Optimizing ultrasound-assisted extraction process of anti-inflammatory ingredients from Launaea sarmentosa: A novel approach

2.1 Plant materials

The stem and leaves of L. sarmentosa (LSm) were gathered from Ben Tre province, Vietnam (9°56′53″N 106°30′51″E), in August 2022. The specimen was identified and deposited in the Laboratory of Medicinal Chemistry at the CTU Hi-tech Building, Can Tho University, Vietnam, under biospecimen code number (VNM00082022CTU). Samples were cleaned and then naturally dried to avoid direct sunlight until a consistent weight was achieved. Afterward, they were ground into a fine powder using a Waring two-speed laboratory blender (Cole-Parmer, USA) and stored in a polyethylene bag at −20°C for further extraction.

2.2 Chemicals and reagents

All chemicals and antibodies were purchased from Sigma-Aldrich (Merck, Germany), except for the Griess-Romijn Nitrite reagent, radioimmunoprecipitation assay (RIPA) buffer, and LPS-Lipopolysaccharide (FUJIFILM-Wako, Tokyo, Japan), TRIzol reagent (Invitrogen, CA, USA), bicinchoninic acid (BCA) assay, and SuperSignal West Atto Ultimate Sensitivity substrate (Thermo Fisher Scientific, USA). Bioassay was conducted using cell counting kit-8 (CCK-8) from Dojindo Molecular Technologies (Rockville, MD, USA), Qiagen RNeasy Kit from Qiagen (Hilden, Germany), and Transcriptor Universal cDNA Master Mix and FastStart Essential DNA Green Master Mix, both obtained from Roche (Mannheim, Germany).

2.3 NO radical-scavenging activity

The nitric oxide (NO^•) radical scavenging activity was evaluated according to the previous studies with slight modification [29,30]. Shortly, an equal volume of 5 mM sodium nitroprusside and various concentrations of LSm extract were mixed under exposure to light for 30 min. Then, 100 μL of the mixture was combined with 100 μL Griess-Romijn nitrite reagent. The color change from transparent to pink was attained after incubating without light at room temperature for 10 min. Absorbance at 540 nm was measured and compared to a reagent blank using a standard sodium nitrite curve (y = 0.0034x + 0.0093; R ² = 0.9992). The IC₅₀ value (half-maximal inhibitory concentration) was determined to assess the radical-scavenging efficiency, while curcumin (y = 30.779ln(x) − 39.421; R ² = 0.996) was compared as a positive control.

2.4 Ultrasound-assisted LSm extraction and single-factor experiments

Single-factor experiments were conducted to investigate the influence of various factors, including sample weight, solvent-to-solid ratio, solvent concentration, temperature, and extraction time, based on the efficiency of NO radical-scavenging activity. In brief, the dried powder of LSm was soaked in ethanol using an Elmasonic S100H ultrasonic water bath, operating at a frequency of 37 kHz and 150 W nominal power. To assess the impact of the sample weight, amounts ranging from 1 to 20 g were applied with the random criteria, including the solvent-to-solid ratio of 20 mL·g⁻¹, ethanol concentration of 100%, and extraction conducted at 30°C for 15 min. Following this, the effect of ethanol concentration (70–100%) was evaluated, keeping the initial weight from the previous step. Then, the optimal temperatures between 30°C and 80°C were investigated for 15 min based on the preceding conditions. Lastly, the effect of extraction time (ranging from 5 to 30 min) was assessed based on the results obtained from the previous condition. After each extraction, the supernatants were centrifuged, filtered, and then concentrated using a Rotavapor R300 (BUCHI, Flawil, Switzerland). The extract was then freeze-dried in an Infistek lyophilizer and kept at −20°C for subsequent analysis. The parameters determined from these single-factor experiments were subsequently applied in an RSM using a BDD [31,32].

2.5 BDD for UAE optimization

BBD was selected to evaluate a second-order polynomial model for UAE, focusing on optimizing NO scavenging activity (Design Expert Software v12.0.0; Stat-Ease Inc., Minneapolis, MN, USA). From the result of single-factor experiments, the response pattern was affected by three parameters, including A: the solvent-to-solid ratio (15–25 mL·g⁻¹), B: temperature (40 – 60°C), and C: extraction time (10–20 min); each one being investigated at three levels, labeled as −1, 0, and +1, respectively. The level of independent variables and experimental design are shown in Table 1. The model was expressed by Eq. 1 [33]:

where Y is the response value; β 0 is the constant effect; β 0 , β jj , and β ij , are the linear, quadratic, and the interaction effect coefficients, respectively, while X i and X j are the independent coded variables, k = 3.

