Ketone bodies as chemical signals for the immune system

Abstract

Ketone bodies are short-chain fatty acids produced by the liver during periods of limited glucose availability, such as during fasting or low carbohydrate feeding. Recent studies have highlighted important nonmetabolic functions of the most abundant ketone body, β-hydroxybutyrate (BHB). Notably, many of these functions, including limiting specific sources of inflammation, histone deacetylase inhibition, NFκB inhibition, and GPCR stimulation, are particularly important to consider in immune cells. Likewise, dietary manipulations like caloric restriction or ketogenic diet feeding have been associated with lowered inflammation, improved health outcomes, and improved host defense against infection. However, the underlying mechanisms of the broad benefits of ketosis remain incompletely understood. In this Perspective, we contextualize the current state of the field of nonmetabolic functions of ketone bodies specifically in the immune system and speculate on the molecular explanations and broader physiological significance.

INTRODUCTION

The biology of ketone bodies has attracted increasing attention across fields, from metabolism to immunology, neurobiology, and even cancer biology. Ketone bodies are best known for their essential role as metabolic fuels when glucose is limited. But the most abundant ketone body, β-hydroxybutyrate (BHB), also has an increasing number of nonmetabolic signaling functions that have been expertly reviewed previously (1, 2). As specific subsets of immune cells are capable of ketone body metabolism and are sensitive to the signaling effects of BHB, the immune system appears to be a nexus for understanding the coordination and partitioning of the metabolic and nonmetabolic roles of ketone bodies.

We propose that immune cells use direct nutrient surveillance to detect perturbed metabolic environments that could indicate infection or cancer. Moreover, we suggest that nutrient uptake and/or sensing represents an alternative alarm system for immune cells that bypasses standard antigen and pattern recognition pathways. For example, the NLRP3 inflammasome has long been postulated to act as a sensor of metabolic homeostasis based on its activation and assembly being induced by a range of metabolic byproducts like ceramides, uric acid crystals, etc. (3). Linking metabolite recognition to immune cell function could be a powerful mechanism for tissue-resident immune cells to instigate homeostatic maintenance and tissue repair mechanisms, or to stimulate immune activation against other novel encounters like tumors or pathogens. Like succinate and lactate, which both have metabolic and signaling functions in immune cells (4–8), we propose ketone bodies like BHB also serve this dual purpose to coordinate immune responses according to the host’s metabolic state (Fig. 1).

Figure 1.

Selected mechanisms of immunometabolic regulation by BHB. β-hydroxybutyrate (BHB) can influence immune cell activity through multiple pathways. 1) Some immune cells express the BDH1 enzyme, critical for BHB and acetoacetate (AcAc) interconversion, allowing ketone bodies to serve as a carbon source and directly contribute to cellular metabolism. 2) BHB stimulates the G-protein coupled receptor GPR109A, and this can alter immune cell function. 3) BHB inhibits histone deacetylases (HDACs), leading to elevated lysine acetylation on histones, which can alter chromatin accessibility and gene expression to modify inflammatory responses. 4) BHB can modify proteins through covalent attachment to lysine residues, a posttranslational modification known as Kbhb, although the consequences of this modification are not known. 5) Imbalanced metabolite pools could potentially indicate tissue damage or infection and serve as a danger signal for immune cells through any of the signaling pathways discussed in this Perspective. Whether immune cells scavenge or store metabolites acquired from the environment is unclear. 6) Changes in the microenvironment, such as hormonal shifts, neuronal inputs, or other metabolites, could impact or synergistically interact with BHB-dependent signals on immune cells. Figure was created with BioRender.com.

Although direct nutrient sensing may not be a conceptual stretch of the imagination, speculating how this might be practically achieved is more complicated. Are immune cells at the whim of their metabolic microenvironment such that metabolite perturbations influence immune cell metabolism and control their activity? Given all the safety mechanisms built into the immune system, this seems unlikely. So perhaps immune cells are not passive, and use acquired metabolites to tailor their own responses according to metabolic cues? This could easily occur through metabolic reprogramming or metabolite signaling properties. Here we will explore the metabolic and nonmetabolic actions of BHB in the context of immune cells and discuss how these different functions might link metabolism and nutrient sensing to immune function at the molecular, cellular, and organismal levels.

Although glucose was long considered the primary fuel source for immune cells, studies over the last 10–15 years established that metabolic programming is dynamic and carefully regulated to control immune function (9). Most recently, ketone bodies have been identified as an important metabolite to control macrophage and T-cell functions. For example, CD8 T cells oxidize ketone bodies BHB and acetoacetate (AcAc; 10, 11), and one study even reported that memory CD8 T cells are themselves capable of ketogenesis (12). Surprisingly, when provided with BHB, CD8 T cells preferentially favor its use for acetyl-CoA production and subsequent histone acetylation, even in the presence of glucose (11). In contrast to T cells, bone marrow-derived macrophages lack expression of BDH1 to interconvert BHB to AcAc and therefore cannot metabolize BHB (13, 14). However, macrophages can oxidize AcAc (13) and conditional ablation of macrophage ketolysis increases liver fibrosis in mice fed a high-fat diet (14). Whether these paradigms extend to macrophages in other organs, particularly in nonketogenic tissues, is still unknown. These studies emphasize the metabolic importance of ketone bodies for immune protection even when glucose is scarce. We speculate this could be especially important for preserving immune function in spite of sickness behavior, a collection of physiological adaptations to infection that include reduced appetite, altered taste or smell, and digestive issues, and which may limit nutrient intake and availability (10, 15).

