Heterokaryon analysis of muscle differentiation: regulation of the postmitotic state.

MM14 mouse myoblasts withdraw irreversibly from the cell cycle and become postmitotic within a few hours of being deprived of fibroblast growth factor (Clegg, C. H., T. A. Linkhart, B. B. Olwin, and S. D. Hauschka, 1987, J. Cell Biol., 105:949-956). To examine the mechanisms that may regulate this developmental state of skeletal muscle, we tested the mitogen responsiveness of various cell types after their polyethylene glycol-mediated fusion with post-mitotic myocytes. Heterokaryons containing myocytes and quiescent nonmyogenic cells such as 3T3, L cell, and a differentiation-defective myoblast line (DD-1) responded to mitogen-rich medium by initiating DNA synthesis. Myonuclei replicated DNA and reexpressed thymidine kinase. In contrast, (myocyte x G1 myoblast) heterokaryons failed to replicate DNA in mitogen-rich medium and became postmitotic. This included cells with a nuclear ratio of three myoblasts to one myocyte. Proliferation dominance in (myocyte x 3T3 cell) and (myocyte x DD-1) heterokaryons was conditionally regulated by the timing of mitogen treatment; such cells became postmitotic when mitogen exposure was delayed for as little as 6 h after cell fusion. In addition, (myocyte x DD-1) heterokaryons expressed a muscle-specific trait and lost epidermal growth factor receptors when they became postmitotic. These results demonstrate that DNA synthesis is not irreversibly blocked in skeletal muscle; myonuclei readily express proliferation-related functions when provided with a mitogenic signal. Rather, myocyte-specific repression of DNA synthesis in heterokaryons argues that the postmitotic state of skeletal muscle is regulated by diffusible factors that inhibit processes of cellular mitogenesis.

Growth factor control of skeletal muscle differentiation: commitment to terminal differentiation occurs in G1 phase and is repressed by fibroblast growth factor.

Analysis of MM14 mouse myoblasts demonstrates that terminal differentiation is repressed by pure preparations of both acidic and basic fibroblast growth factor (FGF). Basic FGF is approximately 30-fold more potent than acidic FGF and it exhibits half maximal activity in clonal assays at 0.03 ng/ml (2 pM). FGF repression occurs only during the G1 phase of the cell cycle by a mechanism that appears to be independent of ongoing cell proliferation. When exponentially growing myoblasts are deprived of FGF, cells become postmitotic within 2-3 h, express muscle-specific proteins within 6-7 h, and commence fusion within 12-14 h. Although expression of these three terminal differentiation phenotypes occurs at different times, all are initiated by a single regulatory "commitment" event in G1. The entire population commits to terminal differentiation within 12.5 h of FGF removal as all cells complete the cell cycle and move into G1. Differentiation does not require a new round of DNA synthesis. Comparison of MM14 behavior with other myoblast types suggests a general model for skeletal muscle development in which specific growth factors serve the dual role of stimulating myoblast proliferation and directly repressing terminal differentiation.

Production of transgenic mice with yeast artificial chromosomes.

Techniques are now available that allow the transfer of intact yeast artificial chromosome (YAC) DNA into transgenic mice. Coupled with the ability to perform mutagenesis on YAC sequences by homologous recombination in yeast, they enable the analysis of large genes or multigenic loci in vivo. This system has been used to study the developmental regulation of the human beta-globin locus.

Transgenic expression of the human amphiregulin gene induces a psoriasis-like phenotype.

Amphiregulin (AR) is a heparin-binding, heparin-inhibited member of the epidermal growth factor (EGF) family and an autocrine growth factor for human keratinocytes. Previous studies have shown that AR expression is increased in psoriatic epidermis. To test the hypothesis that aberrant AR expression is central to the development of psoriatic lesions, we constructed a transgene (K14-ARGE) encoding a human keratin 14 promoter-driven AR gene. Our results indicate that transgene integration and subsequent expression of AR in basal keratinocytes correlated with a psoriasis-like skin phenotype. Afflicted mice demonstrated shortened life spans, prominent scaling and erythematous skin with alopecia, and occasional papillomatous epidermal growths. Histologic examination revealed extensive areas of marked hyperkeratosis with focal parakeratosis, acanthosis, dermal and epidermal lymphocytic and neutrophilic infiltration, and dilated blood vessels within the papillary dermis. Our results reveal that AR exerts activity in the skin that is distinct from that of transgenic transforming growth factor-alpha or other cytokines, and induces skin pathology with striking similarities to psoriasis. Our observations also link the keratinocyte EGF receptor-ligand system to psoriatic inflammation, and suggest that aberrant expression of AR in the epidermis may represent a critical step in the development or propagation of psoriatic lesions.

