Perspective: A Research Roadmap about Ultra-Processed Foods and Human Health for the United States Food System: Proceedings from an Interdisciplinary, Multi-Stakeholder Workshop.

Our objective was to convene interdisciplinary experts from government, academia, and industry to develop a Research Roadmap to identify research priorities about processed food intake and risk for obesity and cardiometabolic diseases (CMD) among United States populations. We convened attendees at various career stages with diverse viewpoints in the field. We held a "Food Processing Primer" to build foundational knowledge of how and why foods are processed, followed by presentations about how processed foods may affect energy intake, obesity, and CMD risk. Breakout groups discussed potential mechanistic and confounding explanations for associations between processed foods and obesity and CMD risk. Facilitators created research questions (RQs) based on key themes from discussions. Different breakout groups convened to discuss what is known and unknown for each RQ and to develop sub-RQs to address gaps. Workshop attendees focused on ultra-processed foods (UPFs; Nova Group 4) because the preponderance of evidence is based on this classification system. Yet, heterogeneity and subjectivity in UPF classification was a challenge for RQ development. The 6 RQs were: 1) What objective methods or measures could further categorize UPFs, considering food processing, formulation, and the interaction of the two? 2) How can exposure assessment of UPF intake be improved? 3) Does UPF intake influence risk for obesity or CMDs, independent of diet quality? 4) What, if any, attributes of UPFs influence ingestive behavior and contribute to excess energy intake? 5) What, if any, attributes of UPFs contribute to clinically meaningful metabolic responses? 6) What, if any, external environmental factors lead people to consume high amounts of UPFs? Uncertainty and complexity around UPF intake warrant further complementary and interdisciplinary causal, mechanistic, and methodological research related to obesity and CMD risk to understand the utility of applying classification by degree of processing to foods in the United States.

Partitioning of caffeine and quinine in oil-in-water emulsions and effects on bitterness.

The bulk vegetable oil-water partition coefficient of caffeine and quinine was determined by a shake-flask method as log Kow  = -1.32 and 2.97. These values were consistent with the effect of oil concentration on the distribution of the bitterants in an oil-in-water emulsion (0-2 and 0-20 wt% oil stabilized with 0.125 and 1 wt% whey protein isolate, respectively). For example, in a 20% o/w emulsion, approximately 90% of the total caffeine remained in the aqueous phase, whereas in a 2% o/w emulsion, only ∼20% of the quinine remained in the aqueous phase. The intensity of the bitter taste of caffeine and quinine in emulsions was assessed by a large cohort (n = 100) of untrained participants. An increase in fat in the emulsions (from 0.5 wt% to 2 wt% oil emulsions stabilized with 0.125 wt% whey protein isolate) caused a significant decrease in perceived bitterness that was accompanied by a decrease in the aqueous concentration of the hydrophobic bitterant quinine Specifically, the bitterness of quinine was reduced ∼13% in the o/w emulsion with more fat, and this drop paralleled a drop in the aqueous concentration and was generally consistent with aqueous dose-response functions published elsewhere. For the hydrophilic bitterant caffeine, there was no significant change in the perceived bitterness or aqueous concentration with changing oil concentration. We conclude that the perceived bitterness of a hydrophobic bitterant like quinine in an emulsion depends on the aqueous concentration rather than the overall concentration.

The effect of dairy proteins on the oral burn of capsaicin.

This study focused on the effect of binding of capsaicin by milk proteins on oral burn. The concentration of free, unbound capsaicin in 5 ppm capsaicin solution containing 0-5% (w/w) micellar casein or whey protein isolate was measured by extraction into hexadecane. The concentration of free capsaicin decreased linearly with protein concentration and the decrease was greater for casein than for whey protein. The intensity of the capsaicin burn in similar solutions was assessed by a large cohort (n = 89) of untrained participants in a time-intensity study. The maximum burn intensity decreased with protein concentration and was lower for samples containing casein than for samples containing whey protein isolate. The maximum burn was linearly related to the free, unbound capsaicin concentration. When protein solutions (1-5% w/w) were used as rinses following exposure to a 5 ppm aqueous capsaicin solution, only the 5% (w/w) micellar casein solution was significantly more effective than the water rinse in reducing oral burn.

