Branched-Chain Amino Acids in relation to Food Preferences and Insulin Resistance in Obese Subjects Consuming Walnuts: A Cross-Over, Randomized, Double-Blind, Placebo-Controlled inpatient physiology study.

Abstract

Background&Aims:

To assess whether the concentrations of circulating Branched-Chain Amino Acids (BCAAs) change after walnut consumption and, whether these changes are associated with alterations in markers of insulin resistance and food preferences.

Methods:

In a crossover, randomized, double-blind, placebo-controlled study, ten subjects participated in two 5-day inpatient study admissions, during which they had a smoothie containing 48 g walnuts or a macronutrient-matched placebo smoothie without nuts every morning. Between the two phases there was a 1-month washout period.

Results:

Fasting valine and isoleucine levels were reduced (p = .047 and p <.001) and beta-hydroxybutyrate levels were increased after 5-days of walnut consumption compared to placebo (p = .023). Fasting valine and isoleucine correlated with HOMA-IR while on walnut (r = 0.709, p = .032 and r = 0.679, p = .044). The postprandial area under the curve (AUC) of leucine in response to the smoothie consumption on day 5 was higher after walnut vs placebo (p = .023) and correlated negatively with the percentage of Kcal from carbohydrate and protein consumed during an ad libitum buffet meal consumed the same day for lunch (r = −0.661, p = .037; r = −0.628, p = .05, respectively).

Conclusion:

The fasting and postabsorptive profiles of BCAAs are differentially affected by walnut consumption. The reduction in fasting valine and isoleucine may contribute to the longer-term benefits of walnuts on insulin resistance, cardiovascular risk and mortality, whereas the increase in postabsorptive profiles with walnuts may influence food preference.

Introduction

Obesity is a disease with continuously increasing prevalence that is associated with higher mortality, mainly due to cardiovascular and diabetes-related complications¹. A key mechanism involved in the detrimental effects of obesity in human health is the development of insulin resistance (IR)¹. IR refers to the lower response of cells to insulin, that leads to reduced glucose uptake from muscle and to inadequate suppression of glucose output from the liver¹. In the short-term, the impaired response to insulin is compensated by higher secretion of insulin from the pancreas, but in the long-term it leads to higher blood glucose levels and diabetes¹.

Walnuts act beneficially in human metabolism by providing cardiovascular protection, improving insulin sensitivity² and reducing the risk for obesity and diabetes³. As we have previously demonstrated, the beneficial effects of walnuts are, at least partially, achieved through induction of satiety^{2, 4}, increase in salience and cognitive control processing of highly desirable food cues² and reduction of lipids that promote atherogenesis and insulin resistance (e.g., small and dense LDL and ceramides)².

Walnuts (Juglans regia) contain polyphenols, fibers, health-promoting fats, proteins and are particularly rich in branched chain amino acids (BCAAs)⁵. In interventional studies, high dietary intake of BCAAs has been associated with improvements in body composition, glycaemia and satiety⁶. In contrast, in observational studies, elevated circulating BCAA levels have been associated with insulin resistance, obesity and type 2 diabetes (T2D)⁶. It has been suggested that the increased BCAA levels in these pathological conditions do not derive from increased dietary uptake but increased protein degradation and impaired BCAA metabolism⁶.

In this context, we aimed in the current study to assess, whether part of the beneficial effects of walnuts in human metabolism are mediated through BCAAs and specifically whether walnut consumption affects BCAA levels and whether the changes observed in BCAA levels with walnuts are associated with alterations in food preferences and markers of insulin resistance.

Materials and Methods

Ten obese individuals were enrolled in a randomized (1:1), double-blind, placebo-controlled, cross-over, inpatient study, with random order of first admission. Participants consumed either 48g of walnuts (similar to the recommended daily dose) or “placebo”. The study was approved by the Beth Israel Deaconess Medical Center (BIDMC) Institutional Review Board. All subjects signed written informed consent. The study design has been previously described (NCT02673281)². Briefly, subjects were admitted at the Clinical Research Center of the BIDMC for 5 days during each phase (walnut or placebo, e.g., subjects who received the placebo smoothie in the first phase received the walnut smoothie in the second phase, or vice-versa). Safflower oil and walnut flavoring replaced walnuts in placebo smoothie. They consumed the walnut and placebo in the form of a smoothie at breakfast for the entire duration of each arm of the inpatient study (5 days), with the same macronutrient composition, allowing double-blinding, as previously described^{7, 8} (Supplemental figure 1). During both inpatient 5-day stays, subjects followed an isocaloric diet to minimize variability (Supplemental Table 1). Baseline measurements and blood draw after an overnight fast were performed on day 1. The same measurements were repeated on day 5 and along with blood draws, at 0, 30, 60, 120, 180 minutes after smoothie consumption. Patients also had an ad libitum or weighed buffet meal to assess caloric consumption and food preferences at day five for lunch⁴. Participants were not allowed to consume nuts during the one-month washout between the two arms of the study.

