Innate Immune Responses to Endosymbiotic Wolbachia Bacteria in Brugia malayi and Onchocerca volvulus Are Dependent on TLR2, TLR6, MyD88, and Mal, but Not TLR4, TRIF, or TRAM¹

Human filarial infections due to Wuchereria bancrofti, Brugia malayi, and Onchocerca volvulus are important public health problems and are the cause of physical disability and socioeconomic disempowerment in many tropical areas of the world. The first two organisms are etiologic agents of lymphatic filariasis and cause inflammation of lymphatic vessels with impaired lymph flow. Chronic infection may lead to massive swelling of the limbs (elephantiasis) and disfigurement of the male genitalia (hydrocele, epididymitis). In onchocerciasis or “river blindness,” caused by O. volvulus, inflammatory responses elicited by the death of larval microfilaria passing through the cornea decreases visual acuity that is manifest in its most severe form by blindness. The mechanisms underlying the inflammation induced by filarial infection are poorly understood, but studies of humans with natural infection and experimental animals exposed to live filaria or soluble extracts of the organisms have implicated a role for both adaptive and innate immunity (1, 2, 3). The remarkable observation that W. bancrofti, B. malayi, and O. volvulus harbor endosymbiotic Wolbachia bacteria (4, 5, 6) has raised the possibility that infected individuals respond to these organisms in addition to nematode ligands and that Wolbachia can initiate or promote inflammation (7, 8, 9, 10, 11, 12). Despite remarkable longevity in the human host (13), adult filarial worms themselves are not strong inducers of inflammation (3, 14). Wolbachia bacteria are in the family Rickettsiacea and are obligatory intracellular bacteria that infect a wide variety of invertebrates, in which they have an important role in sex determination and speciation (15). In filarial worms, Wolbachia bacteria are concentrated in intracytoplasmic vacuoles of hypodermal lateral cords and female reproductive organs and are also detectable in oocytes and microfilaria, reflecting matrilinear inheritance (16). Nematode reproduction, molting, and survival dependent on the endosymbiont as prolonged exposure to antibiotics targeting Wolbachia result in decreased fecundity, impaired molting, and even worm death in vitro as well as in animal and human infections (17, 18, 19, 20). Results of studies of filaria-infected individuals are consistent with the hypothesis that Wolbachia contribute to the pathogenesis of filarial disease. Wolbachia DNA and even Wolbachia bacteria are detectable in the plasma of persons with onchocerciasis or lymphatic filariasis within 1 day of administration of diethylcarbamazine or ivermectin, drugs that rapidly kill microfilaria (21, 22). These acute posttreatment reactions are accompanied by increases in plasma TNF-α, IL-6, and LPS-binding protein (23, 24), suggesting that Wolbachia released into the bloodstream by degenerating or dead microfilaria cause the acute adenolymphangitis and fever that occur after administering antifilarial drugs. Aside from these drug-related side effects, the contribution of Wolbachia to inflammation in infected individuals is unresolved, but several observations suggest they may also play a role in initiating or perpetuating local or systemic inflammatory reactions. First, filarial-induced innate immune cellular activation is ablated if the parasites or infected animal hosts are pretreated with tetracycline antibiotics that partially eliminate intracellular Wolbachia (10, 25, 26). Second, Wolbachia isolated from filaria or from insect cell lines elicit inflammatory responses similar to those of filarial extracts containing Wolbachia (26, 27). Third, in studies using a murine model of ocular onchocerciasis in which O. volvulus extract (containing Wolbachia) is injected into the corneal stroma, it was observed that the neutrophil infiltration and loss of corneal clarity that normally ensues was markedly reduced when Wolbachia-free Acanthocheilonema viteae or Wolbachia-depleted O. volvulus extracts were used (26, 28).

