|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2003 American Society for Clinical Oncology Carbonic Anhydrase IX Expression, a Novel Surrogate Marker of Tumor Hypoxia, Is Associated With a Poor Prognosis in Non–Small-Cell Lung CancerFrom the Departments of Oncology, Pathology, and Epidemiology, University Hospitals of Leicester NHS Trust; Imperial Cancer Research Fund, Molecular Oncology Unit, Institute of Medicine, John Ratcliffe Hospital, Oxford, United Kingdom; and Institute of Virology, Slovak Academy of Sciences, Slovak Republic. Address reprint requests to Kenneth J. O’Byrne, MD, University Hospitals of Leicester NHS Trust, Leicester LE1 5WW, United Kingdom; email: ken.obyrne{at}uhl-tr.nhs.uk.
Purpose: To evaluate carbonic anhydrase (CA) IX as a surrogate marker of hypoxia and investigate the prognostic significance of different patterns of expression in non–small-cell lung cancer (NSCLC). Methods: Standard immunohistochemical techniques were used to study CA IX expression in 175 resected NSCLC tumors. CA IX expression was determined by Western blotting in A549 cell lines grown under normoxic and hypoxic conditions. Measurements from microvessels to CA IX positivity were obtained. Results: CA IX immunostaining was detected in 81.8% of patients. Membranous (m) (P = .005), cytoplasmic (c) (P = .018), and stromal (P < .001) CA IX expression correlated with the extent of tumor necrosis (TN). The mean distance from vascular endothelium to the start of tumor cell positivity was 90 µm, which equates to an oxygen pressure of 5.77 mmHg. The distance to blood vessels from individual tumor cells or tumor cell clusters was greater if they expressed mCA IX than if they did not (P < .001). Hypoxic exposure of A549 cells for 16 hours enhanced CA IX expression in the nuclear and cytosolic extracts. Perinuclear (p) CA IX (P = .035) was associated with a poor prognosis. In multivariate analysis, pCA IX (P = .004), stage (P = .001), platelet count (P = .011), sex (P = .027), and TN (P = .035) were independent poor prognostic factors. Conclusion: These results add weight to the contention that mCA IX is a marker of tumor cell hypoxia. The absence of CA IX staining close to microvessels suggests that these vessels are functionally active. pCA IX expression is representative of an aggressive phenotype.
INCREASED TUMOR cell hypoxia, measured with a polarographic needle, correlates with increased metastatic potential, resistance to radiotherapy, and a worse prognosis in solid organ tumors, including non–small-cell lung cancer (NSCLC).1–3 Carbonic anhydrase (CA) IX is upregulated by hypoxia and has a predominantly perinecrotic pattern of tumor cell expression.4–6 In serial sections taken from biopsies of basal and squamous cell skin and bladder transitional cell carcinomas, a high degree of overlap was found between CA IX and the recognized hypoxic marker pimonidazole.7–10 A prospective study found a close correlation between polarographic measurements of intratumoral oxygen tensions and CA IX immunostaining in cervical carcinoma.11 These findings indicate that CA IX expression may be used as a surrogate marker of cellular hypoxia. CA IX previously has been observed in 80% of resected NSCLC. In bronchial biopsies, negative CA IX expression distinguished preneoplastic lesions from carcinoma-in-situ and microinvasive disease.12 CA IX has been proposed as a tumor marker for renal cell carcinoma and as a biomarker of cervical dysplasia and neoplasia in Papanicolaou smears.13,14 In invasive breast cancer, increased expression of CA IX was associated with markers of an aggressive phenotype and a worse outcome.4,6 CA IX is regulated by the hypoxia-inducible factor (HIF)-1 transcription factor.10 HIF-1 is composed of two subunits, HIF-1 alpha and HIF-1 beta. HIF-1 is stabilized by hypoxia, and activates target genes that contain a hypoxic response element in their promoter region.15 Increased expression of HIF-1 and the related transcription factor HIF-2 have been implicated in tumorigenesis in NSCLC.16 In this study, we examined the prognostic implications and patterns of expression of CA IX in NSCLC.
