Background/Aims: Understanding cellular mechanisms induced by hypoxia is fundamental to reduce blood-brain barrier (BBB) disruption. Nevertheless, the investigation of hypoxia on cellular pathways is complex with true hypoxia because HIF-1α has a short lifetime and rapidly reverts back to a normoxic state. To overcome this difficulty, mimetic agents of the hypoxia pathway have been developed, including the gold standard CoCl2. In this study, we proposed to compare CoCl2 and hydralazine in order to determine a suitable mimetic agent of hypoxia for the study on the BBB. Methods: We studied the cytotoxicity and the impact of these molecules on the integrity of an in vitro BBB model by comparing them to hypoxia controls. Results: We showed that the impact of hypoxic stress in our in vitro BBB model is rather similar between hydralazine and CoCl2. Chemical hypoxic stress led to an increase of BBB permeability either with CoCl2 or hydralazine. Tight junction protein expressions showed that this chemical hypoxic stress decreased ZO-1 but not occluding expressions, and cells had set up a defence mechanism by increasing expression and activity of their efflux transporters. Conclusion: Our results demonstrated that hydralazine is a better mimetic agent and more suitable than CoCl2 because it had the same effect but without the cytotoxic effect on in vitro BBB cells.

The blood-brain barrier (BBB) is a physical and metabolic barrier which separates the central nervous system (CNS) from the blood circulation. It is a multicellular neurovascular unit composed of brain microvascular endothelial cells, astrocytes, pericytes, microglia and extracellular matrix [1, 2]. Brain endothelial cells control the transport of substances between brain and blood through transporters and tight junction (TJ) proteins [3]. TJ proteins regulate the transport of ions via the development of high transendothelial electrical resistance (TEER) as well as the transport of molecules [4, 5], which create a physical blockade to paracellular diffusion [6]. This specialized phenotype is fundamental for protection of the brain. Moreover, this protection is also ensured by efflux ABC family transporter proteins which are ATP dependant and expressed by endothelial cells, like P-glycoprotein (Pgp) and multidrug resistance proteins (MRPs). These active efflux proteins act as a barrier system which reduce the exposure of the CNS to drugs, by pumping compounds out of endothelial cells in blood [4]. These barrier properties are mostly induced and maintained by the close opposition between brain endothelial cells and astrocytes [7, 8]. Since the structural and functional integrity of transporters and TJ proteins are necessary for an intact BBB, alteration of these components is a key event in the BBB’s impairment.

Hypoxia is a major stress factor which induces a BBB’s disruption. Indeed, the brain is a major consumer of oxygen and is therefore sensitive to changes in oxygen levels [9, 10]. The cellular response to hypoxia is mainly driven through activation of the HIF-1 pathway [11]. Under normoxia conditions, oxygen regulates the HIF-1α subunit, which is rapidly degraded by prolyl hydroxylation that targets its degradation in the proteasome. Hypoxia inhibits the prolyl hydroxylase domain leading to stabilization of the HIF-1α subunit in the cytoplasm. Then it is translocated to the nucleus where it binds to hypoxia responsive elements in promoter regions of target genes involved in cellular adaptation to hypoxic stress and induces their expression [12].

Elucidation of the cellular and molecular mechanisms, induced by hypoxic stress, is complex with true hypoxia because HIF-1α has a short half-life (almost five minutes) and is rapidly reverted. In this regard, a wide variety of prolyl hydroxylase domain (PHD) inhibitors, which lead to a stabilization of HIF-1α, have been developed. These inhibitors allow to create a chemical hypoxic stress and represent a useful method to investigate the BBB’s disruption by hypoxia. The PHD inhibitor usually used in studies is CoCl2 [12-14]. Cervellati et al. used CoCl2 to achieve stabilization of HIF-1α because it inhibits PHD by blocking the catalysis of prolyl hydroxylases [15]. Unfortunately, CoCl2 is a rather highly cytotoxic agent and prevents its use in in vitro studies to mimic prolonged cerebral hypoperfusion, such as in ischemic stroke. It is difficult to know whether the observed effect is due to chemical hypoxic stress induced by CoCl2 or its cytotoxicity. Other inhibitors of PHD exist such as hydralazine. Hydralazine is a vasodilator used to treat severe hypertension, congestive heart failure, myocardial infarction and pre-eclampsia [16]. Hydralazine also shows a capacity to induce a transient and physiological HIF-1α overexpression by inhibiting PHD activity [17]. In the literature, hydralazine was only used to mimic a hypoxic state in in vivo and in vitro cancer models [18]. Since hydralazine is already used in in vivo models, it is therefore interesting to test hydralazine as a mimetic agent of hypoxia because others mimetic agents, like CoCl2, are rather toxic to be used on in vivo models. In a recent work, we validated hydralazine as a mimetic agent of hypoxia on an in vitro BBB model in comparison with true hypoxia [19].

