|Home | About | Journals | Submit | Contact Us | Français|
We previously have shown that MyD88 is important for uptake of B. burgdorferi by bone marrow derived macrophages (BMDMs). The mechanism by which MyD88 is involved in uptake of B. burgdorferi is currently is not well characterized. Here, we report that MyD88-mediated defect in the phagocytosis of B. burgdorferi can be complemented by TLR3/TRIF activation in BMDMs from MyD88−/− mice. This effect of TLR3/TRIF activation was not due to its induction of type I interferons, suggesting instead a convergence of signaling pathways downstream of MyD88 and TRIF. In order to characterize signaling pathways involved in MyD88-mediated phagocytosis of B. burgdorferi, BMDMs were treated with specific inhibitors of MAPK, PKC, JAK/STAT, or PI3K. Only inhibition of PI3K resulted in a significant decrease of B. burgdorferi uptake. Consistent with this, B. burgdorferi activation of MyD88- or TLR3-/TRIF-signaling resulted in increased activity of PI3K. In addition, association of B. burgdorferi with actin related protein (Arp2/3) complexes, which facilitate actin rearrangements during phagocytosis, was similarly reduced in MyD88−/− BMDMs and BMDMs treated with a PI3K inhibitor. Taken together, these findings define an essential pathway whereby downstream signals from MyD88 or TRIF converge on PI3K, which triggers actin polymerization to initiate the phagocytosis of B. burgdorferi.
Phagocytosis is a process by which macrophages, dendritic cells, and neutrophils internalize particles. It is an essential component of the innate immune system in defending against microbial pathogens (1). Upon contact with a microbe, phagocytic cells activate complex signaling networks that trigger and coordinate cellular processes as diverse as cytoskeletal rearrangement, alterations in membrane trafficking, activation of microbial killing mechanisms, production of pro- and anti-inflammatory cytokines and chemokines, activation of apoptosis, and production of molecules required for efficient antigen presentation to the adaptive immune system.
Toll-like receptors (TLRs) are pattern-recognition receptors for microbial products that participate in orchestrating the innate immune response to various microbial organisms (2, 3). In addition to activating the release of inflammatory cytokines, TLR signaling has been shown to play an important role in uptake (4–7), phagolysosomal maturation (8) and nitric oxide formation (9) in response to various organisms. The role that TLR signaling plays in phagocytosis may vary depending upon the organism involved and the signaling pathways that are activated. For example, signaling through the LPS receptor, TLR4, has been shown to play an important role in activation of p38 MAPK that is important for phagolysosomal maturation (8).
For Borrelia burgdorferi, the causative agent of Lyme disease, it appears that signaling through the TLR adaptor MyD88 plays an important role in the uptake of the organism (4). Of note, in contrast to E. coli, activation of p38 signaling does not play a major role in the uptake and processing of the organism. Loss of signaling through TLR2, TLR5 or TLR9—each of which is capable of recognizing a specific B. burgdorferi product (lipoproteins, flagellin or CpG DNA respectively) and signals through MyD88—is not sufficient to inhibit phagocytosis of B. burgdorferi (4). It is certainly possible that the phagocytic effects involving MyD88 are mediated through a different TLR that recognizes some other unknown borrelial product or that activation through any one of a number of TLRs is sufficient to activate MyD88 dependent phagocytosis. However, another possibility is that MyD88 signaling is not required for phagocytosis and that defects in uptake seen with MyD88 deficiency are due to developmental defects or a decreased activation state. This hypothesis has been proposed by Yates and Russell et al. (10) where they showed the requirement of MyD88 for phagolysosomal maturation, regardless of the presence of any TLR stimulus.
We are interested in determining the mechanism by which MyD88-mediated signaling plays a role in the uptake of B. burgdorferi. As we will show, the defect in phagocytosis of B. burgdorferi in MyD88−/− cells is not due an intrinsic maturational defect or activation state, but instead is due to a lack of activation of a specific signaling pathway, which can be complemented by activation through an alternative pathway. Here we present our results, identifying the mechanism of MyD88-mediated uptake of B. burgdorferi and the specific signaling pathways involved in the process.
