The Parotoid Gland Secretion from Peruvian Toad Rhinella horribilis (Wiegmann, 1833): Chemical Composition and Effect on the Proliferation and Migration of Lung Cancer Cells

Since Rhinella sp. toads produce bioactive substances, some species have been used in traditional medicine and magical practices by ancient cultures in Peru. During several decades, the Rhinella horribilis toad was confused with the invasive toad Rhinella marina, a species documented with extensive toxinological studies. In contrast, the chemical composition and biological effects of the parotoid gland secretions (PGS) remain still unknown for R. horribilis. In this work, we determine for the first time 55 compounds from the PGS of R. horribilis, which were identified using HPLC-MS/MS. The crude extract inhibited the proliferation of A549 cancer cells with IC50 values of 0.031 ± 0.007 and 0.015 ± 0.001 µg/mL at 24 and 48 h of exposure, respectively. Moreover, it inhibited the clonogenic capacity, increased ROS levels, and prevented the etoposide-induced apoptosis, suggesting that the effect of R. horribilis poison secretion was by cell cycle blocking before of G2/M-phase checkpoint. Fraction B was the most active and strongly inhibited cancer cell migration. Our results indicate that the PGS of R. horribilis are composed of alkaloids, bufadienolides, and argininyl diacids derivatives, inhibiting the proliferation and migration of A549 cells.


Introduction
Toads are widely distributed in the different biotas of Peru with a higher number of species in the Amazonian and eastern Andes [1,2] and are less frequent in the dry coastal ecosystems. Toads were known by the Native Americans both for their role in controlling insects as well as by their toxic defense substances. These toad toxins are secreted from parotoid glands that in contact with oral mucosa or eyes can produce intoxication or death [3][4][5]. Ceramic representations of Rhinella toad species by Peruvian pre-Inca Chavin and Moche cultures [6,7] and current local magic traditions in northern Peru that involved these toads ( Figure 1A,B) suggest the presence of bioactive compounds. Currently, no studies are reporting the chemical composition and biological effects of the parotid glands secretions (PGS) of Rhinella sp. toads of the northern coast of Peru. The PGS of South American toads include alkaloids, argininyl diacids, bufadienolides, and argininyl diacid derivatives from bufadienolides. The biological effects from these secretions include cardiotonic, cytotoxic, and antiproliferative activities, sometimes with a remarkable and selective effect on several cancer cell lines [8][9][10].
Non-small cell lung cancer (NSCLC) is the leading cause of cancer-related deaths globally [11]. The NSCLC tumors have high molecular heterogeneity, exhibiting gene mutations in B-Raf proto-oncogene, serine/threonine kinase (BRAF), KRAS proto-oncogene, and activating mutations of epidermal growth factor receptor (EGFR) [12]. Although recent advances in the development of new therapeutic strategies are focused on the reduction of proliferation and tumor growth [13,14]; reduced efforts have been put into identifying anti-invasive and anti-migratory drugs [15]. Promising uses of microtubule-targeting agents (MTA) have been proposed for NSCLC treatment in advanced stages [13]. Beyond the known effects of MTA on cell cycle progression, the blocking the microtubules stabilization by these anti-cancer drugs produces disruption of cell protrusions and cytoskeleton, triggering inhibition of tumor-induced angiogenesis and cancer cell migration/invasion [16].
Previously, we and others have described that the PGS of Rhinella species produces inhibition of proliferation by cell cycle arrest and induction of apoptosis [9,17,18], being the bufadienolides a promising source of anti-cancer chemical scaffolds [19]. The South American toad defense secretions consist of a complex mixture of argininyl diacids, bufadienolides argininyl diacids, bufadienolides, indolealkylamines, and peptides. The components less investigated so far are the peptides. Bioactive peptides also are considered a potential class of anticancer natural products, which inhibit cancer proliferation [20,21], angiogenesis [22], and induce selective apoptosis [23].
Little is known on the PGS composition and bioactivity of toad defensive compounds for the western South American Andes. No information is available on the secretion constituents of the recently known species of the northern Peru Rhinella horribilis ( Figure 1C), which was extensively confused with Rhinella marina [24], a highly invasive species with well-known toxicological effects and bioactivities [8,18,25,26].
In this work, we report for the first time the chemical composition of the PGS from the Peruvian toad R. horribilis and describe the effect of the whole secretion and their main fractions on human lung cancer cells.  [31][32][33][34][35][36][37][38], and C (FC; 39-57) of PGS from R. horribilis. Adsorbent: silica gel; solvent system: EtOAc:isopropanol:NH4OH (9:7:4). The plate was sprayed with a freshly prepared panisaldehyde solution reagent followed by heating up to 105 °C for maximum visualization of spots.

