Recruitment of Macrophages and Polymorphonuclear Leukocytes in Lyme Carditis

Cultivation of bacteria.A low-passage-number virulent clonal isolate of B. burgdorferi strain N40 was grown to logarithmic phase in Barbour-Stoenner-Kelley II (BSK) medium by incubating for 5 days at 32°C. For experiments, B. burgdorferi was enumerated using a Petroff-Hausser hemocytometer (Hausser Scientific Partnership, Horsham, PA) under dark-field microscopy.

Mouse infection and tissue harvest.Chemokine receptor knockout animals on the C3H/HeJ (C3H) and C57BL/6J (B6) genetic backgrounds were generated by backcrossing as described previously (8) and have now been backcrossed for 10 generations. Wild-type (WT) and CCR2^−/− littermate mice (5 to 10/group) were syringe-inoculated in both hind footpads with 2.5 × 10⁵B. burgdorferi strain N40 organisms in 0.05 ml of BSK II medium (8). At the harvest date, mice were euthanized, and hearts, blood, and bladder were collected. For RNA analysis, hearts were cut in half longitudinally with a sterile blade, rinsed in sterile phosphate-buffered saline, and frozen in RNase- and DNase-free Eppendorf tubes containing 1 ml RNAlater (Ambion, Austin, TX). For histology, hearts were cut in half sagittally and each half was separately fixed in 10% buffered zinc-formalin (Anatech LTD, Battle Creek, MI).

The infection status of mice was confirmed by serum reactivity against a B. burgdorferi lysate by immunoblotting, using sera at a 1:100 dilution. Specific binding was detected by using an alkaline phosphatase-conjugated goat antibody specific for mouse IgG (Vector Laboratories) and was visualized with 5-bromo-4-chloro-3-indoylphosphate and nitroblue tetrazolium substrate (Kirkegaard & Perry Laboratories, Gaithersburg, MD) as described previously (35).

Q-PCR.DNA was first extracted from the urinary bladders and hearts of individual mice using the DNeasy tissue kit (QIAGEN, Valencia, CA) according to the manufacturer's instructions and then eluted in 60 μl double-distilled H₂O. DNA from each tissue was assessed by quantitative PCR (Q-PCR) in two separate determinations, and the average data are presented. Real-time PCRs for B. burgdorferi flaB and for the eukaryotic b-actin gene in murine tissue were performed using the primers and probes we have described previously (25). flaB was amplified using the following primers and probe: forward primer, 5′-AGC TGA AGA GCT TGG AAT GC-3′; reverse primer, 5′-AAC AGC AAT TGC CTC ATC CT-3′; probe, 6-carboxyfluorescein (6FAM)-CTT GAA CCG GTG CAG CCT GAG CA-6-carboxytetramethylrhodamine (TAMRA). For b-actin, the forward primer was 5′-ATC CTG GCC TCA CTG TCC AC-3′, the reverse primer was 5′-GGG CCG GAC TCA TCG TAC G-3′, and the probe was 6FAM-TCC AGC AGA TGT GGA TCA GCA AGC ATA-TAMRA. Two microliters of isolated DNA was amplified in a 50-μl final volume containing High-Fidelity platinum buffer, 5 mM MgSO₄, 200 μM deoxynucleoside triphosphates, 0.2 μM each primer, and 1 U High-Fidelity polymerase (Life Technologies, Gaithersburg, MD). Amplification was performed using an iCycler (Bio-Rad Laboratories, Hercules, CA) for 60 cycles at an annealing temperature of 60°C. Copy numbers for mouse b-actin and B. burgdorferi flaB were calculated using iCycler software (Applied Biosystems). Values for B. burgdorferi flaB were normalized to murine β-actin levels.

