Cell Tissue Res (2003) 314:33–42 DOI 10.1007/s00441-003-0739-8

REVIEW

Ganka Nikolova · Eckhard Lammert

Interdependent development of blood vessels and organs

Received: 8 April 2003 / Accepted: 24 April 2003 / Published online: 1 August 2003
Springer-Verlag 2003

Abstract The cardiovascular system is the first functional organ in the vertebrate embryo, and many organs start to develop adjacent to cells of the cardiovascular system. Endothelial cells (EC) form the inner cell lining of blood vessels and represent the major cell type that interacts with developing organs. On the one hand, EC provide organs with signals. These signals determine the location, differentiation and morphology of an organ. On the other hand, EC receive signals from the organ-specific cell types. Such signals give EC organ-specific features that the organ needs to interact with the circulatory system. This review provides the reader with specific examples of an interdependent development of organs and blood vessels. Keywords Endothelial cells · Organogenesis · Mouse development · Pancreas · Mutual signaling

Introduction The formation of the cardiovascular system begins when hemangioblasts develop from mesodermal cells in the E7.5 mouse embryo. At E8.0 the paired dorsal aortae are visible as endothelial tubes on each side of the notochord. The heart starts to pump shortly thereafter. At E9.0 the heart beats regularly and strongly and it forms together with the major blood vessels the first functioning organ system in the embryo. The development of the other inner organs is initiated shortly after the cardiovascular organ system has been established at E9.0 of mouse development: the liver bud Eckhard Lammert and Ganka Nikolova were supported by the Deutsche Forschungsgemeinschaft DFG (La1216/2–1) G. Nikolova · E. Lammert ()) Max Planck Institute of Molecular Cell Biology and Genetics, Pfotenhauerstraßsse 108, 01307 Dresden, Germany e-mail: [email protected] Tel.: +49-351-2102777 Fax: +49-351-2101289

develops from the ventral gut endoderm (Zaret 2002); the dorsal and ventral pancreases bud from the dorsal and ventrolateral foregut endoderm, respectively (Edlund 2002; Lammert et al. 2003); the lung diverticulum forms from the more anterior ventral foregut endoderm (Hogan and Yingling 1998); and the kidney cords develop from the intermediate plate mesoderm (Kreidberg et al. 1993). Thus, organ development takes place in the presence of either major embryonic blood vessels (such as the dorsal aortae) or endothelial cells (EC) interspersed in the adjacent mesoderm. In the case of the embryonic gut endoderm, the interaction of blood vessels with endoderm happens at a stage when the endoderm has already been pre-patterned. The word "pre-pattern" implies that the endoderm has been divided into domains, which differ in their response to developmental signals such as EC signals. The homeobox genes are good markers for a prepatterned endoderm (Grapin-Botton and Melton 2000). They are expressed in domains along the anteriorposterior axis of the embryonic gut tube, e.g., Pax9 is expressed in more anterior gut regions, from which the esophagus, lung and thyroid gland develop, while Pdx1 is expressed in the region where the pancreas, duodenum and posterior part of the stomach develop, and Hoxa-13 and Hoxd-13 are expressed in the posterior gut region that becomes the large intestine. Pre-patterning embryonic gut endoderm takes place before blood vessels appear. Signals from mesoderm/ectoderm start to establish an anterior-posterior pattern as early as E7.5 of mouse development (Wells and Melton 2000). The pre-pattern of the gut tube continues at E8.5, when cardiac mesoderm initiates hepatic differentiation in the ventral endoderm and the notochord initiates pancreatic development in the dorsal endoderm (Biemar et al. 2001; Douarin 1975; Fukuda-Taira 1981; Gualdi et al. 1996; Hebrok et al. 1998; Kim et al. 1997). A common feature of tissues that induce a pre-pattern is that they often separate from the developing organ. For example, cardiac mesoderm is separated from hepatic endoderm by septum transversum mesenchyme and the notochord is separated from the

34 Fig. 1A–C Blood vessel-organ interactions. Three examples of blood vessel-organ interactions are presented: A pancreatic development; B glomerulus development; and C hepatic development. In each case blood vessels play a critical role in inducing developmental steps during organ formation. The embryonic stages are indicated for mouse organ development. A The aorta endothelium is adjacent to the developing dorsal pancreatic endoderm. The endothelium induces budding and endocrine development. Blood vessels residing in the pancreatic mesenchyme continue to interact with the adjacent pancreatic epithelium throughout development (red vascular endothelium, blue pancreatic endoderm, light blue endocrine cells, yellow nonpancreatic gut endoderm). B Endothelial cells migrate into the assembling kidney nephron to form the glomerulus. The blood flow in the ingrowing capillary induces podocyte coalescence, which results in glomerulus formation (red capillary endothelium, orange podocytes). C Endothelial cells are located next to the hepatic endoderm. They induce proliferation and migration of hepatocytes into the surrounding mesenchyme, the septum transversum (red endothelial cells, brown hepatocytes, yellow nonhepatic gut endoderm)

dorsal pancreatic endoderm by the dorsal aorta. In contrast, blood vessels continue to interact with developing organs throughout embryonic and postnatal development. Blood vessels and organ-specific cells therefore interact with each other continuously. This interaction is mutual in that blood vessels (or vascular EC in particular) and organ-specific cells exchange signals. This coordinated development of EC and organ-specific cell types generates a functional organ with an endothelium adjusted to the needs of the adjacent tissue cells. Indeed, the microvascular endothelium of different organs can be distinguished morphologically and biochemically. For example, an endocrine cell that secretes hormones into the blood is adjacent to a permeable fenestrated endothelium (Bonner-Weir and Orci 1982), while the endothelium is continuous and non-permeable in large parts of the brain for protective reasons (blood-brain barrier) (Rubin and Staddon 1999).

