**3. Results**

#### *3.1. Bioavailability of Flavonoids*

Five flavonoids were quantified in the plasma of the subjects consuming H-ASP (450 mg GAE) or milk vehicle only. The baseline plasma concentration of catechin, quercetin, naringenin, kaempferol, isorhamnetin, and total flavonoid were 17.0 ± 4.0, 7.3 ± 1.2, 11.5 ± 5.7, 10.1 ± 3.2, 2.3 ± 0.4, and 48.1 ± 9.3 ng/mL, respectively. Milk alone did not significantly affect their values. The H dose led to significant increases in plasma catechin, naringenin, and sum of five flavonoids, as compared to the corresponding baseline value (*p* ≤ 0.05). Their maximum concentrations (Cmax) were 44.3 ± 15.6, 19.3 ± 8.2, and 82.3 ± 17.6 ng/mL and times to reach Cmax (Tmax) were 1.4 ± 0.2, 3.3 ± 0.5, and 1.7 ± 0.3 h, respectively (Figure 1). At 10 h, flavonoid concentrations returned to their respective baseline values. The other three measured flavonoids, quercetin, kaempferol, and isorhamnetin, werenotsignificantlyincreasedbytheH-ASP.

**Figure 1.** *Cont.*

**Figure 1.** Time course of plasma flavonoids, catechin ( **A**), quercetin (**B**), naringenin ( **C**), kaempferol ( **D**), isorhamnetin (**E**), and total (**F**) in older adults after acute intake of skim milk vehicle (C) and 450 (H) mg GAE ASP. The data are presented as the percent change from the respective baseline (0 h). The baseline concentrations of catechin, quercetin, naringenin, kaempferol, isorhamnetin, and total were 17.0 ± 4.0, 7.3 ± 1.2, 11.5 ± 5.7, 10.1 ± 3.2, 2.3 ± 0.4, and 48.1 ± 9.3 ng/mL, respectively. Values are expressed as mean ± SEM, *n* = 7. Means with a mark are significantly different from that of the baseline, *p* ≤ 0.05, tested using PROC GLM, followed by Tukey's honestly significant difference (HSD) multi-comparison test. ASP: almond skin polyphenols; GAE: gallic acid equivalents.

#### *3.2. Changes in Plasma Biomarkers of Antioxidant Defense*

Mean baseline plasma values of GSH, GSSG, and their ratio were 1.16 ± 0.14, 0.11 ± 0.01 μmol/L, and 12.1 ± 2.4, respectively. A marked inter-individual variation in GSH, GSSG, and GSH/GSSG was noted. After consumption of skim milk, GSH values tended to decrease, GSSG values tended to increase, and the ratio remained unaltered (Figure 2). H-ASP tended to increase GSH by 25% at 3 h as compared to the respective baseline value and to decrease GSSG by 31% at 2 h. A favorable effect of H-ASP on GSH status was noticeable at 15 min post-consumption but retreated at 45 min and 1 h before the next favorable changes occurred at 2 and 3 h. At 3 h, the GSH/GSSG ratio of H-ASP was 212% of the baseline, which was significantly different from 82% of C at the same time point (*p* = 0.0033). The increased ratio at 3 h after H-ASP consumption was driven by both the increased GSH and the decreased GSSG. The AUC of GSH and GSSG did not differ between C and H-ASPs; the AUC of the GSG/GSSG ratio was significantly increased by 70% by H-ASP as compared to C (Table 1).

**Figure 2.** *Cont.*

**Figure 2.** Time course of plasma glutathione (GSH) (**A**) and oxidized GSH (GSSG) (**B**) and their ratio (**C**) in older adults after acute intake of skim milk vehicle (C) and 450 (H) mg GAE ASP. The data are presented as the percent change from the respective baseline (0 h). Baseline values of GSH, GSSH, and their ratio were 1.16 ± 0.14, 0.11 ± 0.01 μmol/L, and 12.1 ± 2.4, respectively. Values are expressed as mean ± SEM, *n* = 7. Means not sharing the same letter at the same time point are significantly different, *p* ≤ 0.05, tested using PROC GLM, followed by Tukey's HSD multi-comparison test. *p*-values of the dose effect for GSH, GSSG, and their ratio were 0.029, 0.013, and 0.014, respectively.

**Table 1.** The percent change of area under curve (AUC) of plasma glutathione, glutathione peroxidase activity, and oxygen radical absorbance capacity (ORACpca) in older adults after acute intake of 250 mg (L) or 450 mg (H) almond skin polyphenols (ASP) or skim milk (C) 1. GPx: glutathione peroxidase.


