A Critical Role for Type I IFN in Arthritis Development following Borrelia burgdorferi Infection of Mice

Bacterial culture and mouse infection

C3H/HeN mice were obtained from Charles River Laboratories or the breeding facility of the National Cancer Institute (NCI; Frederick, MD), depending on availability. C57BL/6 mice were purchased from National Cancer Institute, and C3H/HeJ and C3H SCID mice were obtained from The Jackson Laboratory. C3H TLR2^−/− mice were originally generated by Deltagen and obtained from Tularik (now Amgen) (26). C57BL/6 MyD88^−/− mice were kindly provided by Dr. S. Akira (Hyogo College of Medicine) (27), and C57BL/6 TLR9^−/− mice were obtained from the Mutant Mouse Resource at the University of California, Davis, CA. C57BL/6 IFN-α/β receptor subunit 1 (IFNAR1)^−/− mice were a gift from Dr. M.-K. Kaja (University of Washington, Seattle, WA). All mice were housed at the University of Utah Animal Research Center (Salt Lake City, UT) and used with strict adherence to all institutional policies and guidelines established for the care and use of laboratory animals in biomedical research.

A clonal isolate of B. burgdorferi strain N40 (cN40), provided by S. Barthold (University of California, Davis, CA) (6), was propagated in BSKII medium (28). For infection experiments, 2000 cN40 spirochetes were intradermally injected into the shaved back of 6-wk-old female mice in a volume of 20 μl. Uninfected control mice received 20 μl of BSKII medium. Mice were sacrificed at 1, 2, or 4 wk postinfection and ankle measurements taken with a metric caliper were compared with preinfection measurements and recorded as the change in ankle measurement (29). For mock-infected control mice as well as animals euthanized at 1 and 2 wk postinfection, all skin was carefully removed, and both rear ankle joints were excised and separately snap frozen in a dry ice-ethanol bath for subsequent RNA extraction and RT-PCR analysis of IFN-inducible genes.

At 4 wk postinfection, the most swollen ankle joint, as assessed by metric caliper measurement, was excised and placed in 10% neutral-buffered formalin (Sigma-Aldrich). Samples were then decalcified, paraffin embedded, sectioned, and stained with H&E by AML Laboratories. Arthritis severity was assessed in a blinded fashion, and overall scores for joints obtained from both mock-infected and B. burgdorferi-infected mice were assigned based on a scale of 0 (no arthritis) to 5 (most severe arthritis) according to the following parameters: overall lesion appearance, degree of neutrophil infiltration, amount of monocytic infiltration, tendon sheath thickness, and the presence of reactive and reparative responses. The least swollen joint was removed at 4 wk postinfection for quantitative PCR assessment (qPCR)³ of B. burgdorferi numbers within the joint.

Abs and depletion protocols

The following Abs were purchased from eBioscience and used for flow cytometry analyses: FITC-conjugated anti-DX5 (CD49b) pan-NK cell Ab, PE-conjugated anti-CD3ε, PE-conjugated anti-Armenian hamster IgG, FITC-conjugated anti-rat IgM, FITC-conjugated anti-CD11c, PE-conjugated anti-B220, and FITC-anti-Hamster IgG. PE-anti-rat IgG2a was purchased from BD Pharmingen.

