Cardenolide-rich fraction of Pergularia tomentosa as a novel Antiangiogenic agent mainly targeting endothelial cell migration

Abstract

Purpose

Angiogenesis related abnormalities underlie several life-threatening disorders. Despite approved therapies, scientists have yet to develop highly efficient, low cost approaches with minimal side effects.

Methods

We evaluated the antiangiogenic activity of 50% hydroalcoholic extracts of Pergularia tomentosa L. root and aerial parts along with their EtOAc and water fractions, in vivo and in vitro. Transgenic zebrafish line Tg(fli1:EGFP) was used for in vivo assay and human umbilical vein endothelial cell (HUVEC) migration test along with possibility of tube formation were performed as in vitro tests. Furthermore, microvasculature in chicken chorioallantoic membrane (CAM) was assessed under P. tomentosa treatment. The fractionation of the 50% hydroalcoholic extracts was led to the identification of the best active fraction in this study. The metabolite profiling of the active fraction was also carried out using LC-HRESIMS analysis.

Results

Pergularia tomentosa markedly inhibited intersegmental vessel (ISV) formation at 48 h post-fertilization (hpf) embryos in zebrafish. The water fraction of root hydroalcoholic extract (PtR2), showed strong antiangiogenic effect with minimal adverse viability impacts. Over 80% of embryos showed more than 50% inhibition in their ISV development at 20 and 40 μg/mL. PtR2 at 20 μg/mL substantially reduced human umbilical vein endothelial cell (HUVEC) migration up to 40%, considerable destruction of the formed tubes in the tube formation and microvasculature in CAM assays. Immunocytochemistry showed a marked reduction in vascular endothelial cadherin (VE-cadherin) abundance at cell junctions concurrent with substantial reduction of phospho-Akt (p-Akt) and β-catenin protein expressions. Phytochemical profile of PtR2 showed a rich source of cardenolide structures, including ghalakinoside, calactin and calotropin derivatives.

Conclusion

Thus, the P. tomentosa cardenolide-rich fraction (PtR2) may hold a considerable promise for an antiangiogenic impact by impairment of endothelial cell (EC) migration and viability.

Graphical abstract

Electronic supplementary material

The online version of this article (10.1007/s40199-020-00356-7) contains supplementary material, which is available to authorized users.

Introduction

Pathological angiogenesis may lead to multiple health problems such as cancer, psoriasis, diabetic retinopathy, atherosclerosis, endometriosis, and rheumatoid arthritis. During angiogenesis, the endothelium undergoes remodeling and paves the way for formation of new capillaries [1]. Extensive research has identified the main signaling molecules and intracellular pathways responsible for angiogenesis. One of the most comprehensively studied molecules is vascular endothelial growth factor (VEGF) and its associated intracellular signaling pathways which has been the main target point for developing therapeutics of angiogenesis-associated disorders [2]. Despite the availability of antiangiogenic agents based on successful translation of the VEGF-A/VEGF receptor blockade, particularly for cancer treatment, a high demand exists for development of therapies with greater efficiencies and lower costs/side effects that target new aspects of angiogenesis signaling pathways, optimize drug administration, and can be combined with available inhibitors of VEGF (signaling) [2, 3]. This research requires full consideration and rapid response due to severe side effects and toxicities of many clinically available antiangiogenic cancer therapies, including cardiotoxicity and hepatotoxicity.

Natural products have been extensively studied as potential antiangiogenic therapeutics. Phenolic compounds especially some of the phenolic acids and flavonoids showed considerable antiangiogenesis activity in vitro and in vivo [4, 5]. The pharmacological properties of many phytochemicals are also discovered followed by pre-clinical and clinical studies of angiogenesis-related disorders which resulted in introduction of different phytochemical groups such as polyphenols, terpenoids, and phytoestrogens [6].

