Colorectal carcinogenesis is a multistep process, from adenoma to invasive carcinoma, for which several separate genetic events have been characterized involving genes such as APC, k-ras, DCC and p53.1 Early changes are believed to cause an increased proliferation resulting in adenoma formation and an enhanced vulnerability for new genetic lesions. However, for each cell division the telomeres will progressively shorten and at a certain critical telomere length the cells are predestined to enter a senescence program. This program can be abolished by reactivation of telomerase, a ribonucleoprotein complex with reverse transcriptase activity, which can maintain telomeres by adding new telomere repeats.2 Telomerase reactivation and cellular immortalization are thereby processes implicated in carcinogenesis and tumor progression, and telomerase activity has also been detected in high frequency in human malignancies.3 There are indications that telomerase activity is associated with a neoplastic phenotype in colorectal epithelium and during the adenoma-carcinoma progress an increased fraction of cases are telomerase activity positive4, 5, 6, 7, 8 and show increasing hTERT RNA expression.9, 10
Although telomerase activation is a common phenomenon in human malignancies and immortal cell lines the regulation of telomerase is not fully understood. hTERT can be regulated by a multitude of factors indicating that different sets of regulators are active in different cell types. One example is the methylation status of the hTERT promoter, which has been reported to repress as well as activate hTERT in various types of cells.11, 12 Histone acetylation has also been found to be involved in the transcriptional regulation of the hTERT gene.13, 14, 15 Furthermore, several proto-oncogene and tumor suppressor gene products have been implicated in the regulation of telomerase activity, including c-Myc, bcl-2, p21, pRb, p53, PKC, Akt/PKB, c-Abl and protein phosphatase 2A (reviewed in 16).
At the gene level, we have previously shown that the size of hTERT minisatellites was unrelated to telomerase expression in colorectal carcinomas.17 Interestingly, amplification (or increase in gene copy number) of the hTERT gene at chromosome 5p15.33 has been demonstrated in a significant fraction of established cell lines and malignant tumors of different origin using FISH or differential PCR.18, 19, 20 Whether this increased gene dosage also results in an increased hTERT RNA expression followed by upregulation of telomerase activity has not been clarified. In a series of embryonal brain tumors, hTERT gene amplification was associated with increased mRNA expression but telomerase activity was not studied.20 In our study, we analyzed a series of colorectal carcinomas for hTERT gene copy number, hTERT RNA levels, telomerase activity, p53 protein expression and DNA ploidy. Our data indicate that increased copy numbers of the hTERT gene is common in colorectal carcinoma and is associated with p53 accumulation and DNA ploidy but not with hTERT expression or telomerase activity.
Material and methods
Diagnostic tumor samples were collected from 64 patients with colorectal carcinoma treated with potentially curative surgery. No patient had been treated with radiation or chemotherapy before sampling. The mean age of the patients was 70 years (range 41–86), and 34 patients were men and 30 were women. Twenty-four tumors were derived from right colon, 22 from left colon and 18 from rectum.
All tumors were directly taken care of after excision and pieces were snap frozen in liquid nitrogen and stored in −80°C until extracts for analysis were prepared. Strictly adjacent tissue pieces were formalin fixed, paraffin embedded and used for routine morphological examination and immunohistochemical staining. From routinely H/E (hematoxylin/eosin)-stained sections, the pathologist staged all tumors according to the classification of Dukes: 6 were Dukes' A, 33 Dukes' B and 25 Dukes' C. All tumors were graded using a 4-graded scale regarding morphological differentiation (well, well-moderate, moderate-poor and poor).
All 64 patients were possible to analyze regarding hTERT gene copy number and telomerase activity. However, only 55 cases were possible to evaluate for hTERT RNA expression, 51 for DNA ploidy, 58 for microsatellite stability status and 63 for p53 expression. The dropping off was mainly due to lack of tissue. For p53, 1 case was excluded due to repetitive loss of tumor tissue during the antigen retrieval procedure.
