs/indels)
or copy number alterations (CNAs), but not both. Here, using a combination of targeted sequencing of 23 CH-related genes and
array-based CNA detection of blood-derived DNA, we have delineated the landscape of CH-related SNVs/indels and CNAs in
11,234 individuals without HM from the BioBank Japan cohort, including 672 individuals with subsequent HM development, and
studied the effects of these somatic alterations on mortality from HM and cardiovascular disease, as well as on hematological
and cardiovascular phenotypes. The total number of both types of CH-related lesions and their clone size positively correlated
with blood count abnormalities and mortality from HM. CH-related SNVs/indels and CNAs exhibited statistically significant
co-occurrence in the same individuals. In particular, co-occurrence of SNVs/indels and CNAs affecting DNMT3A, TET2, JAK2
and TP53 resulted in biallelic alterations of these genes and was associated with higher HM mortality. Co-occurrence of SNVs/
indels and CNAs also modulated risks for cardiovascular mortality. These findings highlight the importance of detecting both
SNVs/indels and CNAs in the evaluation of CH.

T

he presence of clonal components in an apparently normal
hematopoietic compartment, or CH, has drawn increasing
attention in recent years1,2. Although suggested only indirectly by skewed chromosome X inactivation in early studies3–7, CH
has recently been demonstrated by detection of CNAs in peripheral
blood samples from large cohorts of individuals without blood cancer using single-nucleotide polymorphism (SNP) array data from
genome-wide association studies (GWAS)8–11. Showing a substantial
overlap with those characteristic of HM, CNAs were shown to be
associated with an elevated risk of developing HM8,9. More recently,
CH has also been detected by the presence of somatic SNVs/indels
in the peripheral blood of apparently healthy individuals12–15 and
cancer patients16,17 using next-generation sequencing. In addition to
its link to HM, CH as detected by SNVs/indels has been highlighted
by its unexpected association with a significantly increased risk for
cardiovascular diseases (CVD)12,13,18,19.
Regardless of the type of genetic lesions by which it is detected,
CH is strongly age related with an increasing frequency in the
elderly8–13. With substantially improved technologies available to

identify CNAs and somatic SNVs/indels, a complete registry of
CNAs and SNVs/indels associated with CH has been elucidated,
and these are thought to involve virtually every individual among
the most elderly20,21. However, to date, no studies have evaluated
CNAs and SNVs/indels together at comparable sensitivity in a large
cohort of a general population, although they have recently been
investigated in a cancer population in which many patients had
been treated with chemo-/radiotherapy22. What is the landscape of
CH recognized by combining both CNAs and SNVs/indels in a general population? Are there any interactions between SNVs/indels
and CNAs that shape the landscape of CH? How are hematological
phenotypes affected by both CH-related lesions? How does it affect
HM and CVD risks? These are the key questions to be answered for
a better understanding of CH and its implication in HM and CVD.
In the present study, for the purpose of delineating the combined
landscape of driver SNVs/indels and CNAs in CH, we performed
SNP array-based copy number analysis and targeted sequencing
of major CH-related genes on blood-derived DNA from BioBank
Japan (BBJ)23 and which had been SNP-typed for GWAS studies for

Department of Pathology and Tumor Biology, Graduate School of Medicine, Kyoto University, Kyoto, Japan. 2Laboratory for Genotyping Development,
RIKEN Center for Integrative Medical Sciences, Yokohama, Japan. 3Institute for the Advanced Study of Human Biology (WPI-ASHBi), Kyoto University,
Kyoto, Japan. 4Laboratory for Statistical and Translational Genetics, RIKEN Center for Integrative Medical Sciences, Yokohama, Japan. 5Department of
Orthopaedic Surgery, Graduate School of Medicine, Kyoto University, Kyoto, Japan. 6Division of Cellular Signaling, National Cancer Center Research
Institute, Tokyo, Japan. 7Department of Integrated Data Science, M&D Data Science Center, Tokyo Medical and Dental University, Tokyo, Japan.
8
Laboratory of Molecular Medicine, Human Genome Center, Institute of Medical Science, The University of Tokyo, Tokyo, Japan. 9Division of Health
Medical Intelligence, Human Genome Center, Institute of Medical Science, The University of Tokyo, Tokyo, Japan. 10Department of Computational Biology
and Medical Sciences, Graduate School of Frontier Sciences, The University of Tokyo, Tokyo, Japan. 11Division of Molecular Pathology, Institute of Medical
Science, The University of Tokyo, Tokyo, Japan. 12RIKEN Center for Integrative Medical Sciences, Yokohama, Japan. 13Department of Medicine, Centre for
Haematology and Regenerative Medicine, Karolinska Institute, Stockholm, Sweden. ✉e-mail: sogawa-tky@umin.ac.jp
1

Nature Medicine | VOL 27 | July 2021 | 1239–1249 | www.nature.com/naturemedicine

1239

Articles
common diseases, including hypertension, diabetes, autoimmune
diseases and several solid cancers23. We then investigated the combined effect of both CH-related lesions on clinical phenotypes and
outcomes, particularly that on the mortality from HM and CVD.

