Local Production of Interferon Gamma by Invariant Natural Killer T cells Modulates Acute Lyme Carditis

Mice and infection with B. burgdorferi

Jα18-deficient mice (23) were bred at UMass Amherst. The mice have been backcrossed to the C57Bl/6J (B6) background more than 10 generations. Age-matched B6 mice were from Jackson laboratories. IFNγRα-deficient mice in a 129 Sv/J background were purchased from Jackson laboratories and bred at UMass Amherst. Six to 8 week old mice were infected by subcutaneous injection with 10⁵ B. burgdorferi strain 297 in the midline of the back, as previously described (24). At 2 weeks of infection, which represents the acute phase of disease, the mice were sacrificed and analyzed for inflammation and bacterial burdens, as previously described (25). Arthritis and carditis were evaluated histologically in formalin-fixed sections processed for H&E staining. The joints were also decalcified. The hearts were cut in half through bisections across the atria and ventricles prior to sectioning. Signs of arthritis were evaluated as described (25), based on a combined assessment of histological parameters of B. burgdorferi-induced inflammation, such as exudation of fibrin and inflammatory cells into the joints, alteration in the thickness of tendons or ligament sheaths, as well as hypertrophy and hyperplasia of the synovium. Signs of carditis were evaluated based on the cardiac inflammatory infiltrate, including the infiltration of connective tissue with macrophages at the base of the heart, including surrounding the aortic valve and the atria. Carditis was scored on a scale of 0 (no inflammation), 1 (mild inflammation with less than two small foci of infiltration), 2 (moderate inflammation with 2 or more foci of infiltration), or 3 (severe inflammation with focal and diffuse infiltration covering a large area). Both joint and heart tissue sections were blindly examined. The Institutional Animal Care and Use Committee at UMass Amherst approved all procedures involving animals.

Cells and stimulations

Mouse bone marrow cells were collected from the femoral shafts by flushing 3 times with 1 ml of cold RPMI (10% FCS and 1% penicillin/streptomycin) supplemented with 20% L929-conditioned RPMI. The cell suspensions were cultured in 100mm × 15mm petri dishes (Fisher Scientific) in 20% L929-conditioned RPMI for 8 days at 37° C with 5% CO₂. Following incubation, non-adherent cells were eliminated and adherent macrophages scraped, counted and resuspended in serum-free RPMI medium without antibiotics prior to use. RAW264.7 macrophages were cultured in RPMI but washed and resuspended in serum- and antibiotic-free RPMI before use. 10⁶ BMMs or RAW264.7 cells per ml were incubated with recombinant murine IFNγ (rmIFNγ; carrier- and endotoxin-free; eBioscience, Inc., San Diego, CA) (0, 0.5, 5, 50 ng ml^-1) for 12h and analyzed for expression of CD1d (PE-anti-mouse CD1d, eBioscience, Inc.) or stimulated with live B. burgdorferi (strain 297) at a 25:1 multiplicity of infection (m.o.i) for 12h. Following stimulation with B. burgdorferi, culture supernatants were analyzed by ELISA for TNFα and IL-6, as previously described (26).

B. burgdorferi-specific antibodies

Analysis of Total IgG and IgM were determined by coating 96 well plates with 0.5 μg/ml (IgG) or 1.0 μg/ml (IgM) of a B. burgdorferi strain 297 lysate and incubated with serial two-fold dilutions starting at 1:100 followed by incubation with HRP-conjugated anti mouse IgM or IgG (BD Pharmingen) (1:10000 dilution). The reactions were developed using 1-component TMB substrate and stopped with TMB stop solution (KPL, Inc., Gaithersburg, MD).

