CA2071069A1 - Vaccines protective against gram-negative bacteremia and endotoxic shock and related diagnosic assays - Google Patents

Abstract

Abstract Of The Disclosure The present invention is directed to the development of monoclonal anti-idiotypic antibodies (anti-ids) which mimic a common antigenic determinant on the lipid A portion of the Lipopolysaccharite (LPS) of gram-negative bacteria. More particularly, the invention comprises the development of hybridomas A4, H9, and H12 which produce monoclonal anti-Ids that mimic a common antigenic determinant on the lipid A portion of the LPS molecule found on a number of different bacterial strains. The development of the hybridomas was carried out by induction with a monoclonal antibody, 1C9, which recognizes a conserved epitope, that is a target for a bactericidal antibody, or involved in immune clearance mechanisms, and is common to all Enterobacteriaceae gram-negative bacteria.
The anti-Ids contain a protein image of the lipid antigen and are thus capable of stimulating anti-LPS activity.
Therefore, the anti-Ids of the present invention may be used as immunogens in vaccine formulations designed to protect against gram-negative bacteremia and septic shock. In one embodiment of a vaccine formulation in accordance with the present invention, the peptides could be generated based on the identity of the sequences of the idiotopes of anti-Ids. Alternatively, the DNA
encoding the idiotopes which mimic the lipid A antigenic determinant may be used by applying molecular cloning techniques to direct production of the peptides. Additionally, the DNA
itself may be incorporated into various vaccine-related vectors for vaccine applications.

Description

2 ~

VACCINES PROTECTIVE AGAINST GRAM-NEG~TIVE
BACTEREMIA AND ENDOTOXI~ SHOCK AND RELATED
DI~GNOSIC ASSAYS.

_ELD OF l'HE INVENTION
The present invention relates generally to gram-negative bacteremia and endotoxic shock, and more particularly to the development of monoclonal anti-idiotypic antibodies, or peptides thereof, which mimic a common antigenic determinant on the lipid A portion of the lipopolysaccharide.
The anti-idiotypic antibodies may be used as immunogens in vaccine formulations to induce bactericidal activity or elicit an immune clearance response to lipid A, to generate antibodies with application for passive immunization and as reagents in diagnostic assays.
BAC~GROUND OF THE INVENTION
Gram-negative bacteremia with shock is a serious and frequently fatal medical condition with a mortality rate ranging from 30-80% (sraude et al., 1977, J. Infect. 3iseases 136:
(Suppl.):167). It is estimated that in the United States in 1974, approximately 100,000 deaths could be attributed to the invasion of the blood stream of gram-negative bacteria followed by septic (endotoxic) shock (Kreger et al., 1980, Am. J. Med.
1980: 344). Gram-negative bacteremia and septic shock are major problems especially in immunosuppressed patients such as burn patients and patients receiving chemotherapy or irradiation for treatment of malignancies.
Despite recent developments in antibiotic regimens, and advances in health care management, the mortality from septic shock has not improved significantly in the past 20 years~ In fact, antibiotic therapy may actually exacerbate the condition by inducing lysis of bacterial cells resulting in the release into . ~ " , , - : :

2 ~ 7 ~

the blood stream of large amounts of lipopolysaccharide (LPS).
LPS is the major surface glycolipid on gram-negative bacteria, which is responsible for endotoxic shock. The LPS molecule consists of three regions: the O-antigen which is composed of a branching chain(s) of sugar residues, the core region comprising one or more sugar residues attached to a triplet of 2-Keto-3-deoxyoctonate, and lipid A consisting of fatty acid chains. Lipid A, the region of LPS responsible for toxicity, is conserved in all Enterobacteriacea~ gram-negative bacteria (Luderwit-~ et al., 1984, Rev. Infect. Dis. 6:428 and Mutllaria et al., 1984, Infect. and Immun. 45:631). Studies have shown that antisera against the lipid A component and antisera against the O--antigen are protective in models of LPS toxicity and gram-negative bacteremia. However, the advantages of antisera against lipid A over the polysaccharide portion of LPS include lipid A appears to be the most conserved portion of the LPS
molecule arnong different gram negative bacterial species and that ;
the polysaccharide portion of LPS fails to induce a memory response in the immune response. Another approach for the `~
treatment and prevention of this disease would involve stimulating an immune response against LPS. However active immunization with LPS is unacceptable due to inherent toxicity.
The pathophysiological effects of LPS may include fever, leucopenia, leucocytosis, the Shwartzman reaction, disseminated intravascular coagulation, abortion, and, in larger doses, shock and death. Many of these effects, however, are not direct effects of the LPS molecule on the cellular system but are mediated ~ia stimulation of macrophages to release biologically active endogenous factors such as tumor necrosis factor (TNF), some of the interleukins (IL-l and IL-6), prostaglandins of the E2 and F2 alpha type and cytolytic proteases. Lipid A is the portion of LPS mainly responsible for induction of these factors (Luderity et al. 1984, Curr. Topics in Membrane and Transport 2 ~ 7 ~

17:79). Accordingly, there are no currently available vaccines which protect against gram-negative bacteria and endotoxic shock.
Anti-idiotypic antibodies (anti-Ids) are a component of the immune regulatory network theory of N.K. Jerne (1974, "Toward A Net~/ork Theory of the Immune System", Ann. Immunol. 125c:373) that defines the immune system as web of interacting idiotypes.
An idiotype is a set of one or more idiotopes. Idiotopes are regions in or near the antigen recognition sites of immunoglobulin that are themselves capable of acting as antigens and stimulating specific antibody production. Under this theory, the normal immune system features an interlocking network of antibodies directed at one another's idiotopes, i.e. each immunoglobulin molecule not only expressed a nominal antigen-binding site (defined as the paratope) for a particular antigen, but, by virtue of the antigenic determinants expressed on the variable region (idiotopes), could also be recognized by other antibody molecules. Production of any particular idiotype will be suppressed by its corresponding anti-idiotypic antibody (anti-Id). Inherent in this theory is that certain anti-Ids e~press structures of idiotopes which mimic, i.e. are "internal images" of, the structure of external antigens. Thus it is possible for an anti-Id, the antigenic region being protein in composition, to have a similar configuration to an antigen which is biochemically different, such as a lipid. This principle illustrates the potential of anti-Ids in vaccine application.
Su et al., J. of Immunol 145:2994-3001, describes the nature of the host defense, i.e. the idiotypic characterization of the antibody response, to lipid A. This reference discloses LPS-binding properties of anti-LPS monoclonal antibodies produced; studies to determine the presence of lC9 (one of the anti-LPS monoclonal antibodies described) idiotypes in sera from , , '.:' . ' ' :' ' '' ;,',:'. :
, , :' ' ' ' ' 2~7~

different animal species; production of anti-Id antibodies against lC9; and the specificity of binding of the anti-Ids to lC9. The Su et al. reference also describes, in general terms, the potential for use of anti-Ids as vaccines.
The reference alludes to unpublished preliminary evidence that suggests that several of the anti-Id antibodies described have the ability to functionally mimic lipid A by stimulating antibody responses against lipid A. The reference fails to disclose ~ihether the antibodies induced by the anti-Ids (such antibodies are termed "Ab3") stimulated specific immunity against LPS, i.e. are protective against gram-negative bacteremia and endotoxemia. Such Ab3 needs to be able to recognize the antigenic epitope and confer protective immunity if the inducing anti-Id is to be considered as having vaccine potential. Nor does the Su et al. reference disclose if the anti-rds produced functionally mimic lipd A's undesirable biological effects such as mitogen activity; anti-complement activity; and stimulation of macrophages to release biologically active endogenous factors such as tumor necrosis factor, some of the interleukins, prostaglandins of the E2 and F2 alpha type, and cytolytic proteases. In terms of vaccine, it is very important that an anti-Id monoclonal antibody lacks these LPS-associated biological activities.
Anti-idiotype vaccination represents an effective approach in the enhancement or induction of an immune response to ;~
parasite pathogens (Sacks et al., 1985, J. Immunoloyy 135:4155 and Grzych et al., 1985, Nature 316:74); bacterial pathogens (Westerink et al. 1988, Infection and Immunity 156:1120; Stein and Soderstrom, 1984, J. Experimental Medicine, 160:1001; and McNamara et al., 1984, Science 226:1325), and viral pathogens (Gell and Mass, 1985, J. of General Virology 66:1801 and Kennedy et al, 19~6, Science 232:220). Several theories are set forth to explain this idea. First, is that the anti-idiotypic antibody, mimicking an epitope of foreign antigen, may induce populations of T and B cells which can bind antigen. Secondly, a regulatory idiotype is present on most B cell receptors with a required specificity such as the idiotype of an anti-LPS response. The anti-idiotype directed to the regulatory idiotype may stimulate a large proportion of the cells with reactivity to the antigen.
Therefore, a novel vaccine which would elicit anti-LPS
antibodies that would protect against gram-negative bacteremia and endotoxic shock, i.e. LPS toxicity; lack the inherent toxicity of the LPS molecule caused by LPS-associated biological activity; and elicit antibodies against a determinant that is conserved in all Entero~acteriaceae, the family of gram-negative bacteria most often the causative agent of the disease, is highly desirable.
OBJECTS AN~ SUMMARY OF THE rNVENTION
Therefore, the primary object of the present invention is to provide a peptide or protein antigen which stimulates an immune response against the lipid A portion of the lipopolysaccharide of gram-negative bacteria.
Another object of the present invention is to provide such a peptide or protein which comprises a monoclonal anti-l~iotype antibody~ or derivative thereof.
Another object of the present invention is to provide ;~
such a peptide or protein which when used as an immunogen is bacteriocidal or elicits an immune clearance response against gram-negative bacteremia and endotoxic shock.

