| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Articles |
Equipe Mixte Inserm—Université 00-18 (F.S., P.Re., Y.M.), Laboratoire de Biochimie et Biologie Moléculaire, Service d’Endocrinologie (P.Ro. V.R., J.-C.B.), Nutrition et Médecine Interne, Centre Hospitalier Universitaire, F-49033 Angers cedex 01, France
Address all correspondence and requests for reprints to: F. Savagner, Inserm EMI-U 00–18, Laboratoire de Biochimie et Biologie Moléculaire, CHU, 4 rue Larrey, F-49033 Angers cedex 01, France. E-mail: frsavagner{at}chu-angers.fr
Abstract
Serum Tg (sTg) assays are sometimes unsatisfactory for monitoring thyroid cancer because interference caused by anti-Tg antibodies may reduce the sensitivity of the tests during thyroid hormone therapy. We have therefore developed a complementary method using real-time quantitative RT-PCR based on the amplification of Tg mRNA. Two different pairs of primers were used for the determination of the frequency of one of the variants of the alternative splicing of Tg mRNA. The frequency of this variant was as high in patients (n = 40) as in controls (n = 30), accounting for about 33% of the total Tg mRNA. Using appropriate primers, we observed that Tg mRNA values in controls varied according to the volume of thyroid tissue and the TSH concentration. The Tg mRNA values allowed the definition of a positive cutoff point at 1 pg/µg total RNA. This cutoff point, tested on the group of patients treated for thyroid cancer, produced fewer false negative results than those obtained with sTg assays. The standardized, highly sensitive real-time RT-PCR technique may therefore prove useful as a complement to sTg assays, particularly for patients with recurrent thyroid cancer receiving T4 therapy.
ALTHOUGH MUCH PROGRESS has been made in the treatment of differentiated thyroid cancer, the disease still recurs, with progression of the tumor, in 20–40% of patients (1). It is therefore necessary to monitor patients after thyroidectomy, by periodic radioiodine scans and serum Tg (sTg) assays. However, the presence of anti-Tg antibodies limits the accuracy of sTg assays when TSH is suppressed after thyroidectomy. Moreover, an international standard method of sTg measurement, based on purified, stable and certified immunoreactive Tg as a reference, has not yet been developed (2, 3).
Tumor cells circulate in the peripheral blood or lymphatic channels before developing metastases at a distant site. Detection of such cells provides a tool for the early diagnosis of tumor metastases and may lead to better prognosis in cases of recurrence (4). The RT-PCR is a highly sensitive technique for the detection of tissue-specific mRNA transcripts. This method has been tested with the amplification of cancer-specific mRNA from peripheral blood to detect micrometastases in cases of prostate cancer and neuroblastoma (5, 6).
More recently, the RT-PCR method has been used during T4 therapy, for monitoring patients treated for thyroid carcinoma (7). However, because the different protocols used for the end-point RT-PCR have not been standardized, the sensitivity of thyroid cell detection varies according to the number of PCR cycles used (8, 9, 10). Furthermore, attempts to increase the sensitivity of the end-point PCR technique have led to a reduction in specificity (11).
Quantitative RT-PCR has proved to be more adequate for the determination of a sensitive positive threshold for a specific amplification (12). After normal circulating thyroid cells had been identified by the cell sorting method, this technique demonstrated that all normal subjects have detectable circulating Tg mRNA (8). Using the Tg mRNA values of healthy subjects as controls, a positive threshold was fixed such that 38% of the patients with no residual disease and 84% of the patients with either cervical or distant metastases were positive for Tg mRNA (13). However, this cutoff point, determined from studies on healthy subjects, led to false negative values for patients with metastases. Because false negative results should be eliminated as far as possible, the cutoff point needs to be readjusted, taking into account the illegitimate transcription level of Tg mRNA in leukocytes. Positive Tg mRNA in apparently cured patients could represent either false positive values or true positive values with a more sensitive rate of detection.
