| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Iomai Corporation, Gaithersburg, MD 20878
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Heat-labile enterotoxin (LT) of Escherichia coli and the closely related cholera toxin (CT) belong to a group of bacterial exotoxins known to include the most powerful adjuvants in stimulating mucosal and systemic serum Ab responses to coadministered Ags (8, 9, 10). In humans, bacterial exotoxins can be safely used topically on the skin (i.e., transcutaneous immunization) but are precluded from use as oral or injected adjuvants due to diarrhea and local inflammation, respectively (11, 12, 13, 14). Mutant LT derivatives with reduced toxicity have been generated but were found to be ineffective adjuvants when administered orally (11, 15). However, the effects of such LT derivatives upon injection in an i.d. environment are unknown.
Bacterial ADP-ribosylating exotoxins are organized as A:B heterodimers consisting of one A and five B subunits (AB5). To exert its toxic action, binding of the B subunit to the ganglioside GM1 on the cell surface is followed by the A subunit entering the cell. Then, the intracellular ADP-ribosylating activity of the A subunit causes toxicity, such as fluid loss and diarrhea in intestinal epithelia upon oral exposure (16, 17). Application to the nasal mucosal environment in mice resulted in specific translocation of exotoxin (or exotoxin subunits) and coadministered Ags to the CNS, raising concerns about undesirable neurotoxicity (18, 19). To reduce toxicity, modified exotoxins are described in which the A subunit is mutated or deleted to reduce or eliminate ribosylation activity (20, 21). However, these formulations may still retain toxicity due to residual ribosylation activity, or there is an undesirable loss of adjuvanticity (15, 22). Exotoxins with B subunit modifications that interfere with cellular binding are not considered potent adjuvants. Using oral routes of vaccine delivery, attempts to reduce toxicity by mutating the B subunit led to an undesired decrease or change in adjuvant activity (11, 23). In general, interference with in vivo GM1 binding is considered deleterious for vaccine development purposes (24).
The present study investigates ganglioside binding-deficient exotoxins as adjuvants for parenteral immunizations, especially in relation to the current focus on i.d. vaccine development. Possible inflammatory side effects are easily assessed in the skin, and the abundance of immunostimulatory cells, such as cutaneous dendritic cells, may overcome the ineffectiveness observed in oral immunizations.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Female C57BL/6 mice 8–12 wk of age were obtained from Taconic Laboratories. C3HeJ and C3HeOuJ were obtained from The Jackson Laboratory. For complex ganglioside knockout mouse experiments, heterozygous mice (on a C57BL/6 background) with a deletion of 
-1,4-GalNAc-transferase (GM2/GD2 synthase) were provided by Dr. R. Schnaar (Department of Pharmacology and Neuroscience, The Johns Hopkins School of Medicine, Baltimore, MD). A colony was established and maintained under pathogen-free conditions at Biocon, and progeny was genotyped by PCR as described previously (25). Age-matched groups of homozygous and heterozygous females between 8 and 30 wk of age were used. All studies were approved by the Animal Care and Use Committee affiliated with the animal facilities of Gene Logic and BIOCON.
Immunizations
Mice were immunized i.d., i.m., or s.c. with 25-µl injections under the dorsal caudal surfaces (i.d./s.c.) or in the right femoral muscles (i.m.). For i.d and s.c. injections, the back of the mice were shaved 2 days before injection. Ags used were tetanus toxoid (TT; 1000–2000 limits of flocculation (Lf) units per mg of protein; Berna Biotech), OVA (Sigma-Aldrich) containing low endotoxin (<100 endotoxin units/mg), and Fluvirin (trivalent influenza split virus vaccine (TIV); Evans Vaccines) containing the A/New Caledonia/20/99, A/Panama/2007/99, and B/Shangdong/7/97 flu strains. The SIINFEKL peptide was purchased from New England Peptide. LT was supplied by Berna Biotech. LT(G33D) synthesis and purification (>95% purity) is described (11) and was generously provided by Dr. J. Clements (Department of Microbiology and Immunology, Tulane University Medical Center, New Orleans, LA). Endotoxin levels in LT(G33D) preparations were <4000 endotoxin units/mg protein (which equals <2 endotoxin units per injection), as measured by the Limulus amebocyte lysate method (Endochrome kit, Charles River Laboratories). CT and tetanus toxin were obtained from List Laboratories, GM1 and LPS were from Sigma-Aldrich, and alum (Rehydragel, an adsorbent aluminum hydroxide gel) came from Reheis.
