Vitamin D inhibits lymphangiogenesis through VDR-dependent mechanisms

Excessive lymphangiogenesis is associated with cancer progression and renal disease. Attenuation of lymphangiogenesis might represent a novel strategy to target disease progression although clinically approved anti-lymphangiogenic drugs are not available yet. VitaminD(VitD)-deficiency is associated with increased cancer risk and chronic kidney disease. Presently, effects of VitD on lymphangiogenesis are unknown. Given the apparently protective effects of VitD and the deleterious associations of lymphangiogenesis with renal disease, we here tested the hypothesis that VitD has direct anti-lymphangiogenic effects in vitro and is able to attenuate lymphangiogenesis in vivo. In vitro cultured mouse lymphatic endothelial cells (LECs) expressed VitD Receptor (VDR), both on mRNA and protein levels. Active VitD (calcitriol) blocked LEC tube formation, reduced LEC proliferation, and induced LEC apoptosis. siRNA-mediated VDR knock-down reversed the inhibitory effect of calcitriol on LEC tube formation, demonstrating how such inhibition is VDR-dependent. In vivo, proteinuric rats were treated with vehicle or paricalcitol for 6 consecutive weeks. Compared with vehicle-treated proteinuric rats, paricalcitol showed markedly reduced renal lymphangiogenesis. In conclusion, our data show that VitD is anti-lymphangiogenic through VDR-dependent anti-proliferative and pro-apoptotic mechanisms. Our findings highlight an important novel function of VitD demonstrating how it may have therapeutic value in diseases accompanied by pathological lymphangiogenesis.

Vitamin D (VitD) participates in many biological actions in the body, and hence is regarded as "a pleiotropic hormone" 18 . Beyond its cognate function in bone metabolism and calcium homeostasis, "nonclassical" effects of VitD have been shown protective in a variety of pathophysiological conditions such as cardiovascular disease, cancer and immunity 19 . These effects have been mostly linked to the expression of VitD Receptor (VDR) in tissues that are not classically involved in calcium homeostasis 19 . The ability of VitD to modulate the expression of genes involved in biological processes such as cell proliferation, apoptosis, oxidative stress, inflammation, matrix homeostasis and angiogenesis, makes VitD an attractive candidate therapeutic for cancer and cardiovascular diseases 20,21 . Blood vascular endothelial cells express VDR 22 and VitD deficiency has been shown to be associated with increased arterial stiffness and endothelial dysfunction, underscoring the importance of VitD for maintaining endothelial health and function 23 .
Conversely, the data regarding the effect of VitD on angiogenesis is equivocal. Depending on the model, VitD analogues have been shown to have no effect on angiogenesis 24,25 , to have pro-angiogenic effects 26 , or to have anti-angiogenic effects 27 . However, no data have been reported on the potential effects of VitD on lymphatic endothelial cells (LECs) and lymphangiogenesis. The impact of VitD and its analogues on renal pathophysiology has been extensively studied [28][29][30][31][32][33] . Most of the results from these studies suggest that VitD may be considered a generally safe and effective therapeutic in diverse kidney diseases, but the effect of VitD on renal lymphatics and lymphangiogenesis has not yet been described.
Given the beneficial effects of VitD treatment in renal disease, and the deleterious associations of lymphangiogenesis with renal disease, we here tested the hypothesis that VitD has direct anti-lymphangiogenic effects and is able to attenuate lymphangiogenesis.
To address this, we here examined the direct effects of VitD on LEC tube formation, proliferation and apoptosis in vitro, and the effects of VitD treatment on lymphangiogenesis in vivo.

Results
Cultured lymphatic endothelial cells (LECs) express VDR. Immunohistochemistry (IHC) was performed to confirm lymphatic endothelial phenotype of cultured cells. As shown in Fig. 1a, compared to the negative control staining, LECs expressed Prospero homeobox protein 1 (Prox-1), VEGFR3, and Podoplanin. This combination of markers is expressed exclusively by LECs. The presence of VDR protein was confirmed by western blotting using anti-VDR antibody D6 (Fig. 1b). Mouse tubular epithelial cells (MTCs) were considered positive controls since mouse renal tubular epithelium is known to express VDR 34 . Western blot results indicate that LECs clearly express VDR albeit at lower levels compared with MTCs. Immunofluorescent staining with the same antibody confirmed VDR expression in LECs (Fig. 1c).

