Immunological mechanisms contributing to the double burden of diabetes and intracellular bacterial infections

Neutrophils

The role of neutrophils in the host immune response to intracellular bacterial infections is still widely debated. As one of the first phagocytic cells to reach sites of infection, neutrophils are adept at destroying invading pathogens through rapid release of ROS and pre-formed proteolytic granules.⁶³ Clinically, neutrophils are the predominant infected cell type in sputum and bronchoalveolar lavage of patients with active tuberculosis.⁶⁴ There is disparity between the results of in vitro studies regarding the ability of neutrophils to kill M. tuberculosis and B. pseudomallei, probably attributable to both host-specific and organism-specific factors in addition to variability in experimental design.^65–67 In experimental animal models, neutrophils are rapidly recruited to sites of infection where they contribute to early defence against M. tuberculosis and B. pseudomallei (Table4).^27,68 Neutrophil activation by M. tuberculosis influences the host immune response through regulation of surface receptor expression and secretion of chemokines and cytokines to facilitate early leucocyte migration (Table4).^27,40,69 However, while neutrophils may have a beneficial role in the early containment of bacteria, neutrophils harbouring M. tuberculosis may delay the clearance of bacteria during chronic tuberculosis.^70–73 This is consistent with the reduced bacterial loads observed following depletion of neutrophils in animal models of chronic tuberculosis.^64,72 Therefore, the role of neutrophils may largely depend on the stage of infection and their capacity to respond appropriately depending on the virulence of bacteria.

Increased production of inflammatory cytokines and ROS by unstimulated neutrophils has been described in diabetics (Table4), and attributed to direct activation by AGE.^4,41,74 However, neutrophil responses to infection appear to be predominantly suppressed in diabetic hosts.⁷⁵ Decreased pathogen-stimulated ROS production may be related to impaired glucose metabolism through the pentose-phosphate pathway, which produces NADPH, a requirement for optimal NADPH oxidase activity.⁷⁵ Furthermore, impaired activity of glutathione reductase, which also regulates neutrophil-based ROS production and phagocytosis, may be a contributing factor to neutrophil dysfunction in diabetic hosts.⁷⁶ In addition to killing bacteria directly, ROS stimulates the release of neutrophil extracellular traps (NET), another important bactericidal mechanism. Such defects in neutrophil function may favour the ability of intracellular bacteria to ‘hijack’ neutrophils as a means of refuge and dissemination in diabetic hosts.⁷⁰

Diabetes-induced functional defects in neutrophil responses to B. pseudomallei include impairments in phagocytosis, bacterial killing, neutrophil migration, cytokine production, apoptosis and NET formation.^77–79 Diabetes was also associated with attenuated lipopolysaccharide-induced cytokine responses, coinciding with reduced up-regulation of vascular cell adhesion molecule 1 and intercellular adhesion molecule 1, required for leucocyte transmigration into tissue.⁸⁰ Impaired neutrophil transendothelial migration and production of ROS have been attributed to changes caused by activation of the receptor for AGE.⁴¹ The combination of these processes may down-regulate recruitment of phagocytes during the early inflammatory process and impair initial control of bacterial growth. Conversely, increased production of inflammatory cytokines after neutrophil stimulation has also been documented in diabetes.^4,66 Differences in experimental design, infective dose and length of co-culture may account for such discrepancies. Human studies have the added confounding influence of variability in the level of hyperglycaemic control and the use of hypoglycaemic agents, which in many cases are not explicitly defined and may have important immunomodulatory effects. Excessive neutrophil involvement is a significant cause of immunopathology in chronic intracellular bacterial infections and this may be exacerbated by the pro-inflammatory milieu involved in driving diabetic complications.

