Vesikari: A simple question: as much as you talked about IgA, you did not mention what the target antigens of this IgA are. What are they?

Offit: There are a number of studies that have looked at the relative capacity of the different rotavirus proteins to induce virus-specific antibodies. The most compelling of these show that VP4 or VP7, or both, appear to invoke antibodies which neutralize virus infectivity in vitro and appear to protect against challenge in vivo.

Vesikari: I appreciate that answer, but my question was specifically about this IgA known to be protective in many systems.

Saif: A summary of our findings for human rotavirus (human Wa strain infection of gnotobiotic pigs) may shed a bit of light on this. Paul Offit has already presented the first part. If we give virulent human rotavirus orally to pigs, we can induce a high degree of protection against virus shedding and diarrhoea (Saif et al 1996, Ward et al 1996, Yuan et al 1996). This correlates with the high IgA antibody-secreting cell numbers in the duodenum; we see very low numbers in the systemic lymphoid system such as the spleen (Yuan et al 1996). After oral inoculation of pigs with the attenuated Wa rotavirus, one can see that we get similar protection rates against diarrhoea to those obtained in the oral rotavirus human clinical trials. Three doses (63% protection) are better than two doses (36% protection), and these protection data also correlate with IgA antibody-secreting numbers in the duodenum of the small intestine compared with the results in the spleen where we get more IgG antibody-secreting cells (Saif et al 1996, Yuan et al 1996). If we give the inactivated virus to pigs parenterally with adjuvant or orally without adjuvant, we get low numbers of IgA antibody-secreting cells in the intestine and low protection against diarrhoea and virus shedding (Yuan et al 1998). What we have done lately, to address your question about which proteins might be important, is to inoculate the 2/6 Wa rotavirus-like particles (VLPs) intranasally three times with mutant Escherichia coli heat-labile toxin (LT), because we wanted to compare these responses in pigs with the responses generated in mice to these same types of 2/6 VLPs (Yuan et al 2000). We got low numbers of IgA antibody-secreting cells in the small intestine and higher numbers of IgG antibody-secreting cells. In the pig model we could not induce protection against rotavirus diarrhoea or shedding by the intranasal route with mutant LT toxin and 2/6 VLPs. We think that VP6 alone, in terms of the experimental model of antibody-induced protection against diarrhoea, was not effective. However, because of the intussusception reports in terms of the attenuated, live rotavirus vaccine, we tried a combination vaccine (L. Yuan & L.F. Saif, unpublished results). Our first approach was to try the 2/6 Wa VLPs given intranasally twice, followed by attenuated Wa virus given once orally. This did induce some protection against diarrhoea. There were some IgA antibody-secreting cells in the intestinal tract and more IgG antibody-secreting cells in the spleen. This was not optimally protective, but it was a little better than giving the VLPs alone. Our recent studies suggest that if we give one dose of attenuated Wa orally followed by two doses of 2/6 VLPs plus mutant LT intranasally, we get a high degree of protection against virus shedding and diarrhoea and a moderate-to-high number ofIgA antibody-secreting cells in the intestinal tract. This suggests to us that we probably need to prime with an intact virus that has VP4 and VP7 but for some reason we can come back via a different mucosal route, intranasally, and boost effectively only with 2/6 VLPs. The mechanisms involved are unclear, but may involve cross-reactive, primed CD4+ T cells.

Koopmans: In connection with the data that Linda Saif described, we have some experience with work on polio virus. In The Netherlands we use an inactivated polio vaccine whereas almost all other countries use the oral live poliovirus vaccine (OPV). We have shown in human volunteers that people who had prior mucosal infection either with wild-type or OPV, get a very prominent booster IgA response with the injectable vaccine given later. We have shown from sorting circulating lymphocytes, that they are on their way to mucosal surfaces (Herremans et al 1997,1999).

Saif: Was this in children?

Koopmans: This was in adults.

Offit: The notion that one can boost a mucosal immune response by giving a parenteral vaccine only if there has already been a mucosal priming has also been shown for influenza. One of the reasons that the inactivated influenza works so poorly in children less than two years of age is that they haven't already been primed. It is hard to prime the immune system by giving a parenteral vaccine, but the vaccine does have some efficacy in older people because it acts as a booster.

