The total polyphenol content of extracts was determined with Folin-Ciocalteu reagent with a modified version of the protocol described by Papoutsis et al. [28]. We placed 1 mL dry extract (1 mg/mL ethanol) in a 6 mL tube, followed by 0.5 mL of 10% (w/v) Folin-Ciocalteu reagent. The tube was gently shaken for 5 minutes, and we then added 1.5 mL of 2% sodium carbonate solution. The mixture was stirred manually and incubated in the dark at room temperature for 30 minutes. The reaction was stopped by immersing the tubes in a cold bath for 20 minutes. A blank was prepared in a similar manner, but with the extract replaced with ethanol. Absorbance was measured at 760 nm with a UVD 1800 spectrophotometer (Shimadzu, Kyoto, Japan). Gallic acid was used as a standard for the preparation of a calibration curve, with concentrations between 0.01 and 0.20 g/L. The solutions of gallic acid at different concentrations were mixed with Folin-Ciocalteu reagent and subjected to the protocol described for aliquots of the extracts. The results are expressed in mg gallic acid equivalent (GAE) per gram of extract (mg GAE/g extract).

The flavonoid content of extracts was determined according to a modified version of the aluminum chloride method described by Quettier-Deleu et al. [29]. We placed 1 mL of diluted extract (1 mg/mL in ethanol) or 1 mL of a standard solution of rutin in a flask, to which 1 mL of a 2% solution of aluminum chloride in methanol was added. The contents of the flask were mixed and incubated in the dark for 1 hour at room temperature. The absorbance of samples at 415 nm was measured with a UVD 1800 spectrophotometer (Shimadzu, Kyoto, Japan). We used 1 mL of solvent (ethanol) with 1 mL of 2% aluminum chloride in methanol as the blank for sample reading. A calibration curve was prepared with rutin as the standard, at concentrations of 0.005 to 0.090 g/L. Aliquots of rutin solution were mixed with aluminum chloride solution, and the protocol described above was followed. Flavonoid concentration is expressed in milligrams of rutin equivalent (RE) per gram of extract (mg RE/g extract).

A UV 1800 spectrophotometer (Shimadzu, Kyoto, Japan) was used to estimate total carotenoid content by measuring absorbance at 446 nm in a quartz cell with a path length of 1 cm. A wavelength of 471 nm was used for lycopene determinations, and a wavelength of 400 nm was used for $\beta$-carotene determinations. Dried extract (0.1 g) was dissolved in 50 mL ethanol and ultrasonicated for 1 minute. Two calibration curves were plotted from the absorbances measured at 400 and 471 nm for the solutions prepared with concentrations of 0.027 to 0.150 g/mL $\beta$-carotene and lycopene in ethanol. Moreover, total carotenoid content were determined by measuring absorbance at 446 nm and is expressed in mg of $\beta$-carotene equivalent per gram of extract.

The antioxidant activity of the extracts was assessed by determining their free radical scavenging activity against 2,2-diphenyl-1-picryldrazyl (DPPH), as described by Brand-Williams [30]. We dissolved 5 mg DPPH in 100 mL ethanol to obtain an absorbance of DPPH at 516 nm between 0.9 and 1.1. Extracts were diluted in ethanol to obtain concentrations of 500 $\mu$g/mL. We added a 150 $\mu$L aliquot of each extract and 150 $\mu$L of 50 $\mu$g/mL DPPH in ethanol to the wells of a 96-well microplate. The plate was kept in the dark for 40 minutes, with mixing by gentle rotation every 5 minutes. Absorbance was read at 516 nm with a spectrophotometer (Spectrostar, BMG Labtech, Ortenberg, Germany). Percent inhibition was calculated as follows:

(2)$$\mathrm{Inhibition}(\%)=100\times (1-\frac{A}{A_0})$$

where A is the absorption in the presence of the extract, and A${}_0$ is the absorption of the solution obtained by mixing 150 $\mu$L of DPPH solution with 150 $\mu$L of ethanol.

