The Heterogeneous Distribution of BRAF Mutation Supports the Independent Clonal Origin of Distinct Tumor Foci in Multifocal Papillary Thyroid Carcinoma

Abstract

Context: Papillary thyroid carcinoma (PTC) is frequently multifocal. Independent PTC foci may occur either from intraglandular metastases from a single dominant tumor or as unrelated neoplastic clones. In rare cases, the simultaneous presence of PTC foci of different histopathological subtypes points to independent sites of tumor formation.

Objectives: We examined the pattern of BRAF mutations in noncontiguous tumor foci and node metastases from 69 patients affected by multicentric PTC. These included 19 cases characterized by the simultaneous presence of different PTC histopathological variants.

Design: BRAF (exon 15) mutation was examined by PCR-single strand conformational polymorphism followed by DNA sequencing in laser-capture microdissected tissue samples.

Results: Discordant patterns of BRAF mutation were found in about 40% of the multifocal PTCs. In node metastases, BRAF mutations were, in most but not all the cases, concordant with the dominant tumor. A discordant pattern of BRAF mutation was also found in about 50% of the cases in which multiple foci of different histopathological variants were present.

Conclusions: The heterogeneous distribution of BRAF mutations suggests that discrete tumor foci in multifocal PTC may occur as independent tumors. This information has to be considered in the design of targeted therapeutic approaches with BRAF pathway inhibitors.

PAPILLARY THYROID CARCINOMA (PTC) is the most common cancer of the thyroid gland, comprising approximately 90% of all cases (1–3). PTC is characterized by a set of distinctive nuclear features (4, 5). Classical variant PTC (CVPTC) is characterized by a branching architecture and psammoma bodies. Numerous PTC variants are recognized, such as follicular (FVPTC), solid, tall cell (TCVPTC), oncocytic, Warthin-like, diffuse sclerosing, cribriform, and others (4, 5). Quite often (18–87% of the cases), PTCs are multicentric with multiple noncontiguous tumor foci in individual glands (4, 6–8). These cases are in general characterized by a “dominant” tumor and multiple additional smaller foci, often microcarcinomas (<1 cm).

Large tumor size, old age, extrathyroidal invasion, aggressive histological variants, and distant metastases are the most important determinants of a poor outcome in PTC patients (2–4, 7, 9–11). In addition, multifocality has been associated with a poor outcome, increased risks of metastases, and regional recurrence (4, 6–8). In fact, intraglandular dissemination from the dominant tumor may indicate a tumor with metastatic potential and, therefore, a more aggressive phenotype. This should not apply in those cases in which multicentric PTCs occur instead as independent carcinomas (8, 12). By determining the presence of a polymorphism in the X-linked androgen-receptor gene, Shattuck et al. (8) found that in five of the 10 informative female patients, multicentric PTCs were, indeed, independent tumors. Rearrangements of the RET (RET/PTC) receptor are present in a fraction of PTCs (13, 14). Using RET/PTC as a marker, Sugg et al. (15) reported that some multifocal PTCs had different profiles of RET/PTC, suggesting an independent origin of the different tumor foci.

Somatic point mutations in the BRAF gene have been identified as the most common genetic event in PTC (approximately 44% of PTC cases) (13, 14, 16–18). BRAF codes for a serine/threonine kinase that functions in the RAS/MAPK cascade. The T1799A transversion mutation of the BRAF gene, which causes a V600E amino acid substitution in the BRAF protein [the V600E mutation (T1799A)] represents more than 90% of BRAF mutations in PTC. This mutation affects the conformation of the activation loop in the BRAF kinase domain and potentates by more than 500-fold its catalytic activity. Other more rare BRAF mutations have been found. In particular, the K601E (A1801G) mutation has been described in FVPTC cases (16, 18).

Here we show that about 40% of the laser-capture microdissected multifocal PTCs had discordant patterns of BRAF mutation. These results confirm data recently reported by Park et al. (19) who found heterogeneous distribution of BRAF mutation in 39.3% of multifocal PTCs. Importantly, in some cases the BRAF status was discordant in node metastases with respect to the dominant tumor. Moreover, a discordant pattern of BRAF mutation, in some cases with the presence of different mutations was also found in about 50% of the 19 cases in which multiple histopathological variants were dissected. These findings support the notion that a fraction of multifocal PTCs may be independent carcinomas, sustained by different genetic events.

