Killing of Borrelia burgdorferi by Antibody Elicited by OspA Vaccine Is Inefficient in the Absence of Complement

Abstract

A Lyme disease vaccine, based on the Borrelia burgdorferi lipoprotein OspA, has recently undergone phase III trials in humans. The results of one of these trials indicate that vaccine efficacy positively correlates with anti-OspA antibody titer. Spirochete killing within the tick vector midgut, upon which vaccine efficacy appears to depend, may occur chiefly via a mechanism that involves antibody alone, as it has been reported that complement is degraded by tick saliva decomplementing factors. We compared the in vitro killing efficiencies of anti-OspA antibody elicited in rhesus monkeys by the OspA vaccine, in the presence and in the absence of monkey complement. Killing in the absence of complement was between 14 and 3,800 times less efficient than with complement present, depending on the spirochete strain. The relative inefficiency of the complement-independent killing mechanism by anti-OspA antibody may explain why OspA vaccine efficacy is critically dependent on antibody titer.

A vaccine to prevent infection with Borrelia burgdorferi, the spirochete that causes Lyme disease, has recently undergone phase III trials in humans (15, 17). The vaccine’s key component is recombinant lipidated outer surface protein A (OspA) of B. burgdorferi. Two independent phase III trials were conducted by researchers working with SmithKline Beecham Pharmaceuticals and Biologicals (17) and with Connaught Laboratories (15). The results of the SmithKline Beecham trial indicate that vaccine efficacy depends markedly on anti-OspA antibody titer. Thus, levels of antibody to the protective epitope of OspA, which is bound by the so-called LA₂ monoclonal antibody (18), were significantly lower (P < 0.01) at 1 month after the second injection in vaccinated subjects with breakthrough cases of definite Lyme disease than in vaccinated subjects in whom Lyme disease was not confirmed (17). Results of the Connaught trial also indicate that vaccine efficacy depends on factors that may influence antibody titer, as efficacy rates were higher among individuals less than 60 years old or in subjects that had received three rather than two vaccine doses (15).

Notwithstanding the possible reexpression of OspA by B. burgdorferi after spirochetal dissemination (7), it is generally accepted that anti-OspA antibody kills spirochetes mainly within the tick, while OspA is still expressed (4), but not immediately after the spirochetes invade the vertebrate host, when expression of OspA appears to be suppressed (2). Killing within the tick midgut, upon which OspA vaccine efficacy must primarily depend, may occur via a mechanism that involves antibody alone, as it has been reported that the saliva of Ixodes scapularis, the tick that is the most common vector of Lyme disease in the United States, has the ability to inactivate complement (9). Here, we have compared the in vitro killing efficiency of anti-OspA serum antibodies elicited in rhesus monkeys by the OspA vaccine in the presence and in the absence of monkey complement. The relative inefficiency of the complement-independent killing mechanism that is uncovered by our results may explain why OspA vaccine efficacy is critically dependent on anti-OspA antibody titer.

Several strains and species of the B. burgdorferi sensu lato complex were used. B. burgdorferi sensu stricto strains B31 (uncloned, low passage) and HB19 (uncloned, low passage) were obtained from the Centers for Disease Control and Prevention (CDC). A clone of the HB19 isolate was obtained by cloning twice in solid medium, as described previously (13, 16). Borrelia garinii IP90 (high passage) was also obtained from the CDC; strain NBS16 (low passage) and Borrelia afzelii P/Gau (low passage) and ECM-1 (high passage) were obtained from Denee Thomas, University of Texas Health Sciences Center, San Antonio, Tex. As a source of complement, blood samples were collected from uninfected, normal, anesthetized rhesus macaques by femoral venipuncture and were clotted at room temperature for 30 to 45 min. Clotted blood was then kept at 4°C for 2 h. After centrifugation of the samples at 800 × g for 20 min, the sera were decanted, pooled, and stored in small aliquots at −70°C until use. Serum from the animals chosen for this purpose did not contain cross-reactive anti-B. burgdorferi antibodies, as determined by Western blot analysis using whole B. burgdorferi antigens (1). A single pool of normal serum was used as a source of complement for all of the experiments reported herein. Anti-OspA antiserum was pooled from bleeds obtained from four rhesus macaques that were vaccinated with recombinant lipidated OspA (from B. burgdorferi sensu stricto strain ZS7) adsorbed onto aluminum hydroxide. The vaccine formulation and administration protocols, which also had been used in human trials (18), were described previously (11). To perform the antibody-dependent, complement-mediated killing (ADCK) or antibody-mediated killing (AMK) assays, spirochetes were cultured in BSK-H medium (Sigma Chemical Co., St. Louis, Mo.), as previously described (12), until they reached mid-logarithmic phase (about 2 × 10⁷ cells per ml). A total of approximately 5 × 10⁵ spirochetes in 25 μl of BSK-H medium was added to each well of a 96-well plate (Corning, Corning, N.Y.). A volume of 50 μl of a heat-inactivated (56°C, 30 min) anti-OspA antiserum pool appropriately diluted in the same medium had already been dispensed in each well. The plate was incubated at 34°C for 30 min before the addition of 25 μl of complement (normal monkey serum) for ADCK assays or heat-inactivated complement for AMK assays. After 24 h of incubation at 34°C in a humidified atmosphere of 3% CO₂ and 5% O₂, with the balance being N₂, 5 μl of each sample was removed and dead (nonmotile) and live (motile) spirochetes were counted under a dark-field microscope. All of the experiments reported herein were performed at least twice. In selected cases, surviving spirochetes were quantified by the ability to form colonies on solid medium (1). In all cases, the number of colonies obtained matched the number of living spirochetes that had been plated, after correction for the plating efficiency of the HB19 isolate, thus confirming our criterion for spirochete viability.

