INFECTION AND IMMUNITY, Mar. 1997, p. 882–889 0019-9567/97/$04.0010 Copyright q 1997, American Society for Microbiology

Vol. 65, No. 3

Reactivity with a Specific Epitope of Outer Surface Protein A Predicts Protection from Infection with the Lyme Disease Spirochete, Borrelia burgdorferi WILLIAM T. GOLDE,1 JOSEPH PIESMAN,1 MARC C. DOLAN,1 MICHAEL KRAMER,2 PIERRE HAUSER,3 YVES LOBET,3 CARINE CAPIAU,3 PIERRE DESMONS,3, PIERRE VOET,3 DON DEARWESTER,4 4 AND JOSEPH C. FRANTZ * Division of Vector-Borne Infectious Diseases, National Center for Infectious Diseases, Centers for Disease Control and Prevention, Public Health Service, U.S. Department of Health and Human Services, Fort Collins, Colorado 805221; Institut fur Immunologie und Serologie, Universita ¨t Heidelberg, W-6900 Heidelberg, Germany2; SmithKline Beecham Biologicals, B-1330 Rixensart, Belgium3; and Central Research Division, Pfizer Inc., Lincoln, Nebraska 685014 Received 5 June 1996/Returned for modification 15 July 1996/Accepted 11 December 1996

The response to recombinant vaccines for Lyme disease was studied to determine serum antibody levels effective in protecting against tick-transmitted infection. Data presented here demonstrate a significant correlation between antibody to an epitope on outer surface protein A (OspA) and protection against infection with Borrelia burgdorferi in canines and mice. A competitive enzyme-linked immunosorbent assay was developed to measure antibody to a site on OspA, defined by monoclonal antibody LA-2. Comparison of LA-2 titers against infection of canines and mice following vaccination and challenge established a predicted value for LA-2 titers. The statistical relationship between serum antibody levels and protection was calculated by logistic regression analysis. The statistical model predicted that an LA-2 titer of 0.32 mg equivalents (eq) per ml correlated to an 80% predicted probability of protection for both mice and dogs. This value was used to classify mice and dogs as to their protected status at the time of tick exposure. The LA-2 cutoff titer (0.32 mg eq/ml) correctly classified all dogs (n 5 13) and mice (n 5 44) that failed to become infected. By contrast, 20 of 22 dogs and 28 of 31 mice with titers of less than 0.32 mg eq/ml became infected. On the basis of these results, we conclude that an LA-2 titer is a reliable indicator of immune status for estimating immune protection following use of OspA-based vaccines for B. burgdorferi sensu stricto. imal level of LA-2 equivalence required for protection, we applied the model developed in the mouse study to an ongoing vaccine trial with dogs in which the experimental vaccine was formulated with nonspecific protein 1 (NS1)-OspA. These animals were challenged in a laboratory model using field-collected ticks described originally by Appel et al. (1).

One approach to developing of a vaccine for Lyme disease has focused on use of recombinant outer surface protein A (OspA) of Borrelia burgdorferi as an immunogen. Data from published studies show that this recombinant protein, formulated with various adjuvants, elicits a protective immune response (9, 12, 13, 21, 24, 25, 26). In a number of these studies, individuals with sera containing measurable levels of OspAspecific antibody became infected following challenge (5, 9, 13, 14). This finding indicates that some antibodies to OspA failed to contribute to protection. By contrast, certain regions on OspA serve as a useful target for borreliacidal antibody (20, 21). We sought to identify whether there might be specific epitopes on OspA that induce serum antibodies important for protection against infection. A series of monoclonal antibodies specific for OspA have been reported (3, 15, 28). One of these antibodies, LA-2, appears to identify an epitope on OspA that is associated with a protective immune response. LA-2 is a surface-expressed epitope of OspA and serves as a target for antibody that can kill B. burgdorferi sensu stricto. We have developed a competitive enzyme-linked immunosorbent assay (ELISA) that measures serum antibody levels to a site on OspA putatively associated with binding of bactericidal antibody (14). Using sera obtained from individual mice (13), we sought to determine if LA-2 equivalence titers of these serum samples correlated with protection against infection. Once we had established the min-

MATERIALS AND METHODS Vaccines. Mice were vaccinated with four different OspA formulations as described previously (13). Briefly, the gene for OspA was isolated from B. burgdorferi sensu stricto ZS7, recombined, and expressed in Escherichia coli as either a whole lipidated protein or fused protein with NS-1 of influenza virus. The fusion protein is not lipidated. The codons for the N-terminal cysteine of OspA reside directly adjacent to the terminal codons of NS-1. Consequently, a primary amino group on the OspA moiety of the fusion protein is not available. The OspA gene of B. afzellii ACA-1 was also isolated and expressed, fused with NS-1, in E. coli. All but one of the experimental formulations for vaccination of mice were adsorbed to aluminum hydroxide gel. The ratio for adsorption was 1 mg of protein per 100 mg of aluminum hydroxide gel (13). Vaccination protocol. Outbred BALB/c and C3H/HeJ mice received three vaccinations at 2-week intervals. Serum samples were collected 1 week after the last vaccination and 1 week before challenge with infected Ixodes scapularis ticks. Details regarding the numbers of ticks used to challenge mice were reported previously (11). In an attempt to determine the effectiveness of the Lyme disease vaccine formulation in the normal regimen of canine vaccination, we incorporated the OspA vaccine into the normal, recommended canine schedule. Dogs were randomly assigned to one of four groups. At 8 to 9 weeks of age, all dogs received a dose of canine distemper-measles virus vaccine, modified live virus (MLV), and a dose of canine parvovirus vaccine (MLV). At 12 weeks of age, each dog received canine distemper virus, adenovirus type 2, parainfluenza virus, parvovirus (MLV), and Leptospira canicola-icterohaemorrhagiae bacterin. Also at 12 weeks of age, each dog in three groups received a dose of NS1-OspA. A second dose NS1-OspA was administered at 16 weeks of age. One group of dogs (controls) did not receive the NS1-OspA vaccine. Canine vaccinations were administered by the subcutaneous route.

