Lymphangioblasts in the Allantoic Mesoderm
The CAM develops at the inner surface of the egg shell by fusion of the densely vascularized allantois with the avascular chorion. The differentiated CAM is lined by an outer, ectodermal, chorionic epithelium, and an inner, endodermal, allantoic epithelium. The capillary plexus is located within the chorionic epithelium, and the blood-air-barrier of this respiratory organ is, therefore, extremely thin. The larger blood vessels are located in the mesodermal stroma (Leeson and Leeson, 1963). The allantoic (umbilical) artery and vein are originally paired, but the right member in each case regresses. In accordance with previous studies (Wilting et al., 1996; Oh et al., 1997; Papoutsi et al., 2000), we have shown that the arteries and veins of the differentiated CAM are accompanied by lymphatics which are characterized by the expression of two markers: VEGFR-3 and Prox1.
It is well known that the blood vessels of the CAM originate from the allantoic mesoderm (Hamilton, 1965), but the embryonic development of CAM lymphatics has not been studied yet. According to the theories proposed by Sabin (1909), the lymphatics are exclusively derived by sprouting of the early embryonic lymph sacs. Therefore, the lymphatics of the CAM would be derived from the posterior (pelvic) lymph sacs. These lymph sacs are not detectable before day 5 of incubation (Miller, 1913). We have investigated the origin of CAM lymphatics by means of descriptive studies and with the quail/chick chimera model (Le Douarin, 1969). We have grafted the allantoic bud of 3-day-old quail embryos (stage 17–18 HH) homotopically into chick embryos. The hosts were reincubated until day 13 when the CAM is fully differentiated and VEGFR-3 serves as a specific marker of lymphatic endothelial cells in the CAM. Our results unequivocally show that on day 3 the allantoic mesoderm has lymphangiogenic potential. The CAM of the host (chick) contained areas where the endothelium of the blood vessels and the lymphatics were of donor (quail) origin. The lymphatics in these areas are QH1 and VEGFR-3 double-positive, and they are located in their typical position around arteries and veins. The results demonstrate lymphangiogenic potential of the allantoic mesoderm long before the development of the posterior lymph sacs. On day 3, the allantoic mesoderm contains ramified cells that aggregate and form blood islands. In contrast to the blood islands in the yolk sac (Wilting et al., 1995a), the allantoic blood islands are immediately covered by additional mesodermal cells. Neither VEGFR-3 nor Prox1 are markers of the allantoic lymphangioblasts on day 3 (stage 17–18 HH). VEGFR-3 is expressed as soon as blood vascular endothelial networks have formed. During tissue maturation, the receptor expression then becomes restricted to the lymphatic endothelium (Kaipainen et al., 1995; Wilting et al., 1997). However, VEGFR-3 is not expressed in isolated mesenchymal cells, and is, therefore, neither a marker of angioblasts nor of lymphangioblasts. VEGF-C, the ligand of VEGFR-3, is expressed ubiquitously in the early allantoic bud and may serve as an angiogenic growth factor during early development. This finding is in line with the observation that VEGFR-3–deficient mice die of cardiovascular failure during early embryonic development (Dumont et al., 1998). In the differentiated CAM, high levels of VEGF-C expression are restricted to two sites: the allantoic epithelium and the wall of larger blood vessels. The lymphatics of the CAM are located immediately adjacent to the larger blood vessels, and it can be assumed that the expression of VEGF-C in the blood vascular wall serves for the patterning of lymphatics. In the adult, constitutive expression of VEGF-C may then serve as a maintenance factor for the lymphatics (Eriksson and Alitalo, 1999). This function seems to be of major importance for the lymphatic capillaries, because these are not stabilized by a continuous basal lamina and pericytes. We have previously shown that the application of VEGF-C on the differentiated CAM induces development of lymphatics, which are derived by proliferation and growth of the preexisting lymphatics (Oh et al., 1997). VEGF-C, is synthesized as a prepropeptide of 61 kDa and undergoes proteolytic maturation (Korpelainen and Alitalo, 1998). The immature and mature forms of VEGF-C bind VEGFR-3 (flt4, Quek2) with high affinity (Joukov et al., 1996; Lee et al., 1996; Eichmann et al., 1998), whereas only the mature form binds VEGFR-2 (KDR, flk1, Quek1) (Joukov et al., 1997). The two receptors are expressed in the lymphatic endothelium of the CAM (Oh et al., 1997).
