Different of the results of Kesl et al. [8] in Sprague Dawley rats, this study demonstrated that acute oral βHB mineral salts (3 g/kg) are really absorbed and reach the systemic circulation after gavage in Wistar rats and can rapidly increase ketonemia at levels similar to that found during nutritional ketosis [8]. Here, ketosis was maintained even during concomitant glucose intake, which is known to suppress ketogenesis due to the stimulation of endogenous insulin secretion [15], proving that exogenous ketones were the main and possibly the only source of ketone elevation above basal values in blood during the oral test. In the control group, on the contrary, levels of ketone bodies slightly went down during glucose ingestion only. Also, in chronic treatment, despite the ketonemia values have been lower than those achieved during the gavage test and the ketone concentration varied depending on the time of the day the ketone analysis was done (“lights on” values > “lights off”), the ketonemia values were always higher in the treated group. Therefore, this protocol is a useful long-term model to increase ketonemia, which is independent on carbohydrate intake or hepatic substrate metabolic fate (i.e. beta oxidation activation).
It is hypothesized that, once in bloodstream, exogenous βHB, like the endogenous ketones, is taken up by monocarboxylic acid transporters, metabolized to acetoacetate by D-β-hydroxybutyrate dehydrogenase and utilized as an energy source by extrahepatic tissues, such as heart, brain and muscle. However, some important differences between exogenous βHB administration (as βHB salts) and endogenous production (such as in fasting, ketogenic diets) are expected [16]. The strategy of oral supplementation of βHB salts provides the entry of this substrate without the need to synthesize it endogenously dispensing with the several steps of its production which include an increased glucagon/insulin ratio, lipolysis, transport of lipids to the liver and beta-oxidation. Thus, by using this protocol, it is possible to isolate ketone/ketosis effects and investigate whether the beneficial results frequently found during ketogenic diets are due to ketones directly or to other dietary variants, as higher protein/fat intake or carbohydrate restriction.
In addition, in our model, all signaling effects βHB can be expected, such as activation of GPR10A receptor [17]. βHB is an endogenous ligand of GPR10A receptor, which is also a receptor activated by niacin. Niacin is recognized as a pharmacological agent for the treatment of dyslipidemia, used to lower VLDL and LDL lipoproteins, while it raises HDL cholesterol [18]. Here, chronic βHB administration had effects in lipid profile, notably by raising HDL cholesterol and decreasing LDL/HDL ratio, suggesting that ketones may have similar effects of niacin on these parameters.
By sharing the same niacin receptor, this mechanism of βHB action may explain the improvement of blood lipids profile, since it has an inhibitory effect in lipolysis, through Gi protein activation and a subsequent decreasing release of serum lipolytic products, as free-fatty acids (FFA) and glycerol [17]. It is known that serum FFA directly affects other lipid transporters (VLDL/LDL lipoproteins) in blood, since it is a substrate for both triglyceride and cholesterol synthesis [18].
In our study, βHB chronic treatment appears to influence serum lipolytic products because ketonemia was negatively correlated with serum glycerol, a frequently used marker of lipolytic activity in vivo and in vitro. Evidently, our data reflect a transitory influence that might be more intense if considering other periods of the day (remembering that euthanasia and sample collection occurred at ZT 14–16, when ketonemia was lower compared to the levels found at ZT 0–1), and with a significant influence along the 4 weeks of treatment. Thus, studies are necessary to confirm that ketone supplementation runs with low lipolytic rates along the day. Anyway, a former study showed that a partial inhibition of adipocyte lipolytic rates, even without changes in basal glycerol or FFA serum concentrations, improved the animal’s metabolic conditions, such as the insulin tolerance in mice treated with hormone-sensitive lipase inhibitor [19]. It is important to figure out that the fluctuations of serum glycerol and FFA blood levels along the day are quite more complex than their release by adipocyte through lipolysis can explain. Glycerol concentrations also depend on its utilization by liver, where it is utilized as a gluconeogenic substrate. Also, above and beyond being utilized as an energy source, FFA can also be re-esterified back into triglycerides by adipocytes, in a triglyceride/fatty acid futile cycle. These points indicate that the in vivo estimation of lipolysis is not that simple, although an inhibitory effect by βHB can be hypothesized based on former in vitro and in vivo studies [20, 21].