The discordance between the detectable sTg and the undetectable Tg mRNA may be explained by the alternative splicing of Tg mRNA in thyroid cells. The splicing map of human Tg mRNA is composed of at least 16 alternative splicings, located mainly on 4 exonic cassettes (14). Using the PCR method, Tg cDNA molecules of abnormal size were found in benign and malignant tumor tissues as well as in normal tissues, with no particular relationship between the distribution of the alternative forms of Tg mRNA and the pathology (15). However, to our knowledge, the relative proportion of alternative variants of Tg mRNA has not been quantified in thyroid samples from diseased and normal subjects.
To estimate the risk of error in Tg mRNA measurement attributable to alternative splicing, we calculated the percentage of 1 of the variants of the Tg mRNA, with respect to the total Tg mRNA in circulating thyroid cells. We used real-time quantitative RT-PCR to detect the Tg mRNA level in 30 controls, with or without T4 therapy, and having different volumes of thyroid tissue, to define a more accurate threshold value, taking into account the level of illegitimate Tg mRNA transcription. Finally, we tested the validity of the cutoff point thus defined on 40 patients being monitored for thyroid cancer while receiving T4 therapy. Because immunoassays and quantitative RT-PCR measure different aspects of Tg, it may be of interest to consider the extent to which these two methods might be complementary.
Subjects and Methods
Subjects
We studied 30 control subjects (25 women and 5 men) at the University Hospital of Angers, France. Ten of the controls were normal euthyroid subjects (group C1), 12 were on T4 therapy for multinodular goiter (group C2), and 8 on T4 therapy after total surgical thyroidectomy, without postoperative radioiodine ablation for thyroid microcancer (group C3).
Among the 40 patients (31 women and 9 men) followed-up, at the University Hospital of Angers, for thyroid macrocancer, there were 32 cases of papillary carcinoma and 8 cases of follicular carcinoma. All patients had total thyroidectomy, followed by radioiodine therapy, and were evaluated during T4 therapy. Patients were monitored for the presence or absence of disease, by total body scans, after T4 withdrawal. Fifteen patients (group P1) had been free of the disease for more than 5 yr; but in 3 cases, the presence of over 50 U/ml anti-Tg antibodies was suspected of interfering with the Tg immunoassay. Ten patients (group P3) had metastases, proximal in 3 cases and distant in 7 cases. The other 15 patients (group P2) had detectable sTg levels during T4 therapy, with no anti-Tg antibodies, and a negative total body radioiodine scan. In 1 patient who developed pulmonary metastases, Tg mRNA was measured in parallel with sTg, first during T4 therapy, and then after T4 withdrawal (during radioiodine therapy), and again 2 d after therapeutic radioiodine administration.
The sTg levels were determined by immunoradiometric assays (Bio-Rad Laboratories, Inc. Hercules, CA). The TSH concentration, measured by immunoradiometric assays (Immunotech/Beckman Coulter, Inc. Villepinte, France) was undetectable (<0.1 mU/liter) in group C3 and in the 40 patients (P1, P2, and P3). TSH values were 0.38 ± 0.26 mU/liter in group C2 and in the normal range for euthyroid subjects (group C1, 3.1 ± 2.1 mU/liter). Six weeks after T4 withdrawal, whole-body radioiodine scans were performed 48–72 h after the administration of a 2- to 5-mCi dose of 131I. The scans showing no uptake at all of radioiodine, on visual inspection, were considered negative.
RNA extraction and cDNA preparation
Total RNA was isolated from 2 ml whole blood by immediate addition of 6 ml Trizol LS reagent (Life Technologies, Inc., Gaithersburg, MD), following the manufacturer’s recommendations. Using this method, the efficiency of total RNA recovery was satisfactory (4.5–8 µg per 2 ml). To generate cDNA, 1 µg RNA was first denatured at 70 C with 75 pmol random hexamer primers (Promega Corp., Madison, WI) for 5 min before quenching on ice; then 0.5 mM final of each of the 4 deoxynucleotide triphosphates, 10 mM dithiothreitol, 10 U ribonuclease inhibitor, and 200 U superscript II (Life Technologies, Inc.) were added to the 5x buffer to make up a final vol of 20 µl reaction mix. The reaction mix was incubated for 1 h at 42 C. The reverse transcriptase was inactivated at 70 C for 15 min. The quality of the cDNA was tested by the amplification of glyceraldehyde-3 phosphate dehydrogenase in the different cDNA samples obtained.
