Scientist Stories: Arnaud Marchant, Immunobiology of Maternal Immunization

The first talk will be given by Arnaud Marchand, who is a professor at the Université Libre de Bruxelles in Belgium. And Arnaud has his MD and PhD from the same university. And his research has been focused on viral infections and particularly vaccines, and particularly in the very young. And of course, as we all know, vaccines— many vaccines are given to infants whose immune systems are not yet mature. Arnaud was postdoctoral researcher at the Medical Research Council Laboratories in the Gambia and also at the Weatherall Institute in Oxford. And he's currently research director at the Fund for Scientific Research in Belgium and the director of the Institute of Medical Immunology at the Université Libre. And as I've said, he is an expert in the immunology of the newborn and also has similar interests in art and music. And so it's with pleasure I introduce Arnaud Marchand.

Thank you so much, Stanley. Good afternoon. I'm very grateful for the invitation to discuss and share with you some of the ideas and projects we are developing around maternal immunization. So how does this work? So this is like this, backwards. So the idea we generally have about maternal immunization is pretty simple. We administer a vaccine to pregnant women to induce high levels of antibodies that are transferred to the newborn. So they are present at birth at high levels in order to provide protection against those pathogens, those infectious diseases that take place very early during life, first days, first weeks after birth. And so that strategy has, has been efficient in the control of neonatal tetanus. The implementation of maternal immunization against tetanus has had a marked impact on the incidence of this disease, and a number of countries now have eradicated neonatal tetanus. Maternal immunization can also be used to protect pregnant women against those pathogens against which they are highly susceptible, and a good example One example of this is influenza, whether pandemic or seasonal. And in this randomized trial fairly recently conducted in South Africa, pregnant women or maternal immunization of pregnant women provided protection, about 50% protection against influenza. And a similar protection was provided to the infant through the immunization of the mother. A more recently— a more recent target for maternal immunization is pertussis. Following the resurgence of pertussis in many northern countries, maternal immunization has been proposed to prevent those severe cases of whooping cough that are affecting very young children.. And that works. That has proven to be very efficacious in the countries where it has been implemented. So these successes generated a lot of interest and increasing momentum for maternal immunization. And actually, for the first time, we are developing vaccines specifically for pregnant women as the primary target to protect young infants against pathogens affecting them, including respiratory syncytial virus and group B strep. So those are the two main targets for maternal immunization for vaccines that are in development. While developing this vaccine, the question is, should we consider pregnant women as a specific population, or can we consider them as any other adults? And in some ways, pregnant women are like other adults, but in many ways they are unique. And this uniqueness has a lot of implications for maternal immunization, both in terms of clinical research and also in terms of implementation. At the safety level, administering vaccine or any drug to a pregnant woman raises concerns about the safety of the fetus and the outcome of the pregnancy. Yet the classical risk-benefit approach is probably not sufficient for the ethics of maternal immunization, and it seems more and more that an interest-based approach would favor the possibility that pregnant women would decide on the acceptability and the potential interests of immunization for their unborn baby. I will discuss the implications in terms of immunobiology. And so the two points I would like to make are that pregnancy is associated with dynamic changes. It's a period during which immune phenotypes, immune states dynamically change, and we are starting to understand how this could impact vaccine responses. And the second point, perhaps as important, is that maternal and infant immune phenotypes are linked. They are correlated. And so this suggests that if there is a window of opportunity in infancy for immune health for life, probably we should open that window more broadly and include pregnancy. And in the, In the slide that Dr. Verma showed this morning, the human march cartoon, you've seen that the march started with a toddler, right? And not with a pregnant woman. Maybe it should start with a pregnant woman, and that would allow also a woman to make it into the picture. Good. So this is the plan. This is what I will be talking about. And actually, when you think about maternal immunization, it's important to realize that the idea is not new. And actually it's one of the first discoveries in immunology. The idea that we can protect infants, offspring, by immunizing during pregnancy or during lactation was actually discovered by Paul Ehrlich at the end of the 19th century. And actually what he did was to immunize a number of species during pregnancy, during lactation, showed protection against toxins, and he showed that this was associated with, this was correlated with the production of antibodies. Antibodies. Actually, according to Art Silverstein, this is the first description of the kinetics of an immune response ever. So it's really back then. And actually also, Ehrlich had to fight for and to really argue to make this case. Because at the time, people thought— a lot of people thought that if immunity had to be transferred from one generation to the other, obviously it had to be the father. Who had to be at the origin of this. So he had to fight to really show that the mother was actually the source of immunity. So old ideas. And when you look at history, you see that actually little progress has been made in our understanding of the immunobiology of maternal immunization. And recognizing the need to increase our knowledge on this topic, the Gates Foundation funded a landscape analysis to review the current knowledge and also identify gaps in the immunobiology of maternal immunization. And together with