EP0782704A1

EP0782704A1 - Cell separation

Info

Publication number: EP0782704A1
Application number: EP95930638A
Authority: EP
Inventors: John Ugelstad; Per Stenstad; Lars Kilaas; Arvid Berge; Wenche Slettahjell Prestvik
Original assignee: Sinvent AS
Current assignee: Sinvent AS
Priority date: 1994-09-06
Filing date: 1995-09-05
Publication date: 1997-07-09
Also published as: AU3395595A; WO1996007909A1; JPH10505423A; CA2199020A1; GB9417864D0

Abstract

The present invention provides the use of a milk protein to block the non-specific binding of cells to a solid phase selective cell-capture matrix. Generally speaking, the solid phase selective capture matrix will be a solid support adapted for the selective separation of target cells by means of a target-specific binding partner such as an antibody.

Description

Cell Separation

The present invention relates to a method for blocking the non-specific binding of cells in selective cell isolation procedures.

The separation and selective isolation of cells and other viable entities such as viruses represents an area of particular, and indeed, increasing importance in medicine and biochemistry and similar related fields, both from a preparative point of view and as a tool for research. Thus for example, cell separation is a necessary step in the purging of bone marrow of malignant cells or to obtain a purified population of a particular cell type, for example for reconstitution or for subsequent investigation or manipulation, eg. by culture or by the introduction of genetic material. Cell separation is also frequently used in the diagnosis of diseases and infections, for example by isolating pathogenic cells such as bacteria or protozoa or viruses from clinical samples, and in the monitoring of bacterial or other pathogenic contamination, for example in food or waste water. Cell separation thus has .implications in many important areas eg. the treatment and research of serious diseases, diagnosis, blood and tissue typing, food and environmental safety etc.

Many methods for the separation of cells are known. In recent times, these have coirimonly been oased on the selective capture of the desired cell by means of an affinity binding system, that is by means of a pair of binding partners separately attached to the cell to be separated and to some form of support, and which bind when brought into contact thereby enabling the cell to be selectively removed from its environment. This is generally accomplished using an antibody as the target specific binding partner, recognising an antigen on the cell surface, and more recently with magnetic beads as the support thereby enabling ready separation of the bound cells by magnetic aggregation (so-called "immunomagne ic separation") . The use of selective affinity-capture matrices, and particularly magnetic beads, for the selective separation of cells and viruses has been described extensively in the literature (for reviews, see for example Ugelstad et al.. Progress of Polymer Science, 17(1), 87-161, 1992; and Olsvik et al.. Clinical Microbiology Reviews, 43-54, January 1994) .

A problem with this form of separation is the non¬ specific binding of unwanted cells to the support resulting in the separation not only of the desired target cell which binds to the support by means of the target specific binding partner, but also other, undesired cells which bind non-specifically to the surface of the support. The degree of non-specific binding may vary with the nature of the cells, support and affinity-binding system used, but generally speaking, this is a problem with all solid phase affinity capture based-systems, and particularly immunomagnetic separation. This can have potentially serious consequences, for example where it is important that only the desired cell is isolated eg. in bone marrow purging or in typing, and may lead, for example to confusing results in diagnostic systems.

Attempts have been made to reduce the non-specific adhesion of cells to supports by pre-treating the support with bovine serum albumin (BSA) to block the sites at which non-specific binding of cells may occur. A similar effect may be achieved using inert immuno- globulin, that is non-target specific immunoglobulin. However, the reduction of non-specific binding which is attainable using such methods is not complete and in many cases not sufficient to alleviate the problem . Thus, there exists a need for an improved method of blocking such non-specific binding of cells to selective cell capture matrices. The present invention seeks to provide such a method.

Thus, we have now surprisingly found that such non¬ specific cell binding may be drastically reduced, and in many cases completely prevented, by using a milk protein such as casein as blocking agent. Casein has previously been used in the treatment of solid surfaces, eg. polystyrene to prevent physical adsorption of proteins. Such utility has most notably found application in immunoassay procedures such as ELISAs, where plates are blocked by exposure to casein solutions (see for example Vogt et al. , J . Immunol. Methods. 101. 43-50, 1987), but other uses, for example in Western and Southern blotting, have also been described (see e.g. Johnson e_L al. , Gene Anal. Techn. 1, 3-8, 1984). However, it has not previously been proposed that casein might be effective to prevent the non-specific adhesion of cells to solid supports.

In its most general sense, the present invention thus provides the use of a milk protein to block the non-specific binding of cells to a solid phase selective cell-capture matrix.

Alternatively viewed, the invention also provides a method of blocking non-specific binding of cells to a solid phase selective cell-capture matrix, said method comprising adding a milk protein to said matrix.

Generally speaking, the solid phase selective capture matrix will be a solid support adapted for the selective separation of target cells by means of a target-specific binding partner, eg. the solid support may carry a target-specific binding partner.

As mentioned above, such blocking is particularly important in the selective isolation of cells from mixtures where non-specific binding of unwanted cells may be a problem.

According to a further aspect, the present invention thus provides a method of selectively isolating a target cell from a sample, comprising selectively binding said target cells to a solid support by means of a target-specific binding partner and separating said support-bound cells from the sample, characterised in that a milk protein is added to block non-specific binding of non-target cells.

