The Hs27 normal human dermal fibroblast cell line (ATCC 1634-CRL), the HCT116 human colorectal carcinoma cell line (ATCC CCL-247) and the PC3 human prostate cancer cell line (ATCC 11435-CRL) were purchased by American Type Culture Collection (Manassas, Va), while the BMel human melanoma cell line was a gift from Prof. P.G.Natali (“Regina Elena” Institute for Cancer Research, Rome, Italy). All cell types were grown in Dulbecco’s modified Eagle medium, supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 100 IU/mL penicillin and 100 μg/mL streptomycin, and maintained at 37°C in humidified atmosphere with 5% CO₂. Medium was replaced every 3 days and cells were detached and sub-cultured when approximately 90% confluence was reached. All cell culture materials were purchased by Euroclone (Milan, Italy).

Plant extract was provided by “Laboratori Erboristici Caira” (Villa Latina, Italy). Briefly, leaves of B. sempervirens were collected around Villa Latina in September. Fresh leaves were dried and dry leaves (100 g) were macerated in 200 mL of hydroalcoholic solution (50% ethanol) at room temperature and then clarified by filtration. The hydroalcoholic plant extract (50 mg dry weight/mL), not subjected to concentration procedures, was stored at room temperature in the dark and diluted in culture medium when used. The expert who identified the plant was Dr. Marco Leonardi, university researcher of Systematic Botany at the Environmental Sciences Section, Dept. of Life, Health and Environmental Sciences, University of L’Aquila, Italy, being the voucher specimen of the plant specimen deposited in his personal herbarium (Pharmacology Lab) with the acronym PC 001-20.

Trypan blue dye exclusion test (TBDET) and MTS assay [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium; CellTiter 96 Aqueous One Solution reagent, Promega, Madison, WI] were employed in determining viability of the studied cells whose proliferation was assessed following exposure for various times to different concentrations of BSHE. Two vitality tests were adopted since they may have some single limitations due to the manual execution (TBDET) and to the interference with substances present in the system and to an increased number of mitochondria leading to overestimate the number of cells (MTS) (Kok et al., 2007; Wang et al., 2010; McGowan et al., 2011).

For TBDET, cells (in 1.5 mL culture medium/well) were seeded into 6-well plates at approximately 10,000 cells/cm² for Hs27 and PC3 lines, and 5,000 cells/cm² for HCT116 and BMel lines. After 24 h, cells were exposed to scalar concentrations of BSHE, diluted in culture medium, ranging from 3.9 to 500 μg/mL. Cell counting was performed every 24 h by using a Neubauer hemocytometer and a Nikon Eclipse TS 100 inverted microscope equipped with a phase contrast objective. Cell viability was assessed by removing cells from the plates with 0.05% trypsin-0.02% EDTA solution and by combining 20 µL aliquots of this cell suspension with 20 µL of 4% trypan blue dye solution. The dye stained the injured cells but not the uninjured ones, while a diffuse cytoplasmic staining indicated cell death (Tolnai, 1975). Three wells for each BSHE concentration were counted and this procedure was repeated three times. Two controls were predisposed, the first consisting of cells not exposed to BSHE, the second consisting of cells exposed to vehicle alone (i.e., aliquots of hydroalcoholic solutions with a maximum ethanol concentration of 0.5%). No difference in cell proliferation and vitality resulted between the two controls so that only data from the first one were reported.

For MTS assay, cells (in 100 µL culture medium/well) were seeded in 96-well plates (1,500 cells/cm² for HS27 and PC3 lines, 1,000 cell//cm² for BMel and HCT 116 lines), maintained 24 h for attachment and then exposed to suitable concentrations of BSHE. Culture medium was replaced after 24, 48, and 72 h, and 20 µL of MTS reagent were added. After 60 or 120 min, absorbance was determined at 490 nm by a Microplate reader model 680 (Biorad). A blank experiment, detecting background absorbance of cell-free wells, was performed in parallel and the resulting values were subtracted from those detected in cell-containing wells to obtain net absorbance. Each condition was reproduced in quadruplicate and the experiment was repeated three times. The controls were as for TBDET.

