Optimization of the Extraction of Bioactive Compounds from Cabernet Sauvignon Grape Pomace from Querétaro, Mexico, Using MSPD

Abstract

Red wine contains polyphenols which are extracted during the winemaking process. However, winemaking is not an extraction; therefore, the resulting byproducts still have a substantial polyphenol content. The aim of this study was to compare two methods for the extraction of phenolic compounds: maceration and matrix solid-phase dispersion (MSPD). Grape pomace (Vitis vinifera var. Cabernet Sauvignon) from a winery in Querétaro, Mexico was used. The optimal conditions for both methods were identified. Phenolic compounds and antioxidant activity were the response variables. A central composite design was used (Minitab 17) for the extraction by maceration. The maceration time (1, 4, 12, 20, and 24 h) and the ratio of ethanol (50 to 80%) acidified with 1% HCl were the two factors studied. For the MSPD extraction, sea sand was used as a dispersant, and a 2² factorial design was employed for the evaluation, with the elution volume and the ratio of sample/dispersant being the two factors analyzed. The optimal extraction method was MSPD with 96 mL (acidified ethanol) as the elution volume and a 1:2 ratio of sample/dispersant. Using these conditions, 14.01 ± 0.19 mMol TEAC/100 g db (Trolox equivalent of antioxidant capacity) of grape pomace was obtained, whereas the total phenolic content was 2836.73 ± 41.90 mg GAE/100 g db. These values are greater than those obtained by maceration. These conditions are close to those predicted by the model (analysis of variance (ANOVA) with a level of significance of 5% (p < 0.05) and a Tukey comparison test for determining significant differences in the comparison of results).

1. Introduction

In recent years, concerns about the generation of industrial waste have risen, boosting interest in sustainable production. Thus, technologies have been developed to manufacture high-value-added products from waste. Some of the industries that generate the most waste are related to the production and transformation of food [1]. In the present work, residues from the wine industry were used as raw materials because they are a natural source of antioxidants and have various compounds that are beneficial for health [2,3].

There are several studies on the antioxidant and health-promoting effects of secondary metabolites present in plants, including grapes (and wine). It has been reported that the content of phenolic compounds in grapes contributes to the inhibition of the oxidation of low-density lipoproteins, thus helping prevent cardiovascular diseases [4,5]. The edible part of the grape is the pericarp, which is composed of the epicarp (peel or skin), the mesocarp (pulp), and the endocarp (the tissue that surrounds the seeds and is indistinguishable from the pulp and stem) [6,7].

The Mexican wine industry includes producers of table grapes, raisins, concentrated grape juice, wine, and grape liqueurs (brandy) [8]. Wine is the beverage resulting from the complete or partial alcoholic fermentation of grape must in contact or not with pomace [9]. The must is the juice obtained by squeezing and/or draining and/or pressing the grapes [10]. The winemaking process comprises several operations, from harvest to clarification. At the time of harvest, the grape cluster is collected and taken to the winery. The grape harvest is carried out when the fruit has reached an adequate maturity, which is determined by measuring the concentration of sugars and acids [10], and is specific depending on the type of wine [11,12]. Once in the winery, the grapes are squeezed; this process involves breaking the skin so that the juice is easily released. The squeezing process favors the natural inoculation of the juice with the yeast attached to the skin, the aeration of the juice, and the subsequent maceration of the solid parts. For white wine vinification, only the must is vatted, whereas for red wine vinification, both the must and the solid parts are vatted in large tanks—preferably made of stainless steel—which allow for control of the fermentation temperature and are equipped with technical devices that facilitate the processes [13,14]. During fermentation, control of the aeration and temperature is crucial as these parameters modulate the yeast activity and allow for a uniform fermentation process. Once the desired composition of the wine is reached, it is transferred to other tanks, where it will slowly finish fermenting, and in the case of red vinification, the pomace will be separated [15,16].