MyD88−/− mice were maintained as heterozygous breeding pairs at the sixth-generation backcross on the C57BL/6 background. MyD88−/−, MyD88+/+, and MyD88+/− littermates were genotyped as described previously (11). C57BL/6 mice were purchased from the Jackson Laboratory. The procedures used for our animal studies were reviewed and approved by Tufts University Institutional Animal Care and Use committee. Mouse bone marrow-derived macrophages (BMDMs) were recovered from mouse femurs and differentiated as described in (12). In brief, bone marrow cells were flushed from mouse femurs with sterile RPMI media (Cellgro, Vanassas, VA) and cultured on plastic Petri dishes for 5–7 days in medium containing RPMI supplemented with 30% L929 cell conditioned media, 20% fetal bovine serum (FBS) and1% penicillin-streptomycin. BMDMs were harvested from 100×15mm Petri dishes and plated at 0.5x106 macrophages/well in 24 well tissue culture plates. The murine macrophage cell line Raw 264.7 cells (ATCC, Manassas, VA) were grown in DMEM (Cellgro, Vanassas, VA) with L-glutamine, supplemented with 10% FBS and 1% penicillin-streptomycin. Clonal isolates of infectious, low passage B. burgdorferi sensu stricto (strain N40, clone D10E9) were used for all the experiments. B. burgdorferi was cultured in Barbour-Stoenner-Kelly medium at 37°C as previously described (13).
Phagocytosis assays were performed as previously described (4). Briefly, coverslips in 24-well plates were coated with 1% rat collagen in 60% ethanol solution (Acros organics, Morris Plains, NJ) and dried overnight. Fully differentiated BMDMs were plated in RPMI supplemented with 30% L-cell conditioned media, 20% FBS and 1% penicillin-streptomycin. Cells were maintained in this media for 24 hours and then placed into serum-free RPMI overnight prior to use in assays. Serum-free conditions were used for experimentation to provide uniformity in the media and to avoid cross-reaction with bovine cytokines and inhibitors present in serum. B. burgdorferi were added to the cultures at a multiplicity of infection (MOI) of 10. Plates were centrifuged at 1200 rpm at 4°C for 5 min to bring B. burgdorferi in contact with the cells. To initiate phagocytosis, the plates were moved to 37°C (time zero). Coverslips were removed at various timepoints after the addition of B. burgdorferi and washed with cold PBS three times to remove unbound B. burgdorferi. Cells were fixed in 3.7% paraformaldehyde with 5% sucrose in PBS for 20 min at 25°C. Coverslips were washed three times in phosphate buffered saline (PBS) and stored at 4°C until use.
For experiments with poly I:C stimulation, cells were treated with synthetic double-stranded RNA (poly I:C) (Invivogen, San Diego, CA) for 4 hours before phagocytosis assay was performed. For experiments with interferon stimulation, macrophages were primed with either recombinant IFN-β Sigma, St. Louis, MO) or IFN-γ (eBioscience, San Diego, CA) overnight prior to phagocytosis. For experiments with pathway inhibitors, the inhibitors were added to the cells 1 h prior to addition of B. burgdorferi. U0126 (ERK inhibitor), SP600125 (JNK inhibitor), AG490 (JAK/STAT inhibitor), RO31-8220 (PKC inhibitor) and LY294002 (PI3K inhibitor) were purchased from Calbiochem (San Diego, CA). The concentrations (10µM U0126, 20µM SP600125, 10µM AG490, 3µM RO31-8220, and 10µM LY294002) of the inhibitors used were in conformity with earlier published reports and had no visible cytotoxic effect on the BMDMs as judged by trypan blue exclusion (14–16). The activity of the inhibitors at the concentrations was confirmed by testing for known effects of the inhibitors on expression of selected genes by q-rt-PCR.
