Agglutination of Borreliella burgdorferi by Transmission-Blocking OspA Monoclonal Antibodies and Monovalent Fab Fragments

ABSTRACT

Lyme disease vaccines based on recombinant Outer surface protein A (OspA) elicit protective antibodies that interfere with tick-to-host transmission of the disease-causing spirochete Borreliella burgdorferi. Another hallmark of OspA antisera and certain OspA monoclonal antibodies (MAbs) is their capacity to induce B. burgdorferi agglutination in vitro, a phenomenon first reported more than 30 years ago but never studied in molecular detail. In this report, we demonstrate that transmission-blocking OspA MAbs, individually and in combination, promote dose-dependent and epitope-specific agglutination of B. burgdorferi. Agglutination occurred within minutes and persisted for hours. Spirochetes in the core of the aggregates exhibited evidence of outer membrane (OM) stress, revealed by propidium iodide uptake. The most potent agglutinator was the mouse MAb LA-2, which targets the OspA C terminus (β-strands 18 to 20). Human MAb 319-44, which also targets the OspA C terminus (β-strand 20), and 857-2, which targets the OspA central β-sheet (strands 8 to 10), were less potent agglutinators, while MAb 221-7, which targets β-strands 10 to 11, had little to no measurable agglutinating activity, even though its affinity for OspA exceeded that of LA-2. Remarkably, monovalent Fab fragments derived from LA-2, and to a lesser degree 319-44, retained the capacity to induce B. burgdorferi aggregation and OM stress, a particularly intriguing observation considering that “LA-2-like” Fabs have been shown to experimentally entrap B. burgdorferi within infected ticks and prevent transmission during feeding to a mammalian host. It is therefore tempting to speculate that B. burgdorferi aggregation triggered by OspA-specific antibodies in vitro may in fact reflect an important biological activity in vivo.

INTRODUCTION

The spirochete Borreliella burgdorferi is the principal etiologic agent of Lyme borreliosis, the most common tickborne disease in the United States (1, 2). B. burgdorferi resides within midguts of their arthropod vector until the ticks take a blood meal. Afterward, the bacterium proliferates exponentially within the midgut before traversing the midgut epithelium and the basement membrane (3, 4). The spirochetes then migrate to the salivary glands and are eventually deposited into the skin of an impending host, where replication occurs within the dermis. Although humans are incidental hosts, B. burgdorferi infection, if untreated, can spread systemically and result in a range of clinical manifestations involving the central nervous system, joints, and/or heart (2, 5).

B. burgdorferi modulates the expression of its outer surface lipoproteins as a means of adapting to different environmental niches and immune-mediated assaults during tick-to-host transmission (6–8). Outer surface protein A (OspA), for example, is expressed at high levels by spirochetes in the tick midgut where it functions in epithelial cell attachment (3). OspA (31 kDa) is an unusual molecule in that it consists of 21 antiparallel β-strands with a single α-helix at the C terminus (Fig. 1) (9–14). The N terminus of OspA is anchored in lipid rafts in the spirochete’s outer membrane, while the C terminus projects away from the bacterial cell surface and is accessible to host-derived immune factors, including antibodies (11, 15). As the spirochete migrates out of the midgut, the bacterium modifies gene expression and surface lipoprotein profiles to adapt to changing tissue environments and, eventually, attack by the mammalian innate immune responses (6, 7, 16–21).

FIG 1.

Epitopes on OspA recognized by MAbs 221-7, 857-2, 319-44, and LA-2. Schematic of B. burgdorferi OspA (PDB 1FJ4) displayed as surface (A) and ribbon (B) images with the epitopes recognized by 221-7 (light green), 857-2 (green), 319-44 (light pink), and LA-2 (magenta) colored. In B, β-strands 1 to 21 are colored gray and the C-terminal α-helix in yellow. (C) Schematic depicting relative locations of N-terminal domain (NTD), central β-sheet, and the CTD and corresponding β-strands with strand numbers depicted below.

Despite the spirochete’s numerous immune evasion strategies, transmission of B. burgdorferi from the tick to a mammalian host (including humans) is blocked by anti-OspA antibodies (22–31). Both passive and active vaccination results in protection, as demonstrated in numerous experimental models (e.g., mouse, hamster, and nonhuman primates) (23–25, 27–31), as well as artificial feeding chambers with human blood (32, 33). In fact, recombinant OspA was the basis of the human Lyme disease vaccine that was employed in the United States from 1998 to 2002 (34–36). Serological assessment of vaccinated individuals revealed that breakthrough infections were associated with IgG titers below a certain threshold, as defined by a competitive enzyme-linked immunosorbent assay (ELISA) with the protective mouse monoclonal antibody (MAb) LA-2 (9, 24, 35). Having a surrogate measure of immunity was critical in vaccine evaluation, since Johnson and colleagues had reported previously that protection against B. burgdorferi infection in hamsters did not necessarily correlate with total anti-OspA serum IgG titers (27).

