PT958046E - Cell separation composition, kit and method - Google Patents

Description

DESCRIPTION

FIELD OF THE INVENTION The present invention relates to compositions, assemblies and methods of cell separation.

Background of the Invention Cellular and cellular component separation has long been an important role in basic research and diagnostic applications, is increasingly important in a clinical setting. The use of cell transplantation methods, such as bone marrow transplantation, in human therapy, required cellular separation procedures which are not only fast and reproducible, but also produce a viable, safe and non-toxic cell composition. It is also desirable that these isolation procedures be suitable for the isolation of poor cells. For example, following chemotherapy for certain types of cancer, many patients now receive " stem cell transplants " which consist of an enriched fraction of hematopoietic progenitor cells isolated from various tissues, such as bone marrow, peripheral blood or cord blood umbilical. These progenitor cells constitute only about 1% of the total number of cells present in these tissues. Density gradient centrifugation is a popular technique for the separation and isolation of cells and cellular components. This process exploits the phenomenon of cell division in a medium with defined density, according to its fluctuation densities. Such a medium may be a solution or a suspension of microparticles. Examples of media commonly used in density gradient separation include sucrose, dextran, bovine serum albumin (BSA), FICOLL diatrizoate (Pharmacia), FICOLL (Nycomed) metrizoate, Percoll® (Pharmacia), metrizamide and heavy salts such as the cesium chloride. Exposing the cells to many of these media produces a weakened biological function of the cells and / or contamination of preparation by toxic components. For example, BSA and FICOLL are known to cause undesired cell aggregation at a physiological pH. Furthermore, these media are generally not readily sterilized by autoclaving or irradiating in the " final form " - that is, at a concentration, the ionic solution and container that is ready to be used in the isolation of a specific type of cells.

A cell separation medium which has been commonly used in the preparation of blood cell fractions is Percoll® (a trademark of Pharmacia Fine Chemicals). Percoll® is a particle of colloidal silica treated by a curing process to form a coating of polyvinylpyrrolidone on each particle. While Percoll® is quite stable at a physiological pH, there are limitations to its general utility in the isolation of cells for therapeutic purposes. For example the medium is difficult to sterilize, since it is not stable for autoclaving or ionizing radiation after it has been diluted in physiological saline solution. These properties limit the usefulness of the product for clinical applications involving the reintroduction of separate cells into humans, or at any other site requiring sterile material. US 4,927,749 provides an organosilanized colloidal silica (OCS) particle preparation (OCSP) for density gradient separation which overcomes some of the above-mentioned problems. While stable in heat sterilization and ionizing irradiation when diluted in physiological saline solution, when sterilized by ionizing radiation, the OCS particle preparations are toxic to certain rare cell types, such as dendritic cells, natural killing cells, natural suppressor cells, cytotoxic T cells and, in particular, hematopoietic progenitor cells (CD34 +). Since final sterilization by ionizing radiation may be preferred for certain applications, in particular those involving containment of material in plastic consumable containers, the use of this material for clinical applications may be limited. In addition, it has been found that the material causes agglutination of cells or aggregation of cells, resulting in significantly lower yields of these rare cells in functional form. The present invention provides compositions and methods which are particularly suitable for the isolation of " rare " - for example, cell types constituting less than about 1% of the total cells in a population of cells. In particular, the invention includes a colloidal silica cell separation medium which enhances the yield and functional potential of cells, by reducing cell aggregation and cell toxicity. The discovery of the present invention involves the inclusion of a polylactam, such as the polylactam range polyvinylpyrrolidone (PVP), in the solution in which the colloidal silica density gradient material is suspended markedly reduces the aggregation of rare blood cells 3, resulting in better income. A further discovery of the present invention is that the addition of polylactam to the medium prevents the loss of clonogenic potential during the density gradient isolation of certain hematopoietic progenitor cells by reducing the toxicity caused by irradiation of OCS solutions. The invention also provides methods for the isolation of certain cell types which are important in diagnostic and therapeutic methods - including, for example, bone marrow isolated hematopoietic progenitor cells (CD34 +), tumor cells chosen, dendritic cells, natural suppressor cells, (NK), cytotoxic T lymphocytes (CTLs) and natural suppressor cells. The invention also provides for the depletion of cell types that may interfere with a specific clinical outcome, such as CTL.

Summary of the invention

In one aspect, the invention includes compositions for use in cell separation consisting of a cell separation medium based on silanized colloidal silica particles, containing at least about 0.05% polylactam, such as polyvinylpyrrolidone (PVP), present in a concentration between about 0.1 and about 10%. In a preferred embodiment, the separation medium is comprised of a suspension of organosilanized colloidal silica particles. An advantage of this composition is that it can be sterilized either by autoclaving or by ionizing radiation without imparting toxic effects to subsequently separated cells.

Although the particles comprising the medium of the invention may be in the range of about 0.003 to 50 microns, the colloidal silica particles are usually comprised in an average diameter of 10-50 nm.

The compositions of the invention are particularly suitable for use in the isolation of a rare blood cell, selected from a cell mixture. According to methods described herein, in such cases, a means is provided for the particle suspension to have a specific density of about 0.0005 g / ml of said chosen cell. Exemplary rare cell types include CD34 + hematopoietic progenitor cells, dendritic cells, natural exterminating cells, natural suppressor cells, cytotoxic T cells, tumor cells, and nucleated cells of fetal origin.

CD34 + hematopoietic progenitor cells can be isolated from peripheral blood mononuclear cells, in which case the preferred suspension of organosilanized colloidal silica particles has a specific density of 1.0605 g / ml. When CD34 + hematopoietic progenitor cells are isolated from bone marrow cells, the suspension of organosilanized colloidal silica particles preferably has a specific density of 1.0685 g / ml.

Dendritic cells are also isolated using the compositions described above. In this case, the suspensions having densities of 1.0770 g / ml, 1.0720 g / ml, 1.0650 g / ml, 1.0610 g / ml and 1.0565 g / ml may be used. The invention also includes a set for cell separation. The assembly consists of a centrifugal vessel, and a silanized colloidal silica-based cell separation suspension, as described above. 5

A process for sterilizing a silanized silica particle based cell separation medium is also included in the present invention, wherein the separation medium is subjected to autoclaving or ionizing radiation in the presence of about 0.05%, and preferably between about 0.1 and 10%, of polylactam.

More generally, the invention includes a method of isolating a rare blood cell selected from a cell mixture, wherein the mixture is first layered in a cell separation medium based on silanized colloidal silica particles having a specific density of about 0.0005 g / ml of the chosen cell. The mixture is then centrifuged. According to a preferred embodiment, the centrifugation is performed in a so-called " cell trap " (i) side walls and a closed bottom, and (ii) a constriction member disposed within the tube, the constriction tube being placed and constructed to retain fluid in the lower portion of the tube beneath the constriction member, when the tube is inverted. Also, in this configuration, prior to centrifugation the tube is filled with the separating means to a level extending above said constriction member, to a level above an aperture formed by said constriction member, so that the cells that are captured at an interface between the cell separation medium and a lower density medium are discharged after centrifugation with the lower density medium when the tube is inverted. The tube may also include a closed top, a first port at the top through which the fluid material can be introduced into the tube, a second port at the top, and an enclosed fluid channel establishing communication between the second port and the bottom of the tube below the constriction element. 6

This configuration can be used for functional hematopoietic progenitor cells isolated at concentrations of the separation medium of 1.0605 g / ml (from mixtures of peripheral blood cells) and 1.0685 g / ml (from the bone marrow). Tumor cells are also isolated by this method, using a cell sorting medium with a specific density chosen in the range of 1.0490-1.0808 g / ml. Also, dendritic cells can be isolated using separation media having specific densities of 1.0770 g / ml, 1.0720 g / ml, 1.0650 g / ml, and 1.0565 g / ml.

Cytotoxic T lymphocytes can be isolated using a specific density of 1.0720 g / ml, 1.0610 g / ml or 1.0565 g / ml.

The foregoing separation media may be enhanced by adding a suspension which includes a silanized silica particle to which a specific cellular antigen binding molecule is attached. This addition may be used for the depletion of undesired cell types from a preparation, such as T lymphocytes.

Brief description of the figures Figure 1 shows the recovery of functional hematopoietic progenitor cells (colony forming units, CFU) after incubation with 0.5% gamma-irradiated PVP in PBS, density-based gradient material (OCSP) OCS particles subjected to gamma radiation, or density gradient material based on OCS particles subjected to gamma radiation supplemented with 0.5% PVP (OCSP + 0.5% PVP); Figure 2 illustrates the effect of various concentrations of PVP present during OCS particle sterilization (OCSP), on a gamma radiation-induced restriction of hematopoietic progenitor cell (CFU) function; Figure 3 shows the distribution of peripheral blood progenitor cells (PBPC) in interface and pellet fractions in density gradients formed from colloidal silica particles in the absence (OCSP) and presence of 0.5% PVP (OCSP + PVP); Figure 4 shows the distribution of hematopoietic progenitor cells (CD34 +) in interface and fractions of density gradient pellets formed from colloidal silica particles (OCSP) in the absence and presence of 0.5% PVP (OCSP + PVP) ; Figure 5 shows a comparison of nonspecific depletion of density-adjusted cell sorting microspheres (DACS) coated with two types of GAM-IgG preparation peripheral blood mononuclear cells (PBMC); Figure 6 shows a non-specific loss of CD34 + cells with DACS microspheres coated with goat anti-mouse IgG (GAM) or F (ab) 2 GAM-IgG; Figure 7 shows a non-specific loss of CD34 + cells with intact GAM-IgG coated DACS microspheres, in either random or site-specific orientation; Figure 8 illustrates the recovery, in percent, of total cells, CD34 + cells, CTL's and red blood cells (RBC) from the interface + particle combination after centrifugation of bone marrow cells in an OCS gradient; Figure 9 shows the recovery, in percent, of total cells, CD34 +, CTL and red blood cell (RBC) cells from the gradient interface, following centrifugation; Figure 10 illustrates the distribution of the natural suppressor cell activity in fractions with different densities; Figure 11 shows the distribution of natural killing cells in fractions of different density; Figure 12 shows the results of experiments in which the peripheral blood leukocyte layers were depleted of human lymphoma B cells by DACS; Figure 13 illustrates the recovery of CD34 + cells in samples depleted of peripheral blood of the leukocyte layer, of Figure 12; Figure 14 shows a sterilized centrifuge assembly assembled in accordance with the invention; and Figure 15 illustrates a closed system centrifugation device suitable for inclusion in an assembly of the invention.

DETAILED DESCRIPTION OF THE INVENTION The invention encompasses cell density separation means suitable for cell separation, and particularly for separation or isolation to be transplanted to humans. The invention also includes methods for isolating specific types of cells, preferably using the novel media composition. 9

As described in the following sections, the media of the invention were developed in the light of the discovery that the applicants have made the cell toxicity significantly reduced when a solution of a gamma polylactam, such as polyvinylpyrrolidone, is added to a suspension of silica particles organosilanized colloidal (OCS) cells used in density gradient cell sorting.

In particular, the inclusion of polylactam prevents the loss of clonogenic potential in rare populations of hematopoietic progenitor cells that occurs when these cells are suspended in a suspension of sterilized silanized silica particles. In addition, the present invention includes the discovery that the inclusion of polylactams, such as polyvinylpyrrolidone, in the density gradient medium also reduces aggregation and loss of clonogenic potential of haematopoietic progenitor cells occurring without ionizing radiation. I. Definitions " Colloidal silica " refers to an aqueous suspension of colloidal particles formed by the polymerization of monosilicic acid of S102 dissolved in water. " Silanized silica particles " refers to a particulate silica composition to which a silane coating is covalently attached. " Organosilanized colloidal silica (OCS) particles " refers to a colloidal silica composition to which is attached covalently an organosilane coating. U.S. Patent 4,927, 749 describes the preparation of such a composition. " Rare cell populations " refers to cells that make up less than 1% of the total white blood cells 10 (WBC) counted in the blood of a healthy individual. Examples of rare blood cells include CD34 + hematopoietic progenitor cells, which make up about 1% white blood cells (WBC), natural killing cells, natural suppressor cells, dendritic cells and cytotoxic T cells. Also included in this definition are cells from exogenous sources, including fetal cells circulating in maternal blood. " Cell toxicity ", " toxic to cells " and similar phrases refer here to any decrease in cell viability or biological function that is measurable in a cell population. With reference to the hematopoietic progenitor cells described herein, the more common term refers to the decrease in clonogenic potential, as evidenced by the reduced yield of cells capable of forming hematopoietic colonies (CFU).

