Central and peripheral nervous system infection immunity and inflammation in the nonhuman primate model of lyme borreliosis.код для вставкиСкачать
Central and Peripheral Nervous System Infection, Immunity, and Inflammation in the NHP Model of Lyme Borreliosis Andrew R. Pachner, MD,1 Diego Cadavid, MD,1 Gale Shu, BS,1 Donna Dail, MS,1 Sarah Pachner,1 Emir Hodzic, DVM,2 and Stephen W. Barthold, DVM, PhD2 The relationship between chronic infection, antispirochetal immunity, and inflammation is unknown in Lyme neuroborreliosis. In the nonhuman primate model of Lyme neuroborreliosis, we measured spirochetal density in the nervous system and other tissues by polymerase chain reaction and correlated these values to anti-Borrelia burgdorferi antibody in the serum and cerebrospinal fluid, and to inflammation in tissues. Despite substantial presence of Borrelia burgdorferi, the causative agent of Lyme borreliosis, in the central nervous system, only minor inflammation was present there, though skeletal and cardiac muscle, which contained similar levels of spirochete, were highly inflamed. Anti-Borrelia burgdorferi antibody was present in the cerebrospinal fluid but was not selectively concentrated. All infected animals developed anti-Borrelia burgdorferi antibody in the serum, but increased amplitude of antibody was not predictive of higher levels of infection. These data demonstrate that Lyme neuroborreliosis is a persistent infection, that spirochetal presence is a necessary but not sufficient condition for inflammation, and that antibody measured in serum may not predict the severity of infection. Ann Neurol 2001;50:330 –338 Lyme borreliosis is a protean, multisystemic disease caused by the spirochete Borrelia burgdorferi. The organs most often affected are the skin, the joints, the heart, and the central and peripheral nervous system. Neurologic manifestations, known as Lyme neuroborreliosis (LNB), occur in 5 to 20% of North American cases and in a higher percentage of infected Europeans, and include aseptic meningitis, encephalopathy, facial nerve palsy, radiculitis, and peripheral neuropathy.1,2 Lyme borreliosis is currently the most common arthropod-borne disease in the United States, where thousands of cases are reported to the Centers for Disease Control (CDC) every year. Of the several animal models studied, the nonhuman primate (NHP) model is the most representative of human disease3 because it features consistent neurological infection4 –7 as well as involvement of tissues outside of the nervous system. Previous studies have demonstrated disseminated infection but have been limited by relatively small numbers of animals, inadequate tools to assess spirochetal presence, and incomplete assessment of the inflammation and the anti-B. burgdorferi immune response. From the 1Department of Neurosciences, UMDNJ-New Jersey Medical School, Newark, NJ; and 2Center of Comparative Medicine, Schools of Medicine and Veterinary Medicine, University of California, Davis, CA. Received Feb 16, 2001, and in revised form Apr 20. Accepted for publication Apr 20, 2001. 330 © 2001 Wiley-Liss, Inc. The goals of this study were to relate spirochetal presence in nervous system and extraneural tissues in a large group of NHPs to the anti-spirochetal host immune response and to histopathological evidence of inflammation. Specific questions we wished to answer were: (1) Are there sites within the nervous system preferentially infected by the spirochete, and how does spirochetal density in the nervous system compare with spirochetal density elsewhere; (2) within areas in the nervous system, is the presence of inflammation highly correlated with the presence of spirochetes; (3) is spirochetal load correlated with the anti-B. burgdorferi antibody response, ie do animals with more extensive infection mount higher serum antibody responses either in the serum or cerebrospinal fluid (CSF); and (4) is there intrathecal antibody production in the central nervous system? Materials and Methods Animals Eight male Rhesus macaques, Macaca mulatta, 3 to 4 years of age, were anesthetized and underwent cisternal punctures and phlebotomies as previously described.8 Housing and care Published online Jun 18, 2001; DOI: 10.1002/ana.1093 Address correspondence to Dr A. Pachner, Department of Neurosciences, UMDNJ-New Jersey Medical School, 185 South Orange Avenue, Newark, NJ 07103. E-mail: email@example.com was in accordance with the Animal Welfare Act and the Guide for the Care and Use of Laboratory Animals in facilities accredited by the American Association for Accreditation of Laboratory Animal Care. The study was reviewed and approved by the New Jersey Medical School and the University of California, Davis, Animal Care and Use Committees. Complete blood count and chemistries were obtained on all animals at each phlebotomy for monitoring purposes. Three days prior to inoculation, dexamethasone 2mg/kg once daily was given intramuscularly and was continued until 28 days postinoculation, at which time 1mg/kg was administered once daily for 1 week, followed by 0.5mg/kg for 2 weeks, after which no more dexamethasone was given. This dose of dexamethasone is considered relatively mild in NHPs, and no steroid-induced side effects, ie, fluid