Leukemia inhibitory factor is expressed by the preimplantation uterus and selectively blocks primitive ectoderm formation in vitro

Tissue-specific and ontogenetic regulation of LIF protein levels determined by quantitative enzyme immunoassay

To define the physiological role of leukemia inhibitory factor (LIF), it is essential to localize sites of LIF synthesis in vivo. We generated polyclonal antibodies specific for native rat LIF, and developed a two-site immunoassay to detect 10 pg LIF/ml. Using this immunoassay, we determined LIF content of 18 organs, CNS regions, and ganglia throughout postnatal development of rats. High levels of LIF protein (1.0-11.0 ng/g tissue) are present in relatively few tissues: the uterus at late proestrus to estrus and on day 5 of pregnancy, ovary at estrus to early metestrus-1, footpads during early postnatal development and thymus throughout. Intermediate levels (0.5-1.0 ng) are detected in the gut, skin, skeletal muscle, pancreas and lung at one or more postnatal ages. Low levels (0.1-0.5 ng) are observed in most other non-nervous and nervous tissues. LIF protein levels do not completely correspond to reported LIF mRNA levels.

Self renewal of embryonic stem cells in the absence of feeder cells and exogenous leukaemia inhibitory factor

To evaluate the role of leukaemia inhibitory factor (LIF) for maintaining pluripotent embryonic stem (ES) cells in culture, we established several exogenous LIF-independent ES cell lines by continuous passaging in culture. The newly established ES cells, Kli and CBli, sustained their growth and remained undifferentiated in LIF-deficient medium. Analysis of chimaeric animals, produced with the beta-galactosidase transgenic Kli ES cells, revealed that LIF-independent ES cells can contribute to all embryonic germ layers. There was no detectable LIF protein in ES cell conditioned medium, and no upregulation of LIF mRNA was found. The addition of neutralising anti-LIF antibodies was not sufficient to abrogate the self renewal of the Kli ES cells. These studies suggest that the signalling pathway involving diffusible LIF can be bypassed for maintaining the pluripotency in culture, and indicate a considerable heterogeneity in growth factor dependence and differentiation of different ES cells.

Regulation of protein abundance in pluripotent cells undergoing commitment to the neural lineage

The P19 cell line is a widely studied model of neural differentiation When pluripotent P19 cells are cultured as aggregates in the presence of retinoic acid for 4 days, the cells commit to the neural fate, but have not yet undergone overt differentiation. Two-dimensional polyacrylamide gel electrophoresis was used to analyze cellular protein expression during this induction. Approximately 500 abundant polypeptides were analyzed. Seventeen polypeptides were upregulated during induction; several of these were significantly regulated 48 h after the addition of retinoic acid. No downregulations were observed. Fifteen of the 17 polypeptides continued to be expressed throughout terminal differentiation. The upregulation of 14 of the 17 polypeptides requires both retinoic acid and aggregation, which alone do not induce neural differentiation. Furthermore, these regulated polypeptides are expressed in neural tissue, suggesting they are associated with neural function in vivo. Embryonic stem cells, a totipotent line, also neurally differentiate in response to retinoic acid and aggregation. Comparison of embryonic stem cells to P19 cells shows that the two systems regulate a similar set of polypeptides and are thus likely to utilize a similar pathway. These studies are a step toward determining the full extent of regulation involved in the commitment of pluripotent cells to the neural fate.

Maintenance of pluripotency in mouse embryonic stem cells persistently infected with murine coronavirus

A persistently coronavirus-infected embryonic stem (ES) cell line A3/MHV was established by infecting an ES cell line, A3-1, with mouse hepatitis virus type-2. Although almost all A3/MHV cells were found infected, both A3/MHV and A3-1 cells expressed comparable levels of cell surface differentiation markers. In addition, A3/MHV cells retained the ability to form embryoid bodies. These results suggest that persistent coronavirus infection does not affect the differentiation of ES cells.