A " polylactam " is a polymer containing pendant lactam groups (cyclic amide) groups attached to the polymer backbone. The polymer may be a homopolymer or a copolymer. The pendant lactam groups will usually have ring sizes containing from 3-6 carbon atoms. A preferred type of polylactam for use in the present invention is polyvinyl lactam, such as polyvinylpyrrolidone, which has a ring of four carbon atoms.

&Quot; Dendritic cells ", or DC, are immune cells that are negative for expression of CD3, CD4, CD8, CD14, CD16 and CD20, and positive for HLA-DR (i.e., MHC class II) expression. Dendritic cells have a distinct morphology, that is, they are large veiled cells that extend dendrites when cultured in vitro. 11

A " tumor cell antigen " is a molecule, usually a protein, that is exposed on the surface of a specific tumor cell. Examples of antigen from tumor cells include CD9, CD10, CD19 and CD20 present in B cell lymphoma cells; tumor cell antigens that are specific for certain breast tumors include Her2 / Neu and estrogen receptors.

In the context of the present invention, the term " isotonic " means having an osmolality that is within the range tolerated by the cell, usually about 280-340 mOsm / kg H2O, and preferably for human cells, about 280 mOsm / kg H2 O, depending on the type and source of the cell. II. Cell separation media

This section describes various cell separation media that can be used to isolate cells from the cell mix. Preferred cell separation processes described in the following sections include the use of a cell separation material, such as a density gradient material, having a specific gravity between 1.0000 g / ml and 2000 g / ml, preferably between 1.0300 g / ml and 1.2000 g / ml, which is accurate within ± 0.00005 g / ml, preferably ± 0.00002 g / ml of the desired cell-specific gravity. A. Density gradient medium

Preferred density gradient media are colloidal silica suspensions having particles in the range of 0.002 to 50 microns. One useful medium is PERCOLL®, a suspension of colloidal silica particles which is non-silanized and available from Pharmacia Fine Chemicals (Piscataway, N.J.). More preferred is a suspension of organosilanated colloidal silica particles, as described in U.S. Patent 4,927,749 and treated with PVP, as described below.

According to an important feature of the invention, the density gradient solutions should be prepared and adjusted to a predetermined density. In addition, the osmolality should be adjusted in the range of 280 to 320 mOsm / kg H 2 O to preserve cell integrity, and the pH should be in the range of 6.8 to 7.8 and more preferably a pH of 7.4, to maintain a physiologically isotonic density gradient prior to use. The osmolality and pH may vary depending on the specific conditions under which the density gradient separation process is performed. For example, the temperature at which the samples are maintained or centrifuged may necessitate modifications in the osmolality and / or pH of the density gradient material to maintain adequate density. These modifications of osmolality and pH will be apparent to those skilled in the art.

A preferred density gradient material for use in cell separation according to the present invention is a suspension of organosilanized colloidal silica (OCS) particles, as disclosed in U.S. Patent 4,927,749 to Dorn.

In a preferred embodiment, the density gradient material is diluted to a suitable specific density in a physiological saline solution supplemented with polyvinylpyrrolidone (PVP), such as PVP-10 available from Sigma Chemical Co. (St. Louis, MO). According to an important feature, as disclosed below, it has been found that inclusion of at least 0.5% PVP improves the yield of functional rare blood cell types. It has been found that inclusion of at least 0.5% PVP improves the performance of the functional rare blood cell types. In addition, such a concentration of PVP facilitates the sterilization of the OCS density gradient material by ionizing radiation, such as gamma or electron beam radiation. In the absence of PVP, the cells exposed to the medium experience loss of functional capacity in the case of CD34 + cells, evidenced by the loss of capacity to form colonies (reduction of CFU). Generally, the amount of PVP to be added to a suspension of OCS cell separation will depend on the subsequent purpose and treatment of the preparation. To preserve the functional activity of the cells exposed to the OCS preparation for relatively short periods of time, a concentration of 0.5% by weight of PVP is usually sufficient. For higher exposures and exposures to ionizing radiation higher concentrations of PVP may be required, about 3-8% by weight. B. Silanized silica particles

The processes and compositions described herein are applicable to a set of uses of silanized silica particles in conjunction with a cell separation. The particles may be silanized by any of the processes known in the art. US 4,927,749 describes the preparation of OCS particles which are particularly useful in the preparation of compositions used in the present invention. In one embodiment, the OCS particles having a size of 14-18 nm form one between 3-22 nm, and in particular between stable 13-fold suspension which acts as a density gradient medium to separate heterogeneous mixtures of cells based on density.

In addition to forming density gradient materials, the silanized silica particles may also be used in other aspects of cell separation. For example, the particles used in certain positive or negative selection steps can be treated with the processes described herein. These particles are usually silanized silica microspheres or microspheres having diameters up to about 50 microns and conventionally in the range of up to about 1-5 microns. These particles are silanized according to procedures well known in the art, for example using 3-aminopropyltriethoxysilane, (3-iodopropyl) trimethoxysilane, [1-9trimethoxylsilyl) -2- (m- (or p) chloromethyl) phenyl] ethane, or -glycidyl-oxypropyltrimethoxysilane (GPMS). Typically, these particles have specific antigen binding attachments with covalent attachments, such as antibodies, lectins or other cellular ligands. When added to a mixture of cells, the micro-cells bind to cells having a specific antigen or target antigens and modify (increase or decrease) the density of the cells. Polyvinylpyrrolidone (PVP) was added to suspensions of non-silanized colloidal silica particles (Pertoft et al., Exp. Cell Res. 46: 621-623 (1966), Pertoft et al., Exp. Cell Res. 50: 369 (1968), Wolff et al., J. Cell Biol., 55: 579-585 (1972), Pertoft et al., Exp. Cell Res. 110: 449-477 (1977)). However, solution-free PVP has been reported to have undesirable effects on some cell types 15 (Pertoft et al., Exp. Cell Res. 50: 355-368 (1968)). This leads to a development of a thermal curing process which produces a coating of PVP absorbed on the colloidal silica particles, and which is substantially free of free PVP (Pertoft et al., Anal. Biochem 88: 271-282 (1978) ). This composition is commercially available as PERCOLL®.

However, PERCOLL can not be sterilized by conventional means (autoclaving or ionizing radiation) when diluted in physiological saline solution to a final concentration suitable for cell separation. This limits its utility in applications requiring a density gradient material finally sterilized in a physiological medium.

As noted above, it has been found that a preparation based on colloidal silica particles having a covalent organosilane coating has utility in the separation of various blood cells (US Patent 4,927,749). The organosilane coating reduces cellular toxicity and eliminates aggregation of the colloidal silica particles themselves in the presence of a physiological salt and a protein. However, organosilane coated colloidal silica (OCS) particles have certain disadvantages when used to separate or purify specific rare blood cells from blood and other sources. The use of these particles may result in less recovery of these rare cells. This is exemplified below in conjunction with the isolation of an important hematopoietic progenitor cell, antigenically characterized as a CD34 + cell, which may be isolated from blood or bone marrow. Specifically, these rare progenitor blood cells aggregate or otherwise increase in density when suspended in a density material based on OCS particles. In addition, sterilization 16 of a suspension based on OCS particles by ionizing radiation results in a composition that is toxic to these cells. Tests have shown that these results, and the improvements provided by the findings of the present invention are present in the following sections. C. Polylactam protection against radiation toxicity of silanized silica particles There are a number of processes for sterilizing density gradient materials. Density gradient materials based on silica particles may be sterilized by autoclaving or filtering, according to methods known in the art. One process which is particularly useful is ionizing radiation sterilization. This process may be preferred when the bulk density material is to be dispensed into plastic consumable containers, such as in pre-filled centrifuge tubes, bags or syringes, and the like for clinical separations, or when it will be dispensed into any containers sensitive to heat, such as plastic bottles. Unlike other sterilization processes, radiation-based sterilization processes can be performed in the final product container as a one-step process. That is, when it is necessary to sterilize either a container or a liquid or suspension in the container, there is no need to separately sterilize the container itself, as is required when filtering methods are used, and there is no need to transfer the material from one container to another. individual sterilization vessel, which would normally occur with heat sterilization (autoclaving). 17

Techniques for sterilizing medical devices by ionizing radiation are known in the art. The most commonly used and applicable methods of the present invention include gamma radiation and electron beam radiation. Although each process involves a different type of energy source, both methods involve passing the device through a radiation field to achieve a specific dose of radiation to the device. Typically, sterilization of a medical instrument requires a dose between 10 and 30 kiloGray (kGy). However, as described below, when any of these methods are used to sterilize organosilanized colloidal silica, the material becomes toxic to hematopoietic cells, unless a polylactam, such as PVP, is present in the suspension of particles during the radiation.

The polylactams used in the present invention are preferably soluble polymers such as the KOLLIDON® series of polyvinylpyrrolidones (BASF, Ludwigshafen, Germany). These polymers are usually available in polymer molecular weights of between 2000 and 350,000, and are classified by their intrinsic viscosities or " K " values. In tests performed for support of the present invention, preparations having average molecular weights of about 10,000 were used.

It is known that polyvinylpyrrolidone (PVP) is sensitive to gamma radiation. Typically, an increase in the average molecular weight of the polymer is observed when a PVP solution is exposed to 10-30 kGy gamma radiation. The manufacturer therefore does not recommend that the material is exposed to ionizing radiation for sterilization purposes (KOLLIDON: polyvinylpyrrolidone for the pharmaceutical industry, BASF Aktiengesselschaft, Feinchemie, D67056 Ludwigshafen).

In tests performed to support the present invention, it was discovered that incubation of peripheral blood mononuclear cells (PBPC) in a filter-sterilized OCS particle density gradient material (0.2 Âμm) or by heat sterilization ( autoclave, 15 minutes at 120Â °) does not significantly affect the recovery of viable hematopoietic progenitor cells as assessed by a functional cell assay, the colony-forming unit (CFU) assay, which measures the potential of these cells to form hematopoietic colonies . However, when an OCS particle density gradient material was sterilized by exposure to gamma radiation (2.5-3.5 MegaRad) according to Standard procedures, the cells incubated in this medium, according to the protocol described as Method 2 in example 4 (incubation with OCS + 0.5% PVP for 30 minutes plus 10 minutes centrifugation time) showed less functional activity (for approximately 25% of the added CFU activity). The result was even more surprising for cells treated according to the protocol described as Method 4 in Example 4, in which the culture step was not included. Here the cells were incubated for 30 minutes in a density gradient material based on OCS particles, OCS particles + 0.5% PVP or buffer solution + 0.5% PVP, followed by centrifugation for 10 minutes followed by 24 hours incubation in an ISCOVES medium supplemented with 10% fetal cell serum. Inclusion of 0.5% PVP in the suspension of OCS particles during preparative sterilization resulted in a recovery of about 65% of the added CFU (Figure 1), while the cells incubated in a OCS particle density gradient material presented only about 1% of the original functional activity. Although this test used PBMC-derived cells, cells derived from bone marrow and umbilical cord had the same susceptibilities. The amount of polylactam needed to provide protection to the cells has been investigated in other tests performed to support the present invention. Figure 2 shows how the results of the OCS particle density gradient material were complemented with various concentrations of PVP before gamma radiation (11.5-3.5 megaRad). In this study, PBPC cells were processed according to the protocol identified as Method 6 in Example 4 (incubation in the presence of density gradient suspension and 10% fetal calf serum for 24 hours at room temperature). PVP in addition to the final 3% concentration in the density gradient medium provided significant protection against radiation toxicity in the dose (2.5-3.5 megaRad) of applied radiation. Addition of PVP after sterilization of the medium did not provide a similar protective effect.

From the foregoing it will be apparent that the present invention includes a better cell density gradient composition for use in the separation of rare blood cells. The composition of the product is a silanized silica particle treated by the ionizing radiation process in the presence of at least about 0.05% polylactam and more specifically from about 0.1 to 10% polyvinylpyrrolidone. Such a composition is particularly well suited for isolating rare progenitor blood cells, such as hematopoietic progenitor cells (CD34 +) which are sensitive to exposure to irradiated organosilanated colloidal silica. D. Improvement in polylactam from the recovery of rare blood cells from density gradients

According to another aspect of the invention, it has been found that supplementing a suspension of 0.05% (w / v) organosilanized silica particles of polylactam results in improved yields of rare blood cells in functional form. More specifically, the inclusion of polylactam prevents aggregation or agglutination of cells from cells to provide a cell composition having a more uniform density profile and better functional properties. Compared to cells isolated in OCS density gradient material in the absence of a polylactam, these compositions are characterized by better recovery of rare cells in enriched fractions, and better recovery of rare cells having a clonogenic potential.