Toby, Manish, and other colleagues, we conducted this analysis and published it. And so we found gaps that are across pathogens that are generic and some that are specific for the pathogens that I was alluding to. So what we learned from this analysis is that vaccines are immunogenic in pregnant women. When you compare to non-pregnant women, some studies, particularly flu studies, showed slightly decreased antibody levels in pregnant women as compared to non-pregnant women. But overall, vaccines are immunogenic in pregnant women. And that's the reason why we use them in that population. But an important gap in knowledge that we identified is that we have very little understanding, almost no understanding actually, of the impact of pregnancy on the quality of these antibodies. We know virtually nothing about— oops, sorry— about the impact of pregnancy on affinity maturation or antibody repertoire. Nothing about the structure, the impact of pregnancy on the structure of Fc fragments, the subclass, the glycosylation profiles that have profound implications and roles to play in the functional— in the effective functions of antibodies. And we know that most of the diversity of antibodies lies at the Fab fragment, but actually a lot of the diversity also lies at the Fc fragment through IgG subclass and through glycosylation patterns. And what we know is that Pregnancy is one of these states which is associated with changes in the glycosylation profiles of antibodies. Total IgGs have been studied, and consistently people have shown that galactosylation, silylation of antibodies increase during pregnancy with potential implications for inflammation, NK cell activation, complement activation. So it is very intriguing to think about what is the role for these changes. We don't know. It's also intriguing to think about what could be the implications for vaccine responses. We also know that antibodies are transferred across the placenta through syncytiotrophoblasts. They are binding to the near-LFC receptor within endosome at acidic pH, and that binding prevents degradation from the— of the IgGs. What we don't know is what happens what allows the IgG to cross the stroma and also the fetal endothelium. And because these cells don't express the neonatal Fc receptors but do express other Fc gamma receptors, it's likely that these receptors are involved. But we really have to learn about these processes. They remain poorly understood. There is selective transfer of IgGs across the placenta, IgG1s being the most transferred, IgG2 the least transferred. But we are starting to understand now that glycosylation profiles and perhaps functional profiles are also selected by the transport through the placenta. So not all maternal antibodies reach the fetus and the newborn. And when those antibodies get into the fetus and the newborn, the cells with which they will interact are actually newborn or infant cells that are— we are learning more and more— differently programmed as compared to adult cells. And the outcome of this interaction and the basis for pathogen control is virtually unknown. Another outcome of having high levels of maternal antibody in an infant is vaccine interference. For a number of vaccines, high levels of maternal antibodies in the cord blood or in the infant blood are associated with decreased vaccine response in the infant. Again, the basis for that process is largely not understood. The model that's favored today is the— is a model that involves binding of the maternal antibody to the Fc gamma receptor 2B at the level of the naive B cells, which would trigger a negative signal and would decrease the production of antibodies by plasma cells. So many gaps. Limited understanding of how maternal antibodies function in the infants, about the way they are transferred across the placenta, and also about the best way we could induce them, induce optimal antibodies in the pregnant woman. But if these gaps look actually pretty big, they're probably only a fraction of the complexity and the landscape of the links of the parameters that link maternal and infant immunity. The Gates Foundation, together with other institutions, are increasingly recognizing that the health of the infants is very strongly linked to the health of the pregnant woman, of the mother. And if a mother, a healthy mother is more likely to give birth to healthy child, a vulnerable child is much more likely to be born to a diseased mother. And because these diseases affecting infants are largely infectious diseases, it can be proposed that immune fitness of the pregnant woman is a key determinant of immune fitness of the child and that interventions could positively impact immune fitness of both the pregnant woman and the child. And so that allows us to identify a number of windows of opportunities— the pregnancy itself, the transfer of the immunity, the infancy, but perhaps most importantly, the integration of the pregnant woman and the infant, the mother-infant diet. As a group of investigators, we decided that we would explore these opportunities. And so we set up together a program called MADI, for Maternal Determinants of Infant Immunity. Overall, our aims are to unravel the fundamental rules governing immunity in the mother-infant dyad together, and to identify targets for new and improved interventions, including vaccines, but potentially also other interventions. Targeting the changes of immune states taking place during pregnancy, the changes in the microbiome that are taking place during pregnancy. And so we've started using systems biology tools to look at the determinants of vaccine responses during pregnancy. Here, as an example, in a pilot study that we've conducted recently in Belgium, we immunized pregnant women with pertussis, against pertussis, as it is recommended. And what we're seeing here is 7 days post-immunization, we see big changes in the transcriptome gene expression profiles in peripheral blood. And what's reassuring is that we see changes in pathways that have been identified through other systems, vaccinology studies, plasma cells, T cells. And so what we're looking at now are what are the predictors, the correlates, the predictors of the vaccine response in this group, comparing them to non-pregnant women to start identifying potential specificities for the vaccine response in the pregnant woman. But obviously antibodies are not the only maternal immune factors that impact