1 H NMR Analysis
The 1 HNMR spectrum of the fractions 27-30 (Pool A) showed the characteristic signals for the H-21, H-22, and H-23 protons of the bufadienolide α-pyrone ring at δ 7.37-7.48, 7.86-8.00, and 6.25-6.32 ppm, respectively. Minor compounds of the pool presented a singlet at δ 9.67, 9.58, and 9.48 ppm, corresponding to an aldehyde function. The less abundant bufadienolides in the complex mixture displayed a singlet or broad singlet at δ 10.10, 10.08, 10.07, and 10.03 ppm, compatible with a second group of oxidized derivatives with an aldehyde group. According to Kamano et al. [27], the aldehyde signal resonates at δ 10.06 ppm for hellebrigenin and δ 9.49 ppm for resibufagin. The 1 HNMR spectrum of our sample supports the presence of a mixture of compounds belonging mainly to the bufalin and marinobufagin series, with the aldehydes and further oxidized bufadienolides as minor constituents. The main bufadienolides in the fraction are marinobufagin and bufalin argininyl diacids. The HPLC-MS/MS analysis (Section 2.3) confirmed the occurrence of several hellebrigenin and resibufagin argininyl diacids, supporting the NMR assignments. The Pool B, including fractions 31-38, shows a mixture of argininyl diacids, free bufadienolides, and minor amounts of bufadienolides argininyl diacids. The main groups of the components were resibufagin, telocinobufagin, marinobufagin, bufalin, and resibufogenin derivatives. The identity was established by HPLC-MS/MS analysis (see Section 2.3). The 1 HNMR analysis of the fraction 39-57 (Pool C) showed a main compound with H signals at δ 7.32 d (8.4) (1H), 7.12 s (1H), 6.86 d (8.4) (1H), 4.02 br t (5.6) (2H), 3.71 s (6H) and 3.31 ppm (br t, 5.6), in agreement with dehydrobufotenine (DHB) (compound A2). Additional signals support the presence of a mixture of argininyl diacids and minor amounts of bufadienolides. HPLC-MS/MS analysis of the fraction confirmed the occurrence of DHB as the main constituent and the presence of adipyl-, pimeloyl-and suberoylarginine. The 1 HNMR analysis of the pool of fractions 58-69 and 70-95 showed the characteristic signals for a mixture of argininyl diacids and was not further investigated.

HPLC-DAD-Q-TOF-MS/MS Analysis
The crude PGS and the main fraction pools A, B, and C were analyzed by high performance liquid chromatography (HPLC) coupled to a diode-array detector (DAD) and to a quadrupole time-of-flight mass spectrometry/mass spectrometry detector (Q-TOF-MS/MS) to identify the constituents of the complex mixtures (Table 1). Detected and calculated [M+H] + ions were compared and the error was calculated for the tentative identification of the different compounds. Data analyses followed previous work on toad poison secretions [9,10,18,[28][29][30][31]].

Alkaloids and Guanidine Derivatives
The alkaloid guanidinosuccinic acid (A1) and dehydrobufotenine (A2) were detected in the PGS (Figure 3). The fragmentation pattern and errors comparing the detected and expected mass fully support the identification. Dehydrobufotenin was reported in the PGS of R. marina and R. schneideri [9,18].

