Alternative splicing variants of carbonic anhydrase IX in human non-small cell lung cancer

Lung Cancer 64 (2009) 271–276

Contents lists available at ScienceDirect

Lung Cancer journal homepage: www.elsevier.com/locate/lungcan

Alternative splicing variants of carbonic anhydrase IX in human non-small cell lung cancer Francesca Malentacchi a , Lisa Simi a , Caterina Nannelli a , Matteo Andreani b , Alberto Janni b , Silvia Pastorekova c , Claudio Orlando a,∗ a b c

Clinical Biochemistry Unit, Department of Clinical Physiopathology, University of Florence, Viale Morgagni 6, I-50139 Florence, Italy Department of Thoracic Surgery, Azienda Ospedaliera Careggi, Florence, Italy Institute of Virology, Slovack Academy of Sciences, Bratislava, Slovak Republic

a r t i c l e

i n f o

Article history: Received 3 July 2008 Received in revised form 3 October 2008 Accepted 6 October 2008 Keywords: Hypoxia Alternative splicing Cancer progression Pericellular pH regulation Prognosis Real-time RT-PCR

a b s t r a c t In human cancers, carbonic anhydrase IX (CAIX) contributes to maintain intracellular and extracellular pH under hypoxic conditions, but also influences regulation of cell proliferation and tumor progression. CaIX was previously indicated as an independent prognostic marker in non-small cell lung carcinoma (NSCLC). Very recently a CAIX alternative splicing isoform, generating a transcript lacking of exons 8–9, was detected in cancer cells independently from the levels of hypoxia. This alternative splicing (AS) generates a truncated protein lacking the transmembrane region, the intracellular tail and the C-terminal of the catalytic domain and competes with the full-length (FL) isoform in the regulation of the extracellular pH, mainly in a mild hypoxic status. In the present study we measured the mRNA expression of FL and AS CAIX isoforms in 101 NSCLC and in paired not affected tissues. The two isoforms were coexpressed in all NSCLC and normal tissues but while AS mRNA was prevalent in normal tissues (66 ± 3%), the FL isoform was higher in NSCLC (58 ± 2%, p = 0.001). FL mRNA, but not AS, was statistically increased in NSCLC (p = 0.01) and showed a statistical association with lymphnode involvement (p = 0.009) and tumor stage (p = 0.04). Global survival analysis of cancer/related death showed that high levels of FL mRNA were predictive of unfavorable outcome (p < 0.0001) and shorter disease-free survival (p < 0.0001). Multivariate analysis indicated that FL is an independent prognostic factor for overall survival and higher levels of mRNA in NSCLC sensibly increase hazard ratio (∼sixfold). In conclusion, our results seems to indicate that, at least in NSCLC, FL CAIX is the most accurate surrogate of hypoxic stress and represents the only variant with a prognostic role. These data indicate the importance of a separate measurement of the two isoforms in cancer and the need of an accurate re-evaluation of most studies on the clinical role of CAIX in cancer diagnosis. © 2008 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Cancer progression is accompanied by deep alterations of microenvironmental conditions in which tumor cells proliferate, including the progressive reduction of oxygen supply. Hypoxia is a common event in locally advanced solid tumors, frequently associated to cancer progression, genetic instability, selection for resistance to apoptosis, increased risk of invasion and metastasis, poor response to radiation and chemotherapy [1–5]. Hypoxia results in the upregulation of genes that facilitate anaerobic metabolism and promote tumor vascularisation (e.g. vascular

∗ Corresponding author. Tel.: +39 055 4271440; fax: +39 055 4271413. E-mail address: [email protected]fi.it (C. Orlando). 0169-5002/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.lungcan.2008.10.001

endothelial growth factor, VEGF). The transcriptional complex hypoxia-inducible factor-1 (HIF-1) is the major mediator of gene expression in hypoxic cancer cells. HIF-1␣ can regulate the coordinated expression of several genes through the interaction with hypoxic responsive elements in promoters of genes involved in energy metabolism, angiogenesis and with surface transmembrane carbonic anhydrases [6]. The activation of these genes contributes to the selection of more resistant and aggressive cancer clones with increased capacity of invasion and metastasis. Carbonic anhydrase IX (CAIX) was proposed as a surrogate marker of hypoxia, strongly HIF-1 dependent [7–9]. CAIX is a transmembrane glycoprotein, member of a family of zinc metalloenzymes, that reversibly converts carbon dioxide and water to carbonic acid, contributing to acid–base balance. CAIX plays a relevant role in pH regulation and its expression reduces the peri-

