AU706243B2 - Embryonic stem cell isolation - Google Patents

Description

WO 95/20042 PCT/GB95/00140 EMBRYONIC STEM CELL ISOLATION This invention relates to a method for isolating embryonic stem cells.

Mouse embryonic stem (ES) cells are rapidly proliferating, undifferentiated, totipotent cells derived from early embryos. They can be grown in large numbers in vitro and then reintroduced into an early embryo where they can contribute to both somatic and germ cells of a chimeric animal. ES cells are becoming widely used as a route by which genetic alterations wrought in vitro can be introduced into mice and then stably inherited.

A

particularly powerful approach is the generation of predetermined genetic modifications by gene targeting in ES cells. This makes it possible to inactivate, alter, or replace a gene of interest (Bradley et al., Bio/Technology 10 534-539 (1992)). Applications include: fine analysis of regulation of gene expression, analysis of protein structure and function, generation of animal models of human diseases and precise placement of heterologous genes in transgenic animals.

Derivation of an ES cell line from a murine embryo is essentially the release of a normally transient population of undifferentiated cells from the controlling influence of the embryonic environment. ES cells have the innate capacity to proliferate indefinitely without any form of oncogenic transformation.

The standard method of obtaining a murine ES cell line was devised by Evans et al. (Nature 292 154-156 (1981)) and independently by Martin (Proc. Nat'l. Acad. Sci. USA.

78 7634-7638 (1981)) and is described in detail by WO 95/20042 PCT/GB95/00140 Robertson, E.J. "Embryo-derived stem cells" in "Teratocarcinomas and Embryonic Stem Cells, A Practical Approach". Ed. E.J. Robertson. Pub. IRL Press Ltd, Oxford, 1987. Briefly, blastocyst stage embryos are flushed from the uterus and then cultured in medium over a layer of mitotically inactivated fibroblasts, "feeder" cells. Over a few days blastocysts attach and spread onto the feeder layer, exposing cells of the inner cell mass. The inner cell mass of some explants proliferate forming clumps of undifferentiated cells. This portion of each explant is manually isolated, disaggregated and replated onto a new feeder layer. Colonies of cells form, usually composed of several cell types. Any colonies with ES morphology are further isolated, disaggregated and replated. Repeating this through several passages can result in a homogeneous population of ES cells.

A vital part of the derivation process is the physical separation of ES cells away from differentiated tissues, e.g. trophectoderm and endoderm. These arise from differentiated cells of the explanted embryo and also from spontaneous differentiation of ES cells. Endoderm is commonly the first product of ES cell differentiation and cells of this type tend to lie closely apposed to undifferentiated ES cells. The proximity of differentiated cells induces further differentiation of ES cells.

The presence of feeder cells is necessary during the first stages of ES cell derivation to reduce spontaneous differentiation of undifferentiated cells. However, the mechanism involved is only partially defined.

Established ES cell lines can usually be grown WO 95/20042 PCT/GB95/00140 3 successfully in the absence of a feeder layer (Magin et al., Nuc. Acids. Res. 14 3795 (1992)), providing that the culture medium is supplemented with the cytokine leukaemia inhibitory factor, LIF (Smith et al., Nature 336 688-689 (1988) and Williams et al. Nature 336 684-687 (1988)). However soluble LIF alone is not usually a substitute for feeders during ES cell derivation.

The frequency with which ES lines are derived from embryos varies widely. In skilled hands ES lines can be obtained from 5-10% mouse blastocysts using strain 129Sv (Robertson, M.H. Kaufman., Bradley, A and M.J.

Evans. 1983. "Isolation, Properties and Karyotype Analysis of Pluripotential (EK) Cells from Normal and Parthenogenetic Embryos" in: "Teratocarcinoma Stem Cells", Cold Spring Harbor Conferences on Cell Proliferation 10 647-663. Eds. Martin, L. Silver and S. Strickland.

It would be highly desirable to extend the use of gene targeting in ES cells to animals with greater commercial utility than mice, particularly rats and domestic animals. Several researchers have attempted to generate ES cells from farm animals (Handyside et al., Roux's Arch. Dev. Biol 196 185-190 (1987), Notorianni et al., J. Reprod. Fertil. (Suppl.) 41 51-6 (1990), Piedrahita et al., Theriogenology 34 865-877 (1990), Piedrahita et al Theriogenology. 34 879-901 (1990) and Salto et al., Roux's Arch. Dev. Biol. 201 134-141 (1992)), but success has been limited. While ES-like cells have been derived, there are no reports that any are capable of contributing to a developing embryo to form a chimeric animal.