Results

Nature Medicine
of ≤1%, which was below the limit of detection for SNVs/indels.
Thus, smaller clones were detected through CNAs, particularly
copy-neutral loss of heterozygosity (LOH) or uniparental disomy
(UPD), compared with those detected through SNVs/indels
(Supplementary Fig. 1c).
We found 27 significantly recurrent CNAs, many of which are
also commonly seen in HM, supporting a pathogenic link between
CH and leukemogenesis (Extended Data Fig. 4a–c). In accordance
with previous reports8–11, 14qUPD, +21q, del(20q) and +15q were
among the most frequent CNA lesions (Extended Data Fig. 2b,c)
while del(20q), 16pUPD and 17pUPD showed the largest mean
clone size (Supplementary Fig. 2). Several CNAs, including 14qUPD
and +21, showed higher frequencies than those reported in Western
populations, which is probably due to a higher sensitivity for detection of CNAs in this study compared with that in previous studies in Western populations8–11; when confined to lesions with ≥5%
cell fractions, the difference across studies becomes less conspicuous for many CNA targets (Extended Data Fig. 4d,e). Nevertheless,
even considering varying sensitivity, several CNAs, including +15,
del(14q), del(9q), del(20q) and del(13q), still showed a different frequency across studies in both populations21, suggesting an ethnic
difference in positive selection of CH-related CNAs (Extended Data
Fig. 4e), although the exact genetic basis of this difference is largely
unclear for most CNAs.

Identification of CH-related SNVs/indels and CNAs. We enrolled
a total of 11,234 subjects from the BBJ cohort (n = 179,417) in
which SNP array analysis of peripheral blood-derived DNA had
been performed for large-scale GWAS studies for common diseases (Supplementary Tables 1 and 2) (https://biobankjp.org/info/
pdf/sample_collection.pdf)23. Among these, 10,623 were randomly selected from 60,787 cases aged ≥60 years at the time of
sample collection and were confirmed not to have solid cancers
as of March 2013. This randomly selected set included 61 cases
who were known to have developed and/or died from HM as of
March 2017. The remaining 611 consisted of all cases from the
entire BBJ cohort who were confirmed as having developed and/
or died from HM as of the same date but were not included in the
10,623 cases randomly selected. In total, 672 cases in the entire BBJ
cohort were reported to have HM: 215 myeloid, 420 lymphoid and
37 lineage-unknown tumors (Extended Data Fig. 1a). For these
11,234 cases, SNVs/indels in blood were investigated using multiplex PCR-based amplification of exons of 23 CH-related genes, followed by high-throughput sequencing (Methods).24 The sensitivity
of SNV detection according to in silico simulations using known
SNPs was >94% for 3% variant allele frequency (VAF) and >74%
for 2% VAF, but <20% for 1% VAF with a mean depth of ~800×
(Supplementary Fig. 1a,b).
In total, we called 4,056
SNVs/indels (2,750
SNVs and
1,306 indels) in 3,071 (27.3%) subjects, of which 2,312 (20.6%)
had one, 586 (5.2%) two and 173 (1.5%) at least three SNVs/
indels (Fig. 1a). VAF values were widely distributed, from 0.5 to
85.6% with a median of 3.0% (Supplementary Fig. 1c). Age dependence of CH-related SNVs/indels was evident (Fig. 1b). In accordance with previous reports, DNMT3A (13.5%), TET2 (9.5%),
ASXL1 (2.2%) and PPM1D (1.4%) were most frequently mutated
(Extended Data Fig. 2a,c). Several combinations of genes, including TET2/DNMT3A, ASXL1/TET2, ASXL1/CBL, SRSF2/TET2 and
SRSF2/ASXL1, were more frequently comutated than what would
be expected by chance only (odds rstio (OR) = 1.53–6.53, q < 0.05;
Extended Data Fig. 2d). ...

論文の公開元へ

この論文で使われている画像

参考文献

43. Harismendy, O. et al. Detection of low prevalence somatic mutations

in solid tumors with ultra-deep targeted sequencing. Genome Biol. 12,

R124 (2011).

44. Forshew, T. et al. Noninvasive identification and monitoring of cancer

mutations by targeted deep sequencing of plasma DNA. Sci. Transl. Med. 4,

136ra168 (2012).

45. Yoshida, K. et al. Frequent pathway mutations of splicing machinery in

myelodysplasia. Nature 478, 64–69 (2011).

46. Haferlach, T. et al. Landscape of genetic lesions in 944 patients with

myelodysplastic syndromes. Leukemia 28, 241–247 (2014).

47. Suzuki, H. et al. Mutational landscape and clonal architecture in grade II and

III gliomas. Nat. Genet. 47, 458–468 (2015).

48. Shiraishi, Y. et al. An empirical Bayesian framework for somatic mutation

detection from cancer genome sequencing data. Nucleic Acids Res. 41,

e89 (2013).

49. Niida, A., Imoto, S., Shimamura, T. & Miyano, S. Statistical model-based

testing to evaluate the recurrence of genomic aberrations. Bioinformatics 28,

i115–i120 (2012).

50. Arber, D. A. et al. The 2016 revision to the World Health Organization

classification of myeloid neoplasms and acute leukemia. Blood 127,

2391–2405 (2016).

Nature Medicine

Acknowledgements

This work was supported by the Japan Agency for Medical Research and Development

(nos. JP15cm0106056h0005, JP19cm0106501h0004, JP16ck0106073h0003 and

JP19ck0106250h0003 to S.O.; nos. JP17km0405110h0005 and JP19ck0106470h0001 to

H.M.; and no. JP19ck0106353h0003 to Y.N.); the Core Research for Evolutional Science

and Technology (no. JP19gm1110011 to S.O.); the Ministry of Education, Culture, Sports,

Science and Technology of Japan; the High Performance Computing Infrastructure

System Research Project (nos. hp160219, hp170227, hp180198 and hp190158 to S.O.

and S. Miyano) (this research used computational resources of the K computer provided

by the RIKEN Advanced Institute for Computational Science through the HPCI System

Research project); the Japan Society for the Promotion of Science; Scientific Research on

Innovative Areas (nos. JP15H05909 to S.O. and S. Miyano and JP15H05912 to S. Miyano)

and KAKENHI (nos. JP26221308 and JP19H05656 to S.O., JP16H05338 and JP19H01053

to H.M. and JP15H05707 to S. Miyano); and the Takeda Science Foundation (to S.O.,

H.M. and T.Y.). S.O. is a recipient of the JSPS Core-to-Core Program A: Advanced

Research Networks. DNA samples and subjects’ clinical data were provided by BBJ, the

Institute of Medical Science, the University of Tokyo. The supercomputing resource was

provided by the Human Genome Center, the Institute of Medical Science, the University

of Tokyo. We thank K. Matsuo at the Aichi Cancer Center Research Institute (Nagoya,

Japan), who suggested the design of the case-cohort study for estimation of cumulative

mortality from, and incidence of, hematological malignancies. We thank the TCGA

Consortium and all its members for making publicly available their invaluable data.

