- Email: [email protected]
Impact of oxygenation status and patient age on DNA content in cancers of the uterine cervix
Int. J. Radiation Oncology Biol. Phys., Vol. 56, No. 4, pp. 929 –936, 2003 Copyright © 2003 Elsevier Inc. Printed in the USA. All rights reserved 0360-3016/03/$–see front matter
doi:10.1016/S0360-3016(03)00065-8
CLINICAL INVESTIGATION
Cervix
IMPACT OF OXYGENATION STATUS AND PATIENT AGE ON DNA CONTENT IN CANCERS OF THE UTERINE CERVIX ARNULF MAYER, M.D., DR. MED.,* MICHAEL HO¨ CKEL, M.D., DR. MED. HABIL., DR. RER. NAT.,† OLIVER THEWS, M.D., DR. MED. HABIL.,* KARLHEINZ SCHLENGER, DR. RER. NAT.,* AND PETER VAUPEL, M.D., DR. MED. HABIL.* *Institute of Physiology and Pathophysiology, University of Mainz, Mainz, Germany; †Department of Obstetrics and Gynecology, University of Leipzig, Leipzig, Germany Purpose: In carcinomas of the uterine cervix, the tumor oxygenation status has been shown to be a prognostic indicator that is independent of treatment modality. In vitro studies suggest gene amplification and polyploidization to be among the major consequences of hypoxia (with or without consecutive reoxygenation) and to be associated with treatment resistance and tumor progression. This study analyzed whether hypoxia alters net DNA content in uterine cervix cancer cells to the extent that it is identifiable by DNA image cytometry. Methods and Materials: In 64 patients with primary cervical cancer, tumor oxygenation was assessed polarographically and correlated with cell DNA content (DNA image cytometry) in areas adjacent to the oxygen microsensor tracks in which oxygenation measurements were made. Results: No correlation between DNA content (stemline position, Auer classification, and 2c deviation index) and oxygenation status was observed. However, an association between DNA content and patient age and menopausal status was found. Conclusion: Using DNA cytometry, hypoxia-associated genomic changes in uterine cervix cancer cells could not be detected. The impact of tumor hypoxia on the genome may be masked by the effects of alternative mechanisms of genomic instability that can also influence DNA content. © 2003 Elsevier Inc. Tumor oxygenation, DNA content, DNA image cytometry, Gene amplification, Ploidy.
Tumor oxygenation has been shown to be an adverse prognostic factor in primary and recurrent cancer of the uterine cervix (1– 6), as well as in other tumor entities (7, 8). For 50 years, tumor hypoxia has been considered a relevant factor causing “extrinsic” resistance to radiotherapy, because a reduced “oxygen enhancement effect” in hypoxic tumors diminishes the biologic effect of ionizing radiation (9). More recently, a number of studies have presented substantial evidence indicating that the impact of hypoxia on a patient’s prognosis may additionally be attributable to an enhanced “intrinsic” biologic aggressiveness and an acceleration of malignant progression of hypoxic tumors, predominantly because the influence on prognosis has been shown to be independent of treatment modality (2). Hypoxic tumor cells have also been shown to invade tissue faster and reach higher rates of perifocal, local, and distant spread (10). These adverse effects of hypoxia are thought to be mediated through at least two different pathogenetic
mechanisms. First, moderate hypoxia (⬍1% oxygen) has a profound impact on gene expression, which is mainly mediated through the oxygen-regulated transcription factor hypoxia-inducible factor-1␣, which activates genes involved in adaptation to anaerobic metabolism, angiogenesis, and capability of tissue invasion (11). Second, severe hypoxia (⬍0.1% oxygen) seems to contribute to the genomic instability of tumor cells. Hypoxia has been shown to reduce DNA repair mechanisms (12), resulting in point mutations, and to stimulate endonuclease activity (13) suitable to induce chromosomal double-strand breaks (14, 15), thus leading to structural chromosomal aberrations. Double-strand breaks have also been shown to be an important source for gene amplification by the initiation of breakage-fusionbridge cycles (16, 17). Additionally, hypoxia can induce polyploidy (18 –20) and has been suspected of causing aneuploidy by perturbations of the spindle apparatus (21). The chromosomal breaks occurring under hypoxic conditions seem not to be randomly distributed, but to occur
Reprint requests to: Arnulf Mayer, M.D., Institute of Physiology and Pathophysiology, University of Mainz, Duesbergweg 6, Mainz D-55099 Germany. Tel: ⫹49-6131-39 25203; Fax: ⫹49-6131-39 25774; E-mail: [email protected] Acknowledgments—The authors thank Dr. Debra K. Kelleher for
her assistance in preparing this manuscript, and Professor Dr. E. Schulte, Dept. of Anatomy, University of Mainz, for providing technical expertise. Received Oct 10, 2002, and in revised form Dec 26, 2002. Accepted for publication Jan 7, 2003.
INTRODUCTION
929
930
I. J. Radiation Oncology
● Biology ● Physics
preferentially at “fragile sites” (22) (i.e., specific chromosomal sites exhibiting an increased propensity for breakage). These fragile sites are often located in the vicinity of oncogenes or genes mediating drug resistance (23) and thus hypoxia-induced chromosomal aberrations and gene amplification may have some specificity for these crucial genes. Hypoxic cells have accordingly been shown to exhibit an increased resistance to chemotherapeutic agents (18, 24 – 26), as well as a greater metastatic potential (19). Because investigations of these mechanisms have only been carried out in vitro and possible interactions have not been examined, their contribution to the malignant progression of solid tumors in vivo remains unclear. The aforementioned hypoxia-related forms of genomic mutations (gene amplification, polyploidy, and chromosomal aberrations) have been shown to have a distinct effect on DNA content (27, 28). The aim of the present study therefore was to correlate the DNA parameters obtained by image cytometric DNA analysis with the oxygenation status of locally advanced cervical carcinoma. Importantly, with our method, tumor oxygenation and DNA content data originated from adjacent tissue areas. The results presented show that the DNA data are generally independent of the oxygenation status, a finding that is not in agreement with a recent article in this Journal (29). METHODS AND MATERIALS Patients All patients in this study were recruited from a prospective clinical trial for the evaluation of the significance of tumor oxygenation in primary, locally advanced carcinoma of the uterine cervix that has been conducted at the Department of Obstetrics and Gynecology of the University of Mainz since June 1989. Patients gave informed consent before enrollment, and a medical ethics committee approved the study design. Sixty-four patients from this group, for whom cytologic specimens (tumor cell smear preparations) were available, were included in the present study. Tumor oxygenation measurements Tumor oxygenation was assessed pretherapeutically with the computerized Eppendorf histography system (Eppendorf, Hamburg, Germany), using a protocol that has been described in detail earlier (30). In brief, PO2 readings were performed in the conscious patient along linear tracks, first in the s.c. fat of the mons pubis followed by cervical measurements at the 12- and 6-o’clock sites of the macroscopically vital tumor tissue. Within the tumor tissue, up to 35 PO2 measurements were taken in each electrode track (70 readings in total) starting at a tissue depth of 5 mm. The PO2 measuring points were situated 0.7 mm apart from each other, resulting in an overall measurement track length of approximately 25 mm. After oxygenation measurements, needle core biopsies approximately 2 mm in diameter and 20 mm in length were taken from those tumor areas at which the PO2 measurements had been obtained. The intra-
Volume 56, Number 4, 2003
vaginal temperature, blood pressure, heart rate, hemoglobin concentration, hematocrit, and arterial oxyhemoglobin saturation were monitored concomitantly with the PO2 readings. The pretherapeutic PO2 measurements were usually performed 1–5 days before oncologic treatment. Tumor cell preparations Smear preparations on standard histologic slides were obtained from the biopsy cylinders and were spray fixed immediately afterward (Merckofix, Merck, Darmstadt, Germany). In 59 of 64 cases, slides representing both standard positions (6 h and 12 h) of oxygenation measurement could be obtained. In the 5 remaining cases, either only one oxygenation measurement was performed or the biopsy cylinder seemed too fragile for the smearing procedure. Cytochemistry Feulgen staining was performed following a standardized protocol that met the requirements of the European Society for Analytical Cellular Pathology Task Force (31). After rehydration in distilled water, acid hydrolysis was performed using 5N HCl maintained at a constant temperature of 22°C. After rinsing in distilled water, the specimens were stained with Schiff’s reagent (Merck No. 9034-100) for 1 h. Afterward, excess dye was removed by rinsing in SO2 water, followed by dehydration in an ascending alcohol series. Cover slips were fixed using Eukitt (Merck), and the slides were then kept in the dark until measurement. DNA image cytometry The DNA content was measured using the CYDOK system (Technisches Buero Hilgers, Koenigswinter, Germany), consisting of a conventional light microscope (Leica DM LS, Leica Microsystems, Wetzlar, Germany), with a 40⫻ objective, interference filter of 550 ⫾ 8 nm half value, and a diameter of 25 mm (Schott, Mannheim, Germany), and a CCD monochrome camera (Pulnix TM7 CN with 768 ⫻ 494 pixels, Pulnix America, Sunnyvale, CA). Image analysis was performed on a standard personal computer equipped with a framegrabber video processor (MPV-AT, Matrox Electronic Systems, Dorval, Quebec, Canada), connected to a standard cathode-ray display terminal monitor. Measurements were performed and interpreted using the CYDOK software. Slides were taken from light-protected storage immediately before measurement and analyzed in one passage to avoid bleaching from normal daylight or the light of the microscope. Specimens were examined in a meander-like fashion to include all accessible tumor cells. The CYDOK system allows the separation of clustered tumor cells by user interaction, thereby reducing selective bias from this phenomenon, which is frequently found in cytologic preparations of cancer cells of the uterine cervix. The CYDOK software also corrected for glare error (32). From the areas containing the evaluated tumor cells, at least 20 neutrophils were selected as internal reference cells; 200 –300 tumor cells were analyzed in each preparation. For quantitative analysis, DNA histograms were generated and
Oxygenation and DNA content of cervical cancer
Table 1. Clinical tumor stages (FIGO) in 64 patients eligible for DNA cytometry Clinical tumor stage (FIGO)
Patients (n)
Ib IIa IIb IIIa IIIb IVa Unknown
17 1 28 1 15 1 1
Abbreviation: FIGO ⫽ International Federation of Gynecology and Obstetrics.
