ELSEVIER

PII: SO264-410X(98)00076-0

Vaccine, Vol. 16, No. 20, pp. 2069-2076, 1998 0 1998 Elsevier Science Ltd. All rights reserved

Printed in Great Britain 0264-410X/98 $19+0.00

Musosal immunoadjuvant activity of recombinant Escherichia coli heat- labile enterotoxin and its B subunit: Induction of systemic IgG and secretory IgA responses in mice by intranasal immunization with influenza virus surface antigen

Willem R. Venveij*, Lolke de Haan”, Marijke Holtrop”, Etienne Agsteribbe”, Ruud Brands?, Guus J.M. van Scharrenburgt and Jan Wilschut*+

The Escherichia coli heat-labile enterotoxin (LT) is a very potent rnucosal immunogen. LT also has strong adjuvant activity towards coadministered unrelated antigens and is therefore of potential interest for development of mucosal vaccines. However; despite the great demand for such mucosal vaccines, the use of LT holotoxin us an adjuvant is essentially precluded by its toxicity LT is composed of an A subunit, carrying the toxic ADP-ribosylation activity, and a pentamer of identical B subunits, which mediates binding to ganglioside G ,,.,,, the cellular receptor for the toxin. In this paper; we demonstrate that recombinant enzymatically inactive variants of Lr including the LTB pentamer by itsel’ retain the immunoadjuvant activity of LT holotoxin in a murine influenza model. Mice were immunized intranasally (i.n.) with influenza virus subunit antigen, consisting mostly of the isolated surface glycoprotein hemagglutinin (HA), supplemented with either recombinant LTB (rLTB), a nontoxic LT mutant (E112K, with a Glu112-+Lys substitution in the A subunit), or LT holotoxin, and the induction of systemic IgG and local S-&A responses was evaluated by direct enzyme-linked immunosorbent assay (ELISA). Immunization with subunit antigen alone resulted in a poor systemic IgG response and no detectable S-I&4. However; supplementation of the antigen with E112K or rLTB resulted in a substantial stimulation of the serum IgG level and in induction of a strong S-&A response in the nasal cavity. The adjuvant activity of E112K or rLTB under these conditions was essentially the same as that of the LT holotoxin. The present results demonstrate that nontoxic variants of Lz rLTB in particulal; represent promising immunoadjuvants for potential application in an i.n. influenza virus subunit vaccine. Nontoxic LT variants may also be used in i.n. vaccine formulations directed against other mucosal patho- gens. In this respect, it is of interest that LT(B)-stimulated antibody responses after i.n. immunization were also observed at distant mucosal sites, including the urogenital system. This, in principle, opens the possibility to develop i.n. vaccines against sexualb transmitted infectious diseases. 0 1998 Elsevier Science Ltd. All rights reserved

Keywords: E. coli enterotoxin; LT; adjuvant; mucosal immunity; secretory IgA; intranasal influenza vaccine

*Department of Physiological Chemistry, Groningen-Utrecht Institute for Drug Exploration (GUIDE), University of Groningen, Ant. Deusinglaan 1, 9713 AV Groningen, The Netherlands. tSolvay Pharmaceuticals, C.J. van Houtenlaan 36, 1381 CP Weesp, The Netherlands. *Author to whom correspondence should be addressed. Tel: +31-50-3632733/3632740; fax: +31-50-3632728; e-mail:J.C.Wilschut@med.rug.nl (Received 19 August 1997; revised version received 20 January 1998; accepted 24 February 1998)

Vaccine 1998 Volume 16 Number 20 2069

Mucosal immunoadjuvant activity of E. coli LTB: W.R. Vetweij et al.

The heat-labile enterotoxin from Escherichia coli (LT), and its close homologue cholera toxin from Vibrio cholerue (CT), belong to the most potent mucosal immunogens and adjuvants known to date. Both molecules have been demonstrated to greatly enhance mucosal S-IgA as well as systemic antibody responses against coadministered or coupled antigens (for a review see Ref ‘). LT and CT consist of one A subunit and a pentamer of identical B subunits. The molecules exert their toxic effect through ADP-ribosylation of G,,, a regulatory protein of the adenylate cyclase system, resulting in persistent synthesis of CAMP in affected cells. The enzymatic activity of the toxin resides on the A-chain, while the LTB pentamer mediates binding of the molecules to their receptor on epithelial target cells, the ganglioside G,, (for a review, see Ref ‘).

There is an ongoing debate as to whether the mucosal immunoadjuvant activity of LT and CT is intrinsically coupled to their toxic ADP-ribosylation activity. Indeed, if adjuvanticity and toxicity were directly linked, this would essentially preclude the use of unmodified LT or CT holotoxin as adjuvants in mucosal vaccine formulations. Results from early studies suggested that CTB, isolated from CT, would retain the adjuvant properties of the holotoxin’.3. However, recent studies showed that contaminating traces of CT holotoxin in such B subunit preparations are res onsible effect? P

for the observed immunoadjuvant . On the other hand. contradictory results have

been obtained with recombinant B subunit prepara- tions. Whereas several investigators have found that recombinant CTB failed to stimulate S-IgA res

F3 onses

towards admixed antigen?“, Hazama et al. - and Deuce et al.” observed that LTB does possess mucosal immunoadjuvant activity. Another strategy for deter- mining the possible relationship between ADP-ribosylation activity and adjuvant activity involves the use of nontoxic LT mutants. However, also in such studies conflicting results have been obtained. For example, Lycke et al.’ found that an LT mutant, El 12K (Glull2+Lys), devoid of ADP-ribosylation activity, lacked adjuvant activity towards coadministered KLH. However, recently other groups have reported on mutants without ADP-ribosylation activity that did induce antibody responses towards coadministered antigensy.l”.‘~~ls,