.. ... ~.
': ' ~ - :

" . . . i ~ .

: ., :

2~171~

Another object of the inven~ion is to provide such a peptide OL protein which when used as an immunogen, fails to induce LPS-associated mitogenic activity, anti-complement activity, and stimulation of macrophages to secrete tumor necrosis factor (TNF) and interleukin 1 (IL~
The present invention is directed to the development of ~, monoclonal anti-idiotypic antibodies (anti-Ids) which mimic a common antigenic determinant on the lipid A portion of the lipopolysaccharite (LPS) of gram-negative bacteria.
More particularly, the invention comprises the development of hybridomas Ag, H9, and H12 which produce monoclonal anti-Ids that mimic a common antigenic determinant on the lipid A portion of the LPS molecule found on a number of different bacterial strains. The development of the hybridomas was carried out by induction with a monoclonal antibody, lC9, which recognizes a conserved epitope, that is a target for a bactericidal antibody, or involved in imrnune clearance mechanisms, and is common to all E~tero~acte~iaceae gram-negative bacteria.
The anti-Ids contain a protein image of the lipid antigen and are thus capable of stimulating anti-LPS activity.
Therefore, the anti-Ids of the present invention may be used as immunogens in vaccine formulations designed to protect against gram-negative bacteremia and septic shock. In one embodiment of a vaccine formulation in accordance with the present invention, the peptides could be generated based on the identity of the sequences of the idiotopes of the anti-Ids. Alternatively, the DNA encoding the idiotopes which mimic the lipid A antigenic determinant may be used by applying molecular cloning techniques to direct production of the peptides. Additionally, the DNA
itself may be incorporated in~o various vaccine-related vectors for vaccine applications.

2~7~g In another embodiment, a method for mimicking a lipid antigen, such as that found on LPS, by a functional protein antigen is disclosed. The related method for mimicking a toxic antigen, such as lipid A, by a non-toxic antigen which then is amenable ,o vaccine applications is also disclosed.
In a further embodiment the use of monoclonal antibody lC9, or antibodies generated by immunization with anti-Ids produced from hybridornas A4, ~19 and H12, for use in passive immunization, or as reagents in diagnostic assays is disclosed.
The use of the anti-Ids in identifying and selecting for antibodies from an anti-LPS response is also disclosed.
BRIEF DESCRIPTION OF THE DRAWIN{;S
Fig. 1 is a graph illustrating recognition by monoclonal antibody lC9 of LPS from different strains of gram-negative bacteria by ELISA.
Fig. 2 is a graph illustrating the ability of lC9 and an isotype-matched control, HB8~, to inhibit binding of 5I-labeled H12, A4, H9 or C8 to plate bound lC9 Id.
Fig. 3 is a graph illustrating the ability of the different anti-Id to inhibit the binding of 125-labeled lC9 to plate-bound LPS.
Fig. 4 is a graph illustrating the ability of LPS and the control, lipooligosaccharide from Neisseria to inhibit the binding of 125I-labeled lC9 binding to plate-bound anti-Id.
Fig. 5 is a graph illustrating B cell proliferation determined by [3H] thymidine incorporation test and expressed as the proliferation index, i.e., CPM incorporated by stimulated B
cells/cpm incorporated by medium-cultured B cells.

:: , :

:: -: . .
: : :

.

2~7~

Eig. 6 is a yraph illustratiny inhibition vf complement hemolytic activity usiny Dextrain-sRsc and MAb MlD04E system as indicators. SRsC lysis was determined by hemoglobin release measured by spectrophotometer (490am).
Eiy. 7 is a graph illustrating release of IL-l by macrophages after stimulation in vitro. IL-l activity in each culture supernatant was determined by testing [3H] thymidine incorporation by C3H/HcJ thymocytes.
Fig. 8 is a graph illustrating antigen dose response by ELISA.
DETAILED DESCRIPTION OF THE INVENTION
EXAMPLE I `
MATERIALS AND METHODS
l.0 Animals. BALB/c, A/St, and CAF, mice (4 to G week old) were obtained from West Seneca Labs, West Seneca, NY. NZW
rabbits, and chickens were obtained from Springville Labs, Springville, NY. ;
2.0 Animal Immunizations. BALB/c or A/St mice were immunized intraperitoneal (i.p.) two or three times with varying doses (about 100 to 200 ug/mouse) of Re595 LPS or heat-killed Re595 Salmo~ella minnesota in adjuvant. The first immunization was with complete Freunds adjuvant (CFA1, and subsequent ones were with incomplete Freunds adjuvant (IFA). The sera were collected at several time points over a number of weeks. Rabbits or chickens were immunized according to a modified method of Galanos et al, Eur. J. Biochem., 24:116, which disclosure is hereby incorporated by reference. Briefly, an acid-treated Re595 S. minnesota bacterial preparation was emulsified in CFA (the ': :

, . : ~

2 ~ 7 ~

first immu~ ation was with CFA, whereas the subsequent ones were with IFA), and the animals immunized intramuscular (i.m.) (rabbits) or i.p. (chickens) a total of four times. About one month after the final immunization, the animals were bled and serum tested.

3.0 Aq. Wild-type S. minneso~a LPS. LPS from Re595 5. minnesota (prepared by phenol extraction), lipid A (from an acid hydrolysis of Re595 LPS), and KDO were obtained from Sigma Chemical Co., St. Louis, MO.

4.0 Reaqents. Alkaline-phosphatase-conjugated polyclonal goat antisera to mouse Ig H and L chain classes were purchased from Fisher Biotechnology, Pittsburgh, PA. Enzyme substrate (p-nitrophenyl-phosphate), diethanolamine, and glutaraldehyde were all obtained from Sigma. 1~5 I was obtained from ICN Phar-maceuticals, Irvine, CA. Iodogen was obtained from Pierce Chemical Co., Rockford, IL. D-Galactosamine was obtained from Sigma. Guinea pig complement was obtained from GIBCO, Grand Island, NY.

5.0 Monoclonal Antibody (Mab) Preparation. To produce anti-LPS mAb, BALB/c mice were immunized two or three times i.p.
or s.c. in CFA with either heat-killed Re595 S. minnesota bacteria (for lC9 production) or phenol extracted LPS from Re595 -5. minnesota ( for all the rest of the monoclonals). Three days before the fusion, mice received 25 to 50 ug Re595 LPS in saline intravenously (i.v.). Primed spleen cells were fused at a 1:1 ratio with P3-X63-Ag8.653 cells (J. Kearney, University of Alabama at Birmingham, Birmingham AL) in a selection media containing hypoxanthine, aminopterin and thymidine (HAT).
Resulting clones were tested ~or binding to LPS, expanded, and cloned by limiting dilution methods. The majority of monoclonal antibodies were specific for the O/core portion of LPS.

