Uterine cervical cancer, the second most frequently occurring cancer in women worldwide, is tightly associated with the expression of high-risk human papillomavirus [mainly human papillomavirus (HPV)-16 and HPV18] oncogenes E6 and E7 and characteristically exhibits chromosomal instability. However, the mechanisms underlying chromosomal instability in cervical cancer are still not fully understood. In this study, we observed that two of three human cervical epithelial cell lines expressing HPV16 E6E7 became immortalized without extensive chromosomal instability and crisis. The introduction of transforming growth factor (TGF)-β1, a multiple functional cytokine/growth factor, in the culture medium induced crisis, which was associated with massive chromosomal end-to-end fusions and other structural aberrations. The distributions of structural aberrations on individual chromosomes were significantly correlated with the profiles of telomere signal–free ends. The immortalized cells that emerged from the TGF-β1–induced crisis showed multiple clonal structural aberrations that were not observed in cells without TGF-β1 treatment. Overexpression of the catalytic subunit of telomerase (hTERT) abolished the effects of TGF-β1 on chromosomal instability. Interestingly, another HPV16 E6E7–expressing cervical cell line that experienced crisis and telomere dysfunction under ordinary culture condition had a higher level of autocrine TGF-β1 production than the other two crisis-free immortalized cell lines. Blocking the TGF-β1 pathway by an inhibitor of TGF-β1 receptor type I prevented the crisis and telomere-mediated chromosomal instability. In addition, more dramatic telomere shortening was observed in cervical intraepithelial neoplasias having higher expression of TGF-β1 in vivo. These results together suggest an important role of TGF-β1 in the early process of cervical carcinogenesis. [Cancer Res 2008;68(17):7200–9]

Cervical cancer is the second most frequently occurring cancer in women worldwide and causes the highest cancer mortality among women in many developing countries (1). Infection with high-risk human papillomavirus [mainly human papillomavirus (HPV)-16 and HPV18] is a necessary etiologic factor for cervical cancer (1, 2). High-risk HPV types exert their oncogenic potential mainly through deregulated expression of viral oncogenes E6 and E7, which inactivate p53 and Rb proteins, respectively, by accelerating proteolytic degradation (3). However, extensive studies have shown that the expression of the viral oncogenes per se is not sufficient for cancer development. Chromosomal instability, the major form of genomic instability, may be an important additional driving force for cervical carcinogenesis. By continuous generation of new genetic alterations, chromosomal instability provides the cells with selective advantages in acquiring new features in a stepwise process toward cancer (4, 5).

Among multiple sources of chromosomal instability, telomere dysfunction has been revealed as the prime one contributing to cancer initiation under the setting of defective p53 (6, 7). It is known that HPV16 E6 protein induces telomerase activity soon after its expression in several types of human epithelial cells including cervical epithelial cells in vitro (811). This function of E6 is expected to suppress the telomere shortening–mediated chromosomal instability. Yet the paradox is that even the premalignant cervical lesions have significantly shortened telomeres (12, 13) as well as multiple chromosome abnormalities (14, 15).

Chronic inflammation has been identified as a cofactor for HPV-driven cervical carcinogenesis (16, 17). However, the exact mechanisms by which chronic inflammation cooperates with HPV to promote cervical cancer are not clear. Transforming growth factor β1 (TGF-β1) is a polypeptide cytokine that plays essential roles in regulating multiple biological processes such as cell proliferation, differentiation, and inflammation (18). Recently, TGF-β1 has been found to inhibit telomerase activity in several cell types such as immortalized human skin keratinocytes (19) and human breast and lung cancer cells (20, 21). TGF-β1–induced phospho-Smad3 directly interacts with and down-regulates c-Myc, a transcriptional factor of hTERT, which is a rate-limiting protein determining telomerase activity (22). On the other hand, the telomerase activation by HPV16 E6 requires the presence or induction of c-Myc (10, 23). These results prompt us to hypothesize that TGF-β1 can antagonize HPV16 E6–induced telomerase activation, resulting in telomere shortening–mediated chromosomal instability. Thus far, the vast majority of studies on the mechanisms of chromosomal instability associated with high-risk HPV oncogene expression have been done in noncervical cells in vitro. In this study, for the first time, we showed that physiologic levels of TGF-β1 could effectively induce telomere-mediated chromosomal instability in human cervical epithelial cells, the most relevant target of HPV infection, during cellular immortalization, which is an early and indispensable step toward cancer. The association between an increase in TGF-β1 signaling and telomere shortening was also observed in human cervical intraepithelial neoplasia (CIN) specimens.

Cell culture. Primary normal epithelial cells (NC104, NC105, and NC106) were cultured from fresh human cervical tissues donated by three independent individuals aged 46, 51, and 47 y, respectively, after hysterectomy for benign uterine diseases. Informed consents were obtained from the patients before the surgeries. The culture medium was a 1:1 mixture of defined keratinocyte serum-free medium (Life Technologies, Inc.) and EpiLife (Cascade Biologics) with the provided supplements. The cells were subcultured at a split ratio of 1:3 or 1:4 when they reached ∼80% confluency. Population doublings (PD) were calculated after the first subculture from primary cells (PD0). At PD4, the cells were infected with LXSN16E6E7 retroviruses or control LXSN retroviruses expressing neomycin-resistant gene. The viral vectors were a gift from Dr. Denise Galloway (Fred Hutchinson Cancer Research Center, Seattle, WA). Two days after retroviral infection, the cells were selected with 0.2 mg/mL G418 for 10 d. For hTERT overexpression, NC104-E6E7 and NC105-E6E7 cells at PD6 were infected with pBABE-hygro-hTERT viruses or the control pBABE-hygro viruses. The retroviral vectors were a gift from Dr. Robert A. Weinberg (The Whitehead Institute for Biomedical Research, Cambridge, MA). Two days after infection, the cells were selected with 20 μg/mL hygromycin for 4 d. Human recombinant TGF-β1 and SB-431542 were from Calbiochem and Sigma-Aldrich Chemie GmbH, respectively.

Sequential telomere quantitative fluorescence in situ hybridization and spectral karyotyping. Metaphases were harvested and chromosome spreads prepared as described (24). Telomere quantitative fluorescence in situ hybridization (FISH) and spectral karyotyping (SKY) were done sequentially as previously reported (25). At each PD analyzed, 20 and 100 metaphases were analyzed by quantitative FISH and SKY, respectively. Only intact ends were measured for telomere signals to correlate with chromosomal aberrations.

