Quantitative assessment of specific carbonic anhydrase inhibitors effect on hypoxic cells using electrical impedance assays

Abstract Carbonic anhydrase IX (CA IX) is an important orchestrator of hypoxic tumour environment, associated with tumour progression, high incidence of metastasis and poor response to therapy. Due to its tumour specificity and involvement in associated pathological processes: tumourigenesis, angiogenesis, inhibiting CA IX enzymatic activity has become a valid therapeutic option. Dynamic cell-based biosensing platforms can complement cell-free and end-point analyses and supports the process of design and selection of potent and selective inhibitors. In this context, we assess the effectiveness of recently emerged CA IX inhibitors (sulphonamides and sulphocoumarins) and their antitumour potential using an electrical impedance spectroscopy biosensing platform. The analysis allows discriminating between the inhibitory capacities of the compounds and their inhibition mechanisms. Microscopy and biochemical assays complemented the analysis and validated impedance findings establishing a powerful biosensing tool for the evaluation of carbonic anhydrase inhibitors potency, effective for the screening and design of anticancer pharmacological agents.


Introduction
Carbonic anhydrase (CA, EC 4.2.1.1) isoform IX (CA IX) 1 is a zinc transmembrane metalloenzyme recognised as a key component of the hypoxic microenvironment characteristic to many types of solid tumours, where cells are deprived of oxygen due to rapid proliferation and change of metabolism from oxidative phosphorylation to glycolysis. CA IX is involved in processes like tumour proliferation and growth, cell adhesion and intercellular communication, migration, and, most importantly, pH regulation 2 . CA IX maintains the intracellular pH favourable for tumour cell survival and growth while contributing to the acidification of the tumour environment and promoting tumour metastasis 3 . To do so, CA IX regulates pH and ion balance by catalysing the interconversion of carbon dioxide (CO 2 ) to bicarbonate anion (HCO 3 À ) and protons (H þ ) 4 with its extracellularly oriented active site, whereas through its proteoglycan-like domain and extracellular carboxyl terminal tail influences cell-cell adhesion. CA IX is one of the few tumourassociated CA isoenzymes known so far; it is highly overexpressed in a broad range of solid tumours (carcinoma of the cervix, kidney, breast, lung, head and neck and tumours of the brain) 5 yet has only a very limited distribution in normal tissues (e.g. gastrointestinal tractepithelia of the glandular stomach, small intestine and gallbladder) 6 . As such, CA IX has become a validated target for anticancer therapy via the design of specific inhibitors, one such compound, SLC-0111, being in phase I/II clinical trials for the treatment of advanced, metastatic solid tumours (see clinical trials.gov, NCT02215850). Sulphonamides and sulphocoumarins are two wide classes of compounds able to inhibit the catalytic activity of CA IX by different mechanisms 7,8 , and have been proposed as potential antitumour therapeutic compounds. Often, the efficiency of these carbonic anhydrase inhibitors (CAIs) is determined based on a cell-free assessment of enzymatic activity and end-point biochemical cell analyses. In contrast, real-time, long-term analyses are lacking.
Aiming to further the understanding of the dynamics of interaction between CAIs and the target enzyme overexpressed in live cells under hypoxic conditions, we advance a multiparametric cell biosensing platform. In contrast to the end-point analyses, the use of label-free bioanalytical platforms allows dynamic evaluation of discrete, non-lethal effects of external stimuli at the cellular level and enable a quantitative, multiparametric assessment of potential bioeffects 9,10 . As such, they are invaluable tools for the development and preclinical testing of novel therapeutic compounds. To this end, electrical impedance spectroscopy (EIS), a non-invasive, label-free technique that allows real-time monitoring 11 of dynamic cellular responses induced by external biological and chemical compounds 12 , is advanced for testing the effect of several inhibitors (sulphonamides and sulphocoumarins compounds) on HT29 tumour cells. The HT29 colorectal carcinoma cells were selected as a model cell line due to their known overexpression of CA IX under hypoxic conditions and ability to adapt to hypoxic conditions through the interplay between metabolic changes and hypoxia-induced signalling 13 .
