| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Articles |
Department of Molecular Biology, Institute of Gerontology, Nippon Medical School (Y.K., M.E.), 1-396 Kosugi-cho, Nakahara-ku, Kawasaki 211-8533, Japan; Department of Surgery II, Nippon Medical School (Y.K., K.S., S.T.), 1-1-5 Sendagi, Bunkyo-ku, Tokyo 113-8602, Japan; and Ito Hospital (K.I.), 3-4-6 Jingumae, Shibuya-ku, Tokyo 150-8308, Japan
Address all correspondence and requests for reprints to M. Emi, M.D., Ph.D., Department of Molecular Biology, Institute of Gerontology, Nippon Medical School, 1-396 Kosugi-cho, Nakahara-ku, Kawasaki 211-8533, Japan. E-mail: memi{at}nms.ac.jp
Abstract
The genetic mechanisms involved in development of follicular thyroid carcinoma are poorly understood, although allelic losses (LOH) in this type of tumor have been reported in small panels of follicular thyroid carcinomas examined in earlier studies. To clarify the real frequency of allelic loss we carried out a genome-wide allelotyping study of 66 follicular thyroid carcinomas using 39 microsatellite markers representing all nonacrocentric autosomal arms. The mean frequency of LOH was 9.2%, and the mean fractional allelic loss was 0.09. The most frequent allelic losses were detected in 7q (28%), 11p (28%), and 22q (41%). When we compared these results with our previous allelotyping studies using identical markers in other types of thyroid cancers, the 9.2% mean frequency of allelic loss detected in follicular thyroid carcinomas was higher than that in papillary thyroid carcinomas (3%), but not as high as that detected in anaplastic thyroid carcinomas (20%). Frequent allelic losses of markers on chromosomes 7q, 11p, and 22q suggest locations to examine for the presence of suppressor genes associated with the development of follicular thyroid carcinoma.
THYROID CARCINOMAS are categorized as medullary, papillary, follicular, or anaplastic; the last three types originate from follicular cells of the thyroid gland, and medullary carcinomas derive from parafollicular C cells. Papillary carcinoma is the most common histological type of thyroid cancer, the second being follicular carcinoma. Papillary thyroid carcinoma is easily diagnosed clinically by means of fine needle aspiration cytology. However, in most cases of follicular thyroid neoplasia it is impossible to make a diagnosis preoperatively unless distant metastases are present; capsular and vascular invasion by follicular thyroid carcinoma cells, the only conclusive evidence for differentiation of follicular carcinoma from adenoma, is usually diagnosed only by histological examination after surgery (1). Clinicopathological, cytological, histochemical, and imaging approaches to the diagnosis of this disease have been attempted, but accurate preoperative diagnosis will be possible only when genetic features that characterize follicular carcinoma cells have been defined.
Among the genetic alterations known to occur among the major
histological types of thyroid carcinoma are germline or somatic
missense mutations of the RET protooncogene in medullary
carcinomas (2, 3, 4, 5), rearrangements of the RET
gene in papillary carcinomas (6, 7), and mutations of p53
in anaplastic carcinomas (8). Few specific genetic
alterations associated with follicular thyroid carcinoma have been
reported, except for some chromosomal losses (9, 10) and
PAX8-PPAR 
1 fusion gene (11).
Recently, Kroll et al. (11) reported t
(2, 3) (q13; p25), a translocation resulting in fusion of
the DNA-binding domains of the thyroid transcription factor
PAX8 to domains A to F of the peroxisome
proliferator-activated receptor 1 in five of eight follicular
thyroid carcinomas but other histological types of thyroid neoplasms.
Specific allelic losses (LOH) in the cells of a given type of tumor are
considered to signal the positions of critical tumor suppressor genes
associated with that cancer (12). In general, multiple
tumor suppressor genes must be inactivated before cancer will develop.
