Effect on the Antioxidant, Lipoperoxyl Radical Scavenger Capacity, Nutritional, Sensory and Microbiological Traits of an Ovine Stretched Cheese Produced with Grape Pomace Powder Addition

An innovative ovine cheese enriched with red grape pomace powder (GPP) was produced to improve the functional properties of Vastedda cheese typology. Vastedda cheese making was performed adding GPP and four selected Lactococcus lactis strains (Mise36, Mise94, Mise169 and Mise190). For each strain, 40 L of pasteurized ewe’s milk was divided into two aliquots representing control and experimental trials. Control cheese (CC) production did not contain GPP, while the experimental cheese (EC) production was enriched with 1% (w/w) GPP. GPP did not slow down starter development and acid generation. Plate counts and randomly amplified polymorphic DNA (RAPD)-PCR analysis confirmed the dominance of the starters in all trials. The evolution of the physicochemical parameters showed that EC productions were characterized by lower fat content, higher protein content, and higher values of secondary lipid oxidation. Sensory evaluation indicated that the cheeses produced with the strain Mise94 were those more appreciated by the judges. Thus, the last cheeses were investigated for some functional aspects: GPP enrichment significantly increased antioxidant activity and lipoperoxyl radical scavenger capacity, confirming that grape polyphenol inclusion in cheese represents an optimal strategy for the valorization of ovine cheeses as well as winemaking industry by-products.


Introduction
Nowadays, natural antioxidants extracted from plant by-products represent a useful source of active compounds to produce healthy and functional foods more and more requested by consumers [1]. In order to face this increasing demand, in the last years, the food industries also focused the attention toward dairy products. Thus, several efforts are being made to enrich dairy products such as cheeses with functional and antioxidant components [2] because these foods are basically poor in bioactive compounds [3].
The most interesting natural antioxidants belong to the chemical classes of phenols [4]. They are particularly present in some fruits such as red grape [5]. Thus, winemaking industry by-products constitute an interesting source of bioactive phenolic compounds [6]. The main by-products generated during the winemaking process are composed by a mix

Cheese Production and Sample Collection
The experimental cheese making trials were carried out at a dairy pilot plant (Biopek, Gibellina, Italy) following the protocol of production for the stretched cheese "Vastedda" [16]. The experimental plan included two different cheese productions for each LAB strain for a total of eight trials: CP36, control production with L. lactis MISE36; EP36, experimental production with L. lactis MISE36 + 1% of GPP; CP94, control production with L. lactis MISE94; EP94, experimental production with L. lactis MISE94 + 1% of GPP; CP169, control production with L. lactis MISE169; EP169, experimental production with L. lactis MISE169 + 1% of GPP; CP190, control production with L. lactis MISE190; EP190, experimental production with L. lactis MISE190 + 1% of GPP. Each trial was obtained with 20 L of pasteurized (72 • C for 15 s) ewes' milk. Cheese productions were performed in plastic vats after sanitization with a solution of 2% (v/v) of H 2 O 2 and acid adjuvant (PROMOX P900, Leggiuno, Italy) for 15 min to eliminate contaminant bacteria and fungi. Milk (cooled at 38 • C) was firstly inoculated with the corresponding NMSC (200 mL) to reach the final cell densities of 10 7 CFU/mL and secondly added with 6 mL of liquid rennet (Fromase ® 220 TL, DSM Bright Science Brighter Living, Heerlen, Netherlands). After coagulation, the curds were broken with a stainless-steel curd beater until small rice-size grains were obtained. After whey draining, the control curds were immediately put into perforated containers while the experimental curds were added with 1% (w/w) of GPP, which was manually mixed and then transferred into perforated containers. The pH of control and experimental curds was monitored electrometrically by means of the pH-meter pH 70 + DHS (XS Instruments, Carpi, Italy). When curd acidification reached pH values in the range 5. 20-5.40, the curds were stretched under hot (80-85 • C) water and molded into a bun shape. The cheeses were salted in saturated brine for 30 min, dried for 24 h at room temperature, packaged under vacuum, and kept under refrigeration (5 • C) for 15 d. All cheese trials were carried out in duplicate in two consecutive weeks. The sampling points, the number of samples analyzed, and the analyses performed are reported in Table 1.

