From January 1999 to December 2004, 50 cervical cancers tissue specimens were obtained by surgical resection at the Kyung Hee University hospital. Tissue specimens were snap-frozen in liquid N₂ and stored at -70℃ until used. The study group included 38 patients with the International Federation of Gynecology and Obstetrics (FIGO) stage IB, and 12 patients with FIGO stage II disease. The normal cervical tissues from 10 patients who underwent hysterectomy for benign indications were also collected as controls. Clinical outcomes were recorded through retrospective chart review.

One µL of total cellular RNA was converted to cDNA by reverse transcription using random hexamer primers and Moloney murine leukemia virus reverse transcriptase (Life Technologies Inc., Gaithersburg, MD, USA). PCR was initially carried out over a range of cycles (22-40 cycles) using 1:4 diluted cDNA (12.5 ng/50 µL PCR reaction), and PCR undergoing 24-34 cycles were observed to be within the logarithmic phase of amplification with all primers used for TGF-β1, TGF-βRI and TGF-βRII, Smad2, Smad3, Smad4, and an endogenous expression standard gene GAPDH. Primer sequences are available upon request. The cDNA was then subjected to 26-32 cycles of PCR at 95℃ for denaturation (1 min), 58-64℃ for annealing (0.5 min), and 72℃ for extension (1 min) in reaction buffer containing 1.5 mM MgCl₂ (PCR buffer II; Perkin-Elmer Corp., Norwalk, CT, USA). Quantitation of gene expression levels was performed by laser densitometric measurement of the ethidium bromide-stained gel. Absolute area integrations of the curves representing each specimen were compared after adjustment for the expression level of the housekeeping gene, GAPDH. Integration and analysis were performed using Molecular Analyst software program (Bio-Rad, Richmond, CA, USA). Quantitative PCR was repeated at least three times for each specimen and the mean was obtained. The relative levels of expression were represented as the ratio of target genes/GAPDH. We considered less than 50% expression of normal mean as being abnormally low, and more than double times expression of normal mean as abnormally high.

The entire coding regions of the TGF-β1, TGF-βRI, TGF-βRII, Smad2, Smad3, and Smad4 transcripts were amplified by PCR using several sets of primer for amplification of 200-to 280-bp lengths of fragments optimal for single-strand conformation polymorphism (SSCP) analysis. Twenty µL of these nest-PCR products were mixed with 5 µL of 0.5 N NaOH, 10 mM EDTA, 10 µL of denaturing loading buffer (95% formamide, 20 mM EDTA, 0.05% bromphenol blue, and 0.05% xylene cyanol), and 15 ml of double distilled H₂O. After heating at 95℃ for 5 min, samples were loaded in gels wells pre-cooled to 4℃. SSCP analysis was performed using 8% nondenaturing acrylamide gels containing 10% glycerol at room temperature.

Genomic DNA was prepared from tissue specimens using QIAGEN genomic-tip system (Qiagen, Hilden, Germany) according to the manufacturer's instructions. The concentration of extracted DNA was determined by a spectrophotometric measurement (Schimadzu Scientific Instruments Inc., Concord, CA, USA). PCR amplification using intron-specific primers was performed with 200 µg genomic DNA as template for 38 cycles at 95℃ (1 min), 58-62℃ (0.5 min), and 72℃ (0.5 min) in 1.5 mmol/L MgCl₂-containing reaction buffer (PCR buffer II, Perkin Elmer-Cetus, Concord). Ten µL of the PCR products were electrophoresed on 2% agarose gel. Quantitation was achieved by densitometric scanning and absolute area integrations of the curves representing each specimen were compared after adjustment for GAPDH.

To compare means the independent samples T test, and to compare frequencies the Chi-square test, or Fisher's exact test when appropriate, was used. A p-value of less than 0.05 was considered to be significant. All analyses were performed using the statistical software SPSS for Windows ver. 12 (SPSS Inc., Chicago, USA).