Desselberger: With regard to the antigen against which protection might be directed, of course antibodies against VP4 and VP7 would be neutralizing, but there are data from Harry Greenberg's lab (Burns et al 1996), in which they showed that the VP6-specific IgA in the mouse 'backpack' model produced protection from shedding. There are also data from John Herrmann's group who tested a VP6-specific DNA vaccine and produced partial protection (Herrmann et al 1996, Chen et al 1997). The possible mechanism is that VP6-specific IgA antibodies transcytose as shown in Fig. 1 (Desselberger), interact with bilayered particles in the viroplasm and aggregate them. There are some good data from the influenza field by Mazanec et al (1992), who showed that influenza-virus-infected cells that were grown on a semipermeable membrane and were transcytosed with a nucleoprotein-specific IgA produced a significantly decreased titre of virus. I also remember that Harry Greenberg's group produced some data at a recent meeting trying to elucidate this model for rotaviruses.

FIG. 1. (Desselberger) Model of IgA transport and possible intracellular neutralization of rotavirus single-shelled particles in viroplasm by transcytosed VP6-specific IgA. IgA dimers are secreted from B cells located in capillaries or in the submucosa. IgA—SC dimers then bind to the basal plasma membrane of epithelial cells which express SC acting as a polymeric IgA receptor. IgA-SC is endocytosed and found in lysosomal compartments where it loses SC. It can then bind to specific proteins intracellularly; rotavirus VP6-specific IgA is thought to bind to single-shelled subviral particles which are often arranged in paracrystalline arrays in the viroplasm. (Reproduced from Desselberger 1998.)

FIG. 1. (Desselberger) Model of IgA transport and possible intracellular neutralization of rotavirus single-shelled particles in viroplasm by transcytosed VP6-specific IgA. IgA dimers are secreted from B cells located in capillaries or in the submucosa. IgA—SC dimers then bind to the basal plasma membrane of epithelial cells which express SC acting as a polymeric IgA receptor. IgA-SC is endocytosed and found in lysosomal compartments where it loses SC. It can then bind to specific proteins intracellularly; rotavirus VP6-specific IgA is thought to bind to single-shelled subviral particles which are often arranged in paracrystalline arrays in the viroplasm. (Reproduced from Desselberger 1998.)

Estes: We have to be careful when we are interpreting these data as to whether we are talking about model systems where people are measuring protection from infection, or model systems where people are measuring protection from disease.

Bishop: The IgA response in the intestinal tract must be at least partly due to VP6 because it can be measured by ELISA. There is neutralizing antibody there as well which can be measured in a neutralizing antibody assay (Coulson & Masendycz 1990). It is still difficult to identify the components of the intestinal tract response to individual viral proteins, even though it is possible to show a full range of responses in serum to viral proteins (Richardson et al 1993, Johansen et al 1994).

Greenberg: It seems that the number of memory B cells is also relevant to protection. Paul Offit, you are correct that it takes time for memory cells to become effector cells; the more memory cells you have, the faster this appears to occur, because the rate at which IgA is generated increases. In effect, the more memory cells you have in the gut, the better off you are after infection. How memory cells get into the gut and stay in the gut would be another fruitful research direction.

Offit: We have data which I didn't show that are relevant here. We did an in vitro analysis in which we tried to identify the frequency of memory B cells within gut-associated lymphoid tissue at various intervals after immunization with RRV. We found that 16 weeks after immunization, the principal location of virus-specific memory B cells was in the lamina propria. They were much less commonly present in the Peyer's patch, in the spleen, or in the mesenteric node. In fact, both the number and the site ofmemory B cells determined protection against shedding associated with reinfection. You are right; it is not just memory B cell number, it is also where they are. I would say, though, that even in the best of immunization schemes, it still took a number of days to elicit antibody-secreting cells from memory B cells. Once we had infected and allowed that infection to become quiescent in terms of production of IgA in the lamina propria, it still took at least four days in order to detect virus-specific antibody secreting cells derived from memory B cells in the lamina propria.

Greenberg: If the number was five times higher, it might have been three days.

Offit: It is possible; we could never do that.

Greenberg: That in part happens with wild-type infection.

Offit: Even with wild-type virus, we could never get to a point of quiescence; there was always production of IgA in the lamina propria.

