Distinct Mechanisms of Type 3 Secretion System Recognition Control LTB₄ Synthesis in Neutrophils versus Macrophages

Abstract

Leukotriene B4 (LTB₄) is critical for initiating the inflammatory cascade in response to infection. However, Yersinia pestis colonizes the host by inhibiting the timely synthesis of LTB₄ and inflammation. Here, we show that the bacterial type 3 secretion system (T3SS) is the primary pathogen associated molecular pattern (PAMP) responsible for LTB₄ production by leukocytes in response to Yersinia and Salmonella, but synthesis is inhibited by the Yop effectors during Yersinia interactions. Moreover, we unexpectedly discovered that T3SS-mediated LTB₄ synthesis by neutrophils and macrophages require two distinct host signaling pathways. We show that the SKAP2/PLC signaling pathway is essential for LTB₄ production by neutrophils but not macrophages. Instead, phagocytosis and the NLRP3/CASP1 inflammasome are needed for LTB₄ synthesis by macrophages. Finally, while recognition of the T3SS is required for LTB₄ production, we also discovered a second unrelated PAMP-mediated signal independently activates the MAP kinase pathway needed for LTB₄ synthesis. Together, these data demonstrate significant differences in the signaling pathways required by macrophages and neutrophils to quickly respond to bacterial infections.

Introduction

Rapid recruitment of immune cells to sites of infection is essential to control bacterial pathogens. However, sustained or chronic inflammation is detrimental to the host. Therefore, inflammation is a tightly regulated cascade of events controlled by both lipid and protein mediators (1). Initiation of this cascade is primarily mediated by inflammatory lipids, of which the eicosanoid leukotriene B4 (LTB₄) is one of the earliest produced during infection. LTB₄ is a potent pro-inflammatory chemoattractant and immune cell activator necessary for timely recruitment of neutrophils to sites of infections (1–6). LTB₄ is derived from arachidonic acid (AA) upon activation of the enzymes cytosolic phospholipase A2 (cPLA₂) and 5-lipoxygenase (5-LOX). The enzymes are activated by phosphorylation via MAPK signaling and Ca²⁺ binding to translocate to the nuclear membrane or a lipidisome, where a complex is formed with 5-LOX activating protein (FLAP) (7–14). This complex converts AA to LTA₄, which is rapidly converted to LTB₄ by LTA₄ hydrolase (7, 8). Upon subsequent release, LTB₄ is recognized by the high affinity BLT1 receptor on immune cells to promote priming, activation, and chemotaxis. Mouse models that disrupt the host’s ability to produce or respond to LTB₄ are impaired in initiating inflammation and are significantly more susceptible to a variety of infections, demonstrating the importance of LTB₄ in inducing a robust inflammatory response required for effective clearance of pathogens (1, 2, 15, 16).

Plague is an acute infection caused by the bacterial pathogen Yersinia pestis. A hallmark manifestation of plague is a significant delay in the recruitment of immune cells to the sites of infection, allowing for bacterial colonization and replication (17–20). One of the key virulence determinants for Y. pestis to inhibit the immune system and colonize the host is the Ysc type 3 secretion system (T3SS) encoded on the pCD1 plasmid (21, 22). This secretion system mediates direct translocation of bacterial effector proteins called Yops into host cells (22–25). During mammalian infection, Y. pestis primarily targets neutrophils and macrophages for T3SS-mediated injection of the Yops (26–28). Once inside host cells, the Yop effectors target specific host factors to inhibit phagocytosis, reactive oxygen species (ROS) synthesis, degranulation by neutrophils, and inflammatory cytokine and chemokine release that is required to recruit circulating leukocytes to infection sites (29–37). Moreover, the Yop proteins were recently shown to actively inhibit the synthesis of LTB₄ required for the initiation of the inflammatory cascade during pneumonic plague (37, 38). However, in the absence of the Yop effectors, neutrophils interacting with Y. pestis can synthesize LTB₄, but synthesis requires bacterial expression of the T3SS and the YopB/D translocase (38). These data indicate that the T3SS is a pathogen associated molecular pattern (PAMP) recognized by neutrophils that triggers signaling pathways that lead to inflammatory lipid synthesis.

Previous studies in macrophages have established components of the T3SS induce inflammatory cell death pathways. In the absence of the Yops, the T3SS and YopB/D translocon proteins can induce NLRP3-dependent activation of the caspase 1 inflammasome, IL1-β secretion, and pyroptosis, which is inhibited by YopK (31, 39). These studies suggest that T3SS-dependent LTB₄ synthesis by neutrophils may also require NLRP3/CASP1 inflammasome activation. However, LTB₄ synthesis in response to other stimuli is not dependent on inflammasome activation (11, 40, 41). Furthermore, infection of neutrophils with a strain of Y. pestis that lacks all the Yops except YopK does not inhibit LTB₄ synthesis in neutrophils (37, 38), raising an alternative possibility for inflammasome-independent mechanisms leading to T3SS-dependent LTB₄ synthesis. Moreover, Yop effectors other than YopK that have not been directly linked to inflammasome inhibition – YpkA, YopE, YopJ, or YopH – are sufficient to independently inhibit LTB₄ synthesis by neutrophils (37, 38), further supporting the latter hypothesis. Of these four Yop effectors, YopJ and YopH are intimately involved in MAPK and Ca²⁺ signaling required for LTB₄ synthesis. YopJ is an acyltransferase that targets several kinases in the MAPK pathway (42–45), and Pulsifer et al. demonstrated that YopJ inhibition of ERK phosphorylation is sufficient to inhibit LTB₄ synthesis by human neutrophils (37). YopH is a tyrosine phosphatase that has been shown to target multiple proteins of the focal adhesion complex, including SKAP2, SLP-76, PRAM, Vav, and LCK (28, 46–49). Through these interactions, YopH inhibits β1-integrin-mediated Ca²⁺ flux by neutrophils during interactions with Y. pseudotuberculosis, raising the possibility that this kinase signaling hub might be involved in LTB₄ synthesis in response to the T3SS. In this study, we use Y. pestis Yop mutants to define the molecular mechanisms responsible for T3SS-dependent LTB₄ synthesis and discovered significant differences in the host factors required for T3SS-sensing and LTB4 synthesis between neutrophils and macrophages.