During the winemaking process, waste is generated, most of which (80–85%) is organic. This waste is made up of the pomace (62%), lees (14%), stems (12%), and sludge (12%). The pomace is produced when grapes are pressed and is composed of the seeds and skin of the fruit, whereas the lees (the remains of yeast and solids) are generated during the wine clarification process. The stems include the branches and leaves of the vine, and finally, the sludge is generated during the wastewater treatment [17].

The bioactive compounds in the pomace mainly include anthocyanins, catechins, flavonoids, phenolic acids, and stilbenes [18,19]. Hence, grape pomace is considered to be a valuable source of phenolic compounds. Moreover, the high content of dietary fiber suggests the possible nutritive value of grape pomace, which could be used as a functional food ingredient [20,21]. The use of grape waste as a dietary antioxidant supplement has been proposed [22]. Furthermore, several studies [23,24] have reported that grape pomace has a high phenolic content. During the winemaking process, some phenolic compounds are transferred to the wine; however, they are not exhaustively extracted, and the byproduct (grape pomace) still has a substantial content of polyphenols [25,26,27]. Generally, the industrial recovery of grape pomace is aimed at the extraction of tartaric acid or production of ethanol, and pomace can also be used as a fertilizer, animal feed, compost, biomass, and soil amendment, or simply discarded [21]. Therefore, grape pomace represents a cheap and natural source for the extraction of phenolic compounds which may be an alternative to solve environmental problems caused by the disposal of this byproduct [21,28].

The industrial extraction of bioactive compounds from grape pomace is carried out by maceration [29]. Some studies have recently demonstrated that non-thermal technologies such as ultrasound may represent effective alternatives to improve polyphenol extraction. Ultrasound application at 20–35 kHz enhances the extraction of polyphenols from red-grape residues [30]. Although this methodology offers an advantage, the necessary equipment must be implemented to carry it out, which implies extra costs. In addition, it should be considered that this method can also affect the chemical structure of some polyphenols since the high temperatures and pressures generated by the collapse of cavitation bubbles can induce chemical reactions and accelerate some reactions that usually occur during wine aging [16]. In addition, this method can also affect the chemical structure of some polyphenols and modify their antioxidant activity. The introduction of a new technology on the market requires that it perform at least as well as existing commercial processes. In this sense, a patented MSPD method [patent WO 2014013122 A1, ref. [31], with modifications, could be applied for the extraction of bioactive compounds from grape pomace. The novelty of this method is the disruption of the sample with a dispersant, the isolation of the analytes on a solid support, and their subsequent elution with the solvent [32]. The appropriate combination of the dispersant and elution solvent provides adequate recovery percentages and medium selectivity, in addition to having a low cost per extraction and a moderate consumption of organic solvents [32], which are the method’s main advantages. Nevertheless, matrix solid-phase dispersion (MSPD) can be used for the extraction of polar and phenolic compounds from medicinal plants [33]. In this work, the MSPD method for the extraction of bioactive compounds from grape pomace and the maceration method of Acuña et al. [34], with some modifications, were used. The central composite design (CCD) was used to establish the optimal extraction conditions by maceration, whereas the factorial design was applied to identify the optimal extraction conditions by matrix solid-phase dispersion (MSPD).

2. Materials and Methods

Raw material. Grape pomace (Vitis vinifera var. Cabernet Sauvignon) was collected from a winery in Querétaro, Mexico (October 2015 harvest). The pomace samples were frozen and stored before analysis. The grape pomace was dried at 60 °C for 72 h and then pulverized using a Nutribullet^® processor. The method of [34] was used with some modifications. The powdered pomace (250 g) was weighed and passed through a 14–100 mesh sieve. Extraction is affected by various factors; thus, in the present work, we identified the optimal conditions for the extraction of polyphenols from the pomace of Cabernet Sauvignon grapes. The optimization method is described below.

