WO2013024024A2

WO2013024024A2 - Quality assessment of induced pluripotent cells by referring to splicing signatures of pluripotency

Info

Publication number: WO2013024024A2
Application number: PCT/EP2012/065681
Authority: WO
Inventors: Delphine DEBONO-LAGNEAUX; Jean-François DELEUZE; Marie-Christine Multon; Corinne Rocher; Emmanuel SPANAKIS; Gilbert Thill
Original assignee: Sanofi
Priority date: 2011-08-12
Filing date: 2012-08-10
Publication date: 2013-02-21
Also published as: AR087491A1; WO2013024024A3

Abstract

The present invention concerns a method of determining the pluripotency status of a cells comprising looking at the splicing signature in said cells, and correlating said splicing signature with the pluripotency status of the cells.

Description

QUALITY ASSESSMENT OF INDUCED PLURIPOTENT CELLS BY REFERRING TO SPLICING SIGNATURES OF PLURIPOTENCY

BACKGROUND OF THE INVENTION

The present invention relates to a method to check the pluripotency of Induced Pluripotent Cells by looking at some alternative splicing events in suitable genes.

Embryonic stem cells (ES cells) are stem cells established from the inner cell mass of mammalian blastocysts. ES cells can be infinitely grown while maintaining the ability to differentiate into all types of cells excluding trophoblasts (pluripotency). As a result, ES cells are promising donor sources for cell transplantation therapies, which, for examples, comprise treating a patient with myocardial infarction or Parkinson's disease by transplanting myocardial cells or dopaminergic neurones produced in large amounts from ES cells.

However, the use of ESCs is also associated with ethical issues, i.e., the destruction of human embryos, as well as with immunological issues, i.e., rejection reactions after allogenic transplantation of differentiated ESC. It may be possible to overcome these issues by replacing ES cells with Induced Pluripotent Stem cells, commonly abbreviated as iPS cells. iPS cells are a type of pluripotent stem cell artificially obtained from a non-pluripotent cell, typically an adult somatic cell, by exogenous expression of a series of reprogramming transcription factors. The process by which a non-pluripotent cell is transformed into an iPS cell is called reprogramming.

IPS cells are similar to natural pluripotent stem cells, such as ES cells, in many aspects, such as the expression of most of the stem cell genes and proteins, chromatin methylation patterns, doubling time, embryoid body formation, teratoma occurrence, viable chimera formation (for mouse iPSC), and differentiation potential. IPS cells were first generated in 2006 from somatic cells via the transduction of the four transcription factors Oct4, Sox2, Klf4 and c-Myc (Takahashi, K. & Yamanaka, S. Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell 126, 663-676 (2006)). Since then, progress has been made in the reprogramming methods that open the way to the generation of safer iPS with higher efficiency (Trends Mol Med. 2009 Feb, 15(2): 59-68; Cell Death Differ. 2010 Aug, 17(8): 1230-7). Because of this progress and because iPS cells are similar to ES cells without being hampered by the same ethical and immunological concerns, it is now believed that iPS cells, by allowing a patient's own cells to become a source of therapeutic tissue, have the potential to become a platform for personalized medicine. However, several challenges will need to be overcome before this potential is realized, among which is the qualification of the iPS cell lines. Indeed, at the end of the reprogramming of human somatic cells into iPS cells, several colonies that at first sight look like "bona fide" iPS are obtained, but they are not all equivalent as they have not reached the same level of reprogramming and not all of them have acquired the same differentiation potential. Consequently, practitioners of reprogramming will need to characterize the iPS clones that they produce in order to identify the best reprogrammed ones.

Some proposed minimum criteria for evaluation of iPS cells have been proposed (Maherali and Hochedlinger. Guidelines and Techniques for the Generation of Induced Pluripotent Stem Cells Cell Stem Cell. 2008 Dec, 3(6): 595-605).

They consist of :

(i) physical properties, such as cell morphology and self-renewal potential,

Morphology: iPS cells are morphologically similar to embryonic stem cells (ESCs). Each cell has a round shape, large nucleolus and scant cytoplasm. Colonies of iPS cells are also similar to that of ESCs. Human iPS cells form sharp-edged, flat, tightly- packed colonies similar to human ESCs (hESCs) whereas mouse iPS cells form the colonies similar to murine ESCs (mESCs), less flatter and more aggregated colonies than that of hESCs.

Growth properties: Doubling time and mitotic activity are cornerstones of ESCs, as stem cells must self-renew as part of their definition. iPS cells are mitotically active, actively self-renewing, proliferating, and dividing at a rate equal to ESCs.

(ii) genetic and epigenetic properties, such as the expression of key endogenous pluripotency markers, independence from exogenous reprogramming factor and a concomitant down regulation of lineage-specific genes associated with the cell of origin. Pluripotent markers: iPS cells express cell surface antigenic markers expressed on ESCs. Human iPSCs express the markers specific to hESC, including SSEA-3, SSEA-4, TRA-1-60, TRA-1-81 , TRA-2-49/6E, and Nanog. Mouse iPS cells express SSEA-1 but not SSEA-3 nor SSEA-4, similarly to mESCs.

Pluripotent genes: iPS cells express genes expressed in undifferentiated ESCs, including, e.g., Oct-3/4, Sox2, Nanog, GDF3, REX1 , FGF4, ESG1 , DPPA2, DPPA4, and hTERT.

Telomerase activity: Telomerase is necessary to sustain cell division unrestricted by the Hayflick limit of ~ 50 cell divisions. hESCs express high telomerase activity to sustain self-renewal and proliferation, and iPS cells also demonstrate high telomerase activity and express hTERT (human telomerase reverse transcriptase), a necessary component in the telomerase protein complex.

And (iii) functional properties, such as 1) in vitro differentiation, 2) teratoma formation, 3) chimera contribution, 4) germline transmission and 5) tetraploid complementation (direct generation of entirely iPSC-derived mice) (Jaenisch and Young. Stem cells, the molecular circuitry of pluripotency and nuclear reprogramming. 2008, Cell 132, 567- 582).

Pluripotency: iPS cells are capable of differentiation in a similar way to ESCs into fully differentiated tissues. For example, iPS cells injected into immunodeficient mice spontaneously form teratomas after nine weeks. Teratomas are tumors of multiple lineages containing tissue derived from the three germ layers endoderm, mesoderm and ectoderm; this is unlike other tumors, which typically are of only one cell type origin. Teratoma formation is a landmark test for pluripotency. Further, hESCs in culture spontaneously form spheroid embryo-like structures termed "embryoid bodies", which consist of a core of mitotically active and differentiating hESCs and a periphery of fully differentiated cells from all three germ layers. iPS cells also form embryoid bodies that have peripheral differentiated cells.

Blastocyst Injection: hESCs naturally reside within the inner cell mass (embryoblast) of blastocysts, and in the embryoblast, differentiate into the embryo while the blastocyst's shell (trophoblast) differentiates into extraembryonic tissues. The hollow trophoblast is unable to form a living embryo, and thus it is necessary for the embryonic stem cells within the embryoblast to differentiate and form the embryo.

Mouse iPS cells can be injected by micropipette into an irradiated blastocyst, and the "blastocyst" is transferred to recipient females. Chimeric living mouse pups can thus be created, i.e. mice with iPS cell derivatives incorporated all across their bodies with a varying degree of chimerism.

However, because the aforementioned battery of tests is labour intensive and costly, it cannot be applied to all iPS colonies obtained at the end of the reprogramming process. Moreover, experiments for assessing chimera contribution, germline transmission and tetraploid complementation are not applicable to human iPSC. Thus, a pragmatic, i.e., simple, fast and cheap test is needed in order to control the quality of the candidate iPS colonies before more in depth characterization is performed, especially if we envision to build haplotyped iPSC libraries. It is precisely an object of the present invention to offer a pragmatic screening method for assessing the pluripotent status of the candidate iPS colonies. The method consists in determining by RT-PCR the splicing profiles of a handful of cassette exons within the candidate iPS colonies and comparing these splicing profiles with a splicing signature of pluripotency that our team has identified. Only the colonies displaying this splicing signature of pluripotency will be considered for further in depth

characterization.

