The stabilization of telomere lengths by telomerase activation is an important step in carcinogenesis and cell immortalization. Human telomerase reverse transcriptase (hTERT) is the catalytic subunit of this enzyme. The objective of this study was to evaluate the use of real-time reverse transcriptase-polymerase chain reaction (RT-PCR) analysis for the quantification of hTERT in tumor and nontumorous tissue samples.
Matched samples of tumor and adjacent nontumorous mucosa samples from 57 patients with completely resected colorectal carcinoma (International Union Against Cancer Stage I–IV) who underwent complete resection (R0) were quantified for hTERT mRNA expression using real-time RT-PCR. The expression levels were correlated with histopathologic findings and with survival. The median follow-up was 76 months.
hTERT mRNA was expressed in all tumor samples and in all samples of adjacent mucosa. In 12 patients (21%), there was higher hTERT expression in tumor samples compared with nontumorous samples. Compared with tumor samples, the expression of hTERT in samples of nontumorous mucosa decreased with age (P = 0.06). hTERT mRNA expression in both tumor tissue and adjacent mucosa was correlated significantly with the histologic grade of colorectal carcinoma (P < 0.04 and P < 0.05, respectively). Patients with hTERT expression in tumor tissue in relation to the adjacent mucosa of > 0.57 had a significantly poorer overall survival compared with patients with lower hTERT ratios (P < 0.02). In addition to the established prognostic factor lymphatic vessel invasion, the hTERT ratio proved to be of independent prognostic value (P < 0.05).
Telomeres form the ends of eukaryotic chromosomes, preventing irregular chromosomal recombination and loss of genetic information. Telomeres of human somatic cells consist of 500–3000 repeats of a simple 5′-(TTAGGG)-3′ nucleotide sequence.1 They progressively shorten with each successive cell cycle by replication dependent loss of DNA termini, resulting in chromosomal instability.2 Reaching a critical telomere length, somatic cells eventually succumb to crisis, a period of limited proliferative capacity (mitotic clock) and widespread cellular death.3
Telomerase is a reverse transcriptase that compensates for telomeric DNA losses during cell replication by adding telomeric repeats to chromosome ends.4 It is reported that telomerase activity is essential for stabilizing telomere length and may be a critical step in cell immortalization and carcinogenesis.5 Some studies demonstrated increased telomerase activity in colorectal carcinoma tissue6, 7 and suggested that it has prognostic value for patients with colorectal carcinoma.8
The human telomerase complex consists of human telomerase-associated RNA, which provides the template for telomeric repeat synthesis,9 and human telomerase reverse transcriptase (hTERT), which represents the catalytic subunit of the complex.10 In several tumors, hTERT has been identified as a rate-limiting factor for telomerase activity.11–13 It has been reported that hTERT was up-regulated in tumor cells and during cell immortalization.14 It also was shown that the introduction of hTERT cDNA into telomerase negative cells reconstituted telomerase activity15, 16 and extended the life span of these otherwise mortal cells.17 Moreover, the inhibition of hTERT led to telomere loss and limited the growth of human tumor cell lines in vitro and their tumorigenetic capacity in vivo.18
Unlike many other molecular markers that have been investigated for their prognostic relevance, hTERT expression is not just an epiphenomenon that is observed in colorectal carcinoma; it appears to contribute to human oncogenesis for two reasons: First, proliferating cells, like tumor cells, need to compensate for replicative telomere losses by hTERT expression and consecutive activation of telomerase to overcome crisis and preserve their ability to proliferate.19 Second, Hahn et al. identified hTERT expression as one of three fundamental genetic changes for human tumorigenesis. Those authors showed that the expression of hTERT, together with the two oncogenes large-T and H-ras, resulted in direct malignant transformation of normal human epithelial and fibroblast cells.20
In the current study, we quantified hTERT-encoding mRNA by using real-time polymerase chain reaction (PCR) analysis in tumor tissue and adjacent nontumorous mucosa from 57 patients with colorectal carcinoma who underwent complete resection (R0) and compared the expression levels in corresponding tissue samples. We report on the correlation of hTERT with age, histologic grade, and tumor stage and on the prognostic potential of the ratio of hTERT expression in tumor tissue in relation to adjacent nontumorous mucosa.