We appraise the inhibitory capacities of selected inhibitory compounds and reveal the dynamics of the cellular processes that underlie hypoxia and CA IX inhibition, indicating anticancer drug screening as a reliable application of EIS. As model inhibitory compounds we selected: a fluorescent, high affinity CA IX inhibitor, CAI #1, a potent, yet less CA IX specific inhibitor, CAI #2 (proved to have moderate inhibitory capacity for cytosol and membrane bound CA isoforms II and IV), a positively charged, membraneimpermeant derivative that has been reported as a good inhibitor of CA IX and CA II, CAI #3, and a sulphocoumarin demonstrated to be an effective CA IX inhibitor, CAI #4. The clinically used inhibitor acetazolamide (AZA) that indiscriminately inhibits all CAs was used as a control. Complementary biochemical assays confirmed the EIS results, offering additional information regarding cell viability, redox status and actin cytoskeleton organisation as modulated by CAIs under hypoxic conditions. Impedance studies with cells in hypoxic conditions are seldom reported 14,15 , with a focus on barrier function assays. To the best of our knowledge, our study demonstrates for the first time the dynamic evaluation of cellular effects induced by CAIs and indicates a powerful biosensing approach for the assessment of CAIs potency, effective for the screening and design of a wider range of anticancer pharmacological agents.
Cells were seeded at a concentration of 4.5 Â 10 5 cells/ml on IBIDI electric cell-substrate impedance sensing ( ECIS, IBIDI, GmbH, Germany) culture ware (8 W10E PET) and cultured for 2 d before experiments to reach confluence. Exposure to hypoxic (1% O 2 ) and normoxic (21% O 2 ) conditions together with specific treatment compounds was performed after the cells have reached confluence in standard growth conditions. Upon culture media exchange, the medium was aspirated off and replaced with either standard growth medium or medium containing CAIs (100 mm concentration 16 ). Cells were further incubated for 12 h at 37 C in the incubator/hypoxic chamber throughout the experiments.

Hypoxic conditions
Exposure of cells to a hypoxic environment was achieved using a modular incubator chamber (MIC-101, Billups-Rothenberg Inc., San Diego, CA). Cells were placed inside the chamber flushed with a gas mixture containing 94% N 2 , 1% O 2 and 5% CO 2 . Once gas equilibrated, the chamber was sealed and placed inside the incubator and cells were subjected to hypoxia for 12 h. Control cells were kept in normoxic conditions within the same humidified incubator, with 5% CO 2 . Both incubation units were equipped with electrode mounts for impedance evaluation under controlled conditions.

Electrical impedance spectroscopy measurements
The EIS enables multi-parameter, real-time monitoring of the interaction between cells and substrate and the study of cellular and subcellular processes in response to external stimuli 10,22 . A 4294 A Precision Impedance Analyzer (Agilent, now Keysight Technologies, Santa Clara, CA) interfaced with in house multiplexing module for up to eight channels was used for recordings. An AC signal of 100 mV amplitude, zero DC bias, within 100 Hz-100 kHz frequency range (100 frequency points with logarithmic distribution) was applied and spectra were recorded at selected time intervals (every 5 min). Data were collected and processed using a custom developed LabView interface. The whole spectra of the complex impedance Z Ã fr; t ð Þ ¼ Re Z fr; t ð Þ ½ þ i Im Z fr; t ð Þ ½ were analysed and complex fitted with a simplified equivalent circuit to derive time evolutions of specific circuit parameters as function of hypoxic conditions and CAIs' effect. In view of a simplified biosensing tool, single frequency impedance analysis has been applied as well. The imaginary part of impedance at 10 kHz frequency allows direct evaluation of cell attachment and growth and was selected throughout the analysis. Impedance values were normalised using the formula ½V fr; t ð ÞÀ Vðfr; 0Þ=Vðfr; 0Þ where V stands for the imaginary part of the complex impedance. Data analysis was realised using OriginPro 8.5 (OriginLab, Northampton, MA). All values were expressed as the mean ± standard deviation (SD). The statistical significance was assessed using OriginPro 8.5 (OriginLab), Student's t-test, and a p values <.05 was considered statistically significant.