272

F. Malentacchi et al. / Lung Cancer 64 (2009) 271–276

cellular pH, facilitating breakdown of the extracellular matrix [10]. This leads to intracellular alkalosis and extracellular acidosis in the tumor microenvironment, allowing to tumor cells to survive under hypoxic conditions [11,12]. While most CA isoforms are uniformly expressed in normal differentiated cells, CAIX is predominant in cancer cells [13]. In addition, CAIX plays a role in cell proliferation and cellular transformation [14]. Since the response of cells to radiation is dependent on oxygen [15], CAIX has been linked to negative response to radiotherapy [16]. Finally, CAIX expression has been correlated with clinical outcome in several human cancers [17–21]. Several works have shown the role of CAIX in non-small cell lung carcinoma (NSCLC). CAIX is not expressed in normal lung, dysplasia, metaplasia and basal cell hyperplasia, but it is expressed in “in situ” neoplasia and carcinomas [22]. In these cancers, CAIX is variably expressed, ranging from 65% in invasive adenocarcinomas to 100% in invasive squamous cell carcinomas [23]. Furthermore, in NSCLC HIF1-␣ and CAIX are strongly correlated and they appear to be co-localized in the same areas of the tumor [24]. Other authors demonstrated an association between CAIX expression, tumor necrosis factor and Ki67, suggesting a correlation with an increased oxygen consumption [25]. A strong inverse correlation between CAIX and tumor oxygenation was demonstrated, confirming its role as endogenous surrogate of hypoxia marker [26]. Finally, high expression CAIX was related to reduced overall survival and disease-free survival full-length (FL) CAIX l [27,28]. Very recently, besides the expected FL mRNA, a CAIX alternative splicing isoform was detected in normal and in cancer cells, independently from the levels of hypoxia [29]. This alternative splicing (AS) generates a transcript lacking of exons 8–9 and encodes a truncated CAIX protein lacking the transmembrane region, the intracellular tail and the C-terminal part of the catalytic domain. AS CAIX shows a diminished catalytic activity that reduces the capacity of the full-length CAIX protein to acidify extracellular pH of hypoxic cells and to bind carbonic anhydrase inhibitor. Cells transfected with the AS variant generate spheroids that do not form compact cores, suggesting that they fail to adapt to hypoxic stress. The splicing variant can functionally interfere with the FL CAIX in the regulation of the extracellular pH, mainly in cells under mild hypoxic status [29]. Due to the different functional role of the two CAIX isoforms, their separated measurement is mandatory when clinical and prognostic studies are performed. Starting from this consideration, in the present study we measured the mRNA expression of FL and AS CAIX isoforms in 101 NSCLC and in paired not affected tissues, to explore their independent clinical role. 2. Material and methods 2.1. Tissues samples The study group included 101 consecutive patients (age range from 37 to 82 years; median age: 68.5 years) who underwent surgical resection for lung cancer. Clinical and pathological features of patients (see Table 1) were assessed according to the WHO classification [30] and the TNM staging system. NSCLC samples and corresponding non-affected tissues, obtained from the same patient, were immediately snap-frozen in liquid nitrogen. Control tissues were taken at 3–5 cm far from cancer during lobectomy and 5–7 cm in pneumonectomy. Comparable tissue samples were processed for routine histological examination. For RNA extraction, tissues were disrupted by homogenization in 600 ␮l of guanidine-isothiocyanate lysis buffer (QIAGEN, Italy) added ␤-merchaptoethanol. Total RNA was extracted with QIAshedder and RNeasy MiniKit QIAGEN® columns. RNA was eluted from the columns with 50 ␮l of RNAase free water. Samples were

Table 1 Clinical and pathological features of 101 NSCLC. Patients Total

101

Gender Males Females

84 17

Age <55 years ≥55 years

8 93

Smokers Yes Ex Not

45 49 7

Histology Adenocarcinomas Squamous Adenosquamous

40 45 16

Grade Higher Medium Low

21 67 13

T T1 T2 T3 T4

26 63 7 5

Lymphnodes N− N+a

57 44

Stage 1 2 3

53 15 33

a

Node staging was N1 in 18 patients, N2 in 26 patients and N3 in 0 patients.

