DEVICE THAT CAN SERVE AS A HEMATO-ENCEPHALITIC BARRIER MODEL

Abstract

The present invention relates to a device which can serve as a model of a hemato-encephalic barrier (HEB) comprising two compartments in which certain cell types are arranged. The invention also relates to the method for preparing said device and the use thereof as a model of the HEB.

Claims

1. A device comprising two compartments that are separated by a porous synthetic membrane, a luminal compartment comprising endothelial cells and pericytes, and an abluminal compartment comprising astrocytes.

2. A device according to claim 1, wherein the compartment comprising astrocytes additionally comprises microglia.

3. A device according to claim 1, wherein all of the cells originate from the same animal species.

4. A device according to claim 1, wherein one or more cell types are derived from primary cultures.

5. A device according to claim 1, wherein one or more cell types are derived from cultures of immortal cells.

6. A device according to claim 4, wherein the endothelial cells and pericytes are derived from primary cultures from an individual of the same age.

7. A device according to claim 4, wherein one or more of the said cell type(s) derived from primary cultures are disease model cell types.

8. A method for preparing a device according to claim 1, comprising the following steps: a) seeding of astrocytes on one surface and/or on one side of the porous synthetic membrane; b) inserting or depositing of the porous synthetic membrane on the said surface, in a manner such that if the porous synthetic membrane comprises astrocytes, then the side comprising the astrocytes is positioned to be facing the surface; c) seeding of the luminal side of the porous synthetic membrane with endothelial cells and pericytes.

9. A method for preparing a device according to claim 1, comprising the following steps: a) seeding of astrocytes on one side of the porous synthetic membrane; b) seeding of the other side of the porous synthetic membrane with endothelial cells and pericytes, and optionally; c) depositing of the porous synthetic membrane on one surface in a manner such that the side comprising astrocytes is positioned to be facing the surface.

10. Use of the device according to claim 1 as a model of the hemato-encephalic barrier (HEB).

11. The use according to claim 10 for testing the permeability of the model.

12. The use according to claim 10 for testing the permeability of the model to a compound comprising the following steps: a) adding of the said compound to one of the compartments of the device; b) incubation of the device; c) detection and analysis of the presence of the said compound and/or its metabolites in the compartment where the addition of the said compound has not taken place.

13. The use according to claim 10 for studying the physiopathology of a disease, or testing molecules developed with a preventive, therapeutic or diagnostic purpose that target the cellular and molecular actors of the HEB, or testing physical conditions or testing protocols.

14. A device according to claim 6, wherein one or more of the said cell type(s) derived from primary cultures are disease model cell types.

Description

FIGURES

[0077] FIG. 1: Schematic representation of the preparation of a device comprising cultures of primary cells according to the invention.

[0078] At day D1, the astrocytes and microglia are thawed and the endothelial cells and the pericytes are purified from mouse brains. The PBMCs are extracted from mouse peripheral blood.

[0079] At D3, the cell culture medium is renewed, ie replaced with new medium, with cessation of the effect of puromycin in the medium for endothelial cells.

[0080] At D5, the cell culture medium is renewed again.

[0081] At D8, the astrocytes and microglia, represented by dots “.”, are seeded in a culture dish. The culture medium for the other cells is renewed.

[0082] At D10, it is possible to observe the microglia cells.

[0083] At D11, the astrocytes and the microglia are seeded on the porous synthetic membrane.

[0084] At D12, the porous synthetic membrane is deposited in the culture dish in a manner such that the two astrocyte cultures are in contact, thus forming the abluminal compartment. Then the pericytes and the endothelial cells, represented by squares and circles (“□”, “o”) are seeded on the upper surface of the porous synthetic membrane, thus forming the luminal compartment. The treatment of the device with hydrocortisone is initiated.

[0085] At D15, the treatment of the device with hydrocortisone is completed, the PBMCs, represented by crosses “+”, are seeded into the luminal compartment. The model is ready for use.

[0086] FIG. 2: Paracellular permeability of FITC-Dextran on a device according to the invention comprising cultures of primary cells derived from mouse models of Alzheimer's Disease (AD) or wild type (WT) mice.

