Immunobiol., vol. 187, pp. 430-446 (1993) Department of Membrane Biochemistry, Walter Reed Army Institute of Research, Washington, DC, USA
Lipopolysaccharide, Lipid A, and Liposomes Containing Lipid A as Immunologic Adjuvants CARL R. ALVING
Abstract Numerous studies have demonstrated that most or all of the potent adjuvant activity of Gram-negative bacterial endotoxin resides in the lipid A moiety of lipopolysaccharide (LPS). Synthetic analogues of lipid A have provided insights into structure-activity relationships. Several cellular mechanisms of LPS and lipid A adjuvant activities have been identified. Activation of macrophages by LPS or lipid A results in cytokine secretions that enhance the immune response. LPS and lipid A cause recruitment of antigen-presenting cells, particularly macrophages. Liposomes containing lipid A serve as an in vivo adjuvant to recruit increased numbers of macrophages. Liposomal lipid A that has been phagocytized by cultured macrophages also serves as an «intracellular adjuvant» to cause increased immunologic presentation of liposomal antigen by the macrophages to specific T lymphocytes. Lipid A can abolish suppressor T cell activity, resulting in increased immune responses to polysaccharide antigens. Upon combination of lipid A or lipid A analogues with nonionic block polymers, modulation of muri~e antibody isotypes can be achieved with antibodies against a variety of antigens in vivo. Liposomes containing monophosphoryllipid A (MPL) have been utilized in a phase I clinical trial of a proposed malaria vaccine in humans. The liposomal malaria vaccine resulted in very high levels of antibodies against the malarial antigen, and despite the presence of huge amounts of MPL (up to 2.2 mg), the liposomallipid A was nonpyrogenic and safe for use in humans. Lipid A and lipid A analogues, and liposomes or other carriers containing lipid A, have shown considerable promise both as adjuvants for immunization of animals and for human vaccines.
Adjuvant Activities of Lipopolysaccharide and Lipid A The ability of Gram-negative bacterial lipopolysaccharide (LPS) to serve as an adjuvant to enhance immune responses to antigens has been recognized for more than 30 years (1-4). Lipopolysaccharide, which itself is also highly immunogenic (5), has been proposed to be a powerful immunogen because «it is an antigen [polysaccharide] that carries its own adjuvant [lipid A]» (6). Among the numerous biological activities of LPS, the adjuvant activity, one of the most interesting activities from the standpoint of potential practical applications, has been shown to reside in the lipid A moiety of LPS (7). The adjuvant activity of lipid A has been confirmed in many laboratories (8-12), and adjuvant effects have also been demonstrated with synthetic
Lipopolysaccharide, Lipid A, and Liposomes . 431
analogues of lipid A (13-15). Experiments with chemically modified LPS molecules have led to the conclusion that only the lipid A portion is required for adjuvanticity, but adequate amounts of esterified and amidated lipid A fatty acids are also required (9).
Mechanisms of Adjuvant Activity
Insights from synthetic lipid A analogues The availability of synthetic lipid A analogues has permitted analysis of minimal structural requirements for adjuvant activity (16, 17). Studies with synthetic' analogues have suggested that disaccharide and monosaccharide structures are both active as adjuvants. In contrast, the fatty acid composition, including stereospecificity of fatty acids, is extremely critical for expression of adjuvant activity of lipid A (17, 18). The number and composition of lipid A fatty acids would be expected to have an important effect on the structure and solubility of lipid A in an aqueous environment and on the interactions of lipid A with membranes of cells and liposomes or other carriers, and with antigen present in liposomes or other carriers. Lipid A fatty acids therefore might influence the adjuvant effects of liposomal lipid A in vaccine formulations that utilize liposomes as carriers. However, such influences of lipid A fatty acids cannot be easily predicted, and the relative efficacies of lipid A analogues can only be determined by empirical testing in candidate vaccine formulations.
Role of macrophages Early studies with radioactively-labelled LPS and lipid A demonstrated that after parenteral injection both compounds rapidly left the blood and accumulated in macrophages in the liver and spleen (19). Experiments involving transfer of murine macrophages from donor animals to syngeneic recipient animals engendered, at least partly, the present widely held concept that the adjuvant effect of LPS is related to its effect on macrophages. Donor macrophages that had been allowed to ingest antigen in vitro enhanced the immune response to the antigen in the recipient animals, and enhancement was greatly stimulated by pretreating the macn5p hages with LPS (20-22). In retrospect, in the light of current knowledge, these early experiments involving injection of cultured macrophages, even syngeneic macrophages, probably should be considered to be inconclusive evidence because of structural and chemical changes in the macrophages during culture that might cause the macrophages themselves to be viewed as foreign cells by the recipient animals. The adjuvant effects of LPS and lipid A on macrophages have now been well established. Interactions of. LPS with macrophages r~sult in a state qf macrophage activation, a condition which is manifested by myriad synthesis and secretory activities, and signal transduction mechanisms (23). In,
432 . C. R.
addition to the activation of macrophages, it is also recognized that LPS and lipid A stimulate multiple cellular interactions, including the generation of cytokines that causes recruitment of additional macrophages (24). Interactions of macrophages with T lymphocytes in the process of antigen Stimulates or suppresses other T cells, B cells and macrophages; kills infected cells
Differentiates ancJ produces antibodies
IL·2 IL·4 IL·S IL-6 INF"'(
Produces new blood cells Figure 1. Secretion of mediators by an LPS·stimulated macrophage (center) sets in motion further secretory stimuli and biological effects by numerous cell types. Of particular interest for adjuvant development for vaccines are the effects of tumor necrosis factor (TNF), interleukin (IL)-l, IL-2, IL-4, IL-5, and IL-6, growth factor (TGF-~), and interferon-y (IFNy), on proliferation and differentiation of T and B lymphocytes that are crucial elements in the process of antigen presentation. Redrawn and based on information from Ref. 26.
Lipopolysaccharide, Lipid A, and Liposomes . 433
presentation has been proposed as a major basis for adjuvant effects of LPS (25). The cellular interactions induced by LPS and lipid A are mediated by numerous secreted cytokines. Some of the complex interactions that occur between mediators and cells as a result of LPS and lipid A stimulation are illustrated in Figure 1.