The splicing signatures of pluripotency identified are also an object of the invention.

The use of splicing signatures in order to discriminate pluripotent cells has never been considered. It involves also the selection of relevant splicing events and the achievement of a discriminating calculation method.

DETAILED DESCRIPTION OF THE INVENTION

Alternative splicing (AS)

AS is a process by which the exons of the RNA produced by transcription of a gene (a primary gene transcript or pre-mRNA) are reconnected in multiple ways during RNA splicing. The resulting different mRNAs may be translated into different protein isoform; thus, a single gene may code for multiple proteins [Black, Douglas L. (2003). "Mechanisms of alternative pre-messenger RNA splicing". Annual Reviews of

Biochemistry 72 (1): 291-336.] Alternative splicing occurs as a normal phenomenon in eukaryotes, where it greatly increases the diversity of proteins that can be encoded by the genome; in humans, -95% of multiexonic genes are alternatively spliced [Pan, Q; Shai O, Lee LJ, Frey BJ, Blencowe BJ (Dec 2008). "Deep surveying of alternative splicing complexity in the human transcriptome by high-throughput sequencing". Nature Genetics 40 (12):

1413-1415]. There are numerous modes of alternative splicing observed, of which the most common is cassette exon (also called exon skipping). In this mode, a particular exon may be included in mRNAs under some conditions or in particular tissues, and omitted from the mRNA in others. Types of Alternative Splicing Events (AS events):

Five basic types of AS EVENTS are generally recognized. [Black, Douglas L. (2003). "Mechanisms of alternative pre-messenger RNA splicing". Annual Reviews of

Biochemistry 72 (1): 291-336]; [Matlin, AJ; Clark F, Smith, CWJ (May 2005).

"Understanding alternative splicing: towards a cellular code". Nature Reviews 6 (5): 386-398]; [Pan, Q; Shai O, Lee LJ, Frey BJ, Blencowe BJ (Dec 2008). "Deep surveying of alternative splicing complexity in the human transcriptome by high- throughput sequencing". Nature Genetics 40 (12): 1413-1415]; [Michael Sammeth; Sylvain Foissac; Roderic Guigo (2008). Brent, Michael R.. ed. "A general definition and nomenclature for alternative splicing events. PLoS Comput Biol. 4 (8): e1000147]. - Cassette exon (or exon skipping): in this case, an exon may be spliced out of the primary transcript or retained. This is the most common mode in mammalian pre- mRNAs.

- Mutually exclusive exons: One of two exons is retained in mRNAs after splicing, but not both. - Alternative donor site: An alternative 5' splice junction (donor site) is used, changing the 3' boundary of the upstream exon.

- Alternative acceptor site: An alternative 3' splice junction (acceptor site) is used, changing the 5' boundary of the downstream exon.

- Intron retention: A sequence may be spliced out as an intron or simply retained. This is distinguished from exon skipping because the retained sequence is not flanked by introns. If the retained intron is in the coding region, the intron must encode amino acids in frame with the neighboring exons, or a stop codon or a shift in the reading frame will cause the protein to be non-functional. This is the rarest mode in mammals.

Splicing profile of an AS event: Whatever its type, a given AS event gives rise to two different messenger RNA molecules (or splicing isoforms) that differ in size, one being longer than the other because it contains an additional internal sequence corresponding to the alternatively spliced region. In the case of AS events of the cassette exon type, the long isoform contains the alternatively spliced exon and the small one does not. In this application, the splicing profile of an AS event in a sample corresponds to the ratio Ψ= U(L+S), where L is the concentration of the long isoform and S is the concentration of the short isoform in the sample. Ψ equals 1 when only the long isoform is detected and 0 when only the short isoform is detected. For statistical analysis purposes, Ψ is given the value 0.5 when neither isoform is detected in the sample (Ψ is missing). 32 AS events whose splicing profiles were shared between all pluripotent cells examined have been identified as being really associated to the pluripotent state. Since all of these 32 AS events are of the "cassette exon" type and since they are associated to the pluripotent state, they are referred to as PACEs, which is the acronym for Pluripotent Associated Cassette Exon. Overall, 32 PACEs occurring within the 32 genes below were identified:

ACOT9, ADD3, ATP1 1C, BAIAP2, CARM1 , CASK, CD47, CLSTN1 , CLSTN1 , CSNK1 D, DENND1 B, ENAH, ESYT2, EXOC1 , FN 1 , FOXM1 , GIT2, GPR126, TGA6, KIF13A, LRRFIP2, MBD1 , MINK1 , NFYA, NUMA1 , NUMB, PLOD2, SCARB1 , SSBP3, SULF2, TCF12, TMEM63B. These 32 PACEs are annotated in Table 1 , which gives for each PACE: the name of the gene to which it belongs, its numbering within the reference transcript and the identity of the reference transcript, information about how its splicing profile (ψ) varies in ES cells compared with melanocytes, its location within the reference transcript, its size, - its sequence,

Also, each of these PACEs has been renamed from S1 (Splice 1) to S32 (Splice 32).

S1 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO: 1 in ACOT9 gene, said exon having a down (lower) ψ in ES cells compared with melanocytes. S2 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO: 2 in ADD3, said exon having an up (higher) ψ in ES cells compared with melanocytes.

S3 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO: 3 in ATP11 C gene, said exon having an up (higher) ψ in ES cells compared with melanocytes. S4 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO: 4 in BAIAP2 gene , said exon having an up (higher) ψ in ES cells compared with melanocytes .

55 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO: 5 in CARM 1 gene, said exon having an up (higher) ψ in ES cells compared with melanocytes.

56 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO: 6 in CASK gene, said exon having an up (higher) ψ in ES cells compared with melanocytes.

57 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO: 7 in CD47 gene, said exon having an up (higher) ψ in ES cells compared with melanocytes. 58 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO: 8 in CLSTN1 gene, said exon having a down (lower) ψ in ES cells compared with melanocytes

59 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO: 9 in CLSTN1 gene, said exon having a down (lower) ψ in ES cells compared with melanocytes.

S10 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO:

10 in CNSK1 D gene, said exon having a down (lower) ψ in ES cells compared with melanocytes. S11 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO:

1 1 in DENND1 B gene, said exon having an up (higher) ψ in ES cells compared with melanocytes.

512 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO:

12 in ENAH gene, said exon having an up (higher) ψ in ES cells compared with melanocytes.

513 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO:

13 in ESYT2 gene, said exon having an up (higher) ψ in ES cells compared with melanocytes.

514 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO: 14 in EXOC1 gene, said exon having an up (higher) ψ in ES cells compared with melanocytes.

515 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO:

15 in FN 1 gene, said exon having an up (higher) ψ in ES cells compared with melanocytes. S16 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO:

16 in FOXM 1 gene, said exon having a down (lower) ψ in ES cells compared with melanocytes. 517 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO: 17 in GIT1 gene, said exon having a down (lower) ψ in ES cells compared with melanocytes.

518 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO: 18 in GPR126 gene, said exon having an up (higher) ψ in ES cells compared with melanocytes.

519 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO:

19 in ITGA6 gene, said exon having a down (lower) ψ in ES cells compared with melanocytes. S20 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO:

20 in KIF13A gene, said exon having a down (lower) ψ in ES cells compared with melanocytes.

521 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO:

21 in LRRFIP2 gene, said exon having a down (lower) ψ in ES cells compared with melanocytes.

522 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO:

22 in MBD1 gene, said exon having an up (higher) ψ in ES cells compared with melanocytes.

523 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO: 23 in MINK1 gene, said exon having an up (higher) ψ in ES cells compared with melanocytes.

524 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO:

24 in NFYA gene, said exon having a down (lower) ψ in ES cells compared with melanocytes. S25 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO:

25 in NUMA1 gene, said exon having a down (lower) ψ in ES cells compared with melanocytes. 526 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO: 26 in NUMB gene, said exon having an up (higher) ψ in ES cells compared with melanocytes.

527 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO: 27 in PLOD2 gene, said exon having a down (lower) ψ in ES cells compared with melanocytes.

