Page 1 of 9
Accepted Preprint first posted on 20 August 2014 as Manuscript ERC-14-0361
1
Deep sequencing of KIT, MET, PIK3CA, and PTEN hotspots in Papillary
2
Thyroid Carcinomas with distant metastases
3 4
Greta Gandolfi1,#, Dario de Biase2,4#, Valentina Sancisi1, Moira Ragazzi3, Giorgia Acquaviva2, Annalisa
5
Pession2, Simonetta Piana3, Giovanni Tallini4, Alessia Ciarrocchi1,*
6 7
1) Laboratory of Translational Research, Arcispedale S. Maria Nuova - IRCCS, Reggio Emilia 42123, Italy
8
2) Department of Pharmacology and Biotechnology (FaBiT), University of Bologna, 40139 Bologna, Italy
9
3) Pathology Unit, Department of Oncology, Arcispedale S. Maria Nuova - IRCCS, Reggio Emilia 42123,
10
Italy
11
4) Department of Medicine (DIMES) – Anatomic Pathology Unit, Bellaria Hospital, University of Bologna,
12
40139 Bologna, Italy
13 14
#These authors contributed equally to this work
15
*Correspondence should be addressed to:
16
Alessia Ciarrocchi
17
Laboratory of Translational Research,
18
Azienda Ospedaliera “Arcispedale S. Maria Nuova”- IRCCS,
19
Viale Risorgimento 80, 42123 Reggio Emilia, Italy
20
Tel +39 0522 295668 Fax +39 0522 295743
21
[email protected]
22 23
Abbreviated Title: gene mutations in PTCs with distant metastases
24
Keywords: papillary thyroid carcinomas, distant metastases, genetic alterations, next generation sequencing,
25
KIT, MET, PIK3CA, PTEN
26
Word Count: 1343
Copyright © 2014 by the Society for Endocrinology.
Page 2 of 9
27
Dear Editor,
28
Well-differentiated papillary thyroid carcinoma (PTC) accounts for approximately 90% of all thyroid
29
cancers. While the majority of these tumors tend to behave as indolent lesions, a fraction of PTCs is highly
30
aggressive and results in disseminated systemic spread to distant sites (Pellegriti et al. 2013). Numerous
31
studies attempted to define prognostic markers able to discriminate at diagnosis PTCs with more aggressive
32
behavior from those with an indolent course (Handkiewicz-Junak et al. 2010). Nonetheless, the usefulness of
33
genetic analysis in PTC patient management is still controversial, probably due to the fact that our
34
knowledge about the genetic asset of aggressive PTCs is still very limited. The low occurrence of distant
35
metastases and death among the overall PTCs cases, and hence the difficulty in collecting large cohorts of
36
PTCs which developed distant metastases (DM-PTCs), has been an important limitation for studies that
37
attempted to correlate genetic alterations to prognosis and outcome of PTC patients.
38
An additional level of complexity on this issue comes from the genetic heterogeneity of tumors.
39
Mutations that are drivers of aggressiveness are likely to occur as late secondary events during tumor
40
progression and to be confined to small sub-populations of cells, which may compromise the possibility of
41
detecting them by using the standard sequencing technique.
42
Based on this consideration, we investigated a large cohort of 39 DM-PTCs and 20 control-PTCs
43
(without distant metastases), by performing a deep sequencing analysis of 8 mutational hotspots in KIT,
44
MET, PI3KCA, and PTEN genes. These hotspots were chosen since they were associated with aggressive
45
features of various epithelial cancers, but they have been scarcely investigated or rarely found mutated in the
46
general population of non-aggressive well-differentiated PTCs (Garcia-Rostan et al. 2005; Wasenius et al.
47
2005; Hou et al. 2007; Broecker-Preuss et al. 2008; Santarpia et al. 2008; Xing 2010).
48
Next-generation sequencing analysis of these hotspots (KIT exons 9, 11 and 13; MET exons 2 and 14;
49
PIK3CA exons 10 and 21, and PTEN exon 5) was performed on FFPE-extracted DNA using the 454 GS-
50
Junior Next Generation sequencer (Roche Diagnostics, Germany), with the cut-off of 5% mutated read
51
percentage. A total of 79 mutations were detected in the overall cohort of samples. All mutations were single
52
nucleotide variants. The number of total alterations detected in the DM-PTCs was strikingly higher than in
53
the control group. Indeed, 76 mutations were detected in the DM-PTCs while only 5 were observed in the
54
control-PTCs. Two mutations were found in both groups (Figure 1A). Of the 76 mutations found in DM-
Page 3 of 9
55
PTCs, 64 were non-synonymous (NS), including 5 stop-codon mutations and 59 missense mutations. Among
56
the 64 NS mutations, 22 were previously reported in databases (COSMIC and Ensembl) or publications, and
57
42 were not described. Of the 5 mutations found in the control-PTCs group, 4 were missense, including the
58
two mutations shared with the DM-PTCs (METp.E168D and PIK3CAp.R524K). One of the 4 NS mutations
59
was a novel mutation. Most mutations were found in a single sample, and no mutation was found in more
60
than 4 samples.