Study Population The tissue specimens evaluated were taken from 204 consecutive patients with histologic stage I to IIIA NSCLC who had undergone resection of NSCLC with curative intent. Twenty-four patients who died within 61 days of surgery were excluded from the study to reduce the confounding variable of perioperative mortality; five more patients were excluded because of lack of available tissue. The 30-day mortality rate was 6.4%, which is marginally higher than the internationally published 30-day mortality rate (3.9%).17 Of the 175 stained specimens, 125 were from males and 50 were from females. The mean age at surgery was 64.8 years (SD, 8.0; range, 33.8 to 79.1). Positive resection margins were found in 15 specimens (8.5%), the presence of which was associated with stage IIIA disease (n = 8 of 15, P = .011). One patient received adjuvant chemotherapy, and of the 15 patients who were treated with adjuvant radiotherapy, 10 had stage IIIA, six had stage II, and one had stage I disease. The final staging was based on the findings at surgery and the histopathology report. Hospital notes of the patients were reviewed and, if necessary, the local cancer registry or patient’s general practitioner was contacted to complete follow-up. From these data it was established that 124 patients (70.5%) had died by the time of analysis and, of these deaths, 18 (10.2%) were not cancer related.
The specimens studied were routinely processed, formalin-fixed, and paraffin-embedded. Only blocks containing the advancing edge of the primary tumor were evaluated. Tissue sections of 4 µm thickness were cut onto glass slides that were previously treated with 2% 3-aminopropylethoxysilane (in methanol) and dried overnight at 37°C to assist with section-to-slide adhesion. The murine monoclonal antibody (Mab) M75 recognizing the N-terminal domain of MN/CA IX protein has been previously described by Pastorekova et al.18 The specificity of the monoclonal antibody M75 for CA IX was confirmed by Western blots and immunostaining of COS-7 cells transfected with CA IX cDNA.19 The secondary antibody was rabbit antimouse polymer from the Envision kit (Dako, Ely, United Kingdom).
Immunohistochemistry
Investigation of Stromal Cell Type Expressing CA IX
Interpretation of CA IX Staining and Tumor Necrosis Sections were assessed using light microscopy in a blinded fashion by two observers (D.E.B.S. and D.R.). If discrepancies were found, a consensus was reached using a conference microscope. Persistent discrepancies were adjudicated by a third observer (J.L.J). The percentage of tumor cells with membranous (m) CA IX expression and the percentage of tumor cells with cytoplasmic (c) expression were estimated. For subgroup analysis, mCA IX expression was divided into quartiles depending on the percentage of cells stained: negative, 0%; low, less than 5%; moderate,  5% to less than 30%; and extensive, 30%. Perinuclear (p) CA IX expression was classified as positive or negative. There were three distinct patterns classified as pCA IX expression: (1) staining obscuring the nuclear structures, (2) circumferential staining of the nuclear margins, and (3) discrete areas of staining adjacent to and indenting the nucleus (Fig 1 ). Stromal expression was assigned a score of 0 to 3 (0 = no stromal expression, 1 = occasional stromal expression, 2 = moderate stromal expression, and 3 = extensive stromal expression). The degree of tumor necrosis (TN) on the section was assessed by two independent observers and was scored as either limited or extensive.20
Estimation of the Oxygen Pressure (PO2) at Which mCA IX Positivity Starts To assess at what distance from blood vessels mCA IX staining commenced, three specimens were chosen that had extensive expression. Serial sections were stained for CD34 and CA IX, as described above. Measurements were only taken from blood vessels that were surrounded by tumor cells with positivity, so that three distances could be measured from each vessel in different directions. In addition, blood vessels that had been cut through an oval or longitudinal plain were avoided. In total, six blood vessels were suitable, and 18 measurements were made (Fig 2 ). The distance from the CD34 staining endothelium to mCA IX–positive tumor cells was measured using a graticule with a slide wire micrometer, calibrated with a calibration slide by D.E.B.S. This method previously was described by Beasley et al.5
Random Selection of Cells and Measurement of Microvessel to Tumor Cell Distance To compare the blood vessel to tumor cell distance for cells that did and did not express CA IX and for cells from areas of tumor with greater than or equal to and less than median staining, a selection of specimens was double stained for CA IX and CD34. Sections were initially stained with M75, and immunoreactivity was visualized with the Envision technique and DAB. Then the specimens were stained with anti-CD34, and immunoreactivity was visualized with alkaline phosphatase-antialkaline phosphatase and fuchsin (Table 1  ). To ensure random selection, specimens in the absent, low, moderate, and extensive mCA IX subgroups were allocated new identification numbers. A Microsoft Excel random number generator was then used to select pairs of specimens from each subgroup for double staining. Three areas from each specimen that encompassed tumor nests and blood vessels were photographed and printed onto A4 paper. Cells were randomly selected by placing a 10 x 6 square grid in the center of each photograph (Fig 3). Sixteen squares were randomly selected again using Excel (Fig 3). If a tumor cell was at the center of the selected square, the distance to the nearest blood vessel was measured using a pair of dividers (Fig 4). After a total of four measurements were taken from each photo, no further measurements were made. If tumor cells were found at the center of less than four squares, no further squares were chosen, and less than four measurements were accepted for that patient to avoid bias. The measurements were grouped first, as either from mCA IX–stained or unstained cells, and second, as either from areas of greater than or equal to mCA IX staining or from areas of less than median mCA IX staining, irrespective of whether the individual cell measured was stained.
