| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Special Feature |
Department of Endocrinology and Metabolism (R.E., A.V., L.A., E.M., C.N., L.G., A.P.), University of Pisa, 56124 Pisa, Italy; and Section of Endocrinology (F.P.), Department of Internal Medicine, Endocrinology and Metabolism, and Biochemistry, University of Siena, 53100 Siena, Italy
Address all correspondence and requests for reprints to: R. Elisei, M.D., Department of Endocrinology, University of Pisa, Via Paradisa 2, 56124 Pisa, Italy. E-mail: relisei{at}endoc.med.unipi.it.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The present study was designed to verify the sensitivity and specificity of Tg mRNA measurement, performed by quantitative real-time RT-PCR and to compare the usefulness of the Tg mRNA with that of serum Tg, detected by a sensitive chemiluminescent immunometric assay, which has been used in the follow-up of DTC patients for many years offering a good tool to predict their clinical status, being even more sensitive than WBS (6, 7, 8, 9, 10, 11).
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
We studied 80 patients (23 males and 57 females) treated with total thyroidectomy for DTC and 20 normal controls. As shown in Fig. 1
, subjects were enrolled on the basis of their clinical status and divided into four groups: 1) patients (n = 29) who, after a mean follow-up of 9.5 yr appeared disease free as demonstrated by one, or more than one, negative WBS and undetectable serum Tg levels while off L-T4 therapy (<1.0 ng/ml); 2) patients with thyroid remnant (n = 34), some of whom had detectable serum Tg (n = 22) and some with no undetectable serum Tg (n = 12); 3) patients with local or distant metastases (n = 17), all with very high serum Tg levels; and 4) normal controls, with no thyroid disease and detectable serum Tg levels (n = 20). Patients were enrolled on (n = 27) and off (n = 53) L-T4 therapy (Fig. 1). Controls showed a normal thyroid function with a mean TSH value of 1.5 ± 0.87 µU/ml. The patients included in the study were chosen to be negative for anti-Tg antibodies (inclusion criteria).
|
Neck ultrasound was performed in all subjects, both normal and DTC patients. An additional control, with a neck ultrasound examination and a serum Tg evaluation on L-T4 therapy, was performed 24–30 months after the measurement of Tg mRNA.
Patients with elevated blood Tg mRNA and with a follow-up (after the Tg mRNA measurement) of at least 4 yr were also submitted to a Tg stimulation test performed with recombinant human TSH (rhTSH).
Quantitative Tg mRNA RT-PCR
Total RNA was isolated from 3 ml whole blood using the PUREscript kit (Gentra Systems, Minneapolis, MN) according to the manufacturer’s protocol. Total RNA was treated with DNase I, precipitated, and suspended in diethylpyrocarbonate-treated water; it was quantified by measuring absorbance at 260/280-nm UV light, and 1 µg was reverse transcribed to cDNA in 50 µl using 250 pmol of random hexamers, 40 U Rnasin, 20 U avian myeloblastosis virus reverse transcriptase, and 1 mM dNTP. The reaction mixture was incubated 10 min at 65 C, 60 min at 42 C, and 5 min at 95 C. Ten percent of the cDNA was used in the subsequent quantitative PCR. Each sample was assayed in triplicate according to conditions recommended by Applied Biosystems (Foster City, CA); cDNA was added to a mixture of 1x Universal Master Mix, 900 nM each primer, and 200 nM probe in a final volume of 25 µl. Reaction mixtures were incubated for 2 min at 50 C, denatured for 10 min at 95 C, and subjected to 40 cycles of a two-step PCR consisting of a 15-sec denaturation at 95 C and 1 min annealing/extension at 60 C.
