Keywords

Introduction

Aging associates with a systemic inflammatory environment (Franceschi et al. 2005). The circulating levels of pro-inflammatory cytokines, e.g., IL-1β, TNFα, and IL-6, are raised, and concentrations of some anti-inflammatory cytokine such as TGFβ are reduced.

Cells of the innate immune system, particularly neutrophils and monocyte/macrophages, also tend to be persistently activated albeit it at a low level. As these innate immune cells are major producers of oxygen- and nitrogen-centered free radicals, to aid bacterial killing, another important indicator of the pro-inflammatory environment is the accumulation of free radical-mediated oxidative damage to proteins, lipids, and DNA; these oxidation products tend to accumulate within cells and in the circulation with age (Haines et al. 2013). Oxidative damage arises when the rate of reactive oxygen species (ROS) production exceeds its removal and increased ROS can be detected in many cells. The NOX2 enzyme is the main source of the first radical, superoxide anion, to be produced by phagocytic cells. During aging, phagocytic cell superoxide production is increased (Ortega et al. 2000). In the adaptive immune arm, NOX4 is responsible for production of ROS and its primary species in hydrogen peroxide. A third important enzyme family for ROS production is the nitric oxide synthase (NOS) family; inducible NOS (iNOS) catalyzes the conversion of arginine to citrulline and nitric oxide that is used in bacterial killing by inflammatory macrophages. Both the mitochondrion and NOS and NOX isoforms are potential sources of aging-related excessive ROS production (Johannsen and Ravussin 2009; Birch-Machin 2006).

The distribution of cell subsets that secrete inflammatory mediators changes with aging. This is due, in part, to a disturbance in the balance between pro-inflammatory and anti-inflammatory cells within the immune system. With age, the pro-inflammatory cells are more prevalent, more resistant to apoptosis, and less efficient, so they continue to be inflammatory for long (Dunston and Griffiths 2010). In addition, the secretome from senescent cells, which can be found throughout an aging organism, may be enriched in pro-inflammatory cytokines. This is known as the senescence-associated secretory phenotype (SASP) and is under transcriptional and posttranscriptional control (Olivieri et al. 2013).

There is also an important immune-metabolic axis that contributes to a successful inflammation and the adaptive immune response. After receiving a trauma stimulus, leukocyte gene expression profile was shown to be altered by 80% (Xiao et al. 2011). Half of the transcriptome is increased in expression, while the other half is turned over. This genomic storm places an enormous metabolic demand on the leukocytes. The metabolic requirements of monocytes, macrophages, and T cell subsets will change according to whether the cells are proliferating (high demand) or differentiating (low but sustained demand) or quiescent memory cells (very low demand). The efficiency with which the specific energy needs can be met is a measure of the adaptability of the organism to the trauma. Successful metabolic switching is an important factor for immune cell function within both the adaptive (Torrao et al. 2014) and innate immune systems.

This chapter will focus on our current understanding of how less successful adaptation to stress by the innate immune system combines with altered nutrient sensing and nutrient supply to delay the resolution of inflammation in older adults. We will also consider how inefficient metabolic switching of immune cell phenotype plays a role in poor outcomes to sepsis in older adults. Finally, we will reflect on opportunities to shift the immune-metabolic axis to improve outcomes for older adults during sepsis.

Aging and Adaptation to Oxidative Stress

Adaptation is necessary to prevent the accumulation of damage that may emerge from an external stressor. Adaptive capacity is typically lost during aging. This ultimately leads to accumulation of damaged macromolecules and a buildup of senescent cells throughout an organism (Haines et al. 2013).

To achieve successful adaptation, cells must be able to meet the demand for a rapid energy supply; hence, adaptive responses to stress are tightly coupled to cellular metabolism. Glycolysis , the anaerobic metabolism of glucose, provides the most rapid source of glucose and is used by immune cells during the acute phase of inflammation. It effectively generates ATP to supply the cells’ energy needs with or without the presence of oxygen. However, one of the downsides to reliance on glycolytic metabolism is the buildup of toxic lactate. This can be a problem during disease of an acute inflammatory response is sustained (Brand et al. 2016).

Metabolic coupling of a cell’s energy requirement to its genomic adaptation to oxidative stress is an essential requirement to prevent the buildup of damaged molecules (Tweedie et al. 2011). The adaptive process that normally enables healthy tissues to repair is mediated by activation of transcription factors that are sensitive to the environment. Altered gene expression is designed to enable cells to recover from stress. Not surprisingly then, a central feature in the variation in successful human aging has been revealed through GWAS studies which have highlighted the role of cell cycle control (to enable repair) and regrowth genes (Cluett and Melzer 2009).