In the following sections, specific examples of endothelial induction steps in organ formation will be presented (Figs. 1, 2) followed by examples of organspecific endothelial features (Figs. 3, 4). Although our research focuses on the vascular-pancreatic interactions, we will include examples of other organs to provide the reader with a bigger picture. For more detailed information on the EC of other organs, we would like to refer the reader to some excellent reviews (Braet and Wisse 2002; Rubin and Staddon 1999; Schnitzer 2001).

Early pancreatic and vascular development First signs of pancreatic development appear at sites in foregut endoderm located next to major embryonic blood vessels (Figs. 1A, 2A). The expression of the pancreasduodenum homeobox gene Pdx1 is initiated in the dorsal endoderm next to the fusing dorsal aortae and in two

35

Fig. 2A–C Vascular and pancreatic development coincide. Mouse pancreatic development is initially symmetric with three buds forming adjacent to the dorsal aorta and two vitelline veins. Only one of the two ventral pancreatic buds develops into pancreatic tissue, whereas the other bud regresses, coinciding with a developing asymmetry of the vitelline veins. The ventral bud adjacent to the endothelium of the right vitelline (or portal) vein continues to grow and develop, whereas the left ventral bud regresses together with the left vitelline vein. The top panel shows a schematic of blood vessel-gut endoderm interactions at mouse embryonic stages

E8.5 (A), E9.5 (B) and E10.5 (C) (yellow gut endoderm, red blood vessel endothelium, blue pancreatic endoderm). The intermediate panel shows whole-mount images of the respective stages. Pdx1LacZ mouse embryos were cleared to see the pancreatic buds (blue Pdx1 expression, dp dorsal pancreas, vp ventral pancreas). The lower panel shows sections through the respective stages of a Pdx1LacZ mouse embryo stained for LacZ and cell nuclei (hematoxylin) (blue Pdx1 expression, red dots cell nuclei, A aorta, V vitelline, or omphalomesenteric, veins). Chris Wright, Vanderbilt University, kindly provided the Pdx1-LacZ mice (see Offield et al. 1996)

ventrolateral endoderm regions next to the two vitelline (or omphalomesenteric) veins (Fig. 2A). After this direct interaction with the adjacent blood vessel endothelium, the endoderm becomes embedded in mesenchymal tissue (Slack 1995). At this time, the three Pdx1-expressing domains start to form buds (Fig. 2B). Pancreatic buds form by accumulating epithelial cells and are distinct from the rest of the endoderm, which remains an epithelial monolayer. The dorsal bud and two ventrolateral buds are initially arranged almost symmetrically (Fig. 2A, B). However, when the vascular system becomes asymmetric at E9.5–10.5, the arrangement of the pancreatic buds changes coincidentally. The ventrolateral bud adjacent to the endothelium of the expanding right vitelline vein (or “portal vein”) starts to develop into the permanent ventral pancreas, while the other bud regresses together with the adjacent left vitelline vein (Fig. 2C). In conclusion, pancreatic development and vascular development occurs hand-in-hand (Lammert et

al. 2001). During later stages of pancreatic development, vessels form within the pancreatic mesenchyme, and endocrine cell differentiation from pancreatic epithelium can be observed adjacent to major pancreatic blood vessels. Three sets of experiments were performed to show that blood vessel endothelium influences pancreatic development (Lammert et al. 2001). In the first set, E8.5 dorsal endoderm was recombined with different tissues including vascular endothelium, the notochord, neural tube and somites. It was observed that Pdx1 expression was turned on when endoderm was recombined with either vascular endothelium or notochord. This shows that these tissues, which are in close contact with the developing dorsal endoderm, can induce pancreatic development in contrast to more distant tissues such as neural tube or somites. When looking at other aspects of pancreatic development, it turned out that vascular endothelium was the only tissue investigated that could induce tissue growth and insulin