The ability of extracts to inhibit $\alpha$-glucosidase activity was estimated in a modified version of the assay described by Oueslati et al. [31]. We used p-nitrophenyl-$\alpha$-D-glucopyranoside (PNPG) as the substrate for this experiment. Various solutions of the extracts at different concentrations were prepared in dimethyl sulfoxide (DMSO). In brief, 50 $\mu$L of the extract tested (250 $\mu$g/mL in DMSO) was mixed with 100 $\mu$L of the enzyme solution (1 U/mL $\alpha$-glucosidase in phosphate buffer (pH 6.9)) and 50 $\mu$L phosphate buffer (0.1 M, pH = 6.9). The mixture was incubated for 10 minutes at 25 °C, and 50 $\mu$L PNPG solution (5 mM in phosphate buffer, pH = 6.9) was added to the mixture (1% DMSO), which was then incubated for 5 minutes at 25 °C. Absorbance at 405 nm was measured in 96-well microplates with a Thermo Fisher Scientific Multiskan${}^{\text{TM}}$ GO Microplate Spectrophotometer (Vantaa, Finland). Measurements were performed three times. Acarbose, an inhibitor of $\alpha$-glucosidase, was used as a positive control. The percent inhibition (%) reflects the ability of the extract to inhibit $\alpha$-glucosidase activity and was calculated as follows:

(3)$$\begin{array}{c}\hfill \alpha -\text{glucosidase inhibition activity}(\%)=100\times \frac{{\mathrm{A}}_{\text{control}}-{\mathrm{A}}_{\text{sample}}}{{\mathrm{A}}_{\text{control}}}\hfill \end{array}$$

where A${}_{\text{control}}$ is the absorbance (405 nm) of the control prepared with 100 $\mu$L potassium phosphate buffer (replacing $\alpha$-glucosidase) and A${}_{\text{sample}}$ is the absorbance (405 nm) in the presence of the extract.

Inhibitory activity is expressed as the IC${}_{50}$ ($\mu$g/mL), corresponding to the concentration required to inhibit the enzymatic hydrolysis (by $\alpha$-glucosidase) of the p-nitrophenyl-$\alpha$-D-glucopyranoside substrate by 50%.

The ability of extracts to inhibit $\alpha$-amylase activity was estimated by a modified version of the DNS method described by Saidi et al. [32]. Each extract was dissolved in a minimum amount of DMSO and diluted in sodium phosphate buffer (0.1 M, pH 6.9). The concentration of DMSO in the mixture did not exceed 1%. Diluted extract (50 $\mu$L of a 250 $\mu$g/mL dilution) was added to 50 $\mu$L of $\alpha$-amylase solution (3 mg/5 mL of the same buffer). Samples were incubated for 15 minutes at 25 °C, and 100 $\mu$L starch solution (1% in buffer) was then added. The mixture was incubated at 25 °C for 3 minutes and was then mixed with 100 $\mu$L 3,5-dinitrosalicylic acid (DNS) solution. The reaction tube was placed in a boiling water bath for 10 minutes and was then cooled to room temperature, before the addition of 1 mL sodium phosphate buffer. Absorbance was measured at 540 nm in triplicate in 96-well plates with a Thermo Fisher Scientific Multiskan${}^{\text{TM}}$ GO Microplate Spectrophotometer (Vantaa, Finland). Acarbose (1.3 mg/mL), an inhibitor of $\alpha$-amylase, was used as a positive control. The percent inhibition (%) of $\alpha$-amylase was calculated as follows:

(4)$$\begin{array}{c}\hfill \alpha -\text{amylase inhibition activity}(\%)=100\times \frac{{\mathrm{A}}_{\text{control}}-{\mathrm{A}}_{\text{sample}}}{{\mathrm{A}}_{\text{control}}}\hfill \end{array}$$

where A${}_{\text{control}}$ is the absorbance (540 nm) of the control prepared with 100 $\mu$L potassium phosphate buffer in place of $\alpha$-amylase, A${}_{\text{sample}}$ is the absorbance (540 nm) in the presence of the extract. The concentration of the compound providing 50% inhibition (IC${}_{50}$ in $\mu$g/mL) was calculated by plotting percent inhibition against sample concentrations.

The anti-inflammatory activity of extracts was determined with 15-lipoxygenase (15-LOX), as described by Khelifi et al. [33], but with some modifications. In brief, 20 $\mu$L extract (250 $\mu$g/mL) was mixed with 150 $\mu$L sodium phosphate buffer (pH = 7.4), 20 $\mu$L 15-lipoxygenase and 60 $\mu$L linoleic acid (3.5 mM) as a substrate. For the blank, the substrate was replaced with 60 $\mu$L buffer. Then, the mixture was incubated at 25 °C for 10 minutes and absorbance at 234 nm was measured in 96-well microplates with a Thermo Fisher Scientific Multiskan${}^{\text{TM}}$ GO Microplate Spectrophotometer (Vantaa, Finland). Measurements were performed three times. Nordihydroguaiaretic acid (NDGA) at a concentration of 0.5 mg/mL was used as a positive control. The following equation was used to calculate the percent inhibition:

(5)$$I(\%)=100\times \frac{A_{\text{control}}-A_{\text{sample}}}{A_{\text{control}}}$$

where A${}_{\text{control}}$ is the absorbance (234 nm) of the reaction mixture when the extract is replaced with potassium phosphate buffer and A${}_{\text{sample}}$ is the absorbance (234 nm) in the presence of the extract.

