Controlled Feeding of an 8-d, High-Dairy Cheese Diet Prevents Sodium-Induced Endothelial Dysfunction in the Cutaneous Microcirculation of Healthy, Older Adults through Reductions in Superoxide

ABSTRACT

Background

While excess dietary sodium impairs vascular function by increasing oxidative stress, the dietary incorporation of dairy foods improves vascular health. We demonstrated that single-meal cheese consumption ameliorates acute, sodium-induced endothelial dysfunction. However, controlled feeding studies examining the inclusion of cheese, a dairy product that contains both bioactive constituents and sodium, are lacking.

Objectives

We tested the hypothesis that microcirculatory endothelium-dependent dilation (EDD) would be impaired by a high-sodium diet, but a sodium-matched diet high in dairy cheese would preserve EDD through oxidant stress mechanisms.

Methods

We gave 11 adults without salt-sensitive blood pressure (<10 mmHg Δ mean arterial pressure; 64 ± 2 y) 4 separate 8-d controlled dietary interventions in a randomized, crossover design: a low-sodium, no-dairy intervention (LNa; 1500 mg/d sodium); a low-sodium, high-cheese intervention (LNaC; 1500 mg/d sodium, 170 g/d cheese); a high-sodium, no-dairy intervention (HNa; 5500 mg/d sodium); and a high-sodium, high-cheese intervention (HNaC; 5500 mg/d sodium, 170 g/d cheese). On Day 8 of each diet, EDD was assessed through a localized infusion (intradermal microdialysis) of acetylcholine (ACh), both alone and during coinfusion of N^G-nitro-L-arginine methyl ester (NO synthase inhibitor), L-ascorbate (nonspecific antioxidant), apocynin [NAD(P)H oxidase inhibitor], or tempol (superoxide scavenger).

Results

Compared with LNa, microvascular responsiveness to ACh was attenuated during HNa (LNa: −4.82 ± 0.20 versus HNa: −3.21 ± 0.55 M logEC₅₀; P = 0.03) but not LNaC (−5.44 ± 0.20 M logEC₅₀) or HNaC (−4.46 ± 0.50 M logEC₅₀). Further, ascorbate, apocynin, and tempol administration each increased ACh-induced vasodilation during HNa only (Ringer's: 38.9 ± 2.4; ascorbate: 48.0 ± 2.5; tempol: 45.3 ± 2.7; apocynin: 48.5 ± 2.6% maximum cutaneous vascular conductance; all P values < 0.01).

Conclusions

These results demonstrate that incorporating dairy cheese into a high-sodium diet preserves EDD by decreasing the concentration of superoxide radicals. Consuming sodium in cheese, rather than in nondairy sources of sodium, may be an effective strategy to reduce cardiovascular disease risk in salt-insensitive, older adults. This trial was registered at clinicaltrials.gov as NCT03376555.

Introduction

Increased dairy consumption is a lifestyle factor that is associated with a reduced cardiovascular disease (CVD) risk, (1, 2) yet the majority of Americans over the age of 50 do not meet the recommended intake of 3 daily servings of dairy set forth in the 2015 US dietary guidelines (3). Conversely, high dietary sodium intake is a modifiable dietary factor that is associated with an increased CVD risk (4, 5). Because some dairy products—cheese, in particular—are high in sodium, increasing dairy intake in the form of cheese may unintentionally increase dietary sodium consumption. However, it is currently unknown whether dairy cheese consumption mitigates CVD risk, despite its contribution to dietary sodium intake.

High sodium ingestion impairs conduit arterial and microvascular function, independent of its effects on blood pressure (BP) (6, 7). In salt-resistant adults, brachial artery endothelium-dependent dilation (EDD) is significantly reduced following controlled feeding of a high-sodium diet (6). In the cutaneous microcirculation, dietary sodium loading reduces NO-dependent vasodilation to physiological, endothelial NO synthase (eNOS)-specific stimuli (7, 8). Emerging evidence indicates a role for oxidative stress in sodium-induced endothelial dysfunction (9, 10). In animal models, known contributors to sodium-induced oxidative stress include NAD(P)H and xanthine oxidases (11, 12), uncoupled NO synthase (NOS) (13, 14), and impaired superoxide dismutase (SOD) activity (11, 12, 15, 16). In human subjects, the administration of the nonspecific antioxidant ascorbate restores conduit arterial and cutaneous microvascular EDD following elevated sodium ingestion through NO-dependent mechanisms (7, 17). Collectively, these studies demonstrate that the impairments in endothelial function in response to excess sodium intake are mediated by oxidative stress mechanisms. However, specific enzymatic sources of sodium-induced oxidative stress have not been extensively investigated in humans.

