Agent-based model of angiogenesis simulates capillary sprout initiation in multicellular networks

Many biological processes are controlled by both deterministic and stochastic influences. However, efforts to model these systems often rely on either purely stochastic or purely rule-based methods. To better understand the balance between stochasticity and determinism in biological processes a computational approach that incorporates both influences may afford additional insight into underlying biological mechanisms that give rise to emergent system properties. We apply a combined approach to the simulation and study of angiogenesis, the growth of new blood vessels from existing networks. This complex multicellular process begins with selection of an initiating endothelial cell, or tip cell, which sprouts from the parent vessels in response to stimulation by exogenous cues. We have constructed an agent-based model of sprouting angiogenesis to evaluate endothelial cell sprout initiation frequency and location, and we have experimentally validated it using high-resolution time-lapse confocal microscopy. ABM simulations were then compared to a Monte Carlo model, revealing that purely stochastic simulations could not generate sprout locations as accurately as the rule-informed agent-based model. These findings support the use of rule-based approaches for modeling the complex mechanisms underlying sprouting angiogenesis over purely stochastic methods.

Assembly of trunk and limb blood vessels involves extensive migration and vasculogenesis of somite-derived angioblasts

Vascular development requires the assembly of precursor cells into blood vessels, but how embryonic vessels are assembled is not well understood. To determine how vascular cells migrate and assemble into vessels of the trunk and limb, marked somite-derived angioblasts were followed in developing embryos. Injection of avian somites with the cell-tracker DiI showed that somite-derived angioblasts in unperturbed embryos migrated extensively and contributed to trunk and limb vessels. Mouse-avian chimeras with mouse presomitic mesoderm grafts had graft-derived endothelial cells in blood vessels at significant distances from the graft, indicating that mouse angioblasts migrated extensively in avian hosts. Mouse graft-derived endothelial cells were consistently found in trunk vessels, such as the perineural vascular plexus, the cardinal vein, and presumptive intersomitic vessels, as well as in vessels of the limb and kidney rudiment. This reproducible pattern of graft colonization suggests that avian vascular patterning cues for trunk and limb vessels are recognized by mammalian somitic angioblasts. Mouse-quail chimeras stained with both the quail vascular marker QH1 and the mouse vascular marker PECAM-1 had finely chimeric vessels, with graft-derived mouse cells interdigitated with quail vascular cells in most vascular beds colonized by graft cells. Thus, diverse trunk and limb blood vessels have endothelial cells that developed from migratory somitic angioblasts, and assembly of these vessels is likely to have a large vasculogenic component.

Characterization of the vasculogenic block in the absence of vascular endothelial growth factor-A

Vascular endothelial growth factor (VEGF) signaling is required for both differentiation and proliferation of vascular endothelium. Analysis of differentiated embryonic stem cells with one or both VEGF-A alleles deleted showed that both the differentiation and the expansion of endothelial cells are blocked during vasculogenesis. Blood island formation was reduced by half in hemizygous mutant VEGF cultures and by 10-fold in homozygous mutant VEGF cultures. Homozygous mutant cultures could be partially rescued by the addition of exogenous VEGF. RNA levels for the endothelial adhesion receptors ICAM-2 and PECAM were reduced in homozygous mutant cultures, but ICAM-2 RNA levels decreased substantially, whereas PECAM RNA levels remained at hemizygous levels. The quantitative data correlated with the antibody staining patterns because cells that were not organized into vessels expressed PECAM but not ICAM-2. These PECAM+ cell clumps accumulated in mutant cultures as vessel density decreased, suggesting that they were endothelial cell precursors blocked from maturation. A subset of PECAM+ cells in clumps expressed stage-specific embryonic antigen-1 (SSEA-1), and all were ICAM-2(-) and CD34(-), whereas vascular endothelial cells incorporated into vessels were PECAM(+), ICAM-2(+), CD34(+), and SSEA-1(-). Analysis of flk-1 expression indicated that a subset of vascular precursor cells coexpressed PECAM and flk-1. These data suggest that VEGF signaling acts in a dose-dependent manner to affect both a specific differentiation step and the subsequent expansion of endothelial cells. (Blood. 2000;95:1979-1987)

Developmental platelet endothelial cell adhesion molecule expression suggests multiple roles for a vascular adhesion molecule

