Novel application assigned to toluquinol: inhibition of lymphangiogenesis by interfering with VEGF‐C/VEGFR‐3 signalling pathway

Abstract

Background and Purpose

Lymphangiogenesis is an important biological process associated with the pathogenesis of several diseases, including metastatic dissemination, graft rejection, lymphoedema and other inflammatory disorders. The development of new drugs that block lymphangiogenesis has become a promising therapeutic strategy. In this study, we investigated the ability of toluquinol, a 2‐methyl‐hydroquinone isolated from the culture broth of the marine fungus Penicillium sp. HL‐85‐ALS5‐R004, to inhibit lymphangiogenesis in vitro, ex vivo and in vivo.

Experimental Approach

We used human lymphatic endothelial cells (LECs) to analyse the effect of toluquinol in 2D and 3D in vitro cultures and in the ex vivo mouse lymphatic ring assay. For in vivo approaches, the transgenic Fli1:eGFPy1 zebrafish, mouse ear sponges and cornea models were used. Western blotting and apoptosis analyses were carried out to search for drug targets.

Key Results

Toluquinol inhibited LEC proliferation, migration, tubulogenesis and sprouting of new lymphatic vessels. Furthermore, toluquinol induced apoptosis of LECs after 14 h of treatment in vitro, blocked the development of the thoracic duct in zebrafish and reduced the VEGF‐C‐induced lymphatic vessel formation and corneal neovascularization in mice. Mechanistically, we demonstrated that this drug attenuates VEGF‐C‐induced VEGFR‐3 phosphorylation in a dose‐dependent manner and suppresses the phosphorylation of Akt and ERK1/2.

Conclusions and Implications

Based on these findings, we propose toluquinol as a new candidate with pharmacological potential for the treatment of lymphangiogenesis‐related pathologies. Notably, its ability to suppress corneal neovascularization paves the way for applications in vascular ocular pathologies.

Abbreviations

dpf

days post‐fertilization

hpf

hours post‐fertilization

LEC

lymphatic endothelial cell

LYVE‐1

lymphatic vessel endothelial hyaluronan receptor‐1

MTT

3‐(4,5‐dimethylthiazol‐2‐yl)‐2,5‐diphenyltetrazolium bromide

VEGFR

VEGF receptor

Tables of Links

TARGETS
Catalytic receptors a	Enzymes b	ERK1
VEGF receptors	Akt (PKB)	ERK2
VEGFR‐3	Caspase‐3	MMP‐2
	Caspase‐7	PI3K

LIGANDS
Bevacizumab	VEGF‐A
FGF‐2	VEGF‐C
Mitomycin C

These Tables list key protein targets and ligands in this article which are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Pawson et al., 2014) and are permanently archived in the Concise Guide to PHARMACOLOGY 2015/16 (^a,bAlexander et al., 2015a, 2015b).

Introduction

The lymphatic vascular system constitutes a highly specialized part of the vascular system composed of a network of organs, lymph nodes, lymph ducts and thin‐walled and low‐pressure lymphatic vessels (Coso et al., 2014). This vasculature is present in all regions of the human body with the exception of bone marrow, the CNS, cornea and retina, and tissues such as epidermis or cartilage (Schulte‐Merker et al., 2011). The vital role of the lymphatic system is to maintain interstitial fluid homeostasis by absorbing water and macromolecules from the interstitium and transporting them back to the bloodstream. It is also essential for intestinal dietary fat and vitamin absorption and is required for the trafficking of immune cells and immune surveillance (Tammela and Alitalo, 2010; Zheng et al., 2014). Taking into account the importance of lymphatic vessels for normal physiological functions, it is not surprising that abnormalities of the lymphatic vasculature are involved in several human pathologies.

Lymphangiogenesis or the formation of new lymphatic vessels from pre‐existing ones is the major mode of lymphatic growth. This process is very active during embryonic development but is uncommon in adulthood and frequently associated with pathological conditions. Defects in lymphatic function can lead to lymph and fat deposition in tissues, impaired immune responses and tissue swelling, known as lymphoedema. In sharp contrast, excessive lymphangiogenesis is crucially involved in various chronic inflammatory conditions and graft rejection (Dietrich et al., 2010; Alitalo, 2011; Paupert et al., 2011; Karaman and Detmar, 2014). It is also well accepted that the lymphatic vasculature serves as a major route for tumour metastasis and that lymphangiogenesis contributes to cancer cell spreading from primary sites to lymph nodes (Zhang and Lu, 2014). In addition to its pivotal involvement in metastatic dissemination, lymphangiogenesis plays a key role in ocular pathologies and corneal graft rejection, where neovascularization can cause blindness (Abdelfattah et al., 2015).

Members of the VEGF family are commonly considered as key regulators of lymphatic vessel formation through binding to their receptors (VEGFRs), which are expressed at the surface of lymphatic endothelial cells (LECs) (Lohela et al., 2009; Lemmon and Schlessinger, 2010). These growth factors induce dimerization and autophosphorylation of their tyrosine kinase receptors, which then promote the activation of downstream signalling pathways involved in cellular responses (Mäkinen et al., 2001). The VEGF‐C/VEGFR‐3 axis is the best known signalling pathway in lymphangiogenesis. Its activation leads to the phosphorylation of serine kinases, such as phosphoinositide‐3‐kinase (PI3K)/Akt and MAPK (MAPK/ERK), which are crucial for LEC proliferation, tube formation, migration and survival (Zhang et al., 2010). Hence, blocking the VEGFR‐3/VEGF‐C pathway is currently the best way to suppress lymphangiogenesis and the metastatic spread of tumour cells via lymphatic vessels (Norrmén et al., 2011; Alitalo and Detmar, 2012; Quagliata et al., 2014). The inhibition of lymphangiogenesis holds promise for the treatment or prevention of tumour metastasis, but unlike the antiangiogenesis strategy, the antilymphangiogenesis approach is yet to be tested clinically (Witte et al., 2011). However, new drugs with the potential to inhibit lymphangiogenesis have recently been characterized (Bock et al., 2013; Astin et al., 2014; Hos et al., 2014).

Toluquinol, also known as 2,5‐toluenediol, is a 2‐methylhydroquinone isolated from the culture broth of the marine fungus Penicillium sp. HL‐85‐ALS5‐R004. In previous studies, we demonstrated that toluquinol has antiangiogenic activity mediated through suppression of the Akt signalling pathway (García‐Caballero et al., 2013). Toluquinol has been reported to inhibit the proliferation of blood endothelial and tumour cell lines, to suppress the capillary‐like formation and migration of bovine aortic endothelial cells, to decrease the secretion of MMP‐2 by bovine aortic endothelial cells and to induce apoptosis of tumour and blood endothelial cells. Furthermore, toluquinol inhibited angiogenesis in the chicken CAM assay, in the mouse Matrigel plug assay and in the intersegmental vessel formation and caudal fin regeneration assays in the zebrafish model (García‐Caballero et al., 2013). Therefore, toluquinol seems to be a good example of the use of natural compounds, especially those from marine organisms, as potential new sources of therapeutic compounds that could be used for the treatment of cancer or angiogenesis‐related diseases (Bhatnagar and Kim, 2010; Senthilkumar et al., 2013; Khazir et al., 2014).

In the present study, we evaluated the potential of toluquinol to inhibit lymphangiogenesis. To address this issue, a battery of in vitro, ex vivo and in vivo experimental models was used. We demonstrated, for the first time, that toluquinol inhibits LEC proliferation, migration and sprouting of new lymphatic vessels. In addition, toluquinol interfered with the in vivo lymphangiogenic process occurring during zebrafish development, with the lymphatic formation induced by VEGF‐C in mice and with corneal neovascularization. We demonstrated that toluquinol exerts its antilymphangiogenic properties initially by activating the apoptotic cascade and specifically by blocking the VEGF‐C/VEGFR‐3 signalling pathway. Altogether, by showing that toluquinol inhibits several crucial steps in the lymphangiogenic process, our data open novel pharmacological perspectives for using this drug in diseases associated with abnormal lymphangiogenesis.

Methods

Group sizes

For all the in vitro experiments, the group size for each experimental condition was five independent tests.

For the ex vivo lymphatic ring assay, the group size for each experimental condition was 10 independent mouse rings collected from five different mice.

For the in vivo models, the group sizes for each experimental condition were:

• Thoracic duct formation in the zebrafish model: n = 50 embryos per condition.

• Mouse ear sponge assay: n = 5 mice per condition (two sponges per mouse with a total of 10 sponges from five mice per condition). The two samples from each mouse were averaged, and the averaged number was used as a single ‘n’ for statistical analysis.

• Mouse corneal neovascularization model: n = 15 mice per condition (two corneas per mouse with a total of 30 corneas from 15 mice per condition). The two samples from each mouse were averaged, and the averaged number was used as a single ‘n’ for statistical analysis.

Randomization

For all the experimental procedures, control and treated samples or animals were subjected to the same ambient conditions (identical tanks or cages, same incubator, identical behaviour, room lighting, temperature, etc.) and identical experimental settings, in order to obtain differences related only to drug incubation.

Blinding

All the quantifications were blinded. Once the operators performed the experimental battery, they assigned the nomenclature based on the letters A, B, C and so on to the different conditions. Therefore, the analyst did not know the origin of the data during quantification of samples.

