Umami and Satiety

Abstract

This chapter evaluates evidence from human studies that umami taste may enhance satiety. The author elaborates on the idea that humans evolved umami taste to detect and regulate protein intake, providing wider evidence that protein intake is more tightly regulated than other macronutrients and discussing specific evidence that protein is the most satiating. Three strands of evidence that suggest umami may have a role in satiety are evaluated. (1) Evidence from key studies that tested acute effects of manipulated umami taste on satiety in adult volunteers suggests that umami may enhance satiety, especially when coupled with protein intake. (2) A review of studies exploring the role of umami in infant feeding suggests that augmenting umami taste in bottle-fed babies leads to slower growth, implying that the presence of umami taste leads to greater satiety. (3) Evidence from studies exploring responses to umami in relation to protein deprivation suggests that sensitivity to umami varies depending on both acute and habitual protein need state, consistent with a regulatory role for umami involving satiety. This chapter draws these strands of evidence together to suggest two possible models of umami-induced satiety while noting limitations in the data that warrant further investigation.

Umami is today recognized as one of the five basic tastes, complementing the more widely known salty, sweet, sour, and bitter tastes. In all five cases, the taste percept (i.e., the conscious experience of the taste quality) arises from oral detection of nonvolatile chemicals dissolved in saliva that bind to specific taste receptors on the tongue. An important feature of all five basic tastes is that their qualities are unique: that is, it is not possible to generate a different taste quality by mixing other tastes (perceptual independence). In the case of umami, the primary receptor is known to be a heterodimer of two proteins, TAS1R1 + TAS1R3, that has a unique role in taste sensing (Damak et al., 2003; Li et al., 2002), and recent studies suggest strong positive selective pressure for the genes for both TAS1R1 and TAS1R3 in human evolution (Valente et al., 2018). These same receptors are found in other parts of the digestive tract, but only those in the mouth give rise to the perceptual experience of umami taste. A key question, then, is why we have evolved the ability to detect umami? To understand this, we first need to consider the potential evolutionary advantage of umami sensing.

Evolution shapes physiology to adapt to specific challenges for survival. For omnivores, including humans (as cogently argued as the “omnivore’s paradox” in (Rozin, 1976)), the key problem is how to identify which of the many potential foods that occur in nature are safe and nutritious. On that basis, it has long been argued that our ability to taste certain compounds in potential foods evolved to meet that challenge, with the mouth the gatekeeper that directs decisions on whether a potential food is ingested or rejected (Tanaka et al., 2007; Kim et al., 2004), interpreting taste as the “nutritional sense” (Bartoshuk, 2018). Accordingly, our ability to sense bitterness ensures we are cautious when ingesting substances that could be harmful (Shi et al., 2003; Breslin, 2013), and there is clear evidence of how diet shaped the evolution of bitter taste perception (Li and Zhang, 2014). Similarly, sweetness may signal a safe source of energy (Beauchamp, 2016), and most recently sourness has been proposed to help identify key vitamins in states of vitamin deficiency (Teng et al., 2019). From this evolutionary perspective, what may be the reason that humans can detect umami taste? The answer, based on the nature of the foods that contain the compounds that generate umami taste, is that umami may have evolved to help us sense sources of protein, an idea dating back to the original discovery of umami taste (Ikeda, 1908) and then developed by other researchers (e.g., Naim et al., 1991; Kurihara, 2009; Geraedts et al., 2011). But our ability to detect umami compounds in the mouth and later in the gut could have a function beyond the simple sensing of the presence of protein. Sensors in the mouth and gut are also key components of appetite control, and recent evidence suggests that umami taste not only allows us to sense the likely presence of protein but also helps the body regulate protein intake. This chapter evaluates current state of play for research that has tested the role of umami in satiety.

5.1 Why We Taste Umami: Lessons from the Diet

The original idea of umami relating to protein sensing was based on consideration of where in the diet we find the chemicals that activate umami taste receptors. Umami taste can be traced back to the initial isolation of glutamic acid from seaweed and the subsequent discovery in 1908 that it, and particularly its sodium salt, monosodium glutamate (MSG), conveys a unique “flavor.” MSG is now known to be the main molecule that stimulates human umami taste receptors. Subsequent research found that the umami taste generated by MSG was greatly potentiated by the presence of ribonucleotides, most specifically by disodium 5′-inosine monophosphate (IMP) and disodium 5′-guanosine monophosphate (GMP) (Kuninaka, 1981; Fuke & Ueda, 1996). Notably, neither IMP nor GMP elicits a strong umami taste alone, but both act synergistically with MSG to elicit a stronger umami taste than predicted by the sum of the individual components (Yamaguchi & Ninomiya, 2000; Fuke & Ueda, 1996). To understand why we have evolved the ability to taste umami, we need to consider where in our food supply we find MSG, IMP, and GMP: in a wide range of foods that we classify as savory. Indeed, it was Ikeda’s (1908) observation that these foods have a distinct “savory” taste that led to his use of the word umami to describe this.

