This invention relates to the generation of megakaryocyte (MK) lineages from pluripotent stem cells (PSCs).

Megakaryocytes (MK) are blood cells which are uniquely responsible for formation of platelets regulating haemostasis and thrombosis (Kaushansky . Blood. 2008; 111:981-986) . Low blood platelet count, or thrombocytopenia, can originate from different conditions

(including viral and bacterial infections, hereditary syndromes, cancers, medication induced) and may result in severe haemorrhages in extreme cases . In the UK, the National Health Service is

delivering about 240,000 platelet concentrate units per year from voluntary donors for therapeutic and prophylactic purposes. The total platelet cost for NHS is evaluated at £55m per year (£230 per unit in 2009/10) . In addition, platelet transfusion refractoriness due to alloimmune reactions is a serious issue in particular for patients with regular transfusion needs, which imposes additional biological characterization, i.e. HLA geriotyping, of both donors and recipients. Eventually, as for other human biological products, donor platelet transfusion goes with a risk of transmission of contagious age ts.

Human embryonic stem cells (hESC) can be maintained and expanded in culture indefinitely and are able to generate virtually all the different cell types of the organis ( Keller G Genes Dev.

2005;19:1129-1155) . Indeed, they hold great, promise for cell based therapies. Interestingly, it. has been shown that the banking of a small number of ca efully selected hESC lines to include the most common HLA haplotypes in the UK may be sufficient to offer practical benefits, i.e.. matched or single mismatched tissues for a majority of the population (Taylor CJ et al Lancet. 2005;366:2019-2025) .

Moreover, the ability to derive pluripotent stem cells equivalent to hESCs - the so-called induced pluripotent stem cells (iPSCs) - from adult, somatic cells like skin fibroblasts or circulating blood cells, opens new avenues for clinical applications with the

opportunity of generating fully compatible tissues for virtually every single individual (Yam.ana.ka S. Cell Stem. Cell. 2010; 7:1-2; Takahashi K et al Cell. 2007;131:861-872).

Human ESCs, iPSCs and other human stem cells (collectively human pluripotent stem cells, or hPSCs ) may thus offer a valuable option for e -vivo production of biocompatible platelets. Up to now, generation of clinically relevant cell types from hPSCs has been essentially achieved by mimicking in vivo embryonic development based on scientific knowledge gathered from model vertebrate organisms (Murry CE, Keller G. Cell. 2008; 132:661-680). Existing protocols for generation of platelet precursors, the megakaryocytes, use poorly defined culture medium containing fetal calf serum and bone marrow derived murine stromal cell lines (OP9, C3H10T1/2 ) to support megakaryocyte differentiation and platelet production (Gaur M, et al J Thromb Haemost. (2006) 4:436-442; Takayama N, et al . Blood (2008) 111:5298-5306; Takayama N, et al. J Exp Med. (2010)

207:2817-2830) . More recently, protocols using serum free culture conditions to differentiate megakaryocytes from hPSC were described but platelet production was still dependent, on a co-culture step with the OP9 cell line (Lu SJ, et al. Cell Res. 2011; Pick et al PLoS One Feb 2013 10.1371/ ournal , pone .005 530 ) . These methods involve long-term culture of up to 26 days with complicated cell handling - as single haemangioblast colony picking in semi-solid medium - to produce mature megakaryocytes, yet showing low platelet release capacity. Noteworthy, as previously reported in the context of differentiation of other cell lineages, high variability has been found regarding megakaryocyte generation among different hPSC lines .

This invention relates to the development of a process for the efficient forward programming of human pluripotent cells into megakaryocyte progenitor cells (MK-FoP) . This may be useful, for example, in production of mature megakaryocytes and platelets; the modelling of thrombocytopenia and other platelet-associated

conditions; and the development of therapeutics to these conditions . An aspect of the invention provides a method of forward programming pluripotent cells into megakaryocyte progenitor cells; or producing megakaryocyte progenitor cells; the method comprising;

i) providing a population of isolated pluripotent stem cells (PSCs) ,

ii) introducing a combination of transcription factors (TFs) into the population of PSCs, said combination comprising GATA1, FLU and TALI, and;

iii) culturing said population of cells. The combination of transcription factors introduced into the cell population imposes a megakaryocyte progenitor phenotype i.e. one or more cells in the population are forward programmed by the transcription factor combination into megakaryocyte progenitor cells.

A method of producing mammalian cells with a megakaryocyte

progenitor phenotype as described herein may comprise;

i) providing a population of isolated pluripotent ste cells (PSCs) ,

ii) introducing a combination of transcription factors into the population of PSCs, said combination comprising GATAl, FLU and TALI, and;

iii) culturing said population of cells, such that one or more cells in the population displays a megakaryocyte progenitor

phenotype . The population may be cultured under suitable conditio s and for a sufficient period of time, following introduction of the

transcription factors, to allow one or more cells in the population to display a megakaryocyte progenitor phenotype, for example the stable expression of CD61, CD34 and CD41a; and/or the stable expression of CD61, CD235a and CD41a in said cells.

After programming, the megakaryocyte progenitor cells may be maintained in culture, expanded, stored, for example frozen using conventional techniques, or used in therapeutic or other

applications as described herein.

Transcription factors are DNA binding proteins which regulate the expression of genes in cells. Preferably, the transcription factors introduced into the PSCs are human transcription factors,

The combination of transcription factors for programming PSCs to become megakaryocyte progenitors as described herein comprises GATA1, FLU and TALI. The amino acid sequences of GATA1, FLU and TALI are readily available on public databases. For example, the reference amino acid sequence of human GATA1 (GA A binding protein 1; also known as ERYF1 : Gene ID 2623} has the NCBI database entry NP 002040,1 GI : 4503925; the reference amino acid sequence of human FLU (Friend leukaemia virus integration 1, also known as EWSR1 , SIC-1 or ERGB; Gene ID 2313) has the NCBI database entry NP 002008.2 GI : 7110593 and the reference amino acid sequence of human TALI (T cell acute lymphocytic leukemia protein 1; Gene No: 6886) has the NCBI database entry NP_003180.1 GI: 4507363) . In some embodiments, the combination of transcription factors may lack TALI. For example, a method of producing mammalian cells with a megakaryocyte progenitor phenotype as described herein may comprise; i) providing a population of isolated pluripotent stem cells iPSCs),

ii) introducing a combination of transcription factors into the population of PSCs, said combination comprising GATAl and FLU, and;

iii) culturing said population of cells, such that one or more cells in the population displays a megakaryocyte progenitor

phenotype .

GATAl, FLU and TALI may be produced using routine recombinant techniques or may be obtained from commercial suppliers (e.g. R&D Systems, Minneapolis, MM, USA) .

In some embodiments, the combination of transcription factors may consist of GATAl, FLI1 and TALI i.e. the only transcription factors in introduced into the PSCs are GATAl, FLU and TALI.

In other embodiments, the combination of transcription factors may consist of GATAl, FLU and TALI, with optionally, one, two, three or more, additional transcription factors. For example, additional transcription factors may include one or more of the transcription factors shown in Table 1 or one or more of IKZF1, HOXA5, RU X1, ZFPM2, ZFPM1 and GATA2. in some preferred embodiments, additional transcription factors may include one or more of ABLIM1 , FHL1, RU X3, NFIC, NFIL3 , VDR, MESP1, BTBD11, APPL2, MICALl , BATF, SCMH1 and MBP, as shown in table 2.

The amino acid sequences of IKZFl, HOXA5 , RUNX1, ZFPM2, ZFPM1 and GATA2 and ABLIM1, FHL1, RUNX3 , NFIC, NFIL3 , VDR, MESP1, BTBD11, APPL2, MICALl, BATF, SCMH1 and MBP are readily available on public databases.

Suitable transcription factor nucleic acids and proteins may be produced using routine recombinant techniques or obtained from commercial suppliers (e.g. R&D Systems, Minneapolis, MM, USA;

Cellgenix, DE; Life Technologies, USA) .

Suitable transcription factors for use as described herein may comprise the reference database amino sequence or a variant, thereof. A suitable variant may have at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 95%, or at least 98% sequence identity to the reference sequence. Amino acid sequence identity is generally defined with reference to the algorithm GAP ( GCG Wisconsin Package ^'114 , Accelrys, San Diego CA) . GAP uses the Meedleman & Wunsch algorithm (J. Mol. Biol. (48) : 444- 453 (1970)) to align two complete sequences that maximizes the number of matches and minimizes the number of gaps. Generally, the default parameters are used, with a gap creation penalty = 12 and gap extension penalty = 4. Use of GAP may be preferred but other algorithms may be used, e.g. BLAST or TBLASTN (which use the method of Altschul et al . (1990) J. Mol. Biol. 215: 405-410), FASTA (which uses the method of Pearson and Lipman (1988) PNAS USA 85: 2444- 2448), or the Smith-Waterman algorithm (Smith and Waterman (1981) J. Mol Biol. 147: 195-197), generally employing default parameters.

Particular sequence variants may differ from a reference sequence by insertion, addition, substitution or deletion of 1 amino acid, 2, 3, 4, 5-10, 10-20 or 20-30 amino acids.

PSCs are unspecialized, undifferentiated cells that, are capable of replicating or self-renewing themselves and developing into

specialized cells of all three primary germ layers i.e. ectoderm, mesoderm and endoderm. but are not able to develop into all embryonic and extra-embryonic tissues, including trophectoderm (i.e. not totipotent). Preferably, the PSCs are not committed to a

haematopoietic lineage. In preferred embodiments, the PSCs are human pluripotent stem cells.

PSCs include embryonic stem (ES) cells and non-embryonic stem cells, including foetal and adult somatic stem cells and stem cells derived from non-pluripotent cells, for example induced pluripotent (iPS) cells which are derived from non-pluripotent cells . iPS cells are described in more detail below.

PSCs may express one or more of the following pluripotency

associated markers: Oct.4, Sox2, Alkaline Phosphatase, SSEA--3, Nanog, SSEA-4 and Tra-1-60. Preferably, pluripotent stem cells express Oct4. Human PSCs do not express haematopoietic cell or megakaryocyte markers, such as CD61, CD34, CD41, CD42a, CD42b and GPVI . For example, the pluripotent stem cells may have the phenotype CD61-, CD34-, CD41a, CD42a-, CD42b-, GPVI- .