Optical microscopy experiments
Epifluorescence has been used to evaluate the expression of CA IX in cells subjected to hypoxic conditions and treated with the fluorescent CAI #1. Moreover, complementary Differential Interference Contrast (DIC) and Bright Field Reflected Light (BFRL) assays have been used to assess cell morphology and cell-surface contacts. The microscopy set-up contained an AxioObzerver Z1 (Zeiss, Germany) microscope, a 40 Â 0.95 NA objective (Zeiss, Jena, Germany), an ANDOR EMCCD camera, and an environmental control enclosure (CO 2 and temperature, OKOLab, Pozzuoli, Italy). Cells were seeded at a concentration of 5 Â 10 4 cells/ml on Petri dishes with glass bottom (World Precision Instruments, Sarasota, FL) and used for experiments the next day.
Intracellular glutathione (GSH) detection and quantification Intracellular glutathione was stained using CellTracker Green 5-chloromethylfluorescein diacetate (CMFDA; Molecular Probes, Invitrogen). The culture medium was removed and the cells were incubated in FBS-free culture medium with 5 mm CMFDA at 37 C and 5% CO 2 for 30 min. After washing with pre-warmed media, cells were incubated for another 30 min in FBS-free culture medium to allow the hydrolysis of CMFDA to the fluorescent 5-chloromethylfluorescein (CMF) by intracellular esterases and conjugation with GSH or the diffusion of the unconjugated dye. Images were acquired using an inverted fluorescence microscope (Olympus IX71, Tokyo, Japan). The entire cell area was outlined based on the phase contrast images and was transposed on the CMFDA staining images in order to quantify the fluorescence intensity of the reduced GSH-5-CMF (GSH-CMF) complex. The fluorescence was normalised to the corresponding area in order to obtain uniform results regardless of cell size. The quantification of GSH was performed using ImageJ software (NIH, Bethesda, MD) for 200 cells per experimental group, selected from 20 different fields from four independent experiments.

Lysosome staining
Lysosomes were stained with 100 nm LysoTracker Green DND-26 (Molecular Probes, Invitrogen) for 30 min at 37 C, followed by the counterstaining of nuclei for 10 min at room temperature with 2 mg/ml Hoechst 33342 (Molecular Probes, Invitrogen). After washing the cells with PBS, images were acquired using an Olympus IX71 inverted fluorescence microscope.

MTT assay
Cell viability was measured using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; Sigma-Aldrich) assay based on the reduction of the tetrazolium salt to purple, insoluble formazan crystals by the NAD(P)H-dependent cellular oxidoreductase enzymes in the viable cells. The culture medium was removed and the cells were incubated with 1 ml of 1 mg/ml MTT solution for 2 h at 37 C. The formazan crystals formed in the viable cells were then dissolved with 1 ml of 2-propanol (Sigma-Aldrich). Absorbance was measured at 595 nm using a microplate reader (GENios Tecan, Austria).
Preparation of cell lysates HT29 cells were harvested from culture flasks at the end of hypoxia and washed with PBS. Cells were lysed by sonication (30 s for three times) on ice with an ultrasonic processor (Hielscher UP50H, Germany) and centrifuged at 10,000 Â g for 10 min at 4 C. Supernatants were collected for lipid peroxidation assay and Western blot. The protein concentration was determined using the Bradford Reagent (Sigma-Aldrich) and bovine serum albumin as the standard protein.
Lipid peroxidation assay Malondialdehyde (MDA) level was assessed as a marker of lipid peroxidation using a fluorimetric method 23 . A volume of 200 ml cell lysate was mixed with 700 ml of 0.1 N HCl. After 20 min at room temperature, 900 ml of 0.025 M thiobarbituric acid were added and the mixture was incubated for 65 min at 37 C. The MDA concentrations, nmol/mg protein, were obtained using fluorescence recordings (JASCO Spectrofluorometer FP-6300, k ex ¼ 520 nm; k em ¼ 549 nm) and 1,1,3,3-tetramethoxypropane as standard.

