Obesity dysregulates the immune response to influenza infection and vaccination through metabolic and inflammatory mechanisms

Abstract

The COVID-19 pandemic demonstrates that obesity alone, independent of co-morbidities, is a significant risk factor for severe outcomes from infection. This mirrors a similar pattern with influenza infection; that is, obesity is a unique risk factor for increased morbidity and mortality. Therefore, it is critical to understand how obesity contributes to a reduced ability to respond to respiratory viral infections. Herein, we discuss human and animal studies with influenza infection and vaccination that show obesity impairs immunity. We cover several key mechanisms for the dysfunction. These include systemic and cellular level changes that dysregulate immune cell metabolism and function in addition to how obesity promotes deficiencies in metabolites that control the resolution of inflammation and infection. Finally, we discuss major gaps in knowledge, particularly as they pertain to diet and mechanisms, which will drive future efforts to improve outcomes in response to respiratory viral infections in an increasingly obese population.

INTRODUCTION

The increasing prevalence of obesity, which is over 20% in nearly all nations and over 40% in the U.S. alone, contributes toward a broad range of health complications. These complications include the potential for developing cardiovascular diseases, type 2 diabetes, liver steatosis, neurodegenerative disorders, and cancers. An often-overlooked complication of obesity is increased susceptibility to a variety of infections and poor responses to vaccinations. In this review, we first cover clinical, basic, and epidemiological evidence from the past decade to show that increased adiposity drives poor immunological responses to influenza infection and vaccination. Next, we discuss emerging mechanisms by which diet-induced obesity impairs innate and adaptive immunity. These notably include systemic changes in metabolic hormones, such as insulin and leptin, that impact immune cell metabolism and function. In addition, we focus on the metabolism of polyunsaturated fatty acids that are central in the resolution of inflammation. Within the context of obesity, we also discuss the parallels between influenza infection and SARS-CoV-2, the virus responsible for the current COVID-19 pandemic. Finally, we conclude by discussing major gaps in knowledge that will drive future mechanistic and clinical studies.

INFLUENZA OVERVIEW

Influenza virus and disease

There are a number of excellent review articles on influenza infection in humans, should the reader wish more in-depth information. Here, we summarize the viral structure and the pathology induced by infection.

Influenza viruses have been well-studied due to their contribution to significant morbidity and mortality worldwide. There are 4 types of influenza viruses: A, B, C and D, although only influenza A and B cause human disease. Influenza A strains are the causative agents of seasonal epidemics and pandemics. Influenza A (IAV) belongs to the Orthomyxovirade family, and consists of eight single-stranded, negative-sense viral RNA segments. These segments code for 10 proteins, including the two surface glycoproteins, hemagglutinin (HA) and neuraminidase (NA), which stimulate antibody responses. The HA and NA are the most abundant surface proteins. The matrix 2 (M2) is also a surface protein and functions as an ion channel. The internal proteins of the viral particle include M1, nucleoprotein (NP), non-structural proteins (NS) 1 and 2, polymerase acidic (PA), polymerase basic 1 (PB1), and 2 (PB2). The RNA-dependent RNA polymerase complex consists of PA, PB1 and PB2, which associates with each RNA strand coated with NP, forming the viral ribonucleoprotein complex (vRNPs). The virus particle itself is surrounded by a lipid membrane, derived from the infected host cell when the assembled virus exits the cell by budding (9).

Viral particles will bind to the sialic acid residue attached to a galactose of a glycan receptor on lung epithelial cells via the HA viral surface protein. Once endocytosed into the host cell, the virus is released from the endosome and vRNPs translocate to the nucleus of the infected cell where the RNA serves as a template for generating viral mRNAs, which are then exported to the cytoplasm for translation into viral proteins. Packaging of viral particles occurs under the host cell plasma membrane, ultimately resulting in new viral particles budding and becoming surrounded, or enveloped, with the host cell lipid membrane. The NA of the virus mediates cleavage of the HA from the sialic acid on the surface of the infected cell, allowing release of the virus and viral spread to other cells.

Infection with influenza virus is frequently characterized with a sudden onset of fever, myalgia, headache, malaise and cough. Depending on the viral strain, infection with influenza can cause death, particularly in high-risk groups. High-risk groups for severe outcomes from influenza infection include those over age 65, immunocompromised individuals, people with chronic heart or lung conditions or diabetes, pregnant women, individuals infected with HIV, and very young children. In addition to these well-established high-risk groups, the influenza pandemic of 2009 was the first time that people with obesity were listed as an independent high-risk group (66).

Epidemiology

Influenza A viruses are categorized based on the antigenic properties of their HA and NA surface proteins. At this time, there are 18 different HA subtypes and 11 different NA subtypes. Currently, two subtypes are circulating in human populations, H1N1 and H3N2. Seasonal influenza virus strains exhibit frequent point mutations in their HA and NA proteins, which leads to gradual antigenic changes. This process is termed antigenic drift, often necessitating new vaccine formulations each year in order to induce antibodies to recognize the changes. Previous year’s antibodies may not be as effective to antigenically drifted strains.

As influenza has a segmented genome, reassortment of gene segments can result in a major antigenic change, known as antigenic shift. Influenza pandemics arise from these antigenic shifts due to a lack of immunity to the new surface antigens. These antigenic shifts can occur when 2 different strains of influenza mix in a single host, thereby reassorting the segmented viral genome, resulting in a new, 3^rd viral strain containing viral genes from both infecting strains. This mixing can occur as pigs possess the receptors for both avian and human influenza strains, and therefore can be simultaneously infected with 2 different strains of influenza, thereby producing a novel influenza strain, which can then be spread to humans. This is what is thought to have occurred with the H2N2 pandemic in 1968. There have been 3 influenza pandemics in the 20^th century: 1918 caused by an H1N1 strain, 1957 caused by an H2N2 strain and 1968 caused by an H3N2 strain. The first influenza pandemic of the 21^st century was in 2009 (41), discussed in more detail below.