Patients and Methods

Thyroid samples

All the patients were treated at the Institute of Endocrinology and underwent thyroidectomy at the Department of Surgery of the University of Pisa. The initial treatment was total thyroidectomy with central neck and/or laterocervical lymph-node dissection. A series of patients affected by multicentric and metastatic PTC were chronologically selected and collected for the study. Tumor samples were obtained in accordance with protocols approved by the institutional review board, and the informative consent was achieved 1 d before surgery together with the surgical one. As reported in Table 1, 18 CVPTCs, 17 FVPTCs, and 15 TCVPTCs were collected. All the patients showed a dominant larger tumor, and additional smaller tumor foci and concomitant node metastases. No differences were observed regarding average size on PTC variants both in dominant, additional foci as well as in node metastases.

TABLE 1.

Clinicopathological features of the examined multifocal and metastatic PTCs

PTC variant	No. of cases			Mean age ± sd (yr)	Mean size ± sd (cm)
	Total	Male	Female		Dominant tumor	Additional foci	Node metastases
CVPTC	18	5	13	38.7 ± 12.9	1.7 ± 0.9	0.5 ± 0.6	1.2 ± 0.8
FVPTC	17	9	8	41.7 ± 12.9	2.1 ± 1.3	0.6 ± 0.6	1.9 ± 1.1
TCVPTC	15	2	13	39.9 ± 13.9	1.8 ± 1.2	0.5 ± 0.6	1.5 ± 1.9
Total	50	16	34	40.1 ± 13.6	1.88 ± 1.16	0.59 ± 0.62	1.53 ± 1.35

Of the 50 patients with multifocal PTC, nine had multiple clinically evident carcinomas (>1 cm), 27 had one clinically evident carcinoma and a single microcarcinoma or multiple microcarcinomas, and 14 had multiple microcarcinomas. A total of 36 patients had one separate PTC foci, five had two, six had three, and three patients had more than four foci. There were 19 additional patients who showed multiple PTC foci belonging to different histopathological subtypes (Table 2). Histological diagnosis was made independently, in a blinded fashion, by two pathologists (C.U. and F.B.). Tumors were classified according to the histopathological typing of the World Health Organization (20). A concordance rate of 98% was obtained between the two investigators. The few discordant cases were discussed with a third pathologist.

TABLE 2.

BRAF mutations in multicentric cases with different histological variants

Patient no.	Categoriesa	BRAF status
1	FV/TCVPTC	WT/WT
2	FV/TCVPTC	WT/V600E
3	FV/TCVPTC	V600E/WT
4	FV/TCVPTC	K601E/V600E
5	TCV/FVPTC	WT/WT
6	TCV/FVPTC	WT/V600E
7	TCV/FVPTC	V600E/WT
8	TCV/FVPTC	V600E/WT
9	FV/CVPTC	WT/WT
10	FV/CVPTC	WT/V600E
11	FV/CVPTC	WT/V600E
12	FV/CVPTC	V600E/V600E
13	CV/FVPTC	WT/WT
14	CV/FVPTC	V600E/WT
15	PDC/FVPTC	V600E/K601E
16	CV/TCVPTC	V600E/WT
17	CV/TCVPTC	V600E/V600E
18	CV/TCVPTC	V600E/V600E
19	TCVPTC/PDC	V600E/K601E

WT, Wild type.

^a

The morphological type of the dominant (larger) tumor focus precedes that of the additional (smaller) one.