61
As well, the number of mutational events was significantly higher in DM-PTCs than in control-PTCs,
62
with 82 events in DM-PTCs vs. 9 events in control-PTCs (2.10 ± 2.95 vs. 0.45 ± 0.60 events per sample
63
respectively; P=0.02). Within the control-PTCs group, 12 tumor developed lymph node metastasis (LNM)
64
while 8 did not form metastases at any site and were confined to the thyroid. We did not observe difference
65
in the number of mutational events between these two subgroups (0.5 ± 0.5 vs. 0.35 ± 0.74 event per
66
sample, respectively; n.s.).
67
Figure 1B depicts the number and frequencies of PTCs with at least one NS mutation in the DM-
68
PTC and in the control-PTC groups, and the distribution of NS mutations in the four genes analyzed.
69
Twenty-three DM-PTCs (0.59) showed at least one NS mutation in the target gene regions analyzed. Nine
70
DM-PTCs (0.23) showed only one mutation, while 14 DM-PTCs (0.36) presented more than one mutation,
71
up to 13 different mutations (Figure 1C, continuous red line). In the control-PTCs group, 7 out of 20 PTCs
72
showed at least one NS mutation (0.35), of which 6 PTCs showed only one mutation (0.30), and one PTC
73
harbored two mutations (0.05) (Figure 1C, dashed blue line). The average number of NS mutations per
74
sample was 1.77 ± 2.59 in DM-PTCs vs. 0.40 ± 0.60 in control-PTCs (P=0.02).
75
The complete list of the detected mutations is reported in Figure 1D.
76
The percentage of mutated reads of NS mutations ranged from 5% to 69.7% in the DM-PTCs, and
77
from 22.9% to 47.9% in control-PTCs. The average mutated allele percentage was significantly lower in
78
DM-PTCs than in the control group (12.32 ± 10.39% vs. 35.29 ± 11.37%; P
Accepted Preprint first posted on 20 August 2014 as Manuscript ERC-14-0361
1
Deep sequencing of KIT, MET, PIK3CA, and PTEN hotspots in Papillary
2
Thyroid Carcinomas with distant metastases
3 4
Greta Gandolfi1,#, Dario de Biase2,4#, Valentina Sancisi1, Moira Ragazzi3, Giorgia Acquaviva2, Annalisa
5
Pession2, Simonetta Piana3, Giovanni Tallini4, Alessia Ciarrocchi1,*
6 7
1) Laboratory of Translational Research, Arcispedale S. Maria Nuova - IRCCS, Reggio Emilia 42123, Italy
8
2) Department of Pharmacology and Biotechnology (FaBiT), University of Bologna, 40139 Bologna, Italy
9
3) Pathology Unit, Department of Oncology, Arcispedale S. Maria Nuova - IRCCS, Reggio Emilia 42123,
10
Italy
11
4) Department of Medicine (DIMES) – Anatomic Pathology Unit, Bellaria Hospital, University of Bologna,
12
40139 Bologna, Italy
13 14
#These authors contributed equally to this work
15
*Correspondence should be addressed to:
16
Alessia Ciarrocchi
17
Laboratory of Translational Research,
18
Azienda Ospedaliera “Arcispedale S. Maria Nuova”- IRCCS,
19
Viale Risorgimento 80, 42123 Reggio Emilia, Italy
20
Tel +39 0522 295668 Fax +39 0522 295743
21
[email protected]
22 23
Abbreviated Title: gene mutations in PTCs with distant metastases
24
Keywords: papillary thyroid carcinomas, distant metastases, genetic alterations, next generation sequencing,
25
KIT, MET, PIK3CA, PTEN
26
Word Count: 1343
Copyright © 2014 by the Society for Endocrinology.