retention manifested as weight gain, glucose intolerance, change in appearance, or change in behavior resulted from this protocol. Areas of injection were documented, and tissue obtained for analysis were in regions distant from injections. Inoculation Four NHPs, identified as 099, 177, 199, and 383, were injected intradermally with a total of 1ml volume, in multiple 0.1ml aliquots (containing a total of 1 million N40Br strain spirochetes) along the dorsal thoracic midline, as in previous studies.6,8,11,12 Four NHPs, identified as 154, 192, 211, and 242, were infected by tick feeding as below. Because there were no differences between the routes of inoculation on any of the measures used in this study, the eight infected animals will be considered together as B. burgdorferi-infected NHPs. Inoculation of Nonhuman Primates by Feeding of Infected Ticks Ticks were infected by placing uninfected larvae on four C3H/HeJ mice ( Jackson Laboratory, Bar Harbor, ME) that were inoculated intradermally with 105 N40 spirochetes 2 weeks prior to tick feeding, as previously described.25 In brief, replete larval ticks were collected and allowed to molt and harden into nymphs. Thirty ticks were randomly sampled from this population, and individually tested by PCR for infection; all were positive. For analysis, ticks were individually crushed in liquid nitrogen with a plastic grinder, and powder was used for DNA extraction with DNeasy tissue kits, according to the manufacturer’s instruction for insects (Qiagen, Valencia, CA). Infection of NHPs by tick feeding was accomplished by glueing 1.5cm diameter polycarbonate screw-cap chambers with fabric mesh tops onto the skin of the thoracic midline, using Nexaband glue (MWI, Boise, ID). Eight nymphal ticks were placed into each chamber, chambers were sealed, the NHP thorax was wrapped with elastic bandage, then protected in a fabric vest. Each day, NHPs were immobilized with ketamine, chambers opened, and ticks were examined. If fewer than four ticks were found to be attached, three additional ticks were added. On days 3 through 5, replete ticks were collected and tested by PCR, as described above. Necropsy On the morning of the necropsy, food and water were withheld. NHPs were immobilized with intramuscular ketamine for transport, and pentobarbital was given intravenously upon arrival to the necropsy suite. The posterior neck area was shaved, scrubbed, and a cisternal puncture was performed to obtain CSF (3– 8ml). Blood was obtained by direct puncture from the inferior vena cava following laparotomy. A final lethal bolus of pentobarbital was then administered and a cannula was placed in the left ventricle to allow perfusion with saline. The animal was perfused with 1L of saline per 4kg of body weight. Fluids were drained into the right thoracic cavity by an incision in the right atrium and allowed to drain by gravity through an incision in the thoracic wall. After completion of perfusion, and cessation of cardiac function, the base of the skull was entered posteriorly, and a culture obtained of the high cervical spinal cord using a sterile disposable biopsy needle. The head was then removed, transported to a hood, and incisions into the skull made using a Stryker saw, and then brain and brainstem exposed. After cultures were obtained of brain and brainstem structures, the brain and brainstem were sectioned. The spinal cord was then exposed in the dural sheath by sequential removal of vertebrae by rongeuring. Cultures were obtained after sections of spinal cord were exposed and the dura entered using sterile scissors. The spinal cord and cauda equina were sectioned. Samples of heart, urinary bladder, and skeletal muscle were cultured. The heart, aorta, spleen, testes, lymph nodes, urinary bladder, biceps, quadriceps, brachial plexus, median nerve, and sciatic nerve were then removed for processing. These were organs known to be infected in previous studies of NHP Lyme borreliosis.11-13,18 Tissues collected underwent one of three different processes: they were frozen, embedded and frozen in OCT, or fixed for DNA extraction, immunohistochemistry, or histology, respectively. Polymerase Chain Reaction DNA was extracted from tissue and polymerase chain reaction (PCR) performed using OspA and OspB primers as in previous studies.11-13,18 On each PCR a standard curve of positive controls was run. This curve consisted of 400, 200, 100, 50, 25, 12.5fg of B. burgdorferi DNA extracted from cultured spirochetes, each diluted in water; 500ng of DNA extracted from brain of uninfected Rhesus macaques was present in each tube of the standard curve. The level of B. burgdorferi DNA of any specimen was approximated by visually comparing the band on the gel to the readings of the standard curve. Thus, the DNA level of any specimen was expressed as a number of femtograms of B. burgdorferi DNA/500ng of total DNA. The spirochetal load for any given tissue or NHP was determined as the mean of the levels of DNA present within the 31 specimens tested for each animal. Spirochetal load was also analyzed as percentage of samples within a group as being positive, ie, visible as a band on PCR. Each PCR run also included negative controls which included both water samples and DNA