Sheltering of gamma-globin expression from position effects requires both an upstream locus control region and a regulatory element 3' to the A gamma-globin gene.

Integration position-independent expression of human globin transgenes in transgenic mice requires the presence of regulatory elements from the beta-globin locus control region (LCR) in the transgene construct. However, several recent studies have suggested that, while clearly necessary, such elements are not by themselves sufficient to realize this effect. In the case of the human fetal gamma-globin genes, previous results have indicated that additional regulatory information required for sheltering of gamma-globin transgene expression from position effects may reside downstream from the A gamma gene. To investigate this possibility, we established 17 lines of transgenic mice carrying constructs comprising a micro-LCR (microLCR) element, an A gamma-globin gene fragment, and a variable length of 3' sequence information beyond the A gamma 3' HindIII site. gamma-Globin expression during development was studied in 170 individual F2 progeny from these lines. We find that gamma-globin expression becomes sheltered from position effects when the normally position-sensitive microLCR-A gamma construct is extended by 600 bp beyond the 3' HindIII site to include a previously identified regulatory sequence (the A gamma-globin enhancer), the functional significance of which in vivo had heretofore been unclear. The results suggest that the mechanism whereby an upstream LCR achieves sheltering of globin gene expression from position effects involves cooperation with a gene-proximal regulatory element distinct from the promoter region.

Developmental abnormalities in mice transgenic for bovine oncostatin M.

Oncostatin M belongs to the subfamily of hematopoietin cytokines that binds a receptor complex containing gp130. To date, only the human form of oncostatin M has been identified, and its evolutionary conservation is unresolved. We have isolated a bovine gene whose open reading frame encodes a precursor protein that is 58% identical to human oncostatin M. A comparison of the bovine and human amino acid sequences predicts significant similarity, including the four-alpha-helical-bundle structure and the placement of disulfide bridges. As with the human protein, bovine oncostatin M binds specific receptors on human H2981 cells and inhibits the proliferation of human A375 tumor cells and mouse M1 leukemia cells. To identify activities regulated in vivo, we injected bovine oncostatin M fusion genes containing various tissue-specific promoters into mouse embryos. The frequencies of transgenic mice were reduced significantly, suggesting that overexpression of the bovine cytokine is detrimental to normal mouse development. In addition to deaths associated with expression in neurons and keratinized epithelia, bovine oncostatin M caused abnormalities in bone growth and spermatogenesis, stimulated fibrosis surrounding islets in the pancreas, and disrupted normal lymphoid tissue development. This work establishes the existence of a nonprimate oncostatin M gene and provides the first demonstration that this cytokine can function in a pleiotropic manner in vivo. Information regarding bovine oncostatin M may help characterize the structure and function of this cytokine in other vertebrate species.

E-box sites and a proximal regulatory region of the muscle creatine kinase gene differentially regulate expression in diverse skeletal muscles and cardiac muscle of transgenic mice.

Previous analysis of the muscle creatine kinase (MCK) gene indicated that control elements required for transcription in adult mouse muscle differed from those required in cell culture, suggesting that distinct modes of muscle gene regulation occur in vivo. To examine this further, we measured the activity of MCK transgenes containing E-box and promoter deletions in a variety of striated muscles. Simultaneous mutation of three E boxes in the 1,256-bp MCK 5' region, which abolished transcription in muscle cultures, had strikingly different effects in mice. The mutations abolished transgene expression in cardiac and tongue muscle and caused a reduction in expression in the soleus muscle (a muscle with many slow fibers) but did not affect expression in predominantly fast muscles: quadriceps, abdominals, and extensor digitorum longus. Other regulatory sequences with muscle-type-specific activities were found within the 358-bp 5'-flanking region. This proximal region conferred relatively strong expression in limb and abdominal skeletal muscles but was inactive in cardiac and tongue muscles. However, when the 206-bp 5' enhancer was ligated to the 358-bp region, high levels of tissue-specific expression were restored in all muscle types. These results indicate that E boxes and a proximal regulatory region are differentially required for maximal MCK transgene expression in different striated muscles. The overall results also imply that within skeletal muscles, the steady-state expression of the MCK gene and possibly other muscle genes depends on transcriptional mechanisms that differ between fast and slow fibers as well as between the anatomical and physiological attributes of each specific muscle.

