**Branched-Chain Fatty Acids—An Underexplored Class of Dairy-Derived Fatty Acids**

**Victoria M. Taormina 1,**†**, Allison L. Unger 2,**†**, Morgan R. Schiksnis 2, Moises Torres-Gonzalez 3,\* and Jana Kraft 2,4**


Received: 21 August 2020; Accepted: 17 September 2020; Published: 20 September 2020

**Abstract:** Dairy fat and its fatty acids (FAs) have been shown to possess pro-health properties that can support health maintenance and disease prevention. In particular, branched-chain FAs (BCFAs), comprising approximately 2% of dairy fat, have recently been proposed as bioactive molecules contributing to the positive health effects associated with the consumption of full-fat dairy products. This narrative review evaluates human trials assessing the relationship between BCFAs and metabolic risk factors, while potential underlying biological mechanisms of BCFAs are explored through discussion of studies in animals and cell lines. In addition, this review details the biosynthetic pathway of BCFAs as well as the content and composition of BCFAs in common retail dairy products. Research performed with in vitro models demonstrates the potent, structure-specific properties of BCFAs to protect against inflammation, cancers, and metabolic disorders. Yet, human trials assessing the effect of BCFAs on disease risk are surprisingly scarce, and to our knowledge, no research has investigated the specific role of dietary BCFAs. Thus, our review highlights the critical need for scientific inquiry regarding dairy-derived BCFAs, and the influence of this overlooked FA class on human health.

**Keywords:** *anteiso*; branched-chain amino acids; cancer; diabetes; inflammation; *iso*; metabolic diseases; milk; phytanic acid

#### **1. Introduction**

Milk and dairy products have been a staple of the human diet for thousands of years [1] and represent one of the most important agricultural commodities in regard to human nutrition. Dairy products have been instituted in most dietary guidelines around the world as they provide a large variety of essential nutrients and several key shortfall nutrients. Moreover, milk and dairy products hold a unique position among all foods as they are considered the largest single source of natural bioactive components [2]. In particular, dairy fat is the most complex and diverse dietary fat source in nature comprised of an impressive fatty acid (FA) repertoire (>400 different FAs and FA derivatives [3]) that accounts for the myriad of nutritional, organoleptic, and technological characteristics of milk and dairy products. Specifically, dairy fat contains a unique variety of bioactive FAs that are synthesized by rumen microbes (i.e., bacteria and protozoa) and the mammary gland. These FAs consist of short- and medium-chain FAs (4 to 13 carbons), positional and geometric isomers of octadecanoate

(18:1), conjugated linoleic acids, odd-chain FAs (15:0 and 17:0), and branched-chain FAs (BCFAs). Of the repertoire of FAs in dairy fat, about 14% of them are these unique dairy-derived FAs [4] and several are known to impact normal mammalian physiology by functioning as bioactive molecules, exerting beneficial properties that support health and well-being [5–7]. To date, most research efforts in the arena of dairy FAs have been centered on the biological effects of conjugated linoleic acids, specifically rumenic acid (18:2 *c*9,*t*11), which was originally driven by the discovery of its demonstrated anticarcinogenic activity [8]. However, far less work has focused on other bioactive dairy-derived FAs.

BCFAs are an emerging group of bioactive FAs sparking growing research interest within the scientific community due to their biological effects and potential pro-health benefits. Because BCFAs are principally derived from rumen bacteria, milk and dairy products pose unique dietary sources of BCFAs. The objectives of this paper are to review and summarize the current knowledge on BCFAs and specifically to evaluate the possible role of dairy-derived BCFAs in human health.

#### **2. Structure and Origin of BCFAs in Ruminants**

BCFAs are commonly saturated FAs substituted with one (mono-) or more (di-/poly-) methyl branch(es) on the carbon chain. Typically, BCFAs possess either an *iso* structure where the FA has the branch point on the penultimate carbon atom (i.e., one from the end) or an *anteiso* structure where the branch point is located on the antepenultimate carbon atom (i.e., two from the end; Figure 1). More than 50 BCFAs have been identified in ruminant-derived fats [9,10] but *iso*- and *anteiso*-mono-methyl BCFAs with chain lengths from 14 to 17 carbon atoms are quantitatively the most abundant BCFAs in milk fat [4,11–16].

**Figure 1.** Structural differences between straight chain fatty acids and *iso-*, *anteiso-*, and multimethyl branched-chain fatty acids.

In ruminants, monomethyl BCFAs are synthesized by the microorganisms that reside within the rumen, specifically bacteria and protozoa [17], which utilize dietary branched-chain amino acids (BCAAs), i.e., valine, leucine, and isoleucine, to form BCFAs (Figure 2) [18]. Through a common biosynthetic pathway, BCAAs are first transformed into branched-chain α-ketoacids through the removal of the amino group by a BCAA transferase enzyme [19]. These α-ketoacid products, α-ketoisovalerate, α-keto-β-methylvalerate, and α-ketoisocaproate, are subsequently decarboxylated by branched-chain-α-ketoacid dehydrogenase producing the respective branched short-chain carboxylic acids isobutyral-CoA, isovaleryl-CoA, and 2-methylbutyral-CoA [20]. Finally, branched short-chain carboxylic acids are elongated by BCFA synthetase, with malonyl-CoA as the chain extender, to form

*iso-* and *anteiso-*BCFAs [18]. The products of the BCFA biosynthetic pathway are *iso*-14:0 and *iso*-16:0, derived from valine, *iso*-15:0 and *iso*-17:0 from leucine, and *anteiso*-15:0 and *anteiso*-17:0 from isoleucine.

**Figure 2.** Biosynthetic pathway of branched-chain fatty acids from branched-chain amino acids. BCAT: branched-chain amino acid transferase (BCAT) enzyme. BKD: branched-chain-α-ketoacid dehydrogenase.

Phytanic acid (3,7,11,15-tetramethylhexadecanoic acid) and its metabolite pristanic acid (2,6,10,14-tetramethylpentadecanoic acid), two multimethyl BCFAs, are also predominantly synthesized in the rumen (Figure 3) [21,22]. Phytanic acid is derived from the phytol component of chlorophyll found in forages [21]. While mammals are unable to cleave the ring structure from the phytol moiety of chlorophyll, bacteria release phytol in the rumen [23]. The primary synthesis pathway of phytanic acid in ruminants is initiated by the biohydrogenation of phytol to produce dihydrophytol [21,24]. Dihydrophytol is subsequently converted to phytanal before it is synthesized into phytanic acid [21,24]. Notably, a potential secondary synthesis pathway has been described in non-ruminant tissues and marine bacteria [22,24,25] and may be applicable to ruminant animals as well, although there is only limited evidence [26]. In this case, phytol is first converted to phytenal by an alcohol dehydrogenase [22]. Fatty aldehyde dehydrogenase is used to transform phytenal to phytenic acid before it is modified into phytenoyl-CoA by an acyl-CoA synthetase. Subsequently, phytenoyl-CoA is converted by an enoyl-CoA reductase to form phytanoyl-CoA [24,25]. The final synthesis step of phytanoyl-CoA to phytanic acid is thought to occur through α-oxidation [25] due to the enzymatic activity of thioesterase, a protein involved in α-oxidation [24]. In order to form pristanic acid, phytanic acid is activated by phytanoyl-CoA synthetase to create phytanoyl-CoA, which can then undergo α-oxidation to produce pristanic acid.

