**3. Results**

#### *3.1. Cytotoxicity of the Extracts in Gastric Epithelial Cells*

The cytotoxicity of the extracts was assessed in the concentration range 0.1–20 μg/mL in both human gastric epithelial cells (GES-1 and AGS cells) by means of the MTT assay. After 6 h treatment, XA extract showed cytotoxic effects only at 20 μg/mL in both the cell cultures (data not shown), and this concentration was not used for assessing the biological activity of the extract.

#### *3.2. Hydroalcoholic Extracts Inhibit ROS Production in AGS and GES-1 Cells*

The extracts were investigated for their ability to block peroxyl radicals in a cell-free system (ORAC) assay, and to reduce ROS generation in AGS and GES-1 cells. XP and AM extracts were the most active in ORAC assay (around 9.5 and 8.5 μmol Trolox equivalents, respectively), followed by EG and AC extracts (around 6 μmol Trolox equivalents) (Figure 1A).

ROS generation in AGS and GES-1 cells stimulated with hydrogen peroxide (200 μM in AGS and 100 μM in GES-1 cells) was close to 2.5-fold and 3.5-fold compared to the unstimulated controls, respectively. Most of the extracts inhibited ROS production induced by H2O2.

**Figure 1.** Antioxidant activity of the extracts. Oxygen radical absorbance capacity (ORAC) assay (**A**). Data are expressed as ORAC value (μmol Trolox Equivalent/g of sample). Effect of the extracts on intracellular ROS production induced by hydrogen peroxide in human gastric adenocarcinoma (AGS) (**B**) and gastric epithelial (GES-1) (**C**) cells. Antioxidant activity is expressed as μmol Trolox equivalent. Data reported in panels B and C are expressed as percentage versus the stimulated control (grey bar), to which is arbitrarily assigned the value 100%. \* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001.

SZ, MM, TT, and AM exhibited a statistically significant inhibition in AGS cells (*p* < 0.05, Figure 1B), while in GES-1 cells the effects were more pronounced since all the extracts, except for ZL, showed a significant effect approximately halving the levels of ROS (Figure 1C).

#### *3.3. E*ff*ect of the Extracts on the TNF*α*-Induced NF-*κ*B-Driven Transcription in AGS and GES-1 Cells*

The eleven hydroalcoholic extracts were preliminary screened at 10 μg/mL for their ability to impair the TNFα-induced NF-κB driven transcription, in AGS and GES-1 cells. TNFα increased the NF-κB driven transcription around 10-folds compared to unstimulated cells, in both cell lines (Figure 2); XP, TT, DG, AM, and AC extracts significantly impaired the activation in both cell lines, while XA only in AGS cells. The effect was more pronounced in AGS than in GES-1 cells.

**Figure 2.** Effect of the extracts (10 μg/mL) on the NF-κB-driven transcription in human gastric adenocarcinoma (AGS) (**A**) and gastric epithelial (GES-1) (**B**) cells. Data are expressed as percentage versus the stimulated control, which is arbitrarily assigned the value 100% (grey bar). \* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001.

XA, XP, TT, DG, AM, and AC extracts were further investigated for their ability to inhibit the TNFα-induced NF-κB driven transcription in a concentration-dependent manner (Figures S1 and S2). The extracts were tested in the range 0–10 μg/mL, and the corresponding IC50 values were calculated. In AGS cells, all the extracts, except for XA, showed an IC50 below 10 μg/mL (the highest concentration tested) and DG was the most active (IC50 2.1 μg/mL). In GES-1 cells, the IC50s were higher compared to those obtained in AGS cells, and only TT and DG extracts showed IC50 values below 10 μg/mL (Table 1).



#### *3.4. Plant Extracts Inhibit TNF*α*-Induced IL-8 Release and Expression in AGS and GES-1 Cells*

IL-8 is a well-known NF-κB-dependent chemokine strictly involved in the inflammatory process at the gastric level [5,7]; for this reason, the selected extracts (XP, XA, TT, DG, AM, and AC) were assessed for their ability to reduce its release as well as its promoter activity induced by TNF α in AGS and GES-1 cells. All the extracts, at concentrations ranging between 0.1 and 10 μg/mL, were able to reduce IL-8 secretion with di fferent IC50s (Figures S3 and S4), whose values were lower in AGS cells compared to GES-1 cells, as previously observed on the NF-κB driven transcription. The extracts also inhibited the IL-8 promoter activity, in AGS and GES-1 cells (Figures S5 and S6), suggesting that the effects showed on IL-8 release could be due, at least in part, to the impairment of the promoter activity (Table 2).


**Table 2.** Half-maximal Inhibitory Concentration (IC50) (μg/mL) of the extracts on the TNFα-induced IL-8 release and expression in human gastric adenocarcinoma (AGS) and gastric epithelial (GES-1) cells.