Quantitative PCR
Real-time quantitative PCR, monitored by fluorescent SYBR Green I dye (Lightcycler; Roche Diagnostics, Basel, Switzerland) offers a user-friendly, automated technique (16). After setting a threshold value, the Lightcycler software interpolated a straight line through a defined number of data points situated in the log-linear phase for each fluorescence signal obtained with Tg gene-specific primers. For known standards, the crossing points were plotted against the logarithms of the concentrations. The Tg cDNA from different patients was quantified by comparing the crossing points of samples with those of the standards containing 102-105 copies of Tg cDNA as the starting concentration. Standard PCR products were generated from plasmids containing the appropriate cDNA insert as template. The copy number in the final sample was determined by two independent methods, i.e. spectrophotometry and gel analysis after restriction. A sequence-specific standard curve was plotted using serial dilutions of the target gene standard PCR product, and the same primers were used to amplify the cDNA. Template negative controls were included in the amplification reactions. Interassay (n = 5) and intraassay (n = 5) coefficients of variation, calculated for 3 concentrations of the standard, were 9.1% and 3.8%, respectively. The specificity and similarity of SYBR green dye PCR products were attested by plotting the melting curves of products, using the Lightcycler software.
Quantification of assays
Real-time quantitative PCR was used with two sets of primers: TgAf: 5'-GAG CCC TAC CTC TTC TGG CA-3'; TgAr: 5'-GAG GTC CTC ATT CCT CAG CC-3' and TgBf: 5'-AGG AGG TCA GTT GCC CCA TG-3'; TgBr: 5'-GAC CTC TAC CCA TAA AGA CC-3', respectively, amplifying the 320-bp and 199-bp fragments. Different pairs of primers have been designed from the Tg sequence (17) to amplify, from the same cDNA, two nonoverlapping regions encompassing two exons each. The first region encompassed exons 10 and 11 (nucleotides 2631–2951), in which no alternative splicing has yet been described. The second region encompassed exons -7 and -6 (nucleotides 7351–7550), in which alternative splicing has been found (14). Prostate-specific antigen (PSA, located on chromosome 19) was used as a control gene for illegitimate transcription, using primers described elsewhere (18). Each sample was assayed in duplicate under the following reaction conditions: 2 µl master mix containing Taq DNA polymerase, deoxynucleotide triphosphates, and SYBR green I (Fastart DNA Master SYBR Green I Kit, Roche Diagnostics) were incubated with MgCl2 (4 mM), forward and reverse primers (0.5 µM), and 2 µl of template (cDNA or a standard with a known copy number) in a final vol of 20 µl. The reaction mixture was denatured for 5 min at 95 C and subjected to 45 cycles of a three-step PCR consisting of a 1-sec denaturation step at 95 C, a 10-sec annealing step at 58 C, and a 10-sec extension step at 72 C. To confirm product identity, several Tg RT-PCR products were sequenced using a cycling kit (Ceq DTCS Quick Start Kit; Beckman Coulter, Inc., Fullerton, CA) and an automated capillary sequencer (Ceq 2000; Beckman Coulter, Inc., Fullerton, CA).
All results are reported in terms of picograms of Tg mRNA per microgram of total RNA.