Another hallmark of OspA antibodies, as first reported by Sadziene and colleagues (37), is their capacity to induce visible (macro) agglutination of B. burgdorferi cells in culture. Antibody-mediated agglutination of B. burgdorferi was reported to not affect spirochete motility or viability (37). Others noted that mouse MAbs like LA-2 and C3.78 were particularly potent agglutinators of spirochetes in vitro (28, 30, 38, 39). We were intrigued by these observations, given that spirochete agglutination in vivo would be expected to be a major impediment to bacteria attempting to exit the midgut enroute to the salivary glands (4). Moreover, it has been suggested that B. burgdorferi aggregation is driven primarily by antibodies targeting the OspA C terminus (like LA-2), raising the possibility that epitope specificity may be a factor (28, 30, 38, 39). However, issues associated with studying B. burgdorferi agglutination have been confounded by the absence of a collection of OspA MAbs with known binding affinities and epitope specificities.

Wang and colleagues (40) recently generated a large collection of OspA human MAbs from transgenic mice. The binding affinities of the MAbs were determined, as were their complement-dependent borreliacidal activities in vitro. Reactivity with OspA truncations revealed antibodies whose epitopes were scattered within the central and C-terminal domains of OspA (Fig. 1). Ultimately, a subset of MAbs were evaluated for the ability to block transmission of B. burgdorferi from infected ticks to naive mice (40). Complete protection was achieved with MAbs 221-7, 857-2, and 319-44 (40). The passive transfer of MAb 221-7 was subsequently shown to protect Rhesus macaques from tick-mediated B. burgdorferi infection (33).

In this report, we employed a well-characterized collection of OspA human MAbs (221-7, 857-2, and 319-44), alongside LA-2 to identify determinants that promote B. burgdorferi agglutination. We found that the MAbs, individually and in combination, promoted B. burgdorferi agglutination in a dose-dependent and epitope-specific manner. The aggregates formed within minutes and persisted for hours without any apparent loss in cell viability. However, spirochetes within the core of the aggregates exhibit increased outer membrane (OM) permeability and possibly envelope stress. More striking, monovalent Fab fragments of LA-2 were nearly as effective as their respective IgG at promoting B. burgdorferi agglutination and inducing OM permeability. It is tempting to postulate that agglutination may constitute a mechanism by which OspA antibodies impede migration of B. burgdorferi out of the tick midgut and limit transmission to human hosts.

RESULTS

OspA MAbs alter free-swimming behavior of B. burgdorferi.

Existing literature suggests that OspA antisera and certain MAbs are capable of inducing B. burgdorferi agglutination in vitro (37, 41). However, this activity has not been investigated systematically or with OspA MAbs of known epitope specificity. To begin to address this question, we chose four well-characterized MAbs that recognize two immunodominant regions of OspA (Table 1). LA-2 and 319-44 recognize the OspA C-terminal domain (CTD), while 857-2 and 221-7 recognize epitopes within the OspA central β-sheet (CBS). Specifically, LA-2 contacts the loops between OspA β-strands 16 to 17, 18 to 19, and 20 to 21, as well as portions of the strands themselves (Table 1; Fig. 1; see Fig. S1 and S2 in the supplemental material) (9). 319-44 is similar to LA-2 in that it recognizes β-strand 20 and C-terminal a-helix, while 221-7 and 857-2 have nearly identical epitopes involving contact with β-strands 8 to 10 (E. Haque, N. Mantis and D. Weis, manuscript submitted; M. Rudolph and N. Mantis, manuscript in preparation) (33). All four MAbs have been shown to protect against tick-mediated transmission of B. burgdorferi in mouse and feeding chamber models (40). In addition, 221-7 was recently shown to passively protect Rhesus macaques from experimental tick-mediated B. burgdorferi infection (33). The four OspA MAbs have borreliacidal activities in the presence of 20% human complement but no killing activity without complement (E. Haque, N. Mantis and D. Weis, manuscript submitted) (40). Finally, as determined by flow cytometry, each of the four MAbs recognize live B. burgdorferi strain B31, confirming that their respective epitopes on OspA are accessible on the surface of live spirochetes (see Fig. S3 in the supplemental material).

TABLE 1.

OspA MAbs used in this study

MAb	Epitopea	KD (nM)b
221-7	β-Strands 10–11	0.66
857-2	β-Strands 8–10	0.65
319-44	β-Strand 20	83
LA-2	β-Strands 17–19	64

^aManuscript submitted.

^bAs reported in reference 40. K_D, equilibrium dissociation constant.

We initiated studies with LA-2 IgG, as it is the most well-characterized MAb in our collection. By dark field microscopy, exposure of B. burgdorferi B31 to LA-2 (20 μg/mL) resulted in immediate changes in cell swimming behavior. Control cells were observed swimming freely in various directions and exhibiting typical spirochete behavior (e.g., swim in one direction and then reverse; form loops while swimming). In contrast, LA-2 IgG-treated cells frequently formed “lariats” in which one pole of a spirochete would loop back onto itself, even as the opposite pole retained normal rotation (see Fig. S4 in the supplemental material). While lariats were observed occasionally in the control cultures, those events were transient. In the case of LA-2, the lariats persisted for several minutes and then served as nuclei for agglutination with other bacteria (see Videos S1 and S2 in the supplemental material). With time, the LA-2 IgG-treated cells formed large aggregates with dense cores in which individual cells were not readily differentiated from each other. At the periphery of the aggregates, the spirochetes appeared long and straight, whereby the length of the spirochete confined to the center of the aggregate appeared blunted. A large fraction of the spirochetes emanating from the aggregate exhibited normal flat-wave morphology, accompanied by normal motile behaviors (e.g., wriggling, and writhing motility along the length of the spirochete body visibly protruding from the aggregate center (see Video S3 in the supplemental material)). Even 24 h later, spirochete motility was observed within the bacterial aggregates, while free-swimming spirochetes not integrated into the aggregates remained motile (see Video S4 in the supplemental material). At 24 h post-LA-2 exposure, the aggregates had amassed to a point where they frequently occluded the entire field of view. Cells treated with 319-44 IgG were similar to those treated with LA-2 IgG, although there were generally fewer aggregates when visualized across multiple fields of view. In the case of B. burgdorferi treated with 857-2 IgG, lariat formation was rarely observed, and the spirochetes appeared to associate sporadically.