36

Fig. 3A–E Organ-specific endothelial cell morphology. Depending on the adjacent organ-specific cell types, endothelial cells (EC) can be fenestrated or continuous. The number of caveolae and transport vesicles and the presence or absence of a basement membrane (BM) also distinguishes the endothelial cells (EC) of different tissues. A In pancreatic islets, EC are separated from pancreatic beta cells by a BM. The EC contains numerous fenestrae with a permeable diaphragm (arrowheads). The beta cell possibly releases insulin from the secretory granules (SG) into the blood stream through the basement membrane and endothelial fenestrae. B In many parts of the brain, capillaries are found adjacent to astrocytes. A BM separates EC from astrocytes. The endothelium is continuous without fenestrae and few caveolae and transport vesicles. It is possible that this EC morphology contributes to the blood-brain barrier. C Liver parenchymal cells are located next to a fenestrated

endothelium. Basement membrane and diaphragm of fenestrae are missing. Parenchymal microvilli (MV) reach into the space of Disse (SD) that separates the liver parenchyma from the EC. The fenestration (arrowheads) and space between EC and parenchyma possibly enables the exchange of lipid particles between the liver sinusoidal blood and the parenchymal cells. D Lung epithelium is found next to a continuous endothelium, which is characterized by a high number of caveolae (arrowheads). These caveolae are involved in transport of macromolecules (such as albumin) through the endothelium to the lung epithelium. E The foot processes (FP) of the kidney podocytes are connected with the glomerular BM and a fenestrated endothelium. The fenestrae (arrowheads), the BM and the podocytic foot processes form a sieve, which acts as the filtration barrier between blood plasma and urinary space

expression in dorsal pancreatic endoderm. These in vitro experiments suggested that during embryonic development Pdx1 expression in dorsal endoderm is initiated by signals from the notochord and then upregulated by signals from the dorsal aorta endothelium, which replaces the notochord at E8.5. The endothelium also provides signals necessary for the budding and endocrine differentiation of the dorsal pancreatic endoderm. These experiments point to the stepwise nature of organ induction and show that endothelial cells (EC) can induce certain aspects of organ development. Moreover, the in vitro experiments demonstrated that inductive events could be mediated by vascular endothelium, independent of blood flow or serum factors.

In the second experiment, the EC precursors of the dorsal aorta were removed from frog embryos. Frog embryos survive for some time without aorta and regenerate the aorta, possibly due to EC precursors that migrate to the midline from other sides of the embryo. The aorta-less embryos were shown to be unable to turn on endocrine pancreatic gene expression. However, Pdx1 expression was detectable, most likely as a result of the signals coming from the notochord, which had not been removed during the surgery. The in vivo experiment with the aorta-less frog embryos demonstrated that the aorta as an additional signaling center is required for endocrine pancreatic development. In the third experiment, the vascular endothelial growth factor VEGF-A was expressed under the control

37

of the Pdx1 promoter in transgenic mouse embryos. As a result, EC accumulated next to non-pancreatic endoderm such as the posterior part of the stomach and the duodenum early during embryonic development. The ectopic EC were found next to ectopic insulin-expressing cells in stomach and duodenum, thus indicating that EC can instruct Pdx1-expressing foregut endoderm to adapt a pancreatic cell fate. Moreover, the experiment strongly suggested that the coincidence of vascular and pancreatic development as seen in Fig. 2 is due to instructive endothelial signals, which drive pancreatic development within the pre-patterned foregut endoderm. In conclusion, blood vessel endothelium has a profound effect on diverse aspects of pancreatic development including endocrine differentiation and pancreatic morphogenesis.

Early hepatic and vascular development During liver development, cardiac mesoderm initiates hepatic cell differentiation in ventral endoderm (Douarin 1975; Fukuda-Taira 1981; Gualdi et al. 1996). At the same time, endoderm is required for the development of cardiac mesoderm (Zaret 2002). This demonstrates an early interaction between the liver endoderm and cardiovascular system. After liver specification at E8.5, the septum transversum mesoderm surrounds the liver endoderm, and the liver forms a multilayered epithelium. EC within the mesoderm are located next to the liver and form vessels at the time when cells delaminate from the liver epithelium and grow into the surrounding mesenchyme (Fig. 1C). This sequence of events is reminiscent of pancreatic development: like pancreatic development, liver development starts with an interaction of endoderm with cardiovascular tissue. This is followed by endoderm growth into the surrounding mesoderm, which gives rise to a multilayered epithelium in the case of the liver and a bud in the case of the pancreas. Finally, blood vessels form within the mesoderm next to pancreas and liver and these vessels co-develop with pancreatic and hepatic epithelium (Fig. 1A, C). Two sets of experiments were performed to show that EC play an important role in liver development (Matsumoto et al. 2001). In the first experiment, embryos were analyzed, which were deficient for VEGFR2, the main signaling receptor for the vascular endothelial growth factor, VEGF-A (Shalaby et al. 1995). These mice still harbor endothelial precursors, but the EC are incapable of forming blood vessels. When the mice were analyzed for liver defects, it was observed that liver endoderm forms a multilayered epithelium, but is unable to delaminate and grow into the surrounding septum transversum mesoderm. To exclude the possibility that the observed growth arrest resulted from the absence of serum factors or the early lethality of VEGFR2-deficient embryos, liver buds were explanted and cultured in vitro. Because the growth of VEGFR2-deficient liver buds was also inhibited in vitro when compared with control explants, it was concluded