treated with RNase free DNase set QIAGEN® to eliminate DNA. Total RNA concentration were determinated with NanoDrop® ND1000 Spectrophotometer. The integrity of total RNA was verified in all samples with Agilent 2100 bioanalyzer. GAPDH mRNA was measured with the Pre-Developed TaqMan Assay Reagent, GAPDH endogenous control kit from Applied Biosystems (Foster City, CA, USA). One hundred nanogram of RNA samples were reversetranscribed in 20 ␮l of final volume in a reaction mixture containing 2 ␮l of TaqMan RT buffer 10×, 5.5 mM MgCl2 , 500 ␮M of dNTPs, 2.5 ␮M of random hexamers, 0.4 U/␮l RNase inhibitors and 1.25 U/␮l Multiscribe reverse transcriptase. The profile of the one-step reverse transcription reaction was 10 min at 25 ◦ C, 30 min at 48 ◦ C and 2 min at 95 ◦ C. 2.2. Qualitative reverse transcriptase PCR To evaluate the presence of the CAIX isoforms we performed a qualitative RT-PCR, using the same three sets of primers as previously described [29]. The h7S–h8A set identified the FL isoform, whereas the h7S–h10/7A set detected the AS variant. To confirm the results, we used the h7S–h11A primers to amplify the two isoforms, simultaneously. PCR reactions were performed in a final volume of 25 ␮l with 10 ␮g of cDNA, PCR buffer 1×, MgCl2 1.5 mM, dNTPs 200 ␮M, each primer 1 ␮M and 0.05 U Taq Gold (Applied Biosystems) at the following conditions: 10 min at 95 ◦ C, 30 s at 95 ◦ C, 30 s at 62 ◦ C and 20 s at 72 ◦ C for 40 cycles and a final extension at 72 ◦ C for 10 min. Products were analyzed on 2% agarose gel. The two amplification products obtained from PCR reaction with h7S–h11A

F. Malentacchi et al. / Lung Cancer 64 (2009) 271–276

273

primers were extracted from agarose gel accordingly to the instruction of NucleoSpin Extract II (Macherey-Nagel, Italy). The products were then sequenced to confirm their identity and localize the splicing site. 2.3. Quantification of CAIX mRNA Measurement of total and full-length form of CAIX mRNA was performed by quantitative real-time RT-PCR method, based on TaqMan technology. Probe and primers were selected by the software “Primer Express” (Applied Biosystems, Monza, Italy). The following sets of probe and primers were chosen (NCBI accession code NM 001216): for the total mRNA the probe was 5 -CAG GGA CAA AGA AGG GGA TGA CCA-3 , labelled with FAM, forward primer 5 -CCT CAA GAA CCC CAG AAT AAT GC-3 and reverse primer 5 CCT CCA TAG CGC CAA TGA CT-3 ; for the full-length isoform the probe was: 5 -ACA CCC TCT CTG ACA CC-3 , labelled with TET, forward primer 5 -CAA TAT GAG GGG TCT CTG ACTA–3 , reverse primer: 5 -CTC AAT CAC TCG CCC ATT CAA A–3 . PCR reaction was performed with 25 ng of cDNA in a final volume of 12.5 ␮l, in a reaction mix containing 6.25 ␮l of TaqMan Universal Master Mix, 300 nM of forward and reverse primers, for each set and 200 nM of each fluorescent probe. Plates were incubated 2 min at 50 ◦ C, 10 min at 95 ◦ C and than submitted to 40 cycles of amplification at 95 ◦ C for 15 s and 60 ◦ C for 60 s in ABI Prism 7700 Sequence Detector PE (Applied Biosystems, Monza, Italy). External reference curve was obtained by serial dilution, from 1 × 107 to 1 × 102 copies, of a plasmid vector containing the full-length isoform of CAIX (pGEX-3X-CA9) [29]. Results for CAIX mRNA were expressed as copies/␮g total RNA. To test the specificity of our assays we randomly sequenced 10% (n = 11) products obtained by the amplification with the same primers used in real-time RT-PCR. In all cases the sequence of amplified products corresponded to expected one (data not shown). 2.4. Statistical analysis Statistical analysis was carried out using the SPSS software package (SPSS inc., Chicago, IL). Statistical differences between groups were assessed by t-test and ANOVA analysis. For analysis of follow-up data, life table curves were calculated using Kaplan–Meier method and survival distribution were compared by log-rank statistics. The primary end point was cancer-related survival, as measured from the date of surgery to the time of the last follow-up or cancer-related death. The joint effects with already recognized prognostically relevant variables were examined via Cox proportional hazard analysis. Clinical parameters were entered stepwise forward into the model to test these covariables for possible joint effects with high/low levels of CAIX mRNA expression and the relative expression of the full-length respect to the alternative splice variant. Differences were considered statistically significant with p > 0.05.