Macrophages

Macrophages play a critical role in providing early host defence against intracellular bacterial infections. Important effector functions of macrophages include the phagocytosis of bacteria and clearance of apoptotic and necrotic neutrophils to contain infection. Recruitment and activation of circulating monocytes to sites of infection, where they differentiate into macrophages, are facilitated by neutrophil-derived cytokines and chemokines, such as TNF-α and CCL2.⁶⁹ In addition to phagocytic and antibacterial mechanisms, the cytokine profile of macrophages is crucial for driving effective cell-mediated immunity and protection against intracellular bacteria. M1 macrophage polarization in response to intracellular bacterial infections induces up-regulation of co-stimulatory molecules, inducible nitric oxide synthase and inflammatory cytokines, including TNF-α, IL-12 and IL-18. Production of IL-12 and IL-18 is essential for eliciting an IFN-γ response from NK cells and T cells in the establishment of T helper type 1 (Th1) cell-mediated immunity.⁸¹ Both IFN-γ and TNF-α activate macrophages and promote killing of intracellular bacteria by stimulating inducible nitric oxide synthase and NADPH oxidase, as recently reviewed by MacMicking.⁸² Clinical and experimental studies have confirmed the importance of IFN-γ and TNF-α in the control of M. tuberculosis infection.^81,83,84 However, excessive cytokine production may contribute to tissue damage, especially if chronically elevated by an unresolved infection.⁸⁵ Therefore, the inflammatory response requires precise regulation to achieve this balance between protection and injury.

Activated inflammatory macrophages are closely linked to many diabetic complications through the generation of significant levels of pro-inflammatory cytokines and ROS (Table4).^86,87 While inflammatory cytokine production by unstimulated macrophages is higher in individuals with diabetes, infection-induced cytokine production tends to be impaired compared with non-diabetic individuals.^88,89 This may be associated with reduced macrophage migration to sites of infection as suggested by lower levels of CCL2 in lung lysates in experimental models of diabetes and tuberculosis.^88,89 In addition to impaired recruitment, clinical and experimental evidence indicates that monocytes from individuals with diabetes have reduced phagocytic and antibacterial activity against M. tuberculosis and B. pseudomallei in vitro.^90–92 Reduced phagocytosis may be associated with defects in complement factors or receptor expression required for bacterial opsonization and internalization.⁹⁰ As well as providing an intracellular niche that facilitates bacterial persistence, impaired phagocytic and antibacterial activity of macrophages may have downstream effects on the activation of the cell-mediated immune responses necessary for host protection. Reduced secretion of IL-12 and IFN-γ by peripheral blood mononuclear cells from individuals with diabetes has been reported following stimulation with intracellular bacteria.⁴⁴ This is supported by in vivo evidence of lower levels of IL-12, IFN-γ and TNF-α in experimental animal models of diabetes following acute infection with intracellular bacteria.^27,93,94 These diabetes-induced changes in macrophage responses may contribute to poor containment of intracellular bacteria in the critical early stages of infection and subsequent alterations in the type of T-cell response initiated.

It has been suggested that immunological dysregulation associated with diabetes is a direct consequence of impaired glycaemic control.^44,95 The epidemiological data linking poor glycaemic control to increased risk of active tuberculosis lends support to this theory.^11,96 High glucose concentrations have been shown to inhibit lectin binding, contributing to poor pathogen recognition and impaired bacterial phagocytosis in diabetic hosts.⁹⁵ Reduced immune recognition of intracellular bacteria and altered cellular interactions potentially facilitate increased bacterial persistence.⁹⁵ Phagocytic dysfunction may be mediated directly through impaired glucose metabolism or indirectly through increased endoplasmic reticulum stress and accumulation of misfolded proteins.^97,98 These mechanisms may also contribute to the decreased expression of cell surface receptors and altered secretion of cytokines and other immunomodulatory proteins, representing an area for further research.

