Carbonic anhydrase IX in oligodendroglial brain tumors

Oligodendrogliomas account for approximately 5–18% of all intracranial gliomas and they occur primarily in the frontal or temporal lobes of the cerebrum [1, 2]. It has been suggested that oligodendrogliomas were previously underdiagnosed, and there are several novel studies where their incidence is considerably higher, up to 33% of gliomas [3]. Accurate diagnosis is important in the case of oligodendrogliomas because the pathophysiology, treatment options and prognosis vary from that of diffusely infiltrating astrocytomas[1].

There are at least two different genetic pathways for the tumorigenesis of oligodendrogliomas. Loss of chromosomal regions on 1p and 19q is characteristic of oligodendrogliomas, but there is also evidence for another subset of oligodendrogliomas that have amplification of EGFR oncogene, loss of heterozygosity on chromosome 10 and homozygous deletion of the CDKN2A tumor suppressor gene. This seems to exclude the more common genetic changes in 1p and 19q [4, 5].

Antioxidant enzymes and related proteins (AOEs) are part of the cellular protection mechanisms against functional and structural damage caused by reactive oxygen species [6]. Increased oxidant stress and/or diminished levels of AOEs lead to multiple injurious consequences in living cells, including susceptibility to genetic alterations and carcinogenesis. AOEs such as manganese superoxide dismutase (MnSOD), glutathione associated enzymes (GLCL-C and GLCL-R) and thioredoxin-thioredoxin reductase (Trx, TrxR) have been shown to correlate with tumor grade, metastasis and poor prognosis in invasive carcinomas such as lung and gastrointestinal malignancies and we have earlier studied them in oligodendroglial brain tumors [7‐10].

Carbonic anhydrase IX (CA IX) is a tumor-associated metalloenzyme that belongs to the physiologically important family of at least 13 different mammalian carbonic anhydrases, CAs [11‐13]. It is localized in the plasma membrane and like other enzymatically active CA isoenzymes (CA I-IV, VA, VB, VI, VII, XII-XV), it contains four important histidine residues: three residues for the coordination of a zinc ion in the active site and one for a proton shuttle. CAs catalyze the reversible interconversion between CO₂ and HCO₃ ^-, one of the most fundamental chemical reactions in cells. This reaction is essential for organisms as it influences respiration, pH regulation and homeostasis, exchange of electrolytes and several metabolic biosynthetic pathways [12‐15]. Aberrant changes in this fine machinery are implicated in many diseases, including cancer.

CA IX has a special role among human CA isoenzymes because it can only be found in few normal tissues, but it is abundant in several tumors, such as colorectal, bladder, cervical, lung and breast carcinomas [16‐18]. Even though the expression of CA IX in these carcinomas is evident, the tissues from which the carcinomas are originally derived are known to be CA IX-negative or they show only low enzyme expression. Furthermore, the few normal tissues or cell types that express CA IX, such as gastrointestinal and gallbladder epithelial cells, have been reported to lose the expression of CA IX during carcinogenesis [19‐22]. This quite exceptional phenomenon makes CA IX an interesting tumor-associated protein.

CA IX expression is strongly induced by hypoxia. This transcriptional activation is accomplished via the HIF-1 transcription factor, which accumulates in tissue under the hypoxic condition that is often present in growing tumors. That, in turn, is an outcome of poorly organized and insufficient vasculature in uncontrollably growing malignant tissue. The HIF-1 transcription factor is a trigger for several hypoxia-regulated genes linked to cell survival, proliferation, apoptosis, angiogenesis and metabolism in tumor cells. The activation of these genes helps the cell to adapt to the stress caused by low oxygen level in a particular tissue. Earlier studies have shown that CA IX expression is induced as early as two hours after HIF activation and persists for several days, even if HIF-1 expression has ceased. This means that CA IX reflects both previous and present hypoxia in cells [23‐25]. In addition to the important correlation between hypoxia and CA IX expression, it has been found that CA IX expression correlates with poor prognosis in various tumors [26‐29]. It has also been suggested that CA IX has a direct role in tumor progression and the regulation of pH balance during tumorigenesis. CA IX has effects on cell adhesion, too, and it has been suggested to play a role in tumor invasion through weakening of cell-cell adhesion as E-cadherin is competing for β-catenin [30]. However, the effect of CA IX on tumor cell invasion is still under debate as some recent results have suggested that there is no evidence of such a correlation [31]. Even though we do not know the exact mechanisms of CA IX in cancer development, it is undoubtedly a promising target for anticancer treatments. It is also known that CA IX is strongly inhibited by aromatic and heterocyclic sulfonamides, and several other potent inhibitors for CAs have already been designed [32, 33]. These facts prompted us to investigate the expression of CA IX in a series of histopathologically and genetically verified oligodendroglial brain tumors.

