Assessment of Photo-Induced Cytotoxic Activity of Cachrys sicula and Cachrys libanotis Enriched-Coumarin Extracts against Human Melanoma Cells

Photochemotherapy is one of the most interesting current therapeutic approaches for the treatment of melanoma. Different classes of naturally occurring phytochemicals demonstrated interesting photoactive properties. The aim of this study was to evaluate the photocytotoxic potential of two Cachrys species from Southern Italy: C. sicula and C. libanotis (Apiaceae). The enriched-coumarin extracts were obtained from aerial parts through both traditional maceration and pressurized cyclic solid-liquid (PCSL) extraction using Naviglio extractor®. Qualitative and quantitative analyses of furanocoumarins were performed with GC-MS. The photocytotoxic effects were verified on C32 melanoma cells irradiated at a dose of 1.08 J/cm2. The apoptotic responses were also assessed. Moreover, phenolic content and the in vitro antioxidant potential were estimated. Xanthotoxin, bergapten, and isopimpinellin were identified. All the samples induced concentration-dependent photocytotoxic effects (IC50 ranging from 3.16 to 18.18 μg/mL). The C. libanotis sample obtained with Naviglio extractor® was the most effective one (IC50 = 3.16 ± 0.21 μg/mL), followed by C. sicula sample obtained with the same technique (IC50 = 8.83 ± 0.20 μg/mL). Both Cachrys samples obtained through PCSL induced up-regulation of apoptotic signals such as BAX (Bcl2-associated X protein) and PARP (poly ADP-ribose polymerase) cleavage. Moreover, these samples proved to be more photoactive, giving a greater upregulation of p21 protein in the presence of UVA radiation. Obtained results suggest that investigated species could be promising candidates for further investigations aimed to find new potential drugs for the photochemotherapy of skin cancer.


Introduction
The World Health Organization (WHO) indicates that cancer is a dominant public health issue with a high mortality rate. Among skin cancer, melanoma is the most aggressive form, with an average of 232,000 new cases/year [1]. It is possible to identify three types of skin cancer, whose names derive from the kind of involved cells: Basal cell carcinoma, squamous cell carcinoma, and malignant melanoma. Basal and squamous types belong to the group of non-melanocytic skin cancer while malignant melanoma is the most aggressive and metastatic type of skin cancer which develops from melanocytes [2].
MAP kinases and phosphatidylinositol-3-kinase (PI3K) are the major signal transduction pathways involved in the development of melanoma. Currently, there are many treatment strategies aiming at blocking the migratory and invasive capacity of cancer cells [3]. The cellular response to most drugs is usually associated with high cell resistance,

Phytochemical Profile
In this work, aerial parts of C. sicula and C. libanotis species from Southern Italy were extracted with methanol through two different extraction techniques: Traditional maceration (TM) and pressurized cyclic solid-liquid (PCSL) extraction. As regards the raw extracts obtained through PCSL technique, achieved yields were 7.05% and 12.56% for C. sicula L. (CSN) and C. libanotis L. (CLN), respectively (Table 1). Higher yields were instead obtained with the classical maceration method: 19.48% and 17.76 for C. sicula (CSM) and C. libanotis (CLM), respectively. Total phenolic and total flavonoid contents of Cachrys spp. raw extracts were expressed as equivalents of chlorogenic acid or quercetin per g of dry plant material, respectively. The highest amounts were detected in the extracts obtained through maceration. CLM showed the highest amounts of phenolic compounds (25.0 ± 0.2 mg/g, Table 1) and flavonoids (1.29 ± 0.04 mg/g), followed by CSM (23.1 ± 0.7 and 0.65 ± 0.04 mg/g). Lower amounts were observed for the other two samples both for phenolic content (12.8 ± 0.1 and 11.4 ± 0.2 mg/g for CLN and CSN, respectively) and flavonoid content (0.09 ± 0.01 and 0.41 ± 0.02 mg/g).
C. sicula also demonstrated a higher content in terpenes: Limonene, fenchone, estragole, anethole, and neophytadiene were identified. On the contrary, estragole was the only compound found in C. libanotis.
C. sicula also demonstrated a higher content in terpenes: Limonene, fenchone, estragole, anethole, and neophytadiene were identified. On the contrary, estragole was the only compound found in C. libanotis.
Given the importance of furanocoumarins for the phototoxic potential of investigated samples, the quantification of the three psoralens xanthotoxin, bergapten, and isopimpinellin was then accomplished (Table 3). Xanthotoxin was the most abundant compound (concentration ranging from 2.08 to 4.98 mg/mL). CLN showed the highest content (4.98 ± 0.21 mg/mL), followed by CSN obtained through the same extraction method (4.10 ± 0.23 mg/mL), while traditional maceration allowed to reach lower amounts. A similar trend was observed for bergapten, with a higher content in CLN and CSN (0.59 ± 0.08 and 0.58 ± 0.04 mg/mL, respectively) compared to CLM and CSM (0.27 ± 0.02 and 0.17 ± 0.01 mg/mL, p < 0.05, Bonferroni post-hoc test). No statistical differences were detected among samples as regards the isopimpinellin content.