Markers expressed by a cell, including pluripotency associated markers and haematopoietic cell markers, may be identified using standard techniques, such as flow cytometry, PGR, western blotting, immunocytochemistry and in situ hybridisation.

In some embodiments, the PSCs are ES cells, for example human ES cells and non-human ES cells. Suitable ES cells may be obtained from a cultured hES cell line, such as Edi2, H9 or hSF-6. Further examples of suitable human embryonic ste cells are described in (Thomson JA et al Science 282: 1145-1147 (1998) Reubinoff et al. Nat Biotechnol 18:399-404 (20Q0); Cowan, C.A. et al. N. Engl. J. Med. 35Q, 1353-1356(2004), Gage, F.H., et al . Ann. Rev. Neurosci. 18 159-192 (1995); and Gotlieb (2002) Annu. Rev. Neurosci 25 381-407); Carpenter et al. Stem Cells. 5(1) : 79-88 (2003); see also: the NIH stem cell registry which is accessible online. Potentially clinical grade hESCs are described in Klimanskaya, I. et al . Lancet. 365, 1636-1641 (2005); and Ludwig,T.E. et al . Nat. Biotechnol. 24, 185- 187 (2006) . In some embodiments, the PSCs are not hES cells.

In some embodiments, the ES cells may be obtained by methods which do not involve the destruction of a human embryo or the use of a human embryo for an industrial or commercial purpose. For example, hES cells may be obtained by biastomere biopsy techniques

(Klimanskaya (2013) Semin Reprod Med. 31 (1) : 9-55; Klimanskaya et al Nature (2006), 444(7118)481-5; Chung et al Cell Stem. Cell 2008, 2(2), 113-117; US7893315) . In other embodiments, the pluripotent stem cells are iPS cells, for example human iPS cells. iPS cells are pluripotent cells which are derived from non- pluripotent ancestor cells, for example somatic cells, such as fibroblasts. Ancestor cells are typically reprogrammed into iPS cells through the introduction of reprogramming factors Oct4, Sox2, Klf4 and c-Myc into the cell. Other suitable reprogra ming factors and combinations of reprogramming factors for inducing pluripotency are known in the art. (see, for example, Yu et al Science 318 200/ 1917-1920, Tesar, P. J. et al . Nature 448, 196-199 (2007); Nichols, J. & Smith, A. Cell Stem Cell 4, 487-492(2009); Ying, Q. L. et al . Nature 453, 519-523 (2008) , Hanna J, et al Proc Natl Acad Sci U S A. 2010 May 18;107 (20) : 9222-7; Han DW, et al Nat Cell Biol. 2011 Jan; 13 (1) : 66-71; Silva J et al Cell. 2009 Aug 21; 138 ( ): 722-37 ) .

Reprogramming factors and techniques for the production of iPS cells are well-known in the art and include introducing reprogramming factors by plasmid or viral transfection, direct protein delivery or direct delivery of nucleic acid, such as mRNA. (Yamanaka et al Nature (2007); 448:313-7; Yamanaka 6 (2007) Jun 7 ; 1 ( 1 ) : 39-49. Kim et al. Nature. (2008) Jul 31; 454 (7204 ): 646-50; Takahashi Cell. (2007) Nov 30; 131 (5 ): 861-72. Park et al Nature . (2008) Jan 10;

451(7175) :141-6; Kimet al Cell Stem Cell. (2009) Jun 5;4 (6) : 472-6; Vallier, L. , et al . (2009) Stem Cells 27, 2655-66. ) .

The non-pluripotent ancestor cells for use in the production of iPS cells may be obtained from an individual . The individual may be healthy (i.e. without any disease condition) or may have a disease condition. For example, iPS cells may be derived from a sample of cells obtained from an individual with a haematological condition, for example a thrombocytopenic or other platelet-related condition, including essential thrombocytosis and congenital amegakaryocytic thrombocytopenia (CAMT) , Thrombocytopenia-absent radius syndrome

(TAR), Bernard Soulier syndrome (BSS), Gray platelet syndrome (GPS), and Glanzmann thrombasthenia . IPS cells obtained from an individual with a haematological condition may be used to generate

megakaryocyte progenitor cells or mature megakaryocytes using the methods described herein for modelling a haematological condition; for the treatment of an individual with a haematological condition or for the generation of platelets for the treatment of an

individual with a haematological condition (Cell Mol Life Sci. 2012 Apr 2 ) .

A population of pluripotent stem cells for use in the present methods, for example human pluripotent stem cells, may be obtained by culturing cells from a pluripotent cell line, using conventional techniques (Vallier, L. et al Dev. Biol. 275, 403-421 (2004) , Cowan, C.A. et al. N. Engl. J. Med. 350, 1353-1356 (2004) ,

Joannides, A. et al . Stem Cells 24, 230-235 (2006) Klimanskaya, I. et al. Lancet 365, 1636-1641 (2005) , Ludwig, T.E. et al. Nat.

Biotechnol . 24, 185-187 (2006) ) . For example, human pluripotent cells suitable for use in the present methods may be conventionally cultured in a culture dish on a layer of feeder cells, such as irradiated mouse embryonic fibroblasts (MEF) , at an appropriate density (e.g.10 ⁵ to 10 ⁶ cells/ 60mm dish), or on an appropriate substrate with feeder conditioned or defined medium. Human

pluripotent cells for use in the present methods may be passaged by enzymatic or mechanical means - Suitable culture media for human pluripotent cells include SC medium (Knockout Dulbecco' s Modified Eagle' s Medium (KO-DMEM) supplemented with 20% Serum. Replacement, 1% Non-Essential Amino Acids, ImM L-Glutamine, 0. ImM β-mercaptoethanol and rig/ml to lOng/ml human bFGF) and ES medium (DMEM/F12

supplemented with 20% knockout serum replacement (KSR) , 6 ng/ml FGF2 (PeproTech) , ImM L-Gln, 100 μιη non-essential amino acids, 100 μΜ 2- mercaptoethanol, 50 U/ml Penicillin and 50 mg/ml Streptomyci ) ,

A population of pluripotent stem cells for use in the present methods, for example human pluripotent stem cells, is preferably substantially free from one or more other cell types. Before introduction of feeder cells, the population of isolated pluripotent stem cells may be expanded. For example, the human pluripotent stem cells may be cultured in a monolayer under

conditions that simulate FGF2 signalling. In some embodiments , the cells may be cultured in a culture medium supplemented with FGF2 (e.g. 5 to 20 ng/ml FGF2, preferably lOng/ml). Suitable culture media include the SC and ES media described above, which may be MEF- conditioned and supplemented with FGF2.

Any mammalian FGF2 may be employed, preferably human fibroblast growth factor 2 ( FGF2 ) (NCBI Gene ID: 2247, nucleic acid sequence MM 002006.3 GI : 41352694, amino acid sequence MP 001997.4 GI :

41352695). FGF2 may be produced using routine recombinant techniques or obtained from commercial suppliers (e.g. R&D, Minneapolis, MM, USA) .

As described above, pluripotent cells are typically cultured and maintained on MEF feeder cells and may be separated from the feeder cells by any suitable technique. For example, the cells may be briefly (e.g. one hour) cultured on gelatin, and then the human pluripotent cells, which do not adhere to the gelatin separated from the MEFs which do adhere to the gelatin. In most experimental settings, human pluripotent stem cells have been cultivated in CDM on gelatin coated dishes (Vallier et al Curr Protoc Stem Cell Biol. 2008 Mar; Chapter l.Unit 1D.4.1-1D.4.7) in presence of FGF2 and Activin-A.

In some embodiments, the PSCs may have a defined genotype, such as a defined HLA haplotype . For example, the PSCs may be or have been subjected to biological characterisation, such as HLA genotyping . A method described herein may comprise providing a population of PSCs having a defined genotype, such as a defined HLA haplotype. For example, a population of individuals may be HLA/ABO typed and skin biopsy or peripheral blood from selected individuals in the

population with particular HLA/ABO genotypes may be used to generate iPSC lines for forward programming as described herein. Suitable methods of HLA/ABO typing are well-known in the art. Forward programming is the direct imposition of a more

differentiated phenotype on a pluripotent stem cell or other precursor cell which bypasses normal differentiation pathway; i.e. the cell does not pass through intermediate stages of

differentiation. For example, a PSC which is forward programmed into a megakaryocyte progenitor does not differentiate through all of the mesoderm progenitor, haemogenic endothelium progenitor and

hematopoietic progenitor stages before displaying the megakaryocyte progenitor phenotype. In other words, during step (iii), the cells said population may not progressively display each of the mesoderm progenitor, haemogenic endothelium progenitor and hematopoietic progenitor phenotypes. In some embodiments , a PSC may be

differentiated into a mesoderm progenitor cell as described herein and the mesoderm progenitor cell may be forward programmed into a megakaryocyte progenitor ,

Mature megakaryocytes are non-proliferative bone marrow cells which are responsible for the production of platelets. Mature

megakaryocytes may have the phenotype CD34+/-, CD61+, CD41a+,

CD42a+, CD42b+, GPVI+ or CD61+, CD41a+, CD42a+, CD42b+, GPVI+, CD235a+/-. Mature megakaryocytes are large (20-100um) polyploid cells (4-128N) which eventually produce platelets through pro- platelet formation. They include megakaryoblast, pro-megakaryocyte and megakaryocyte stages as described in Journal of Thrombosis and Hae ostasis, 5 (Suppl. 1) : 318-327.

Megakaryocyte progenitor cells are proliferative precursors of mature megakaryocytes which undergo a final differentiation step to form mature megakaryocytes. Megakaryocyte progenitors may have the phenotype CD235a +/-, CD34+/-, CD61+ , CD41a+, CD42a ^~ , CD42b-, GPVI- . They include BFU-MKs (burst forming units-megakaryocytic) , CFU-MKs (colony forming units-megakaryocytic) and PMKBs (promegakaryoblasts } as described in Journal of Thrombosis and Haemostasis, 5 ( Suppl . 1) : 318-327.

A megakaryocyte progenitor phenotype may include surface expression of CD235a, CD34, CD61 and CD41a or CD34, CD61 and CD41a and may not include expression of CD42a, CD42b and GPVI.