In epidemiological studies, dairy consumption (3 servings per day) is associated with improved measures of vascular health, including lower BP and reduced arterial stiffness (18, 19). Independent of BP reductions, dairy-derived bioactive proteins (e.g., lactoferrin, lactotripeptides, caseinophosphopeptides) protect vascular endothelial function through multiple putative mechanisms, including acting as free radical scavengers (20, 21), reducing NAD(P)H oxidase (22), inhibiting lipid peroxidation (23), and improving antioxidant enzyme capacity through increased expression and activity (24, 25). These studies in isolated cell and animal models suggest that dairy proteins preserve endothelial function by limiting reactive oxygen species (ROS); however, specific mechanistic data in humans are lacking.

Recent work from our laboratory has demonstrated that single-meal cheese consumption protects against sodium-induced impairments in NO bioavailability through antioxidant mechanisms (26). However, studies incorporating dairy cheese as a source of bioactive peptides and sodium in a sustainable dietary pattern are needed to examine the mechanisms underlying these vascular effects. Thus, the objectives of this study were: 1) to establish the vasoprotective effects of an 8-d controlled feeding with a high-cheese diet on sodium-induced endothelial dysfunction in healthy, older adults; and 2) to identify the vasoprotective antioxidant mechanisms of cheese on sodium-induced oxidative stress. To address these aims, we measured vascular responsiveness to the endothelium-dependent agonist acetylcholine (ACh) in the human cutaneous microcirculation, using a validated model for assessing microvascular endothelial function (27, 28). We hypothesized that a high-sodium diet, compared with a low-sodium diet (5500 mg/d versus 1500 mg/d, respectively) would impair EDD, but a high-cheese diet (4 servings or 170 g per day) would preserve endothelial function despite providing high-sodium content. We further hypothesized that a high-cheese diet would protect against sodium-induced oxidative stress by reducing the accumulation of superoxide radicals.

Methods

All protocols were approved by the Institutional Review Board at The Pennsylvania State University and complied with the guidelines in the Declaration of Helsinki. The study was registered at clinicaltrials.gov as NCT03376555. All participants voluntarily provided written and verbal consent prior to enrollment. Prior to participation, subjects underwent a health screening that included a medical history, resting BP assessment, and fasting blood chemistry (Quest Diagnostics). Subjects (55–75 y old) were required to have an office-seated systolic BP (SBP) between 120 and 140 mmHg and diastolic BP (DBP) between 70 and 90 mmHg. Subjects were non-smokers and were not taking dietary supplements or prescription medications that might alter vascular function (e.g., antidepressants, antihypertensives, or statins). Subjects were screened for cardiovascular, neurological, metabolic, and renal diseases. Women taking hormone replacement therapy were excluded from the study.

Dietary protocol

Subjects were enrolled in 4 separate 8-d controlled feeding interventions in a randomized, crossover study design. All food and beverages were prepared by a registered dietitian in the Pennsylvania State Clinical Research Center Metabolic Kitchen. Investigators involved in data collection and analysis were blinded to the dietary treatment. The 4 dietary treatments were as follows: 1) a low-sodium (1500 mg) diet devoid of dairy products (LNa); 2) a low-sodium diet (1500 mg) containing 4 servings (170 g) of cheese per day (LNaC); 3) a high-sodium (5500 mg) diet devoid of dairy products (HNa); and 4) a high-sodium diet (5500 mg) containing 4 servings (170 g) of cheese per day (HNaC). The LNaC and HNaC diets consisted of a variety of natural cheeses, including cheddar, Swiss, Parmesan, Monterey Jack, Muenster, mozzarella, and provolone cheeses. The foods and quantities consumed during each dietary treatment are presented in Supplemental Tables 1–4. Dietary treatment blocks were separated by a minimum 1-week washout period. Dietary treatments were matched for total energy and macronutrients (30% fat, 50% carbohydrate, 20% protein). Subjects were permitted to drink water ad libitum, use sodium-free seasonings, and drink coffee or tea with minimal amounts of non-dairy, sodium-free creamer throughout each study arm. At the time of enrollment, participants met with a registered dietitian to identify their eucaloric energy requirements and were prescribed an energy level that achieved weight maintenance (1900, 2300, or 2700 kcal/d). The nutrient compositions of the dietary interventions providing 2300 kcal/d are presented in Table 1.

TABLE 1.

Macronutrient and micronutrient composition of 4 dietary interventions, which varied in sodium content and source¹

	LNa	LNaC	HNa	HNaC
Calories, kcal	2335	2348	2320	2289
Fat, g	84.2	91.0	81.7	82.5
Saturated	16.4	37.5	19.2	38.6
Monounsaturated	33.9	31.5	30.5	25.4
Polyunsaturated	26.2	13.2	23.3	10.7
Carbohydrate, g	288.9	258.4	288.4	272.0
Protein, g	114.0	113.1	112.8	117.5
Sodium, mg	1514	1475	5477	5496
Potassium, mg	3255	2454	3165	2433
Calcium, mg	894	1624	708	1836
Magnesium, mg	441	305	331	294

¹HNa, high-sodium diet; HNaC, high-sodium, high-cheese diet; LNa, low-sodium diet; LNaC, low-sodium, high-cheese diet.