The term "cells" is used herein to include all prokaryotic and eukaryotic cells and other viable entities such as viruses and mycoplasmas, and sub- cellular components such as organelles. Representative "cells" thus include all types of mammalian and non- mammalian animal cells, plant cells, protoplasts, bacteria, protozoa and viruses.

The term "blocking" as used herein includes reducing as well as preventing said non-specific binding. As mentioned above, the reduction in non¬ specific binding obtainable by the present invention is much higher than could have been predicted and significantly better than that obtained with BSA and other proteins. Indeed, treatment with milk protein of solid supports already pre-treated with BSA, leads to a marked reduction in non-specific binding compared to supports treated with BSA alone.

The milk protein may be any of the milk proteins known and described in the literature for blocking of non-specific adsorption in immunoassays, including for example, casein, whey protein concentrate, non fat dry milk and skim milk. Any of those milk proteins may be used, alone or in combination. Casein represents a convenient choice. Also included are derivatives of milk proteins such as partial hydrolysates or cross- linked proteins.

The solid support may be any of the well-known supports or matrices which are currently widely used or proposed for immobilisation, separation etc. These may take the form of particles, sheets, gels, filters, membranes, or microtitre strips, tubes or plates etc. and conveniently may be made of a polymeric material. Particulate materials eg. beads are generally preferred, due to their greater binding capacity, particularly polymeric beads, a wide range of which are known in the art. To aid manipulation and separation, magnetic beads are preferred. The term "magnetic" as used herein, means that the support is capable of having a magnetic moment imparted to it when placed in a magnetic field, and thus is displaceable under the action of that field. In other words, a support comprising magnetic particles may readily be removed by magnetic aggregation. Preferably such magnetic particles are superparamagnetic to avoid magnetic remanence and hence clumping, and advantageously are monodisperse to provide uniform kinetics and separation. The preparation of superparamagnetic monodisperse particles is described by Sintef in EP-A-106873. The monodisperse polymeric superparamagnetic beads sold as DY ABEADS by Dynal AS (Oslo, Norway) are exemplary of commercially available magnetic particles which may be used or modified for use according to the invention.

Non-specific binding of cells to such supports occurs by physical adsorption of the cells and has been observed with all types of solid surfaces which may be used in cell separation procedures, notably polymer supports, including both hydrophilic and hydrophobic surfaces and all such supports are included within the present invention. There have been several reports however, that the problem may be exacerbated with hydrophobic supports. For a number of reasons hydrophobic surfaces are favoured for the coupling of specific binding partners such as antibodies (or at least that the surface be hydrophobic at the time of the coupling reaction) . The reason for this is that the binding of antibodies to supports by chemical coupling in some cases is a slow reaction. A spontaneous and high degree of adsorption of the antibody to the support surface results in a very large increase in the concentration of the antibody at the reaction site and thereby leads to a higher rate of chemical coupling. The use of relatively hydrophobic surfaces therefore strengthens the need for effective blocking of non¬ specific adhesion of cells to the surface of the support.

As mentioned above, the advantage of milk protein in hindering non-specific binding of cells is clearly and unambiguously noticeable with any type of support surface, including rather hydrophilic surfaces but it is especially marked with hydrophobic surfaces, where casein is dramatically better than BSA when using optimal amounts of the latter.

The target-specific binding partner may be any grouping capable of recognising and binding to the target particle and conveniently may comprise any such binding partner as is conventionally used in separation and immobilisation techniques. In this context the binding partner may comprise a single entity capable both of binding selectively to the target cell and to the support, or a series of entities, comprising at one end a component capable of binding selectively to the target cell, and at the other end a component which binds to the support. At its simplest, such a series may comprise a primary binding partner recognising the target cell and a secondary binding partner, binding to the primary partner and to the support. Typically the binding partner will comprise an antibody, or antibody fragment, recognising an antigen on the surface of the cell, virus particle etc. The antibody may be mono- or polyclonal or chimaeric and may be used in the form of a fragment which retains binding activity, eg. F(ab)₂, Fab or Fv fragments (the Fv fragment is defined as the "variable" region of the antibody which comprises the antigen binding site) . Alternative binding partners include proteins such as avidin, lectins, or ligands binding to receptors on the surface of the cells. Depending on the particle desired to be separated, and the environment from which it is desired to separate it, the binding partner may be chosen to recognise selectively surface epitopes specific to the particle, eg. surface antigens expressed only by a particular type of cell, or the binding partner may be of more general reactivity eg. capable of recognising a range of cells.

Because of their selectivity and ready availability antibodies and their fragments are generally the preferred binding partner, particularly IgG and IgM antibodies. Monoclonal antibodies can readily provide desired target specificity. Thus, such antibodies or their fragments may be coupled directly to the support, for binding to the target cell, or they may be coupled indirectly, as mentioned above, as part of a binding partner series. Conveniently, such indirect coupling may be via a secondary antibody which is itself bound to the support and which binds to the target-specific, primary antibody at a site, eg. the Fc portion, which leaves the primary antibody free to bind the target. Alternatively, such a secondary antibody may be replaced by an immunoglobulin-binding protein eg. protein A. The support may carry functional groups such as hydroxyl, carboxyl or amino groups which by well known methods may be activated to allow covalent binding of antibodies or of suitable ligands for attachment of antibodies. Examples of suitable ligands are avidin or streptavidin which may be coupled directly to the support for subsequent binding with a biotinylated target-specific binding partner. The support may also carry groups such as epoxy or aldehyde groups, in which case no activation is needed for establishment of covalent coupling of the antibodies or ligands to the support.