2.1. Optimization Process with Experimental Designs

The optimization of the experimental design helps find the optimal conditions for the responses of interest. In the case of the phenolic extracts, both the quality and quantity are affected by several factors. However, it is not possible to identify the effects of all parameters at the same time; thus, parameter grouping is necessary [23]. However, because the factors considered in the extraction by maceration are continuous but the factors considered in the MSPD extraction are discrete, two experimental designs were needed: a response surface design for extraction by maceration and a factorial design for extraction by MSPD.

Design of experiments with a factorial design. A factorial design is a type of experiment devised to study the effects that various factors may have on a response. When conducting an experiment, varying the levels of all factors at the same time instead of one at a time allows the interactions between factors to be studied [35].

Design of experiments with a response surface design. According to [36], the response surface methodology is a set of mathematical techniques used in the treatment of problems in which a response of interest is influenced by various factors. The objective of these techniques is to design an experiment that provides reasonable values for a response variable and determine the mathematical model that best fits the data obtained. Thus, the values of the factors that optimize the value of the response variable are established. The mathematical model of the response surface is shown in Equation (1).

$$y={\beta }_0+{{\displaystyle \sum }}_{i=0}^k{\beta }_i{x}_i+{{\displaystyle \sum }}_{i=1}^k{\beta }_{ii}{x}_i^2+{{\displaystyle \sum }}_i{{\displaystyle \sum }}_j{\beta }_{ij}{x}_i{x}_j$$

(1)

where: y: Answer

x_i: Significant factors.

β₀: Constant term.

β_i: Coefficient of the model that affects the factor x_i.

β_ij: Interaction coefficient between factor x_i and factor x_j.

β_ii: Coefficient that explains the curvature of the factor $x_i^2$.

Therefore, for the extraction by maceration, the proportion of ethanol and the maceration time were evaluated, whereas the sample/dispersant ratio and the elution volume were selected as factors for evaluation in the extraction by MSPD. The phenolic compounds and antioxidant activity were the response variables.

2.2. Extraction by Maceration

Extraction by maceration is a solid–liquid extraction technique, where the sample is soaked in an appropriate solvent which penetrates the tissues, softening and dissolving the soluble portions until a concentration in equilibrium with that remaining in the sample is reached [37,38]. The speed and efficiency of the extraction are affected by various factors, mainly those related to the solubility of the compounds to be extracted. The temperature, solvent concentration, particle size, porosity, and agitation are the main factors that affect the extraction. Each variable adds different properties to the process; thus, to determine the optimal extraction method, it is necessary to study these variables [39]. For the extraction of total phenolic compounds from grapes used in wine production, different solvents have been used but the most commonly used are ethanol and methanol. Although in some works it has been reported that methanol is more efficient for the extraction of bioactives, this will depend on the type of sample since these may contain compounds with greater affinity to a particular solvent. Moreover, the fermentation process of the grapes to obtain the wine causes changes in the chemical composition due to the process conditions, such as time, temperature, grape variety, etc. Other factors to consider are the origin of the grapes and their variety on which the quantity and type of bioactives depend [34]. On the other hand, the use of ethanol is safer, especially for extracts intended for human consumption.

Procedure: To identify the optimal extraction conditions for bioactive compounds, the central composite design in Minitab 17 was used. This design comprised 13 experiments performed in triplicate. Two factors with five levels each were analyzed (

Table 1. Factors and levels applied in the extraction by maceration of phenolic compounds.

Factors
Levels	Ethanol (%)	Time (h)
−1.414	44	1
−1	50	4:22
0	65	12:30
1	80	20:38
1.414	86	24

).

For the above, ethanol was selected as the solvent as it has been widely used for the extraction of bioactive compounds from grape pomace in proportions ranging from 50% to 80% [21,40]. Ethanol is recognized as safe by the FDA [19]. The maceration time (1, 4, 12, 20, and 24 h) and the ratio of ethanol acidified with 1% HCl were the two factors studied. Each determination was performed in an amber vial at 23 °C, using 1 g of sample in 10 mL of solvent. The sample was mixed manually and left to stand at room temperature for the time indicated in the sample matrix design. Once the extraction time elapsed, each extract was centrifuged for 10 min at 3000 rpm and filtered through organza. The extractions were carried out according to the design generated in Minitab 17. As response variables, the total phenol content (TPC) and antioxidant capacity (AC) were measured as described below.