Despite these findings, our understanding of how Wolbachia interact with mammalian innate immune pathways is incomplete. Examination of this problem is complicated by difficulties of interpretation related to distinguishing between the effects of Wolbachia and filarial molecules and the multiple TLRs and adaptor molecules involved in activation and regulation of innate immunity. In the current study, we extend previous findings and describe TLR and adaptor molecule usage by Wolbachia using human cell lines transfected with specific TLRs and macrophages from TLR and adaptor molecule gene knockout mice. Our results indicate that the inflammatory responses to Wolbachia are mediated primarily by engagement of TLR2 and the coreceptor TLR6, and are dependent on the adaptor molecules MyD88 and Toll/IL-1R domain-containing adaptor protein (TIRAP)³/MyD88 adaptor-like (Mal).

C57BL/6 mice were obtained from The Jackson Laboratory. TLR2-, TLR4-, MyD88-, Mal-, Toll/IL-1R domain-containing adaptor-including IFN-β (TRIF)-, and TRIF-related adaptor molecule (TRAM)-deficient mice were provided by Dr. S. Akira (Osaka University, Osaka, Japan), TLR2/4^−/− mice were generated in-house from single TLR2^−/− and TLR4^−/− strains. Animals were housed in filter-covered microisolator cages in the animal facility of Case Western Reserve University (Cleveland, OH).

Live B. malayi adult worms (mixed sex) were obtained from infected Mongolian gerbils maintained by the National Institute of Allergy and Infectious Diseases repository at the University of Georgia (Athens, GA). Worms were also collected by aseptic necropsy from infected gerbils maintained at Case Western Reserve University after at least 3 mo of treatment with tetracycline drinking water (2.5 mg/ml) to reduce filarial Wolbachia. Control worms were collected from infected gerbils maintained at Case Western Reserve University without antibiotics. To remove potential contaminating exogenous bacteria, the worms were washed extensively with sterile PBS then cultured in the presence of antibiotics not active against Wolbachia for 10–14 days with daily medium changes. Live parasites were washed extensively and homogenized in sterile medium using endotoxin-free coarse glass tissue grinders (Duall size 20; Kimble-Kontes) followed by external sonication on ice for 10 min. Coarse material was removed by centrifugation. All procedures were performed on ice to minimize protein degradation, and strict aseptic techniques were used with endotoxin-free glassware and reagents. Preparations contained <0.25 ng/ml endotoxin detectable by Limulus amoebocyte lysate assay (BioWhittaker). O. volvulus adult worms were recovered from s.c. nodules of infected individuals in Southwestern Côte d’Ivoire as previously described (27). Adult worms of A. viteae were obtained from TRS Laboratories and processed as described. Protein concentration was determined using a commercial kit (Pierce) and all extract preparations were used at concentrations of 200 μg/ml. The filarial preparations were characterized by quantitative PCR for the single copy Wolbachia surface protein (WSP) gene (Brugia GenBank AJ252061; O. volvulus GenBank AJ276496) relative to 5 S (GenBank D87037). Primer sequences are available upon request.

The mosquito cell line Aa23 derived from Aedes albopictus infected with Wolbachia pipientis was maintained in 1:1 Mitsuhashi Maramorosh medium (Promo Cell) and Schneider’s Insect medium (Sigma-Aldrich) with 10% FCS at 26°C. To isolate Wolbachia bacteria, infected cells were centrifuged at 4100 × g at 4°C for 5 min, and the pellet was resuspended in saline and stored at −80°C. An immunofluorescent Ab test was used to assess the purity of the preparation by staining with a rabbit polyclonal serum against WSP of B. malayi Wolbachia. To isolate Wolbachia bacteria from filarial nematodes, adult male and female Onchocerca ochengi worms were coarsely chopped, then homogenized in a glass homogenizer in 0.85% sodium chloride with 0.001% Nonidet P-40 under sterile conditions, filtered through a 70-μm sieve to remove large debris, and then centrifuged at 350 × g for 25 min at 4°C to pellet nematode nuclei. The supernatant was centrifuged at 4100 × g at 4°C for 25 min to pellet Wolbachia. The preparation was characterized by quantitative PCR for the single copy wsp gene (29) and by immunofluorescent staining and counting on a FACSVantage flow cytometer (BD Biosciences).