In the present study, we wanted to determine a suitable mimetic agent of hypoxia which is not cytotoxic, reproducible and controllable. For that, we compared hydralazine and CoCl2 to study the impact of prolonged chemical hypoxia on an in vitro transwell BBB model. We used a co-culture in contact model composed of the immortalized cell line bEnd.3 and the C6 cell line (which display astrocytic properties [20]). This approach allowed interaction between endothelial cells and astrocytic cells. Then impact of prolonged chemical hypoxic stress, induced by hydralazine or CoCl2, was assessed by studying endothelial paracellular permeability with TEER measurements and absolute membrane permeability was determined with sodium fluorescein (Na-F) [21, 22]. Evaluation of transport was also investigated on expression and activity of two efflux transporters (Pgp and MRP-1) and on expression of two TJ proteins (ZO-1 and occludin). Our results showed that hydralazine represented a suitable and reproducible way to create an in vitro chemical hypoxic environment without cytotoxicity. Moreover we demonstrated that hydralazine induced the same effect on our cells than true hypoxia.

Chemicals and reagents

Hydralazine, CoCl2, BCECF-AM, probenecid, verapamil and Na-F were purchased from Sigma-Aldrich (St. Quentin Fallavier, France). MTT (methyl-thiazolyl-tetrazolium) kit and DMEM (Dulbecco’s Modified Eagle’s Medium) were also purchased from Sigma. LDH release test was purchased from Promega Corporation (Madison, USA). Antibodies and reagents for the detection of HIF-1α were products of R&D systems (Lille, France). All compounds for Ringer-Hepes buffer were purchased from Sigma.

Occludin and ZO-1 antibodies were purchased from Life Technologies (Saint Aubin, France), Pgp and MRP-1 antibodies were purchased from GeneTex (San Antonio, Texas, U.S.A) and Santa Cruz Biotechnology (Dallas, Texas, U.S.A), respectively.

bEnd.3 cells and C6 cells were obtained from the ATCC (Manassas, VA, U.S.A).

Cell culture inserts for 24-well (0.4 µm pore diameter size, transparent PET membrane) were purchased from Corning distributors (Sigma). EVOM voltohmmeter system was purchased from World Precision Instruments (Hertfordshire, UK).

In vitro cytotoxicity assay

Cytotoxicity of hydralazine or CoCl2 was measured by the MTT method. Growing cells were seeded at 10,000 cells/well using a 96-well microplate that was supplemented with DMEM. Cells were allowed to grow for 48 h before they were exposed to drugs. Then cells were treated with various concentrations of hydralazine or CoCl2. Phosphate buffer saline (PBS) 1X was used to dissolve drugs. After the treated cells were incubated for 24 h, MTT solution was added and the plates were incubated at 37°C for 4 h. To dissolve formazan, 100 µl DMSO was added and the plates were measured at 570 nm with a spectrophotometer. The least cytotoxic values were determined by plotting the drug concentrations versus the survival ratio of treated cells.

Then we also evaluated cell death with the LDH release method to confirm the least cytotoxic concentrations given by MTT results. For that, cells were seeded in 96-well plates at 10,000 cells/well. 100 μM of hydralazine or CoCl2 was added during 24 h. After the exposure period, a lysis solution was used in control wells to generate a maximum LDH release. Then cells were incubated with cytotox-ONETM reagent during 10 min. A stop solution was added before recording fluorescence with an excitation wavelength of 560 nm and an emission wavelength of 590 nm. The average fluorescence values of the culture medium background (wells without cells) were subtracted from fluorescence values of experimental wells. Then percent cytotoxicity was given by the following equation:

Exposure to true hypoxia

Our in vitro BBB cells were exposed to 2% 02 into a hypoxic chamber during 24h [19].

Effect of hydralazine or CoCl2 treatment on hypoxia metabolic pathway

Effect of hydralazine or CoCl2 treatment was validated by the study of HIF-1α expression (the key regulation of hypoxic response) to determine the effective dose where HIF-1α was induced. We also exposed cells to 2% O2 as hypoxia controls. These controls allowed us to verify that hydralazine or CoCl2 mimic hypoxic cellular environment. For that, cells were seeded onto a 96-well plate at a density of 10,000 cells/well. Then cells were treated with either hydralazine or CoCl2 (50 and 100 µM; the least cytotoxic doses defined by cytotoxic test) during 24 h. Finally, cells were fixed to the support with 4% formaldehyde and the expression of HIF-1α was determined by whole cell-ELISA. For that, anti-HIF 1α antibodies were added at 4°C overnight (diluted at 1/100), then cells were incubated with secondary antibodies for 2 h at room temperature (diluted at 1/500). Fluorescence was measured with a fluorescence spectrophotometer with 540 nm excitation and 600 nm emission wavelengths.

In vitro BBB model set up

The BBB model was composed of bEnd.3 cells (immortalized mouse brain endothelial cells) and C6 cells (rat malignant glioma cells). All cells were used before passage number 30 (passages 12-28 for bEnd.3 cells and 10-30 for C6 cells), which corresponded to the time where these cells may begin to lose their BBB properties, as cited by the supplier and in a previous study [23].

Once the cells reached subconfluency, they were placed onto cell culture inserts for 24-well plates. Figure 1 shows a schematic description of the in vitro BBB model.

Fig. 1.

Schema of the in vitro blood-brain barrier model.

Fig. 1.

Schema of the in vitro blood-brain barrier model.

Close modal

For the procedure of contact co-culture, the transwell filter was inverted and C6 cells were seeded onto the abluminal side of the filter at a density of 4 x 104 cells/cm2. Then cells were placed at 37°C for 6 h (time necessary for the cells to adhere to the membrane of the insert). Afterward, the insert was flipped back and C6 cells were cultured for two days in DMEM. At the end of two days, bEnd.3 cells were seeded onto the luminal side of the transwell filter at a density of 4 x 105 cells/cm2 and co-cultured with C6 cells for eight days.