Normalization

In our work, data normalization was employed in the cases described as follows:

• Cell growth analysis assessed by the 3‐(4,5‐dimethylthiazol‐2‐yl)‐2,5‐diphenyltetrazolium bromide (MTT) dye reduction test. With this approach, increasing the amounts of viable cells results in increased purple colouring, and the cellular proliferation is proportional to the absorbance measured at 550 nm. In control wells (absence of drug), where the absorbance values were the highest, normalization was applied, and the cellular proliferation was considered to be 100%. Thus, from the absorbance obtained in the wells with different drug concentrations, the percentage of viable cells was determined. Finally, the IC₅₀ values were calculated as those concentrations of drug yielding 50% cell survival.

• Viability/cytotoxicity assay. At each time point, numerical data were normalized according to the values obtained in control wells (untreated cells), which were considered to be 100% for cell viability and 0% for cell cytotoxicity.

• Scratch migration assay. In order to investigate the migratory capability of LECs, the surfaces of wounded areas at time 0, without migrated cells, were considered to have 0% of recovered areas. After different times of incubation, the surfaces occupied by migrated cells were determined by image analysis, and the percentages of recovered areas were calculated by normalizing them to their respective values at time 0.

• Caspase‐3/‐7 assay. With this approach, the luminescence measured in control wells (non‐treated cells) was the lowest. Data were normalized considering that the luminescence in control wells was 100.

• Western blotting analyses. For all the protein phosphorylation tests, the phosphorylated protein/total protein ratio was calculated. The maximum phosphorylation levels were always obtained in control conditions after stimulation with FBS or VEGF‐C, VEGF‐C156S or VEGF‐A, in the different proteins studied. The value of 100% was assigned to these ratios, and the phosphorylation levels in presence of the compound were normalized to their respective values in stimulated control conditions.

• Mouse corneal neovascularization and ear sponge models. The data depicted in the Supporting Information Fig. S5 show the percentage of area, length, branching and end‐point density inhibition in toluquinol‐treated corneas. For these measurements, normalizations were applied by considering that the growth in control corneas was 100%, and results are expressed as percentage of inhibition. For similar quantifications in the ear sponge samples, the vascular density measured in sponges soaked with VEGF‐C and in the absence of toluquinol was considered as 100%, and the normalized percentages of vascular inhibition upon toluquinol treatment are displayed in the Supporting Information Fig. S5.

Validity of animal species or model selection

In this work it was considered that the research on animals carried out was an excellent way of characterizing the effects of a drug on the lymphangiogenesis phenomenon. Zebrafishes and mice were used for the following reasons:

• Zebrafish: although human pathologies have mostly been modelled using higher mammalian systems like mice, the lower vertebrate zebrafish has a good reputation as a model system. The advantages of zebrafish over classical vertebrate models are numerous and include high genetic and organ system homology to humans, high fecundity, external fertilization, ease of genetic manipulation, and transparency through early adulthood that enables powerful imaging modalities. Additionally, several stable lines of transgenic zebrafish are available. In this study, the transgenic Fli:eGFPy1, which represents endothelial cells labelled with GFP, was used. This model is very useful because it allows the fast and easy evaluation of the effect of compounds on the blood and lymphatic vascular systems. We focused on the effect of toluquinol on the development of the thoracic duct, which is a lymphangiogenesis‐dependent process.

• Mouse: the use of mice in research has led to major advances in our ability to know and treat a number of serious human diseases. Some of the advantages of the research on mice are the similarities to the human genome, the availability of a unique battery of sophisticated molecular and genetic tools and the animal's small size. Taken together these facilitate large‐scale/high‐throughput studies and make it a cost‐efficient model, providing functional information about several diseases and allowing the testing of new drugs, among others. In this study, mice were used for the lymphatic ring, ear sponge and corneal neovascularization assays. The lymphatic ring assay is an ex vivo test that bridges the gap between in vitro and in vivo systems and allows the evaluation of compounds on different stages of lymphangiogenesis, such as proliferation, migration and tubulogenesis. The ear sponge assay is a new model that uses sponges that can be embedded with different factors (stimulators and/or inhibitors) and can be easily placed into mouse ears, where there are a large number of lymphatic vessels. This approach was used to analyse the ability of toluquinol to inhibit VEGF‐C‐stimulated lymphangiogenesis. The corneal neovascularization induced by thermal cauterization is an in vivo model that allowed us to study the formation of new lymphatic and blood vessels in the avascular cornea. Thus, it is also possible to evaluate the interference of drugs with this vascular network. The corneal test was applied to investigate the ability of toluquinol to suppress lymph/angiogenesis under inflammatory conditions.

Animals

• Transgenic Fli1:eGFPy1 adult zebrafishes (Danio rerio), aged 12–14 months, were from the International Resource Centre (ZIRC, Eugene, Oregon) and were maintained in the GIGA‐Zebrafish Facility (University of Liège, Belgium). The average body weight for the adult animals (used to obtain embryos) was ~3 g. Male and female Fli1:eGFPy1 embryos at 24 h post‐fertilization (hpf) were incubated in the presence of the test compound, and were photographed at 5 days post‐fertilization (dpf). The average body weight for the embryo at 5 dpf was ~0.15 g.

• Female C57BL/6 mice aged 8 weeks and 3 months were purchased from Janvier (Saint Berthevin, France). The 8‐week‐old C57BL/6 mice, with an average body weight of 16.5 g, were used for the ear sponge and corneal neovascularization assays. The 3‐month‐old mice, with an average body weight of 25 g, were used to perform the ex vivo thoracic duct test.

Housing and husbandry

Transgenic zebrafishes, as well as mice used for the animal procedures, were maintained in the GIGA‐Cancer Zebrafish and GIGA‐Cancer Mouse Facilities, respectively, where there was an enriched environment with adequate temperature, humidity, food and so on. The adult zebrafish were housed using a rack and tank system. They were maintained in 3 L polycarbonate aquarium tanks (20 zebrafish per tank) with a recirculating water system; the water having optimal physical, chemical and biological characteristics. They were kept on a 14/10 h light/dark cycle, at a temperature of 27–28°C, and ZM‐Fish Food with artemia nauplii were supplied twice per day. The transgenic eggs were obtained by natural mating when males and females were transferred to a 1.5 L breeding tank, with a ratio of two males to two females. Embryos were incubated at 28.5°C.

Mice were housed in polycarbonate cages of 18 cm × 35 cm (five mice per cage; according to the Royal Decree of 29/05/2013 from the Belgian Animal Welfare Legislation) with floors covered with sawdust bedding. They were maintained under a 12/12 h light/dark cycle with controlled room temperature (~26°C) and humidity (35–75%) and were allowed ad libitum access to a diet of standard laboratory chow and water. Caging equipment was sterilized, and the food irradiated and water filtered.

Interpretation

Experiments involving animals were conducted with rigour, with the implementation of the 3Rs principles, as defined by the National Centre for the Replacement, Refinement and Reduction of Animals in Research (NC3R). Efforts were made to reduce the number of animals used and minimize animal suffering. Furthermore, animals were anaesthetised when it was likely they could be subjected to pain, and they were killed by a method that ensured the least effect on their welfare.

Ethical statement

The animal procedures considered in this project were performed in strict compliance with the European Communities Council Directive 2010/63/EU and the Belgium legislation for the animal experimentation. The Local Animal Ethics Committee at the University of Liège approved the ethical and legal aspect of our experimental protocol before the start of the research (number of ethical approval: 13/1522). Researchers working with animals received a specific training to reach Category C from the Federation of European Laboratory Animal Science Associations (FELASA) before animal manipulation. All studies involving animals or animal tissue are reported in accordance with the ARRIVE guidelines (Kilkenny et al., 2010; McGrath & Lilley, 2015), and the recommendations of the Basel Declaration Society, who are seeking to internationalize good practice (McGrath et al., 2015).

Experimental procedures

In vitro experiments

Cell growth and cell viability/cytotoxicity assays

The MTT (Sigma‐Aldrich) dye reduction assay was applied to LECs (3 × 10³ cells per well) in 96‐well microplates (García‐Caballero et al., 2013). For determining cell growth, the IC₅₀ values were calculated as concentrations of toluquinol yielding 50% cell survival, taking the values obtained for control as 100%. For viability/cytotoxicity tests, LECs were plated in 96‐well plates (15 000 cells per well) and treated with the highest toluquinol concentration (5 μM) applied in in vitro experiments, for different incubation times. Then, two differential fluorogenic protease biomarkers, GF‐AFC and bis‐AAF‐R110, provided in the ApoTox‐GloTM Triplex Assay kit (Promega Corporation, Madison, WI, USA) were simultaneously added in a single nonlytic reagent. The fluorescence was recorded with a Victor X3 Multilabel Plate Reader (PerkinElmer, Waltham, MA, USA). All determinations were carried out in quadruplicate, and five independent experiments were performed.

LEC migration assay

For the scratch migration assay (Liang et al., 2007), confluent monolayers were wounded with pipette tips and supplied with EGM‐2 medium containing 2% FBS, mitomycin C (0.1 μg·mL⁻¹; Sigma‐Aldrich) and different toluquinol concentrations. Wounded areas were photographed at different times of incubation using a phase contrast microscope (Axiovert 25; Zeiss Microscopy, Zaventem, Belgium). The percentage of recovered area at 24 and 48 h was determined by image analysis, normalizing them with their respective values at time 0. The NIH Image 1.6 software was used for quantification, and five independent experiments were evaluated.

