Supplemental Watermelon Juice Attenuates Acute Hyperglycemia-Induced Macro-and Microvascular Dysfunction in Healthy Adults

ABSTRACT

Background

Acute hyperglycemia reduces NO bioavailability and causes macro- and microvascular dysfunction. Watermelon juice (WMJ) is a natural source of the amino acid citrulline, which is metabolized to form arginine for the NO cycle and may improve vascular function.

Objectives

We examined the effects of 2 weeks of WMJ compared to a calorie-matched placebo (PLA) to attenuate acute hyperglycemia-induced vascular dysfunction.

Methods

In a randomized, placebo-controlled, double-blind, crossover trial, 6 men and 11 women (aged 21–25; BMI, 23.5 ± 3.2 kg/m²) received 2 weeks of daily WMJ (500 mL) or a PLA drink followed by an oral-glucose-tolerance test. Postprandial flow-mediated dilation (FMD) was measured by ultrasound (primary outcome), while postprandial microvascular blood flow (MVBF) and ischemic reperfusion were measured by near-infrared spectroscopy (NIRS) vascular occlusion test (VOT).

Results

The postprandial FMD area AUC was higher after WMJ supplementation compared to PLA supplementation (838 ± 459% · 90 min compared with 539 ± 278% · 90 min; P = 0.03). The postprandial MVBF (AUC) was higher (P = 0.01) following WMJ supplementation (51.0 ± 29.1 mL blood · 100 mL tissue^–1 · min^–1 · 90 min) compared to the PLA (36.0 ± 20.5 mL blood · 100 mL tissue^–1 · min^–1 · 90 min; P = 0.01). There was a significant treatment effect (P = 0.048) for WMJ supplementation (71.2 ± 1.5%) to increase baseline tissue oxygen saturation (StO₂%) when compared to PLA (65.9 ± 1.7%). The ischemic-reperfusion slope was not affected by WMJ treatment (P = 0.83).

Conclusions

Two weeks of daily WMJ supplementation improved FMD and some aspects of microvascular function (NIRS-VOT) during experimentally induced acute hyperglycemia in healthy adults. Preserved postprandial endothelial function and enhanced skeletal muscle StO₂% are likely partially mediated by increased NO production (via citrulline conversion into arginine) and by the potential antioxidant effect of other bioactive compounds in WMJ.

Introduction

Emerging evidence suggests that postprandial hyperglycemia is a more significant risk factor for developing cardiovascular disease (CVD) than fasting hyperglycemia (1–3). Even in healthy, young people, a 75-g oral glucose tolerance test (OGTT) causes postprandial hyperglycemia capable of reducing endothelium-dependent vasodilation, measured by flow-mediated dilation (FMD) (4–6). Skeletal muscle and adipose tissue microvascular dysfunction also occur in healthy people undergoing an oral-glucose challenge and in those with type 2 diabetes (T2DM) (7–10). The insulin response to glucose facilitates glucose disposal into the skeletal muscle and adipose tissue in part by promoting microvascular recruitment in those tissues to increase nutritive blood flow (11). However, insulin resistance can also occur in vascular endothelium and contribute to reduced postprandial vasodilation and glucose delivery to insulin-sensitive tissues.

The responsiveness of the vascular endothelium is primarily modulated by NO derived from endothelial nitric oxide synthase (eNOS). The eNOS enzyme generates NO by the conversion of arginine into citrulline. NO promotes vasodilation by causing vascular smooth muscle relaxation through cyclic guanosine monophosphate–mediated phosphorylation of smooth muscle myosin (12). The reduction of NO bioavailability is a hallmark characteristic of CVD and T2DM (13). Work from Mah et al. (5) demonstrated that during an OGTT, arginine is reduced and its competitive inhibitor, asymmetric dimethylarginine (ADMA), is increased, leading to overall reduced NO bioavailability. Therefore, a commonly pursued therapy to improve vascular function is to increase NO bioavailability by providing supplemental arginine (14, 15). Arginine is the direct substrate for eNOS function; however, oral arginine supplementation is subject to a high degree of first-pass extraction and conversion to ornithine and urea in the liver (16, 17). Furthermore, randomized clinical trials using long-term arginine supplementation have failed to improve NO synthesis or vascular function (18, 19). Interestingly, supplemental citrulline appears to be a more efficient source of arginine than arginine itself (16, 20). When consumed orally, citrulline does not undergo extraction in the gastrointestinal tract or liver, but is absorbed by the kidney and converted to arginine by a truncated urea cycle process (21).