the DNA parameters stemline position, S-phase fraction (percentage of DNA values ⱖ2.5c and ⱕ3.8c), and the square deviation index of DNA values from the normal 2c-peak (2c deviation index [2cDI]) were calculated. For interpretation, DNA data were dichotomized into diploid (from 1.80c to 2.20c) vs. aneuploid (⬍1.80c or ⬎ 2.20c) stemlines. In addition, the DNA histograms were assigned to one of four classes according to the criteria defined by Auer et al. (33). Auer I is defined as a peak at 2c (DNA value corresponding to diploid cells) and a minor proportion of cells with a DNA content of 4c. Less than 5% of all cells are found between the 2c and 4c peaks (S-phase fraction). Histograms classified as Auer II showed either a single peak at 4c (Auer IIa) or peaks at 2c and 4c, with ⬎20% belonging to the latter peak (Auer IIb). Auer III histograms are similar to Auer I or II, but with ⬎5% of the cells lying between the 2c and 4c peaks. Auer IV histograms contain aneuploid peaks (for definition of aneuploid stemlines, see above). Statistical analysis The results are expressed as the mean ⫾ SEM. The differences between the groups were assessed by the twotailed Wilcoxon test for unpaired samples and Fisher’s exact test, as appropriate. The significance level was set at ␣ ⫽ 5% for all comparisons. Linear correlation between two parameters was described by Pearson’s correlation coefficient (r). RESULTS Cytologic specimens from 64 patients with primary cervical cancers who had undergone tumor oxygenation measurement were available for assessment of DNA content. The median patient age was 53 years (range 26 – 80). The groups of pre- and postmenopausal patients were equally large (32 patients). Histologic examination showed squamous cell carcinoma in 49 (77%), adenocarcinoma in 12 (19%), and other types in 3 cases (4%). The median of the maximal clinical diameter of the tumors was 40 mm (range 15–150). The distribution of clinical tumor stages is shown in Table 1.
● A. MAYER et al.
931
Auer classification Table 2 shows the associations between the Auer classification and various tumor and patient characteristics. Most tumors fell into Auer classes III (23.4%) and IV (64.1%). The Auer classification correlated significantly with patient age (p ⫽ 0.0025) and menopausal status (p ⫽ 0.019). The specimens from patients ⱖ54 years, as well as from postmenopausal patients, were classified as Auer IV significantly more often than those of patients ⱕ53 years, which tended to belong to Auer classes I–III. None of the “older” or postmenopausal patients were denoted as Auer I. In the subgroup of squamous cell carcinomas, the same correlations were found, albeit with a lower p value for the association with menopausal status (p ⫽ 0.005). The DNA classification was independent of all other investigated tumor and patient variables. No correlation of Auer classification with tumor oxygenation status was found. As Table 3 and Fig. 1 illustrate, hypoxic (median PO2 ⬍10 mm Hg) and less hypoxic (median PO2 ⱖ10 mm Hg) tumors were distributed almost equally among Auer classes I–IV. DNA stemline position Statistical analysis of the stemline position data showed a significant correlation with patient age (p ⫽ 0.0163). In the group of younger patients, stemline position was found to be significantly lower compared with patients ⬎53 years of age. Significantly lower stemlines were also found in the group of premenopausal women (p ⫽ 0.029). Both correlations were more pronounced in the subgroup of squamous cell carcinoma (age: p ⫽ 0.0008; menopausal status: p ⫽ 0.0005). In the histologic analysis, squamous cell carcinomas showed significantly lower stemline positions compared with adenocarcinomas (p ⫽ 0.01). The stemline position (divided into diploid and aneuploid tumors) was not affected by the oxygenation status (Table 3 and Fig. 2). 2c deviation index When considering the whole patient group, the 2cDI showed no correlation with any of the parameters mentioned, including oxygenation status. Because DNA parameters have been found to be strongly associated with patient age and menopausal status, possible correlations in different age subgroups (⬍54 years and ⬍41 years) were also analyzed. In these subgroups, the Auer classification and stemline groups again failed to show any significant correlation. However, an association was found between the 2cDI and median tumor PO2 in the group of patients ⬍54 years (r ⫽ 0.46; p ⫽ 0.01). This association was more pronounced in patients ⬍41 years (r ⫽ 0.83; p ⫽ 0.0002; Fig. 3). In the latter age group, an association was also noted between the 2cDI and the fraction of hypoxic PO2 values ⱕ2.5 mm Hg (p ⫽ 0.07) and ⱕ5 mm Hg (p ⫽ 0.04). Tumors with better oxygenation status showed higher 2cDI values. DISCUSSION In the present study, possible associations between the oxygenation status of primary, locally advanced cancers of
932
I. J. Radiation Oncology
● Biology ● Physics
Volume 56, Number 4, 2003
Table 2. Auer classification and tumor/patient data with p values for impact of patient parameters on Auer classification Auer classification Parameter Menopausal status Pre Post Patient age (y) ⱕ53 ⬎53 Smoker Yes No Parity Nullipara Para Clinical tumor stage (FIGO) Ib IIa, IIb IIIa, IIIb IVa, IVb Grading I II III Maximal diameter (cm) ⬍5 ⱖ5 Histologic subtype SCC AC
I (9.4%)
II (3.1%)
III (23.4%)
IV (64.1%)
Correlation (p)
6 0
1 1
9 6
16 25
0.019 (0.005)
6 0
0 2
11 4
17 24
0.0025 (0.003)
4 2
1 1
6 9
15 25
NS
2 4
0 2
2 12
7 34
NS
2 3 1 0
1 1 0 0
4 7 3 0
10 18 12 0
NS
2 2 2
0 0 2
1 7 3
4 20 7
NS
4 2
1 1
10 5
20 21
NS
6 0
1 1
12 1
30 10
NS
Abbreviations: FIGO ⫽ International Federation of Gynecology and Obstetrics; SCC ⫽ squamous cell carcinoma; AC ⫽ adenocarcinoma. Values in parentheses represent subgroup of squamous cell carcinomas.
the uterine cervix and changes in DNA content were examined. When the patient cohort was taken as a whole, all DNA data were found to be independent of oxygenation status. Hypoxic (median PO2 ⬍10 mm Hg) and “normoxic” (less hypoxic) tumors (median PO2 ⱖ10 mm Hg) were almost equally distributed between the different categories of the DNA data according to the Auer classification and also between the stemline position groups (Figs. 1 and 2). A
10-mm Hg cutoff corresponded to the mean of the median oxygenation values of the entire study population and was also used in our earlier publications (1–3). These results are not in agreement with recent experimental data (14, 18 –20, 22, 29), especially because we unexpectedly found that higher values for the 2cDI correlated (r ⫽ 0.83) with better oxygenation of the tumors in a subgroup of younger patients (see below).
Table 3. Distribution of hypoxic and normoxic (less hypoxic) tumors between Auer classes and dichotomized stemline positions (DNA diploid and DNA aneuploid) Hypoxic (PO2 ⬍10 mm Hg)
Less hypoxic (PO2 ⱖ10 mm Hg)
4 (9.4) 1 (3.1) 10 (23.4) 26 (64.1) 41
2 1 5 15 23
14 (34.2) 27 (65.8) 41
7 16 23
Auer class I II III IV Total Stemline position ⱕ2.20c (DNA diploid) ⬎2.20c (DNA aneuploid) Total Numbers in parentheses are percentages.
Oxygenation and DNA content of cervical cancer
Fig. 1. Relative frequency of hypoxic (median PO2 ⬍10 mm Hg) and less hypoxic (median PO2 ⱖ10 mm Hg) tumors in four Auer classes I–IV.
However, several possible explanations for the lack of a correlation between the DNA content and oxygenation data found in this study have to be considered. First, the abovementioned mechanisms of hypoxia-driven genomic changes were observed in in vitro experiments, and the described processes may not necessarily take place in human cancers of the uterine cervix. Alternatively, if the mutational effects of hypoxia are indeed active in carcinomas of the uterine
Fig. 2. Relative frequency of hypoxic (median PO2 ⬍10 mm Hg) and less hypoxic (median PO2 ⱖ10 mm Hg) tumors in DNAdiploid and DNA-aneuploid tumors.
● A. MAYER et al.