Immunofluorescence studies were performed as previously described (4) with the following modifications. Briefly, the coverslips were incubated three times for 5 min in blocking buffer (PBS containing 2% goat serum) at room temperature. All antibody incubations were continued for 1 h at 37°C in a humidified incubator. After blocking, the coverslips were incubated for 1 h at 37°C with an anti-B. burgdorferi polyclonal rabbit antibody (a kind gift of Dr. Jenifer Coburn) diluted 1:10,000 in blocking buffer. Coverslips were then washed three times with blocking buffer and incubated with a FITC-conjugated goat anti-rabbit IgG antibody (Molecular Probes, Eugene, OR). Samples were again washed three times in phosphate-buffered saline (PBS) for 5 min and then permeabilized with chilled methanol for 10 sec. After incubating three times for 5 min in blocking buffer, the coverslips were again incubated with the anti-B. burgdorferi rabbit antibody. After washing three times for 5 min in blocking buffer, samples were incubated simultaneously with a Texas Red conjugated goat anti-rabbit IgG antibody (Molecular Probes, Eugene, OR). For studies of Arp3 localization, Arp2/3 complexes were detected by rabbit anti-Arp3 antibody (1:200), a generous gift of Dr. Ralph Isberg (Tufts University, MA) and B. burgdorferi were identified by using mouse anti-OspA (outer surface protein A) antibody (1:5,000) (Maine Biotechnology, Portland, ME). After washing, coverslips were mounted by using DAPI Vectashield mounting medium (Vector Laboratories, Burlingame, CA) and examined by differential interference contrast (DIC) and fluorescence microscopy by using a Zeiss Axioplan 2 microscope (Carl Zeiss Microscopy, Jena, Germany). Images were captured with a digital CCD camera (Hamamatsu, Hamamatsu-City, Japan). Analysis of co-localization of the fluorescent labels was performed by using OpenLab software (Improvision Inc., Lexington, MA) with or without three dimensional reconstruction and deconvolution as indicated.
For quantitative analyses, the percentage of cells with one or more internalized B. burgdorferi particles were counted by examining sequential fields from minimum three independent experiments (minimum 3 sets of 100–200 cells/set). Cells containing any internalized B. burgdorferi particles or cells containing internalized/intact B. burgdorferi were counted and expressed as a percent of the total number cells examined. The mean percent of minimum three independent experiments were plotted over time and the statistical significance between groups was analyzed by using the nonparametric Mann-Whitney U test.
After incubation with B. burgdorferi, cells were washed with phosphate buffered saline (PBS) and RNA extracted by using Trizol as per the manufacturer’s instructions (Invitrogen, Carlsbad, CA). First-strand synthesis of cDNA from total RNA was performed by using Improm II (Promega, Madison, WI) as per the manufacturer’s instructions. Quantification of cDNA was performed by quantitative PCR (iCycler, Biorad) by using Sybrgreen technology (Quantitect SYBR green PCR mix; Qiagen, Valencia, CA). Cycling parameters were 60°C for 5 min and 95°C for 15 min, followed by 40 cycles of 95°C for 30 sec and 60°C for 1 min. The following primers were used: IFN-β forward: AGCTCCAAGAAAGGACGAACAT, IFN-β reverse: GCCCTGTAGGTGAGGTTGATCT. TNF-α forward: ATGAGCACAGAAAGCATGATC, TNF-α reverse: TACAGGCTTGTCACTCGAATT. The specificity of each reaction was checked by melt curve analysis and by agarose gel electrophoresis of PCR products. Expression of target genes was referenced to expression of β-actin. Calculations of expression were normalized by using the ΔΔCt method where the amount of target, normalized to an endogenous reference and relative to a calibrator, is given by 2−ΔΔCt, where Ct is the cycle number of the detection threshold.
Raw 264.7 cells were transiently transfected with a dominant-negative (DN) mutant of MyD88 (a generous gift of Dr. Katherine Fitzgerald, UMASS Medical School, MA) (17) or pCDNA3-GFP plasmid (Invitrogen, Carlsbad, CA), by using a 4:1 lipid/DNA ratio of Lipofectamine 2000 transfection reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s protocol (18). The transfection mix was added to cells in DMEM serum-free media and incubated at 37°C. After 6 hours, the media was replaced with 10% FBS added DMEM, and 24 hours later, we performed phagocytosis assay as described. We estimated transfection efficiency of Raw 264.7 cells by randomly choosing 10 fields and counting both total cells and cells expressing GFP after transient transfection of cells with pCDNA3-GFP plasmid. Estimated transfection efficiency for all experiments was approximately 70–80%.
