The outer surface protein A (OspA) vaccine against Lyme disease: efficacy in the rhesus monkey Vaccine 1997 Volume 15 Number 17/18, pages 1872-1887. Mario T. Philipp*¦¦, Yves Lobet†, Rudolf P. Bohm Jr.*, E. Donald Roberts*, Vida A. Dennis*, Yan Gu*, Robert C. Lowrie Jr.*, Pierre Desmons†, Paul H. Duray‡, John D. England§, Pierre Hauser†, Joseph Piesman¶ and Keyu Xu* *Tulane University Medical Center, Tulane Regional Primate Research Center, Covington, LA, USA. †SmithKline Beecham Biologicals, Rixensart, Belgium. ‡National Cancer Institute, National Institutes of Health, Bethesda, MD, USA. §Department of Neurology, Louisiana State University, New Orleans, LA, USA. ¶Division of Vector-Borne Infectious Diseases, Centers for Disease Control and Prevention, Fort Collins, CO, USA. ¦¦To whom correspondence should be addressed. Tel.: (504) 892-2040; fax: (504) 893-1352; e-mail: philipp@tpc.tulane.edu. (Received 19 February 1997; revised version received 12 April 1997; accepted 6 May 1997) The efficacy of an outer surface protein A (OspA) vaccine in three different formulations was investigated in the rhesus monkey. The challenge infection was administered using Ixodes scapularis ticks that were infected with the B31 strain of Borrelia burgdorferi. Protection was assessed against both infection and disease, by a variety of procedures. Some of the animals were radically immune suppressed, as an attempt to reveal any putative low level infection in the vaccinated animals. The significant difference found between the spirochaetal infection rates of ticks that had fed on vaccinated vs. control monkeys, lack of seroconversion in the vaccinated animals, and the absence of spirochaetal DNA in the skin of vaccinated animals in the weeks following the challenge, indicate that vaccinated monkeys were protected against tick challenge. The post-mortem immunohistochemical and polymerase chain reaction analyses, however, suggest that these monkeys may have undergone a low-level infection that was transient. 1997 Elsevier Science Ltd. Keywords: vaccine efficacy; OspA; Borrelia burgdorferi; Lyme disease Recombinant outer surface protein A (OspA) is currently the most developed, molecularly-defined vaccine candidate to prevent Lyme borreliosis. This disease, which is caused by the spirochaete Borrelia burgdorferi and is transmitted to humans by ticks, continues to be the most frequently-reported arthropod-borne malady in the United States and Europe (1,2). An infection with B. burgdorferi may affect multiple organs, including the dermal, musculoskeletal, ocular, reticulo- endothelial, cardiac or nervous systems, and although timely administration of appropriate antibiotics is usually curative, long courses of therapy may be required if the infection is allowed to become chronic, and in some patients there is no response to therapy at all (3). This occasional refractoriness to treatment substantiates the need for immunoprophylactic strategies. The safety and immunogenicity of recombinant OspA vaccines have been evaluated in both naive and B. burgdorferi-infected rhesus monkeys (4), as well as in human subjects without (5,6) and with (7) a previous history of Lyme disease but with no evidence of active infection. These evaluations have consistently indicated that OspA-based vaccines are safe. Proof of the 'principle' that OspA is a protective antigen was achieved with both passive and active immunisation experiments in rodents (8,9) (see reference (10) for a review). More recently, vaccine formulations that are compatible with human use were also shown to be efficacious in mice (11,13). In addition, phase III human trails are underway (14). The present study of efficacy of OspA vaccines in the rhesus monkey was undertaken to ascertain if different vaccine formulations, including one currently used in the phase III human trials, were able, at the very least, to induce protection against disease and, at best, to elicit sterile immunity in the rhesus monkey. The animal model was chosen because it was shown previously that rhesus monkeys infected with B. burgdorferi develop disease signs that mimic both the acute and chronic phases of Lyme disease in humans. These signs include erythema migrans, arthritis, myocarditis and neuroborreliosis (15-19). At the serological level, too, the time course and specificity spectrum of the antibody response to B. burgdorferi in rhesus monkeys have been shown to be similar to their counterparts in humans (15). Thus, by choosing the rhesus monkey for this study we sought to retain the case with which evidence of infection may be gathered in an animal model, without relinquishing the possibility of assessing protection against disease manifestations that most closely resemble human Lyme disease. EXPERIMENTAL DESIGN Distribution of information Personnel responsible for the clinical examinations, sample collection, vaccine administration, and assays described below were blinded with respect to which animal had received a vaccine or a placebo injection. Study population and animal examination schedule The study population consisted of 16 male 2-3 year old Chinese Macaca mulatta (rhesus). It was divided into four groups of four animals each, according to the vaccine formulation received. Animals were examined bi-weekly during the 12-week period before the challenge infection, and weekly thereafter. Vaccine formulations and vaccine administration protocol Vaccine formulations and the vaccine administration protocol were as described previously (4). Animals L458, L594, L971 and M585 received NS1-OspA/Al(OH)3. The NS1-OspA fusion protein is composed of recombinant OspA (cloned from the B. burgdorferi sensu stricto isolate ZS7) that lacks the amino-terminal cysteine and is fused to a fragment of 81 N-terminal amino acids from the nonstructural influenza virus protein NS1. Animals L476, L712, L642 and M106 received NS1-OpsA/4'- monophosphoryl lipid A (MPL)/Al(OH)3; animals M243, L537, M219 and M107 received lipidated OspA/Al(OH)3, a formulation that was used also in the Phase III human trial (14); animals L457, L549, M021, and M581 received only Al(OH)3. Inoculation with B. burgdorferi All animals were inoculated with the B31 strain of B. burgdorferi by the natural route of tick bite using infected Ixodes scapularis nymphal ticks. Assessment of vaccine efficacy Detection of spirochaetes by direct immunofluorescence (DFA) in ticks after infection. Since immunity to B. burgdorferi elicited by OspA appears to be effected chiefly within the tick (20), ticks used for the challenge infection were analysed for the presence of residual spirochaetes by DFA. Engorged ticks that had fed on each animal were divided into two equal groups. Ticks in one of the groups were dissected 4-5 days post-engorgement and ticks in the second group were dissected 3 weeks post-engorgement. Western blot analysis of serum samples collected before and after challenge infection. Seroconversion of the animals was investigated by Western blot using detergent extracts of whole B. burgdorferi as antigen. Serum samples were obtained from all of the animals longitudinally throughout the study period. In virto culture of skin biopsy samples. Skin biopsy samples were collected from each animal before the administration of the first vaccine dose and 1-4 weeks after the challenge infection (post- challenge, PC) and were cultured for 8 weeks. Polymerase chain reaction (PCR) applied to skin samples. PCR of skin biopsies was performed on samples collected as above. Histology and immunohistochemistry of skin biopsy samples. Skin biopsies collected as above were processed for histology and immunohistochemistry. Immune suppression The purpose of the immune suppression (IS) experiment was to assess whether a putative low-level B. burgdorferi infection would be rendered apparent by radically immune suppressing a subgroup of vaccinated monkeys. Five of the vaccinated animals were involved in the experiment. Three had been vaccinated with NS1-OspA/Al(OH)3 (L594, M585, L971) and the other two (L712, M106) with NS1-OspA/Al(OH)3/MPL. Four animals were immune suppressed and one (L971) was left as control and not immune suppressed. IS began at 50 weeks PC and continued for 6 weeks, at which point the animals were sacrificed. IS was assessed longitudinally by performing complete blood cell counts (CBC) and peripheral blood mononuclear cell (PBMC) counts, and by measuring blastogenic responses of PBMCs to concanavalin A (Con A), pokeweed mitogen (PWM), and whole killed B31 B. burgdorferi spirochaetes. The well-being of the animals during IS was assessed by a weekly clinical examination, and through the results of CBC, urinalysis and serum chemistries performed either weekly or bi-weekly. Electrocardiograms (EKG) were also performed, 1 week and 1 month after IS had begun. Possible reactivation of infection subsequent to IS was assessed as above, by in vitro culture, PCR, histology and immunohistochemistry of skin biopsy samples, postmortem analyses and, in addition, in