The earliest known marker of the lymphatic endothelium is the homeobox-containing transcription factor Prox1. In mice, Prox1 mRNA is expressed in a subpopulation of endothelial cells of the early jugular vein and in the endothelium of the lymph sacs. Expression is then found in differentiated lymphatics, but not in blood vessels (Oliver et al., 1993; Wigle and Oliver, 1999). We have observed the same expression pattern in chick embryos (unpublished data). In Prox1-deficient mice, the development of the lymphatics is arrested shortly after the lymph sacs have formed, and the embryos show severe edema (Wigle and Oliver, 1999). These mice die during early stages of development, probably due to malformations of the liver, which also expresses Prox1 (Wigle et al., 1999). The expression pattern of Prox1 in the chick is identical to that of the mouse (Tomarev et al., 1996). Additionally, in the differentiated CAM Prox1 is a specific marker of lymphatic endothelial cells. No Prox1-positive cells can be observed in the allantoic bud of stage 18 (HH) chick embryos (day 3) neither on protein nor on mRNA levels, although the experiments demonstrate the existence of lymphatic precursors in this tissue. However, 1 day later, Prox1 is expressed in scattered cells of the allantoic mesoderm of stage 21 (HH) embryos and there is no indication of a localized budding or sprouting of the lymphatics from the allantoic veins. Unlike the jugular vein, there is no expression of Prox1 in restricted areas of the allantoic veins, which might have indicated a venous origin of the allantoic lymphatics. Furthermore, early lymph sacs are retrogradely filled with blood and have been identified by their dark, stagnant blood (Clark and Clark, 1920). Such features have never been observed in the allantois. In contrast, the scattered Prox1-positive cells of the allantois soon form lymphatic networks, which are preferentially located adjacent to the large blood vessels of the CAM of day 5 embryos. It seems to be of great functional importance that the two vascular systems do not fuse with each other. Only in the coccygeal region do lymphovenous anastomoses exist and are characterized by the development of specialized contractile structures, the lymph hearts (Berens von Rautenfeld and Budras, 1981; Wilting et al., 1999). Later, these anastomoses become severed and the thoracic duct takes up function. Our results strongly suggest that lymphangioblasts are at first Prox1-negative, but expression of the gene then starts during commitment of mesenchymal cells to the lymphatic endothelial lineage. Our results demonstrate that lymphangioblasts are present in the allantois, and, therefore, the mechanisms driving angiogenesis and lympangiogenesis are very much alike. Lymphangiogenesis seems to be a complex process involving both the development of lymphatics from specialized parts of specific veins, and the fusion and tube formation of lymphangioblasts located in the mesoderm.
Circulating Angioblasts From the Allantois?
In our experiments, we have not been able to detect either blood vascular or lymphatic endothelial cells in places other than the CAM. The emergence of circulating angioblasts from the allantoic bud has been reported recently (Caprioli et al., 1998). We have observed QH1-positive cells, most likely representing macrophages, in numerous intraembryonic sites such as the walls of arteries, connective tissue, and bone marrow. That we could not detect circulating angioblasts could mean that this population of cells is either very small, and we have, therefore, missed them, or these cells are not all present in early avian development. The latter seems to be supported by the fact that circulating angioblast were not detected in quail/chick parabiosis studies in which the chick becomes perfused by quail blood cells (Kurz and Christ, 1998, Kurz et al., 2001). The major difference of our studies and those by Caprioli et al. (1998) resides in the fact that these authors have grafted the allantoic bud into the coelomic cavity, whereas we have performed homotopic grafting. Several studies of various groups have shown that angioblasts grafted into the coelomic cavity migrate through the mesenchyme and integrate into the lining of blood vessels (Christ et al., 1990; Pardanaud and Dieterlen-Lièvre, 1999). By this means, the cells may also have become integrated into the bone marrow vasculature, where they have been observed by Caprioli et al. (1998). However, the emergence of circulating angioblasts in avian development calls for further investigations.
In summary, our study shows that lymphangioblasts are present in the splanchnic mesoderm of the allantoic bud. This finding is in line with the observation of lymphangioblasts in the paraxial mesoderm (Wilting et al., 2000). Angioblasts and lymphangioblasts are obviously derived from identical mesodermal subcompartments. Lymphangioblasts are at first Prox1-negative, but start expressing this transcription factor while still located in the mesenchyme.