Microbiological Analyses
Cell suspensions of milk samples were directly subjected to decimal serial dilutions in Ringer's solution, while GPP, curd and cheese samples (15 g) were first homogenized with Ringer's solution (135 mL) by the stomacher Bag-Mixer 400 (Interscience, Saint Nom, France) for 2 min at the maximum speed (blending power 4) and then 1:10 serially diluted.
Cell suspensions of GPP were used to enumerate the following microbial groups: Total mesophilic microorganisms (TMM) on plate count agar (PCA), after incubation at 30 • C for 72 h; LAB rods and cocci on de Man-Rogosa-Sharpe (MRS) agar and M17 agar, respectively, both incubated at 30 • C for 48 h in anaerobiosis; members of the Enterobacteriaceae family, detected on violet red bile glucose agar (VRBGA), after incubation at 37 • C for 24 h; yeasts on dichloran rose bengal chloramphenicol (DRBC) agar, incubated at 28 • C for 24 h; coagulase-positive and coagulase-negative staphylococci (CPS and CNS, respectively) on Baird Parker (BP) agar added with rabbit plasma fibrinogen, incubated at 37 • C for 48 h; Listeria monocytogenes on Listeria selective agar base (LSAB) added with SR0140E supplement, incubated at 37 • C for 48 h; Escherichia coli and Salmonella spp., detected on Hektoen enteric agar (HEA), incubated at 37 • C for 24 h. Anaerobiosis occurred in hermetically sealed jars added with the AnaeroGen AN25 system (Oxoid, Milan, Italy).
Cell suspensions of raw milk and pasteurized milk were subjected to plate count for the enumeration of TMM on Skim milk agar (SMA) ISO 6610 [17] incubated aerobically at 30 • C for 72 h; mesophilic rod-shaped LAB on MRS agar, acidified to pH 5.4 with lactic acid (5 mol/L) and incubated anaerobically for 48 h at 30 • C; thermophilic LAB rods were pour plated on whey-based agar medium (WBAM) prepared as described by Settanni et al. [18]; mesophilic and thermophilic coccus-shaped LAB on M17 agar, incubated anaerobically for 48 h at 30 and 44 • C, respectively.
Milk inoculated with each NMSC, curd and cheese samples were analyzed only for TMM and L. lactis on SMA and M17 agar, respectively. The last media were incubated as reported above.
All media and supplements were purchased from Biotec, except HEA provided by Microbiol Diagnostici (Uta, Italy). Plates counts were performed in duplicate.

Persistence of Added Strains and Identification of the Survival Indigenous Milk LAB
The persistence of the strains added as starter cultures and their dominance over the indigenous milk LAB resistant to pasteurization process was performed by randomly amplified polymorphic DNA (RAPD)-PCR technique as reported by Alfonzo et al. [19]. The colonies developed from the highest dilutions on the agar media used for LAB counts were randomly picked up, purified by successive sub-culturing, and tested for Gram reaction and catalase activity [20]. The presumptive LAB cultures, grown overnight, were subjected to DNA extraction using the DNA-SORB-B kit (Sacace Biotechnologies Srl, Como, Italy) following the protocol provided by the manufacturer. RAPD-PCR was performed by means of Swift™ MaxPro Thermal Cycler (Esco Health Care Pte. Ltd., Singapore) and the amplicons were separated by electrophoresis. The recognition of the added strains was performed by comparing the polymorphic profiles obtained from the isolates of a given trial with that of the pure culture added using the software Gelcompare II version 6.5 (Applied-Maths, Sin Marten Latem, Belgium).
The identification of the different LAB strains isolated from pasteurized bulk milk before NMSC inoculation was carried out by 16S rRNA gene sequencing as reported by Weisburg et al. [21]. The resulting DNA fragments were purified and sequenced by Eurofins Genomics (Ebersberg, Germany). The sequences were compared with the sequences available in GenBank/EMBL/DDBJ (http://www.ncbi.nlm.nih.gov) and Ez-Taxon-e (http://eztaxon-e.ezbiocloud.net/) databases.

