| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Clinical Cancer Genetics Program (F.W., K.A.W., C.E.), Human Cancer Genetics Program (F.W., C.P., K.A.W., C.E.), Comprehensive Cancer Center (C.E.), and Division of Human Genetics (C.E.), Department of Internal Medicine, Division of Human Cancer Genetics, Department of Molecular Virology, Immunology, and Medical Genetics (F.W., K.A.W., C.P., C.E.), Department of Pathology (C.D.M.), The Ohio State University, Columbus, Ohio 43210; Department of General Surgery and Transplantation (F.W., A.F., C.E.B.), University of Essen, D-45122 Essen, Germany; Division of Medical Genetics (M.A.A.), University of Leicester, Leicester LE1 7RH, United Kingdom; and Cancer Research U.K. Human Cancer Genetics Research Group (C.E.), University of Cambridge, Cambridge CB2 1XZ, United Kingdom
Address all correspondence and requests for reprints to: Charis Eng, M.D., Ph.D., Human Cancer Genetics Program, The Ohio State University, 420 West 12th Avenue, Suite 690 TMRF, Columbus, Ohio 43210. E-mail: eng.25{at}osu.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Total RNA from a panel of 15 FTCs and its variants, 14 follicular adenoma (FAs), seven PTCs, and five normal frozen thyroid tissues was subjected to global gene expression analysis as described previously (2). The histological classification of these neoplasias was as follows: seven classic FTCs, six oncocytic variants of FTC [Hürthle cell carcinomas (HCC)], two insular FTCs, and seven classic PTCs. Two of the six HCCs and two of the seven classic FTCs were classified as minimally invasive. No atypical variant or Hurthle cell adenoma was included in the set of 14 FAs. Furthermore, DNA from selected samples was obtained for sodium bisulfite sequencing (see below). For two of the 15 FTCs, matching normal tissue was available. In addition, we extracted DNA from paraffin-embedded archival material of an independent sample set consisting of 17 FTCs, seven FAs, and five multinodular goiter, all with matching normal tissue, for LOH analysis.
All samples were obtained as anonymized materials without linked identifiers, with the approval of The Ohio State University’s Institutional Review Board for Human Subjects’ Protection.
RNA extraction
Total RNA was isolated from 0.2 g of snap-frozen tissue using the TRIzol reagent (Invitrogen, Carlsbad, CA) and purified with the RN easy kit (Qiagen, Valencia, CA) according to the manufacturer’s recommendation. Aliquots of 1 µg total RNA were pretreated with DNAse I (Invitrogen); 500 ng were then reverse transcribed into cDNA using the SuperScript II system (Invitrogen) and a random hexamer anchored primer (Roche, Indianapolis, IN).
RT-PCR
Quantitative RT-PCR was performed using the iQ SYBR Green RT-PCR system (Bio-Rad Laboratories, Hercules, CA) on an iCycler Instrument (Bio-Rad) using the comparative threshold cycle (Ct) method (9). Equal efficiency of the reference and target amplification was determined by a validation experiment for all reference and target genes. Samples were analyzed in triplicate for the target gene and normalized to the average Ct value of the two reference genes, ß-actin and glyceraldehyde-3-phosphate dehydrogenase, the latter two of which were analyzed in duplicate. All quantitative RT-PCR products were further visualized on a 2% agarose gel to ensure a single product. 
Ct was determined by normalizing to the average Ct of five normal thyroid samples. The fold difference between FTC and FA is calculated by 2 to the power of the absolute difference in Ct between the two groups (e.g. FTC and FA). All values are given as means and SD of each group. We used the two-tailed Student’s t test for independent samples assuming equal variance to determine statistical significance between the two groups.