Quantitative analysis of IFNγ expression

The relative expression of IFNγ in cardiac tissue was determined using total RNA extracted from the base of the heart using the TRIzol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. The RNA was treated with DNase I (Promega, Madison, WI), and reverse transcribed using random primers (Invitrogen) and SuperScript II reverse transcriptase (RT) (Invitrogen). The cDNA was amplified using primers specific for glyceraldehyde 3-phosphate dehydrogenase (gadph; 5′-CCA TCA CCA TCT TCC AGG AGC GAG-3′ and 5′-CAC AGT CTT CTG GGT GGC AGT GAT -3′) and ifnγ (5′-GCG TCA TTG AAT CAC ACC-3′ and 5′-GGA CCT GTG GGT TGT TGA CC-3′) in an Mx3005P® QPCR System (Stratagene, La Jolla, CA) and SYBR green-containing reaction buffer (Roche, Nutley, NJ). Relative expression of the gene is referred to control-infected mouse heart tissue, as described (27).

Detection of iNKT cells

DNA from the hearts of control and infected mice were utilized to perform a nested-PCR reaction using primers specific for the Vα14-Jα18i TCR. For the first reaction, non-specific TCR (5′-GTC CTC AGT CCC TGG TTG TC-3′ and 5′-CTG CCT CCG AGG TAG TGA C-3′) was amplified using 5 μl of DNA in a total volume of 50 μl with the following conditions: 94°C 2 min, (94°C 1 min, 55°C 1.5 min, 72°C 1.5 min) x 30 and 72°C 2 min. A second reaction using primers specific for the Vα14-Jα18i TCR (5′-GAC AGT CCT GGT TGA CC-3′ and 5′-AAT GCA GCC TCC CTA AG-3′) was performed using 5 μl of template from reaction 1 with the same conditions, except that 35 cycles were used.

Confocal microscopy

At sacrifice, hearts were collected and snap-frozen in O.C.T. freezing medium (Sigma) and stored at -80°C. Hearts were grossly cut through the ventricles parallel to the atria, sectioned using a Triangle Biomedical Sciences Cryostat (Durham, NC) and every sixth 10 micron section through the aortic sinus was placed on Fisherbrand Superfrost microscope slides (Fisher Scientific, Pittsburgh, PA). Sections were fixed in -20 °C acetone for 5 min, rehydrated in distilled water for 3 min, blocked with Fc Block (anti-mouse CD16/32, clone 2.4G2; BD Pharmingen, Franklin Lakes, NJ) and stained with the following monoclonal antibodies: Alexa Fluor⁴⁸⁸ anti-mouse CD3 (clone 145-2C11), Alexa Fluor⁶⁴⁷ anti-mouse CD11b (clone M1/70), PE-anti-mouse CD1d (clone 1B1; BD Pharmingen). Nuclear staining was done with DAPI (Roche). The sections were imaged using a Zeiss LSM 510 META confocal laser scanning microscope (Zeiss Microimaging, Thornwood NY) operating in the multitrack mode, as described before (28).

Analysis of phagocytosis

BMM or RAW264.7 cells were activated with different doses of rmIFNγ for 12 hour in serum- and antibiotic-free RPMI prior to the addition of GFP-expressing B. burgdorferi strain 297 at an m.o.i. of 10 for 4h. To distinguish between binding and internalization of spirochetes, the mixtures were incubated at 4 °C and 37 °C, in parallel. The incubation at 4 °C results in the binding of particles, but not their internalization (29). After the incubation, the cells were washed and analyzed by flow cytometry. For Apotome epifluorescence analysis of phagocytosis, RAW264.7 cells were activated with 50 ng ml^-1 of rmIFNγ and challenged with GFP-B. burgdorferi at 37° C, as described. The cells were then washed, cytospun and stained with the macrophage cell membrane marker F4/80 (clone BM8; rat anti-mouse F4/80, eBioscience, Inc.) followed by a secondary stain with Alexa Fluor⁵⁹⁴ goat anti-rat IgG (Invitrogen).