MATERIALS AND METHODS
Patient Group and Tumor Specimens
Our study group consisted of 57 patients who underwent complete resection (R0) for colorectal carcinoma between 1993 and 1996. None of the patients received neoadjuvant chemotherapy or radiation therapy. Twenty-seven patients (47%) were female, and 30 patients (53%) were male. The mean patient age (± standard deviation) was 64.6 ± 13.5 years. Of the 57 adenocarcinomas, 27 tumors (47%) were located in the colon, and 30 tumors (53%) were located in the rectum. The resection procedures for rectal carcinomas included 10 abdominoperineal resections and 20 anterior resections. Of the patients with colon carcinomas, 1 patient underwent anterior resection, 6 patients underwent sigmoid resections, 2 patients underwent left hemicolectomies, 15 patients underwent right hemicolectomies, and 3 patients underwent subtotal colectomies. The single patient with Stage IV disease underwent primary resection for stenosing rectal carcinoma and resectable solitary lung metastasis. The TNM classification and the absence of residual tumor after resection (R0 resection) were classified according to the International Union Against Cancer (UICC) classification system (Table 1).21 Two tumors (3%) were well differentiated (Grade 1), 30 tumors (53%) were moderately differentiated (Grade 2), and 25 tumors (44%) were poorly differentiated (Grade 3). Lymphatic vessel invasion was present in 14 patients (25%). From the resected specimen of each patient, matched samples from the tumor tissue and the adjacent mucosa were obtained immediately after resection. All tissue samples were shock frozen in liquid nitrogen within 1 hour of resection and stored at − 80 °C until use.
Table 1. Correlation of Human Telomerase Reverse Transcriptase mRNA Expression in Tumor Tissue and Adjacent Mucosa with Histopathologic Parameters for 57 Patients with Colorectal Carcinoma who Underwent Complete Resection (RO)
Follow-up was carried out at periodic intervals according to a standardized protocol and was complete as of April 15, 2001, for all 57 patients. One patient (2%) was lost for follow-up after 54 months. The median follow-up was 75.5 months (range, 52–87 months). All 16 patients with Stage I disease and 18 of 20 patients with Stage II disease did not receive any adjuvant therapy. Two patients with Stage II rectal carcinoma received adjuvant radiochemotherapy (5-fluorouracil [5-FU]/folic acid, 50 grays [Gy]; one patient in combination with intraoperative radiation with 15 Gy). Among the patients with Stage III disease, eight patients received adjuvant chemotherapy (5-FU/folic acid), four patients received adjuvant radiochemotherapy (5-FU/folic acid, 50 Gy; in combination with intraoperative radiation with 15 Gy in two patients), and eight patients did not receive any adjuvant therapy because of a reduced state of health or patient refusal. The single patient with Stage IV disease refused adjuvant therapy.
Nineteen patients (33%) developed recurrent disease, with distant metastases in 12 patients (10 patients with liver metastases, 1 patient with lung metastases, and 1 patient with brain metastases) and local recurrence in 7 patients (12%). Seventeen of 19 patients died of recurrent disease during follow-up. Three patients (5%) died of postoperative complications within 3 months after surgery, and 8 patients (14%) who did not develop tumor recurrences died of causes unrelated to their disease during follow-up. Both groups were considered censored by statistical survival analysis.
Total RNA was extracted from 20 16-μm cryostat sections from all samples, corresponding to 20–25 mg of tissue, using the High Pure RNA Tissue Kit (Roche Diagnostics, Mannheim, Germany). Before and after these 20 sections, a 7-μm cryostat section from each tissue sample was stained with hematoxylin and eosin for histopathologic analysis. Two authors (M.W. and H.N.) microscopically identified more than 80% of the tumor samples as colorectal carcinoma and ruled out any contamination of the adjacent mucosa samples with carcinoma cells. A 1:100 diluted aliquot of the extracted RNA yield was quantified spectrophotometrically at 260 nm and 280 nm wave length for RNA concentration and purity. Integrity of the extracted RNA was determined by electrophoresis through agarose gels with ethidium bromide and visualization of typical 18S and 28S RNA bands under ultraviolet light. In addition, the housekeeping gene PBGD showed adequate reverse transcriptase-PCR (RT-PCR) amplification in all samples.
Quantification of hTERT Expression
Kinetic PCR quantification of hTERT-encoding mRNA was performed in a real-time, one-step RT-PCR using the LightCycler TeloTAGGG hTERT Quantification Kit (Roche Diagnostics) according to the manufacturer's instructions (Fig. 1). For each normal mucosa and tumor sample, 300 ng of RNA were analyzed for both hTERT and porphobilinogen deaminase (PBGD) in separate RT-PCRs. A 198-base pair fragment of the generated hTERT cDNA was amplified with specific primers. HTERT mRNA expression in unknown samples was calculated on a standard curve that was constructed from standards supplied with the kit. For quantification, hTERT values were corrected against the housekeeping gene PBGD and are expressed as the ratio of copy numbers of hTERT mRNA to copy numbers of PBGD mRNA. To monitor the reaction, total RNA from two cell lines expressing hTERT (control RNA as supplied with the kit and the colon carcinoma cell line HT-29) were included as positive controls in each run and performed as expected with an interrun variation coefficient of 15%. For appropriate negative controls, the RNA template was replaced with nuclease free water without detection of hTERT mRNA copies in all samples.