Physicochemical Analyses of Cheeses
Cheeses after 15 d of refrigerate storage were assessed for external and internal color, measured in duplicate by a Minolta Chroma Meter CR300 (Minolta, Osaka, Japan) using the illuminant C; results are expressed as lightness (L*, from 0 = black, to 100 = white), redness (a*, from red = +a, to green = −a), and yellowness (b*, from yellow = +b, to blue = −b), according to the CIE L* a* b* system. Cheese hardness was evaluated with an Instron 5564 tester (Instron, Trezzano sul Naviglio, Milan, Italy) measuring the maximum resistance to compression (compressive stress, N/mm 2 ) of samples (2 cm × 2 cm × 2 cm) kept at room temperature (22 • C).

Volatile Organic Compounds
Five grams of the final cheese samples were finely chopped and placed into 25 mL glass vials sealed with silicon septum. The headspace solid phase microextraction SPME (DVB/CAR/PDMS, 50 mm, Supelco) fiber was exposed to the cheese under continuous stirring at 60 • C for 15 min. After sampling, the SPME fiber was thermally desorbed for 1 min through a splitless GC injector at 250 • C. A gas chromatograph (Agilent 6890) equipped with a mass selective detector (Agilent 5975 c) and a DB-624 capillary column (Agilent Technologies, 60 m, 0.25 mm, 1.40 µm) was used for the chromatographic analysis. Chromatographic conditions applied were helium carrier gas at 1 mL/min and an oven temperature program with a 5 min isotherm at 40 • C followed by a linear temperature increase of 5 • C min up to 200 • C, where it was held for 2 min. The MS scan conditions were interface temperature was 230 • C acquisition mass range was 40-400; acquisition mode was scan. The volatile organic compounds were also determined for GPP.
The identification of VOC compounds was performed through the comparison of the MS spectra with commercial library NIST05. The relative proportions of the identified constituents were expressed as percentages obtained by GC-MS peak area normalization with total area of significant peaks. Three replicates of each sample were analyzed.

Sensory Evaluation
The cheeses were also evaluated for their sensory traits. All cheeses were judged by 12 assessor members including 6 men and 6 women (aged between 21-65 years old) familiar with the sensory analysis of cheese. All panelists were specifically trained for cheese attribute evaluation following the ISO 8589 [27] indications. The cheeses were acclimated at about 20 • C for 1 h, cut into cubes (3 cm × 3 cm × 3 cm) and then coded and served in a random order. Twelve descriptive attributes were judged as evaluated by Costa et al. [28].
In particular, for each cheese, the evaluation considered the following aspects: Intensity of odor and aroma, sweet, salt, bitter, acid, fiber, friability, adhesiveness, hardness, humidity, and the overall assessment. Each aspect was scored using a line scale from 0 to 7 (cm) as reported by Faccia et al. [29].

In Vitro Gastrointestinal Digestion
Simulated in vitro human digestion, mimicking physicochemical and biochemical changes of the food in the upper gastrointestinal (GI) tract, was performed (n = 3) as reported by Attanzio et al. [30]. Samples of 7.5 g of cheese were added in 20 mL of a buffered pH 6.8 solution simulating saliva [NaCl (0.126 g), KC1 (0.964 g) KSCN (0.189 g), KH 2 PO 4 (0.655 g), urea (0.200 g), Na 2 SO 4 ·10H 2 O (0.763 g), NH 4 Cl (0.178 g), CaCl 2 ·2H 2 O (0.228 g), and NaHCO 3 (0.631 g) in 1 L of distilled water]. The mixture was blended for 3 times × 15 s in a semi-micro stainless-steel blender jar using a Waring blender (Waring, New Hartford, CT, USA) miming the oral phase of digestion. Post-oral samples were then transferred into a bottle, acidified at pH = 2.0 with HCl before porcine pepsine addition (8 mg/mL; 3200-4500 units/mg, Sigma-Aldrich). The bottle was sealed and incubated in a water bath (type M 428-BD, Instruments s.r.l., Bernaggio, Mi, Italy) with shaking (100 rpm), at 37 • C for 2 h, in the dark. Post-gastric digested were brought to a final pH 7.5 adding 200 mM NaH 2 PO 4/ Na 2 HPO 4 buffer and 5 M NaOH. Then porcine bile extract (2.4 mg/mL, Sigma) and pancreatin from hog pancreas (0.4 mg/mL, Sigma) were added. The bottle was sealed and incubated in the shaking water bath at 37 • C for 2 h, in the dark. The digestion was stopped by ice bath immersion. Post-intestinal digested was centrifuged at 167,000× g for 35 min at 4 • C (Beckman Optima TLX ultracentrifuge, Beckman Instruments, Inc., Palo Alto, CA, USA) to separate supernatant (Bioaccesible fraction) from the particulate material. Aliquots of samples from each digestion step were withdrawn and centrifuged at 1500× g for 10 min at 4 • C. Supernatants and bioaccesible fraction were brought at pH 2.0 to stabilize polyphenols and stored at −80 • C until analysis.