Many cell specific antibodies are known and commercially available. Many such antibodies and their sources are listed in Linccott's Directory (available from 40 Glen Drive, Mill Valley, California, USA) . As representative of such antibodies may be mentioned, antibody B1-3C5 against human pluripotential precursors, available from Sera Lab Ltd, Sussex, UK. ,* antibody BU10 against human dendritic cells, available from Binding Site Ltd, Birmingham, UK; antibodies B721 and L243 against human HLA DR and DR histocompatibility antigens, available from Becton Dickinson Immunocytometry Systems, CA, USA; antibody MAB1273 against mitochondria, available from Paesel GmbH, Frankfurt, Germany,* antibody CA14-50 against Candida albicans. available from Chemunex S.A, Maisons Alfort, France; antibody HBC170-4 against the hepatitis B virus core antigen, available from Biosoft, Paris, France; and antibody 4D2 against Staphylococcus aureus. available from Biodesign Inc. Maine, USA.

As a further example a large number of antibodies are available against specific markers expressed on cells of the haematopoietic system (see for example the range of antibodies against CD antigens available from Dako, Copenhagen) . In addition, the anti-CD34 antibodies 12.8 and B1-3C5 useful for selecting early haematopoietic cells are available from Biosys S.A, France.

The literature also contains descriptions of numerous antibodies suitable for selecting infectious agents ^'including bacteria, protozoa and viruses. Thus for example, antibodies against the K88 (F4) fimbrial antigen of E.Coli are described by Lund et al in J.Clin.Microbiol. 2£.-2572-2575; Skjerve and Olsvik used a commercial polyclonal goat IgG in the immumagnetic separation of Salmonella (J.Fod Microbiol. 14.-H-18, 1989) ,* sero-group specific monoclonal antibodies against Salmonella are described by Widjojoatmodo et al.. (Eur. J. Clin, Microbiol.Infect.Dis.10:935-938, 1991); Skjerve et al. , describe a monoclonal antibody against Listeria monocytogenes (Appl. Environ. Microbiol. 5_£:3478:3481) ,- Morgan et al. describe a monoclonal antibody against Pseudomonas putida (Appl. Environ.Microbiol. 67:503-509. 1991) ,- and antibodies against the CD4 antigen, useful for detecting HIV-infected CD4 cells are described by Brinkmann et al. (J.virol. £≥.:2019-2023, 1991) .

Alternatively, poly or monoclonal antibodies of the desired specificity may be obtained using standard techniques.

The binding partner may be chosen to bind wanted or unwanted particles ie. to achieve positive or negative selection, for example either to isolate a desired population of cells or to purge unwanted particles from a system. The method of the invention has been found to be particularly useful in the positive selection of various desired cell populations, for example from blood, plasma or other body fluids or clinical samples, or from cell culture or other biological or artificial media etc.

Many methods are known for attaching binding partners such as antibodies and their fragments to supports to provide affinity capture matrices and are widely described in the literature. The supports may be provided with a range of functional groups which may be activated to give a covalent bond between the support and the antibody by reaction of the activated group with amino or SH groups in the antibody. The antibody so bound may, as mentioned above, be a target-specific antibody, or it may be a secondary antibody, binding the primary, target-specific antibody. Examples of the most common methods are: 1. Coupling of antibodies to supports containing -CH₂-0H groups by activating with sulphonyl chlorides which gives sulphonyl esters on the supports which in turn react with amino groups or -SH groups on the antibody to give covalently coupled antibody with -CH₂-NH- and -CH₂-S- bonding respectively. 2. Coupling of antibodies to supports containing COOH groups by activation of the carboxylic groups with carbodiimide and N-hydroxysulphosuccinimide whereby amide bonds are formed between the support and antibody. 3. Coupling of antibodies to supports containing amino groups which have been activated with glutaraldehyde and thereby react to form covalent bonds with the amino groups of the antibody. 4. Coupling of antibodies to supports containing epoxy groups takes place without further activation as the epoxy groups react directly with amino groups and -SH groups on the antibody. Analogous methods may be used to couple other binding partner proteins.

The above-mentioned magnetic beads or other supports may be provided with a range of functional groups for attachment of the binding partner, (eg. hydroxyl, carboxyl, aldehyde, epoxy or amino groups) or they may be modified, eg. by surface coating, to introduce a desired functional group. US-A-4654267, 4336173 and 4459378 describe the introduction of many such surface coatings.

The binding of the target cells to the support may take place by a number of steps. For example all the reagents, including the specific binding partner may be bound stepwise to the support, which is then contacted with a sample containing the target cells. Alternatively the target cell population may be contacted, in a separate step, with the specific binding partner e.g. a monoclonal antibody, before exposure to an appropriate insoluble support. Thus, different parts of the linkage, particularly where this is a complex indirect linkage, may be constructed on each of the target cell and support, before being brought together for binding.

One preferred method is to bind an antibody or other target-specific binding partner to the support, prior to contacting with the target cells. Alternatively an antibody (as target-specific binding partner) may be bound to the target cells in a first step. In a separate step, a secondary antibody capable of binding to the Fc region of the primary antibody is attached to the support, following which the cells and support are brought into contact.