Statistical analysis was performed using the SPSS software package (SPSS, Inc., Chicago, IL). Differences in hTERT mRNA expression between matched tissue samples were determined by the Wilcoxon test for matched pairs. Differences in hTERT levels among various groups of patients discriminating for histopathologic parameters were analyzed by the Kruskall–Wallis test and the Mann–Whitney two-sample test. All tests were performed at a significance level of P ≤ 0.05. Group-oriented curves for overall survival were calculated according to the Kaplan–Meier model.22 To determine the relative prognostic impact of hTERT expression compared with established prognostic factors, overall survival was analyzed according to a Cox proportional hazards model.23 For univariate and multivariate Cox regression analyses, continuous variables were recoded to binary variables. The classification and regression trees (CART) technique was used to determine the optimal cut-off values for hTERT expression in tumor tissue and for the hTERT ratio of our study group.24
hTERT-encoding mRNA was found in all 57 colorectal carcinoma tissue samples and in all 57 nontumorous adjacent mucosa samples. The median hTERT expression levels (the ratio of copy numbers of hTERT mRNA to copy numbers of PBGD mRNA) were 23.2 (range, 4.6–214.3) in tumor tissues and 41.4 (range, 12.5–209.9) in adjacent mucosal tissues (Fig. 2). Patient-by-patient comparisons of matched tissue samples showed a significant difference in hTERT expression in tumor tissue and adjacent mucosa (P < 0.001). Higher hTERT mRNA levels in tumor tissue compared with adjacent mucosa were found in 12 patients (21.1%), whereas 45 patients (78.9%) expressed less hTERT mRNA in tumor tissue compared with the matched adjacent mucosa. Compared with the 30 male patients (53%), the 27 female patients (47%) had significantly higher hTERT mRNA levels in tumor tissue samples (P < 0.02) but not in the adjacent mucosa samples (Table 1).
In the adjacent mucosa samples, hTERT mRNA levels decreased with increasing age, approaching statistical significance (r = − 0.248; P = 0.06). There was no correlation between hTERT expression in colorectal carcinoma tissue and age (data not shown). To adjust this age dependent variation of hTERT values and to outline the individual differences between hTERT expression levels in tumor tissue and in adjacent mucosa, the ratio of tumor tissue to adjacent mucosa for hTERT expression levels was calculated for each patient. This hTERT ratio showed a median of 0.5 (range, 0.1–3.3).
Correlation with Histopathologic Parameters
There was a significant positive correlation between hTERT mRNA expression in tumor tissue and increasing histologic grade of colorectal carcinoma (P < 0.04; Table 1). Well-differentiated (Grade 1) and moderately differentiated (Grade 2) colorectal carcinomas (n = 32 tumors) expressed significantly lower hTERT levels compared with poorly differentiated tumors (Grade 3; n = 25 tumors; P < 0.04). Moreover, hTERT mRNA expression levels in the adjacent nontumorous mucosa samples also were correlated significantly with the histologic tumor grade (P < 0.05). The 32 patients with well-differentiated and moderately differentiated tumors expressed significantly lower hTERT mRNA levels in adjacent mucosa compared with the 25 patients with poorly differentiated tumors (P < 0.02).
Our data revealed a positive correlation between hTERT mRNA levels in tumor tissues and increasing depth of invasion (pT classification) of the primary tumors that, nonetheless, was not statistically significant (Table 1). The expression of hTERT was not correlated with tumor site, disease stage, lymph node involvement (pN), or lymphatic vessel invasion in tumor tissue or in the adjacent mucosa.
For hTERT mRNA expression in tumor tissue, an optimal cut-off value of 37 was determined for our study group using the CART technique (P = 0.04). The Kaplan–Meier survival curve in Figure 3A illustrates the increased hazard rate for 13 patients (23%) who had hTERT expression levels > 37, with a 5-year survival rate of 44.9% ± 14.1%, compared with 44 patients (77%) who had hTERT levels ≤ 37 and a 5-year survival rate of 72.1% ± 7.7% (P = 0.05).