Platelet endothelial cell adhesion molecule (PECAM) is used extensively as a murine vascular marker. PECAM interactions have been implicated in both vasculogenesis and angiogenesis. To better understand the role of PECAM in mammalian development, PECAM expression was investigated during differentiation of murine embryonic stem (ES) cells and in early mouse embryos. Undifferentiated ES cells express PECAM, and as in vitro differentiation proceeds previously unidentified PECAM-positive cells that are distinct from vascular endothelial cells appear. PECAM expression is gradually restricted to endothelial cells and some hematopoietic cells of differentiated blood islands. In embryos, the preimplantation blastocyst contains PECAM-positive cells. PECAM expression is next documented in the postimplantation embryonic yolk sac, where clumps of mesodermal cells express PECAM before the development of mature blood islands. The patterns of PECAM expression suggest that undifferentiated cells, a prevascular cell type, and vascular endothelial cells express this marker during murine development. PECAM expression in blastocysts and by ES cells suggests that PECAM may function outside the vascular/hematopoietic lineage.

Expression trapping: identification of novel genes expressed in hematopoietic and endothelial lineages by gene trapping in ES cells

We have developed a large-scale, expression-based gene trap strategy to perform genome-wide functional analysis of the murine hematopoietic and vascular systems. Using two different gene trap vectors, we have isolated embryonic stem (ES) cell clones containing lacZ reporter gene insertions in genes expressed in blood island and vascular cells, muscle, stromal cells, and unknown cell types. Of 79 clones demonstrating specific expression patterns, 49% and 16% were preferentially expressed in blood islands and/or the vasculature, respectively. The majority of ES clones that expressed lacZ in blood islands also expressed lacZ upon differentiation into hematopoietic cells on OP9 stromal layers. Importantly, the in vivo expression of the lacZ fusion products accurately recapitulated the observed in vitro expression patterns. Expression and sequence analysis of representative clones suggest that this approach will be useful for identifying and mutating novel genes expressed in the developing hematopoietic and vascular systems.

Yolk sac-derived murine macrophage cell line has a counterpart during ES cell differentiation

Macrophages are phagocytic hematopoietic cells involved in several immune processes, but they are also present early in mammalian development and may participate in embryonic tissue remodeling. We have isolated and characterized a cell line, Py-YSA, from the mouse yolk sac. Py-YSA cells have several functional properties of macrophages, including uptake of acetylated low density lipoprotein and phagocytic capability. They express the murine macrophage markers F4/80 and Mac-1, and they express RNA for the c-fms receptor. Their expansion in culture requires fibroblast conditioned medium or exogenous monocyte-colony stimulating factor. Murine ES (embryonic stem) cell cultures that undergo in vitro differentiation recapitulate yolk sac development, and during this process cells arise that express both Mac-1 and F4/80 and morphologically resemble the Py-YSA cells. The kinetics and distribution pattern of the Mac-1+ cells during a time course of ES cell differentiation suggest that they originate in the blood islands, and that they subsequently leave the blood islands and disperse to tissue sites. Both F4/80 and Mac-1 are first expressed in primary cultures from day 9.5 yolk sacs. The Py-YSA cultured cells thus resemble embryonic tissue macrophages by several criteria, and they share a marker profile with a cell type found in yolk sacs and differentiating ES cells. Py-YSA cells will be a useful reagent for further understanding the role of embryonic tissue macrophages in development.

Murine endothelial cells support fetal liver erythropoiesis and myelopoiesis via distinct interactions

Endothelial cells are an important component of the haemopoietic microenvironment. To investigate how endothelial cells are involved in haemopoiesis, two established murine endothelial cell lines were assayed in stromal cultures with fetal liver haemopoietic cells. Both endothelial cell lines allowed for the proliferation and differentiation of erythroid and monocyte-macrophage precursors, suggesting that support for haemopoiesis is a general property of endothelial cells. Erythropoiesis was dependent on the addition of erythropoietin (Epo), whereas myelopoiesis was independent of added Epo. Haemopoietic colonies developed in close contact with the endothelial cells. Erythroid colonies did not develop when transwell filters were used between the stroma and haemopoietic cells, or when conditioned medium was used in place of stromal cells. In contrast, monocyte-macrophage colonies formed in the presence of transwell filters or conditioned medium. Thus close cell contact is necessary for erythropoiesis but not myelopoiesis under these conditions. These results suggest that different molecular mechanisms are used by endothelial cells to support erythroid development and myeloid development in the mouse fetal liver.