Cellular lysates of mouse macrophages were prepared by lysis buffer (0.05M Tris, pH 7.4, 0.15M NaCl, 0.5 mM PMSF, 50µg/ml aprotinin, 10µg/ml leupeptin, µg/ml pepstatin, 0.4mM sodium orothovanadate, 10mM sodium fluoride, and 10mM sodium pyrophosphate) and then separated by SDS-PAGE on 4–12% acrylamide gels and transferred to a polyvinyldifluoride membrane. The membrane was incubated in blocking buffer (5% milk in 0.2M Tris base, 1.36M NaCl with 0.1% Tween20 (TBS/T)) for 1 hour at room temperature and washed three times for 5 minutes each with 15ml of TBS/T. Membranes were incubated with the primary antibody overnight at 4°C. Phospho-Akt (Ser 473) antibody and total Akt antibody were purchased from Cell Signaling (Danvers, MA). After washing three times with TBS/T, the membranes were incubated with anti-rabbit IgG HRP-conjugated secondary antibody for one hour at 25°C. After washing three times with TBS/T, the membrane was incubated with LumiGlo substrate and exposed to the film.
Experiments were repeated three times as indicated. The statistical significance between groups was analyzed by using the nonparametric Mann-Whitney U test. Differences were considered statistically significant when the p values were equal to or less than 0.05.
We previously reported that MyD88 is required for uptake of B. burgdorferi, but not for E. coli (4). Among the differences between innate immune recognition of B. burgdorferi and E. coli is the fact that B. burgdorferi lipoproteins are recognized by TLR2, while E. coli lipopolysaccaride (LPS) is recognized through TLR4. One potential implication of this difference is that TLR4, in addition to utilizing MyD88 for activation of signaling pathways, can also activate MyD88-independent pathways through the use of TRIF adaptor pathway. In order to determine whether signaling through TRIF can complement the loss of MyD88 and restore phagocytosis of B. burgdorferi in MyD88 deficient cells, we stimulated both WT and MyD88−/− BMDMs with the TLR3 ligand, poly I:C. Among TLRs, TLR3 is unique in that it is the only identified TLR that does not utilize MyD88 and activates pathways solely through recruitment and activation of TRIF.
We first confirmed the effect of poly I:C on activation of MyD88−/− cells by evaluating mRNA expression of type I interferon (IFN-β) and tumor necrosis factor (TNF-α). Poly I:C stimulation induced similar mRNA expression of IFN-β and TNF-α for both WT and MyD88−/− macrophages, indicating that MyD88-independent signaling pathways remained intact in both cells types as would be anticipated (Fig. 1A). The addition of poly I:C in MyD88−/− cells significantly increased uptake of B. burgdorferi to WT levels at 20 and 60 min post infection (p=0.009, p=0.016) (Fig. 1B). Poly I:C did not affect the phagocytosis of B. burgdorferi in WT BMDMs (p=0.602). Similar complementation of the phagocytic defect for B. burgdorferi with the addition of LPS to MyD88−/− cells was also seen (data not shown).
TLR3 signaling results in the induction of type I IFN, such as IFN-α and -β. Both type I and type II IFNs are known activators of BMDMs (19, 20). To determine whether the effect of poly I:C in restoring phagocytosis to MyD88−/− BMDMs is due to cellular activation through IFNs or whether it is the result of activation of more specific pathways that converge downstream of MyD88 and TRIF, we studied the effects of activation of cells with IFN-β on the phagocytosis of B. burgdorferi.
BMDMs were first pre-incubated with recombinant IFN-β (100 Units/ml) overnight to activate macrophages and phagocytosis assays were performed the next day. We evaluated phagocytosis of B. burgdorferi by WT and MyD88−/− cells with and without IFN-β stimulation. In contrast to results with the addition of poly I:C, priming MyD88−/− macrophages with IFN-β did not increase the phagocytosis of B. burgdorferi and at 20 min and 60min post infection, there were still fewer cells containing internalized spirochetes, compared to WT cells primed with IFN-β (Fig. 2). There was no significant increase in numbers of cells containing internalized B. burgdorferi, even in the presence of IFN-β priming in MyD88 deficient cells (32.87±6.22% MyD88−/− BMDMs alone vs. 29.09±2.90% MyD88−/− BMDMs primed with recombinant IFN-β at 60min post infection; Fig. 2). We also tested higher concentrations of IFN-β(1000 and 5000 Units/ml) which also showed no effect (data not shown). This data suggest that poly I:C mediated increase of B. burgdorferi uptake in MyD88 deficient cells is not due to TLR3-mediated induction of type I interferon. Of note, we also observed similar results with priming BMDMs with recombinant IFN-γ, which is often used as an activator of macrophages for killing of intracellular organisms, but which is not induced by TLR3 activation (data not shown).