The most concentrated levels of free glutamate are found in a wide variety of foods, including some vegetables (e.g., mushrooms, potatoes, cabbage, corn, onion, spinach), fruits (e.g., tomatoes, avocado), some fish and seafood (e.g., prawn, clam, crab, oyster), seaweed, egg yolks, ripened cheeses, human breast milk, and fermented soya beans and in lower concentrations in commonly consumed meats (e.g., beef, pork, chicken) (Kurihara, 2009; Ninomiya, 1998; Drake et al., 2007; Fuke & Shimizu, 1993; Hajeb & Jinap, 2015; Ghirri & Bignetti, 2012). From a nutritional standpoint, all of these provide valuable nutrients, but at first sight there appears to be no obvious strong nutritional link across these foods that could have driven the evolution of umami as a preferred taste.

However, ribonucleotides are distributed rather differently. IMP in particular is a constituent of most of the meat sources that provide most of the protein humans consume (beef, lamb, pork, chicken, fish, and goat). Moreover, although the protein found in those foods has a relatively weak taste profile in its natural state, once hydrolyzed, the protein has a characteristic umami taste carried predominantly by glutamate and ribonucleotides. Notably, the processes of aging, heating, and curing these foods releases both free MSG and IMP, greatly enhancing the level of umami taste (Sasaki et al., 2007; Rotola-Pukkila et al., 2015; Maga, 1994). This has led to important speculation that our penchant for umami taste evolved at a time when humans were switching to cooking as the primary process to prepare food for ingestion (Wrangham, 2017; Valente et al., 2018). Thus, in the form that we consume these foods, umami may provide an oral signal of the likely presence of protein, an idea first noted by Ikeda (1908).

It is therefore possible to develop a hypothesis that we evolved umami taste as a way of detecting protein in cooked foods. However, unlike other basic tastes, umami is strongly influenced by other flavor elements, making the sensing of umami more complex than sensing MSG and ribonucleotides alone (Mouritsen et al., 2017). This complexity of umami taste, with the strongest umami taste found only when sources of MSG and IMP/GMP are combined with other flavor elements (McCabe & Rolls, 2007), may have had a particularly profound impact on the development of cuisine worldwide, underlying the use of umami-rich seasonings (Mouritsen et al., 2017) and shaping how we combine ingredients to better enable sensing of protein (Ninomiya, 2015). The general idea that the overall savory character of a food is then related to protein content has been supported by some studies that asked participants to rate the savory character of foods and then explored how these ratings related to actual protein content (van Dongen et al., 2012; Martin et al., 2014). Other studies, however, suggest this relationship is relatively weak (Teo et al., 2018) or not significant (Buckley et al., 2018). Overall, both the content of free glutamate and umami-related nucleotides in foods and consumer evaluations of food products suggest umami taste may function, at least in part, to locate protein in the diet, although more work is needed to fully clarify this relationship.

5.2 Umami and the Regulation of Protein Intake

The argument developed in this chapter is that umami taste may go beyond a mechanism for simply sensing potential sources of protein in our diet to actually playing a key role in the regulation of protein intake. To understand this, the evidence that protein intake is regulated, followed by the effects of protein on satiety, is reviewed. Then the literature that has examined how the presence of umami taste modifies our experience of appetite is explored, focusing in particular on evidence that umami may enhance satiety when protein is ingested. Studies that tested liking for umami tastes as a function of protein need state provide further evidence for a role of umami in satiety, suggesting that umami is liked more when we are in short-term need of protein and, conversely, that liking is suppressed when we are protein replete.

5.2.1 Regulation of Protein Intake

Nutritionally, protein is one of the three main macronutrients and is an essential component of the human diet, needed for the formation and repair of muscle tissue and to provide essential amino acids to produce key neurotransmitters, hormones, enzymes, and receptors, among other key cellular functions. Both fat and carbohydrate can be stored as reserves if intake exceeds current energetic needs, but protein cannot be stored in the body. This inability may explain why the proportion of protein in the diet is remarkably similar in human cultures with very different dietary habits. An analysis of the percentage of energy consumed as protein across 175 countries (Fig. 5.1: based on data published in Ritchie & Roser, 2017) gave an average of 11.1%, with more than 50% of countries having protein intake between 10.5% and 12.5% of total energy intake and only 10% of countries having more than 12.6% of total energy from protein. Remarkably similar overall protein intake was seen in countries where more than 65% of protein intake comes from animal sources or 85% comes from plants. These data provide striking evidence of how tightly controlled human protein intake is and adds credence to the idea that oral sensing of protein in food, likely involving umami taste, is a critical part of appetite control.