Results

Neutrophil LTB₄ synthesis in response to Salmonella enterica Typhimurium is dependent on the SPI-1 T3SS

LTB₄ synthesis in response to Y. pestis is dependent on the expression of the T3SS and the YopB/D translocon (38). Moreover, S. enterica Typhimurium, which encodes two T3SSs (SPI-1 and SPI-2), also induces an LTB₄ response in neutrophils, but whether synthesis required expression of the T3SSs was not tested (50). To determine if leukocyte sensing of the T3SSs of S. enterica Typhimurium was responsible for LTB₄ synthesis, bone marrow derived murine neutrophils (BMNs) were infected with S. enterica Typhimurium LT2 (ST+) or a mutant lacking both the Salmonella pathogenicity island 1 (SPI-1) and SPI-2 encoded type 3 export apparatuses (ΔSPI1/2). After 1 h of infection, LTB₄ synthesis was significantly elevated in ST+ infected BMNs (Fig 1A, ST+ vs. UI, p≤0.0001) but it was not elevated in cells infected with the ΔSPI1/2 mutant (Fig 1A, p= 0.1513). To determine the contribution of the individual T3SSs, BMNs were next infected with individual SPI-1 or SPI-2 mutants. No significant differences in recovered LTB₄ were observed between cells infected with the ΔSPI1/2 and ΔSPI1 mutant strains, but LTB₄ concentrations were significantly elevated in the ΔSPI2 mutant infected cells (Fig 1A, ΔSPI1/2 vs. ΔSPI2, p≤0.0001). Together these data support that neutrophils not only sense the Y. pestis T3SS but also the SPI-1 T3SS of S. enterica Typhimurium to initiate a robust LTB₄ response.

Fig. 1. T3SS-depedent LTB₄ synthesis is conserved in S. enterica Typhimurium and BMDMs.

(A) Murine neutrophils (BMNs) or (C) Murine macrophages (BMDMs) were infected with S. enterica Typhimurium LT2 (ST+), mutants that lacked Salmonella T3SS SPI-1, SPI-2, or both SPI-1/2. (B) BMDMs were infected with Y. pestis, Y. pestis T3E, Y. pestis T3⁽⁻⁾, a yopB mutant in the T3E background (ΔB), or ΔB complemented with yopB (ΔB::B). (D) BMDMs infected with Y. pestis, Y. pestis T3E, Y. pestis T3⁽⁻⁾, or Y. pestis strains expressing only one Yop (+A = YpkA; +E = YopE; +H = YopH; +J = YopJ; +K = YopK; +M = YopM; or +T = YopT). (A) BMNs were infected at an MOI of 20 for 1 h. (B-D) BMDMs were infected at an MOI of 20 for 4 h. (A-D) LTB₄ was measured from supernatants by ELISA. Each symbol represents an independent biological infection, and the box plot represents the median of the group ± the range. UI = uninfected. ns = not significant. One-way ANOVA with Tukey’s post hoc test compared to each condition for A, C, and D, or Dunnett’s post hoc test compared to uninfected for B. *=p≤0.05, ***=p≤0.001, ****=p≤0.0001. For panel D, p values when compared to uninfected denoted as a=p≤0.05 or b=p≤0.0001, and when compared to Y. pestis T3E as c=p≤0.0001.

Requirement for interactions with the YopB/D translocase for LTB₄ synthesis is conserved in macrophages

While we have previously shown that LTB₄ synthesis by both neutrophils and M1-polarized macrophages in response to Y. pestis is dependent on the presence of the T3SS, we only demonstrated a requirement for the YopB/D translocase in neutrophils (38). To determine whether the YopB/D translocase is also required for LTB₄ synthesis by macrophages, M1-polarized bone marrow derived macrophages (BMDMs) were infected with Y. pestis, a Y. pestis strain expressing the T3SS but lacking all Yop effectors (Y. pestis T3E), a Y. pestis strain lacking the pCD1 plasmid encoding the entire Ysc T3SS [Y. pestis T3⁽⁻⁾], or a Y. pestis T3E yopB mutant that is defective in expression of the translocase that directly interacts with the host cell plasmid membrane (51–53). As previously observed for neutrophils (38), BMDMs also did not synthesize LTB₄ in response to Y. pestis T3⁽⁻⁾ or Y. pestis T3E yopB (Fig 1B). Normal LTB₄ synthesis was restored by yopB complementation (Fig 1B, yopB::cyopB). To confirm that the S. enterica Typhimurium SPI-1 T3SS is also required for LTB₄ synthesis by macrophages, BMDMs were infected with ST+ or the ΔSPI1/2, ΔSPI1, or ΔSPI2 mutants. As observed for BMNs, BMDM synthesis of LTB₄ was dependent on the presence of the SPI-1 T3SS but not SPI-2 system (Fig 1C). These data show that like neutrophils, macrophages respond to bacterial T3SSs by rapidly synthesizing LTB₄.

Only YopJ is sufficient to inhibit LTB₄ synthesis by macrophages

We have previously shown that YpkA, YopE, YopJ, or YopH are individually sufficient to inhibit LTB₄ synthesis in neutrophils (37, 38). To determine if these Yop effectors could inhibit LTB₄ synthesis by macrophages, LTB₄ was measured from BMDMs infected with Y. pestis strains that expressed only one Yop effector. BMDMs infected with strains expressing YpkA, YopE, YopH, YopK, and YopT showed significant decreases in LTB₄ compared to those infected with Y. pestis T3E, but still produced significantly more LTB₄ than Y. pestis infected or uninfected cells (Fig 1D). However, BMDMs infected with a strain expressing only YopJ appeared to be completely inhibited in their ability to synthesize LTB₄, producing concentrations similar to those recovered from cells infected with Y. pestis expressing all of the Yop effectors (Fig 1D; +J vs. Yp). YopM was the only effector that did not appear to have any impact on LTB₄ synthesis. When compared to the ability of individual Yop effectors to inhibit LTB₄ synthesis in neutrophils, these data suggest that the signaling pathways leading to LTB₄ in response to the T3SS may differ between macrophages and neutrophils.