Natural killer cells and natural killer T cells

Natural killer cells play an important role in innate immune responses to pathogens and interest into their contribution to protection against intracellular bacterial infections has gained momentum over the past decade.²⁵ Natural killer cells are regulated by a series of inhibitory and activating receptors.⁹⁹ Experimental studies of M. tuberculosis infection have demonstrated that NK cells are recruited to sites of infection where they contribute to IFN-γ production and lysis of M. tuberculosis-infected target cells.^99–101 The NK cells also modulate T-cell responses to M. tuberculosis, favouring Th1 effector functions and contributing to CD8⁺ T cell-derived IFN-γ production and cytolytic activity.¹⁰² Down-regulation of NK cell activating receptors in patients with tuberculosis coincides with impaired IFN-γ levels and reactivation of disease.¹⁰³ Although higher numbers of NK cells have been documented in patients with diabetes before infection, decreased expression of activating receptors, NKp46 and NKG2D, has also been observed.^104,105 Impaired activation of NK cells may dampen IFN-γ production and the cytolytic activity required for the early containment and killing of intracellular bacteria.¹⁰⁵ Given the importance of NK cells in innate immunity, there is a need for research to understand the clinical relevance of diabetes-induced alterations in NK cell function and the direct effect on intracellular bacterial infections.

Natural killer T (NKT) cells are a unique subset of NK cells that also possess T-cell receptors. They respond to glycolipid rather than peptide antigens and have the potential to augment a range of immune responses.¹⁰⁶ There is evidence that NKT cells contribute to host protection in M. tuberculosis infection by inhibiting intracellular bacterial growth through cytolytic mechanisms, enhancing maturation and activation of antigen-presenting cells (APC) and modulating the type of immune response generated.^106–108 The involvement of NKT cells in adipose tissue inflammation and glucose intolerance has been described in experimental models of diabetes.^104,109 Increases in NKT cell numbers are observed in patients with tuberculosis and are higher in the blood and bronchoalveolar lavage of patients with co-morbid diabetes than those without.¹¹⁰ This may be a direct consequence of the increased bacillary burden observed in these patients and has been suggested as a useful marker for active tuberculosis.¹¹⁰ Whether diabetes causes functional defects in NKT cell activity or otherwise biases the immunomodulatory response to intracellular bacterial infections is an area worthy of further research.

Dendritic cells

Dendritic cells represent an important link between innate and adaptive immune responses. Mature DC are potent immune-modulators and APC for priming specific lymphocyte responses.^26,111 At the onset of infection with intracellular bacteria, DC accumulate at the site of infection to participate in bacterial uptake and antigen processing. Antigen presentation takes place following migration of mature DC to draining lymph nodes.^26,28 Dendritic cells also modulate the lymphocyte profile generated through production of immunoregulatory cytokines, such as IL-12 and IL-18, essential for effective Th1 cell-mediated immune clearance of intracellular bacteria.^26,28 Suppression of DC trafficking to lymph nodes has been suggested as a mechanism by which M. tuberculosis evades the early host immune response.²⁶ Defects in DC maturation, migration and interaction with T cells may also contribute to intracellular bacterial persistence within the host.²⁶

Increased expression of activation markers on unstimulated DC from diabetic individuals has been documented.^112,113 Despite efficient trafficking of DC to regional lymph nodes, an initial delay in the recruitment of myeloid cells to the pulmonary site of infection was observed in diabetic mice following infection with M. tuberculosis.⁸⁸ This coincided with reduced levels of CCL2 and CCL5, chemokines involved in the recruitment of macrophages and DC.⁸⁸ Reduced early recruitment of APC to the primary site of infection may account for delayed induction of protective T-cell-mediated immune responses.¹¹⁴ In an experimental animal model of diabetes, DC phagocytosis of intracellular bacteria was also impaired.⁹¹ However, there were no differences in the up-regulation of DC markers involved in antigen presentation and co-stimulation of naive T cells. Impaired phagocytosis and delayed kinetics of antigen presentation at the onset of infection potentially contributes to poor early control and downstream alterations in lymphocyte activation.

