Regulation of anterior/posterior patterning of the axial skeleton by growth/differentiation factor 11.

The bones that comprise the axial skeleton have distinct morphological features characteristic of their positions along the anterior/posterior axis. We previously described a novel TGF-beta family member, myostatin (encoded by the gene Mstn, formerly Gdf8), that has an essential role in regulating skeletal muscle mass. We also identified a gene related to Mstn by low-stringency screening. While the work described here was being completed, the cloning of this gene, designated Gdf11 (also called Bmp11), was also reported by other groups. Here we show that Gdf11, a new transforming growth factor beta(TGFbeta) superfamily member, has an important role in establishing this skeletal pattern. During early mouse embryogenesis, Gdf11 is expressed in the primitive streak and tail bud regions, which are sites where new mesodermal cells are generated. Homozygous mutant mice carrying a targeted deletion of Gdf11 exhibit anteriorly directed homeotic transformations throughout the axial skeleton and posterior displacement of the hindlimbs. The effect of the mutation is dose dependent, as Gdf11+/- mice have a milder phenotype than Gdf11-/- mice. Mutant embryos show alterations in patterns of Hox gene expression, suggesting that Gdf11 acts upstream of the Hox genes. Our findings suggest that Gdf11 is a secreted signal that acts globally to specify positional identity along the anterior/posterior axis.

Regulation of left-right patterning in mice by growth/differentiation factor-1.

The transforming growth factor-beta (TGF-beta) superfamily encompasses a large group of structurally related polypeptides that are capable of regulating cell growth and differentiation in a wide range of embryonic and adult tissues. Growth/differentiation factor-1 (Gdf-1, encoded by Gdf1) is a TGF-beta family member of unknown function that was originally isolated from an early mouse embryo cDNA library and is expressed specifically in the nervous systemin late-stage embryos and adult mice. Here we show that at early stages of mouse development, Gdfl is expressed initially throughout the embryo proper and then most prominently in the primitive node, ventral neural tube, and intermediate and lateral plate mesoderm. To examine its biological function, we generated a mouse line carrying a targeted mutation in Gdf1. Gdf1-/- mice exhibited a spectrum of defects related to left-right axis formation, including visceral situs inversus, right pulmonary isomerism and a range of cardiac anomalies. In most Gdf1-/- embryos, the expression of Ebaf (formerly lefty-1) in the left side of the floor plate and Leftb (formerly lefty-2), nodal and Pitx2 in the left lateral plate mesoderm was absent, suggesting that Gdf1 acts upstream of these genes either directly or indirectly to activate their expression. Our findings suggest that Gdf1 acts early in the pathway of gene activation that leads to the establishment of left-right asymmetry.

Mice lacking ornithine-delta-aminotransferase have paradoxical neonatal hypoornithinaemia and retinal degeneration.

Deficiency of ornithine-delta-aminotransferase (OAT) in humans causes hyperornithinaemia and gyrate atrophy (GA), a blinding chorioretinal degeneration. Surprisingly, OAT-deficient mice produced by gene targeting exhibit neonatal hypoornithinaemia and lethality, rescuable by short-term arginine supplementation. Post-weaning, these mice develop hyperornithinaemia similar to human GA patients. Subsequent studies in one human GA infant also showed transient hypoornithinaemia. Thus, the OAT reaction plays opposite roles in neonatal and adult mammals. Over several months, OAT-deficient mice develop a retinal degeneration with involvement of photoreceptors and pigment epithelium. OAT-deficient mice appear to be an excellent model of human GA.

Targeted disruption of the mouse factor VIII gene produces a model of haemophilia A.

Haemophilia A is a classic X-linked disease which affects 1 in 5-10,000 males in all populations and is caused by defects in coagulation factor VIII. Roughly 60% of patients have severe disease with factor VIII activity < 1% of normal; they have frequent spontaneous bleeding into joints, soft tissues, muscles and internal organs. These patients usually require regular injections of plasma-derived or recombinant human factor VIII. Because this is expensive and can potentially lead to life-threatening complications, other forms of therapy, including gene therapy, have been proposed. Natural canine models of factor VIII and factor IX deficiency have been available for many years, and gene therapy attempts on these dogs have met with partial success. However, a small animal model of the disease is desirable for studies of factor VIII function and gene therapy. Using gene targeting, we have made a mouse with severe factor VIII deficiency.