Validation of the Optimal Extraction Conditions by Maceration

Once the optimal conditions (time and proportion of ethanol) of the extraction by maceration were found, the TPC and AC were determined in triplicate. The error with respect to the predicted value was calculated and the residuals were plotted. We also verified that the results obtained (TPC and AC) were within the prediction interval of the model at a 95% confidence level.

2.3. Matrix Solid-Phase Dispersion (MSPD)

The appropriate combination of the dispersant and elution solvent provides adequate recovery percentages and medium selectivity, in addition to having a low cost per extraction and a moderate consumption of organic solvents [41], which are the method’s main advantages. As in the extraction by maceration, the extraction by MSPD is affected by various factors that must be optimized. The factors that affect the extraction can be grouped into two types: those related to the elution and those related to the dispersant and the sample. The selection of the solvent is related to the nature of the sample and the polarity of the analyte; mixtures of organic solvents are generally used [42]. With the elution sequence, the aim is to isolate the analyte or eliminate interfering substances from the column. MSPD columns allow the isolation of analytes of different polarity or entire classes of chemical compounds in a single solvent or in solvents of different polarity passing through the column, facilitating multi-residue isolation and analysis on a single sample [43]. Minuscule particles (3–10 µm) require long elution times as well as high pressure or high vacuum to achieve elution. Good results have been reported for silica particles in the 40–100 μm range, in addition to their low cost [44]. Classic MSPD applications use reversed-phase adsorbents for the dispersion. Generally, dispersant/sample ratios of 1:1 to 1:4 are used, the latter being the most common; however, the ratio varies from application to application. Most of the protocols that use materials with lipophilic bonded phases (C18, C8) mix 2 g of dispersant with 0.5 g of sample. Nevertheless, this relationship is application-dependent and should be evaluated during method development to optimize it [42,43].

Matrix effect: All the components of the sample are dispersed through the column, covering a large part of the surface of the dispersant (i.e., the packing of the column), thus creating a new phase that can have critical effects on the isolation of the analyte; these effects vary from one matrix to another [45].

Procedure: The MSPD extraction was carried out as described by [31], with modifications. The dispersant used for the evaluation was sea sand (Merck, mesh size of 60–100) [46]. The reagents and solvents were from the following brands: Sigma Aldrich, Fermont, J.T. Baker, Hycel, and Alfimex. According to [31], in the first stage of the MSPD extraction, the pomace is ground together with the dispersant and the appropriate solvent and is left to macerate; then, it is used to pack and elute the column. In the present work, the maceration time and the ethanol proportion of the solvent were selected based on the optimal extraction conditions for maceration. This was the first stage in the MSPD extraction. To identify the optimal conditions for the extraction of the bioactive compounds, a 2² factorial design was used (Minitab 17), with the elution volume and the ratio of sample/dispersant being the two factors analyzed in four experiments performed in duplicate. Two factors with two levels each were analyzed (

Table 2. Factors and levels applied in the extraction by MSPDSand of phenolic compounds.

Factores
Levels	Ratio Sample/Dispersant	Elution Volume (mL)
−1	1:2	48
1	1:4	96

). The TPC and AC were measured as response variables.

The extractions were carried out according to the generated design (

Table 3. Factorial design for the selection of the best extraction. Conditions by MSPD of phenolic compounds.