Since IC50s on IL-8 release were not always comparable with those on promoter activity, we investigated the inhibitory mechanism at mRNA level; concentrations corresponding to the IC50 values obtained on IL-8 release by each extract were tested for this purpose.

In GES-1 cells, the extracts with an IC50 >10 μg/mL (XA and AM) were tested at 10 μg/mL. As shown in Figure 3, in AGS cells, the *IL-8* mRNA levels were around 40–50% lower with respect to stimulated control; however, this e ffect was not statistically significant. In GES-1 cells, XP, DG, and TT extracts were able to reduce the *IL-8* mRNA levels of around 60%, while the activity of AC extract was around 40% lower. These results confirmed that the activity on IL-8 release could be attributed to an effect at transcriptional level (Figure 3).

**Figure 3.** Effect of the extracts on *IL-8* mRNA levels in human gastric adenocarcinoma (AGS) (**A**) and gastric epithelial (GES-1) (**B**) cells. Data are expressed as fold changes versus stimulated control (grey bar). \* *p* < 0.05; \*\*\* *p* < 0.001.

#### *3.5. E*ff*ect of the Extracts on the TNF*α*-Induced IL-6 Release and Expression in AGS and GES-1 Cells*

IL-6 is a cytokine dependent on the NF-κB activation involved in *H. pylori*-induced gastric inflammation [30]. Thus, the possible ability of the extracts to counteract gastric inflammation through the decrease of IL-6 release and expression was investigated as well. In these studies, just GES-1 cells were considered since TNFα did not induce any detectable IL-6 secretion in AGS cells (data not shown). The extracts were tested on IL-6 release in the range 0–10 μg/mL (Figure S7), while the activity on *IL-6* gene expression was screened at 10 μg/mL. XP and DG extracts inhibited IL-6 release both with an IC50 of 3.5 μg/mL, whereas the effect of TT and AC was lower (4.9 and 5.1 μg/mL, respectively); XA and AM exhibited an IC50 higher than 10 μg/mL (Table 3).


**Table 3.** Half-maximal Inhibitory Concentration (IC50) (μg/mL) of the extracts on the IL-6 release in gastric epithelial (GES-1) cells.

The effect of DG extract (10 μg/mL) on *IL-6* gene expression was statistically significant, as displayed in Figure 4; XP, DG, and TT also reduced *IL-6* mRNA levels; however, this effect was not statistically significant. XA and AM extracts did not show any activity (Figure 4).

**Figure 4.** Effect of the extracts on IL-6 mRNA levels in gastric epithelial (GES-1) cells. Results are expressed as fold changes versus stimulated control (grey bar). \* *p* < 0.05; \*\*\* *p* < 0.001.

#### *3.6. Tetrapleura Tetraptera Extract Inhibits PTGS2 (COX-2) Gene Expression in AGS and GES-1 Cells*

The gene expression, induced by TNF<sup>α</sup>, of the two cyclooxygenases, *PTGS1* (*COX-1*) and *PTGS2* (*COX-2*), was investigated at different times (1, 3, and 6 h) in AGS and GES-1 cells. *PTGS1* (*COX-1*) expression, as expected, was not induced by the stimulus in both cell lines; *PTGS2* (*COX-2*) mRNA was considerably increased in GES-1 cells at 6 h, while a mild increased after 1 h stimulation was observed in AGS cells (data not shown). The activity of the extracts on the gene expression of the two isoforms of cyclooxygenase was evaluated in GES-1 cells after 6 h treatment, in basal and inflammatory conditions. The *PTGS2* (*COX-2*) mRNA level of the cells treated with TT extract and TNFα was the only condition non-statistically different with respect to unstimulated control (Figure 5), suggesting the ability of this extract to reduce TNFα-induced *PTGS2* (*COX-2*) gene expression; the effect was exerted without affecting the gene expression of the *PTGS1* (*COX-1*) isoform. TT extract was also inactive on *PTGS2 (COX-2*) and *PTGS1* (*COX-1*) mRNA levels in the basal conditions (Figure 5). Interestingly, TT extract did not affect the enzymatic activity of PTGS1 (COX-1) and PTGS2 (COX-2) isoforms in a cell-free system (data not shown).

**Figure 5.** Effect of the extracts on prostaglandin endoperoxide synthase 2 *PTGS2* (*COX-2*) and prostaglandin endoperoxide synthase 1 *PTGS1* (*COX-1*) mRNA levels in the presence (**A**,**B**) or in the absence (**C**,**D**) of TNF<sup>α</sup>, in gastric epithelial (GES-1) cells. Data are expressed as fold changes versus control (grey bar). \* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001.