ABTS + Radical Cation Decolorization Assay
Radical scavenging activity was evaluated using the ABTS + radical cation decolorization assay as described by Attanzio et al. [31]. Samples were analyzed in duplicate, at three different dilutions, within the linearity range of the assay. The assay was standardized with the water-soluble vitamin E analog Trolox (Sigma), and results were expressed as µmol Trolox equivalents/g cheese.

Lipid Peroxidation Assay
Pig's brain microsomes were prepared from a tissue homogenate in 10 mM phosphate buffer saline, pH 7.4 (PBS), by differential centrifugation. Microsomes in PBS (2 mg protein/mL) were pre-incubated for 5 min at 37 • C either in the absence (control) or in the presence of variable amounts of the bioaccessible fraction of cheeses. Lipid oxidation was induced by 20 mM 2,2 -azobis (2-amidino-propane) dihydrochloride] (AAPH, Sigma). Lipid hydroperoxides formation was monitored after reaction with thiobarbituric acid (TBA), as TBA-reactive substances (TBARS). Aliquots (1 mL) of the reaction mixture were added to 2 mL of a solution containing 15% TCA (v/v), 0.375% TBA (w/v), 0.25 N HCl, and 0.02% BHT (w/v), to prevent formation of non-specific TBA-RS and decomposition of AAPH during the subsequent boiling. The mixture was incubated in a boiling water bath for 30 min. After cooling, samples were clarified by centrifugation, and TBA-RS in the supernatant determined at 532 nm. The results are expressed as nmol of MDA equiv/mg of protein, using the molar extinction coefficient of 156,000.
Proteins in microsomal preparations were determined by the Bio Rad colorimetric method [32].

Statistical Analyses
Data of microbiological analyses and physicochemical traits of 15 d cheeses were statistically analyzed using the generalized linear model (GLM) procedure in SAS 9.2 (SAS, 2010) to evaluate the effects of cheese trial (1, 2), treatment (TR) with grape pomace powder (control, experimental), starter culture (NMSC: Mise36, Mise94, Mise169, Mise190) and the interaction TR*NMSC. When a statistically significant effect (p ≤ 0.05) of NMSC and TR*NMSC interaction was detected, means were compared using p-values adjusted according to the Tukey-Kramer multiple comparisons test. The results of in vitro digestion of cheeses were made using one-way ANOVA test, with Tukey's correction for multiple comparisons by Instat-3 statistical software (GraphPad Software Inc., San Diego, CA, USA). Comparison between individual group means was performed by unpaired Student's t-test. In all cases, significance was accepted if the null hypothesis was rejected at the p < 0.05.

Acidification Kinetics of Curds
The evolution of curd pH is shown in Figure 1. Statistically significant differences were found between control and experimental curds for all strains inoculated. The initial values of pH of control curds were at ca. 6.7, while lower values (ca. 6.4) were registered for the experimental curds. These differences (p < 0.0001) were registered until the end of the acidification process.
The lower values of pH in experimental curds are imputable to the presence of organic acids such as tartaric acid, malic acid, and citric acid in GPP [7]. However, all curds reached the value comprised in the range 5.20-5.40, which represents the optimum level of acidity allowing the stretching of the curd [33], about 24 h after curdling ( Figure 2).