Vascular endothelial growth factor-toxin conjugate specifically inhibits KDR/flk-1-positive endothelial cell proliferation in vitro and angiogenesis in vivo

Inhibition of tumor neovascularization has profound effects on the growth of solid tumors. An endothelial cell-specific cytotoxic conjugate was prepared by chemically linking recombinant vascular endothelial growth factor (VEGF165) and a truncated diphtheria toxin molecule (DT385). The treatment of subconfluent cultures of human umbilical vein endothelial cells and human microvascular endothelial cells with the VEGF165-DT385 conjugate resulted in a selective, dose-dependent inhibition of growth. Parallel experiments with either the free toxin or a mixture of VEGF and the toxin polypeptide did not affect proliferation (DNA synthesis) of these cells. The selective cytotoxicity correlated with the appropriate receptor expression (KDR/flk-1 positive) on the target cells. VEGF-toxin conjugate inhibited the growth of a murine hemangioma-derived endothelial cell line (Py-4-1), which was positive for flk-1 expression. Under similar conditions, the conjugate did not affect the proliferation of a receptor-negative ovarian cancer cell line in vitro. In an in vivo model of angiogenesis, the VEGF165-DT385 conjugate blocked basic fibroblast growth factor-induced neovascularization of the chick chorioallantoic membrane. These studies demonstrate the successful targeting of a cytotoxic polypeptide to proliferating vascular endothelial cells (normal and tumorigenic) and the potential utility of such conjugates in blocking tumor neovascularization.

Blood island formation in attached cultures of murine embryonic stem cells

Differentiation of murine embryonic stem cells in suspension culture results in the formation of cystic embryoid bodies that develop blood islands. In this study pre-cystic embryoid bodies were attached to a substratum, and the program of differentiation was monitored. The attached ES cell cultures formed blood islands on a cell layer that migrated out from the center of attachment and beneath a mesothelial-like cell layer. Morphological and in situ marker analysis showed benzidine-positive hematopoietic cells surrounded by vascular endothelial cells that expressed PECAM and took up DiI-Ac-LDL. Waves of morphological differentiation were evident, suggesting a graded response to differentiation signals. Electron microscopy of the blood islands showed that they were similar to blood islands of cystic embryoid bodies and mouse yolk sacs, and cell-cell junctions were evident among the blood island cells. RNA expression analysis was consistent with the presence of hematopoietic precursor cells of several lineages and a primitive vascular endothelium in the cultures. Thus a program of vascular and hematopoietic development can be elaborated in attached ES cell cultures, and these blood islands are accessible to experimental manipulation.

Structure and evolution of the human IKBA gene

IkappaBalpha belongs to a gene family whose members are characterized by their 6-7 Ankyrin repeats, which allow them to interact with members of the Rel family of transcription factors. We have sequenced a human IkappaBalpha genomic clone to determine its gene structure. The human IkappaBalpha gene (IKBA) has six exons and five introns that span approximately 3.5 kb. This genomic organization is similar to that of other members of the Ankyrin gene family. The human IKBA gene shares similar intron/exon boundaries with the human BCL3 and NFKB2 genes, which is consistent with their conserved Ankyrin repeats. To examine further the evolutionary relationship between human IkappaBalpha and other members of its gene family, we performed a phylogenetic analysis. Although the resulting phylogenetic tree does not identify a common ancestor of the IkappaBalpha and other members of its gene family, it indicates that this family diverges into two groups based on structure and function.

Expression and inducibility of vascular adhesion receptors in development

ICAM-1 and VCAM are cell adhesion receptors expressed on vascular endothelial cells, and their expression is up-regulated as part of the inflammatory response. Tumor-derived Py-4-1 cells were a source of murine endothelial cells, and they showed increased lymphocyte binding when incubated with TNF alpha. Py-4-1 cells stimulated with TNF alpha or LPS also had elevated levels of ICAM-1 and VCAM RNAs, with maximum levels at 2 h. Developmental expression of adhesion receptors was assayed in murine yolk sacs and in cystic embryoid bodies (CEBs) that differentiate in vitro from murine embryonic stem cells. ICAM-1 and VCAM RNAs were expressed in unstimulated yolk sacs and CEBs. Expression was inducible in CEBs by exposure to LPS at all developmental stages, beginning at day 6. The time course of RNA induction in CEBs was similar to that of the Py-4-1 endothelial cell line. ICAM-1 protein expression was localized to vascular and hematopoietic blood islands in the CEBs. These results show that embryonic cells respond to inflammatory mediators by up-regulating expression of vascular adhesion receptors, and they imply an early ontogeny for the endothelial cell signal transduction pathway necessary for leukocyte recruitment.