Calcitriol inhibits tube formation of LECs in a VDR-dependent manner.
To investigate a potential lymphangiostatic effect of calcitriol on LECs, tube formation assays were performed. Interestingly, calcitriol appeared to have lymphangiostatic effect, and dose-dependently decreased lymphatic capillary tube formation (Fig. 2a). Quantitative analyses expressed as both total tube length ( Fig. 2b) and total number of branching points (Fig. 2c) revealed reduced tube formation in the presence of calcitriol. These data indicate that the VitD analogue calcitriol exerts anti-lymphangiogenic effects in vitro. We next examined whether the observed effects on tube formation were VDR-dependent. At mRNA level, qRT-PCR indicated that VDR expression in LECs transfected with VDR siRNA was reduced remarkably compared to scramble siRNA knock-down after 48 h (81% and 62% reduction compared to scramble in siRNA 1 and siRNA 2, respectively) (Fig. 3a). Immunofluorescent staining showed that VDR siRNA transfected cells had virtually no VDR protein expression when compared with non-transfected and scramble siRNA transfected cells (Fig. 3b). Using VDR and scramble siRNA transfected cells in the tube formation assay, we showed that VDR knock-down almost completely abolished the inhibitory effect of calcitriol on tube formation (Fig. 4a). Quantification showed that the significant reduction of tube formation induced by calcitriol was abolished by VDR siRNA, both in terms of tube length ( Fig. 4b) but also numbers of branching points (Fig. 4c). These data indicate that calcitriol-induced impairment of tube formation is VDR-dependent.
Effects of calcitriol on LEC proliferation and apoptosis. Next we sought to identify the mechanism underlying VitD/VDR-mediated reduction in tube formation, focusing on LEC proliferation and induction of apoptosis. We studied the effect of calcitriol on LEC proliferation in vitro using Cell Proliferation Reagent WST-1. No effect of calcitriol on proliferation was observed after 24 h incubation (Fig. 5a). However, after 48 h, the lower calcitriol concentrations (0.1 nM, 1 nM, and 10 nM) showed a marked increase in proliferation compared with control. However, the highest concentration (100 nM), which had a clear inhibitory effect on tube formation, also significantly reduced this increase of proliferation to the level of placebo control. We next analyzed possible induction of apoptosis. To accomplish this, LECs were plated and after 24 h of attachment and subsequent 24 h serum starvation, cells were treated with calcitriol (100 nM) or vehicle alone for 6 h. Untreated cells served as negative control (medium). To detect apoptotic cells staining for Annexin V and 7-AAD was performed followed by flow cytometry. Apoptotic cells are Annexin V + 7-AAD − . Figure 5b shows representative dot plots of non-treated and calcitriol-treated LECs. Quantitative analysis indicated that 100 nM calcitriol induces significant LEC apoptosis (Fig. 5c). Paricalcitol treatment prevents lymphangiogenesis in vivo. To translate our in vitro findings to the in vivo setting, we analyzed whether the VitD analogue paricalcitol is able to attenuate proteinuria-associated lymphangiogenesis in a rat model. We previously showed that in the Adriamycin Nephrosis rat model, massive lymphangiogenesis occurs between weeks 6 to 12 after induction of disease 35 . Therefore, we treated animals in this time frame with paricalcitol to test whether this can impair adriamycin-induced lymphangiogenesis in the kidney. The results of LV quantification in the kidney of proteinuric rats receiving paricalcitol (160 ng/kg three times per week p.o.) showed a significant inhibitory effect on renal lymphangiogenesis compared to vehicle-treated proteinuric rats (Fig. 6a,b). This experiment confirms, as a proof of principle, that treatment with a VitD analogue is also able to prevent lymphangiogenesis in vivo.