2.6 Cell viability assay

RAW 264.7 macrophage cell line (ATCC – TIB-71) was cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% (v/v) fetal bovine serum (FBS), supplemented with reagents including 1% (v/v) penicillin–streptomycin, kept in a humidified incubator of 5% CO₂ at 37°C. Afterward, cells (a density of 2 × 10⁵ cells/well) were seeded in 96-well plates before reaching 80% cell density. Then, cells were divided into marked groups and pre-incubated with a concentration of LSm with 1 µg·mL⁻¹ LPS (lipopolysaccharide from Escherichia coli O55:B5) or without LPS for 24 h. After removing the media, each cell group was added to a mixture of 90 µL DMEM and 10 µL CCK-8. Finally, the cell cytotoxicity was assessed at 450 nm using a microplate spectrophotometer (Epoch, BioTek, USA) [34].

2.7 Quantification high-performance liquid chromatography (HPLC) analysis

A Prominence-I series HPLC system (LC-2030, Shimadzu, Tokyo, Japan), equipped with a C18 analytical column (5 µm, 250 mm × 4.6 mm, Shimadzu, Tokyo, Japan) and a diode array detector (Shimadzu SPD-M20 A, Tokyo, Japan), were applied to analyze the active compounds from LSm extract, following a previously described method with minor modifications [35]. Instrument control and data analysis were performed using LabSolution software (version 5.90). Specifically, the dried extract was dissolved in methanol to a 1 mg·mL⁻¹ concentration and, subsequently, filtered through a 0.45 μm nylon syringe filter. The sample injection volume was set to 10 μL, with a 1.0 mL·min⁻¹ flow rate at 25°C oven temperature. The mobile phase gradient was comprised of 2% acetic acid (solvent A; HPLC grade, ≥99.8%) and methanol (solvent B; HPLC grade, ≥99.8%). A gradient elution program was employed as follows: 12% of solvent B up to 10 min, subsequently changed to achieve 20%, 30%, 50%, 60%, and 70%, followed by reductions to 20% and 10% of solvent B from 20 to 80 min, with each stage maintained for 10 min. Peaks were identified by comparing the retention times and UV spectra of pure standards, including quercetin, rutin, esculetin, and luteolin-7-O-β-d-glucoside, as summarized in Table 4. Data were expressed as µg·g⁻¹ off dried extract with three replications.

2.8 NO detection

Cells were pre-incubated in 12-well plates with varying concentrations of LSm extract, followed by washing with phosphate-buffered saline (PBS), then stimulated with 1 µg·mL⁻¹ LPS (1 µg·mL⁻¹) for 24 h. Griess-Romijn nitrite reagent was used to observe the NO secretion, whereas N-nitro-l-arginine methyl ester (l-NAME; 100 µM) was indicated as a positive control [34]. Subsequently, an equal volume (100 µL) of supernatant medium and Griess reagent was mixed and incubated in the dark at room temperature for 10 min. Absorbance was then measured at 540 nm using a microplate spectrophotometer and quantified against a sodium nitrite standard curve.

2.9 Detection of an inflammatory marker

Cells were seeded in six-well plates and pretreated with the indicated concentration of LSm extract. Then, cells were washed with PBS twice and stimulated in the presence/absence of LPS for 18 h. TRIzol reagent and the Qiagen RNeasy Kit were used to extract and purify total RNA, following the manufacturer’s instructions, to ensure the RNA quality for downstream experiments. Finally, iNOS, TNF-α, and IL-6 mRNA expression was evaluated using quantitative reverse transcription polymerase chain reaction (qRT-PCR), following the protocol summarized in our previous study [34]. β-Actin was used for normalization to ensure accurate and reliable qRT-PCR results. The specific details of all primers utilized in the study are presented in Table S1.

