TERT mRNA Expression as a Novel Prognostic Marker in Papillary Thyroid Carcinomas

Abstract

Background:

Telomerase reverse transcriptase (TERT) promoter mutations have been found in a subset of papillary thyroid carcinomas (PTCs) and are associated with tumor aggressiveness and worse prognosis. However, little is known about the status of TERT mRNA expression and its relationship between TERT promoter mutations and clinicopathological features.

Methods:

We analyzed 159 PTC samples for TERT promoter mutations using direct DNA sequencing. TERT expression was measured using quantitative reverse transcription polymerase chain reaction. To examine low allelic frequency of TERT promoter mutations with high sensitivity, we used droplet digital polymerase chain reaction (ddPCR). The relationship between the status of the TERT promoter mutation/expression and clinicopathological features including recurrence risk was statistically analyzed.

Results:

TERT promoter mutations were found in 20 cases (12.6%). However, TERT expression was observed not only in the mutation-positive tumors but also in 56 of 139 (40.3%) mutation-negative tumors. Among them, we detected low allelic frequency of TERT promoter mutations in three samples (5.4%) using ddPCR. We confirmed a significant association between TERT promoter mutations and aggressive clinicopathological features in this series. The risk of recurrence of TERT mutation-negative/expression-positive tumors was significantly higher than that of the mutation-negative/expression-negative tumors, suggesting that TERT expression even in absence of a mutation confers a negative influence on PTCs. Moreover, when we reclassified the mutation-negative cases into two groups based on the TERT expression levels: expression-negative/expression levels <80th percentile and expression levels >80th percentile because minimal expression may have a negligible clinical impact, a higher hazard ratio for recurrence was observed. Interestingly, TERT expression levels in the mutation-negative PTCs were inversely correlated with patient age and the presence of BRAF mutations.

Conclusions:

We confirm a strong correlation between the presence of TERT promoter mutations and aggressive clinicopathological features in this PTC series. In addition, there were PTCs showing high TERT mRNA expression even in the absence of TERT promoter mutations. These cases also showed a significantly higher recurrence rate. Since the TERT promoter mutations are observed only in elderly patients, TERT mRNA expression can be a useful prognostic marker especially in younger PTC patients.

Introduction

The incidence of papillary thyroid carcinoma (PTC) has been increasing worldwide (1). PTC has generally a favorable prognosis; however, ∼10–15% of patients have recurrences either locally or/and at distant sites, some of which become refractory to treatment (2,3). To distinguish between high-risk and low-risk cases, there have been many studies evaluating the value of molecular markers to predict PTC aggressiveness and prognosis.

The BRAF^V600E mutation is the most frequent genetic change in PTCs. Its prevalence varies from 30% to 80% (4), probably depending on the population. Many studies have indicated an association between the presence of the BRAF^V600E mutation and aggressive clinicopathological features; however, its prognostic value, especially as an independent marker, is still debated. According to our results, the mutation rate in Japan is high (∼80%), and we did not find any correlation between the presence of the BRAF^V600E mutation and aggressive clinicopathological features and worse prognosis in our series (5).

Recently, mutations in the promoter region of the telomerase reverse transcriptase (TERT) gene have been found in many types of cancers including thyroid carcinomas. There are two hot spots, called C250T (chr5: 1,295,250C>T) and C228T (chr5: 1,295,228C>T) (6,7), and these are mutually exclusive. The average rate of the presence of these mutations in PTC has been reported to be ∼10% (8), and there seems to be no major differences between populations. The mutations create a binding motif for the E26 transformation-specific (ETS) transcription factors and upregulate TERT mRNA expression, especially when the ETS family members are activated (e.g., by BRAF^V600E) (9 –11).

While the primary function of TERT is to maintain telomere length, there is increasing evidence regarding its telomerase-independent oncogenic functions through NF-κB, Wnt/β-catenin, and MYC pathways (12 –14). In PTCs, many studies have demonstrated that the coexistence of the BRAF^V600E mutation and TERT promoter mutations is strongly associated with aggressive features and worse prognosis (5,15 –27). Moreover, TERT promoter mutations seem to be also associated with anaplastic transformation (28).

However, regarding the impact of TERT promoter mutations, two recent findings should be considered. First, Paulsson et al. have reported that there is a subset of cases showing TERT mRNA expression in the absence of TERT promoter mutations in follicular thyroid tumors (adenomas, follicular tumors with uncertain malignant potential, and carcinomas) (29). In that study, the authors have also demonstrated that tumors with positive TERT mRNA expression showed a shorter time to recurrence than TERT expression-negative carcinomas (29). However, many of the TERT-expressing tumors harbored TERT promoter mutations, and the prognostic value of TERT mRNA expression independent of the presence of TERT promoter mutations has not been analyzed.

Muzza et al. also demonstrated TERT protein expression in PTCs without TERT promoter mutations (30). However, the number of analyzed cases was limited, and a correlation analysis between presence of TERT expression and clinicopathological findings was not presented (30).

Second, according to The Cancer Genome Atlas (TCGA) data, TERT promoter mutations were not clonal in PTCs (31). The mutant allele frequency varied from 5% to 50% (average 23%) and there were cases with low allelic frequency that may not be detectable by regular Sanger sequencing (31).

In PTCs, little is known about the relationship between TERT mRNA expression and clinicopathological features. Therefore, we measured TERT mRNA expression levels and allelic frequency of the TERT promoter mutations in PTCs in this study, and analyzed their impact on clinicopathological features, especially as a prognostic marker.

Materials and Methods

PTC samples and patient information

We collected 159 PTC samples operated between November 2001 and December 2017 at Nagasaki University Hospital (Nagasaki, Japan) and Kuma Hospital (Kobe, Japan). Clinicopathological data were collected from the patients' medical records. Patient age at operation ranged 14–81 years old (median age: 54 years old, 17.0% male). For staging, the AJCC/TNM staging system (8th edition) was used (32,33). The histological subtypes were 146 classic PTCs (25 were microcarcinomas), 10 follicular variant of PTCs (4 were microcarcinomas), 2 diffuse sclerosing variant of PTCs, and 1 tall cell variant of PTCs.