A particularly advantageous coupling system for selectively binding a target cell to a solid support is described in our co-pending international patent application No. W094/20858. This is based on a hydroxyboryl/cis-diol linkage between a target-specific binding partner and the solid support. As described above, the linkage may take the form of a binding partner system comprising a series of entities, and either the cell-recognising primary target-specific binding partner, or a secondary binding partner, binding the primary partner is attached to the support in this . manner ie. a primary or secondary antibody may be the subject of the hydroxyboryl/cis-diol linkage.

The hydroxyboryl-based system has been found to be particularly effective in the coupling, and subsequent release of IgM antibodies and the use of an IgM binding partner coupled to a support by means of a hydroxyboryl/ cis-diol linkage, represents a preferred aspect of the invention.

We have surprisingly found that where binding partners are coupled to a support by virtue of a hydroxyboryl-based linkage, a favourable orientation of the binding partner on the support is obtained, without detracting from its selective binding. This leads in turn to highly efficient and reliable binding of the target cell to the support. As will be discussed in more detail below, an additional advantage of the hydroxyboryl-based system is that under certain conditions the binding of the hydroxyboryl/cis-diol residues may readily be reversed under mild conditions, thereby liberating the target cell in a simple and non¬ destructive manner. However, although from the above- mentioned point of view, the hydroxyboryl/cis-diol linking method is advantageous, it does seem to suffer particularly from the problem of non-specific cell binding, which appears to be especially marked with this method. In particular, marked non-specific binding of B cells and monocytes has been observed, threatening severely to limit or complicate the use of the technique where such cells are present. It has now been found that the use of milk protein according to this invention is particularly effective to block non-specific cell binding with solid supports provided with hydroxyboryl- based linkages and this represents one particularly preferred aspect of the invention.

In cell separation procedures, it may or may not be desirable to detach the cells from the support following separation, depending on the target cell and the application. Thus, in some cases removal of the support is necessary, for example in the isolation of pure cell fractions for clinical use or for functional studies, whereas in other cases (eg. purging or other negative selection procedures) there is no need to do so. A range of detachment methods are known in the art, and different methods may be appropriate for different supports. Where the specific binding partner is an antibody, the DETACHaBEAD system of Dynal AS, Oslo, Norway may be used (see eg. Rasmussen et al. in J. Immunol. Methods, 146, 195-202, 1992). Other detachment methods include competitive displacement reactions using eg. excess free antigen, or cleavage of disulphide linkages using thiol reagents such as dithiothreitol. Hydroxyboryl-based linkages may be broken simply and effectively by adding a competing cis-diol reagent as described in WO94/20858.

The time of addition of the milk protein, and the length of time the support is exposed or incubated with it will depend on the nature of the support, and linkage and also whether or not it is desired to liberate the target cells following separation. In all cases, the addition of the milk protein must be such as to permit a sufficient linkage of the binding partner to the support to take place,* in addition to preventing non-specific cell binding the milk protein will also have the effect of reducing binding of the binding partner to the surface. Generally speaking therefore, the milk protein is added after the binding partner is linked or bound to the support. Where the binding partner is a multi- component system, this will generally be after the support-binding component eg. the secondary antibody in the systems described above, is bound.

Some milk protein may be added, if desired, during the binding partner attachment step, but generally speaking, the main portion is added after the binding partner (binding partner component) is bound. The inclusion of additional protein during the binding partner attachment step may be advantageous in promoting binding of the binding partner in the correct orientation, and milk protein may be used to provide such additional protein.

As explained above, proteinaceous binding partners such as antibodies tend to form hydrophobic interactions with hydrophobic surfaces which strengthen their linkage to the support. This may be advantageous in cases where subsequent detachment is undesired or where it is particularly desirable to avoid leakage of the binding partner from the support eg. in typing reactions. Thus, in such cases it may be desirable to avoid exposing the support to the milk protein for a longer period of time after the coupling reaction, to ensure that a stronger hydrophobic bond is formed. In such a situation the blocking step of incubating the support with milk protein may conveniently take place at any time sufficient for establishing a sufficiently strong linkage between the binding partner and the support, without any upper time limit. If, however, subsequent detachment of the binding partner from the support is required, it may be preferable to carry out the blocking step as soon as the binding partner has been sufficiently firmly linked to the support, but before a too strong hydrophobic interaction between the support and the binding partner has been established which will tend to lower the degree of detachment which may subsequently be obtained. This time varies strongly with the type of support and the type of binding partner. In principle, the blocking agent may be added at any time after that the binding partner has been firmly linked to the support, which may vary from 15 minutes to 24 hours or longer. As an example, the milk protein blocking agent may be added after 1 to 4 hours.

After the blocking agent has been added, the support with binding partners attached may be stored for any length of time prior to use, especially if no detachment of the binding partner is desired. For systems where at some stage release of the binding partner from the support may be desired, as for instance in the selective capture of cells, it may in some cases be preferable to use the support as soon as possible because of possible structural changes in the binding partner which may lead to non-reversible binding. However, the presence of milk protein at the surface of the support will diminish the hydrophobic binding of the binding partner to the support, so that a sufficient release of the binding partner, eg. antibody, from the support may be achieved even after several weeks of storage at 4°C.