Anaphase bridge analysis. Cells were cultured in chamber slides for at least 72 h, fixed with 1:3 acetic acid/methanol for 30 min without metaphase arrest, air-dried with nitrogen gas, and stained with 4′,6-diamino-2-phenylindole. Anaphase bridges were identified as having strings of chromatin connecting the two poles of anaphase cells.

Western blotting. Twenty-microgram protein from whole-cell extracts was separated by SDS-PAGE and blots were prepared on a polyvinylidene fluoride membrane (Amersham). Primary antibodies against c-Myc (C-33), HPV16 E7 (ED 17), Rad51, and actin were from Santa Cruz Biotechnology; anti–TGF-β (56E4) and anti–phospho-Smad3(Ser423/425) were from Cell Signaling Technology; and anti-p53 was from DakoCytomation. Blots were probed with secondary antibody against peroxidase-conjugated mouse, rabbit, or goat IgG and visualized by the enhanced chemiluminescence Western blotting system (Amersham).

Autocrine TGF-β1 detection. Cells were cultured in fresh medium for 48 h to reach ∼80% confluence. The conditioned medium was collected for autocrine TGF-β1 detection using the Human Cytokine Antibody Array kit (RayBiotech).

Reverse transcription-PCR for HPV16 E6 and E7. Total RNA was isolated using the Qiagen RNeasy Mini Kit and immediately used for cDNA preparations using SuperScript III reverse transcriptase (Invitrogen Life Technologies) according to the manufacturer's recommendation. Primers for the HPV16 E6 gene amplified a 120-bp specific fragment and primers for the HPV16 E7 gene amplified a 133-bp specific fragment (26). Primers for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) amplified a 372 bp specific fragment as an internal control. The PCR reactions were done at 94°C for 5 min, followed by 30 cycles of 45 s at 94°C, 45 s at 60°C (E6, E7)/64°C (GAPDH) and 45 s at 72°C, and then extension for 5 min at 72°C. The reverse transcription-PCR (RT-PCR) products were visualized by agarose gel electrophoresis with SYBR Gold (Molecular Probes) staining.

Telomerase activity measurements. Telomerase activity was monitored by the telomerase repeat amplification protocol using TRAPEZE telomerase detection kit (Chemicon). For each PCR, 0.1-μg protein was used. PCR was run for 33 cycles. The products after PAGE were stained with SYBR Green (Invitrogen).

Immunohistochemistry and telomere FISH on CIN specimens. Serial sections (5 μm thick) of formalin-fixed, paraffin-embedded CIN specimens from 20 patients (3 CIN I and 17 CIN II–III) were obtained from the archives of the Department of Pathology, The University of Hong Kong. Immunohistochemistry was done using a microwave-based antigen retrieval technique (27). The primary antibodies mouse anti–c-Myc, rabbit anti–TGF-β1, and rabbit anti–phospho-Smad2/3(Ser423/425) were from Santa Cruz Biotechnology. The secondary antibodies (horseradish peroxidase–conjugated rabbit anti-mouse IgG and swine anti-rabbit IgG) were from DakoCytomation. Immunostaining was visualized with 3,3′-diaminobenzidine tetrahydrochloride/hydrogen peroxide as chromogen and the sections were counterstained with hematoxylin. For telomere FISH, the deparaffinized sections were microwave heated for 10 min in 0.01 mol/L citric buffer (pH 6.0), then treated with 0.1 mg/mL RNase at 37°C for 1 h, followed by digestion with 0.5 mg/mL proteinase K for 30 min at 37°C. Other procedures for slide and probe denaturation, incubation, washes, and telomere signal quantification were done as described (25).

Telomerase activity and chromosomal instability in HPV16 E6E7–expressing cervical epithelial cells. Previous studies have shown that some human epithelial cells including cervical epithelial cells expressing HPV16 E6E7 (9, 11, 28, 29) or whole-length HPV16 (30) can be immortalized without crisis, which is characterized by the halt of cell population increase due to massive cell death. In this study, we have established three cervical epithelial cell lines by expression of HPV16 E6E7 in primary normal cervical epithelial cells derived from three independent donors. We also observe crisis-free immortalization of two cell lines (NC104-E6E7 and NC105-E6E7; Supplementary Fig. S1A and B), but another cell line, NC106-E6E7, had a severe crisis lasting for ∼6 weeks (Supplementary Fig. S1C). The three cell lines have undergone >100 PDs and are regarded as immortalized. All of our control cells infected with the empty vector LXSN had a finite in vitro life span of ≤12 PDs. Telomerase activity in the two crisis-free cell lines (NC104-E6E7 and NC105-E6E7) was readily detectable at an early PD (PD6, just 2 PDs after the selection for E6E7-expressing cells; Fig. 1A). The telomerase activity also showed moderate progressive increase with PDs. Notably, NC106-E6E7 cells had remarkably lower telomerase activity than the two crisis-free cell lines at early PDs (PD6 and PD20). Because only critically short telomeres that usually lack telomere signals detected by quantitative FISH, and not average telomere lengths, are crucial to chromosomal instability (25, 31, 32), the frequencies of telomere signal–free ends were quantitated. We found low frequencies (0–0.2%) of telomere signal–free ends in both NC104-E6E7 and NC105-E6E7 cells throughout the whole process of immortalization (Table 1). Spectral karyotyping (SKY) analyses showed that the frequencies of chromosomes involved in de novo (nonclonal) end-to-end fusions in both cell lines ranged from 0 to 8 per 100 metaphases at multiple PDs (Table 1). The frequencies of other nonclonal structural aberrations including nonreciprocal translocations, deletions, duplications, and intra-arm dicentrics ranged from 12 to 16 per 100 metaphases. These results indicate that only minor structural chromosomal instability persisted during immortalization of the two crisis-free cervical epithelial cell lines expressing HPV16 E6E7. During the process of immortalization, some cells may gain growth advantage and undergo clonal expansion after acquiring certain aberrations. In NC104-E6E7 cells, a clonal derivative chromosome generated by a rearrangement involving two chromosomes was found at PD14 and the same sole clonal aberration was observed at PD70 (Table 1; Supplementary Fig. S1D). Surprisingly, no clonal structural aberration was found in NC105-E6E7 cells until PD70, at which only a clonal chromosome arm loss was detected (Table 1; Supplementary Fig. S1D). This further shows the low extent of structural chromosomal instability in the two crisis-free cell lines during immortalization.