Bacterial phagocytosis enhances LTB₄ synthesis by macrophages but not neutrophils

One important consequence of Yop intoxication of leukocytes is the inhibition of phagocytosis via the action of YpkA and YopE (54, 55). Moreover, an important function of the SPI-1 T3SS of S. enterica Typhimurium is to induce phagocytosis (56). Because phagocytosis is required for LTB₄ synthesis by leukocytes interacting with crystalline silica (11) and Y. pestis and the ΔSPI-1 mutant do not elicit LTB₄ synthesis in both neutrophils and macrophages, we next asked if phagocytosis of Y. pestis T3E or ST+ was required for T3SS-dependent LTB₄ synthesis by treating BMNs and BMDMs with the phagocytosis inhibitor cytochalasin D (cytoD) prior to infection. In BMNs, while treatment with cytoD inhibited phagocytosis of Y. pestis T3E (S1 Fig A–B), it did not alter the ability to synthesize LTB₄ (Fig 2A and B), indicating that Yop-mediated inhibition of phagocytosis is not responsible for the inhibition of T3SS-mediated LTB₄ synthesis during Y. pestis infection in neutrophils. Additionally, as previously reported by Golenkina et al. (50), cytoD treatment increased LTB₄ synthesis by BMNs infected with ST+ (Fig 2C; p≤0.05), indicating that induction of phagocytosis by S. enterica Typhimurium is not required for T3SS-mediated LTB₄ synthesis. However, in contrast to neutrophils, when phagocytosis was inhibited in BMDMs (S1 Fig C–D), LTB₄ synthesis was reduced in response to both Y. pestis T3E (Fig 2E; p≤0.05) and ST+ (Fig 2F; p≤0.059). Interestingly, LTB₄ levels were still higher in the cytoD infected BMDMs than the uninfected BMDMs (Fig. 2D). Together, these data suggest that phagocytosis is not required for T3SS-mediated LTB₄ synthesis by neutrophils but enhances synthesis by macrophages.

Fig. 2. Phagocytosis is not required for LTB₄ synthesis in BMNs but enhances LTB₄ synthesis by macrophages in response to Y. pestis.

(A-C) BMNs or (D-F) BMDMs were untreated (green circles) or pretreated with cytochalasin D (10 μM, purple circles) for 30 min and were either (A,D) uninfected or infected with (B,E) Y. pestis T3E or (C,F) ST at an MOI of 20 for (A-C) 1 h or (D-F) 4 h. (A-F) LTB₄ was measured from supernatants by ELISA. Each symbol represents an independent biological infection, and the box plot represents the median of the group ± the range. UI = uninfected. ns = not significant. T-test with Welch’s post hoc test. *=p≤0.05.

PLC signaling is required for LTB₄ synthesis in neutrophils but not macrophages

Ca²⁺ flux is required for the activation of cPLA₂ and 5-LOX (7, 14). However, it is unclear if the T3SS induces Ca²⁺ flux through Ca²⁺ migration through the YopB/D translocase pore or via conventional Ca²⁺ signaling. Phospholipase C (PLC) is the central mediator of conventional Ca²⁺ signaling in the cell (57–60), and chemical inhibitors of PLC have been well characterized (61, 62). Therefore, to determine if Ca²⁺ signaling is required for T3SS-dependent LTB₄ synthesis, leukocytes were treated prior to infection with U73122, which inhibits PLCβ and PLCγ (61, 63, 64). When PLC signaling was inhibited, BMNs infected with the Y. pestis T3E mutant were no longer able to synthesize LTB₄ compared to untreated BMNs (Fig 3A), suggesting that PLC-mediated Ca²⁺ signaling is required for LTB₄ synthesis. To ensure that U73122 treatment did not have off target effects on cPLA₂ or 5-LOX, U73122-treated BMNs were incubated with the Ca²⁺ ionophore, A23187, which induces Ca²⁺ flux and LTB₄ synthesis independent of PLC signaling (65). Within 10 min of A23187 treatment, U73122-treated BMNs were able to synthesize LTB₄, supporting that U73122 treatment is not affecting the activity of cPLA₂ or 5-LOX (Fig 3A). In contrast to BMNs, U73122 treatment of Y. pestis T3E-infected BMDMs only modestly inhibited LTB₄ synthesis compared to untreated cells (Fig 3B; p≤0.05), suggesting PLC is not the primary source of Ca²⁺ flux in macrophages. Together, these data suggest that the T3SS activates PLC-mediated Ca²⁺ flux in neutrophils needed for LTB₄ synthesis, but alternative mechanisms are required for T3SS-induced LTB₄ synthesis in macrophages.

Fig. 3. PLC signaling is required for LTB₄ synthesis in BMNs.

(A) BMNs or (B) BMDMs were infected with Y. pestis, Y. pestis T3E, or Y. pestis T3⁽⁻⁾ at an MOI of 20 for (A) 1 h or (B) 4 h. Leukocytes pretreated with PLC inhibitor (U73122; (A) 5 μM or (B) 20 μM) for 30 min (purple circles). (A) BMN supernatants were replaced with fresh media and cells were treated with A23187 (1 μM) for 10 min after treatment with U73122 for 30 min (orange circles). (A-B) LTB₄ was measured from supernatants by ELISA. Each symbol represents an independent biological infection, and the box plot represents the median of the group ± the range. UI = uninfected. One-way ANOVA with Tukey’s post hoc test compared to each condition. *=p≤0.05, ****=p≤0.0001.