Author contributions

R.S., H.M. and S.O. designed the study. K.M., Y. Kamatani, T.M. and Y. Murakami provided

DNA samples and clinical data. Y. Kuroda and S. Matsuda provided bone marrow samples.

C.T. and Y. Kamatani performed copy number analysis. Y. Momozawa and M.K. performed

sequencing. M.M.N. performed cell sorting and single-cell analysis. R.S., M.M.N., Y.O.,

T.Y., Y.S., K.C., H.T., A.N., S.I. and S. Miyano performed bioinformatics analysis. R.S., Y.N.,

M.M.N., Y.O., T.Y., H.M. and S.O. prepared the manuscript. All authors participated in

discussions and interpretation of the data and results.

Competing interests

The authors declare no competing interests.

Additional information

Extended data is available for this paper at https://doi.org/10.1038/s41591-021-01411-9.

Supplementary information The online version contains supplementary material

available at https://doi.org/10.1038/s41591-021-01411-9.

Correspondence and requests for materials should be addressed to S.O.

Peer review information Nature Medicine thanks Daniel Link, Duane Hassane, Todd

Druley and the other, anonymous, reviewer(s) for their contribution to the peer review

of this work. Michael Basson was the primary editor on this article and managed its

editorial process and peer review in collaboration with the rest of the editorial team.

Reprints and permissions information is available at www.nature.com/reprints.

Nature Medicine | www.nature.com/naturemedicine

Articles

Nature Medicine

CH(+)

Case-control study for all HM

Case

(n=672)

Total

376

296

672

160

215

100

CNA alone

Both

All CH(+)

154

115

107

AML

MDS

MPN

CML

Others

Case

Myeloid

Lymphoid

191

229

420

B-NHL

123

143

266

T-NHL

CLL

ALL

Control

2,177

1,399

633

4,209

6,353

10,562

Total

2,331

1,514

740

4,585

6649

11,234

SNV alone

CNA alone

Both

All CH(+)

CH( )

Total

AML

MDS

Control

(n=10,562)

MM/PCT

Others

Linage Unknown

CH( )

SNV alone

Subcohort

CH(+)

Case-cohort study for HM death

Hematogical malignancy (+)

Myeloid

Target cohort (n=43,662 * )

(≥60 y.o. and no cancer history)

MPN

CML

Others

B-NHL

T-NHL

CLL

Lymphoid

Subcohort

(n=7,937 ** )

ALL

MM/PCT

Others

Hematological malignancy ( )

1,614

1,036

447

3,097

4,785

7,882

Total

1,628

1,047

454

3,129

4,808

7,937

Linage Unknown

Case

(n=401)

overlap

(n=55)

Case (Death from HM)

CH(+)

Among 60,787 cases aged ≥60 years and

confirmed not to have solid cancers as of March

2013, 43,662 had the follow up data for survival.

SNV alone

CNA alone

Both

All CH(+)

CH( )

Total

109

239

162

401

107

146

AML

MDS

MPN

CML

Others

122

244

B-NHL

148

T-NHL

CLL

ALL

Hematogical malignancy (+)

Myeloid

Among 10,623 cases randomly selected from the

60,787 cases, 7,937 had the follow up data for

survival.

Lymphoid

MM/PCT

Others

Linage Unknown

Hematological malignancy ( )

Total

109

239

162

401

Extended Data Fig. 1 | Design of case-control and case-cohort study. a, Design of case-control study (Left). Diagnosis of hematological malignancies

(HM) in subjects with or without CH enrolled in the case-control study (Right). b, Design of case-cohort study for death from HM (Left). Diagnosis of

HM in subjects with or without CH enrolled in the case-cohort study (Right). AML, acute myeloid leukemia; MDS, myelodysplastic syndromes; MPN,

myeloproliferative neoplasms; CML, chronic myeloid leukemia; B-NHL, B-cell non-Hodgkin lymphoma; T-NHL, T-cell non-Hodgkin lymphoma; CLL, chronic

lymphoid leukemia; ALL, acute lymphoblastic leukemia; MM, multiple myeloma; PCT, plasma cell tumor.