Fig. 5.1

Average protein intake, as a percentage of total caloric intake, in 175 different countries and dependencies (numbers at tops of columns). (Data from Ritchie and Roser (2017))

The evidence that protein intake is regulated more tightly than are fat and protein has important implications for understanding our habitual food intake. According to the protein leverage hypothesis, the absolute need for a certain level of protein in our diet could drive overconsumption where available levels of protein are reduced (Simpson & Raubenheimer, 2005; Gosby et al., 2014). A full evaluation of that hypothesis is outside the scope of this chapter, but in the present context, the key message is that protein intake is tightly regulated, and for that to be true, the body has to have a mechanism for sensing and monitoring protein intake. Umami taste offers a potential sensory component that could achieve this.

5.2.2 Protein: The Most Satiating Macronutrient?

Satiety could potentially be explained by a broader sensing mechanism that monitors overall energy intake relative to energetic needs. At its simplest level, that proposition suggests that energy is the key controlled variable, and the prediction then follows that intake of the three main sources of energy (fat, carbohydrate, and protein) would have satiating effects, once matched for energy. However, humans generally experience greater satiety, that is, stronger suppression of appetite and subsequent less ingestion at the next meal, after consuming protein-rich foods than after energy-matched fat- or carbohydrate-rich foods (see Anderson & Moore, 2004; Dhillon et al., 2016; Johnstone, 2013; Morell & Fiszman, 2017; Paddon-Jones et al., 2008; Veldhorst et al., 2008 for reviews).

In this context, most short-term studies examining satiety use the preload design (Yeomans, 2018), where participants ingest a fixed amount of different versions of a test food, as the main manipulation, and their appetite is tested afterward to assess satiety. A large number of preload studies using different foods and drinks, different protein sources, and varying times between the preload and the ad libitum test meal have shown greater satiety after protein than after other energy sources (e.g., Anderson & Moore, 2004; Bertenshaw et al., 2008; Chungchunlam et al., 2012; Martens et al., 2013). Notably, however, a number of well-designed studies failed to find any greater satiety for protein than for other macronutrients (e.g., Geliebter, 1979; Raben et al., 2003) and in some cases even found no evidence that any macronutrient source generated short-term satiety relative to a low-energy control (e.g., de Graaf et al., 1992). Although small methodological differences (study power, satiety measures, etc.) could have influenced these outcomes, the variability in outcome itself could suggest something important about how protein is sensed and controlled. Crucially, to enable clear claims about protein, most protein-satiety studies take great care to disguise the presence of protein at a sensory level. In doing so, these studies may inadvertently remove the very signals the body uses to sense and control protein intake—including, critically, umami taste.

This idea that the sensory characteristics of protein-rich products at least partially explain the higher levels of satiety they generate was tested explicitly (Bertenshaw et al., 2013). Their earlier studies had confirmed both that a whey-protein-enhanced beverage was more satiating than an equicaloric carbohydrate-enhanced beverage (Bertenshaw et al., 2008) and that the effects of whey protein on satiety were dose dependent (Bertenshaw et al., 2009). However, the flavor characteristics of whey protein (which includes an umami component) made it impossible to fully disguise the addition of protein. Thus, they surmised that perhaps the greater effectiveness of protein than carbohydrate to induce satiety may have been due to the sensory characteristics cuing an expectation of protein, which in turn generated the stronger satiety response. To test this, they contrasted the effects of the same high-whey-protein preload with two alternative high-energy preloads: a carbohydrate preload adjusted to the same perceived thickness, with cream flavoring added to match sensory characteristics of the whey preload, and a second protein preload using a different form of whey protein that lacks these sensory characteristics (Bertenshaw et al., 2013). Notably, when sensory matched, the high-sensory whey and sensory-matched carbohydrate induced the same levels of satiety (evidenced by lower-energy intake at a test meal; Fig. 5.2), but the equicaloric low-sensory protein preload produced significantly less satiety. However, this study did not include any specific manipulation of umami taste. Thus, studies that examine whether satiety is greater in the context of umami taste can test the umami-satiety hypothesis.

Fig. 5.2

Recovery of hunger after consuming one of six versions of a naturally low-glutamate soup (Masic & Yeomans, 2013): lower-energy (126 kcal) soup (a), higher-energy soup (310 kcal) primarily with added protein (b), or a similar higher-energy soup (303 kcal) with added carbohydrate (c), either with (broken lines) or without (solid lines) added MSG

5.3 A Role for Umami in Protein-Based Satiety?

How might the possible effects of umami on appetite be tested? To date, a number of different approaches have been used, and although the picture is far from complete, several lines of evidence suggest that umami may have some role in generating postmeal satiety. This research is described in the following sections, grouped by the approach taken: human experimental preload-satiety studies with adult volunteers, insights from the studies of feeding by human infants, and consideration of how acute protein need state modifies responses to umami taste.

5.3.1 Umami-Enhanced Satiety: Human Experimental Studies

To date, 10 studies conducted with human adult volunteers, reporting 13 experiments and representing research from 6 different research groups, have tested the effects of added MSG or a combination of MSG with nucleotides including IMP on satiety (see Table 5.1). These studies provide the most direct test of the idea that umami enhances satiety and so warrant detailed evaluation. These studies are presented in chronological order since the findings changed as the design of studies became more sophisticated.