extracted from the brain of uninfected NHPs. The false positivity rate was 2%. Runs in which a false positive was detected were assumed to be contaminated, and data from those runs were not included in the data analysis. Levels of spirochetal DNA within tissue determined by analyzing the density of the band on gel were confirmed in selected samples by analysis of the PCR product, using capture Pachner et al: Lyme Neuroborreliosis 331 Fig 1. Spirochetal DNA in tissues. Spirochetal density is expressed as the mean log 10 of the femtograms per 500ng of extracted DNA of tissue. All samples from each tissue were used in the calculation, including negative samples which were assigned a score of 0. enzyme-linked immunosorbent assay (ELISA) as previously described.13,18 Statistical analyses were performed using DataDesk software (DataDesk, Ithaca, NY). Interference with B. burgdorferi Polymerase Chain Reaction by DNA Extracted from Tissue The 500ng of DNA extracted from tissue was added to tubes containing 400, 200, 100, 50, 25, or 12.5fg of B. burgdorferi DNA. In each experiment, a positive control was included in which water was added instead of tissue DNA. The degree of interference was measured by how many logs of 2 the sensitivity curve was pushed to the right relative to the sensitivity of the standard curve without tissue DNA. For example, an absence of signal at 12.5, 25, and 50, but positivity at 100fg, was read as an interference of 3, when the sensitivity of the standard curve performed in known noninterferring DNA was 12.5fg. ELISA/Immunoblot Serum and CSF antibody ELISAsa and immunoblotting were performed as previously described8,9,18 Each plate contained a positive control, the necropsy serum of NHP 099, at dilutions of 1 to 5000, 1 to 15000, 1 to 45000, and 1 to 135000. Sera other than the positive control were run at a dilution of 1 to 5000. Commercial B. burgdorferi sensu stricto nitrocellulose strips (Microbiology Reference Laboratory, Cypress, CA) were used. Pathology Inflammation was assessed by examination of hematoxylin and eosin (H and E) stained tissue as described previously,11 with a score of 0 to 5, ranging from no inflammatory cells (a score of 0) to confluent inflammation (a score of 5). Results Clinical Manifestations of Infection As in previous studies in this model,8,12,13 NHPs tolerated this infection well without obvious clinical signs. Behavior monitored by Primate Center personnel appeared normal, level of activity was similar to uninfected NHPs, and weight was maintained. Neuropsychological or activity measures were not performed. Spirochetal Presence in Tissue Spirochetal presence in tissue was assessed by PCR of DNA extracted from tissues obtained at the necropsy 3 months after infection. The sensitivity of the PCR was generally 12.5fg of B. burgdorferi DNA for the OspA target. Results for groups of tissue are shown in Table Table 1. PCR for B. burgdorferi in Tissues Percent Positive Samples by PCR Mean Density (fg B. burg. NDA/500ng Extracted DNA Tissue Source Needle Inoculated Tick Inoculated Both Groups Needle Inoculated Tick Inoculated Both Groups CNS PNS Heart Tissue Other Tissues All Tissues 22/63 (35%) 11/26 (42%) 4/17 (24%) 5/18 (28%) 42/124 (34%) 20/65 (31%) 10/27 (37%) 5/15 (33%) 6/21 (29%) 41/128 (32%) 42/128 (33%) 21/53 (39%) 9/32 (28%) 11/39 (28%) 82/252 (33%) 38 64 44 26 42 34 59 52 31 41 36 61 47 29 41 CNS ⫽ central nervous system; PNS ⫽ peripheral nervous system, including muscle; Other Tissues ⫽ bladder, skin, spleen, testis, lymph. 332 Annals of Neurology Vol 50 No 3 September 2001 1; the mean level in the third column is the mean for all tissues in the group, including those samples which were negative. Within the cerebrum, samples obtained from the frontal and anterior parietal areas were more frequently positive with higher levels than samples from the posterior parts of the cerebrum. Brainstem and spinal cord had less organism than cerebrum. The highest concentration of spirochetes occurred in peripheral nervous system tissue, particularly in the skeletal muscles. Results of the PCR analysis tissue by tissue are shown in Figure 1. To confirm the results with OspA target, PCRs on all specimens were performed using the OspB target. Results were more than 90% concordant, with lack of concordance being mainly the result of samples in which OspA was positive and OspB negative. Interference with Polymerase Chain Reaction by Extracted DNA Extraction of tissue specimens using the above techniques yielded tissue extracts of high purity as assessed by OD 260/OD 280 ratio. When 500ng of these extracts were added to PCR of a standard curve of B. burgdorferi DNA, interference with PCR signal was observed, but the interference varied among extracts. Those tissue samples that were positive on the original PCR displayed less ability to interfere. This was observed primarily with the OspA PCR in which 7 of 12 DNA samples negative on the B. burgdorferi PCR interfered with the standard curve, while only 2 of 16 initially positive DNA samples interferred. There was no correlation between OD 260 of OD 280 ratio, a crude assay for DNA purity, and the degree of interference (data not shown). Interference was not affected by pretreatment of the interferring