Skin observations
The skin of mice injected i.d. was inspected regularly for redness and swelling during 14 days after injection. Periodically, the diameter of swelling was measured by calipers, and in some cases full-thickness skin biopsies were taken from the site of injection. Biopsies were fixed in Formalin and further processed (embedded, cut, mounted to slides, and stained with H&E) by Gene Logic. Slides were inspected under a Nikon Y-FL microscope, and representative fields were photographed with a digital camera (Nikon Coolpix 990).
Lung wash
Three weeks after the final immunization, lung washes were collected as previously described (26). Briefly, mice were exsanguinated, a polypropylene tube was inserted into the trachea, and 0.5 ml of buffer (PBS containing 3 µg/ml PMSF; Sigma-Aldrich) was infused to inflate the lungs. The infused material was then withdrawn and centrifuged, and the supernatant was stored at –80°C.
ELISA
Ag-specific Ig (IgG) titers were detected in sera and lung washes by ELISA on 96-well ELISA plates (Immulon-2HB from Dynex Technologies). Plates were coated overnight with Ag (1 µg/well) and subsequently blocked with 0.5% casein-Tween 20 for 2 h and washed. Samples were serially diluted (2-fold) on the plates and incubated overnight at 4°C. IgG levels were detected by using HRP-conjugated goat anti-mouse IgG (Bio-Rad) as secondary Ab and ABTS substrate (Kirkegaard & Perry Laboratories). Ab titers are reported as the OD at 405 nm (OD405) or ELISA units (corresponding to the inverse dilution of the serum that yielded an OD405 of 1.0).
In vivo tetanus toxin challenge
Tetanus toxin was diluted in 1:1 (v/v) PBS/nutrient broth, and a lethal dose of 1 ng was given s.c. to each group of mice as described previously (27). Mice were monitored daily for paralysis and death. In case of paralysis, mice were immediately sacrificed.
Hemagglutination inhibition (HAI) assay Ab titers
In experiments with Fluvirin, serum samples were analyzed for neutralizing Abs with a HAI assay as previously described (28). HAI titers are expressed as the highest serum dilution factor causing complete hemagglutinin (HA) inhibition.
Spleen and lymph node cell isolation
Three weeks after the final immunization, spleen and inguinal draining lymph nodes (DLNs) were removed and pooled per group (for ELISPOT) or individually processed (for in vivo CTL analysis). Single cell suspensions were prepared by passing the tissue through a 100-µm mesh screen. For spleen cell suspensions, erythrocytes were lysed with Tris-buffered ammonium chloride, and the remaining cells were extensively washed in Hanks’ buffer (Invitrogen Life Technologies). Eventually, cells were resuspended in complete medium containing RPMI 1640 (BioWhittaker) supplemented with 10% (v/v) FBS (Invitrogen Life Technologies), 2 mM L-glutamine (Invitrogen Life Technologies), 100 U/ml penicillin (Invitrogen Life Technologies), and 100 µg/ml streptomycin (Invitrogen Life Technologies).
ELISPOT
The ELISPOT assay was used to determine the number of cells producing IFN-
or IL-4. MultiScreen-HA 96-well membrane plates (Millipore) were coated with mAbs specific for IFN- (BioSource International) or IL-4 (BD Biosciences) overnight at 4°C and subsequently blocked by 1% BSA in PBS for 1 h at 37°C. Cells isolated from spleen or DLNs were plated at 106 cells/well in triplicate. Cells were cultured overnight in complete medium alone, medium with peptide (for CD8 T cell responses), or medium with whole protein (for helper T cell responses). Spots were resolved as described (29) and enumerated by a Bioreader 4000 PRO-X ELISPOT reader (Bio-Sys).