To examine the role of other potential mediators, we studied the requirement for IL-1 in phagocytosis of B. burgdorferi. IL-1 is an important cellular activator. IL-1β is induced from BMDMs by the presence of B. burgdorferi through activation of MyD88 (4). In addition, IL-1 receptor (IL-1R), similar to TLRs and IL-18R family members, utilizes the MyD88 adapter protein to initiate signaling (21). We previously reported that phagocytosis of B. burgdorferi is not dependent on the presence of individual TLRs, such as TLR-2, 5, or 9 (4). Previous reports have suggested the IL-18 does not have a role in the inflammatory response to B. burgdorferi or in control of infection (22). IL-1R has been shown to promote neutrophil recruitment and control clearance of the organisms via MyD88 signaling in an effective innate immune response against Staphylococcus aureus infection (23). Therefore, we sought to examine whether IL-1R is also important for uptake of B. burgdorferi.
We performed phagocytosis assays by using BMDMs from IL-1R−/− mice as described above. WT control BMDMs ingested and degraded B. burgdorferi within phagolysosomes of macrophages by 20 min with almost no B. burgdorferi seen extracellularly in association with cells. The absence of IL-1R did not affect phagocytosis of B. burgdorferi and at 20 min and 60min, almost all the organisms were degraded with the same percentage of cells containing degraded B. burgdorferi as WT control BMDMs (71.29±9.14% WT control BMDMs vs. 67.88±8.8% IL-1R−/− BMDMs at 60 min post infection; Fig. 3). Similar results were seen using BMDMs from mice deficient in IL-1α, IL-1β or IL-1α/β (data not shown).
Since the defect in phagocytosis of B. burgdorferi by MyD88−/− BMDMs did not appear to be due to a lack of activation that could be complemented by TLR3-dependent pathway, we began to examine signaling pathways that are activated downstream of both MyD88 and TRIF and/or have been shown to be activated by the presence of B. burgdorferi. We and other labs have shown that B. burgdorferi induces multiple signaling pathways, such as MAPK (p38, JNK, and ERK) (14, 24, 25), PKC (15), and JAK/STAT (14, 26). We have previously shown that inhibition of p38 MAPK does not suppress uptake and degradation of B. burgdorferi (4) despite the important role that p38 activation has been shown to play for phagocytosis of other bacteria through its role in phagolysosomal maturation (8). To determine which signaling pathway(s) is/are involved in MyD88-mediated phagocytosis, we used pharmacological inhibitors of specific signaling pathways to investigate downstream targets of MyD88 in phagocytosis.
BMDMs from WT mice were pre-incubated with U0126 (ERK1/2 inhibitor), SP600125 (JNK inhibitor), AG490 (JAK/STAT inhibitor) or RO31-8220 (PKC inhibitor) for 1 hour prior to the addition of B. burgdorferi (Fig. 4). Concentrations of the inhibitors were chosen based on previously published studies showing optimal inhibition and specificity for the targeted receptors in the concentration range used without causing any cytotoxicity (14–16). The activity of each inhibitor was confirmed by examining the effect of inhibitors on the induction of downstream cytokines known to be associated with that pathway (data not shown).
Although activation of ERK, JNK, JAK/STAT or PKC signaling molecules are important for the induction of inflammatory signaling pathways, inhibition of these pathways did not affect phagocytosis of B. burgdorferi and by 60 min, almost all the organisms were degraded with the same percentage of cells containing degraded B. burgdorferi as vehicle treated controls (Fig. 4A). This suggests that these pathways are either not involved in phagocytosis of B. burgdorferi or that the level of pathway activation required to support phagocytosis is far less than that needed for cytokine induction.
PI3K has been shown to play an important role in the phagocytosis of large particles (1, 27–29). PI3K activation has been shown to occur downstream of TLR signaling (16, 30, 31), and few studies have reported its importance for TLR-mediated phagocytosis (6, 32). We performed phagocytosis assays in the presence of PI3K inhibitor, LY294002. BMDMs from WT mice were pre-incubated with LY294002 for 1 hour prior to the addition of B. burgdorferi (Fig. 4B). As in Fig. 4A, in the vehicle (DMSO) treated controls, B. burgdorferi were found to be degraded and associated with phagolysosomes of WT BMDMs by 60 min with almost no B. burgdorferi seen extracellularly in association with cells (data not shown). In contrast, when WT BMDMs were pre-treated with LY294002 before incubation with B. burgdorferi, phagocytosis was significantly inhibited (66.48±4.79% vehicle treated cells vs. 29.08±12.83% LY290402 treated cells (p≤0.05, compared with vehicle treated controls; (Fig. 4B), similar to what is seen in MyD88−/− BMDMs. This data demonstrates that PI3K activation is required for the uptake of spirochetes.