Photocytotoxic Activity
To evaluate the anticancer activity of Cachrys samples on melanoma, C32 cells (human melanoma) were treated with different concentrations of each sample for 30 min of incubation and 1 h of irradiation, and cell viability was determined after 48 h. All samples from the two Cachrys species, obtained with two extraction methods, affected cell viability in a concentration-dependent manner after irradiation for 1 h at a dose of 1.08 J/cm 2 (IC 50 ranging from 3.16 to 18.18 µg/mL, Figure 2, Table 4).

Photocytotoxic Activity
To evaluate the anticancer activity of Cachrys samples on melanoma, C32 cells (human melanoma) were treated with different concentrations of each sample for 30 min of incubation and 1 h of irradiation, and cell viability was determined after 48 h. All samples from the two Cachrys species, obtained with two extraction methods, affected cell viability in a concentration-dependent manner after irradiation for 1 h at a dose of 1.08 J/cm 2 (IC50 ranging from 3.16 to 18.18 μg/mL, Figure 2, Table 4).  Obtained results showed that the extraction with Naviglio extractor ® allowed a better chemical composition for the antitumor activity than the traditional maceration technique. The best activity was observed for CLN sample (C. libanotis Naviglio), with an IC50 of 3.16 μg/mL, a very interesting value if compared with the positive control bergapten (p < 0.05, Bonferroni post-hoc test). The C. sicula sample obtained with Naviglio extractor ® also showed a very good inhibitory activity on cell line viability (IC50 = 8.83 ± 0.20 μg/mL). Even if both these samples, at the highest concentration tested, also induced cytotoxic effects in the dark, the IC50 values observed for unirradiated cells were significantly higher than those referred to UVA-treated plates (IC50 values equal to 55.20 ± 1.65 and 26.72 ± 1.18 μg/mL for CLN and CSN, respectively). A lower, but still noteworthy, photocytotoxicity was exerted by the two extracts obtained through maceration (IC50 values of 14.57 ± 0.95  Obtained results showed that the extraction with Naviglio extractor ® allowed a better chemical composition for the antitumor activity than the traditional maceration technique. The best activity was observed for CLN sample (C. libanotis Naviglio), with an IC 50 of 3.16 µg/mL, a very interesting value if compared with the positive control bergapten (p < 0.05, Bonferroni post-hoc test). The C. sicula sample obtained with Naviglio extractor ® also showed a very good inhibitory activity on cell line viability (IC 50 = 8.83 ± 0.20 µg/mL). Even if both these samples, at the highest concentration tested, also induced cytotoxic effects in the dark, the IC 50 values observed for unirradiated cells were significantly higher than those referred to UVA-treated plates (IC 50 values equal to 55.20 ± 1.65 and 26.72 ± 1.18 µg/mL for CLN and CSN, respectively). A lower, but still noteworthy, photocytotoxicity was exerted by the two extracts obtained through maceration (IC 50 values of 14.57 ± 0.95 and 18.18 ± 1.33 µg/mL, for C. sicula and C. libanotis, respectively), which did not affect cell viability in the dark. Figure 3 shows cell morphology 48 h after UVA irradiation as captured on a digital camera. The incubation of cell cultures with a concentration of 25 µg/mL of both C. sicula and C. libanotis PCSL samples and UVA radiation significantly affected cell viability compared to control (untreated irradiated cells).
Plants 2021, 10, x FOR PEER REVIEW 6 of 13 and 18.18 ± 1.33 μg/mL, for C. sicula and C. libanotis, respectively), which did not affect cell viability in the dark. Figure 3 shows cell morphology 48 h after UVA irradiation as captured on a digital camera. The incubation of cell cultures with a concentration of 25 μg/mL of both C. sicula and C. libanotis PCSL samples and UVA radiation significantly affected cell viability compared to control (untreated irradiated cells).