specimens with RNase. multiple bands visualized, the strongest being at 17, 18, 39, 41, 58, 60, and 93 kD (see Fig 2B). Anti-B. burgdorferi Antibody Response in Cerebrospinal Fluid ELISA. Anti-B. burgdorferi antibody was detectable in the cerebrospinal fluid (CSF) beginning in the second month postinoculation (data not shown). When the serum and CSF were adjusted to the same concentration of IgG, values were similar, consistent with absence of significant intrathecal IgG antibody production. The total immunoglobulin, IgG, IgA, and IgM, and total protein concentrations in the CSF were normal for cisternal fluid. There were no unique bands seen in the CSF relative to those seen in the serum, and the pattern, when corrected for immunoglobulin concentration, was similar to that in the serum (data not shown). IMMUNOBLOT. H and E of Target Tissues Inflammation was present primarily in skeletal and cardiac muscle as shown in Table 2. Brachial plexus and peripheral nerves had a low level of inflammation but significantly more than the central nervous system or those tissues sampled in the group “other tissues,” including aorta, skin, spleen, bladder, testis, and lymph node. No inflammation was present in the tissues of uninfected NHPs, ie scores for these animals in all tissues were 0. Examples of inflammation observed in brain and nerve are shown in Figures 3A and B, respectively. Anti-B. burgdorferi IgG antibody began to appear at a significant level in the second month of infection, and rose through the third month (Fig 2A). The pattern of development of IgG antibody was similar in most animals except for NHP 192, which had a higher antibody level at all time points than the other NHPs. Correlations Between Spirochete Levels in Nonhuman Primates and Antibody Response There was no significant relationship between spirochete load and anti-spirochetal antibody response. Although NHP 192 had higher antibody levels than the other animals, the spirochete load in NHP 192 was similar to the other animals, being in midrange for the group of eight NHPs. Spirochete levels were analyzed in relationship to band densities on immunoblots, and again there were no significant correlations between the spirochete densities and any of the bands visualized on the blots. The immunoblot patterns for the anti-B. burgdorferi IgG antibody response of all animals were similar. In both groups, the immunoblots of the preinoculation sera were negative, and an increasing number of bands appeared with increasing time postinoculation. At the time of necropsy 3 months postinoculation, sera from all animals in both groups had fulfilled CDC criteria for immunoblot positivity with Correlations Between Spirochete Levels in Tissue and Inflammation Correlations were assessed separately for individual animals in addition to the combined group of eight animals. Spirochetal DNA was present at similar levels in all tissues, but inflammation was much more marked in peripheral and skeletal muscle. A scatterplot was generated from the spirochetal density and the inflam- Anti-B. burgdorferi Antibody Response in Serum ELISA. IMMUNOBLOT. Pachner et al: Lyme Neuroborreliosis 333 Fig 2. (A) Anti-B. burgdorferi IgG antibody in tick- versus needleinoculated NHPs. Sera were diluted 1 to 2,500, and run on the same day for comparison purposes on a routine anti-B. burgdorferi IgG ELISA. (B) Immunoblot patterns for IgG antibody at necropsy in tick- and needle-inoculated NHPs. Strips 1 and 2 reacted with positive sera from infected humans with strong and weak IgG antibody responses, respectively, and strip 3 represents a negative control human serum. Strips 4 to 7 are from the needle-inoculated NHPs (99, 177, 199, and 383, respectively) and strips 8 to 11 are from the tickinoculated NHPs (154, 192, 211, and 242, respectively). All sera were used at a dilution of 1 to 100. matory index. The correlation coefficient was weakly positive at 0.191 (Spearman Rank). It appeared from the scatterplot that the majority of the variability was present in those tissues without inflammation, ie an inflammatory index of 0. When these values were deleted from the analysis, the correlation coefficient became more significant at 0.311. Discussion Neurological involvement in the NHP model is unique3 among animal models of Lyme borreliosis. Meningitis, neuritis, and myositis are prominent features of NHP Lyme borreliosis.5,8,12,14 –16 B. burgdorferi has been cultured from the spinal cord and CSF, PCRs are positive in skeletal muscles, peripheral nerves, nerve root and brachial plexi, and spirochetes can be visualized by immunohistochemistry.11 In this study, 334 Annals of Neurology Vol 50 No 3 September 2001 the nervous system continued to be a favored site for spirochetal invasion with PCR positivity consistently detectable in sites throughout the neuraxis. The NHP model, as an excellent model of human disease,3 can be used to test hypotheses about neurological involvement that cannot be answered in rodent models in which the nervous system is not involved. This manuscript presents for the first time a combined analysis of spirochetal load, immunological response to the spirochete in the CSF and serum, and inflammation in infected tissues in a large group of animals