Endocrine and metabolic characteristics of polyoma large T transgenic mice that develop ACTH-producing pituitary tumors

The authors have analyzed several endocrine and metabolic parameters in polyoma large T transgenic mice (PyLT-1) that develop adrenocorticotropic hormone (ACTH)-immunoreactive pituitary tumors and in nontransgenic mice with tumor transplants. All clinically ill PyLT-1 mice (13 to 16 months of age) had pituitary macroadenomas and elevated plasma ACTH levels. Compared to PyLT-1 transgenic mice, the ACTH plasma concentrations in immunocompetent mice with transplant tumors were markedly raised. In these animals, a secondary effect of hypercorticotropism was documented by a moderate hyperglycemia. Furthermore, mice with transplant tumors had a pathological weight increase from the time the tumor was palpable. The present study supports and extends the authors' previous morphological documentation of the similarity between the ACTH-producing tumors in this mouse model and human Cushing's disease.

Lens expression of TGF alpha in transgenic mice produces two distinct eye pathologies in the absence of tumors

Transgenic mice carrying the rat TGF alpha minigene linked to the alpha A crystallin promoter were generated to investigate the effects of expression of this growth factor in the lens of the eye. All transgenic mice exhibited eye abnormalities in the absence of overt tumors, and two distinct and heritable phenotypes were observed. Five lineages produced 'squinting' transgenic mice characterized by microphthalmic eyes with severe lens and retinal dysplasia, and four lineages produced 'bulged' transgenic mice with eyes that exhibited enlarged globes, lens epithelial hyperplasia, anterior segment dysgenesis, and in some cases retinal dysplasia. The eye perturbations of both phenotypes were evident histologically by 1 week of age, and the eyes of squinting mice were abnormal during embryonic development. The squinting phenotype was dominant over the bulged phenotype in intercrosses, suggesting that position effects from the transgene integration site resulted in differences in TGF alpha expression between the two phenotypes. In situ hybridization showed that TGF alpha transgene expression was confined to the lens or lens rudiment of all transgenic eyes despite the involvement of non-lenticular tissues in the pathology. These results show that inappropriate expression of TGF alpha in the eye can disrupt the communication required to coordinate normal eye development.

Perturbations in the fibrinolytic pathway abolish cyst formation but not capillary-like organization of cultured murine endothelial cells

Fibrinolytic activity and its relation to morphogenesis was investigated in several transformed murine endothelial cell lines and primary cultures of endothelial cells. Two in vitro systems, fibrin gels and Matrigel (Collaborative Research, Bedford, MA), were used. Fibrin gels model a fibrin-rich extracellular matrix that frequently supports neovascularization in vivo, and Matrigel models the basement membrane surrounding quiescent endothelial cells in vivo. The transformed endothelial cell lines have higher levels of plasminogen activator (PA) mRNA than primary cultures of endothelial cells, and an increased PA-mediated proteolytic activity was correlated with formation of cysts in fibrin gels. Addition of neutralizing anti-urokinase antibodies, plasminogen depletion, or addition of a plasmin inhibitor prevented cyst formation. Addition of plasminogen restored the ability to form cysts in the plasminogen-depleted system. Normal endothelial cells organized into capillary-like structures in fibrin gels regardless of manipulations affecting the fibrinolytic pathway. In Matrigel, both transformed and primary cultures of endothelial cells rapidly formed a capillary-like network that was not affected by plasminogen depletion or addition of plasmin inhibitors. Thus, elements of the fibrinolytic pathway necessary for cyst formation are not critical in capillary-like structure formation on a reconstituted basement membrane. These results suggest that plasmin is essential for hemangioma formation but is not critical to the organizational behavior of normal endothelial cells.