MAR1-5A3, a mouse anti-mouse IgG1 mAb directed against IFNAR1, was used to inhibit type I IFN responses in C3H/HeN mice (30). The mouse IgG1 mAb MOPC-21 (BioExpress Cell Culture Services) or the mouse anti-human IFNGR1 Ab GIR-208 (31) was used as an isotype control. Groups of five age-matched, female C3H/HeN mice received a i.p. injection of PBS or 2.5 mg of either MAR1-5A3 or an appropriate isotype control 1 day before B. burgdorferi infection. IFN-γ was neutralized using the hamster H22 mAb (32). The hamster anti-GST mAb PIP was used as an isotype control. C3H/HeN mice were injected i.p. with 0.5 mg of either H22 or PIP or an equal volume of PBS 2 days before B. burgdorferi infection. H22 (0.25 mg) or PIP (0.25 mg) or PBS was administered on the day of infection (day 0), and booster injections were given every 7 days thereafter for the duration of the study. To block both type I IFN and IFN-γ responses, an i.p. injection of either 2.5 mg of MAR1-5A3 plus 0.5 mg of H22 or 2.5 mg of MOPC-21 plus 0.5 mg of hamster IgG or an equal volume of PBS was administered 1 day before B. burgdorferi infection. The rabbit polyclonal Ab anti-asialo GM1 (Wako Chemicals) was used to deplete NK cells from C3H mice (33). Groups of five female, age-matched mice received 50-μl (1.8 mg) i.p. injections of α-asialo GM1, rabbit IgG (Sigma-Aldrich), or an equal volume of PBS on the day of infection (day 0) and on day 4. Anti-asialo GM1 administration resulted in a 75% depletion of NK and NK T cells from the spleen as assessed by flow cytometry using the FITC-conjugated anti-DX5 (CD49b) pan-NK cell Ab. Plasmacytoid dendritic cells (pDCs) were blocked with the anti-pDC Ag 1 (PDCA-1) mAb (Miltenyi Biotec). Groups of 10 female, age-matched mice were injected i.p. with 500 μg of anti-PDCA-1, rat IgG, or an equal volume of PBS 1 day before B. burgdorferi infection. Animals were then boosted with 250 μg of anti-PDCA-1 or appropriate controls on the day of infection.

RNA extraction and RT-PCR analysis

Rear ankle joints excised from B. burgdorferi- and 16 mock-infected mice were homogenized in chilled guanidine thiocyanate (Roche) and RNA was extracted via cesium chloride ultracentrifugation (34). RNA was extracted from cultured cells with TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. RNA samples were subsequently reverse transcribed into cDNA as described elsewhere (15) and LightCycler-based PCR amplification was conducted using the LightCycler FastStart DNA Master^Plus SYBR Green I kit for the LightCycler 2.0 instrument (Roche), as previously described (35), or the LightCycler 480 (LC480) SYBR Green I master mix for the LightCycler 480 machine (Roche). The number of transcript copies present in the starting template sample were calculated and normalized to 1000 copies of the mouse β-actin gene as mentioned elsewhere (35). The following primer pairs span introns and were used to amplify cytokines and IFN-responsive transcripts: Oasl2 forward 5′-TGGGAAGAAGAAAGGGATGGG-3′ and Oasl2 reverse 5′-GGGTTTGTCCAGAGTGTCCAATC-3′; Cxcl9 forward 5′-TTGGGCATCATCTTCCTGGAGCAG-3′ and Cxcl9 reverse 5′-GAGGTCTTTGAGGGATTTGTAGTGG-3′; Il-6d forward 5′-ACAACCACGGCCTTCCCTACTT-3′ (36) and Il-6b reverse 5′-TCTCTGAAGGACTCTGGCTTTGTC-3′ (LightCycler only); Il-6 forward 5′-ACCGCTATGAAGTTCCTCTCTGC-3′ and Il-6d reverse 5′-CCAGAAGACCAGAGGAAATTTTC-3′ (LightCycler 480 only); Il-12p40 forward 5′-AGATGACATCACCTGGACCT-3′ and Il-12p40 reverse 5′GCCATGAGCACGTGAACCGT-3′. The primer sequences for β-actin gene, Igtp, Iigp1, and 16S rRNA have been reported previously (15, 37).

DNA extraction and qPCR analysis

DNA was extracted from mouse tissues as described elsewhere (38) using sequential digestions with a 0.1% collagenase A solution and a 0.2 mg/ml proteinase K solution. Following a series of phenol/chloroform extraction and ethanol precipitation steps, the purified DNA was diluted to 5 μg/ml for use in qPCR assays. Continuous fluorescence monitoring of PCR with SYBR Green I as the detection dye (Molecular Probes) was conducted with the ntm17 primer pair and normalized to 1000 copies of the mouse nidogen gene as previously described (38).