The tumor sections taken strictly adjacent to the fresh frozen tissue were evaluated regarding the amount of tumor cells per unit of area. The evaluation was performed using a semiquantitative 4-graded scale. No significant relationship between the amount of tumor cells and hTERT RNA expression and telomerase activity was observed (data not shown).
Fluorescence in situ hybridization (FISH)
The hTERT gene status was analyzed by FISH performed on 4 μm paraffin embedded tumor sections as previously described.18 The hTERT locus specific probe was isolated from a PAC clone RPCI-6 135 M06 (Pieter de Jo, Roswell Park Cancer Institute, Buffalo, NY) and labeled with Spectrum Orange fluorophore-conjugated dUTP by nick translation.18 The specificity of the hTERT probe has been corroborated by using PCR to analyze hTERT copy number in tumor cell lines (reference 18 and unpublished data). Signals of both hTERT and a Spectrum Green-dUTP labeled 5q31 marker probe (685A18, PAC library RPC14, Pieter de Jo, Roswell Park Cancer Institute, Buffalo, NY) were counted from at least 100 nonoverlapping nuclei. Increased copy number was defined as ≥ 3 hTERT gene copies/nucleus in at least 20 % of the cells.
The tissues in paraffin sections can vary somewhat in different areas due to differences in fixation and section thickness giving variability in hybridization conditions leading to heterogeneous FISH signals. This natural feature of paraffin embedded material can hamper a proper evaluation of the FISH signal and analysis of spatial heterogeneity. In order to further study this issue, we analyzed a subset of 6 tumors using nuclear suspensions from enzymatically disintegrated paraffin sections. These tumors were from the first sampling year of our tumor collection, and 2 cases with normal gene copy number and 4 with increased gene copy numbers in the FISH analyses on tumor sections with tissue available for disintegration were selected. Nuclear spreads were air-dried, incubated in 75 mM KCl (20 min, 37°C) and fixed in Carnoy's solution for 5 min at room temperature. Thereafter the slides were RNAse (100 μg/ml) treated for 1 hr and washed 3 times in 2× SSC. Finally, the slides were incubated in 100 μg/ml pepsin in 0.01 M HCL for 10 min, followed by PBS for 5 min at room temperature and dehydration in alcohol (70%, 80% and 95 %). The Spectrum Orange labeled RPCI-6 135 M06 clone (see above) was used for hTERT detection and the 5q31 probe (Spectrum Green) as a control probe. Ten microliters of probe containing 100 ng specific DNA and 5 μl Cot-1 DNA in 60% formamide was pre-incubated for 1 hr at 37°C and then applied to each slide. Probe and target DNA were denatured simultaneously for 3 min at 72°C. Slides were hybridized overnight at 37°C in a humid chamber (HYBrite, Abbott Laboratories, North Chicago, IL), washed in 2× SSC containing 0.3% NP-40 and counterstained in DAPI. In each case, 2 independent investigators counted hTERT and 5q31 signals in 200 nuclei. Analysis was performed using an Axioplan 2 microscope (Zeiss, Inc., Thornwood, NY). Digital images were captured and stored using Cytovision software, version 3.0 (Applied Imaging, Inc., Newcastle, UK).
Preparation of extracts was performed as described and the placental ribonuclease inhibitor was added to the lysis buffer in order to protect against endogenous ribonuclease activity.21, 22 Telomerase activity was measured by the TRAPeze method (TRAPeze Telomerase Detection Kit, Oncor, Inc., Gaithersburg, MD) according to the guidelines given by the supplier. The relative telomerase activity level was expressed as units of TPG (Total Product Generated). The samples were analyzed at 0.5 μg protein of extract/assay and the level of telomerase expression was expressed as the mean of 2 or more analyses. Good linearity was confirmed between 0.1–2.5 μg protein per assay both for samples used in the study as well as for cell lines (not shown in figures).
hTERT RNA expression
Total RNA was extracted according to standard procedures. The amount of hTERT mRNA was quantified using the Light Cycler Telo TAGGG hTERT Quantification kit using 200 ng total RNA according to the manufacturer's protocol (Roche, Basel, Switzerland). hTERT RNA levels were expressed as a ratio between the expression level of hTERT RNA and a housekeeping gene RNA (porphobilinogen deaminase) according to the manufacturer's protocol. The total material was run as single tests at 2 different occasions with < 2 years interval and the correlation between the 2 runs was highly significant (rs =0.93, p<0.001), and based on this the data set generated in the first analysis was used.