Mice with hemangiomas induced by transgenic endothelial cells. A model for the Kasabach-Merritt syndrome

Inoculation of an established endothelial cell line from transgenic mouse hemangiomas (Py-4-1) into histocompatible mice induced vascular tumor formation at the site of infection with 100% frequency. Histological and hematological studies revealed that the mice developed hemangiomas with hematological changes similar to those found in the Kasabach-Merritt syndrome in humans, including hemolytic anemia and thrombocytopenic purpura. Modifications in the red blood cell count, hemoglobin concentration, hematocrit, and platelet count were directly correlated with the size of the hemangioma. Thus transgenic endothelial cell injection into histocompatible mice provides an in vivo model system to study the pathobiology of hemangiomas as well as the investigation of angiogenesis inhibitors.

Developmental analysis of bone tumors in polyomavirus transgenic mice

BACKGROUND

Transgenic mice carrying the polyoma (Py) early region gene develop both vascular and bone tumors that express the transgene (R. Wang and V. L. Bautch, J. Virol. 65:5174-5183, 1991). To determine the correlation between bone tumor formation and transgene expression, mice of two lines that showed differences in tumor pattern were analyzed.

EXPERIMENTAL DESIGN

Py DNA encoding the early region gene was inserted into B6D2F1 mouse embryos. Transgenic mice were sacrificed at appropriate times in development, and the histopathology of the skulls and genotype analyses were performed independently. Py transgene expression was assayed in individual skulls or small groups of skulls.

RESULTS

In mice of the Py-4 line, Py oncogene expression was detectable in skulls from 14.5 days of development. These mice had normal skulls at birth, and they developed lesions histologically resembling osteosarcoma synchronously 1 to 2 weeks after birth. Mice of the Py-3 line did not express detectable levels of transgene at 3 weeks of age, and they did not develop detectable skull pathology at this age. Less than 20% of adult Py-3 mice developed calvarial osteosarcoma-like lesions, and a similar number of these mice expressed the transgene in the skull.

CONCLUSIONS

These data show that Py transgene expression in bone of the skull is differentially regulated in mice of the two lines, and they indicate that epigenetic changes may be sufficient to complement expression of the Py oncogene in neoplastic bone formation during development.

[Transgenic mice as animal model of Cushing's syndrome]

Transgenic mice were generated with the polyoma early region promoter linked to cDNA encoding polyoma large T antigen (PyLT). Light microscopic examination showed up to 5 mm large pituitary adenomas in clinically ill transgenic mice. The tumour cells showed positive ACTH immuno-reactivity. The adrenal glands of clinically ill mice showed an increase in weight and exhibited medullary hyperplasia. These findings were unexpected, and might have been caused by transgene expression in the neuroendocrine cells of the adrenal medulla. Plasma ACTH measurements showed significantly increased levels in clinically ill PyLT-1 transgenic mice.

ACTH-producing pituitary tumors in transgenic mice. An ultrastructural and immunoelectron microscopic study

Ultrastructural and electron microscopic immunohistochemical features of corticotropic pituitary tumors arising in polyoma large T transgenic mice and the corresponding tumor transplants in non-transgenic mice are reported. Spherical, irregular and drop-shaped secretory granules measuring 150-450 nm in diameter, were seen in all tumors. Both in tumors from transgenic mice and in tumor transplants immunoreactivity for adrenocorticotropic hormone (ACTH) and beta-endorphin was expressed in the majority of the secretory granules, whereas growth hormone (GH) immunoreactivity was demonstrated only in a small number of cells in tumors from transgenic mice. In addition, positive immunostaining for neuron-specific enolase (NSE) and synaptophysin was found in the two pituitary tumor transplants tested. The study shows that the pituitary tumors from transgenic mice and their tumor transplants have features similar to human corticotropic pituitary tumors, and may therefore be a valuable model for experimental studies of the tumorigenesis.