[0087] The test of permeability of the devices is performed with 4 kD FITC-dextran. The culture media are replaced by 1 mL of HBSS with Ca.sup.2+/Mg.sup.2+ in the abluminal compartment and 500 μL of FITC-dextran in the luminal compartment (that is to say 2.10.sup.−6 moles). Samples of 50 μL in the luminal and abluminal compartments are taken at 0 min, 10 min, 20 min, 30 min, 1 hr and 1 hr 30 min and deposited in the wells of a 96-well black plate, read on the Varioskan microplate reader (Thermo Scientific). The excitation wavelength (kex) of FITC-dextran is 485 nm and the emission wavelength (λem) is 515 nm.

[0088] FIG. 2 shows the results obtained for the abluminal compartment in the form of a curve. *p<0.05, **p<0.01 in relation to the control device which corresponds to a device without cells but coated with the same “coating” matrix as the other devices studied (AD versus WT).

[0089] The permeability of the device AD remains higher as compared to the device WT (46%), thus indicating a lower impermeability of the pathology related device as compared to the healthy device. The statistical test used is the Kruskal-Wallis test followed by the Dunn test for multiple comparisons.

[0090] FIG. 3: Paracellular permeability coefficient of FITC-dextran on a device according to the invention comprising cultures of primary cells from wild mice (WT).

[0091] The calculation of the permeability coefficient Pe is done by using this formula:

Pe=dQ/(dT*A*Co)

[0092] Pe: coefficient of permeability (cm/s)

[0093] dQ: quantity transported (mol)

[0094] dT: time of incubation (second)

[0095] A: surface area of the porous synthetic membrane (here 1.12 cm.sup.2)

[0096] Co: initial concentration (4.10.sup.−6 mol/cm.sup.3)

[0097] The determination of dQ is carried out based on the following calculation:

[0098] (Abluminal Fluorescence Intensity at 1 h)*2.10.sup.−6/(Luminal Fluorescence Intensity at T0)

[0099] The fluorescence intensity is proportional to the amount of FITC-dextran present in each compartment (abluminal and luminal).

[0100] The permeability coefficient is shown in FIG. 3 and was calculated after one hour. The results represent the mean±SEM (mean standard deviation) of the permeability coefficient of 3 to 4 devices in each group.*p<0.01 in relation to the control device which corresponds to a device without cells but covered with the same “coating” matrix as the device WT. The statistical test used is the Mann Whitney test.

[0101] FIG. 4: Functionality of the P-glycoprotein in the device according to the invention.

[0102] This functionality test is carried out with rhodamine 123, because it is known that this molecule as a substrate for the P-glycoprotein is expelled by the P-glycoprotein into the luminal compartment, which thus limits its passage through the device. The culture media are replenished: 1 mL in the abluminal compartment and 250 μL in the luminal compartment either containing or not containing the Zosuquidar inhibitor of P-glycoprotein P at 5 μM (Dantzig et al. 1996). The devices are incubated for 2 hrs in the incubator. Then, 250 μL of medium containing 2 μM rhodamine 123 with or without 5 μM Zosuquidar are added into the luminal compartment and incubated for 1 hr in the incubator. Samples of 50 μL in the luminal and abluminal compartments are taken just after the addition of the medium containing rhodamine 123 and after 1 hour of incubation at 37° C. The samples are placed in the wells of a 96-well black plate. The reading of the fluorescence intensity is done on the same apparatus as for the permeability test (described above). The λex of the rhodamine is 500 nm and the λem is 524 nm.

[0103] In FIG. 4, the results represent the mean±SEM of the fluorescence intensity after 1 hr of incubation of the cells with Rhodamine 123 (n=4 to 8 in each group). The passage of the rhodamine 123 through the device WT decreases by 72.6% as compared to the device without cells (**p<0.01). The presence of the specific P-glycoprotein inhibitor inhibits the efflux by 57.3%. The statistical test used is the Kruskal-Wallis test followed by the Dunn test for multiple comparisons.

[0104] FIG. 5: Trans-endothelial electrical resistance (TEER) in the murine devices according to the invention, having 12 and 24 wells.

[0105] In FIG. 5, the results represent the mean±SEM. For the 12-well format (FIG. 5A), n=16 controls and n=18 devices having the 12-well format. For the 24-well format (FIG. 5B), n=12 controls and n=74 devices having the 24-well format. The statistical test used is the Mann Whitney test: ****p<0.0001.

[0106] FIG. 6: Permeability coefficient of FITC-Dextran on murine devices according to the invention in 24- and 96-well format.

[0107] The coefficient of permeability is represented in FIG. 6 and in the same manner as in FIG. 3 after one hour. In FIG. 6, the results represent the mean±SEM. For the 24-well format (FIG. 6 A), n=12 controls and n=50 devices. For the 96-well format (FIG. 6B), n=8 controls and n=54 devices. The statistical test used is the Mann Whitney test: ****p<0.0001.

