A TRIPLE CO-CULTURE MODEL OF THE BLOOD-BRAIN BARRIER USING PRIMARY PORCINE BRAIN ENDOTHELIAL CELLS, PORCINE PERICYTES AND PORCINE ASTROCYTES

Technical field of the invention

The present invention relates to a blood brain barrier (BBB) model. In particular, the present invention relates to a BBB model comprising three different porcine cell types. Background of the invention

Brain endothelial cells (BECs) denote the blood-brain barrier (BBB) and form a major physical restraint on the transport into the brain of several molecules present in blood plasma for transport. The BECs are non-fenestrated, rich in mitochondria, high in concentrations of drug- and nutrient metabolizing enzymes, but low in vesicles involved in endocytotic and transcytotic activity. The BECs are also closely connected with intermingling tight junctions and adherence junctions. Pericytes and end-feet of astrocytes form close contact with the BECs and participate in the formation, regulation and maintenance of the integrity of the BBB.

Modelling the BBB has been an important issue for decades. Experimental conditions are often more controllable in vitro than in vivo, and they are overall also more ethically acceptable due to the reduced number of animals applied per study when performed in vitro. Although the BBB formed in vitro lacks the full complexity of the BBB in vivo, many parameters of the in vivo conditions can still be determined, e.g. tight junction expression and luminal to abluminal transport of large molecules. Both primary and immortalized cells are being used for in vitro studies of the BBB. Primary BECs have been isolated and cultured from most mammals with the foremost originating from rats, mice, pigs, cows, and even humans.

The mechanisms that induce polarization of the BECs are not fully understood, but astrocytes are known to secrete a number of substances that participates in the induction of the BBB, e.g. basic fibroblast growth factor (bFGF) and angiopoietin 1 (ANG1). Primary BECs are often isolated from small mammals.

In the past, PBEC's have mainly been cultured in either monoculture or in co- culture with primary rat astrocytes isolated from neonatal rats or rat astrocytes cell lines e.g. C6 glioma. Such co-cultures were therefore, constructed from two different species, which could influence the barrier function and gene expression of the PBEC's. Furthermore, rat astrocytes are most often derived from rat pups due to their ability to grow faster than astrocytes obtained from older animals. Deriving cells from laboratory animals is very expensive and requires large amounts of animal sacrifices only for this single purpose.

Hence, an improved BBB model would be advantageous, and in particular, a more efficient and/or reliable BBB model derived entirely from porcine cells would be advantageous.

Summary of the invention

An in vitro BBB model based on porcine brain endothelial cells (PBEC's) has several advantages, when compared to those of in vitro rodent BBB models: i) higher cellular yield per animal, ii) PBEC's retain many of the important BBB features, iii) human and porcine genome, anatomy, physiology, and disease progression are more comparable, iv) porcine brains are by-products from the abattoir and, therefore, inexpensive and their usage for research more ethically acceptable. In this study, PBEC's, astrocytes and pericytes were isolated from 6 months old domestic pig brains donated and considered a waste product by an abattoir. The aim of the present study was to establish a triple co-culture based entirely on porcine cells i.e. PBEC's, astrocytes and pericytes and to determine if this preferable cellular combination for BBB formation would compare to PBEC's co- cultured with rat astrocytes and pericytes isolated from new-born rat pups.

Porcine or rat astrocytes and pericytes were cultured in both contact and non- contact co-culture with PBEC's to examine their effects on the PBEC's for barrier formation as revealed by formation of trans-endothelial electric resistance (TEER), loss of passive permeability, and expression patterns of BEC specific proteins. The results show that primary porcine astrocytes and pericytes are useable for triple co-culture with PBEC's instead of primary rat astrocytes and pericytes, as equally high TEER values, low passive permeability and expression of hallmarks of BECs were observed. Because astrocytes may need culturing for two-three weeks before being co-cultured in most in vitro BBB models, they are often isolated from a different animal than the PBEC's and pericytes, which may impair the overall function of the model.

Thus, an object of the present invention relates to the provision of a blood-brain barrier model constituted entirely of porcine cells. In particular, it is an object of the present invention to provide a BBB model that solves the above-mentioned problems of the prior art with sacrificing animals for the single purpose of creating such models. Thus, one aspect of the invention relates to an isolated triple co-culture

comprising porcine primary brain endothelial cells (BECs), porcine astrocytes and porcine pericytes. In a preferred embodiment, the isolated triple co-culture according to any of the preceding claims, comprises (or is provided in)

- a culture plate (2), comprising the culture of porcine astrocytes or porcine pericytes;

- a porous culture membrane (3), comprising the culture of BECs (5) on one side and a culture of the porcine pericytes or the porcine astrocytes on the opposite side; with the proviso that the triple co-culture comprises porcine astrocytes and porcine pericytes.

Another aspect of the present invention relates to the use of the isolated triple co- culture according to the invention as an in vitro model of a mammalian blood- brain barrier.

Yet another aspect of the present invention is to provide a method of

measuring/evaluating blood-brain barrier permeability of a substance, the method comprising

a) providing an isolated triple co-culture according to the present invention; b) incubating said substance at one side of a porous culture membrane; and c) measuring the amount of said substance permeating across the porous culture membrane or binding to the BEC's. Still another aspect of the present invention is to provide a method for producing an isolated triple co-culture according to the present invention, the method comprising

a) providing porcine brain cells;

b) proliferating from said porcine brain cells viable BEC's, viable porcine

astrocytes and viable porcine pericytes; and

c) growing said BEC's, porcine astrocytes and porcine pericytes into a triple- co-culture.

A further aspect of the present invention is to provide a method for

determining/evaluating the toxicity of a substance toward the blood brain barrier, the method comprising incubating said substance with the triple cell co-culture according to the invention and subsequently assessing the viability of one or more of the cell types in the triple cell co-culture. A yet further aspect of the present invention is to provide an isolated triple co- culture comprising porcine primary brain endothelial cells (BEC's), porcine astrocytes and porcine pericytes obtained/obtainable by a process according to the method for producing an isolated triple co-culture comprising porcine primary brain endothelial cells (BEC's), porcine astrocytes and porcine pericytes.

The present invention serves to provide technical progress in several important areas compared to prior art. A porcine triple co-culture blood brain barrier model is a more anatomical correct model than competing technologies based on cultures of lower mammals. Consequently, it is to be expected that the present invention will yield results that are more reliable than competing models.

Furthermore, the present invention is constructed from pig brain material obtained from pigs sacrificed at an abattoir. Such pig brain material is usually perceived as a waste product and the present invention therefore represents a new ethical benchmark compared to technologies in which pigs are bred with the sole purpose of harvesting the brain. The ability to use pig brain material from local abattoirs was only possible because the inventors surprisingly found that the pig brain material could be cultured subsequent to cooled transport from the abattoir. Moreover, the inventors surprisingly discovered that the pig brain material did not have to originate from piglets, as commonly presumed, but also animals as old as 6 months yielded viable brain cell cultures.

Together, the technical progress provided by the present invention offer not only a more anatomically correct in vitro model of the human blood brain barrier compared to competing technologies, but also supplies it in a more ethical acceptable manner.

Brief description of the figures

Figure 1

Figure 1A shows an in-vitro blood-brain barrier model in different cell

combinations. Thirteen different in-vitro BBB model combinations were

constructed. 1) Mono culture of PBEC's, 2) Non-contact co-culture of PBEC's and porcine astrocytes, 3) Non-contact co-culture of PBEC's rat astrocytes, 4) Non- contact co-culture of PBEC's and porcine pericytes, 5) Non-contact co-culture of PBEC's and rat pericytes, 6) Contact co-culture of PBEC's and porcine astrocytes, 7) Contact co-culture of PBEC's and rat astrocytes, 8) Contact co-culture of PBEC's and porcine pericytes, 9) Contact co-culture of PBEC's and rat pericytes, 10) Triple culture of PBEC's, porcine astrocytes and porcine pericytes, 11) Triple culture of PBEC's, rat astrocytes and porcine pericytes, 12) Triple culture of PBEC's, porcine astrocytes and rat pericytes, and 13) Triple co-culture of PBEC's, rat astrocytes and rat pericytes.

Figure IB also shows a schematic representation of the in-vitro blood-brain barrier model with a preferred combination and placement of the different porcine cell types, corresponding to number 10 from 1A. (1) The overall BBB model, (2) astrocytes, (3) porous culture membrane, (4) pericytes, (5) BEC's, and (6) culture plate. Figure 2

Figure 2 shows characterization of primary cell cultures by immunocytochemistry. PBEC's were found to express the tight junction proteins Claudin-5 (light grey) and Zonula occludens 1 (light grey) at the cell borders. Porcine mixed glial cells (C) were found to mainly consist of astrocytes, which express GFAP (light grey). Rat astrocytes (D) were found to express GFAP (light grey). Porcine pericytes (E) and rat pericytes (F) were found to express alpha-smooth muscle actin (light grey). Porcine pericytes cultured in monoculture (G) and porcine pericytes cultured in a triple culture with porcine pericytes and PBECs in a triple co-culture (H) was stained for PDGFR-beta (light grey, without filament

structures) and alpha-smooth muscle actin (light grey with filament structures). Only a few of the porcine pericytes, which were cultured in triple co-culture, expressed alpha-smooth muscle actin compared to the porcine pericytes cultured in monoculture. Cell nuclei were stained with DAPI (grey round structure). Scale bar = 20μηι.