TEER measurements

To characterize the formation of a tight endothelial cell monolayer, TEER was recorded using an EVOM resistance meter. One electrode was placed on the luminal side and the other electrode on the abluminal side. These two electrodes were separated by the endothelial layer. The TEER measurements of blank filters (without cells) were subtracted from those of filters with cells. Then the resulting value was multiplied by the membrane area to obtain the TEER measurements in Ω.cm2.

Na-F permeability measurements

Endothelial paracellular barrier function was also evaluated by measuring the permeability of cells to Na-F (MW = 376 Da). First the medium was removed and cells were washed with Ringer-Hepes prewarmed buffer (5 mM Hepes, 5.2 mM KCl, 2.2 mM CaCl2, 0.2 mM MgCl2, 6 mM NaHCO3 and 2.8 mM glucose). Then Ringer-Hepes buffer containing 10 µg/ml of Na-F was loaded onto the luminal side of the insert and incubated at 37°C for 1 h. Samples were removed from the abluminal chamber at 10, 20, 30, 40, 50 and 60 min and immediately replaced with fresh Ringer-Hepes buffer. The concentration of fluorescein was determined using a fluorescence multiwell plate reader (Ex (λ) 485 nm; Em (λ) 530 nm). Transendothelial permeability coefficient (Pe) was calculated as previously described by Deli et al. [24]. The increment in cleared volume was calculated by dividing the amount of compound in the receiver compartment by the drug concentration in the donor compartment. The volume cleared was plotted versus time and the slope estimated by linear regression analysis. Then the average cleared volume was plotted versus time, and permeability x surface area product value for endothelial monolayer (PSe) was calculated as follows:

PSendothelial divided by the surface area (A) in cm2 (0.33cm2 for transwell-24) generated the endothelial permeability coefficient (Pe in 10-6cm.s-1) and calculated as follows:

Whole cell ELISA of Pgp, MRP1, ZO-1 and occludin

Co-cultured cells were washed with 1% BSA in PBS at pH 7.4 and fixed for 20 min at room temperature with 4% paraformaldehyde in PBS at pH 7.4. Then inserts were washed again and overlaid with 3% H2O2 in methanol for 30 min to block endogenous peroxidase, followed by 20% normal goat serum to block unspecific immunoglobulin binding. Cells were incubated all night at 4°C with either monoclonal mouse anti-Pgp (2 µg/ml), monoclonal mouse anti-MRP-1 (10 µg/ml), rabbit anti-ZO-1 (4 µg/ml) or rabbit anti-occludin (1 µg/ml) antibodies. Then cells were washed and secondary antibodies were added for 2 h at room temperature (diluted at 1/500). After washing several times, 0.1% of o-phenylenadiamine and 0.002% H2O2 in 0.05 M citrate buffer at pH 4.5 was added for 10 min at room temperature and in the dark. The colour reaction was measured with a spectrophotometer at 490 nm, as described by Cioni et al. [5].

Drug transporter functional assays

The functionality of Pgp and MRP-1 was tested by assessing the release of the substrate BCECF-AM (specific for Pgp and MRP-1) in the absence or presence of specific inhibitors such as verapamil for Pgp and probenecid for MRP-1. The co-cultured cells were washed and cultured in DMEM. Then cells were preincubated for 15 min in the absence or presence of inhibitors at 37°C (40 µM verapamil and 1 mM probenecid; concentrations were determined previously [25]). Inhibitors were added in the luminal compartment to study the transport from the luminal to the abluminal side and conversely. The inserts were incubated with 1 mM of BCECF-AM for 1 h at 37°C. Finally, 100 µl was sampled from each compartment and fluorescence was measured with a fluorescence spectrophotometer at 493 nm excitation and 515 nm emission wavelengths.

Statistical analysis

The values are expressed as the means ± s.e.m. Each test was repeated three times and made in triplicate (n = 9). Statistical analysis was performed using Student’s t-test. One-way and two-way analyses of variance (ANOVA) followed by Tukey-Kramer’s tests were applied to multiple comparisons. Statistical analysis was performed using Statview software. The differences between means were considered to be significant when P values were less than 0.05.

In vitro cytotoxicity assay

bEnd.3 cells were exposed to a range of concentrations of hydralazine or CoCl2 (25 – 300 µM) according to the literature. We studied the effect of hypoxic stress with an incubation time of about 24 h because the impact of prolonged hypoxia on BBB is less described than short exposure and difficult to monitor with chemical agents like CoCl2, which has an important cytotoxicity [15]. After 24 h of exposure to drugs, cytotoxicity activity was evaluated by the MTT method (Fig. 2A). IC50 values of drugs were 200 µM for hydralazine and 100 µM for CoCl2, respectively. So the least cytotoxic concentrations for hydralazine and CoCl2 are 50 and 100 µM. We verified the cytotoxicity of these concentrations by evaluating cell death with the LDH release method (Fig. 2B). We showed that hydralazine had a percentage of cytotoxicity about 2.36 ± 0.4% for 50 µM and 17.77 ± 4.9% for 100 µM; whereas CoCl2 had a percentage of cytotoxicity about 50.84 ± 4.9% for 50 µM and 68.96 ± 7% for 100 µM.