Reverse transcriptase Q-PCR.RNA was extracted from hearts by using the RNAeasy kit (QIAGEN, Valencia, CA) according to the manufacturer's instructions. cDNA was synthesized using reverse transcriptase (DNASTAR, Stratagene, La Jolla, CA), and Q-PCR was performed as we have described previously (25, 35). Primers and probes for mouse cytokines were as follows. For interleukin-1 (IL-1), the forward primer was 5′-GCA CCT TCT TTT CCT TCA TCT TTG-3′, the reverse primer was 5′-CCA GCA GGT TAT CAT CAT CAT CC-3′, and the probe was 6FAM-AGC CCA TCC TCT GTG ACT CAT GGG A-TAMRA. For macrophage inflammatory protein 2 (MIP-2), the forward primer was 5′-CAT CCA GAG CTT GAG TGT GAC-3′, the reverse primer was 5′-CTT TTT GAC CGC CCT TGA GAG T-3′, and the probe was 6FAM-ACG CCC CCA GGA CCC CAC TGC GCC CAG ACA GAA GTC A-TAMRA. For monocyte chemoattractant protein 1 (MCP-1), the forward primer was 5′-GTT GGC TCA GCC AGA TGC A-3′, the reverse primer was 5′-AGC CTA CTC ATT GGG ATC ATC TTG-3′, and the probe was 6FAM-TTA ACG CCC CAC TCA CCT GCT GCT ACT-TAMRA.

Semiquantitative assessment of tissue histopathology.Mice were sacrificed at peak infection (21 days), and hearts were harvested for histopathologic examination. Hearts (5 per group) were bisected sagittally through both atria and ventricles and were immersed in 10% buffered zinc-formalin (Anatech LTD). All tissues were processed to slides by dehydration in graded ethanol, cleared in xylene, and embedded in paraffin (Blue Ribbon; Surgipath Medical Industries, Inc., Richmond, IL). Each heart was serially sectioned at a thickness of 5 μm, and the 5th, 10th, 15th, and 20th sections (80 total) were stained with hematoxylin and eosin, followed by placement of coverslips. Observers were blinded to the study conditions until after assessment of the histopathologic features.

Each histopathologic parameter was assessed and scored separately using a semiquantitative criterion-based methodology adapted from a published method (41). Heart tissues were assessed for inflammation by microscopic examination at low (×2.5)-, intermediate (×10)-, and high (×40, ×60)-power magnification and were scored for severity of inflammation (carditis, vasculitis) according to the percentage of inflammation at the heart base upon examination at low power (×2.5). Scores of 0 (none), 1 (minimal; less than 5%), 2 (mild; between 5% and 20%), 3 (moderate; between 20% and 35%), 4 (marked; between 35% and 50%), and 5 (severe; greater than 50%) were assigned for the severity of inflammation. The character of the inflammatory infiltrate (neutrophil, macrophage, multinucleated giant cell, lymphocyte) was assessed by examination of a minimum of 5 random fields at the level of the heart base at ×40 and ×60. Digital light microscopic images were recorded using a Zeiss Axioskop microscope, an AxoCam MRC camera, and AxioVision 4.4 imaging software (Carl Zeiss Micro Imaging, Inc., Thornwood, NY).

Statistics.Statistical significance was assessed by the Mann-Whitney test (two-tailed) using GraphPad Prism software. The significance level was set at a P value of <0.05.

Disseminated spirochete burden does not reflect heart infection.To assess the importance of macrophage-lineage cells in the host defense against B. burgdorferi, we infected CCR2^−/− mice with spirochetes. We quantified disseminated infection in WT and CCR2^−/− animals by Q-PCR assessment of spirochete loads in urinary bladders (Fig. 1A). Bladders of several strains of mice become infected with B. burgdorferi, can remain infected for 6 to 12 months, and are especially representative of disseminated burden, because they can be harvested intact and do not have an inflammatory infiltrate (6, 25, 44). The level of disseminated infection reflected in bladders is lower for CCR2^−/− than for WT B6 animals, although levels are similar for C3H WT and CCR2^−/−mice. The B. burgdorferi burdens in CCR2^−/− animals of both strains show that the overall efficient clearance of B. burgdorferi is maintained in these animals despite the reduction in macrophage recruitment, suggesting that additional mechanisms may be employed.

• Open in new tab
• Download powerpoint

FIG. 1.

B. burgdorferi burdens in tissues of CCR2^−/− mice. Mice were infected via footpad inoculation, and organs were harvested at 21 days. The B. burgdorferi loads in urinary bladders (A) and hearts (B) were determined by Q-PCR using flaB for spirochetes and normalizing to murine β-actin. Data are representative of one of two separate experiments, each using 6 to 10 mice/group, Bars indicate median values. Data trends are similar in replicate experiments and reach statistical significance for CCR2^−/− versus WT animals for bladders (P = 0.004) and hearts (P = 0.02) of B6 mice and for C3H mice (P = 0.03).