Semiquantitative duplex RT-PCR was performed in a thermal cycler PTC-200 machine (MJ Research Inc., Waltham, MA) for a total of 35 cycles using an annealing temperature of 55 C in a 30-µl reaction mixture containing 15 µl Multiplex master mix, 6 µl 5x Q-Solution, and 1 µl each primer (20 µM) (Invitrogen). All primer sequences are shown in Table 1
. We employed dot spot densitometry using the ChemiImager 4000 (Alpha Innotech, San Leandro, CA) to obtain signal ratio between ARHI and ß-actin. The two-tailed Fisher exact test was used to determine difference between groups.
|
To determine methylation status of the three identified CpG Islands, we first used the technique of combined bisulfite restriction analysis (data not shown). However, we found that this technique is rather ambiguous and not sufficient to facilitate a conclusive insight into the biological situation. This might be due to the limitations of the restriction enzymes used (BstUI, TaqI). These show a considerable margin of error when cutting at a single, unmethylated CG site. Furthermore, with this technique, only one CG site among numerous in each island, is evaluated. We therefore decided to employ sodium bisulfite genomic sequencing, which enables us to evaluate the complete CpG island for changes in methylation status.
Genomic DNA was extracted from snap-frozen tissue with the DNA minikit (Qiagen) according to the manufacturer’s recommendation. Methylation analysis was performed by bisulfite sequencing as described previously (10). In brief, 1 µg genomic DNA was denatured and incubated with 30 µl 10 mM hydroquinone and 520 µl 3 M sodium bisulfite for 16 h at 50 C. The samples were then desalted, desulfonated by 0.3 M NaOH, and purified with the Qiaquick gel extraction kit (Qiagen). Then 1.5 µl aliquots were amplified in a 30-µl reaction mixture containing 15 µl HotStar master mix, 6 µl 5x Q-Solution, and 1 µl each primer (20 µM) (Invitrogen). Primers used for bisulfite-treated DNA were designed as previously described (Table 1
) (5). Complete bisulfite conversion was assured by having less than 0.01% of the cytosines in non-CG dinucleotides unconverted in the final sequence. The PCR products were than cloned into the pCR2.1-TOPO vector system by using the TOPO TA cloning method, according to the manufacturer’s recommendations (Invitrogen). Five to 10 separate clones were selected for further analysis. Sequencing was performed using dGTP technology and the ABI 3730 analyzer (Applied Biosystems, Perkin-Elmer Corp., Norwalk, CT) according to the manufacturer’s recommendation. The Sequencher software package (version 4.2, GeneCodes Corp., Ann Arbor, MI) was used for sequence analysis.
LOH analysis
Genomic DNA from paraffin-embedded tumor tissue and paired normal tissue was obtained as described previously (11). Microsatellite markers for LOH analysis were selected from Genethon and Marshfield genetic maps according to their physical location and analyzed as described previously (2). These markers were D1S1603, D1S2829, and D1S198, spanning a region 1.5 cM centromeric and 1.5 cM telomeric of ARHI on 1p31.2. DNA fragments were separated by capillary electrophoresis, and the signal was detected with an semiautomated DNA sequencer 3700 (Applied Biosystems, Perkin-Elmer Corp.). The results were analyzed by automated fluorescence detection using the GeneScan collection and analysis software (Applied Biosystems, Perkin-Elmer Corp.). Scoring of LOH was performed by inspection of the GeneScan analysis output. A ratio of peak heights of alleles between normal and tumor DNA of 0.67 was employed to define LOH as described by us and others previously (11, 12, 13, 14). The two-tailed Fisher exact test was used to determine statistical difference between the two groups.
Quantitative PCR (Q-PCR)
From frozen tissue that lacked matching normal material, we extracted genomic DNA with the DNA minikit as described above (Qiagen). We performed Q-PCR using the iQ SYBR Green RT-PCR system (Bio-Rad) on an iCycler instrument (Bio-Rad) employing the comparative Ct method as described above (11, 12). Primers were designed to amplify untranslated regions of ARHI and ß-actin. All primer sequences are shown in Table 1.
Mutation analysis
PCR products were sequenced using Big Dye terminator technology (version 3.1) and the ABI 3730 analyzer (Applied Biosystems, Perkin-Elmer Corp.) according to the manufacturer’s recommendation for mutation analysis as described previously (2). PCR consisted of 38 cycles using an annealing temperature of 55 C in a 30-µl reaction mixture containing 15 µl HotStar master mix, 6 µl 5x Q-Solution (Invitrogen), and 1 µl (20 µM) each of each primer. All primer sequences are shown in Table 1.