**Figure 3.** Synthesis of phytanic and pristanic acid via rumen microorganisms derived from chlorophyll within forages NAD: Nicotinamide adenine dinucleotide. NAD+: the oxidized form of NAD. NADH: the reduced form of NAD (protonated with a hydrogen).

BCFAs are key structural lipid constituents of the bacterial membranes [17] where they play an important regulatory role in membrane fluidity and permeability [27]. Importantly, bacterial membrane lipids make an important contribution to ruminant milk fat, as bacterial cells leaving the rumen pass to the duodenum where their membrane FAs are absorbed and subsequently incorporated into milk fat and other tissues [21].

#### **3. Occurrence and Metabolism of BCFAs in Humans**

BCFAs are common constituents of microbial lipids present in abundant quantities but have also been found in many other organisms (e.g., *C. elegans*, [28]) including mammals [29–31], although in much lower amounts. In humans, BCFAs have been detected in various tissues and fluids including vernix caseosa, the biofilm covering the skin of the fetus [32,33], colostrum and mature breast milk [34–36], adipose tissue [37], and serum [38]. Early work by Nicolaides and co-authors [39–41] established that the meibomian and sebaceous glands of the human skin produce BCFAs secreted into meibum and sebum, respectively. Subsequent research in vitro and in vivo using mouse models confirmed the endogenous synthesis of BCFAs from their respective BCAA precursors (i.e., valine, leucine, isoleucine) as a result of the BCAA catabolic pathway [30,31,42,43]. BCAA degradation in humans is comparable to that in bacteria (Figure 2). Despite the endogenous synthesis, dietary intake of BCFAs is presumably the principal source of BCFAs in the human body.

The intake of BCFAs depends on the type of food and the fat content of the food consumed. While ruminant-derived foods (i.e., milk and meat and their respective products) are the chief source of BCFAs, fish and non-dairy fermented foods (e.g., kimchi, sauerkraut, miso, tempeh) may also supply BCFAs, although only at very small quantities. The mean daily intake of BCFAs in the United States has been estimated to range between 220 mg/day (for milk only [44]) and 500 mg/day (from dairy and beef products combined [11]). The contribution of total dairy products to the BCFA intake has been calculated at 317 mg/day [11]. There is no requirement for BCFAs per se but it is conceivable that the daily dietary intake of BCFAs needs to be higher than the current estimates to achieve potential health benefits.

Very little is known about the fate and metabolism of dietary BCFAs in humans, thus representing an important area for future research. While information is scarce, it can be presumed that digestion, absorption, and transport of BCFAs in humans is comparable to that of other long-chain FAs. Our knowledge of BCFA uptake by cells is based largely on work in cell lines, specifically from the intestine (Caco-2), fetal small intestine (H4), and breast cancer (SKBR-3 and MCF-7), following exposure to one

or a mixture of BCFAs [45–49]. From this limited perspective, it is thought that the cellular uptake of BCFAs is dependent on BCFA length and configuration. Using Caco-2 cells, Yan et al. [45] demonstrated that BCFAs can be taken up and further metabolized into their elongation or chain-shortened products. Other work indicates that in fetal intestinal cells incubated with identical concentrations of BCFAs *(*i.e., *iso*-16:0, *anteiso*-17:0, *iso*-18:0, and *iso*-20:0), *anteiso*-17:0 exhibited the greatest uptake efficiency, followed by *iso*-16:0, *iso*-18:0, then *iso*-20:0, with significant differences, and a specific hierarchy, between these BCFAs [47]. In contrast, MCF-7 human breast cancer cells accumulated greater amounts of *iso*-15:0 and *iso*-17:0 compared to their *anteiso* analogs [49]. Dietary supplementation of BCFAs to neonatal rat pups demonstrated that BCFAs incorporate into ileal phospholipids as well as liver tissue and serum [50]. Research conflicts as to whether BCFAs are preferentially incorporated into cellular triacylglycerols or phospholipids [47,48], and whether the FA structure plays a determinant role requires further investigation.

#### **4. Occurrence of BCFAs in Dairy Products**

#### *4.1. Content and Composition of BCFAs in Milk*

Within dairy foods, research to date has mainly focused on the content and composition of BCFAs derived from cow's milk. From this work, studies established that cow's milk is a significant source of BCFAs, typically comprising 1.7–3.4% BCFAs of total FAs (Table 1). When considered on a per serving basis, three servings of whole milk (3.25% milk fat) per day can thus provide 367–763 mg of BCFAs (Table 1). In milk fat, monomethyl BCFAs are the principal BCFAs present, with a chain length of 14–17 carbons and the methyl group occurring in either an *iso* or *anteiso* configuration. In general, total *iso*- and *anteiso*-BCFAs isomers occur in an approximate ratio of 1:1 (Table 1). Short-chain (4:0–6:0) and medium-chain (7:0–13:0) *iso*- and *anteiso*-BCFA isomers are also present in cow's milk but in rather a minor abundance, accounting for approximately 0.01 and 0.12% of total BCFAs, respectively [51].


**Table 1.** Branched-chain fatty acid composition and content of cow's milk.

<sup>1</sup> FA <sup>=</sup> fatty acid. <sup>2</sup> Averages were estimated by calculating the median of reported FA values when appropriate. <sup>3</sup> Content of FAs per serving was calculated from FA values listed (% of total FAs) as described by Bainbridge et al. [13], assuming 3.25% fat per serving whole milk, then multiplied by three. <sup>4</sup> Raw data obtained from previously published work [4]. <sup>5</sup> BCFAs = branched-chain FAs; sum of *iso* and *anteiso* BCFA isomers (13:0–18:0).

Multimethyl BCFAs, namely phytanic and pristanic acid, are consistent constituents within milk fat, but are often overlooked or ignored, likely because of their very low content and analytical limitations. In cow's milk, the content of phytanic acid typically ranges between 0.10 and 0.50% of total FAs, equivalent to 7–37 mg/serving whole milk [12,23,52–56]. The presence of phytanic acid in milk is specifically dependent on the feed type of the dairy cow, with higher amounts associated with the intake of grass-based rations rich in chlorophyll [23,52,53,55]. Pristanic acid, a metabolite of phytanic acid, occurs in very low amounts in milk fat, at approximately 0.04–0.06% (3–4 mg/serving whole milk) [23].