Collagenase release of ankle cells

To release cells from the ankle tissue, both rear ankles were excised following skin removal from uninfected or B. burgdorferi-infected C3H or C57BL/6 mice 7 or 14 days postinfection and incubated for 2 h at 37°C in RPMI 1640 (Invitrogen) containing 10% FCS and 1 mg/ml collagenase A (Roche) as described previously (16). Two-color flow cytometry analysis was conducted to assess whether the numbers of conventional dendritic cells (cDCs) (CD11c⁺B220⁻) or pDCs (CD11c⁺B220⁺) increase in the ankle joints following B. burgdorferi infection. The increase in cDC and pDC cell numbers was calculated relative to values obtained for uninfected mice.

Cell culture

Bone-marrow derived macrophages were isolated from the femurs and tibias of mice as previously described (39) by culture in RPMI 1640 medium (Invitrogen) supplemented with 30% L929 conditioned medium and 20% horse serum (HyClone) demonstrated to be free of Mycoplasma with the MycoProbe Mycoplasma detection kit (R&D Systems). Harvested macrophages were replated in 6-well dishes at a density of 6 × 10⁵/ml in medium lacking serum and containing 1% Nutridoma. After an overnight incubation, the medium was removed and replaced with medium alone or medium containing 7.4 × 10⁶/ml live B. burgdorferi cN40 (washed twice in PBS to remove BSKII components). Macrophages were stimulated at 37°C in 5% CO₂ for 24 h, after which the cells were harvested for RNA extraction.

To select for cDCs using a low-density protocol, bone marrow cells were cultured in RPMI 1640 medium (Invitrogen) containing GM-CSF (provided by R. Daynes, University of Utah), as previously described (40). Harvested cells were magnetically labeled with anti-CD11c microbeads, applied to MACS columns (Miltenyi Biotec), and the CD11c⁺ cells retained by the column were eluted. The purity of these cells was confirmed via flow cytometry staining using anti-CD11c and anti-B220 Abs. On average, 90% of the cells recovered using this protocol were CD11c⁺B220⁻. Monocyte-derived dendritic cells (1.5 ml; 1.5 × 10⁶ cells) and 1.5 ml of medium only, medium plus 1000 U/ml rIFNβ (Pestka Biomedical Laboratories InterferonSource), 7.4 × 10⁶/ml live B. burgdorferi, B. burgdorferi plus IFNβ, 10 μg/ml anti-OspA (determined to be negative for Mycoplasma) (41), or B. burgdorferi plus anti-OspA were added to 17 × 100-mm snap-cap tubes (Falcon) and incubated for 24 h at 37°C in 5% CO₂. Cells were then collected by centrifugation and RNA was extracted.

pDCs were isolated from the spleens of mice that had received daily i.p. injections of rFlt3L (BioExpress Cell Culture Services) for 9 days (42). Spleens were removed and dissected into RPMI 1640 medium containing 2 mg/ml collagenase A and 20 μg/ml DNase I. pDCs were magnetically labeled with anti-PDCA1 microbeads (Miltenyi Biotec) and recovered via separation over a MACS column, followed by elution of PDCA1+ cells. Purity was assessed via flow cytometry analysis using anti-CD11c and anti-B220 Abs. On average, 95% of the cells recovered using this protocol were CD11c⁺B220⁺. pDCs were seeded into 24-well plates at 7 × 10⁵/ml in 1 ml of either complete medium or medium containing the following stimuli: 100 U/ml rIFN-α (Pestka Biomedical Laboratories InterferonSource), 7.4 × 10⁶/ml live B. burgdorferi, B. burgdorferi plus IFN-α, 10 μg/ml anti-OspA (41), or B. burgdorferi plus anti-OspA. Following a 24-h incubation at 37°C in 5% CO₂, cells were harvested for RNA extraction.

Statistics

The nonparametric Mann-Whitney U test was used for data values obtained from PCR, RT-PCR, and histopathological scoring of arthritis lesions. One-way ANOVA with the Bonferroni posttest was used to assess the significance of data obtained from ankle measurements and collagenase-digested ankle joints. All statistical analyses were conducted using GraphPad InStat3 version 3.0b for MacIntosh (GraphPad Software). For all tests, A value of p < 0.05 was considered significant.