In previous studies, we have shown that the ADP-ribosylation activity of LT is not essential for the immunogenicity of the molecule upon intranasal (i.n.) administration to mice, even though the presence of the LTA subunit, either enzymatically active or inactive, does appear to stimulate systemic IgG and secretory IgA (S-IgA) antibody responses against LTBlh.“. In the present study, we show that nontoxic variants of LT, including rLTB by itself, significantly stimulate IgG responses against an admixed unrelated antigen and also induce a strong local S-IgA response, upon i.n. administration to mice. As an antigen we chose the influenza virus subunit vaccine used for human flu vaccination purposes. Currently, this flu vaccination involves intramuscular or subcutaneous injection of the subunit material, which consists mainly of isolated hemagglutinin (HA), the main viral surface glycoprotein. The vaccination procedure is quite

2070 Vaccine 1998 Volume 16 Number 20

effective in that it elicits high systemic levels of virus- neutralizing IgG”-‘I’. These antibodies will prevent systemic spread of the virus and also, through trans- udation to the lungs, provide protection against viral pneumonia. However, systemic IgG does not usually offer protection of the mucosal surfaces of the upper respiratory tract against infection’xm”“. Local secretion of S-IgA antibodies, as induced by a natural influenza infection, appears to be essential for complete protec- tion of the upper airways”-‘“. Thus far, mucosal vaccin- ation by i.n. or oral administration of inactivated influenza virus or subunit vaccine alone, without a mucosal immunoadjuvant, has not been very successfulx,“-‘3. Moreover, in general it would appear that direct administration of nonreplicating protein antigens to the epithelial surface of the upper respira- tory tract does not result in induction of a substantial S-IgA response3’.35. Recently, Katz and coworkers’7.33 demonstrated that oral immunization with wild-type LT, in conjunction with inactivated influenza virus, induces both systemic IgG and mucosal S-IgA antibodies as well as cell-mediated immune responses, and offers protection against a nasal challenge with infectious virus.

Our present results demonstrate that nontoxic variants of LT, simply mixed with influenza virus subunit antigen, strongly stimulate systemic IgG as well as local mucosal S-IgA responses in mice after i.n. immunization. It is of particular importance that rLTB by itself, in the absence of traces of holotoxin, was found to be an equally effective mucosal immunoadju- vant as the LT holotoxin. Furthermore, S-IgA responses were found to spread throughout the common mucosal immune system. It is concluded that nontoxic variants of LT, rLTB in particular, represent promising immunoadjuvants for use in i.n. vaccine formulations against influenza or other pathogens that invade their host through mucosal surfaces.

MATERIALS AND METHODS

Production and characterization of recombinant LT, LT mutant E112K, and LTB

Construction of the plasmids encoding wild-type LT (PROFIT-LT), LTB subunit (PROFIT-LTB), and the LT mutant E112K (PROFIT-E112K) has been described previously’h,‘7. Recombinant proteins were overexpressed in E. coli MC1061’” and subsequently purified from periplasmic extracts using immobilized o-galactose (Pierce,) ,_column chromatography as described previously ‘J “. To avoid cross contamina- tion, each protein was purified on a separate pristine column. Column fractions containing purified protein were pooled, dialysed against PBS, and stored at 4°C. The protein preparations were found to be free of contaminating endotoxin in a Limulus amoebocyte lysate assay. The enzymatic and biological activity of LT, E112K and LTB was determined in a direct ADP-ribosylation assay, and a Chinese hamster ovary (CHO) cell morphology assay, respectively, as described previously”.“‘. ADP-ribosylation activity of wild-type LT, E112K and rLTB was determined using diethylamino-benzylidine-aminoguanidine (DEABAG) as a substra&“. In short, 750 ng of toxin was incubated

Mucosal immunoadjuvant activity of E. coli LT5: W. R. Vetweij et al.

in 200 mM phosphate buffer, pH 7.5, containing 20 mM D’IT, 0.1 mg ml-’ BSA, 0.1% Triton X-100 and 1.75 mM DEABAG. The reaction was started by addition of 25 ~1 of 40 nM NAD+. After incubation for 2 h at 30°C unreacted DEABAG was removed by adsorption to DOWEX resin and 500 /tl of the eluate was assayed for fluorescence at 440 nm. ADP-ribosylation activity was expressed as pmol DEABAG converted per nmol protein. The biological effect of the wild-type and mutant LT preparations and rLTB was determined in a CHO cell morphology assay as described previously’7.3X. In short, 4 x 10” CHO cells were seeded in six-well plates (Corning, Acton, USA), and cultured in Dulbecco’s Modified Eagle Medium (DMEM, Gibco, Glasgow, UK) containing 10% bovine calf serum, 100 U ml-’ penicillin G and 100 pgmll’ streptomicin sulphate. LT and LTB preparations were added in final concentralions of 10 ng, 50 ng or 1 pg. In control wells PBS was added. Cell cultures were grown for 5 days, after which the cells were washed, fixed with methanol and stained ,with a 5% GIEMSA, 1.25% crystal violet solution in water for 60 min. The cells were washed three times with water and subsequently examined microscopicall,y at an 800-fold magnification. Biological effects were Iscored positive or negative by the presence or absence of cell elongation at the different concentrations of the preparations respectively.