To determine whether PI3K was responsible for the phagocytic defect in MyD88−/− BMDMs, we sought to determine the relationship between B. burgdorferi activation of MyD88 and PI3K.
BMDMs from either WT or MyD88−/− mice were infected with B. burgdorferi and then harvested (Fig. 5). Western blots of cellular lysates show that there is an increase in Akt phosphorylation in WT BMDMs by 20 min (Fig. 5A). In contrast, there was substantially less phosphorylation of Akt in BMDMs from MyD88−/− mice. To confirm this data, we transiently transfected plasmids expressing a dominant negative MyD88 (MyD88-DN) or an empty vector control into the mouse macrophage cell line Raw 264.7. The effect of abrogating MyD88 signaling in these cells was confirmed by reduced TNF-α and IL-6 mRNA expression in response to B. burgdorferi (data not shown). MyD88-DN and control-transfected cells were incubated with B. burgdorferi for 20min and then harvested (Fig. 5B). Western blots of cellular lysates show that there is an increase in phosphorylation of Akt, a phosphorylation target of PI3K, in Raw cells transfected with control vector by 20 min. In contrast, there was a reduction in phosphorylation of Akt in Raw cells transfected with MyD88-DN. These data demonstrate that incubation of B. burgdorferi induces activation of PI3K and Akt phosphorylation at least in part through MyD88 signaling.
To determine if TLR3-mediated complementation of phagocytic defect by MyD88−/− cells also requires the activation of PI3K, we examined first, whether stimulation of TLR3 dependent pathways by using poly I:C induces phosphorylation of Akt, and second, whether blocking PI3K pathway after rescuing phagocytic defect in MyD88−/− cells with poly I:C stimulation leads to suppression in B. burgdorferi uptake.
BMDMs were treated with various concentrations of poly I:C (0–100 µg/ml) and phosphorylation levels of the Akt was examined by Western blotting. Phosphorylation of the Akt was substantially increased upon poly I:C stimulation and this was dependent on the concentration used (Fig. 6A). Therefore, this suggests that TLR3 induces phosphorylation of Akt in BMDMs in a concentration-dependent manner.
Next, we tested whether blocking PI3K pathway inhibits uptake of B. burgdorferi in MyD88−/− BMDMs pre-treated with poly I:C. BMDMs from WT or MyD88−/− mice were pre-treated with poly I:C for 4 hours and control or the PI3K inhibitor, LY294002 (PI3Ki), was added for 1 hour prior to B. burgdorferi incubation. Similar to results in Fig. 1, stimulation of MyD88−/− cells with poly I:C resulted in a marked increase of B. burgdorferi uptake at 60min post infection (Fig. 6B). However, in the presence of PI3Ki, both WT and MyD88−/− cells internalized statistically significant less B. burgdorferi, compared to control-treated cells (64.60±5.04% WT + poly I:C with a control treatment vs. 23.83±2.70% WT + poly I:C with a LY290402 treatment (p≤0.05); 69.82±1.70% MyD88−/− cells + poly I:C with a control treatment vs. 21.13±4.15% MyD88−/− cells + poly I:C with a LY290402 treatment (p≤0.05). Thus, PI3K is activated by both TRIF and MyD88 signaling pathways. These data demonstrate that PI3K activation is a critical pathway that is required for the optimal phagocytosis of B. burgdorferi.