To perform the quantitative RT-PCR for Tg mRNA we used the real-time sequence detection system 7700 (PE Applied Biosystems, Foster City, CA), in which detection of PCR products is obtained during the amplification process using a fluorogenic probe. The amount of PCR product is measured by determining the first cycle of the logarithmic phase of amplification, called threshold cycle, which is dependent on the cDNA initial concentration. The standard curve was constructed using triplicates of serial dilution of cDNA (1–1000 pg) obtained from a pool of normal control thyroid tissues. The PCR primers spanned between two exons, and the probe sequence covered the junction of the two exons; their sequences were as follows: sense, 5'-GTGCCAACGGCAGTGAAGT-3'; antisense, 5'-TCTGCTGTTTCTGTAGCTGACAAA-3'; probe, 5'-FAM-ACAGACAAGCCACAGGCCGTCC T-TAMRA-3'. Samples omitting reverse transcriptase or cDNA were included in each run as controls of potential laboratory and/or assay contamination. cDNAs derived from cell lines (CHO, ARO, and FRO), which are known to not express Tg mRNA, were used to determine the cutoff for Tg mRNA values.
Before performing the quantitative RT-PCR for Tg mRNA, all cDNAs were amplified for a fragment (254 bp) of a ubiquitous gene (N-Ras) to verify their quality. Mean circulating Tg mRNA values are expressed in picogram equivalents per microgram of thyroid RNA where 1 pg equivalent means that the amount of Tg mRNA present in the sample is equivalent to the Tg mRNA amount present in 1 µg of normal thyroid RNA. This is a relative quantification due to the use of normal thyroid RNA as standard according to standard methods described previously (26).
Serum Tg, TSH, and anti-Tg antibody measurements
Serum Tg was measured using a commercial chemiluminescent immunometric assay (Diagnostic Products Corp., Los Angeles, CA) with the lower detection limit of 0.2 ng/ml and a functional sensitivity of 0.9 ng/ml. In our laboratory, the intra- and interassay coefficients of variation of the method are 4.3 and 7%, respectively. Based on the functional sensitivity of the assay we selected 1 ng/ml as the cutoff value discriminating undetectable from detectable Tg levels.
Serum Tg of disease-free patients was also measured with an ultrasensitive ELISA method (RSR Ltd., Cardiff, UK), with a functional sensitivity of 0.03 ng/ml.
Anti-Tg autoantibodies were measured in all sera by an immunoradiometric assay (ICN Pharmaceuticals, Inc., Asse Relegem, Belgium). Patients with circulating anti-Tg antibodies were not enrolled in the study.
Serum TSH was measured using an ultrasensitive commercial chemiluminiscent immunometric assay (Diagnostic Products).
Statistical analysis
Statistical analysis of data was performed using StatView 4.5 software (Abacus Concepts, Inc., Berkeley, CA). Depending on data to be analyzed, we used Student’s t, Mann-Whitney U, Kruskall Wallis, or regression tests. P < 0.05 was considered as significant.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
A strong linear relationship (r = 0.988–0.999) was constitutively observed in all standard curves. The mean interassay coefficient of variation (CV) for the threshold cycle of the total thyroid RNA calibrators (1, 10, 100, and 1000 pg) assayed in triplicate was 1.0–1.5%; the intraassay CV for the threshold cycle of the calibrators assayed in triplicate was 0.0–2.0% with no major differences among different concentrations of total thyroid RNA tested. The mean interassay CV of two unknown samples that were expected to have a low and a high Tg mRNA concentration were 25 and 7.7%, respectively. As expected, the nonhuman cell line CHO (Chinese hamster ovary) consistently failed to demonstrate detectable PCR product at 40 cycles. Because the two human anaplastic thyroid carcinoma cell lines (ARO and FRO), expected to be negative for Tg, became positive at 36 and 37 threshold cycles, respectively, the corresponding level of 6 pg equivalents was arbitrarily used as the cutoff value. To rule out possible accidental laboratory contamination of RT-PCR, negative controls without reverse transcriptase or without cDNA were performed and were always negative.
Patient analysis
As shown in Fig. 2, we found a wide spread of Tg mRNA values in all groups of analyzed subjects. Mean ± SD, range, and median values of circulating Tg mRNA in the four different groups of enrolled subjects are shown in Table 1. The mean Tg mRNA of the disease-free patients’ group (87.8 ± 125.9) was higher than that of the with-remnant patients’ group (64.1 ± 88.6), both in those with detectable and those with undetectable serum Tg (73.1 ± 103.8 and 47.5 ± 50.4, respectively). On the contrary, the mean Tg mRNA of the disease-free patients’ group was lower than the mean Tg mRNA of the metastatic patients (195.3 ± 278.8), but this difference was not statistically significant (P = 0.08). The only significant difference was found between the Tg mRNA mean values of the group with remnant and the group with metastatic disease (P = 0.015). Mean values of Tg mRNA in normal controls was similar to those of the first two groups (disease-free and with-remnant patients) and lower, but not significantly, than that of metastatic patients (P = 0.07). We found that disease-free patients had a lower median circulating Tg mRNA value with respect to metastatic patients, whereas no difference was found when they were compared with patients with remnant. In any case, statistical analysis showed that the difference between the median value of Tg mRNA in disease-free and metastatic patients was not significant (P = 0.16).