Effects of LPS on antigen presentation
In recent years the concept of the macrophage as an antigen-presenting cell (APC) has emerged (26). A key element in the process of antigen presentation is the participation of major histocompatibility gene complex (MHC) molecules. For induction of an immune response against foreign antigens, class II MHC molecules (also known as la molecules) are expressed by the APC, and recognition by helper T cells of a complex of processed antigen with la on the surface of the APC is the initial event leading to an immune response (27). It has been demonstrated that LPS can stimulate induction of la expression by murine macrophages (28-31). Cultured macrophages that were in.cubated in the absence of cytokines did not exhibit la expression. It was th~refore proposed that induction of high levels of la-positive macrophages were one consequence of the adjuvant action of LPS and this effect required participation of another cell found in athymic mice, along with secretion of appropriate mediators (31). Induction of la-positive macrophages therefore promotes the increased numbers of antigen-presenting cells and thereby contributes to the overall adjuvant effect of LPS (31). Among other adjuvants tested, it is of interest that beryllium, complete Freund's adjuvant (31), and a nonionic block copolymer (32) also induced the appearance of la-positive macrophages, but appearance of la cells was not promoted by muramyl dipeptide, poly I-C, dextran sulfate, or incomplete Freund's adjuvant (31). Lipid A as an intracellular adjuvant A novel hypothetical mechanism for understanding some of the adjuvant effects of lipid A has been recently proposed. It was demonstrated that liposomal lipid A can serve as an intracellular adjuvant for promoting enhanced antigen presentation by macrophages. An in vitro model system was utilized in which bone marrow-derived macrophages were allowed to phagocytize liposomes containing both encapsulated malaria antigen (R32NS1) and lipid A (from Salmonella minnesota R595) (33). Based on immunogold electron microscopy with a monoclonal antibody by R32NSl it was concluded that phagocytized liposomal R32NSl became distributed in large intracellular vacuoles, following which liposomes in the vacuoles apparently fused with the vacuolar membranes resulting in the extrusion of large amounts of liposomal antigen into the cytoplasm (33). After phagocytosis of the liposomal R32NSl by macrophages, the cells were incubated with T cells from a cell line that recognized R32NSl, and antigen presentation was determined by proliferation of the T cells (24). As shown in Figure
434 . C. R.
< Ii:M~ ::;).
w!!: >< 9 E ::;:0.
~ ~ 20 ~
0.4 0.8 1.2 LlPOSOMAL ANTIGEN (!lg/ml)
Cell Density: 1
x 10 6 cells
o E ~ ~10
L(Ag+LA) L(Ag) L(LA) (control)
LlPOSOMAL ANTIGEN (!lg/ml)
Figure 2. Immunologic presentation of phagocytized liposome-encapsulated malaria by murine bone marrow-derived macrophages in viero, as determined by proliferation of cultured T lymphocytes specific for the malaria antigen. C47BL/I0 bone marrow-derived macrophages were seeded at 105 (A) or 106 (B) cells per well, and then pulsed with liposomes containing antigen [L(Ag)], or liposomes containing both antigen and lipid A [L(Ag + LA)], for 90 min prior to addition of the specific T cell clone. During the last 16 h of the 72 h incubation period, 1 !-lCi of eH]thymidine was added. See (24) for further details.
2, presentation of lphagocytized liposomal antigen to T cells was markedly enhanced by the presence of lipid A in the liposomes. Presentation of liposomal antigen was inversely related to the density of macrophages that were originally seeded in the culture. At the highest cell density (1 x 106 cells), presentation of liposomal R32NS1 was more than twenty-fold higher when the phagocytized liposomes contained l1pid A, thus leading to the concept of lipid A as an intracellular adjuvant. ,
Lipid A as an extracellular adjuvant: recruitment of antigen-presenting cells
Liposomes containing or lacking lipid A were injected'intraperitoneally into mice to be tested for the ability to recruit polymdtphonudear leukocytes (PMNs) and macrophages. As shown in Table 1, lipbsomes containing R32NS1 antigen induced modest increases in PMNs and peroxidase-posi-
3.0 ±2.0 4.0 ±4.0 56.0 ± 1.0 4.0 ± 0.8 11.0 ± 6.0 55.0 ±6.0
0 4.0 ± 1.0 11.0 ± 5.0
0.1±0.1 0.8 ± 0.2 7.0 ± 0.2
% of peroxidase-positive cells/totalleukocytes
r -B' o
r -B' is.:
0.7 ± 0.6 17.0 ± 4.0 60.0 ± 14.0
% of peroxidase-positive macrophages/total macrophages
"Five C3H/HeN mice per group were injected intraperitoneally with 0.1 ml volumes of PBS, liposomes containing R32NS1 [L(Ag)], and liposomes containing R32NS1 and lipid A [L(Ag + LA)]. Data are means ± standard deviations. All values for the three groups were significantly different at 6 h (p < 0.05; two-tailed Student's t test). From Ref. 24.
54.8 ± 7.6 60.6 ± 5.6 79.3 ±.1.6
% of polymorphonuclear leukocytes/totalleukocytes
0 100 100
0 0 20
Spleen wt (mg)
0 0.1 ± 0.1 0.4 ± 0.1
Table 1. Recruitment of cells after intraperitoneal injection of liposomes containing or lacking lipid LA"
436 . C. R.
tive macrophages. However, when the liposomes contained lipid A, massive recruitment of both cell types occurred. Recruitment of new macrophages does not guarantee increased antigen presentation since, as noted earlier, presentation can vary inversely with cell density (Fig. 2). It therefore appears that liposomallipid A has dual effects: (a) recruitment of a high density of macrophages, a process that also may actually cause decreased presentation by each cell; and (b) stimulation of increased presentation by APCs at high cell density, a process that occurs through an intracellular adjuvant effect of liposomallipid A. In addition to recruiting and stimulating macrophages to enhance the immune response against antigens, LPS also elicits macrophages that cause increased antigen catabolism. This latter process can lead to down-regulation of the immune response (34).