This test is based on the reduction of MTT (a yellow water-soluble tetrazolium salt) by the mitochondrial dehydrogenase of intact cells to generate purple formazan. The anti-proliferation activity of the extracts against HCT-116 cells (a human colon cancer cell line) was estimated with a modified version of the protocol described by Dawra et al. [34]. The cells were purchased from ATCC, and cultured in RPMI 1640 medium at 37 °C in a humidified incubator (Thermo Fischer Scientific, Illkirch, France) under an atmosphere containing 5% CO${}_2$. The HCT-116 cells were dispensed in a 96-well microplate, with 3 $\times$ 10${}^4$ cells/well, in 100 $\mu$L of appropriate culture medium, and were incubated for 24 hours at 37 °C. The test compounds were suspended in DMSO and diluted to obtain solutions with a DMSO concentration below 1%. The microplate was incubated at 37 °C for 48 hours. The supernatant was then removed, 20 $\mu$L of MTT solution was added to each well and the plates were incubated for a further 40 minutes. The MTT reagent was removed and 80 $\mu$L DMSO was added to dissolve the formazan crystals. Absorbance was measured at 605 nm with a microplate reader (Multiskan${}^{\text{TM}}$ GO, F1-01620, Thermo Fisher Scientific, Vantaa, Finland). Measurements were performed three times. Tamoxifen at a concentration of 100 $\mu$M was used as a positive control. The percent inhibition of cellular activity was calculated as follows:

(6)$$\mathrm{Inhibition}(\%)=100\times \frac{A_{\text{blank}}-A_{\text{sample}}}{A_{\text{blank}}}$$

A${}_{\text{blank}}$ is the absorbance of the cell suspension without the extract.

For flow cytometry, the A549 cells were cultured in 96-well plates and infected with ZIKV${}^{\text{GFP}}$, in the presence or absence of extracts, for 24 hours at a multiplicity of infection (MOI) of 1. The cells were then treated with trypsin and fixed by incubation with 3.7% PFA for 15 minutes. The cells were then subjected to a flow cytometry analysis in a Cytoflex machine (Beckman, Indianapolis, Indiana, USA). The results were analyzed with cytexpert software

We used the $\alpha$-glucosidase and $\alpha$-amylase enzymes to assess antidiabetic activity. Acarbose, an $\alpha$-glucosidase inhibitor prescribed for the treatment of diabetes, was used as a positive control. It has an IC${}_{50}$ of 48.8 $\mu$g/mL for these two enzymes. The percent inhibition of $\alpha$-glucosidase and $\alpha$-amylase achieved with the various extracts is presented in Figs. 9,10. These extracts clearly outperformed the positive control (with IC${}_{50}$ = 15.1 and 8.7 $\mu$g/mL for anti $\alpha$-glucosidase and $\alpha$-amylase activities, respectively), probably due to synergy between the various molecules present in the extracts.

Fig. 9.

Anti-$\alpha$-glucosidase activity of TP extracts (tested at 250 $\mu$g/mL) obtained by maceration at 20 °C for 24 hours or with SC-CO${}_{\mathrm{2}}$ at 40 °C and 200 bars. CPT, chemical pretreatment; EPT1, enzymatic pretreatment with cellulase; EPT2, enzymatic pretreatment with pectinase; EtAc, ethyl acetate. Blue histograms correspond to extracts obtained with pretreatment; Red histograms correspond to extracts obtained without pretreatment, columns with at the same letters are not significantly different (p $>$ 0.05).

Fig. 10.

Anti-$\alpha$-amylase activity of TP extracts (tested at 250 $\mu$g/mL) obtained by maceration at 20 °C for 24 hours or with SC-CO${}_{\mathrm{2}}$ at 40 °C and 200 bars. CPT, chemical pretreatment; EPT1, enzymatic pretreatment with cellulase; EPT2, enzymatic pretreatment with pectinase; EtAc, ethyl acetate. Blue histograms correspond to extracts obtained with pretreatment. Red histograms correspond to extracts obtained without pretreatment), columns with the same letters are not significantly different (p $>$ 0.05).