[0108] FIG. 7: Functionality of the P-glycoprotein in the murine device according to the invention in 24- and 96-well format.

[0109] In FIG. 7, the results represent the mean±SEM. For the 24-well format (FIG. 7A), n=12 controls and n=9 devices in the 24-well format without Zosuquidar and n=10 devices in the 24-well format with Zosuquidar. For the 96-well format (FIG. 7 B), n=7 controls and n=6 devices in the 96-well format without Zosuquidar and 9 devices in the 96-well format with Zosuquidar. The statistical test used is the Kruskal-Wallis test followed by a Dunn test for multiple comparisons: *p<0.05, **p<0.01, ***p<0.001.

[0110] FIG. 8: Coefficient of permeability of 8 molecules on murine devices according to the invention in 24 and 96-well format.

[0111] The molecules were added into the luminal compartment and incubated for a period of 48 hours. The luminal and abluminal media were then collected for the assay of the molecules. In FIG. 8 the different permeability coefficients calculated are represented (the results represent the mean±SEM), the results obtained for the 24-well format are shown in FIG. 8A, and the results obtained for the 96-well format in FIG. 8B.

[0112] FIG. 9: Trans-endothelial electrical resistance (TEER) and permeability coefficient of FITC-DEXTRAN in the commercially available device.

[0113] Represented in FIG. 9A is the Trans-endothelial Electrical Resistance (TEER) (n=14 controls and n=36 24-well devices). Represented in FIG. 9B is the permeability coefficient of FITC-DEXTRAN (n=8 controls and n=23 24-well devices). The results represent the mean±SEM. The statistical test used is the Mann Whitney test: ****p<0.0001.

[0114] FIG. 10: Functionality of P-glycoprotein in the commercially available device.

[0115] In FIG. 10, the results represent the mean±SEM of the fluorescence intensity after 1 hour of incubation of the cells with Rhodamine 123 (n=4 to 8 in each group). n=3 controls and n=4 commercially available 24-well devices with or without Zosuquidar. The statistical test used is the Kruskal-Wallis test followed by a Dunn test for multiple comparisons: *p<0.05.

[0116] FIG. 11: Trans-endothelial electrical resistance (TEER) in 24-well formats after cryopreservation.

[0117] Represented in FIG. 11 is the trans-endothelial electrical resistance (TEER) (n=3-4 controls and n=2-15 devices in the 24-well format). The 24-well format devices were cryopreserved for 7, 15 or 30 days, and then used 4, 5 or 6 days post-thawing. The results represent the mean±SEM. The statistical test used is the Mann Whitney test: *p<0.05, **p<0.01.

[0118] FIG. 12: Permeability coefficient of FITC-DEXTRAN in 24-well formats after cryopreservation.

[0119] In FIG. 8 the different permeability coefficients calculated are shown (n=8 controls and n=2-5 devices in the 24-well format). The results represent the mean±SEM). The devices in 24-well format were cryopreserved for 7, 15 or 30 days, and then used 4, 5 or 6 days post-thawing. The statistical test used is the Mann Whitney test: *p<0.05, **p<0.01.

[0120] FIG. 13: Trans-endothelial electrical resistance (TEER) in murine devices with immortalised cell lines according to the invention in 12- and 24-well format.

[0121] In FIG. 13, the results represent the mean±SEM. For the 12-well format (FIG. 13A), n=20 controls and n=20 devices in the 12-well format. For the 24-well format (FIG. 13B), n=10 controls and n=97 devices in the 24-well format. The statistical test used is the Mann Whitney test: ****p<0.0001.

[0122] FIG. 14: Permeability coefficient of FITC-dextran on murine devices with immortalised cell lines according to the invention in 12-, 24- and 96-well formats.

[0123] In FIG. 14, the results represent the mean±SEM. For the 12-well format (FIG. 14A), n=10 controls and n=6 devices in the 12-well format. For the 24-well format (FIG. 14B), n=17 controls and n=92 devices in the 24-well format. For the 96-well format (FIG. 14C), n=12 controls and n=63 devices in the 96-well format. The statistical test used is the Mann Whitney test: **p<0.01, ****p<0.0001.

[0124] FIG. 15: Functionality of P-glycoprotein on murine devices with immortalised cell lines according to the invention in 12-, 24- and 96-well format.

