| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Departments of Medicine (S.K.G.G., B.M., I.D.H.), Biochemistry (R.B.J., N.L.E.) and Molecular Biology, Laboratory Medicine and Pathology (J.R.G., R.B.J.), and Surgery (C.S.G.), Mayo Clinic and Mayo Foundation, Rochester, Minnesota 55905; the Department of Medicine, University of California-San Diego (P.S.-C.W.), La Jolla, California 92093; the Department of Internal Medicine, School of Medicine de Ribeirao Preto (L.M.Z.M.), 14049 Ribeirao Preto Sp, Brazil; and the Division of Medical Oncology, University of Colorado Cancer Center (H.A.D.), Denver, Colorado 80262
Address all correspondence and requests for reprints to: Norman L. Eberhardt, Ph.D., 4–407 Alfred, Saint Mary’s Hospital, Mayo Clinic, Rochester, Minnesota 55905. E-mail: eberhardt{at}mayo.edu
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Previous cytogenetic and molecular genetic studies of follicular thyroid cancer (FTC) have indicated high rates of loss of heterozygosity (LOH) on the short arm of chromosome 3 (4) and the long arm of chromosome 10 (5), implicating potential tumor suppressor genes in these regions. However, these losses have not been mapped precisely, and only a small number of other chromosomes have been studied. Furthermore, the role of known tumor suppressors in the deleted regions has not been assessed, nor has the presence of LOH on specific chromosomal regions been correlated with the biological behavior of the tumor or with clinical outcome.
We examined 14 FTC (10 Hurthle cell carcinomas and 4 nonoxyphilic FTC), 14 papillary thyroid carcinomas (PTC), and 7 follicular adenomas (FA) for LOH on several chromosomes and assessed the presence of mutations in the known tumor suppressor genes VHL and p53, which reside in regions of high LOH rate in FTC. We also correlated LOH data with clinical data to determine whether LOH at any of the locations examined predicts outcome in thyroid cancer.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Tumor and peripheral blood specimens were collected from 35 patients undergoing surgery for primary (n = 25) or recurrent (n = 10) follicular cell-derived thyroid tumors. These patients included all patients with FTC (n = 14) seen by one of the investigators (I.D.H.) between January 1987 and June 1991. The remaining patients (PTC and FA) were selected at random from the patients seen by the same investigator during this period. Follow-up data were obtained for all patients in June 1996. Tumor type, size, extent of spread, histological grade, and DNA ploidy as well as patient age and sex were also recorded. These studies were approved by the Mayo Clinic institutional review board, and all patients gave informed consent before tissue and blood samples were collected.
Methods
Tumor specimens were snap-frozen in liquid nitrogen at the time of surgery and stored at -70 C. All were reviewed by a single pathologist (J.R.G.) to confirm the diagnosis and grade the tumor, according to previously published criteria (6). Frozen sections of the cryopreserved tumors were examined to ensure a minimum neoplastic cell content of 70%. Paraffin-embedded tumor material was stained with propidium iodide, and DNA ploidy was assessed by fluorescence-activated flow cytometry (FACScan, Becton Dickinson, Mountain View, CA) (7). DNA was extracted from blood and homogenized tumor tissue by ionic detergent lysis, proteinase K digestion, repeated phenol-chloroform extraction, and ethanol precipitation. Extracted DNA was dissolved in TE buffer (10 mmol/L Tris-HCl, pH 7.4, and 1 mmol/L ethylenediamine tetraacetate) and stored at 4 C.
LOH analysis was performed by both restriction fragment length
polymorphism (RFLP)- and PCR-based microsatellite polymorphic markers
(Table 1
). For RFLP analysis, eight
paired restriction digests (BglII, TaqI,
BamHI, MspI, HindIII, RsaI,
EcoRI, and PvuII; New England Biolabs, Beverley,
MA) were made of each tumor and matching normal DNA sample. Digests
were separated by agarose gel electrophoresis, blotted onto nylon
membranes, and probed with specific probe (25 ng) labeled with 50 µCi
[
-32P]deoxy (d)-ATP (6000 Ci/mmol; Amersham Corp.,
Arlington Heights, IL) using a commercial kit (Multiprime Labelling
Kit, Amersham Corp.). For microsatellite markers, DNA from tumor and
blood samples were amplified in separate paired optimized PCR
reactions, using the appropriate primer set. Reactions contained
1.5–3.5 mmol/L Mg2+, 50 µmol/L dNTPs, 2.5 U
Taq polymerase (Ampli-Taq, Perkin-Elmer, Norwalk, CT), 1
mmol/L each of forward and reverse primers, 10 mmol/L Tris-HCl (pH
8.3), 50 mmol/L KCl, 0.001% gelatin (wt/vol), and 2 µCi
[
-32P]dATP (3000 Ci/mmol; Amersham Corp.). PCR
products were separated by 6% PAGE.
