Physiologically, lymphatic vasculatures in small intestine and its mesentery serve as essential conduits for the absorption and transport of lipids from the intestine to the thoracic duct and into the blood circulation at the left subclavian vein (Saladin,2004; Zawieja,2005). In the villi of the small intestine, digested and hydrolyzed carbohydrates and amino acids are absorbed into blood capillaries, while digested and hydrolyzed lipids are absorbed as chyle into lymphatic capillaries, so-called “lacteals” (Saladin,2004). The fine structure, distribution, and function of lymphatic vessels in the small intestine have been examined with several classical methods in adults (Ohtani,1987; Unthank and Bohlen,1988) and during the postnatal period (Shimoda et al.,2001). The growth of lymphatic vessels has also been studied during regeneration of the muscle coat after transection (Shimoda et al.,2004). However, little is known about the developmental dynamics of intestinal lymphatic vasculature during embryonic development.
To date, the molecular and cellular regulation, control of development, differentiation, and maintenance of lymphatic vessels are only starting to be elucidated (Hong et al.,2004; Alitalo et al.,2005; Oliver and Alitalo,2005). Recent studies of gene-targeted mice have greatly advanced our understanding of the molecular mechanisms regulating lymphatic vessel formation (Hong et al.,2004; Alitalo et al.,2005; Oliver and Alitalo,2005). By around embryonic day (E) 10.5 in the embryonic mouse, the lymphatic vessels develop centrifugally by budding from central veins in response to a lymphangiogenic growth factor, vascular endothelial growth factor-C (VEGF-C), and its lymphatic endothelium-specific receptor, VEGF receptor-3 (VEGFR3) (Hong et al.,2004; Karkkainen et al.,2004; Alitalo et al.,2005; Oliver and Alitalo,2005). These lymphatics subsequently give rise to lymphatic vessel plexuses and networks by regulation of a key transcription factor, Prox-1 (homeobox domain related transcription factor) (Wigle and Oliver,1999; Oliver,2004; Hong et al.,2004; Alitalo et al.,2005; Oliver and Alitalo,2005). Later in development, around E11.5 to E14.5, these primary lymphatic vessels separate from the blood vessels by control of the tyrosine kinases Syk and adaptor protein SLP76 (Abtahian et al.,2003). Then, further prenatal and postnatal development allows patterning and remodeling of the lymphatic vessels to form a superficial capillary plexus and collecting lymphatic vessels in response to angiopoietin-2, ephrinB2, Foxc2, podoplanin, and unidentified factors (Gale et al.,2002; Schacht et al.,2003; Petrova et al.,2004; Makinen et al.,2005; Shimoda et al.,2007). Thus, molecular control of the formation and development of lymphatic vessels is more precisely regulated in space and time than we previously envisaged. However, many cellular mechanisms still remain unsolved concerning the generation and development of lymphatic vessels in each organ during embryonic development.
In this study, we examined the development of intestinal lymphatic and blood vasculatures during mouse embryonic development by immunostaining of recently discovered molecular markers for lymphatic endothelial cells, such as lymph vessel endothelial hyaluronan receptor-1 (LYVE-1), VEGFR3, Prox-1, and podoplanin (Banerji et al.,1999; Wigle and Oliver,1999; Hong et al.,2004; Oliver,2004; Alitalo et al.,2005; Oliver and Alitalo,2005). Our data reveal that the temporal and spatial dynamics of development and maturation of lymphatic vasculatures are different from those of blood vasculature. Our findings provide useful information for understanding the underlying cellular mechanisms of developmental abnormalities of the lymphatic vasculature in the intestine.