Run	Proportion Pomace/Dispersant	Elution Volume (mL)
1	1:4	96
2	1:2	48
3	1:2	96
4	1:2	96
5	1:4	48
6	1:4	48
7	1:4	96
8	1:2	48

). A sample of 3 g of powdered pomace was weighed in an amber bottle and was mixed with the corresponding proportion of sand (1:2 or 1:4) [45] and 30 mL of solvent (ethanol: water: hydrochloric acid, 57.5: 41.5: 1%, v/v/v). The mixture was ground in a mortar for 1 min, quantitatively transferred to a 250 mL amber vial, and left to macerate for 24 h at room temperature.

The blend was then transferred to a glass column (450 × 17.89 mm (id)) and packed. Before packing the glass column, a fiberglass filter was placed, followed by four 1.7 cm diameter discs of organza and four discs of the same diameter of fiberglass, which prevented the column from being covered and retained the pomace/dispersant mixture. Once the extraction time elapsed, the pomace/dispersant mixture was quantitatively transferred to the glass column, using 10 mL of solvent. The packing of the column was carried out by gravity, taking care that the material did not form paths or become compacted or compressed. The elution was carried out under vacuum, with a total volume of 48 or 96 mL, divided into three parts (i.e., three elutions of 16 or 32 mL as appropriate, according to the experimental design shown in Table 3).

Finally, the eluate was adjusted to 100 mL when the elution volume was 48 mL, and to 150 mL when it was 96 mL. Figure 1 illustrates the packed column used for the extraction. The optimum extraction conditions for each experimental design were determined using a general function optimization. To verify the validity and adequacy of each prediction model, three extractions were performed under optimal conditions and the relative error (%RE) was calculated (%RE = (observed value − predicted value)/(observed value) × 100).

2.4. Total Phenolic Compounds (TPC)

The Folin–Ciocalteu reagent was used according to the methodology of [47]. A 0.4 mL aliquot of the extract was diluted with 2 mL of the Folin–Ciocalteu reagent (0.5 mol/L) and sonicated for 10 min. Subsequently, the reaction was neutralized with 2 mL of 75 mg/mL saturated Na₂CO₃. The absorbance was measured at 765 nm after incubation for 2 h at room temperature in the dark (Mod. Lambda 35 UV-Visible spectrophotometer Perkin-Elmer). The results were expressed as milligram equivalents of gallic acid (mg GAE/100 g wet base (wb)).

2.5. Antioxidant Activity by DPPH

The antioxidant capacity was quantified by preparing a standard curve using Trolox and measured based on the stable radical 2,2-diphenyl-1-picrylhydrazyl (DPPH), following the methodology of [48], with some modifications. DPPH was prepared at a concentration of 0.1 mM in 80% methanol. A 0.1 mL aliquot of the extract, previously obtained for the determination of phenols, was mixed with 3.9 mL of the DPPH solution, vigorously shaken, and sonicated for 10 min, and the absorbance of the sample was immediately measured at 515 nm. The antioxidant capacity of the extract was expressed as millimoles equivalent of Trolox/100 g of dry base pomace (mM TE/100 g db).

3. Results

3.1. Extraction by Maceration

The equations for the phenolic content and the antioxidant capacity of the extracts were obtained under the conditions generated by the central composite design produced in Minitab 17. The equations used coded terms for each of the responses, which were adjusted to a quadratic model (

Table 4. Coefficients of determination and equations in coded terms of the responses of the composite central design.

Respuesta	R2	Equation in Coded Terms (MACERATION)
TPC	0.8356	TPC = +2346.73 + 157.99 A − 285.92 B − 58.42 A2 − 239.02 B2 + 1.77 AB
AC	0.9560	AC = +11.11 + 1.13 A − 1.39 B − 0.20 A2 − 1.51 B2 + 0.043 AB

TPC = Total phenolic compounds; AC = Antioxidant capacity; A: Time; B: Proportion of ethanol.

). The significant values for each equation are shown in bold.