Cell Culture, Lysis, Nuclear and Cytoplasmic Differential Extraction, and Immunoblotting The cell line was A549 human pulmonary adenocarcinoma from American Type Culture Collection (ATCC, Manassas, VA). Cells were grown in DMEM (Sigma-Aldrich Company Ltd, Dorset, England), supplemented with 10% fetal calf serum (Gibco Life Technologies, Paisley, United Kingdom), L-glutamine (2 µM), penicillin (50 IU/mL), and streptomycin sulfate (50 µg/mL). Studies of inducible gene expression were performed on cells approaching confluence. Parallel incubations were performed on aliquots of cells in normoxia (humidified air with 5% CO2) and hypoxia. Hypoxic conditions were generated in a NAPCO 7001 incubator (Precision Scientific, Winchester, VA) with 0.1% O2, 5% CO2, and the balance of N2 unless otherwise specified. Experimental exposures were performed in normal growth medium for 16 hours. Whole cell protein extracts were prepared from tissue culture cells prepared by 30 seconds of homogenization in denaturing conditions. For Western immunoblotting, aliquots were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to Immobilon-P membranes (Millipore [U.K.] Ltd, Stonehouse, Gloucestershire, United Kingdom). CA IX was detected using M75 (1:50 for 16 hours at 4°C). Horseradish peroxidase–conjugated secondary antimouse antibody 1:2,000 in phosphate-buffered saline containing 5% fat-free skimmed milk (Marvel Premier Brand, Spalding, Lincolnshire, United Kingdom) and 0.1% Tween 20 (Dako) was applied at room temperature for 1 hour. Bound antibodies were detected using the chemiluminescent substrate ECL+Plus (Amersham Pharmacia Biotech, Amersham, United Kingdom).
Statistics
CA IX Tumor Cell Expression Tumor cell CA IX immunostaining was detected in 81.8% (142) of patients. Mixed patterns of cytoplasmic and membranous expression were seen in the majority of patients (Table 2 ) (Fig 5). The median percentage for both mCA IX and cCA IX staining patterns was 5% (range, 0% to 100%). mCA IX and cCA IX expression was observed in carcinoma-in-situ and entrapped bronchial epithelium within the tumor mass (Fig 6).
pCA IX expression was observed in occasional discrete tumor cells that always had mCA IX expression. Of pCA IX–positive patients (n = 46), 95.6% had high mCA IX expression and 66.7% were in the extensive mCA IX subgroup, demonstrating a close association between these two staining patterns. Three patients with pCA IX expression also expressed nuclear CA IX. Positive pCA IX expression also correlated with increased cCA IX expression (P < .001) but not with stromal CA IX expression.