933
cervix, hypoxia-induced changes of the DNA content may not be registered by the image cytometric method used. The latter possibility has two main aspects. First, the magnitude of hypoxia-induced DNA changes may lie below the detection threshold of the DNA image cytometry system. The systematic error due solely to the coefficient of variation of the reference cells alone is up to 3%. Furthermore, preliminary studies on normal tissue have shown that normal diploid cervical epithelial cells could vary in their measured DNA content by up to almost 10% from the normal 2c value. For this reason, the cutoff between DNA “diploid” and DNA “aneuploid” tumors was set at 2.2c and 1.8c, respectively. The second aspect concerning the possible insensitivity to hypoxia-induced changes of the DNA content of the method used is probably of even greater importance: that various other factors are also known to have an impact on DNA content in cancer cells, independent of tumor hypoxia. At the same time, assessment of the DNA content of cancer cells is a net measure, which cannot distinguish between individual underlying mutations. Each mutation that affects DNA content, such as aneuploidy, polyploidy, gene amplification, and structural chromosomal aberrations, may itself have multiple causes that are not necessarily mutually exclusive. A few examples to illustrate this point follow. Polyploidy may be caused by inactivation of the p53 tumor suppressor protein alone (34). Chromosomal aberrations and gene amplification have also been shown to occur when p53 is inactivated (35). In the vast majority of cancers of the uterine cervix, p53 is at least partially inactivated by accelerated proteasome degradation mediated by the E6 oncoprotein of “high-risk” human papillomaviruses (HPVs) (36), in rare cases through mutations (37, 38). Aneuploidy and corresponding changes in DNA content also occur independently of tumor hypoxia, because these changes are even found in premalignant lesions of various tumor entities. Cancers of the uterine cervix are a good example of this, because premalignant lesions have been studied thoroughly, and “aneuploid” DNA content has constantly been found even in early dysplastic lesions (39). Multiple defects may lead to aneuploidy (40). For carcinomas of the uterine cervix, Skyldberg et al. (41) showed that centrosome aberrations and aneuploid DNA content are correlated and both can be found in early dysplastic lesions (“CIN II”). Centrosome disturbances may again be in part the consequence of p53 inactivation (42), but additional events are likely to be required (43), because p53 inactivation alone has been shown not to result in aneuploidy (44). Aneuploidy and corresponding changes in the DNA content of cancer cells may also be attributable to the consequences of critical shortening of telomeres in epithelial cells with a long history of mitotic activity, because shortened telomeres have been shown to behave like “sticky ends,” causing chromosome fusion and subsequent breakage (45). Fitting with this, critical “telomere attrition” has been shown to lead to aneuploid cancers in a telomerase-knockout mouse model (46). This mechanism is also of special interest because in this study aneuploidy was found to be associated
934
I. J. Radiation Oncology
● Biology ● Physics
Volume 56, Number 4, 2003
Fig. 3. Correlation between 2cDI and median tumor PO2 in patients ⬍41 years with squamous cell carcinomas of the uterine cervix.
with patient age (see below), a finding that has also been documented by other groups (47, 48). The only other known study of possible correlations between DNA content (determined by DNA flow cytometry) and tumor oxygenation status was performed by Haensgen et al. (29). They reported a (nonsignificant) tendency toward higher DNA stemlines in hypoxic tumors. They also found a significant correlation between a higher DNA stemline position and positive immunohistochemical staining for p53 (p ⫽ 0.05). The authors interpreted this positive staining as a correlate of either p53 mutation or inactivation by “high-risk”-HPV E6. Therefore, on the basis the pathogenetic model first mentioned by Graeber et al. (49), describing the selection of p53-deficient cells under hypoxic conditions, they combined p53 and oxygenation status for correlation with DNA content. Hypoxic cases that at the same time showed positive immunohistochemical staining for p53 [termed “t-p53/HF5(⫹)” by Haensgen et al. (29)] had “the highest (DNA) stemlines.” Although this correlation was again not statistically significant, the authors interpreted this finding to be the result of an increased genomic instability in apoptosis-deficient, p53-inactivated cells, selected under hypoxic conditions. The values for the DNA indexes in the “t-p53/HF5(⫹)” group, which encompassed 34 of the 55 patients available for DNA analysis, differed substantially from those presented in major studies on DNA content in cervical carcinoma (47, 50, 51). The interpretation of the pathogenetic causality of the association among hypoxia, p53 status, and DNA data in the study of Haensgen et al. (29) also depends decisively on the interpretation of positive immunohistochemical staining for
p53. They used the antibody DO-7, which binds p53 irrespective of mutational status, and found 85% of their cases to be positive. According to Haensgen and coworkers (29), immunohistochemical detectability of p53 is either the consequence of a prolongation of its half-life period due to mutations that interfere with its “physiologic” degradation or of its stabilization due to interaction with the “high-risk” HPV-E6 oncoprotein. Because it is widely accepted that mutations of the p53 gene are very rare events in cervical cancer (37, 38), they alone cannot account for the high rate of p53 positivity found by Haensgen et al. (29) (⬃85%). Interaction with “high-risk” HPV on the other hand, does not lead to stabilization but to ubiquitin-dependent targeting of p53 to proteasome degradation (36, 52), that is, a shorter half-life. In conclusion, it is unlikely that interaction with HPV oncoproteins or physical mutation can sufficiently explain the finding of 85% p53-positive cases. In cervical carcinoma, therefore, many authors regard positive immunohistochemistry for p53 as an indication of overexpression of wild-type p53 (53–58). Data supporting this concept can also be drawn from the work of Haensgen et al. (29), who reported a significant association between the occurrence of hypoxia and increased staining for p53. This finding is compatible with the interpretation of positive p53 staining as being an indication of overexpression of functional wildtype p53, because hypoxia is known to trigger p53-dependent apoptosis (49) and because hypoxia has been shown to increase p53 levels by inhibition of “high-risk” E6-mediated degradation and by downregulation of mdm2 (59). In the present study, a significant correlation of both Auer classification and stemline position with patient age was
Oxygenation and DNA content of cervical cancer
evident when the latter was evaluated in the form of two age groups (ⱕ53 vs. ⬎53 years), together with a correlation with menopausal status. These results are in agreement with those of other studies (47, 48). Ly et al. (60) presented data supporting a causal relationship between age and the incidence of aneuploidy. Because of this age dependence, possible associations between DNA parameters and oxygenation status in different age groups were examined. Here, a significant correlation of oxygenation status with the 2cDI in the group of patients ⬍41 years was found (r ⫽ 0.83, p ⫽ 0.0002). The highest values for 2cDI were found in 2 patients with well-oxygenated tumors. Although the significance of the correlation depends primarily on the data from these 2 cases, this relationship certainly deserves further investigation. If the above-mentioned argument concerning a reactivation of p53 function under hypoxic conditions is considered (liberation from the interaction with “high-risk” E6), the conclusion that hypoxia may even antagonize genetic instability could be reached. However, even if the p53 protein is “protected” under hypoxic conditions, this may not apply to its function, because hypoxia may affect p53’s
● A. MAYER et al.
935
downstream effectors. The findings of Kim et al. (61) are in accordance with this argument, because they showed that hypoxia leads to increased levels of p53 in cervical carcinoma cell lines (CaSki and SiHa) but not to apoptosis, which is commonly regarded as one of p53’s main effects under hypoxic conditions in this cell type. Correlations between the DNA content and oxygenation status of tumors have to date only been reported by Haensgen et al. (29). Their findings are not in agreement with those of the present study. However, an alternative interpretation of their data may allow this contradiction to be at least partially resolved. The correlation found in our study in the group of “younger” patients must be interpreted cautiously, owing to the small number of patients in the respective age group. The investigation of genomic changes induced by tumor-hypoxia warrants additional research, although single-time measurements of DNA content appear to be too unspecific to allow the identification of correlations with hypoxia that may reveal clues to the pathogenesis of hypoxia-induced genomic changes.
REFERENCES 1. Ho¨ ckel M, Knoop C, Schlenger K, et al. Intratumoral pO2 predicts survival in advanced cancer of the uterine cervix. Radiother Oncol 1993;26:45–50. 2. Ho¨ ckel M, Schlenger K, Aral B, et al. Association between tumor hypoxia and malignant progression in advanced cancer of the uterine cervix. Cancer Res 1996;56:4509–4515. 3. Ho¨ ckel M, Schlenger K, Ho¨ ckel S, et al. Tumor hypoxia in pelvic recurrences of cervical cancer. Int J Cancer 1998;79: 365–369. 4. Fyles AW, Milosevic M, Wong R, et al. Oxygenation predicts radiation response and survival in patients with cervix cancer. Radiother Oncol 1998;48:149–156. 5. Knocke TH, Weitmann HD, Feldmann HJ, et al. Intratumoral pO2-measurements as predictive assay in the treatment of carcinoma of the uterine cervix. Radiother Oncol 1999;53:99– 104. 6. Sundfor K, Lyng H, Trope CG, et al. Treatment outcome in advanced squamous cell carcinoma of the uterine cervix: Relationships to pretreatment tumor oxygenation and vascularization. Radiother Oncol 2000;54:101–107. 7. Brizel DM, Sibley GS, Prosnitz LR, et al. Tumor hypoxia adversely affects the prognosis of carcinoma of the head and neck. Int J Radiat Oncol Biol Phys 1997;38:285–289. 8. Nordsmark M, Alsner J, Keller J, et al. Hypoxia in human soft tissue sarcomas: Adverse impact on survival and no association with p53 mutations. Br J Cancer 2001;84:1070–1075. 9. Gray LH, Conger AD, Ebert M, et al. The concentration of oxygen dissolved in tissues at the time of irradiation as a factor in radiotherapy. Br J Radiol 1953;26:638–648. 10. Ho¨ ckel M, Vaupel P. Tumor hypoxia: Definitions and current clinical, biologic, and molecular aspects. J Natl Cancer Inst 2001;93:266–276. 11. Semenza GL. HIF-1: Mediator of physiological and pathophysiological responses to hypoxia. J Appl Physiol 2000;88: 1474–1480. 12. Yuan J, Narayanan L, Rockwell S, et al. Diminished DNA repair and elevated mutagenesis in mammalian cells exposed to hypoxia and low pH. Cancer Res 2000;60:4372–4376. 13. Stoler DL, Anderson GR, Russo CA, et al. Anoxia-inducible
14.
15. 16.
17. 18.
19. 20. 21. 22.
