Reprod. Fertil. Dev., 1992, 4 , 449-58
The Effect of Leukaemia Inhibitory Factor (LIF) on Embryogenesis
Richard C. Fry Victorian Institute of Animal Science, Werribee, Vic. 3030, Australia.
Abstract Leukaemia inhibitory factor (LIF) was originally identified as a haemopoetic factor that induced the differentiation of certain myeloid leukaemia cell lines. In contrast to this action, LIF was subsequently shown to inhibit the spontaneous differentiation of murine embryonic stem cells in culture, thus maintaining their pluripotency and ability to contribute to the germline of chimaeric mice. In the mouse, mRNA for LIF is expressed by the endometrial glands of the uterus coincident with the time of blastocyst implantation and receptors have been found on the preimplantation blastocyst. The signal for LIF expression appears to be of maternal origin, perhaps regulated by oestradiol. Recombinant LIF improves the development of murine and ovine blastocysts in culture although there is some species specificity with respect to the type of LIF that is bioactive. It is proposed here that LIF acts on the trophectoderm of the rapidly expanding blastocyst and improves the implantation rate of otherwise compromised embryos. Further studies in livestock should elicit therapeutic uses for LIP in embryo culture, embryo transfer and embryo survival in vivo.
Introduction Reproductive performance is a major determinant of both the rate of genetic gain and the productivity of livestock enterprises. Under natural mating conditions, about 30% of embryos die before implantation (Wilmut et al. 1986). With the advent of embryo technology, the survival of the embryo has even greater significance. In addition to the high value placed upon the embryo, most embryo manipulative processes such as embryo transfer, splitting, in vitro maturatiodin vitro fertilization of the oocyte (IVMAVF) and nuclear transplantation are detrimental to the normal development and implantation of the embryo. The low success rates of these techniques, often in the range of 10-208, are prohibitive to their commercialization (Bondioli et al. 1990). Adequate development of the embryo and the uterus, resulting in the successful establishment of pregnancy, depend on a complex interaction between the uterus and conceptus. Only through understanding these interactions at the maternal-embryo interface will new methods of increasing pregnancy rates and in vitro culture techniques become available. This short review concentrates on one recently identified molecule, leukaemia inhibitory factor (LIF), that may play a significant role in the development of the embryo. LIF: the Molecule In their search for haemopoetic factors that may control the growth of cancer, Hilton et al. (1988a) isolated three candidates secreted by Krebs I1 ascites tumour cells that induced the terminal differentiation of various leukaemic cell lines. Two of these substances were identified as granulocyte colony-stimulating factor (G-CSF) and granulocyte-macrophage 1031-3613/92/040449$05,00
R. C. Fry
colony-stimulating factor (GM-CSF), but the other was a novel factor and was named 'leukaemia inhibitory factor' (LIE). LIE is a heavily glycosylated single-chain polypeptide containing three disulfide links. It has an apparent molecular mass ranging between 32 and 67 kDa in the natural form depending on the degree of glycosylation, and reducing to about 20 kDa after removal of the carbohydrate moieties (Hilton et al. 19886; Smith et al. 1988). LIE is characterized by its ability to suppress the proliferation of the murine-derived M1 leukaemia cell line (MI) by inducing the irreversible differentiation of the cells to macrophage cells. It is an extremely potent molecule causing a noticeable reduction in cell proliferation within 6 h of exposure and obvious differentiation of the cells within 48 h. In vitro, 50% of M1 cells differentiate at a dose of g L-' (defined as 50 U mL-') with approximately 20% receptor binding (Metcalf et al. 1988). LIE is known to be secreted by a number of cell types, including various ascites, fibroblast, lymphocyte, spleen and liver cells (for review see Gough and Williams 1989). Murine, human and ovine LIE (m,h,oLIF) have been sequenced with the highest degree of homology occurring between oLIF and hLIF (88%) and the least homology between mLIF and oLIF (74%) (Gearing et al. 1988; Stahl et al. 1990; Willson et al. 1992). Both mLIF and hLIF have been expressed in E. coli and yeast host cell systems and can therefore be produced in commercial