[0125] In FIG. 15, the results represent the mean±SEM. For the 12-well format (FIG. 15 A), n=4 controls and n=2 devices in the 12-well format with or without Zosuquidar. For the 24-well format (FIG. 15B), n=7 controls and n=9 and 12 devices in the 24-well format with or without Zosuquidar. For the 96-well format (FIG. 15C), n=12 controls and n=13 and 14 devices in the 96-well format with or without Zosuquidar.

[0126] The statistical test used is the Kruskal-Wallis test followed by a Dunn test for multiple comparisons: *p<0.05, **p<0.01, ***p<0.001.

EXAMPLE 1: PREPARATION OF A DEVICE ACCORDING TO THE INVENTION

[0127] This device comprises primary cultures of endothelial cells, pericytes, and PBMCs harvested from mouse brains.

[0128] The development and preparation of this device involved going through the following technical steps:

[0129] The endothelial cells and pericytes were purified by using magnetic beads in order to exclude myelin and a Percoll gradient to dissociate the endothelial cells from the pericytes.

[0130] The PBMCs were extracted from the peripheral blood of the same mice on a Ficoll gradient identical to that used in human medical haematology and then recovered by centrifugation.

[0131] A primary co-culture stock of astrocytes and microglia was created as follows.

[0132] Primary mouse astrocyte and microglia cultures were obtained from brains of newborn mice between days D1 and D3. Cell dissociation of the brain tissue was performed mechanically. The astrocytes and microglia were selected by means of a selective culture medium. One week after the seeding of a newborn brain dissociated and cultured in a 75 cm2 Vial coated with Poly-L-Lysine, astrocytes forming a confluent mat were detached by using trypsin and cryopreserved. The thawing of a cone of cells was performed in a 25 cm.sup.2 Vial, thus making it possible for the astrocytes to be used for the assembling of the device 72 hrs thereafter. These cells may be subcultured 3 times.

[0133] The cells were cultured in selective media for each cell type until a cell mat covering the entire surface of the selected culture dish was obtained.

[0134] The astrocytes and the microglia were seeded in a culture dish (30,000 cells per well for a 12-well dish).

[0135] Then the astrocytes and microglia were also seeded under a porous synthetic membrane and incubated for 24 hours (see FIG. 1).

[0136] The membrane was deposited in the culture dish containing the astrocytes and the microglia in a manner such that the two astrocyte cultures were in contact, thereby forming the abluminal compartment. These cells thus model the glial cerebral parenchyma.

[0137] The endothelial cells and the pericytes were seeded onto the membrane (endothelial cells=10.sup.6 cells/membrane, and pericytes=350,000 cells/membrane in a 12-well dish).

[0138] The device was incubated for 72 hours in the presence of hydrocortisone so as to promote P-glycoprotein expression in the endothelial cells. In fact P-glycoprotein serves as an important efflux pump with respect to the functionality of the HEB.

[0139] After incubation, the model is ready. In particular, it is able to receive PBMCs.

