Published Ahead of Print on May 18, 2017, as doi:10.3324/haematol.2017.166124.
Copyright 2017 Ferrata Storti Foundation.
Mutated ASXL1 and number of somatic mutations as possible
indicators of progression to chronic myelomonocytic leukemia of
myelodysplastic syndromes with single or multilineage dysplasia
by Ana Valencia-Martinez, Alessandro Sanna, Erico Masala, Elisa Contini, Alice Brogi,
Antonella Gozzini, and Valeria Santini
Haematologica 2017 [Epub ahead of print]
Citation: Valencia-Martinez A, Sanna A, Masala E, Contini E, Brogi A, Gozzini A, and Santini V.
Mutated ASXL1 and number of somatic mutations as possible indicators of progression to chronic
myelomonocytic leukemia of myelodysplastic syndromes with single or multilineage dysplasia
Haematologica. 2017; 102:xxx
doi:10.3324/haematol.2017.166124
Publisher's Disclaimer.
E-publishing ahead of print is increasingly important for the rapid dissemination of science.
Haematologica is, therefore, E-publishing PDF files of an early version of manuscripts that
have completed a regular peer review and have been accepted for publication. E-publishing
of this PDF file has been approved by the authors. After having E-published Ahead of Print,
manuscripts will then undergo technical and English editing, typesetting, proof correction and
be presented for the authors' final approval; the final version of the manuscript will then
appear in print on a regular issue of the journal. All legal disclaimers that apply to the
journal also pertain to this production process.
Mutated ASXL1 and number of somatic mutations as possible indicators of progression to
chronic myelomonocytic leukemia of myelodysplastic syndromes with single or multilineage
dysplasia.
Ana Valencia-Martinez1, Alessandro Sanna1, Erico Masala1, Elisa Contini2, Alice Brogi1,3,
Antonella Gozzini1, and Valeria Santini1
1
Department of Hematology, Università degli Studi di Firenze, AOU Careggi, Florence, Italy.
2
SOD Diagnostica Genetica. AOU Careggi. Florence, Italy.
3
Department of Biotechnologies, University of Siena. Siena, Italy
Address for correspondence:
Valeria Santini, MD.
Functional Unit of Haematology
AOU Careggi. University of Florence.
Largo Brambilla 3, 50134. Florence, Italy.
Tel.: +390557947296; Fax: +390557947343
E-mail: [email protected]
Running title: Role of mutations in the progression MDS-to-CMML
Conflict of interest: The authors declare no conflict of interest
1
Myelodysplastic syndromes (MDS) are heterogeneous clonal disorders characterized by ineffective
hematopoiesis and transformation in acute myeloid leukemia (AML).1However, disease progression
is not always to overt AML: rarely MDS may develop myeloproliferative features, and move to
overlap-diseases (MDS/MPN) like chronic myelomonocytic leukemia (CMML).
Some authors reported unexpected transformation of de novo MDS to CMML, independently of the
presence of relative monocytosis (>10%) at initial diagnosis of MDS.2,3 In these reports, apart from
an evident increment of total white blood cells count (WBC) and monocytosis, alterations of other
clinical parameters and/or cytogenetic and molecular abnormalities failed to be associated directly
with progression.3 Deep sequencing demonstrated the high frequency of somatic mutations in MDS
as well as the acquisition of additional mutations during disease history.4,5 Some somatic mutations
are often shared by MDS and MDS/MPN and contribute differently to the prognosis of these
disorders.6-8
Padron et al.4 identified similar pattern of mutations at diagnosis and in the myeloproliferative stage
of MDS patients that transformed to CMML, with maintained variant allele frequency (VAF). The
acquisition of new mutations was nevertheless evident at progression (ETV6, NRAS, CBL, and
RUNX1). Differently, but in the same line of multi-step clonal evolution, it was observed that MDS
progressing to AML acquired mutations of RUNX1, ETV6, PHF6, NRAS and KRAS.5
We molecularly describe here a cohort of lower risk MDS patients that evolved to CMML. All
consecutive cases, diagnosed at Azienda Ospedaliero Universitaria Careggi (Firenze, Italy) as MDS
in accordance with WHO 2008 criteria between April 2010 and September 2016, were revised. Out
of 360 cases analysed, we identified 9 patients reclassified with revised WHO 20169 presenting
diagnosis of MDS with single lineage dysplasia (MDS-SLD) or MDS with multilineage dysplasia
(MDS-MLD) who progressed to overt CMML after variable follow-up periods (Table 1) (referred
as Group 1). In all patients, absolute and relative monocytosis was absent during the entire duration
of MDS and the median time to CMML transformation was 21 months (range, 10-41) (Table 1).
