TIE1-dependent lymphatic vascular remodeling is mediated by its second tyrosine kinase domain

Abstract

Mutations in ANG2 and TIE1 are associated with primary lymphedema in humans, but the mechanisms of ANG/TIE signaling in the lymphatic vasculature remain incompletely understood. We document that TIE2 is not detected in lymphatic endothelial cells (LECs) before E14.5 but is expressed in collecting vessels from later embryonic stages, in contrast to robust TIE1 expression in all LECs from early stages. Accordingly, only LEC-specific deletion of Tie1 but not Tie2 resulted in defective lymphatic development and abnormal valve function. We discovered that defects of Tie1 lymphatic knockout mice were largely rescued by simultaneous loss of Foxo1. In addition, FOXC2 expression was abolished in Tie1-deficient lymphatics but restored by simultaneous loss of Foxo1, indicating that FOXO1, regulating FOXC2, might be a direct downstream effector of TIE1 signaling in the lymphatic system. Further, we generated point mutations in each tyrosine kinase domain of TIE1 and found that the second, but not the first, tyrosine kinase domain of TIE1 is essential for its function in the lymphatic system. Thus, our results suggest new avenues for manipulation of TIE1 signaling to enhance therapeutic lymphangiogenesis.

Keywords: FOXC2; FOXO1; Lymphatic system; Mouse; TIE1; TIE2; Tyrosine kinase domain.

Fig. 1.

TIE2 and TIE1 expression in LECs of developing and mature lymphatics. (A-L) Transverse frozen sections of E16.5 WT mouse embryos were immunostained for LYVE1 or PROX1 (LEC markers) and TIE2 or TIE1. Both TIE1 and TIE2 were co-expressed with LYVE1 and PROX1 in dermal and mesenteric LECs at E16.5. However, single red channel images (D-F,J-L) show that relative to the strong staining of TIE1 in LECs and TIE2 in blood vessel ECs, TIE2 staining in LECs is much weaker, especially in dermal lymphatic capillaries (D,E). Arrow, lymphatics; arrowheads, blood vessels. (M-T) Whole-mount (M-P) and frozen section (Q-T) immunostaining of WT mesenteries at P7 revealed that, similar to TIE1, TIE2 is co-expressed with PROX1 on LECs of mesenteric collecting lymphatic vessels and valves (arrows). Single red channel images (O,P,S,T) show TIE1 or TIE2 staining only. Arrowheads, blood vessels. DAPI was used to counterstain nuclei in A-C,G-I,Q and R. Scale bars: 25 μm (A-L,Q-T); 50 μm (M-P).

Fig. 2.

LEC-specific deletion of Tie1 but not Tie2 results in defective lymphatic development. (A) Tamoxifen injection protocol for embryonic deletion of Tie1 and Tie2. (B-D) Compared to control (B, Tie1^f/f or Tie2^f/f), early embryonic LEC-specific deletion of Tie1 via Prox1CreER^T2 results in dorsal subcutaneous edema (C, arrow) at E14.5, while Prox1CreER^T2;Tie2^f/f embryos (D) treated with the same procedure appeared as normal as control. (E-G) Whole-mount immunostaining of mesenteric vessels from Prox1CreER^T2;Tie1^f/f (F), Prox1CreER^T2;Tie2^f/f (G) and their control embryos (E) at E18.5 with PROX1(red) and CD31(green). b, blood vessel; l, lymphatic vessel. (H) Quantification of the total valves per millimeter from each mesentery of control, Prox1CreER^T2;Tie1^f/f (Tie1^lecKO) and Prox1CreER^T2;Tie2^f/f (Tie2^lecKO) embryos. (I) Schematic of WT mesenteric lymphatic vascular hierarchy at E18.5, demonstrating vessels denoted as primary, secondary and tertiary based on branching and distance from the intestine. (J) Quantitation of lymphatic vascular hierarchy in E18.5 control, Tie1^lecKO and Tie2^lecKO embryos. Lymphatic vessel width was measured by the widest portion (yellow bar) at the three levels of the lymphatic vascular tree as indicated in I. Data are mean±s.e.m. N=5 per genotype. Two-tailed unpaired Student's t-test; **P<0.01; NS, not significant. Scale bars: 100 μm.

Fig. 3.

Postnatal LEC-specific deletion of Tie1 results in defects in lymphatic collecting vessels. (A) Tamoxifen injection procedure for postnatal deletion of Tie1. (B-E) Direct fluorescence imaging of GFP (green) of freshly dissected mesenteries (proximal and middle jejunum) from the Prox1Cre-GFP-expressing control (Tie1^f/f) and Prox1CreER^T2;Tie1^f/f pups at P8. Yellow arrows indicate lymphatic valves. D and E are high-magnification images of lymphatic valves with leaflets (white arrows) or without leaflet (white arrowhead). (F,G) Quantification of the diameter of lymphatic collecting vessels (F) and total valves, including mature and immature valves per millimeter from each mesentery (G). Data are mean±s.e.m. N=5 per genotype. Two-tailed unpaired Student's t-test; *P<0.05, **P<0.01; NS, not significant. Scale bars: 500 μm (B,C); 50 μm (D,E).