* Corresponding author. Present address: Pfizer Inc., Central Research, Eastern Point Road, Groton, CT 06340-5146. Phone: (860) 441-8728. Fax: (860) 441-8739. E-mail: [email protected]. 882

VOL. 65, 1997

PREDICTIVE ELISA FOR LYME DISEASE VACCINE

883

Testing of experimental canine vaccine. Vaccines administered to dogs were tested by ELISA to measure the content of NS1-OspA. Briefly, to a 96-well plate (Nunc-Immuno Plate MaxiSorp; Nunc, Roskilde, Denmark) was added 100 ml of capture monoclonal antibody 1-2 (1.2 mg/ml) in 0.01 M sodium borate (pH 9.4). The plate was incubated overnight at 48C and then washed with phosphatebuffered saline (PBS). Then 200 ml of PBS containing 0.1% Tween 20 and 5% nonfat dry milk was added, and the mixture was incubated for 1 h at 378C. The plate was washed with PBS, highly purified NS1-OspA reference antigen was applied, and the mixture was incubated for 1 h at 378C. A monoclonal antibody having bacterial killing activity similar to that of LA-2 was used as an indicator antibody. This antibody (1-5) also competitively inhibited binding between LA-2 and OspA. One hundred microliters of indicator antibody (monoclonal antibody 1-5–biotin conjugate; 60 ng/ml) was added to the wells. After 1 h at 378C, the plate was washed with PBS. Then 100 ml of peroxidase-labeled streptavidin (1 ml/ml; Kirkegaard & Perry Laboratories, Gaithersburg, Md.) diluted in nonfat dry milk was added. The plate was incubated for 1 h at 378C and washed with PBS, and 100 ml of 2,29-azinobis(3-ethylbenzthiazolinesulfonic acid) (ABTS; Kirkegaard & Perry) was added. The concentration of NS1-OspA was determined by interpolation from a reference curve constructed from the absorbance (405 nm) of a series of wells containing dilutions of a highly purified reference standard of NS1-OspA. The reference curve was constructed by using fourparameter logistics. Tests for coparallelism were performed as described by Plikaytis et al. (19). Challenge procedure for dogs. Adult ticks (I. scapularis) were collected from a region of Westchester County, New York, where Lyme disease is known to be endemic (8). Ticks were segregated based on sex, placed into preservation vials, and stored at 218C in 97% relative humidity until applied to dogs. A tick containment capsule was applied to the rib cage of each dog the day prior to exposure to ticks (day 21). As described by Appel et al. (1), 15 adult female and 7 adult male ticks were placed into the capsule on the day of tick infestation (day 0), and the capsule was tightly sealed. No efforts were made to determine a more suitable ratio of males to females than the ratios tested. Tick attachment and feeding was inspected no earlier than day 6 of infestation. Capsules were removed when adult females had fed to repletion, or no later than day 14. Immunoblotting. Immunoblotting was performed on canine sera only. Blood was collected in vacuum tubes (Becton Dickinson Vacutainer Systems, Rutherford, N.J.) by drawing samples from the jugular vein. Sera were prepared from whole blood by centrifugation at approximately 600 3 g for 20 min at ambient room temperature. The supernatant fraction (serum) was decanted, heated at 568C for 30 min, and stored at 2208C until analyzed. Prior to testing, serum samples were gently thawed at ambient room temperature. Briefly, immunoblotting was performed on commercial strips (MarDx, Inc., San Diego, Calif.) as follows. Nonspecific binding sites were blocked prior to use by the manufacturer. Control antibodies (monoclonal antibodies specific for OspA and P39) or unknown canine sera were diluted in 10 mM Tris–150 mM NaCl–0.1% Triton X-100 (TBST) and added to individual strips. Antibodies were allowed to incubate for 0.5 h with shaking at ambient room temperature. Following three washes in TBST, a cocktail of second antibody conjugates (anti-mouse immunoglobulin G–phosphatase or anti-canine immunoglobulin G [heavy plus light chain]–phosphatase) was diluted in TBST and added to appropriate strips. The blot paper and antibody conjugates were incubated at ambient room temperature for 15 min. Following three washes in TBST and one wash in distilled water, 5-bromo4-chloro-3-indolylphosphate–nitroblue tetrazolium (Kirkegaard & Perry) substrate was added. The reaction was terminated by discarding the substrate and adding deionized water to the blot. Blots were preserved in a cool, dark place until photographed. The results for individual dogs were scored based on reactivity against an array of B. burgdorferi proteins by using criteria similar to those proposed by Dressler et al. (7). Apparent molecular masses of the 10 proteins in the array for evaluating canine sera were 18, 21, 30, 35, 39, 41, 45, 58 to 60, 71, and 93 kDa. An individual dog was determined to be infected if serum from that dog reacted against 5 of 10 proteins in the array (7). Bactericidal assay. The bactericidal assay measures antibodies in sera that can recognize and kill live B. burgdorferi. Serial dilutions of sample sera were prepared in BSK-II medium containing 6% young rabbit serum (PelFreez, Rogers, Ark.). Guinea pig complement (Gibco-Bethesda Research Laboratories, Grand Island, N.Y.) was added to diluted serum at 267 U/ml. One hundred microliters of each dilution of sera was transferred to four replicate wells of a 96-well microtiter plate (Costar, Cambridge, Mass.). A 3- to 4-day-old culture of highpassage (more than 60 in vitro passages upon isolation from a tick) B. burgdorferi B31 was diluted to give 1 3 108 to 2 3 108/ml in fresh BSK-II medium containing sample sera. Known positive and negative sera were used as assay controls. A back titration of the challenge was prepared by inoculating wells containing 100 ml of growth medium with various concentrations of B. burgdorferi B31. The plates were incubated at 328C for 3 to 4 days. The plates were read spectrophotometrically at 560 nm. Change in the indicator dye (phenol red) from red to purple represented inhibition of growth or killing of B. burgdorferi by sample sera. Endpoint titers were determined by comparison of absorbance values in each well to the value from wells containing 2 3 107 B31 cells/ml, which reflected the absorbance value for 90% killing. LA-2 competitive ELISA. Antibody having binding activity to OspA equivalent to that to LA-2 was measured by a competitive ELISA format as described previously (14). Recombinant OspA was purified as described previously (13).