in the NHP model of LNB. The work showed that B. burgdorferi is widely disseminated throughout the central and peripheral nervous system, a strong host immune response attacks the spirochete but is unable to clear the organism, and there is widespread in- Table 2. Inflammation in Tissues of Infected NHPs Tissue Brain Brain stem Spinal cord Cauda equina Brachial plexus/peripheral nerves Skeleton muscle Cardiac muscle Other tissues (aorta, skin, spleen, bladder, testis, lymph node) Degree of inflammationa ⬍0.1 (2/188) ⬍0.1 (2/47) ⬍0.1 (3/80) ⬍0.1 (1/13) 0.3 (14/46) 2.8 (67/24) 1.8 (74/41) ⬍0.1 (7/81) a Total inflammation score/number of tissues sampled. flammation in which presence of spirochete is necessary but not sufficient to cause inflammation. The presence of spirochetal infection was demonstrated by identifying B. burgdorferi DNA in multiple tissues by PCR. PCR is the preferred method of detection of chronic infection in the NHP model, as other techniques such as culture, mouse infectivity testing, and immunohistochemical localization have inadequate sensitivity.13 The level of spirochetal load in this group of NHPs varied near the level of detection of B. burgdorferi DNA in tissue, ie 5 to 15fg in 500ng extracted DNA. Some tissue samples were strongly positive by gel electrophoresis of PCR products indicating a density 3 to 100 times higher than this cutoff value, while other tissue samples were negative. It appears likely that some of these “negative” tissues were false negative, since many samples of DNA extracted from PCRnegative tissues had substances that interferred with the PCR. This problem is common in PCR studies of tissue where pathogen DNA represents an extremely small percentage of total DNA assayed, and has been previously reported in the NHP model of Lyme borreliosis.17 The density of spirochetes in a given tissue was determined in two ways: first, by a semiquantitative measure of samples within a tissue based on comparison to a standard curve, and second, by sampling numerous tissues and determining the percentage of tissue samples positive, ie; above the level of detection. Our first hypothesis was that spirochetes would be found by PCR diffusely through the central and peripheral nervous system, and this hypothesis was supported by the results. Based on data from chronically immunosuppressed NHPs,18 which had demonstrated that the densities of spirochetes were 100- to 1,000-fold higher in skeletal and cardiac muscle than in CNS tissue, we additionally hypothesized that the same ratio would hold true in the immunocompetent NHPs but that the density of spirochetes would generally be less. The density of spirochetes per milligram of tissue, or per microgram of extracted DNA, was indeed lower in our Fig 3. Samples of inflammation in the brain (A) and sciatic nerve (B) of infected NHPs. (A) A focal area of meningeal thickening and inflammation in the frontal lobe. A nest of subperineurial inflammatory cells in the sciatic nerve is seen. immunocompetent NHPs relative to the immunosuppressed NHPs,18 but the densities were surprisingly uniform in all tissues, with CNS spirochetal load being essentially identical to that in other tissues, and at levels similar to those in immunosuppressed NHPs. These data provide evidence for the theory that spirochetes in CNS tissue are protected from immune-mediated clearance relative to spirochetes in peripheral nerve, skeletal muscle, and heart. This may be particularly relevant to chronic CNS Lyme disease,19 –22 in which the organism may persist in the CNS for months to years. The studies described in this manuscript were of animals infected for about 3 months; longer studies are now in progress. The primary immune defense mechanism against B. burgdorferi infection, as with other spirochetal infections,23 is the specific humoral immune response.24,25 The anti-B. burgdorferi antibody response in our stud- Pachner et al: Lyme Neuroborreliosis 335 ies displayed characteristics seen previously.8,9 There was an initial IgM response followed by a switch to IgG. The amplitude of the IgG response by ELISA and the number of B. burgdorferi proteins identified on immunoblot increased with increasing duration of infection. The height and complexity of the IgG response in NHP Lyme borreliosis is likely important for spirochetal clearance because NHPs treated with steroids who mount a high titer IgM response with a negligible IgG response have very high levels of spirochetes18 in tissues other than central nervous system, relative to animals who mount a typical strong IgG response to disseminated infection. One of the questions posed at the outset of these studies was whether spirochetal load was consistently correlated with the anti-B. burgdorferi antibody response. A reasonable hypothesis would be that because the presence of the spirochete theoretically drives the specific antibody response, the higher the level of antibody, the greater the spirochetal load. The only animal with an antibody response significantly different than the other NHPs was NHP 192, which had a higher antibody response detected by ELISA of the serum. However, the spirochetal load in this NHP was essentially identical to the others. Conversely, NHP 177, the only NHP