Statistical analysis
Statistical analysis was performed using SPSS, Inc. (Chicago, IL) software (version 9.0). The differences in the Tg mRNA values, between the different groups studied, were analyzed using the Mann-Whitney U test and considered statistically significant for P 
0.05. All the numerical values below are expressed in terms of the mean ± SD.
Results
Proportion of Tg mRNA alternative splicing in total Tg mRNA
In patients monitored for thyroid cancer (n = 40), compared with controls (n = 30), alternative variants in exons -7/-6 were identified in 20 and 22% of the samples, respectively. When present, alternative Tg mRNA variants accounted for 33% and 34% of the total Tg mRNA in patients and controls, respectively.
Evaluation of quantitative RT-PCR assay in controls
Depending on the therapy and the presence of thyroid tissue, there was a significant difference between the control subgroups: 1) In group C1 (n = 10), the Tg mRNA measurement, evaluated with the pair of primers encompassing exons 10 and 11, was 10.6 ± 3.1 pg/µg total RNA. 2) Group C2 (n = 12) had significantly lower circulating levels of Tg mRNA than group C1 [4.1 ± 2.4 pg/µg (P 0.001)]. Group C3 (n = 8) had the lowest level of Tg mRNA (1.6 ± 0.3 pg/µg), significantly lower than that of group C2 (P 0.001) or group C1 (P 0.001). Tg mRNA levels were correlated with TSH levels (mean TSH values varied from 3.1 mU/liter for group C1 to 0.38 mU/liter for group C2 and were undetectable for group C3). Figure 1
shows the results of Tg mRNA for the three control groups.
|
Evaluation of quantitative RT-PCR assay for monitoring patients after thyroid cancer
Using the cutoff value of 1 pg Tg mRNA/µg RNA, 18% of cured patients (group P1) had positive Tg mRNA levels, compared with 100% of patients with metastases (group P3). The mean Tg mRNA levels were, respectively, 1.3 and 15.4 pg/µg total RNA for P1 and P3 groups. The mean Tg mRNA level was significantly higher in group P3 than in group P1 (P 0.001). Figure 2 shows the results for patients divided into groups according to their radioiodine diagnostic scans and sTg assays. Patients with distant metastases seemed to have a higher mean Tg mRNA level (18.7 pg/µg total RNA) than patients with proximal metastases (7.5 pg/µg total RNA). However, these two groups, comprising only seven patients and three patients respectively, were too small for statistical analysis.
|
Figure 3 shows the different kinetics of sTg and Tg mRNA assays for one patient who was studied prospectively for the treatment of pulmonary metastases. The levels of sTg and Tg mRNA varied according to the TSH value and to the efficacy of treatment, with more rapid kinetics for Tg mRNA than for sTg levels.
|
0.02) but not from that of group P3 (P 0.44).
Table 1 represents Tg mRNA values, indicating percentages of patients considered positive for Tg mRNA with a cutoff value 1 pg Tg mRNA/µg total RNA.
|
RT-PCR has been shown to be a useful method for monitoring the eradication of cells expressing a tumor-specific transcript. In cases of thyroid cancer, this method has been demonstrated to be more sensitive than sTg assays, especially during T4 therapy or in the presence of anti-Tg antibodies (13). However, the existence of several alternative splicings of the Tg mRNA has raised the question of the frequency of these variants in thyroid tissues in diseased and normal subjects. In fact, the transcription pattern of Tg mRNA is heterogeneous in thyroid cancer tissues, as well as in normal thyroid tissues, and no specific pattern has been found in differentiated thyroid carcinoma (19). The examination of the pairs of primers used in different studies (7, 8, 9, 10, 11, 13) reveals that the amplified sequences were located in regions affected by alternative Tg mRNA splicings in all but one report (13). In this report, quantitative Tg mRNA assays were based on primers located in a region that did not seem to be affected by alternative splicing. The quantity of Tg mRNA may therefore have been underestimated by the RT-PCR method in the majority of studies. By determining the frequency of one of the alternative variants of Tg mRNA, we found that the default in Tg mRNA measurement could represent up to 33% of the value, on the average, in 20% of the cases, whatever the course of the thyroid cancer in the patients.