The effect of the maceration time (A) and the proportion of ethanol (B), as well as the B² interaction, were significant for the phenolic content. The determination coefficient (R²) was 0.8356, which indicates that the model explained 83.56% of the variability; this value is higher than the 0.725 and 0.759 reported by [49] for the extraction by maceration of phenolic compounds from Cabernet Sauvignon grapes, using acetone–water and methanol–water as solvents. In the case of antioxidant capacity (

Table 5. Real values of the total phenol content and antioxidant capacity of the extracts carried out under optimal conditions and values predicted by the model.

Total Phenol Content (mg GAE/100 g Pomace d.b)
Extract	Predicted Value	Real Value	Error	% de Error
1	2535.28	2528.81	6.46	0.25
2	2535.28	2689.66	−154.39	−6.09
3	2535.28	2567.67	−32.40	−1.28
Antioxidant Capacity (mMol TE/100 g Pomace d.b)
Extract	Predicted Value	Real Value	Error	% de Error
1	12.59	12.27	0.33	2.59
2	12.59	12.71	−0.11	−0.91
3	12.59	13.34	−0.75	−5.97

), the constant, the time (A), the proportion of ethanol (B), and the A² and B² interactions were significant. The R² value of 0.956 indicates that the model explained 95.60% of the variability and was greater than the 0.866 reported by [49] for the extraction of antioxidant compounds from Cabernet Sauvignon pomace. For the extraction by maceration, the experimental data were fitted to obtain a second-degree regression, and three-dimensional response surface plots were constructed (Figure 2).

The TPC ranged from 1342.7975 to 2613.7556 mg GAE/100 g pomace db, and the AC ranged from 6.2537 to 13.0020 mMol TE/100 g pomace db. For the extraction of phenolic compounds, the maceration time increased their content, whereas the proportion of ethanol exhibited a quadratic tendency (i.e., the maximum point was achieved at an ethanol proportion of 56% and a maceration time of 23 h). For the AC, time had a positive effect (positive sign of term A in the equation). In other words, as the maceration time increased, the AC increased. On the other hand, the proportion of ethanol exhibited a quadratic trend, with a maximum point at 24 h of maceration with a proportion of ethanol of 58%, which corresponds to the blue region of the graph.

An optimization analysis of multiple responses was carried out, looking for the maximum values of both variables; the results are shown in Figure 3. The global desirability was acceptable [50], as it had a value of 0.9388, which means that the model has a 93.88% probability of being optimal [51]. The optimal extraction conditions were the use of a solvent composed of ethanol: water: hydrochloric acid (57.5: 41.5: 1.0%, v/v/v) for 24 h. By introducing these conditions into the model (Figure 3), the AC (12.59 mMol TE/100 g pomace db) and TPC (2535.28 mg GAE/100 g pomace db) were within the prediction interval (at a 95% confidence level) for the phenolic content (2221.1–2849.4 mg GAE/100 g pomace db) and AC (11,708–13,473 mmol TE/100 g of pomace db).

Validation of the Optimal Extraction Conditions by Maceration

As mentioned above, the model was validated by performing the extraction in triplicate under optimal conditions. Once the extracts were obtained, the TPC and AC were analyzed. The results, together with the theoretical values and the error percentages, are shown in Table 5.

The percentage of error between the predicted and the real value for TPC was between −6.09 and 0.25%, whereas for AC, it was between −5.97 and 2.59%. These values are similar to the error reported by [51], who extracted phenolic antioxidants from white tea. All the real values were found in the prediction interval for both TPC and AC. After calculating the error between the predicted and real value, the residuals were plotted (Figure 3B); these were fitted to a normal distribution and the real value was within the 95% confidence interval delimited by the lines. Therefore, the model correctly predicted the extraction of phenolic compounds and antioxidant capacity. The optimal conditions of the extraction by maceration were used for the first stage of the extraction by MSPD. Subsequently, the optimization of the MSPD was carried out as described below.

3.2. MSPD Extraction

The determination coefficients (R²) and the equations in coded terms for each of the responses are presented in

Table 6. Coefficients of determination and equations in coded terms of the responses of the factorial design.

