Turning foes into permissive hosts: manipulation of macrophage polarization by intracellular bacteria

Abstract

Macrophages function as tissue immune sentinels and mediate key antimicrobial responses against bacterial pathogens. Yet they can also act as a cellular niche for intracellular bacteria, such as Salmonella enterica, to persist in infected tissues. Macrophages exhibit heterogenous activation, or polarization, states that are linked to differential antibacterial responses and bacteria permissiveness. Remarkably, recent studies demonstrate that Salmonella and other intracellular bacteria inject virulence effectors into the cellular cytoplasm to skew the macrophage polarization state and reprogram these immune cells into a permissive niche. Here, we review mechanisms of macrophage reprogramming by Salmonella and highlight manipulation of macrophage polarization as a shared bacterial pathogenesis strategy. In addition, we discuss how the interplay of bacterial effector mechanisms, microenvironmental signals, and ontogeny may shape macrophage cell states and functions. Finally, we propose ideas of how further research will advance our understanding of macrophage functional diversity and immunobiology.

Introduction:

Macrophages are mononuclear phagocytes that act as tissue immune sentinels and perform critical homeostatic functions, including phagocytosing spent cells, recycling nutrients, remodeling tissues, and resolving inflammation(1, 2). Armed with arsenals of cell surface and cytosolic pattern recognition receptors (PRRs) that sense pathogen associated molecular patterns (PAMPS), during bacterial infection macrophages readily phagocytose invading pathogens, release inflammatory mediators to facilitate innate and adaptive immune responses, and kill bacteria via production of reactive metabolites and multistep inflammasome activation(3, 4). Paradoxically macrophages can also serve as a cellular niche for intracellular bacterial pathogens, such as Salmonella enterica, Bartonella henselae, and Mycobacterium tuberculosis (Mtb), to survive within infected tissues(5-7). Macrophages exhibit differential polarization, or activation states, and these functional states have distinct cellular phenotypes and functions(8-10). Macrophage polarization has been described using a conceptually simplified M1 and M2 framework(11-13). The classical, or M1, macrophage activation paradigm occurs in microenvironments upon recognition of PAMPs, such as LPS, and inflammatory immune signaling, such as IFN-γ. These stimuli activate downstream transcriptional regulators, including NFκB and STAT1, to generate inflammatory and antibacterial functional states(10, 11). On the other hand, the alternatively activated, M2 polarized state prototypically entails macrophage activation from the type-2 immunity associated cytokines such as IL-4 and IL-13 or IL-10 that trigger STAT6, STAT3, and Peroxisome Proliferator-Activated Receptors (PPARs) among other regulators, to elicit macrophage activities involved in resolving inflammation and tissue repair. The dichotomous M1 and M2 framework does not, however, fully capture the heterogeneous functional states of tissue macrophages in vivo. Increasingly macrophage polarization and heterogenous phenotypes have been recognized as a spectrum of functional states, with overlapping cellular features that are shaped by a multitude of factors and dependent on pathophysiological contexts(10, 14, 15).

Accumulating studies over the past decade have linked macrophage polarization states to differential antibacterial capacity and permissiveness, with an M2-like functional state being more permissive for intracellular bacterial replication and survival(6, 16-18).

Macrophage ontogeny and microenvironmental factors influence their functional phenotypes, such as antibacterial responses(19). Intracellular bacteria that exploit macrophages as a cellular niche to establish persistent infection, such as Salmonella enterica serovar Typhimurium (STm), express macromolecular secretion systems to inject virulence effector proteins into the host cell cytoplasm to co-opt cellular activities(5, 20, 21). Thus, in addition to evading antibacterial innate immune responses and exploiting existing favorable macrophage functional states, a fascinating question for some time is whether intracellular bacteria have specific mechanisms to actively skew macrophage polarization toward more permissive, M2-like states. Herein, we review recent research that uncover mechanisms by which Salmonella and other intracellular bacteria employ injected virulence effectors to manipulate macrophage polarization and reprogram macrophage functional states. We highlight how the STAT3 pathway has emerged as a critical cellular target for intracellular bacteria to reprogram macrophages and may represent a convergent bacterial pathogenesis strategy. These fascinating findings from studies of intracellular bacterial infections have broadened our perspectives on macrophage heterogeneity and generate exciting questions for future research that will further our understanding of macrophage immunobiology.