Flow cytometric analysis of B. burgdorferi following OspA MAb treatment.

The observation that LA-2 IgG induced B. burgdorferi agglutination prompted us to quantitate antibody-induced cell aggregates by flow cytometry (Fig. 2). Mid-log-phase cultures of B. burgdorferi B31 (5 × 10⁶ cells/mL) grown in Barbour-Stoenner-Kelly II (BSK-II) medium (without gelatin supplement) at 23°C were treated with various concentrations of LA-2 (1, 20, and 100 μg/mL) or an equivalent amount of an isotype control (PB10) and incubated for 2 h at 37°C. The cells were then subjected to flow cytometric analysis by forward scatter (FSC) and side scatter (SSC). Spirochetes were gated to remove small debris particles, and a total of 50,000 events were collected. In preliminary studies, we defined agglutination as the events localized in the upper right quadrant.

FIG 2.

OspA MAbs promote dose-dependent agglutination of B. burgdorferi. (A) Mid-log-phase cultures of B. burgdorferi B31 were incubated with 1, 20, or 100 μg/mL of OspA MAbs or isotype control MAbs for 2 h at 37°C. Fifty-thousand events were collected on a FACSCalibur instrument (BD Biosciences, San Jose, CA) by forward scatter (FSC) and side scatter (SSC) to quantify aggregates based on size and granularity. Gating was set on untreated B. burgdorferi and defined by FSC- and SSC-positive cells (Q1 + Q2). Flow plots are representative of three biological samples run during three separate events, with three technical replicates performed three independent times. (B) Bar graph quantification of the percentage of aggregated events by each MAb at the corresponding concentration of MAb used. Data represent the mean ± SD of three biological replicates. Significance was determined at each MAb concentration by a two-way analysis of variance (ANOVA) with Tukey’s multiple-comparison test. ****, P < 0.0001; **, P < 0.01; *, P < 0.05.

For control B. burgdorferi B31 cells, the total number of events in the upper right quadrant was <1%, confirming that the growth conditions alone did not induce bacterial autoagglutination (42). No agglutination was evident upon addition of an IgG1 isotype control (PB10), as the distribution of cells in the four quadrants was indistinguishable from control cells (Fig. 2). In contrast, the addition of LA-2 IgG1 resulted in a dose-dependent increase in agglutination, as evidenced by an increased number of events in the upper right quadrant (Fig. 2). At the highest dose of LA-2 tested, ~20% of cells were agglutinated, while the percent agglutination in cultures treated with 20 or 1 μg/mL was 17% and 6.7%, respectively. MAb 319-44 also induced dose-dependent agglutination of B. burgdorferi, albeit to a slightly lesser extent than LA-2 (e.g., 13%, versus 20% at 100 μg/mL). 857-2 IgG induced B. burgdorferi agglutination to a level similar to 319-44 IgG, whereas 221-7 was notably less effective. Even large amounts of 221-7 IgG (100 μg/mL) induced only ~5% cell agglutination (Table 2). These results suggest that OspA epitope specificity, rather than binding affinity, is a determinant of spirochete agglutination.

TABLE 2.

Agglutination of B. burgdorferi by OspA MAbs

MAb (μg/mL)a	% agglutination (± SD)b of:
	PBS	PB10	221-7	857-2	319-44	LA-2
1	0.79 (0.20)	1.1 (0.27)	2.4 (0.7)	3.13 (1.03)	3.36 (0.6)	6.67 (1)
20	1.3 (0.4)	1.35 (0.23)	3.1 (0.65)	6.5 (1.3)	10.8 (1.6)	19 (3.27)
100	1.41 (1.17)	2.6 (0.64)	5 (0.96)	12.6 (0.6)	12.5 (0.6)	20.3 (3.51)

^aMid-log-phase B31 cultures were incubated with the indicated amounts of OspA MAbs for 2 h at 37°C. A total 50,000 events per sample were acquired and analyzed on a BD FACSCalibur instrument by forward and side scatter (FSC, and SSC, respectively) to quantify aggregate size and granularity after-MAb exposure.

^bResults represent data from three separate biological experiments acquired at three separate times.