Response	R2	Equation in Coded Terms (MSPD)
TPC	0.6744	TPC = +2833.19 − 60.70 A + 83.45 B − 60.84 AB
AC	0.8641	AC = +13.72 − 0.070 A + 0.86 B − 0.23 AB

TPC = Total phenolic compounds; AC = Antioxidant capacity; A: Ratio Sample/dispersant; B: Elution volume.

; the significant factors for each equation are shown in bold.

Figure 4a shows that the elution volume (B) had a positive effect (the greater the volume, the greater the CA), which is reflected in the positive sign for term B of the equation. By contrast, the sample/dispersant ratio (A) had a negative effect (a negative sign is observed in the equation for term A).

This may be due to the fact that decreasing the amount of dispersant favors the disruption of the sample and the release of bioactive compounds, and on the other hand, by increasing the dispersant, the pomace is surrounded by a greater amount of sand and it is less available for the solvent to penetrate and extract the compounds.

Figure 4b shows that the elution volume had a notable effect on the antioxidant capacity (a standardized effect greater than 2.76). The constant term and the elution volume had a significant effect on the antioxidant capacity. The coefficient of determination (

Table 7. Real values of the total phenol content and antioxidant capacity of the extracts made under the best conditions and values predicted by the model.

Total Phenol Content (mg GAE/100 g Pomace db)
Extract	Predicted Value	Real Value	Error	% de Error
1	3038.18	2869.48	168.70	5.55
2	3038.18	2789.51	248.67	8.18
3	3038.18	2851.18	187.00	6.15
Antioxidant Capacity (mMol TE/100 g Pomace db)
Extract	Predicted Value	Real Value	Error	% de Error
1	14.88	14.17	0.72	4.81
2	14.88	14.08	0.80	5.38
3	14.88	13.80	1.08	7.28

) was 0.8641, which indicates that the model explained 86.41% of the variability, and is similar to the 0.866 reported by [23] for the extraction of antioxidant compounds from Cabernet Sauvignon pomace.

Figure 5a,b show the TPC and AC obtained using the MSPD with the data obtained by the 2² factorial design with each tested condition (n = 2). The total phenol content was between 2657.73 and 3163.95 mg EAG/100 g pomace db, whereas the antioxidant capacity was in the range of 12.57–15.54 mMol ET/100 g pomace db. A higher value could be obtained if a 1:2 sample/dispersant ratio and an elution volume of 96 mL were used. This suggests that using a smaller amount of sand favors the disruption of the sample and the release of bioactive compounds while occupying a larger elution volume, washing the column more efficiently, and resulting in greater extraction of bioactive compounds. Under Figure 5 are the results of the ANOVA which was carried out. The one-way ANOVA (p = 0.05) of the Antioxidant Capacity values showed that there was no significant difference between the two sample/dispersant ratios (p = 0.726); for the elution volume, there was a significant difference with a p value of 0.006 and there was no lack of model fit (p value = 0.259), obtaining a value of R² = 80.54%. Regarding the Total Phenol Content, no significant difference was found between the sample/dispersant ratio (p value = 0.245) and the elution volume (p value = 0.130).

By introducing the optimal conditions of both variables into the model (Figure 6), an antioxidant capacity of 14.88 mMol TE/100 g pomace db and a phenolic content of 3038.18 mg GAE/100 g pomace db were expected. The global desirability (0.7647) was acceptable [50], as the obtained value means that the probability that the model is optimal is 76.47% [51]. The prediction interval at a 95% confidence level for the antioxidant capacity was between 13,181 and 16,580 mMol TE/100 g pomace db, and for the phenolic content, it ranged between 2637.9 and 3438.5 mg GAE/100 g pomace db.