The adaptive immune response is not required for induction of IFN profile genes in C3H mice

Previous studies demonstrated that a large number of IFN-responsive genes were significantly up-regulated by 1 wk after B. burgdorferi infection within the rear ankle joints of arthritis-susceptible C3H mice, but not in mildly arthritic C57BL/6 mice, (15). Elevated transcription of IFN-responsive genes could be linked to the severe arthritis phenotype manifested by C3H but not C57BL/6 mice at later time points. Alternatively, triggering of these genes might occur as a result of early infiltrating B and T lymphocytes that have been reported to secrete IFN-γ in response to B. burgdorferi (43, 44, 45, 46). To distinguish between these two scenarios, C3H scid mice, which develop severe arthritis (9), were infected with B. burgdorferi, sacrificed at 1 wk postinfection, and quantitative RT-PCR of selected IFN-responsive genes was conducted on RNA isolated from both rear ankle joints. Fig. 1⇓ depicts transcript levels obtained for four IFN-inducible genes: the IFN-γ inducible GTPase gene Igtp (Fig. 1⇓A); the IFN-inducible GTPase 1 gene Iigp1 (Fig. 1⇓B); the oligoadenylate synthase-like 2 gene, Oasl2 (Fig. 1⇓), and the T cell recruiting chemokine gene Cxcl9 (Mig) (Fig. 1⇓D). Although the majority of these genes have been previously annotated as IFN-γ inducible, increased transcription of these genes can also be triggered by type I IFN (47, 48, 49). These four genes were selected as representatives of the IFN-responsive profile, as Affymetrix microarray analyses revealed these to be among the most highly up-regulated genes at 1 wk postinfection. For example, both Igtp and Iigp1 expression levels were elevated >100-fold vs values recorded for uninfected mice (15). Each of the four examined IFN-responsive gene transcripts were elevated within the ankle joints of C3H scid mice to similar levels, relative to uninfected mice, as those obtained for wild-type C3H mice (Fig. 1⇓). These data indicate that early recruitment/stimulation of adaptive lymphocytes is not responsible for the transcription of IFN-responsive genes within the ankle joints of C3H mice at 1 wk postinfection. Instead, induction of these genes is mediated by cells of the innate immune response.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

B and T lymphocytes do not contribute to the induction of IFN-responsive genes. RT-PCR analysis of ankle joint RNA obtained from C3H scid mice at 1 wk after B. burgdorferi infection is shown. Igtp (A), Iigp1 (B), Oasl2 (C), and Cxcl9 (D) transcripts were all elevated to similar levels, relative to uninfected animals, as those obtained for wild-type C3H mice. Transcript levels were normalized to 1000 copies of β-actin and values shown represent the average ± SD for four mice per group.

IFNAR1 blockage results in decreased arthritis severity and diminished IFN-responsive gene expression, but does not alter host defense

