During natural infection with the agent of Lyme disease, Borrelia burgdorferi, polymorphonuclear leukocytes (PMNL) are the first cells of the innate immune system to arrive at the site of spirochete deposition in the skin. This study examined the degree of spirochete clearance likely to occur with PMNL or mononuclear cells before the development of the secondary immune response. Without specific antibody in vitro, there was very limited uptake of spirochetes by PMNL or monocytes and no intracellular colocalization of PMNL granule products with spirochetes. Most of the killing of spirochetes by PMNL was extracellular. In contrast, mature macrophages ingest and kill spirochetes avidly with or without specific antibody. Once the spirochetes are opsonized, PMNL clear them rapidly. These findings may be relevant to the initial survival of spirochetes introduced into the host.
Lyme disease is a multisystem disorder caused by infection with the spirochete Borrelia burgdorferi sensu lato, delivered by Ixodid ticks . Spirochetes may persist in the skin and organs of infected hosts, evading immune surveillance and leading to complications of ongoing infection . We previously showed that macrophages in vitro rapidly ingest and kill B. burgdorferi in large numbers, with or without opsonization and participation of Fc receptors; ingested spirochetes arrive in macrophage lysosomes with a clearance half time of ∼20 min [2–4].
We recently examined the possibility of down-regulation of immune responsiveness during infection in murine Lyme borreliosis and showed that there is not a global impairment of macrophage function in the infected host . Indeed, macrophages in hearts of mice with Lyme carditis produce increased levels of mRNA for proinflammatory cytokines, reflecting appropriate macrophage activation. Thus, no evidence of pressure toward immune down-regulation was detected in infected animals .
The initial response of phagocytes to spirochetes in Lyme disease may determine the course of the infection. Indeed, when the site of inoculation is excised within 48 h of tick bite, no systemic infection occurs . Disseminated infection indicates that the innate immune cells have been unsuccessful at eliminating the spirochetes at the site of entry. This may be due in part to the inhibitory effects of the delivery vehicle, tick saliva, which inhibits some phagocyte functions . However, the efficiency of clearance of spirochetes by polymorphonuclear leukocytes (PMNL), the first phagocytes to arrive at the site of spirochete inoculation [1, 9], has not been examined in detail.
PMNL in vitro bind Borrelia organisms in various experimental systems [10–17]. In addition to conventional phagocytosis, 2 specialized techniques of ingestion have been described: tube phagocytosis, where a thin cellular protrusion surrounds the spirochete lengthwise , and coiling phagocytosis, which involves pseudopods of cell membrane wrapping around the target in a spiral pattern [13, 15, 18]. Phagocytosis of borrelia by PMNL was enhanced by treatment of the spirochetes with immune serum, but little or no contribution by complement factors was noted in several studies [10–12], although such complement sensitivity may vary with different strains [19, 20].
We recently quantified mechanisms employed by PMNL to eliminate spirochetes and showed that B. burgdorferi are susceptible to both oxidative and nonoxidative killing mechanisms and that PMNL lysates are as effective as intact cells in killing opsonized spirochetes, suggesting the particular relevance of PMNL granule proteins in vivo . We also observed that PMNL are substantially more efficient at killing spirochetes that have been opsonized with immune serum . In this study, we examine the clearance of B. burgdorferi by phagocytic cells before the development of specific immunity.
Materials and Methods
Cell isolation and culture. PMNL were isolated from heparinized blood from healthy volunteers by using 3% dextran sedimentation and hypotonic lysis to remove red cells, as described elsewhere . PMNL were plated in Krebs Ringer phosphate buffer with 5.4 mM glucose (KRPG)/10% heat-inactivated (HI) human serum at 5 × 105 per 12-mm round glass coverslip (immunofluorescence) or 2–5 × 106 per 22 × 22-mm coverslip (killing assays). Monocytes were isolated by use of Ficoll-Paque (Pharmacia) and plated in 0.1 mL of RPMI 1640 medium/20% HI serum at the same concentrations as for PMNL. Nonadherent cells were rinsed away gently with warm RPMI 1640 medium after 1–2 h of incubation at 37°C in 5% CO2. Monocytes were used immediately or cultured for 6–8 days to obtain macrophages.
Spirochete culture. A low-passage clonal isolate of B. burgdorferi strain N40 was cultivated in Barbour-Stoenner-Kelly II medium at 33°C, as described elsewhere . Spirochetes were opsonized for 30 min with 1%–10% HI serum in PBS from a well-characterized Lyme disease patient or a normal volunteer. Donor serum reactivities were documented by use of ELISA and Western blot against B. burgdorferi lysate .