Argininyl Diacids of Bufadienolides
In the toad secretions, bufadienolides occur as free and as argininyl diacid derivatives. The identity of the bufadienolide moiety followed from the analysis of the neutral loss of water from the [M+H] + ion and subsequent fragmentation leading to the arginine on at m/z 175. The neutral loss of 368 amu from the [M+H] + ion was used to identify the homologous series of bufalin derivatives. Compounds 7, 8, and 9 were identified as 3-(N-adipoylargininyl)-, 3-(N-pimeloylargininyl)-and 3-(N-suberoylargininyl) bufalin, respectively, in agreement with the literature [9]. The telocinobufagin derivatives, with an additional hydroxyl function at C-5, showed a constant neutral loss of 384 amu, as can be observed for the mass spectra of compounds 16-18. While the identity of the argininyl diacid moiety can be unambiguously assigned, telocinobufagin and its isomer gamabufotalin, with a hydroxyl function at C-11 instead of C-5, show the same molecular formula and fragments. In the R. horribilis poison secretion, two compounds with the molecular formula C 36 H 54 N 4 O 9 were detected at R t 49.8-50.1 and 57.9-58.2 min; two products with the molecular formula C 37 H 56 N 4 O 9 at R t 53.3-54.1 and 62.4-64.7 min; and two compounds presenting the molecular formula C 38 H 58 N 4 O 9 at R t 66.4-67.2 and 68.6-72.0 min, and were assigned to the adipoyl-(compound 10 and 16), pimeloyl-(11 and 17) and suberoyl-(compounds 12 and 18) esters of gamabufotalin and telocinobufagin, respectively. According to Cao et al. [28] for argininyl diacids from gamabufotalin and telocinobufagin, the homologous series have higher Rt for the telocinobufagin than for the gamabufotalin argininyl diesters. Therefore, in the pairs with the same molecular formula and fragments, the compound with lower Rt was tentatively identified as the gamabufotalin argininyl ester. The same observation regarding Rt values is found for the genines by Ye and Guo [29] and Ren et al. [33]. Three compounds with molecular formula C 38  3-(N-suberoylargininyl) hellebrigenin (hellebritoxin) was reported from the PGS of R. marina from the Peruvian Amazon [18] and Paraguayan Rhinella species [9].  The occurrence of hellebrigenin and resibufagin argininyl diacids in our samples was suggested by the 1 HNMR analysis of the PGS and fractions. The isomer pairs with the molecular formula C 36 H 52 N 4 O 10 , C 37 H 54 N 4 O 10, and C 38 H 56 N 4 O 10 , were assigned as 3-(N-adipoyl argininyl) arenobufagin 13 and 3-(N-adipoyl argininyl) hellebrigenin 20; 3-(N-pimeloyl argininyl) desacetylcinobufotalin 14 and 3-(N-pimeloyl argininyl) hellebrigenin 21; 3-(N-suberoyl argininyl) desacetylcinobufotalin 15 and 3-(N-suberoyl argininyl) hellebrigenin 22, respectively. The assignment of the arenobufagin and hellebrigenin argininyl diacids in our samples of PGS follows the elution sequence previously reported for the genines and argininyl esters of both compounds, with hellebrigenin isomers eluting before the corresponding arenobufagin derivatives [33]. Ren et al. [33] reported that arenobufagin elutes before hellebrigenin in their PGS samples. Furthermore, the loss of 29 amu in the hellebrigenin derivatives was observed for compounds 20-22 in agreement with Ye and Guo [29] (Figure 4).
The compound 23 showed the neutral loss of 414 amu and a base peak in agreement with pimeloylarginine and was assigned as 3-(N-pimeloyl argininyl) hydroxyhellebrigenin. Two compounds with the same molecular formula C 38 H 56 N 4 O 11 and R t 48.8-49.3 and 58.1-58.4 min occur in the PGS. The mass spectrum of the compound, described here as compound 24, showed the neutral loss of 414 amu, leading to the suberoylargininyl ion at m/z 331. The compound was assigned as 3-(N-suberoyl argininyl) hydroxyhellebrigenin. The placement of the additional hydroxyl function in compound 24 remains to be established, being C-12 or C-16 the most probable options. The isomers were tentatively identified as 3-(N-suberoyl argininyl) hydroxyhellebrigenin 24 isomer 1 and 2, respectively. The resibufogenin derivatives (compounds 25, 26 and 27) were assigned based on the loss of 366 amu from the [M+H] + ion, leading to the argininyl diacids and were identified as adipoyl-, pimeloy-and suberoylresibufogenin, respectively. The compounds were previously reported from the Amazonian R. marina [18] and Paraguayan Rhinella species [9] (Figures 4 and 5).
The compounds 28, 29, 34, and 35 were tentatively assigned by the neutral loss of the bufadienolide from the [M+H] + ions, leading to the argininyl esters. The compounds 28 and 29 with a neutral loss of 380 amu are compatible with the aldehyde derivatives of resibufogenin (resibufagin) and were assigned as pimeloylargininyl resibufagin 28 and suberoylargininyl resibufagin 29, respectively. The compounds 34 and 35 showed the neutral loss of 396 amu in agreement with aldehyde derivatives of marinobufagin (bufotalinin), being assigned as pimeloylargininyl bufotalinin 34 and suberoylargininyl bufotalinin 35, respectively. The compounds 30, 31, and 32 showed the neutral loss of 382 amu, with base peak for adipoyl-, pimeloyl-, and suberoylarginine and were identified as marinobufagin argininyl diacids, in full agreement with spectrometric data reported for South American Rhinella species [10,18]. Marinobufagin is the 5-hydroxy derivative of resibufogenin. The compound 33, with the molecular formula C 36 H 50 N 4 O 9 shows one additional unsaturation than compound 30, it presents the neutral loss of 380 amu and the base peak for adipoyl arginine. It was tentatively assigned as 3-(N-adipoyl argininyl) marinobufagin-9,11-ene 33. Two compounds with the molecular formula C 38 H 54 N 4 O 11 eluted at R t 52.1-52.4 and 54.7-54.9 min, respectively, and showed the neutral loss of 412 amu, suggesting the occurrence of hydroxybufotalinin derivatives differing in the placement of the hydroxyl group. For hydroxybufotalinin, the most common hydroxylation places are either at C-11, C-12, or C-16. However, we are not able to fully assign the most probable isomers. The compound 36 was tentatively identified as 3-(N-suberoyl argininyl) hydroxybufotalinin 36 isomers 1 and 2, respectively (Figures 4  and 5).
In this study, we identified a high number of constituents (55 compounds) from PGS of R. horribilis compared to the 29 compounds previously reported from three Peruvian Amazon populations of R. marina [18]. Although this difference might be explained by the different environments, access to food or the sensibility of the instrument used, we are not excluding the hypothesis of unique chemical composition patterns for both species. This is especially relevant if is considered that both R. marina and, previously confused R. horribilis, have geographical, genetic, and morphological differences [24,34]. The main compounds in the PGS are marinobufagin and bufalin derivatives, with aldehydes as minor constituents of the mixture. Oxidation of the methyl group at C-10 led to the primary alcohols and then to the corresponding aldehydes, giving origin to the bufotalinin derivatives starting from marinobufagin, resibufagin derivatives from resibufogenin, and hellebrigenin from telocinobufagin, respectively. Further oxidation of the steroid moiety led to alcohol and ketone functions mainly at C-11, C-12, or C-16. The unambiguous identification of the minor constituents requires isolation by preparative HPLC and extensive spectroscopic and spectrometric studies to identify the exact position and stereochemistry of the functional groups. Some differences with the Peruvian R. marina are the occurrence of arenobufagin derivatives and a higher number of oxidized compounds in R. horribilis compared with the Amazonian species. Interestingly, most of the compounds reported in the R. horribilis PGS have been previously described from South American Rhinella species [9,10,18] and/or from the crude drug Venenum Bufonis or Ch'an Su from China and Eastern Asia [27,30,31], which exhibit anti-cancer effects in several cancer cell lines [35][36][37][38][39][40]. Based on this, we propose to evaluate the anti-cancer effect of crude PGS of R. horribilis and their fractions A, B, and C.  30, 31, and 32 showed the neutral loss of 382 amu, with base peak for adipoyl-, pimeloyl-, and suberoylarginine and were identified as marinobufagin argininyl diacids, in full agreement with spectrometric data reported for South American Rhinella species [10,18]. Marinobufagin is the 5-hydroxy derivative of resibufogenin. The compound 33, with the molecular formula C36H50N4O9 shows one additional unsaturation than compound 30, it presents the neutral loss of 380 amu and the base peak for adipoyl arginine. It was tentatively assigned as 3-(N-adipoyl argininyl) marinobufagin-9,11-ene 33. Two compounds with the