23.
24.
endonuclease activity as a potential basis of the genomic instability of cancer cells. Cancer Res 1992;52:4372–4378. Russo CA, Weber TK, Volpe CM, et al. An anoxia inducible endonuclease and enhanced DNA breakage as contributors to genomic instability in cancer. Cancer Res 1995; 55:1122–1128. Reynolds TY, Rockwell S, Glazer PM. Genetic instability induced by the tumor microenvironment. Cancer Res 1996; 56:5754–5757. Toledo F, Le Roscouet D, Buttin G, et al. Co-amplified markers alternate in megabase long chromosomal inverted repeats and cluster independently in interphase nuclei at early steps of mammalian gene amplification. EMBO J 1992;11: 2665–2673. Ma C, Martin S, Trask B, et al. Sister chromatid fusion initiates amplification of the dihydrofolate reductase gene in Chinese hamster cells. Genes Dev 1993;7:605–620. Rice GC, Hoy C, Schimke RT. Transient hypoxia enhances the frequency of dihydrofolate reductase gene amplification in Chinese hamster ovary cells. Proc Natl Acad Sci USA 1986; 83:5978–5982. Young SD, Marshall RS, Hill RP. Hypoxia induces DNA overreplication and enhances metastatic potential of murine tumor cells. Proc Natl Acad Sci USA 1988;85:9533–9537. Rofstad EK, Johnsen NM, Lyng H. Hypoxia-induced tetraploidisation of a diploid human melanoma cell line in vitro. Br J Cancer (Suppl) 1996;27:S136–S139. Gaulden ME. Maternal age effect: The enigma of Down syndrome and other trisomic conditions. Mutat Res 1992;296: 69–88. Coquelle A, Toledo F, Stern S, et al. A new role for hypoxia in tumor progression: Induction of fragile site triggering genomic rearrangements and formation of complex DMs and HSRs. Mol Cell 1998;2:259–265. Coquelle A, Pipiras E, Toledo F, et al. Expression of fragile sites triggers intrachromosomal mammalian gene amplification and sets boundaries to early amplicons. Cell 1997;89: 215–225. Vaupel P, Briest S, Ho¨ ckel M. Hypoxia in breast cancer:
936
25.
26. 27. 28. 29. 30. 31. 32. 33. 34.
35. 36. 37. 38. 39.
40. 41. 42. 43. 44.
I. J. Radiation Oncology
● Biology ● Physics
Pathogenesis, characterization and biological/therapeutic implications. Wien Med Wschr 2002;152:334–342. Vaupel P, Ho¨ ckel M. Tumor hypoxia and therapeutic resistance. In: Nowrousian MR, editor. Recombinant human erythropoietin (rhEPO) in clinical oncology. Wien: Springer; 2002. p. 127–146. Vaupel P, Thews O, Hoeckel M. Treatment resistance of solid tumors: Role of hypoxia and anemia. Med Oncol 2001;18: 243–259. Serra M, Tarkkanen M, Baldini N, et al. Simultaneous paired analysis of numerical chromosomal aberrations and DNA content in osteosarcoma. Mod Pathol 2001;14:710–716. Sreekantaiah C, Bhargava MK. Double minute chromatin bodies in carcinoma of the human cervix uteri. Cancer Genet Cytogenet 1992;58:134–140. Haensgen G, Krause U, Becker A, et al. Tumor hypoxia, p53, and prognosis in cervical cancers. Int J Radiat Oncol Biol Phys 2001;50:865–872. Ho¨ ckel M, Schlenger K, Knoop C, et al. Oxygenation of carcinomas of the uterine cervix: Evaluation by computerized O2 tension measurements. Cancer Res 1991;51:6098–6102. Bo¨ cking A, Giroud F, Reith A. Consensus report of the ESACP task force on standardization of diagnostic DNA image cytometry. Anal Cell Pathol 1995;8:67–74. Kindermann D, Hilgers CH. Glare-correction in DNA image cytometry. Anal Cell Pathol 1994;6:165–180. Auer GU, Caspersson TO, Wallgren AS. DNA content and survival in mammary carcinoma. Anal Quant Cytol 1980;2: 161–165. Gualberto A, Aldape K, Kozakiewicz K, et al. An oncogenic form of p53 confers a dominant, gain-of-function phenotype that disrupts spindle checkpoint control. Proc Natl Acad Sci USA 1998;95:5166–5171. Honma M, Momose M, Tanabe H, et al. Requirement of wild-type p53 protein for maintenance of chromosomal integrity. Mol Carcinog 2000;28:203–214. Scheffner M, Werness BA, Huibregtse JM, et al. The E6 oncoprotein encoded by human papillomavirus types 16 and 18 promotes the degradation of p53. Cell 1990;63:1129–1136. Kessis TD, Slebos RJ, Han SM, et al. p53 gene mutations and mdm2 amplification are uncommon in primary carcinomas of the uterine cervix. Am J Pathol 1993;143:1398–1405. Lee JH, Kang YS, Koh JW, et al. p53 gene mutation is rare in human cervical carcinomas with positive HPV sequences. Int J Gynecol Cancer 1994;4:371–378. Heselmeyer K, Macville M, Schrock E, et al. Advanced-stage cervical carcinomas are defined by a recurrent pattern of chromosomal aberrations revealing high genetic instability and a consistent gain of chromosome arm 3q. Genes Chromosomes Cancer 1997;19:233–240. Lengauer C, Kinzler KW, Vogelstein B. Genetic instabilities in human cancers. Nature 1998;396:643–649. Skyldberg B, Fujioka K, Hellstrom AC, et al. Human papillomavirus infection, centrosome aberration, and genetic stability in cervical lesions. Mod Pathol 2001;14:279–284. Fukasawa K, Choi T, Kuriyama R, et al. Abnormal centrosome amplification in the absence of p53. Science 1996;271: 1744–1747. Zhou H, Kuang J, Zhong L, et al. Tumour amplified kinase STK15/BTAK induces centrosome amplification, aneuploidy and transformation. Nat Genet 1998;20:189–193. Bunz F, Fauth C, Speicher MR, et al. Targeted inactivation of p53 in human cells does not result in aneuploidy. Cancer Res 2002;62:1129–1133.
Volume 56, Number 4, 2003
45. Blackburn EH. Telomere states and cell fates. Nature 2000; 408:53–56. 46. Artandi SE, Chang S, Lee SL, et al. Telomere dysfunction promotes non-reciprocal translocations and epithelial cancers in mice. Nature 2000;406:641–645. 47. Kristensen GB, Kaern J, Abeler VM, et al. No prognostic impact of flow-cytometric measured DNA ploidy and S-phase fraction in cancer of the uterine cervix: A prospective study of 465 patients. Gynecol Oncol 1995;57:79–85. 48. Jarrell MA, Heintz N, Howard P, et al. Squamous cell carcinoma of the cervix: HPV 16 and DNA ploidy as predictors of survival. Gynecol Oncol 1992;46:361–366. 49. Graeber TG, Osmanian C, Jacks T, et al. Hypoxia-mediated selection of cells with diminished apoptotic potential in solid tumours. Nature 1996;379:88–91. 50. Jakobsen A, Bichel P, Kristensen GB, et al. Prognostic influence of ploidy level and histopathologic differentiation in cervical carcinoma stage IB. Eur J Cancer Clin Oncol 1988; 24:969–972. 51. Kimmig R, Kapsner T, Spelsberg H, et al. DNA cell-cycle analysis of cervical cancer by flow cytometry using simultaneous cytokeratin labelling for identification of tumour cells. J Cancer Res Clin Oncol 1995;121:107–114. 52. Hengstermann A, Linares LK, Ciechanover A, et al. Complete switch from mdm2 to human papillomavirus E6-mediated degradation of p53 in cervical cancer cells. Proc Natl Acad Sci USA 2001;98:1218–1223. 53. Oka K, Suzuki Y, Nakano T. Expression of p27 and p53 in cervical squamous cell carcinoma patients treated with radiotherapy alone: Radiotherapeutic effect and prognosis. Cancer 2000;88:2766–2773. 54. Avall-Lundqvist EH, Silfversward C, Aspenblad U, et al. The impact of tumour angiogenesis, p53 overexpression and proliferative activity (MIB-1) on survival in squamous cervical carcinoma. Eur J Cancer 1997;33:1799–1804. 55. Park CS, Joo IS, Song SY, et al. An immunohistochemical analysis of heat shock protein 70, p53, and estrogen receptor status in carcinoma of the uterine cervix. Gynecol Oncol 1999;74:53–60. 56. Huang LW, Chou YY, Chao SL, et al. p53 and p21 expression in precancerous lesions and carcinomas of the uterine cervix: Overexpression of p53 predicts poor disease outcome. Gynecol Oncol 2001;83:348–354. 57. Giannoudis A, Herrington CS. Differential expression of p53 and p21 in low grade cervical squamous intraepithelial lesions infected with low, intermediate, and high risk human papillomaviruses. Cancer 2000;89:1300–1307. 58. Tervahauta AI, Syrjanen SM, Mantyjarvi R, et al. Detection of p53 protein and Ki-67 proliferation antigen in human papillomavirus (HPV)-positive and HPV-negative cervical lesions by immunohistochemical double-staining. Cytopathology 1994;5:282–293. 59. Alarcon R, Koumenis C, Geyer RK, et al. Hypoxia induces p53 accumulation through mdm2 down-regulation and inhibition of E6-mediated degradation. Cancer Res 1999;59:6046– 6051. 60. Ly DH, Lockhart DJ, Lerner RA, et al. Mitotic misregulation and human aging. Science 2000;287:2486–2492. 61. Kim CY, Tsai MH, Osmanian C, et al. Selection of human cervical epithelial cells that possess reduced apoptotic potential to low-oxygen conditions. Cancer Res 1997;57:4200– 4204.