Ascites fluid containing monoclonal antibody LA-2 was the kind gift of Marcus Simon, Max Planck Institute, Freiburg, Germany. LA-2 was purified and conjugated with biotin as previously described (14). Briefly, purified OspA was diluted to 100 ng/ml in 0.1 M borate buffer (pH 9.3). One hundred microliters was added to desired wells of 96-well ELISA plates (Nunc-Immuno Plate MaxiSorp). Plates sensitized with OspA were incubated overnight at 48C. Studies to optimize the blocking reagent established that background absorbance was negligible for ELISA components in wells not coated with OspA. Contents of the wells were discarded and filled with 300 ml of PBS containing 0.1% Tween 20 (Sigma Chemical Co., St. Louis, Mo.) and 1.0% very low viscosity polyvinyl alcohol (Aldrich Chemical Company, Inc., Milwaukee, Wis.) (PBS-Tween 20-PVA). Following incubation for 1 h at 378C, plates were washed once with PBS-Tween 20 and twice with PBS. For measurement of murine sera, the next component added was either nonconjugated LA-2 or unknown murine serum. For analysis of canine sera, the next component in the reaction was either reference canine serum or unknown canine serum. Reference canine sera were tested and determined to contain 10.406 mg equivalents (eq) of LA-2 per ml. Reference canine sera, unknown canine or murine sera, and monoclonal antibody LA-2 used as reference for construction of a competition curve were diluted in PBS-Tween 20-PVA containing 0.1% nonimmune mouse serum and added in 100-ml volumes to triplicate wells. Standard concentrations of reference canine sera included nine twofold serial dilutions beginning with 520.3 ng eq/ml. Absorbance values from wells containing reference canine sera were used to construct a reference curve. Unknown sera were added at dilutions that ensured absorbance values within a range that could be interpolated from the reference curve. A control serum sample was included on each plate to determine day-to-day and plate-toplate variation. Nonconjugated antibody was omitted from wells designed to establish maximum absorbance values for that plate. Following incubation for 1 h at 378C, plates were washed as described above. Biotin-conjugated LA-2 was diluted to approximately 380 ng/ml, and 100 ml was added to all wells. Following incubation for 1 h at 378C, plates were washed as described above. From a concentrate (500 mg/ml), peroxidase-labeled streptavidin was diluted to 1 mg/ml in PBS-Tween 20-PVA and added in 100-ml volumes to all wells. Following a final 1-h incubation at 378C, plates were washed as described above. Peroxidase substrate (ABTS; Kirkegaard & Perry) was added in 100-ml volumes to all wells. Absorbance values were determined when absorbance (405 nm) of the wells devoid of nonconjugated antibody was approximately 1.0 absorbance unit. For analysis of murine sera, the mean absorbance values for wells that received standard concentrations of nonconjugated LA-2 were used to construct a reference curve. The points of the standard curve were fitted by using exponential regression analysis. Studies established the slope of LA-2 competition (20.639 6 0.065, n 5 26). Values for LA-2 microgram equivalents for mouse sera were estimated by interpolation from the historical reference curve for polyclonal sera. This reference curve was constructed to originate from a point on the x axis by projecting the LA-2 competition curve for each test plate through the abscissa. From this point on the abscissa, the slope for polyclonal sera (20.298 6 0.074) was used to construct the reference curve for polyclonal mouse sera. For analysis of canine sera, the mean absorbance values for wells that received standard concentrations of reference canine sera were used to construct a reference curve. The points of the standard curve were fitted by using four-parameter logistics. Data from all plates were considered acceptable if the results of the competition curve met three criteria: (i) the correlation coefficient for the reference curve had to be $0.98, (ii) the value for the control sample interpolated from the reference curve had to be within 20% of a historical mean, and (iii) analysis for coparallelism had to conform to criteria described by Plikaytis et al. (19). Serum antibody levels for unknown sera were estimated by interpolation from the reference curve and were expressed as microgram equivalents per milliliter. Direct culture of skin biopsy samples. Skin biopsy samples were collected at 3-week intervals beginning 1 week following the completion of tick feedings. Each biopsy sample was tested for the presence of spirochetes. With 2% lidocaine used as a local anesthetic, one or more biopsy samples were taken. Unless otherwise specified, biopsy samples were taken within 10 cm of the tick attachment site. Culture attempts were performed following surface decontamination of the biopsy sample. All samples were exposed to fresh, sterile equipment or containers to avoid cross-contamination. Each biopsy sample was minced with a sterile razor blade, and the pieces of tissue were transferred to approximately 4 ml of BSK-II culture medium. Cultures from these tissue samples were observed weekly for at least 6 weeks by dark-field microscopy for viable spirochetes. Direct culture of ticks. After ticks were removed from challenged dogs, they were cultured directly in BSK-II medium. Each live female replete or feeding tick was tested for the presence of spirochetes. By using 0.05% Wescodyne (AMSCO, Medical Products Division, Erie, Pa.), the surface of each tick was decontaminated for approximately 5 min. No attempt was made to culture dead female ticks. Culture attempts were performed immediately following surface decontamination. All ticks were exposed to fresh, sterile equipment or containers to avoid cross-contamination. Each tick was pierced with a sterile 20-gauge needle in the area of the body containing host blood. Liquid extracted from the midgut of each tick, or the entire tick, was then placed in a 3.5-ml snap-cap tube containing BSK-II culture medium and incubated at 308C. Samples were examined weekly for at least 6 weeks by dark-field microscopy for viable spirochetes.

884

GOLDE ET AL.

INFECT. IMMUN.

TABLE 1. LA-2 equivalence titers of sera from outbred mice Mouse

3075 3076 3077 3078 3079 3070 3071 3072 3073 3074 3065 3066 3067 3068 3069 3060 3061 3062 3063 3064 3055 3056 3057 3058 3059 a b

Immunization

Alum

NS1-OspA (ZS7) 1 alum

Whole OspA (ZS7)

Whole OspA (ZS7) 1 alum

NS1-OspA (ACA-1)

Prechallenge OspA titera

Ear biopsy culture

LA-2 ELISA (mg eq/ml)

,125 125 250 125 250 8,000 8,000 32,000 4,000 8,000 16,000 32,000 4,000 16,000 128,000 64,000 64,000 128,000 16,000 32,000 2,000 2,000 4,000 32,000 2,000

1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 2 1

0.035 0.046 0.035 0.033 0.023 2.088 1.695 1.002 NDb 1.125 0.103 0.588 0.466 0.626 0.719 ND 0.406 1.355 4.628 3.073 0.091 0.027 0.045 0.027 0.064

Data published previously by Golde et al. (13). ND, not determined.