Pergularia tomentosa L. (P. tomentosa) is a plant from the Asclepiadaceae family, which is valuable for local inhabitants of Southern and Southeastern Iran. Its vast biological activities make it a promising plant for several industrial pharmaceutical applications. Traditionally, P. tomentosa has been used for depilation, treatment of constipation, abortions [7, 8], tuberculosis treatment [9], fighting molluscs [10] and as an insect repellant [11]. Previous research studies have reported that it has cytotoxic [12], antioxidant [13], fungicidal [14], and antiproliferative activities [15]. Cardenolides are one of the major phytochemicals in both the roots and leaves of P. tomentosa [16]. Cardenolides or cardioactive steroids are plant toxins [17] well-known for their therapeutic effects for cardiac failure which are sometimes referred to as cardiac glycosides (CGs) [18]. Due to the major anticancer and antiproliferative activities of the Asclepidaceae family, including P. tomentosa [15], and well-known anticancer properties of CGs [19], the CG constituents of this family have also been extensively studied for use as potential anticancer agents. However, to the best of our knowledge, there have not been any studies on the effects of P. tomentosa on angiogenesis, which indirectly attenuates tumorigenesis. In the current study, we screened P. tomentosa extracts using a zebrafish model to explore its antiangiogenic activity, followed by in vitro and ex ovo experiments to discover the precise mechanisms of action of the active fraction(s) in an attempt to find a novel candidate for angiogenesis-related diseases.

Materials and methods

Plant materials, extraction and fractionation

The roots and aerial parts of P. tomentosa were collected during its growth stage in April 2016 from Kahnouj, Kerman Province in Southeastern Iran. The plant material was identified by Mr. Pourmirzaei and deposited in Kerman Agricultural and Natural Resources Research and education Center with voucher specimen no. 8644. The air-dried plant materials (10 g) were separately powdered and extracted using 100 mL of ethanol/water (50%) by sonication (30 min, room temperature [RT]), and then filtered through Whatman filter paper (no. 1). The obtained extracts were evaporated under reduced pressure using a rotary evaporator and finally freeze-dried. Powder extracts from the root and aerial parts of this plant were subjected to a zebrafish screening bioassay. According to the positive result, they were partitioned with water and ethyl acetate (EtOAc) to obtain more and less polar fractions, respectively. We suspended 100 mg of hydroalcoholic extract from the root (PtR) in 100 mL of distilled water, and partitioned it in a separating funnel with EtOAc (3 × 35 mL) to obtain the EtOAc (PtR1) and water (PtR2) fractions. The same process was carried out for the hydroalcoholic extract of the aerial part (PtS) to obtain the EtOAc (PtS1) and water (PtS2) fractions. All of the plant extracts and fractions were prepared as 1.66 mg/mL stock solutions in E3 or EGM-2 medium (CC-4147, Lonza). For full description of chemicals preparation, see supporting information.

LC-HRESIMS analysis

The active fraction from P. tomentosa was analyzed using liquid chromatography (LC) coupled to electrospray ionization and high resolution mass spectrometry (LC-HRESIMS) with an LTQ Orbitrap XL mass spectrometer. Detailed description of the method can be found in supporting information.

Assay of plant extracts using Tg(fli1:EGFP) zebrafish

The transgenic zebrafish line Tg(fli1:EGFP) was subjected to serial concentrations of extracts or fractions and their intersegmental vessel (ISV) formation was analyzed. For detailed description, see supporting information.

Cell culture and in vitro experiments

Human umbilical vein endothelial cells (HUVECs) were isolated from aseptic human umbilical cords, which were received from Arash Hospital, Tehran, after obtaining written consent from volunteer couples. HUVECs were cultured in EGM-2 supplemented with 10% fetal bovine serum (FBS; 10,270, Gibco). All in vitro experiments including cell viability, wound healing and tube formation assays, acetylated low-density lipoprotein (Dil-Ac-LDL) uptake as well as protein expression analysis assays including immunostaining and western blot were performed using passages 3–6 HUVECs according to standard methods and their detailed descriptions can be found in supporting information.

Chorioallantoic membrane (CAM) assay

Fertilized eggs from Hy-line W-36 hens were obtained from a commercial farm. The eggs were cracked under a sterile laminar flow hood and the contents were transferred to a petri dish. Each embryo with the yolk was then transferred to a surrogate shell, which was 3–4 g heavier than the egg shell, sealed with plastic wrap, and allowed to incubate in a forced-air incubator for 60 h at 37 °C and 60% humidity. The embryonic day when eggs were placed in the incubator was considered to be embryonic day 0 (ED0). On ED2.5, the yolk embedded embryo was transferred to a second surrogate shell, which was 35 to 40 g heavier than the egg shell, sealed with plastic wrap, and allowed to incubate for another 5 days. Dead or infected embryos were removed daily to avoid further contamination. The CAM angiogenesis assay was performed as previously described [20]. Briefly, O-ring paper filters soaked in EGM-2 that contained PtR2 were deposited on the intact CAMs at ED9, at a location distal from the embryo and proximal to the major blood vessels. The embryos were maintained in the incubator for 48 h. At ED11, the embryos were transferred to the stage of a SZX16 Wide Zoom Versatile Stereo Microscope (Olympus) and images were taken from inside the O-rings. The number of branches were calculated for 5 random images in each treatment and averaged.