Fig. 2.

Cytotoxicity effect of drugs in bEnd.3 cells. Cells were incubated with various concentrations of hydralazine or CoCl2 for 24 h. Cytotoxicity was measured by an MTT assay (A) and LDH release method (B). The results are presented as a mean value for triplicates.

Fig. 2.

Cytotoxicity effect of drugs in bEnd.3 cells. Cells were incubated with various concentrations of hydralazine or CoCl2 for 24 h. Cytotoxicity was measured by an MTT assay (A) and LDH release method (B). The results are presented as a mean value for triplicates.

Close modal

As expected, CoCl2 exposure produced a high cytotoxicity. On the contrary, hydralazine was not cytotoxic and can be used for 24 h of exposure.

Effective dose of drugs induced HIF-1 pathway

Hydralazine and CoCl2 are chemical agents which could mimic hypoxia since they inhibited PHD, which negatively regulates HIF-1. First we determined the effective dose for each drug which induced HIF-1α overexpression. For that, bEnd.3 cells were exposed to 50 and 100 µM of hydralazine and CoCl2 during 24 h. The concentration 50 µM of hydralazine and CoCl2 did not induce HIF-1α (data not shown). After 24-h treatment with 100 µM hydralazine, the level of HIF-1α significantly increased by 226% (p < 0.05); whereas 100 µM of CoCl2 significantly induced HIF-1α after 24 h of treatment, with an increase of 239% (p < 0.05) (Fig. 3). Consequently, we used the concentration of 100 µM in our studies. In parallel, cells were exposed to 2% 02 and validated that hydralazine or CoCl2 induced HIF-1α in a same way as true hypoxia. Indeed, our results showed that the level of HIF-1α significantly increased by 245% (p < 0.05).

Fig. 3.

HIF-1 expression induced by each drug and true hypoxia. bEnd.3 cells were treated with 100 µM hydralazine, CoCl2 or exposed to 2% O2. Results were presented as mean value ± s.e.m (n = 9) and normalized. * P < 0.05 normoxia vs. exposure (hypoxia or drugs).

Fig. 3.

HIF-1 expression induced by each drug and true hypoxia. bEnd.3 cells were treated with 100 µM hydralazine, CoCl2 or exposed to 2% O2. Results were presented as mean value ± s.e.m (n = 9) and normalized. * P < 0.05 normoxia vs. exposure (hypoxia or drugs).

Close modal

Impact of CoCl2 and hydralazine on BBB integrity

On TEER measurements. At the optimal TEER measurement (D6 established at the laboratory), cells were exposed 24 h to either CoCl2, hydralazine or true hypoxia. Figure 4 shows the impact of addition of these molecules on TEER in an in vitro BBB model. A decrease in TEER was observed after 24 h for each molecule. For CoCl2, TEER significantly decreased by 55%. TEER values varied from 83 ± 2 to 37 ± 4 Ω.cm2 (p < 0.05). For hydralazine, TEER significantly decreased by 58%. TEER values varied from 83 ± 5 to 34± 4 Ω.cm2 (p < 0.05). Hypoxia controls showed that TEER significantly decreased by 46% (p < 0.05).

Fig. 4.

TEER measurement after cells of the blood-brain barrier model were exposed to hydralazine, CoCl2 or true hypoxia during 24 h. Results are presented as mean value ± s.e.m (n = 9). * P < 0.05 normoxia vs. exposure (hypoxia or drugs).

Fig. 4.

TEER measurement after cells of the blood-brain barrier model were exposed to hydralazine, CoCl2 or true hypoxia during 24 h. Results are presented as mean value ± s.e.m (n = 9). * P < 0.05 normoxia vs. exposure (hypoxia or drugs).

Close modal

On membrane permeability. Absolute membrane permeability was determined after 24-h incubation with hydralazine, CoCl2 or exposed to 2% 02, and results are illustrated by Fig. 5. Coefficient permeability of fluorescein in the BBB model significantly increased from 6.06± 0.6x10-6 to 6.66 ± 0.46x10-4 cm.s-1 for CoCl2 (p < 0.05). Cells of the BBB model treated with hydralazine showed a significant increase in fluorescein permeability (p < 0.05). The permeability coefficient of fluorescein varied from 9.09 ± 0.1x10-6 to 1.57 ± 0.3x10-4 cm.s-1. Moreover, the permeability coefficient of fluorescein significantly increased after exposure to true hypoxia (2.98 ± 0.9x10-4 cm.s-1).There was no significant difference between the three exposures.

Fig. 5.

Absolute membrane permeability measurement during 24 h of hydralazine, CoCl2 or true hypoxia exposure versus normoxia. The membrane permeability of fluorescein is expressed in cm.s-1. Results are presented as mean value ± s.e.m (n = 9). * P < 0.05 normoxia vs.exposure (hypoxia or drugs).

Fig. 5.

Absolute membrane permeability measurement during 24 h of hydralazine, CoCl2 or true hypoxia exposure versus normoxia. The membrane permeability of fluorescein is expressed in cm.s-1. Results are presented as mean value ± s.e.m (n = 9). * P < 0.05 normoxia vs.exposure (hypoxia or drugs).