The 1918 influenza pandemic is notable for high lethality, in which one-third of the world’s population were infected, resulting in at least 500 million deaths. This pandemic had a unique feature in which younger (age 20–40) healthy individuals experienced a high mortality rate. The 1957 influenza pandemic, in which H2N2 influenza virus was now circulating, was less lethal than the 1918 pandemic, yet was still responsible for 1.1 million deaths worldwide. The 1968 pandemic saw the replacement of the H2N2 strain with an H3N2 strain, although only the hemagglutinin changed (from H2 to H3) as the N2 was the neuraminidase from the 1957 strain (103). This pandemic was responsible for an estimated 1–4 million deaths worldwide.

In 2009, the first influenza pandemic of the 21^st century occurred, which was designated (H1N1)pdm09. This H1N1 strain was very different from the currently circulating H1N1 viruses of the time. Of note, for the first time, obesity was recognized as an independent risk factor for hospitalization and increased mortality due to infection with (H1N1)pdm09. Since then, cohort studies have indicated that obesity is an independent risk factor for influenza strains beyond (H1N1)pdm09 (132), indicating that the increased susceptibility observed in individuals with obesity was not unique to the pandemic viral strain of 2009. The (H1N1)pdm09 pandemic was responsible for 151,700–575,400 deaths worldwide. Another aspect of this pandemic, in addition to the increased susceptibility of individuals with obesity, was that 80 percent of the deaths occurred in people younger than age 65. For this strain of influenza, an older age was protective, which was speculated to be due to pre-existing cross-reactive protection gained from exposure to pre-1950 H1N1 viruses (110).

Influenza vaccination is the most effective public intervention to prevent influenza infection, hospitalization, and mortality (83). However, vaccine effectiveness can vary considerably from year to year (10% to 60%), based on how well the match between circulating virus and vaccine strains. As will be discussed below, recent studies have demonstrated that individuals with obesity have increased risk of influenza vaccine failure.

OBESITY AND INFLUENZA

Influenza morbidity and mortality in obese populations

Several studies have pointed to the increased risk of poor outcomes from influenza infection in individuals with obesity. Concerning (H1N1)pdm09, a study from California reported more than half of the hospitalized adult patients had obesity, and 61% of the mortality occurred in this group (66). As mentioned earlier, although the (H1N1)pdm09 strain identified obesity for the first time as an independent risk factor for increased severity of influenza infection, individuals with obesity are also at higher risk from non-pandemic, seasonal strains of influenza. A number studies have demonstrated that obesity results in increased likelihood of hospitalization, a longer hospitalization, increased length of stay in the ICU, increased need for mechanical ventilation and increased mortality from infection with influenza (28). Two studies using systematic review and meta-analysis also demonstrated that obesity increased the risk for severity and mortality (93).

What is the mechanism by which obesity increases the risk of a poor outcome from infection with influenza virus? We suggest that impairment of the immune response, due to metabolic and inflammatory alterations is the cause for the increased risk for poor outcome to influenza virus infection in individuals with obesity. Below, and in Figures 1 and 2, we detail the immune dysfunction to influenza virus and/or vaccination that occurs with obesity (Figure 1), followed by a discussion of the metabolic and inflammatory dysfunction that is likely driving the impaired immunity (Figure 2). We conclude with future directions that will drive the field forward.

Figure 1. The response to influenza infection/vaccination in the obese host.

Red circles demonstrate a decrease (minus) or increase (plus) response to infection under conditions of obesity. IFN, interferon; T_eff, T effector cell; NK, natural killer. Figure adapted from reference 135. [**Note to Annual Reviews: I am an author of this article; the publisher grants authors the right to use their own figures without permission**]

Figure 2: Potential mechanisms for increased influenza morbidity and mortality and decreased influenza vaccine effectiveness in a host with obesity.

These mechanisms include dysregulation driven by leptin and insulin, impaired immunometabolism, and chronic uncontrolled inflammation.

Effects of obesity on innate immunity to influenza infection

The innate immune system is both a critically important first line of defense against influenza virus infection and an important step in the activation of adaptive immune cells to influenza, which will be discussed in the next section (94). Pathogen recognition receptors to the influenza virus are found on multiple innate immune cells leading to the involvement of these cells in the early immune response to infection. Innate immune cells shown to have a role in the immune response to influenza include natural killer (NK) cells, innate lymphoid cells, neutrophils, macrophages, monocytes, and dendritic cells (56; 94). These cells respond to influenza infection through actions such as phagocytosis, antigen presentation, release of granular contents, and production and secretion of key cytokines that limit viral replication and activate other immune cells (56).

Obesity is associated with chronic inflammation characterized by changes in both innate and adaptive immune cells as well as tissue-localized and circulating cytokines. Therefore, at baseline, prior to viral infection, the innate immune system is already altered in obesity, and thus, it is not surprising that the innate immune cell response to viral infection is likewise aberrant. For example, obese mice have a higher proportion of pro-inflammatory M1-like macrophages than lean mice (12; 33). Following influenza infection, macrophage migration to the lungs has been shown to be impaired in obese animals. Moreover, lung resident alveolar macrophages in obese mice have decreased expression of type 1 IFN receptor and IFN gene expression than lung resident alveolar macrophages from lean mice (44).