WO 95/20042 PCT/GB95/00140 A common difficulty in the derivation and maintenance of ruminant ES cells is the progressive differentiation of undifferentiated cells which leads to eventual extinction of the cell line. It is now proposed that the differentiation of ES cells is in large part induced by adjacent differentiated cells. Hitherto, as explained above, differentiated cells have, where possible, been manually separated from ES cells. The present invention is based on the realisation that the ablation of differentiated cells, that is to say the specific destruction of differentiated cells in situ, increases the efficiency with which ES cells may be derived from embryos.

According to a first aspect of the present invention there is provided a method for selecting embryonic stem (ES) cells from an embryo in culture, the method comprising selectively killing differentiated cells of the embryo.

The differentiated cells may be derived from the embryo as explanted into the (generally in vitro) culture or may result from the differentiation of ES cells within the embryo. The intention will generally be to selectively kill all differentiated cells.

The invention has application to the production not only of murine stem cells but also stem cells of other animals, particularly other placental mammals, such as rats (Ratus spp.), other rodents, rabbits and domesticated animals, particularly farm animals such as cattle, pigs, sheep and goats. The invention is not, however, limited to any particular species and may be found to have a very wide range of applicability across the animal kingdom.

WO 95/20042 PCT/GB95/00140 Explanted embryos will generally contain undifferentiated (ES) cells as well as differentiated cells such as trophectoderm and endoderm cells. The explanted embryo may be obtained, as outlined by Robertson (1987) loc.

cit., by abstracting an embryo from the uterus (for example at the blastocyst stage) and culturing in a suitable medium over a layer of feeder cells, which may be mitotically inactivated fibroblasts, as previously described. The explanted embryo on which the selective killing step of the invention is performed may have gone through one or more passages in which a clump of predominantly undifferentiated cells is isolated, disaggregated and replated onto a new feeder layer.

In one preferred embodiment of the invention, the differentiated cells are killed by means of a drug selection regime. The regime may be used to kill differentiated cells during early passages of the explanted embryo. Undifferentiated ES cells may express a resistance marker for a drug which is toxic to both differentiated and undifferentiated cells, so that when the colony is exposed to the drug only the differentiated cells are killed. For this embodiment, the use of a promoter specific or substantially specific for undifferentiated cells is important, and the embryo cells can be made transgenic for, or otherwise be able to express, a construct comprising such a promoter operatively coupled to a DNA sequence encoding the drug resistance marker.

There have been two published reports of genes which are strongly down-regulated on the differentiation of ES cells. These are the transcription factor Oct-3/4 (Okazawa et al., EMBO J. 10 2997-3005 (1991)) and the WO 95/20042 PCT/GB95/00140 growth factor FGF-4 (Ma et al Dev. Biol. 154 45-54 (1992)). The examples which appear below use the Oct-3/4 promoter, although the FGF-4 promoter could also be used, as could any other suitable ES-specific promoters.

The transcription factor Oct-3/4 was first identified as present in undifferentiated mouse embryonal carcinoma (EC) cells (closely related to ES cells) but not in their differentiated derivatives (Okamoto et al., Cell 60 461- 472 (1990)). Analysis of Oct-3/4 mRNA expression in mouse showed that it was restricted to undifferentiated ES and EC cells (Ben-Shusan et al., Mol. Cell. Biol. 13 891-901 (1993)), embryo inner cell mass and primordial germ cells (Rosner et al., Nature 345 686-92 (1990)).

The Oct-3/4 promoter is capable of conferring Oct-3/4 specific expression on a heterologous reporter gene in transgenic mouse embryos (Okazawa et al., EMBO J. 2997-3005 (1991)).

Several genes are available for conferring drug resistance on non mutant cells, e.g. G418 selection for the neo gene (Colbere-Garapin et al., J. Mol. Biol. 150 1-14 (1981)), hygromycin selection for the hygro gene (Santerre et al.,Gene 30 147-156 (1984)), histidinol selection for the his gene (Hartman et al., Proc. Nat'l.