Hyperglycemia-induced vascular dysfunction is mediated by a reduction in plasma arginine, leading to reduced NO function (22). Watermelons are a rich natural source of citrulline and are potentially capable of improving cardiometabolic health (21). Watermelon juice (WMJ) supplementation increases plasma arginine and NO bioavailability (23, 24). In the current study, we aimed to test the hypothesis that a natural source of citrulline, WMJ, would attenuate the vascular dysfunction that occurs during postprandial hyperglycemia. Since conditions like T2DM are associated with both macro- and microvascular dysfunction, we performed FMD of the brachial artery with concurrent measurements of skeletal muscle microvascular blood flow (MVBF) and tissue oxygen saturation in response to ischemic reperfusion during an OGTT.

Methods

Participants and study design

The study was approved by the Pennington Biomedical Research Center and Louisiana State University Institutional Review Board. Eligible participants provided written informed consent prior to enrolling in the study. This trial was registered at clinicaltrials.gov under NCT04092439. Figure 1A provides an overview of the study design. Participants completed a single baseline (prerandomization) fasted blood draw and body composition analysis by DXA (Hologic, Horizon A), after which they were randomly assigned to 2 weeks of daily fresh WMJ or a placebo (PLA) drink followed by an OGTT with simultaneous macro- and microvascular measurements (Figure 1B). Each participant completed a 2-week washout period followed by the alternative supplement (WMJ or PLA) for an additional 2 weeks and the final OGTT. Due to the production of WMJ and PLA drinks, the first 6 participants received treatment A first and the next 6 participants received treatment B first. The remaining 5 participants were randomized to receive either treatment A or B first.

FIGURE 1.

Study design and experimental procedures. FMD, flow-mediated dilation; NIRS, near-infrared spectroscopy; OGTT, oral glucose tolerance test; PLA, calorie-matched placebo; WMJ, watermelon juice supplement.

Supplementation intervention

Study participants were required to consume either 500 mL of fresh WMJ or 500 mL of calorie-matched watermelon flavored drink (PLA) per day, prepared by the Louisiana State University Agriculture Center (Co-Investigator, Losso). Table 1 provides a breakdown of the nutritional composition of the WMJ drink. The PLA drink was a red-colored, watermelon-flavored water matched for calories (140 kcals) and containing 35 g of fructose. Participants were required to consume the study beverage on site during the weekdays and were provided with study drinks over the weekends, and were required to return the bottles when arriving for study visits. Participants consumed the last study drink at least 10–12 hours before the OGTT visit. Throughout the study, participants received dietary counseling to maintain their daily carbohydrate intake (∼150 g/day) and avoid increasing intakes of nitrate-rich foods or NO enhancing supplements, which may confound the data interpretation.

TABLE 1.

Energy, nutrient, and amino acid content of supplemental WMJ¹

	WMJ
Nutrient2
Energy, kcal	140
Total carbohydrate, g	35.3
Total sugar, g	29.0
Glucose, g	7.4
Fructose, g	15.7
Sucrose, g	5.7
Protein, g	2.8
Total fat, g	0.7
Sodium, mg	4.7
Potassium, mg	523
Magnesium, mg	46.6
Phosphorus, mg	51.4
Calcium, mg	32.7
Vitamin A, μg	130.8
Vitamin C, mg	37.8
Vitamin E, mg	0.20
Beta-carotene, mg	1.45
Lycopene, mg	21.2
Amino acids,3 mg
Alanine	27.4
Arginine	195
Asparagine	33.7
Aspartic acid	51.2
Citrulline	795
Glutamine	133.
Glutamic acid	13.4
Glycine	6.89
Histidine	20.1
Isoleucine	41.4
Leucine	17.3
Lysine	21.4
Methionine	21.8
Ornithine	19.3
Phenylalanine	48.2
Proline	15.3
Serine	38.1
Threonine	11.8
Tryptophan	19.3
Tyrosine	7.8
Valine	26.3

¹WMJ, watermelon juice

²Energy and nutrient compositions were obtained from Nutritionist Pro Software.