Western blotting
Cell lysates corresponding to 40 mg of protein were separated on 10% sodium dodecyl sulphate-polyacrylamide gel electrophoresis under reducing conditions and transferred onto 0.4 mm poly(vinylidene difluoride) membrane (Millipore, Billerica, MA) in a wet transfer system (Bio-Rad, Hercules, CA). The membranes were blocked with blocking solution included in the WesternBreeze Chromogenic kit (Invitrogen) for 30 min at room temperature. Primary mouse polyclonal antibodies anti-human p53 and antib-actin (dilution 1:250; Santa Cruz Biotechnology, Santa Cruz, CA) were used for protein detection. The membranes were processed in accordance with the manufacturer's instructions, using antimouse secondary antibodies coupled with alkaline phosphatase and 5-bromo-4-chloro-30-indolylphosphate/nitroblue tetrazolium as the chromogenic substrate. The resulting bands were imaged with a ChemiDoc MP system (Bio-Rad) and quantified using Image Lab software (Bio-Rad).

Statistical analysis
All results were expressed as the mean ± SD of five independent experiments and were represented as a percentage relative to control. The statistical significance was assessed using GraphPad Prism software (La Jolla, CA), one-way ANOVA test followed by a post-hoc Bonferroni test. A p values <.05 was considered statistically significant.

Cellular dynamics under hypoxic conditions revealed by EIS measurements
EIS measurements of cells in the frequency range of 100 Hz-100 kHz were performed to monitor cell growth and to evaluate in real-time and non-invasively the cell monolayer response to hypoxia and CAIs effects. Depending on the number, morphology, degree of attachment and electrical properties, interfacial effects and membrane processes of the cells attached to the electrodes shape the frequency dependence of system's impedance. An equivalent circuit, presented in inset Figure 1(a), comprising a constant phase element (CPE) accounting for the cell monolayer-electrode interface, a resistor (Rp) that defines the paracellular route in parallel with a capacitor (Cc) which describes the cellular monolayer and a resistor for the bulk solution (Rsol) is suitable for fitting the complex impedance spectra in the 100 Hz-100 kHz domain. Analysis of the impedance spectra in respect to the equivalent circuit is presented in Supporting information. Rp and CPE values were demonstrated to be the best reporters of cellular state and electrode coverage and hence suitable for cell assessment. Alternatively, impedance values at single frequencies can be used for monitoring system's response. The 10 kHz frequency is suitable 24 for monitoring the adhesion of cells onto electrodes, provides information regarding the membrane impedance and was found to be optimal for the particular electrode configuration (i.e. 10 circular 250 lm diameter active electrodes connected in parallel on a common gold pad) of the culture ware used in the experiments. At this frequency, the time evolution of the impedance (in particular, the imaginary part of impedance) is able to show reliably changes of cell monolayer integrity and of cell-support interface for both cell growth and in response to hypoxic conditions (Figure 1(a,b)).
Upon cellular confluence, achieved when impedance values reached a plateau (Figure 1(a), arrow), the cells were exposed to hypoxic conditions. In response to a first (12 h) phase of hypoxia, the impedance values have a distinct evolution from the one corresponding to normal growth conditions. Cellular effects due to pH changes determined by the onset of hypoxic conditions shape the dynamics of the values of impedance during the first hours of hypoxia. Tumour cells (HT29 including) were demonstrated 25 to overexpress in hypoxic conditions CA IX and CA XII, transmembrane isoforms of CA generally considered key pH regulators 26 . The impedance data reveal a transient drop in normalised impedance values consistent with cellular uncoupling at the surface of the electrodes and progressive expression of CA IX.
As expected, the overexpression of CA IX induced by hypoxic conditions is able to promote tumour cell survival as well as tumour-specific pH homeostasis. This fact is reflected in the subsequent increase in impedance values (Figure 1(b) Hypoxia I up to first 6 h) and is confirmed based on fluorescent CA IX inhibitor binding (inset). Increased cell mobility and changes in cellular status (including cell death) are evident (decrease of the values in the following 6 h) throughout the hypoxic exposure, yet cells are still able to maintain their cell-surface contacts that enable adequate impedance assessment. Notably, HT29 cells ability to adapt to hypoxic conditions through the interplay between metabolic changes and hypoxia-induced signalling 13 is revealed by impedance assays during these first 12 h of hypoxia.