Detection of phagocyte killing of spirochetes. Coverslips with adherent PMNL, monocytes, or macrophages were incubated with 1–2 × 107B. burgdorferi cells per coverslip for a 10:1 ratio of B. burgdorferi:phagocyte in a 0.1-mL bubble of PBS containing physiologic glucose (5.4 mM) and 2% human serum albumin (HSA) (PBS/glu/HSA) at 37°C. Cells and spirochetes were concentrated at the center of the coverslips and were completely covered by a 0.1-mL volume of PBS/glu/HSA. At 15 and 60 min of incubation, coverslips were washed gently with warmed PBS/glu/HSA and labeled for spirochete viability by using a vital stain, Live/Dead kit, according to the manufacturer's instructions (Molecular Probes). Cells were examined on a Zeiss LSM 510 scanning laser confocal microscope equipped with an argon/krypton laser . The 2 dyes contained in the kit represent an advance over the single dye, acridine orange, as the phagocytes stain very brightly and independent adjustment of signals in 2 channels allows better visualization of spirochetes. Under these conditions, live spirochetes stain green, and heat-killed organisms stain red.
Cell imaging. For immunofluorescent studies, adherent PMNL or monocyte-macrophage samples were incubated with preparations of spirochetes inKRPG/10%HI-pooledNHS for the times indicated. Samples were fixed in 4% paraformaldehyde (PFA) and blocked in PBS/10% goat serum. All incubations contained 0.01% saponin to permeabilize the cells. Primary antibodies directed against PMNL components: myeloperoxidase (MPO), CD15, elastase (Dako), tolllike receptor (TLR)-2 and TLR-4 (both from eBioscience), and the actin stain fluorescein isothiocyanate (FITC)-phalloidin (Molecular Probes) were used to label fixed cells. PMNL MPO was not well detected after MeOH fixation, so PFA fixation was used for all colocalization studies. Rabbit polyclonal anti-B. burgdorferi serum (gift of Fred Kantor, Yale University School of Medicine, New Haven, CT) was used to detect spirochetes . Antibodies were detected with appropriate FITC- or tetramethyl rhodamine isothiocyanate-conjugated secondary antibodies (Tago Immunologicals, Biosource International), and samples were imaged by confocal microscopy as described above . These studies were performed with cells from > 6 different donors.
Freeze frame images were obtained from video sequences by using leukocytes from fresh heparinized whole blood isolated after sedimenting in tubes at a 60° angle at room temperature for 1 h. Cells were concentrated in a microcentrifuge (Costar) for 30 s at 5585 g and resuspended in 1 mL of autologous heparinized plasma with spirochetes, sealed under glass coverslips, and examined on a 33°C stage of a Zeiss phase-contrast photomicroscope. Video images were collected on amicroscope video camera (model C2400; Hamamatsu Photonics K.K.) and a time lapse video recorder (Panasonic AG6720; Matsushita Electric Industrial). The orientation and trajectory of leukocytes were observed by using a 25× objective and recorded in time-lapse video microscopy (16× real time), as described elsewhere .
PMNL and monocytes infrequently recognize unopsonized spirochetes. We previously showed that spirochetes are rapidly ingested by macrophages without opsonization . However, PMNL are the first innate immune cells to arrive in the skin after tick bite, when spirochetes first enter the host [1, 9]. To reflect these initial interactions, we followed the fate of spirochetes incubated with PMNL in vitro and showed that PMNL infrequently bind unopsonized spirochetes (figure 1A and 1C).
Dramatic differences were noted in the presence of specific opsonizing antibody (figure 1B and 1D) at both 5 and 60 min of incubation. In the presence of spirochetes opsonized with specific antibody, the PMNL were extensively spread out on the coverslips, and nearly every PMNL had attached spirochetes. The percentages noted for unopsonized versus opsonized uptake were 10.6% versus 49.3% at 5 min and 38.3% vs. 51.0% at 60 min (data are the mean of 4 experiments with different donors; P = .04 for unopsonized at 5 vs. 60 min). When unopsonized, spirochetes remained elongated at both time points (100% and 92%, respectively), whereas opsonized spirochetes were 66% and 59% elongated at 5 and 60 min of incubation, respectively, suggesting ingestion. This is in contrast to macrophage uptake, where elongated spirochetes are rarely observed after 20 min of incubation . However, we did not observe at either time point colocalization of spirochetes with the azurophilic granule component, MPO, suggesting that the spirochetes were not internalized by the PMNL. In addition, colocalization was not noted when cells were double labeled for spirochetes and for the PMNL surface markers, CD11b or CD15, for another azurophilic granule component, elastase, or for actin visualized by FITC-phalloidin.