For the ratio of hTERT expression in tumor tissue in relation to adjacent mucosa, an optimal cut-off value of 0.57 was determined for our study group using the CART technique (P = 0.01). Twenty-six patients (46%) with hTERT ratios > 0.57 had significantly poorer overall survival, with a 5-year survival rate of 56.5% ± 10.3% compared with a 5-year survival rate of 82.1% ± 7.3% for 31 patients (54%) with hTERT ratios ≤ 0.57 (P < 0.02) (Fig. 3B).
Univariate and Multivariate Cox Regression Analyses
In univariate Cox regression analysis, depth of invasion (pT), lymph node status (pN), lymphatic vessel invasion, hTERT ratio, and hTERT expression in tumor tissue were correlated significantly with overall survival (Table 2). Gender, age, tumor site, and histologic grade showed no prognostic impact on overall survival.
Table 2. Relative Risk of Death as Assessed by Univariate and Multivariate Cox Regression Analysis for 57 Patients with Colorectal Carcinoma who Underwent Complete Resection (R0)
In multivariate analysis, the hTERT ratio proved to be an independent prognostic factor for overall survival (P < 0.05; Table 2). The relative risk of death was 2.9 times higher in 26 patients (46%) who had hTERT ratios > 0.57 compared with 31 patients (54%) who had hTERT ratios ≤ 0.57 (confidence interval, 1.0–8.4). In addition, it was shown that lymphatic vessel invasion was of independent prognostic value for overall survival, with a relative risk of 5.6 and a confidence interval of 2.1–15.0 (P < 0.002). Forty-three patients (75%) without lymphatic vessel invasion had significantly better overall survival, with a 5-year survival rate of 81.5% ± 6.31%, compared with a 5-year survival rate of 20.5% ± 12.0% for 14 patients (25%) with lymphatic vessel invasion.
The current study on hTERT mRNA quantification by real-time RT-PCR analysis demonstrates the prognostic potential of the telomerase subunit hTERT in patients with colorectal carcinoma. Our data reveal that the hTERT ratio, which reflects differences in expression levels between tumor tissue and nontumorous tissue, is correlated significantly with overall survival. To date, all of the available studies on hTERT expression in patients with colorectal carcinoma have failed to relate hTERT mRNA expression to prognosis.25–27 Poremba et al. are the only authors who have reported on the prognostic significance of quantitative hTERT mRNA expression, as shown in a study of 67 patients with neuroblastoma.28 In studies that used qualitative methods for detecting hTERT mRNA, results on the prognostic relevance of hTERT expression in patients with sporadic breast tumors and nonsmall cell lung carcinoma were inconsistent.29, 30 Further validation of hTERT as a new prognostic marker, including the statistically calculated cut-off values, may have implications for improved tumor staging with possible advantages over established prognostic factors, like lymphatic vessel invasion,31 or parameters with potential prognostic impact, like histologic grade. First, endoscopically harvested tumor tissue samples are evaluable for both hTERT expression and histologic grade but not for evaluation of lymphatic vessel invasion. Second, unlike both histologic grade and lymphatic vessel invasion, hTERT measurement also is feasible in normal mucosa, in adenomas, and in all stages of multistep carcinogenesis with the potential for observing changes throughout this process.
The detection of hTERT mRNA in all colorectal tumor samples and their adjacent mucosa samples is in line with earlier studies using both real-time PCR32 and conventional PCR.25–27 In situ hybridization of normal colorectal mucosa confirmed that hTERT mRNA was expressed by epithelial cells within the proliferative zone of the crypts.33
In our study, 79% of the histologically normal adjacent mucosa samples expressed more hTERT mRNA compared with the corresponding colorectal carcinoma samples. Using semiquantitative, conventional RT-PCR, Rohde et al. also observed equal or more hTERT mRNA in normal renal tissue compared with corresponding renal carcinoma tissue, although telomerase activity was detected in the tumors only and not in the normal specimens.34 Other studies found that hTERT was expressed in nontumorous colorectal, renal, and ovarian tissue without any detectable telomerase activity.27, 35, 36 Several factors have been discussed to explain the high hTERT levels in normal mucosa. First, although hTERT mRNA was identified as a determining factor for telomerase activity in tumor tissues like colorectal carcinoma,25–27 no reports have found a positive correlation between hTERT expression and telomerase activity in nontumorous tissues. Thus, hTERT well may be transcripted in nontumorous tissues, even at higher levels than the levels in tumor tissue, without resulting in telomerase activity. Detecting lower telomerase activity levels than expected from hTERT mRNA expression levels, Nakamura et al. hypothesized that telomerase activity not only may be controlled by the transcription of hTERT but also may be influenced at the posttranscriptional, translational, and posttranslational levels or by telomerase inhibitors.25 Second, alternate splicing of hTERT transcripts reportedly prohibited the formation of hTERT protein that contained functional reverse transcriptase domains37 and inhibited telomerase activity in nontumorous cells.38, 39 Finally, detected hTERT levels in whole tissue not only may be due to hTERT expression in epithelial or tumor cells but also may be influenced by hTERT transcription in other cells, like infiltrating activated lymphocytes.33 To obtain pure populations of epithelial cells, Liu et al. performed laser-capture dissection and found that, in contrast to our data, there was no hTERT expression in isolated nontumorous prostate epithelial cells.12 To date, studies using single-cell PCR to examine the origin of hTERT expression in nontumorous colorectal tissue have not been performed.