The study protocol was approved by the institutional review boards of Nagasaki University and Kuma Hospital. Written informed consent was obtained from each patient. Fresh tumor tissue samples were obtained during surgical operations, snap frozen in liquid nitrogen, and stored at −80°C. DNA and total RNA were extracted at the same time using ISOGEN reagent (Nippon Gene) according to the manufacturer's protocol.

Direct DNA sequencing

The mutational status of BRAF (around V600) and the promoter region of TERT were analyzed by direct DNA sequencing (the Sanger method) as described previously (5).

Quantitative reverse transcription polymerase chain reaction

Total RNA was reverse transcribed using a High Capacity RNA-to-cDNA kit (Applied Biosystems). The following polymerase chain reactions (PCRs) were done using SYBR Premix Ex TaqII (TaKaRa Bio) in a Thermal Cycler Dice real-time system (TaKaRa Bio). The cycle threshold value, which was determined using the second derivative method, was used to calculate relative expression. The TERT mRNA levels were normalized using TATA-binding protein (TBP) mRNA expression as a reference. Primer sequences are TERT ex6–7 F, 5′-AGCCACGTCTCTACCTTGAC-3′ and TERT ex7–8 R, 5′-CTCATTCAGGGAGGAGCTCT-3′; TBP ex2 F, 5′-CCTGCCACCTTACGCTCAG-3′ and TBP ex3 R, 5′-TGGTGTTCTGAATAGGCTGTGG-3′.

Droplet digital PCR

Droplet digital PCR (ddPCR) was performed using ddPCR Supermix for Probes (Bio-Rad Laboratories) in a QX100 droplet generator (Bio-Rad Laboratories), a C1000 Touch thermal cycler (Bio-Rad Laboratories), and a QX100 droplet reader (Bio-Rad Laboratories). Probes used for the ddPCR were TERT mut, 5′-/56-FAM/C+CC+C+T+TC+CGG/3IABkFQ/-3′, and TERT wt 228, 5′-/5HEX/C+CC+C+C+TC+CGG/3IABkFQ/-3′ (a base preceded by + is Locked Nucleic Acid). Primers were same as those used in direct DNA sequencing.

It is possible to discriminate between C228T and C250T in the two-dimensional (2D) display using the mentioned two probes at the same time because the TERT mut probe can bind to both C228T and C250T, but the wt probe can bind to only C228. In the case of the C250T mutation, both FAM and HEX signals are detected, while only a FAM signal is detected when the mutation is C228T.

Recurrence as an endpoint

Disease recurrence was defined as a surgically removed and pathologically verified local lesion or regional metastasis/distant metastasis detected by ultrasonography, scintigraphy, or other imaging not earlier than 6 months after initial treatment. The time to recurrence was calculated based on the date of reoperation or on the date of recurrence detection by medical imaging.

Statistical analysis

Univariate Fisher's or Fisher–Freeman–Halton exact test was used for categorical data. Pairwise statistical comparisons of proportions in more than two groups were performed with the COMPPROP macros in SAS. Nonparametric Mann–Whitney or Kruskal–Wallis test followed by Dwass, Steel, Critchlow-Fligner multiple comparison procedure for continuous data was used to compare characteristics in different PTC subgroups.

To analyze recurrence-free survival (RFS), the Kaplan–Meier method and log-rank test were used. Factors affecting RFS were assessed in Cox proportional hazard models. To determine a threshold of the TERT expression level based on the concept that minimal expression compared with relatively higher expression by the TERT promoter mutations could have negligible effect on clinical behavior, we first calculated hazard ratios (HRs) in serial optimal Cox models for each cutoff percentile (mRNA expression-negative plus expression levels below each percentile cases vs. expression levels above the percentile cases) in five percentile increment. We next ran Kaplan–Meier analyses using percentiles that demonstrated relatively higher HR. Then optimal threshold was selected based on the log-rank p-value.

Multivariable logistic regression models were used to identify factors associated with extrathyroidal extension or pT category. Analyses with very small number of outcomes (<5 per cell) or when quasi-complete separation was observed were conducted using Firth's approach to bias-reducing penalized maximum likelihood fit. Nonautomatic model optimization was routinely performed using the Akaike information criteria. Stepwise variable selection was applied to the models amendable to automatic optimization. Once the most appropriate model was determined, the maximum likelihood estimates of the respective parameters and their Wald-type 95% confidence intervals (CIs) were calculated. Statistical assessments were performed using the 3.71 release of SAS Studio for the 9.4M5 version of SAS (SAS Institute) or IBM SPSS Statistics Version 24 software (IBM). Graphs were drawn using GraphPad Prism 6 (GraphPad). All p-values were two-sided and considered significant if p < 0.05.

Results

Mutational status of the TERT promoter region and TERT expression

We first screened 159 PTC samples for mutations of the TERT promoter region (C228T and C250T) by direct DNA sequencing. TERT promoter mutations were found in 20 (12.6%) samples, all of which were C228T, and there was no C250T substitution in the current series. We next examined TERT mRNA expression by real-time quantitative reverse transcription PCR. TERT expression was confirmed in all of the TERT promoter mutation-positive samples. Interestingly, even among 139 mutation-negative samples, 56 (40.3%) showed TERT expression. We then explored the possibility that there are tumors with low allelic frequency of TERT promoter mutations that are not detectable by regular sequencing.

To investigate the presence of the mutations with high sensitivity, we used ddPCR. First, the detection limit of the ddPCR for the two TERT promoter mutations was determined using serial dilutions of the PCR product of the TERT promoter region containing C228T or C250T in the PCR product of the wild-type promoter. In our hands, the detection limit of the mutant allele frequency was ∼0.25% (Supplementary Fig. S1). We then analyzed all of the 56 TERT-expressing samples using ddPCR. We identified the mutation with low allelic frequency in three samples (5.4%), hereafter PTCs A, B, and C (Fig. 1A). According to the 2D display, all harbored a C228T mutation.