. ~
.. .
, ~'' ' '' ~ ' '' '' ~

: ' :

~7~6~

Anti--idiotypic antibodies against lC9 were prepared, using a modified method of Lieberman et al., J. Exp. Med., 139:983 which disclosure is hereby incorporated by reference, by immunizing A/St mice with 100 ug doses of purified lC9 (in CFA
adjuvant) i.p. and subcutaneously (s.c.) two times (in IFA) at 3 to 4 week intervals. Three days before the fusion, mice received 50 ug lC9 i.v. Fusions were then performed as above, and clones tested by binding to plate-bound lC9. Goat anti-mouse Ig (H
chain class-specific) (Fisher Biotechnology, Pittsburgh, PA) was used to identify the anti-Id bound to lC9. Anti-idiotypic and anti-isotypic clones were distinguished by simultaneous screening for binding to lC9 and HB86, an isotype-matched control antibody.
Hybridoma cell line A~ has been deposited at the SUNY Monoclonal Antibody Center, State University of New York at Buffalo, Buffalo, ~iew York 14214, having Accession No. T2-S2-Cr-RW.

6.0 Antibody Purification. Anti-LPS monoclonal antibodies lC9, 8H, llF, and 2C were expanded in BAL~/c mice as ascites. lC9 (IgG2a) was purified from the ascitic fluid on a protein A-Sepharose (Sigma) column. The other anti-LPS
antibodies (IgM) were purified on a goat anti-mouse-IgM Sepharose ~B affinity column. The anti-lC9 mAb A4, ~12, H9 and C8 were expanded in CAF, mice or BALB/c mice and affinity purified on a lC9-coupled Sepharose 4B column. HB86 mAb was obtained from the American Type Culture Collection, Bethesda, MD and the ascitic fluid purified on a protein A column. After purification, the antibodies were tested for LPS or lC9 binding.

7.0 ELISA for Determininq Antiqen Specificity. Sera and cell culture supernatants were assayed for anti-LPS or anti-lipid A antibody specificity by enzyme linked immunsorbent assay (ELISA). The 96-well polystyrene plates were coated with S. MinneSota Re595 LPS (10 ug/ml in Tris buffer plus 2.3 mM

2Q7~

MgC12, p~l 7.4), wild type S. mi~esota LPS (5 ug/ml in Tris buffer plus 2.3 m~ MgCL2, pH 7.4) or lipid A (2 ug/ml in carbonate buffer, pH 8~0). After an overnight incubation at 4C, the unbound antigen was washed away with Tris buffer and the plates were incubated with 1% BSA in Tris buffer for 2 hours at 37C. The plates were rewashed with Tris buffer and the sample (diluted in Tris buffer plus 1% ~SA) added to each well. After overnight incubation at 4C, the plates were washed with Tris buffer and alkaline phosphatase-labeled goat anti-Ig H and L
chain reagents added. After 2 hours at 37C, the plates were washed with Tris buffer and developed by adding alkaline phosphatase substrate in diethanolamine buffer, pH 9Ø The plates were scanned in an Artek ELISA scanner coupled to an Apple II computer.

8.0 Anti-Id Characterization Assays. Anti-Id binding to anti-LPS mAb was performed by coating 96-well plates with 5 ug/ml concentration of antibody in Tris buffer, p~ 7.4, overnight at 4C. 125 I-labeled anti-Id were then reacted with the plate-bound Id for about 14 hours, at which time binding was assessed by measuring cpm bound. The 125 I-labeled antibodies used in these assays were labeled using the Iodo-Gen method as described by Seon et al, J. Irnmunol, 127:2580, which disclosure is hereby incorporated by reference using the 1, 3, 4, 6-tetrachloro-3(alpha), 6(alpha)--diphenylglycoluril reagent from Pierce Chemical Co. (Rockford, IL). Specificity of the anti-Id for lC9 Id was assessed by an inhibition assay, in which the assay described above was inhibited by coincubation of the labeled anti-Id with lC9 or an isotype-matched control antibody.
Epitope specificity of the anti-Id was assessed in two different assays. The first assay tested the ability of antigen to compete with lC9 binding to the anti-Id. One to two ug/ml ,;
. : , .

: ' - 12 - 2~71~9 concentrations of the anti-Id were bound to 96-well plates.
After ~ashing away the unbound anti-Id, graded doses of antigen were then preincubated with 125 I-labeled lC9 for 4 hours, followed by overnight coincubation in the anti-Id coated wells.
After ~ashing, the wells were dried and the counts per minute (cpm) were counted in a Packard gamma-counter (Packard Instrument Co., Downers Grove, IL).
In the second (reciprocal) assay, the ability of anti-Id to compete with the binding of lC9 to LPS were tested.
LPS (lO ug/ml) was coated to 96-well plates as described above, followed by overnight coincubation of graded doses of anti-Id and 125 I-labeled lC9. A~ter washing, the wells were dried and counted in a Packard gamma-counter. Epitopic cross-reactivity of the anti--LPS monoclonals were also studied using this assay, in which the anti-LPS antibodies substituted for the anti-Id inhibitors.
Anti-idiotypic cross-reactivity was assessed using an inhibition solid-phase radioimmunossay (RIA). As described above, anti-Id were coated to 96-well plates, followed by overnight coincubation of unlabeled anti-Id and 125 I-labeled lC9. After washing, the plates were dried and counted in a Packard gamma-counter. The concentration of anti-Ids needed to inhibit 50% of the binding to labeled lC9 (IDso doses) were then compared.

9.0 Idiotope Expression Assays. To determine expression of specific lC9 idiotopes in serum defined by the anti-Id reagents, a solid phase inhibition RIA was used. lC9 antibody (5 ug/ml) was coated to a 96-well plate. After washing and blocking with Tris buffer plus sSA 125 I-labeled anti-Id bindin~ to the plate bound lC~ was inhibited by serum dilutions or dilutions o~ supernatants to be tested. Inhibition represents :

2 ~

the presence of a specific idiotope. To ensure specificity, various positive and negative controls were used (including anti-isotype reagents); and samples to be tested were also absorbed for 24 hours on an LPS plate to remove anti-LPS
reactivity, and retested in the inhibition test to make sure the Id detected was of anti-LPS specificity. ~ -10.0 Bacteria. Wild--type S. minnesota, S. minnesota Re595, and Salmonella typhimurium were ob-tained from American Type Culture Collection, Rockville, MD. The bacteria were propagated in trypticase soy broth, Becton ~ickinson, Cockeysville, MD. ~acteria colony forming units (CFU) were enumerated by plating on trypticase soy agar plates supplemented with red blood cells (RBC). The bactericidal assay used was a modification of one described by Markham et al., J. Immunol, 133:962, -~hich disclosure is hereby incorporated by reference.
Briefly, S. typhimurium bacteria were harvested in log phase growth, counted, and plated in 96-well plates at a concentration of 2 x lC CFU/ml (100 ul/well). Next, various dilutions of the antibodies to be treated (50 ul) and a 1/10 dilution of guinea pig complement were added to the wells. The 96-well plates were then incubated at 37C for ~ hours, after which a 50 ul aliquot was removed, and plated out on a trypticase soy agar-RBC plate.
Eighteen hours later, a colony count was performed. Experiments were performed in duplicate, and controls for each experiment included guinea pig complement alone or an irrelevant antibody.

11 0 5. t~ph.imurium bacteremia. An LDso of S. typhimurium infection to 8ALB/c mice experiement was performed according to the method of Reed and Muench, Am. J. Hyg., 27:493, which dislosure is hereby incorporated by reference. Mice were infected with graded doses of S. ~yphimurium bacteria i.v., and deaths recorded for 7 days. The dose of bacterial suspension , ' ', ' '' ' ~ "
~. ' ' .

': . ' ' ,. .

! . :

-` 2~7~ ~6~

that kills 50 percent of the infected mice (LDso) is 3 x 107 bacteria per mouse. ~or the protection experiment, groups of five BALs/c mice received 200 ug of antibody i.v. Thirty-six hours after immunization, mice were challenged with 1X or lOx LD~o dose of bacteria and deaths recorded on a daily basis.

12.0 Endotoxemia, sALB/c mice were rendered susceptible to LPS-mediated toxic shock by the administration of 20 ug D-galactosamine/mouse to lessen the amount of LPS used. The L350 dose of Re595 LPS was 8 ug. In the protection study, groups of five mice received 200 ug antibody i.v. Then 3 to 4 days later, 20 ug galactosamine, admixed with 8 or 20 ug LPS, was administered i.v. to recipient mice and deaths were recorded for 1 week.