Dendritic cell number and function is also altered in the response to influenza in the context of obesity. In human studies, circulating dendritic cells were significantly decreased in individuals with obesity compared to normal weight individuals; moreover, dendritic cells generated from individuals with obesity had decreased functional response following activation in vitro (81). In animal studies, obese mice had decreased recruitment of mononuclear cells to the lung, decreased dendritic cell number, and impaired antigen presentation following influenza infection compared to infected lean mice (112). NK cells are innate lymphocytes that can kill infected cells without prior activation and are early producers of key pro-inflammatory cytokines in the response to viral infection. Multiple studies have examined the effect of obesity on NK cells and in general the results have been conflicting. However, in the context of influenza infection, NK cells isolated from obese mice infected with influenza virus had reduced cytotoxicity as measured by cytotoxicity assay compared to NK cells isolated from lean infected mice (111). Altogether, baseline changes in innate cell number and function in obesity, paired with deficiencies in the functional response to influenza infection, contribute to the impaired immune response to viral infection in obesity leading to increased morbidity and mortality.

Effects of obesity on influenza specific adaptive immunity

Obesity influences both B cell and T cell responses to the influenza virus. In naïve individuals, if viruses have escaped the innate immune response, a peak viral load will occur in 24–72 hours post infection, which is highly dependent on the dose of the exposure. Seven-to-ten days later, antibody titers peak in response to the infection. In a naïve individual, the generated antibody repertoire is not as highly specific as it would be for those experiencing a second or third infection, or following multiple vaccinations.

There are few studies that examine the effects of obesity on humoral responses to influenza infection. Studies using diet-induced obese mice have demonstrated lower antibody titers to influenza infection when compared with lean mice, which was associated with a decrease in B cells in the bone marrow of the obese mice (54). Cho et al. demonstrated that diet-induced obese mice, compared with lean animals, had reduced production of influenza-specific IgG antibody (measured by ELISA) following priming with an intramuscular injection of inactivated influenza virus (17). Another study demonstrated that late memory, or double negative B cells are increased in the blood of individuals with obesity, and that these cells do not proliferate, nor do they make antibodies to influenza antigens, although they secrete autoimmune antibodies (31). In a vaccination model, Park et al. demonstrated reduced levels of neutralizing antibody titers in obese mice compared with lean mice (90). An adjuvanted influenza vaccine in obese mice also failed to prevent the antibody deficiency (48), suggesting that obesity can significantly alter immune function.

In human studies, at 30 days post influenza vaccination, adults with obesity have a similar antibody titer compared to healthy weight and overweight individuals. However, at one year post vaccination, there is a steeper decline in the antibody response in individuals with obesity compared with healthy weight or overweight (85). Interestingly, this effect is not specific to influenza, as studies with tetanus and hepatitis vaccines also demonstrate impaired antibody production and maintenance in people with obesity (26; 86).

T cell responses are also impaired in both obese mice and humans with obesity (Figure 1). Using the diet-induced obese mouse model, our laboratory published a number of papers demonstrating that obesity impairs both the primary and the memory T cell response to influenza infection (49; 50; 73; 98; 111). Memory T cells generated in the primary response can fully protect lean mice from a lethal challenge with a second strain of influenza. Obese mice, however, are less protected, with 50–60% of the mice dying from the challenge, along with increased viral titer, increased lung inflammation, and damage and alterations in the lung metabolome. Importantly, this increase in mortality is due to decreased memory CD8+ T cell generation, maintenance, and function (49; 50). We have also demonstrated that the increase in mortality is due to the obesity itself, not the diet, as hyperphagic obese mice that consume excess low-fat chow die from influenza infection at the same rate as high-fat fed obese mice (73). Critically important resident memory influenza specific CD4+ and CD8+ T cells were found to be reduced in the lungs of obese mice (98). Park et al. found low level of influenza specific effector memory CD8+ T cells following vaccination with inactivated influenza virus (90).

In humans, in a study of influenza-vaccinated 28 lean and 27 adults with obesity (matched for age, sex, race), following in vitro stimulation with influenza virus, the frequency of CD4⁺ and CD8⁺ T cells expressing activation markers CD69, CD28, and CD40L as well as the functional markers IFNγ and granzyme B, were all significantly lower in adults with obesity (85). This was not associated with impairments in activation or function of dendritic cells. In addition, the total number of CD4⁺ and CD8⁺ T cells with and without influenza stimulation were not different between lean adults and adults with obesity. Vγ9Vδ2 (γδ) T cells were found to be reduced in the peripheral blood of adults with obesity and were much less responsive to an in vitro challenge with influenza virus, with both the number of cells producing both fewer cells producing IFNγ, and the amount of IFNγ per cell (19).

In addition to the documented impaired vaccine responses in adults with obesity, we also reported that, compared with influenza-vaccinated healthy weight individuals, influenza vaccinated individuals with obesity are 2X more likely to develop laboratory confirmed or influenza-like illness (77), despite a robust serological response equivalent to healthy weight individuals.

POTENTIAL MECHANISMS BY WHICH OBESITY DYSREGULATES IMMUNITY

Systemic changes in metabolic hormones that influence immune cell function

Obesity is associated with changes in multiple nutritional hormones that regulate appetite, metabolism, and body fat distribution. Among these nutritionally regulated hormones, leptin and insulin are notable for also having well-described roles in influencing immune cell function and infection response and will be discussed here (Figure 2).

Leptin is primarily secreted by adipocytes in proportion to adipose tissue mass and is therefore increased in individuals with obesity. Leptin is well-known for its role in regulating appetite and systemic metabolism through signaling via leptin receptors expressed in the hypothalamus; however, leptin is also a key regulator of immunity. This was first discovered in individuals born with rare mutations in the genes expressing leptin or leptin receptor (74). Individuals with congenital leptin deficiency have severe obesity due to impaired signaling in the hypothalamus leading to dysregulated appetite; however, these individuals were also observed to have increased risk of intracellular infections and striking abnormalities in T cell number and function (84). Both the metabolic and immune abnormalities observed in individuals with congenital leptin deficiency are reversed by treatment with systemic leptin (27). These immune cell deficiencies observed in humans with mutations in leptin or leptin receptor have been studied thoroughly in the mouse models of leptin deficiency (ob/ob mouse) and leptin receptor deficiency (db/db mouse) (13; 69). From both mouse and human studies, we now know that multiple innate and adaptive immune cells are leptin responsive (51).