The importance of macrophages makes it of particular interest to examine Lyme carditis in CCR2^−/− animals. When we examined B. burgdorferi burdens in these mice, we found that levels of B. burgdorferi were higher in the hearts of CCR2^−/− B6 mice than in those of WT B6 mice (Fig. 1B) and lower in the hearts of CCR2^−/− C3H mice than in those of WT C3H mice (Fig. 1B). These data suggest that for B6 mice, in the absence of recruitment of macrophages, B. burgdorferi clearance from the heart is impaired or delayed. However, our studies are from a single harvest time point and thus cannot reflect the kinetics of disease progression (Fig. 1B). Interestingly, at 21 days of infection, WT B6 mice have a lower B. burgdorferi burden in hearts than WT C3H mice (P = 0.005), which may contribute to the overall resistance of B6 mice to the development of Lyme carditis (3). This may reflect either a difference in the kinetics of macrophage-mediated clearance or possibly a natural defect in the recruitment or function of macrophages in C3H mice. In addition, the differences in B. burgdorferi burden between bladders and hearts suggest the importance of tissue-specific control of spirochetes.

Cytokine and chemokine expression in B. burgdorferi-infected hearts.The lack of signaling through CCR2 is likely to have a significant effect on the recruitment and activation of inflammatory cells following B. burgdorferi infection. To examine this effect, we first quantified the expression of the proinflammatory cytokine IL-1 in the hearts of infected mice by reverse transcriptase Q-PCR at 21 days postinfection. WT and CCR2^−/− animals of both strains expressed IL-1 in infected hearts. Slight increases were noted in CCR2^−/− groups from both strains, but these were not statistically significant (Fig. 2A). The similarity of these levels in WT and CCR2^−/− hearts suggest that the production of IL-1 may depend largely on resident cells of the heart. Of note, somewhat lower levels of IL-1 were detected in both groups (WT and CCR2^−/−) of C3H animals than in B6 animals. The higher level of IL-1 produced by B6 mice may contribute to the more efficient resistance of the B6 strain to B. burgdorferi infection (3).

• Open in new tab
• Download powerpoint

FIG. 2.

Cytokine production in hearts. Animals were infected with B. burgdorferi and sacrificed at 21 days postinfection. Levels of cytokines and chemokines in infected hearts were assessed by Q-PCR analysis. (A) IL-1; (B) MIP-2; (C) MCP-1. Two separate experiments were performed with 5 to 11 mice/group. Data trends are similar in replicate experiments. Differences between CCR2^−/− and WT animals reach statistical significance for MIP-2 in B6 (P = 0.049) and C3H (P = 0.009) mice and for MCP-1 in B6 mice (P = 0.006). Bars indicate median values.

To assess the role of inflammatory cell recruitment to the chemokine milieu of the infected heart, we investigated the levels of chemokines in infected tissues in the absence of CCR2 expression. When we examined MIP-2 (CXC chemokine) and MCP-1 (CC chemokine) levels in infected hearts, we observed increases in the levels of both chemokines in the hearts of infected CCR2^−/− mice compared to those in WT mice (Fig. 2B and C) for both disease-resistant (B6) and disease-susceptible (C3H) strains of mice. Thus, in the absence of CCR2 signaling, there was a dramatic elevation in the production of chemokines, which may reflect the presence of a feedback mechanism for cell recruitment to infected hearts.

Differential representation of inflammatory cells in the absence of CCR2.To determine the functional role of CCR2-recruited cells in the resolution of Lyme carditis, we compared the severity of cardiac inflammation by histopathologic analysis of hearts from infected WT and CCR2^−/− mice of resistant (B6) and sensitive (C3H) strains. We noted that B6 animals had very minimal inflammatory infiltrates following B. burgdorferi infection (Fig. 3). The severity of inflammation tended to be higher for WT than for CCR2^−/− mice of each strain but was mild for the B6 mice, reaching an inflammation score of only 0.6 on a scale of 1 to 5 (Table 1). In contrast, C3H animals had dramatically higher inflammatory responses, and C3H WT mice were even more severely inflamed than C3H CCR2^−/− mice (Fig. 3).