Influenza subunit antigen

The influenza subunit antigen was prepared from either A/Johannesburg/33/94 virus (A/Johannesburg) or B/Harbin/7/94 virus (BiHarbin), grown in embryonated chicken eggs3’. The influenza subunit material, equiva- lent to the vaccine us,ed for human immunization purposes, was a gift from Solvay Duphar (Weesp, The Netherlands). The potency of the subunit vaccine preparations, expressed as pg of HA per ml, was deter- mined in a single-radial diffusion test “.

Immunization and collection of sera and mucosal lavages

Female BALB/c mice, aged 6-8 weeks, were obtained from Harlan/Winkelmann (Germany). Groups of four animals were immunized i.n., without anaesthesia, on day 0 by application of a 20 ~1 sample on the external nares. Booster immunizations following the same procedures were given on day 7 and day 14. Mice received either 5 /lg of subunit antigen alone or admixed with either 2.9 pg LT holotoxin, LT mutant, or 2.0 /lg LTB (equimoi,ar amounts of LTB). Control mice received the same volume of PBS. Sera were collected from mice sacrificed on day 28 by severing the vena porta under pentobarbital anaesthesia. Mucosal lavages of the nasal cavity and u;;&enital tract were performed as described previously ’ . In short, nasal lavage fluids were collected by flushing 0.5 ml of PBS retrograde via the nasopharynx to the upper part of the trachea, flushing back and collecting the lavage fluid at the nostrils. Lavages of the urogenital tract were conducted by introducing and withdrawing a 100 ~1 volume of PBS ten times into the vagina using a pipettor tip. Sera were stored at -20°C and mucosal washes at 4°C.

ELISA

Influenza subunit antigen-specific antibodies were measured in a direct enzyme-linked immunosorbent assay (ELISA) as described previously’6.4’. In short, 96-well ELISA plates (Greiner) were coated overnight at 37°C with 200 ng of the antigen in coating buffer (0.05 M carbonate-bicarbonate, pH 9.6). Plates were washed and blocked by a 1% milk powder solution in coating buffer for 1 h at 37°C. Subsequently, plates were washed with PBS/Tween (PBS containing 0.5 M NaCl and 0.05% Tween20, pH 7.6) and stored at -20°C. Serum samples and mucosal washings were transferred to the plates and serially diluted twofold in PBS/Tween. After incubation for 1.5 h at 37°C plates were washed with PBSlTween and incubated with a peroxidase-conjugated goat antibody directed against either mouse IgG (Nordic) (1:6000) or mouse IgA (Southern Biotechnology Associates, Inc.) (1:4000) in PBS/Tween for 1 h at 37°C. Plates where then washed with PBS/Tween, and PBS. Specific antibodies were detected by staining with o-phenyl-diamine dihydro- chloride, as described previously’b.J . Titers were deter- mined arbitrarily as the reciprocal of the calculated sample dilution corresponding with an A492 of 0.2 above the background. Antibody titers are expressed as the geometric mean titers f SD. Comparisons between experimental groups were made by Student’s t-test. Probability (P) values co.05 were considered significant.

RESULTS

Enzymatic and biological activity of LT, E112K and rLTB

Recombinant LT, E112K and LTB were overexpressed in E. coli MC1061 and purified using immobilized o-galactose column chromatography as previously described’6s’7. All proteins were found to retain GM’-binding activity”. The enzymatic activity of LT, the E112K mutant and rLTB was investigated in a direct ADP-ribosylation assay using DEABAG as an artificial substrate (Table I). As expected, rLTB did not have any detectable ADP-ribosylation activity. In agreement with our previous observations, E112K also gave no conversion of the DEABAG substrate”. In the CHO cell assay the absence of enzymatic activity in these two preparations was confirmed since neither rLTB nor E112K showed cell elongation at either of the doses used. Wild-type LT on the other hand showed extensive CHO cell elongation at the lowest dose (10 ng) used. Our results concerning the El 12K mutant” are in agreement with the findings of Lycke et al.‘, who also found the E112K mutant to be enzymatically inactive.

Table 1 ADP-ribosylation activity of LTB, L and El 12K

ADP-ribosylation activity Toxin (pmol DEABAG/nmol protein) CHO cell elongation

Control (PBS) 0.54fO.l _

LTB 0.48kO.2 _

LT 14.6 f 0.9 +ab E112K 0.51 +0.1 _a

a No CHO cell elongation at the 10 ng and 1 pg dose. b Extensive CHO elongation already at the 10 ng dose.

Vaccine 1998 Volume 16 Number 20 2071

Mucosal immunoadjuvant activity of E. coli LTB: W.R. Vetweij et al.

Systemic antibody responses to influenza subunit vaccine using E112K, rLTB, or LT as mucosal adjuvants

The immunoadjuvant properties of these molecules were studied upon i.n. administration to mice. Groups of four mice were immunized in. without anaesthesia with 5 klg influenza subunit antigen derived from A/Johannesburg virus. The antigen was given alone or together with either 2.9 /lg of LT mutant E112K or 2.0 ~(8 rLTB (equimolar amounts of LTB) in a volume

of 20 ~11. As a positive control, a separate group received subunit antigen together with 2.9 /cg LT holotoxin. Negative controls received 20 1’1 of PBS. Booster immunizations were given on days 7 and 14, and mice were sacrificed on day 28, when serum and mucosal samples were collected. Serum IgG antibody responses were determined by ELISA.