Acidification Kinetics of Curds
The evolution of curd pH is shown in Figure 1. Statistically significant differences were found between control and experimental curds for all strains inoculated. The initial values of pH of control curds were at ca. 6.7, while lower values (ca. 6.4) were registered for the experimental curds. These differences (p < 0.0001) were registered until the end of the acidification process. The lower values of pH in experimental curds are imputable to the presence of organic acids such as tartaric acid, malic acid, and citric acid in GPP [7]. However, all curds reached the value comprised in the range 5.20-5.40, which represents the optimum level of acidity allowing the stretching of the curd [33], about 24 h after curdling ( Figure 2).

Microbiological Analyses
The microbiological loads of the samples collected during the cheese-making trials are reported in Table 2. The microbiological analyses of GPP did not evidence the presence of any of the microbial groups object of investigation. The absence of microorganisms in GPP is undoubtedly due to the oven-drying stabilization [34]. Raw ewes' milk hosted levels of TMM of 6.51 Log CFU/mL, higher than the limit (<10 5 CFU/mL) established in Europe for raw ewes' milk for cheese production, with mesophilic rod and coccus LAB at 10 6 CFU/mL and thermophilic rod and coccus LAB at 10 4 CFU/mL. The high levels probably depend on the conservation conditions [35]. After pasteurization, TMM, thermophilic rod, and coccus LAB were recorded at 10 3 CFU/mL, while mesophilic

Microbiological Analyses
The microbiological loads of the samples collected during the cheese-making trials are reported in Table 2. The microbiological analyses of GPP did not evidence the presence of any of the microbial groups object of investigation. The absence of microorganisms in GPP is undoubtedly due to the oven-drying stabilization [34]. Raw ewes' milk hosted levels of TMM of 6.51 Log CFU/mL, higher than the limit (<10 5 CFU/mL) established in Europe for raw ewes' milk for cheese production, with mesophilic rod and coccus LAB at 10 6 CFU/mL and thermophilic rod and coccus LAB at 10 4 CFU/mL. The high levels probably depend on the conservation conditions [35]. After pasteurization, TMM, thermophilic rod, and coccus LAB were recorded at 10 3 CFU/mL, while mesophilic rod and coccus LAB were lower than 10 2 CFU/mL, showing the ability of the indigenous thermoduric LAB to survive during the thermal treatment [36]. However, the effect of pasteurization depends on the initial cell densities of microorganisms in the raw milk [37] and the temperatures applied [38]. According to Tukey's test, statistically significant differences were found only for the levels of TMM in curd samples. The levels of TMM were almost superimposable to those of mesophilic coccus LAB in all samples analyzed. After inoculation with each NMSC, all milks showed approximately 7 Log cycles of TMM and almost the same levels of mesophilic coccus LAB, confirming that the inoculums occurred at 10 7 CFU/mL and the starter LAB cocci dominated the microbial community of cheeses. After coagulation, an increase of about 1 Log cycle was registered for the LAB levels in all control and experimental curds as a consequence of whey draining [39]. The levels of TMM and mesophilic coccus LAB reached values of about 9 log CFU/g in all acidified curds and these levels remained almost constant in control and experimental cheeses soon after production as well as after 15 d of refrigerated storage as previously observed by Gaglio et al. [40,41] during the production of Vastedda della valle del Belìce PDO cheeses. The results highlighted that the addition of 1% (w/w) of GPP did not influence the fermentation process carried out by the four L. lactis (Mise36, Mise94, Mise169, and Mise190) strains used individually.

