VEGF-C gene therapy augments postnatal lymphangiogenesis and ameliorates secondary lymphedema

Abstract

Although lymphedema is a common clinical condition, treatment for this disabling condition remains limited and largely ineffective. Recently, it has been reported that overexpression of VEGF-C correlates with increased lymphatic vessel growth (lymphangiogenesis). However, the effect of VEGF-C-induced lymphangiogenesis on lymphedema has yet to be demonstrated. Here we investigated the impact of local transfer of naked plasmid DNA encoding human VEGF-C (phVEGF-C) on two animal models of lymphedema: one in the rabbit ear and the other in the mouse tail. In a rabbit model, following local phVEGF-C gene transfer, VEGFR-3 expression was significantly increased. This gene transfer led to a decrease in thickness and volume of lymphedema, improvement of lymphatic function demonstrated by serial lymphoscintigraphy, and finally, attenuation of the fibrofatty changes of the skin, the final consequences of lymphedema. The favorable effect of phVEGF-C on lymphedema was reconfirmed in a mouse tail model. Immunohistochemical analysis using lymphatic-specific markers: VEGFR-3, lymphatic endothelial hyaluronan receptor-1, together with the proliferation marker Ki-67 Ab revealed that phVEGF-C transfection potently induced new lymphatic vessel growth. This study, we believe for the first time, documents that gene transfer of phVEGF-C resolves lymphedema through direct augmentation of lymphangiogenesis. This novel therapeutic strategy may merit clinical investigation in patients with lymphedema.

Figure 1

(a–d) Rabbit ear model of lymphedema: effect of phVEGF-C gene therapy. (a) Postoperative appearance of the dorsal surface of the rabbit ear. Lymphedema surgery leaves a gap of cartilage crossed only by the skin bridge. (b) A view under a surgical microscope after lifting up the skin bridge showing neurovascular bundle. Lymphatic vessels were visualized as blue lines (arrows) due to the uptake of Evans blue. Ear thickness (c) and volume (d) show consistent differences between the VEGF-C and saline groups over 12 weeks. *P < 0.05; **P < 0.01. (e–i) Decreased skin thickness after phVEGF-C transfer in a rabbit lymphedema model. Photos show cross sections of the skin after elastic-tissue trichrome staining 8 weeks after lymphedema surgery. Compared with normal ears (e and g), operated ears (f and h) had fibrofatty tissue deposition and thus greater skin thickness. The phVEGF-C–transfected ear shown in h shows less fibrosis and decreased thickness compared with the saline-injected ear (f), which demonstrates other characteristic features of lymphedema, such as profound epidermal hyperplasia and papillomatosis. (i) Measurement of skin thickness from histologic sections shows a significant difference between the saline and VEGF-C groups (P < 0.05). *P < 0.05; **P < 0.01. Scale bar, 500 μm. Normal-S and Normal-V indicate unoperated ears from the saline and VEGF-C groups, respectively; LE indicates lymphedema-operated ears.

Figure 2

Temporal changes of lymphatic function visualized by lymphoscintigraphy. (a and b) Orientation of the lymphoscintigraphic images. In normal ears, lymphatic flow assumes a linear pattern and the draining LNs are clearly visible. In the operated ear, the lymphatic passages were blocked, resulting in backward diffusion and no visualization of LNs. (c and d) Temporal changes in the saline group. Even at 12 weeks (d), lymphoscintigraphy demonstrates substantial impairment of lymphatic drainage of the saline-injected ear, indicated by dermal backflow and faint visualization of the LNs. (e and f) Temporal changes in the VEGF-C group. In the phVEGF-C–transfected ears, there was remarkable improvement of draining function. At 12 weeks, a linear passage of radiotracer, decreased dermal backflow, and increased uptake by LNs were observed. (g and h) Representative lymphoscintigraphic images and calculation of radioactivity index from the saline (g) and VEGF-C group (h). To quantitatively compare lymphatic drainage, the radioactivity within the ear was counted. Net radioactivity of the ear was obtained by subtracting γ counts at injection sites (arrows) from the total counts of the ear. The radioactivity index is the ratio of radioactivity of the operated ear divided by the radioactivity of the normal ear; this was used to compare lymphatic drainage function of the lymphedema ears. Higher ratios indicate more persistent radioactivity and less lymphatic drainage. (i) Comparison between the saline and VEGF-C groups shows the values were consistently lower in the VEGF-C group at 4, 8, and 12 weeks. *P < 0.05; **P < 0.01.