For quantification of drug response, the number of viable cells (or surrogates such as net A_490nm using MTS assay) in the presence of BSHE divided by cell counts for untreated controls was fitted to a sigmoidal curve to compute the concentration of drug at which the cell count is half of the control (IC₅₀). However, this traditional parameter (such as AUC and E_max) is confounded by the different numbers of cell divisions taking place during an assay because of natural differences in proliferation rate, variations in growth conditions or changes in the duration of an experiment, thus varying dramatically independently from any changes in the underlying biology. Therefore, we used a recently proposed new method for parameterizing drug response, i.e., the normalized growth rate inhibition (GR), which is based on comparing growth rates in the presence and absence of drug. In this regard, GR at time t in the presence of drug at concentration c will be: G R c = 2 k c / k 0 − 1 where k(c) is the growth rate of drug-treated cells and k(0) is the growth rate of untreated (or vehicle-treated) control cells. In practice, growth rate can be evaluated by considering both number of cells at the beginning of the treatment as just prior to drug exposure (x ₀ ) and number of cells at the end of the assay in an untreated (or vehicle-treated) control well [x (0)] and in a drug-treated well [x(c)] (Hafner et al., 2016; 2017; Clark et al., 2017). The GR value is thus: G R c = 2 log 2 x c / x 0 log 2 x 0 / x 0 − 1

Therefore, the GR value represents the ratio between growth rates under treated and untreated conditions normalized to a single cell division. Such value relates directly to response phenotype indicating partial growth inhibition (when it lies between 0 and 1), complete cytostasis (when it equals 0) or cell death (when it lies between 0 and −1). Parameterization of GR dose-response curves yields, inter alia, GR₅₀, analogous to EC₅₀, representing the concentration at which GR(c) = 0.5 and the pertubagen has half of its maximal effect on cell growth, while being largely independent of cell division rate and assay duration. On the other hand, GR metrics appear to be more robust than traditional metrics in assessing the effects of drugs on cell signalling and growth. In the present study, cell number was determined in TBDET by direct counting or, in MTS assay, by using the net absorbance at 490 nm as surrogate for direct cell counting. Online GRcalculator tools ( http://www.grcalculator.org ) were employed for calculation, analysis and visualization of dose-response data using GR approach and for comparison of GR and traditional metrics (Clark et al., 2017).

Cells were seeded in T75 culture flasks at a density of l × 10⁴ cells/cm², maintained 24 h for attachment and then treated or not with BSHE, whose GR₅₀ concentrations were referred to the 48 h-exposure).

After 24 and 48 h of incubation, control and treated cells were harvested and lysed in RIPA buffer [containing 1 mM PMSF (phenyl-methyl-sulfonyl fluoride) and (1 μg/mL each) aprotinin, pepstatin A and chymostatin (SIGMA-Aldrich)] by passing cell suspensions through a syringe equipped with 15 G, 25 G and 30 G needles (10 times each). The soluble protein fraction was then collected after centrifugation at 13,000 g for 30 min at 4°C and total protein content was assessed by Pierce BCA Protein Assay Kit (Thermo Scientific). Samples were then diluted (1:1 v/v) in the sample buffer (125 mM Tris-HCl, 20% glycerol, 10% β-mercaptoethanol, 4% sodium dodecyl sulphate, bromophenol blue, pH 6.8); denatured samples (15 µg total proteins) were run through 12% polyacrylamide denaturing gels and proteins were transferred onto methanol-activated polyvinylidene fluoride (PVDF) sheets (Millipore) by wet electrophoretic transfer (Laemmli, 1970; Towbin et al., 1979). Non-specific binding sites were blocked with 5% non-fat dried milk in TBS-T (10 mM Tris-HCI, 150 mM NaCl and 0.05% Tween-20, pH 7.5) for 60 min. Afterwards, samples were incubated overnight at 4°C with different primary antibodies (dilution l:1,000 each) in 5% non-fat dried milk in TBS-T. Primary antibodies were a) from Cell Signaling Technology, Inc, Danvers, MA, United States, i.e., anti-p62/SQSTM1Ab (cat. 88588), anti-p21 Waf1/Cip1 Ab (cat. BK 2947T), anti-survivin Ab (cat. BK 2808T), anti-p27 Kip1 Ab (cat. BK3698T) and anti-beclin-1 Ab (cat. BK 3495T) b) from Immunological Sciences, Rome, Italy, i.e., anti-cyclin B1 Ab (cat. AB84359), anti-p53 Ab (cat. AB83528), anti-Bcl₂ Ab (cat. AB 82346), anti-BAX Ab (cat. 82329) and anti-cyclin D1 Ab (cat. AB82346) c) from Thermo Fisher Scientific Waltham, MA, United States, i.e., anti-cyclin E Ab (cat. MA5-14336) d) from Novus biologicals Littleton, CO, United States, i.e., anti-LC3B Ab (cat. NB100-2220) e) from Santa Cruz Biotechnology, Inc, Santa Cruz, CA, United States, i.e., anti-β-actin Ab (cat. sc47778).