Stable lines of HEK293 cells expressing surface protein TLR constructs were obtained under Material Transfer Agreement from Dr. D. Golenbock (University of Massachusetts Medical School, Worchester, MA). These HEK cells naturally lack most TLRs (constitutively express TLR1, TLR5, and TLR6), and genetic complementation with TLR constructs renders the cells responsive to the respective TLR ligands (30, 31). The cells were maintained in DMEM (Cellgro) supplemented with 10% low endotoxin FCS (Atlanta Biologicals) in a 5% saturated CO₂ atmosphere at 37°C. TLR4 signaling requires MD2, a small glycosylated protein, to optimally sense LPS, therefore the HEK-TLR4 cells were activated in the presence of soluble human MD2. Nontransfected HEK293 or HEK-TLR3 cells were also stimulated. To activate, cells were seeded into 96-well plates at a density of 5 × 10⁴ cells/well in triplicate. The cells were allowed to adhere for a minimum of 2 h then stimulated with filarial and Wolbachia reagents and TLR controls (Ultrapure LPS (Escherichia coli O111:B4) (TLR4/MD2); Pam₃CysK₄ or peptidoglycan (a preparation that likely also contains contaminating lipoteichoic acid, a strong TLR2 ligand) (TLR2); poly(I:C) (TLR3) (InvivoGen)), and activation was measured by secreted IL-8 using a commercially available ELISA (R&D Systems). Results are reported as fold activation (IL-8 response to test ligand/medium alone). Each experiment was repeated at least three times with similar results.

Peritoneal macrophages were isolated by adherence following peritoneal lavage (72 h after i.p. injection of sterile thioglycolate solution (REMEL)) and stimulated at a density of 125,000 cells per well in 96-well flat-bottom tissue culture plates in RPMI 1640 (Cambrex) with 10% low endotoxin FCS. After an overnight stimulation, the supernatants were removed and assayed for the proinflammatory cytokines TNF-α, IL-6, and RANTES using commercially available reagents (R&D Systems). Specific inhibition of TLR activation of murine macrophages was conducted using anti-TLR2 Abs (T2.5 clone, which was a gift from C. Kirschning, Technical University of Munich, Munich, Germany) or isotype control. All animal studies have been reviewed and approved by the Case Western Reserve University Institutional Animal Care and Use Committee Studies using MyD88, Mal, TRIF, and TRAM knockout mice were performed under approved protocols in the laboratories of D. Golenbock at the University of Massachusetts Medical School (Worcester, MA).

The corneal stroma of control and TLR gene knockout mice were injected with 4 μg of soluble O. volvulus extract as previously described (27). After 24 h, eyes were removed, snap frozen in liquid nitrogen, and 5-μm sections were incubated for 2 h with anti-macrophage Ab F4/80 diluted 1/100 in 1% FCS/PBS as described (9). After washing, corneal sections were incubated with FITC-conjugated rabbit anti-rat Ab (Vector Laboratories) diluted 1/200 in 1% FCS/TBS. Slides were mounted in Vectashield containing 4′,6′-diamidino-2-phenylindole (DAPI) (Vector Laboratories), and the number of macrophages per section was counted by fluorescence microscopy.

Data from the mouse macrophage experiments were analyzed using commercial software (OriginLab) and the two-sample independent Student’s t test was used to compare the responses of gene knockout mice compared with the control mice. A value of p < 0.05 was considered statistically significant. The fold change in human IL-8 production by transfected HEK cells was calculated by dividing the average human IL-8 produced by replicate wells exposed to a test ligand by the average human IL-8 produced by replicate wells exposed to the medium alone. This value controls for the differences in background human IL-8 produced by the different cell lines. Statistical comparisons were made between the fold change in human IL-8 produced in response to a specific ligand and the background (medium only) using the Student two-sample independent t test.