Tumor inoculation and measurements
B16-derived tumor cells genetically engineered to express OVA (MO5 cells) were a gift from Dr. K. Rock (Dana-Farber Cancer Institute, Boston, MA). Tumor cells were grown in batches in vitro in complete medium. First, at day 0 mice were s.c. injected on their shaven backs with 105 MO5 cells per mouse. Then, 3 and 15 days after tumor inoculation the mice were immunized i.d. with OVA with or without adjuvant. Outgrowth of s.c. tumors was measured with calipers periodically in individual mice.
In vivo CTL assay
To determine the ability of CTL to kill specific targets in vivo, a recently developed assay has been described (30). Briefly, splenocytes from naive mice were labeled with PKH26 (Sigma-Aldrich) and either 50 or 500 nM CFSE (Molecular Probes). The cells labeled with 50 nM CFSE were then coated with the OVA immunospecific peptide SIINFEKL (31, 32). Both cell populations (PKH-CFSE500 and PKH-CFSE50-SIINFEKL) were transferred i.v. (2.5 x 106 of each population) into the immunized mice. The next day, individual spleens and DLNs were isolated, and cell suspensions were generated as described above. The two target cell populations were identified by PKH26 staining and distinguished from another based on CFSE staining by multivariant FACScan analyses (FACSCalibur from BD Biosciences). The percentage of specific target cell killing was calculated as previously described (30).
APC labeling and analysis
FITC (Sigma-Aldrich) was reconstituted at 500 mg/ml in DMSO and then diluted into PBS. Skin at the dorsal caudal side was injected i.d. with 10 µg of FITC alone in PBS or admixed with LT(G33D). The next day, cells from DLNs (inguinal) were isolated and suspended in ice-cold PBS containing 0.5% (w/v) BSA. Cells were stained with fluorescently conjugated Abs (BD Biosciences) specific for mouse CD11c (allophycocyanin) and CD86 (PE) or H-2 IAb (PE). After 30 min the cells were washed twice with ice-cold 0.5% BSA in PBS and analyzed using a FACSCalibur flow cytometer.
Statistical analysis
Data were analyzed statistically using a two-tailed Student’s t test with the levels of significance indicated. Ab titers were compared by paired analysis of the logarithmic values of the measured ELISA units.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Exotoxins such as LT and CT are highly inflammatory when injected into the dermis or s.c. tissues. Upon injection they elicit self-limited local erythema and swelling at the site of injection, which may persist for weeks to eventually resolve without sequelae. Local skin inflammation was evaluated by measuring skin swelling (formation of skin nodules) in response to i.d. injection. Injection i.d. of 0.5 µg of LT caused inflammatory skin nodules in all mice (Fig. 1A) lasting 
3 wk. In contrast, LT preadsorbed to soluble GM1 ganglioside (1:16 molar ratio equals 0.5 µg of LT plus 0.25 µg of GM1) before injection caused minimal swelling (Fig. 1A). Similarly, i.d. injection of CT, a related GM1-binding exotoxin, caused skin inflammation, whereas a preadsorbed CT-GM1 complex did not (data not shown).
|
Full thickness biopsies were taken for histological examination following i.d. injection with LT, LT-GM1, or LT(G33D). Consistent with the gross observations, biopsies obtained from mice 2 days after injection were edematous with a large diffuse polymorphonuclear leukocyte infiltrate throughout the epidermis, dermis, and s.c. tissues (Fig. 1C) that resolved in 2–3 wk. In contrast, histological examination of LT(G33D)-injected skin exhibited normal skin architecture and an absence of an inflammatory infiltrate (Fig. 1D). Skin injected with LT-GM1 also resembled control skin (not shown).
