TERT Expression and Clinical Outcome in Pulmonary Carcinoids

Abstract

Purpose: The clinical course of pulmonary carcinoids ranges from indolent to fatal disease, suggesting that specific molecular alterations drive progression toward the fully malignant state. A similar spectrum of clinical phenotypes occurs in pediatric neuroblastoma, in which activation of telomerase reverse transcriptase (TERT) is decisive in determining the course of disease. We therefore investigated whether TERT expression defines the clinical fate of patients with pulmonary carcinoid.

Methods: TERT expression was examined by RNA sequencing in a test cohort and a validation cohort of pulmonary carcinoids (n = 88 and n = 105, respectively). A natural TERT expression cutoff was determined in the test cohort on the basis of the distribution of TERT expression, and its prognostic value was assessed by Kaplan-Meier survival estimates and multivariable analyses. Telomerase activity was validated by telomere repeat amplification protocol assay.

Results: Similar to neuroblastoma, TERT expression exhibited a bimodal distribution in pulmonary carcinoids, separating tumors into TERT-high and TERT-low subgroups. A natural TERT cutoff discriminated unfavorable from favorable clinical courses with high accuracy both in the test cohort (5-year overall survival [OS], 0.547 ± 0.132 v 1.0; P < .001) and the validation cohort (5-year OS, 0.788 ± 0.063 v 0.913 ± 0.048; P < .001). In line with these findings, telomerase activity was largely absent in TERT-low tumors, whereas it was readily detectable in TERT-high carcinoids. In multivariable analysis considering TERT expression, histology (typical v atypical carcinoid), and stage (≤IIA v ≥IIB), high TERT expression was an independent prognostic marker for poor survival, with a hazard ratio of 5.243 (95% CI, 1.943 to 14.148; P = .001).

Conclusion: Our data demonstrate that high TERT expression defines clinically aggressive pulmonary carcinoids with fatal outcome, similar to neuroblastoma, indicating that activation of TERT may be a defining feature of lethal cancers.

FIG 1.

TERT expression and telomerase activity in pulmonary carcinoids. (A) Expression of TERT in neuroblastoma lacking TMM (NB:TMM–), ALT-positive neuroblastoma (NB:ALT), telomerase-positive neuroblastoma (NB:TERT), pulmonary carcinoids (CA), SCLC, and LCNEC sample. Expression levels are given as TERT expression score derived from RNA-seq data. Boxplots show the median, and first and third quartiles (boxes), with whiskers indicating the minimum and maximum of the data within 1.5× the IQR. (B) To determine a natural cutoff that discriminates TERT-high and TERT-low expression in pulmonary carcinoids, a mixture model with two components was applied. The distribution of TERT expression values in the test cohort is shown by the histogram (left axis), while curves indicate normal distributions fitted to tumors with low and high TERT expression using a mixture model (right axis). The threshold at a TERT expression score of 8.17 was defined as the lowest expression value having a posterior probability ≥95% to fall within the distribution on the right (ie, TERT-high cases), thereby separating TERT-high (>8.17) and TERT-low (≤8.17) cases. (C) Telomerase enzymatic activity was determined in TERT-high and TERT-low pulmonary carcinoid samples by TRAP enzyme-linked immunosorbent assay. ALT, alternative lengthening of telomeres; CA, pulmonary carcinoid; LCNEC, large-cell neuroendocrine carcinoma; NB, neuroblastoma; SCLC, small cell lung cancer; sFPKM, significant Fragments Per Kilobase of transcript per Million mapped reads; TERT, telomerase reverse transcriptase; TMM, telomere maintenance mechanisms; TRAP, telomeric repeat amplification protocol.

FIG 2.

Kaplan-Meier plots of OS in patients with pulmonary carcinoids according to TERT expression. OS of patients was assessed in subgroups defined by TERT-high (TERT expression score >8.17) and TERT-low (TERT expression score ≤8.17) expression in (A) the test cohort, (B) the validation cohort, (C) the cohort of patients with typical carcinoids, (D) the cohort of patients with atypical carcinoids, (E) the cohort of patients with stage ≤IIA, and (F) the cohort of patients with stage ≥IIB. Censored data are indicated by tick marks. OS, overall survival; TERT, telomerase reverse transcriptase.

FIG 3.

DNA methylation at CpG site cg116250005 and gene expression patterns at the TERT locus in pulmonary carcinoids. (A) DNA methylation at CpG site cg116250005 as examined by DNA methylation arrays was compared between cases with high (TERT expression score >8.17) and low (TERT expression score ≤8.17) TERT expression, and (B) correlation of DNA methylation and TERT expression was determined. (C) The average expression difference of genes in proximity to the TERT locus between TERT-high and TERT-low subgroups was evaluated (n = 118, TERT-high = 39, TERT-low = 79). Average gene expression occurring in TERT-high tumors was compared with that of TERT-low tumors set as baseline. sFPKM, significant Fragments Per Kilobase of transcript per Million mapped reads; TERT, telomerase reverse transcriptase.