Tubulogenesis assay

LECs (5 × 10⁵ per well) were embedded in a 1:1 mix of collagen R solution (Serva, Heidelberg, Germany) and HBSS (Gibco, New York, NY, USA) and cultured in 2% FBS‐supplemented EGM‐2 medium with or without toluquinol. Tube formation was monitored in the next 24 h. Pictures were photographed with an inverted and phase contrast microscope. Results are expressed as the area density of tubes, defined as the area occupied by tubes divided by the total area of the field studied. Five independent assays were carried out, and the MatLab R2013a4 (8.1.0.604) image analysis toolbox software was used for tube quantification.

Spheroid assay

In order to generate multicellular microspheres, 1.5 × 10³ LECs per well were precultured in EBM‐2 medium containing 1% FBS and 0.24% high‐viscosity methyl cellulose (Sigma‐Aldrich) for 24 h. Spheroids were then embedded in collagen gels containing or not toluquinol and cultured in 2% FBS‐supplemented EGM‐2 medium with or without the drug. After 24 h, photos of spheroids were taken using a phase contrast microscope, and a computer‐assisted quantification was used as described in Blacher et al., 2014. In some assays, cells were stained for 30 min, with green or orange CellTracker Dye (Life Technologies, Thermo Fisher Scientific, Belgium). Green and orange stained cells were mixed in a 1:1 proportion to generate spheroids as described above. Cell sprouting was analysed with an epifluorescence microscope (Nikon Eclipse Ti) after 12 h of incubation. For statistical analysis, at least eight spheroids per group were analysed in each individual experiment, and five independent experiments were performed.

Apoptosis and cell cycle experiments

After treatment with the indicated toluquinol concentrations for 14 h, apoptosis assays were carried out by staining of nuclei with Hoechst, analysis of cell cycle by flow cytometry and trough the evaluation of caspase‐3/‐7 activity. To analyse nuclear morphological changes, cells were treated, washed with PBS, fixed with formalin solution (Sigma‐Aldrich, Madrid, Spain) and stained with Hoechst 33258 (1 μg·mL⁻¹; Sigma‐Aldrich) in PBS. Subsequently, cells were mounted on slides using DAKO Cytomation Fluorescent Mounting Medium (Dako, Denmark) and observed under a fluorescence microscope (TCS‐NT; Leica, Heidelberg, Germany). For cell cycle analysis, LECs were washed with PBS, fixed with ice‐cold 70% ethanol for 1 h on ice, centrifuged and suspended in propidium iodide staining solution (40 μg·mL⁻¹ propidium iodide and 0.1 mg·mL⁻¹ RNase‐A in PBS; Sigma‐Aldrich, Madrid, Spain). After 1 h of incubation with shaking while protected from light, percentages of subG1, G1 and S/G2/M populations were determined using a MoFlo Dakocytomation cytometer (Dako) and the Summit 4.3 software. For the determination of caspase‐3/‐7 activity, LECs were plated in 96‐well plates (15 000 cells per well) and treated with or without different toluquinol concentrations for different times of incubation. Then, Caspase‐Glo® 3/7 reagent, provided in the ApoTox‐GloTM Triplex Assay Kit, was added to the wells according to the manufacturer's instructions, and the luminescence was recorded at 30 min with a PerkinElmer's VICTORX3 Multilabel Plate Reader. Five independent experiments were analysed for each of the three different approaches.

Cells and drug

Human adult dermal microvascular LECs were purchased from Lonza (HMVEC‐dLy.Ad; Braine‐l'Alleud, Belgium) and cultured in complete endothelial growth microvascular (EGM2‐MV) medium (Lonza, Walkersville, MD, USA) composed of EBM‐2 and singleQuotes (Detry et al., 2013). Toluquinol, provided by Biomar Institute S.A (León, Spain), was dissolved in DMSO at a concentration of 80 mM and stored in aliquots at −20°C until use. DMSO was used in controls at the same proportion used in experimental conditions.

Ex vivo experiments

Three‐dimensional mouse lymphatic ring cultures

Lymphatic ring cultures were performed as previously described (Bruyère et al., 2008). Briefly, 3‐month‐old female C57BL/6 mice were killed by decapitation after the animals had been made unconscious in a saturated‐isoflurane bell‐jar, the mouse skin was sterilized with ethanol, 70%, denatured, and the thoracic duct was dissected with microdissecting forceps and iridectomy scissors sterilized by dry heat. After removal of fibroadipose tissue, thoracic ducts were cut into 1‐mm‐long fragments, and embedded in a collagen gel. All the above mentioned steps were carried out in a horizontal laminar flow hood (Cleanair; VWR, Belgium) to ensure sterile conditions. Explants were cultured in MCDB131 medium supplemented with 2% Ultroser (Pall Life Sciences, Port Washington, New York, USA), 1% l‐glutamine, 25 mM NaHCO3, 100 U·mL⁻¹ penicillin and 100 μg·mL⁻¹ streptomycin and in the presence of toluquinol or vehicle (0.01% DMSO). The cultures were kept at 37°C in 5% O₂, 5% CO₂ and 90% N₂ for 7 days. For whole‐mounted immunostainings, at the end of the incubation, lymphatic rings were harvested from the agar cylinder, washed for 1 h in PBS, fixed for 30 min with 80% methanol and kept at 4°C in 70% ethanol until used. Then, gels were washed three times with PBS and blocked in 1.5% BSA‐3% Gloria milk for 1 h at room temperature. The gels were incubated overnight with the appropriate primary antibody, rabbit antilymphatic vessel endothelial hyaluronan receptor‐1 (LYVE‐1)/Biotin (1/500; ReliaTech GmbH, Germany) or mouse anti‐Thy1.1 (1/100; Chemicon, Billerica, MA, USA). After washing with PBS, gels were incubated with streptavidin/Alexa Fluor488 (1/100; Invitrogen, ThermoFisher, Belgium) or an Alexa Fluor 555‐coupled goat anti‐mouse secondary antibody (1/40; Invitrogen, ThermoFisher), for 2 h at room temperature. After three washes with PBS, gels were mounted on a microscope slide with Vectashield‐DAPI mounting medium (Vector Laboratories, Burlingame, CA, USA) and visualized with fluorescence and confocal microscopes (FSX100 Olympus, Hamburg, Germany; Confocal Leica TCS SP5, Leica Microsystems, Wetzl, Germany).

In order to verify the lack of toxicity of toluquinol concentrations used in the lymphatic explants, the LIVE/DEAD® Viability/Cytotoxicity Kit (Molecular Probes, Invitrogen, Carlsbad, CA, USA) was applied at the end of the assay, according to the manufacturer's instructions.

Computerized quantifications were performed on binary images, and a grid of concentric rings was generated by successive increments at fixed intervals of thoracic duct boundary (Bruyère et al., 2008). The number of microvessel–grid intersections was counted and plotted versus the distance to the ring to determine microvessel distribution. For the quantifications, 10 rings per group were used.

In vivo experiments

Thoracic duct formation in zebrafish

Transgenic Fli1:eGFPy1 zebrafish embryos at 24 hpf were manually dechorionated with sterile forceps in a horizontal laminar flow workstation (Airstream, Esco Technologies, USA). Then, they were incubated in zebrafish water with the indicated concentrations of the tested compound at 28.5°C for 4 days. DMSO (0.02%) was used as both carrier of drug and control vehicle. To visualize the zebrafish thoracic duct after drug treatment, embryos were anaesthetized with tricaine methanesulfonate 0.02% for 5 min (Sigma‐Aldrich, Diegem, Belgium), mounted in 3% methyl cellulose and observed under a fluorescence Eclipse 90i microscope (Nikon, Japan) with acquisition software NIS‐Elements 3.00 (Nikon). Embryos were killed by incubation with tricaine methanesulfonate 0.02% for 10 min. Quantification of defective thoracic duct formation at 5 dpf was expressed as percentages of embryos with severe (no vessel, 0), drastic (5–25% of normal thoracic duct length), moderate (25–90% of normal thoracic duct length) and no (100% of normal thoracic duct length) lymphatic defects, as previously described (Detry et al., 2012). A total of 50 embryos was evaluated for each experimental condition.