Fig. 4.

LEC-specific deletion of Tie1 at adult stage results in lymphatic valve function defects. (A) The morphology of representative valves isolated from the 8-12 week old control (Tie1^f/f) and Tie1 mutant (Prox1CreER^T2;Tie1^f/f) mice treated with tamoxifen in the valve function test. (B-D) Summary data of the average back leak (cmH₂O) for each genotype during the low pressure back-leak tests (0.5-10 cmH₂O), the high pressure back-leak tests (10-100 cmH₂O) and closure tests. Data are mean±s.e.m. of n=9 (control) or 27 (mutant) valves from at least five animals of each genotype. Two-tailed unpaired Student's t-test. (E) Proportion of abnormal valves in each experimental group presented in B-D, showing impaired function of mutant valves. Scale bars: 40 μm.

Fig. 5.

FOXO1 expression in lymphatic system. (A-F) Transverse cryosections of WT mouse embryos at E11.5 (A), E12.5 jugular region (B), E16.5 intestine (C), E16.5 thoracic duct (D) and P7 mesentery (E,F) were co-immunostained for FOXO1 (red) and PROX1 or VEGFR3 (both labeling LECs, green), showing that FOXO1 was strongly expressed in the LEC nuclei during developmental lymphangiogenesis, including LEC progenitors (A) and LECs in embryonic lymphatic sacs (B), lymphatic valves (yellow arrows) and collecting vessels (C-F), in contrast to cytoplasmic staining of FOXO1 in vascular ECs. The majority of PROX1- or VEGFR3-positive LECs are positive for FOXO1 (white arrows), but a portion of LECs were negative for FOXO1 (arrowheads). (G) Quantification of FOXO1-negative LECs in embryonic jugular lymphatic sacs (JLS) at E11.5, E12.5 and the thoracic duct (TD) at E16.5. Five mice and three consecutive section pictures from each were measured. Data are mean±s.e.m. N=5 per genotype. One-way ANOVA; **P<0.01. (H,I) Transverse cryosections of WT and hypomorphic Tie1^neo/neo mouse embryos at E13.5 (through the jugular region) were co-immunostained for FOXO1 (red) and VEGFR3 (labeling LECs, green), showing that Tie1 deficiency did not alter subcellular localization of FOXO1 in LECs. b, blood vessel; cv, cardinal vein; jls, jugular lymphatic sacs; l, lymphatic vessel. DAPI was used to counterstain nuclei. Scale bars: 50 μm.

Fig. 6.

Lymphatic defects of Ncre;Tie1^f/f were largely rescued by loss of Foxo1. (A) Survival curve analysis of WT (n=32), Ncre;Tie1^f/f (n=36) and Ncre;Tie1^f/f;Foxo1^f/f (n=45) mice. (B-F) Direct fluorescence imaging of GFP (green) of freshly dissected mesenteries (proximal and middle jejunum) from Prox1-GFP reporter-expressing control (Tie1^f/f or Tie1^f/f;Foxo1^f/f), Ncre;Tie1^f/f and Ncre;Tie1^f/f;Foxo1^f/f pups at P0 (B-D) and P8 (E,F). Arrows, mature valves; arrowheads, immature valves. (G-I) Quantification of valves (including mature, immature and total valves) per millimeter from each mesentery (G), the average number of the branches located in the thinner precollecting lymphatic vessels close to the intestinal wall from ten large collecting vessels in each mesentery (H), and the diameter of lymphatic collecting vessels (I). Data are mean±s.e.m. N=6 per genotype. One-way ANOVA; *P<0.05, **P<0.01; NS, not significant. Scale bars: 200 μm.

Fig. 7.

Dermal lymphatic phenotype in mice with Tie1 TK1/TK2 point mutations. (A-C) Schematic overview of intracellular TK domains of the murine TIE receptors and comparison of their partial sequence. (A) TIE2 has three phosphotyrosine residues (1101, 1107, 1112) in TK2, which are known to associate with important signaling proteins GRB2, p85, Dok-R and SHP2, whereas TIE1 has only two (1113 and 1124). The equivalent to TIE2 pTyr1107 is missing. CM, cell membrane; TK, tyrosine kinase. (B) Alignment of partial sequence in TIE2/TIE1 TK1 domain. (C) Alignment of partial sequence in TIE2/TIE1 TK2 domain. (D) Macroscopic images of Tie1 WT, TK1^−/− and TK2^−/− embryos at E14.5 and E18.5. Arrow, edema. (E) Direct fluorescence imaging of GFP (green) of freshly dissected ventral skin and limb skin samples from the Prox1-GFP-expressing control (Tie1^+/+, TK1^+/− or TK2^+/−), TK1^−/− and TK2^−/− embryos at E18.5. The asterisks mark the ventral midline crossing the umbilicus (U). Arrows indicate valves (GFP^high cell clusters). Scale bars: 500 μm.