STIM1-mediated flux of extracellular Ca²⁺ is required for LTB₄ synthesis in both neutrophils and macrophages

PLC-mediated Ca²⁺ flux is required for LTB₄ synthesis in neutrophils, but PLC signaling can lead to both Ca²⁺ eflux from the ER and influx from the extracellular space (57–60, 66). To determine if intracellular Ca²⁺ eflux is sufficient to induce T3SS-dependent LTB₄ synthesis, BMNs were pre-treated with EGTA to chelate extracellular Ca²⁺ prior to infection with Y. pestis T3E – if intracellular Ca²⁺ eflux is sufficient for cPLA₂ and 5-LOX activation, then EGTA should not inhibit LTB₄ synthesis. As observed during PLC inhibition, EGTA chelation of extracellular Ca²⁺ significantly reduced LTB₄ production compared to untreated BMNs (Fig 4A; p≤0.0001). Influx of extracellular Ca²⁺ also requires the cell to maintain a membrane potential by eflux of intracellular potassium (K⁺) (67, 68). Therefore, if extracellular Ca²⁺ is required, disrupting the K⁺ gradient should also inhibit LTB₄ synthesis. As predicted by EGTA treatment, increasing the extracellular K⁺ concentration significantly inhibited LTB₄ synthesis by Y. pestis T3E-infected BMNs (Fig 4A; p≤0.0001). Finally, we also treated BMNs with a pharmacological inhibitor of STIM1 (SKF), which activates SOC-mediated Ca⁺ flux across the plasma membrane (69), and observed a significant decrease in LTB₄ synthesis (Fig 4A; p≤0.0001). Together, these data support that that extracellular Ca²⁺ flux is required for the T3SS-dependent LTB₄ synthesis, and it is mediated by the PLC-STIM1 pathway in neutrophils. Interestingly, while PLC does not appear to be required for LTB₄ synthesis by macrophages (Fig. 3), there remains a requirement for STIM1 and extracellular Ca²⁺ flux (Fig. 4B), further indicating that alternative pathways are involved in triggering Ca²⁺ flux needed for LTB₄ synthesis in macrophages.

Fig. 4. Influx of extracellular Ca²⁺ is required for Y. pestis T3SS-dependent LTB₄ synthesis.

(A) BMNs or (B) BMDMs were infected with Y. pestis, Y. pestis T3E, or Y. pestis T3⁽⁻⁾. Leukocytes pretreated with EGTA (1 mM; purple circles) for 30 min, with (A) 50 mM or (B) 100 mM KCl (orange circles) for 30 min, or with STIM1 inhibitor (SKF, 50 μM; white circles) for 2 min prior to infection with Y. pestis T3E. Leukocytes were infected at an MOI of 20 for (A) 1 h or (B) 4 h. LTB₄ was measured from supernatants by ELISA. Each symbol represents an independent biological infection, and the box plot represents the median of the group ± the range. UI = uninfected. One-way ANOVA with Tukey’s post hoc test compared to each condition. ****=p≤0.0001.

SKAP2 is required for LTB₄ synthesis by neutrophils but not macrophages

YopH inhibits PLC-mediated Ca²⁺ flux in neutrophils during Y. pseudotuberculosis infection by modifying proteins of the SKAP2/SLP-76/PRAM/Vav signaling hub (28, 48, 49). Because YopH independently inhibits LTB₄ synthesis in neutrophils, we sought to determine if this hub is also required for T3SS-dependent LTB₄ synthesis by measuring LTB₄ synthesis by leukocytes from SKAP2^−/− mice infected with Y. pestis, Y. pestis T3E, or Y. pestis T3⁽⁻⁾. Unlike BMNs from C57BL/6J mice, SKAP2^−/− BMNs did not synthesize LTB₄ in response to any of the strains tested (Fig 5A). To ensure that SKAP2^−/− BMNs were not generally defective in LTB₄ synthesis, SKAP2^−/− BMNs were treated with the Ca²⁺ ionophore A23187, and cells robustly produced LTB₄ (Fig 5B). Complementing the PLC inhibitor data, SKAP2^−/− BMDMs were not impaired in LTB₄ synthesis (Fig 5C; p≤0.0001). Together, these data demonstrate that T3SS-dependent LTB₄ synthesis by neutrophils requires SKAP2 but is independent of SKAP2 in macrophages.

Fig. 5. SKAP2 signaling required for LTB₄ synthesis in BMN. SKAP2^−/−.

(A) BMNs or (C) BMDMs were infected with Y. pestis, Y. pestis T3E, or Y. pestis T3⁽⁻⁾ at an MOI of 20 for (A) 1 h or (C) 4 h. (B) SKAP2^−/− BMNs were treated with A23187 (1 μM; purple circles) for 1 h. (A-C) LTB₄ was measured from supernatants by ELISA. Each symbol represents an independent biological infection, and the box plot represents the median of the group ± the range. UI = uninfected. One-way ANOVA with Dunnett’s post hoc test compared to uninfected. ****=p≤0.0001.

Activation of MAPK signaling required for LTB₄ synthesis is independent of the T3SS

In addition to Ca²⁺ flux, LTB₄ synthesis requires MAPK signaling to phosphorylate cPLA₂ and 5-LOX (7, 70–72). While we previously showed that p38 and ERK1/2 are phosphorylated in human neutrophils infected with a high MOI (100 bacteria/cell) of Y. pestis T3⁽⁻⁾ (37), the MAP kinases responsible for T3SS-dependent LTB₄ synthesis have not been defined. Therefore, to determine whether p38 and ERK1/2 are phosphorylated during interactions with Y. pestis T3E, leukocytes were infected with Y. pestis, Y. pestis T3E, or Y. pestis T3⁽⁻⁾ at an MOI of 20. Both p38 and ERK1/2 were phosphorylated in BMNs infected with Y. pestis T3E but not Y. pestis (Fig 6A–B). For BMDMs, we observed elevated basal levels of p38 and ERK1/2 phosphorylation in uninfected cells, which has been reported in M1 polarized macrophages by others (73–76), but phosphorylation of both kinases was significantly lower in Y. pestis infected cells (Fig 6C–D). While p38 phosphorylation was slightly elevated in BMDMs infected with Y. pestis T3E, ERK1/2 phosphorylation appeared to decrease slightly, suggesting that p38 phosphorylation may be driving LTB₄ synthesis in BMDMs. However, in both leukocytes we observed similar trends in phosphorylation in cells infected with Y. pestis T3E or Y. pestis T3⁽⁻⁾, indicating that activation of MAPK signaling is in response to a PAMP unrelated to the T3SS.