Persistence of LAB Inoculums
Three hundred and ninety-six colonies of presumptive LAB were isolated during cheese making, from pasteurized bulk milk until stored cheeses. After Gram and catalase tests, 360 Gram-positive and catalase-negative cultures were still considered presumptive LAB cultures. The determination of starter persistence by means of RAPD patterns comparison is a common technique to monitor the dynamics of the added strains and their dominance over indigenous bacteria during cheese productions [33,40]. Figure 3 shows the dendrogram with only 44 of the 360 LAB strains analyzed, which are those isolated at least once from the different samples.   Four major RAPD clusters were identified. Each cluster included one L. lactis strain used for the preparation of the NMSCs. Four LAB strains (BK286, BK291, BK296, and BK299) collected from pasteurized bulk milk before inoculum were subjected to the 16S rRNA gene sequencing and identified as Enteroccus faecium (Ac. No. MW283889-MW283890), Leuconostoc mesenteroides (Ac. No. MW283891) and Leuconostoc pseudomesenteroides (Ac. No. MW283892). These species are part of the common dairy starter LAB and non-starter LAB cultures [42]. However, these strains were not detected in any of the samples analyzed after milk inoculation with NMSCs, highlighting the dominance of the selected L. lactis strains (Mise36, Mise94, Mise169 and Mise190) over the indigenous thermoduric milk LAB (Figure 2).

Physicochemical Analyses of Cheeses
The physicochemical parameters of the final cheeses are reported in the Table 3. As expected, the inclusion of GPP modified markedly both external and internal color indexes. Indeed, the experimental cheeses acquired an evident pink color, as indicated by the strong redness increase, to which corresponded the reduction of lightness and yellowness, whereas the indexes recorded on control cheeses were in line with those reported for ovine stretched cheeses by Todaro et al. [43].
The chemical composition of cheeses was affected by both GPP inclusion and the starter cultures. GPP are characterized by lower lipid content than milk determining a decrease of fat and, consequently, an increase of protein in GPP-enriched cheeses. The increase of ash can be referred to the higher ash level of GPP compared to milk. Similar trends for lipid, protein, and ash concentrations were observed by Marchiani et al. [7] in cheeses fortified with GPP of different origin. Regarding the starter strains, a different behavior was observed for L. lactis MISE94 and MISE190 that in experimental cheeses induced a stronger fat decrease, which was accompanied by a protein increase, suggesting a more intense recourse to fatty acids to sustain their energy metabolism [44]. Nevertheless, the levels recorded for each chemical component were within the ranges observed in other investigations for Vastedda cheeses [43].
GPP inclusion was also responsible for the hardness increase of the cheese paste, evaluated as resistance to compression, presumably due to the effect of GPP in increasing DM or lowering fat content in experimental cheeses; however, in accordance to the results of sensory evaluation, the difference for hardness was almost negligible when the strain L. lactis MISE36 strain was used, an aspect that can be referred to the lower difference in fat content together with a similar DM level between the corresponding control and GPP-enriched cheeses.
Moreover, higher TBARS values were observed in experimental rather than control cheeses, indicating that the presence of GPP did not inhibit the lipid oxidation, despite their interesting content in phenolic compounds with antioxidant activity [7,44]. Since the experimental cheeses were lower in fat, this higher presence of oxidation products could be linked to their enrichment with the polyunsaturated fatty acids that characterizes the lipid profile of GPP [44] and are more sensitive to oxidation.

Volatile Organic Compounds
Volatile profiles are generated by complex biochemical processes derived from the hydrolysis or metabolism of carbohydrates, proteins, and fats, along with compounds added during processing or directly from the milk due to the activity of LAB [45]. VOCs were identified in cheese samples and GPP by SPME-GC/MS technique (Table 4). A total of 21 VOCs were detected in GPP, whereas 21 and 14 VOCs were identified in cheese samples with and without GPP addition, respectively. Volatile compounds belonging to alkanes, aldehydes, monoterpenes, esters, acids, and ketones were identified in all cheese samples. Acetic acid and 2-hydroxy4-methyl pentanoic acid were the most abundant acids detected in cheese sample without GPP addition. Acetic acid may be produced by carbohydrate catabolism by LAB, 2-hydroxy-4-methyl pentanoic acid is formed enzymatically from the corresponding amino acid (L-leucine) [46]. In addition, the free fatty acids (FAA) identified in this study were hexanoic, butyric, and 3-methyl-butanoic acid. Butyric and acetic acids were the main acids identified in cheese sample with GPP addition. FFA contribute to the formation of cheese flavor directly and as precursors of ketones, alcohols, aldehydes, and esters [47,48]. Similar FFA profiles were also observed in other cheeses produced from sheep's milk [41,46]. Two aldehydes (hexenal and heptanal) and one ester (octanoinc acid ethyl ester) were detected in the experimental cheese samples. No ester was identified in the control cheeses. Aldehydes and esters are poorly represented in Vastedda cheeses [49], probably because they are oxidized to acids or reduced to nalkanols due to the enzymatic activities of microorganisms [50]. Isoamyl alcohol, 2-butanol, 2-pentanol, and phenyl-ethyl-alcol were the only alcohols detected in cheese samples with isoamyl alcohol being the major alcohol present in all cheeses. A total of eight hydrocarbons and terpenes were emitted from the cheeses. Terpenes and hydrocarbons originated from the secondary metabolism of plants [51], therefore they can be rapidly transferred into milk fat by forages or other substances added to cheese. In light of the GPP VOCs profile, we can assess that the addition of GPP contributed to increase the amount of isoamyl alcohol and to determine the presence of D-limonene and 2-phenylethanol in the experimental cheeses.

