The microenviroment within a solid tumor has an extensive influence on the outcome of cancer treatment and the prognosis of the disease. Tumor hypoxia affects the behaviour of tumor cells and is associated with poor prognosis and reduced overall survival . This fact reveals the need for a detailed study of biological effects under reduced oxygen levels. The most common technique used to investigate in vitro tumor hypoxia is the hypoxia chamber. However, this approach has limitations. The method requires special technical equipment while it has been shown that it leads to a slow onset of hypoxia that might influence the correlation between changes in oxygen concentration and kinetic of hypoxia dependent biological events.
An alternative to hypoxia chamber represents the enzymatic GOX/CAT system, which has been shown to rapidly induce in vitro hypoxia. The GOX/CAT system has been employed in the past in various studies. In particular, Baumann et al. have applied the enzymatic system for investigation of the effects of the hypoxia-targeted prodrug KS119 [9, 17]. Furthermore, Zitta et al. used GOX/CAT for rapidly induction of hypoxia and investigated the influence of mild hypothermia and postconditioning with catalase on hypoxia-mediated cell damage , as well as the potential cytoprotective properties of different sevoflurane conditioning strategies on a human neuronal cell culture model . In addition, Owegi et al. applied the GOX/CAT technique to test macrophage activity under various O2 and H2O2 concentrations, as presented under infection conditions . All these studies have demonstrated a rapid decrease of oxygen concentration using glucose oxidase and catalase but provided only limited comparisons to the established hypoxia chamber technique. Therefore, in the present study we evaluated the enzymatic GOX/CAT system in direct comparison to the established hypoxia chamber technique for investigation of different biological events, including gene expression, glucose uptake and radioresistance at a defined O2 concentration.
The conditions for in vitro generation of hypoxia at a level of 2% were carefully chosen in concert with the results of previous studies. In particular, evaluation of oxygen concentration using a computer-driven oxygen electrode revealed that at the conditions used for our experiments 2% hypoxia was rapidly induced within 15 min and maintained over 24 h . Since oxygen transport studies using hypoxia chambers have revealed time periods of more than 3 h for equilibration of pO2 between the medium inside the plate and the gas outside of it, which even accelerated in the presence of cells , evaluation of both systems was performed after 24 h cell cultivation under hypoxic conditions to ensure that the observed biological events are not a result of differences in the oxygenation level. We further chose for our investigation a head and neck squamous cell carcinoma (HNSCC) cell line because there is strong evidence that hypoxia is an important microenvironment factor, which influences the response of HNSCC to therapy  and because the role of low oxygen tension has been extensively investigated for this cancer entity both in preclinical and in clinical studies [22, 23].
Our experiments demonstrated comparable trends for both systems in regard to gene expression, glucose uptake and resistance towards radiation therapy. In particular, investigation of hypoxia related genes using microarray chip analysis in our study revealed a similar regulation trend for most known HIF-1 target genes for both the rapid enzymatic GOX/CAT system and the hypoxia chamber after 24 h of hypoxia (Figure 1). The expression of prominent hypoxia dependent genes, such as carbonic anhydrase IX (CA9) and lysyl oxidase (LOX) was additionally to microarray analysis quantified by real time PCR. These genes were chosen for analysis not only because it is known that they are hypoxia regulated, but also because various studies have reported prognostic values for them in head and neck squamous cell carcinoma [24, 25]. Microarray analysis in our study indicated CA9 and LOX activation both in the chamber and the enzymatic system after 24 h, while CA9 showed stronger activation than LOX. Quantification through real time PCR demonstrated different kinetic patterns between the two hypoxia systems (Figure 4). Particularly, although both genes were upregulated under hypoxic conditions the upregulation peak was reached earlier for the rapid enzymatic GOX/CAT system and decreased thereafter, compared to the hypoxia chamber that showed a continuous increase of gene expression over 24 h. Our results are in concert with the results of previous studies using the GOX/CAT system . Millonig et al. have shown that a fast onset of hypoxia using the enzymatic system leads to rapid induction of HIF-1 that later disappears although the cells remain under stable hypoxia. In contrast, cell exposure to the same oxygen concentration using a conventional hypoxia chamber causes a late onset and continuous upregulation of HIF-1 over a time period of 24 h. These results led the authors to the conclusion that HIF-1 responds rather to oxygen decrements than to absolute hypoxia, a hypothesis that might also explain the different kinetic patterns of the HIF-1-target genes CA9 and LOX as demonstrated in our study.
In regard to glucose metabolism, uptake experiments of fluorodeoxyglucose (FDG) revealed an enhanced cellular uptake for both the enzymatic and the chamber system (Figure 5), which increased with time progression (Table 1). This result is expected, since it is known that hypoxia is associated with a reprogrammed cellular metabolism, characterized by enhanced uptake of glucose for use as anabolic and catabolic substrate. The enhanced FDG uptake is supported by a HIF-1 dependent activation of the transcription of SLC2A1 and SLC2A3 genes, which encode the glucose transporters GLUT1 and GLUT3 respectively. Furthermore, HIF-1 activates the transcription of the HK1 and HK2 genes, which encode for hexokinase, an enzyme that phosphorylates FDG and represents the first enzyme of the Embden-Meyerhoff (glycolytic) pathway [26, 27]. The role of HIF-1 in further metabolisation of glucose has been extensively investigated in previous studies. In particular, it has been shown that glycolytic enzymes which metabolize glucose to pyruvate, and lactate dehydrogenase A (LDHA) which further converts pyruvate to lactate are regulated by HIF-1, promoting ATP production through increased anaerobic glycolysis under hypoxic conditions . The results of our study demonstrate that the new enzymatic GOX/CAT system affects glucose metabolism in a similar trend like the established hypoxia chamber. FDG uptake was increased for both systems, result that is in concert with the microarray analysis, which shows an upregulation of genes involved in glucose metabolism, such as SLC2A1, SLC2A3, HK1, HK2 and LDHA. The slower increase of FDG uptake for the hypoxia chamber, compared to GOX/CAT (Table 1) might be explained by different kinetics in the expression of HIF-1 target genes that are involved in glucose metabolism, considering the fact that further HIF-1 target genes, such as CA9 and LOX showed different expression kinetics for the two systems.