Effects on suppressor T cells The immune respo~se to bacterial polysaccharide is greatly amplified by the use of monophosphoryl lipid A (MPL) (35-38). At least part of the enhancing effect of MPL was attributed to the inactivation of suppressor T cell activity, and this may represent a further adjuvant mechanism for lipid A. Although the immune response to pneumococcal polysaccharide was equally strong in LPS-responsive (C3H/HeN) and LPS-nonresponsive (C3H/HeJCR) strains of mice, abolishment of expression of suppressor T cell activity occurred only in the LPS responsive strain of mice (36). Synthetic analogs of lipid A were employed to determine the minimum structure of lipid A required for abolishment of suppressor T cell activity (37). It was concluded that in order to retain this activity lipid A (a) must contain a glucosamine disaccharide, (b) must have at least one phosphate, and (c) must have at least five fatty acyl groups. It was further concluded that the chain length of the nonhydroxylated fatty acid and the location of acyloxyacyl groups (2' vs 3' position) may play important roles.
Modulation of antibody isotype LPS and lipid A can be used as adjuvants for influencing the isotype of induced murine antibodies. In mice, IgG2a is the most desirable antibody isotype and IgG 1 is the least desirable isotype. The Ribi adjuvant system (RAS) (consisting of MPL and trehalose dimycolate in a 2 % oil-in-water emulsion with Drakeol, 0.2 % Tween 80/PBS, and antigen) induced predominantly IgG 1 antibodies against human serum albumin (39). Lesser levels of IgG2a, IgG2b, and IgG3 were obtained. When compared with other adjuvants, RAS was equivalent to Freund's adjuvant (complete followed by incomplete) in the production of IgGl and IgG2b antibodies. The isotype of specific antibodies against LPS itself is strongly influenced by the carrier system (40). Mixtures of LPS with phospholipids or hydrophobic proteins induced predominantly IgG 1 antibodies, but mixtures of LPS with hydrophilic proteins induced predominately IgG2 antibodies.
Lipopolysaccharide, Lipid A, and Liposornes . 437
Nonionic block copolymers in oil-in-water emulsions (squalane-inwater) have been proposed as adjuvants for a variety of hapten, peptide, and polysaccharide immunogens conjugated to protein carriers (41-45). One of the most interesting properties of these copolymers is that the IgG subclass of specific murine antibodies can be strongly influenced both by the copolymer/oil emulsion alone, and by copolymer/oil emulsion in combination with detoxified LPS or a variety of lipid analogues. Different LPS or lipid A analogues had variable effects on the antibody titer, but most of the preparations had beneficial effects on isotype in that they tended to increase the proportions of IgG2a and IgG2b while suppressing IgGl (42-44). Although the LPS or lipid A analogues influenced the isotype pattern in the manner described when used alone, a synergistic effect was found when they were combined with the copolymer emulsions.
Safety of lipid A as an adjuvant for vaccines Many of the most successful vaccines developed in the past two hundred years, including smallpox (18th century) and tetanus toxoid (19th century) vaccines have had annoying but tolerable reactogenic characteristic (46). One of the most frequently observed adverse reactions to vaccines is pyrogenicity, a property that is commonly associated with LPS. Although the evolution of modern vaccine technology has now led to proliferation of myriad peptide and recombinant protein antigens that have inherently high levels of safety because of their purity, simplified soluble protein antigens do not necessarily have high levels of inherent immunogenicity. It is well known that different biological effects of LPS and lipid A can be related to different chemical moieties in the structure of LPS and lipid A. It has been argued that certain so-called «beneficial» effects may be identified with chemical structures that are different than the structures associated with «toxic» effects (47). Although the complexity of LPS, lipid A, and their biological effects precludes simplified generalizations, the concept that nontoxic structures might be produced that retain the potent adjuvant effects of lipid A has generated optimism and interest in lipid A among investigators in the field of vaccinology. RIBl and his colleagues have demonstrated that a monophosphoryl lipid A (MPL) formulation exhibits reduced lethality and pyrogenicity but still retains antitumor and adjuvant activities (11, 12, 48, 49). In the same context, it has also been a major goal in the lipid A biosynthesis field to produce lipid A analogues that would retain beneficial biological activities, including adjuvanticity, but lack endotoxic activity (13, 15, 17, 18, 50y: Intravenously administered MPL has been tested in a phase I trial in humans (51). The maximal safe intravenous dose in humans was estimated to be 100 f,tg/m2, but even at 25 f,tg/m2 minor episodes of fever, gastrointesti-, nal symptoms, and chills were often observed (51). Most vaccine formulations are given intramuscularly or subcutaneously', ~nd it is possible that these latter routes of administration might result in less reactogenicity than the intravenous route. " 01: 1• •
438 . C. R.
Reduction of toxicity by using Jiposomes containing lipid A One approach that has been employed in an effort to widen the gap between reactogenicity and adjuvanticity has been to incorporate LPSand lipid A into liposomes. Among the activities of lipid A that are reduced by incorporation into liposomes are included: Limulus lysate coagulatlOn in vitro (52-54), neutropenia in rabbits (52), stimulation of both interleukin-1 (54), and tumor necrosis factor (55) by a macrophage cell line, pyrogenicity both in rabbits (56, 57) and in humans (58) and lethal toxicity to mice (59). As noted earlier, one of the most common complaints associated with many vaccines is the occurrence of fever. In many cases pyrogenic reactions are due to the presence of LPS or lipid A, either as a contaminant or as' an antigen, in the vaccine. Preclinical studies demonstrated that a formulation of liposomal lipid A, in which the liposomes containing lipid A were adsorbed to aluminum hydroxide (alum), had markedly reduced pyrogenic activity in rabbits but retained potent adjuvant activity. This basic formulation was therefore proposed as a carrier and adjuvant for human vacdnes (60). Liposomes as Carriers of Vaccines Because of the central role played by macrophages in adjuvant mechanisms of LPS and lipid A, a logical approach would be to promote increased delivery of LPS or lipid A to macrophages. One of the most effective methods currently being employed to deliver therapeutic substances to macrophages is to use liposomes as carriers (61-63). Liposomes have been highly successful as carriers of antigens and adju-, vants (64). The theoretical basis for interactions of liposomes with ~the immune system and for enhancement of immune responses against liposomal protein antigens has been extensively reviewed elsewhere· ,(64, 65). Liposomes provide a means to reconstitute hydrophobic antigens, or:to encapsulate soluble antigens or peptides. They also provide a solid surface for the chemical attachment of antigens. Intravenously-injected liposomes are rapidly removed from the blood and are avidly ingested by macrophages in the liver and spleen where they undergo gradual degradation. The fate of intramuscularly or subcutane~' ously injected liposomes is dependent to a great extent on the scavenging, characteristics of local tissue macrophages or macrophages recruited to the site in response to an inflammatory or adjuvant stimulus. For example, liposomal antigen (adenovirus hexon) that was injected intramuscularly was slowly released from the injection site with a half-life for clearance of 30-44 hours (66). It is therefore evident that a transient sustained release or depot effect contributes to the mechanism by which intramuscularly-injected~ liposomal antigen interacts with the immune system. The depot effect of liposomes probably does not substantially enhance the adjuvant effect of lipid A compared to that observed with lipid A alone since it has been
Lipopolysaccharide, Lipid A, and Liposomes . 439
reported that subcutaneously injected lipid A alone has a half-life at the injection site of more than 15 days (67). As noted earlier, liposomes also attenuate the toxic effects of lipid A but the adjuvant effect of lipid A is retained even in the absence of pyrogenicity or other types of reactogenicity. It is therefore possible that the blocking of the toxicity of lipid A by liposomes is partly due to retardation of release of lipid A into the environment and also due to degradation of liposomallipid A along with the liposomes in macrophages.