Nature Medicine | www.nature.com/naturemedicine

Articles

Number of subjects

Nature Medicine

1500

1000

500

Number of subjects

TE A

A T2

PP L1

SF 53

SR 1

G L

JA 1

U 2

G 1

EZ S

RU H2

K 1

N S

ET S

M V6

1500

1000

500

de 21q

l(2

+1 )

1p 5q

1q PD

de PD

de (5q

l(1 )

11 3q

qU )

9q PD

9p PD

6p PD

17 PD

4q PD

de PD

l(1

+2 )

de 2q

l(6

de q

l(1 )

17 +8q

FDR

< 0.01

< 0.001

< 0.01

> 100

< 0.1

cooccurr

in ≥5 cases

0.6

0.5

0.4

0.3

0.2

0.1

0.2

0.3

0.4

78.0

12qUPD

TET2

p.I1873T

39.9

79.8

+21

Subject 2

DNMT3A

p.W313X

57.1

100.0

del(20q)

14UPD

7.3

Subject 3

TP53

p.G205S

51.3

100.0

del(6q)

9.2

Subject 4

DNMT3A

p.R882H

34.8

69.6

ETV6

Subject 8

0.1

0.5

0.6

0.1

SNV/indels

p.G375R

VAF (%) Cell fractions (%) CNA

0.3

0.2

0.1

0.2

5.5

p.R166Q

19.9

39.8

p.S34F

35.1

70.2

SF3B1

p.K700E

27.5

55.0

TET2

p.R1216X

U2AF1

p.S34Y

p.T188A

2.2

4.4

34.0

68.0

31.1

1.9

67.4

del(13q)

57.0

0.3

62.2

SRSF2

p.P95R

30.2

TET2

p.V1900A

32.4

64.8

JAK2

p.V617F

64.8

100.0

TET2

p.L748X

48.0

96.0

p.E184X

41.5

83.0

0.5

0.3

0.2

0.1

0.3

0.4

0.5

0.6

0.1

TCRA

0.2

0.3

0.4

0.5

0.6

VAF of TET2

0.5

0.4

0.3

SNV/indel

Both

CNA

21 %

13 %

0.2

0.1

0.6

0.1

0.2

0.3

0.4

0.5

0.6

VAF of CBL

Position on chr14 (Mb)

Proportion of subjects (%)

TP53

DNMT3A

TET2

JAK2

ASXL1

del(20q)

PPM1D

2.3

9p+

60.4

60.7

61.2

del(20q)

54.6

9pUPD

28.5

SF3B1

Number of alterations

GNB1

3 ≥4

CBL

60.4

TET2

0.4

VAF of SRSF2

11.0

RUNX1

U2AF1

RUNX1

1.4

0.4

0.6

Cell fraction (%)

21.4

0.2

VAF of ASXL1

0.4

0.1

0.5

VAF of TET2

0.6

0.5

39.0

Subject 7

0.2

0.6

TE A

A T2

PP XL1

TP D

SF 53

SR B1

G L

JA 1

U K2

G F1

EZ S

ID 2

2p 2

4q OH

17 O

pL H

1p OH

1q PD

6p PD

9p PD

9q PD

11 UP

q D

14 UP

q D

16 UP

p D

17 UP

qU D

de PD

de 5q

de l(6q

l( )

d 11

de el(1 q)

l(1 3q

4q )

de 1

l(2 1)

VAF of ASXL1

p.V927fs

Subject 6

0.3

VAF of TET2

Gene

TET2

Subject 5

0.4

Subject 1

0.5

2pLOH

4qLOH

17pLOH

1pUPD

1qUPD

6pUPD

9pUPD

9qUPD

11qUPD

14qUPD

16pUPD

17qUPD

del(5q)

del(6q)

del(11q)

del(13q)

del(14q11)

del(20q)

+15

+21

+22

0.6

VAF of SRSF2

Odds ratio

VAF of DNMT3A

TET2

ASXL1

PPM1D

TP53

SF3B1

SRSF2

CBL

GNB1

JAK2

U2AF1

GNAS

EZH2

IDH2

VAF of ASXL1

SRSF2

Deletion

TET2 SNV/indels

U2AF1

GNAS

Extended Data Fig. 2 | Landscape of genetic alterations in CH. a-b, The number of subjects with individual SNVs/indels (a) and CNAs (b). The vertical

axis represents the number of subjects with indicated alterations. Unclassifiable CNAs are not included in (b). c, Landscape of SNVs/indels and CNAs

in 11,234 subjects. Those without CH-related alterations are omitted. d, The correlations between individual genetic alterations. Combinations seen in 5

or more cases are indicated by asterisks. e-i, VAF of cooccurring SNVs/indels in diagonal plot. Dots above the dashed line fulfill ‘pigeonhole principle’. j,

Venn diagram illustrating the overlap between subjects with SNVs/indels and those with CNAs. Frequencies within all subjects in whom SNVs/indels and

CNAs were examined (n = 11,234) are indicated. k, Subjects in whom cooccurring SNVs/indels and CNAs were suspected to coexist in the same cells on

the basis of ‘pigeonhole principle.’ l, A magnified illustration of microdeletions around TCRA locus (14q11.2). A gray bar represents gene body of TCRA.

Blue horizontal bars represent microdeletions. Cooccurring TET2 SNVs are indicated by red dots. Genomic coordinates in hg19 are indicated above. m,

Proportions of subjects with different number of cooccurring alterations within those who harbor SNVs/indels in the indicated genes. The proportions of

subjects with 1, 2, 3, and ≥4 CNAs are depicted by different colors.

Nature Medicine | www.nature.com/naturemedicine

Articles

Nature Medicine

Chr1

Chr2

Chr3

Chr4

TET2

DNMT3A

Chr5

Chr6

GNB1

Chr8

Chr7

TNFAIP3

TCRB

EZH2

HLA

Chr9

Chr10

Chr11

Chr12

Chr13

miR-15a

miR-16-1

CBL

Chr16

Chr17

Chr18

Chr15

TCRA

ATM

JAK2

Chr14

Chr19

Chr20

Chr21

NF1

Chr22

CHEK2

TP53

50 subjects

RUNX1

Type of CNAs

Cell fraction (%)

Duplication

0.1

10 100

Deletion

0.1

10 100

UPD

0.1

Unclassifiable

Cooccurring SNV/indels

10 100

Extended Data Fig. 3 | Distribution of CNAs in all chromosomes. Distributions of CNAs on all chromosomes are illustrated. Loci of known driver genes are

indicated by arrows. Each horizontal bar represents one CNA. Cooccurring SNVs/indels are indicated by red dots. Types of CNAs are depicted by different

colors as indicated in the annotations.