Twenty-four-hour urine and blood pressure

On Day 7 of each 8-d intervention, participants collected a 24-h urine sample that was analyzed for urinary sodium and potassium concentrations (SmartLyte Electrolyte Analyzer, Diamond Diagnostics), urine osmolality (3320 Micro-Osmometer, Advanced Instruments), urine specific gravity, and total volume. On Day 7, subjects also underwent 24-h ambulatory BP monitoring (Ambulo 2400, Tiba Medical), during which BP was measured every 20 min during waking hours and every hour during sleep. Subjects that exhibited more than a 10 mmHg increase in 24-h mean arterial pressure (MAP) from the low-sodium to high-sodium diet were classified as salt-sensitive and excluded from the analysis (29–31).

Assessment of micro- and macrovascular function

Subjects abstained from vigorous physical activity and caffeine for at least 12 h prior to each measurement. On the morning of Day 8, subjects arrived at the laboratory, following an overnight fast, for a comprehensive evaluation of microvascular and macrovascular function. Subjects remained fasted throughout the duration of laboratory testing.

Microvascular: ACh-induced vasodilation

Measurements were obtained in a thermoneutral environment with subjects in a semi-supine position. Using a sterile technique, 5 intradermal microdialysis fibers (10 mm, 55 kDa cutoff membrane; CMA Microdialysis) were placed in the ventral forearm, as previously described (32). Red blood cell flux, an index of skin blood flow, was measured directly over each microdialysis membrane by a laser-Doppler flowmetry probe placed in a local heating unit (MoorLab, Temperature Monitor SH02; Moor Instruments), which was set to a thermoneutral temperature of 33°C, unless noted otherwise. Brachial BP was recorded (Cardiocap; GE Healthcare) every 5 min during the protocol.

All pharmaceutical perfusates were prepared immediately prior to infusion (33). Prior to data collection, 60–90 min were allowed for the resolution of any hyperemia caused by the insertion of the fibers, during which fibers were randomly assigned and perfused (2 μL/min; Baby Bee microinfusion pumps; Bioanalytical Systems) with either: 1) lactated Ringer's, to serve as a control; 2) 15 mmol/L of N^G-nitro-L-arginine methyl ester (L-NAME; Calbiochem, EMD Millipore), to non-selectively inhibit NO synthase; 3) 10 mmol/L of L-ascorbate (Sigma-Aldrich), to serve as a nonspecific antioxidant; 4) 100 μmol/L of apocynin (Tocris Bioscience), to inhibit NAD(P)H oxidase; or 5) 10 μmol/L of tempol (Sigma-Aldrich), to scavenge superoxide (FDA Investigational New Drug 129552). The efficacy of all perfusates at the concentrations utilized has been verified in previous studies (26, 33).

Following 20 min of baseline data collection, increasing concentrations of ACh (10⁻¹² to 10⁻¹ mol/L; USP) were perfused alone or in combination with site-specific pharmacological agents. After completion of the ACh dose-response, 28 mmol/L of sodium nitroprusside (USP) was perfused at all sites and the local skin temperature was increased to 43°C to induce maximal vasodilation (7, 26, 32).

Brachial artery endothelium-dependent and -independent vasodilation

According to standardized guidelines (34), after 10 min of quiet, supine rest, flow-mediated dilation (FMD) was assessed with a high-frequency ultrasound. A BP cuff was placed on the right forearm, distal to the ultrasound probe. Longitudinal images of the brachial artery, ∼5–10 cm above the elbow, were obtained with ultrasound imaging (Acuson Aspen 128XP, 10-mHz linear array transducer). Images were obtained during a 1-min baseline period, 5-min arterial occlusion via cuff inflation at 250 mmHg (Hokanson), and 2-min reactive hyperemia. Blood velocity (duplex-pulsed Doppler) was recorded during 10–15 s of baseline and immediately following the cuff release, for a calculation of reactive hyperemia. To assess endothelium-independent dilation (EID), ultrasound images were recorded during a 1-min baseline period and the 5 min following nitroglycerin administration (sublingual; 0.4 mg).

Blood analyses

On Day 8 of each diet, blood samples were collected in EDTA-treated tubes. Hemoglobin and hematocrit were analyzed directly from whole blood (Quest Diagnostics). Plasma was isolated by centrifugation and stored at −80°C until analysis. Plasma malondialdehyde was quantified using a Shimadzu LC-20XR system, equipped with a RF-10AXL fluorescence detector and vitamin E (as α- and γ-tocopherols), using a Thermo Scientific Dionex UltiMate 3000 HPLC-electrochemical system, as described previously (35). Plasma total nitrite and nitrate (NO_x) were quantified spectrophotometrically on a Biotek Synergy H1 microplate reader, using a commercial assay, in accordance with the manufacturer's instructions (Cayman Chemical). All plasma samples were measured in triplicate.