528 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO:

28 in SCARB1 gene, said exon having a down (lower) ψ in ES cells compared with melanocytes. S29 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO:

29 in SSBP3 gene, said exon having an up (higher) ψ in ES cells compared with melanocytes.

530 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO:

30 in SULF2 gene, said exon having a down (lower) ψ in ES cells compared with melanocytes.

531 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO:

31 in TCF12 gene, said exon having an up (higher) ψ in ES cells compared with melanocytes.

532 refers to an AS event involving the splicing of an exon of sequence SEQ ID NO: 32 in TMEM63B gene, said exon having a down (lower) ψ in ES cells compared with melanocytes.

The above AS events (S1 to S32) are sometimes referred to as "AS events according to the invention" throughout the present specification.

Splicing signature of pluripotency In this application, the term "splicing signature" refers to a set of splicing profiles (Ψ values). The term "splicing signature of pluripotency" refers to a set of splicing profiles (Ψ values) expressed in pluripotent cells that collectively distinguish said pluripotent cells from non pluripotent cells.

Determination by RT-PCR of the splicing profile of an AS event within a given sample

A method to determine the splicing profile of an AS event within a given sample is fully described in Venables et Al., Nat Struct Mol Biol, 2009; 16(6), 670-6.

Briefly, total RNA (or messenger RNA) from said sample is first extracted using any method well known by those skilled in the art. A reverse transcription primed with a mixture of an oligo-dT and random hexamers is then performed on the isolated RNA in order to convert it into cDNA. An end-point PCR reaction is then done on the cDNA using a pair of PCR primers that anneal the flanking regions on both sides of the AS event. By doing so, the two splicing isoforms generated by the AS event are co- amplified in the same reaction. The two amplification products, also called amplimers, are then analysed by capillary electrophoresis on an Applied Biosystem instrument (3730X DNA Analyser) which, through a dedicated software, determines the size and offers a quantification of the amplimers that allows to determine the Ψ value and hence the splicing profile of the AS event.

Use of the discriminant functions: For each cell population to be classified, the splicing profile, i.e., the Ψ value, of each predictor gene of a given splicing signature is determined by RT-PCR or any other methods adapted for this purpose. Each of these Ψ values is then multiplied with the corresponding unstandardized coefficient and the products so generated are sum up with the constant coefficient of the function. The result (discriminant score) is then compared with the centroid coordinates of the pluripotent and differentiated groups. The test specimen is thus classified as pluripotent if its score is closer to the centroid of the pluripotent group than to the centroid of the differentiated group; otherwise the specimen is classified as non pluripotent.

Some combinations of alternative splicing events have been identified as being specific of a pluripotent status. IPS samples (see table 4):

All the iPS cell lines used in this work have been derived from human fibroblasts, except Mel1_iPS and Mel2_iPS (derived from human melanocytes), B_iPS_CB32 and B_iPS_CB3 (derived from human cord blood), and H9 IPS (derived from the H9 MSC cell line, which is an MSC cell line obtained in vitro by differentiation from the H9 human ES cell line)

Also, all the iPS cell lines used in this work have been generated by infecting the parental cells with a combination of recombinant retroviruses coding for a cocktail of transcription factors. All the iPS cell lines were made with a cocktail comprising Oct4, Sox2, Klf4 and c-Myc, except B_iPS_CB32 (made with Oct4, Sox2 and Klf4),

B_iPS_CB3 (made with Oct4 and Sox2) and H9 IPS (made with Oct4, Sox2, Nanog and Lin28).

An embodiment of the invention is a splicing signature comprising or consisting of at least one splicing profile of an alternative splicing event (AS event) selected from the group consisting of S15, S9, S18, S19, S23, S26, S30 and S32. More specifically, this splicing signature is a splicing signature of pluripotency.

In one embodiment, this splicing signature comprises or consists of at least one splicing profile of an alternative splicing event selected from the group consisting of: S15, S23, S26, S30 and S32. In one embodiment, this splicing signature comprises or consists of at least one splicing profile of an alternative splicing event selected from the group consisting of: S15, S30 and S32.

In one embodiment, this splicing signature comprises or consists of the splicing profiles of the alternative splicing events S15, S30 and S32. In one embodiment, this splicing signature comprises or consists of the splicing profiles of the alternative splicing events S15, S23, S26, S30 and S32.

In one embodiment, this splicing signature comprises or consists of the splicing profiles of the alternative splicing events S9, S15, S18, S19, S23, S26, S30 and S32. Another embodiment of the invention is a splicing signature comprising or consisting of:

- a splicing profile of an alternative splicing event selected from the group consisting of S9, S15, S18, S19, S23, S26, S30 and S32, and - at least one splicing profile of an alternative splicing event comprised in the group consisting of S1 to S32.

In a specific embodiment, the splicing signature comprises or consists of the splicing profiles of alternative splicing events S1 to S32.

The above splicing signatures are sometimes referred to as "splicing signatures according to the invention" throughout the present specification.

Another embodiment of the invention is the use of a splicing signature according to the invention for determining the pluripotency status of cells, by in vitro experiments.

Another embodiment of the invention is the use of an alternative splicing event according to the invention or of a combination of alternative splicing events as disclosed hereabove as a biomarker for determining the pluripotency status of cells.

As used herein, the term "biomarker" has its usual meaning in the art, i.e., it refers to a biological marker that is an indicator of a biological state. In the present case, the biological marker is the AS event or a combination thereof, and the biological state the pluripotency status of a cell. The invention concerns also a method of determining the pluripotency status of cells comprising looking at the splicing signature in said cells, and correlating said splicing signature with the pluripotency status of cells.

Another aspect of the invention is a method of determining the pluripotency status of a cell, comprising: a) determining/calculating the splicing signature of said cell, b) comparing said splicing signature with a splicing signature of pluripotent cells, and c) determining that said cells are pluripotent if said splicing signature is similar to, or is not statistically different from, said splicing signature of pluripotent cells.

In one embodiment, a splicing signature is "similar to or not statistically different from" another splicing signature if each of the splicing profiles of the first signature have a value that is no more than 25%, 20%, 15%, 10%, 5% or 2% different from the corresponding splicing profile of the second signature.

The invention also pertains to a method of determining the pluripotency status of cells or of groups of cells to be tested comprising the steps of: i) For each cell or group of cells, determining the splicing profile (i.e., the Ψ value) of each gene of the splicing signature (e.g. by RT-PCR, or by any other methods adapted for this purpose); ii) Calculating the discriminant score (e.g. by multiplying each of the Ψ

values with the unstandardized coefficient, and sum up the generated values with the constant coefficient of the function); and iii) Comparing the discriminant score with the centroid coordinates of one or more control pluripotent group(s) and of one or more differentiated control group(s); wherein the cells or groups of cells to be tested are classified as pluripotent if their score is closer to the centroid of the control pluripotent group(s) than to the centroid of the control differentiated group(s), and wherein they are classified as non-pluripotent otherwise.

The invention further relates to a method of determining the pluripotency status of cells comprising the detection of an alternative splicing event or of a combination of alternative splicing events in said cells (e.g. by RT-PCR). This method can further comprise a step of determining/calculating the splicing signature. And also another step maybe present: correlating the splicing signature and the pluripotency status of the cells, e.g. by comparing it to the splicing signature of pluripotent cells and/or by calculating the discriminant score (as explained in steps (ii) and (iii) above). The splicing signatures, the splicing profiles and the splicing events can notably be determined by RT-PCR. Methods for calculating the value of splicing signatures and of determinant scores are explained in details in the examples. The above methods of determining the pluripotency status of cells may be carried out in vitro or ex vivo. Another object of the invention is a diagnostic kit comprising means for the

determination of a splicing signature in cells, wherein these means can be primers useful for amplification of exons during a RT-PCR. The determination of the splicing signature is carried out in vitro.

The invention concerns also a composition comprising a reagent capable of detecting an alternative splicing event or of a combination of alternative splicing events in cells.

