TMCnet News
Physiological intestinal oxygen modulates the Caco-2 cell model and increases sensitivity to the phytocannabinoid cannabidiol [In Vitro Cellular & Developmental Biology](In Vitro Cellular & Developmental Biology Via Acquire Media NewsEdge) Abstract The Caco-2 cell model is widely used as a model of colon cancer and small intestinal epithelium but, like most cell models, is cultured in atmospheric oxygen conditions (?21%). This does not reflect the physiological oxygen range found in the colon. In this study, we investigated the effect of adapting the Caco-2 cell line to routine culturing in a physiological oxygen (5%) environment. Under these conditions, cells maintain a number of key characteristics of the Caco-2 model, such as increased formation of tight junctions and alkaline phosphatase expression over the differentiation period and maintenance of barrier function. However, these cells exhibit differential oxidative metabolism, proliferate less and become larger during differentiation. In addition, these cells were more sensitive to cannabidiol-induced antiproliferative actions through changes in cellular energetics: from a drop of oxygen consumption rate and loss of mitochondrial membrane integrity in cells treated under atmospheric conditions to an increase in reactive oxygen species in intact mitochondria in cells treated under low-oxygen conditions. Inclusion of an additional physiological parameter, sodium butyrate, into the medium revealed a cannabidiol-induced proliferative response at low doses. These effects could impact on its development as an anticancer therapeutic, but overall, the data supports the principle that culturing cells in microenvironments that more closely mimic the in vivo conditions is important for drug screening and mechanism of action studies. Keywords Colon . Cannabinoid . Normoxia . Oxidative metabolism . Butyrate Introduction Primary tissue samples are an invaluable tool in research and its procurement is important as it can provide crucial ex vivo data for a range of tissue specific pharmacokinetics and pharmacodynamics of potential drug compounds. However, fresh tissue can be difficult to obtain and often the amount of tissue available is limited. In particular, human gut primary tissue is notoriously difficult to culture; once it is removed from the body's microenvironment and no longer has the required external cues, its viability is limited (Rocha and Whang 2004). Due to these difficulties, traditional in vitro cell models have been fundamental in providing knowledge of pathway interaction and potential drug targets. However, the simplicity of basic cell models is a disadvantage as they are generally artificial, and do not represent the majority of crucial microenvironment parameters. Considerable progress has been made with 3D organoids (reviewed in (Roeselers et al. 2013) and organ-on-a-chip technology is a tantalizing prospect (Chiu et al. 1975). A simple environmental parameter often overlooked in routine cell maintenance and cell model development is the availability of oxygen. For example, intestinal cell lines, like many other cell lines, are cultured in atmospheric oxygen (O2) concentrations (~21%). This parameter is rather artificial as intestinal cells usually exist at an oxygen gradient of approximately 3-6% in vivo (Espey 2013). The intestinal lumen is thought to be anaerobic with a large number of resident commensal organisms and intestinal cells are provided oxygen from the submucosal tissue surrounding the crypt structures (Bauer et al. 2006). The relevance of this can be seen throughout the literature as it is often that low oxygen in culture is cited as 'hypoxic'. It should rather be considered, 'physiological O2' or 'normoxia' as it is what the cells from that site in the body are accustomed to. Primary stem cells cultured in oxygen levels that are close to the physiological level have been shown to have different differentiation, proliferation and gene expression profiles to primary cells which are grown in atmospheric oxygen levels (Csete et al. 2005). The Caco-2 adenocarcinoma model is a widely used in vitro intestinal cell model and can be utilized as a model for colon cancer in its proliferative state. These cells have the ability to spontaneously differentiate in culture over a period of 14-21 d, forming a polarized monolayer culture which mimics the intestinal epithelial cells of the small intestine forming crypt like structures and expressing brush border enzymes (Pinto et al. 1983;Haraetal.1993). This model, although useful, does not take into account a number of physiological intestinal parameters leaving it relatively artificial. One parameter repeatedly overlooked is the availability of oxygen (O2). Using Caco-2 cells, we investigated morphologic, metabolic and proliferative parameters under both atmospheric and physiological oxygen. As a proof of principle, a drug with known effect on these cells was used to test the validity of the model. Cannabidiol (CBD) is a nonpsychoactive constituent of the Cannabis sativa plant. The action of CBD is being examined as potential anticancer agent for a range of cancers including colorectal cancer (CRC). CBD has been shown to reduce cellular proliferation in Caco-2 cells, with a higher potency than a range of other tested cannabinoids including ?9-THC and cannabigerol. In CRC mouse models, CBD treatment significantly reduces polyps, aberrant crypt foci and tumours. CBD has also been shown to reduce tumour invasiveness and proliferation, although this has been shown in the breast cancer cell line MDA-MB231 and not CRC cells (Ligresti et al. 2006; McAllister et al. 2011; Aviello et al. 2012). We show that in our physiological Caco-2 model CBD actions are comparable to published data, but that there is a left-shift in dose responses and the mechanism of action is dependent on oxidative metabolism in the mitochondria. In addition, inclusion of the commensal-derived short chain fatty acid, butyrate, in the culture medium revealed a proliferative capacity of low dose CBD. We believe that physiological cell models with highthroughput potential for testing candidate drug compounds should include maintenance of parameters within the microenvironment of the cells' origin because drug screening and testing using more relevant in vivo conditions, may not only improve the understanding of how compounds work in vivo, but also reveal potential failures earlier in the drug discovery and development pipeline. Materials and Methods Unless otherwise stated, materials were purchased from Sigma-Aldrich Company Ltd, Dorset, UK. Cell culture. The human epithelial colorectal adenocarcinoma cell line, Caco-2 was obtained from the European Collection of Cell Cultures (ECACC) in 2009. Cells were reconstituted and stock early passages were stored in liquid nitrogen. Cells were routinely subcultured from these stock aliquots for up to a maximum of 40 passages and were tested for mycoplasma contamination twice during this time period, using the MycoFluor(TM) Mycoplasma Detection Kit (Life Technologies, Paisley, UK). Once cells reached 70-80%, confluence flasks were washed twice with 1x DPBS without CaCl2 or MgCl2 and incubated with 0.25% Trypsin with EDTA at 37°C, to detach cells from the flask. Cells were maintained in Minimal Essential Medium (MEM) supplemented with 8% Foetal Bovine Serum (FBS) and Non-Essential Amino Acids (NEAA) (all medium components from Life Technologies, Paisley, UK). Cells were incubated in either atmospheric conditions (21% O2 and 5% CO2), denoted "AtmosO2", or routinely maintained in a more physiological environment (5% O2 and 5% CO2), denoted "PhysO2" using the H35 Hypoxystation from Don Whitley Scientific, Shipley, UK. Cells moved from the AtmosO2 conditions to PhysO2 workstation (or vice versa) for the duration of the experiment (no longer than 21 d) are denoted "Atmos-PhysO2" or "Phys-AtmosO2" to indicate direction of transfer (see Table 1). For cell counts, 1x105 cells were plated in each well of 6 well plates at and periodically collected and counted using trypan blue exclusion. Cell division analysis. Proliferating cells were starved for 48 h, then detached and stained using carboxyfluorescein diacetate succinimidyl ester, CFSE (Life Technologies, Paisley, UK), as per the manufacturer's instructions. Day 3 samples were also accompanied by a day 3 sample treated with mitomycin (1 µg/ml). After 3, 7, 10 and 14 d, cells were detached and fixed in 1% paraformaldehyde (PFA). Cells were stored at 4°C before FITC intensities were recorded using Flow Cytometry on the FACS Canto II (BD Biosciences, Oxford, UK) (FACS overlays were achieved using CyFlogics software from CyFlo Ltd, Turku, Finland). Proliferation assays using cannabidiol. Cells were starved for 48 h in MEM without FBS. Cells were then detached and seeded 1 x 104cells per well of a 96 well plate in titrations of (-)-cannabidiol (Tocris Bioscience, R&D Systems Europe Ltd, Abingdon, UK) 10 µM-5 nM. Proliferation was assessed using either XTT (Roche Diagnostics Ltd, West Sussex, UK) or CyQuant (Life Technologies, Paisley, UK) analysis over 72 h, as per manufacturers' instructions. Data is presented as percentage of no treatment controls. Alkaline phosphatase activity and WGA size analysis using confocal microscopy. Cells were plated onto glass coverslips and maintained in MEM with FBS and NEAA. At days 3, 7, 14 and 21 cells were fixed in 1% PFA at RT for 45 min. Cells were stored in