FIG. 1.

Low allelic frequencies of the TERT promoter mutation in three samples, PTC A, PTC B, and PTC C. (A) Droplet digital PCR results. Each dot represents a positive droplet of a mutant allele or a wild-type allele. (B) Sanger sequencing chromatograms of indicated samples. The hot spots of the TERT promoter mutation are shown as arrowheads. All of PTC A, PTC B, and PTC C had the C228T mutation. hetMut, both mutant and wild-type control (monoallelic); NTC, nontemplate control; PCR, polymerase chain reaction; PTC, papillary thyroid carcinoma; TERT, telomerase reverse transcriptase; Wt, wild-type control. Color images are available online.

The allelic frequencies of the mutant were 17%, 10%, and 5% in PTCs A, B, and C, respectively (Supplementary Table S1). Since tumor tissues consist not only of tumor cells but also of stromal, endothelial, and blood cells, the allelic frequency of the mutant in the tumor cells was corrected for tumor purity based on the ratio of the TERT promoter mutation to the BRAF^V600E mutation because the BRAF^V600E mutation is considered a clonal monoallelic mutation in all tumor cells according to the data obtained using next-generation sequencing and immunohistochemistry (34,35). However, one should note that there is also some debate about the allelic frequency of the BRAF^V600E mutation. Guerra et al. reported variable BRAF mutant allelic frequencies in PTCs using pyrosequencing (36). For this study, we assumed that all tumor cells were BRAF^V600E positive. After the correction, the allelic frequencies of the TERT promoter mutation in tumor cells were as follows: PTC A, 14%; PTC B, 4%; and PTC C, 3%, which means that 28%, 8%, and 6% of the tumor cells had the TERT promoter mutation in PTCs A, B, and C, respectively (Supplementary Table S1).

We then retrospectively checked the chromatograms of the direct DNA sequencing results of these samples. There were discrete peaks of C228T (Fig. 1B), but it was impossible to confidently discriminate these signals from background signals. The results of direct DNA sequencing, expression analysis, and ddPCR are summarized in Figure 2A–C.

FIG. 2.

Summary of the TERT mutational and expression status in the current series. (A) The flowchart of the TERT mutation/expression screening. (B) Pie chart of the results. (C) Relative TERT expression level in each group classified using the mentioned status. (D) Correlation between the TERT expression level and patient age. The mutation-negative/expression-positive cases are plotted. Solid line represents the linear regression model with 95% confidence intervals indicated by dotted lines. Arrow heads indicate cases with TERT mRNA expression >80th percentile shown in (C).

Taken together, TERT expression was observed in all of the tumors with the TERT promoter mutation as expected, even in the cases with low allelic frequency; however, TERT expression was also detected in 38% of mutation-negative cases, suggesting that there are other mechanisms to upregulate TERT expression.

Relationship between TERT mutational/expression status and clinicopathological features

We analyzed the relationship between the status of the TERT promoter mutation/expression and clinicopathological features. We classified the 159 cases into three groups: the TERT promoter mutation-negative/mRNA expression-negative group (mut−/exp−), the TERT promoter mutation-negative/mRNA expression-positive group (mut−/exp+), and the TERT promoter mutation-positive group (mut+/exp+). As shown in Table 1, tumors with the mutation (mut+/exp+) showed statistically significant differences in age, extrathyroidal extension, stage II/III/IV, compared with the other two groups (mut−/exp− and mut−/exp+) (1 vs. 3 and 2 vs. 3). These findings suggest that the TERT promoter mutation confers aggressive properties to PTCs.

Table 1.

Association Between TERT Mutational/Expression Status and Clinicopathological Features

TERT mutational/expression status Parameter	Whole group Mean ± SD (range) or n (%)	mut−/exp−(1)	mut−/exp+(2)	mut+/exp+(3)	p(1 vs. 2)	p(1 vs. 3)	p(2 vs. 3)
Number of cases	159	83	53	23
Age ± SD (range)	52.0 ± 15.8 (14–81)	49.2 ± 15.7 (16–78)	50.3 ± 15.6 (14–76)	66.3 ± 7.0 (55–81)	ns	<0.001	<0.001
Sex F/M, ratio	132/27 (4.9:1)	68/15 (4.5:1)	47/6 (7.8:1)	17/6 (2.8:1)	ns	ns	ns
Tumor size, mm	21.5 ± 13.0 (0.3–62)	21.2 ± 11.9 (0.3–60)	21.8 ± 15.5 (3–62)	21.9 ± 10.7 (6–45)	ns	ns	ns
pN	111 (69.8%)	59 (71.1%)	37 (69.8%)	15 (65.2%)	ns	ns	ns
M	4 (2.5%)	2 (2.4%)	1 (1.9%)	1 (4.3%)	ns	ns	ns
Extrathyroidal extension	82 (52.6%)	41 (51.3%)	23 (43.4%)	18 (78.3%)	ns	<0.05	<0.05
pT3 and pT4	89 (56.0%)	46 (55.4%)	25 (47.2%)	18 (78.3%)	ns	ns	<0.05
Stage
I	88 (55.3%)	49 (59.0%)	37 (69.8%)	2 (8.7%)	ns	<0.05	<0.05
II	61 (38.4%)	29 (34.9%)	14 (26.4%)	18 (78.3%)	ns	<0.05	<0.05
III	7 (4.4%)	4 (4.8%)	1 (1.9%)	2 (8.7%)	ns	ns	ns
IV	3 (1.9%)	1 (1.2%)	1 (1.9%)	1 (4.3%)	ns	ns	ns
Stages II, III, and IV	71 (44.7%)	34 (41.0%)	16 (30.2%)	21 (91.3%)	ns	<0.05	<0.05
Recurrence^a	17 (11.0%)	4 (4.9%)	6 (11.5%)	7 (31.8%)	ns	<0.05	ns
Mean recurrence time [95% CI, months^a	115.2 [108.6–121.8]	119.5 [114.1–124.8]	104.7 [92.8–116.7]	91.6 [68.3–114.9]	<0.05	<0.001	ns
Total thyroidectomy	89 (56.0%)	48 (57.8%)	26 (49.1%)	15 (65.2%)	ns	ns	ns
LN dissection	147 (92.5%)	76 (91.6%)	49 (92.5%)	22 (95.7%)	ns	ns	ns
BRAF mutation	111 (69.8%)	57 (68.7%)	33 (62.3%)	21 (91.3%)	ns	ns	<0.05

Four cases with distant metastasis and 20 cases that were followed for <6 months were not included.