Fig. 1. RT-PCR analysis of CAIX isoforms in paired NSCLC and normal mucosa. cDNA from samples was amplified separately for FL CAIX (primers h7S and h8A: 154 bp) and for AS CAIX (primers h7S and h10/7A: 140). AS and FL were amplified simultaneously with h7S and h11A primers (FL: 412 bp and AS: 241 bp).

amplification products: for FL the length was 154 bp and 140 bp for AS variant, respectively. Finally we amplified simultaneously AS and FL isoforms in a PCR using h7S and h11A primers. The RTPCR analysis of CAIX isoforms indicated that FL and AS variants were simultaneously detectable in NSCLC and paired apparently not affected control tissues as reported in Fig. 1. To confirm the localization of the splicing site, PCR products were resolved by gel electrophoresis and the two bands corresponding to expected FL and AS amplicons were recovered from gel, extracted and sequenced. As reported in Fig. 1, the AS isoform resulted from the loss of exons 8 and 9, whereas the FL product contained the entire coding sequence of CAIX gene. 3.2. Real-time quantification of mRNA CAIX isoforms in NSCLC and paired normal tissues The expression of total CAIX mRNA in 101 NSCLC specimens was significantly higher (p = 0.02) than in the paired apparently not affected control tissues. Similarly, when the absolute expression of the two isoforms was evaluated separately, we found that FL CAIX mRNA was significantly more elevated in NSCLC in comparison to paired normal tissues (p = 0.01), whereas AS CAIX was not statistically different (see Fig. 2) We also evaluated the mRNA expression of the two variants as the percentage of total CAIX expression and we clearly demonstrated that the distribution of FL and AS isoforms was statistically opposite between NSCLC and related controls. The AS variant was prevalent in normal tissues (66 ± 3% of the total), whereas in NSCLC the same isoform was significantly reduced and FL the most expressed variant (58 ± 2%) (p = 0.001) (see Fig. 3).

3. Results 3.1. CAIX isoforms in NSCLC and paired normal tissues For RT-PCR the same primer sets previously described [29] were used to perform three separate PCR experiments. Since it was previously reported that AS CAIX does not contain the exons 8 and 9, we used the forward primer located on exon 7 (h7S) which is common for both the transcripts, in combination with primer h8A, present only on FL variant. Conversely, we amplified the AS fragment from a PCR with h7S and h10/7A primers. From these separated PCRs we obtained in NSCLC and paired normal tissues the two expected

Fig. 2. Absolute levels of FL and AS CAIX isoforms in 101 NSCLC and corresponding not affected mucosae. n.s. = not statistically different.

274

F. Malentacchi et al. / Lung Cancer 64 (2009) 271–276

Fig. 3. Percent distribution of AS and FL CAIX in 101 NSCLC and paired normal mucosae. The AS variant was 66 ± 3% of the total in not affected tissues, whereas in NSCLC FL CAIX 58 ± 2% (2 test, p = 0.001).

3.3. CAIX isoforms and clinical features We compared the absolute levels of FL and AS CAIX isoforms in patients classified on the basis of conventional clinical–pathological parameters. We did not find any difference for FL and AS mRNA levels according to patient age (more or less 55 years), gender, tumor grade and T. Also the stratification according to tumor histology did not reveal significant differences among adenocarcinomas, squamous cell carcinomas and adenosquamous cell carcinomas for FL and AS mRNA expression. Conversely, we found that in node positive patients FL CAIX mRNA was significantly higher (p = 0.009) than in node negative patients, whereas AS mRNA levels were not statistically different (see Fig. 4). No difference was found comparing N1 and N2 patients for FL CAIX mRNA expression. Finally, patients in advanced stage (stage 2 or 3) had a significantly higher expression of FL mRNA (p = 0.04), but not of the AS variant, in comparison to patients with a stage 1 NSCLC. 3.4. CaIX isoforms and patient outcome To evaluate the prognostic value of CAIX isoform expression, patients were stratified according to an arbitrary cut-off corresponding to the median levels of their expression in NSCLC (488 for AS and 1280 copies/␮g total RNA for FL, respectively). Kaplan–Meier analysis was performed by comparing patients with high mRNA expression (>median value) vs. low expressing NSCLC (≤median