Analysis of muscle creatine kinase gene regulatory elements in skeletal and cardiac muscles of transgenic mice.

Regulatory regions of the mouse muscle creatine kinase (MCK) gene, previously discovered by analysis in cultured muscle cells, were analyzed in transgenic mice. The 206-bp MCK enhancer at nt-1256 was required for high-level expression of MCK-chloramphenicol acetyltransferase fusion genes in skeletal and cardiac muscle; however, unlike its behavior in cell culture, inclusion of the 1-kb region of DNA between the enhancer and the basal promoter produced a 100-fold increase in skeletal muscle activity. Analysis of enhancer control elements also indicated major differences between their properties in transgenic muscles and in cultured muscle cells. Transgenes in which the enhancer right E box or CArG element were mutated exhibited expression levels that were indistinguishable from the wild-type transgene. Mutation of three conserved E boxes in the MCK 1,256-bp 5' region also had no effect on transgene expression in thigh skeletal muscle expression. All these mutations significantly reduced activity in cultured skeletal myocytes. However, the enhancer AT-rich element at nt - 1195 was critical for expression in transgenic skeletal muscle. Mutation of this site reduced skeletal muscle expression to the same level as transgenes lacking the 206-bp enhancer, although mutation of the AT-rich site did not affect cardiac muscle expression. These results demonstrate clear differences between the activity of MCK regulatory regions in cultured muscles cells and in whole adult transgenic muscle. This suggests that there are alternative mechanism of regulating the MCK gene in skeletal and cardiac muscle under different physiological states.

Developmental specificity of the interaction between the locus control region and embryonic or fetal globin genes in transgenic mice with an HS3 core deletion.

The human beta-globin locus control region (LCR) consists of five erythroid-lineage-specific DNase I-hypersensitive sites (HSs) and is required for activation of the beta-globin locus chromatin domain and globin gene expression. Each DNase I-HS of the LCR consists of a highly conserved core element and flanking sequences. To analyze the functional role of the core elements of the HSs, we deleted a 234-bp fragment encompassing the core of HS3 (HS3c) from a beta-globin locus residing on a 248-kb beta-locus yeast artificial chromosome and analyzed its function in F2 progeny of transgenic mice. Human epsilon-globin gene expression was absent at day 10 and severely reduced in the day 12 embryonic erythropoiesis of mice lacking HS3c. In contrast, gamma-globin gene expression was normal in embryonic erythropoiesis but it was absent in definitive erythropoiesis in the fetal liver. These results indicate that the core element of HS3 is necessary for epsilon-globin gene transcription in embryonic cells and for gamma-globin gene transcription in definitive cells. Normal gamma-globin gene expression in embryonic cells and the absence of gamma-globin gene expression in definitive cells show that different HSs interact with gamma-globin gene promoters in these two stages of development. Such results provide direct evidence for developmental stage specificity of the interactions between the core elements of HSs and the promoters of the globin genes.

Oncostatin M stimulates excessive extracellular matrix accumulation in a transgenic mouse model of connective tissue disease.

Oncostatin M (OM), a member of the IL-6 gene family, stimulates a variety of functions implicated in wound repair. Transgenic mice that express this cytokine in islet beta-cells develop a connective tissue disorder that typifies excessive healing with severe fibrosis and lymphocytic infiltration. To compare this phenotype with the normal progression of connective tissue disease, we measured the expression patterns of genes encoding proinflammatory cytokines, fibrogenic cytokines, and ECM components by in situ hybridization. To test whether the OM effect was caused by its ability to regulate IL-6, we crossed the OM transgene into IL-6-deficient mice. Our data suggest that the fibrosis in these animals is not a secondary consequence of inflammation, or IL-6 expression, but is a direct effect by OM on extracellular matrix production. In a separate experiment, we observed that OM could regulate vasoactive intestinal peptide gene expression in the neurons that innervate the transgenic pancreas. This nerve healing response, in combination with its fibrogenic activity, suggests that OM functions downstream of inflammation in the wound repair cascade. These transgenic mice represent a useful model in which the fibroproliferative phase of connective tissue disease is uncoupled from inflammation.