Damage to macromolecules is more common with age, either because there is less capacity to prevent damage or because the damage is not effectively removed. Oxidative damage has been associated with aging across a range of species. The simplest explanation for the increase in observed oxidative damage is that it arises from an imbalance between the rate of ROS formation and the rate of their removal. The key ROS and the enzymes responsible for their removal are shown in Table 1. In neutrophils, NOX2 produces ROS to kill bacteria and requires NADPH as a cofactor. NADPH is supplied, and its levels are sustained when the pentose phosphate pathway (PPP) is active. The PPP is a pathway that can be used by cells that are oversupplied with the sugar, glucose, as a fuel. If glucose is being heavily metabolized through glycolysis, the lactate levels rise (causing cramp in muscles). Glucose is shunted from glycolysis toward the PPP as lactate concentrations rise ensuring that NADPH supply is sufficient for NOX2 activity and that lactate levels are controlled.

Table 1 The major ROS produced by innate immune cells and the enzymes that regulate their removal
Full size table

In the majority of cells, including macrophages, the largest source for ROS production is the mitochondria . The mitochondrial superoxide combines spontaneously with nitric oxide to produce peroxynitrite that is important for bacterial killing by macrophages. The mitochondrial respiratory chain will not be very efficient after peroxynitrite has been formed. Peroxynitrite induces nitrosylation of the iron sulfur proteins in the electron transport chain and inhibits electron transfer (Mailloux et al. 2014). So it is important that macrophages involved in bacterial cell killing can use glycolysis for energy.

In order to maintain the redox state and minimize damage from ROS in conditions of bacterial killing and inflammation, there is concomitant upregulation of antioxidant enzymes. For example, the canonical TLR-4 signaling pathway activates NFκB which triggers the expression of a number of antioxidant genes including SOD2 to catalyze the removal of superoxide anion in the mitochondrion. It is also interesting to note that FOXO, a transcription factor family that mediates metabolic responses, also promotes the expression of SOD2 and catalase (Storz 2011). In this way, the increase in ROS as by-products of mitochondrial respiration will be mitigated by SOD2 which is co-regulated by the members of the FOXO family.

Genotoxic damage arising from oxidative stress was originally proposed by Harman in the 1950s as a cause of aging (Harman 1956). The accumulation of 8-oxodG in mitochondria, a commonly oxidized nucleotide, induces the senescent phenotype. Several studies have attempted to link an increase in free radical damage to macromolecules with the age-related decline in health. However, overexpression of the key antioxidant enzymes in animal models has, for the most part, shown the expected decrease in oxidative damage but without any improvement in life-span (Jang et al. 2009). Some have interpreted this to mean that ROS and oxidative damage are unimportant for aging. More recent evidence suggests that in fact, organismal health during aging is improved by enhancing adaptive capacity in the “3Rs” to reduce ROS formation, remove ROS with antioxidant enzymes, and repair oxidized macromolecules. Consistent with this hypothesis, younger tissues are more adaptive to oxidative stress than the same tissues from an older animal (Davies 2016). The same is true in humans up to the age of 70. Interestingly, healthy octogenarians, nonagenarians, and centenarians tend to have an improved capacity for adaptation to stress (Lattanzi et al. 2014).

More recent studies in model animals including the nematode worm and drosophila have consistently identified the daf-2 homologue, the insulin growth factor, working with the daf-16 homologue, FOXO , as a critical component of the aging process; together they control metabolic pathways and regulate the response to stress and resistance to bacterial pathogens (Henis-Korenblit et al. 2010). It is evident that an effective adaptive response requires the mobilization of energy reserves and/or the generation of ATP through intermediary metabolism. A good example of adaptation is during low oxygen tension. As oxygen tension is reduced, there is an increase in mitochondrial peroxynitrite leading to stabilization of HIF1α through S-nitrosation at Cys533 within the oxygen-dependent degradation domain of the transcription factor, HIF1α (Marinho et al. 2014). The chaperone, VHL, is unable to bind to the oxidized PHD domain, and VHL becomes ubiquitinated and degraded. Correspondingly, the half-life of HIF1α is extended, and it binds to the constitutively expressed HIF1b. Dimeric HIF1 then directs an increase in expression of genes that promote survival under hypoxia. Of course, during aging, there is also an increase in mitochondrial ROS because of accumulating mitochondrial DNA damage and reduced expression of the terminal electron transporter cytochrome c. There is no chemical distinction between the ROS generated by hypoxia and those produced during aging; hence, HIF1α is also likely to be stabilized by ODD nitrosylation during aging. The importance of redox signaling as a control mechanism for macrophage activation has recently been extensively reviewed by Brune et al. (2013).