Mouse ear sponge assay

The mouse ear sponge assay was carried out following the protocol recently described (Lenoir et al., 2014). Sterile gelatin sponges (Gelfoam; Pfizer, Puurs, Belgium) were cut into small pieces, incubated with serum‐free DMEM containing or not recombinant VEGF‐C (1 μg·mL⁻¹) and different concentrations of toluquinol or DMSO (0.04%) and then embedded in interstitial type I collagen gel. After the mouse had been anaesthetized with i.p. injection of ketamine hydrochloride (100 mg·kg⁻¹ body weight) and xylazine (10 mg·kg⁻¹ body weight), ear skin was sterilized with ethanol, 70% denatured, and small incisions were made on the upper side of the ear of 8‐week‐old female C57BL/6 mice. Sponges were implanted between the two skin layers and every 2 days; serum‐free DMEM containing toluquinol (0.5, 2 or 4 nmol) or DMSO (controls) was injected in the apex of the ear. This manipulation was performed in a horizontal laminar flow hood Cleanair. It is worth mentioning that a pilot study was first conducted using different toluquinol concentrations (0.2, 0.5, 2, 4 and 6 nmol), in order to establish the appropriate concentration range demonstrating a dose–response for the inhibition. After 3 weeks, mice were killed by cervical dislocation followed by decapitation, and ears were excised, embedded and frozen in tissue optimal cutting temperature compound (VWR Chemicals, Fontenay‐sous‐Bois, France). For detection of lymphatic vasculature, sections were immunostained with polyclonal goat anti‐mouse LYVE‐1 (1/200; R&D, Abingdon, UK) and Alexa Fluor 488‐coupled rabbit anti‐goat (1/200; Molecular Probes, Gent, Belgium) antibodies. For blood vessel detection, a primary monoclonal rat anti‐mouse antibody against CD31 (1/200; BD Biosciences Pharmingen, San Jose, CA, USA) and a secondary Alexa Fluor 546‐coupled goat anti‐rat antibody (1/200; Molecular Probes) were used. Images were acquired with a NanoZoomer 2.OHT scanner (Hamamatsu, Mont‐Saint‐Guibert, Belgium), and the vessel area density and lymphatic normalized frequency were determined using image analysis toolbox of MatLab R2013a (8.1.0.604) software. Sections from the two ears of each mouse were used to obtain the results, providing a unique averaged value, and a total of five mice were used per group.

Mouse corneal neovascularization assay

The mouse corneal assay was performed as previously described (Detry et al., 2013). Briefly, the mice were anaesthetized with ketamine hydrochloride (100 mg·kg⁻¹ body weight) and xylazine (10 mg·kg⁻¹ body weight) injected i.p., the local anaesthetic (Unicaïne 0.4%; Thea Pharma, Wetteren, Belgium) was applied, and the central cornea was thermally cauterized using an ophthalmic cautery (Optemp II V; Alcon Surgical, Fort Worth, TX, USA). One day after cauterization, 300 μL of PBS or PBS containing toluquinol (75 nmol) was injected i.p. daily until day 9, or day 40 post‐cauterization in the case of vascular regression analyses. After the mice had been killed by cervical dislocation followed by decapitation, corneas were dissected with sterile microdissecting instruments, whole mounted, fixed in ethanol 70%, blocked with 3% BSA‐3% Gloria milk and immunostained as detailed for ear sponge assay. Corneas were visualized by using a fluorescent microscope Olympus FSX100 or a Leica TCS SP2 inverted confocal microscope. The whole cornea picture was reconstituted with the Microsoft Image Composite Editor software, and the lymphatic and blood vasculatures were quantified with the toolbox of MatLab R2013a (8.1.0.604) software. Corneas from the same animal were quantified and averaged. A total of 15 mice per group were evaluated for the statistical analysis.

Western blotting analyses

In order to detect VEGFR‐3 and VEGFR‐2 and phosphorylated VEGFR‐3 and VEGFR‐2, and downstream targets, such as p‐Akt, Akt, p‐ERK and ERK, subconfluent LEC cultures were incubated in serum‐free EBM‐2 medium for 24 h. Then, toluquinol or vehicle (0.02% DMSO) was added to the cultures for 2 h and the samples stimulated for 30 min with either recombinant human VEGF‐C (400 ng·mL⁻¹; R&D Systems, Abingdon, UK), recombinant human VEGF‐C156S (500 ng·mL⁻¹; R&D Systems) or recombinant human VEGF‐A (100 ng·mL⁻¹; R&D Systems). After stimulation, cells were rinsed with ice‐cold PBS and protein lysates were obtained by scrapping the cells in RIPA buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 1% Triton X‐100, 1% sodium deoxycholate, NP40 1% and SDS 0.1%) containing phosphatase and protease inhibitors (Roche, Mannheim, Germany). Cell extracts were centrifuged at 17200x g for 15 min at 4°C, evaluated for protein concentration and stored at −80°C until the moment of analysis. Afterwards, samples were denatured for 5 min at 95°C and subjected to 10% SDS‐PAGE electrophoresis. After electrotransfer on PVDF membranes, phosphorylated and total proteins were detected by 4°C overnight incubation with the indicated antibodies followed by 1 h incubation at room temperature with horseradish peroxidase‐coupled secondary antibody (Cell Signalling, San Diego, CA, USA). The immunoreactive bands were detected using a chemiluminescence system (ECL Western Blotting Substrate, Pierce, Rockford, IL, USA) with an imaging system (LAS4000 imager; Fujifilm, Tokyo, Japan). The following antibodies were used: rabbit monoclonal p‐VEGFR‐3 (1/1000; Cell Application, San Diego, CA, USA), rabbit monoclonal p‐Akt, Akt, p‐ERK1/2, ERK1/2, p‐VEGFR‐2 and VEGFR‐2 (1/1000; Cell Signaling) and mouse monoclonal VEGFR‐3 (1/1000; Millipore, Carrigtwohill, Ireland). Membranes were incubated with an anti‐GAPDH primary antibody (1/2000; Millipore) to ensure equal loading. At least, five western blots were carried out, and the phosphorylated protein/total protein ratios were calculated from five independent tests.

Statistical analysis

Results are given as mean ± SEM. Statistical analysis was carried out with the GraphPad Prism 5.0 programme to compare the differences between control conditions in the absence of compound and the different toluquinol concentrations. Mann–Whitney–Wilcoxon tests were used to decide if the differences among means were statistically significant and P < 0.05 constituted the threshold for the significance level. Significant differences are indicated in the figures as ^* P < 0.05. Data and statistical analysis comply with the recommendations on experimental design and analysis in pharmacology (Curtis et al., 2015).

Results

Toluquinol inhibits LEC growth and decreases LEC viability

We first examined the effect of toluquinol (Figure 1A) on LEC proliferation using the MTT assay. Toluquinol inhibited the growth of actively growing LECs in a dose‐dependent manner with an IC₅₀ value of 6.2 ± 2 μM (Figure 1B). Moreover, to evaluate the cell viability/cytotoxicity in the presence of toluquinol 5 μM (highest toluquinol concentration used in vitro), the ApoTox‐Glo™ Triplex Assay kit was used. Toluquinol decreased the cell viability by 34% after 14 h of treatment, but had low cytotoxicity (<13%) after 48 and 72 h of drug exposure (Figure 1C).

Figure 1.

Toluquinol inhibits LEC proliferation and affects LEC viability. (A) Chemical structure of toluquinol. (B) Representative curve with the dose‐dependent effect of toluquinol on the in vitro growth of LECs. Cell proliferation is represented as a percentage of untreated cells. Each point represents the mean of quadruplicates; SD values were typically lower than 10% of the mean values and are omitted for clarity. The IC₅₀ value was calculated from dose–response curves as the concentration of compound yielding 50% of control cell survival. It is expressed as means ± SEM of five independent experiments. (C) Percentage of cell viability and cell toxicity after toluquinol treatment for 14, 24, 48 and 72 h. Results are expressed as the percentage of control (untreated cells) viability and toxicity (mean ± SEM of five independent experiments). Controls (100% of viability and 0% cytotoxicity) are not presented for each time point for clarity. Mann–Whitney–Wilcoxon test was used to decide if the differences among control (untreated cells) and toluquinol‐treated cells were statistically significant. ^* P < 0.05 versus control.

Toluquinol inhibits LEC migration

A scratch assay was carried out to examine the effect of toluquinol on LEC migration (Figure 2). To determine whether a putative inhibitory effect could be related to an effect on cell proliferation, mitomycin C was always added to the culture medium after wounding, and pictures were taken at time 0 and after 24 and 48 h (Figure 2A). In control conditions, 50% of the scratch area was covered by migrating LECs after 24 h. After treatment with toluquinol 1, 2.5 and 5 μM, 45.4, 9.1 and 5.5% of the wounded areas were covered after 24 h respectively (Figure 2B). This marked inhibition of migration was also observed after 48 h of scratching with toluquinol at 2.5 and 5 μM (Figure 2B). Thus, toluquinol exerted a dose‐dependent inhibitory effect on LEC migration after 24 and 48 h of scratching. Additional experiments revealed that the effect of toluquinol is reversible, because LECs pretreated with 5 μM toluquinol for 24 h and incubated in EGM‐2 medium containing 2% FBS (in the absence of drug) were able to migrate and close the wounded areas in a similar way to control cells after 48 h (Supporting Information Fig. S1).

Figure 2.

Toluquinol inhibits LEC migration. Confluent LEC monolayers were wounded, and fresh culture medium was added in either the absence or presence of the indicated concentrations of toluquinol. Photographs were taken at the beginning of the assay and after 24 and 48 h of incubation. All assays were performed in the presence of mitomycin C to avoid a proliferative effect on LEC. (A) Representative pictures of cell migration. Black drawings in pictures delineate the initial (time 0) wound edges (bar = 100 μm). (B) Quantification of LEC migration. Results are expressed as the percentage of the initial wounded area recovered by endothelial cells after 24 or 48 h (mean ± SEM of five independent experiments). Mann–Whitney–Wilcoxon test was the test used to analyse if the differences among control (untreated cells) and toluquinol‐treated cells were statistically significant. ^* P < 0.05 versus control.