with a substantially lower spirochetal load relative to the others, had a similar antibody level to the other NHPs. Thus, it appeared that there was no correlation between spirochetal load and level of antibody as measured by ELISA. Interestingly, the immunoblots of the serum antibody response in all NHPs looked very similar. Thus, the elevated antibody seen on the ELISA of the serum of NHP 192 may have resulted from antibodies directed against nonprotein antigens not detected on immunoblots, an antibody reactivity previously described in Lyme borreliosis.26 Another question we wished to answer was: Within areas in the nervous system, is the presence of inflammation highly correlated with the presence of spirochetes? These data confirm an association, but it is not a close one. Thus, although the presence of spirochetes is a necessary condition, it is not sufficient. For instance, the cerebrum had a large load of spirochetes relative to other organs, but had no inflammation. Other organs, such as skeletal and cardiac muscle, had very high levels of inflammation with spirochetal levels similar to that of brain. However, it is still possible that a correlation does exist between spirochetal load and inflammation, as skeletal and cardiac muscle may have had much higher levels of spirochetes prior to immune-mediated clearance. Thus, the low levels in muscle relative to the levels seen in immunosuppressed NHPs18 may indicate that the necropsy tissue assayed in this experiment was the result of substantial, but incomplete, clearance of the spirochete from these tissues. For example, spirochete numbers and inflamma- 336 Annals of Neurology Vol 50 No 3 September 2001 tion may have been more highly correlated at earlier time points postinfection. The inflammatory response was assessed in tissues obtained at necropsy by examining muliple organs stained with hematoxylin and eosin. The inflammation was similar to that described in previous studies in the NHP model11,27 for most tissues. Inflammatory infiltrates in cardiac muscle have been composed of T cells and plasma cells and to a lesser degree B cells. There have been few studies of inflammation in tissues of humans with Lyme borreliosis, but those published have been consistent with the NHP model in showing multifocal collections of lymphocytes and plasma cells.27 Cardiac muscle inflammation has been consistent in the NHP model of Lyme borreliosis, similar to the mouse model,28,29 while skeletal muscle inflammation, though not as prominent previously in NHPs injected with N40Br, has been seen in other studies of LNB in NHPs infected with other B. burgdorferi strains.15 Why inflammation tends to occur more in cardiac and skeletal muscle relative to other organs is unclear but may be an initial preferential tropism of the spirochete to collagen-rich muscle followed by immune-mediated incomplete clearance of the spirochete. B. burgdorferi is known to have an affinity for collagen.28,30 In the immune response to other pathogens, intrathecal antibody production is associated with improved clearance of organisms from the CNS.31,32 Another hypothesis that we tested in this study was that selective intrathecal IgG antibody production would not be prominent, despite the presence of large amounts of B. burgdorferi within the central nervous system. This hypothesis was based on the fact that N40Br is a B. burgdoferi sensu stricto strain of the pathogen, and in human studies of Lyme neurological involvement this genogroup of the pathogen, intrathecal IgG antibody production33 has not been prominent, in contrast to the pathogen of Lyme meningitis in Europe, B. garinii, in which selective concentration of intrathecal IgG antibody is a consistent feature.34,35 Previous to this study, we had found anti-B. burgdorferi IgG antibody in the CSF of our infected NHPs,8,13 but had not tested specifically whether this was more than would be predicted by passive diffusion from the serum. In the current work we documented that intrathecal IgG antibody production was not present and that the antibody present in the CSF was likely because of diffusion from the serum; these data are consistent with the human disease. The factors that determine which types of infections in the nervous system induce intrathecal antibody production are unknown. Studies are currently underway in our laboratory to assess the presence of intrathecal antibody production in NHPs infected with a B. garinii strain of B. burgdorferi, a European strain. Our studies also compared two different routes of inoculation, needle and tick, and found them to be comparable. The optimum route for inoculation in animal models of Lyme borreliosis has been controversial, with evidence that both needle inoculation and tick inoculation result in systemic infection,36,37 although immune responses to the organism may be different38 – 40 and vaccine studies may show different results for protection for the challenge by the two routes.41 Tick inoculation has significant methodological disadvantages relative to needle inoculation. Investigators who perform tick inoculations must have experience with the insects and facilities that will support ticks. The time required for tick inoculations is greater than for needle