Effect of alkyl chain length on the antioxidant activity of alkylresorcinol homologues in bulk oils and oil-in-water emulsions.

The antioxidant cut-off theory details the importance of fine-tuning antioxidant hydrophobicity to optimize antioxidant effectiveness for a given food system; however, previous research has utilized synthetic antioxidant homologues which fail to align with the food industry's demand for natural ingredients. Alkylresorcinols represent a natural homologous series of phenolipid antioxidants. The antioxidant activities of individual alkylresorcinol homologues were investigated in bulk oils and oil-in-water emulsions. In oils, antioxidant activity decreased as alkyl chain length increased and there was no effect on rate of loss. In emulsions, optimum antioxidant activity was observed at intermediate alkyl chain length (C21:0) and longer homologues were lost more rapidly. Radical scavenging capacity decreased as alkyl chain length increased but alkylresorcinols were unable to chelate iron. This suggests that intrinsic properties (e.g. radical scavenging capacity) are responsible for the antioxidant activity of alkylresorcinols in oils while physicochemical phenomena (e.g. partitioning) drive antioxidant activity of alkylresorcinols in emulsions.

Effect of high-pressure-jet processing on the physiochemical properties of low-fat ice cream mix.

The objective of this study was to use high-pressure-jet (HPJ) processing to produce functional properties in a low-fat (4.5% fat) ice cream mix similar to those seen when emulsifiers are used. Ice cream mix or serum (nonfat portion of the ice cream mix) were subjected to 200 or 400 MPa HPJ processing and compared with a non-HPJ-treated control. A similar non-HPJ-treated formulation but containing polysorbate 80 (0.075% wt/wt) was also used as a control. The mix samples were characterized in terms of their particle size, density, flow properties, stability, crystallization kinetics, and fat-protein interactions. The sample from the mix subjected to 400 MPa HPJ processing (HPJ-M-400) had increased consistency coefficient (5°C; 228 ± 102.7 mPa·s) and particle size (D[4,3]; 16.0 ± 2.5 µm) compared with the non-HPJ-treated control sample, with viscosity and particle size (volume-moment mean diameter, D[4,3]) values of 7.5 ± 0.4 mPa·s and 0.50 ± 0.1 µm, respectively. These differences were attributed to an increase in casein-fat interactions and casein-casein interactions caused by the 400 MPa HPJ treatment, which were observed using confocal scanning laser microscopy and inferred from an increase in protein and fat concentrations in the sediment after ultracentrifugation. Interestingly, the density of HPJ-M-400 was also lower (0.79 ± 0.17 g/mL) than that of the control (1.04 ± 0.00 g/mL) because bubbles were trapped within these complexes. The large casein-fat complexes formed in the HPJ-M-400 sample also appeared to act as steric barriers that slowed ice crystal growth during quiescent freezing. The alterations in physiochemical properties and apparent ice crystal growth induced by the 400 MPa treatment of low-fat ice cream mix have many potential applications, including clean-label confections.

Antioxidant activity of a winterized, acetonic rye bran extract containing alkylresorcinols in oil-in-water emulsions.

Naturally derived antioxidants are in high demand as the food industry strives to meet consumer preferences for non-synthetic additives. Alkylresorcinols (ARs) represent a novel class of natural antioxidants that can be derived from a natural waste stream (bran) and have the potential to inhibit lipid peroxidation given their phenolic structure. The antioxidant activity of rye bran extract containing ARs was investigated in an oil-in-water emulsion and was found to inhibit lipid oxidation reactions. The concentration of ARs in the continuous phase of emulsions was measured to understand partitioning behavior, as this is known to impact antioxidant activity. It was found that a majority of the ARs were associated with the lipid phase and those in the continuous phase were associated with surfactant micelles, perhaps inhibiting their interaction with water-soluble pro-oxidants. These results show that a rye bran extract containing ARs can function as a radical scavenging antioxidant in lipid dispersions.