immunity in the child. Essentially through animal mouse studies, we have evidence that beyond antibodies, other factors including immune complexes, perhaps free antigens, microbes, microbial products, cytokines themselves, nutrients, and even maternal cells impact either the development of the immune system or immune responses in early life. And so to start addressing the possibility that those maternal factors could impact immunity in the child, in the infant, we've started looking at correlations between immune phenotypes in the mother and in the newborn. And as you would expect,, we see huge correlations in antibody titers. So vaccine antibodies are highly correlated at birth. Maternal and newborn antibodies are highly correlated. So that was expected. What was less expected and actually quite surprising was that immune phenotypes— so the phenotype of innate cells in peripheral blood— are also very correlated. Here we're looking at the expression of CD40, a marker of monocyte activation. At the level of classical monocytes in mothers and infants. And we see high correlations at birth. We see persistent correlations at 6 months of age. And this is under physiological state. It is also true in diseased conditions, in HIV-infected mothers and their babies. And we started exploring further these correlations looking at the proteome, serum proteome of mothers and babies. And we see massive correlations. We see in some cohorts dozens of proteins that are correlated, in some cohorts hundreds. Of proteins that are heavily correlated at birth between the mother and the infants. Many of these proteins are immune proteins, and some of them are actually candidates for the regulation of the myeloid cells and the phenotypes that we see in peripheral blood. So why do we think, or what is the evidence that these correlations might be important? Actually, some of this evidence comes from a recent a study that we conducted recently in babies born to HIV-infected mothers. So these babies are not themselves infected by the virus, but the literature indicates that they have, despite the fact that they are not infected with HIV, they have an increased susceptibility to infection. So we wanted to look at what were the determinants of this susceptibility in this prospective cohort. So this is about 250 children. And so we comparing the incidence of hospitalization for infection during the first year of life in HIV-exposed and infected, HEU children, and HEU, so HIV unexposed, so the controls. And what you see is that these HIV-exposed children have about 2-fold increased risk of being hospitalized for infection during the first 3 months of life. And the main determinant, the major risk factor for hospitalization for infection was the time at which antiretroviral therapy was initiated in the mother. Essentially, all the risk was observed in children born to mothers who had initiated therapy only during pregnancy, whereas those children who had— who were born to mothers who had initiated therapy before pregnancy and maintained, obviously, therapy during pregnancy were at similar risk as compared to the controls. 2 quite fundamental implications here. The first one is that maternal health matters for the health of the infant. And the second is that you can intervene. By controlling the viral infection, you can improve the outcome of the child. Next, we, we thought, can we identify immune markers that are associated with this clinical phenotype or clinical outcome? As you may expect, we saw immune activation in HIV-infected mothers. And the highest activation was, again, as you may expect, observed in those mothers who initiated therapy later, so during pregnancy. What was less expected was that this immune activation was also seen in their newborn. The children born to HIV-infected mothers have immune activation at birth. And you see that at the level of classical monocytes, intermediate monocytes, or non-classical monocytes. So this is a phenotype that's quite consistently observed. In association with this immune activation, we see reduced transfer of maternal antibodies. So those children born to HIV-infected mothers have reduced transfer of antibodies against a number of specificities. These are the ratios. So cold blood to maternal antibody ratios. So pathogens or vaccine-specific antibodies. Antigens, we see decreased transfer. And the lowest transfer, again, is observed in the highest-risk groups, so those children born to mothers who had lower control of the disease. So what we observe here is that the children who are at the highest risk of clinical outcome also have the strongest immune alterations. So we then decided that we should try and model these data and see whether the immune factors, the immune parameters measured at birth, could predict the risk of hospitalization. And that's what we saw. By modeling the data in the highest risk group, so immune parameters at birth, you can distinguish those children who will be hospitalized in the following months as compared to the children who will not be hospitalized. And you reach high classification accuracy. I think this is the first study identifying immune parameters at birth predicting the risk of infection. In the following— in the next months of life. So definitely maternal health matters. So I think I tried to make a case for a better understanding of the immunobiology of pregnancy as a way to optimize and improve the way we immunize pregnant women and protect young infants. I'm quite convinced that there are many more determinants of infant immunity within the mother. So beyond antibodies, a number of maternal determinants. I think it will be very, very important to integrate these parameters. And antibodies don't work in isolation. They work in the context of a complex immune system that can be influenced by the maternal immune system itself. And eventually, the idea is that improving maternal immune fitness whether it is through immunization or through new interventions that we haven't identified yet, we have the potential to improve infant immunity to pathogens. So let me quickly thank the people who have participated in the clinical study that I've just shown you, and perhaps focus on Toby, with whom we've been discussing those issues for a while, and Dishnu Das in Galitz's lab. Has been really key in the analysis of the data and really opening new avenues for us to look at clinical studies. Thank you so much.