Effect of R. horribilis PGS on Proliferation, Clonogenic Capacity and Reactive Oxygen Species (ROS) Levels of Lung Cancer Cells
To evaluate the anti-proliferative activity of crude PGS (L) of R. horribilis and their fractions A, B, and C, lung cancer A549 cells were treated with increasing concentrations by 24 and 48 h and the IC 50 values were calculated. As Table 2 shows, the crude PGS had an IC 50 value of 0.031 ± 0.007 µg/mL at 24 h, and fractions A, B, and C were close to 3.08, 4.4, and 1.26 folds more active than crude PGS, respectively and this was a concentration-dependent effect ( Figure 6A). At 48 h, fraction C exhibited a strong increase in its anti-proliferative activity.   The colony formation is the ability of a single cancer cell to proliferate indefinitely, growing into a colony [41]. Since that the concentrations 0.1 and 0.5 µg/mL shown significant differences in viability ( Figure 6A), we evaluate if the crude PGS and theirs fraction A, B, and C irreversibly affect the clonogenic potential of A549 cancer cells. These cells were treated for 24 h with 0.1 and 0.5 µg/mL of crude PGS and their fractions and then, the culture medium was washed and replaced by a fresh and PGS-free culture medium for seven days. As Figure 6B shows, fraction B evoked a strong and irreversible inhibition of the clonogenic potential and fraction C was only active at 0.5 µg/mL. Moreover, we evaluate the effect of the R. horribilis poison secretion and their fractions on reactive oxygen species (ROS) levels in A549 cancer cells. As Figure 7 shows, all the conditions produced a significant increase in the intracellular ROS levels. Taken together, these results suggest that the PGS of R. horribilis increases ROS production and inhibits the proliferation and clonogenic capacity of lung cancer cells, being the fraction B the most active.