Data acquisition and statistical analysis

ACh-induced vasodilation

We determined by an a priori power analysis (power = 0.8; α = 0.05) on our primary outcome variable [cutaneous vascular conductance (CVC)] that 8 subjects would be sufficient to detect a physiologically meaningful difference between microdialysis treatments. Data were collected with Windaq (Dataq Instruments) at a frequency of 40 Hz. CVC was calculated as red blood cell flux, divided by MAP. ACh dose concentrations were log-transformed and CVC was expressed as a percentage of maximum to account for differences in vessel densities due to heterogeneity of the cutaneous microcirculation (8, 36). Pharmacological curve modeling was performed using a 4-parameter, nonlinear regression with a variable slope to detect between-diet differences in microvascular sensitivity [logarithm of the half maximal effective concentration (logEC₅₀) of the group mean curve] to ACh (Prism v7.01, GraphPad). Although the current study was not powered to detect sex differences, we performed a preliminary analysis to examine the potential influence of sex on microvascular sensitivity to ACh across diets. For this analysis, we used a 4-parameter, nonlinear regression with a variable slope to detect between-diet differences in vascular sensitivity to ACh with separate group mean curves for men (n = 5) and women (n = 6; Prism v7.01, GraphPad). We used 2-way repeated-measures ANOVAs to further examine the independent and interactive effects of sodium and cheese (sodium, cheese, and sodium x cheese) on vasodilation responses to ACh administration, biochemical variables, and hemodynamic parameters, and to compare vasodilation responses to ACh administration across local microdialysis treatments (site, dose, and site x dose; SAS; Version 9.4). Bonferroni's post hoc corrections were performed to account for multiple comparisons.

Endothelium-dependent dilation and endothelium-independent dilation

Automated edge detection software (Brachial Analyzer) was used to measure the brachial artery diameter continuously throughout the protocol. The baseline diameter was an average of all images obtained during the 1-min baseline period. The peak artery diameter was the largest diameter obtained during the 2-min reactive hyperemia and the 5-min post-nitroglycerin period, for the calculation of FMD and EID, respectively. FMD and EID were calculated as percentage changes from the baseline diameter. The arterial blood flow was calculated as the velocity time integral × heart rate × arterial cross-sectional area. Reactive hyperemia was expressed as a percentage change in blood flow following the cuff release, and was calculated as [(hyperemic flow volume − baseline flow volume)/baseline flow volume] × 100. A mixed-model ANOVA (SAS; Version 9.4) with subject as a random effect was used to detect differences across dietary treatments on FMD, EID, and hemodynamic parameters of the FMD response. Significance was accepted using α = 0.05. Unless otherwise indicated, all values are presented as means ± SEMs.

Results

There were 14 subjects who participated in the study, 2 of which were excluded for salt sensitivity and 1 of which was for noncompliance with the diets. The subject characteristics of the 11 subjects (64 ± 2 y; 5 men, 6 women) included in the data analysis are presented in Table 2.

TABLE 2.

Baseline characteristics of the study participants¹

	Values
Sex (M, F), n	(5, 6)
Age, y	64 ± 2
BMI, kg m2	26.6 ± 1.1
SBP, mmHg	125 ± 1
DBP, mmHg	81 ± 1
Serum total cholesterol, mg dL−1	198 ± 6
Serum HDL, mg dL−1	61 ± 5
Serum LDL, mg dL−1	117 ± 6
Serum fasting glucose, mg dL−1	97 ± 3
Whole blood HbA1c, %	5.5 ± 0.1

¹Values are shown as means ± SEMs unless otherwise indicated, n = 11. DBP, diastolic blood pressure; HbA1c, glycated hemoglobin; SBP, systolic blood pressure.

Hemodynamic and biochemical parameters

Mean hemodynamic and biochemical variables for each dietary intervention are displayed in Table 3. The 24-h SBPs, DBPs, and MAPs were not different across diets. As expected, urinary sodium excretion was significantly higher during the high-sodium diets, compared to the low-sodium diets (P < 0.001), providing evidence of subject adherence to the dietary interventions and the appropriate classification of included subjects as salt-insensitive.

TABLE 3.

Hemodynamic and biochemical parameters in healthy, older adults who consumed 4 separate 8-d diets, which varied in sodium content and source¹