Several genes that are upregulated by HIF1 form part of the glycolytic pathway. Glycolysis results in rapid production of ATP without the need for oxidative phosphorylation in the mitochondria. Thus, HIF1 ensures that the obvious cellular energy demand can be met, e.g., for ATP-dependent calcium ion gradients which are essential for survival. As oxygen tension is normalized, HIF1α is subsequently degraded, glycolytic enzyme production declines, and the energetic requirements are met by metabolically efficient mitochondria using oxidative phosphorylation (OX PHOS). However, there is no efficient switchback to OX PHOS in aged cells unless damaged mitochondria are removed by mitophagy and the low-level ROS production is eliminated (Palikaras et al. 2015).

ATP production during an adaptation phase to stress is the result of a more efficient mitochondrial respiratory chain. Expression of a number of OX PHOS genes is increased following FOXO activation so there is less leakage of ROS (Storz 2011). This is paralleled by increased expression of antioxidant enzymes, increased expression of DNA oxidation repair enzymes (e.g., OGG1), protein oxidation repair (MSR1), 26S proteasome activity (degrades oxidized proteins), and chaperone synthesis (supports mildly oxidized protein refolding). Many of the enzymes that are upregulated during adaptation have an essential requirement for ATP to catalyze removal of damaged molecules and dysfunctional organelles such as mitochondria, e.g., in a process known as mitophagy. In support of this link between energy metabolism and repair, the accumulation of DNA damage is known to suppress the insulin/insulin-like growth factor 1 (IGF-1) signaling pathway, possibly as an adaptive response to redirect energy toward DNA/cell repair pathways instead of proliferation after DNA damage accumulation (Niedernhofer et al. 2006).

Insulin Signaling, Metabolism, and Aging Immune Cells

Caloric restriction (CR) remains the only proven life-extending intervention; it can delay the onset of age-related diseases and metabolic disorders through reduction in the activity of the insulin signaling pathway. Defined as an intake of ~40% of ad libitum calories, CR is an effective experimental manipulation that increases life-span across a range of species from unicellular organisms such as yeast through to primates (Blagosklonny 2010). There has been significant interest in the underlying mechanisms of CR from pharma- and nutraceutical industries that are seeking to intervene in the aging process as a strategy to delay the onset of age-related disease without the need for strict adherence to dietary restriction.

Manipulation of the insulin/IGF-1 signaling pathway extends life-span in candida, C. elegans, drosophila, and mice. Indeed, IGF-1+/− animals are more resistant to oxidative stress, are phenotypically younger, are longer lived, and have improved stress resistance (Kondo et al. 2003). Variants in three genes that are involved in glucose, lipid, and ROS metabolism relate to longevity in humans (Barbieri et al. 2012). Specific alleles of IGF-1, IRS, and UCP2 in unrelated >600 Italians with an average age of 62 were linked to improved survival and health over a 6-year follow-up period. Healthy long-lived adults also tend to express the same alleles and have conserved insulin sensitivity. However, inhibition of mTOR and IGF-1 signaling will also impair activation of the PPAR transcription factor family and result in decreased mitochondrial function (Scarpulla 2011). This may reflect and impair the adaptive catabolic phase of an immune response. But cells of the immune system are not immortal postmitotic cells that receive aging stimulus over the lifetime – how do they really age? Because of the short-lived nature of some immune cells, it is useful to consider that bone marrow stem cell precursors of the myeloid lineage will receive environmental aging signals over the liver. Bone marrow-derived stem cells typically exhibit impaired proliferation, senescence, and chondrogenic response (Beane et al. 2014).

GWAS studies have highlighted the importance of the insulin signaling pathway for longevity in humans (Deelen et al. 2013; van der Spoel et al. 2015; Teumer et al. 2016). Early work by Franceschi et al., using a targeted approach, had shown a relationship between immune function decline and the insulin signaling pathway in centenarians (Franceschi et al. 2005). Despite this, few studies have attempted to study the relationship between IGF signaling and the aging immune system. One of these showed that IGFBP3 is increased in centenarians and correlates positively with the naïve CD8 T cell pool (Chen et al. 2010); the authors suggested that IGFBP3 may improve metabolism directly within the thymus to maintain the naïve CD8 T cell pool during later life. In support of this, in aging mice with telomere dysfunction, there was an enhanced requirement for glucose to maintain energy homeostasis (Missios et al. 2014). Oral glucose was shown to enhance IGF-1/mTOR-dependent mitochondrial biogenesis in hematopoietic cells and promote repair. There is an important and necessary trade-off between energy requirements for tissue repair, essential cellular bioenergetics, and insulin signaling during aging which is likely to be cell type specific and relate to expression quantitative trait loci.