Actin polymerization has been well characterized to be a driving force for the formation and extension of membrane protrusions, which is important for the successful phagocytosis of microbial organisms (33). PI3K signaling has been shown to play an important role in actin polymerization through activation of Rac (34). The Rho family GTPases, Rac1 and CDC42, subsequently recruit Arp2/3 to form the actin complex (35, 36). To determine whether the defect in B. burgdorferi uptake by MyD88−/− BMDMs was due to a loss of PI3K directed actin polymerization, we examined the localization of the Arp2/3 complex of actin with B. burgdorferi. The cellular distribution of Arp2/3 complexes was evaluated by using an antibody directed against the 50 kDa Arp3 subunit of the Arp2/3 complex (37, 38). At 5 min post B. burgdorferi infection, Arp2/3 was found clearly associated with contact points where B. burgdorferi were adhered to the WT cell surface and throughout the entire length of organisms as they are been taken up into WT cells (Fig. 7). In contrast, recruitment of Arp2/3 co-localized with B. burgdorferi attached to the surface of MyD88−/− cells was not observed. Similarly, BMDMs treated with the PI3K inhibitor also did not show co-localization of Arp2/3 with attached B. burgdorferi. This suggests that MyD88 signaling is important for the coordination of actin polymerization and efficient recruitment of Arp2/3 required for uptake of B. burgdorferi. These data provide further evidence that PI3K signaling pathway, by directing cellular distribution of Arp2/3 complexes, is required for MyD88-dependent phagocytosis of B. burgdorferi.
A role for MyD88 in different aspects of phagocytosis, including effects on uptake, phagolysosomal maturation, and oxidative killing, has been proposed (4, 5, 8, 9). In this study, we investigated the mechanisms by which MyD88 participates in the phagocytosis of B. burgdorferi. We have previously shown that MyD88 plays an important role in uptake, but not phagolysosomal processing of B. burgdorferi (4). There have only been a few reports on the role of TLR signaling on the uptake of organisms (4–7). A study by Doyle et al. suggested that the role of MyD88 in uptake of organisms occurs through up-regulation of specific phagocytic receptors, such as scavenger receptors (5). Up-regulation of specific scavenger receptors including scavenger receptor-A (SR-A), macrophage receptor with a collagenous structure (MARCO), and lectin-like oxidized low density lipoprotein receptor-1 (LOX-1), does occur in response to B. burgdorferi infection (Shin, unpublished data). However, consistent with the results seen for induction of scavenger receptors by other organisms (5), up-regulation of these receptors by B. burgdorferi appears to occur at a time-point after uptake of the organism into the cells, suggesting that scavenger receptors are not major contributors to the early uptake of B. burgdorferi seen in our phagocytic assays.
Instead, we have shown that the uptake of B. burgdorferi is mediated by downstream signaling events activated in response to the organism. We found that the role of MyD88 activation in phagocytosis can be replaced by activation of the other major TLR signaling adaptor, TRIF (Fig. 1B). By pre-treating MyD88−/− cells with a TLR3 ligand, poly I:C, which is able to activate downstream signaling through TRIF without the involvement of MyD88, we were able to restore the ability of MyD88−/− cells to phagocytose B. burgdorferi. The ability to restore phagocytosis with the addition of poly I:C confirms that there is not an intrinsic defect in the ability of MyD88−/− cells to take up B. burgdorferi and provides clues as to the possible downstream pathways responsible for controlling phagocytosis of B. burgdorferi.
Activation downstream of TRIF occurs along two major pathways: 1) activation of TRAF3, which leads to a subsequent induction of type I interferon and activation of interferon-responsive genes and 2) activation of TRAF6 (also utilized by MyD88) which leads to downstream activation of many signaling pathways and translocation of NFκB. Activation of macrophages by type I and type II IFNs has been shown to increase phagocytic capacity of these cells (19, 20, 39). However, unlike poly I:C, addition of IFN-β was unable to restore phagocytosis of B. burgdorferi in MyD88−/− cells (Fig. 2), making it unlikely to be the mechanism by which TRIF activation complements the loss of MyD88.
Thus, we focused on pathways directly downstream of TRAF6 as well as those that may be activated indirectly as a result of TRAF6 activation. We examined downstream pathways that can be activated by recognition of B. burgdorferi products including p38, ERK, JNK, PKC, JAK/STAT and PI3K using chemical inhibitors. Of these, only inhibition of PI3K blocked uptake of B. burgdorferi (Fig. 4B) (results for p38 were previously reported in (4)). PI3K is a major regulator for phagocytosis of large particles (28). Inhibition of PI3K can block new membrane formation at the site of particle internalization and reduce membrane extension and fusion of the bound particle (27–29). Previous studies have shown the involvement of PI3K in the induction of phagocytosis of the parasite Cryptosporidium parvum, and of some intracellular pathogens such as Legionella pneumophila, Listeria monocytogenes, and Chlamydia pneumoniae (40–43). Consistent with these findings, Wang et al. have shown that expression of TRAF6 increases PI3K-dependent cytoskeletal changes required for reorganization of actin cytoskeleton (44).