Confirmation of the Points Selected as the Optimal Extraction Conditions by MSPDSand

The model was validated by performing the extraction in triplicate under optimal conditions. Once the extracts were obtained, the TPC and AC were analyzed. Table 7 presents the results obtained together with the predicted values and the error percentages. The predicted value and the real value were similar to those reported in [51] for the optimization of the extraction of phenolic antioxidants from white tea. In the present study, the real value was lower than the predicted value for both responses (Table 7), which may be due to the fact that, as mentioned by [52], it is difficult for all the responses to meet their maximum values at the same time. However, although the predicted values were higher than those obtained, they were within the prediction interval for both the TPC (2637.9–3438.5 mg GAE/100 g pomace db) and AC (13,181–16,580 mmol TE/100 g pomace db) at a 95% confidence level.

The TPC obtained in the present study agrees with some studies (186–7475 mg GAE/100 g db [25,26]; however, the value obtained here is lower than those obtained by [23,28,53,54], as those authors reported 7475 ± 222, 5500, 5304 ± 482, and 3300 ± 350 mg GAE/100 g db in Cabernet Sauvignon grape pomace from Brazil, Lebanon, Italy, and France, respectively. However, ref. [27,28] found mean TPC values of 186 GAE/100 g db for the mixture of Cabernet Sauvignon, Syrah, and Tempranillo from Mexico, and 1230 ± 0.4 mg GAE/100 g db for Cabernet Sauvignon pomace from Colombia. Furthermore, for the AC, ref. [53] reported values of 50.55 ± 46.2 and 48.542 ± 37.2 mMol TE/100 g db, measured with the DPPH and ABTS assays, respectively. These values are higher than those found in the present work.

In the present investigation, the best conditions were found for the extraction of phenolic compounds from grape pomace from a producing region in Mexico. However, work must continue to improve the extraction conditions of the different components of wine waste. This can be done considering that an important feature of MSPD is that process efficiency can be easily optimized in a few simple stages by fitting an appropriate sorbent type and determining the sample mass-to-sorbent mass ratio, blending time, eluent composition, and/or its volume. Occasionally, to increase the MSPD yield during blending, modifiers such as acids, bases, salts, chelators, and co-sorbents are added [55]. MSPD is a unique sample preparation method that can be directly applied for semi-solid, solid, and viscous samples. MSPD does not need the solvent extraction step and the samples can be blended with sorbents directly to obtain a homogeneous mixture; then, the mixture is transferred and packed in the SPE cartridge and is washed and eluted with liquid solvents. Therefore, MSPD eliminates steps of pretreatment such as centrifugation and filtration. [56].

4. Conclusions

The central composite design allowed us to establish the optimal extraction conditions by maceration, whereas with the factorial design, it was possible to identify the optimal extraction conditions by matrix solid-phase dispersion (MSPD). Through model predictions, we found that the total phenolic content (TPC) and antioxidant capacity (AC) of the extracts were within the prediction intervals at a 95% confidence level. By maceration, the percentage of error between the predicted and the real value for TPC was between −6.09 and 0.25%, whereas for AC, it was between −5.97 and 2.59%. Regarding the MSPD method, the prediction interval at a 95% confidence level for the antioxidant capacity was between 13,181 and 16,580 mMol TE/100 g pomace db, and for the phenolic content, it ranged between 2637.9 and 3438.5 mg GAE/100 g pomace db. However, although the predicted values were higher than those obtained, they were within the prediction interval obtained.

Therefore, the models could be used to predict the response values. Moreover, when the extraction was performed by MSPDSand, greater TPC concentration and AC were obtained. Thus, MSPDSand was chosen for further analysis. The TPC and AC of the extracts demonstrated that Cabernet Sauvignon grape pomace has great potential as a source of antioxidant compounds for food applications. Moreover, the content of bioactive compounds was comparable to that reported for Cabernet Sauvignon grape pomace from other countries. Both the maceration method and MSPD proved to be efficient for the extraction of phenolic compounds from waste from the winemaking process. However, when using the MSPD method, higher values of both TPC and AC were obtained under the optimal conditions. Future research may evaluate the use of other solvents and dispersion materials with this method to obtain bioactive compounds (tannins, anthocyanins, etc.) with different polarities.