Table 5.1 Preload-satiety studies investigating effects of manipulated umami taste in human adult volunteers

The first study that directly assessed the effects of umami using a laboratory-based satiety test in human volunteers looked at the effects of a relatively high concentration of MSG (20% w/w) added to a minimal-energy beef consommé in three related experiments (Rogers & Blundell, 1990). There was no evidence that addition of MSG-enhanced satiety in these three experiments, with no effect on the experience of appetite over the ensuing 2 h (experiment 1) and no significant effect of 10% or 20% MSG on test meal intake (experiments 2 and 3). The only significant effect on any of the measures of appetite was for more rapid recovery of hunger after consuming the soup with 20% MSG followed 10 mins later by a meal (experiment 3). Thus, these experiments provide no evidence of any effects of MSG on satiety and, indeed, found instead limited evidence that MSG could stimulate appetite, although the levels of MSG used were notably high and outside the physiological range consumers would experience from normal foods.

The next study (Luscombe-Marsh et al., 2009) likewise failed to find any evidence that umami caused satiety. Here, appetite and intake at a test meal were contrasted between a water control and a high-protein meal, with four variations: unaltered, with added MSG (0.6%), with added MSG + IMP, or with the MSG/IMP version sham-fed (i.e., chewed but not swallowed). There were no significant differences in changes in appetite ratings across time between the three high-protein meals, all of which generated greater satiety (i.e., reduced rated appetite) than did the water control or the sham-fed preload. Contrary to the predictions of enhanced satiety, more was consumed after the meal with added MSG (but not added MSG + IMP) than after the unaltered meal. Thus, the conclusion from these first two studies would be that umami does not enhance satiety; rather, to the contrary, MSG might cause a small increase in appetite. However, in the next study (Carter et al., 2011), when a low-energy broth was consumed twice prior to a test meal, with umami manipulated by the addition of MSG or MSG + IMP, MSG slowed the recovery of hunger after the second soup preload. Although the umami manipulations had no significant effects on food intake, this study did suggest a possible, albeit small, effect of umami on satiety, and notably, this was seen at much lower levels of added MSG than in the earlier studies. The approach in the next study was different: Finlayson et al. (2012) tested whether consumption of a preload that was bland, sweet, or savory in flavor modified food selection and intake at a subsequent buffet meal. The relevance here was that MSG (level not reported) was used to modify the flavor of the savory condition. Overall, there was little effect of the sensory manipulations, but notably, intake of high-fat foods was significantly lower after the MSG-enhanced savoury than the sweet preload.

Building on the findings of reduced recovery of hunger after an umami-enriched soup (Carter et al., 2011), and the finding that carbohydrate could have the same effects on satiety as protein if the sensory characteristics were similar (Bertenshaw et al., 2013), Masic and Yeomans (2013) tested the effects of a realistic level of MSG added to a low-energy, low-glutamate soup, and the same soup with either added protein or added carbohydrate. As in an earlier study (Carter et al., 2011), there was evidence that MSG could slow the recovery of hunger after ingestion, but only after consuming the soup with added protein (Fig. 5.2), in contrast to the results of Carter et al. (2011), who only manipulated MSG content of a low-energy (15 kcal) broth. The implication is that the MSG signal appears to be most effective when it is experienced in the context of actual protein ingestion. This might suggest that umami acts as a signal of the likely presence of protein, thus aiding more efficient processing of ingested protein, an idea further supported by a follow-up study that examined test meal intake after low- and higher-energy soup preloads with and without added MSG (Masic & Yeomans, 2014a). In that study, addition of protein to a low-energy, low-glutamate soup reduced test meal intake more than did an equicaloric carbohydrate preload. To further characterize satiety, the satiating efficiency of the two macronutrient manipulations, with and without added MSG, was calculated (Kissileff, 1984; Bellisle & Blundell, 2013). Satiating efficiency is calculated by determining what percentage of the energy difference between a treatment and control preload is compensated for by reduced intake at subsequent meals. Hypothetically, imagine someone consuming a low-energy (e.g., 100 kcal) preload on 1 day and a high-energy (e.g., 400 kcal) on a second day. If they then consumed 300 kcal less at lunch after the high-energy than after the low-energy preload, the satiating efficiency of the high-energy preload can be said to be 100%. In practice, perfect compensation in preload-satiety studies is very rare (Almiron-Roig et al., 2013; Chambers et al., 2015). Notably, when these values were calculated (Masic & Yeomans, 2014a), compensation for the added energy was significantly greater after the protein-rich soup with added MSG (62%) than in the equivalent carbohydrate condition (24%). This study therefore added to the evidence that umami can enhance satiety but that it does so particularly in the context of protein intake.

