WO2011049439A1

WO2011049439A1 - Method for selecting bone forming mesenchymal stem cells

Info

Publication number: WO2011049439A1
Application number: PCT/NL2010/050690
Authority: WO
Inventors: Jan De Boer
Original assignee: Universiteit Twente
Priority date: 2009-10-19
Filing date: 2010-10-19
Publication date: 2011-04-28

Abstract

The invention relates to a method for determining the bone forming potential of a mesenchymal stem cell comprising determining the level of expression of a marker gene, characterised in that said gene is CADM1.

Description

Title: Method for selecting bone forming mesenchymal stem cells

FIELD OF THE INVENTION

The invention is in the field of bone tissue engineering and transplantation. More particularly, the invention relates to methods for selecting an osteogenic mesenchymal stem cell, a method for selecting an individual for a treatment, a method for treatment of a patient in need of a mesenchymal stem cell, a method for testing the bone forming potential of a cell and a method for screening compounds interfering with bone formation.

BACKGROUND OF THE INVENTION

Many human diseases are caused by failure of tissue function, with well know examples such as diabetes, damage inflicted by myocardial infarcts and degeneration of the hip joint. The disciplines of tissue engineering and cell therapy aim at restoring worn-out or diseased tissues for which the patient's body represents a source of suitable cells, i.e, by autologous cell therapy.

For instance, a much used source of autologous cells in the field of bone tissue engineering are human mesenchymal stem cells (hMSCs), also referred to as multipotent stromal cells. Because hMSCs are easily isolated from bone marrow aspirates and expanded in vitro, these cells are used for various cell-based therapeutic approaches.

hMSCs are multipotent cells which are able to differentiate, depending on the stimulus, into several lineages including the osteogenic, chondrogenic and adipogenic lineage in vitro. Osteogenic differentiation of hMSCs can be characterized by a cuboidal morphology, expression of alkaline phosphatase (ALP), reactivity with anti- osteogenic cell surface monoclonal antibodies, modulation of osteocalcin mRNA production and the formation of a mineralized extracellular matrix containing hydroxyapatite. Growth factors and other diffusible molecules such as dexamethasone (dex), cyclic adenosine 3', 5'— monophosphate (cAMP), 1,25-dihydroxy- vitamin D (vit D3) and bone morphogenetic protein 2 (BMP- 2) are used to drive osteogenic differentiation of hMSCs. For bone tissue engineering, ectopic bone formation was demonstrated by seeding hMSCs onto porous calcium phosphate scaffolds followed by subcutaneous implantation into immune- deficient mice. Although it is clear that hMSCs can in principle be used for bone tissue engineering in animal models, clinical application is hampered by the fact that there is a large variation in in vivo bone forming potential (ultimately visible as bone formation) of hMSCs between different donors.

Therefore, it would be beneficial to predict the in vivo bone forming potential of mesenchymal stem cells. Current methods include the use of collagen type I and osteoprotegerin or ALP expression as markers for bone forming potential. However, these parameters are no strong predictive markers for the true performance of mesenchymal stem cells when used as a graft in bone tissue engineering. There is still a need for an easy and reliable method for determining the bone forming potential of a mesenchymal stem cell in general. SUMMARY OF THE INVENTION

The present inventors have now found an in vitro diagnostic marker which is able to predict the in vivo bone-forming capacity of hMSCs. The present inventors have used a microarray based approach on isolated RNA to find in vitro diagnostic markers which are able to predict the in vivo bone-forming capacity of hMSCs.

Therefore, a bank of hMSCs was developed from 62 different donors, various in vitro differentiation assays were performed and the corresponding in vivo bone formation was analyzed. In addition, the gene expression profile of the hMSCs from the different donors was determined using microarray technology to correlate with the in vivo bone-forming capacity. Here, it is shown that expression of the CADM1 gene correlates positively with in vivo bone forming potential of hMSCs.

The invention provides in a first aspect a method for determining the bone forming potential of a mesenchymal stem cell comprising determining the level of expression of a marker gene, characterised in that said gene is CADM1.

In a preferred embodiment, said bone forming potential is the in vivo bone forming capacity.

In another preferred embodiment, said mesenchymal stem cell is obtained from bone marrow or blood. In a preferred embodiment of a method according to the invention, said level of expression of CADM1 is determined by qPCR.

The invention further provides a method for selecting a mesenchymal stem cell based on bone forming potential of said cell, comprising performing the method of the invention and selecting said bone forming mesenchymal stem cell based on a comparison of said level of expression with a control level.

The invention further provides a method for enriching a sample of

mesenchymal stem cells based on bone forming potential of said cells, comprising performing a method for selecting a bone forming mesenchymal stem cell within a plurality of cells of said sample, and separating the selected cell from said sample. Such a separation step may for instance be performed by using suitable cell sorting equipment, such as a fluorescence associated cell sorter (FACS). Such methods are well known in the art, and described in the examples below.

The invention further provides a method for selecting an individual suitable as a donor of bone forming mesenchymal stem cells, comprising the steps of providing a sample comprising at least one bone forming mesenchymal stem cell from said individual, subsequently determining the bone forming potential of said mesenchymal stem cell in said sample using the method for determining the bone forming potential of a mesenchymal stem cell determining the bone forming potential of a mesenchymal stem cell of the invention and selecting said individual as a suitable donor based on a comparison of said level of expression with a control level.

The invention further presents a method for providing a graft of bone forming mesenchymal stem cells comprising the steps of selecting an individual suitable as a donor of bone forming mesenchymal stem cells by performing the method for selecting an individual suitable as a donor of bone forming mesenchymal stem cells of the invention, obtaining a sample of mesenchymal stem cells from said individual, and optionally enriching said sample based on bone forming potential of said cells by performing a method for enriching a sample of bone forming mesenchymal stem cells of the invention and/or performing further treatments on said sample to make it suitable for use as a graft.