Effect of fat content on the physical properties and consumer acceptability of vanilla ice cream.

Ice cream is a complex food matrix that contains multiple physical phases. Removal of 1 ingredient may affect not only its physical properties but also multiple sensory characteristics that may or may not be important to consumers. Fat not only contributes to texture, mouth feel, and flavor, but also serves as a structural element. We evaluated the effect of replacing fat with maltodextrin (MD) on select physical properties of ice cream and on consumer acceptability. Vanilla ice creams were formulated to contain 6, 8, 10, 12, and 14% fat, and the difference was made up with 8, 6, 4, 2, and 0% maltodextrin, respectively, to balance the mix. Physical characterization included measurements of overrun, apparent viscosity, fat particle size, fat destabilization, hardness, and melting rate. A series of sensory tests were conducted to measure liking and the intensity of various attributes. Tests were also conducted after 19 weeks of storage at -18°C to assess changes in acceptance due to prolonged storage at unfavorable temperatures. Then, discrimination tests were performed to determine which differences in fat content were detectable by consumers. Mix viscosity decreased with increasing fat content and decreasing maltodextrin content. Fat particle size and fat destabilization significantly increased with increasing fat content. However, acceptability did not differ significantly across the samples for fresh or stored ice cream. Following storage, ice creams with 6, 12, and 14% fat did not differ in acceptability compared with fresh ice cream. However, the 8% fat, 6% MD and 10% fat, 4% MD ice creams showed a significant drop in acceptance after storage relative to fresh ice cream at the same fat content. Consumers were unable to detect a difference of 2 percentage points in fat level between 6 and 12% fat. They were able to detect a difference of 4 percentage points for ice creams with 6% versus 10%, but not for those with 8% versus 12% fat. Removing fat and replacing it with maltodextrin caused minimal changes in physical properties in ice cream and mix and did not change consumer acceptability for either fresh or stored ice cream.

Binding of Caffeine and Quinine by Whey Protein and the Effect on Bitterness.

Many drugs and phytochemicals are bitter, leading to noncompliance with prescriptions and avoidance of healthy foods and a need to suppress their taste. The goal of this study was to investigate the binding of bitterants (quinine and caffeine) by whey protein isolate (WPI) and the effect on perceived bitterness. Caffeine interacted minimally with WPI, while the proportion of unbound quinine decreased exponentially with protein concentration. Molecular modeling was used to show the energy of the quinine-Β-lactoglubulin interaction was an order of magnitude greater than the caffeine-Β-lactoglobulin interaction. Untrained assessors were used to assess the bitterness of caffeine (1.8, 5.7, and 18 mM) and quinine (0.056, 0.10, and 0.18 mM) solutions with 0% or 1% WPI. There was no significant effect of protein on the bitterness of caffeine solutions, but WPI decreased the bitterness of quinine relative to the same concentration in water. This is generally consistent with our hypothesis that higher binding results in lower bitterness; however the magnitude of reduction was not large and the bitterness of the protein-quinine solutions was greater than would be expected for the unbound quinine present.

Essential Oils Against Pathogen and Spoilage Microorganisms of Fruit Juices: Use of Versatile Antimicrobial Delivery Systems.

Essential oils (EO) are increasingly used as natural antimicrobial compounds, however the effect of delivery system to enhance their antimicrobial activity has not been widely studied. Limonene (0 to 10 µL/mL) was added to microbial suspensions (∼105 CFU/mL) of selected foodborne pathogens (Listeria monocytogenes Scott A, Salmonella enterica Typhimurium, Escherichia coli and Staphylococcus aureus), and spoilage microorganisms (Lactobacillus plantarum, Saccharomyces cerevisiae, and Candida albicans). S. aureus was found to be the most sensitive foodborne pathogen while Salmonella enterica showed continued growth under all concentrations. Stable nanoemulsions and solid lipid nanoparticles (SLN) (d ∼ 170 nm) were prepared using an alkane carrier oil (n-tetradecane and n-eicosane, respectively). Interfacial effects and homogenous distribution of limonene in nanoemulsions improved its (8 and 12 µL/mL) antimicrobial effect against S. aureus. Higher aqueous concentrations as a result of expulsion from SLN further enhanced the antimicrobial activity pronounced at higher limonene concentrations. Therefore, our findings confirm that the emulsion-based delivery systems are able to effectively distribute limonene inside a microbial suspension to improve its antimicrobial activity.

