Different Clonal Origin of Bilateral Papillary Thyroid Carcinoma, with a Review of the Literature

Abstract

Papillary thyroid carcinoma (PTC) often presents with two or more anatomically separate foci. A long-standing argument is whether this multifocality is the result of multiple independent tumors (“multicentricity”) or of intrathyroidal spread originating from a single tumor mass, presumably through permeation of intrathyroidal lymph vessels. We reexamined this issue with a clonality assay and compared our results with those in the literature. A total of 27 nodules from 11 female patients with bilateral PTC treated with total thyroidectomy were investigated for clonality using the HUMARA assay. Eight of 11 cases were informative (72.7%). All but one of tumor foci showed a monoclonal population. The outlier sample gave a value indicative of balanced X-inactivation in one nodule. The monoclonality was concordant in three patients, discordant in three, and mixed in two (with both concordant and discordant results). Interestingly, in both of the latter cases (composed of over two samples per case), the contralateral nodules were discordant. Moreover, all four ipsilateral nodules were concordant. The results of our study suggest that some cases of multifocal PTC are the result of true multicentricity, whereas others are the consequence of intrathyroid spread by an originally single tumor mass. These conclusions support those made in the past years on the basis of morphologic considerations. Specifically, the incidental finding of two or more microscopic foci of PTC widely separate from each other was felt to favor multicentricity, whereas the finding of multiple ipsilateral foci of PTC within vascular spaces, often accompanied by multiple lymph node metastases, suggested intrathyroid spread; the most striking manifestation of this phenomenon being seen in the diffuse sclerosing variant of PTC.

Introduction

Papillary thyroid carcinoma (PTC) is the most common type of thyroid carcinoma, accounting for over 85% of all malignant thyroid tumors. A common feature of this neoplasm is the presence of two or more anatomically separate foci; the incidence of this phenomenon ranges from 18 to 87% in the literature [1–6]. This multifocality has been associated with an increased risk of nodal and distant metastases, persistence of local disease after initial treatment, and regional recurrence [7]. A long-standing argument regarding multifocal PTC is whether this occurrence is the result of multiple independent tumors (“multicentricity”) or of intrathyroidal spread originating from a single tumor mass, presumably through permeation of intrathyroidal lymph vessels. The molecular genetic studies that have been carried out to answer this question have given rise to contradictory results. We therefore thought of reexamining the issue by determining the clonality of the lesions in microdissected tumor samples with the use of the HUMARA assay. We selected cases with bilateral disease in order to leave little doubt about the fact that we were dealing with anatomically separate tumor foci.

Material and Methods

Cases

Archived, anonymized cases of bilateral PTCs from 11 female patients treated with total thyroidectomy were obtained from the files of the Laboratory of Pathology, Arcispedale Santa Maria Nuova, Reggio Emilia, Italy. All available haematoxylin–eosin slides were reviewed. The clinicopathological findings in these cases are summarized in Table 1. The patients ranged in age from 29 to 69 years, with an average of 49 years. All patients had bilateral disease, and two had known regional lymph node metastases. A total of 27 tumor nodules were studied. The size of the greater tumor nodule ranged from 0.5 to 2.5 cm. Microscopically, seven of the main tumors were of the classic type, and four belonged to the follicular variant.

Table 1.

Clinicopathological findings

Case	Age (years)	n	T1 Size (cm), side, and histotype	T2 Size (cm), side, and histotype	T3 Size (cm), side, and histotype	T4 Size (cm), side, and histotype	pN	Nonneoplastic parenchyma
1	42	2	1.5, LL, PTC	0.6, RL, PTC			N0	AH
2	43	3	0.5, RL, FVPC	0.4, RL, PTC	0.4, LL, PTC		N0	AH
								HT
3	29	2	2.0, LL, PTC	0.4, RL, FVPC			N1	Npa
4	62	2	2.3, LL, PTC	0.3, RL, FVPC			N0	Npa
5	53	3	0.6, LL, PTC	0.5, RL, TALL	0.1, LL, PTC		N0	HT
6	45	3	1.0, RL, PTC	0.5, RL, PTC	0.4, LL, PTC		N0	HT
7	56	2	1.3, RL, PTC	0.5, LL, FVPC			N0	AH
8	69	2	1.5, LL, FVPC	0.1, RL, PTC			N0	AH
9	54	2	1.0, RL, PTC	0.4, LL, PTC			N0	AH
10	56	4	0.9, LL, FVPC	0.5, RL, FVPC	0.4, RL, PTC	0.3, LL, PTC	N0	HT
11	33	2	2.5, RL, FVPC	0.1, LL, FVPC			N2	HT

n number of tumor nodules, LL left lobe, RL right lobe, pN pathological lymph nodes stage, AH adenomatous hyperplasia, HT Hashimoto thyroiditis, Npa no pathologic alterations, PTC papillary carcinoma, classic, FVPC papillary carcinoma, follicular variant