|
|
). A discrepancy was also found in the metastatic group in which only 88% of patients were Tg mRNA positive, whereas 100% of them were serum Tg positive, as expected (Table 1). From these data, the sensitivity and the specificity of the method are 82.3 and 24.2%, respectively, with a positive predictive value and a negative predictive value of 65.6 and 43.7%, respectively. Moreover, when we compared the Tg mRNA and serum Tg values in each patient of each group, we did not find any correlation (Fig. 3).
|
). All disease-free patients had undetectable levels of serum Tg both in the group analyzed on L-T4 (n = 18) and in the group analyzed off L-T4 (n = 11), as assessed by chemiluminescent immunometric assay. The measurement of serum Tg with the ultrasensitive ELISA method confirmed undetectable levels (<0.03 ng/ml) in eight cases and revealed very low levels of serum Tg (0.03–0.5 ng/ml) in the other 21 cases. However, these Tg values did not correlate with the values of blood Tg mRNA (Table 2).
|
All disease-free patients showed undetectable levels of serum Tg (<1.0 ng/ml) and a negative neck ultrasound at the control performed 24–30 months after their enrolling in this study. The nine patients of this group who submitted to a rhTSH stimulation test for serum Tg 4 yr after the blood Tg mRNA measurements were confirmed to have both basal and stimulated negative serum Tg.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The alternative method of Tg mRNA measurement has been proposed to determine the persistence of the DTC in the above described situations (15, 16, 17). In our hands, the Tg mRNA quantitative measurement showed a good sensitivity (82.3%) but, unfortunately, a very low specificity (24.2%).
According to Ringel et al. (17), we found an evident difference among the median values of the analyzed groups, with the highest value in the group of metastatic patients, but these differences were not statistically significant. It is worth noting that all but two metastatic patients had detectable levels of Tg mRNA, suggesting that circulating thyroid cells might be present in the blood of these patients and contribute to the Tg mRNA detection. However, 75.8% of disease-free patients, who according to our protocols can be considered definitively cured, were positive for circulating Tg mRNA to various degrees of expression (
6–535 pg equivalents/µg thyroid RNA). These patients did not show any residual thyroid tissue or derived lymph node metastases at the neck ultrasound, either at the time of the study or at the 24–30 months control, that could justify the Tg mRNA-positive results. The long follow-up (median, 7 yr; range, 3–29 yr) of these patients, without any recurrence in their clinical history, does not support the idea that Tg mRNA detectable levels could be predictive of future recurrences as postulated by Ringel et al. (17) to explain their similar results found in 38% of patients without any 131I uptake. Our interpretation is also supported by the negativity of the rhTSH stimulation test for serum Tg in nine patients of the disease-free group who showed positive levels of Tg mRNA (11.8–336 pg equivalents/µg thyroid RNA) 4 yr before.
Recently, other authors (21, 22, 23, 24, 25) have shown that the Tg mRNA expression is not specific for thyroid tissue and not correlated with the diagnosis of thyroid cancer. At variance with us, in two of these studies (21, 23), the Tg mRNA was detected with a qualitative and a semiquantitative method, respectively. Takano et al. (22) used a real-time quantitative RT-PCR method very similar to us, but they employed a set of primers located in regions affected by alternative Tg mRNA splicing. However, these primers can result in an underestimation of the Tg mRNA, whereas, similarly to us, Takano et al. found Tg mRNA in all samples obtained after total thyroidectomy, including four medullary thyroid carcinomas. On the contrary, in preliminary experiments devoted to setting up the method, we found a set of primers located on exons 7 and 8 that gave origin to a PCR product also in samples that are known to be negative for Tg expression (data not shown), thus raising the question of the presence of a Tg pseudogene whose amplification could result in false positive cases. In any case, to overcome the above described problems, in the present study we used a set of primers that have been reported to be the only one not affected by alternative Tg mRNA splicing (14, 19) and that, in our preliminary experiments, gave no amplification in the CHO cell line.