We evaluated the mRNA expression profile of TGF-β1, TGF-βRI, and TGF-βRII in a total of 60 cervical tissue specimens (Fig. 1). Ten specimens were normal cervical tissues from hysterectomies performed for non-malignant conditions, 44 were squamous cell carcinoma, 4 were adenocarcinoma, and 2 were adenosquamous carcinoma.

Quantitative RT-PCR was repeated at least three times for each specimen, and the mean expression level of TGF-β1, TGF-βRI, and TGF-βRII in noncancerous tissues was determined as 0.18, 0.77, and 1.15, respectively. Expression levels (Target/GAPDH) of TGF-β1, TGF-βRI, and TGF-βRII in cervical cancer tissues were observed within the range of 0.05-0.89 (mean 0.34), 0.65-0.94 (mean 0.77), and 0.16-1.27 (mean 0.92), respectively.

Compared with normal controls, RT-PCR analysis demonstrated an abnormal overexpression of TGF-β1 mRNA and decreased expression of TGF-βRII mRNA expression in the cervical cancer tissues (Table 1). We arbitrarily classified tumors with expression levels of less than a half of normal means as abnormally low expression. Abnormal overexpression of TGF-β1 and abnormal reduction of TGF-βRII were identified in 36% (18 of 50) and 20% (10 of 50) of cervical cancer tissues, respectively.

Next, we evaluated Smads expression in 50 cervical cancer tissues and 10 noncancerous tissues (Fig. 2). Our findings indicate that Smad2 and Smad4 mRNA expression is significantly related to the condition of the cervix (normal, cancer). As a result of RT-PCR analysis, cervical cancer tissues showed variable expression of Smad2, Smad3, and Smad4 (0.17-1.18, 0.79-1.21, and 0.18-1.12, respectively), and their levels were markedly reduced in Smad2 and Smad4 (mean; 0.84 and 0.85, respectively) compared to that of noncancerous controls (mean; 1.02 and 0.97, respectively). According to our arbitrary classification, 22% (11 of 50) in Smad2 and 14% (7 of 50) in Smad4 revealed tumor specific mRNA reduction of less than a half of normal means (Table 2)

To analyze the allelic status of TGF-βRII, Smad2, and Smad4 in cervical cancers, we first determined the genomic level of TGF-βRII, Smad2, and Smad4 using quantitative PCR. As shown in Table 3, the absence and marked reduction of TGF-βRII, Smad2, and Smad4 were detected in abnormal low expressed cervical cancer tissues which have homozygous deletion or LOH of the gene, while normal levels of TGF-β RII, Smad2, and Smad4 were detected in noncancerous tissues and normally expressed cancer tissues, indicating that genomic level determined by our quantitative PCR is consistent with the allelic status of the gene. Nine of the 10 (90%), 7 of the 11 (63.6%), and 4 of the 7 (57.1%) of cervical cancer tissues showing low mRNA expression represented a detectable reduction of TGF-βRII, Smad2, and Smad4 levels, respectively. In contrast, none of the noncancerous tissues and cervical cancer tissues showing normal mRNA expression cell lines had absence or significantly low frequencies of genomic reduction (Table 4).

To explore the presence of mutational alteration of TGF-β RII and Smads, we performed RT-PCR-SSCP analysis for TGF-βRII and Smads transcripts expressed from 50 primary tumor tissues. The entire coding region for TGF-βRII and Smads transcript was amplified using seven sets of exon specific primer sets, and the RT-PCR products were subjected to SSCP analysis. To improve mutation detection sensitivity, the same PCR products were digested using different restriction endonuclease(s), and SSCP was carried out under two different running conditions. However, we failed to detect any types of mutation leading to amino-acid substitutions or frameshifts of TGF-βRII and Smad3 in cervical cancer tissues, whereas Smad2 and Smad4 mutations were detected in 2 of 50 (4%) and 1 of 50 (2%), respectively (Figs. 3 and 4).