Gut microbiota utilize immunoglobulin A for mucosal colonization.

The immune system responds vigorously to microbial infection while permitting lifelong colonization by the microbiome. Mechanisms that facilitate the establishment and stability of the gut microbiota remain poorly described. We found that a regulatory system in the prominent human commensal Bacteroides fragilis modulates its surface architecture to invite binding of immunoglobulin A (IgA) in mice. Specific immune recognition facilitated bacterial adherence to cultured intestinal epithelial cells and intimate association with the gut mucosal surface in vivo. The IgA response was required for B. fragilis (and other commensal species) to occupy a defined mucosal niche that mediates stable colonization of the gut through exclusion of exogenous competitors. Therefore, in addition to its role in pathogen clearance, we propose that IgA responses can be co-opted by the microbiome to engender robust host-microbial symbiosis.

Rotavirus vaccines and vaccination potential.

The development of a successful rotavirus vaccine is a complex problem. Our review of rotavirus vaccine development shows that many challenges remain, and priorities for future studies need to be established. For example, the evaluation of administration of a vaccine with OPV or breast milk might receive less emphasis until a vaccine is made that shows clear efficacy against all virus serotypes. Samples remaining from previous trials should be analyzed to determine epitope-specific serum and coproantibody responses to clarify why only some trials were successful. Detailed evaluation of the antigenic properties of the viruses circulating and causing illness in vaccinated children also should be performed for comparisons with the vaccine strains. In future trials, sample collection should include monitoring for asymptomatic infections and cellular immune responses should be analyzed. The diversity of rotavirus serotype distribution must be monitored before, during, and after a trial in the study population and placebo recipients must be matched carefully to vaccine recipients. Epidemiologic and molecular studies should be expanded to document, or disprove, the possibility of animal to human rotavirus transmission, because, if this occurs, vaccine protection may be more difficult in those areas of the world where cohabitation with animals occurs. We also need to have an accurate assessment of the rate of protection that follows natural infections. Is it realistic to try to achieve 90% protective efficacy with a vaccine if natural infections with these enteric pathogens only provide 60% or 70% protection? Subunit vaccines should be considered to be part of vaccine strategies, especially if maternal antibody interferes with the take of live vaccines. The constraints on development of new vaccines are not likely to come from molecular biology. The challenge remains whether the biology and immunology of rotavirus infections can be understood and exploited to permit effective vaccination. Recent advances in developing small animal models for evaluation of vaccine efficacy should facilitate future vaccine development and understanding of the protective immune response(s) (Ward et al. 1990b; Conner et al. 1993).

Recombinant Norwalk virus-like particles as an oral vaccine.

Viruses that infect cells in the gastrointestinal tract are well suited for examining the immune response to oral delivery of antigen and for exploring the advantages and pitfalls of oral vaccines. Norwalk virus (NV) (family Caliciviridae, genus Calicivirus) causes acute gastroenteritis in all age groups. The NV capsid is composed of 180 copies of a single 58000 molecular weight protein which spontaneously forms virus-like particles (VLPs) that can be purified in extremely high yields (22 mg per 300 ml culture) when produced using the baculovirus expression system. We are testing the potential of these recombinant NV (rNV) particles for use as an oral vaccine by administering them to mice and volunteers. Mice were orally inoculated four times with rNV particles in concentrations ranging from 5 to 500 micrograms in the absence of adjuvant or from 5 to 200 micrograms with 10 micrograms of cholera toxin. Serum IgG and fecal IgA immune responses were monitored. rNV particles were found to be immunogenic when orally given to mice with or without adjuvant. These particles also were safe and immunogenic when orally given to volunteers. These studies show that rNV particles are an excellent model to test the oral delivery of mucosal immunogens in general, and that rNV particles are ideal candidates for vaccine development in particular.

Rotavirus subunit vaccines.