PSCs are forward programmed to become megakaryocyte progenitors in the methods described herein through the introduction of a specific combination of transcription factors, which causes the intracellular levels of the transcription factors in the PSCs to be increased.

The combination of transcription factors may be introduced into the PSCs in the form of nucleic acids (Warren L et al . Cell Stew, Cell. 2010 Nov 5;7 (5) : 618-30) or proteins (Zhou H, et al Cell Stem Cell. 2009 May 8 ; 4 (5 ) : 381- ) by any suitable technique, including plasmid or more preferably, viral transfection, direct protein delivery or direct delivery of nucleic acid, such as mR A. Following

introduction of the reprogramming nucleic acids or proteins, the population of treated cells may be cultured. The combination of transcription factors, for example GATA1, FLU and TALI and optionally one or more additional transcription factors, may be introduced into the PSCs by expressing nucleic acid encoding the combination of transcription factors in the PSCs. For example, the nucleic acid may be operably linked to inducible or non-inducible regulatory elements within a suitable vector, for example a retroviral or lentiviral vector, for expression within the cells. Vectors containing the nucleic acid are then transfected into the PSCs. Any convenient technique for the transfection may be employed. Following transfection, the combination of transcription factors is expressed in the PSCs and programs the PSCs to become megakaryocyte progenitors .

In some embodiments, transposon-mediated or other random integration transgenesis techniques may be employed. Reprogramming cells through expression of nucleic acid encoding one or more

transcription factors is well-known in the art (Takahashi et al

2007; Takahashi et al 2007; Seki et al 2010; Loh et al 2010; Staerk et al 2010) . In some preferred embodiments, the PSCs may be programmed to become megakaryocyte progenitors with minimal or no genetic modification to the cells, Suitable techniques are known in the art and include the use of excisable lentiviral and transposon vectors; repeated application of transient plasmid, episomal and adenovirus or adeno- associated vectors or; the use of small molecules, synthetic mRNA and/or microRNAs (Sidhu KS . Expert Opin Biol Ther. (2011) May;

11 (5) : 569-79; Woltjen K et al (2009) Nature 458 (7239) : 766-70; Chou BK et al. Cell Res. 2011 21 ( 3 ) : 518-29 ) .

In other embodiments, the combination of transcription factors, for example GATA1, FLU and TALI and optionally one or more additional transcription factors, for example one or more transcription factors from Table 1 and/or Table 2, may be introduced into the PSCs by contacting transcription factor proteins or transcription factor nucleic acids, such as mRNAs encoding transcription factors, with the population of PSCs. Programming cells though contact with transcription factor nucleic acids (Warren L et al . Cell Stem Cell. 2010 Nov 5;7 (5) : 618-30) or proteins (Zhou H, et al Cell Stem Cell. 2009 May 8 ; 4 (5 ) : 381- ) is well-known in the art and any suitable technique may be employed. For example, the combination of

transcription factor proteins or nucleic acids may be cultured in the presence of the PSCs under conditions which allow for entry of the proteins or nucleic acid into the cell. In some embodiments, entry of transcription factor proteins into the cell may be

facilitated by a membrane penetrating peptide, which may be linked or attached to the transcription factor proteins . The combination of transcription factor proteins or nucleic acids may be introduced into the PSCs by traditional methods such as lipofeetion,

electroperation, calcium phosphate precipitation, particle

bombardment and/or microinjection, or may be delivered into cells by a protein delivery agent. For example, the combination of

transcription factor proteins or nucleic acids can be introduced into cells by covalently or non-covalently attached lipids, e.g. a myristoyl group.

Transcription factor nucleic acids for direct delivery into PSCs may be translatable by endogenous translation factors within the cell. Suitable synthetic mRNAs may be modified. For example, 5- methylcytidine may be substituted for cytidine, and pseudouridine for uridine, followed by phosphatase treatment to produce the transcription factor nucleic acids (Zhou H, et al 2009). In other embodiments, the combination of transcription factors, for example GATA1, FLU and TALI and optionally one or more additional transcription factors, for example one or more transcription factors from Table 1 and/or Table 2, may be introduced into the PSCs by activating expression of endogenous nucleic acid sequences encoding the transcription factors in the population of PSCs. Suitable techniques for endogenous gene activation include Zinc Finger or Transcription like Activator (TAL) techniques and are well

established in the art (see for example Hum Gene Ther. 2012 May 15; Zhang P et al. Hum Gene Ther. 2012 Nov; 23 ( 11 }: 1186-99) .

In preferred embodiments, the PSCs are forward programmed in a chemically defined medium (CDM) . A CDM is a nutritive solution for culturing cells which contains only specified components, preferably components of known chemical structure. A CDM is devoid of

components which are not fully defined, for example serum or proteins isolated therefrom, such as Foetal Bovine Serum ( FBS ) , Bovine Serum Albumin (BSA) , and feeder or other cells. In some embodiments, a CDM may be humanised and may be devoid of components from non-human animals. Proteins in the CDM may be recombinant human proteins Suitable CDMs are well known in the art and described in more detai1 be 1ow ,

Media and ingredients thereof may be obtained from commercial sources (e.g. Gibco, Roche, Sigma, Europabioproducts , Cellgenix,

Life Sciences) . In a humanised CDM, for example BSA may be replaced in CDM by Polyvinyl alcohol (PVA), human serum albumin, Plasmanate™ (human albumin, alpha-globulin and beta globulin: Talecris

Biotherapeutics MC USA) or Buminate ^VM (human albumin: Baxter

Healthcare), all of which are available from commercial sources.

Suitable CDMs include Knockout (KS) medium supplemented with 4 ng/ml FGF ₂ ; Knockout Dulbecco's Modified Eagle's Medium (KO-DMEM)

supplemented with 20% Serum Replacement, 1% Non-Essential Amino Acids, ImM L-Glutamine, 0. ImM β-mercaptoethanol and 4ng/ml to

10ng/ml human FGF2 ; and DMEM/F12 supplemented with 20% knockout serum replacement (KSR) , 6 ng/ml FGF2 (PeproTech) , ImM L-Gln, 100 μιπ non-essential amino acids, 100 μΜ 2-mercaptoethanol , 50 U/ml penicillin and 50 mg/ml streptomycin and TeSR (Ludwig et al Nat Biotech 2006 24 185). Other suitable CDM which may be used in accordance with the present methods are known in the art (e.g. N12 medium, Johansson and Wiles CDM; Johansson and Wiles (1995) Mol Cell Biol 15, 141-151) . Suitable humanised CDMs may comprise a basal culture medium, such as IMDM and/or F12 supplemented with insulin, for example at Ο.δμς/ιηΐ to 7( g/ml, transfer in, for example at a concentration of 1.5p.g/ml to 15Q g/ml, an antioxidant, such as 1-thiolglycerol, for example at a concentration of 45μΜ to 4.5mM, lipids, and one or more of human serum albumin, polyvinyl alcohol (PVA) , Plasmanate™ (human albumin, alpha-globulin and beta globulin: Talecris Biotherapeutics NC USA) or Buminate™ (human albumin: Baxter Healthcare) , for example at a concentration of 0.5 mg/ml to 50 mg/ml. For example, humanised CDM include humanised Johansson and Wiles CDM, which consists of: 50% IMDM (Gibco) plus 50% F12 NUT-MIX (Gibco) 7μσ/πι1 insulin; Ιδμς/ιηΐ transferrin; 5 mg/ml human serum albumin, polyvinyl alcohol (PVA), Plasmanate™ or Eliminate™ ; 1% chemically defined lipid concentrate ( Invitrogen) ; and 450μΜ 1-thiolglycerol. Another suitable chemically defined medium may comprise 50% IMDM, 50% F12 NUT-MIX, lug/nil insulin, 15μ9 πι1 transferrin, 1% chemically defined lipid

concentrate, 5 mg/ml human serum albumin or Polyvinyl Alcohol (PVA) and 450μΜ 1-thiolglycerol . Another suitable chemically defined medium is CellGRO SCGM ^lM which is commercially available (Cellgenix, DE ) .

Following the introduction of the combination of reprogramming factors into the PSCs, the cells may be cultured in a pluripotency cell culture medium, for example CDM with Activin-A and FGF2, or mesodermal cell culture medium, for example CDM supplemented with BMP4 (e.g. rh-BMP4 at. lOng/ml) and/or FGF2 and/or LY294002 for 1 or more, 2 or more, 3 or more, 4 or more, or 5 or more days, preferably about 2 days .

The cells may then be cultured in an appropriate megakaryocyte (MK) programming medium. For example, population of cells may be cultured in a chemically defined medium (CDM) supplemented with TPO (Thrombopoietin) and/or SCF (Stem Cell Factor) and /or IL1B, (all preferably recombinant, human proteins) . The amino acid sequences of TPO, SCF and IL1B are readily available on public databases. For example, the reference amino acid sequence of human TPO (Thrombopoietin: also known as THPO: Gene ID 7066) has the NCBI database entry NP 000451.1 GI: 4507493; the reference amino acid sequence of human SCF(Stem Cell Factor: also known as KITLG: Gene ID 4254) has the NCBI database entry NP Q00890.1 GI : 505175 and the reference amino acid sequence of human IL1B (Interleukin 1 beta: Gene ID 3553) has the NCBI database entry NP_000567.1

GI:10835145.

TPO, SCF and IL1B may be produced by synthetic or recombinant means or obtained available from commercial suppliers (e.g. R&D Systems, Minneapolis, MINI, USA; Sigma-Aidrich Co. LLC USA, EMD Milli ore MA USA) . For example, the pluripotent cells may be cultured by a method comprising ;

(i) culturing said pluripotent cells in mesoderm medium comprising FGF2 and BMP4 ; and,

(ii) further culturing said cells i MK programming medium comprising PO and/or SCF, preferably TPO and SCF.

Suitable media are known in the art (Cell Stem Cell. 2011 Aug

5; 9 (2) : 144-55; Dev Cell. 2011 May 17 ; 20 ( 5 ) : 597 -609 ; Blood. 2008 Jun

1; 111 (11) : 5298-306) and are described in more detail below.