CI, confidence interval; LN, lymph node; mut−/exp−, TERT promoter mutation-negative/mRNA expression-negative; mut−/exp+, TERT promoter mutation-negative/mRNA expression-positive; mut+/exp+, TERT promoter mutation-positive/mRNA expression-positive; ns, not significant, p ≥ 0.05; SD, standard deviation; TERT, telomerase reverse transcriptase.

RFS was evaluated using the Kaplan–Meier method and Cox proportional hazard models. In this analysis, we excluded four cases that had distant metastasis at the time of operation and additional 20 cases that were followed for <6 months. The survival curves of the three groups separated, and there was a statistically significant trend (Fig. 3A; log-rank trend, p < 0.001).

FIG. 3.

Kaplan–Meier curves of RFS. The vertical tick marks correspond to censored data. (A) The RFS curves of the indicated groups. p-Value of a log-rank trend test is shown. (B, C) Only mutation-negative cases were analyzed. p-Values of log-rank tests are shown. RFS, recurrence-free survival.

The HRs for recurrence of the mut+/exp+ group relative to the mut−/exp− group and the mut−/exp+ group after adjustment for age, sex, tumor size, extrathyroidal extension, and lymph node metastasis were 20.47 ([95% CI 4.54–114.1], p < 0.001) and 5.38 ([95% CI 1.14–30.32], p = 0.046), respectively (Table 2; 1st and 2nd comparisons). In the optimal models, the HRs of the mut+/exp+ group relative to the mut−/exp− group and the mut−/exp+ group were 23.39 ([95% CI 4.49–121.85], p < 0.001) and 6.24 ([95% CI 1.44–27.13], p = 0.015), respectively (Table 2; 1st and 2nd comparisons).

Table 2.

Hazard Ratios of Disease Recurrence

	HR	[95% CI]	p	HR	[95% CI]	p	HR	[95% CI]	p	HR	[95% CI]	p
mut−/exp−	1.00			1.00			1.00			1.00
mut+/exp+	7.53	[2.39–26.69]	0.001	20.25	[4.54–109.26]	<0.001	20.47	[4.54–114.1]	<0.001	23.39	[4.49–121.85]	<0.001
Adjustment					Age, sex			Age, sex, size, Ex, N			Optimal model (age)^a
mut−/exp+	1.00			1.00			1.00			1.00
mut+/exp+	2.34	[0.80–6.99]	0.131	5.07	[1.24–24.03]	0.034	5.38	[1.14–30.32]	0.046	6.24	[1.44–27.13]	0.015
Adjustment:					Age, sex			Age, sex, size, Ex, N			Optimal model (age)^a
mut−/exp−	1.00			1.00			1.00			1.00
mut−/exp+	3.27	[0.98–11.87]	0.067	4.65	[1.31–18.51]	0.026	4.25	[1.15–17.71]	0.041	4.24	[1.13–15.90]	0.032
Adjustment					Age, sex			Age, sex, size, Ex, N			Optimal model (age)^a
mut−/exp− and exp <80th percentile	1.00			1.00			1.00			1.00
mut−/exp >80th percentile	5.72	[1.39–19.23]	0.010	5.09	[1.21–17.81]	0.022	3.36	[0.80–11.75]	0.095	4.34	[1.11–16.94]	0.035
Adjustment:					Age, sex			Age, sex, size, Ex, N			Optimal model (age)^a

In the optimal model, only “age” was used for adjustment.

HR, hazard ratio; mut−/exp >80th percentile, TERT mRNA expression levels >80th percentile of the mut−/exp+ cases; mut−/exp− and exp <80th percentile, TERT promoter mutation-negative/mRNA expression-negative and mRNA expression levels <80th percentile of the mut−/exp+ cases; mut−/exp−, TERT promoter mutation-negative/mRNA expression-negative; mut−/exp+, TERT promoter mutation-negative/mRNA expression-positive; mut+/exp+, TERT promoter mutation-positive/mRNA expression-positive.

In two-group analysis, based on the mutational status (mut−/exp− and mut−/exp+ vs. mut+/exp+) and the expression status (mut−/exp− vs. mut−/exp+ and mut+/exp+), the Kaplan–Meier curves and the HRs are shown in Supplementary Figure S2A and B and Supplementary Table S2, respectively.

Regarding the grouping based on the expression status (regardless of the mutational status), we attempted to set a threshold based on the concept that minimal expression could have a negligible effect on the clinical behavior because many of the mut–/exp+ tumors showed very low expression compared with the mut+/exp+ group (Fig. 2C).

First, we calculated HRs in serial optimal Cox models for each cutoff percentile (mRNA expression-negative plus expression levels below each percentile cases vs. expression levels above the percentile cases) (Supplementary Fig. S3A). We then ran Kaplan–Meier analyses between the 50th and 70th percentiles showing high HRs (Supplementary Fig. S3B). This allowed us to identify the optimal threshold, the 65th percentile (Supplementary Fig. S3B). Based on this, recurrence was the only parameter that was statistically significant (Supplementary Table S3), and the HR adjusted for age, sex, tumor size, extrathyroidal extension, and lymph node metastasis was 4.12 ([95% CI 1.55–10.72], p = 0.005). In the optimal model, the HR was 4.44 ([95% CI 1.71–11.53], p = 0.002) (Supplementary Table S2; 5th comparison). Other results using different adjustments are listed in Supplementary Table S2.