We evaluated rotavirus subunit vaccines for use in humans and animals. Insect cells were co-infected with combinations of individual baculovirus recombinants expressing human, bovine or simian rotavirus VP2, VP4, VP6 or VP7 to produce virus-like particles (VLPs). To determine whether immunization with VLPs could induce active protective immunity, VLPs were administered parenterally to rabbits, and the immune response and protection from rabbit ALA rotavirus challenge were evaluated. Complete or partial protection was attained, showing that parenteral immunization with VLPs induces active protective immunity. We also examined whether heterotypic immune responses could be induced with a limited number of broadly reactive VP7 proteins or with chimeric particles (multiple VP7 types on individual particles). The feasibility of this approach was determined by immunizing mice with VLPs containing a G3 VP7 or G1 VP7 and chimeric G1/G3 VLPs. Broadly reactive neutralizing antibody was induced by the G1 VLPs. VLPs also have been successfully used to boost lactogenic (colostral and milk) immunity in dairy cows. Taken together, these results show that VLPs can be effective immunogens in rabbits, mice and dairy cattle when administered parenterally, a limited number of VLPs may be sufficient to produce a broadly protective vaccine, and G3 VLPs may serve as an effective subunit vaccine for use in bovines.

Evaluation of rotavirus vaccines in small animal models.

The high morbidity and mortality of rotavirus (RV) infections has spurred the development of RV vaccines (1-13). Although children naturally infected with RV commonly undergo multiple infections, primary infections in children generally induce disease, and children are normally protected against severe disease during subsequent infections (1,2,6-8,14). For RV, the immunologic mechanisms responsible for protection are poorly understood, but antibody (Ab) in the intestine appears to be the primary mechanism of protection (2,15,16). Because RV is a localized enteric infection, and induction of intestinal mucosal immune responses was expected to be required for protection, live orally administered vaccines were pursued first. Vaccine development of the live attenuated vaccines proceeded to clinical trials in humans without prior animal testing. In August 1998, Rotashield™, a three dose, live attenuated tetravalent (TV), rhesus rotavirus (RRV) vaccine produced by Wyeth Lederle Vaccines and Pediatrics (West Henrietta, NY), was licensed. This vaccine shows promise, eliciting ∼80% protection against severe disease (6,7,12,17-19). The recent detection or emergence of new RV serotypes in humans suggests that incorporation of additional P-and G-serotypes into this vaccine may be necessary in the future (20-22). Additional concerns with the use of live attenuated vaccines include interference of vaccine replication by other enteric pathogens (common in children from the underdeveloped world); neutralization by maternal Ab; limited replication competence of animal strains, because of the host range restriction observed with RVs; and safety, because of the possibility of producing new virulent virus, emerging by reassortment of circulating wild-type (WT) virus with the vaccine virus.Development and testing of nonreplicating RV vaccines have also been pursued, and will be the focus of this chapter. The use of nonreplicating immunogens presents additional challenges beyond that of developing RV vaccines effective in young children against potential infection by multiple serotypes of RV. The nonreplicating immunogen must be able to induce protective immune responses against the target virus. Traditionally, nonreplicating vaccines have been thought to be poor inducers of mucosal immune responses and protection of the mucosa. Without amplification of the vaccine virus by replication, a high dose of nonreplicating immunogen may be required. To enhance the immune response to nonreplicating immunogens, development and testing of new adjuvants and/or delivery systems, and alternative routes of immunization to boost immunogenicity and protective efficacy, are needed. If administered orally, the nonreplicating immunogens must be stable in the digestive environments of the stomach and intestine.

Virus-like particle vaccines for mucosal immunization.

Viruses which infect the gastrointestinal tract are well suited for examining the immune response(s) to oral delivery of antigen and exploring the advantages and pitfalls of oral vaccines. We have used recombinant DNA techniques to produce nonreplicating self-assembled virus-like particles (VLPs) from two gastrointestinal viruses, rotavirus and Norwalk virus. Both of these viruses normally cause acute gastroenteritis in man or animals. The VLPs are morphologically and antigenically similar to the native virus and quite stable, features which are advantageous for their use as subunit vaccines. In addition, these VLPs could be useful as carriers of foreign epitopes from heterologous pathogens or of drugs which need to be delivered to the gastrointestinal track. This paper briefly reviews the properties of these VLPs made in insect cells and data showing their potential as subunit vaccines for parenteral or oral delivery.

Rotavirus infection in foals.

Fecal samples from 86 foals with diarrhea were examined by electron microscopy during a 2.5 year period. Of these, 26 (30%) were positive for rotavirus. All of the cases were found in epizootic areas. The disease was produced in an experimental foal by inoculation via stomach tube of a bacteria-free fecal filtrate containing rotavirus. Examination of postmortem tissues from the duodenum and jejunum of 2 naturally infected foals and an experimentally infected foal revealed replicating virus in the intestinal epithelial cells. A limited survey of complement-fixing antibody to rotavirus in horses from Kentucky, Virginia, and France indicated that all horses had antibody to the virus. The sole exception was 1 foal from which blood samples were collected prior to suckling. These results were presumptive evidence that rotavirus is a major cause of diarrhea in foals, and the presence of antibody in horses from diverse areas is evidence for the ubiquitousness of this infection.

