Elevated Expression of hTERT Is Associated with Dysplastic Cell Transformation during Human Oral Carcinogenesis in Situ

Tissue Samples.

We obtained NHOM (n = 20) and tissues samples with varying degrees of dysplasia: mild (n = 7), moderate (n = 5), and severe (n = 3) dysplasia, and SCC (n = 18). These samples were paraffin-embedded biopsy specimens archived at the Section of Oral Pathology, University of California Los Angeles School of Dentistry. The SCC tissues used in this study contained >80% of epithelial cells with atypical nuclei. Thin (6 μm) paraffin sections of each tissue were stained with H&E, and the grade of epithelial dysplasia was determined according to the standard criteria (30) . These tissue samples were used either to extract total RNA for quantitative determination of hTERT expression or to identify hTERT expression in situ.

cDNA Synthesis.

Total RNA was isolated from two 20-μm thick sections of NHOM and SCC tissues using Paraffin Block RNA Isolation kit (Ambion Inc., Austin, TX) according to the manufacturer’s guidelines. The isolated RNA was dissolved in 7.5 μl of H₂O, and reverse-transcription reaction was performed in the first strand buffer (Life Technologies, Inc., Rockville, MD), containing 200 units of Superscript II (Life Technologies, Inc.), 40 units of RNase inhibitor (Perkin-Elmer, Foster City, CA), 10 μm of dithiotrietol, 250 ng of random hexamer (Perkin-Elmer), and 2.5 μm deoxynucleotide triphosphate. The annealing reaction occurred for 10 min at 25°C, and cDNA synthesis was performed for 50 min at 42°C, followed by incubation for 10 min at 70°C for enzyme inactivation.

Real-Time PCR.

The principle of real-time PCR was first described by Heid et al. (31) . Briefly, amplification of the target sequence is monitored per PCR cycle by detecting the fluorescence signal emitted by the internal probe that is degraded by the 5′ nuclease activity of the Taq polymerase. The emission signal accumulates in each sample, and the C_t required to reach a given fluorescence threshold is determined. Thus, the C_t value of a sample inversely correlates to the quantity of the starting cDNA. Using the cDNA of known quantity, a standard curve can be generated and used to determine the starting amount of cDNA based on the C_t value of each sample.

In our study, we constructed the standard curves for hTERT and 18S rRNA using serially diluted (0.008–5 ng) cDNA of HOK-16B, an immortalized HOK, which highly expresses telomerase activity (16) . The PCR reaction for the standard curves and individual samples (18 SCC and 20 NHOM tissues) was carried with 5 μl of each RT samples and 20 μl of the master mix prepared with 1× TaqMan Universal Master Mix (Perkin-Elmer) and 200 nm of primers and the internal probe. We have used the identical primer and probe sequences as described by Bieche et al. (32) for amplification of hTERT. Amplification of 18S rRNA was performed with the primers and the internal probe provided by Perkin-Elmer. The obtained data for 18S rRNA was used to normalize the sample-to-sample variation in the amount of input cDNA and also to evaluate the quality of the isolated RNA and RT efficiency. The thermal cycling and collection of the fluorescence emission spectra were performed with an ABI Prism 7700 Sequence Detection System (Perkin-Elmer) under the parameters described elsewhere (31, 32, 33) .

In Situ RT-PCR.