Ward et al. (9) summarized previous studies on
allelic losses in follicular thyroid carcinomas and concluded that
although allelic loss in more than 25% of cases had been reported in
multiple chromosomes (2, 3, 4p, 8, 9, 11p, 15q, 17p, and 22q), it
remained unclear whether those data were significant, because only a
limited number of DNA markers had been examined in a small number of
tumors (at most 28 tumors) in any documented study.
To clarify the importance of chromosomal losses in follicular thyroid carcinomas, it is obvious that a large number of cases must be examined. Therefore, we carried out a genome-wide LOH analysis (all 39 acrocentric chromosome arms) in 66 follicular thyroid carcinomas, using the same polymorphic microsatellite markers as those we used previously for allelotyping anaplastic and papillary thyroid carcinomas (13, 14). This approach has enabled us to make a rational comparison of allelotypes among different histological types of thyroid carcinoma.
Subjects and Methods
Patients and tumors
We examined tumors from 66 consecutive patients with follicular
thyroid carcinoma who were treated at Ito Hospital (Tokyo, Japan).
Histological diagnosis was made according to the criteria of the
Japanese Society of Thyroid Surgery (1). Ten oxyphilic
cell carcinomas and 20 widely invasive carcinomas were included in this
panel of tumors. Informed consent in the formal style of the hospital
was obtained from each patient before surgery, and their clinical data
are given in Table 1
. Eleven patients
were men, and 55 were women; the mean ages at initial treatment were
46.8 (range, 21–83) yr. Forty patients underwent tumor resection or
hemithyroidectomy, and 26 underwent subtotal or total thyroidectomy.
Anterior and paratracheal lymph nodes (central node) were dissected in
12 patients, and modified neck dissections were carried out in 25
patients. Fifty-six patients had a careful follow-up; the median
follow-up time was 115.5 months (range, 3–159). Fifteen patients had
tumor recurrence, including 4 patients who died of thyroid carcinoma;
they had lymph node metastases in the neck, recurrences in the remnant
thyroid, or distant metastases. The remaining 41 patients did not have
tumor recurrence.
|
Carcinoma samples and corresponding normal tissues were obtained from formalin-fixed, paraffin-embedded tissue blocks prepared at the time of surgery. From each block, 10–15 slides of 10-µm tissue sections were prepared. After a careful comparison between hemotoxylin-eosin-stained slides and the corresponding tissue sections, tumor and nontumor tissues were isolated with a fine scalpel under microscopic inspection to avoid contamination by normal thyroid tissue. Dissected tumors were judged to contain at least 60% of tumor cells by microscopic observation. Tumor and nontumor DNAs were extracted using DEXPAD (Takara, Tokyo, Japan) according to the manufacturer’s instructions and purified by phenol-chloroform extraction.
LOH analysis
Allelotyping was carried out by PCR-based LOH analyses using 39
highly polymorphic microsatellite markers, i.e. 1 marker
from each acrocentric autosomal arm (Table 2). We selected the markers that would
represent either a locus for a known tumor suppressor gene or a region
where frequent allelic losses had been described in some human tumors
(15). We chose markers whose allele sizes were less than
140 bp for amplification by PCR, as genomic DNAs from our panel of
follicular thyroid carcinomas had been extracted from archival
paraffin-embedded tissues. Our selection criteria also included markers
whose reported heterozygosity exceeded 0.8, to obtain maximum
information from each LOH analysis.
|
-32P]ATP end-labeled forward primer,
2.5 pmol reverse primer, and 0.25 U Taq polymerase. Cycle
conditions were 94 C for 3 min, then 30 cycles of 94 C for 30 sec, an
appropriate annealing temperature (55–64 C) for 30 sec, and 72 C for
30 sec, with a final extension of 72 C for 3 min, in a GeneAmp PCR 9600
System (Perkin-Elmer Corp./Cetus, Norwalk, CT). PCR
products were resolved by electrophoresis on 6% polyacrylamide
denaturing gels (36% formamide and 8 M urea) at
2000 V for 2–3 h. Size-separated alleles were visualized by
autoradiography. Definition of LOH
LOH was assessed by visual inspection of autoradiographs and by measuring signal intensities of alleles using a GS-370 densitometry scanning system (Hoefer Scientific, San Francisco, CA). LOH was scored if we found a reduction in signal intensity of more than 50% from one allele of tumor DNA compared with that in the corresponding allele from nontumor DNA of the same patient.
