Chlamydia pneumoniae Infection of the Central Nervous System Worsens Experimental Allergic Encephalitis

Chlamydia pneumoniae belongs to a family of intracellular organisms that typically causes a self-limiting respiratory infection (1). More recently, C. pneumoniae has been linked to a number of chronic human diseases including those that involve the central nervous system (CNS; reference 2). Infections are known to play a role in the development and progression of a number of autoimmune diseases, and we have suggested that multiple sclerosis (MS) may be linked to infections with C. pneumoniae (3–7).

Although the etiology of MS is not known, clinical and pathologic observations suggest a close interplay between an infectious agent(s) and an autoimmune response to myelin antigens in the development of the disease (8, 9). Epidemiologic studies have implicated environmental factors and most likely infectious agents as a necessary element in the development of MS (10). The autoimmune basis for MS stems from similarities between MS and the animal model, EAE (11). In view of the possible association between chlamydial infection and the development of MS, we examined the effect of systemic infection of C. pneumoniae on the development and progression of EAE.

Female SJL/J and C57BL/6 mice were purchased from The Jackson Laboratory. Guinea pig myelin basic protein (MBP) and mouse spinal cord homogenate was prepared as described previously and myelin oligodendrocyte protein (MOG) peptide (p35–55: MEVGWYRSPFSRVVHLYRNGK) was synthesized by Genemed Synthesis, Inc. Fluorophenicol was a gift from Dr. F. DeGraves, University of Auburn, Auburn, AL. IFN-γ was measured by a commercial kit obtained from R&D Systems. Concentrated C. pneumoniae elementary bodies were obtained by growing C. pneumoniae (VR-1310; American Type Culture Collection) and C. trachomatis (HAR13; American Type Culture Collection) on a monolayer of HL cells. The number of infectious forming units of chlamydial bodies were estimated using the HL indicator cell line. The GD-11 strain of S. flexnerii, a gift from J. Bright (Vanderbilt Medical Center, Nashville, TN), was grown in soft agar. Heat killed C. pneumoniae was prepared by boiling the organism in a water bath for 5 min.

Lymphocytes isolated from draining lymph node cells were cultured in RPMI 1640 complete medium in a 96-well microtiter plate under the atmosphere of 5% CO₂ and 95% air at 37°C, and the proliferation assay was done as described previously (12).

Active and adoptively transferred EAE and the clinical scoring of paralyzed mice was done as described previously (12, 13). To examine the effect of chlamydial infection on EAE, mice were injected with live C. pneumoniae or C. trachomatis organisms. 0.5 × 10⁶ infectious units (in 0.5 ml of PBS) intraperitoneally, 7 d after receiving the first immunization.

The spinal cords were isolated from the mice after perfusion with 4% paraformaldehyde in PBS on day 18, fixed in 10% formalin, and embedded in paraffin. The presence with C. pneumoniae antigens in the sections was detected by immunohistochemistry using anti-chlamydial LPS antibody (mAB 807; Chemicon) following the protocol for the M.O.M. mouse blocking kit (Vector Laboratories). Using antigen retrieval techniques the spinal cord sections were incubated with mAB 807 (recognizes all chlamydial LPS), was added at a dilution of 1:750, and placed on a rocking platform overnight at 4°C.) The slides were then washed and biotin conjugated goat anti–mouse antibody was added and the color developed using the Envision kit (DakoCytomation). Anti Escherichia coli LPS antibody (Fitzgerald Industries) and isotype matched IgG2a antibodies (Sigma-Aldrich) were used as control antibodies in all the staining procedures.