To determine TLR activation by filarial extracts containing Wolbachia, HEK cells transfected with human TLRs were incubated with extracts of B. malayi and O. volvulus, which harbor Wolbachia, or extracts of the rodent nematode A. viteae, which do not naturally carry Wolbachia. TLR activation is expressed as a fold change in human IL-8 production by cells exposed to the ligand compared with human IL-8 production by cells incubated with medium alone. HEK cells naturally lack TLR2, TLR3, and TLR4. However, coreceptors required for TLR2 activation (including TLR1 and TLR6) are constitutively expressed. Incubation with specific control ligands for TLR2, TLR3, or TLR4 (peptidoglycan/lipoteichoic acid, poly(I:C), and LPS, respectively) stimulated only the cells expressing the appropriate receptors. Extracts of B. malayi strongly activated TLR2-expressing cells but not cells expressing TLR3 or TLR4. Furthermore, none of the cell lines were activated by A. viteae that did not harbor Wolbachia. Enriched Wolbachia preparations made from an insect cell line containing Wolbachia pipientis showed activation of cells expressing TLR2 but not TLR3 or TLR4 (Fig. 1,A). Extracts of O. volvulus also activated the TLR2-expressing cell line, but not cells expressing TLR3 or TLR4 (Fig. 1 B). Together, these data show that filarial extracts selectively activate TLR2 and that activation requires the presence of Wolbachia.

Previous studies showed that the presence of Wolbachia was essential for filarial-induced macrophage activation (25). To determine whether TLR2 is also important in the induction of inflammatory responses by immunocompetent C57BL/6 macrophages, we used a neutralizing Ab directed against the epitope binding C terminus region of TLR2. Fig. 2,A shows that the TNF-α response by C57BL/6 macrophages stimulated with B. malayi extract containing Wolbachia was completely inhibited by preincubation with an Ab directed against TLR2 compared with an isotype control Ab. Similar results were obtained for IL-6 and RANTES (Fig. 2, B and C).

To examine further the role of TLRs in macrophage activation, peritoneal macrophages isolated from TLR2^−/− and TLR4^−/− mice were stimulated with soluble extracts of B. malayi, O. volvulus, and A. viteae, and cytokine production was measured by ELISA. Because many TLR2 ligands require either TLR6 or TLR1 as a coreceptor for activation (32, 33), we also examined the effect of filarial extracts on TLR6^−/− macrophages. The TNF-α cytokine response to filarial extracts is shown in Fig. 3,A. The soluble extract of B. malayi stimulated TNF-α production by wild-type C57BL/6 and TLR4^−/− macrophages, but not TLR2^−/− or TLR6^−/− macrophages. In contrast, a soluble extract of A. viteae worms that does not contain Wolbachia failed to induce TNF-α production by macrophages from any of the mouse strains. In a separate experiment O. volvulus-induced extracts containing Wolbachia induced a similar response to B. malayi extract in activating C57BL/6 and TLR4^−/− macrophages, but not TLR2^−/− or TLR6^−/− macrophages. Fig. 3,B shows that TLR2 and TLR6 were also required for B. malayi and O. volvulus IL-6 production. RANTES production was also dependent on TLR2/6, as shown in Fig. 3 C. Control TLR ligands showed the expected TLR dependence (ultrapurified E. coli LPS responses are abrogated in the TLR4^−/− macrophages, and triacylated synthetic ligand Pam₃CysK₄ responses are abrogated in TLR2^−/− macrophages).

Enriched Wolbachia preparations were made from the nematode O. ochengi as described. As shown in Fig. 3, cytokine production induced by isolated Wolbachia was comparable between TLR4^−/− and wild-type macrophages. In contrast, isolated Wolbachia from O. ochengi did not stimulate TNF-α, IL-6, or RANTES responses by macrophages isolated from TLR2^−/− or TLR6^−/− mice. Together, these findings demonstrate that nematode endosymbiotic Wolbachia activates TLR2 and TLR6, but not TLR4 expressed on murine macrophages.

The adaptor molecule TIRAP/Mal is integral to TLR2 signaling via MyD88, whereas TRIF and TRAM are associated with TLR3 signaling and with LPS-induced up-regulation of IFN-αβ via TLR4 (34, 35, 36). To determine the relative roles of these adaptor molecules in filaria and Wolbachia-induced inflammation, peritoneal macrophages were isolated from mice deficient in MyD88, Mal, TRIF, or TRAM and then stimulated overnight with TLR control ligands, soluble filarial extracts, or isolated Wolbachia bacteria from O. ochengi. Secreted cytokines were measured in the supernatants by ELISA.