There was significantly higher positive pCA IX and high mCA IX expression in the squamous carcinomas than in other NSCLC histologic subtypes (P = .003 and 0.008, respectively) (Table 3
CA IX Stromal Cell Expression CA IX stromal cell expression was observed in 111 patients. Increased stromal expression was associated with high mCA IX (P = .003) but not with either pCA IX or cCA IX. Stromal cell CA IX was significantly associated with advanced tumor stage (P = .011). Serial sections were stained with CD31 and CD34, markers for vascular endothelium, and vimentin and smooth muscle actin, which are markers for myofibroblasts21–23 (see Table 1 ). A clear overlap was seen between CA IX, vimentin, and smooth muscle actin, indicating that the CA IX–positive stromal cells were myofibroblasts.
Associations With Tumor Necrosis
Serial sections demonstrated that there was little or no CA IX expression adjacent to vascular endothelium expressing CD31 and CD34, in contrast to the frequent perinecrotic expression (Fig 2 ).
Distance From Vascular Endothelium to the Start of CA IX Tumor Positivity
Mean Tumor Cell to Blood Vessel Distances for Randomly Selected Cells There was a significant difference between the distances from tumor to blood vessel for mCA IX–positive tumor cells (mean = 117.1 µm, standard deviation [STD] 54.2 µm, n = 26) and mCA IX–negative tumor cells (mean distance = 43.2 µm, STD = 34.1 µm, n = 61) (P < .001). There was also a significant difference between the distances from tumor to blood vessel for tumor cells from areas of less than median staining (mean = 37.7 µm, STD = 21.6 µm, n = 47) and from areas of greater than or equal to median staining (mean = 103.6 µm, STD = 53.7 µm, n = 40) (P < .001), irrespective of whether the actual cell selected was mCA IX–positive.
A549 Cell Lines Express Nuclear CA IX
Correlations With a Poor Prognosis Positive pCA IX (P = .036), extensive TN (P = .003), male sex (P = .034), increasing tumor-node-metastasis stage (P < .001), and high platelet count (P = .034) were univariate indicators of a poor prognosis. There was a trend toward a poor prognosis for high mCA IX patients (P = .11) and positive resection margins (P = .15) (Table 5 ). cCA IX and stromal expression were not associated with outcome (Fig 9 and Table 5). In multivariate analysis, increasing tumor stage (P < .001), high platelet count (P = .010), positive pCA IX expression (P = .004), male sex (P = .022), and extensive TN (P = .040) were independent poor prognostic factors (Table 6).
We have shown that pCA IX expression is an independent poor prognostic marker in NSCLC. pCA IX expression was only seen in cells with mCA IX staining, predominantly in patients with high mCA IX expression. Furthermore, high mCA IX expression is associated with extensive TN. Finally, mCA IX–positive cells are seen at an increased distance from the microvasculature compared with mCA IX–negative cells, in keeping with the postulate that hypoxia induces the enzyme. The frequency (81.8%) and predominant focal pattern of CA IX expression in this cohort is similar to that found in a previous series of 65 patients of NSCLC (80%).13 The perinecrotic pattern of expression is consistent with findings from studies in other tumor types.4–6 The proposal by Wykoff10 and Loncaster11 that CA IX is a marker of tumor cell hypoxia is supported first by the associations between the extent of TN and mCA IX, cCA IX, and stromal CA IX expression. Second, CA IX positivity only starts when the estimated PO2 is hypoxic (5.77 mmHg or 0.76% O2). This estimated PO2 is in keeping with previous studies in head and neck cancer using a similar method and is comparable to in vitro work demonstrating that CA IX is increasingly upregulated as the PO2 is decreased from 2.5% to 0.1%.5,10 An obvious criticism of the deduction that CA IX is a marker of hypoxia on the basis of the strength of the estimated PO2, is that the method used incorporates bias by only selecting tumor sections that have extensive expression and by excluding blood vessels that have been cut along their longitudinal axis. To counter this bias, we randomly selected tumor cells and measured the distance to the nearest blood vessel. The mean distances for positive cells and for cells from areas of tumor with high positivity found with this method were similar but