Legend of the figures:

Figure 1 : Hierarchical clustering of the pluripotent and the differentiated samples based on the splicing profiles of 32 candidate exons. Dendrogram of alternative splicing profiles (Ψ) using Word linkage. The samples are indicated on the vertical axis using their case number along with their group designation as in Tables 2 and 3. Note the definition of clusters E (for ectoderm origin), M (mesoderm), EM

(endoderm/mesoderm) and P (pluripotent stem cells). The horizontal axis represents a distance (dissimilarity) scale.

Figure 2: Box-plot representation of the distributions of Ψ (vertical axis) of five genes (horizontal axis) in differentiated (blank) and pluripotent (patterned) cells. The circles represent values beyond the inner fences, and the stars, values beyond the outer fences of the distributions. Note that none of the genes could be reliably used as a marker of pluripotency because the distributions of Ψ always overlap between the two groups. Figure 3: Box-plot representation of the distributions of the discriminant score from the 5-gene function (vertical axis) in differentiated (1 ; light grey) and pluripotent (2; dark grey) cells. Note that the discriminant score computed from the Ψ values of the 5 genes could be reliably used to classify pluripotent cells (low values) against differentiated cells (high values) since the distributions of this score in the two groups do not overlap. Figure 4: Scatter plot of discriminant scores (vertical axis) of the samples belonging to the four clusters as defined from the dendrogram of Figure 1. The discriminant scores were calculated from the optimal, 5-gene function with P for entrance<=0.01 and P for removal >0.05. The horizontal axis separates the four clusters: 1 =Mesoderm cluster, 2=Endoderm/mesoderm cluster, 3=Pluripotent cell cluster and 4=Ectoderm cluster. Circles represent differentiated cells of mesoderm, endoderm/mesoderm or ectoderm origin and squares represent pluripotent cells. The dark-grey dashed line represents the centroid of the differentiated-cell group, and the light-grey one, the centroid of the pluripotent-cell group. Figure 5: Scatter plot of discriminant scores (vertical axis) of the samples belonging to the four clusters as defined from the dendrogram of Figure 1. The discriminant scores were calculated from the robust, 8-gene function with P for entrance<=0.05 and P for removal >0.10. The horizontal axis separates the four clusters: 1 =Mesoderm cluster, 2=Endoderm/mesoderm cluster, 3=Pluripotent cell cluster and 4=Ectoderm cluster. Circles represent differentiated cells of mesoderm, endoderm/mesoderm or ectoderm origin and squares represent pluripotent cells. The dark-grey dashed line represents the centroid of the differentiated-cell group, and the light-grey one, the centroid of the pluripotent-cell group.

Figure 6: Scatter plot of discriminant scores (vertical axis) of the samples belonging to the four clusters as defined from the dendrogram of Figure 1. The discriminant scores were calculated from the economical, 3-gene function with P for entrance<=0.001 and P for removal >0.01. The horizontal axis separates the four clusters: 1=Mesoderm cluster, 2=Endoderm/mesoderm cluster, 3=Pluripotent cell cluster and 4=Ectoderm cluster. Circles represent differentiated cells of mesoderm, endoderm/mesoderm or ectoderm origin and squares represent pluripotent cells. The dark-grey dashed line represents the centroid of the differentiated-cell group, and the light-grey one, the centroid of the pluripotent-cell group.

The object of the invention is then showed but not limited by the following examples. Examples:

Example 1 : Description of the candidate approach:

Because cancer cells and pluripotent cells share a lot of characteristics (Nature. 2009 Aug, 460: 1085-6), a set of AS events previously associated with cancer were firstly evaluated in the IPS / melanocyte model (Mel1_iPS/Melano_P3; see table 4). More precisely, a list of 57 cancer associated AS events selected from the literature [Venables et al, Nat Struct Mol Biol. 2009; 16(6): 670-6] was drawn and the profiles of these 57 AS events within the two cell lines examined (Mel1_iPS and Melano_P3) was determined. 17 out of the above 57 AS events were found to be highly differentially expressed between the melanocytes (Melano_P3) and the iPS cell line (Mel1_iPS).

Example 2: Description of the genome-wide approach:

With approximately four probes per exon and roughly 40 probes per gene, the GeneChip® Human Exon 1.0 ST Array (developed by Affymetrix, Inc., Santa Clara, CA, USA) enables an exon-level analysis on a whole-genome scale, allowing the detection of splicing differences between various samples.

Using the aforementioned exon array platform, a list of 323 putative regulated exons between the iPS cell line (Mel1_iPS) and the melanocytes (Melano_P3) was drawn up. Then the splicing profiles of 109 of these 323 exons within the two situations examined were determined by RT-PCR and 20 of them were found as really differentially expressed.

Taken together, these two approaches described above identified 37 AS events differentially expressed between the melanocytes (Melano_P3) and the iPS cell line (Mel1_iPS).

In order to check whether these 37 AS events were really associated to pluripotency, it was examined how consistent their splicing profiles were across a panel of 13 pluripotent cell lines (1 ES cell line and 12 iPS cell lines, see table 4). Good reproducibility was retrieved for 32 splicing events on the 37. Only the 32 AS events whose splicing profiles were shared between all pluripotent cells examined were considered as really associated to the pluripotent state. Since all of these 32 AS events are of the "cassette exon" type and since they are associated to the pluripotent state, they are referred to as PACEs, which is the acronym for Pluripotent Associated Cassette Exon.

Overall, 32 PACEs occurring within the 32 genes below were identified:

ACOT9, ADD3, ATP1 1C, BAIAP2, CARM1 , CASK, CD47, CLSTN1 , CLSTN1 , CSNK1 D, DENND1 B, ENAH, ESYT2, EXOC1 , FN1 , FOXM1 , GIT2, GPR126, TGA6, KIF13A, LRRFIP2, MBD1 , MINK1 , NFYA, NUMA1 , NUMB, PLOD2, SCARB1 , SSBP3, SULF2, TCF12, TMEM63B.

These 32 PACEs are annotated in Table 1 , which gives for each PACE: the name of the gene to which it belongs, its numbering within the reference transcript and the identity of the reference transcript, - information about how its splicing profile (ψ) varies in ES cells compared with melanocytes, its location within the reference transcript, its size, its sequence, Also, each of these PACEs has been renamed from S1 (Splice 1) to S32 (Splice 32).

In order to determine whether some of these 32 PACEs could be specific of the pluripotent state and hence be considered as true Markers of Pluripotency (MoP), their splicing profiles were examined across a commercial (Clontech Laboratories, Inc. Mountain View, CA, USA) panel of 37 normal and 7 cancerous human tissues (see table 4). From these profiling experiments, it was concluded that none of the 32 PACEs tested is individually eligible as MoP because for each of these, the splicing profile associated to pluripotency was also observed in at least one sample of non- pluripotent type. It was assumed that in the absence of true individual MoPs, it could still be possible to derive from the set of 32 PACEs a splicing signature of pluripotency, i.e., a subset of AS events whose combined profiles might distinguish pluripotent cells from all other types of cells. Indeed, such splicing signatures of pluripotency have been identified by using the method explained below.

Example 3: Statistical identification of molecular signatures

For each gene and sample analysed, the variable Ψ was computed as the ratio U(L+S), where L and S are the intensities of the electrophoretic bands representing the Long and Short splicing isoforms, respectively. Other transformations, such as the LogRatio of the signals (Log₁₀(LJS)), a non-parametric ratio variable with values 1 , if L>S, or -1 , if L<S as well as the z-scores of LogRatio or that of Ψ, were also analysed (results not shown). However, only Ψ proved useful for robust classification. Missing Ψ points (i.e. when none of the splicing variants was detected for a gene in a sample) were given the value 0.5. All statistical analyses were performed using the IBM SPSS version 19 software (IBM Corporation, Somers, NY).

Identification of natural molecular classes of cell types as defined by splicing variation.