Acad. Sci. 85 8047-8051 (1988)), methotrexate selection for the dhfr gene (Wigler et al., Proc. Nat'l. Acad. Sci.

USA 77 3567-3570 (1980)), aminopterin/mycophenolic acid selection for the gpt gene (Mulligan et al., Proc. Nat'l.

Acad. Sci. USA 78 2072-2076 (1981)), methionine sulphoximine selection for the glutamine synthetase (gs) gene (Hayward et al., Nucl. Acids Res. 14 999-1008 (1986)) and deoxycoformicin selection for the adenosine deaminase (ada) gene (Kaufman et al., Proc. Nat'l. Acad.

WO 95/20042 PCT/GB95/00140 7 Sci. USA 83 3136-3140 (1986)). The examples below show the use of the neo (aminoglycoside phosphotransferase) gene, although the hygro, his, dhfr, gpt, gs or ada genes could also be used, as could any other suitable drug resistance gene. It should be noted that the use of the term "gene" in this context does not imply that natural genomic DNA has to be used, although that may be preferred, cDNAs may be at least as suitable, as may "minigenes" which contain some, but not all, of the introns which may naturally be present in the gene.

In another important embodiment, a DNA sequence whose expression gives rise to cell death (for example a toxin gene) is selectively expressed in differentiated cells.

For this embodiment, the use of a promoter specific or substantially specific for differentiated cells is important, and the embryo cells can be made transgenic for, or otherwise be able to express, a construct comprising such a promoter operatively coupled to the lethal DNA sequence (for example that encoding the toxin).

One suitable promoter is that of the transforming growth factor P-2 gene, which is activated on differentiation of ES cells independent of the cell type formed (Mummery et al., Dev. Biol. 137 161-170 (1990)).

Genes encoding various toxins may be placed under the control of a promoter active in differentiated cells.

For example, the Diphtheria toxin subunit-A gene (Maxwell et al., Mol. Cell. Biol. 7 1576-1589 (1987)) or the Ricin toxin subunit-A gene (Landel et al., Genes and Dev. 2 1168-1178 (1988)) can be placed under the control of a promoter expressed only in differentiated cells. Again, WO 95/20042 PCT/GB95/00140 8 either the natural cene or a non-natural seauence encoding the toxin may be used.

Embryos which can express either a selectable marker gene under the control of a promoter expressed specifically in undifferentiated ES cells and/or a toxin gene under the control of a promoter expressed specifically in differentiated cells can be generated using standard methods. Such embryos may be transgenic, in that a heterologous DNA construct may be stably integrated in the embryonic genome, or they may simply contain non-integrated expressible DNA. ES cells may then be derived from these transgenic, injected or transfected embryos by culture under selective conditions.

Embryos suitable for use in the above method are also within the scope of the invention, according to a second aspect of which there is provided an embryo comprising expressible DNA comprising a drug resistance gene (or other DNA sequence conferring drug resistance) under the control of a promoter expressed specifically in undifferentiated ES cells and/or a DNA sequence whose expression gives rise to cell death (such as a toxin gene or other DNA sequence encoding a toxin) under the control of a promoter expressed specifically in differentiated cells. The embryo may be transgenic, in which case a transgene construct comprising the expressible DNA specified under or above may be integrated into the embryonic genome.

A transgenic embryo as in above may be prepared either de novo or may be derived from a transgenic parent. A transgenic embryo as in above would be WO 95/20042 PCT/GB95/00140 9 prepared de novo, in the absence of special measures, as expression of a toxin construct in differentiated cells would be lethal to a whole animal. Transgenic animals therefore also form a part of the present invention, according to a third aspect of which there is provided a transgenic, generally non-human, animal having integrated in its genome a transgene construct comprising a drug resistance gene (or other DNA sequence conferring drug resistance) under the control of a promoter expressed specifically in undifferentiated ES cells.