A previous report by Brown and Reiner demonstrated that C3H IFN-γ^−/− mice develop severe arthritis following B. burgdorferi infection, indicating that IFN-γ is not required for Lyme arthritis development (17). However, much controversy has remained in the field about the impact of IFN-γ on arthritis development and resolution (10, 50, 51). When combined with our observation that IFN-responsive gene induction in C3H scid mice mimics that seen in wild-type mice (Fig. 1⇑), it is highly probable that transcription of any IFN-inducible gene associated with arthritis development would be triggered by innate immune cells and could be a consequence of type I IFN. In line with this hypothesis, the contribution of type I IFN to Lyme arthritis severity in C3H mice at 4 wk postinfection was assessed using a mouse mAb directed against murine IFNAR1 administered 1 day before infection. Overall, mice injected with the anti-IFNAR1 mAb exhibited reduced ankle swelling compared with values obtained for mice treated with PBS or IgG1. Individually, a broad spectrum of arthritis phenotypes was noted for joints obtained from IFNAR1-blocked mice, ranging from a drastic reduction in ankle swelling for two mice, an intermediate phenotype for two mice, and moderately severe ankle swelling for one mouse (Fig. 2⇓A). There was significant concordance between ankle swelling measurements and histopathology, as the two mice exhibiting the least amount of ankle swelling also received greatly reduced lesion scores. A significant reduction in lesion pathology was noted in anti-IFNAR1-treated mice when compared with IgG1-treated mice (Fig. 2⇓B). Anti-IFNAR1-injected mice displayed decreased edema, fewer neutrophils, and less thickening of the cranial tibial tendon sheath (Fig. 3⇓D). These observations were in stark contrast to those noted for PBS- or IgG1-treated mice, both of which exhibited extensive edema, neutrophil infiltration, and thickening of the tendon sheath (Fig. 3⇓, B and C), features that are classic hallmarks of Lyme arthritis (5, 6). Similar results were also obtained in another experiment. To determine whether blockage of type I IFN impacts immunological clearance of B. burgdorferi, qPCR was conducted on DNA extracted from ankle joints of anti-IFNAR1-injected mice. There was no significant difference in spirochete numbers within the joints of IFNAR1-inhibited mice vs those detected in PBS- or IgG1-treated mice (Fig. 2⇓C).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Anti-IFNAR1-treated C3H mice exhibit decreased arthritis severity at 4 wk after B. burgdorferi infection without impacting host defense. Anti-IFNAR1-treated C3H mice exhibit decreased ankle swelling (A) and histopathological lesion scores (B) relative to PBS- and IgG1-treated animals; p < 0.05 vs both control groups as determined by one-way ANOVA with Bonferroni as the posttest in A; in B, ∗, p < 0.05 vs IgG1 and PBS (B), by Mann-Whitney U test. C, Spirochete numbers were detected within joint tissues analyzed from each Ab treatment group by qPCR using RecA and normalized to the single-copy mouse gene nidogen. Data shown are the averages ± SD for five mice per group obtained from one experiment and are representative of results obtained from two separate experiments.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Blockage of IFNAR1 suppresses many key histopathological features of Lyme arthritis. H & E-stained sections of rear ankle joints obtained from uninfected (A) or, at 4 wk postinfection, from PBS-treated (B), IgG1-treated (C), or anti-IFNAR1 (α-IFNAR1)-treated (D) mice were examined for key features of Lyme arthritis, including thickening of the cranial-tibial tendon sheath and neutrophil infiltration. Although extensive tendon sheath thickening and neutrophil infiltration are evident in both the PBS (B) and IgG1-treated (C) mice, animals receiving anti-IFNAR1 Ab (D) have significantly reduced tendon sheath thickening and neutrophil infiltration. All images were obtained using a ×4 magnification. The arrow points to the cranial tibial tendon sheath. Data shown were obtained from one experiment and are representative of results obtained from two separate experiments.

To determine whether the decrease in arthritis severity observed as a result of IFNAR1-blockage was linked to a suppression of IFN-inducible genes, the effect of anti-IFNAR1 pretreatment in the triggering of IFN-responsive genes at 1 wk postinfection was also assessed. Transcript levels for all four examined IFN-responsive genes were significantly suppressed within the ankle joints of C3H mice that received the anti-IFNAR1 blocking Ab when compared with message levels obtained for infected mice treated with either PBS or IgG1. The diminished induction of transcription observed for these genes in the presence of anti-IFNAR1 was striking, as message levels were reduced by 68% for Cxcl9, 70% for Iigp1, and 81% for Igtp. Intriguingly, induction of Oasl2 transcription was completely suppressed, as message levels in anti-IFNAR1-treated mice were reduced to the baseline level obtained for uninfected C3H mice (Table I⇓). Together, these data provide strong evidence that type I IFN is involved in induction of the IFN-responsive profile at 1 wk postinfection. The above results also strongly link type I IFN with arthritis development and indicate that type I IFN may partially exert its proarthritic effects through enhanced transcription of IFN-responsive genes by innate immune cells residing within ankle tissue. The kinetics of B. burgdorferi dissemination within anti-IFNAR1-blocked mice was also monitored at 1 wk postinfection to assess whether the spirochete burden within the rear ankle joints differed between these mice and those receiving IgG1 or PBS before infection. Quantitative RT-PCR using primers specific for 16S rRNA revealed low but overlapping B. burgdorferi numbers within the joints of mice from all treatment groups at 1 wk postinfection (data not shown), indicating that administration of anti-IFNAR1 did not alter bacterial numbers in the joints.