vitro cultivation of urine and bronchoalveolar lavage samples, nerve conduction studies (NCS) and xenodiagnosis. Post-mortem analyses All of the animals from the group that received Al(OH)3, three of the animals that were vaccinated with lipidated OspA/Al(OH)3 and one of the animals that were given NS1-OspA/MPL/Al(OH)3 were sacrificed between week 30 and 40 PC. Animal M219 of the 'lipidated OspA' group and animal L458 of the 'NS1-OspA/Al(OH)3 group died unexpectedly while under anaesthesia, by 2 weeks and 19 weeks PC, respectively. The remaining five animals, which participated in the IS experiment, were sacrificed 56 weeks PC. The following evaluations were performed postmortem. (1) Assessment of gross pathologic changes. This was done in all animals in the following organ systems: thyroids and parathyroids, respiratory system, heart and vessels, spleen, liver, biliary system, pancreas, musculo-skeletal system, gastrointestinal tract, adrenal glands, urinary system, genital system, peripheral lymph nodes, and peripheral and central nervous systems. (2) Assessment of histopathological changes and immunohistochemical evidence of present or past infection with B. burgdorferi-specific monoclonal antibodies. These were assessed in samples taken from skin, synovium, heart, brain, central and peripheral nerves, lungs, kidney, ureter and bladder. (3) In vitro culture and PCR with primers hybridising to a B. burgdorferi chromosomal DNA fragment. Both were performed on a similar set of tissue samples as in (2). (4) Morphometric quantification of cells that bind the anti-Ospa mAb LA31 in heart and lungs. The four control animals and six of the vaccinated animals were involved in this study. (5) Silver-staining (for spirochaetes) of sections from the brain. Brain sections from all of the control monkeys, seven of the vaccinated monkeys and three monkeys that were not challenged with B. burgdorferi were employed in this study. MATERIALS AND METHODS Animal care and housing Animals were cared for and housed as described previously (15). Challenge infection and xenodiagnosis Ixodes scapularis nymphs infected with the B31 strain of B. burgdorferi were used to deliver a challenge infection. For this purpose, ten ticks were placed in each of two capsules. The capsules were already attached to both scapular areas of each monkey by a procedure described previously (15). For xenodiagnosis; performed 4 weeks after the beginning of the IS protocol, 20 uninfected nymphs were enclosed in each capsule and both capsules were placed on each of the five animals involved in the immune suppression experiment. Ticks were left on the animals for 5 days, both for the challenge infection and the xenodiagnosis. Xenodiagnosis ticks were analysed by DFA 3 weeks after removal from the capsule. As a positive control for the xenodiaganosis procedure, a rhesus macaque was infected with B. burgdorferi and subsequently xenodiagnosed as above. Detection of spirochaetes in ticks by DFA DFA was performed as described previously (21). Negative preparations were scanned completely under the fluorescent microscope; DFA-positive specimens were scanned only until spirochaetes were observed. Western blotting The procedure used was described previously (22) but B. burgdorferi strain B31 (4th passage) was used instead of strain JD1. Sodium dodecyl sulphate (SDS) extracts of whole B. burgdorferi B31 were electrophoresed on SDS polyacrylamide gels (15% acrylamide). Serum samples from all animals were incubated with the nitrocellulose blot at a final concentration of 1:50. Antibodies that bind to OspA also appear to bind to other components of greater and lesser molecular mass, which span the whole length of the gel (4). To avoid these confusing bands, serum samples were preincubated with just enough recombinant non-lipidated OspA to permit the detection of OspA as a 31 kDa band (166 µg per ml OspA for serum samples from animals vaccinated with lipidated OspA, and 33 µg per ml for all other vaccinated animals). To verify that the preincubation of serum with OspA did not inhibit binding of antibody to any other B. burgdorferi antigens, blots of whole B. burgdorferi antigens were incubated with serum obtained from control animals at 15 week PC with and without added OspA (166 µg per ml). No differences were observed in the antigen patterns obtained with and without added OspA (not shown). PCR Ante-mortem biopsy samples were collected aseptically. Necropsy instruments such as forceps, scissors and blades were flamed before being used to collect samples. Skin was sterilised with 70% ethanol. Samples from different tissues/organs were kept separately in sterile cryogenic vials with 0.5 ml TE buffer (10 mM Tris-HCl, 0.1 mM EDTA, pH7.4) in each vial. All samples were kept at -70 degrees C until use. Before performing the PCR, samples were thawed and a sterile needle was used to move the sample out from the vial to a weighing boat, to be cut into approximately 20 mg pieces with a sterile blade. Blade, needle and weighing boat were discarded after processing each sample. Each 20mg piece was processed for PCR as described previously (23). The readout of this protocol is a Southern blot. Hence, whenever reference is made in the text to a positive PCR reaction, a positive Southern blot is implied and amplicons rated as positive are therefore specific for the amplified DNA fragment. From each of the skin biopsy samples collected before the challenge infection and on weeks 1-4 PC, only one aliquot of lysate was amplified. From each tissue lysate of samples collected post-mortem from animals that were part of the immune suppression experiment, two aliquots were amplified, and three from all of the tissue lysates of tissues collected post-mortem from all of the remaining animals. Between four and seven negative control tubes (containing no tissue lysate) were interspersed and processed in parallel with each of ten tubes with lysate subjected to the PCR. In vitro cultivation of skin biopsy samples The procedure employed was described previously (15). However, two changes were introduced, namely, BSK-H (Sigma, Chemical Co., St. Louis, Missouri) was used instead of BSK-II, and tissue samples were incubated at a larger ratio of tissue to medium volumes. Thus, whereas in previous experiments, samples were incubated at a ratio of 1:1000-1:3000 (15), in the present experiments this ratio was typically 1:100. Histopathological and immunohistochemical analyses of skin biopsy samples and organ tissues collected post-mortem Skin punch biopsies (8mm in diameter) were taken from sites adjacent to the tick feeding areas and fixed in Streck Tissue Fixative, a non- formalin fixative, for 1 week, according to the manufacturer's instructions (Streck Tissue Fixative, Omaha, NB [sic]) and then processed using a standardised paraffin embedding technique. Tissues were sectioned at 3-6 µm thickness. For immunohistochemistry, the primary antibodies used were either a monoclonal antibody (mAb) directed to a B. burgdorferi 7.5 kDa lipoprotein (24) or an anti-OspA mAb (LA31) (25). Bound antibody was revealed with anti-mouse IgG labelled with alkaline phosphatase, and fast-red as chromogen (Kirkegaard & Perry Laboratories, Gaithersburg, Maryland). Samples of organs collected post-mortem were processed in the same way. Morphometric quantification of cells that bind the anti-OspA mAb LA31 Tissue. The areas of full cross-sections of both right and left sections of the heart, septum and each of the lung lobes were measured. The variations in areas measured reflect differences in tissue size. All tissue sections measured were 3-4 µm thick paraffin sections. They were immunostained as above, using the LA31 anti-OspA mAb and cells that stained red were counted. Controls. A pellet of cultured B. burgdorferi (B31) was fixed by the same procedure employed for the tissue sections and used as a positive control. Negative controls were: (1) a section of the heart or lungs from an uninfected animal, incubated with the LA31 mAb; (2) a section of the heart or lungs from the animal being studied, incubated with the second antibody but not the LA31 mAb; (3) a 'positive control' slide incubated with the second antibody but not the LA31 mAb. No cells were stained on negative control slides. Measurements. All measurements were made using a computer-based morphometric device (Bioscan, Optimas Corporation, Edmonds, WA). The system was calibrated with a stage micrometer. Area was determined using a 1.2x objective, and the numbers of stained cells were determined with a 10x and/or a 20x objective. Silver staining of brain sections Brain samples collected post-mortem and fixed as described above (histopathology section) were processed using a modified form of the original Steiner silver-staining method. The modified Steiner procedure has been described elsewhere (26). Immune suppression protocol The IS protocol consisted of a combination of surgical splenectomy and administration of cyclophosphamide (CP) and prednisone sodium