Leukemia inhibitory factor (LIF) and LIF receptor expression in human endometrium suggests a potential autocrine/paracrine function in regulating embryo implantation

The uterine expression of leukemia inhibitory factor (LIF) is essential for embryo implantation in the mouse. Here, we describe the expression of LIF, related members of this group of cytokines, oncostatin M and ciliary neurotrophic factor, and the LIF receptor beta and glycoprotein gp130 in normal human tissues and in the endometrium of fertile women. Our results show that LIF is the only one of these factors expressed at detectable levels in the endometrium of women of proven fertility. LIF expression is restricted to the endometrial glands during the secretory/postovulatory phase but is not present in the endometrium during the proliferative/preovulatory phase. The LIF receptor beta is expressed during the proliferative and secretory phases of the cycle and is restricted to the luminal epithelium. The associated signal-transducing component of the LIF receptor, gp130, is also expressed in both the luminal and glandular epithelium throughout the cycle. These results suggest that uterine expression of LIF in humans, like mice, may have a role in regulating embryo implantation, possibly through an autocrine/paracrine interaction between LIF and its receptor at the luminal epithelium.

Differential hormonal regulation of leukemia inhibitory factor (LIF) in rabbit and mouse uterus

Leukemia inhibitory factor (LIF) has been shown to be essential for the implantation of mouse blastocysts. The present study was designed to determine how LIF protein was hormonally regulated in rabbit and mouse uterus using immunohistochemistry. In unmated rabbits, LIF protein was at a low level in the uterine epithelium and glands, and up-regulated by progesterone alone or estradiol-17 beta and progesterone combined. Estradiol-17 beta alone had no apparent effect. In ovariectomized mice, the level of LIF protein was very low in the uterine epithelium and glands, and was up-regulated by estradiol-17 beta alone or estradiol-17 beta and progesterone combined. Progesterone alone had no apparent effect. These results suggest that LIF protein is differentially regulated in rabbit and mouse uterus by progesterone and estrogen, respectively. This would explain the high level of LIF protein observed in uterine epithelium and glands prior to blastocyst implantation in the two species with different hormonal requirements for implantation.

Leukemia inhibitory factor, LIF receptor, and gp130 in the mouse uterus during early pregnancy

Leukemia inhibitory factor (LIF) has been shown to play an important role in the implantation of mouse blastocysts. The present study was designed to investigate the changes of LIF protein, LIF receptor, and gp130 in the mouse uterus during the early pregnancy. LIF protein and LIF receptor were at high levels in the mouse uterus near the time of ovulation and on day 4 of pregnancy. gp130 was highest on days 3 and 4 of pregnancy. Both LIF receptor and gp130 showed strong staining in the stroma of the day 5 uterus, at a time when LIF protein was low. The presence of LIF receptor and gp130 in the luminal epithelium on day 4 and in the stroma on day 5 may indicate the site of the high affinity LIF receptor. The coexistence of a high level of LIF protein, LIF receptor, and gp130 in the day 4 uterus is consistent with the previously observed high level of uterine LIF mRNA on the same day and the importance of LIF for the blastocyst implantation in mouse.

Expression of c-MYC under the control of GATA-1 regulatory sequences causes erythroleukemia in transgenic mice

To study oncogenesis in the erythroid lineage, we have generated transgenic mice carrying the human c-MYC proto-oncogene under the control of mouse GATA-1 regulatory sequences. Six transgenic lines expressed the transgene and displayed a clear oncogenic phenotype. Of these, five developed an early onset, rapidly progressive erythroleukemia that resulted in death of the founder animals 30-50 d after birth. Transgenic progeny of the sixth founder, while also expressing the transgene, remained asymptomatic for more than 8 mo, whereupon members of this line began to develop late onset erythroleukemia. The primary leukemic cells were transplantable into nude mice and syngeneic hosts. Cell lines were established from five of the six leukemic animals and these lines, designated erythroleukemia/c-MYC (EMY), displayed proerythroblast morphology and expressed markers characteristic of the erythroid lineage, including the erythropoietin receptor and beta-globin. Moreover, they also manifested a limited potential to differentiate in response to erythropoietin. Studies in the surviving transgenic line indicated that, contrary to our expectations, the transgene was not expressed in the mast cell lineage. That, coupled with the exclusive occurrence of erythroleukemia in all the transgenic lines, suggests that the GATA-1 promoter construct we have used includes regulatory sequences necessary for in vivo erythroid expression only. Additional sequences would appear to be required for expression in mast cells. Further, our results show that c-MYC can efficiently transform erythroid precursors if expressed at a vulnerable stage of their development.