Effect of R. horribilis PGS and Their Fractions on Etoposide-Induced A549 Cell Death
Etoposide (Eto) is a known topoisomerase II inhibitor used in the treatment of several solid cancer, inducing DNA damage, growth arrest, and apoptosis in a concentration-dependent manner [42]. In several cancer cell lines, Eto produces cell cycle arrest in G2-phase [43] through p53 stabilization [44], promoting apoptosis [45]. On the other hand, we have described that the poison secretions of Rhinella schneideri, R. ornata, and R. marina produce G1-or S-phase arrest [9,18], avoiding the cell cycle progression to the G2-phase checkpoint. Therefore, to determine the action of poison secretions of R. horribilis on phases of the cell cycle, we evaluate its effect (at 0.5 µg/mL) on G2/M arrest-induced cell death by Eto. We hypothesized that whether the R. horribilis poison secretion acts on the cell cycle in an earlier phase different to G2/M phase, it will produce a systematic decrease of Eto-induced cell death.
At 48 h of treatment, 25 µM Eto produced 45.50 ± 1.38% of positive annexin V subpopulation (p < 0.001 vs. control), suggesting apoptotic cell death as described [46,47]. As Figure 8 shows, only fraction B induced 22.51 ± 3.32% of apoptotic cancer cells and crude PGS and other fractions lacked cytotoxic effect. As it was hypothesized, the combination of the R. horribilis poison secretion or their fractions plus 25 µM Eto prevented the apoptosis induced by the treatment of only Eto (Figure 8), suggesting that anti-proliferative effect of the R. horribilis poison secretion is produced by cell cycle blocking before of G2/M phase. Effect of fractions and crude extract of parotoid gland secretions of R. horribilis on ROS production in A549 cancer cells. Cells were exposed 24 h to DMSO (control), crude extract (L), and fractions (A, B, and C), and changes in the ROS levels were determined using dihydroethidium (DHE) probe by flow cytometry. Data shown are the mean ± SEM of three independent experiments. ** p < 0.01, *** p < 0.001, vs. control (DMSO). n.s. not significant.

Effect of R. horribilis PGS and Their Fractions on Etoposide-Induced A549 Cell Death
Etoposide (Eto) is a known topoisomerase II inhibitor used in the treatment of several solid cancer, inducing DNA damage, growth arrest, and apoptosis in a concentration-dependent manner [42]. In several cancer cell lines, Eto produces cell cycle arrest in G2-phase [43] through p53 stabilization [44], promoting apoptosis [45]. On the other hand, we have described that the poison secretions of Rhinella schneideri, R. ornata, and R. marina produce G1-or S-phase arrest [9,18], avoiding the cell cycle progression to the G2-phase checkpoint. Therefore, to determine the action of poison secretions of R. horribilis on phases of the cell cycle, we evaluate its effect (at 0.5 µg/mL) on G2/M arrest-induced cell death by Eto. We hypothesized that whether the R. horribilis poison secretion acts on the cell cycle in an earlier phase different to G2/M phase, it will produce a systematic decrease of Eto-induced cell death.
At 48 h of treatment, 25 µM Eto produced 45.50 ± 1.38% of positive annexin V subpopulation (p < 0.001 vs. control), suggesting apoptotic cell death as described [46,47]. As Figure 8 shows, only fraction B induced 22.51 ± 3.32% of apoptotic cancer cells and crude PGS and other fractions lacked cytotoxic effect. As it was hypothesized, the combination of the R. horribilis poison secretion or their fractions plus 25 µM Eto prevented the apoptosis induced by the treatment of only Eto (Figure 8), suggesting that anti-proliferative effect of the R. horribilis poison secretion is produced by cell cycle blocking before of G2/M phase.