Generally speaking, the support is stored in the presence of the blocking agent, and washed immediately prior to use. Washing is desirable, to remove any binding partner which may have leaked from the support.

Temperatures and other conditions during the blocking step are not especially critical and conventional laboratory procedures and conditions may be used. A temperature of 4°C during the blocking step has been found to be convenient. The concentration of milk protein used may also be varied according to circumstance, and choice. Concentrations of 0.01 to 0.1 % w/v are for example suitable, more particularly 0.05 to 0.1 % w/v.

As mentioned above, the method of the invention may be applied to the isolation of any prokaryotic or eukaryotic cells, or other viable entities from biological or artificial media including whole blood, buffy coat and cell suspensions obtained by density gradient centrifugation.

The various reagents and materials required to perform the methods of the invention may conveniently be supplied in kit form.

A further aspect of the invention thus provides a kit for use in the method of the invention, comprising:

(i) a solid support as hereinbefore defined;

(ii) a target-specific binding partner capable of binding to said support and a target cell, as hereinbefore defined,* and (iii) a milk protein.

The cell separation method of the invention may have many uses, for example in bone marrow purging, depletion of normal T-cells in allografts, isolation of stem cells e.g. for reconstitution, isolation of pure cell sub-populations for functional studies, tissue typing, and diagnosis, for example detection of bacterial pathogens.

Non-specific cell binding may be a problem in such cell separation procedures as described for example in the situations below:

1. Contamination of target cells with non-target cells in direct positive cell selection steps. Such contamination with non-target cells is especially a serious problem in cases were the target cells are present in low concentration, as for instance in the positive separation of stem cells. The concentration of stem cells is usually so low that even a relatively low amount of other cells bound non-specifically to the beads may represent a larger percentage based on the isolated stem cells.

2. Loss of target cells when isolating target cells by negative removal of cells. This method of selective cell separation involves removal of all non-target cells from the cell suspension by means of a support carrying a binding partner directed against antigens on the non- target cells which one wants to remove. Non-specific binding of target cells represents a loss of target cells which in some cases may be significant. As a particular example may be mentioned the removal of parts of the stem cell fraction in the bone marrow when purging bone marrow for cancer cells in connection with bone marrow separation. In this case, the marrow which has been cleaned for cancer cells by immunomagnetic purging is reinjected into the patient. A successful transplantation result is dependent upon the presence of a sufficient number of active stem cells in the rein used bone marrow. Therefore, a high non-specific binding to the magnetic particles of stem cells during the process of cancer cell depletion is unwanted.

3. Contamination of isolated bacteria or viruses with cells and non-target bacteria in the immunomagnetic isolation of target bacteria. The presence of contaminants in the isolated target bacteria may complicate the subsequent determination of the isolated target bacteria.

The invention will now be described in more detail below, with reference to the following, non-limiting Examples. EXAMPLES

The effect of different substances and methods for reducing non-specific binding of cells to the particles may be demonstrated in different ways.

Below we give results of non-specific binding of cells to various particle surface obtained by different methods.

Example 1

NON-SPECIFIC BINDING OF CELLS BY DIRECT MEASUREMENT OF THE ATTACHMENT OF CELLS TO MAGNETIC PARTICLES

A direct method which in a simple but consistent way may be used to demonstrate the effect of blocking agents on the non-specific attachment of cells to the magnetic particles is based on measurements of cells which are so strongly attached to the magnetic beads that they are transported along with a large given number of the beads when these under given conditions were moved by a magnet.

This procedure was carried out as follows: A suspension of mononuclear cells obtained by gradient centrifugation of a buffycoat and containing 20 x 10⁶ cells per ml was incubated with magnetic particles in an amount corresponding to 3 particles per cell. After 30 minutes of incubation at 4°C with gentle rotation, the tubes containing the particles and the cells were diluted 10 times with a PBS buffer, pH 7.4, and the particles with any cells attached to them were separated from the suspension with a magnet. The number of cells attached to the beads were estimated in a fluorescence microscope using a Burker haemocytometer after staining with acridine orange and ethidium bromide. The number of cells attached to the particles were expressed as % of total cells in the suspension. Although, as expected, the absolute number of cells attached to the beads varied from one cell suspension to another the relative values of non-specific binding of cells on the different particle surfaces were remarkably consistent and in all cases gave a clear preference for casein as the most effective blocking agent in preventing non¬ specific binding of cells.

Casein was also applied to improve particles which had previously been covered with BSA. In this case the casein was used as an aftertreatment of particles which were previously treated with BSA. The efficiency of this treatment was estimated as described above.

The type of beads which were investigated for non¬ specific binding of cells by the method described in Example 1 were:

a. M-450 particles (DYNAL AS)

These particles which have been described in several papers (e.g. J. Ugelstad et al. in Progress of Polymer Science, 87-161, 1992) are magnetic particles covered with an epoxy resin and have a moderately hydrophilic surface^'. These particles were used for direct measurements of physical adsorption of cells, both naked and after treatment with BSA and casein. In short time experiments where the measurements of non-specific binding of cells were obtained after one hour attachment of protein, the measurements of cell addition may be considered as taking place on particles having the protein physically adsorbed. In cases when the particles where treated for a day or more with the proteins before measurements of cell adhesion the results may be considered as a case of adsorption of cells to particles covered with covalently bound proteins .

b. Particles with BSA covalently attached to the surface

These particles were made by incubation of M-450 particles with BSA solution containing 0.1% BSA over 20 hours at room temperature. Non-specific adsorption of cells to these particles were measured directly and after treatment of the particles with 0.1 % casein solution at 4°C for 1 hour at pH 7.4. Contrary to the procedure given in point a) the particles were in the present case always covered with chemically bound BSA before addition of casein and the measurements thus shows the effect of casein in reducing the non-specific binding of cells to particles covered with BSA

c. Particles with phenyl groups attached to the surface.

