3.2. Cheese Physico-Chemical Traits
Table 2 shows the physico-chemical parameters of cheeses manufactured from individual milk. Cheese yield tended to be lower with milk from FSFL ewes, despite its higher casein content. This result can be attributed in part to the higher milk yield with FSFL diet that, in comparison with the other diets, contributed to the reduction of milk fat (4.63%,
Table 1), although at a non-significant level; in this regard, the superior casein content of FSFL (4.03%,
Table 1) was not able to balance the lower fat. However, also the higher water loss, indicated by the higher DM level recorded at a non- significant level in FSFL cheeses (54.72%), could have contributed to reducing their cheese yield. Nevertheless, the higher protein content of FSFL cheeses (46.21% DM), which resulted in a significant difference (
p = 0.0488) in comparison with SHL and DSF2 cheeses, seems to reflect the higher casein content of the corresponding milk.
In general, the cheese components with antioxidant effect were detected in major levels in cheeses from diets based on green forage than from diets based on dried forages. In particular, vitamin A showed the highest content in cheeses from FSFL and intermediate levels with FSF4 and DSF2 diets, whereas vitamin E was higher in cheeses from DSF2, FSFL and FSF4 diets. It can be noticed that with dehydrated forage the cheese content of vitamin E was almost double than that with SHL, and comparable to that with FSF4 and FSFL, in line with the content of vitamin E in the different forages (4.96, 23.63 and 22.81 mg/kg DM in SH, DSF and FSF, respectively) reported in Gannuscio et al. [
28]; these results demonstrate as the dehydration process, contrarily to hay-making, did not alter the presence of vitamin E, and evidence also as the presence of vitamin E in dairy products is strictly related to the ingested amount by diet.
Total polyphenols resulted in higher levels in FSFL than in SHL cheeses, whereas intermediate contents were observed with the other diets. The same trend was not found for CT, which were higher with FSF4 diet than with DSF2 diet; however, despite the high intake and digestibility recorded with the FSF4 and FSFL diets characterized by the higher proportions of FSF, as observed in Gannuscio et al. [
28], the levels detected in cheeses from all diets were very low when compared to the respective total polyphenols ingested by ewes. This aspect suggests that CT of sulla forage could be degraded at ruminal level and metabolized along the intestine, thus modified with respect to their original structure; in this way, the derived metabolites, of lower molecular weight, are carried by circulation and incorporated into milk [
4,
56] where are presumably detected among polyphenols. Thus, the presence of polyphenols in cheeses, due to the transfer of dietary tannic or non-tannic compounds and/or their metabolites to milk, demonstrates a certain degree of their bioaccessibility by which it is possible to exploit their antioxidant properties, in accordance with other authors [
4,
9,
22].
The antioxidant capacity of cheeses, expressed as TEAC, was higher with all diets based on green forage and comparable between cheeses obtained using dried forages. This improvement in the antioxidant activity induced by FSF can be attributed to the synergy effect of vitamins A and E and polyphenols. Whereas the only vitamins content in DSF2 cheeses was not able to induce a marked improvement of their antioxidant capacity, although the latter was statistically comparable to that of FSFL cheeses.
The oxidative stability of cheese fat was assessed by determining POV and TBARS as indexes of primary and secondary lipid oxidation, respectively. POV was lower in cheeses manufactured with milk from ewes fed with SHL and FSFL diets; presumably, the better oxidative stability of SHL cheeses was due to their lower PUFA content, as discussed later, whereas that of FSFL cheese can be explained by their higher antioxidant protection exerted by vitamins and polyphenols. Instead, as emerged from the higher TBARS value, the secondary oxidation interested more the SHL cheeses, which were less rich in antioxidant molecules.