We performed additional controls to investigate the specificity of antibody-mediated spirochete agglutination. For example, we confirmed that agglutination was dependent on surface expression of OspA, as an ospA mutant of B. burgdorferi (B313), did not aggregate following treatment with 20 or 100 μg/mL LA-2 (see Fig. S5 in the supplemental material). Moreover, heat-killed B. burgdorferi cells (56°C for 30 min) remained monodispersed following treatment with LA-2 (20 or 100 μg/mL), as determined by both dark-field microscopy (data not shown) and flow cytometry (see Fig. S6 in the supplemental material), although these studies should be interpreted with caution considering that heat treatment itself may induce spirochete membrane damage. Finally, we found that B. burgdorferi cells incubated at 4°C prior to LA-2 treatment did not measurably agglutinate (<1%), whereas parallel cultures at 37°C had >20% agglutination (see Fig. S7 in the supplemental material). This observation was confirmed by dark-field microscopy (data not shown). As OspA is known to associate with cholesterol rafts in the bacterial outer membrane (43, 44), this experiment would suggest that OspA OM mobility may be involved in LA-2-OspA interactions.

Impact of OspA MAbs on B. burgdorferi OM integrity.

While none of the OspA MAbs used in this study (LA-2, 319-44, 857-2, and 221-7) have borreliacidal activity in the absence of a complement, we were interested in the possibility that they impact OM integrity within bacterial aggregates. To examine whether OspA MAb treatment impacts B. burgdorferi OM integrity, mid-log-phase cultures of B. burgdorferi B31 (5 × 10⁶ cells/mL) grown in BSK at 23°C were treated with PB10 (isotype control) or LA-2, followed by incubation with propidium iodide (PI) (1.5 μM) for 10 min prior to analysis by flow cytometry (45). Treatment of B. burgdorferi B31 with PB10 resulted in <1% PI⁺ events (Fig. 3A and B). In contrast, treatment with LA-2 (20 μg/mL) resulted in ~15% PI⁺ events. PI⁺ cells localized exclusively to the upper right quadrant in the flow cytometry plots, indicating that changes in OM integrity are restricted to cell aggregates. The same held true for cells treated with 319-44, 857-2, and 221-7, even if the total number of aggregates differed for each MAb. In all cases, PI⁺ cells were evident only within the upper right quadrant corresponding to agglutinated cells (Fig. 3B). There were no PI⁺ cells when B. burgdorferi B31 was incubated at 4°C prior to LA-2 treatment, indicating that cell mobility and/or OM fluidity is required for PI staining to occur (Fig. S6).

FIG 3.

OspA MAb-mediated agglutination of B. burgdorferi alters outer membrane permeability. (A) Mid-log-phase cultures of B. burgdorferi B31 were incubated with 20 μg/mL of OspA MAbs or isotype control antibodies for 2 h at 37°C. Fifty-thousand events were collected on a FACSCalibur instrument (BD Biosciences, San Jose, CA) by forward scatter (FSC) and side scatter (SSC) to quantify aggregates based on size and granularity. Gating was set on untreated B. burgdorferi and defined by FSC- and SSC-positive cells (Q1 + Q2). Flow plots are representative of three biological replicates, with three technical replicates performed three independent times. (B) Bar graph quantification of the percentage of aggregated events by each MAb at the corresponding concentration of MAb used. Data represent the mean ± SD of three biological samples run during three separate events. Significance was determined at a fixed concentration comparing each treatment to the isotype by a Brown-Forsythe and Welch ANOVA with Dunnett’s T3 multiple-comparison test. ***, P < 0.001; **, P < 0.01; *, P < 0.05. PI was added for a final concentration of 1.5 μM prior to analysis by flow cytometry and was detected in the FL3 channel. The numerical value in the upper right quadrant (Q2) is indicative of the % of aggregated cells per treatment. The red population of cells depicted in the dot plot demonstrates the localization of PI⁺ events.

To further investigate this observation, we added a membrane permeable dye, SYTO9, to the cell mixture, as per the manufacturer’s instructions. Epifluorescent widefield microscopy was used to determine whether single cells outside the aggregates were indeed devoid of PI stain and to visualize if cells embedded within the aggregate appeared to be altered morphologically. We observed PI⁺ staining within the core of large LA-2-induced spirochete aggregates, while single B. burgdorferi cells not integrated within the mass were not detectably PI⁺ (see Fig. S8 in the supplemental material). We noted alterations in B. burgdorferi morphologies within the aggregates, including condensed or spheroid-shaped cells reminiscent of “round bodies” that form during starvation in vitro (46–49) and under certain conditions in vivo (50) or even spheroplasts that occur when B. burgdorferi is exposed to the OspB antibody CB2 (51). We conclude that LA-2 IgG-mediated aggregation has consequences for B. burgdorferi OM permeability.

Combinations of MAbs enhance B. burgdorferi agglutination and PI staining.

Antibody-mediated agglutination is generally thought to be the result of interbacterial cross-linking or lattice formation (52–54). If that were the case here, we reasoned that combining MAbs that target different epitopes on OspA would enhance overall B. burgdorferi agglutination compared with any individual MAb alone. To test this possibility, we combined LA-2 (10 μg/mL) and equimolar amounts of PB10 (isotype control) or one of the two central β-sheet MAbs, namely, 857-2 or 221-7. B. burgdorferi cells were then treated with individual MAbs (20 μg/mL) or combinations of MAbs (20 μg/mL total) and evaluated by flow cytometry. Cells were also stained with PI to assess changes in OM permeability. We found that B. burgdorferi treated with LA-2 and the isotype control MAb PB10 resulted in <15% aggregated events. The combination of LA-2 and 857-2, in contrast, resulted in ~28% agglutination, compared with LA-2 alone (21% agglutination) (Fig. 4). The same was observed when LA-2 was combined with 221-7 (Fig. 4). Enhanced agglutination was accompanied by elevated PI⁺ staining (Fig. 4), which was restricted to cell aggregates localized in quadrants Q1 and Q2.