RESULTS Serologic response in mice and response to challenge. Data in Table 1 show antibody levels in sera taken from outbred mice following vaccination with the various formulations of OspA. Also shown are postchallenge infection results for individual mice. As we reported previously, these results show that vaccination with OspA protects mice against infection after challenge with ticks harboring a strain of B. burgdorferi sensu stricto homologous for OspA. By contrast, protection was not observed when the OspA used as the immunogen was derived from B. afzellii. When serum samples collected before tick challenge were assayed for total antibody to OspA (ZS7), the response did not seem to compare favorably with protection. For instance, sera from mice vaccinated with ACA-1 OspA reacted prominently against B31 OspA. However, only one mouse (mouse 3058) remained uninfected and had a very high level of antibody against B31 OspA. The other mice from this group had significant titers of 1:2,000 to 1:4,000 yet were not protected. Alternatively, there appeared to exist a good agreement between LA-2 equivalence titers and protection against infection. One mouse, 3065, had a low LA-2 equivalence titer and remained uninfected following challenge. In the case of the inbred BALB/c mice, again LA-2 equivalence titers identified individuals that remained uninfected after challenge. Here, the difference between control, shamvaccinated animals and homologous OspA-vaccinated mice was even more striking. As with outbred mice, the heterologous vaccination with OspA derived from strain ACA-1 (B. afzellii) did not induce significant LA-2 titers, and these mice became infected upon tick challenge (Table 2). The same result was observed in the inbred strain C3H/HeJ (Table 3). This strain was included in this study because these mice developed arthritis and carditis subsequent to B. burgdorferi infection (4). Interestingly, we observed one vaccination

TABLE 2. LA-2 equivalence titers of sera from BALB/c mice Mouse

3191 3192 3193 3194 3195 3196 3197 3198 3200 3176 3177 3178 3179 3180 3181 3182 3183 3184 3186 3187 3188 3189 3190 a

Immunization

Alum

NS1-OspA (ZS7) 1 alum

Whole OspA (ZS7)

Whole OspA (ZS7) 1 alum

NS1-OspA (ACA-1) 1 alum

Prechallenge OspA titera

Ear biopsy culture

LA-2 ELISA (mg eq/ml)

,125 125 250 125 ,125 16,000 16,000 8,000 8,000 2,000 8,000 8,000 16,000 8,000 4,000 16,000 8,000 8,000 250 125 ,125 250 250

1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1

0.143 0.169 0.175 0.220 0.147 22.910 33.970 2.514 0.485 1.775 0.972 2.072 0.540 0.516 0.713 3.269 2.546 3.244 0.072 0.036 0.065 0.047 0.075

Data published previously by Golde et al. (13).

failure. Serum from mouse 3217 had no detectable antibody against OspA and a negative LA-2 equivalence titer. This animal also became infected upon challenge. For the remaining animals in this group, there existed a clear correlation between LA-2 equivalence titer and resistance to infection from ticktransmitted B. burgdorferi. Tick feedings on dogs and direct culture of ticks. Tickfeeding results are summarized in Table 4. In total, 36 dogs

TABLE 3. LA-2 equivalence titer of sera from C3H/HeJ mice Mouse

3251 3252 3253 3254 3255 3256 3257 3258 3259 3215 3216 3217 3218 3223 3224 3225 3261 3262 3263 3264 3265 a b

Immunization

Alum

NS1-OspA (ZS7) 1 alum

Whole OspA (ZS7)

Whole OspA (ZS7) 1 alum NS1-OspA (ACA-1) 1 alum

Prechallenge OspA titera

Ear biopsy culture

LA-2 ELISA (mg eq/ml)

125 125 125 ,125 ,125 64,000 32,000 64,000 64,000 32,000 16,000 125 8,000 16,000 16,000 16,000 2,000 250 250 ,125 500

1 1 1 1 1 2 2 2 2 2 2 1 2 2 2 2 1 1 1 1 1

0.023 0.019 0.018 0.015 NDb 4.628 5.197 10.910 8.390 0.489 0.466 0.026 0.226 23.812 9.421 5.614 0.069 0.131 0.036 0.036 0.091

Data published previously by Golde et al. (13). ND, not determined.

VOL. 65, 1997

PREDICTIVE ELISA FOR LYME DISEASE VACCINE

885

TABLE 4. Summary of feeding and spirochete infection rates for adult female ticksa Dose (mg)

Total no. of ticks applied (F/M)

Avg no. of ticks applied (F/M)

Avg no. of females fed or replete

% Replete and feeding females

No. of ticks culture positive/no. cultured (%)

10 1 0.1 0 (controls)

136/63 136/63 136/63 135/63

15.1/7.0 15.1/7.0 15.1/7.0 15.0/7.0

13.8 13.5 13.3 13.9

91.2 89.7 88.2 92.6

0/115 (0.0) 3/117 (2.5) 6/113 (5.3) 21/121 (17.3)

a

Each group contained nine dogs. F, female; M, male.

were challenged with adult ticks as described above. Each dog was exposed to at least 15 adult female ticks. At least 13 female ticks fed on each dog. Each feeding tick removed from a dog was cultured for viable spirochetes. A representative sampling from the pool of ticks used for this challenge study indicated a 44% infection rate of B. burgdorferi prior to challenge, as measured by an antigen capture ELISA designed to detect the presence of spirochetes in ticks (6). Following challenge, 17.3% of the cultivable female ticks removed from nonvaccinated dogs provided a positive culture. One dog in the group that received 10 mg of NS1-OspA became infected. Our attempt to culture ticks removed from that dog failed because of overgrowth by contaminant organisms and were not counted or scored as negative. Ticks removed from the remaining dogs from the group that received 10 mg of NS1-OspA did not provide a positive culture. By contrast, ticks from dogs vaccinated with 1 or 0.1 mg of NS1-OspA provided 2.5 and 5.3% positive cultures, respectively. Direct culture of canine skin biopsy samples. Spirochetes were recovered from skin samples beginning 30 days after tick challenge. The results are summarized in Table 5. Skin samples from all nonvaccinated control dogs provided positive cultures at each of the 3-month periods of observation. One dog that received 10 mg of NS1-OspA provided a positive culture from a skin biopsy sample taken 30 days after challenge and remained infected throughout the observation period. Some dogs that received 1 or 0.1 mg became culture positive during month 2 or 3 of observation (Table 5). Immunoblotting of canine sera. Results for immunoblot analysis of sera taken 90 days after challenge are shown in Table 6. At 90 days postchallenge, dogs 401 and 402 (vaccinated with 1 mg) tested positive for spirochetes by culture of skin biopsy samples; however, sera from these two dogs did not react by immunoblotting until 5.5 months after challenge. Two dogs (no. 453 [0.1-mg dose group] and no. 409 [a control]) became immunoblot positive during the first month of observation and remained so throughout the period of observation. For the rest of the dogs, the onset of a positive immunoblot occurred during month 2 after challenge and persisted throughout the observation period. Borreliacidal antibody in canine sera. As described in Materials and Methods, the borreliacidal antibody test measured killing of B. burgdorferi by serum and complement. The results are expressed as the reciprocal of the last dilution causing 90% bacterial lysis as determined spectrophotometrically. Borreliacidal antibody levels were measured prior to each vaccination and 2 weeks following the second dose of vaccine. Geometric mean titers (GMTs), by group, are provided in Table 7. Ten micrograms of NS-1 OspA induced the greatest response. Lower doses of NS1-OspA gave rise to a corresponding lower response. LA-2 ELISA. Levels of antibody to OspA were measured by LA-2 ELISA prior to each vaccination and throughout the observation period. Antibody levels (GMTs) at 2 weeks fol-