1 Values are mean ± SEM, *n* = 7. a,b Means in the same column not sharing the same letter differ, *p* ≤ 0.05, using an ANOVA *t*-test, followed by Tukey's HSD test. 2 Plasma was treated with perchloric acid (PCA) first before testing by ORAC assay. 3 Percent change from the respective baseline value (0 h) of each visit is calculated to construct the area under the curve (AUC) using the linear trapezoidal integration.

Mean baseline plasma GPx activity was 197 ± 11 U/L. Skim milk did not affect GPx activity. L- and H-ASP up-regulated GPx activity in a two-phase mode with an initial increase occurring between 15 and 45 min, followed by the second one at 2 h (Figure 3). The effect of ASP on GPx activity was independent of the dose. The magnitude of the ASP-induced increase in GPx activity ranged between 26% and 35% from the baseline at 15, 30, and 45 min and 2 h, while skim milk slightly increased the activity during the same period. The AUC of GPx activity did not differ among three treatments (Table 1). Plasma ORAC value was not affected by ASP and milk up to 10 h. Mean baseline ORACpca value was 896 ± 28 μmol/L TE (Table 1).

**Figure 3.** Time course of the percent change of plasma GPx activity in older adults after acute intake of skim milk vehicle (C), 225 mg (L), or 450 mg (H) GAE ASP. The data are presented as the percent change from the respective baseline (0 h). The mean baseline activity was 197.0 ± 10.6 U/L. Values are expressed as mean ± SEM, *n* = 7. Means not sharing the same letter at the same time point are significantly different, *p* ≤ 0.05, using PROC GLM followed by Tukey's HSD multi-comparison test. *p*-Values for dose and time effect and their interaction were 0.039, ≤0.001, and 0.001, respectively.

#### *3.3. Changes in LDL Resistance to Oxidation*

The mean baseline lag time of LDL oxidation was 45.1 ± 1.7 min, and its value was not extended by either ASP dose. The AUC of lag time was comparable between treatments. The addition of 6 μmol/L α-tocopherol prior to the initiation of Cu2+-induced LDL oxidation increased lag time to 95.7 ± 2.6 min at baseline (Figure 4). At 3 h, the lag time with added α-tocopherol after intake of L- and H-ASP was 144.7 ± 13.1 and 165.2 ± 25.0% of that at baseline, respectively, as compared to the 102.2 ± 2.4% observed after the skim milk (*p* ≤ 0.05). There was no difference in the lag time between the two ASP doses. To reduce the influence of the variation of the baseline values, the percentage change from the baseline was calculated to assess the change obtained with in vitro addition of α-tocopherol.

**Figure 4.** Time course of lag time of LDL oxidation with in vitro addition of 6 μmol/L α-tocopherol in older adults after acute intake of skim milk vehicle (C), 225 mg (L), or 450 mg (H) GAE ASP. The data are presented as the percent change from the respective baseline (0 h). The mean baseline lag time was 95.7 ± 2.6 min. Values are expressed as mean ± SEM, *n* = 7. Means not sharing the same letter at the same time point are significantly different, *p* ≤ 0.05, tested using PROC GLM followed by Tukey's HSD multi-comparison test. *p*-Values of the dose and time effect and their interaction were 0.001, 0.014, and 0.049, respectively.

#### *3.4. Changes in Plasma Biomarkers of Oxidative Stress*

The mean baseline plasma MDA value was 2.4 ± 0.3 μmol/L. Skim milk and H-ASP did not affect MDA up to 10 h (data not shown) or its AUC (Table 2). Mean baseline plasma F2α-isoprostanes were 5.0 ± 0.2 ng/mL. Given that there were marked inter-individual variations, F2α-isoprostanes in plasma (data not shown) and their AUC (Table 2) were not affected by skim milk and H-ASP.

**Table 2.** The percent change of area under curve (AUC) of oxidative stress biomarkers in older adults after acute intake of 250 mg (L) or 450 mg (H) almond skin polyphenols (ASP) or skim milk (C) 1.


1 Values are mean ± SEM, *n* = 7. a,b Means in the same column not sharing the same letter differ, *p* ≤ 0.05, using an ANOVA *t*-test, followed by Tukey's HSD test. 2 α-Tocopherol (6 μmol/L) was added to LDL before the initiation of oxidation; 3 Percent change from the respective baseline value (0 h) of each visit is calculated to construct the area under the curve (AUC) using the linear trapezoidal integration.