However, in addition to what might be taking place in our protocol, a special attention is needed on the morphological changes in visceral fat considering the results from the lipid profile. The enlargement of the visceral fat cells denotes that an increased coming of FFA is occurring specifically through the mesenteric circulation toward the portal-hepatic system, and it is an important risk factor to ectopic liver fat deposition and hepatic metabolic dysfunctions [22, 23]. Therefore, our results indicate that both the smaller visceral fat cell volume found after chronic treatment and the decreased rates of lipolysis elicited by βHB treatment may have contributed to improvement of serum lipid profile. To reinforce this idea, enlarged adipocytes inversely correlate with insulin sensitivity [24] and positively with FFA release, particularly in visceral cells [25].
Regarding the HDL rise observed in βHB treated rats (Fig. 3a), it can be hypothesized that βHB can act like niacin by reducing the HDL catabolism and prolonging its half-life due to an inhibition of the hepatocyte lipoprotein uptake [26]. If the correctness of this supposition is confirmed, it could explain our data shown in Fig. 3. On the other hand, visceral fat deposition also has a well-known negative correlation with HDL cholesterol [27], and this is another way βHB treatment could be indirectly influencing the HDLc concentrations.
An additional biological effect of GPR10A activation is the increment of adiponectin secretion by adipocytes [17]. Like niacin, βHB is able to induce adiponectin secretion in primary adipocytes, although there are no data on whether this effect also happens in vivo when stimulated by ketones alone. On the other hand, ketogenic diet elevated adiponectinemia in children on a trial comparing caloric restriction and ketogenic diet [7]. Thus, these are enough evidences to lead us to hypothesize that βHB salts supplementation could increase adiponectin secretion. Furthermore, our verification of smaller fat cells in adipose tissue of βHB supplemented rats reinforce this supposition, since adiponectin synthesis and secretion is more intense in smaller fat cells. Fat depots containing a larger proportion of small adipocytes appear to have a higher ability to secrete adiponectin [28]. However, hyperadiponectinemia was not reported in our chronic protocol.
Blood glucose was another unaffected parameter in this study. In Kesl et al. [8], increased ketonemia was negatively correlated with blood glucose, suggesting a hypoglycemic effect. Indeed, Kesl et al. [8] obtained a higher ketonemia than we got here with the exogenous ketone supplements. Still, it is possible to note in Kesl’s work that ketone esters and specially MCT oil, which were the ketone supplements that most increased ketonemia, presented relevant suppression of blood glucose, perhaps demonstrating the need for a greater ketosis than that found in our experiment to achieve the same result. However, MCT oil could induce a lower glycemia by interfering with the dietary carbohydrate digestion and absorption rate or by a strong reduction in food consumption after gavage instead of its direct effect.
In our experiment, daily food consumption as well caloric intake did not change significantly (−9%, p = 0.08). Also different from βHB salts treated animals in Kesl et al. [8], our animals did not lose weight, although the percentage of visceral fat mass tended to be decreased (−16%, p = 0.0529). It should be noted that in our protocol rats received less βHB salts than in Kesl’s study (~6 g/kg vs. 5–10 g/kg, respectively). The different result between Kesl’s study and ours is possibly due to the different rat model in consideration (Sprague Dawley vs Wistar) as well the way of BHB administration (daily gavage vs drink water).
However, gathering all the results and comments plus the decreased fat cell volume found, interesting elements come out: ketogenic diets have already shown advantages in important clinical trials, at least in short/medium term, when carbohydrate ingestion is low enough to increases ketonemia [1, 2]; in mice, ketogenic diets decrease body weight and fat mass as well as a caloric restriction diet, even those with a higher caloric content per gram [29]; as Srivastava, et al. [30] reported, a shrinkage of body fat in mice fed a ketone ester/carbohydrate free diet and an increase resting energy expenditure and brown fat activity; rats fed on a diet enriched with 1,3-butanediol (a ketogenic alcohol) got a reduced body fat mass [31]. In our protocol, differently from the others above, since we did not make any change in animal diet and the caloric intake remained the same, an isolated ketotic effect was reached on visceral fat mass and tissue morphology. This data may help to explain the metabolic advantage (i.e. a better body weight/fat loss) found during ketogenic diets. Still, a dose-dependent effect of βHB salts might be possible, at least on ketonemia. For that, maybe other models of administration would be interesting, as βHB salts on food, as is being done by D’Agostino’s group [32]. Lastly, it is important to emphasize that although the amount of βHB salts supplemented here in rats was a useful tool for our experimental purposes, a translational approach to humans must be taken with care regarding the sodium and potassium ingestion and some concern must be considered. For example, in an average adult man (~70 kg), to keep the same proportions βHB salts, an ingestion of ~50 g of sodium and potassium would occur which is evidently well above the dose recommended by RDA.