Figure I shows the observed subunit antigen-specific serum IgG responses. In agreement with earlier obser- vation&4’, administration of influenza subunit vaccine without adjuvant gave poor systemic antibody responses. Supplementation of the antigen with LT enhanced serum antibody responses by almost three orders of magnitude, to a level similar to that seen in mice recovered from an influenza infection or mice immunized parenterally with subunit antigen”. Surpris- ingly, not only E112K but also rLTB stimulated these responses to the same extent. These results suggest that ADP-ribosylation activity is not involved in the adjuvant properties of LT. Furthermore, the presence of the A subunit does not appear to be required for adjuvanticity towards i.n. administered influenza subunit antigen.

T

HA LT rLTB E112K control

Figure 1 Serum IgG antibody responses in BALB/c mice upon

i.n. immunization with A/Johannesburg/33/94 influenza virus subunit antigen. BALB/c mice (groups of four animals) were immunized in. with either 5 119 subunit antigen alone or admixed with 2.9llg wild-type LT (LT), 2.Opg recombinant LTB (rLTB), or 2.9 /(g LT mutant El 12K. Booster immunizations were given on

day 7 and 14, and animals were sacrificed on day 28. Antibody titers are expressed as the reciprocal serum dilution with an Adg2 value of > 0.2. Admixing either LT, rLTB or El 12K significantly enhanced antibody responses compared to mice immunized with subunit antigen alone (P<O.O5); differences between the LT , rLTB and El 12K groups were not statistically significant (P > 0.05; Student’s t test). Control mice, that received PBS in.. showed no detectable influenza-specific antibody titer.

Induction of mucosal S-&A responses to influenza subunit vaccine using E112K, rLTB, or LT as adjuvants

Although high serum IgG titers against influenza are important for preventing systemic spread of the virus and protection of the lungs against infection, local S-IgA antibodies are crucial for protection of the upper respiratory tract”,“.‘5.‘h. To investigate the ability of E112K and rLTB to evoke influenza subunit antigen- specific S-IgA responses nasal washes of the above mice were analyzed.

As shown in Figure 2, rLTB induced strong nasal S-IgA responses against the subunit antigen. Interest- ingly, the observed antibody titers were comparable with those obtained with wild-type LT. Brisk influenza- specific nasal antibody responses were also induced by El 12K.

S-IgA responses at a distant mucosal effector site

Induction of a local immune response may not only result in S-IgA secretion at the site of administration, but also at other mucosal effector sites. The concept of an interconnected common mucosal immune system (CMIS)34.42 postulates that induction of a mucosal immune response at a particular site, leads to migra- tion of primed B cells, which disseminate throughout the body to secrete antigen-specific IgA at other mucosal effector sites. Although of no direct interest for influenza infections, we examined the induction of influenza-specific S-IgA antibodies in the genital tract of the above mice. Figure 3 shows that E112K and rLTB proved as effective as LT holotoxin in inducing S-IgA responses at this distant mucosal site. These data indicate that for influenza subunit vaccine prepared from strain A/Johannesburg, the rLTB has equally strong adjuvant activity as LT, capable of inducing both serum and mucosal anti-influenza antibody responses when coadministered via the i.n. route.

HA LT rLTB E112K control Figure 2 Nasal S-ISA antibody responses in BALB/c mice upon in. immunization with A/Johannesburg/33/94 influenza virus subunit antigen alone or admixed with LT, rLTB, or E112K. For details, see legend to Figure 7. Antigen-specific antibody

responses in the nasal cavity were significantly enhanced when either LT, rLTB, or E112K were admixed as mucosal adjuvants (P ~0.05); differences between these three groups were not signifi-

cant (P>O.O5). Control mice, given PBS i.n., and mice that received subunit antigen alone showed no detectable influenza- specific antibody titer.

2072 Vaccine 1998 Volume 16 Number 20

Mucosal immunoadjuvant activity of E. coli LTB: W.R. Vetweij et al.

HA LT rLTB E112K control

Figure 3 Vaginal S-ISA antibody responses in BALB/c mice upon in. immunization with A/Johannesburg/33/94 influenza virus subunit antigen alone or admixed with LT or rLTB. For details, see

legend to Figure 7. Antigen-specific antibody responses in the

vagina were significantly enhanced when either LT, rLTB, or

El 12K were admixed as mucosal adjuvant (PcO.05); differences

between these three groups were not significant (P > 0.05).

Control mice, given PBS i.r., and mice that received subunit antigen alone showed no detectable influenza-specific antibody titer.

Strain-specificity of the adjuvant activity of rLTB

The above experiments demonstrate that, in conjunction with a subunit vaccine derived from an influenza A virus, rLTB and El 12K have strong mucosal immunoadjuvant activity, very similar to that of the native LT holotoxin. In order to rule out a possible virus strain specificity, we investigated the adjuvant activity of rLTI3 also with a subunit prepara- tion from an influenza B virus (BiHarbin).

Figure 4 shows the serum IgG and mucosal S-IgA responses of mice immunized with B/Harbin subunit vaccine alone, in conjunction with rLTB or, as a control, with LT holotoxin. Both LT and rLTB induced high BiHarbin-specific systemic IgG responses (Figure 4A) and mucosal S-IgA responses in the nasal cavity (Figure 4B) as well as in the vagina (Figure 4C). Furthermore, the observed antibody responses were comparable to those obtained with the A/Johannesburg subunit vaccine, indicating that the adjuvant effect of LT and rLTB is not strain-specific. We conclude, there- fore, that rLTB represents a promising mucosal immunoadjuvant for i.n. administered influenza subunit vaccines.