Thank you, Arnaud. Very interesting data. And a very simple question. There's some slight controversy about when during pregnancy to immunize the mother. That is, which trimester, which week. But what is your opinion?

I think that the controversy is related to the lack of data. And I think that data are being accumulated which will help us defining this better. I think the data you're referring to suggests that by immunizing earlier during the second trimester, you reach higher level in the cord blood. And I think that makes sense because you have more time for antibody transfer. But I think that people are looking at that more systematically now and will have better data soon.

Thank you.

Other— yes?

I realize it's not your area, but I wonder if there's any key reflections you would draw from the Zika experience? In terms of? In terms of maternal and child immunity and passage of virus or protection.

Yeah, I think that's what I've been trying to discuss will apply to many pathogens. And the congenital infections, like for CMV, which is at least as important as Zika, I believe, is another example. Indeed, you have a maternal immune system which has its own characteristics and influences the immune system of the fetus. So the determinants of disease effects in the fetus are not clear, I don't think, for CMV or Zika. And I think the question of whether it's pathology induced by the virus or there is a component of immunopathology is open. So I think those dimensions are potentially relevant.

Yes, Marcelo.

[INAUDIBLE] —can transfer an activation of monocytes in the mother to an activation of monocytes in the baby? I hope to be able to answer that question, you know, in the coming years. We don't know. I think that the possibility that this is, you know, passive transfer and that, you know, cytokines or mediators cross like that, I think that this possibility is unlikely. Placental perfusion models indicate that cytokines don't cross. The notion that's emerging is that the placenta itself is a source of mediators, and I start to see the placenta as a computer capturing signal from the mother and releasing new signals to the fetus. That's the way I would imagine things work. Whether everything derives from the placenta, I don't know, but I think it's a good start to try and modelize this.

Yes, last question there.

Is there any evidence for maternal-fetal transfer of immune memory?

Maternal cells— I think there are several ways to answer that question. Maternal cells do cross the placenta, and a number of groups are looking at that in animal models. And it could well be that, you know, the number of cells could provide some basis for that. And I think more philosophically, yes, antibodies, you know, transfer immunological memory at the population level. So the immunological memory of the mother is indeed transferred through the offspring through antibodies. And I think that the fact that those antibodies are present will influence the way the infant will respond to pathogens. So this is a number of papers. I think that makes sense. You know, immunological memory, at least at the antibody level, is at the population level, is transmitted. Whether it happens with cells as well, I think needs more data.

I know.

So Yeah, this is a difficult topic. So there is evidence that you do have IgMs and perhaps T cells induced in utero. People interpret that as transfer of the antigen and priming of the fetal immune system. Robustness of that, I don't think is there yet.

All right, thank you very much, Arnaud. So—