Furthermore, during the second phase of hypoxia (another 12 h) the impedance values decrease abruptly suggesting a looser contact with the surface and a disturbed cellular state upon repeated/prolonged hypoxic stress.
Biochemical assays performed after the first and the second phase of hypoxia confirmed impedance findings offering additional information regarding viability, oxidative stress and p53 levels. The results, presented in Supporting information ( Figure S1) revealed a stronger effect of the second stage of hypoxia in comparison to cells kept in hypoxic conditions for only 12 h. Cellular responses to hypoxia include metabolic changes induced by the lack of oxygen and an increased generation of reactive oxygen species (ROS) which attack preferentially the membrane lipids. The prolonged hypoxia induces oxidative injury to unsaturated fatty acids through ROS, resulting in compromised lipid membrane matrix dynamics, altered cell morphology with important effects on the actin cytoskeleton and actin-dependent cellular processes, and decreased cell viability. The tumour suppressor p53 is sensitive to hypoxia-mediated oxidative stress 27 and its activation starts the process leading to apoptosis if the DNA damage cannot be repaired.
As such, a decrease in cell viability down to 85% ±5 for the first phase of hypoxia and 62% ±2 for the second phase in comparison to control is accompanied by specific increases in MDA and p53 levels (Supporting information Figure S1). The increase in MDA levels indicates 28 an enhanced lipid peroxidation and indirectly, the generation of ROS. The activation of p53-induced apoptosis 29 is revealed by a two-fold increase in expression of this protein for the second stage of hypoxia.
Moreover, the effect of hypoxia on cell morphology and adhesion has been confirmed based on DIC and BFRL (Supporting information Figure S2A) and is highlighted by an altered organisation of the actin cytoskeleton (Supporting information Figure S2B).
Since functional CAIs would determine an augmentation of the hypoxic conditions effect at the cellular level, the evaluation of the two hypoxia steps was a prerequisite for establishing the suitable "sensitivity window" for our assay. The extent of the cellular effects of the second phase of hypoxia is significant, hindering detailed electro-optic assessment; therefore, to increase the sensitivity of the assay, the effect of CAIs was further interrogated only for cells exposed to the first phase (the first 12 h) of hypoxia.

Characterisation of the inhibitory capacity revealed by EIS measurements
To test the capability of EIS assays to reveal the inhibitory potency, five model inhibitory compounds among sulphonamides and sulfocoumarins 21,30,31 (Figure 2(a)) were chosen.
Sulphonamides are a group of inhibitors that target efficiently CA IX 32 and their mechanism of inhibition is well-understood 26 . Among the sulphonamides model inhibitory compounds used in the analysis, CAI #1 is fluorescent and has a high affinity for CA IX 18 , CAI #2 has moderate inhibitory capacity in the case of CA II and IV 30,31 (cytosol and membrane-bound respectively), CAI #3 is a positively charged, membrane-impermeant derivative that has been reported as an inhibitor of CA IX and, intriguingly, of CA II 33 . CAI #4 is a 6-substituted sulphocoumarin reported as an effective CA IX inhibitor 8 . The clinically used inhibitor AZA which indiscriminately inhibits all CAs (16 mammalian CA isoforms identified) 34 and has been extensively used in CA inhibition studies, was considered herein a control.
Although HT29 cells show intermediate CA IX expression under normoxic conditions, literature reports 35 document that only during hypoxia CA IX becomes catalytically active and able to bind sulphonamides 35 . In this context, CA IX expression in cells subjected to normoxic and hypoxic conditions was evaluated by epifluorescence assay based on the incubation of cells with the fluorescent CAI#1. A fluorescent signal was detected only in hypoxic cells, being absent in the normoxic cells (Figure 1(b) inset and Supporting information Figure S2A). The lack of the fluorescence signal in the cells kept in normoxic conditions indicates CAI #1 selectivity for CA IX expressed during hypoxia.