IgA is important for clearance and critical for protection from rotavirus infection.

Based on a lack of severe phenotype in human immunoglobulin A (IgA) deficiency syndromes, the role of IgA in controlling respiratory and gastrointestinal (GI) infections has not been clearly defined. C57BL/6 and BALB/c mice lacking IgA (IgA(-/-)) were developed and used to address this question. When exposed to a common GI virus, rotavirus, IgA(-/-) mice exhibited a substantial and significant delay in clearance of the initial infection compared with wild-type mice. IgA(-/-) mice excreted rotavirus in stool up to 3 weeks after the initial exposure compared with 10 days observed in wild-type mice. Importantly, IgA(-/-) mice failed to develop protective immunity against multiple repeat exposures to the virus. All IgA(-/-) mice excreted virus in the stool upon re-exposure to rotavirus, whereas wild-type mice were completely protected against re-infection. These findings clearly indicate a critical role for IgA in the establishment of immunity against a GI viral pathogen.

Rabbit model of rotavirus infection.

A new small animal model was developed to study parameters of rotavirus infections, including the active immune response. Seronegative New Zealand White rabbits (neonatal to 4 months old) were inoculated orally with cultivatable rabbit rotavirus strains Ala, C11, and R2 and with the heterologous simian strain SA11. The course of infection was evaluated by clinical findings, virus isolation (plaque assay and enzyme-linked immunosorbent assay), and serologic response. All four strains of virus were capable of infecting rabbits as determined by isolation of infectious virus from intestinal contents or fecal samples, by seroconversion, or by a combination of these methods. The responses differed depending on the virus strain used for inoculation. Rabbits remained susceptible to primary infection to at least 16 weeks of age (upper limit examined). Virus excretion in intestinal contents was detected from 6 h to 7 days postinoculation. RNA electropherotypes of inocula and viruses isolated from rabbits were the same in all samples tested. Transmission of Ala virus and R2 virus but not SA11 virus from inoculated animals to uninoculated controls also occurred. In a challenge experiment with Ala virus, 74- and 90-day-old rabbits were rechallenged with Ala 5 weeks after a primary infection with Ala. Virus was excreted in feces from 2 to 8 days after the primary infection but was not excreted after challenge. These results indicate that the rabbit provides an ideal model to investigate both the primary and secondary active immune responses to rotavirus infections and to evaluate candidate vaccines.

Determination of the duration of a primary immune response and the ID50 of ALA rabbit rotavirus in rabbits.

The rabbit model of rotavirus infection has been used to examine the immune response to rotavirus infection and to evaluate strategies for rotavirus vaccine development. To determine the 50% infectious does (ID50) of tissue culture adapted ALA virus, rabbit were orally inoculated with 10(1)-10(3) PFU of ALA rotavirus. The ID50 of ALA virus was determined to be 1.7 x 10(2) PFU. The immune response induced in rabbits infected at low virus doses (10(2)-10(3) PFU) was of similar magnitude to the immune responses induced with a high dose (10(6) PFU) inoculum, indicating that the immune response to ALA rotavirus in rabbits is not dose dependent. To determine if a single rotavirus inoculation would induce a long lasting immune response, four rabbits were inoculated once with ALA virus (3.5 x 10(5) PFU) and their serologic and mucosal antirotavirus titers were monitored at intervals for 1.5-2 years. The infected rabbits maintained serologic and mucosal rotavirus antibody titers until the final time point more than 700 days post inoculation. These data are important because they indicate that the antigenic load achieved following a single oral inoculation is sufficient to achieve long lasting immunity, the goal of any potential vaccine.

Comparative amino acid sequence analysis of the outer capsid protein VP4 from four lapine rotavirus strains reveals identity with genotype P[14] human rotaviruses.

