Intranasal or oral immunization of inbred and outbred mice with murine or human rotavirus VP6 proteins protects against viral shedding after challenge with murine rotaviruses

Vaccine 20 (2002) 3310–3321

Intranasal or oral immunization of inbred and outbred mice with murine or human rotavirus VP6 proteins protects against viral shedding after challenge with murine rotaviruses Anthony H. Choi a , Monica M. McNeal a , Mitali Basu a , Jason A. Flint a , Susan C. Stone a , John D. Clements c , Judy A. Bean b , Stacey A. Poe b , John L. VanCott a , Richard L. Ward a,∗ a

c

Division of Infectious Diseases, Children’s Hospital Medical Center, Cincinnati, OH 45229, USA b Department of Biostatistics, Children’s Hospital Medical Center, Cincinnati, OH 45229, USA Department of Microbiology and Immunology, Tulane University Medical Center, New Orleans, LA 70112, USA Received 21 January 2002; received in revised form 14 May 2002; accepted 8 June 2002

Abstract Intranasal (i.n.) administration of an Escherichia coli-expressed chimeric VP6 protein from the EDIM strain of murine rotavirus to adult BALB/c (H-2d ) mice along with LT(R192G), an attenuated mutant of the mucosal adjuvant E. coli heat-labile toxin, has been found to consistently stimulate ca. 99% reductions in rotavirus shedding after subsequent EDIM challenge. This study was designed to determine the robustness of this protection, i.e. can VP6 immunization consistently protect against shedding in this model, thus, providing an indication of its potential as a vaccine. Intranasal immunization with two 8.8 ␮g doses of EDIM VP6 and 10 ␮g of LT(R192G) was found to stimulate 99% reductions in EDIM shedding in four additional strains of inbred mice belonging to three haplotypes, i.e. DBA/2 (H-2d ), C57BL/6 (H-2b ), 129 (H-2b ) and C3H (H-2k ). Protection stimulated against EDIM antigen shedding following i.n. immunization with VP6 from the human CJN strain was less (P = 0.02) than induced by EDIM VP6 (86% versus 99%), but no further loss of protection was observed when the dose of CJN VP6 was reduced 100-fold. Protection against EDIM shedding was also maintained after i.n. immunization of three strains of outbred mice (CF-1, CD-1 and Swiss Webster) with either EDIM or CJN VP6, i.e. EDIM VP6 immunization reduced EDIM shedding by 99% while CJN VP6 immunization produced reductions of 86–96%. Protection stimulated by oral immunization of BALB/c mice with two 8.8 ␮g doses of either VP6 chimera plus LT(R192G) was not significantly different from that induced by i.n. immunization. Finally, protection found after either oral or i.n. immunization with EDIM or CJN VP6 was no different when the mice were challenged with McN, another strain of murine rotavirus. These results support further evaluation of VP6 as a vaccine. © 2002 Elsevier Science Ltd. All rights reserved. Keywords: Rotavirus; VP6 vaccines; Mouse model

1. Introduction Rotaviruses are the primary cause of severe gastroenteritis in young children and are responsible for approximately 500,000 deaths in the world annually [1]. Vaccines evaluated for their abilities to prevent rotavirus disease have all been live, attenuated rotavirus strains that are delivered orally [2]. None have provided consistent protection, even against severe rotavirus disease. Furthermore, side effects associated with these vaccine candidates may limit their usage. The one licensed rotavirus vaccine was removed from the USA market in 1999, less than 1 year after its introduction into the routine childhood vaccination series, due to its association with intussusception [3]. To avoid the potential difficulties ∗

Corresponding author. Tel.: +1-513-636-7628; fax: +1-513-636-7682. E-mail address: [email protected] (R.L. Ward).

of live vaccines, non-living rotavirus vaccine candidates are being evaluated in animal models. Many passive protection studies have been performed following immunization with the rotavirus neutralization proteins, VP4 and VP7, or their peptides, but these have provided limited information concerning active immunity. More recently, active immunization studies performed in rabbits or mice immunized parenterally with either inactivated rotavirus particles or virus-like particles (VLPs) have provided evidence that mucosal exposure is not a prerequisite for protection elicited by non-living rotavirus vaccines. Intramuscular injection of rabbits with either inactivated rotavirus strain SA11 or VLPs lacking both neutralization proteins (i.e. 2/6 VLPs), along with appropriate adjuvants, stimulated partial protection against rotavirus shedding following oral challenge with a lapine rotavirus [4,5]. Mice immunized parenterally with either triple- or double-layered

0264-410X/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 2 6 4 - 4 1 0 X ( 0 2 ) 0 0 3 1 5 - 8

A.H. Choi et al. / Vaccine 20 (2002) 3310–3321

inactivated rotavirus particles of homologous or heterologous origin were partially, and sometimes, fully protected against rotavirus shedding following challenge with a murine rotavirus [6,7]. Again, inclusion of adjuvant during immunization significantly increased the level of protection [7]. It was later shown that intranasal (i.n.) immunization of mice with 2/6 VLPs or double-layered inactivated rotavirus particles lacking the outer capsid neutralization proteins, along with a powerful mucosal adjuvant, stimulated nearly complete protection against murine rotavirus shedding [8–10]. In contrast, i.n. immunization of gnotobiotic piglets with 2/6 VLPs and adjuvant elicited significant protection against diarrhea induced by a human rotavirus only if the piglets were first primed by infection with an attenuated human rotavirus strain [11–13], a result that has controversial explanations. The most immunogenic protein in both 2/6 VLPs and double-layered rotavirus particles is VP6, the protein that comprises the intermediate capsid layer of the rotavirus particle. Because these particles elicited such dramatic protection of mice against rotavirus shedding, an Escherichia coli-expressed chimeric protein containing the VP6 protein of murine rotavirus strain EDIM was evaluated as a new vaccine candidate in the mouse model. Intranasal immunization of BALB/c mice with this protein along with a genetically-attenuated E. coli heat-labile toxin [LT(R192G)] as adjuvant induced nearly complete protection against rotavirus shedding after a subsequent EDIM inoculation [14]. Before further evaluating this candidate vaccine in other animal models or in humans, it was imperative to demonstrate its general utility in the mouse model. In the studies reported here, we measured the protection against EDIM shedding in mice belonging to five inbred and three outbred strains after i.n. inoculation with EDIM VP6 together with LT(R192G). We also assessed the level of protection after oral immunization of mice with EDIM VP6 and determined whether protection against EDIM shedding was altered when the immunizations were performed with a VP6 protein from a human rotavirus. The results of these studies substantiate the need to further evaluate the rotavirus VP6 protein as a vaccine candidate in other animal models. 2. Materials and methods 2.1. Rotavirus strains used to challenge mice Two stains of wild type murine rotaviruses, EDIM and McN, were used to challenge adult mice in these studies. These two strains were found to have distinctly different genotypes and phenotypes (McNeal et al., in preparation). Sequence analysis of the VP7, VP4 and VP6 protein genes revealed 16, 37, and 8 amino acid differences, respectively, between these rotavirus strains. Furthermore, the EDIM strain was found to belong to subgroup II while McN was typed as subgroup I. The EDIM strain was originally obtained from the stool of an infected mouse and provided by M. Collins (Microbiological Associates, Bethesda, MD)