Figure 1.

Telomerase activity detected by telomerase repeat amplification protocol assay. The internal control (IC) was used for semiquantitative analysis of telomerase activity; the bands were weak or almost invisible in cells with excessively strong telomere activity due to the semicompetitive amplification of the telomerase repeat amplification protocol products and the internal control.

Figure 1.

Telomerase activity detected by telomerase repeat amplification protocol assay. The internal control (IC) was used for semiquantitative analysis of telomerase activity; the bands were weak or almost invisible in cells with excessively strong telomere activity due to the semicompetitive amplification of the telomerase repeat amplification protocol products and the internal control.

Close modal
Table 1.

Effects of TGF-β1 on telomere shortening and chromosomal instability

Cell linePDTelomere signal–free ends (%)Chr. in end-fusions/100 cellsChr. in other nonclonal structural aberrations/100 cells*ABF (%)Chr. in clonal structural aberrations/100 cells
NC104-E6E7 14 0.2 12 
 28 0.1 14 
 40 0.1 15 
 70 16 
NC105-E6E7 14 0.1 16 
 24 0.06 14 
 40 0.04 12 
 70 15 
NC104-E6E7(TGF-β1) 14 3.3 74 23 13 
 28 9.0 356 174 30 
 40 0.1 18 10 
 70 0.1 10 20 10 
NC105-E6E7(TGF-β1) 14 2.6 82 31 11 
 24 8.1 307 212 35 
 40 0.2 10 16 11 
 70 0.1 14 13 
Cell linePDTelomere signal–free ends (%)Chr. in end-fusions/100 cellsChr. in other nonclonal structural aberrations/100 cells*ABF (%)Chr. in clonal structural aberrations/100 cells
NC104-E6E7 14 0.2 12 
 28 0.1 14 
 40 0.1 15 
 70 16 
NC105-E6E7 14 0.1 16 
 24 0.06 14 
 40 0.04 12 
 70 15 
NC104-E6E7(TGF-β1) 14 3.3 74 23 13 
 28 9.0 356 174 30 
 40 0.1 18 10 
 70 0.1 10 20 10 
NC105-E6E7(TGF-β1) 14 2.6 82 31 11 
 24 8.1 307 212 35 
 40 0.2 10 16 11 
 70 0.1 14 13 

Abbreviations: Chr., chromosomes; ABF, anaphase bridge frequency.

*

Including nonreciprocal translocations, deletions, duplications, and intra-arm dicentrics.

Crisis.

In sharp contrast, extensive telomere-mediated chromosomal instability occurred in NC106-E6E7 cells during crisis, evidenced by the involvement of 419 and 191 chromosomes in nonclonal end-to-end fusions and other nonclonal structural aberrations, respectively, in 100 metaphases (Table 2). In particular, the detailed distributions of nonclonal structural aberrations were highly significantly correlated with the profiles of telomere signal–free ends (P < 0.001; Fig. 2A). Moreover, in post-crisis cells, eight of nine chromosomes that were involved in clonal structural aberrations (Fig. 3A) were among those having the highest frequencies of telomere signal–free ends during crisis. These results show that telomere dysfunction played a significant role in the induction of massive chromosomal instability during crisis as well as in the complex karyotype outcome in post-crisis NC106-E6E7 cells, manifesting a striking difference compared with the two crisis-free cell lines.

Table 2.

Telomere dysfunction and chromosomal instability in NC106-E6E7 cells

Cell linePDTelomere signal–free ends (%)Chr. in end-fusions/100 cellsChr. in other nonclonal structural aberrations/100 cells*ABF (%)Chr. in clonal structural aberrations/100 cells
NC106-E6E7 14 2.6 96 37 12 
 26 8.4 419 191 25 
 40 0.3 22 15 
 70 0.07 13 
NC106-E6E7 (SB-431542) 14 0.1 14 
 26 0.2 10 18 
 40 0.1 12 16 
 70 0.05 12 
Cell linePDTelomere signal–free ends (%)Chr. in end-fusions/100 cellsChr. in other nonclonal structural aberrations/100 cells*ABF (%)Chr. in clonal structural aberrations/100 cells
NC106-E6E7 14 2.6 96 37 12 
 26 8.4 419 191 25 
 40 0.3 22 15 
 70 0.07 13 
NC106-E6E7 (SB-431542) 14 0.1 14 
 26 0.2 10 18 
 40 0.1 12 16 
 70 0.05 12 
*

Including nonreciprocal translocations, deletions, duplications, and intra-arm dicentrics.

Crisis.

Figure 2.

Frequencies of telomere signal–free ends and nonclonal structural aberrations (including end-to-end fusions) on individual chromosome arms during crisis. A, NC106-E6E7; the correlation coefficient between the two parameters was 0.95 (P < 0.001). B, NC105-E6E7 cells treated with TGF-β1; the correlation coefficient between the two parameters was 0.92 (P < 0.001). C, NC104-E6E7(TGF-β1) cells; the correlation coefficient between the two parameters was 0.94 (P < 0.001).

Figure 2.

Frequencies of telomere signal–free ends and nonclonal structural aberrations (including end-to-end fusions) on individual chromosome arms during crisis. A, NC106-E6E7; the correlation coefficient between the two parameters was 0.95 (P < 0.001). B, NC105-E6E7 cells treated with TGF-β1; the correlation coefficient between the two parameters was 0.92 (P < 0.001). C, NC104-E6E7(TGF-β1) cells; the correlation coefficient between the two parameters was 0.94 (P < 0.001).

Close modal
Figure 3.