Toluquinol suppresses tube formation and LEC sprouting in a collagen matrix

In an initial 3D model, LECs were seeded between two collagen gels in order to generate tube‐like structures after 24 h of culture. In the presence of toluquinol, the tubular network was disorganized (Figure 3A). Toluquinol inhibited lymphatic tubulogenesis in a dose‐dependent manner with an optimal effect reached at 5 μM (Figure 3B). As a second 3D model, LEC spheroids were embedded in a collagen matrix to mimic the sprouting, migration and tube formation of lymphatics. Within 24 h, untreated spheroids sprouted from the initial microsphere core (Figure 4A). In sharp contrast, LECs from 5 μM toluquinol‐treated spheroids failed to spread out (Figure 4A). A dose‐dependent inhibition of cell sprouting was again observed, as assessed by the determination of various quantitative parameters (Blacher et al., 2014). Firstly, the convex envelope area, corresponding to the area occupied by the spheroid core and the migrated cells, was determined, and its value after 5 μM toluquinol treatment was significantly lower (0.09 mm²) than in the control condition (0.39 mm²) (Figure 4B). Secondly, an estimation of the number of migrated cells per area also revealed a maximal inhibition at 5 μM toluquinol. Finally, the quantification of cell distribution around the spheroid core further supported the marked inhibition of LEC migration induced by toluquinol when used at the highest concentration (Figure 4B). The maximal distance of LEC migration (Lmax) was 0.26 mm in the presence of 5 μM toluquinol and 0.4 mm in the other conditions. In order to check the reversibility of the inhibitory effect elicited by toluquinol, some mixed spheroids were generated with green and orange stained cells (1:1 proportion). After 12 h, in non‐treated mixed spheroids, green and orange stained LECs migrated to a similar extent (‘1’ in Figure 4C). In contrast, when green and orange cells were mixed and embedded in a collagen matrix containing 5 μM toluquinol, no sprouting was detected (‘2’ in Figure 4C). However, when spheroids were generated with green LEC pretreated with toluquinol (5 μM), both orange and green stained LECs sprouted from the spheroid (‘3’ in Figure 4C). These findings demonstrate the reversibility of these effects of toluquinol.

Figure 3.

Toluquinol suppresses LEC tubulogenesis. LECs were embedded in a collagen gel, maintained in culture for 24 h in order to form tubes and photographed under a phase contrast microscope. Toluquinol inhibited LEC tubulogenesis in a dose‐dependent manner at non‐toxic doses. (A) Representative pictures at different toluquinol concentrations (bar = 200 μm). (B) Quantification of the density of the tube area. Values are expressed as mean ± SEM of five independent assays, and Mann–Whitney–Wilcoxon test was used to evaluate statistically significant differences between control (untreated cells) and toluquinol‐treated cells. ^* P < 0.05 versus control.

Figure 4.

Toluquinol inhibits LEC sprouting from spheroids. LEC spheroids embedded in a collagen‐methyl cellulose gel extended protrusion sensing the environment and formed tube‐like structures. Spheroids were treated or not (control) with toluquinol. (A) Cell sprouting is inhibited by increasing concentrations of toluquinol (bar = 100 μm). (B) Graphs correspond to the quantification of (i) the convex envelope area (area of minimal convex polygon containing the spheroid core as well as the migrated cells) (left panel) and (ii) the area occupied by migrating cells (middle panel). The graph at right displays LEC density as a function of the distance to the spheroid centre. Values are expressed as mean ± SEM of five tests with at least eight spheroids quantified from each one. Mann–Whitney–Wilcoxon test was used to decide if the differences among control (untreated spheroids) and toluquinol‐treated spheroids were statistically significant. ^* P < 0.05 versus control. (C) Cells stained orange with CellTracker Orange CMRA Dye and cells stained green with CellTracker Green CMFDA Dye were mixed in a 1:1 proportion before generating spheroids. After 12 h, cell sprouted from spheroids (1: non‐treated cells). The treatment of mixed spheroids (green + orange cells) with toluquinol during the migration assay led to a blockade of cell migration (2: toluquinol‐treated cells). When LECs were treated with toluquinol (5 μM) for 24 h (pretreated LECs), washed with PBS (toluquinol elimination), stained in green, mixed with orange untreated cells and treated without toluquinol during the migration assay (3: toluquinol‐pretreated cells), migration occurred in a similar way as in control conditions, revealing the reversibility of the inhibition induced by toluquinol (bar = 100 and 50 μm on higher magnification).

Toluquinol induces LEC apoptosis

To further determine whether toluquinol induces LEC apoptosis, the morphology of chromatin was analysed in treated and untreated cells stained with Hoechst 33258 (Figure 5A). Toluquinol promoted chromatin condensation at 2.5 and 5 μM after 14 h of treatment in 18 and 48% of cells respectively (Figure 5B). Less than 4% of apoptotic nuclei was observed with toluquinol used at 1 μM. A dose‐dependent induction of apoptosis was thus observed with the toluquinol doses used. Cell cycle analyses by flow cytometry were next conducted in control and toluquinol‐treated LECs after propidium iodide staining. The apoptotic subG1 cell subpopulation was significantly increased upon toluquinol treatment for 14 h (Figure 5C), leading to a threefold increase in the % of apoptotic LEC with 5 μM toluquinol compared with the control condition (Figure 5D). In addition, the ‘effector’ caspase‐3/‐7 activity, which plays a key role in the induction of apoptosis, was determined in LEC cultures. LEC caspases were activated upon treatment with 2.5 and 5 μM toluquinol for 14 h (Figure 5E). However, this initial 1.5‐fold enhancement of caspase activation observed upon 5 μM toluquinol treatment was not detected after 48 and 72 h of incubation (Figure 5F).

Figure 5.

Toluquinol induces LEC apoptosis after 14 h of treatment. LECs were cultured in the presence of different toluquinol concentrations for 14 h, and then apoptosis was evaluated by Hoechst staining, flow cytometry and caspase‐3/‐7 activity. (A) Representative pictures showing the effect of toluquinol on nuclei morphology after Hoechst staining. The white circles delineate the chromatin condensation of apoptotic nuclei (bar = 100 μm). (B) Percentage of cells showing chromatin condensation per field. Values are expressed as mean ± SEM of the counts evaluated in 10 fields from five independent experiments (chromatin‐condensed cells or apoptotic cells were counted by fluorescence microscopy; total cells were counted by bright field microscopy). (C) Representative histograms showing the effect of toluquinol on LEC cycle distribution. After incubation with toluquinol, cells were stained with propidium iodide, and percentages of subG1, G1 and S/G2/M subpopulations were determined using a MoFlo DakoCytomation cytometer. (D) The graph corresponds to the distribution of cell subpopulation percentages expressed as means ± SEM of five independent assays. (E) Effect of toluquinol treatment for 14 h on LEC caspase‐3/‐7 activity. Results are expressed as mean ± SEM of five independent assays. (F) Effect of toluquinol 5 μM treatment on LEC caspase‐3/‐7 activity over time. Results are expressed as mean ± SEM of five independent assays. Controls (caspase‐3/‐7 activity = 100) are omitted for clarity. Statistically significant differences between control (untreated cells) and toluquinol‐treated cells were determined with the Mann–Whitney–Wilcoxon test. ^* P < 0.05 versus control.

Toluquinol suppresses VEGFR‐3 phosphorylation, and MAPK/ERK1/2 and PI3K/Akt downstream activation in LECs

A mechanistic exploration of the inhibitory effect of toluquinol on LECs was next conducted by focusing on VEGFR‐2 and VEGFR‐3, the two main receptors involved in lymphangiogenesis, and their downstream signalling pathways PI3K/Akt and MAPK/ERK1/2 (Olsson et al., 2006; Bahram and Claesson‐Welsh 2010; Deng et al., 2015). We used VEGF‐C, VEGF‐C156S and VEGF‐A in order to examine the effect of toluquinol on VEGFR‐3‐ versus VEGFR‐2‐induced LEC activation. As expected, stimulation with VEGF‐C resulted in VEGFR‐3 activation, as well as Akt and ERK1/2 phosphorylations, as assessed by western blotting (Figure 6A). A pretreatment with toluquinol 5 μM decreased VEGFR‐3, Akt and ERK1/2 phosphorylation levels. Similarly, VEGFR‐3, Akt and ERK were activated after stimulation with VEGF‐C156S, a mutant VEGF‐C form, which binds VEGFR‐3, but not VEGFR‐2. Interestingly, under this experimental condition, a stronger effect of toluquinol on p‐VEGFR‐3, p‐Akt and p‐ERK inhibition was seen at 2.5 and 5 μM (Figure 6B). However, when VEGF‐A was used as the more potent VEGFR‐2 stimulator, toluquinol failed to interfere with VEGFR‐2 activation, and no reduction in VEGFR‐2 phosphorylation levels was observed in toluquinol‐treated LECs. It is worth noting that p‐Akt and p‐ERK levels were diminished in toluquinol‐treated, VEGF‐A‐stimulated LECs, suggesting a modulation of downstream effectors (Figure 6C). The results are expressed as phosphorylated protein/total protein ratios and are displayed in the Supporting Information Fig. S2, and complete western blot membranes are presented in the Supporting Information Fig. S3.

Figure 6.