Marked effects of the diet were found for all colour parameters detected on the external surface of cheeses: lightness (L*) was more apparent in the cheeses obtained with FSF2 diet, while the cheeses produced with DSF2, FSF4 and FSFL diets showed higher values of yellowness (b*) and corresponding lower values of redness (a*), indicating shifts towards yellow and green colorations, respectively. These results may be attributable to the transfer in cheeses of carotenoid pigments [
57], especially lutein which, according to Rufino-Moya et al. [
58], is presumably present in greater amounts in the diets that contain higher levels of sulla forage, either fresh or dehydrated. The detection of internal colour on the cheese paste revealed a single difference for redness (a*), lower in cheeses produced with FSFL diet; similarly for the external surface, the lower redness of FSFL cheeses corresponded to a higher yellowness, although the latter did not differ significantly, and may also be linked to the major content of carotenoids derived from the green forage in the diet.
The greater hardness, expressed as resistance to compression, was found in cheeses obtained with FSFL diet and reflects the major consistency of their paste, justified by their lower humidity and then by their higher DM percentage, as previously evidenced.
Parity did not influence the physico-chemical parameters of cheese, with the exception of hardness, higher in cheeses from FL ewes, presumably linked to the concomitant effects of their higher DM and lower fat, although both parameters were not significantly different.
3.3. Cheese Fatty Acid Profile
The fatty acid composition of cheese, reported in
Table 3,
Table 4 and
Table 5, was greatly influenced by the diet, whereas was not affected by parity.
Regarding the short and medium chain FA (
Table 3), the higher levels of C6:0, C8:0, C10:0, C12:0 and C14:0 acids were recorded with the diet based on DSF, while the C16:0 was lower in the diet with exclusive FSF provided ad libitum. As known, FA with short- and medium-chain, and partially the C16:0, are synthesized ex novo in the udder tissues from acetic acid, that is their precursor formed in the rumen by the microbial fermentation of cellulose; their major presence in the DSF2 cheeses can be related to the high fiber intake of ewes fed DSF, favoured by the high dotation of dehydrated forage in NDF and cellulose characterized by a smaller encumbrance, as evidenced in Gannuscio et al. [
28].
Among long-chain FA (
Table 4), the stearic acid (C18:0) was not affected by the diet, whereas C20:0 and C22:0 acids were higher with the diet FSF4. It has been widely recognized that dietary saturated FA (SFA) contributes to enhancing serum cholesterol and, consequently, the risk of cardiovascular diseases (CVD) [
50]. According to Santos-Silva et al. [
52], the hypercholesterolemic effect is absent for the stearic acid (C18:0), low for lauric (C12:0) and palmitic (C16:0) acids, and higher for the myristic acid (C14:0). However, it has to be noticed how current evidence based on most recent data meta-analyses do not support the recommendations encouraging high consumption of n-3 FA and PUFA and suggesting to limit the saturated fat intake to prevent the risk of CVD [
58,
59]; particularly, high intake of SFA has been linked to a reduced risk of stroke [
60], whereas no association was found between consumption of whole-fat dairy products and increased risks of CVD [
61].
The cheeses obtained with SHL diet showed greater single (
Table 3: C13:0 iso, C15:0 iso, C15:0 anteiso, C17:0 iso) and total amounts (
Table 5) of branched-chain FA. These FA are synthesised by ruminal bacteria and their presence is conseguence of their activity, favoured by a higher forage:concentrate ratio and then by the fiber level in the diet [
62]; this explains as the higher content of branched-chain FA was recorded for the SHL group that ingested highly lignified fiber [
28]. The branched-chain FA showed to have cytotoxic effects on breast cancer cells [
63] and, in particular, the form C15:0 iso exerts an inhibitory action inducing apoptosis in cancer cells as those of prostate cancer, leukemia and breast adenocarcinoma [
64]; this antitumoral activity is considered to be equal to that recognized for CLA [
62].
The oleic acid (C18:1 c9), the main of monounsaturated FA (MUFA), as well as C16:1 c9, showed higher levels with the SHL diet, whereas the other C18:1 cis FA resulted higher with FSFL diet; since oleic acid derives also from the rumen biohydrogenation of PUFA, its increase, together with that of branched chain FA, can be attributed to the low content of CT in the sulla hay that was not able to limit the activity of microflora in the rumen, as presumably occurred with the other diets containing CT from FSF.