lowing vaccination are provided in Table 8. On the basis of GMTs, the vaccination regimen induced a dose-dependent response. Shown in Fig. 1 are the LA-2 titers throughout the vaccination and postchallenge observation periods. Interestingly, exposure to B. burgdorferi by means of tick challenge did not induce an increased response against the LA-2 epitope (Fig. 1). Following challenge, antibody levels either decreased or remained at very low levels throughout the observation period. At 3 months after challenge, GMT for the groups that received 10 and 1 mg decreased to 0.454 and 0.412 mg eq/ml, respectively. The results of LA-2 ELISA and borreliacidal antibody determinations were compared by regression analysis. The correlation between the LA-2 ELISA and borreliacidal assay was indicative of a relatively close relationship (r 5 0.6733). Correspondence between antibody to OspA and protection. Animals were declared infected on the basis of direct culture of skin punch biopsy samples and, in dogs, analysis of sera by immunoblotting. Every dog that provided a positive skin culture eventually reacted by immunoblotting. In dogs, GMTs were dose dependent; however, individual responses for dogs within each group varied. In mice, the dose administered induced a broad range of responses. It was of interest, therefore, to compare serum antibody levels (LA-2 ELISA) against infection status for individual animals to test for a correlation between infection and LA-2 titer. The predicted probability of protection (PPP) was estimated by logistics regression analysis. On the basis of this statistical method, an 80% PPP corresponded to LA-2 titers of 0.3196 mg eq/ml for dogs and 0.3230 mg eq/ml for mice. Classification of animals based on LA-2 titer. Using the LA-2 titer that corresponded with an 80% PPP, we classified dogs and mice as to their status of infection. The 80% PPP (LA-2 titer of $0.32 mg eq/ml) correctly classified 13 of 13 and 44 of 44 uninfected dogs and mice, respectively, following challenge (Table 9). An 80% PPP (LA-2 titer of #0.32 mg eq/ml) correctly classified 20 of 22 and 28 of 31 infected dogs and mice, respectively. Two dogs had LA-2 titers greater than the background and #0.32 mg eq/ml but did not become infected. Individual LA-2 titers in sera from these two dogs were

TABLE 5. Direct culture of skin biopsy samples for 3 months after challenge

Dose (mg)

10 1 0.1 0 (controls)

No. of dogs/group

9 8 9 9

Culture result for sample taken at indicated month postchallenge (no. positive/total) 1

2

3

1/9 1/8 6/9 9/9

1/9 1/8 7/9 9/9

1/9 3/8 7/9 9/9

886

GOLDE ET AL.

INFECT. IMMUN.

TABLE 6. Recognition of B. burgdorferi protein band by canine sera taken monthly after tick challenge Dose (mg)

10

Total no. positive 1

Total no. positive 0.1

Total no. positive Controls

Total no. positive a b

Dog

Recognition of B. burgdorferi protein by canine sera p18

p21

403 417 428 430 441 451 455 476 482

1

401 402 407 410 422 463 472 500 502

(1)b (1)

408 424 425 431 445 453 474 498 504

1 1 1 1 1 1 1 7

4

404 406 409 419 427 439 443 461 478

p30

1

p35

p39

p41

p45

p58

p71

p93

1

1

1

1

1

1

1

1

1

1

1

1

1

1

(1) (1)

(1) (1)

(1) (1)

(1) (1)

(1) (1)

(1) (1)

(1) (1)

1 1

1

0

1 1

1

3

1 1 1 1 1 1 1 1 1 9

1 1 1 6

1 1

1

1

1

1

1

3

3

3

3

3

3

3

1 1

1 1

1 1 1

1 1

1 1

1 1 1 1 1 1

1 1 1 1 1 1

1 1 1 1 1 1

1

1 1

1 1 1 1 1 1

1 1 1 1

2

4

1 6

1 7

6

1 7

1 7

5

0

1 1 1 1 1 1 1 1 1 9

1 1 1 1 1 1 1 1 1 9

1 1 1 1 1 1 1 1 1 9

1 1 1 1 1 1 1 1 1 9

1 1 1 1 1 1 1 1 1 9

1 1 1 1 1 1 1 1 1 9

0

1 1 1 5

1 1

1 1

1 1 1 5

1 1 1 1 1 1 1 1 8

Total no. positivea

Positive skin culture

8 1 0 1 1 0 1 1 1

Yes No No No No No No No No 1/9

8 8 1 2

Yes Yes No No Died Yes No No No 3/9

7 1 1 1 7 8 9 6 9 9 1 1 5

Yes Yes Yes Yes Yes Yes No No Yes 7/9

9 8 9 7 8 9 9 9 9

Yes Yes Yes Yes Yes Yes Yes Yes Yes 9/9

Dogs were scored infected if their sera reacted against five or more marker proteins. (1), immunoblot reaction for sera collected 5.5 months after challenge. Sera collected prior to 5.5 months after challenge failed to react on immunoblots.