So far all of the umami-satiety studies in our laboratory had relied solely on the manipulation of MSG content only. However, earlier it was noted that cuisine favors combinations of ingredients that result in the presence of MSG + IMP, and the first study to suggest there may be a role of umami in satiety included an MSG + IMP manipulation (Carter et al., 2011). Therefore, Masic and Yeomans (2014b) examined the effects of MSG + IMP on protein-induced satiety by contrasting appetite and intake after consuming a low-glutamate soup with added energy (principally as whey protein) and a combination of MSG + IMP. In this case, there was evidence both for the expected effect of protein on satiety, with less consumed after the high-energy protein soup than after the low-energy soup, and an overall effect of umami, with less consumed when umami was enhanced by the addition of MSG + IMP in both the low-energy and protein-enriched preloads (Fig. 5.3). They again calculated satiating efficiency and found that added protein plus MSG + IMP produced significantly greater compensation (70%) than did added protein alone (44%).

Fig. 5.3

(a) Intake at lunch after consuming a low-energy (12 kcal) control soup (left) or a higher-energy (310 kcal) soup, with energy mainly from protein (right), either with (white columns) or without (gray columns) added MSG/IMP (Masic & Yeomans, 2014b). (b) Percentage of the preload energy compensated for by reduced lunch intake after the high- relative to the low-protein preload in the control and MSG/IMP conditions. * p < 0.05

Since the series of studies in our laboratory, we are aware of a further three publications that have further investigated the possible effects of umami on satiety using the preload-satiety model. In the first of these (Imada et al., 2014), participants consumed a low-energy (8 kcal) chicken broth either alone (control), with added MSG (0.5%), or with added MSG + IMP prior to a multi-item buffet meal. Overall energy intake at the meal was lower after consuming the soup with added MSG, but not with added MSG + IMP, than in the control condition, due primarily to the reduced selection and intake of sweet and high-fat snacks. Thus, this study did suggest some direct satiating effects of a low concentration of MSG but found this in the context of a minimal-energy broth. A subsequent study by the same group tried to extend the findings from the study by Carter et al. (2011) using the two-soup intervention model in overweight individuals, with a soup naturally low in glutamate (Miyaki et al., 2016). Again, addition of 0.5% MSG to this soup reduced subsequent food intake at a test meal and also reduced rated hunger between soup ingestion and the start of the meal.

In the most recent two experiments to examine the effects of umami on satiety, Anderson et al. (2018) attempted to replicate and extend the earlier findings of Masic and Yeomans (2014a). The key differences were the use of a slightly larger preload (500 mL instead of 300 mL), a longer gap between preload and test meal (120 min instead of 45), and measurement of satiety-related hormones as additional measures of satiety. In their first experiment, the high-protein soup with MSG sustained fullness for a longer period than did all other treatments, but MSG had no significant effects on intake at the test meal. In their second experiment, subjective appetite was significantly lower after the protein soup with added MSG than after all other conditions, and intake at the next meal was significantly lower for the protein + MSG but not protein-alone conditions than in the lower-energy and water control conditions. Both experiments thus found some evidence to support a role for umami in satiety, replicating earlier studies. The second experiment also found some changes in physiological markers that further supported these behavioral satiety findings: the protein + MSG preload but not the protein-alone preload resulted in increased levels of insulin and C-peptide (an early-stage marker for insulin synthesis) and also lowered blood glucose.

At face value, of the 13 experiments examining effects of MSG on satiety, 9 reported at least some evidence supporting a role for umami in satiety, and of the 4 that did not, 3 were from the first study to examine this question. Closer inspection of the designs of these studies gives clues to why some did not find evidence of umami-enhanced satiety. The first key design issue appears to be the dose of MSG used to manipulate umami: the dose used in the earliest three experiments (Rogers & Blundell, 1990) was considerably higher (10–20%) than in the more recent experiments finding evidence of satiety (typically 0.5–1.0%). The levels of MSG in those early studies were considerably higher than those found in normal food products, and it is notable that studies using more realistic levels of MSG found evidence of satiety enhancement. This further implies that satiety is not enhanced as a linear function of MSG concentration. The second design issue may be the size of the preload: all of the successful studies used relatively low-energy preloads (maximum of ~300 kcal), nearly always as a soup manipulation. In contrast, the study by Luscombe-Marsh et al. (2009), which found no significant effect, used a much larger preload, comprising soup plus one or two savory rolls, with the portion size adjusted to deliver 20% of the estimated energy requirements for each participant. The effects of umami may have been masked by the much stronger satiety from the larger preload. This in itself raises important questions about the nature of the satiety signal umami generates, discussed in depth further below. A final issue worth noting is that many studies did not fully control the level of umami taste in the control conditions of these studies, which hinders interpretation. For example, if the level of umami in the control was already sufficient to signal the possible presence of protein, then additional MSG might not add any new information that could affect appetite.

5.3.2 Glutamate and Satiety in Human Infants

Although it is less possible to conduct direct studies of the role of umami in satiety in infants and children, further insights into the potential impacts of umami on satiety have been provided by studies exploring differences in weight gain associated with breast-feeding and bottle-feeding practices in humans. In relation to umami taste, one of the most surprising observations is that human breast milk has as much as 19 times the free glutamate content of cow’s milk (Van Sadelhoff et al., 2018). That observation allowed researchers to consider what role umami might have in control of infant feeding (Mennella et al., 2011; Ventura et al., 2012, 2015), in the context of the well-established finding that human infants who are bottle-fed a diet of standard cow milk formula (CMF) typically gain more weight than do babies who are breast-fed (reviewed by Appleton et al., 2018).