The FSFL diet induced in cheeses the increase of PUFA (
Table 4), until now interesting for their presumed health benefits, such as linoleic acid (LA, C18:2 n-6), ALA (C18:3 n-3), and RA (C18:2 c9 t11) together with the trans vaccenic acid (TVA, C18:1 t11) that is its precursor in the rumen; these changes in the FA profile of cheese fat can be attributed to the high presence of PUFA in FSF, as well as to the inhibitory effect of CT on the rumen biohydrogenation of PUFA.
The TVA is the precursor of RA, and represents the intermediate product of biohydrogenation of PUFA, especially LA and ALA, carried out by ruminal bacteria [
65]. Indeed, the presence of TVA was favoured in cheeses from the diet based on exclusive FSF, due to its high content of both ALA and CT in FSF. Once again, the CT of FSF showed to inhibit partially the rumen biohydrogenation of PUFA, favouring the production of TVA. Since the RA in ruminants milk originates mainly from the enzymatic desaturation of TVA obtained by the activity of Δ9-desaturase in the udder tissues [
65], the content of TVA and RA is highly related, as confirmed by the major presence of both FA in the FSFL cheeses.
A significant increase in the ALA content in cheese occurred also with DSF2, that allowed to reach a level comparable to that with FSFL (
Table 4); this result was certainly due to the ALA level, which was higher in both fresh and dehydrated forage than in SH (6.24, 7.04 and 0,54 g/kg DM, respectively; [
28]). However, the ALA increase in DSF2 cheese did not correspond to an increase in TVA and RA, as occurred in FSFL cheeses. This result can be attributed to the reduction of CT to which FSF was subjected during dehydration and pelleting (from 17.9 to 5.4 g delphinidin equivalent (DE)/kg DM; [
28]) as a consequence of the effect of heater temperatures [
66]; indeed, the lower CT content detected in DSF can explain its moderate effect in inhibiting the ruminal biohydrogenation and forming TVA, and then RA.
The RA. the main and the most abundant among the isomers of CLA, can be synthetised in the rumen for isomerization of LA, but the most part is formed by desaturation of TVA in the tissues of ruminants, then also in the udder during milk secretion; thus, RA is present mainly in the fat of meat and dairy products from ruminants. In the last years, the RA has been the subject of interest for its health properties; in particular, it is recognized for its antitumoral activity, but is retained active in the prevention of atherosclerosis, therefore against the onset of CVD, by reducing cholesterolemia (LDL, low-density lipoprotein) and the triglycerides plasma levels, and shows also immunomodulatory and antidiabetic functions, reduces oxidative stress, and contributes to osteogenesis and in the control of obesity [
67].
Among the factors affecting the RA content in milk, the feeding of dairy animals is determinant since it provides the precursors (LA and ALA) from which it derives; accordingly, feeding based on fresh forage with high dotation of PUFA, in particular ALA, is known to be responsible for the major RA enrichment of dairy products [
68,
69]. However, since FSF is rich in ALA as well as in CT, this explains because the FSFL diet, in which the forage component is exclusively represented by FSF, was associated to the higher level of TVA and RA in cheese, other than to the marked increase of total PUFA; indeed, the inhibiting action of PUFA biohydrogenation exerted by CT allowed the PUFA passage and absorption in the intestine, and then the transfer in milk and cheese.
Mainly, the exclusive presence of FSF in the diet improved the FA profile of cheeses in relation to health quality, at least until the current guidelines will be reassessed in accordance with the new scientific evidence about the effective role of SFA and PUFA intake in favouring or preventing the risks of CVD [
61]. In this investigation, the FSFL diet contributed to reducing SFA and raising PUFA, thus determining a marked improvement of PUFA/SFA and n-6/n-3 ratios, as well as of the estimated health indexes (
Table 5). In particular, the PUFA/SFA ratio of FSFL cheeses increased towards the threshold of 0.45 recommended for foods to control the level of serum cholesterol to preserve human health [
70].