Nature Medicine | www.nature.com/naturemedicine

Articles

Nature Medicine

Duplication

Current Study

Loh et al. 2020

+1q

Current Study

6pUPD

9pUPD

9qUPD

+15q

+18

+21q

+22q

+12

11qUPD

+14q

14qUPD

17qUPD

del(6q16-24) TNFAIP3

del(7q32-36)

TCRB, EZH

11pUPD

del(11q14-23) ATM

del(13q13-31) miR-15a miR-16-1

12qUPD

13qUPD

16p,16qUPD

15qUPD

del(14q11.2) TCRA

del(17p13-11) TP53

del(20q11-13)

22qUPD

del(4q23-24)

TET2

del(5q14-32)

del(3p13)

4qUPD

Loh et al. 2020

del(2p23)

DNMT3A

+3q

Current Study

1qUPD

Deletion

Loh et al. 2020

1pUPD

UPD

del(21q)

del(8p23.1)

del(10q25-26)

FRA10B

del(22q12) CHEK2

1.3

1.2

1.1

Frequency (%)

1.0

0.9

0.8

Cell fraction <5%

This study

Loh et al, 2020

Laurie et al, 2012

Jacobs et al, 2012

0.7

0.6

0.5

0.4

0.3

0.2

0.1

−q

l(1

l(2

de 1−1

l(1 3)

de +1

l(2 5

1p 3)

de P

l(9 D

de 14 31)

l(1 qU

de q13 D

l(5 −3

de 14 1)

l(6 −3

q1 2)

6−

9p q

1q 1

11 PD

16 PD

de 12 PD

l(1 qU

de q14 D

l(7 −2

q3 3)

de −36

l(3 )

de p1

l(2 3)

9q 2)

l(1 qU

7p PD

17 11)

6p D

11 18

de +

l(2 12

4q 2)

16 PD

13 PD

15 PD

0.4

Frequency (%)

0.3

Enriched in the current study

Cell fraction ≥5%

Enriched in Loh et al. 2020

0.2

0.1

q2 +3

5− q

l(2

de 1−1

l(1 3)

de +1

l(2 5

1p 3)

de P

l(9 D

de 14 31)

l(1 qU

3q PD

de 13

l(5 −3

de q14 1)

l(6 −3

q1 2)

6−

9p q

1q 1

11 PD

16 PD

de 12 PD

l(1 qU

de q14 D

l(7 −2

q3 3)

de −36

l(3 )

de p1

l(2 3)

9q 2)

de 22 22

l(1 qU

7p PD

17 11)

6p D

11 18

de +

l(2 12

4q 2)

16 PD

13 PD

15 PD

Extended Data Fig. 4 | Chromosomal regions significantly affected by CNAs. a-c, Chromosomal regions significantly affected by duplications (a), UPDs

(b), and deletions (c) in a Japanese cohort (current study) and in a British cohort11. Statistical significance for recurrence of CNAs were evaluated by PART49.

Dashed lines indicate thresholds for statistical significance (FDR = 0.25). d-e, Comparison of frequencies of individual CNAs between the current and

previous studies8,9,11. Comparisons were performed in those aged 60-75 years. In (d) or (e), CNAs in <5% or ≥5% cell fractions were taken into account,

respectively. CNAs significantly enriched in either cohort (FDR < 0.1) were indicated by asterisks in (e).

Nature Medicine | www.nature.com/naturemedicine

Articles

Nature Medicine

542

0.3

SNV/indels

CNAs

0.2

200

SNV/indels + CNA

SNV/indel alone

CNA alone

Age (Years)

Frequency (%)

Count

DNMT3A

TET2

ASXL1

PPM1D

TP53

JAK2

SF3B1

SRSF2

GNAS

GNB1

CBL

U2AF1

IDH2

MYD88

EZH2

KRAS

NRAS

20q

13q

14q

12q

11q

15q

22q

11p

17q

others

−9

−6

Number of alterations

−8

−7

0.1

−6

Frequency

Number of subjects

400

−8

600

chr1

chr2

chr3

chr4

MYD8

TET2

chr5

chr6

chr7

chr8

chr9

EZH2

JAK2

chr10

chr11

chr12

chr13

chr14

chr15

SNV/indel

Missense

Inframe indel

Splice−site

Frameshift indel

Stop−gain

Multiple

chr16

chr17

UPD

Unclassifiable

chr19

chr20

chr21

chr22

Cell fraction (%)

100

Duplication

Deletion

chr18

0.1

CNA

TP53

Deletion

UPD

Unclassifiable

SNV/indels

Extended Data Fig. 5 | Analysis of SNVs/indels and CNAs in peripheral blood samples in TCGA cohort. a, Distribution of the number of genetic

alterations in each subject. Subjects with SNVs/indels alone, with CNAs alone, or with both of them are illustrated by different colors. b, Solid lines indicate

the prevalence of CH-related SNVs/indels and CNAs, according to age. Colored bands represent the 95% confidence intervals. c, The landscape of

CH-related SNVs/indels and CNAs. Each row represents genetic alterations or affected chromosomal arms, and each column represents subjects. Subjects

without any alterations are omitted. Types of SNVs/indels and CNAs are depicted by different colors. d, Distributions of CNAs on all chromosomes are

illustrated. Loci of cooccurring SNVs/indels are indicated by arrows. Each horizontal bar represents one CNA. Cooccurring SNVs/indels are indicated by

red asterisks. Types of CNAs are depicted by different colors.