3311

in 1980. The EDIM preparation used for challenge was passed solely in neonatal mice. To make this preparation, BALB/c mice, approximately 4 days of age, were orally inoculated and their stools were collected twice daily into Earle’s balanced salt solution (EBSS) for the subsequent days until diarrhea abated. Later, this stool material was thawed, combined, mixed with an equal volume of cold freon, and blended. After collection of the aqueous phases, they were combined and centrifuged to pellet the virus. The virus pellet was re-suspended in tissue culture medium and stored in aliquots at −70 ◦ C. The McN strain was obtained from the stool of an infected mouse housed at Cincinnati Children’s Hospital Animal Facility. The preparation of wild type McN used for challenge was made in the same manner as described for the EDIM strain. 2.2. Chimeric rotavirus VP6 proteins used to immunize mice The genes for the VP6 proteins used to immunize mice in this study were obtained from EDIM and the human rotavirus strain CJN. The wild type EDIM strain was adapted to grow in cell culture [15] and the VP6 gene of a triply plaque-purified preparation of the culture-adapted virus was cloned for expression in E. coli as described previously [14]. The CJN strain, a G1P[8] (subgroup II) rotavirus, was obtained from the stool of an 8-month-old child hospitalized at Cincinnati Children’s Hospital Medical Center in 1982. This strain was also adapted to grow in cell culture as previously described [16]. 2.3. Cloning of CJN VP6 To clone the VP6 gene of CJN, a triply plaque-purified preparation of the virus was grown in the MA104 monkey kidney cell line and the double-stranded (ds) RNA genome segments were purified by methods already described [14]. Double-stranded cDNAs of gene 6 were generated by RT/PCR using the ThermoScript RT-PCR System Plus Platinum Taq DNA Polymerase kit (Invitrogen Life Technologies, Carlsbad, CA). Briefly, 1 ␮g of purified ds CJN genomic RNA was denatured (3 min, 94 ◦ C) in 25 ␮l of distilled water containing 200 nM of a forward primer (5 -GGC TTT AAA ACG AAG TCT TCG AC) and of a reverse primer (5 -GGT CAC ATC CTC TCA CTA CAT C) along with an RNase inhibitor (20 units of RnaseOut, Invitrogen). The primers used were based on the terminal sequences of the 5 and 3 non-coding regions (NCR) of gene 6 of rotavirus human strain Wa [17]. The primers were allowed to rapidly anneal to denatured viral RNA by cooling the suspension on ice. Two times ThermoScript Reaction mix and ThermoScript Plus/Platinum Taq Enzyme mix (25 ␮l each) were added to the denatured RNA. Reverse transcription was carried out at 50 ◦ C for 60 min and immediately followed by 30 cycles (1 min, 94 ◦ C; 2 min, 50 ◦ C; and 1.5 min, 72 ◦ C) of PCR. After the last cycle, a final extension step

3312

A.H. Choi et al. / Vaccine 20 (2002) 3310–3321

(10 min, 72 ◦ C) was included. The nucleotide sequence of the CJN VP6 gene was then determined using ds cDNA generated by RT/PCR as templates. Sequencing was performed by Cleveland Genomics (Cleveland, OH) using an ABI DNA Sequencer (Applied Biosystems, Foster City, CA). The primer pair used for generating dsDNA was also used to sequence the initial 500 nucleotides from the ends of the positive and the negative cDNA strands. New primer pairs were designed using the sequence generated until both cDNA strands were completely sequenced. The gene sequence was determined twice using PCR products generated in separate reactions using both strands as templates. 2.4. Construction of recombinant expression plasmids containing CJN VP6 Standard cloning procedures were then used to construct recombinant expression plasmids. The open reading frame (ORF) in the sequenced CJN VP6 sequence was identified using the Omiga nucleotide analysis software (Accelrys, version 2.0, Princeton, NY). A forward primer (ATG GAG GTT TTA TAC TCA TTG) corresponding to the first 21 nucleotides and a reverse primer (TCA CTT AAT CAA CAT GCT TC) corresponding to the last 20 nucleotides of the ORF of CJN VP6 sequence were used to generate ds cDNA containing the ORF by PCR. The generated DNA were cloned into the XmnI site of the expression plasmid pMAL/c2X (New England Biolabs, Beverly, MA) according to methods already described [14]. The cloned VP6 sequence was placed downstream from the E. coli malE gene, which encodes the maltose-binding protein (MBP), and immediately after the factor Xa proteolytic cleavage site, which consists of the amino acid sequence Ile–Glu–Gly–Arg. pMAL/c2X utilizes the strong tac promoter and the malE translation initiation signals for expression of fusion proteins. The plasmid contains the gene for ampicillin resistance to recombinant bacteria and a lacZ-␣ gene sequence. Insertional inactivation of lacZ-␣ allows blue-to-white selection of recombinants with inserts. Following ligation of cDNA and XmnI-digested pMAL/c2X, recombinant pMAL/c2X plasmids were transformed into protease-deficient E. coli BL21 cells (Stratagene, La Jolla, CA) which were grown on agar plates. Numbers of white colonies of bacteria grown in the presence of IPTG (isopropyl-b-d-thiogalactopyranoside) and X-Gal (5-bromo-4-chloro-3-indolyl-b-d-galactopyranoside) on replicate plates were noted, and the corresponding clones were selected from replicate plates for further screening by PCR for gene identity and orientation. Recombinant plasmids were sequenced to ultimately confirm the authenticity of the rotavirus gene sequences. 2.5. Expression of recombinant CJN VP6 in E. coli Expression of recombinant VP6 in E. coli was induced essentially as previously described [14]. Single colonies