Embryos containing selectable DNA constructs as in (a) above may be obtained either by mating adult animals, at least one of which is transgenic, followed by embryo removal from the reproductive tract or by de novo transgenesis performed on the embryo. Transgenesis may be achieved, for example, by DNA microinjection or transfection. Microinjection methods involve DNA microinjection into zygotes or early cleavage stage embryos. Methods for transgenic mice are described in detail by Hogan et al., "Manipulating the Mouse Embryo: A Laboratory Manual", Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, 1986). DNA microinjection using essentially the same techniques has been successfully applied to other species, including farm animals, albeit with reduced efficiency (reviewed by Wall et al., Theriogenology 38 337-357 (1992) and Wilmut et al., Reprod. Fert. Suppl. 43 265-275 (1991)).

Transgenic animals may be identified by DNA analysis of tissue biopsies, and appropriate matings carried out to produce transgenic embryos.

It is not envisaged that such transgene constructs will necessarily be the only transgene constructs integrated WO 95/20042 PCT/GB95/00140 into the animal's genome, although in many cases they will be the only ones.

Not all embryos in accordance with the second aspect of the invention will be transgenic. An alternative is for an expression construct simply to be introduced into the embryo cell(s) by any suitable means, such as microinjection (see, for example, Burdon et al., Mol.

Reprod. and Dev. 33 436-442 (1992)). Although such a construct may not survive for long in the absence of replication sequences (and in the absence of integration) it may persist for long enough for the method of the first aspect of the invention to be practised on the embryo. In any event, the relatively short duration of a non-integrating, non-replicating construct may actually be an advantage, as no particular steps are needed to remove the construct if its continued presence is either not desired, or where its removal is obligatory. The latter case would apply to ES cells containing a toxin gene under the control of a promoter causing specific expression in differentiated cells. Such a construct is normally to be removed before ES cells can participate in embryo development.

A third possibility is that the expression construct may be present in the embryonic cells coupled to sequences that give rise to episomal replication. Episomal vectors of this type have been derived from bovine papilloma virus (Mathias et al., EMBO J. 2 1487-1492 (1983)) and adenovirus (Quantin et al., Proc. Nat'l. Acad. Sci. 89 2581-2584 (1992)). The advantage of using an episomal vector is that most embryos and their cultured derivatives will contain the selective cassette for an extended period after microinjection.

WO 95/20042 PCT/GB95/00140 11 Selection strategies for ES cell isolation depend upon expression of ES-specific salvage transgenes which preserve the ES lineage under generally lethal selection conditions or differentiated cell-specific expression of toxin transgenes. Integration and maintenance of the transgene after ES isolation is probably unimportant and may be undesirable. A short term transfection system such as lipofection (Felgner et al., Nature 337 387-388 (1989)) is therefore suited to this approach in large animals particularly where transient infection can be effected under near-physiological conditions (Brunette et al., Nucl. Acids Res. 20 1151 (1992)). Lipofection involves the spontaneous association of DNA with a liposome containing a cationic lipid, which then fuses with the cell membrane, leading to internalisation of DNA. Suitable liposome preparations are commercially available.

There are a number of advantages of lipofection as a means of transgenesis in the context of the present invention. First, as a procedure, it takes far less effort and time than microinjection; therefore, large numbers of embryos can be treated. Secondly, lipofection is a relatively benign treatment which can be applied repeatedly without toxicity. Repeated lipofection will therefore help ensure the presence of DNA in a high proportion of explanted embryos. Thirdly, selectable DNA will be present mainly as non-integrated, non-replicating molecules; it should therefore be easy to isolate nontransgenic ES cells once selection is removed.

DNA constructs useful in the preparation of embryos as described above are also within the scope of the invention, according to a fourth aspect of which there is WO 95/20042 PCT/GB95/00140 provided a DNA construct comprising a drug resistance gene (or other DNA sequence conferring drug resistance) under the control of a promoter expressed specifically in undifferentiated ES cells and/or a DNA sequence whose expression gives rise to cell death (such as a toxin gene or other DNA sequence encoding a toxin) under the control of a promoter expressed specifically in differentiated cells.

Once an ES cell line has been established the selectable marker or toxin gene construct can, if necessary or desirable, be removed. In the case of non-integrating, non-replicating DNA the removal is, in effect, automatic.