2.10 Immunofluorescence staining

For this method, cells (a density of 2 × 10⁴ cells/well) were seeded in Nunc Lab-Tek chamber slides (Thermo Fisher Scientific, MA, USA). Then, the cell was treated with an indicated concentration of extract, either with or without LPS, for 30 min. The treatment followed established experimental protocols for assessing cellular response [36]. In sequence, rabbit anti-p65 NF-кB antibody (1:800) and Alexar-488 secondary antibody (1:1,000) were stained to detect the p65 expression, followed by 4′,6-diamidino-2-phenylindole nuclear staining for 20 min. Then, samples were observed under a fluorescence scanning microscope FV10i (Olympus, Tokyo, Japan). The results were analyzed using MetaMorph analysis software (Molecular Devices, Sunnyvale, CA, USA).

2.11 Western blotting

Cells were pretreated with an appropriate extract concentration for 24 h before LPS stimulation. Whole proteins extracted using RIPA buffer were normalized by the BCA assay. Following 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (v/v) separation, proteins were transferred to polyvinylidene difluoride membranes, blocked with OneStep Blocker (Bio-Helix, Taiwan) for 1 h, and incubated with anti-phospho-p65 primary antibodies. Protein bands were visualized using SuperSignal West Atto Ultimate Sensitivity substrate. Band intensities were quantified by ImageJ analysis software (National Institute of Health, Maryland, USA) and normalized to β-actin.

2.12 Green assessment impact

In this work, Ecochain Mobius (Ecochain Technologies B.V., Amsterdam, Netherlands), a platform designed to conduct LCA, examined the environmental impact of the extraction process [37]. The platform integrates Ecoinvent v3.9.1, a comprehensive database, compliant with ISO 14044 standards [38]. This integration allows for a structured and reliable approach to conducting LCA, ensuring accuracy and compliance in the environmental impact analysis. The Ecochain Mobius platform has been certified by the International Organization for Standardization (ISO), specifically ISO 9001:2015 for quality management and ISO 27001:2017 for information security management. In this work, the function unit was 1 g of initial raw materials for applying the LCA. The system boundary. The system boundary encompassed LSm pretreatment (i.e., electricity consumption), solvent and extraction process (Figure 1). Herein, the LCA gate-to-gate approach was employed to assess these environmental impacts of the extraction process at the lab scale. The relevant input variables for evaluating the extraction process included electrical energy consumption, solvent usage, and waste generation. From this, a streamlined LCA was carried out, covering 25 different environmental impact categories, to evaluate the broader ecological effects of the process. This suggests an overview of the method and points out the essential modifications that should be implemented in order to achieve sustainable development. The current study is focused on the UAE method, which has been extensively applied to enhance the yield of bioactive ingredients extracted from plant materials or agri-by-products [39]. Moreover, rather than using organic or toxic solvents, the UAE utilizes food-grade solvents like water or ethanol to disrupt the vegetal matrix [40]. Additionally, most studies confirmed that the UAE became the eco-friendliest option, enabling a substantial reduction in energy consumption (e.g., electricity consumption) and environmental impacts (e.g., climate change and human carcinogenic toxicity) compared to other methods [39,41,42]. Furthermore, as mentioned above, the FDA approves ethanol for specific uses in food products that meet the criteria of “Generally Recognized as Safe” (GRAS). This confirmed the safe solvent level in production, signifying low risk to health and the environment. Hence, the selected extraction technology and solvent were well-suited to minimize environmental impacts, aligning with the sustainable industrial development trend. As shown in Figure 1, the detailed flowchart of the extraction process was compiled and characterized, including materials input, energy consumption for extraction, and solvent. However, at the lab scale, plant materials’ by-product was disregarded due to their being produced as biowaste. The subsequent drying, transport, residue disposal, and sale processes were not considered. Due to the small volume of solvent used, the reuse was not considered in this work. Nevertheless, while beneficial, solvent reuse typically requires a regeneration or recovery process (addition of energy input, etc.), which should be a concern on a larger scale. Therefore, electricity and solvent consumption, the main factors, were analyzed using LCA with the Environmental Footprint 3.1 methodology (EF 3.1 Method), followed by a database from Ecoinvent v3.9.1 from the Ecochain Mobius platform. Another issue is that the moisture content and scale-up mass of raw materials for industrial applications significantly impact energy consumption. High moisture or larger mass samples often need drying for initial processing, typically at 40–55°C, which consumes high electricity [39]. However, in the present work, a naturally dried sample that avoids direct sunlight was applied, and it did not affect the LCA model. Regarding the impact allocation methods, in this work, Ecochain Mobius LCA software was used, and Mobius itself did not automate allocation. Nevertheless, the physical allocation (mass, volume, energy usage) was manually applied by dividing environmental burdens based on quantifiable physical properties. Besides, scalability is reasonable when processes are linear (energy/material usage per unit is constant), no threshold effects exist (e.g., same machinery, same efficiency at small or large scales), and market effects are ignored (price elasticity, supply chain shifts).