Release Kinetics of Nisin from Chitosan-Alginate Complex Films.

Understanding the release kinetics of antimicrobials from polymer films is important in the design of effective antimicrobial packaging films. The release kinetics of nisin (30 mg/film) from chitosan-alginate polyelectric complex films prepared using various fractions of alginate (33%, 50%, and 66%) was investigated into an aqueous release medium. Films containing higher alginate fractions showed significantly lower (P < 0.05) degree of swelling in water. Total amount of nisin released from films into an aqueous system decreased significantly (P < 0.05) with an increase in alginate concentration. The mechanism of diffusion of nisin from all films was found to be Fickian, and diffusion coefficients varied from 0.872 × 10-9 to 8.034 ×10-9 cm2 /s. Strong complexation was confirmed between chitosan and alginate polymers within the films using isothermal titration calorimetry and viscosity studies, which affects swelling of films and subsequent nisin release. Complexation was also confirmed between nisin and alginate, which limited the amount of free nisin available for diffusion from films. These low-swelling biopolymer complexes have potential to be used as antimicrobial packaging films with sustained nisin release characteristics.

Assessing Interactions between Lipophilic and Hydrophilic Antioxidants in Food Emulsions.

Dietary lipids containing high concentrations of polyunsaturated fatty acids are considered to be beneficial to human health, yet their incorporation within formulated foods is complicated by their susceptibility to oxidation. Lipid oxidation in foods is inhibited through the incorporation of antioxidants, yet the list of antioxidants approved for food use is small, and consumers frequently demand foods without synthetic additives. As a consequence, food processors are now tasked with improving the efficacy of approved, "natural" (i.e., nonsynthetic) antioxidants; a rational strategy for doing so involves localizing the antioxidants at the interface where oxidation usually occurs and regenerating the consumed antioxidants after the oxidation event has occurred. The present study describes a procedure to evaluate antioxidant interactions in oil-in-water food emulsions, which is based on controlled oxidation reactions induced in the dispersed oil phase by the lipophilic radical generator, 2,2'-azobis(2,4-dimethylvaleronitrile). The extent of lipid oxidation is measured spectroscopically by following the loss of an oxidatively labile, lipophilic probe (methyl eleostearate), the synthesis of which is described here. Using this procedure, the ability of various aqueous phase solvated antioxidants (ascorbic acid, gallic acid, (-)-epicatechin, (-)-epigallocatechin-3-gallate) to regenerate lipid phase solvated α-tocopherol was evaluated. In all cases, the test compounds were able to inhibit oxidation reactions; however, these effects were not profoundly synergistic, and the maximum synergistic interaction observed was only ∼ 3% using ascorbic acid.

Effect of lipophilization on the distribution and reactivity of ingredients in emulsions.

HYPOTHESIS: The reactivity of small molecules in emulsions is believed to depend on their partitioning between phases, yet this is hard to verify experimentally in situ. In the present work, we use electron paramagnetic resonance (EPR) spectroscopy to simultaneously measure the distribution and reactivity of a homologous series of lipophilized spin probes in an emulsion. EXPERIMENTS: 4-Hydroxy-2,2,6,6-tetramethylpiperidin-1-oxyl (TEMPOL) was derivatized with saturated fatty acids to create a series of spin probes with increasing lipophilicity (C4-, C8-, C12-, and C16-TEMPO). The probes were added to a 10 wt.% tetradecane-in water emulsions (d32∼190 nm) stabilized with sodium caseinate (1 wt.% in the aqueous phase, pH 7). The distribution of the probes between phases was measured by electron paramagnetic resonance (EPR) spectroscopy. FINDINGS: TEMPOL partitioned into the aqueous phase, C4-TEMPO distributed between the lipid and aqueous phases (69% and 31% respectively) while the more lipophilic probes dissolved exclusively within the lipid droplets. Interestingly, the more lipophilic probes initially precipitated upon their addition to the emulsion, and only slowly redistributed to the droplets over hours or days, the rate of which was dependent on their carbon chain length. The reactivity of the probes with aqueous an aqueous phase reductant (ascorbate) generally depended on the proportion in the aqueous phase (i.e., TEMPOL>C4-TEMPO>C8-TEMPO∼C12-TEMPO∼C16-TEMPO).