As shown in Fig. 4,A, TNF-α production induced by B. malayi and O. volvulus extracts as well as isolated Wolbachia was completely abrogated in MyD88^−/− and Mal^−/− macrophages. Similarly, IL-6 responses to these Wolbachia containing reagents were not detected from MyD88^−/− and Mal^−/− macrophages (Fig. 4,B). In contrast, A. viteae extracts did not stimulate TNF-α or IL-6 responses from wild-type or adaptor molecule knockout macrophages. Inflammatory responses to Brugia or Onchocerca extracts were comparable for wild-type and TRIF knockout macrophages. Also, there was no deficiency in TNF-α or IL-6 responses by TRIF^−/− macrophages in response to isolated Wolbachia bacteria. Similarly, macrophages from TRAM^−/− mice showed equal or higher cytokine responses to wild-type macrophages in response to isolated Wolbachia or Wolbachia-containing filaria (Fig. 4, C and D). Overall, these data are consistent with Wolbachia signaling through TLR2/TLR6 with the use of the adaptor molecules MyD88 and Mal.

Peritoneal macrophages from wild-type C57BL/6 mice were stimulated overnight with soluble filarial extracts from adult worms collected from antibiotic-treated (to reduce Wolbachia) or control gerbils. As shown in Fig. 5, antibiotic treatment reduced the TNF-α response to undetectable levels comparable to A. viteae, which do not contain Wolbachia. Untreated filarial extracts induced a robust cytokine response. These findings confirm the role of the endosymbiont Wolbachia rather than nematode ligands in the activation of macrophages expressing TLR2.

Quantitative assessment of macrophage recruitment to the cornea in response to stromal injection of Wolbachia containing O. volvulus extract was performed in immunologically naive animals using our previously published protocol (27). As shown in Fig. 6, the number of macrophages per 5-μm corneal section is comparable between C57BL/6 and TLR4-deficient mice after intrastromal injection of O. volvulus. In contrast, the macrophage recruitment in TLR2-deficient mice is significantly reduced. Interestingly, there is no additional loss of macrophage recruitment in the double knockout mice, TLR2/4^−/− compared with the single TLR2-deficient mice. However, a significant loss of macrophage recruitment was observed when comparing TLR4^−/− to TLR2/4^−/− mice. Representative corneal sections stained with hematoxylin (to demonstrate morphology), anti-F4/80 (macrophages), or DAPI (nuclear stain to demonstrate cellular infiltrates) are shown in Fig. 7. Overall increased cellularity in wild-type and TLR4^−/− mice compared with TLR2^−/− and TLR2/4^−/− mice can be appreciated on DAPI stained section (Fig. 7, B, E, and H). Reduced macrophage recruitment in the TLR2^−/− and TLR2/4^−/− mice can be seen with F4/80 staining (Fig. 7, C, F, and I). These results show a critical role for TLR2 for macrophage recruitment to the cornea in an animal model of keratitis and no role for TLR4 either singly or in combination with TLR2.

Our findings demonstrate that innate inflammatory responses induced by the endosymbiotic intracellular bacteria Wolbachia are mediated via TLR2 and TLR6, with no apparent role for TLR4 because we found that HEK cells transfected with human TLR2 but not TLR4 are activated by isolated Wolbachia bacteria and filarial extracts containing Wolbachia. Furthermore, macrophages from TLR2^−/− mice fail to respond to soluble filarial extracts or Wolbachia bacteria, whereas responses by TLR4^−/− macrophages are intact. Additionally, wild-type macrophage inflammatory responses are blocked by anti-TLR2 Ab. TLR6, a coreceptor for TLR2, is required for Wolbachia-induced inflammatory cytokine production. Although we cannot completely exclude a possible role for TLR4 and other receptors in more complex biological systems such as whole animal model of filarial disease (37) or in the development of adaptive immune responses (38, 39, 40), we did not observe a role for TLR4 in a mouse model of keratitis.