Transgenic mice that develop pituitary tumors. A model for Cushing's disease

Transgenic mice that developed adrenocorticotropic hormone (ACTH)-producing pituitary tumors were generated with the polyoma early region promotor linked to a cDNA encoding polyoma large T antigen (PyLT). Light microscopic examination of the pituitaries showed normal morphology at 4 months of age, either unremarkable morphology or microadenoma formation at 9 months of age, and up to 5 mm large adenomas in clinically ill transgenic mice at 13-16 months of age. At age 9 months, transgenic mice weighed significantly more than corresponding control mice, but they began wasting at approximately 1 year of age. The adrenal glands of these older PyLT-1 mice showed a weight increase and exhibited a medullary hyperplasia. Subcutaneous transplants of transgenic pituitary tumors to nontransgenic, immunocompetent mice resulted in tumors with a morphology and ACTH immunoreactivity similar to the primary tumor. The effects of hypercorticotropism were more enhanced and occurred with a shorter latency in the mice carrying transgene pituitary transplants than in the PyLT-1 transgenic mice themselves. Moreover, these transplanted mice showed a weight increase with an axial deposition pattern and hypertrophy of the adrenal cortex that resembled the findings in human Cushing's disease. Plasma ACTH levels were significantly increased in clinically ill transgenic mice and even higher levels were found in the transplant mice. Thus, both murine models should be useful for studying Cushing's disease.

Embryonic stem cell-derived cystic embryoid bodies form vascular channels: an in vitro model of blood vessel development

Murine embryonic stem cells can differentiate in vitro to form cystic embryoid bodies (CEB) that contain different structures and cell types. The blood islands are one such structure that consist of immature hematopoietic cells surrounded by endothelial cells, the first identifiable vascular cells. CEBs differentiated in vitro developed blood islands initially, and subsequently these blood islands matured to form vascular channels containing hematopoietic cells. Phase contrast microscopy demonstrated the presence of channels in mature CEBs grown in suspension culture, and high resolution light and electron microscopy showed that the cells lining these channels were endothelial cells. The channels appeared less organized than the vasculature of the mature yolk sac. The hematopoietic cells were occasionally seen ‘flowing’ through the CEB channels, although their numbers were reduced relative to the yolk sac. Analysis of primary CEB cultures showed the presence of cells with two characteristics of endothelial cells: approximately 30% of the cells labelled with fluorescent acetylated low density lipoprotein and a small number of cells were positive for von Willebrand's factor by immunostaining. Thus we conclude that a primitive vasculature forms in CEBs differentiated in vitro, and that not only primary differentiation of endothelial cells but also some aspects of vascular maturation are intrinsic to this cell culture system. CEBs are therefore a useful model for the study of developmental blood vessel formation.

Isolation and characterization of an established endothelial cell line from transgenic mouse hemangiomas

A murine endothelial cell line was isolated from hemangiomas induced by expression of the polyoma early region gene in transgenic mice. After two cell sortings using acetylated low-density lipoprotein with a fluorescent label (Dil-Ac-LDL), a pure population of endothelial cells has been carried for more than 60 passages from the animal. The cells retain endothelial cell properties such as a characteristic cobblestone appearance at confluency, contact-inhibited growth, and active uptake of Ac-LDL. Expression analysis shows that the cells express both the polyoma transgene and the von Willebrand factor, an endothelial cell marker. Subcutaneous injection of the cultured endothelial cells into nontransgenic histocompatible mice or nude mice led to hemangioma formation, and endothelial cells were re-isolated by cell sorting from these secondary hemangiomas. This cell line represents a renewable source of murine endothelial cells derived from transgenic mice that can be studied both in vitro and by reintroduction into a host.

The polyomavirus early region gene in transgenic mice causes vascular and bone tumors

Transgenic mice carrying the entire polyomavirus (Py) early region consistently develop both vascular and bone tumors. This tumor spectrum represents a subset of the tumors found in mice infected with Py and an expansion of the vascular tumor spectrum seen in Py middle T antigen (MT) transgenic mice (V. L. Bautch, S. Toda, J. A. Hassell, and D. Hanahan, Cell 51:529-538, 1987). Transgenic mice of three independent lineages develop these pathologies, and mice of individual lineages also develop lymphangiomas and fibrosarcomas. All tumors are of mesenchymal origin, and all tumor tissues express the Py transgene. Some unaffected tissues, including the testes of mice of all lineages, also express the Py transgene. The number of transgene expression sites in mice of a given lineage correlates with the severity and latency of the tumor phenotype in these animals. Analysis of transgene transcripts indicates that RNAs for Py large T antigen (LT), MT, and small T antigen (ST) are present in both tumors and testes of transgenic mice. The ratio of LT RNA to MT and ST RNAs, however, is higher in testes than in tumors and other unaffected tissues, indicating that tissue-specific differences in the relative amounts of the alternatively spliced Py RNAs exist. The finding that some sites of Py transgene expression are susceptible to tumor formation while other expressing tissues such as testes are refractory to tumorigenesis suggests that mechanisms of viral pathogenesis are influenced by a tissue specificity in the effects of the Py early region.