For transgenic embryos and embryos in which the expression construct is episomally replicating, more deliberate steps have to be taken. The use of site specific recombination systems to generate precisely defined deletions in cultured mammalian cells has recently been demonstrated. Gu et al. (Cell 73 1155-1164 (1993)) describe how a deletion in the immunoglobulin switch region in mouse ES cells was generated between two copies of the bacteriophage P1 loxP site by transient expression of the Cre site-specific recombinase, leaving a single loxP site. Similarly, yeast FLP recombinase has been used to precisely delete a selectable marker defined by recombinase target sites in mouse erythroleukemia cells (Fiering et al., Proc. Nat'l. Acad. Sci. USA 8469-8473 (1993)). The Cre lox system is exemplified below, but other site-specific recombinase systems could be used.

A construct used in the Cre lox system will usually have the following three functional elements: WO 95/20042 PCT/GB95/00140 13 1. The expression cassette; 2. A negative selectable marker Herpes simplex virus thymidine kinase (TK) gene) expressed under the control of a ubiquitously expressed promoter phosphoglycerate kinase (Soriano et al., Cell 64 693-702 (1991)); and 3. Two copies of the bacteriophage P1 site specific recombination site loxP (Baubonis et al., Nuc. Acids. Res. 21 2025-2029 (1993)) located at either end of the DNA fragment.

This construct can be eliminated from established ES cell lines containing it by means of site specific recombination between the two loxP sites mediated by Cre recombinase protein which can be introduced into the cells by lipofection (Baubonis et al., Nuc. Acids Res. 21 2025-2029 (1993)). Cells which have deleted DNA between the two loxP sites are selected for loss of the TK gene (or other negative selectable marker) by growth in medium containing the appropriate drug (ganciclovir in the case of TK).

Embryonic stem cells isolated by the selection procedure can be tested for totipotency by assessing their capacity to form adult tissues, most importantly germ cells.

Totipotent embryonic stem cells can be used as a means of manipulating an animal's genome either by simple introduction of a transgene or by the more extensive modifications possible using gene targeting technology.

Preferred features of each aspect of the invention are as for each other aspect mutatis mutandis.

WO 95/20042 PCT/GB95/00140 14 The invention will now be illustrated by the following examples. The examples refer to the drawings, in which: FIGURES 1A to 1F show representative examples of embryoid bodies obtained as described in Example 2; and FIGURE 2 shows the structure of the OctneoTL construct.

EXAMPLE 1 Generation of Octneo DNA Construct A 1.94 kb fragment containing the 5'-flank and promoter region of the Oct-3/4 gene was amplified from mouse genomic DNA by the polymerase chain reaction. The sequence of primers used corresponded to the Oct-3/4 published sequence (Okazawa et al., EMBO J. 10 2997-3005 (1991)) with additional restriction enzyme cleavage sites to facilitate cloning and construction. The Oct-3/4 fragment was used to generate a construct, designated Octneo, designed to confer resistance to the drug G418 on undifferentiated ES cells.

Octneo was constructed from three DNA fragments: 1. A 1.94 kb Oct-3/4 5' flanking fragment (1940bp to 34bp upstream of the translation initiation site) with 5' SalI and 3' HindIII sites derived from the PCR primers; 2. A 1.7 kb HindIII/EcoRI fragment containing a modified aminoglycoside phosphotransferase (neo) gene and human growth hormone polyadenylation signal (Selfridge et al., Som. Cell and Mol. Gen. 18 325- 336 (1992)); and 3. pUC18 cut with EcoRI and Sall.

WO 95/20042 PCT/GB95/00140 EXAMPLE 2 ES Cell Specific Expression of Octneo Gene The tissue specificity of Oct 3/4-directed neo expression was investigated in two experiments: 2.1 Octneo expression in ES cells and fibroblasts.

Octneo DNA was transfected into NIH 3T3 differentiated fibroblasts and HM-1 undifferentiated ES cells (Magin et al., Nuc. Acids. Res. 14 3795 (1992)). A control construct, designated MTneo, in which neo expression is directed by the metallothionein promoter, was transfected in parallel. Colony numbers observed after G418 selection are shown in Table 1: Table 1.

Cell line Octneo MTneo NIH3T3 0 31 HM-1 11 9 This indicates that the Oct 3/4 promoter is active in ES cells and inactive in fibroblasts.

2.2 Effects of Octneo expression on in vitro differentiation of ES cells under G418 selection Octneo and MTneo constructs were transfected by electroporation into HM-1 ES cells and G418-resistant clones isolated. A suspension culture method of inducing differentiation of ES cells was then used. The method was very similar to that described by Robertson, E.J.