Fig. 6. Activation of MAPK signaling required for LTB₄ synthesis is independent of the T3SS.

(A-B) BMNs or (C-D) BMDMs were infected with Y. pestis, Y. pestis T3E, Y. pestis T3⁽⁻⁾ at an MOI of 20 for (A-B) 1 h or (C-D) 30 min. (A-B) BMNs were infected Y. pestis TE3 expressing only one Yop effector (+H = YopH; +J = YopJ) at an MOI of 20 for 1 h. Densitometry and representative WB images for (A,C) p-p38 or (B,D) p-ERK1/2 from whole cell lysates normalized to beta actin. (A-D) Each symbol represents an independent biological infection, and the box plot represents the median of the group ± the range. UI = uninfected. One-way ANOVA with Dunnett’s post hoc test compared to uninfected. *=p≤0.05, **=p≤0.01, ****=p≤0.0001.

YopH inhibits ERK phosphorylation in neutrophils

In BMNs, YopH and YopJ can independently inhibit LTB₄ synthesis. However, both effector proteins have different targets within the host cell, which can lead to inhibiting both Ca²⁺ and/or MAPK signaling (22, 29). While we have shown that YopJ inhibition of ERK1/2 phosphorylation is sufficient to block LTB₄ synthesis by human neutrophils (37), and Shaban et al. previously showed that YopH from Y. pseudotuberculosis inhibits ERK1/2 phosphorylation in neutrophils (49), whether YopH can also sufficiently inhibit ERK1/2 phosphorylation in our model has not been tested. Therefore, we measured ERK1/2 and p38 phosphorylation in BMNs infected with Y. pestis T3E strains expressing only YopH or YopJ. As predicted from our work with human neutrophils (37), phosphorylation of both MAP kinases was inhibited in BMNs infected with the YopJ expressing strain (Fig 6A and B; +J). However, infection with the YopH expressing strain appeared to only inhibit ERK1/2 phosphorylation and did not significantly impact p38 phosphorylation (Fig 6A and B; +H samples). These data suggest that YopH can block both Ca²⁺ and ERK1/2 phosphorylation in neutrophils during interactions with Y. pestis.

Inflammasome activation enhances LTB₄ synthesis in macrophages but not neutrophils

Previous studies with Y. pseudotuberculosis indicate that the T3SS translocase is recognized by NLRP3, leading to activation of the caspase 1 inflammasome and pyroptosis (39, 77). Inflammasome activation has been linked to LTB₄ synthesis in response to some PAMPS but is dispensable for others (11, 40, 41). Therefore, to determine if the NLRP3/CASP1 inflammasome contributes to LTB₄ synthesis in response to the T3SS, leukocytes isolated from NLRP3^−/− or CASP1/11^−/− mice were infected with Y. pestis, Y. pestis T3E, or Y. pestis T3⁽⁻⁾, and LTB₄ synthesis was compared to cells isolated from WT mice. Absence of NLRP3 or CASP1/11 did not diminish the neutrophil response to Y. pestis T3E (Fig 7A). Furthermore, treatment of Y. pestis T3E infected BMNs or human PMNs with the pan-caspase inhibitor zVAD also did not decrease LTB₄ synthesis (Fig 7B and C). In contrast, LTB₄ synthesis was significantly lower in both NLRP3 and CASP1/11 deficient BMDMs (Fig 7D; p≤0.0001). However, LTB₄ synthesis by Y. pestis T3E infected cells was still greater than uninfected, Y. pestis, or Y. pestis T3⁽⁻⁾ infected cells, indicating that the NLRP3/CASP1 inflammasome enhances the LTB₄ response by macrophages but not neutrophils. Because CASP1/11^−/− macrophages still synthesized lower levels of LTB₄ in response to the Y. pestis T3E strain, we next asked if this synthesis was dependent on PLC signaling, and observed only a moderate decrease in LTB₄ when infected cells were treated with the PLC inhibitor U73122 (Fig 7E). Interestingly, we observed a much more significant decrease if we inhibited phagocytosis by treatment with cytoD or during infection with a Y. pestis T3E strain expressing only YopE, which is a potent inhibitor of phagocytosis (Fig 7E; p≤0.0001), suggesting that phagocytosis is key to initiating LTB₄ synthesis in macrophages.

Fig. 7. Inflammasomes enhance LTB₄ synthesis in BMDMs but are dispensable in BMNs.

(A) BMNS or (D) BMDMs from WT, NLRP3^−/−, CASP1/11^−/− mice were infected with Y. pestis, Y. pestis T3E, or Y. pestis T3⁽⁻⁾. (B) C57BL/6J BMNs or (C) human PMNs (hPMNs) pretreated with zVAD inhibitor (100 μM; purple circles) for 30 min prior to infection with Y. pestis T3E. (E) BMDMs from CASP1/11^−/− mice were infected with Y. pestis T3E or Y. pestis T3E strain expressing only YopE (+E). BMDMs were either left untreated (green circles) or pretreated with PLC inhibitor (U73122, 20 μM; purple circles) or Cytochalasin D (10 μM; purple circles) for 30 min prior to infection with Y. pestis T3E. Leukocytes were infected at an MOI of 20 for (A-C) 1 h or (D-E) 4 h. (A-E) LTB₄ was measured from supernatants by ELISA. Each symbol represents an independent biological infection, and the box plot represents the median of the group ± the range. UI = uninfected. UT = untreated. ns = not significant. One-way ANOVA with Tukey’s post hoc test compared to each condition for A-D., or with Dunnett’s post hoc test compared to untreated/T3E for E. *=p≤0.05, **=p≤0.01, ***=p≤0.001, ****=p≤0.0001.