Energy Supply and Inflammation

Metabolism refers to the pathways that provide energy from a variety of sources. The principal nutrient sources for energy come from carbohydrates and lipids, but during starvation and in times of energy crisis, proteins can become important energy suppliers. Organisms respond to energy availability by altering the balance between energy-producing catabolic processes and energy-consuming anabolic processes. During catabolism , carbohydrates are metabolized through glycolysis and the PPP in the cytosol and feed the citric acid cycle in the mitochondria to generate ATP and reducing equivalents. Fatty acids are shuttled into the mitochondria and undergo beta-oxidation to produce reducing equivalents, e.g., NADH and NADPH. In addition to the role of NADH and FADH2 in feeding oxidative phosphorylation by mitochondria to generate ATP, reducing equivalents such as NADPH are essential cofactors for ROS-generating NADPH oxidase and for antioxidant enzymes, e.g., glutathione reductase.

During starvation, proteins are degraded to provide energy. One of the pathways involved is glutaminolysis and this takes place in all proliferating cells. Glutaminolysis is particularly important for lymphocyte proliferation (Costa Rosa et al. 1992), and a study of rat lymphocytes and macrophages during aging showed that glycolysis was impaired but that glutaminolysis was intact (Costa Rosa et al. 1993).

The metabolic response to nutrients is tightly coordinated by nutrient- and energy-sensing signaling pathways. Metabolic regulation is achieved by a combination of substrate and product availability as well as the expression level of enzymes that are rate limiting in each pathway. The rate-limiting enzyme in glycolysis is phosphofructokinase (PFK). PFK is allosterically inhibited by ATP (the downstream product), but conversely it is activated to drive glycolysis when AMP is in excess.

The mammalian target of rapamycin (mTOR) is a key feature of the nutrient-sensing signaling network that controls cellular metabolism (Howell and Manning 2011). There are two mTOR family members which integrate nutrient and insulin signals, and we know most about TORC1 ; during conditions of nutrient depletion, TOR protein kinase activity within TORC1 is inhibited, but during nutrient-rich conditions, TORC1 is activated and promotes anabolic cell growth. Conversely, TORC2 responds to declining glutamine catabolites in order to restore metabolic homeostasis (Moloughney et al. 2016). mTOR is central to the process of adaptation to nutrients, and the benefits of caloric restriction against aging appear to be mediated by reductions in mTOR activity (Blagosklonny 2010).

The more mTOR is activated, the faster an organism ages. Why does mTOR inhibition delay aging? The mTOR inhibitor rapamycin has shed some light onto functions of the complex. mTORC1 activates protein synthesis which is required for cells to grow and proliferate. This is achieved by regulating the nutrient availability for mRNA translation to begin. mTOR is ubiquitous, has kinase activity, and catalyzes the phosphorylation and activation of the transcription factors that control mitochondrial biogenesis and lipid metabolism (Duvel et al. 2010; Howell and Manning 2011; Kim and Chen 2004). Maintenance of healthy mitochondria is important in healthy aging and requires careful coordination of biogenesis and mitophagy. This ensures that ROS production is minimized from leaky respiratory chain and that energy production is efficient because damaged mitochondria have been removed by mitophagy (Palikaras et al. 2015). Activation of glycolysis, the pentose phosphate pathway, and de novo lipid biosynthesis occur in response to TORC1 activation and are mediated via HIF1α and sterol regulatory element-binding proteins (Duvel et al. 2010).