FIG 4.

Combinations of OspA-specific MAbs enhance B. burgdorferi agglutination. Combinations of OspA MAbs (10 μg/mL each for a total of 20 μg/mL) were tested for their ability agglutinate live B. burgdorferi. PI (1.5 μM) was added to cells prior to analysis by flow cytometry and detected in the FL3 channel. In the two-parameter density plots, the numerical value in upper right quadrant (Q2) is indicative of the % of aggregated cells per treatment, while PI⁺ events are depicted in red on the dot plots and localized to Q2. Data are representative of three biological replicates.

Impact of OspA monovalent Fabs on B. burgdorferi swimming behavior and agglutination.

Up to this point, our studies have involved only OspA IgGs. However, certain OspA Fabs have been shown to be sufficient to block experimental tick-mediated transmission of B. burgdorferi in a mouse model (41). Therefore, we investigated what effect monovalent Fab fragments derived from LA-2, 319-44, and 857-2 might have on B. burgdorferi agglutination and/or swimming behavior. Mid-log-phase cultures of B. burgdorferi B31 (5 × 10⁶ cells/mL) were treated with LA-2 Fabs at low or high dose (20 or 100 μg/mL) and examined by flow cytometry. We observed that ~7% of the cells were agglutinated following low-dose LA-2 Fab treatment (20 μg/mL) and ~13% of the cells with high-dose LA-2 Fab treatment (100 μg/mL) (Table 3; Fig. 5A and B). 319-44 Fab fragments also induced B. burgdorferi agglutination, whereas 857-2 did not (Table 3; Fig. 5A and B). Labeling B. burgdorferi cells with PI revealed that agglutinated cells were PI positive, which is indicative of increased OM permeability (Fig. 6A). Combinations of LA-2 or 319-44 Fabs with 857-2 Fabs did not enhance the overall degrees of agglutination (Fig. 6B) or PI positivity (Fig. 6B). Nonetheless, these results demonstrate that OspA-specific monovalent Fab fragments can promote B. burgdorferi agglutination, albeit to a lesser degree than bivalent IgG.

TABLE 3.

Agglutination of B. burgdorferi by OspA monovalent Fabs

Faba (μg/mL)	% agglutination (mean Q1 + Q2 ± SD)b
	PBS	PB10	857-2	319-44	LA-2
20c	1.4 (0.25)	1.3 (0.4)	1.1 (0.32)	5.6 (2.41)	6.9 (2.20)
100	1.33 (1.03)	1.65 (0.47)	1.21 (0.19)	12.3 (3.21)	13.1 (1.01)

^aB31 cultures were exposed to the indicated Fab equimolar concentrations for 2 h at 37°C. A total of 50,000 events per sample were acquired any analyzed by FSC and SSC.

^bResults are from three biological repeats.

^cData at 20 μg/mL are reported for three biological replicates.

FIG 5.

OspA-specific monovalent Fabs promote agglutination of B. burgdorferi. Mid-log-phase cultures of B. burgdorferi B31 were incubated with 20 μg/mL (A) or 100 μg/mL (B) of OspA Fab fragments for 2 h at 37°C. Flow plots are representative of three biological replicates, with three technical replicates each. (B) Represents two biological replicates with three technical replicates per experiment. The numerical value in upper right quadrant (Q2) is indicative of the % aggregated cells.

FIG 6.

OspA-specific monovalent FAbs promote bacterial agglutination and alter B. burgdorferi outer membrane permeability. (A) OspA monovalent Fabs (20 μg/mL) were tested for their ability to induce OM permeability. (B) Fabs LA-2 or 319-44 (10 μg/mL each) were tested for their ability to enhance agglutination and increase OM permeability of B. burgdorferi when combined with Fab 857-2 (10 μg/mL). In the two-parameter density plots, the numerical value in upper right quadrant (Q2) is indicative of the % of aggregated cells per treatment, while PI⁺ events are depicted in red on the dot plots and localized to Q2, as described in the legend to Fig. 3.

We used dark-field microscopy to directly visualize the effects of monovalent LA-2 Fabs interacting with B. burgdorferi. We observed that treatment with C-terminal domain Fabs 319-44 and LA-2 resulted in agglutination, although to a reduced degree, as reflected in the flow cytometry data. Additionally, the aggregate cores appeared less dense than cells treated with the corresponding IgG (see Fig. S9C and D in the supplemental material). In contrast, Fab 857-2, which targets the central β-sheet, failed to promote spirochete aggregation (Fig. S9A and B). Collectively, these data reveal that OspA antibodies can induce B. burgdorferi agglutination by two distinct (nonexclusive) mechanisms, as follows: the first involving bivalent antibody interactions and possibly intercellular cross-linking and the second involving a type of autoagglutination triggered by engagement with the OspA C-terminal domain by monovalent Fabs derived from antibodies like LA-2.

