**4. Discussion**

From production and processing to storage, retailing, and consumption, waste is produced in all the phases of food life cycle [10]. According to the Food and Agriculture Organization of the United Nations [35], approximately one-third of the food produced in the world for human consumption is lost or wasted annually. For example, ~30% of nonedible products of vegetables and some fruits, mainly skins and seeds, are commonly wasted and discarded [9]. Similarly, almonds, the most commonly consumed tree nut, can generate ≥4% waste as skin during the production of blanched almonds. This type of food waste has historically been used as low-value livestock feed or compost. More recently, the agri-food industry has made substantial progress in utilizing waste by-products for development of novel ingredients or products [36,37]. Successful examples include the recovery of oil from olive kernel, the production of essential oils, flavonoids, and pectin from citrus peel, and the recapture of protein concentrates from cheese whey [12]. Some of these plant-based wastes contain a variety of phytochemicals, particularly polyphenols, which are often abundant in skins and seeds [9]. For example, almond skins contain an array of flavonoids and phenolic acids, with isorhamentin-3-rutinoside being the main polyphenol [13,17]. The result of several experimental studies suggests the potential for utilizing this by-product of almond processing as a value-added ingredient useful in the development of functional foods or nutraceuticals because of its antioxidant, anti-microbial, anti-viral, neuroprotective, photoprotective, and/or prebiotic activities [13–15,23,24,38–42]. Within the last decade and with availability of new equipment, the almond blanching process has evolved to use substantially less water and more steam, reducing the loss of polyphenols into the blanch water and allowing a higher polyphenol content of the almond skin after processing. We demonstrate here the acute bioavailability and antioxidant actions of polyphenols derived from almond skins in humans without apparent untoward side effects.

Flavonoid bioavailability is dependent on a wide array of factors, e.g., type of flavonoid, food matrix, co-consumed food components, polymorphism of detoxification mechanisms, and aging. Using a simulated human digestion system, Mandalari et al. [43] demonstrated that ASP were bioaccessible for the absorption in the upper gastrointestinal tract. In this study, the subjects consumed ASP at the dose of either 225 mg or 450 mg GAE, and plasma flavonoids were monitored over 10 h. As previously characterized in almonds, isorhamnetin is one of its principal flavonoids [13,17]; however, we found only a modest trend toward an increase in its plasma concentration. This result is in contrast with the marked increase we found in our hamster study [24]. Nonetheless, our inability to detect isorhamnetin in plasma is consistent with the report by Bartolomé et al. [44] who found it undetectable at 2.5 h post consumption of 884 mg GAE total phenols of almond skin extract. The dose employed in that study was about twice our highest dose. Bartolomé et al. study [44] estimated this dose is ~8-times higher than the dietary intake (102–121 mg/person/d) of nut polyphenols in the Spanish diet. While reports on the clinical pharmacokinetics of isorhamnetin following the consumption of isorhamnetin-rich foods or supplements are limited, Schulz et al. [45] observed a significant increase in plasma isorhamnetin with Cmax in the 10-ng/mL range in 18 healthy men consuming one dose of 900 mg dry extract of St. John's wort. Besides isorhamnetin, we monitored catechin, naringenin, kaempferol, and quercetin values and found significant increases in plasma catechin and naringenin with the Tmax at 2 and 3 h, respectively. Similarly, Garrido et al. [46] reported that consumption of almond skin extract containing 884 mg GAE total phenols increased urinary excretion of epicatechin and naringenin conjugates derived from phase II metabolism. We also noted a marked inter-individual variation in the concentration of these flavonoids, which is consistent with a range of other reports that ascribe this phenomenon to the wide range of endogenous and exogenous factors mentioned above [47,48]. We did not examine urinary flavonoids or phenolic acids derived from bacteria-catalyzed flavonoids in this study. However, Llorach et al. [49] noted a plethora of phenolic acids in urine of 24 people consuming an ASP extract. This study underscores the significant role of gu<sup>t</sup> microbiota on the catabolism of flavonoids in the formation of phenolic acids from flavonols, e.g., hydroxyphenylvaleric, hydroxyphenylpropionic, and hydroxyphenylacetic acids. Together with low concentrations of detected flavonoids in plasma, future research is warranted to examine the effect of phenolic acids derived from polyphenols via catabolism of colonic microflora on human health.