|
Screening for p53 mutations was performed by PCR amplification of the genomic locus containing exons 5–9, using specific primers and the PCR conditions described above. These exons contain at least 90% of all known p53 mutations (8). Individual exons were then amplified from the product of this first round PCR. VHL mutation screening involved PCR amplification of the transcribed exons (no. 1–3) of the VHL gene. All reactions involving VHL exon 1 included 10% dimethylsulfoxide, because of the high GC content of this exon. Amplified exons were screened for mutations by dideoxy fingerprinting (DDF; Cyclist Exo-Pfu, Stratagene, La Jolla, CA) (9). In this procedure, separate forward and reverse strand PCR reactions, employing nested primer sets and including dideoxy-CTP, generate sets of dideoxy-terminated single-stranded DNA fragments. These fragments were then separated by low voltage nondenaturing 6% PAGE. Differences in the fragment migration patterns between tumor and normal DNA samples indicated possible mutations. All potential mutants were sequenced by end-labeled primer dideoxy cycle sequencing in both forward and reverse directions and resolved on 8% denaturing PAGE. Sequencing was performed using the same primers as those used for DDF, except for VHL exon 1, for which a total of 5 forward and 3 reverse primers were used to allow reliable sequencing of this GC-rich region. The sequences of primers employed in the p53 and VHL mutational screening are available from the authors on request. All 35 tumors were screened for p53 mutations; all 14 FTC and the 1 FA and 4 PTC showing LOH on 3p were screened for VHL mutations.
Statistical analysis was performed using 
2 and Fisher
exact tests for all comparisons. Kaplan-Meier and Cox regression
analysis were used to analyze survival curves. For all statistical
comparisons, P < 0.05 was considered significant.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Patient demographic and tumor pathology data are summarized in
Table 2. Ten of the 14 FTC and 1 of the
14 PTC were of oxyphilic cell type, reflecting the nature of our
tertiary endocrine practice. Oxyphilic FTC may be more aggressive than
the more common nonoxyphilic variant (10), and our data may be biased
toward more aggressive tumors. The histological grades and pathological
TNM (tumor, nodes, metastases) stages of FTC and PTC were similar; most
were grade 1 and were either stage I or II. Significantly more of the
FTC than either FA or PTC were DNA aneuploid (P <
0.05).
|
Probes were chosen on chromosome arms 1p, 3p, 3q, 10p, 10q, 11p,
11q, 13q, 17p, and 17q and are summarized in Table 1
. Chromosomes 1 and
13 have shown low rates of LOH in follicular thyroid tumors in previous
studies (4, 11, 12) and were chosen to assess the background rate of
LOH in these tumor specimens. One probe on chromosome 13 (p68RS2.0) was
chosen to map to the retinoblastoma tumor suppressor gene
(RB1). Probes on chromosomes 10, 11, and 17 were chosen to
include the region of the ret protooncogene (10q11.2) (13),
the multiple endocrine neoplasia type I locus (11q13) (14), and the
p53 gene (17p13) (15) as well as at least one other region
from each of these chromosomes. Chromosome 3, previously shown by us to
exhibit high rates of LOH in FTC (4), was more intensively mapped
(Table 1 and Fig. 1), using a total of 21
probes covering the renal cell carcinoma breakpoint (3p14.2) (16);
ACY1 (3p21), a site of frequent LOH in small cell lung
cancer (17); and distal regions of 3p close to the VHL gene
(3p25) (18) and possibly other tumor suppressor genes (19).
|
, categorized by chromosome arm. The
overall rate of LOH was higher in FTC than in either FA or PTC
(P < 0.03) and involved at least one chromosome in
86% of FTC, 56% of FA, and 43% of PTC. Multiple (>2) chromosomal
locations were involved in 72% of FTC with LOH, but only 1 (7%) of
the PTC (P < 0.001) and none of the FA
(P < 0.005).
|
In FA, the only chromosome exhibiting LOH in more than a single tumor was 10q (28%). This observation is consistent with previous suggestions that this may be an early site of allelic loss in the development of follicular neoplasia (5). Of the PTC, eight showed no LOH at any site examined, three (21%) showed LOH at a single site, two (14%) showed LOH at two sites, and only one (the oxyphilic variant) showed loss at more than two sites.