Fig. 8.

Mesenteric lymphatic phenotype in mice with Tie1 TK1/TK2 point mutations and EBD visual lymphangiography. (A-J) Direct fluorescence imaging of GFP (green) of freshly dissected mesenteries (proximal and middle jejunum) from the Prox1-GFP-expressing control (Tie1^+/+, TK1^+/− or TK2^+/−), TK1^−/− and TK2^−/− pups at E18.5, P0, P4 and P9. Arrows, mature valves; arrowheads, immature valves. (K-M) Quantification of number of lymphatic valves (K), number of branches (L) and diameter of lymphatic collecting vessels (M) from each mesentery of control, TK1^−/− and TK2^−/− pups at P0 and P9. While the severity of lymphatic developmental defects is variable in TK1^−/− pups, only those with defective mesentery lymphatics are shown and measured for quantification. Data are mean±s.e.m. N=5 per genotype. One-way ANOVA; *P<0.05, **P<0.01; NS, not significant. (N-P) Lymphangiography by intradermal injection of EBD into embryonic forelimbs at E17.5. Sites of injection are indicated by dashed circles. EBD injection into all 15 control mice (N) after 10 min resulted in uptake of EBD by lymphatic capillaries (arrowheads) and unidirectional drainage and convergence of the dye into collecting vessels (arrows). In the majority (8 of 12) of TK1^−/− embryos (O), transport of EBD in collecting vessels was attenuated when compared with that seen in control mice, while the dye spread into collecting vessels of all five TK2^−/− embryos (P) was totally missing. (Q,R) Following hind paw injection of EBD, in both control (n=5) and TK1^−/− (n=8) mice at P9 the dye was restricted to the thoracic duct (arrow) as it moved unidirectionally cephalad. Scale bars: 200 μm (A-J); 500 μm (N-R).

Fig. 9.

Lacteal development was disrupted in Tie1 TK2 point mutant mice. (A-I) Whole-mount (A-C, duodenum) or cross cryosection (D-I, proximal jejunum) immunostaining of small intestine highlights lacteal lymphatic vessel (LYVE1, green; arrows) and blood capillaries (CDH5, red) in control, TK1^−/− and TK2^−/− pups at P0. (G-I) Omitting the red channel to show LYVE1 expression (green channel). Lacteal filopodia (inset) were often observed in villi of control and TK1^−/− pups, but missing in TK2^−/− pups. (J,K) Comparisons of absolute and relative lacteal lengths in duodenum among control, TK1^−/− and TK2^−/− mice at P0. Each dot indicates mean value of five to ten villi in a mouse (n=6 mice/group). Data are mean±s.e.m. One-way ANOVA; **P<0.01; NS, not significant. DAPI was used to counterstain nuclei in D-I. Scale bars: 50 μm.

Fig. 10.

FOXC2 expression was abolished in Tie1-deficient lymphatics but restored by simultaneous loss of Foxo1. (A-R) Cryosections of control (TK1^+/−, TK2^+/− or Tie1^f/f), TK1^−/−, TK2^−/− and Ncre;Tie1^f/f mouse mesenteries at E17.5 were co-immunostained for LYVE1 and TIE1, TIE2 or FOXC2 as indicated. (D-F,J-L,Q,R) Omitting the green channel to show TIE1, TIE2 or FOXC2 expression only (red channel). Compared to control, the expression level of TIE1, TIE2 or FOXC2 in blood vessels appeared to be normal among all those Tie1 mutants. In addition, the expression level of TIE1 in lymphatic vessels of TK1^−/− and TK2^−/− was also similar to that in control. However, TIE2 staining was abolished in TK1^−/− (H,K) and TK2^−/− (I,L) lymphatics, and FOXC2 staining was almost complete lost or dramatically reduced in Ncre;Tie1^f/f (O,Q) and TK2^−/− (P,R) lymphatics, relative to the controls. Intestine wall is to the left. (S,T) Cryosections of Prox1-GFP-expressing control (Tie1^f/f;Foxo1^f/f) and Ncre;Tie1^f/f;Foxo1^f/f mouse mesenteries at P5 were co-immunostained for GFP and FOXC2, showing normal expression of FOXC2 in LECs of lymphatic collecting vessels including valve leaflets (orange arrows) in both genotypes. b, blood vessels; l, lymphatics. White arrows, positive staining in lymphatics; arrowheads, no or reduced staining in lymphatics. DAPI was used to counterstain nuclei. Scale bars: 50 μm.