Macrophage heterogenous polarization states and antibacterial capacities

The mechanisms involved in early steps of macrophage recognition of and activation by PAMPs through PRRs during bacterial infections that lead to macrophage inflammasome activation, antibacterial immune responses, and bacterial killing have been active areas of investigations for many years and are extensively reviewed(4, 22). Although macrophage polarization states have been more widely used to describe differential phenotypes and functional properties associated with inflammatory immune responses, metabolic activities, tissue repair, and anti-tumor responses, they have also been linked with differential antibacterial capacities. Early experiments examining the functional impact of polarization states showed pre-treatment with either IL-4 or IFN-γ differentially affected the antimicrobial response of macrophages to Mtb infection, such as production of reactive nitrogen species(23). Using single-cell RNA-sequencing (scRNA-seq) and bulk population analyses, the expression levels of surface macrophage IL-4Rα, which is a canonical functional marker of alternatively activated or M2-like macrophages, have been shown to tightly correlate with macrophage capacity to support replicating and persistent intracellular STm in murine bone marrow-derived macrophages (BMDMs) in vitro and tissue macrophages in vivo(17, 24-26). IL-4Rα is the shared subunit of IL-4 and IL-13 receptors that transduce signals activating STAT6. In macrophages, IL-4 stimulated DNA binding of the transcriptional regulator STAT6 turns on an anti-inflammatory, M2-like functional program and also represses inflammatory responses(27, 28). IL-4 treatment results in significantly higher intracellular bacterial levels in STm-infected BMDMs, compared to IFN-γ treatment(29). Furthermore, conditional knockout mice in which Il4ra is deleted in myeloid cells, including monocytes and macrophages, by the LysM Cre recombinase activity have reduced bacterial levels in infected tissues during persistent STm infection(29). In addition to IL-4 and IL-13/IL-4R/STAT6 axis, the IL-10/IL-10R/STAT3 pathway is a key cytokine signaling axis that promotes M2-like macrophage polarization states and anti-inflammatory programming(11). During mycobacterial infection, the IL-10 mediated, anti-inflammatory response of macrophages is thought to limit excessive tissue damage from inflammation, but this host response may be exploited by bacteria for survival. IL-10^−/− mice had been shown to have enhanced protection from Mtb infection, which was associated with more robust Th1 response(30). However, the protective effect from disabling IL-10 signaling may vary with the ability to induce IL-10 production among Mtb strains, host genetic differences, and other factors(30, 31). In addition, a recent study found that induced Pluripotent Stem Cells (iPSC)-derived human macrophages lacking IL-10R have higher intracellular bacterial levels when infected with STm and this phenotype is prostaglandin E-dependent, suggesting additional factors controlling intracellular bacteria survival within macrophages are independent of IL-10 signaling capacity(32).

Increasing evidence suggests macrophage polarization and functional states are interrelated with their cellular metabolic activities(10, 11). During intracellular bacterial infections, M2-like metabolic states have been associated with increased macrophage capacity to promote bacterial replication and survival. PPARα, PPARβ/δ, and PPARγ are a family of nuclear receptor proteins that regulate carbohydrate and lipid metabolism, energy balance, and inflammation. Their expression and activities can be induced by upstream IL-4, IL-13, and IL-10 signals and have been recognized as a metabolic signature of M2-like polarization states(10, 11). In BMDMs, PPARβ/δ and PPARγ enhance intracellular STm and Brucella abortus bacterial levels, respectively(29, 42). PPAR-driven permissiveness of macrophages for intracellular STm and B. abortus is dependent on macrophage glucose availability and the abilities of bacteria to utilize glucose, linking macrophage metabolic activities with intracellular bacteria metabolism and survival. Furthermore, PPARδ^−/− and PPARγ^−/− mice infected with STm and B. abortus, respectively, have lower abundances of M2-like macrophages and reduced bacterial levels in infected tissues, such as spleens and mesenteric lymph nodes, during persistent infection(29, 42). Similarly, macrophages also exhibit differential metabolic activities and bacteria permissiveness during Mtb infection, with PPAR expression and function correlating with higher capacity to support intracellular bacterial growth and persistence(43-45). The growing body of studies on macrophage metabolism and its functional impacts is being reviewed by Avraham R. et. at. in this issue. A summary of key host factors influencing macrophage polarization during Salmonella infection is provided in table 1.