#### *4.2. BCFAs in Milk across Ruminant Species*

A survey of the recent literature regarding the BCFA composition of ruminant milk shows that milk from sheep and goats comprises approximately 1.8–3.1% and 1.2–2.4% BCFAs of total FAs, respectively, and is thus comparable to cow's milk (1.7–3.4% BCFAs of total FAs; Table 1). Yet, the content of total and individual BCFAs in milk varies considerably based on the ruminant species of origin (Table 2). In particular, on a per serving basis, milk from sheep contains considerably more BCFAs per serving compared to milk from cows (204–502 mg/serving versus 123–254 mg/serving). In the U.S., cow's milk is standardized to 3.25% milk fat (i.e., whole milk), however, this practice is less common for specialty milk such as goat and sheep milk. Differences in the BCFA content of ruminant-derived milks therefore appear to be chiefly reflective of milk fat standardization practices in retail milk, rather than species differences in BCFA occurrence.

#### *4.3. Comparison of BCFAs among Dairy Products*

The content of BCFAs in dairy also differs depending upon the type of dairy product (Table 2). For example, Ran Ressler et al. [11] found that yogurt contained an equivalent of 152 mg BCFAs/serving, while butter contained an equivalent of 195 mg/serving. Limited work assessing the BCFA content in cheese demonstrates that cheese type appears to be an important consideration as well (82–322 mg/serving depending upon cheese variety). One notable source of variation in the content of BCFAs per serving among cheese types is the differing fat content. For example, one serving (28.35 g) of low-moisture Mozzarella (22.1% fat) contains 82 mg BCFAs, whereas Cheddar (33.0% fat) contains 148 mg BCFAs [11]. One serving of butter (14.2 g), another dairy-derived product high in fat (<80%), can contain more than 200 mg of BCFAs (137–204 mg/serving; Table 2).

#### **5. Human Trials Assessing Health E**ff**ects of BCFAs**

To date, epidemiological research indicates that dairy fat consumption is neutral or protective against cardiometabolic diseases. However, the specific role of BCFAs is less understood, as these studies largely rely upon either food frequency questionnaires and/or the measurement of specific dairy FA biomarkers (i.e., 15:0, 17:0, and 16:1 *t*9) in blood or tissues [57–60], but not BCFAs.

#### *5.1. Obesity*

Epidemiological studies have noted that the intake of full-fat dairy products can be beneficial for long-term weight maintenance [61,62]. Other research which specifically focused on the consumption of BCFAs suggests that BCFAs have an important role in energy homeostasis. For example, one study found that total BCFAs in serum were higher in non-obese women than obese women and that *iso*-BCFAs were inversely associated with body mass index [38]. Similar results have been reported in a recent study by Pakiet et al. [63]. Another study found that the adipose tissue of lean subjects had higher proportions of total BCFAs and individual BCFAs than obese individuals [37].

One niche of scientific interest is the relationship between tissue concentrations of BCFAs and weight status in individuals post-gastric bypass surgery. In a longitudinal study, the BCFAs of adipose tissue in obese subjects was assessed at baseline and one year following Roux-en-Y gastric bypass surgery [37]. This study demonstrated that, after one year, the proportion of total BCFAs increased after surgery-induced weight loss. A follow-up study found similar results, showing that serum BCFAs increased in individuals after one-anastomosis gastric bypass surgery [63]. Of note, a similar study found no changes in serum BCFAs two weeks after one-anastomosis gastric bypass surgery, suggesting that a minimum time period may be needed to observe changes [64].



Averages by calculating reported appropriate; per serving reported (% FAs), assuming 93.3% of FAs in milk fat (correction for glycerol) [77]. 2 FA = fatty acid. 3 Cow-derived; based on the assumption of 3.25% fat, 244 g per serving [13]. 4 Cow-derived; based on reported % fat and serving sizes (g) listed by FoodData Central for each cheese variety [78–89] with the exception of Ricotta cheese (62 g/serving). 5 Cow-derived; based on reported % fat and serving size (245 g) listed by FoodData Central [90]. 6 Cow-derived; based on reported % fat (80%) and serving size (14.2 g) listed by FoodData Central [91]; when FA values were reported as μg/g butter [70], FAs per serving were calculated by assuming 14.2 g per serving [91] without correction factor for glycerol. 7 Based on reported % fat (median of values calculated when appropriate) or % fat listed by FoodData Central [92] if not reported; based on serving size (245 g) listed by FoodData Central [92]. 8 Based on reported % fat (median of values calculated when appropriate) or % fat listed by FoodData Central [93] if not reported; based on serving size (244 g) listed by FoodData Central [93]. 9 Includes Ricotta, Romadur, Cottage cheese, Camembert, Brie, Limburger, Feta, and Bavaria Blue. 10 Includes Provolone, low-moisture Mozzarella, Emmental, Cheddar, Montasio, Gouda, and Butter cheese (Butterkäse). 11 Includes Swiss, American, Alpine cheese, curd cheese, and Mozzarella (varieties without su fficient information reported to categorize). 12 Raw data underlying previously published work [4]. 13 BCFAs =branched-chain FAs; sum of *iso* and*anteiso* BCFA isomers (13:0–18:0).

1

#### *5.2. Insulin Sensitivity*

Serum collected from fasted individuals showed a modest inverse correlation between serum BCFA concentrations and an index of insulin resistance (homeostatic model of assessment of insulin resistance), suggesting that BCFAs may promote insulin sensitivity [63]. Similarly, total BCFAs in adipose tissue have been associated with measurements of insulin sensitivity (i.e., insulin-stimulated glucose rate of disappearance) [37]. Moreover, Mika et al. [38] observed that two BCFAs, *anteiso-*15:0 and *iso-*17:0, were negatively correlated with fasted serum insulin levels but not with the homeostatic model assessment of insulin resistance.

#### *5.3. Limitations of Current Evidence Available in Humans*

Research suggests that BCFAs promote weight maintenance and metabolic health [37,38,63], however, more studies are needed to ascertain the specific impact of dietary BCFAs, and to what extent dietary BCFAs contribute to these outcome measures. Research assessing the metabolic effects of dairy-derived BCFAs is still lacking. Therefore, for this review, we considered studies evaluating the role of BCFAs on health, regardless of the source. While dietary patterns are the plausible source of variation in tissue BCFA occurrence in humans, BCFAs derived from an individual's gut microbiota or metabolism from BCAAs cannot be discounted. Future work evaluating the relationship between diet and disease risk should consider the inclusion of BCFAs as biomarkers of dairy fat consumption.

#### **6. Role of Dairy-Derived BCFAs in Health: Potential Mechanisms**

#### *6.1. Inflammation*

Inflammation is an important underlying factor contributing to the pathogenesis of many metabolic diseases [94–96]. In particular, gastrointestinal health is critical for proper immune function and regulation of inflammation [97,98]. Research utilizing gastrointestinal cell lines has demonstrated the potent anti-inflammatory potential of dietary BCFAs. For example, exposure of Caco-2 cells to BCFAs reduced the lipopolysaccharide-induced gene expression of important proinflammatory mediators (i.e., IL-8, TLR-4, and NF-κB) [45]. In addition, certain shorter chain BCFAs (*i.e*., *iso*-14:0, *iso*-16:0, *anteiso*-13:0) improved the lipopolysaccharide-induced decrease in cell viability. Similar results were reported in a follow-up study [46] when Caco-2 cells were treated with monoacylglycerols or free FAs comprised of ~30% BCFAs isolated from the vernix.