Suitable cell culture conditions are well known in the art. (Vallier, L. et al Dev. Biol. 275, 403-421 (2004), Cowan, C.A. et al . N. Engl. J. Med. 350, 1353-1356 (2004), Joannides, A. et al . Stem Cells 24, 230-235 (2006) Kli anskaya, I. et al. Lancet 365, 1636-1641 (2005), Ludwig,T.E. et al. Nat. Biotechnol. 24, 185-187 (2006)).

Methods for culturing mammalian cells are well-known in the art (see, for example, Basic Cell Culture Protocols, C. Helgason, Humana Press Inc. U.S. (15 Oct 2004) ISBN: 1588295451; Human Cell Culture Protocols (Methods in Molecular Medicine S.) Humana Press Inc., U.S. (9 Dec 2004) ISBN: 1588292223; Culture of Animal Cells: A Manual of Basic Technique, R. Freshney, John Wiley & Sons Inc (2 Aug 2005) ISBN: 0471453293, Ho WY et al J Immunol Methods. (2006) 310:40-52, Handbook of Stem Cells (ed. R. Lanza) ISBN: 0124366430) . Media and ingredients thereof may be obtained from commercial sources (e.g. Gibco, Roche, Sigma, Europa bioproducts, R&D Systems). Standard mammalian cell culture conditions may be employed, for example 37°C, 21% Oxygen, 5% Carbon Dioxide. Culture medium is preferably changed every two days and cells allowed to settle by gravity.

The population of pluripotent stem cells may be cultured for at least. 7 days after introduction of the combination of transcription factors . Megakaryocyte progenitors may be identified in the cell culture after at least 4, 5, 6, or 7 or more days.

In some embodiments, a method may comprise identifying or confirming the identity of the megakaryocyte progenitor cells or mature megakaryocytes in the culture.

In some embodiments, cells may be tested for presence of cell markers associated with the megakaryocyte progenitor cells, for example to identify or confirm their identity. Cells which express the markers may be identified as megakaryocyte progenitor cells. For example, megakaryocyte progenitor cells may identified by expression of CD34 and CD41a as described above but no expression of CD42a and CD42b.

Megakaryocyte progenitor cells do not express the pluripotency associated markers, such as Oct , Sox2, Alkaline Phosphatase, SSEA- 3, anog, SSEA-4 and Tra-1-60, which are expressed by PSCs or display reduced expression relative to PSCs.

A method may further comprise isolating and/or purifying the forward programmed megakaryocyte progenitor cells. Megakaryocyte progenitors may be separated from other cell types in the population using any technique known to those skilled in the art, including those based on the recognition of extracellular epitopes by antibodies and/or magnetic beads or fluorescence activated cell sorting (FACS), including the use of antibodies against extracellular regions of characteristic markers . The megakaryocyte progenitor cells may be cultured and/or expanded to generate a homogenous or substantially homogenous population of cells. Suitable techniques for mammalian cell culture are well known in the art and described elsewhere herein. A method may comprise monitoring or detecting the expression of one or more megakaryocyte progenitor cell markers and/or one or more plu.ripot.ent cell markers in cells in the population. This allows the extent, of forward programming in the population to be determined as it is cultured.

At least 5%, at least 10%, at least 20%, at least 30%, at least 40%, at least 50% or at least 60% of the population of PSCs may become megakaryocyte progenitor cells following forward programming as described herein.

Megakaryocyte progenitor cells produced by the present methods may be substantially free from other cell types. For example, a population of megakaryocyte progenitors produced by a method described herein may contain 80% or more, 85% or more, 90% or more, or 95% or more megakaryocyte progenitor cells, following culture. The population of megakaryocyte progenitor cells may be cultured and/or expanded and optionally stored.

Following the production of a population of megakaryocyte progenitor cells by forward programming, the methods described herein may further comprise allowing the population of megakaryocyte progenitor cells to differentiate into mature megakaryocyte cells, for example by culture in a megakaryocyte maturation medium.

For example, megakaryocyte progenitor cells may be passaged, for example from day 7 or day 10, into suspension culture plastic dishes in a megakaryocyte maturation medium comprising CDM (e.g. CellGRO SCGM™) supplemented by TPO and/or ILl-beta and/or SCF; and then cultured for an additional 5-7 days or 5-15 days, up to 10-15 days. In some embodiments, the megakaryocyte progenitor cells may be cultured in a megakaryocyte maturation medium comprising CDM (e.g. CellGRO SCGM™} supplemented by TPO and/or ILl-beta and/or SCF, preferably TPO and ILl-beta, from day 10 onwards for an additional 5-7 days or 8-12 days, preferably 10 days.

Typically, the cells are cultured for 18 to 22 days from TF

transduction, preferably about 20 days, to produce mature

megakaryocyte cells . The mature megakaryocyte cells may be isolated, purified and/or stored according to standard techniques.

In preferred embodiments, a population of mature megakaryocyte cells produced by a method described above may be pure or substantially pure and may not require further sorting or purification. For example, at least 80%, at least 90% or at least 95% of cells in the population of mature megakaryocyte cells may express CD41a (i.e. CD41a+ cells) . At least 30%, at least 40% or at least 50% of cells in the population may express CD42a (i.e. CD 2a+ cells). In some preferred embodiments, at least 95% of cells in the population express CD41a and at least 50% of cells in the population may express CD42a.

In some embodiments , mature megakaryocyte cells produced as

described herein may be used in the production of platelets. For example, a method described herein may comprise allowing one or more of the mature megakaryocyte cells to produce platelets.

Another aspect of the invention provides a population of

megakaryocyte progenitor cells or mature megakaryocytes produced as described herein. Populations of megakaryocyte progenitor cells and mature

megakaryocytes produced by forward programming as described herein are described in more detail above.

Megakaryocyte progenitor cells produced by forward programming as described herein may be highly proliferative (for example over 28 days or more ) .

Mature megakaryocyte cells produced by forward programming and differentiation as described herein may generate and/or release functional platelet like particles (PLPs) in vitro .

Another aspect of the invention provides the use of a population of megakaryocyte progenitor cells or mature megakaryocytes produced as described herein in the production of platelets.

Another aspect, of the invention provides megakaryocyte progenitor cells or mature megakaryocytes produced as described herein for use in methods of treatment of haematological conditions as described herein and methods of treatment of haematological conditions which comprise administering megakaryocyte progenitor cells or mature megakaryocytes produced as described herein to an individual in need thereof .

Megakaryocyte progenitor cells or mature megakaryocytes produced as described herein may also be useful in screeni g. Screening may include drug or small molecule screening. For example, the isolated programmed cells may be contacted with a test compound and the effect of the test compound on the cells is determined. Screening may also include functional genomic screening. For example, a gene may be suppressed, knocked out or otherwise inactivated in the isolated reprogrammed cells and the effect of the inactivation on the cells determined.

In some embodiments, megakaryocyte progenitor or mature

megakaryocytes cells may be produced from iPS cells as described herein. The iPS cells may be derived from normal differentiated cells or from differentiated cells having a disease phenotype or genotype, for example from an individual with a disease condition. After programming, the megakaryocyte progenitor cells or mature megakaryocytes may express a detectable reporter or display an observable cellular phenotype which differs between disease-affected cells and normal cells. The megakaryocyte progenitor cells or mature megakaryocytes may be exposed to test compounds and the effect of the test compound on the reporter expression or observable cellular phenotype determined. Compounds which cause the megakaryocyte progenitor cells or mature megakaryocytes to revert from disease cell state to the normal state may be identified. Alternatively, the one or more genes in the megakaryocyte progenitor cells or mature megakaryocytes may be inactivated, for example by targeted mutation or RNAi suppression, and the effect of the inactivation on the reporter expression or observable cellular phenotype determined. Genes whose inactivation causes the cells to revert from disease cell state to the normal state may be identified.

Screening may include toxicology screening. For example, the isolated megakaryocyte progenitor cells may be contacted with a test compound at various concentrations that mimic abnormal /normal concentrations in vivo. The effect of the test compound on the cells may be determined and toxic effects identified. Toxicology screening is well known in the art (see for example Barbaric I et al. Biochem Soc Trans. 2010 Aug; 38 ( } : 10 6-50 ) . Forward programmed megakaryocyte progenitor cells (and cells derived from the programmed cells, such as mature megakaryocytes and platelets) may also be used for the treatment of an individual, for example for the treatment of a platelet or megakaryocyte related condition. The individual may be the same individual from whom the original IPS cells were obtained.

In some embodiments, the forward programmed megakaryocyte progenitor cells or mature megakaryocytes may, for example, be admixed with a pharmaceutical acceptable carrier in a pharmaceutical composition. The composition may be administered to the individual ((Leukemia. 2008 Jan; 22 (1) :203-8) . Forward programmed megakaryocyte progenitor cells and cells derived from the programmed cells, such as mature megakaryocytes, may also be used for disease modelling. For example, cells may be programmed into megakaryocyte progenitors which are affected in a disease condition, either directly or by differentiating into the affected megakaryocytes or platelets. The effect of the mutation on the cellular phenotype may be studied and the genetic and/or biochemica interactors that contribute to the cellular pathology of the diseas may be identified and/or characterised. Another aspect of the invention provides a method of screening for compound useful in the treatment of a disease condition, in particular a haematological condition, for example a

thrombocytopenic or other platelet-related condition, including essential thrombocytosis, congenital amegakaryocytic

thrombocytopenia (CAMT) , Thrombocytopenia-absent radius syndrome

(TAP.), Bernard Soulier syndrome (BSS), Gray platelet syndrome (GPS ) and G1anzmann throrabasthenia , compris ing;

contacting a population of megakaryocyte progenitor cells produced by a method described above with a test compound, and; determining the effect of the test compound on said cells and/or the effect of said reprogrammed cells on the test compound.

Suitable forward programmed megakaryocyte progenitor cells are described above.

The megakaryocyte progenitor cells may display a disease phenotype and the effect of the test compound on one or more disease

pathologies in the reprogrammed cells may be determined. A decrease or amelioration of one or more disease pathologies in the

reprogrammed cells in the presence, relative to the absence of test compound is indicative that, the test, compound may be useful in the treatment of the disease in the individual.

Suitable disease conditions and phenotypes are described above.