Sensory Test
All cheeses, after 15 days of refrigerated storage, were subjected to the sensory analysis and the results comparing control and experimental samples per each L. lactis strain are reported in Figure 4. The addition of GPP modified the sensory attributes of cheeses. Generally, the addition of fibers or by-products exert a strong effect on the sensory parameters of dairy products [37,52]. In this study, except for sweet and bitter that scored similar values in all trials, all other attributes evaluated were consistently influenced by the addition of GPP. In detail, the addition of GPP increased odor and aroma intensity, acid perception, fiber sensation, friability, adhesiveness, and humidity while influenced negatively sweet and hardness attributes. Similar results were observed by Costa et al. [28] who tested white and red wine grape pomace to fortify bovine Primosale cheese. The selected LAB added as starter cultures determined different effects on the sensory attributes of cheeses (Figure 4a-d). The cheeses produced with the strain Mise94 (with and without GPP) were mostly appreciated by the judges.

Functional Properties
GPP is a rich source of polyphenols, which are the most abundant antioxidants in the human diet with potential health benefits [53]. The antioxidant capacity of GPP-added cheese was investigated only for the cheeses mostly appreciated by the judges; for this reason, ECh 15 94 and CCh 15 94 were analyzed through ABTS + radical cation decolorization assay. Both cheeses were submitted to the in vitro gastro-intestinal digestion and the ability of reduction of the radical was evaluated from the different digestion steps (Table 5).
Post-oral fraction of the control cheese showed reducing activity that did not significantly (p > 0.05) vary after gastric digestion (0.202 ± 0.012 µmol TE/g). Milk components, mainly high molecular weight casein, are considered responsible for the antioxidant capacity of the dairy products [54]. Moreover, it was reported that pepsin digestion does not increase radical scavenging activity of purified casein [55]. After intestinal digestion, the antioxidant activity of control CCh 15 94 cheese was about 50% higher than that measured after gastric phase, possibly due to the antioxidant fat-soluble vitamins released in the bile salts micelles or to the amino acids with reducing ability [56,57]. When GPP-added cheese (ECh 15 94) was submitted to the in vitro digestion, the antioxidant potential of the post-oral fraction was not recorded different from that of the relevant fraction of control cheese (p > 0.05), indicating that dilution and homogenization in simulated saliva fluid did not cause polyphenols release (Table 5). Interestingly, a net rise of the reducing activity (+41%) was measured in the samples after gastric digestion. As reported by Tagliazucchi et al. [58], casein digestion by pepsin can affect the binding with incorporated polyphenols, resulting in their solubilization. The antioxidant activity of ECh 15 94 cheese increased also after the intestinal step (Table 5). Finally, all reducing components released through the digestion of both cheeses were solubilized in the bioaccessible fraction, i.e., the soluble fraction of the digesta available for absorption (Table 5). Overall, during the digestion process, with the exception of the oral phase, all fractions obtained from the digestion of the GPP containing cheese showed a much higher antioxidant capacity than GPP-free cheese. The reducing activity of the bioaccessible fraction of ECh 15 94 accounted for 175% of that measured in the corresponding fraction of CCh 15 94. Our results demonstrated the important contribution of the polyphenols of GPP to the antioxidant capacity of the cheese and highlighted the functional potential of the fortified product. Unregulated food lipids oxidation during digestion can lead to a postprandial oxidative stress condition, which negatively affects the human health [59]. Consequently, antioxidants contained in foods that are able to react with and detoxify lypoperoxyl radicals can be considered of interest to reduce the oxidative phenomena during digestion. Peroxyl radical scavenger activity of the bioaccessible fractions of the cheeses was compared against lipid oxidation of pig brain microsomes. When lipid peroxidation was induced in microsomes by AAPH in the absence of cheese, TBA-RS production started immediately reaching a maximum at 90 min incubation (Figure 4). Bioaccessible fraction from 0.1 g or 0.2 g of CCh 15 94 slightly slowed down TBA-RS formation in a dose-dependent manner. Interestingly, bioaccessible fraction from GPP-enriched cheese (ECh 15 94) caused a stronger dose-dependent inhibition of the lipid peroxidation, evidenced by the net delay in the initial TBA-RS formation and a higher reduction of TBA-RS production throughout the incubation period ( Figure 5).
These results demonstrated that GPP caused 60% increase of the lipoperoxyl radical scavenger capacity of the cheese measured after digestion, conferring a potential healthy value against postprandial oxidative stress.