"Embryo-Derived Stem Cells" in: "Teratocarcinomas and Embryonic Stem Cells, A Practical Approach", Ed. E.J.

Robertson, Pub. IRL Press Ltd., Oxford, 1987. Such differences from the Robertson method that there were are detailed in the following description.

WO 95/20042 PCT/GB95/00140 16 Differentiation of transfected clones under selective conditions was investigated as now described.

Day 0: Two Octneo clones (Octneo 2,4) and one MTneo clone (MTneo 6) were grown as undifferentiated ES cell monolayers on gelatinised tissue culture plastic in ES culture medium as described by Selfridge et al. (Som.

Cell and Mol. Gen. 18 325-336 (1992)). G418 was also added at 0.2mg/ml. This concentration of G418 is double that necessary to kill untransfected HM-1 cells.

Day 1: ES cell monolayers passaged onto non-gelatinised plastic in ES culture medium to induce aggregation of cells.

Day 2: ES cells passaged as multicellular aggregates into suspension culture in bacterial petri dishes. The culture medium used was ES medium with reduced serum foetal, 2.5% newborn serum) and LIF omitted. After a few days under these conditions aggregates of ES cells tend to form embryoid bodies composed of an internal mass of undifferentiated cells and an outer layer of endoderm.

Cells from each clone were divided into four groups each with different G418 concentrations in the culture medium: Omg/ml, 0.5mg/ml, 1.0mg/ml, Day 8: Embryoid bodies were induced to attach to a substrate by transfer to gelatinised tissue culture dishes. This was to allow assessment of the extent of differentiation which had occurred in suspension culture.

However, reattachment itself provides a stimulus to differentiate and so G418 selection was maintained for a further week to allow newly differentiated cells to be WO 95/20042 PCT/GB95/00140 17 killed. During this time growth was in ES culture medium including LIF. The presence of LIF does not affect the viability of differentiated cells.

Day 15: Attached embryoid bodies were fixed with methanol/acetic acid and stained with crystal violet.

Differentiation was scored by the presence of a halo of differentiated cells, predominantly endoderm, spreading out onto the substrate.

Results are summarised in Table 2 and representative examples of embryoid bodies are shown in Figures 1A to 1F.

Table 2. Differentiation of embryoid bodies under selective conditions.

G418 concn.

Clone Omg/ml 0.5mg/ml Img/ml MTneo 6 Octneo 2 Octneo 4 Extensive differentiation.

Little or no differentiation.

These results show that differentiated ES derivatives of Octneo transfected clones are killed at G418 concentrations of 0.5mg/ml and Img/ml, while differentiated derivatives of the MTneo control survive.

At higher levels of G418 a more general toxicity in both differentiated and undifferentiated cell types was observed.

WO95/20042 PCT/GB95/00140 18 EXAMPLE 3 ES cell derivation from Octneo transgenic murine embryos The effect of drug selection on the efficiency of ES cell derivation was investigated. Octneo DNA was microinjected into mouse zygotes and 10 transgenic mouse lines (C57B1 CBA hybrid) were generated by standard techniques (Hogan et al., loc cit).

Transgenic male offspring of founder animals from each transgenic line were mated naturally with strain 129SV females. Blastocysts were collected and explanted intact onto feeder layers using the procedure described by Robertson (loc cit). Feeder layers were mitomycin -C inactivated STO-neo cells, which are an established fibroblast line transfected with a metallothionein-neo construct conferring resistance to G418. Each flush (ie the contents of a single uterus) was divided in a ratio of 1:2 into non-selective and selective treatments.

Selection was carried out with 100 gg/ml G418 included in the culture medium. This concentration had been previously established as the minimum necessary to kill non-transgenic embryos within 10 days. Primary embryo explants were cultured as pools of 2-5 embryos in single wells of 48 well plates. The culture medium consisted of BHK21 Glasgow MEM, supplemented with 10% newborn and fetal serum and murine leukaemia inhibitory factor. In all other respects ES cell isolation proceeded as described by Robertson (loc cit).