The extract obtained with the ethanol:water (50:50) mixture yielded the lowest IC${}_{50}$ values (5.9 and 4.9 $\mu$g/mL). This extract contains flavones (molecules 21, 25 and 26) and phenolic acids (molecules 1, 5 and 7), which are known to have antidiabetic activity. The contribution of 5,7-dihydroxy-4-propyl coumarin (molecule 20) to several biological activities has been investigated with African mustard extracts [42]. In their review, Khalid Al-Ishaq et al. [43] described the antidiabetic activity of flavonoids, including flavones, together with their mechanism of action. Gheonea et al. [17] also reported an inhibitory effect (79.9%) against $\alpha$-amylase for lycopene extracted from tomato skin at a concentration 8.5 mg/mL.

Nordihydroguaiaretic acid (NDGA) was used as a positive control in the 15-lipoxygenase (15-LOX) inhibition assays, with an IC${}_{50}$ of 4 $\mu$g/mL. The IC${}_{50}$ values of the extracts ranged from 9.3 to 12.3 $\mu$g/mL (Fig. 11), indicating clear activity against 15-lipoxygenase whatever the solvent used for extraction. The extract obtained with ethyl acetate after chemical pretreatment had one of the lowest IC${}_{50}$ values recorded (i.e., 9.3). Kaempferol (molecule 23), cinnamyl-3,4-dihydroxy-$\alpha$-cyanocinnamate (molecule 24) and rutin hydrate (molecule 14) are probably involved in this bioactivity, as all these molecules have already been reported to have anti-inflammatory properties. Devi et al. [44] described the mechanisms of action of kaempferol, and Pergola et al. [45] showed that molecule 24 is a potent inhibitor of 5-LO that is effective in vivo.

Fig. 11.

Anti-inflammatory activity of TP extracts (tested at 250 $\mu$g/mL)against the 15-lipoxygenase enzyme, after maceration at 20 °C for 24 hours or with SC-CO${}_{\mathrm{2}}$ at 40 °C and 200 bars. CPT, chemical pretreatment; EPT1, enzymatic pretreatment with cellulase; EPT2, enzymatic pretreatment with pectinase; EtAc, ethyl acetate. Blue histograms correspond to extracts obtained with pretreatment. Red histograms correspond to extracts obtained without pretreatment, columns with the same letters are not significantly different (p $>$ 0.05).

Based on recent data from in vitro studies, Maleki et al. [46] highlighted the possibility that different subclasses of flavonoids (including rutin hydrate) might modulate different stages of inflammation, thereby playing key roles in diseases such as diabetes, asthma, cardiovascular diseases and cancer. Finally, the phenolic acids present in the ethanol-based extracts have also been described as potent inhibitors of 15-lipoxygenase [47, 48].

The cytotoxicity of extracts (50 $\mu$g/mL) was evaluated against a variant HCT 116 tumor cell line (Fig. 12). Tamoxifene (100 $\mu$M), used as a positive control, gave 93% growth inhibition. The SC-CO${}_2$ extract had a clear anticancer effect, yielding 64.5% inhibition. Four other extracts in ethanol had a moderate anticancer effect, with inhibition rates between 31% and 37%. These four extracts had high total polyphenol contents (Fig. 4), with high levels of chlorogenic acid (molecule 5) and several flavones in common. Their activities may be linked to several molecules previously reported to have anticancer properties. Trihydroxethylrutin (molecule 17), used in a pure state, has been shown to be an effective inhibitor of the growth of HCT-116 cells [33]. Chlorogenic acid extracted from decaffeinated coffee beans was found to be cytotoxic in the MTT test on HCT 116 cancer cells, with an IC${}_{50}$ of 0.84 mg/mL [49]. Lycopene (molecule 29) increased the inhibition of HCT116 cell growth by 58.3% when used at a concentration of 100 $\mu$g/mL [50]. Moreover, 5,6,7-trihydroxyflavone (i.e., baicalein, molecule 21) extracted from the fruits of Oroxylum indicum, widely used in traditional Chinese herbal medicine for its anti-inflammatory properties, inhibited the growth of HL 60 cells by 50% at concentrations of 25 to 30 mM [51]. Flavones from tomato skin have also been reported to have inhibitory activity against SCC 9 human oral cancer cells [52, 53].