Increased oxidative stress in diabetic hosts may also influence the profile of cytokines secreted by APC during intracellular bacterial infections. The reduced intracellular GSH : GSSG ratio in APC from diabetic individuals alters the secreted cytokine profile due to the immunomodulatory properties of GSH.^115–117 Consistent with this, peripheral blood mononuclear cells from patients with poorly controlled diabetes had defects in IL-12 production in response to intracellular bacterial infection, which could be reversed by replenishing GSH levels.⁴⁴ The exact mechanisms by which GSH influences IL-12 production are under investigation but may involve the modulation of intracellular redox status and glutathionylation of signalling intermediates or transcription factors.⁴⁴ The therapeutic potential of agents that together improve the IL-12/IFN-γ axis and decrease oxidative stress represents an exciting avenue to pursue.

T-cell-mediated immunity is critical to host protection against intracellular bacterial infections. Diabetes-induced alterations in the immunomodulatory nature of DC may influence the type of T-cell response elicited, which is an important determinant in the long-term outcome of intracellular bacterial infections. Effective antigen presentation in secondary lymphoid organs, together with early secretion of IL-12 and IFN-γ, is essential for the priming and differentiation of Th1 cells involved in host protection. Development of active pulmonary or extrapulmonary tuberculosis, a common finding in those with co-morbid diabetes (Fig.1), has been linked to impairments in Th1 cell-mediated immunity.^20,118,119 Potential alterations in the development of specific T-cell responses may be a secondary complication of the diabetes-induced innate immune defects already described (Fig.3). In an experimental model of co-morbid tuberculosis and diabetes, reduced levels of chemokines and cytokines were associated with delayed priming of T cells.^88,114 This was followed by a higher pulmonary M. tuberculosis burden and an exaggerated inflammatory response during the latter stages of infection after specific adaptive immunity was established (Fig.3).^88,114

Lymphocytes

Host protection against intracellular bacterial infections relies on a strong T-cell-mediated response.¹²⁰ T cells differentiate into a range of subtypes (Th1, Th2, Th17, Treg), which elicit distinct types of immunity fundamentally based on secreted cytokine profiles. An early influx of IFN-γ-producing Th1 cells is a significant determinant of protection against intracellular bacterial infections.¹²¹ There is strong evidence to indicate an initial delay in activation of Th1 cell-mediated immunity in diabetic hosts.^44,88,114 However, there is also clinical and experimental evidence that the late inflammatory response during chronic tuberculosis is enhanced (Table5), although it may come too late to rescue diabetic hosts from bacterial dissemination.^122,123 It is possible that this late hyper-inflammatory response is a direct result of increased antigenic stimulus, as a consequence of impaired innate immune control, or a cumulative build-up adding to the chronic inflammation underlying the immunopathology of diabetes itself.¹²⁴ Increased circulating levels of Th1- and Th17-associated cytokines have been described in patients with tuberculosis and co-morbid diabetes.¹²⁵ In vitro stimulation of whole blood with M. tuberculosis antigens resulted in elevated frequencies of CD4⁺ Th1 cells and Th17 cell subsets.¹²² However, lower production of IFN-γ by CD4⁺ T cells from patients with tuberculosis and poorly controlled diabetes has also been documented, consistent with reduced expression of IL-12 by APC.^44,126 The inconsistencies between findings may be attributed to differences in study design and cell culture, including the absence of additional leucocyte interactions and removal from hyperglycaemic conditions. Furthermore, the Th1 response in diabetic hosts may have limited efficacy because of impaired cellular interactions and an inability to mount an effective antibacterial response, leading to intracellular bacterial persistence.

There is still a paucity of research on the role of other T-cell subsets in co-morbid diabetes and intracellular bacterial infections. While Th1-mediated immunity plays a crucial role in host protection against intracellular bacteria, the functional significance of Th17 responses is less clear.¹²⁷ Experimental evidence indicates that Th17 responses may facilitate dissemination of M. tuberculosis, potentially through IL-17 secretion and its role in neutrophil recruitment.¹²⁸ Bacterial dissemination may be further exacerbated by the functional defects in neutrophil and macrophage bactericidal mechanisms described in diabetic hosts. Without appropriate regulation, exaggerated Th17 responses may also contribute to immune-mediated pathology.¹²⁹