Introduction and expression of the 400 kilobase amyloid precursor protein gene in transgenic mice [corrected].

Overexpression of the gene encoding the beta-amyloid precursor protein (APP) may have a key role in the pathogenesis of both Alzheimer's disease (AD) and Down Syndrome (DS). We have therefore introduced a 650 kilobase (kb) yeast artificial chromosome (YAC) that contains the entire, unrearranged 400 kb human APP gene into mouse embryonic stem (ES) cells by lipid-mediated transfection. ES lines were generated that contain a stably integrated, unrearranged human APP gene. Moreover, we demonstrate germ line transmission of the APP YAC in transgenic mice and expression of human APP mRNA and protein at levels comparable to endogenous APP. This transgenic strategy may prove invaluable for the development of mouse models for AD and DS.

Linkage of two pseudogenes from V kappa 1 and V kappa 9 murine immunoglobulin families.

As an initial step towards the molecular analysis of the murine V kappa locus, a cosmid library from BALB/cJ mouse liver DNA was screened with probes representing 10 V kappa families. Of eight cosmids that were isolated from the initial screen, five contained a single restriction fragment that hybridized to the probes. Two cosmids contained two fragments that hybridized to the same probe, V kappa 4, indicating that some V kappa 4 gene segments are linked. One cosmid had two genes that belonged to different families, V kappa 1 and V kappa 9. The two gene segments were located within 12 kb of each other and lay in the same transcriptional orientation. Linkage of gene segments from the V kappa 1 and V kappa 9 families is consistent with a genetic map of the locus, and provides physical evidence for the first time that two genes from different families are closely linked in the murine kappa locus. Sequence analysis revealed that both genes are pseudogenes: the V kappa psi 1.7 gene segment has eight mutations, including termination codons, insertions, and deletions, and the V kappa psi 9B.8 gene segment has two mutations of an insertion and altered RNA splice site. Both genes have the potential to rearrange based on the sequence of their heptamer-nonamer motifs. The identification of pseudogenes raises the question of how many nonfunctional genes are present in the murine germline repertoire.

CD34 expression by embryonic hematopoietic and endothelial cells does not require c-Myb.

CD34 is a cell-surface glycoprotein expressed in a developmental, stage-specific manner by bone marrow stem and progenitor cells. In this study we explored a possible role for c-Myb in CD34 regulation during developmental hematopoiesis. The results indicate that c-Myb can induce CD34 expression in hematopoietic and nonhematopoietic cells, and that murine CD34 promoter activity is enhanced in myeloid cells transgenic for c-Myb. To test whether c-Myb is necessary for CD34 expression during developmental hematopoiesis in vitro, c-Myb-null D3 embryonic stem (ES) cells were analyzed for their ability to develop CD34+ hematopoietic cells in vitro. CD34 promoter activity in transient transfections and CD34 upregulation during ES cell differentiation into embryoid bodies was identical in wild-type and c-Myb-null ES cells, indicating that c-Myb is not required for CD34 expression. CD34 protein is expressed on both hematopoietic and endothelial cells of the E8.5 blood islands during the development of c-Myb-null embryos, and expression is nearly identical in wild-type and c-Myb-null embryos. However, in E12.5 c-Myb-null embryos, the majority of identifiable CD34+ cells in the developing liver are endothelial rather than hematopoietic, which is consistent with the absence of colony-forming units in c-Myb-null embryos and developing ES cells. These data indicate that c-Myb is not required for CD34 expression in endothelial or primitive hematopoietic cells in the yolk sac, but is necessary for definitive hematopoiesis.

Targeted disruption of Gnas in embryonic stem cells.