Sexually dimorphic sterility phenotypes in Hoxa10-deficient mice

The Abdominal B (AbdB) genes constitute a distinct subfamily of homeobox genes that exhibit posterior domains of expression, including the genital imaginal disc in Drosophila and the developing urogenital system in vertebrates. We have mutated the AbdB gene Hoxa10 in mice. We report here that homozygotes are fully viable and show an anterior homeotic transformation of lumbar vertebrae. All male homozygotes manifest bilateral cryptorchidism resulting in severe defects in spermatogenesis and increasing sterility with age. Female homozygotes ovulate normally, but about 80% are sterile because of death of embryos between days 2.5 and 3.5 post coitum. This coincides spatially and temporally with expression of maternal Hoxa10 in distal oviductal and uterine epithelium. These results indicate a role for AbdB Hox genes in male and female fertility and suggest that maternal Hoxa10 is required to regulate the expression of a factor that affects the viability of preimplantation embryos.

Mat-8, a novel phospholemman-like protein expressed in human breast tumors, induces a chloride conductance in Xenopus oocytes

We recently identified a novel 8-kDa transmembrane protein, Mat-8, that is expressed in a subset of murine breast tumors. We have now cloned a cDNA encoding the human version of Mat-8 and show that it is expressed both in primary human breast tumors and in human breast tumor cell lines. The extracellular and transmembrane domains of Mat-8 are homologous to those of phospholemman (PLM), the major plasmalemmal substrate for cAMP-dependent protein kinase and protein kinase C in several different tissues. PLM, which induces chloride currents when expressed in Xenopus oocytes, contains consensus phosphorylation sites for both cAMP-dependent protein kinase A and protein kinase C in its cytoplasmic domain. In contrast, the cytoplasmic domain of Mat-8 contains no such consensus phosphorylation sites and is, in fact, unrelated to the cytoplasmic domain of PLM. RNA blot analysis reveals that Mat-8 and PLM exhibit distinct tissue-specific patterns of expression. We show that expression of Mat-8 in Xenopus oocytes induces hyperpolarization-activated chloride currents similar to those induced by PLM expression. These findings suggest that Mat-8 and PLM, the products of distinct genes, are related proteins that serve as Cl- channels or Cl- channel regulators but have different roles in cell and organ physiology.

Fas(CD95)/FasL interactions required for programmed cell death after T-cell activation

Receptor crosslinking of T-cell hybridomas induces cell activation followed by apoptosis. This activation-induced cell death requires de novo synthesis of RNA and proteins, but the actual gene products that provide the death signal have not been identified. We show here that receptor crosslinking induces Fas ligand and upregulates Fas, and that the ensuing engagement of Fas by Fas ligand activates the cell-death programme. Cell death, but not activation, can be selectively prevented by a soluble Fas-immunoglobulin fusion protein. Thus, Fas and Fas ligand are the death-gene products, and their interaction accounts for the molecular mechanism of activation-induced T-cell death.