Promoter for the regulatory type I beta subunit of the 3',5'-cyclic adenosine monophosphate-dependent protein kinase directs transgene expression in the central nervous system.

Cyclic AMP-dependent protein kinase (cAPK) modulates synaptic transmission and influences memory and learning. Among the various isoforms of regulatory and catalytic subunits that comprise mammalian cAPK, only the regulatory type I beta (RI beta) subunit is unique to nervous tissue. The requirement for RI beta in neurons is presently unknown. Previous studies demonstrate that holoenzyme containing RI beta activates at lower concentrations of cAMP compared to other forms of cAPK. Thus, neurons that induce RI beta expression may become more sensitive to subsequent hormonal signals and maintain more long-term phosphorylation events. To further elucidate the function of this novel protein, we have begun to investigate its gene. Here we report the isolation of the mouse RI beta promoter as determined by S1 nuclease analysis and transgenic mouse expression. A beta-galactosidase fusion gene containing 1.5 kilobases of 5'-nontranscribed RI beta DNA and 2 kilobases of intron 1 was expressed preferentially in the cortex and hippocampus of the brain and within the spinal cord. In addition to mimicking the location of endogenous RI beta expression, the transgene was activated at a similar time (embryonic day 11.5) during mouse fetal development. Isolation of the RI beta promoter will help identify the elements that direct transcription in a subset of neurons and illuminate the physiological conditions that may regulate RI beta expression. This promoter can also be used to target the expression of wild type and mutant cAPK subunit genes in order to investigate synaptic plasticity in animals.

Generation of transgenic mice from yeast artificial chromosome DNA that has been modified by gene targeting.

Yeast artificial chromosome (YAC) vectors permit the cloning of up to megabase fragments of human genomic DNA in yeast (Saccharomyces cerevisiae). Efficient recombination between homologous segments of DNA is one of the hallmark genetic features of yeast. This characteristic facilitates the introduction of specific mutations into YACs by gene targeting. Gene targeting has been used recently to introduce specific mutations into YACs spanning the human ß-globin locus and the human apolipoprotein (apo)-B gene, and the mutated YAC DNA has been used to generate transgenic mice. This approach has been useful for the study of the regulatory elements controlling ß-globin gene expression and for the study of apo-B structure and function. This article reviews the techniques for introducing mutations into YACs, for the purpose of expression in transgenic mice.

Oncostatin M transforms lymphoid tissue function in transgenic mice by stimulating lymph node T-cell development and thymus autoantibody production.

Oncostatin M (OM) is a member of the IL-6 subfamily of cytokines that is expressed in primary lymphoid tissues such as bone marrow and thymus, as well as in secondary lymphoid tissues and activated leukocytes. We produced transgenic mice that overexpressed the human, bovine, or mouse OM genes and compared their relative ability to modulate lymphopoiesis. Each species of cytokine induced a similar extrathymic pathway of T-cell development involving the accumulation of immature T cells within lymph nodes. Reconstitution experiments utilizing lethally irradiated athymic mice indicated that OM had caused hematopoietic precursors within fetal liver and bone marrow to initiate lymph node T-cell development in the absence of a thymic environment. Breeding experiments with IL6-/- and IL-7r(alpha)-/- deficient mice, indicated that induction of this extrathymic pathway by the OM transgene occurred in the absence of IL-6, but was strictly dependent on IL-7 receptor signaling. Separately, OM stimulated the accumulation of immature B cells within the transgenic thymus and caused the subcapsular regions of the thymus to expand with mature B and T cells. This thymus conversion to secondary lymphoid tissue was responsible for a lethal autoimmune-like disease marked by high titers of circulating autoantibodies, proteinuria, and glomerulonephritis. The conserved phenotypes elicited by these three forms of OM indicate that this potent hematopoietic cytokine can regulate lymphoid tissue function and morphogenesis.

Thymic overexpression of CD40 ligand disrupts normal thymic epithelial organization.