Microdissection and DNA Extraction

Two to five 5 μm-thick sections were obtained from formalin-fixed paraffin-embedded blocks and stained with hematoxylin and eosin. Selected tumor cells were procured from each case by manual microdissection with a 30-gauge needle under microscopic visualization. Cells from nonneoplastic thyroid parenchyma were also microdissected and used as normal control. The procured tissue was resuspended in a lysis buffer containing 0.5 mg/ml proteinase K, 0.5% Tween-20, 1 mM EDTA pH 8.0, and 50 mM Tris–HC1, pH 8.5, and incubated at 60°C for 48 h. The samples were then heated at 95°C for 5 min, inactivate proteinase K and stored at –20°C until use.

Clonality Analysis by HUMARA Assay

The HUMARA assay takes advantage of restriction sites for two different restriction enzymes (in our method, Rsal and Hhal), which are located in the first exon less than 100 bp 5′ to a highly polymorphic trinucleotide repeat (CAG) of the human androgen receptor gene (HUMARA), and which has an informativeness of >90%, allowing differentiation between the paternal and the maternal alleles. HUMARA also exhibits a consistent pattern of methylation (the inactive allele is methylated, whereas the active is unmethylated) that correlates well with the X chromosome inactivation pattern. Briefly, the DNA is digested by a methylation-sensitive endonuclease (Hhal) that cuts all the unmethylated (active) alleles from different cells near the CAG repeat region. Therefore, only the inactive (methylated) alleles provide an intact template for PCR amplification when primers that span a region containing the polymorphic CAG repeat and the Hhal restriction site are used.

We adapted the assay of Busque et al. [8] to investigate the X chromosome inactivation pattern of HUMARA. All samples were analyzed in duplicate. In each case, 6 μl of DNA was initially digested at 37°C for 6–8 h with Rsal (40 U) and the methylation-sensitive restriction enzyme Hhal (40 U). Each sample was also processed in parallel without enzymes for comparison. To prevent the restriction buffer from affecting subsequent PCR amplification, the samples were then phenol–chloroform-extracted, ethanolprecipitated, and dissolved in 8 μl of water.

Three microliters of restricted DNA was amplified in a final reaction volume of 10 μl of buffer containing 10 mM Tris–HCl pH 8.3, 50 mM KCl, 1.5 mM MgCl₂, 0.001% (v/w) gelatin, 200 μM each deoxyribonucleotide triphosphate (dNTP), 5% DMSO, 5 pmol of each primer, 0.1 μl [α⁻³²P] dCTP (6,000 Ci/mmol, New England Nuclear, Boston, MA), and 0.5 U AmpliTaq Gold DNA polymerase (Perkin Elmer, Foster City, CA). We included the 7-deaza-2–dGTP in the dNTD mix instead of the dGTP to improve resolution of the allelic signals. Primer sequences were 5′-TCCAGAATCTGTTCCAGAGCGTGC-3′ (sense, labeled with 5′-hexachlorofluorescein phosphoaramidite) and 5′-GCTGTGAAGGTTGCTGTTCCTCAT-3′ (antisense) as described [9]. PCR reaction were carried out in a Perkin Elmer Thermal Cycler 480 with the following conditions modified from Mutter et al. [10]: initial denaturation at 95°C for 9 min, 3 cycles at 95°C for 2 min, 55°C for 45 s, 72°C for 90 s, followed by 24–28 cycles at 95 for 45 s, 55°C for 45 s, and 72°C for 90 s. All PCRs were finally elongated at 72°C for 7 min. We size-separated the PCR products on the ABI PRISM® 310 Genetic Analyzer and analyzed the results using GeneScan 3.1.2 software (Applied Biosystems), with the allele peak height serving as a semiquantitative measure of the amount of PCR product amplified from each allele.

Polyclonal tissues with random X-inactivation patterns would be expected to have allelic cleavage ratios (ACRs) equal or close to 1.0. We took into account the possibility of preferential amplification of one allele, which occurs if the alleles differ markedly in the length of their CAG repeats or if the DNA contains, by chance, a deletion in the X chromosome that includes the amplified region. In order to correct for this potential preferential amplification, we calculated a corrected ACR for each HhaI-digested and undigested negative control sample pair. The corrected ACR was derived by dividing the ratio of undigested and digested allele 1 by the ratio of undigested and digested allele 2. We considered any value between 0.7 and 1.5 of this corrected ratio as indicative of a balanced methylation and therefore of a polyclonal population, and any value less than 0.7 or above 1.5 as indicative of an unbalanced methylation and therefore of a monoclonal cell population.