of recombinant bacteria expressing CJN VP6 were grown as overnight cultures (37 ◦ C) in 50 ml of rich broth containing 100 mg ampicillin/l. On the following day, 10 ml of the overnight cell culture were inoculated into 1 l of rich broth. When the optical density (A600 ) reached approximately 0.6, IPTG was added to give a final concentration of 0.3 mM to induce expression of fusion proteins. At 3 h post-induction, an aliquot was withdrawn and pelleted in a micro-centrifuge (2 min, 4 ◦ C). The cell pellet was re-suspended in SDS-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer and analyzed by SDS-PAGE. The remaining cell suspension was centrifuged (4000 × g, 20 min, 4 ◦ C) to harvest the cells, which were washed in phosphate-buffered saline (PBS), centrifuged again, and frozen at −20 ◦ C. To prepare soluble chimeric proteins, frozen bacteria containing expressed chimeric MBP::VP6 were processed according to the method of Jarrett and Foster [18]. In short, the bacterial pellets were thawed and re-suspended in 50 ml of buffer L (5 mM NaH2 PO4 , 10 mM Na2 HPO4 , 30 mM NaCl, 10 mM ␤-mercaptoethanol, 0.2% Tween 20, 1 mM phenylmethylsulfonyl fluoride, 25 mM benzamidine, 200 mg lysozyme/l). After digestion, the suspensions were sonicated in an ice water bath. NaCl and RNase A (final concentrations of 26.5 mg/ml and 5 ␮g/ml, respectively) were then added. The lysates were centrifuged (54,000 × g, 30 min) to separate insoluble cell debris from supernatants (soluble fraction) which contained rotavirus proteins. 2.6. Purification of recombinant VP6 fusion proteins Fusion proteins in the soluble fractions were purified by affinity chromatography. Amylose resin (New England Biolabs) was used to purify chimeric proteins containing MBP. The resin was prepared by washing twice with 8 vol. of buffer C (buffer L containing 0.5 M NaCl). For each wash, the mixture was rocked for 30 min at 4 ◦ C and the resin was recovered by centrifugation (2100 × g, 5 min). The supernatants, which contained the fusion proteins, were mixed with amylose resin for 2 h. After centrifugation (2100 × g, 5 min), the resin was recovered and re-suspended in 50 ml of buffer C, rocked for 30 min, and centrifuged to recover the resin. The resin was washed in this manner three times and finally washed overnight with 500 ml of buffer C. On the following day, the resin was recovered by centrifugation (2100 × g, 5 min), re-suspended in 50 ml of buffer D (50 mM Tris–HCl [pH 7.5], 50 mM NaCl, 1 mM EDTA, 10 mM ␤-mercaptoethanol, 1 mM phenylmethylsulfonyl fluoride), and rocked for 30 min. The resin was pelleted by centrifugation (2100 × g, 5 min) and the bound fusion protein was eluted from the resin with 250 ml of 15 mM maltose in buffer D for 2 h. The resin was removed by centrifugation (2100 × g, 5 min), and the supernatant containing the fusion proteins was subjected to buffer exchange to PBS while simultaneously being concentrated by ultrafiltration with a stirred-cell concentrator (model 8400; Amicon Inc.,

A.H. Choi et al. / Vaccine 20 (2002) 3310–3321

Beverly, MA). The concentrations of purified proteins were measured by the Bradford method. The quality of the purified proteins was then analyzed by Western blot in the same manner as reported previously [14] with essentially identical results. That is, the MBP-containing purified product was found to be highly truncated, ranging in size from 87.7 kDa (the size of the complete MBP::VP6 protein) down to the size of MBP itself (42.7 kDa) as was found previously for chimeric EDIM VP6 [14]. 2.7. Properties of the adjuvant LT(R192G) The adjuvant used in all immunizations with the VP6 chimeric proteins in this study was an attenuated mutant of the E. coli heat-labile toxin (LT). In the intestine, wild type LT is normally activated by protease cleavage that simultaneously potentiates its toxigenic properties. A mutant of LT [i.e. LT(R192G)] was developed in which arginine 192 was replaced by glycine, thus, eliminating the trypsin cleavage site and attenuating the protein. The mechanisms by which LT(R192G) potentiates immune responses are unclear. Some suggestions include enhanced mucosal permeability [19], selective induction of Th2-mediated antibody responses [20], induction of antigen-specific T-cell responses [21,22], and increased antigen presentation [23]. 2.8. Mice Five inbred and three outbred mouse strains were used in this study. The inbred strains included BALB/c (H-2d ; Harlan–Sprague–Dawley, Indianapolis, IN), C57BL/6 (H-2b ; Jackson Laboratories, Bar Harbor, ME), 129 (H-2b ; Taconic, Germantown, NY), DBA/2 (H-2d ; Taconic), and C3H (H-2k ; Taconic). The outbred strains were CD-1 (Harlan–Sprague–Dawley), Swiss Webster (Harlan– Sprague–Dawley), and CF-1 (Charles River Laboratories, Wilmington, MA). All mice were pathogen-free and none showed evidence of a previous rotavirus infection as determined by the absence of serum rotavirus antibody. Experiments were conducted only with adult, female mice, 6–10 weeks of age at the time of their first immunization. The mice were housed in micro-isolation cages (four per cage) and all procedures were conducted in accordance with a protocol approved by Cincinnati Children’s Hospital Research Foundation Institutional Animal Care and Use Committee.

3313

along with 10 ␮g of LT(R192G). When performed i.n., immunization was done by gradual inoculation of the nares of the mice (maximum of 30 ␮l per nostril). All mice were administered a second dose 2 weeks after the first. Four weeks later, blood specimens were again collected for rotavirus IgG measurements and 1 day later, the mice were orally (gavage) challenged with 1000 shedding doses (SD50 ) of either wild type EDIM or McN rotavirus strains. Stools were collected for the 7 days after challenge (two pellets per mouse placed into 1.0 ml of EBSS and stored at −20 ◦ C until analyzed) to be examined for the quantity of rotavirus shed. Groups of control mice in these studies were either unimmunized (not inoculated) or were inoculated with adjuvant only. Because the adjuvant alone had no effect on protection, only the results found with the unimmunized mice are presented as controls unless otherwise noted. 2.10. Determination of rotavirus IgG titers in serum specimens Rotavirus IgG titers were determined by ELISA using the blood specimens collected before and after immunization by methods essentially as already described [10]. The only difference is that the concentration of rotavirus IgG was determined from a standard curve that measured this antibody in nanograms rather than units per milliliter. The limit of detection was 100 ng/ml. Although both serum and stool rotavirus IgA titers were evaluated before and after immunization in every experiment, little or none was found. Therefore, the IgA results are not included with the presented data. 2.11. Measurement of rotavirus antigen titers in stools Stool samples collected after rotavirus challenge were thawed, homogenized, and centrifuged (1500 × g, 5 min, 4 ◦ C). The quantities of rotavirus antigen in the fecal suspensions were then determined by ELISA and expressed as nanograms per milliliter of stool specimen using methods described previously [10]. Protection due to immunization was determined as the percentage reduction in the average quantity of rotavirus shed per mouse per day during the 7 days after challenge in each of the groups of immunized mice relative to that found in the group of unimmunized control mice for each experiment. 2.12. Statistical methods

2.9. Immunization and rotavirus challenge of mice Blood (retroorbital capillary plexus puncture) specimens were collected for rotavirus IgG measurements from all mice in a study and serum separated from these specimens was stored at −70 ◦ C until analyzed. Groups containing from four to eight mice (specific numbers are listed later for each of the experiments) were then immunized intranasally under light sedation or orally by gavage using 8.8 ␮g (unless otherwise specified) of either the EDIM or CJN VP6 chimera

The mean of the rotavirus antigen shed in stool samples collected from each mouse over the 7-day post-challenge period was used for the antigen computations. The log of the rotavirus IgG titers was used for the antibody computations. Multiple two sample t-tests were performed between mouse groups to determine whether statistically significant differences could be found in either the mean of the rotavirus shed or the log of the post-vaccination IgG titers. Differences between samples were considered statistically

3314

A.H. Choi et al. / Vaccine 20 (2002) 3310–3321

significant when the probability levels (P) were <0.05. SAS version 8.2 was used for the analyses.