Adjuvant effects of liposomes containing lipid A Early studies demonstrated that lipid A enhanced the murine immune responses against liposomal haptens consisting of lipid derivatives of either 2,4-dinitrophenyl or fluorescein groups (8, 68). It was soon suggested that liposomes containing lipid A (or liposomes containing LPS) could stimulate immune responses against liposomal proteins (10, 69), or even against lipid A as an antigen (10). Lipid A proved to be such a potent adjuvant that antibodies were even induced against both the liposomes themselves and liposomal phospholipid constituents (70-71).
10,000 9000 8000
5000 4000 3000 2000 1000
WEEKS AFTER PRIMARY IMMUNIZATION Figure 3. Immune response against cholera toxin (CT) alone, CT associated with complete Freund's adjuvant (CFA), liposomes, or liposomes containing lipid adjuvant. Each line is the mean of two rabbits immunized either with 5!-1g CT alone (i.v. injection), or with 5!-1g CT either emulsified with CFA (s.c. injection), attached to the surface of liposomes with adjuvant, or attached to the surface of liposomes containing either lipid A or one of two forms of lipophilic MDP, CH3(CHz)16CO-MDP (stearoyl-MDP, or L18-MDP) or [CH 3(CH z)13]ZCHCO-MDP (B30-MDP) (i.v. injection). The maximum measurable level of RIA units in the assay was 107 • Therefore for those data points for CFA and lipid A in which this latter value is illustrated the actual values achieved were probably higher. From Ref. 55.
440 . C. R. ALVING
Liposomes containing lipid A have been used as adjuvants for enhancing immune responses to: cholera toxin (10, 72, 73); herpes simplex antigens, including glycoprotein-enriched herpes antigens (74) and a herpes peptidepalmitic acid conjugate (75); Epstein-Barr virus membrane antigens (76); Plasmodium falciparum sporozoite antigens, including a Plasmodium peptide-protein conjugate (72), and a recombinant protein containing Plasmodium epitopes (56, 77); an unconjugated 25 amino acid peptide from the active site of acetylcholinesterase (78); and tumor-associated glycosphingolipids (79). Recently, Neisseria meningitidis LPS was used as a liposomal antigen, resulting in increased immunogenicity and lOOo-fold reduction of pyrogenicity of the LPS (57). One of the difficulties that is often encountered in adjuvant research lies in evaluating the relative efficacy of an adjuvant in comparison with other adjuvants. There are no standardized techniques available for this purpose because of the existence of multiple adjuvant mechanisms and immunological techniques. Different chemical and physical structures of adjuvants often precludes standardization of all of the different possible variables. Comparative evaluation of adjuvants such as lipid A, analogues of lipid A, LPS, and muramyl dipeptide can be a heroic undertaking (80). However, some of these difficulties can be overcome with liposomes. Many variables can be simultaneously controlled by utilizing liposomes containing both the antigen and the adjuvant. In one study, direct comparison of the adjuvant effects of liposomal lipid A with liposomal lipophilic MDP derivatives demonstrated that, under the conditions employed, lipid A had stronger adjuvant effects than lipophilic MDP (Fig. 3) (72).
Human liposomal vaccine The first injectable liposomal vaccine was administered to human volunteers in a phase I trial in October, 1989 (58). The liposomes contained a recombinant protein antigen having repeat sequence epitopes from the circumsporozoite protein of Plasmodium falciparum. Monophosphoryl lipid A was included in the liposomes as an adjuvant, and the liposomes containing antigen and MPL were also adsorbed to aluminum hydroxide to prolong the depot effect and provide further adjuvant activity. At the highest dose tested, 2.2 mg of liposomal MPL was administered with each intramuscular injection of vaccine (at 0, 12, and 20 weeks). Despite the extremely high doses of liposomal MPL - approx. 12-fold greater than the maximum safe dose of intravenous free MPL in humans (51) - the liposomal MPL exhibited virtually no acute toxic effects. As mentioned earlier, MPL is less pyrogenic than native lipid A when tested in a rabbit pyrogenicity model. In experiments designed to explore the roles of physical environment and chemical structure on pyrogenicity, MPL was approximately 40-fold less pyrogenic than the native lipid A from which the MPL was derived (max. nonpyrogenic dose of 0.32 compared to 0.008 [Lg/kg) (Table 2). Liposomes further reduced the pyrogenicity by at least 25-fold (Table 2), thus causing liposomal MPL to be lOoo-fold less
Lipopolysaccharide, Lipid A, and Liposomes . 441 Table 2. Effect on pyrogenicity of incorporating native lipid A and monophosphoryllipid A in liposomes"
Lipid A type
s. minnesota R 595 MP lipid A (MPL),,"'f".