RESULTS AND DISCUSSION
Presence of Characteristic LYVE-1+ Lymphatics and Macrophages in the Intestine of E20.5 and E16.5 Mouse
To examine whether lymphatic vessels are already formed in the intestinal villi before birth, the ileac region of intestines and mesenteries of E20.5 and E16.5 mice were whole-mounted and immunostained with anti-LYVE-1 antibody (Fig. 1). At E20.5 mouse, LYVE-1+ lymphatic capillaries (lacteals) were clearly identified in the intestinal villi. Abundant and relatively well-organized LYVE-1+ lymphatic plexuses and networks were observed in the intestinal wall (Fig. 1A and B). Moreover, these lymphatic plexuses in the intestinal wall were connected to lymphatics in the mesentery. Interestingly, LYVE-1+ macrophages were evenly distributed in the mesenteric membrane, while some LYVE-1+ macrophages were densely clustered (Fig. 1B). At the intestinal-mesenteric border of E16.5 mice, LYVE-1+ lymphatic vessels of mesentery had many branches extending to the outside and inside of the intestinal walls (Fig. 1C), suggesting that intestinal lymphatics in the wall and villi are derived from mesenteric lymphatics through an active branching process. In addition, variably-shaped LYVE-1+ and F4/80+ macrophages were observed close to or apart from the branching lymphatics (Fig. 1C and D), suggesting that macrophages may play a role in the formation of lymphatic branching through secretion of lymphangiogenic factors. In fact, recent reports (Schledzewski et al.,2006; Jackson,2004; Cho et al.,2007) have indicated that a subset of LYVE-1+ macrophages are present in wound healing tissues, implanted malignant tumor tissues, and the tip region of epididymal adipose tissue, and these macrophages are involved in angiogenesis and lymphangiogenesis in these pathologic and physiologic environments. Thus, a set of LYVE-1+ macrophages in the developing intestine could be involved in proper angiogenesis and lymphangiogenesis in intestine and mesentery.
Active Formation of Lymphatic Branching in the Intestinal Wall at E16.5
The findings above led us to examine in more detail the developmental changes of lymphatic and blood vessels in intestine during embryonic development. Immunostaining with anti-LYVE-1, anti-VEGFR3, and anti-PECAM-1 (selective marker for blood endothelial cells) in the whole-mounted intestines revealed that there were few or no lymphatic vessels in the intestinal tubes of E13.5 and E14.5 mice, although immature lymphatics and a few scattered LYVE-1+ macrophages were observed in the mesentery (Fig. 2). In comparison, blood vessels had already formed with well-organized plexuses and plentiful networks in the intestinal wall of E13.5 and E14.5 mice (Fig. 2). These observations clearly indicate that vasculogenesis or angiogenesis precedes lymphangiogenesis during intestinal development. In the intestinal tube of E16.5 mouse, there was a dramatic increase of lymphatic branching in the outside and inside of intestinal walls with plentiful and variably shaped LYVE-1+ macrophages (Fig. 2). In the intestinal wall of E19.5 mouse, a relatively well-organized lymphatic plexus and networks were observed (Fig. 2). Thus, the lymphatic plexus and networks in the intestinal wall are formed through an active branching process in the last trimester of development.
We next examined developmental changes in mesenteric lymphatic and blood vessel by immunostaining with anti-LYVE-1, anti-VEGFR3, and anti-PECAM-1 antibodies in the whole-mounted mesenteries that were connected with the intestine. In the mesentery of E13.5 mouse, we identified immature lymphatic vessels and plentiful and variably shaped LYVE-1+ macrophages; nearby blood vessels were mature and had abundant branching into the intestinal wall (Fig. 3A). In the mesenteries of E15.5 and E17.5 mice, VEGFR3 immunostaining clearly identified the mature lymphatic vessels that had valves (Fig. 3A). In the mesenteric–intestinal border of 17.5 mice, plentiful radial and perpendicular branching of VEGFR3+ lymphatics were observed, suggesting that fat absorptive lacteals and collecting lymphatics in the intestinal tube could be formed by active lymphangiogenesis through dynamic activation of the VEGF-C/VEGF-D/VEGFR3 system during this period (Fig. 3B). Recently, one report (Maruyama et al.,2005) raises a possibility that macrophages could participate to form lymphatic tubes; these macrophages express Prox-1 during inflammation. Therefore, we examined the possibility of whether macrophages may participate to form lymphatics during intestinal development. Because lymphatic endothelial cells, but not macrophages, express both LYVE-1 and Prox-1, we could distinguish lymphatic endothelial cells from macrophages by double immunofluorescent staining. Our extensive analysis revealed that