ERG Regulates Lymphatic Vessel Specification Genes and Its Deficiency Impairs Wound Healing-Associated Lymphangiogenesis

Abstract

Objective: Rarefaction of blood and lymphatic vessels in the skin has been reported in systemic sclerosis (SSc) (scleroderma). E26 transformation-specific-related factor (ERG) and Friend leukemia virus-induced erythroleukemia 1 (FLI-1) are important regulators of angiogenesis, but their role in lymphatic vasculature is lesser known. The goal of this study was to determine the role of ERG and FLI-1 in postnatal lymphangiogenesis and SSc lymphatic system defects.

Methods: Immunofluorescence was used to detect ERG and FLI-1 in skin biopsy samples from patients with SSc and healthy controls. Transcriptional analysis of ERG or FLI-1-silenced human dermal lymphatic endothelial cells (LECs) was performed using microarrays. Effects of ERG and FLI-1 deficiency on in vitro tubulogenesis in human dermal LECs were examined using a Matrigel assay. ERG and FLI-1 endothelial-specific knockouts and ERG lymphatic-specific knockouts were generated to examine vessel regeneration in mice.

Results: ERG and FLI-1 protein levels were reduced in the blood and lymphatic vasculature in SSc skin biopsy samples. ERG levels were shown to regulate genes involved in lymphatic vessel specification, including vascular endothelial growth factor receptor 3/FLT-4, lymphatic vessel endothelial hyaluronan receptor 1, SOX-18, and prospero homeobox 1 (PROX-1), whereas FLI-1 enhanced the function of ERG. The ERG-FLT-4 pathway regulated in vitro tubulogenesis in human LECs. Deficiency of ERG or FLI-1 similarly impaired the function of blood vessels in mice. However, only ERG deficiency affected the regeneration of lymphatic vessels during wound healing.

Conclusion: ERG and FLI-1 are essential regulators of blood and lymphatic vessel regeneration. Deficiency of ERG and FLI-1 in SSc endothelial cells may contribute to the impairment of blood and lymphatic vasculature in patients with SSc.

Figure 1:. Reduced ERG and FLI1 expression in dermal BEC and LEC in SSc skin

A,B. Representative immunofluorescent staining of ERG (A) and FLI1 (B) in BEC (green, CD31+ PDPN−) and LEC (red, CD31 weak, PDPN+) in healthy control (HC) and SSc skin. Lower panels show DAPI and ERG alone (scale bar = 20 um). The same vessels (serial sections) are shown in A and B. C,D. A thresholded area for ERG+ and FLI1+ pixels was normalized to each individual vessel area and each data point represents the average of all vessels for one biopsy. E,F. Average mean ERG or FLI1 nuclear intensity for all BEC and LEC within one biopsy is plotted. Asterisks in C-F mark values for the two skin biopsies shown in A-B. Mean and standard deviations are plotted with the p values determined from Mann-Whitney U test.

Figure 2:. Increased extranuclear ERG and FLI1 expression in BEC and LEC in SSc skin.

A. Examples showing tight nuclear location of ERG in HC and both nuclear and extracellular localization in SSc. B-C (scale bar = 20 um). The ratio of cytoplasmic to nuclear ERG intensity for each BEC or LEC nucleus is plotted as a function of nuclear circularity (1 = perfect circle). D. Average cytoplasmic/nuclear ERG ratio for BEC and LEC for each patient. E-G. Similar analysis for FLI1 localization as per B-D. Mean and standard deviations are plotted with the p values determined from Mann-Whitney U test. E,G asterisks mark the samples shown in figure 1A–B.

Fig. 3:. ERG and FLI1 regulate key genes involved in lymphatic vessel specification.

A. PCA analysis of cultured healthy donor dermal LECs treated with siSCR, siERG, siFLI1, or siERG/FLI1 for 72 hours B. Heatmap of differentially expressed genes determined by microarray. The top 500 genes were used to perform unsupervised hierarchical clustering. C-E. Volcano plots of differentially expressed genes. F. LECs were treated with siERG, siFLI1, and siERG/FLI1 and mRNA levels of indicated genes were measured by qPCR. Each dot represents an independent experiment. *p<0.05. **p<0.01, ***p<0.001. G. Representative immunoblot of indicated genes. H. Quantification of immunoblots. Values are means±SEM of the relative protein levels by densitometry, n=3–5/group, *p<0.05, **p<0.01.

Fig. 4:. FLT4 mediates ERG-dependent tubulogenesis.

A. SSc sera reduce ERG and FLT4 levels in LECs. Dermal human LECs were treated with 5% each FBS, healthy control (HC) serum or SSc serum. Immunoblot of indicated genes (representative blot of 3 independent HC or SSc sera is shown). B. Quantification of immunoblots, each point is one patient. Values are means±SEM of the relative protein levels by densitometry, n=6/group, *p<0.05, **p<0.01. C,D. In vitro tube formation in Matrigel was evaluated in healthy donor human LECs treated with scrambled, ERG, FLI1 1, or ERG + FLI11 (EF) siRNAs. Representative image of capillary network formed in Matrigel 24 hours after plating and quantitative analysis of the number of tubes. E,F. Tube formation in Matrigel was evaluated in LECs treated with scrambled and FLT4 siRNAs. G,H. Tube formation in Matrigel was evaluated in LECs treated with scrambled and ERG or FLI1 siRNAs in the presence or absence of AdenoFLT4. I,J. Tube formation in Matrigel was evaluated in LECs treated with scrambled and ERG + FLI1 (EF) siRNAs in the presence or absence of AdenoFLT4. Each dot represents an independent experiment. *p<0.05. **p<0.01, ****p<0.001, n=5–8.

Fig. 5:. Impaired neo-angiogenesis in Erg-and Fli1CKO mice.

A. Endothelial cells (CD31⁺CD45⁻) were purified from WT and ErgCKO and Fli1CKO mice by FACS as previously described (40). Immunoblot of Erg and Fli1 proteins in WT and ErgCKO and Fli1CKO mutant mice. Representative images from 3 mice/group are shown. B,C. Representative images of wound margins in Erg- and Fli1CKO mice. D. Bar graph representation of the days required for full re-epithelization of control and ErgCKO or Fli1CKO mice (POD, postoperative day). *<0.05, 3–8 mice per group. E,G. The vascular network in the wound of ErgCKO, Fli1CKO and littermate controls visualized after re-epithalization using FITC injection (orange circle indicates wound area). Panels F,H show skin areas adjacent to the wounds. I-L. Vessel diameter was measured in CD31+ vessels in healed wounds from control and Erg- and Fli1CKO mice, n=3–5 and diameters were counted in each section in three random views, *p<0.05.

Fig. 6:. Impaired neo-lymphanangiogenesis in Erg deficient mice.

A. Lymphatic vessels in the re-epithalized wound areas of ErgCKO and Fli1CKO mice were evaluated by double staining with CD31 and PDPN. Arrows indicate double-positive cells. The bar graphs depict number of double-positive vessels per mm², **p<0.01. B. Whole mount fluorescence images of axillary lymph nodes at 1 hr after FITC-dextran dye injection in the skin of ProxErgCKO mice and control littermates at day 7 post wounding. C. Using image J, a threshold was chosen that highlights the bright areas of fluorescence and graph quantifies the percent thresholded area within the entire lymph node. *<0.05, (n=4 mice per group). D. Representative images of lymphatic vessels in the skin of ProxErgCKO/TmG and WT controls at day 7 post wounding. E. The bar graphs depict number of GFP+ vessels per mm². *<0.05, (n= 5 controls and 3 ProxErgCKO/TmG mice).