Discussion
The present study reveals an inhibitory effect of vitamin D (VitD) on lymphangiogenesis. The results demonstrate that in the presence of active VitD (calcitriol), LEC tube formation and proliferation were attenuated whilst apoptosis was induced in vitro. VDR knock-down in LECs reversed the inhibitory effect of calcitriol on LEC tube formation, confirming that this effect is VDR-dependent. In vivo, paricalcitol treatment significantly decreased lymphangiogenesis in the kidneys of adriamycin-induced proteinuric rats. This inhibitory effect of VitD may  Numerous studies have highlighted the fundamental role of lymphatic vessels and lymphangiogenesis in physiological and pathological conditions 1-3 . Both blocking and promoting lymphangiogenesis have been suggested to be advantageous, and therefore imply to be highly context-dependent. For example, inducing lymphangiogenesis has been shown to decrease interstitial fluid accumulation in experimental models of lymphedema 36,37 , and also reported to be of advantage in the resolution of inflammation 11,12,38 . A recent study by Chi et al. 39 showed that promoting lymphangiogenesis by VEGF-C156S increased clearance of interstitial hyaluronan (HA), and could improve graft outcomes in an animal model of lung transplantation. Contrary to these findings, inhibiting lymphangiogenesis in several experimental transplantation models improved graft survival [6][7][8][9][10] . Whether lymphangiogenesis is a 'friend or foe' in kidney transplantation, is still unclear 4 . Lymphangiogenesis has also sparked a great deal of interest in cancer therapy. Numerous studies have demonstrated the feasibility of anti-lymphangiogenic therapy to prevent tumor metastasis 1-3,16,40 . Nevertheless, despite the long history of research on the mechanism of lymphangiogenesis and encouraging experimental results, clinically approved anti-lymphangiogenic drugs are still not available.
VitD has become a focus of intense interest both because of its vital roles in health, but also as an important immune modulator. The beneficial therapeutic effects of VitD and VitD analogues have been proposed in many conditions, ranging from proteinuria, fibrosis, atherosclerosis and inflammation to cancer therapy [28][29][30][31][32][33]41 . It has been found that VDR regulates at least 229 genes through binding to 2776 genomic DNA binding sites. These genes target anti-proliferative, pro-apoptotic, anti-inflammatory, angiostatic, immune regulatory and pro-differentiation functions, and affect many cell functions in tissue-and cell-specific manners 41 .
Here we demonstrated the expression of VDR in mouse LECs in vitro, both at the mRNA and protein expression levels. Consistent with these findings, the inhibitory effect of LEC tube formation was completely blocked after gene knock-down of VDR in LECs, highlighting the dependence of functional VDR in mouse LECs for reduced lymphangiogenesis. Taken together, these experimental findings shown that calcitriol blocks LEC tube formation in a VDR-dependent manner. Blood endothelial cells also express VDR 22 , and active VitD has been shown to induce both angiogenic and anti-angiogenic effects on blood endothelial cells [24][25][26][27][42][43][44][45] . The effect of VitD on blood endothelial cell proliferation is also contradictory. Calcitriol has been shown to inhibit the proliferation of human umbilical vein endothelial cells (HUVEC) 46 , while another report found an induction of proliferation through a nitric oxide-dependent mechanism in the same cell type 26 . These inconsistencies might reflect differences in treatment concentrations or origin of cells/tissues that were used.
The effect of calcitriol on LEC proliferation and apoptosis was studied in order to gain additional insights into mechanisms underlying the VitD-mediated suppression of LEC tube formation. Our results showed a concentration-dependent effect of calcitriol on LEC proliferation. While the lower concentrations promoted proliferation, the highest used concentration significantly prevented this increase of proliferation. Such a biphasic effect of the VitD on proliferation has also been reported in mouse epidermal keratinocytes 47,48 where low concentrations of VitD promoted proliferation while higher concentration (10-1000 nM) inhibited proliferation. This result may provide important information regarding physiological and pharmacological actions of VitD, respectively. The same calcitriol concentration (100 nM) also significantly blocked LEC tube formation, which may explain, at least in part, the underlying mechanism(s). One might argue that the high (100 nM) VitD might have toxic effects and therefore impairs tube formation and proliferation, and promotes apoptosis. This is however unlikely since VDR-knock down fully restored tube formation even in the presence of 100 nM calcitriol.
Programmed cell-death or apoptosis, a key mechanism in cancer therapy, can be also induced by VitD through repressing the expression of anti-apoptotic proteins such as Bcl2 and Bcl-X, as well as by inducing the expression of pro-apoptotic proteins such as BAX, BAD, and BAK [49][50][51][52] . Although in current study we did not delve deeply into detailed mechanisms of apoptosis, we did find that calcitriol significantly induced apoptosis in LECs, which might be another reason for capillary tube suppression in the presence of calcitriol.