The invention further provides a method for treatment of a patient in need of a graft of mesenchymal stem cells, comprising the steps of providing a graft of mesenchymal stem cells by performing a method for providing a graft of mesenchymal stem cells of the invention, and administering said graft to said patient.

The invention further provides a method for testing whether a compound interferes with the bone forming capacity of a mesenchymal stem cell comprising steps of providing said compound to said mesenchymal stem cell in vitro; determining the expression of CADM1 in said mesenchymal stem cell and determining whether a compound interferes with the bone forming capacity based on said expression.

In a preferred embodiment of a method of the invention said step of determining said level of expression further comprises a step of performing a bone formation bioassay.

In a preferred embodiment of a method for determining the bone forming potential of a mesenchymal stem cell, a method for selecting a bone forming mesenchymal stem cell, a method of enriching a sample of mesenchymal stem cells based on bone forming potential, a method for selecting an individual suitable as a donor of bone forming mesenchymal stem cells, a method of providing a graft of bone forming mesenchymal stem cells or a method for treatment of a patient in need of a graft of bone forming mesenchymal stem cells according to the invention, said control level is determined by an in vitro assay wherein the level of expression of CADM1 is correlated to the bone forming potential based on a bone formation bioassay.

Preferably, said bioassay comprises a mineralization assay.

In a preferred embodiment of any method of the invention, said mesenchymal stem cell (MSC) is a human MSC (hMSC). Preferably, said mesenchymal stem cell is proliferated in vitro, differentiated in vitro or otherwise cultured in vitro prior to determining the level of expression of said marker gene.

DESCRIPTION OF THE DRAWINGS

Fig. 1 depicts the identification of hMSCs. The identity of the hMSCs was confirmed according to the set of standards proposed by the Mesenchymal and Tissue Stem Cell Committee of the ISCT. A) The hMSCs as indicated in the Examples were adherent to tissue culture plastic; B) The cells were able to differentiate into the adipogenic (Oil red O staining), osteogenic (Alizarin Red staining) and chondrogenic (Alcian Blue staining) lineage in vitro; C) Histograms showing flow cytometric analyses of the hMSCs, demonstrating expression of cell surface markers,

representative for three different donors. Percentages of positive cells are indicated.

Fig. 2 shows the characterization of hMSCs: A) Proliferation; hMSCs were cultured and counted when reaching 70-80% confluence. Frequency of population doublings (PD) per day was calculated, passage 0- 1, for 61 donors in total. B) hMSCs were cultured in basic (con) or osteogenic (dex) medium during seven days. The percentage of ALP positive cells was determined using flow cytometry. Error bars represent the standard deviation. C) Mineralization; hMSCs were seeded at 5000/cm² and cultured in mineralization medium for three weeks. HC1 was used to release calcium and calcium deposition was measured and expressed as mg/dl sample. Error bars represent the standard deviation. D) Adipogenesis; hMSCs were cultured in adipogenic medium for three weeks. Adipogenic differentiation was visualized by staining with Oil red O, the color was extracted and measured spectrophotometrically. Error bars represent the standard deviation. E) In vivo bone formation of hMSCs; hMSCs were cultured on biphasic calcium phosphate (BCP) particles (200.000 cells/3 particles) in osteogenic medium for seven days and implanted s.c. in nude mice for six weeks. Out of 62 donors, 35 showed bone formation. Histomorphometric analysis demonstrated the large variation between donors.

Fig. 3 shows the correlation between clinical and cell biological data labels and the in vivo bone formation. A correlation curve was created by plotting the different parameters against the in vivo bone formation. An example is shown of proliferation doublings per day (A) or percentage of ALP positive cells (dex) (B). No correlation was found in all data labels. C) Receiver Operating Characteristic (ROC) curves to represent the trade off between the false negative and false positive rates for every possible cut-off. The AUC (area under curve) was calculated for all labels, which is a measure of the probability that a classifier based on this label would rank a randomly chosen positive donor higher than a randomly chosen negative donor.

Fig. 4 shows the classification analyses using AUC scores, describing for each clinical label the performance of the classifier for different bone are/ scaffold area thresholds. A) AUC score, giving the probability that a classifier based on this label would rank a randomly chosen positive donor higher than a randomly chosen negative donor. B) P- value, describing the significance of the AUC-score, based on permutation tests (10.000 permutations). Fig. 5 shows the classifier performance on the bone labels. Performance was determined using ROC curves, obtained by using the posterior probabilities on the validation samples of the leave-one-out cross-validation procedure. Features (probe sets) were ranked by using the SAM test statistic within the cross-validation loop. A) AUC (area under ROC curve) score for the microarray dataset, using a fixed number of features (probe sets) from the top of the ranked list. The top probe set in every cross-validation fold was CADMl. B) ROC performance curves of classifiers using either clinical features, microarray probe sets or the CADMl qPCR data. No feature selection (and thus cross-validation) was necessary for CADMl qPCR. For the other two datasets, we used the procedure as described above. The optimal number of genes to select from the top of the ranked list was determined using inner cross-validation.

Fig. 6 shows the classification analyses using AUC scores, describing for each CADMl probe the performance of the classifier for different bone are/ scaffold area thresholds. A: AUC score, giving the probability that a classifier based on this label would rank a randomly chosen positive donor higher than a randomly chosen negative donor. B: P- value, describing the significance of the AUC-score, based on permutation tests (10.000 permutations). C: Idem as B in black white presentation.

Figure 7 shows the mRNA sequence of CADMl.

Figure 8 shows the decreasing gene expression levels of CADMl between passage 0 and 2 of hMCSs. The gene expression was determined by means of PCR.