Page 2 of 9
27
Dear Editor,
28
Well-differentiated papillary thyroid carcinoma (PTC) accounts for approximately 90% of all thyroid
29
cancers. While the majority of these tumors tend to behave as indolent lesions, a fraction of PTCs is highly
30
aggressive and results in disseminated systemic spread to distant sites (Pellegriti et al. 2013). Numerous
31
studies attempted to define prognostic markers able to discriminate at diagnosis PTCs with more aggressive
32
behavior from those with an indolent course (Handkiewicz-Junak et al. 2010). Nonetheless, the usefulness of
33
genetic analysis in PTC patient management is still controversial, probably due to the fact that our
34
knowledge about the genetic asset of aggressive PTCs is still very limited. The low occurrence of distant
35
metastases and death among the overall PTCs cases, and hence the difficulty in collecting large cohorts of
36
PTCs which developed distant metastases (DM-PTCs), has been an important limitation for studies that
37
attempted to correlate genetic alterations to prognosis and outcome of PTC patients.
38
An additional level of complexity on this issue comes from the genetic heterogeneity of tumors.
39
Mutations that are drivers of aggressiveness are likely to occur as late secondary events during tumor
40
progression and to be confined to small sub-populations of cells, which may compromise the possibility of
41
detecting them by using the standard sequencing technique.
42
Based on this consideration, we investigated a large cohort of 39 DM-PTCs and 20 control-PTCs
43
(without distant metastases), by performing a deep sequencing analysis of 8 mutational hotspots in KIT,
44
MET, PI3KCA, and PTEN genes. These hotspots were chosen since they were associated with aggressive
45
features of various epithelial cancers, but they have been scarcely investigated or rarely found mutated in the
46
general population of non-aggressive well-differentiated PTCs (Garcia-Rostan et al. 2005; Wasenius et al.
47
2005; Hou et al. 2007; Broecker-Preuss et al. 2008; Santarpia et al. 2008; Xing 2010).
48
Next-generation sequencing analysis of these hotspots (KIT exons 9, 11 and 13; MET exons 2 and 14;
49
PIK3CA exons 10 and 21, and PTEN exon 5) was performed on FFPE-extracted DNA using the 454 GS-
50
Junior Next Generation sequencer (Roche Diagnostics, Germany), with the cut-off of 5% mutated read
51
percentage. A total of 79 mutations were detected in the overall cohort of samples. All mutations were single
52
nucleotide variants. The number of total alterations detected in the DM-PTCs was strikingly higher than in
53
the control group. Indeed, 76 mutations were detected in the DM-PTCs while only 5 were observed in the
54
control-PTCs. Two mutations were found in both groups (Figure 1A). Of the 76 mutations found in DM-
Page 3 of 9
55
PTCs, 64 were non-synonymous (NS), including 5 stop-codon mutations and 59 missense mutations. Among
56
the 64 NS mutations, 22 were previously reported in databases (COSMIC and Ensembl) or publications, and
57
42 were not described. Of the 5 mutations found in the control-PTCs group, 4 were missense, including the
58
two mutations shared with the DM-PTCs (METp.E168D and PIK3CAp.R524K). One of the 4 NS mutations
59
was a novel mutation. Most mutations were found in a single sample, and no mutation was found in more
60
than 4 samples.
61
As well, the number of mutational events was significantly higher in DM-PTCs than in control-PTCs,
62
with 82 events in DM-PTCs vs. 9 events in control-PTCs (2.10 ± 2.95 vs. 0.45 ± 0.60 events per sample
63
respectively; P=0.02). Within the control-PTCs group, 12 tumor developed lymph node metastasis (LNM)
64
while 8 did not form metastases at any site and were confined to the thyroid. We did not observe difference
65
in the number of mutational events between these two subgroups (0.5 ± 0.5 vs. 0.35 ± 0.74 event per
66
sample, respectively; n.s.).
67
Figure 1B depicts the number and frequencies of PTCs with at least one NS mutation in the DM-
68
PTC and in the control-PTC groups, and the distribution of NS mutations in the four genes analyzed.
69
Twenty-three DM-PTCs (0.59) showed at least one NS mutation in the target gene regions analyzed. Nine
70
DM-PTCs (0.23) showed only one mutation, while 14 DM-PTCs (0.36) presented more than one mutation,
71
up to 13 different mutations (Figure 1C, continuous red line). In the control-PTCs group, 7 out of 20 PTCs
72
showed at least one NS mutation (0.35), of which 6 PTCs showed only one mutation (0.30), and one PTC
73
harbored two mutations (0.05) (Figure 1C, dashed blue line). The average number of NS mutations per
74
sample was 1.77 ± 2.59 in DM-PTCs vs. 0.40 ± 0.60 in control-PTCs (P=0.02).
75
The complete list of the detected mutations is reported in Figure 1D.
76
The percentage of mutated reads of NS mutations ranged from 5% to 69.7% in the DM-PTCs, and
77
from 22.9% to 47.9% in control-PTCs. The average mutated allele percentage was significantly lower in
78
DM-PTCs than in the control group (12.32 ± 10.39% vs. 35.29 ± 11.37%; P