DISCUSSION

The results presented in this paper demonstrate that recombinant LTB and an enzymatically inactive LT mutant, El 12K, retain the mucosal immunoadjuvant activity of native LT holotoxin when administered intranasally to mice in conjunction with influenza virus surface antigen. A strong stimulation of systemic IgG and local S-IgA responses directed against the viral antigen was observed, irrespective of whether rLTB, E112K, or LT holotoxin was used as an adjuvant. The induced antibody titers, both serum IgG and mucosal S-IgA, are comparable to those found in mice recovered from an influenza infection4”. The serum IgG

titers also compare favorably to IgG levels induced by i.m. immunization with subunit vaccine alone”, the current route of human flu vaccination, which fails to induce a mucosal S-IgA response. The S-IgA responses induced by LT(B)-supplemented subunit antigen are not restricted to the site of antigen application but are also found in distant mucosal tissues, including the urogenital tract.

LT and CT are known for their strong mucosal immunoadjuvant activity. As a consequence, much effort has been invested in delineating the potential of

HA LT rLTB control

J

/B

HA LT rLTB control

a cn2

B 0 r(

1 HA LT rLTB control

Figure 4 Serum IgG and mucosal S-ISA antibody responses in BALB/c mice upon in with B/Harbin/7/94 influenza virus subunit antigen alone or admixed with wild-type LT (LT), or recombinant

LTB (rLTB). For details see legend to Figure 7. (A) Serum IgG; (B) Nasal S-IgA) (C) Vaginal S-ISA. Admixing either LT or rLTB signifi- cantly enhanced antibody responses compared to mice

immunized with subunit antigen alone (P<O.O5); differences between the adjuvant admixed groups were not significant (P > 0.05). Control mice that received PBS in. showed no detect- able influenza-specific antibody titer.

Vaccine 1998 Volume 16 Number 20 2073

Mucosal immunoadjuvant activity of E. coli LTB: W.R. Vetweij et al.

these toxins as adjuvants in vaccines directed against pathogens which invade their host through mucosal surfaces. In this regard, there is an ongoing debate as to whether the toxic ADP-ribosylation activity of LT and CT is functionally linked to their mucosal immunoadjuvant activity. A number of studies support the view that indeed such a coupling exists. For example, Lycke et al.” used the enzymatically inactive E112K LT variant and found this mutant to lack adjuvant activity towards coadministered KLH. In the same paper, it was reported that recombinant CTB did not stimulate antibody responses towards orally administered KLH either”. The involvement of the A chain in adjuvanticity was further evidenced by the studies of Czerkinsky et a1.5 and Tamura et al.h.7, who observed that the addition of a trace of holotoxin to recombinant CTB or LTB was essential for adjuvant activity. On the other hand, several investigators have now reported on ADP-ribosylation-deficient mutants which appear to retain adjuvant activityy.“‘,‘“~‘5. Our results are in agreement with these observations. Previously, we have found that LT mutants with altered or without ADP-ribosylation activity, including E112K, are as immunogenic as the LT holotoxin upon i.n. administration to mice”. Our present study extends these observations in that now, not only E112K, but also rLTB appears to retain the mucosal immunoadju- vant activity of the holotoxin when administered i.n. to mice in conjunction with influenza virus surface antigen. It is not directly evident why in the studies of Lycke et al.’ the E112K mutant or recombinant CTB appeared inactive, whereas in our studies E112K and rLTB exhibit clear-cut adjuvant activity. We presume that the route of antigen/adjuvant administration may be crucial, intranasal in our work vs. oral in the studies of Lycke et al.“. It is interesting to note that, quite recently, Yamamoto et al.’ reported on two CT mutants, including CT E112K, that lacked ADP-ribosylation and other toxic activities, but retained immunoadjuvant activity towards ovalbumin upon subcutaneous administration to mice.

The most striking aspect of our present results relates to the potent mucosal immunoadjuvant activity of recombinant LTB pentamer alone, and raises the question by which mechanism LTB exerts its immune- stimulatory activity. In a number of studies, conjugates of B pentamer and foreign antigenic epitopes have been shown to elicit serum and/or mucosal antibody responses against the coupled antigen44-46. In these cases, it seems likely that the B pentamer functions as a carrier, targeting the coupled antigen to mucosal epithelial cells through binding to the ganglioside GM,. On the other hand, there are only few studies where antigen and recombinant B pentamer have been simply mixed to result in stimulation of the antibody response towards the coadministered antigen’.“,“. Our study demonstrates that i.n. administered recombinant LTB acts as a mucosal immunoadjuvant when freely mixed with influenza subunit vaccine. This finding is incon- sistent with a carrier function of LTB. Evidently, LTB does possess bona fide immunoadjuvant activity which is expressed clearly upon i.n. administration together with an unrelated antigen. We have recently found that this adjuvant activity is not restricted to influenza subunit antigen, but is also evident with other

2074 Vaccine 1998 Volume 16 Number 20

antigensJ7. Furthermore, using an LTB mutant lacking the capacity to bind to GM,, we have established that G,,-binding affinity is crucial for the immunoadjuvant activity of LTB4’.