Unpaired t-tests were performed for the different passages per donor which showed a statistical significant decrease in 4 out of 6 donors tested (* p<0.05, ** p<0.01). It was observed that CADMl gene expression in hMSCs decreased during cell passaging. This phenomenon is analogous to the expression of previously identified specific mesenchymal stem cell markers like STRO-1 and NGFR (CD271). Both cell surface markers provide a useful tool for selecting hMSCs with high multipotential differentiation ability from human bone marrow. CADMl reveals a similar decreasing expression pattern over cell passaging as these two markers. Furthermore, it was observed that the entire population of hMSCs isolated from bone marrow, using the common markers set by the International Society for Cellular Therapy (ISCT), express CADMl. Based on the above, CADMl may further suitably be used as a marker to select mesenchymal stem cells from bone marrow samples disclosing their biological activity, eg. wherein the marker expression is a measure of the biological activity.

DETAILED DESCRIPTION OF THE INVENTION

Definitions

The term "bone forming mesenchymal stem cell" as used herein refers to a mesenchymal stem cell from which bone develops in vivo.

The term "bone forming potential" as used herein, refers to the predicted ability of a bone forming mesenchymal stem cell to form bone in vivo, or the chance that a bone forming cell will contribute to the formation of bone, either in terms of quantity and/or in terms of speed, when placed under the suitable conditions for bone formation, optionally preceded by conditions for differentiation, preferably in vivo. The term is used interchangeably with the term "osteogenic potential".

Bone as used herein refers to cortical bone, as well as to trabecular bone. The term "mesenchymal stem cell" (abbreviated "MSC"), as used herein, refers to a multipotent stem cell that is capable of differentiating into a variety of cell types. Cell types that MSCs have been shown to differentiate into in vitro and in vivo include osteoblasts, chondrocytes, myocytes, adipocytes, endotheliums, and beta- pancreatic islets cells. Preferably, human MSCs express CD105, CD73 and CD90. Preferably, said MSC are cells as defined using the standards proposed by the Mesenchymal and Tissue Stem Cell Committee of the ISCT. Preferably, said standard include the following three minimal criteria. First, a MSC must be plastic- adherent when maintained in standard culture conditions. Second, a MSC must express CD105, CD73 and CD90, and lack expression of CD45, CD34, CD14 or CD lib, CD79a or CD 19 and HLA-DR surface molecules. Third, MSC must differentiate to osteoblasts, adipocytes and chondroblasts in vitro. Highly preferred MSCs are osteogenic MSCs.

The term "sample of mesenchymal stem cells" as used herein refers to a biological sample containing a plurality of mesenchymal stem cells as defined above. Preferred biological samples comprise blood samples or bone marrow samples, preferably derived from the acetabulum or the iliac crest. Preferably, said sample contains more than 95% of cells expressing CD105, CD73 and CD90, preferably as measured by flow cytometry. More preferably less than 2% of the cells of said sample express CD45, CD34, CDllb, CD19 and HLA-DR, preferably as measured by flow cytometry. Preferably, the cells of said sample are able to differentiate into the osteogenic, adipogenic and chondrogenic lineage under standard in vitro

differentiation conditions, as described herein. Preferably, said ability to differentiate into the osteogenic, adipogenic and chondrogenic lineage is determined as described herein.

The term "gene", as used herein refers to a DNA sequence including but not limited to a DNA sequence that can be transcribed into mRNA which can be translated into polypeptide chains. The term refers to any DNA sequence comprising several operably linked DNA fragments such as a promoter region, a 5' untranslated region (5' UTR), a coding region (which may or may not code for a protein), and an untranslated 3' region (3' UTR) comprising a polyadenylation site. Typically, the 5' UTR, the coding region and the 3' UTR are transcribed into an RNA of which, in the case of a protein encoding gene, the coding region is translated into a protein. The gene usually comprises introns and exons and thus a gene may include additional DNA fragments such as, for example, introns.

"Expression" refers to the transcription of a gene into structural RNA (rRNA, tRNA) or messenger RNA (mRNA) with subsequent translation into a protein.

The term "level of expression" as used herein refers to the level of nucleic acids or to the level of protein.

The term "nucleic acid" as used herein, includes reference to a

deoxyribonucleotide or ribonucleotide polymer, i.e. a polynucleotide, in either single-or double-stranded form, and unless otherwise limited, encompasses known analogues having the essential nature of natural nucleotides in that they hybridize to single- stranded nucleic acids in a manner similar to naturally occurring nucleotides (e. g., peptide nucleic acids). A polynucleotide can be full-length or a subsequence of a native or heterologous structural or regulatory gene. Unless otherwise indicated, the term includes reference to the specified sequence as well as the complementary sequence thereof. Thus, DNAs or RNAs with backbones modified for stability or for other reasons are "polynucleotides" as that term is intended herein. Moreover, DNAs or

RNAs comprising unusual bases, such as inosine, or modified bases, such as tritylated bases, to name just two examples, are polynucleotides as the term is used herein. It will be appreciated that a great variety of modifications have been made to DNA and RNA that serve many useful purposes known to those of skill in the art. The term polynucleotide as it is employed herein embraces such chemically, enzymatically or metabolically modified forms of polynucleotides, as well as the chemical forms of DNA and RNA characteristic of viruses and cells, including among other things, simple and complex cells.

As used herein, the term "cell adhesion molecule 1 (CADM1) gene" refers to the gene encoding the CADM1 protein. The CADM1 gene is also known as"BL2", "ST17", "IGSF4", "NECL2", "RA175", "TSLCl", "IGSF4A", "Necl-2", "SYNCAM", "sglGSF", "sTSLC-1", "synCAMl", "MGC51880", "MGC149785", "DKFZp686F1789". The CADM1 gene (also identified herein by the HUGO Gene Nomenclature Committee *Cambridge, UK as HGCN:5951), as referred to herein, is also identified herein as a gene having the nucleic acid sequence as depicted in Figure 7 and naturally occurring allelic variants thereof. This molecule contains 3 immunoglobulin domains, a C- terminal transmembrane domain, and a short cytosolic tail with an intracellular type II PDZ binding domain and band 4.1 binding domain. This cell adhesion molecule is identified to have a role in multitude of different biological functions with

extracellular recognition. It is known for driving synapse formation in developing neurons, for mediating nerve-mast cell attachment, for being responsible for the direct contact between spermatogenic and Sertoli cells which is crucial for spermatogenesis, for its function as tumor suppressor gene in non- small-lung cancer.