Leukemia inhibitory factor (LIF) inhibits angiogenesis in vitro

Using an in vitro model in which endothelial cells can be induced to invade a three-dimensional collagen gel to form capillary-like tubular structures, we demonstrate that leukemia inhibitory factor (LIF) inhibits angiogenesis in vitro. The inhibitory effect was observed on both bovine aortic endothelial (BAE) and bovine microvascular endothelial (BME) cell, and occurred irrespective of the angiogenic stimulus, which included basic fibroblast growth factor (bFGF), vascular endothelial growth factor (VEGF), the synergistic effect of the two in combination, or the tumor promoter phorbol myristate acetate. LIF inhibited bFGF- and VEGF-induced proliferation in BAE and BME cells. In addition, LIF inhibited BAE but not BME cell migration in a conventional two-dimensional assay. Finally, LIF decreased the proteolytic activity of BAE and BME cells and increased their expression of plasminogen activator inhibitor-1. These results demonstrate that LIF inhibits angiogenesis in vitro, an effect that can be correlated with a LIF-mediated decrease in endothelial cell proliferation, migration and extracellular proteolysis.

Evidence for involvement of activin A and bone morphogenetic protein 4 in mammalian mesoderm and hematopoietic development

Xenopus in vitro studies have implicated both transforming growth factor beta (TGF-beta) and fibroblast growth factor (FGF) families in mesoderm induction. Although members of both families are present during mouse mesoderm formation, there is little evidence for their functional role in mesoderm induction. We show that mouse embryonic stem cells, which resemble primitive ectoderm, can differentiate to mesoderm in vitro in a chemically defined medium (CDM) in the absence of fetal bovine serum. In CDM, this differentiation is responsive to TGF-beta family members in a concentration-dependent manner, with activin A mediating the formation of dorsoanterior-like mesoderm and bone morphogenetic protein 4 mediating the formation of ventral mesoderm, including hematopoietic precursors. These effects are not observed in CDM alone or when TGF-beta 1, -beta 2, or -beta 3, acid FGF, or basic FGF is added individually to CDM. In vivo, at day 6.5 of mouse development, activin beta A RNA is detectable in the decidua and bone morphogenetic protein 4 RNA is detectable in the egg cylinder. Together, our data strongly implicate the TGF-beta family in mammalian mesoderm development and hematopoietic cell formation.

Leukaemia inhibitory factor and the regulation of pre-implantation development of the mammalian embryo

This paper reviews the evidence that certain growth factors, particularly leukaemia inhibitory factor (LIF), play a crucial role in regulating the development of the pre-implantation mammalian embryo. LIF was originally implicated in regulating the early development of the mouse embryo because it inhibited the differentiation of embryonic stem (ES) cells, pluripotential cells derived from the inner cell mass of the blastocyst. Subsequent studies on its role in vivo revealed, surprisingly, that it is essential for the growth rather than the differentiation of the blastocyst. In vivo, overtly normal blastocysts can be produced in a LIF-deficient environment that are capable of forming viable fertile adults. However, in the absence of LIF, they fail to implant and enter into a state resembling that exhibited by blastocysts undergoing delayed implantation, which is characterized by a cessation of cell proliferation. This failure to implant occurs because the principle sites of LIF production are the endometrial glands of the uterus. These synthesize and secrete LIF at implantation, with LIF synthesis essential for implantation. Preliminary evidence indicates that LIF synthesis is required both by the uterus for it to undergo decidualization and by the blastocyst for implantation. These data indicate that the maternal environment plays a crucial role in the development and growth of the pre-implantation embryo, by supplying factors that regulate these processes in the embryo.

Identification of two elements involved in regulating expression of the murine leukaemia inhibitory factor gene