Close modal

On paracellular transport mediated TJ proteins occludin and ZO-1. Determination of occludin and ZO-1 expressions in our in vitro BBB model were established after exposure to hydralazine, CoCl2 or true hypoxia (Fig. 6). Expression of ZO-1 significantly decreased after these exposures (p < 0.05). Expression of ZO-1 decreased by 37% with hydralazine, by 35% with CoCl2 and by 46% with true hypoxia. However, there was no significant difference for occludin’s expression with these three exposures. The concentration of occludin varied from 3.78 to 3.04 ± 0.63 µg/ml with hydralazine, from 3.78 to 2.84 ± 0.35 µg/ml with CoCl2 and from 3.78 to 3.20 ± 0.52 µg/ml with true hypoxia. There was no significant difference between these exposures.

Fig. 6.

Concentrations of ZO-1 and occludin after cells of the blood-brain barrier model were exposed to hydralazine, CoCl2 or true hypoxia and determined by cell-ELISA. Results are presented as mean value ± s.e.m (n = 9). * P < 0.05 normoxia vs. exposure (hypoxia or drugs).

Fig. 6.

Concentrations of ZO-1 and occludin after cells of the blood-brain barrier model were exposed to hydralazine, CoCl2 or true hypoxia and determined by cell-ELISA. Results are presented as mean value ± s.e.m (n = 9). * P < 0.05 normoxia vs. exposure (hypoxia or drugs).

Close modal

On transendothelial transport mediated efflux transporters. Transendothelial transport in our model was evaluated with two efflux transporter pumps, Pgp and MRP-1. It was evaluated by studying the transport of a substrate of Pgp and MRP-1, i.e., BCECF-AM, with a method developed in our laboratory [26]. BCECF-AM is cleaved by intracellular esterase into a fluorescent component BCECF. In a previous work, we demonstrated that pumps were more active on apical membranes of endothelial cells (transport A to B), since BCECF was more efflux [19]. So we investigated if the transport of BCECF from A to B was changed after hydralazine, CoCl2 or true hypoxia exposure (Fig. 7A). We noticed that efflux transporters prevented entry of BCECF in cells, since an important transendothelial transport was observed after each exposure. A significant increase of extracellular BCECF was also observed after hydralazine exposure: 5.13 ± 0.97 µg/ml versus 8.80 ± 0.78 µg/ml (p < 0.05) together with CoCl2, where extracellular BCECF varied from 5.13 ± 0.97 µg/ml to 8.88 ± 0.41 µg/ml (p < 0.05). We also observed a significant increase of extracellular BCECF with true hypoxia (9.91 µg/ml). In parallel, we observed no significant difference between the two chemical agents and hypoxia controls. The relative importance between Pgp and MRP-1 was evaluated with two specific inhibitors of Pgp and MRP-1, i.e., verapamil and probenecid, respectively. We observed no significant difference between these two transporters.

Fig. 7.

(A) Release of BCECF during transport on apical membranes of endothelial cells, after exposure to hydralazine, CoCl2 and true hypoxia. (B) Concentrations of MRP-1 and Pgp after cells of the blood-brain barrier model were exposed to hydralazine, CoCl2 and 2% 02 during 24 h. Results are presented as mean value ± s.e.m (n = 9). * P < 0.05 BCECF vs. inhibitors and* P < 0.05 normoxia vs. exposure (hypoxia or drugs).

Fig. 7.

(A) Release of BCECF during transport on apical membranes of endothelial cells, after exposure to hydralazine, CoCl2 and true hypoxia. (B) Concentrations of MRP-1 and Pgp after cells of the blood-brain barrier model were exposed to hydralazine, CoCl2 and 2% 02 during 24 h. Results are presented as mean value ± s.e.m (n = 9). * P < 0.05 BCECF vs. inhibitors and* P < 0.05 normoxia vs. exposure (hypoxia or drugs).

Close modal

In a second time, we studied the expression of Pgp and MRP-1 proteins in our in vitro BBB model after the three exposures, in order to understand the increase in the pump’s observed activity (Fig. 7B). MRP-1 expression significantly increased after incubation with each exposure. For hydralazine MRP-1 increased by 91% (p < 0.05), by 136% (p < 0.05) for CoCl2 and by 90% for true hypoxia (p < 0.05). There was no significant difference between hydralazine, CoCl2 and hypoxia controls. Pgp expression significantly increased by 439% (p < 0.05), 502% (p < 0.05) and 371% for hydralazine, CoCl2 and true hypoxia, respectively. There was no significant difference between these exposures.

This study on transendothelial transport mediated efflux transporters showed that the functionality of Pgp and MRP-1 increased after chemical and true hypoxic stress, which can be explained by an overexpression of Pgp and MRP-1 proteins.