EXAMPLE 2: OBSERVATION BY MEANS OF IMMUNOLABELLING OF DEVICES MADE ACCORDING TO EXAMPLE 1

[0140] After the permeability and functionality tests, the porous synthetic membranes of the devices developed according to Example 1 are washed twice for a period of 5 min with 500 μL of Phosphate Buffered Saline (PBS). This is followed by addition of 500 μL of a paraformaldehyde solution (4% PFA) to the abluminal and luminal compartments for 15 min at ambient temperature. Two further 5-minute washes are carried out with 500 μL of PBS. The cells are then blocked and permeabilised with 500 μL of PBS/Triton 0.5%/Bovine Serum Albumin (BSA) 5% in the abluminal and luminal compartments for 1 hour at ambient temperature. On paraffin plastic film (parafilm) stretched over a petri dish, 30 μL of primary antibodies (Ad) are deposited. The membranes are then cut and subsequently deposited in a manner such that the cells are in contact with the primary antibodies (Ad). The antibodies used are anti-Zonula Occludens Protein 1 (ZO-1) antibodies (1/50 dilution, marker for tight junction proteins, used as marker for endothelial cells), anti-Alpha Smooth Muscle Actin (αSMA) (1/50 dilution, marker for pericytes), anti-von Willebrand Factor (vWF) (1/50 dilution, marker for endothelial cells) and anti-Glial Fibrillary Acidic Protein (GFAP) (1/100 dilution, marker for astrocytes). Incubation occurs over a period of one night at 4° C. in a humidity chamber. The next day, each membrane is gently placed in a 12-well dish with the side having the cells being studied on top and subjected to two 5 min washes with PBS with no agitation. 30 μL of Arthrobacter aurescens chondroitinase AC-II are deposited on a new paraffin plastic film before being incubated for 1 hour at ambient temperature with the membranes. The Ac II solution contains anti-mouse II secondary antibodies coupled to the fluorochrome Rhodamine Red-X (RRX) (red fluorescence) and anti-rabbit II secondary antibodies coupled to the fluorochrome Alexa 488 (green fluorescence) all diluted to 1:50. After 1 hr of incubation, the washes are done under the same conditions as above and then the membranes are incubated with 30 μL of DAPI solution (4,6-Diamino-2-Phenylindole) for 15 min at ambient temperature, protected from light, on paraffin plastic film in a wet chamber in order to mark the nuclei of the cells. The membranes are again recovered in order to undergo three 5 min washes with H2O UHQ to remove the salts. The membranes are then glued on a glass slide with DAPI glue in a manner such that the bonded side corresponds to that of the non-immunolabelled cells. A glass slide is then glued onto the membrane. The slides are observed under the epifluorescence microscope (Olympus BX 51).

[0141] The immunolabelling renders visible the cell types that make up the HEB model. In addition, it was observed that endothelial cells organise themselves into vessels with tight junctions being formed therebetween (ZO-1 tagging), thus spontaneously reproducing important characteristics of the HEB in vivo. The pericytes organise around the endothelial cells by establishing points of contact.

EXAMPLE 3: USE OF A DEVICE ACCORDING TO THE INVENTION AS A MODEL OF THE HEB IN THE CASE OF CELLS DERIVED FROM MOUSE MODELS OF ALZHEIMER'S DISEASE (AD)

[0142] A device was prepared according to Example 1, here referred to as the Alzheimer device (device AD), in which the endothelial cells and pericytes were prepared from 4 to 8 week old Alzheimer mice (APPswePS1dE9, AD) and the astrocytes and microglia were prepared from wild mice (WT). This device was compared to a similar device formed completely from cells from wild mice (WT), here referred to as the wild device, and to a similar cell-free device, here referred to as the control device.

The results are shown in FIG. 2.

[0143] The presence of cells serves to decrease the passage of FITC-dextran (FIG. 2). In fact, it was observed that there was a significant 82% decrease in the passage of FITC-dextran through the wild device at 1 hour, and then a 78% decrease at 1.5 hours as compared to the control device without cells.

[0144] This difference in permeability is also observed with the device AD where there is a 60% decrease obtained at 1 hour and 1.5 hours as compared to the control. However, these results are not statistically significant.

[0145] In addition, although the observed difference in permeability between the AD and wild-type devices was not statistically significant, there was a 47% decrease observed in the passage of FITC-dextran through the wild-type device as compared to the device AD at 1.5 hours.

EXAMPLE 4: TESTING THE PERMEABILITY AND FUNCTIONALITY OF A DEVICE ACCORDING TO THE INVENTION

[0146] In order to test the permeability of the device according to the invention, a device was prepared according to Example 1, and FITC-dextran was added to the luminal compartment of the device. The presence of FITC-dextran was detected and analysed by means of fluorescence.

[0147] In FIG. 3 it is noted that the coefficient of permeability is higher with the control device. A significant decrease in the permeability coefficient of 98% is observed for the device WT. This difference is statistically significant. It demonstrates the relative permeability of the device according to the invention.

[0148] In order to test the functionality of the device according to the invention, a device was prepared according to Example 1, and rhodamine 123 was added to the luminal compartment of the said device in the presence or absence of Zosuquidar. Indeed it is known that rhodamine 123 as a substrate for glycoprotein P is expelled by the glycoprotein P into the luminal compartment, which limits the passage thereof through the device. Zosuquidar is an inhibitor of P-glycoprotein, therefore its use should limit the efflux of rhodamine 123 into the luminal compartment.

[0149] In FIG. 4, it is observed that the passage of Rhodamine 123 to the abluminal compartment decreases by 72.6% for the wild device as compared to the control device (cell-free insert). These results are statistically significant. The presence of the specific inhibitor of glycoprotein P inhibits the efflux by 57.3%. These results demonstrate the functionality of the device according to the invention.