Table 1.

Key cellular regulators of macrophage polarization in Salmonella infection

Protein (gene)	Function	Gene/protein manipulation	Macrophage polarization effect	References
PAMPS/PRR pathways(various)	various	various	promoting M1-like polarization	(4, 10, 22)
IFNκ	cytokine	cytokine stimulation	promoting M1-like polarization	(29, 33)
IL-10	cytokine	complete knockout, cytokine stimulation	promoting M2-like polarization	(25, 34, 35)
IL-4 and IL-13	cytokines	cytokine stimulation	Promoting M2-like polarization	(29, 33)
IL-4Rα	IL-4 and IL-13 signaling	conditional knockout (LysM-Cre)	Promoting M2-like polarization	(17, 29)
TNF	cytokine	complete knockout, cytokine neutralization	promoting M1-like phenotypes and restraining M2-like polarization	(26, 36)
PPARδ	transcription factor	complete knockout	Promoting M2-like polarization	(29)
STAT1	transcription factor	complete knockout	promoting M1-like polarization	(37, 38)
STAT3	transcription factor	siRNA, chemical inhibitors	Promoting M2-like polarization	(25, 26, 39)
GSK3	protein kinase	complete knockout, chemical inhibitors	Promoting M2-like polarization	(25, 39)
IRF5	Transcription factor	siRNA	Promoting M1-like polarization	(40)
KDM6B	epigenetic regulator	siRNA, chemical inhibitors	Promoting M2-like polarization	(41)

Macrophage polarization has functionally been tied to infection outcomes. In infected tissues, intracellular bacterial pathogens are contained and persist within granulomas, which are immunological microstructures comprised of macrophages and diverse types of immune and non-immune cells(26, 46-49). In silico, computational modeling studies suggest that the relative M1 and M2 activities may predict infection containment or dissemination in Mtb-infected lungs of nonhuman primates(16). Transcriptomics analyses of STm-infected BMDMs have identified an inflammatory, M1-like transcriptional signature associated with bystander macrophages, which were exposed to bacteria but not infected, and macrophages containing intracellular, host-killed bacteria. In contrast, an M2-like transcriptional signature was associated with macrophages containing persistent and growing bacteria(17, 24). During persistent infection in mice, the levels of STm tissue persistence were linked to relative abundances of different macrophage polarization and functional states, which were dependent on TNF, a pleotropic cytokine that retrains M2-like polarization among its many pathophysiologic activities(26, 27). Additionally, STAT6, which is a potent transcriptional regulator of M2-like polarization state, was recently found to control granuloma macrophage epithelial reprogramming and bacterial levels during Mycobacterium marinum infection in zebrafish(50). Inhibition of IL-4R had a lesser effect(50). Although examination of Il4ra^−/− mice infected with Mtb reported no major bacterial control and histopathological defects, in vitro experiments with primary macrophages infected with Mtb showed IL-4 impaired bacterial restriction(51, 52).