The forward programmed megakaryocyte progenitor cells may display a normal phenotype and the effect of the test compound on the growth, differentiation or viability of the reprogrammed cells or the ability of the reprogrammed cells to perforin, one or more cell functions may be determined. In some embodiments, cells may be modified to express reporters that can be used to measure particular cell functions or attributes. A decrease in growth, viability or ability to perform one or more cellular functions may be indicative that the compound has a cytotoxic effect (see for example, Barbaric I et al Biochem Soc: Trans. 2010 Aug; 38 ( ): 1046-50) .

The data set out herein shows that GATA1 and FLU are also able to drive forward programming in the absence of TALI. Other aspects and embodiments of the invention provide all of the aspects and

embodiments described above with the transcription factor TALI omitted from the combination of transcription factors.

Other aspects and embodiments of the invention provide the aspects and embodiments described above with the term, "comprising" replaced by the term "consisting of" and the aspects and embodiments

described above with the term "comprising" replaced by the term "consisting essentially of".

Various further aspects and embodiments of the present invention will be apparent to those skilled in the art in view of the present disclosure . All documents and database entries which are mentioned in this specification are incorporated herein by reference in their entirety for all purposes .

The transcript ion fact.or symbo1s and names set out herein are the unique official HUGO Gene Nomenc ^" Lature Committee (HGNC) symbols and names that have been assigned to that transcription factor (see Gray KA et al Nucleic Acids Res. 2013 Jan 1; 41 (Dl ) : D545-52 ; and the HGNC Database, HUGO Gene Nomenclature Committee (HGNC), EMBL Outstation - Hinxton, European Bioinformatics Institute, Wellcome Trust Genome Campus, Hinxton, Cambridgeshire , CB10 ISO, UK) .

"and/or" where used herein is to be taken as specific disclosure of each of the two specified features or components with or without the other. For example "A and/or B" is to be taken as specific

disclosure of each of (i) A, (ii) B and (iii) A and B, just as if each is set out individually herein. Unless context dictates otherwise, the descriptions and definitions of the features set out above are not limited to any particular aspect or embodiment of the invention and apply equally to all aspects and embodiments of the invention which are described. Thus, the features set out above are disclosed for use in the invention in all combinations and permutations .

Certain aspects and embodiments of the invention will now be illustrated by way of example and with reference to the figures and tables described herein.

Figure 1 shows the internal biological interactions of the top 20 TF candidates identified using VisANT, highlighting the 9 tested genes in this study (in dark) and the GA A centred network (grey shade). Tested combinations are indicated in the adjacent table.

Figure 2 shows a comparison between 14TF and 9TF combinations for MK-FoP of the H9 ES line (hESC#l) . Figure 3 shows transgene expression levels quantified by RT-QPCR in flow cytometry sorted CD41a+ cells generated 7 days after lentiviral transduction of the hESC#l line with the 9 TFs on fibronectin coated plates and maintained in pluripotent medium ( FGF2+Activin-A) for 2 days followed by MK medium (TPO+SCF) for 5 days.

Figure 4 shows the fold increase of CD41a+ cells relative to the 9- TFs combination (mean ± sem, n=3/3/l/2 respectively) in hESC#l cells transduced with different combinations of TFs using a 2 day mesoderm induction and measured by flow cytometry at day 7. 3TFs combination is shown as the most effective for MK-FoP.

Figure 5 shows the fold increase of CD41a+ cells relative to the 9- TFs in hESC#l cells transduced with different combinations of TFs under pluripotent conditions without mesoderm induction and measured by flow cytometry at day 7. 3TFs combination is shown as the most effective for MK-FoP.

Figure (5 shows the fold increase of CD41a+ cells relative to the 3- TFs combination (mean ±sem, n=2 ) in hiPSC#l cells transduced with all permutations of the 3-TFs combination using 2 days mesoderm induction and measured by flow cytometry at day7. Figure 7 shows real time quantitative PGR analysis of MK gene expression at day 7 in an unsorted population: HES3 (hESC#2) ESC line before and after forward programming protocol (21% CD 1a+ at day7 ) . ES: pluripotent ESC line; GFP: GFP transduced cells were cultivated in the same conditions as for 3TFs transduced cells,

Figure 8 shows the megakaryocytic potential of cells sorted by flow cytometry based on CD41a expression at day7 following transduction of hiPSC#2 line with the 3-TFs combination. The megakaryocytic potential of sorted cells was tested using the Megacult clonogenic assay (1,000 cells in duplicate). A representative megakaryocyte colony obtained from a CD41a+ cell and co-expressing CD41a and CD42b as detected by immunofluorescence is shown. Figure 9 shows real-time quantitative PGR analysis of MK gene expression at day 7 in sorted populations following forward

programming of the hESC#l line.

Figure 10 shows a time course flow analysis of the mean values +/- SD of surface marker expression using forward programming protocol on the hiPSC#l and #2 lines.

Figure 11 shows the results of cytospin and Romanowsky staining: polyploid MK are produced from, forward programmed hiPSC#l and#2 lines (BOB and BBHX ) and are similar to cord blood derived MKs .

Figure 12 shows the percentages of CD41a expressing cells produced when human pluripotent stem cells sown on fibronectin coated plates were transduced in parallel with the 3-TFs and kept for 2 days in pluripotent ( FGF2+Act.ivin-A) or mesoderm (FGF2+BMP4+LY294002) medium followed by MK medium. (TPO+SCF) for 6 days.

Figure 13 shows histograms representing the CD41a+ cells fold increase relative to the hiPSC input (mean ±sem, n=4/3 respectively) when the hiPSC#2 line was sown in fibronectin coated plates (2-D culture) or induced to form embryoid bodies by forced aggregation of single cells and subsequently transduced with the 3-TFs, Cells were kept 2 days in mesoderm medium and 5 days in MK medium. Figure 14 shows the morphology of day20 hiPSC-MKs and CB-MKs analysed after Romanowsky staining. Arrowheads point to cells showing several nuclei. Scale bars, lOOu . Figure 15 upper panel shows histograms representing the percentages of CD41a+ cells generated at daylO (mean ±sem, n=ll/6/2/l

respectively) from a variety of hiPSC lines using the optimised 3- TFs forward programming protocol. Lower panel shows cell fold expansion at day20 for total and mature megakaryocytes (CD41a and

CD42b positive cells respectively) relative to the hiPSC input (mean +sem, n=6/6/2 respectively) . MD not done.

Figure 16 shows histograms representing the amount of PLPs generated per hiPSC-MK input after hiPSC#l day 20 MKs were sowed in parallel on different stromal cell lines .

Table 1 shows 46 candidate TPs selected upon analysis of protein- protein interaction network using VisANT software, integrating interactions with chromatin remodelling factors and level of gene expression. 14 transcription factor candidates cloned into

lentiviral vector backbones are highlighted.

Table 2 shows 13 TPs identified from differential expression between cord blood or peripheral blood derived megakaryocytes and hiPSC-MKs

Table 3 shows an enrichment analysis for biological processes using the DAVID bioinformatics resource which indicates significant enrichment for megakaryocyte/platelet function genes in hiPSC-MKs compared to the starting hiPSCs . Input gene expression dataset from CD42b sorted hiPSC-MKs (line #1 and #2) and hISPC pluripotnet culture analysed on an Illumina Human HT-12 v4 BeadArray (hiPSC#l- MK, n=4; hiPSC#2-MK, n=2; hiPSC#l, n=2) . 1. Materials and Methods

1.1 Cell lines

The human embryonic stem cell lines HES3 and H9 (from ES Cell

International, Singapore and Wicell, Madison respectively) and the human iPSC lines #1-4 (A1ATD1 , BBHX8, iPS40 and S4-SF5 respectively, obtained from the Cambridge Biomedical Research Centre iPSC Core

Facility) are cultivated at 37C/5%C02 in chemically defined culture conditions as described previously (Curr Protoc Stern. Cell Biol. 2008 Mar; Chapter l:Unlt ID .4.1-lD .4.7) , Briefly, cells are maintained in a chemically defined basal medium (CDM) supplemented with

recombinant human FGF2 (12ng/ml, University of Cambridge) and

Activin-A (15ng/ml, University of Cambridge) on feeder free gelatin coated wells as previously described {Curr Protoc Stem Cell Biol, 2008 Mar; Chapter l:Unlt ID , 4.1-lD .4 , 7) with medium changes daily. Subculture is performed every 5-7 days by detaching pluripotent colonies by incubation in a dispase/collagenase-IV mix ( Img/ml , Sigma Aldrich) for 45 minutes at 37°C, collecting detached colonies, breaking them down into small clumps carefully pipetting up and down with a P1000 tip and plating them onto new gelatin coated plastic dishes .

Human iPSC line derivation has been performed under appropriate ethical approval and volunteer consent were obtained (Ethics reference no. 08/H0311/201; R&D no. A091485) . The iPSC lines have been derived from adult dermal fibroblasts using murine

oncoretroviral vectors (hiPSC #1-3) or Sendai vectors (hiPSC #4) expressing the human OCT4, SOX2 , KLF4 and MYC reprogramming factors and following subculture steps on irradiated mouse embryonic feeder cells in 20% SR medium supplemented by rh-FGF2 as previously described .