Figure 1

A schematic framework of anti-inflammatory ingredients extracted from L. sarmentosa using an ultrasound-assisted method (per function unit of 1 g initial raw material).

2.13 Statistical analysis

Data was evaluated by one-way analysis of variance (ANOVA) with GraphPad Prism Software v8.0.1 (GraphPad Software Inc., San Diego, CA, USA), represented as mean ± SD of triplicate (whereas p < 0.05 indicated a statistically significant difference).

3.1 Impact of individual factorst

3.1.1 Effect of sample weight

First, while maintaining all other factors constant, the weight of raw materials was varied across a range from 1 to 20 g to investigate its impact. The raw materials were derived from a blend of various components, particularly the stems and leaves of LSm. Therefore, it was crucial to consider the uniformity of the chemical composition in the extract concerning the effectiveness of its biological activity [43,44]. On the other hand, bioactivity is impacted by the inconsistency in the material’s composition throughout its mass, which may complicate the creation of accurate models for large-scale industrial applications [45]. As depicted in Figure 2a, the NO scavenging activity efficiency did not vary significantly, even as the sample weight increased. This indicated that the weight of raw material did not affect the anti-inflammatory activity, likely due to the uniformity of the main active ingredients. Thus, the sample weight was set at 1 g for subsequent investigations.

Figure 2

Effect of single factors on the NO˙ scavenging efficiency of LSm extract. (a) Sample weight impact; (b) solvent-to-solid ratio impact; (c) solvent concentration impact; (d) temperature impact; and (e) time impact. Values were expressed as mean ± SD (n = 3), compared using one-way ANOVA with Tukey test; n.s.: not statistically significant; **p < 0.01, ****p < 0.0001.

3.2 Optimization procedure and model fitting

According to the initial investigation of single factors, three influencing factors (i.e., the solvent-to-solid ratio from 5 to 30 mL·g⁻¹, temperature from 30°C to 80°C, and extraction time from 5 to 30 min) have directly impacted the extraction process based on NO removal efficiency. Therefore, to optimize the extraction process, the Box–Behnken matrix design with three corresponding independent variables (i.e., A, B, and C, respectively) and the results concerning the IC₅₀ value of NO removal efficiency were applied in this study, as presented in Table 1. The experimental design consisted of 17 experiments, with 5 central points that exhibited the highest NO scavenging activity (Table 2). All experiments were repeated with at least three replications and were in randomized order. Particularly, the highest efficiency was observed in condition number 17 (204.71 ± 12.67 µg·mL⁻¹), while condition number 6 exhibited the lowest activity, reaching 328.58 ± 3.00 (µg·mL⁻¹). The central point was consistently marked from conditions 13 to 17, with variables A (20 mL·g⁻¹), B (50°C), and C (15 min).

Response values were estimated from experimental results using polynomial regression via the least squares method [60]. The following second-order polynomial equation describes each response based on the IC₅₀ value in terms of coded factors as follows, Eq. 2:

Statistical analysis (ANOVA), according to Design Expert 12.0 software, was conducted to evaluate the optimization model (Table 3). The results indicated that all regression models were highly significant (p ≤ 0.01), and no evidence exhibited a lack of fit at the 5% significance level [61]. The regression equation formed a quadratic function with an F-value of 78.98, demonstrating a robust correlation with statistical variance analysis. The p-value was less than 0.001, implying a 0.01% probability of noise leading to the above significant F value. Consistent with previous studies, the model’s lack of fit (p = 0.1337) was not substantial in terms of statistics (p-value >0.05), indicating that model errors were negligible. The strong correlation coefficient (R ² = 0.948) also exceeded the commonly accepted 0.8 threshold. The response model demonstrated high reliability during evaluation, as indicated by a low coefficient of variation (CV) of 4.13%. This value is well below the acceptable threshold of 10%, confirming the stability of the model’s outputs. This observation proved the practical applicability of this model with high dependability [48,62].