higher than those found with the initial method. This discrepancy would be expected because the initial method measured the distance from the nearest tumor cell to the blood vessel, whereas the second method selected a random cell and measured the distance. Importantly, a highly significant difference was found between tumor cells that expressed CA IX and those that did not, strengthening the proposal that CA IX is a hypoxic marker. We have also demonstrated that hypoxia enhanced CA IX expression in the cytosolic and nuclear extract of A549 cells. The presence of a range of distances between microvessels and CA IX–positive tumor cells, as opposed to a set distance, has two explanations. First, there may be differences in tissue oxygen use throughout the tumor. Second, blood flow through often-fragile tumor microvessels may be intermittent, which would cause variations in the degree of hypoxia at varying distances from the microvessels. Squamous carcinomas are known to be more necrotic than other NSCLC histologic subtypes, suggesting that they are more hypoxic.25 In keeping with this observation, and the correlation between TN and mCA IX expression in this study, a higher expression of mCA IX was found in squamous tumors. The association between pCA IX and squamous carcinomas is not so readily explained; pCA IX was not associated with TN. However, pCA IX was only observed in mCA IX–positive cells, which suggests that hypoxia is in part responsible for pCA IX expression. Alternatively, a secondary process may be responsible for pCA IX expression that is more active in squamous carcinomas. Entrapped nonmalignant bronchial epithelium and, in particular, carcinoma-in-situ, often expressed mCA IX, which reflects the hypoxic nature of these tissues in contrast to adjacent normal lung tissue that never stained for CA IX. The lack of CA IX expression around microvessels suggests that these tissues are relatively well oxygenated and, by inference, that the microvessels are functionally active. Certain observations were made that were inconsistent with the proposal that hypoxia was the sole regulator of CA IX expression. On occasion, CA IX expression was observed in areas that would be expected to be well oxygenated, such as adjacent to blood vessels, and in patients with little or no TN. Furthermore, subgroup analysis of mCA IX expression found no significant difference between the extent of TN in the extensive and moderate subgroups and between the low- and no-expression subgroups. Finally, as previously stated, pCA IX expression did not correlate with TN even though all cells that expressed pCA IX coexpressed mCA IX. The transcription factor HIF-1 regulates CA IX; hence, aberrant regulation of HIF-1 would be expected to increase expression of CA IX. Disregulation of HIF-1 could be caused by mutation of the von Hippel-Lindau (VHL) tumor suppressor gene or activation of the epidermal growth factor receptor (EGFR). Mutated VHL prevents appropriate normoxic degradation of HIF-1 and, as such, CA IX is heavily expressed in mutated VHL-related tumors, independently of the extent of TN.10 VHL gene mutations have previously been demonstrated in a proportion of cell lines derived from thoracic malignancies.26 VHL mutations may be present in the subgroup of patients with extensive mCA IX expression and focal TN. With regard to EGFR signaling, in vitro studies show that activation of EGFR stimulates the phosphatidylinositol 3-kinase pathway, which in turn can upregulate HIF-1 alpha independently of hypoxia.27 Recent work has demonstrated that EGFR activation is able to augment hypoxic induction of another HIF-1 alpha transcription target, vascular endothelial growth factor (VEGF), via the phosphatidylinositol 3-kinase pathway.28 This suggests that EGFR activation may also augment induction of CA IX. The pCA IX group, which consists of patients with predominantly extensive mCA IX, may represent such processes acting in concert with tumor hypoxia. Conversely, TN was commonly observed in the absence of CA IX staining. This may represent infarction of nonhypoxic tissue caused by an acute ischemic event. Ausprunk et al29 showed that tumor microvasculature is extremely fragile and can easily be damaged. Under such circumstances an acute thrombus may occur, resulting in TN. The structures responsible