DISCUSSION

The mechanism(s) by which OspA antibodies block the transmission of B. burgdorferi from the lumen of the tick midgut during the course of a blood meal to a mammalian host remains poorly understood (33–35). In this report, we demonstrated that OspA mouse and human MAbs induce dose-dependent and epitope-specific agglutination of B. burgdorferi, in vitro. While this phenomenon was reported in the literature dating back more than 30 years (55), we are the first to examine the interaction quantitatively using a well-characterized panel of OspA MAbs whose epitopes have been defined at high resolution (E. Haruque, N. Mantis and D. Weis, manuscript submitted) (33, 40). Bacterial agglutination occurred within minutes following antibody treatment and resulted in the formation of large multicellular aggregates that persisted in culture for >24 h. Although spirochetes remained visibly motile within antibody-induced aggregates, the cells at the core were likely experiencing a form OM stress, as they stained positive with PI, a known indicator of B. burgdorferi membrane permeability (43). Combinations of MAbs, namely, LA-2, which targets OspA’s C-terminal domain, plus 857-2 or 221-7, which each target the OspA central β-sheet, resulted in enhanced agglutination beyond that observed with any one given MAb at the same concentration. Based on these findings, we propose that agglutination represents a mechanism by which anti-OspA antibodies entrap B. burgdorferi within the tick midgut and impede transmission of the bacterium into a mammalian host.

Within our panel of MAbs, LA-2 was the most potent agglutinator of B. burgdorferi. LA-2 recognizes the loops formed between the OspA terminal β-strands 16 to 21 (PDB 1FJ1) (9). 319-44 IgG was significantly less potent at promoting B. burgdorferi agglutination than LA-2, even though the two MAbs have similar apparent binding affinities for OspA (60 to 80 nM) (40). 319-44 differs from LA-2 in that its epitope is restricted to β-strand 20. In contrast, MAb 3-24, whose epitope encompasses OspA β-strands 16 to 20, was as potent as LA-2 in stimulating bacterial aggregation (A. Frye, unpublished results). Taken together, these results indicate that B. burgdorferi agglutinating activity by a given MAb is determined at the level of epitope specificity, rather than binding affinity or surface accessibility, even within the C-terminal domain. Moreover, maximal agglutinating activity appears to be associated with contact of β-strands 16 to 19 and their associated interstrand loops.

That said, agglutination of B. burgdorferi was not necessarily restricted to MAbs targeting the OspA C-terminal domain as some have suggested (30, 39), considering that 857-2, which recognizes the central β-sheet (β-strands 8 to 10) was as effective as 319-44 in aggregating spirochetes. In this respect, it is curious that MAb 221-7 was so ineffective at promoting of B. burgdorferi agglutination considering its epitope is nearly identical to 857-2, at least at the level of resolution afforded by hydrogen-deuterium exchange (E. Haruque et al., manuscript submitted). In a separate study, we recently solved the X-ray crystal structure of 857-2 Fab bound to OspA, which will allow the identification of how it differs from 221-7 at the structural level (M. Rudolph et al., unpublished results).

An intriguing observation from our study is that monovalent LA-2 Fabs (as well as 319-44 Fabs) were capable of inducing B. burgdorferi agglutination, albeit to a lesser extent than the parental IgG. Agglutination was not due to contaminating IgG or F(ab’)₂, as the same preparations were subjected to Protein A-affinity chromatography and employed in X-ray crystallography studies with no evidence of contaminating factors. The Fab-mediated agglutination was largely indistinguishable from IgG agglutination. For example, by dark-field microscopy, LA-2 IgG (20 μg/mL) induced B. burgdorferi lariat formation within a matter of minutes. Thereafter, the bacteria begin to form aggregates that persisted for >24 h. Spirochetes within the aggregates appeared to remain motile. By flow cytometry, LA-2 IgG (20 μg/mL) induced ~20% agglutination and was accompanied by PI⁺ staining that was restricted to aggregates. By dark-field microscopy, monovalent LA-2 Fab fragments also induced lariat formation within 10 min. However, the sheer number of aggregates formed by Fabs was reduced compared with that by the IgG, which was also reflected in the flow cytometry studies (7% at 20 μg/mL; 13% at 100 μg/mL). Moreover, the aggregates themselves induced by LA-2 Fabs were less dense than those induced by LA-2 IgG.

These observations challenge the notion that OspA-mediated agglutination is the sole the result of antibodies forming molecular “bridges” between spirochetes. Rather, IgG and monovalent Fab fragments of LA-2 and LA-2-like antibodies would appear to trigger spirochete autoagglutination. Autoagglutination of borrelia is not uncommon. It was noted decades ago that B. burgdorferi tends to grow in large aggregates upon initial isolation from tick midguts and subsequent early passages in culture (56). Others subsequently observed the formation of aggregates with established B. burgdorferi strains under certain culture conditions (e.g., high cell density, late-log/stationary growth phase, low pH, and high temperature) (42). It is also tempting to speculate that spirochete agglutination may in fact be linked to the nature of OspA itself. For example, OspA shares structural similarities to several bacterial self-associating autotransporters such as the Escherichia coli antigen 43, which is known to function as molecular “Velcro” in promoting cell-cell adhesion (57). Either way, the formation of spirochete aggregates within the tick midgut, would be expected to impede bacterial egress across the epithelial basement membrane.