These particles which were meant to represent a type of support having a relatively highly hydrophobic surface were made as follows:

Dynabeads M450, 4.5 μm magnetisable beads (Dynal AS, Oslo, Norway) (10 g) carrying free epoxy groups were reacted with bis- (3-aminopropyl) a ine (100 g) for 5 hours at 70°C to yield surface -NH₂ groups on the beads. The beads were then washed seven times with 500 ml diglyme (diethyleneglycol dimethyl ether) . The beads (10 g) were then reacted with glycidyl methacrylate (200 g) at 70°C for 20 hours to yield surface vinyl groups. After washing the beads 7 times with 500 ml acetone, 10 g of the beads were reacted with 50 g acrylic acid and 2g of AIBN (azobisisobutyronitrile) dispersed in 300 g isopropanol, at 70°C for 20 hours, leading to copolymerisation and formation of surface -COOH groups. The -COOH beads (0.5 g) were dispersed in 7.5 ml diglyme, and a solution of 2-ethoxy-l-ethoxycarbonyl- 1,2-dihydrochinolin (EEDQ) (0.5 g) in 7.5 ml diglyme. After 10 minutes of stirring at 20°C, 1 ml of aniline was added, and this suspension was stirred 16 hours at 30°C. Then the beads were washed five times with 50 ml diglyme, once in 50 ml methanol and at last three times in 50 ml distilled water.

Non-specific cell adhesion to these particles were measured as described above. The non specific adhesion of cells was measured with various concentrations of proteins and with various incubation times for the pretreatment of the beads with protein solution.

d. Particles with boronic acid groups on the surface. These particles were made according to a procedure described in International Patent Application No. PCT/GB94/00473. The COOH beads described above (5 g) were washed once with 100 ml 0.1M phosphate buffer pH 7.3. Then the beads were redispersed in 50 ml of the same buffer, and a solution of 3-aminophenyl boronic acid (3 g) in 0.1M phosphate buffer ((50 ml) was added. A solution of L-ethyl-3 (3-aminopropyl) carbodiimide hydrochloride (EDO (3 g) and N-hydroxysulfosuccinimide sodium salt (0.5 g) in 0.1M phosphate buffer pH 7.3 (100 ml) , was added dropwise (20 ml/min) under stirring at 10 °C. After additional stirring for 5 minutes, the temperature was raised to 20°C and the particle suspension was stirred at 20°C for 20 hours. Then the beads were washed; twice with 150 ml 0.1M phosphate buffer pH 7.3, twice with 150 ml 1M NaCl, twice with 150 ml 0.1M carbonate buffer, pH 10.5, 1 hour under stirring for each wash. Finally the beads were washed three times with 200 ml distilled water. Measurements of non¬ specific adhesion of cells to the particles after treatment with different proteins were carried out in the same way as described above for particles under a,b,c and d. e. Particles with BSA or casein coupled to the surface by help of glutaraldehyde.

Attachment of BSA or casein to the surface of the magnetic beads by help of glutaraldehyde was carried out as follows: M450 beads (100 mg) were dispersed in 2 ml PBS buffer, pH 7.4, containing either 0.1% BSA or 0.1% casein, and incubated for 20 hours at 20°C. Then the particles were isolated by a magnet and washed whereafter 2 ml of a glutaraldehyde solution (0.5 M) in 0.1 M phosphate buffer, pH 7.7, was added. This reaction took place at 20°C for 20 hours. After isolation and washing, 2 ml PBS buffer, pH 7.4, with either 0.1% BSA or 0.1% casein was added. After incubation for 20 hours at 20°C the beads were again isolated and washed. Then the beads were incubated for 4 hours at 20°C in a PBS buffer, pH 7.4, (2 ml), containing 0.1 M NaBH₄, in order to transform aldehyde groups into hydroxy groups. Thereafter the beads were washed 6 times in 2 ml distilled water. These particles could also be directly prepared by crosslinking BSA or casein which previously was attached to magnetic particles with a hydrophobic surface.

The results of the measurements of non-specific binding of cells to the particles described under a, b, c, d and e are given in Table i.

TABLE I NON-SPECIFIC ADHESION OF CELLS

Treatment None BSA Casein

Concentration 0.1 1.0 0.01 0.1 (%)

Incubation time 1 20 20 1 1 20 (hours) a. M450 15.0 6.0 4.5 2.8 2.4 2.2 1.8 b. 450 /BSA 4.5 1.9 c. Phenyl groups 19.0 14.7 11.3 6.0 3.0 2.5 d. "Bor" particle 42.0 29.0 26.0 15.0 8.0 5.4 4.0

Coupled with BSA Coupled with casein

e. Glutaraldehyde 9.8 !> 4.5 1

The values given are % of total cells attached to the beads, and are average values based upon more than 5 experiments.