Effect of Fractions B and C of the R. horribilis PGS on the Migration of Lung Cancer Cells
To evaluate the effect of components of R. horribilis poison secretion on the migration of cancer cells, we selected both the most active (fraction B) and the less active (fraction C) on inhibition of proliferation. The lung cancer cells were exposed to both fractions by 24 h and the effect on fibronectin-dependent migration was measured. At 0.1 and 0.5 µg/mL, fraction B reduced the migrated cells to 0.79 ± 0.03 and 0.28 ± 0.04 folds of Control and fraction C decreased the migration to 0.95 ± 0.28 and 0.60 ± 0.05 folds of Control, respectively. As Figure 9 shows, fraction B was more active inhibiting the migration of A549 cells.

Effect of Fractions B and C of the R. horribilis PGS on the Migration of Lung Cancer Cells
To evaluate the effect of components of R. horribilis poison secretion on the migration of cancer cells, we selected both the most active (fraction B) and the less active (fraction C) on inhibition of proliferation. The lung cancer cells were exposed to both fractions by 24 h and the effect on fibronectin-dependent migration was measured. At 0.1 and 0.5 µg/mL, fraction B reduced the migrated cells to 0.79 ± 0.03 and 0.28 ± 0.04 folds of Control and fraction C decreased the migration to 0.95 ± 0.28 and 0.60 ± 0.05 folds of Control, respectively. As Figure 9 shows, fraction B was more active inhibiting the migration of A549 cells.
Although certain bufadienolides present in the poison secretion of toads produce cell cycle arrest in S-phase [48], reduce the activity and protein levels of topoisomerases [49,50] and intercalates with DNA [51]; other DNA damage-independent mechanisms have also been recently described [8,26]. For example, the major constituents of the PGS of Rhinella species such as bufalin, marinobufagin, telocinobufagin, and hellebrigenin exhibit strong binding affinities to tubulin in vitro and in vivo conditions, producing dysregulated microtubule arrangement and G2/M cell cycle arrest in cancer cells [8]. In this work, we identified these same tubulin-inhibiting bufadienolides from R. horribilis poison secretions and their fractions, producing irreversible inhibition of proliferation and clonogenic capacity, increasing the ROS production, and inhibiting the fibronectin-dependent migration of human lung cancer A549 cells. An interestingly mechanistic explanation of these effects may be proposed, taking into consideration the effects of current MTA on cancer cells mentioned above [16]. Although certain bufadienolides present in the poison secretion of toads produce cell cycle arrest in S-phase [48], reduce the activity and protein levels of topoisomerases [49,50] and intercalates with DNA [51]; other DNA damage-independent mechanisms have also been recently described [8,26]. For example, the major constituents of the PGS of Rhinella species such as bufalin, marinobufagin, telocinobufagin, and hellebrigenin exhibit strong binding affinities to tubulin in vitro and in vivo conditions, producing dysregulated microtubule arrangement and G2/M cell cycle arrest in cancer cells [8]. In this work, we identified these same tubulin-inhibiting bufadienolides from R. horribilis poison secretions and their fractions, producing irreversible inhibition of proliferation and clonogenic capacity, increasing the ROS production, and inhibiting the fibronectin-dependent migration of human lung cancer A549 cells. An interestingly mechanistic explanation of these effects may be proposed, taking into consideration the effects of current MTA on cancer cells mentioned above [16].

Conclusions
The composition and biological effects of the PGS of the Amazonian R. marina has been extensively studied [8,18,26], while that of R. horribilis remained unknown. In this study, some 55 constituents, including argininyl diacids, bufadienolides argininyl diacids, bufadienolides, and alkaloids were identified/tentatively identified in the PGS of R. horribilis for the first time. The number of constituents of this species historically confused with R. marina [24], was higher than the 29 compounds reported from three Peruvian Amazon populations of R. marina [18].