Animal work is limited, but also supports that BCFAs attenuate inflammation and related diseases. Ran-Ressler et al. [50] examined the effect of a diet containing a mixture of BCFAs (*iso*-14:0, *anteiso-*15:0, *iso-*16:0, *anteiso*-17:0, *iso*-18:0, and *iso*-20:0) on necrotizing enterocolitis in neonatal Sprague–Dawley pups. BCFA-fed pups had a lower incidence of necrotizing enterocolitis and an enhanced gene expression of IL-10, an anti-inflammatory cytokine. Moreover, the cecal samples of pups fed the BCFA-enriched rat formula had a greater abundance of *Bacillus subtilis* and *Pseudomonas aeruginosa* compared to those fed standard rat formula, which is notable as both are known to contain BCFAs in their membranes [18,50,99]. Of note, an inverse association has been found in humans for serum *iso*-BCFAs (i.e., *iso-*15:0, *iso-*16:0, *iso-*17:0, and *anteiso-*15:0) and serum C-reactive protein [38], an important inflammatory marker for type 2 diabetes risk.

#### *6.2. Anticarcinogenic Properties*

One of the first beneficial effects attributed to BCFAs were their anticancer properties discovered 20 years ago. Specifically, *iso-*15:0 is known to inhibit the growth of T-cell non-Hodgkin lymphoma (Jurkat, Hut78, and EL4 cell lines [100]) and various types of carcinoma cell lines including, breast (MCF-7 and SKBR-3), prostate (DU145), lung (NCI-H1688), pancreas (BxPC-3), liver (SNU-423), bladder (T24, 5637, and UM-UC-3), leukemia (K-562), and gastric (NCI-SNU-1) and colorectal carcinoma (HCT 116) in a dose- and/or time-dependent manner by inhibiting proliferation and inducing apoptosis [48,49,101,102]. In a comparison of eight *iso*-BCFA species of varying carbon chain

lengths (*iso*-12:0 through *iso*-20:0), Wongtangtintharn et al. [103] demonstrated that the anticarcinogenic activity of BCFAs are dependent on the chain length, with *iso*-16:0 exerting the highest cytotoxic activity. Furthermore, a more recent study showed that BCFAs with an *iso* structure (i.e., *iso*-15:0 and *iso*-17:0) were more potent than their *anteiso* counterparts (i.e., *anteiso*-15:0 and *anteiso*-17:0) [49]. Similarly, in vivo experiments have shown that BCFAs inhibit tumor growth in both mouse xenograft tumor models in which cancer cells were co-grafted [100,101] and in the orthotopic VX2 squamous cell carcinoma model in rabbits [104].

BCFAs may also induce beneficial, location-specific effects on angiogenesis. Treatment with *iso-*15:0 promoted angiogenesis post cerebral ischemia/reperfusion injury [105], while *anteiso*-15:0 suppressed angiogenesis to aid corneal recovery [106]. These effects appear to be structure specific, but studies in this area are still very limited, hence more research is needed. Moreover, the biological significance of the role of BCFAs on angiogenesis in the context of cancer growth is not yet established.

#### *6.3. Energy and Glucose Homeostasis*

Previous research has suggested that the body content, and specifically the liver content, of BCFAs reflects the ingestion of their respective BCAA precursors [30,107]. Furthermore, Brooks et al. [107] and Garcia-Caraballo et al. [30] observed that the body or hepatic content of BCFAs were inversely correlated with the body triacylglycerol content, as shown in *C. elegans*, or the liver triacylglycerol content, as shown in mice, respectively. Likewise, a recent in vitro study, using a human fatty liver cell line model (i.e., L02 cells), concluded that individual BCFAs (*iso-*15:0 and *iso-*18:0) reduced the cellular triacylglycerol content and were associated with an upregulation of multiple genes involved in lipid catabolism [108]. These observations are in accordance with human trials which showed that, in serum, total BCFAs [63] and specific *iso*-BCFAs [38] had an inverse relationship with triacylglycerols. Thus, the results from these studies indicate that BCFA may be directly or indirectly involved in the regulation of fat storage in the body.

As described above, BCFAs may favorably influence insulin sensitivity in humans, however, studies assessing potential mechanisms involved are scarce. One study performed with rat insulinoma INS-1 β-cells demonstrated that *iso-*17:0 may beneficially modulate β-cell function via the upregulation of Pdx1 and PPAR-γ transcription factors [109]. Yet, more in vivo research is necessary to confirm these results.

#### *6.4. Biological Functions of Polymethyl BCFAs: Phytanic Acid*

The health effects of dietary phytanic acid, both positive and negative, have been extensively examined in a recent review [110]. In particular, Roca-Saavedra et al. [110] summarize the biological complications that arise due to an excess of phytanic acid within the body, such as neurological injury, oxidative stress, and cancers. To that end, recent in vitro work continues to show that phytanic acid, when provided at concentrations exceeding normal physiological ranges, is neurotoxic [111,112]. Importantly, the nutritional relevance of such studies is called into question for the general population, as the accumulation of phytanic acid due to impaired lipid metabolism (e.g., Refsum disease) is thought to be rare in terms of prevalence [113]. Nakanishi et al. [114] demonstrated that when cells are treated with concentrations of phytanic acid comparable to that found in the plasma of healthy individuals, phytanic acid exerted beneficial immunomodulatory effects in a PPAR-α-dependent manner. Furthermore, there is no clinical evidence that the consumption of phytanic acid in naturally occurring amounts, e.g., phytanic acid derived from milk and dairy products, detrimentally impacts health in individuals with normal lipid metabolism.

Roca-Saavedra et al. [110] also describe favorable effects of phytanic acid on glucose and lipid metabolism, energy expenditure, as well as immune and anticancer functions. Indeed, efforts to ameliorate or prevent obesity have been gaining momentum, and phytanic acid has been attracting scientific interest as an active compound for potential novel therapeutic drugs. Emerging research identified phytanic acid as a potent agonist of PPAR-α [115], a transcription factor expressed in metabolically active tissues controlling cellular FA oxidation and regulating energy homeostasis [116]. An et al. [115] showed that phytol treatment in a high-fat diet-induced mouse model of obesity resulted in greater contents of phytanic acid in liver and brown adipose tissue as well as the activation of PPAR-α and its target genes. Additionally, a recent study reported that phytanic acid promotes beige adipogenesis in murine 3T3-L1 (white) adipocytes (also called browning) in a PPAR-α ligand-dependent manner [117]. Taken together, these novel findings add support to phytanic acid possibly playing a role in the regulation of energy homeostasis.