With regard to influenza vaccines in particular, the potential use of LTB or CTB as mucosal adjuvants to stimulate intranasal S-IgA antibody induction has been thoroughly COWOrkerS~,h.‘.~‘-~l,,~~.~~~~

investigated by Tamura and Using recombinant LTB or CTB,

these investigators found that supplementation of a trace of holotoxin was essential for adjuvant activity of the B pentamer6.‘. Our present observations, which clearly demonstrate that rLTB alone markedly stimu- lates antibody responses to admixed influenza subunit antigen, are in disagreement with these results. It is not clear what the reason for the discrepancy is. It is unlikely that it is due to differences in antigen dose and/or immunization regimen. Contamination of the subunit antigen with neuraminidase or with endotoxin could possibly be responsible’“. However, the subunit vaccine from Solvay Duphar (Influvac) has extremely low contents of neuraminidase and endotoxin’“, thereby minimizing a possible role for these components in the observed mucosal antibody responses. It is also unlikely that endotoxin contamination of our LTB preparation is critically involved. As indicated above, we have recently characterized the adjuvant activity of an LTB mutant lacking binding affinity for GM,; this mutant was prepared in the same way as the native LTB and contained an equal (low) endotoxin content in a Limulus assay, but lacked adjuvant activity. Another possibility is that differences in the precise nature of the B subunit preparations used underlie the discrep- ancy. While our LTB gene is derived from an E. coli strain of porcine origin, Tamura and coworkers have used either CTB, or LTB derived from a human E. co/i

strain6.7. We are currently investigating the adjuvant activity of B subunits of different origin towards i.n. administered influenza subunit antigen.

Induction of S-IgA in the nasal cavity is important for protection against infection of the upper respiratory tract with influenza virush,7.‘9.23-3”. The current, intra- muscular, flu vaccination induces high levels of influ- enza-specific systemic IgG, which provides adequate protection against potential serious consequences of an influenza infection such as viral pneumonia. However, it fails to induce an S-IgA response. Clearly, the efficacy of current influenza subunit vaccines would be improved if these vaccines were to induce a local mucosal immune response in addition to systemic IgG, thereby providing a better protection against primary infection of the upper respiratory tract. An influenza subunit preparation simply mixed with rLTB would meet such requirements and, therefore, would repre- sent a readily accessible and suitable vaccine formula- tion. Furthermore, it would be quite convenient, since the vaccine would be administered intranasally, for example by nose drops. The protective efficacy of the rLTB-supplemented influenza subunit vaccine in mice is currently being evaluated in a challenge study in our laboratory. Since rLTB also has the capacity to induce S-IgA responses in the vagina, we are also exploring the possibilities of developing vaccines against genital and sexually transmitted infectious diseases using i.n. administered vaccines.

Mucosal immunoadjuvant activify of E. coli LTB: W.R. Verweq et al.

ACKNOWLEDGEMENTS

This study was supported by the Netherlands Organiza- tion for Scientific Research (NWO) under the auspices of the Council for Mledical Research (GB-MW) (Research Grant 900-515-048, supporting WRV).

REFERENCES

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

Snider, D.P. The mucosal adjuvant activities of ADP-ribosylating bacterial enterotoxins. Crit Rev. /mmunol. 1995, 15,317-348. Spangler, B.D. Structure and function of cholera toxin and the related Escherichia co/i heat-labile enterotoxin. Microbial. Rev. 1992,56,622-647. Wilson, A.D., Clarke, C.J. and Stokes, CR. Whole cholera toxin and B subunit act synergistically as an adjuvant for the mucosal immune response of mice to keyhole lympet haemo- cyanin. &and. J. Immunol. 1990,31,443-451. Tamura, S-l., Samegai, Y., Kurata, H., Nagamine, T., Aizawa, C. and Kurata, T. Protection against influenza virus infection by a vaccine inoculated with cholera toxin B subunit. Vaccine 1988,6,409-415. Czerkinsky, C., Russell, M.W., Lycke, N., Lindblad, M. and Holmgren, J. Oral administration of a streptococcal antigen coupled to cholera toxin B subunit evokes strong antibody responses in salivary glands and extramucosal tissues. Infect. Immun. 1989, 57,1072-l C77. Tamura, S-I., Yamanaka, A., Shimohara, M., Tomita, T., Komase, K., Tsuda, Y., Suzuki, Y., Nagamine, T., Kawahara, K.. Danbara, H., Aizawa, C., Oya, A. and Kurata, T. Synergistic action of cholera toxin B subunit (and fscherichia co/i heat- labile toxin B subunit) and a trace amount of cholera whole toxin as an adjuvant for nasal influenza vaccine. Vaccine 1994, 12,419-426. Tamura, S.-I., Asanuma, H., Tomita, T., Komase, K., Kawahara, K., Danbara, H., Hattori, N., Watanabe, K., Suzuki, Y., Nagamine, T., Aizawa, C., Oya, A. and Kurata, T. Escheri- chia co/i heat-labile enterotoxin B subunits supplemented with a trace amount of the l-olotoxin as an adjuvant for nasal influenza vaccine. Vaccine 1994, 12, 1083-1089. Lycke, N., Tsuji, T. and Holmgren, J. The adjuvant effect of Vibrio cholerae and Escherichia co/i heat-labile enterotoxins is linked to their ADP-ribosyiransferase activity. Eur. J. fmmunol. 1992, 22,2277-2281. Yamamoto, S., Takeda, Y., Yamamoto, M., Kurazono, H., Imaoka, K., Yamamoto, M., Fujihashi, K., Noda, M., Kiyono, H. and McGhee, JR. Mutants in the ADP-ribosyltransferase cleft of cholera toxin lack diarrheagenicity but retain adjuvanticity. J. Exp. Med. 1997, 185, 1203-1210. Yamamoto, S., Kiyono, H., Yamamoto, M., Imaoka, K., Yamamoto, M., Fujihashl, K., VanGinkel, F.W., Noda, M., Takeda, Y. and McGhee, JR. A nontoxic mutant of cholera toxin elicits Th-2 type responses for enhanced mucosal immunity. Proc. Nat/. Acad. Sci. USA. 1997, 94, 5267-5272. Deuce, G., Fontana, M., IPizza, M., Rapuoli, R. and Dougan, G. lntranasal immunogenicity and adjuvanticity of site- directed mutant derivatives of cholera toxin. Infect. Immun. 1997,65,2821-2828. Hazama, M., Mayumi-Aoro, A., Miyazaki, T., Hinuma, S. and Fujisawa, Y. lntranasal immunization against herpes simplex virus infection by using a. recombinant glycoprotein D fused with immunomodulating proteins, the B subunit of Escherichia co/i heat-labile enterotoxin and interleukin-2. immunology 1993,78,643-649. Dickinson, B.L. and Clements, J.D. Dissociation of Escheri- chia co/i heat-labile enterotoxin adjuvanticity from ADP-ribosyltransferase activity. Infect. Immun. 1995, 63, 1617-1623. Di Tommaso, A., Saletti, G., Pizza, M., Rappuoli, R., Dougan, G., Abrignani, S., Deuce, G. and De Magistris, M.T. Induction of antigen-specific antibodies in vaginal secretions by using a nontoxic mutant of heat-labile enterotoxin as a mucosal adjuvant. Infect. Immun. 1996, 84, 974-979. Deuce, G., Turcotte, C., Cropley, I., Roberts, M., Pizza, M., Domenghini, M., Rappuoli, R. and Dougan, G. Mutants of Escherichia co/i heat-labile toxin lacking ADP-ribosyl-

