Vaccination of neonates and infants is problematic because of two unsolved problems: the immature immune system of neonates and the presence of inhibitory maternal antibodies. A number of studies have determined that the immaturity of the immune system is most pronounced after birth and is overcome as the child develops. The immaturity (inability to fully respond to an antigenic stimulus) of the neonatal immune system has been observed in humans (1) and a number of other species, e.g., pig (2), cow (3, 4), and horse (5), and in experimental rodent models like mouse (1), rat (6), and cotton rat (7, 8).

Maternal antibodies are transferred from mother to child and protect neonates and infants during the time of maturation of their immune system. The vast majority of maternal antibodies are of the IgG isotype. In humans, maternal antibodies are preferentially transferred before birth transplacentally, and in animals of veterinary importance, preferentially through uptake of IgG in the intestine from colostrum within the first 24 h after birth. These passively acquired antibodies enter the bloodstream of offspring and act as a protective shield throughout the body in the same way as actively produced antibodies. Sometimes IgA antibodies contained in breast milk are also referred to as maternal antibodies. However, there are important differences in the action of passively transferred IgG and IgA antibodies. Upon transfer after birth, IgG antibodies are present in the bloodstream of the neonate in a finite amount that declines over time. These IgG antibodies suppress vaccine-induced immune responses. In contrast, IgA antibodies are continuously supplied through breast milk from the mother and protect the gastro-intestinal tract against pathogens without having an effect on the immune response. For the purpose of this review, the term “maternal antibodies” will be used for passively transferred IgG antibodies.

Maternal antibodies are very effective in protecting neonates and infants against most infectious diseases. The most impressive example is the protection of children with agammaglobulinemia (deficiency in the production of antibody) against bacterial infection for up to 6 months (9). Other documented examples of the ability of maternal antibodies to fully or partially protect are the amelioration of infection with respiratory syncytial virus (RSV) (10) or influenza virus (11) in humans, canine distemper virus in dogs (12), and infection with avian leukosis virus in chickens (13). Over time, maternal antibody titers decline because antibodies are being metabolized and do not protect any longer. However, even low, non-protective titers of maternal antibodies are still able to inhibit vaccination against infectious diseases of humans and animals. It is this phase of decaying maternal antibodies that presents a window of opportunity for infection by pathogens encountering the neonatal child or animal.

In humans, maternal antibodies wane over a period of 6–12 months (14–17). The kinetics of maternal antibody decline is correlated to the amount of maternal antibody present in the neonate after birth in that higher titers persist for a longer time. In contrast to humans, the duration of maternal antibodies in agriculturally important animal species is usually 3–6 months (18–20) and in chicken only 4–7 days (21). In contrast, maternal antibodies in bats persist similar to humans for 6–12 months (22–24). Maternal antibodies in all species have been reported to reduce or abolish antibody generation after vaccination. The reduction or lack of antibody typically results in reduced or absence of protection against disease (Tables 1 and 2). It is of interest to note that all types of vaccines (live-attenuated, inactivated, subunit, and experimental vaccines) have been reported to be inhibited. For a number of experimental vaccines immunization with variable success in the presence of maternal antibodies has been reported but these reports have to be evaluated based on criteria set out in Section “Evaluation and Development of Experimental Vaccines.”

The best studied example in human and veterinary medicine for the inhibition of vaccination by maternal antibody is measles vaccination. The measles vaccine virus (Edmonston strain) was developed in the late 1950s and early 1960s by attenuating a wild type virus on human and chicken embryo fibroblasts [for review see Ref. (62)]. Today various derivatives of the Edmonston strain are used as vaccine viruses worldwide (63). Immunization by both the standard subcutaneous and more experimental respiratory routes has been successful in protecting seronegative children (64). The protective immune response induced by vaccination consists of neutralizing antibodies against the two glycoproteins, with the (receptor-binding) hemagglutinin protein being the major target and the fusion protein the minor one (65). In addition, CD4 as well as CD8 T cell responses are induced. Case reports describing persistent MV infection in T cell deficient patients support the notion that T cells are required to clear infection but do not protect against infection (66). This role of CD8 T cells in virus clearance has been confirmed in the Rhesus macaques model (67). In cotton rats, CD4 T cells have no role in protecting or clearing virus from the respiratory tract (68). However, in mice they have a role in clearing virus infection from brain tissue through production of IFNγ (69, 70).