However, with the primers used in our study, a few patients showed a discrepancy between detectable sTg levels after T4 withdrawal and undetectable Tg mRNA levels. This was the case in 4 of 15 patients from the group suspected of recurrence (group P2). Although 1 patient with proximal metastases received a second radioiodine therapy, the sTg levels remained unchanged at 1.7 ng/ml for 4 months during T4 therapy. The remaining patients have had no recurrence of thyroid cancer for over 10 yr. All of them presented sTg values between 0.5 and 2.0 ng/ml during T4 therapy, but it has been suggested that these values reflect false positive immunoassay results (20). On the other hand, Tg-secreting cells may still be present although not circulating at the time of blood collection. The analysis of serial blood samples would clear up this point.
Among 9 of the 15 patients of group P2 with positive Tg mRNA levels, 2 displayed radioiodine uptake in the thyroid bed, 9 and 11 months later. We thus confirm the report that Tg mRNA detection is more sensitive than sTg measurement during T4 therapy, especially in cases of local metastases (13). For the remaining patients, no detectable metastases were discovered during the last 6 months. However, studies on patients with prostate cancer have shown that surgical tissue removal leads to transient hemoconversion from RT-PCR-negative to RT-PCR-positive in about 30% of the cases (21). We therefore recommend systematic Tg mRNA measurements, in parallel with sTg assays, for monitoring the development of local or regional metastases in cases of thyroid cancer.
The variation in the Tg mRNA level in the different groups of controls suggests that the number of differentiated circulating thyroid cells is related to the presence of residual thyroid tissue. Because controls receiving suppressive T4 therapy (group C2) have lower Tg mRNA levels than healthy controls (group C1), these variations are likely to reflect specific thyroid cell expression rather than illegitimate transcription of Tg mRNA. This is further confirmed by absence of variation in PSA mRNA in the different groups. As previously reported (9), epithelial cells expressing Tg may circulate in the peripheral blood of healthy subjects. We have shown that T4 therapy, administered to control subjects, leads to a decrease in the transcription rate of the circulating thyroid cells. However, this rate may vary in individual circulating cells (22). As indicated in a previous study (23), the cutoff value we have proposed corresponded to the detection of one thyroid cell per milliliter of whole blood, and we have shown that real-time quantitative PCR was well adapted for detection of a unique cell (24). The variation of the Tg mRNA levels in the different control groups may thus reflect either the differences in the number of circulating thyroid cells or the differences in TSH-dependent transcriptional activity in a constant number of cells. Therefore, the results of Tg mRNA assays should always be interpreted taking into account the TSH level, as in the case of immunoreactive Tg assays.
In patients with recurrent thyroid cancer, evidenced by sTg measurements and radioiodine scans, Tg mRNA was detected in 100% of the cases (group P3). We found that the Tg mRNA levels were higher for distal metastases (seven cases) than for proximal metastases (three cases). However, the number of patients will have to be increased to allow for statistical analysis. All the patients with proximal metastases were free of anti-Tg antibodies. Conversely, all patients, except one with distal metastases, developed Tg-antibody levels ranging from 11–77 U/ml. These preliminary results disagree with studies associating the presence of anti-Tg antibodies and local metastases (25, 26). However, Tg mRNA seemed to be correlated with the presence of metastases, whatever the level of the antibodies. Because no threshold anti-Tg antibody level could be established below which no interference on sTg immunoassay occurs (27), the Tg mRNA assay represents the best method for monitoring patients with anti-Tg antibodies. The detection of 100% of the cases with metastases suggests that the volume of blood collected is sufficient for satisfactory detection of Tg mRNA, with no need for tissue-specific enrichment, after the elimination of interference attributable to illegitimate Tg mRNA transcription. However, the blood sample conditions we used, which are essential for satisfactory yield of RNA extraction, could limit the clinical applicability of the assay in some medical centers.
The measurement of the sTg and Tg mRNA kinetics in one of the patients undergoing radioiodine therapy confirmed that the Tg mRNA levels vary according to the TSH values. Moreover, the variation of the Tg mRNA values was faster than the variation of the sTg values, reflecting cell damage (decrease of transcription and release of Tg). This is in accordance with the half-life of 65.2 h for the Tg protein (28).