there were no Prox-1+ macrophages in the mesentery (Fig. 3C) or intestinal tube (data not shown) of E17.5 mouse. These data indicate that macrophages in the gut could not be transdifferentiated into lymphatic endothelial cells during mouse embryonic development. Moreover, LYVE-1 is structurally related to CD44 and other hyaluronan-binding proteins, and is known to be selectively expressed in lymphatic endothelial cells, sinusoidal endothelium in liver and spleen (Banerji et al.,1999; Jackson,2004), and a set of macrophages (Schledzewski et al.,2006; Cho et al.,2007). The mechanisms of LYVE-1 expression and the functional significance of LYVE-1 beyond hyaluronan transport have not been defined in detail. However, recent reports (Gale et al.,2007; Huang et al.,2006) showed that LYVE-1 knockout mice displayed an apparently normal phenotype, with no significant alterations in hyaluronan metabolism, development and function of lymphatic vessels, or inflammatory processes. Therefore, the role of LYVE-1 in lymphatic endothelial cells and macrophages needs to be determined within a range beyond hyaluronan transport in the future. Podoplanin is a mucin-type transmembrane glycoprotein that is expressed not only in lymphatic endothelial cells but also in renal podocytes, keratinocytes, and alveolar type I cells in lung (reviewed in Hong et al.,2004; Alitalo et al.,2005; Oliver and Alitalo,2005). Podoplanin is first expressed around E11.0 in Prox-1-positive lymphatic progenitor cells, and podoplanin (−/−) mice die at birth due to respiratory failure and have defects in lymphatic structure and function (Schacht et al.,2003). Therefore, we examined the expression pattern of podoplanin in the intestinal–mesenteric border of E17.5 mice during formation of active lymphatic branching. Both podoplanin and LYVE-1 were strongly expressed in the larger lymphatic vessels of mesentery, but LYVE-1 was more highly expressed in most of the branching lymphatics of the intestinal wall. In comparison, both VEGFR3 and LYVE-1 were strongly expressed in branching lymphatic vessels, and even in the macrophages. Thus, as suggested by previous reports (Schacht et al., 2004; Makinen et al.,2005), podoplanin plays an important role in the differentiation of lymphatic vessels, whereas LYVE-1 and VEGFR3 play an important role in the branching of lymphatic vessels during intestinal lymphatic development.
Lymphatic Capillaries Are Formed in Intestinal Villi at E15.5–17.5
To examine the timing of the formation of the lymphatic capillaries (lacteals) in intestinal villi, cross-sections of intestine were immunostained with anti-LYVE-1, anti-PECAM-1, and anti-Prox-1 (a selective transcription factor for lymphatic vessels). In the intestine of E13.5 mouse, there were no villi in the intestinal tube, but a few LYVE-1+/Prox-1- and LYVE-1-/Prox-1+ cells were in the surface, but not in the inside (Fig. 4A). In contrast, the plexus of blood vessel endothelial cells in the intestinal tube extended to the surface of the plexus of blood vessel endothelial cells at the surface of the intestine via endothelial bridges (Fig. 4A); this structure has been well described in a recent report (Wilm et al.,2005). In the intestine of E14.5 mouse, there were a few villi, rare LYVE-1+ cells, and almost no Prox-1+ cells, but plexuses of blood vessel endothelial cells in the middle portion as well as in the intestinal villi were observed (Fig. 4A). In the intestine of E15.5 mouse, we can observe many villi; the PECAM-1+/LYVE-1+ blood vessel endothelial cells are organized in vessel shapes and located in the middle portion as well as in villi; a few LYVE-1+ and Prox-1+ lymphatic endothelial cells are present in the middle portion (Fig. 4A). In the intestine of E17.5 mouse, mature shapes of villi, plexuses of blood vessel endothelial cells (PECAM-1+ cells), and lymphatic endothelial cells (LYVE-1+ and Prox-1+ cells or only LYVE-1+ cells) were observed in the middle portion and villi (Fig. 4B). These data indicate that: (1) the formation of the blood vessel plexuses and capillaries precedes the formation of lymphatic vessel plexuses and capillaries, (2) an organized lymphatic plexus and capillaries are detected around E17.5, and (3) blood endothelial cells transiently expressed both PECAM-1 and LYVE-1 around E15.5. Moreover, it is reported that the intestinal mesodermal origin of blood vascular plexus in the gut at E9.5–10.5 extends to the blood vascular plexus that originated from the serosal mesothelium at E13.5–E16.5 (Wilm et al.,2005). In comparison, our study reveals that there is no intestinal mesodermal origin of lymphatic vascular plexus in the intestine. Lymphatic vessels in the intestine originate from extension of mesenteric lymphatic vessels through an active branching process.