for pCA IX positivity are unknown. However, this occasional pattern of staining may represent CA IX protein in the endoplasmic reticulum during periods of increased synthesis.30 Supporting this is the observation that pCA IX was predominantly found in patients with high mCA IX expression. Immunoblotting of whole cell extracts with M75 has previously been reported to demonstrate a doublet of apparent molecular weights 54 and 58 kd in cultured HeLa cells grown in dense cultures and tumorigenic HeLa cross fibroblast hybrids. Using immunoelectron microscopy, the CA IX staining was localized to the surface microvilli and the nucleus, particularly in the nucleoli in the HeLa cells.18,31 The presence of nuclear CA IX in three patients in our series shows that this is a rare but reproducible feature in NSCLC. In relation to this observation, a nuclear protein with CA enzymatic activity has been identified in human and rat testis and rat lymphocytes. This protein appeared to be identical to the transcription factor NonO/p54nrb on amino acid sequencing. NonO/p54nrb lacks the structural elements heretofore considered essential for zinc binding, which is essential for CA activity and as such has been classified as a nonclassical CA. The role of nuclear CA IX is as yet undetermined, but the observation of a known transcription factor with CA activity suggests that CA IX may also act as a transcription factor.32 In NSCLC, breast, and cervical cancer CA IX expression is detected in premalignant lesions, which suggests that hypoxia and CA IX are involved at an early stage in tumorigenesis.12 Our observation of CA IX expression in carcinoma-in-situ supports this hypothesis. The finding that positive resection margins approached significance as a univariate indicator of a poor prognosis is possibly the result of the small number of patients. Of note is the association between stage IIIA (N2) disease and the presence of positive resection margins. Previous studies of larger populations report that the prognosis for positive resection margins is similar to that of stage IIIA (N2) disease. However, studies differ about whether the poor prognosis conferred by positive margins is independent of that conferred by advanced stage of disease.33–36 CA IX may be a therapeutic target for cancer. Inhibition of CA isoenzymes with acetazolamide or sulfonamides results in either reduced tumor invasiveness or blocked tumor growth, respectively.37,38 Furthermore, CA isoenzyme antagonism has been observed to augment the cytotoxic effects of various chemotherapeutic agents, including platinum-based drugs.39 Additional work is required to investigate whether these results are reproducible in vivo. Work is under way investigating the possibilities of using CA IX as a target for antigen-specific immunotherapy in renal carcinoma, where it is widely expressed in the tumor but not in normal tissue.40 In summary, pCA IX expression has been identified as a prognostic variable in NSCLC. The associations between TN and different patterns of CA IX expression, the demonstration of a significant difference in the distance from blood vessels of tumor cells and tumor areas with and without CA IX staining, and the estimation of the PO2 at which CA IX positivity starts suggest that CA IX expression can be used as a surrogate marker of hypoxia in NSCLC. The lack of expression adjacent to microvessels indicates that these areas are well oxygenated and highlights the functional activity of these vessels. A small subset of tumors that have high mCA IX expression in the presence of low or absent TN indicates regulation of this enzyme by factors other than hypoxia in a proportion of patients. More work is required to identify the structures expressing pCA IX and to elucidate all the regulatory processes.
We thank Medisearch for funding, the Institutes of Cancer Studies and Lung Health for their support during this work, and Salli Muller, FRCPATH, and Catherine Richards, MRCPATH, for the use of their equipment and the benefit of their expertise.
1. Choi N, Baumann M, Flentjie M, et al: Predictive factors in radiotherapy for non-small-cell lung cancer: Present status. Lung Cancer 31:43–56, 2001[CrossRef][Medline]
