An enigmatic feature of Lyme disease is the slow resolution of musculoskeletal symptoms that can continue after treatment, with some patients developing an inflammatory arthritis that becomes refractory to antibiotic therapy. Using intravital microscopy and the mouse model of Lyme borreliosis, we observed that Borrelia burgdorferi antigens, but not infectious spirochetes, can remain adjacent to cartilage for extended periods after antibiotic treatment. B. burgdorferi was not recovered by culture or xenodiagnosis with ticks after antibiotic treatment of WT mice and all but one of the immunodeficient mice with heightened pathogen burden due to impaired TLR responsiveness. Amorphous GFP+ deposits were visualized by intravital microscopy in the entheses of antibiotic-treated mice infected with GFP-expressing spirochetes and on the ear cartilage surface in sites where immunofluorescence staining detected spirochete antigens. Naive mice were not infected by tissue transplants from antibiotic-treated mice even though transplants contained spirochete DNA. Tissue homogenates from antibiotic-treated mice induced IgG reactive with B. burgdorferi antigens after immunization of naive mice and stimulated TNF-α production from macrophages in vitro. This is the first direct demonstration that inflammatory B. burgdorferi components can persist near cartilaginous tissue after treatment for Lyme disease. We propose that these deposits could contribute to the development of antibiotic-refractory Lyme arthritis.
Linda K. Bockenstedt, David G. Gonzalez, Ann M. Haberman, Alexia A. Belperron
Submitter: Linda K. Bockenstedt | firstname.lastname@example.org
Authors: David G. Gonzalez, Ann M. Haberman, and Alexia A. Belperron
Yale University School of Medicine
Published September 10, 2012
Among the theories advanced to explain post-Lyme disease syndromes, direct evidence for persistence of Borrelia burgdorferi (Bb) antigens had been lacking prior to our study in mice (1). Importantly, bacterial antigens were present for months after antibiotic treatment had eliminated infection. Bacterial antigen persistence, however, does not preclude concomitant alterations in immune regulation or autoreactivity in the clinical expression of disease.
Antibiotic-refractory LA differs from other HLA-DR-associated autoimmune diseases in that it has an identifiable infectious trigger, affects only the one (or two) joints inflamed during the period of untreated infection, and eventually resolves over months to years (2). In Guillain-Barre syndrome, the inciting pathogen does not have to infect the nervous system where autoimmune pathology occurs (3). The site-specific inflammation of antibiotic-refractory LA suggests that features unique to the previously infected joint are required for disease expression. Bb antigen persistence may be such a feature, and we now have evidence that antigen deposits occur in joints of wild-type mice with low pathogen burdens (unpublished data). A decline in Bb-specific B and T cell responses is not by itself proof that bacterial antigens are no longer present. The spectrum of bacterial antigens that remain and their effects on joint biology are unknown.
IL-10 is a key regulatory cytokine for inflammation and live Bb stimulate its production by innate immune cells (4-7). Although Bb-infected IL-10-/- mice not unexpectedly develop heightened IFNγ responses, IL-10 deficiency does not quite reproduce the pathology of human LA. Joints of infected IL-10-/- mice have greater numbers of CD4+ T cells and NK cells, but neutrophils remain the predominant cell type (7). Similar enhancement of T cell infiltrates into mouse joints occurs when the innate defense against Bb is interrupted in other ways, such as in the setting of TLR2-deficiency (8). Long-term studies will be necessary to determine whether IL-10-/- mice develop proliferative synovitis with lymphoplasmacytic infiltrates characteristic of human LA that can persist in the absence of live spirochetes or their debris.
Our own view is that autoimmunity per se falls short of explaining the site-specific, self-limited nature of antibiotic-refractory LA, which may take as long as 4 years to resolve. Recognition that bacterial antigens can persist near cartilage adds an important dimension to consider in the pathogenesis of this form of LA. Rather than autoimmunity suddenly “righting” itself (2), disposal of these foreign antigens may tip the balance in favor of disease resolution.
The authors have no conflict of interest.
Submitter: Elise E. Drouin | email@example.com
Authors: Allen C. Steere
Massachusetts General Hospital/Harvard Medical School
Published August 21, 2012
In their recent report (1), Bockenstedt et al. showed convincingly that Borrelia burgdorferi (Bb) antigens are retained near cartilage and patellae entheses after antibiotic therapy in C57BL/6 (B6) MyD88-/- or C3H MyD88-/- mice, respectively. MyD88-/- results in high pathogen loads because of impaired pathogen clearance. However, it is a leap to assume that these findings explain post-infectious phenomena in human Lyme disease.
First, subjective musculoskeletal pain symptoms, reminiscent of fibromyalgia, following antibiotic therapy for Lyme disease probably have a different etiology than antibiotic-refractory Lyme arthritis (LA), a syndrome of proliferative synovitis with synovial pathology similar to that in rheumatoid arthritis. Second, in contrast with MyD88-/- mice, human antibiotic-responsive and antibiotic-refractory arthritis are both associated with low spirochetal burdens in joints (2). Although Bb-specific immune responses may be higher in the refractory group during infection, both T and B cell responses to Bb antigens decline similarly after antibiotic therapy in refractory or responsive patients (3, 4), but inflammatory mediators, particularly IFN-γ, remain high or even increase in the refractory group during the post-antibiotic period (5). If Bb antigens were continuing to induce synovial inflammation, we would expect Bb-specific immune responses to remain high in the post-antibiotic period. Human antibiotic-refractory LA seems most similar to the arthritis in B6 IL-10-/- mice, which have a low pathogen load, but high levels of inflammatory mediators, particularly IFN-γ (6). It will be important to learn whether retained spirochetal antigens are found in B6 wild-type or in B6 IL-10-/- mice, and whether tissue homogenates from the cartilage of these mice induce Bb-specific immune responses in naïve mice.
The greatest known genetic risk factor for antibiotic-refractory LA is certain HLA-DR alleles (e.g. DRB1*0101 and 0401) (7), a risk factor common to many autoimmune diseases (8). Although inflammatory processes resulting from spirochetal antigens or autoimmunity are not mutually exclusive, we hypothesize that infection-induced autoimmunity is important in antibiotic-refractory LA (9). Bockenstedt et al. concluded that autoimmunity cannot explain this outcome since antibiotic-refractory arthritis resolves eventually. In patients with the Guillian-Barre syndrome, an infection-induced autoimmune disease, most patients recover fully in 3-6 months (10, 11). In LA, live Bb are killed by antibiotics or the immune response, even in patients with refractory arthritis (2). Despite the presence of activated autoreactive immune cells (9), we think that the immune system eventually regains homeostasis in the absence of the adjuvant effect of live spirochetes, which allows for the resolution of arthritis.
The authors declare that no conflict of interest exists.