Mouse leukaemia inhibitory factor (LIF) is a polyfunctional cytokine which exhibits multiple functions in vitro and in vivo. Two forms of LIF cDNA, differing at their 5' ends, have been described encoding either diffusible (D-LIF) or matrix-associated (M-LIF) forms of the protein [Rathjen, Toth, Willis, Heath and Smith (1990) (Cell 62, 1105-1114]. The present report describes the DNA sequence and functional characterization of the murine LIF gene and its surrounding transcriptional regulatory elements. Transient transfection of constructs containing the LIF gene and various amounts of 5'-non-coding sequence failed to give detectable levels of expression, suggesting the presence of inhibitory sequences within the LIF gene. Stable cell lines were produced by transfection of experimental constructs containing various lengths of 5'-non-coding sequence of the LIF gene, or the heterologous phosphoglycerate kinase promoter, linked to an LIF/neomycin-resistance-hybrid-coding sequence. The frequency of recovery of stable clones indicated that sequences located in the first intron between the transcriptional start sites for D-LIF and M-LIF act to suppress expression of the gene in most genomic locations. This region is rich in GC residues and has been shown to be hypomethylated in vitro [Kaspar, Dvorak and Bartunek (1993) FEBS Lett. 319, 159-162]. Analysis of the LIF/neomycin-resistance transgene expression in these stable cell clones demonstrated that transcripts containing the M-LIF or D-LIF exons required the presence of sequences located between -1200 and -3200 in the LIF gene. In the absence of these sequences, transcription is initiated elsewhere within the first intron. These sequences can be replaced by the heterologous phosphoglycerate kinase promoter. Deletion of the GC-rich region between the D-LIF and M-LIF transcriptional start sites results in the appearance of transcripts that do not splice out the first intron of the LIF gene. These may result from gene or promoter trapping of the LIF gene. Sequence analysis of the region between -1200 and -3200 revealed a number of minimal steroid-response elements, regions of similarity to DNAase I-hypersensitive sites in the uteroglobin gene and a region of alternating purine/pyrimidine sequence. This study therefore defines two important regulatory regions in the LIF gene: a GC-rich region in the first intron and a distal 'enhancer' located between -3200 and -1200.

Growth factor regulation of mouse primordial germ cell development

This chapter focused on three key regulators of PGC survival and proliferation; SLF, LIF, and bFGF. The survival of all animal cells may require multiple polypeptide factors and PGCs seem to be no exception (Fig. 7). A number of lines of evidence suggest that membrane-bound forms of SLF may be required for PGC survival. These data suggest an exquisite mechanism for controlling both PGC survival and migration. Thus PGCs that stray from the normal migratory pathway might be eliminated through programmed cell death. SLF, together with LIF, can stimulate PGC proliferation in culture and it seems likely that LIF or a related cytokine may function in vivo to regulate PGC survival and proliferation. Animals doubly deficient in LIF and its relatives may soon allow the roles of these cytokines in PGC development to be determined. Although bFGF is a potent PGC mitogen in vitro, whether PGCs ever encounter bFGF in vivo remains questionable since in culture it alters both the proliferative and developmental potential of PGCs. TGF beta or MIS may be important negative regulators of PGC development, and mice lacking these factors should allow their role in PGC development to be assessed.

Expression of the Brachyury gene during mesoderm development in differentiating embryonal carcinoma cell cultures

When aggregated and treated with dimethyl sulfoxide (DMSO), P19 embryonal carcinoma cells differentiate into cell types normally derived from the mesoderm and endoderm including epithelium and cardiac and skeletal muscle. The Brachyury gene is expressed transiently in these differentiating cultures several days before the appearance of markers of the differentiated cell types. The expression of Brachyury is not affected by DMSO but is induced by cell aggregation, which requires extracellular calcium. Expression of Brachyury is also induced by various members of the TGF beta family such as activin and bone morphogenetic proteins. D3 is a mutant clone of P19 cells selected for its failure to differentiate when aggregated in DMSO. Aggregated D3 cells express Brachyury mRNA suggesting that the mutation(s) responsible for the phenotype of D3 cells is downstream of the chain of events initiated by Brachyury expression.