3. Results 3.1. Equal protection is induced in mouse strains belonging to H-2b , H-2d , and H-2k haplotypes after i.n. immunization with EDIM VP6 and LT(R192G) We previously reported that two i.n. immunizations of BALB/c (H-2d ) mice with 8.8 ␮g doses (separated by a 2-week interval) of a chimera containing maltose-binding protein and murine rotavirus strain EDIM VP6 plus 10 ␮g of genetically-attenuated E. coli heat-labile toxin LT(R192G) induced reductions of approximately 99% in rotavirus shedding when mice were challenged with 1000 SD50 of unpassaged EDIM 4 weeks after the second immunization [14]. In a subsequent experiment, using the same methods, we found that when the time of EDIM challenge was delayed for 1 year, protection remained at 99% (results not shown). Based on the finding that the primary effectors of protection under these conditions appear to be CD4 cells [24], protection could be restricted to mice belonging only to certain haplotypes. Furthermore, protection could even be restricted to BALB/c mice since no other mouse strain belonging to the H-2d haplotype had been evaluated. The initial study, therefore, was to determine if protection against EDIM shedding could be stimulated in other inbred mouse strains belonging either to the H-2d haplotype or different haplotypes following two i.n. immunizations with EDIM VP6 plus LT(R192G). Four weeks after the second immunization, mice belonging to the BALB/c, DBA/2 (H-2d ), C57BL/6 (H-2b ), 129 (H-2b ), or C3H (H-2k ) strains were challenged with 1000 SD50 of wild type EDIM and protection against rotavirus shedding was evaluated during the subsequent 7 days. Using post-vaccination serum rotavirus IgG titers as indicators of immune responsiveness in the different mouse strains, i.n. inoculation with EDIM VP6 and LT(R192G) was highly immunogenic in all five inbred strains examined (Table 1). Table 1 Serum rotavirus IgG responses in inbred mouse strains following i.n. immunization with 8.8 ␮g doses of EDIM VP6 and 10 ␮g of LT(R192G) Mouse strain

Haplotype

Na

IgG titersb (ng/ml)

BALB/c DBA/2 C57BL/6 129 C3H

H-2d H-2d H-2b H-2b H-2k

8 5 8 6 6

31374 436541 60416 139838 42810

± ± ± ± ±

35535 212313 49991 50310 128423

a Groups of six unimmunized or adjuvant-immunized mice were included as controls for each group of animals. b Geometric mean titers of serum rotavirus IgG ± standard deviation. Note that the rotavirus IgG titers for every mouse in the control groups were less than the limit of detection of 100 ng/ml.

Although the quantities of rotavirus shed during the 7 days after challenge in the unimmunized control mice varied greatly from one strain to the next, immunization consistently reduced shedding by 99% as was found in BALB/c mice under the same conditions (Fig. 1). No protection was found in mice immunized with MBP or LT(R192G) only (results not shown). Therefore, VP6-induced protection was not restricted by either strain or haplotype based on this study with inbred mice. 3.2. BALB/c mice are protected against EDIM shedding after i.n. immunization with a chimeric protein containing VP6 of a human rotavirus The VP6 protein of group A rotavirus is highly conserved which has permitted it to be classified as the group antigen. The greatest difference in amino acid sequence identity between the VP6 proteins of any two strains of mammalian group A rotavirus that have been examined is approximately 13% [25]. This suggests that immunization with the VP6 protein of any group A rotavirus may stimulate protection against most, and possibly all, mammalian group A rotaviruses. To begin to evaluate this possibility, the VP6 gene of the human G1 rotavirus strain CJN, obtained from the stool of a child hospitalized with severe gastroenteritis in Cincinnati in 1982, was used to immunize mice prior to EDIM challenge. Based on its coding sequence, the CJN VP6 protein differs from that of EDIM VP6 by 9%, i.e. 35 amino acids (Fig. 2), which represents nearly the greatest difference between any two group A rotavirus VP6 proteins that has been reported. The VP6 gene of CJN was first cloned into the bacterial plasmid pMAL/c2X, then expressed in E. coli as a chimeric protein containing MBP at its amino terminus. Following purification by affinity chromatography with amylose resin, 8.8 ␮g of CJN VP6 together with 10 ␮g of LT(R192G) were used for i.n. immunization of BALB/c mice (two doses separated by a 2-week interval) and the results were compared with those obtained after immunization with EDIM VP6 under the same conditions. When the immune responses stimulated by these immunogens were compared just prior to challenge with wild type EDIM 4 weeks after the second immunization, inoculation with either EDIM or CJN VP6 was found to elicit large but not significantly different serum rotavirus IgG responses (Table 2). After EDIM challenge (1000 SD50 ), rotavirus shedding was found to be reduced 98.6 or 86.0% relative to the unimmunized control mice during the week following challenge in the EDIM or CJN VP6 immunized mice, respectively. Although the amount of protection induced by CJN VP6 was significantly (P = 0.02) less than induced with EDIM VP6, it was still highly significant (P < 0.001) relative to that found in the unimmunized control group. This suggests that cross-protection against rotaviruses belonging to multiple serotypes may be feasible using a single VP6 protein as the immunogen.

A.H. Choi et al. / Vaccine 20 (2002) 3310–3321

3315

Fig. 1. (A–E) Effect of i.n. immunization with EDIM VP6 on rotavirus shedding with different strains of inbred mice belonging to different haplotypes. Different strains of inbred mice (the numbers per group and the haplotypes are designated in Table 1) were immunized i.n. with two doses of 8.8 ␮g of EDIM VP6 and 10 ␮g of LT(R192G) separated by a 2-week interval. At 4 weeks after the second dose, the mice were challenged with 1000 SD50 of wild type EDIM and mean shedding of rotavirus antigen (ng per mouse per day) was measured during the subsequent 7 days. The percentage reduction in shedding in the immunized mice (䊉) during these 7 days was determined relative to that found in unimmunized control mice (䊏). Standard errors of the mean are shown for shedding on each day for the mice within a group.