Maximum non pyrogenic dose*'f Decrease in - - - - - - - - - - - - - - pyrogenicityof Liposomal Lipid A Free Lipid A liposomal compared to free lipid A 0.008 0.32
" Data obtained from Ref. 56. 'f"· Maximum nonpyrogenic dose was the maximum lipid A dose in !!g/kg that did not cause an increase in temperature of 0.6 °C or greater in any of three rabbits over a 3-h period. 'f"·"·MP lipid A is a monophosphoryllipid A fraction derived from s. minnesota R 595 lipid A.
pyrogenic than free native lipid A. Vaccine-grade liposomal MPL prepared for human use according to Good Manufacturing Practices (GMP) as promulgated by the U.S. Food and Drug Administration, was 8-fold less pyrogenic yet (max. nonpyrogenic dose of 66 Ilg of liposomal MPLlkg) (unpublished data). Inclusion of aluminum hydroxide (alum) as a nonliposomal adjuvant reduced the pyrogenicity of GMP liposomal MPL still further (approx. lO-fold more) (unpublished). It is therefore estimated that GMP liposomal MPL adsorbed to alum may be approximately 80,OOO-fold less pyrogenic than native lipid A alone. Despite the virtual elimination of pyrogenicity, liposomal MPL still served as an extremely potent adjuvant for inducing humoral immunity against liposomal protein antigens both in animals (56) and humans (58) (Fig. 4). As shown in Figure 4, the highest vaccine dose employed in humans induced a geometric mean level of 33 Ilg of IgG anti-NANP antibodies/m!. In contrast, the same antigen adsorbed to aluminum hydroxide induced only 2.3 Ilg of IgG/ml (dashed line in Fig. 4). The previously expressed concern that lipid A in liposomes would serve as an adjuvant only at pyrogenic concentrations, and that lipid A would thus be unacceptable for use in humans (81), has therefore been proven to be unfounded. The strong adjuvant effects of nonpyrogenic formulations of liposomallipid A are also consistent with the observations mentioned earlier that adjuvant effects of certain synthetic lipid A analogues can be dissociated from pyrogenic effects. The success of the human liposomal vaccine in providing exceptionally high levels of serum antibodies validates the concept that lipid A can be combined with other types of carriers and adjuvants to provide additive adjuvant activities in humans. The high doses of MPL employed also demonstrates that potent adjuvant effects can be achieved in the absence of toxicity. Because of the success of this initial vaccine for inducing enhanced immunity, additional candidate human liposomal vaccines employing other antigens and MPL have been proposed, and two additional malaria vac-
442 . C. R. ALVING 300 ,.... 250
S 200 ........ till
Geometric Mean: 23
40 30 20 10
Group Number Figure 4. R32-specific IgG levels in human subjects 8 weeks after the last of three doses of MPL-bearing liposome-encapsulated R32NSl sl . Each vertical bar represents a single volunteer studied by R32-specific ELISA. The groups and doses injected were as follows: groupl, 1:100 dilution of liposomes, containing 6.3 f-Ig of R32NSl sl and 22 f-Ig of MPL; group 2, 1:10 dilution of liposomes, containing 63 f-Ig of R32NSl sl and 220 f-Ig of MPL; group 3, 1:4 dilution of liposomes, containing 158 f-Ig of R32NSl sl and 550 f-Ig of MPL; group 4, 1:2 dilution of liposomes, containing 315 f-Ig of R32NSl sl and 1100 f-Ig of MPL; group 5, undiluted liposomes containing 630 f-Ig of R32NSl sl and 2200 f-Ig of MPL. Dashed line indicates the geometric mean peak R32-specific IgG level (2.3 f-Ig/ml) attained in a previous trial after four injections of 1150 f-Ig of R32NSl sl adsorbed to AI(OH3) alone. The geometric mean value for each dose group in the present study is indicated above the bar graph. From Ref. 58.
cines, and an HIV (gp 120) vaccine are scheduled for initiation of phase I clinical trials in late 1992 and early 1993.
References 1., JOHNSON, A. G., S. GAINES, and M. LANDY. 1956. Studies on the 0 antigen of Salmonella
4. 5. 6.
typhosa. V. Enhancement of antibody response to protein antigens by the purified lipopolysaccharide.]. Exp. Med. 103: 225. FRANZL, R. E. and P. D. McMASTER. 1986. The primary immune response in mice. 1. The enhancement and suppression of hemolysin production by a bacterial endotoxin. ]. Exp. Med. 127: 1087. ADA, G. L., P. G. LANG, and G. PLYMIN. 1968. Antigen in tissues. V. Effect of endotoxin on the fate of, and on the immune response to, serum albumin and to the albuminantibody complexes. Immunology 14: 825. SEPpALA, 1. ]. T. and O. MAKELA. 1984. Adjuvant effect of bacterial LPS and/or alum precipitation in responses to polysaccharide and protein antigens. Immunology 53: 827. RUDBACH,]. A. 1971. Molecular immunogenicity of bacterial lipopolysaccharide antigens: establishing a quantitative system. ]. Immuno!. 106: 993. CHiLLER,]. M., B.]. SKIDMORE, D. C. MORRISON, and W. O. WEIGLE. 1973. Relationship of the structure of bacterial lipopolysaccharides to its function in mitogenesis and adjuvanticity. Proc. Natl. Acad. Sci. USA 70: 2129. ALVING, C. R. 1992. Lipid A and liposomes containing lipid A as adjuvants for vaccines.
