Enhanced survival in Pseudomonas aeruginosa septicemia associated with high levels of circulating antibody to Escherichia coli endotoxin core

Abstract

We studied the relationship between serum antibodies to the cross-reactive endotoxin core of Escherichia coli and survival following Pseudomonas aeruginosa septicemia. Core glycolipid was purified from the outer cell membrane of a uridine diphosphate galactose 4-epimerase-deficient rough mutant E. coli (J5 strain), characterized, and used as the antigen in a quantitative enzyme-linked immunosorbent assay (ELISA) to measure core-specific IgG and IgM antibodies. 43 patients with Pseudomonas septicemia, among whom there was a mortality of 42%, were evaluated. Core-specific antibody concentrations in acute sera ranged from 1 to 49 micrograms/ml in the case of IgG and from 1 to 200 micrograms/ml for IgM. Core-specific antibodies of both isotypes were higher in patients who survived compared with those who succumbed to their septicemias (mean, microgram/ml +/- SEM, 26 +/- 3 vs. 14 +/- 4, P = 0.005 for IgG, and 55 +/- 12 vs. 18 +/- 5, P = 0.009 for IgM). Although total IgG levels were also higher in acute sera from survivors compared with nonsurvivors (mean, mg/dl +/- SEM, 1,120 +/- 99 vs. 694 +/- 119, P = 0.004), total IgM levels were virtually identical in the two groups (146 +/- 23 vs. 148 +/- 48, P = 0.52). Conversely, patients with core-specific IgG levels greater than 10 micrograms/ml at the onset of septicemia had better survival than those with levels less than 10 micrograms/ml (79 vs. 14%, P less than 0.001), and patients with core-specific IgM levels greater than 30 micrograms/ml had better survival than those with levels less than 30 micrograms/ml (81 vs. 44%, P = 0.01). In comparison, patients with total IgG levels greater than 1,000 mg/dl also had better survival than those with levels less than 1,000 mg/dl (82 vs. 42%, P = 0.01), while those with total IgM levels greater than 150 mg/dl showed somewhat less improvement in survival compared with those with levels less than 150 mg/dl (71 vs. 50%, P = 0.12). Core-specific IgM was highly correlated with core-specific IgG (r = 0.52), but not with type-specific anti-lipopolysaccharide (r = 0.13) or anti-toxin A (r = 0.12) antibodies, or with total IgG (r = 0.28) or IgM (r = 0.31). In contrast, core-specific IgG correlated somewhat more closely with type-specific antibodies (r = 0.36), and with total IgG (r = 0.51) and IgM (r = 0.52). Stepwise linear discriminant analysis indicated that type-specific antibody levels were the best predictor of outcome, among those antibodies examined, followed by anti-core IgM. Although anti-core IgG, anti-toxin A, and total IgG levels all correlated individually with survival, none augmented the prognostic power of type-specific antibodies in combination with anti-core IgM, which together predicted outcome accurately 73.5% of the time. Host factors not significantly associated with anti-core antibody levels included rapidly fatal underlying disease, age, sex, leukopenia, and prior treatment with cytotoxic drugs. In contrast, prior steroid therapy was associated with low levels of both core-specific IgG and IgM (P < 0.05). These data suggest cross-protective activity against P. aeruginosa septicemia of naturally occurring antibodies to the endotoxin core of E. coli. Anti-core antibodies, particularly of the IgM isotype appear to augment the more specific protective immunity engendered by antibodies to the O-specific side chains of Pseudomonas lipopolysaccharides. This cross-protective immunity likely applies to other Gram-negative pathogens as well.