Statistical analysis

All data were expressed as mean ± standard error of the mean (SEM). Each experiment was performed in triplicate for three independent biological replicates unless otherwise stated. Comparisons among groups were performed using an unpaired t-test. P < 0.05 was considered statistically significant.

Results

A transgene model of fli1:EGFP is widely used for angiogenic studies in zebrafish. Development of inter-segmental vessels tagged with GFP are used for real-time visualization of angiogenesis. A primary screening of P. tomentosa extract was performed at the concentrations listed in Table S1. At 10 to 12 hpf, the zebrafish embryos were incubated with serial concentrations of the hydroalcoholic extract of the root (PtR) and the hydroalcoholic extract of the aerial part (PtS). Based on ISV formation results and the mortality rate, we chose the 10, 20, 40, and 80 μg/mL for further studies (Fig. 1S). PtR and its fractions (PtR1 and PtR2) inhibited ISV formation at 48 hpf embryos (Fig. 1S A). Moreover, PtR treated embryos showed less than 30% mortality, while PtS resulted in less than 50% viable embryos and in some concentrations the mortality was almost 100% (Fig. 1S B). Of note, PtR mortality rates at 10 μg/mL (14%), 40 μg/mL (24%), and 80 μg/mL (0%) were lower than 2 mM SU5416 (28%) which is an indoline synthetic derivatives, selectively inhibits the vascular endothelial growth factor receptor and used as an antiangiogenic agent (Fig. 1SB).

As stated above, both PtR and PtS extracts were screened in zebrafish, and further confirmed inhibition of ISV formation by PtR (Fig. 1S). We compared mortality rates in PtR1 and PtR2, and also considered our developmental observations, and finally selected PtR2 due to its better inhibition of ISV formation (Fig. 1SA) and higher number of viable embryos (91.7%) versus 39.3% for PtR1. Further screening with PtR2 bioassays showed that PtR2 at 20 and 40 μg/mL attenuated intersegmental vasculogenesis to a large extent, so that 33.7% and 63% of 48 hpf embryos were determined to have score 1 angiogenesis, respectively, while only 0% and 2.3% had fully formed ISVs (Fig. 1a, b). In contrast, 92 ± 4.6% of untreated embryos showed fully formed ISVs and had a score of 4 (Fig. 1a, b). When compared to SU5416, PtR2 exerted the same effect on ISV formation, with a score 1 of 56.7% (SU5416), 33.7% (20 μg/mL), and 63% (40 μg/mL). There was also a reduced mortality rate of 12% (20 μg/mL) and 10.7% (40 μg/mL) compared with 19% in SU5416 (Fig. 1c). PtR2 markedly weakened ISV formation compared to VEGF-A, which is a proangiogenic molecule (score 4).

Fig. 1.

PtR2 inhibited the outgrowth of intersegmental vessels (ISVs). A) Embryos at 48 h’ post-fertilization (hpf). Left: anterior; top: dorsal. Intersegmental vessel (ISV) formation was normal in the control and VEGF-A groups, while the outgrowths were inhibited in the SU5416, 20 μg/mL PtR2 (PtR2–20) and 40 μg/mL PtR2 (PtR2–40) groups. White arrows show partial ISV outgrowth formation. Scale bar; 500 μm. B) Percentage of larvae classified into 4 categories of ISV formation that range from undeveloped to fully developed (1 to 4) in each group. C) Mortality rate of embryos in each group. Error bars indicate SEM of three independent experiments. **P < 0.01, ***P < 0.001 compared to Ctrl

Angiogenesis occurs through complex interactions between several cell types, in which migration and tube formation of ECs are two major steps. We used the wound-healing migration and tube formation assays to assess the effect of PtR2 on ECs for both phenomena. We observed considerable reduction in migration at all concentrations used (Table S1) compared to the control, thus we performed the rest of in vitro experiments with two minimum concentrations i.e. 10 and 20 μg/mL. While 20 μg/mL of PtR2 reduced EC migration by about 64%, SU5416 inhibited migration by 24% compared to untreated ECs after 24 h (Fig. 2a, c). In order to test vessel formation capacity of P. tomentosa treated ECs, HUVECs were stimulated by plating them on the surface of a growth factor enriched Matrigel matrix which provides sufficient proangiogenic stimuli. PtR2 substantially reduced formation of tube-like structures over 6 h, as shown by the decreased number of tubes (39% of control) and branches (82% of control) (Fig. 2b, d, e). Of note, PtR2 markedly accelerated the destruction of tube-like structures after 18 h compared to untreated ECs (Fig. 2b). These findings showed that PtR2 had a negative impact on the angiogenic capacity of HUVECs in vitro.