Intracellular bacteria effector mechanisms that reprogram macrophage polarization states and antibacterial capacities

Salmonella, mycobacteria, and other intracellular bacteria that exploit host cells as a niche to survive and establish persistent infection have membrane-spanning, multiprotein secretion systems to inject virulence effector proteins into the host cytoplasm to modulate cellular activities. STm, which is commonly utilized as model Salmonella enterica serovar for experimental infections, encodes two type 3 secretion systems (T3SSs) on the Salmonella Pathogenicity Islands (SPI)1 and 2 in its genome. These T3SS-1 and 2 are required to inject various effectors that facilitate cellular invasion, immune evasion, and intra-macrophage replication and persistence(20, 53). For example, STm deploys multiple injected effectors, to inhibit NFκB signaling, including SpvD, which interferes with nuclear import of the NFκB subunit p65 and SteA, which suppresses ubiquitination and degradation of the NFκB inhibitor IκB(20, 54, 55). Similarly, B. abortus and Mtb employ various secreted effectors to co-opt host pathways for forming bacteria-containing vacuoles, interfering with antigen presentation, inhibiting NFκB, and impairing reactive oxygen species production(5, 56).

Recent studies demonstrate that intracellular bacterial pathogens have specific effector mechanisms to actively skew macrophage polarization and reprogram macrophages into a permissive niche. It had been known that exposure to viable STm induces STAT3 phosphorylation in BMDMs, though how much of this induction is driven by host signaling and/or directly mediated by bacterial mechanisms was not defined(34). In macrophages, STAT3 can be activated by IL-6 and IL-10 signaling and regulates expression of a number of target genes, including Il4ra, which encode gene products that collectively trigger and reinforce anti-inflammatory responses to limit inflammation during mycobacterial infection and other pathophysiologic settings(57-61). Using gene deleted BMDMs, Lin et. al. found deletion of Il6 or Il10 upstream cytokine signaling did not abolish STm-induced STAT3 phosphorylation(34). Clues that STm may inject effector(s) to manipulate macrophage polarization emerged when the SPI2 T3SS effector SteE (also known as SarA) was shown to be necessary and sufficient for STAT3 phosphorylation in lymphoblastoid cell lines(62). Subsequently studies demonstrated that SteE drives STAT3 phosphorylation and IL-4Rα expression to promote M2-like polarization and bacteria permissiveness of primary and tissue macrophages(24-26). Intriguingly, Panagi et. al. showed that SteE binding to the host serine/threonine Glycogen Synthase Kinase 3 (GSK3) changes its substrate binding pocket and specificity, enabling it to phosphorylate and activate STAT3 at the residue Tyr-705(25). In addition, Gibbs et. al. found that SteE/SarA C-terminus biochemically and functionally mimics the cytoplasmic domain of the common glycoprotein 130 (gp130) subunit of type I cytokine receptors such as IL-6R(39, 63). SteE contains a YXXQ motif that when phosphorylated binds to STAT3 with greater affinity than the YXXQ motif of gp130(39). SteE-driven polarization of macrophages to M2-like, permissive states counteracts TNF-mediated restriction and promotes STm persistence in granulomas and infected spleens in mice, demonstrating how manipulation of macrophage polarization is an important bacterial pathogenesis mechanism facilitating tissue persistence despite host immune responses(26).

Previous studies with the STm effector SteE raises the possibility that other intracellular bacteria may similarly polarize and reprogram macrophages to permissive states using injected virulence effectors. Recent studies with the intracellular bacterium Bartonella henselae have identified an effector BepD, which is translocated into host cell through a Type 4 Secretion System (T4SS), plays a role in reprogramming macrophages to M2-like, permissive states(64). Using mononuclear phagocyte cell lines and macrophages, Sorg et. al. showed that when injected into the cellular cytoplasm, BepD serves as a scaffold to bring together the c-ABL tyrosine kinase and STAT3 and facilitates phosphorylation of STAT3 at the Tyr-705 residue in a c-ABL dependent manner(64). This BepD-mediated activation of STAT3 promotes cellular IL-10 production and impairs TNF secretion, skewing macrophages toward an M2-like functional state. Additionally, Mtb infection and the Mtb T7SS effector ESAT-6 have been found to induce STAT3 phosphorylation in BMDMs(65, 66). However, whether Mtb virulence effectors directly mediate STAT3 phosphorylation remain to be defined. Interestingly, orthologous coding DNA sequence of SteE is absent in Salmonella enterica serovar Typhi(20), which is a human-restricted pathogen that has predilection for causing persistent infection(53). However, it is possible that S. Typhi employs a different virulence effector other than SteE to skew macrophage polarization or dampens the antibacterial responses of macrophages via mechanisms that are not dependent on active manipulation of polarization. Collectively, these studies suggest that the ability of intracellular bacterial pathogens to manipulate macrophage polarization and functional states represents a convergent evolution strategy to persist in the mammalian hosts (Table 2) (Figure 1).