Discussion

Establishing a non-inflammatory environment during the early stages of plague is crucial for the progression of disease (20). We and others have shown that Y. pestis subverts the host innate immune response by inhibiting leukocyte chemotaxis (78), phagocytosis (32, 33), neutrophil degranulation (36, 37), neutrophil ROS production (33, 35), and inflammatory lipid, cytokine, and chemokine release (34, 37, 38). Despite the T3SS being a PAMP, the Yop effectors are highly efficient at preventing immune cell activation, including inhibiting the synthesis of LTB₄ needed for a proper inflammatory host response (31, 37, 38). In this study, we sought to understand how the molecular signaling pathways stimulated by the T3SS lead to the synthesis of LTB₄ and to define how the pathogen uses specific Yop effectors to block this response. Using these data, we have developed a working model showing a differential response to the T3SS between neutrophils and macrophages needed for LTB₄ synthesis – with the SKAP2/PLC signaling pathway as key for a rapid response in neutrophils and the NLRP3/CASP1 inflammasome more important in macrophages (Fig 8).

Fig. 8. T3SS triggered LTB₄ synthesis differs between neutrophils and macrophages.

LTB₄ synthesis requires an increase in intracellular Ca²⁺ and activation of MAPK signaling. Activation of MAPK signaling required for LTB₄ synthesis appears to be independent of the T3SS, and instead is initiated by a currently unknown PAMP and signaling pathway(s) (Signal 1). In response to the T3SS (Signal 2), neutrophils require Ca²⁺ signaling through SKAP2/PLC/STIM1 to produce LTB₄. Macrophages require Ca²⁺ influx from the extracellular space, with only a partial requirement of STIM1 but not SKAP2 or PLC. Inflammasome activation is not required in neutrophils for LTB₄ synthesis but enhances the induction in macrophages. Phagocytosis also enhances LTB₄ induction in macrophages, but not in neutrophils. Red boxes indicate factors involved in LTB₄ synthesis examined in this study. Blue boxes indicate potential or unknown factors they may be involved but have yet to be tested.

One important discovery made here was that recognition of the T3SS leading to LTB₄ synthesis is not specific to the Y. pestis T3SS. While the S. enterica Typhimurium SPI-1 T3SS varies structurally from the Y. pestis secretion system (79), S. enterica Typhimurium still triggers LTB₄ in a SPI-1 T3SS-dependent manner (Fig 1). Additionally, LTB₄ synthesis by S. enterica Typhimurium infected leukocytes appears to be more responsive to the SPI-1 T3SS than the SPI-2 system. Like the Y. pestis T3SS, the SPI-1 T3SS engages with the host cell through the plasma membrane, while SPI-2 engages through the Salmonella containing vacuole (80, 81), suggesting that the host cells have evolved to sense T3SS interactions across the plasma membrane and respond by synthesizing LTB₄. This scenario is attractive, as interactions with the plasma membrane represent the earliest interaction that leukocytes would have with pathogens and allow for rapid synthesis of the lipid as soon as contact is made. It also appears that unlike Y. pestis, S. enterica Typhimurium has not evolved effector proteins to inhibit this response, or at minimum, not to the degree that the Y. pestis Yop effectors can inhibit LTB₄ synthesis, as the WT S. enterica Typhimurium LT strain produces LTB₄ at levels similar to the Y. pestis T3E mutant. This difference supports that the inhibition of LTB₄, and initiation of the inflammatory cascade is an important aspect in the virulence and lifestyle of Y. pestis that is not required for S. enterica Typhimurium.

A key virulence strategy mediated by the T3SSs of both Y. pestis and S. enterica Typhimurium is the manipulation of phagocytosis (46, 82). Because phagocytosis of crystalline silica is required for LTB₄ synthesis (11), determining the contribution of phagocytosis to LTB₄ synthesis was another critical aspect of understanding the leukocyte response to these pathogens. In neutrophils, phagocytosis was not required for LTB₄ synthesis, and cytochalasin D treatment induced even greater LTB₄ synthesis in response to S. enterica Typhimurium. However, inhibiting phagocytosis reduced LTB₄ synthesis by macrophages, providing the first evidence that the host cell molecular mechanisms leading to cPLA₂ and 5-LOX activation and LTB₄ synthesis differ between these two cell types. It is important to note that for both bacteria, LTB₄ synthesis by macrophages was not completely inhibited by cytochalasin D treatment, indicating that synthesis is not completely dependent on phagocytosis in macrophages. Moreover, data from infections with the Y. pestis T3⁽⁻⁾ mutant, which is as readily phagocytosed as the Y. pestis T3E mutant, indicates that phagocytosis alone is not sufficient to trigger LTB₄ synthesis in the absence of the T3SS.

Ca²⁺ flux is a critical step in the enzyme activation required for LTB₄ synthesis (7, 14), and thus understanding the mechanisms leading to Ca²⁺ flux during host cell interactions with the T3SS expressing bacteria is key to understanding how cells are sensing this PAMP. YopB and YopD insert into the plasma membrane to form a pore needed for effector translocation into the host cell (51–53). Others have shown that the pore formed by the translocase can result in the diffusion of molecules larger than Ca²⁺, but YopN appears to limit diffusion through the YopB/D translocase (83). While the Y. pestis T3E strain used in these studies retains YopN, it is still possible that Ca²⁺ diffusion through the translocase could occur in infected cells. However, using pharmacological inhibitors, we have shown that PLC and store-operated Ca²⁺ signaling, and not diffusion of Ca²⁺ through the translocase pore, is likely the primary driver of Ca²⁺ flux needed for LTB₄ synthesis in neutrophils. This is further supported by the ability of YopH to inhibit LTB₄ synthesis, which blocks Ca²⁺ flux by inhibiting PLC phosphorylation and signaling and not ion diffusion through the translocase (28, 49), and the requirement for SKAP2, which activates PLC in response to signaling from a variety of tyrosine kinase receptors (84, 85). Identifying which tyrosine kinase receptors are phosphorylating SKAP2 in response to the T3SS is an ongoing research direction for us and will help us to better understand the mechanisms directly sensing the T3SS and/or YopB/D translocase. Contrarily, because PLC pharmacological inhibition and YopH had minimal effects on LTB₄ synthesis in macrophages, the contribution of Ca²⁺ diffusion through the translocase in macrophages is less clear. It appears that LTB₄ synthesis in macrophages still requires activation of STIM1 and store-operated calcium channels (Fig. 5), indicating that diffusion of Ca²⁺ through the pore is not sufficient to activate cPLA₂ and 5-LOX in macrophages, but also that the signal to activate STIM1 is independent of PLC. PLC-independent activation STIM1 has been reported, but activation requires Ca²⁺ store depletion from the ER (86, 87), suggesting that any Ca²⁺ diffusion via the translocase pore is not directly responsible for STIM1 activation. Currently it is unclear how STIM1 is activated during macrophage interactions with the Y. pestis T3E strain and additional research is needed to completely understand the PLC-independent pathway(s) activated in macrophages.