					P value
	LNa	LNaC	HNa	HNaC	Sodium	Cheese	Sodium × cheese
24-h SBP, mmHg	124 ± 2	123 ± 3	126 ± 3	126 ± 3	0.39	0.84	0.96
24-h DBP, mmHg	77 ± 1	76 ± 2	78 ± 2	77 ± 1	0.36	0.52	0.94
24-h MAP, mmHg	92 ± 1	91 ± 2	94 ± 2	93 ± 2	0.30	0.62	0.98
24-h Na+ excretion, mmol/24 h	68.9 ± 8.2	69.6 ± 11.8	252.9 ± 18.2	216.6 ± 17.4	<0.001	0.33	0.32
24-h K+ excretion, mmol/24 h	35.6 ± 3.4	28.8 ± 3.0	42.0 ± 2.7	38.3 ± 2.5	0.022	0.10	0.60
Urine osmolality, mOsm kg H2O−1	423 ± 60	446 ± 46	534 ± 64	568 ± 65	0.08	0.64	0.93
Urine-specific gravity	1.012 ± 0.001	1.013 ± 0.002	1.020 ± 0.007	1.015 ± 0.002	0.21	0.49	0.44
Urine flow rate, mL/min	1.39 ± 0.20	1.21 ± 0.13	1.49 ± 0.13	1.45 ± 0.19	0.33	0.51	0.69
Hemoglobin, g/dL	13.9 ± 0.3	13.9 ± 0.3	13.3 ± 0.3	13.3 ± 0.2	0.056	0.93	0.87
Hematocrit, %	41.9 ± 0.8	40.9 ± 0.8	39.6 ± 0.6	39.5 ± 0.6	0.024	0.44	0.57
Plasma MDA, μmol/L	1.60 ± 0.05	1.57 ± 0.05	1.67 ± 0.06	1.56 ± 0.05	0.58	0.18	0.48
Plasma γ-tocopherol, μmol/L	3.69 ± 0.44	2.60 ± 0.37	3.45 ± 0.37	2.79 ± 0.25	0.95	0.040	0.57
Plasma α-tocopherol, μmol/L	29.2 ± 1.9	30.5 ± 1.6	29.3 ± 1.5	29.0 ± 1.9	0.70	0.76	0.64
Plasma NOx, μmol/L	70.3 ± 4.7	101.2 ± 10.8	63.7 ± 8.7	71.5 ± 7.7	0.054	0.042	0.20

¹Values are shown as means ± SEMs, n = 11. DBP, diastolic blood pressure; HNa, high-sodium diet; HNaC, high-sodium, high-cheese diet; LNa, low-sodium diet; LNaC, low-sodium, high-cheese diet; MAP, mean arterial pressure; MDA, malondialdehyde; NO_x, total nitrite and nitrate; SBP, systolic blood pressure.

ACh-induced vasodilation

Microvascular sensitivity to ACh was reduced during HNa, compared to during LNa and LNaC [Figure 1A; LNa: −4.82 ± 0.20 M versus HNa: −3.21 ± 0.55 M logEC₅₀ (P = 0.032); LNaC: −5.44 ± 0.20 M versus HNa: −3.21 ± 0.55 M logEC₅₀ (P = 0.002)]. Conversely, sensitivity to ACh during HNaC (−4.46 ± 0.50 M logEC₅₀) was not different from during LNa (P = 0.92) or LNaC (P = 0.31). In men, the reduction in microvascular sensitivity to ACh during HNa approached or reached significance, compared to during LNa (LNa: −5.22 ± 0.26 M versus HNa: −2.48 ± 1.08 M logEC₅₀; P = 0.054) and LNaC (LNaC: −5.80 ± 0.25 M versus HNa: −2.48 ± 1.08 M logEC₅₀; P = 0.01). These differences were not significant in women. No differences in microvascular sensitivity to ACh during concurrent L-NAME perfusion were observed across diets (Figure 1B; LNa: −3.64 ± 0.41 M; HNa: −2.75 ± 0.77 M; LNaC: −4.46 ± 0.35 M; HNaC: −4.20 ± 0.88 M logEC₅₀; P-diet = 0.12), suggesting that high dietary sodium impairs EDD by decreasing NO-dependent vasodilation, and that the inclusion of cheese in a high-sodium diet preserves endothelial function, at least in part, through an increase in NO bioavailability.

FIGURE 1.

CVC (% maximum) in response to perfusion of exogenous ACh (A) alone and (B) with concurrent perfusion of the NO synthase inhibitor, N^G-nitro-L-arginine methyl ester, in healthy, older adults who consumed 4 separate 8-d diets varying in sodium content and source. Values are means ± SEMs, n = 11. *P < 0.05, HNa versus LNa; ^†P < 0.05, HNa versus LNaC; ^‡P < 0.05, HNa versus HNaC. ACh, acetylcholine; CVC, cutaneous vascular conductance; HNa, high-sodium diet; HNaC, high-sodium, high-cheese diet; LNa, low-sodium diet; LNaC, low-sodium, high-cheese diet.

Local ascorbate, apocynin, and tempol administration each improved ACh-induced vasodilation during HNa (Figure 2; all P values < 0.01). As a result, microvascular sensitivity to ACh with a coinfusion of ascorbate (−4.65 ± 0.19 M logEC₅₀), tempol (−4.25 ± 0.41 M logEC₅₀), or apocynin (−4.59 ± 0.41 M logEC₅₀) during HNa was no longer different from that during LNa, LNaC, or HNaC. Ascorbate, tempol, or apocynin administration did not improve ACh-induced vasodilation during LNa, LNaC, or HNaC.

