Slit-Robo signaling mediates lymphangiogenesis and promotes tumor lymphatic metastasis

Abstract

The Slit family of guidance cues binds to Roundabout (Robo) receptors to modulate neuronal, leukocytic, and endothelial migration. Slit-Robo signaling had been reported to function as chemoattractive signal for vascular endothelial cells during angiogenesis. In this study, we found that Robo1 was expressed in lymphatic endothelial cells to mediate the migration and tube formation of these cells upon Slit2 stimulation, which were specifically inhibited by the function-blocking antibody R5 to Slit2/Robo1 interaction. To further explore the lymphangiogenic effect and significance mediated by Slit-Robo signaling, we intercrossed Slit2 transgenic mice with a non-metastatic RIP1-Tag2 mouse tumor model, and found that transgenic overexpression of Slit2 significantly enhanced tumor lymphangiogenesis and subsequently promoted mesenteric lymph node metastasis of pancreatic islet tumors. Taken together, our findings reveal that through interacting with Robo1, Slit2 is a novel and potent lymphangiogenic factor and contributes to tumor lymphatic metastasis.

Fig. 1

Robo1 is expressed in primary lymphatic endothelial cells. (A-B) Semiquantitive real-time RT-PCR analysis of the expression of Robo1 in hLECs relative to HUVECs. PCR products were run in agarose gel to confirm the proper size of the amplified fragments, with β-actin as internal control (A). NTC is no template negative control. (C) Robo1 (green) was detected on the membrane and cytoplasm of hLECs, with LYVE-1 (red, upper) or Prox-1 (red, lower) as markers of hLECs. Nuclei were identified by DAPI. Scale bar: 50 μm. Results are representatives of three independent experiments.

Fig. 2

Slit2 induces tube formation and migration of lymphatic endothelial cells, which are inhibited by function-blocking antibody R5 to Slit2/Robo1 interaction. (A) Affinity purified recombinant Slit2 was silver stained and immunoblotted with S1 (an IgG_2a mAb to the NH₂-terminal portion of human Slit2), and recombinant hRobo1-Fc was analyzed by Coomassie Brilliant Blue (CBB) staining and immunoblotted with HRP-conjugated secondary antibody. (B) Recombinant Slit2 interacts with hRobo1-Fc in an ELISA assay. Plate wells were coated with recombinant Slit2 in different concentrations or buffer control (−). The experiment was carried out in triplicate and was repeated three times. (C) Tube formation of hLECs on Matrigel visualized by phase-contrast microscopy. Cells were respectively treated with VEGF-C, Slit2 or Slit2 plus function-blocking Ab R5. Scale bar, 50 μm. (D) Normalization analysis of the effects of recombinant Slit2 and VEGF-C on the tube formation of hLECs, with R5 for inhibition and mouse IgG for control. Results were calculated as mean ± SD from measurements of photographed tubular structures. (E-F) The effects of recombinant Slit2 and VEGF-C on hLECs migration were measured using the Boyden chamber assay. Recombinant Slit2 induced migration in a dose-dependent manner (E). Comparing to control IgG, R5 effectively inhibited Slit2-induced but not VEGF-C-induced migration of hLECs (F). Results were calculated as mean ± SD from measurements of three separate experiments. * P < 0.05; ** P < 0.01.

Fig. 3

Overexpression of Slit2 promotes lymphangiogenesis and lymphatic invasion of pancreatic islet tumors in Slit2;RIP1-Tag2 mice. (A) Schematic display of FLAG-tagged human Slit2 (hSlit2) transgene construct. The FLAG sequence was inserted in front of the Slit2 cDNA and the transgene was under the control of CMV promoter. The lengths of the DNA fragment are given in base pairs (bp). (B) Slit2 expression. Western blot analysis of pancreas extracts from age-matched wide-type (WT) versus Slit2 transgenic mice and RIP1-Tag2 versus Slit2;RIP1-Tag2 mice. Tubulin was used as loading control. (C) Representative immunohistochemical staining of LYVE-1 in pancreas of 14-week-old wild-type (WT) or Slit2 transgenic (Slit2 Tg) mice and the corresponding normalization analysis of pancreas lymphatics (right). Results are mean ± SD values of the total lymphatic lengths from a minimum of four mice per group. * P < 0.05. Scale bar: 100 μm. (D) Representative images of tumor cells invading the intra-tumoral (left) or peri-tumoral lymphatic vessels (right) of 14-week-old Slit2;RIP1-Tag2 mice. Arrow heads indicate LYVE-1-labeled lymphatic vessels with invading tumor cells. Scale bar: 100 μm. (E) Robo1 (green) was expressed in LYVE-1-positive (red) tumor lymphatics. Arrows indicate co-localization of Robo1 with LYVE-1. T: tumor; LV: lymphatic vessel. Scale bar: 40 μm. (F) Representative immunohistochemical staining of LYVE-1 in pancreas of 14-week-old RIP1-Tag2 (left) or Slit2;RIP1-Tag2 (middle) mice and the corresponding normalization analysis of tumor lymphatics (right). Results are mean ± SD values of the total tumoral lymphatic lengths from a minimum of four mice per group. * P < 0.05. Scale bar: 100 μm.

Fig. 4

Slit2;RIP1-Tag2 mice have increased regional lymph node tumor metastasis and decreased survival rate compared with RIP1-Tag2 mice. (A) Tumor burdens and tumor numbers (mean ± SD) of Rip1-Tag2 (n = 8) and Slit2;Rip1-Tag2 (n = 9) mice at 14 weeks showed no significant differences. (B) Identification of islet β-cell metastases. Macroscopic image of primary islet tumor and metastases was shown (a). Arrow head points to a primary tumor. Arrows point to a mesenteric lymph node metastasis and an intestinal metastasis. Microscopic images indicate tumor metastasis (arrows) colonized into the submucosal layers (arrow head) of the small intestine (b) and the mesenteric lymph nodes (c). (C) Kaplan-Meier survival curves of RIP1-Tag2 (n = 35) and Slit2;RIP1-Tag2 (n = 13) mice. P = 0.0117 analyzed by log-rank test. (D) Blood insulin levels in 14-week-old RIP1-Tag2 (n = 5) and Slit2;RIP1-Tag2 (n = 5) mice showed no significant difference, determined by ELISA.