#### **7. Conclusions**

BCFAs are a class of saturated FAs that comprise a significant portion of total FAs in milk and dairy products. While humans can derive a limited amount of BCFAs from endogenous synthesis from BCAAs, dietary BCFAs remain the most important source of BCFAs within the body. Recent human trials indicate that BCFAs in tissues have a beneficial influence on metabolism. Mechanistic research validates these studies by demonstrating that BCFAs possess a wide range of structure-specific functions that may favorably influence health at the cellular and systemic level. Despite the promise of BCFAs as pro-health dietary constituents, our narrative review reveals a critical scarcity in the scientific understanding of the relationship between diet-derived BCFAs and disease risk. Of note, research has been slow to utilize these FAs as markers of dairy fat intake. Here we propose that BCFAs may be a useful and complimentary biomarker of dairy fat intake for future studies. It is clear that more research is needed to clarify the diverse biological roles of BCFAs in vivo, particularly through clinical and epidemiolocal studies that evaluate the relationship between BCFA consumption, food source, BCFA tissue concentrations, and disease.

**Author Contributions:** Conceptualization, J.K and M.T.-G.; Data curation, V.M.T., A.L.U., M.R.S., and J.K.; Writing—original draft preparation, V.M.T., A.L.U., M.R.S., and J.K.; Writing—review and editing, M.T.-G. and J.K.; Visualization, M.R.S. and J.K.; Supervision, J.K.; Project administration, M.T.-G. and J.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** M.T.-G. is employee of National Dairy Council. J.K. has received research funding from National Dairy Council.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Review* **Dairy Consumption and Metabolic Health**

**Claire M. Timon 1, Aileen O'Connor 2,3, Nupur Bhargava 2,3, Eileen R. Gibney 2,3,\* and Emma L. Feeney 2,3**


Received: 26 August 2020; Accepted: 1 October 2020; Published: 3 October 2020

**Abstract:** Milk and dairy foods are naturally rich sources of a wide range of nutrients, and when consumed according to recommended intakes, contribute essential nutrients across all stages of the life cycle. Seminal studies recommendations with respect to intake of saturated fat have been consistent and clear: limit total fat intake to 30% or less of total dietary energy, with a specific recommendation for intake of saturated fat to less than 10% of total dietary energy. However, recent work has re-opened the debate on intake of saturated fat in particular, with suggestions that recommended intakes be considered not at a total fat intake within the diet, but at a food-specific level. A large body of evidence exists examining the impact of dairy consumption on markers of metabolic health, both at a total-dairy-intake level and also at a food-item level, with mixed findings to date. However the evidence suggests that the impact of saturated fat intake on health differs both across food groups and even between foods within the same food group such as dairy. The range of nutrients and bioactive components in milk and dairy foods are found in different levels and are housed within very different food structures. The interaction of the overall food structure and the nutrients describes the concept of the 'food matrix effect' which has been well-documented for dairy foods. Studies show that nutrients from different dairy food sources can have different effects on health and for this reason, they should be considered individually rather than grouped as a single food category in epidemiological research. This narrative review examines the current evidence, mainly from randomised controlled trials and meta-analyses, with respect to dairy, milk, yoghurt and cheese on aspects of metabolic health, and summarises some of the potential mechanisms for these findings.

**Keywords:** dairy; health; matrix; metabolism; nutrient; composition; saturated fats

#### **1. Contribution of Dairy to a Balanced Diet**

Milk and dairy foods are naturally rich sources of a wide range of nutrients such as proteins, fats, oligosaccharides and micronutrients including vitamins A, D, E and K and Ca, Mg, P and Zn [1], Figure 1. Milk proteins are of high biological value, not only because they contain essential amino acids but also because of their high digestibility and bioavailability. Approximately 80% of milk protein is casein and the remaining 20% is serum, or whey protein for cow's milk [2]. Fat, mainly in the form of triacylglycerols (98%), is present in milk as globules which are surrounded by a membrane (or milk fat globule membrane (MFGM)). This component of milk fat has been suggested to elicit favourable lipid and low-density lipoprotein (LDL) cholesterol response to dairy consumption [3], and will be discussed later. With respect to micronutrients, milk is considered a major source of calcium in the diet [4]. As well as being a rich source of this nutrient, the bioavailability of calcium from dairy sources has also been shown to be higher compared to other dietary sources [5,6]. Furthermore, modelling of dietary

intake data has indicated that, without consuming dairy products, less than half of the dietary calcium requirements would be met [7]. The authors of that study also noted that nutrients from dairy foods are difficult to replace and modelled removal and replacement with available alternatives, which resulted in lower amounts of several nutrients including protein, phosphorus, riboflavin, zinc and vitamin B12 [7]. While there is some concern that avoidance of dairy may have implications for some nutrients, this does not suggest that adequate nutrient intake from a low-dairy or dairy-free diet is unattainable, but rather indicates that dairy can significantly contribute towards a healthy diet.

**Figure 1.** Nutrient content and associated health benefits of dairy consumption.

When consumed according to recommended intakes of national guidelines, milk and dairy products contribute essential nutrients across all stages of the life cycle [1]. For example, milk and dairy products are an important part of a young child's diet as they are a good source of energy and protein and contain a wide range of vitamins and minerals, especially calcium, that young children need for healthy bones and teeth [8]. In Europe, it is reported that milk contributes proportionally more to the diets of young children than to adults [9,10]. Data from cross-sectional studies [11] and intervention studies [12] have reported the positive effect of milk consumption in childhood and adolescence on bone mineral content and bone mineral density. In addition, some research studies have indicated that the consumption of milk and milk products during adolescence is associated with neutral or reduced risk of adiposity [13,14]. During pregnancy, dairy products can be an important means of providing adequate calcium and other key nutrients in the diet [15]. Evidence from prospective cohort studies suggests that moderate milk consumption compared to none or low intakes during pregnancy is positively associated with foetal growth and infant birth weight in healthy, Western populations [16]. Finally, several studies point to the benefits of milk and dairy products in diets of the elderly and highlight that, in combination with physical activity, milk and dairy products can improve muscle mass and function resulting in a lower risk of sarcopenia and vertebral fractures [16], although another recent review of the area did not find strong evidence for a benefit of milk on muscle health in older adults [17].

Current consumption patterns of dairy are in a period of considerable change, with reported decreases in dairy consumption in countries who have traditionally consumed large quantities, potentially due to increased intake of 'dairy-free' alternative products [18] and reported increases in some global regions, where dairy has not been commonly consumed [19]. These changes are reported to be due to several factors, but are predominantly driven by consumer perception of the health effects of dairy consumption and the environmental impact of dairy production [20]. Particular nutrients of concern, when considering intake of dairy, are sodium and fat. Whilst dairy

has been shown to contribute beneficially to the diet, dairy foods also contribute significantly to sodium and saturated fat intakes (Figure 1). High dietary salt intake is a prominent factor for the development of hypertension, a strong risk factor for cardiovascular disease [21,22]. In the US and UK, dairy products are significant contributors to dietary salt intakes, providing approximately 11% and 8% of overall intakes, respectively [23,24]. Dietary efforts to reduce hypertension include the well-established DASH diet (Dietary Approaches to Stop Hypertension), where low-fat dairy products are important characteristics [25]. In relation to dairy fat, the 2011–2014 National Health and Nutrition Examination Survey (NHANES) demonstrated that dairy foods contributed 26% of saturated fat and 14.2% of total fat to the diets of US adults [26].