Additional studies investigated the potential use of LT(G33D) and the LT-GM1 complex by other parenteral routes, namely i.m. and s.c. As observed with i.d. injection, wild-type LT elicited the formation of a large nodule with s.c. injection and swelling of the thigh muscle with i.m. injection. In contrast, i.m. and s.c. injections of 0.5 µg of LT(G33D) or 0.5 µg of LT-GM1 elicited no visible signs of an inflammatory response and were similar to injection of the PBS vehicle control (data not shown).
Lack of inflammation after i.d. injection of LT in GM2/GD2 synthase knockout mice
To further investigate the role of in vivo GM1 ganglioside binding as the underlying mechanism of LT-mediated toxicity, GM2/GD2 synthase knockout mice were studied (25). These mice are unable to synthesize complex gangliosides, including GM1, and therefore lack high affinity receptors. Injection i.d. of 0.5 µg of LT in the homozygous knockout mouse did not cause skin nodules to develop within 2 days (Fig. 2A) or later in time (not shown). However, injection of the same LT dose into heterozygous littermates produced a typical inflammatory nodule at the injection site (Fig. 2A). Histological examination of the injection site showed an absence of an inflammatory infiltrate or edema in the knockout mice (Fig. 2B), whereas tissues from the heterozygous littermates showed LT-induced edema and inflammation following i.d. injection (Fig. 2C). These results indicate that in vivo i.d. toxicity of LT is mediated through binding to complex ganglioside receptors such as GM1.
|
The use of non-GM1 ganglioside binding exotoxins to adjuvant immune responses to coadministered Ags when injected was investigated. Groups of 10 mice were i.d. injected with a suboptimal dose of 0.01 Lf TT alone or mixed with soluble GM1 ganglioside control. Other groups were immunized with TT adjuvanted with 0.005–0.5 µg of LT, LT(G33D), or LT-GM1 (1:16 molar ratio). Two weeks after three rounds of immunization (days 0, 14, and 28), serum samples were collected and analyzed for Ab titers to TT. Mice immunized with TT alone (or admixed with soluble GM1) already generated significant anti-TT titers (Fig. 3A). However, Ab titers in mice immunized with TT adjuvanted with 0.5 µg of LT, LT(G33D), or preadsorbed LT-GM1 were all significantly increased (Fig. 3A). Examination of the injection sites showed that only the group injected with all concentrations of wild-type LT developed local edema and inflammation at the site of injection as previously described.
|
To extend our findings to other GM1-binding exotoxins, the ability of detoxified CT to adjuvant humoral immune responses was tested. CT preadsorbed to GM1 increased Ab titers to coadministered TT similarly as did CT (Fig. 3C), but without inflammation.
Intradermal LT(G33D) induces complete protection against in vivo lethal tetanus toxin challenge
To confirm that the TT-specific Ab responses were functional, a highly lethal in vivo challenge model was applied (27). Groups of 10 mice were immunized i.d. two times, 2 wk apart, with 0.0025 Lf TT and/or 0.5 µg of LT(G33D). This low dose of TT by itself induced low anti-TT serum Ab titers (<1,500 ELISA units in 80% of the mice), whereas LT(G33D) induced anti-TT Ab titers above 25,000 ELISA units. Three weeks after the last immunization, 1 ng of tetanus toxin was injected s.c., and mice were monitored for 10 days. Severe paralysis or death occurred within 48 h in naive mice or mice immunized with adjuvant only (Fig. 4). Thirty percent of the mice immunized with TT alone survived, whereas mice immunized with TT plus LT(G33D) were completely protected (Fig. 4), demonstrating in vivo efficacy of the adjuvant-induced immune responses.
|
The use of LT(G33D) as adjuvant was further studied with other types of Ags. OVA is a well-studied protein commonly used as a model Ag to evaluate adjuvant efficacy in mice. Two weeks after three rounds of immunization, serum samples were collected and analyzed for Ab titers to OVA. Mice immunized with OVA alone generated low anti-OVA titers (Table I). Mice immunized with LT(G33D)-adjuvanted OVA generated 100-fold higher serum Ab titers compared with the nonadjuvanted group.