Paired DNA samples of the MDS and CMML phases were isolated in all 9 cases. Bone marrow
(BM) mononuclear cells were collected and cryopreserved at diagnosis (MDS phase) and at
transformation (CMML phase). Informed consent was provided in accordance with the Declaration
of Helsinki Principles and local Ethics Committee’s approval. Twenty-two candidate genes were
selected for sequencing analysis (See Supplementary data for methods), referred in literature as
most frequently mutated in MDS and CMML.10-12 ASXL1, EZH2, TP53, SF3B1, U2AF1, SRSF2,
TET2, DNMT3A, ETV6, RUNX1, NPM1, FLT3, CBL, SETBP1, CSF3R, CEBPA, IDH1, IDH2,
2
JAK2, MPL, CARL, NRAS. Characteristics of the patients are presented in Supplementary Table 1.
In brief, five were males, and the median age at diagnosis was 67 years (range, 56–76 years). At
diagnosis, four patients had MDS-SLD and five patients had MDS-MLD. At transformation, three
patients presented a CMML-0, four patients a CMML-1 and two patients a CMML-2 according the
WHO classification 2016. WBC, absolute monocyte counts (AMC) and blast count in CMML
phase were statistically higher than in MDS phase (WBC mean, 3.64 × 109/L vs. 13 × 109/L, P=
0.001; AMC, 0.21 × 109/L vs. 1,6 × 109/L, P= 0.004; blast count, 0% versus 6%; P= 0.0001). Mean
hemoglobin level (10.85 g/dL versus 9.7 g/dL; P= 0.42) and platelet count (147× 109/L versus 86×
109/L; P= 0.11) were not significantly different between the MDS and CMML phases. After the
initial diagnosis of MDS, all patients were treated exclusively with erythropoiesis-stimulating
agents and all patients with and without stable disease were followed every three months. In both
phases, five patients presented a normal karyotype, one patient a monosomy 7, one patient a trisomy
8 and one patient a trisomy 14. Only case #1 presented a new trisomy 8 at progression (Table 1).
After sequencing, coverage across targeted regions was 2500X (minimum 316 and maximum
9877). In the MDS phase, somatic mutations were identified in 9 genes, resulting in 29 variants. All
mutations were heterozygous, and comprised 10 missense, 10 nonsense and 9 frameshift mutations.
Eight of the 9 CMML-progressing cases presented three mutations or more (median number of
mutations: 3, range 2-5) (Supplementary Figure 1). The mutational analysis of the CMML phase
confirmed the presence of all the 29 variants found at diagnosis (Figure 1). Generally, mutation
VAF did not change significantly, while some mutated clones were marginally expanded during
progression. An acquisition of two missense variants was presented in the CMML samples: case #3
with a new variant of the DNMT3A gene and case #7 with a new variant of the gene NRAS (Figure
1). In our cohort of patients, the acquisition of new somatic mutations was not closely correlated
with progression, as in 7 of the 9 cases studied was absent. We then compared the mutational
profile of Group 1 with a second cohort of lower risk MDS patients that did not present CMML
transformation or progression to higher risk MDS or AML (referred as Group 2). Clinical data of
Group 2 are summarized in Supplementary Table 1. Group 1 and 2 were matched by age, sex and
IPSS risk; Group 2 had a longer OS than Group 1 (67 vs. 38 months) (Supplementary Figure 2).
After sequencing of Group 2, coverage across the targeted regions of 3441X (minimum 410 and
maximum 11401) was obtained. Thirteen cases (28%) did not present any mutation in the 22 genes
analysed, and 33 cases (72%) presented at least one mutation (median number of mutations: 1,
range 1-3). The first important observation is that Group 1 cases presented with a higher number of
mutated genes: 88% cases of Group 1 showed ≥ 3 mutations compared with 2 cases (4%) of Group
2 (Figure 2 and Supplementary Figure 1). The second observation is that the most frequently
3
mutated genes in Group 1 were ASXL1 and TET2 that were mutated in all cases and in 6/9 cases,
respectively (Figure 2A). Consistently to literature, TET2, SF3B1 and DNMT3A, were the most
frequently mutated genes in Group 2 cases (19%, 24% and 28%, respectively),10,11 while ASXL1
was present in only three cases (7%): in two cases as an isolated mutation and in one case in
combination with TET2 and SRSF2 (Figure 2B). The latter case has been censored because received
an early hematopoietic stem cell transplant, and therefore was not possible to evaluate progression.