Fig. 4A, B Endothelial cell morphology in pancreatic islets. A A thin endothelial cell (EC, surrounded by a black line) embraces the capillary lumen (see black line) and separates the blood from the pancreatic beta cells (B). B A higher magnification shows that the fenestrated endothelium along with a thin BM separates the secretory granules (SG) from the blood. Crystalline insulin granules are visible in the secretory granules of the pancreatic beta cell. The fenestrae are cytoplasmic holes, in which a permeable diaphragm is located (arrowheads). Black bars 1 mm. These images were generated in collaboration with Margaret McLaughlin and Dennis Brown, Massachusetts General Hospital and Harvard Medical School, Boston, USA

that endothelial signals were required for early liver morphogenesis. In the second set of experiments, the growth of liver endoderm was shown to be impaired in wild-type liver buds, when vessel formation was inhibited using the angiogenesis inhibitor NK4. This experiment demonstrates that liver development could also be arrested in explants isolated from embryos that had undergone a normal development. In conclusion, liver morphogenesis and growth requires the interaction of liver endoderm with blood vessel endothelium.

38

In mammals, kidney development takes place in three stages: pronephros, mesonephros and metanephros (Kreidberg et al. 1993). The definitive kidney develops from the metanephros at E10.5 of mouse development, and it is preceded by the development of the rudimentary pronephros and the temporary mesonephros. In fish, unlike mammals, the pronephros serves as an excretory organ during embryonic (larval) stages (Drummond 2000). The pronephros therefore has a functional pronephric glomerulus, which performs the essential functions of blood filtration and osmoregulation. In mice, the kidney glomerulus develops when EC start to migrate into the kidney nephron and form capillaries adjacent to the podocytes (Fig. 1B). In the zebrafish mutant cloche the number of EC is reduced. In these fish mutants, differentiation of podocytes proceeds normally, despite the absence of endothelial cells (Majumdar and Drummond 1999). However, podocyte assembly into a glomerulus does not occur without EC. Zebrafish lacking blood circulation were generated and shown to have a similar defect of glomerular assembly despite the presence of EC, suggesting that both EC and blood flow are required for glomerular morphogenesis (Serluca et al. 2002). Specifically, the expression of matrix metalloproteinase-2 (MMP-2) in renal EC depends on blood flow. If MMP2 is inhibited in zebrafish by injection of its inhibitor, tissue inhibitor of metalloproteinase-2 (TIMP-2), glomerulus formation does not take place, even when blood flow is normal. These data therefore suggest that blood flow induces MMP-2 expression in EC. The endothelial MMP2, in turn, enables the assembly of podocytes, possibly by degrading extracellular matrix components (Serluca et al. 2002). In conclusion, these experiments provide a specific example of how EC and their matrix proteases enable organ morphogenesis to take place.

precursors within the adult hippocampus reside within a thin lamina termed the subgranule zone. Proliferative cells reside in tight juxtaposition to the capillaries within the subgranule zone, and their mitogenic recruitment is accompanied by a concurrent angiogenic response. Precursors in other areas of the hippocampus are also proliferative, but they are not found within this vascular niche and do not generate neurons. This has led to the hypothesis that the angiogenic microenvironment (or “vascular niche”) may be important for adult hippocampal neurogenesis and ultimately for hippocampus-mediated cognitive function (Palmer et al. 2000). The idea of a vascular niche that promotes neuronal proliferation and regeneration gained momentum with the investigation of songbirds. Many male temperate-zone songbirds sing at high rates in the spring as compared to other seasons. In these species seasonal differences are correlated with seasonal changes in hormone levels (in particular testosterone levels) in the blood plasma. Coincident with these changes the volume of several song nuclei in the songbird brain changes accordingly (Ball et al. 2002). In particular the higher vocal center (HVC) in the adult songbird brain is largest, when testosterone levels are highest, and cell death in the HVC is associated with low testosterone levels (Nottebohm 1981). It has now been shown that testosterone upregulates the expression of the vascular endothelial growth factor VEGF-A within the HVC (Louissaint et al. 2002). This leads to dramatic EC proliferation. EC, in turn, secrete the neurotrophin BDNF (brain-derived neurotrophic factor), which stimulates migration of the neuroblast and drives the maturation and survival of the newborn HVC neurons. The entire angiogenic process is selfregulated and resolves within 2 weeks, despite the continuous presence of testosterone (Louissaint et al. 2002). In summary, the songbird brain is a specific example of adult tissue regeneration in which blood vessels play a profound role.

Songbirds—a brief intermezzo

Organ-specific endothelial cell development

The embryonic EC-tissue interactions lead to the question of whether these interactions also play a role in the adult, for example during tissue regeneration. It is obvious that during regeneration both vascular and nonvascular cells develop and differentiate. Whether similar or different inductive EC signals play a role during tissue regeneration is unclear at the moment. However, reports on the coordinated interaction of neurons and blood vessels in the brain need to be mentioned as an example of endothelial signals during adult tissue remodeling and regeneration (Louissaint et al. 2002; Palmer et al. 2000). In mammals, the hippocampus is responsible for learning and memory, and neurons residing in the hippocampus therefore have to be continuously remodeled (in part by cell death and growth). Adult neurogenesis is mediated by stem cell-like precursors that remain in the hippocampus. Anatomically, the proliferative