Figures 3 and 4 show the result of experiment in which PBPCs isolated according to standard methods were loaded on an organosilanized colloidal silica density gradient (OCS particle density gradient material, 1.0605 g / ml, 280 mOsm / kg H 2 O, pH 7.4) and centrifuged according to the protocol detailed in example 3. Figure 3 shows that, in the absence of PVP, most of the cells present in the starter material were recovered from the pellet. Visual inspection revealed that this pelletisation of cells appeared to be the result of aggregation or agglutination of cells' cells. In contrast, when the OCS density material was supplemented with 0.5% PVP, a larger percentage of cells remained at the interface, as would be expected based on the predetermined density of these cells.

This observation was most pronounced when the same cell preparations were analyzed for the presence of hematopoietic progenitor cells (CD34 + cells) according to the procedures detailed in examples 5 and 6 (FACS and CFU analysis). The results shown in Figure 4 show that, in the absence of PVP, about 90% of CD34 + cells migrated into the pellet. The presence of 0.5% PVP in the density gradient suspension reversed this pattern. Here, most CD34 + cells remained at the interface, as is desirable for the purposes of isolation and enrichment of these cells for transplantation and other clinical purposes. III. Methods of isolating specific cell types

In view of the foregoing, it may be appreciated that the present invention includes, not only a density gradient composition which improves the recovery of rare blood cells, but also a process of isolation of rare functional blood cells, and particularly cells such as hematopoietic progenitor cells (e.g., CD34 + cells), dendritic cells, natural killing cells, natural suppressor cells, and the like of a cell mixture.

In general, hematopoietic progenitor cells may be isolated from bone marrow, umbilical cord blood or peripheral blood samples derived from a subject, in particular a human. It is desirable, first, to mobilize the cells in the subject by administering agents such as granulocyte colony stimulating factor (G-CSF), VP16 or a combination of the two factors, according to protocols well known in the art. Following this treatment, the patient is subjected to apheresis to collect PBPC or to aspirate the bone marrow to collect bone marrow.

According to the present process, the collected cells are deposited in layers on a density gradient material and preferably in a cell separation density gradient suspension based on organosilanized colloidal silica particles having a specific density which is approximately equal of the cells to be isolated. The medium preferably includes at least 0.05% PVP and, more preferably, at least about 0.5% PVP. The suspension is then centrifuged into pelleted cells having a specific density that is higher than the specific density of the selected hematopoietic cell. Such a density gradient composition also has the advantage of being sterilized by the methods described above. The above process may also be improved by adding to the initial cell mix a suspension comprising an organosilanized silica particle to which is attached an antigen-specific binding molecule for negative or positive selection. As an example of the negative selection, the antigen-specific binding molecule attaches an unwanted cell type in the cell mixture to force the cell to migrate to the pellet. As an example, it is sometimes desirable to empty the T lymphocytes from a cell mixture, such as a hematopoietic stem cell mixture, before reintroducing the chosen cells into the body. These T lymphocytes can be emptied, for example, by treating the cell mixture with 23 microspheres to which a specific T lymphocyte antigen is attached. The microspheres are then separated from the desired cell mixture by centrifugation. Examples of the T lymphocyte specific antigen binding reagents that can be attached to the microspheres include anti-CD3, anti-CD4 and anti-CD8 mouse monoclonal antibodies.

Any tubes suitable for use in centrifugation may be used in the practical embodiment of the invention. In a preferred embodiment of the present invention, the " trap " described in WO 96/07097 will be used for the density separation of CD34 + cells. A. Isolation of CD34 + hematopoietic progenitor cells PCT / 95/11169 application from WO 96/07097 describes studies in which PERCOLL® with a physiological osmolality of 280 ± 10 mOsm / kg H2 O and pH of 7.4 was used as a density gradient material to isolate CD34 + cells from the separated blood (apheresis). According to the invention described herein, in this osmolality, recovery of CD34 + cells was optimal with a specific density of 1.0605 g / ml.

In studies performed to support the present invention, bone marrow cells were harvested by aspiration according to standard clinical procedures from healthy donors obtained from the Bone Marrow Transplantation Laboratory of the School of Medicine of Stanford University, Paio Alto, CA. CD34 + cells were enriched from the bone marrow cell fraction by centrifugation over an OCS gradient having a specific density of 1.0685 g / ml with 280 mOsm / kg H2 O, pH 7.4, prepared and formed from according to the processes of the present invention, as described in example 10.

Figures 8 and 9 summarize the results of isolation studies on bone marrow cells processed in the OCS suspension of 1.0685 g / ml. Figure 8 shows the recovery in percentage of total cells, CD34 + cells, T lymphocytes and red blood cells (RBC) of the whole gradient (interface + pellet). Figure 18 shows the percent recovery of the same cell fractions from the gradient interface following centrifugation. Data presented in both figures were calculated by dividing each fraction by the total number of cells initially loaded onto the gradient.

From Figure 8 it can be seen that more than 90% of the cells initially loaded in the gradient were recovered from the interface and pellet. Since the loss of CD34 + cells was comparable to the approximate loss of total cells (10%), there was no preferential loss of these cells. Figure 9 shows that while only 26 10% of the total cells were recovered from the interface, this fraction contained 91 ± 9% of CD34 + cells, 10% of T lymphocytes and essentially no red blood cells. These studies demonstrate a substantive enrichment of CD34 + cells coupled with T lymphocyte depletion through the density gradient procedure of the present invention. As will be discussed in the following section, these features recommend the preparation for use in transplantation of allogeneic stem cells. B. Natural suppressor cells 25

In vitro studies show that human bone marrow contains cells with low density that block in vitro all responses in the mixed lymphocyte reaction (MLR). Based on the fact that this suppressor activity is not restricted to HLA, it has been reported in the literature as a natural suppressor (NS) activity. A PERCOLL density gradient was adjusted to a density of 1.0605 ± 0.0005 g / ml to test its ability to enrich cells with NS activity. Separated blood samples (apheresis) from lymphoma patients and from normal subjects receiving G-CSF treatment were centrifuged in a five-layer batch gradient, and the interfaces and pellet were evaluated for their suppression potential of the mixed lymphocyte culture. Figure 10 shows that cells with NS activity had a density of less than or equal to 1.0605 g / ml. Accordingly, more than 90% of the NS activity was present in the final cell preparation after centrifugation on a 1.0605 g / ml gradient. C. Natural exterminating cells

Natural killer (NK) cells have been shown to exterminate autologous tumor cells. From a clinical perspective, it may be beneficial to have a greater number of NK cells at the transplant to reduce tumor relapse. In experiments for the support of the present invention, NK cell density was determined on a five-layer PERCOLL® batch gradient. NK cells also showed a density of 1.0605 g / ml or less. Accordingly, more than 90% of the NK cells were present in the final cell preparation after centrifugation in a gradient of 1.0605 g / ml, as shown in figure 11. D. Dendritic cell enrichment and cytotoxic T lymphocytes

Dendritic cells (cells bearing antigens) are cells derived from blood fractions. These cells act in vivo to present the antigen to the T lymphocytes to elicit a major cytotoxic T lymphocyte (CTL) response. These cells have clinical utility for use in inducing or ameliorating immune cell responses.

According to another aspect of the present invention, dendritic cells and cytotoxic lymphocytes are isolated using defined density gradients in a three-step process. This process includes centrifugation through three different density gradient solutions, as detailed in Example 12. In accordance with the present invention, the OCS density gradient material having densities of 1.0770, 1.0720 and 1, 0610 or 1.0720, 1.0610 and 1.0565 g / ml, each having a pH of 7.4 and an osmolality of 280 mOsm / kg H2 O are used in three phases, which are conveniently performed, but not necessarily, in a " trap " tube. Those skilled in the art will recognize that equivalent densities of other media may also be used. For example, step 1 can be run in a Lymphoprep (Nycomed Laboratories, Oslo, Norway) or a PERCOLL® equivalent to Ficoll (density 1.0770, 310 mOsm / kg H20, pH 7.4); step 2 can be run in a PERCOLL® solution (density 1.0650, pH 7.4, 300 mOsm / kg H2 O, and step 3 can be run in a PERCOLL® solution having a density of 1.0800 g / ml, pH 7.4, mOsm / kg H2 O, or in a PERCOLL® solution having a density of 1.0550 g / ml, 290 mOsm / kg H2 O, pH 7.4). However, the densities referred to above and determined using an OCS-based cell sorting and enrichment methods of the present invention are advantageous in that they are performed with physiological osmolality, rather than under previously reported hyperosmotic conditions. E. Density Adjusted Cell Classification (DACS) The present invention also provides for the clearance of bone marrow or other tumor cell cells by contaminating the source of progenitor cells or other undesirable cells. For purposes of clearance, the density of a given tumor cell is determined by centrifugation of a sample containing a tumor in a discontinuous density gradient in the range of 1.0490 to 1.0640 g / ml. Most undesirable, non-tumor cells represent cells of the immune and hematopoietic system and have a density in the range of 1.0610 to 1.0770. The density of the tumor cells usually falls within a density range of 1.0490 to 1.0580 g / ml. The density of the cell separation medium is determined to be within ± 0.0005 g / ml, preferably ± 0.0002 g / ml of the specific gravity of the desired cells. Thus, tumor clearance can be achieved in a two step process in which the desired cells are first pelleted through a cell separation medium capable of retaining tumor cells. Alternatively, the clearance technique may be used to remove tumor cells in a one-step process, according to the methods described below. 28

Experiments performed to support the present invention demonstrated that CD34 + cells were enriched using the DACS procedure in conjunction with the density method of the invention, removing contaminating cells with an anti-CD45 mAb coupled to a heavy carrier (such as magnetic microspheres or microspheres aminopropyl glass). The total number of cells was reduced by 82% and the CD34 yield was about 40%. The purity of CD34 was increased from 2% to approximately 20%. Since the anti-CD45 antibody also removed some of the CD34 + cells, it is to be understood that this process could be improved by using a mixture of other antibodies for the depletion of non-stem cells.

To examine non-specific cell losses during DACS experiments were performed using peripheral blood mononuclear cells (PBM) of the leukocyte layer and DACS microspheres in the absence of monoclonal antibodies. As detailed in Example 10, the DACS microspheres were made using either intact GAM-IgG or F (ab) 2 GAM-IgG to test the theory that Fc receptor-mediated cell binding to IgG coated DACS microspheres contributed to a non-specific depletion. The results in Figure 5 demonstrate that there was no nonspecific cell depletion when the DAMS-coated DACS F (ab) 2r microspheres were used when compared to DACS microspheres with intact GAM-IgG. The GAM-IgG coated DACS microspheres bound through the glycosylated site in the Fc zone also demonstrated less nonspecific cell depletion when compared to DACS microspheres with GAM-IgG in a random orientation. These results demonstrated that removal or blocking of the Fc zone of GAM-IgG leads to a reduction in nonspecific cell depletion. 29

The above results show that FC receptor (FcR) mediated binding to DACS microspheres may be responsible for some of the non-specific depletion of CD34 + cells observed during DACS treatment. This is further illustrated by site specific coupling which results in a reduction of nonspecific cell loss due to blockage of the glycosylated areas of the Fc portion of the immunoglobulin molecule, as described above. However, the use of BSA-coated DACS microspheres has demonstrated that FcR-mediated binding is not normally responsible for all non-cell-specific depletion. It is likely that the ionic interactions between the cell and the surface of the microsphere are also involved. To investigate this possibility, PEG-coated DACS microspheres sandwiched between the microsphere surface and GAM-IgG were made. This was done to explore the ability of PEG to reduce ionic interactions between proteins and charged surfaces. In an experiment comparing PEG-GAM-IgG-coated DACS microspheres with DACS microspheres with GAM-IgG in the absence of PEG, it was found that the use of PEG is associated with a reduction in nonspecific cell depletion. Reduction in nonspecific depletion by modification of DACS microspheres is particularly important in the context of the enrichment of rare cells such as CD34 + cells, NK cells, natural suppressor cells, dendritic cells, nucleated cells of fetal origin, and the like, in which it is important to maximize the yield of the cells. For example, experiments to define the non-specific loss of CD34 + cells demonstrate that the use of F (ab) 2 coated DACS microspheres leads to a reduction in the non-specific loss of CD34 + cells (Figure 6). Likewise, using GAM-IgG1-coated DACS microspheres ligated through the Fc-zone glycosylation site also reduced the non-specific depletion of CD34 + cells as compared to DACS microspheres with GAM-IgG in a random orientation (Figure 7).