³The amino acid concentration was obtained through our HPLC analysis.

OGTT visits

After completing the 2-week WMJ or PLA supplementation period, participants were required to arrive fasted (10–12 hours), with no previous exercise or caffeine (48 hours), to undergo an OGTT. Upon arrival, participants were instructed to void, and a flexible i.v. catheter was inserted into the antecubital vein of the left arm. Blood samples were collected at 0 minutes and, following ingestion of 75 g (Glucola, Mercedes Scientific), at 15, 30, 60, 90, and 120 minutes. Follow-up measures of body composition were obtained via DXA for PLA and WMJ OGTT visits.

FMD

FMD measurements were taken according to the standard procedures provided by Thijssen et al. (25). A blood pressure cuff (Delphi) was placed on the upper forearm distal to the ultrasound probe. The brachial artery of the right arm was identified using color and Doppler modes and scanned using B-mode on the ultrasound (GE LogiqE9). The ultrasound probe was held in place by an adjustable ultrasound clamp (QUIPU) to ensure image quality and reproducibility. Before cuff inflation, the brachial artery diameter was recorded for 2 minutes. Then, the cuff was rapidly inflated (Hokanson E20) to 250 mm Hg for 5 minutes. Before deflating the cuff, the participant was reminded to remain still, and videos of the brachial artery were attained immediately after deflation and at 1, 3, and 5 minutes after deflation. The peak vessel diameter was measured using automated edge-detection software (Brachial Analyzer, Mia-LLC). FMD measurements were performed at 0, 30, 60, and 90 minutes after glucose ingestion.

Near-Infrared Spectroscopy (MVBF) and Vascular Occlusion Test

The near-infrared spectroscopy (NIRS) MVBF measurement and vascular occlusion test (VOT) were performed at 0, 30, 60, and 90 minutes after glucose ingestion. The NIRS system was calibrated using a calibration phantom before each study visit, after allowing the system to run for approximately 30 minutes. The NIRS optodes (Oxymon MKIII, Artinis Medical Systems) were placed longitudinally on the flexor digitalis superficialis (forearm) proximal to its insertion and measured to ensure reproducibility. The optodes were secured in place with 2-sided adhesive tape and further anchored in place with additional medical tape. NIRS signals for oxygenated hemoglobin/myoglobin (O₂Hb/O₂Mb), deoxygenated hemoglobin/myoglobin (HHb/HMb), total hemoglobin/myoglobin (tHb/tMb), and tissue oxygen saturation (StO₂, %) were continuously monitored through the MVBF and VOT. The rapid inflation cuff was placed proximal to the optodes. After ensuring correct probe placement, the participants laid supine at 80° and were instructed to remain still during the measurements. MVBF was initially monitored by performing a rapid venous occlusion (60 mm Hg) for 2 seconds each for 3 times, separated by a 2-minute washout period as previously described (26). The MVBF response was measured by calculating the slope of the linear increase in total hemoglobin (tHb):

(1)

The NIRS-VOT was performed in parallel with the FMD test, as detailed by Soares et al. (27) (Figure 2). The baseline StO₂ was calculated from an average of 2 minutes before the arterial occlusion. The ischemic-reperfusion slope (i.e., slope 2, %/s^–1) was calculated from the slope of the linear increase during a 10-second window after cuff pressure deflation. The peak StO₂ (%) represents the peak O₂% achieved after the 5-minute arterial occlusion period. The reperfusion AUC is calculated from the 8-minute period when the StO₂ increased above the baseline StO₂ during the reperfusion period.

FIGURE 2.

Example of components of the NIRS-VOT oxygen saturation signal. The baseline StO₂ is calculated from the 2 minutes before arterial cuff occlusion. The arterial occlusion period lasted for 5 minutes, followed by rapid cuff deflation. Slope 2 is calculated from the 10-second window after cuff deflation. The peak StO₂ is measured as the point of highest StO₂ achieved after cuff deflation. The reperfusion AUC is calculated from the trapezoidal method from the 8 minutes of StO₂ overshoot above baseline StO₂. NIRS, near-infrared spectroscopy; StO₂, tissue oxygen saturation; VOT, vascular occlusion test.

Clinical chemistries

Blood samples were collect using EDTA and sodium fluoride/oxalate tubes and were placed on ice. Serum samples were collected using silicone-coated tubes. Blood glucose was analyzed using a bedside measurement system (GL5, Analox Instruments LTD). Plasma insulin was measured using an ELISA kit (Millipore Sigma). Serum nonesterified fatty acids (NEFA) were quantified using an enzymatic colorimetric assay (Wako Chemical). Plasma and WMJ amino acid concentrations were measured by HPLC analysis as previously reported (28).

Statistical analysis

The statistical analysis was performed using JMP version 15 statistical software (SAS Institute Inc.). Sample size calculations using G*Power suggest that 18 participants are required to detect a ∼15% difference in the FMD response to an OGTT, assuming 80% power, an α of 0.05, an SD of 1.65% (4), and a 0.5 correlation between measurements. Linear mixed-effect models were used to assess mean differences between treatments and time points, after adjusting for visit order. In the primary models, treatment (PLA compared with WMJ), time (0–120 minutes for OGTT and 0–90 minutes for vascular tests), and visit order were treated as fixed effects, while the subject ID was treated as a random effect. Treatment, time, and treatment-by-time interaction effects for postprandial plasma (glucose, insulin, and NEFA), FMD, and NIRS-VOT responses were tested using these linear mixed-effect models, adjusting for visit order. Post hoc differences in least square means were determined by Tukey's Honest Significant Difference test. The AUC was calculated using the trapezoidal rule over the 120 minutes (plasma markers) and 90 minutes (vascular measurements). Comparisons of AUCs and fasted plasma amino acids (0-minute time point) between treatments were performed using a linear mixed-effects model with treatment as the fixed effect and subject ID as a random effect, after adjusting for visit order. No visit order effects were statistically significant (P > 0.05; data not shown). Data are reported as means ± SDs unless otherwise indicated. Statistical significance was determined as a P value < 0.05,  and values between 0.05 and 0.10 were considered trends.

Results

Participants

Seventeen healthy, young (aged 21–25 years) adults (6 men and 11 women) were enrolled and completed the study. Table 2 provides body composition data at baseline (prerandomization) and at the time of the follow-up OGTT for PLA and WMJ treatments.

TABLE 2.

Body composition of healthy young adults who consumed WMJ and PLA daily, each for 2 weeks¹

Parameter	Baseline	PLA	WMJ
BMI, kg/m2	23.7 ± 3.2	23.6 ± 3.2	23.6 ± 3.3
Fat mass, kg	19.0 ± 7.1	19.8 ± 7.9	19.3 ± 7.0
Fat mass, %	27.7 ± 8.1	28.0 ± 8.8	28.2 ± 8.1
Fat free mass, kg	47.2 ± 10.2	48.5 ± 11.8	46.9 ± 10.4
Fat free mass, %	69.9 ± 8.2	69.7 ± 8.7	69.6 ± 7.9
Visceral fat mass, g	256 ± 113	261 ± 114	244 ± 108

¹Data were analyzed using a linear mixed-effects model with subject ID as random effect and treatment and visit order as fixed effects. No treatment or visit order effects (P > 0.05) were observed. Values are means ± SDs; n = 17. PLA, placebo; WMJ, watermelon juice.