Relationship between TERT expression and clinicopathological features in the mutation-negative cases

Next, we focused on the TERT promoter mutation-negative cases. Interestingly, there were no statistical differences between the mut−/exp− group and the mut−/exp+ group in all clinicopathological parameters except mean recurrence time (Table 1; 1 vs. 2). The Kaplan–Meier curve showed a significant difference (Fig. 3B, p = 0.046), and Cox proportional hazard models also demonstrated statistical significance after adjustment for covariates (Table 2; 3rd comparison).

The HR of the mut−/exp+ group relative to the mut−/exp− group adjusted for age, sex, tumor size, extrathyroidal extension, and lymph node metastasis was 4.25 ([95% CI 1.15–17.71], p = 0.041) (Table 2; 3rd comparison). In the optimal model, the HR was 4.24 ([95% CI 1.13–15.90], p = 0.032) (Table 2; 3rd comparison). Since the unadjusted HR was not significant (p = 0.067), we analyzed the relationship between age and the amount of TERT expression. Surprisingly, TERT expression was inversely correlated with age (Fig. 2D, p = 0.002), a finding that is opposite to the presence of TERT promoter mutations. These results indicate that TERT expression, even without presence of a TERT promoter mutation, has a negative influence on PTC prognosis.

Because many cases of the mut−/exp+ tumors showed very low TERT expression compared with the mut+/exp+ group (Fig. 2C), we also attempted to reclassify all mut− cases into two new groups based on the TERT expression level using the same method already described (Supplementary Fig. S3C). Kaplan–Meier analyses were run between the 75th and 85th percentiles, and the optimal threshold was determined to be the 80th percentile (Supplementary Fig. S3D). Using this threshold, the unadjusted HR was 5.72 ([95% CI 1.39–19.23], p = 0.01) (Table 2; 4th comparison), which was higher than the HR calculated based on the presence or absence of the expression (exp− vs. exp+). In the optimal model, the HR was 4.34 ([95% CI 1.11–16.94], p = 0.035) (Table 2; 4th comparison).

Using this grouping, there were significant differences in age (the group of expression levels >80th percentile was younger), stage, recurrence, recurrence time, and the prevalence of the BRAF mutation (Table 3). Interestingly, all cases with an expression level >80th percentile were stage I and displayed a significantly lower rate of the BRAF mutation (30.0% vs. 69.9%) (Table 3). Even though tumors in which the expression levels were >80th percentile were present in patients with younger age, lower stage, and a lower BRAF mutation rate, their prognosis was worse. The Kaplan–Meier curve also demonstrated a significant difference (Fig. 3C, p = 0.007).

Table 3.

Association Between TERT Expression and Clinicopathological Features in the TERT Promoter Mutation-Negative Cases

TERT expression	Whole group Mean ± SD (range) or n (%)	TERT exp− and exp <80th percentile	TERT exp >80th percentile	Univariate p
Number of cases	136	126	10
Age	49.6 ± 15.6 (14–78)	50.5 ± 15.6 (14–78)	38.8 ± 12.6 (20–52)	<0.05
Sex F/M, ratio	115/21 (5.5:1)	106/20 (5.3:1)	9/1 (9.0:1)	ns
Tumor size, mm	21.4 ± 13.4 (0.3–62)	21.1 ± 13.1 (0.3–62)	24.7 ± 16.9 (4–55)	ns
pN	96 (70.6%)	88 (69.8%)	8 (80.0%)	ns
M	3 (2.2%)	3 (2.4%)	0	ns
Extrathyroidal extension	64 (48.1%)	58 (47.2%)	6 (60.0%)	ns
pT3 and pT4	71 (52.2%)	65 (51.6%)	6 (60.0%)	ns
Stage
I	86 (63.2%)	76 (60.3%)	10 (100.0%)	<0.05
II	43 (31.6%)	43 (34.1%)	0	<0.05
III	5 (3.7%)	5 (4.0%)	0	ns
IV	2 (1.5%)	2 (1.6%)	0	ns
Stages II, III, and IV	50 (36.8%)	50 (39.7%)	0	<0.05
Recurrence^a	10 (7.5%)	7 (5.7%)	3 (30.0%)	<0.05
Mean recurrence time [95% CI], months^a	115.2 [109.4–121.0]	117.6 [112.1–123.0]	86.3 [53.9–118.7]	<0.05
Total thyroidectomy	74 (54.4%)	67 (53.2%)	7 (70.0%)	ns
LN dissection	125 (91.9%)	116 (92.1%)	9 (90.0%)	ns
BRAF mutation	90 (66.2%)	87 (69.0%)	3 (30.0%)	<0.05

Three cases with distant metastasis and 20 cases that were followed for <6 months were not included.

ns, not significant, p ≥ 0.05; TERT exp− and exp <80th percentile, TERT mRNA expression-negative and mRNA expression levels <80th percentile; TERT exp− >80th percentile, TERT mRNA expression levels >80th percentile.

Discussion

First, in this study, we have successfully reconfirmed the strong correlation between the presence of TERT promoter mutations and aggressive clinicopathological features in this PTC series. Second, we have demonstrated that there is a subset of PTCs that express TERT mRNA even in the absence of the TERT promoter mutation (mut−/exp+). In this case, TERT expression conferred a significant negative impact on PTC prognosis, which was, however, not as high as in the cases with a TERT promoter mutation. Since TERT promoter mutations are only observed in elderly patients, TERT expression may be a promising marker in younger patients.

Among TERT mut−/exp+ cases (determined by regular sequencing), there seems to be a small number of tumors with low allelic frequency of the TERT promoter mutation. Hence, in addition to a low allelic frequency of TERT promoter mutations, there are other mechanisms that upregulate TERT expression in the majority of the TERT mut−/exp+ cases. In other types of cancers, it has been reported that amplification, structural variants, alternative splicing, and promoter methylation also upregulate TERT expression (37 –44). Of note, the allelic frequency depends on tumor cell purity and sensitivity of a detection method. It is still unclear whether the low allelic frequency of the TERT promoter mutation has a significant influence on tumor aggressivity because the number of such cases in the current series was too small. Further studies are needed to fully understand the mechanisms of the TERT mRNA upregulation in PTCs.