|
|
Adjuvanticity of i.d. LT(G33D) in mice with pre-existing anti-LT serum Ab titers
Apart from adjuvanting immune responses to coadministered Ags, exotoxins also induce Ab responses to themselves, which may preclude repeated use as an effective adjuvant. Analyses of LT-specific serum Ab titers in mice after a standard regimen of three i.d. injections with 0.5 µg LT(G33D) resulted in an average geometric mean anti-LT Ab titer of 50,000 ELISA units. These mice were subsequently immunized i.d. (two times, 2 wk apart) with 0.1 µg of TIV with or without 0.5 µg LT(G33D). Two weeks after the last immunization, serum samples were analyzed for influenza-specific Abs. LT(G33D) still potently enhanced specific Ab titers to all three influenza strains (Fig. 6) despite pre-existing serum anti-LT Abs.
|
LT(G33D) preparations contained low levels of endotoxin (1–2 endotoxin units per injection). To exclude a role for endotoxin in our observations, adjuvanticity of LT(G33D) was investigated in C3H/HeJ mice, which are hyporesponsive to endotoxin due to a defect in the endotoxin receptor TLR4 (34, 35). Serum analysis after two and three immunizations with TT and LT(G33D) showed adjuvant-induced enhancement of Ab titers in C3H/HeJ mice (Fig. 7A), similar to LT(G33D) adjuvanticity in normal endotoxin-sensitive C3H mice (C3H/HeOuJ).
|
Induction of T cell responses including functional CTLs by i.d. LT(G33D)
The ability of LT(G33D) to induce T cell responses was investigated in the spleens and DLNs of immunized mice. Groups of mice (n = 5) were immunized i.d. with Ag alone (100 µg OVA or 0.l µg of TIV) or with 0.5 µg of LT(G33D). Three weeks after three rounds of i.d. immunizations, mice were sacrificed and the spleens and DLNs were collected and pooled for each group. The number of IFN-- and IL4-secreting immune cells was measured by ELISPOT analysis after in vitro restimulation with whole protein Ag. In general, LT(G33D) increased both Ag-specific IFN-- and IL4-secreting immune cells in spleens and DLNs, indicating induction of mixed Th1/Th2 type T cell responses (Fig. 8, A and B).
|
-secreting immune cells was measured by ELISPOT analyses after in vitro restimulation with the OVA-specific CTL peptide SIINFEKL. LT(G33D) strongly induced CD8 T cell responses in the spleen and DLN (Fig. 8C). To validate the functionality of the induced CD8 T cell responses, the killing of specific target cells in vivo was measured (30). In short, SIINFEKL-pulsed targets were fluorescently labeled and injected in the immunized mice. The next day, the presence of injected peptide-pulsed and unpulsed targets in the spleens of individual mice was measured by FACS analyses, reflecting specific target cell killing in vivo. LT(G33D) strongly induced OVA-specific target cell killing in vivo (Fig. 8D).
In another therapeutic approach, mice inoculated in the skin with MO5 cancer cells (B16 melanoma-derived cells genetically engineered to express OVA) were subsequently immunized with OVA alone or with LT(G33D). With the adjuvant, the outgrowth of tumors was significantly delayed (Fig. 8E), reflecting the killing of tumor cell targets by the induced cellular immune responses. The eventual outgrowth of tumors in therapeutic immunization models is a well-established phenomenon explained by a variety of mechanisms such as down-regulation or elimination of Ag presentation by the tumor cells, induction of T cell anergy, defective expression of MHC class I, or production of apoptosis-inducing mediators (36, 37, 38, 39, 40, 41, 42, 43). Taken together, inclusion of LT(G33D) in i.d. immunizations induced strong cellular immune responses to coadministered Ags capable of destroying Ag-specific target cells.