E. Depletion of T lymphocytes Graft vs Host disease (GvHD) is caused by T lymphocytes that are present in donor allografts, although the complete removal of these cells also results in graft failure and tumor relapse. Thus, the presence of a limited number of T lymphocytes for allogeneic transplantation may be required successfully. In experiments performed to support the present invention, DACS microspheres were used to empty the T lymphocytes from separate PBSC preparations (apheresis) from donors donated with G-CSF. T lymphocytes were emptied from a mobilized PBSC preparation by adding to the preparation of the anti-CD3, anti-CD4, and anti-CD8 mouse monoclonal antibody mixture, followed by rat anti-mouse IgG DACS microspheres, and centrifugation through a OCS gradient having a density of 1.0605 g / ml at 280 mOsm, pH 7.4. T lymphocytes were monitored by reaction with anti-CD2 monoclonal antibodies to avoid possible artifacts associated with the use of antibodies that recognize the same epitopes on differentiation antigens. The DACS T lymphocyte depletion method referred to above resulted in a 97% reduction in T lymphocytes with only a 70% reduction in CD34 + cell yield. 60% of the CD34 + cells were retained at the interface after DACS treatment. The results demonstrated that the combined cell enrichment 31 with the DACS method specifically reduced the T lymphocyte content of allograft material by 97%, with a loss of less than 40% of CD34 + cells. This preparation is suitable to prevent GVHD in a transplant material. F. Enrichment and clearance of tumor cells 1. Enrichment of tumor cells for diagnosis

Mammary tumor cells are used to exemplify the enrichment methods of the present invention. To enrich these cells, a gradient should be adjusted to a density of 1.0490 to 1.0580 g / ml, depending on the specific type of mammary tumor cells, a physiological osmolality of 270-290 mOsm / kg H20 and physiological pH of 6.8-7.8. In a specific embodiment, by way of example, mammary tumor cells are loaded directly into a centrifuge trap tube containing a suitable cell separation medium, such as OCS particulate cell separation media, or filled PERCOLL® solution to a level above the constriction.

In the specific examples described herein, successful enrichment of tumor cells was performed on a PERCOLL density gradient material that was adjusted to the specific gravity between 1.0490 to 1.0580 ± 0.0002 g / ml, depending on the specific cell type of the mammary tumor, osmolality of 280 mOsm / kg H20 and pH 7.4. The density of the PERCOLL solution can be adjusted in a densitometer to accurately define its accuracy.

Cells enriched by the methods described above can be subsequently examined for the presence of mammary tumor cells. The resulting yield of cells isolated or enriched from the cell separation methods of the present invention may be used for diagnostic purposes, for example morphological, molecular, biochemical or immunophenotypic assays. For example, the DNA may be prepared from the harvested cells and subjected to the polymerase chain reaction (PCR), as described in the section for clearance of tumor cells below, or the harvested cells can be evaluated morphologically, thus avoiding the use of invasive and costly surgical procedures on which such determination is based. Various mammary tumor antigens and mammary tumor markers are known to those of skill in the art or are available commercially, including, but not limited to, cathepsin D, EGF-R, estrogen receptor, Ki-67, progesterone receptor and TGF-α. Antibodies directed to these antigens or markers may be used to assess the tumor type of the harvested cells.

A major advantage of the methods described herein is that a large volume of whole blood can be placed directly in the density gradient. Peripheral blood can be collected in tubes containing anticoagulant or by apheresis or leucoferesis. Complete blood does not need to be processed or diluted prior to centrifugation. However, since the methods enrich the mammary tumor cells based on their specific flotation density, it is important that the cells be subjected to separation after a relatively short period of time after collection from an in vivo source , because the density of the cells changes according to their culture or storage conditions. Thus, in order to achieve optimum enrichment of mammary tumor cells from the blood, it is preferred that the blood samples be used within 48 hours after collection. More preferably, the blood samples should be subjected to a density gradient centrifugation up to a few hours after collection. 2. Debugging Tumor Cells Tumor cell clearance is conveniently performed using the tumor cell enrichment methods referred to above, or more preferably, the above-described DACS enhanced cell separation methods. As an example, when the radiolabeled mammary tumor cells were admixed with a peripheral blood leukocyte layer, up to 80% were retained by centrifugation of the mixture on a gradient having a specific gravity of 1.0580 g / ml and an osmolality of 280 mOsm / kg H2O. In addition, only a small fraction (< 10% of the initial number of cells) of non-tumor cells was discovered in the tumor fraction collected. PCT / 95/11169 patent application WO 96/07097 discloses the enrichment of mammary tumor cells according to the processes of the invention. This process may also be used in conjunction with the DACS procedure described herein. For example, whole blood can be directly incubated with anti-CD45 antibodies coated with carrier particles that react with most of the leukocytes. Since breast tumor cells do not react significantly with anti-CD45, the vast majority of non-red blood cells, leukocytes and other cells are weighted than the density and pellet material during centrifugation, whereas tumor cells remain in the upper compartment. A variety of other cell-type specific binding agents may be used to label specific types of cells in the blood, as referred to in the preceding sections.

Alternatively, according to a preferred clearance procedure, the DACS positive selection procedure may be used to cause tumor cells to heavier than their normal densities, so that they are pelleted during centrifugation. Cell-type specific binding agents useful in the positive selection procedure include, but are not limited to, antibodies to mammary tumor antigens and antibodies to breast tumor markers, for example HER2 / Neu, CA 15-3, CA 549, cathepsin D, EGF-R, estrogen receptor, Ki-67, progesterone receptor, and TGF-α. Many of these antibodies are available commercially.

Cells enriched by the processes described above may subsequently be screened for the presence of mammary tumor cells. The resulting yield of cells isolated or enriched from the cell separation methods of the present invention may be used for diagnostic purposes, for example morphological, molecular, biochemical or immunophenotypic assays. For example, DNA can be prepared from harvested cells and subjected to the polymerase chain reaction (PCR) or the harvested cells can be morphologically evaluated, thus avoiding invasive and costly surgical procedures, hitherto used for such a determination. Alternatively, tumor cells can be identified by the presence of tumor cell markers, known to those skilled in the art, as described above. Antibodies directed to these antigens or markers may be used to evaluate the tumor type of harvested cells. 3. Tumor clearance of peripheral blood samples The presence of tumor cells in stem cell grafts and the inability to specifically empty tumor cells from the bone marrow of patients with non-Hodgkin's lymphoma may be an important indicator in predicting disease relapse after a autologous transplantation. The cell separation processes according to the present invention may be applied to remove tumor cells from the peripheral blood products prior to the reintroduction of these products into the patient.

In studies performed to support the present invention, the SU-DHL4 cell line of B-cell lymphoma was used to model peripheral blood tumor cell contamination. This cell line contains a reciprocal translocation between the long arms of human chromosomes 14 and 18, resulting in a juxtaposition of the bcl-2 proto-oncogene to the contiguous zone of the immunoglobulin (Ig) heavy chain gene. This translocation, leading to high levels of the bcl-2 gene product, is found in 85% of small cell cleaved follicular lymphomas and 25% of large cell diffuse lymphomas. The breakpoint of translocation in the bcl-2 gene is located either in the major breakpoint region (MBR) spreading approximately 150 bp in the 3 '(UTR) region of the bcl-2 gene or in the minor cluster region (MCR) of 500 bp located approximately 20 kb downstream of the bcl-2 gene. Using pairs of translocation-specific primers and quantitative PCR limiting the techniques of dilution analysis known in the art, the depletion of SU-DHL4 cells in the DACS clearance experiments was monitored.

Likewise, the human adenocarcinoma cell line SK-BR3 is used to model the infiltration of peripheral blood breast cancer cells. This cell line reacts with the mouse anti-human monoclonal antibody HR2 / Neu. Prior to being transformed into normal leukocyte layers, SK-BR3 cells were preloaded with calcein AM, a fluorogenic substrate, to be enumerated by fluorescence activated cell sorting (FACS). Details of the experimental methods used in performing the tumor clearance experiments are provided in the example.

A PCR assay was used to monitor tumor cell levels in the blood cell mixtures. Initially, mixtures of SU-DHL4 genomic DNA and a negative translocation (14; 18) human cell line (K562) were used as start-up models to assess the sensitivity level of the PCR assay. Agarose gel electrophoresis was used to separate PCR products, either one turn or two rounds of amplification, as shown. PCR products derived specifically from SU-DHL DNA were identified on the gel. The results of these preliminary tests showed that a PCR with internal indicators can detect approximately one SU-DHL4 cell per 150,000 cells.

In addition, genomic DNA was prepared from frozen blood aliquots (apheresis) of 24 patients with a non-Hodgkin's lymphoma (NHL) who received chemotherapy and G-CSF in accordance with standard methods known in the art. PCR products with internal markers using specific primer pairs for major breakpoint region 37 (MBR) translocation were found in four patients, demonstrating that this assay can detect cells carrying the translocation of authentic clinical specimens.

Further experiments were performed to validate the use of PCR analysis under limited dilution conditions for quantification of tumor cell frequency in a sample. Replicates of PCR reactions were performed using known amounts of input SU-DHL4 genomic DNA. PCR primer pairs with internal indicators for t (14; 18) (a target sequence present once per SU-DHL4 cell) and p53 (a target sequence present twice per SU-DHL4 cells) were used. The observed frequency of positive PCR reactions demonstrated that the experimental frequencies are similar to frequencies expected based on the Poisson statistics. Thus, it has been shown that PCR using limited dilution analysis provides a precise and quantitative method for monitoring tumor cells.

To model the contamination of PBPC tumor cells, separate blood samples (apheresis) from G-SCF mobilized patients were fed a known number of SU-DHL4 cells. This material was used to evaluate the removal of tumor cells by DACS. This procedure involves incubating the blood with mouse monoclonal antibodies specific for surface antigens expressed on the target cell (CD9, CD10, CD19, CD20). This step was then followed by incubation with goat anti-mouse Ig coated with high density microspheres (DACS microspheres). The subpopulation of cells attached to these microparticles was then separated from unbound cells by centrifugation through an OCS suspension supplemented with 0.5% PVP at a specific density of 1.0605 g / ml and sterilized by gamma radiation. Cells 38 were recovered from the interface (IF) and pellet (PT) after gradient centrifugation and quantified by gel electrophoresis, as detailed in Example 5. The results of a representative depuration experiment are shown in Figure 12 and Figure 12 shows the number of SU-DHL4 cells in: A) unfractionated sample (determined by visual counting), B) unfractionated sample, C) interface without DACS, D) interface after use of DACS GAM microspheres -IgG, and E) Interface after use of DACS GAM-IgG microspheres and a cocktail of monoclonal antibodies (mAbS) on anti-B cells. The number of SU-DHL4 cells in B, C, D and E was determined by quantitative PCR. Figure 13 depicts the recovery of CD34 + cells at the interface (IF) and pellet (PT) of C, D and E. Tumor cell depletion was quantified by PCR analysis with limited dilution for B-E samples, as shown . Recovery of CD34 + hematopoietic cells was monitored by tricolor flow cytometry. The recovery of whole cells and stem cells are illustrated with respect to the recovery of cells from the interface. A single clearance cycle using an antibody cocktail resulted in a log 2-3 specific depletion of SU-DHL4 cells. The recovery of CD4 + cells after DACS was greater than 70% and comparable to the number of CD34 + cells recovered from the interface in the absence of DACS Peripheral blood mononuclear cells are reduced by 70-80% by density centrifugation with or without DACS.

In another model for contamination of tumor cells, SK-BR3 cells were seeded in normal leukocyte layers at a frequency of 2%. Tumor cells from the AM calcein labeled cells were removed from the cell suspension by centrifugation in a suspension of OCS density having a specific density of 1.0605 g / ml with DACS using a HER-2 / Neu anti-human mAbs. Depletion was determined by FACS analysis. The cells labeled with AM calcein were monitored and their percentage determined using the LYSYS II program. While SK-BR3 cells had not been emptied of the leukocyte layer after processing in an OCS suspension in the presence or absence of DACS microspheres, they were almost completely removed from the cell mix when using DACS microspheres in combination with the anti-Her2 / Neu mAb specific for breast cancer. These data show that DACS treatment reduces tumor cell contamination by 1 or 2 logs (10-100 times) when the FACS percentage is calculated back to the initial number of tumor cells in the sample.