Group 1 cases shared the constant presence of ASXL1 mutations. Based on average sequencing
depth versus VAF, ASXL1 mutations in Group 1 appear both as clonal and subclonal.Moreover,
VAF of ASXL1 in cases #3 and #8 did not vary significantly while other mutations presented in the
same samples increased in size (Figure 1). It seems that together with the presence of ASXL1
mutations, the increase in size of other clones could be related with the disease progression. ASXL1
mutation is present in 45% of CMML cases and has adverse prognostic impact on OS.12,13 TET2
mutations, the second most frequent in Group 1, are widely prevalent in CMML (46%) and their
prognostic relevance remains unclear14 ; in our series were not present in the totality of our MDSto-CMML progressing cases. Co-mutation of ASXL1 and TET2 is associated with very poor
survival in CMML.14 In order to evaluate the weight of ASXL1 and TET2 mutation positive clones
in this peculiar MDS-to-CMML progression, we compared VAF data of the MDS and CMML
phases using the paired sign test. VAF of ASXL1 was significantly higher in CMML than in the
corresponding MDS phase (42% versus 29%, respectively; P= 0.039) while VAF of TET2 was not
significantly varied in both phases (55% versus 46%, respectively; P= 0.38) (Supplementary Figure
3A-B), suggesting a determinant and prominent role of early presence and increase in the burden of
mutated ASXL1, during progression of the disease to myeloproliferative features. To be stressed that
in matched lower risk MDS cases of Group 2 without any kind of progression, the presence of
ASXL1 mutation was a rare event (7%). Although limited by the number of cases evaluated, no
significant changes were observed in the mutation profile during progression to CMML. Admitting
that the presence of acquired mutations in other genes not included in this study cannot be excluded,
nevertheless a strong suggestion of the possibility to identify a “pre-CMML state” derives from our
observations. All “pre-CMML” cases were characterized by worsening anemia that required
treatment, in line with the role of ASXL1 in erythropoiesis.15 These peculiar “pre-CMML”-MDS
cases seem to be well characterized by enhanced genetic instability, justifying multiple comutations, and by the constant feature of early ASXL1 mutation. To note, none of the patients with a
“pre-CMML” MDS presented TP53 mutations, very rarely found in CMML. This suggests that the
presence of ASXL1 mutations and TP53 wild type could be part of this definition of “pre-CMML”
state. Although this kind of progression MDS-to-CMML is quite rare, it looks worth pursuing
4
studies with more such cases to confirm whether this kind of evolution is predictable and therefore
justify inception of earlier therapy in lower risk MDS with ASXL1 mutations.
CONFLICT OF INTEREST
The authors declare no conflict of interest.
ACKNOWLEDGMENTS FOR RESEARCH SUPPORT: This work was supported by
Associazione Italiana Ricerca sul Cancro (AIRC: IG 2015 Id 16861).
REFERENCES
1. Steensma DP, Bennett JM. The myelodysplastic syndromes: diagnosis and treatment. Mayo Clin
Proc. 2006;81(1):104-130.
2. Rigolin GM, Cuneo A, Roberti MG, et al. Myelodysplastic syndromes with monocytic
component: hematologic and cytogenetic characterization. Haematologica. 1997;82(1):25-30.
3. Wang SA, Galili N, Cerny J, et al. Chronic myelomonocytic leukemia evolving from preexisting
myelodysplasia shares many features with de novo disease. Am J Clin Pathol. 2006;126(5):789797.
4. Padron E, Yoder S, Kunigal S, et al. ETV6 and signaling gene mutations are associated with
secondary transformation of myelodysplastic syndromes to chronic myelomonocytic leukemia.
Blood. 2014;123(23):3675-3677.