EXAMPLE 5: VALIDATION OF A DEVICE ACCORDING TO THE INVENTION IN 24 AND 96 WELL PLATE FORMAT

[0150] The device preparation method based on these two new HEB formats is identical to that described in Example 1. Only the densities of each cell type had to be adapted.

TABLE-US-00001 TABLE 1 Cell densities of devices in 12, 24 and 96-well formats Cell densities (cells/wells) Astrocytes at Astrocytes Endothelial Format bottom of well below insert cells Pericytes 12 wells 30,000 74,000 1,000,000 350,000 24 wells 15,000 21,765 350,000 117,000 96 wells 2,500 9,450 250,000 85,000

[0151] The relevance of these formats was validated by using four tests: trans-endothelial electrical resistance (TEER), paracellular permeability, Glycoprotein G (P-gp) functionality and selectivity with respect to 8 molecules.

[0152] 1—Trans-Endothelial Electrical Resistance (TEER)

[0153] The Trans-endothelial Electrical Resistance was only measured on the 12 and 24-well format HEBs because the electrode is not suitable for the 96-well format. It is expressed in ohm.Math.cm.sup.2 taking into account the surface area of the insert: 1.12 and 0.33 cm.sup.2 for the 12 and 24-well inserts, respectively. It was measured with an ohm meter (Millicell Electrical Resistance System-2, Millipore—[Molsheim] France) using two STX01 electrodes: the larger one is placed in the abluminal compartment and the smaller one in the luminal compartment. The system carrying the two electrodes is connected to the ohm meter to measure the electrical resistance between the two compartments. The value displayed on the device is expressed in ohm and then multiplied by the surface area of the insert to obtain the results in ohm.Math.cm.sup.2.

[0154] FIG. 5 shows that the TEER is significantly increased in devices according to the invention in relation to the control (that is to say a device without any cells).

[0155] 2—Paracellular Permeability

[0156] As in Example 4, the passage of FITC-dextran was studied and the permeability coefficient (Pe) of this molecule as a control for paracellular permeability was calculated (FIG. 6).

[0157] The results obtained for both the 24-well and 96-well formats show a permeability coefficient of less than 4.10.sup.−6 cm/s as for the 12-well format (see FIG. 3).

[0158] 3—Functionality of the P-Glycoprotein (P-Gp)

[0159] As in Example 4, the passage of rhodamine 123 in the presence or absence of Zosuquidar was studied.

[0160] The P-gp pumps are functional in both 24- and 96-well format devices because Rhodamine123 is effluxed significantly towards the luminal side and therefore passes very little on to the abluminal side. This functionality is indeed inhibited by Zosuquidar known as a specific inhibitor of P-gp. This shows that as with the 12-well format, the endothelial cells expressing P-gp are indeed polarised, the addition of Rhodamine123 into the abluminal compartment results in passage thereof to the luminal side that is comparable to cell-free HEBs.

[0161] The results are shown in FIG. 7.

[0162] 4—Selectivity with Respect to 8 Molecules

[0163] On the 24 and 96-well devices, the passage of 8 molecules known to either be able or unable to pass through the HEB was studied.

[0164] Dopamine (DA), Levodopa (L-DOPA), Bromazepam (BROMO), Caffeine (CAF), Sucrose (SUC), Cyclosporin A (CYCLA), Zosuquidar (ZOSU), and Mitotane (MITO) were tested at a physiological and/or therapeutic concentration known in the literature in humans.

[0165] Indicated in Table 2 here below are the concentrations chosen for each molecule, as well as whether they are known to be able to pass through the HEB (+) or unable to pass through the HEB (−).

TABLE-US-00002 TABLE 2 Concentration of molecules used in the test and ability to pass through the HEB Name of Molecule CAF SUC BROMO L-DOPA DA CYCLA ZOSU MITO Passage + + + + − − − − through the HEB Concentration 40 2000 1 160 50 70 0.3 20 (μg/mL) in the insert

[0166] The molecules were added to the luminal compartment at the indicated concentration and incubated for a period of 48 hours. The luminal and abluminal media were then removed for the assay of the molecules.

[0167] For each molecule, the coefficient of permeability was calculated. The results are shown in FIG. 8. The results validate the passage of these molecules through the devices of the invention which serve here as a model of the murine HEB, that is to say the first 4 molecules pass through whereas the last 4 do not (hence a very low coefficient of permeability for the last 4 molecules).