Greenberg: It becomes a little bit of an 'apples and oranges' scenario. It has to do with numbers. You also have a hugely increased number of memory cells with wild-type.

Offit: Absolutely. The problem is that it is hard to sort out memory from effector cells if there is an ongoing effector cell response.

Greenberg: Cells become enteric memory cells because they have molecules on their surface that allow them to go to the gut. These molecules have classically been the integrin a4^7, but there is a whole set of new chemokines being found that also mediate intestinal trafficking. The most widely studied of these to date is CCR9 and its counter-receptor TECK. The ability to manipulate the way in which B or T cells develop those molecules on their surface would be another mechanism by which you could influence mucosal immunity. At the moment the only way we know how to do this is to immunize at the mucosal surface, but this simply means that there is some form of signalling that occurs at the mucosal surface that doesn't happen when you immunize parenterally. Presumably one could figure this out and induce memory cells that think they want to travel to the gut after systemic immunization. Nelly Koklin and I have asked the question, what is more important: the ability of B cells to secrete IgA antibody, or the ability of B cells to home to the intestine? We have looked at this in a combination of knockout mice in which we can delete either IgA or ^7. We asked, if you had to have one phenotype, which would rather have after a natural infection? The answer was interesting: mice that had IgG-producing cells that could home to the intestine were better protected than mice that had IgA-producing memory B cells that could not home to the intestine.

Offit: One always has to insert some level of caution in work with knockout mice. Margaret Connor recently published a paper in which she stated that IgG could be protective against challenge with rotavirus, in what was essentially an IgA knockout mouse (O'Neal et al 2000). What was interesting was that in her normal animals that were not IgA knockout mice, she was unable to detect virus-specific IgG at the intestinal mucosal surface. However, in her IgA knockout mice she was able to detect IgG at this surface. This makes you wonder whether or not these knockout mice had developed some compensatory mechanism to allow for the secretion of IgG onto the intestinal surface.

Greenberg: That is the point of what I was saying. In the IgA knockout mouse, those IgG B cells become a4^7 positive because the mouse compensates and allows its IgG cells to express high levels of the gut homing receptor a2^7.

Offit: It is still a polymeric immunoglobulin receptor at the basolateral surface.

Greenberg: I would rather have IgA and homing.

Offit: What gets IgG to the intestinal surface when it is a monomeric molecule? That is what is interesting.

Estes: We all know that knockout mice often have compensatory components, so we do have to interpret data from them with caution. Since you have raised the issue about IgG, we shouldn't forget its importance at the mucosal surface. It is clear from most of the natural studies looking at live rotavirus infection in animals or children, that the primary response that is measured is IgA. This doesn't exclude the fact that there is some IgG there, which is also a component of protection. If we are thinking about future vaccines, IgG at the mucosal surface could also be important.

Offit: I disagree. I am still not convinced that IgG plays an important role in protection against reinfection in the gut. I think it is true in the lower respiratory tract.

Greenberg: The critical question was what can play an important role. In natural rotavirus infection there is no doubt that in all the model systems tested to date, IgA appears to be a more important effector mechanism than IgG. This doesn't mean that IgG is not capable of being a fully powerful effector mechanism, depending on how one immunizes.

Offit: It still has to get to the intestinal mucosal surface.

Estes: It does. There is now evidence from a group at Harvard that there is a receptor in the intestine that will take IgG to the surface.

Matson: We evaluated children hospitalized for diarrhoea, but who did not have rotavirus diarrhoea at presentation (O'Ryan et al 1995). All were monitored and upon subsequent monitoring some developed nosocomial rotavirus diarrhoea and others did not. These children, whose mean age was 4 months, didn't have any IgA antibody, but they did have IgG. This IgG antibody was transplacental and they were not breastfeeding.

Offit: Did they have IgA antibodies in their blood?

Matson: No.

Offit: The question is, was the blood predictive of events that occurred in the intestinal mucosal surface?

Matson: They did not have serum IgA, but serum IgG titres predicted protection. This tells me that in natural infections transplacental IgG has some sort of a role in protection in the absence of IgA.