5. Pellagatti A, Roy S, Di Genua C, et al. Targeted resequencing analysis of 31 genes commonly
mutated in myeloid disorders in serial samples from myelodysplastic syndrome patients showing
disease progression. Leukemia. 2016;30(1):247-250.
6. Maxson JE, Gotlib J, Pollyea DA, et al. Oncogenic CSF3R mutations in chronic neutrophilic
eukemia and atypical CML. N Engl J Med. 2013;368(19):1781-1790.
7. Meggendorfer M, Roller A, Haferlach T, et al. SRSF2 mutations in 275 cases with
chronicmyelomonocytic leukemia (CMML). Blood. 20012;120(15):3080-3088.
8. Papaemmanuil E, Cazzola M, Boultwood J, et al. Somatic SF3B1 mutation in myelodysplasia
with ring sideroblasts. N Engl J Med. 2011;365(15):1384-1395.
9. Arber DA, Orazi A, Hasserjian R, et al. The 2016 revision to the World Health Organization
classification of myeloid neoplasms and acute leukemia. Blood. 2016;127(20):2391-2405.
5
10. Papaemmanuil E, Gerstung M, Malcovati L, et al. Clinical and biological implications of driver
mutations in myelodysplastic syndromes. Blood. 2013;122(22):3616-3627.
11. Haferlach T, Nagata Y, Grossmann V, et al. Landscape of genetic lesions in 944 patients with
myelodysplastic syndromes. Leukemia. 2013;28(2):241-247.
12. Itzykson R, Kosmider O, Renneville A, et al. Prognostic score including gene mutations in
chronic myelomonocytic leukemia. J Clin Oncol. 2013;31(19):2428–2436.
13. Patnaik MM, Itzykson R, Lasho TL, et al. ASXL1 and SETBP1 mutations and their prognostic
contribution in chronic myelomonocytic leukemia: a two-center study of 466 patients. Leukemia.
2014;28(11):2206–2212.
14. Patnaik MM, Lasho TL, Vijayvargiya P, et al. Prognostic interaction between ASXL1 and
TET2 mutations in chronic myelomonocytic leukemia. Blood Cancer J. 2016;6:e385.
15. Shi H, Yamamoto S, Sheng M, et al. ASXL1 plays an important role in erythropoiesis. Sci Rep.
2016;6:28789.
6
Table 1. Clinical and cytogenetic characteristics of MDS patients that progress to CMML
Case#
**
4
5
6
7
8
9
70
76
63
67
71
72
64
Female
Male
Male
Mal e
Mal e
Male
Female
Female
12,1
21, 2
39,2
10,5
17,0
20,6
41, 6
27,6
MDS-
MDSMDS-SLD
MDS-SLD
MDS-SLD
MDS-MLD
MLD
MLD
Int-1 / Int
Int-1 / Int
1
2
Age
63
56
Sex
Female
20,8
3
Ti me to transformati on
CMML (months)
*WHO classification
MDSMDS-MLD
MDS-SLD
MLD
Low / Very
Int-1 /
Low /
low
Low
Very low
MDS
4,04
2,24
10, 3
2,79
5,7
3,64
1,74
4,54
2,40
CMML
13,2
12,1
59, 1
24,4
13
4,9
33,3
12, 1
4,5
MDS
2,2
1,7
6,8
2,01
3,4
2,1
1,2
2,66
1,9
CMML
9,4
8,2
56, 2
10,5
8,7
2,7
26,4
6,3
2,7
0,48
0,12
0,72
0,25
0,30
0,44
0,06
0,17
0,03
CMML
2,1
1,6
1,48
8,6
3,60
1,60
5,3
1,4
1,5
MDS
9,4
5,4
6,9
8,9
5,2
9,3
3,5
3,7
1,3
15,9
13,2
2,5
35,2
27,7
32,6
16
11, 5
33,3
0
0
0
4
4
0
4
0
0
12
4
6
9
19
5
9
4
0
11,2
12,4
8
5,6
14,7
10,8
10,4
10, 5
13,3
CMML
7,4
9,7
8,8
7,2
15
11,1
9,1
10, 1
12,3
MDS
145
78
173
208
86
145
37
354
149
80
82
94
110
45
142
30
86
104
Normal
-7
Normal
+14
+8
Normal
Normal
Normal
Normal
CMML
+8
-7
Normal
+14
+8
Normal
Normal
Normal
Normal
MDS
No
No
No
No
No
No
No
No
No
CMML
Yes
No
Yes
Yes
Yes
No
No
Yes
No
8,8
NA
NA
NA
6,9
NA
NA
3,1
NA
IPSS/IPSS-R
Low /
Low /
Int-1 / Int
Very l ow
Low / Low
Very low
WBC count
9
(x10 /L)
ANC count
9
(x10 /L)
MDS
AMC count
9
(x10 /L)
Monocytes (%)
CMML
MDS
BM blasts (%)
CMML
MDS
Hemoglobi n
(g/dL)
Platelets
(x109/L)
CMML
MDS
Cytogenetics
Organomegal y
Ti me AML evolution
(months)
* At initial diagnosis of MDS
7
** To note that for case #3, criteria for diagnosis of CMML are not completely fulfilled. In fact in
this case, the percentage of monocytes is < 10%, level required in WHO2016 definition of CMML.