13.0 Antibody-protein coniuqate antiqens: Hemocyanin (Hy) from ~imulus polyphemus hemolymph (Sigma Chemical Co.? was used as a carrier pro-tein for MAb lC9, A4, H9, H12 and control antibodies. Mab-Hy was prepared as described by Mishell and Shiigi, 1980, in Selected Methods in Cellular Immunoloqy; which disclosure is hereby incorporated by reference. A 1 ml solution containing 5 mg antibody and 25 mg Hy in 0.1 M PBS, pH 6.8, was prepared. 0.1 ml of O.S% glutaraldehyde (fresh) was added at room temperature and allowed to react for 1 hour. The reaction was stopped by dialysis in 0.1 M ammonium bicarbonate for 3-4 hours at 5C. The solution was dialyzed overnight in PsS, pH
7.5. The next day, the solution was centrifuged at 10,000 rpm for 15 min ~o remove precipitate.

14.0 Anti-Id immunization in mice, rabbits and chickens: BALB/c mice were immunized two times with 10 ug, 50 ug, 100 ug or 200 ug A~-Hy, H9-Hy or H12-Hy i.p. twice at one month intervals. Rabbits or chickens were immunized with 1 mg A4-Hy, H9-Hy or H12-Hy s.c. four times at one month intervals.

2~7~

The immunogens were mixed with CFA for the first immunization, with IFA for the second immunization, and in saline for succeeding immunizations. One week after each immunization, the animals were bled. Control iso-type-matched Ig-Hys were used as immunogens.
15.0 Vaccine immunization: BALB/c mice were immunized with 100 ug A4-Hy, 50 ug H9-Hy or 10 ug Re595 S~ minnesota LPS
i.p. two times at one month interval. The immunogen preparations were emulsified in CFA and IFA respectively. Control ~ -isotype-matched Ig-Hys were also prepared and injected in parallel. Ten to fourteen days after the second immunization, the mice were challenged with pathogens~

16.0 Endotoxemia and active Protection studyo In this model, BALB/C mice were rendered susceptible to Re595 LPS-mediated toxic shock by the administration of 15 mg D-ga]actosamine/mouse to lessen the a.nount of LPS used. Using this model, an LDso dose of Re595 LPS was determined as described by Reed and Muench, 1938, American Journal of Hygiene 133:962, whicil disclosure is hereby incorporated by reference). Groups of 6-8 BALB/c mice were challenged with graded doses of Re595 LPS
plus 15 mg D-galactosamine i.v. and deaths were recorded, Here the LDso was determined to be 0.05 ug Re595 LPS per mouse.
Next, an active protection study was performed. BALB/c mice were immunized with different immunogens (see Vaccine immunization method above). Ten to 14 days after the second immunization, mice were challenged i.p. with 15 mg D-galactosamine plus 0.05 ug (1 X LDso) or 0.5 ug ~10 X LDso) of Re595 LPS. Mouse survival was followed for a period of ten days.

:
' ~,:;, .

- 2 ~

17.0 E.coli Olll:B170 bacteremia and active protection study: An LDso of E.coli Olll:B170 infection in BALB/c mice was determined to be 3.3 X108 bacteria per mouse. The mice were immunized as described above, and ten to fourteen days after second imlllunization, mice were challenged i.v. with 2 X LDso bacteria. Mouse survival was followed over a period of ten days.

18.0 Mitoqenic activity assay: This assay was a modiEication of a method descr-ibed by Vukaljlovich and Morrison, 1985, J. of Immunology 135:1345, which disclosure is hereby incorporated by reference. B cells were prepared from BALB/c mouse spleen and cultured in 96-well plates (2 X 105/well) in RPMI 16~0(100 ul/well) supplemented with 10% fetal calf serum (FCS), L-glutamine (2 X 10 3M), penicillin (100 u/ml), streptomycin (100 ug/ml) and gentamicin (20 ug/ml) in the presence of normal mouse Ig (3 mg/ml) for one hour at 37C, in a 5% C2 incubator. Then the tested samples (100 ul/well), i.e. S.
min~esota Re595 LPS, S. minnesota wi ld type LPS, S. minnesota lipid A, O/core portion of S. minnesota LPS, Dextran (m.w.
9,400), monoclonal antibodies (A4, H9, H12) and control IgGl and IgG3 (final concentration 3 X 10 6M, 3 X 10 7M and 3 X 10 8M) were added to their appropriate wells. After 54 hours, 3H-thymidine (0.5 uCi/well in 20 ul RPMI :L640 medium) was added.
After 18 hours, the cells were harvested with a 96-well channel harvester (Skation Inc.) and the incorporation of 3H-thymidine into the cultured cells was measured using a incorporation liquid scintillation counter. The assay was done in triplicate. The medium-cultured s cells incorporated 1252 + 102 cpm of 3H
thymidine. B cell proliferation index, i.e. cpm incorporated by stimulated B cells over cpm incorporated by medium-stimulated B cells ~as reported.

' 2~7~ ~9 19.0 The compLement fixation test: The method used for the complement fixation test is described by Galanos et al., 1971, European Journal of Biochemistry 19:143, which disclosure is hereby incorporated by reference. 0.3 ml Dextran-SRBC (0.3%) and 0.3 ml M104E ascites (1/64) were mixed for one hour at 37C, followed by overnight incubation at 4C. 15 ul of the sample to be tested was mixed with 15 ul of dilu~ed guinea pig serum for one hour at 37C. The solutions were mixed and incubated for 90 minutes at 37C. The sample was centrifuged and the absorbance of the supernatant was measured at 410 nm. The experiment was performed in triplicate. Anticomplement activity was expressed as percent inhibition of hemolysis obtained with the titrated tested sample.
Percentage inhibition of hemolysis was calculated as follows:
% inhibition of hemolysis =
1 _ O D. in presence of inhibitor - O.D. of backqround X 100%
O.D. in absence of inhibitor - O.D. of background 20.0 Preparation o_ Peritoneal macrophaqes: Peritoneal macrophages were prepared according to a method described by Hanazawa et al., 1985, Infection and Immunity 50:262, which disclosure is hereby incorporated by reference. BALB/c mice were injected i.p. with 3 ml thioglycolate medium. At day four, the peritoneal exudate cells were harvested by washing the peritoneal cavity with cold RPMI 1640 medium containing heparin (50 units/ml). The harvested cells were washed and suspended in macrophage culture medium, i.e. RPMI 1640 medium supplemented with 10% FCS, L-glutamine (2 X 10 3M), penicillin (100 u/ml), streptomvcin (100 ug/ml) and gentamicin (20 ug/ml). The cell suspension (106 cell/ml) was placed in Falcon dishes and , .:

2~7~

incubated for two hours at 37C in a 5% CO2 incubator. Th~
nonadherent cells were removed by washing three times with the warm RPMI 1640 mediurn. The adherent cells were judged to be about 90O macrophages by morphology and trypan blue staining.

21.0 TNF assay: The procedure was a modification of a method described by Mannel and Falk, 1989, Infection and Immunity 57:1953, which disclosure is hereby incorporated by reference.
Macrophages (2 X 105 cells/0.2 ml, in macrophage culture medium) were cultured in the presence of graded doses of tested agents for 24 hrs at 37C in 5O CO2 incubator. 0.1 ml supernatant was transferred from each well to a L929 mouse fibrosarcoma cell cultured well [4 x 104 cells/well, in 0.1 ml MEM alpha medium (GIBCO, Lab) supplemented with 5% FCS], in the presence of actinomycin D (0.5 ug/ml). After 20 hours, the surviving cells were fixed and stained with crystal violet (0~5% crystal violet, 3% formaldehyde, 0.17% NaCl, 22.3% ethanol) for 15 min. Excess dye was washed away and the remaining dye was solubilized in 33%
acetic acid. The absorbance was measured, and the percentage of cells surviving inhibition was calculated. Concentration of TNF
was e.Ypressed by the reference of standard curve or the titrated murine-rTNF dose (Asachi Chemical Industry Co, 3.1 x 10 6 units/mg)O `

22.0 IL-l assaY: The procedure was a modification of a method described previously Hanazawa et al., 1985, Infection and Immunity 50:262. Mouse thymus was removed aseptically from 8-12 week old C3H/HeJ mouse and teased in RPMI 1640 medium. The thymocytes were collected and washed with the same medium, and cultured in a 96 well plate (106 cells/well in 0.1 ml macrophage '!
culture medium in the presence of 1:2000 PHA). Supernatants from macrophage cultures were prepared as described in the method of the TNF assay. 0.1 ml of the supernatant was transferred from each well to the thymocyte cultured well.