The ADCK₅₀ and AMK₅₀ values, defined as the antibody dilution at which 50% of the spirochetes are killed in 24 h in the presence and absence, respectively, of complement, for all of the strains used are shown in Table 1. Reciprocal geometric mean titers (GMT) of duplicate determinations are listed, together with the titer range. The differences in killing efficiencies between the complement-dependent and complement-independent mechanisms are striking. For spirochete species other than B. burgdorferi sensu stricto, such as B. garinii and B. afzelii, ADCK₅₀ GMT ranged between 1:84.8 and 1:240. In contrast, the AMK₅₀ values for such species were consistently below 1:10 in all cases. For strains of the B. burgdorferi sensu stricto species, such as B31, ADCK₅₀ GMT were higher than for the other two species, as would be expected from the higher level of identity (99.6%) between the mature OspA sequence of B. burgdorferi ZS7 (GenBank accession no. X16467) and that of B31 (GenBank accession no. X14407), compared to that found between ZS7 and the available sequence of IP90 (GenBank accession no. U65821), which was 95% over 80 amino acids of the mature protein. However, this higher ADCK₅₀ value was not ostensibly paralleled by the AMK₅₀ value, which remained below 1:10 for the B31 strain. Even for spirochetes of the HB19 strain, whose mature OspA is 100% identical to that of ZS7 (16), the measured AMK₅₀ GMT was 1:1.85 whereas the ADCK₅₀ GMT was as high as 1:7,000. Since we had previously found that OspA escape mutants were present amidst spirochetes of our uncloned HB19 isolate (16), we utilized for this experiment an HB19 wild-type clonal population. This was relevant, since we had demonstrated that these escape mutants, which had been selected in the presence of a rhesus monkey serum pool similar to the one used herein, were by definition resistant to AMK but were nonetheless sensitive to ADCK (16). Hence, the uncloned isolate of HB19 could have shown an unusual resistance to AMK due to the presence of these mutants within the wild-type population. All of the mutants that were isolated exhibited deletions within the ospAB operon which resulted in the expression of either a chimeric protein that contained the N terminus of OspA and the C terminus of OspB or native OspA together with an OspB molecule lacking a portion of its C terminus (16).

TABLE 1.

In vitro killing efficiencies of different B. burgdorferi sensu lato species with anti-OspA antiserum in the presence and absence of rhesus monkey complement

Species	Strain	GMT (range)a		% Killingb	Serum dilutionc
		1/ADCK50	1/AMK50
B. burgdorferi sensu stricto	B31	1,250 (NA)	<10	17	1:1,300
	HB19 (clone)	7,000 (6,950–7,050)	1.85 (1.5–2.3)	NDd	ND
B. garinii	IP90	240 (235–245)	<10	10	1:430
	NBS16	84.8 (80–90)	<10	4	1:140
B. afzelii	P/Gau	127.4 (125–130)	<10	7	1:235
	ECM-1	130 (NA)	<10	5	1:220

^aNA, not applicable (i.e., titer range was zero). 

^bPercent killing at a 1:10 serum dilution, with no complement. 