Hierarchical clustering of the samples was performed with the Ψ values of 32 candidate exons of Table 1 using the Ward's linkage method of clustering and the squared Euclidean distance as interval measure. This analysis identified groups of cell types with similar splicing patterns. For the interpretation of the clusters we grouped the samples by their presumed ontogenetic origin (when known), by tissue or by phenotype. The dendrogram resulting from this analysis is shown in Figure 1. In this dendrogram the samples are indicated by their serial number as well as with their group designation as described in Tables 2 and 3. There were two major clusters corresponding to the differentiation state of the cells: a cluster of non-pluripotent cells (cluster 1 ; Differentiated) and a cluster of pluripotent cells, i.e. ESC-IPSC (cluster 2; Pluripotent). Further, the cluster of differentiated cells was divided into 3 sub-clusters. In the 4-cluster classification solution, cluster 3 contained only pluripotent cells. The other clusters were more or less homogeneous and contained samples of mesoderm origin only (cluster 1), of endoderm/mesoderm origin (cluster 2), or ectoderm origin (cluster 4; Table 3). This application is not concerned with those differentiated sub- clusters (clusters 1 , 2 and 4) although the four- group clustering is used in the figures showing the distributions of discriminant scores (the calculation of discriminant scores is described below).

Table 1 : Annotation of the 32 PACEs

Table 2: Symbols indicating the ontogenetic origin of the samples of the four clusters¹

¹ The symbols used represent the ontogenetic origin of the majority of the samples within each cluster

Table 3: : Cluster membership of the samples according to Ward hierarchical cluster analysis applied to Ψ values for 32 exons (A priori sample classification and hierarchical clustering)

Differencluster Origin/ cluster

Sample Presumed

Number Sample name tiation membership ^b phenotype membership⁰ origin⁸

value label value label

1 adipose tissue Mesoderm 1 Differentiated 1 Mesoderm

2 adrenal gland Endoderm 1 Differentiated 2 Endoderm/ mesoderm

3 adrenal gland P2 Endoderm 1 Differentiated 2 Endoderm/ mesoderm

4 aorta Mesoderm 1 Differentiated 1 Mesoderm

5 B iPS CB3 IPS 2 Pluripotent 3 Pluripotent

6 B iPS CB32 IPS 2 Pluripotent 3 Pluripotent

7 B iPS F1 IPS 2 Pluripotent 3 Pluripotent

8 bladder Mesoderm 1 Differentiated 1 Mesoderm

9 blood Mesoderm 1 Differentiated 2 Endoderm/ mesoderm

10 bone marrow Mesoderm 1 Differentiated 2 Endoderm/ mesoderm

11 bone marrow P2 Mesoderm 1 Differentiated 2 Endoderm/ mesoderm

12 brain cerebellum Ectoderm 1 Differentiated 4 Ectoderm P2

13 brain fetal Ectoderm 1 Differentiated 4 Ectoderm

14 brain whole P2 Ectoderm 1 Differentiated 4 Ectoderm

15 breast tumor Mesoderm 1 Differentiated 2 Endoderm/ mesoderm colon Endoderm 1 Differentiated 2 Endoderm/ mesoderm colon tumor Endoderm 1 Differentiated 2 Endoderm/ tumor mesoderm dorsal root Dorsal root 1 Differentiated 1 Mesoderm ganglion ganglion

EB1 1 CI8 EB1 1 1 Differentiated 4 Ectoderm

H Fibro cl5 Fibroblast 1 Differentiated 1 Mesoderm

H iPS cl5 IPS 2 Pluripotent 3 Pluripotent

H9 H9 ESC 2 Pluripotent 3 Pluripotent

H9 IPS H9 IPS 2 Pluripotent 3 Pluripotent

H9 MSC H9 MSC 1 Differentiated 1 Mesoderm heart Mesoderm 1 Differentiated 1 Mesoderm heart infarctus Mesoderm 1 Differentiated 1 Mesoderm heart normal Mesoderm 1 Differentiated 1 Mesoderm heart P2 Mesoderm 1 Differentiated 1 Mesoderm iPS43 A2 IPS 2 Pluripotent 3 Pluripotent iPS43 B2 IPS 2 Pluripotent 3 Pluripotent iPS43 D6 IPS 2 Pluripotent 3 Pluripotent iPS57 A5 IPS 2 Pluripotent 3 Pluripotent iPS57 A7 IPS 2 Pluripotent 3 Pluripotent iPS57 B7 IPS 2 Pluripotent 3 Pluripotent kidney Mesoderm 1 Differentiated 2 Endoderm/ mesoderm kidney P2 Mesoderm 1 Differentiated 2 Endoderm/ mesoderm kidney tumor Mesoderm 1 Differentiated 2 Endoderm/ tumor mesoderm liver Endoderm 1 Differentiated 2 Endoderm/ mesoderm liver fetal Endoderm 1 Differentiated 2 Endoderm/ mesoderm liver P2 Endoderm 1 Differentiated 2 Endoderm/ mesoderm lung Endoderm 1 Differentiated 2 Endoderm/ mesoderm lung tumor Endoderm 1 Differentiated 2 Endoderm/ tumor mesoderm lung whole P2 Endoderm 1 Differentiated 4 Ectoderm mammary gland Mesoderm 1 Differentiated 2 Endoderm/ mesoderm

Mel1 iPS IPS 2 Pluripotent 3 Pluripotent

Mel2 iPS IPS 2 Pluripotent 3 Pluripotent

Melano_P3 Melanocyte Differentiated 2 Endoderm/

¹

s mesoderm mTilys_P49 Neuronal Differentiated 4 Ectoderm

¹

procursor

mTilys_P53 Neuronal Differentiated 4 Ectoderm

¹

procursor

Neuro31 Neuronal Differentiated 4 Ectoderm

¹

procursor

ovary Mesoderm 1 Differentiated 1 Mesoderm ovary tumor Mesoderm Differentiated 2 Endoderm/

¹

tumor mesoderm pancreas Endoderm Differentiated 2 Endoderm/

¹

mesoderm placenta Placenta Differentiated 2 Endoderm/

¹

mesoderm placenta P2 Placenta Differentiated 2 Endoderm/

¹

mesoderm prostate Mesoderm Differentiated 2 Endoderm/

¹

mesoderm prostate P2 Mesoderm Differentiated 2 Endoderm/

¹

mesoderm retina Retina 1 Differentiated 4 Ectoderm salivary gland Salivary Differentiated 2 Endoderm/

¹

gland mesoderm salivary gland P2 Salivary Differentiated 2 Endoderm/

¹

gland mesoderm skeletal muscle Mesoderm 1 Differentiated 1 Mesoderm skeletal muscle Mesoderm Differentiated 1 Mesoderm

¹

P2

small intestine Endoderm Differentiated 2 Endoderm/

¹

mesoderm smooth muscle Mesoderm 1 Differentiated 1 Mesoderm spinal cord Spinal cord 1 Differentiated 4 Ectoderm spinal cord P2 Spinal cord 1 Differentiated 4 Ectoderm spleen Mesoderm Differentiated 2 Endoderm/

¹

mesoderm 8 stomach Endoderm 1 Differentiated 2 Endoderm/ mesoderm9 stomach tumor Endoderm 1 Differentiated 2 Endoderm/ tumor mesoderm0 testis Testis 1 Differentiated 4 Ectoderm1 testis P2 Testis 1 Differentiated 4 Ectoderm2 thymus Endoderm 1 Differentiated 2 Endoderm/ mesoderm3 thymus P2 Endoderm 1 Differentiated 2 Endoderm/ mesoderm4 thyroid gland Endoderm 1 Differentiated 2 Endoderm/ mesoderm5 thyroid gland P2 Endoderm 1 Differentiated 2 Endoderm/ mesoderm6 tonsil Tonsil 1 Differentiated 2 Endoderm/ mesoderm7 trachea Endoderm 1 Differentiated 2 Endoderm/ mesoderm8 trachea P2 Endoderm 1 Differentiated 2 Endoderm/ mesoderm9 uterus Mesoderm 1 Differentiated 1 Mesoderm0 uterus P2 Mesoderm 1 Differentiated 1 Mesoderm1 uterus tumor Mesoderm 1 Differentiated 1 Mesoderm ^a Samples are pooled into an ectoderm, endoderm or mesoderm group when their ontogenetic origin is known. Otherwise, samples are pooled into tissue or cell-type groups or not pooled at all.

b Samples are classified into two clusters according to their patterns of alternative splicing of 29 genes. The clusters are labeled according to the phenotype of the majority of samples in them.