2. Hockel M, Schlenger K, Aral B, et al: Association between tumor hypoxia and malignant progression in advanced cancer of the uterine cervix. Cancer Res 56:4509–4515, 1996 3. Stadler P, Becker A, Feldmann HJ, et al: Influence of the hypoxic subvolume on the survival of patients with head and neck cancer. Int J Radiat Oncol Biol Phys 44:749–754, 1999[CrossRef][Medline]
4. Wykoff CC, Beasley N, Watson PH, et al: Expression of the hypoxia-inducible and tumor-associated carbonic anhydrases in ductal carcinoma in situ of the breast. Am J Pathol 158:1011–1019, 2001
5. Beasley NJ, Wykoff CC, Watson PH, et al: Carbonic anhydrase IX, an endogenous hypoxia marker, expression in head and neck squamous cell carcinoma and its relationship to hypoxia, necrosis, and microvessel density. Cancer Res 61:5262–5267, 2001
6. Chia SK, Wykoff CC, Watson PH, et al: Prognostic significance of a novel hypoxia-regulated marker, carbonic anhydrase IX, in invasive breast carcinoma. J Clin Oncol 19:3660–3668, 2001 7. Raleigh JA, Chou SC, Arteel GE, et al: Comparisons among pimonidazole binding, oxygen electrode measurements, and radiation response in C3H mouse tumors. Radiat Res 151:580–589, 1999[Medline] 8. Haustermans K, Hofland I, Van de Pavert L, et al: Diffusion-limited hypoxia estimated by vascular image analysis: Comparison with pimonidazole staining in human tumors. Radiother Oncol 55:325–333, 2000[CrossRef][Medline] 9. Varia MA, Calkins-Adams DP, Rinker LH, et al: Pimonidazole: A novel hypoxia marker for complementary study of tumor hypoxia and cell proliferation in cervical carcinoma. Gynecol Oncol 71:270–277, 1998[CrossRef][Medline]
10. Wykoff CC, Beasley NJ, Watson PH, et al: Hypoxia-inducible expression of tumor-associated carbonic anhydrases. Cancer Res 60:7075–7083, 2000
11. Loncaster JA, Harris AL, Davidson SE, et al: Carbonic anhydrase (CA IX) expression, a potential new intrinsic marker of hypoxia: Correlations with tumor oxygen measurements and prognosis in locally advanced carcinoma of the cervix. Cancer Res 61:6394–6399, 2001 12. Liao SY, Stanbridge EJ: Expression of MN/CA9 protein in Papanicolaou smears containing atypical glandular cells of undetermined significance is a diagnostic biomarker of cervical dysplasia and neoplasia. Cancer 88:1108–1121, 2000[CrossRef][Medline]
13. Vermylen P, Roufosse C, Burny A, et al: Carbonic anhydrase IX antigen differentiates between preneoplastic malignant lesions in non-small-cell lung carcinoma. Eur Respir J 14:806–811, 1999
14. Liao SY, Aurelio ON, Jan K, et al: Identification of the MN/CA9 protein as a reliable diagnostic biomarker of clear cell carcinoma of the kidney. Cancer Res 57:2827–2831, 1997
15. Tsuzuki Y, Fukumura D, Oosthuyse B, et al: Vascular endothelial growth factor (VEGF) modulation by targeting hypoxia-inducible factor-1 alpha -> hypoxia response element -> VEGF cascade differentially regulates vascular response and growth rate in tumors. Cancer Res 60:6248–6252, 2000 16. Giatromanolaki A, Koukourakis M, Sivridis E, et al: Relation of hypoxia inducible factor 1 alpha and 2 alpha in operable non-small cell lung cancer to angiogenic/molecular profile of tumours and survival. Br J Cancer 85:881–890, 2001[CrossRef][Medline] 17. Rubins JB, Ewing SL, Leroy S, et al: Temporal trends in survival after surgical resection of localized non-small cell lung cancer. Lung Cancer 28:21–27, 2000[CrossRef][Medline] 18. Pastorekova S, Zavadova Z, Kostal M, et al: A novel quasi-viral agent, MaTu, is a two-component system. Virology 187:620–626, 1992[CrossRef][Medline]
19. Saarnio J, Parkkila S, Parkkila AK, et al: Immunohistochemical study of colorectal tumors for expression of a novel transmembrane carbonic anhydrase, MN/CA IX, with potential value as a marker of cell proliferation. Am J Pathol 153:279–285, 1998 20. Swinson DE, Jones JL, Richardson D, et al: Tumour necrosis is an independent prognostic marker in non-small cell lung cancer: Correlation with biological variables. Lung Cancer 37:235–240, 2002[CrossRef][Medline] 21. O’Byrne KJ, Koukourakis MI, Giatromanolaki A, et al: Vascular endothelial growth factor, platelet-derived endothelial cell growth factor and angiogenesis in non-small-cell lung cancer. Br J Cancer 82:1427–1432, 2000[Medline]
22. Cox G, Jones JL, Andi A, et al: A biological staging model for operable non-small cell lung cancer. Thorax 56:561–566, 2001