Cell culture
The FTC133 human follicular thyroid cancer cell line was grown in a mixture of DMEM, Ham’s F12, and MCDB104 (2:1:1) with 10% fetal bovine serum (Invitrogen, Grand Island, NY). 5-AzA-dC was dissolved in dimethylsulfoxide as a 10-mM stock solution and stored in aliquots at –80 C. Trichostatin A (TSA) was dissolved in 100% ethanol at a stock concentration of 1 mg/ml (3.3 mM) and stored at –20 C. All drugs were purchased from Sigma Chemical Co. (St. Louis, MO). Cells were plated at 106 per 100-mm dish and grown for 24 h before treatment. 5-AzA-dC was added in concentrations from 1 to 10 µM for 48 h with adding fresh 5-AzA-dC daily. TSA was added after the initial 48 h of culture at concentrations from 400 nM to 1 µM for 16 h before harvesting the cells.
DNA was extracted with the DNA minikit (Qiagen) according to the manufacturer’s recommendation, and methylation analysis was performed as described above. For RNA extraction, the RNAeasy minikit (Qiagen) was used according to the manufacturer’s recommendation, and RT-PCR was performed as described above.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). Detailed analysis of the FTC samples revealed that five of six HCCs also showed strong suppression of ARHI expression, a proportion similar to that of classic FTC. Notably, ARHI was also markedly underexpressed in all four minimally invasive FTCs (two classic FTCs and two HCCs) (Fig. 1). Duplex semiquantitative RT-PCR, which measures levels of ARHI transcript normalized to the housekeeping gene ß-actin, revealed significantly reduced average ARHI expression levels when compared with that of normal thyroid tissue (P = 0.022) or FAs (P = 0.0091). In addition, we analyzed seven PTCs and found underexpression of ARHI mRNA in only two of seven samples (28.6%), a frequency similar to what is seen in FA samples but different from FTCs (P = 0.022, compared with FTC) (Fig. 1).
|
Ct –2.97 ± 2.02) and FAs (mean Ct 0.95 ± 2.4), with FAs showing transcript levels similar to normal thyroid tissue (P < 0.0001, Fig. 2). ARHI expression was similar between FTCs (Ct –2.88 ± 2.09) and HCCs (Ct –3.06 ± 2.14, P = 0.88). Furthermore, the minimally invasive FTCs had a Ct of –2.3 (± 1.43), which is not statistically different from the remaining FTCs (–3.37 ± 2.16, P = 0.41).
|
). Notably, when LOH was present, all informative markers for a particular sample detected this. In addition, we used the Q-PCR approach to evaluate allelic imbalance for samples that lacked DNA from matching normal tissue (Fig. 3). Similar to traditional LOH analysis, a threshold value of 0.67 was considered indicative of allelic loss, whereas a value above 1.5 indicates gain of copy number (11, 12). Nine of 14 FTCs (64.3%), which have been used for ARHI expression analysis by duplex RT-PCR and quantitative RT-PCR, showed allelic loss, compared with 1 of 10 FAs (10%; P = 0.013) (four FAs could not be evaluated due to lack of frozen material for DNA extraction), thus independently confirming the data obtained with a distinct sample set and LOH analysis (above). Furthermore, one of four FTCs with normal copy number also had retained ARHI expression. We were able to validate our Q-PCR approach by analyzing the only two FTC samples that also had matching normal tissue (Fig. 4). Sample 2 (in Table 2), which had LOH for all three markers, also showed allelic loss identified by Q-PCR (Fig. 3) accompanied by underexpression of ARHI (0.4, Fig. 4). In contrast, sample 1 (in Table 2), which retained heterozygosity at all three markers, showed no change in copy number by Q-PCR (Fig. 3) as well as retained ARHI expression (0.7, Fig. 4).