Functional Properties
GPP is a rich source of polyphenols, which are the most abundant antioxidants in the human diet with potential health benefits [53]. The antioxidant capacity of GPP-added cheese was investigated only for the cheeses mostly appreciated by the judges; for this reason, ECh1594 and CCh1594 were analyzed through ABTS + radical cation decolorization assay. Both cheeses were submitted to the in vitro gastro-intestinal diges-  CCh 15   These results demonstrated that GPP caused 60% increase of the lipoperoxyl radical scavenger capacity of the cheese measured after digestion, conferring a potential healthy value against postprandial oxidative stress.

Conclusions
This study provided, for the first time, an extended analysis of the microbiological, physicochemical, sensory, and functional aspects of GPP-enriched ovine cheese. The addition of GPP did not alter the microbiological parameters during the fermentation carried out with four strains of L. lactis used as single inoculums. The chemical composition of cheeses was affected by GPP addition. GPP-enriched cheeses were characterized by lower fat content, higher protein content, and higher values of secondary lipid oxidation. The differences in VOCs detected between control and experimental cheeses are a direct consequence of GPP addition. Sensory evaluation indicated that the GPP-enriched cheeses were characterized by a general appreciation by judges and, in particular, the higher values of overall acceptance was detected for the cheeses produced with the strain MISE94. In particular, from the functional point of view, the cheeses enriched with GPP, submitted to simulated human digestion, showed an increased antioxidant activity and lipoperoxyl radical scavenger capacity. Degradation of dairy matrix appears a requisite to release the incorporated polyphenols in GPP-enriched cheese and increase the reducing activity of the intestinal digesta. However, due to the importance of using mixed strain starters to ensure the fermentation process, further studies are necessary to validate this technology for industrial applications.

Conclusions
This study provided, for the first time, an extended analysis of the microbiological, physicochemical, sensory, and functional aspects of GPP-enriched ovine cheese. The addition of GPP did not alter the microbiological parameters during the fermentation carried out with four strains of L. lactis used as single inoculums. The chemical composition of cheeses was affected by GPP addition. GPP-enriched cheeses were characterized by lower fat content, higher protein content, and higher values of secondary lipid oxidation. The differences in VOCs detected between control and experimental cheeses are a direct consequence of GPP addition. Sensory evaluation indicated that the GPP-enriched cheeses were characterized by a general appreciation by judges and, in particular, the higher values of overall acceptance was detected for the cheeses produced with the strain MISE94. In particular, from the functional point of view, the cheeses enriched with GPP, submitted to simulated human digestion, showed an increased antioxidant activity and lipoperoxyl radical scavenger capacity. Degradation of dairy matrix appears a requisite to release the incorporated polyphenols in GPP-enriched cheese and increase the reducing activity of the intestinal digesta. However, due to the importance of using mixed strain starters to ensure the fermentation process, further studies are necessary to validate this technology for industrial applications.