Finally, PVDF sheets were washed three times with TBS-T (10 min each), incubated with appropriate horseradish peroxidase-conjugated secondary antibody for 2 h at room temperature (dilution 1:1,000), and rinsed 3 times (5 min each) with TBS-T and once with TBS (i.e., TBS-T without 0.05% Tween-20).

Images of the specific immune complexes were revealed, acquired and analyzed by using Enhanced Chemi-luminescent Substrate Kit (Euroclone, Milan, Italy), Alliance Q9 hardware (UVItec Limited, Cambridge, United Kingdom), and Total Lab TL120 software (TotalLab, Newcastle upon Tyne, United Kingdom). Densitometry values were normalized for β-actin and experiments were performed in triplicate. Results were given as protein expression relative to control (relative intensity).

Cells were seeded in coverslips dishes at a density of l × 10⁴ cells/cm², maintained 24 h for attachment and then treated or not with BSHE (whose GR50 concentrations were those referred to the 48 h-exposure). After 48 h-incubation, cells were washed twice with PBS, stained with 4 μg/mL AO for 10 min and washed again with PBS. The coverslips were overturned on microscope slides and immediately observed under fluorescence microscopy (Axio Imager A2 Zeiss). Images were acquired using a color camera (DFC320 Leica). AO is a metachromatic dye whose luminescence wavelength is strongly dependent on its concentration. In live cells, AO is accumulated by acidic vesicles yielding strong orange or red signals; it is also provided with affinity for nucleic acids yielding less intense green fluorescence (Pierzynska-Mach et al., 2014). The vesicle red fluorescence intensity was quantified by using Fiji-ImageJ image processing software (Schindelin et al., 2012). At least 50 cells per each experimental condition were analysed, while corrected total cell fluorescence (CTCF), of both BSHE-untreated and BSHE-treated cells, was obtained as follows: CTCF = Integrated Density - (Area of selected cell x Mean fluorescence of background readings). (https://theolb.readthedocs.io/en/latest/imaging/measuring-cell-fluorescence-using-imagej.html#measuring-cell-fluorescence-using-imagej).

Cells were seeded in coverslips dishes at a density of 1 × 10⁴ cells/cm², maintained 24 h for attachment and then treated with BSHE extract (whose GR₅₀ concentrations were those referred to the 48 h-exposure).

After 48 h-incubation, cells were washed twice with PBS, fixed with 4% paraformaldehyde and permeabilized with PBS containing 0.5% Triton X-100. Non-specific binding sites were blocked following a 30 min-exposure to 3% bovine serum albumin (BSA) in PBS-T (PBS and 0.05% Tween-20). Afterwards, cells were incubated with rabbit anti-LC3B antibody (dilution 1:200 in PBS and 1% BSA) at 4°C overnight. Cells were then washed three time with PBS-T and incubated with secondary Dy Light 488-conjugated anti-rabbit antibodies (dilution 1:1,000 in PBS and 1% BSA) in the dark at 4°C for 1 h and washed again with PBS-T.

Coverslips (overturned on microscope slides) were mounted with Pro Long Diamond Antifade containing DAPI (4′,6-Diamidino-2-Phenylindole, Dihydrochloride) (Thermo Fisher Scientific) and observed under fluorescence microscopy (Axio Imager A2 Zeiss). Images were acquired using a black and white camera (Leica, DFC350FX).

Kolmogorov-Smirnov test of goodness of fit (p-value >0.05) indicated no significant departure from normal distribution for the obtained data. Therefore, parametric tests were used for detecting differences at the 0.05 level of significance. Data were expressed as mean ± S.E.M. or S.D. of n independent experiments. Differences between mean values were estimated by the Student’s t-test. Post-hoc comparisons were performed using the Dunnett test of GraphPad Prism ver. 6.1 software (United States).