0.208 and 0.122 mg eq/ml, respectively. Similarly, three mice with LA-2 titers of ,0.32 mg eq/ml did not become infected. Titers for these animals were 0.226, 0.103, and 0.027 mg eq/ml (Table 1). Response of dogs to vaccine dose. On the basis of antibody levels (LA-2 ELISA) at 2 weeks after the second vaccination, the experimental vaccine induced a dose-dependent response (Fig. 2). The range of responses for each dose administered traversed an 80% PPP for each group of dogs except for the nonvaccinated control group. Despite a broad scatter of responses in each group, the GMTs decreased in proportion to the dose administered. DISCUSSION The aim of this study was to induce in animals a range of responses that would traverse and establish a threshold for immune protection. It was of particular interest to investigate

the influence of minimal antibody levels on establishment of infection. It was fortuitous that the number of mice and the range of their responses permitted a comparison of LA-2 titer and infection by logistics regression analysis. In dogs, the objective of inducing a dose-dependent response was achieved. In

TABLE 7. GMTs for borreliacidal antibody levels in dogs Borreliacidal antibody (GMT)a Dose (mg)

Before 2nd vaccination

2 wks after 2nd vaccination (challenge)

10 1 0.1 0 (control)

1:36 ,1:39 ,1:25 ,1:16

1:401 1:141 1:37 ,1:16

a

In all cases, titers before the first vaccination were ,1:16.

VOL. 65, 1997

PREDICTIVE ELISA FOR LYME DISEASE VACCINE

TABLE 8. GMTs for antibody levels in canine sera to OspA as measured by ELISA Antibody (GMT) to OspA (mg of LA-2/ml) Dose (mg)

10 1 0.1 0 (control)

887

TABLE 9. Classification of dogs and mice on the basis of antibody against an epitope on OspA as defined by monoclonal antibody LA-2 No. infected/total

Before 1st vaccination

Before 2nd vaccination

2 wks after 2nd vaccination (challenge)

Animal

0.045 0.046 0.054 0.048

0.281 0.099 0.080 0.066

1.977 0.496 0.163 0.045

LA-2 titer $0.32 mg eq/ml

LA-2 titer #0.32 mg eq/ml

Dogs Mice

0/13 0/44

20/22 28/31

both mice and dogs, responses to challenge were determined by direct culture. In dogs, immunoblot results were compared with direct culture as an indicator of infection. These data were compared against antibody levels to determine if a correlation existed between LA-2 titer and immune protection. Uncontrollable factors included variable responses for individual animals, the number of spirochetes transmitted during feeding, and mechanisms for immune killing other than antibody against OspA. Consequently, it was important that each group contain a sufficient number of animals with an adequate number of ticks to ensure that each susceptible dog or mouse would become infected. On the basis of indicators of infection (skin culture and immunoblot), all nonvaccinated dogs and mice became infected. Accordingly, we were then able to observe the outcome of challenge in the presence of low levels of circulating antibody. We computed the product moment correlation between LA-2 ELISA titers and OspA titers. This correlation was 0.61 with a P value of 0.002. When we ran this analysis, we included both the prechallenge OspA titers and LA-2 titers for prediction of protection. Our analyses demonstrated that the prechallenge OspA titers did not improve the fit, as both values tended to demonstrate the same relationship between serology and protection. The LA-2 ELISA provided a continuum of data, much more compatible with statistical analysis. We wanted a conservative estimate of a protective value by this assay and selected an 80% probability of protection as a cutoff value. Under this standard, we expected errors to result in the case of animals not becoming infected when their LA-2 titers were less than the cutoff. This expectation was realized experimentally. The design of the LA-2 ELISA allowed us to make a preliminary comparison of the effects of antigen and adjuvant on

FIG. 1. GMTs of antibody to outer surface protein A (OspA) by LA-2 ELISA throughout vaccination and challenge for dogs that received 10 mg ({), 1 mg (h), or 0.1 mg (Ç) of NS1-OspA and for the nonvaccinated control group (E).

immune responses of dogs and mice. In dogs, immune responses induced by NS1-OspA and OspA did not differ appreciably. The effects of adjuvant, however, were marked. Dogs vaccinated with NS1-OspA, but not OspA, in the absence of adjuvant did not form measurable levels of antibody to the LA-2 epitope (data not shown). Adjuvant also influenced responses of mice. Mice vaccinated with OspA without adjuvant responded more poorly than mice inoculated with OspA adsorbed to aluminum hydroxide gel (Table 1). The responses of inbred mice to the LA-2 epitope were up to eightfold greater than those of outbred mice (Table 1). Overall, the responses to OspA exceeded the responses to NS1-OspA except for BALB/c mice. From these preliminary results, we conclude that additional studies are required to optimize the ratio of antigen to adjuvant. The protocol used in these challenge experiments relied on natural tick transmission of infection. We believed that it was extremely important to analyze efficacy of this vaccine in a manner similar to how animals will be exposed in the field. Although we compromised control of the challenge dose in this system, we incorporated vitally important effects of tick transmission and the related aspects of a tick-borne disease. Further, tick transmission has been closely studied and reported for both mice and dogs (1, 8, 11, 12, 14, 16, 24) and is the preferred method for experimental infection when resources are available. Prior to challenge of dogs, immunofluorescent staining of dissected tick gut smears from a representative group of ticks demonstrated that 44% harbored B. burgdorferi. On that basis, each dog received six to seven infected female ticks, and approximately 90% of these fed (Table 4). Following removal of

FIG. 2. Serum antibody response to experimental vaccine. Dogs received experimental product formulated to contain 10, 1, or 0.1 mg of NS1-OspA or were nonvaccinated (controls). Shown are serum antibody levels for dogs that did not become infected following challenge (■) and for dogs that became infected (h) and the GMT for each treatment group (F). An 80% PPP corresponded to an antibody titer against an epitope on OspA defined by LA-2 of 0.3196 mg eq/ml.

