Outer Surface Protein A and Arthritis in Hamsters

In the February issue of Infection and Immunity, Croke et al. (1) report that hamsters vaccinated with recombinant OspA (rOspA) following infection with Borrelia burgdorferi develop inflammatory arthritis. The authors suggest that these reports are important because an rOspA vaccine was approved by the Food and Drug Administration for use in humans.

We vigorously disagree that the findings of Croke et al. have implications for potential vaccinees on the basis of the study design, methodologies, and the lack of confirmation that this model is predictive and applicable to humans.

In the study, hamsters (weighing ≤150 g) were administered one, two, or four times the human adult recommended dose. The formulation of the OspA vaccine administered to the hamsters was different (nonlipidated, fused to a carrier protein, and adsorbed to a different aluminum hydroxide preparation) than the currently marketed SmithKline Beecham (SB) product, LYMErix [Lyme disease vaccine (recombinant OspA)]. The hamsters were infected with 10⁷ organisms, which is 10-fold higher than that stated in the Discussion section and three logs higher than that required to repeatedly induce infection in our laboratory. In some cases it was apparently necessary to reinoculate the hamsters, introducing yet another variable.

Table 1 is misleading in its statement that no nonvaccinated hamster developed arthritis; in footnote b to Table 1, Results, and Fig. 1, it is stated that an unknown percentage of the nonvaccinated, challenged animals developed arthritis. Since the control hamsters developed arthritis, the induction of inflammatory arthritis is therefore independent of vaccination status. In addition, if this were an immune response-mediated arthritis, one would expect systemic involvement including other joints, which the authors have not demonstrated. The data do not support an autoimmune process, but even if cross-reactivity were noted in hamsters, extrapolation to other species should not be assumed due to protein differences. Inexplicably, the animals that received the highest dose of OspA had the lowest percentage of swelling among vaccinees. The statistical evaluation of the swelling is based on the gross morphologic size and not on a validated microscopic pathological scale. Due to the small sample sizes (n = 4), no definite conclusion regarding the differences can be ascertained.

The same laboratory has also reported the induction of arthritis due to infection with B. burgdorferi alone using the same strain of hamsters (2). Induction of arthritis may be a species-specific effect since these changes cannot be reliably produced in other species, including other rodents, dogs, and monkeys.

Available human data contradicts the authors' results. In the pivotal efficacy trial involving about 11,000 subjects, an outside, independent data safety monitoring board (DSMB) reviewed the safety data and found no evidence of an association between LYMErix and autoimmune arthritis. Moreover, neither the independent DSMB nor SB found evidence of a higher incidence of arthritis among the recipients of LYMErix than among the recipients of a placebo. No evidence of an association between vaccination with LYMErix and inflammatory arthritis has been found in any of SB's other clinical trials. Since launch, over 1,000,000 doses of LYMErix have been distributed and no unusual patterns of adverse events have been noted in postmarketing surveillance.

The study by Croke et al. demonstrates that challenging a specific strain of hamsters with an overwhelming number of organisms induces local synovial inflammation, independent of vaccination status. We believe that the data obtained from the model cannot be extrapolated to humans and that there are currently no data to support the authors' contention that these results have implications for vaccinees.