The BBB is a fundamental structure for the CNS because it confers a protection for the brain. This protection is mainly ensured by TJ proteins and efflux transporter proteins [27]. On the one hand, TJ proteins create a tightness barrier and reduce paracellular diffusion. On the other hand, efflux transporter proteins actively pump compounds out of endothelial cells to reduce the entry of potential cytotoxic substances. However, this BBB can be disrupted in some cases. Hypoxia is a key factor of the BBB’s disruption and is involved in some neurological diseases. The cellular response to hypoxia is driven by the HIF-1 pathway [12]. Nevertheless, elucidation of cellular mechanisms induced by hypoxic stress is rather difficult with true hypoxia because HIF-1α has a short half-life, is rapidly reverted and thus less reproducible. In this regard, we present in this study a suitable method of chemical hypoxic stress induction via mimetic agents of the hypoxia pathway. We decided to compare the capacity of a potential mimetic agent, hydralazine, with the classical agent used in the literature, CoCl2 [14, 28], to determine the suitable mimetic agent for the study of chemical hypoxic stress on the BBB. Moreover, we validated their effects by comparing with hypoxia controls. This induction method was set up to understand cellular mechanisms involved by chemical hypoxic stress on an original in vitro BBB model. We exposed our model to hydralazine or CoCl2 during 24 h to study the effect of prolonged chemical hypoxic stress on the BBB to mimic prolonged cerebral hypoperfusion, like in ischemic stroke. In a first step, we demonstrated that these two molecules were able to induce a chemical hypoxic state on our cells since we observed an increase of HIF-1α protein expression, similar to the expression observed with hypoxia controls. Nevertheless, we showed that 100 µM of CoCl2 was the IC50 for bEnd.3, meaning that at this concentration there were 50% dead cells and at a lower concentration there was no induction of HIF-1α. Cervellati et al. explained that this decrease of viability in cells induced by CoCl2 must be due to activation of caspase-3, which leads to apoptosis [15]. This decrease in viability of cells could be a problem because we could not determine whether the observed effect was due to the inhibition of the PHD induced by CoCl2 or its cytotoxicity.This has also been confirmed by other researchers using these agents [14, 15, 17]. Moreover, we observed the same result for HIF-1 expression after 24-h exposure to true hypoxia. So we decided to use this concentration to investigate the consequences of chemical hypoxic stress on our in vitro BBB model. Our results showed that chemical hypoxic stress, induced by hydralazine or CoCl2, significantly decreased TEER value and significantly increased membrane permeability to fluorescein. Taken together, these results demonstrated that the BBB lost its physical barrier properties after chemical hypoxic stress since ionic compounds (measured by TEER value) and molecules (measured by Na-F) could cross cell membranes.

To understand this loss, we investigated the expression of TJ proteins since they are in charge of these physical barrier properties. Chemical hypoxic stress significantly decreased ZO-1 but not occludin expression.This result confirmed those obtained by Engelhard et al. with a CoCl2 and true hypoxia approach [12] and must be explained by the fact that ZO-1 has an important role in the development and barrier maintenance of the BBB [29]. However, an increase of VEGF via HIF-1 pathway decreased ZO-1 protein expression and lead to a rearrangement of ZO-1 in brain endothelial cells [30, 31]. It was shown that hydralazine induced a stabilization of HIF-1α and increased VEGF secretion [17] So,this suggested that HIF-1α signalling impacts ZO-1 localization via VEGF pathway but has no effect on occludin localization. Animal experiments (knockdown for occludin) have shown that occludin was not essential for the establishment of TJ complexes, despite the fact that a loss of this protein was involved in many diseases. Occludin would participate in the regulation of the BBB and more particularly in the regulation of calcium flux through the BBB [32].

We then investigated the impact of prolonged chemical hypoxic stress on efflux transporters as they conferred a protection for the brain by rejecting potential dangerous compounds. We showed that Pgp and MRP-1 were significantly active since they rejected more BCECF during chemical hypoxic stress; but we noticed no significant difference between the two pumps. This result was shown in tumors cells which are rather hypoxic cells. In these cells, HIF-1α was overexpressed and was associated with an increase of Pgp and MRP-1 activity, which explained a chemoresistance for these cells [33]. Moreover, this result confirmed those obtained with cells exposed to 2% O2. Finally, we showed that this increase in activity was linked to an overexpression of Pgp and MRP-1 proteins after chemical hypoxic stress. Taken together, these results demonstrated that cells established a defence mechanism to protect the brain from chemical hypoxic stress, but it would create a resistance to therapy, as it has been described for anti-epileptic drugs and chemotherapy [34]. This should be taken into consideration during the development of therapeutics.

To conclude, this study was undertaken to define a suitable mimetic hypoxia agent in order to investigate and understand cellular and molecular impacts of hypoxic stress on an in vitro BBB model. Our in vitro BBB model is composed of immortalized cell lines and presents a lower TEER values than primary cells, which reproduce more physiologically the cellular and structural characteristics of an in vivo BBB. Nevertheless, the permeability and the expression of TJ proteins respond to the expected characteristics of an in vitro BBB model [35]. Moreover, these immortalized cells allow to obtain a reproducible model, and to study the impact of hypoxic stress over time, which is not necessary the case with primary cells. Our results presented hydralazine as a suitable candidate to create a chemical hypoxic state in our in vitro BBB model, since it was able to induce HIF-1α without any cytotoxic effect. Moreover, we showed that hydralazine induced the same effect on our in vitro BBB model than hypoxia controls. This result is in line with those described in the literature [17, 19]. CoCl2 could be a useful tool to create a hypoxic state, but it was too cytotoxic for our cells. Hence induction of chemical hypoxic stress by hydralazine is a suitable mimetic agent of hypoxia to understand the early stages of the consequences of hypoxic stress on the BBB. This induction method has many advantages on true hypoxia because chemical hypoxic stress is standardization controlled, easy to produce, reproducible and reversible. It is also less expensive. We are confident that this induction method by hydralazine will allow the understanding of cellular and molecular mechanisms activated at the BBB during hypoxic events like repeated hypoxic stress.