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

transferase activity act as nontoxic, mucosal adjuvants. hoc.

Nat/. Acad. Sci. USA 1995, 92, 1644-l 648. De Haan, L., Holtrop, M., Verweij, W.R., Agsteribbe, E. and Wilschut, J. Mucosal immunogenicity of the Escherichia co/i heat-labile enterotoxin: role of the A subunit. Vaccine 1996, 14,260-266. De Haan. L.. Verweij, W.R., Feil, I.K., Lijnema, T.H., HOI, W.G.J., Agsteribbe, E. and Wilschut, J. Mutants of the Escher- ichia co/i heat-labile enterotoxin with reduced or without ADP-ribosylating activity retain the immunogenic properties of the native holotoxin. infect. Immun. 1996, 64, 5413-5416. Clancy, R.L. Mucosal vaccines for the prevention of influenza. Drugs 1995,50, 587-594. Ramphal, R., Cogliano, R.C., Shands, J.W. and Small, P.A. Jr. Serum antibody prevents lethal murine influenza pneumonitis but not tracheitis. infect. Immun. 1979, 25, 992-997. Small, P.A. Jr., Waldman, R.H., Bruono, J.C. and Giford, G.E. Influenza infection in ferrets: Role of serum antibody in protection and recovery. Infect. Immun. 1976, 13, 417-424. Clements, M.L., Betts, R.F., Tierney. E.L. and Murphy, B.R. Serum and nasal wash antibodies associated with resistance to experimental challenge with influenza A wild-type virus. J. C/in. Microbial. 1986, 24, 157-160. De Haan, A., Van Scharrenburg, G.J.M., Masihi, K.N., Bender, B.S., Small, P.A. Jr. and Wilschut, J. Protective immunity induced by liposome-supplemented influenza virus subunit antigen administered intranasally to mice. Vaccine Res. 1996, 4,207-216. Liew, F.Y., Russell, S.M., Appleyard, G., Brand, C.M. and Beale, J. Cross-protection in mice infected with influenza A virus by the respiratory route is correlated with local IgA rather than serum antibody or cytotoxic T cell reactivity. 1984. Eur. J. Immunol. 1984, 14,350-356. Meitin, C.A., Bender, B.S. and Small, P.A. Jr. Influenza immunization: intranasal live vaccinia recombinant contrasted with parenteral inactivated vaccine. Vaccine 1991, 9, 751-756. Renegar, K.B. and Small, P.A. Jr. Passive transfer of local immunity to influenza virus infection by IgA antibody. J. Immunol. 1991,148,1972- 1978. Renegar, K.B. and Small, P.A. Jr. lmmunoglobulin A media- tion of murine nasal anti-influenza virus immunity. J. Viral. 1991,65,2146-2148. Tamura, S.-l., Funato, H., Hirabayashi, Y., Kikuta, K., Suzuki, Y., Nagamine, T., Aizawa, C., Nakagawa, M. and Kurata, T. Functional role of respiratory tract haemagglutinin-specific IgA antibodies in protection against influenza. Vaccine 1990, 8, 1479-l 485. Tamura, S.-l., Funato, H., Hirabayashi, Y., Suzuki, Y., Nagamine, T., Aizawa, C. and Kurata, T. Cross-protection against influenza A virus infection by passively transferred respiratory tract IgA antibodies to different hemagglutinin molecules. Eur. J. Immunol. 1991.21, 1337-1344. Tamura, S-l., Ito, Y., Asanuma, H., Hirabayashi, Y., Suzuki, Y., Nagamine, T., Aizawa, C. and Kurata, T. Cross-protection against influenza virus infection afforded by trivalent inacti- vated vaccines inoculated intranasally with cholera toxin B subunit. J. Immunol. 1992, 149, 981-988. Tomoda, T., Morita, H., Kurashige, T. and Maassab, H.F. Prevention of influenza by the intranasal administration of cold-recombinant, live-attenuated influenza virus vaccine: Importance of interferon-y production and local IgA response. Vaccine 1995, 13, 185-190. Glathe, H. and Lange, W. Influenza vaccination in older patients: immunogenicity, epidemiology and available agents. Drugs and Aging 1995,6,368-387. Katz, J.M., Lu, X., Young, S.A. and Galphin, J.C. Adjuvant activity of the heat-labile enterotoxin from enterotoxigenic Escherichia co/i for oral administration of inactivated influenza virus vaccine. J. Infect. Dis. 1997, 175, 352-363. Katz, J.M., Lu, X., Galphin, J.C.. Clements, J.D. 1996. Heat- labile enterotoxin from Escherichia co/i as an adjuvant for oral influenza vaccination. In: Options for the Control of influenza (Eds Brown, L.E., Hampson, A.W. and Webster R.G.). Elsevier Science, New York. McGhee, JR., Mestecky, J., Dertzbaugh, M.T., Eldridge, J.H., Hirasawa, M. and Kiyono, H. The mucosal immune system: from fundamental concepts to vaccine development. Vaccine 1992, 10,75-88.