The nucleic acid sequence of the CADM1 gene is known in the art. When herein reference is made to the CADM1 gene, it is meant to refer to all sequence variations of the CADM1 gene which may result in the expression of a functional CADM1 protein. There are two isoforms of the CADM1 protein, the amino acid sequences of which are known and can be found in, for example, GenBank accession number GL148664190 and GL148664211. The nucleotide sequence of the two transcript variants can be found in, for example, Genbank accession number

NM_014333.3 GL148664189 and NM_001098517.1 GL148664210. Preferably, the CADM1 gene according to the invention has a nucleic acid sequence having at least 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99% sequence identity with the sequence as depicted in Figure 7. The terms "stringency" or "stringent hybridization conditions" refer to hybridization conditions that affect the stability of hybrids, e.g., temperature, salt concentration, pH, formamide concentration and the like. These conditions are empirically optimized to maximize specific binding and minimize non-specific binding of primer or probe to its target nucleic acid sequence. The terms as used include reference to conditions under which a probe or primer will hybridise to its target sequence, to a detectably greater degree than other sequences (e.g. at least 2-fold over background). Stringent conditions are sequence dependent and will be different in different circumstances. Longer sequences hybridize specifically at higher

temperatures. Generally, stringent conditions are selected to be about 5°C lower than the thermal melting point (Tm) for the specific sequence at a defined ionic strength and pH. The Tm is the temperature (under defined ionic strength and pH) at which 50% of a complementary target sequence hybridizes to a perfectly matched probe or primer. Typically, stringent conditions will be those in which the salt concentration is less than about 1.0 M Na+ ion, typically about 0.01 to 1.0 M Na+ ion concentration (or other salts) at pH 7.0 to 8.3 and the temperature is at least about 30°C for short probes or primers (e.g. 10 to 50 nucleotides) and at least about 60°C for long probes or primers (e.g. greater than 50 nucleotides). Stringent conditions may also be achieved with the addition of destabilizing agents such as formamide. Exemplary low stringent conditions or "conditions of reduced stringency" include hybridization with a buffer solution of 30% formamide, 1 M NaCl, 1% SDS at 37°C and a wash in 2x SSC at 40°C. Exemplary high stringency conditions include hybridization in 50% formamide, 1 M NaCl, 1% SDS at 37°C, and a wash in O.lx SSC at 60°C. Hybridization procedures are well known in the art and are described in e.g. Ausubel et al, Current Protocols in Molecular Biology, John Wiley & Sons Inc., 1994.

Methods of the invention can in principle be performed by using any nucleic acid amplification method, such as the Polymerase Chain Reaction (PCR; Mullis 1987, U.S. Pat. No. 4,683,195, 4,683,202, en 4,800,159) or by using amplification reactions such as Ligase Chain Reaction (LCR; Barany 1991, Proc. Natl. Acad. Sci. USA 88:189- 193; EP Appl. No., 320,308), Self- Sustained Sequence Replication (3SR; Guatelli et al, 1990, Proc. Natl. Acad. Sci. USA 87:1874-1878), Strand Displacement Amplification (SDA; U.S. Pat. Nos. 5,270,184, en 5,455,166), Transcriptional Amplification System (TAS; Kwoh et al, 1989, Proc. Natl. Acad. Sci. USA Feb;86(4):1173-7), Q-Beta Replicase (Lizardi et al., 1988, Bio/Technology 6: 1197), Rolling Circle Amplification (RCA; U. S. Pat. No. 5,871,921), Nucleic Acid Sequence Based Amplification (NASBA), Cleavase Fragment Length Polymorphism (U.S. Pat. No. 5,719,028), Isothermal and Chimeric Primer-initiated Amplification of Nucleic Acid (I CAN), Ramification- extension Amplification Method (RAM; U. S. Pat. Nos. 5,719,028 and 5,942,391) or other suitable methods for amplification of DNA. Generally, in order to detect gene expression in the form of mRNA, the mRNA is first reverse transcribed into cDNA by reverse transcription using methods well known in the art, for instance based on the use of M-MLV reverse transcriptase from the Moloney murine leukemia virus or AMV reverse transcriptase from the avian myeloblastosis virus. Subsequently, the cDNA is then amplified by using for instance the PCR reaction.

In order to amplify DNA with a small number of mismatches to one or more of the amplification primers, an amplification reaction may be performed under conditions of reduced stringency (e.g. a PCR amplification using an annealing temperature of 38°C, or the presence of 3.5 mM MgC ). The person skilled in the art will be able to select conditions of suitable stringency.

The primers used for amplification of nucleic acids are selected to be

"substantially" complementary (i.e. at least 65%, more preferably at least 80% perfectly complementary) to their target regions present on the different strands of each specific sequence to be amplified. It is possible to use primer sequences containing e.g. inositol residues or ambiguous bases or even primers that contain one or more mismatches when compared to the target sequence. In general, sequences that exhibit at least 65%, more preferably at least 80% sequence identity with the target DNA oligonucleotide sequences, are considered suitable for use in a method of the present invention. Sequence mismatches are also not critical when using low stringency hybridization conditions.

The detection of the amplification products can in principle be accomplished by any suitable method known in the art. The detection fragments may be directly stained or labelled with radioactive labels, antibodies, luminescent dyes, fluorescent dyes, or enzyme reagents. Direct DNA stains include for example intercalating dyes such as acridine orange, ethidium bromide, ethidium monoazide or Hoechst dyes.