Figure 3

Figure 3 shows trans-endothelial electric resistance (TEER) measurement made across PBEC's in thirteen co-culture combinations. The mean TEER value of monocultures (n = 35) is significantly lower ($$$, P<0.001) than the mean TEER values for all other culture combinations (n = 13-31). Significant difference was also found between the mean TEER value for PBEC's cultured in contact co-culture with porcine pericytes (n = 24) compared to PBEC's co-cultured in contact co- culture with rat pericytes (n = 23) (**, P<0.01) (n equals numbers of inserts).

Figure 4

Figure 4 shows mannitol permeability measurements on PBEC's in thirteen co- culture combinations as a function of their TEER. TEER was measured just before the permeability experiment was conducted. The Papp mannitol where measured on n=3 for each culture condition. Each point represents one hanging culture insert with PBEC's.

Figure 5

Figure 5 shows gene expression of A) claudin-5, B), Occludin, C), transferrin, D) BCRP, and E) P-gp, in PBEC's. RT-qPCR was performed on the PBEC's from all thirteen different culture combinations. The relative gene expression of Claudin-5, Occludin, P-gp, BCRP and transferrin receptor- 1 is shown for each culture combination. The results are given as relative expression normalized to beta-actin using the Pfaffl method (n= 3-6 replicates and one replicate consist of RNA from PBEC's from 4-6 inserts). Figure 6

Functionality of the P-gp efflux transporter using rhodamine-123 (R123) as a Pgp- substrate, and C4 and Verapamil as P-gp inhibitors (n = 5-6 inserts per substrate).

Figure 7

Passive permeability of the fluorescent molecule sodium fluorescein (FLU) in co- and triple cultures (n = 3 inserts per culture condition).

The present invention will now be described in more detail in the following.

Detailed description of the invention

Definitions

Prior to discussing the present invention in further details, the following terms and conventions will first be defined :

Blood brain barrier model

The blood brain barrier (BBB) is a permeability barrier formed from brain endothelial cells (BEC's) connected by tight junctions that limits the transport of material from the blood into the brain extracellular fluid. The BBB is highly selective and thereby protects the brain from entry of foreign substances such a bacterial material. In addition to BEC's, both astrocytes and pericytes play crucial supporting roles in the establishment of the BBB. Consequently, in the present context, a BBB model comprises three types of cells; BEC's, astrocytes and pericytes.

Triple co-culture

In the present context, a triple co-culture is a conglomerate of three distinctly different cell types cultured together in a single medium. For the present invention, these three cell types are porcine brain endothelial cells, porcine astrocytes and porcine pericytes. It is noted that the different cells are preferably grown at physically distinct locations in the medium.

Confluent monolayer

In the present context, a confluent monolayer is a single layer of cells that covers an entire surface, such as a culture plate. Confluency is the fraction of a surface covered by cells and is measured in percent; such that 50% confluency means that half the area of a surface is covered with cells and 100% confluency means that the entire surface is covered with cells. It is preferred that the cells are 100% confluent, especially the BEC cells should be confluent.

Porous culture membrane

In the present context, a porous culture membrane is a membrane comprising pores of a defined size and is suitable for culturing cells on. The size of the pores decides what size of molecules will be allowed to pass the membrane, and accordingly serves as an effective molecular cut-off. The porous culture

membrane can be part of an insert used together with a culture plate, which enables the seeding of triple co-cultures. When used together with a culture plate, all transport of molecules from one side of the culture plate medium to the other side is restricted to be conveyed through the porous culture membrane and the cells attached to it.

In order to provide a more anatomically correct and reliable in vitro model of the human blood brain barrier, the present inventors have constructed a BBB model comprising a porcine triple co-culture that fulfil the benchmarks of electrical resistance, permeability and expression of proteins. Furthermore, the inventors surprisingly found that they could produce this model from pig brain material considered a waste product from the abattoir. Together, the present invention solves both a technical need of an anatomically correct BBB model and the ethical dilemma of sacrificing piglets with the sole purpose of harvesting their brain material.

From here onwards, several aspects and embodiments of the present invention will be described. It should be noted that embodiments and features described in the context of one of the aspects of the present invention also apply to the other aspects of the invention.

Cell culture

The cell culture used for production of the in vitro blood brain barrier (BBB) is a central element of the present invention. The ultimate object is to produce a BBB model that closely imitates the real life human (or pig) BBB. To achieve that goal it is necessary to build a culture comprising not only brain endothelial cells but also astrocytes and pericytes.

Consequently, a first aspect of the invention relates to an isolated triple co-culture comprising porcine primary brain endothelial cells (BEC's), porcine astrocytes and porcine pericytes. In the present invention, one reason for preferring a porcine triple co-culture BBB model to a competing rat triple co-culture BBB model has been to obtain a model that anatomically and physiologically mimic the human BBB as closely as possible.

The native BBB is shielding the brain from foreign substances in the blood by a single layer of brain endothelial cells connected by tight junctions. For formation of a selective and tight BBB model, it is therefore necessary to regulate the growth of cells in the triple co-culture to one single layer of cells.

Thus, in one embodiment of the invention, at least the BEC's of the isolated triple co-culture are present as a monolayer, such as a confluent monolayer.

In another and preferred embodiment, each type of porcine cells of the isolated triple co-culture is present as individual monolayers, such as confluent

monolayers.

To construct a native-like BBB model it is of importance that the cells when cultured are not stressed and depleted, but are viable and set to participate in the formation of the in vitro BBB model. Thus, in one embodiment of the invention, the BEC's, the astrocytes and the pericytes of the isolated triple co-culture are viable.

The most common method of creating BBB models from co-cultures is by seeding the different cells in a culture plate with an insert containing a porous culture membrane. Co-cultures can be grown as either contact or non-contact co- cultures, whereas triple co-cultures only grow as contact co-cultures. For triple co- cultures one type of cells are seeded at the culture plate and the two remaining types of cells are seeded on the either side of the porous culture membrane.

Therefore, in one embodiment of the invention, the different cell types of the isolated triple co-culture are grown at different locations in the co-culture.

In another embodiment of the invention, the isolated triple co-culture, comprises

- a culture plate (2), comprising the culture of porcine astrocytes or porcine pericytes;

- a porous culture membrane (3), comprising the culture of BEC's (5) on one side and a culture of the porcine pericytes or the porcine astrocytes on the opposite side; with the proviso that the triple co-culture comprises porcine astrocytes and porcine pericytes.

In a preferred embodiment of the invention, the isolated triple co-culture comprises

- a culture plate (2), comprising the culture of porcine astrocytes or porcine pericytes;

- a porous culture membrane (3), placed above the culture plate, comprising the culture of BEC's (5) on the side facing away from the culture plate and a culture of porcine pericytes or porcine astrocytes on the side facing towards the culture plate; with the proviso that the triple co-culture comprises porcine astrocytes and porcine pericytes. Figure 1 shows a schematic overview of such co-cultures comprising both a culture plate and a culture membrane. To imitate the real life BBB it is a requirement that molecules within the model can only transfer from one side of the porous culture membrane to the other side by crossing of the porous culture membrane on which confluent monolayers of cells are residing. This ensures that measurements of the electrical resistance and permeability characteristics of the BBB model resembles the path through the native BBB.

Thus, in one embodiment of the invention, the triple co-culture is an in vitro blood-brain-barrier (BBB) model (1) or is suitable for use as an in vitro blood- brain-barrier (BBB) model (1). Figure IB shows a schematic overview of a preferred embodiment according to the invention, where the different cell types are indicated in their preferred location in the model.

In another embodiment, the isolated triple co-culture are arranged such that substances passing from one side of the porous culture membrane to the other side pass through the porous culture membrane. In a preferred embodiment, the porous culture membrane is a hanging cell culture insert.

In nature, the brain endothelial cells are lined by pericytes, with astrocyte endfeets placed adjacent to pericytes and brain endothelial cells. It is therefore desirable to imitate this configuration of the cells in the BBB model.

Thus, in one preferred embodiment of the invention, the culture plate (2) comprises a culture of porcine astrocytes (6), and said porous culture membrane (3) comprises a culture of porcine pericytes (4) on the side facing towards the culture plate.

To enable the growth of a triple co-culture and creation of an in vitro BBB model, a hanging culture can be inserted in a regular culture plate. At this porous membrane, brain endothelial cells and either astrocytes or pericytes can be seeded to form the actual cellular permeability barrier. Conditions can be adjusted to facilitate simultaneous growth of all three cell types and ensure that each type of cells form a satisfactory monolayer of cells. Therefore, in one embodiment of the invention, the porous culture membrane (3) is a hanging culture that are not in direct physical contact with the culture plate (2). In another embodiment of the invention, the porcine triple co-culture is immersed in a liquid medium, such as a growth medium allowing for growth of all three cell types.

In yet another embodiment of the invention, the BEC's, the astrocytes and the pericytes are monolayers of 50-100% confluency, such as 70-100% confluency, such as 90-100% confluency, preferably 100% confluency.

The BBB model can be further optimized by changing parameters such as the material of the porous culture membrane or the size of the membrane pores. Therefore, in one embodiment of the invention, the porous culture membrane (3) is of a material selected from the group consisting of polycarbonate,

nitrocellulose, cellulose, collagen and fiberglass.