Figure 3

Increased expression of VEGF-C protein and VEGFR-3 mRNA in the phVEGF-C–transfected ears. (a and b) Western blot of VEGF-C protein from skin. VEGF-C was detected in its 58-kDa (a) and 31-kDa forms (b). VEGF-C protein expression was significantly higher at and around the phVEGF-C–transfected lymphedema skin. Prox, Mid, and Dist represent samples obtained from ear skin proximal to the skin bridge, skin from the bridge itself, and intact ear skin just distal to the skin bridge of the phVEGF-C–transfected ear, respectively. Neg, samples from the skin bridge of saline-injected lymphedema ear. NL, samples from the bridge site of unoperated contralateral ear. (c and d) Using degenerate oligonucleotides, RT-PCR was performed for total RNA extracted from mesentery (Mes), lung, kidney, and LNs. The PCR product (470 bp) from the kidney sample was sequenced. At the protein level, the rabbit (Rb) VEGFR-3 clone displayed 92.9%, 93.6%, and 94.3% identity with human (Hu), bovine (Bo), and mouse (Mo) VEGFR-3, respectively. (e) New primer sets were designed from the sequenced rabbit VEGFR-3 DNA, yielding a single PCR product of 362 bp. (f) Representative semiquantitative RT-PCR showing higher expression of VEGFR-3 in the lymphedema skin transfected with phVEGF-C than in the saline-injected or unoperated skin. (g) Quantification of VEGFR-3 mRNA levels. (*P < 0.001. **P < 0.01). (h and i) The effect of phVEGF-C gene transfer on tyrosyl phosphorylation of VEGFR-3 (h) and VEGFR-2 (i) by immunoprecipitation with anti-phosphotyrosine Ab followed by Western blot analysis with anti–VEGFR-3 or anti–VEGFR-2 Ab’s, respectively. Samples transfected with phVEGF-C revealed similar levels of phosphorylated VEGFR-2 compared with the control groups (saline and LacZ).

Figure 4

(a) Gene transfer of phVEGF-C decreases lymphedema in a mouse tail model of lymphedema. Tail thickness was significantly greater in the operated tail than in the unoperated tail during the entire 5 weeks. In the VEGF-C group, compared with the saline, LacZ, and VEGF₁₆₅ groups, the tail thickness was significantly smaller at 3–5 weeks (*P < 0.05). No-op, no operation. (b–m) phVEGF-C induces lymphangiogenesis in a mouse tail model of lymphedema. (b–k) Immunohistochemistry using markers of lymphatic endothelium, LYVE-1 (b–f), and VEGFR-3 (g–k), in normal (b and g) and operated (3 weeks after lymphedema surgery) skin sections from the saline (c and h), LacZ (d and i), VEGF-C (e and j), and VEGF₁₆₅ (f and k) groups. Lymphatic vessels are seen as brown color (black arrows). Note the abundance of hyperplastic lymphatic vessels in phVEGF-C–transfected sections (e and j). l and m show quantification of LYVE-1– and VEGFR-3–positive lymphatic vessels. Compared with normal and control (saline, LacZ, and VEGF₁₆₅) groups, the VEGF-C group showed significantly higher lymphatic vessel density. *P < 0.05 vs. normal; **P < 0.01 vs. LE-saline and LE-LacZ. Scale bar, 100 μm.

Figure 5

(a–j) phVEGF-C induces proliferation of lymphatic endothelial cells. Double immunohistochemistry using LYVE-1 and Ki-67 in active lymphangiogenesis site from skin sections. In a, d, and g, LYVE-1 staining of lymphatic vessels (arrows) in the dermis. In b, e, and h, green fluorescence (white arrowheads) depicts the nuclear staining of Ki-67. In c, f, and i, double fluorescence (yellow arrowheads) demonstrates Ki-67⁺ nuclei (green) in lymphatic vessels (red). Lymphatic vessels in normal skin (c) are shown negative for Ki-67. In the LacZ group, some of the lymphatic vessels contain Ki-67⁺ nuclei (f). White arrows in f show Ki-67^– lymphatic vessels. In phVEGF-C–transfected skin, most of the LYVE-1–positive lymphatic vessels are positive for Ki-67 (i), indicating that active cell division occurs in the lymphatic vessels. (j) Number of Ki-67⁺ nuclei are 2.5 times higher in the VEGF-C group. *P < 0.01 compared with normal; **P < 0.01 compared with saline and LacZ. Scale bar, 100 μm. (k–u) phVEGF-C does not increase capillary density in two animal models of lymphedema. Immunohistochemistry with CD31 (PECAM-1) in a rabbit ear (k–n) and a mouse tail (p–t) model of lymphedema on skin sections from the normal (k and p), saline (l and q), LacZ (m and r), VEGF-C (n and s), and VEGF₁₆₅ (t) groups. Vascular endothelial cells are stained red (black arrows). o and u show quantification of capillary density. Only the VEGF₁₆₅ group in the mouse tail model demonstrated significantly higher capillary density than the other groups. *P < 0.01 vs. saline, LacZ, and VEGF-C. Scale bar, 100 μm.