Early invasiveness characterizes metastatic carcinoid tumors in transgenic mice

The conversion of a normal cell into a metastatic tumor is thought to occur in a stepwise progression of genetic changes that affect both growth control and interactions with the extracellular environment. The development of invasiveness allows tumor cells to escape from their primary site. We have investigated transgenic mice that develop both invasive intestinal neuroendocrine tumors and noninvasive tumors of the pancreatic beta-cells. Visual inspection and gene expression studies indicate that the beta-cell tumors rarely metastasize. In contrast, intestinal tumors that first appear in submucosal areas metastasize with high frequency to the lymph nodes and liver. No evidence of preneoplastic mucosal lesions was seen in the intestine, indicating that invasiveness is acquired early in the tumorigenic progression of these cells. Comparison of intestinal and pancreatic neuroendocrine tumors in transgenic mice suggests that an early requirement for invasiveness may contribute to metastatic potential.

Development of neuroendocrine tumors in the gastrointestinal tract of transgenic mice. Heterogeneity of hormone expression

Expression of hormones in endocrine tumors and derived cell lines of transgenic mice carrying insulin-promoted oncogenes has been investigated by histochemical, immunohistochemical, ultrastructural, and radioimmunologic means. Tumors of the pancreas, small intestine, mesentery, and liver were examined. Insulin-immunoreactive cells were prevalent in pancreatic tumors, with a significant subpopulation of pancreatic polypeptide-immunoreactive elements. Conventional ultrastructural and immunogold analysis identified insulin-storing beta granules in pancreatic tumor cells. In contrast, the largest immunoreactive subpopulation of intestinal tumors expressed secretin (53% of total cells), followed by proglucagon-related peptides (15%), glucose-dependent insulinotropic polypeptide (7%), gastrin (7%), pancreatic polypeptide (2%), neurotensin (2%), and somatostatin (1%). No detectable immunoreactivity for either insulin or serotonin was observed. Electron microscopy and immunogold labeling showed that intestinal tumor cells contained secretin-storing S-type granules. Lymph node and liver tumors contained secretin-immunoreactive cells with ultrastructural features similar to those of intestinal tumors. In addition, high levels of circulating insulinlike and secretinlike immunoreactants were detectable. Analogous hormone profiles were identified in tumor cell lines and culture media. Large T-antigen immunoreactivity was detected in all the nuclei of neoplastic cells, as well as in insulin-immunoreactive elements of non-neoplastic islets and pancreatic ducts and in some secretin-immunoreactive cells of small intestinal mucosa. These data indicate that neuroendocrine tumors arise both in beta cell and S-cell subpopulations of transgenic mice.

Effects of polyoma virus oncogenes in transgenic mice

The effects of polyoma (Py) oncogenes in transgenic mice are reviewed, with emphasis on an analysis of mice carrying individual Py early region genes linked to either the Py early region promoter or to the rat insulin II promoter. The perturbations documented in this series of animals varied with both the Py gene and the promoter. Py promoter-driven transgenes led to the development of pituitary tumors and vascular tumors in a gene-dependent manner, suggesting that the Py T antigens have different tissue specificities of transforming activity. The only pathology found in insulin promoter-driven transgenic mice was a beta cell tumor seen in insulin-promoted Py large T antigen-carrying mice. These results also suggest that the Py oncogenes have different tropisms relative to a specific cell type. Analysis of the Py oncogene-induced perturbations may provide both insights into the molecular events controlling tumorigenesis and access to rare or dispersed cell types for further analysis.