FIGURE 2.

CVC (% maximum) in response to perfusion of exogenous ACh alone and with concurrent perfusion of ascorbate (nonspecific antioxidant), apocynin [NAD(P)H oxidase inhibition], or tempol (superoxide scavenger) in healthy, older adults who consumed the following 8-d diets: (A) LNa, (B) LNaC, (C) HNa, or (D) HNaC. Values are means ± SEMs, n = 11. *P < 0.05, Ringer's versus ascorbate; ^†P < 0.05, Ringer's versus apocynin; ^‡P < 0.05, Ringer's versus tempol. ACh, acetylcholine; CVC, cutaneous vascular conductance; HNa, high-sodium diet; HNaC, high-sodium, high-cheese diet; LNa, low-sodium diet; LNaC, low-sodium, high-cheese diet.

Plasma concentrations of malondialdehyde (P-sodium * cheese = 0.48) and α-tocopherol (P-sodium * cheese = 0.64) were unaffected by the dietary treatments (Table 3). The plasma concentration of γ-tocopherol, however, was lower with the inclusion of cheese in the diet (P = 0.040). Plasma NO_x was increased with the inclusion of cheese in the diet (P = 0.042).

FMD and nitroglycerin-induced dilation

Brachial artery FMD was not statistically different across diets (Table 4; P-sodium * cheese = 0.19). No differences were observed in the brachial artery diameter, average blood velocity, blood flow volume before and after cuff occlusion, or reactive hyperemia (Table 4). EID was also not different across diets (Table 4; P-sodium * cheese = 0.25).

TABLE 4.

Hemodynamic parameters of the flow-mediated dilation response in healthy, older adults who consumed 4 separate 8-d diets, which varied in sodium content and source¹

					P value
	LNa	LNaC	HNa	HNaC	Sodium	Cheese	Sodium × cheese
FMD, %	5.07 ± 0.91	4.21 ± 0.77	3.45 ± 1.04	5.91 ± 1.35	0.97	0.45	0.19
EID, %	18.8 ± 1.2	17.3 ± 2.7	13.3 ± 2.1	17.2 ± 1.4	0.23	0.62	0.25
Baseline blood flow, mL/min	123 ± 14	123 ± 16	120 ± 14	122 ± 25	0.90	0.97	0.95
Baseline blood velocity, cm/s	15.6 ± 0.5	16.4 ± 2.1	14.5 ± 1.9	15.0 ± 1.1	0.49	0.69	0.91
Baseline artery diameter, mm	4.64 ± 0.35	4.51 ± 0.25	4.86 ± 0.38	4.53 ± 0.30	0.69	0.42	0.78
Hyperemic blood flow, mL/min	963 ± 85	944 ± 97	1017 ± 79	915 ± 106	0.90	0.53	0.67
Hyperemic blood velocity, cm/s	88.7 ± 9.5	93.6 ± 9.6	92.7 ± 13.2	88.0 ± 5.1	0.94	0.99	0.66
Hyperemic artery diameter, mm	4.86 ± 0.35	4.69 ± 0.25	5.00 ± 0.36	4.65 ± 0.28	0.87	0.62	0.57
Reactive hyperemia, %	714 ± 64	713 ± 73	798 ± 83	733 ± 69	0.50	0.66	0.69

¹Values are shown as means ± SEMs. FMD and hemodynamic parameters: LNa, n = 9; LNaC, n = 9; HNa, n = 8; and HNaC, n = 5. EID: LNa, n = 4; LNaC, n = 6; HNa, n = 9; and HNaC, n = 6. EID, endothelium-independent dilation; FMD, flow-mediated dilation; HNa, high-sodium diet; HNaC, high-sodium, high-cheese diet; LNa, low-sodium diet; LNaC, low-sodium, high-cheese diet.

Discussion

The findings of this rigorously controlled study in salt-insensitive, older adults demonstrate that the consumption of a high-cheese diet, despite its high-sodium content, protects against endothelial dysfunction by limiting oxidative stress. To mechanistically examine a potential vasoprotective role of dairy cheese against sodium-induced impairments in endothelial function, we measured cutaneous vascular responsiveness to the endothelium-dependent agonist ACh during 8-d high- or low-sodium diets that were either devoid of dairy or contained 4 daily servings (170 g/d) of cheese. Our data demonstrate that high dietary sodium ingestion impairs cutaneous vascular responsiveness to ACh through reductions in NO bioavailability. Importantly, the inclusion of cheese into a high-sodium diet prevented the reduction in vascular sensitivity to ACh that was otherwise observed with high dietary sodium ingestion, which may be mediated by an increase in NO-dependent vasodilation. Further, the acute administration of either a nonspecific antioxidant, NAD(P)H oxidase inhibitor, or superoxide scavenger improved ACh-induced vasodilation during a high-sodium diet, but did not further improve this response during a low-sodium diet or a high-sodium diet containing cheese. These findings suggest that sodium-induced endothelial dysfunction is mediated by increased superoxide and that the inclusion of 4 daily servings (170 g/d) of dairy cheese into a high-sodium diet protects against microvascular dysfunction by reducing the detrimental effects of sodium-induced superoxide radicals.