We characterized the distribution of CD40 and CD40 ligand (CD40-L) in the adult and developing murine thymus. Before birth, CD40 was almost exclusively localized to scattered foci of medullary cells. By birth there was a dramatic upregulation of CD40 expression by cortical epithelial cells, which was accompanied by a consolidation of medullary epithelial foci. CD40-L+ thymocytes displayed a medullary location. Analysis of mice deficient in CD40-L expression indicated that CD40-L/CD40 interactions were not required for development of the medullary compartment. Overexpression of CD40-L targeted to thymocytes altered thymic architecture, as reflected by a dramatic loss of cortical epithelial cells, expansion of the medullary compartment, and extensive infiltration of the capsule with a mixture of CD3+ cells, B-cells, and macrophages/dendritic cells. Reconstitution of lethally irradiated normal mice with lck CD40-L bone marrow cells also resulted in loss of cortical epithelium and expansion of the medullary compartment. Disruption of the normal pattern of thymic architecture and epithelial differentiation as a consequence of increased intrathymic levels of CD40-L expression points to a role for CD40-L/CD40 interactions in the normal pattern of epithelial compartmentalization/differentiation within the thymic environment.

Genetic characterization of a brain-specific form of the type I regulatory subunit of cAMP-dependent protein kinase.

An isoform (RI beta) of the regulatory type I subunit gene of cAMP-dependent protein kinase (EC 2.7.1.37) has been characterized in mouse. The open reading frame of the RI beta cDNA is 72% identical in nucleotide sequence with the previously cloned RI gene, now referred to as RI alpha. Both genes code for a protein of 380 amino acids and their proteins are 82% identical in amino acid sequence. Sequence similarity is highest in the regions that form the pseudosubstrate-binding site of the catalytic subunit and the two cAMP binding domains. The amino-terminal portion shows the greatest dissimilarity, suggesting that the isoforms may differ in their dimerization properties or interaction with other proteins. In contrast to RI alpha, which is constitutively expressed in all tissues, RI beta is expressed in a highly tissue-specific manner. Brain and spinal cord contained significant levels of RI beta mRNA, testis RNA gave a detectable signal, and all other tissues tested were negative. Expression of a RI beta cDNA in NIH 3T3 cells resulted in the appearance of a RI subunit protein that migrated more slowly than RI alpha after NaDodSO4/PAGE. The native form of RI beta in brain could also be distinguished from RI alpha by its abnormal migration on NaDodSO4/PAGE. RI beta protein produced in 3T3 cells was shown to be functional by its ability to form a cAMP-dependent holoenzyme with the catalytic subunit.

Regulation of EGR-1, c-jun, and c-myc gene expression by oncostatin M.

Oncostatin M (OM) is a cytokine that shares a structural and functional relationship with interleukin 6, leukemia-inhibitory factor, and granulocyte colony-stimulating factor. In this report, we tested for correlations between immediate-early gene expression and some of the cellular responses elicited by OM. We determined that OM stimulated a rapid and transient elevation of EGR-1, c-jun, and c-myc mRNA in human fibroblasts prior to their proliferation. OM also stimulated a transient induction of these genes in M1 leukemic cells that differentiated into nonreplicating, macrophage-like cells. The expression of c-myc, however, decreased significantly as the cells stopped dividing. Interestingly, OM had no detectable effect on the expression of EGR-1, c-jun, and c-myc during the cell cycle arrest of human A375 melanoma cells. Our results indicate that an early nuclear event associated with OM action is the regulation of immediate-early gene expression. We suggest that the transcription factors encoded by the EGR-1, c-jun, and c-myc genes are utilized in both cell proliferation and differentiation but are not part of the mechanism by which OM inhibits A375 cell growth.

In vitro positive selection and anergy induction of class II-restricted TCR transgenic thymocytes by a cortical thymic epithelial cell line.