SKY karyotypes. A, post-crisis NC105-E6E7 cells. Red arrows, chromosomes with structural aberrations. Karyotype description: 44,XX,del(10)(p11),der(11;15)(q10;q10),der(16;20)(q10;q10),-18,+20, der(21)t(8;21)(q21;p11),der(21)t(8;21)(p/q?;qt). B, NC105-E6E7 cells treated with TGF-β1. Left, typical examples of chromosome end-to-end fusions. Note the lack of telomere signals at both fusion points (arrows). Right, SKY image of the corresponding metaphase. C, a separate metaphase of NC105-E6E7 treated with TGF-β1 showing the lack of telomere signals on intact chromosome ends by sequential quantitative FISH and SKY. The four chromosome ends indicated by white arrows on the left were the same as those involved in fusions shown on the right of B. Also shown in this partial metaphase is the lack of telomere signals on 3p and 8p (green arrows), which were frequently involved in end-fusions shown in Fig. 2B. D, SKY karyotype of TGF-β1–treated post-crisis NC105-E6E7 cells. Arrows, chromosomes with structural aberrations. Karyotype description: 46,XX,der(3;21)(q10;q10),der(8;19)(q10;p10),+19, der(19)t(19;20)(pt;q?)t(17;20)t(q22;q?),+20,der(20)dup(20)(q11q12)t(17;20)(q23;qt),der(22)t(8;22)(q21;qt).

Figure 3.

SKY karyotypes. A, post-crisis NC105-E6E7 cells. Red arrows, chromosomes with structural aberrations. Karyotype description: 44,XX,del(10)(p11),der(11;15)(q10;q10),der(16;20)(q10;q10),-18,+20, der(21)t(8;21)(q21;p11),der(21)t(8;21)(p/q?;qt). B, NC105-E6E7 cells treated with TGF-β1. Left, typical examples of chromosome end-to-end fusions. Note the lack of telomere signals at both fusion points (arrows). Right, SKY image of the corresponding metaphase. C, a separate metaphase of NC105-E6E7 treated with TGF-β1 showing the lack of telomere signals on intact chromosome ends by sequential quantitative FISH and SKY. The four chromosome ends indicated by white arrows on the left were the same as those involved in fusions shown on the right of B. Also shown in this partial metaphase is the lack of telomere signals on 3p and 8p (green arrows), which were frequently involved in end-fusions shown in Fig. 2B. D, SKY karyotype of TGF-β1–treated post-crisis NC105-E6E7 cells. Arrows, chromosomes with structural aberrations. Karyotype description: 46,XX,der(3;21)(q10;q10),der(8;19)(q10;p10),+19, der(19)t(19;20)(pt;q?)t(17;20)t(q22;q?),+20,der(20)dup(20)(q11q12)t(17;20)(q23;qt),der(22)t(8;22)(q21;qt).

Close modal

TGF-β1 inhibited telomerase activity and induced telomere-mediated chromosomal instability in NC104-E6E7 and NC105-E6E7 cells. Clearly, the low extent of chromosome instability and few clonal structural aberrations at late PDs in the two crisis-free cervical epithelial cell lines are not compatible with the complex chromosomal aberrations in premalignant cervical lesions and cervical cancers. We then tested the hypothesis that TGF-β1 may promote chromosomal instability in HPV16 E6E7–infected cervical epithelial cells. NC104-E6E7 and NC105-E6E7 cells were fed with medium containing TGF-β1 (0.3 and 0.5 ng/mL for NC105-E6E7 and NC104-E6E7 cells, respectively) every other day for ∼10 months. The said doses were chosen because they were well within the physiologic range of active TGF-β1 (0.03–4.0 ng/mL) in healthy subjects (33, 34). As shown in Fig. 1B, both NC104-E6E7 and NC105-E6E7 cells treated with TGF-β1 showed remarkably lower telomerase activity than the cells without TGF-β1 treatment at early PDs (PDs 12 and 22). In addition, the TGF-β1–treated cells, designated NC104-E6E7(TGF-β1) and NC105-E6E7(TGF-β1), succumbed to a crisis lasting ∼1 month at PD28 and PD24, respectively (Supplementary Fig. S1A and B). Concomitantly, the frequencies of telomere signal–free ends, end-to-end fusions, and other nonclonal structural aberrations increased progressively until the cells survived crisis (Table 1). The frequency of chromosomes involved in end-to-end fusions and other nonclonal structural aberrations in NC105-E6E7(TGF-β1) cells during crisis increased by 50- and 15-fold, respectively, compared with untreated cells at the same PD. Similar results were obtained in NC104-E6E7(TGF-β1) cells. Because anaphase bridges are associated with, although not specific for, telomere dysfunction (35), we quantitated the frequencies of anaphase bridges, which increased to 30% and 35% in NC104-E6E7(TGF-β1) and NC105-E6E7(TGF-β1) cells, respectively, during crisis, from the original ≤5% in both cell lines without TGF-β1 treatment (Table 1).

More specifically, the chromosome arms involved in structural aberrations in both TGF-β1–treated cell lines were significantly (P < 0.001) correlated with the specific profiles of telomere signal–free ends during crisis (Fig. 2B and C). In fact, further examinations showed that ∼80% of fusion points lacked telomere signals. Typical examples of chromosomal end-to-end fusions were shown in Fig. 3B. The chromosome arms involved in end-fusions in Fig. 3B had barely detectable telomeres at intact ends as detected in a separate metaphase shown in Fig. 3C. These results show that physiologic levels of TGF-β1 effectively induced telomere-mediated chromosomal instability up to cellular crisis in HPV16 E6E7–expressing cervical epithelial cells.

Effect of TGF-β1 on the karyotype outcomes in post-crisis cells. Both NC104-E6E7(TGF-β1) and NC105-E6E7(TGF-β1) cell lines eventually survived crisis in the presence of TGF-β1. Strong telomerase activity was detected in the TGF-β1–treated post-crisis cells measured at PD30 and PD60 (Fig. 1B) and showed little difference from that in the untreated cells. The frequencies of telomere signal–free ends, end-to-end fusions, other nonclonal structural aberrations, and anaphase bridges in the post-crisis cells declined significantly (P < 0.5) compared with crisis cells (Table 1), indicating the attenuation of ongoing structural instability in the cells that emerged from crisis. Strikingly, ≥10 chromosomes were involved in multiple clonal structural aberrations in those post-crisis cells (Table 1; Fig. 3D; Supplementary Fig. S2). The sets of chromosomes involved in the clonal structural aberrations were mainly among those carrying critically short telomeres during crisis, such as chromosomes 8, 10, 12, 14, 21, and 22 in NC104-E6E7(TGF-β1) and chromosomes 3, 8, 19, 20, 21, and 22 in NC105-E6E7(TGF-β1) cells. We conclude that TGF-β1–induced telomere dysfunction in the early process of immortalization (up to cellular crisis) had a profound long-term effect on the complex karyotype outcomes in post-crisis (immortalized) cells after TGF-β1 treatment.