In addition to assessing microvascular function, we also examined arterial function using FMD and EID. FMD is a noninvasive method to measure peripheral endothelial function that is largely NO-dependent and predictive of future CVD (37). In the current study, brachial artery FMD and EID were not statistically different across diets. However, it should be noted that this study was not powered to detect differences in macrovascular function and that, despite a limited data set, we observed a large effect size (η_p² = 0.26) in FMD. Thus, while we were unable to make definitive conclusions on the effects of sodium and cheese ingestion on macrovascular function, our data support previous findings of sodium-induced impairments in macrovascular endothelial function (6, 38, 39) and further suggest that the inclusion of a dairy source of sodium—namely, cheese—into a diet may prevent the reduction in brachial artery EDD that is otherwise observed with high dietary sodium.

A large body of literature suggests that increases in oxidative stress mediate the impairments in endothelial function that are induced by high dietary sodium (9). A potential source of excess superoxide production during high sodium intake is NAD(P)H oxidase, as NAD(P)H oxidase inhibition restores arterial vascular function and NO bioavailability in animals fed a high-salt diet (12). Additionally, endogenous SOD is downregulated in multiple vascular beds of high salt–fed rodents (15, 40, 41). The administration of the SOD mimetic tempol improves markers of eNOS activation (42) and NO bioavailability (12) in high salt–fed rats. These data suggest that reductions in SOD expression and/or activity contribute to impaired EDD and NO signaling under high dietary sodium conditions.

Excess superoxide, due either to upregulated NAD(P)H oxidase and/or downregulated SOD, can react with NO to form peroxynitrite, a potent oxidizer of the essential NOS cofactor tetrahydrobiopterin (BH₄) (43). BH₄ supplementation normalizes EDD during elevated sodium intake in middle-aged and older adults (17), which supports the idea that excess superoxide mediates the impairments in endothelial function caused by high dietary sodium. In agreement with this previous work, we observed improvements in ACh-induced vasodilation with ascorbate, apocynin, or tempol administration during high-sodium intake, but not low-sodium intake, suggesting that superoxide—particularly NAD(P)H oxidase-derived superoxide—contributes to sodium-induced endothelial dysfunction.

In controlled studies examining the vasoprotective effects of dairy consumption, the ingestion of dairy or milk-derived proteins reduced measures of oxidative stress. In mice, 3 weeks of a high-dairy diet reduced NAD(P)H oxidase expression and intracellular ROS concentrations (22). Similarly, in humans, the consumption of dairy proteins or nonfat dairy milk significantly increased plasma antioxidant capacities (44) and reduced markers of oxidative stress, including oxidized LDL and the lipid peroxidation byproduct, malondialdehyde (45). We recently provided further evidence for the antioxidant properties of dairy by demonstrating that acute (single meal) dairy cheese consumption prevented sodium-induced endothelial dysfunction through a reduction of ascorbate-sensitive oxidants (26). Our current data extend these findings, and show that the inclusion of cheese (4 servings or 170 g per day) into a high-sodium diet preserves EDD through reductions in oxidative stress. In the present study, plasma concentrations of malondialdehyde were not statistically lower during the high-cheese diets, although the effect size was large (η_p² = 0.20). Unexpectedly, plasma concentrations of the specific antioxidant γ-tocopherol were significantly lower during the diets containing cheese, compared to the diets devoid of dairy. However, this finding may be explained by the lower dietary γ-tocopherol content of the diets containing cheese, compared to the dairy-free diets. In addition, we observed main effects of sodium and cheese on total NO_x, but we did not observe a significant sodium by cheese interaction. Our functional data suggest that cheese increases NOS-dependent dilation in the presence of high sodium, but not low sodium. Conversely, cheese consumption increased total NO_x—an indirect measure of total NO production/bioavailability—independent of sodium intake. It should be noted that NO_x is not specific to endothelium-derived NO and is not a direct measure of functional NO (46). Whether cheese consumption alters NO homeostasis independent of eNOS requires further investigation. Further, cheese consumption may reduce ROS and, thus, improve NO bioavailability, but in the absence of high sodium, there may be no functional deficit to correct in this population.

Cheese is a dairy product that is high in saturated fat. Saturated fat is a macronutrient that has been associated with an increase in CVD risk; however, saturated fat may have different effects on CVD risk, depending on the food source. The consumption of saturated fat from dairy sources—specifically, cheese—has been associated with a reduced CVD risk, whereas the consumption of saturated fat from non-dairy sources (e.g., meat) has been associated with an increased CVD risk (47). In the current study, a high-cheese diet improved microvascular function, despite a higher saturated fat content, which supports this notion that the saturated fat consumed in dairy may not have the same detrimental effects on vascular health as the saturated fat in non-dairy foods.