Table 2.

Intracellular bacterial effector mechanisms and macrophage polarization

Bacterium	Injected effector	Host target	Macrophage polarization effect	References
S. Typhimurium	SteE	GSK3/STAT3	promoting M2-like polarization by mediating STAT3 phosphorylation	(24, 25, 26, 39, 62)
S. Typhimurium	SteA	CUL-1/IκB	dampening M1-like polarization by interfering with IκB degradation	(55)
S. Typhimurium	SseK1	FADD/TRADD	dampening M1-like polarization by inhibiting NFκB signaling	(20, 67)
S. Typhimurium	SseK3	TRIM32/TRADD	dampening M1-like polarization by inhibiting NFκB signaling	(20, 67)
S. Typhimurium	SpvD	XP02/NFκB	dampening M1-like polarization by interfering with NFκB subunit nuclear localization	(20, 54)
S. Typhimurium	GogB	SKP1/FBXO22	dampening M1-like polarization by inhibiting NFκB signaling	(20, 68)
B. henselae	BepD	STAT3/c-ABL	promoting M2-like polarization by mediating STAT3 phosphorylation	(64)
M. tuberculosis	ESAT-6?	unknown	may modulate macrophage polarization by targeting STAT3 phosphorylation	(65, 66)

Figure 1: Intracellular bacteria manipulate macrophage polarization via STAT3 signaling.

Inside macrophages, S. Typhimurium reside within modified, Salmonella-containing vacuoles and inject the virulence effector SteE through T3SS into cellular cytoplasm. SteE promotes phosphorylation and activation of STAT3 by GSK3. Host IL-6 and IL-10 signaling also induce STAT3 phosphorylation and activation. Phosphorylated STAT3 translocates into the cellular nucleus and regulates expression of target genes, including Il4ra and Arg1, culminating in M2-like, anti-inflammatory responses that are more bacteria permissive, such as enhanced IL-4R/STAT6 signaling and reduced production of Reactive Nitrogen Species (RNS). Similarly, during B. henselae infection, vacuole-residing intracellular bacteria inject the effector BepD through T4SS to mediate phosphorylation of macrophage STAT3 by c-ABL. Whether a macrophage ultimately acting as and antibacterial cell or a permissive niche for bacteria may vary depending on the summative strengths of microenvironmental signals such as of IL-10 and IL-6, other macrophage antimicrobial properties, and the robustness of bacterial effector mechanisms to skew macrophage polarization.

Impacts of ontogeny and microenvironment on macrophage functional states and antibacterial capacities

The functional phenotypes and programing of macrophages are shaped by lineage determining transcriptional factors and signal dependent transcriptional factors that are driven by their microenvironment milieu(69). In mammalian tissues, a fraction of macrophages originate from embryonic precursors and others share a hematopoietic developmental origin with monocytes(1,2, 15). A recent sc-RNAseq study of acute STm infection in mice showed that a non-classical monocyte-derived macrophage population harbors more intracellular bacilli than other types of macrophages in the infected spleens, including red pulp macrophages, which have been known to originate from embryonic precursors(70, 71). Others have found that localization to microenvironments with varying surrounding cell-cell interactions and signaling milieu in infected tissues influences the antibacterial capacities of macrophages. In the spleens of mice infected with STm, CXCL9/10+ macrophages preferentially localize to the periphery of granulomas, closer to T-cell rich areas during persistent infection(47). These macrophages are less likely to harbor intracellular bacteria compared to macrophages that occupy the center of granulomas and express high levels of the inducible Nitric Oxide Synthase (iNOS), which is involved in production of Reactive Nitrogen Species (RNS) and commonly associated with M1-like, antibacterial macrophage functions(47, 72). However, many iNOS+ granuloma macrophages are not infected and many macrophages harboring bacteria have undetectable iNOS levels, underscoring the spectrum and functional state heterogeneity of macrophages in infected tissues.