The paraffin-embedded tissue samples were sectioned at 6-μm thickness. Three consecutive sections of each sample were placed per one In Situ PCR Glass Slide (Perkin-Elmer). The slides were baked at 60°C for 3 h to immobilize the sections, deparaffinized in xylene, and air-dried after dehydrating them in graded alcohol. To allow access to nucleic acids, the sections were digested with 2 mg/ml pepsin (Sigma Chemical Co., St. Louis, MO) in 0.1 n HCl for 30 min at 37°C in a humid chamber, and two of the three sections were treated with 20 units of DNase (Stratagene, La Jolla, CA) overnight. One remaining section was maintained untreated, and, thus, contained intact genomic DNA to control the efficiency of PCR amplification in situ. Subsequently, each section was submerged in 50 μl of one-step RT-PCR reaction solution containing 0.3 mm deoxynucleotide triphosphates, 0.06% BSA, 40 units of RNase inhibitor, 1.2 μm primers, 12 nm digoxigenin-dUTP (Roche Molecular Biochemicals, Indianapolis, IN), 2.5 mm Mn(OAc)₂, and 5 units of rTth DNA polymerase. The slides were assembled with AmpliCover Discs (Perkin-Elmer) and reverse-transcribed at 65°C for 30 min. The thermal cycling was then allowed in a GeneAmp In situ PCR System 1000 (Perkin-Elmer) at 94°C/3 min once and for 20 cycles of 94°C/45 s and 60°C/1.5 min. To control genomic DNA contamination, thermal cycling was performed in one of the two DNase I-treated sections in the absence of RT reaction. hTERT sequence was amplified with the primers 5′-ACTTTGTCAAGGTGGATGTGACGG-3′ (forward) and 5′-AAGAAATCATCCACC AAACGCA GG-3′ (reverse), yielding 493-bp amplicons (34) . The forward primer is located on exon 6 and the reverse on exon 10 (35) . To visualize hTERT amplification in situ, the slides were blocked in 0.1 × SSC/0.2% BSA at 45°C for 15 min. The incorporated digoxigenin-dUTP was probed with antidigoxigenin antibody using the DIG Nucleic Acid Detection kit (Roche Molecular Biochemicals).

hTERT Expression Was Elevated in Human Oral SCC Tissues.

We have quantitatively analyzed the level of hTERT expression in paraffin-embedded biopsy specimens, 18 of which were histologically diagnosed with oral SCC and 20 with NHOM. The validity of the real-time RT-PCR for quantitative amplification of hTERT has been established previously (31, 32, 33) . In our study, hTERT amplification was first assessed with serially diluted cDNA (0.008–5 ng) of cultured HOK-16B, an immortalized HOK cell line (36) , which strongly expresses telomerase activity (16) . During the amplification reaction, the emission spectra were collected at every cycle, and the C_t value, i.e., the minimal cycle number at which the fluorescence threshold was reached for individual amplification reaction, was determined for each sample with an increasing amount of input cDNA (Fig. 1A)⇓ . Comparison between the hTERT C_t values and the amount of cDNA input yielded a remarkable correlation (r = 0.995; P < 0.001; Fig. 1A⇓ ). A similar result (r = 0.993; P < 0.001) was also obtained for amplification of 18 s rRNA that served as an internal control with which we normalized the variation in the starting amount of cDNA in different samples. This abundantly expressed rRNA was also used to evaluate the quality of the total RNA extracted from each tissue sample.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Quantitative amplification of hTERT transcript isolated from paraffin-embedded tissue specimens by real-time RT-PCR. A, a standard curve was constructed with serially diluted (0.008–5 ng) cDNA of HOK-16B cell line and respective C_t values (▴) for hTERT amplification. The input cDNA amount and the C_t values were negatively correlated with remarkable consistency (r = −0.995; P < 0.001). This standard curve was found to be inclusive of all C_t values of individual samples (•) and was used to convert the C_t numbers to the input cDNA amount equivalent to the known cDNA quantity of HOK-16B. B, fluorescence emission spectra were collected from each sample once per PCR cycle, and the kinetics of hTERT amplification was monitored over the entire thermal cycling reaction. The C_t values of individual samples were determined at the minimal number of cycles needed to reach the fluorescence threshold.

The C_t values obtained from amplification of hTERT in each SCC and normal samples were converted to nanograms of input HOK-16B cDNA using the standard curve shown in Fig. 1A⇓ (Table 1)⇓ . In some SCC (6/18) and normal tissues (7/20), amplification of 18 s rRNA did not reach the fluorescence threshold, indicating the failure to extract enough RNA of detectable quality from these specimens, and they were excluded from data analysis. The nanograms corresponding to hTERT C_t values of the remaining specimens were then corrected for the sample-to-sample variation against that of 18 s rRNA to yield normalized hTERT expression. Most normal epithelial tissues exhibited detectable hTERT expression with the mean (± SE) normalized value of 7.0 ± 2.3 without preference to the tissue origin nor to the age or sex of the patients (Table 1)⇓ . hTERT expression was detected in 11 of the tested 13 SCC tissues (48.6 ± 17.5) and was elevated to varying degrees in 10 SCC tissues compared with that of NHOM, resulting in the overall fold-induction by a factor >6.9.

hTERT Activation Occurred in Moderate Dysplasia during Human Oral Carcinogenesis in Situ.