In summary, the above model example experiments using human B cell lymphoma cells and quantitative PCR demonstrate that a 2-3 log reduction in tumor cells can be achieved in a DACS loop, retaining the majority of CD34 + hematopoietic stem cells . Using a human adenocarcinoma cell line and FACS analysis a 1-2 log reduction was achieved in tumor cells. In general, with the present invention it has been found that this process can be used to purge blood or populations of therapeutically enriched cells from undesired tumor cells. IV. Sterilized cell separation kit

An important application of the discovery of the present invention is the sterile assembly for cell separation. Such a kit is particularly useful for clinical cellular separations. Such a kit will typically include a centrifugal vessel and a density gradient material to separate cells, which has been sterilized as a full unit by exposure to ionizing radiation. The centrifugal container will normally be a consumable and disposable unit made of plastic, such as polypropylene, polycarbonate, polysulfone, polymethylpentene, polyethylene, polyalomer, polystyrene and the like. However, as will be noted below, the advantages of sterilizing the assembly as a pre-filled unit also make this procedure desirable for assemblies including tubes made from other materials. In a preferred embodiment, the container will be closed, as shown below, to facilitate the aseptic pick up of starter material and the preservation and manipulation of the material in a sterile condition. Figure 14 shows an assembly 10 formed in accordance with the present invention. As shown, the assembly 10 includes a centrifuge tube 12, shown here filled with density gradient material 14. The density gradient material is formed by a suspension of organosilanized silica particles, such as particles 16. The gradient material of density also includes at least 0.05% PVP, and preferably 0.1% to 10% PVP, to reduce the toxicity to isolated rare blood cells in the pool, thereby improving the yield of these cells as a whole. The container and the density gradient contained therein are sterilized by ionizing radiation, such as gamma radiation or electron beam. The resulting assembly may also include a protective cap, such as a cap 18, to protect the tube contents from contamination after sterilization. The assembly may also include a holder, such as a tube holder 20, to prevent displacement of the contents of the tube during shipping and storage. The assembly, as described above, may be prepared according to the specifications required for separation of a set of cell types. The size of the tube used, as well as its structural and chemical resilience, may be altered within the limits of the present invention, for use in ultracentrifugation, as well as in relatively low speed centrifugation applications, as described herein. These modifications will be apparent to those skilled in the art.

Similarly, the array can be designed for different cell types by changing the composition of the density gradient material. Example 9 describes a kit which is particularly formulated for isolation of CD34 + hematopoietic progenitor cells from PBPC. In the described embodiment, the OCS particle density gradient suspension is adjusted to a specific density of 1.0605 g / ml and an osmolality of 280 mOsm / kg H 2 O. This density is approximately equal to the density of the desired CD34 + cells, so, upon centrifugation, these cells will " float " at the interface between the filler material and the density material present in the tube.

Taking into account that enrichment of other rare blood cell types, such as dendritic cells, T cytotoxic cells, natural exterminating cells, natural suppressor cells, and nucleated cells of fetal origin can be achieved by adjusting the specific density and osmolality of the material of density gradient. This adjustment may be made so that the desired cells then float or pelletize, according to the desired cell density and the relative density of undesired cells present in the cell mixture from which the cells will be isolated.

A preferred centrifuge tube configuration for use in the assembly and method of the invention is one having a constriction member disposed within the tube, forming a " trap ". Such a tube is described in the patent application PCT / 95/11169 (publication number WO 96/07097). A modified form of this tube is illustrated as part of a closed system in Figure 14, described below.

In the trap tube, the constriction member is positioned and constructed to retain fluid at the bottom of the tube below the constriction member, when the tube is inverted. The tube also includes cell separation means which is contained in the lower portion of the tube and which extends over the constricting member to a level above the aperture formed by the constriction member. Cells that are captured at an interface between the cell separation medium and a lower density medium after centrifugation are discharged with the lower density medium when the tube is inverted. These easy cell collection means are particularly convenient in the context of the present invention, whereby minimal manipulation of the cells is desirable.

Also consider that, for clinical applications, the set can be configured as a " closed system " to restrict access to pathogens external to biological materials as described herein. Figure 14 shows a centrifugal container having a constricting member incorporated in a closed system. As shown, the closed system 70 includes a reservoir 71, shown as a sterile bag containing blood 72 previously collected by known techniques, being connected by a sterilized connecting tubing 74 to the centrifuge container 76, shown as a centrifuge tube having an enclosed top 77. The closed top will have at least 1, and preferably at least two inlet ports, useful for sample introduction and removal, and for ventilation, as described below. In the illustrated embodiment, a solid rib 79 protruding upwardly from the closed top 77 is included to form a protective barrier for the entry ports, and as a securing point for a removable protective cover for the apparatus, which serves to potential contamination during shipment and storage.

Still with reference to figure 14, the tubing 74 is secured to the container 76 through the inlet port 78, fitted with a fitting 80, which may be any type of locking tip adapted for sterilized connection, for example a syringe connecting element Luer-Lock ™. Alternatively, the housing 80 may be a sterile septum adapted for attachment to sterile fluid bags and tubes, for example a SAFSITE ™ small wire extension kit with reflux valve and Spin-Lock adapter available from Burron Medical Inc., Bethlehem , Pennsylvania. To facilitate the flow of fluid to the centrifuge tube 76, the basket contains a venting port 82. As shown, the air filter 84 is attached to the port 82 to prevent contamination.

According to a preferred embodiment of the invention, the container 76 includes a constriction element 88. Although a set of configurations is possible, as shown, the constriction element 88 is in the form of a funnel on its upper surface . As shown, the constriction member 88 is formed integrally with the container 76 forming a toothed cutout 89 on the outer surface of the tube. The constriction member is supported by supports 90 to prevent compression during centrifugation. The container also contains an OCS particle density gradient material 91 disposed in the tube either above or below the constriction member 88. The entire system, as shown, is sterilized by gamma radiation or electron beam according to with Standard methods.

A further features of the container 76 is the inlet port 92. This inlet port communicates through the closed fluid channel 94 with the lower portion 95 of the container 76. The inlet port and channel are used to fill the lower portion of the tube 95 , for example, with a density gradient cell separation medium prior to sterilization. Alternatively, the port and channel may be used to remove materials, including cellular pelletizing materials, from the bottom of the tube following centrifugation. This feature of the tube is not required by, nor is it restricted to, the context of the closed system in which it is illustrated. The fluid flow from the reservoir 70 to the vessel 76 may be initiated by applying suction to the ventilation inlet port 82, or may be initiated by other means known in the art, including the action of gravity. The flow rate is adjusted either by changing the pressure between the two containers or by regulating a valve placed in the tube or inlet port at a point between the reservoir 70 and the port 78. The flow rate is optimally adjusted to fill or partially fill , the top of the container 76, above the level of the density gradient solution. When a sufficient sample of fluid has entered the tube, the flow may be interrupted through any of the set of means known in the art, such as regulation by a valve or by lowering the pressure. The tube is then removed from the basket, port 78 is sealed, and the closed container is subjected to centrifugation, as described above.

In another preferred embodiment which is particularly useful in carrying out the method of the present invention, the constricted tube may be in the form of a centrifugal syringe, such as the centrifuge syringe described in PCT / US Pat. No. 95/1162 (WO 96/06679). These embodiments are well suited for use in a closed system, as described above.

Another configuration of the device that is suitable for use in the invention is a bag that can be centrifuged. Such a configuration is particularly suitable for use in the processing of clinical blood samples which can be collected directly into the bag.

The following examples illustrate, but are not in any way intended to limit, the present invention.

EXAMPLES

Materials A. Density gradient materials

Organosilanated colloidal silica particles (size range 15-30 nm) used in the examples below were prepared according to the processes described in U.S. Patent 4,927,749 using gamma-glycidoxypropyltrimethoxysilane (GPMS). Polyvinylpyrrolidone (PVP-10) was purchased from Sigma (St. Louis, MO). B. Monoclonal antibodies

Monoclonal antibodies (mAb) directed against specific surface antigens of hematopoietic progenitor cells (anti-CD34; anti-HPCA-2) and leukocytes (anti-CD45; anti-HLE-1) were purchased from Becton Dickinson, Inc (San Jose, , CA). The different monoclonal antibodies were labeled directly with fluorescein isothiocyanate (FITC) or phycoerythrin (PE). Ig-labeled polyclonal IgG1 control antibodies from the PE-labeled isotope were purchased from Becton Dickinson, Inc. (San Jose).

Example 1

Preparation of the density gradient solution The OCS particle density gradient material, prepared as described in " Materials " above, was diluted in a buffered saline phosphate (PBS) to provide a suspension having a defined osmolality and a specific density with a pH of 7.4, as indicated. The density gradient material based on OCS particles was diluted in water to approximately the required specific gravity. Solid salts were added to provide a final physiological saline solution (a physiological saline or Dulbecco's buffered saline phosphate, Gibco). The suspension was supplemented with different final concentrations (0% - 10% w / w) of polyvinylpyrrolidone (PVP-47 10) by adding the adjusted solid powder to the final suspension density. The diluted suspension was finally sterilized by ionizing radiation (gamma ray and electron beam, 2.5-3.5 megaRad, Isomedix, Morton Grove, IL), autoclaving (15 minutes at 120 °) or filtering through a 0.2 microns according to standard methods known in the art and used at room temperature.

Example 2

Preparation of hematopoietic progenitor cells A. Blood cell mixtures

Peripheral blood mononuclear cells (PBPC) were collected by apheresis from non-Hodgkins lymphoma (NHL), Hodgkins lymphoma (HL) and breast cancer patients at Bone Marrow Transplantation Laboratory at Stanford University School of Medicine, Paio High, California, USA. PBPCs were mobilized in NHL patients by treatment with cyclophosphamide (4 g / m 2, intravenous) followed by a granulocyte colony stimulating factor (G-CSF) (10 g / kg, intravenous, daily). PBPCs were mobilized in breast cancer patients by treatment with XXX (VP-16; 2g / m 2, intravenous) followed by G-CSF (10æg / kg, intravenous daily). PBPCs were mobilized in breast cancer patients by treatment with etoposide (VP-16, 2 g / m 2, intravenous) followed by G-CSF (10 pg / kg, intravenous daily). PBPCs were mobilized in HL patients by treatment with G-CSF alone. Mobilization protocols are standard protocols known in the art. 24 hours after the final injection, each patient underwent apheresis. PBPCs were collected from the separated blood (apheresis) according to standard methods. B. Cellular bone marrow mixtures

Aspirated bone marrow cells were slowly layered in an OCS suspension prepared as described above at a density of 1.0608 g / ml and an osmolality of 280 mOsm / kg H 2 O. A maximum of 2x109 cells was deposited in layers and processed through centrifuge tube. The tube was centrifuged for 30 minutes at 850 x g at room temperature. The interface and pellet cells were collected separately from a trap tube, in accordance with the present invention. Cell fractions were characterized by FACS analysis using anti-CD34 (anti HPCA-2), anti-CD45 (anti-Hle-1), anti-CD3 (Leu-4) conjugated to FITC and PE, and IG- obtained from Becton Dickinson, Inc. (San Jose, CA). For analysis, the cells were labeled with the LDS 751 nuclear dye (Exciton, Inc., Dayton, OH). Statistical analyzes were performed on flow type 104 events using a FACSCAN system equipped with a LYSYS II program (Becton Dickinson, Inc. (San Jose, CA)).

Example 3

Density gradient separation of PBPC

PBPC cells (0.25 ml) collected according to example 2 and PBS diluted to a concentration of approximately 5x107 cells / ml were layered in a density gradient suspension based on OCS particles (15 ml in a tube of 50 ml centrifuge). The suspension had a specific gravity of 1.0605 g / ml an osmolality of 280 mOsm / kg H 2 O and a pH of 7.4. Layer deposition was performed slowly to avoid mixing the sample with the solution. A maximum of 2x109 cells was deposited in layers per tube. Centrifugation was performed for 30 min at 850 x g at room temperature. To avoid mixing the cells and density gradient solution, the centrifuge was stopped without braking effort. Following centrifugation, the cells were divided into a low density fraction located at the interface and in a high density fraction forming the pellet. Each cell fraction was collected and transferred to another 50 ml polypropylene centrifuge tube. Cells were washed once and stored in a Dulbecco's buffered saline phosphate free from CA ++ and MG ++ (D-PBS) at room temperature until further manipulation. The number and functionality of haematopoietic progenitor cells (CD34 + cells) were determined either in the cell fractions by FACS analysis or in clonogenic assays (CFU), as detailed in examples 5 and 6 below.

Example 4

Effect on the cells of the blend with silanized colloidal silica particles The evaluation of the mixing effect and subsequent incubation of human peripheral blood in silanized colloidal silica silica particles density gradient solutions was performed by one or more of the processes described below.

