Rabbit antisera to cross-reactive organisms--where's the problem.

SIR, The ability of rabbit antisera to certain enteric bacteria to recognise an HLA-B27 associated marker on the cells of the majority of HLA-B27 positive patients with ankylosing spondylitis (B27+AS+) has been extensively documented by our group during the past six years,' 2 but despite two confirmatory reports3 (one in preparation) the phenomenon remains controversial. The reasons for the inability of several other groups4 5 to confirm our observations have not yet been fully identified, though factors such as the use of ill-defined bacterial strains and failure to adhere to published experimental protocol may explain many of the 'negative' reports. Another explanation for our success, advanced privately by some investigators, is that the colony of rabbits used for the production of antisera is somehow unique and that we have been fortunate in eliciting antibodies to 'cross-reactive' bacterial antigens, i.e., antigens, antibodies to which lyse B27+AS+ cells. We feel this explanation is unlikely as we have successfully raised 'cross-reactive' antisera not only in 29 outbred rabbits from several different colonies but also in nine guinea pigs and 10 BALB/c mice (Table 1). Antisera have been successfully raised against (a) formalin-killed whole bacteria and sonicated bacteria (administered with either complete or incomplete Freund's adjuvant' 2 6); (b) bacterial culture filtrates (unfractionated or partially purified7 8; and (c) culture supernatant from B27+AS+ Epstein-Barr virus-transformed lymnphoblastoid cell lines (Sullivan and Geczy, unpublished observations). In our initial studies6 9 the first inoculum was given intravenously (10' to 5x 10' organisms) and subsequent injections were either subcutaneous or intramuscular, or both (108 to 101" organisms). More recently2 "' we have abandoned the intravenous route and all immunisations are now given subcutaneously or intramuscularly, or both (10' to 10"' organisms). With the exception of one rabbit which died 24 hours after an intravenous injection, all animals survived for at least six months, despite regular boosts at about monthly intervals, and all immunisation schemes were equally effective in eliciting specific 'cross-reactive' antibodies. These antibodies, in the presence of prescreened complement from a number of different commercial sources (e.g., Pel-Freeze, Behring Institut), reacted specifically with B27+AS+ cells.' All antisera were frozen in aliquots of 2-3 ml and thawed once only. During the past seven years over 50 rabbits have been immunised with either 'cross-reactive' or 'non-crossreactive' bacteria and only three animals have been encountered whose naturally occurring serum cytotoxins gave unacceptable cytotoxicity values (> 10% 3'Cr release) when tested on a panel of human lymphocytes. Table 1 Animals in which 'cross-reactive' antibodies have been raised