In another embodiment of the invention, the porous culture membrane (3) has pores of a diameter in the range 0.2-2 μηι, such as 0.5-2 μηι, such as 0.5-1.5 μηι or such as 0.7-1.2.

In the vicinity of the native BBB, the environment is not static but dynamic with blood flow transporting substances around the body. The flow of liquid introduces forces on the BBB that can be difficult to capture in a static model. Thus, in one embodiment of the invention, the cells are exposed to a flow of liquid, such as a continuous flow.

In a further embodiment of the invention, the liquid of the continuous flow is selected from the group consisting of blood, plasma, serum, cell growth media or cerebrospinal fluid.

Typically, for construction of BBB models, animals are bred with the sole purpose of harvesting the brain cells of the animals. These animals, such as piglets and rat pups, are then sacrificed at a very early age to supply cells that have not taken damage over time. In addition, it usually takes more than 10 rat pups to obtain enough brain material to construct a single BBB model. Pigs gives a higher cellular yield per animal and thereby alleviates some of the ethical concerns of sacrificing animals with the sole purpose of harvesting brain material. However, it would be preferable if viable brain cells could be retrieved in a more acceptable manner.

Consequently, in one embodiment of the invention, the porcine astrocytes (4), the porcine BEC's (5), and the porcine pericytes (6) are obtained from one or more pigs, which is at least two months old, such as at least, 4 months old or such as at least six months old, e.g. 4-12 months old, such as 4-10 months old or such as 5- 7 months old.

In another embodiment of the invention, the porcine astrocytes (4), the porcine BEC's (5), and the porcine pericytes (6) are obtained from the same pig.

In a preferred embodiment of the invention, the porcine astrocytes (4), the porcine BEC's (5), and the porcine pericytes (6) are obtained from the same pig, which is at least 4 months old, preferably at least 6 months old. In yet another embodiment of the invention, the cells of the triple co-culture are obtained from a pig sacrificed at an abattoir/slaughterhouse, a laboratory, or an animal facility, preferably from an abattoir/slaughterhouse.

Pig brains are considered to be a waste product of many abattoirs, which essentially means that all the ethical concerns of the prior art are circumvented when utilizing such waste material for the construction of in vitro BBB models.

It is a general conception that brain cells used for production of artificial BBB models a required to be fresh when culturing of each cell type is initiated. The inventors surprisingly found that pig brain samples could be transported in a cooled, but not frozen, format before culturing of the individual cell types was initiated and still produce high quality BBB models. This finding is critical for use of pig brain material from abattoirs, since these abattoirs will usually not be located next to the laboratory. Consequently, in one embodiment of the invention, the cells of the isolated triple co-culture are obtained from cooled pig brains.

In another embodiment of the invention, the cooled pig brains have been cooled to a temperature below 10°C, such as below 5°C, such as below 1°C, such as in the range 0.1-10°C, such as 2-10°C, or such as 5-10°C. Such cooling could e.g. take place by placing a pig head on ice, and storing it in an insulated container, e.g. under transport to a location for further processing. In yet another embodiment of the invention, the cooled pig brains are instantly cooled after sacrificing the pig, such as within 4 hours from death of the animal, such as within 2 hours from the death of the animal, preferably within 30 minutes from the death of the animal. In a further embodiment of the invention, the cells of the isolated triple co-culture are harvested from the cooled pig brains for further processing within 24 hours from the death of the animal, such as within 12 hours from the death of the animal, such as 4 hours from the death of the animal, preferably within 3 hours from the death of the animal.

In a preferred embodiment of the invention, the cooled pig brains are instantly cooled, within 30 minutes from the death of the animal, to a temperature in the range of 0.1-10°C and the cells of the isolated triple co-culture are harvested from the cooled pig brains for further processing within 3 hours from the death of the animal.

Two main parameters that are typically used to assess the integrity of artificial BBBs are the trans-epithelial electrical resistance (TEER) and the passive permeability of the BBB. The values are widely used within the field of science to determine the quality of the BBB model. While the TEER value is a well-defined standard, the passive permeability may be measured for a wide range of different substances. The sugar alcohol mannitol is a substance that has been used to facilitate the transport of therapeutic molecules into the brain, by shrinking the brain endothelial cells, thereby inducing a stretch of the tight junctions. Consequently, BBB passive permeability to mannitol is a valid benchmark for the integrity of the BBB model.

Thus, in one embodiment of the invention, the isolated triple co-culture has a TEER value of at least 700 Ω x cm ² , such as in the range of 700-2000 Ω x cm ² , or such as 700-1200 Ω x cm ² across the porous culture membrane.

In another embodiment of the invention, the TEER value is measured by using Millicell epithelial-volt-ohm meter and chopstick electrodes (Millipore), the TEER value is calculated as the measured values minus measurements of coated but cell free culture inserts and the difference is multiplied with the area of the culture insert.

In yet another embodiment of the invention, the isolated triple co-culture has a passive permeability of mannitol across the porous culture membrane below 4.0xl0 ⁶ cm*s ^_1 , such as 0-4.0xl0 ^~6 cm*s ^_1 , preferably below 2.5xl0 ^~6 cm*s ^_1 , such as 0-2.5xl0 ⁶ cm*s ^_1 . In example 1, it is described how TEER is measured. Further data is provided in example 4. In a further embodiment of the invention, the passive permeability of [3H] mannitol is measured by addition of radioactive labelled mannitol, [3H] mannitol, to the luminal compartment of the BBB and measuring radioactivity in the lower, abluminal compartment. The measurements are performed on a scintillation counter. In example 1, it is described how passive permeability of mannitol is measured. Further data is provided in example 5.

In a preferred embodiment of the invention, the isolated triple co-culture has a TEER value of at least 700 Ω x cm ² and a passive permeability of mannitol across the porous culture membrane below 2.5xl0 ^~6 cm*s ^_1 .

As shown in example 11 and figure 7, the isolated triple co-culture also has a low permeability towards FLU. Thus, in an embodiment, the isolated triple co-culture of the invention, has a passive permeability of FLU across the porous culture membrane below 7.0*10 ^~6 Papp (Cm-s or such as below 6.0*10 ^~6 Papp (Cm-s ^~ For ease of handling, it is preferable if the isolated triple co-culture can be stored before use. Many pharmaceutical products are stored in frozen forms or as freeze- dried products. For the present invention, a routine for storing the triple co- culture within the culture plate and porous culture membrane will be

advantageous.

Therefore, in one embodiment of the invention, the isolated triple co-culture is in a frozen state, such as below -20°C, such as below -50°C or such as below -80°C. Besides the integrity of the BBB model, an integral part of a native BBB is the expression pattern of proteins in the cells comprising the BBB. Some proteins may play crucial roles in anchoring the brain endothelial cells to each other via tight junctions and other proteins may be part of signaling pathways through transport across the BBB. The expression of these proteins can be dependent on the composition of cells in the BBB model, as astrocytes and pericytes within a triple co-culture can influence the protein expression pattern of brain endothelial cells. Consequently, the expression of a set of proteins may be used as an additional benchmark of the quality of the BBB model. Thus, in one embodiment of the invention, the BEC's of the isolated triple co- culture express all of the proteins selected from the group consisting of claudin 5, occluding, P-glycoprotein, breast cancer related protein, and transferrin receptor. In the example section, expression pattern in different cultures are determined. Uses

The isolated triple co-culture of the present invention is of porcine origin, which has a great anatomical and physiological resemblance to higher mammals. This similarity translates directly into its usefulness as an in vitro model of the blood brain barrier of higher mammals.

Consequently, a second aspect of the invention relates to the use of the isolated triple co-culture as an in vitro model of a mammalian blood-brain barrier. In a preferred embodiment of the invention, the mammal is selected from human and pig. Test method

The present invention comprises an isolated porcine triple co-culture that can be used as an in vitro BBB model, which both meets the criteria of high TEER values and low passive permeability, but also express proteins important to the native functioning of the BBB, such as transporters and receptors. Therefore, it is possible to utilize the present invention to screen substances for their ability to cross or bind the BBB model.

Thus, a third aspect of the invention, relates to a method of measuring/evaluating blood-brain barrier permeability of a substance, the method comprising

a) providing an isolated triple co-culture;

b) incubating said substance at one side of porous culture membrane; and c) measuring the amount of said substance permeating across the porous culture membrane or binding to the BEC's.

In principle, the substance (such as a compound or drug) can be added at either side of the porous culture membrane, but preferably, it is added at the side facing towards the BEC's. Another embodiment of the invention relates to the method of measuring BBB permeability of a substance, wherein said substance is intended to be able to cross the mammalian blood-brain barrier.

Yet another embodiment of the invention relates to the method of measuring BBB permeability of a substance, wherein said substance is not intended to be able to cross the mammalian blood-brain barrier.

A further embodiment of the invention relates to the method of measuring BBB permeability of a substance, wherein said substance is a molecule intended to bind one or more constituents of the blood-brain barrier.

The blood brain barrier protects the brain from exposure to foreign and unwanted substances in the blood stream. Consequently, diseases coupled to the central nervous system are in many cases difficult to treat or ameliorate due to the challenge with delivery of pharmaceutical substances across the highly selective and tight BBB. To arrest disease progression in the brain it is central to identify pharmaceutical lead structures capable of traversing the BBB. To this end, high quality BBB models are important. Therefore, one embodiment of the invention relates to the method of measuring BBB permeability of a substance, wherein said substance is a medicament or a compound intended to be a medicament.