doi:10.1016/S0360-3016(03)00065-8
CLINICAL INVESTIGATION
Cervix
IMPACT OF OXYGENATION STATUS AND PATIENT AGE ON DNA CONTENT IN CANCERS OF THE UTERINE CERVIX ARNULF MAYER, M.D., DR. MED.,* MICHAEL HO¨ CKEL, M.D., DR. MED. HABIL., DR. RER. NAT.,† OLIVER THEWS, M.D., DR. MED. HABIL.,* KARLHEINZ SCHLENGER, DR. RER. NAT.,* AND PETER VAUPEL, M.D., DR. MED. HABIL.* *Institute of Physiology and Pathophysiology, University of Mainz, Mainz, Germany; †Department of Obstetrics and Gynecology, University of Leipzig, Leipzig, Germany Purpose: In carcinomas of the uterine cervix, the tumor oxygenation status has been shown to be a prognostic indicator that is independent of treatment modality. In vitro studies suggest gene amplification and polyploidization to be among the major consequences of hypoxia (with or without consecutive reoxygenation) and to be associated with treatment resistance and tumor progression. This study analyzed whether hypoxia alters net DNA content in uterine cervix cancer cells to the extent that it is identifiable by DNA image cytometry. Methods and Materials: In 64 patients with primary cervical cancer, tumor oxygenation was assessed polarographically and correlated with cell DNA content (DNA image cytometry) in areas adjacent to the oxygen microsensor tracks in which oxygenation measurements were made. Results: No correlation between DNA content (stemline position, Auer classification, and 2c deviation index) and oxygenation status was observed. However, an association between DNA content and patient age and menopausal status was found. Conclusion: Using DNA cytometry, hypoxia-associated genomic changes in uterine cervix cancer cells could not be detected. The impact of tumor hypoxia on the genome may be masked by the effects of alternative mechanisms of genomic instability that can also influence DNA content. © 2003 Elsevier Inc. Tumor oxygenation, DNA content, DNA image cytometry, Gene amplification, Ploidy.
Tumor oxygenation has been shown to be an adverse prognostic factor in primary and recurrent cancer of the uterine cervix (1– 6), as well as in other tumor entities (7, 8). For 50 years, tumor hypoxia has been considered a relevant factor causing “extrinsic” resistance to radiotherapy, because a reduced “oxygen enhancement effect” in hypoxic tumors diminishes the biologic effect of ionizing radiation (9). More recently, a number of studies have presented substantial evidence indicating that the impact of hypoxia on a patient’s prognosis may additionally be attributable to an enhanced “intrinsic” biologic aggressiveness and an acceleration of malignant progression of hypoxic tumors, predominantly because the influence on prognosis has been shown to be independent of treatment modality (2). Hypoxic tumor cells have also been shown to invade tissue faster and reach higher rates of perifocal, local, and distant spread (10). These adverse effects of hypoxia are thought to be mediated through at least two different pathogenetic
mechanisms. First, moderate hypoxia (⬍1% oxygen) has a profound impact on gene expression, which is mainly mediated through the oxygen-regulated transcription factor hypoxia-inducible factor-1␣, which activates genes involved in adaptation to anaerobic metabolism, angiogenesis, and capability of tissue invasion (11). Second, severe hypoxia (⬍0.1% oxygen) seems to contribute to the genomic instability of tumor cells. Hypoxia has been shown to reduce DNA repair mechanisms (12), resulting in point mutations, and to stimulate endonuclease activity (13) suitable to induce chromosomal double-strand breaks (14, 15), thus leading to structural chromosomal aberrations. Double-strand breaks have also been shown to be an important source for gene amplification by the initiation of breakage-fusionbridge cycles (16, 17). Additionally, hypoxia can induce polyploidy (18 –20) and has been suspected of causing aneuploidy by perturbations of the spindle apparatus (21). The chromosomal breaks occurring under hypoxic conditions seem not to be randomly distributed, but to occur
Reprint requests to: Arnulf Mayer, M.D., Institute of Physiology and Pathophysiology, University of Mainz, Duesbergweg 6, Mainz D-55099 Germany. Tel: ⫹49-6131-39 25203; Fax: ⫹49-6131-39 25774; E-mail: [email protected] Acknowledgments—The authors thank Dr. Debra K. Kelleher for
her assistance in preparing this manuscript, and Professor Dr. E. Schulte, Dept. of Anatomy, University of Mainz, for providing technical expertise. Received Oct 10, 2002, and in revised form Dec 26, 2002. Accepted for publication Jan 7, 2003.
INTRODUCTION
929
930
I. J. Radiation Oncology
● Biology ● Physics
preferentially at “fragile sites” (22) (i.e., specific chromosomal sites exhibiting an increased propensity for breakage). These fragile sites are often located in the vicinity of oncogenes or genes mediating drug resistance (23) and thus hypoxia-induced chromosomal aberrations and gene amplification may have some specificity for these crucial genes. Hypoxic cells have accordingly been shown to exhibit an increased resistance to chemotherapeutic agents (18, 24 – 26), as well as a greater metastatic potential (19). Because investigations of these mechanisms have only been carried out in vitro and possible interactions have not been examined, their contribution to the malignant progression of solid tumors in vivo remains unclear. The aforementioned hypoxia-related forms of genomic mutations (gene amplification, polyploidy, and chromosomal aberrations) have been shown to have a distinct effect on DNA content (27, 28). The aim of the present study therefore was to correlate the DNA parameters obtained by image cytometric DNA analysis with the oxygenation status of locally advanced cervical carcinoma. Importantly, with our method, tumor oxygenation and DNA content data originated from adjacent tissue areas. The results presented show that the DNA data are generally independent of the oxygenation status, a finding that is not in agreement with a recent article in this Journal (29). METHODS AND MATERIALS Patients All patients in this study were recruited from a prospective clinical trial for the evaluation of the significance of tumor oxygenation in primary, locally advanced carcinoma of the uterine cervix that has been conducted at the Department of Obstetrics and Gynecology of the University of Mainz since June 1989. Patients gave informed consent before enrollment, and a medical ethics committee approved the study design. Sixty-four patients from this group, for whom cytologic specimens (tumor cell smear preparations) were available, were included in the present study. Tumor oxygenation measurements Tumor oxygenation was assessed pretherapeutically with the computerized Eppendorf histography system (Eppendorf, Hamburg, Germany), using a protocol that has been described in detail earlier (30). In brief, PO2 readings were performed in the conscious patient along linear tracks, first in the s.c. fat of the mons pubis followed by cervical measurements at the 12- and 6-o’clock sites of the macroscopically vital tumor tissue. Within the tumor tissue, up to 35 PO2 measurements were taken in each electrode track (70 readings in total) starting at a tissue depth of 5 mm. The PO2 measuring points were situated 0.7 mm apart from each other, resulting in an overall measurement track length of approximately 25 mm. After oxygenation measurements, needle core biopsies approximately 2 mm in diameter and 20 mm in length were taken from those tumor areas at which the PO2 measurements had been obtained. The intra-
Volume 56, Number 4, 2003
vaginal temperature, blood pressure, heart rate, hemoglobin concentration, hematocrit, and arterial oxyhemoglobin saturation were monitored concomitantly with the PO2 readings. The pretherapeutic PO2 measurements were usually performed 1–5 days before oncologic treatment. Tumor cell preparations Smear preparations on standard histologic slides were obtained from the biopsy cylinders and were spray fixed immediately afterward (Merckofix, Merck, Darmstadt, Germany). In 59 of 64 cases, slides representing both standard positions (6 h and 12 h) of oxygenation measurement could be obtained. In the 5 remaining cases, either only one oxygenation measurement was performed or the biopsy cylinder seemed too fragile for the smearing procedure. Cytochemistry Feulgen staining was performed following a standardized protocol that met the requirements of the European Society for Analytical Cellular Pathology Task Force (31). After rehydration in distilled water, acid hydrolysis was performed using 5N HCl maintained at a constant temperature of 22°C. After rinsing in distilled water, the specimens were stained with Schiff’s reagent (Merck No. 9034-100) for 1 h. Afterward, excess dye was removed by rinsing in SO2 water, followed by dehydration in an ascending alcohol series. Cover slips were fixed using Eukitt (Merck), and the slides were then kept in the dark until measurement. DNA image cytometry The DNA content was measured using the CYDOK system (Technisches Buero Hilgers, Koenigswinter, Germany), consisting of a conventional light microscope (Leica DM LS, Leica Microsystems, Wetzlar, Germany), with a 40⫻ objective, interference filter of 550 ⫾ 8 nm half value, and a diameter of 25 mm (Schott, Mannheim, Germany), and a CCD monochrome camera (Pulnix TM7 CN with 768 ⫻ 494 pixels, Pulnix America, Sunnyvale, CA). Image analysis was performed on a standard personal computer equipped with a framegrabber video processor (MPV-AT, Matrox Electronic Systems, Dorval, Quebec, Canada), connected to a standard cathode-ray display terminal monitor. Measurements were performed and interpreted using the CYDOK software. Slides were taken from light-protected storage immediately before measurement and analyzed in one passage to avoid bleaching from normal daylight or the light of the microscope. Specimens were examined in a meander-like fashion to include all accessible tumor cells. The CYDOK system allows the separation of clustered tumor cells by user interaction, thereby reducing selective bias from this phenomenon, which is frequently found in cytologic preparations of cancer cells of the uterine cervix. The CYDOK software also corrected for glare error (32). From the areas containing the evaluated tumor cells, at least 20 neutrophils were selected as internal reference cells; 200 –300 tumor cells were analyzed in each preparation. For quantitative analysis, DNA histograms were generated and
Oxygenation and DNA content of cervical cancer
Table 1. Clinical tumor stages (FIGO) in 64 patients eligible for DNA cytometry Clinical tumor stage (FIGO)
Patients (n)
Ib IIa IIb IIIa IIIb IVa Unknown
17 1 28 1 15 1 1
Abbreviation: FIGO ⫽ International Federation of Gynecology and Obstetrics.