Beyond the adaptive immune system, other cells such as adipocytes also play a role in inflammation. In adipocytes, TORC1 S6kinase upregulates PPARγ to direct lipid synthesis and adipocyte differentiation (Kim and Chen 2004). This is consistent with the observation that energy supplies are also stored into different organellar depots with age (Pararasa et al. 2015). For example, the composition of fatty acids in the liver of aging rats was not changed; however, in the spleen (but not other lymphoid organs), there was an increase in fatty acid unsaturation index with age (Guimaraes et al. 1995). An increase in the polyunsaturated fatty acid (PUFA) content may influence cellular function in several ways. First, PUFA are much more prone to oxidation by ROS, and the lipid oxidation products that are formed can be inflammatory (Awada et al. 2012). Second, some of the PUFA that are incorporated into membrane phospholipid bilayers affect membrane fluidity and can increase phagocytic activity in macrophages (Calder et al. 1990). Finally, PUFA are important substrates in the production of inflammatory eicosanoids (from arachidonic acid) and anti-inflammatory resolvins (from omega-3 fatty acids). Eicosanoids and resolvins are produced by nonimmune cells such as epithelial cells and can promote the inflammation and its resolution according to the PUFA supply available to the cell (Ali et al. 2016).

A further level of metabolic control that has implications for aging is mediated by the sirtuin (SIRT), family of seven nicotinamide adenine dinucleotide (NAD+)-dependent deacetylase members. SIRT share homology with the yeast silent information regulator 2 (Sir2) deacetylase protein; Sir2 expression increases yeast life-span by 30%, whereas Sir2 knockout yeast cells have a 50% reduction in life-span. Calorically restricted (glucose-deprived) yeast cells have increased life-span which associates with upregulation of the Sir2 protein.

SIRT1 is the most closely related of the mammalian SIRT family to yeast sir2. SIRT1 catalyzes the deacetylation of the histone proteins H3 and H4 at lysine residues K9 and K16, respectively, the NFκB subunit P65 (Yeung et al. 2004), the peroxisome proliferator gamma coactivator 1 alpha (PGC-1α) (Rodgers et al. 2005), and the FOXO proteins FOXO1 and FOXO3 (Brunet et al. 2004). SIRT1 has been identified as a suppressor of inflammation in several tissue types by inhibiting the NFκB pathway including the RelB noncanonical pathway. Nuclear p65 is acetylated by p300 to activate transcription of the inflammatory response genes. SIRT1 deacetylates and binds to the P65 subunit of the NFκB complex preventing phosphorylation and activation (Yeung et al. 2004).

PGC-1α is a transcriptional coactivator that interacts with several different transcription factors increasing glucose and fatty acid metabolism and mitochondrial biogenesis. The patterns of metabolic genes whose expression is regulated by SIRT1-dependent deacetylation are implicated in aging, and as expected, CR is ineffective in SIRT1 knockout mice (Mercken et al. 2014). In contrast, moderate overexpression of SIRT1 was not able to mimic the beneficial effects of intermittent fasting, and the improve glucose metabolism and insulin sensitivity occur via different pathways (Boutant et al. 2016).

As SIRT1 is NAD+ dependent, intracellular NAD+ availability is a deacetylation rate-limiting factor. NAD+ is also an important metabolic intermediate; its availability is increased during low nutrient availability due to increased flux of AMP through the NAD+ salvage pathway and low metabolic flux producing reducing equivalents. The NAD+ salvage pathway comprises of nicotinamide phosphoribosyltransferase (NAMPT) and nicotinamide mononucleotide adenylyltransferase (NMNAT). NAMPT converts nicotinamide (NAM), the product of NAD+-dependent deacetylation, to nicotinamide mononucleotide (NMN) which is subsequently converted to NAD+ by NMNAT. In contrast with this, during nutrient excess, increased glycolysis leads to more pyruvate formation, increased intracellular acetyl-CoA which is the substrate for acetylation (Haigis and Sinclair 2010). The dependence of SIRT1 on NAD+ directly links acetylation to the energy status of the cell.

SIRT3–5 are found in the mitochondria but have very different activities. SIRT3 regulates the acetylation of key metabolic enzymes such as acetyl-CoA synthetase, whereas SIRT5 has demalonylase and desuccinylase activity (Newman et al. 2012). SIRT4 has no detectable deacetylase activity; however, it appears to inhibit the activity of PPARα through nuclear SIRT1; SIRT4 appears to control NAD+ availability and uses NAD+ in retrograde signaling to the nucleus where energy sensing is critical for the regulation of metabolic gene expression (Laurent et al. 2013). Nuclear SIRT1 has been shown to activate RELB to differentially induce SIRT3 expression and increase mitochondrial biogenesis; this coordinates and supports bioenergetics adaptation and highlights the central regulatory node of SIRT1 (Liu et al. 2015).

By acting in response to cellular nutrient availability and subsequently altering such biological processes such as inflammatory response, fatty acid and glucose metabolism, mitochondrial biogenesis, DNA repair pathways, and cellular aging, the SIRT family is an important link between metabolism, inflammation, and aging.