Glucose, insulin, and NEFAs

The fasting plasma glucose levels did not differ (P = 0.75) between the PLA (5.19 ± 0.38 mmol/L) and WMJ (5.23 ± 0. 38 mmol/L) interventions. Fasting plasma insulin (PLA 3.6 ± 2.0 μU/mL compared with WMJ 4.1 ± 1.9 μU/mL; P = 0.20) and serum NEFA (PLA 0.37 ± 0.20 mmol/L compared with WMJ 0.48 ± 0.23 mmol/L; P = 0.17) values did not differ between groups. As expected, there were significant time effects for glucose (P < 0.0001), insulin (P < 0.0001), and NEFA (P < 0.0001) during the OGTT (Figure 3A–C), but these parameters did not differ by treatment. The AUCs were also not different in response to the OGTT for glucose (PLA 757 ± 140 mmol/L · min compared with WMJ 818 ± 133 mmol/L · min; P = 0.14), plasma insulin (PLA 5320 ± 3720 μU/mL · min compared with WMJ 5890 ± 3960 μU/mL · min; P = 0.40), and NEFA (PLA 22.4 ± 9.3 nmol · min compared with 24.1 ± 12.3 nmol · min; P = 0.57).

FIGURE 3.

The postprandial response to an OGTT in healthy young adults after 2 weeks of daily WMJ and PLA supplementation for (A) glucose, (B) insulin, (C) free fatty acid, and (D) FMD. Data were analyzed using a linear mixed-effect model with subject ID as a random effect and fixed effects for treatment, time, and treatment-by-time. Data are means ± SEMs; n = 17. *Time point is different at a P value < 0.05 from the 0-minute time point. FMD, flow-mediated dilation; NEFA, nonesterified fatty acids; OGTT, oral-glucose-tolerance test; PLA, placebo; WMJ, watermelon juice.

FMD

The resting/fasting brachial artery diameters did not differ between the WMJ supplementation (3.53 ± 0.65 mm) and PLA (3.31 ± 0.65 mm) periods (P = 0.12). There were no significant effects for time (P = 0.23) or treatment (P = 0.07) for postprandial FMD after 2 weeks of WMJ supplementation or PLA (Figure 3D). There was also no time-by-treatment interaction (P = 0.84). However, WMJ supplementation did increase the total AUC (838 ± 459% · 90 min) for the postprandial FMD (P = 0.03) when compared to PLA (539 ± 278% · 90 min).

NIRS MVBF

The fasted blood flow values did not differ between groups (PLA 0.51 ± 0.36 mL blood · 100 mL tissue^–1 · min^–1compared with WMJ 0.53 ± 0.27 mL blood · 100 mL tissue^–1 · min^–1). There were no significant time (P = 0.21), treatment (P = 0.49), or time-by-treatment interaction (P = 0.72) effects detected for postprandial MVBF (Figure 4A). The total AUC for postprandial MVBF (Figure 4B) was significantly greater after WMJ supplementation (51.0 ± 29.1 mL blood · 100 mL tissue^–1 · min^–1 · 90 min; P = 0.01) when compared to the PLA (36.0 ± 20.5 mL blood · 100 mL tissue^–1 · min^–1 · 90 min).

FIGURE 4.

The postprandial response to an OGTT in healthy young adults after 2 weeks of daily WMJ and PLA supplementation on skeletal muscle microvascular blood flow. (A) The analysis time course of blood flow response was done using a linear mixed-effect model with subject ID as a random effect and fixed effects for treatment, time, and treatment-by-time. (B) The differences in AUC for blood flow were determined using a linear mixed-effect model with subject ID as a random effect and fixed effects for treatment and visit order. Data points are means ± SEMs; n = 17. *P < 0.05 difference between PLA and WMJ. OGTT, oral-glucose-tolerance test; PLA, placebo; WMJ, watermelon juice.

NIRS StO₂ and VOT

Using a linear mixed-effects model, we observed significant time (P = 0.05) and treatment (P = 0.048) effects, but no interaction effect (treatment-by-time; P = 0.28), demonstrating an increased baseline StO₂ during the postprandial period after WMJ supplementation (mean baseline StO_2, 71.2 ± 1.5%) compared with PLA (mean baseline StO_2, 65.9 ± 1.7%; Figure 5A). An analysis of the AUC revealed that WMJ supplementation (6400 ± 956% · 90 min) increased (P = 0.01) the baseline StO₂ compared to the PLA intervention (5800 ± 651% · 90 min). There was a significant time effect (P = 0.001), but no effect for treatment (P = 0.83) or interaction (treatment-by-time, P = 0.58), for postprandial slope 2 (Figure 5B). The WMJ treatment demonstrated a significant treatment effect (P = 0.02) for the postprandial peak StO₂ (Figure 5C) but did not differ by time (P = 0.68) or treatment-by-time interaction (P = 0.95). The total AUC for the peak StO₂ was higher (P = 0.01) after WMJ supplementation (7700 ± 581% · 90 min; mean Peak StO₂, 86.2 ± 0.78%) compared to PLA (7210 ± 555% · 90 min; mean Peak StO₂, 82.1 ± 0.89%)_. We observed time (P < 0.001) and treatment (P = 0.04) effects for reperfusion AUC, but no treatment-by-time interaction (P = 0.33; Figure 5D).