Table 2 Serum rotavirus IgG responses and protection against EDIM shedding after i.n. immunization of BALB/c mice with EDIM or CJN VP6 and LT(R192G) Immunogen

Dose (␮g)

N

IgG titersa (ng/ml)

Quantity of antigen shedb

%Protection from shedding

None EDIM VP6 CJN VP6 CJN VP6

– 8.8 8.8 0.088

14 8 6 5

<100 31374 ± 35535 117789 ± 89867 8113e ± 17666

363 ± 180 5±4 51d ± 35 17 ± 17

0c 98.7 86.0 95.4

Geometric mean titers of serum rotavirus IgG ± standard deviation. Average quantity of rotavirus antigen shed ± standard deviation during the 7 days after challenge (ng per mouse per day). c Shedding in the unimmunized control group was used to determine 0% protection in this study. d Significantly (P = 0.02) more shedding than in mice immunized with EDIM VP6. e Significantly (P = 0.006) less than mice immunized with 8.8 ␮g of CJN VP6. a

b

3316

A.H. Choi et al. / Vaccine 20 (2002) 3310–3321

Fig. 2. The predicted amino acid sequences of the VP6 proteins of the EDIM strain of murine rotavirus and the CJN strain of human rotavirus. All 35 amino acid changes in the two proteins are highly conservative. The Genbank accession numbers for the nucleotide sequences of the EDIM and CJN VP6 protein genes are U65988 and AF461757, respectively.

It should be noted that the immune responses, as measured by serum rotavirus IgG titers, elicited by i.n. inoculation with 8.8 ␮g of CJN VP6 were substantially (P = 0.006) diminished when the dose administered was reduced to 0.088 ␮g (Table 2), yet protection was not reduced. Therefore, no greater protection would be expected if the i.n. doses were increased beyond the 8.8 ␮g level. Likewise, in a separate study, i.n. immunization of BALB/c mice with only a single dose of 1.8 ␮g of EDIM VP6 stimulated the same level of protection as two doses of 8.8 ␮g (results not shown). Therefore, it is anticipated that two doses of 8.8 ␮g of either VP6 chimera will provide maximal protection if delivered i.n. with 10 ␮g of LT(R192G). This immunization protocol was used for the delivery of the two chimeras in the subsequent studies unless otherwise noted. 3.3. Intranasal immunization with chimeric VP6 proteins protects outbred mice against EDIM shedding In order for the VP6 vaccine to be useful in humans, it must protect individuals with diverse genetic backgrounds. The results already presented demonstrated that the EDIM VP6 chimera protected several inbred mouse strains against EDIM shedding when delivered i.n. with LT(R192G). To determine whether this protection could be extended to mice with highly divergent genetic backgrounds and be effective when the VP6 used as the immunogen is different from that contained in the challenge virus, outbred mice strains were immunized i.n. with either CJN or EDIM VP6 prior to EDIM challenge. Two i.n. doses of these immunogens were administered to groups of mice belonging to three out-

bred strains. Four weeks after the second immunization, the mice were challenged with 1000 SD50 of wild type EDIM and shedding of rotavirus was monitored for the subsequent 7 days. Immune responses to the vaccines, as determined by serum rotavirus IgG titers just prior to challenge, varied between mouse strains but there were no significant variations in these responses between mice immunized with the CJN versus EDIM VP6 proteins in the same strain of mice (Table 3). Likewise, all mice belonging to each of the three outbred strains examined were significantly protected against EDIM shedding after immunization with either the EDIM or CJN VP6 chimera (Fig. 3). The least protection measured in any one mouse (i.e. one of the eight CF-1 mice immunized with CJN VP6) was 69.7% when determined relative to the average shedding of the eight mice in the unimmunized control groups. Although overall protection in the CF-1 mice was significantly (P = 0.03) better after immunization with EDIM VP6 compared with mice administered CJN VP6, no differences in protection were found in the other two outbred stains between groups given the different immunogens. Therefore, all outbred mice examined were significantly protected against rotavirus shedding after EDIM challenge whether immunized with EDIM or CJN VP6, and in most mice protection was >95%. 3.4. Oral immunization with VP6 chimeras and LT(R192G) can be as effective as i.n. immunization It is unclear which routes of immunization will be most desirable in the future, but at this time, oral delivery appears

A.H. Choi et al. / Vaccine 20 (2002) 3310–3321

3317

Fig. 3. (A–C) Protection against rotavirus (EDIM) shedding in outbred mice immunized i.n. with two 8.8 ␮g doses of EDIM or CJN VP6 and 10 ␮g of LT(R192G). The quantities of rotavirus antigen shed during the 7 days after EDIM challenge (4 weeks after the second immunization) are shown for mice immunized with either EDIM (䊉) or CJN (䉱) VP6 along with the percentages reductions relative to the unimmunized control mice (䊏). Asterisk denotes significantly (P = 0.03) less shedding was found in mice immunized with EDIM VP6 than CJN VP6 in the CF-1 strain. Standard errors of the mean are shown for shedding on each day for the mice within a group. The numbers of mice in each group are noted in Table 3.

3318

A.H. Choi et al. / Vaccine 20 (2002) 3310–3321

Table 3 Serum rotavirus IgG responses developed after i.n. immunization of outbred mice with EDIM or CJN VP6 and LT(R192G) Mouse strain

VP6 immunogen

Na

IgG titersb (ng/ml)

CF-1 CF-1 CD-1 CD-1 Swiss Webster Swiss Webster

EDIM CJN EDIM CJN EDIM CJN

8 8 6 8 6 8

20271 ± 57224 53218c ± 52425 131933 ± 82814 247466 ± 475879 454889d ± 378441 331364 ± 278076

a Groups of eight unimmunized or adjuvant-immunized mice were included as controls in each experiment. b Geometric mean titers of rotavirus IgG ± standard deviation. Post-immunization titers of serum rotavirus IgG in all the control mice were below the limit of detection of 100 units/ml. c Significantly (P < 0.03) less than the other two mouse strains given CJN VP6. d Significantly (P < 0.02) more than the other two mouse strains given EDIM VP6.

to be preferred over intranasal delivery. Therefore, we examined the effectiveness of oral immunization with VP6 chimeras in prevention of rotavirus shedding in mice. Two oral doses of 8.8 ␮g of either EDIM or CJN VP6 along with 10 ␮g of LT(R192G) were administered to groups of BALB/c mice 4 weeks prior to EDIM challenge. Additionally, a group of mice was orally administered three times the normal dose of EDIM VP6 (i.e. 26.4 ␮g) along with 10 ␮g of LT(R192G) during the two immunizations to determine if delivery of higher concentrations of the immunogen would increase its effectiveness. Oral immunization with 8.8 ␮g of either expressed VP6 protein stimulated very poor immune responses as determined by titers of serum rotavirus IgG (Table 4). Although immunization with the larger amount of EDIM VP6 (i.e. 26.4 ␮g/dose) elicited a higher GMT of serum rotavirus IgG than the lower quantity of this immunogen, the difference was not significant. The protection stimulated by the larger EDIM VP6 dose was also greater (i.e. 99.0% versus 93.2%), but the difference remained insignificant. Oral immunization with 8.8 ␮g doses of CJN VP6 elicited protection (i.e. 85.5%) that was also not significantly different from that found after immunization with the same concentration of EDIM VP6 (Table 4). Likewise, the protection levels elicited after immunization by the oral versus i.n. route with either the CJN or EDIM VP6 was not significantly different (see