Fig. 2.

PtR2 inhibited human umbilical vein endothelial cell (HUVEC) migration and accelerated destruction of formed tubes. A) Representative images of wound healing assay at 0, 6, 12 and 24 h showing PtR2 inhibition of human umbilical vein endothelial cell (HUVEC) migration at 10 and 20 μg/mL in contrast to wound closure in the control group. Scale bar: 500 μm. B) Representative images of tube formation by HUVECs on Matrigel at 2 and 18 h post-seeding. Scale bar:200 μm. C-E) Quantification of the results in wound healing and tube formation assays by analyzing the percentage of wound closures (C), number of tubes (D), and number of branches (E). Error bars indicate SEM of three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001 compared to Ctrl

A popular angiogenesis model is the chorioallantoic membrane (CAM), a highly vascularized membrane located beneath the eggshell. Its location provides easy access for application of diverse substances and monitoring vessel growth in response to these substances in real-time. We further examined the effect of 10 and 20 μg/mL of PtR2 on the CAM vasculature and observed substantial microvessel destruction after 48 h (Fig. 3a). PtR2 reduced the branches by 80% at 10 μg/mL and 85% at 20 μg/mL compared to aged-matched untreated eggs. SU5416 was not as efficient as PtR2 in microvessel destruction (Fig. 3a and b).

Fig. 3.

CAM assay which showed diminution of the adult arterioles and venules on the CAM vascular bed. A) Representative images of vascular bed expanded on chicken chorioallantoic membrane (CAM) in the presence and absence of PtR2. Black arrows show main arteries, arterioles, and venules in the control and SU5416 groups, and their absence following PtR2 treatment. B) Quantitative analysis of the number of branches on the CAM vascular bed. Error bars indicate SEM of three independent experiments. **P < 0.01, ***P < 0.001 compared to Ctrl

Cell migration of angiogenic ECs depends on the rearrangements at cell junctions in which vascular endothelial cadherin (VE-cadherin) plays an important role. Upregulation of VE-cadherin has been shown to result in proangiogenesis in vitro and in vivo [21, 22]. Interestingly, when the HUVECs were subjected to 10 and 20 μg/mL of PtR2 for 18 h, we observed a relevant reduction in VE-cadherin at the cell junctions, which was shown by immunolabeling of this junctional protein (Fig. 4a). While untreated HUVECs showed VE-cadherin staining at the plasma membrane, the abundance of this protein was very low or even absent after treatment with PtR2 at indicated concentrations. Of note, an increase in cell size and cell death was observed after treatment of HUVECs with 10 and 20 μg/mL of PtR2 for 18 h as reflected by lower density and larger sizes in immunostaining. However, cell death effect was further analyzed by Annexin V/PI and MTS assays which will be explained in the following. Application of SU5416 did not affect membrane staining of VE-cadherin (Fig. 4a).

Fig. 4.

PtR2 reduced vascular endothelial cadherin (VE-cadherin) abundance at cell junctions. A) Representative images of vascular endothelial cadherin (VE-cadherin) immunostaining in human umbilical vein endothelial cells (HUVECs) show a high abundance of this adhesion molecule at the cell junctions of the control and SU5416 groups, but not the PtR2-treated HUVECs. B) PtR2 at 10 and 20 μg/mL were added to HUVECs for 24 h and analyzed for Akt phosphorylation (p-Akt) and β-catenin protein expression by Western blotting. C) Quantitative analysis of Western blot indicates significant reduction of p-Akt and β-catenin protein expression by PtR2 treatment. The CTNNB1 gene encodes β-catenin. Error bars indicateSEM of three independent experiments. *P < 0.015, **P < 0.01 compared to Ctrl

As shown above, PtR2 inhibited HUVECs migration and tube formation, and accelerated tube-destruction. To further study the intracellular regulators of this observation, we performed protein expression analyses of phospho-Akt (p-Akt) and β-catenin, two important regulators of the angiogenic response in ECs. The protein abundance of both p-Akt and β-catenin substantially decreased when cells were treated with 10 and 20 μg/mL of PtR2 (Fig. 4b, c). In contrast, SU5416 did not affect the expressions of these two proteins in HUVECs.