value). According to this division we found that the elevated expression of FL mRNA was significantly related to a reduced disease-free survival (DFS) (log-rank, Mantel-Cox, 2 = 17.6 and p < 0.0001) and a strong predictor of adverse overall survival (OS) (2 = 17.8 and p < 0.0001). On the contrary, the expression of AS mRNA was not related to DFS (p = 0.07) and showed a reduced correlation with patient OS (2 = 6.4 and p = 0.011) (see Fig. 5). Following univariate analysis to evaluate the prognostic value of clinical and experimental markers on cancer-related survival we found that lymphnode involvement and clinical stage were significantly associated to a worse prognosis of our patients, as well as the presence of high levels of mRNA expression of total CAIX and its two isoforms, with a maximal significance for FL in comparison to AS variants (see Table 2). Conversely, the multivariate analysis showed that only FL CAIX mRNA maintained its significance. The relative risk of cancerrelated death was more than sixfold increased in NSCLC with high FL expression in comparison to low expressing cancers. On the contrary AS expression did not maintain any clinical value in multivariate analysis. All patient in stages 3A and 3B (n = 33) underwent adjuvant therapy and 12 of them (T3) received neoadjuvant therapy. However we did not find any significant relation between patient survival and therapy (data not shown). None of the other patients received extra-surgical treatment. 4. Discussion Hypoxia is the natural consequence of altered microcirculation in cancers [31]. Since a correct blood supply is required for exploiting anti-cancer effects of radiotherapy or chemotherapy, tumor hypoxia was regarded as a predictor of poor response in cancer treatment [32,33]. The presence of hypoxia has been also correlated with a more aggressive phenotype in a subset of cancer cells, selected on the basis of hypoxia-induced changes and increased survival under hypoxic microenvironment. These clones have a more aggressive phenotype resulting in invasion and metastasis [31,34]. CAIX is a transmembrane enzyme frequently overexpressed in a variety of tumors, and associated with hypoxia, tumor aggressiveness and worse prognosis [35–37]. A recent study indicated the existence of two variants of CAIX mRNA. The FL transcript, coding for the entire sequence deriving from the native CAIX gene with 11 exons and a variant form without exons 7–8, that codes for a truncated protein lacking of the transmembrane region, the

Fig. 4. Absolute levels of FL and AS CAIX isoforms in 101 NSCLC classified according to tumor stage (T1, n = 26; T2, n = 63; T3, n = 12) and the presence of lymphnodes metastasis (node negative N−, n = 57 and node positive N+, n = 44).

F. Malentacchi et al. / Lung Cancer 64 (2009) 271–276

275

Fig. 5. Kaplan–Meier analysis estimates of overall survival (OS) and disease-free survival (DFS) in 101 NSCLC. Patients with high FL CAIX mRNA expression (≥median value = 1280 copies/␮g total RNA) have a significant reduction of OS probability (panel A, p < 0.0001) and DSF interval (panel B, p < 0.0001). In patients with high AS CAIX mRNA expression (≥median value = 488 copies/␮g total RNA) the significance is reduced for OS (panel C, p = 0.01) and absent for DFS (panel D, p = n.s.).