After the respective incubations, as detailed low, the cells were collected and washed once with PBS. Their ability to form hematopoietic colony forming erythroids (CFU-E, BFU-E), granulocytes / macrophages (CFU-6M) and mixed colonies (CFU-GEMM) according to methods known in the art, and as described in examples 6 below, were filtered at the pH of each method. In carrying out these experiments, Methods 1, 3 and 5 are incubation control conditions for methods 2, 4 and 6, respectively. Method 1: An aliquot containing 107-108 cells in 5 ml of 1xD-PBS (free of CA ++, MG ++) was added to a 15 ml polypropylene centrifuge tube. The initial number of cells was noted. The cells were incubated for 30 minutes at room temperature with a gentle mixture within 15 minutes apart. Method 2: an aliquot containing 107-108 cells in 5 ml of OCS density gradient suspension was added to a 15 ml polypropylene centrifuge tube. The initial number of cells was noted. The cells were incubated for 30 minutes at room temperature with a gentle mixture within 15 minutes apart. Method 3: An aliquot containing 107-108 cells in 5 ml of 1xD-PBS (free from CA ++, MG ++) was added to a 15 ml polypropylene centrifuge tube. The initial number of cells was noted. The cells were incubated for 30 minutes at room temperature with a gentle mixture within 15 minutes apart. Following centrifugation, the supernatant was discarded and the pellet resuspended in 511 ml Iscove medium supplemented with 10% fetal-calf. The cell mixture was then incubated for 24 h at room temperature. Method 4: an aliquot containing 107-108 cells in 5 ml of OCS density gradient suspension was added to a 15 ml polypropylene centrifuge tube. The initial number of cells was noted. The cells were incubated for 30 minutes at room temperature with a gentle mixture within 15 minutes apart. Following centrifugation, the supernatant was discarded and the pellet resuspended in 5 ml Iscove medium supplemented with 10% fetal-calf. The cell mixture was then incubated for 24 h at room temperature. Method 5: an aliquot containing 107-108 cells in 5 ml of 1xD-PBS (free of CA ++, MG ++) supplemented with 10% of fetal calf was added to a 15 ml polypropylene centrifuge tube. The initial number of cells was noted. The cell mixture was then incubated for 24 h at room temperature. Method 6: an aliquot containing 107-108 cells in 5 ml of OCS density gradient suspension supplemented with 10% fetal-calf was added to a 15 ml polypropylene centrifuge tube. The initial number of cells was noted. The cell mixture was then incubated for 24 h at room temperature.

Example 5

Spotting and quantification of CD34 + cells by FACS The amount of CD34 + cells was determined by fluorescence activated cell sorting (FACS) after cells were labeled with a nuclear dye and mAbs directed to CD34 and CD45. The percentage of CD34 cells present was determined on the nucleated cell portal. This approach was chosen to avoid interference of non-nucleated particulate material with the accuracy of CD34 cell analysis in FACS.

A suspension of 2 x 107 cells / ml was made using D-PBS without CA ++ and MG ++ as diluent. 50 ml of this cell suspension containing 1 x 10 6 cells was added to each well of a microtiter plate 96. 50 microliters of a 20% solution of heat inactivated rabbit serum / D-PBS and 10 microliters of 10æg / ml of LDS 751 (nuclear dye) in a D-DPS solution were also added to each well, followed by mixing. The plate was film-coated and incubated for 30 minutes at room temperature. 10 microliters of IgGI-DPE was added to each control well. For each test well, 10 microliters of anti-CD34-PE was added, followed by mixing. The plate was film-coated and incubated for 15 minutes at 4 ° C. The plate was then centrifuged at 850 x g for 5 minutes at 4 ° C. The supernatant was then removed by stirring.

Each pellet was resuspended in 200 μΐ cold (4 ° C) lxDPS (free of CA ++ and MG ++ 100). The plate was then centrifuged at 850 x g for 5 minutes at 4 ° C, and the resulting supernatant was removed by rapid stirring of the plate. Each pellet was resuspended in 50 ml of a 20% solution of heat inactivated rabbit serum. Anti-CD45-FITC (10 μΐ) was added to each control and test well and mixed. The plate was covered with film and incubated for 30 minutes at 4 ° C. 100 μl of cold (4 ° C-free) CA.sub.1 and DC ++ lxDPS were then added to all control and test wells, and the plate was centrifuged at 850 x g for 5 minutes at 4 ° C. The plate was shaken rapidly to remove the supernatants, and each pellet was resuspended in 200 μl cold (4 ° C) lxDPS (free of CA ++ and MG ++ 100). The plate was centrifuged at 850 x g for 5 minutes at 4 ° C, and then stirred rapidly to remove the supernatant. Each pellet was resuspended in 200 ml of 1% paraformaldehyde (4 ° C). The plate was then covered with film and stored at 4Â ° C until FACS analysis of the samples.

FACS analyzes were performed on 104 flow events using a FACSStar Plus system, equipped with a LYSYS II statistical analysis program (Becton Dickinson, Inc.). For the purpose of analysis, a port (Region 1) was placed around the nucleated cells determined by spotting LDS 751 in FL3 to make separate red blood cells, platelets and debris. FL1 and FL2 were used as a dot plot using Region 1. One port (Region 2) was placed around the cell population that stained with the anti-CD45 and anti-CD34 monoclonal antibodies. For the purpose of analysis, the percentage of cells staining with the anti-CD34 (FL2) and anti-CD45 (FL1) monoclonal antibodies was determined in Region 2. This represents the number of CD34 positive cells as a percentage of the total number of cells nucleated. The total number of CD34 positive cells was determined on the unprocessed sample and the cell fraction obtained from the interface and pellet after processing of the PCPB samples in the density gradient solution. 54

Example 6

Colony forming assay (CFU)

Functional characteristics of CD34 + cells in a cell sample were determined by the colony forming assay (CFU). This assay provides quantification of the number of hematopoietic progenitor cells engaged in the cellular solution.

Cells were diluted to a concentration of 105 cells in commercially available semi-solid methylcellulose containing various colony stimulating factors and erythropoietin (H4433 MethoCult medium, Terry Fox Laboratories, Vancouver).

After 14 days of culture at 37øC, erythroid colonies (CFU-E, BFU-E), granulocytes / macrophages (CFU-GM) and mixed colonies (CFU-GEMM) were counted under an inverted microscope (40x) with Standard methods. The total number of CFUs present in a cell mix was defined by totalizing the different types of counted colonies.

Example 7

Determination of total cell recovery The percentage of cell recovery was determined by the formula:% Recovery = 100 x

The percent recovery of CD34 cells was determined by the formula:% Recovery = 100

Number of CD34 cells at start The percentage recovery of clonogenic cells was determined by the formula:% Recovery = 100 x

Nr. Clonogenic cells (CFUI after manipulation No. of clonogenic cells (CF ll) at start

Example 8

Effect of radiation on hematopoietic cell recovery The effect of organosilanated colloidal silica particles on the integrity of hematopoietic cells was tested using the protocol described in Example 4 as " Methods 3 & 4 ". OCS particles, OCS particles supplemented with 0.5% PVP, or PBS supplemented with 0.5% PVP (5 ml total volume) were subjected to radiation in a 50 ml conical tube centrifuge by means of a dose of gamma radiation (2.5-3.5 megaRad). In these experiments, the density of the OCS particle suspension used was 1.0605 g / ml, a density which is lower than that of the cells to be isolated, to facilitate subsequent pelletisation and quantification of cells present in the tube.

An aliquot containing 3x107 PBPC cells, isolated as described in example 1, was added to each tube, and the initial number of cells was noted. Cells were incubated for 30 minutes at room temperature, with gentle mixing at 15 minute intervals. The cell mixture was then centrifuged at 550 x g for 10 minutes. Following centrifugation, the supernatant was discarded and the pellet was resuspended in 5 ml Iscove medium supplemented with 10% fetal calf serum. The cell mixture was then incubated for 24 hours at room temperature and then tested in the CFU assay described in Example 6. The results of this experiment are illustrated in Figure 1.

Example 9

Sterilized centrifuge assembly

A sterile centrifuge kit was prepared to separate CD34 + hematopoietic progenitor cells from peripheral blood mononuclear cells (PBPC) as follows: a density gradient suspension based on OCS particles was prepared as described in Example 1. An stock solution was prepared by mixing 12 parts of the OCS particle suspension with a part of calcium and magnesium-free buffered saline phosphate (PBS) and concentrated PVP (PVP-10; Sigma) to prepare a medium having a density of 1.0605 ± 0.0005 g / ml, pH 7.4, an osmolality of 280 mOsm / kg H2 O, a concentration of 4% PVP. The density of the OCS particle solution can be adjusted in a densitometer to accurately define its accuracy up to at least the fourth decimal place. 15 ml of this suspension was added to a 50 ml polypropylene conical centrifuge tube (Baxter). The tube was sealed with the standard screw cap supplied with the tube. The tube and its contents were then exposed to gamma radiation (2.5-3.5 megaRad) by placing it in a chamber. The resulting sterile centrifugation kit was used in the separation of CD34 + cells from PBPC, as described in example 4 herein.

Example 10

Enrichment of CD34 + cells from cell mixtures A. Preparation of density gradients A.l. PERCOLL® density gradient

A PERCOLL® solution was purchased from Pharmacia Biotech (Uppsala, Sweden) and stored at 4 ° C according to the seller's recommendation. A stock solution was prepared by mixing 12 parts of PERCOLL with 1 part of calcium-free buffered saline phosphate (PBS) and 10 x (PBS). The pH of the solution was adjusted to 7.4 and the osmolality to 280 mOsm / kg H2 O. For use in separating CD34 + cells in a cell mix, the stock solution was further diluted with PBS without calcium and magnesium at a density of 1.0605 ± 0.0005 g / ml, and used at room temperature. It was crucial to adjust the gradient density to an accuracy of ± 0.0005 g / ml of 1.0605 g / ml to ensure reproducibility and accuracy of cell separation. This was done with a high precision digital densitometer such as DMA48 (Anton PAAR USA, Ashland, VA). All procedures were performed under sterile conditions and at room temperature. 58 A. 2. Density gradient of organosilanized colloidal silica (OCS)

A suspension of OCS particles is a suspension of organosilanized colloidal silica particles prepared as described by Dorn (US Patent 4,927,749). OCS particles having a size in the range of 3-22 nm, and particularly in the range of 13-18 nm, form a stable suspension which acts as a density gradient medium to separate heterogeneous mixtures of cells based on the floc density. Unless otherwise noted, the OCS particle suspension was prepared in physiological saline to have a density of 1.0605 g / ml and an osmolality of 280 mOsm / kg H2 O as defined for cell enrichment CD34 + from PBSC samples. The suspension was produced under GMP conditions and sterilized by gamma radiation in the presence of at least 0.05% polyvinylpyrrolidone (PVP-10; Sigma Chemical Co., St. Louis, MO). PVP was added as a dry powder to a suspension of OCS particles preset to approximately the final desired specific gravity. PVP concentrations of at least 0.05% and up to 10% are preferred for use in the methods, device and assembly of the invention. B. Density gradient of leukocyte layers of blood and bone marrow Separated peripheral blood (apheresis) was applied directly on the density gradient. However, unless otherwise noted, aspirates of whole blood 59 and bone marrow were processed to a leukocyte layer (removal of red blood cells) according to Standard methods, before being applied to the density gradient.

Separated blood samples (apheresis) were layered on a pre-adjusted PERCOLL® gradient at a density of 1.0605 ± 0.0005 g / ml, an osmolality of 280 mOsm / kg H2 O, and a pH of 7.4 in a 50 ml conical trap tube or in a commercially available tube. The trap tube contained a constriction at one site so that approximately 15 ml of PERCOLL® was in the lower compartment and 5 ml of PERCOLL® over the constriction. It was critical to completely fill the volume under the constriction with PERCOLL® to prevent the formation of air bubbles. Typically, 20 ml of separate blood samples (apheresis) are layered over this gradient. The tube was centrifuged at 850 x g for 30 minutes at room temperature. Cells housed at the interface of the gradient, i.e. above PERCOLL®, were collected by pouring the entire contents of the upper tube compartment into another 50 ml tube. Cell pellet leakage in the zone under the constriction was avoided when the tube was inverted.

To compare the cell separation method described in the preceding paragraph with conventional separation methods, the test samples were also layered in a Ficoll-Hypaque (Pharmacia). The density of the FICOLL stock solution was 1.077 ± 0.01 g / ml and the osmolality at 320 mOsm / kg H2 O as published by the seller.

Bone marrow cells were prepared as described in example 2. C. Density-Adjusted Cell Sorting (DACS) C.l. Preparation of carrier particles

Silica microspheres (1.4 micrometers) obtained from Bangs Laboratories, Carmel, IN were washed with concentrated HCl for 2 hours at room temperature and vortexed every 15 minutes to agglutinate. After washing, the microspheres were centrifuged at 850 g for 5 minutes. The HC1 containing the supernatant was decanted and the microspheres were washed with deionized H2 O subjected to intense vortex to split the agglutination.

The microspheres were incubated at room temperature overnight in concentrated nitric acid with constant stirring using a magnetic stirrer. The microspheres were then centrifuged at 850 g for 5 minutes and washed 3 times with deionized water using 50 ml deionized H2 O at each step. The microspheres were subjected to intense vortexing between washes to avoid agglutination of microspheres. To prevent microbacterial contamination, the microspheres were stored at 0-4øC in deionized H 2 O until further use.