Mutations in the gene encoding the stimulatory G protein of adenylyl cyclase (G alpha(s)) are present in subjects with Albright hereditary osteodystrophy, a syndrome of characteristic developmental defects and, in some patients, resistance to multiple hormones that stimulate cAMP accumulation (pseudohypoparathyroidism type Ia). As the first step in generating a model of Albright hereditary osteodystrophy, the gene encoding G alpha(s) (Gnas) was disrupted in mouse embryonic stem (ES) cells by homologous recombination. Northern blot analysis and immunoblot analysis demonstrated that steady-state levels of G alpha(s) messenger RNA and G alpha(s) protein in targeted ES cells were approximately 50% of levels in untargeted ES cells. In response to 10 microM forskolin and to various concentrations of isoproterenol (0.1-3.0 microM), cAMP accumulation was reduced in the G alpha(s) knockout ES cell lines, relative to wild-type ES cells and to five of six ES cell lines with randomly integrated targeting vector. These results support the role of G alpha(s) haploinsufficiency in reducing the ability of hormones to generate cAMP in subjects with pseudohypoparathyroidism type Ia. The targeted disruption of Gnas in mouse ES cells establishes an in vitro system for further studies of the role of G alpha(s) and cAMP coupled signal transduction in differentiation and development.

Regulation of skeletal muscle mass in mice by a new TGF-beta superfamily member.

The transforming growth factor-beta (TGF-beta) superfamily encompasses a large group of growth and differentiation factors playing important roles in regulating embryonic development and in maintaining tissue homeostasis in adult animals. Using degenerate polymerase chain reaction, we have identified a new murine TGF-beta family member, growth/differentiation factor-8 (GDF-8), which is expressed specifically in developing and adult skeletal muscle. During early stages of embryogenesis, GDF-8 expression is restricted to the myotome compartment of developing somites. At later stages and in adult animals, GDF-8 is expressed in many different muscles throughout the body. To determine the biological function of GDF-8, we disrupted the GDF-8 gene by gene targeting in mice. GDF-8 null animals are significantly larger than wild-type animals and show a large and widespread increase in skeletal muscle mass. Individual muscles of mutant animals weigh 2-3 times more than those of wild-type animals, and the increase in mass appears to result from a combination of muscle cell hyperplasia and hypertrophy. These results suggest that GDF-8 functions specifically as a negative regulator of skeletal muscle growth.

Characterization of GDF-10 expression patterns and null mice.

Growth/differentiation factor-10 (GDF-10) is a TGF-beta family member highly related to bone morphogenetic protein-3. In order to determine the biological function of GDF-10, we carried out a detailed analysis of the expression pattern of GDF-10 and characterized GDF-10-null mice that we generated by gene targeting. During embryogenesis GDF-10 is expressed prominently in developing skeletal structures both in the craniofacial region and in the vertebral column. In adult animals, GDF-10 is expressed at high levels in the brain, where GDF-10 is localized primarily to cells in the Purkinje cell layer of the cerebellum, and in the uterus, where the expression levels of GDF-10 are regulated both during the menstrual cycle and during pregnancy. Despite the high levels of GDF-10 expression in these tissues, we found no obvious abnormalities in GDF-10-knockout mice with respect to the development of these tissues. These findings suggest either that GDF-10 plays no regulatory role in these tissues or that its function is redundant with that of other growth factor-like molecules.

Adult B-cell repertoire is biased toward two heavy-chain variable-region genes that rearrange frequently in fetal pre-B cells.

Fetal pre-B cells rearrange a very restricted set of immunoglobulin variable genes for the heavy chain (VH). To determine whether the adult B-cell repertoire is similarly skewed, we first identified the genes that rearrange in pre-B cells from BALB/c mice and then determined their frequency of rearrangement in adult B cells. In fetal pre-B cell lines, two genes, VH81X from the 7183 subfamily and VHOx2 from the Q52 subfamily, comprise 75% of the rearranged alleles of an estimated 1000 genes (P less than 0.001). Sequencing analyses revealed that rearrangements involving the two genes were both productive and nonproductive. The biased rearrangement of these two VH genes persists in B-cell hybridomas from adult mice at a frequency of 22%, as determined by Southern gel analysis and RNA sequencing. The sequence of one VHOx2 rearrangement from a hybridoma shows that the rearrangement is productive, suggesting that the gene encodes an antibody that could participate in the immune response. The data indicate that the adult B-cell repertoire is not random concerning usage of individual VH genes, and it may be shaped by the unknown mechanisms that cause preferential rearrangement of certain genes early in ontogeny.

Craniofacial abnormalities resulting from targeted disruption of the murine Sim2 gene.