quantities (Gearing et al. 1989). Recombinant mLIF has been investigated for biological action in many cell and tissue culture systems and has a diverse range of actions, depending on the target cell. For example, in contrast to inducing differentiation of M1 myeloid leukaemia cells, mLIF will inhibit differentiation of embryonic stem (ES) cells, thus maintaining their pluripotent potential (see next section). Furthermore, mLIF has been shown to sustain the survival of murine primordial germ cells in culture (DeFelici and Dolci 1991), induce the proliferation of myoblasts in culture (Austin and Burgess 1991), potentiate the production of megakaryocytes in vitro (Metcalf et al. 1991), induce the expression of the cholinergic phenotype in sympathetic neurones in preference to noradrenergic function (Yamamori et al. 1989), have an anabolic effect on bone formation in vitro and affect bone growth and remodelling (Lorenzo et al. 1990; Reid et al. 1990), inhibit lipoprotein lipase activity in adipocytes (Mori et al. 1989), and inhibit the maturation of kidney cells (Tomida et al. 1990). The administration of near-toxic doses of recombinant mLIF to mice, either through the grafting of LIE-producing cells (Metcalf and Gearing 1989) or by repeated injection (Metcalf et al. 1990)' produced a range of biological changes consistent with these known actions in vitro. LIE is therefore a multifunctional protein with therapeutic potential in the treatment of diseases such as muscular dystrophy, thrombocytopenia, CNS diseases, osteoporosis, inflammation and atherosclerosis. Use of LIF in the Culture of Embryonic Stem Cells
The first indication that LIE could be an embryotrophic agent came from the observation made concurrently by Smith et al. (1988) and Williams et al. (1988) that LIE could be the differentiation inhibiting activity (DIA) produced by Buffalo rat liver that was used to maintain murine ES cells in culture. ES cells are the pluripotent cell lines derived from the inner cell mass (ICM) of blastocysts that can contribute to all cell lines, including the germline, when reincorporated into the murine blastocyst. The transfer of these blastocysts to recipient females allows the production of chimaeric mice that can then be crossbred to segregate the specific traits of the ES cells. Consequently, they are an invaluable tool in the study of factors that control the differentiation of early embryos and in providing a vector for the introduction of genes to produce transgenic animals (Robertson et al. 1986; Doetschman et al. 1987). Unless ES cells are cultured with feeder cells such as embryonic fibroblasts or in media conditioned with DIA, they will differentiate and die over a period of 3-6 days (Evans and Kaufman 1981; Wobus et al. 1984; Hooper et al. 1987). Highaffinity LIE receptors have been found on ES cells (Smith et al. 1988; Williams et al. 1988)
Effect of LIF on Embryogenesis
and the addition of mLIF to the culture medium will substitute for feeder cells and allow the long-term maintenance of ES cells for over 24 passages or for more than 2 months, with high repeatability (Pease and Williams 1990). It has since been confirmed that the feeder layers secrete LIF (Rathjen et al. 1990b) and that DIA is analogous to LIF (Smith et al. 1988). Media supplemented with 1000 U m ~ - 'of recombinant mLIF are generally used to isolate ES cell colonies and maintain them in an undifferentiated state, and concentrations of up to 10000 U mL-' have not proved cytotoxic (Williams et al. 1988; Gearing et al. 1989; Pease et al. 1990). It appears that LIF has no effect on the proliferation of these ICM-derived cells, as their total number remains constant over a wide range of LIF concentrations (Fig. 1). Irrespective of the divergent action of LIF on MI leukaemic and ES cells, the EDs0 is the same for the actions of LIF on both cell lines (50-100 U m ~ - ' ) , although ES cells must have continual exposure to LIF to prevent differentiation to fibroblasts (Pease et al. 1990) whereas M1 cells require only a brief exposure to LIF to be irreversibly committed to differentiation (Metcalf et al. 1988). Gough and Williams (1989) proposed that LIF may hold the genetic switch for cell differentiation in the 'off' position for ES cells while tripping it to the 'on' position for M1 cells. Whether these cell lines have similar or different receptors for LIF is unknown; however, the recent cloning of the LIF receptor (Gearing et al. 1991) should now clarify the biological mechanisms involved in its actions.