IFNγ blockage has no effect on arthritis development or host defense to B. burgdorferi but does suppress induction of IFN-responsive gene transcripts at 1 and 2 wk postinfection

As an additional test of the selective effect of type I IFN on arthritis development, we decided to inhibit type II IFN. Mice were administered anti-IFN-γ blocking Ab, PBS, or hamster IgG at various intervals both before and following B. burgdorferi infection. Blockage of IFN-γ did not impact arthritis development in mice examined at 4 wk postinfection (data not shown), consistent with a previous report that IFN-γ^−/− C3H mice develop severe Lyme arthritis similar to that noted in wild-type mice (17). The effect of IFN-γ inhibition on B. burgdorferi numbers within ankle tissues at 4 wk postinfection was assessed by qPCR and there were no significant differences in the number of spirochetes detected within the ankle joints of anti-IFN-γ treated mice, when compared with values obtained for control treatment groups (data not shown). These data support the previous conclusion of Brown and Reiner that the control of spirochete numbers within the ankle joints of infected C3H mice is IFN-γ independent (17).

Although arthritis proceeded uninhibited in IFN-γ-blocked mice, anti-IFN-γ-treated mice exhibited a striking 100% reduction in the up-regulated expression of the IFN-responsive genes Igtp and Cxcl9 when compared with mice receiving PBS or hamster IgG at 1 wk postinfection. Iigp1 transcript levels were 97% suppressed in anti-IFN-γ-blocked mice (Table II⇓). Oasl2 gene induction can be mediated by either type I IFN or IFN-γ (47, 48). Interestingly, IFN-γ inhibition caused a 73% suppression of Oasl2 (Table II⇓). In light of the 100% reduction in Oasl2 message levels obtained with the anti-IFNAR1 blocking Ab (compare Table I⇑ and Table II⇓), these data suggest that Oasl2-induced expression is more effectively reduced by type I IFN blockage. Our previous microarray study found that although the expression of IFN-responsive genes declined by 2 wk after B. burgdorferi infection, message levels for many transcripts were still severalfold higher than those obtained for uninfected mice (15). For this reason, the effect of IFN-γ blockage was also assessed at 2 wk postinfection. Markedly reduced transcript levels (90–100% suppression) were obtained for Igtp, Iigp1, and Cxcl9, whereas anti-IFN-γ treatment had little effect on Oasl2, suppressing its expression by only 43% when compared with PBS- and hamster IgG- treated mice (Table II⇓). Analyses of 16S rRNA transcript levels indicated that similar low levels of B. burgdorferi were present within the joints of mice from all treatment groups at both 1 and 2 wk postinfection (data not shown), indicating that IFN-γ blockage did not alter the kinetics of bacterial dissemination. Finally, we assessed the effect of simultaneous blockage of both type I IFN and IFN-γ on the induction of IFN-responsive genes at 1 wk postinfection. Induced expression of all four examined IFN-responsive transcripts was 100% abolished within the joints of C3H mice injected with both anti-IFNAR1 and anti-IFN-γ, as message levels were suppressed below the baseline values obtained for untreated, uninfected mice (Table III⇓). Together, these results support the unique role of type I IFN in arthritis development but also provide compelling evidence that induction of IFN-responsive genes within the joints of C3H mice at 1 wk postinfection reflects the contribution of both type I and type II IFNs.

pDCs and cDCs traffic to the joints following infection but are not the initial source of B. burgdorferi-induced type I IFN