PBS. Cells were stained using Vector® Red Alkaline Phosphatase Substrate Kit (as per the manufacturer's instructions, Vector Laboratories Ltd, Peterborough, UK) and an alkaline phosphatase inhibitor, levamisole, for 30 min at RT. For size analysis, the membrane marker wheat germ agglutinin (WGA), and the nuclear stain propidium iodide (PI), were used. Cells were grown on coverslips and at days 1 and 14 were fixed in PFA. Cells were stained with 5 µ g/ml of WGA for 20 min RT and washed in PBS. Then, cells were permeabilised with 0.2% Triton(TM) X-100 for 10 min and then incubated with 10 µg/ml PI and RNAse A for 45 min at 37°C and washed with PBS. In both experiments, coverslips were then fixed onto slides with VECTASHIELD® Mounting Media and sealed. Stained cells were viewed on a Ziess Confocal at x63 magnification. Tight junction analysis. Cells were grown in 0.4 µMinserts for21d.Membraneswerefixedin3%PFAfor45minatRT. CellswerethenwashedinPBS(3x5mineach).Cellswere incubated with 0.5% Triton(TM) X-100 for 10 min. Cells were then washed twice with PBS and blocked in 5% BSA in PBS for 1 h at RT. Cells were then incubated for 2 h at RT with an apical application of a 1:100 dilution of Mouse ZO-1 Alexa Fluor 488 conjugated antibody (Life Technologies, Paisley, UK). Membranes were washed with PBS (3 x 5 min) and membranes were removed from the membrane insert using a scalpel blade, and fixed onto slides using VECTASHIELD® Mounting Media. Imaging was completed using confocal microscopy, as above. Phalloidin cytoskeletal staining. Cells were grown on 8 well chamber slides (BD Bioscience, Oxford, UK) for 21 d and fixed in 4% PFA for 45 min at RT. Cells were then washed in PBS before permeabilisation with 0.5% Triton(TM) X-100. Cells were blocked in 5% BSA in PBS for 20 min at RT and incubated with phalloidin tetramethylrhodamine B isothiocyanate (TRITC)-conjugated solution 5 Units/ml (a kind donation from Dr Alan Shirras, Faculty of Health and Medicine, Lancaster University, UK) for 20 min at RT. Cells were washed 3 times with PBS and mounted with VECTASHIELD® Mounting Media containing DAPI. Imaging was completed using confocal microscopy, as previous. Immunoblotting. Cells were grown in 6 well culture plates and, at the times indicated, washed in PBS and lysed in RIPA buffer (150 mM NaCl, 1.0% IGEPAL® CA-630, 0.5% sodium deoxycholate, 0.1% SDS, and 50 mM Tris, pH 8.0) with protease inhibitor cocktail (#P8340). Protein content of lysates was determined using Bradford reagent. Fifteen micrograms of protein was then processed on a precast 10% SDS-PAGE gel and transferred onto nitrocellulose membrane (BioRad, Hertfordshire, UK) before blocking with 5% non-fat milk in Tris-buffered saline with 0.1% Tween (TBST) at RT for 1 h. Primary ?-actin antibody (Cell Signaling, New England Biolabs, Hertfordshire, UK) at 1:10,000 was incubated overnight at 4°C. This was then bound to 1:10,000 Anti-rabbit IgG, HRP-linked (Cell Signaling, New England Biolabs, Hertfordshire, UK) secondary antibody incubated at room temperature for 1 h. Blots were exposed using Clarity(TM) Western ECL Substrate (BioRad, Hertfordshire, UK) and imaged with the BioRad ChemiDoc(TM) XRS system. Oxygen consumption rate measurements. Caco-2 cells were plated at 5 x 104 cells/well and maintained overnight in medium supplemented with FBS and NEAA at 37°C. Cells were then washed twice with PBS and changed into unbuffered serum free DMEM medium (Sigma-Aldrich Company Ltd, Dorset, UK) and placed in a CO2 free 37°C incubator before the experiment, whilst the injector plate equilibrated. An XF Flux analyser from Seahorse Bioscience (Copenhagen, Denmark) was used to measure the oxygen consumption rate (OCR) of cells in response to CBD on proliferating and differentiated Caco-2 cells over 12 h, alongside wells for testing mitochondrial function using an XF Cell Mito Stress Test Kit (Seahorse Bioscience Europe). Data was generated in picomoles of oxygen per minute and percentage baseline values were used to normalise against control cells for each treatment, data is presented as percentage of control cells. Mitochondrial staining. Cells grown in complete medium were plated onto coverslips in six-well plates at 1x 106 cells/ well and left to adhere overnight. Cells were stained in MEM without FBS for 30 min at 37°C in each environment with a suspension of 500 nM MitoTracker® Red CM-H2 XRos and 150 nM MitoTracker® Green FM (Life Technologies, Paisley, UK). Cells were washed with prewarmed MEM and then incubated with 1 and 10 µM of CBD for 2 h and then imaged using a confocal microscope. Data analysis. Data was analysed and graphs constructed using Microsoft Excel 2010. Paired T tests were used for the statistical analysis of cell number data and