To minimize the problem of the low specificity of our method, we might choose a higher cutoff based on the positivity of disease-free patients, but in contrast, we would lose several positive metastatic patients because of the great overlap of the Tg mRNA values in the two groups. The choice of the cutoff is fundamental but also very arbitrary; we decided to use the value at which human cell lines, proven not to be able to express Tg, became positive for Tg mRNA, assuming that this value could be due to a background noise. Some authors (22) did not indicate any cutoff, but their samples are all positive. We also would obtain similar results if we eliminated the cutoff of 6 pg equivalents/µg thyroid RNA.
Another explanation for the detection of Tg mRNA in the 75.8% of disease-free patients might be the illegitimate transcription that is a well known phenomenon consisting of very low transcription levels of any gene in any cell type (29). As quantitative real-time RT-PCR is a very sensitive technique, we can suppose that it is able to detect the low amount of Tg mRNA present in nonthyroid blood cells. Because we extracted RNA from whole blood and nonthyroid cells are the great majority, we can expect that illegitimate transcription creates a noise covering the effective Tg mRNA derived from circulating thyroid cells. In preliminary experiments, we separated by Ficoll the mononuclear (lymphocytes and monocytes) from the polymorphonuclear (granulocytes) fraction. By qualitative RT-PCR, we found that both fractions expressed Tg mRNA (data not shown). On this basis, we cannot rule out that our quantitative RT-PCR results are not affected by illegitimate transcription. In contrast, we believe that this cannot be the explanation of our results, because we found different values, even not statistically significant, among the groups of analyzed patients, especially between cured and metastatic patients, whereas illegitimate transcription should produce similar results in all subjects. Moreover, the introduction of a control gene for illegitimate transcription (19) could be a useful strategy to reduce this problem.
Because the sensitivity of our method is mainly based on the analysis of the results obtained in the disease-free patients, who by definition are serum Tg free, one could raise the question of whether these patients are really cured or whether the positivity of Tg mRNA suggests a false negative result of serum Tg. In a retrospective study (6), we demonstrated that only 0.6% of a group of 315 patients with at least one negative WBS and negative serum Tg off L-T4 developed a recurrence of the thyroid carcinoma after a mean follow-up of 8.51 yr. According to these results, it is hard to believe that 75.8% of patients with negative clinical findings and defined as disease free will develop a recurrence of the thyroid disease in the future. The theoretical influence of the TSH levels on the Tg mRNA levels reported in other studies (19) seems to not affect our results because no difference has been found in Tg mRNA results between disease-free patients analyzed on or off L-T4 therapy.
It has been demonstrated that both tumor and normal cells circulate in the blood (30, 31), and in particular thyroid cells have been found in blood smears of normal subjects (16). These results suggest that this method might be a useful tool for the management of cancer patients, but, at least in our hands, the method seems to need additional improvements looking for potential causes of its low specificity. At the moment, the technique might be not yet well refined, and it is known that several steps, with particular regard to the RT efficiency and the choice of primers, may greatly affect the sensitivity of the method. These technical problems need to be accurately evaluated before starting with clinical application.
In our opinion, the recent suggestion (19, 20) to use both methods in the clinical practice is not convenient because the Tg mRNA measurement is expensive and time consuming with respect to the serum Tg determination. Furthermore, the opposite results between the two methods (from 38–75% of disease-free patients according to different studies) can mix up the choice of therapeutic strategy. In conclusion, because serum Tg measurement is a very sensitive and specific marker for differentiated thyroid cancer, at this moment, we recommend that the follow-up of patients affected by DTC must be performed using serum Tg rather than blood Tg mRNA measurement.
|
Footnotes |
|---|
Abbreviations: DTC, Differentiated thyroid carcinoma; rh, recombinant human; Tg, thyroglobulin; WBS, whole-body scan.
Received August 4, 2003.
Accepted October 9, 2003.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
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]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