Moreover, treatment with the VitD analogue paricalcitol in the in vivo proteinuric adriamycin-induced nephrosis model significantly blocked development of renal lymphangiogenesis compared to vehicle-treated rats. Paricalcitol is a synthetic VitD 2 agonist of the VDR, and is considered a selective VDR activator 53,54 . Although the affinity of paricalcitol for the VDR is reported to be about three times less than that of calcitriol, calcemic and phosphatemic effects of paricalcitol are 10 times lower 55 . Consequently, we preferred to use paricalcitol which has more selective VDR activity and lower risk of side effects such as increase in calcium concentrations or hypercalcemia (in very high concentrations). Our in vivo observations are again consistent with our data on the inhibitory effect of calcitriol in vitro.
In summary, in this work we show for the first time the expression of VDR by lymphatic endothelial cells. Our observations provide further information that calcitriol-induced inhibition of LEC tube formation may reflect the anti-proliferative and pro-apoptotic effects of VitD in this model. More in depth and mechanistic studies are still required to elucidate the molecular mechanisms responsible for the observed inhibitory effects of VitD on LECs. In keeping with our in vitro observations, we also showed the anti-lymphangiogenic effect of paricalcitol in the adriamycin-induced nephrosis rat model. These findings describe potentially important novel therapeutic effects of VitD since blocking lymphangiogenesis could be useful in treating conditions exhibiting detrimental, unfavorable lymphangiogenesis.

Western blotting.
To determine VDR expression in MTCs and LECs at the protein level, western blotting was performed. Cells were washed with the Tris-buffered saline (TBS, 20 mM Tris, pH 7.5, 150 mM NaCl) and homogenized in Radio immunoprecipitation assay buffer (50 mM Tris-HCl, 150 mM NaCl, 0.1% SDS, 0.1% Igepal in 0.5% sodium deoxycholate with 1 tablet of protease inhibitor cocktail and 1 tablet of phosphatase inhibitor; Roche Diagnostics, Mannheim, Germany). The lysates were mixed with loading buffer to make a total volume of 20 μ l. Samples were then spun down, boiled at 100 °C for 5 min in a thermo block, spun down again and put on ice. 20 μ l of every sample was loaded on a denaturizing 7.5% polyacrylamide gel. On every gel, a protein ladder was included to indicate band size. The gel was run for 15 min at 40 V and 75 min at 120 V. After running, the gel was soaked for 10 minutes in blotting buffer. The gel was then transferred on a nitrocellulose membrane followed by blotting at 150 mA for 45 min. Afterwards the membrane was blocked with TBS with 5% milk powder and 0.1% Tween 20 for 2 h at 4 °C. The membrane was incubated overnight at 4 °C with mouse anti-VDR D6 (1:500) (Santa Cruz Biotechnology, Inc) diluted in TBS-T (TBS with 0.1% Tween 20) with 1% milk powder. The membrane was washed twice followed by incubation for 2 h at RT with rabbit anti-mouse Ig horseradish peroxidase (HRP) (DAKO, Heverlee, Belgium) diluted 1:1000 in TBS-T with 1% milk powder. The membrane was washed twice and HRP activity was visualized with a chemiluminescence kit (NEL103E001EA, PerkinElmer ® ) according to the manufacturer's instructions. The chemiluminescence membrane was recorded on an ImageQuant ™ LAS 4000 system.
IHC was performed on paraffin sections using the same protocol as described above but HRP-conjugated secondary antibodies were diluted in PBS + 1% normal rat serum. HRP-activity was visualized using 3,3′ Scientific RepoRts | 7:44403 | DOI: 10.1038/srep44403 -diaminobenzidine (DAB, Sigma-Aldrich, USA) as chromogen with hematoxylin counterstaining. Slides were scanned on a Nano Zoomer HT whole slide scanner (Hamamatsu Photonics K.K., Shizuoka Pref., Japan) and sections were analyzed using Aperio Image Scope software (Aperio Technologies Inc, Vista, CA, USA).