Another embodiment of the invention relates to the method of measuring BBB permeability of a substance, wherein said substance is for or intended for the treatment or amelioration of diseases in the central nervous system.

Yet another embodiment of the invention relates to the method of measuring BBB permeability of a substance, wherein said diseases in the central nervous system are selected from the group consisting of ADHD, autism, Alzheimer's disease, Parkinson's disease, multiple sclerosis, Krabbe's disease, Huntington's disease, stroke and brain cancer.

Critical parameters for pharmaceutical substances include their release profiles, bio-distribution and half-lives, all of which are time dependent characteristics. It is therefore important to keep in mind over what time spans permeability and binding of substances are measured.

Thus, one embodiment of the invention relates to the method of measuring BBB permeability of a substance, wherein the amount of permeating substances are determined after a predetermined period of time, such as within 24 hours.

Another embodiment of the invention relates to the method of measuring BBB permeability of a substance, wherein the amount of said substance permeating across the culture insert is measured by a technique selected from the group consisting of radio-imaging, fluorescence, HPLC, FPLC, NMR, MS, ELISA, PCR and/or Western blot.

Yet another embodiment of the invention relates to the method of measuring BBB permeability of a substance, wherein the amount of said substance binding to the BEC's is measured by a technique selected from the group consisting of radio- imaging and fluorescence.

A fundamental objective for all development of pharmaceutical lead structures is to minimize adverse side effects that could abolish the usability of the substance as a medicament. One key indicator of side effects is the toxicity of a substance to the affected cells.

Toxicity testing

Thus, a fourth aspect of the invention relates to a method for

determining/evaluating the toxicity of a substance toward the blood brain barrier, the method comprising incubating said substance with the triple cell co-culture according to the invention and subsequently assessing the viability of one or more of the cell types in the triple cell co-culture.

An embodiment of the invention relates to a method for determining/evaluating the toxicity of a substance toward the blood brain barrier, wherein the viability of one or more of the cell types in the triple cell co-culture is measured by an assay selected from the group consisting of life/dead assays, Elisa, fluorescence spectrometry, TEER, permeability measurements,.

Production of triple co-culture

The present invention also provides a method for producing the triple co-culture necessary for establishment of an efficient BBB model. The method revolves around a triple co-culture in which brain endothelial cells and pericytes are in direct contact through the pores of a porous culture membrane, while astrocytes are located in the bottom of the culture plate. The astrocytes and pericytes stimulate the expression of proteins integral to the formation of the BBB. Thus, a fifth aspect of the invention relates to a method for producing an isolated triple co-culture comprising porcine primary brain endothelial cells (BEC's), porcine astrocytes and porcine pericytes, the method comprising

a) providing porcine brain cells; b) proliferating from said porcine brain cells viable BEC's, viable porcine astrocytes and viable porcine pericytes; and

c) growing said BEC's, porcine astrocytes and porcine pericytes into a triple- co-culture.

An embodiment of the invention relates to the method for producing an isolated triple co-culture, further comprising seeding primary cultured brain endothelial cells onto one surface side of a porous culture membrane; seeding primary cultured pericytes or primary cultured astrocytes onto the other surface side of the porous culture membrane; seeding primary cultured astrocytes or primary cultured pericytes onto the inside surface of a culture plate; and co-culturing these cells in a culture medium.

A preferred embodiment of the invention relates to the method for producing an isolated triple co-culture, wherein primary cultured pericytes or primary cultured astrocytes are seeded on culture plate and primary cultured pericytes or primary cultured astrocytes are seeded on the side of the porous culture membrane facing towards the culture plate. As described previously, the inventors surprisingly found conditions at which pig brain material could be recycled from abattoirs and used for establishment of a high quality BBB model. These conditions gives rise to several embodiments of the method for production of the porcine triple co-culture. Thus, one embodiment of the invention relates to the method for producing an isolated triple co-culture, wherein the porcine astrocytes (4), the porcine BEC's (5), and the porcine pericytes (6) are obtained from one or more pigs, which is at least two months old, such as at least, 4 months old or such as at least six months old.

Another embodiment of the invention relates to the method for producing an isolated triple co-culture, wherein the porcine astrocytes (4), the porcine BEC's (5), and the porcine pericytes (6) are obtained from the same pig. Yet another embodiment of the invention relates to the method for producing an isolated triple co-culture, wherein said cells are obtained from a pig sacrificed at an abattoir/slaughterhouse, a laboratory, or an animal facility, preferably from a an abattoir/slaughterhouse.

Furthermore, conditions also involve handling the pig material at adequate temperatures and periods of time. Prior art teaches that cell material for use in the production of BBB models should be cultured as close as possible to the time of harvesting the cells from the pig brains. However, the present invention provides several conditions at which cell material may be stored or transported over longer time spans and still yield cellular material suitable for culturing.

Thus, one embodiment of the invention relates to a method for producing an isolated triple co-culture, wherein the cells are obtained from cooled pig brains.

Another embodiment of the invention relates to a method for producing an isolated triple co-culture, wherein said cooled pig brains have been cooled to a temperature below 10°C, such as below 5°C, such as below 1°C, such as in the range 0.1-10°C, such as 2-10°C, or such as 5-10°C.

Yet another embodiment of the invention relates to a method for producing an isolated triple co-culture, wherein said cooled pig brains are cooled after sacrificing the pig, such as within 4 hours from death of the animal, such as within 2 hours from the death of the animal, preferably within 30 minutes from the death of the animal.

A further embodiment of the invention relates to a for producing an isolated triple co-culture, wherein cells are harvested from the cooled pig brains for further processing within 24 hours from the death of the animal, such as within 12 hours from the death of the animal, such as 4 hours from the death of the animal, preferably within 3 hours from the death of the animal.

According to the typical way of culturing astrocytes, it may be beneficial to seed the astrocytes in advance of the BEC's and pericytes. Thus, one embodiment of the invention relates to a method for producing an isolated triple co-culture, wherein the astrocytes are seeded before the BEC's and the pericytes, such as 1-10 weeks before, such 1-4 weeks before, such as 1-3 weeks before, such as 14-21 days before.

For long-term storage of brain cell types it may be preferable to cryoprotect either of the cell types. This may be relevant if one cell type is to be seeded in advance of any other cell type, such as the astrocytes being seeded in advance of the BEC's and pericytes. Cryoprotecting of cells may for example be achieved by slowly cooling the cell material or quickly by immersion of the cell material into liquid nitrogen. Cell material is typically frozen together with a cryoprotectant, such as DMSO, to preserve the sample.

Therefore, one embodiment of the invention relates to a method for producing an isolated triple co-culture, wherein one or more of the BEC's, astrocytes and pericytes are cryoprotected as a monoculture after step b) and optionally before step c).

Another embodiment of the invention relates to a method for producing an isolated triple co-culture, wherein one or more of the BEC's, astrocytes and pericytes are provided from a cryoprotected frozen monoculture before step c). A further embodiment of the invention relates to a method for producing an isolated triple co-culture, wherein the BEC's or pericytes are provided from a cryoprotected frozen monoculture.

Obtained isolated triple co-culture

A sixth aspect of the invention is an isolated triple co-culture comprising porcine primary brain endothelial cells (BEC's), porcine astrocytes and porcine pericytes obtained/obtainable by a process according to the method for producing an isolated triple co-culture comprising porcine primary brain endothelial cells

(BEC's), porcine astrocytes and porcine pericytes

It should be noted that embodiments and features described in the context of one of the aspects of the present invention also apply to the other aspects of the invention. All patent and non-patent references cited in the present application, are hereby incorporated by reference in their entirety. The invention will now be described in further details in the following non-limiting examples.

Examples

Example 1 - Materials and methods

Isolation of PBEC's

PBEC's derived from 6 months old domestic pig brains were obtained from the local abattoir (Danish Crown, DK), which are obligated to follow the Danish regulations within animal welfare and are under constant supervision by the Danish and European Food Standard Agency. The brains were collected and transported on ice to the Laboratory of Neurobiology, Aalborg University,

Denmark. The isolation of the PBEC's was started within 2-3 hours from

termination of the animal. Meninges were removed and approximately 12-15 g cortex, containing as little white matter as possible (approximately 20%), were collected in DMEM-F12 (Life Technology, Naerum, Denmark, DK) and cut into small pieces using scalpels. The tissue was digested in collagenase II (Life

Technology) and DNase I (Roche, Hvidovre, Denmark, DK) for 75 min at 37 °C, and purified in 20% BSA, followed by a second enzyme treatment with

collagenase/dispase (Roche) and DNase I for 50 min at 37°C. Microvessels were collected using a 33% Percoll gradient (Sigma-Aldrich, Brondby, Denmark, DK). The isolated microvessel fragments were finally plated on to 60 mm ² plastic dishes coated with collagen IV (Sigma-Aldrich) and fibronectin (Sigma-Aldrich). PBEC's were maintained in DMEM/F12 supplemented with 10% plasma-derived serum (First Link, Wolverhampton, United Kingdom, UK), basic fibroblast growth factor (Roche), heparin (Sigma-Aldrich), insulin, transferrin, sodium selenite (Roche) and gentamicin sulphate (10 pg/ml) and cultured in an incubator with humidified 5 % CO 2 / 95 % air at 37 °C. Puromycin (Sigma-Aldrich) was added to the media for the first 3 days to obtain a pure culture of PBEC's. After 3 days, the cells were passaged and either seeded on to 1.12 cm ² Millicell hanging culture inserts with 1 μηι pore size (Millipore) in a density of 100.000 cells per insert or frozen until seeded as just described. All experiments on PBEC'S were conducted on passage 1.