FIGURE 5.

The postprandial response to an OGTT in healthy young adults after 2 weeks of daily WMJ and PLA supplementation on the components of the NIRS-VOT. (A) BL StO₂, (B) slope 2, and (C) peak StO₂ values and (D) the reperfusion AUC were analyzed using a linear mixed-effect model with subject ID as a random effect and fixed effects for treatment, time, and treatment-by-time. Data are means ± SEMs; n = 17. BL, baseline; OGTT, oral-glucose-tolerance test; NIRS, near-infrared spectroscopy; PLA, placebo; VOT, vascular occlusion test; WMJ, watermelon juice.

Plasma amino acids

We analyzed fasting plasma amino acid concentrations in a subset of participants (n = 12; 6 males; Table 3). Fasting alanine concentrations were significantly (P = 0.007) reduced, while fasting arginine concentrations trended (P = 0.08) to increase for WMJ supplementation when compared to PLA. The ratio of ADMA to arginine was not affected after WMJ compared with PLA supplementation (P = 0.28). The plasma citrulline concentration was not different between the PLA and WMJ supplementation periods (P = 0.86).

TABLE 3.

Fasted plasma amino acid concentrations of healthy young adults who consumed WMJ and PLA daily, each for 2 weeks¹

Amino acid, μmol/L	PLA	WMJ	P value
Alanine	290 ± 55	241 ± 35	0.007
Arginine	58 ± 10	64 ± 14	0.08
Asparagine	72 ± 11	68 ± 12	0.42
Aspartic acid	2.0 ± 1.3	1.4 ± 0.4	0.12
Citrulline	24 ± 4.6	24 ± 4.5	0.86
Glutamic acid	23 ± 9.5	19 ± 7.4	0.06
Glutamine	511 ± 67	488 ± 70	0.06
Glycine	191 ± 28	185 ± 44	0.56
Histidine	72 ± 9.3	71.9 ± 10	0.65
Isoleucine	53 ± 12	51 ± 12	0.53
Leucine	101 ± 20	95 ± 20	0.44
Lysine	143 ± 36	134 ± 22	0.53
Methionine	20 ± 5.1	19 ± 3.5	0.45
Ornithine	35 ± 6.7	34 ± 10	0.79
Phenylalanine	44 ± 8.4	42 ± 6.7	0.44
Serine	81 ± 18	77 ± 8.3	0.40
Threonine	127 ± 28	117 ± 19	0.31
Tryptophan	40 ± 6.0	38 ± 6.5	0.33
Tyrosine	45 ± 14	40 ± 8.7	0.11
Valine	177 ± 22	176 ± 35	0.92
ADMA	0.50 ± 0.5	0.48 ± 0.05	0.67
ADMA:arginine	0.009 ± 0.002	0.008 ± .002	0.28

¹Data were analyzed using a linear mixed-effects model with subject ID as random effect and treatment and treatment order as fixed effects. P values are for the main effect of treatment. No visit order effects were observed (P > 0.05). Values are means ± SDs; n = 12. ADMA, asymmetric dimethylarginine; PLA, placebo; WMJ, watermelon juice.