To monitor the effect of CAIs in conjunction with the hypoxic conditioning of cells (i.e. overexpression of CA IX) we have compared, using EIS, the behaviour of the cells incubated with CAIs in normoxic versus hypoxic conditions during 12 h of hypoxia (Figure 2(b)). The cells treated with CAIs and kept in normoxic conditions did not suffer significant changes as compared to control normoxic conditions but had highly different dynamics from the ones in hypoxic conditions. CAIs application revealed a characteristic time evolution of the impedance value (the imaginary part at 10 kHz) in comparison to normoxia and hypoxia alone. The pronounced decrease (Figure 2(b)) of impedance values in hypoxic conditions as a function of applied CAI, confirms an augmentation of the deleterious effects at the cellular level of the hypoxic conditions, due to impaired tumour pH homeostasis upon CA inhibition. Each inhibitor presented a characteristic behaviour indicating a different inhibitory capacity. As such, the cellular impact of CA inhibition is dependent on the type of CAI and can be used to assess the inhibition potency: for CAI #1 the drop in impedance had fast kinetics and the decreased value was the biggest, corresponding to extensive electrode uncovering. For CAI #4 and CAI #2, although the kinetics is fast at the beginning, the inhibitory capacity is lower in comparison to that of CAI #1 as demonstrated by an incomplete electrode uncoverage. Both CAI #3 and AZA seem to have the same, low inhibitory capacity.
Complex fitting of the impedance spectra (100 Hz-100 kHz) with the equivalent circuit presented in inset Figure 1(a) generates specific dynamics of the characteristic equivalent circuit components during 12 h of hypoxia as a function of model CAIs exposure. Consistent with effects on cell monolayer integrity as a function of CAIs inhibitory potency, the derived Rp values (Figure 2(c)), corresponding to the tightness of the cell monolayer, confirm the characteristic decrease for efficient CA inhibition also highlighted by the dynamics of the Im[Z] at 10 kHz. The characteristic dynamics of the other circuit parameters are provided in Supporting information (Figures S3-S6) and confirm the effects of CAIs on cell attachment (CPE) and cell membrane (Cc).
For biosensing purposes, the evaluation of impedance changes after 6 h of combined exposure (hypoxia and inhibitory compound) enables clear discrimination of CAIs potency as revealed in Figure 2(d). The high statistical significance is obtained versus both hypoxia and between specific CAIs (inset Figure 2(d)).
In conjunction with Supporting information (Figures S3-S6) data, these results indicate monolayer disruption, cell detachment from the substrate and subsequent cell death, as CA IX is being inhibited and the cells are no longer able to adapt to a hypoxic environment and, as such are important quantitative predictors of inhibitory potency.
Endpoint biochemical assays complement impedance measurements to highlight the effect of CAIs on cell viability, redox status and actin cytoskeleton organisation. Cell viability decreases for cells incubated with CAIs during hypoxia in comparison to cells kept only in hypoxic conditions (Figure 3(a)). This decrease is dependent on the type of CAI used and is well correlated to inhibitor potency. The lowest viability was recorded for cells undergoing combined hypoxia and CAI #1 treatment confirming CAI #1 as having the highest inhibitory potential.
A decreased capacity of cells to maintain pH homeostasis under combined hypoxia and CAIs treatment is further confirmed based on the evaluation of intracellular GSH content (Figure 3(b)). Intracellular GSH concentration provides information regarding the induction and progression of oxidative stress and has been associated 36 with the mitochondrial generation of ROS. In the process of oxidative stress alleviation, GSH acts as a barrier against excessive oxidation of redox-sensitive molecules 37 and a lower intracellular GSH content suggests an altered redox status. The reduction in intracellular GSH level of HT29 cells incubated with CAIs during hypoxic conditions compared to that of cells subjected only to hypoxia (Figure 3(b)) follows the trend of viability data, indicating the induction of oxidative stress and depletion of GSH in all cellular compartments. The dynamics of normalised impedance data during 12 h of hypoxia as a function of model CAIs exposure (100 lm). Data are represented as mean ± SD for n ¼ 6. (c) The dynamics of a characteristic equivalent circuit component derived from the complex fitting of impedance data during 12 h of hypoxia as a function of model CAIs exposure (100 lm). (d) Characteristic impedance changes at a selected time point (after 6 h of treatment) revealing the CAIX inhibitors potency on HT29 cells under normoxic and hypoxic conditions. Data are represented as mean ± SD for n ¼ 6. Statistical significance Ã p < .05, ÃÃ p < .01 and ÃÃÃ p < .001 versus hypoxic conditions based on paired t-test. Inset: Statistical significance of impedance data for discrimination between CAI's potency Ã p < .05, ÃÃ p < .01 and ÃÃÃ p < .001.