PtR2 not only reduced Dil-Ac-LDL uptake, which is a measure of HUVECs functionality (Fig. 5a and b), but also increased apoptosis, as confirmed by the Annexin V/PI and MTS assay results at most of the applied concentrations, including 10 and 20 μg/mL (Fig. 5c and Fig. 2S). Annexin V/PI can assess viable, apoptotic, and necrotic cells based on plasma membrane integrity and permeability. PtR2 substantially reduced the percentage of viable cells to ~80% at 10 μg/mL and 75% at 20 μg/mL. This finding was also seen with Annexin V staining. However, the population of apoptotic cells did not differ markedly between untreated and PtR2 treated HUVECs (Fig. 5c). This might originate from the inhomogeneous distribution of apoptotic cells at both the early and late apoptosis stages. The percentage of necrotic cells also remained unchanged (Fig. 5c).

Fig. 5.

PtR2 reduced the uptake of acetylated low-density lipoprotein (Dil-Ac-LDL). A) Representative images of fluorescently labeled Dil-AC-LDL uptake by human umbilical vein endothelial cells (HUVECs). Red fluorescent dots show LDL versus nuclear staining with 4′, 6-diamidino-2-phenylindole (DAPI). LDL uptake was markedly reduced in PtR2-treated HUVECs as reflected by the decreased red fluorescent intensity. B) Quantitative analysis of fluorescent intensity related to labeled LDL uptake in each group. C) Quantitative analysis of Annexin V/propidium iodide (PI) staining assay representing the percentage of live (L), early apoptotic (EA), late apoptotic (LA) and necrotic (N) cells. Error bars indicateSEM) of three independent experiments. ***P < 0.001 compared to Ctrl

The Bioassay showed a promising bioactivity of PtR2. In order to identify the metabolites of PtR2, we performed the metabolite profiling procedure or dereplication using LC-HRESIMS analysis. Figure 3S represents the chromatogram of analysis in negative ion mode of the mass spectrometry. The interpretation of the data revealed that PtR2 is a rich source of cardenolides and flavonoids. Table 1 shows all the identified compounds in this bioactive fraction. The chemical formula, observed mass and MS/MS fragmentation along with references for each compound have been shown in the table. PtR2 mainly consists of known cardenolides (4, 8, 9 and 17) and doubly linked cardenolides (1–3, 5, 7, 11–16, and 18–21), along with two flavonol glycoside (6, 10) structures. Chemically, the doubly linked cardenolides belong to the two calactin and calatropin derivative groups, which differs in the configuration of C-3′.

Table 1.

All the identified compounds in the active fraction of Pergularia tomentosa, PtR2, using reverse phase LC coupled to HRESIMS with an LTQ Orbitrap XL mass spectrometer