Every organ has a different interaction with the circulatory system. The endothelium of a kidney glomerulus must enable filtration of blood plasma, the endothelium in pancreatic islets must enable secretion of polypeptide hormones such as insulin into the blood, and the liver endothelium must allow the exchange of lipid particles between liver and blood. It therefore is not surprising that blood vessels display organ-specific features. This is true in particular for microvessels, which have the most intimate contact with the surrounding organ-specific cells (Fig. 3). The microvascular endothelium in pancreatic islets consists of thin, fenestrated EC separated from pancreatic beta cells by a basement membrane (BM) (Fig. 3A). Electron microscopy of islet endothelium shows that EC form single-cell tubes lined up to form a capillary (Fig. 4A). Such "seamless" endothelia were first described

Glomerular development and the need of blood flow

39

for brain capillaries (Wolff and Bar 1972). “Seamless” vessels do not form through an endothelial coat that is buttoned up by tight junctions on one side, but instead the vascular lumen perforates the EC. An additional feature of pancreatic EC is the presence of fenestrae. In the case of pancreatic islets, fenestrae are channels that pass through the EC cell body. These channels or pores enable the exchange of molecules between blood plasma and basement membrane (Bearer and Orci 1985; Lombardi et al. 1986). EC next to beta cells are perforated with numerous fenestrae (Fig. 4B). In general, the diameter of fenestrae ranges from 50 to 200 nm. The fenestrae in islets are small fenestrae and always contain a diaphragm (Bearer and Orci 1985). The diaphragm is composed of fibrils that form a permeable mesh. The permeable structure of the islet endothelium and basement membrane (BM) was demonstrated by injection of horseradish peroxidase into the circulatory system of rats (Like 1970). Within 40 s after the injection, the peroxidase was visible in the fenestrae and the entire BM, showing that this protein could easily permeate the islet endothelium and BM. Based on these findings, it has been suggested that fenestrae represent channels through which insulin could be secreted from the beta cell into the blood stream. However, to date a role of endothelial fenestrae for insulin secretion or blood glucose regulation has not been directly demonstrated and it is not known how these fenestrae form. Because islet cells express all VEGF ligands (Inoue et al. 2002), it is possible that some of these VEGF ligands induce fenestrae in the adjacent endothelium to enable fast insulin delivery into the blood stream in response to elevated blood glucose levels. The microvascular endothelium of the brain is significantly different from that described for pancreatic islets (Fig. 4B). The EC in brain tissue are continuous and do not have any fenestrae (Rubin and Staddon 1999). Moreover, the endothelium is characterized by few endocytic vesicles that limit the amount of transcellular flux, and brain EC are coupled by tight junctions that severely restrict the amount of paracellular flux. This combination of features constitutes the blood-brain barrier, which impedes passive entry from blood to brain of virtually all molecules, except those that are small and lipophilic. There are hydrophilic molecules that pass the brain endothelium though, but these molecules are transported by active rather than passive transport. For essential nutrients such as glucose, specific transporter proteins are present in high concentration in brain EC. Classical quail-chick experiments were performed to show that the brain harbors signals to induce an impermeable vascular endothelium (Stewart and Wiley 1981). In these experiments, it was observed that gut tissue transplanted into the brain was vascularized by EC from brain that became leaky in their new location. On the other hand, brain tissue transplanted to gut became vascularized by gut EC that formed a rather impermeable vascular endothelium. This experiment clearly showed that the microenvironment is involved in specifying at least some endothelial features. The cell closest to brain

capillaries is the astrocyte, whose processes cover much of the endothelial basal surface (Fig. 3B). EC culture experiments with astrocyte-conditioned medium or coculture experiments with astrocytes have suggested that astrocytes can reduce the EC permeability, thus pointing to astrocytes as a signaling source for inducing the bloodbrain barrier in vascular endothelium (Rubin et al. 1991). In summary, these experiments show that the organspecific cells adjacent to the microvascular endothelium are (at least partially) responsible for the features of the endothelium. Moreover, it seems as if these organspecific cells change the EC in a way that matches the needs of the specific organ. The liver sinusoidal EC (LSEC) are another example of a tissue- or organ-specific endothelial cell type (Braet and Wisse 2002). The LSEC are characterized by fenestrae, which do not contain any diaphragm in contrast to the fenestrated EC found in other organs (Fig. 3C). The space between the LSEC and the liver parenchymal cells, the space of Disse (SD), is not filled with any basement membrane, but represents a more or less empty space between the endothelium and the parenchyma. Sometimes lipid particles are found in this extracellular space, and this observation points to a possible role of these endothelial structures in lipoprotein metabolism. Dietary fats, absorbed by the epithelium of the small intestine, are assembled with specific apolipoproteins to form specific lipid particles, called chylomicrons. After entering the blood circulation, triglycerides are hydrolyzed to fatty acids in the capillaries of adipose tissue and muscle through the action of lipoprotein lipase present on the capillary EC. The hydrolyzed lipid particles, called chylomicron remnants, enter the space of Disse most likely through the LSEC fenestrae (Fig. 3C) and are taken up by the chylomicron remnant (Apo E) receptors of the liver parenchyma. Thus, LSEC fenestrae are suggested to serve as sites where lipid particles can be exchanged between the blood circulation and liver parenchyma (Braet and Wisse 2002). Studies on LSEC fenestrae have demonstrated that fenestrae are pores or cytoplasmic holes that can contract and dilate, so that the diameter of fenestrae can be changed in response to a variety of different agents (including serotonin and ethanol). A fenestrae-associated cytoskeleton ring was shown to open and close like fenestrae and possibly enables the fenestral changes. Actin and myosin filaments are found adjacent to LSEC fenestrae and represent good candidates for regulating the size of endothelial fenestrae (Braet et al. 1995). It can be anticipated that much of the cell biology of fenestrae will be learned from LSEC. The capillary endothelium in the lung is continuous, but lung EC—unlike brain EC—contain many caveolae and are very active in transcytosis (Fig. 3D). By electron microscopy, caveolae appear as smooth, approximately 50- to 100-nm-diameter flask-shaped invaginations of the plasma membrane (Schnitzer 2001). Unlike fenestrae, caveolae are morphological features found to varying extents in a variety of different cell types including