Lysosomes are important cellular organelles involved in the degradation of macromolecules and in the induction of apoptosis and autophagy. Previous studies have shown that hypoxia induces lysosomal autophagy and mitochondrial dysfunction in a manner dependent on expression of hypoxia inducible factor 38,39 . Therefore, to complete the picture of CAIs effect on hypoxic cells, we have evaluated their accumulation. In accordance with the decreased GSH level (Figure 3(b)), an accumulation of lysosomes was noticed after the incubation of HT29 cells with CAIs ( Figure  3(c,d)) suggesting an increasingly oxidised environment. The accumulation of lysosomes in the case of cells incubated with CAI #1 was higher than for those incubated with the other CAIs indicating that more persistent damaged molecules have undertaken the degradation pathway.
The organisation of actin cytoskeleton filaments in HT29 cells kept in hypoxic conditions was also altered after CA IX inhibition.
The presence of cortical ring-shaped actin observed in the cells grown in normoxia and simple hypoxia was significantly reduced (Figure 3(e)solid arrows) as a result of the enhanced cellular stress induced by hypoxic conditions and CA IX inhibition. This feature was observed along with a diffuse pattern of actin staining (organisation of actin in the form of filaments was disrupted), the punctuate staining displayed being a characteristic of toxic effects on the cytoskeleton (Figure 3(e)arrow heads).

Discussion
Hypoxic environments are characteristic for a fast growing and/or solid tumours, being associated with aggressive tumour proliferation, resistance to therapy and a poor prognosis. CA IX is considered an endogenous marker of hypoxia, coordinating the carefully orchestrated tumour cell response to hypoxia that provides cancerous cells both a metabolic advantage as well as the capability to metastasise. This role in tumour progression and high incidence of metastasis is due to its capacity to maintain tumour intracellular pH homeostasis while acidifying the extracellular environment. Another important aspect of CA IX functionality, the potential to decrease cell-cell adhesion by destabilisation of E-cadherin links to the cytoskeleton, may contribute to the acquisition of increased tumour aggressiveness 6 . Therefore, CA IX is regarded as a clinically relevant, validated target for anticancer treatment as well as hypoxic tumour imaging.
Sulphonamides inhibit the catalytic activity of CA by binding to the zinc ion from the CA active site, whereas sulphocoumarins possess a diverse inhibition mechanism: they are hydrolysed within the enzyme active site to sulphonic acids which thereafter anchor to the water molecule bound to the zinc ion 34 . When cells grow in hypoxic conditions 34 , sulphonamide inhibitors are able to Data are expressed as mean ± SD for n ¼ 6; ÃÃÃ p < .001, ÃÃ p < .01 and Ã p < .05 versus hypoxic conditions. reduce the medium acidity by inhibiting the catalytic activity of CA IX enzyme, and thus the generation of hydrogen ions. Generically named CAIs, these compounds have been proposed as novel therapeutic/diagnostic agents 34,40 . Sulphocoumarins are among the latest CAI class introduced and have been shown to be effective CA inhibitors as well 8,[41][42][43] .
The efficiency of CAIs is often predicted based on enzymatic activity and end-point biochemical cell analyses, yet the real-time, continuous assessment of cellular responses brings new opportunities to assist in the design of powerful, isoform selective inhibitors.
To this end, by electrically monitoring the dynamic changes of cellular responses induced by hypoxia and CA(IX)Is, we hereby show the capability of the EIS cell substrate sensing to assess the potency of anticancer drugs in hypoxic conditions. This study expands EIS proficiency as a label-free, real-time and non-invasive method to appraise a variety of cellular processes e.g. cell differentiation, cell spreading, cell motility, toxicology, cytotoxicity, cell cycle progression 12,44 providing a quantitative analytical tool for evaluation of anticancer drug efficacy.