Apoptotic Responses on C32 Cells
The two Cachrys species were also evaluated to assess apoptotic responses on C32 cells. The Naviglio extracts CSN and CLN increase the cyclin-dependent kinase inhibitor p21 protein, with respect to control, with greater up-regulation under the combination with UV. Both CSN and CLN samples stimulate the pro-apoptotic BAX protein. Consistently, PARP (poly ADP-ribose polymerase) signature fragments, biomarkers of cell death, are specifically induced under CSN and CLN, with or without UV exposure ( Figure 4).
The other Cachrys samples, CSM and CLM, do not elicit the same results mentioned above. In fact, no increase of p21 protein levels are evidenced in C32 cells that underwent treatment with the maceration products of the Cachrys, and if anything, under UV there is a reduction of the protein with respect to single treatment and control. In contrast, the same treatments cause a significant up-regulation of BAX, the levels of which remain equally high in the presence of co-treatment with UV rays, no significant PARP cleavage is evident under the same experimental conditions ( Figure 5).

Apoptotic Responses on C32 Cells
The two Cachrys species were also evaluated to assess apoptotic responses on C32 cells. The Naviglio extracts CSN and CLN increase the cyclin-dependent kinase inhibitor p21 protein, with respect to control, with greater up-regulation under the combination with UV. Both CSN and CLN samples stimulate the pro-apoptotic BAX protein. Consistently, PARP (poly ADP-ribose polymerase) signature fragments, biomarkers of cell death, are specifically induced under CSN and CLN, with or without UV exposure ( Figure 4).
The other Cachrys samples, CSM and CLM, do not elicit the same results mentioned above. In fact, no increase of p21 protein levels are evidenced in C32 cells that underwent treatment with the maceration products of the Cachrys, and if anything, under UV there is a reduction of the protein with respect to single treatment and control. In contrast, the same treatments cause a significant up-regulation of BAX, the levels of which remain equally high in the presence of co-treatment with UV rays, no significant PARP cleavage is evident under the same experimental conditions ( Figure 5).

In Vitro Antioxidant Activity
Furthermore, the in vitro antioxidant activity of the four Cachrys extracts was evaluated. Two methods were employed: The DPPH and the β-carotene bleaching test. All the Cachrys samples demonstrated concentration-dependent radical scavenging potency. The two traditional alcoholic macerates (IC 50 = 102.13 ± 0.79 and 112.73 ± 0.88 µg/mL for CLM and CSM, respectively, Table 5) exerted a better radical scavenging potency than two samples obtained through PCSL extraction (IC 50 = 163.80 ± 2.63 and 212.80 ± 6.91 µg/mL, p < 0.05, Bonferroni post-hoc test). The same samples, CLM and CSM, also demonstrated an interesting antioxidant activity in the β-carotene bleaching method, with IC 50 values equal to 16.77 ± 1.43 (CSM) and 19.22 ± 1.07 µg/mL (CLM) after 30 min of incubation. After 60 min, these extracts were still effective, with IC 50 values of 31.07 ± 1.71 and 27.52 ± 1.73 µg/mL. The samples obtained though Naviglio extractor ® were also effective, even if to a lesser extent.