Discussion

The current study's results demonstrate that 2 weeks of WMJ supplementation maintains macro- and microvascular function during experimentally induced hyperglycemia (OGTT) in healthy, young adults. Specifically, the present data suggest postprandial endothelial function is maintained to a greater degree after participants received chronic WMJ supplementation. Additionally, we are the first to show that aspects of microvascular function are improved when participants are treated with WMJ compared to a calorie-matched PLA. Using NIRS, we measured skeletal muscle oxygen saturation during discrete components of the NIRS-VOT. Our data show a clear reduction in the baseline StO_2, and the reperfusion AUC during postprandial hyperglycemia, which is attenuated after WMJ supplementation compared to PLA. The postprandial peak StO₂ level remained significantly elevated in the WMJ compared with the PLA condition. Since many macro- and microvascular functions are thought to be reduced in a NO-mediated manner during hyperglycemia, we measured fasting plasma amino acid levels. In a subset of participants (n = 12), we demonstrated a trend (P = 0.08) for WMJ supplementation to increase fasting plasma arginine, which has been shown to be important for postprandial vascular responses (29).

Reduced endothelial function is a common feature in T2DM that is considered highly predictive of future CVD events. The loss of NO bioavailability is implicated at a mechanistic level. As stated previously, even in healthy people, acute hyperglycemia reduces circulating arginine levels (5). The reduction in arginine and the increase in the competitive inhibitor of eNOS, ADMA, results in eNOS uncoupling, leading to reduced NO and endothelial function as measured by FMD. In our healthy, normal-BMI, nonsmoking population, we did not detect a significant decline in postprandial FMD (Figure 3D). Other studies have shown a small and nonsignificant reduction in FMD during an OGTT (4, 30). Nonetheless, when the AUC was calculated for FMD during the OGTT, we did observe a positive effect for WMJ compared with PLA.

Skeletal muscle is the primary site for glucose disposal in the postprandial state. The microvasculature is responsible for the transport of nutrients to the myocytes. Therefore, skeletal muscle MVBF plays a critical role in maintaining postprandial glucose homeostasis. A 50-g glucose challenge reduced the skeletal muscle blood flow compared to a carbohydrate-matched liquid mixed meal (9). We detected overall differences in the postprandial blood flow AUC between WMJ and PLA treatments during the OGTT (Figure 4B). Skeletal muscle MBVF was reduced by ∼20% at peak glucose (7.8 ± 2.0 mmol/L at the 30-minute time point) after PLA supplementation and by only ∼12% during a similar peak glucose (8.0 ± 1.5 mmol/L) after WMJ supplementation. The ischemic-reperfusion slope, referred to as slope 2 (Figure 2), is considered a microvascular comparison to brachial artery FMD. Previous studies provided data to indicate that slope 2 increases in healthy, normal-weight participants, whereas obese participants display a blunted increase in slope 2 after glucose ingestion. In the current study, we did detect a similar response in slope 2 after glucose ingestion (time effect, P = 0.001; Figure 5D). The consistency between our findings and those of others (31, 10) is interesting, considering we performed our NIRS testing related to a different anatomical site. Soares et al. (10) conducted their measurements at the tibialis anterior muscle, whereas we performed the NIRS measurement at the forearm. The vascular responsiveness in the arms is generally considered to be greater than that in the legs (32). To our knowledge, this is the first report of the measurement of ischemic reperfusion during the OGTT using the NIRS-VOT at the forearm site. Future studies should consider the measurement site when assessing the impact of metabolic challenges and microvascular function.

Microvascular dysfunction is considered a major mediator of cardiometabolic disease and is a functional component of insulin resistance (33–35). The NIRS-VOT provides information regarding the microvascular responsiveness to shear stress and hypoxia (36). After glucose ingestion, the StO₂ reperfusion AUC is reduced in healthy, normal-weight, and obese individuals (31, 10). Hyperglycemia promotes the generation of free radicals that quench NO, disrupt endothelial-dependent vasodilation, and reduce reactive hyperemia (5). Here, we show that when compared to PLA, chronic WMJ supplementation rescues this reduction in microvascular reoxygenation (StO₂ reperfusion AUC; Figure 5D) at similar blood glucose and insulin levels (Figure 3A and B). While we were unable to provide data regarding the arginine-eNOS function during the OGTT, we show that fasting plasma arginine trended to increase after WMJ supplementation compared to PLA treatment (Table 3). Fasting plasma arginine is an essential factor for determining the benefits of arginine-increasing supplementation (29). Our functional NIRS data provide evidence that baseline, peak, and mid-reperfusion (StO₂ AUC) oxygen saturation levels are generally reduced during the OGTT, but WMJ supplementation attenuates this reduction.