Lipopolysaccharide, Lipid A, and Liposomes . 443
In: Bacterial Endotoxic Lipopolysaccharides, Vo!' II: Immunopharmacology and Pathophysiology, J. RYAN and D. C. MORRISON (eds.). CRC Press, Boca Raton, p. 429. DANCEY, G. F., T. YASUDA, and S. C. KINSKY. 1977. Enhancement of liposomal model membrane immunogenicity by incorporation of lipid. A. J. Immuno!. 119: 1868. CHO, Y., K. TANAMOTO, Y. OH, and J. Y. HOMMA. 1979. Differences of chemical structures of Pseudomonas aeruginosa lipopolysaccharide essential for adjuvanticity and antitumor and interferon-inducing activities. FEBS Lett. 105: 120. ALVING, C. R., B. BANER]I, J. D. CLEMENTS, and R. L. RICHARDS. 1980. Adjuvanticity of lipid A and lipid A fractions in liposomes. In: Liposomes and Immunobiology, B. H. TOM and H. R. SIX. Elsevier/North-Holland, Amsterdam, p. 67. JOHNSON, A. G., M. TOMAI, L. SOLEM, L. BECK, and E. RIB!. 1987. Characterization of a nontoxic monophosphoryllipid A. Rev. Infect. Dis. 9 (Supp!. 5): S512. BAKER, P. J., J. R. HIERNAUX, M. B. FAUNTLEROY, P. W. STASHAK, B. PRESCOTT, J. L. CANTRELL, and J. A. RUDBACH. 1988. Ability of monophosphoryllipid A to augment the antibody response of young mice. Infect. Immun. 56: 3064. KUMAZAWA, Y., M. MATSUURA, J. Y. HOMMA, Y. NAKATSURU, M. KISO, and A. HASEGAWA. 1985. B cell activation and adjuvant activities of chemically synthesized analogues of the nonreducing sugar moiety of lipid A. Eur. J. Immuno!. 15: 199. KOTANI, S., H. TAKADA, 1. TAKAHASHI, T. OGAWA, M. TSU]IMOTO, H. SHIMAUCHI, T. IKEDA, H. OKAMURA, T. TAMURA, K. HARADA, S. TANAKA, T. SHIBA, S. KUSUMOTO, ana T. SHIMAMOTO. 1986. Immunobiological activities of synthetic lipid A analogs with low endotoxicity. Infect. Immun. 54: 673. TAKADA, H. and S. KOTAN!. 1989. Structural requirements of lipid A for endotoxicity and other biological activities. CRC Crit. Rev. Microbio!. 16: 477. YASUDA, T., S. KANEGASAKI, T. TSUMITA, T. TADAKUMA, J. Y. HOMMA, M. INAGE, S. KUSUMOTO, and T. SHIBA. 1982. Biological activity of chemically synthesized analogues of lipid A. Demonstration of adjuvant effect in hapten-sensitized liposomal system. Eur. J. Biochem. 124: 405. TAKAHASHI, 1., S. KOTANI, H. TAKADA, M. TSU]IMOTO, T. OGAWA, T. SHIBA, S. KUSUMOTO, M. YAMAMOTO, A. HASEGAWA, M. KISO, M. NISHI]IMA, F. AMANO, Y. AKAMATSU, K. HARADA, S. TANAKA, H. OKAMURA, and T. TAMURA. 1987. Requirement of a properly acylated f3(1-6)-D-glucosamine disaccharide bisphosphate structure for efficient manifestation of full endotoxic and associated bioactivities of lipid A. Infect. Immun. 65: 57. KUMAZAWA, Y., M. NAKATSUKA, H. TAKIMOTO, T. FURUYA, T. NAGUMO, A. YAMAMOTO, J. Y. HOMMA, K. INADA, M. YOSHIDA, M. KISO, and A. HASEGAWA. 1988. Importance of fatty acid substituents of chemically synthesized lipid A-subunit analogs in the expression of immunopharmacological activity. Infect. Immun. 56: 149. HOWARD, J. G., D. ROWLEY, and A. C. WARDLAW. 1958. Investigations on the mechanism of stimulation of non-specific immunity by bacteriallipopolysaccharides. Immunology 1: 181. UNANUE, E. R., B. A. ASKONAS, and A. C. ALLISON. 1969. A role of macrophages in the stimulation of immune responses by adjuvants. J. Immuno!. 103: 71.
21. SPITZNAGEL, J. K. and A. C. ALLISON. 1970. Mode of action of adjuvants: effects on antibody responses to macrophage-associated bovine serum albumin. J. Immuno!. 104: 128. 22. ALLISON, A. C. 1979. Mode of action of immunological adjuvants. J. Reticuloendothelial Soc. 26 (Dec. Supp!.): 619. 23. KARNOVSKY, M. L. and J. K. LAZOINS. 1978. Biochemical criteria for activated macrophages. J. Immuno!. 121: 809. 24. VERMA, J. N., M. RAo, S. AMSELEM, U. KRZYCH, C. R. ALVING, S. J. GREEN, and N. M. W ASSEF. 1992. Adjuvant effects of liposomes containing lipid A: enhancement of liposomal antigen presentation and recruitment of macrophages. Infect. Immun. 60: 2438. 25. MCGHEE, J. R., J. J. FARRAR, S. M. MICHALEK, S. E. MERGENHAGEN, and D. L. ROSENSTREICH. 1979. Cellular requirements for lipopolysaccharide adjuvanticity. A roYe for
444 . C. R. ALVING
26. 27. 28. 29. 30.
31. 32. 33. 34. 35. 36.
both T lymphocytes and macrophages for in vitro responses to particulate antigens. J. Exp. Med. 149: 793. OLD, 1. J. 1988. Tumor necrosis factor. Sci. Am. 258: 59. UNANUE, E. R. and P. M. ALLEN. 1987. The basis for the immunoregulatory role of macrophages and other accessory cells. Science 236: 551. WENIWORTH, P. A. and H. K. ZIEGLER. 1987. Induction of macrophage la expression by lipopolysaccharide and Listeria monocytogenes in congenitally athymic nude mice. J. Immunol. 138: 3167. MARSHALL, N. E. and H. K. ZIEGLER. 1989. Role of lipopolysaccharide in induction of Ia expression during infection with gram-negative bacteria. Infect. Immun. 57: 1556. WENIWORTH, P. A. and H. K. ZIEGLER. 1989. Modulation of macrophage Ia expression by lipopolysaccharide: stem cell requirements, accessory lymphocyte involvement, and lA-inducing factor production. Infect. Immun. 57: 2028. BEHBEHANI, K., D. I. BELLER, and E. R. UNANUE. 1985. The effects of beryllium and other adjuvants on la expression by macrophages. J. Immunol. 134: 2047. HOWERTON, D. A., R. 1. HUNTER, H. K. ZIEGLER, and I. J. CHECK. 1990. Induction of macrophage la expression in vivo by a synthetic block copolymer, L81. J. Immunol. 144: 1578. VERMA,J. N., N. M. WASSEF, R. A. WIRTZ, C. T. ATKINSON, M. AIKAWA, 1. D. LOOMIS, and C. R. ALVING. 1991. Phagocytosis of liposomes by macrophages: intracellular fate of liposomal malaria antigen. Biochim. Biophys. Acta 1066: 229. CLUFF, C. W. and H. K. ZIEGLER. 1987. An early response to lipopolysaccharide is the elicitation of macrophages specialized for antigen degradation with negative regulatory effects on the induction of specific immune responses. Infect. Immun. 55: 1346. BAKER, P. J., J. R. HIERNAUX, M. B. FAUNTLEROY, B. PRESCOTT, J. 1. CANTRELL, and J. A. RUDBACH. 1988. Inactivation of suppressor T-cell activity by nontoxic monophosphoryllipid A. Infect. Immun. 56: 1076. EKWUNIFE, F. S., C. E. TAYLOR, M. B. FAUNTLEROY, P. W. STASHAK, and P. J. BAKER. 