Alternatively, the DNA fragments may be detected by incorporation of labeled dNTP bases into the synthesized DNA fragments. Detection labels which may be associated with nucleotide bases include e.g. fluorescein, cyanine dye or BrdUrd. mRNA expression analysis may also be performed by expression profiling, using DNA microarrays by methods well known in the art.

Alternatively, the expression level of the CADM1 gene may be detected by measuring the level of the CADM1 protein. Methods of measuring the level of a protein are well known in the art and need not be explained in great detail.

The present invention provides in one aspect a method for determining the bone forming potential of a mesenchymal stem cell. This method provides as a result an indication of the capacity of an osteoblast or a progenitor cell thereof as defined herein to develop an extracellular matrix that mineralizes by calcium phosphate deposition when placed in the body of a vertebrate subject. The method comprises the step of determining the amount of gene product, such as RNA or protein, from the gene CADM1. Hence, a CADM1 expression product is used herein as a marker of the osteogenic potential.

The osteogenic or bone forming potential in particular relates to the ability of the cells to develop bone in vivo. However, this capacity also expresses the potential when used in vivo.

The mesenchymal stem cell, subject of the methods of the present invention may be a mesenchymal stem cell such as present in or obtained from bone marrow or blood. Methods for obtaining mesenchymal stem cells from marrow or blood are well known to the skilled person. Usually, an aspirate is obtained from bone marrow using a syringe.

The expression level of the CADM1 gene may be determined by any method available to the skilled person. In this context, reference is made to the Examples below, wherein both microarray methods and qPCR are indicated as very suitable. qPCR has the advantage that the measurements can be made specific to CADM1 expression products only, and allow accurate quantitation of the expression level.

The present invention relates in another aspect to a method for selecting a bone forming mesenchymal stem cell. This method uses the above-mentioned method for determining the bone forming potential and allows the selection of cells having both an improved bone forming potential or a reduced bone forming potential as compared to a reference cell. The reference cell may suitably be a positive (bone forming) control cell, or, alternatively, may be a negative (non-bone forming) control cell. Also, a cell having normal bone forming capacity can be used as a control. The control or reference cell is preferably a cell of the same type (MSC), and of the same species (human, etc). Whether a cell has an improved bone forming capacity relative to a control cell can be easily determined via performing a bone formation bioassay. Such a bioassay includes the assessment of the capacity of the cell to mineralize. Such assays are well known in the art. Reference is also made to the Examples as described herein below under the section "in vivo bone formation assay".

The bone formation assay may for instance comprise the seeding of a population of mesenchymal stem cells on a scaffold material and the implantation of the thus seeded scaffold in a living vertebrate, for instance subcutaneously in a mouse or rat. Following a suitable incubation period, the implants are examined for bone formation. These data can be used to prepare a calibration curve. Such a calibration curve correlates bone forming potential to marker expression. Hence, based on the calibration curve (which could in principle comprise a single reference point, but preferably comprises data from mesenchymal stem cells of various bone forming potential), the skilled person can relate the expression level of CADMl as obtained for a cell of which the bone forming potential is to be determined, to a certain bone forming potential.

If the cell has the desired potential it may hereafter be selected, for instance by separating it from other cells or by removing the cells with undesirable potential. This method of selecting may be performed on a plurality of cells from a single sample, to the effect that said sample is enriched for a cell type having a particular potential.

The present invention further provides a method of treatment based on the selection of a graft of mesenchymal stem cells with a higher chance of resulting in bone formation in the patient receiving said graft. This is achieved by virtue of the provision of a method for selecting an individual suitable as a donor of bone forming mesenchymal stem cells, to the exclusion of individuals that are less suitable for the intended purpose. The method for selecting an individual as a donor for a graft is based on the determination of the bone forming potential of mesenchymal stem cell in a sample of mesenchymal stem cells from a candidate donor using CADMl as a marker for the bone forming potential of the cells in said sample as described herein above. Following the selection of the candidate donor as a suitable donor individual, the tested sample of mesenchymal stem cells, or an additional sample obtained from said individual, may be used as a graft.

The graft material may be further enriched as described above, or may be treated to make it (more) suitable for use as a graft. With this is meant that growth factors or carriers are added to the cell sample. Also, the cells in the sample can be further expanded ex vivo by cultivation under suitable conditions. Further additional cells may be added. The cells of the graft may further be seeded onto a scaffold. Scaffolds for use in bone tissue engineering are well known in the art and need not be described in great detail here.

The treatment method as proposed herein comprises the administration of the graft as described herein to a patient in need thereof. The graft can be administered via injection, surgery, or any available medical technique for placing a bone graft into a patient.

The present invention further provides a method for screening compounds for their ability to inhibit, stimulate or otherwise alter the bone forming capacity of a mesenchymal stem cell. This method suitably comprising the steps of providing a (sample of) mesenchymal stem cell(s), preferably under ex vivo or in vitro conditions, and bringing a test compound in contact with said cell(s). Upon exposure of the cell to the compound it is determined whether the expression of CADMl in the cell is affected. Hence, the level of expression of CADMl is preferably determined at two different points in time, notably, before and after said exposure. In case the expression of CADMl in the cell before and after exposure is different, it is positively determined that said compound has a increased probability of affecting the bone forming potential of said cell. This can then be confirmed by performing a bone formation bioassay, preferably as described herein.

The present invention will now be illustrated by way of the following non- limiting examples.