cannabidiol proliferation data was converted to area under the curve (AUC) and compared by paired T test (p < 0.05). Data collected using the Seahorse XF flux analyser was converted to percentage of control and a one-way ANOVA with Tukey'spost-hoctest was performed on AUC data using SPSS v21 (p <0.05). Results PhysO 2 environment decreases cell number, cell divisions and size over the differentiation period. The Caco-2 cell line has been reported as proliferating during its differentiation period before culminating in a plateau post-confluence (Pinto et al. 1983). However, stem cell studies have shown that cells grown at lower oxygen levels result in reduced proliferation (Shin et al. 2012). To assess the effect of PhysO2 on Caco-2 proliferation, trypan blue exclusion counts were conducted over 21 d. Cells reached confluence between days 3 and 7 and continued to proliferate during the differentiation period up to 21 d. Cells were able to reach confluence by day 7 when transferred to the PhysO2 environment after plating, but did not proliferate beyond day 10 (Fig. 1a). AtmosO2 cells were significantly higher in number by day 10 than cells transferred to the PhysO2 environment showing a 1.4-fold difference (±0.07) and a final fold difference of 3.7 (±1.1) by day 21. This proliferation profile was comparable to cells adapted to the PhysO2 environment and was reversible; in that, cells transferred back into the AtmosO2 environment regained their proliferative capacity showing a 3.09 fold difference (±0.11) by day 21 (Fig. 1b). In order to establish whether the difference in cell number was related to fewer cell divisions, cell division analysis using the cell division tracker dye, CFSE, showed that PhysO2 undergo fewer divisions in the same time as AtmosO2 cells and that the effect was reversible (Fig. 1c). Proliferating and differentiated cell membranes were stained with WGA and nuclei stained with PI to assess whether this effect was related to a difference in cell size. Proliferating cells from both environments have no marked difference in cell size (Fig. 1d,left and right top panels), whereas the nuclei of differentiated PhysO2 cells were two to three times larger by day 14 (Fig. 1d, compare left and right bottom panels). Caco-2 cells grown in PhysO 2 conditions maintain a number of key model characteristics. The classical Caco-2 model differentiates over a period of 14-21 d, during this time Caco-2 cells show an increase in intestinal epithelial cell markers, including alkaline phosphatase and form a cell monolayer with tight junctions providing a model which mimics intestinal barriers (Pinto et al. 1983;Haraetal.1993). These characteristics render the Caco-2 model useful for uptake and permeability studies. Staining cells with an alkaline phosphatase dye showed that AtmosO2 and PhysO2 maintain the ability to differentiate in the PhysO2 environment, demonstrated by the increase in fluorescence from day 3 to 14 (Fig. 2a). PhysO2 cells also maintain their ability to express the tight junction protein, ZO-1 (Fig. 2b) and there is no significant difference in the permeability of 21 d PhysO2 cells to the fluorescent FD4 dye when compared to AtmosO2 monolayers (Fig. 2c). It is noteworthy that although the PhysO2 cells maintain these key characteristics they are larger in size than AtmosO2 Caco-2 cells. Using TRITCphalloidin, which binds to the cytoskeletal protein F-actin, we again observed this difference in size (Fig. 2d). AtmosO2 and PhysO2 cells have similar cytoskeletal structure, possibly with a more cytoplasmic distribution in some PhysO2 cells (Fig. 2d, left and right top panels respectively), but the DAPI staining of the nucleus (Fig. 2d, left and right bottom panels respectively) suggests the PhysO2 cells have more cytoplasmic space as the AtmosO2 nucleus fills the majority of the size of the cell. This is in keeping with the WGA staining (Fig. 1d) that shows PhysO2 cells are larger in size. Western blot analysis of ?-actin (Fig. 2e) indicates there is no significant increase in ?