Tick Saliva Reduces Adherence and Area of Human Neutrophils

ABSTRACT During natural infection with the agent of Lyme disease, Borrelia burgdorferi, spirochetes are delivered with vector saliva, which contains anti-inflammatory and antihemostatic activities. We show here that the saliva of ixodid ticks reduces polymorphonuclear leukocyte (PMN) adhesion via downregulation of β2-integrins and decreases the efficiency of PMN in the uptake and killing of spirochetes. Inhibition of integrin adhesion and signaling reduces anti-inflammatory functions of PMN. These effects may favor the initial survival of spirochetes in vivo.


MATERIALS AND METHODS
Cell isolation and imaging. PMN were isolated from the heparinized blood of healthy volunteers by 3% dextran sedimentation, and red blood cells were removed by hypotonic lysis (10). Informed consent was obtained from all blood donors in accordance with the guidelines of the Human Investigations Committee of Yale University. For immunofluorescent staining, PMN were plated at a concentration of 5 ϫ 10 5 cells per 12-mm round glass coverslip in Krebs Ringer phosphate buffer with 5.4 mM glucose (KRPG) containing 10% heat-inactivated pooled human serum. Adherent PMN samples were incubated with spirochetes at a 100:1 ratio in KRPG containing 10% heat-inactivated pooled normal human serum as described previously (17). At times up to 60 min, samples were fixed in 4% paraformaldehyde and blocked in phosphate-buffered saline (PBS) containing 10% goat serum and 0.01% saponin. PMN were labeled with primary antibodies directed against the PMN azurophilic granule component myeloperoxi-dase (DAKO Corp., Carpinteria, Calif.). Rabbit polyclonal anti-B. burgdorferi serum, a kind gift from Fred Kantor, Yale University School of Medicine, was used to label spirochetes (18). Antibodies were detected with appropriate fluorescein isothiocyanate-or tetramethyl rhodamine isocyanate-conjugated secondary antibodies (Tago Immunologicals, Biosource Intl., Camarillo, Calif.), and samples were imaged by using multitracking with a Zeiss LSM 510 scanning laser confocal microscope equipped with an argon-krypton laser (17).
For measurement of cell area, freshly isolated PMN in PBS containing Ca 2ϩ , Mg 2ϩ , and 0.1% bovine serum albumin (BSA) were plated at a concentration of 1 ϫ 10 5 cells per glass coverslip and incubated for 1 h at 37°C to allow adherence. Adherent cells were then incubated for an additional 1 h in a humidified chamber either alone or in dilutions of saliva (1:10, 1:20, and 1:100). Cells were fixed in 4% paraformaldehyde in PBS, mounted in Mowiol, and examined with a Zeiss Axiovert 200 M microscope (Carl Zeiss Microimaging, Inc., Thornwood, N.Y.). Digital images were taken from 100 to 200 cells under each condition at a magnification of ϫ40, and the diameter of each cell (in arbitrary units) was determined from a print of the digital image.
For videomicroscopy, 5-l portions of cells in PBS-2% BSA were sealed under glass coverslips with a mean final thickness of 5.7 m so that the cells were compressed between the slides and coverslips (13,15). PMN were examined by phase-contrast microscopy by using a ϫ40 objective on a 33°C stage. The orientation and trajectory of PMN was recorded before, during, and after their response to a chemotactic stimulus created by ruby laser microbeam destruction of an erythrocyte (wavelength, 694.3 min; duration of flash, 0.5 ms) (13). Freezeframe images were collected with a Hamamatsu C2400 microscope video camera (Hamamatsu Photonics K.K., Hamamatsu City, Japan) and a Panasonic AG6720 time-lapse video recorder (Matsushita Electric Industrial Co., Osaka, Japan). Reagents for treatment of cells were obtained from Sigma-Aldrich Fine Chemicals (St. Louis, Mo.) Spirochete culture. A low-passage clonal isolate of B. burgdorferi strain N40 was cultivated in Barbour-Stoenner-Kelly (BSK) II medium at 33°C as described previously (10). Spirochetes were opsonized for 30 min with 1 to 10% heatinactivated serum from a well-characterized Lyme disease patient or a normal volunteer in PBS; donor serum reactivities were documented by an enzymelinked immunosorbent assay and Western blotting against B. burgdorferi lysate.
Saliva preparation. Adult Ixodes scapularis ticks were allowed to feed on naïve rabbits for 5 to 7 days until they were engorged. Saliva was harvested from the ticks following pilocarpine stimulation as described previously (28) and was stored at Ϫ80°C until it was used.
Detection of killing of spirochetes. B. burgdorferi was quantified by using a modified regrowth assay which measures spirochete uptake of [ 3 H]adenine over 48 h (10). PMN in these experiments did incorporate some [ 3 H]adenine, but as they did not divide, it was a small amount. Briefly, 5 ϫ