Tube formation assay. Matrigel tube formation assays were performed to investigate lymphangiogenesis in vitro. The endothelial cell tube formation assay on basement membrane (including Matrigel) is widely used to study angiogenic potential in vitro (reviewed in ref. 56). Although vascular endothelial cells are most commonly used in this assay, it has also been described useful to study lymphangiogenesis 57 . Although not visible at the microscopical level, studies performed at the ultrastructural level (i.e. electron microscopy) revealed that the tubes formed when using either vascular endothelial cells 58 or LECs 59 contain a lumen. Based on this data the LEC tube formation assay appears to be a suitable model to study lymphangiogenesis in vitro. Here 15-well μ -Slide Angiogenesis (Cat. No. 81501, ibidi, Martinsried, Germany) and Growth Factor Reduced Matrigel (Cat. No. 356230, BD Biosciences, Bedford, MA) were used. 10 μ L of Matrigel was placed per well and incubated at 37 °C for 30 minutes. 50 μ L of 1.4 × 10 5 cells/mL LECs incubated in DMEM were added to each well with or without 1α ,25-Dihydroxyvitamin D 3 (calcitriol) (D1530, Sigma Aldrich, USA). Calcitriol was added to the medium to reach a final concentration of 0.1 nM, 1 nM, 10 nM or 100 nM (performed in triplicates). Cells were incubated for 4 h at 37 °C in humidified air with 5% CO 2 . Tube formation assay was also performed on VDR knock-down cells. No siRNA-transfected, VDR siRNA-transfected and scramble siRNA-transfected LECs were seeded into μ -Slide Angiogenesis pre-coated with 10 μ l growth factor-reduced Matrigel. Simultaneously, cells were treated with 100 nM calcitriol or PBS. After 4 h incubation the capillary tubes were analyzed. Representative images were captured with an Olympus inverted phase-contrast microscope (Olympus Optical C., Melviller, NY, USA) equipped with the Quick Imaging System. From each well 5 images were taken (at 100 x magnification) at fixed areas (starting at the center of the well and then above, below, right, left of the center). To perform quantifications, both the total cumulative tube length (i.e. total cumulative length of all tubes visible in the photomicrographs) and the total number of branching points were determined in all 5 pictures per well and corrected by the image surface using Image J 1. Proliferation assay. The effect of 1α ,25-Dihydroxyvitamin D 3 (calcitriol) (D1530, Sigma Aldrich, USA) on proliferation of LECs was determined by plating cells in 96-well gelatin-coated plates (2,000 cells per well) in 200 μ l medium. The cells were incubated at 37 °C for 3.5 hours for adherence and then treated with vehicle (PBS) or various concentrations of calcitriol for 24 h or 48 h. For cells cultured for 48 h (with or without calcitriol), medium was replenished with new medium containing the appropriate calcitriol concentration after 24 h. Cell proliferation was quantified using Cell Proliferation Reagent WST-1 (100 μ l/well) according to the manufacturer's instructions and absorption was measured after three hours of WST-1 incubation.
Apoptosis assay. LECs were plated in 6-well culture plates in DMEM at a density of 1.5 × 10 5 cells/ml. Cells were allowed to attach for 24 h, starved for additional 24 h and subsequently stimulated with 100 nM calcitriol for 6 h. Non-stimulated cells (no calcitriol, medium only) served as negative control. A positive control for the staining procedure was obtained by addition of 0.4 mM H 2 O 2 to LECs for 6 h to induce cell death. Cells were trypsinized, washed in PBS twice and stained with anti-Annexin V Apoptosis Detection Kit and 7-AAD (Cat: 640922, Uithoorn, The Netherlands) according the supplier's protocol. Fluorescence was assessed on a FACS Calibur flow cytometer (BD Bioscience, Breda, The Netherlands) within the same day. At least 10,000 gated events were collected per sample and data was analyzed using Flowjo software (Tree Star, Inc., Ashland, OR, USA).

Induction of proteinuria in rats. Adriamycin Nephrosis (AN) was induced in Wistar rats (Harlan, The
Netherlands) by a single injection of 1.5 mg/kg adriamycin (Doxorubicin ® ) into the tail vein as previously described 37 . After 6 weeks animals were placed in 24-hour metabolic cages for urine collection, and then rats were stratified according to the level of proteinuria into two experimental groups. Subsequently, at six weeks after adriamycin injection, renal biopsy was taken via abdominal incision in order to evaluate kidney histology and numbers of LVs. Animals were treated then with paricalcitol 160 ng/kg three times per week p.o. (paricalcitol/ ethanol dissolved in water, n = 8) or vehicle (ethanol diluted in water, n = 6) from week 6 to 12. At the end of the study, blood and kidneys were collected for further analysis. All animals received care in compliance with the recommendations of ARRIVE, the Directive 2010/63/EU and the Dutch Law on Experimental Animal Care, and were approved by the animal ethics committee of the University of Groningen.