In both PTC (3 cases) and FTC (10 cases) with LOH at 17p, loss was
restricted to the most telomeric loci (mapping to 17p13.3 and
17p13.3–13.1), with no LOH detected at more centromeric markers on
this chromosome arm. Similarly, LOH on 3p appeared to cluster in
the telomeric region (3p21-ter), although a high rate of LOH (73%) was
also observed at D3S1300, mapping to 3p14.2 in the region of the
fragile chromosome 3 breakpoint (FRA3B) (16). The pattern of chromosome
3 LOH is shown in Fig. 1 for the 12 FTC, 4 PTC, and 1 FA exhibiting LOH
at 1 or more 3p loci.
We identified six regions of significant LOH on chromosome 3
(Fig. 1). The five significant regions of LOH in FTC involved D3S86
(3p24.3–23), the THRB locus (3p24), ACY1 (3p21.2–21.1), D3S1300
(3p14.2), and D3S30 (3p13-cen), respectively (Fig. 1). The most
telomeric region corresponding to D3S1038 (3p26–25) represents the
only site of significant loss in PTC, but was lost in only two of
the FTC.
Mutation screening of p53 and VHL
DDF of all 35 tumors detected 3 possible mutations of
p53 (2 in FTC and 1 in FA), 2 of which (1 FTC and 1 FA) were
confirmed by sequencing. Figure 2 shows
an example of a p53 point mutation detected by altered DDF
banding and confirmed by direct sequencing. This mutation in an FTC
involved exon 7, the most common site for p53 mutations (8),
resulting in a C to T point mutation at base 14,069, which would change
codon 248 from Arg to Trp. This mutation is homozygous, as no evidence
of the wild-type allele was observed in the sequencing gels. In
contrast, the second p53 mutation, which occurred in a FA,
was heterozygous and involved a C to G mutation at base 13,400, leading
to a codon 214 change from His to Asp.
|
Clinical outcomes
Follow-up data were available for all 35 patients and ranged from 0.17–32.75 yr (median, 6.4 yr) after primary surgery. Twenty-six of the 35 patients were followed for at least 5 yr. During the follow-up period, 7 patients died of thyroid cancer, 3 died from other causes, 4 were alive with residual disease, and 21 were alive and free of disease at the time of the last follow-up.
The median survival in PTC (24.5 yr) was not statistically different
from that in FTC (11.1 yr; Fig. 3A).
Tumor grade was significantly correlated with outcome (Fig. 3B)
independent of tumor type, with grades 2 or 3 carrying a relative risk
of cause-specific mortality of 6.3 (95% confidence interval,
1.5–25.9) compared to grade 1 tumors (P < 0.009).
Similarly, patients with DNA aneuploid tumors were at higher risk than
those with diploid tumors; no patient with a DNA diploid tumor died,
whereas 7 of the 16 patients with DNA aneuploid tumors died (44%),
with a median life expectancy in these patients of 11.1 yr
(P < 0.03).
|
). LOH on chromosomes 3p and 10q was not significantly
correlated with cause-specific mortality. However, all seven
patients who died of thyroid cancer (five with FTC and two with PTC)
exhibited LOH on chromosome 17p (median survival of 11.1 yr for
patients with 17p LOH), whereas all the patients without 17p LOH
survived their thyroid cancer (P < 0.005; Fig. 3D).
Five of the seven patients who died showed LOH on 3q, which also was
correlated with survival. A single patient with LOH on 3q, but no LOH
on 17p, remains alive, whereas two patients with LOH on 17p but no 3q
LOH have died, suggesting that LOH on 17p, rather than on 3q, is the
better predictor of outcome. |
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
All four PTC with 3p LOH showed loss at D3S1038, which maps to 3p26–25. Interestingly, two receptor tyrosine kinase genes (RET and NTRK1) have been reported to be inverted or translocated in up to 50% of PTC (23, 24), and one such NTRK1 translocation gene (TRK-T3) has been mapped recently to chromosome 3 (20).