The link between ontogeny and tissue microenvironments to macrophage functional states and phenotypes has also been observed in mycobacterial infections. In Mtb-infected nonhuman primate lungs, macrophages that distribute in different regions of granulomas exhibit varying functional features, such as iNOS and arginase expression(73). In mice, tissue-resident alveolar macrophages, which originate from embryonic precursor cells, have been shown to be a more favorable replicative niche for Mtb bacilli than monocyte-derived interstitial macrophages and this functional difference may be due in part to distinct metabolic programming of the two macrophage populations(45, 74). However, alveolar macrophages can re-localize to the lung interstitial region, which is dependent on IL-1R/MyD88 signaling(74). Furthermore, during and after an inflammatory state, depleted tissue alveolar macrophages can be replaced by hematopoietic, monocyte-derived macrophages(75, 76). Thus, ontogeny, microenvironmental signals, and pathogen factors may dynamically interact to shape a spectrum of macrophage functional states and antibacterial capacities in infected tissues (Figure 1). Consistent with this notion, recent scRNA-seq studies showed that both alveolar macrophages and interstitial macrophages are comprised of multiple distinct subsets with varying capacities to restrict intracellular bacteria(77). Approaches to defining macrophage heterogenous polarization and functional states in infected tissues often depend on biased cellular markers and bulk population analysis. Emerging scRNA-seq studies will be critical for identifying key cellular features of macrophage heterogenous functional states in an unbiased manner that capture the full heterogeneous spectrum and interplay of many host and pathogen factors influencing macrophage diverse phenotypes and functions(72, 77).

Perspective:

Defining the factors and mechanisms that shape tissue macrophage capacities to produce antibacterial responses and eliminate invading pathogens or act as a cellular niche for diverse intracellular bacteria to persist infected tissues is critical for the understanding of microbial pathogenesis and macrophage biology. Although studies over the years have provided insights and expanded the concept of macrophage cell states and functional diversity substantially from an early M1 and M2 framework, findings on how intracellular bacteria employ virulence effectors to actively reprogram macrophage functional states and turn these immune cells into permissive cellular niches have advanced the mechanistic and conceptual principles underlying macrophage heterogeneity and diverse functions. A number of intriguing questions remains that will be exciting subjects of further inquiries. Firstly, do other intracellular bacteria, such as S. Typhi, use injected virulence factors to reprogram macrophage cell states? What are the host pathways that these bacterial factors co-opt? We propose that actively reprogramming macrophages into a permissive cellular niche represents evolutionarily convergent microbial pathogenesis strategy for diverse intracellular bacteria to establish persistent infection. Another interesting question is how macrophages integrate bacteria-driven effects and host signaling effects on the same pathway, such as activation of STAT3 by SteE and upstream IL-6 or IL-10 cytokine signals? Are there differences in downstream transcriptional and functional outcomes when STAT3 is activated by a bacteria-driven mechanism versus host signaling? Since the bacterial levels and tissue microenvironmental milieu dynamically change over the course of infection, bacterial effects on macrophage polarization states and reprogramming likely vary throughout the course of infection. Reprograming of macrophages by intracellular bacteria may also vary significantly across different macrophage populations in infected tissues. For example, SteE may differentially affect red pulp macrophages versus monocyte-derived macrophages in STm-infected spleens. Finally, how stable are the effects of bacterial reprogramming on macrophages and what are the long-term fates of the macrophages that have been reprogrammed to be a permissive niche? Elucidating these outstanding questions will further our understanding of macrophage biology and bacterial pathogenesis, as well as enable us to more precisely target and harness macrophage pathways for therapeutic purposes.