Conclusions
The composition and biological effects of the PGS of the Amazonian R. marina has been extensively studied [8,18,26], while that of R. horribilis remained unknown. In this study, some 55 constituents, including argininyl diacids, bufadienolides argininyl diacids, bufadienolides, and alkaloids were identified/tentatively identified in the PGS of R. horribilis for the first time. The number of constituents of this species historically confused with R. marina [24], was higher than the 29 compounds reported from three Peruvian Amazon populations of R. marina [18].
The alkaloids from the venoms are mainly tryptophan derivatives and some of them present anti-inflammatory effects, as shown by Zhang, et al. [52] on Chinese samples. Several studies on the composition of PGS have been undertaken with the methanol soluble fraction of the crude venom. Recently, the effect was reported of PGS of Bufo bufo and R. marina on metastatic melanoma, finding that the bufadienolide bufalin was the best antiproliferative agent in the samples [53]. Sinhorin, et al. [54] described the composition of the Amazonian Rhinella margaritifera PGS and reported argininyl diacids, bufotalin, and cinobufagin argininyl diacids as well as the bufadienolides bufotalin, cinobufagin, and telocinobufagin as main compounds.
The composition of the PGS in the Peruvian species R. horribilis differs from that of R. margaritifera by the higher ratio of argininyl diesters, a more complex array of oxidized bufadienolides argininyl diesters, the absence of cinobufagin, and the presence of resibufogenin. When compared with the Rhinella species from the Paraguayan river basin area, the toad from the northern Peruvian coast has a much lower content of marinobufagin and its argininyl diacids, that are the main compounds in R. schneideri and R. ornata, while cinobufagin argininyl diesters are relevant for R. scitula [9]. Marinobufagin and its argininyl diesters are the main compounds in R. marina from the Peruvian Amazon basin [18]. The chemical composition of the Peruvian coast species shows a different profile that may be associated with a different diet and genetic factors. An update on the chemistry, bioactivity, and quality control of the Asian toad Bufo gargarizans has been recently published [55] and provides relevant information to be considered for further studies on the South American species.
Although we describe the anti-cancer effect of the R. horribilis PGS, some issues remain to be explored, such as the action of the isolated constituents of crude secretion and the effect on non-tumoral cells and in vivo studies. The R. horribilis poison secretions produced irreversible inhibition of proliferation and clonogenic capacity, increasing the ROS production, and inhibiting the fibronectin-dependent migration of human lung cancer A549 cells. The fraction B was the most active on proliferative and migration of cancer cells, suggesting that certain constituents present only in this fraction as minor compounds or the combination of the many different compounds acting in synergy may be attractive candidates for further anti-cancer studies.

Parotid Glands Secretions (PGS) Samples
Twenty female Rhinella horribilis (Wiegmann, 1833) toads were collected in the Region of Piura, on the north coast of Peru, between December 2018 to January 2019 with herpetologists from Universidad Nacional de Trujillo. The size and weight of the animals ranged between 15.2 to 19.8 cm and 145 to 310 g, respectively. The animals were gently placed in cardboard boxes and then transferred to the Pharmacognosy Laboratory of Universidad Nacional de Trujillo, Peru. The skin secretions were obtained by mild electric stimulation using platinum electrodes (4-ms pulse width, 50 Hz, 5V) (Memex) by ten seconds, as described by Thirupathi, et al. [56]. Next, the parotid glands of the toads were gently pressed with the hands to remove the PGS, which was placed in a beaker. All animals were released after collecting the samples. The fresh PGS (5.4 g) was resuspended in MeOH and was sonicated for 5 min, filtered, taken to dryness under reduced pressure, and lyophilized, affording 900 mg of the crude extract of PGS. The w/w extraction yield of the MeOH extract was 16.7%.
The obtention of peptides from the toad venom needs a different protocol than that used in our work. The venoms should be extracted with a saline buffer [22] or with acidified water and not with organic solvents. The clean-up requires methods used for protein isolation and identification. The identification of the peptides also requires different analytical tools, MALDI-TOF detection, microsequencing, and synthesis [23].

Thin-Layer Chromatography Analysis
For the thin-layer chromatography (TLC) analysis, commercial plates coated with silica gel F 254 as the stationary phase were used. For visualization of the TLC profile, the plates were first observed immediately after the TLC at wavelengths (λ) 254 and 365 nm. Afterward, the plates were sprayed with a p-anisaldehyde solution reagent freshly prepared (2.0 mL p-anisaldehyde in 340 mL ethanol, 40 mL glacial acetic acid, and 20 mL 97% sulfuric acid) followed by heating up to 105 • C for maximum visualization of spots.