In the current series, all of the TERT promoter mutation-positive cases were 55 years of age or older, and among the 8 recurrent cases in this age group, 7 had a TERT promoter mutation. Thus, the presence of a TERT promoter mutation is probably a good marker in elderly patients. In contrast, since there was no TERT promoter mutation-positive case in the patients younger than 55 years, this mutation is not useful in younger patients presenting with PTCs. However, in these patients, TERT expression appears to have prognostic value.

In the TERT mut−/exp+ group, there were many tumors with a very low amount of TERT expression compared with the levels of TERT expression in the TERT promoter mutation-positive cases. Such a minimal expression may not have clinical significance; therefore, we attempted to seek the best threshold based on the TERT expression levels. In this study, the highest HR was obtained when the mutation-negative cases were categorized into two groups: cases with expression levels >80th percentile and others. However, it is necessary to analyze a much larger number of cases to determine a proper threshold. In addition, the threshold may be influenced by the detection method. This is a limitation of this study. Moreover, since an active surveillance management approach has been used for low-risk micro-PTCs in Japan, current cases may not be fully representative of the whole PTC spectrum. It should rather be considered as a proof of principle that high TERT expression is associated with risk for recurrence.

The TERT expression levels >80th percentile cases were associated with larger tumor size, more frequent lymph node metastasis and extrathyroidal extension, and more T3 and T4 tumors than those in the other group, yet these differences were not statistically significant (Table 3), which could be due to the very small number of cases (only 10) in the group. For the same reason, it was difficult to compare these cases with the mutation-positive cases. Intriguingly, in the mutation-negative cases, TERT expression levels were inversely correlated with patient age and the frequency of the BRAF mutation. We have reported that tumor size is inversely correlated with the presence of the BRAF mutation in pediatric and adolescent PTCs (45). Taken together, tumors without the BRAF mutation may likely have higher TERT expression levels and more aggressive features in young PTC patients.

In summary, as reported by others, we confirm the association between presence of TERT promoter mutations and aggressive clinicopathological characteristics in PTCs. Moreover, high TERT expression levels were observed in PTCs even in TERT promoter mutation-negative tumors in patients of all ages, and TERT expression was associated with worse prognosis. Since TERT promoter mutations are only found in elderly patients, TERT expression can also be a useful marker, especially in younger patients with PTCs.

Footnotes

Acknowledgment

This study was supported, in part, by Grants-in-Aid for Scientific Research (JSPS KAKENHI) Grant Nos. 26293222, 26293142, and 16K09804.

Author Disclosure Statement

No competing financial interests exist.

Supplementary Material

Supplementary Figure S1

Supplementary Figure S2

Supplementary Figure S3

Supplementary Table S1

Supplementary Table S2

Supplementary Table S3

References

Wiltshire

, Drake

, Uttley

, Balasubramanian

. 2016. Systematic review of trends in the incidence rates of thyroid cancer. Thyroid, 26:1541–1552.

Schlumberger

. 1998. Papillary and follicular thyroid carcinoma. N Engl J Med, 338:297–306.

Ito

, Miyauchi

, Kihara

, Fukushima

, Higashiyama

, Miya

. 2018. Overall survival of papillary thyroid carcinoma patients: a single-institution long-term follow-up of 5897 patients. World J Surg, 42:615–622.

Xing

. 2005. BRAF mutation in thyroid cancer. Endocr Relat Cancer, 12:245–262.

Matsuse

, Yabuta

, Saenko

, Hirokawa

, Nishihara

, Suzuki

, Yamashita

, Miyauchi

, Mitsutake

. 2017. TERT promoter mutations and Ki-67 labeling index as a prognostic marker of papillary thyroid carcinomas: combination of two independent factors. Sci Rep, 7:41752.

Landa

, Ganly

, Chan

, Mitsutake

, Matsuse

, Ibrahimpasic

, Ghossein

, Fagin

. 2013. Frequent somatic TERT promoter mutations in thyroid cancer: higher prevalence in advanced forms of the disease. J Clin Endocrinol Metab, 98:E1562–E1566.

Liu

, Bishop

, Shan

, Pai

, Liu

, Murugan

, Sun

, El-Naggar

, Xing

. 2013. Highly prevalent TERT promoter mutations in aggressive thyroid cancers. Endocr Relat Cancer, 20:603–610.

Alzahrani

, Alsaadi

, Murugan

, Sadiq

. 2016. TERT promoter mutations in thyroid cancer. Horm Cancer, 7:165–177.

Vinagre

, Almeida

, Populo

, Batista

, Lyra

, Pinto

, Coelho

, Celestino

, Prazeres

, Lima

, Melo

, da Rocha

, Preto

, Castro

, Pardal

, Lopes

, Santos

, Reis

, Cameselle-Teijeiro

, Sobrinho-Simoes

, Lima

, Maximo

, Soares

. 2013. Frequency of TERT promoter mutations in human cancers. Nat Commun, 4:2185.

10.

Huang

, Hodis

, Xu

, Kryukov

, Chin

, Garraway

. 2013. Highly recurrent TERT promoter mutations in human melanoma. Science, 339:957–959.

11.

Horn

, Figl

, Rachakonda

, Fischer

, Sucker

, Gast

, Kadel

, Moll

, Nagore

, Hemminki

, Schadendorf

, Kumar

. 2013. TERT promoter mutations in familial and sporadic melanoma. Science, 339:959–961.

12.

Koh

, Khattar

, Leow

, Liu

, Muller

, Ang

, Li

, Franzoso

, Li

, Guccione

, Tergaonkar

. 2015. Telomerase regulates MYC-driven oncogenesis independent of its reverse transcriptase activity. J Clin Invest, 125:2109–2122.

13.

Low

, Tergaonkar

. 2013. Telomerase: central regulator of all of the hallmarks of cancer. Trends Biochem Sci, 38:426–434.