2 ~ 9 The thymocytes were cultured for 48 hrs at 37C in a 5%
C2 incubator. The cells were pulsed with 3H-thymidine (0.8 uCi/well in RPMI 1640 medium) for 24 hrs, and harvested with a 96-channel cell harvester (Skation Inc.). 3H-thymidine incorporation was measured in a liquid scintillation counter.
EXAMPLE II
INDUCTION OF ANTI-LPS MONOCLONAL ANTIBODIES
The development of monoclonal antibody lC9 was prepared by immunizing Balb/C mice intraperitoneally or subcutaneously in complete Freunds adjuvant with a heat-killed suspension of Re595 5. ~innesota bacteria. Three days before the fusion, mice received intravenously 25 to 50 ug Re595 S. Minnesota :~
phenol-extracted LPS. The spleen cells were removed and the primed splenocytes were fused at a 1:1 ratio with Balb/C
plasmacytoma cells. The fusion procedure and media used is described by Kennett et al., 1979, Methods Enzymol. 58:345, which disclosure is hereby incorporated by reference.
EXAMPLE III
ELISA FOR DETERMINING ANTIGEN SPECIFICITY
Supernatants from the initial plating of the fusion were screened for production of antibody by ELISA using 96-well polystyrene plates coated with S Minnesota Re595 LPS, 10 ug/ml in Tris buffer plus 2.3 mM MgC12, pH 7.4; or lipid A, 2ug/ml in carbonate buffer, pH 8Ø After an overnight incubation at 4C, the unbound antigen was washed away with Tris buffer and plates were incubated with 1% BSA in Tris buffer for 2 hours at 37C.
The plates were rewashed with Tris buffer and the supernatants, diluted in Tris buffer plus 1% BSA, were added to each well.
After overnight incubation at 4C, the plates were washed with .
, 2~7~

Tris bu~fer and alkaline phosphatase-labeled goat anti-mIg H and L chain reagents added. After 2 hours at 3~C, the plates were washed with Tris buffer and developed hy adding alkaline phosphatase substrate in diethanolamine buffer, pH 9Ø The plates were scanned for reactivity in an ELISA scanner. The hybridomas which producted monoclonal antibodies reactive in the ELISA, were cloned by limiting dilution~ The results show that lC9, a monoclonal antibody of IgG2a isotype, is an anti-lipid A
antibody. Monoclonal lC9 was expanded in Balb/C mice as ascites fluid and purified by using a protein A-Sepharose column.
Importantly, for the purpose of vaccine development, this anti-lipid A monoclonal antibody (MAb) was able to bind to LPS preparations from a number of different gram-negative strains, including rough and wild type S. minnesota, some strains of E. coli, S. abortus eq~i, Pseudomonas aeruginosa, Serratia me~cescence and S. ryphimurium but not from ~eisseria gonorrhea as shown in Figure 1. Figure 1 illustrates the Mab lC9 in culture medium tested for its binding to various strains of LPS
coated plate by ELISA. The titiated culture medium (on the x axis) is plotted against the O.D. (405 nm) which was read 2 hours after adding phosphatase substrate. One possible reason that lC9 did not bind to lipooligosaccharide (LOS) of Neisseria gonorrhoeae is that the core region of LOS is highly branched, unlike the more linear structure present in the LPS of most gram-negative strains. This highly branched core region may prevent lipid A of LOS from being accessible to lC9.
EXAMPLE IV
PASSIVE PROTECTION STUDIES
The monoclonal antibody lC9 was tested in an in-vi-tro assay of bacterial killing using S. t~phimurium. Such assays : ' ':

2 ~ J ~ ~ ~ e~

provide evidence of epitopes important in antibody-mediated protection against gram-negative infections. The bactericidal assay use~ was a modification of the procedure outlined elsewhere (Markham et al., 1984, J. rmmunol. 133:962). To perform the in-vitro assay, graded doses of monoclonal antibody lC9 and S.
typhimurium bacteria at a concentration of 2 x 104 CFU/ml (100 ul per well) were incubated in 96 well plates for 4 hours at 37C in the presence of guinea pig complement. After incubation, an aliquot was removed from the wells and plated out on trypticase soy agar plates. The plates were incubated for 18 hours and then a colony count was performed. As shown in Table I, lC9, but not the negative control antisera M104E, was able to bind the bacteria specifically and induce complement-mediated lysis. This data provides indirect evidence that this lipid A determinant is exposed on the bacterial cell wall, a point of some controversy as described by sraude et al., 1972, J. Immunol, 108:505; Braude et al., 1973, J. Infect. Dis 128 (Suppl):157; Greisman et al., 1973, J. Immunol. 111:1349; Mullan et al., 1974, Infect. and Immun. 10;1195, the disclosures of which are hereby incorporated by reference.
Table I - In~vitro assay of bacterial killing Approximate number Bacteria Percent of bacteria per sample Antibody Surviving + SEM Killed 400no antibody, 375+62 0 no complement (C') -400C' only 375+40 8 400M104E + C' 400+41 0 400lCg + C' 166+19 56 ;'. ~ '', ' ' Next, the ability of lC9 to protect in an in-vitro S typhimurium infection model was tested. An LDso of S. typhimurium infection in BALB/C mice experiment was performed according to the procedure outlined elsewhere (Reed et al., 1938, Am.J.Hyg. 27:493). I'hese experiments indicated that the in-vivo LDso was 3 x 107 bacteria per mouse. Two groups of five salb/C
mice were immunized intravenously with antibodies. One group was immunized with the negative control sera M10~, and the other group with lC9. Thirty-six hours after immunization, the mice were challenged with a lethal dose of bacteria and monitored daily for death in each experiment up to 7 days. In one experiment the mice were challenged with a 10xLDso dose of 5.
typhimuri~m and in another experiment a lx LDso dose. The results of the protection studies, following in-vivo challenge, are summarized in Table II.

TABLE II
Passive protection against in-vivo challenge with 5. typhimuri~
-Immunization Percent Lethality protocol Challenge + SEM

no antibody 10 x LD50 100 1 x LD~o 47+10 200ug M104E,i.v. 10 x LD50 100 1 x LD50 ~8+10 200ug lC9, i.v. 10 x LD50 63~14 These results further support the observation that this lipid A
determinant is exposed on the bacterial cell membrane, and reveal that lC9 confers protection against bacterial infection, in-vitro and in-vivo.

, : . ' : :
..

. . :. : ~ ~ .. ::
:

:

2 ~

To further study the ability of monoclonal antibody lC9 to protect against gram-negative infections, in particular against endotoxemia, the antibody was tested in an endotoxemia model. In this model, salb/C mice were rendered susceptible to LPS-mediated toxic shock by the administration of D-galactosamine. Using, this model the LDso doses of Re595 LPS
were determined to be 8ug. Each group of five mice received 200ug of antibody intravenously. Seventy-two hours later 20 ug of galactosamine, admixed with either 8 or 20 ug LPS, was administered intravenously to recipient mice. The mice were checked daily and deaths were recorded for 1 week. The results of these e~periments (Table III) show that lC5 could also passively protect mice from the lethal toxicity of LPS, i.e.
confers protection against endotoxemia.
I'ABLE III ;~
Passive protection against LPS - induced shock Immunization Percent Lethality protocol Challenge + SEM

no antibody 20ug galactosomine (gal) 100~0 + 20ug LPS
no antibody 20ug gal + 8ug LPS5~+18 lC9 20ug gal + 20ug LPS13+10 H386* 20ug gal + 20ug LPS100-~0 lC9 20ug gal + 8ug LPS0~0 H~86 20ug gal + 8ug LPS100+0 *HB86 is a negative control serum.