Results
Representative autoradiographs are shown in Fig. 1. Figure 2
displays schematically the LOH frequencies that are indicated in Table 2
at each locus. Informativeness of the markers ranged from 46–89%,
with an average of 75%. LOH on individual arms was documented in up to
41% of informative cases; the mean percentage was 9.2%, with an
SD of 8.2; 55 of the 66 tumors (83%) revealed LOH on at
least 1 chromosome arm, with the highest frequency (16 of 39, 41%) at
D22S284, a marker mapped at 22q13.1–13.2. Allelic losses
were also frequent on chromosome arms 7q (D7S2431 at
7q21–22; 10 of 36, 28%) and 11p (D11S922 at 11p15; 12 of
43, 28%).
|
|
Discussion
According to some reports follicular thyroid carcinomas exhibit chromosomal instability, as frequent allelic losses have been detected in multiple chromosome arms (9). The largest previous study was reported by Tung et al. (10), who analyzed 28 follicular thyroid carcinomas for LOH with microsatellite markers. In that study the mean percent LOH of each chromosome arm was 20.4, and the mean FAL was 0.21. We carried out genome-wide LOH analyses in a larger number (n = 66) of follicular thyroid carcinomas, and recorded much lower average frequencies of allelic losses: 9.2 as the mean percent LOH and 0.09 as the mean FAL. The cases that Tung’s group studied included an unusually high proportion of oxyphilic cell carcinomas (39%) and widely invasive tumors (50%) compared with the numbers one would expect to find in an ordinary panel of follicular thyroid carcinomas (17). Those pathological findings suggested that the cases in the Tung panel were skewed in favor of more aggressive cancers. In fact, Segev et al. (18) reported that oxyphilic cell carcinomas show much higher frequencies of allelic loss than nonoxyphilic tumors. In general, oxyphilic cell cancers and widely invasive subtypes account, respectively, for 10–20% and 30–40% of follicular thyroid carcinomas (17, 19, 20, 21, 22). The reason behind the higher frequencies of allelic loss reported by others than those found in the present study appears to reflect selective biases toward specific subtypes of follicular thyroid carcinoma that entail poorer prognosis. The small number of cases examined in former studies also complicates accurate evaluation.
In the study reported here we documented frequent allelic losses in chromosome arms 7q, 11p, and 22q, suggesting that these regions are likely to harbor genes associated with tumorigenesis of follicular thyroid carcinoma. Hemmer et al. (23) reported frequent chromosomal gains in 1q and 17q as well as frequent losses in 1p, 13q, and 22q among 20 follicular thyroid carcinomas after comparative genomic hybridization analysis. As chromosomal amplifications were not always distinguishable from chromosomal losses by the PCR-based LOH analysis used in the present study, it is possible that the chromosomal imbalances we have described as LOH might sometimes reflect aneuploidy or allelic amplification in addition to allelic loss.
Losses at chromosome bands 7q21 and 7q31 have been reported in myeloid neoplasms, ovarian cancers, uterine leiomyomas, and head and neck squamous cell carcinomas (24, 25, 26, 27). Zhang et al. (28) and Travato et al. (29) also reported that loss of chromosome 7q was associated with the development of follicular thyroid carcinoma. In our previous study that compared allelotypes between patients who survived papillary thyroid carcinoma postoperatively and those who died, frequent loss at 7q appeared to be associated with postoperative death (14). However, we did not find 7q loss to be associated with postoperative recurrence of the 56 follicular thyroid carcinomas that had careful follow-up. With regard to the other two areas of frequent allelic loss in our panel of follicular thyroid carcinomas, 11p harbors at least two known tumor suppressor loci, the WT1 gene (11p13) and a locus at 11p15 associated with breast cancer (30). Tung et al. (10) reported loss of the WT1 locus in 33% of their cases of follicular thyroid carcinoma, and we had also recorded frequent loss of 11p in anaplastic thyroid carcinomas (13). In addition, we detected loss of chromosome 22q in up to 41% of our panel of follicular thyroid carcinomas. Chromosome 22q carries the putative tumor suppressor genes NF2 (22q12) and SNF5/INI1 (22q11.2) (31). Frequent loss of 22q has been reported in several other studies of follicular thyroid carcinoma; Tung et al. (10) found 22q loss in 30% of cases by PCR-based microsatellite analysis, and Hemmer et al. (23) detected it in 40% by comparative genomic hybridization.