Another interesting parallel worth noting relates to OspB and the fact that LA-2 like MAbs, such as CB2, have complement-independent borreliacidal activity (43, 58). OspA and OspB, which are coexpressed at the genetic level, share ~50% amino acid identity and a high degree of structural similarity (59). Both OspA and OspB associate with lipid raft-like microdomains in the bacterial OM (44). LaRocca and colleagues demonstrated that CB2 targeting an epitope on the C terminus of OspB promotes changes to B. burgdorferi OM permeability that ultimately results in osmotic lysis (58). That same interaction was preceded by lipid raft reorganization and bacterial agglutination (T. LaRocca, personal communication) (43). With this in mind, we postulate that LA-2 evokes a similar (albeit milder) effect on the B. burgdorferi OM, possibly accounting for self-association/agglutination.

Spirochete agglutination may explain, at least in part, key observations by de Silva and colleagues regarding how OspA antisera and the mouse MAb C3.78 inhibits tick-to-mouse transmission of B. burgdorferi. C3.78 is an LA-2-like mouse MAb that recognizes an epitope on the C terminus of OspA (39). de Silva and colleagues (60) demonstrated that passive transfer of low doses of C3.78 IgG “equivalents” of OspA antisera (as defined in a competitive ELISA) was sufficient to protect mice from B. burgdorferi infection without eradicating the spirochetes from the feeding tick. Inspection of those ticks revealed that low-dose OspA antisera restricted bacterial migration to the salivary glands, possibly due to entrapment in the midgut. In a follow-up study, Gipson and de Silva (41) demonstrated that a host complement was not required for the transmission blocking activity of C3.78 IgG. Furthermore, C3.78 monovalent Fab fragments were as effective as C3.78 IgG in restricting spirochete egress from the feeding tick. We postulate that the transmission blocking activity of C3.78 in the absence of bacterial killing is due, at least in part, to antibody-mediated aggregation of B. burgdorferi within the tick midgut.

While our focus in this report has been on agglutination of B. burgdorferi by LA-2 and others MAbs, there are numerous other possible mechanisms by which anti-OspA antibodies inhibit spirochete egress from its vector. For example, 221-7 does not promote B. burgdorferi agglutination and yet it blocks tick transmission of B. burgdorferi to mice and nonhuman primates (33, 40). It is possible that anti-OspA antibodies interfere with as of yet unidentified OspA receptors that are necessary for spirochete migration from the midgut to hemocoel or beyond (8). Ultimately, identifying molecular mechanisms by which antibody block transmission of tick-borne diseases like Lyme borreliosis will contribute to efforts to mitigate disease though vaccines and passive antibody therapies (34, 36).

MATERIALS AND METHODS

Chemicals and biological reagents.

Chemicals and reagents were obtained from ThermoFisher, Inc. (Waltham, MA), unless noted otherwise. Phosphate-buffered saline (PBS) was prepared by the Wadsworth Center Cell and Tissue Culture Core Facility.

Recombinant OspA, human monoclonal antibodies (MAbs), and Fab fragments.

B. burgdorferi OspA (residues 18 to 273) was expressed in E. coli strain BL21(DE3) with an N-terminal decahistidine and SUMO tag and purified as described (33). OspA MAbs used in this study are shown in Table 1 and were prepared as described by Wang et al. (40). We used a chimeric version of LA-2 in which the mouse V_H and VL elements were grafted on to human IgG1 Fc and kappa chain backbones, respectively, and expressed in 293 cells (40). Fab fragments were generated using papain digestion followed by affinity depletion of the Fc fragment by protein A fast protein liquid chromatography (FPLC) chromatography (33).

B. burgdorferi strains and growth conditions.

B. burgdorferi strains used in these studies are described in Table 4. Frozen aliquots of B. burgdorferi were inoculated into modified BSK-II lacking gelatin, supplemented with 6% rabbit serum (Pel-Freeze Biologicals, Rogers, AR) (61). BSK-II medium was prepared by the Wadsworth Center Tissue and Media Core Facility and filter sterilized (0.2 μm) prior to use. Cultures were maintained at 23°C with 2% CO₂ under microaerophilic conditions and passaged no more than three times by dilution (1:10,000) into fresh BSK-II medium. For flow cytometry and microscopy studies, B. burgdorferi cells were collected in mid-logarithmic phase (~1 × 10⁷ to 3 × 10⁷ spirochetes per mL). Spirochetes were heat killed by incubation at 56°C for 30 min, as described (26).

TABLE 4.

B. burgdorferi strains used in this study

Strain	Description	Reference	Source
B31	B. burgdorferi strain ATCC 35210	62	ATCC
B313	B. burgdorferi strain B31 lacking ΔospAB	63	Yi-Pin Lin

OspA ELISA.