Tables 2 and 4). These results indicate that even though the overall immune responses, as determined by titers of serum rotavirus IgG, are much less after oral than after i.n. immunization with 8.8 ␮g doses of the VP6 immunogens, the protection stimulated by the two routes was not significantly different. 3.5. Protection against EDIM shedding observed after either oral or i.n. immunization with EDIM or CJN VP6 and LT(R192G) is also observed when mice are challenged with a different strain of murine rotavirus All experiments up to this time had utilized EDIM as the challenge virus following immunization. To determine whether these results have broader applicability, it was necessary to show that comparable results could be obtained with another challenge virus. Our laboratory recently obtained a second murine rotavirus, named McN, and has developed it for use in these studies. The electrophoretic mobilities of 8 out of the 11 double-stranded RNA genome segments of McN differed from those of EDIM (McNeal et al., in preparation). Likewise, many of the phenotypic properties of the two virus strains were distinctly different. Even so, the predicted sequence of the 327 amino acids contained in the VP7 proteins of the two viruses differed in only 16 positions (5.2%), thus, allowing both to be classified as the same G type, i.e. G3. One of the phenotypic characteristics of the McN strain is its ability to consistently produce large quantities of rotavirus antigen in every mouse strain examined, including C57BL/6 and genetically-altered strains on this background. This is not true of the EDIM strain. Therefore, C57BL/6 mice were included along with BALB/c mice in evaluating protection against McN shedding after VP6 immunization. Two doses of either EDIM or CJN VP6 administered orally or i.n. with LT(R192G) elicited excellent protection against rotavirus shedding when either mouse strain was challenged with 1000 SD50 of wild type McN 4 weeks after the second immunization (Table 5). Significantly greater protection was stimulated by i.n. immunization with EDIM VP6 in both mouse strains (P < 0.001 for BALB/c and P = 0.05 for C57BL/6 mice, respectively). However, EDIM VP6 did not induce significantly greater protection than CJN VP6 after oral immunization of C57BL/6 mice, the only mouse strain in which this comparison was made. These results are

Table 4 Serum rotavirus IgG responses and protection against EDIM shedding after oral immunization of BALB/c mice with EDIM or CJN VP6 and LT(R192G) Immunogen

Dose (␮g)

N

IgG titersa (ng/ml)

Quantity of antigen shedb

None EDIM VP6 EDIM VP6 CJN VP6

– 8.8 26.4 8.8

6 6 6 6

<100 209 ± 3038 1530 ± 12525 217 ± 2915

621 42 7 90

± ± ± ±

586 68 4 143

%Protection from shedding 0c 93.2 99.0 85.5

Geometric mean titers of rotavirus IgG ± standard deviation. Average quantity of rotavirus antigen shed ± standard deviation during the 7 days after challenge (ng per mouse per day). c Shedding in the unimmunized control group was used to determine 0% protection. a

b

A.H. Choi et al. / Vaccine 20 (2002) 3310–3321

3319

Table 5 Protection against rotavirus shedding stimulated by i.n. or oral immunization of mice with EDIM or CJN VP6 and LT(R192G) after challenge with the McN strain of rotavirus Mouse strain

N

Immunogen

Immunization route

Quantity of antigen sheda

%Protection from sheddingb

BALB/c BALB/c BALB/c BALB/c BALB/c C57BL/6 C57BL/6 C57BL/6 C57BL/6 C57BL/6 C57BL/6

4 6 5 6 6 5 5 6 6 6 6

LT(R192G) EDIM VP6 CJN VP6 LT(R192G) EDIM VP6 LT(R192G) EDIM VP6 CJN VP6 LT(R192G) EDIM VP6 CJN VP6

Intranasal Intranasal Intranasal Oral Oral Intranasal Intranasal Intranasal Oral Oral Oral

9177 ± 2844 129 ± 79 673c ± 122 9841 ± 2698 149 ± 104 9118 ± 1379 160 ± 102 852d ± 685 14313 ± 4369 2429 ± 3869 1134 ± 724

0 98.6 92.7 0 98.5 0 98.2 90.4 0 83.0 92.1

Average quantity of rotavirus antigen shed ± standard deviation during the 7 days after challenge (ng per mouse per day). protection is determined relative to the unimmunized control groups. c Significantly (P < 0.001) more shedding in mice immunized i.n. with CJN versus EDIM VP6. d Significantly (P = 0.05) more shedding in mice immunized i.n. with CJN versus EDIM VP6. a

b Percentage

essentially identical to those found after EDIM challenge of VP6 immunized BALB/c mice (see Tables 2 and 4). Also, there were no significant differences in the protection against McN shedding elicited by i.n. versus oral immunization with either immunogen (Table 5), a result already noted for mice challenged with the EDIM strain. Therefore, excellent protection was stimulated after either i.n. or oral immunization with EDIM or CJN VP6 against a second strain of murine rotavirus that was both genetically and phenotypically distinct from EDIM.

4. Discussion Several live, orally deliverable rotavirus vaccine candidates have been evaluated in humans, but all have inherent properties that make them potentially undesirable for administration to infants. Non-living, second-generation rotavirus vaccine candidates have, therefore, been developed and their abilities to elicit active immunity against rotavirus in animal models are being evaluated. Although protection in these models, particularly in the mouse model, has been generated by parenteral immunization with non-living rotavirus vaccines [4–7], mucosal immunization may be more effective against this enteric pathogen. Intranasal immunization with rotavirus 2/6 VLPs or inactivated double-layered particles, each of which lacks the outer capsid VP4 and VP7 neutralization proteins, has been found to be particularly effective against rotavirus shedding in mice when the immunogens were administered with a powerful adjuvant such as attenuated E. coli heat-labile toxin LT(R192G) [8–10]. Because the immunodominant protein in these particles is VP6, this protein was expressed in E. coli and evaluated for its ability to elicit protection after intranasal immunization of mice [14]. When the VP6 protein from the murine rotavirus strain EDIM was administered i.n. to adult BALB/c mice with LT(R192G), it consistently stimulated

nearly complete protection against EDIM shedding. Before evaluating this protein in another, more complex animal model, it was imperative to determine the breadth of its effectiveness in the mouse model. This study was designed to make this determination. We have found that CD4 T cells generated in response to i.n. inoculation with EDIM VP6 and LT(R192G) are the only lymphocytes needed for protection [24]. This suggested that the protective epitopes on VP6 might be MHC-II restricted and possibly, therefore, not all inbred mouse strains would be protected. However, we also observed that peptides derived from multiple regions of the EDIM VP6 protein induced at least partial protection of BALB/c mice against EDIM shedding [26]. This suggested either that multiple CD4 epitopes within VP6 can elicit protective responses or that effectors other than CD4 cells are involved in protection. In either case, VP6-induced protection would not be expected to be restricted to a single mouse haplotype. When inbred mice belonging to either another strain of the same haplotype as BALB/c (i.e. H-2d ) or other haplotypes (i.e. H-2b or H-2k ) were immunized i.n. with EDIM VP6 and LT(R192G), protection against EDIM shedding was identical to that found in BALB/c mice. Thus, VP6-induced protection was not restricted to BALB/c or even H-2d mice. It is of interest to note that the quantity of EDIM antigen produced after infection of unimmunized mice in this study varied greatly between mouse strains, an observation already reported when comparing mice on BALB/c versus C57BL/6 backgrounds [27]. Even so, the level of protection induced after EDIM VP6 immunization was essentially identical. Because the quantity of EDIM proteins produced in the immunized mice after challenge parallels that produced in unimmunized mice of the same inbred strain, it follows that this level of protection (99%) is the maximum attainable using this immunization procedure. One potential advantage of the VP6 protein as an immunogen is that it is highly conserved within all mammalian