1991. Differential effects of monophosphoryllipid A on expression of suppressor T cell activity in lipopolysaccharide-responsive and lipopolysaccharide-defective strains of C3H mice. Infect. Immun. 59: 2192. SCHNEERSON, R., A. FATTOM, S. C. Szu, D. BRYLA, J. T. ULRICH, J. A. RUDBACH, G. SCHIFFMAN, and J. B. ROBBINS. 1991. Evaluation of monophosphoryllipid A (MPL) as an adjuvant. Enhancement of the serum antibody response in mice to polysaccharide-protein conjugates by concurrent injection with MP1. J. Immunol. 147: 2136. BAKER, P. J., T. HRABA, C. E. TAYLOR, K. R. MYERS, K. TAKAYAMA, N. QURESHI, P. STUETZ, S. KUSUMOTO, and A. HASEGAWA. 1992. Structural features that influence the abiliry of lipid A and its analogs to abolish expression of suppressor T cell activiry. Infect. Immun. 60: 2694. KENNEY, J. S., B. W. HUGHES, M. P. MASADA, and A. C. ALLISON. 1989. Influence of adjuvants on the quantiry, affiniry, isotype and epitope specificity of murine antibodies. J. Immunol. Meth. 121: 157. KARCH, H. and K. NIXDORFF. 1983. Modulation of the IgG subclass responses to lipopolysaccharide by bacterial membrane components: Differential adjuvant effects produced by primary and secondary stimulation. J. Immunol. 131: 6. HUNTER, R., M. OLSEN, and S. BUYNITZKY. 1991. Studies on non-ionic block copolymer adjuvants. IV. Effect of molecular weight and formulation on titre and isorype of antibody. Vaccine 9: 250. TAKAYAMA, K., M. OLSEN, P. DATTA, and R. 1. HUNTER. 1991. Adjuvant activiry of nonionic block copolymers. V. Modulation of antibody isorype by lipopolysaccharides, lipid A and precursors. Vaccine 9: 257. KALISH, M. 1., I. J. CHECK, and R. 1. HUNTER. 1991. Murine IgG isorype responses to the Plasmodium cynomolgi circumsporozoite peptide (NAGGls. I. Effects of carrier, copolymer adjuvants, and lipopolysaccharide on isotype selection. J. Immunol. 146: 3583. VAN DE WIJGERT, J. H. H. M., A. F. M. VERHEUL, H. SNIPPE, I. J. CHECK, and R. 1. HUNTER. 1991. Immunogenicity of Streptococcus pneumoniae type 14 capsular polysac-
Lipopolysaccharide, Lipid A, and Liposomes . 445 charide: Influence of carriers and adjuvants on isotype distribution. Infect. Immun. 59: 2750. 45. MILLET, P., M. L. KALISH, W. E. COLLINS, and R. L. HUNTER. 1992. Effect of adjuvant formulations on the selection of B-cell epitopes expressed by a malaria peptide vaccine. Vaccine 10: 547. 46. EDELMAN, R. 1980. Vaccine adjuvants. Rev. Infect. Dis. 2: 370. 47. NowoTNY, A. 1987. Review of the molecular requirements of endotoxic actions. Rev. Infect. Dis. 9 (Supp!. 5):S503. 48. TAKAYAMA, K., E. RIBI, and J. L. CANTRELL. 1981. Isolation of a nontoxic lipid A fraction containing tumor regression activity. Cancer Res. 41: 2654. 49. AMANO, K., E. RIBI, and J. L. CANTRELL. 1983. Structural requirements of endotoxic glycolipid for antitumor and toxic activity. J. Biochem. 93: 1391. 50. TSUJIMOTO, M., S. KOTANI, T. OKUNAGA, T. KUBO, H. TAKADA, T. KUBO, T. SHIBA, S. KUSUMOTO, T. TAKAHASHI, Y. GOTO, and F. KINOSHITA. 1989. Enhancement of humoral immune responses against viral vaccines by a non-pyrogenic 6-0-acyl-murarnyldipeptide and synthetic low toxicity analogues of lipid A. Vaccine 7: 39. 51. VOSIKA, G. J., C. BARR, and D. GILBERTSON. 1984. Phase-I study of intravenous modified lipid-A. Cancer Immuno!. Immunother. 18: 107. 52. RAMSEY, R. B., M. B. HAMNER, B. M. ALVING, J. S. FINLAYSON, C. R. ALVING, and B. L. EVATT. 1980. Effects of lipid A and liposomes containing lipid A on platelet and fibrinogen production in rabbits. Blood 56: 307. 53. RICHARDSON, E. c., B. BANERJI, R. C. SEID JR., J. LEVIN, and C. R. ALVING. 1983. Interactions of lipid A and liposome-associated lipid A with Limulus polyphemus amoebocytes. Infect. Immun. 39: 1385. 54 DIJKSTRA, J., J. W. MELLORS, J. L. RYAN, and F. C. SZOKA. 1987. Modulation of the biological activity of bacterial endotoxin by incorporation into liposomes. J. Immuno!. 138:2663. 55. DIJKSTRA, J., J. W. LARRICK, J. L. RYAN, and F. C. SZOKA. 1988. Incorporation of LPS in liposomes diminishes its ability to induce tumoricidal activity and tumor necrosis factor secretion in murine macrophages. J. Leukocyte Bio!. 43: 436. 56. RiCHARDS, R. L., G. M. SWARTZ JR., C. SCHULTZ, M. D. HAYRE, G. S. WARD, W. R. BALLOU, J. D. CHULAY, W. T. HOCKMEYER, S. L. BERMAN, and C. R. ALVING. 1989. Immunogenicity of liposomal malaria sporozoite antigen in monkeys: adjuvant effects of aluminum hydroxide and non-pyrogenic liposomallipid A. Vaccine 7: 506. 57. PETROV, A. B., B. F. SEMENOV, Y. P. VARTANYAN, M. M. ZAKIROV, V. P. TORCHILIN, V. S. TRUBETSKOY, N. V. KOSHKINA, V. L. L'vov, 1. K. VERNER, 1. V. LoPYREv, and B. A. DMITRIEV. 1992. Toxicity and immunogenicity of Neisseria meningitidis lipopolysaccharide incorporated into liposomes. Infect. Immun. 60: 3897. 58. FRIES, L. F., D. M. GORDON, R. L. RICHARDS, J. E. EGAN, M. R. HOLLINGDALE, M. GROSS, C. SILVERMAN, and C. R. ALVING. 1992. Liposomal malaria vaccine in humans: A safe and potent adjuvant strategy. Proc. Nat!' Acad. Sci. USA 89: 358. 59. DIJKSTRA, J., J. W. MELLORS, and J. L. RYAN. 1989. Altered in vivo activity of liposomeincorporated lipopolysaccharide and lipid A. Infect. Immun. 57: 3357. 60. ALVING, C. R. and R. L. RICHARDS. 1990. Liposomes containing lipid A: a potent nontoxic adjuvant for a human malaria sporozoite vaccine. Immuno!. Lett. 25: 275. 61. ALVING, C. R. 1983. Delivery of liposome-encapsulated drugs to macrophages. Pharmac. Ther. 22: 407. 62. FIDLER, 1. J. 1985. Macrophages and metastasis-a biological approach to cancer therapy: presidential address. Cancer Res. 45: 4714. 63. ALVING, C. R. 1988. Macrophages as targets for delivery of liposome-encapsulated antimicrobial agents. Adv. Drug Deliv. Rev. 2: 107. 64. ALVING, C. R. 1991. Liposomes as carriers of antigens and adjuvants. J. Immuno!. Meth. 140: 1. 65. ALVING, C. R. 1992. Immunologic aspects of liposomes: Presentation and processing of liposomal protein and phospholipid antigens. Biochim. Biophys. Acta (Rev. Biomembranes). 1113: 307.