Similarly, Feeney et al. observed that, within the Irish diet, dairy foods contributed 12.8% of total fat to the diet, and 19.8% of the saturated fat [27]. For this reason, many healthy eating guidelines recommend 3–5 portions of dairy daily, with consumption of "low/reduced-fat dairy" when possible [28,29]. However, research investigating the importance of the food source of saturated fatty acids (SFA) suggests that although SFAs from meat and processed-meat are associated with detrimental health effects [30], SFA intake from dairy sources may be associated with either neutral [31] or beneficial effects on cardiovascular health markers [32,33]. Further, the individual dairy sources may have different impacts. Much work in this area is underway, which is summarised below.

#### **2. Dairy Fat and the Link to Health**

Since the seminal Seven Countries Study and other subsequent studies [34,35], recommendations with respect to intake of saturated fat have been consistent and clear: limit total fat intake to 30% or less of total dietary energy, with a specific recommendation for intake of saturated fat to less than 10% of total dietary energy [36–38]. However, recent work has re-opened the debate on intake of saturated fat in particular [39]. In a large review and meta-analysis, de Souza et al. examined associations between intake of total fat, saturated fat and trans-unsaturated fat with all-cause mortality and differing morbidities. The authors concluded that, contrary to previous evidence, saturated fat intake was not associated with all-cause mortality, cardiovascular disease (CVD) mortality, total coronary heart disease (CHD), ischemic stroke or type 2 diabetes [39]. Drouin-Chartier et al. in 2016 also examined the impact of dairy consumption and dairy fat on cardiometabolic disease risk factors, and also reported that the purported detrimental effects of SFAs on cardiometabolic health may in fact be nullified when they are consumed as part of complex food matrices such as those in cheese and other dairy foods [40]. Similarly, Alexander et al. (2016) completed a meta-analysis of prospective studies and the intake of dairy products and CVD risk. These authors more cautiously concluded that although for some individual dairy products risk estimates below 1.0 were observed, additional data are needed to more comprehensively examine potential dose–response patterns [41].

#### *2.1. Dietary Guidelines—Nutrient-Based vs. Food-Based*

Despite these more recent findings, the most recent review of published literature by the UK Scientific Advisory Committee on Nutrition (SACN) concluded that there is a significant body of evidence demonstrating a relationship between intake of saturated fats and CVD and CHD events, but not CVD and CHD mortality [28], and noted that, irrespective of the lack of evidence for an effect on mortality, non-fatal CVD and CHD events have a serious adverse impact on health and quality of life, and that existing public health recommendations for saturated fat to be <10% of total dietary energy intake were still valid [42]. This mirrors recommendations in other countries and regions of the world [36–38]. However, some criticisms on the continued support for such recommendations note that such policies are based on evidence from total dietary saturated fat intake, and may have not considered the source of fat, whereby the food source, or matrix may in fact have a differing influence on metabolism and subsequently, health [43]. In addition, different food sources contain different types and amounts of SFA, and the continued promotion of an overall <10% of total dietary energy recommendation may be perceived to overlook this [43]. Further, by focusing on SFA content alone, foods that are nutrient-dense but also high in SFA may be excluded from the diet and inadvertently result in a reduced intake of important micronutrients [43]. For this reason, many are advocating for food-based guidelines rather than nutrient-based advice. As such, evidence may need to be considered at a food-item level within the dairy food group, rather than together. In light of a reported shift in recent dairy consumption [19], where instances of decreased dairy consumption may be potentially explained by an increase in dairy-free alternatives due to the perceived impact of dairy on health [18], it is important to consider some of the recent evidence at both a total-dairy level and for individual dairy products cheese, milk and yoghurt, and then further discuss potential mechanisms influencing the metabolic response to consumption.

#### *2.2. Total Dairy vs. Specific Products*

Dehghan et al. (2018) specifically examined the associations between total dairy and specific types of dairy products with mortality and major cardiovascular disease. Dietary intakes of dairy products for 136,384 individuals were recorded using country-specific validated food frequency questionnaires and associations with mortality or major cardiovascular events were examined [44]. The authors reported that a higher intake of total dairy (>2 servings per day compared with no intake) was associated with a lower risk of the composite of mortality or major cardiovascular events, total mortality, non-cardiovascular mortality, cardiovascular mortality, major cardiovascular disease and stroke. No significant association with myocardial infarction was observed. Higher intake (>1 serving vs. no intake) of milk and yoghurt was associated with lower risk of the composite outcome, whereas cheese intake was not. Butter intake was low and was not significantly associated with clinical outcomes. The authors concluded that dairy consumption was associated with lower risk of mortality and major cardiovascular disease events in a diverse multinational cohort [44].

Fontecha and colleagues [45] specifically examined evidence regarding the influence of dairy product consumption on the risk of major cardiovascular-related outcomes and how various doses of different dairy products affected such responses. In this overview of 12 meta-analyses involving randomised controlled trials (RCTs), as well as the updated meta-analyses of RCTs, increasing consumption of dairy products did not result in significant changes of known risk biomarkers such as systolic and diastolic blood pressure and total cholesterol and LDL cholesterol. They concluded that consumption of total dairy products (either regular or low-fat content), did not adversely affect the risk of CVD [45].

Considering total dairy consumption initially, characteristics of fifteen published RCTs that studied the effects of overall dairy consumption on markers of metabolic health and CVD risk on variable age groups are included in Table 1. RCTs included in the table were either parallel or crossover trials and the participants' ages ranged from 20 to 75 years. Duration of interventions varied widely from 4 to 8 weeks in some to 12 to 24 weeks in the others, with a washout period included in some of the trials. Most of the interventions were based on consumption of different dairy products such as milk, cheese or butter. Zemel et al. (2010) compared the effects of consumption of soy-based smoothie to a dairy-based smoothie and observed suppressed inflammatory and stress markers for the latter [46]. Groups that consumed a low-fat or non-fat dairy diet showed decreased levels of total cholesterol (TC) and low-density cholesterol (LDL) as compared to the diet containing conventional levels of fat. Most recently, Vasilopoulou et al. (2020) examined the impact of modified dairy fat consumption, through the provision of products with modified monounsaturated fatty acid (MUFA) content in adults at moderate CVD risk—the RESET study (controlled REplacement of SaturatEd fat in dairy on Total cholesterol). The authors concluded that consumption of a high-fat diet containing modified dairy products with reduced saturated fatty acids, and enriched monounsaturated fatty acids showed beneficial effects on fasting LDL cholesterol and endothelial function compared with conventional dairy products [47]. Finally, although not an intervention study, Drouin-Chartier (2019), recently examined dairy intake and risk of diabetes, and reported positive impacts for yoghurt and reduced-fat milk, but a negative association for cheese [48].