EXAMPLES

Description of donor population

To find a predictive marker for bone-forming capacity of hMSCs we isolated cells from bone marrow aspirates (5-20 mL) which were obtained from 62 donors undergoing orthopedic surgery. Aspirates were drawn from either the acetabulum or the iliac crest. Of the patients, 48 were female and 15 male; their age varied from 17- 84 with an average of 56. We confirmed the identity of the cells according to the set of standards proposed by the Mesenchymal and Tissue Stem Cell Committee of the ISCT. Our hMSCs were adherent to plastic (Fig. la) and more than 95% of the cells expressed CD105, CD73 and CD90, and less than 2% expressed CD45, CD34, CDllb, CD 19 and HLA-DR, as measured by flow cytometry in three different donors (Fig. lb). The cells were able to differentiate into the osteogenic, adipogenic and chondrogenic lineage under standard in vitro differentiation conditions, demonstrated by histological staining (Fig. lc). For all donors, the proliferation rate was calculated (proliferation doublings per day from passage 0 to 1) (Fig. 2a). For osteogenic differentiation, we used dex which enhances the expression of ALP in vitro, an early osteogenic marker. As reported by us previously, the basic expression of ALP (without induction of dex) has a large donor variation, even as the dex-induced expression. In our case, ALP expression (as determined by flow cytometry) in the control group ranged from 0.21-38.96% of ALP positive cells and in the dex-induced group from 0.26-46.61%, with an average of 12.49% in 44 different donors (Fig. 2b). Next, we evaluated the in vitro mineralization capacity of hMSCs, by culturing cells in mineralization medium. After three weeks, calcium deposition, which is a late marker for in vitro osteogenesis, was quantified. Again, we found a huge donor variation ranging from 0-20.3 mg/dl calcium with an average of 8.3 (Fig. 2c) in 21 different donors. To determine the adipogenic differentiation capacity of hMSCs, we cultured cells in adipogenic medium for three weeks. Lipid formation was quantified by Oil red O staining, extraction of color and measurement of absorbance. The optical density (OD) was ranging from 1-6.70 with an average of 3.31 in 18 different donors (Fig. 2d). To assess the in vivo bone formation we used the ectopic bone formation model in immune- deficient mice, which is widely used to evaluate the bone-forming capacity of hMSCs. Cells of 62 different donors were seeded onto porous calcium phosphate ceramics and cultured for one week in osteogenic medium, prior to implantation. After six weeks, scaffolds were explanted and bone formation was quantified. Out of 62 donors, 35 did show bone formation ranging from 0.01-4.56% of scaffold area/bone area demonstrating a large donor-variability (Fig. 2e). Correlation between clinical or cell biological data labels and bone formation

To explain the large donor- variability and find a marker for the in vivo bone formation we tried to correlate the bone forming capacity to either the clinical labels such as surgery type, site of aspiration, gender and age or to cell biological labels such as the number of mononuclear cells per ml of bone marrow, the rate of proliferation of the hMSCs or the differentiation parameters such as percentage of ALP positive cells, mineralization or the quantification for adipogenic differentiation. For each data label, we created a correlation curve where we plotted the parameter against in vivo bone formation (depicted for proliferation rate and percentage ALP positive cells in Figs. 3a and 3b). In addition, we visualized the probability in a Receiver Operating Characteristic (ROC) curve which is a widely used standard for describing and comparing the accuracy of diagnostic tests. It represents the trade-off between the false negative and false positive rates for every possible cut-off (again depicted for proliferation rate and percentage ALP positive cells in Fig. 3c). Using the ROC curves, we calculated the AUC (area under curve) for all labels (Fig. 4a), which is a measure of the probability that a classifier based on this label would rank a randomly chosen positive donor higher than a randomly chosen negative donor. AUC=1 is a perfect ranking classifier and AUC=0.5 depicts complete randomness. In addition, for all labels the p-value was calculated, describing the significance of the AUC-score based on permutation tests (10.000 permutations) (Fig. 4b). Significance was not found for any single parameter after multiple testing, neither when we used the presence or absence as a cut off, nor when the threshold was based on percentage of bone.

Aspirate location versus surgery type had a relatively low p-value and the labels were highly correlated, which can be explained since the aspiration site is determined by the surgery type. Next, we tried to build a classifier using the clinical and biological data, which means that multiple labels combined may provide a higher predictive value (see Materials and Methods section). We calculated the AUC for all labels and the best AUC we obtained for imputation based strategies (i.e. classifier on all donors) was around 0.55, which is slightly above random.

A single gene bone classifier based on whole genome gene expression profiling

Since we were not able to find a marker for in vivo bone formation by hMSCs, we performed a microarray to find genes which' expression could be correlated with the in vivo bone formation. Therefore, we isolated RNA from undifferentiated hMSCs during the expansion phase in passage 2, and hybridized it to the Human Genome U133A 2.0 Array (Affymetrix). We determined if we could predict bone forming capacity using the microarray measurements by making a diagnostic bone forming classifier. To this end we trained a nearest-mean classifier, using the SAM test statistic to select features (see Materials and Methods section). Since we want to identify diagnostic markers which can be used to distinguish bone forming from non- bone forming donors, we decided to focus only on genes which display a large difference between different arrays (standard deviation >0.4), which was observed in 1653 out of 22277 genes. The resulting ROC curve was created (Fig. 5b), giving the result for different threshold values of the classifier. This is a much better result (compared to the clinical labels), which is also shown by the area under the ROC curve (AUC) score of 0.7563. This performance can already be reached using only one probeset (Fig. 5a). We found that in all folds of the cross-validation, the top probeset was the same, which detects the CADMl gene. CADMl expression was confirmed by qPCR, an ROC curve was created (Fig. 5b) with a corresponding AUC score of 0.8434, thus resulting in an increased specificity. Furthermore, in the microarray experiment for each gene multiple probes are used. For CADMl, we checked the performance of multiple probes by calculating the AUC score for each probe (Fig. 6a). Most of the probes show similar AUC scores, however, some of them show lower values which might be due to the existence of splice-variants. All p-values are calculated (Fig. 6b) showing near significance between the different probes.