After perfusion with PBS, spinal cords from mice were collected, and total RNA extracted using TRI Reagent (Sigma-Aldrich) in accordance with manufacturer's protocol. The following sense and antisense oligonucleotide PCR primers were used: C. trachomatis 16S RNA, 5′-ATT TGG GCA TCC GAG TAA CG (sense) and 5′ CCA CGC GGT ATT AAC CGT CT, C. pneumoniae 16S RNA, 5′-GCT AAT ACC GAA TGT AGT GTA A (sense) and 5′-ATC TAT CCT CTA GAA AGA TAG TT, and GAPDH, 5′-TGA AGG TCG GTG TGA ACG GAT TTG GC (sense) and 5′-CAT GTA GGC CAT GAG GTC CAC CAC. 4 μg of total RNA was reverse transcribed to cDNA using GeneAmp RNA PCR kit with oligo d(T)₁₆ primers (Roche). PCR amplification of each cDNA target was performed along with GAPDH, which served as an internal control for RNA quantity. Each PCR reaction contained 5 μl of cDNA, 2 μl of 10× PCR buffer (Perkin Elmer), 1 μl of 25 mM MgCl₂, 0.5 μl of each dNTP (10 mM), 0.5 μl of sense and antisense target gene-specific primers (50 pmole/μl), 0.25 μl AmpliTaq DNA polymerase (5 U/μl) (PerkinElmer), and 13.75 μl of nuclease-free H₂O, and was performed in PTC-200 Peltier Thermal Cycler (MJ Research, Inc.). PCR products were resolved on 1.5% agarose in TAE containing 0.5 μg/ml of ethidium bromide and visualized under UV light.

C. pneumoniae infection in mice was induced by intraperitoneal inoculation of infectious elementary bodies after the second immunization with mouse spinal cord homogenate (MSCH). Control mice received equal numbers of infectious C. trachomatis or an equal number of colony forming units of S. flexnerii (GD-11 strain). As shown in Fig. 1 A, mice immunized with MSCH and injected with live C. pneumoniae had a mean maximal disease severity score of 3.1, while those that received PBS, C. trachomatis or S. flexnerii, had a mean maximal clinical score of 1.8, 2.1, and 1.2, respectively (P < 0.05). To determine if enhancement of EAE was specific to the immunogen, MSCH, we examined the effect of C. pneumoniae infection in MOGp35–55 induced EAE. The mean clinical score in mice that received live C. pneumoniae after immunization with MOG p35–55 was 2.3; in contrast, the mean clinical score in vehicle treated mice was 1.3, suggesting that the effect of C. pneumoniae infection on EAE was not specific to the immunizing antigen (Fig. 1 B). A comparison of viable (i.e., infectious) elementary bodies with heat-killed (i.e., noninfectious) elementary bodies revealed that worsening of EAE requires viable C. pneumoniae organisms (Fig. 1 C). Also, worsening of EAE was dependent upon the actual number of infectious elementary bodies in the intraperitoneal inoculum (Fig. 1 D). To establish whether the effect of C. pneumoniae infection on EAE was seen after T cell priming to neural antigens, we examined infection with C. pneumoniae after adoptive transfer of MBP reactive T cells. As shown in Fig. 1 E, mean clinical score of mice that received live C. pneumoniae and MBP primed lymphocytes was 3.6. In mice that received vehicle or heat killed C. pneumoniae the mean clinical scores were 2.6 and 2.4, respectively. The sum of these results show that live C. pneumoniae can worsen EAE induced by different neural antigens and also after priming of T cells.

Three groups of mice (five/group) were immunized with MSCH in CFA and infected with either C. pneumoniae or C. trachomatis on day 7. The third group of mice was left untreated. 18 d after immunization, the mice were killed, and an RT-PCR analysis of RNA obtained from the spinal cord were performed. As shown in Fig. 2, all four mice from which RNA was available showed an RT-PCR signal for the presence of replicating C. pneumoniae. None of the mice infected with C. trachomatis showed a signal to primers that were specific for the C. trachomatis 16sRNA gene. These studies suggest that C. pneumoniae but not C. trachomatis is capable of infecting the CNS in mice with EAE.