Sim2 is a member of the basic helix-loop-helix PAS transcription factor gene family and is evolutionarily related to the Drosophila single-minded gene, a key regulator of central nervous system midline development. In an effort to determine the biological roles of Sim2 in mammalian development, we disrupted the murine Sim2 gene through gene targeting. Mice homozygous for the disrupted allele (Sim2 -/-) exhibit a cleft of the secondary palate and malformations of the tongue and pterygoid processes of the sphenoid bone. These craniofacial malformations are the most probable cause of aerophagia (air swallowing with subsequent accumulation of air in the gastrointestinal tract) and postnatal death exhibited by Sim2 -/- mice. The developing palates of the Sim2 -/- mice are hypocellular, and at embryonic day 14.5 contain excess extracellular matrix component hyaluronan (HA) compared with heterozygotes and homozygous wild-type littermates. HA plays an important role in the regulation and mechanics of palate development. Its premature accumulation in Sim2 -/- animal palates suggests a regulatory role for Sim2 in HA synthesis and in the establishment of craniofacial architecture.

Early rearrangements of genes encoding murine immunoglobulin kappa chains, unlike genes encoding heavy chains, use variable gene segments dispersed throughout the locus.

Immunoglobulin heavy-chain variable region (TH) gene segments located closest to the joining (JH) gene segments are preferentially rearranged during ontogeny, indicating that chromosomal position influences the frequency of rearrangement. In addition, certain VH gene segments are repeatedly rearranged, suggesting that the DNA sequence or structure surrounding these segments may increase the probability of rearrangement. To determine whether there is similar based rearrangement of kappa variable (V kappa) gene segments, 25 rearrangements were sequenced from murine fetal and neonatal B-cell hybridomas and from subclones of a pre-B cell line that rearranged V kappa genes during in vitro culture. Four gene segments were isolated twice and one gene segment was isolated three times, suggesting that the process that targets individual variable gene segments for repeated rearrangement operates on both the VH and V kappa loci. Based on a current map of the V kappa locus, the rearranged gene segments belong to nine families that are dispersed throughout the locus. Thus, in these cell types, V kappa rearrangements use germ-line gene segments located across the entire locus, whereas the corresponding VH rearrangements use gene segments proximal to the JH gene segments. Heterogeneity of V kappa rearrangements would add diversity to the biased pool of VH rearrangements, producing a broad repertoire of antibodies early in development.

Growth retardation and neonatal lethality in mice with a homozygous deletion in the C-terminal domain of RNA polymerase II.

The C-terminal domain (CTD) of the largest subunit of RNA polymerase II consists of tandem repeats of the consensus heptapeptide YSPTSPS. Deletion studies in tissue culture cells have indicated that the CTD plays an essential role in transcription, although the nature of this essential function remains unclear. About half of the CTD can be deleted without affecting the viability of cells in tissue culture. Paradoxically, the dispensable CTD repeats are precisely conserved among all mammals whose CTD sequences are known. To determine whether the mammalian CTD is important in transcription during mouse development, we developed a gene targeting approach to introduce deletions into the CTD coding region of mouse embryonic stem (ES) cells. To maintain a functional Rpo2-1 gene, the neo marker in the targeting vector was positioned outside of the Rpo2-1 transcribed region, 1.2 kb from the site of the CTD deletion. G418-resistant clones were screened for co-integration of the CTD deletion, and the resulting ES lines were used to create germline chimeric mice. Stable heterozygous lines were established and mated to produce animals homozygous for the CTD deletion. We show here that mice homozygous for a deletion of thirteen of the 52 heptapeptide repeats are smaller than wild-type littermates and have a high rate of neonatal lethality. Surviving adults, although small, appear morphologically normal and are fertile. This result suggests that the CTD plays a role in regulating growth during mammalian development. The gene targeting approach described here should be useful for making further deletions in the CTD and may be of general applicability where it is desirable to engineer specific mutations in the germline of mice.

Alterations of yeast artificial chromosome transgenic sequences in stretched embryonic stem-cell chromatin visualized by fluorescence in situ hybridization.

Transgenic mice have been generated from embryonic stem (ES) cells carrying functional genes cloned within yeast artificial chromosomes (YACs). Information on the integrity and organization of the inserted sequences, including the number of copies and their orientation to each other, is still limited by current methods. We have applied fluorescence in situ hybridization to stretched chromatin preparations from YAC-transfected ES cells to analyze the organization and copy number of the integrated sequences.