The expression of p53 protein was analyzed with immunohistochemistry. Routinely formalin fixed and paraffin embedded tissue was immunohistochemically stained in a semiautomatic staining machine (Ventana ES, Ventana, Inc., Tucson, AZ) as previously described.23 Antigen retrieval was performed in citrate buffer (pH 6.0). A primary monoclonal antibody against p53 (Ab-6, Oncogene Research Inc.) was used in a dilution of 1:400. The antibody recognizes both wild-type and mutated p53 protein. The p53 immunoreactivity was evaluated by classifying the tumors in 2 categories; (−) corresponding to labeling index (LI) <5%, (+) ≥5% according to McKay et al.24 where (+) designate tumors with abnormal p53 accumulation. Only nuclear staining of the p53 antibody was considered.
DNA ploidy analysis
DNA-staining was performed according to Vindeløv et al.25 A FACScan flow cytometer was used (Becton Dickinson Immunocytometry Systems, Inc., San Jose, CA) and the DNA index (DI) was calculated. A tumor was denominated diploid (DI = 1) when only 1 peak was detected and aneuploid when 2 (or more) separate peaks were found.
DNA extraction and microsatellite instability (MSI) analysis
Fresh frozen tissue was available from 58 out of 64 tumors and DNA was extracted from tumor and corresponding normal tissue using the NucleonTMST kit (Amersham Life Science, England) according to the manufacturer's instructions. In 7 cases, control DNA had to be extracted from formalin-fixed and paraffin-embedded material since snap frozen normal tissue was missing. Ten micrometer sections (5 to 10) were deparaffinized in xylene, centrifuged and incubated in alcohol. The retrieved material was incubated overnight in a buffer containing 50 mM Tris-HCl, pH 8.0, 1 mM EDTA, 0.50% Tween 20 and proteinase K (0.4 mg/ml) at +55°C. Inactivation of proteinase K was achieved by incubation at +95°C, 8 min.
Microsatellite status was determined using PCR and as suggested by the National Cancer Institute Workshop,26 5 different microsatellite markers including 2 mononucleotides repeats and 3 dinucleotide repeats were used. The loci used was D2S123 and D17S250 as (CA)n repeats and BAT25, BAT26 and BAT34C4 as poly-A repeats. According to Boland et al.,26 a tumor was scored as high-frequency MSI (MSI-H) if at least 2 out of 5 markers showed instability, while those with 1 out of 5 markers was scored as low-frequency MSI (MSI-L). Tumors with no marker showing instability were defined as microsatellite stable (MSS). The PCR reaction was performed using standard conditions with 1 of the primers in each pair being fluorescence-dye labeled. The PCR products were analyzed in an ABI 377 Automated Genetic Analyzer (Applied Biosystems, Foster City, CA) with the help of GeneScan software.
Comparison between groups in cross tabulations was performed using the Fisher's Exact test. Differences between groups regarding continuous variables were examined using the Mann Whitney U test. Spearman's correlation coefficient (rs) was used to compare sets of continuous variables. Kaplan-Meier's method was used to estimate cancer specific survival, and comparison between groups was tested with the log-rank test. The time measured from primary tumor resection to death is recorded as the survival time, in which death with known locoregional or distant metastases were processed as an event. If no event occurred, the patient was censored at the time of last clinical follow up or death from other causes. The median follow up time of surviving patients ranged from 1–104 months (median 32). All calculations were performed using SPSS version 11.0 (SPSS, Inc., IL).
hTERT gene copy number in colorectal cancer
The hTERT gene demonstrated an increase in copy number in 31 out of 64 cases (48%) (Fig. 1). In tumors with increased copy number by FISH on sections more than 80% of suspended nuclei from the enzymatically disintegrated paraffin sections showed a similar copy number increase, indicating a low grade of heterogeneity regarding hTERT copy number distribution (Fig. 2a,b).