In FTC, 3p LOH was seen most frequently in the regions 3p24–25, 3p21.1, 3p14.2, and 3p13-cen. Other than 3p14.2, these regions of minimal loss remain large and require further fine mapping before candidate tumor suppressor genes will be identifiable. In contrast, the 3p14 band shows only low to moderate LOH, with the exception of a single marker (D3S1300) that maps close to FRA3B (16) and exhibits 73% LOH in FTC. This narrow region of loss suggests possible involvement of the fragile histidine triad gene (FHIT), a recently described putative tumor suppressor in lung and other tumor types (21).
Regions 3p21 and 3p24–25 have also previously been implicated as sites of possible tumor suppressor genes in lung carcinoma (17), renal cell cancer (22), squamous head and neck cancer (23), and pancreatic islet cell tumors (19), although the gene(s) responsible has yet to be identified. The Von Hippel Lindau tumor suppressor (VHL) has been localized to 3p25 (24). However, we did not detect any VHL mutations in either FTC or PTC tumors with LOH on 3p, and all VHL exons were amplifiable in these tumors, suggesting an absence of homozygous deletions. Although hypermeth-ylation accounts for up to 20% of VHL inactivation in renal cell carcinoma (25), inactivation of this gene in other tumors results most commonly from a combination of gene deletion and mutation (23, 24). It, therefore, seems unlikely on the basis of our data that VHL plays a significant role in the pathogenesis of FTC or PTC.
Similarly, p53 does not appear to be involved in differentiated thyroid cancer. We found p53 mutations in only 2 of 35 tumors, with only a single homozygous inactivation, as shown by an absence of the wild-type allele on the sequencing gel, and LOH on 17p. This overall low rate (6%) of p53 mutations confirms previous studies of differentiated thyroid cancer (26) and contrasts strikingly with findings in other solid neoplasms, in which p53 mutation rates of around 40% are commonly reported (27).
Nevertheless, LOH at 17p13, affecting 13 of our patients (10 FTC and 3 PTC) was the strongest univariate predictor of cause-specific mortality. A similar finding has been reported in breast cancer (28), whereas LOH on 17p13, distal to the p53 gene, has also been shown in several other solid neoplasms, suggesting the presence of other tumor suppressors in this region (29, 30, 31, 32). The predictive power of 17p13 LOH in our tumors did not derive from its association with follicular histology, which was not itself predictive of outcome in this small group of patients, nor was it merely a reflection of overall LOH, which was more commonly seen on 3p, where loss did not correlate with outcome.
However, the number of patients in this study remains insufficient to confirm 17p LOH as an independent risk factor in multivariate analysis. In addition, several of our tumor samples were obtained from locally recurrent, rather than primary, disease, biasing our data toward more aggressive tumors. This is also suggested by the high proportion of oxyphilic tumors in our study group and the high observed overall mortality rate (25% at a median follow-up of 6.4 yr). These findings are consistent with a tertiary care institution selection bias and may limit the generalizability of our findings. Further studies with larger numbers of patients are clearly indicated.
The differences in prognostic significance of LOH on 3p and 17p suggest that these events are separated on the tumorigenic pathway, with LOH on 3p preceding that on 17p. Furthermore, LOH on 3p is rare in FA, but very common in FTC, consistent with a role for this region in progression from FA to FTC. In contrast, LOH on 10q is seen in both FA and FTC and may represent an even earlier step in tumorigenesis.
The high degree of DNA aneuploidy and the widespread LOH seen in FTC indicate widespread genomic instability in these tumors. However, it is unlikely that this instability is due to more advanced disease, as three of four primary FTCs exhibited 3p LOH. The presence of distinct regions with higher than average rates of LOH on chromosome 3p suggests such regions have a local propensity to LOH (e.g. at a fragile site) or contain a tumor suppressor that contributes to tumor growth. Moreover, the correlation of 17p LOH with mortality is not explained by more widespread LOH, as LOH occurs more frequently on 3p, but is not correlated with mortality.