888

GOLDE ET AL.

female ticks from each dog, direct culture demonstrated the aftereffects of feeding on a host having antibody to OspA. Several factors influenced the efficiency of tick culture. Although isolation media contained antibiotics, a number of cultures became overgrown with normal flora, presumably from the tick gut. Second, some ticks that fed to repletion detached from the host and succumbed to desiccation, making culture impossible. Lastly, spirochete levels in ticks are not at their highest immediately after feeding (6). Despite these complicating factors, we successfully performed a direct culture on the vast majority of ticks. The percentage of infected ticks gathered from each group of dogs following challenge appeared to depend on the dose of experimental vaccine that each group received (Table 4). Factors influencing direct culture of replete or partially replete ticks most likely accounted for a detectable infection rate of 17.3%, which was lower than the infection rate (44%) measured prior to feeding. Several investigators have reported that serum and serum components, ingested by ticks, promote killing before the spirochete can be transmitted to the host (2, 10–13, 23, 27). Further, several investigators report that spirochetes do not migrate from the tick to the host for at least 36 to 48 h after the tick begins a blood meal (16–18, 22). These early stages of feeding provide sufficient opportunity for blood components to kill spirochetes. Results from borreliacidal assays (Table 7) established that sera from dogs vaccinated with NS1-OspA contained bactericidal antibody. Collectively, the results of this study corroborate the finding of others and provide strong evidence for spirochete killing before transmission to the host. Serum from every dog that provided a positive skin culture eventually reacted positively on immunoblots (Tables 5 and 6). Typically, positive skin cultures preceded an immunoblot response by approximately 30 days. It is unlikely that numbers of organisms transmitted to susceptible dogs would be sufficient to elicit the formation of antibody without replication. Consequently, we submit that the antibody response observed by immunoblotting resulted from multiplication and dissemination of the spirochete. Two dogs (401 and 402) did not seroconvert by immunoblotting until 5.5 months after challenge. At challenge, both dogs had antibody titers (LA-2 ELISA) appreciably above the levels of cohort nonvaccinated controls. The levels of antibody in these two dogs, however, were lower than that associated with protection. The delayed response in these two dogs indicated that low levels of antibody either retarded spirochete multiplication or reduced the numbers of spirochetes transmitted. The major element of the study design was to determine if a correlation existed between antibody against OspA (LA-2 ELISA) and protection against infection. The ELISA design exploited an interaction between a borreliacidal antibody designated LA-2 (20, 21) and the LA-2 binding site on OspA. The test measured competition between antibodies in unknown samples and LA-2 for binding to OspA. By measuring antibodies against a defined site, we hoped to measure only those relevant to a protective immune response. We have previously published a study of mice that used a number of formulations of OspA as the vaccine. Sera from that study provided a sufficient number of samples to perform a comprehensive analysis for a correlation between LA-2 titer and protection against tick-transmitted infection. Once established, we could then apply this model to studies using other animals, including dogs. By vaccinating dogs with a broad range of NS1-OspA concentrations, we also hoped to induce antibody responses above and below the threshold of protection. The response of individual dogs, as shown in Fig. 2, indicates that this study objective was achieved. On the basis of

INFECT. IMMUN.

logistics regression analysis, an LA-2 titer of 0.3196 mg eq/ml corresponded to a PPP of 80%. Retrospectively, we classified dogs on the basis of an LA-2 titer of 0.32 mg eq/ml. The 80% PPP correctly classified all uninfected dogs and mice. Conversely, the cutoff correctly classified 20 of 22 and 28 of 31 infected dogs and mice, respectively. Serum antibody levels for two dogs and three mice that did not become infected were below the 80% PPP but well above values for negative serum samples. These results indicated that some animals having antibody levels above a negative response could be protected. Another measure of the LA-2 ELISA was a correlation with an in vitro bacterial killing assay. ELISA offers the advantage of providing a continuum of data and fewer variables than the borreliacidal assay. By contrast, killing assay results are not continuous and are reported as the reciprocal of a twofold dilution. Other variables that distinguish the killing assay from ELISA may include the passage of the organism and activity of complement. Despite these variables, we calculated a correlation between the LA-2 ELISA and borreliacidal assay (r 5 0.6733). We believe that this correlation demonstrates a close agreement for these two tests and suggests that LA-2-like antibodies mediate bacterial killing. The LA-2 assay is predictive of protection as described here under conditions in which relatively high levels of total antibody to OspA did not appear to predict protection. These results do not exclude the existence of other OspA epitopes that correlate to a protective immune response. We screened several monoclonal antibodies for bactericidal activity. Only those that reacted with the LA-2 epitope kill B. burgdorferi sensu stricto. It should be noted, however, that immunization with OspA from one genospecies rarely protects against another (11, 12) and that LA-2 is specific for B. burgdorferi sensu stricto (15). Accordingly, we believe that the results presented here establish an excellent correlation between antibody to OspA, as measured by the LA-2 ELISA, and immune status at the time of exposure to B. burgdorferi-infected ticks. ACKNOWLEDGMENTS We acknowledge and thank Becki Sandell, Robert Lohse, and Xioxu Jiang for excellent technical support. We also thank Robert Fanton for superb veterinary medical advice and David Keller for statistical analysis of the data. REFERENCES 1. Appel, M. J., S. Allan, R. H. Jacobson, T. L. Lauderdale, Y. F. Chang, S. J. Shin, J. W. Thomford, R. J. Todhunter, and B. A. Summers. 1993. Experimental Lyme disease in dogs produces arthritis and persistent infection. J. Infect. Dis. 167:651–664. 2. Aydintug, M. K., Y. Gu, and M. Philipp. 1994. Borrelia burgdorferi antigens that are targeted by antibody-dependent, complement-mediated killing in the rhesus monkey. Infect. Immun. 62:4929–4937. 3. Barbour, A. G., S. L. Tessier, and W. J. Todd. 1983. Lyme disease spirochetes and ixodid tick spirochetes share a common surface antigenic determinant defined by a monoclonal antibody. Infect. Immun. 41:795–804. 4. Barthold, S. W., E. Fikrig, L. K. Bockenstedt, and D. H. Persing. 1995. Circumvention of outer surface protein A by host-adapted Borrelia burgdorferi. Infect. Immun. 63:2255–2261. 5. Bockenstedt, L. K., E. Fikrig, S. W. Barthold, F. S. Kantor, and R. A. Flavell. 1993. Inability of truncated recombinant Osp A proteins to elicit protective immunity to Borrelia burgdorferi in mice. J. Immunol. 151:900–906. 6. Burkot, T. R., R. A. Wirtz, B. Luft, and J. Piesman. 1993. An OspA antigencapture enzyme linked immunosorbent assay for detecting North American isolates of Borrelia burgdorferi in larval and nymphal Ixodes dammini. J. Clin. Microbiol. 31:272–278. 7. Dressler, F., J. A. Whalen, B. N. Reinhardt, and A. C. Steere. 1993. Western blotting in the serodiagnosis of Lyme disease. J. Infect. Dis. 167:392–400. 8. Falco, R. C., and D. Fish. 1988. Prevalence of Ixodes dammini near the homes of Lyme disease patients in Westchester County, New York. Am. J. Epidemiol. 127:826–830. 9. Fikrig, E., S. W. Barthold, F. S. Kantor, and R. A. Flavell. 1990. Protection of mice against the Lyme disease agent by immunizing with recombinant OspA. Science 250:553–556.