0 Samples are classified into four clusters according to their patterns of alternative splicing of 29 genes. The clusters are labeled according to the ontogenetic origin of the majority of samples in them.

Discriminant functions

A stepwise discriminant function analysis was employed to identify sets of splicing events that confidently predict the cluster membership (Differentiated or Pluripotent) for each sample. A discriminant function is a linear combination of the Ψ values of multiple exons that results in a discriminant score (D) for each sample according to the linear model:

Equation 1: D = β₀ + βιΨι + β₂Ψ₂ +■■■ + β_ηΨ_η

The aim of the algorithm of discriminant function analysis is to select enough Ψ variables (from the candidate exons of Table 1 ) and to compute the β coefficient for each Ψ such that the distributions of the discriminant scores for each sample group (in this case, Pluripotent vs. Differentiated cell types) do not overlap (Figures 2 and 3). According to this classical algorithm, discriminant variables are selected for entry into the equation on the basis of how much they lower the overall Wilk's lambda when they enter the model. The variable that minimises this statistic enters the model at each step. All variables within the model are re-evaluated and any variable with nonsignificant corrected effect is removed. At a first instance, the probability criterion for inclusion (P_in) in the model was set to a P_/n<=0.01 and the probability for removal (P_out) was set to P_OUf>0.05. Selection continues until all variables in the model are significant and all the non-included variables are non-significant. Prior probabilities of 2-group membership were set equal (P=0.5). The covariance matrix within groups was used for the calculation of the discriminant function. The leave-one-out cross-validation method was employed in order to estimate the performance of the discriminant function in classifying unknown samples. Only models with 100% cross-validated correct classification were retained. The samples used for discriminant analyses are those listed in Table 4 and their discriminant scores for the three alternative discriminant functions computed here are shown in Table 5.

Table 4: The cell lines and commercial samples used and their presumed ontogenetic origin: Four-cluster membership and discriminant scores of the samples

Sample Presumed origin³ Cluster Four-cluster

membership¹³

1 adipose tissue Mesoderm 1 Mesoderm

2 adrenal gland Endoderm 2 Endoderm

3 adrenal gland P2 Endoderm 2 Endoderm

4 aorta Mesoderm 1 Mesoderm

5 B_iPS_CB3 IPS 3 Pluripotent B_iPS_CB32 IPS 3 Pluripotent

B_iPS_F1 IPS 3 Pluripotent bladder Mesoderm 1 Mesoderm blood Mesoderm 2 Endoderm bone marrow Mesoderm 2 Endoderm bone marrow P2 Mesoderm 2 Endoderm brain cerebellum P2 Ectoderm 4 Ectoderm brain fetal Ectoderm 4 Ectoderm brain whole P2 Ectoderm 4 Ectoderm breast tumor Mesoderm 2 Endoderm colon Endoderm 2 Endoderm colon tumor Endoderm tumor 2 Endoderm dorsal root ganglion Dorsal root ganglion 1 Mesoderm

EB1 1 CI8 EB1 1 4 Ectoderm

H_Fibro_cl5 Fibroblast 1 Mesoderm

H_iPS_cl5 IPS 3 Pluripotent

H9 H9 ESC 3 Pluripotent

H9 IPS H9 IPS 3 Pluripotent

H9 MSC H9 MSC 1 Mesoderm heart Mesoderm 1 Mesoderm heart infarctus Mesoderm 1 Mesoderm heart normal Mesoderm 1 Mesoderm heart P2 Mesoderm 1 Mesoderm iPS43_A2 IPS 3 Pluripotent iPS43_B2 IPS 3 Pluripotent iPS43_D6 IPS 3 Pluripotent iPS57_A5 IPS 3 Pluripotent iPS57_A7 IPS 3 Pluripotent iPS57_B7 IPS 3 Pluripotent kidney Mesoderm 2 Endoderm kidney P2 Mesoderm 2 Endoderm kidney tumor Mesoderm tumor 2 Endoderm liver Endoderm 2 Endoderm liver fetal Endoderm 2 Endoderm liver P2 Endoderm 2 Endoderm lung Endoderm 2 Endoderm lung tumor Endoderm tumor 2 Endoderm lung whole P2 Endoderm 4 Ectoderm mammary gland Mesoderm 2 Endoderm

Mel1_iPS IPS 3 Pluripotent

Mel2_iPS IPS 3 Pluripotent

Melano_P3 Melanocytes 2 Endoderm mTilys_P49 Neuronal procursor 4 Ectoderm mTilys_P53 Neuronal procursor 4 Ectoderm

Neuro31 Neuronal procursor 4 Ectoderm ovary Mesoderm 1 Mesoderm ovary tumor Mesoderm tumor 2 Endoderm pancreas Endoderm 2 Endoderm placenta Placenta 2 Endoderm placenta P2 Placenta 2 Endoderm prostate Mesoderm 2 Endoderm prostate P2 Mesoderm 2 Endoderm retina Retina 4 Ectoderm salivary gland Salivary gland 2 Endoderm salivary gland P2 Salivary gland 2 Endoderm skeletal muscle Mesoderm 1 Mesoderm skeletal muscle P2 Mesoderm 1 Mesoderm small intestine Endoderm 2 Endoderm smooth muscle Mesoderm 1 Mesoderm spinal cord Spinal cord 4 Ectoderm spinal cord P2 Spinal cord 4 Ectoderm spleen Mesoderm 2 Endoderm 68 stomach Endoderm 2 Endoderm

69 stomach tumor Endoderm tumor 2 Endoderm

70 testis Testis 4 Ectoderm

71 testis P2 Testis 4 Ectoderm

72 thymus Endoderm 2 Endoderm

73 thymus P2 Endoderm 2 Endoderm

74 thyroid gland Endoderm 2 Endoderm

75 thyroid gland P2 Endoderm 2 Endoderm

76 tonsil Tonsil 2 Endoderm

77 trachea Endoderm 2 Endoderm

78 trachea P2 Endoderm 2 Endoderm

79 uterus Mesoderm 1 Mesoderm

80 uterus P2 Mesoderm 1 Mesoderm

81 uterus tumor Mesoderm 1 Mesoderm

Samples are pooled into an ectoderm, endoderm or mesoderm group when their ontogenetic origin is known. Otherwise, samples are pooled into tissue or cell-type groups or not pooled at all.

Samples are classified into four clusters according to their pattern of alternative splicing of 32 candidate genes. The clusters are labeled according to the ontogenetic origin of the majority of sample members.

Table 5: Discriminant scores of the samples for the 3 discriminant functions presented (Discriminant scores of the samples for the functions with 5, 8 or 3 exons).

B_iPS_CB32 -5.0 -5.9 -3.5

B_iPS_F1 -4.6 -5.6 -3.1 bladder 2.2 1.7 1.9 blood -0.3 0.7 0.4 bone marrow 1.8 3.0 2.6 bone marrow P2 0.8 0.9 2.3 brain cerebellum P2 0.6 1.1 0.3 brain fetal 0.1 -0.5 -0.3 brain whole P2 0.7 1.0 0.4 breast tumor -0.6 0.5 -1.0 colon 0.4 0.5 -0.1 colon tumor -0.4 -0.3 -0.8 dorsal root ganglion 1.0 2.5 0.7

EB11CI8 -0.3 0.3 -0.5

H_Fibro_cl5 2.9 3.0 2.5

H_iPS_cl5 -5.5 -5.5 -5.7

H9 -6.9 -8.0 -6.1

H9 IPS -6.3 -7.1 -5.6

H9 MSC 1.0 0.7 1.0 heart 0.7 1.2 0.6 heart infarctus 0.7 0.6 0.6 heart normal 0.6 0.5 0.4 heart P2 0.9 1.3 0.8 iPS43_A2 -5.4 -6.4 -4.8 iPS43_B2 -7.2 -8.5 -5.9 iPS43_D6 -6.5 -7.3 -5.7 iPS57_A5 -4.2 -5.3 -4.1 iPS57_A7 -7.0 -7.7 -6.5 iPS57_B7 -6.1 -6.8 -5.7 kidney 1.6 2.6 1.0 kidney P2 1.1 1.0 0.8 kidney tumor 1.5 1.5 1.5 liver 1.2 1.2 1.1 liver fetal 1.3 0.8 2.6 liver P2 0.9 0.8 1.0 lung 2.4 2.4 2.3 lung tumor -0.3 -0.4 -0.3 lung whole P2 0.7 1.1 0.9 mammary gland 0.6 1.3 0.5

