During the last 80 years developmental biology of the vascular system has focused on its cardiovascular part. Our knowledge about the development of the lymphatic system is based almost entirely on studies done at the beginning of the 20th century. The landmark studies of lymphatic development have been done in very diverse species such as pig, frog and chicken, which might explain some of the controversial findings. Despite being the standard model organism, the mouse is still one of the less well characterized species concerning embryonic lymphatic development, although the situation is rapidly changing. Among the nowadays frequently used model organisms only the mouse has a large number of lymph nodes as do humans. However, it is still unknown whether the mouse can serve as a truly good model for all aspects of lymphatic research, due to its small size (creating only a minute amount of hydrostatic pressure), differences in the lymphatico-venous connections and many differences at the molecular level.
In the mouse the development of the lymphatic system starts at E10.5 (corresponding to 6.5-7 weeks of human embryonic development and E4.5 in the chick). By that time the cardiovascular system is already fully functional (Clark and Clark 1920; van der Putte 1975a; b). A discrete population of endothelial cells expressing the lymphatic-specific transcription factor Prox-1 can be already identified at E9.5. They are located on one side of the anterior cardinal vein, and at E10.5 the first lymphatic outgrowths (lymphatic primordia) can be identified at that location (Wigle and Oliver 1999; Wigle et al. 2002). It is not understood what induces the outgrowth of these lymphatic primordia. The lymphatic primordia remodel and finally fuse into lymphatic plexuses (lymph sacs). There is considerable inter-species variance in the number and exact location of the lymphatic primordia and lymph sacs, although the jugular region seems always to be the main area of lymphatic induction. In mammalian embryos eight lymph sacs have been described: the paired jugular, subclavian and posterior lymph sacs, the unpaired retroperitoneal sac and the cisterna chyli (Sabin 1909; van der Putte 1975a). Two major contradicting theories have emerged about the events that follow the above-mentioned formation of the lymph sacs.
According to Sabin the peripheral lymphatic system develops from the embryonic lymph sacs exclusively by the sprouting of endothelial cells into the surrounding tissues and organs (Sabin 1902; Clark 1912). Most recent data favors this theory, including expression studies of lymphatic-specific markers (Kaipainen et al. 1995; Kukk et al. 1996) and the Prox-1 knock-out mouse (Wigle et al. 2002). However, there is no agreement among the advocates of the centrifugal sprouting theory about the relationship of the embryonic lymph sacs to the adult lymphatico-venous communications. According to Huntigton and McClure (1910) all lymph sacs lose their connections with the veins and the adult jugular lymphatico-venous communication is a secondary development. Alternatively, the jugular communication of the adult animal is a persisting embryonic communication and data from the mouse argues for this theory (van der Putte 1975b).
McClure and Huntington proposed a vasculogenic mechanism for the establishment of the peripheral lymphatic system. In the mesenchyme lymphatic spaces would arise independently from the veins, fusing into a primitive lymphatic network, which subsequently would spread centripetally and connect to the venous system. The centripetally sprouting lymphatics would either integrate or replace the embryonic lymph sacs (Huntington and McClure 1910; Kampmeier 1912). It is true that the luminal continuum of the lymphatic primordia to the early veins is often not seen (van der Putte 1975a), but a venous origin does not in itself require such, as individual cells might migrate to form the lymphatic primordia.
A model that incorporates both sprouting from lymph sacs and in-situ differentiation of mesenchymal precursors was already proposed in 1932 by van der Jagt and support for this model has been recently gathered. The lymphatics of the avian chorioallantoic membrane (CAM) and perhaps also the wing are apparently not only sprouts from the lymph sacs but also in-situ derivations from mesenchyme. Homotopic grafting of Prox-1 negative, day 2 quail allantoic buds into chick hosts and day 3.5 chick wing buds into quail hosts resulted in lymphatics composed of both donor and host endothelial cells in the graft area (Papoutsi et al. 2001; Schneider et al. 1999).
The development of the lymphatics, just like the one of the blood vessels, is restrained by their evolutionary origin. Much the same as the transient aortic arches in mammalian development are the lymph hearts in birds, which are formed and functional during embryogenesis but disappear by adulthood. In the adult organism lymphatic endothelial cells are normally quiescent, but angiogenic processes - both pathological and physiological - are often accompanied by lymphangiogenesis (Clark and Clark 1932; Ohtani et al. 1998; Paavonen et al. 2000; Mimura et al. 2001).