3320

A.H. Choi et al. / Vaccine 20 (2002) 3310–3321

group A rotavirus strains that have been examined. The largest reported difference in amino acid sequence between any two strains is 13% even though much greater differences are found in nucleotide sequences of VP6 genes [25]. This suggests that protection elicited by VP6 should not be restricted by the serotype of its outer capsid VP4 and VP7 neutralization proteins [2]. To test this hypothesis, VP6 from the human rotavirus G1 strain CJN was expressed in E. coli and used to protect against EDIM shedding. The CJN VP6 protein has 35 out of its 397 amino acids (9%) different from that of EDIM VP6, nearly the greatest difference found between any two mammalian rotavirus strains. Even so, both EDIM and CJN were found to belong to subgroup II. Others have reported that the VP6 protein can be classified into four subgroups (i.e. I, II, I/II, and non-I/II) based on recognition by subgroup-specific monoclonal antibodies [28–30], but only a single amino change is required to cause a shift in subgroup [30]. Although protection of BALB/c mice against EDIM shedding by i.n. immunization with CJN VP6 and LT(R192G) was highly significant, this protection was less (P = 0.02) than stimulated by EDIM VP6. Thus, excellent protection can be stimulated by immunization with heterologous VP6 proteins, but the best protection may be species-specific. To determine whether the range of protection induced by EDIM and CJN VP6 could be extended beyond that of inbred strains, three outbred strains were immunized i.n. and challenged with EDIM. Shedding by every immunized mouse, relative to the non-immunized control mice, was reduced a minimum of 69.7%. Again, CJN VP6 was somewhat less effective than EDIM VP6 and in one mouse strain (i.e. CF-1) the difference was significant (P = 0.03). The efficiencies with which EDIM and CJN VP6 elicit immunity after i.n. versus oral inoculation was next examined. Based on titers of serum rotavirus IgG, oral immunization produced very poor immune responses. Furthermore, no serum or stool rotavirus IgA was detectable after oral immunization, yet protection elicited by oral immunization with VP6 was not significantly different from that induced by the i.n. route. We have previously reported that protection elicited after intranasal immunization with VP6 and LT(R192G) occurs in the absence of antibody [14] but was dependent on the presence of CD4 T cells [24]. This contrasts with results reported after oral immunization of animals with live rotaviruses where dependence on antibody has been consistently indicated [31–38], thus providing strong evidence for different effectors in protection stimulated by mucosal immunization with live virus vaccines versus non-living rotavirus vaccines together with powerful adjuvants. Human rotavirus illnesses are caused by a large number of rotavirus strains but all rotavirus studies we have conducted with adult mice have utilized the same murine strain as the challenge virus. To determine the potential of the VP6 protein as a vaccine candidate, we compared these results with those obtained using a second murine rotavirus recently

isolated in our research facility. This strain (i.e. McN) was distinct from EDIM based on the electropherotype, subtype, amino acid sequences of VP4, VP7 and VP6 proteins, and several phenotypic traits, amongst which was the level of rotavirus shedding following challenge of mice on C57BL/6 backgrounds (McNeal et al., in preparation). The two strains were even found to be serotypically distinct even though both were classified G3. In spite of these differences, protection against McN shedding induced in BALB/c and C57BL/6 mice after oral or i.n. immunization with either EDIM or CJN VP6 was essentially identical to that observed after EDIM challenge. Although the results reported here demonstrate the effectiveness of mucosal immunization with VP6 and LT(R192G) in suppressing murine rotavirus shedding in mice, they do not provide direct information on the potential usefulness of this vaccine candidate in humans. Other investigators have reported that 2/6 VLPs administered mucosally with cholera toxin, LT, or LT(R192G) to mice also effectively suppresses murine rotavirus shedding after challenge [8,9]. However, this outcome was not observed in gnotobiotic piglets using a similar protocol. Protection against human rotavirus shedding or diarrhea induced by i.n. immunization with 2/6 VLPs and LT(R192G) was observed only if the piglets were first primed by immunization with an attenuated human rotavirus [11–13]. An important difference between the mouse and piglet studies is that even though the quantity of antigen administered was increased from 10 to 250 ␮g to compensate for the differences in the sizes of the animals, 10 ␮g per dose of LT(R192G) was given to mice while only 5 ␮g was administered to the piglets. Also, the piglets were gnotobiotic and it is well established that mucosal immune responses are suppressed in germ-free animals [39,40]. Even so, significant quantities of rotavirus-specific memory B cell responses were stimulated in the piglets and the quantity of each B cell isotype was greater when the adjuvant was included during immunization. Whether VP6 immunization with adjuvant will prevent rotavirus disease in piglets, in another animal model, or in vaccinated humans remains to be determined.

Acknowledgements This work was supported in part by NIH-NIAID contract NO1 AI 45252 to Children’s Hospital Medical Center and a grant (NIH-NIAID 1 R43 AI50326-01) to A.H. Choi. References [1] Parashar UD, Hummelman EG, Bresee JS, Miller MA, Glass RG. In: Proceedings of the Conference on Vaccines for Enteric Diseases (Abstract), Tampere, Finland, 2001. [2] Bernstein DI, Ward RL. Rotaviruses. In: Feigin RD, Cherry JD, editors. Textbook of pediatric infectious disease IV. Philadelphia: Saunders, 1998. p. 1901–21.