Saliva treatment reduces attachment of B. burgdorferi by PMN.
We have recently shown that PMN are inefficient at binding spirochetes in the absence of specific antibody (17). In our examination of PMN efficiency, we noted that salivatreated PMN appeared to be less well attached to coverslips, which suggested that they might have an impaired ability to bind spirochetes. When control PMN on coverslips were incubated with opsonized spirochetes, a procedure that generally results in efficient binding, we observed well-spread cells with numerous attached spirochetes (Fig. 1A). In contrast, salivatreated PMN were significantly less well spread and bound fewer spirochetes, even when the spirochetes had been opsonized with specific antibody (Fig. 1B). A fivefold reduction in the number of PMN with attached B. burgdorferi cells was noted when saliva-treated PMN were compared to control cells (P ϭ 0.09; n ϭ 3) (Fig. 1C). Reduced binding was observed both in the presence of serum and under serum-free conditions (data not shown).
Saliva reduces the cellular area of treated PMN. When we examined coverslips with adherent PMN that had been incubated for 1 h in the presence of saliva, we observed an apparent reduction in the number of cells. While counts of cells on grid coverslips indicated that there was a modest reduction in the number of cells after saliva treatment (data not shown), there was a dramatic difference in the area of the remaining cells ( Fig. 2). We quantified the cellular areas of untreated PMN and PMN incubated with dilutions of saliva (1:10, 1:20, and 1:100) from printed digital images of 100 to 200 fixed cells per condition (n ϭ 2). As shown in Table 1, there was a dosedependent reduction in the cellular area after treatment of PMN with saliva for 1 h; the difference is statistically significant (P ϭ 0.0001). We found a similar reduction in the cellular area of PMN treated with saliva in suspension for 1 h prior to plating on glass coverslips (data not shown).
Saliva reduces expression of PMN ␤2-integrins. The reduced attachment and spreading of PMN in the presence of saliva suggested that there was altered expression of PMN ␤2-integrins, the molecules that mediate adhesion of the cells. To examine this, we incubated PMN with saliva and quantified expression of several surface antigens. As expected, PMN stimulated in suspension with 15 ng of TNF-␣ per ml for 30 min  showed significant upregulation of CD18 integrin expression (Table 1) (16). In contrast, a dramatic decrease in CD18 (␤2integrin) expression was detected in saliva-treated cells. This reduction was dose dependent, and higher concentrations of saliva resulted in more profound reductions (Fig. 3). The reduction in CD18 expression on saliva-treated PMN was apparent in the presence or absence of TNF-␣ activation and was statistically different from CD18 expression by PMN not exposed to saliva (for 1:10 dilution of saliva, P ϭ 0.0004 without TNF-␣ and P ϭ 0.08 in the presence of TNF-␣; n ϭ 4) ( Table  1). Expression of the PMN surface marker CD15 (siayl Lewis X antigen) was not appreciably changed by saliva treatment or by TNF-␣ stimulation. We examined PMN surface marker expression after treatment with reagents similar to components of tick saliva. No significant differences from untreated PMN were detected for PMN expression of the integrin molecules Cd11a, CD11b, CD11c, and CD18 or of surface protein CD15 after 1 h of incubation in the presence of the reagent apyrase (10 U/ml), prostaglandin E 2 (PGE 2 ) (20 g/ml), or calcitonin gene-related peptide (1 ϫ 10 Ϫ8 M) at concentrations reported to be effective for activity (22,23,26,27). As expected, TNF-␣ produced a statistically significant increase in expression of all integrins tested (P ϭ 0.04), regardless of treatment, but not of CD15 (n ϭ 3) (data not shown).
Saliva treatment impairs spirochete killing in suspension. Saliva-treated PMN were dramatically less adherent (Fig. 1) and thus were not suitable for direct microscopic observation of spirochete killing by vital staining. To measure killing, we quantified the recovery of spirochetes in suspension using a well-characterized [ 3 H]adenine regrowth assay (10). PMN were preincubated in suspension with saliva at a 1:10 dilution for 1 h before addition of spirochetes at a B. burgdorferi/PMN ratio of 5:1 or 2:1. In three experiments in which PMN killed unopsonized spirochetes, a modest increase in the level of viability was noted in saliva-treated PMN (Fig. 4); the reduction was significant at a B. burgdorferi/PMN ratio of 2:1 (P ϭ 0.04). While this suspension assay does not require adherence of the PMN to a surface, the increased survival of spirochetes in the presence of treated PMN is in keeping with the observed decrease in attachment of spirochetes to PMN.
Tick saliva does not affect PMN orientation or chemotaxis. In preparations in which PMN have access to only one surface and adhesive properties are important, the absence of integrins greatly reduces locomotion (13). We have found that com- pressed PMN in very thin slide preparations exhibit normal random locomotion and chemotaxis even in the presence of antibodies that disable their integrins. Normal locomotion was also noted in compressed PMN from a patient whose PMN lack ␤2-integrins (13,14). In the present study, as expected, control PMN were well attached to the glass slides and locomoted flat on the slide before and after stimulation. In contrast, saliva-treated PMN were less adherent and often rounded up (data not shown). Saliva-treated cells resembled cells treated with EDTA or anti-CD18 (␤2-integrin) antibodies, conditions that disable integrins (13,14). To determine which aspects of chemotaxis were affected by saliva, we examined saliva-treated PMN in compressed slide preparations, in which integrins were not required for a PMN response to a chemotactic target. Compressed PMN showed no effect of tick saliva at a dilution of 1:10 and responded appropriately to the chemotactic gradient created by laser irradiation of erythrocytes (Fig. 5), indicating that orientation and directed locomotion were intact.
In the presence of saliva, both in pretreatment and throughout the interaction, we saw the disaggregation of platelets described previously (24). As noted above for surface marker integrin expression as assessed by FACS, apyrase (1 to 10 U/ml) and PGE 2 (2 to 20 g/ml), reagents similar to salivary components, had no effect on random locomotion or chemotaxis in thin or thick preparations (data not shown). PMN responded normally to the chemotactic stimulus even in the presence of 200 g of PGE 2 per ml.