Mel1_iPS -6.6 -7.2 -6.2

Mel2_iPS -6.7 -7.3 -5.9

Melano_P3 1.9 1.9 2.6 mTilys_P49 -1.2 -0.8 -0.8 mTilys_P53 -0.3 0.0 0.0

Neuro31 1.2 1.9 0.9 ovary 2.5 2.1 1.9 ovary tumor 2.6 3.5 2.3 pancreas 2.3 2.2 2.6 placenta 2.0 2.0 2.2 placenta P2 0.9 1.0 0.7 prostate 2.5 2.7 1.8 prostate P2 1.6 1.9 1.0 retina 1.0 0.3 0.7 salivary gland 1.9 1.6 1.4 salivary gland P2 1.1 0.8 0.8 skeletal muscle 0.7 0.2 0.6 skeletal muscle P2 1.0 0.8 0.8 small intestine 1.9 1.8 1.4 smooth muscle 2.0 1.5 1.8 spinal cord 2.0 3.1 1.6 spinal cord P2 1.8 1.5 1.3 spleen 1.8 1.7 2.3 stomach 1.3 0.8 0.6 stomach tumor 1.4 1.3 0.6 testis 2.3 2.4 0.8 testis P2 2.1 2.1 0.6 thymus 0.5 1.2 0.2 thymus P2 0.4 0.9 0.6 thyroid gland 1.2 1.6 0.9 thyroid gland P2 0.8 1.5 0.7 tonsil -0.8 -0.6 -0.3 trachea 1.0 1.1 0.5 78 trachea P2 1 .3 1 .3 0.9

79 uterus 2.6 3.0 2.3

80 uterus P2 2.7 1 .6 2.3

81 uterus tumor 0.0 1 .5 0.3 a'^b'° Scores of the samples for three discriminant functions with 5, 8 or 3 AS events, respectively, as shown in Figure 4, 5 and 6. P_in is the probability value criterion for a splicing event entering a model discriminant function. P_0Uf is the probability value criterion for a splicing event being removed from a model discriminant function. Of course, when different probability criteria are used for inclusion and/or removal of variables or when only a subset of the candidate exons is considered for entry, a different discriminant function may be computed. From all the possible discriminant functions resulting from the entire set of the 32 candidates we are particularly interested in 3 functions: (/) an 'optimal', 5-exon function which results with the P_m and P_out settings as mentioned above (Figure 4); (/^'/) a 'robust', 8-exon function computed with P_/n<=0.05 and P_OUf>0.1 which is relatively more difficult to evaluate experimentally but confers maximal confidence of classification (Figure 5); (Hi) an 'economical', 3- exon function computed with P_/n<=0.001 and P_OUf>0.01 which may confer somewhat limited confidence of classification in the long term (with many new samples) but provides for an extremely simple experimental design leaving only few doubtful cases to be tested for additional exons (Figure 6). The unstandardized coefficients (β) and the group centroids (means of discriminant scores for each group of samples) for the three mentioned functions are shown in Table 6 through to Table 11 . Table 12 shows the cross-validation results which are the same for all the three functions since a 100% cross-validated correct classification was our criterion for retaining a

discriminant function.

Table 6: Unstandardized coefficients of the discriminant function including 5 AS events: Coefficients³ of Discriminant Function 1 (Optimal; 5 exons)

a Unstandardized coefficients

Table 7: Group centroids for the discriminant function including 5 AS events: Function 1 (Optimal; 5 exons) at Group Centroids³

a Unstandardized canonical discriminant functions evaluated at group means

Table 8: Unstandardized coefficients of the discriminant function including 8 AS events: Coefficients³ of Discriminant Function 2 (Robust; 8 exons)

³ Unstandardized coefficients Table 9: The centroid coordinates from the discriminant function including 8 AS events: Function 2 (Robust; 8 exons) at Group Centroids³

a Unstandardized canonical discriminant functions evaluated at group means

Table 10: Unstandardized coefficients of the discriminant function including 3 AS events: Coefficients³ of Discriminant Function 3 (Economical; 3 exons)

a Unstandardized coefficients

Table 11 : The centroid coordinates from the discriminant function including 3 AS events: Function 3 (Economical; 3 exons) at Group Centroids³

a Unstandardized canonical discriminant functions evaluated at group means

As mentioned above, cluster analysis of the AS events suggested the existence of 2 major clusters, one of which consisted exclusively of pluripotent cells (ES and iPSC cell types) and the other of non-pluripotent differentiated cells.

Three major signatures of pluripotency have been identified as being very discriminant: they are consisting in respectively 5, 8 and 3 PACEs. Example 4:

An object of the invention is the splicing signature comprising the PACEs S15, S23, S26, S30 and S32. The discriminant function of those 5 PACEs distinguished perfectly the two clusters. The unstandardized coefficients of that discriminant function are shown in Table 6, the group centroids in Table 7 and the classification results in

Table 12.

Table 12: Classification and cross-validation results for either of the three discriminant functions presented: Results of classification using any of the three discriminant functions^{a b c}

a Cross validation is done only for those cases in the analysis. In cross validation, each case is classified by the functions derived from all cases other than that case. ^b 100.0 % of original grouped cases correctly classified.

c 100.0 % of cross-validated grouped cases correctly classified. Figure 2 shows box-plot representations of the distributions of Ψ for each of the 5 genes of the signature while Figure 3 shows the distribution of the discriminant score as calculated using the coefficients of the discriminant function from Table 6. The box- plots represent the median, the upper and lower quartiles (hinges) and the upper and lower adjacent values of each distribution as well as the outliers (values beyond the inner or outer fences of the distribution). In Figure 2, the distributions of Ψ for all of the five genes overlap between the Pluripotent and the Differentiated cell groups. Some outliers from one group may be considered as belonging to the other group. In Figure 3, the distributions of the corresponding discriminant scores are well apart. This representation confirms that none of the AS events can, alone, be used for discriminating between pluripotent and differentiated cells; instead, the discriminant score identifies the two groups very efficiently.

Example 5:

Another object of the invention is the signature constituted of the 8 following PACEs: S9, S15, S18, S19, S23, S26, S30 and S32. Using always the Wilk's lambda minimization algorithm, we relaxed the criteria for entry and removal of variables. For example, when the P-value for entry was set to 0.05 and for removal to 0.10, a longer discriminant signature was produced including 8 AS events. The results of this analysis are presented in Tables 8 and 9 and in Figure 4. The information gathered from such a signature would be rather excessive since the estimated rate of correct classification cannot be improved (being already 100% with only 5 AS events).

However, the distance between the distributions of the discriminant scores for the two groups is increased, making class prediction more comfortable (Figure 5). A longer signature may become useful in practice with respect to some future specimens which would be mapped too close to the discriminant border between the two categories when using the 5-gene function.

Another way to produce different discriminant functions is to remove one, or more, important discriminant gene(s) from the data set and run the algorithm again. The removed genes are replaced by one, or more, 'second choice' markers each. The so produced alternative discriminant functions may also achieve 100% cross-validated correct classification with more or less confident discriminant score mapping. In fact, there are innumerable possible discriminant functions one can compute from a set of 32 potential predictors using various algorithms, various criteria for predictor entry and/or removal and various data subsets; it is impossible to describe them all here. Example 6:

A third object of the invention is the splicing signature comprising the PACEs S15, S30 and S32. This alternative function presents particular interest because it consists of only 3 AS events (Tables 10 and 11 and Figure 6). This function was obtained either by removing MINK1 from the data set or by setting the entry and removal criteria to P_/n<=0.001 and P_ouf>0.01 , respectively. Note that, when MINK1 is intentionally removed from the data, then NUMB is not selected. This 3-predictor combination is considered as a minimal function still resulting in 100% cross-validated correct classification; because, when we removed any one of those three predictors from the data, none of the remaining 29 AS events could be used in its place without dropping the classification efficiency below 100%. However, although most concise, this minimal function may have limited applicability to the extent that the distance between the centroids of the Pluripotent and the Differentiated cell groups is shorter compared to that computed from the other two functions described above. This short signature may be cost-effective for preliminary screenings or identification

experiments.