^cSerum dilution at which the percent killing given in the previous column was obtained in the presence of complement. 

^dND, not done. 

As an additional indication of how much more demanding in terms of anti-OspA antibody titer the AMK mechanism is, we tabulated the percent killing obtained at a 1:10 serum dilution in the absence of complement and the dilution at which the same percent killing was obtained when complement was present (Table 1). The serum dilution could be between 14 and 130 times higher and still result in the same killing efficiency by ADCK.

The mechanism of ADCK of B. burgdorferi requires antibody to facilitate the interaction of the bacterial membrane with the complement C5b-9 membrane attack complex. The latter originated from the classical pathway C5 convertase (8). This “facilitation” is possible even with Fab monomer (8) and, as mentioned above, can be elicited by the binding of anti-OspA antibody to OspA-OspB chimeras and thus does not require a structurally intact OspA molecule (16). Binding of anti-OspA antibody to any surface-exposed epitope appears to be sufficient to trigger the chain of events that leads to the effective formation of a membrane attack complex and subsequent spirochetal death. In contrast, AMK with anti-OspA antibodies appears to require not only a structurally intact OspA molecule (13) but also an OspB that retains at least its C terminus, as we were able (with our anti-OspA antiserum) to select escape mutants which expressed a full-length wild-type OspA and a truncated OspB derived from an ospB gene that lacked 220 C-terminal base pairs comprised of bp 1256 to 1477 of the ospAB operon (16). A pertinent implication of this finding is that for AMK to be effective both OspA and OspB need to be bound by anti-OspA antibodies. Since the 121 C-terminal amino acids of OspA and OspB are 52% identical (16), this conjecture is not only possible but entails also that more anti-OspA antibody is required for AMK to operate as effectively as ADCK. Binding of a C-terminal epitope of OspB is necessary for the complement-independent killing of B. burgdorferi with anti-OspB monoclonal antibodies (3, 5). However, on Western blots of whole-cell extracts of HB19 spirochetes, the monkey anti-OspA antiserum which we used in this study reacted only with OspA, not with OspB (our unpublished data). Thus, the postulated requirement of dual binding of OspA and OspB for AMK to occur may hold only if the OspB epitope(s) bound by the anti-OspA antibody is conformational and is destroyed by the Western blotting procedure.

The fact that anti-OspA AMK appears to require structural integrity of both OspA and OspB further suggests that this AMK mechanism acts by inhibiting the function of OspA. While this function remains unknown (10), it is very likely that, for it to be manifest, both OspA and OspB must be present simultaneously in the spirochetal membrane. Such a hypothesis would be consistent with the fact that both molecules are part of an operon and are thus transcribed and possibly expressed in unison. It stands to reason that to fully inhibit the activity of the putative OspA-OspB functional unit in a manner that leads to spirochetal death, a higher antibody titer might be required than for complement-dependent killing. Both OspA and OspB are abundantly expressed by B. burgdorferi either when these organisms are cultivated in vitro or when they are present in unfed ticks, and many molecules might have to be bound by antibody before their function is inhibited such that spirochetal survival is impaired. Functional inhibition of the putative OspA-OspB unit may depend, moreover, on the binding of antibody to a particular OspA epitope such as LA₂. Indeed, anti-OspA antibody reactivity with this epitope strictly correlates with protection against tick challenge (6). Binding of the LA₂ epitope as a condition for effective AMK would explain the need of a higher polyclonal anti-OspA antibody titer for AMK than for ADCK. The latter mechanism appears to be triggered, as mentioned above, by the binding of any surface-exposed epitope. The need for a structurally intact OspB could still be invoked if the latter molecule is part of the postulated OspA-OspB functional unit, but the binding of OspB by anti-OspA antibody would not be necessary. It should be noted, however, that for the “functional inhibition” hypothesis to be tenable, other molecules must be able to fulfill the function of the OspA-OspB unit in mutant spirochetes that do not express any of these lipoproteins or express OspA-OspB chimeras. Both types of mutants are viable when cultivated in vitro (13, 16), and the latter type has been isolated from ticks (14).

If, as reported by Mather et al. (9), complement is rendered nonfunctional by decomplementing factors present in tick saliva, our data indicate that efficacious anti-OspA antibody titers may be all the more difficult to achieve in the absence of complement and may thus contribute to explaining why OspA vaccine efficacy is critically dependent on anti-OspA antibody titer (17).