In an attempt to identify the cell type(s) responding to type I IFN to trigger arthritis development, we next questioned whether dendritic cells, either resident in or trafficking to the joint following infection, are involved in the production of type I IFN and/or in Lyme arthritis development. pDCs have long been recognized as the major producers of and responders to type I IFN, but cDCs also possess the ability to both synthesize and react to type I IFN (52, 53). To this end, we assessed whether pDCs and/or cDCs traffic to the joint following B. burgdorferi infection. Cells released from collagenase-digested ankle joints excised from C3H mice at 7 and 14 days postinfection were analyzed for CD11c and B220 surface expression and compared with the values obtained from uninfected ankle joints. By 7 days postinfection, an increase in the numbers of both CD11c⁺B220⁻ (myeloid or cDCs) and CD11c⁺B220⁺ (pDCs) cells was observed in C3H ankle joints, with even greater numbers observed at 14 days postinfection (Fig. 4⇓). However, blockage of pDCs in C3H mice using anti-PDCA-1 Ab had no impact on arthritis severity (data not shown). Taken together, these results demonstrate that infiltration of both cDCs and pDCs by 1 wk postinfection may occur as a response to infection and inflammation but that pDCs do not mediate arthritis development in C3H mice.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

pDCs and cDCs traffic to B. burgdorferi-infected mouse joints as early as 7 days postinfection and are further elevated at 14 days postinfection. Collagenase-released ankle cells obtained from three C3H mice at each time point were stained with CD11c and B220 and analyzed by flow cytometry. The numbers of CD11c⁺B220⁻ (cDC) and CD11c⁺B220⁺(pDC) cells were determined at each time point. For pDCs, ∗ represents p < 0.05 at 14 days postinfection vs uninfected mice, and for cDCs ∗ represents p < 0.05 at 14 days postinfection vs uninfected mice by one-way ANOVA with Bonferroni as the posttest. Data shown are the averages ± SD obtained from one experiment and are representative of results obtained from three separate experiments.

However, because increased numbers of DCs were detected in infected joint tissue as early as 1 wk postinfection, a time frame coincident with that of IFN-responsive gene induction, we also investigated the ability of both splenic pDCs and bone marrow-derived cDCs to induce expression of the IFN profile in response to in vitro B. burgdorferi stimulation. Although both cDCs and pDCs exhibited elevated IL-6 and/or IL-12 transcripts and protein (data not shown) in the presence of live B. burgdorferi, IFN-responsive transcripts were not induced in response to bacterial stimulation alone. In addition, pDCs stimulated in the presence of IFN-α and B. burgdorferi also failed to up-regulate IFN-inducible genes. However, a synergistic increase in expression was detected in cDCs administered IFNβ and live spirochetes, suggesting that priming may be required for optimal cDC contribution to the IFN response (Table IV⇓). It has been reported that pDCs exhibit a diminished capacity for phagocytosis of Ags relative to other APCs, such as cDCs and macrophages (52). To enhance the ability of pDCs to internalize live bacteria, Ab to the abundantly expressed outer surface protein OspA was added to pDCs or cDCs before the addition of B. burgdorferi. Unexpectedly, incubation of cDCs with an anti-OspA Ab resulted in induction of all four examined IFN-responsive genes, but not in IL-12 or IL-6 transcripts. In the presence of opsonizing Ab, neither Borrelia-stimulated pDCs nor cDCs exhibited further up-regulation of IFN-responsive genes (Table IV⇓). Taken together, the above data strongly imply that whereas DCs may contribute either to the amplification of IFN profile genes or to the resolution of this response in vivo, they are unlikely to be the first cell type driving the transcription of type I IFN-responsive genes in response to B. burgdorferi infection.

Bone marrow-derived macrophage induction of IFN-responsive genes in response to in vitro B. burgdorferi-stimulation is type I IFN-dependent and occurs via a MyD88-independent signaling pathway