|
|
|
). The two benign samples showed a symmetric pattern of methylation, i.e. representing one presumptively imprinted, thus hypermethylated, allele and the other unmethylated allele (Fig. 5). In contrast, one FTC underexpressing ARHI (sample K, Figs. 1 and 3) showed only one pattern of hypermethylation over all three CpG islands, whereas the second FTC (sample H, Figs. 1 and 3), which retained ARHI transcript levels, displayed the symmetric pattern, similar to the benign lesions (Fig. 5).
|
Ct 0.65 ± 1.43), we were able to show that the FTC133 cell line had a monoallelic deletion of ARHI and therefore serves as a good in vitro model of studying ARHI epigenetics. Our data showed that inhibition of histone deacetylation with TSA alone (2.5-fold induction, compared with untreated control ± 0.06, P = 0.0007) or in combination with demethylation by 5-AzA-dC (3.4-fold ± 0.08, P = 0.0002) can reactivate or induce expression of ARHI in this cell line in a dose-dependent manner, whereas demethylation with 5-AzA-dC alone (1.12 fold ± 0.13) failed to increase ARHI expression (Fig. 6).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Our observation that relatively few FAs but the majority of FTCs, including minimally invasive FTCs, show marked ARHI mRNA underexpression suggests two things. First, ARHI down-regulation is primarily by deletion, perhaps secondarily by chromatin modulation, occurs relatively early, likely sometime before minimal invasion occurs. Second, invasive FTC derives from genetic/epigenetic evolution from FAs and minimally invasive FTCs.
Our hypothesis, that the first event is hypermethylation of the ARHI promoter (maternal imprinting) followed by physical loss of the nonmethylated allele, can be proven in two ways, structurally and functionally. Ideally, functional proof would entail demonstration of monoallelic expression of ARHI. Unfortunately, because of the absence of polymorphisms in all our samples, we cannot conclusively prove allelic distribution of the transcript and have to rely on previously published evidence (5, 6, 7). Nonetheless, maternal genomic imprinting has been proven for breast and ovarian tissue (5, 6, 7). Importantly, our bisulfite sequencing data provide structural evidence that indeed loss of the nonmethylated ARHI allele, leaving the imprinted allele, occurs. Specifically, benign thyroid lesions that retained heterozygosity and normal mRNA expression of ARHI display the expected symmetrical methylation pattern of parental imprinted genes (19), one being hypermethylated and one nonmethylated and the latter transcribed. These data are corroborated by the fact that this symmetrical pattern of methylation is shared by one FTC with retention of heterozygosity and retained ARHI gene expression, whereas the FTC with ARHI silencing and LOH displays only hypermethylation over all three CpG islands. This latter observation strongly suggests that FTC with down-regulated ARHI expression achieves this state by losing its remaining nonmethylated allele.
Another interesting, yet unelucidated, aspect of our data is that allelic imbalance with gain in copy number can occur and subsequently can lead to increase in ARHI mRNA expression as seen in one FTC and four of our FA samples. This finding is supported by recent observations employing comparative genomic hybridization that show gain in copy number of chromosome 1 in follicular thyroid neoplasia (8). However, our findings indicate that other aspects might also influence ARHI expression. The almost mosaic-like methylation pattern in CpG Island I, seen in one of our samples, has been reported in breast cancer to a much higher frequency (5). Such aberrant methylation might be of particular interest because it might portray relaxation or loss of imprinting. The possibility of a high degree of interaction between imprinted genes and their regulatory mechanisms over a larger distance and even across alleles has been reported for an imprinted domain on 11p15, which harbors, among others, the imprinted tumor-suppressor gene CDKN1C that shows similarities to ARHI in the mechanism of how silencing occurs (20, 21, 22, 23). In this respect, ARHI and the 1p31 region will be very interesting for further research to determine whether the genetic and epigenetic events reported by us might influence other genes located on 1p31 or vise versa. However, we currently have no evidence for such an interaction.
