Diphosphoryl lipid A (DPLA) obtained from the nontoxic lipopolysaccharide (LPS) of Rhodopseudomonas sphaeroides ATCC 17023 did not induce interleukin-1 release by murine peritoneal macrophages. However, it blocked this induction by toxic deep-rough chemotype LPS (ReLPS) from Escherichia coli D31m4. Previously, we obtained similar results on the induction of tumor necrosis factor (TNF) by macrophages. These results showed that DPLA is able to block in vitro the induction of two important mediators of gram-negative bacterial sepsis. We then wanted to determine whether DPLA could also block the induction of TNF by LPS in animals. Mice were treated with 100 ,ug of R. sphaeroides DPLA and challenged 60 min later with 1.0 ,g of ReLPS from E. coli. The serum TNF level was measured after 60 min. Treatment of mice with this DPLA blocked the rapid and transient rise of TNF caused by ReLPS. This result suggested that R. sphaeroides DPLA might be able to protect animals against endotoxin shock caused by gram-negative bacterial infection.

The immunogenic and adjuvant properties of Brucella abortus and Escherichia coli lipopolysaccharides (LPSs) were studied in endotoxin-responsive, athymic, and euthymic BALB/c mice and in responsive C3H/HeAu mice and congenic nonresponsive C3H/HeJ mice. Consistent with previous reports, E. coli LPS did not stimulate significant primary or secondary antibody responses in C3H/HeJ mice and induced the production of immunoglobulin M (IgM) and low levels of IgG in C3H/HeAu mice. In contrast, B. abortus smooth and rough LPS stimulated primary and secondary antibody responses and induced the production of IgM and high levels of IgG in both responsive and nonresponsive strains of C3H/He mice and in nude mice. When used as adjuvant, B. abortus LPS augmented the IgG plaque-forming-cell response of C3H/HeAu and BALB/c euthymic mice to the T-dependent antigen sheep erythrocytes. E. coli LPS augmented only the IgM plaque-forming-cell response in the same mouse strains. Neither B. abortus nor E. coli LPS was adjuvant for C3H/HeJ or nude mice. The dichotomy between the antibody and adjuvant responses of both C3H/HeJ mice and athymic mice to B. abortus LPS may be a function of the true thymus independence and dependence of these responses. In addition, the refractiveness of C3H/HeJ and nude mice to B. abortus LPS as adjuvant, but not as mitogen or polyclonal B cell activator, clearly dissociates these phenomena.

Lipopolysaccharide (LPS) of Brucella abortus strain 45/20 was purified using a novel method. Monophosphoryl lipid A (MPLA) was prepared from this LPS, methylated, and purified by high performance liquid chromatography. Chemical, mass spectral, and nuclear magnetic resonance analyses showed that MPLA consists of heptaacyl lipid As with molecular weights of 2095, 2123, 2151 and 2179. They contained the β-1,6-linked 2,3-diamino-2,3,-dideoxy-glucose disaccharide backbone and a phosphate group at the 4' position. Bisphosphoryl lipid A was also prepared and completely O-deacylated. It contained an additional phosphate group, and either 2 hydroxyhexadecanoic, 1 hydroxytetradecanoic, 1 hydroxydodecanoic acids or 2 hydroxyhexadecanoic and 2 hydroxydodecanoic acids, all in amide linkage. The predominant ester-linked fatty acyl group in acyloxyacyl linkage was hexadecanoate. The purified LPS, bisphosphoryl lipid A, and MPLA from B. abortus showed about 14%, 3% and 1%, respectively, of the B cell mitogen activity of ReLPS from Escherichia coli at 1.0 μg/ml.