TGF-β1 did not affect HPV16 E6E7 expression. We next asked whether telomere dysfunction induced by TGF-β1 was related to HPV16 E6E7 expression. Although TGF-β1 has been shown to reduce E6 and E7 mRNA expression in a full-length HPV16-immortalized cervical cell line (36), we did not observe a similar phenomenon in either NC104-E6E7 or NC105-E6E7 cell line (Fig. 4A) at various time points ranging from 24 to 72 h after TGF-β1 treatment in pre-crisis cells, which showed reduced telomerase activity. In addition, we found no significant differences in HPV16 E7 protein expression between the TGF-β1–treated and untreated cells (Fig. 4B). Because we could not detect HPV16 E6 protein expression using commercially available antibodies from several companies, yet E6 targets p53 protein for degradation (3, 37), we thus monitored p53 protein expression and found no significant changes after TGF-β1 treatments (Fig. 4B). By contrast, reduced p53 protein expression was readily detected in E6E7-expressing cells compared with normal cervical epithelial cells infected with empty vector. These results exclude the possibility that our HPV16 E6E7 retroviral gene expression system per se was subject to the regulation of TGF-β1. For comparison, TGF-β1 increased phospho-Smad3 and decreased c-Myc protein expression in the pre-crisis cells (Fig. 4B), in agreement with the results in noncervical cells (22).

Figure 4.

Protein and mRNA expression. A, RT-PCR analysis for HPV16 E6 and E7 in cells with (+) or without (−) TGF-β1 treatment. GAPDH serves as internal control. Cervical cancer cell line CaSki and normal cervical epithelial (NL) cells served as positive and negative controls, respectively, for E6 and E7 expression. B, Western blot analysis for HPV16 E7, p53, phospho-Smad3 (P-Smad3), and c-Myc. NL, normal proliferating NC104 (left) and NC105 (right) cells at PD6 for comparison. Actin served as an internal control. C, RT-PCR analysis for E6 and Western blot analysis for p53, TGF-β, phospho-Smad3, c-Myc, and actin after 24-h TGF-β1 treatment. D, autocrine TGF-β1 in conditioned media (after 48-h cell culture) analyzed using human cytokine antibody array. POS, positive controls.

Figure 4.

Protein and mRNA expression. A, RT-PCR analysis for HPV16 E6 and E7 in cells with (+) or without (−) TGF-β1 treatment. GAPDH serves as internal control. Cervical cancer cell line CaSki and normal cervical epithelial (NL) cells served as positive and negative controls, respectively, for E6 and E7 expression. B, Western blot analysis for HPV16 E7, p53, phospho-Smad3 (P-Smad3), and c-Myc. NL, normal proliferating NC104 (left) and NC105 (right) cells at PD6 for comparison. Actin served as an internal control. C, RT-PCR analysis for E6 and Western blot analysis for p53, TGF-β, phospho-Smad3, c-Myc, and actin after 24-h TGF-β1 treatment. D, autocrine TGF-β1 in conditioned media (after 48-h cell culture) analyzed using human cytokine antibody array. POS, positive controls.

Close modal

Rad51 expression was not affected by TGF-β1 treatment. A high concentration (5 ng/mL) of TGF-β1 was shown to down-regulate the DNA repair protein Rad51 in mink epithelial cells (38). We analyzed Rad51 protein expression level in our cervical epithelial cell lines under the condition of relatively low levels (0.3–0.5 ng/mL) of TGF-β1 used in this study. No reduction was detected compared with the untreated cells (Supplementary Fig. S3), suggesting that the TGF-β1–induced chromosomal instability observed in this study was not through Rad51 down-regulation.

Ectopic hTERT overexpression abolished TGF-β1–induced chromosomal instability. To confirm the direct role of telomerase inhibition by TGF-β1 in the induction of chromosomal instability, we tested whether ectopic overexpression of hTERT could abolish TGF-β1–induced chromosome aberrations. NC104-E6E7 and NC105-E6E7 cells were infected with a retroviral vector encoding hTERT at PD8. Both early and late PD cells showed strong telomerase activity, which was no longer inhibited by TGF-β1 (Fig. 1B). The cells treated with TGF-β1 had undergone >100 PDs without crisis (Supplementary Fig. S1A and B), and their chromosomes presented no telomere signal–free ends in 20 metaphases and no end-to-end fusions in 100 metaphases analyzed at PDs 14, 40, and 70. The frequencies of anaphase bridges and other nonclonal structural aberrations in E6E7 + hTERT–expressing cells remained low (≤5%) after TGF-β1 treatment (Supplementary Table S1). Moreover, the number of chromosomes (≤2) involved in clonal structural aberrations did not increase after long-term TGF-β1 exposure (Supplementary Table 1). These results showed that ectopic hTERT overexpression in HPV16 E6E7–expressing cervical epithelial cells completely abolished the effects of TGF-β1 on telomere-mediated chromosomal instability throughout the long-term cell culture.

Elevated autocrine production of TGF-β1 is responsible for the crisis and chromosomal instability in NC106-E6E7 cells. A question was raised about why NC106-E6E7 cells differed from NC104-E6E7 and NC105-E6E7 cells in telomere dysfunction. We first checked if NC106-E6E7 cells had lower E6 expression, which might lead to lower telomerase activity, than the other two cell lines in early-passage cells. Our results of E6 mRNA expression showed that this was not the case, which was also supported by the similar levels of p53 protein expression (Fig. 4C). Western blotting analysis of whole-cell lysate showed that NC106-E6E7 had significantly higher expression of TGF-β, the precursor of TGF-β1, than the other two cell lines (Fig. 4C). Because TGF-β1 needs to be transported outside the cells where it exerts its function, we then collected the conditioned media of the three cell lines for detection of TGF-β1 using a cytokine antibody array. The array was used because of its high sensitivity in protein expression detection. It was found that the conditioned medium from NC106-E6E7 cells had a remarkably higher level of TGF-β1 than the other two cell lines (Fig. 4D). In line with this, NC106-E6E7 cells had much higher phospho-Smad3 and lower c-Myc protein expression than the other two cell lines (Fig. 4C).