Flavonoids are regarded as strong antioxidants, acting via scavenging (reducing) reactive oxidants, chelating transient metals, and/or modulating endogenous antioxidant defense mechanisms. However, the efficacy of such mechanisms of action post absorption has been questioned because of the low concentrations of flavonoids in blood and tissues as compared to other abundant endogenous antioxidants [50]. Previously, we demonstrated in hamsters that absorbed ASP enhance LDL resistance against ex vivo Cu2+-induced oxidation [24] and worked with the in vitro addition of vitamin E to further bolster the LDL resistance. Instead, we found the effect of the absorbed ASP on the protection of LDL against oxidation was only unmasked when α-tocopherol (in a dose-dependent manner) was added in a physiologically relevant concentration. Thus, absorbed ASP may incorporate into LDL particles and then exert antioxidative actions or/and stabilize LDL structure to enhance LDL resistance. This speculation is also based on our in vitro study showing sand α-tocopherol work in a synergistic manner to stabilize LDL conformation during oxidation [25]. While the magnitude of polyphenol bioavailability and circulating flavonoid metabolites vary between species, this study extends the putative benefit of ASP on LDL resistance to oxidation from hamsters to humans.

Flavonoids are regarded as a class of beneficial phytonutrients. However, they also display characteristics as xenobiotics and are subject to phase I, II, and III detoxification metabolism known to be involved in drug clearance from the body. In addition, they may modulate the expression of cytochrome P450 monooxygenases, phase II conjugation enzymes, and/or on membrane transporters [51]. These actions are likely attributed to the effect of flavonoids on activating xenobiotic response elements and/or antioxidant/electrophil response elements (AREs/EpREs) [51,52]. For example, via these signal transduction pathways, quercetin enhanced the expression and activities of GSH reductase, GPx, and catalase [53,54]. We determined plasma GSH and GPx activity in support of the effect of ASP on endogenous antioxidant defense systems, a result that could implicate the activation of ARE by the absorbed ASP. We found that the high-dose ASP tended to increase plasma GSH between 1 and 5 h after consumption as compared to the skim milk vehicle, suggesting the absorbed ASP either increased its production in the liver via up-regulating γ-glutamylcysteine synthetase [51] or decreased its utilization or excursion. Further, the GSH/GSSG ratio, regarded as an index of redox status, was elevated at 3 h post-consumption of the high-dose ASP as compared to the skim milk vehicle, as well as the significantly larger AUC of the ratio, substantiating the absorbed

ASP could boost antioxidant defense capacity. Similarly, there was an early, transient effect of the absorbed ASP on plasma GPx activity. Although GSH is the substrate for the GPx reaction in the reduction of H2O2 and lipid peroxides to water and alcohols, we did not find a correlation between plasma GPx activity and GSH, GSSG, or their ratio (data not shown). In a previous human study, we also found that polyphenol avenanthramides from oats enabled a similar effect on GSH and GPx in humans [28]. However, neither blackberry nor cranberry juice, both rich in anthocyanins, affected erythrocyte GPx activity [55,56]. Thus, the effect of polyphenols on GPx activity and GSH status appears to depend on sample type (plasma vs. erythrocytes), flavonoid class (anthocyanins vs. flavonols), and study design (chronic vs. acute). Establishing an efficient extraction protocol to produce a commercial, high-quality polyphenol-rich product from almond skins for human use remains a challenge, though use of almond skin powder can already be found in the marketplace [57].

There are some limitations to this study, including its small sample size and absence measures of the oxidative modification of DNA and proteins. The inclusion of only healthy older adults also limits the generalizability of the results. There were significant inter-individual temporal differences in the pattern changes of plasma flavonoids and biomarkers of oxidative stress. This common phenomenon suggests marked individual genetic variation in enzymes that influence polyphenol absorption, metabolism, disposition, and excretion. Future studies are warranted to better understand interactions between genomics and flavonoid metabolism and utilization and their impact on biomarkers of oxidative stress and health outcomes.