Comparison of LT(G33D) as i.d. adjuvant to alum
In humans, alum is the only widespread accepted adjuvant strategy for many injected vaccines. However, in the skin alum is reactogenic, causing the formation of skin nodules (44). In addition, alum works only with certain Ags and is ineffective in generating IFN--dependent cellular immune responses (45). In a comparison study, alum adsorption was compared with LT(G33D) for its ability to generate Ab and cellular immune responses using OVA and influenza as Ags. Alum adsorption was achieved by premixing Ag with 50% (v/v) alum hydroxide gel (Rehydragel) before injection. Serum samples were analyzed after three immunizations. As shown in Table II, LT(G33D) induced significantly higher Ab titers than alum after i.d. immunization. Further functional analyses showed that LT(G33D)-induced immune responses correlated with higher neutralizing Ab titers (HAI) than alum immunization (Table III). LT(G33D) also generated higher HAI Ab titers than alum, using other parenteral routes such as i.m. injection (data not shown).
|
|
-secreting CD8 T cell responses were measured in OVA-immunized mice by ELISPOT analyses after in vitro restimulation with SIINFEKL. Strong CD8 T cell responses were induced in the spleen and DLN with LT(G33D) but not with alum (Fig. 9A). Validation of these CD8 T cell responses with the above-described in vivo CTL assay demonstrated strong induction of OVA-specific target cell killing in vivo with LT(G33D) adjuvant, but not with alum (Fig. 9B). In general, LT(G33D) induced stronger and broader immune responses than the standard alum adjuvant strategy.
|
To investigate the durability of the induced immune responses, OVA-immunized mice with or without LT(G33D) were evaluated for Ab and cellular immune responses several months after the final immunization. CTL activity was measured in individual spleens and DLNs by the in vivo CTL assay 12 wk after either a single immunization or a standard regimen of three i.d. injections. Significant long-term CTL responses were induced by LT(G33D) even after one immunization (Fig. 10A). For Ags such as OVA and TIV (see Fig. 5A), a prime-boost regimen is required to induce high levels of serum Ab responses. Mice immunized three times with OVA and LT(G33D) were analyzed for serum Abs 10 wk after the final immunization. Persisting high levels of OVA-specific serum Abs were present in mice immunized with adjuvant (Fig. 10B).
|
Previous results show that interference with GM1 binding did not compromise the adjuvanticity of exotoxins, raising the question of whether an alternative biological mechanism may exist. To study this question, mice lacking complex gangliosides were immunized i.d. As shown in Fig. 2, injection of LT in the skin is not inflammatory in these knockout mice, similarly as the injection of GM1 binding-deficient LT in wild-type mice (Figs. 1 and 2). Complex ganglioside knockout mice and age-matched heterozygous littermates were immunized with 0.5 µg of wild-type LT and coadministered 100 µg of OVA. Serum was analyzed after two and three immunizations. In both knockout and heterozygous mice, comparable high levels of serum anti-OVA Abs were generated in the presence of LT (Fig. 11A). Three weeks after the last immunizations, spleens and DLN were harvested for analyses of cellular responses by ELISPOT assays. Upon in vitro peptide restimulation, CD8 T cell responses were detected in the DLNs and spleens of both knockout and heterozygous mice (Fig. 11B). Overall, Ab and cellular immune responses upon i.d. immunization were induced in complex ganglioside knockout mice, revealing a GM1-independent mechanism of adjuvanticity underlying LT-induction of immune responses.