Cerebral porcine pericytes

Cerebral porcine pericytes were obtained by culturing a cell fraction obtained from the PBEC's isolation protocol. When the microvessels were collected from the Percoll gradient, the underlying cell fraction in the gradient was collected as pericytes. Pericyte survival and proliferation were favored over PBEC's by i) using uncoated dishes, ii) addition of puromycin, and iii) DMEM supplemented with 10% fetal calf serum and gentamicin sulphate. Only passage 1 or 2 of primary porcine pericytes were used in this study. It is noted that (like astrocytes) pericytes can also be cryoprotected and frozen before use. Porcine brain endothelial cells can also be cryoprotected and frozen before use. Thus, the cerebral endothelial cells, astrocytes and pericytes can all be cryoprotected and frozen before they are used in a triple culture. Therefore, cells from the same animal can be stored and later thawed for new/different purposes without being limited by a short timeframe.

Rat pericytes were derived from 2-3 weeks old Sprague Dawley rats as previously described by Nakagawa et al. A new blood-brain barrier model using primary rat brain endothelial cells, pericytes and astrocytes. Neurochem Int. 2009;54: 253- 263. The rats were deeply anesthetized by a subcutaneous injection of 0.5 ml / 10 g body weight of Hypnorm/Dormicum (Fentanyl/Fluanisone mixed with Midazolam and sterile water in a ratio of 1 : 1 :2). The rat heads were rinsed with 70 % ethanol and 10 % poly(vinylpurrolidone)-iodine complex. The head was separated from the body by scissor. The brains were gently removed from the scull, and the forebrain collected in ice-cold PBS. The meninges and any visible white matter were carefully removed. From here on the protocol for isolation of porcine pericytes was used. Porcine glia cells - astrocytes

Mixed cultures of porcine glia cells were additionally obtained from the brain of the 6 months old domestic pigs. Approximately 1.5-2 g of cortical pieces were collected and mechanically dissociated in DMEM supplemented with 10% fetal bovine serum and gentamicin sulphate. Dissociated cells were seeded into culture flasks until they reached confluence, frozen in media supplemented with DMSO and FCS in a -80 °C freezer for 24 hours, and then moved to a -140°C freezer until use. It was evidenced by immunocytochemistry that the mixed glial cell cultures consisted mainly of astrocytes and only a few microglial cells, and therefore the mixed glial cell population is referred to below as porcine astrocytes. The porcine astrocytes were thawed two-three weeks before establishment of co- culture models and seeded in 12 well dishes to obtain a confluent layer for co- culture. During the first three days of culturing of the porcine cells, the antibiotic chloramphenicol (Sigma-Aldrich) was added to the culture medium due to the high occurrence of methicillin resistance staphylococcus aureus (MRSA) (CC398) in Danish pigs.

Mixed cultures of rat glial cells were also isolated from neonatal Sprague Dawley rats as previously described by Nakagawa et al. The rats were obtained from the Animal Facility at Aalborg University Hospital. Their mothers were fed and housed under a 12/12 h dark/light cycle and had free access to food and water until they were euthanized. All animals were put down following the directions given by the Danish Experimental Animal Inspectorate.

The rats were rapidly decapitated by scissor, and their brains removed from the scull. From here on the procedure for isolation of astrocytes and pericytes described above was followed.

In vitro BBB model construction

Thirteen different in vitro BBB models were constructed using the five different primary cell types, i.e PBEC's, porcine astrocytes, porcine pericytes, rat

astrocytes, and rat pericytes (Fig. 1). The thirteen different models were subdivided into four different types of in vitro BBB models. The first and simplest in vitro BBB model was a monoculture of PBEC's, in which the PBEC's were cultured on the upper side of the hanging culture inserts. The second type of in vitro BBB model was a non-contact co-culture model in which the culture insert containing PBEC's were cultured together with porcine astrocytes, rat astrocytes, porcine pericytes or rat pericytes, which were located on the bottom of the 12 well culture dish. The third type was a contact co-culture models in which porcine astrocytes, rat astrocytes, porcine pericytes or rat pericytes was cultured on the bottom of the culture insert, together with PBEC's, which were cultured on the upper side of the culture insert. The fourth and final type of in vitro BBB model was a triple co-culture model. In this model, the PBEC's were cultured on the upper side of the culture inserts, while porcine or rat pericytes were cultured on the bottom side of the culture inserts and porcine or rat astrocytes were seeded on the bottom of the culture dish. The PBEC's in all thirteen in vitro BBB models were supplied once with 550nM hydrocortisone (Sigma-Aldrich), 250 μΜ cAMP (Sigma-Aldrich) and 17.5μΜ RO-201724 (Sigma-Aldrich) to further induce BBB characteristics, when the PBEC's had reached confluence approximately 24 hours after seeding.

When constructing the contact co-cultures the astrocytes or pericytes were seeded in a density of 80.000 cells per insert. The hanging cell culture insert was turned upside down in a large petri dish and coated with poly-l-lysine for seeding astrocytes. The appropriate amount of cells was resuspended in 100 μΙ media per insert and seeded on the insert. The closed petri dish with the hanging cell culture inserts were then placed in an incubator for 3-4 hours until attached. The inserts were then placed hanging into a 12 well culture dish supplied with media in both insert and well and incubated for three days, until PBEC's were seeded in the inserts as described previously.

For construction of non-contact co-cultures with pericytes a density of

20.000cells/cm ² was seeded into a 12 well culture dish and incubated for 2-3 weeks before co-culture studies was conducted. Astrocytes cultured in non- contact co-cultures were seeded as described previously.

Evaluation of barrier integrity

The barrier integrity of the different in vitro BBB models was evaluated by measurement of TEER and permeability to radiolabeled mannitol (Pelkin Elmer, Skovlunde, Denmark, DK).

TEER

TEER was measured using a Millicell epithelial-volt-ohm meter and chopstick electrodes (Millipore). The TEER value was calculated as the measured values minus measurements of coated but cell free culture inserts for monoculture and contact co-culture or coated inserts with either astrocytes or pericytes on the bottom of the insert for contact-co-cultures. The difference was multiplied with the area of the culture insert (1.12cm ² ), resulting in a TEER value given as a mean in Ω x cm ² ± standard deviation. TEER values were obtained from 35 culture inserts with PBEC's cultured in monoculture (n = 35) and from 13-31 culture inserts with PBEC's cultured in co-cultures and triple co-cultures (n= 13- 31).

Passive permeability

Passive permeability was analyzed by the addition of 1 μθ 3H-D-Mannitol (Specific activity 14.2 Ci/mol) to the upper chamber of a culture insert. The passive permeability was performed on three individual culture insert of each of the thirteen different in vitro models (13 x n=3). The culture plate was placed on a rocking table at 37 °C for 120 min. Samples of 100 μΙ were collected from the upper chamber at 0 and 120 min, and from the lower chamber at 0, 15, 30, 60 and 120 min. The samples were replaced with 100 μΙ fresh culture medium.

Samples were added with Ultima Gold™ liquid Scintillations fluid (Pelkin Elmer) and counted in a liquid scintillations counter. The total amount of millimoles transported in each well was plotted against time. The flux at steady state was calculated as the slope of the straight line divided by the area of the culture insert (1.12 cm ² ). Finally, the apparent permeability (Paap) was calculated by dividing the flux at steady state with the initial concentration in the donor upper compartment. The calculated Paap data were plotted against TEER values for each individual culture insert. Data from TEER and passive permeability were analyzed by the GraphPad Prism 5.0 software using a 1-way ANOVA with Bonferroni's multiple comparisons test. Immunocytochemistry

All primary and secondary antibodies were dissolved in PBS (1 : 200) prior to labeling. The PBEC's, astrocytes and pericytes were fixed in 4% paraformaldehyde and blocked in PBS supplemented with 0.2 % Triton-X-100 and 3 % bovine serum albumin for 1 hour. The PBEC's were stained with polyclonal rabbit anti-claudin-5 (Sigma-Aldrich, cat. no. SAB4502981, lot 310145) and polyclonal rabbit anti-ZO-1 (Invitrogen, cat. no. 617300, lot 1087989A). Mixed glial cells were stained with rabbit anti-glial fibrillary acidic protein (GFAP)(DAKO, DK, cat. no. Z0334, lot 20003791) and Texas Red labeled Lycopersicon Esculentum (Tomato) Lectin (Vector Labs, Peterborough, United Kingdom, cat. no. TL1176, lot W0812).

Pericytes were stained with monoclonal mouse anti-a-smooth muscle actin (a- SMA) (Sigma-Aldrich, cat. no. A5228, lot 091M4832), polyclonal rabbit anti-ZO-1 and rabbit anti-platelet-derived growth factor receptor-beta (PDGFR-β) (Santa Cruz, cat.no.Sc-432, lot K1113). For detection, the cells were subsequently stained with goat anti-rabbit Alexa 488 or goat anti-mouse Alexa 585 (Invitrogen) as the secondary antibodies. All cells were counterstained with DAPI. The Millicell membranes were cut out of the inserts and mounted on glass slides in fluorescent mounting media (Dako, Denmark) and cover slips were placed upon the membranes.