We investigated the expression of CA IX in a series of oligodendroglial brain tumors. We found that CA IX is expressed in most of these tumors and that it has an independent prognostic value, predicting a poor outcome. In addition, CA IX(+) tumors had decreased cell proliferation compared to CA IX(-) tumors.

Chromosomal loss of 1p and 19q is a common genotypic change in oligodendrogliomas. It is commonly used as a tool to make more precise decisions concerning the treatment. Recently, it has become clear, however, that defects in single chromosomes, genes, or even the sequential acquisition of mutations, are not the entire story in tumorigenesis, given that many cancer susceptibility genes show a high degree of tissue specificity in their association with neoplastic transformation. Also in oligodendrogliomas and other malignant brain tumors, the acquisition of a complete neoplastic phenotype is caused by the activation of several oncogenes and the inactivation of several growth-suppressor genes, and therefore, a number of target molecules are required to be involved in an optimal treatment strategy. There are several characteristics which are important for such a protein target: The molecule (a) is expressed only in tumor cells, and in the majority of tumor cells; (b) is well accessible on the surface of the cell; (c) is not encoded from a mutated gene, but is derived from a stable chromosomal area; (d) can be targeted with a safe drug or inhibitory antibody; (e) is easy to determine in normal laboratory samples and (f) the costs of diagnostic and therapeutic uses are reasonable.

CA IX is an extremely attractive molecule in cancer research because it has several important properties that make it completely unique. First of all, it is mainly expressed in cancerous tissue. Second, it is expressed in various solid tumor types, including some very common ones (for example breast, lung and cervical carcinomas), which can be quite resistant to chemotherapy and radiotherapy. Third, the CA IX antigen is exposed to the surface of the malignant cell, thus making it accessible for antibodies or cancer drugs. Fourth, the structure of the CA IX molecule is amenable to targeting by drugs and it also has good selectivity. Fifth, CA IX expression can be easily verified from tissues with common laboratory protocols. Sixth, CA inhibitors are already being used in both clinical applications and new trials [33, 39] and more effective and safer inhibitors are being developed [40]. Therefore, one could conclude that CA IX indeed exhibits several features that are necessary for a good target molecule in anticancer therapy, and at least for now it is clear that in most cases good cancer treatment includes the best combination of different treatments supporting one another. CA IX and its inhibitors could definitely be considered as part of that kind of treatment strategy in the near future.

CA IX has proved to be a surrogate marker for hypoxia and it has also been found to correlate with prognosis in several cancers (including our present tumor material) [17, 26, 27, 29, 41]. CA IX expression could thus be useful in the prognostic assessment of certain tumors. For example, the previous study by Haapasalo et al showed that it has a strong association with tumor grade and poor prognosis in astrocytic brain tumors [28]. CA IX immunopositivity was observed in 78% of astrocytic brain tumors (compared to 80% of oligodendrogliomas in our study). In addition, both in astrocytic brain tumors and in oligodendrogliomas perinecrotic accentuation is quite commonly seen. A study by Preusser et al shows that CA IX has a positive correlation with the presence of necrosis [42]. When necrosis or microvascular proliferation was present, perinecrotic or microvessel staining was commonly seen in our present study, too. This could reflect the severe changes disturbing the tissue (hypoxia, insufficient microvasculature etc.) leading finally to necrosis but also enhancing the expression of CA IX. Although the expression of CA IX is common in both astrocytomas and oligodendrogliomas, it was more clearly associated to tumor malignancy grade in astrocytic brain tumors. However, the unique features of CA IX molecule could be useful when designing improvement of therapeutic interventions for both of these tumor groups.