Although telomere shortening with increasing age is a well-known and well-explained phenomenon in normal somatic cells, including normal colorectal mucosa,2, 40 we are the first investigators to show that hTERT mRNA expression decreased with increasing age in nonmalignant tissue. Because no correlation between hTERT mRNA expression and age was found in tumor tissue, colorectal carcinoma seems to escape age-related hTERT expression control during the process of carcinogenesis. Thus, the expression of hTERT mRNA in tumor tissue appears to be a specific characteristic of each individual tumor. In addition, the correlation of hTERT in nontumorous samples with increasing age supports the hypothesis that the measured hTERT levels correspond to the age-influenced colorectal epithelial cells rather than the probably age-independent hTERT-expressing blood cells such as, e.g., activated lymphocytes.
Our study also is the first report on significant correlations between hTERT expression and grading in both tumor tissue and nontumorous tissue. To date, an increase in hTERT mRNA with increasing histologic grade of malignant clones has been shown only for some other tumor entities29, 41–43 but not for colorectal carcinoma7, 8 and never for corresponding normal tissue. The high hTERT expression in normal mucosa samples from the two patients with Grade 1 tumors cannot be interpreted reliably because of the small number of Grade 1 tumors (n = 2 tumors). Statistical significance for correlation of hTERT with histologic grade was achieved for the differences in hTERT expression between patients with Grade 2 tumors (n = 30 patients) and Grade 3 tumors (n = 25 patients; P < 0.02; Table 1). Because a correlation of hTERT mRNA with tumor grade was found in both tumor tissue and adjacent mucosa in our study group, the grade of differentiation of an emerging colorectal carcinoma seems to be determined already by the level of hTERT mRNA expression in normal colorectal mucosa. We therefore hypothesize that colorectal mucosa expressing high levels of hTERT mRNA may be more susceptible to the development of poorly differentiated colorectal carcinoma. This supports the theory that initiating events in carcinogenesis may occur within target cells that already express elevated levels of hTERT.44 Cells may be selected continuously for high hTERT levels as they acquire further genetic changes associated with carcinogenesis, like overexpression of large-T and mutation of H-ras,20 and may start to proliferate invasively. In this context, we found that hTERT mRNA expression in colorectal carcinoma tissue also increased with the depth of local invasion (pT). Currently, there are no other data available on hTERT mRNA expression and the correlation with depth of invasion, lymph node status, or the presence of metastasis in patients with colorectal carcinoma. To explain the correlation of hTERT only with histologic grade and not with tumor stage, it may be hypothesized that hTERT, as an intracellular factor for telomere regulation, is more likely to be related to the grade of differentiation and growth potential of a tumor cell itself than to the invasiveness (T) or even the metastatic potential (N,M) of the tumor. To grow invasively and form metastases, the primary tumor has to deal with numerous extracellular factors (immunologic tumor-host interactions).45 The difference in hTERT expression between males and females cannot be explained reliably based on our results and the current literature. A possible hypothesis is that gender specific hormones may interfere with the regulation of hTERT expression.
In conclusion, the different hTERT expression levels in tumor tissue and nontumorous tissue and their correlation with histologic grade imply characteristic changes of hTERT expression between nontumorous tissue and tumor tissue. We are the first authors to report on the prognostic potential of the hTERT expression pattern in patients with colorectal carcinoma: It may serve as a molecular marker for biologic and prognostic staging. Further studies should aim to correlate hTERT expression with telomerase activity and telomere length in malignant tissues and matched, nonmalignant tissues to further elucidate the mechanisms of telomere regulation, its therapeutic options,46 and the prognostic potential of hTERT expression.
The authors thank C. Marthen and D. Poehlmann for their expert technical assistance.