Physical approaches to masking bitter taste: lessons from food and pharmaceuticals.

Many drugs and desirable phytochemicals are bitter, and bitter tastes are aversive. Food and pharmaceutical manufacturers share a common need for bitterness-masking strategies that allow them to deliver useful quantities of the active compounds in an acceptable form and in this review we compare and contrast the challenges and approaches by researchers in both fields. We focus on physical approaches, i.e., micro- or nano-structures to bind bitter compounds in the mouth, yet break down to allow release after they are swallowed. In all of these methods, the assumption is the degree of bitterness suppression depends on the concentration of bitterant in the saliva and hence the proportion that is bound. Surprisingly, this hypothesis has only rarely been fully tested using a combination of adequate human sensory trials and measurements of binding. This is especially true in pharmaceutical systems, perhaps due to the greater experimental challenges in sensory analysis of drugs.

Localization and reactivity of a hydrophobic solute in lecithin and caseinate stabilized solid lipid nanoparticles and nanoemulsions.

The distribution and reactivity of the lipophilic spin probe 4-phenyl-2,2,5,5-tetramethyl-3-imidazoline-1-oxyl nitroxide (PTMIO) in tetradecane (C14)- and eicosane (C20)-in-water emulsions and solid lipid nanoparticles (SLN) respectively, were investigated by electron paramagnetic resonance (EPR) spectroscopy. The lipid phase (10 wt% C14 or C20) was emulsified into either caseinate solutions (1 wt%) or lecithin+bile salt dispersions (2.4 wt%+0.6 wt%) at 70-75 °C. In C14 emulsions stabilized with lecithin+bile salt, three populations of PTMIO were observed: a population in the lipid phase (~60%, a(N)~13.9 G), an aqueous phase population (~20%, a(N)~15.4 G) with high mobility, and an immobilized surface layer population (~20%, a(N)~14.2 G) with low mobility. However, in C14 emulsions stabilized by caseinate, only two distinct populations of PTMIO were seen: a lipid phase population (~70%, a(N)~13.8 G) and an aqueous phase population (~30%, a(N)~15.5 G) with high mobility. In C20 SLN stabilized with either lecithin+bile salt or caseinate, PTMIO was excluded from the lipid phase. In lecithin+bile salt-stabilized C20 SLN, the majority of the probe (~77%) was in the interfacial layer. For both surfactant systems the rate of PTMIO reduction by aqueous iron/ascorbate was greater for C20 SLN than C14 emulsions. Lecithin affects the properties of emulsions and SLN as delivery systems by providing a distinct environment for small molecules.

Solute distribution and stability in emulsion-based delivery systems: an EPR study.