|
Previous studies describe the ability of wild-type LT to induce migration and activation of APCs, such as skin dendritic cells, to DLNs (46). We postulate that LT(G33D) retains this ability in a GM1-independent manner. To address the effect of i.d. LT(G33D) on APC migration and activation, we used an aqueous solution of FITC to track APCs in the DLN. Following i.d. injection of 10 µg of FITC alone or with 0.5 µg of LT(G33D), DLN cells were isolated and analyzed by flow cytometry. For characterization of APCs in DLNs, large granular cells were gated and analyzed for surface expression of CD11c. LT(G33D) increased the number of CD11c+ FITC+ cells 5-fold (Fig. 12, A and B). Of note, size-excluded cells such as lymphocytes were negative for FITC (not shown). The expression of costimulatory molecules on the FITC+ cells was investigated after costaining with anti-CD86 (Fig. 12, C and D) or MHC class II Abs (Fig. 12, E and F). The FITC+ cells demonstrated high levels of costimulatory molecules, suggesting that LT(G33D) promoted APC activation and maturation.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
In general, the potent adjuvant effects of LT and related toxins are predominantly due to the ADP-ribosylating activity of the A subunit (15, 47). Elimination of enzyme activity by deletion or mutation of the A subunit compromises adjuvanticity (15, 22). For intranasal application, the mechanism of action remains unclear after some studies reported adjuvanticity independent of ADP-ribosylation activity (48). As has been suggested, a nonspecific presence in the adjuvant preparations of contaminating endotoxin may have modulated intranasal immune responses (49, 50). Endotoxin levels in our studies were routinely monitored and found to be <10 endotoxin units per injection. This very low level of endotoxin by itself did not have significant adjuvant activity in our studies. Moreover, i.d. LT(G33D) remained a potent adjuvant in endotoxin-tolerant mice (C3H/HeJ), comparable to endotoxin-sensitive C3H strains. Thus, endotoxin was ruled out as a significant factor underlying LT(G33D)-mediated adjuvanticity in our studies.
To exert its intracellular enzymatic activity, LT (or CT) relies on binding of the B subunit to the complex ganglioside GM1, which is widely present on cellular surfaces (17, 51). Previously, prevention of B subunit binding to in vivo GM1 abolished adjuvant activity in oral immunization protocols (11). We show that parenteral adjuvanticity is not dependent on GM1. The results in the complex ganglioside knockout mouse indicate the existence of binding/uptake mechanisms other than complex gangliosides. For some time, at least for LT, other undefined binding molecules have been thought to exist, such as lactosylceramide and certain galactoproteins (52, 53). Receptors with lower affinity and/or lower expression levels may reduce adjuvant uptake to a level sufficient for immunostimulation without causing cellular injury, or other receptors may not be ubiquitously expressed but be restricted to certain target cell populations with immunostimulatory capacity, thereby leaving bystander cells unharmed and preventing widespread tissue damage. Future studies are underway to further resolve the biological mechanisms underlying the adjuvanting effects of GM1 binding-deficient exotoxins.
The present study demonstrates effective adjuvanticity of GM1-deficient LT in parenteral settings with an emphasis on i.d. immunization. The skin environment has been recognized as highly potent in stimulating immune responses. Successful i.d. vaccines have been developed against bacillus Calmette-Guérin and rabies, and recent reports on influenza highlight the advantages of this immunization route over standard i.m. injection (1, 2, 3). The presence of numerous professional APCs throughout the skin, such as epidermal Langerhans cells and dermal dendritic cells, enable induction of strong primary immune responses (54). With LT(G33D), the absence of skin inflammation raises the question of what danger signal, if any, is responsible for activating the immune system (55). Perhaps very low levels of inflammatory signaling are sufficient, or perhaps exotoxins themselves act as danger molecules by activating APCs directly and cause subsequent APC migration to the DLNs where coadministered Ags, picked up in the process, are presented to the immune system (46). We hypothesize that, whereas injected wild-type LT nonselectively binds to all epithelial cells via GM1, LT(G33D) will not do so but is still taken up by the abundant skin APCs, where the intact A subunit activity selectively activates the APCs, resulting in adjuvanticity without local inflammation.