446 . C. R. ALVING 66. KRAMP, W. J., H. R. SIX, and J. A. KASEL. 1982. Postimmunization clearance of liposome entrapped adenovirus type 5 hexon. Proc. Soc. Exp. BioI. Med. 169: 135. 67. YOKOCHI, T., Y. INOUE, J. YOKOO, Y. KIMURA, and N. KATO. 1989. Retention of bacterial lipopolysaccharide at the site of subcutaneous injection. Infect. Immun. 57: 1786. 68. TAMAUCHI, H., T. TADAKUMA, T. YASUDA, T. TSUMITA, and K. SAITO. 1983. Enhancement of immunogenicity by incorporation of lipid A into liposomal model membranes and its application to membrane-associated antigens. Immunology 50: 605. 69. VAN ROOI]EN, N. and R. VAN NIEUWMEGEN. 1980. Endotoxin enhanced adjuvant effect of liposomes, particularly when antigen and endotoxin are incorporated within the same liposome. Immunol. Comm. 9: 747. 70. SCHUSTER, B. G., M. NEIDIG, B. M. ALVING, and C. R. ALVING. 1979. Production of antibodies against phosphocholine, phosphatidylcholine, sphingomyelin, and lipid A by injection of liposomes containing lipid A. J. Immunol. 122: 900. 71. ALVING, C. R. 1986. Antibodies to liposomes, phospholipids and phosphate esters. Chern. Phys. Lipids 40: 303. 72. ALVING, C. R., R. L. RICHARDS, J. Moss, L. I. ALVING, J. D. CLEMENTS, T. SHIBA, S. KOTANI, R. A. WIRTZ, and W. T. HOCKMEYER. 1986. Effectiveness of liposomes as potential carriers of vaccines: applications to cholera toxin and human malaria sporozoite antigen. Vaccine 4: 166. 73. PIERCE, N. F., J. B. SACCIJR., C. R. ALVING, and E. C. RICHARDSON. 1984. Enhancement by lipid A of mucosal immunogenicity of liposome-associated cholera toxin. Rev. Infect. Dis. 6: 563. 74. NAYLOR, P. T., S. LARSEN, L. HUANG, and B. T. ROUSE. 1982. In vivo induction of "'anti-herpes simplex virus immune response by type 1 antigens and lipid A incorporated into liposomes. Infect. Immun. 36: 1209. 75. BRYNESTAD, K., B. BABBITT, L. HUANG, and B. T. ROUSE. 1990. Influence of peptide acylation, liposome incorporation, and synthetic immunomodulators on the immunogenicity of a 1-23 peptide of glycoprotein D of herpes simplex virus: implications for subunit vaccines. J. Virol. 64: 680. 76. MORGAN, A. J., M. A. EpSTEIN, and J. R. NORTH. 1984. Comparative immunogenicity studies on Epstein-Barr virus membrane antigen (MA) gp340 with novel adjuvants in mice, rabbits, and cotton-top tamarins. J. Med. Virol. 13: 281. 77. RICHARDS, R. L., M. D. HAYRE, W. T. HOCKMEYER, and C. R. ALVING. 1988. Liposomes, lipid A, and aluminum hydroxide enhance the immune response to a synthetic malaria sporozoite antigen. Infect. Immun. 56: 682. 78. OGERT, R. A., M. K. GENTRY, E. C. RICHARDSON, C. D. DEAL, S. N. ABRAMSON, C. R. ALVING, P. TAYLOR, and B. P. DOCTOR. 1990. Studies on the topography of the catalytic site of acetylcholinesterase using polyclonal and monoclonal antibodies. J. Neurochem. 55: 756. 79. BRODIN, T., J. THURIN, N. STROMBERG, K. KARLSSON, and H. O. SJOGREN. 1986. Production of oligosaccharide"binding monoclonal antibodies of diverse specificities by immunization with purified tumor-associated glycolipids inserted into liposomes with lipid A. Eur. J. Immunol. 16: 951. 80. KOTANI, S., H. TAKADA, M. TSU]IMOTO, T. OGAWA, Y. MORI, M. SAKUTA, A. KAWASAKI, M. INAGE, S. KUSUMOTO, T. SHIBA, and N. KASAl. 1983. Immunobiological activities of synthetic lipid A analogs and related compounds as compared with those of bacterial lipopolysaccharide, Re-glycolipid, lipid A, and muramyl dipeptide. Infect. Immun. 41: 758. 81. ALLISON, A. C. and N. E. BYARS. 1986. An adjuvant formulation that selectively elicits the formation of antibodies of protective isotypes and of cell-mediated immunity. J. Immunol. Meth. 95: 157.
Dr. CARL R. ALVING, Dept. of Membrane Biochemistry, Walter Reed Army Institute of Research, Washingoton, DC 20307-5100, USA