Embryos from 5 of the transgenic mouse lines all died under selective conditions as shown below: WO 95/20042 PCT/GB95/00140 Lines Embryos Explants ES lines Explants ES lines S (+G418) (+G418) (-G418) (-G418) 72-92 19 13 0 6 0 72-28 13 9 0 4 0 71-12 23 16 0 7 0 71-10 0 0 0 0 0 71-19 30 20 0 10 0 Total 85 58 0 27 0 However, embryos from the other 5 lines gave rise to G418 resistant ES cell lines. These data are shown below: Line Embryos Explants ES lines Explants ES lines (+G418) (+G418) (-G418) (-G418) 72-9 25 17 4 8 0 72-49 17 10 1 7 2 71-2 12 10 2 2 0 71-11 43 26 2 17 0 72-10 20 11 1 9 2 Total 117 74 10 43 4 As the stud males used were hemizygous for the Octneo transgene, half of the explanted embryos did not contain the Octneo transgene and so could not survive G418 selection. When comparing the rate of ES cell derivation under selective and non selective conditions, it is therefore necessary to double the efficiency of ES cell derivation under selection to take account of this mortality.

Thus: ES derivation efficiency of in the absence of G418: 9% Non adjusted efficiency in the presence of G418: 13.5%, True efficiency, accounting for death of non-transgenic embryos: 27% Statistical significance (by unrelated t test), p< 0.05.

WO 95/20042 PCT/GB95/00140 These results indicate that G418 selection of Octneo transgenic embryos significantly improves the efficiency of ES cell line derivation.

EXAMPLE 4 Derivation of ES Cells which do not Contain Octneo DNA In many cases it is desirable that ES cells do not contain experimentally introduced DNA sequences.

A

modified drug selection construct can be used to establish an ES cell line and may then be removed from the host genome. This example uses the bacteriophage P1 site-specific recombinase Cre and its target site loxP.

The construct used to exemplify this is designated OctneoTL and is composed, as is shown in Figure 2, of three functional elements: 1. The drug selection cassette (Octneo); 2. A negative selectable marker Herpes simplex virus thymidine kinase (TK) gene) expressed under the control of a ubiquitously expressed promoter phosphoglycerate kinase (Soriano et al., Cell 64 693-702 (1991)); and 3. Two copies of the bacteriophage P1 site specific recombination site loxP (Baubonis et al., Nuc. Acids. Res. 21 2025-2029 (1993)) located at either end of the DNA fragment.

A strain of Octneo TL transgenic animals are generated as in Example 3 and ES cells are derived from transgenic embryos by culture with G418 selection. Established

ES

cell lines can be rendered non-transgenic by means of site specific recombination between the two loxP sites WO 95/20042 PCT/GB95/00140 21 mediated by Cre recombinase protein which can be introduced into the cells by lipofection (Baubonis et al., Nuc. Acids Res. 21 2025-2029 (1993)). Cells which have deleted DNA between the two loxP sites are selected for loss of the TK gene by growth in medium containing the drug ganciclovir.

EXAMPLE 5 Derivation of ES Cells from Large Animal Species The generation of a transgenic strain for the purpose of ES cell derivation is impractical in domestic livestock such as cattle because of the lengthy generation time.

In this case, the selectable construct is microinjected or transfected into embryos which are then placed directly into culture and ES cells derived.

The selective cassette can be introduced into livestock embryos as a DNA fragment as described in Examples 3 or 4, or alternatively in a vector designed to replicate independently of the host genome. Episomal vectors of this type have been derived from bovine papilloma virus (Mathias et al., EMBO J. 2 1487-1492 (1983)) and adenovirus (Quantin et al., Proc. Nat'l. Acad. Sci. 89 2581-2584 (1992)). The advantage of using an episomal vector is that most embryos and their cultured derivatives will contain the selective cassette for an extended period after microinjection. In contrast, only 10-15% of embryos microinjected with a conventional DNA fragment will contain stably integrated DNA.

Unintegrated DNA persists in other embryos for approximately one week after injection.

Removal of the episome from established lines can be achieved by ganciclovir selection against the TK marker gene.