Biological profile of bone-forming hMCS

By analyzing the microarray results, we determined a list of genes correlating with the in vivo bone formation of hMSCs. We confirmed the gene expression of CADMl, WISP1, IGF1, HOXB7 and DKK1 (Fig. 7a) which were all in the top of listed genes (Fig. 7b). Materials and methods

Isolation and culture of hMSCs

Bone marrow aspirates (5-20 mL) were obtained from donors with written informed consent, and hMSCs were isolated and proliferated as described previously (Both et al., 2007, Tissue Eng 13(l):3-9). Briefly, aspirates were resuspended using a 20-gauge needle, plated at a density of 500.000 cells/cm² and cultured in hMSC proliferation medium containing a-MEM (Gibco), 10% foetal bovine serum (Biowhittaker,

Australian), heat-inactivated, 0.2 mM ascorbic acid (Sigma), 2 mM L-glutamin (Gibco), 100 U/ml penicillin with 100 mg/ml streptomycin (Gibco) and 1 ng/ml basic fibroblast growth factor (Instruchemie, Delfzijl, The Netherlands). The serum batch was selected on proliferation rate and osteogenic differentiation potential and used for all experiments in this paper. Cells were grown in a humid atmosphere with 5% CO2. Medium was refreshed twice a week, and cells were used for further subculturing or cryopreservation upon reaching near confluence. hMSC basic/control medium was composed of hMSC proliferation medium without basic fibroblast growth factor, hMSC osteogenic medium was composed of hMSC basic medium supplemented with lO ⁸ M dex (Sigma), and hMSC mineralization medium was composed of basic medium supplemented with 10^~8 M dex and 0.01 M 6— glycerolphosphate (Sigma).

In vivo bone formation

To evaluate the bone forming capacity of hMSCs, cells were seeded onto porous biphasic calcium phosphate (BCP) ceramic granules of approximately 2-3 mm, prepared and sintered at 1150 °C as described previously (Yuan et al., 2002, J Mater Sci Mater Med 13(12):1271-5). In total, 200.000 cells per three particles were seeded, in osteogenic medium. After one week of culturing, tissue-engineered constructs were implanted s.c. in immune -deficient mice (Hsd-cbp:NMRI-nu, Harlan, n=6, in some cases n=5). The mice were anesthetized by intramuscular injection of 0.05 mL of 0.5 mg/mL of anesthetic (1.75 mL of 100 μg/ml ketamine, 1.5 mL of 20 mg/mL xylazine and 0.5 mL of 0.5 mg/mL atropine). Four s.c. pockets were made dorsally and each pocket was implanted with three particles and closed with a vicryl 5-0 suture.

Animals were housed at the Central Laboratory Animal Institute (Utrecht University, Utrecht, The Netherlands), and experiments were approved by the local animal care and use committee.

Histology and histomorphometry

After six weeks, the mice were sacrificed by using CO2 and samples were explanted, fixed in 1.5% glutaraldehyde (Merck) in 0.14 M cacodylic acid (Fluka) buffer (pH 7.3), dehydrated and embedded in methyl methacrylate (LTI) for sectioning. Sections were processed on a histological diamond saw (Leica SP1600). Sections were etched with an HCl/ethanol mixture and sequentially stained to visualize bone, with 1% methylene blue (Sigma) and 0.03% basic fuchsin (Sigma), which stained cells blue and bone pink. Histological sections were analyzed by using a light microscope (E600 Nikon).

Histo morphometry was performed by making low- magnification images from three sections per sample. Scaffold and bone were pseudocolored, and image analysis was performed with KS400 software (Zeiss Vision). A custom-made program (University of Utrecht) was used to measure percentage of cartilage or bone compared to scaffold area.

ALP analysis by flow cytometry

To analyze ALP expression, hMSCs were seeded at 1000 cells/cm² in 6-wells plates, in basic or osteogenic medium in triplicate. After one week, cells were trypsinized and incubated for 30 min in block buffer (PBS with 5% bovine serum albumin (BSA, Sigma) and 0.05% NaNa), incubated with primary antibody (anti-ALP, B4-78, Developmental Studies Hybridoma Bank, University of Iowa), diluted in wash buffer (PBS with 1% BSA and 0.05% NaN₂) for 30 min or with isotype control antibodies. Cells were then washed two times with wash buffer and incubated with secondary antibody (goat-anti-mouse IgG PE, DAKO) for 30 min. Cells were washed and suspended in 500 iL wash buffer with 10 iL Viaprobe (Pharmingen, Uppsala, Sweden) for live/dead cell staining, and only living cells were used for further analysis. ALP expression levels were analyzed on a FACS Calibur (Becton Dickinson). The percentage of ALP positive cells was calculated compared to untreated cells and expressed as relative ALP expression compared to respective controls.

Mineralization

To determine the mineralization capacity and calcium deposition, hMSCs were seeded in T25 flasks at 5000 cells/cm². Cells were cultured in mineralization medium for three weeks, in triplicates. The total calcium deposition was analyzed by using a Calcium Assay Kit (Quantichrom, BioAssay Systems) according to manufacturers protocol. Briefly, medium was aspirated and cells were washed twice with calcium and magnesium free PBS (Gibco) and incubated overnight with 0.5 N HC1 on an orbital shaker. The supernatant was collected for direct absorbance measurement or stored at -20 °C. The calcium content was measured at 620 nm (ELx808, Bio-tek Instruments, Burlington, VT) and expressed as mg/dl calcium. Adipogenesis

Adipogenic differentiation capacity of hMSCs was investigated by seeding 25.000 cells/well in a 24-wells plate, in triplicates, in adipogenic medium composed of DMEM (Gibco), 10% FBS (Biowhittaker) 100 U/ml penicillin +100 μg/ml streptomycin (Gibco), 0.2 mM Indomethacin (Sigma), 0.5 mM 3-Isobutyl- l-methylxanthine (Sigma), lO ⁶ M dex (Sigma) and 10 μg/ml Insulin (human, Sigma). After three weeks, lipid formation was visualized by staining with Oil red O. Briefly, cells were fixed for at least 4h in formol (3.7% formalin with lg/100 mL CaC * 2 H2O), rinsed with water, incubated for 5 min in 60% isopropanol, and stained for 5 min in freshly filtered Oil red O solution (stock: 500 mg of Oil red O (Sigma), 99 mL of isopropanol, 1 mL of water; stain: 42 mL of stock and 28 mL of water), and rinsed with water. After staining, Oil Red O is quantified by extraction with 4% Igepal (Sigma) in isopropanol for 15 minutes by shaking at room temperature and measuring absorbance at 540 nm (ELx808, Bio-tek Instruments, Burlington, VT). Microarray analysis