-actin expression in whole cell extracts of AtmosO2 cells, Atmos-PhysO2 cells or PhysO2 cells up to and including 21 d. Sensitivity to cannabidiol differs in proliferating and differentiating Caco-2 cells. CBD has been shown to have antiproliferative effects on a number of cancer cell lines, including the Caco-2 cell line (Massi et al. 2004; Aviello et al. 2012). In our experiments, with cells grown in both 1% and 5% serum conditions, PhysO2 cells are significantly more sensitive to the antiproliferative effect of CBD (Fig. 3a, b) (up to 10 µM) than AtmosO2 cells, with a left hand shift in concentration effect over 72 h. In low-serum conditions (1%) there is an increase in proliferation between 0.1 and 1 µM CBD followed by a decrease in proliferation at 5-10 µM CBD in AtmosO2 cells. However, in the PhysO2 cells, there is an increase in proliferation up to 0.1 µM and a decrease between 0.5 and 10 µM(Fig.3a). In 5% serum, there is no pro-proliferative effect in cells grown in either environment, but there is a resulting decrease in proliferation at 5-10 µMinAtmosO2 cells and a decrease in PhysO2 cells at 1-10 µM(Fig.3b). In differentiated cells however, CBD does affect cell viability in either environment or differing serum conditions (Fig. 3c, d). The small chain fatty acid, butyrate, is an abundant carbon source in the colon produced by respiring microbial flora and important for healthy gut function (Bugaut 1987). Butyrate enemas have been used to alleviate inflammation and irritation in colitic patients (Scheppach 1994), and has antiproliferative and apoptotic effects on cancerous cells, possibly through its ability to function as an inhibitor of deacetylation (Yu et al. 2010). Recently, it has been reported that low doses of butyrate can increase proliferation in intestinal cells if the Warburg effect is inhibited (Donohoe et al. 2011). We wanted to know how butyrate inclusion under PhysO2 conditions would impact on cell growth and whether it would alter the CBD effects. Thus, over 72 h, butyrate has a dose-dependent antiproliferative effect in PhysO2 cells in both complete medium (8% serum, MEM with glucose) and low glucose medium (8% serum, glucose free DMEM) (Fig. 3e). At 1.25 and 2.5 mM butyrate under PhysO2 conditions, cells in low glucose medium were less sensitive to the effects of butyrate than complete medium, as demonstrated by the left hand shift of the dose response (Fig. 3e). PhysO2 cells treated with a low dose of butyrate (0.5 mM) that does not affect proliferation (from data in Fig. 3e) in low glucose medium, with CBD showed increased proliferative capacityupto1µM(Fig.3f). Previous studies have indicated a metabolic-driven action for cannabinoids (Chiu et al. 1975;Athanasiouetal.2007). Using the Seahorse XF Flux analyser, proliferating AtmosO2 Caco-2 cells show a 3.28-fold (±0.81) drop in oxygen consumption by 2.5 h when treated with 10 µM CBD. Although 10 and 100 nM were unable to cause a drop in oxygen consumption, 1 µM CBD was able to produce a significant 1.17-fold (±0.04) decrease by 2.5 h (Fig. 4a). CBD (10 nM- 10 µM) was unable to decrease the oxygen consumption rate of differentiated AtmosO2 cells in comparison to the controls (Fig. 4b). Due to limitations of Seahorse technology we could not assay the PhysO2 cells, therefore, the effect of CBD on oxidative metabolism using MitoTracker Red CM-H2XRos was investigated. In AtmosO2 conditions, CBD (10 µM) induces a loss of fluorescence within 2 h. This is related to the loss of mitochondrial membrane integrity (Fig. 4c,upperRH panel), whereas in the PhysO2 conditions, there is an increase in oxidative metabolism, also within 2 h (Fig. 4c,lowerRH panel). Discussion In vitro cell models are useful research tools as they are readily available, can be utilized for high-throughput studies and their microenvironments can be tightly controlled (Rocha and Whang 2004). The Caco-2 model is a widely used in vitro intestinal cell model and often used in barrier studies (Van De Walle et al. 2010). However, like many in vitro models, the microenvironment is limited and lacks the range of intestinal cues such as sheer stress, blood flow, and appropriate concentration gradients, carbon sources and oxygen levels (Young and Beebe 2010). The need for cell models that more closely mimic the physiological environment is more pressing, with increasing attention being paid to 3D scaffolding and multi-cell models, however very few areas have focussed on the oxygen requirement of cells (Dutta and Dutta 2009). Physiological oxygen levels being implemented in stem cell cultures have shown differences in their ability to survive and form colonies and promote differentiation, which may be beneficial to cell and organ modelling (Morrison et al. 2000; Shipp et al. 2012). Intestinal cells exist in the body at a range of 3-6% O2, yet cell and tissue models are being cultured at 21% O2. Moving the colon cancer adenocarcinoma cell line, Caco-2, from 21% O2 (AtmosO2)to5%O2 (PhysO2)resulted in an inability to maintain the proliferative numbers achieved by AtmosO2 grown cells. The proliferative profile from AtmosO2-Phys O2 cells was comparable to PhysO2 cells and was reversible, thus showing that diminished ability to maintain the cellular proliferation was due to the PhysO2 environment. This is comparable to previous research that has shown in stem cells, low-oxygen environments can lower the proliferation of some cell types and in addition