Fig. 1. Dose-response curves for the effect of mLIF [U mL-'1 on embryonic stem cell differentiation and proliferation. (Compiled from data in Williams et al. 1988, Gearing et al. 1989 and R. L. Williams, unpublished data, with permission.) Note: 50 U mL-I LIF is defined as the dose causing a 50% inhibition of differentiation in the M1 cell bioassay (Metcalf et al. 1988).
Log [LIF]
Thus, the use of purified recombinant mLIF as a direct additive to the culture medium is an important advancement in the culture and maintenance of these murine ES cells. LIF abolishes the need to establish feeder layers or to provide preconditioned media, both of which complicate the use of these ES cells for biochemical and genetic studies. Furthermore, such use of LIF should simplify and standardize the isolation of stem cells from livestock species. Attempts have been made with LIF or feeder layers to isolate lines of ES cells from the pig (Evans et al. 1990; Piedrahita et al. 1990; Strojec et al. 1990), sheep (Handyside et al. 1987; Li and Trounson 1990) and cow (Evans et al. 1990; Hassan-Hauser et al. 1990)' but confirmation through the production of chimaeric offspring is still required.
R. C. Fry
Action of LIF on Murine Embryos Following the demonstration that LIF had an action of embryo-derived cells, Bhatt et al. (1991) investigated the physiological role of LIE during embryogenesis in the mouse. Although low levels of LIF mRNA are expressed in many tissues of both the fetus and adult, high levels are expressed by the endometrial glands of the uterus specifically on the 4th day of pregnancy which coincides with the time that the blastocyst hatches from the zona pellucida and implants on the uterine wall (Finn and McLaren 1967). Either very low or undetectable levels of LIF expression occurred in postimplantation embryos, which is in broad agreement with the findings of an earlier study by Conquet and Brulet (1990). Furthermore, LIF receptors have been found on the trophectoderm of the expanded blastocyst (S. Gabrielle and L. Williams, unpublished data). These data imply that the principal function of LIF may be to regulate blastocyst growth and implantation. LIF is also expressed in D a y 4 pseudopregnant mice and, in response to an oestradiol injection, in female mice undergoing delayed implantation (Bhatt et al. 1991); this suggests that its expression is under maternal control, with the higher concentrations of circulating oestrogen that occur on Days 3 to 4 of pregnancy (McCormack and Greenwald 1974) perhaps being the signal. The timing of the expression and binding of LIF in vivo indicates that its action could be to increase the proliferation of the trophoblast cells (at the time of rapid expansion of the blastocyst) thereby improving the embryo's chance to implant and possibly also exerting some developmental control on the embryo itself. In support of this hypothesis, Lavranos and Seamark (1989) showed that although the addition of 1000 U m ~ - 'of hLIE to the culture medium had little influence on the early development of 8-cell mouse embryos through to the blastocyst stage of development, it did increase the proportion of blastocysts that hatched from the zona pellucida and attached to the surface of the culture dishes. In another study, these authors found that blastocysts cultured in media containing LIF had a larger area of trophectoderm than controls but a similar ICM area, indicating that the site of action was most likely the trophectoderm (Table 1; Robertson et al. 1990). Whether this action is through direct stimulation of cell division or prevention of cell death is unknown. Although LIF receptors have been found on ES cells (Gough and Williams 1989), labelled LIF does not bind to the ICM of whole embryos (Gabrielle and Williams, unpublished data); this suggests either that the trophectoderm provides an impermeable barrier, that expression of receptors in vivo is inhibited or that the receptors are fully occupied. It is still possible that LIF influences the ICM development through receptormediated endocytosis, as has been described for insulin (Heyner et al. 1989). In the murine studies of Conquet and Brulet (1990) and Rathjen et al. (1990a), high levels of LIF expression have been found in the extraembryonic tissues, suggesting that there may be
Table 1. The effect of hLIF on the development of mouse embryos (percentage developing) and on area (arbitrary units) of inner cell mass and trophoblast of Day-5 mouse embryos From Lavranos and Seamark (1989) and Robertson et al. (1990) with permission. *P
The Effect of Leukaemia Inhibitory Factor (LIF) on Embryogenesis
Richard C. Fry Victorian Institute of Animal Science, Werribee, Vic. 3030, Australia.