To further prove that an inflammatory response in the brain was necessary in order to set an infection with C. pneumoniae, naive mice were infected with either C. pneumoniae or C. trachomatis and the presence of chlamydia in the CNS was examined. As shown in Table I, the RT-PCR signal for the 16S gene of C. pneumoniae and C. trachomatis was noted in the spleen, lung, and lymph nodes of naive mice, infected intraperitoneally with either C. pneumoniae or C. trachomatis. An RT-PCR signal for the 16S gene was not seen in the CNS in mice infected with either C. trachomatis or C. pneumoniae on days 3, 7, and 11 after infection.

To directly demonstrate the presence of C. pneumoniae in the CNS, spinal cords were obtained from mice which were infected with C. pneumoniae and induced to develop EAE. Spinal cords were stained with anti-chlamydial antibodies using standard immunohistochemical techniques. Intracellular staining of perivascular mononuclear and CNS parenchymal cells with anti-chlamydial antibodies was seen in at least one section of all four mice examined (Fig. 3, c and d). No staining of elementary bodies were seen in mice immunized with MSCH and inoculated with C. trachomatis (Fig. 3 b). The monoclonal antibody 807 recognizes LPS and MOMP antigens of both C. trachomatis and C. pneumoniae. No staining of chlamydial antigens was seen in mice with EAE and infected with C. pneumoniae after incubation with an irrelevant control antibody (anti–E. coli LPS antibody, unpublished data). These results corroborate our observations on RT-PCR and offer direct evidence of infection by C. pneumoniae of the CNS.

If infection with C. pneumoniae directly affects the development of EAE we predicted that antibiotic therapy aimed at C. pneumoniae would attenuate the disease. Treatment with Fluorophenicol reduced the severity of EAE. The mean maximal clinical score decreased from 2.3 in untreated mice to 1.4 in mice that received Fluorophenicol. The severity of EAE in antibiotic treated mice was similar to mice immunized with MSCH alone. Fluorophenicol did not affect the course of EAE in mice immunized with MSCH that did not receive intraperitoneal inoculation with infectious C. pneumoniae elementary bodies suggesting that the effect was unlikely to be due to any immunomodulatory effect of the antibiotic (Fig. 4).

We next examined if worsening of EAE in mice inoculated with C. pneumoniae was due to an increase in activation of MBP reactive Th1 cells. We determined the effect of in vivo infection with live chlamydia on lymphocyte proliferation and IFN-γ production to MBP in, in vitro cultures. Proliferation counts to MBP, obtained from mice immunized with MBP and infected with C. pneumoniae increased from background levels of 4,305 ± 120 to 28,854 ± 1,154 cpm to in the presence of 50 μg/ml of MBP. Background counts in C. trachomatis mice was 4,196 ± 1,492 cpm, which increased to 19,308 ± 932 cpm in the presence of MBP. In uninfected mice, proliferative response to MBP increased from 3,916 ± 186 to 15,068 ± 815, in the presence of MBP. The proliferative response to MBP was higher in both C. pneumoniae and C. trachomatis infected mice when compared with uninfected controls suggesting that concurrent infection with chlamydia can amplify an autoimmune response. A proliferative response to MBP was not seen in mice infected with C. pneumoniae alone, suggesting a lack of cross reactivity between C. pneumoniae antigens and MBP (Fig. 5). IFN-γ levels in lymphocyte culture supernatants were similarly higher in C. pneumoniae (1,020 pg/ml) and C. trachomatis infected mice (950 pg/ml) when compared with uninfected mice (612 pg/ml; P < 0.05, Fig. 5). These observations suggest that unlike its different effects on paralytic EAE, both C. trachomatis and C. pneumoniae infections are capable of enhancing T cell proliferation and IFN-γ production in response to MBP over that seen in uninfected controls.