Results

The results obtained are summarized in Table 2. A total of 27 tumor nodules were investigated: two in seven cases, three in three cases, and four in one case (in addition to control normal parenchyma in every case). The DNA obtained was of poor quality in two cases, which were therefore discarded. One case was homozygous for the HUMARA allele and was considered non-informative. This resulted in a total of eight informative cases (72.7%), on which the conclusions of this paper are based. In all of these cases, the control samples of normal thyroid tissue showed an ACR in the range of 0.7–1.3, indicating random inactivation and therefore a polyclonal nature. By contrast, all but one of the informative tumor samples gave an ACR outside the normal range, indicative of a monoclonal population. The outlier sample gave a value indicative of balanced inactivation (0.9).

Table 2.

X chromosome inactivation analysis of multifocal papillary thyroid carcinomas

Case	Parenchyma ACR	Tumor	Side	Methylated allele	ACR
1	1.3	T1	LL	Upper	2.3
		T2	RL	Lower	0.4
2	1.0	T1	RL	Lower	0.2
		T2	RL	Lower	0.3
		T3	LL	Upper	2.8
3	1.0	T1	LL	Upper	3.9
		T2	RL	Upper	>4.0
4	NA	T1	LL	NA	NA
		T2	RL
5	0.8	T1	LL	Lower	0.2
		T2	RL	Lower	0.2
		T3	LL	Lower	<0.2
6	1.0	T1	RL	Upper	4.6
		T2	RL	Upper	2.6
		T3	LL	Upper	>4.0
7	NA	T1	RL	NA	NA
		T2	LL
8	1.2	T1	LL	Upper	3.2
		T2	RL	Lower	<0.1
9	Homozygous	T1	RL	NI	NI
		T2	LL
10	0.7	T1	LL	Lower	0.5
		T2	RL	Upper	1.6
		T3	RL	Upper	2.5
		T4	LL	Balanced	0.9
11	0.9	T1	RL	Lower	0.6
		T2	LL	Upper	2.2

Two HUMARA alleles—a lower allele and an upper allele—are distinguishable in every tumor nodule

ACR corrected allelic cleavage ratio, LL left lobe, RL right lobe, NA not available, NI not informative

Among the cases in which the tumor samples gave results indicative of monoclonality, three were concordant (cases 3, 5, and 6), three were discordant (cases 1, 8, and 11), and two showed both concordant and discordant results (cases 2 and 10) (Figs. 1 and 2). Interestingly, in the latter two cases (composed of over two samples per case), the contralateral nodules were discordant. Moreover, all four ipsilateral nodules were concordant.

Fig. 1.

Classic papillary carcinoma, left lobe. High power (a). Follicular variant of papillary carcinoma, right lobe (b). A series of four electropherograms of HUMARA analysis in case 3. In the upper tracings (left lobe), two major peaks are shown representing the two human androgen receptor alleles. The numbers underneath the peaks indicate the peak height and fragment size. Below (right lobe), the same alleles are shown after incubation with the restriction enzyme HhaI. Because HhaI digests only unmethylated DNA fragments, the remaining peaks represent methylated alleles. A series of two ratios in which the proportion of alleles in the lower tracing is compared with the proportion of alleles in the upper tracing will determine the clonality of the sample. In this case, the lower allele is unmethylated, and the upper allele is methylated (c). Both tumors have the same inactivated (methylated) X chromosome suggesting a common clonal origin

Fig. 2.

Left lobe showing a papillary carcinoma with follicular predominance. Note the red colloid, nuclear grooves, nuclear clearing, and some nuclear pseudoinclusions (a). Classic papillary carcinoma in the right lobe (b). The HUMARA analysis in this case (case 8) showed that in the left lobe nodule, the upper allele was methylated and the lower allele was unmethylated (c), whereas in the right lobe nodule, only the lower allele appeared methylated (d). We concluded that these two tumors belonged to different clones since they had a different pattern of X chromosome methylation

Discussion

The results of our study suggest that the multiple tumor nodules present in some cases of PTC are the result of true multicentricity, whereas in others, they are the consequence of intrathyroid spread by an originally single tumor mass. Interestingly, all ipsilateral PTC foci showed the same X chromosome inactivation pattern, supporting a common clonal origin, whereas five out of eight contralateral nodules showed a different X-inactivation.