Vaccine 1998 Volume 16 Number 20 2075

Mucosal immunoadjuvant activity of E. coli LTB: W.R. Vetweij et al.

35

36

37

38

39

40

41

42

43

Revillard, J.P., Cozen, G. and Czerkinsky, C. Oral administra- liposome-supplemented influenza subunit antigen: tion of immunomodulators and the mucosal immune system. comparison with responses induced by conventional Dev. Biol. Stand. 1992, 77, 31-37. immunization or influenza infection. In press. Meissner, P.S., Sisk, W.P. and Berman, M.L. Bacteriophage lambda cloning system for the construction of directional cDNA libraries. froc. Nat/. Acad. Sci. USA 1987, 84, 4171-4175.

44

Uesaka, Y., Otsuka, Y., Lin, Z., Yamasaki, S., Yamaoka, J., Kurazone, H. and Takeda, Y. Simple method of purification of Escherichia co/i heat-labile enterotoxin and cholera toxin using immobilized galactose. Microb. Path. 1994, 16, 71-76. Feil, I.K., Reddy, R., De Haan, L., Merrit, E.A., Van den Akker, F., Storm, DR. and HOI, W.G.J. Protein engineering studies indicate the crucial importance of the A chain loop 47 to 56 for cytotoxicity of E. co/i heat-labile enterotoxin. Mol. Micro- biol. 1996, 20, 823-832. Chaloupka, I., Schuler, A., Marschall, M. and Meier-Ewert, H. Comparative analysis of six european influenza vaccines. Eur. J. Mi&obio/. Infeci. Dis. 1996, 15, 121-127. Wood. J.M.. Schild, G.C.. Newman, R.W. and Seaaroatt. V. An improved single radial immunodiffusion technic@ for the assay of influenza hemagglutinin antigen: Application for potency determination of inactivated whole virus and subunit vaccines. J. Biol. Stand. 1997, 5, 237-241. De Haan, A., Geerligs, H.J., Huchshorn, J.P., Van Scharren- burg, G.J.M., Palache, A.M. and Wilschut, J. Mucosal immunoadjuvant activity of liposomes: Induction of systemic IgG and secretory IgA responses in mice by intranasal immunization with an influenza subunit vaccine and coadmi- nistered liposomes. Vaccine 1995, 13, 155-l 62. McDermott, M.R. and Bienenstock, J. Evidence for a common mucosal immunologic system. I. Migration of B immunoblasts into intestinal, respiratory, and genetical tissues. J. Immunol. 1979,122, 1892-1898. De Haan, A., Van Scharrenburg, G.J.M., Masihi, K.N. and Wilschut. J. Systemic and local IgG and IgA antibody responses in mice induced by intranasal immunization with

45

46

Hajishengallis, G., Hollingshead, S.K., Koga, T. and Russell, M.W. Mucosal immunization with bacterial protein antigen genetically coupled to cholera toxin A2/B subunits, J. Immunol. 1995,154, 4322-4332. Lipscombe, M., Charles, LG., Roberts, M., Dougan, G., Tite, J. and Fairweather, N.F. lntranasal immunization using the B subunit of the Escherichia co/i heat-labile toxin fused to an epitope of the Bordetella pertussis P.69 antigen. Mol. Micro- biol. 1991, 5, 1385-1392. Stok, W., Van der Heijden, P.J. and Bianchi, A.T.J. Conver- sion of orally induced suppression of the mucosal immune response to ovalbumin into stimulation by conjugating ovalbumin to cholera toxin or its B subunit. Vaccine 1994, 12, 521-526.

47

48

49

50

Dettaan, L., Feil, I.K., Verweij, W.R., Holtrop, M., HOI, W.G.J., Agsteribbe, E. and Wilschut, J. Mutational analysis of the role of ADP-ribosylation activity and G (Ml) binding activity in the adjuvant properties of the Escherichia co/i heat-labile enter- otoxin towards intranasally administered keyhole limpet hemocyanin. Euro. J. Immunol. 1998, 28, 1243-1250. Hirabayashi, Y., Kurata, H., Funato, H., Nagamine, T., Aizawa, C., Tamura, S-l., Shimada, K. and Kurata, T. Comparison of intranasal inoculation of influenza HA vaccine combined with cholera toxin B subunit with oral or parenteral vaccination. Vaccine 1990, 8, 243-248. Kikuta, K.. Hirabayashi, Y., Nagamine, T., Aizawa, C., Ueno, Y.. Oya, A., Kurata, T. and Tamura, S.-l. Cross-protection against influenza B type virus infection by intranasal inocula- tion of the HA vaccines combined with cholera toxin B subunit. Vaccine 1990, 8, 595-599. Deroo, T., Min You, W. and Fiers, W. Recombinant neuramini- dase vaccine protects against lethal influenza. Vaccine 1996, 14,561-569.

2076 Vaccine 1998 Volume 16 Number 20