Statistical Analysis

The results are presented as means ± SE. Significant differences between means were calculated with the Student’s t test. P values of 0.05 or less were considered statistically significant.

iNKT cells modulate the pathogenesis of Lyme carditis

The recent identification of B. burgdorferi-derived diacylglycerol glycolipids as natural antigens to iNKT cells (7), prompted us to examine the role of iNKT cells in the pathogenesis of Lyme disease. To determine whether iNKT cells contribute to the host response to B. burgdorferi, we syringe-infected wild type and Jα18-deficient C57Bl/6 mice with 1 × 10⁵ spirochetes in the midline of the back and allowed infection to progress for 2-3 weeks, which represents the acute phase of infection (25). Histological evaluation of the joints and hearts of the infected mice revealed that Jα18-deficient animals are more susceptible to both arthritis (Fig. 1A) and carditis (Fig. 1B, C). Although there was a trend towards worse arthritis in mice that lacked Jα18, the difference was not significant (Student t test; p = 0.07; n = 15). In contrast, carditis severity in Jα18-deficient mice was increased significantly (p < 0.01).

Infection of Jα18-deficient mice results in increased spirochetal burdens in the heart

Inflammation during experimental infection with B. burgdorferi is dependent on the ability of spirochetes to evade local clearance mechanisms in cardiac and joint tissue (16). Thus, we evaluated the ability of the mice to clear B. burgdorferi in the absence of iNKT cells by assessing bacterial loads in the heart tissue of the infected animals. Compared to wild type mice, Jα18-deficient mice harbored significantly more B. burgdorferi in their heart (Student’s T test; p = 0.05), while the bacterial burden in the ear was unaffected by the loss of iNKT cells (Fig. 2A). iNKT cells have been shown to potentiate the antibody response to various pathogens (30, 31). Since the antibody response to B. burgdorferi is critical for clearance of spirochetal infection (32, 33), we assessed the B. burgdorferi-specific IgM and IgG responses 2 weeks following infection. The serum levels of antigen-specific IgM and IgG in wild type and Jα18-deficient mice were similar (Fig. 2B and 2C). Moreover, the antigen specificity between the groups of mice was indistinguishable as assessed by immunoblotting (data not shown). These results demonstrate an important role for iNKT cells in the control of B. burgdorferi at the tissue level where acute inflammation occurs and show that differences in B. burgdorferi control and pathology between the Jα18-deficient and wildtype mice are not due to distinct antibody responses.

iNKT cells and CD1d-positive macrophages migrate to the inflamed heart of B. burgdorferi-infected mice

iNKT cells have been shown to migrate to sites of inflammation in the lung, liver and spleen (17, 34-37). We therefore analyzed the presence of iNKT cells in the hearts of uninfected and B. burgdorferi-infected mice using a nested PCR approach (38, 39). While control mice did not harbor iNKT cells in the heart, the presence of the invariant TCR alpha chain was readily detected in the inflamed cardiac tissue of wild type, but not of Jα18-deficient mice (Fig. 3A). We next used confocal immunofluorescence microscopy to characterize the infiltrating cells in the infected hearts. The infected hearts contained, as expected, CD11b⁺ cells (a macrophage marker) as well as CD3-expressing lymphocytes, (Fig. 3B), and confirmed (40) a predominant macrophage infiltration during Lyme carditis. The inflamed hearts also showed increased expression of CD1d compared to uninfected controls (Fig. 3B). Because CD1d expression is a requirement for APC-mediated activation of iNKT cells, we next determined the identity of the CD1d-expressing cells. CD1d expression colocalized with CD11b expression (Fig. 3C), thereby identifying macrophages as the predominant CD1d-expressing immune cell.

iNKT cells are responsible for the production of IFNγ in the infected heart

B. burgdorferi glycolipid-pulsed dendritic cells induce IFNγ production by iNKT cells in vivo (7). In addition, IFNγ-producing CD4⁺ T cells facilitate the regression of murine Lyme carditis (41), while Stat1-deficient mice infected with B. burgdorferi develop exacerbated Lyme carditis, compared to control mice (42). We therefore addressed the contribution of iNKT cells to the pool of cardiac IFNγ in B. burgdorferi-infected mice. B. burgdorferi infection of wild type mice resulted in increased expression of the ifnγ gene in the cardiac tissue compared to uninfected controls, as determined by quantitative RT-PCR (Fig. 4A). Of particular importance, the level of ifnγ gene expression in the hearts of B. burgdorferi-infected mice was reduced in Jα18-deficient mice at the peak of disease (Student’s t test, p < .05; Fig. 4A). These results indicate that iNKT cells are critical for generating an IFNγ response in the heart, which, in turn, is required for local control of spirochete burdens.