intracellular tail and the catalytic domain [29]. This difference influences the cellular localization of the variants: whereas FL CAIX is typically a plasma membrane protein, AS has also a cytoplasmatic localization. In addition AS seems to compete functionally with FL, by reducing the ability of FL to promote extra-cellular acidification. In conclusion AS CAIX expression seems to be not related to the hypoxic adaptation of neoplastic cells during cancer progression and, for some extents, to act as a natural competitor of this mechanism of aggressive clone selection. Due to their opposite role, it is evident that for the use of CAIX as surrogate marker of hypoxia and cancer prognosis we must evaluate separately the expression of the two isoforms. Several previous studies reported a significant correlation between CAIX iperexpression and reduced overall survival in NSCLC [10,23–25,38], mostly based on immunohistochemical (IHC) detection of secreted CAIX protein. Most of these studies used M75

antibody which does not differentiate between two variants [29], because it binds to a linear epitope localized in N-terminal, not affected by splicing. However IHC studies predominantly evaluated the membrane staining of CAIX, we can assume that the membrane staining might have similar prognostic value as the detection of the FL mRNA. Unfortunately IHC cannot be used in transcriptional profiling of tumors with the prognostic intent and accurate quantitative mRNA assays should improve patient stratification. In our previous study [28], the measurement of CAIX mRNA was performed in a cDNA sequence (exons 1–3) that appears now as common to both CAIX isoforms. In the present study we evaluated the clinical role of FL and AS CAIX mRNA expression and their relation with clinical and pathological features of 101 NSCLC. Both isoforms were expressed in NSCLC and paired not affected tissues. Whereas the absolute expression of FL mRNA was significantly increased in cancers (p < 0.01), the mean levels of AS were

Table 2 Univariate and multivariate analysis in 101 NSCLC. Risk factor

Univariate analysis (p-value)

Multivariate analysis Relative risk

Gender Age (<55 vs. ≥55) Smoker (yes vs. no) Histology Grade pT status pN status Tumor stage Total CAIX mRNA AS CAIX mRNA FL CAIX mRNA

0.973 0.153 0.457 0.793 0.236 0.108 0.008 0.003 0.001 0.013 0.000

– – – – – – 1.077 1.589 0.394 1.757 6.178

95% CL – – – – – – 0.395–2.941 0.891–2.835 0.085–1.823 0.805–3.837 1.565–24.378

p-value – – – – – – 0.884 0.117 0.233 0.157 0.009

276

F. Malentacchi et al. / Lung Cancer 64 (2009) 271–276

not statistically different between the two series. The differences in the distribution of the two variants were more evident when their expression was proportionally calculated. As expected the AS is strongly prevalent in not affected tissues, representing about the 70% of total CAIX expression. This means that FL is expressed also in non-cancer lung cells. This finding can be explained with the presence of low but easily detectable FL transcript in normal cells as previously reported [29]. However it is important to remark that most of our NSCLC patients were smokers (44%) or ex-smokers (49%). The occurrence of a mild hypoxic condition in these patients cannot be excluded even in apparently normal epithelial cells and this could be the explanation of a mechanism of adaptation to a reduced O2 supply. The presence of AS mRNA in cancer is less surprising: even in this case the presence of this mRNA can be a physiological features and, in addition, the presence of normal cells in cancer tissue, can represent the source of this variant. However, independently from their origin, the FL represents the predominant isoform of CAIX in NSCLC. In addition, higher levels of FL mRNA were significantly associated to the presence of lymphnode metastases and higher stage NSCLCs, whereas AS was not related to any clinico-pathological index. The survival analysis of cancer-related death in our group of patients suggested that high levels of FL expression were highly related to patient outcome, both for OS and DFS (p < 0.0001). In comparison, the prognostic role of AS isoform appeared negligible for OS (p = 0.01) and absent for DFS. Finally, multivariate analysis revealed that only FL is an independent prognostic factor for overall survival and that the presence of higher levels of mRNA in NSCLC can sensibly increase the hazard ratio. In conclusion our results seems to indicate that, at least in NSCLC, FL CAIX is the most accurate surrogate of hypoxic stress in cancer cells, representing the only variant with a prognostic role. These data indicate the importance of a separate measurement of the two isoforms in cancer diagnosis and the need of an accurate re-evaluation of most studies on the clinical role of CAIX in cancer diagnosis. Conflict of interest No conflict of interest. References [1] Brizel DM, Scully SP, Harrelson JM, Layfield LJ, Dodge RK, Charles HC, et al. Radiation therapy and hyperthermia improve the oxygenation of human soft tissue sarcomas. Cancer Res 1996;56:5347–50. [2] Graeber TG, Osmanian C, Jacks T, Housman DE, Koch CJ, Lowe SW, et al. Hypoxia-mediated selection of cells with diminished apoptotic potential in solid tumours. Nature 1996;379:88–91. [3] Höckel M, Schlenger K, Mitze M, Schäffer U, Vaupel P. Hypoxia and radiation response in human tumors. Semin Radiat Oncol 1996;6:3–9. [4] Reynolds TY, Rockwell S, Glazer PM. Genetic instability induced by the tumor microenvironment. Cancer Res 1996;56:5754–7. [5] Kim CY, Tsai MH, Osmanian C, Graeber TG, Lee JE, Giffard RG, et al. Selection of human cervical epithelial cells that possess reduced apoptotic potential to low-oxygen conditions. Cancer Res 1997;57:4200–4. [6] Sly WS, Hu PY. Human carbonic anhydrases and carbonic anhydrase deficiencies. Annu Rev Biochem 1995;64:375–401. [7] Wykoff CC, Beasley NJP, Watson PH, Turner KJ, Pastorek J, Sibtain A, et al. Hypoxia-inducible expression of tumor-associated carbonic anhydrases. Cancer Res 2000;60:e7075–83. [8] Beasley NJ, Wykoff CC, Watson PH, Leek R, Turley H, Gatter K, 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 2001;61:5262–7. [9] Watson PH, Chia SK, Wykoff CC, Han C, Leek RD, Sly WS, et al. Carbonic anhydrase XII is a marker of good prognosis in invasive breastcarcinoma. Br J Cancer 2003;88:1065–70. [10] Giatromanolaki A, Koukourakis MI, Sivridis E, Pastorek J, Wykoff CC, Gatter KC, et al. Expression of hypoxia-inducible carbonic anhydrase-9 relates to angiogenic pathways and independently to poor outcome in non-small cell lung cancer. Cancer Res 2001;61:7992–8.