Saif: We have seen IgG antibody-secreting cells in the intestinal lamina propria which, in agreement with Paul Offit's comment, doesn't necessarily equate with IgG antibodies being secreted onto the luminal surface of the intestine. Just looking at the IgA literature on the secretory immune system, I would agree with the view that high levels of IgG antibodies on the surface of the mucosa is not highly desirable in resolving most natural infections, because IgG antibodies have the ability to fix complement and create inflammatory conditions there. I am not sure that this is the ideal situation. IgG antibodies can come in to play in situations where they act as a second line of defence but I think it is not desirable to have them in there in a long-term situation such as we see in the serum after infection to provide sterilizing immunity.

Estes: That is the classical interpretation of IgG being present in the intestine, but my feeling from all the experiments that Dr Conner has done in rabbits, where IgG is present at the mucosal surface, there is no evidence of any inflammatory reaction taking place there (Conner et al 1993, Ciarlet et al 1998, O'Neal et al 1998). There may be other regulatory mechanisms that, particularly in rotavirus infections, may be different from what has been seen in the past in other models.

Bishop: We may be confusing ourselves here with species differences, and perhaps with age differences. Grimwood et al (1988) looked at children with rotavirus infection, and compared duodenal mucosal antibody and serum antibody responses. With very few exceptions there was no detectable IgG response in gut contents at any level. However, circumstantial evidence implicates anti-rotavirus IgG in protection of newborn babies (R. F. Bishop & G. L. Barnes, unpublished results).

Estes: Let me clarify: in rabbits the primary response to live infection is an IgA response, but if you give inactivated virus or VLP it is an IgG response.

Greenberg: There are two critical questions. First, what happens in a natural infection. This is very important because as Paul Offit has said, natural infections do lead to protection. As you all know, however, many immunizations are not exactly the same as natural infection. Inactivated polio vaccination is not the same as wild-type polio infection, and it works just as well in terms of protection. It is wrong to assume that every strategy of vaccination has to recapitulate the exact form ofimmunity that occurs in natural infection. Ifwe mix up these two concepts, we will be talking at cross-purposes. A virus-like particle administered systemically may or may not protect, but if it does protect it sure as heck is not going to protect by the same mechanism as a natural intestinal infection. We need to understand what the mechanism is. We will hear more about intussusception later in this meeting. It is not clear how big a problem this actually is, but to the degree that it is a problem, we need to think about what might be an alternative immunization strategy if this problem persists with all orally given replicating vaccine candidates.

Saif: However, the polio vaccine story may not be the best example for enteric viral vaccines (Saif 1990). With polio vaccines, one can either block viral infection locally at the primary site of infection in the intestine by induction of local immunity, or one can use inactivated vaccines parenterally to induce systemic immunity which blocks the systemic spread the virus to the secondary site of infection, the nervous system. Induction of local intestinal immunity appears essential to prevent the infection of enterocytes by enteric viruses such as rotavirus which do not induce secondary systemic infections outside the intestine.

Kapikian: Paul Offit, do we need a vaccine that protects better than natural infection? The elegant longitudinal studies of Velazquez and others of rotavirus infections in a cohort of 200 children studied from birth to 2 years of age show that after two natural rotavirus infections, symptomatic or asymptomatic, no child developed moderate-to-severe diarrhoea (Velazquez et al 1996). We also know from the rhesus rotavirus-based vaccine studies that it is possible to achieve 100% protection against hospitalization and severe diarrhoea in certain settings. If a controlled vaccine infection mimics the wild-type infection and thereby induces protection, why do we need to achieve better protection than occurs under natural conditions?

Offit: I agree. I would say that the goal of the rotavirus vaccine is to prevent the moderate-to-severe disease that is associated with hospitalization and death. The point I was trying to make was that in that second year after immunization, when children are still getting severe disease, it may be of value to give a booster dose to keep up that level of IgA antibody.

Kapikian: If you vaccinate a child three times, at 2, 4 and 6 months of age, the memory should be maintained.

Offit: Memory only offers you so much. But I agree: the goal of the vaccine is to keep the children out of the hospital and keep them from dying, and this is a consequence of natural infection.

Kapikian: Dr Chiba and his colleagues deserve credit for elucidating a mechanism of protection against rotavirus illness. They showed in an infant home in Japan that a serum neutralizing antibody level of 1:128 was protective against illness caused by a homotypic virus. I believe that this was the first study to demonstrate a clear-cut effect of serotype-specific protection in a naturally occurring rotavirus outbreak in humans (Chiba et al 1986).