Telomerase activity, hTERT RNA expression and hTERT gene copy number
Telomerase activity data were obtained from 64 tumors and 49 of these (77 %) were found to be telomerase activity positive. In 15 cases, no telomerase activity could be detected in repeated experiments. In the positive tumors the relative telomerase activity levels varied considerably between individual tumors, from 2 to 700 TPG. The mean relative telomerase activity was 52.0 TPG (SEM ±12.4). Expression of hTERT RNA was detected in 50 out of 55 included cases (91 %) with relative RNA levels ranging from 0.004 to 0.28.
A significant, but rather weak, correlation was demonstrated between telomerase activity and hTERT RNA expression (rs =0.29; p = 0.030) (not shown in figures), but neither of these parameters were significantly associated with the hTERT gene copy number. For further details, see Table I.
Table I. hTERT Gene Copy Number in Relation to Telomerase Activity and hTERT RNA Expression
hTERT gene copy number
hTERT mRNA expression
p53 protein expression and hTERT gene copy number
Forty-five out of the 63 cases (71 %) evaluated by immunohistochemistry showed abnormal nuclear p53 protein accumulation. A strong relationship was found between p53 status and hTERT gene copy number status (normal vs. increased copy number) (p=0.002). Only 3 out of 31 cases (9.7%) harboring increased hTERT gene copies/nucleus had a normal p53 status in contrast to 15 out of 32 (47%) cases with a normal hTERT gene copy number (Table II).
Table II. hTERT Gene Copy Number in Relation to p53 Protein Expression, DNA ploidy and Microsatellite Instability1
hTERT gene copy number
Number of cases in each group is displayed.
DNA ploidy, microsatellite instability and hTERT gene copy number
The vast majority of the tumors with increased copy numbers of the hTERT gene were aneuploid (22/27, 81 %), whereas the tumor group with normal copy number had an equal distribution of diploid and aneuploid tumors (p=0.036). There was no association between microsatellite instability (MSI) and copy number of the hTERT gene (p=0.61), but only 4 cases with MSI (all classified as MSI-H) were detected.
Increase in hTERT gene copy number was not associated with patient prognosis (p=0.43) (Fig. 3). Excluding the MSI cases from the analysis hTERT gene copy number was still not significantly associated with patient prognosis (p=0.15).
Tumor development includes a series of genetic events leading to a malignant phenotype. Escape from replicative senescence resulting in an infinite proliferative potential (“immortalization”) can be an important step during cancer development and progression. The abrogation of life-span checkpoints is a complex process requiring cell-type specific changes including abnormal expression of cell-cycle checkpoint molecules and derepression or upregulation of telomerase followed by telomere length stabilization. The regulation of telomerase and its catalytic subunit hTERT is a complex and cell type dependent process where a multitude of factors have been implicated,16 including proto-oncogene and tumor suppressor gene products, some known to be associated with colorectal carcinoma progression.