40 Table 1 Blood vessel-derived signals mediate a variety of diverse processes. Examples of signals and signaling processes are listed (EC endothelial cell, SMC vascular smooth muscle cell, BMP bone morphogenetic protein, TGF transforming growth factor, MMP matrix metalloproteinase, NO nitric oxide, BDNF brain-derived neurotrophic factor, PDGF platelet-derived growth factor, ICAM intercellular adhesion molecule, VCAM vascular cell adhesion molecule, E-selectin endothelial selectin)

Process

Signal

Signaling cell

Recipient cell

Outcome

Differentiation

BMP2 BMP4 BMP7 TGF-beta MMP2 NO BDNF Artemin

EC

Neural crest cells

Autonomic neuronal differentiation

EC EC EC EC SMC

Mesenchymal cells Podocytes SMC Brain neurons Sympathetic neurons

EC EC

Mural cells Leukocytes

Mural cell specification Glomerular assembly Relaxation Survival Chemoattraction of neuroblasts Chemoattraction Adhesion

Morphogenesis Survival Attraction

Inflammation

PDGF-BB ICAM-1 VCAM-1 E-selectin

adipocytes, muscle cells, fibroblasts and type I alveolar epithelial cells. Lung EC have been shown to transport proteins (including albumin) by transcytosis using a caveolae-mediated mechanism (McIntosh et al. 2002; Predescu et al. 1998). Interestingly, both capillary endothelium and type I alveolar epithelium contain caveolae (Schnitzer 2001). This points to a possible transport method for proteins from the blood circulation to the alveolar space and vice versa. Indeed circulatory proteins such as albumin and immunoglobulin G were found within the fluid lining the alveolar epithelial surface (Bignon et al. 1976). The physiologic function of this protein transport has not been identified yet, but it is possible that it serves as waste removal from the blood and immune surveillance in the lung. The last example of an organ-specific endothelium are the EC of the kidney glomerulus (Fig. 3E). The glomerulus is the filtration unit of the kidney and its fenestrated endothelium is found adjacent to the kidney podocytes (Drummond 2000). The podocytes are specialized epithelial cells with an elaborate network of interdigitated cellular processes, called foot processes (Fig. 3E). The interdigitated foot processes (FP) form a fine sieve, which acts as the filtration barrier between the blood plasma and the urinary space. The permeable fenestrated EC are likely to enable waste products to exit the plasma and enter the urinary space.

Endothelial signals in tissue formation The previous paragraphs lead to the question of the molecular mechanisms of EC-tissue interactions. The formation of organs with specific nonvascular cell types as well as specific vascular cells has been suggested to be the outcome of signals exchanged between vascular and nonvascular cells during organ formation (Lammert et al. 2001). Because a review has been recently written about EC signals (Lammert et al. 2003), this section only briefly describes the vascular signals inducing developmental programs in nonvascular tissue cells. Some of these vascular signaling molecules are listed in Table 1. The

signals to which EC respond (such as the VEGFs) are described elsewhere in this issue of Cell and Tissue Research and are not mentioned here, although they play a critical role in determining the tissue-specific properties of EC. As shown in Table 1, a variety of different processes are induced in nonvascular tissue cells by vascular signals. These signals include differentiation signals of the TGF-beta family (Hirschi et al. 1998; Reissmann et al. 1996; Shah et al. 1996). EC-induced morphogenesis involves the upregulation and activation of proteases that degrade the matrix to allow tissue cells to adapt a specific morphology (Serluca et al. 2002). Moreover, endothelial signals such as BDNF support the survival of adjacent cells in organs such as the brain where cell death plays a critical role during development (Leventhal et al. 1999). Another class of vascular signals are those that attract tissue cells to the vessels. These signals include factors such as PDGF (platelet-derived growth factor) and attract mural cells such as pericytes and smooth muscle cells to capillaries and macrovessels, respectively (Hirschi et al. 1998; Leveen et al. 1994; Lindahl et al. 1997). Furthermore, the factor Artemin as a member of the GDNF (glial cell line-derived neurotrophic factor) family acts as a guidance factor that makes sympathetic neuroblasts migrate and project axons to blood vessels (Honma et al. 2002). During an inflammation, EC express or activate various cell adhesion molecules on their apical side (Table 1). These molecules help leukocytes to adhere to EC in an inflamed tissue (Libby and Pober 2001). This example shows that an EC can send signals from both its apical and basal plasma membranes. On the basal side EC signal to the adjacent tissue cells and on the apical side EC signal to hematopoietic cells such as leukocytes.