[11] Stubbs M, McSheehy PM, Griffiths JR, Bashford CL. Causes and consequences of tumour acidity and implications for treatment. Mol Med Today 2000;6:15–9. [12] Thiry A, Dogne JM, Masereel B, Supuran CT. Targeting tumor-associated carbonic anhydrase IX in cancer therapy. Trends Pharmacol Sci 2006;27:566–73. [13] Potter CP, Harris AL. Diagnostic, prognostic and therapeutic implications of carbonic anhydrases in cancer. Br J Cancer 2003;89:2–7. [14] Pastorek J, Pastorekova S, Callebaut I, Mornon JP, Zelnik V, Opavsky R, et al. Cloning and characterization of MN, a human tumor-associated protein with a domain homologous to carbonic anhydrase and a putative helix–loop–helix DNA binding segment. Oncogene 1994;9:2877–88. [15] Gray Lh, Conger Ad, Ebert M, Hornsey S, Scott Oc. The concentration of oxygen dissolved in tissues at the time of irradiation as a factor in radiotherapy. Br J Radiol 1953;26:638–48. [16] Chia S, Ragaz J, Makretsov N, Cheang M, Hayes M, Gilks B, et al. Carbonic anhydrase IX (CA IX) as a predictive marker for benefit from post-mastectomy adjuvant locoregional radiation (RT): results from the British Columbia (BC) randomized radiation trial. J Clin Oncol 2005;23:16S. [17] Chia SK, Wykoff CC, Watson PH, Han C, Leek RD, Pastorek J, et al. Prognostic significance of a novel hypoxia-regulated marker, carbonic anhydrase IX, in invasive breast carcinoma. J Clin Oncol 2001;19:3660–8. [18] Loncaster JA, Harris AL, Davidson SE, Logue JP, Hunter RD, Wycoff CC, 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 2001;61:6394–9. [19] Kaanders JH, Wijffels KI, Marres HA, Ljungkvist AS, Pop LA, van den Hoogen FJ, et al. Pimonidazole binding and tumour vascularity predict for treatment outcome in head and neck cancer. Cancer Res 2002;62:7066–74. [20] Swinson DE, Jones JL, Richardson D, Wykoff C, Turley H, Pastorek J, et al. Carbonic anhydrase IX expression, a novel surrogate marker of tumour hypoxia, is associated with a poor prognosis in non-small-cell lung cancer. J Clin Oncol 2003;21:473–82. [21] Hussain SA, Palmer DH, Ganesan R, Hiller L, Gregory J, Pastorek J, et al. Role of hypoxic marker CA IX in transitional cell carcinoma of the bladder. Oncol Rep 2004;11:1005–10. [22] McKierman JM, Buttyan R, Bander NH, Stifelman MD, Katz AE, Chen MW, et al. Expression of the tumour-associated gene MN: a potential biomarker for human renal cell carcinoma. Cancer Res 1997;57:2362–5. [23] Vermylen P, Roufosse C, Burny A, Verhest A, Bosschaerts T, Pastorekova S, et al. Carbonic anhydrase IX antigen differentiates between preneoplastic malignant lesions in non-small cell lung carcinoma. Eur Respir J 1999;14:806–11. [24] Kim SJ, Rabbani ZN, Dewhirst MW, Vujaskovic Z, Vollmer RT, Schreiber EG, et al. Expression of HIF-1␣, CAIX, VEGF and MMP-9 in surgically resected non-small cell lung cancer. Lung Cancer 