In addition to the difference in free glutamate, there is a notable difference in overall protein content of the two main milk sources fed to human infants: CMF has more protein overall than does human breast milk, and consequently, formula-fed infants have much higher protein intake (55–80% more, adjusted for body weight; Alexy et al., 1999) than do breast-fed infants. Because protein is more satiating than other macronutrients, we might expect that CMF would be more satiating than is breast milk and that the faster weight gain seen in CMF-fed infants cannot be attributed to a lack of satiety from feeding. However, while CMF has higher overall protein content, it has lower overall levels of free amino acids, most notably glutamate. When the growth rate of human infants was contrasted between standard CMF and a formula with protein that had been extensively hydrolyzed (EHF), which consequently had a much higher free amino acid content (Mennella et al., 2011), those fed on EHF had more normative weight gain than those fed standard CMF, and their growth rates were in line with predictions for breast-fed infants (for further details, see Chap. 5 in this volume). Hydrolyzing the protein in the formula had a particularly large effect on levels of free glutamate, with 106.5 mg/100 mL in the hydrolyzed formula compared to just 1.8 mg/100 mL in CMF. These short-term effects were subsequently verified in a larger randomized controlled trial to more fully determine the direct impact of standard CMF and EHF on infant growth (Mennella et al., 2018). Replicating the initial finding, that study found evidence of sustained slower growth in body weight for infants fed with EHF, which persisted across the first year of life, because infants ingested more CMF than EHF, driven at least partly by a reduced intake of EHF, which was more satiating.

More remarkably, the same group extended this work to contrast whether the presence of free amino acids affected actual infant feeding (Ventura et al., 2012). Infants attended the laboratory on three occasions to consume two meals with each of three formula diets: CMF, EHF, and CMF with added MSG. Infants fed both with EHF and MSG-enhanced formula consumed less at the first meal than they did with CMF, and a longer interval elapsed before they signaled readiness for their second meal, and they did not compensate by consuming more at that second meal. Thus, both reduced immediate intake and delayed demand for the second meal are consistent with an effect of umami taste on satiety in human infants, adding further weight to the idea that umami has a clear role in regulating human appetite. A subsequent study explored the feeding style of infants fed standard and CMF with added glutamate (Ventura et al., 2015). Infants consumed less of the glutamate-enhanced CMF and tended to feed for shorter times. Although detailed analysis of feeding behavior (from video analysis) found few differences between diets, there was evidence that the distinct end-of-meal behaviors seen later in meals became evident sooner with the glutamate-enhanced formula. The reduction in intake, shorter feeding duration, and earlier switch to end-of-meal behaviors might be consistent with increased satiety from the glutamate-enhanced CMF.

5.3.3 Protein Need State, Satiety, and Liking for Umami Taste

The evidence to date suggests a likely role for umami taste in satiety in humans but is far from conclusive. The inconsistent results among studies that explicitly explored umami taste and satiety in particular (see Table 5.1) suggest that any such role of umami may be specific to very defined conditions, and so whether this realistically contributes to satiety in real life may be questionable. However, if umami acts as a signal for the likely presence of protein, then biologically it would make sense for sensitivity to that umami signal to also vary depending on individual protein need state. If so, the variability in study outcomes could relate to differences in acute and habitual protein intake across study designs and individual participants.

As discussed earlier, regulation of protein intake appears much more tightly controlled than fat and carbohydrate, and this may in turn mean that sensitivity to umami taste, including its effects on satiety, may also vary with both habitual protein intake and acute protein need. Thus, the final set of studies explored in this chapter tested how responses to umami taste and/or wider savory evaluations vary in relation to both acute and longer-term protein intake. The focus again is on human studies because of the suggested evolution of a specific role for umami coinciding with when humans first started to cook food (Hartley et al., 2019; Valente et al., 2018).