1.2 Transcription factor candidate selection

We performed a differential expression analysis focused on DNA binding protein coding genes (PANTHER) from whole genome expression data generated in the H9 hESC line (internal data, Illumina HumanWG- 6 v3 ) and human cord blood derived Ks {HaemAtlas, Blood,

2009; 113 :el-9, Illumina HumanWG-6 v2 ) . The list of 116 MK specific genes generated was further refined by removal of 21 histone coding genes and addition of 6 genes with known or potential roles in megakaryopoiesis . Using the VisANT web-based software (Nucleic Acids Res. 2005; 33:^352-357) for analysing networks of biological interactions, the resulting 101 candidate genes were subsequently ranked based on number of 1) internal protein interactions, 2) interaction with epigenome modifiers (HAC, HDAC, DNMT) and 3) differential expression levels. Weakly differentially expressed genes (log2 (MK-ESC) <1 ) were excluded from the candidate list. The final gene candidate list is eventually made of 46 factors ranked (from highest to lowest value) on 1) number of internal interactions 2) number of nodes with chromatin remodelling factors 3)

differential expression value. Exceptions in the final list, are the MLL and MLL3 genes which were removed since they have coding sequences of over lOkb and are incompatible with lentiviral vectors. MCM7 was also removed since it is not known to act as a

transcription factor (mini-chromosome maintenance complex) . CLOCK has been added despite a low expression value (0.89) because it has reported Histone Acetyl Transferase activity. A heatmap of differentially expressed transcription factors between cord blood or peripheral blood derived megakaryocytes and hiPSC-MKs was generated using Heatmap Builder vl .1 and a row normalised sorting algorithm (cut-off >2-fold increase, FDR1% ) . Each line of the heatmap represented a gene specific probe on the Illumina Human HT-12 v4 Beadarrays. Mean values of CB-MK (n=4) , PB-MK (n=2), hiPSC#l (n=4 ) and hiPSC#2 (n=2) . The candidate list was then reduced to the 13 TF genes showing very low/null expression in hiPSC-MKs (log2<8) . These factors may play an important role in further maturation of hiPSC-MKs

1.3 Transcription factor cloning into lentiviral backbone

The full human coding sequences of the 9 candidate genes (CDS of transcript variants 1, NCBI Refseq) including the 5' Kozak concensus sequence were generated by PGR using cDNA from cord blood derived

MKs as template. PGR fragments were subsequently individually cloned into the pWP lentiviral vector backbone ( Trono laboratory _/ Adcigene 12255) in place of the eG-FP coding sequence in-between the Mlul and Sail restriction sites, downstream of the human EFl-alpha ubiquitous promoter. Final constructs has been checked for sequence integrity against the NCBI Refseq library before use for lentiviral vector production. Replication deficient lentiviral vector particles (LVPs) were produced by transient co-transfection of HEK 293T/17 cells (ATCC CRL-11268) with the generated pWPT constructs along with the psPAX2 and pMD2. G helper plasmids (Trono laboratory,, Addgene 12260 and 12259) using TransIT-LT1 transfection reagent (MirusBIO) . Crude supernatants containing LVPs were collected 48 hours after

transfection, filtered through a 0.45um membrane and DNasel treated before concentration by ultracentrifugation or PEG-based

precipitation [LentiX-concentrator, Clontech) . Functional LVPs titres were determined by QPCR measurement of provirus copy number in genomic DNA of transduced HCT116 cells (ATCC CCL-247) and were in the range of 1-10E+8 TU./ml for all vectors. Human PSC lines were routinely transduced using multiplicity of infection of 20 in presence of 10 micrograms Protamine Sulfate (Sigma, P4505) per millilitre of culture medium leading to 60-80% transduction

efficiencies through e periments.

1.4 MK Forward programming

Optimised embryoid body based protocol : on transduction day (day 0), sub confluent (50-80%) human pluripotent stem cells were dissociated to single cells using TrypLE (Life Technologies) for 5 min at 37C and viable cells counted on a haemocytometer . Embryoid body (EB) formation was initiated with 5x10 ^' to lx 10' viable cells per well of an Aggrewell™400 plate (Stemcell Technologies, France) in order to obtain embryoid bodies (EB) of 400 to 800 cells per EB following spin aggregation (detailed protocol in Aggrewell™ Technical manual} .

Importantly, lentiviral transduction was performed concomitantly to the aggregation phase. Briefly, cells were added to the well in CDM supplemented by Y-27632 (lOmicroM, Sigma Aldrich) , rh-BMP4 (lOng/ l, R&D) and protamine sulfate (10pg/ml, Sigma) . Concentrated lentiviral vectors individually coding for each forward programming factor were added to the well to MOI20 (multiplicity of infection) .

Subsequently, Aggrewell™ plates were centrifuged at lOOg for

3minutes and put into the incubator (37C/5%C02) for 24 hours. After 24h (dayl), transduced EBs were collected and sowed in ultralow adherent cell culture plates (Corning) at a density of 600 EB per 10cm2 dish in CDM supplemented with s rh-BMP4 (lOng/ml, R&D) and rh-FGF2 (5ng/ml, University of Cambridge). EBs were collected 24 hours later (day2) and further cultivated in ultralow adherent plates in Cellgro SCGM medium (Cellgenix, Germany) supplemented with rh-TPO (lOOng/ml, Cellgenix) and rh-SCF (25ng/ml, Life Technologies). At daylO, EBs were dissociated to single cells using CollagenaselV and Dispasell (Img/ml, Gibco) followed by enzyme free cell

dissociation buffer treatment (Gibco) . Collected cells were

cultivated on suspension culture plates {Grenier) for an additional 10 days in Cellgro SCGM supplemented with TPO (lOOng/ml) and IL1- beta (lOng/ml, Miltenyi Biotec) for further MK maturation.

Culture medium is changed every three days by aspirating half of the volume and adding 2 times cytokine concentrated fresh medium on top. Adherent cell protocol: Small cell clumps were generated from sub- confluent hPSC cultures using a CollagenaselV/Dispasell mix (lmg/ml) and sowed on human fibronectin coated (50ug/ml, Millipore) tissue culture plates in CDM with FGF (12ng/ml) and Activin-A (15ng/ml) at an approximated density of 2-5E+5 cells,/10cm2. Cells were transduced the day after with MOI20. The culture media used for the first two days were devised for pluripotency maintenance (as above) or mesoderm induction (FLyB; Bernardo et al.Cell Stem Cell. 2011 Aug 5; 9 (2) : 144-55) depending on experiments. The following days, cells were maintained in Cellgro SCGM supplemented with TPO (lOOng/ml) and SCF (25ng/ml) until analysis.

1.5 Cord blood derived megakaryocytes

Cord blood was obtained after informed consent under a protocol approved by the National Research Ethics Service. CDS4-positive cells (≥98%) isolated by magnetic cell sorting {Miltenyi Biotec) were seeded at 1E+5 cells/ml in Cellgro SCGM with TPO (lOOng/ml) and ILl-beta (lOng/ml) and incubated for 10 days. We routinely obtained 70-90% CD41a+ and 20-60% CD42a+ cells by the end of the culture.

1.6 Flow cytometry analysis

Flow cytometry experiments were performed on a CyAn ADP (Beckman Coulter) . Single cell suspensions were generated using

CollagenaselY/Dispasell and/or enzyme free dissociation buffer when needed. Cells were stained for 20-30' at room, temperature in PBS 0.5%BSA. 2mM EDT.A. using combinations of FITC, PE and APC conjugated antibodies (all from BD Pharmingen except anti-GP6 antibody from NHSBT-Brlstol) . Background fluorescence was set against matched isotype control antibodies and compensation matrix defined using single-colour stained cells. Flow count beads (Flow count

fluorospheres , Beckman Coulter) and DAP I were used to determine viable cell count in samples. 1.7 Immunofluorescence analysis

MKs were cultivated on human fibrinogen coated (50ug/ml, Millipore) tissue culture plates for 48 hours to monitor proplatelet formation. Cells were fixed with 2% PFA and permeabilised/blocked with 0.1% Saponin/0.2% Gelatin. Cells were incubated with primary antibodies (anti alpha-Tubulin, Sigma} anti vwf, Dako; anti P-selectin, NHSBT- Bristol) at room temperature for 2 hours and secondary antibodies conjugated with Cy3 or A.lexa-488 fluorochromes {Invitrogen Molecular Probes) for 45 minutes. Cell nuclei were stained with DAP I before image acquisition on a confocal Zeiss Axiovert 200M microscope.

1.8 Cell morphology analysis

Cells were spun on a glass slide using cytofunnels at. 400g for 5' , methanol fixed and Romanov/sky stained (eosin and methylene blue). Cells were observed on a phase contrast Axiovert Zeiss microscope (630x magnification). 1.9 Transmission electron microscopy

MKs were fixed in 2% glutaraldehyde 0.1M phosphate buffer for 60' at room temperature. After washing with phosphate buffer, the samples were post-fixed with 1% osmium tetroxide in phosphate buffer for 60' on ice, ethanol dehydrated and infiltrated with and embedded in lipoxy resin. Ultrathin sections (50nm) were cut and stained with 2% uranyl acetate in methanol and Reynolds' lead citrate. Samples were read using a FBI Tecnai 12 (Philips) transmission electron

microscope .

1.10 Gene expression analysis by RT-QPCR

Total RNA was extracted with RNeasy kits according to the

manufacturer's instructions (Qlagen) including DNase treatment. cD A was prepared from 250 -500ng RNA using Maxima First Strand cD A Synthesis Kit and random hexamers (Fennentas) . QPCR reactions were performed in duplicates using recommended SYBR green based PGR mixes on ΑΒΪ 7500HT or Mx30Q0P real time thermal cyclers using 2-step amplification protocols {Applied Biosystems, Agilent Technologies)■ Relative gene expression was calculated with the 2 ^"aelt ° ^c~ method using HMBS for normalisation. Primer pairs were designed to amplify cDNA only, have no reported off targets after blasting against human Refseq and showed 80-120%PCR efficiencies. Endogene specific primers were designed to amplify UTR regions absent from transgene sequences while transgene specific primer pairs have 3' primers binding to viral sequences.

1.11 Whole genome expression study

DNA free total RNA (RNeasy, Qiagen) was extracted from sorted CD42b+ cells (EasySEP, Stem cell Technologies; >95% purity) and 500ng were hybridized to Illumina Human HT-12 v4 BeadArrays , Data import. Raw Illumina bead-level output was imported to the R statistical programming environment using functions of the beadarray package for the Bioconductor software suite. Data processing. Signal intensities were background corrected, summarized and converted to log2

expression units using functionality of the beadarray package.

Probe-sets without signal deemed significantly above background level in all profiles of at least one sample group (Illumina signal detection statistic p < 0.01) were removed. Quantile normalization, implemented in the li ma package for Bioconductor, was employed to equalize summarized expression intensity distributions across all sample profiles. Probe sets were annotated to gene targets using information available from the manufacturer. Data analysis . The statistical overrepresentation of gene categories among genes deemed differentially expressed between sample group profiles was assessed using the DAVID bioinformatics resource. Gene set enrichment analyses were performed using web tools from the Broad Institute using Hae Atlas data as input gene set. Differential gene expression between two sample groups was assessed through the output of a moderated t-test and significance P-values obtained converted to corrected q-values using the FDR method.