• Open in new tab
• Download powerpoint

FIG. 3.

Inflammation in hearts of mice with Lyme borreliosis. Animals (5/group) were infected with B. burgdorferi and sacrificed at 21 days as described in the text. Histological analysis of tissue sections of hearts showed minimal inflammation for B6 mice (A and B) and extensive inflammatory infiltrates for C3H mice (C and D). At higher magnifications, macrophages are apparent as the predominant cell type in C3H WT mice (E and G), while more PMN (indicated by arrows) are present in CCR2^−/− animals (F and H). Bars, 200 μm for panels A to D, 50 μm for panels E and F, and 20 μm for panels G and H.

In addition, the nature of the inflammatory infiltrate differed by strain and genotype. All mice had a mixture of lymphocytes and macrophages within the adventitia where the carditis was observed (Fig. 3E to H). Although the severity of inflammatory infiltrates was lower for B6 than for C3H mice, macrophages tended to predominate over lymphocytes in the WT mice and macrophages and lymphocytes were equivalent in the CCR2^−/− mice of both strains. There was a significant difference in the level of PMN infiltration by genotype. PMN were essentially absent from the inflammatory response for B6 mice: 9/10 had no PMN infiltration. For the single WT B6 mouse with infiltrating PMN, they comprised less than 10% of the inflammatory infiltrate. The susceptible C3H mice overall had a markedly higher percentage of PMN in the inflammatory infiltrate: 3/4 of the WT C3H mice had infiltrates with as much as 30% PMN (Fig. 3E and G; Table 1). For the C3H CCR2^−/− mice, the percentage of PMN was even higher: 3/4 of the mice had 40% or more (as much as 60%) PMN (Fig. 3F and H). One WT C3H mouse had numerous multinucleated giant cells as a component of the inflammatory response, and two WT C3H mice had small numbers of lymphocytes in the ventricular epicardium.

Differential clearance of spirochetes in the absence of CCR2.We examined the course of B. burgdorferi infection during the resolution phase for WT and CCR2^−/− mice. B. burgdorferi burdens for all groups were dramatically lower at day 45 than at peak infection, as expected (Fig. 4). Among C3H mice, however, CCR2^−/− mice had higher B. burgdorferi burdens in bladders than did WT mice (Fig. 4A) (P = 0.0037), suggesting that levels of disseminated spirochetes remain elevated in the absence of CCR2. However, B. burgdorferi loads in bladders of the resistant B6 mice were equivalent for WT and CCR2^−/− animals, in keeping with the efficient clearance of spirochetes noted at peak infection for that strain (Fig. 1). No increase in B. burgdorferi loads was apparent in the hearts of CCR2^−/− mice of either strain at day 45 (Fig. 4B), suggesting that clearance of spirochetes and resolution of carditis proceed via alternate mechanisms in the absence of CCR2. B. burgdorferi burdens in bladders did not correlate with levels in hearts for C3H or B6 mice during resolution, as was noted previously at peak disease.

• Open in new tab
• Download powerpoint

FIG. 4.

B. burgdorferi burdens in hearts during resolving infection. Animals were infected with B. burgdorferi as described in the text and were sacrificed at 45 days. (A) Bladders; (B) hearts. Q-PCR analysis of spirochete numbers shows a higher burden of B. burgdorferi in bladders of C3H CCR2^−/− mice (P = 0.004) but equivalent clearance of B. burgdorferi from B6 bladders and from WT and CCR2^−/− hearts for both strains. Data, in semilogarithmic plots, are from one experiment using 7 to 10 mice per group. Bars indicate median values.

At peak disease, increased levels of chemokines were detected in CCR2^−/− animals of both strains (Fig. 2). At day 45, levels of IL-1, MIP-2, and MCP-1 were similar for WT and CCR2^−/− animals of both strains (data not shown). Thus, the absence of effective macrophage recruitment does not dramatically alter the clearance of B. burgdorferi or delay the resolution of carditis in those animals, suggesting that the resolution of carditis proceeds via alternate mechanisms in the absence of macrophage recruitment.