Based on the collective data from these protection studies, 2~7~

monoclonal antibody lC9 was chosen to induce the production of anti-IDs.
EXAMPLE V
I UCTION ~F MONO~LONAL ANTI-IDS
Anti-idiotypic antibodies against monoclonal antibody lC9 were prepared using the modified procedures outlined elsewhere (Lieberman et al., 1974, J. Exp. Med. 139:983). A/St mice were immunized with 100 ug dose of purified lC9 (in CFA
adjuvant) intraperitoneally, and subcutaneously two times (in IFA) at 3 to 4 week intervals. Three days before the fusion, mice received 50 ug of monoclonal antibody lC9 intravenously.
Fusions were then performed as described previously.
EXAMPLE Vl D~TERMIN~TION OF SPECIFICITY OF MONOCLO~AL ANTI-IDS
Monoclonal antibodies produced from immunizations with lC9 were tested for anti-Id binding to lC9 by ELISA and ELISA
inhibition assays. lC9, 5ug/ml in Tris buffer pH7.4, was used to coat the 96 well plates 125 I-labelled anti-Ids were then reacted with the plate bound lC9 for approximately 14 hours at which time binding was assessed by measuring the counts bound. Labelling of the antibodies was done using the Iodo-Gen procedure outlined elsewhere (Seon et al., 1981, ~. Immunol. 127:2580). Speci~icity of the anti-Ids for the lC9 Id was assessed by an inhibition assay, in which the assay described above was inhibited by coincubation of the labelled anti-Id with lC9 or an isotype-matched control antibody, HB86. Specifically, and with reference to Figure 2, Id specificity of the four monoclonals using solid phase inhibition RIA was conducted. The binding of plate-bound lC9 to 125I-anti-Id was inhibited by lC9, and not by , :~
: ' 2~71 ~

an isotype-matched control HB86. This demonstrates Ld specificity for the monoclonals. With reference to Figure 3, H9, H12 and H4, but not C8, were able to inhibit LPS-lC9 binding. A4 appeared to have the highest affini-ty for lC9. Even at inhibitor concentrations as low as 3 x 10 1M(0.05gml), A~ could inhibit 60% of the binding of LPS to lC9. These results, demonstrate the specificity of anti-Id monoclonal antibodies A4, H9, and H12 for lC9 idiotopes. The binding of the anti-Ids to lC9 has been exploited by affinity purification of the monoclonal antibodies from ascitic fluid over a lC9-coupled Sepharose column.
For purposes of vaccine development, the anti-Id which most closely mimics the LPS antigen, recognized by lC9, would probably be the most beneficial for vaccine applications. To determine if the anti-Ids could mimic an epitope on the LPS
antigen, the anti-Ids were tested for their ability to inhibit LPS-binding to lC9 in two types of inhibition ELISAs. The first assay tested the ability of the different anti-Ids to inhibit the binding of 125I-labeled lC9 to plate-bound LPS. LPS, 10 ug/ml, was used to coat 96-well plates as described above, followed by overnight coincubation of graded doses of the individual anti-Ids and 125I-labeled lC9. After washing, the wells were dried and counts were quantitated. As seen in Fig~re 3, Ag, H9, and H12 anti-Ids were able to inhibit LPS-lC9 binding, with A4 appearing to have the highest affinity for lC9, i.e., could inhibit up to 60% of the binding of LPS to lC9.
The second assay tested the ability of LPS antigen to compete with 125I-labeled lC9 binding to the plate-bound anti-Ids. The individual anti-Ids were bound to 96-well plates using the method described above. After washing away the unbound anti-Id, graded doses of the LPS Ag were then preincubated with 125I-labeled lC9, followed by overnight coincubation in the :

2~7~.~6~

anti-Id coated wells. The negative control used was LOS antigen, an oligosaccharide from N.gonorrhea which does not bind to lC9.
After washing, the wells were dried and the counts were quantitated. With reference to Figure 4, the results, as shown, indicate that low doses of LPS can inhibit the binding of of A4, H9, and H12 anti-Ids to lC9. In contrast, the LOS negative control shows no inhibition. The data from both assays indicate that anti-Ids A4, H9, and H12 recognize the LPs binding site-related epitopes of the anti-LPS monoclonal antibody lC9.
These results, showing that the anti-Ids mimic the LP~ antigen, confirm the potential for using these anti-Ids, or derivatives thereof, in vaccine applications.
To determine if A4, H9, and H12 anti-Ids were reactive to the same or different idiotopes on lC9, graded doses of each of the differen-t anti-Ids was tested for its ability to inhibit the bindiny of 125I-labeled lC9 to the individual plate-bound anti-Ids using the method as described above. The data is expressed in concentrations (molarities) of anti-Ids antibodies that inhibit 50% of the binding of labeled lC9 to pla-te-bound anti-Ids (IDso). The results suygest that A4, H9, and H12 recognize the same idiotope or distinct, but closely related in terms of locale, idiotope on lC9 (Table IV).
Table IV
IDso of anti-Id binding to lC9 Plate-Bound Inhlb1tors Anti--Id A4 H9 H12 A4 4x10 llM 5.5x10 4M 7.5x10 4M
A9 2.7x10 3M 5x10 7M 1.7x10 7M
H12 5.5x10 10M 3x10 4M 6.5x10 7M

:: . . :
, , :
.. . .

.

~7~9 EXAMPLE V I I
IDIOTYPIC ANALYSIS OF ANTI-LIPID A RESPONSE
In producing monoclonal antibodies to the LPS of Re595 S~min~esota bacteria, four anti-lipid A monoclonals were identified: lC9, llF, 8H and 2C. These four monoclonal antibodies were compared for epitopic and idiotopic cross-reactivity. If one or more of the three monoclonals shared the specificity of lC9, these results would suggest that this epitope on the lipid A antigen, recognized by these antibodies, is immunodominant; i.e. a disproportionately large part of the immune response to lipid A is directed against this epitope.
Immunodominance is a desirable Eeature of an antigen used in vaccine applications. In the first experiment these antibodies were tested for their ability to cross inhibit each other's binding to plate-bound LPS. Various doses of the antibodies were preincubated in wells, that contain bound LPS, for 2 to 4 hours at 37C followed by addition of 125I-labeled lC9 for an overnight incubation. After washing, the wells were dried and the counts were quantitated. Values given are the molar quantities of the antibodies need to inhibit 50% of -the binding of lC9 to the plate-bound LPS (IDso). The results of this assay indicate that two of the four monoclonal antibodies, lC9 and IlF, recognize the same or closely related epitopes on lipid A (Table V).

.. . .
, ' ' -~, TABLE V
Epitopic cross-reactivity of IlF and lC9 Plate-sound Antibody inhibitor Ag lC9 IlF 8H 2C
. . _ .

LPS 3xlO 6 6xlO 6 no no inhibition inhibition In the second experiment the same antibodies were tested for idiotypic cross-reactivity by testing their ability to inhibit-the binding of 125I-labeled anti-Ids A4, H9, or H12 to plate-bound lC9. lC9, 5ug/ml, was used to coat 96 well plates.
Graded concentrations of the antibody inhibitors were added at the same time as thc 125I-labeled anti-Ids and incubated overnight. After washing, the wells were dried and the counts were quantitated. Molar quantities of the antibodies which inhibited significant binding of plate-bound lC9 to 125I-labeled :
anti-Ids are shown in Table VI. These results show that lC9 and IlF share idiotopic specificity. These findings taken together :~
with the results showing epitopic specificity, suggest that lC9 may represent a dominant idiotype in the response to lipid A.

.` ~ ' ' ~ , ' '' ' ' ~
' 2~7~ ~9 TABLE VI

diotypic cross-reactivity of anti-LPS antibodies -5I Anti-~d Inhibitors lC9 8H 2C IlF

A4 4X10 1OM 1X10 7M lo~6M 2x10 8M
H12 5x10 10M lx10 7M lx13 7M lx10 8M
H9 3x10 10M lx10 6M lx10 7M lx10 8M

EXAMPLE VIII
PRESENCE OF lC9 IDIOTOPES IN SERA FROM DIFFERENT SPECIES
By analyzing the anti-LPS response in different species, the finding of lC9 idiotopes would pro~ide further evidence that lC9 may be a dominant Id recognizing an immunodominant epitope on lipid A. The anti-LPS response of mice, rabbits, and chickens were studied for idiotopic specificity. Balb/c or A/St mice were im~unized 2 to 3 times with heat-killed Re595 S. ~innesota bacteria at 2 to 3 week intervals. Three weeks after the last immunization the immunized mice were bled. Rabbits and chickens were immunized following procedures outlined elsewhere (Galanos et al., Eur.J.Biochem.24:116.) and bled one month after the last immunization. Animal sera was tested for anti-lipid A titer by ELISA using a 96-well plate coated with lipid A. These sera were then tes~ed for the presence of lC9 idiotopes, as defined by A4, H9, and H12 idiotopes, by inhibition RIA. Various serum :.. , , ~ .

dilutions were tested for their ability to inhibit the binding of 125I-labeled anti-Ids to plate-bound lC9. The results are expressed as reciprocal titer + standard error of the mean (SEM).
The results (Table VII) show that the anti-lipid A sera of the different species tested, expressed individual idiotopes defined by A4, H9, and H12.