In support of responses to MSG varying with protein need state, a number of older studies in humans reported stronger preferences for foods with higher MSG content in both adults (Murphy, 1987) and children (Vazquez et al., 1982) when in protein-deficient states or with poor nutritional status compared to well-nourished controls. Some studies have also reported that liking for savory flavors varied with acute protein status. In this context, one study in particular suggested that participants’ ability to respond to experimentally manipulated acute protein deficit (by provision of low-protein breakfasts) has to be learned: participants who were acutely protein deprived came to prefer a dessert flavor that was previously paired with the delivery of more protein relative to participants tested and trained in the absence of acute protein lack (Gibson et al., 1995). That study did not specifically manipulate umami taste, but the demonstration that acute protein need heightened sensitivity to sensory characteristics of a protein-rich product fits the wider idea that umami promotes protein choice in the protein-deprived state. Specific evidence that acute protein deprivation modifies the response to MSG came from a study that examined responses to umami, salty, and sweet tastes in participants in relation to acute protein need (by manipulating earlier protein intake) and habitual protein intake (Masic & Yeomans, 2017). In that study, acute protein deprivation increased liking regardless of MSG concentration (Fig. 5.4a), and crucially, this effect depended on habitual protein intake (Fig. 5.4b). Here, liking for the strongest (1%) MSG soup decreased as a function of habitual protein intake after consuming a protein-rich breakfast, but not after the low-protein or baseline (habitual) breakfast. Because the high-protein breakfast would have been predicted to generate higher satiety, the implication is that expression of liking for umami is suppressed by protein-induced satiety. This suggests that umami’s role in appetite control is fine-tuned to control protein intake, with increased preference for umami driving protein intake in the context of either habitual low-protein intake or acute protein need. Moreover, if the body monitors protein intake, at least in part, through experience of umami taste, then increased exposure to umami taste may in turn result in reduced preference for umami, an idea supported by a recent study where repeated exposure to umami resulted in a generalized decline in liking for umami taste (Noel et al., 2018).

Fig. 5.4

How acute and habitual protein intake modified hedonic responses to umami taste (Masic & Yeomans, 2017). (a) Overall changes in liking from baseline for a low-glutamate soup with 0% (white columns), 0.6% (gray columns), or 1.0% (dotted columns) added MSG evaluated 2 h after consumption of a low- or high-protein breakfast. * p < 0.05. (b) Absolute liking for soup with 1% added MSG as a function of habitual protein intake on the baseline (normal breakfast) day (black line) and after days when participants consumed a specific low-protein (dark gray line) or high-protein (light gray line) breakfast

5.4 Possible Models for Umami-Enhanced Satiety in Humans

If umami taste does have a role in short-term regulation of food intake, particularly associated with satiety in the context of regulation of protein intake, a key question is how this might work. The original idea that umami taste evolved to allow us to sense protein in our diet (Ikeda, 1908) would not necessarily suggest that umami should play a role in satiety: umami might direct food choice to protein-rich sources, but the control of intake could be achieved through other satiety processes. However, as discussed in detail above, intake of protein appears to be much more tightly regulated than intake of other macronutrients. Thus, the effects of umami on satiety might suggest umami taste has a role beyond food choice to help regulate protein intake. How might that regulation be achieved?

Protein has been shown to be more satiating than other macronutrients, but this is most evident where the sensory characteristics of the ingested food include cues that predict the likely presence of protein (Bertenshaw et al., 2013). Umami taste may be one of the most reliable predictors of protein in the context of the cooked foods that most humans consume. This is supported by the positive findings of most studies that examined the effects of umami taste on satiety and included a protein-rich manipulation (Table 5.1). Although these effects were relatively subtle, these studies found at least some evidence (in terms of altered intake and/or enhanced subjective satiety) that the combination of MSG + protein was more satiating than would be predicted from the sum of the separate effects of protein and MSG alone. Although how umami taste magnifies the satiating effects of protein is not clear, two possible models are presented here that might be explored in future research.

The first model builds on increasing evidence that the orosensory experience of foods generates consumer expectations about the level of satiety they will experience postingestion (Chambers et al., 2015). There is now a wide variety of evidence that both sensory and cognitive characteristics of ingested products can modify these satiety expectations and that these expectations then interact with actual nutrient consumption to generate satiety. Beyond umami, at a sensory level, two examples are (1) where perceived thickness and creaminess of a beverage generated predictable expectations about satiety (McCrickerd et al., 2012), and these same expectations then modified the degree of satiety generated by a disguised nutrient load (Yeomans & Chambers, 2011), and (2) where differences in oil-droplet size likewise altered satiety expectations (Lett et al., 2015) and actual satiety (Lett et al., 2016). To date, studies have not explicitly tested how manipulation of umami modifies satiety expectations, but the prediction would be that the increasing savory characteristics of umami-rich foods would generate stronger satiety expectations. However, if umami simply generated satiety expectations, then any nutrient-rich food that has some level of umami taste should be more satiating than the same nutrients without umami taste. What is more complicated with umami is that the effects seem to be protein specific, which argues against a generalized expected satiety model. However, if the expectations generated by umami were confirmed by protein sensing in the gut postingestion, this model offers a plausible account. It is now well established that the TAS1R1 and TAS1R3 receptor units that underlie umami taste when glutamate is detected in the mouth are also present throughout the gastrointestinal tract (Kondoh et al., 2009; San Gabriel & Uneyama, 2013), although as with other postoral taste receptors, activation of these receptors does not result in our “tasting” food in our gut. Instead, it is widely believed that these gut-based receptors monitor nutrient levels (e.g., Shirazi-Beechey et al., 2014; Raka et al., 2019), perhaps fine-tuning digestive processes depending on the balance of ingested nutrients.