Abstract Leukaemia inhibitory factor (LIF) was originally identified as a haemopoetic factor that induced the differentiation of certain myeloid leukaemia cell lines. In contrast to this action, LIF was subsequently shown to inhibit the spontaneous differentiation of murine embryonic stem cells in culture, thus maintaining their pluripotency and ability to contribute to the germline of chimaeric mice. In the mouse, mRNA for LIF is expressed by the endometrial glands of the uterus coincident with the time of blastocyst implantation and receptors have been found on the preimplantation blastocyst. The signal for LIF expression appears to be of maternal origin, perhaps regulated by oestradiol. Recombinant LIF improves the development of murine and ovine blastocysts in culture although there is some species specificity with respect to the type of LIF that is bioactive. It is proposed here that LIF acts on the trophectoderm of the rapidly expanding blastocyst and improves the implantation rate of otherwise compromised embryos. Further studies in livestock should elicit therapeutic uses for LIP in embryo culture, embryo transfer and embryo survival in vivo.
Introduction Reproductive performance is a major determinant of both the rate of genetic gain and the productivity of livestock enterprises. Under natural mating conditions, about 30% of embryos die before implantation (Wilmut et al. 1986). With the advent of embryo technology, the survival of the embryo has even greater significance. In addition to the high value placed upon the embryo, most embryo manipulative processes such as embryo transfer, splitting, in vitro maturatiodin vitro fertilization of the oocyte (IVMAVF) and nuclear transplantation are detrimental to the normal development and implantation of the embryo. The low success rates of these techniques, often in the range of 10-208, are prohibitive to their commercialization (Bondioli et al. 1990). Adequate development of the embryo and the uterus, resulting in the successful establishment of pregnancy, depend on a complex interaction between the uterus and conceptus. Only through understanding these interactions at the maternal-embryo interface will new methods of increasing pregnancy rates and in vitro culture techniques become available. This short review concentrates on one recently identified molecule, leukaemia inhibitory factor (LIF), that may play a significant role in the development of the embryo. LIF: the Molecule In their search for haemopoetic factors that may control the growth of cancer, Hilton et al. (1988a) isolated three candidates secreted by Krebs I1 ascites tumour cells that induced the terminal differentiation of various leukaemic cell lines. Two of these substances were identified as granulocyte colony-stimulating factor (G-CSF) and granulocyte-macrophage 1031-3613/92/040449$05,00
R. C. Fry
colony-stimulating factor (GM-CSF), but the other was a novel factor and was named 'leukaemia inhibitory factor' (LIE). LIE is a heavily glycosylated single-chain polypeptide containing three disulfide links. It has an apparent molecular mass ranging between 32 and 67 kDa in the natural form depending on the degree of glycosylation, and reducing to about 20 kDa after removal of the carbohydrate moieties (Hilton et al. 19886; Smith et al. 1988). LIE is characterized by its ability to suppress the proliferation of the murine-derived M1 leukaemia cell line (MI) by inducing the irreversible differentiation of the cells to macrophage cells. It is an extremely potent molecule causing a noticeable reduction in cell proliferation within 6 h of exposure and obvious differentiation of the cells within 48 h. In vitro, 50% of M1 cells differentiate at a dose of g L-' (defined as 50 U mL-') with approximately 20% receptor binding (Metcalf et al. 1988). LIE is known to be secreted by a number of cell types, including various ascites, fibroblast, lymphocyte, spleen and liver cells (for review see Gough and Williams 1989). Murine, human and ovine LIE (m,h,oLIF) have been sequenced with the highest degree of homology occurring between oLIF and hLIF (88%) and the least homology between mLIF and oLIF (74%) (Gearing et al. 1988; Stahl et al. 1990; Willson et al. 1992). Both mLIF and hLIF have been expressed in E. coli and yeast host cell systems and can therefore be produced in commercial quantities (Gearing et al. 1989). Recombinant mLIF has been investigated for biological action