EXAMPLE 6: COMPARISON OF THE DEVICE ACCORDING TO THE INVENTION WITH A COMMERCIALLY AVAILABLE DEVICE

[0168] The device in Example 1 in 24-well format was compared with a commercially available model (BBB Kit™ (RBT-24H) from Pharmaco-Cell®) which corresponds to a primary HEB model prepared from rat brain cells. In this commercially available device the endothelial cells are seeded on the insert, the pericytes under the insert, and the rat astrocytes at the bottom of the well.

[0169] The measurement of the TEER in this model shows a good transendothelial electrical resistance averaging 247±17.76 Ω.Math.cm.sup.2 (see FIG. 9 A). However, the permeability coefficient of FITC-dextran is higher than that of the device according to the invention (see FIG. 9 B) (25.850±2.308×10.sup.−6 cm/s versus the device of the invention 3.867±0.333×10.sup.−6 cm/s). The impermeability would therefore be better on the device of the invention by a factor of 6.7 as compared to that of the commercially available model.

[0170] Here again, the functionality of the commercially available model was evaluated by studying the passage of Rhodamine 123 in the presence or absence of the specific inhibitor of the P-gp glycoprotein, Zosuquidar.

[0171] The results show that the pump is present on both the luminal and abluminal sides, thus the cells are not properly polarised. Thus there is as much Rhodamine123 effluxed on the luminal side as on the abluminal side when it is deposited either in the luminal or in the abluminal compartment (97% efflux regardless of the side where it is deposited).

[0172] Thus, no effect is observed when Zosuquidar inhibits the P-gp pump in the commercially available model as shown in FIG. 10 contrary to the device according to the invention (see FIG. 7). Thus, the device according to the invention makes it possible to obtain a model with greater similarity to the HEB than the commercially available model that was tested here.

EXAMPLE 7: CRYOPRESERVATION OF THE DEVICE ACCORDING TO THE INVENTION IN 24-WELL FORMAT

[0173] Cryopreservation of the device would provide the means for on-demand preparation and delivery of the frozen device to the client.

[0174] The device in 24-well format was cryopreserved with a CRYOSTOR® solution marketed at Sigma® (REF: C2874-100ML). A measurement of the TEER was performed before the cryopreservation at day D15 of the assembly of the device. CRYOSTOR volumes of 100 and 200 were selected for the cryopreservation of cells in the luminal and abluminal compartments, respectively.

[0175] Thawing was performed at days D7, D15, and D30 post-cryopreservation according to the following protocol and the impermeability of the thawed devices was investigated 4, 5, and 6 days after thawing (TEER and permeability coefficient of FITC-dextran). [0176] On the day of thawing, preheat the complete Endogro™ culture medium with serum to 37° C., [0177] During the entire thawing process, do not touch the membrane of the insert with the pasteur pipettes or tips. Do not move the inserts. Handle and treat on a device by device basis, [0178] Once the culture medium has been heated, immediately add 150 and 300 μL of Endogro™ complete culture medium with serum, on the luminal and abluminal sides, respectively, [0179] Incubate for 3 hours at 37° C. under 5% CO.sub.2. [0180] After 3 hours, gently aspirate the culture media (CM) from the luminal and abluminal sides and add 200 and 500 μL of CM from the luminal and abluminal sides, respectively, [0181] Incubate at 37° C. under 5% CO.sub.2, [0182] The day after thawing, replace the culture medium, [0183] At days D4, D5 and D6, tests were carried out to measure the TEER and the coefficient of paracellular permeability. The results are shown in FIGS. 11 and 12 respectively.

[0184] Compared to the non-cryopreserved 24-well format device, it is observed that the TEER resistance after 7 days of freezing is of the same order of magnitude as that measured just before the freezing. For 15 and 30 days of freezing, an average decrease of 26% is observed but remains insignificant.

[0185] As for the paracellular permeability, it is found that the permeability coefficients are also of the same order of magnitude after 7 days of freezing as those obtained on the non-cryopreserved HEBs (3.278±0.925 versus 3.867±0.333 10.sup.−6 cm/s, respectively). For 15 and 30 days of freezing, the coefficient of permeability is between 6 and 20.10.sup.−6 cm/s.

EXAMPLE 8: MURINE HEB DEVICE WITH IMMORTALISED CELL LINES

[0186] Cell lines of murine endothelial cells and murine pericytes after immortalisation of the primary cells were obtained by using a method already published in the literature (Burek et al. 2012).