In addition to Ca²⁺ flux, cPLA₂ and 5-LOX phosphorylation via MAPK signaling is also essential for LTB₄ synthesis (7, 70–72). Depending on the stimulus, 5-LOX phosphorylation is mediated by p38, ERK1/2, or JNK (11, 70, 88, 89), but while p38 and ERK1/2 are both phosphorylated in human neutrophils during interactions with Y. pestis T3E, only ERK1/2 was essential for LTB₄ synthesis (37). While we confirmed that both p38 and ERK1/2 are also phosphorylated in murine neutrophils and macrophages (Fig 6), we also discovered that MAPK activation is not dependent on the presence of the T3SS, as we observed significant phosphorylation of the kinases in cells infected with the Y. pestis T3⁽⁻⁾ strain that lacks the pCD1 plasmid and the entire T3SS. These data indicate that a different PAMP(s) is triggering the MAPK pathway and that the host factors recognizing the T3SS and triggering Ca²⁺ flux differ from those required for MAPK phosphorylation. A requirement for two signals is an important check point for other responses to bacteria, for example inflammasome activation, to limit premature inflammatory responses that could be detrimental to the host if stimulated too easily. Thus, to avoid chronic inflammation, a similar two signal check point may have also evolved to limit premature LTB₄ responses to bacteria. It is also currently unclear if this is specific for bacteria with T3SSs. Studies with other pathogens are needed to determine if multiple PAMPs/signals are also required for LTB₄ responses to bacteria that do not express T3SSs. Importantly, these data support that the primary signal governing whether LTB₄ is produced in recognition of the T3SS is Ca²⁺ flux, not the MAPK activation.

While BLT1-LTB₄ signaling has been shown to enhance inflammasome activation in gout (90), asthma (6), and Staphylococcus aureus skin infection models (91), evidence in the literature that inflammasome activation induces LTB₄ synthesis is limited. However, Von Moltke et al. showed that direct delivery of the NAIP5 ligand flagellin to the cytosol of resident peritoneal macrophages, but not M2 polarized BMDMS, triggers leukotriene and prostaglandin synthesis, which was dependent on NLRC4, NAIP5, and CASP1 (40). Moreover, they demonstrated that resident peritoneal macrophages produce prostaglandin E2 (PGE2) in response to S. enteric Typhimurium, but they did not report LTB₄ synthesis or test dependence on the SPI1 T3SS (40). The new data reported here further support that polarization or activation state significantly impacts the potential of macrophages to synthesize eicosanoids during infection. Moreover, we have now identified a second example of a PAMP that triggers eicosanoid synthesis in an inflammasome-dependent manner and supports that inflammasome activation not only results in the release of protein mediators of inflammation (i.e., IL-1β and IL-18) but also lipid mediators of inflammation. The mechanisms leading to T3SS-dependent inflammasome-mediated LTB₄ synthesis are still unclear, but the MAPK data suggests that it is involved in triggering Ca²⁺ flux, either directly or through the activation of STIM1 and the store-operated calcium channels.

In conclusion, we have shown that leukocytes have evolved to recognize the T3SS as part of two signal cascade that leads to LTB₄ synthesis during bacterial interactions. However, the molecular mechanisms resulting in LTB₄ synthesis differ between neutrophils and macrophages, the former dependent on the SKAP2/PLC signaling pathway and the latter on the inflammasome. Together, these data provide us with a better understanding of the early response of leukocytes to bacterial pathogens and provide an example of how a pathogen, in this case Y. pestis, has evolved mechanisms to subvert these responses and evade the immune recognition to cause disease.

Material and Methods

Ethics statement

All animal work was approved by the University of Louisville Institutional Animal Care and Use Committee (IACUC Protocol #22157). Use of human neutrophils was approved by the University of Louisville Institutional Review Board guidelines (IRB #96.0191) and written consents for use were obtained.

Bacterial strains

Bacterial strains used in this study are listed in S1 Table. Y. pestis was cultured with BHI broth for 15–18 h at 26°C in aeration. Cultures were then diluted 1:10 in fresh, warmed BHI broth containing 20 mM MgCl₂ and 20 mM Na-oxalate and cultured at 37°C for 3 h with aeration to induce expression of the T3SS. S. enterica Typhimurium was cultured with LB broth for 15–18 h at 37°C in aeration. Cultures were then diluted 1:10 in fresh, warmed LB broth and cultured at 37°C for 3 h with aeration to reach logarithmic growth. Bacterial concentrations were determined using a spectrophotometer and diluted to desired concentrations in fresh medium.