Toluquinol suppresses VEGFR‐3 phosphorylation and downstream signalling targets. LECs were cultured in serum‐depleted conditions for 24 h, and incubated or not with different toluquinol concentrations for 2 h and then stimulated for 30 min with VEGF‐C (A), VEGF‐C156S (B) or VEGF‐A (C). Cell lysates were collected, and western blot analyses were performed. (A) Representative western blots showing the effects of toluquinol treatment on the phosphorylated VEGFR‐3, Akt and ERK1/2 in LECs stimulated with VEGF‐C (400 ng·mL⁻¹). (B) Illustrative western blots showing the impact of toluquinol treatment on VEGFR‐3, Akt and ERK1/2 phosphorylations in LECs stimulated with VEGF‐C156S (500 ng·mL⁻¹). (C) Representative western blots showing the interference of toluquinol treatment on VEGFR‐2, Akt and ERK1/2 phosphorylations in LECs stimulated with VEGF‐A (100 ng·mL⁻¹).

Toluquinol decreases the ex vivo outgrowing from mouse thoracic explants

To validate the antilymphangiogenic activity of toluquinol found with the in vitro assays, next, we tested it in an ex vivo assay. Explants from mouse thoracic ducts were cultured under hypoxic conditions to analyse the ex vivo lymphatic outgrowth. A noticeable outgrowth of LECs from control rings (with DMSO) was observed after 7 days of incubation (Figure 7A). In sharp contrast, explants exposed to 2.5 and 5 μM toluquinol exhibited a significantly reduced outgrowth, and only a limited LEC migration was observed in the presence of toluquinol at 7.5 μM (Figure 7A). It is worth noting that the influence of toluquinol in the lymphatic ring assay was examined in the presence of 2% Ultroser G to ensure an optimal lymphangiogenic reaction in all the explants. For computerized quantification, a grid of concentric circles was used, and the number of microvessel intersections with this grid was estimated at different distances to the lymphatic ring. A dose‐dependent reduction in cell migration was observed with toluquinol (Figure 7B). The number of microvessel intersections at 0.25 mm from the lymphatic ring reached the value of 280 for control rings and 170, 98 and 43 for 2.5, 5 and 7.5 μM toluquinol respectively (Figure 7B). The absence of toluquinol toxicity was checked at the end of this assay with the highest concentration used (Supporting Information Fig. S4A). Additionally, the LEC origin of sprouting cells was assessed by immunostaining of whole‐mounted explants, revealing the LYVE‐1 positivity of outgrowing cells (Supporting Information Fig. S4B). To verify the absence of outgrowing fibroblasts, an anti‐Thy1.1 antibody was used, which showed the absence of migrating fibroblasts (Supporting Information Fig. S4C–E). Only one cell appeared positive (stained in red), but it was located in the ring (Supporting Information Fig. S4C).

Figure 7.

Toluquinol inhibits lymphatic outgrowth in mouse lymphatic ring assay. Mouse lymphatic duct explants embedded in type I collagen gel were cultured in the absence (control) or presence of different doses of toluquinol for 7 days under hypoxic conditions. (A) Representative micrographs of lymphatic rings (bar = 500 μm). (B) For quantification, binarized images from 10 rings per condition were subjected to a grid corresponding to successive increments at fixed intervals of thoracic duct boundary, and the number of microvessel grid intersections (Ni) at day 7 of incubation was calculated. Values are expressed as means ± SEM of 10 different rings. The graph (at right) corresponds to LEC density at a distance (d) = 0.25 mm to the ring border. Mann–Whitney–Wilcoxon test was used to determine if the differences among control (untreated rings) and rings incubated with toluquinol were statistically significant. ^* P < 0.05 versus control.

Toluquinol inhibits zebrafish lymphangiogenesis in vivo

In zebrafish, the thoracic duct generated by lymphangiogenesis is located between the dorsal aorta and the posterior cardinal vein (Figure 8A). Its development starts around 30 to 50 hpf and is complete at 3 to 6 dpf. Embryos from a transgenic fli1:eGFPy1 zebrafish line (which drives the GFP expression in the endothelium) were treated with different amounts of toluquinol added to the zebrafish water. With this approach, the direct addition of the drug to the water justified the use of a different range of concentrations from that used in vitro and in vivo, but similar to doses applied previously (García‐Caballero et al., 2013) for the antiangiogenic effect of toluquinol in zebrafish. Thus, after 4 days of incubation with toluquinol, we analysed the morphology of the fish with a special focus on the thoracic duct. Embryos subjected to toluquinol treatment displayed a normal morphology with normal body size and trunk circulation. No obvious tissue malformation, retarded development or other toxic signs were detected. Toluquinol concentration ≥15 μM induced significant thoracic duct defects in zebrafish (Figure 8B). The % of embryos with severe (no vessel, lymphatics abolished), drastic (5–25% of normal thoracic duct length), moderate (25–90% of normal thoracic duct length) or no (100% of normal thoracic duct length) lymphatic defects were determined (Figure 8C). At 5 dpf, the thoracic duct failed to form in 20% of 20 μM toluquinol‐treated zebrafishes, exceeded its normal length by 5–25% in 22% of embryos, and only 34% of embryos had normal length thoracic duct. In the presence of toluquinol at doses lower than 15 μM, the antilymphangiogenic effect was less apparent (Figure 8C). Table 1 summarizes the data and the number of embryos of the abovementioned phenotypes.

Figure 8.

Toluquinol blocks thoracic duct (TD) development in zebrafish model. Transgenic Fli1:eGFPy1 zebrafish embryos were incubated in zebrafish water with the indicated concentrations of toluquinol at 28.5°C for 4 days, and then, thoracic duct length was analysed in the anaesthetized embryos. (A) Schematic drawing of a 5 dpf zebrafish embryo with the main anatomical features. Abbreviations used on this drawing denote as follows: dorsal longitudinal anastomotic vessels (DLAV), thoracic duct (TD), dorsal artery (DA), posterior cardinal vein (PCV), parachordal line (PAC), intersegmental vessels (ISV), arterial intersegmental vessels (aISV), venous intersegmental vessels (vISV) and intersegmental lymphatic vessels (ISLV). (B) Representative pictures of untreated and treated zebrafishes are shown. Zebrafish thoracic duct is indicated with white arrowheads on lower magnification pictures, and it is included in white rectangles on higher magnification pictures. Abbreviations indicate the main zebrafish embryo anatomical structures (bar = 100 and 70 μm on higher magnification). (C) Quantification of the defective thoracic duct formation at 5 dpf determined by the percentages of embryos with severe, drastic, moderate and no lymphatic defects, expressed as % of thoracic duct length. A total of 50 embryos were analysed in each experimental condition.

Table 1.

Inhibition of zebrafish TD formation by toluquinol

Phenotype	Control, %	10 μM toluquinol, %	15 μM toluquinol, %	20 μM toluquinol, %
No vessels	4 (2 embryos)	4 (2 embryos)	6 (3 embryos)	20 (10 embryos)
5–25% of TD	6 (3 embryos)	10 (5 embryos)	20 (10 embryos)	22 (11 embryos)
25–90% of TD	8 (4 embryos)	16 (8 embryos)	18 (9 embryos)	24 (12 embryos)
100% of TD	82 (41 embryos)	70 (35 embryos)	56 (28 embryos)	34 (17 embryos)

TD formation in zebrafish embryos was analysed to test the in vivo antilymphangiogenic effect of toluquinol. The defective TD formation at 5 dpf was determined by the percentages of embryos with severe (no vessel, 0), drastic (5–25% of normal TD length), moderate (25–90% of normal TD length) and no (100% of normal TD length) lymphatic defects. A total of 50 embryos was analysed in each experimental condition.

TD, thoracic duct.

Toluquinol affects in vivo mouse VEGF‐C‐induced lymphangiogenesis

Next, we extended our analyses by using the mouse ear sponge assay, allowing the concomitant study of lymphangiogenesis and angiogenesis. In this murine model, fragments of gelatin sponges were embedded in collagen containing either DMEM (negative control) or VEGF‐C (positive control), or a mix of VEGF‐C and different amounts of toluquinol. These sponges were implanted into mice ears. Every 2 days, DMEM with the vehicle or DMEM supplemented with toluquinol was injected in the apex of the ear for 21 days. The blood and lymphatic networks were analysed by CD31 and LYVE‐1 immunostainings respectively. A huge amount of blood vessels and only few lymphatic vessels were observed in sponges soaked with DMEM (Figure 9A). As expected, sponges soaked with VEGF‐C (1 μg·mL⁻¹) exhibited a fourfold increase in the infiltrating lymphatic vessels, with slight effects on the blood vasculature (Figure 9A, B). The in vivo administration of toluquinol to mice induced a clear dose‐dependent reduction in the area occupied by lymphatic vessels (Figure 9A). Indeed, twofold and fourfold decreases in the lymphatic area were detected after administration of toluquinol at 2 and 4 nmol respectively (Figure 9B). Upon drug administration, most lymphatic vessels remained located close to the sponge border, and the Lmax in sponges embedded with 4 nmol of drug did not exceed 1.5 mm from the sponge edge (Figure 9B). In contrast, in VEGF‐C‐embedded sponges without drug, Lmax was more than 2 mm (Figure 9B). Notably, no significant angioinhibitory effect of toluquinol was seen at the doses tested, as assessed by CD31 stainings (Figure 9A, B), and although toluquinol inhibited lymphangiogenesis by up to 80%, the angiogenic response was only reduced by about 30% (Supporting Information Fig. S5A).