Because of the existence of several histologic variants of follicular neoplasias, how each variant relates to one another at the molecular level in the context of carcinogenesis has been studied at great length. Whereas it is widely accepted that FTCs and PTCs are distinct at both the morphological and molecular pathway levels (1, 16, 24, 25), it remains controversial whether HCCs, the oncocytic variant of FTC, is distinct from classic FTC. Recent reports propose that HCCs should be considered self-contained because unique molecular alterations, clearly differentiating them from FTCs and PTCs, have been discovered (26, 27, 28). Our observations clearly show that not only both classic FTCs and HCCs underexpress ARHI, plausibly due to genomic deletion causing loss of the nonimprinted paternal allele, but also, more importantly, that this holds true for minimally invasive FTCs and minimally invasive HCCs. These findings clearly suggest that loss of ARHI mRNA expressions serves as an initiator of follicular thyroid carcinogenesis, likely in conjunction with other events yet to be described.
One limitation of this study is that we cannot conclusively prove that ARHI gene transcriptional levels are translated proportionately into protein expression. Nonspecific binding of the commercially available antibody against ARHI (Invitrogen) was so severe that a consistent and reproducible immunohistochemical result was not possible in thyroid tissue. However, the translation of ARHI gene expressional changes into protein expression has been shown by Yuan et al. (5) for breast cancer (70% of 20 specimens analyzed), and our own Western blot experiments also suggest that this is true for thyroid tissue (data not shown).
In conclusion, we have shown that down-regulation of ARHI occurs at a very early stage in the malignant transformation of follicular thyroid epithelial cells. Because the putative maternally imprinted tumor-suppressor ARHI is only monoallelically expressed (3, 5, 7), a corollary of the two hit hypotheses of Knudson (29) may be adopted as a model for follicular thyroid oncogenesis. Because the genomically imprinted ARHI allele is already silenced (first hit) during normal development, the remaining expressed allele is open to somatic genetic or epigenetic events (second hit), early in neoplasia as is likely the case in follicular carcinogenesis. Further analysis using a larger set of FAs, and especially atypical variants of FAs, need to be performed to conclusively answer the question whether indeed down-regulation of ARHI expression might serve to identify a population of FAs that might be prone to further transformation to a variety of follicular carcinomas including classic FTCs and HCCs. Additional analysis will help to determine whether up-regulation of ARHI in FAs is, in contrast, associated with inability to transform to FTCs.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: ARHI, Tumor suppressor gene ras homolog I; Ct, threshold cycle; FA, follicular adenoma; FTC, follicular thyroid carcinoma; HCC, Hürthle cell carcinoma; LOH, loss of heterozygosity; PTC, papillary thyroid carcinoma; Q-PCR, quantitative PCR; TSA, trichostatin A.
This work was supported by the Comprehensive Cancer Center, The Ohio State University (Columbus, OH), which is supported in part by Grant P30 CA16059 from the National Cancer Institute, Bethesda, MD. C.E. is a recipient of the Doris Duke Distinguished Clinical Scientist Award.
Received July 22, 2004.
Accepted November 4, 2004.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J J L Wong, N J Hawkins, and R L Ward Colorectal cancer: a model for epigenetic tumorigenesis Gut, January 1, 2007; 56(1): 140 - 148. [Full Text] [PDF]
|
|
|
|
 |
|
 F. Weber, R. E. Teresi, C. E. Broelsch, A. Frilling, and C. Eng A Limited Set of Human MicroRNA Is Deregulated in Follicular Thyroid Carcinoma J. Clin. Endocrinol. Metab., September 1, 2006; 91(9): 3584 - 3591. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. S. Sarquis, F. Weber, L. Shen, C. E. Broelsch, S. M. Jhiang, J. Zedenius, A. Frilling, and C. Eng High Frequency of Loss of Heterozygosity in Imprinted, Compared with Nonimprinted, Genomic Regions in Follicular Thyroid Carcinomas and Atypical Adenomas J. Clin. Endocrinol. Metab., January 1, 2006; 91(1): 262 - 269. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Weber, L. Shen, M. A. Aldred, C. D. Morrison, A. Frilling, M. Saji, F. Schuppert, C. E. Broelsch, M. D. Ringel, and C. Eng Genetic Classification of Benign and Malignant Thyroid Follicular Neoplasia Based on a Three-Gene Combination J. Clin. Endocrinol. Metab., May 1, 2005; 90(5): 2512 - 2521. [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 |