phosphate (PSP). Three doses had been considered: low (CP 2.6 mg per kg; PSP 4.1 mg per kg), medium (CP 6.5 mg per kg; PSP 8.3 mg per kg), and high (CP 10.6 mg per kg; PSP 12 mg per kg). The high dose was chosen after assessing the level of IS elicited in three rhesus monkeys, each of which received a low, medium or high dose. Administration of both drugs began 1 week after splenectomy and continued for a total of 5 weeks. CP was given every other day by orogastric intubation while the animal was anaesthetised. PSP was given by intramuscular injection once every day. Bronchoalveolar lavage and urine collection in immune suppressed animals Bronchoalveolar lavage and urine sample collection began 1 week after the first dose of immune suppressive drugs and continued on a weekly basis until sacrifice. For bronchoalveolar lavage, animals were preanaesthetised with glycopyrolate and acepromazine and anaesthesia induced with ketamine HCl. A fibreoptic bronchoscope was manipulated into the trachea after applying local anaesthetic (lidocaine) to the larynx. After directing the scope into the right primary bronchus, a volume of 20 ml of warmed normal saline was instilled through the bronchoscope. The saline was then collected by aspiration and the animal was given O2 by face mask for 5 min before being returned to its cage. Urine was collected by cystocentesis. Nerve conduction studies Peripheral nerve conduction velocity and amplitude of the response to stimulation was determined in both motor and sensory ulnar and median nerves and the tibial motor nerve (one side only), 1 week before and 1 month after initiation of the IS protocol. Possible influence of IS on nerve conduction properties was ruled out by performing NCS once before and twice after the initiation of drug administration on the three animals participating in the trail IS study. The method employed for NCS was described previously (19). Statistical analyses The frequency of appearance of PCR amplicons in organs collected post-mortem from control and vaccinated animals was compared by a nonparametric X˛ [chi-square] test. The mean numbers of monocyte-type cells that stained with the anti-OspA mAb LA31 in the heart and lungs of control and vaccinated animals were compared using Kruskal-Wallis ANOVA by ranks or MANOVA, as appropriate. RESULTS Detection of spirochaetes by DFA in ticks, after infection Of the 320 infective nymphs placed on the 16 animals, 188 (59%) were either fully (132) or partially (56) engorged with blood. The number that fed on each monkey ranged from two (animal M219) to 18 (animal L971), the mean being 11.8 (Table 1). The rationale for dissecting half of the ticks 4-5 days after their removal from the capsules and the other half 3 weeks after removal entailed the assumption that, if there were quantitative differences in the spirochaete load of ticks that had fed on control and on vaccinated animals, these differences could be detected early, but not late after tick capsule removal, since appreciable differences in spirochaete numbers would no longer exist after 3 weeks, due to spirochaetal multiplication. Table 1 Detection of spirochaetes by direct immunofluorescence in the infected ticks that were used for the challenge infection -------------------------------------------------------------- Vaccine/ Total no. of animal no. DFA+ ticks DFA- ticks ticks engorged(a) -------------------------------------------------------------- Al(OH)3 L457 10 1 11 L549 16 - 16 M021 6 1 9 M581 16 - 17 Lipidated OspA M243 - 5 5 L537 - 12 12 M219 - 2 2 M107 - 9 9 NS1-OspA L458 - 11 12 L594 - 11 11 L971 - 18 18 M585 1 8 11 NS1-OspA/MPL L476 - 11 12 L712 - 13 13 L642 - 15 16 M106 - 14 14 -------------------------------------------------------------- (a)Only live ticks were dissected and examined by DFA. On the other hand, if only very small numbers of spirochaetes were detectable in any tick 4-5 days after tick capsule removal, multiplication over a 3-week period would permit detection by DFA. For each of the animals, no differences were observed in the number of spirochaetes per tick in ticks dissected early or late, thus indicating that spirochaetes may not have multiplied appreciably in the days following engorgement. Hence, only the pooled results are shown for each animal in Table 1. All but two of the ticks that fed on animals that received placebo and that were dissected (43) harboured large numbers of spirochaetes. Conversely, all but one of the ticks that fed on OspA-vaccinated animals and that were dissected (121) harboured no detectable spirochaetes (Table 1). In one of the vaccinated animals (M585), eight blood engorged ticks were DFA negative, but one nymph was DFA positive. This animal was fed upon by 11 nymphs, two of which died before a DFA test could be performed. The spirochaete density on this one preparation was at least tenfold lower (about ten fluorescent spirochaetes on the entire slide) than the number generally observed in the contents of nymphs that fed on control monkeys (Figure 1). [Image] Figure 1 DFA analysis of ticks after the challenge infection. Direct immunofluorescence view of spirochaetes within ticks that fed on (a) a control animal and (b) a vaccinated animal (M585) Western blot analysis of serum samples For the purpose of vaccine efficacy analysis, Western blots were used primarily to investigate whether vaccinated animals had seroconverted to showing antibodies to B. burgdorferi antigens other than OspA. Western blots of whole antigen extracts of B. burgdorferi B31 were incubated with serum samples collected from all animals immediately before and at 4, 6, 9, 15, 27 and 40 or 45 weeks PC. Only the results obtained at weeks 40 or 45 PC are shown (Figure 2). The pattern of reactivity of all serum samples collected after the challenge infection was compared with the reactivity before challenge. Only in one instance, by week 15 PC, one animal, L594, showed one band that was not present before challenge infection. However, by week 45 PC this band had disappeared (Figure 2, tracks 4 and 5). In marked contrast with the essentially complete absence of detectable anti-B. burgdorferi antibodies in OspA- vaccinated animals, serum samples from the four animals that had received Al(OH)3 alone reacted with several antigens by weeks 6 and 15 PC, and with even more numerous bands by weeks 40-45 PC (not shown). Anti-P39 antibodies were detectable in all control animals. To better visualise the gradual increase in the number of bands recognised by serum from control animals over time, which is indirect evidence of an active infection in these animals, serum samples obtained at weeks 4, 8, 12, 16, 20, 24, 28, 33, 36 and 40 PC were compared on the same Western blot (Figure 3(A)). Bands were counted and plotted as a function of time PC (Figure 3(B)). The most pronounced increases in the number of antigens recognised were observed between weeks 4 and 8 PC. However, in some animals, the number of bands continued to increase after this time interval. [Image] Figure 2 IgG antibody responses to whole B. burgdorferi antigens in vaccinated animals before and after the challenge infection. Serum samples were obtained from each animal immediately before and several weeks after the challenge infection. For each animal, a pair of lanes of the Western blot was employed. The first lane of each pair was incubated with the pre-challenge serum and the second lane with the serum obtained post-challenge (PC). Animal L458, lanes 2 and 3; L594, lanes 6 and 7; L971 lanes, 12 and 13; M585, lanes 17 and 18; L476, lanes 4 and 5; L712, lanes 10 and 11; L642, lanes 8 and 9; M106, 15 and 16; L537, lanes 19 and 20; M107, lanes 21 and 22; M243, 23 and 24; M219, lanes 25 and 26. Monoclonal antibodies to P41, P39, OspB and OspA, lanes 1 and 14. Extracts of whole B. burgdorferi B31 were electrophoresed on 15% acrylamide gels. Serum samples from all animals were used at a final dilution of 1:50. L458, week 19 PC; L594, L971, M585, L476, L712, L642, M106, week 45 PC; L537, M107, M243, week 40 PC; M219, week 2 PC. [Image] Figure 3 (A) Longitudinal analysis of the IgG antibody responses of control animals to whole B. burgdorferi antigens. Animal L457, lanes 1- 9 sera from weeks 0 (pre-challenge), 4, 8, 12, 16, 20, 24, 28 and 33 PC, respectively; L549 lanes 10-18, sera from same weeks; M021 lanes 20-30, sera from same weeks, plus weeks 36 and 40 PC; M581, lanes 31-41, sera from same weeks as M021. Monoclonal antibodies to P41, P39, OspB and OspA, lanes 19 and 42. Extracts of whole B. burgdorferi B31 were electrophoresed on 15% acylamide gels. Serum samples from all animals were incubated with the nitrocellulose blot at a final concentration of 1:100. (B) Plot of the number of bands appearing on the Western blot of (A) as a function of time (in weeks) PC In vitro culture of skin biopsy samples Skin samples collected weekly from all animals between 1 and 4 weeks PC were cultured in BSK-H medium as described in the Methods section. Samples were kept for 8 weeks, and checked