14.

, Tergaonkar

. 2014. Noncanonical functions of telomerase: implications in telomerase-targeted cancer therapies. Cancer Res, 74:1639–1644.

15.

Xing

, Liu

, Murugan

, Zhu

, Zeiger

, Pai

, Bishop

. 2014. BRAF V600E and TERT promoter mutations cooperatively identify the most aggressive papillary thyroid cancer with highest recurrence. J Clin Oncol, 32:2718–2726.

16.

Qasem

, Murugan

, Al-Hindi

, Xing

, Almohanna

, Alswailem

, Alzahrani

. 2015. TERT promoter mutations in thyroid cancer: a report from a Middle Eastern population. Endocr Relat Cancer, 22:901–908.

17.

George

, Henderson

, Williams

, Roberts

, Hei

, Lai

, Clayman

. 2015. Association of TERT promoter mutation, but not BRAF mutation, with increased mortality in PTC. J Clin Endocrinol Metab, 100:E1550–E1559.

18.

Bullock

, Ren

, O'Neill

, Gill

, Aniss

, Sywak

, Sidhu

, Delbridge

, Learoyd

, de Vathaire

, Robinson

, Clifton-Bligh

. 2016. TERT promoter mutations are a major indicator of recurrence and death due to papillary thyroid carcinomas. Clin Endocrinol (Oxf), 85:283–290.

19.

Lee

, Hwang

, Choi

, Han

, Kim

, Jang

, Kim

, Yang

. 2016. Prognostic significance of TERT promoter mutations in papillary thyroid carcinomas in a BRAF(V600E) mutation-prevalent population. Thyroid, 26:901–910.

20.

, Siraj

, Divya

, Kong

, Parvathareddy

, Al-Rasheed

, Al-Obaisi

KAS

, Victoria

, Al-Sobhi

, Al-Dawish

, Al-Dayel

, Al-Kuraya

. 2018. Telomerase reverse transcriptase mutations are independent predictor of disease-free survival in Middle Eastern papillary thyroid cancer. Int J Cancer, 142:2028–2039.

21.

Melo

, da Rocha

, Vinagre

, Batista

, Peixoto

, Tavares

, Celestino

, Almeida

, Salgado

, Eloy

, Castro

, Prazeres

, Lima

, Amaro

, Lobo

, Martins

, Moura

, Cavaco

, Leite

, Cameselle-Teijeiro

, Carrilho

, Carvalheiro

, Maximo

, Sobrinho-Simoes

, Soares

. 2014. TERT promoter mutations are a major indicator of poor outcome in differentiated thyroid carcinomas. J Clin Endocrinol Metab, 99:E754–E765.

22.

Liu

, Qu

, Liu

, Sheng

, Shi

, Zhu

, Murugan

, Guan

, Yu

, Wang

, Sun

, Shan

, Teng

, Xing

. 2014. TERT promoter mutations and their association with BRAF V600E mutation and aggressive clinicopathological characteristics of thyroid cancer. J Clin Endocrinol Metab, 99:E1130–E1136.

23.

Jin

, Chen

, Dong

, Cai

, Zhang

, Zhou

, Zeng

, Yang

, Pan

, Liu

, Wu

, Xing

, Zhang

, Wang

. 2016. BRAF and TERT promoter mutations in the aggressiveness of papillary thyroid carcinoma: a study of 653 patients. Oncotarget, 7:18346–18355.

24.

Sun

, Zhang

, Lu

, Gao

, Ren

, Teng

, Duan

, Lin

, Li

, Zhang

, Liang

. 2016. BRAF V600E and TERT promoter mutations in papillary thyroid carcinoma in chinese patients. PLoS One, 11:e0153319.

25.

Kim

, Kim

, Ahn

, Kim

, Ki

, Oh

, Kim

, Yun

, Park

, Choe

, Kim

, Chung

. 2016. TERT promoter mutations and long-term survival in patients with thyroid cancer. Endocr Relat Cancer, 23:813–823.

26.

Liu

, Bishop

, Zhu

, Zhang

, Ladenson

, Xing

. 2016. Mortality risk stratification by combining BRAF V600E and TERT promoter mutations in papillary thyroid cancer: genetic duet of BRAF and TERT promoter mutations in thyroid cancer mortality. JAMA Oncol, 3:202–208.

27.

Penna

, Pestana

, Cameselle

, Momesso

, de Andrade

, Vidal

APA

, Araujo Junior

, Melo

, Fernandes

, Corbo

, Vaisman

, Sobrinho-Simoes

, Soares

, Vaisman

. 2018. TERTp mutation is associated with a shorter progression free survival in patients with aggressive histology subtypes of follicular-cell derived thyroid carcinoma. Endocrine, 61:489–498.

28.

Oishi

, Kondo

, Ebina

, Sato

, Akaishi

, Hino

, Yamamoto

, Mochizuki

, Nakazawa

, Yokomichi

, Ito

, Ishikawa

, Katoh

. 2017. Molecular alterations of coexisting thyroid papillary carcinoma and anaplastic carcinoma: identification of TERT mutation as an independent risk factor for transformation. Mod Pathol, 30:1527–1537.

29.

Paulsson

, Mu

, Shabo

, Wang

, Zedenius

, Larsson

, Juhlin

. 2018. TERT aberrancies: a screening tool for malignancy in follicular thyroid tumours. Endocr Relat Cancer, 25:723–733.

30.

Muzza

, Colombo

, Rossi

, Tosi

, Cirello

, Perrino

, De Leo

, Magnani

, Pignatti

, Vigo

, Simoni

, Bulfamante

, Vicentini

, Fugazzola

. 2015. Telomerase in differentiated thyroid cancer: promoter mutations, expression and localization. Mol Cell Endocrinol, 399:288–295.

31.