In regard to glucose metabolism, assignment of gene expression results to biological function gene ontology terms (GO-terms), demonstrated glycolysis to be the most probably regulated GO-term for both systems (Figure 3). This is expected since glycolysis is known to be the preferred route for energy production under conditions of oxygen deficiency. Although our results provide strong indications of glycolytic metabolism, further investigation of the ratio between lactate production and glucose consumption is needed in order to assess the balance between glycolytic and oxidative metabolism under normoxia and hypoxia using the GOX/CAT system. This is important, considering the fact that cancer cells are known to use glycolysis even under normoxic conditions. Since glycolysis can produce ATP at higher rates than oxidative phosphorylation  and tumor cells require fast energy production in order to support cell growth and survival, metabolic alterations in favour of glycolysis is noticed even under normoxia , demonstrating the complexity of pathways and mechanisms in respect to microenvironment adaptation of tumor cells.
The enzymatic GOX/CAT system has however a critical limitation that needs to be considered in experiments investigating glucose metabolism. Glucose oxidase (GOX) does not only consume oxygen but also leads to depletion of glucose in the incubation medium. Previous studies investigating in vitro FDG uptake in various cell lines have revealed that hypoglycemic conditions lead to an increased FDG uptake [31, 32]. Furthermore, it has been shown that the enhanced transport activity caused by hypoglycemia is attributed to an increased expression of GLUT1 in the cell membrane . Therefore, the GOX mediated glucose depletion might bias the results of metabolic experiments. The substrate consumption at various settings of the GOX/CAT system has been extensively evaluated . Under our conditions, hypoxia could be stably maintained for about 24 h without replacing the medium and reagents, leading to a glucose decrease of about 10% . For comparison, 5% equals the 24-hour glucose consumption of about 90 million exponentially growing tumor cells .
Subphysiologic levels of oxygen in the tumor lead to an up to 3-fold increase of resistance against antineoplastic strategies, such as radiation therapy . The enhanced radioresistance is explained through a reduced production of cytotoxic reactive species and promotion of the upregulation of genes that protect the cells from irradiation . Within our study we performed proliferation experiments after irradiation of the cells in order to investigate whether the enzymatic GOX/CAT system could be used for in vitro investigation of hypoxia related radioresistance. The comparison with the established hypoxia chamber revealed that at O2 concentration of 2% only a slight resistance increase was noticed for the hypoxia chamber system, while the GOX/CAT system showed a higher resistance to photon irradiation (Figure 6). The enhanced radioresistance for the rapid hypoxic strategy could be explained by an increased growth arrest in the G0/G1 phase of the cell cycle. It has been shown in the past that one of the genes that promote growth arrest in the G0/G1 phase via upregulation of p21 is heme oxygenase 1 (HMOX1) . Our gene expression analysis revealed a strongly increased expression of HMOX1 for the GOX/CAT system compared to the hypoxia chamber (Log2 of 3.5 and 0.2, respectively), result that offers a possible explanation for the enhanced cytoprotection that needs to be further investigated.
The hypothesis of growth arrest through rapid hypoxia is supported by the results of viability experiments irrespective of irradiation. These experiments showed lower cell numbers for the hypoxic systems compared to normoxia. Trypan blue and microscopy analysis revealed however that the reduced cell number was not attributed to cell death. Our results might be explained by a reduced cell division and DNA synthesis, which has been described in previous studies using the enzymatic model .
The GOX/CAT system has some limitations. Besides the fact that GOX causes glucose depletion and therefore the results might be affected by substrate deprivation, the activity of GOX also leads to the production of D-gluconolactone, which may cause culture medium acidification. pH measurement during our studies with HNO97 cells revealed no significant acidification of the DMEM medium for the investigated time period of 24 h. However, using the GOX/CAT system for hypoxia induction on human umbilical vein endothelial cells (HUVEC) a rapid pH decrease to a level of about 4.0-4.5 was noticed, leading to RNA degradation and cell death (data not shown). The extracellular pH of malignant tumors is known to be acidic, within a range of 6.5 to 7.0, as a consequence of increased glucose metabolism and poor perfusion , promoting tumor cell invasion via several matrix remodeling systems, including metalloproteinases, lysosomal proteases and hyaluronidase [40, 41]. However, the strong acidosis measured on HUVEC cells using the GOX/CAT system, can not only be attributed to physiologically induced acidocis. A possible explanation is a low buffer capacity of the HUVEC cell culture medium, which needs to be considered in the design of experiments using the GOX/CAT system. In order to minimize substrate depletion and gluconolactone production two strategies can be applied for incubation periods longer than 24 h. The first strategy is the replacement of the incubation medium by fresh, preequilibrated medium and the second is the use of larger volumes of medium, which will in turn increase the time to reach stable hypoxia .
Finally, it should be mentioned that the GOX/CAT system allows the additional generation and control of hydrogen peroxide independently of the degree of hypoxia [34, 42]. Since reactive oxygen species play an important role during tumor growth and radiation therapy of tumors, this option may be highly interesting when studying the role of transcription factors such as HIF-1 that are both responsive to hypoxia but also reactive oxygen species.