Several i.m. administered adjuvants are licensed with their vaccines for human use. Of those, aluminum compounds (alum) are the most widely used adjuvants worldwide and the only ones licensed in North America. The major advantage of alum is its decades-long track record of safety and ability to induce significant Ab responses (44, 45). Shortcomings of alum adjuvants include the following: 1) the induction of persistent inflammatory nodules, especially in i.d. and s.c. applications; 2) induction of Abs being limited to certain Ags; and 3) inability to elicit cell-mediated immune responses (47, 48, 56, 57). Other adjuvants for human use in some parts of the world are microfluidized oil/water emulsions (MF59) and virosomes (58). Ongoing clinical trials attempt to elucidate the balance between inflammatory reactogenicity and adjuvanticity of these formulations. None of the adjuvants are evaluated specifically for their use in i.d. vaccines. Our extensive preclinical evaluation of LT(G33D) as potential i.d. adjuvant demonstrated an unusual combination of attractive characteristics such as: 1) adjuvanticity in the absence of reactogenicity; 2) potent induction of Ab responses to a variety of Ags like bacterial (TT), viral (influenza), and protein (OVA); 3) induction of functional Ag-neutralizing Abs; and 4) induction of cellular immune responses, including CTLs capable of destroying Ag-specific targets. Taken together, these findings may favor LT(G33D) as a broad spectrum candidate adjuvant for injection over traditional or experimental adjuvants currently in development.
Apart from potent immunostimulation in the absence of inflammatory side effects, other criteria apply to finding an ideal adjuvant such as an affordable, reproducible manufacturing process and long-term storage stability. In this regard, a protein such as LT(G33D) can be produced in bulk quantities in bacterial fermentation reactors and, subsequently, highly purified using inexpensive scalable column chromatography. In vitro cell assays are being developed to test and standardize potency and toxicity of the final product (59, 60). If necessary, short-term animal studies (e.g., i.d. mouse studies) quickly ensure adjuvant safety and efficacy. LT derivatives such as LT(G33D) have long-term stability in dry form at room temperature, are water soluble, and can be easily mixed with vaccine preparations of interest. Other adjuvants are often difficult to reproduce in a chemically and biologically defined way or require refrigeration and oil emulsions to maintain integrity (58).
Intradermal vaccines offer a great promise to improve on current vaccine strategies such as dose sparing for influenza or enhancement of immune responses in the elderly. To fully exploit the potential of the i.d. route, safe and potent immunostimulating compounds are needed to meet future challenges of infectious diseases in vulnerable populations. GM1 binding-deficient exotoxins are excellent candidate adjuvants for further evaluation in human i.d. vaccine trials because of their general ability to induce a wide range of effective immune responses without inflammation in preclinical animal models.
|
Disclosures |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Footnotes |
|---|
1 Address correspondence and reprint requests to Dr. J. Paul Zoeteweij, Iomai Corporation, 20 Firstfield Road, Gaithersburg, MD 20878. E-mail address: pzoet{at}iomai.com 
2 Abbreviations used in this paper: i.d., intradermal(ly); CT, cholera toxin; DLN, draining lymph node; HA, hemagglutinin; HAI, hemagglutination inhibition; Lf, limits of flocculation; LT, heat-labile enterotoxin; TIV, trivalent influenza split virus; TT, tetanus toxoid.
Received for publication August 31, 2005. Accepted for publication April 17, 2006.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
) or LT-GM1 (
). B, Swelling one day after i.d. injection of LT or LT(G33D). Data represent mean ± SD of 7–10 mice; *, p < 0.0005. C and D, Skin biopsies were taken two days after i.d. injection of the injection site for histological analyses. Sections were mounted on slides and stained with H&E. Skin injected with LT (C) or LT(G33D) (D); magnification, x100. Arrows denote the area between muscle and upper dermis. The area designated by the larger arrow in C indicates significant dermal edema compared with normal skin (D). An inflammatory infiltrate is present in the edematous area in C as indicated by the numerous purple dots.
), with bars indicating the geometric mean titers for each group. B, Immunization by different parenteral routes (i.d., i.m., and s.c.) with TT alone or mixed with 0.5 µg of LT(G33D). Mice were immunized on days 0 and 14, and serum was collected 2 wk after immunization. C, Parenteral immunization with TT alone or mixed with 0.5 µg of CT or preadsorbed CT-GM1 (0.5 µg of CT plus 0.25 µg of GM1). Mice were immunized on days 0 and 14, and serum was collected 2 wk after immunization. *, p < 0.005 vs TT alone; 
, p < 0.05 (A–C). Many individual titers in the nonadjuvanted groups were below 100 ELISA units (B and C).