During their first year of life, children are protected by neutralizing maternal antibodies against MV infection. Over time, these antibody titers wane and eventually do not protect against wild type infection [for review see Ref. (71)]. Even these low, non-protective antibody titers inhibit seroconversion after both subcutaneous and intranasal immunization in cotton rats as well as humans. The lack of seroconversion severely reduces protection after vaccination with the live-attenuated vaccine virus. In the cotton rat model, this leads to a complete lack of protection (68). In clinical studies, immunization in the presence of maternal antibodies leads to partial reduction in mortality (72) and morbidity (73) (even in seronegative children). However, a broad clinically desirable protective immunity (protective T and B cell responses, and no clinical symptoms after infection rather then reduction in morbidity and mortality) is not established after immunization in the presence of maternal antibodies.

In contrast to the antibody response, MV-specific T cell proliferation is usually measurable after immunization in the presence of maternal antibodies (74–76). In spite of the inhibition of seroconversion, it was observed that vaccination in the presence of maternal antibodies leads to priming of B cells. Although no antibodies were produced, children immunized with measles vaccine in the presence of maternal antibodies responded with a higher immune response to booster vaccination after maternal antibodies had declined than children who had not been vaccinated at all (77). This study indicated that inhibition of B cell responses by maternal antibodies is a temporary effect and not due to the induction of anergy. However, it has also been observed that in children who were immunized in the presence of maternal antibodies long term antibody titers were reduced (even after boosting) compared to children who were seronegative at the time of immunization (76).

Since no current vaccine formulation is fully effective in the presence of maternal antibodies, two approaches have been used clinically to address the problem: one, the use of a high titer vaccine, and the other, the determination of the earliest time point possible for successful vaccination. The high titer vaccine (>10^4.7 pfu) had a 10- to 50-fold higher viral titer than the normal vaccine and induced some level of protection after immunization in the presence of maternal antibodies (29, 78). However, the use of this vaccine was associated with increased mortality (particularly in females) (79–81), which was attributed to immune suppression by the vaccine and its use was discontinued. A possible explanation for higher mortality in girls might be that girls have been reported to have lower maternal antibody titers (82) that were overwhelmed by a higher vaccine virus dose.

In a second approach, children were immunized at different times after birth (in the face of declining maternal antibodies). These studies have shown that a low level of maternal antibody correlates best with vaccination success. At the age of 6 months, maternal antibody titers are still high enough to suppress seroconversion but at the age of 9 months vaccination campaigns are relatively successful (31, 83). However, the complete disappearance of antibody at the age of 12 months seems to be optimal for immunization (32, 74, 84–86). Additional studies into the levels of maternal antibodies demonstrated that strong regional differences exist in that in a number of countries maternal antibodies might have disappeared by 6 months of age, and sometimes immunization at 4.5 months of age might be partially successful (87, 88). Regional differences might be due to a number of factors such as the exposure to wild type virus infection of mothers, which results in higher antibody titers and higher transfer rates into children than in vaccinated mothers (89). Maternal antibody levels in children of vaccinated mothers are lower and decline earlier than in children from naturally infected mothers.

Numerous reports detail the inhibition of antibody responses after vaccination in the presence of maternal antibodies. The important question is whether this is clinically relevant and whether clinical solutions exist to overcome the problem. For measles virus vaccination, maternal antibodies clearly inhibit the generation of neutralizing antibodies and therefore severely restrict protective immunity. However, the clinical outcome might vary depending on the specific vaccine and infectious disease under study. A recent study found that suppression of antibody responses against the tetanus and pneumococcal vaccines was more pronounced than against hemophilus influenza B and pertussis vaccines (25). Theoretically, immunization should be successful when maternal antibodies have declined below the threshold of detection. In practice, it is not feasible to accurately predict this time point as it depends on the amount of maternal antibody transferred, region, gender, nutritional status, and species. Therefore, the measurement of antibody levels would be required before immunization, which is not feasible in a clinical setting. The clinical practice to immunize children and animals repeatedly because most immunizations are based on a prime-boost principle, also ensures that children and animals are immunized when maternal antibodies have been metabolized. Immunization in the presence of maternal antibodies does not interfere with later vaccinations, and this practice ensures that eventually the individual is immunized when maternal antibodies have disappeared. However, a variable period of time without active immune protection will remain and therefore this vaccination schedule is most successful in regions with a low prevalence of infection. The other important question for vaccination is whether antibodies or T cells are a correlate of protection. If T cells play a major role in protection [e.g., Aujeszky virus (90)], early immunization will afford protection because the T cell response is affected very little by maternal antibodies. If, in contrast, neutralizing antibodies are required for protection, early immunization is not likely to succeed.