23. Ishizaki M, Zhu G, Haseba T, et al: Expression of collagen I, smooth muscle alpha-actin, and vimentin during the healing of alkali-burned and lacerated corneas. Invest Ophthalmol Vis Sci 34:3320–3328, 1993 24. Thomlinson RH, Gray LH: The histological structure of some human lung cancers and the possible implications for radiotherapy. Br J Cancer 9:539–549, 1955[Medline] 25. Underwood JCE, Hunter J: General and systematic pathology. Churchill Livingstone, New York, 1994, p 343 26. Sekido Y, Bader S, Latif F, et al: Molecular analysis of the von Hippel-Lindau disease tumor suppressor gene in human lung cancer cell lines. Oncogene 9:1599–1604, 1994[Medline]
27. Zhong H, Chiles K, Feldser D, et al: Modulation of hypoxia-inducible factor 1 alpha expression by the epidermal growth factor/phosphatidylinositol 3-kinase/PTEN/AKT/FRAP pathway in human prostate cancer cells: Implications for tumor angiogenesis and therapeutics. Cancer Res 60:1541–1545, 2000 28. Clarke K, Smith K, Gullick WJ, et al: Mutant epidermal growth factor receptor enhances induction of vascular endothelial growth factor by hypoxia and insulin-like growth factor-1 via a PI3 kinase dependent pathway. Br J Cancer 84:1322–1329, 2001[CrossRef][Medline] 29. Ausprunk DH, Folkman J: Vascular injury in transplanted tissues: Fine structural changes in tumor, adult, and embryonic blood vessels. Virchows Arch B Cell Pathol 31–44, 1976 30. Alberts A, Watson JD, Roberts K, et al: Molecular Biology of the Cell (ed 3). London, UK, Garland Publishing, 1989, pp 433–451 31. Zavada J, Zavadova Z, Pastorekova S, et al: Expression of MaTu-MN protein in human tumor cultures and in clinical specimens. Int J Cancer 54:268–274, 1993[Medline]
32. Karhumaa P, Parkkila S, Waheed A, et al: Nuclear NonO/p54(nrb) protein is a nonclassical carbonic anhydrase. J Biol Chem 275:16044–16049, 2000
33. Hofmann HS, Taege C, Lautenschlager C, et al: Microscopic (R1) and macroscopic (R2) residual disease in patients with resected non-small cell lung cancer. Eur J Cardiothorac Surg 21:606–610, 2002 34. Deinemann H, Trainer C, Hoffmann H, et al: Incomplete resections in bronchial carcinoma: Morbidity and prognosis. Chirurg 68:1014–1019, 1997[CrossRef][Medline]
35. Ghiribelli C, Voltolini L, Paladini P, et al: Treatment and survival after lung resection for non-small cell lung cancer in patients with microscopic residual disease at the bronchial stump. Eur J Cardiothorac Surg 16:555–559, 1999 36. Liewald F, Hatz RA, Dienemann H, et al: Importance of microscopic residual disease at the bronchial margin after resection for non-small cell carcinoma of the lung. J Thorac Cardiovasc Surg 104:408–412, 1992[Abstract] 37. Parkkila S, Rajaniemi H, Parkkila AK, et al: Carbonic anhydrase inhibitor suppresses invasion of renal cancer cells in vitro. Proc Natl Acad Sci USA 5:2220–2224, 2000 38. Supuran CT, Briganti F, Tilli S, et al: Carbonic anhydrase inhibitors: Sulfonamides as antitumor agents? Bioorg Med Chem 9:703–714, 2001[CrossRef][Medline] 39. Teicher BA, Liu SD, Liu JT, et al: A carbonic anhydrase inhibitor as a potential modulator of cancer therapies. Anticancer Res 13:1549–1556, 1993[Medline] 40. Nakagawa Y, Uemura H, Shimizu K, et al: The role of MN/CA IX antigen in carcinogenesis and metastasis of renal cell carcinoma. Hinyokika Kiyo 47:809–814, 2001[Medline] Submitted November 27, 2001; accepted October 1, 2002. This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
INTRODUCTION
2 test was used to analyze the associations between categorical variables. The statistical significance of trends between the increasing values of different categorical variables was analyzed using the linear-by-linear association test. Survival curves were plotted using the Kaplan-Meier method, and the statistical significance was assessed using the log-rank test. A Cox proportional hazards regression model was used to identify statistically significant differences in survival and estimate hazard ratios and 95% confidence intervals. The assumption of proportional hazards was assessed graphically by plotting log-minus-log of survival against time for each of the prognostic groups. Prognostic variables identified by univariate analysis with P < .1 were entered into a Cox multivariate analysis model. A forward, stepwise selection procedure was used, with variables added to the model according to a partial likelihood ratio test, using an entry criterion of P < .05. 