Together, these data reveal that the lymphatics are detected in mesentery at E13.5, and organized lymphatic vessel plexuses and capillaries are formed in the intestinal tube and villi around E17.5, corresponding with a high expression of VEGFR3 in lymphatic endothelial cells and possible involvement of LYVE-1+ macrophages during mouse intestinal development (Fig. 4C). Although formation of lymphatic vasculature follows formation of blood vasculature in the intestinal tract, the mature form of lymphatic vasculature is already formed before birth, ready to absorb lipids from milk (Fig. 4C). Our data may be useful to clarify defective cellular mechanisms of developmental intestinal lymphatic vasculatures observed in primary intestinal lymphangiectasia and in knockout mice that have intestinal lymphatic defects during embryonic development.
Animals and Treatment
Both male and female specific pathogen-free C57BL/6J mice were purchased from Jackson Laboratory (Jackson Labs, Bar Harbor, ME), and the mice were bred in our pathogen-free animal facility. Embryos were used from timed mating, with the day of plug formation being E0.5. Animal care and experimental procedures were performed under approval from the Animal Care Committees of KAIST.
Histological and Morphometric Analysis
Pregnant mice were anesthetized by intramuscular injection of a combination of anesthetics (80 mg/kg ketamine and 12 mg/kg xylazine). Embryos of the indicated ages were harvested from the amniotic cavities, and intestines from the embryos were carefully harvested and fixed by 1% paraformaldehyde in PBS (for wholemount) or 4% paraformaldehyde in PBS (for cryosection). After blocking with 5% donkey serum in PBST (0.3% Triton X-100 in PBS) for 1 hr at room temperature, the whole-mounted or cross sectioned tissues were incubated with one or more of the following primary antibodies: anti-mouse PECAM-1 antibody, hamster clone 2H8, 1:1,000 (Chemicon International, Temecula, CA); anti-mouse LYVE-1 antibody, rabbit polyclonal, 1:1,000 (Upstate, Lake Placid, NY); anti-mouse LYVE-1 antibody, rat monoclonal, 1:1,000 (Aprogen, Daejeon, Korea); anti-mouse VEGFR3 antibody, goat polyclonal, 1:1,000 (R&D, Minneapolis, MN); anti-human Prox-1 antibody, rabbit polyclonal, 1,1000 (Reliatech, Braunschweig, Germany), and anti-mouse podoplanin antibody, goat polyclonal, 1:1,000 (Santa Cruz Biotechnology). After several washes in PBST, the samples were incubated for 2 hr at room temperature with the following secondary antibodies. For 3,3′-diaminobenzidine (DAB) immunostaining, samples were incubated with the following antibodies: HRP-conjugated anti-rat antibody (Jackson ImmunoResearch, West Grove, PA), HRP-conjugated anti-rabbit antibody (Amersham, Piscataway, NJ), HRP-conjugated anti-hamster antibody (Jackson ImmunoResearch) or HRP-conjugated anti-goat antibody (Jackson ImmunoResearch), and developed with DAB substrate kit according to manufacturer's instructions (Vector, Burlingame, CA). For immunofluorescent staining, the samples were incubated with the following antibodies. FITC- or Cy5-conjugated anti-rat antibody, 1;1,000 (Jackson ImmunoResearch); Cy3-conjugated anti-hamster IgG antibody, 1:1,000 (Jackson ImmunoResearch); FITC- or Cy5-conjugated anti-rabbit antibody, 1;1,000 (Jackson ImmunoResearch); Cy3-conjugated anti-goat antibody, 1:1,000 (Jackson ImmunoResearch). For control experiments, the primary antibody was omitted or substituted with preimmune serum. DAB and fluorescent signals were visualized, and digital images were obtained using a Zeiss Apotome microscope and a Zeiss LSM 510 confocal microscope equipped with argon and helium-neon lasers (Carl Zeiss).
We thank Jennifer Macke for editing of the manuscript. This work was supported by the Korea Science and Engineering Foundation (KOSEF) through the National Research Lab. Program funded by the Ministry of Science and Technology (2004-02376, G.Y.K.).