[0187] The immortal nature of these lines was verified by performing a karyotype showing a change in the number of chromosomes. Under normal circumstances and conditions, in mice, 2n=40 chromosomes. In the immortal lines, the reading of 16 metaphasic plates shows several anomalies in the number of chromosomes validating immortalisation with 2n=39 to 77 chromosomes depending on the plates read.

[0188] These lines have a replication time of 48 hours. These lines may be cryopreserved. The culture media used are the same as those used for the primary cultures. Their immortal nature leads to very high cell adhesion.

[0189] The assembly of the device with these cells follows the following kinetics: [0190] D1 seeding of astrocytes/microglia I under the insert and at the bottom of the well [0191] D2 seeding of endothelial cells and pericytes on the insert [0192] D3-D4 incubation of the model for 48 hours in the incubator [0193] At D4 the device is ready for any experimentation.

[0194] Indicated in Table 3 below are the cell densities used for each cell type and for each format of the device (12-, 24-, and 96 wells).

TABLE-US-00003 TABLE 3 Cell densities of devices in 12-, 24-, and 96-well formats Cell densities (cells/wells) Astrocytes at Astrocytes Endothelial cell Pericytes cell Format well bottom under insert line line 12 wells 95,000 149,333 170,000 57,000 24 wells 47,500 44,000 50,000 17,000 96 wells 8,000 3,575 22,000 7,500

[0195] On the 12 and 24 well plate formats, the TEER was measured (see FIG. 13) and on all formats the paracellular permeability (see FIG. 14) and the functionality of the P-gp efflux pump (see FIG. 15) were evaluated.

[0196] For this HEB model, the co-immunolabelling was performed to render visible the expression of molecular markers specific to each cell type. The inventors thus observed that the immortalised endothelial cells express the following in the device: [0197] P-glycoprotein, [0198] the tight junction proteins: ZO-1, Claudine-5, [0199] the vonWillebrand factor (vWF), [0200] the LRP-1 receptor, and [0201] the transferrin receptor.

[0202] They do not express the pericyte markers α-SMA, NG2, platelet-derived growth factor β receptor (PDGFβR), indicating that the culture is pure. On the contrary, pericytes indeed express these 3 markers, and LRP-1, and do not express GFAP and vWF, indicating a pure culture of pericytes uncontaminated by endothelial cells and astrocytes.

[0203] Impermeability of the HEBs

[0204] FIG. 13 shows the results of the TEER for the 12- and 24-well plate formats.

[0205] The TEER on the 12-well format is quite comparable to that obtained on the device with primary cultures (see FIG. 5: 12-well TEER mean 218±7.17 Ω.Math.cm.sup.2 and here for the model with immortal cell lines the mean TEER is 184.60±6.65 Ω.Math.cm.sup.2). For the 24-well format, the mean is 63.30±1.35 Ω.Math.cm.sup.2 while the HEB model with primary cultures had a mean TEER of 125.50±2.34 cm.sup.2.

[0206] For the paracellular permeability, the permeability coefficient values for FITC-dextran are shown in FIG. 14 for each of the 3 formats (12-, 24-, and 96 wells, FIG. 14 A, B and C respectively). Regardless of the format used with the immortalised endothelial cell lines and pericytes in place of primary cultures, the results show that the devices are impermeable and the mean permeability coefficient is 12.15±0.92, 10.19±0.44, 9.66±0.50.10.sup.−6 cm/s for the 12-, 24-, and 96-well formats, respectively.

[0207] Functionality of the HEBs

[0208] The 12-, 24-, and 96 well plate formats are functional as they significantly limit the passage of Rhodamine.sup.123 on the abluminal side by more than 60% (see FIG. 15). Contrary to the HEBs prepared with primary cultures, the specific Glycoprotein P inhibitor does not show any effect on all the formats, this may be related to a different efflux pump of the “Multi drug resistance” type very often encountered in cell lines.

REFERENCE

[0209] Burek M, Salvador E, Förster C Y. Generation of an immortalised murine brain microvascular endothelial cell line as an in vitro blood brain barrier model. J Vis Exp. 2012 Aug. 29; (66):e4022. doi: 10.3791/4022 [0210] Dantzig A H, Shepard R L, Cao J, Law K L, Ehlhardt W J, Baughman T M, Bumol T F, Starling J J. Reversal of P-glycoprotein-mediated multidrug resistance by a potent cyclopropyldibenzosuberane modulator, LY335979. Cancer Res. 1996 Sep. 15; 56(18):4171-9.