Results reported in this study using intact Wolbachia bacteria are similar to those reported for the major WSP, which was shown to mediate inflammation via TLR2 and to a lesser extent TLR4 (41). Since we found no dependence on TLR4, the differences between these observations may be due to use in our study of whole bacteria or native soluble extracts compared with a recombinant protein (produced in E. coli), which may have low levels of endotoxin or other TLR ligand contaminations. It is also possible that the native protein of the intact bacteria binds differently to TLR2 than the soluble recombinant protein. Lower levels of activation of the transfected HEK cells induced by isolated bacteria compared with filarial extracts is likely due to a dose effect, as the approximated numbers of bacteria contained in the extracts is 10⁸–10⁹ organisms per well (by quantitative PCR for wsp, a single copy gene in the Wolbachia genome (42)) compared with 10³ for the bacterial preparations. We currently show that a single innate immune receptor complex (TLR2/TLR6) predominates in inflammation induced by isolated bacteria. Our studies using Wolbachia-containing filarial extracts also show only TLR2/TLR6-dependent inflammation, whereas Wolbachia-free filaria (both A. viteae, which does not naturally contain Wolbachia, and B. malayi, which is treated with antibiotics to reduce Wolbachia) failed to activate any inflammatory response, implying that the endosymbiotic bacteria dominate the host inflammatory responses. One explanation is that parasitic nematodes actively suppress inflammation to escape immune detection in the host. This suggestion is compatible with the predominant lack of clinical disease seen in human populations with a high transmission intensity of filarial infection (43). However, it remains to be determined how current observations fit with reports of other studies that demonstrated that ES-62, a secreted product of A. viteae, induces a TLR4–dependent immunosuppressive effect (44, 45).

Our earlier observations using in vitro and in vivo models suggested that innate immune responses to Wolbachia contained in the filarial nematodes B. malayi and O. volvulus induced inflammation via TLR4 because C3H/HeJ mice expressing a nonfunctional TLR4 due to a point mutation in tlr4 (46) showed decreased responses relative to C3H/HeN mice with the functional wild-type TLR4 (25, 28). These earlier observations did not unequivocally exclude endotoxin contamination as a potential TLR4 receptosome ligand. In the current study, we have used more stringently pre-pared filarial preparations to reduce possible endotoxin contamination and found no difference in the degree of proinflammatory cytokine responses induced in macrophages from C3H/HeJ and C3H/HeN mice (data not shown). Additionally, we now have used mice genetically deficient in TLR2 and TLR4 (but on the same genetic background) in our model of keratitis and did not find a role for TLR4 in inflammatory responses to Wolbachia containing filarial extracts in the cornea. Furthermore, there was no difference in responses between mice deficient in TLR2 and mice deficient in both TLR2 and TLR4, thereby excluding any apparent cooperative role for TLR4 in inflammatory cell recruitment to the cornea.

In addition to TLRs, we examined adaptor molecule usage by Wolbachia and filaria extracts. Signaling by all TLRs originates from the conserved Toll-IL-1R domain that mediates recruitment of members of a family of Toll-IL-1R domain-containing adaptor molecules. Recruitment of the common adaptor molecule, MyD88 (47, 48, 49) leads to the interaction and activation of the IL-1R-associated kinase family members (50) and subsequent activation of TNFR-associated factor-6 (51, 52), resulting in NF-κB activation via the IκB kinase complex. Other Toll-IL-1R domain-containing adaptor molecules have been identified, including TIRAP/Mal (53, 54, 55), TRIF (56, 57, 58), and TRAM (59, 60). These adaptors mediate TLR signaling either alone or in combination with MyD88, and confer pathogen-associated molecular pattern stimulus specificity of inflammatory responses. Recently, it was shown that 76% of >900 genes induced in LPS-treated wild-type macrophages are induced via a MyD88-independent pathway (61). Additionally, Fitzgerald et al. (62) demonstrated that LPS or poly(I:C) treatment of macrophages from MyD88-deficient and wild-type mice induced comparable levels of the chemokine RANTES (CCL5) and activation of IFN regulatory factor-3, a transcription factor necessary for the expression of IFN-β and RANTES genes.