Response surface analysis, holding other factors at their central values, revealed the interaction affecting the NO removal efficiency of LSm. Figure 3a and b indicates that the optimal NO scavenging activity, equivalent to the IC₅₀ minimal value, was achieved with the solvent-to-solid ratio of 19–21 mL·g⁻¹, approximately 15 min for extraction time. Similar trends were observed for other factors reached near their central field (Figure 3c–f). In comparison with F-values of the difference factor and surface curvature, the solvent-to-solid ratio exerted the greatest impact on the extraction rate, followed by extraction time and temperature.

Figure 3

Response surface to NO˙ scavenging activity of LSm. (a) and (b) Response surface plots and contour plots of solvent-to-solid ratio and temperature; (c) and (d) response surface plots and contour plots of solvent-to-solid ratio and time; and (e) and (f) response surface plots and contour plots of temperature and time.

Last but not least, Figure 4 depicts the optimal condition for the extraction procedure according to data analysis. The predicted IC₅₀ value of NO removal efficiency is suggested at 209.68 μg·mL⁻¹, followed by the optimum conditions such as the solvent-to-solid ratio of 20.81 mL·g⁻¹, approximately 15.72 min at 51.80°C, and 99.8% ethanol. Remarkably, the actual data was obtained at 206.56 ± 2.80 μg·mL⁻¹, which aligned closely with the predicted data and was lower than curcumin with an IC₅₀ value of 17.69 ± 0.32 μg·mL⁻¹. Besides, based on our current knowledge, there are no reports confirming the NO scavenging activity from LSm. Another similar information indicated that Launaea procumbens and Launaea taraxacifolia, the same genus Launaea, exerted stronger NO removal radical efficiency than LSm, reaching IC₅₀ values of 59.4  ±  2.42 and 123.2 ± 1.5 μg·mL⁻¹, respectively [63,64]. As previously reported, biological activity differs largely depending on specific characteristics of the soil, the prevailing climate, the temperature, the distribution area, and the biodiversity of plant species [65].

Figure 4

Predicted experimental value of the extraction process at optimized conditions.

3.3 Anti-inflammatory activity of LSm extract

3.3.1 Quantification of NO production

To confirm the anti-inflammatory activity of LSm through the optimization process, NO secretion was quantified under LPS-stimulated RAW264.7 macrophages. For this observation, the LSm extract was produced under optimal conditions (LSm-oc), and its part obtained from random conditions (LSm-rc) was chosen to conduct this comparison. First, an initial assessment of cell cytotoxicity was conducted to establish the concentration levels that ensure minimal toxicity and safety. As illustrated in Figure 5a, the cell viability in the presence of LSm-oc and LSm-rc was not significantly different compared to the control group. In contrast, since exposure to LPS, cell survival was reduced by over 5%, suggesting that LPS caused cell death (Figure 5b). Indeed, LPS-induced cell damage formed the phenomenon of macrophages, underscoring its impact on cell structural features [14]. This also indicated that the observation of apoptosis and pyroptosis through the morphology transformation implied abnormalities during infection occurred. Nevertheless, co-incubation with LSm-oc and LSm-rc at 5 and 10 g·mL⁻¹ concentrations considerably improved cell survival in LPS-stimulated. Therefore, this issue implies that the presence of active ingredients through the extraction process may protect cells against infections by LPS.

Figure 5

The effect of LSm extract on (a and b) cell viability and (c) NO production under LPS-stimulated RAW264.7 macrophages. One-way ANOVA with Sidak’s multiple comparison test was applied for statistical analysis. Data were expressed as the mean ± SD (n = 6); n.s., not significant; p < 0.05 vs control without LPS; ****p < 0.0001 vs LPS group. LSm-oc as LSm extract under optimum conditions; LSm-rc as LSm extract under random conditions (i.e., solvent-to-solid ratio of 25 mL·g⁻¹; extraction temperature of 60°C; extraction time of 15 min).