If the impact of umami on satiety was mediated through expectations, then we would predict that umami taste would enhance behavioral measures of satiety expectations and have measurable effects in brain areas associated with these expectations. To date, there has been no detailed evaluation of effects of umami on expected satiety, and likewise, the neural representation of expected satiety has not been elucidated. However, a number of studies with human volunteers have looked at neural responses to umami taste, which allow some inference of the plausibility that umami enhances protein-induced satiety by generating expected satiety. Early studies established that umami taste is represented in the primary and secondary gustatory cortex, incorporating the anterior insula, frontal operculum, and the orbitofrontal cortex (OFC: e.g., Schoenfeld et al., 2004; McCabe & Rolls, 2007; Singh et al., 2015). The OFC has also been shown to have a role in sensory-mediated satiety (O’Doherty et al., 2000). More importantly, the OFC has been shown to have a critical role in predicting reward value (e.g., O’Doherty, 2007; Hare et al., 2008; McDannald et al., 2012), and because satiety expectations are neural predictions of the potential impact of ingestion on satiety, it would be predicted (but as yet not tested) that OFC neurons would play a critical role in satiety expectations. Thus, although no studies to date have formally tested an expected satiety model of umami-enhanced satiety, current knowledge of encoding of umami taste and broader prediction are both consistent with an expectation-based model.

An alternative to the expectation-mediated effects of umami on satiety would be a more direct effect of umami on the physiological systems known to be involved in satiety. Of principal interest are the specific hormones released in the gut in response to nutrient ingestion (see Chaudhri et al., 2008; for reviews Zanchi et al., 2017), which include cholecystokinin (CCK), pancreatic polypeptide, glucagon-like peptide, and polypeptide-YY. CCK is of particular interest, since studies in vivo and in nonhuman animals have provided some evidence that stimulation of the intestinal TAS1R1 + TAS1R3 complex can stimulate CCK release (Daly et al., 2013; Tian et al., 2019). If this is the cause of the enhanced satiety seen in some human appetite studies, then it would be predicted that circulating CCK levels would be raised after consuming umami-tasting foods, particularly in the context of protein consumption. In this context, it is notable that one study in human volunteers found evidence of reduced hunger, but not increased CCK, following intragastric infusion of umami tastants (Van Avesaat et al., 2015). A more recent study examined effects of umami taste and ingestion on glucagon-like peptide and again found no evidence of any change in hormone levels (Anderson et al., 2018). Thus, the only two studies that examined changes in satiety hormones after umami did not find evidence in humans to support the suggested effects of umami on CCK release reported elsewhere. Overall, while a direct effect of umami taste on gut-based satiety signaling cannot be discounted, to date there is no evidence for this.

Finally, although the expectation-based and gut signaling-based models of umami-based satiety are discussed here as separate potential models, in practice control of appetite is multifaceted and integrated. Notably, there is increasing evidence of top-down control of gut-based satiety signaling (e.g., Crooks et al., 2021) and, in particular, that sensory characteristics of products that generate satiety expectations may also enhance gut-based release of CCK (Yeomans et al., 2016).

5.5 Uncertainties and Future Directions

This chapter reviews evidence that umami may enhance satiety in humans and identifies important shortcomings in our current knowledge. The wider idea that umami is an alimentary taste (Keast et al., 2021) that may have evolved in humans in the context of our use of fire to modify protein prior to ingestion complicates the investigation of umami-induced satiety since it implies that the role of umami in humans may differ from that in other species. Usually findings in humans are verified by more detailed physiologically based studies in nonhuman animals, but in this case such studies may not be appropriate tests of the role of umami in satiety in humans. Instead, more detailed human studies are needed to meet specific shortcomings in the literature, including (but not limited to) the effects of umami taste on satiety expectations, more detailed evaluations of neural responses to umami taste in relation to satiety, and more highly powered and sustained studies of the effects of umami-enriched products on satiety. A further complication is that many products already have appropriate levels of satiety signals that match the relevant nutrient content; adding umami taste in that context could lead to a mismatch between the sensed taste and nutrient content, an issue that has been noted in the context of sweet taste and carbohydrate-based calories (Veldhuizen et al., 2017).

5.6 Conclusions

This chapter evaluates the evidence that umami taste may impact satiety in humans. The evidence that umami predicts protein, as originally suggested by Ikeda (1908), does not look convincing until the cooked state of the food humans ingest is taken into consideration. In the context of cooked food, the idea that umami predicts protein is plausible and also fits with evidence of how the system of sensing umami taste evolved. Although not all studies reviewed found evidence of enhanced satiety, the balance of evidence, particularly where the level of umami enhancement is physiologically realistic, does suggest a subtle effect of umami on satiety. In particular, the evidence suggests that umami increases the degree to which protein is satiating, which may explain why protein has been found in general to be more satiating than other macronutrients. Future research is needed, however, to understand the mechanisms through which umami enhances satiety. This chapter offers the suggestion to focus on two models, one based on prediction error and expectations and the other on physiological satiety cues.