2005;49:325–35. [25] Kim SJ, Rabbani Z, Vollmer RT, Schreiber EG, Oosterwijk E, Dewhirst MK, et al. Carbonic anhydrase IX in early stage non small cell lung cancer. Clin Cancer Res 2004;10:7925–33. [26] Le QT, Chen E, Salim A, Cao H, Kong CS, Whyte R, et al. An evaluation of tumor oxygenation and gene expression in patients with early state non-small cell lung cancers. Clin Cancer Res 2006;12(5):1507–14. [27] Kon-no H, Ishii G, Nagai K, Yoshida J, Nishimura M, Nara M, et al. Carbonic anhydrase IX expression is associated with tumor progression and a poor prognosis of lung adenocarcinoma. Lung Cancer 2006;54:409–18. [28] Simi L, Venturini G, Malentacchi F, Gelmini S, Andreani M, Janni A, et al. Quantitative analysis of carbonic anhydrase IX mRNA in human non-small cell lung cancer. Lung Cancer 2006;52:59–66. [29] Barathova M, Takacova M, Holotnakova T, Gibadulinova A, Ohradanova A, Zatovicova M, et al. Alternative splicing variant of the hypoxia marker carbonic anhydrase IX expressed independently of hypoxia and tumour phenotype. Br J Cancer 2008;98:129–36. [30] Travis WD, Colby TV. International histological classification of tumor. 3rd ed. Geneva, Switzerland: Springer/World Health Organization; 1999. [31] Hockel M, Vaupel P. Tumor hypoxia: definitions and current clinical, biologic, and molecular aspects. J Natl Cancer Inst 2001;93:266–76. [32] Brizel DM, Sibley GS, Prosnitz LR, Scher RL, Dewhirst MW. Tumor hypoxia adversely affects the prognosis of carcinoma of the head and neck. Int J Radiat Oncol Biol Phys 1997;38:285–9. [33] Smith E, Stratford IJ, Adams GE. Cytotoxicity of adriamycin on aerobic and hypoxic Chinese hamster V79 cells in vitro. Br J Cancer 1980;42:568–73. [34] Harris AL. Hypoxia—a key regulatory factor in tumour growth. Nat Rev Cancer 2002;2:38–47. [35] Airley RE, Loncaster J, Raleigh JA, Harris AL, Davidson SE, Hunter RD, et al. GLUT-1 and CAIX as intrinsic markers of hypoxia in carcinoma of the cervix: relationship to pimonidazole binding. Int J Cancer 2003;104:85–91. [36] De Schutter H, Landuyt W, Verbeken E, Goethals L, Hermans R, Nuyts S. The prognostic value of the hypoxia markers CA IX and GLUT 1 and the cytokines VEGF and IL 6 in head and neck squamous cell carcinoma treated by radiotherapy +/− chemotherapy. BMC Cancer 2005;5:42. [37] Koukourakis MI, Bentzen SM, Giatromanolaki A, Wilson GD, Daley FM, Saunders MI, et al. Endogenous markers of two separate hypoxia response pathways (hypoxia inducible factor 2 alpha and carbonic anhydrase 9) are associated with radiotherapy failure in head and neck cancer patients recruited in the CHART randomized trial. J Clin Oncol 2006;24:727–35. [38] Swinson DE, Cox G, O’Byrne KJ. Coexpression of epidermal growth factor receptor with related factors is associated with a poor prognosis in non-small-cell lung cancer. Br J Cancer 2004;91:1301–7.