Oil-in-water emulsions and related systems are often used to deliver hydrophobic solutes in foods, personal care products, and pharmaceuticals. Recent work has considered the use of crystalline lipid carrier particles (i.e., solid lipid nanoparticles, SLN) to control the availability of the solute; however, there is little direct evidence for the localization of small molecules in these systems. Alkanes (10 wt.% tetradecane or eicosane) containing the spin probe 4-phenyl-2,2,5,5-tetramethyl-3-imidazoline-1-oxyl (PTMIO, 200 ppm) were homogenized into sodium caseinate solution (1 wt.%) to produce fine or coarse droplets (0.2 µm or 1.3 µm, respectively) and cooled to 21.5 °C where eicosane is crystalline and tetradecane is liquid. Analysis of the resulting EPR spectra revealed populations of probe in two discrete environments (i.e., aqueous and lipid). PTMIO is largely hydrophobic with 77% and 70% present in the coarse and fine liquid lipid droplets (i.e., tetradecane droplets), respectively. In the solid droplets (i.e., eicosane), all of the probe was excluded from the droplets into the aqueous environment. In all cases, the mobility of the probe in both lipid and aqueous environments was affected by the droplet surface; thus, we hypothesize that the majority of the probe molecules are associated with the droplet interface. The PTMIO was reduced to an EPR-silent form by the addition of iron/ascorbate to the aqueous phase, and the apparent rate constant of the reaction was proportional to the fraction of the spin probe in the aqueous phase. Based on these findings, we propose that droplet crystallization excludes solute molecules from the droplet core to the aqueous environment where they interact with the droplet surface.

Ultrasonic characterization of lactose crystallization in gelatin gels.

Ultrasonic velocity and attenuation measurements (2.25 MHz center frequency) were used to follow bulk crystallization of lactose (43% and 46%) from gelatin (1.5% and 3%) gels at 25 °C, and compared to turbidity (500 nm) and isothermal calorimetric measurements. Ultrasonic velocity decreased slightly (approximately 0.5%) during crystallization while ultrasonic attenuation was low in the absence of lactose crystals and increased progressively during crystallization. The lag time before the onset of crystallization decreased and the maximum rate of increase in attenuation during crystallization increased with increasing lactose supersaturation but was not affected by gelatin concentration (P < 0.05). Similar results were seen in turbidity and isothermal calorimetric measurements. Ultrasonic attenuation measurements have the potential to measure crystallization kinetics in complex food matrices and to be applied on-line. Practical Application: Many foods contain crystals that affect their taste and texture (for example, lactose crystals can give a grainy defect in ice cream). The formation of crystals is often hard to predict so methods to measure the development of crystals inside real foods are useful. In this study, we show that as lactose crystallizes in a gelatin gel the ultrasonic attenuation--capacity to absorb sound--increases and can be related to the amount of crystals present. Ultrasound is easier to apply in real food processing than the existing methodologies.

Effects of droplet crystallization and melting on the aroma release properties of a model oil-in-water emulsion.

Aroma compounds partition between the dispersed and the continuous phases in emulsions, and phase transitions in the lipid droplets profoundly affect the position of the equilibrium. In the present study, the release of ethyl butyrate, ethyl pentanoate, ethyl heptanoate, and ethyl octanoate from a series of sodium caseinate-stabilized, n-eicosane emulsions was investigated as a function of solid and liquid lipid droplet concentration. For all compounds, headspace volatile concentrations above the solid droplet emulsions were higher than those above the liquid droplet emulsions. The interaction with liquid droplets could be modeled in terms of volume-weighted bulk partition coefficients while the more nonpolar volatiles bound to the surface of solid lipid droplets. The amount of volatiles bound to solid surfaces increased with aqueous concentration up to a critical point and then rapidly increased. The critical point corresponds to the dissolution of the solid lipid in a phase of adsorbed volatile. The binding of volatiles to both solid and liquid eicosane droplets is reversible.

Chemical and sensory analysis of strawberry flavoured yogurt supplemented with an algae oil emulsion.

A yogurt mix (2 g fat and 17g solids/100 g) was supplemented with an algae oil emulsion to provide 500 mg omega-3 fatty acids per 272 g serving of yogurt white mass. The emulsion was added to the yogurt mix either before or after the homogenization step and prior to pasteurization. It was then flavoured with a strawberry fruit base and fermented and stored for up to three weeks. The oxidative deterioration of the products was determined by hydroperoxide measurements and by trained and consumer sensory evaluations. The hydroperoxide content of the supplemented yogurts increased over the storage treatment and was unaffected by the stage of addition. The trained panel could distinguish a stronger fishy flavour in both of the supplemented yogurts after 22 days storage, but the consumer panel rated both control and supplemented samples similarly, as 'moderately liked'.