It is known that TGF-β1 functions by binding to its transmembrane type II receptor, which phosphorylates and activates its type I receptor, TβRI (39). The activated TβRI phosphorylates Smad3, which then enters into the nucleus and directly inhibits c-Myc and telomerase activity (22). To further examine whether the elevated autocrine TGF-β1 in NC106-E6E7 cells was truly responsible for the induction of chromosomal instability, we used a specific TGF-β1 type I receptor inhibitor, SB-431542 (40), to block the TGF-β1 signaling pathway to see if this could reduce the telomere-mediated chromosomal instability. After SB-431542 (10 μmol/L) treatment for 24 h, phospho-Smad3 protein was decreased and c-Myc was increased as expected (Fig. 4C). Meanwhile, elevated telomerase activity was detected at PD6 and PD20 under continuous SB-431542 treatment compared with that at respective PDs without the treatment (lanes 16 and 17 versus lanes 12 and 13 in Fig. 1A). Moreover, the SB-431542–treated cells have undergone 75 PDs without crisis (Supplementary Fig. S1C). The frequencies of telomere signal–free ends, chromosome end-to-end fusions, and other nonclonal structural aberrations, as well as anaphase bridges, at PDs 14 and 26 decreased dramatically compared with those without the treatment at the same PDs (P ≤ 0.5 for each parameter; Table 2). Throughout the long-term culture, the sole clonal structural aberration was a single derivative chromosome (Supplementary Fig. S4A), observed at PDs 26, 40, and 70 in 6%, 21%, and 74% of metaphases, respectively, indicating the expansion of the stable clone. Amazingly, a completely normal karyotype (46,XX) was still observed in 7% of the metaphases at PD70, although the percentage decreased from 38% at PD14. Another population of cells had a gain of chromosome 20 but no clonal structural aberration (Supplementary Fig. S4B). We conclude that blocking the TGF-β1 pathway significantly reduced the frequencies of telomere dysfunction–induced structural chromosome aberrations. Taken together, these data show a crucial role of autocrine TGF-β1 in the induction of telomere dysfunction and its related chromosome instability in this cell line.

TGF-β1 expression was associated with telomere shortening in CIN specimens. Because the TGF-β1 signaling pathway is often inactivated in malignancies and our results in vitro showed that TGF-β1 induced telomere dysfunction in pre-crisis cells before immortalization, we anticipated that the role of TGF-β1 in telomere shortening could be observed during the early process of tumor development. To test this, we evaluated the expression of TGF-β1, phospho-Smad2/3, and c-Myc, as well as of telomere signals, in CIN specimens from 20 patients. Strong immunostaining for TGF-β1, predominantly in the cytoplasm, was found in 17 of the specimens. This was associated with a marked activation of downstream TGF-β1 signaling as indicated by nuclear immunoreactivity for phospho-Smad2/3 in the majority of CIN cells. Interestingly, c-Myc staining was mainly observed in the cytoplasm with few nuclear positive in these cells (see typical images in Supplementary Fig. S5A; only nuclear c-Myc is expected to be functional). The average ratio of telomere signal intensity in these CIN cells to that in stromal cells was 0.32 ± 0.08 (Supplementary Fig. S5B; Table S2), showing significant telomere shortening in these CINs. However, we also observed weak TGF-β1 immunostaining in CIN regions from the other three patients (exemplified in Supplementary Fig. S5C). In these regions, <30% of cells were positive for nuclear staining of phospho-Smad2/3, whereas >50% of cells were positive for nuclear c-Myc staining; the average ratio of telomere signal intensity in the CIN to that in stromal cells was 0.54 ± 0.08, showing a significantly lesser extent of telomere shortening (P ≤ 0.05) than CINs with strong TGF-β1 expression (Supplementary Fig. S5D; Table S2). These results together reveal that up-regulation of TGF-β1 and activation of its downstream signaling pathway are associated with more dramatic telomere shortening in the carcinogenesis of cervical cancer.

To our knowledge, this is the first report providing direct evidence that TGF-β1 promotes telomere-mediated chromosomal instability during immortalization of human epithelial cells. TGF-β1 is not only widely distributed throughout the body in normal physiologic conditions but is also one of the major cytokines and growth factors involved in many pathologic processes including cell proliferation, differentiation, inflammation, and carcinogenesis (18, 29). Because chronic inflammation is an important cofactor for human cervical cancer (16, 17), our finding that TGF-β1 induces chromosomal instability through telomere dysfunction may provide a novel link between chronic inflammation and human cervical carcinogenesis.

Cervical carcinogenesis is tightly associated with the expression of high-risk HPV E6 and E7, which have been found to be able to cause not only numerical chromosomal instability but also DNA damage independent of telomere dysfunction in foreskin epithelial cells (41). We observed a similar phenomenon in our cervical epithelial cells expressing HPV16 E6E7 or HPV16 E6E7 + hTERT (data not shown). Interestingly, however, only minor structural chromosomal instability was detected in the cervical epithelial cell lines expressing HPV16 E6E7 + hTERT and two of three cell lines expressing HPV16 E6E7. These findings are in agreement with several previous studies on noncervical cells types (28, 42, 43).

It seems that besides DNA damage caused by high-risk HPV E6 and E7, there must be some other prominent inducers of chromosomal instability during the process of cervical carcinogenesis. Here we have shown that chronic exposure to physiologic levels of TGF-β1 induced severe cellular crisis in the two cell lines that originally became immortalized without crisis. The TGF-β1–induced crisis was associated with extensive telomere dysfunction evidenced by the frequent formation of chromosomal end-to-end fusions and other structural aberrations preferentially on chromosomes carrying critically short telomeres. The early activation of telomerase by ectopic overexpression of hTERT completely abolished the effects of TGF-β1, supporting that the telomerase inhibition by TGF-β1 was directly responsible for the induction of chromosomal instability. Because the post-crisis cells from both NC104-E6E7(TGF-β1) and NC105-E6E7(TGF-β1) cell lines eventually had strong telomerase activity after the acquisition of multiple chromosome aberrations, they became resistant to TGF-β1 with respect to further telomere-mediated chromosomal instability (Table 1).