Ciliary neurotrophic factor maintains the pluripotentiality of embryonic stem cells

Ciliary neurotrophic factor was discovered based on its ability to support the survival of ciliary neurons, and is now known to act on a variety of neuronal and glial populations. Two distant relatives of ciliary neurotrophic factor, leukemia inhibitory factor and oncostatin M, mimic ciliary neurotrophic factor with respect to its actions on cells of the nervous system. In contrast to ciliary neurotrophic factor, leukemia inhibitory factor and oncostatin M also display a broad array of actions on cells outside of the nervous system. The overlapping activities of leukemia inhibitory factor, oncostatin M and ciliary neurotrophic factor can be attributed to shared receptor components. The specificity of ciliary neurotrophic factor for cells of the nervous system results from the restricted expression of the alpha component of the ciliary neurotrophic factor receptor complex, which is required to convert a functional leukemia inhibitory factor/oncostatin M receptor complex into a ciliary neurotrophic factor receptor complex. The recent observation that the alpha component of the ciliary neurotrophic factor receptor complex is expressed by very early neuronal precursors suggested that ciliary neurotrophic factor may act on even earlier precursors, particularly on cells previously thought to be targets for leukemia inhibitory factor action. Here we show the first example of ciliary neurotrophic factor responsiveness in cells residing outside of the nervous system by demonstrating that embryonic stem cells express a functional ciliary neurotrophic factor receptor complex, and that ciliary neurotrophic factor is similar to leukemia inhibitory factor in its ability to maintain the pluripotentiality of these cells.

IP-10, a -C-X-C- chemokine, elicits a potent thymus-dependent antitumor response in vivo

IP-10 is a member of the -C-X-C-chemokine superfamily of proinflammatory cytokines whose secretion is induced by interferon gamma (IFN-gamma) and lipopolysaccharide (LPS). To date no function has been described for IP-10. We have genetically engineered tumor cells to secrete high levels of murine IP-10 and demonstrate that while IP-10 has no effect on the growth of these tumor cells in culture, it elicits a powerful host-mediated antitumor effect in vivo. The IP-10 antitumor response is T lymphocyte dependent, non-cell autonomous, and appears to be mediated by the recruitment of an inflammatory infiltrate composed of lymphocytes, neutrophils, and monocytes. These results document an important biologic property of IP-10 and raise the possibility that some of the T cell-directed effects of IFN-gamma and LPS may be mediated by this chemokine.

Prosomatostatin processing in Neuro2A cells. Role of beta-turn structure in the vicinity of the Arg-Lys cleavage site

Proline residues located near the processing sites of human prosomatostatin were previously shown to be important for cleavage of the precursor into somatostatin 28 and somatostatin 14 [Gomez, S., Boileau, G., Zollinger, L., Nault, C., Rholam, M. & Cohen, P. (1989) EMBO J. 8, 2911-2916]. In this study, site-directed and regional mutagenesis of the human prosomatostatin cDNA coupled with analysis by circular-dichroism and Fourier-transform-infrared spectroscopies of the native and mutated peptide sequences were used to elucidate the role of proline in proteolytic processing. Glycine was substituted for proline a position -5 and the beta-turn-promoting sequence Pro-Arg-Glu-Arg, located near the somatostatin-14 cleavage site and predicted to form a beta-turn structure, was replaced by Ser-Ser-Asn-Arg or Tyr-Lys-Gly-Arg, which have been shown by X-ray diffraction to form beta turns in other proteins. Analysis of the prosomatostatin-derived peptides produced by expression of the mutated cDNA species in Neuro2A cells indicated that while Pro-5-->Ala abolished cleavage at the dibasic site, the formation of mutants [Gly-5] prosomatostatin, [Ser-5, Ser-4, Arg-3] prosomatostatin and [Tyr-5, Lys-4, Gly-3] prosomatostatin did not affect cleavage at the dibasic site but produced modifications in both the relative proportions of the generated hormones and in precursor processing efficiency. Moreover, spectroscopical analysis showed that whereas these substitutions did not modify the presence of a beta turn structure in the corresponding peptide sequences, replacement of Pro-5-->Ala resulted in a dramatic increase in alpha-helix accompanied by the significant decrease of other structures including beta turn. The data support the hypothesis that the proline residue near the processing site for somatostatin-14 production is an important structural feature for conferring on the cleavage domain the adequate conformation for accessibility to processing enzymes and permitting production of equivalent amounts of both hormones.