**1.**Randomisedcontrolledtrials(RCT)demonstratingffectofoveralldairyconsumptionmarkersofmetabolichealthandcardiovascular


**Table 1.** *Cont.*



#### *2.3. Cheese*

Focusing specifically on cheese, several published RCTs have demonstrated a beneficial effect of cheese consumption on markers of metabolic health and CVD risk, summarised in Table 2. Brassard et al. (2017) compared the impact of consuming equal amounts of SFAs from cheese and butter on cardiometabolic risk factors [61]. In this multicentre, crossover, randomised controlled trial, participants were assigned to a randomised sequence of five isoenergetic diets of 4-week duration (separated by 4-week washout periods). The diets were rich in SFAs from either cheese or butter, or a monounsaturated fatty acid (MUFA)–rich diet, a polyunsaturated fatty acid (PUFA)–rich diet and a low-fat, high-carbohydrate diet. The authors reported that serum HDL-cholesterol concentrations were similar after the cheese and butter diets but were significantly higher in comparison to response after the carbohydrate diet. Comparing cheese and butter, LDL-cholesterol concentrations after the cheese diet were lower than after the butter diet but were higher than after all of the other diets. Some variation in response was noted. Work conducted by this research group previously has both supported and added to the existing evidence [32]. A 6-week randomised parallel intervention involving 164 volunteers who received ~40 g of dairy fat/d, in 1 of 4 treatments: 120 g full-fat Irish cheddar cheese (group A), 120 g reduced-fat Irish cheddar cheese + butter (21 g) (group B); butter (49 g), calcium caseinate powder (30 g) and Ca supplement (CaCO3) (500 mg) (group C) or 120 g full-fat Irish cheddar cheese, for 6 weeks following completion of a 6-week "run-in" period, where this group excluded all dietary cheese before commencing the intervention (group D). This study found that a stepwise-matrix effect was observed between the groups for total cholesterol (TC) (P = 0.033) and LDL cholesterol (P = 0.026), with significantly lower post-intervention TC and LDL cholesterol when all of the fat was contained within the cheese matrix (Group A), compared with Group C when it was not. These findings suggest that dairy fat, when eaten in the form of cheese, appears to differently affect blood lipids compared with the same constituents eaten in different matrices, with significantly lower total cholesterol observed when all nutrients are consumed within a cheese matrix. This 'dairy matrix' concept, whereby the nutrients within a dairy food interact with the overall structure providing different health effects, is becoming increasingly studied, [62], and is discussed further in Section 3, below.


 **2.** Randomised controlled trials (RCT) demonstrating an effect of cheese consumption on markers of metabolic health and CVD

**Table**

 risk.




**Table 2.** *Cont.*

Other groups have also looked at postprandial response to cheese consumption, as the postprandial response to lipid consumption is considered an independent indicator of CVD risk [75]. Drouin-Chartier (2017) reported minor differences in postcirculating TAG concentrations, in a postprandial RCT where participants ingested 33 g fat from a firm cheese (young cheddar), a soft cream cheese (cream cheese) or butter (control) incorporated into standardised macronutrient-matched meals. They conclude that the study demonstrates that the cheese matrix modulates the impact of dairy fat on postprandial lipemia in healthy subjects [75].

Hansson et al. 2019, examining postprandial response to sour cream, whipped cream, cheese and butter, noted that sour cream resulted in a larger postprandial triacylglycerol (TAG) area under the curve (AUC), compared to whipped cream, butter and cheese (P = 0.05). Intake of sour cream also induced a larger HDL cholesterol AUC compared to cheese. Intake of cheese induced a 124% larger insulin AUC compared to butter. Hansson et al. concluded that high-fat meals containing similar amount of fat from different dairy products induce different postprandial effects on serum TAGs, HDL cholesterol and insulin in healthy adults [76].

#### *2.4. Milk*

Characteristics of seven RCTs that studied the effects of milk consumption on markers of metabolic health and CVD risk are described in Table 3 [77–83]. RCTs included in the table were either parallel or crossover trials and the participants' ages ranged from 20–85 years. Duration of interventions ranged from 4 to 16 weeks with a washout period included in some of the studies. Gardner et al. (2007) compared the effects of soy-based drinks to dairy milk consumption on metabolic health markers, and found that LDL was significantly lower after consuming soy milk compared to dairy milk, but no significant differences between groups were observed for HDL, triacylglycerols, insulin or glucose [77]. In 2016, Lee et al. examined the impact of milk consumption (400 mL per day) compared to habitual intake on markers of metabolic health, and found no significant differences in body mass index, blood pressure or lipid profile [78]. Hidaka et al. compared intakes of full-fat vs. non-fat milk intakes and showed lower plasma triglyceride and phospholipid levels in the no-fat group [79].

#### *2.5. Yoghurt*

Table 4 presents a summary of RCTs that specifically studied the effects of yoghurt consumption on markers of metabolic health and CVD risk. Most of these studies focused specifically on the microbial content of yoghurts, where consumption of probiotic yoghurt or modified-bacterial-strain-containing yoghurts or comparisons between commonly available varieties of yoghurts such as non-fat yoghurt, natural yoghurt and heated yoghurt were examined. In general, when compared to participants on a controlled diet (diet with zero consumption of fermented products), the participants who consumed probiotic or conventional yoghurt showed significant decrease in TC and LDL cholesterol [84–92].


Randomisedcontrolledtrials(RCT)demonstratinganeffectofmilkconsumptiononmarkersofmetabolichealthandcardiovasculardisease(CVD)



**Author**

Gardner et al.

2007 [77]

USA

 Crossover

 **Country**

 **Study Design**

 **Population**

*n* = 28 (6 male;

22 female),

30–65

hypercholesteraemic

 **Age (Years)**

 **Intervention** 3 arm: 32 oz/d whole soy bean drink.

**(B)** Soy protein isolate drink, 28 oz/d. **(C)** Dairy milk, 18.5 oz/d, (all volumes were

 **Duration** Three 4-week

LDL was significantly lower after consuming soy milk in groups **A** + **B** compared to dairy milk (group

**C**). No significant differences between groups were

periods, separated

by 4-week

 **Main Findings**

**et al. 2010** [86]

Iran

 Parallel

healthy volunteers

19–49

**(C)** Control, no consumption of fermented

products

6 weeks

group A + B compared with the control as well as a

decrease in TC:HDL ratio. HDL increased in group A

compared with the control.


**Table 4.** *Cont.*



#### **3. The 'Dairy Matrix'**

While dairy products are often considered together as a food category in nutritional epidemiology, they vary considerably in terms of their content and structure and how these interact with other food components, which describes the 'dairy matrix' concept [93]. Values from the Composition of Foods Integrated Dataset (CoFID) were used to compare nutrient composition across the range of commonly available dairy products (summarised in Figure 2). While they represent a wide range of products, only plain, unflavoured versions with no added sugar were included in this analysis. Figure 1 demonstrates how nutrient composition across the range of products varies greatly, as does their overall matrix or structure, depending on the product type. For example, whole milk contains 3.6% fat in a liquid oil-in-water emulsion with lactose and protein (both casein and whey) while cheeses mainly consist of casein proteins and fat, in a solid matrix, with only trace levels of lactose and whey [62]. Butter is an emulsion of water-in oil, and contains mostly fat and water, with no protein or carbohydrate, while (liquid) cream is also a water-in-oil emulsion, and contains low levels of protein (approx. 2%) [62] and lactose (approx. 3%) (in 35%-fat cream). The various processing steps that different products undergo from raw milk to final foodstuff impact the level of the different nutrients and the overall macro and microstructure of these foods. The nature of these differences may result in the different health outcomes associated with their consumption.