Conclusions The cardiovascular system is the first functional organ that penetrates the developing vertebrate embryo. It therefore is easy to imagine that many (if not all)

41

developing organs interact with the preexisting vessels and receive vascular signals. To date, endothelial signaling has been shown for some organs such as the pancreas and the liver, and it is likely that—in the future—other organs will be added to the list of organs that depend on vascular signaling. Organ-specific cells also signal to the blood vessels to adapt their properties, so the vessels fit the demands of the organ. Mutual signaling events between the blood vessels and the tissue are likely to ensure that the nonvascular and vascular cells form functional units such as the pancreatic islets or the kidney glomeruli. In future research, the mutual signaling events between vascular and nonvascular cells need to be better studied in order to understand how organs form. In addition, it will be interesting to see how the interaction between organ-specific cell types and EC is disturbed during the disease of an organ. Possible targets for disease treatment might result from such investigations. In summary, the interdependent development of blood vessels and organs is a young research field, and it will be exciting to see how this research contributes to a better understanding of organ development, regeneration and disease.

References Ball GF, Riters LV, Balthazart J (2002) Neuroendocrinology of song behavior and avian brain plasticity: multiple sites of action of sex steroid hormones. Front Neuroendocrinol 23:137–178 Bearer EL, Orci L (1985) Endothelial fenestral diaphragms: a quick-freeze, deep-etch study. J Cell Biol 100:418–428 Biemar F, Argenton F, Schmidtke R, Epperlein S, Peers B, Driever W (2001) Pancreas development in zebrafish: early dispersed appearance of endocrine hormone expressing cells and their convergence to form the definitive islet. Dev Biol 230:189–203 Bignon J, Jaurand MC, Pinchon MC, Sapin C, Warnet JM (1976) Immunoelectron microscopic and immunochemical demonstrations of serum proteins in the alveolar lining material of the rat lung. Am Rev Respir Dis 113:109–120 Bonner-Weir S, Orci L (1982) New perspectives on the microvasculature of the islets of Langerhans in the rat. Diabetes 31:883– 889 Braet F, Wisse E (2002) Structural and functional aspects of liver sinusoidal endothelial cell fenestrae: a review. Comp Hepatol 1:1 Braet F, De Zanger R, Baekeland M, Crabbe E, Van Der Smissen P, Wisse E (1995) Structure and dynamics of the fenestraeassociated cytoskeleton of rat liver sinusoidal endothelial cells. Hepatology 21:180–189 Douarin NM (1975) An experimental analysis of liver development. Med Biol 53:427–455 Drummond IA (2000) The zebrafish pronephros: a genetic system for studies of kidney development. Pediatr Nephrol 14:428–435 Edlund H (2002) Pancreatic organogenesis—developmental mechanisms and implications for therapy. Nat Rev Genet 3:524–532 Fukuda-Taira S (1981) Hepatic induction in the avian embryo: specificity of reactive endoderm and inductive mesoderm. J Embryol Exp Morphol 63:111–125 Grapin-Botton A, Melton DA (2000) Endoderm development: from patterning to organogenesis. Trends Genet 16:124–130 Gualdi R, Bossard P, Zheng M, Hamada Y, Coleman JR, Zaret KS (1996) Hepatic specification of the gut endoderm in vitro: cell signaling and transcriptional control. Genes Dev 10:1670–1682