in many cell and tissue culture systems and has a diverse range of actions, depending on the target cell. For example, in contrast to inducing differentiation of M1 myeloid leukaemia cells, mLIF will inhibit differentiation of embryonic stem (ES) cells, thus maintaining their pluripotent potential (see next section). Furthermore, mLIF has been shown to sustain the survival of murine primordial germ cells in culture (DeFelici and Dolci 1991), induce the proliferation of myoblasts in culture (Austin and Burgess 1991), potentiate the production of megakaryocytes in vitro (Metcalf et al. 1991), induce the expression of the cholinergic phenotype in sympathetic neurones in preference to noradrenergic function (Yamamori et al. 1989), have an anabolic effect on bone formation in vitro and affect bone growth and remodelling (Lorenzo et al. 1990; Reid et al. 1990), inhibit lipoprotein lipase activity in adipocytes (Mori et al. 1989), and inhibit the maturation of kidney cells (Tomida et al. 1990). The administration of near-toxic doses of recombinant mLIF to mice, either through the grafting of LIE-producing cells (Metcalf and Gearing 1989) or by repeated injection (Metcalf et al. 1990)' produced a range of biological changes consistent with these known actions in vitro. LIE is therefore a multifunctional protein with therapeutic potential in the treatment of diseases such as muscular dystrophy, thrombocytopenia, CNS diseases, osteoporosis, inflammation and atherosclerosis. Use of LIF in the Culture of Embryonic Stem Cells
The first indication that LIE could be an embryotrophic agent came from the observation made concurrently by Smith et al. (1988) and Williams et al. (1988) that LIE could be the differentiation inhibiting activity (DIA) produced by Buffalo rat liver that was used to maintain murine ES cells in culture. ES cells are the pluripotent cell lines derived from the inner cell mass (ICM) of blastocysts that can contribute to all cell lines, including the germline, when reincorporated into the murine blastocyst. The transfer of these blastocysts to recipient females allows the production of chimaeric mice that can then be crossbred to segregate the specific traits of the ES cells. Consequently, they are an invaluable tool in the study of factors that control the differentiation of early embryos and in providing a vector for the introduction of genes to produce transgenic animals (Robertson et al. 1986; Doetschman et al. 1987). Unless ES cells are cultured with feeder cells such as embryonic fibroblasts or in media conditioned with DIA, they will differentiate and die over a period of 3-6 days (Evans and Kaufman 1981; Wobus et al. 1984; Hooper et al. 1987). Highaffinity LIE receptors have been found on ES cells (Smith et al. 1988; Williams et al. 1988)
Effect of LIF on Embryogenesis
and the addition of mLIF to the culture medium will substitute for feeder cells and allow the long-term maintenance of ES cells for over 24 passages or for more than 2 months, with high repeatability (Pease and Williams 1990). It has since been confirmed that the feeder layers secrete LIF (Rathjen et al. 1990b) and that DIA is analogous to LIF (Smith et al. 1988). Media supplemented with 1000 U m ~ - 'of recombinant mLIF are generally used to isolate ES cell colonies and maintain them in an undifferentiated state, and concentrations of up to 10000 U mL-' have not proved cytotoxic (Williams et al. 1988; Gearing et al. 1989; Pease et al. 1990). It appears that LIF has no effect on the proliferation of these ICM-derived cells, as their total number remains constant over a wide range of LIF concentrations (Fig. 1). Irrespective of the divergent action of LIF on MI leukaemic and ES cells, the EDs0 is the same for the actions of LIF on both cell lines (50-100 U m ~ - ' ) , although ES cells must have continual exposure to LIF to prevent differentiation to fibroblasts (Pease et al. 1990) whereas M1 cells require only a brief exposure to LIF to be irreversibly committed to differentiation (Metcalf et al. 1988). Gough and Williams (1989) proposed that LIF may hold the genetic switch for cell differentiation in the 'off' position for ES cells while tripping it to the 'on' position for M1 cells. Whether these cell lines have similar or different receptors for LIF is unknown; however, the recent cloning of the LIF receptor (Gearing et al. 1991) should now clarify the biological mechanisms involved in its actions.