There are several reported studies looking at this issue with various molecular genetic techniques [11–15]. The figures obtained varied substantially from study to study, but it is significant that all but one of the authors concluded—as we did—that the multiplicity of thyroid carcinoma could be the result of either true multicentricity or of intrathyroidal spread (Table 3). In fact, Moniz et al. [13], Shattuck et al. [11], and Park et al. [14] observed a nearly equal number of cases with either the same or different molecular characteristics in concurrent PTC nodules, concordant with their common or independent clonal origin, respectively. Moreover, both McCarthy et al. [15] and Wang et al. [16] found molecular homogeneity among the nodules in most cases, consistent with a common clonal origin of the coexisting PTCs, i.e., intrathyroidal spread. In contrast to such findings, Sugg et al. [12] concluded that in most cases, the nodules were different at the molecular genetic level and therefore presumably independent.

Table 3.

Literature review summary

Authors (year)	Probe	Results	Authors’ conclusion
Sugg et al. (1998) [12]	RET/PTC rearrangement	Different rearrangement in 15/17 cases Same rearrangement in 2/17 cases	Most multiple PTCs arise independently
Moniz et al. (2002) [13]	X chromosome inactivation (HUMARA assay)	Concordant inactivation in 5/8 cases Discordant inactivation in 3/8 cases	Multiple PTCs may arise from intrathyroid spread or independently, with a slight predominance of the former
Shattuck et al. (2005) [11]	X chromosome inactivation (HUMARA assay)	Concordant inactivation in 5/10 cases Discordant inactivation in 5/10 cases	Multiple PTCs may arise from intrathyroid spread or independently
Park et al. (2006) [14]	BRAFV600E mutation	24/61 of the cases showed mutation in some of the nodules but not in others	Multiple PTCs may arise from intrathyroid spread or independently, with a slight predominance of the latter
McCarthy et al. (2006) [15]	1. LOH for 3 microsatellite polymorphic markers 2. X chromosome inactivation (HUMARA assay)	1. Concordant allelic loss pattern in 20/23 cases 2. Concordant X chromosome inactivation in 13/13 informative cases	Nearly all multiple PTCs arise from the same clone and are the result of intrathyroid spread
Wang et al. (2010) [16]	1. BRAFV600E mutation 2. X chromosome inactivation (HUMARA assay)	1. Concordant BRAF status in 18/21 cases 2. Concordant X chromosome inactivation in 9/11 informative cases	Nearly all multiple PTCs arise from the same clone and are the result of intrathyroid spread

Some of these differences may be attributable to the tumor population studied. In particular, Sugg et al. [12] limited their study to cases of multiple papillary microcarcinoma (i.e., cases in which none of the tumor nodules measured over 1 cm in diameter), a subtype of PTC in which an independent origin would seem much more likely on pathologic grounds. Park et al. [14] studied only Korean patients, an ethnic group in which BRAF mutations are particularly common. However, it is likely that most of the differences encountered are the result of the different techniques used and their inherent shortcomings. The RET rearrangement analysis chosen by Sugg [12] and the BRAF-activating mutations searched by Park et al. [14] seem logical probe choices due to the proven role of these two genes in thyroid carcinogenesis. However, they may not represent the initiating mutational event, but rather, they occur later during tumor progression—a possibility supported by the report of several different types of RET/PTC transcripts and BRAF^V600E mutations within a single tumor nodule, or when comparing a primary PTC with its metastasis.

The clonality assay of the HUMARA gene, employed by Moniz et al. [13], Shattuck et al. [11], McCarthy et al. [15], and by us, is based in the “lyonization” of the X chromosome [17]. In each female cell, one of the two X chromosomes is randomly inactivated early in embryogenesis by methylation of deoxycytosine residues. Once established, this inactivation is transmitted to the progeny so that the adult female theoretically has inactivated the paternal X chromosome in half of the cells and the maternal X chromosome in the other half. A monoclonal population of cells is that in which all cells contain the same inactivated X chromosome (either paternal or maternal), whereas a polyclonal population is made up of cells with similar numbers of inactive paternal and maternal X chromosomes. Since inactivation of the X chromosome is independent of neoplastic selection and occurs before cell transformation, determination of X chromosome inactivation can accurately determine whether a group of tumor cells originates from a single precursor (monoclonal) or from multiple precursors (polyclonal). Separate nodules composed of monoclonal cells with different inactivated X chromosome represent strong evidence that they have arisen as separate events from different progenitor cells. The finding of the same inactivated X chromosome in different nodules is less definitive. Such a finding is certainly consistent with a shared clonal origin of the nodule, but it is also possible that it happened by chance alone. Naturally, this becomes less of an issue as a larger number of nodules from the same case are studied. To wit, if two nodules from the same case share the same pattern of X chromosome inactivation, the chances of this being due to chance alone are 50 %, whereas if five nodules from the same case share the same pattern, the odds of this having happened by chance alone are 3.1%. In practical terms, a different X-inactivation pattern among two cell populations means two different clones, but the reverse is not necessarily true, unless one studies a large number of nodules (which few of the authors cited did).