1) Average values of 2 experiments

Table I reveals that the advantage of casein as compared to BSA in decreasing the non-specific binding of cells is most pronounced for the case of particles with boronic acid coupled to the surface.

Also a comparison of the a) and c) particles in Table I indicates that the advantage of casein is more pronounced in the case of hydrophobic surfaces than for the more hydrophillic ones. It is noteworthy that one by use of casein as blocking agent may obtain a decrease of the non-specific binding of the hydrophobic c) particles (particles with phenyl groups) which make them less susceptible to adhesion of cells than may be obtained with the far more hydrophilic M450 particles treated with BSA. Example 2

NON-SPECIFIC BINDING IN THE CASE OF POSITIVE SELECTION OF CELLS AS ESTIMATED FROM CONTAMINATION OF THE TARGET CELLS WITH INDIFFERENT CELLS IN THE CELL FRACTION DETACHED FROM THE PARTICLES

The particles used for measurements of non-specific cell contamination in target cells released after positive cell separation were particles with boronic acid covalently bound to the surface. The preparation of these boronic acid particles for use in positive cell separation was similar to that described in Example 1 for boronic acid particles used for direct measurement of the non-specific adhesion of cells. Positive separation of cells applying these particles has so far been carried out for T4 cells, T8 cells B cells and monocytes. The effect of casein as compared with BSA in preventing non-specific binding of cells and thereby contamination of non-target cells in the cell fraction obtained after removal of the particles from the isolated cells is described in detail for the case of T4 cells. Similar results were obtained by positive isolation of the other cells given above.

Coupling of anti-CD4 antibodies to thc» "Bor-particles"

"Bor-particles" prepared as described were incubated with an IgGl, anti-CD4 monoclonal antibody, ST4 (Biosys, Compiegne, France) . The incubation of the particles (10 mg/ml) with antibody (100 μg/ml) took place in a PBS buffer, pH 7.4, at 4°C for one hour. The particles were then isolated with a magnet and subsequently washed once with a PBS buffer. Then the particles were incubated for 1 hour at 4°C in a PBS buffer, pH 7.4, containing either 0.1% casein or 0.1% BSA and subsequently washed twice with the same PBS buffer at 4°C, immediately before use

Preparation of the cell suspension.

Peripheral blood mononuclear cells (PBMC) were isolated from platelet-depleted buffycoat by Lymphoprep (Nycomed Pharma AS, Oslo, Norway) and washed four times with a PBS buffer, pH 7.4, containing 0.6% sodium citrate.

The PBMC suspension was then incubated with Dynabeads M450, coated with monoclonal antibodies against CD14 (5- 10 particles/target cell) (Dynal AS, Oslo, Norway) . The particle concentration was at least 20 x 10⁶ beads/ml. The incubation took place in a PBS buffer, pH 7.4, containing 2% foetal calf serum, at 4°C for 30 minutes. Then the magnetic beads, and thereby the monocytes were removed with a magnet and the PBMC depleted for monocytes was collected.

Isolation of T4 cells.

The boronic acid particles coated with anti-CD4 antibodies were incubated with PBMC depleted for monocytes. The number of particles was adjusted to 10 - 20 particles per target cell, and in addition the particle concentration was kept at 25 x 10⁶ beads/ml. The incubation took place in a PBS buffer, pH 7.4, containing either 0.1 % BSA or 0.1 % casein at 4°C for 30 minutes. After incubation the suspension was diluted 10 times with the incubation buffer. Then the target cells with attached particles and excess particles were isolated by magnetic aggregation by help of an external magne . The isolated cells were resuspended in PBS buffer, pH 7.4, with either 0.1 % BSA or 0.1 % casein at 4°C and again the cells with attached magnetic particles isolated by help of a magnet. This process of isolation and resuspension was repeated four times. Detachment of the beads.

To achieve detachment of the beads from the cells, the isolated cells were resuspended in a PBS buffer, pH 7.4, containing 50% foetal calf serum, 0.3% sodium citrate and 0.2 M sorbitol at 20°C. The tubes with the cells were then placed on a Rock and Roller (Labinco, Breda, The Netherlands) for 2 hours. Then the beads were removed with a magnet. This process results in detachment of 80 to 90% of the cells from the beads. To improve the yield, the beads were resuspended in the detachment buffer and the suspension was pipetted 10 times. Again, the beads were removed by help of a magnet and the cells were collected and added to the cell suspension from the first step.

Cell counting.

The number of cells in the different cell fractions was determined by use of a Coulter Multisiser II (Coulter Electronics Ltd., Luton, England) .

Flow cγtometry.

Fluorochrome conjugated antibodies against CD3, CD4, CD8, CD19 and CD56 (Beeton Dickinson, Mountain View, California, USA) were incubated with cells from the following three cell fractions: PBMC depleted for monocytes, the remaining cell fraction after T4 cell isolation and from the isolated cells depleted for magnetic beads. The cells were analyzed using a FACScan flow cytometer (Becton Dickinson) .