The genes encoding the outer capsid VP4 proteins of four lapine rotavirus strains, three isolated in the US (ALA, C-11 and BAP-2) and one isolated in Japan (R-2) were sequenced, and the predicted amino acid (aa) sequence was compared to all known rotavirus genotypes. A high degree of aa identity (96.8-98.9%) was found among the American lapine strains, while the Japanese rotavirus strain R-2 shared less aa identity (89.5-90.0%) with the American strains. The four lapine rotaviruses shared the closest aa identity (90.6-94.9%) with the P[14] genotype, consisting of viruses isolated from humans in Italy, Finland and Thailand. These results indicate that the VP4 protein of the four lapine strains are genotype P14, and that among lapine strains there are possibly two subtypes, one represented by the American lapine strains and the other by the Japanese R-2 strain.

Simian rhesus rotavirus is a unique heterologous (non-lapine) rotavirus strain capable of productive replication and horizontal transmission in rabbits.

Simian rhesus rotavirus (RRV) is the only identified heterologous (non-lapine) rotavirus strain capable of productive replication at a high inoculum dose of virus (>10(8) p.f.u.) in rabbits. To evaluate whether lower doses of RRV would productively infect rabbits and to obtain an estimate of the 50% infectious dose, rotavirus antibody-free rabbits were inoculated orally with RRV at inoculum doses of 10(3), 10(5) or 10(7) p.f.u. Based on faecal virus antigen or infectious virus shedding, RRV replication was observed with inoculum doses of 10(7) and 10(5) p.f.u., but not 10(3) p.f.u. Horizontal transmission of RRV to one of three mock-inoculated rabbits occurred 4-5 days after onset of virus antigen shedding in RRV-infected rabbits. Rabbits infected at 10(7) and 10(5), but not 10(3), p.f.u. of RRV developed rotavirus-specific immune responses and were completely (100%) protected from lapine ALA rotavirus challenge. These data confirm that RRV can replicate productively and spread horizontally in rabbits. In attempts to elucidate the genetic basis of the unusual replication efficacy of RRV in rabbits, the sequence of the gene encoding the lapine non-structural protein NSP1 was determined. Sequence analysis of the NSP1 of three lapine rotaviruses revealed a high degree of amino acid identity (85-88%) with RRV. Since RRV and lapine strains also share similar VP7s (96-97%) and VP4s (69-70%), RRV might replicate efficiently in rabbits because of the high relatedness of these three gene products, each implicated in host range restriction.

Relationship between anti-Tax antibody responses and cocultivatable virus in HTLV-I-infected rabbits.

The presence of anti-Tax antibody responses in human T cell leukemia virus type I (HTLV-I)-infected individuals has been correlated with increased proviral load, increased risk of transmitting infection, and increased risk of developing tropical spastic paraparesis/HTLV-I-associated myelopathy (TSP/HAM). In this study, a rabbit model of HTLV-I infection was used to determine whether anti-Tax antibody responses could predict the presence of virus with the potential to replicate. Seven of 14 HTLV-I-infected rabbits developed anti-Tax antibody responses. The onset of Tax reactivity was variable, but once detected remained constant throughout the remainder of the 60-week course of the study. All anti-Tax antibody positive rabbits produced virus as measured by p19 expression upon coculture, while p19 was detected in only one of the Tax antibody negative animals. Thus the presence of an anti-Tax antibody response correlates with p19 expression following cocultivation, and may be a useful predictor of virus replication in HTLV-I infected individuals.

Molecular characterization of three rabbit rotavirus strains.

We report biochemical (RNA and protein patterns and gene-coding assignments) and serologic (serotype and subgroup) properties of three strains of rabbit rotaviruses--Ala C11 and R2. The RNA electropherotypes were a standard "short" pattern for R2, an unusual "short" pattern for Ala, and an unusual "long" pattern for C11. Serologic studies indicated that these viruses were all group A serotype 3 rotaviruses. In addition, the Ala and C11 viruses were found to possess subgroup I specificity, whereas the R2 virus possessed subgroup II specificity. In contrast to their distinctive RNA patterns, the polypeptide patterns of the rabbit viruses were similar to those of SA11. To identify cognate genes and determine gene-coding assignments for the rabbit rotaviruses, cDNA probes of individual SA11 genes were hybridized to rabbit rotaviral genomic RNA segments that had been separated by polyacrylamide gel electrophoresis and transferred to filters (Northern blots). The order of genome segments 7-11 for each of the rabbit rotaviruses was unique and differed from that of SA11 genes. These differences were possibly due to rearrangements of the RNA sequences within these individual genome segments. Sequence analysis of the individual RNA segments will confirm whether genome rearrangements are the molecular basis for these novel migration patterns.