the DNA parameters stemline position, S-phase fraction (percentage of DNA values ⱖ2.5c and ⱕ3.8c), and the square deviation index of DNA values from the normal 2c-peak (2c deviation index [2cDI]) were calculated. For interpretation, DNA data were dichotomized into diploid (from 1.80c to 2.20c) vs. aneuploid (⬍1.80c or ⬎ 2.20c) stemlines. In addition, the DNA histograms were assigned to one of four classes according to the criteria defined by Auer et al. (33). Auer I is defined as a peak at 2c (DNA value corresponding to diploid cells) and a minor proportion of cells with a DNA content of 4c. Less than 5% of all cells are found between the 2c and 4c peaks (S-phase fraction). Histograms classified as Auer II showed either a single peak at 4c (Auer IIa) or peaks at 2c and 4c, with ⬎20% belonging to the latter peak (Auer IIb). Auer III histograms are similar to Auer I or II, but with ⬎5% of the cells lying between the 2c and 4c peaks. Auer IV histograms contain aneuploid peaks (for definition of aneuploid stemlines, see above). Statistical analysis The results are expressed as the mean ⫾ SEM. The differences between the groups were assessed by the twotailed Wilcoxon test for unpaired samples and Fisher’s exact test, as appropriate. The significance level was set at ␣ ⫽ 5% for all comparisons. Linear correlation between two parameters was described by Pearson’s correlation coefficient (r). RESULTS Cytologic specimens from 64 patients with primary cervical cancers who had undergone tumor oxygenation measurement were available for assessment of DNA content. The median patient age was 53 years (range 26 – 80). The groups of pre- and postmenopausal patients were equally large (32 patients). Histologic examination showed squamous cell carcinoma in 49 (77%), adenocarcinoma in 12 (19%), and other types in 3 cases (4%). The median of the maximal clinical diameter of the tumors was 40 mm (range 15–150). The distribution of clinical tumor stages is shown in Table 1.
● A. MAYER et al.
931
Auer classification Table 2 shows the associations between the Auer classification and various tumor and patient characteristics. Most tumors fell into Auer classes III (23.4%) and IV (64.1%). The Auer classification correlated significantly with patient age (p ⫽ 0.0025) and menopausal status (p ⫽ 0.019). The specimens from patients ⱖ54 years, as well as from postmenopausal patients, were classified as Auer IV significantly more often than those of patients ⱕ53 years, which tended to belong to Auer classes I–III. None of the “older” or postmenopausal patients were denoted as Auer I. In the subgroup of squamous cell carcinomas, the same correlations were found, albeit with a lower p value for the association with menopausal status (p ⫽ 0.005). The DNA classification was independent of all other investigated tumor and patient variables. No correlation of Auer classification with tumor oxygenation status was found. As Table 3 and Fig. 1 illustrate, hypoxic (median PO2 ⬍10 mm Hg) and less hypoxic (median PO2 ⱖ10 mm Hg) tumors were distributed almost equally among Auer classes I–IV. DNA stemline position Statistical analysis of the stemline position data showed a significant correlation with patient age (p ⫽ 0.0163). In the group of younger patients, stemline position was found to be significantly lower compared with patients ⬎53 years of age. Significantly lower stemlines were also found in the group of premenopausal women (p ⫽ 0.029). Both correlations were more pronounced in the subgroup of squamous cell carcinoma (age: p ⫽ 0.0008; menopausal status: p ⫽ 0.0005). In the histologic analysis, squamous cell carcinomas showed significantly lower stemline positions compared with adenocarcinomas (p ⫽ 0.01). The stemline position (divided into diploid and aneuploid tumors) was not affected by the oxygenation status (Table 3 and Fig. 2). 2c deviation index When considering the whole patient group, the 2cDI showed no correlation with any of the parameters mentioned, including oxygenation status. Because DNA parameters have been found to be strongly associated with patient age and menopausal status, possible correlations in different age subgroups (⬍54 years and ⬍41 years) were also analyzed. In these subgroups, the Auer classification and stemline groups again failed to show any significant correlation. However, an association was found between the 2cDI and median tumor PO2 in the group of patients ⬍54 years (r ⫽ 0.46; p ⫽ 0.01). This association was more pronounced in patients ⬍41 years (r ⫽ 0.83; p ⫽ 0.0002; Fig. 3). In the latter age group, an association was also noted between the 2cDI and the fraction of hypoxic PO2 values ⱕ2.5 mm Hg (p ⫽ 0.07) and ⱕ5 mm Hg (p ⫽ 0.04). Tumors with better oxygenation status showed higher 2cDI values. DISCUSSION In the present study, possible associations between the oxygenation status of primary, locally advanced cancers of
932
I. J. Radiation Oncology
● Biology ● Physics
Volume 56, Number 4, 2003
Table 2. Auer classification and tumor/patient data with p values for impact of patient parameters on Auer classification Auer classification Parameter Menopausal status Pre Post Patient age (y) ⱕ53 ⬎53 Smoker Yes No Parity Nullipara Para Clinical tumor stage (FIGO) Ib IIa, IIb IIIa, IIIb IVa, IVb Grading I II III Maximal diameter (cm) ⬍5 ⱖ5 Histologic subtype SCC AC
I (9.4%)
II (3.1%)
III (23.4%)
IV (64.1%)
Correlation (p)
6 0
1 1
9 6
16 25
0.019 (0.005)
6 0
0 2
11 4
17 24
0.0025 (0.003)
4 2
1 1
6 9
15 25
NS
2 4
0 2
2 12
7 34
NS
2 3 1 0
1 1 0 0
4 7 3 0
10 18 12 0
NS
2 2 2
0 0 2
1 7 3
4 20 7
NS
4 2
1 1
10 5
20 21
NS
6 0
1 1
12 1
30 10
NS
Abbreviations: FIGO ⫽ International Federation of Gynecology and Obstetrics; SCC ⫽ squamous cell carcinoma; AC ⫽ adenocarcinoma. Values in parentheses represent subgroup of squamous cell carcinomas.
the uterine cervix and changes in DNA content were examined. When the patient cohort was taken as a whole, all DNA data were found to be independent of oxygenation status. Hypoxic (median PO2 ⬍10 mm Hg) and “normoxic” (less hypoxic) tumors (median PO2 ⱖ10 mm Hg) were almost equally distributed between the different categories of the DNA data according to the Auer classification and also between the stemline position groups (Figs. 1 and 2). A
10-mm Hg cutoff corresponded to the mean of the median oxygenation values of the entire study population and was also used in our earlier publications (1–3). These results are not in agreement with recent experimental data (14, 18 –20, 22, 29), especially because we unexpectedly found that higher values for the 2cDI correlated (r ⫽ 0.83) with better oxygenation of the tumors in a subgroup of younger patients (see below).
Table 3. Distribution of hypoxic and normoxic (less hypoxic) tumors between Auer classes and dichotomized stemline positions (DNA diploid and DNA aneuploid) Hypoxic (PO2 ⬍10 mm Hg)
Less hypoxic (PO2 ⱖ10 mm Hg)
4 (9.4) 1 (3.1) 10 (23.4) 26 (64.1) 41
2 1 5 15 23
14 (34.2) 27 (65.8) 41
7 16 23
Auer class I II III IV Total Stemline position ⱕ2.20c (DNA diploid) ⬎2.20c (DNA aneuploid) Total Numbers in parentheses are percentages.
Oxygenation and DNA content of cervical cancer
Fig. 1. Relative frequency of hypoxic (median PO2 ⬍10 mm Hg) and less hypoxic (median PO2 ⱖ10 mm Hg) tumors in four Auer classes I–IV.
However, several possible explanations for the lack of a correlation between the DNA content and oxygenation data found in this study have to be considered. First, the abovementioned mechanisms of hypoxia-driven genomic changes were observed in in vitro experiments, and the described processes may not necessarily take place in human cancers of the uterine cervix. Alternatively, if the mutational effects of hypoxia are indeed active in carcinomas of the uterine
Fig. 2. Relative frequency of hypoxic (median PO2 ⬍10 mm Hg) and less hypoxic (median PO2 ⱖ10 mm Hg) tumors in DNAdiploid and DNA-aneuploid tumors.
● A. MAYER et al.