To analyze the gene expression profile of hMSCs, cells were seeded at 1000 cells/cm² and upon reaching near confluence RNA was isolated using an RNeasy mini kit (Qiagen) and DNase treated on column with 10U RNase free DNase I (Gibco) at 37 °C for 30 minutes. DNase was inactivated at 72 °C for 15 minutes. The quality and quantity of RNA was analyzed by gel electrophoresis and spectrophotometrically. The RNA was hybridized to the Human Genome U133A 2.0 Array (Affymetrix) and scanned with a GeneChip G3000 scanner (Affymetrix). The microarray experiments were performed in three batches. Although this was done at the same microarray facility using arrays from the same production batch, there were still noticeable batch effects. To normalize the measurements, we first applied robust multichip analysis (RMA) background subtraction (Irizarry et al., 2003, Biostatistics 4(2):249-64), followed by quantile-quantile normalization (Bolstad et al., 2003, Bioinformatics 19(2):185-93). Next, we calculated a residual signal for each probe by fitting a probe level model (Irizarry et al., 2003, Nucleic Acids Res 31(4):el5), with the different batches added as an extra effect. We mapped these residual probe signals to the location of the probe on the array and, using a median box filter of size 11, we determined and removed array location specific effects. After this we obtained the final summarized probe set signals by refitting the probe level model. To determine the most significant genes with respect to a label-set, we determined a (two-sided) p- value for each gene using a permutation test. As test statistic we used the significant analysis of microarrays (SAM) test statistic (Tusher et al., 2001, Proc. Natl. Acad. Sci. USA 98(9):5116-5121), for class labels and the F-test for continuous labels. In total, for each label-set, we performed 22277 permutations (equal to the number of genes in the dataset). Genes were sorted on their estimated p-value. For further analysis, we also calculated gene set enrichments using gene sets from the database of molecular signatures (MsigDB, Subramanian et al., 2005, Proc. Natl. Acad. Sci. USA

102(43): 15545- 15550). In addition to this, we trained a classifier for the binary bone label (bone or no bone), predicting if bone formation would occur for a certain donor or not. We applied a Nearest-Mean classifier (available as part of PRTools), and performance was estimated using leave-one-out cross-validation. The probesets to be used as features were selected by taking those with the highest SAM test statistic value on the training set. To determine the optimal number of probesets an inner leave-one-out cross-validation loop was performed. An area under curve (AUC) score of the Receiver Operating Characteristic (ROC) was constructed by combining results for the different validation sets using the classifier class probability (determined using maximum likelihood posterior probabilities).

Claims

1. A method for determining the bone forming potential of a mesenchymal stem cell comprising determining the level of expression of a marker gene, characterised in that said gene is the gene encoding cell adhesion molecule 1 (CADM1).

2. Method according to claim 1, wherein said bone forming potential is the in vivo bone forming capacity.

3. Method according to claim 1 or 2, wherein said mesenchymal stem cell is obtained from bone marrow or blood.

4. Method according to any one of claims 1-3, wherein said level of expression of CADM1 is determined by qPCR.

5. A method for selecting a mesenchymal stem cell based on bone forming potential of said cell, comprising performing the method of any one of claims 1-4 and selecting said mesenchymal stem cell based on a comparison of said level of expression with a control level.

6. A method of enriching a sample of mesenchymal stem cells based on bone forming potential of said cells, comprising performing a method of claim 5 on a plurality of mesenchymal stem cells of said sample, and separating the selected cell from said sample.

7. A method for selecting an individual suitable as a donor of bone forming mesenchymal stem cells, comprising the steps of:

- providing a sample comprising at least one bone forming mesenchymal stem cell from said individual,

- determining the bone forming potential of said mesenchymal stem cell in said sample using the method of any one of claims 1-4, - and selecting said individual as a suitable donor based on a comparison of said level of expression with a control level.

8. A method of providing a graft of bone forming mesenchymal stem cells comprising the steps of:

- selecting an individual suitable as a donor of bone forming mesenchymal stem cells by performing a method according to claim 7,

- obtaining a sample of mesenchymal stem cells from said individual, and

- optionally enriching said sample based on bone forming potential of said cells by performing a method of claim 6 and/or performing further treatments on said sample to make it suitable for use as a graft.

9. A method for treatment of a patient in need of a graft of bone forming mesenchymal stem cells, comprising the steps of providing a graft of bone forming mesenchymal stem cells by performing a method according to claim 8, and administering said graft to said patient.

10. A method for testing whether a compound interferes with the bone forming capacity of a mesenchymal stem cell comprising steps of:

- providing said compound to said mesenchymal stem cell in vitro;

- determining the expression of CADM1 in said mesenchymal stem cell and;

- determining whether a compound interferes with the bone forming capacity based on said expression.

11. Method according to any of the preceding claims, wherein the step of determining said level of expression further comprises a step of performing a bone formation bioassay.

12. Method according to any one of claims 5-9 and 11, wherein said control level is determined by an in vitro assay wherein the level of expression of CADM1 is correlated to the bone forming potential based on a bone formation bioassay.

13. Method according to claim 12, wherein said bioassay comprises a mineralization assay.

14. Method according to any one of the preceding claims, wherein said mesenchymal stem cell is a human cell.

15. Method according to claim 1, wherein said mesenchymal stem cell is proliferated in vitro, differentiated in vitro or otherwise cultured in vitro prior to determining the level of expression of said marker gene.