HPC-DAD-Q-TOF-MS/MS Analysis
HPLC-DAD-MS/MS analysis was performed using a UHPLC/HPLC-DAD Bruker Elute LC system coupled in tandem with a Q-TOF Compact (Bruker Daltonik, Bremen, Germany). The same above-mentioned reverse phase column and a linear gradient solvent system (HPLC-DAD analysis, Section 4.3) were performed. The volume injected was 5 µL of a 10 mg/mL solution.
The data were acquired in the full scan mode (range of m/z 50-1500) in positive electrospray ionization mode using capillary voltage at 4.5 kV; nebulizer and drying gas (N 2 ) at 4.0 bar and 9.0 L/min, respectively; dry temperature of 200 • C. The MS/MS spectra were acquired in auto mode using variable collision energy of 250 to 100% of 10 eV. Internal calibration of the instrument was accomplished using 10 mM sodium formate solution introduced to the ion source via a 20 µL loop at the beginning of each analysis. The compounds were identified by comparison of the fragmentation pattern, UV profile, and mass spectra.

NMR Analysis
The NMR spectra were recorded on a Bruker Avance 400 spectrometer (Bruker, Rheinstetten, Germany) at 400 MHz for 1 H and 100 MHz for 13 C in CD 3 OD. Chemical shifts are given in ppm with residual methanol as the internal standard.

Cell Culture Conditions
Human lung cancer A549 cells were grown in Dulbecco's modified Eagle's medium (DMEM), containing 25 mM glucose and 4 mM glutamine supplemented with 10% fetal bovine serum (FBS), penicillin (100 IU/mL), and streptomycin (100 µg/mL). No exogenous pyruvate supplementation was added, and cells were maintained in a humidified atmosphere at 37 • C and 5% CO 2 .

Colony Formation
For the colony assay, A549 cells were seeded in 6-well plates at 500 cells per well according to Franken et al. [41] and incubated for 24 h. The cells were treated with the crude extract (L) and fractions A, B and, C of R. horribilis secretions for 24 h. After treatment, the medium was replaced with fresh medium, and the cells were incubated for 7 days to allow colony formation. Colonies were stained with 0.5% crystal violet solution in 20% methanol and washed with tap water. Colony formation was analyzed with ImageJ software (NIH, Bethesda, MD, USA) as we described [57].

Determination of ROS Levels
The generation of intracellular oxidative stress was determined using the dihydroethidium (DHE) probe as previously described [58]. In brief, A549 cells were grown in complete media, seeded in 12-well plates (50.000 cells/mL) allowed overnight to attach. Cells were exposed for 24 h to DMSO (Control) or 0.1 and 0.5 µg/mL of crude poison secretion (L) or A, B, and C fractions. Then, culture media was replaced by a solution containing 5 µM DHE in Hank's balanced salt solution (HBSS) and incubated for 20 min in the dark. After this time, cells were washed, trypsinized, and resuspended in 200 µL of HBSS and measured by FACSCanto flow cytometer.

Migration Assay
Cell migration was evaluated in Boyden Chamber assays (Transwell Costar, 6.5-mm diameter, 8-µm pore size) as previously reported [58]. The cancer cells were exposed to fractions B and C (0.1 and 0.5 µg/mL) for 24 h. Then, cancer cells (50,000 cells/mL) were re-suspended in serum-free medium, plated on top of the chamber insert coated with fibronectin (2 µg/mL), and incubated at 37 • C for 2 h. The inserts were removed, washed, and the bottom side of the inserts was stained with 0.5% crystal violet solution in 20% methanol. Cells from eight different frames were counted for each condition in an inverted microscope.

Annexin V/Propidium Iodide Staining
The cell death was evaluated using annexin V/propidium iodide (AV/PI) dual staining, following the instructions of the Annexin V-FITC Apoptosis Detection Kit (Abcam, Cambridge, UK), as previously reported [57]. In brief, A549 cancer cells were seeded (50.000 cells/mL) in 12-well plates, treated with 0.5 µg/mL of the crude extract (L) and fractions for 24 h. The fluorescence was measured by flow cytometry.

Statistics Analysis
All statistical analyses were performed using Graph Pad Prism 4.03 (GraphPad Software, San Diego, California, USA). The data are expressed as mean ± SEM of three independent experiments. Statistical analysis was performed using one-way or two-way ANOVA with Bonferroni's post-test for pairwise comparisons. The data were considered statistically significant when p < 0.05. Funding: This work was supported by ANID PCI-Biotechnology #Redbio0027, FONDEQUIP EQM 170023, and "líneas de apoyo a la investigación financiadas por ICBM (2020)". J.P.de A. thanks the Universidad de Talca for a Postdoctoral grant and CP thanks ANID for Master fellowship #22191223.

Conflicts of Interest:
The authors declare no conflict of interest.