some research has found low-O2 grown cells will return to a similar behaviour as high-O2 grown cells when transferred back to high O2, which is in keeping with our results (Prasad et al. 2009;Shinetal.2012). Results also indicated that PhysO2 cells underwent fewer divisions than AtmosO2 grown cells. This, however, can be attributed to the size of the cell, as WGA and PI staining showed that although there is no discernible difference of cell size in the proliferating state, by the time they reach day 21, the cells are larger in the PhysO2grown environment. This increase in size is also reflected in the cytoskeletal phalloidin staining showing the F-actin in the cells. Both the nucleus and cytoplasmic space are increased in PhysO2 cells. This change in karyoplasmic ratio may be related to genome content or changes in availability of membrane components from the endoplasmic reticulum (Webster et al. 2009), but expression of the 'housekeeping' protein ?- actin showed no significant variability in the low-oxygen environment. This interesting phenomenon was not explored further in our current study. A key property of Caco-2 cells is their ability to differentiate over a period of 21 d. Differentiation is characterised by the appearance of a brush border accompanied by an increase in enzymes, such as alkaline phosphatase, lactase and sucrose isomaltase (Chantret et al. 1988). Using an alkaline phosphatase staining kit on fixed cells from day 3 and 14 results showed that the cells did maintain their ability to differentiate in the PhysO2 environment. The change in intensity of the staining between day 3 and day 14 is less intense in the PhysO2 cells than the AtmosO2 cells, but once again, due to the PhysO2 cells being bigger, there are less tightly packed cells to emit the same fluorescence seen in AtmosO2 cells. Immunohistochemistry indicated that the PhysO2 cells have ZO-1 tight junctional proteins in both AtmosO2 and PhysO2 cells. In addition to this, there was no significant difference in FD4 flux through the membranes. This is an important characteristic that makes the Caco-2 model useful for intestinal barrier studies and it is maintained in the PhysO2 line. Further mechanistic studies on cannabinoid-induced changes in barrier function are ongoing in our lab. CBD has shown antiproliferative effects in a number of cancer cell lines, including intestinal, neuronal, immune and breast cancer lines (Massi et al. 2004; McKallip et al. 2006; McAllister et al. 2011; Aviello et al. 2012). In this study, dose response curves over a period of 72 h were used to determine whether the effects of CBD would differ at a low oxygen tension. Treatment of proliferating Caco-2 cells in 1% and 5% serum showed that PhysO2 cells were significantly more sensitive to the antiproliferative effects of CBD than AtmosO2 grown cells. Although in low-serum conditions there was a significant increase in proliferation at low doses of CBD in the PhysO2 cells this was lost by the addition of 5% serum. Recently, transfer of glioma cells from atmospheric conditions into low oxygen (1%) incubation for 24 h induced the hypoxiainducible factor (HIF)-1? (Solinas et al. 2013) and the authors state that no difference was seen in the CBD-induced antiproliferative effect, although this data was not shown. We would suggest that this may be because these cells were not adapted to the low-oxygen environment and only subjected to short-term hypoxia. Indeed, physiological oxygen levels in the brain can be quite heterogeneous (in the range of 0.5% to 7%) (Ivanovic 2009) and it would be interesting to validate CBD-induced antiproliferative effects of adapted glioma cells at the low oxygen tension. In addition, PhysO2 cells in full serum medium conditions without glucose and with low dose sodium butyrate show a biphasic response, in that low doses of CBD enhance proliferation and high doses inhibit proliferation. This is potentially important because sodium butyrate is thought to be the primary energy source for colonic epithelial cells and if CBD acts in this way in vivo, it reveals a potential pitfall in the development of CBD as an anticancer therapeutic. A proproliferative response by low concentrations of CBD in breast cancer cells not deprived of serum has previously been mentioned(Ligrestietal.2006), but we found in our system, that this pro-proliferative effect may be due to differences in levels of serum and glucose content throughout the experimental period. This highlights the need to consider what the in vivo nutrient environment might comprise before preclinical drug testing in vitro or ex vivo. Actions of phytocannabinoids on cells and tissue have been linked directly to decreases in oxidative metabolism, generation of