The efficiency of the blocking agents were determined by measurements of the contamination of non-target cells among the cells set free after isolation of cells and removal of the magnetic particles. Results: The number of target cells isolated from the cell suspension was practically equal with BSA and casein as blocking agents. With BSA as blocking agent the cell fraction released from the beads contained 85% of the target cells (T4 cells) . The contamination was mainly B cells which amounted to 13% of the isolated cells.

With casein as blocking agent the cell fraction released contained 98% of T4 cells.

Example 3

REMOVAL OF TARGET CELLS FROM CELL SUSPENSIONS USING BSA OR CASEIN AS BLOCKING AGENTS.

In all cases of cell isolation, in addition to target cells small amounts of non-target cells are removed from the cell suspension. In the case where during isolation of target cells both magnetic particles are removed in excess as well as agglomerates of particles and target cells a noticeably higher amount of non-target cells are co-removed than is the case with magnetic beads with the same surface treatment but with an indifferent antibody on the particles. Thus the isolation of target cells along with the beads clearly implies a higher tendency to non-specific binding of non-target cells than do the beads alone. Removal of target cells was carried out with different particles with different coupling methods for the monoclonal antibodies and with BSA or casein as blocking agents. In the case of M-450 beads the beads were first incubated with monoclonal antibody, ST4, for 20 hours at 20°C. Then the particles were incubated with 0.1% BSA or 0.1% casein for another 20 hrs. Both the beads covered with BSA and the beads covered with casein were used for removing of T4 cells from the same type of cell suspension as described in Example 2. Both beads with BSA and beads with casein were capable of removing more than 95% of the target cells from the cell suspension. The non-specific binding of cells was estimated from determination of non-target cells in the cell suspension before removal of T4 cells and after removal of the T4 cells by use of flow cytometry as described above. In this case also there is a clear difference between the beads with BSA and casein in non¬ specific binding of cells with a clear preference for the latter in giving less non-specific binding of non- target cells.

Similar procedures were applied also for beads which were activated with sulphonyl chlorides for subsequent binding of the antibody. Also in this case there was observed a significant decrease in the removal of non¬ specific cells when BSA is replaced with casein.

In the case of particles which were covered with BSA or casein crosslinked with glutaraldehyde, the IgG monoclonal antibody was coupled to the beads by means of excess aldehyde groups after which the system was treated with sodium borohydride in order to reduce the Schiff bases and the remaining aldehyde groups. Both the particles with BSA and casein were effective in selective removal of target cells. Again the particles with casein tended to give less removal of non target cells along with the target cells.

Claims

1. The use of a milk protein to block the non-specific binding of cells to a solid phase selective cell-capture matrix.

2. A method of blocking non-specific binding of cells to a solid phase selective cell-capture matrix, said method comprising adding a milk protein to said matrix.

3. The use of claim 1 or the method of claim 2, wherein the solid phase selective cell-capture matrix is a solid support adapted for the selective separation of target cells by means of a target-specific binding partner.

4. A method of selectively isolating a target cell from a sample, comprising selectively binding said target cells to a solid support by means of a target- specific binding partner and separating said support- bound cells from the sample, characterised in that a milk protein is added to block non-specific binding of non-target cells.

5. The use or method of any one of claims 1 to 4, wherein the milk protein is casein.

6. The use or method of any one of claims 1 to 5, wherein the solid phase or support is particulate.

7. The use or method of any one of claims 1 to 6, wherein the solid phase or support is magnetic.

8. The use or method of any one of claims 3 to 7, wherein the target-specific binding partner comprises a single entity capable both of binding selectively to the target cell and to the support, or a series of entities, comprising at one end a component capable of binding selectively to the target cell, and at the other end a component which binds to the support.

9. The use or method of any one of claims 3 to 8, wherein the target-specific binding partner comprises one or more antibodies or fragments thereof.

10. The use or method of claim 9, wherein the antibody is IgM, or a fragment thereof.

11. The use or method of any one of claims 3 to 10, wherein the target-specific binding partner is coupled to said support by a linkage comprising a hydroxyboryl/cis-diol bond..

12. The use or method of any one of claims 3 to 11, wherein the target-specific binding partner carries cis- diol groups and is bound directly, via a said hydroxyboryl/cis-diol linkage, to hydroxyboryl groups on the support.

13. The use or method of any one of claims 1 to 11, wherein the target-specific binding partner is indirectly bound to the support by means of a secondary binding partner, binding said primary, target-specific binding partner, said secondary binding partner carrying cis-diol groups and being bound, via a said hydroxyboryl cis-diol linkage, to hydroxyboryl groups on the support.

14. The use or method of any one of claims 3 to 13, wherein the milk protein is added to the solid phase or support after the target-specific binding partner, or a support binding component thereof, is bound to the support. - so -

is. The use of method of any one of claims 1 to 14, wherein the milk protein is used at a concentration of 0.01 to 0.1% w/v.

16. The method of any one of claims 4 to 15, further comprising detaching said separated target cells from said support.

17. The method of claim 16, wherein said target specific binding partner is coupled to said support by means of a hydroxyboryl/cis-diol linkage, and said detachment is achieved by adding a competing cis-diol- containing reagent to cleave said hydroxyboryl/cis-diol linkage.

18. A kit for use in the method of any one of claims 2 to 17, comprising:

(i) a solid support as defined in any preceding claim,-

(ii) a targe -specific binding partner capable of binding to said support and a target cell, as defined in any preceding claim; and (iii) a milk protein.