Single frequency impedance representation provides a straightforward biosensing tool. With appropriate frequency tuning (supported by analysis of impedance spectra using an equivalent circuit), the impedance data revealed changes at cell-cell and cellsurface level induced by the application of CAIs and allowed the differentiation between the inhibitory capacities of the compounds investigated. The interfacial capacitance, revealed by the imaginary part of impedance at a single frequency (10 kHz) provides a straightforward indicator of the overall cell-substrate adhesion. The results showed cell monolayer effects (including cellular detachment from the surface) as a function of effective and CA IX- selective inhibition. As such, cell adhesion and cytoskeletal activity changes, which are valuable indicators for understanding the cancer cell behaviour (such as proliferation, migration and invasion processes) 24,45,46 were highlighted in the context of cell response profiles and cellular dynamics alterations as a result of the interaction of CAI compounds with cells. The different inhibitory capacities of the compounds could also be evaluated: CAIs #1 stands out as having a more potent effect than CAIS #4, #2, #3 and AZA, the least potent one. Complementary end-point cell viability assay indicated that indeed CAIs could successfully suppress the adaptive mechanism of hypoxic HT29 cells towards maintaining cell proliferation, promoting cell detachment and cellular death instead. As confirmed through complementary biochemical data, the exposure of cells with functionally altered CA IX (i.e. nonhypoxic HT29 cells or cells grown in the presence of potent inhibitors) to hypoxic conditions determines, in the context of altered pH homeostasis, the induction of cellular effects characteristic to oxidative stress: low GSH levels, accumulation of lysosomes and increase of their function, which may subsequently lead to the initiation of autophagy and cellular death.
Evolutions of the equivalent circuit parameters presented in Supporting information (Figures S3-S6) complement the one of trans-epithelial resistance Rp in Figure 2(c) highlighting subtle differences between the tested compounds in the mechanism of CA inhibition and possibly in the effects on the intracellular structures. The evolutions of CPE _T values, indicating cell-electrode interface, are clustered for hypoxia and low-efficiency inhibitors (AZA and CAI #3) and CAI #2 and #4 respectively, yet the biphasic evolution for CAI #4 is notable: there is an initial increase of the CPE _T values, that could correspond to a transient tightening of cell attachment for sulphocoumarin binding. These differences may, in fact, reflect the different inhibition mechanisms of the sulphonamides (zinc-binding inhibitors) on one hand and of the sulphocoumarins (prodrug inhibitors, hydrolysed first and then anchoring to the zinc-coordinated water molecule) on the other 34 .
Rearrangements of actin filaments and stress fibres formation are dynamic adaptive changes induced by hypoxia, in particular through the activity of Rho-GTPases 47 . We noticed the formation of stress fibres and the disruption of actin filaments when cells were treated with CAIs which could further have important consequences for cell morphology, adhesion and migration, changes highlighted by impedance data. Particular alterations of the cell membrane revealed by dynamics of the Cc element of the equivalent circuit could be used as additional indicators of CAIs antitumoural efficiency. Therefore, these results pinpoint once more the capability of the electrical impedance technique to sensitively assess cellular dynamics in conjunction with in vitro assays. In association with end-point biochemical data, the evaluation based on EIS and equivalent circuit analysis sheds light on the complex cellular effects determined by specific inhibitors and hypoxic conditions and stands out as an important biosensing tool.

Conclusions
This study bridges the gap between classic end-point analyses and dynamic cell assessment and highlights process parameters relevant for assessing CA inhibition effectiveness. Potent CAIs are able to augment cellular changes characteristic for hypoxia progression: oxidative injury and compromised lipid membrane matrix dynamics, ROS accumulation, altered cell morphology with important effects on the actin cytoskeleton and actin-dependent cellular processes and decreased cell viability. The study demonstrates exquisite sensitivity of electric cell substrate impedance for the evaluation of CAIs potency towards a biosensing platform for novel and promising anticancer drugs development with clinical applications. Both the continuous cell dynamics and early time point (6 h) assessment were demonstrated to enable quantitative assessment of specific CAIs effect on live hypoxic cells.