Although multiple immune cell types are responsive to leptin signaling, the role of leptin in modulating immune cell number and function is best described in T cells, which are highly leptin responsive. Leptin is required for normal T cell development in the thymus (34; 53; 102) as well as normal peripheral T cell proliferation (65). Leptin also promotes the production of pro-inflammatory cytokines by CD4+ Th1 and Th17 cells, as well as promoting the differentiation into Th17 cells (34; 65). One mechanism by which leptin promotes Th17 cell differentiation may be through increased expression of the transcription factor RAR-related orphan receptor gamma (RORγt), a critical transcription factor for Th17 differentiation (51; 131). Another key mechanism by which leptin regulates CD4+ T cells is through regulation of cellular metabolism. Promotion of CD4+ T cell metabolism via leptin signaling is required for the proliferation and function of activated CD4+ T cells (102). Specifically, leptin is required for activated CD4+ T cells to increase glucose uptake and metabolism and thereby fuel effector T cell function (102), but this is only true of pro-inflammatory Th1 and Th17 cell subsets and is not observed in Treg cells (34). One mechanism for this is through the ability of leptin to promote expression of HIF-1alpha, a critical regulator of Th17 versus Treg cell differentiation and a key promotor of T cell glucose metabolism (34). The role of T cell metabolism in influencing T cell function in the context of obesity will be discussed in greater detail below.

Leptin also influences B cell number, reduces B cell apoptosis and promotes B cell cycling (55). Moreover, leptin alters the function of B cells by increasing inflammatory cytokine production and reducing class switching and IgG production (30). Lastly, leptin has effects on innate immune cells including macrophages and monocytes, mast cells, dendritic cells, neutrophils, basophils, eosinophils, NK cells, and innate lymphoid cells, primarily in a pro-inflammatory manner, but with distinct effects on each innate immune cell type (51).

These leptin-driven changes in immune cell number and function have implications in obesity-associated diseases, including infection. This may be partly due to the fact that obesity is associated with resistance to leptin signaling at the level of the hypothalamus. Whether or not immune cells likewise become leptin resistant in obesity has not been determined; however, altered levels of leptin and/or leptin action in obesity could certainly affect immune outcomes. Following the observations that individuals with obesity had increased morbidity and mortality in response to H1N1 influenza, leptin receptor deficient db/db mice were infected with H1N1 influenza A virus to test the hypothesis that impaired response to leptin receptor signaling might explain poor outcomes to infection in obesity (95). The db/db obese mice had decreased clearance of influenza virus from lungs, compared to normal weight mice, as well as increased mortality to infection. However, when this was tested in tissue specific leptin receptor knockout mice, including lung epithelial or alveolar macrophage specific knockout of leptin receptor, those mice had better viral clearance and improved survival compared to global knockout mice (95). These results indicate that the response to leptin of other immune cells, and perhaps particularly adaptive immune cells such as T cells and B cells, may be critical for clearance of influenza virus in this mouse model of obesity.

Another key metabolic hormone that is altered in obesity, with known effects on immune cells, is insulin. Insulin is produced by beta cells from pancreatic islets in response to elevated blood glucose concentration. Insulin is best known for its ability to promote glucose uptake into metabolic cells and tissues and increase the rate of glycolysis and glycogen synthesis. Insulin also plays key roles in protein and lipid metabolism, increasing muscle mass and lipogenesis and decreasing lipolysis, making insulin a primarily anabolic hormone. In addition to its effects on metabolic tissues, insulin can also influence immune cell activation and function (121). Indeed, insulin receptors have long been reported to be expressed on the surface of activated immune cells (42); however, these early studies only measured the binding of radiolabeled insulin onto the surface of the activated lymphocytes, and not specifically insulin receptor expression. Following those early observations, many subsequent studies pointed to insulin as an anti-inflammatory agent through its ability to suppress pro-inflammatory transcription factors and cytokines, decrease hyperglycemia toxicity and free fatty acid mediated inflammation, and alleviate toll-like receptor-mediated inflammatory injury (114).

More recent evidence points to a role for insulin in promoting immune cell function in a cell-specific manner. One study using an inducible knockdown of the insulin receptor gene in rats found impaired T cell function and decreased CD4+ T cell glucose uptake and glycolysis (29). This study was limited, however, by use of a global knockout and the concurrent hyperglycemia that occurred. In a subsequent study by another group, using a T cell specific insulin receptor knock out mouse, insulin receptor signaling was found to have a critical role in promoting Teff cell proliferation and cytokine production (118). In the same study, mice with T cell deficiency of insulin receptor infected with H1N1 influenza virus had decreased numbers of influenza specific CD4+ T cells within the lungs and draining lymph nodes at nine days post infection, decreased CD4+ T cell cytokine production, and decreased CD8+ T cell response, compared to wildtype mice infected with influenza (118), demonstrating that insulin receptor signaling is critical for both CD4+ and CD8+ T cell response to the respiratory virus. Moreover, insulin receptor signaling was found to be critical for both glucose uptake and amino acid transport in activated T cells (118). This finding that insulin promotes T cell glucose uptake is not altogether surprising given the role of insulin signaling on other metabolic cells, as well as earlier reports that insulin increased glucose uptake in thymocytes and splenocytes from rats with metabolic syndrome (10), but it still highlights the role for insulin in regulating immune cell metabolism and function.

Systemic insulin levels increase in obesity in response to insulin resistance and beta cell dysfunction. Similar to leptin, it is unclear if the insulin resistance observed in metabolic tissues in obesity is also observed in immune cell signaling. Moreover, obesity-associated inflammation itself leads to insulin resistance within adipose tissue and systemically, as inflammation inhibits the insulin signaling activity in adipocytes and hepatocytes through several mechanism involving decreased insulin receptor substrate-1 signaling, inhibition of PPARγ function, and increased free fatty acid availability (128). Altogether, changes in insulin levels and insulin action in the context of obesity can lead to altered immune response to respiratory viral infections.