Microdissection and DNA extraction

Serial 5-μm sections were taken from paraffin blocks for histological examination on glass slides and for DNA extraction on membrane slides (Nikon, Firenze, Italy). Presence of the tumor tissue was confirmed in the first and last section for each section series. Unstained sections were deparaffinized with Bio-Clear (Bio-Optica, Milano, Italy), rehydrated in graded ethanol, and stained with hematoxylin and eosin. Microdissections were performed using the laser-assisted SL μcut Microtest (MMI GmbH distributed by Nikon). For each sample, three to five microareas of 5 μm were obtained. Each area contained 200–500 cells. Particular care was taken in microdissecting areas of the dominant tumor, additional foci, and node metastases. Samples from the nonneoplastic thyroid parenchyma, generally from the contralateral lobe, were dissected as control reference. The microdissected cells were placed in the SL μcut transfer film (Nikon), and the DNA was extracted overnight in a humidified chamber at 56 C in 200 μl tissue lysis buffer (ATL DNeasy Tissue kit; QIAGEN GmbH, Hilden, Germany) containing 20-μl proteinase K. DNA was isolated by QIAGEN spin column; carrier tRNA was added to improve DNA recovery. Finally, DNA was eluted in 40 μl Tris EDTA buffer and immediately processed for PCR amplification. A mock control in which no tissue was added was processed in parallel with each sample.

Detection of BRAF mutation by PCR-single strand conformational polymorphism (SSCP) and DNA sequencing

PCR-SSCP screening of BRAF mutations was performed by amplifying exon 15 according to a standard procedure (21). Two microliters of DNA were used as a template in a 20-μl PCR mixture containing 10 mm Tris-HCl, 50 mm KCl, 1.5 mm MgCl₂ (pH 8.3), 0.2 mm deoxynucleotide triphosphate, 8 pmol amplimers, and 1 U AmpliTaq DNA Polymerase (Perkin-Elmer Cetus Instruments, Foster City, CA). PCR primers for the BRAF exon 15 were as follows: 5′(F)-tcctttacttactacacctcagat-3′ and 5′(R)-agtggaaaaatagcctcaat-3′. The amplicon size was 167 bp. Cycling conditions were as follows: initial denaturation (94 C, 5 min), then 35 cycles (denaturation, 94 C for 40 sec; annealing, 55 C for 40 sec; and synthesis, 72 C for 40 sec), followed by a final extension of 5 min. All PCR products were visualized by electrophoresis in 2% agarose gel and purified using PCR purification kit (QIAGEN, Crawley, West Sussex, UK). Purified products were then diluted 1:1 with denaturing solution (1% xylene cyanol, 1% bromophenol blue, 0.1 mm EDTA, and 99% formamide), boiled for 5 min, and immediately placed in ice to prevent the annealing of single strand products. SSCP screening was performed on the GenePhor Electrophoresis Unit using GeneGel Excel 12.5/24 (12.5% T, 2% C), according to manufacturer’s instructions (Pharmacia Biotech, Inc., Piscataway, NJ). Electrophoresis (600 V, 25 mA, 15 W) was performed at 18 C for 100 min. Gels were stained with PlusOne Silver Staining Kit (Amersham Biosciences, Freiberg, Germany), according to manufacturer’s instructions. Altered migration patterns in two or three independent PCR-SSCP runs were indicative of DNA mutations. Purified PCR products were then sequenced by an ALF II automated sequencer (Amersham Biosciences) using the Thermo Sequenase Cy5 Dye Terminator Cycle Sequencing Kit (Amersham Biosciences). DNA sequences were compared with those of the normal BRAF gene exon 15 using the Basic Alignment Search Tool (BLAST) software available at the National Center for Biotechnology Information (Bethesda, MD). As control, two human thyroid cancer cell lines, ARO and TPC, heterozygous and negative for the BRAF mutation, respectively, were used.

Statistical analysis

Data were analyzed using the Mann-Whitney U, χ², and bivariate tests. A P value < 0.05 denoted the presence of a significant difference.

Results

BRAF mutation in the dominant tumors

The dominant (larger) tumors of the 50 patients affected by multifocal PTC were analyzed for the presence of exon 15 BRAF mutation. The V600E mutation was found in 31 of 50 (62%) dominant tumors. In particular, BRAF was mutated in 10 of 18 (55.5%) CVPTCs, nine of 17 (52.9%) FVPTCs, and 12 of 15 (80%) TCVPTCs (Table 3). The percentage of V600E BRAF mutation in our series of FVPTC is surprisingly high in comparison with the literature data (16). We believe that this phenomenon could be due to the selection of the patients because all FVPTCs analyzed are multifocal and metastatic, suggesting an intrinsic aggressive behavior of these cases.