The significant first-pass metabolism of oral arginine intake limits its efficacy to increase systemic circulating arginine levels. Citrulline and citrulline-rich watermelon juice can increase the circulating arginine pool and increase the activation of NO synthase to a greater degree than arginine itself (23, 37). Oral citrulline ingestion alone promotes increased NO bioavailability (38, 39). However, the capacity for citrulline to increase blood flow and vascular responsiveness is equivocal (38, 39). Watermelon is rich in antioxidants (lycopene and β-carotene) and potassium, in addition to citrulline (40). Glutathione is also a potent antioxidant enriched in watermelon. When combined with citrulline, glutathione increases NO synthesis markers greater than citrulline alone (41). Therefore, it is conceivable that the additional antioxidants and elevated arginine concentrations are sufficient to induce the physiological changes in blood flow with WMJ compared to citrulline alone. The role of citrulline, independent of its relationship to NO formation, is not well understood. Some data suggest that citrulline reduces reactive oxygen species by scavenging the hydroxyl radicals generated during hyperglycemia or hyperlipidemia (42). Recently, citrulline has been identified as a potential biomarker for metabolic health and disease (43, 44). The role of citrulline supplementation in targeting metabolic dysfunction (i.e., obesity and insulin resistance) warrants further investigation.

Some limitations of our study include the inability to acquire data regarding shear stress and pulse wave velocity and the limited control over dietary patterns and physical activity. The lack of a Doppler velocity signal limits our ability to account for blood velocity (shear rate) and prevents us from reporting normalized FMD (FMD/shear) data. Furthermore, the normalization of NIRS-derived data from the measurement of the shear rate would provide important mechanistic data. We asked participants to repeat their 3-day food diary before their follow-up OGTT. We also required participants to refrain from exercise for 3 days before OGTT visits. Since diet and exercise do play roles in glucose tolerance and vascular responsiveness, a greater level of normalization may be required to reduce the confounding effects of these factors. Another limitation we must acknowledge is the lack of statistical power to investigate sex as a biological variable. Future studies are required to inspect the sex-related difference in microvascular function. Despite these limitations, our study has several strengths. We utilized a common method, OGTT, to induce acute hyperglycemia, which has been shown to promote acute vascular dysfunction in healthy individuals. To that point, we monitored brachial artery and skeletal muscle microvascular functions in a simultaneous manner, repeatedly, during the OGTT. We also used a natural source, WMJ, to increase circulating arginine levels. This natural product approach increases the accessibility of our supplementation protocol.

In conclusion, our data suggest that WMJ can preserve brachial artery FMD (endothelial function) and skeletal muscle microvascular oxygen saturation and blood flow during experimentally induced hyperglycemia (OGTT). While we lack some details regarding oxidative stress statuses during hyperglycemia, we show a trend for an increase in fasting plasma arginine after WMJ supplementation, which suggests that eNOS activity may be increased.

Notes

This publication was supported by the National Watermelon Promotional Board, National Center for Complementary and Integrative Health, and the Office of Dietary Supplements of the NIH under Award Number P50AT002776, U54 GM104940. TDA is supported by NCCIH T32 AT004094. BAI is supported by NIH R21AG058181.

Author disclosures: TDA is supported by NCCIH T32 AT004094. BAI is supported by NIH R21AG058181. All other authors report no conflicts of interest.

The presented work was supported in part by a grant from the National Watermelon Promotional Board, Orlando, FL, USA. The funding sponsor had no role in the decision of the study design, data collection, analysis, interpretation, writing of the manuscript, or the involvement in the publication process.

Abbreviations used: ADMA, asymmetric dimethylarginine; CVD, cardiovascular disease; FMD, flow-mediated dilation; eNOS, endothelial nitric oxide synthase; MVBF, microvascular blood flow; NEFA, nonesterified fatty acids; NIRS, near-infrared spectroscopy; OGTT, oral-glucose-tolerance test; PLA, placebo; StO₂, tissue oxygen saturation; T2DM, type 2 diabetes; VOT, vascular occlusion test; WMJ, watermelon juice.