Changes in immune cell metabolism in obesity

Another potential mechanism for impaired immune cell response to influenza in individuals with obesity is metabolic dysfunction. A number of studies have established that immune cell function and metabolism are closely linked, and changes in immune cell metabolism can result in altered differentiation and function. Here we will discuss the details of those findings in the context of lymphocytes. Activated T cells have a very high metabolic demand that is essential to fuel growth, proliferation, and function (35; 70; 122). To meet this increased energy demand, the metabolic profile of a T cell is altered upon activation with increased expression of the glucose transporter Glut1, increased glucose uptake and metabolism (glycolysis), as well as increased glutamine metabolism to promote growth, survival, and effector function (47; 67; 68), although there are some nuances between effector subtypes. In contrast, naïve T cells, regulatory T cells (Treg), and memory T cells (Tmem) utilize fatty acid oxidation to fuel suppressive function and immune surveillance (72; 106). Multiple studies have shown that changes to T cell metabolism can influence differentiation and function (72; 82; 120), but it has only recently been shown that obesity induces changes in T cell metabolism.

The effects of obesity on T cell metabolism have been studied in the mouse model. Activated CD4⁺ T cells from high fat diet-induced obese mice had an increase in glucose uptake and oxygen consumption rate (OCR; a measure of mitochondrial oxidation), but no change in fatty acid uptake or extracellular acidification rate (ECAR, a surrogate of lactate production), resulting in an increase in the OCR:ECAR ratio, and demonstrating a switch in the metabolic phenotype of activated T cells from obese mice compared to lean mice. This represents a unique cell metabolic phenotype of glucose oxidation that is not utilized by naive, memory, or activated T cells from lean animals and may mechanistically explain obesity-associated T cell dysfunction (2). Similar changes in T cell metabolism were observed in obese mice following activation in vivo in a mouse model of influenza reinfection. Using influenza A/HKx31 (X31) as the primary infection to induce T cell memory, followed by a secondary challenge infection with influenza A/Puerto Rico/8/34 (PR8), obese mice had impaired cross-reactive memory T cell responses, as discussed above, as well as altered T cell metabolism (98). Specifically, CD4⁺ and CD8⁺ T cells isolated from obese mice 7 days following influenza reinfection displayed increased OCR and OCR:ECAR ratio compared to T cells isolated from lean reinfected mice (98). In a separate study, metabolic impairment was observed in CD8⁺ T cells from the lungs of obese mice infected with a sublethal dose of influenza virus (37). Lung CD8⁺ T cells from obese mice had suppressed OCR and ECAR at 10 days post infection, compared to CD8+ T cells from lean controls, although there was no difference in glucose or fatty acid uptake in these cells (37).

Further studies demonstrated that targeting T cell oxidative metabolism could improve influenza survival in mice. These experiments used the well-described antidiabetic drug metformin, which is known to target mitochondrial oxidation, presumably through inhibition of complex I of the electron transport chain (99). Preliminary studies first showed that wildtype CD4⁺ T cells from lean C57BL/6 mice activated in the presence of metformin in vitro had a decrease in OCR and OCR:ECAR, as well as changes in activation marker and cytokine expression, demonstrating a direct effect of metformin on T cell metabolism and function (2). Metformin was also able to decrease T cell oxidation when administered to obese mice systemically (2). Diet-induced obese mice were then treated with metformin in the drinking water for 6 weeks and compared to lean controls or mice undergoing weight loss through diet switch; treatment with metformin, but not weight loss, was found to both decrease T cell oxidation in vivo and improve survival following an influenza challenge (2). Altogether, these results show that T cells from obese mice have an altered metabolic profile characterized by increased glucose oxidation, which can be targeted using metabolic drugs to improve survival against influenza infection.

In a 2017 human study, influenza vaccine responses were compared in type 2 diabetes patients taking metformin or not (25). Metformin was found to reverse the impaired B cell function and vaccine responses seen in obesity and type 2 diabetes. Moreover, B cells isolated from untreated patients with type 2 diabetes were activated in culture in the presence or absence of metformin, and the B cells treated with metformin had reduced inflammation and increased antibody responses. This was the first study to show an effect of metformin on B cells.

Obesity and inflammation

Obesity is well established to drive a state of chronic inflammation, often termed ‘meta-inflammation’. A range of factors contribute toward the development and progression of chronic inflammation in obesity (58). Notably, white adipose tissue becomes inflamed through several overlapping mechanisms with increasing adiposity. These mechanisms include, but are not limited to, secretion of pro-inflammatory cytokines by adipocytes, increased oxidative stress and hypertrophy of adipocytes. Furthermore, there is infiltration of pro-inflammatory macrophages (i.e. typically termed M1 macrophages) in addition to many other immune cell populations such as invariant natural killer cells, Th17 cells, CD4⁺ T cells, CD8⁺ T cells, and CD19⁺ B cells that contribute toward the inflammatory tone (7; 15; 24; 79; 80; 124–126; 133). Inflammation is not limited to the adipose tissue as the liver, gut, central nervous system, skeletal muscle, and pancreas also contribute toward systemic meta-inflammation in obesity (57; 101; 129).

There are likely many factors related to meta-inflammation that contribute toward the dysregulated immune response to influenza. These include nutritionally regulated hormones such as leptin and adiponectin and pro-inflammatory cytokines such as TNF, IL-6, and CRP. We have already discussed the role of leptin in regulating immune cell function in obesity above, and the role of nutritionally regulated cytokines have been recently reviewed (3). Here, we turn our attention to one set of key lipid metabolites that control the resolution of inflammation and infection.