RT-qPCR analysis

All reagents for RT-qPCR were obtained from Thermo Scientific (Slangerup, Denmark, DK), except primers that were purchased from TAG Copenhagen (Frederiksberg, Denmark, DK). RNA was isolated from PBEC's from all thirteen in vitro BBB model setups and obtained in 3-6 replicates and one replicate consisted of PBEC's from 4-6 inserts. RNA was isolated using the GeneJet RNA purification kit. The RNA samples were treated with DNase I to eliminate genomic DNA and 100 ng RNA was converted to cDNA using the RevertAid First Strand cDNA

Synthesis Kit. The expression profile of endothelial cell characteristic proteins was assessed with the qPCR technique using primers specific for claudin-5, occludin, transferrin receptor, p-glycoprotein (P-gp) and breast cancer resistance protein (BRCP). Beta actin was used for normalization purpose (Table 1). Each qPCR reaction was performed by mixing 2.5 ng cDNA and 10 pmol of each primer with the Maxima™ SYBR Green qPCR Mastermix. Each sample was performed in triplicates, while non-reversed RNA and water served as negative controls. The qPCR reactions were 95 °C for 10 min, 40 cycles of 95 °C for 30 sec, 60 °C for 30 sec and 72 °C for 30 sec, which were performed using the Stratagene Mx3000P™ QPCR system (Agilent Technologies, Horsholm, Denmark, DK). The relative expression of mRNA was calculated and analysed in the GraphPad Prism 5.0 software using a 1-way ANOVA with Tukey's multiple comparisons post hoc test. Table 1. The table displays the reference sequence numbers and primer sequences of the six primers used in this study.

Reference

Target Forward primer Reverse primer sequence

Claudin NM_0011616 GTCTTGTCTCCAGCCATGGG GTCACGATGTTGTGGTCC 5 36.1 TTC (SEQ ID NO: 1) AGGAAG

(SEQ ID NO: 2)

GCCCATCCTGAAGATCAGGT CTCCACCATATATGTCGTT

NM_0011636

Occludin GAC GCTGGG

47.2

(SEQ ID NO: 3) (SEQ ID NO: 4)

Transfer TTGATGATGCTGCTTTCCCTT CCATTCTGTTCAACTGAGG

NM_214001.

rin- TCCT AACCCT

1

receptor (SEQ ID NO: 5) (SEQ ID NO: 6)

CGATGGATCTTGAAGAAGGC CCAGTTTGAATAGCGAAAC

XM_0031302

Pgp CGAAT ATGGCA

05.2

(SEQ ID NO: 7) (SEQ ID NO: 8)

GCTATCGAGTGAAAGTGAAG AACAACGAAGATTTGCCTC

NM_214010.

BCRP AGTGGCT CACCTG

1

(SEQ ID NO: 9) (SEQ ID NO: 10)

CAGAGCGCAAGTACTCCGTG GCAACTAACAGTCCGCCT

XM_0031242

β-Actin TGGAT AGAAGCA

80.2

(SEQ ID NO: 11) (SEQ ID NO: 12)

Example 2 - Cell cultures

PBEC's, astrocytes and pericytes were isolated from 6 months old domestic pigs. Thereby, the present in vitro BBB model gives the advantage of having all three cell types forming an in vitro BBB derived from the same species. Because astrocytes need culturing for two-three weeks before being co-cultured in most in vitro BBB models, they are often isolated from a different animal than the PBEC's and pericytes. This disadvantage can be avoided in the present model by freezing PBEC's and pericytes, and thereby the present in vitro BBB model contained all three cell-types from the same animal. The average yield of PBEC's per isolation from 12-15 grams of brain tissues was 8-10 x 10 ⁶ cells and for porcine pericytes the yield was on average 5.0 x 10 ⁶ cells. Concerning porcine astrocytes, 5.0 x 10 ⁷ cells were isolated from 2 grams of brain. Such high cellular yield allows for many experiments on the in vitro BBB model using cells from the same isolation. The porcine astrocytes grew slowly in the first week of culture compared to rat astrocytes, but after two weeks of continued culture no difference in growth was detected between the porcine astrocytes and rat astrocytes. When grown on collagen/fibronectin coated hanging cell culture inserts PBEC's acquired the characteristic morphology of brain endothelial cells, seen as a tightly connected polarized monolayer. Conclusion

It is possible to construct a BBB model comprising a triple co-culture of PBEC's, porcine astrocytes and porcine pericytes wherein all three cell-types may be derived from the same animal. Example 3 - Immunocytochemical stain of PBEC's, porcine astrocytes and pericytes

The PBEC's contained claudin-5 and ZO-1 (Fig. 2a+b). Claudin-5 was abundant at cell borders, but also in the cytosol. ZO-1 formed a continuous border between the cells. Pericyte contamination in the PBEC monolayers was reduced with the addition of puromycin to the PBEC's for the first three days of culture. Only a minor fraction of a-SMA positive pericytes ocurred below the PBEC monolayer, but this did not disrupt the PBEC monolayer, which was confirmed by high TEER and consistent expression of claudin-5 and ZO-1. The PBEC monoculture was without GFAP immunoreactivity indicating absence of astrocytes. The porcine mixed glial cells mainly consisted of GFAP positive astrocytes (Fig. 2c). The rat mixed glial cells mainly consisted of GFAP positive cells (Fig. 2d) and a few microglia (data not shown).

The porcine pericytes stained positive for a-SMA and PDGFR-β, when cultured in monoculture (Fig. 2g). Porcine pericytes co-cultured in a triple co-culture with PBEC's and astrocytes stained positive for PDGFR-β but only a minority of the pericytes were a-SMA positive (Fig. 2h). These observations are in good accordance with studies on differential stages of pericytes, which have shown that pericytes always express PDGFR-β, irrespective of differential stage but turn into a-SMA negative pericytes when subjected to bFGF. Furthermore, a-SMA negative pericytes have been shown to induce higher TEER than a-SMA positive pericytes. In the present study the pericytes were first isolated and cultured in monoculture in bFGF free media resulting in a-SMA positive (Fig 2e+g). When the pericytes were co-cultured with PBEC's or in a triple co-culture with PBEC's and astrocytes bFGF was added to the media resulting in a-SMA negative pericytes (Fig. 2h). In the pericyte cultures, very small clusters of PBECs lying on top of the pericyte monolayer could be found by ZO-1 staining. The immunocytochemical stainings all indicated that the three cell types have been successfully isolated and expressed cell-specific proteins. Rat pericytes were a-SMA positive in monoculture (Fig. 2f) and in co-culture or triple co-culture fewer of the rat pericytes were a-SMA positive (data not shown).

Example 4 - trans-endothelial electrical resistance of PBECs in thirteen culture conditions

Methodological considerations

TEER measurements denote a valid real-time monitor of the BBB integrity.

Reportedly, primary PBECs form TEER values between 70-1800 Ω x cm ² depending on their culture conditions. PBECs were previously mainly cultured as pure monocultures or as co-cultures with primary rat astrocytes or astrocytic cell lines like C6 glioma. We have successfully isolated porcine astrocytes and pericytes from 6 months old domestic pigs, and would, therefore, like to

investigate whether a good in vitro BBB model could be obtained by PBECs co- cultured with porcine astrocytes and pericytes, and if this model would be just as good an in vitro BBB model as PBECs co-cultured with rat astrocytes and pericytes. We established thirteen different culture combinations with PBECs in co-culture with porcine astrocytes, porcine pericytes, rat astrocytes and/or rat pericytes (Fig. 1). The TEER values monitored as a measure for tightness (Fig. 3) were on average 344±25 Ω x cm ² for the PBECs in the monocultures, which is within the range of TEER values found in other studies on PBECs in monoculture. The TEER values of the present study was measured by chopstick electrodes, and the PBECs grown on hanging cell culture inserts with a pore size of 1 μηη, which differs from the before mentioned studies, which all used a pore size of 0.4 μηι and two of the studies measured TEER with an Endohm electrode chamber and the rest with chopstick electrodes. As the pore size and TEER measuring devices affect TEER values, this should be accounted for when comparing studies. Porcine brain endothelial cells in monoculture

The TEER values of the monoculture were found to be significantly lower

(P<0.0001) than PBEC's cultured in co-culture and triple co-culture (Fig. 3). That co-culture with rat astrocytes beneficially impacts TEER was shown in other studies on PBEC's [22,32,34] It has also been shown that TEER values in rat BECs can be increased by co-culture with primary rat pericytes and triple co-culture with primary rat astrocytes and pericytes increases the TEER even higher [18], Too our knowledge the present study is the first to show that TEER can be significantly increased by co-culture and triple co-culture with primary porcine pericytes and porcine astrocytes.

Porcine brain endothelial cells in contact co-cultures

For the contact co-cultures, TEER of PBEC's varied from 831±29 Ω x cm ² for the co-contact porcine pericytes to 1192±113 Ω x cm ² for the contact rat astrocytes. This indicates that rat astrocytes are better at inducing high TEER in PBEC's if they are cultured in contact co-culture, which has also been found by Malina et al with a mean TEER at approximately 1100 Ω x cm ² [35].