Figure 9.

Toluquinol impairs the in vivo VEGF‐C‐stimulated lymphangiogenesis in mouse ear collagen sponges. Gelatin sponges were soaked with either DMEM containing the vehicle (DMSO) as negative control or VEGF‐C (1 μg·mL⁻¹) as a positive control or VEGF‐C and different concentrations of toluquinol, (indicated with ‘T’ in the graphs). Sponges were implanted between the two skin's layers of mice ears for 3 weeks. (A) Lymphatic and blood vasculatures were examined by LYVE‐1 (green) and CD31 (red) immunostainings respectively. DAPI staining was used to detect cell nuclei (blue) (bars = 1500 and 500 μm on higher magnification). (B) The graphs represent the computerized quantification of the densities of lymphatic (left panel) or blood (middle panel) vessels, defined as the area occupied by vessels divided by the area of sponge section. Data are expressed as means ± SEM of five mice. The right graph corresponds to vessel distribution from the sponge edge to its centre. The arrows indicated the maximal distance of LEC migration (Lmax). Mann–Whitney–Wilcoxon test was used to evaluate statistically significant differences between sponges soaked with VEGF‐C + toluquinol and control sponges with VEGF‐C. ^* P < 0.05 versus VEGF‐C‐stimulated sponges.

Toluquinol reduces the corneal neovascularization in mice

Corneal neovascularization was induced in mice daily injected i.p. with 300 μL of three different toluquinol doses. Based on this first test and according to the 3Rs principles, we used 0.25 mM toluquinol (75 nmol toluquinol) as the optimal dose, in order to repeat the study by increasing the group size.

At day 9 post‐injury, in cauterized corneas (Figure 10A–D), lymphatic (stained in green) and blood (stained in red) vasculatures grew out of the limbus to the central part of the cornea. A clear reduction in both lymphatic and blood vascular networks was detected in corneas dissected from treated mice (Figure 10A–D). The computer‐assisted quantification revealed that toluquinol reduced the area, length, branching and end‐point densities of both lymphatic (Figure 10G) and blood (Figure 10H) vasculatures. Lmax was also significantly reduced (Lmax for lymphatic vessels =1.99 in controls vs. 1.5 in toluquinol‐administered mice; Lmax for blood vessels =1.65 in controls vs. 1.36 in toluquinol‐administered mice) (Figure 10I). Notably, toluquinol was more potent on lymphangiogenesis than on angiogenesis for all of the parameters quantified. Indeed, as shown in Supporting Information Fig. S5B, higher percentage of inhibition (except for the end‐point density) was found for corneal lymphatic vasculature as compared with that of blood.

Figure 10.

Toluquinol reduces corneal neovascularization in cauterized mouse corneas. Corneal lymph/angiogenesis was induced by thermal cauterization, and mice were administered 75 nmol toluquinol i.p. daily (treated mice) or PBS (control mice). Corneas were flat mounted at day 9 post‐injury and immunostained for detecting lymphatic vessels (LYVE‐1 positive, in green) and blood vessels (CD31 positive, in red). Whole‐mounted corneas were observed under a fluorescent (A–D) or a confocal (E–F) microscope. (A–D) Representative pictures of corneas from non‐treated (A, B) and treated (C, D) mice. Lymphatic vessels and blood vessels appear in green and in red respectively (bars = 1000 μm in A and C; 500 μm in B and C). (E–F) Representative pictures of filopodia‐like structures (white arrowheads) displayed by migrating LECs in corneas from control (E) and toluquinol‐treated mice (F) (bars = 10 μm). (G–H) Computer‐assisted quantification was based on the splitting of green and red channel to dissociate lymphatic from blood networks. The vessel area (area covered by neoformed vessels), total length (cumulative length of the vessels), branching (number of bifurcations) and end‐point (number of sprout tips) densities were calculated after image binarization. All results were divided by the total cornea area. Results are expressed as the mean ± SEM of 15 mice. (I) A grid was applied on each cornea picture to establish the distribution curves of capillaries around the limbal vessels, and it is presented as the number of intersections (Ni) versus the distance to the limbus. (J) Graphs represent the number (left) and length (right) of filopodia‐like structures in a total length of 25 μm at the end of the lymphatic vessel. For these quantifications, 10 pictures were quantified per mouse, and corneas from five mice were evaluated. Mann–Whitney–Wilcoxon test was the statistical test used to evaluate significant differences between control mice (PBS) and toluquinol‐injected mice. ^* P < 0.05 versus control mice.

We next visualized the corneas under a confocal microscope in order to analyse the migrating cells at the tip of the lymphatic buds, characterized by numerous cytoplasmic extensions or filopodias. Interestingly, the number and length of filopodias were higher in control corneas (~7 filopodia‐like extensions and with a length of ~8.2 μm) than in toluquinol‐treated corneas (~3 filopodia‐like extensions and with a length of ~4.7 μm) (Figure 10E, F, J).

Because toluquinol induced apoptosis in vitro after 14 h of treatment, we wondered whether this drug was able to provoke apoptosis and vessel regression in vivo. We thus determined blood and lymphatic vessel densities in mice treated or not with toluquinol (75 nmol) for 40 days post‐cauterization, and a clear reduction in both neovasculatures was detected in treated mice (Supporting Information Fig. S6A). As expected, the lymphatic and blood vascular densities were lower at day 40 than at day 9, in control and treated mice, and the regression of the blood vasculature was more pronounced than that of the lymphatic network (Supporting Information Fig. S6B). Nevertheless, no significant difference in vessel regression was seen between control and treated mice after 40 days post‐cauterization (Supporting Information Fig. S6B).

Discussion

Lymphatic vessels are critical for health and are involved in fluid drainage, lipid transport and inflammation. Because excessive lymphangiogenesis favours cancer cell metastasis, inflammation and graft (renal and corneal) rejection, inhibiting lymphangiogenesis has potential as a treatment for different pathological conditions (Alitalo, 2011). Here, we provide in vitro, ex vivo and in vivo evidence that toluquinol (isolated from the marine fungus Penicillium) is a potent inhibitor of lymphangiogenesis. Mechanistically, this antilymphangiogenic effect relies on the inhibition of VEGFR‐3 phosphorylation and the downstream signalling pathway. We previously reported that toluquinol can exert antiangiogenic effects (García‐Caballero et al., 2013). Indeed, we found that toluquinol inhibited in vitro tumour proliferation, endothelial capillary tube‐like formation and migration and in vivo angiogenesis by targeting Akt signalling pathway. We now assign new antilymphangiogenic activities to toluquinol, which are supported by the use of a panel of in vitro, ex vivo and in vivo models replicating the different steps of the complex lymphangiogenic process.

The present study sheds light on the antilymphangiogenic potential of toluquinol and the molecular mechanisms involved in this process. Toluquinol significantly decreased the proliferation and migration of LECs, and lymphatic tube formation in a dose‐dependent manner with concentrations ranging from 1 to 5 μM with an IC₅₀ value of 6.2 μM. It is worth mentioning that the effective dose of toluquinol in the in vitro assays is in agreement with the effective dose of toluquinol found to be an angiogenic inhibitor (García‐Caballero et al., 2013). Importantly, an apoptotic effect of toluquinol was clearly detected after 14 h of treatment, because LECs exhibited some typical features of apoptosis, including increased chromatin condensation, an increase in the percentage of cells with subdiploid DNA content (subG1 phase) and an enhancement of the effector caspase‐3/‐7 activity. However, 3 days after the addition of toluquinol, the caspase activity had returned to basal levels. Accordingly, a decrease in LEC viability was detected at an early time point, but not after 72 h of treatment. Therefore, the apoptotic effect of toluquinol is likely to contribute to the growth inhibitory effect and the antilymphangiogenic activity exhibited by this compound on LECs, but only in its initial antilymphangiogenic action. This observation is also in agreement with the reversibility of the effects of toluquinol demonstrated in the migration assays.

The potent impact of toluquinol on LEC migration was demonstrated in 2D and 3D models of migration, namely, the wound healing, the tubulogenesis and the spheroid assays. The spreading of lymphatic cells from a pre‐existing vessel was also assessed by the thoracic duct ring assay, revealing similar inhibitory effects, with a partial inhibition at 5 μM and complete suppression of lymphatic outgrowth at 7.5 μM. In a developmental model with zebrafish embryos, toluquinol impaired thoracic duct formation at 15 μM, leading to lymphatic abnormalities in 66% of embryos treated with 20 μM toluquinol. Examination of the embryos revealed the inhibitory effect of toluquinol on the lymphangiogenic process without any evidence of toxicity, developmental delay or other unhealthy signs. However, because of the ability of toluquinol to interfere with the angiogenic process (García‐Caballero et al., 2013), we cannot exclude an indirect effect on lymphangiogenesis in this developmental model. The antilymphangiogenic action of toluquinol is supported in vivo by the strong inhibitory effects observed in gelatin sponges soaked with VEGF‐C and by the findings in the mouse corneal assay.