Resolution of inflammation is impaired in obesity and may contribute toward poor outcomes upon influenza infection or vaccination

At a molecular level, the initiation and resolution of inflammation, including infectious inflammation, are highly orchestrated events (105). The initiation of inflammation is triggered, in part through the biosynthesis of metabolites synthesized from the n-6 polyunsaturated fatty acid (PUFA) known as arachidonic acid. Arachidonic acid, upon cleavage by phospholipase A2 from membrane phospholipids, serves as a substrate for enzymes such as cyclooxygenases that generate metabolites of inflammation initiation. Examples of these molecules include prostaglandins and leukotrienes, which are essential for triggering molecular pathways that lead to vasodilation, recruitment of specific immune cells to the site of injury, clearance of debris, and blood clotting.

There is an increasing appreciation that the failure to resolve inflammation contributes toward complications of obesity. The resolution phase of inflammation occurs upon ‘lipid class switching’, in which the biosynthesis of arachidonic acid-derived mediators is decreased with a parallel increase in the biosynthesis of metabolites referred to as specialized pro-resolving mediators (SPM) (104). SPMs known as resolvins, protectins and maresins are enzymatically synthesized from the conditionally essential long chain n-3 polyunsaturated fatty acids eicosapentaenoic (EPA) and docosahexaenoic (DHA) acids. SPMs can also be synthesized from arachidonic acid such as the lipoxin family. SPMs bind specific G-protein coupled receptors and thereby initiate signaling pathways which promote the resolution of inflammation and infection (105).

It is established that inflammation promoting metabolites, such as prostaglandin E2 or leukotriene B4, are elevated in obesity (76). To exemplify, leukotriene B4 is a potent chemoattractant that drives key immune cell populations into white adipose tissue and inhibiting this mechanism genetically or pharmacologically improves inflammatory tone and thereby metabolic status (60). Data are now emerging to suggest that obesity is not just increasing inflammatory metabolites, but may also be driving a state of SPM deficiency (88). This deficiency may be contributing toward an inability to resolve inflammation.

SPM deficiencies appear to be highly tissue specific and are found in both obese mice and in humans with obesity (summarized in Table 1) (5; 18; 64; 78; 87; 91; 117). For instance, mass spectrometry studies reveal that DHA-derived metabolites of the SPM family are decreased in the spleens of obese male, but not female mice, relative to lean controls (20). Similar results, albeit modest, were also found in the bone marrow (20). The spleen and bone marrow are relevant for infectious responses as they are enriched in a wide range of immune cells and therefore these tissues are vital for the response to influenza and other viruses. Similarly, white adipose tissue and liver, which also accumulate various immune populations with increasing adiposity, are deficient in DHA- and EPA-derived SPMs in murine obesity models of type 2 diabetes (18; 78; 87; 117).

Table 1:

Summary of studies to show that obesity and type 2 diabetes are associated with a deficiency in the concentration of SPMs and their metabolic intermediates.

Model System	Outcome	References
-Adipose tissue, spleen, bone marrow, and liver analyses of C57BL/6J mice. -Similar results are observed with adipose tissue analyses using genetic models of obesity, compared to controls	Several SPMs and/or their precursors derived from eicosapentaenoic acid and/or docosahexaenoic acid are decreased with diet-induced obesity or models of type 2 diabetes	(20,54,78)
-Adipose tissue macrophages from obese C57BL/6J mice	Most, but not all, docosahexaenoic acid-derived SPMs of the resolvin family were decreased with obesity	(5)
-Hypothalamus analyses of Swiss mice	Hypothalamic docosahexaenoic acid-derived SPM is reduced in obese mice	(91)
-Plasma analyses of humans with type 2 diabetes compared to lack of diabetes	Docosahexaenoic acid derived SPM is decreased compared to controls	(71)
-Analyses of adipose tissue isolated from humans with obesity compared to lean controls	Several SPMs and their precursors are lowered with obesity	(117)
-Analyses of plasma and leukocyte analyses of subjects with morbid obesity compared to lean controls	Several SPMs and/or their precursors derived from arachidonic acid, eicosapentaenoic acid, and docosahexaenoic acid are decreased in subjects with morbid obesity	(64)
Serum analysis of individuals with obesity compared to overweight or lean subjects	Serum 14-hydroxydocosahexaenoic acid levels are decreased with increased adiposity compared to lean controls	(59)

SPMs are also deficient in humans with obesity. A key study by the Claria laboratory showed that humans with morbid obesity, compared to lean controls, showed a significant reduction in the concentration of the SPM precursor known as 17-hydroxydocosahexaenoic acid (17-HDHA) (64). This SPM precursor, synthesized in response to DHA serving as a substrate for 15-lipoxygenase, is notable as it appears to have a role in controlling antibody levels in the context of influenza infection and vaccination (96; 97). In another study, 14-hydroxydocosahexaenoic acid was decreased in the serum of individuals with obesity compared to lean subjects (59). 14-HDHA is of interest as it boosts antibody levels upon influenza infection (54). Finally, the DHA-derived SPM known as maresin 1 was decreased in subjects with type 2 diabetes compared to controls (71). Taken together, obesity appears to be driving a state of SPM deficiency, and replenishing these SPMs through dietary or pharmacological approaches, may be a therapeutic strategy for improving infectious outcomes.

SPMs regulate viral replication and antibody production upon influenza virus infection

There is increasing evidence that SPM administration improves outcomes related to viral, bacterial, and parasitic infections. Notably, the DHA biosynthetic pathway for the synthesis of SPMs appears to be a lead for improving outcomes related to influenza infection (Figure 3). A seminal study established that mice infected with a lethal dose of influenza H5N1 were largely protected from death in response to administration of the DHA-derived SPM protectin DX (75). A combination of cell culture and animal studies showed that protectin DX, unlike other SPMs, inhibited H5N1 replication by preventing viral transcripts from being exported through a specific transporter (75). A notable aspect of this study was the protectin DX enhanced survival even under conditions where canonical anti-viral drugs failed to provide any protection.