Successful phagocytosis relies on the rearrangements of the actin cytoskeleton and the plasma membrane to engulf bacteria or particles. Arp2/3 complexes are driving forces to bring together actin monomers to form a new nucleus for actin polymerization and localized in the lamellipodia at branch points between actin filaments (35, 36). PIP2 has been suggested to regulate actin remodeling during phagocytosis, probably through its regulation of cytoskeletal proteins such as Wiskott-aldrich syndrome protein (WASP) (45). Uptake of B. burgdorferi has been associated with formation of f-actin rich structures, driven by activation of WASP and Arp2/3 complex, which are recruited to the engulfment structures (46). Thus, we wanted to examine if a loss of MyD88 or inhibition of PI3K signaling pathway would have any impact on the localization and distribution of Arp2/3 complexes at the sites of B. burgdorferi entry into the cells. We show that both MyD88−/− BMDMs and WT BMDMs treated with a PI3K inhibitor failed to recruit Arp3 to sites of B. burgdorferi, suggesting a potential role of MyD88/PI3K for initiating actin polymerization (Fig 7).
Interestingly, although we initiated our studies of poly I:C complementation of MyD88 deficiency because of the observation that E. coli uptake is not affected by MyD88 deficiency and the hypothesis that E. coli LPS could activate TRIF through TLR4, bypassing the requirement for MyD88, uptake of E. coli does not appear to be inhibited by the addition of a PI3K inhibitor (Shin, data not shown) and thus likely does not require either MyD88 or TRIF. Other investigators have also found that E. coli uptake, and that of other small extracellular bacteria such as Brucella and Salmonella, does not require PI3K (7, 47, 48). One possibility for the differences in requirement for PI3K activation by different bacteria may be differences in the mechanisms of phagocytosis due to the size of the organisms.
Some investigators have reported that B. burgdorferi is primarily taken up through coiling phagocytosis rather than conventional phagocytosis (46, 49). Coiling phagocytosis is a process whereby a single pseudopod extends and engulfs the spirochete. It has been suggested that processing of B. burgdorferi internalized by coiling phagocytosis differs from conventional phagocytosis in that degradation occurs through a non-lysosomal mechanism (50). This consistent with the lack of a role for p38 MAPK in killing of B. burgdorferi as non-lysosomal mediated killing may not utilize p38 MAPK to control phagosome acidification and maturation. However, the phenomenon of coiling phagocytosis remains somewhat controversial and more recent studies have suggested that B. burgdorferi can be internalized through either coiling or conventional phagocytosis and that, regardless of the mechanism of internalization, all of the degraded particles are localized within lysosomes (50).
In summary, we have shown that MyD88-mediated phagocytosis of B. burgdorferi can be initiated by TRIF and is dependent upon activation of PI3K. We have demonstrated that inhibition/loss of either MyD88 or PI3K results in a failure to appropriately recruit Arp2/3 complexes to the bacteria/cell membrane interface to initiate phagocytosis. Furthermore, activation of MyD88 increases the activity of PI3K. Thus, the MyD88/PI3K pathway is an essential mechanism that controls uptake and phagocytosis of B. burgdorferi. The role of PI3K in MyD88-mediated phagocytosis of B. burgdorferi differs from prior reports demonstrating an important role for p38 in MyD88-mediated phagocytosis of other organisms and suggests that different pathways mediate this process depending on the infecting organism (8). Taken together, the identification of the MyD88/PI3K as a key pathway for phagocytosis of B. burgdorferi provides new insights into the complex TLR signaling pathways that govern the phagocytic response, which may not only have important implications in mechanisms of host defense against B. burgdorferi but also in infections caused by and other bacterial pathogens.
We wish to thank Dr. Ralph Isberg, Dr. Jenifer Coburn, Dr. Alexander Poltorak, and Dr. Philip Tsichlis, Dr. Tanja Petnicki-Ocwieja, Dr. Andrew Heilpern, Ms. Meghan Marre, and Mr. Deb Bhattacharya for their helpful discussions in preparing this manuscript. We also thank Mr. Alex Stone for help with genotyping MyD88−/− mice.