We first examined the effect of BSHE on cell proliferation under assessment of cell viability by either TBDET or MTS assay. As reported in Materials and Methods, the antiproliferative effect of BSHE in the studied cell lines was evaluated not only by IC50 (concentration at which the cell number is half the control) but also by a novel approach consisting in the evaluation of GR, i.e., normalized growth rate inhibition that takes into account cell growth rate in the presence and absence of drug treatment. GR allows to overcome the interference that the different growth speeds of different cell lines exert on traditional metrics (like IC₅₀, E_max and AUC) not taking into account the number of recurring cell divisions. Parameterization of GR data through the online tool GR calculator (see above) yields, inter alia, GR₅₀, GR_max and h_GR (Hill Slope) values (minimally dependent on cell division rate and assay duration) along with traditional metric parameters (such as relative cell vitality and IC₅₀) and concentration-response curves from which GR₅₀ and IC₅₀ were obtained (Figures 1A, B; Table 1). In this regard, such curves obtained with GR metrics or traditional metrics had a high coefficient of determination (R ² > 0.98).

BSHE inhibited proliferation of cells in a concentration- and time-dependent manner, but the studied cell lines exhibited different susceptibilities. Table 1 summarizes the GR₅₀ and IC₅₀ values calculated in these cell lines for BSHE at different treatment times under assessment of cell vitality with TBDET and MTS assays. As expected, GR₅₀ values were less dependent on BSHE exposure times than IC₅₀ ones. For example, in HS27 normal fibroblasts, the GR₅₀ value (MTS assay) was 106 μg/mL at 24 h and 71 μg/mL at 72 h of exposure, being the corresponding IC₅₀ values 260 and 86 μg/mL, respectively. The values of GR₅₀ and IC₅₀ obtained for the studied cell lines differed depending on the above vitality assays. However, the three tumor lines were more responsive, whatever the vitality assay, than the normal HS27 line as shown by values of GR₅₀ and IC₅₀ for all times of exposure to BSHE to which, among tumor cells, BMel cells were the most responsive. In particular, the GR₅₀ values at 48 h-exposure to BSHE (MTS assay) were 72, 48, 38, and 32 μg/mL for HS27, HCT 116, PC3 and BMel cells, respectively, while the corresponding IC₅₀ values were 111, 46, 73, and 46 μg/mL (Table 1). Subsequent experiments were carried out by exposing the same cells for 24 and 48 h to these GR₅₀ concentrations. At such concentrations, cells were 95% vital, being BSHE active mainly by inhibiting cell proliferation. It is to be pointed out that a 48 h-exposure to 125 μg/mL of BSHE caused detachment of all BMel and HCT 116 cells, which could be found only as fragments in the supernatant. On the other hand, HS27 and PC3 cells, when treated in the same way, remained viable in percentages of 47% and 33%, respectively, in comparison with the corresponding untreated cells, in the absence of quantizable cell fragments in the supernatant.

Cells, exposed for 48 h to BSHE (GR₅₀) and observed under inverted phase-contrast microscopy, showed vesicle accumulation in the cytoplasm, mainly around nuclei (Figure 2). Following AO staining, these vesicles emitted red fluorescence revealing their acidic nature (Figure 3). In this respect, AO is known to freely enter cells, in which it is trapped and accumulated in acidic vesicles where, in a lower pH environment, its molecules are protonated, thus becoming unable to cross the vesicle membrane. At high concentrations, AO molecules form stacks and emit red fluorescence, whereas nucleic acid-bound AO characteristically emits green fluorescence (Pierzynska-Mach et al., 2014). In our experimental conditions, the red fluorescence in BSHE-treated cells increased significantly as compared to controls. In particular, such fluorescence was 5.4, 4.5, 3.5, and 2.9 higher in HCT, BMel, HS27 and PC3 cells, respectively (Figure 4).

In order to assess whether the above vesicles were autophagosomes, we considered the microtubule-associated light chain-3 protein (LC3) that represents the most reliable marker protein for autophagy. LC3 is present in two forms: LC3I, distributed through the cytoplasm, and LC3II, phosphatidylethanolamine conjugate that is recruited in autophagosome membranes (Zhang et al., 2016). Immunofluorescent localization of LC3 in control cells showed that it was homogeneously spread within the cytoplasm, thus indicating the presence of the unprocessed form (LC3I). On the other hand, in all cells exposed to BSHE (GR₅₀) for 48 h there were many dots close to nuclei corresponding to the processed, autophagosome-specific isoform of LC3 (LC3II) (Figure 5).