Expression of alternative forms of differentiation inhibiting activity (DIA/LIF) during murine embryogenesis and in neonatal and adult tissues

Differentiation inhibiting activity/leukaemia inhibitory factor (DIA/LIF) is a pleiotropic cytokine which has been implicated in a variety of developmental and physiological processes in mammals due to its broad range of biological activities in vitro. A role in very early development is suggested by the requirement for DIA/LIF to support the self-renewal of cultured embryonic stem (ES) cells. Other data point to potential roles in the establishment and maintenance of primordial germ cells, in osteogenesis and in haematopoiesis, and possibly in neuronal specification. DIA/LIF may also act as a mediator of the hepatic acute phase response. In the present study the expression of DIA/LIF transcripts during murine development and in adult mice has been determined using a highly sensitive ribonuclease protection analysis. In contrast to previous reports, it is apparent that DIA/LIF transcripts are present at low levels in many adult mouse tissues. Higher levels of expression are observed in skin, lung, intestine, and uterus. Elevated amounts of mRNA are also found in certain foetal tissue during late gestation and neonatally. In earlier embryogenesis, however, DIA/LIF mRNA is produced primarily in extraembryonic tissues. The alternative transcripts which produce either soluble or matrix-associated DIA/LIF exhibit overlapping but non-identical patterns of expression, consistent with the proposition that the two isoforms may have distinct biological functions. These findings are suggestive of widespread roles for DIA/LIF in vivo and are discussed in the light of available data on the phenotype of homozygous DIA/LIF-deficient mice.

Transforming growth factor-beta in the early mouse embryo: implications for the regulation of muscle formation and implantation

In a search for functions of transforming growth factor-beta during early embryonic development we used two different experimental approaches. In the first we made use of embryonic stem (ES) cells. ES cells in culture differentiate to derivatives of all three germ layers and mimic some aspects of organogenesis when grown as aggregates in suspension to form embryoid bodies. Differentiation proceeds further when the embryoid bodies attach to suitable substrates. Muscle and neuronal cells are among the most readily identified cell types then formed. We examined the effect of all-trans retinoic acid (RA) and members of the transforming growth factor-beta family (TGF-beta 1, TGF-beta 2) under these conditions in an assay where single aggregates formed in hanging microdrops in medium supplemented with serum depleted of lipophilic substances which would include retinoids. Endoderm-like cells formed under all conditions tested. RA at concentrations of 10(-8) M and 10(-7) M induced the formation of neurons but in the absence of RA or at concentrations up to 10(-9) M, neurons were not observed. Instead, beating muscle formed in about one-third of the plated aggregates; this was greatly reduced when RA concentrations increased above 10(-9) M. Immunofluorescent staining for muscle specific myosin showed that two muscle cell types could be distinguished: elongated, non-contractile myoblasts and mononucleate flat cells. The mononucleate flat cells appeared to correspond with rhythmically contracting muscle. The number of non-contractile myoblasts increased 3-fold over controls in the presence of 10(-9) M RA. TGF-beta s increased the number of contractile and non-contractile muscle cells by a factor 3 to 7 over controls, depending on the TGF-beta isoform added and the muscle cell type formed. TGF-beta 2 also invariably increased the rate at which contracting muscle cells were first observed in replated aggregates. The stimulatory effect of TGF-beta s on the formation of mononucleate flat cells was completely abrogated by RA at 10(-9) M while the number of myoblasts under similar conditions was unchanged. These data suggest that a complex interplay between retinoids and TGF-beta isoforms may be involved in regulation of differentiation in early myogenesis. In the second approach, neutralizing polyclonal rabbit antibodies specific for TGF-beta 2 were injected into the cavity of mouse blastocysts 3.5 days post coîtum (pc). After 1 day in culture, embryos were transferred to pseudopregnant females. The number of decidua, embryos and resorptions were counted at day 8.5-9.5 pc.(ABSTRACT TRUNCATED AT 400 WORDS)