Discussion
Among skin cancer with rapid growth, the melanoma represents the most invasive and metastatic one. Due to its strong aggressiveness, there are no effective treatments to antagonize cell growth, and often the cancer cells over time develop resistance to chemotherapy. In patients with metastatic melanoma, several immune therapies have been approved, however, with unsatisfactory responses compared to those obtained with targeted therapies [22].
In this context, more recently, there is the great attention towards natural products since they have been useful in antagonizing chemotherapy resistance as well as they lack major side effects [23].
In the present paper, we investigated the effect of two species of Cachrys (C. sicula and C. libanotis), subjected to different extraction processes, on the apoptotic response of the C32 human melanoma cell line.
In particular, as previously reported, by comparing two methods of extraction on aerial parts of Cachrys, the results indicated that pressurized cyclic solid-liquid (PCSL) extraction yields biologically more active products since they address significant apoptotic effects in C32 cells than the traditional maceration (TM).
The solid-liquid dynamic extractor is an innovative technology of solid-liquid extraction that allows to carry out the extraction in a short time, compared to the other currently used techniques. While most of them are based on osmosis and diffusion principles, PCSL works on Naviglio's Principle, based on an increase of pressure of the extracting liquid on the solid plant material and the generation of a negative gradient pressure from the inside to the outside of the plant matrix [20,21].
In line with the other results discussed above, the CSN and CLN extracts are able to increase p21 protein, a well-known cyclin-dependent kinase inhibitor (CKI), that is able to affect all cyclin/CDK complexes [24]. Therefore, Cachrys that underwent the PCSL extraction technique is really stronger against C32 cells. In fact, C. libanotis and C. sicula extracted with Naviglio extractor ® lead proliferative arrest and up-regulation of apoptotic signals such as BAX and PARP cleavage. In addition, the above extracts are more photosensitive since in the presence of UV rays they give a greater upregulation of p21, while the increase in BAX and PARP remain equally sustained, like in the absence of photoactivation.
Thus, malignant melanoma C32 cell line was less responsive to apoptotic induction when incubated with Cachrys samples underwent to traditional maceration (TM), than extracts by Naviglio extractor ® .
These data confirm those of cellular phototoxicity, in which it has been seen that the Naviglio extractor ® allows us to obtain a better chemical composition of Cachrys for its antitumor activity. Similarly, gas chromatography-mass spectrometry analyses allowed us to identify eight coumarins, including three furanocoumarins, the class of compounds known for their ability to act as photosensitizers [25]. All the four Cachrys samples were particularly rich in these tricyclic aromatic compounds made of a furan ring fused to a coumarin (α-benzopyrone) system: Xanthotoxin, bergapten, and isopimpinellin were detected, and quantitative analyses were also carried out. The exact molecular mechanism of these compounds relies upon their chemical structure, which in turn depends on the combination of the furan ring and coumarin backbone in an angular or linear structure, and on the number, type, and location of the attached substituents [26]. The CH 3 presence at C5 and C8 improves the tumor properties [27]. The C. libanotis extract obtained through PCSL technique using Naviglio extractor ® (CLN) showed the highest content of xanthotoxin, (CH 3 at C8) followed by C. sicula obtained through the same extraction method.
Samples CLN and CSN also showed a higher content in bergapten (CH 3 at C5). No statistical differences were observed among the samples for the isopimpinellin content.
On the contrary, the two Cachrys macerates showed higher amounts of polar constituents. For both Cachrys species, these extracts showed a higher phenolic content compared to the extracts obtained through PCSL technique, with amounts equal to 25.0 ± 0.2 mg/g and 23.1 ± 0.7 for CLM and CSM, respectively. The same trend was observed for the total flavonoid content.
Consistently, samples obtained through maceration showed a better antioxidant activity, evaluated by the DPPH radical scavenging assay and the β-carotene bleaching test. Even if all the four Cachrys extracts demonstrated concentration-dependent biological activity, the two alcoholic macerates (IC 50 values equal to 102.13 ± 0.79 and 112.73 ± 0.88 µg/mL) exerted a better radical scavenging potency than the other two samples (IC 50 = 163.80 ± 2.63 and 212.80 ± 6.91 µg/mL), and no significant statistical differences were observed among the two species for this kind of activity. The same trend was maintained in the β-carotene bleaching test, with macerates being significantly more effective than samples realized with PCSL extraction.
In conclusion, the obtained results showed that extraction with the Naviglio extractor ® allowed a better chemical composition for the antitumor activity than the traditional maceration technique, and suggest that the investigated species could be promising candidates for further studies with the aim to find new potential drugs useful in the photochemotherapy of skin cancer. Furanocoumarins, probably, are the compounds directly involved in this activity, but the most important aspect is how these compounds are present together in the tested sample showing a high activity.