1.12 In vitro platelet study.

Co-culture on feeder cells. In order to promote platelet like particle (PLP) production, daylO CB and day20 hiPSC derived MKs were further cultivated for 48h in Cellgro SCGM on gamma-irradiated stromal cells (OP9, ATCC CRL2749; C3H10T1/2, Riken Institute} HBMEC, courtesy of Dr. Weksler) sowed on gelatine coated tissue culture plates at 1E+4 cells/cm ² .

PLP flow analysis . Crude supernatant containing the PLPs were analysed by flow cytometry after addition of 1/9 volume of acid citrate dextrose (ACD, Sigma Aldrich) and cell removal by

centrifugation at 150g for 10' . Antibodies against human CD41a and CD42a were added directly to the medium {BD Pharmingen, APC/PE conjugated respectively, used at 1:50 dilution) and flow count fluorospheres used for quantification. Human platelets were analysed from fresh whole blood collected in citrate buffer.

Washed platelet preparation . Human platelet rich plasma (PRP) and hiPSC PLPs collected as above were washed twice in pH7.4 modified Tyrode-HEPES buffer (lOmM HEPES, 12mM NaHC03, 138mM NaCl, 5.5mM glucose, 2.9mM KC1, and ImM MgC12 ) using 800g/10' centrifugation steps after an initial addition of prostaglandin El (1 M) and apyrase (lU/ml) to prevent activation. Washed platelet counts were subsequently determined by flow cytometry.

Thrombus formation in laminar flow. A defined amount of washed platelets or PLPs are mixed with 1ml of mouse blood collected in ACD and the participation of human platelets to collagen induced mouse thrombi subsequently monitored by immunofluorescence. The procedure was modified from Auger JM, ATVB, 2008. Briefly, glass slides were locally coated with Honti collagen spots (lOQug/ml) and mounted into a flow chamber placed under a fluorescent microscope (EVOS system, Advanced Microscopy Group) . The blood was then perfused through the chamber at 1600s-l (7.2ml/hr) for 3 minutes allowing thrombi formation on collagen spots. A perfusion of 1:100 diluted anti CD41a-FITC ( BD Pharmingen) and anti P-selectin-PE (internal NHSBT- Bristol) was then run on the clots for 2 minutes and then washed before pictures were taken. 1.13 Gene expression analysis by quantitative PGR

For gene expression analysis, cultivated cells are treated by TRIzol and RNA extracted from the aqueous phase following published protocol (Life Technologies). Alternatively, RNA is extracted and purified using Qiaprep RNeasy mini columns (Qiagen) including an on column DNA digestion step. Subsequently, cDNA is synthesized from

250-50Qng of purified RNA using the Maxima Reverse Transcriptase kit and random hexamers (Fermentas) . The PGR reaction is performed using a SYBR green based PCR mix (Applied Biosysterns, FastSYBR green) on a real time thermal cycler analyser (ABI 7500HT) following a fast 2- step amplification protocol. Relative gene expression quantification is calculated using the 2 ^{" cc} method using HMBS endogene expression as a reference. Primer pairs specific for any given gene has been carefully design using the NCBI primer design website in order to be separated by at least one intron on the corresponding gDNA and with no identified potential off target after a BLAST on the human RefSeq repository. Furthermore, all primer pairs have been tested for PCR efficiency between 80-120%. Primer pair specific for endogene expression has been designed in the 5' or 3'UTR of corresponding transcript. Primer pair specific for transgene expression are made of a reverse primer specific for the lentiviral backbone and a forward primer specific for a given transgene.

1.14 Cell morphology analysis

Cell morphology is observed after sedimentation on a glass slide by cytospin (400g/5min, 2,000-20,000 cells per slide) and Romanowsky staining ( Eosin/Methylene blue) after Methanol fixation.

2. Results

2.1 A combination of GATA1, FLU and TALI induces megakaryocyte differentiation from human pluripotent stem cells

We used a rational whole genome expression data driven process to select a list of candidate genes to test in the context of

megakaryocyte forward programming (MK-FoP) . Briefly, from the initial set. of 116 genes coding for DNA binding proteins specifically expressed in cord-blood derived megakaryocytes (CB-MKs) compared to the H9 hESC line (hESC#l), 101 genes were retained after histone gene removal. 46 candidates were then selected upon analysis of protein-protein interaction network using VisANT software, integrating interactions with chromatin remodelling factors and level of gene expression. The visualization notably showed GATAl as a core factor for 11 interacting partners (Figure 1) ,

1 transcription factor candidates were then cloned into lentiviral vectors for experimental assessment of their MK-FoP potential (Table 1) . All 14 factors together were tested in parallel to a combination of 9 factors (higher rank factors) (Figure 1). The 9TFs combination was found to be better than the 14TFs combination (Figure 2).

The transduction of the hESC#l line with the 9 TFs concurrently generated a well-defined population of CD41a positive cells

(integrin alpha- lib, megakaryocyte marker) identified by flow cytometry at day 7 (1.710.8%) .

The CD41a+ population (9TFs forward-programmed cells) was then flow sorted at day 7 and individual transgene expression measured by RT- QPCR. Expression of all 9 TFs was detected in the CD41a negative cell population.

CD41a expressing cells showed a clear dominance in GATAl, FLU and TALI transgene expression providing indication that the combined expression of these 3 TFs was instrumental in the differentiation process (Fig.3). Indeed, the 3-TFs combination showed a better

CD41a+ cells outcome when compared with the 9-TFs or a reduced 5-TFs combination including GATAl interactors (Fig.4) The potential of the GATAl+FLII combination was further explored by systematic addition of one of the remaining 7 factors (from 9TFs combination) . The 14TFs, 9TFs and 5TFs (top ranked core network factors: GATA1+FLI1+TAL1+SPI1+ZBTB16 were tested in parallel in hESCs cells under pluripotent conditions without mesoderm induction (Figure 5) . The 3TFs combination was identified as the best one to drive forward programming. However, the GATA1+TAL1 combination was tested in a separate experiment and was not able to drive forward programming . We validated further the requirement for a concurrent expression of the 3 TFs by testing all permutations of the latter (Fig.6).

Interestingly, the combination of GATAl and FLU also generated a significant amount of CD41a+ cells under some conditions, although with a reduced efficiency providing indication that TALI might act more as a promoter than a primary inducer of forward programming.

In an initial MK-FoP protocol, hPSCs were dissociated with

Collagenase IV in order to generate small cell clumps subsequently seeded on human fibronectin coated plates (around 5E+5 cells per 10cm2) - The day after (dayO) , cell clumps were transduced by lentiviral vectors expressing t GATA1 , TALI and FLI1, using MOI20 in presence of protamine sulfate. Cells were kept in pluripotency medium (chemically defined with Activin-A and FGF2 ) for two days, then in MK medium (chemically defined with TPO and SCF) for five days .

In a modified MK-FoP protocol, hPSCs were seeded as single cells after dissociation by TrypLE and allowed to attach on fibronectin coated plates in pluripotency medium supplemented with rock

inhibitor Y-27632 (inhibition apoptosis) for 24 hours before transduction. In addition, cells were cultivated in mesoderm inducing conditions (FGF2, BMP and LY294002) for the first two days, then in MK medium (chemically defined with TPO and SCF) for five days .

Another modified MK-FoP protocol, which achieved the best cell yield, used an embryoid body culture approach in chemically defined conditions (Fig 9). On transduction day (dayO), sub- confluent (50- 80%) human pluripotent stem cells in pluripotency medium are dissociated to single cells using TrypLE (Life Technologies} for 5 min at 37°C and viable cells counted on a haemocytometer . Desired amount, of cells (e.g. 1E+6 cells) is added to Aggrewell™400 plates (Stemcell Technologies, France) in order to obtain embryoid bodies

(EB) of 300 cells each following spin aggregation (detailed protocol in Aggrewell ¹ '" ¹ Technical manual) . Importantly, lentiviral

transduction is performed concomitantly to the aggregation phase. Briefly, cells are added to the well in CDM supplemented by Y-27632 (lOuM, Sigma Aldrich) , rh-BMP4 (lOng/ml, R&D) and protamine sulfate (8ug/ml, Sigma). Concentrated lentiviral vectors individually coding for each forward programming factor are added to the well to MOI20 (multiplicity of infection) (2E+7 TU) . Subsequently, Aggrewell™ plates are centrifuged at lOOg for 3minut.es and put into the incubator (37C/5%C02) for 24 hours. Transduced EB are collected the day after and rinse twice with PBS before being seeded in ultralow adherent cell culture plates (Corning) at a density of 600 EB per 10cm2 dish in CDM plus rh-BMP4 (lOng/ml, R&D) and rh-FGF2 (5ng/ml, University of Cambridge) (i.e. mesoderm induction medium). Twenty four hours later, EB are collected, rinse with PBS and further cultivated in ultralow adherent plates in Cellgro SCGM medium

(Cellgenix, Germany) supplemented with rh-TPO (lOOng/ml, Cellgenix) and rh-SCF (25ng/ml, Life Technologies) (MK programming medium) until collection for analysis. Culture medium was changed every three days by aspirating half of the volume and adding 2 times cytokine concentrated fresh medium on top. We performed additional analyses to confirm the megakaryocyte identity of the emerging CD41a+ population. The expression of key megakaryocyte genes was measured by QPCR at day ^' 7. Forward programing of the hESC#2 line using the 3TF combination (FLU, GATA1 and TALI) was shown to induce expression of the megakaryocyte genes MPL

(coding for the thrombopoietin (TPO) receptor), ZFPM1, RUNX1 and late differentiation markers like NFE2, MEIS1 and MEF2C, as well as endogenous expression of GATA1, TALI and F'LIl (Fig 7). Moreover, functional megakaryocyte progenitors were limited to the CD41a+ population at day 7 as demonstrated by clonogenic colony forming assays (Fig 8} . Indeed, these cells were found to be able to form mature MK colonies expressing CD41a and CD42b in semi-solid collagen cultures; interestingly, CD41a- cells did not. show such potential. Importantly, key MK gene expression was shown to be restricted to the CD41+ cell population at day7 (Figure 9) .