Molecular biology and immunology of rotavirus infections.

Rotaviruses were first recognized about 15 years ago in association with diarrhea in children and animals. Since then, rotaviruses have been determined to be the most important viral agent that causes clinically significant diarrhea in children and a need for an effective vaccination program has been recognized. This article reviews the progress which has been made in understanding the molecular biology of rotaviruses and summarizes information on the immune responses to rotavirus infections obtained in a new animal model in rabbits. This model is useful to systematically evaluate active protective immunity following infection of seronegative animals.

Antigenic and molecular analyses reveal that the equine rotavirus strain H-1 is closely related to porcine, but not equine, rotaviruses: interspecies transmission from pigs to horses?

We have sequenced the genes encoding the inner capsid protein VP6 and the outer capsid glycoprotein VP7 of the subgroup (SG) I equine rotavirus strain H-1 (P9[7], G5). The VP6 and VP7 proteins of the equine rotavirus strain H-1 shared a high degree of sequence and deduced amino acid identity with SG I porcine strains and serotype G5 porcine strains, respectively. Previous sequence analyses of the genes encoding the outer capsid spike protein VP4 and the nonstructural proteins NSP1 and NSP4 of equine H-1 strain also revealed a high degree of sequence and deduced amino acid homology with the prototype porcine rotavirus strain OSU (P9[7], G5). We have also confirmed and extended the VP4 and VP7 antigenic relatedness of equine rotavirus strain H-1 to porcine strains of P9[7] and G5 serotype specificities isolated in the United States, Venezuela, Argentina, and Australia based on cross-neutralization studies. In addition, the pathogenicity of tissue culture-adapted equine H-1, H-2, FI-14, FI-23, and L338, and porcine OSU rotavirus strains was compared in the neonatal mouse model. The 50% diarrhea dose (DD50) of equine H-1 was similar to that of porcine OSU and equine H-2 and L338 strains, while the DD50 of equine H-2 was > or = 50 or 315-fold lower than those of equine FI-14 or FI-23, respectively. Our sequence comparison of NSP4 of the rotavirus strains tested potentially identified amino acid residue 136, within the variable region spanning amino acids 130 to 141, as playing a role in virulence. Taken together, there is strong support to suggest that the equine rotavirus strain H-1 may represent an example of interspecies transmission from pigs to horses.

Serologic and mucosal immune response to rotavirus infection in the rabbit model.

We examined the humoral immune response to rotavirus infection in specific pathogen-free rabbits inoculated and challenged orally with rabbit Ala rotavirus (7.5 x 10(5) to 1 x 10(7) PFU). The humoral immune response in both serologic and mucosal samples was monitored by using total antibody enzyme-linked immunosorbent assays (ELISAs), isotype-specific ELISAs, and plaque reduction neutralization assays. Following a primary infection, all rabbits shed virus and serologic and mucosal antibody responses were initially detected by 1 week postinoculation. Intestinal immunoglobulin M was detected by 3 days postinoculation, and secretory immunoglobulin A was detected by 6 days postinoculation. Following challenge, rabbits were protected (no detectable virus shedding) from infection. An anamnestic immune response was observed only with mucosal neutralizing antibodies, and all serologic and mucosal immune responses persisted at high levels until at least 175 days postchallenge (204 days postinoculation). Detection of neutralization responses was influenced by the virus strain used in the neutralization assay; all inoculated rabbits developed detectable serum and intestinal neutralizing antibodies against the infecting (Ala) virus strain. Neutralization activity in both serum and mucosal samples was generally, but not exclusively, homotypic (VP7 serotype 3) after both primary and challenge inoculations with Ala virus. Heterotypic serum neutralization activity was observed with serotype 8 (9 of 12 rabbits) and 9 (12 of 12 rabbits) viruses and may be based on reactivity with the outer capsid protein VP4 or on a shared epitope in the C region of VP7. Comparisons of heterologous (serotype 3) and heterotypic neutralizing responses in mucosal and serologic samples revealed that 43% (21 of 49) of the responses were discordant. In 19 of 49 (39%) of these cases, a heterotypic serologic response was seen in the absence of a heterotypic mucosal response, but in 2 of 49 (4%) instances, a heterotypic mucosal response was seen in the absence of a concomitant serologic response. These results provide insight into factors which may affect detection of heterotypic responses.

Stimulation of local immunity and protection in mice by intramuscular immunization with triple- or double-layered rotavirus particles and QS-21.