A.H. Choi et al. / Vaccine 20 (2002) 3310–3321 [3] Murphy TV, Gargiullo PM, Massoudi MS, et al. Intussusception among infants given an oral rotavirus vaccine. N Engl J Med 2001;344:564–72. [4] Conner ME, Crawford SE, Barone C, et al. Rotavirus vaccine administered parenterally induces protective immunity. J Virol 1993;67:6633–41. [5] Ciarlet M, Crawford SE, Barone C, et al. Subunit rotavirus vaccine administered parenterally to rabbits induces active protective immunity. J Virol 1998;72:9233–46. [6] McNeal MM, Sheridan JF, Ward RL. Active protection against rotavirus infection of mice following intraperitoneal immunization. Virology 1992;191:150–7. [7] McNeal MM, Rae MN, Conner ME, et al. Stimulation of local immunity and protection in mice by intramuscular immunization with triple-or double-layered rotavirus particles and QS-21. Virology 1998;243:155–66. [8] O’Neal CM, Crawford SE, Estes MK, et al. Rotavirus virus-like particles administered mucosally induce protective immunity. J Virol 1997;71:8707–17. [9] O’Neal CM, Clements JD, Estes MK, et al. Rotavirus 2/6 viruslike particles administered intranasally with cholera toxin, Escherichia coli heat-labile toxin (LT), and LT-R192G induce protection from rotavirus challenge. J Virol 1998;72:3390–3. [10] McNeal MM, Rae MN, Bean JA, et al. Antibody-dependent and-independent protection following intranasal immunization of mice with rotavirus particles. J Virol 1999;73:7565–73. [11] Yuan L, Geyer A, Hodgins DC, et al. Intranasal administration of 2/6-rotavirus-like particles with mutant Escherichia coli heat-labile toxin (LT-R192G) induces antibody-secreting cell responses but not protective immunity in gnotobiotic pigs. J Virol 2000;74:8843–53. [12] Yuan L, Iosef C, Azevedo MP, et al. Protective immunity and antibody-secreting cell responses elicited by combined oral attenuated Wa human rotavirus and intranasal Wa 2/6-VLPs with mutant Escherichia coli heat-labile toxin in gnotbiotic pigs. J Virol 2001;75:9229–38. [13] Iosef C, Nguyen TV, Jeong K-O, et al. Systemic and intestinal antibody secreting cell responses and protection in gnotobiotic pigs immunized orally with attenuated Wa human rotavirus and Wa 2/6-rotavirus-like-particles associated with immunostimulating complexes. Vaccine 2002;20:1741–53. [14] Choi AH, Basu M, McNeal MM, et al. Antibody-independent protection against rotavirus infection of mice stimulated by intranasal immunization with chimeric VP4 or VP6 protein. J Virol 1999;73:7574–81. [15] Ward RL, McNeal MM, Sheridan JF. Development of an adult mouse model for studies on protection against rotavirus. J Virol 1990;64:5070–5. [16] Ward RL, Bernstein DI, Young EC, Sherwood JR, Knowlton DR, Schiff GM. Human rotavirus studies in volunteers: determination of infectious dose and serological response to infection. J Infect Dis 1986;154:871–9. [17] Both GW, Siegman LJ, Bellamy AR, et al. Comparative sequence analysis of rotavirus genomic seqment 6—the gene specifying viral subgroups 1 and 2. J Virol 1984;51:97–101. [18] Jarrett HW, Foster JL. Alternate binding of actin and calmodulin to multiple sites on dystrophin. J Biol Chem 1995;270:5578–86. [19] Lyke N, Karlsson U, Sjolander A, et al. The adjuvant effect of cholera toxin is associated with an increased permeability for luminal antigens. Scand J Immunol 1991;33:691–8. [20] Snider DP, Marshall JS, Perdue MH, et al. Production of IgE antibody and allergic sensitization of intestinal and peripheral tissues after oral immunization with protein Ag and cholera toxin. J Immunol 1994;53:647–57.

3321

[21] Hornquist E, Lycke N. Cholera toxin adjuvant greatly promotes antigen priming of T cells. Eur J Immunol 1993;23:2136–43. [22] Xu-Amano J, Jackson RJ, Fujihashi K, et al. Helper Th1 and Th2 cell responses following mucosal or systemic immunization with cholera toxin. Vaccine 1994;12:903–11. [23] Brumander AK, Kjerrulf M, Holmgren J, et al. Cholera toxin enhances alloantigen presentation by cultured intestinal epithelial cells. Scand J Immunol 1993;37:452–8. [24] McNeal MM, VanCott JL, Choi AH, et al. CD4 T cells are the only lymphocytes needed to protect mice against rotavirus shedding after intranasal immunization with a chimeric VP6 protein and the adjuvant LT(R192G). J Virol 2002;76:560–8. [25] Tang B, Gilbert JM, Matsui SM, et al. Comparison of the rotavirus gene 6 from different species by sequence analysis and localization of subgroup-specific epitopes using site-directed mutagenesis. Virology 1997;237:89–96. [26] Choi AH, Basu M MM, et al. Functional mapping of protective domains and epitopes in the rotavirus VP6 protein. J Virol 2000;74:11574–80. [27] Choi AH, McNeal MM, Basu M, et al. Immunity to homologous rotavirus infection in adult mice. Trends Microbiol 2000;8:52. [28] Gorziglia M, Hoshino Y, Nishikawa K, et al. Comparative sequence analysis of the genomic segment 6 of four rotaviruses each with a different subgroup specificity. J Gen Virol 1988;69:1659–69. [29] Lopez S, Espinosa R, Greenberg HB, et al. Mapping the subgroup epitopes of rotavirus protein VP6. Virology 1994;204:153–62. [30] Tang B, Gilbert JM, Matsui SM, et al. Comparison of the rotavirus gene 6 from different species by sequence analysis and localization of subgroup-specific epitopes using site-directed mutagenesis. Virology 1997;237:89–96. [31] Franco MA, Greenberg HB. Role of B cells and cytotoxic T lymphocytes in clearance of and immunity to rotavirus infection in mice. J Virol 1995;69:7800–6. [32] McNeal MM, Barone KS, Rae MN, et al. Effector functions of antibody and CD8+ cells in resolution of rotavirus infection and protection against reinfection in mice. Virology 1995;214:387–97. [33] Yuan L, Ward LA, Rosen BI, et al. Systemic and intestinal antibody-secreting cell responses and correlates of protective immunity to human rotavirus in a gnotobiotic pig model of disease. J Virol 1996;70:3075–83. [34] Moser CA, Cookinham S, Coffin SA, et al. Relative importance of rotavirus-specific effector and memory B cells in protection against challenge. J Virol 1998;72:1108–14. [35] O’Neal CM, Harriman GR, Conner ME. Protection of the villus epithelial cells of the small intestine from rotavirus infection does not require immunoglobulin A. J Virol 2000;74:4102–9. [36] Yuan L, Geyer A, Saif LJ. Short-term immunoglobulin A B-cell memory resides in intestinal lymphoid tissues but not in bone marrow of gnotobiotic pigs inoculated with Wa human rotavirus. Immunology 2001;103:188–98. [37] Chang KO, Vandal OH, Yuan L, et al. Antibody-secreting cell responses to rotavirus proteins ingnotobiotic pigs inoculated with attenuated or virulent human rotavirus. J Clin Microbiol 2001;39:2807–13. [38] Moser CA, Offit PA. Distribution of rotavirus-specific memory B cells in gut-associated lymphoid tissue after primary immunization. J Gen Virol 2001;82:2271–4. [39] Umesaki Y, Setoyama H. Structure of the intestinal flora responsible for development of the gut immune system in a rodent model. Microbes Infect 2000;2:1343–51. [40] Savage DC. Mucosal microbiota. In: Ogra PL, Mestecky J, Lamm ME, Strober W, Bienenstock J, McGhee JR, editors. Mucosal immunolgy. 2nd ed. San Diego: Academic Press, 1999. p. 19–30.