Based on our experience of crisis-free immortalization of NC104-E6E7 and NC105-E6E7 under ordinary culture condition (without exogenous TGF-β1), it was surprising that another cervical epithelial cell line, NC106-E6E7, derived from an independent donor, showed a severe crisis under the same culture condition. Interestingly, a previous article also reported that HPV16-immortalized cervical epithelial cell lines derived from one of three individuals experienced cellular crisis for unknown reasons (30). Our discovery that exogenous TGF-β1 induced chromosomal instability motivated us to test whether this NC106-E6E7 cell line had strong autocrine TGF-β1 that might cause cellular crisis and chromosomal instability. Our analysis showed that the level of TGF-β1 in the conditioned medium after NC106-E6E7 cell culture was indeed higher than that after NC104-E6E7 and NC105-E6E7 cell culture. Concomitantly, NC106-E6E7 cells had higher expression of phospho-Smad3, lower expression of c-Myc, lower telomerase activity, and more severe telomere dysfunction–mediated chromosomal instability than the other two cell lines. Conversely, blocking the TGF-β1 pathway using a specific TGF-β1 type I receptor inhibitor (SB-431542) reversed these effects. The role of differential levels of autocrine TGF-β1 in modulating immortalization and chromosomal instability may provide a novel explanation to the mystery of why some epithelial cell lines expressing HPV16 E6E7 could be immortalized without crisis but some could not, even under the same culture conditions.

The finding that TGF-β1 induced telomere shortening and made the cells enter into cellular crisis may have important implications to the fate of premalignant cervical lesions. Indeed, we found that higher expression of TGF-β1 and activation of its downstream signaling pathway were associated with more dramatic telomere shortening in CIN specimens. About one to two thirds of women with premalignant high-grade CIN will develop invasive cervical cancer after decades of latency (44). The telomere-mediated cellular crisis in vitro might be reminiscent of an important part of long-term precancerous state in vivo, from which some lesions may progress to cancer. Although crisis is a potent barrier to immortal growth, it is recognized that the extensive chromosomal instability in this state may well be the mechanism by which rare cells acquire enough genetic alterations that provide the impetus for further malignant transformation (6). We therefore propose that TGF-β1 may have an important effect on the progression of premalignant cervical lesions through the induction of telomere dysfunction. Because immortalization in vivo may well represent a stage of no-return in cervical carcinogenesis (45), our finding on the role of TGF-β1 in the production of complex clonal chromosomal aberrations in immortalized cells may shed some light on the complex karyotypes in cervical cancers. It is tempting to speculate that TGF-β1 may modulate cervical cancer risk considering its tight association with telomere erosion and chromosomal instability. Further in vivo studies on this issue are warranted.

No potential conflicts of interest were disclosed.

Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

Grant support: Research Grants Council of Hong Kong Special Administrative Region, China, Project No. HKU 7556/06M (A.L.M. Cheung); University of Hong Kong CRCG #200507176189 (W. Deng and A.L.M. Cheung).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Drs. P.M. Lansdorp and S.S.S. Poon (BC Cancer Research Center, Canada) for the quantitative FISH software (TFl-TELO); the Department of Pediatrics and Adolescent Medicine, The University of Hong Kong, for use of SKY facilities; and Chee-Wai Chua, Jenny Cheung, Tony Chan, and Alla Li, Department of Anatomy, for technical assistance.