On the other hand, this patient cannot be diagnosed as atypical CML because the first condition
according to WHO for such diagnosis namely “PB leukocytosis due to increased numbers of
neutrophils and their precursors (promyelocytes, myelocytes, metamyelocytes) comprising ≥10% of
leukocytes” is lacking. MDS/MPN-unclassifiable is also not applicable because AMC in the
progression to MDS/MPN was > 1000/uL.
Abbreviations: CMML, chronic myelomonocytic leukemia; WHO, world health organization;
MDS-SLD, myelodysplastic syndrome with single lineage dysplasia; MDS-MLD, myelodysplastic
syndrome with multilineage lineage dysplasia; IPSS, international prognostic score system; IPSS-R,
revised international prognostic score system; Int-1, intermediate-1; Int, Intermediate; WBC, white
blood cells; ANC, absolute neutrophils count; AMC, absolute monocyte count; BM, bone marrow;
AML, acute myeloid leukemia; NA, not applicable.
FIGURE LEGENDS
Figure 1. Representation of mutations and VAF (%) in the paired DNA samples of the MDS
and CMML phases in the Group 1 of patients. Every graph represents cases #1 to #9 of Group 1.
Genes are colour coded and VAFs (axis y) and protein change (axis z) are represented for each
variant. Our panel gene was as follows but only the first nine genes were mutated in the Group1 of
patients at MDS diagnosis: ASXL1, EZH2, RUNX1, SETBP1, SF3B1, U2AF1, SRSF2, TET2,
DNMT3A, TP53, ETV6, NPM1, FLT3, CBL, CSF3R, CEBPA, IDH1, IDH2, JAK2, MPL, CARL
Figure 2. Mutated genes in patients included in Group 1 (A) and Group 2 (B). (A) All nine
patients of Group 1 presented at least two mutations. (B) In Group 2, 13 cases (28%) did not present
any mutation and are all represented in the last grey column. Genes are colour coded and every case
is represented in a column
8
Supplementary data for methods
The SureDesing of Agilent Technologies was used to design a custom enrichment of the candidate
genes: ASXL1, EZH2, TP53, SF3B1, U2AF1, SRSF2, TET2, DNMT3A, ETV6, RUNX1, NPM1,
FLT3, CBL, SETBP1, CSF3R, CEBPA, IDH1, IDH2, JAK2, MPL, CARL, NRAS. Genes were
selected after a manually curated literature screening of the most commonly mutated genes in MDS
and CMML. Library preparation was performed using the HaloPlex target enrichment protocol
(Agilent technologies, Santa Clara, USA). The genomic DNA input for amplicon library
preparation was 225 ng for each sample according to manufacturer’s instructions. All sample
libraries were equimolarly pooled and sequenced on the Illumina MiSeq Sequencer (Illumina, San
Diego, CA, USA) with a default 150 bp paired-end reads protocol, according to the manufacturer’s
instructions.
Data analysis and processing
Reads
quality
was
checked
with
FastQC
(http://www.bioinformatics.babraham.ac.uk/projects/fastqc) and reads were aligned to the reference
human genome hg19 with BWA mem [Li, Heng. "Aligning sequence reads, clone sequences and
assembly contigs with BWA-MEM." arXiv preprint arXiv:1303.3997 (2013)].Genome Analysis
Toolkit (GATK) version 3.3 was used to recalibrate base qualities and realign aligned reads around
indels. Coverage statistics was performed by DepthOfCoverage utility of GATK. BASH and R
custom scripts were used to obtain the list of low coverage (≤100X) regions per sample.