reactive oxygen species (ROS), apoptosis and cyclooxygenase (COX)-2 (McAllister et al. 2011; Ruhaak et al. 2011; Shrivastava et al. 2011). We chose to measure the oxygen consumption rate first in proliferating AtmosO2 cells and found a rapid and steep reduction of OCR within 2.5 h of 10 µM CBD that lead to AtmosO2 cells being metabolically incapable of oxygen consumption. Unfortunately we were unable to measure this effect in the PhysO2 cells because the Seahorse XF Analyzer does not include maintenance of alternative gaseous microenvironments. The drop in oxygen consumption rate implicates a direct effect on mitochondrial function and cellular energetics. Thus, by measuring total mitochondria along with oxidative metabolism using Mitotracker dyes, AtmosO2 cells show a loss of oxidative metabolism at 10 µM CBD, which is comparable to the Seahorse data showing a loss of OCR. Oxidative metabolism in adapted cells from both environments do not appear that different at baseline, but CBD induced a mitochondrial production of ROS in PhysO2 cells, suggesting that the cellular redox environment can influence how CBD induced antiproliferative effects in PhysO2 to AtmosO2 cells. Conclusions Cells in the body have distinct metabolic profiles dependent on the location in which they reside. This can (and should) be modelled in vitro because AtmosO2 and PhysO2 cells in our study reveal a differential mechanism of action for CBD based on two parameters. The differences that we see in cell size, cell number and drug sensitivity highlight the importance of creating the right cellular context to model cellular function. Adopting this strategy can improve predictions of drug actions and contribute to reduced attrition rate of new drugs in development. Acknowledgments T. Macpherson is partly funded by a Novartis Pharmaceuticals UK Ltd (Horsham, UK) studentship with matched funding from the Biotechnology and Biological Sciences Research Council (BBSRC, Swindon, UK). K. Wright is partly funded by a Peel Trust Lectureship from The Dowager Countess Eleanor Peel Trust. We would also like to thank Dr Dave Clancy for useful discussions on data interpretation and statistics and Dr Jane Andre for assistance with the confocal imaging (both from Faculty of Health and Medicine, Lancaster University). References Athanasiou A.; Clarke A.; Turner A.; Kumaran N.; Vakilpour S.; Smith P.; Bagiokou D.; Bradshaw T.; Westwell A.; Fang L.; Lobo D.; Constantinescu C.; Calabrese V.; Loesch A.; Alexander S.; Clothier R.; Kendall D.; Bates T. Cannabinoid receptor agonists are mitochondrial inhibitors: a unified hypothesis of how cannabinoids modulate mitochondrial function and induce cell death. Biochem. Biophys. Res. Commun. 364: 131-137; 2007. Aviello G.; Romano B.; Borrelli F.; Capasso R.; Gallo L.; Piscitelli F.; Di Marzo V.; Izzo A. Chemopreventive effect of the non-psychotropic phytocannabinoid cannabidiol on experimental colon cancer. J. Mol. Med. 90: 925-934; 2012. Bauer E.; Williams B. A.; Smidt H.; Verstegen M. W.; Mosenthin R. Influence of the gastrointestinal microbiota on development of the immune system in young animals. Curr. Issue Intest. Microbiol. 7: 35-51; 2006. Bugaut M. Occurrence, absorption and metabolism of short chain fatty acids in the digestive tract of mammals. Comp. Biochem. Physiol. B 86: 439-472; 1987. Chantret I.; Barbat A.; Dussaulx E.; Brattain M. G.; Zweibaum A. Epithelial polarity, villin expression, and enterocytic differentiation of cultured human colon carcinoma cells: a survey of twenty cell lines. Cancer Res. 48: 1936-1942; 1988. Chiu P.; Karler R.; Craven C.; Olsen D. M.; Turkanis S. A. The influence of delta9-tetrahydrocannabinol, cannabinol and cannabidiol on tissue oxygen consumption. Res. Commun. Chem. Pathol. Pharmacol. 12: 267-286; 1975. Csete M.; Ourednik J.; Ourednik V.; Sakaguchi D.; NilsenHamilton M. Oxygen in the cultivation of stem cells. Stem Cell Biol. Develop Plast. 1049: 1-8; 2005. Donohoe D. R.; Garge N.; Zhang X.; Sun W.; O'Connell T. M.; Bunger M. K.; Bultman S. J. The microbiome and butyrate regulate energy metabolism and autophagy in the mammalian colon. Cell Metab. 13: 517-526; 2011. Dutta R. C.; Dutta A. K. Cell-interactive 3D-scaffold; advances and applications. Biotechnol. Adv. 27: 334-339; 2009. Espey M. G. Role of oxygen gradients in shaping redox relationships between the human intestine and its microbiota. Free Radic. Biol. Med. 55: 130-140; 2013. Hara A.; Hibi T.; Yoshioka M.; Toda K.; Watanabe N.; Hayashi A.; Iwao Y.; Saito H.; Watanabe T.; Tsuchiya M. Changes of proliferative activity and phenotypes in spontaneous differentiation of a colon cancer cell line. Jpn. J. Cancer Res. 84: 625-632; 1993. Ivanovic Z. Hypoxia or in situ normoxia: the stem cell paradigm. J. Cell. Physiol. 