Inspired by growing evidence of extrahepatic ketone body production, we investigated whether immune cells might also synthesize their own ketone bodies through ketogenesis, as has been shown in intestinal stem cells (16) and adipocyte precursors (17). There is not strong evidence for hematopoietic HMGCS2 expression, the rate-limiting ketogenic enzyme, but some immune cells do express HMGCL, the terminal enzyme in ketogenesis, and humans with inborn genetic mutations in Hmgcl have lower HMGCL enzymatic activity in their peripheral blood leukocytes (18). However, ablation of HMGCL (3-hydroxy-3-methylglutaryl-CoA lyase) in macrophages and neutrophils only modestly improved glucose tolerance in aged mice, but worsened glucose tolerance in obese mice (19). In contrast, we have previously shown that innate immune cells are extremely sensitive to exogenous BHB, which inhibits activation of the NLRP3 inflammasome through metabolism-independent biophysical mechanisms in macrophages and neutrophils from both mice and humans (20, 21). Altogether, these data indicate that circulating BHB is more likely to modify innate immune inflammatory responses, rather than cell-intrinsic ketogenesis. We should note that data from human whole blood immune activation experiments do not fully reproduce or even contradict the NLRP3 inhibitory effects of BHB and suggest that host metabolic state and/or stimulation conditions may be important determinants of the anti-inflammatory effect (22, 23). Collectively, these data highlight that nonmetabolic actions of BHB are also critical for immune modulation and that translational potential between preclinical and human models requires additional investigation.

One perplexing question that arises from the prior work is why an immune cell like a macrophage might take up BHB if it is unable to metabolize BHB. Based on our nutrient surveillance hypothesis, one explanation for this is the newly described ability of BHB to form covalent attachment to lysine residues, a process known as β-hydroxybutyrylation (adduct referred to as Kbhb based on epigenetics nomenclature; 24). Immune cells are circulating in the presence of high BHB concentrations (∼0.1–2 mM in fed and healthy ketotic states, respectively) and can also be resident in ketogenic and ketolytic organs. Therefore, most immune cells will be intermittently exposed to high BHB concentrations that can increase even more during periods of infection or certain dietary manipulations. This is important to consider because the intracellular abundance of Kbhb increases as BHB concentrations rise (24, 25). As Kbhb is present on histones, this modification could conceivably alter immune cell differentiation programs, although additional tools and more investigation are required to understand the importance of histone Kbhb (26). Moreover, BHB can inhibit class I histone deacetylases (27) by reversing their enzymatic activity to actually catalyze Kbhb formation (28), and BHB oxidation increases acetyl-CoA availability, further contributing to the modulation of gene expression. Notably, only modest effects of BHB have been reported on transcriptional profiles in immune and nonimmune cells (29) so this axis of immune-metabolic regulation requires additional exploration. Remarkably, Kbhb-modified proteins can be found in nearly every intracellular compartment and the full consequences of this, and whether this could skew immune responses, are unknown (24, 25).

Beyond metabolic utilization or protein modification, extracellular BHB can also impact immune cells through stimulation of the mouse and human orthologues of G protein-coupled receptor GPR109A (the nicotinic acid receptor, also known as the hydroxy-carboxylic acid receptor 2, HCA2) at physiological ketotic concentrations (30, 31). The involvement of GPR109A/HCA2 signaling in inflammation, and specifically its role in immune cells, has been an area of investigation for over 10 years (32–34). Numerous studies have documented anti-inflammatory responses in GPR109A-stimulated macrophages (35, 36), whereas T cells lacking this receptor have increased ROS, increased apoptosis, and fail to cause graft versus host inflammation in a model of allotransplantation (37). The importance of BHB-stimulated GPR109A activation will also surely depend on the phase of the immune response when different cell types and different effector functions will be active.

This brings us to our final consideration in this short Perspective: how might ketone bodies synergize with other physiological cues to control immune function? We previously showed that γδ T cells in the lung and adipose tissue proliferate in response to ketogenic diet feeding and this was important for protection against influenza virus infection in mice (38, 39). However, increasing circulating BHB levels through 1,3-butanediol feeding was not sufficient to reproduce this effect, suggesting BHB does not directly stimulate protective γδ T cells and that additional aspects of ketogenic metabolism like increased fatty acids, other dietary adaptations, or even endocrine changes, may be important (40–42). One such possibility where synergy or costimulatory signals may be important for BHB’s immune modulation is adrenergic signaling, which regulates both metabolic and immune functions (43–45). β-adrenergic stimulation induces lipolysis and release of free fatty acids from adipose tissue that can be used for hepatic ketogenesis (46, 47). Both the Immgen (48) and ImmPRes (49) databases confirm that immune cells also express β-adrenergic receptors and their expression may be dynamic. Indeed certain immune functions like T-cell polarization (50), T-cell exhaustion (51), and bone marrow egress (52, 53) are controlled through adrenergic signaling. Given that these signals will coincide with ketogenesis during periods of fasting/starvation, a feature of infection-induced sickness behavior, but also during physiological responses to stress and exercise, it is reasonable to suspect that the effects of both adrenergic signaling and ketone bodies on immune cells could occur simultaneously, potentially leading to complementary or synergistic outcomes. How metabolite and other physiological signals converge on immune cells to control their function is an important aspect of physiology that is vastly understudied.

In summary, ketone bodies, and particularly BHB, represent a potent endogenous signal that can control immune function. Nonmetabolic functions of ketone bodies have stepped into a spotlight, and we believe there is much more to discover about this immune-metabolic axis.

GRANTS

The Goldberg Lab is funded, in part, by NIH grants R00AG058801 and DP2AI175641 and the Chan Zuckerberg Biohub.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.B.G. and E.G. drafted manuscript; edited and revised manuscript; approved final version of manuscript.