Claims (18)

1. A method for selecting embryonic stem (ES) cells from an embryo in culture, wherein the embryo cells contain a construct comprising a promoter specific or substantially specific for undifferentiated cells operatively coupled to a DNA sequence encoding a drug resistance marker, and/or a construct comprising a promoter specific or substantially S 10 specific for differentiated cells operatively coupled to a DNA sequence whose expression gives rise to cell death, S0the method comprising selectively killing differentiated cells of the embryo, ,000 •wherein the construct is subsequently removed from embryo cells by use of a 006 •site-specific recombination system. ooo•

2. A method as claimed in claim 1, wherein the promoter specific or substantially specific for undifferentiated cells is the Oct-3/4 promoter. 0:0: 3. A method as claimed in claim 1, wherein the promoter specific or 20 substantially specific for undifferentiated cells is the FGF-4 promoter. 0. 0

4. A method as claimed in claim 1, 2 or 3, wherein the DNA sequence 009o •whose expression gives rise to cell death encodes a toxin. 000 *0 0 •0 25 5. A method as claimed in claim 4, wherein the toxin is diphtheria toxin or ricin toxin. 0 0

6. A method as claimed in any of claims 1 to 5, wherein the embryo is transgenic for the construct.

7. A method as claimed in any of claims 1 to 5, wherein the construct is capable of episomal replication within the embryo.

8. A method as claimed in any of claims 1 to 5, wherein the construct is present within the embryo in non-integrated and non-replicating form. 23

9. A method as claimed in any of claims 1 to 8, wherein the site specific recombination system is the Crelox system or the FLP system. A method as claimed in any of claims 1 to 9, wherein the removal of the construct is selected for by means of a negatively selectable marker, which may be under the control of constitutive promoter.

11. A method as claimed in claim 10, wherein the negatively selectable marker is the Herpes Simplex Virus thymidine kinase (TK) gene.

12. A method as claimed in claim 10 or 11, wherein the constitutive promoter is the phosphoglycerate kinase promoter.

13. A transgene construct suitable for integration into the genome of an e 15 embryo cell, comprising a DNA sequence conferring drug resistance under the control of a promoter expressed specifically in undifferentiated ES cells, and/or a DNA sequence whose expression gives rise to cell death under 0. 20 the control of a promoter expressed specifically in differentiated cells, :i"which transgene construct also comprises a site-specific recombinase system which enables the removal of the DNA construct from the genome. 25 14. A construct as claimed in claim 13, wherein the site-specific recombinase system is the Crelox system or the FLP system. 0 A construct as claimed in claim 13 or 14, which site-specific recombinase system also comprises a negatively selectable marker, which may be under the control of constitutive promoter.

16. A construct as claimed in claim 15, wherein the negatively selectable marker is the Herpes Simplex Virus thymidine kinase (TK) gene.

17. A construct as claimed in claim 15 or 16, wherein the constitutive promoter is the phosphoglycerate kinase promoter.

18. A non-human animal embryo comprising expressible DNA comprising a DNA sequence conferring drug resistance under the control of a promoter expressed specifically in undifferentiated ES cells, and/or a DNA sequence whose expression gives rise to cell death under the control of a promoter expressed specifically in 10 differentiated cells and (ii) a site-specific recombinase system which enables the removal of the expressible DNA from the genome of the embryo. S 15 19. An embryo as claimed in claim 18, wherein the site-specific recombinase system is the Crelox system or the FLP system. An embryo as claimed in claim 18 or 19, which site-specific recombinase system also comprises a negatively selectable marker, which 20 may be under the control of constitutive promoter.

21. An embryo as claimed in claim 20, wherein the negatively selectable marker is the Herpes Simplex Virus thymidine kinase (TK) gene. 0* *S 25 22. An embryo as claimed in claim 20 or 21, wherein the constitutive promoter is the phosphoglycerate kinase promoter.

23. An embryo as claimed in any of claims 18 to 22, wherein the embryo is transgenic for the expressible DNA.

24. An embryo as claimed in any of claims 18 to 22, wherein the expressible DNA is capable of episomal replication within the embryo. An embryo as claimed in claim 24, wherein the expressible DNA is present on a bovine papilloma virus expression vector.

26. An embryo as claimed in claim 24, wherein the expressible DNA is present on an adenovirus expression vector.

27. An embryo as claimed in claim 18, wherein the expressible DNA is present within the embryo in non-integrated and non-replicating form. Dated this twelfth day of April 1999 PPL THERAPEUTICS (SCOTLAND) LIMITED Patent Attorneys for the Applicant: FBRICE CO 9* 90 C@ 0