injection, because initially mice need to be infected by needle; the ticks then need to be fed on the infected mice prior to putting the ticks on NHPs. Finally, the inoculum entering the NHP cannot be quantitated, and may vary from one NHP to the next. Because of these problems, most investigators working with animal models of Lyme borreliosis use needle inoculation, which is simple and fast and requires no special facilities. However, tick inoculation has some theoretical benefits. For instance, some investigators feel that infection with laboratory-reared ticks is more “natural” than needle inoculation. In addition, others feel that expression of B. burgdorferi proteins in the infected animal is likely to be more similar to expression in ticks rather than culture medium. However, in the only study relevant to this issue, B. burgdorferi from laboratory-reared ticks behaved more similarly to cultured spirochetes than to field-sampled ticks,42 indicating significant complexities in delivery of infection by ticks. In this study, the NHPs underwent transient immunosuppression with the corticosteroid dexamethasone during the initial phase of the infection, at a dose considered mild for NHPs. This technique has been utilized to maximize the yield of infection because in previous studies in this model, some immunocompetent NHPs spontaneously cleared infection within the first few months, presumably from the rapid development of high-titered anti-B. burgdorferi IgG.18 This mild immunosuppression protocol, previously used in this model,8,13 resulted in all eight NHPs in these 2 groups becoming infected, and allowed the use of a minimum of experimental animals for these studies. These results in the NHP model have direct relevance to the diagnosis and therapy of Lyme neuroborreliosis in humans.2 The height of the antibody response to B. burgdorferi within an outbred population such as humans or NHPs, infected with the same strain and amount of spirochete, will be variable. Thus, the relative amplitude of the antibody response is likely not to be useful clinically to determine severity of inflammation. The magnitude of inflammation bears only a slight association with spirochete density, and CNS tissues had minimal inflammation despite sub- stantial spirochetal densities; thus, measures of inflammation such as CSF pleocytosis are not highly correlated with spirochete load. Our studies were performed with an American isolate of N40Br. Because the clinical characteristics of Lyme neuroborreliosis are different in Europe,35 and the strains of spirochete are substantially different there, the above conclusions may not be able to be applied to European LNB. Studies are now in progress in NHPs infected with a European strain. In summary, the studies outlined identify the central and peripheral nervous system as major reservoirs of spirochetal infection and demonstrate that a strong, well-developed anti-B. burgdorferi humoral immune response does not clear spirochetes from the animal during months of infection, especially in the nervous system. Spirochetal presence is a necessary but not sufficient condition, for inflammation. Despite the presence of the spirochetes in CNS tissue, there is little inflammation and no intrathecal antibody production in the CNS, in contrast to strong systemic antibody production and marked inflammation in muscle. Supported by the National Institutes of Health (NO1-AI 95358 and RO1-NS34715 to A.R.P.). References 1. Pachner AR, Steere AC. The triad of neurological manifestations of Lyme disease. Neurology 1985;35:47–53. 2. Pachner AR. Borrelia burgdorferi in the nervous system: the new “Great Imitator.” Ann NY Acad Sci 1988;539:56 – 64. 3. Coyle PK. Neurological Lyme disease: is there a true animal model? Ann Neurol 1995;38:560 –562. 4. Philipp MT, Aydintug MK, Bohm RP, et al. The early and early disseminated phases of Lyme disease in the Rhesus monkey: a model for the infection in humans. Infect Immun 1993;61:3047–3059. 5. Roberts ED, Bohm RP, Cogswell FB, et al. Chronic Lyme disease in the Rhesus monkey. Lab Invest 1995;72:146 –160. 6. Pachner AR, Amemiya K, Delaney E, et al. Interleukin 6 is expressed at high levels in the CNS in Lyme neuroborreliosis. Neurology 1997;49:147–152. 7. Roberts ED, Bohm RP, Lowrie RC, et al. Pathogenesis of Lyme neuroborreliosis in the rhesus monkey: the early disseminated and chronic phases of disease in peripheral nervous system. J Infect Dis 1998;178:722–732. 8. Pachner AR, Delaney E, O’Neill T, Major E. Inoculation of non-human 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. 9. Pachner AR, Ricalton N, Delaney E. Comparison of polymerase chain reaction with culture and serology in murine experimental Lyme borreliosis. J Clin Microbiol 1994;31:208 –214. 10. Dressler F, Whalen JA, Reinhardt BN, Steere AC. Western blotting in the serodiagnosis of Lyme disease. J Infect Dis 1993; 167:392– 400. 11. Cadavid D, O’Neill T, Schaefer H, Pachner AR. Borrelia burgdorferi localizes to the leptomeninges, perineurium, and perimysium in the non-human primate model of Lyme neuroborreliosis. Lab Invest 2000;80:1043–1054. 