Porcine brain endothelial cells in non-contact co-cultures

The PBEC's cultured in non-contact co-cultures varied from 778±33 Ω x cm ² for the non-contact porcine pericytes to 1093±60 Ω x cm ² for the non-contact porcine astrocytes. This means that porcine astrocytes should be preferred if PBEC's are cultured in non-contact co-cultures. Studies made on PBEC's cultured in non- contact co-culture with primary rat astrocytes has reported mean TEER values from ~400-800 Ω x cm ² , which is lower than the mean TEER value of 881±33 Ω x cm ² obtained in the present study.

Porcine brain endothelial cells in triple co-cultures

The TEER values for PBEC's cultured in triple culture varied from 1052±55 Ω x cm ² for the triple porcine culture to 1171±55 Ω x cm ² for the triple culture with rat astrocytes and rat pericytes (Fig. 3). The small non-significant differences found in TEER between the PBEC's cultured in the four different triple co-cultures (Fig. 1) indicates that rat astrocytes and pericytes cannot be preferred over porcine astrocytes and pericytes. Comparison of the influence of porcine versus rat pericytes and astrocytes on TEER values

The TEER values obtained on PBEC's co-cultured with either rat or porcine astrocytes or pericytes were also compared. There was no significant difference between the TEER obtained on PBEC's cultured in non-contact co-cultures when comparing the inductive properties of rat and porcine astrocytes or pericytes. Only TEER values of PBEC's that had been cultured in contact co-cultures with porcine pericytes were significantly lower (P<0.05) than TEER values obtained from PBEC's cultured in contact with rat pericytes (Fig. 3). Rat pericytes are therefore better at inducing high TEER in PBEC's when cultured in contact co-culture in comparison with porcine pericytes. This is though not the case when PBEC's are cultured in non-contact co-culture with porcine or rat pericytes, where no significant difference was found. As earlier mentioned there were no significant differences within the four different triple cultures suggesting that the enhancing impact that rat pericytes might have on PBEC's in contact co-culture did not seem to have an enhancing effect when astrocytes are also included in the culture.

Conclusion

The results show that an in vitro model established from a triple co-culture of PBEC's, porcine astrocytes and porcine pericytes is just as tight as an in vitro model based on co-culture of PBEC's with rat astrocytes and pericytes. Primary rat astrocytes and pericytes are derived from brains from either rat pups or two-three weeks old rats. This method is expensive and not as ethically to use compared to the use of porcine astrocytes and pericytes that are derived from pig brains that are a waste product in the meat industry.

Therefore, the TEER results suggest that researchers could benefit from using an in vitro BBB model based solely on porcine cells.

Example 5 - Apparent permeability of PBEC's in thirteen culture

conditions

Optimal properties of an in vitro BBB model are reflected in high expression of tight junction proteins that do not just lead to a high TEER, but also lead to a low permeability of e.g. sodium fluorescein or mannitol from the luminal to the abluminal side of the in vitro BBB model. The apparent mannitol permeability was measured on PBEC's in thirteen different culture conditions and plotted against TEER values measured on the same PBEC's just before the permeability

experiments was initiated (Fig. 4).

The permeability showed an inverse relation to TEER. The lowest permeability and hence highest integrity was found in PBEC's cultured in co-contact with rat astrocytes, which had an average steady state mannitol permeability of 0.87 ± 0.04 x 10 ^"6 cm x s ^"1 . Also, this culture setup had the highest TEER before the permeability experiment of an average of 1994 ± 79 Ω x cm ² . All the co- and triple co-culture setups were in the range of 4.10 - 0.87 x 10 ^~6 cm x s ¹ and no significant difference was found within these culture conditions. The highest permeability and hence lowest integrity was found in the monocultures, which had an average steady state mannitol permeability of 8.06 ± 2.4 xl0 ^~5 cm x s ^_1 . This was consistent with the TEER value hence the monoculture had the lowest TEER with an average of 535 ± 32 Ω x cm ² . The apparent mannitol permeability measured on the monoculture was significantly higher (P<0.0001) than on all the PBEC's cultured in co- and triple co-cultures. Zhang et al have reported a comparable apparent mannitol permeability of 9.4 x 10 ^~5 cm x s ¹ for PBEC's in non-contact co-cultures with mean TEER of ~500 Ω x cm ² . Franke et al reported a permeability coefficient to mannitol in PBEC's in monoculture as low as 1.8 x 10 ^~6 cm x s ^_1 , but with corresponding high peak TEER values of ~1500 Ω x cm ² . This corresponds well with our permeability values of PBEC's with a TEER of ~1500 Ω x cm ² (Fig. 4). Likewise Naklband and Ohmidi reported a mannitol permeability coefficient for PBEC's in non-contact co-culture with rat C6 glioma cells at

2.31xl0 ⁶ cm x s ¹ with corresponding TEER of 900 Ω x cm ² , and Patabendige et al reported a Papp to mannitol of PBEC's in a non-contact co-culture with rat astrocytes at a range of 0.1-2.6 x 10 ^~5 cm x s ¹ and TEER values of ~800 Ω x cm ² . All permeability values and corresponding TEER values seems to be in the same range as the ones found in the presents study (Fig. 4) and, therefore, it seems that the inverse relationship is the same regardless of isolation procedures for PBEC's and to some extend also co-culture conditions.

Gaillard and de Boer found that there was an inverse relation between Papp and TEER, and found that at a certain level the Papp would not decrease any further despite of increasing TEER. Similarly, Franke et al found that PBEC's with a TEER of 600 Ω x cm ² or above had reached a low permeability that did not decrease with further increasing TEER. Patabendige et al found that the apparent

permeability was independent of TEER when TEER was above 200 Ω x cm ² . In the present model low and relatively steady permeability seems to be reached in between a TEER value of 606 Ω x cm ² and 704 Ω x cm ² . TEER should therefore preferably be above 700 Ω x cm ² to ensure a low permeability with the PBEC's in the present study.

Example 6 - Differences in relative mRNA expression in PBEC's in the thirteen different culture conditions

The isolated porcine PBEC's were co-cultured with porcine and rat astrocytes and pericytes in the thirteen different culture conditions found in Fig. 1, and RNA was isolated from PBEC's in each setup. Generally for PBEC's cultured in all conditions, relative mRNA expression of claudin-5, occludin, transferrin receptor, P-gp and BCRP was confirmed by RT-qPCR (Fig. 5). These are all BBB relevant tight junction proteins, receptors and efflux transporters and their expression indicates that important BBB features have been maintained in the PBEC's in culture. No significant differences were found in the relative mRNA expression of all the tested genes between PBEC's grown in triple co-culture with porcine astrocytes and pericytes and PBEC's cultured in triple co-culture with rat astrocytes and pericytes. Therefore, a triple co-culture model consisting entirely of porcine cells could be preferred based on relative gene expression.

Claudin-5

Expression of claudin-5 was significantly increased in PBEC's when they were cultured in contact co-culture with either porcine astrocytes (P<0.01), porcine pericytes (P<0.0001) and triple culture with rat astrocytes and porcine pericytes (P<0.05). The claudin-5 expression was significantly higher when PBEC's were cultured in a contact co-culture with porcine pericytes (P<0.0001) compared to rat pericytes. This indicates that claudin-5 expression in PBEC's depends on induction from porcine pericytes, which cannot be substituted by rat pericytes. Unfortunately, it also seems that when porcine astrocytes were included in the triple porcine culture, the inductive properties of the porcine pericyte were reversed, which was not seen when rat astrocytes were used instead for a triple culture with PBEC's and porcine pericytes. The results reveals that porcine astrocytes should be in contact with the PBEC's for inducing the expression of claudin-5 due to a significant increase in expression (P<0.05) seen when comparing with claudin-5 expression in PBEC's in non-contact co-culture with porcine astrocytes. Malina et al found similar results in PBEC's in co-culture with rat astrocytes where the protein expression of claudin-5 was significantly increased by co-culturing the astrocytes in a contact co-culture with the PBEC's instead of a non-contact co-culture. Porcine pericytes should also be cultured in contact rather than non-contact co-culture with the PBEC's to induce a significant increase (P<0.05) in claudin-5 expression. Overall, we conclude that porcine pericytes are the most important cell type for increasing the expression of claudin- 5 in PBEC's when cultured in a contact co-culture.

Occludin

Expression of occludin was only significantly increased by non-contact co-culture with rat pericytes (P<0.0001) and in triple co-culture with rat astrocytes and rat pericytes (P<0.05) when compared to monoculture. Furthermore, in non-contact co-culture rat pericytes significantly increased the occludin expression in PBEC's compared to porcine pericytes in non-contact co-culture (P<0.0001). It should though be noted that there were no significant difference between the occludin expression in PBEC's cultured in contact co-culture with either porcine or rat pericytes. It can, therefore be concluded that occludin expression in PBEC's is highly upregulated by co-culture with rat pericytes in a non-contact co-culture. Malina et al showed that the protein expression of occludin was increased although not significant when PBEC's where co-cultured in contact with the astrocytes compared with PBEC's in non-contact co-cultures with rat astrocytes. However, in the present study PBEC's co-cultured with rat astrocytes in non- contact co-culture have a higher relative mRNA expression of occludin than if they were co-cultured together in a contact co-culture. This difference could be due to differences in isolation and culture methods and this indicates that it is very important to have knowledge of exactly how the PBEC's react in different culture setups to determine which conditions are the most optimal for different research purposes.