We next tested the direct contribution of IFNγ to local inflammation and spirochete clearance in Lyme carditis by infecting IFNγ receptor alpha (IFNγRα)-deficient 129Sv/J mice. Like Jα18-deficient mice, IFNγRα-deficient mice developed increased cardiac inflammation following syringe infection with B. burgdorferi (Student t test; p < .05; Fig. 4B), supporting a role for IFNγ in the protection against Lyme carditis. The quantification of B. burgdorferi burdens by real-time PCR in the heart of the infected mice showed a significant (p < .006) increase in the levels of bacteria in the absence of IFNγ-mediated signals (Fig. 4C).

IFNγ signaling enhances macrophage responses to B. burgdorferi

The mechanism underlying the control of cardiac inflammation during murine Lyme borreliosis by IFNγ has not been elucidated. IFNγ enhances cellular responses, including macrophages, to various pathogens (43). Indeed, activation of bone marrow-derived macrophages (BMMs) with IFNγ enhances spirochete recognition, as evidenced by augmented TNF and IL-6 production elicited against live B. burgdorferi (Fig. 5A and 5B). Increased B. burgdorferi burdens in the heart of IFNγRα- and Jα18-deficient mice during the presence of a strong antibody response suggested a defect in phagocytic clearance of the spirochete by infiltrating macrophages. We therefore defined the contribution of IFNγ to macrophage-mediated clearance of B. burgdorferi by phagocytosis. First, we determined by flow cytometric analysis that BMMs readily phagocytose GFP-expressing infectious B. burgdorferi (Fig. 5C). Activation with recombinant murine IFNγ increased the already high phagocytic capacity of BMMs for B. burgdorferi (Fig. 5C). We also performed this assay using the macrophage cell line RAW264.7. Incubation of RAW264.7 cells with GFP-B. burgdorferi resulted in a modest amount of phagocytosis, compared to BMMs, confirming previous reports indicating their relatively low phagocytic capacity (44) (Fig. 5D, 5E). In separate experiments, a few RAW264.7 cells showed internalized B. burgdorferi by fluorescence microscopy (data not shown). The activation of these cells with IFNγ markedly enhanced the percentage of cells that had internalized GFP-expressing B. burgdorferi, as determined by flow cytometric analysis and flourescence microscopy (Fig. 5D, 5E). These results suggested that iNKT cells control the levels of spirochetes in the infected heart through the production of IFNγ and increased phagocytosis.

IFNγ enhances the surface expression of CD1d in macrophages

Previous reports have suggested that IFNγ can increase CD1d surface expression by macrophages (2, 45). We thus assessed the effect of IFNγ on the expression of CD1d in primary macrophages and RAW264.7 cells. We stimulated BMMs (Fig. 6A) and RAW264.7 cells (Fig. 6B) with 50 ng/ml of recombinant IFNγ for 16 h. The cells were subsequently assessed for surface CD1d expression by flow cytometry. IFNγ treatment increased the expression of CD1d in both BMMs and RAW264.7 cells (Fig. 6A, 6B). Since B. burgdorferi induces the upregulation of CD1 molecules on human monocytes (46), we also assessed the effect of live spirochetes on CD1d expression by BMMs and RAW264.7 cells. In contrast to human monocytes, the stimulation of both BMMs and RAW264.7 cells with live B. burgdorferi resulted in a minimal increase of CD1d expression (data not shown). These results suggest that during infection with B. burgdorferi, iNKT-derived IFNγ is self-renewing through its capacity to induce CD1d expression and presentation to iNKT cells.