Previously we had carried out LOH analyses in anaplastic and papillary thyroid carcinomas (13, 14) using the identical microsatellite markers as those used in the study reported here. The mean percent LOH and the mean FAL in each histological type were, respectively, 20.1% and 0.19 in the anaplastic carcinomas, 10.4% and 0.10 in the postoperative-deceased papillary carcinomas, 2.9% and 0.03 in the postoperative-surviving papillary carcinomas, and 9.2% and 0.09 in the follicular carcinomas. Prognosis is favorable for most patients with papillary thyroid carcinoma (32); fewer than 10% of all papillary thyroid carcinoma patients die of the disease (33). Therefore, the higher frequency of LOH among follicular thyroid carcinomas than in the papillary type of thyroid tumor was not unexpected, nor was the observation that allelic loss in follicular carcinomas was much less frequent than in anaplastic carcinoma.
Extrathyroidal invasion, older age, and distant metastasis at the time of diagnosis are clinico-pathological factors affecting the prognosis of individual follicular thyroid carcinomas (17, 34, 35). However, few reports have discussed genetic prognostic factors for this disease. Hemmer et al. (23) reported that 22q deletion was found in widely invasive and fatal follicular carcinomas, but not in the papillary carcinomas of their series, suggesting that 22q loss had prognostic value for follicular thyroid carcinoma only. On the basis of our results, however, it is clear that chromosome 22q loss occurs frequently in all histological types of thyroid cancers originating in follicular cells; we have documented 22q loss in 38%, 19–33%, and 41% of anaplastic, papillary, and follicular carcinomas, respectively (13, 14). Moreover, we found no association of 22q loss with postoperative prognosis in our panel of 66 follicular thyroid carcinomas. Therefore, allelic loss of chromosome 22q might play a fundamental role in the development of thyroid carcinoma as a common, early genetic event, but would not serve as a useful prognostic factor.
Recently, Kroll et al. (11) reported that the
PAX8-PPAR gene fusion via chromosomal translocation, t (2, 3) (q13; p25), might be the major etiological step in genesis of
follicular thyroid carcinomas, as it was found in five of eight
follicular thyroid carcinomas. On the other hand, several researchers
observed specific LOH at several chromosomal loci. In addition to the
frequent observations of LOH on 7q, 11p, and 22q as discussed above,
LOH at 10q and 11q were also noted by Zendius et al.
(36), Marsh et al. (37), and Nord
et al. (38). However, previous findings of LOH
in this type of tumor were based on small sample size and few markers
that only covered a limited arm of human chromosomes. In the present
study we examined the largest panel of follicular thyroid carcinomas
with a comprehensive set of genome-wide polymorphic markers
representing all of the human chromosomal arms and established the view
that inactivation of multiple tumor suppressor genes represents another
essential part of etiology for the development and progression of this
type of cancer.
Acknowledgments
Footnotes
This work was supported in part by special grants-in-aid for priority areas cancer and genome science from the Ministry of Education, Science, Sports, and Culture of Japan; a Research Grant from the Ministry of Health and Welfare of Japan; and a Research for the Future Program Grant of the Japan Society for the Promotion of Science.
Abbreviations: FAL, Fractional allelic loss.
Received December 28, 2000.
Accepted May 9, 2001.
References
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