During inflammation, macrophages, multitasking immune cells, adapt to the immune response by activating signaling cascades and then releasing pro-inflammatory cytokines [66]. As a result, NO secretion is a crucial signaling, as mentioned above, recognized when the defense mechanisms are activated against infectious pathogens [67]. Here, NO production was remarkably increased in the LPS-incubated group compared to the control and l-NAME groups. Similarly, treating LSm-oc and LSm-rc significantly suppressed the NO secretion (Figure 5c). More interestingly, at the same concentration (100 μg·mL⁻¹), LSm-oc exhibited the highest activity compared to the LSm-rc group. This aimed to re-evaluate the effectiveness of LSm-oc with inflammatory activity. Therefore, this result validated the efficiency of the optimum extraction process, emphasizing the highest concentration of anti-inflammatory components in LSm-oc.

3.3.2 Quantitative RT-PCR

Regarding the approach to inhibiting NO synthesis, the level of the inducible form of nitric oxide synthase (iNOS) was assessed using qRT-PCR. Due to NO persistent secretion under inflammatory responses, the overexpression of iNOS was initially emphasized by its role in catalyzing the creation of NO production [68]. Thus, quantitative NO secretion and iNOS levels provide an indicator of inflammation’s severity and can assess pharmacological agent efficacy [14]. The mRNA expression of iNOS was up-regulated following LPS-activated RAW264.7 macrophages, whereas it was down-regulated upon pre-incubation with LSm-oc and LSm-rc incubation at 100 μg·mL⁻¹ (Figure 6a). Another key marker, tumor necrosis factor-alpha (TNF-α), is a cytokine that drives inflammatory processes by activating the transcription of genes associated with inflammation and promoting cell death through indirect mechanisms. Besides, interleukin-6 (IL-6) expressed later than TNF-α, functions as a cytokine with diverse roles in acute inflammation [14]. Similar to the iNOS inhibitory level, the presence of LSm-oc strongly suppressed the mRNA expression of both TNF-α and IL-6 compared to LSm-rc (Figure 6b and c). Consistent with previous results, LSm-oc demonstrated higher efficiency than LSm-rc. This evidence proves the inhibition efficiency of LSm-oc on NO-releasing and pro-inflammatory cytokines due to the optimum active components.

Figure 6

The pro-inflammatory and iNOS expression levels under LPS-stimulated RAW264.7 macrophages. (a) mRNA iNOS expression level; (b) mRNA TNF-α expression level; and (c) mRNA IL-6 expression level. One-way ANOVA with Sidak’s multiple comparison test was applied for statistical analysis. Data were expressed as the mean ± SD (n = 6); ^# p < 0.05 vs control without LPS; ****p < 0.0001 vs LPS group.

3.3.3 Immunostaining detection and Western blotting

To further investigate the impact of LSm-oc and LSm-rc on inflammatory responses, an immunofluorescent analysis of NF-кB p65 was conducted to clarify this mechanism. As mentioned above, LPS stimulates the innate immune system, thereby initiating a series of complex intracellular signaling cascades. Among these signaling pathways, NF-кB signaling exerts a central role in inflammation due to promoting the release of pro-inflammatory cytokines through IкB degradation and phosphorylation of p65. In there, NF-кB subunits are sequestered in the cytoplasmic form of p50/p65 heterodimers at standard conditions, whereas infectious conditions promote their liberation and subsequently translocate into the nucleus [69]. As depicted in Figure 7a, representative micrographs of fluorescent staining demonstrated that these extracts strongly inhibited the translocation of NF-кB p65 during LPS activation. This detail has clarified the greater influence of LSm-oc compared to LSm-rc, as evidenced by the reduced intensity of p65 translocation (Figure 7b).

Figure 7

Inhibition of LSm-oc and LSm-rc on NF-кB p65 signaling pathway during LPS-activated RAW264.7 macrophages. (a) Immunology fluorescein-staining observation of NF-кB p65 subunit translocation; (b) nucleus fluorescent intensity was analyzed using MetaMorph software (n = 30); and (c) expression of phosphorylated-p65 following 100 µg·mL⁻¹ of both extracts was analyzed by Western blot (n = 3). One-way ANOVA with Sidak’s multiple comparison test was applied for statistical analysis. Data were expressed as the mean ± SD; ^# p < 0.05 vs control without LPS; ****p < 0.0001 vs LPS group.

Furthermore, adding LSm-oc dramatically reduced the phosphorylation of p65 compared to LSm-rc, while the p-p65 protein level was significantly enhanced in LPS-stimulated macrophages (Figure 7c and d). Overall, LSm-oc strongly exerted its anti-inflammatory role by deactivating the NF-кB signaling pathway, consistent with our previous study [4]. Furthermore, this result confirmed the high presence of anti-inflammatory ingredients from the optimum extraction process.