This work was supported by grants from Jean Monnet University of Saint-Etienne.

The authors thank the Dr Anne Briançon-Marjollet (INSERM 1042 HP2 Laboratory of Grenoble, France) for her help in performing physical hypoxia experiments.

The authors declare no conflict of interest.

1.
Hawkins BT, Davis TP: The blood-brain barrier/neurovascular unit in health and disease. Pharmacol Rev 2005; 57:173–185.
2.
Daneman R, Prat A: The blood-brain barrier. Cold Spring Harb Perspect Biol 2015; 7:a020412.
3.
Helms HC, Abbott NJ, Burek M, Cecchelli R, Couraud P-O, Deli MA, Förster C, Galla HJ, Romero IA, Shusta EV, Stebbins MJ, Vandenhaute E, Weksler B, Brodin B: In vitro models of the blood-brain barrier: An overview of commonly used brain endothelial cell culture models and guidelines for their use. J Cereb Blood Flow Metab 2016; 36:862-890.
4.
Kido Y, Tamai I, Nakanishi T, Kagami T, Hirosawa I, Sai Y, Tsuji A: Evaluation of blood-brain barrier transporters by co-culture of brain capillary endothelial cells with astrocytes. Drug Metab Pharmacokinet 2002; 17:34–41.
5.
Cioni C, Turlizzi E, Zanelli U, Oliveri G, Annunziata P: Expression of Tight Junction and Drug Efflux Transporter Proteins in an in vitro Model of Human Blood-Brain Barrier. Front Psychiatry 2012; 3:47.
6.
Brown RC, Davis TP: Hypoxia/aglycemia alters expression of occludin and actin in brain endothelial cells. Biochem Biophys Res Commun 2005; 327:1114–1123.
7.
Al Ahmad A, Taboada CB, Gassmann M, Ogunshola OO: Astrocytes and pericytes differentially modulate blood-brain barrier characteristics during development and hypoxic insult. J J Cereb Blood Flow Metab 2011; 31:693–705.
8.
Dehouck MP, Méresse S, Delorme P, Fruchart JC, Cecchelli R: An easier, reproducible, and mass-production method to study the blood-brain barrier in vitro. J Neurochem 1990; 54:1798–1801.
9.
Ogunshola OO, Al-Ahmad A: HIF-1 at the blood-brain barrier: a mediator of permeability? High Alt Med Biol 2012; 13:153–161.
10.
Keller A: Breaking and building the wall: the biology of the blood-brain barrier in health and disease. Swiss Med Wkly 2013; 143:w13892.
11.
Fandrey J, Gassmann M: Oxygen sensing and the activation of the hypoxia inducible factor 1 (HIF-1)–invited article. Adv Exp Med Biol 2009; 648:197–206.
12.
Engelhardt S, Al-Ahmad AJ, Gassmann M, Ogunshola OO: Hypoxia selectively disrupts brain microvascular endothelial tight junction complexes through a hypoxia-inducible factor-1 (HIF-1) dependent mechanism. J Cell Physiol 2014; 229:1096–1105.
13.
Yang S-J, Pyen J, Lee I, Lee H, Kim Y, Kim T: Cobalt chloride-induced apoptosis and extracellular signal-regulated protein kinase 1/2 activation in rat C6 glioma cells. J Biochem Mol Biol 2004; 37:480–486.
14.
Wang Y, Tang Z, Xue R, Singh GK, Liu W, Lv Y, Yang L: Differential response to CoCl2-stimulated hypoxia on HIF-1α, VEGF, and MMP-2 expression in ligament cells. Mol Cell Biochem 2012; 360:235–242.
15.
Cervellati F, Cervellati C, Romani A, Cremonini E, Sticozzi C, Belmonte G, Pessina F, Valacchi G: Hypoxia induces cell damage via oxidative stress in retinal epithelial cells. Free Radic Res 2014; 48:303–312.
16.
Rodrigues SF, de Oliveira MA, dos Santos RA, Soares AG, de Cássia Tostes R, Carvalho MHC, Fortes ZB: Hydralazine reduces leukocyte migration through different mechanisms in spontaneously hypertensive and normotensive rats. Eur J Pharmacol 2008; 589:206–214.
17.
Knowles HJ, Tian Y-M, Mole DR, Harris AL: Novel mechanism of action for hydralazine: induction of hypoxia-inducible factor-1alpha, vascular endothelial growth factor, and angiogenesis by inhibition of prolyl hydroxylases. Circ Res 2004; 95:162–169.
18.
Horsman MR, Nordsmark M, Høyer M, Overgaard J: Direct evidence that hydralazine can induce hypoxia in both transplanted and spontaneous murine tumours. Br J Cancer 1995; 72:1474–1478.
19.
Chatard M, Puech C, Roche F, Perek N: Hypoxic Stress Induced by Hydralazine Leads to a Loss of Blood-Brain Barrier Integrity and an Increase in Efflux Transporter Activity. PloS One 2016; 11:e0158010.
20.
Fischer S, Wobben M, Kleinstück J, Renz D, Schaper W: Effect of astroglial cells on hypoxia-induced permeability in PBMEC cells. Am J Physiol Cell Physiol 2000; 279:C935-944.
21.
Hayashi K, Nakao S, Nakaoke R, Nakagawa S, Kitagawa N, Niwa M: Effects of hypoxia on endothelial/pericytic co-culture model of the blood-brain barrier. Regul Pept 2004; 123:77–83.
22.
Eigenmann DE, Xue G, Kim KS, Moses AV, Hamburger M, Oufir M: Comparative study of four immortalized human brain capillary endothelial cell lines, hCMEC/D3, hBMEC, TY10, and BB19, and optimization of culture conditions, for an in vitro blood-brain barrier model for drug permeability studies. Fluids Barriers CNS 2013; 10:33.