Biochemical Measurements

Blood samples drawn through venipuncture were processed for plasma or serum and stored at −80 °C until assayed in duplicate. Measurements were acquired using the following techniques: BCAAs, alanine, glucose, citrate and ketone bodies (β-hydroxybutyrate, acetoacetate and acetone) were measured using a Vantera Clinical Analyzer, a 400-MHz proton nuclear magnetic resonance (NMR) spectrometer, as previously described^{9, 10}.

Statistical analysis

The Statistical Package for Social Sciences (SPSS), v.19, was used for statistical analysis. A general linear mixed model followed by LSD post hoc analysis and Pearson correlations was used. The sample size has been calculated on a previously published fMRI outcome². The analysis is described in detail in Appendix S1.

Results

General characteristics of the subjects (6 men and 4 women, mean age 50.7±2.3 years) are shown in supplemental table 2.

On the 5^th day of walnut consumption and despite the high amounts of BCAAs included in walnuts (supplemental Figure 1), fasting blood levels of valine and isoleucine were significantly decreased in the walnut group compared to placebo (p=.047 and p<.001 respectively; Table 1). In contrast, fasting β-hydroxybutyrate concentration was significantly increased in the walnut group (p=.023; Table 1) while acetoacetate and acetone concentration were not different between the two groups (Table 1). While on walnuts, fasting valine and isoleucine were significantly and positively correlated with the HOMA-IR calculated on the same day (r=0.709, p = .032 and r=0.679, p=.044, respectively, Figure 1 A–B).

Table 1.

Fasting levels (before and after 5 days of treatment) and postprandial AUCs (after smoothie consumption on day 5) of BCAA, ketone bodies, glucose and citrate with walnut vs placebo treatment.

	Placebo		Walnut		p
	Day 1	Day 5	Day 1	Day 5
Fasting measurements
Amino Acid Concentrations
Branched-Chain Amino Acids (BCAA) (μmol/L)	375.00 ± 19.85	390.3 ± 22.42	379.9 ± 23.61	374.9 ± 29.23	.180
Valine (μmol/L)	221.00 ± 10.64	228.00 ± 12.19	217.00 ± 11.65	213.60 ± 14.82	.047
Leucine (μmol/L)	118.44 ± 10.35	117.30 ± 6.30	113.00 ± 10.00	115.60 ± 12.52	.651
Isoleucine (μmol/L)	38.63 ± 6.02	49.56 ± 6.12	53.89 ± 4.20	45.90 ± 6.69	.001
Alanine (μmol/L)	310.56 ± 18.79	382.10 ± 23.26	333.90 ± 30.95	388.20 ± 25.39	.931
Ketone Bodies
Ketone Body Concentrations (μmol/L)	179 ± 28.73	106.61 ± 29.96	114.94 ± 13.19	114.71 ± 16.80	.160
Beta-hydroxy-butyrate (μmol/L)	119.88 ± 23.63	73.42 ± 21.99	69.01 ± 8.01	80.62 ± 12.90	.023
Aceto-acetate (μmol/L)	20.93 ± 5.00	8.44 ± 2.24	16.83 ± 3.73	10.32 ± 4.50	.928
Acetone (μmol/L)	38.19 ± 4.61	24.75 ± 8.38	29.11 ± 6.96	23.79 ± 3.87	.772
Small Molecule Metabolites
Glucose (mg/dL)	95.34 ± 2.49	91.37 ± 2.69	97.94 ± 2.82	93.08 ± 4.03	.918
Citrate (mg/dL)	100.11 ± 13.93	109.30 ± 7.64	99.50 ± 8.31	113.00 ± 10.77	.863
Smoothie Postprandial AUC measurements
Amino Acid Concentrations
Branched-Chain Amino Acids (BCAA) (μmol/L*min)		52256 ± 2384		57137 ± 3061	.004
Valine (μmol/L*min)		33156 ± 1255		34461 ± 1394	.060
Leucine (μmol/L*min)		14528 ± 1044		17163 ± 1305	.023
Isoleucine (μmol/L*min)		5538 ± 675		5801 ± 854	.350
Alanine (μmol/L*min)		66543 ± 2366		67875 ± 3756	.630
Ketone Bodies
Ketone Body Concentrations (μmol/L*min)		23316 ± 2503		20325 ± 2114	.069
Beta-hydroxy-butyrate (μmol/L*min)		14876 ± 1986		12774 ± 1941	.091
Aceto-acetate (μmol/L*min)		3949 ± 507		3810 ± 344	.790
Acetone (μmol/L*min)		4535 ± 621		4027 ± 449	.150
Small Molecule Metabolites
Glucose (mg/dL*min)		18123.6 ± 476		17740.8 ± 350	.292
Citrate (mg/dL*min)		24198 ± 1851		24606 ± 1665	.737

Figure 1.

Correlation plots of HOMA-IR with fasting levels of valine (A) and isoleucine (B) after 5 days of walnut consumption. Circulating levels of BCAA (C) and leucine (D) after smoothie intake (placebo or walnut) on the 5^th day of treatment. Correlation plots of AUCs of leucine after walnut smoothie intake with percentage of Kcal from carbohydrate (E) and protein (F) in an ad libitum lunch the same day (5^th day of treatment). Means0020± standard error (SE) of the mean are demonstrated for (C) and (D); r = correlation coefficient. *, ** indicate p<.05 and <.01 respectively in Fisher’s LSD test for walnut vs placebo for each timepoint in C and D.