3.4 Identification and quantification of active compounds

The presence of active compounds for both LSm-oc and LSm-rc was identified through HPLC-DAD chromatograms, as presented in Figure 8 and Figure S1. Polyphenols (i.e., luteolin-7-O-β-d-glucoside, rutin, quercetin) and coumarin (i.e., esculetin) representatives were analyzed and compared to highlight their distinguishing characteristics. These substances are popular due to their potential anti-inflammatory effects, leading to their widespread use as natural remedies [70–72]. The HPLC results revealed that LSm-oc exhibited a broader spectrum of bioactive compounds in its composition than LSm-rc (Figure 8). This indicated a more comprehensive recovery of bioactive compounds from the source material than the LSm-rc extraction. Specifically, all four compounds were identified in LSm-oc, while only luteolin-7-O-β-d-glucoside and rutin were detected in LSm-rc (Figure S1). The appearance of luteolin-7-O-β-d-glucoside and rutin in LSm-oc with amounts of 14.77 and 15.83 µg·mg⁻¹ dried extract, respectively, further works synergistically to enhance the extract’s properties against inflammation, exerting a vital role in its overall effectiveness (Table 4). Besides, quercetin, a common compound with diverse biological effects, may have limited individual contributions at low concentrations in LSm-oc. However, its synergistic effects with other bioactive compounds could contribute to the overall efficacy.

Figure 8

Representative HPLC-DAD profile of LSm optimized extract (LSm-oc). 1: esculetin; 2: luteolin-7-O-β-d-glucoside; 3: rutin; 4: quercetin.

Among the identified flavonoids, esculetin, which belongs to coumarin derivatives, appeared in LSm-oc at 2.32 µg·mg⁻¹ dried extract compared to LSm-rc (Table 4). The presence of esculetin may contribute to the diversity of anti-inflammatory components from LSm-oc, consistent with our recent study [73]. Besides, other anti-inflammatory ingredients, particularly lutein, succinic acid, and 2(4-hydroxyphenyl)acetic acid, not belonging to the polyphenol group, were isolated from L. sarmentosa extract in our recent studies [14,74]. Remarkably, lutein, as a carotenoid with an IC₅₀ value for both NO scavening activity and NO inhibition at 23.65 and 29.23 µg·mL⁻¹, respectively, served as a typical anti-inflammatory compound isolated from LSm.

In view of the findings above, this evidence offers a promising opportunity for a new approach to maximally extract anti-inflammatory agents derived from plant materials. Another issue to consider is that, despite the correlation between polyphenols or flavonoids and NO inhibitory/anti-inflammatory activity mentioned, several pieces of information indicate that NO inhibition can still be observed in plant species with low levels of their content [75,76]. In particular, good correlations between total phenolic/flavonoid contents and NO scavenging activity were obtained at 0.63 and 0.73, respectively. However, for example, Parrotia persica Mey leaves exert high phenolic/flavonoids as 139.2 mg gallic acid equivalent g⁻¹ extract and 28.7 mg quercetin equivalent g⁻¹ extract but lower NO removal activity (IC₅₀, 3.43 mg·mL⁻¹) compared to Alcea hyrcana Grossh leaves (IC₅₀, 0.457 mg·mL⁻¹) with phenolic/flavonoids at 14.7 and 28.3 mg gallic acid and quercetin equivalent g⁻¹ extract. Moreover, unfortunately, recent research implies that polyphenols, like any chemical substance, depend on the interaction, dose, and conditions, and still exert harmful effects on human health. It could possibly block iron uptake, inhibit digestive enzymes, interact with drugs, and impact hormonal balance [77]. Therefore, a strategy for developing targeted natural health products derived from herbs to replace mixed-ingredient products should be considered. Finally, these results reinforce the evidence supporting the aim of this study, emphasizing NO scavenging activity as a crucial factor for optimizing extraction.