219: 271-275; 2009. Ligresti A.; Moriello A. S.; Starowicz K.; Matias I.; Pisanti S.; De Petrocellis L.; Laezza C.; Portella G.; Bifulco M.; Di Marzo V. Antitumor activity of plant cannabinoids with emphasis on the effect of cannabidiol on human breast carcinoma. J. Pharmacol. Exp. Ther. 318: 1375-1387; 2006. Massi P.; Vaccani A.; Ceruti S.; Colombo A.; Abbracchio M. P.; Parolaro D. Antitumor effects of cannabidiol, a nonpsychoactive cannabinoid, on human glioma cell lines. J. Pharmacol. Exp. Ther. 308: 838-845; 2004. McAllister S. D.; Murase R.; Christian R. T.; Lau D.; Zielinski A. J.; Allison J.; Almanza C.; Pakdel A.; Lee J.; Limbad C.; Liu Y.; Debs R. J.; Moore D. H.; Desprez P. Y. Pathways mediating the effects of cannabidiol on the reduction of breast cancer cell proliferation, invasion, and metastasis. Breast Cancer Res. Treat. 129: 37-47; 2011. McKallip R. J.; Jia W.; Schlomer J.; Warren J. W.; Nagarkatti P. S.; Nagarkatti M. Cannabidiol-induced apoptosis in human leukemia cells: a novel role of cannabidiol in the regulation of p22phox and Nox4 expression. Mol. Pharmacol. 70: 897-908; 2006. Morrison S. J.; Csete M.; Groves A. K.; Melega W.; Wold B.; Anderson D. J. Culture in reduced levels of oxygen promotes clonogenic sympathoadrenal differentiation by isolated neural crest stem cells. J. Neurosci. 20: 7370-7376; 2000. Pinto M.; Robineleson S.; Appay M.; Kedinger M.; Triadou N.; Dussaulx E.; Lacroix B.; Simonassmann P.; Haffen K.; Fogh J.; Zweibaum A. Enterocyte-like differentiation and polarization of the human-colon carcinoma cell-line Caco-2 in culture. Biol. Cell. 47: 323-330; 1983. Prasad S.; Czepiel M.; Cetinkaya C.; Smigielska K.; Weli S. C.; Lysdahl H.; Gabrielsen A.; Petersen K.; Ehlers N.; Fink T. Continuous hypoxic culturing maintains activation of Notch and allows longterm propagation of human embryonic stem cells without spontaneous differentiation. Cell Prolif. 42: 63-74; 2009. Rocha F. G.; Whang E. E. Intestinal tissue engineering: from regenerative medicine to model systems. J. Surg. Res. 120: 320-325; 2004. Roeselers G.; Ponomarenko M.; Lukovac S.; Wortelboer H. M. Ex vivo systems to study host-microbiota interactions in the gastrointestinal tract. Best Pract. Res. Clin. Gastroenterol. 27: 101-113; 2013. Ruhaak L. R.; Felth J.; Karlsson P. C.; Rafter J. J.; Verpoorte R.; Bohlin L. Evaluation of the cyclooxygenase inhibiting effects of six major cannabinoids isolated from Cannabis sativa. Biol. Pharm. Bull. 34: 774-778; 2011. Scheppach W. Effects of short chain fatty acids on gut morphology and function. Gut 35: S35-S38; 1994. Shin E.; Forsyth N. R.; Fricker R. A. The effect of physiological oxygen levels on GABAergic neuronal differentiation from mouse embryonic stem cells. Stem Cell Stud. 2: e3; 2012. Shipp C.; Derhovanessian E.; Pawelec G. Effect of culture at low oxygen tension on the expression of heat shock proteins in a panel of melanoma cell lines. PLoS ONE 7: e37475; 2012. Shrivastava A.; Kuzontkoski P. M.; Groopman J. E.; Prasad A. Cannabidiol induces programmed cell death in breast cancer cells by coordinating the cross-talk between apoptosis and autophagy. Mol. Cancer Ther. 10: 1161-1172; 2011. Solinas M.; Massi P.; Cinquina V.; Valenti M.; Bolognini D.; Gariboldi M.; Monti E.; Rubino T.; Parolaro D. Cannabidiol, a nonpsychoactive cannabinoid compound, inhibits proliferation and invasion in U87-MG and T98G glioma cells through a multitarget effect. PLoS ONE 8: e76918; 2013. Van De Walle J.; Hendrickx A.; Romier B.; Larondelle Y.; Schneider Y. Inflammatory parameters in Caco-2 cells: effect of stimuli nature, concentration, combination and cell differentiation. Toxicol. in Vitro 24: 1441-1449; 2010. Webster M.; Witkin K. L.; Cohen-Fix O. Sizing up the nucleus: nuclear shape, size and nuclear-envelope assembly. J. Cell Sci. 122: 1477- 1486; 2009. Young E. W.; Beebe D. J. Fundamentals of microfluidic cell culture in controlled microenvironments. Chem. Soc. Rev. 39: 1036-1048; 2010. Yu D.; Waby J. S.; Chirakkal H.; Staton C. A.; Corfe B. M. Butyrate suppresses expression of neuropilin I in colorectal cell lines through inhibition of Sp1 transactivation. Mol. Cancer 9: 276; 2010. Received: 29 August 2013 /Accepted: 2 December 2013 /Published online: 24 January 2014 / Editor: T. Okamoto © The Society for In Vitro Biology 2014 T. Macpherson * K. L. Wright () Faculty of Health and Medicine, Division of Biomedical and Life Sciences, Lancaster University, Lancaster LA1 4YG, UK e-mail: [email protected] J. A. Armstrong * D. N. Criddle NIHR Liverpool Pancreas Biomedical Research Unit, Department of Molecular and Clinical Cancer Medicine, Royal Liverpool University Hospital, Liverpool L69 3GA, UK (c) 2014 Society for In Vitro Biology |