weekly for presence of B. burgdorferi by darkfield microscopy. No spirochaetes were found. These uniformly negative results were probably an artefact due to the low ratio of medium to tissue volumes, as explained in the Discussion section. Clinical examination of the skin, histopathology, immunohistochemistry and PCR analysis of skin biopsy samples Erythema migrans was not observed clinically in any animal. Skin biopsy samples were obtained, fixed, sectioned, and stained as described in Materials and Methods. Two of the control animals, L457 and M581, showed histopathological and immunohistochemical evidence of infection in skin samples taken at weeks 3 and 4 PC, respectively (Table 2). The dermatitis was scored as 3, which corresponds to a deep perivascular cellular infiltrate involving vascular structures at the panniculus- dermis junction. The infiltrates consisted of lymphocytes, macrophages and plasma cells. B. burgdorferi antigen was detected with the anti-7.5 kDa antibody in both skin biopsies. Three of the animals vaccinated with lipidated OspA (M243, M107 and M219) showed histopathological evidence of infection and M219 also showed reactivity with the anti-7.5 kDa mAb. One of the animals immunised with NS1-OspA/MPL (L712) showed both histopathological and immunohistochemical evidence of infection. The lesion observed in this animal was rated type 2, which corresponds to a dermatitis involving the superficial perivascular structures with an infiltrate of primarily macrophages and the occasional neutrophil. No lesions or evidence of B. burgdorferi antigen were observed in animals immunised with NS1-OspA (Table 2). All skin biopsies obtained before the beginning of the vaccination schedule were negative. Skin biopsy samples collected weekly both from vaccinated and control animals between weeks 1 and 4 PC were processed for PCR as described in Materials and Methods. The results are shown in Table 2. Only the control animals showed presence of detectable B. burgdorferi DNA. Tubes without template or with lysates of skin collected before the challenge infection did not show a positive signal on Southern blots. Table 2 PCR histopathological and B. burgdorferi-specific immunohistochemical evaluation of skin biopsy samples obtained 1-4 weeks after the challenge infection ------------------------------------------------------------------------ Week 1 Week 2 Week 3 Week 4 ------------ ------------ ------------ ------------ Vaccine/ PCR HP IH PCR HP IH PCR HP IH PCR P H animal no. (a) (b) (c) [sic] ------------------------------------------------------------------------ Al(OH)3 L457 - - - + - - + D3 + - - - L549 - - - - - - + - - - - - M021 - - - - - - - - - + - - M581 - - - + - - - - - + D3 + Lipidated OspA M243 - - - - D2 - - - - - - - L537 - - - - - - - - - - - - M107 - - - - - - - D2 - - - - M219 - - - - D2 + ND ND NS1-OspA L458 - - - - - - - - - - - - L594 - - - - - - - - - - - - L971 - - - - - - - - - - - - M585 - - - - - - - - - - - - NS1-OspA/MPL L476 - - - - - - - - - - - - L712 - - - - D3 + - - - - - - L642 - - - - - - - - - - - - M106 - - - - - - - - - - - - ------------------------------------------------------------------------ (a) Polymerase chain reaction: -, amplicon absent; +, amplicon present. (b) Histopathology: -, normal skin; D2, superficial dermatitis; D3, deep dermatitis, severe. (c) Immunohistochemistry: positive (+) or negative (-), using mAb anti- 7.5 kDa lipoprotein, and alkaline phosphatase-labelled indicator antibodies. ND, not done. Immune suppression of vaccinated animals The number of PBMC decreased rapidly after the initiation of the drug administration protocol in all of the treated animals, and remained relatively low (<6 x 10 superscript(6) PBMC per ml) throughout the study period (Figure 4(A)). Proliferative responses of PBMC to heat-killed B. burgdorferi B31 cells and to Con A also decreased rapidly and uniformly for all suppressed animals (Figure 4(B) and Figure 4(C), respectively), whereas the response to PWM took longer to decrease and in one case (M106), remained virtually unaffected (Figure 4(D)). Post-mortem, the PBMC responses to B. burgdorferi, Con A and PWM remained uniformly depressed (not shown). In contrast, blastogenic responses of mesenteric and inguinal lymph node cells obtained post-mortem did not appear to be uniformly suppressed compared with the control animal that was not immune suppressed (L971). For example, animal M106 had a mesenteric cell response to Con A similar to that of L971 (Figure 5(A)). Responses to PWM appeared to be unaffected (Figure 5(B)). Time points labelled ND (not determined) correspond to those cases where it was not possible to collect a sufficient number of cells. [Image] Figure 4 Evolution of immune suppression as a function of time after the initiation of the immune suppression protocol. Immune suppressed animals: L594, L712, M106, M585. Control (vaccinated but not immune suppressed), L971. Number of PBMCs per ml of blood (A), and blastogenic responses to heat-killed B. burgdorferi (B31) cells (B), Con A (C) and PWM (D). NCPM, net counts per minute (counts per min incorporated by stimulated cells minus counts per min of unstimulated cells). Day 0 is the day the splenectomy was performed [Image] Figure 5 Immune suppression assessed post-mortem in PBMCs, and mesenteric and inguinal lymph node cells. Blastogenic responses to Con A (A) and PWM (B). NCPM, net counts per minute (counts per min incorporated by stimulated cells minus counts per min of unstimulated cells) Physical examinations, clinical laboratory analyses and EKG of immune suppressed animals No abnormal findings were noted throughout in physical examinations performed weekly during the 6-week IS procedure. All animals developed profound lymphopaenia within 2 days after the beginning of immune suppressive drug administration. Lymphopaenia persisted throughout the study in all animals. Absolute leukopaenia followed with onset at 14-27 days after initiation of therapy. Elevation of aspartate aminotransferase and alanine aminotransferase levels occurred in all animals. Onset ranged from day 8 to day 27 post-immunosuppression. These elevations usually occurred during one sampling date only. The same findings have been seen in other animals similarly immune suppressed and probably result from a drug effect. All electrocardiographic findings were within normal limits. Evidence of reactivation of a residual B. burgdorferi infection In vitro culture. Urine samples and cells collected from lung lavages were cultivated over an 8-week period. No B. burgdorferi spirochaetes were recovered. The rationale for culturing lung lavage specimens in the immune suppressed animals was arrived at after noticing that all of the controls and several of the vaccinated animals showed immunohistochemical evidence of B. burgdorferi antigens in the lungs (see below). Xenodiagnosis. An attempt was made to recover spirochaetes from immune suppressed animals by xenodiagnosis. Engorged nymphs were dissected 3 weeks after removal of the tick capsules and analysed by DFA. An average of 18 nymphs fed on each of the five animals participating in the immune suppression experiment. None contained spirochaetes that were detectable by DFA. In contrast, of the 18 ticks that engorged while feeding on the positive control animal, nine were dissected and all nine were DFA positive. Nerve conduction studies. The baseline values of nerve conduction remained unchanged in all animals throughout the study, indicating that if immune suppression had led to a reactivation of infection, this had not affected the peripheral nerves investigated. Histology and immunohistochemistry of skin biopsy samples. Skin biopsy samples were collected weekly after the initiation of the IS protocol. Two animals, L594 and L971 (IS control) showed a superficial dermatitis on one occasion. In these perivascular lymphocytic infiltrates there were macrophage-type cells that stained positive with the anti-7.5 kDa mAb. Skin biopsies of animals L712, M585 and M106 were normal at all sampling periods. Post-mortem analyses Gross pathology. Neither control nor vaccinated animals showed abnormalities attributable to the B. burgdorferi infection in any organ system at the gross (macroscopic) level. The only exception to this generalisation is perhaps the finding of a slight meningeal congestion in vaccinated animal M107. However, no immunohistochemical or PCR- derived evidence of infection of the central nervous system was obtained for this animal post-mortem (see below). All organ systems of animals participating in the immune suppression experiment were grossly normal, with exception of the lungs, which showed pneumonia presumably related either to a cytomegalovirus infection secondary to immune suppression or to the weekly lung lavages. Histopathology and immunohistochemistry. Histopathology and immunohistochemistry of sections stained with B. burgdorferi-specific mAbs were assessed in samples taken from the skin, lungs (anterior, middle and