In our PMNL studies, no colocalization was noted for spirochetes with markers of the endosomal pathway, the lysosomal membrane glycoprotein (LAMP-1) or with antibodies specific for the signaling molecules, TLR-2 and TLR-4 (data not shown). In samples incubated in the absence of serum, material reactive with antibodies to granule components could be stained, and spirochetes bound avidly to the granules (data not shown). B. burgdorferi have been reported to bind granulocyte membrane fractions .
When freshly isolated monocytes were incubated with spirochetes, an initial diffidence to spirochetes at early time points was largely overcome by 60 min in the presence and absence of specific antibodies (figure 2). Monocytes are seen binding spirochetes, which colocalize with LAMP-1, although with less efficiency than mature macrophages (figure 2A and 2B). In contrast to PMNL, human monocyte-derived macrophages were capable and thorough in binding and ingesting B. burgdorferi at every time point examined, regardless of antibody pretreatment, as noted elsewhere with murine macrophages .
Spirochetes are active targets that rapidly move both rotationally and longitudinally. The experiments above suggest that in the absence of antibody, PMNL may contact but not retain spirochetes. Video images of the live interaction show spirochetes in contact with PMNL subsequently moving out of range of the cell's pseudopods (figure 3A). In the absence of antibody, PMNL entering the visual field attempt to ingest a large clump of spirochetes. Up to 13 PMNL cannot dispatch the spirochetes, which can be seen escaping from the field. In contrast, in the presence of opsonizing antibody, spirochetal targets are efficiently ingested; a single PMNL removed 5 of 6 spirochetes in 4 min (figure 3B), and we have recorded a PMNL ingesting 14 spirochetes. An apparent chemoattractive effect of the spirochetes was noted as well. Figure 4 shows a PMNL moving toward a partially consumed opsonized spirochete protruding from a monocyte. Once the spirochete has been fully engulfed by the monocyte tube, the chemoattractive effect is apparently masked, and the PMNL moves away.
Killing of spirochetes reflects binding ability. By use of vital staining to examine killing of spirochetes by confocal microscopy, we observed that phagocytic killing of spirochetes parallels the binding activity revealed in figures 1–3. When incubated in the presence of unopsonized spirochetes, PMNL and monocytes are adjacent to live spirochetes, not visualized when cell imaging is optomized, although they kill some organisms (figure 5A and 5B). In contrast, macrophages ingest large numbers of spirochetes and kill them (figure 5C).
PMNL in vitro are activated by spirochetes to produce an oxidative burst [13, 14, 24]. Indeed, stimulation of the PMNL oxidative burst by B. burgdorferi varies with the number of spirochetes used but not with the virulence of the strain [13, 14, 24], and its outer surface protein (Osp)A activates PMNL directly . Activation in the absence of specific antibody or prior to antibody production has been reported to depend on the presence of complement , perhaps mediated by the binding of spirochetes to the CR3 (Mac-1) integrin on the PMNL surface [27, 28]. But the data shown here suggest that such activation is inefficient for killing without specific antibody to hold spirochetes in close contact with PMNL.
PMNL ingest microorganisms such as staphylococci  and, in some cases, may harbor infectious organisms . Yet the absence of colocalization with PMNL markers in the present study suggests that much of the antibody-independent killing of spirochetes occurs extracellularly. This is supported by the significant role of PMNL lysate components in spirochetal killing  and is in contrast to the macrophage's voracious antibody-independent consumption of spirochetes.
In the bloodstream, the rapid spinning movements of the spirochetes may aid them in avoiding phagocytosis and thus they may survive more easily than those in tissue. Indeed, spirochetes can survive in blood products for 6 weeks at 4°C [30–32]. Activated platelets bind B. burgdorferi ; however, the skin is the main reservoir for spirochetes .
In tissue, PMNL are activated to kill by extracellular matrix proteins , and reactive oxygen intermediates are produced by adherent PMNL at the site of attachment . This suggests that PMNL may be more effective in killing B. burgdorferi in tissue. Indeed, adherence of macrophages is a determining factor in bacterial susceptibility, either enhancing survival  or, for alveolar macrophages, leading to enhanced ingestion of Candida species .
While we cannot yet explain spirochetal persistence in tissues, we previously found that the clinically observed intermittent “immune invisibility” is not achieved by immune down-regulation [5, 6]. Late spirochetal persistence in tissues may arise from alteration of spirochetal surface antigens, masking of proteins on the organism's surface , coating with host proteins, or close association with cells . However, the current results suggest that early persistence of spirochetes may be a question not of visibility but of inefficient clearance by phagocytic cells.
We thank Rita Palmarozza for excellent technical assistance and Philippe Male for expert photography.