Based on studies in animal models, parenteral immunization has become recognized as a potential vaccination strategy against rotavirus. Using an adult mouse model, the effects of the saponin adjuvant QS-21 on protection against murine rotavirus (strain EDIM) infection was determined following two intramuscular (i.m.) immunizations with purified EDIM particles including triple-layered (tl) infectious particles, tl particles inactivated with psoralen/UV, and double-layered (dl) inactivated particles. All three particles stimulated large serum rotavirus IgG responses and small amounts of serum rotavirus IgA, but undetectable stool rotavirus IgA. Inclusion of QS-21 during immunization increased the serum responses approximately 2- to 10-fold and also stimulated low levels of stool rotavirus IgA. Protection based on reduced shedding of rotavirus following EDIM challenge was significant (P < 0.001) with each immunized group and was enhanced (P < 0.001) by inclusion of QS-21 during immunization. Mice immunized with either live or inactivated tl particles and QS-21 were almost fully protected. Furthermore, animals inoculated with dl particles and the adjuvant shed significantly (P = .02) less virus following challenge than mice immunized with inactivated tl particles even though the latter induced measurable titers of neutralizing antibody to EDIM. These results demonstrate significant protection against rotavirus following i.m. immunization with both dl and tl EDIM particles which is consistently enhanced with QS-21.

Rotavirus 2/6 viruslike particles administered intranasally with cholera toxin, Escherichia coli heat-labile toxin (LT), and LT-R192G induce protection from rotavirus challenge.

We have shown that rotavirus 2/6 viruslike particles composed of proteins VP2 and VP6 (2/6-VLPs) administered to mice intranasally with cholera toxin (CT) induced protection from rotavirus challenge, as measured by virus shedding. Since it is unclear if CT will be approved for human use, we evaluated the adjuvanticity of Escherichia coli heat-labile toxin (LT) and LT-R192G. Mice were inoculated intranasally with 10 microg of 2/6-VLPs combined with CT, LT, or LT-R192G. All three adjuvants induced equivalent geometric mean titers of rotavirus-specific serum antibody and intestinal immunoglobulin G (IgG). Mice inoculated with 2/6-VLPs with LT produced significantly higher titers of intestinal IgA than mice given CT as the adjuvant. All mice inoculated with 2/6-VLPs mixed with LT and LT-R192G were totally protected (100%) from rotavirus challenge, while mice inoculated with 2/6-VLPs mixed with CT showed a mean 91% protection from challenge. The availability of a safe, effective mucosal adjuvant such as LT-R192G will increase the practicality of administering recombinant vaccines mucosally.

Rotavirus virus-like particles administered mucosally induce protective immunity.

We have evaluated the immunogenicity and protective efficacy of rotavirus subunit vaccines administered by mucosal routes. Virus-like particles (VLPs) produced by self-assembly of individual rotavirus structural proteins coexpressed by baculovirus recombinants in insect cells were the subunit vaccine tested. We first compared the immunogenicities and protective efficacies of VLPs containing VP2 and VP6 (2/6-VLPs) and G3 2/6/7-VLPs mixed with cholera toxin and administered by oral and intranasal routes in the adult mouse model of rotavirus infection. VLPs administered orally induced serum antibody and intestinal immunoglobulin A (IgA) and IgG. The highest oral dose (100 microg) of VLPs induced protection from rotavirus challenge (> or = 50% reduction in virus shedding) in 50% of the mice. VLPs administered intranasally induced higher serum and intestinal antibody responses than VLPs administered orally. All mice receiving VLPs intranasally were protected from challenge; no virus was shed after challenge. Since there was no difference in immunogenicity or protective efficacy between 2/6- and 2/6/7-VLPs, protection was achieved without inclusion of the neutralization antigens VP7 and VP4. We also tested the immunogenicities and protective efficacies of 2/6-VLPs administered intranasally without the addition of cholera toxin. 2/6-VLPs administered intranasally without cholera toxin induced lower serum and intestinal antibody titers than 2/6-VLPs administered with cholera toxin. The highest dose (100 microg) of 2/6-VLPs administered intranasally without cholera toxin resulted in a mean reduction in shedding of 38%. When cholera toxin was added, higher levels of protection were achieved with 10-fold less immunogen. VLPs administered mucosally offer a promising, safe, nonreplicating vaccine for rotavirus.