Figure 3: Biosynthetic pathway by which specialized pro-resolving mediators (SPM) are synthesized from docosahexaenoic acid (DHA).

DHA serves as a substrate for differing enzymes such as the 12- and 15- lipoxygenase (LOX) to generate downstream metabolites known as maresins, protectins, and resolvins. Key SPM pathway intermediates and precursors are 17-HDHA and 14-HDHA. In particular, 17-HDHA is an SPM precursor for downstream resolvins. These metabolites are a lead for improving outcomes related to influenza infection and potentially in response to SARS-CoV-2.

A series of studies show that metabolites of the DHA-derived SPM family or their parent compounds improve humoral immunity upon influenza infection or vaccination. Initially, it was demonstrated that dietary DHA, but not EPA, increased antibody levels in obese mice, particularly in response to T-independent antigens (38; 39; 54; 115; 116). Subsequently, it was reported that in vitro treatment of CD19⁺ B cells with 17-HDHA increased antibody levels relative to a vehicle control (97). Follow up work from the same laboratory demonstrated that 17-HDHA improved influenza-specific antibody levels and survival upon H1N1 influenza vaccination and infection in lean mice (96). Mechanistically, 17-HDHA increases the abundance of CD138⁺ long-lived antibody secreting cells by increasing the expression of key transcription factors involved in B cell differentiation such as BLIMP-1 (96). Another study demonstrated that dietary DHA could also increase influenza-specific antibody levels in obese mice, which was through the biosynthesis of downstream metabolites 17-HDHA and 14-HDHA (54). Again, the mechanism of action was through an increased abundance of bone marrow CD138⁺ cells. Furthermore, there is also evidence that metabolites of the DHA family can increase antibody levels in mouse models of vaccination (54). 17-HDHA was also shown to be an effective adjuvant to improve survival in an influenza vaccination model (96). However, there remains a need to study SPMs as adjuvants or a treatment modality in the context of obesity.

There is also emerging evidence that SPMs, besides those derived from DHA, have a role in controlling the response to influenza infection. For instance, lipoxin B4 that is synthesized from arachidonic acid, can increase memory B cell IgG levels upon in vitro treatment of cells isolated from individuals that were vaccinated for the influenza virus (52). The mechanism of action appears to be similar to the DHA-derived mediators as expression levels of BLIMP-1 were increased with lipoxin B4 treatment compared to controls.

SPMs are not just improving the response to influenza virus but also other infections. For instance, aspirin-triggered resolvin D1 biosynthesis can improve outcomes upon murine pulmonary pneumonia and thereby the need for antibiotics were lowered (1). As another example, a study showed that pharmacological inhibition of a key receptor for SPMs known as formyl peptide receptor 2 (ALX/FPR2) led to an impaired ability of the mouse to resolve pulmonary inflammation upon pneumococcal pneumonia infection (107). Finally, there is emerging evidence that conjugates of SPMs known as cysteinyl-SPMs have a role in resolution of infectious inflammation and remain to be studied in the context of influenza infection or vaccination (16).

A major gap in knowledge is how SPMs potentially control the immune response to infection with SARS-CoV-2. Several recent reviews have highlighted the potential for SPMs and their parent PUFA compounds to improve the cytokine storm upon SARS-CoV-2 infection (88; 100; 109). One pilot study suggested that increased circulating levels of EPA and DHA, the parent compounds for SPMs, could potentially be associated with a lower risk of death upon SARS-CoV-2 infection (4). However, this study was underpowered and only showed a trend, which warrants further investigation. More importantly, there are data emerging on the relationship between circulating levels of SPMs, their metabolic intermediates and COVID-19. For instance, plasma levels of specific SPMs were decreased in critically ill patients infected with SARS-CoV-2 compared to those with severe disease (89).

SIMILARITIES BETWEEN INFLUENZA AND OBESITY STUDIES WITH SARS-CoV-2 AND OBESITY

As with influenza, obesity is also a risk factor for severe outcome from infection with SARS-CoV-2. Recent studies have shown an increase in the risk of COVID-19 in obesity, even in people under the age of 60. In a study from Mexico, the odds of having COVID-19 among patients with obesity (BMI ≥ 30 kg/m²) was 61% higher than that among patients without obesity (6). A case-control study showed that obesity increased the risk by 67% (22). Another study, which used UK Biobank data (n=285,817), showed that obesity almost doubled the risk, adjusting for age, sex, ethnicity, and socioeconomic status (43). Among patients with symptoms, those with severe or critical conditions had much higher BMI and prevalence of obesity than the normal population or SARS-CoV-2 negative patients (8; 14; 43; 61–63; 92; 108; 127). Thus, it is quite clear that obesity results in a higher risk of increased severity of infection with SARS-CoV-2. These findings mirror what is reported with influenza infection—obesity independently increases risk for influenza severity and death.

In a large observational retrospective analysis, the neutralizing antibody titer was associated with BMI, with the highest titers in patients with severe obesity. The authors speculated that this finding may be due to increased severity in obese patients driving a higher antibody response. In addition, selective bias may have occurred, with patients with more severe disease dying before an antibody titer could be obtained (113). However, this finding is in contrast to several other studies in which COVID-19 patients with obesity had lower antibody titers to SARS-CoV-2 compared with non-obese subjects (32; 123).

Using a cohort of COVID-19 hospitalized and non-hospitalized subjects matched for age, sex, race/ethnicity, and date of symptom onset, Yu et al. studied both antibody and T cell responses in convalescent samples (130). They found that the breadth and magnitude of both the antibody and T cell responses were higher in hospitalized patients, compared with non-hospitalized patients. However, although co-morbidities were higher in the hospitalized patients, the authors did not include obesity as a co-morbidity. Another study reported that obesity, together with a decline in CD8+T cells, predicted a poor prognosis (119).