Altogether, we identified GATA1, FLU and TALI as a minimal

combination of TFs inducing efficient megakaryocyte forward

programming from hPSCs, TALI acting as an enhancer while GATA1 and FLU were instrumental to the programming process

Time course analysis of surface marker expression by flow cytometry showed that human iPSC lines forward programmed using the 3TFs started to express CD41a from day4 and CD42a from day? (Fig 10) . By day 14, cells kept in the MK culture medium TPO+SCF have reached near homogeneity for CD41a expression (91%) and a clear CD42a+ population has developed (22%), generating on average 13+/-7 millions mature MK per million iPSC input. The morphology of dayl4 cells obtained by 3TF forward programming as observed by Romanowsky staining is very similar to in vitro cord blood derived MK regarding cell size, cytoplasm and nuclei content (Fig 11) . Mature MKs co- expressed CD41a, CD42b, GPVI, CD61 and CD42a. To improve forward programming efficiency, we tested different chemically defined in vitro culture condition settings following 3- TF transduction of various hPSC lines . We measured the effect of commitment to mesoderm - the embryonic germ lineage from which the haematopoietic system originates - when induced simultaneously to 3- TF expression. We observed that mesoderm induction by exposition of hPSCs to BMP4 , FGF2 and LY294002 for two days after viral

transduction significantly increased the number of CD41a+ cells at day 7 in a variety of cell lines (Fig.12) . Importantly, the

transduction efficiencies were similar for hPSCs maintained in piuripotent or mesoderm conditions indicating that mesoderm cells are indeed more responsive to the 3-TF driven forward programming. Additionally, we tested different hPSCs sowing techniques for their impact on forward programming at day 7. When comparing hPSCs sowed and transduced as clumps or single cells in tissue culture plates with those induced to form embryoid bodies (EB) by forced

aggregation, we observed a significant improvement of the CD41a+ cell number with the latter (Fig.13). Overall, we showed that the 3- TF driven MK-FoP efficiency was significantly improved by mesoderm commitment and cultivation as spin aggregated EB . The latter also offered the advantage of a standardised compact suspension culture and had been consequently used for the remaining experiments of this study. 2.2 A chemically defined optimised protocol generates high number of mature megakaryocytes

Using the optimised chemically defined protocol comprising viral transduction of spin aggregated EBs at day 0, culture in mesoderm medium comprising FGF2 and BMP4 and LY294002 for two days after viral transduction; culture in MK medium comprising TPO and SCF until day 10 and dissociation of embryoid bodies showing cystic structures and actively growing cell aggregates to single cells at daylO and further cultivated in MK maturation medium (TPO+ILlb) for an additional 10 days, we analysed megakaryocyte maturation of 3-TF forward programmed hiPSCs compared to cord blood derived

megakaryocytes. We observed a gradual increase of CD41a+

megakaryocyte progenitor cells followed by the progressive

acquisition of the maturation marker CD42b (glycoprotein lb, part of the MK specific GPIb/V/IX receptor complex) over cultivation time mimicking normal megakaryocyte differentiation.

Interestingly, the culture reached megakaryocyte purity (95±2% CD41a+, n=12) with more than half mature cells (56+4% CD42b+, n=12) by day 20 post-transduction without additional sorting procedure. Further maturation could be obtained by maintaining cells in culture for longer periods (>80% CD42b+) but the cell yield is reduced by higher cell death as observed in prolonged cord blood cultures. In addition to CD41a and CD42b, we also observed expression of

additional key surface proteins involved in megakaryocyte and platelet functions (itgb3 (CD61), gp6 and gp9 (CD42a) ) by flow cytometry on hiPSC forward programmed day20 cells and cord blood derived megakaryocytes. Similar expression profiles between cord blood and hiPSC derived megakaryocytes were observed.

Moreover, forward programmed cells showed typical size,

morphological and ultrastructural features of megakaryocytes. We observed numerous polyploid cells with sizes ranging from 15 to 30um and large cytoplas s (Fig.14) a d using transmission electron microscopy, cells displaying characteristic structures of maturing megakaryocytes as multi vesicular bodies - precursors of platelet granules - and developing demarcation membrane systems. Importantly, we showed that the 3-TF forward programming protocol efficiently generated mature megakaryocytes from different hiPSC lines with up to 33 fold increase in mature CD42b+ cell number compared to the hiPSCs i put (Fig, 15) .

We further characterised the megakaryocyte identity of the forward programmed cells by whole genome expression analysis. The tissue enrichment analysis between the parental hiPSCs and derived 3-TF forward programmed cells showed most, significance for platelets (P=l .03E-62) while the top 5 enriched biological processes are related to haemostasis and platelet activity confirming on a genome wide level the megakaryocyte phenotype (Table 3) . A gene set.

enrichment analysis ( (GSEA web tools, Broad Institute) on the genes differentially expressed between hiPSC-MKs and the parental

pluripotent stem cells showed a significant enrichment for

megakaryocyte specific genes (dataset generated from HaemAtlas, MK compared to other blood lineages) and confirmed further the MK identity of forward programmed cells amongst other blood cell types (NES=1.47, data not shown). Intriguingly, we observed that MK-FoP applied to two different hiPSC lines, although showing variability in. cell number outcome (Fig.15), generated highly similar expression profiles in the differentiated mature MKs (R2=0.99) .

We identified 13 TFs whose expression is missing in hiPSC-MKs compared, to CB or peripheral blood differentiated MKs (Table 2) . These genes may be determinants for post-natal phenotype acquisition and may further improve the maturation of hiPSC derived MKs .

In summary, we demonstrated that the 3-TF forward programming approach was able to efficiently generate genuine megakaryocytes sharing key features with their cord blood derived counterpart.

2.3 Forward programmed megakaryocytes produce functional plateletlike particles in vitro

We tested the ability of forward programmed megakaryocytes to generate functional platelet like particles (PLPs) in vitro.

Platelet production happens by a process of proplatelet formation by mature MKs, Immunofluorescence analysis of hiPSC forward programmed day 20 megakaryocytes cultivated for an additional 48 hours on fibrinogen coated plates showed pro-platelet like cytoplasmic protrusions in vitro showing bulbous structures expressing von Willebrand factor (vwf) and P-selectin along alpha-tubulin positive cytoplasmic filaments, von Willebrand factor (vwf) and P-selectin which are key proteins embedded in platelet granules .

The release of platelet -like particles (PLPs) in the culture supernatant ^" was highly improved by co-culture of hiPSC derived MKs on the OP9 stromal cell line for 48 hours. In such conditions, we detected significant amount of PLPs by flow cytometry identified as human platelet size particles co-expressing CD41a and CD42a. They were produced at an average of 110.3 million per million hiPSC-MK input after 48 hours of co-culture on various feeder cells (Fig.16} . e explored further the functionality of the generated PLPs by monitoring their contribution to in vitro collagen induced mouse thrombi under flow. Human particles were detected by immunostaining after clot formation on collagen fibres following mixing of PLPs or human platelets with mouse blood (3E+6 and 5E+6 per ml

respectively) . Activated platelets co-expressing CD41a and P- selectin on their surface were also identified. This shows that the hiPSC derived PLPs in formed mouse platelet, clots displayed granule content, on their surface, demonstrating functional activation.

We describe above a method for generating megakaryocytes from hPSCs using a forward programming strategy based on combined forced expression of the three transcription factors GATAl, FLU and TALI. The generated cells were genuine megakaryocytes able to release functional platelet like particles in vitro. Indeed, this method offers several advantages compared to previously described methods. First, the differentiation and maturation of megakaryocytes is achieved in chemically defined conditions without supporting stromal cells. This improves reproducibility and will greatly help the transition to clinically compatible procedures. By acting directly at the level of the gene regulatory network, the forward programming strategy allows an efficient megakaryocyte differentiation using a reduced cytokine combination and minimal cell handling. The cell yield of forward programming matches the best described approaches so far; leading in 20 days to up to 50 megakaryocytes per hiPSC input and provides a pure megakaryocyte population without the need for sorting (>95% CD41a+ and >50%

CD42a+) . An additional benefit, of the protocol described above is the use of suspension culture only (embryoid bodies followed by single cells) which greatly reduces the footprint of the experiment and should facilitate its transfer to large scale production systems . A mesoderm inducing treatment concomitant to transgene expression was found to be beneficial to forward programming, providing indication that, the epigenome and /or transcriptional profile of mesoderm cells were more amenable to respond to the programming factors. The forward programming happens very rapidly since markers of MK commitment are detected as early as four days after 3 TFs transduction, the MK potential is restricted from day 7 to the CD41a+ population and cells show an early dependency to

haematopoietic cytokines . In conclusion, our study demonstrates forward programming to generate MKs from hiPSCs. The generated hiPSC-megakaryocytes shared key features with their cord blood derived counterpart at the genetic and ultrastructural levels, and are able to release

functional platelet like particles in vitro. This novel

differentiation approach will have broad applications in both basic research and clinical development of hiPSC derived transfusion products .

Table 1 Factor Name Gene ID Reference Sequence

ABLIM1 Actin binding LIM 3983 NP 001003407.1 G1 : 51173713 protein 1

FHL1 Four and a half LIM NP 001153174.1 G1 : 228480211 domains 1

RUNX3 Runt-related 864 NP 001026850.1 GI: 72534652 transc iption factor 3

NFIC Nuclear factor I/C NP 001231931.1 GI : 350529396

(CCAAT --binding- trans cription factor)

NFIL3 Nuc1ear factor, 4783 NP 005375.2 GI: 52630429 interleukin 3 regulated

VDR Vitamin D (1,25- 7421 NP 000367.1 GI: 4507883

dihydroxyvitamin D3 )

receptor

KESPl Mesoderm posterior 1 55897 NP 061140 ,1 GI : 14149724

BTBD11 TB (POZ) domain NP 001018082.1 GI: 65786661 containing 11

APPL2 Adaptor protein, 55198 NP 060641.2 GI: 24586663 phosphotyrosine

interaction, PH domain

and leucine zipper

containing 2

M1CAL1 Microtubule associated 64780 NP 073602.3 GI: 205360947 monooxygenase, calponin

and LIM domain

containing 1

BATE' Basic leucine zipper 10538 NP 006390.1 GI: 5453563

transcription factor,

ATF-like

SCMH1 Sex comb on midleg 22955 NP 001026864.1 GI: 72534680 homolog 1

MBP Myelin basic protein 4155 NP 001020252.1 GI : 68509930

Table Top 5 enriched Biological Processes

(hiPSC-MK vs. hiPSC)

Table 3