Hebrok M, Kim SK, Melton DA (1998) Notochord repression of endodermal Sonic hedgehog permits pancreas development. Genes Dev 12:1705–1713 Hirschi KK, Rohovsky SA, D’Amore PA (1998) PDGF, TGF-beta, and heterotypic cell-cell interactions mediate endothelial cellinduced recruitment of 10T1/2 cells and their differentiation to a smooth muscle fate. J Cell Biol 141:805–814 Hogan BL, Yingling JM (1998) Epithelial/mesenchymal interactions and branching morphogenesis of the lung. Curr Opin Genet Dev 8:481–486 Honma Y, Araki T, Gianino S, Bruce A, Heuckeroth R, Johnson E, Milbrandt J (2002) Artemin is a vascular-derived neurotropic factor for developing sympathetic neurons. Neuron 35:267–282 Inoue M, Hager JH, Ferrara N, Gerber HP, Hanahan D (2002) VEGF-A has a critical, nonredundant role in angiogenic switching and pancreatic beta cell carcinogenesis. Cancer Cell 1:193–202 Kim SK, Hebrok M, Melton DA (1997) Notochord to endoderm signaling is required for pancreas development. Development 124:4243–4252 Kreidberg JA, Sariola H, Loring JM, Maeda M, Pelletier J, Housman D, Jaenisch R (1993) WT-1 is required for early kidney development. Cell 74:679–691 Lammert E, Cleaver O, Melton D (2001) Induction of pancreatic differentiation by signals from blood vessels. Science 294:564– 567 Lammert E, Cleaver O, Melton D (2003) Role of endothelial cells in early pancreas and liver development. Mech Dev 120:59–64 Leveen P, Pekny M, Gebre-Medhin S, Swolin B, Larsson E, Betsholtz C (1994) Mice deficient for PDGF B show renal, cardiovascular, and hematological abnormalities. Genes Dev 8:1875–1887 Leventhal C, Rafii S, Rafii D, Shahar A, Goldman SA (1999) Endothelial trophic support of neuronal production and recruitment from the adult mammalian subependyma. Mol Cell Neurosci 13:450–464 Libby P, Pober JS (2001) Chronic rejection. Immunity 14:387–397 Like AA (1970) The uptake of exogenous peroxidase by the beta cells of the islets of Langerhans. Am J Pathol 59:225–246 Lindahl P, Johansson BR, Leveen P, Betsholtz C (1997) Pericyte loss and microaneurysm formation in PDGF-B-deficient mice. Science 277:242–245 Lombardi T, Montesano R, Furie MB, Silverstein SC, Orci L (1986) Endothelial diaphragmed fenestrae: in vitro modulation by phorbol myristate acetate. J Cell Biol 102:1965–1970 Louissaint A Jr, Rao S, Leventhal C, Goldman SA (2002) Coordinated interaction of neurogenesis and angiogenesis in the adult songbird brain. Neuron 34:945–960 Majumdar A, Drummond IA (1999) Podocyte differentiation in the absence of endothelial cells as revealed in the zebrafish avascular mutant, cloche. Dev Genet 24:220–229 Matsumoto K, Yoshitomi H, Rossant J, Zaret KS (2001) Liver organogenesis promoted by endothelial cells prior to vascular function. Science 294:559–563 McIntosh DP, Tan XY, Oh P, Schnitzer JE (2002) Targeting endothelium and its dynamic caveolae for tissue-specific transcytosis in vivo: a pathway to overcome cell barriers to drug and gene delivery. Proc Natl Acad Sci U S A 99:1996– 2001 Nottebohm F (1981) A brain for all seasons: cyclical anatomical changes in song control nuclei of the canary brain. Science 214:1368–1370 Palmer TD, Willhoite AR, Gage FH (2000) Vascular niche for adult hippocampal neurogenesis. J Comp Neurol 425:479–494 Predescu D, Predescu S, McQuistan T, Palade GE (1998) Transcytosis of alpha1-acidic glycoprotein in the continuous microvascular endothelium. Proc Natl Acad Sci U S A 95:6175–6180 Reissmann E, Ernsberger U, Francis-West PH, Rueger D, Brickell PM, Rohrer H (1996) Involvement of bone morphogenetic protein-4 and bone morphogenetic protein-7 in the differenti-

42 ation of the adrenergic phenotype in developing sympathetic neurons. Development 122:2079–2088 Rubin LL, Staddon JM (1999) The cell biology of the blood-brain barrier. Annu Rev Neurosci 22:11–28 Rubin LL, Hall DE, Porter S, Barbu K, Cannon C, Horner HC, Janatpour M, Liaw CW, Manning K, Morales J, et al. (1991) A cell culture model of the blood-brain barrier. J Cell Biol 115:1725–1735 Schnitzer JE (2001) Caveolae: from basic trafficking mechanisms to targeting transcytosis for tissue-specific drug and gene delivery in vivo. Adv Drug Deliv Rev 49:265–280 Serluca FC, Drummond IA, Fishman MC (2002) Endothelial signaling in kidney morphogenesis: a role for hemodynamic forces. Curr Biol 12:492–497 Shah NM, Groves AK, Anderson DJ (1996) Alternative neural crest cell fates are instructively promoted by TGFbeta superfamily members. Cell 85:331–343 Shalaby F, Rossant J, Yamaguchi TP, Gertsenstein M, Wu XF, Breitman ML, Schuh AC (1995) Failure of blood-island

formation and vasculogenesis in Flk-1-deficient mice. Nature 376:62–66 Slack JM (1995) Developmental biology of the pancreas. Development 121:1569–1580 Stewart PA, Wiley MJ (1981) Developing nervous tissue induces formation of blood-brain barrier characteristics in invading endothelial cells: a study using quail-chick transplantation chimeras. Dev Biol 84:183–192 Wells JM, Melton DA (2000) Early mouse endoderm is patterned by soluble factors from adjacent germ layers. Development 127:1563–1572 Wolff JR, Bar T (1972) ’Seamless’ endothelia in brain capillaries during development of the rat’s cerebral cortex. Brain Res 41:17–24 Zaret KS (2002) Regulatory phases of early liver development: paradigms of organogenesis. Nat Rev Genet 3:499–512