Plant Materials and Extraction Procedures
The aerial parts of C. sicula and C. libanotis were collected in Southern Italy. Plants were identified by Filomena Conforti, and a voucher specimen has been deposited at the Herbarium of natural History Museum of Calabria. The air-dried aerial parts of C. sicula and C. libanotis were extracted with methanol through maceration (72 h × 3 times, plant to solvent ratio 1:10 g/mL) and also by means of Naviglio extractor ® (Atlas Filtri SrL, Limena, PD, Italy) (plant:solvent ratio 1:10 g/mL × 2 cycles). Obtained total extracts (CSM and CLM for samples obtained with maceration; CSN and CLN for samples obtained with Naviglio) were concentrated under reduced pressure at 40 • C. The extracts were dried, weighed, and stored at 4 • C for experimental use.

Determination of Total Phenolic and Flavonoid Content
The total phenolic content was determined using the Folin-Ciocalteu method [28]. Briefly, 50 mg of the extracts were dissolved in 25.0 mL of an extraction solution (acetone: methanol: water: acetic acid, 40:40:20:0.1) and heated for 1 h at 60 • C. Then 200 µL of samples, 1.0 mL of Folin-Ciocalteu's reagent and 1.0 mL of sodium carbonate (7.5%) were mixed and absorbance was measured at 726 nm 2 h later. The total phenolics content was calculated from the calibration curve, and the results were expressed as mg of chlorogenic acid equivalents per g of dried plant material.
The total flavonoid content of extracts was determined by a colorimetric method based on the use of aluminum chloride [29]. One mL of each extract (2 mg/mL in 80% EtOH) was added to 1 mL of 2% AlCl 3 . The absorbance was measured at 430 nm 15 min later. Results were calculated from calibration curve of the standard quercetin, and data were reported as mg of standard equivalent per g of dry plant material.

GC-MS Analysis
The coumarins, terpenes, and fatty acids content was assessed by means of gas chromatography-mass spectrometry (GC-MS). Qualitative GC-MS analyses were carried out using a gas chromatograph (Hewlett-Packard 6890) equipped with an SE-30 capillary column coupled to a selective mass detector (Hewlett-Packard 5973). Helium was utilized as the carrier gas, and a programmed temperature from 60 to 280 • C (with a rate of 16 • /min) was used to realize the analyses. The injector and detector were set, respectively, at 250 • and 280 • C [30]. Compounds were identified by comparing the GC retention factors with those of standards and the mass spectra with the Wiley 138 library data.
Quantitative GC analyses were performed to determine the content of furanocoumarins xanthotoxin, bergapten, and isopimpinellin in the extracts. Bergapten (10-0.16 mg/mL in MeOH/CHCl 3 ) was used as the external standard, and the linear regression equation was y = 1E + 07x + 2E + 06 (R 2 = 0.9949), 1 µL of each tested extract was injected in the GC-MS system at a final concentration of 50 mg/mL. All the analyses were carried out in isothermal conditions at 235 • C, and GC peaks area was determined in triplicate.