VOL. 65, 1997

PREDICTIVE ELISA FOR LYME DISEASE VACCINE

10. Fikrig, E., S. R. Telford, S. W. Barthold, F. S. Kantor, and A. Spielman. 1992. Elimination of Borrelia burgdorferi from vector ticks feeding on OspAimmunized mice. Proc. Natl. Acad. Sci. USA 89:5418–5421. 11. Fikrig, E., S. R. Telford, R. Wallich, M. Chen, Y. Lobet, F. R. Matuschka, R. B. Kimsey, F. S. Kantor, S. W. Barthold, A. Spielman, and R. A. Flavell. 1995. Vaccination against Lyme disease caused by diverse Borrelia burgdorferi. J. Exp. Med. 181:215–221. 12. Gern, L., O. Rais, C. Capiau, P. Hauser, Y. Lobet, E. Simoen, P. Voet, and J. Petre. 1994. Immunization of mice by recombinant OspA preparations and protection against Borrelia burgdorferi infection induced by Ixodes ricinus tick bites. Immunol. Lett. 39:249–258. 13. Golde, W. T., T. R. Burkot, J. Piesman, M. C. Dolan, C. Capiau, P. Hauser, G. Dequesne, and Y. Lobet. 1994. The Lyme disease vaccine candidate outer surface protein (OspA) in a formulation compatible with human use protects mice against natural tick transmission of B. burgdorferi. Vaccine 13:435–441. 14. Johnson, B. J. B., S. L. Sviat, C. M. Happ, J. J. Dunn, J. C. Frantz, L. W. Mayer, and J. Piesman. 1995. Incomplete protection of hamsters vaccinated with unlipidated OspA from Borrelia burgdorferi infection is associated with low levels of antibody to an epitope defined by Mab LA-2. Vaccine 13:1086– 1094. 15. Kramer, M. D., U. E. Schaible, R. Wallich, S. E. Moter, D. Petzoldt, and M. M. Simon. 1990. Characterization of Borrelia burgdorferi associated antigens by monoclonal antibodies. Immunobiology 181:357–366. 16. Piesman, J. 1993. Dynamics of Borrelia burgdorferi transmission by nymphal Ixodes dammini ticks. J. Infect. Dis. 167:1082–1085. 17. Piesman, J., T. N. Mather, R. J. Sinsky, and A. Spielman. 1987. Duration of tick attachment and Borrelia burgdorferi transmission. J. Clin. Microbiol. 25:557–558. 18. Piesman, J., G. O. Maupin, E. G. Campos, and C. M. Happ. 1991. Duration of adult female Ixodes dammini attachment and transmission of Borrelia burgdorferi, with description of a needle aspiration isolation method. J. Infect. Dis. 163:895–897. 19. Plikaytis, B. D., P. F. Holder, L. B. Pais, S. E. Maslanka, L. L. Gheesling, and G. M. Carlone. 1994. Determination of parallelism and nonparallelism in bioassay dilution curves. J. Clin. Microbiol. 32:2441–2447.

20. Schaible, U. E., M. D. Kramer, K. Eichmann, M. Modolell, C. Museteanu, and M. M. Simon. 1990. Monoclonal antibodies specific for the outer membrane surface protein A (OspA) of Borrelia burgdorferi prevent Lyme borreliosis in severe combined immuno deficiency (scid) mice. Proc. Natl. Acad. Sci. USA 87:3786–3772. 21. Schaible, U. E., R. Wallich, M. D. Kramer, L. Gern, J. F. Anderson, C. Museteanu, and M. M. Simon. 1993. Immune sera to individual Borrelia burgdorferi isolates or recombinant OspA thereof protect SCID mice against infection with homologous strains but only partially or not at all against those of different OspA/OspB genotype. Vaccine 11:1049–1054. 22. Shih, C.-M., R. J. Pollack, S. R. Telford, and A. Spielman. 1992. Delayed dissemination of Lyme disease spirochetes from the site of deposition in the skin of mice. J. Infect. Dis. 166:827–831. 23. Shih, C.-M., R. Spielman, and S. R. Telford. 1995. Short report: mode of action of protective immunity to Lyme disease spirochetes. Am. J. Trop. Med. Hyg. 52:72–74. 24. Simon, M. M., U. E. Schaible, M. D. Kramer, C. Eckerskorn, C. Museteanu, H. K. Muller-Hermelink, and R. Wallich. 1991. Recombinant outer surface protein A from Borrelia burgdorferi induces antibodies protective against spirochetal infection in mice. J. Infect. Dis. 164:123–132. 25. Stover, C. K., G. P. Bansal, M. S. Hanson, J. E. Burlein, S. R. Palaszynski, J. F. Young, S. Koenig, D. B. Young, A. Sadziene, and A. G. Barbour. 1993. Protective immunity elicited by recombinant bacille Calmette-Guerin (BCG) expressing outer surface protein A (OspA) lipoprotein: a candidate Lyme disease vaccine. J. Exp. Med. 178:197–209. 26. Telford, S. R., E. Fikrig, S. W. Barthold, L. R. Brunet, A. Spielman, and R. A. Flavell. 1993. Protection against antigenically variable Borrelia burgdorferi conferred by recombinant vaccines. J. Exp. Med. 178:755–758. 27. Wang, H., and P. A. Nuttall. 1994. Excretion of host immunoglobulin in tick saliva and detection of IgG-binding proteins in tick haemolymph and salivary glands. Parasitology 109:525–530. 28. Wilske, B., A. G. Barbour, S. Bergstrom, N. Burman, B. I. Restrepo, P. A. Rosa, T. Schwan, E. Soutschek, and R. Wallich. 1992. Antigenic variation and strain heterogeneity in Borrelia spp. Res. Microbiol. 143:583–596.

Editor: J. R. McGhee

889