Monocyte and Macrophage Metabolism: The Effect of Aging

The main function of monocytes lies in the recognition, removal, and repair of infected or damaged tissue. Over the last 10 years, monocytes have been described in three subsets: classical monocytes (CD14++CD16-) Mon1, intermediate (CD14++CD16+) Mon2, and nonclassical monocytes (CD14+CD16++) Mon3 (Ziegler-Heitbrock et al. 2010). Several studies have established that the ratio of these subsets is altered during chronic and acute inflammation including aging, sepsis, liver fibrosis, and rheumatoid arthritis with bias toward an increase of Mon2 monocytes relative to other subsets (Motwani and Gilroy 2015). It is gradually emerging that the remodeling of first-line defense cells, such as monocytes and granulocytes, is a complex phenomenon in the elderly. Some functions of innate immunity are depressed, while many other functions are upregulated. For instance, while chemotaxis and phagocytosis, as well as antigen presentation, are depressed in the elderly, cell activation and the secretion of inflammatory cytokines, such as IL-1, IL-6, and TNF, are markedly increased (De Martinis et al. 2004). In a comparative study between frail elderly and middle-aged and young adults, the frail elderly had significantly fewer Mon1 and Mon3, but not Mon2 monocytes. The function of monocytes from the frail elderly was also skewed toward inflammation; resting monocytes and after TLR-2 and TLR-4 agonist activation produced more TNF than from middle-aged adults (Verschoor et al. 2014). These observations highlight several unanswered questions: can monocyte subsets switch phenotype? Is phenotypic switching a feature of aging? Does altered subset frequency contribute to inflammaging ?

Macrophages are versatile, malleable cells that respond to environmental signals in tumorigenesis, wound healing, and the resolution of inflammation. The transcriptional basis for determining macrophage polarization is better described than for monocytes. The bacterial product, lipopolysaccharide (LPS), and host response to viral infection, IFNγ, are costimulatory signals that are transduced to activate NFκB and AP1 which then drive classic M1 macrophage polarization. The alternatively activated M2 state is triggered by IL-4 and IL-13 (typically activated by parasitic infection and damage signals) and regulated at the transcriptional level by STAT6 which can activate PGC-1α (Olefsky and Glass 2010).

The links between metabolism and the innate immune response were first drawn by Bustos and Sobrino; they identified that corticosteroids were a powerful anti-inflammatory but were also inhibitors of glycolysis (Bustos and Sobrino 1992). Glycolysis provides a very rapid (although not very efficient) supply of energy which is ideally suited to the function of the early recruited M1 monocyte/macrophage s. Figure 1 illustrates some of the key redox sensitive steps that promote glycolysis following TLR engagement. TLR-4 is coupled to NOX2 in macrophages, causing an increase in extracellular superoxide that can dismutate spontaneously to hydrogen peroxide outside the cell. Hydrogen peroxide is lipophilic and diffuses into the cytoplasm where it performs an important second messenger function. Hydrogen peroxide is on the one hand a powerful inhibitor of AMPK and on the other, through the inhibition of PTEN, is an activator of AKT/PKB. Hydrogen peroxide also promotes NFκB activation through activation of IKK and degradation of IkB. NFκB promotes the transcription of many pro-inflammatory genes including iNOS. Nitric oxide can diffuse throughout the cytoplasm and into mitochondria where it has the potential to form peroxynitrite with the superoxide anion.

Fig. 1
figure 1

M1 metabolic pathways. Pro-inflammatory stimuli drive glycolytic metabolism and inhibit oxidative phosphorylation. Red lines are inhibitory pathways and green lines indicate activation (Adapted from Griffiths et al. 2017)

Full size image

Downstream of activation of AKT/PKB by hydrogen peroxide is the activation of the aging node, mTOR (Fig. 1). mTOR promotes HIF1α expression. HIF1α is stabilized following NFκB-mediated iNOS expression, by the increase in peroxynitrite which nitrosylates HIF1α. FOXO activity is also increased by virtue of increased MAPK signaling that prevents the AKT-driven export of FOXO from the nucleus. FOXO cooperates together with HIF1α to target the upregulation of inflammatory genes, glycolytic enzymes, and the glucose transporter GLUT1. Correspondingly, mitochondrial respiration is inhibited through nitrosylation of the iron sulfur proteins of the electron transport chain. In an iterative loop, LPS increases succinate concentrations derived from glutamine-dependent anerplerosis to stabilize hypoxia-inducible factor 1α through succinylation (Tannahill et al. 2013). The sum of these events is that the M1 macrophage will be pro-inflammatory and glycolytic. Importantly, in older adults there is an unwanted circulating nutrient excess in the blood due to insulin resistance – this results in an increase in circulating saturated fatty acids and glucose, both of which will fuel the M1 phenotype.