Cellular Phototoxicity
The photocytotoxic activity was determined as earlier described [31]. Human melanoma cells (C32) were grown in RPMI-1640 medium supplemented with L-glutamine, penicillin/streptomycin, and fetal bovine serum (1%, 1%, and 10%, respectively). For the experiments, 100 µL of medium containing 3.8 × 10 4 cells were placed in each well of a 96-well tissue culture microtiter plate. After 24 h incubation, the medium was removed and 100 µL of samples (0.625-100 µg/mL), dissolved in MeOH and diluted with Hanks' Balanced Salt Solution (HBSS, pH 7.2), were added to each well. Plates were incubated at 37 • C for 30 min and then irradiated. UV irradiation was carried out at 365 nm with an HPW 125 Philips lamp. The spectral irradiance of the source was 0.3 mW cm −2 as measured by a radiometer equipped with a 365-CX sensor (Cole-Parmer Instrument Company, Niles, IL, USA). Cells were irradiated for 1 h at a dose of 1.08 J/cm 2 . To prevent contamination from other light sources, the assay was performed in a dark room. Then, the solution was replaced with new medium, and microtiter plates were incubated for further 48 h. Bergapten was used as positive control.

Immunoblotting Analysis
The melanoma cells (C32) were harvested and lysed in 500 µL of RIPA buffer for total protein extraction. A 10% SDS-polyacrylamide gel was used for the resolution of the proteins that were transferred to a nitrocellulose membrane. Then, the filter was probed with p21, BAX, PARP, and GAPDH antibodies (Santa Cruz, Biotechnology, Heidelberg, Germany). The antibody antigen complex was detected with a secondary antibody conjugated to horseradish peroxidase and revealed with the ECL System (Amersham Pharmacia Biotech, Cologno Monzese, Italy) [33].

Free Radical Scavenging Activity (FRSA) Assay
The hydrogen atom or electron-donation ability of the extracts was evaluated based on the bleaching of a purple solution of 1,1-diphenyl-2-picryl-hydrazyl (DPPH), as earlier described [34]. Briefly, 200 µL of each extract (5-1000 µg/mL) were added to 800 µL of a 0.1 mM solution of DPPH and incubated in the dark. The absorbance was measured at 517 nm 30 min later. Ascorbic acid was used as a positive control.

β-Carotene-Linoleate Bleaching Assay
Antioxidant activity was determined using the β-carotene bleaching method [35]. β-carotene (0.2 mg in 1 mL chloroform) was added to 0.02 mL of linoleic acid and 0.2 mL of 100% Tween 20. Chloroform was removed and 100 mL of water were added. Then, 5 mL of the emulsion were added to 0.2 mL of samples in methanol (concentrations ranging from 0.25 to 100 µg/mL) and placed in a water bath at 45 • C for 60 min. Propyl gallate was used as standard. The absorbance was measured at 470 nm at the initial time and after 30 and 60 min. The antioxidant activity was measured in terms of the successful prevention of β-carotene bleaching.

Statistical Analysis
Experiments involving cell lines were carried out in quadruplicate, while the other assays were run in triplicate. Results were expressed as means ± SEM. Data were checked for normality and homogeneity of variances (D'Agostino-Pearson's K2 test and Levene's test, respectively). Biological data were fitted through nonlinear regression in order to calculate the IC 50 values using GraphPad Prism Software (San Diego, CA, USA).
Statistical differences between treated groups and the control were estimated by one-way analysis of variance (ANOVA) followed by Dunnett's multiple comparison test, while the statistical significance of differences among treated group means were tested by Bonferroni post-hoc test (p ≤ 0.05).

Data Availability Statement:
The data presented in this study are available on request from the corresponding authors.

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