**Figure 2.** Boxplot showing average fat and mineral content in dairy products per 100 g. Values from the Composition of Foods Integrated Dataset (CoFID). Only plain, unflavoured products with no added sugar were included in this analysis. Nutrients shown are: (**A**) total fat (g), (**B**) calcium (mg), (**C**) magnesium (mg) and (**D**) phosphorus (mg), calculated from *n* = 4 butter, *n* = 43 cheese, *n* = 8 cream, *n* = 20 milk and *n* = 4 yoghurt samples.

Cheese in particular is associated with lower levels of blood cholesterol than other dairy products and especially when compared to butter (see Table 2 for an overview of studies in this area). Cheese structures contain aggregated casein micelles [94] which may impact the ability of lipases to break down the fat contained within the matrix, compared to the same fat contained within other food matrices (e.g., milk and butter). The structure of cheese, including the degree of hardness and cohesiveness, can result in it being more physically resistant to digestion than other matrices [94] which affects the degree to which the fat can be digested and absorbed. There may be additional effects from the calcium contained within this matrix, reacting with the fatty acids to form insoluble calcium

soaps [95], as well as the separate textural effects from calcium that increase the cohesiveness [94] which may result in enhanced digestive resistance. This mechanism also appears to be supported by a higher faecal fat excretion following cheese consumption compared to other sources, in postprandial studies [67] although is not fully confirmed [68]. In addition to the calcium content of cheese, the phosphorus content is also implicated in the reduction of fat digestibility, by affecting the ability of cheese constituents to form insoluble soaps during digestion (calcium phosphate). This is thought to further increase fat excretion via the adsorption of bile acids to the surface, and has been implicated in the reduction of LDL-c [62,95]. Both phosphorus and calcium are particularly concentrated in cheese compared to other dairy products [96].

Cheese is a fermented dairy product, and the fermentation process may also be one of the contributors to the cardiometabolic protective nature of cheese via a number of mechanisms. Lactic acid bacteria found in fermented dairy products can result in platelet-activating factor (PAF)-inhibitory lipid production [97,98]. Further, as cheeses ripen and age, shorter peptides are produced and in some cases there is a release of latent bioactives, as some of these peptides have specific bioactivity that is not apparent in the intact 'parent' protein [99,100]. In cheeses, antihypertensive peptides are produced during fermentation, including V-P-P and I-P-P, which are tripeptides that exert their effects via inhibition of the angiotension converting enzyme (ACE) pathway [101,102]. Some cheeses have also been found to have bioactivity related to glycaemic control [103], which may also contribute to the cardio-metabolic benefits from cheese consumption. The starter culture used in the cheesemaking process can have an additional impact on the inherent bioactivity produced during fermentation [104,105]. Finally, the form in which fat is contained in cheese compared to butter may also result in some of the differences observed between these two products. The polar lipids found in dairy products in general appear to have anti-inflammatory properties compared to other oxidised dietary lipids [106–108] and they are mostly contained in a bioactive envelope surrounding the fat, known as the milk fat globule membrane (MFGM) [109]. Cheese contains particularly high levels of polar sphingolipids that are not present in the same levels in butter, since the membrane is disrupted during the churning process [62,110]. Studies suggest that polar lipids can impact blood lipid levels in the acute postprandial period, with lower lipaema (and insulin) observed following a liquid meal of palm fat when an MGFM-rich dairy fraction was added, compared to the same meal without this addition [111].

With research strongly suggestive of postprandial hyperlipidema as an independent risk factor for CVD [112], this could be a further explanation for the growing list of studies showing a protective effect of cheese consumption on CVD risk [32,60,61,68,72,113–115] despite the relatively high SFA content.

#### **4. Conclusions**/**Future Directions**

This paper summarises and discusses the evidence that examines the link between dairy intake and CVD risk. Whilst the evidence is mixed for some dairy foods (milk), it is more consistent for others (cheese/yoghurt), and supports the concept that the source of saturated fat intake has an important impact on cardiometabolic response to consumption. This is not a new concept, but whilst this evidence is growing, more research is needed before any significant change in public health recommendations are implemented.

While the link between dairy foods and metabolic health has been well-studied, research gaps still remain, and must be considered in future work. Many of the RCTs completed are short-term studies with single products, and there is a need to consider combinations of foods in a dietary pattern, considered by only a few studies to date [116,117]. The manner of the food consumption also needs to be considered, as dairy is often consumed in many forms (heated, melted) and as part of meals or recipes or in sweetened beverages. In addition, the amount of dairy product given in many of the randomised controlled trial studies to date are largely outside of recommended portion size intakes, as such, caution is required in interpreting and generalising findings.

Looking to the future, to date many of the studies have focused on traditional markers of cardiometabolic risk, particularly circulating lipid levels, which may not show subtle changes between products and/or further elucidate the cause/mechanisms for difference in response following consumption of specific foods. It has been suggested that the use of such traditional markers alone may limit the ability to predict health outcomes from the fat in dairy products, since many other components in dairy may have impact on CVD risk [118]. For this reason, more recent studies are also examining novel biomarkers that include vascular function (arterial stiffness, flow-mediated vasodilation (FMD) [47] and LDL-c particle size distribution [119]. It is important that we more fully understand the mechanisms underlying the variance in response to consumption, and the use of such novel markers will help develop this knowledge.

In conclusion, dairy foods are diverse in their structure and their nutrient content, resulting in differing biological responses and associated health outcomes. Thus, continuing to treat them as a single food category in food-based healthy eating guidelines may obscure the individual effects of these foods. As diets transition, there is an urgent need to understand the impact of different dairy foods, their preparation methods and how they are consumed, within the overall patterns of dietary intake in different cultural groups.

**Author Contributions:** Conceptualization, C.M.T., E.R.G. and E.L.F.; investigation, writing—original draft preparation, writing—review and editing, C.M.T., N.B., A.O., E.R.G. and E.L.F.; supervision, project administration, and funding acquisition, E.R.G. and E.L.F. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by the Food for Health Ireland project, funded by Enterprise Ireland, grant number TC20180025.

**Conflicts of Interest:** E.R.G. and E.L.F. have previously received travel expenses and speaking honoraria from the National Dairy Council. E.R.G. and E.L.F. have received research funding through the Food for Health Ireland project, funded by Enterprise Ireland, grant number TC20180025. The other authors are currently, or have been in the past, employed on the Food for Health Ireland project, funded by Enterprise Ireland, grant number TC20180025. The funders had no role in the design of the study; in the collection, analyses or interpretation of data; in the writing of the manuscript or in the decision to publish the results.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