In conclusion, real-time quantitative RT-PCR offers all the advantages of a user-friendly, automated technique. Moreover, it eliminates the large variations observed with the end-point RT-PCR method. We have shown that the choice of appropriate primers is essential for any meaningful comparison of quantitative results. This method has led to excellent results with patients receiving T4 therapy, especially in the presence of anti-Tg antibodies that frequently perturb results of immunoassays. Finally, the information obtained from the Tg mRNA and sTg assays, each method having its specific kinetics, may be considered to be complementary. We therefore recommend the use of Tg mRNA assays, in addition to sTg measurements, for a better monitoring of selected patients followed for a recurrence of thyroid cancer and receiving T4 suppressive therapy.
Acknowledgments
We are grateful to C. Savagner for statistical analysis and to K. Malkani for his critical reading of the manuscript. We thank A. Coutolleau for her continuous support during the study.
Footnotes
This work was supported by grants from the French Association for Research on Cancer and from the Regional Delegation for Clinical Research of the University Hospital of Angers (PHRC PL 99-08).
Abbreviations: PSA, Prostate-specific antigen; sTg, serum Tg.
Received July 11, 2001.
Accepted October 16, 2001.
References
This article has been cited by other articles:
 |
|
 |
|
  S.-Y. Chia, M. Milas, S. K. Reddy, A. Siperstein, M. Skugor, J. Brainard, and M. K. Gupta Thyroid-Stimulating Hormone Receptor Messenger Ribonucleic Acid Measurement in Blood as a Marker for Circulating Thyroid Cancer Cells and Its Role in the Preoperative Diagnosis of Thyroid Cancer J. Clin. Endocrinol. Metab., February 1, 2007; 92(2): 468 - 475. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Wagner, R. Arciaga, A. Siperstein, M. Milas, I. Warshawsky, S. Sethu, K. Reddy, and M. K. Gupta Thyrotropin Receptor/Thyroglobulin Messenger Ribonucleic Acid in Peripheral Blood and Fine-Needle Aspiration Cytology: Diagnostic Synergy for Detecting Thyroid Cancer J. Clin. Endocrinol. Metab., April 1, 2005; 90(4): 1921 - 1924. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Chinnappa, L. Taguba, R. Arciaga, C. Faiman, A. Siperstein, A. E. Mehta, S. K. Reddy, C. Nasr, and M. K. Gupta Detection of Thyrotropin-Receptor Messenger Ribonucleic Acid (mRNA) and Thyroglobulin mRNA Transcripts in Peripheral Blood of Patients with Thyroid Disease: Sensitive and Specific Markers for Thyroid Cancer J. Clin. Endocrinol. Metab., August 1, 2004; 89(8): 3705 - 3709. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. D. Ringel Molecular Detection of Thyroid Cancer: Differentiating "Signal" and "Noise" in Clinical Assays J. Clin. Endocrinol. Metab., January 1, 2004; 89(1): 29 - 32. [Full Text] [PDF]
|
|
|
|
|
|
R. Elisei, A. Vivaldi, L. Agate, E. Molinaro, C. Nencetti, L. Grasso, A. Pinchera, and F. Pacini Low Specificity of Blood Thyroglobulin Messenger Ribonucleic Acid Assay Prevents Its Use in the Follow-Up of Differentiated Thyroid Cancer Patients J. Clin. Endocrinol. Metab., January 1, 2004; 89(1): 33 - 39. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. K. Gupta, L. Taguba, R. Arciaga, A. Siperstein, C. Faiman, A. Mehta, and S. S. K. Reddy Detection of Circulating Thyroid Cancer Cells by Reverse Transcription-PCR for Thyroid-stimulating Hormone Receptor and Thyroglobulin: The Importance of Primer Selection Clin. Chem., October 1, 2002; 48(10): 1862 - 1865. [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 |