Endothelial cell tumors develop in transgenic mice carrying polyoma virus middle T oncogene

Inoculation of newborn mice with the murine polyoma (Py) virus leads to tumor formation in a wide range of tissues. In order to investigate viral oncogenesis, we generated transgenic mice carrying either the Py large T antigen (LT) gene or the Py middle T antigen (MT) gene linked to Py early region regulatory sequences. While Py LT mice exhibit no phenotype, Py MT mice develop multifocal tumors of the vascular endothelium. These hemangiomas are lethal to the animals and can be passaged in vivo. Transgene RNAs and protein are present in both hemangiomas and the testes of these mice, and the Py middle T protein in both tissues is complexed to a cellular tyrosine kinase. The expression of complexed middle T protein in both tumorigenic endothelial cells and unperturbed testes implies that endothelial cells may be particularly susceptible to the action of the middle T oncogene. These observations indicate that Py middle T disrupts the normal strict controls on vascular growth, and suggest that Py MT transgenic mice will provide a model for studying the control of angiogenesis.

Deletion and duplication of DNA sequences is associated with the embryonic lethal phenotype of the t9 complementation group of the mouse t complex

We have analyzed the genomic structure of three mouse t haplotypes of the t9 complementation group. Each of these t haplotypes, tw18, t4, and tks1, is known to have resulted from a rare recombination event between a complete t haplotype and a wild-type chromosome. Using molecular probes that identify sequences in the distal portion of the t complex, we have shown that each of these t haplotypes contains a similar (perhaps identical) deletion of one group of t complex sequences, and duplication of another group. These data suggest that the recombination events that produced these three t haplotypes involved similar unequal crossovers within the distal inversion. The deletion and duplication of genetic material associated with all members of the t9 complementation group tested provides a molecular explanation for the recessive lethal mutation associated with these t haplotypes.

Genetic background affects integration frequency of ecotropic proviral sequences into the mouse germ line

Germ line acquisition of ecotropic proviruses occurs at a high frequency in the progeny of SWR/J-RF/J hybrid mice carrying two genetically linked RF/J ecotropic proviral loci, Emv-16 and Emv-17 (N. A. Jenkins and N. G. Copeland, Cell 43:811-819, 1985). To determine if genetic background affects proviral integration frequency, I analyzed a series of crosses in which the two RF/J proviral loci were transferred onto different provirus-negative background strains. Unlike SWR/J-RF/J hybrid progeny, few CBA/CaJ-RF/J hybrid mice were identified that carried new germ line proviral loci. These results indicate that genetic factors other than the linked RF/J proviral loci contribute to the increased frequency of germ line provirus integration seen in the SWR/J-RF/J hybrids. The frequency of proviral acquisition appeared to increase when females carrying Emv-16, Emv-17, and at least one new proviral locus were further backcrossed, suggesting that integration frequency can be increased by genetic manipulation. The breeding data are consistent with the hypothesis that virus from the mother infects the egg or the early embryo. Analysis of the transmission frequency and cosegregation patterns of new proviral loci indicated that viral integration occurs after the first round of DNA replication and before the germ line is set aside during embryogenesis, with a majority of viral integrations occurring at the two-cell stage of development, and independent viral integrations can occur in the same or in different cells of the embryo.

Identification of a cytoplasmic tropomyosin gene linked to two muscle tropomyosin genes in Drosophila

A Drosophila cytoplasmic tropomyosin gene has been identified and is located on the same genomic DNA clone as two Drosophila muscle tropomyosin genes previously identified. A positive hybrid-selection translation assay using the subcloned gene and RNA from non-muscle cell sources yielded a protein with a size (Mr, 31,000) and isoelectric point (5.0) similar to vertebrate cytoplasmic tropomyosin. A modified protocol for the purification of vertebrate cytoplasmic tropomyosin was used to partially purify cytoplasmic tropomyosin from the Drosophila Kc cell line. The Kc cell protein was identified as a cytoplasmic form of tropomyosin on the basis of its size, isoelectric point, and crossreactivity with a polyclonal vertebrate antitropomyosin antibody in a two-step binding assay. The Kc cell cytoplasmic tropomyosin comigrates in two dimensions with the hybrid-selected in vitro translation product of the region 3 gene, and both proteins show a mobility shift in NaDodSO4/urea/polyacrylamide gels that is characteristic of vertebrate tropomyosins. The cytoplasmic tropomyosin gene hybridizes to both Drosophila muscle tropomyosin genes under decreased stringency conditions. This cross-hybridization spans several internal restriction endonuclease sites in each muscle tropomyosin gene and indicates an overall partial homology among the three Drosophila tropomyosin genes. These results show that Drosophila tropomyosins are encoded by a closely linked family of differentially regulated genes.