DISCUSSION
We found that PMN are less adherent, bind fewer spirochetes, and have a reduced cellular footprint after saliva treatment. The mechanism for reduced spirochete attachment after saliva treatment is unclear and may include the reduced surface area or the absence or masking of a specific receptor. FACS analysis of saliva-treated PMN revealed a dose-dependent decrease in expression of CD18, the common monomer of the three ␤2-integrins, evident in the presence or absence of stimulation with TNF-␣. Moreover, killing of spirochetes was reduced in the presence of saliva, in keeping with evidence that integrins mediate binding of unopsonized spirochetes to leukocytes (3,4). Although we did not distinguish between decreased appearance and increased removal of CD18 from the cell surface, the behavior of saliva-treated PMN during chemotaxis suggests that there is aggregation of integrin molecules that are swept to the rear of moving PMN. In compressed slide preparations, in which adhesion was less important for locomotion (13), tick saliva had no effect on the random locomotion of PMN or on their orientation or chemotaxis.
After a tick bite, a local cement plug or absorbent reservoir forms, which contains high concentrations of saliva that are replenished by intermittent bursts of salivation during feeding (1). Generally, attachment of ticks for 48 h is necessary for efficient transmission of B. burgdorferi to a host (25), which provides a critical window of time for saliva to exert its modulatory effects on the host immune response.
Besides their apparent contribution to the adhesion of PMN to B. burgdorferi and killing of spirochetes as demonstrated here, integrins are known to be important in adherence and locomotion and other aspects of phagocyte function; the absence of these molecules results in severe complications of infectious diseases and wound healing (8). For example, masking of CD18 blocked cytokine stimulation of H 2 O 2 production by PMN (19) and delayed apoptosis (5). The downregulation of PMN integrins by saliva compromises the abilities of the responding PMN and provides spirochetes with critical time to establish themselves in situ before they disseminate to cause systemic illness.

ACKNOWLEDGMENTS
This work was supported in part by grants from the National Institutes of Health (grants AI 43558, AR 10493, AR 48513, and AR 07107), from the Eshe Fund, from the Mathers Foundation, and initially from the Centers for Disease Control and Prevention. D.L. was a fellow of the Arthritis Foundation.
We are grateful for ticks supplied by John Anderson and Durland Fish and for the technical assistance of Rita Palmarozza and Deborah Beck.