Fig. 1. Dose-response curves for the effect of mLIF [U mL-'1 on embryonic stem cell differentiation and proliferation. (Compiled from data in Williams et al. 1988, Gearing et al. 1989 and R. L. Williams, unpublished data, with permission.) Note: 50 U mL-I LIF is defined as the dose causing a 50% inhibition of differentiation in the M1 cell bioassay (Metcalf et al. 1988).
Log [LIF]
Thus, the use of purified recombinant mLIF as a direct additive to the culture medium is an important advancement in the culture and maintenance of these murine ES cells. LIF abolishes the need to establish feeder layers or to provide preconditioned media, both of which complicate the use of these ES cells for biochemical and genetic studies. Furthermore, such use of LIF should simplify and standardize the isolation of stem cells from livestock species. Attempts have been made with LIF or feeder layers to isolate lines of ES cells from the pig (Evans et al. 1990; Piedrahita et al. 1990; Strojec et al. 1990), sheep (Handyside et al. 1987; Li and Trounson 1990) and cow (Evans et al. 1990; Hassan-Hauser et al. 1990)' but confirmation through the production of chimaeric offspring is still required.
R. C. Fry
Action of LIF on Murine Embryos Following the demonstration that LIF had an action of embryo-derived cells, Bhatt et al. (1991) investigated the physiological role of LIE during embryogenesis in the mouse. Although low levels of LIF mRNA are expressed in many tissues of both the fetus and adult, high levels are expressed by the endometrial glands of the uterus specifically on the 4th day of pregnancy which coincides with the time that the blastocyst hatches from the zona pellucida and implants on the uterine wall (Finn and McLaren 1967). Either very low or undetectable levels of LIF expression occurred in postimplantation embryos, which is in broad agreement with the findings of an earlier study by Conquet and Brulet (1990). Furthermore, LIF receptors have been found on the trophectoderm of the expanded blastocyst (S. Gabrielle and L. Williams, unpublished data). These data imply that the principal function of LIF may be to regulate blastocyst growth and implantation. LIF is also expressed in D a y 4 pseudopregnant mice and, in response to an oestradiol injection, in female mice undergoing delayed implantation (Bhatt et al. 1991); this suggests that its expression is under maternal control, with the higher concentrations of circulating oestrogen that occur on Days 3 to 4 of pregnancy (McCormack and Greenwald 1974) perhaps being the signal. The timing of the expression and binding of LIF in vivo indicates that its action could be to increase the proliferation of the trophoblast cells (at the time of rapid expansion of the blastocyst) thereby improving the embryo's chance to implant and possibly also exerting some developmental control on the embryo itself. In support of this hypothesis, Lavranos and Seamark (1989) showed that although the addition of 1000 U m ~ - 'of hLIE to the culture medium had little influence on the early development of 8-cell mouse embryos through to the blastocyst stage of development, it did increase the proportion of blastocysts that hatched from the zona pellucida and attached to the surface of the culture dishes. In another study, these authors found that blastocysts cultured in media containing LIF had a larger area of trophectoderm than controls but a similar ICM area, indicating that the site of action was most likely the trophectoderm (Table 1; Robertson et al. 1990). Whether this action is through direct stimulation of cell division or prevention of cell death is unknown. Although LIF receptors have been found on ES cells (Gough and Williams 1989), labelled LIF does not bind to the ICM of whole embryos (Gabrielle and Williams, unpublished data); this suggests either that the trophectoderm provides an impermeable barrier, that expression of receptors in vivo is inhibited or that the receptors are fully occupied. It is still possible that LIF influences the ICM development through receptormediated endocytosis, as has been described for insulin (Heyner et al. 1989). In the murine studies of Conquet and Brulet (1990) and Rathjen et al. (1990a), high levels of LIF expression have been found in the extraembryonic tissues, suggesting that there may be
Table 1. The effect of hLIF on the development of mouse embryos (percentage developing) and on area (arbitrary units) of inner cell mass and trophoblast of Day-5 mouse embryos From Lavranos and Seamark (1989) and Robertson et al. (1990) with permission. *P