Cell isolation and cultivation

Murine leukocytes were isolated from bone marrow of 7–12-week-old mice that were either C57BL/6J, C57BL/6J Tyr^−/− (B6 albino; Jackson Laboratory 000058), C57BL/6J Tyr^−/− NLRP3^−/−, C57BL/6J Tyr^−/− Caspase1/11^−/−, or BALB/c SKAP2^−/−.Murine neutrophils were isolated using an Anti-Ly-6G Microbeads kit (Miltenyi Biotec Cat. No. 130–120-337) per the manufacturer’s instructions. Neutrophil isolations yielded ≥ 95% purity and were used within 1 h of isolation. Macrophages were differentiated from murine bone marrow in DMEM supplemented with 1 mM Na-pyruvate, and 10% FBS for 6 days. Macrophages were polarized with 20 ng/mL of GM-CSF (Kingfisher Biotech Cat. No. RP0407M) throughout the differentiation. The medium was replaced on days 1 and 3 (adapted from (92)). Human neutrophils were isolated from the peripheral blood of healthy, medication-free donors, as described previously (93). Neutrophil isolations yielded ≥ 95% purity and were used within 1 h of isolation.

Leukocyte infections

Human neutrophils were cultured in Kreb’s buffer (w/ Ca²⁺ & Mg), mouse neutrophils (BMNs) were cultured in RPMI + 5% FBS, and macrophages (BMDMs) were cultured in DMEM + 10% FBS. Human neutrophils were adhered to 24-well plates for 30 min that were coated with pooled human serum prior to infection. BMNs were adhered to 24-well plates for 30 min that were coated with FBS prior to infection. Plates were washed twice with 1 × DPBS prior to plating the cells. BMDMs were adhered to 24-well plates 1 day prior to infection. Leukocytes were infected at a multiplicity of infection (MOI) of 20 and incubated for 1 h (neutrophils) or 4 h (macrophages) at 37°C with 5% CO₂. All infections were synchronized by centrifugation (200 × g for 5 min). Supernatants and cells were then collected from wells to an Eppendorf tube and centrifuged for 1 min at 6,000 × g. Supernatants devoid of cells were transferred to a fresh tube and stored at −80°C until ELISA. When applicable, cell pellets were directly prepped for western blot analysis.

Treatments and inhibitors

Prior to infection, leukocytes were treated with the following for the times and concentrations indicated in the figure legends: phagocytosis inhibitor cytochalasin D (VWR; Cat. No. 100507–376), calcium ionophore A23187 (Sigma-Aldrich; Cat. No. C7522), PLC inhibitor U73122 (Abcam; Cat. No. ab120998), STIM1 inhibitor SKF-96365 (VWR; Cat. No. 89156–792), 1 mM EGTA, 50 mM or 100 mM KCl, or the pan-caspase inhibitor Z-Vad-FMK (Enzo; Cat. No. ALX-260–020). At the time of infection, bacteria were added for a 500 μL final volume.

Measurement of LTB₄ by enzyme-linked immunosorbent assay

Supernatants of neutrophils and macrophages were collected and measured for LTB₄ by ELISA per manufacturer’s instructions (Cayman Chemicals; Cat. No. 502390).

Western blots

Pellets were lysed over ice in 1x Novex lysis buffer and processed through Qiashredders (Qiagen; Cat. No. 79654). Samples were boiled for 10 min, and 10 μL was separated on a 10% SDS-PAGE gel. Samples were immunoblotted with polyclonal anti-p-p38 (Cell Signaling; Cat. No. 9211S), anti-p38 (Cell Signaling; Cat. No. 9228), anti-p-p44/42 (ERK1/2) (Cell Signaling; Cat. No. 9101s), anti-p44/42 (Cell Signaling; Cat. No. 4696), anti-beta-actin (Cell Signaling; Cat. No. 3700s). All antibodies were diluted 1:1000 or except anti-p44/42, which was diluted 1:2,000. Anti-rabbit (Sigma-Aldrich; Cat. No. A9169) or anti-mouse (ThermoFisher Scientific; Cat. No. 31430) IgG HRP secondary antibodies were diluted to 1:20,000. SuperSignal West Femto maximum-sensitivity substrate (ThermoFisher Scientific; cat. no. 34095) was used to detect antigen-antibody binding. Densitometry was performed using ImageJ software to quantify bands, normalized to beta actin.

Confocal microscopy

BMNs infected at an MOI of 10 with a GFP expressing Y. pestis T3E strain were pretreated with 10 μM cytochalasin D or DPBS. After 1 h of infection, cells were then fixed with 4% PFA (Sigma Aldrich; cat. no. P6148–500G), blocked with 3% BSA-PBS (Sigma; A4503–100G), stained with rabbit anti-Yersinia pestis sera (1:1,000; lot UL25, 9/14/2013) overnight at 4°C, followed by donkey anti-rabbit Alexa Fluor 647 (1:1,000; JacksonImmuno Research; cat. no. 711–605-152) for 2 hours at room temperature, and finally with Hoechst (1:350; cat. no. ThermoScientific; 62249) at room temperature for 15 minutes. Cells were then mounted in Prolong Gold (Invitrogen; cat. no. P36980) and visualized with z-stack images using a confocal Olympus Fluoview FV3000 UPlanxApo. To quantify the rates at which bacteria were phagocytosed, 3D volume Pearson correlation coefficients were calculated for eGFP and Alexa647.

Statistics

For all studies, male and female mice or human donors were used and no sex biases were observed for any phenotype. For all experiments, each data point represents data from biologically independent experiments performed on different days. Where appropriate and as indicated in the figure legends, statistical comparisons were performed with Prism (GraphPad) using one-way analysis of variance (ANOVA) with Dunnett’s or Tukey’s post hoc test, or T-test with Mann-Whitney’s post hoc test. P values ≤ 0.05 were considered statistically significant and reported.

Significance.

The production of inflammatory lipid mediators by the host is essential for timely inflammation in response to invasion by bacterial pathogens. Therefore, defining how immune cells recognize pathogens and rapidly produce these lipids is essential for us to understand how our immune system effectively controls infection. In this study, we discovered that the host signaling pathways required for leukotriene B4 (LTB₄) synthesis differ between neutrophils and macrophages, highlighting important differences in how immune cells respond to infection. Together, these data represent a significant improvement in our understanding of how neutrophils and macrophages rapidly react to bacteria and provide new insights into how Yersinia pestis manipulates leukocytes to evade immune recognition to cause disease.