3.5 Green assessment for the extraction process

Over the decades, rapid industrial development has posed an enormous challenge in terms of the global environment due to its severe impacts. The role of technical innovation and research activities in developing sustainable manufacturing industries addresses the balance between profitability and sustainability by adopting “greener” practices into industry processes [37]. As a result, several observations related to “eco-friendliness,” such as the consumption of less energy and solvents with biodegradable properties and/or replacement technologies that ensure efficient productivity and safety, should be considered throughout the entire production cycle. A range of environmental impact variables needs to be evaluated and promptly adjusted to assess the “really green” extraction process compared to conventional extraction techniques, especially in the current context of ongoing sustainable development efforts. For this, LCA is frequently employed to assess the relative greenness or overall sustainability of new technologies and materials. This quantitative sustainability metric is highly valuable for lab research, enabling optimization at manageable control or allowing carefully refined experimental parameters [42]. This approach enables process review and is helpful for initial small-scale environmental impact assessments (e.g., raw material use, energy, solvent, and/or emissions) to inform comprehensive, larger-scale upgrades, especially the extraction process.

The growing focus on plant-based alternatives to animal products, such as functional foods, raises critical environmental questions. As these industries expand, addressing their environmental impact and ensuring sustainable practices is crucial for long-term viability and sustainable industrial development [78]. Therefore, this study aimed to address the unexplored limitations in optimizing UAE and assessing its environmental impacts for obtaining anti-inflammatory components from L. sarmentosa. Additionally, the initial raw material per 1 g of L. sarmentosa was selected as the functional unit for LCA, as mentioned above. This depends on the purpose of the LCA, focusing on input materials when the quality of the output products has been controlled, especially those related to composition and bioactivity, as reported in prior research [79].

As illustrated in Table 5, LCA results suggested several impacts, including acidification, climate change group (CC), ecotoxicity group (EC), human toxicity (HT), ionizing radiation, and ozone depletion (OD). These impact categories were frequently analyzed due to their severity on ecosystems, human health, and global warming. Regarding acidification, the transfer from CO₂ to carbonic acid (H₂CO₃) formed, reaching 90% dissolved in seawater, quickly dissociates HCO₃ ⁻, CO₃ ⁻, and H ⁺ ions are released. An increase in H⁺ concentration leads to reduced pH (lower than 8.1), associated with a reduction in calcium carbonate solubility and harming marine life, especially calcifying organisms like corals and shellfish [80]. Moreover, soil acidification resulting from acid rain, as the transformation of SO₂ and NO _x composition, has significantly altered soil conditions, adversely affecting crops [81]. This indicates severe impacts on the environment that are a necessary concern through the control of acidification levels. LCA results showed that the 1.04 × 10⁻³ mol H⁺ equiv of acidification level from the extraction process was equivalent to 7.94 × 10⁻⁴ kg SO₂ equiv and 5.9 × 10⁻⁴ kg NO _x equiv, respectively. Notably, these levels were significantly lower than in a previous study at 0.006 kg SO₂ equiv and 0.03 kg NO _x equiv [42]. Another portion of CO₂ emissions in the atmosphere contributes to the acceleration of global warming. Hence, much more attention is paid to reducing CC, as expressed by quantifying CO₂ emissions. The CC value was obtained at 1.25 × 10⁻¹ kg CO₂ equiv, which was insignificant and 8.4-fold lower than similar criteria in flavonoid extraction from Ginkgo biloba leaves, reaching 1.056 kg CO₂ equiv [42]. Similarly, the freshwater and marine eutrophication (EU) at 1.07 × 10⁻⁵ kg P equiv and 1.35 × 10⁻⁴ kg N equiv were significantly lower than the values reported in this study. The EU typically attributed emissions exceeding environmental limits to freshwater or marine pollution [82]. In addition, using ethanol minimized harmful impacts on both ecosystems and human health, particularly the EC and HT, which have an approximate value of 2.64 × 10⁻¹ CTUe (comparative toxic units ecotoxicity) and 2.53 × 10⁻¹¹ CTUh (comparative toxic unit for human). Conversely, the CTUh metric can increase under the impact of organic solvents such as benzene, etc., potentially leading to cancer or serious illnesses [83]. Finally, the OD metric, as an essential value for assessing overall environmental impacts, reflects the extent of stratospheric ozone depletion due to human chemical emissions. Based on LCA evaluation, this metric with a reference unit kg CFC11 equiv was 7.42 × 10⁻¹⁰, while the allowable baseline level is 1.0 [84]. Overall, other impacts were relatively minor and did not pose any harmful effects on the environment. These results contributed to the initial assessment of the extraction process to reduce harmful environmental impacts and minimize unnecessary costs.