Instead, DSF determined a slight increase in SFA, for the contribution of C10:0 that has no effect on human health, and implied also a reduction in MUFA, due to the reduction of oleic acid; moreover, DSF induced a level of PUFA lower than that with FSFL diet and analogous to those of other diets, but greatly reduced the n-6/n-3 ratio and the corresponding ratio based on their precursors, LA/ALA, for the effect of the high level of total n-3 FA, especially ALA. However, with diets based on DSF and FSF, the level of n-6/n-3 ratio was always below the threshold (<5) recommended by FAO/WHO [
71] in the human diet to prevent and treat chronic diseases, whereas this threshold was highly exceeded with the SHL diet (6.35).
With regard to the other estimated health indexes (
Table 5), the FSFL diet favoured the reduction of the thrombogenic index, and the raising of the Health Promoting Index, while the h/H index was maximum with SHL diet, the latter due to the high content of oleic acid in the corresponding cheeses. The DSF worsened the Health Promoting Index of cheese fat, which reached the minimum level, but induced values of thrombogenic and h/H indexes which were similar to those of FSF2 and FSF4 cheeses.
The GHIC showed a progressively increasing trend passing from diets based on dried forages to diets with fresh forages, with the maximum value recorded in cheeses from ewes fed FSF ad libitum. This trend is in accordance with Giorgio et al. [
53] who proposed the GHIC that, associating the contribution of FA (RA, total PUFA and n-3 PUFA), polyphenols and antioxidant capacity (TEAC), is able to score specifically the health-promoting value of cheeses obtained from animals fed pasture or fresh forage. The GHIC values of cheeses were above the range (16–23) recorded for sheep cheeses by Di Trana et al. [
72], with exception of that of SHL cheeses, confirming the fundamental role of fresh forage-based diets to enhance the health properties of cheeses.
Finally, in FSFL cheeses, the Δ-9 desaturase indexes of C14:1 and C16:1 were lower with slight differences, whereas the index related to TVA desaturation to form RA was markedly lower, and presumably associated to the higher content in TVA.
3.4. Cheese Biomarkers of Animals’ Feeding Regime
The components of cheeses, as well as their groupings or ratios which resulted significantly affected by the diet, were tested for their ability to differentiate cheese for the feeding regime of animals. Accordingly, linear and quadratic regressions were performed using data of fresh forage intake (% diet DM), reported by Gannuscio et al. [
28], and the cheese traits from SHL, FSF2, FSF4, FSFL experimental groups and farming ewes fed pasture-based diet.
Table 6 reports the relationships of those 7 parameters of cheese better related (R
2 > 0.50) to ewes’ fresh forage intake. Among these traits, which are all comprised within the lipid fraction of cheese, the highest R
2 coefficients were obtained with the quadratic regressions of ALA (0.9435) and LA/ALA ratio (0.9213), represented in
Figure 1.
The ability of these 7 cheese components to be used as biomarkers of animals’ diet was explored developing their box plots, which allow to display their range in relation to the diets (
Figure 2).
Observing the box plots, it is possible to appreciate as the levels of ALA and LA/ALA ratio were more effective in distinguishing cheeses in relation to animals’ feeding regime. In particular, the levels of ALA in cheeses from SHL diet without fresh forage showed overlaps only with the levels of cheeses from FSF2 diet, whereas the levels emerged with higher fresh forage intake were above. The same trend can be observed for n-3 PUFA, but not for total PUFA and C17:0 anteiso.
An analogous effective discriminating ability emerged for the ratio LA/ALA, confirming also for this trait the robustness of the quadratic regression. The same potential was not detected for the n-6/n3 and PUFA/SFA ratios, whose levels in SHL cheeses showed no overlap only with the products obtained from ewes fed exclusively at pasture, corresponding to 100% of fresh forage intake.
Thus, both ALA and LA/ALA ratios have revealed a promising role as biomarker of cheeses produced from milk of animals fed fresh forage-based diets. Whereas also Maniaci et al. [
26] reported ALA among the promising biomarkers for traceability of cheese production season, Segato et al. [
25] did not find ALA among the predictors of the cheese production system.
Finally, it can be noticed also that a complete separation of cheeses obtained from DSF2 due to the levels of ALA and LA/ALA ratio occurred only with the cheeses from pasture. This circumstance represents further confirmation that dehydration may contribute to preserving the FA profile of green forage.