23.
Brown RC, Mark KS, Egleton RD, Huber JD, Burroughs AR, Davis TP: Protection against hypoxia-induced increase in blood-brain barrier permeability: role of tight junction proteins and NFkappaB. J Cell Sci 2003; 116:693–700.
24.
Deli MA, Abrahám CS, Kataoka Y, Niwa M: Permeability studies on in vitro blood-brain barrier models: physiology, pathology, and pharmacology. Cell Mol Neurobiol 2005; 25:59–127.
25.
Perek N, Le Jeune N, Denoyer D, Dubois F: MRP-1 protein expression and glutathione content of in vitro tumor cell lines derived from human glioma carcinoma U-87-MG do not interact with 99mTc-glucarate uptake. Cancer Biother Radiopharm 2005; 20:391–400.
26.
Perek N, Denoyer D, Dubois F, Koumanov F: Malignant gliomas display altered plasma membrane potential and pH regulation–interaction with Tc-99m-MIBI and Tc-99m-Tetrofosmin uptakes. Gen Physiol Biophys 2002; 21:381–404.
27.
Berezowski V, Landry C, Dehouck MP, Cecchelli R, Fenart L: Contribution of glial cells and pericytes to the mRNA profiles of P-glycoprotein and multidrug resistance-associated proteins in an in vitro model of the blood-brain barrier. Brain Res 2004; 1018:1–9.
28.
Wang G, Hazra TK, Mitra S, Lee HM, Englander EW: Mitochondrial DNA damage and a hypoxic response are induced by CoCl(2) in rat neuronal PC12 cells. Nucleic Acids Res 2000; 28:2135–2140.
29.
Katsuno T, Umeda K, Matsui T, Hata M, Tamura A, Itoh M, Takeuchi K, Fujimori T, Nabeshima Y, Noda T, Tsukita S, Tsukita S: Deficiency of zonula occludens-1 causes embryonic lethal phenotype associated with defected yolk sac angiogenesis and apoptosis of embryonic cells. Mol Biol Cell 2008; 19:2465–2475.
30.
Fischer S, Wobben M, Marti HH, Renz D, Schaper W: Hypoxia-induced hyperpermeability in brain microvessel endothelial cells involves VEGF-mediated changes in the expression of zonula occludens-1. Microvasc Res 2002; 63:70–80.
31.
Yeh WL, Lu DY, Lin CJ, Liou HC, Fu WM: Inhibition of hypoxia-induced increase of blood-brain barrier permeability by YC-1 through the antagonism of HIF-1alpha accumulation and VEGF expression. Mol Pharmacol 2007; 72:440–449.
32.
Saitou M, Furuse M, Sasaki H, Schulzke JD, Fromm M, Takano H, Noda T, Tsukita S: Complex phenotype of mice lacking occludin, a component of tight junction strands. Mol Biol Cell 2000; 11:4131–4142.
33.
Chen ZS, Lee K, Walther S, Raftogianis RB, Kuwano M, Zeng H, Kruh GD: Analysis of methotrexate and folate transport by multidrug resistance protein 4 (ABCC4): MRP4 is a component of the methotrexate efflux system. Cancer Res 2002; 62:3144–3150.
34.
Xiao-Dong L, Zhi-Hong Y, Hui-Wen Y: Repetitive/temporal hypoxia increased P-glycoprotein expression in cultured rat brain microvascular endothelial cells in vitro. Neurosci Lett 2008; 432:184–187.
35.
Naik P, Cucullo L: In vitro blood-brain barrier models: current and perspective technologies. J Pharm Sci 2012; 101:1337–1354.

N. Perek and F. Roche contributed equally to this work.

Open Access License / Drug Dosage / Disclaimer
This article is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License (CC BY-NC-ND). Usage and distribution for commercial purposes as well as any distribution of modified material requires written permission. Drug Dosage: The authors and the publisher have exerted every effort to ensure that drug selection and dosage set forth in this text are in accord with current recommendations and practice at the time of publication. However, in view of ongoing research, changes in government regulations, and the constant flow of information relating to drug therapy and drug reactions, the reader is urged to check the package insert for each drug for any changes in indications and dosage and for added warnings and precautions. This is particularly important when the recommended agent is a new and/or infrequently employed drug. Disclaimer: The statements, opinions and data contained in this publication are solely those of the individual authors and contributors and not of the publishers and the editor(s). The appearance of advertisements or/and product references in the publication is not a warranty, endorsement, or approval of the products or services advertised or of their effectiveness, quality or safety. The publisher and the editor(s) disclaim responsibility for any injury to persons or property resulting from any ideas, methods, instructions or products referred to in the content or advertisements.