To silanize the microspheres, either 3-aminopropyltriethoxysilane, (3-iodopropyl) trimethoxysilane or [1-9trimethoxysilyl) -2 (m- (or p) chloromethyl) phenyl] ethane were used. 40 ml of silane solution (a 10% solution in ethanol in 95% ethanol / deionized water) were added per 4 g of microspheres. The microsphere bead was rotated vertically for 1 hour at room temperature. The microspheres were centrifuged at 850 g for 5 minutes and the excess silane was removed by washing using 95% ethanol / deionized H 2 O in a volume of 100 ml. The microspheres were subjected to intense vortexing between each washing step to prevent agglutination of microspheres. After the washing step, the microspheres can be dried and stored. Alternatively, the microspheres can be stored in 95% ethanol / deionized H 2 O in the cold, which prevents agglutination of the microspheres.

The silanized microspheres were incubated overnight in 2.5% glutaraldehyde at room temperature. The next day the microspheres were centrifuged at 850 g for 5 minutes and the free glutaraldehyde was removed by washing with deionized H 2 O in a volume of 100 ml per 5 g of microspheres. The microspheres were subjected to intense vortexing between each washing step to avoid agglutination of microspheres. The antibody was added to the excess aminopropyl microspheres at least 3 mg / m 2 of the total surface of the microsphere and rotated vertically overnight at room temperature. The next day the microspheres were centrifuged at 850 g for 5 minutes and the free protein was removed by washing with 100 ml deionized H2 O. The microspheres were subjected to intense vortexing between each washing step to avoid agglutination of microspheres. The microspheres were stored in deionized H2 O containing 0.1% sodium azide in the cold. The final suspension of microspheres should contain 70-90% of single microspheres and 10-30% of predominantly double and triple microspheres. The binding efficiency of the antibody-conjugated microspheres (in terms of the percentage of microspheres that are coated) can be determined using flow cytometric analysis and a fluorosceinated antibody directed to the coupled antibody. Alternatively, the antibody may be added directly to the silanized microspheres without the binding and glutaraldehyde. C.2. Anti-CD45 coated microspheres The separated blood product (apheresis) was incubated with 1.4 μm aminopropyl glass microspheres (Bangs Laboratories Inc., Carmel, IN) which were coated with glutaraldehyde with an anti-CD45 antibody (Clone ALB- 12, Biodesign International, Kennebunk, ME) for 45 minutes at room temperature. The whole blood cell mixture was layered into OCS + 0.5% Pip (1.0605 ± 0.0005 g / ml, 280 mOsm / kg H20, pH 7.4) in a 50 ml tube. C.3. Microspheres coated with goat anti-mouse IgG

Cells from the PBSC sample were counted by trypan blue exclusion, washed and resuspended in PBS supplemented with 0.1% bovine serum albumin (BSA) at a concentration of 107 cells / ml. Anti-CD3, CD4 and CD8 mAbs were added to give a final concentration of 3-10 kg / ml. The cells were incubated for 30 minutes on ice and then centrifuged. The pellet was resuspended in 0.1% BSA in PBS at a cell concentration of 107 cells / ml. DACS microspheres coated with goat anti-mouse IgG (GAM-IGG) were counted and added to the cells to give microsphere ratios: cells between 4: 1 and 8: 1. Microspheres and cells were incubated for 15 minutes. During incubation, the cell: microsphere mixture was resuspended gently every 5 minutes. The cell: microsphere mixture was then layered in the OCS density gradient suspension and centrifuged at room temperature at 850 x g for 30 minutes. The interface cells were harvested by decanting the spin tube, counted and prepared for FACS analysis. To examine non-specific depletion, a similar procedure was used, but incubation with the monoclonal antibodies was omitted.

Experiments to examine the efficacy of DACS for T lymphocyte clearance were performed using mobilized PBSC preparations and anti-CD3, CD4 and CD8 mAbs. T lymphocyte yield was monitored using an anti-CD2 mAb to avoid artifacts associated with mAbs recognizing the same epitopes on the differentiation antigens. D. Conjugation of cell attachment molecules to carrier particles D.l. Monoclonal antibodies

Phycoerythrin (PE-CD34) conjugated monoclonal antibodies (hematopoietic progenitor cell marker) and fluorescein isothiocyanate (FITC-CD45) conjugated anti-45 monoclonal antibodies (pan-leukocyte marker) were obtained from Becton, Dickinson, Inc., San Jose, CA). Unconjugated antibodies directed to CD45, CD16 (granulocytes, monocytes), CD3 (T lymphocytes), CD14 (monocytes) were prepared in the laboratory, according to methods well known in the art.

Anti-CD3 mAbs (Leu4 clone), CD4 and CD8 were obtained from Dr. E. E. Engleman of Stanford University Hospital Blood Bank. In addition, FITC-CD2, PE-CD34 and FITC-CD45 were purchased from the same seller. 64

Antibodies were conjugated to either goat anti-mouse magnetic microspheres or goat anti-mouse coated aminopropyl glass microspheres by incubation overnight at room temperature. Alternatively, the antibodies could be attached directly to such microspheres without the goat anti-mouse bridge, or could be coupled via an avidin-biotin coupling reaction. In addition, the antibodies could be cleaved into Fab2 fragments to reduce nonspecific binding of cells to the microspheres through their Fc portion. D. Goat anti-mouse IgG (GAM-IGG) IgGI mouse conjugated either to FITC or to PE were purchased from Becton Dickinson (San Jose, CA) and used as an isotope control. Several different procedures were used to prepare goat-coated DACS anti-mouse IgG (GAM-IgG) microspheres (Biodesign International). The aim was to maximize specific depletion and minimize non-specific depletion during DACS. The DACS microspheres were covalently coated either with GAM-IgG or F (ab) 2 GAM-IgG (East Coast Biologicals) in a random orientation or GAM-IgG in a specific orientation to maximize the availability of the Ig binding sites for binding to the antigen. In addition, some microsphere preparations were coated with polyethylene glycol (PEG) (Shearwater) to reduce non-specific affinity of the microspheres for the protein or cells. The different chemical formulas are defined below. GAM-IgG or F (ab> 2 random GAM-IgG 65

The silica microspheres were coated with aminopropyltriethoxysilane (Sigma) having -NH2 groups available to be conjugated to the -CHO of glutaraldehyde (Sigma). According to the chemical formula used, glutaraldehyde then bridges and binds the -NH2 group on lysines along the GAM-IgG or F (ab) 2 GAM-IgG molecules, when the antibody and glutaraldehyde were added to the preparation. The orientation of the protein molecules relative to the microsphere was considered random, based on the fact that any lysine with an -NH2 side chain potentially conjugates with the glutaraldehyde -CHO. This gives an ATES-GLUT-IgG chain with GAM-IgG or F (ab) 2 in a random orientation. GAM-IgG in a specific guideline

The silica microspheres were coated with aminopropyltriethoxysilane having -NH2 groups available to be conjugated to the -CHO group of the carbohydrate in the glycosylated region of the FC portion of the GAM-IgG molecule. The GAM-IgG molecule is only glycosylated in the Fc portion so that the GAM-IgG binding to the surface of the microsphere along this region has the effect of directing the IgG molecule with the Fc portion near the surface of the microsphere and the IgG binding sites away from the surface. This treatment theoretically increases the number of binding sites available to bind antigens and decreases the availability of the FC moiety to bind Fc receptors. This provides an ATES-CHO-IgG chain with IgG in a specific orientation. 66 GAM-IGG randomized PEG-aldehyde

The silica microspheres were coated with aminopropyltriethoxysilane having -NH2 groups available to be conjugated to a -CHO group of an aldehyde-PEG-aldehyde molecule in which the PEG is sandwiched between two aldehyde residues. The GAM-IgG then binds to the other -CHO residue to give an ATES-CHO-PEG-CHO-IgG chain with IGG in a random orientation. E. Analyzes Antibody staining, FACS analysis and CF assays were performed as described in examples 5 and 6. F. Long-term culture initiator cell assay (LTC-IC) / functional determination of non-committed CD34 + cells The number of progenitor hematopoietic cells not committed in a cell mix was determined by the long-term initiation culture. The cells were seeded in a radiation-treated stroma feeder layer and a CFU determination was made as a function of time. Hematopoietic stem cells were able to self-renew and give rise to CFUs in this system for a period exceeding 5 weeks. Long-term cultures of bone marrow stroma were started in 96 well plates (106 cells in 200 μΐ / well) in α-MEM medium supplemented with 12.5% horse serum, 12.5% calf serum 2 mM 1-glutamine, 0.2 mM i-inositol, 20 μl folic acid, 10-4 M 2-mercaptoethanol and maintained at 33 ° C in a humidified atmosphere. At 67 week intervals, half of the medium was removed and replaced with an equal volume of fresh medium. After two weeks of culture, the confluent stroma layers were gamma irradiated (2000 rad) to kill endogenous hematopoietic cells. Unfractionated samples and cell preparations after separation were seeded into the stromal radiation layers in the same medium supplemented with 10-6 M hydrocortisone. After 5 weeks of culturing, the adherent and non-adherent cells were collected and filtered in the CFU assay, as in part G, above. G. Natural killing cell (NK) assay

K562 target cells were labeled with 100 pCi51Cr for 1 hour at 37 ° C, and then washed 4 times and counted. The target cells were co-cultured for 4 hours in multi-well plates of a V-bottom 96 well with separate blood (apheresis) and cells from different fractions after cell sorting. The effector and target cells were mixed in different ratios of 100: 1, 50: 1, 25: 1 and 12: 1. For example, the 100: 1 ratio contained 5x105 cells and 5x103 target cells. After the incubation period, 100 μΐ of the supernatant was collected and counted in a scintillation counter. The maximal and spontaneous release of 51CR was measured by counting both the 50 μg of the target stock and the supernatant of the effectors per se respectively. Percent cytotoxicity was determined according to the formula:

Percent Cvtotoxicitv = with experiment - with spontaneous release_ 3 3 cpm maximai release - cpm spontaneous release 68 J. Mixed lymphocyte culture and natural suppressor cell activity

Cells from two different leukocyte layers were mixed in a 96-well smooth bottom multi-well plate with 105 cells each. One of the leukocyte layers received 3000 rad and was referred to as " stimulators ". The other leukocyte layer was used untreated and referred to as

" responders ". Separated peripheral blood products (apheresis) (APBL) or cells of fractions of different density were added to these co-cultures at 105 cells per well. These cells are referred to as " suppressors " and received 1500 rad before being added to the MLR. Cells were cultured for 5 days and then pulsed with [H] -thymidine (1æCi per well). 18 hours later, the cells were harvested and the amount of incorporated thymidine determined in a scintillation counter. Percent suppression induced by suppressor cells was determined by the formula:

Percent Suppression with control - with exoeciment cpm experiment

Example 11

Debug peripheral blood tumor cells The human cell lymphoma SU-DHL4 cell line (Stanford University Blood Bank, Stanford, California) was used as a standard experimental model for modeling peripheral blood tumor cell contamination. This cell line has a reciprocal translocation between the long arms of human chromosomes 14 and 18, resulting in a juxtaposition of the bcl-2 proto-oncogene in the contiguous zone of the immunoglobulin (Ig) heavy chain gene. This translocation, which leads to high levels of the bcl-2 gene product, is found in 85% of small cell cleaved follicular lymphomas and 25% of diffuse large cell lymphomas. The breakpoint of translocation in the bcl-2 gene is located either in the major breakpoint region (MBR) spreading approximately 150 bp in the 3 '(UTR) region of the bcl-2 gene or in the minor cluster region (MCR) of 500 bp located approximately 20 kb downstream of the bcl-2 gene. Using pairs of translocation-specific primers and quantitative PCR limiting the techniques of dilution analysis known in the art, depletion of SU-DHL4 cells in the DACS clearance experiments can be monitored. The human adenocarcinoma cell line SK-BR3 (American Type Culture Collection (ATTC), Rockville, MD) was used to model the infiltration of peripheral blood breast cancer cells. These cells react with the mouse anti-human monoclonal antibody HR2 / Neu, indicating that they possess Her2 / Neu antigen. Prior to being transformed into normal leukocyte layers, SK-BR3 cells were preloaded with calcein AM, a fluorogenic substrate, to be enumerated by fluorescence activated cell sorting (FACS) analysis. Alternatively, or in addition, breast tumor cells contain estrogen receptors and react with antibodies to these receptors and other receptor binding molecules such as estrogen and estrogen agonists. For purposes of the present invention, such cells may be captured, as in the methods of clearance of tumor cells described herein, by any molecule that binds with sufficient affinity for such a tumor cell antigen including, but not limited to, antibodies , estrogen or estrogen agonists. In this context, a sufficient binding affinity is an affinity approximately equal to or comprised within 2 log units of binding affinity of Her2 / Neu mouse anti-human monoclonal antibody (HB 8694, 520C9; ATTC Rockville, MD).