Salt sensitivity is associated with an increased risk of hypertension and cardiovascular mortality (48). However, vascular endothelial dysfunction—a precursor to the development of clinical CVD—is evident during high sodium intake, even in the absence of an increase in BP (6, 7). To separate the direct effects of sodium on endothelial function from the effects of BP, we excluded salt-sensitive participants from the analysis. As a result, we did not observe a significant increase in BP across diets. Thus, our findings demonstrate that the alterations in oxidant status and endothelial function induced by high dietary sodium and the beneficial effects of cheese on sodium-induced endothelial dysfunction occur independently of any changes in BP.

There is evidence in the existing literature that the impairment in endothelial function during high dietary sodium intake is greater in men, compared to women (49, 50). In the current study, the decrease in vascular responsiveness during the high-sodium diet, relative to the low-sodium diet, was evident in men but not women. Thus, although this study was not powered to examine sex differences, our data appear to support previous findings of greater vascular sensitivity to dietary sodium in men, compared to women, and suggest that the inclusion of cheese into a high-sodium diet may have a greater protective effect in men than in women. Future studies are required to further elucidate sex differences in the vascular responses to dietary sodium and cheese intake.

Limitations

Although there is robust evidence for beneficial, bioactive properties of milk-derived peptides, we cannot attribute the vascular benefits observed in the present study solely to the dairy proteins in cheese. The dietary interventions were matched for macronutrient contents (i.e., total protein, fat, and carbohydrates), but not micronutrient contents (e.g., calcium, potassium, and magnesium). Dietary potassium intake is associated with improved measures of NO bioavailability and may partially mitigate sodium-induced endothelial dysfunction (51). The ratio of dietary sodium-to-potassium intakes is becoming increasingly recognized as a predictor of CVD, even more so than dietary sodium consumption alone (52). Interestingly, we observed an improvement in EDD with the inclusion of cheese in a high-sodium diet, despite a lower potassium content than the diets devoid of dairy products. Thus, the findings of the present study suggest that the beneficial effects of dairy on sodium-induced endothelial dysfunction, along with evidence from our pharmacological dissection of ROS mechanisms, are likely due to the well-documented antioxidant properties of bioactive proteins in cheese.

The local administration of apocynin improved microvascular endothelial function during a high-sodium diet, but not a high-sodium, high-cheese diet, which suggests that cheese prevents the increase in NAD(P)H oxidase-derived superoxide that occurs during high sodium intake. We recognize that there are limitations of using apocynin to inhibit NAD(P)H oxidase. First, apocynin is a non-specific inhibitor of NAD(P)H oxidase and does not identify which isoform contributes to alterations in ROS. Second, there is evidence that apocynin may scavenge hydrogen peroxide and that the inhibition of NAD(P)H oxidase by apocynin involves the activation of myeloperoxidase (53, 54). However, hydrogen peroxide is known to cause vasodilation in the microvasculature (55, 56); thus, apocynin-mediated reductions in hydrogen peroxide likely do not explain the results of the current study.

Perspectives

The average intake of dairy in the United States remains significantly lower than the USDA's recommended 3 servings per day (3). Further, the average dietary sodium intake in the United States is significantly greater than the recommended upper limit of 2300 mg of sodium per day set forth by the American Heart Association (57), and only a small percentage (∼5%) of the total sodium consumption is in the form of dairy foods. Thus, although cheese consumption may contribute to sodium intake, our findings suggest that natural dairy cheese prevents sodium-induced vessel dysfunction through reductions in oxidative stress and that consuming sodium in cheese, rather than in non-dairy sources of sodium, may be an effective strategy to reduce CVD risk in healthy, older adults without salt-sensitive blood pressure.

Notes

This research was supported by the National Dairy Council.

Author disclosures: BKA, AES, PD, RSB, WLK, and LMA, no conflicts of interest.

Supplemental Tables 1-4 are available from the “Supplementary data” link in the online posting of the article and from the same link in the online table of contents at https://academic.oup.com/jn.

Abbreviations used: ACh, acetylcholine; BH₄, tetrahydrobiopterin; BP, blood pressure; CVC, cutaneous vascular conductance; CVD, cardiovascular disease; DBP, diastolic blood pressure; EDD, endothelium-dependent dilation; EID, endothelium-independent dilation; eNOS, endothelial NO synthase; FMD, flow-mediated dilation; HNa, high-sodium diet; HNaC, high-sodium, high-cheese diet; LNa, low-sodium diet; LNaC, low-sodium, high-cheese diet; L-NAME, N^G-nitro-L-arginine methyl ester; MAP, mean arterial pressure; NOS, NO synthase; NO_x, total nitrite and nitrate; ROS, reactive oxygen species; SBP, systolic blood pressure; SOD, superoxide dismutase.