This study shows the effect of live C. pneumoniae infection on the amplification of the autoimmune disease after immunization with three different neural antigens (MSCH, MBP, and MOG). The accentuation of EAE was seen in the setting of direct infection of the CNS by C. pneumoniae. A causal association between C. pneumoniae infection and accentuated EAE can be inferred not only from the direct presence of replicating organism in the CNS, but also from the attenuation of EAE after therapy with Fluorophenicol. We believe that infection of C. pneumoniae in the CNS is a requisite for worsening of EAE. Systemic infection with C. trachomatis enhanced the in vitro proliferative response to MBP that were higher than controls. Unlike C. pneumoniae, C. trachomatis did not infect the CNS in mice, which we believe is important to cause worsening of EAE.

One mechanism by which infections can potentially induce autoimmune disease is through molecular mimicry. After immunization with chlamydial peptides that show homology with MBP, rats developed severe EAE (14). In view of the link between chlamydial antigens and heart disease, sequence homology between heart myosin and C. pneumoniae antigens were screened and mice were immunized with antigens that showed homology with myocardial antigens (15). These two studies showed that cross reactive epitopes between chlamydial and self antigens are capable of inducing different forms of autoimmune disease.

Our study did not show evidence of molecular mimicry between C. pneumoniae and neural antigens in SJL mice. Mice infected with C. pneumoniae alone did not show a lymphocyte proliferative response to MBP. Although we cannot fully exclude the expansion of autoreactive T cells that cross-react with C. pneumoniae, we think this to be unlikely. Animals showed worsening of EAE induced by three different classes of encephalitogenic antigens, MBP, MOG, and MSCH. It is unlikely that molecular mimicry is present between chlamydial antigens and three different neural antigens.

Infectious agents are well known to rapidly expand the pool of immune cells that recognize the invading pathogen. An increase in population of T cells that recognize other antigens including those that react to self-proteins may occur consequent to the secretion of cytokines and may be sufficient to cause disease (16). In mouse hepatitis model of CNS infection, prominent demyelination mediated by CD8 T cells was seen in mice in the absence of cognate antigen in the CNS, suggesting that demyelination did not require the presence of antigen-specific T cells (17). In the mouse keratitis model induced by HSV-1, both antigenic mimicry and bystander activation are thought to be responsible for tissue injury. In trying to reconcile the views of molecular mimicry and bystander activation in autoimmunity, it was proposed that these two processes may not necessarily be exclusive and may depend upon the circulating levels of autoreactive cells (18).

There has been considerable interest in the role of infectious agents in the development of MS. We have proposed that chlamydial infections should be considered as a potential candidate agent in MS (2). Our current studies suggest that C. pneumoniae can infect the CNS in mice. A number of case reports have suggested that C. pneumoniae can cause acute CNS infections in humans (19, 20). Persistence of C. pneumoniae in the CNS is likely to provide an environment which can lead to the activation of autoreactive T cells and contribute to the pathogenesis of a chronic disease such as MS. The lack of worsening of EAE in mice receiving intraperitoneal inoculations of infectious C. trachomatis elementary bodies suggests that direct infection of CNS tissues is necessary to enhance EAE. C. pneumoniae and C. trachomatis disseminate to lymphoid organs and lung after parental administration of the pathogen (21–23). We did not observe dissemination of C. pneumoniae to the CNS in naive mice. We propose that infection of the CNS is a necessary for accentuation of EAE which may be facilitated in the presence of an ongoing CNS inflammation. We suggest that a similar scenario may occur in MS, in which a ubiquitous pathogen may amplify an autoimmune response. We predict that if an infectious agent can persist and amplify an immune response, it can modify the expression of a T cell–mediated autoimmune disease in an organ specific manner. A direct interplay between an infectious agent and autoimmunity is also likely to have immediate therapeutic implications.

We thank Dr. C. Stratton for a critical review of the manuscript.

This study was supported by the National Multiple Sclerosis Society, and the Dr. T. West and W. Weaver research funds.