LC3, as above reported, is the most widely used autophagosome marker to evaluate the autophagic flux. During autophagy, LC3I, the cytoplasmic form, is conjugated with phosphatidylethanolamine to form LC3II that is recruited in autophagosome membranes. Both LC3II/LC3I ratio and LC3II level increase with autophagosome formation (Klionsky et al., 2016).

Densitometric analyses of western blots revealed that, when treated with BSHE (GR₅₀) for 24 or 48 h, all cell lines showed statistically significant increases of LC3II protein level and LC3II/LC3I ratio, both representing a sensitive index of autophagy (Figure 6). There were no significant differences between the different cell lines nor between 24 and 48 h of BSHE exposure.

Sequestosome-1 or p62, an autophagic cargo protein, is another marker of autophagic flux. Autophagy induction involves, besides the increase of LC3II, a concurrent decrease of p62, which undergoes degradation during the autophagic process (Lao et al., 2014). In all cell lines treated for 24 or 48 h with BSHE, immunoblot analysis showed a statistically significant increase of p62 suggesting an inhibition of its autolysosome degradation with resulting autophagy unbalance (Klionsky et al., 2016) (Figure 6).

Also Beclin-1 is an important regulator of autophagy (Levine and Klionsky, 2004) being essential for formation and extension of the pre-autophagosomal structure. Immunoblot analysis showed that, in HS27 cells, Beclin-1 was upregulated following 24 h-treatment with BSHE, whereas its expression at 48 h did not differ from that of control. This finding is likely to be consistent with a role of Beclin-1 in early events of autophagy. However, in all other cell lines, no changes in the expression levels of Beclin-1 by BSHE were observed according to studies dealing with Beclin-1-independent mechanisms of autophagy in cancer cells (Smith et al., 2010; Athamneh et al., 2017; Pan et al., 2019) (Figure 6).

Progression through the cell cycle is regulated by cyclins and cyclin-dependent kinases (CDKs), being p21 (Wafl/Cip1) a CDK inhibitor that can induce G1 arrest. p21, known as a tumor suppressor, is a key mediator of p53-dependent cell cycle arrest after DNA damage that is transcriptionally controlled by p53 and p53-independent pathways (Gartel et al., 1996; Kreis et al., 2019). In the present study, immunoblot analysis showed, in comparison with the untreated cells, a statistically significant increase of p21 in HS27 fibroblasts and BMel melanoma cells exposed for 24 h or 48 h to BSHE. In HCT 116 cells, p21 increased following 24 h of treatment with BSHE without changes after 48 h. Moreover, no changes of p21 levels were observed in BSHE-exposed PC3 cells (Figure 7).

With regard to p53, it was expressed in all cell lines with the exception of PC3 ones according to the reported deletion of such protein in this cell line (Nguyen et al., 2005). However, treatment with BSHE did not modify p53 expression in the studied cells (Figure 7).

As shown in Figure 7, p27 (Kip l), another CDK inhibitor, did not vary significantly in all studied cell lines after treatment with BSHE.

When considering cyclins, we evaluated the effects of BSHE treatment on levels of cyclin D1, cyclin E and cyclin B1. Immunoblot analysis revealed that cyclin B1 decreased significantly in HCT116 cells after 24 h and 48 h of treatment and only after 48 h in BMel and PC3 cells. These data agreed with the BSHE-induced inhibition of proliferation. On the other hand, an unexpected but not significant increase of cyclin D1 by BSHE was found in all tested cells, while the expression of cyclin E did not vary significantly in all BSHE-treated cells (Figure 8).

We also evaluated the expression of survivin, an anti-apoptotic protein. It is highly expressed in most human cancers and during fetal development, while being present at low levels or completely absent in healthy cells and tissues. Its expression levels correlate with more aggressive disease and poor clinical outcome. Within cell cycle, survivin is highly expressed at G2/M phase and declines rapidly in G1 phase (Jaiswal et al., 2015). Immunoblot analysis revealed that HCT116, BMel and PC3 cell lines, when treated with BSHE (GR₅₀) for 24 h or 48 h, showed a significant decrease in survivin expression, whereas HS27 human fibroblasts did not express such protein. BAX and Bcl₂ (proapoptotic and antiapoptotic proteins, respectively) did not vary following BSHE treatment of the studied cells (Figure 9).