posterior lung lobes), urinary system (kidney, ureter and bladder), heart (left and right heart, and septum), spleen, nervous system (brain and central and peripheral nerves), and joints and synovial membranes (Table 3). In contrast with what had been observed in skin biopsy samples taken between weeks 1 and 4 PC, the skin of both vaccinated and control animals had no detectable type 3 dermatitis (deep perivascular cellular infiltrate at the panniculus-dermis junction) by the time the animals were sacrificed. Similarly, no spirochaetes or spirochaetal antigens were detectable by immunohistochemistry, except in samples from animal M219. It should be noted, however, that these samples were obtained 2 weeks PC, at the time this animal died by anaesthesia (Table 3). Sections of the anterior, middle and posterior lung lobes of all of the sacrificed/dead animals were explored for presence of histopathological changes and for immunohistochemical evidence of B. burgdorferi. Lymphocytic hyperplasias (LH) were prevalent both in control and vaccinated animals (Table 3). In addition, most of the animals with LH also showed immunohistochemical evidence of B. burgdorferi, including four of the vaccinated animals that were not immune suppressed and all of the controls. Among the immune suppressed animals, the only immunopositive lung samples were those of L971 and L594 (Table 3). However, all of these monkeys showed pneumonia, which was identified by intranulcear basophilic inclusion bodies. As mentioned above, pneumonia may have been caused by the weekly lung lavages or as a consequence of CMV infection secondary to IS. Two of the control animals, L457 and M021, showed lymphocytic hyperplasias in the ureter and kidney, respectively, that stained positive for B. burgdorferi (Table 3). In contrast, only one of the vaccinated animals (L971) had detectable pathologic changes in the urinary system. This animal had ureteritis, and the lesion contained cells that stained positive for the 7.5 kDa B. burgdorferi antigen (Table 3). Perivascular lymphocytic infiltrates of the renal pelvis were also observed in animal L971. Three of the four control animals and one of the vaccinated animals (M219) had myocarditis (Table 3). Perivascular lymphocytic infiltrates of the renal pelvis were also observed in animal L971. Three of the four control animals and one of the vaccinated animals (M219) had myocarditis (Table 3). Perivascular lymphocytic infiltrates were present, often with immunohistochemical evidence of the 7.5 kDa B. burgdorferi antigen. All heart sections were normal in all of the animals that participated in the immune suppression experiment. The spleen showed evidence of infection but only in the control animals, with macrophages staining positively with the anti-B. burgdorferi mAb. The spleens of animals that participated in the immune suppression experiment were not analysed (Table 3). Sections of joint and synovium samples from the hip, knee, tarsus, shoulder, elbow and carpus (left and right, in all cases) of all of the dead/sacrificed animals were assessed for histopathological changes and for immunohistochemical evidence of the presence of B. burgdorferi organisms (or their antigens). One control animal, M021, showed arthritis, which also showed positive staining for B. burgdorferi. Microscopic joint lesions of the synovial membrane were observed in two of the animals that participated in the immune suppression experiment (L971 and L594) (Table 3). These lesions consisted of synovial cell hyperplasia with perivascular lymphocyte infiltrates. Synovial membrane sections from both animals stained with anti-B. burgdorferi mAbs. Several parts of the central and peripheral nervous system were sectioned and analysed for histopathological changes and for immunohistochemical evidence of the presence of B. burgdorferi organisms (or their antigens). The brain cortex, midbrain and stem were investigated, as well as the entire spinal cord, spinal nerves and ganglia. Peripheral nerves examined were the sciatic, median and facial nerves. The most frequent changes observed were in the peripheral nerves of the control animals, which showed perivascular lymphocytic infiltrates ranging from mild to marked and immunohistochemical evidence of the presence of B. burgdorferi (Table 3). In addition, one of the control animals also had a moderate cellular infiltrate and positive immunostaining for B. burgdorferi in the spinal cord. Three of the vaccinated animals, M219, M107 (lipidated OspA),, and L971 (NS1-OspA) also showed histopathological evidence of infection in central (ganglioneuritis, M219, L971) and in peripheral nerves (perivascular lymphocytic infiltrate, M107). Macrophages within the infiltrate of animal M107 stained positive for B. burgdorferi (Table 3). Morphometric quantification of cells that bound the anti-OspA mAb LA31. We observed previously (16) that some mononuclear cells from perivascular infiltrates that were formed in tissues concomitantly with the presence of B. burgdorferi would stain either with the anti-7.5 kDa mAb or with the LA31 (anti-OspA) mAb. Antigen that bound to either mAb appeared to be present within the cytoplasm of the stained mononuclear cells (16). Since the number of stained cells was larger the more intense the cellular infiltrates, we sought to quantify inflammation in the heart and lungs indirectly, by counting the mononuclear-type cells that bound the anti-OspA LA31 mAb. These cells were quantified on sections of the heart and right lung of all of the control animals, all of the animals vaccinated with lipidated OspA and one animal of each of the NS1-OspA and NS1-OspA/MPL groups. There were marked differences in the numbers of stained cells (Table 4), as well as in the severity of the mycoarditis (Figure 6) in control vs. vaccinated animals. All of the control animals had between 24 and 106 cells per cm˛ (mean of 63) in the right heart, the section of the heart with the most cells, whereas vaccinated animals had a range of 0-25 (mean of 8), and only two of the six animals assessed showed such cells. In the lung's anterior lobe, the range was 12-155 (mean of 77) in control animals, and 0-36 (mean of 7) in vaccinated monkeys. A similar distribution of cells between control and vaccinated animals was seen in the left heart, septum and posterior lung lobe (Table 4). A comparison of the mean number of stained cells in control and vaccinated animals by organ (Kruskal-Wallis ANOVA by ranks) or overall (MANOVA) indicated that the difference between the two groups was either significant (P<0.05, right heart, septum and lung lobes) or highly significant (P<0.01, left heart and overall). [Image] Figure 6 Myocarditis in vaccinated and in control monkeys. (A) Right heart from monkey M107 (lipidated OspA group). Note the one cell staining red with the mAb that recognises the 7.5 kDa lipoprotein of B. burgdorferi. When compared with the right heart of monkey M021 (B) which is a control monkey, there is a greatly reduced inflammatory infiltrate, as well as fewer stained cells in the myocardium of the vaccinated animal. Magnification x 400 (A and B) Silver staining of brain sections. Brain sections from six vaccinated animals, the four controls and three animals that had not been challenged with B. burgdorferi were stained as described in Materials and Methods. One of the four control monkeys (M021) and one of the four animals vaccinated with lipidated OspA (L537) showed stained filaments consistent with the appearance of silver-stained B. burgdorferi spirochaetes. The remainder of the brain sections either stained negative or showed stained filaments that did not unequivocally resemble the morphology of silver stained spirochaetes (Table 3). One of the three animals that had not been exposed to B. burgdorferi also showed a staining pattern consistent with that of spirochaetes. PCR of tissue samples. B burgdorferi DNA could be amplified by PCR from the organs of all four control animals. The lung and the bladder were the organs most frequently positive. Of a total of 114 determinations, 11 (9.6%) were found to be positive in control animals. Of the six vaccinated, non immune suppressed animals whose organs were analysed by PCR, four showed one or two positive results. The lung was also the most frequently positive organ in these animals. Of a total of 150 determinations, seven (4.7%) were positive. In the four immune suppressed vaccinated animals, all but one (M106) had one or two organs from which an amplicon could be detected. Of a total of 64 determinations in these animals, four (6.3%) were positive (Table 3). Tubes without template did not show a positive signal on Southern blots. Although the frequency of appearance of PCR amplicons was higher in control than in vaccinated animals, comparison of values by a nonparametric X˛ [chi-square] test indicated that the difference was not significant (P>0.05). Similarly, the frequency of appearance of PCR amplicons in the immune suppressed animals did not differ significantly from that of the other two groups. In vitro culture. Tissue samples from the same organs that were analysed by PCR were subjected to in vitro culture. No spirochaetes were recovered. Table 3 Evaluation of tissue samples obtained post-mortem from several organ systems by PCR, histopathology (HP), B. burgdorferi-specific immunohistochemistry (IH), and silver staining (SS) ------------------------------------------------------------------------ Skin Lung Heart Kidney Vaccine/ ------------ ------------ ------------ ------------ animal no. PCR HP IH PCR HP IH PCR HP IH PCR HP IH ------------------------------------------------------------------------ Al(OH)3 L457 0/3 - - 3/3 LH + 0/3 M + 0/3 - - L549 0/3 D2 - 0/3 LH + 0/3 M + 0/3 - - MO21 0/3 - - 0/3 LH + 0/3 M + 1/3 LH + M581 0/3 - - 1/3 LH + 0/3 - - 0/3 - - Lipidated OspA M243 0/3 - - 0/3 LH - 0/3 - - 0/3 - - L537 0/3 - - 1/3 LH + 0/3 - - 0/3 - - M219 ND D2 + ND LH + ND M - ND - - M107 0/3 - - 2/3 LH + 0/3 - - 0/3 - - NS1-OspA L458 0/2 - - 0/2 - - 0/2 - - 0/2 - - L594 0/2 D2 + 0/2 P + 0/2 - - 0/2 - - L971 0/2 D3 + 0/2 P + 0/2 - - 0/2 ND - M585 0/2 D2 - 0/2 P - 1/2 - - 0/2 - - NS1-OspA/MPL L476 0/3 - - 1/3 LH + 0/3 - - 0/3 - - L712 0/2 - - 0/2 P - 0/2 - - 0/2 - - L642 ND ND ND ND ND ND ND ND ND ND ND ND M106 0/2 D2 - 0/2 P - 0/2 - - 0/2 - - ------------------------------------------------------------------------ ------------------------------------------------------------------------ Ureter Bladder Liver Spleen Vaccine/ ------------ ------------ ------------ ------------ animal no. PCR HP IH PCR HP IH PCR HP IH PCR HP IH ------------------------------------------------------------------------ Al(OH)3 L457 ND - + 0/3 - - 0/3 ND ND 0/3 - + L549 ND - - 1/3 - - 0/3 ND ND 0/3 - + MO21 ND - - 2/3 - - 0/3 ND ND 0/3 - + M581 ND - - 1/3 - - 0/3 ND ND 0/3 - - Lipidated OspA M243 ND - - 0/3 - - 0/3 ND ND 0/3 - - L537 ND - - 0/3 - - 0/3 ND ND 0/3 - - M219 ND - - ND - - ND ND ND ND - - M107 ND - - 0/3 - - 1/3 ND ND 0/3 - - NS1-OspA L458 ND - - ND - - ND ND ND 0/2 - - L594 ND - - 0/2 - - ND - - ND ND ND L971 ND U + 0/2 - - ND - - ND ND ND M585 ND - - 0/2 - - ND - - ND ND ND NS1-OspA/MPL L476 ND - - 0/3 - - 0/3 ND ND 0/3 - - L712 ND - - 0/2 - - ND H - ND ND ND L642 ND ND ND ND ND ND ND ND ND ND ND ND M106 ND - - 0/2 - - ND - - ND ND ND ------------------------------------------------------------------------ ------------------------------------------------------------------------ Spinal nerves+ Synov./Joint Brain Spinal cord ganglia Vaccine/ ------------ --------------- ------------ ----------- animal no. PCR HP IH PCR HP IH SS PCR HP IH PCR HP IH ------------------------------------------------------------------------ Al(OH)3 L457 0/3 - - 0/3 - - - ND - - ND - - L549 0/3 - - 0/3 - - - ND - - ND - - MO21 0/3 SH + 0/3 - - + ND - - ND - - M581 0/3 - - 2/3 - - ? ND I(3)+ ND - - Lipidated OspA M243 0/3 - - 0/3 - - ? ND - - ND - - L537 0/3 - - 0/3 - - + ND - - ND - - M219 ND - - ND - - ? ND - - ND GN* - M107 0/3 - - 0/3 - - ? ND - - ND - - NS1-OspA L458 0/2 - - 0/2 - - ND ND - - ND - - L594 1/2 SH* + 1/2 - - ND ND - - ND - - L971 0/2 SH* + 0/2 - - - ND - - ND GN* - M585 0/2 - - 0/2 - - ND ND - - ND - - NS1-OspA/MPL L476 0/3 - - 0/3 - - - ND - - ND - - L712 0/2 - - 1/2 - - ND ND - - ND - - L642 ND ND ND ND ND ND ND ND ND ND ND ND ND M106 0/2 - - 0/2 - - ND ND - - ND - - ------------------------------------------------------------------------ ------------------------------------------------------------------------ Periph. nerves Vaccine/ ----------------- animal no. PCR HP IH ------------------------------------------------------------------------ Al(OH)3 L457 ND I(2) + L549 ND I(4) + MO21 0/3 - - M581 0/3 I(2) + Lipidated OspA M243 1/3 - - L537 1/3 - - M219 ND - - M107 0/3 I(2) + NS1-OspA L458 ND - - L594 0/2 - - L971 0/2 - - M585 0/2 - - NS1-OspA/MPL L476 0/3 - - L712 0/2 - - L642 ND ND ND M106 0/2 - - ------------------------------------------------------------------------ LH, lymphocytic hyperplasia; M, myocarditis; SH, synovial cell hyperplasia; SH*, SH with perivascular lymphocyte infiltrate; I(2) mild perivascular lymphocytic infiltrate; I(3), moderate; I(4), marked; N, nephritis; U, ureteritis; P, pneumonitis; D2, superficial dermatitis; D3, deep dermatitis; H, hepatitis; PCR, denominator, number of aliquots amplified from a given tissue lysate; numerator, number of PCR-positive aliquots; SS, silver staining; ?, probably positive. Animal numbers in italics correspond to animals that were immune suppressed. The mAb anti-7.5 kDa was used for the immunohistochemistry. Table 4 Morphometric quantification of cells that bind anti-OspA mAb LA31 in the heart and lungs ------------------------------------------------------------------------ Lung* Lung* Vaccine/ anterior posterior animal no Right heart Left heart Septum lobe lobe ------------------------------------------------------------------------ Al(OH)3 L457 40/35 5/153 2/126 12/25 3/83 L549 24/197 5/99 24/107 75/110 34/230 M021 106/57 15/166 12/105 155/116 100/145 M581 80/81 29/100 20/91 64/59 22/252 Lipidated OspA M219 0/51 0/97 0/83 0/41 0/131 M243 0/62 0/77 0/113 0/48 0/101 L537 22/62 0/112 0/63 36/126 19/157 M107 25/58 2/94 3/90 8/45 17/168 NS1-OspA L458 0/81 0/120 0/148 0/94 0/176 NS1-OspA/MPL L476 0/51 0/76 0/69 0/89 0/120 ------------------------------------------------------------------------ Numerator: number of B. burgdorferi-positive cells per cm˛; denominator, cross-sectional area measured (mm˛); mAb, LA31 anti-OspA. *Right lung. DISCUSSION The mechanisms whereby an infection with B. burgdorferi causes Lyme arthritis or neuroborreliosis, the two most morbid manifestations of Lyme disease, are not known. In fact, there is as yet no clear understanding of whether disease is caused by the infection itself, by the immune response to it, to antigens left in the tissues after bacterial demise, or by all of these factors acting in concert. An important and inevitable implication of this lack of understanding is that thus far no bacterial molecules have been incriminated in disease pathogenesis. Vaccines against disease are therefore not feasible yet, and efforts in Lyme disease immunoprophylaxis have focused on the prevention of infection. Unfortunately, the development of a vaccine to prevent infection with B. burgdorferi is constrained by the notion that no less than sterile immunity may be required to avoid long-term sequelae. The scientific literature on syphilis contains accounts of human infections caused by inoculation with as few as ten Treponema pallidum spirochaetes (27). In rabbits, similar low level treponemal infections are too sparse to elicit an antibody response (J. Miller, personal communication) and, as in humans (29), such infections may lie quiescent for a long time, only to become clinically apparent much later (J. Miller, personal communication and ref. 29). B. burgdorferi infections can be caused in mice by intradermal injection of very few organisms (ID subscript(50)=10) (30), and once an infection becomes chronic, spirochaetes are present in the skin and other organs at much lower densities than during the acute phase (30). Similarly, spirochaetes may be cultivated from the skin of recently infected humans, but very rarely from the cerebrospinal fluid of patients with chronic neuroborreliosis (31). Notwithstanding such low level infections, disease is clinically apparent. Could a low level infection occur in an OspA-vaccinated host? The mechanism whereby B. burgdorferi spirochaetes succumb to the immunity elicited by OspA has been under intense scrutiny. A two-tiered mode of protection by anti-OspA antibody, the key mediator of protective immunity, had initially been postulated, one acting on the spirochaetes within the tick prior to transmission and the second upon delivery within the host skin (32). Evidence gathered subsequently, indicated that the expression of OspA is interrupted when spirochaetes reach the tick's salivary gland, en route to the vertebrate host (20, 33). A corollary of this finding is that anti-OspA antibody can kill only within the tick midgut, while OspA is still expressed, but not thereafter (20). One can envisage that within the midgut of an infected nymph that is feeding on an OspA-vaccinated host, a process of selection may ensue whereby spirochaetes with the least surface density of OspA molecules, and endowed perhaps with other attributes such as enhanced motility and ability to penetrate the tick midgut wall, may escape antibody-dependent killing. Two recent discoveries make this scenario more than just formally possible. First, I. scapularis saliva has the ability to inactivate complement (34). Hence, killing within the midgut could occur via a mechanism that involves antibody alone, a mechanism that appears to be slower in effecting spirochaetal death than that enacted by antibody and complement together. Second, although OspA is an abundant B. burgdorferi protein, only a minor fraction of OspA molecules is exposed on the outer surface of the spirochaete (35). This entails that small variations in the absolute number of OspA surface molecules could cause significant differences in the spirochaete killing rate. Thus, one can argue that low level infections may occur in an OspA-vaccinated host even in the absence within the tick of variant spirochaetes that do not express OspA at all (36). By analogy with syphilis, such low level infections could be pathogenic. We therefore built into this efficacy trial multiple ways of finding evidence of infection in the vaccinated animals and, moreover, we attempted to uncover putative low-level infections by radically immune suppressing a portion of the study population. DFA analysis of the ticks that were used for the challenge infection indicated that spirochaetal survival in the face of the anti-OspA serum antibody titre present at the time of challenge was minimal. Only 0.8% of the engorged ticks that were dissected and that had fed on vaccinated animals (1/121) contained detectable spirochaetes, whereas 95% (41/43) of such ticks that fed on control animals contained spirochaetes (Table 1). This result indicates that most, but not all, of the spirochaetes that remained in the ticks which fed on vaccinated animals were destroyed within the first week of exposure to anti-OspA antibody. No estimate is possible, of course, of the number of spirochaetes that were 'missing' from the midgut because they escaped the initial onslaught of anti-OspA antibody, but it is reasonable to assume that this number must be either very small, or zero. The latter alternative, which is equivalent to sterile immunity, is supported by the results of PCR applied to skin biopsy samples collected during the first 4 weeks PC. During this period, spirochaete DNA was detectable in the skin of the four control monkeys but in none of the vaccinated animals (Table 2). This result is a strong indication of absence of infection in the vaccinated animals. Moreover, by the time of the challenge infection all of the vaccinated animals had an LA2 antibody titre, i.e. a titre of anti-OspA antibody that is capable of effecting antibody-dependent killing of B. burgdorferi, that was between 100-fold (NS1-OspA/Al(OH)3 group) and 400-fold (lipidated OspA/Al(OH) group), the titre known to be effective in killing the spirochaete in vivo, which is 1-5 LA2 µg equivalents per ml 4 [sic]. On the other hand, two vaccinated monkeys (and two of the control animals) had immunohistochemically detectable spirochaetal antigens in the skin during the early localised phase of infection (Table 2) and in several organs post-mortem (Table 3). More convincingly, several of the vaccinated animals and all of the control monkeys exhibited post-mortem, spirochaetal DNA in organs such as the lungs, brain, heart, bladder and peripheral nerves (Table 3). These remote sites could have been reached only after spirochaetal dissemination. Although the frequency of appearance of PCR amplicons was higher in control than in vaccinated animals, the difference was not significant (nonparametric X˛ [chi-square] test, P>0.05). Unfortunately, silver staining of brain tissues yielded a result from which no conclusion may be drawn at this point, for spirochaete-like filaments also were observed in samples from an uninfected animal, in addition to one of the controls and one vaccinated monkey. If there was a low-level disseminated infection, it remained below the threshold of seroconversion, for no vaccinated animal seroconverted, as assessed by Western blot. In contrast, all of the control animals seroconverted within the first 6 weeks PC and exhibited a gradual increase in the number of serum antibody specificities that is consistent with an active infection. All of the control monkeys satisfied the Dressler criteria for Western blot diagnosis of a B. burgdorferi infection in humans (37). The putative low-level infection was not increased by the immune suppression protocol employed, as evidenced by the following findings: (1) the frequency of post-mortem detection of spirochaetal DNA by PCR was not significantly different from that obtained in the vaccinated animals that were not immune suppressed; (2) no spirochaetes were recovered from the immune suppressed animals by xenodiagnosis, whereas a recently infected and otherwise normal animal was successfully xenodiagnosed using a similar procedure; (3) no changes were detected in peripheral nerve conduction velocity or in amplitude of the response to stimulation during nerve conduction studies, whereas five of eight B. burgdorferi-infected rhesus monkeys (and none of ten uninfected controls) showed such changes (19); (4) cultivation of bronchoalveolar lavage and urine samples yielded no B. burgdorferi spirochaetes. It is doubtful that these results arose from insufficient immune suppression. T-cell function was clearly diminished, insofar as the blastogenic responses of PBMC and lymph node cells to Con A were severely reduced. The failure to affect B-cell function, as indicated by the persistence of the response to PWM, is probably irrelevant since the 'low-level infection' was serologically non-immunogenic and thus unlikely to have been under the control of antibody. However, surveillance of infection by anti-OspA antibodies may have occurred if OspA is re-expressed by the spirochaete in an organ-specific manner (10), as the frequent finding of anti-OspA antibodies in patients with Lyme arthritis (38) appears to suggest. The apparent failure to enhance the putative low-level infection admits, of course, an explanation other than absence of the infection per se. The working hypothesis of the IS experiment, that the B. burgdorferi infection burden was kept low by an immune surveillance mechanism, may be wrong. Indeed, the notion that spirochaetes localise to immune-privileged sites, a notion often invoked to explain why it is that a chronic B. burgdorferi infection can coexist with bactericidal serum antibodies elicited during the infection process, is inimical to the concept of immune surveillance. It is most unlikely that a low-level infection could have been detected by in vitro cultivation of tissue specimens. In any case, our inability to recover spirochaetes from the control animals implies that the in vitro culture results are moot. This outcome contrasted with results from two previous studies in which we had been able to cultivate spirochaetes from skin samples of nine out of nine infected animals (15) and four out of five infected animals (16). A potentially relevant change had been introduced in the in vitro culture procedure used in the current study, in that skin samples of a volume eight to ten times larger were cultured in the same volume of medium as before, under the naive expectation that this would increase the chances of recovering spirochaetes. As we tried to troubleshoot the culture procedure we assessed whether spirochaetes could grow in BSK-H medium in the presence of added normal skin at a low ratio of medium to skin sample volume (100:1) and also at higher ratios. Live spirochaetes were observed only in tubes where the ratio of medium to skin volume was at least 3000:1 (Y. Gu et al., unpublished observation). Skin biopsy samples obtained from several rhesus macaques infected with B. burgdorferi (strains JD1 and NT1) subsequently to the present study, unfailingly yielded cultivable spirochaetes when incubated at a medium to skin volume ratio of about 3000:1 (Y. Gu et al., unpublished observation). Could a non-immunogenic B. burgdorferi infection be pathogenic? Some of the vaccinated animals showed dermatitis shortly after the challenge infection, and lung lymphocytic hyperplasia, myocarditis and PNS and CNS involvement post-mortem. On a quantitative basis, however, pathologic changes were significantly less intense, as demonstrated by the morphometric quantification of inflammation in the heart and lungs (P<0.05) (Table 4). The results of our efficacy study thus appear to fit the following interpretation. As a consequence of their exposure to infected ticks, vaccinated animals, regardless of the vaccination protocol, were probably infected by a small number of spirochaetes. These spirochaetes, in contrast to those received by control animals, were too sparse to be immunogenic, but sufficient to disseminate, become detectable by PCR and, in some organs, by immunohistochemisty. The putative low-level infection received by the vaccinated animals was clearly less pathogenic than the full infectious burden given to the control monkeys. Since the low-level infection appeared not to be under T (or B)-cell immune surveillance, it is puzzling that it remained at a low level, unless, as mentioned before, OspA was re-expressed selectively by B. burgdorferi in certain organs. If so, one possible solution to the riddle is that the infection was transient, terminated eventually by remaining anti-OspA antibodies. Another possible, even probable explanation is that the spirochaetes that putatively evaded the anti- OspA antibody in the tick midgut were low-virulence mutants that do not express, or express only a portion of, OspA. An infection with such mutants could be self-limiting. We are currently investigating the infectivity of OspA escape mutants in mice to assess this hypothesis. Thus, while our results do not allow us to discriminate unequivocally between transient and still-exant low-level infections, they underscore more convincingly the former possibility and thus suggest that sterile immunity was ultimately achieved. ACKNOWLEDGEMENTS The authors would like to express their appreciation for the skilful technical assistance of Barbara Adams, Maryjane Dodd, Maurice J. Duplantis, Robert Frantz and Barbara Lasater (TRPRC). The excellent secretarial help of Christie Trew, as well as the photographic skill of Murphy Dowouis. TRPRC, are also acknowledged with thanks. This work was supported by SmithKline Beecham Biologicals and by NCRR/NIH grant RR00164. A portion of the results presented in this paper was communicated at the VII LDF Annual Scientific Conference on Lyme Borreliosis held in Vancouver, Canada, 28-29 April 1995. REFERENCES 1 Centers for Disease Control. Lyme Disease. United States. Morbidity and Mortality Weekly Report, 1994, 1995, 44, 459-462. 