Nature Medicine | www.nature.com/naturemedicine

Articles

Nature Medicine

HM(-)

HM(+)

TP53

TET2

JAK2 DNMT3A

GNB1

CBL

RUNX1

SNV alone (n=1,357)

SNV+CNA, different loci (n=332)

Cumulative mortality from

cardiovascular diseases

SNV+CNA, same loci (n=42)

0.2

P = 0.34

0.1

No alteration (n=4,097)

Proportion of SNV/indels

associated with CNAs in the same gene (%)

Number of cases with

SNV/indels and CNAs in the same gene

0.3

EZH2

TP53

TET2

JAK2

DNMT3A

GNB1

RUNX1

CBL

EZH2

100

120

140

Months

0.2

P value

No alterations (n=4.947)

SNV/indel alone (n=1,723)

0.12

8.3×10

3.9×10

0.8

-5

-3

SNV+CNA, different genes/loci (n=450)

SNV+CNA, same genes/loci (n=64)

0.08

SNV/indel + CNA in the same genes/loci

other than TP53 + 17p (n=46)

TP53 + 17pLOH (n=18)

0.04

100

120

1.6×10

1.5×10

0.6

0.4

No alteration (n=4,097)

0.047

SNV alone (n=1,357)

0.2

SNV+CNA, different loci (n=332)

SNV+CNA, same loci (n=42)

Number of subjects within the case-cohort design.

P = 0.16

140

100

120

140

Months

-2

-5

19 21

20 22

WBC Hb

Plt Ht

#CNA

P = 0.0085

0.15

Cumulative mortality from

hematological maligancies

Overall survival

Cumulative mortality from

hematological maligancnies

0.16

TP53 + 17p alt.

(n=29) *

0.1

TP53 without 17p alt.

(n=81) *

0.05

100

120

140

Months

of subjects within the case-cohort design.

* Number

SNVs/indels of TP53 include those detected by ddPCR.

17p alt.

#CNA ≥3

del(5q)

Number of TP53 SNVs

Copy-number alterations

Blood counts abnormality

#CNAs

Single TP53 SNV

Duplication

UPD

Low

Normal

≥3 CNAs

Multiple TP53 SNV

Deletion

Unclassifiable

High

Unknown

<3 CNAs

100

Odds ratio for mortality from MDS in

subjects with TP53-involving SNVs/indels (n=165)

Extended Data Fig. 6 | Interplay between SNVs/indels and CNAs. a, Number of subjects with SNVs/indels and CNAs involving the same genes/loci. b,

Proportion of SNVs/indels associated with CNAs in the same genes/loci. c, Cumulative mortality from hematological malignancies. d, Cumulative mortality

from cardiovascular diseases. e, Survival curves for overall survival. f, Profiles of CNAs in subjects with SNVs/indels in TP53. Abnormally high or low blood

counts (WBC, Platelet, hemoglobin, and hematocrit) are indicated by red or blue, respectively. Numbers of cooccurring CNAs are indicated on the right

side (#CNA), where subjects with ≥3 CNAs were highlighted by purple. Subjects without any CNA are abbreviated. g, Mortality from hematological

malignancies in TP53-mutated cases with or without CNAs in 17p. h, Odds ratio for mortality from MDS calculated by multivariate logistic regression in

subjects with TP53-involving SNVs/indels. Error bars indicate 95% confidence intervals. We included unclassifiable CNAs involving 17p in 17p alterations

(17p alt.) in panel (g-h) because they are most likely to be LOH (UPDs or deletions). TP53-involving SNVs/indels in panel (f-h) included those detected by

ddPCR (Supplementary Fig. 3).

Nature Medicine | www.nature.com/naturemedicine

Articles

Nature Medicine

WBC (low)

WBC (high)

Hb (low)

Hb (high)

Plt (low)

Plt (high)

Cytopenia (All)

Cytopenia (Multi)

AS T2

PP L1

SF 3

SR 1

GN L

U2 2

Any abnormality

WBC (low)

Odds ratio

WBC (high)

> 10

Hb (low)

Hb (high)

Plt (low)

FDR

Plt (high)

<0.001

Cytopenia (All)

<0.01

Cytopenia (Multi)

<0.1

Extended Data Fig. 7 | See next page for caption.

Nature Medicine | www.nature.com/naturemedicine

1p 15

1q D

de D

de 5q)

l(1

11 3q)

9q D

9p D

l(2

>0.1

Any abnormality

Articles

Nature Medicine

Extended Data Fig. 7 | Genetic alterations in CH and abnormalities in blood counts. a, Landscape of SNVs/indels and CNAs in subjects without

abnormalities in blood counts (left), in those with any abnormalities in blood counts (middle), and in those with no available blood counts (right). Each

row represents a genetic alteration while each column represents a subject. Subjects without any alteration are omitted. Different types of mutations and

CNAs are depicted by different colors. b, Enrichment of genetic alterations in subjects with abnormalities in blood counts. Sizes of rectangles indicate

significance of enrichment. Colors of rectangles indicate odds ratios. The enrichment of alterations was examined by Fisher exact test. Cytopenia (All),

subjects with cytopenia in at least one lineage; Cytopenia (Multi), subjects with cytopenia in ≥2 lineage. WBC, white blood cell; Hb, hemoglobin; Plt,

platelet. c, Distribution of blood cell counts in subjects with different CH-related alterations. In all box plots, the median, first and third quartiles (Q1 and

Q3) are indicated, and whiskers extend to the furthest value between Q1 – 1.5×the interquartile range (IQR) and Q3 + 1.5×IQR. Numbers of subjects (n)

are indicated below the names of alterations. d, Relationships between blood cell counts and VAF of SNVs/indels or cell fractions of CNAs. P values are

calculated by two-sided t test in multivariate linear regression models, taking the effect of age and gender into account. Correction for multiple testing is

not performed.