Results from the current study show that the primary adaptor molecules used by filarial and Wolbachia preparations are MyD88 and Mal, which is consistent with a TLR2/TLR6-dependent response. Peritoneal macrophages from MyD88- and Mal-deficient mice stimulated with Wolbachia containing filarial extracts or isolated Wolbachia showed a complete loss of TNF-α and IL-6 responses. In contrast, the TLR4 receptor complex signaling uses adaptor molecules MyD88 and Mal in the classical pathway, and TRIF and TRAM in the MyD88-independent signaling pathway, leading to the up-regulation of RANTES and IFN-inducible genes (56, 57, 59). We show that TRIF and TRAM do not have a role in filaria and Wolbachia-induced up-regulation of TNF-α and IL-6, indicating that the MyD88-independent signaling pathway is not involved in these responses.

In some experiments, we used enriched Wolbachia preparations from the filarial nematode O. ochengi, a species of filarial worm that infects cattle. It is closely related to the human pathogen O. volvulus on the basis of 16S ribosomal RNA analyses (18) and shares the same vector, Simulium damnosum (63, 64, 65, 66). Wolbachia bacteria belonging to two distinct phylogenetic clades, those that are endosymbionts of the filarial nematode O. ochengi and those from a mosquito cell line Aedes albopictus, induced remarkably similar innate immune receptors and intracellular signaling pathways. Wolbachia from the filaria and from the insect cell line used exclusively TLR2/TLR6.

We propose a model for the pathogenesis of early innate immune responses in filarial diseases whereby Wolbachia ligands, such as the major surface protein WSP or other molecules are released through either secretion from the nematode or the death of either adult nematodes or the larger biomass of circulating microfilaria, releasing the endosymbiont Wolbachia. These molecules activate TLR2/TLR6 on the surface of resident innate immune cells in the tissues such as corneal fibroblasts in ocular onchocerciasis, or resident macrophages and endothelial cells in the lymphatic and perilymphatic tissues in lymphatic filariasis. TLR2/TLR6 signaling primarily through MyD88 and Mal stimulates production of proinflammatory and chemotactic cytokines that leads to the release of vasoactive molecules, vascular endothelial cell activation leading to capillary leakiness, recruitment of neutrophils, macrophages, and other effector cells to the affected tissue, and the propagation of the inflammatory response. Continued and repeated inflammation in the perilymphatic tissues in lymphatic filariasis may induce changes in lymphatic vessel architecture and lymphatic dilatation and eventual lymphatic scarring and lymphedema.

In onchocerciasis, chronic cellular infiltration and activation lead to permanent structural changes in the cornea such as scarring and opacification, associated with visual impairment and blindness (67). The role of potential anti-inflammatory factors released by the filaria and the role of adaptive host responses are not addressed in this study and potentially add to the complexity of the pathogenesis of filarial diseases. Turner et al. (68) have recently shown that immune tolerance induced by Wolbachia containing Brugia extracts to a range of TLR ligands is dependent on TLR2 and MyD88 but not TLR4. In our recent study of O. volvulus induced neutrophil recruitment in the cornea, Wolbachia–induced production of proinflammatory and chemotactic cytokines, development of corneal haze, and neutrophil activation were completely abrogated in MyD88^−/− mice (27), consistent with the data presented in this study. In summary, results from the current study extend our previous observations by demonstrating that the predominant innate receptors for Wolbachia and Wolbachia containing filaria are TLR2/TLR6 and not TLR4, and that signaling is entirely MyD88/Mal dependent. Host genetic polymorphisms in these key innate immune molecules involved in the inflammatory response to Wolbachia may play a critical role in the pathogenesis of clinical filarial diseases. Knowledge of the critical innate pathways involved in inflammatory responses to filaria and Wolbachia will assist in understanding the host pathogen interactions involved in the pathogenesis of these debilitating filarial diseases.

We are grateful to Dr. Shizuo Akira (Osaka University, Osaka, Japan) for the gifts of TLR2, TLR4, and TLR6 and adaptor molecule MyD88, Mal, TRIF, and TRAM-deficient mice.

The authors have no financial conflict of interest.