Another innate immune candidate cell, the bone marrow-derived macrophage (BMMφ), was then evaluated for its ability to up-regulate IFN-inducible genes in response to B. burgdorferi. Twenty-four hours after the addition of B. burgdorferi, transcripts for Igtp (Fig. 5⇓A), Iigp1 (Fig. 5⇓B), Oasl2 (Fig. 5⇓C), and Cxcl9 (Fig. 5⇓D) were up-regulated in BMMφ isolated from both C57BL/6 and C3H/HeN mice. BMMφ from C3H/HeJ mice, which lack a functional Tlr4 gene (54), also displayed up-regulation of IFN-responsive transcripts, indicating that transcription of these genes was not due to low-level LPS contamination of our B. burgdorferi grown in complex medium (Fig. 5⇓). To assess whether borrelial lipoproteins or unmethylated CpG DNA were being sensed by TLRs 2 and 9, respectively, to enact transcription of IFN-inducible genes, BMMφ isolated from TLR2^−/− and TLR9^−/− mice were stimulated with B. burgdorferi. Transcription of IFN-inducible genes proceeded in the absence of TLR 2 or 9, indicating that these TLRs are not used by B. burgdorferi to up-regulate IFN-responsive transcripts. Induction of IFN-responsive transcripts is also MyD88 independent, as MyD88^−/− BMMφ stimulated with B. burgdorferi up-regulated IFN profile genes. Intriguingly, BMMφ isolated from IFNAR1^−/− mice failed to induce IFN-responsive genes as a result of B. burgdorferi stimulation. This observation provides strong evidence that type I IFN modulates transcription of IFN-responsive genes in macrophages and underscores the utility of the macrophage as a model cell to examine the mechanisms of IFN-responsive gene induction following B. burgdorferi infection.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

BMMφ isolated from both C3H and C57BL/6 mice induce IFN-responsive genes following B. burgdorferi stimulation in a type I IFN-dependent, MyD88-independent fashion. RT-PCR was performed on BMMφ prepared as described in Materials and Methods from the following strains of mice: C57BL/6 (B6), C3H/HeN, C3H/HeJ, C3H TLR2^−/−, B6 TLR9^−/−, B6 MyD88^−/−, and B6 IFNAR1^−/− mice. Transcript levels for Igtp (A), Iigp1 (B), Oasl2 (C), and Cxcl9 (D) were normalized to β-actin. Data shown are the averages ± SD obtained from one experiment and are representative of results obtained from three separate experiments.

NK cell depletion has no effect on arthritis development or host defense to B. burgdorferi at 4 wk postinfection, but NK cells are major IFN-γ contributors to induction of the IFN-responsive profile at 1 wk postinfection

Our above results demonstrating that IFN-γ blockage suppressed IFN-responsive gene transcription without impacting arthritis severity was intriguing (Table I⇑), especially in light of conflicting data in the field concerning a role for IFN-γ in Lyme arthritis development (17, 50, 51). Our finding that IFN-responsive genes are still induced in C3H scid mice (Fig. 1⇑) indicated that cells other than T lymphocytes were responsible for both the type I- and IFN-γ-inducible portions of the profile. NK cells are potent producers of IFN-γ and have been previously shown to secrete this cytokine in response to B. burgdorferi (55) (56). NK cells were depleted from C3H mice using anti-asialo GM1 to determine whether these cells impact arthritis development or host defense to B. burgdorferi. Anti-asialo GM1 was used to neutralize NK cells because C3H mice produce the NK1.2 allele, to which there are no commercially available Abs. In agreement with data previously obtained with the footpad-injection model (56), ankle swelling measurements and histopathological lesion scores obtained for anti-asialo GM1-injected mice did not differ significantly from those recorded for PBS- or vehicle-treated animals (data not shown). Similar numbers of spirochetes were present both in the joints of anti-asialo GM1-treated mice and in control mice at 4 wk postinfection (Table V⇓), indicating that NK cell depletion does not affect the ability of the host immune system to respond to B. burgdorferi challenge. We also examined whether NK cells are a major source of the IFN-γ driving the induction of the IFN-profile at 1 wk after B. burgdorferi infection. Strikingly, mice treated with anti-asialo GM1 exhibited a significant reduction (77–98%) in the expression levels of Igtp, Iigp1, and Cxcl9 when compared with message levels obtained for PBS- and rabbit IgG-treated mice. In contrast, Oasl2 message levels were only decreased by 54% within the joints of anti-asialo GM1 treated mice (Table V⇓), an observation that supports other data presented in this report indicating that the magnitude of Oasl2 expression shows greater dependence on type I IFN. These data clearly demonstrate that NK cells are a major IFN-γ source contributing to induction of IFN-responsive profile genes at 1 wk postinfection. However, as we have reported here in confirmation of previously published reports (17, 56), NK cells and IFN-γ production do not contribute to arthritis development, reinforcing the idea that the entire IFN profile is not required for arthritis development.