Mechanistically, we have provided evidence that toluquinol interferes with VEGF‐C‐induced VEGFR‐3 phosphorylation, but not VEGF‐A‐mediated VEGFR‐2 phosphorylation. Interestingly, toluquinol appears to be more potent at blocking VEGFR‐3 homodimers than VEGFR‐3/VEGFR‐2 heterodimers, because it is more efficient on cells stimulated with the mutated form of VEGF‐C (VEGF‐C156S), which only binds to the homodimer, than on cells treated with the wild‐type VEGF‐C form, which is able to stimulate both the homodimer and the heterodimer (Dixelius et al., 2003; Lohela et al., 2003; Deng et al., 2015). After dimerization and autophosphorylation, VEGFR‐3 serves as a docking site for downstream signalling molecules involving mainly Akt and ERK1/2 pathways (Zhang et al., 2010). Importantly, we also found a repression of downstream signalling mediators (Akt and ERK) in stimulated LECs, suggesting that components of these signalling cascades could also be major targets in the molecular mechanism of toluquinol.

One important finding of our work is that toluquinol is a more potent inhibitor of lymphangiogenesis than angiogenesis. Indeed, no significant effect on angiogenesis was detected in the sponge assay at the concentrations tested (up to 4 nmol), although a reduction in the lymphatic vasculature was evident. We previously reported that toluquinol displayed antiangiogenic activity in the in vivo Matrigel plug assay, in which we used 30 nmol, a 7.5‐fold higher concentration than that used here in the gelatin sponges (García‐Caballero et al., 2013). In addition, it is worth mentioning that in this assay, FGF‐2 was used instead of VEGF‐C to stimulate angiogenesis. In the corneal assay, toluquinol was again more potent on the neoformed lymphatic vasculature than on that of the blood. It is also worth noting the capacity of toluquinol to decrease the number of filopodia‐like structures at the end of the corneal lymphatic bud, suggesting an effect on the composition of the tip cells. Altogether, the more potent antilymphangiogenic effect of toluquinol fits with its ability to primarily inhibit VEGFR‐3 phosphorylation without affecting VEGFR‐2 activation.

Through its ability to block the VEGF‐C/VEGFR‐3 axis, toluquinol appears to be a good candidate for various therapeutic applications. Several antitumour drugs display lymphangiogenic inhibitory properties and regulate the VEGFR‐3 signalling pathway; these include deguelin (Hu et al., 2010), endostar (Dong et al., 2011), liposomal honokiol (Wen et al., 2009), Ki23057 (Yashiro et al., 2009), etodolac (Iwata et al., 2007), MMI270 (Nakamura et al., 2004), CSDA (Matsumoto et al., 2010) and norcantharidin (Liu et al., 2012). Blocking the VEGFR‐3 cascade has been reported to impair metastatic dissemination and tumour growth in lymph nodes (Zwaans and Bielenberg, 2007; Sleeman and Thiele, 2009). Thus, it has been experimentally demonstrated that lymph node‐specific antibody‐mediated inhibition of VEGFR‐3 activation in response to VEGF‐C draining to the nodes from metastatic breast tumours inhibits the outgrowth in these nodes (Quagliata et al., 2014). As well as cancer and metastasis, ocular diseases are also associated with abnormal lymphangiogenesis. Although the cornea is normally avascular, some infectious, traumatic, degenerative and inflammatory diseases can induce corneal neovascularization, leading to diminished visual acuity (Epstein et al., 1987; Bock et al., 2013). Indeed, corneal neovascularization is the second highest cause of blindness worldwide, and it is associated with a risk of corneal graft rejection (Bachmann et al., 2010). Therefore, pharmacological inhibition of corneal lymph/angiogenesis represents an attractive therapeutic option, and drugs that target lymph/angiogenesis, like bevacizumab, are now available in the clinic to treat ocular pathologies (Amadio et al., 2016). Lymphangiogenesis is also associated with skin inflammatory disorders like psoriasis (Varricchi et al., 2015) and atherosclerosis (Aspelund et al., 2016). In this context, our results suggest a promising role for toluquinol in multiple biomedical fields, such as inflammatory diseases and malignancies associated with excessive lymphangiogenesis.

Conclusions

The data presented here bring new insights into the antilymphangiogenic properties of toluquinol; we demonstrated, for the first time, that toluquinol suppresses lymphangiogenesis by down‐regulating the VEGF‐C/VEGFR‐3 cascade. The present work emphasizes the potential value of toluquinol in the pharmacological treatment of pathologies with excessive lymphangiogenesis, because it is able to block LEC proliferation, migration and tubulogenesis and to induce LEC apoptosis. In line with our previous study, these findings extend the applicability of toluquinol for inhibiting angiogenesis and lymphangiogenesis. Hence, targeting VEGFR‐3 appears to be a realistic therapeutic strategy for inhibiting tumour‐induced lymphangiogenesis and lymph node metastasis, making toluquinol a promising candidate for further preclinical testing. Furthermore, because of its ability to suppress corneal neovascularization, toluquinol could be pursued as a therapeutic approach for the prevention and treatment of vascular pathologies of the cornea and to increase cornea graft survival.

Author contributions

M.G.‐C. and J.P. performed the experiments. A.N., S.B. and M.G.‐C. analysed the data. A.N., M.G.‐C., A.R.Q. and M.A.M. interpreted the data. A.N., M.G.‐C. wrote the manuscript.

Conflict of interest

The authors declare no conflicts of interest.

Declaration of transparency and scientific rigour

This Declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research recommended by funding agencies, publishers and other organizations engaged with supporting research.

Supporting information

Figure S1. Toluquinol does not inhibit cell migration in pretreated LECs after 48 h of treatment. LECs were incubated in presence of toluquinol 5 μM for 24 h, washed with PBS, scratched and fresh culture medium containing 2% FBS, and mitomycin C was added. Photographs were taken at the beginning of the assay and after 24 and 48 h of incubation.

Figure S2. Toluquinol provokes a decrease in the phosphorylated VEGFR‐3, Akt and ERK1/2 protein levels of treated LEC stimulated with VEGF‐C (A), VEGF‐C156S (B) or VEGF‐A (C). Western blot quantification by densitometry. Results are expressed as the percentage of phosphorylated proteins/total protein ± SEM of five western blots. Mann–Whitney–Wilcoxon test was used to determine if the differences among control (stimulated LECs without toluquinol treatment) and toluquinol‐treated and stimulated LECs were statistically significant. ^* P < 0.05 versus control.

Figure S3. Toluquinol affects VEGFR‐3 and downstream target phosphorylations. Serum‐depleted LEC cultures were incubated or not with different toluquinol concentrations for 2 h, stimulated for 30 min with different growth factors, and cell extracts were subjected to western blotting analyses. Complete western blot membranes with the bands obtained for the toluquinol‐treated and stimulated cells with VEGF‐C (A), VEGF‐C156S (B) or VEGF‐A (C) are displayed. Inside rectangles indicated the specific protein detected in each membrane.

Figure S4. Evaluation of toluquinol toxicity in lymphatic ring explant (A) at day 7 of culture by using a LIVE/DEAD kit (in green viable cells; in red dead cells). Characterization of 3D lymphatic ring cultures by immunostainings in whole‐mounted explants after 7 days of culture (B–E): outgrowing LECs were immunopositive (in green) for LYVE‐1 (B) and immunonegative (in red) for Thy1.1 (C). Higher magnification and other field with outgrown LECs in C are detailed in D and E. Blue staining corresponds to nuclear staining with Dapi (bar = 250 μm in A, C and E; bar = 50 μm in B and D).

Figure S5. Toluquinol is a more potent lymphangiogenic inhibitor than angiogenic one. (A) Graph represents the percentage of lymphatic and blood vessel density inhibition by toluquinol in the mouse ear sponge model. Data are obtained by comparison between the total averaged area density value in sponges soaked with VEGF‐C (control) and that value in sponges soaked with VEGF‐C + toluquinol. (B) Graph shows the percentage of lymphatic and blood vessel area, length, branching and end‐point density inhibition by toluquinol in the mouse corneal neovascularization model. Data are obtained by comparison between the total averaged density value in control mice (without toluquinol injections) and that value in mice daily injected with this drug.

Figure S6. Toluquinol reduces corneal neovascularization in mice injected with toluquinol (75 nmol) until day 40 post‐cauterization, but it does not accelerate vessel regression. Corneal lymph/angiogenesis was induced by thermal cauterization and mice were daily i.p. administered with 75 nmol of toluquinol (treated mice) or PBS (control mice). At Day 9 or 40 post‐injury, corneas were immunostained for lymphatic vessels and blood vessels, whole mounted and automatic quantifications were performed. (A) Quantification of lymphatic and blood area densities at Day 40 post‐injury in control and toluquinol‐treated mice. (B) Graphs represent the lymphatic and blood area densities in control and injected mice at Day 9 and 40 post‐cauterization. Differences between densities observed in both days reflect vessel regression. Results are expressed as the mean ± SEM of 15 mice. Mann‐Whitney‐Wilcoxon test was the statistical test used to evaluate significant differences between control mice (PBS) and toluquinol‐injected mice. **P<0.01, ***P<0.001 versus control mice.

Supporting info item

García‐Caballero, M. , Blacher, S. , Paupert, J. , Quesada, A. R. , Medina, M. A. , and Noël, A. (2016) Novel application assigned to toluquinol: inhibition of lymphangiogenesis by interfering with VEGF‐C/VEGFR‐3 signalling pathway. British Journal of Pharmacology, 173: 1966–1987. doi: 10.1111/bph.13488.