During tumor progression, increasing genetic instability is a common phenomenon. Gross genetic alterations, both losses and gains, can be detected by comparative genomic hybridization (CGH) and in colorectal carcinomas aberrations, e.g., amplifications, involving a large number of chromosomes, have been identified using this technique.27, 28, 29 Theoretically, amplification of the hTERT gene can be 1 mechanism for upregulation of telomerase. In our study, we could show a high frequency of colorectal cancer samples with increased copy number of the hTERT gene located on chromosome 5p15–33. In FISH analysis, about half of the cases demonstrated ≥ 3 hTERT gene copies per nucleus. These data are in agreement with previous studies showing that hTERT amplification is a common finding in epithelial and brain tumors,18, 19, 20 but the association between this feature and telomerase activity has not been evaluated. In our study of colorectal cancer, no significant association between hTERT gene copy number and hTERT RNA expression or telomerase activity levels could be demonstrated, indicating that this factor is of no or minor importance for the expression of telomerase in colorectal cancer. It is fully possible that the hTERT gene copy number can be of significance in this respect for other tumors since in embryonic tumors of the central nervous system an association between hTERT amplification, studied by differential PCR, and hTERT RNA expression has been described.20 No data on telomerase activity levels were however given. The collected data on hTERT gene copy numbers are in line with CGH analyses, showing that gene amplifications are common in colorectal carcinoma.29 The same study also demonstrated that consequence of amplification in general is not increased gene expression within the amplicons, which also give support to our data on hTERT.29
Other possible confounding factors can be tumor heterogeneity and different proportions of stromal cells. which theoretically could conceal virtual correlations between hTERT gene copy number and RNA expression and telomerase activity. The FISH signal varied somewhat in paraffin sections, but this phenomenon was most likely due to differences in section thickness and fixation efficiency. Support for this was obtained by our further analysis of hTERT copy number in disintegrated nuclei from paraffin embedded material where copy number increase involved most cells within a sample, indicating a low degree of heterogeneity. Neither did the amount of stromal component influence the hTERT RNA expression or telomerase activity data (not shown in results).
However, we found a statistically significant association between nuclear p53 protein accumulation, indicating inactivated p53 protein, and hTERT gene copy number. This is in accordance with a previous study of a small set of breast cancers.17 We could also demonstrate a significant correlation between increased hTERT gene copy number and aneuploidy. The relationship in colorectal cancer between p53 abnormalities and aneuploidy is well known30, 31 and aneuploidy and gene amplification is closely related.27 Furthermore, inactivation of p53 has been suggested as a requirement for admitting gene amplification.32, 33 With disruption of the p53 pathway cells with DNA, abnormalities are allowed to proceed in the cell cycle, which can lead to establishment of clonal cell populations with DNA errors, including gene amplifications. As a consequence, tumors with p53 mutations have shown more genomic aberrations as detected by CGH, indicating that gross genomic instability is associated with p53 inactivation.27 In contrast to aneuploid tumors, the few cases in our material demonstrating microsatellite instability were all diploid, and all but 1 showed normal hTERT gene copy number. However, p53 accumulation was observed in all but 1 MSI case. Our results partly support the suggestion that colorectal cancers can evolve along 2 separate genetic pathways.34 One pathway in this scenario is characterized by gross genomic instability and is associated with aneuploidy, p53 inactivation and CGH detected aberrations, 1 example of which is hTERT gene amplification. The second pathway is exemplified by diploid tumors with more subtle genetic mutations, 1 example of which is microsatellite instability. Nevertheless, many clinical tumor samples show mixed patterns making this distinction difficult in individual cases.
In our rather small set of patients with colorectal cancer, hTERT gene copy number was not associated with survival, neither after exclusion of the MSI cases who are known to have a better prognosis than microsatellite stable tumours.35 The hTERT amplification, and the possible cellular effects thereof, is expected to occur rather early in colorectal tumorigenesis rather than being involved in late metastatic processes. From this perspective the lack of prognostic significance of hTERT gene dosage is not very surprising.
In summary, one possible mechanism for telomerase upregulation is amplification of the hTERT gene, but in the present material no correlation was present between gene amplification/copy increase and hTERT RNA or telomerase activity levels. Thus, genetic changes at the chromosomal level are frequent but the functional effects seem less significant, one example of which seems to be hTERT gene copy number and telomerase expression in colorectal carcinoma.
The authors are grateful to U.-B. Westman, K. Näslund and B Bäcklund for their skillful technical assistance.