For the detection of low abundance variants: SNVs were identified using MuTect version 1.1.4 with
standard parameters and GATK IndelGenotyperV2 was used to detect InDels. Genomic and
functional annotation of detected variants was made by Annotate Variation Software (ANNOVAR,
version 14 Dec, 2015). Detected variants were distilled on the basis of their exonic function, allele
frequency, the presence in variants databases, such as the current release of dbSNP
(http://www.ncbi.nlm.nih.gov/projects/SNP/), 1000 Genomes Project (http://ftp.ncbi.nih.gov/),
Exome Sequencing Project (http://evs.gs.washington.edu/EVS/) and several prediction scores
(SIFT, MutationTaster, PolyPhen2, Gerp and PhyloP). Allelic depths for the reference and altered
alleles were used to calculate the variant allele frequency (VAF) of all variants. All variants
showing a VAF >5% were extracted and annotated with ANNOVAR. Candidate somatic variants
were filtered using the following criteria:
1. Removal of all synonymous variants and noncoding variants more than 6 bases from
splice junctions were discarded;
2. Removal of missense SNVs which were positioned in coding regions and filtered as
polymorphisms when corresponding to known SNPs in public and private databases,
including dbSNP144, ESP6500 (version march 2015), the 1000 genomes project
(version 1 000g2015aug), and our in-house database at a population frequency of 1% or
more;
3. Removal of variants present within regions prone to sequence context specific artifacts,
including regions enriched for reads of low mapping quality that harbor multiple
mismatches;
4. Removal of variants generated by incorrect calling after a manually revision using The
Integrative Genomics Viewer (IGV)
5. Retention of variants present in the COSMIC (v70), ClinVar (v 20150629) and in-house
databases as known somatic variation in myeloid and other common malignancies.
6. Remaining variants were the somatic origin was confirmed after and accurate and
extensive research of the PubMed database (www.pubmed.gov).
Supplementary Table 1. Clinical and cytogenetic characteristics of patients of Group 2
Characteristics
Age, years
Sex
Male
Female
WBC count (x109/L)
ANC count (x109/L)
AMC count (x109/L)
Haemoglobin (g/dL)
Platelets (x109/L)
WHO Classification
MDS-SLD
MDS-RS-SLD
MDS-MLD
del(5q)
Cytogenetics
Normal
Aberrant
IPSS risk group
Low
Intermediate-1
IPSS-R risk group
Very Low
Low
Intermediate
Median (range)
n = 46
%
26
20
56
44
17
2
24
3
37
4
52
7
32
14
69
31
28
18
61
39
17
20
9
36
42
19
65 (43-88)
4.15 (2.7-15.7)
2.3 (0.8-14.8)
0.4 (0.02-0.70)
10.25 (8.1-13.11)
140 (17-722)
Abbreviations: AMC, absolute monocyte count; WHO, world health organization; MDS-SLD,
myelodyplastic syndrome with single lineage dysplasia; MDS-RS-SLD, myelodyplastic syndromering syderoblasts with single lineage dysplasia; MDS-MLD, myelodyplastic syndrome with
multilineage lineage dysplasia; IPSS, international prognostic score system; IPSS-R, revised
international prognostic score system
Supplementary Figure 2. Overall survival curves for Group 1 (dashed line) and 2 (solid line):
median OS values, 38 months vs. 67 months, respectively; P=0.015.
Supplementary Figure 3. (A) VAF value of ASXL1 and (B) TET2 mutations in case #1 to #9 of
Group 1 of patients in the MDS phase (grey bars) and CMML phase (black bars). (A) The median
VAF value of ASXL1 mutations was significantly higher in CMML than in the corresponding MDS
phase (42% vs. 29%, respectively; P= 0.039) while the median VAF value of TET2 mutations (B)
did not significantly varied in both phases (46% vs. 55%; P=0.38).
A
70
ASXL1 VAF (%)
60
50
40
30
20
10
0
#1
#2
#3
#4
#5
#6
#7
#8
#9
B
90
80
TET2 VAF (%)
70
60
50
40
30
20
10
0
#1
#2
#3
#4
#5
#6
#7
#8
#9