1
Woodman CB, Collins SI, Young LS. The natural history of cervical HPV infection: unresolved issues.
Nat Rev Cancer
2007
;
7
:
11
–22.
2
zur Hausen H. Papillomaviruses and cancer: from basic studies to clinical application.
Nat Rev Cancer
2002
;
2
:
342
–50.
3
Munger K, Howley PM. Human papillomavirus immortalization and transformation functions.
Virus Res
2002
;
89
:
213
–28.
4
Duesberg P. Does aneuploidy or mutation start cancer?
Science
2005
;
307
:
41
.
5
Lengauer C, Kinzler KW, Vogelstein B. Genetic instabilities in human cancers.
Nature
1998
;
396
:
643
–9.
6
Maser RS, DePinho RA. Connecting chromosomes, crisis, and cancer.
Science
2002
;
297
:
565
–9.
7
Stewart SA, Weinberg RA. Telomeres: cancer to human aging.
Annu Rev Cell Dev Biol
2006
;
22
:
531
–57.
8
Klingelhutz AJ, Foster SA, McDougall JK. Telomerase activation by the E6 gene product of human papillomavirus type 16.
Nature
1996
;
380
:
79
–82.
9
Kiyono T, Foster SA, Koop JI, et al. Both Rb/p16INK4a inactivation and telomerase activity are required to immortalize human epithelial cells.
Nature
1998
;
396
:
84
–8.
10
Veldman T, Horikawa I, Barrett JC, Schlegel R. Transcriptional activation of the telomerase hTERT gene by human papillomavirus type 16 E6 oncoprotein.
J Virol
2001
;
75
:
4467
–72.
11
Baege AC, Berger A, Schlegel R, Veldman T, Schlegel R. Cervical epithelial cells transduced with the papillomavirus E6/E7 oncogenes maintain stable levels of oncoprotein expression but exhibit progressive, major increases in hTERT gene expression and telomerase activity.
Am J Pathol
2002
;
160
:
1251
–7.
12
Meeker AK, Hicks JL, Iacobuzio-Donahue CA, et al. Telomere length abnormalities occur early in the initiation of epithelial carcinogenesis.
Clin Cancer Res
2004
;
10
:
3317
–26.
13
Zhang A, Wang J, Zheng B, et al. Telomere attrition predominantly occurs in precursor lesions during in vivo carcinogenic process of the uterine cervix.
Oncogene
2004
;
23
:
7441
–7.
14
Umayahara K, Numa F, Suehiro Y, et al. Comparative genomic hybridization detects genetic alterations during early stages of cervical cancer progression.
Genes Chromosomes Cancer
2002
;
33
:
98
–102.
15
Costa C, Fuste P, Alameda F, et al. Study of chromosomal abnormalities in 11 cases of cervical dysplasia using comparative genomic hybridization on cotton-lint cervical samples.
Cancer Genet Cytogenet
2006
;
164
:
61
–5.
16
Castle PE, Giuliano AR. Chapter 4: Genital tract infections, cervical inflammation, and antioxidant nutrients-assessing their roles as human papillomavirus cofactors.
J Natl Cancer Inst Monogr
2003
;
31
:
29
–34.
17
Hammes LS, Tekmal RR, Naud P, et al. Macrophages, inflammation and risk of cervical intraepithelial neoplasia (CIN) progression-clinicopathological correlation.
Gynecol Oncol
2007
;
105
:
157
–65.
18
Coussens LM, Werb Z. Inflammation and cancer.
Nature
2002
;
420
:
860
–7.
19
Cerezo A, Kalthoff H, Schuermann M, Schafer B, Boukamp P. Dual regulation of telomerase activity through c-Myc-dependent inhibition and alternative splicing of hTERT.
J Cell Sci
2002
;
115
:
1305
–12.
20
Yang H, Kyo S, Takatura M, Sun L. Autocrine transforming growth factor β suppresses telomerase activity and transcription of human telomerase reverse transcriptase in human cancer cells.
Cell Growth Differ
2001
;
12
:
119
–27.
21
Katakura Y, Nakata E, Miura T, Shirahata S. Transforming growth factor β triggers two independent-senescence programs in cancer cells.
Biochem Biophys Res Commun
1999
;
255
:
110
–5.
22
Li H, Xu D, Li J, Berndt MC, Liu JP. Transforming growth factor β suppresses human telomerase reverse transcriptase (hTERT) by Smad3 interactions with c-Myc and the hTERT gene.
J Biol Chem
2006
;
281
:
25588
–600.
23
McMurray HR, McCance DJ. Human papillomavirus type 16 E6 activates TERT gene transcription through induction of c-Myc and release of USF-mediated repression.
J Virol
2003
;
77
:
9852
–61.
24
Deng W, Tsao SW, Lucas JN, Leung CS, Cheung AL. A new method for improving metaphase chromosome spreading.
Cytometry A
2003
;
51
:
46
–51.
25
Deng W, Tsao SW, Guan XY, et al. Distinct profiles of critically short telomeres are a key determinant of different chromosome aberrations in immortalized human cells: whole-genome evidence from multiple cell lines.
Oncogene
2004
;
23
:
9090
–101.
26
Zhang H, Jin Y, Chen X, et al. Papillomavirus type 16 E6/E7 and human telomerase reverse transcriptase in esophageal cell immortalization and early transformation.
Cancer Lett
2007
;
245
:
184
–94.
27
Lan HY, Mu W, Nikolic-Paterson DJ, Atkins RC. A novel, simple, reliable, and sensitive method for multiple immunoenzyme staining: use of microwave oven heating to block antibody cross-reactivity and retrieve antigens.
J Histochem Cytochem
1995
;
43
:
97
–102.
28
Klingelhutz AJ, Qian Q, Phillips SL, et al. Amplification of the chromosome 20q region is associated with expression of HPV-16 E7 in human airway and anogenital epithelial cells.
Virology
2005
;
340
:
237
–44.
29
Glick AB, Perez-Lorenzo R, Mohammed J. Context-dependent regulation of cutaneous immunological responses by TGFβ1 and its role in skin carcinogenesis.
Carcinogenesis
2008
;
29
:
9
–14.
30
Woodworth CD, Bowden PE, Doniger J, et al. Characterization of normal human exocervical epithelial cells immortalized in vitro by papillomavirus types 16 and 18 DNA.
Cancer Res
1988
;
48
:
4620
–8.
31
Hemann MT, Strong MA, Hao LY, Greider CW. The shortest telomere, not average telomere length, is critical for cell viability and chromosome stability.
Cell
2001
;
107
:
67
–77.
32
Der-Sarkissian H, Bacchetti S, Cazes L, Londono-Vallejo JA. The shortest telomeres drive karyotype evolution in transformed cells.
Oncogene
2004
;
23
:
1221
–8.
33
Dziadzio M, Smith RE, Abraham DJ, Black CM, Denton CP. Circulating levels of active transforming growth factor β1 are reduced in diffuse cutaneous systemic sclerosis and correlate inversely with the modified Rodnan skin score.
Rheumatology (Oxford)
2005
;
44
:
1518
–24.
34
Ayatollahi M, Geramizadeh B, Samsami A. Transforming growth factor β-1 influence on fetal allografts during pregnancy.
Transplant Proc
2005
;
37
:
4603
–4.
35
Rudolph KL, Millard M, Bosenberg MW, DePinho RA. Telomere dysfunction and evolution of intestinal carcinoma in mice and humans.
Nat Genet
2001
;
28
:
155
–9.
36
Rorke EA, Zhang D, Choo CK, Eckert RL, Jacobberger JW. TGF-β-mediated cell cycle arrest of HPV16-immortalized human ectocervical cells correlates with decreased E6/E7 mRNA and increased p53 and p21(WAF-1) expression.
Exp Cell Res
2000
;
259
:
149
–57.
37
Howley PM. Role of the human papillomaviruses in human cancer.
Cancer Res
1991
;
51
:
5019
–22s.
38
Kanamoto T, Hellman U, Heldin CH, Souchelnytskyi S. Functional proteomics of transforming growth factor-β1-stimulated Mv1Lu epithelial cells: Rad51 as a target of TGFβ1-dependent regulation of DNA repair.
EMBO J
2002
;
21
:
1219
–30.
39
Massague J. How cells read TGF-β signals.
Nat Rev Mol Cell Biol
2000
;
1
:
169
–78.
40
Inman GJ, Nicolas FJ, Callahan JF, et al. SB-431542 is a potent and specific inhibitor of transforming growth factor-β superfamily type I activin receptor-like kinase (ALK) receptors ALK4, ALK5, and ALK7.
Mol Pharmacol
2002
;
62
:
65
–74.
41
Duensing S, Munger K. The human papillomavirus type 16 E6 and E7 oncoproteins independently induce numerical and structural chromosome instability.
Cancer Res
2002
;
62
:
7075
–82.
42
Mizumoto Y, Kyo S, Ohno S, et al. Creation of tumorigenic human endometrial epithelial cells with intact chromosomes by introducing defined genetic elements.
Oncogene
2006
;
25
:
5673
–82.
43
Akimov SS, Ramezani A, Hawley TS, Hawley RG. Bypass of senescence, immortalization, and transformation of human hematopoietic progenitor cells.
Stem Cells
2005
;
23
:
1423
–33.
44
Baseman JG, Koutsky LA. The epidemiology of human papillomavirus infections.
J Clin Virol
2005
;
32
Suppl 1:
S16
–24.
45
Steenbergen RD, de WJ, Wilting SM, et al. HPV-mediated transformation of the anogenital tract.
J Clin Virol
2005
;
32
Suppl 1:
S25
–33.