While various studies were reported in the early 1980's which used different individual LPS such as E.coli J5 or Salmonella minnesota Re to measure anti-endotoxin-core antibodies, our studies with these antigens amongst a panel of other LPS antigens did not convince us that they were ideal for this purpose.
Our first difficulty was in obtaining reproducibility for the ELISA. The R-LPS was reluctant to bind stably to polystyrene ELISA microtiter plates. We explored various techniiques, and finally tried complexing the R-LPS with polymyxin. We found that for around 50% of individuals the antibodies in their plasma bound equally well to R-LPS whether coated directly or whether complexed with polymyxin. However, for the other 50% their antibodies bound well only to R-LPS complexed with polymyxin
We interpreted this as indicating that certain antibodies could only bind to LPS when the structure was opened out when bound to polymyxin, but not in the form (probably self associated complex) that purified LPS takes up. Purified LPS does not occur in nature - but LPS complexed with many other things does. Since the antibodies were natural - already existing in plasma - we assume they are important anti-LPS antibodies. Apparently to detect all anti-LPS-core antibodies the R-LPS complexed with polymyxin is preferable to R-LPS alone. The polymyxin also stabilised the binding of the LPS to the plates, allowing good reproducibility.
We conducted a number of studies using different LPS (complexed with polymyxin). We found that around 50% of individuals have anti-LPS-core antibodies which bind to Re-LPS, the least complete core structure comprised of lipid A and KDO sugars only (see LPS structures - under construction). The remaining individuals have anti-LPS-core antibodies which bind only to the Rc-LPS, which in addition to the Re structure has the heptose sugars and one hexose (complete inner core). These antibodies bind equally well to the complete core R-LPS (Ra LPS).
The colours are coded according to whether these same sera cross-reacted on E.coli J5 LPS and E.coli O111 LPS (green) or reacted only with E.coli O111 LPS (blue), See Barclay & Scott (Infect. Immun. 1987).
Thus, to pick up all LPS-core cross-reactive antibodies, the minimal structure required is the Rc structure. Complete core (Ra) structures are to be avoided, since they contain the more heterogeneous outer cores (e.g. R1, R2, R3, R4 and K12 for E.coli) to which specific antibodies can undoubtedly by found in different individuals, so incomplete core structures such as Rc are sufficient and best. However, we wished to have some element of cross-species LPS-core cross reactivity (we had noted E.coli - pseudomonas aeruginosa - Salmonella core cross reactivity in screening). We therefore sought incomplete outer-core R-LPS with intact inner-core (lipid A - KDO - heptose) from each of E.coli; Pseudomonas aeruginosa; Salmonella; and Klebsiella for the ELISA. A single suitable R-LPS from each species was complexed with polymyxin, combined in equimolar proportions in a cocktail, and used to coat the ELISA plates for the EndoCAb ELISA.