As the inflammatory response progresses, ATP is consumed resulting in lactate, NAD+, and AMP buildup. These metabolites may provide sufficient of a nutrient-deficit signal to switch the metabolic phenotype of the macrophage and in turn its activity toward M2 macrophages (Fig. 2). The accumulating NAD+ is an essential cofactor for SIRT1 and SIRT6. SIRT6 is reported to deacetylate and inactivate glycolytic enzymes. This will shut down glycolysis. On the other hand, SIRT1 catalyzes the deacetylation and activation of PGC-1α and components of the electron transport chain, favoring mitochondrial biogenesis and an efficient electron transport chain. In addition, deacetylation of the p65 component of NFκB will inhibit its activity and reduce pro-inflammatory gene transcription (Yeung et al. 2004). Our own studies on substrate availability have shown that the profile of available fatty acids in the plasma can prime monocytes to polarize either M1 or M2 macrophages, with saturated fatty acids favoring the M1 and monounsaturated fatty acids favoring M2 (Pararasa et al. 2016). We also showed that the circulating fatty acid profile in older adults differs from younger people with higher levels of long-chain saturated fatty acids. There is a possibility that the change in fat storage with age predisposes to a change in the extracellular nutrient supply toward favoring the M1 phenotype and inflammaging. If this proves to be the case, it would be expected that in addition to the well-established anti-inflammatory activity of omega-3 fatty acids which are the precursors of prostaglandins, a switch to an anti-inflammaging phenotype could be achieved by increasing dietary and therefore plasma monounsaturated fatty acids. Interaction with the CD36 receptor would promote activation of the PPAR family and downstream expression of arginase andante-inflammatory cytokines.

Fig. 2
figure 2

M2 metabolic pathways. Anti-inflammatory stimuli drive oxidative phosphorylation and inhibit glycolytic metabolism. Red lines are inhibitory pathways and green lines indicate activation (Adapted from Griffiths et al. 2017)

Full size image

The timing on metabolic switching toward the resolution phase is critical – too soon – and there may not be effective removal of pathogen, and if the switch comes too late, a persistent systemic inflammation may ensue. This is the conundrum during sepsis which in older adults leads to mortality and morbidity.

Sepsis, Metabolism, and the Aging Immune System

Sepsis is a complex pathology that results from dysregulated host inflammatory responses to systemic bacterial infection (Hotchkiss et al. 2009). It is a leading cause of death in intensive care units worldwide, but the immunological and molecular basis of this syndrome are poorly understood. Sepsis is 12 times more prevalent following bacterial infection in older adults (>65 years) compared to younger adults (<65 years). In addition, the classical symptoms of systemic infection, such as fever, are often absent in older adults making early diagnosis and treatment less likely. Sepsis causes hyper-permeability of blood vessels, loss of blood pressure (attributed to excessive nitric oxide production), and organ failure due to anoxia. Given the high mortality rate from sepsis, it is imperative to understand more about its mechanism in older adults.

There has been a significant increase in the incidence of sepsis during the past two decades, with an increasing number of deaths occurring despite a decline in overall in-hospital mortality. The prevalence of sepsis in emergency department and ward in the UK was 15% with a significant number of patients developing severe sepsis. Mortality from sepsis ranges between 30% and 50% and is rising because of drug-resistant organisms, a growing elderly population, and an increased incidence of immunosuppression. The long-term effect of sepsis on survivors is significant; sepsis survivors show significant morbidity and mortality, with 5-year mortality rates of more than 70% (Deutschman and Tracey 2014; Iwashyna et al. 2010; Valdes-Ferrer et al. 2013). Although the underlying mechanisms remain unclear, high IL-6, HMGB1, and inflammatory monocytes have been implicated based on studies in human and mice studies (Valdes-Ferrer et al. 2013; Yende et al. 2008).

During sepsis, lactate levels rise and the ensuing acidosis poses a significant risk to patients. Patients also become hyperglycemic due to a stress response and impaired liver function. Interestingly, tight control of blood glucose has been shown to worsen outcome from sepsis. Can this be related to immune phenotype? There is supportive evidence from this from in vitro studies that looked at how the antidiabetic drug, metformin , influenced the macrophage response to a fungal infection; metformin inhibited mTOR and protection by innate immune system (Cheng et al. 2014).