2 Jaenson, T.G.T. The epidemiology of Lyme borreliosis. Parasitology Today 1991, 7, 39-45. 3 Steere, A.C. Lyme disease. New England Journal of Medicine 1989, 321, 586-596. 4 Philipp, M.T., Lobet, Y., Bohm, R.P. Jr., Conway, M.D., Dennis, V.A. and Desmons, P. et al. Safety and immunogenicity of recombinant outer surface protein A (OspA) vaccine formulations in the rhesus monkey. Journal of Spirochetal and Tick-borne Diseases 1996, 3, 67-79. 5 Keller, D., Koster, F.T., Marks, D.H., Hosbach, P., Erdile, L.F. and Mays, J.P. Safety and immunogenicity of a recombinant outer surface protein A Lyme vaccine. JAMA 1994, 271, 1764-1768. 6 Van Hoecke, C., Comberbach, M., DeGrave, D., Desmons, P., Fu, D., Hauser, P., Lebacq, E., Lobet, Y. and Voet, P. Evaluation of the safety, reactogenicity and immunogenicity of three recombinant outer surface protein (OspA) Lyme vaccines in healthy adults. Vaccine 1996, 14, 1620-1626. 7 Schoen, R.T., Meurice, F., Brunet, C.M., Cretella, S., Krause, D.S., Craft, J.E. and Fikrig, E. Safety and immunogenicity of an outer surface protein A vaccine in subjects with previous Lyme disease. Journal of Infectious Diseases 1995, 172, 1324-1329. 8 Schaible, U.E., Kramer, M.D., Eichmann, K., Modolell, M., Museteanu, C. and Simon, M.M. Monoclonal antibodies specific for the outer surface protein A (OspA) of Borrelia burgdorferi prevent Lyme Borreliosis in severe combined immunodeficiency (scid) mice. Proceedings of the National Academy of Sciences 1990, 87, 3768-3772. 9 Fikrig, E., Barthold, S.W., Kantor, F.S. and Flavell, R.A. Protection of mice against the Lyme disease spirochete agent by immunizing with recombinant OspA. Science 1990, 250, 553-556. 10 Philipp, M.T. and Johnson, B.J.B. Animal models of Lyme disease: pathogenesis and immunoprophylaxis. Trends in Microbiology 1994, 2, 431-437. 11 Gern, L., Rais, O., Capiau, C., Hauser, P., Lobet, Y., Simoen, E., Voet, P. and Pętre, J. Immunization of mice by recombinant OspA preparations and protection against Borrelia burgdorferi infection induced by Ixodes ricinus tick bites. Immunology Letters 1994, 39, 249-258. 12 Telford, S.R., Kantor, F.S., Lobet, Y., Barthold, S.W., Spielman, A., Flavell, R.A. and Fikrig, E. Efficacy of human Lyme disease vaccine formulations in a mouse model. Journal of Infectious Diseases 1995, 171, 1368-1370. 13 Golde, W.T., Burkot, T.R., Piesman, J., Dolan, M.C., Capiau, C., Hauser, P., Dequesne, G. and Lobet, Y. The Lyme disease vaccine candidate outer surface protein A (OspA) in a formulation compatible with human use protects mice against natural tick transmission with Borrelia burgdorferi, Vaccine 1995, 13, 435-441. 14 Wormser, G.P. Prospects for a vaccine to prevent Lyme disease in humans. Clinical Infection and Disease 1995, 21, 1267-1274. 15 Philipp, M.T., Aydintug, M.K., Bohm, R.P. Jr., Cogswell, F.B., Dennis, V.A., Lanners, N., Lowrie, R.C. Jr., Roberts, E.D., Conway, M.D., Karaçorlu, M., Peyman, G.A., Gubler, D.J., Johnson, B.J.B., Piesman, J. and Gu, Y. Early and early- disseminated phases of Lyme disease in the rhesus monkey: a model for infection in humans. Infection and Immunity 1993, 61, 3047- 3059. 16 Roberts, E.D., Bohm, R. Jr., Cogswell, F.B., Lanners, H.N. Lowrie, R.C. Jr., Povinelli, L., Piesman, J. and Philipp, M.T. Chronic Lyme disease in the rhesus monkey. Laboratory Investigation 1995, 72, 146-160. 17 Pachner, A.R., Delaney, E., O'Neill, T. and Major, E. Inoculation of nonhuman primates with the N40 strain of Borrelia burgdorferi leads to a model of Lyme neuroborreliosis faithful to the human disease. Neurology 1995, 45, 165-172. 18 Pachner, A.R., Delaney, E. and O'Neill, T. Neuroborreliosis in the non-human primate: Borrelia burgdorferi persists in the CNS. Annals of Neurology 1995, 38, 667-669. 19 England, J.D., Bohm, R.P., Roberts, E.D. and Philipp, M.T., Mononeuropathy multiplex in rhesus monkeys with chronic Lyme disease. Annals of Neurology 1997, 41, 375-384. 20 De Silva, A.M., Telford, III, S.R., Brunet, L.R., Barthold, S.W. and Fikrig, E., Borrelia burgdorferi OspA is an arthropod-specific transmission-blocking Lyme disease vaccine. Journal of Experimental Medicine, 1996, 183, 271-275. 21 Piesman, J., Mather, T., Sinsky, R. and Spielman, A. Duration of tick attachment and Borrelia burgdorferi transmission. Journal of Clinical Microbiology 1987, 25, 557-558. 22 Aydintug, M., Gu, Y. and Philipp, M. Borrelia burgdorferi antigens that are targeted by antibody-dependent, complement-mediated killing in the rhesus monkey. Infection and Immunity 1994, 62, 4929-4937. 23 Cogswell, F.B., Bantar, C.E., Hughes, T.G., Gu, Y. and Philipp, M.T. Host DNA can interfere with the detection of Borrelia burgdorferi in skin biopsy specimens by PCR. Journal of Clinical Microbiology 1996, 34, 980-982. 24 Katona, L.I., Beck, G. and Habicht, G.S. Purification and immunological characterization of a major low-molecular-weight lipoprotein from Borrelia burgdorferi. Infection and Immunity 1992, 60, 4995-5003. 25 Leuba-Garcia, S., Kramer, M.D., Wallich, R. and Gern, L. Characterization of Borrelia burgdorferi isolated from different organs of Ixodes ricinus ticks collected in nature. Int. J. Med. Microbiol. Virol. Parasitol. Infect. Dis. 1994, 280, 468-475. 26 Sparks, C.K., Genest, D.R. and Crum, C.P. Rapid, reliable microwave modification of Steiner silver impregnation method for diagnosis of spirochetes and non-filamentous bacteria. Journal of Histotechnology 1995, 18, 315-318. 27 Magnuson, H.J., Thomas, E.W., Olansky, S., Kaplan, S.I., de Mello, L. and Cutler, J.C. Inoculation of syphilis in human volunteers. Medicine (Baltimore) 1956, 35, 33-82. 29 Hook, III, E.W. and Marra, C.M., Acquired syphilis in adults. New England Journal of Medicine 1992, 326, 1060-1069. 30 Barthold, S.W. Infectivity of Borrelia burgdorferi relative to route of inoculation and genotype in laboratory mice. Journal of Infectious Diseases 1991, 163, 419-420. 31 Halperin, J.J., Logigian, E.L., Finkel, M.F. and Peral, R.A. Practice parameters for the diagnosis of patients with nervous system Lyme borreliosis (Lyme disease). Neurology 1996, 46, 619- 627. 32 Fikrig, E., Telford, III, S.R., Barthold, S.W., Kantor, F.S., Spielman, A. and Flavell, R.A., Elimination of Borrelia burgdorferi from vector ticks feeding on OspA-immunized mice. Proceedings of the National Academy of Sciences of the United States of America, 1992, 89, 5418-5421. 33 Schwan, T.G., Piesman, J., Golde, W.T., Dolan, M.C. and Rosa, P.A. Induction of an outer surface protein on Borrelia burgdorferi during tick feeding. Proceedings of the National Academy of Sciences of the United States of America, 1995, 92, 2909-2913. 34 Mather, T.N., Yeh, M.T., Ribeiro, J.M.C., Scorpio, A., Nelson, D.R. and Coughlin, R.T., Ixodes saliva: vector competence for Borrelia burgdorferi and potential vaccine strategies. VII International Congress on Lyme Borreliosis, San Francisco, CA, 1996, Abstracts. 35 Cox, D.L., Akins, D.R., Bourrel, K.W., Lahdenne, P., Norgard, M.V. and Radolf, J.D. Limited surface exposure of Borrelia burgdorferi outer surface lipoproteins. Proceedings of the National Academy of Sciences of the United States of America 1996, 93, 7973-7978. 36 Sadziene, A., Rosa, P.A., Thompson, P.A., Hogan, D.M. and Barbour, A.G. Antibody-resistant mutants of Borrelia burgdorferi: in vitro selection and characterization. Journal of Experimental Medicine 1992, 176, 799-809. 37 Dressler, F., Whalen, J.A., Reinhardt, B.N. and Steere, A.C. Western blotting in the serodiagnosis of Lyme disease. Journal of Infectious Diseases 1993, 167, 392-400. 38 Kalish, R.A., Leong, J.L. and Steere, A.C. Association of treatment-resistant chronic Lyme arthritis with HLA-DR4 and antibody reactivity to OspA and OspB of Borrelia burgdorferi. Infection and Immunity 1993, 61, 2774-2779. ----- [end]