Rt* (min)	Molecular formula	Δ ppm	[(M + HCOOH)-H]−	[M-H]−	MS/MS	compound	reference
1. 8.85	C30H44O13	−1.21	611.2691		564, 467, 421	ghalakinoside	J. Nat. Prod. (2019), 82(1), 74–79
2. 10.24	C30H42O12	−1.59	593.2583		547, 529, 401	6′-hydroxycalactin	J. Nat. Prod. (2019), 82(1), 74–79
3. 11.48	C32H44O14	−1.10	651.2690		606	6′-hydroxy-16α-acetoxycalactin	J. Nat. Prod. (2019), 82(1), 74–79
4. 11.54	C35H50O15	−1.32		709.3056	529, 305,	calotropagenin	Chem. Pharm. Bull. (1992), 40(11), 2917–20
5. 11.64	C30H42O12	−1.30	593.2585		547, 419	16α-hydroxycalotropin	J. Nat. Prod. (2019), 82(1), 74–79
6. 11.64	C21H20O12	−1.41		463.0865	301	quercetin-3-O-β-D-glucopyranoside	J. Nat. Prod. (2019), 82(1), 74–79
7. 11.69	C29H39O11	−0.64		563.2480	519, 347	12β,6′-dihydroxycalotropin.	J. Nat. Prod. (2019), 82(1), 74–79
8. 11.79	C30H46O12	−1.36	597.2897		551, 373	glucocoroglaucigenin	J. Nat. Prod. (2019), 82(1), 74–79
9. 12.52	C29H42O11	−0.08		565.2638	403, 385, 359	Strophanthidin-glicoside	Chem. Pharm. Bull. (2012), 60(10) 1275–1282
10. 12.88	C22H22O12	−1.37	477.1021		357, 315	isorhamnetin-3-O-β-D-glucopyranoside	J. Nat. Prod. (2019), 82(1), 74–79
11. 12.98	C30H42O12	−1.17	593.2586		529, 419, 401	6’β-hydroxycalotropin	J. Nat. Prod. (2019), 82(1), 74–79
12. 13.30	C30H44O12	−1.03	595.2743		549, 387	12′-dehydroxyghalakinoside	J. Nat. Prod. (2019), 82(1), 74–79
13. 13.40	C30H42O12	−1.30	593.2585			12β-hydroxycalactin	J. Nat. Prod. (2019), 82(1), 74–79
14. 13.66	C30H44O12	−1.28	595.2741		549	6′-dehydroxyghalakinoside	J. Nat. Prod. (2019), 82(1), 74–79
15. 13.92	C36H52O16	−1.30	739.3162		693	3-O-β-glucopyranosiylcalactin	J. Nat. Prod. (2019), 82(1), 74–79
16. 14.59	C36H52O16	−1.30	739.3162		693, 531	glucopyranosiylcalactin isomer	–
17. 14.95	C30H46O11	−1.44	581.2948		535, 391,373	desglucouzarin	J. Nat. Prod. (2019), 82(1), 74–79
18. 16.20	C32H44O13	−1.60	635.2488			16α-acetoxycalotropin	J. Nat. Prod. (2019), 82(1), 74–79
19. 16.77	C29H40O10	−1.07		547.2532	529, 401, 357	16’α-hydroxycalactin	J. Nat. Prod. (2019), 82(1), 74–79
20. 16.77	C30H42O11	−2.00	577.2632		531, 373, 271	calotropin	J. Nat. Prod. (2019), 82(1), 74–79
21. 18.01	C30H42O11	−1.47	577.2683		531, 373, 271	calactin	J. Nat. Prod. (2019), 82(1), 74–79

*Retention time (R_t)

Discussion

Pathological angiogenesis, especially in the tumor microenvironment, directly or indirectly influences disease progression. There are multiple factors involved in the process of new blood vessel formation. Thus, a rational approach that considers various aspects and signaling factors involved in suppressing this remodeling process may significantly increase treatment efficiency. Consequently, finding novel antiangiogenic agents that target multiple aspects of this phenomenon, especially within natural compounds with minimally-induced side effects, has gained significant attention [23, 24]. Most currently identified antiangiogenic phytochemical groups and their derivatives, target VEGF-A/VEGFR-2 signaling pathways [6]. To improve their effects, it is necessary to find other targets in corresponding intracellular signaling.

P. tomentosa is a medicinal plant widely used as an antitumor medication in eastern traditional medicine [25]. We used an in vivo whole-organism model to check whether this plant could target angiogenesis in the context of its antitumor function. A primary screening was performed and resulted in selection of PtR2 as an antiangiogenic fraction that could markedly inhibit ISV development in 48 h embryos. The 20 and 40 μg/mL concentrations of PtR2 had the least mortality and effect on other traits such as body axis and head morphology.