Our data suggest that late in the course of cancer development, tumors that have accumulated multiple genetic alterations may suffer loss of an as yet unidentified tumor suppressor gene on chromosome 17p13, enhancing their malignant behavior and hastening the patient’s death. FTC, with its greater predisposition to suffer genomic abnormalities, may be more likely to reach this stage, but loss of the same tumor suppressor may also occur in some PTC and may carry with it similar prognostic implications in this tumor type. These data provide further support for the concept that the pathogenesis of thyroid cancer is a complex multistep process that involves several tumor suppressors and/or oncogenes (33).
|
Footnotes |
|---|
Received May 7, 1997.
Revised July 9, 1997.
Accepted July 15, 1997.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  B. Joseph, M. Ji, D. Liu, P. Hou, and M. Xing Lack of Mutations in the Thyroid Hormone Receptor (TR) {alpha} and Genes but Frequent Hypermethylation of the TR Gene in Differentiated Thyroid Tumors J. Clin. Endocrinol. Metab., December 1, 2007; 92(12): 4766 - 4770. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Sunde, K. C. Y. McGrath, L. Young, J. M. Matthews, E. L. Chua, J. P. Mackay, and A. K. Death TC-1 Is a Novel Tumorigenic and Natively Disordered Protein Associated with Thyroid Cancer Cancer Res., April 15, 2004; 64(8): 2766 - 2773. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. A. Aldred, M. E. Ginn-Pease, C. D. Morrison, A. P. Popkie, O. Gimm, C. Hoang-Vu, U. Krause, H. Dralle, S. M. Jhiang, C. Plass, et al. Caveolin-1 and Caveolin-2,Together with Three Bone Morphogenetic Protein-related Genes, May Encode Novel Tumor Suppressors Down-Regulated in Sporadic Follicular Thyroid Carcinogenesis Cancer Res., June 1, 2003; 63(11): 2864 - 2871. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. N. Nikiforova, R. A. Lynch, P. W. Biddinger, E. K. Alexander, G. W. Dorn II, G. Tallini, T. G. Kroll, and Y. E. Nikiforov RAS Point Mutations and PAX8-PPAR{gamma} Rearrangement in Thyroid Tumors: Evidence for Distinct Molecular Pathways in Thyroid Follicular Carcinoma J. Clin. Endocrinol. Metab., May 1, 2003; 88(5): 2318 - 2326. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. A. French, E. K. Alexander, E. S. Cibas, V. Nose, J. Laguette, W. Faquin, J. Garber, F. Moore Jr, J. A. Fletcher, P. R. Larsen, et al. Genetic and Biological Subgroups of Low-Stage Follicular Thyroid Cancer Am. J. Pathol., April 1, 2003; 162(4): 1053 - 1060. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Bodmer, W. van den Hurk, J. J. M. van Groningen, M. J. Eleveld, G. J. M. Martens, M. A. J. Weterman, and A. Geurts van Kessel Understanding familial and non-familial renal cell cancer Hum. Mol. Genet., October 1, 2002; 11(20): 2489 - 2498. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Farrand, B. Delahunt, X.-l. Wang, B. McIver, I. D. Hay, J. R. Goellner, N. L. Eberhardt, and S. K. G. Grebe High Resolution Loss of Heterozygosity Mapping of 17p13 in Thyroid Cancer: Hurthle Cell Carcinomas Exhibit a Small 411-Kilobase Common Region of Allelic Imbalance, Probably Containing a Novel Tumor Suppressor Gene J. Clin. Endocrinol. Metab., October 1, 2002; 87(10): 4715 - 4721. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Loss of Heterozygosity of the Long Arm of Chromosome 7 in Follicular and Anaplastic Thyroid Cancer, but Not in Papillary Thyroid Cancer J. Clin. Endocrinol. Metab., September 1, 1999; 84(9): 3235 - 3240. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Lesueur, M. Stark, T. Tocco, H. Ayadi, M. J. Delisle, D. E. Goldgar, M. Schlumberger, G. Romeo, and F. Canzian Genetic Heterogeneity in Familial Nonmedullary Thyroid Carcinoma: Exclusion of Linkage to RET, MNG1, and TCO in 56 Families J. Clin. Endocrinol. Metab., June 1, 1999; 84(6): 2157 - 2162. [Abstract] [Full Text]
|
|
|
|
|
|
S. Hemmer, V.-M. Wasenius, S. Knuutila, K. Franssila, and H. Joensuu DNA Copy Number Changes in Thyroid Carcinoma Am. J. Pathol., May 1, 1999; 154(5): 1539 - 1547. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
Trp) at
nucleotide position 14,069 (arrow).