The above data clearly indicated a quantitative difference in the level of hTERT expression between NHOM and SCC tissues. It was additionally necessary to delineate the hTERT expression level in individual cells of the tissues and to identify the clinical stage at which hTERT activation occurred. For this purpose, we determined hTERT expression by in situ RT-PCR in NHOM and in resected tissues with varying degrees of histopathology.

We obtained paraffin-embedded biopsy specimens with mild (n = 7), moderate (n = 5), and severe (n = 3) dysplasia, and invasive carcinoma (n = 4). The grades of epithelial dysplasia and histological diagnosis were made according to the standard criteria (30) . In particular, we evaluated cytological atypia, atypical mitoses, or nuclear atypia, and other cellular alterations along the basal and parabasal layers. Using these tissue specimens, the sections were first deparaffinized and digested with DNase I to degrade genomic DNA. Subsequently, hTERT mRNA was reverse transcribed and amplified in a one-step reaction containing digoxigenin-dUTP. hTERT amplicon was then detected in situ by direct immunohistochemical staining using antidigoxigenin antibody coupled with alkaline phosphatase. The use of this technique was validated by in situ staining of the sample without DNase I digestion, which showed intranuclear staining where amplification of genomic DNA occurred (Fig. 2)⇓ . However, after the enzyme digestion, the PCR reaction without reverse transcription yielded no visible amplification, indicating that DNase I digestion in situ was sufficient to prevent amplification of cognate genomic DNA sequence. The above two controlled reactions were carried out in parallel with individual samples of interest for each in situ RT-PCR reaction on the same slide.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

In situ RT-PCR of hTERT in NHOM. A, thin sections (6 μm) of NHOM were stained for H&E. B, a consecutive section of the same tissue was used for in situ RT-PCR using hTERT primers (see “Materials and Methods”) after DNase I digestion of genomic DNA. C, in the absence of DNase I digestion, the cognate genomic sequence was amplified with hTERT primers and was shown as intranuclear staining. The area within the box was magnified ×4 (inset). D, without reverse-transcription and with DNase I digestion, no amplification of hTERT occurred. Bar, 50 μm.

None of the 10 tested NHOM showed hTERT amplification (Table 2)⇓ , albeit some normal tissues with focal epithelial hyperplasia contained positively stained cells primarily along the basal cell layer (Fig. 3A)⇓ . The signals detected in these cells, in contrast with the genomic DNA amplification (Fig. 2C)⇓ , were localized to the cytoplasmic region resembling a “donut-shaped” staining pattern. hTERT amplification remained undetectable in tissues of mild dysplasia (Fig. 3B)⇓ until moderate dysplasia in which discrete amplification signal was visible in the cytoplasmic region (Fig. 3C)⇓ . Similar pattern and intensity of staining was detected in tissues of severe dysplasia and invasive carcinoma (Fig. 3, D and E)⇓ . Thus, hTERT expression and, hence, telomerase activation appears to occur during moderate dysplasia.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Expression of hTERT during multistep human oral carcinogenesis in situ. hTERT transcripts were detected in situ by RT-PCR in paraffin-embedded human oral tissues with varying degree of histopathology: basialar hyperplasia (A), mild dysplasia (B), moderate dysplasia (C), severe dysplasia (D), and squamous carcinoma (E). The left half of each panel shows the histopathology by H&E staining, and the right half the in situ RT-PCR of the same tissue for hTERT expression. Arrows, examples of positive hTERT staining; arrowhead, an example of mitotic figure at the basal cell layer. Bar, 50 μm.