Fig. 12.

Cytotoxic activity of TP extracts (tested at 50 $\mu$g/mL) against HCT 116 cancer cell lines, after maceration at 20 °C for 24 hours or with SC-CO${}_{\mathrm{2}}$ at 40 °C and 200 bars. CPT, chemical pretreatment; EPT1, enzymatic pretreatment with cellulase; EPT2, enzymatic pretreatment with pectinase; EtAc, ethyl acetate. Blue histograms correspond to extracts obtained with pretreatment. Red histograms correspond to extracts obtained without pretreatment, columns with the same letters are not significantly different (p $>$ 0.05).

The potential antiviral activities of the six extracts prepared without TP pretreatment were investigated against Zika virus (ZIKV), a mosquito-borne enveloped arbovirus from the Flavivirus genus that constitutes a major health burden, due to the diverse human cells it can infect [54]. For this assay, human A549 cells were infected with ZIKV carrying a reporter gene encoding green fluorescent protein (GFP) [55] for 24 hours in the presence or absence of various nontoxic concentrations of the extracts. The concentrations inhibiting infection by 50% were determined. The results of antiviral activity assays are presented in Fig. 13.

Fig. 13.

Extracts display antiviral activity against ZIKV. (A) A549 cells were incubated with two-fold serial dilutions (400 to 6.25 $\mu$g/mL) of plant extracts for 24 hours. Cellular mitochondrial activity was evaluated in an MTT assay. (B) A549 cells were infected with ZIKV${}^{\text{GFP}}$ at an MOI of 1 in the presence of different concentrations of extracts. Flow cytometry analysis of GFP fluorescence was performed 24 hours post-infection. (C) Schematic representation of time-of-drug-addition assays used to characterize the antiviral activity of extracts 1, 3 and 6 (100 $\mu$g/mL). (D) Flow cytometry analysis of GFP expression in A549 cells infected with ZIKV${}^{\text{GFP}}$ for 24 hours at an MOI of 1 under the various experimental conditions shown in (C). The data shown are the means $\pm$ SD of four independent experiments performed in triplicate, and are expressed relative to vehicle-treated/infected cells. Statistical analysis consisted of one-way ANOVA followed by Dunnett’s test for multiple comparisons with p 0.05 considered significant.

Cellular mitochondrial activities were preserved below 20 $\mu$g/mL (Fig. 13A). Above this concentration, cell viability was conserved for extracts 1 (ethanol) and 4 (ethanol and water), with values of 80–90% at 200 $\mu$g/mL relative to vehicle-treated control cells. In antiviral activity assays, extracts 1, 3 and 6 (prepared with ethanol, ethanol:ethyl acetate and SC-CO${}_2$, respectively), inhibited ZIKV infection in a dose-dependent manner at concentrations above 10 $\mu$g/mL (Fig. 13B). At a concentration of 200 $\mu$g/mL these three extracts inhibit ZIKV infection in human cells up to 75%. Time-of-drug-addition assays (Fig. 13C) showed that the most active extracts (1, 3 and 6) impeded ZIKV entry into cells (Fig. 13D). Gaafar et al. [56] showed that an ethanol extract of tomato pomace (400 $\mu$g/$\mu$L) inhibited avian influenza A H5N1 virus by 63%. The epicatechin (molecule 6) detected in extracts 1 and 3 was described in the review by Behl et al. [57] as having antiviral activity against Zika virus. Baicalin (a trihydroxyflavone, molecule 20), which was also detected in ethanol extracts, has been reported to downregulate ZIKV replication [57, 58]. Isoquercitrin (quercetin-3-O-glucoside, molecule 16), which was detected in extracts 1, 2 and 3, has been identified as a potent inhibitor of ZIKV in various cell types of human origin, as this molecule prevent the initiation of host-cell infection [59]. Some flavonoids have been reported to have anti-Zika virus activities, and several possible mechanisms of action have been suggested [54, 60]. Finally, hamametannin (molecule 4 in extracts 2 and 3) has been described as a potential SARS-CoV-2 inhibitor, based on molecular docking data [61]. These classes of molecules may account for the observed antiviral activities of some TP extracts against ZIKV. They probably act in synergy, as the ethanolic fraction obtained after maceration in ethyl acetate, which would have concentrated the extract in polar molecules only (extract 5), had lower bioactivities.