Conversely to what was observed in fasting conditions, blood levels of BCAAs were higher right after consuming a walnut smoothie compared to a placebo one, as assessed by the total area under the curve (AUC) (p<.004, Table 1, Figure 1C). Specifically, BCAA levels decreased rapidly after the first 30’ post smoothie consumption in the placebo group while they remained significantly higher overtime in the walnut group (Table 1, Figure 1C). Among BCAAs, the post smoothie leucine profile had the most significant increase in the walnut group compared to placebo (p<.023, Table 1, Figure 1D). The placebo group exhibited lower leucine levels until the end of the 180-minute experimental period. While valine concentration showed a tendency towards increase after the walnut smoothie consumption compared to placebo (p=.06), post smoothie isoleucine and alanine were not different between the groups (Table 1). Importantly, leucine AUC was negatively correlated with the percentage of Kcal deriving from carbohydrate and protein consumed during the ad libitum buffet meal (r=−0.661, p=.037; r=−0.628, p=.05, respectively, Figure 1 E–F). No significant difference was found in glucose and citrate concentrations between the two groups.

Discussion

We demonstrate herein that after 5 days of walnut consumption, fasting levels of BCAAs and specifically of valine and isoleucine are lower compared to placebo and are associated with HOMA-IR. In contrast, directly after consumption of a walnut smoothie for breakfast on the 5^th day, blood levels of leucine are higher compared to placebo and they are negatively associated with carbohydrate and protein intake in the subsequent ad libitum buffet meal consumed for lunch.

Our results extend previous findings reporting a significant role of walnut consumption on appetite control^{2, 4}, suggesting that single amino acids naturally present in walnuts may play a role in macronutrient preference thus leading to health-promoting food choices. Protein sources rich in BCAAs elicit increased satiety in rodents and humans, supporting the indication that blood leucine concentration may deliver information regarding the anabolic value of the food to modulate the appetite control in the brain. Direct leucine infusion into the duodenum was shown to induce satiety in normal-weight individuals, especially at higher infusion rates compared to placebo, whereas animal data suggest that these effects are mediated through activation of leucine-sensitive or POMC and NPY&AGRP neurons in the hypothalamus^{11, 12}. The increased leucine concentration after walnut consumption, through the aforementioned central mechanisms, could inform the brain about protein content of the meal, and consequently, may guide the subject in choosing less carbohydrate and protein-based food afterwards.

Circulating BCAA levels usually are only shortly affected by dietary intake, whereas their fasting levels are mainly defined by tissue protein degradation and by rates of BCAA disappearance⁶. Insulin is a known potent inhibitor of muscle tissue protein degradation. Thus, resistance to its function may explain the increase in protein degradation and the impairmentin BCAA excretion observed in obesity, T2D and cardiovascular disease⁶. Consequently, the reduction in fasting BCAA levels with walnuts shown in our study may reflect a decrease in protein degradation and, an enhancement of BCAA excretion through an improvement in insulin resistance. The above conclusion is further supported by the association of fasting valine and isoleucine levels with HOMA-IR. Finally, we observe a significant increase in the blood concentrations of ketone bodies and specifically of β-hydroxybutyrate. This is also indicative of improved BCAA metabolism, although lower carbohydrate intake or increased fatty acid oxidation may also contribute to higher ketone body formation.

Our study has some significant strengths and limitations. This study is strengthened by the use of a placebo/walnut smoothie delivery system, which has been previously validated and tested, allowing for double-blinding^{7, 8}. Another strength is the inpatient setting, with standardized study conditions, which ensured compliance. The cross over study of the same subjects, in random order, minimizes subject related biological variability, increases power and decreased potential baseline differences and confounders between the subjects. Considering the small number of subjects and the a priori power calculation, this study produced statistically significant results. Laboratory assays were run blindly by laboratory personnel that was not aware of the study hypotheses. In this study, many variables have been analyzed; however, they were treated as discrete hypotheses. We herein applied, for the first time, an NMR analysis to measure BCAAs to investigate the mechanisms underlying the previously published positive effect of dietary walnut consumption on appetite in obese subjects.

Conclusion

Our findings 1) indicate the relevance of the BCAAs in the postprandial state in regulating walnut-related food preference controlling mechanisms and, 2) may indicate that the reduction of fasting valine and isoleucine after five days of walnut consumption could represent an early positive effect of possible future decreased insulin resistance and diabetes risk. Given the exploratory nature of our study, future, broader studies are needed to confirm results herein presented on the walnut-mediated mechanisms that improve appetite control.

Abbreviations:

AGRP

Agouti-Related Peptide

BMI

Body Mass Index

fMRI

functional Magnetic Resonance Imaging

GLP-1

Glucagon-Like Peptide-1

HOMA-IR

Homeostatic Model Assessment for Insulin Resistance

NMR

Nuclear Magnetic Resonance

NPY

Neuropeptide Y

POMC

Pro-Opiomelanocortin

PYY

Peptide YY