Regulatory T (Treg) cells are critical for preventing exaggerated inflammatory responses to limit host tissue damage, although this may also limit host immunity and pathogen clearance.¹³⁰ In experimental models of tuberculosis, Treg cells impair immune protection by delaying recruitment of effector CD4⁺ and CD8⁺ T cells to sites of infection.¹³⁰ The Th2-mediated responses are also correlated with increased susceptibility to intracellular bacterial infections.¹³¹ Decreased numbers of Treg and Th2 cells are an underlying feature of diabetes.^119,120,122 Interestingly, elevated systemic levels of immunosuppressive Treg and Th2-type cytokines, IL-10, TGF-β and IL-4, and an overall lower Th1 : Th2 ratio have been documented in patients with tuberculosis and diabetes.^122,131 This may contribute to susceptibility in diabetic hosts by abrogating the protective mechanisms afforded by Th1-type cytokines and enhancing intracellular bacterial persistence.¹³¹ The diabetes-induced dysregulation of T-cell responses to intracellular bacteria is no doubt complex and requires further clarification.

While T-cell activity is critical to the protective immune response to intracellular bacterial infections, the definitive role of B cells is still debated.¹³² Until recently, B cells were generally considered to be of little benefit during intracellular bacterial infections, given the limited protection afforded by humoral immunity in general. However, it is now appreciated that B cells are present in tuberculosis granulomas in active follicle-like centres, where they influence local inflammatory responses (Table5).¹³³ In particular, B cells have a newly defined role in regulating neutrophil migration to the site of infection. Neutrophilia is a compensatory response to B-cell deficiency during M. tuberculosis infection.¹³³ Hyperglycaemia-induced functional defects in B cells have been described in vitro, specifically resulting in impaired immunoglobulin production, although the clinical relevance of this remains to be shown.¹³⁴ Whether diabetes also delays the kinetics of B-cell activation and how this may influence the immunoregulatory role of B cells remains to be discerned. In this respect, defects in B-cell function in diabetic hosts may have significant immunomodulatory consequences on the functional response of other leucocytes, although this is yet to be determined.

Granuloma formation

When absolute bacterial clearance cannot be achieved, dynamic cellular interactions at sites of infection lead to the formation of granulomas, a characteristic feature of tuberculosis. Granulomas depend on the organized and complex interaction of many immune cell populations, including macrophages, DC and T and B cells, as recently reviewed by Guirado and Schlesinger.¹³⁵ The precise balance and kinetics of cytokine and chemokine production and appropriate cellular function are necessary for proper granuloma formation.^29,136 For the past decade, granulomas were primarily considered a protective mechanism, providing a barrier against bacterial dissemination and containing inflammatory processes to limit collateral tissue damage.¹³⁷ However, cavitary granulomatous lesions can increase bacterial dissemination and are associated with destruction of lung parenchyma.¹³⁸

The influence of diabetes on granuloma formation in tuberculosis is not well understood. Clinical observations of an increased frequency of cavitary lung lesions in patients with tuberculosis and co-morbid diabetes may indicate alterations in granuloma formation.^20,32 Larger granulomas were also observed in an experimental model of diabetes and tuberculosis.²⁷ This coincided with reduced production of TNF-α, IL-12 and nitric oxide by alveolar macrophages.²⁷ Up-regulation of IFN-γ and TNF-α is critical for control of M. tuberculosis infection and in facilitating appropriate granuloma formation.^81,83,84,136 It is possible that delayed pulmonary migration of macrophages, DC and activated T cells, caused by reduced CCL2 and CCL5, also contributes to altered structural organization of granulomas in individuals with co-morbid diabetes. Cavitary lesions are associated with increased degenerative macrophages and infiltration of neutrophils, which is typical of mice lacking IFN-γ, TNF-α and CCL2.^101,139 Activated macrophages play a pivotal role in containment of bacilli within the granuloma, so incomplete macrophage activation or impaired microbicidal mechanisms in diabetes may contribute to M. tuberculosis escape and increased dissemination.