TABLE 3.

BRAF (V600E) distribution according to the histological variant of the dominant tumor

Variant	Total	BRAF V600E positive cases, n (%)
CVPTC	18	10 (55.5)
FVPTC	17	9 (52.9)
TCVPTC	15	12 (80)
Total	50	31 (62)

Heterogeneous distribution of BRAF mutation among the dominant tumor, additional foci, and metastatic nodes

In addition to the dominant tumor, we analyzed BRAF in all histologically recognized neoplastic foci (ranging from one to five) and in all metastatic nodes (ranging from one to five). Representative examples of the microdissection and BRAF mutation detection method are shown in Figs. 1 and 2, respectively. Multiple microdissected samples (3–5) were examined for each individual lesion. The obtained results are summarized in Table 4. Seven different conditions were verified; four of them (1–4 in Fig. 3) were quite common (44 of 50 cases), while the last three (5–7 in Fig. 3) were more rare (six cases).

TABLE 4.

BRAF V600E positivity in the dominant tumor, additional foci, and corresponding lymph node metastases

BRAF V600E positive cases
Sample	No. of CVPTCs	%	No. of FVPTCs	%	No. of TCVPTCs	%	Total	%
P	4	22.2	1	5.9	5	33.3	10	20
F	0	0.0	1	5.9	0	0.0	1	2
N	2	11.1	1	5.9	0	0.0	3	6
P+F	0	0.0	2	11.8	0	0.0	2	4
P+N	0	0.0	5	29.4	2	13.3	7	14
F+N	0	0.0	0	0.0	0	0.0	0	0
P+F+N	6	33.3	1	5.9	5	33.3	12	24
NEG	6	33.3	6	35.3	3	20.0	15	30
Total	18		17		15		50

Presence of BRAF mutation in the dominant tumor (P), tumor foci (F), node metastases (N), in dominant and in tumor foci, simultaneously (P+F), dominant and node metastases, simultaneously (P+N), in tumor foci and in node metastases, simultaneously (F+N), and in the dominant tumor, additional foci and metastasis (P+F+N). Absence of BRAF mutation in the dominant tumor, additional foci, and metastasis (NEG).

Fig. 1.

Laser-assisted microdissection of multifocal PTC. Left, Hematoxylin and eosin-stained sections of the dominant tumor (A1), smaller focus (A2), and node metastasis (A3). Middle, Dissection of tumor cells from the dominant tumor (B1), smaller focus (B2), and node metastasis (B3). Right, Areas lacking tumor cells after the dissection. Squared areas in A series represent approximately the same region of the B series.

Fig. 2.

a, SSCP run of five PTC dominant samples (P), additional focus (F), metastasis (M), and normal control tissue (N). Altered migration is indicated (*). Positive (+) and negative (−) BRAF V600E mutation controls are shown. b, Representative sequence analysis of one BRAF V600E mutant. Arrows indicate the T1799A transversion, in forward (upper) and reverse (lower) strand. L, 100-bp ladder.

Fig. 3.

Schematic representation of the variable distribution of the BRAF V600E mutation in the 50 patients affected by multifocal PTC. Shaded shapes indicate presence of BRAF mutation, and clear shapes indicate wild-type (WT) BRAF. Number: 1, absence of BRAF mutation in the dominant tumor, additional foci and metastasis; 2, presence of BRAF mutation in the dominant tumor, additional foci and metastasis; 3, presence of BRAF mutation only in the dominant tumor; 4, presence of BRAF mutation in the dominant tumor and the corresponding metastasis; 5, presence of the BRAF mutation exclusively in the node metastasis; 6, absence of the BRAF mutation in the node metastasis; and 7, presence of the BRAF mutation exclusively in the additional focus.