Landa

, Ibrahimpasic

, Boucai

, Sinha

, Knauf

, Shah

, Dogan

, Ricarte-Filho

, Krishnamoorthy

, Xu

, Schultz

, Berger

, Sander

, Taylor

, Ghossein

, Ganly

, Fagin

. 2016. Genomic and transcriptomic hallmarks of poorly differentiated and anaplastic thyroid cancers. J Clin Invest, 126:1052–1066.

32.

Amin

, Edge

, Greene

, Byrd

, Brookland

, Washington

, Gershenwald

, Compton

, Hess

, Sullivan

, Jessup

, Brierley

, Gaspar

, Schilsky

, Balch

, Winchester

, Asare

, Madera

, Gress

, Meyer

. 2017. AJCC Cancer Staging Manual, 8th edition. Springer International Publishing, New York.

33.

Tuttle

, Haugen

, Perrier

. 2017. Updated American joint committee on cancer/tumor-node-metastasis staging system for differentiated and anaplastic thyroid cancer (Eighth Edition): what changed and why?. Thyroid, 27:751–756.

34.

Ghossein

, Katabi

, Fagin

. 2013. Immunohistochemical detection of mutated BRAF V600E supports the clonal origin of BRAF-induced thyroid cancers along the spectrum of disease progression. J Clin Endocrinol Metab, 98:E1414–E1421.

35.

Nikiforova

, Wald

, Roy

, Durso

, Nikiforov

. 2013. Targeted next-generation sequencing panel (ThyroSeq) for detection of mutations in thyroid cancer. J Clin Endocrinol Metab, 98:E1852–E1860.

36.

Guerra

, Fugazzola

, Marotta

, Cirillo

, Rossi

, Cirello

, Forno

, Moccia

, Budillon

, Vitale

. 2012. A high percentage of BRAFV600E alleles in papillary thyroid carcinoma predicts a poorer outcome. J Clin Endocrinol Metab, 97:2333–2340.

37.

Barthel

, Wei

, Tang

, Martinez-Ledesma

, Hu

, Amin

, Akdemir

, Seth

, Song

, Wang

, Lichtenberg

, Hu

, Zhang

, Zheng

, Verhaak

. 2017. Systematic analysis of telomere length and somatic alterations in 31 cancer types. Nat Genet, 49:349–357.

38.

Castelo-Branco

, Choufani

, Mack

, Gallagher

, Zhang

, Lipman

, Zhukova

, Walker

, Martin

, Merino

, Wasserman

, Elizabeth

, Alon

, Zhang

, Hovestadt

, Kool

, Jones

, Zadeh

, Croul

, Hawkins

, Hitzler

, Wang

, Baruchel

, Dirks

, Malkin

, Pfister

, Taylor

, Weksberg

, Tabori

. 2013. Methylation of the TERT promoter and risk stratification of childhood brain tumours: an integrative genomic and molecular study. Lancet Oncol, 14:534–542.

39.

Fan

, Lee

, Wu

, Easton

, Yergeau

, Dummer

, Vogel

, Kirkwood

, Barnhill

, Pappo

, Bahrami

. 2016. Telomerase expression by aberrant methylation of the TERT promoter in melanoma arising in giant congenital nevi. J Invest Dermatol, 136:339–342.

40.

Seynnaeve

, Lee

, Borah

, Park

, Pappo

, Kirkwood

, Bahrami

. 2017. Genetic and epigenetic alterations of TERT are associated with inferior outcome in adolescent and young adult patients with melanoma. Sci Rep, 7:45704.

41.

Peifer

, Hertwig

, Roels

, Dreidax

, Gartlgruber

, Menon

, Kramer

, Roncaioli

, Sand

, Heuckmann

, Ikram

, Schmidt

, Ackermann

, Engesser

, Kahlert

, Vogel

, Altmuller

, Nurnberg

, Thierry-Mieg

, Mariappan

, Heynck

, Mariotti

, Henrich

, Gloeckner

, Bosco

, Leuschner

, Schweiger

, Savelyeva

, Watkins

, Shao

, Bell

, Hofer

, Achter

, Lang

, Theissen

, Volland

, Saadati

, Eggert

, de Wilde

, Berthold

, Peng

, Zhao

, Shi

, Ortmann

, Buttner

, Perner

, Hero

, Schramm

, Schulte

, Herrmann

, O'Sullivan

, Westermann

, Thomas

, Fischer

. 2015. Telomerase activation by genomic rearrangements in high-risk neuroblastoma. Nature, 526:700–704.

42.

Valentijn

, Koster

, Zwijnenburg

, Hasselt

, van Sluis

, Volckmann

, van Noesel

, George

, Tytgat

, Molenaar

, Versteeg

. 2015. TERT rearrangements are frequent in neuroblastoma and identify aggressive tumors. Nat Genet, 47:1411–1414.

43.

Davis

, Ricketts

, Wang

, Yang

, Cherniack

, Shen

, Buhay

, Kang

, Kim

, Fahey

, Hacker

, Bhanot

, Gordenin

, Chu

, Gunaratne

, Biehl

, Seth

, Kaipparettu

, Bristow

, Donehower

, Wallen

, Smith

, Tickoo

, Tamboli

, Reuter

, Schmidt

, Hsieh

, Choueiri

, Hakimi

, The Cancer Genome Atlas Research

, Chin

, Meyerson

, Kucherlapati

, Park

, Robertson

, Laird

, Henske

, Kwiatkowski

, Park

, Morgan

, Shuch

, Muzny

, Wheeler

, Linehan

, Gibbs

, Rathmell

, Creighton

. 2014. The somatic genomic landscape of chromophobe renal cell carcinoma. Cancer Cell, 26:319–330.

44.

Wong

, Wright

, Shay

. 2014. Alternative splicing regulation of telomerase: a new paradigm?. Trends Genet, 30:430–438.

45.

Mitsutake

, Fukushima

, Matsuse

, Rogounovitch

, Saenko

, Uchino

, Ito

, Suzuki

, Yamashita

. 2015. BRAF(V600E) mutation is highly prevalent in thyroid carcinomas in the young population in Fukushima: a different oncogenic profile from chernobyl. Sci Rep, 5:16976.