PC Analysis

A high molecular weight genomic DNA was created using the QIAmp Blood Kit (Qiagen, Chatsworth, Ca.) and quantified by optical density at 260 nm. PCR reactions were done using a Perkin-Elmer Model 9600 thermal cycler for 35 cycles. Each amplification cycle consisted of 94Â ° C (15Â ° C), 55Â ° (60Â °) and 72Â ° C (15Â °), with a 72Â ° C extension (6 minutes) after the final cycle. For reactions with internal primers, 1 ml of the first-round reaction was used as the template. The following oligonucleotides were used for sample analysis: (i) 5-CAGCCTTGAAACATr-GATGG-3 (MBR) (ii) 5-ACCTGAGGAGACG-GTGACC-3 (JH Consensus) (iii) 5-ATGCTGT-GGTTGATATTTCG-3 (MBR) (iv) 5-ACCTGAGGAGACGGTGACC-3 (JH Consensus s). Amplification products were visualized by agarose gel electrophoresis.

FACS Analysis

SK-BR3 cells were preloaded by incubating them in AM calcein and then seeded into normal peripheral blood mononuclear cells with a final frequency of 2%. The DACS was run as described below, using the mouse anti-human monoclonal antibody HR2 / Neu. Aliquots of the experimental samples were analyzed by FACS analysis using the LYSYS II program. For analysis, 100,000 events were collected for each sample.

Density-Adjusted Cell Classification (DACS)

Peripheral blood progenitor cells (PBPC) and leukocyte layers were obtained from Bone Marrow

Transplantation Laboratory, Paio Alto, CA, Stanford University Blood Bank, Paio Alto, CA. Monoclonal antibodies to CD9, CD10; CD19 and CD20 were obtained from Caltag Laboratories (So. San Francisco, CA). The anti-CD-20 mAb (IF5 clone) and Her2 / Neu anti-human mAb (HB 8694, 520C9) and the SK-BR3 adenocarcinoma cell line were obtained from the ATCC (Rockville, MD). IgG1 mouse and fluorescein isothiocyanate-conjugated anti-human CD45 mouse (FITC), as well as the phycoerythrin-conjugated mouse IgG1 and mouse anti-human mouse CD34 antibodies were obtained from Becton Dickinson (San Jose, CA). FACS procedure and flow cytometric analysis of samples were performed as described above.

Debugging and electrophoretic analysis of peripheral blood cells

A peripheral blood leukocyte layer, spiked (5%) with ot (14:18) + B cell lymphoma SUDHL-4 cell line was incubated with an anti-B cell mAb cocktail for CD9, CD10, CD19 and CD20. After washing, the cell mix was incubated with GAM-gGG-coated DACS microspheres in a ratio of 8: 1 microspheres: cells and then loaded into an OCS suspension with densities ranging from 1.0605 to 1.0720 with comparable results. Following centrifugation for 30 minutes at 850 g, the DNA was prepared from the cells at the interface and analyzed by standard gel electrophoretic methods with different concentrations (1 mg, 0.1 mg, 0.01 mg, 0.001 mg) for 2 rounds of PCR using T (14; 18) using specific primers. The number of translocations t (14:18) was further determined by limiting the PCR dilution. As a control, the same protocol was used on non-spiked cells, spiked SUDHL-4 sample and spiked SUDHL-4 sample treated with DACS microspheres without prior incubation with cocktail anti-B cell mAbs. Centrifugation of the cell mixture on the OCS suspension did not result in a reduction of the PCR signal. While the DACS microspheres alone induced a small reduction of the PCR signal, the PCR signal was completely absent in the spiked SUDHL-4 sample, which was treated with either the mAb cocktail of anti-B cells or with DACS microspheres. The determination of the number of translocations t (14; 18) by limiting the PCR dilution showed that in the last experimental group the number of SUDHL-4 cells was reduced by 2-3 logs (100-1000).

Preliminary experiments indicate that an additional round of DACS in this context further decreases the number of SUDHL-4 cells by approximately one additional log.

Example 12

Dendritic cell isolation 73

Leukocyte layers prepared from a blood unit of healthy, healthy volunteer donors HLA-A02 01 were obtained from the Stanford University Blood Center (Stanford, CA). Cells were harvested from leukopacs, diluted to 60 ml using CA ++ / MG ++ free buffered saline phosphate (D-PBS; Gibco Laboratories, Grand Island, NY) and layered onto two 15 ml columns of OCS separation medium 1.0720 g / ml, pH 7.4, 280 mOsm / kg H2 O) in 50 ml centrifuge tubes, preferably trap tubes. The tubes are centrifuged at 1000 x g for 35 minutes at room temperature. The spin batch is allowed to stop without braking and peripheral blood mononuclear cells (PBMC), present at the interface, are harvested.

The PBMCs are resuspended in D-BPS, centrifuged once at 650 x g for 10 minutes and twice more at 200 x g for 5 minutes to remove the platelets. Platelet-free PBMCs are resuspended in 60 ml of D-PBS, placed over two 15 ml columns of OCS (density 1.0610 g / ml, 280 mOsm / kg H20) in a trap tube of the present invention and centrifuged at 650 xg for 25 minutes at 4 ° C without braking. The resulting interface (mainly monocytes) and cell pellets (mainly lymphocytes) are harvested and washed with D-PBS by centrifugation at room temperature (once at 650 x g for 10 minutes and then twice at 200 x g for 5 minutes).

In situations where dendritic cells are used to generate peptide-specific cytotoxic T lymphocytes (CTLs) to elucidate their antigen presentation function, the interface fraction (mostly monocytes) is resuspended in a cold human AB serum (Irvine Scientific, Santa Ana, CA) to which an equal volume of 80% AB serum and 20% dimethyl sulfoxide (DMSO) (Sigma Chemical Company, St. Louis, MO) is added dropwise. The resulting cell suspension is aliquoted into cryovends and frozen in liquid nitrogen. Monocytes can be used for a new CTL stimulation for expansion. The pelleted fraction is resuspended in 100 ml of AB Culture Medium, inoculated into T7-5 tissue culture flasks and cultured in a humidified 5% CO 2 incubator for 40 hours. Following incubation, the non-adherent cells are harvested moderately with a pipette, washed and resuspended at a concentration of 2-5 x 10 6 cells / ml in AB culture medium. The cell suspension is placed over four columns of a 4.0 ml OCS separation medium (density 1.0565 g / ml, pH 7.4, 280 mOsm / kg H2 O) in an AB culture medium and centrifuged at 650 x g for 20 minutes at room temperature, without braking. The interface and pelleted cells are collected and washed in AB Culture Medium (Basal medium RPMI-1640, Gibco Laboratories, Grand Island, NY) by centrifugation once at 650 x g for 10 minutes and then twice at 200 x g for 5 minutes at the temperature environment. The yield and viability of both cell fractions is estimated by counting in a hemocytometer using trypan blue exclusion. The purity of the dendritic cells in the interface fraction is quantified following analysis on a flow cytometer (FACS). Dendritic cells are characterized as negative for CD3 (T lymphocytes), CD14 (monocytes), CD16 (NK cells) and CD20 (B cells) phenotype markers and positive for HLA class II expression using double HLA-DR staining FTTC channel) and a cocktail of CD3, CD14, CD16, CD20 (at 75

channel PE). Double staining with IgG2a in the FITC and PE channels can be used as isotype control. Cell morphology can also be assessed using photomicroscopy. The enriched DC fraction contains veined large cells with cytoplasmic processes extending from the cell surface, characteristic of DC.

Lisbon, 20 December 2011. 76

Claims (26)

A composition for use in cell separation, comprising a cell separation medium based on silanized colloidal silica particles containing at least about 0.05% polylactam, wherein said separation medium containing said polylactam is sterilized . The composition of claim 1, wherein said cell separation medium based on colloidal silica particles is comprised of organosilanated silica particles. A composition according to claim 1 or 2, wherein said composition is sterilized by autoclaving or by ionizing radiation. The composition of claim 3, wherein said ionizing radiation is selected from gamma rays and electron beam radiation. Composition according to any one of claims 1 to 4, wherein said polylactam is polyvinylpyrrolidone present in a concentration of from about 0.1 to about 10%. A composition according to any one of claims 1 to 5, wherein said particle is a microsphere having a diameter between about 0.003 and 50 micrometers. 1 A composition according to any one of claims 1 to 6 for use in the isolation of a rare blood cell selected from a cell mixture, wherein said suspension of silanized colloidal silica particles has a specific gravity of about ± 0, 0005 g / ml of said chosen cell. The composition of claim 7, wherein the rare blood cell is selected from the group consisting of CD34 + hematopoietic progenitor cells, dendritic cells, natural killing cells, natural suppressor cells, cytotoxic T cells, tumor cells and nucleated cells from fetal origin. The composition of claim 8, wherein said rare blood cell is a CD34 + hematopoietic progenitor cell, said cell mix comprises peripheral blood mononuclear cells, and said suspension of organosilanized colloidal silica particles has a specific density of 1.0605 g / ml. The composition of claim 8, wherein said rare blood cell is a CD34 + hematopoietic progenitor cell, said cell mix consisting of bone marrow cells, and said suspension of organosilanized colloidal silica particles has a specific density of 1.0605 g / ml. The composition of claim 8, wherein said rare blood cell is a dendritic cell, said cell mix comprises peripheral blood mononuclear cells, and said suspension of organosilanized colloidal silica particles has a specific density chosen from between the group consisting of 1.0770 g / ml, 1.0720 g / ml, 1.0650 g / ml, 1.0610 g / ml and 1.0565 g / ml. A cell separation assembly comprising a centrifugal vessel and a silanized colloidal silica particulate cell separation suspension according to any one of claims 1 to 11. A method of sterilizing a cell separation medium based on silanized silica particles, comprising subjecting said separation medium to autoclaving or ionizing radiation in the presence of at least 0.05% by weight polylactam. A process according to claim 13, wherein said polylactam is polyvinylpyrrolidone present at a concentration of from about 0.1 to about 10%. A process according to any one of claims 13 or 14, wherein said silica particle is an organosilanized colloidal silica particle. A method of isolating a rare blood cell selected from a cell mixture, comprising: layering the mixture containing said rare hematopoietic cell in the cell separation medium based on silanized colloidal silica particles according to any one of one of claims 1 to 6, and centrifuging said mixture with a centrifugal force sufficient for pelletizing the cells having a specific density that is higher than the specific density of the chosen cell. A process according to claim 16, wherein said mixture containing said rare cell and said cellular separation means is contained in a centrifugation tube having side walls and a closed bottom, a constriction element disposed within the tube , said constriction member placed and constructed to retain the liquid in the lower portion of the tube below the constriction member when the tube is inverted, and wherein said cell separation means has a specific density which is within ± 0 , 0005 g / ml of the specific density of the cell type chosen and said medium is contained in the lower part of the tube and extends above said constriction element to a level above an opening formed by said constriction element, so that the cells that are captured at the level of an interface between the cell separation medium and the lower density medium after centrifugation are discharged with the lower density medium when the tube is inverted. 4 A method according to claim 17, wherein said tube comprises an enclosed top, a first port at said top through which liquid material may be introduced into the tube, a second port at said top, and a liquid channel closed manner by establishing communication between the second port and the bottom of the tube beneath said constricting member. A method according to claim 16, in which the rare cell chosen is a functional hematopoietic progenitor cell, the cell mixture is a mixture of peripheral blood mononuclear cells, and the specific density of the cell separation density gradient suspension is of 1.0605 g / ml. A method according to claim 16, wherein the rare cell chosen is a functional hematopoietic progenitor cell, the cell mixture is a bone marrow cell mixture, and the specific density of the cell separation density gradient suspension is 1 , 0685 g / ml. A method according to claim 16, wherein the rare blood cell chosen is a tumor cell, the cell mixture is derived from blood or bone marrow, and wherein the cell separation medium has a specific density selected from the range of 1.0490 to 1.0580 g / ml. A method according to claim 16, wherein the rare blood cell chosen is a dendritic cell and wherein said cell separation medium has a specific density selected from the group consisting of 1.0770 g / ml, 1.0720 g / ml, 1.0650 g / ml, 1.0610 g / ml and 1.0565 g / ml. A method according to claim 16, wherein the rare blood cell chosen is a cytotoxic T lymphocyte and wherein said cell separation medium has a specific density selected from the group consisting of 1.0720 g / ml, 1.0610 g / ml and 1.0565 g / ml. A process according to any one of claims 16 to 23, further comprising adding to the cell mix a suspension comprising a silanized silica particle to which a specific binding molecule of a cellular antigen is attached. A method according to claim 24 for use in the depletion of T lymphocytes, wherein said specific binding molecule of a cellular antigen binds to an antigen present on said T lymphocytes. The method of claim 25, wherein said specific binding molecule of a cellular antigen is selected from the group consisting of mouse anti-CD3, anti-CD4 and anti-CD8 monoclonal antibodies. Lisbon, 20 December 2011. 6