12. Pachner AR, Delaney E, O’Neill T. Neuroborreliosis in the Pachner et al: Lyme Neuroborreliosis 337 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. non-human primate: Borrelia burgdorferi persists in the central nervous system. Ann Neurol 1995;38:667– 679. Pachner AR, Zhang W, Schaefer S, O’Neill T. Detection of active infection in nonhuman primates with Lyme neuroborreliosis: comparison of PCR, culture and a bioassay. J Clin Microbiol 1998;36:3243–3247. Philipp MT, Johnson BJB. Animal models of Lyme disease: pathogenesis and immunoprophylaxis. Trends Microbiol 1994; 2:431– 437. England JD, Bohm RP, Roberts ED, Philipp MT. Lyme neuroborreliosis in the rhesus monkey. Semin Neurol 1997;17:53– 56. England JD, Bohm RP, Roberts ED, Philipp MT. Mononeuropathy multiplex in rhesus monkeys with chronic Lyme disease. Ann Neurol 1997;41:375–384. Cogswell FB, Bantar CE, Hughes TG, et al. Host DNA can interfere with detection of Borrelia burgdorferi in skin biopsy specimens by PCR. J Clin Microbiol 1996;34:980 –982. Pachner AR, Amemiya K, Bartlett M, et al. Lyme borreliosis in rhesus macaques—effect of corticosteroids on spirochetal load and isotype switching of anti-Borrelia burgdorferi antibodies. Clin Diagn Lab Immunol 2001;8:225–232. Ackermann RE, Gollmer E, Rehse-Kupper B. Progressive Borrelien-Enzephalomyelitis. Chronische manifestation der Erythema-chronicum-migrans Krankeit am Nervensystem. Dtsch Med Wochenschr 1985;110:1039 –1042. Logigian EL, Kaplan RF, Steere AC. Chronic neurological manifestations of Lyme disease. N Engl J Med 1990;323:1438 – 1444. Pachner AR, Duray P, Steere AC. Central nervous system manifestations of Lyme disease. Arch Neurol 1989;46:790 –795. Logigian EL, Kaplan RF, Steere AC. Successful treatment of Lyme encephalopathy with intravenous ceftriaxone. J Infect Dis 1999;180:377–383. Cadavid D, Pachner AR. Spirochetal infections of the CNS. In: Griggs RC, Joynt RJ, eds. Clinical neurology. Philadelphia: Lippincott-Raven, 1998. Johnson RC, Kodner C, Russel M. Passive immunization of hamsters against experimental infection with the Lyme disease spirochete. Infect Immun 1986;53:713–714. Barthold SW, Feng S, Bockenstedt LK, et al. Protective and arthritis-resolving activity in serum from mice infected with Borrelia burgdorferi. Clin Infect Dis 1997;25:S9 –S17. Wheeler CM, Garcia Monco JC, Benach JL, et al. Nonprotein antigens of Borrelia burgdorferi. J Infect Dis 1993;167:665– 674. England, JD, Bohm RP, Roberts ED, Philipp MT. Mononeuropathy multiplex in Rhesus monkeys with chronic Lyme disease. Ann Neurol 1997;41:375–384. Pachner AR, Basta J, Hulinska D. Localization of Borrelia burgdorferi in murine Lyme borreliosis by electron microscopy. Am J Trop Med Hyg 1995;52:128 –133. 338 Annals of Neurology Vol 50 No 3 September 2001 29. Barthold SW, deSouza MS, Janotka JL, et al. Chronic Lyme borreliosis in the laboratory mouse. Am J Pathol 1993;143: 959 –972. 30. Sigal LH. Lyme disease: a review of aspects of its immunology and immunopathogenesis. Annu Rev Immunol 1997;15:63–92. 31. Griffin D, Levine B, Tyor W, et al. The role of antibody in recovery from alphavirus encephalitis. Immunol Rev 1997;159: 155–161. 32. Tyor WR, Wesselingh S, Levine B, Griffin DE. Long term intraparenchymal Ig secretion after acute viral encephalitis in mice. J Immunol 1992;149:4016 – 4020. 33. Steere AC, Berardi VP, Weeks KE, et al. Evaluation of the intrathecal antibody response to Borrelia burgdorferi as a diagnostic test for Lyme neuroborreliosis. J Infect Dis 1990;161:1203– 1209. 34. Wilske B, Scierz G, Preac-Mursic V, et al. Intrathecal production of specific antibodies against Borrelia burgdorferia in patients with lymphocytic meningoradiculitis. J Infect Dis 1986; 153:304 –314. 35. Hansen K, Lebech AM. The clinical and epidemiological profile of Lyme neuroborreliosis in Denmark, 1985–1990: a prospective study of 187 patients with Borrelia burgdorferi-specific intrathecal antibody production. Brain 1992;115:399 – 423. 36. Levin M, Levine JF, Yang S, et al. Reservoir competence of the southeastern five-lined skink (Eumeces inexpectatus) and the green anole (Anolis carolinensis) for Borrelia burgdorferi. Am J Trop Med Hyg 1996;54:92–97. 37. Shih CM, Liu LP. Accelerated infectivity of tick-transmitted Lyme disease spirochetes to vector ticks. J Clin Microbiol 1996; 34:2297–2299. 38. Roehrig JT, Piesman J, Hunt AR, et al. The hamster immune response to tick-transmitted Borrelia burgdorferi differs from the response to needle-inoculated, cultured organisms. J Immunol 1992;149:3648 –3653. 39. Golde WT, Kappel KJ, Dequesne G, et al. Tick transmission of Borrelia burgdorferi to inbred strains of mice induces an antibody response to P39 but not to outer surface protein A. Infect Immun 1994;62:2625–2627. 40. Aydintug MK, Gu Y, Philipp MT. Borrelia burgdorferi antigen that are targeted by antibody-dependent, complement-mediated killing in the rhesus monkey. Infect Immun 1994;62:4929 – 4937. 41. Hagman KE, Yang X, Wikel SK, et al. Decorin-binding protein A (DbpA) of Borrelia burgdorferi is not protective when immunized are challenged via tick infestation and correlates with the lack of DbpA expression by B. burgdorferi in ticks. Infect Immun 2000;68:4759 – 4764. 42. Brunet LR, Spielman A, Fikrig E, Telford SR III. Heterogeneity of Lyme disease spirochetes within individual vector ticks. Res Microbiol 1997;148:437– 445.