Rent1, a trans-effector of nonsense-mediated mRNA decay, is essential for mammalian embryonic viability.

The ability to detect and degrade transcripts that lack full coding potential is ubiquitous but non-essential in lower eukaryotes, leaving in question the evolutionary basis for complete maintenance of this function. One hypothesis holds that nonsense-mediated RNA decay (NMD) protects the organism by preventing the translation of truncated peptides with dominant negative or deleterious gain-of-function potential. All organisms studied to date that are competent for NMD express a structural homolog of Saccharomyces cerevisiae Upf1p. We have now explored the consequences of loss of NMD function in vertebrates through targeted disruption of the Rent1 gene in murine embryonic stem cells which encodes a mammalian ortholog of Upf1p. Mice heterozygous for the targeted allele showed no apparent phenotypic abnormalities but homozygosity was never observed, demonstrating that Rent1 is essential for embryonic viability. Homozygous targeted embryos show complete loss of NMD and are viable in the pre-implantation period, but resorb shortly after implantation. Furthermore, Rent1(-/-) blastocysts isolated at 3.5 days post-coitum undergo apoptosis in culture following a brief phase of cellular expansion. These data suggest that NMD is essential for mammalian cellular viability and support a critical role for the pathway in the regulated expression of selected physiologic transcripts.

A mouse model for X-linked adrenoleukodystrophy.

X-linked adrenoleukodystrophy (X-ALD) is a peroxisomal disorder with impaired beta-oxidation of very long chain fatty acids (VLCFAs) and reduced function of peroxisomal very long chain fatty acyl-CoA synthetase (VLCS) that leads to severe and progressive neurological disability. The X-ALD gene, identified by positional cloning, encodes a peroxisomal membrane protein (adrenoleukodystrophy protein; ALDP) that belongs to the ATP binding cassette transporter protein superfamily. Mutational analyses and functional studies of the X-ALD gene confirm that it and not VLCS is the gene responsible for X-ALD. Its role in the beta-oxidation of VLCFAs and its effect on the function of VLCS are unclear. The complex pathology of X-ALD and the extreme variability of its clinical phenotypes are also unexplained. To facilitate understanding of X-ALD pathophysiology, we developed an X-ALD mouse model by gene targeting. The X-ALD mouse exhibits reduced beta-oxidation of VLCFAs, resulting in significantly elevated levels of saturated VLCFAs in total lipids from all tissues measured and in cholesterol esters from adrenal glands. Lipid cleft inclusions were observed in adrenocortical cells of X-ALD mice under the electron microscope. No neurological involvement has been detected in X-ALD mice up to 6 months. We conclude that X-ALD mice exhibit biochemical defects equivalent to those found in human X-ALD and thus provide an experimental system for testing therapeutic intervention.

Further characterization of factor VIII-deficient mice created by gene targeting: RNA and protein studies.

Previously we created two strains of factor VIII-deficient mice by insertion of a neo gene into (1) the 3' end of exon 16 and (2) exon 17 of the factor VIII gene. Affected mice of both strains have no plasma factor VIII activity, yet are healthy with no spontaneous bleeding. Factor VIII-deficient females bred with affected males survive pregnancy and delivery. We used reverse transcriptase-polymerase chain reaction of liver RNA to characterize factor VIII mRNA processing. Factor VIII mRNA of the exon 16 knockout strain contains neo sequences plus 17 bp of intron 16 due to use of a cryptic donor site in intron 16. All factor VIII mRNA of the exon 17 knockout strain lacks exon 17 and neo sequences. In skipping exon 17, the intron 16 donor site or a cryptic donor site 46 bp 3' to the intron 16 donor site are used. Thus, factor VIII deficiency in exon 16 knockout mice is due to truncated protein, while in exon 17 knockout mice it is due to either truncated or partially deleted protein. After immunizing exon 16 knockout mice with human recombinant factor VIII, two monoclonal antibodies were obtained that recognize < 100 pg of mouse factor VIII light chain. Assay of cryoprecipitate from the plasma of affected mice failed to show factor VIII light chain.