Thymic epithelial cell lines isolated from hyperplastic thymi of transgenic mice over-expressing human papilloma viral oncogenes E6 and E7 constitutively displayed a phenotype consistent with a cortical origin. Exposure to IFN-gamma induced class II MHC and ICAM-1 expression, and up-regulated expression of VCAM-1 and class I MHC molecules. CD40 expression was maximally induced by a combination of IFN-gamma and IL-1, with lower levels of induction observed with a mixture of IFN-gamma and tumor necrosis factor (TNF)-alpha or TNF-alpha alone. B7-1 or B7-2 was not expressed constitutively or in response to cytokines. These stromal cells supported the development of CD4 single-positive (SP) cells in reaggregate co-cultures with CD4+ CD8+ thymocytes from TCR transgenic mice, but did not stimulate class II MHC-restricted, moth cytochrome c (MCC)-reactive T cells in vitro. The behavior of the culture system was consistent with positive selection, i.e. increased numbers of CD4 SP cells, gain of antigen responsiveness, and requirement for epithelial class II MHC products. Some variants of these stromal cell lines required exogenous MCC peptide in the reaggregation cultures (RC) for positive selection to occur. While a low concentration of MCC peptide (0.01-0.1 microM) significantly enhanced the accumulation of CD4 SP cells, higher concentrations of peptide (1-10 microM) resulted in recovery of predominantly CD4- CD8- and CD4(low) CD8- cells. Thymocytes recovered from RC containing low, but not high concentrations of peptide responded to MCC peptide in secondary cultures with splenic antigen-presenting cells.

Promoter sequences in the RI beta subunit gene of cAMP-dependent protein kinase required for transgene expression in mouse brain.

Neural-specific expression of the mouse regulatory type-I beta (RI beta) subunit gene of cAMP-dependent protein kinase is controlled by a fragment of genomic DNA comprised of a TATA-less promoter flanked by 1.5 kilobases of 5'-upstream sequence and a 1.8-kilobase intron. This DNA contains a complex arrangement of transcription factor binding motifs, and previous experiments have shown that many of these are recognized by proteins found in brain nuclear extract. To identify sequences critical for RI beta expression in functional neurons, we performed a deletion analysis in transgenic mice. Evidence is presented that the GC-rich proximal promoter is responsible for cell type-specific expression in vivo because RI beta DNA containing as little as 17 base pairs (bp) of 5'-upstream sequence was functional in mouse brain. One likely regulatory element coincides with the start of transcription and includes an EGR-1 motif and 3 consecutive SP1 sites within a 21-bp interval. Maximal RI beta promoter activity required the adjacent 663 bp of 5'-upstream DNA where most, but not all, of the regulatory activity was localized between position -663 and -333. A 37-bp direct repeat lies within this region that contains 2 basic helix-loop-helix binding sites, each of which are overlapped by two steroid hormone receptor half-sites, and a shared AP1 consensus sequence. Intron I sequences were also tested, and deletion of a 388-bp region containing numerous Sp1-like sequences lowered transgene activity significantly. These results have identified specific regions of the RI beta promoter that are required for the expression of this signal transduction protein in mouse neurons.

Control of mouse myoblast commitment to terminal differentiation by mitogens.

Regulation of the transition of mouse myoblasts from proliferation to terminal differentiation was studied with clonal density cultures of a permanent clonal myoblast cell line. In medium lacking mitogenic activity, mouse myoblasts withdraw from the cell cycle, elaborate muscle-specific gene products, and fuse to form multinucleated myotubes. Addition of a purified mitogen, fibroblast growth factor, to mitogen-depleted medium stimulates continued proliferation and prevents terminal differentiation. When mitogens are removed for increasing durations and then refed, mouse myoblasts irreversibly commit to terminal differentiation: after 2-4 h in the absence of mitogens, myoblasts withdraw from the cell cycle, elaborate muscle-specific gene products, fuse in the presence of mitogens that have been fed back. Population kinetics of commitment determined with 3H-thymidine labeling and autoradiography suggests the following cell-cycle model for mouse myoblast commitment: 1)if mitogens are present in the extracellular environment of myoblasts in G1 of the cell cycle, the cells enter S and continue through another cell cycle; 2) if mitogens have been absent for 2 or more hours, cells in G1 do not enter S; the cells commit to differentiate, permanently withdraw from the cell cycle (will not enter S if mitogens are refed), and they subsequently elaborate acetylcholine receptors and fuse (even if mitogens are refed); 3) cells in other phases of the cell cycle continue to transit the cell cycle in the absence of mitogens until reaching the next G1. the commitment kinetics and experiments with mitotically synchronized cells suggest that the commitment "decision" is made during G1. Present results do not, however, exclude commitment of some cells in other phases of the cell cycle.