:, 2 ~

TABLE VII
Ldiotope distribution in different LPS - primed species Animal species Anti~Lipid A lC9 Idiotope Titer + SEM
titer A~ H9 H12 . _ _ BALB/c mice 500 13+2 50+8 50+8 +Re595 s.minn.2650 160+8500+83 90+8 A/st mice 200 100+15300+83150+42 +Re 595 s.minn. 1600360+42 100+0 120+17 N2B rabbit 1300 2 20 400 +Re 595 s.minn. 2000 200 175 1200 N2B rabbit 2250 17 _ _ +Re 595 s.minn. 2500 2 NZs rabbit 3250 2 11 6 +Re595 s.minn.2500 90 180 500 Chicken 1 150 12 8 7 +Re 595 s.minn. 2500 320 60 80 Chicken 2 175 20 3 4 ~Re595 s.minn.3000 250 50 55 The observation that lC9 is likely an interspecies Idx, suggests that this idiotope will be present in the human anti-LPS
response. These results, taken together with the other results, are relevant to the development of a vaccine against LPS
endotoxin.

EX~MPLE IX
INTRODUCTION OF ANTI-LIPID A ANTIBODIES IN DIFFERENT SPECIES
As a potential anti-idiotype vaccine, it is important that this monoclonal anti-idiotype antibody, ei-ther alone or coupled to a carrier protein, induce anti-antigen responses in different species. lCg idiotype, detected by A4, H9 and H12 anti-Ids, was present in mice, rabbits and chickens. Therefore . . ' 2 ~ 7 ~ 9 it is possible that A4, H9 and H12 anti-Ids might be able to induce an anti-lipid A response in these animals. I'his possibility was tested by immunizing BAI.B/C mice, rabbits and chickens, with A4, H9 and H12 anti-Ids coupled to a carrier protein, hemocyanin (Hy). BALB/C mice were immunized twice, at one month intervals. The rabbits and chickens were immunized 4 times, at one month intervals. The animals were bled one week after each immunization. The anti-lipid A titer of the sera was tested on lipid A coated plates by ELISA. The anti-lipid A
specificity was confirmed by inhibition ELISA. Table VILI shows the anti-lipid A response titer from the sera bled one week after the last immunization. The titers were significantly increased in BALB/C mice by immunization of these 3 anti-idiotypes coupled to Hy, and also increased in rabbits and chickens by immunization of A4 and H9, but not H12, coupled to Hy. None of the isotype control antibodies were able to induce increased anti-lipid A
titer in these animals.

. ~ : , - ~

1: ' , '' '' ~ ' : ,.

2~7~

Table VIII
Anti-Lipid A Antibody Response Titer in Different Species Preimmune SerumImmune Serum (s.e.m.) Balb/C mice A~-Hy ~0 750 (+41) Hg-Hy 70 650 (+20) ~-Hl2-HY 70 525 (+83) Control IgGl-Hy 70 50 (+ 8) Control IgG3-Hy 70 75 (+ 8) NZB Rabbit A~-Hy 250 2500 Hg-Hy 150 550 H12-Hy 200 100 Control IgG1-Hy 150 200 Chicken A~-Hy 200 1000 Hg-Hy 200 ~25 H12-Hy 200 400 Control IgG1-Hy 200 250 Control IgG3-Hy 175 200 EXAMPLE X
OTHER LPS-ASSOCIATED BIOLOGICAL ACTIVI _ STUDY
As shown above, these 3 anti-lds could mimic epitopes of lipid A in inducing an immune response. Next these anti-Ids were examined for their ability to mimic other LPS-associated biological activities, many of which are related to its toxicity.
Ideally, the anti-Ids, used as a vaccine, should not induce any LPS toxicity.
. .
To show different types of biological activity were studied: (l) mitogen activity, (2) anti-complement activity, and (3) stimulation of macrophages to secrete TNF and IL-1.

- ' .; ' , , :

2 ~ 9 - 3~ -(1) Mitoqen Activity ssay: The [3H]-thymidine incorporation was measured to compare the proliferation of sALs/c spleen s cells which were stimulated by different doses of immunogens in adjuvant, i.e., twice at one month intervals. One week after the 2nd injection, the animals were bled and the sera tested ~or anti-lipid A titer by ELISA, using a 96 well plate coated with lipida. Preimmune sera titer was equal to 3-4 fold increased over B5A background. The results in Figure 5 show that Re595 LPS, wild type LPS and lipid A could significan-tly stimulate s cell proliferation. Lipid A was a little weaker than wild type LPS in inducing mitogenic activity. It is important to note that all 3 anti-idiotypic antibodies exhibited no mitogellic activity, similar in effect as the isotype controls.
(2) Anti-Complement Activity: The assay design was based on the observation that LPS could bind to complement and inhibit its hemolytic activity. The system, including Dextran conjugated SRBC and Dextran-specific MAb M104E in ascites, was used as an indicator to test the hemolytic activity of complement. SRBC lysis was determined by measuring hemoglobin release using a spectrophotometer (410 nm). The stronger inhibitor would, eventually, lead to less hemoglobin release from SRBC. The results in Figure 6 demonstrate a strong inhibition of complement hemolytic activity by mutant and wild type LPS, and lipid A. The inhibition ability was greatest for Re595 LPS
followed by, in order of decreasing ability, lipid A and wild type LPS. No obvious inhibition is seen with the anti-idiotypic antibodies, or the O/core portion of LPS, or control antibodies.
(3) Stimulation of Macrophaqes to secrete TNF and IL-l: There have been reports that LPS was able to stimulate macrophages to secrete TNF and IL-1, both of which were potential mediators of LPS toxicity (see: Chen et al, 1990 J. Immunology ,,~: " '' ,..

::

.

2~711 ~

145:8, which disclosure is hereby incorporated by reference). It is important to examine whether the 3 anti-Ids, mimicking an epitope on lipid A, stimulate macrophages to secrete these two cytokines. Thioglycolate medium-treated peritoneal adherent cells were used as a source of macrophage. The supernatant of the stimulated macrophages was tested for TNF or IL-1 activity.
The L929 mouse fibrosarcoma cell line is sensitive to TNF, and hence used for indicating TNF activity. In the presence of the tested supernatant, and Actinomycin D which is an inhibitor of DNA proliferation, L929 cell viability was determined. Commercial murine rTNF was used to produce a standard curve. 50% inhibition of L929 cell survival required ;~
95 units/rnl rTNF.
The resu]ts in Table IX show that none of the 3 anti-lds stimulated macrophages to secrete detectable amounts of TNF. Re595 was the most effective stimulator. Lipid A alone was a weaker stimulator in this assay.
For testing IL-l activity in the supernatant, [3H]-thymidine incorporation by thymocytes f rom C3H/HeJ mice was used as indicators. C3H/HeJ mouse strain is unresponsive to l.PS
in all biological activities including immunological activity, because this strain has a mutant allele (LPSd) at the genetic locus that controls responsiveness to LPS. Thus, the proliferation of lymphocytes oE the C3H/HeJ mouse strain is refractory to the effect of LPS. Here the proliferation of C3H/HeJ thymocytes, in the presence of phytohaemagglutinins (PHA) which is a thymocyte mitogen depending mainly on the existence of IL-l, even though LPS was not removed out of the tested supernatant.

2~7~$~

The results in Figure 7, showed that the supernatant from these anti-Id-stimulated macrophage culture, contained no more than IL-l that from non-stimulated macrophage culture, since their measured cpm was similar. Significantly increased [3H]-thymidine incorporation was seen in the C3H/HeJ thymocytes which were cultured with different LPS or lipid A stimulated macrophage superna-tants. Still, Re595 LPS was the most effective stimulator. The O/core portion did not stimulate macrophages to secrete either TNF or IL-l.
Collectively, the above results demonstrate that A4, H9 and H12 anti-rds do not mimic LPS in stimulating LPS-associated biological activities including mitogen activity, anti-complement activity, and stimulation of macrophages to secrete TNF and IL-l.
Lipid A was sho~n here to be the portion of LPS mainly responsible for the biological activities, since the O/core portion alone was not able to show any of these activities.