Absence of BRAF mutation in the dominant tumor, additional foci, and metastasis

There were 15 cases negative for BRAF mutation in all the examined samples (dominant tumor, additional foci, and node metastases) (example 1, in Fig. 3). Multiple additional foci were examined in three of these patients, and all scored negative. Two independent metastatic nodes were examined in 13 of these patients, and all scored negative. Thus, these patients were not informative, and no conclusion could be drawn relative to the origin of their multifocal PTC.

Presence of BRAF mutation in the dominant tumor, additional foci, and metastasis

There were 12 cases that resulted homogeneously positive for BRAF V600E in the dominant tumor, additional foci, and corresponding node metastasis (example 2, in Fig. 3). Multiple additional foci, all BRAF V600E positive, were examined in four of these patients. Two independent metastatic nodes were examined in eight of these patients, and all scored positive. Such a homogeneous presence of the BRAF mutation is compatible with either a shared or independent clonal origin. In the last case, independent mutational events should have occurred in distinct neoplastic clones.

Presence of BRAF mutation only in the dominant tumor

There were 10 cases that showed the BRAF V600E mutation in the dominant tumor, but neither in the additional foci nor in the node metastasis (example 3, in Fig. 3). In two of these patients, three distinct additional foci were all negative for the mutation. Two independent metastatic nodes were examined in seven of these patients, and all scored negative. These findings are consistent with independent sites of cancer formation in the glands of these patients. If the (remote) possibility that the mutated BRAF allele is lost during the metastatic process is excluded, it can be inferred that in these patients, the node metastases could not derive from the dominant tumor.

Presence of BRAF mutation in the dominant tumor and corresponding metastasis

Seven cases were positive for the mutation in both the dominant tumor and metastatic node (two independent node metastases in six of them), but not in the additional smaller focus (example 4, in Fig. 3). In one of these cases, multiple independent additional foci were all negative. These findings are also consistent with independent sites of cancer formation. The common positivity for BRAF of the dominant tumor and the node metastases is compatible with either a shared or independent clonal origin.

Presence of the BRAF mutation exclusively in the node metastasis

In three patients while the node metastases (two independent nodes in two of them) were BRAF V600E positive, both the dominant tumor and smaller focus were negative (example 5, in Fig. 3). A similar condition was already reported (22). These findings are consistent with the de novo formation of BRAF mutation during the neoplastic progression rather than tumor initiation.

Absence of the BRAF mutation in the node metastasis

In two patients the BRAF mutation was present in the dominant tumor, in one additional focus, but neither in the second additional focus nor the node metastasis (example 6, in Fig. 3). One of these cases had multiple node metastases, all negative for BRAF mutation. Similar to condition no. 3 (Fig. 3), this rare condition is consistent with metastases not deriving from the dominant tumor.

Presence of the BRAF mutation exclusively in the additional focus

In one single case, the mutation was confined to the smaller focus (example 7, in Fig. 3). In this case the dominant tumor was a FVPTC, while the focus was a nonotherwise specified PTC. Three node metastases were all negative.

BRAF mutation in multifocal PTC of different histopathological variants

The coexistence of distinct PTC foci of different histopathological subtypes points to independent sites of tumor formation. We could examine 19 of these patients (a total of 38 carcinomas) (Table 2), namely: 15 FVPTCs, nine CVPTCs, 12 TCVPTCs, and two poorly differentiated carcinomas (PDCs). Altogether, three of 15 FVPTCs showed the V600E and two of 15 the K601E mutation; eight of nine CVPTCs showed the V600E, and seven of 12 TCVPTCs harbored the V600E mutation. The two examined PDCs were both mutated (one V600E and one K601E). Regarding the concordant or discordant nature of the BRAF genetic status among these patients, four were not informative because negative for the mutation. In the remaining 15 cases, quite rarely (three cases), we found the same mutation in the two tumors. More frequently, one of the two tumors was positive and the other negative (nine cases). More interestingly, there were three patients in which two different BRAF mutations (V600E and K601E) were found in the two tumor components (Table 2). In one case, the V600E mutation was associated to the TCVPTC, while the K601E mutation to the FVPTC; in a second case, the K601E was associated to the FVPTC and the V600E to the PDC. In a third case, the V600E was associated to the TCVPTC, while the K601E to the PDC.