933
cervix, hypoxia-induced changes of the DNA content may not be registered by the image cytometric method used. The latter possibility has two main aspects. First, the magnitude of hypoxia-induced DNA changes may lie below the detection threshold of the DNA image cytometry system. The systematic error due solely to the coefficient of variation of the reference cells alone is up to 3%. Furthermore, preliminary studies on normal tissue have shown that normal diploid cervical epithelial cells could vary in their measured DNA content by up to almost 10% from the normal 2c value. For this reason, the cutoff between DNA “diploid” and DNA “aneuploid” tumors was set at 2.2c and 1.8c, respectively. The second aspect concerning the possible insensitivity to hypoxia-induced changes of the DNA content of the method used is probably of even greater importance: that various other factors are also known to have an impact on DNA content in cancer cells, independent of tumor hypoxia. At the same time, assessment of the DNA content of cancer cells is a net measure, which cannot distinguish between individual underlying mutations. Each mutation that affects DNA content, such as aneuploidy, polyploidy, gene amplification, and structural chromosomal aberrations, may itself have multiple causes that are not necessarily mutually exclusive. A few examples to illustrate this point follow. Polyploidy may be caused by inactivation of the p53 tumor suppressor protein alone (34). Chromosomal aberrations and gene amplification have also been shown to occur when p53 is inactivated (35). In the vast majority of cancers of the uterine cervix, p53 is at least partially inactivated by accelerated proteasome degradation mediated by the E6 oncoprotein of “high-risk” human papillomaviruses (HPVs) (36), in rare cases through mutations (37, 38). Aneuploidy and corresponding changes in DNA content also occur independently of tumor hypoxia, because these changes are even found in premalignant lesions of various tumor entities. Cancers of the uterine cervix are a good example of this, because premalignant lesions have been studied thoroughly, and “aneuploid” DNA content has constantly been found even in early dysplastic lesions (39). Multiple defects may lead to aneuploidy (40). For carcinomas of the uterine cervix, Skyldberg et al. (41) showed that centrosome aberrations and aneuploid DNA content are correlated and both can be found in early dysplastic lesions (“CIN II”). Centrosome disturbances may again be in part the consequence of p53 inactivation (42), but additional events are likely to be required (43), because p53 inactivation alone has been shown not to result in aneuploidy (44). Aneuploidy and corresponding changes in the DNA content of cancer cells may also be attributable to the consequences of critical shortening of telomeres in epithelial cells with a long history of mitotic activity, because shortened telomeres have been shown to behave like “sticky ends,” causing chromosome fusion and subsequent breakage (45). Fitting with this, critical “telomere attrition” has been shown to lead to aneuploid cancers in a telomerase-knockout mouse model (46). This mechanism is also of special interest because in this study aneuploidy was found to be associated
934
I. J. Radiation Oncology
● Biology ● Physics
Volume 56, Number 4, 2003
Fig. 3. Correlation between 2cDI and median tumor PO2 in patients ⬍41 years with squamous cell carcinomas of the uterine cervix.
with patient age (see below), a finding that has also been documented by other groups (47, 48). The only other known study of possible correlations between DNA content (determined by DNA flow cytometry) and tumor oxygenation status was performed by Haensgen et al. (29). They reported a (nonsignificant) tendency toward higher DNA stemlines in hypoxic tumors. They also found a significant correlation between a higher DNA stemline position and positive immunohistochemical staining for p53 (p ⫽ 0.05). The authors interpreted this positive staining as a correlate of either p53 mutation or inactivation by “high-risk”-HPV E6. Therefore, on the basis the pathogenetic model first mentioned by Graeber et al. (49), describing the selection of p53-deficient cells under hypoxic conditions, they combined p53 and oxygenation status for correlation with DNA content. Hypoxic cases that at the same time showed positive immunohistochemical staining for p53 [termed “t-p53/HF5(⫹)” by Haensgen et al. (29)] had “the highest (DNA) stemlines.” Although this correlation was again not statistically significant, the authors interpreted this finding to be the result of an increased genomic instability in apoptosis-deficient, p53-inactivated cells, selected under hypoxic conditions. The values for the DNA indexes in the “t-p53/HF5(⫹)” group, which encompassed 34 of the 55 patients available for DNA analysis, differed substantially from those presented in major studies on DNA content in cervical carcinoma (47, 50, 51). The interpretation of the pathogenetic causality of the association among hypoxia, p53 status, and DNA data in the study of Haensgen et al. (29) also depends decisively on the interpretation of positive immunohistochemical staining for
p53. They used the antibody DO-7, which binds p53 irrespective of mutational status, and found 85% of their cases to be positive. According to Haensgen and coworkers (29), immunohistochemical detectability of p53 is either the consequence of a prolongation of its half-life period due to mutations that interfere with its “physiologic” degradation or of its stabilization due to interaction with the “high-risk” HPV-E6 oncoprotein. Because it is widely accepted that mutations of the p53 gene are very rare events in cervical cancer (37, 38), they alone cannot account for the high rate of p53 positivity found by Haensgen et al. (29) (⬃85%). Interaction with “high-risk” HPV on the other hand, does not lead to stabilization but to ubiquitin-dependent targeting of p53 to proteasome degradation (36, 52), that is, a shorter half-life. In conclusion, it is unlikely that interaction with HPV oncoproteins or physical mutation can sufficiently explain the finding of 85% p53-positive cases. In cervical carcinoma, therefore, many authors regard positive immunohistochemistry for p53 as an indication of overexpression of wild-type p53 (53–58). Data supporting this concept can also be drawn from the work of Haensgen et al. (29), who reported a significant association between the occurrence of hypoxia and increased staining for p53. This finding is compatible with the interpretation of positive p53 staining as being an indication of overexpression of functional wildtype p53, because hypoxia is known to trigger p53-dependent apoptosis (49) and because hypoxia has been shown to increase p53 levels by inhibition of “high-risk” E6-mediated degradation and by downregulation of mdm2 (59). In the present study, a significant correlation of both Auer classification and stemline position with patient age was
Oxygenation and DNA content of cervical cancer
evident when the latter was evaluated in the form of two age groups (ⱕ53 vs. ⬎53 years), together with a correlation with menopausal status. These results are in agreement with those of other studies (47, 48). Ly et al. (60) presented data supporting a causal relationship between age and the incidence of aneuploidy. Because of this age dependence, possible associations between DNA parameters and oxygenation status in different age groups were examined. Here, a significant correlation of oxygenation status with the 2cDI in the group of patients ⬍41 years was found (r ⫽ 0.83, p ⫽ 0.0002). The highest values for 2cDI were found in 2 patients with well-oxygenated tumors. Although the significance of the correlation depends primarily on the data from these 2 cases, this relationship certainly deserves further investigation. If the above-mentioned argument concerning a reactivation of p53 function under hypoxic conditions is considered (liberation from the interaction with “high-risk” E6), the conclusion that hypoxia may even antagonize genetic instability could be reached. However, even if the p53 protein is “protected” under hypoxic conditions, this may not apply to its function, because hypoxia may affect p53’s
● A. MAYER et al.
935
downstream effectors. The findings of Kim et al. (61) are in accordance with this argument, because they showed that hypoxia leads to increased levels of p53 in cervical carcinoma cell lines (CaSki and SiHa) but not to apoptosis, which is commonly regarded as one of p53’s main effects under hypoxic conditions in this cell type. Correlations between the DNA content and oxygenation status of tumors have to date only been reported by Haensgen et al. (29). Their findings are not in agreement with those of the present study. However, an alternative interpretation of their data may allow this contradiction to be at least partially resolved. The correlation found in our study in the group of “younger” patients must be interpreted cautiously, owing to the small number of patients in the respective age group. The investigation of genomic changes induced by tumor-hypoxia warrants additional research, although single-time measurements of DNA content appear to be too unspecific to allow the identification of correlations with hypoxia that may reveal clues to the pathogenesis of hypoxia-induced genomic changes.
REFERENCES 1. Ho¨ ckel M, Knoop C, Schlenger K, et al. Intratumoral pO2 predicts survival in advanced cancer of the uterine cervix. Radiother Oncol 1993;26:45–50. 2. Ho¨ ckel M, Schlenger K, Aral B, et al. Association between tumor hypoxia and malignant progression in advanced cancer of the uterine cervix. Cancer Res 1996;56:4509–4515. 3. Ho¨ ckel M, Schlenger K, Ho¨ ckel S, et al. Tumor hypoxia in pelvic recurrences of cervical cancer. Int J Cancer 1998;79: 365–369. 4. Fyles AW, Milosevic M, Wong R, et al. Oxygenation predicts radiation response and survival in patients with cervix cancer. Radiother Oncol 1998;48:149–156. 5. Knocke TH, Weitmann HD, Feldmann HJ, et al. Intratumoral pO2-measurements as predictive assay in the treatment of carcinoma of the uterine cervix. Radiother Oncol 1999;53:99– 104. 6. Sundfor K, Lyng H, Trope CG, et al. Treatment outcome in advanced squamous cell carcinoma of the uterine cervix: Relationships to pretreatment tumor oxygenation and vascularization. Radiother Oncol 2000;54:101–107. 7. Brizel DM, Sibley GS, Prosnitz LR, et al. Tumor hypoxia adversely affects the prognosis of carcinoma of the head and neck. Int J Radiat Oncol Biol Phys 1997;38:285–289. 8. Nordsmark M, Alsner J, Keller J, et al. Hypoxia in human soft tissue sarcomas: Adverse impact on survival and no association with p53 mutations. Br J Cancer 2001;84:1070–1075. 9. Gray LH, Conger AD, Ebert M, et al. The concentration of oxygen dissolved in tissues at the time of irradiation as a factor in radiotherapy. Br J Radiol 1953;26:638–648. 10. Ho¨ ckel M, Vaupel P. Tumor hypoxia: Definitions and current clinical, biologic, and molecular aspects. J Natl Cancer Inst 2001;93:266–276. 11. Semenza GL. HIF-1: Mediator of physiological and pathophysiological responses to hypoxia. J Appl Physiol 2000;88: 1474–1480. 12. Yuan J, Narayanan L, Rockwell S, et al. Diminished DNA repair and elevated mutagenesis in mammalian cells exposed to hypoxia and low pH. Cancer Res 2000;60:4372–4376. 13. Stoler DL, Anderson GR, Russo CA, et al. Anoxia-inducible
14.
15. 16.
17. 18.
19. 20. 21. 22.
23.