P-glycoprotein

P-gp (ABCBl) expression in PBEC's was not significantly increased by co-culture or triple co-culture with either porcine or rat astrocytes or pericytes. P-gp expression was increased significantly in PBEC's by non-contact co-culture with rat pericytes when compared to non-contact co-culture with porcine pericytes

(P<0.0001). The data also shows that rat pericytes in contact co-culture with PBEC's significantly decreased the P-gp expression compared to non-contact co- culture with rat pericytes (P<0.0001). It therefore seems that the highest expression of P-gp is achieved by culturing the PBEC's in a non-contact co-culture with rat pericytes.

Breast cancer resistance protein

BCRP (ABCG2) expression was significantly increased in PBEC's in non-contact co- culture with rat astrocytes (P<0.01) and rat pericytes (P<0.05). Furthermore, BCRP expression was significantly increased in all four types of triple co-culture setups (P<0.05). There was no significant difference between co-culturing PBEC's with rat or porcine astrocytes or pericytes. Although not significant, the BCRP expression is upregulated in PBEC's when they are cultured in co-culture and the expression is even higher when they are cultured in triple co-cultures. This up regulation seems to be independent of astrocyte and pericyte origin.

Transferrin receptor

Transferrin receptor expression was only significantly decreased in PBEC's by non- contact co-culture with rat astrocytes (P<0.01) and in triple co-culture with rat astrocytes and porcine pericytes (P<0.05).

Example 7 - Pericytes impact on TEER, gene expression and permeability in co-culture and triple culture.

Nakagawa and colleagues found that rat pericytes significantly increased TEER and decreased permeability of primary rat BECs in contact co-culture and triple culture with rat pericytes in contact with the endothelial cells and astrocytes in the bottom of the well when compared to rat BECs in either monoculture or in co- culture with only rat astrocytes.

In the present study no significant difference was found between PBEC's in co- culture with porcine/rat astrocytes and PBEC's in a triple culture with porcine/rat astrocytes and porcine/rat pericytes (contact) either regarding gene expression, TEER or permeability. The porcine pericytes significantly increase the claudin-5 expression in contact culture with PBEC's when compared to PBEC's in non-contact co-culture with porcine astrocytes (P<0.01) and in triple culture with porcine astrocytes and porcine pericytes (contact) (P<0.01) (Fig. 5). This indicates that the porcine pericytes are favorable to use in a co-culture with PBEC's if claudin-5 should be highly expressed. Furthermore, rat pericytes are also important for obtaining a high expression of occludin and P-gp in PBEC's when cultured in a non-contact co- culture (Fig. 5).

Example 8 - Impact of astrocytes vs pericytes on PBEC's in co-culture.

When comparing TEER values, rat and porcine astrocytes do not have higher inductive skills on TEER in PBEC's than rat and porcine pericytes do (Fig. 3).

Regarding permeability no significant difference was found between co-cultures with either porcine or rat pericytes and astrocytes (Fig. 4).

Furthermore, no significant difference between co-cultures with either rat or porcine astrocytes and pericytes was found regarding gene expression of claudin- 5, occludin, Pgp, BCRP and the transferrin receptor.

Conclusion

It is concluded that astrocytes and pericytes of either porcine or rat origin are equally good at inducing high TEER, low permeability and gene expression of the five investigated genes in the present study.

Example 9 - Is monoculture, co-culture or triple co-culture the most optimal culture condition for PBEC's for establishing an in vitro BBB model?

TEER and apparent permeability to mannitol was significantly increased

(P<0.0001) by co-culture and triple co-culture compared to monoculture of PBEC's (Fig. 2 and 3). Statistically TEER values and permeability are not more optimal, when culturing PBEC's in triple co-culture as opposed to co-culture alone. In gene expression, a tendency towards a higher expression of occludin (P<0.01), claudin-5, P-gp (P<0.05) and BCRP and lower expression of transferrin receptor was found, when the PBEC's were cultured in triple cultures with rat astrocytes and rat pericytes, when compared to co-culture alone with either non-contact culture with rat astrocytes or contact co-culture with rat pericytes (Fig. 5). It seems more complicated when comparing the mRNA expression in PBEC's in the triple porcine culture with PBEC's in either non-contact co-culture with porcine astrocytes or contact co-culture with porcine pericytes. The only significant difference was found in the expression of claudin-5 (P<0.01) which was increased in PBECs cultured in contact co-culture with porcine pericytes compared to PBECs in triple porcine cultures.

Example 10 - Investigations of the functionality of the P-gp efflux transporter using rhodamine-123 (R123) as a Pgp-substrate, and C4 and Verapamil as P-gp inhibitors.

Introduction

The P-gp efflux transporter is very important for a functional BBB. If the P-gp transporter is inactive, drugs that would normally be effluxed by the P-gp transporter would instead be able to pass through the BBB and get access to the brain. The in vitro BBB model will therefore not be a useful model for predicting drug permeability if the P-gp efflux transporter is not active. We have detected the presence of the P-gp efflux transporter in the PBECs by RT-qPCR, but it is also essential to investigate the functional activity of the efflux transporter. This was investigated by adding verapamil, which is an efflux transporter inhibitor, and C4, which is a P-gp specific inhibitor to the PBECs cultured in the triple culture setup. The two inhibitors were used to assess the permeability of the P-gp substrate R123. If the P-gp efflux transporter is functional, R123 should be effluxed from the PBECs, and would therefore not able to pass through the BBB, but if this efflux is inhibited by applying C4 or Verapamil, R123 would be able to cross the BBB.

Materials and methods

Functionality of the Pgp-efflux transporter was analyzed by the addition of 20 μΜ of the Pgp-substrate R123 in the upper chamber with or without a Pgp-inhibitor (10μΜ C4 or 25μΜ verapamil) in order to block the efflux transport. The passive permeability was performed on five to six biological replicates for each fluorescent molecule (2 x n = 5-6). Each replicate with a previously measured TEER value was washed once in PBS and then transferred to a 12well plate containing 1.5ml Ringer-Hepes buffer in each well. The insert was the added 1ml Ringer-Hepes buffer with the calculated amount of R123 and either C4 or verapamil and incubated on a rocking table at 37 °C for two hours. Following the incubation, samples were collected. The samples were diluted 1 : 100 in Ringer-Hepes buffer and loaded in black 96well plate (ΙΟΟμΙ pr. well) before analyzed using

fluorescence spectrometry (Enspire, Perkin Elmer). Data was accumulated in Microsoft Excel, using an adapted method by Culot et al. (Lens, France) and mean Papp (Cm-s ^_1 ) ± SEM was plotted using GraphPad Prism 5.0 software. The Papp values were normalized with the Papp of R123 set to 100%.

Results and discussion

The accumulation of R123 was measured in the wells with or without addition of C4 and verapamil. The accumulation of R123 was increased to 212±50.80% by inhibiting the P-gp efflux transporter with C4. Verapamil inhibited the P-gp efflux transporter slightly by increasing the accumulation of R123 to 125±40.14%

(Figure 6).

Conclusion

Based on the results it can be concluded that the porcine in vitro BBB model has a functional P-gp efflux transporter.

Example 11 - Passive permeability using sodium fluorescein (FLU).

Introduction:

The tightness of the in vitro BBB model is determined by measuring the permeability of molecules with different sizes. The BBB should only be permeable to very small molecules ~70 Da, therefore larger molecules like Mannitol (182 Da) and sodium fluorescein (FLU) (376 Da) should not be able to easily pass through the BBB. Mannitol permeability has previously been described (see e.g. example 5).

Material and Methods:

Passive permeability of the PBECs cultured in either a non-contact co-culture with porcine astrocytes or in a porcine triple-culture was analyzed by the addition of 1 Mg/ml FLU in the upper chamber of a culture insert, in order to investigate the amount of detectable fluorescence in the bottom chamber following a one hour period of incubation. The passive permeability was performed on three biological replicates ( n = 3). Each replicate with a previously measured TEER value was washed once in PBS and then transferred to a 12well plate containing 1.5ml Ringer-Hepes buffer in each well. The insert was then added 1ml Ringer-Hepes buffer with the calculated amount of each transport molecule and incubated on a rocking table at 37 °C for 15 min. After the first 15 min of incubation, each insert was transferred to a new well containing 1.5ml Ringer-Hepes buffer and again incubated for 15 min until one hour had passed. After 60 min, the insert was transferred once more to a well with buffer constituting the final step. Afterwards samples were diluted 1 : 100 in Ringer-Hepes buffer and loaded in black 96well plates (ΙΟΟμΙ pr. well) before analyzed using fluorescence spectrometry (Enspire, Perkin Elmer).

Data was accumulated in Microsoft Excel using a method developed by Culot et al. (Lens, France) and mean Papp (Cm-s ^_1 ) ± SEM was plotted using GraphPad Prism 5.0 software.

Results and discussion:

Both PBECs cultured in co-culture and in triple culture had a low permeability to FLU. The PBECs cultured in co-culture had a higher permeability to FLU

(8.67±2.42*10 ⁶ Papp (Cm-s ¹ )) than the PBECs cultured in triple culture

(5.69±0.53*10 ^"6 Papp (Cm-s ¹ )) (Figure 7). This indicates that the PBECs cultured in triple culture are tighter than the co-culture. Conclusion:

The porcine in vitro BBB has a low permeability to FLU, which has a weight of 376 Da. The PBECs cultured in triple porcine culture are less permeable compared to the non-contact co-culture.