Vesikari: He showed that serotype-specific neutralizing antibody protects against reinfection, but the great mystery of all rotavirus vaccine studies is why is it that these vaccines protect regardless of the serotype of the surface antigen? I think some light was starting to be shed on this question when Harry Greenberg discovered this intracellular-working IgA. Then the question is, what is the mechanism of the IgA that acts intracellularly against VP6? I was very impressed by Dr Prasad's talk last year in Amherst when he showed the VP6 antibody effect to be through blocking the channels of nucleic acid release. It is an extremely important question as to how a non-neutralizing antibody can actually block rotavirus infection.

Prasad: We have also done the structure of VLPs bound with IgA. The IgA binding to VP6 causes a conformational change down below. It is affecting the transcription, and particularly the translocation of the transcripts through the type I channel.

Greenberg: I want to remind all of you of the old literature, some from Ruth Bishop and Barbara Coulson, some by Ian Holmes, some by me, and some from Koki Tanaguchi, all of which showed that of the monoclonal antibodies to VP4 and VP7, some of the molecules were serotype specific, but in all groups there were antibodies that were not serotype specific and had varying degrees of heterotypic reactivity. While VP6 is a very interesting target, the fact is that both VP4 and VP7 also represent targets that are capable of generating both type-specific and cross-reactive antibodies. There are molecular mechanisms to account for heterotypic immunity at the level of VP4, VP7 and VP6 on a humoral basis. We keep forgetting this.

Saif: In regard to VP6, there may be some mechanism whereby it is moderating the infection, but apparently in the neonatal gnotobiotic pig model it is not enough to control the disease. We did see some degree of reduced virus shedding, although it wasn't significantly different from controls with the 2/6 VLPs alone (Yuan et al 2000). In terms of the adult mouse backpack model in John Burns' and Harry Greenberg's work, the problem there is that you had a physiologically unusually large amount of antibody to get to those sites that you would never be able to reproduce under normal conditions. In addition, in subsequent studies by Ruggeri et al (1998), IgA and IgG hybridomas to rotavirus VP6 implanted as backpack tumours in neonatal BALB/c mice failed to protect mice against rhesus rotavirus-induced diarrhoea, whereas hybridomas secreting IgA monoclonal antibodies to rotavirus VP8* protected the neonatal mice.

Greenberg: The most convincing part of the VP6 data is not the backpack but the 2/6 VLPs that have also all been shown to protect.

Saif: This is in an adult model of virus shedding, not in the neonatal model of rotavirus diarrhoea.

Greenberg: Absolutely correct; that is why I said there is a basis for heterotypic immmunity at the level of VP4 and VP7.

Glass: Paul Offit, in your presentation, you didn't distinguish that one key feature of the epidemiology of rotavirus is the fact that newborns get asymptomatic infections. I would guess this might be mediated by maternal antibody and raises another mechanism of immunization. I am intrigued that the vaccine developed by Drs Ward and Bernstein, strain 89-12, which is a serotype 1 human strain, has caused diarrhoea in volunteers and is only partially attenuated. By giving the vaccine to infants under 3 months old, it may be maternal antibody that prevents any diarrhoea. What does this say about the role of IgG and maternal antibodies?

Offit: You are trying to distinguish, then, the maternal antibody that is passively transferred placentally from the maternal antibody which is passively transferred in milk. I think this is not a practical strategy for immunization in the USA, because disease occurs commonly in the 6—24 month old, a group not commonly breastfed.

Glass: The strategy there is to give the natural Wa infection to children who are protected by maternal antibody transplacentally.

Offit: It would surprise me ifthat immunization strategy could induce the kind of protective response that would be important up to two years of age.

How To Bolster Your Immune System

How To Bolster Your Immune System

All Natural Immune Boosters Proven To Fight Infection, Disease And More. Discover A Natural, Safe Effective Way To Boost Your Immune System Using Ingredients From Your Kitchen Cupboard. The only common sense, no holds barred guide to hit the market today no gimmicks, no pills, just old fashioned common sense remedies to cure colds, influenza, viral infections and more.

Get My Free Audio Book

Post a comment