Table IX
Release of Tumor Necrosis Factor (TNF) by Macrophage After Stimulation In Vitro Stimulating Agents Concentration (M) TNF(U/ml,SEM) 5. Minnesota Re545 LPS 10 7 95(~3) 10-9 93(+4) 5. Minnesota LPS 10 7 401+4) 10-9 7(+3) 5. Minnesota Lipid A 10 9 15(+6) O/Core from S. Min. LPS 10 79 N.D. ;
N.D.
A4 10-7 N.D.
10 9 N.D.
H9 10 7 N.D.
10 9 N.D.
H12 10 7 N.D.
10 9 N.D.

t . .

:: ~
.' ,'."; '. ' ,, ' '. ' ~ :

N.D. = not detectable EXAMPLE XI
ACT I VE P ROTECT I VE_STUD I ES

Next, the ability of the anti-Ld antibodies to function as a vaccine was examined, since it has been demonstrated that these 3 binding-site related anti-Ids could antigenically mimic epitopes on lipid A, but biologically they lacked its toxic activity. Since two of three anti-Ids were able to induce an an immune response against lipid A in different species, these two were chosen as the vaccine components to be tested.
Consistent with other idiotype vaccines reported by Westerink et al., Infection and Immunity 155:1120, coupling an ;
anti-Id to a potent immunogenic carrier to receive full T-cell help is important. Here Hy was used as such a carrier. Optimal doses of the immunogen were investigated. BALB/c mice (5 per ;~
group) were immunized intraperitoneally with different doses of immunogen, twice at a one month interval. The first time, the immunogens were mixed with CFA, and the second time, with rFA.
One week later, the mice were bled. The sera were titrated and tested on a lipid A coated plate by ELISA. The results in Figure 8 show that the best dose of A4-HY to induce a lipid A
specific immune response was 100 ug per mouse, and that the best dose for H9-Hy was 50 ug per mouse. The induced antibody titers were much higher compared to control groups. The optimal dose range of these antigens for immunization was consistent with other reports (see: Galuton et al., 1986, J. Immunology 137:2930;
and Francotte et al., 1984, J. Experimental Medicine 160:1485), whereas the optimal dose of ReS95 LPS was 10 ug per mouse.

... . .

.. ,- , .,: : ~

2~7~

Next, the ability of anti-Id antibodies, as well as Re595 LP~ to stimulate an immune response to protect BALB/c mice against LPS endotoxemia and gram--negative bacteremia was tested.
In order to set up a LPS endotoxemia protection assay, it was necessary to first determine the LDso of Re595 LPS, which was found to be 0.05 ug per mouse in the presence of 15 mg galactosamine. Galactosamine was used to decrease the endotoxin lethal dose via a high accumulation of UDP-gaiactosamine deriva-tives in liver and a depletion of hepatic UTP. For the protection assay, mice were immunized with 10 ug Re595 LPS, 50 ug H9-Hy, 100 ug A4-Hy, or dose-matched control antibodies-ffy, per mouse. Ten to fourteen days after the second immunization, the mice were challenged with 1 X LD50 (0.05 ug) or 10 X LD50 (0.5 ug) of Re595 LPS in the presence of 15 mg galactosamine.
Mouse survival was followed for a period of ten days. Most deaths occurred at one day after the LPS challenge.
The result in Table X showed that immuniza-tion with either Re595 LPS, H9-Hy or A4-Hy could equally effect protection of mice against LPS endotoxemia. Statistical analysis showed no <~
significant difference among these 3 groups (p 0.05 X2), but control groups (IgG3-Hy and IgG-Hy) showed little protection over the saline immunized group.
Next, the anti-Ids were tested in a gram-negative bacteremia protection assay. Enteropathogenic bacteria E. coli Olll:B170 was used to challenge mice immunized with the anti-Ids.
The LDso was determined to be 3.3 X 108 bacteria in normal BALB/c mice. The mice wer-e immunized by using the same immunization protocol as in the LPS endotoxemia pr-otection assay. Ten to fourteen days after the second immunization, mice were challenged with 2 ;C LDso (6.6 X 108) E. coli Olll:B170, intravenously.
These mice were observed for the following 10 days. Most deaths . . . . .
...
!

.'' ~ ~;' ' '''' ' ,~ , ,; ' '"' '"

2~7~9 were seen in the second to third days. The results, depicted in Table XI, show control IgG3-Hy and IgG-Hy did not protect mice at all; whereas H9-Hy and A4-Hy protected mice very well, i.e. 45%
and 74% mice survived, respectively in these two groups.
Interestingly, although Re595 LPS could induce a comparable immune response against lipid A it only barely protected mice against gram-negative bacteremia (10% mice survived).
Collectively, these experiments reveal that H9-Hy and A4-Hy were very effective in protecting mice against gram-negative bacteremia and endotoxemia, while Re595 LPS only induced protective function in the endotoxemia protection assay, but poorly protected mice against bacteremia.
The monoclonal antibody A-4 is further being characterized as to indentification of the encoding genomic DNA
sequence from hybridoma A4. Identification of the nucleic acid sequences encoding that portion of monoclonal antibody A4 which mimics an epitope on the lipid A portion of LPS, will allow production of that representative portion by recombinant DNA
technology. The range of hosts known to those skilled in the art and used for production of recombinant proteins or peptides, include a variety of bacteria, yeasts, fungi, and some eukaryotic cells. There are plasmid and viral vectors that are commercially available, and of common knowled~e to those in the ar-t, which can be used as expression vectors for production of recombinant proteins and peptides. Therefore, it is ~ithin the scope of the present invention to utilize recombinant DNA technology, including use of restriction enzymes, ligation, and selection, to insert the nucleic acid sequence which encodes the portion of A4 that mimics lipid A epitope into an expression vector. ~his expression could then be introduced into the appropriate host cell for large scale production of the protein or peptide.

, 2 ~

- ~o -Further, this protein or peptide can be purified from the expression system by immunoaffinity chromatography in preparation for use in a composition for immunization purposes.

,'' : ,i .:

..

2 ~

TABLE X
Protection Against In Vivo Challenqe with LPS Endotoxin Immunoqen LPS Challenqe~t Alive/Total % Survival Saline LD50 5/12 42 10 X LD500/12 o (Re595) LPS LD50 12/12 100 H9-Hy LD50 12/12 100 A4-Hy LD50 12/12 100 Control IgG3-Hy LD50 9/12 75 Control :
IgGl-HY LD50 8/12 67 2~7~69 TABLE XI
Protection Against In-Vivo Challenge with E. coli Olll:Bl7~ Bacteremia Immunoqen# Alive/Total % Survival Saline 0/20 0 H9-Hy 9/20 45 A4-Hy 14/90 74 IgG3-HY 0/20 0 IgGl-Hy 0/20 0 . . ; . ~ . ~ ., ;

. : :, .:

2 ~
- ~3 -It is understood that the E~amples of the present inventior, described herein are for illustrative purposes only, and any changes or modifications as will become apparent to one of ordinary skill in the art from the foregoing description and accompanying figures are intended to be included within the spirit of this application and within -the scope of the appended claims.

- ~ . .
:. . .

~.

Claims (10)

1. A composition useful in the prophylactic treatment of gram-negative bacteremia and endotoxemia, said composition comprising a monoclonal anti-idotype antibody, or binding fragment or portion thereof, wherein said antibody functions as a molecular mimic of an epitope on the lipid A portion of lipopolysaccharide of gram-negative bacteria and functions to stimulate specific immunity against lipopolysaccharide.

2. The composition of claim 1, wherein said specific immunity stimulated comprises anti-lipid A antibodies which bind to said epitope and are bactericidal or elicit an immune clearance in the presence of active complement or neutrophils.

3. The composition of claim 1, wherein said gram-negative bacteria are members of the family Enterobacteriaceae.

4. A pharmaceutical composition according to claim 1 further comprising a physiologically acceptable carrier.

5. A subunit vaccine formulation in which the immunogen comprises an effective amount of the composition of claim 1 mixed with a pharmaceutical carrier.

6. A recombinant vector comprising a DNA sequence coding for an antigenic determinant derived from an idiotope of the monoclonal anti-idiotype antibody of claim 1, said determinant mimics an epitope on the lipid A portion of lipopolysaccharide and is capable of stimulating specific immunity against lipopolysaccharide.

7. A bacterium containing the recombinant vector of claim 6.

8. A unicellular organism containing the recombinant vector of claim 6.

9. A composition comprising anti-lipid A antibodies induced by immunization with the composition according to claim 1.

10. A composition comprising of anti-lipid A antibodies induced by immunization with the composition according to claim 6.