Nature Medicine | www.nature.com/naturemedicine

Articles

Nature Medicine

Gender

(1.5%)

0.08

0.12

# of SNVs (n)

Cumulative moretality from

hematological maligancnies

Age

(14.5%)

CH (84%)

Attributable proportions of

CH-associated increase in HM mortality

≥3 (129)

1 (1,683)

0 (6,046)

≥3 (50)

7.4×10-4

2 (425)

0.06

# of CNA (n)

2 (229)

1.5×10-3

1 (1,334)

0.08

1.3×10-3

0 (6,670)

9.5×10-4

5.6×10-4

1.1×10-3

0.04

0.02

100 120 140

Time (month)

Cumulative mortality from

hematological malignancies

1 SNV

0.06

0.048

SNV+CNA (n=348)

SNV alone (n=1,335)

0.08

SNV+CNA (n=121)

P = 1.8×10-4

0.06

0.024

0.04

0.012

0.02

100 120 140

Time (month)

0.03

Both of SNV and CNA

(n=272)

P = 0.39

Either of SNV or CNA

(n=450)

No alteration

(n=4,947)

0.018

3 alterations

Both of SNV and CNA

(n=159)

0.048

P = 0.93

No alteration

(n=4,947)

0.024

100 120 140

Time (month)

No alteration

(n=4,947)

0.072

0.012

P = 0.71

Either of SNV or CNA

(n=16)

0.006

100 120 140

Both of SNV and CNA

(n=54)

0.096

0.048

4 alterations

Either of SNV or CNA

(n=87)

0.036

0.12

0.024

0.012

Time (month)

0.06

0.024

100 120 140

Time (month)

2 alterations

P = 0.047

SNV alone (n=68)

0.036

SNV+CNA (n=32)

P = 0.31

SNV alone (n=304)

100 120 140

3 SNV

0.1

0.036

Cumulative moretality from

hematological maligancnies

2 SNV

0.06

Time (month)

100 120 140

Time (month)

100 120 140

Time (month)

Extended Data Fig. 8 | Impact of CH on mortality from HM stratified by number of alterations. a, Pie chart showing the proportions of difference in

mortality from hematological malignancies (HM) between subjects with or without CH (Fig. 4a) which are attributable to each prognostic factor (Online

methods). b-c, Cumulative mortality from HM in subjects with different number of SNVs/indels (b), or CNAs (c). d-f, Cumulative mortality from HM in

subjects with both SNVs/indels and CNAs or in those with SNVs/indels alone. Subjects with 1 (d), 2 (e), or ≥3 alterations (f) are separately shown. g-i,

Cumulative mortality from HM in subjects with both SNVs/indels and CNAs or in those with either of them. Subjects with 2 (g), 3 (h), or 4 alterations (i)

are separately shown. Throughout the figure, P values were calculated by two-sided Wald test and not adjusted for multiple comparison.

Nature Medicine | www.nature.com/naturemedicine

Articles

All HM

Alteration

SNV

VAF <5%

VAF 5−10%

VAF >10%

#SNV 1

#SNV 2

#SNV >=3

3071

1960

485

626

2312

586

173

CNA

CF <5%

CF 5−10%

CF >10%

#CNA 1

#CNA 2

#CNA >=3

2254

1870

162

222

1841

322

all CH

SNV alone

CNA alone

4585

2331

1514

SNV+CNA

Same loci

Different loci

740

648

0.1

0.01

Alteration

Nature Medicine

Myeloid

100

0.01

0.1

All HM

Lymphoid

Odds ratio

100

0.01

0.1

Myeloid

Alteration

0.1

100

0.01

0.1

100

Lymphoid

100

0.1

Case

(n=67)

** )

overlap

(n=11)

Among 60,787 cases aged≥60 years and confirmed not to

have solid cancers as of March 2013, 52,472 had the follow

up data for development of HM.

Among 10,623 cases randomly selected from the 60,787

cases, 9,147 had the follow up data for development of HM.

Cumulative incidence of

hematological malignancies

100

0.01

1,214

198

835

17pUPD

16qUPD

20qUPD

11pUPD

9pUPD

13qUPD

del(13q)

1pUPD

+21q

14qUPD

165

0.1

100

Hazard ratio

CNA (CF≥5%)

0.0024

0.0012

0.0006

100

Time (month)

100

Both

Either

SNV alone

CNA alone

No alteration

0.0012

0.1

0.003

0.0024

0.01

0.003

0.0018

100

Hazard ratio

0.0018

0.1

CNA (CF<5%)

P = 0.019

GNAS

U2AF1

RUNX1

EZH2

JAK2

GNB1

TP53

DNMT3A

SF3B1

ASXL1

TET2

0.0036

0.1

Alteration

3,630

1,865

1,222

543

SNV (VAF<5%)

No alteration

0.1

All CH

SNV alone

CNA alone

SNV+CNA

0.006

SNV (VAF≥5%)

0.01

Any CNA 1,765

CF <5% 1,482

CF 5−10%

123

CF >10%

160

#CNA 1 1,477

#CNA 2

241

#CNA ≥3

0.0048

100

Any SNV 2,408

VAF <5% 1,591

VAF 5−10%

374

VAF >10%

443

#SNV 1 1,840

#SNV 2

454

#SNV ≥3

114

Subcohort

(n=9,147

Lymphoid

Odds ratio

Alteration

Case-cohort study for HM development

Target cohort (n=52,472 )