A recent study by Birner et al in 60 oligodendroglial tumors with 1p aberrations showed that CA IX expression is often accompanied with the expression of hypoxia-inducible factor 1α (HIF-1α) [43]. They concluded that when expressed together, they represent a true tissue hypoxia and not just oncogenic activation, which might otherwise be the case, at least for HIF-1α. It is already known that HIF-1α is one of the key factors regulating cellular O₂ homeostasis and its activation represent a key step in angiogenesis and adaptation to hypoxia and thus the vitality of the tumor. Overexpression of HIF-1α is shown to be an independent prognostic factor in oligodendroglial tumors and interestingly, it also correlates with the microvessel density in these tumors [44]. They suggested that as tissue hypoxia is known to diminish the efficacy of radiotherapy and thus influence the adjuvant therapy given to the patients, the evaluation of tissue hypoxia using this combination of hypoxia-markers would be helpful for recruitment of patients for individualized therapy strategies, e.g. identification of hypoxic tumors for hyperbaric oxygenation preceding radiotherapy.

In the present study we found that CA IX was expressed more often in MnSOD-positive tumors. In our previous study on AOEs in partly the same subgroup of oligodendroglial brain tumors, MnSOD was associated with the histopathological component (pure oligodendrogliomas vs. mixed astrocytomas), but failed to show a significant correlation with patient survival [10]. The association reported here with CA IX is interesting because CA IX is known to be induced by hypoxia, and MnSOD in turn is induced among other factors by hyperoxia. It seems that CA IX expression and hypoxia are quite consistently associated, while factors that induce MnSOD are quite complex and numerous [7, 45]. Another possible explanation for this peculiar association might be that MnSOD and CA IX expression are associated with different phases of tumorigenesis in oligodendrogliomas. MnSOD has been suggested to be a tumor suppressor gene [46, 47]. The increased expression of MnSOD may reflect the imbalance during tumorigenesis where cells attempt to control the tumor environment. As the process advances, the environment becomes increasingly hostile for normal cells. Rapidly growing, poorly organized and improperly proliferated cancer cells lead to insufficient microvasculature and low pH, suppressing the growth of normal cells even further, while selectively stimulating the growth of malignant ones. This leads to the activation of hypoxia response elements and consequently induces a number of genes, including the CA9 gene, involved in the adaptation to the stress caused by low oxygen. The association between MnSOD and CA IX may thus represent partly overlapping periods of tumorigenesis during which both proteins are present in neoplastic tissue. Furthermore, the basic reaction of MnSOD where reactive oxygen species superoxide O₂- is converted to H₂O₂ by MnSOD is known to affect to HIF-1α, by letting it accumulate into the cytosol. Thus, MnSOD, HIF-1α and CA IX expressions are representing an exciting and complicated continuum.

We also found an association between CA IX and proliferation index (MIB-1). In our material, CA IX(+) tumors had significantly lower proliferation index than their CA IX(-) counterparts. This could reflect the severe hypoxia that is present in the tumor tissue, thus affecting to cells capability to proliferate. A study by Proescholdt et al shows that the relationship between CA IX and proliferation index is quite complicated [48]. They examined the expression of CA IX and CA XII in different kind of brain tumors and also in the normal brain. They found an association between CA IX positivity and increasing proliferation index. However, their material was quite heterogeneous, and oligodendrogliomas were not included at all. Despite of that, it can be said that the correlation between CA IX and proliferation is interesting and should be studied more thoroughly.

In conclusion, CA IX is expressed in the majority of oligodendroglial tumors and its expression correlates independently with patient survival. It is known to be induced by hypoxia in the tissue, partly by the same elements that induce angiogenesis, and it affects the survival of cancer cells. CA IX seems to play a major role in the pathophysiology of several malignancies, including our material of oligodendroglial tumors, and it may contribute to the abnormal cellular environment of malignant brain tumors by shifting the pH gradient towards higher acidity. Because of its many unique properties, CA IX can be considered a good candidate for a target molecule when designing new anticancer therapies.