Angiogenesis develops through a complicated remodeling comprising EC growth, migration, sprouting, pruning, and tubulogenesis. In accordance with what we observed in zebrafish, the PtR2 treated ECs showed diminished migration and tube formation. Furthermore, tube-like structures underwent earlier deterioration in comparison to the control group. Previously, the cytotoxic and antiproliferative effects of P. tomentosa extract have been reported in different cancerous cell lines [15, 26]. Further experiments using Annexin V/PI revealed that PtR2 has a proapoptotic effect which induced cytotoxicity in HUVECs, with no significant mortality in zebrafish. The presence of proapoptotic compounds without significant toxicity in zebrafish, has been reported [27]. CAM treatment showed that PtR2 has a significant diminution effect on the adult arterioles and venules expanded on this vascular bed. Migration defects encouraged us to study the major endothelial adhesion molecule, VE-cadherin, which controls cellular junctions and new blood vessel formation [28]. Interestingly, PtR2 treatment led to a reduction of VE-cadherin at the HUVEC plasma membranes, which might be a possible cause of uncoordinated interactions of cell junctions and, consequently, migration defect. It has been previously shown that VE-cadherin gene (CDH5) knockout resulted in disruption of vascular lumen development in mice embryos [29]; however, a partial reduction in CDH5 expression by an inducible gene inactivation system promoted an enhanced angiogenic sprouting using both in vitro HUVECs [30] and in vivo murine retina models [31]. Thus, an adequate cell junction abundance of VE-cadherin is necessary for normal angiogenesis. In addition, we observed a marked reduction in β-catenin protein levels after PtR2 treatment, which might explain the antiangiogenic phenotype observed by this fraction because β-catenin/VEGF-R/PI3K/Akt signaling has been shown to promote angiogenesis [32]. β-catenin transduced HUVECs had an angiogenic phenotype, both in tube formation and matrigel plug assays, by upregulation of VEGF-A and VEGF-C. On the other hand, VEGFR-2 phosphorylation was promoted by β-catenin, and resulted in a VEFG-R/PI3-kinase association and subsequent activation of Akt signaling, which is required for angiogenesis [32].

Cardiac glycosides comprise a large family of naturally derived compounds. Their core structures contain a steroid nucleus with a five-membered or a six-membered lactone ring and sugar moieties which is named as cardenolides and bufadienolides, respectively. Clinical evidences suggest that breast cancer patients who received digitalis treatment have a significantly lower mortality rate, and their cancer cells have more benign characteristics compared to no treatment group (6% vs. 34%) [33]. Interestingly, the anticancer effects of these drugs are exerted at non-toxic concentrations [34]. Several mechanisms of action, including inhibition of cancer cell proliferation, induction of apoptosis, and chemotherapy sensitization, have been reported supporting the potential use of these compounds for cancer treatment [12, 35]. Major plant-derived cardenolides are extracted from the families Scrophulariaceae, Apocynaceae, and Asparagaceae. The extracted cardenolides can influence angiogenesis processes in addition to anticancer functions [36, 37]. Accordingly, the root and aerial parts of P. tomentosa, belonging to Apocynaceae, is cardenolide-rich with potential anticancer effect by inhibition of Na⁺/K⁺-ATPase function [12, 15, 26]. The structure of these compounds, with the single sugar moiety attached in a unique dioxanoid structure and the A/B rings of transfused to the steroidal skeleton, cause the bioactivity and the antiproliferative function of these phytochemicals [15]. Cardenolides and the doubly linked cardenolides occur in different parts of Pergularia species and some other genera from the Asclepiadaceae family, particularly Calotropis and Asclepias, and is rarely found in other families [38]. Metabolite profiling showed a high frequency of cardenolides in the PtR2 fraction with antiangiogenic effects.

According to the results, P. tomentosa and specially its bioactive cardenolide-rich fraction, may be a potential anticancer therapeutics by a combined approach of targeting cancerous cells as well as tumor angiogenesis, through apoptosis promotion and migration suppression.

Non-standard abbreviations

CAM

Chorioallantoic membrane

CGs

Cardiac glycosides

Dil-Ac-LDL

Acetylated low-density lipoprotein

EC

Endothelial cells

EtOAc

Ethyl acetate

HUVECs

Human umbilical vein endothelial cells

ISV

Intersegmental vessel

LC

Liquid chromatography

LC-HRESIMS

LC coupled to electrospray ionization and high resolution mass spectrometry

VE-cadherin

Vascular endothelial cadherin

VEGF

Vascular endothelial growth factor

Authors’ contributions

MH performed the experiments and wrote the manuscript. MA helped for phytochemical parts, plant extracts, fractions and manuscript writing. AM performed the Western blot experiments. SPI and AnC performed the LC-HRESIMS analyses and metabolite profiling. AlC helped in design of zebrafish screenings. SPA designed and supervised the research project and wrote the manuscript.

Funding information

This work was supported financially by Research and Technology Council of Royan Institute.

Compliance with ethical standards

Conflicts of interest/competing interests

The authors declared no conflicts of interest.

Ethics approval

All animal studies were performed in accordance with guidelines approved by the Ethics Committee of Royan Institute in conformity with the NIH Guide for the Care and Use of Laboratory Animals.

Consent to participate

Not applicable.

Consent for publication

Written informed consent for publication was obtained and attached.