Discussion

Our analysis of the BRAF mutational status supports the possibility that individual tumors in multifocal PTC could occur as independent tumors in 20 of 50 cases (40%). In fact, as reported in Fig. 3 (category nos. 3, 4, 6, and 7), BRAF mutations were present in one or more lesions (dominant, foci, or lymph node metastases), but not in all, suggesting that at least one of these lesions could occur independently from the others. Consistent with the histopathological appearance, an even larger fraction [30 of 50 (60%)] of genetically heterogeneous cases was found among multifocal PTCs in which individual tumor foci belong to different morphological subtypes. Differently, BRAF mutation in single tumors of mixed architecture was relatively homogeneous (23).

Overall, our data confirm the results reported by Shattuck et al. (8) and Park et al. (19). By analyzing X chromosome inactivation or, as in our study, the BRAF mutation rate, these authors reported discordant genetic patterns in 50% and 39.3% of the multifocal PTC, respectively. However, the possibility of intrathyroidal metastasis remains in those cases in which multiple foci had concordant (either wild-type or mutated) BRAF status. This possibility was supported by the study of McCarthy et al. (24), which in contrast to the other cited and also our own study, reported a high concordance rate of X chromosome inactivation and loss of heterozygosity among different foci of multifocal PTC. Although the reasons for this discrepancy remain unknown at present, such an intraglandular dissemination may occur through the abundant network of lymphatic vessels present in the thyroid (19). The distinction between independent tumors and intraglandular dissemination of the same tumor might be important in terms of biology of the disease; in fact, the latter condition may anticipate an aggressive behavior with metastatic ability.

That a large fraction of multifocal PTCs is the result of the parallel development of independent tumors might be not surprising, considering that the tumor foci additional to the dominant tumor were often microcarcinomas. Because microcarcinomas are overall highly prevalent in the general population, it is possible that they are found only by chance in a thyroid gland otherwise affected by a full-blown PTC. Also not surprising is the finding that the BRAF status in the node metastases in most of the cases (35 of 50) matched that of the dominant tumor. In three patients we found that the BRAF mutation was present in the metastasis, but not in the gland, suggesting that it had occurred during the metastatic process.

More puzzling was the observation that in 12 cases, the node metastasis was concordant with the smaller thyroid focus rather than the dominant one. In all these cases, the metastasis was negative for the BRAF mutation. It is unlikely that the mutated allele was lost during the metastatic process. Thus, in these cases, the possibility has to be considered that the metastasis occurred from the small additional focus rather than the dominant tumor.

The finding that multiple PTC foci may occur independently has implications for the pathogenesis and suggests that thyrocytes in some individual patients may have a significant propensity to undergo neoplastic transformation. On the other hand, this genetic heterogeneity should be considered when the BRAF mutation is exploited for therapeutic purposes. Indeed, BRAF or its downstream MAPK kinase is in fact being explored as a suitable molecular target for the development of novel compounds for cancer treatment (17, 25–27). The genetic analysis of the selected molecular target is a necessary prerequisite for the correct design of targeted therapeutic approaches. Thus, patient genotyping may be impaired by the heterogeneous presence of BRAF mutations, and this may render difficult the stratification of patients to be directed to the proper clinical trial as well as the analysis of the results of the trial.

Acknowledgments

This study was supported by the Associazione Italiana per la Ricerca sul Cancro, Italy, and by Grant COFIN 2003 no. 2003069778 of the Italian Minister of University and Scientific Research.

Disclosure Statement: The authors have declared no conflict of interest.

Abbreviations

 
• CVPTC,

  Conventional variant PTC;

 
• FVPTC,

  follicular variant PTC;

 
• PDC,

  poorly differentiated carcinoma;

 
• PTC,

  papillary thyroid carcinoma;

 
• SSCP,

  single strand conformational polymorphism;

 
• TCVPTC,

  tall cell variant PTC.