Although the challenges of sepsis are not new, still significant control over sepsis has not been achieved. Sepsis is a leading cause of death in intensive care units (ICUs). Sepsis reflects a detrimental host response to infection in which bacteria or membrane fragments such as LPS act as potent activators of innate immune cells, including monocytes and macrophages.

During sepsis, a prolonged and intensive activation of the innate immune system causes a “cytokine storm” resulting physiological shock and multiple organ failure. The ensuing anti-inflammatory period – the compensatory anti-inflammatory response syndrome (CARS) – is now viewed as the primary clinical concern. This phase is linked with microangiopathy, coagulopathy, and catabolic metabolism leading to multi-organ failure. Indeed, the immune component of CARS, described as immune paralysis (Hotchkiss et al. 2009), might pose an equal if not a greater threat to the host than the initial “cytokine storm.”

There is frequently overlapping coexistence of inflammatory and immunosuppressive processes during sepsis. This is feasible and may reflect the time and naïve exposure of innate immune cells to stressors in the bone marrow. Indeed, immunosuppression might contribute to increasing mortality risk in most sepsis patients that prevents individuals from responding to secondary infections (Hotchkiss et al. 2009). Patients may experience oscillations between a hyper-inflammatory and hypo-inflammatory state largely based on cytokine profile (Motwani and Gilroy 2015). In a cross-sectional study, Park et al. performed differential counts in up to 16 leukocyte populations (Park et al. 2014). They showed that CD16− monocytes decreased in size and predicted poor outcome in severe sepsis. Another study has identified that platelet monocyte aggregate (PMA) numbers and associated IL-6 and IL-8 levels in plasma predicted outcome in older septic patients but not in those less than 65 years of age (Rondina and Garraud 2014). Both of these studies link monocyte phenotype to outcome.

Other than intervention with antibiotics and low-dose steroid therapy, in most cases, septic patient care is supportive. Mortality is rising because of an increasing number of drug-resistant organisms, a growing elderly population, and an increased incidence of immunosuppression. Recent analysis has revealed failure in over 100 drug trials for sepsis based on targeting cytokines, using receptor antagonists and antibodies to LPS (Marshall 2014). New targets are being sought that can be manipulated during the complex interactions of inflammatory mediators during the different phases of sepsis.

Considering the evidence for the relationship between monocyte phenotype to outcome, the blocking of HIF1α activity has been proposed as a therapeutic target for LPS-induced sepsis; it triggers expression of IRAKM, a negative modulator of TLR signaling, leading to an endotoxin-tolerant monocyte that is characterized by antibacterial and wound-healing properties (Shalova et al. 2015). A more subtle target than HIF1 has recently been proposed; pyruvate dehydrogenase (PDH) oxidizes pyruvate from glycolysis into citrate within the mitochondria. Citrate itself is a precursor for an antimicrobial metabolite known as itaconate and a substrate for lipogenesis. In spite of HIF1α activation, PDH flux is maintained and so may be an important target in sustained M1 activation (Meiser et al. 2016). SIRT1 is another potential target for modulating the outcomes in sepsis; nuclear SIRT1 has been shown to activate RELB to differentially induce SIRT3 expression and increase mitochondrial biogenesis (Liu et al. 2015) that may direct an anti-inflammatory response.

Conclusions and Future Directions

Metabolism is an important pathway for regulating the inflammatory response. According to substrate availability, the inflammatory response maybe sustained or resolved. Given the advances in metabolomics and stable isotope trace labeling to study nutrient and metabolic flux in humans in vivo, over the next few years, we are likely to gain a better understanding of the kinetic complexity of the human metabolome. During chronic inflammation and sepsis, fatty acids are mobilized from the adipose tissue and liver, and the profile of fatty acids that are released can predict distinct outcomes.

It is envisaged that a parenteral nutrition program could be developed on a patient by patient basis according to their inflammatory state (Taylor et al. 2016) that would at different times suppress M1 to deflect the cytokine storm with lactate toxicity and excessive nitric oxide production, then later to suppress M2 so enabling later response to persistent pathogens. Some of the key nodes of transition between pro- and anti-inflammatory phenotypes may serve as important targets such as SIRT1 (Liu et al. 2015). We are not there yet, and there remains much work to be done in understanding the time course of sepsis and its associated metabolome.

Cross-References