96-well microtiter plates (ThermoFisher) were coated overnight with recombinant OspA (0.1 μg/well in PBS) at 4°C. MAbs were diluted as needed in PBS. Plates were washed four times with PBS containing 0.1% Tween 20 (PBS-T) and blocked at room temperature for 2 h in PBS-T containing 2% goat serum. After plates were blocked, primary antibodies were diluted in block and incubated for 1 h at room temperature. Plates were washed with PBS-T and incubated with horseradish peroxidase (HRP)-labeled goat anti-human IgG (0.5 μg/mL; SouthernBiotech, Birmingham, AL) and washed in and developed with SureBlue TMB 1-component microwell peroxidase substrate (100 μL/well; SeraCare, Milford, MA). Plates were analyzed with a SpectraMax iD3 spectrophotometer equipped with Softmax Pro 9 software (Molecular Devices, Sunnyvale, CA).

B. burgdorferi whole-cell ELISA.

To determine MAb reactivity to OspA on the spirochete surface, Immulon 96-well microtiter plates were coated with poly-l-lysine (10 μg/mL PBS) overnight at 4°C. B. burgdorferi grown at 23°C was washed twice in PBS and placed into each well of the microtiter plate. The plates were centrifuged four times at 500 × g for 10 min (rotating 180° for the subsequent spins). Subsequent steps were repeated as described above for ELISAs.

Spirochete surface staining by flow cytometry.

Viable B. burgdorferi (5 × 10⁶ cells) were incubated with OspA MAbs (20 μg/mL) in BSK-II medium for 1 h at 37°C. The spirochetes were then washed twice in PBS and incubated with goat anti-human IgG-647-conjugated secondary antibody (ThermoFisher; 1:1,000) for 1 h at room temperature. Cells were washed twice, and the resulting fluorescence intensity of spirochetes was measured by flow cytometry using a FACSCalibur instrument (BD Biosciences, San Jose, CA). Unstained B. burgdorferi strain B31 was applied to ensure accurate gating. Additionally, the spirochetes incubated with only the secondary antibody were a control experiment to ascertain the specificity of the primary antibody to bind to the cognate antigen. An irrelevant MAb (PB10) was included as a control. Bar graphs depict the geometric mean fluorescence of 50,000 positive events ± standard deviation derived from 3 independent experiments.

Measurement of B. burgdorferi agglutination by flow cytometry.

Mid-log-phase cultures of B. burgdorferi were incubated with OspA MAbs (1, 20, or 100 μg/mL) or an isotype control (PB10) for 2 h at 37°C. Experiments evaluating outer membrane fluidity were performed the same way except at 4°C. Prior to analysis, 250 μL of phenol-free BSK-II medium was added to each sample to remove the unbound antibody. Samples were not centrifuged prior to analysis to prevent artificial aggregation of spirochetes. Live samples were analyzed on a BD FACSCalibur instrument in which 50,000 events per sample were acquired by forward scatter (FSC) and side scatter (SSC) to quantify aggregates based on size and granularity, using CellQuest Pro (BD Biosciences). Gating and thresholding were set on untreated B. burgdorferi and defined by FSC- and SSC-positive cells (Q1 + Q2). All samples were run at a low fluid speed to prevent artificial aggregation which can occur with samples run on a high fluid speed. For experiments evaluating outer membrane permeability of B. burgdorferi, propidium iodide (PI) was added (to a final concentration of 1.5 μM) to each sample at 10 min prior to analysis by flow cytometry. Detection of PI-positive events was performed in the FL3 channel. Samples were analyzed using GraphPad Prism 8. Shown are the geometric mean number of aggregates from Q1 + Q2 ± standard deviation of three independent experiments performed during three separate events.

Evaluation of B. burgdorferi morphology and motility by dark-field microscopy.

Dark-field microscopy of live B. burgdorferi was performed on a Trinocular DF microscope (AmScope) equipped with a microscopy camera with reduction lens (AmScope SKU: MU1603) using a 40× dry darkfield condenser (AmScope; DK-DRY200). Live B. burgdorferi was routinely inspected for culture viability and motility during in vitro culture maintenance prior to the initiation of any experiments. Spirochetes were enumerated by darkfield microscopy using a Petroff-Hausser counting chamber (Hausser Scientific, Horsham, PA). Image acquisition at post-antibody exposure at the indicated times was performed in parallel with flow cytometry, where 15 μL was aspirated from center of media suspension and spotted onto uncoated glass slides and approximately five fields of view were inspected on five slides per treatment.

Dark-field video microscopy.

B. burgdorferi B31 cells were imaged using an AmScope inverted DF scope with attached video camera for DF image and video acquisition. Videos were acquired with a dry condenser at a 40× objective. Mid-log cultures were treated with the indicated MAb concentrations and then were spotted (15 μL) onto glass microscope slides and imaged immediately at room temperature. Motility was determined by visual analysis. The video microscopy experiments were conducted over the course of three independent sessions and run in parallel with the flow cytometry. A total of five images were acquired across five slides per treatment sample.

Evaluation of B. burgdorferi morphology by fluorescence microscopy.

Fluorescent microscopy of live B. burgdorferi treated with PI was performed on a Nikon TI inverted fluorescence microscope equipped with a CoolSnap HQ2 camera (Photometrics, Tucson, AZ) and imaged using a 60× oil objective. In subsequent studies, spirochetes were treated with the membrane permeable dye SYTO9 (ThermoFisher), per the manufacturer’s instructions, and incubated at room temperature for 15 min to evaluate whether single spirochetes outside aggregates were truly devoid of PI staining. The experiments were performed on two independent occasions; within each occasion, the images were captured from five fields of view across five separate prepared slides.