24.
endonuclease activity as a potential basis of the genomic instability of cancer cells. Cancer Res 1992;52:4372–4378. Russo CA, Weber TK, Volpe CM, et al. An anoxia inducible endonuclease and enhanced DNA breakage as contributors to genomic instability in cancer. Cancer Res 1995; 55:1122–1128. Reynolds TY, Rockwell S, Glazer PM. Genetic instability induced by the tumor microenvironment. Cancer Res 1996; 56:5754–5757. Toledo F, Le Roscouet D, Buttin G, et al. Co-amplified markers alternate in megabase long chromosomal inverted repeats and cluster independently in interphase nuclei at early steps of mammalian gene amplification. EMBO J 1992;11: 2665–2673. Ma C, Martin S, Trask B, et al. Sister chromatid fusion initiates amplification of the dihydrofolate reductase gene in Chinese hamster cells. Genes Dev 1993;7:605–620. Rice GC, Hoy C, Schimke RT. Transient hypoxia enhances the frequency of dihydrofolate reductase gene amplification in Chinese hamster ovary cells. Proc Natl Acad Sci USA 1986; 83:5978–5982. Young SD, Marshall RS, Hill RP. Hypoxia induces DNA overreplication and enhances metastatic potential of murine tumor cells. Proc Natl Acad Sci USA 1988;85:9533–9537. Rofstad EK, Johnsen NM, Lyng H. Hypoxia-induced tetraploidisation of a diploid human melanoma cell line in vitro. Br J Cancer (Suppl) 1996;27:S136–S139. Gaulden ME. Maternal age effect: The enigma of Down syndrome and other trisomic conditions. Mutat Res 1992;296: 69–88. Coquelle A, Toledo F, Stern S, et al. A new role for hypoxia in tumor progression: Induction of fragile site triggering genomic rearrangements and formation of complex DMs and HSRs. Mol Cell 1998;2:259–265. Coquelle A, Pipiras E, Toledo F, et al. Expression of fragile sites triggers intrachromosomal mammalian gene amplification and sets boundaries to early amplicons. Cell 1997;89: 215–225. Vaupel P, Briest S, Ho¨ ckel M. Hypoxia in breast cancer:
936
25.
26. 27. 28. 29. 30. 31. 32. 33. 34.
35. 36. 37. 38. 39.
40. 41. 42. 43. 44.
I. J. Radiation Oncology
● Biology ● Physics
Pathogenesis, characterization and biological/therapeutic implications. Wien Med Wschr 2002;152:334–342. Vaupel P, Ho¨ ckel M. Tumor hypoxia and therapeutic resistance. In: Nowrousian MR, editor. Recombinant human erythropoietin (rhEPO) in clinical oncology. Wien: Springer; 2002. p. 127–146. Vaupel P, Thews O, Hoeckel M. Treatment resistance of solid tumors: Role of hypoxia and anemia. Med Oncol 2001;18: 243–259. Serra M, Tarkkanen M, Baldini N, et al. Simultaneous paired analysis of numerical chromosomal aberrations and DNA content in osteosarcoma. Mod Pathol 2001;14:710–716. Sreekantaiah C, Bhargava MK. Double minute chromatin bodies in carcinoma of the human cervix uteri. Cancer Genet Cytogenet 1992;58:134–140. Haensgen G, Krause U, Becker A, et al. Tumor hypoxia, p53, and prognosis in cervical cancers. Int J Radiat Oncol Biol Phys 2001;50:865–872. Ho¨ ckel M, Schlenger K, Knoop C, et al. Oxygenation of carcinomas of the uterine cervix: Evaluation by computerized O2 tension measurements. Cancer Res 1991;51:6098–6102. Bo¨ cking A, Giroud F, Reith A. Consensus report of the ESACP task force on standardization of diagnostic DNA image cytometry. Anal Cell Pathol 1995;8:67–74. Kindermann D, Hilgers CH. Glare-correction in DNA image cytometry. Anal Cell Pathol 1994;6:165–180. Auer GU, Caspersson TO, Wallgren AS. DNA content and survival in mammary carcinoma. Anal Quant Cytol 1980;2: 161–165. Gualberto A, Aldape K, Kozakiewicz K, et al. An oncogenic form of p53 confers a dominant, gain-of-function phenotype that disrupts spindle checkpoint control. Proc Natl Acad Sci USA 1998;95:5166–5171. Honma M, Momose M, Tanabe H, et al. Requirement of wild-type p53 protein for maintenance of chromosomal integrity. Mol Carcinog 2000;28:203–214. Scheffner M, Werness BA, Huibregtse JM, et al. The E6 oncoprotein encoded by human papillomavirus types 16 and 18 promotes the degradation of p53. Cell 1990;63:1129–1136. Kessis TD, Slebos RJ, Han SM, et al. p53 gene mutations and mdm2 amplification are uncommon in primary carcinomas of the uterine cervix. Am J Pathol 1993;143:1398–1405. Lee JH, Kang YS, Koh JW, et al. p53 gene mutation is rare in human cervical carcinomas with positive HPV sequences. Int J Gynecol Cancer 1994;4:371–378. Heselmeyer K, Macville M, Schrock E, et al. Advanced-stage cervical carcinomas are defined by a recurrent pattern of chromosomal aberrations revealing high genetic instability and a consistent gain of chromosome arm 3q. Genes Chromosomes Cancer 1997;19:233–240. Lengauer C, Kinzler KW, Vogelstein B. Genetic instabilities in human cancers. Nature 1998;396:643–649. Skyldberg B, Fujioka K, Hellstrom AC, et al. Human papillomavirus infection, centrosome aberration, and genetic stability in cervical lesions. Mod Pathol 2001;14:279–284. Fukasawa K, Choi T, Kuriyama R, et al. Abnormal centrosome amplification in the absence of p53. Science 1996;271: 1744–1747. Zhou H, Kuang J, Zhong L, et al. Tumour amplified kinase STK15/BTAK induces centrosome amplification, aneuploidy and transformation. Nat Genet 1998;20:189–193. Bunz F, Fauth C, Speicher MR, et al. Targeted inactivation of p53 in human cells does not result in aneuploidy. Cancer Res 2002;62:1129–1133.
Volume 56, Number 4, 2003
45. Blackburn EH. Telomere states and cell fates. Nature 2000; 408:53–56. 46. Artandi SE, Chang S, Lee SL, et al. Telomere dysfunction promotes non-reciprocal translocations and epithelial cancers in mice. Nature 2000;406:641–645. 47. Kristensen GB, Kaern J, Abeler VM, et al. No prognostic impact of flow-cytometric measured DNA ploidy and S-phase fraction in cancer of the uterine cervix: A prospective study of 465 patients. Gynecol Oncol 1995;57:79–85. 48. Jarrell MA, Heintz N, Howard P, et al. Squamous cell carcinoma of the cervix: HPV 16 and DNA ploidy as predictors of survival. Gynecol Oncol 1992;46:361–366. 49. Graeber TG, Osmanian C, Jacks T, et al. Hypoxia-mediated selection of cells with diminished apoptotic potential in solid tumours. Nature 1996;379:88–91. 50. Jakobsen A, Bichel P, Kristensen GB, et al. Prognostic influence of ploidy level and histopathologic differentiation in cervical carcinoma stage IB. Eur J Cancer Clin Oncol 1988; 24:969–972. 51. Kimmig R, Kapsner T, Spelsberg H, et al. DNA cell-cycle analysis of cervical cancer by flow cytometry using simultaneous cytokeratin labelling for identification of tumour cells. J Cancer Res Clin Oncol 1995;121:107–114. 52. Hengstermann A, Linares LK, Ciechanover A, et al. Complete switch from mdm2 to human papillomavirus E6-mediated degradation of p53 in cervical cancer cells. Proc Natl Acad Sci USA 2001;98:1218–1223. 53. Oka K, Suzuki Y, Nakano T. Expression of p27 and p53 in cervical squamous cell carcinoma patients treated with radiotherapy alone: Radiotherapeutic effect and prognosis. Cancer 2000;88:2766–2773. 54. Avall-Lundqvist EH, Silfversward C, Aspenblad U, et al. The impact of tumour angiogenesis, p53 overexpression and proliferative activity (MIB-1) on survival in squamous cervical carcinoma. Eur J Cancer 1997;33:1799–1804. 55. Park CS, Joo IS, Song SY, et al. An immunohistochemical analysis of heat shock protein 70, p53, and estrogen receptor status in carcinoma of the uterine cervix. Gynecol Oncol 1999;74:53–60. 56. Huang LW, Chou YY, Chao SL, et al. p53 and p21 expression in precancerous lesions and carcinomas of the uterine cervix: Overexpression of p53 predicts poor disease outcome. Gynecol Oncol 2001;83:348–354. 57. Giannoudis A, Herrington CS. Differential expression of p53 and p21 in low grade cervical squamous intraepithelial lesions infected with low, intermediate, and high risk human papillomaviruses. Cancer 2000;89:1300–1307. 58. Tervahauta AI, Syrjanen SM, Mantyjarvi R, et al. Detection of p53 protein and Ki-67 proliferation antigen in human papillomavirus (HPV)-positive and HPV-negative cervical lesions by immunohistochemical double-staining. Cytopathology 1994;5:282–293. 59. Alarcon R, Koumenis C, Geyer RK, et al. Hypoxia induces p53 accumulation through mdm2 down-regulation and inhibition of E6-mediated degradation. Cancer Res 1999;59:6046– 6051. 60. Ly DH, Lockhart DJ, Lerner RA, et al. Mitotic misregulation and human aging. Science 2000;287:2486–2492. 61. Kim CY, Tsai MH, Osmanian C, et al. Selection of human cervical epithelial cells that possess reduced apoptotic potential to low-oxygen conditions. Cancer Res 1997;57:4200– 4204.