Example 7: Validity of the splicing signatures of plunpotency

The validity of the three splicing signatures of pluripotency specified in this application was tested on additional cell samples which were not used for the calculation of the discriminant functions. These samples were two human embryonic cell lines (clones H1 and H7) and two different sources of human cord blood CD34+ progenitors (see table 13).

To do this, for each splicing signature tested, the corresponding discriminant scores of the four samples were calculated (see table 14,) and then compared with the centroid coordinates of the pluripotent and differentiated groups (see table 15).

From the above analysis, the applicant was able to verify that the three signatures tested classified correctly the four new samples. Indeed, the two human embryonic cell lines H 1 and H7 were rightly classified as "pluripotent" and the two CD34+ samples were also rightly classified as non "pluripotent".

To sum-up, each of the three splicing signatures of pluripotency have passed the above test.

Table 13: The new samples used for validation of splicing signatures of pluripotency

Sample Product Supplier

ESC_H1 Human Embryonic Stem Cell; clone H1 WiCell Research Institute

ESC_H7 Human Embryonic Stem Cell; clone H1 WiCell Research Institute

HSC1 Human Cord Blood CD34 + progenitors Praxcell

HSC2 Human Cord Blood CD34 + progenitors Lonza Table 14: The discriminant scores of the four samples for the three signatures of pluri potency

DS: Discriminant Score

SS8: Splicing Signature comprising the 8 following AS events: CLSTN1 e1 1 , FN1 , GPR126, ITGA6, MINK1 , NUMB, SULF2, TMEM63B

SS5: Splicing Signature comprising the 5 following AS events: FN1. MINK1. NUMB, SULF2. TMEM63B

SS3: Splicing Signature comprising the 3 following AS events: FN1 , SULF2,

TMEM63B

Table 15: The centroid coordinates of the pluripotent and differenciated groups

In order to illustrate how precisely the discriminant scores are calculated, the calculation of the discriminant score of sample HSC1 associated with the SS5 (SS5 refers to the splicing signature comprising the following 5 PACEs: S15, S23, S26, S30 and S32) splicing signature is presented below:

Let DSss5(X) be the discriminant score of a given sample X associated with the SS5 splicing signature.

The general equation for calculating DS_sss(X) is as follow:

(1) DS_ss5(X) = -6.6 - 3.8Ψ_ΡΝ1 - 3.4Ψ_ΜΙΝΚΙ + 3.1Ψ_ΝΥΜΒ + 6.6Ψ_5Υ_Ρ2 + 3.5Ψ_ΤΜΕΜ63Β with -6.6 corresponding to the constant coefficient of the discriminant function and ΨΡΝΙ , ΨΜΙΝΚΙ , ΨΝυΜΒ, Ψευι^ and Ψ_ΤΜΕΜ63Β corresponding to ίϊι^βΨ values in sample X of the 5 predictor genes of the SS5 splicing signature.

The Ψ values for the HSC1 sample have been determined and are as follow:

Ψ_ΡΝ1= 0

ΨΜΙΝΚΙ= 0.54 Ψ_ΝυΜΒ = 0.05

¾ULF2 = 0.89

ΨΤΜΕΜ63Β =0.97

By introducing the βϋονβΨ values in equation (1), we have,

DS_ss5(HSC1) = -6.6 -(3,8x0) -(3.4x0.54) + (3.1x0.05) + (6.6x0.89) + (3.5x0.97) which eventually gives the following result

DS_ss5(HSC1) = 1.03.

Claims

1. Splicing signature comprising at least one splicing profile of an alternative splicing event selected from the group consisting of S15, S9, S18, S19, S23, S26, S30 and S32.

2. Splicing signature according to claim 1 , wherein said splicing signature comprises at least one splicing profile of an alternative splicing event selected from the group consisting of: S15, S23, S26, S30 and S32.

3. Splicing signature according to claim 1 or 2, wherein said splicing signature comprises at least one splicing profile of alternative splicing event selected from the group consisting of: S15, S30 and S32.

4. Splicing signature according to any one of claims 1 to 3, wherein said splicing signature comprises or consists of the splicing profiles of the alternative splicing events S15, S30 and S32.

5. Splicing signature according to any one of claims 1 to 3, wherein said splicing signature comprises or consists of the splicing profiles of the alternative splicing events S15, S23, S26, S30 and S32.

6. Splicing signature according to any one of claims 1 to 3, said splicing signature comprises or consists of the splicing profiles of the alternative splicing events S9, S15, S18, S19, S23, S26, S30 and S32.

7. Splicing signature according to claim 1 , comprising:

a splicing profile of an alternative splicing event selected from the group consisting of S9, S15, S18, S19, S23, S26, S30 and S32, and

at least one splicing profile of an alternative splicing event selected from the group consisting of S1 to S32.

8. Splicing signature according to any one of the preceding claims, wherein said splicing signature is a splicing signature of pluripotency.

9. Use of the splicing signature as defined in any one of the preceding claims for determining the pluripotency status of cells.

10. Use of an alternative splicing event or of a combination of alternative splicing events as defined in any one of claims 1 to 8 as a biomarker for determining the pluripotency status of cells.

11. A method of determining the pluripotency status of cells or of several groups of cells to be tested comprising the steps of: i) for each cell or group of cells, determining the splicing profile of each gene of the splicing signature; ii) calculating the discriminant score; and iii) comparing the discriminant score with the centroid coordinates of a

pluripotent control group and of a differentiated control group; wherein the cells or groups of cells to be tested are classified as pluripotent if their discriminant score is closer to the centroid of the pluripotent control group than to the centroid of the differentiated control group, and wherein they are classified as non- pluripotent otherwise.

12. Method of determining the pluripotency status of cells comprising:

a) determining the splicing signature as defined in one of claims 1 to 8 in said cells, and

b) correlating said splicing signature with the pluripotency status of the cells.

13. Method of determining the pluripotency status of cells according to claim 12, wherein step b) comprises comparing the splicing signature with a splicing signature of pluripotent cells, wherein the determination that said splicing signature is similar to, or not statistically different from, the splicing signature of pluripotent cells indicates that said cells are pluripotent.

14. Method of determining the pluripotency status of cells comprising the detection of an alternative splicing event or of a combination of alternative splicing events as defined in any one of claims 1 to 8 in said cells.

15. The method of determining the pluripotency status of cells according to claim 14, further comprising the step of determining the splicing signature as defined in one of claims 1 to 8.

16. The method of determining the pluripotency status of cells according to claim 15, further comprising the step of correlating the splicing signature with the pluripotency status of the cells.

17. The method of determining the pluripotency status of cells according to claim 16, wherein said step of correlating the splicing signature with the pluripotency status of the cells comprises either of:

i) comparing the splicing signature with a splicing signature of pluripotent cells, wherein the determination that said splicing signature is similar to, or not statistically different from, the splicing signature of pluripotent cells indicates that said cells are pluripotent; or

ii) calculating the discriminant score and comparing said discriminant score with the centroid coordinates of a pluripotent control group and of a differentiated control group, wherein the cells or groups of cells to be tested are classified as pluripotent if their discriminant score is closer to the centroid of the pluripotent control group than to the centroid of the differentiated control group, and wherein they are classified as non- pi uri potent otherwise

18. The method of determining the pluripotency status of cells according to any one of claims 11 to 17, wherein the splicing signature as defined in any one of claims 1 to 8 is determined by RT-PCR.

19. A diagnostic kit comprising means for the determination of a splicing signature as defined in any one of claims 1 to 8 in cells.

20. A diagnostic kit as claimed in claim 19, wherein the means for the determination of a splicing signature are primers useful for amplification of the exons.

21. Composition comprising a reagent capable of detecting an alternative splicing event or of a combination of alternative splicing events as defined in any one of claims 1 to 8 in cells.