There are similarities in the obesity-associated immune dysfunction to both influenza virus and SARS-CoV-2. The fact that the influenza vaccine is less effective in an obese population warrants close attention to this population with monitoring vaccine responses and protection from COVID-19. Will obesity lead to a faster and/or steeper decline in protective antibodies? Will the memory T cell response be less protective from a re-infection?

Interestingly, a retrospective analysis of 25,326 subjects tested for COVID-19 between February and June of 2020 was performed at a U.S. tertiary care center to determine subject characteristics associated with increased mortality. While both obesity and diabetes increased mortality to COVIOD-19 (p<0.0001 for both), metformin treatment prior to diagnosis of COVID-19 significantly reduced mortality in subjects with diabetes (p=0.0210) (21). Other observational studies have also demonstrated decreased mortality in individuals taking metformin (46). This finding suggests that lessons learned in the study of influenza and obesity may be applicable to the study and treatment of COVID-19.

GAPS IN KNOWLEDGE AND FUTURE DIRECTIONS

There are many major gaps in knowledge that need to be addressed to drive basic and clinical studies focused on obesity and respiratory infections such as influenza. We focus on key gaps that pertain to diet and mechanisms. First, obesity is a highly heterogenous disease. This heterogeneity comes from variety of factors, which include host genetics, host microbiome status, baseline diet, geography, race, age, and sex, length of time in the obese state, amongst other factors (36). These variables are difficult to account for but may have an important role when dissecting how diet-induced obesity drives dysregulated innate and adaptive immunity. Therefore, futures studies will need to account for these variables in pre-clinical and clinical studies of obesity and influenza. This will ultimately drive precision approaches to improving outcomes in specific populations with obesity.

A second major gap in knowledge is the specific role of dietary fat in controlling the immune response. A similar argument can also be made for other macronutrients, but here we focus on dietary fat given the discussion above on fatty acid-derived metabolites of the SPM family. The evidence discussed above on PUFAs underscores the importance of fatty acid composition in controlling humoral immunity and inflammation upon viral infection. For instance, it remains unclear if SPM deficiencies can be overcome with increased n-3 PUFA consumption, either as pharmacological supplements or increased intake of foods enriched in n-3 PUFAs, to improve the response to influenza infection or vaccination in select clinical populations. Very recently, there is emerging evidence that monounsaturated fatty acids are critical in driving B cell development and germinal center formation in mouse models of influenza infection (134). Therefore, a systematic analysis of how differing fat sources control influenza infection and vaccination will drive efforts to improve infectious outcomes.

Another area for future investigation is to determine how changes in dietary patterns may control outcomes upon influenza infection and vaccination. There are compelling studies to demonstrate that a shift away from the high fat/high sugar western diet can lead to sustained improvements in glucose homeostasis and inflammation. In particular, the Mediterranean diet, which is enriched in health fats and high quality carbohydrates exerts sustained anti-inflammatory effects (11). There is also increasing interest in how fasting could potentially impact immunity upon vaccination. As an example, fasting reduced protection from influenza vaccination, which was mechanistically driven by leptin deficiency (23). Given the role of dysregulated leptin metabolism with obesity, this is an area for future investigation.

Epigenetic regulation is another area for study. The inability of weight loss to restore T cell function suggests an epigenetic effect that occurred in the obese state (98). Studies to determine the role of epigenetic alterations as well as how to reverse these effects are important areas of investigation.

Finally, there remains a critical need for continuing to investigate underlying cellular and molecular mechanisms of obesity-associated immune dysfunction (Figures 1 and 2). There are many areas for further investigation as they relate to diet and nutrition. Although it is beyond the scope of this review to cover all of these, a starting point is to investigate how chronic low-grade inflammation driven by the diet is contributing toward immunological dysfunction through changes in the microbiome and ultimately infection. There is evidence that differing components of the western diet drive dysbiosis in the gut microbiome and permeability of the intestinal barrier, which then drives an increase in systemic LPS and activation of innate immunity (45). There is a need to study how systemic inflammation mediated by the gut is driving poor responses to respiratory tract infections. As an example, disruption of the gut microbiome with antibiotics in humans led to strong changes in the vaccine response to influenza (40). Of course, the role of the lung microbiome is yet another area of investigation.

CONCLUSIONS

Obesity contributes toward poor responses to influenza infection and vaccination in humans and mouse models of disease. This is mediated through multiple complex mechanisms including dysregulated innate and adaptive immunity. Systemically, obesity leads to changes in key metabolic hormones that have a significant role in regulating immune cell function and the infection response. At a molecular level, diet-induced obesity dysregulates the metabolism of differing immune cells that are critical in the response to viral infection. In addition, an inability to resolve chronic inflammation, which is directly linked to polyunsaturated fatty acid metabolism, may be increasing susceptibility to infection. Collectively, there remains a need for further investigation into how obesity dysregulates the immune response to influenza infection and vaccination, which will impact the strategy for improving outcomes to influenza and other respiratory tract infections in a population that is increasingly obese. These gaps in knowledge, as they pertain to diet and nutrition include the need to investigate how the heterogeneity of obesity contributes toward immune dysregulation, the role of dietary fat on inflammation and infection, and the contribution of the underlying gut and lung microbiome profile of the host on disease susceptibility.

ABBREVIATIONS

EPA

eicosapentaenoic acid

DHA

docosahexaenoic acid

14-HDHA

14-hydroxydocosahexaenoic acid

17-HDHA

17-hydroxydocosahexaenoic acid

NK

natural killer

ECAR

extracellular acidification rate

OCAR

oxygen consumption rate

SPM

specialized pro-resolving mediators