(≥60 y.o. and no cancer history)

0.01

Odds ratio

Myeloid

19pUPD 19

+9 16

del(7q) 19

+3q 13

17pUPD 26

13qUPD 17

+1q 14

del(13q) 41

del(1p) 10

22qUPD 18

+12 17

3pUPD 11

+8 26

del(17p) 13

12qUPD 21

11qUPD 40

+14 12

14qUPD 223

+1p 11

del(4q) 12

4qUPD 31

17qUPD 32

del(14q) 28

del(21q) 13

del(5q) 42

9pUPD 35

9qUPD 38

+21 109

1qUPD 53

del(2p) 19

del(3p) 10

1pUPD 81

+18 19

2qUPD 12

del(11q) 29

+17 12

7qUPD 10

15qUPD 14

11pUPD 21

6pUPD 32

+15 84

del(20q) 98

del(6q) 28

16qUPD 22

20qUPD 18

+22 29

16pUPD 24

19qUPD 12

2pUPD 20

U2AF1

EZH2

RUNX1

SRSF2

143

TP53

JAK2

159

PPM1D

SF3B1 115

CBL

TET2 1067

ASXL1 247

IDH2

DNMT3A 1521

GNB1

GNAS

KRAS

NRAS

0.01

All HM

Time (month)

100

Time (month)

Extended Data Fig. 9 | See next page for caption.

Nature Medicine | www.nature.com/naturemedicine

Nature Medicine

Articles

Extended Data Fig. 9 | Association of CH-related SNVs/indels and CNAs with hematological malignancies. a, Odds ratios for the events (death and/

or development) of hematological malignancies in case-control study (Extended Data Fig. 1a). Error bars indicate 95% confidence intervals. b, Design of

case-cohort study for development of hematological malignancies. c, Hazard ratios for development of hematological malignancies. Error bars indicate

95% confidence intervals. d-f, Effect of SNVs/indels (d), CNAs (e), and combined SNVs/indels and CNAs (f) on the cumulative incidence of development

of hematological malignancies. P values are calculated by two-sided Wald test. n, number of cases with the indicated alterations; SNV + CNA,

cooccurrence of both SNVs/indels and CNAs; #SNV, number of SNVs/indels; CF, cell fraction of CNAs; #CNA, number of CNAs.

Nature Medicine | www.nature.com/naturemedicine

Articles

Nature Medicine

Overall Survival

0.8

0.6

0.4

CNA

P value

SNV/indel

VAF <5% (n=1,146)

No SNV/indel (n=4966)

0.31

0.2

9.5×10-8

Cell fracrtion <5% (n=1,056)

100

120

140

No alteration (n=4,097)

100

120

140

Any SNV vs. No SNV

Any CNA vs. No CNA

Both vs SNV (VAF>5%) alone

CNA (CF<5%) vs. No CNA

Both vs CNA alone

SNV (VAF≥5%) vs. No SNV

CNA (CF≥5%) vs. No CNA

Both vs Either

0.1

0.4

P = 0.20

≥2 alterations

0.35

SNV+CNA (n=86)

0.4

SNV alone (n=332)

≥1 SNVs/indel + ≥1 CNA

(Max VAF>5%) (n=143)

0.28

P = 0.04

0.3

0.21

0.2

0.14

0.1

0.07

100

120

140

Time (month)

100

120

0.35

0.28

0.21

P = 0.041

0.07

0.06

Time (month)

120

140

120

140

P = 0.37

100

120

P = 0.90

0 (n=4,097)

0.18

No CH (n=4,097)

0.12

100

P = 0.014

2 (n=554)

1 (n=1,810)

0.14

≥3 (n=236)

0.24

0.14

0.3

3 SNV/indels

(Max VAF>5%) (n=28)

No CH (n=4,097)

Number of alterations

2 SNVs/indels + 1 CNA

or 1 SNVs/indels + 2 CNA

(Max VAF>5%) (n=50)

2 SNVs/indels

(Max VAF>5%) (n=96)

3 alterations

P = 0.091

0.28

Time (month)

0.35

0.21

1 ...

参考文献をもっと見る

分野

大学

学位論文種類・取得年

言語

Combined landscape of single-nucleotide variants and copy number alterations in clonal hematopoiesis

概要

この論文で使われている画像

関連論文

EYS is a major gene involved in retinitis pigmentosa in Japan: Genetic landscapes revealed by stepwise genetic screening.

Long-Read Sequence Confirmed a Large Deletion Including MYH6 and MYH7 in an Infant of Atrial Septal Defect and Atrial Arrhythmias.

Real-world application of next-generation sequencing-based test for surgically resectable colorectal cancer in clinical practice (本文)

Prenatal diagnosis of severe mitochondrial disease caused by nuclear gene defects: a study in Japan

Identification of a novel uterine leiomyoma GWAS locus in a Japanese population (本文)

参考文献

分野

大学

学位論文種類・取得年

言語

コピーが完了しました

URLをコピーしました

Combined landscape of single-nucleotide variants and copy number alterations in clonal hematopoiesis

概要

この論文で使われている画像

関連論文

EYS is a major gene involved in retinitis pigmentosa in Japan: Genetic landscapes revealed by stepwise genetic screening.

Long-Read Sequence Confirmed a Large Deletion Including MYH6 and MYH7 in an Infant of Atrial Septal Defect and Atrial Arrhythmias.

Real-world application of next-generation sequencing-based test for surgically resectable colorectal cancer in clinical practice (本文)

Prenatal diagnosis of severe mitochondrial disease caused by nuclear gene defects: a study in Japan

Identification of a novel uterine leiomyoma GWAS locus in a Japanese population (本文)

参考文献