The transfer of maternal antibodies from mother to offspring may occur during pregnancy (from the maternal blood via transplacental transfer) and within 24 h after birth (from colostrum via the small intestine). The amount of antibody transferred by these two mechanisms differs by species depending on the type of placenta (http://placentation.ucsd.edu/placenta.html). Just prior to formation of the placenta, there are a total of six layers of tissue separating maternal and fetal blood. There are three layers of fetal extraembryonic membranes in the chorioallantoic placenta of all mammals, all of which are components of the mature placenta: endothelium lining allantoic capillaries, connective tissue in the form of chorioallantoic mesoderm, and the chorionic epithelium, the outermost layer of fetal membranes derived from trophoblasts. There are also three layers on the maternal side, but the number of these layers that are retained – that is, not destroyed in the process of placentation – varies greatly among species (see Table 3). Overall, species with few layers between maternal and fetal blood have a higher rate of transport of maternal antibodies transplacentally [through Fc-receptors (FcR)]. In species with more layers between maternal and fetal blood, the transport of maternal antibodies preferentially occurs through colostrum. Although the differences between species are important for the study of placentation and transfer of nutrients, for the study of vaccination in the presence of maternal antibodies the structure of the placenta and the mode of transmission of IgG is only relevant if intrapartum studies are performed. Most studies, however, assess the inhibition of vaccination sometime after birth and the relevant parameter in these experimental settings is the amount of IgG being transferred to the offspring.

Inhibition of vaccination has been studied in the presence of natural maternal antibodies or in the presence of passively transferred homologous or heterologous antibodies. Overall, inhibition of seroconversion after vaccination has been similar using both methods. For the induction of natural maternal antibodies, dams are immunized during pairing with the male and will transfer maternal antibodies to the pups. In rodents, differences can be found in antibody levels between pups of the same litter, which may occur because of the suckling hierarchy. The supposed advantage of natural maternal antibodies is the fact that antibodies from the same species have a predetermined fit for the FcR present in the species whereas heterologous IgG does not always interact with FcR (113, 114). However, studies in mice and humans have shown that subclasses of homologous IgG differ in their interaction with FcR (115, 116).

Often, the question of inhibition of vaccination by maternal antibodies will have to be separated from the immaturity of the neonatal immune system experimentally. One possible resolution is the transfer of homologous immune serum into adolescent animals (117). Using this experimental model, it was shown that suppression of vaccine responses correlates with amount of antibody present at the time of immunization (118). The injection of antibodies allows us to determine the amount of antibody transferred and ensures equal titers in all animals of the experimental group. In addition, often higher titers of passively transferred antibody can be achieved than would be possible through the induction of natural maternal antibodies. A disadvantage of homologous antibodies (whether natural maternal antibodies or passively transferred) is that they are indistinguishable from antibodies induced actively via immunization. The transfer of heterologous antibodies provides the advantage that actively induced and passively transferred antibodies can be distinguished by ELISA. Heterologous antibodies degrade faster than homologous antibodies, and e.g., in the cotton rat model the half-life of cotton rat IgG was estimated to be 7 days versus a half-life of 3.5 days for injected human IgG (Niewiesk, unpublished). The rate of decay should not matter for vaccination studies, as the amount of antibody at the time of vaccination is important for inhibition, not the duration of antibody levels. However, one should not inject heterologous antibodies repeatedly, as that will lead to the induction of regulatory T cells, which might interfere with the immune response after vaccination (119). Another caveat is warranted when monoclonal antibodies are used. It is important to test whether specific subclasses will interact with the FcγRIIB receptor in the respective species and be able to suppress B cell responses. In cotton rats, for example, mouse IgG1 antibodies do not bind to the FcγIIB receptor and do not interfere with vaccination whereas mouse IgG2a antibodies do (114).

A number of studies claim vaccine efficacy after immunization in the presence of maternal antibodies for both approved vaccines and vaccine candidates. For experimental vaccines successes as well as failures have been reported for a variety of vector systems expressing the vaccine antigen simultaneously with or without different cytokine (combinations). The results from these studies are contradictory and it helps to clearly define the experimental conditions and study design. A vaccine can most easily be proven to be successful if it is used in the presence of low titers of maternal antibodies at the time of immunization, if T cell responses are measured and if surrogate markers like histological changes (in animal models) are used for protective efficacy. To prove convincingly vaccine efficacy in the presence of maternal antibodies, levels of maternal antibodies at the time of vaccination have to be high, neutralizing antibodies should be measured as an immunological parameter, and protection should be measured as clinical efficacy by the absence of clinical symptoms or a significant reduction in viral/bacterial titers. By these standards, very few examples of successful immunization in the presence of maternal antibodies exist, thus necessitating further research into this area. When considering the development and testing of vaccines for immunization in the presence of maternal antibodies, three options seem to emerge as being potentially successful: inoculation at a site with low IgG, continuous expression of antigen, and inoculation with adjuvants stimulating type I interferon secretion. Inhibitory maternal antibodies are of the IgG isotype, and IgG is present in respiratory and ocular fluids and saliva (148, 149) although at lower concentrations than in the blood. Occasionally, vector systems have been used successfully in animal model in the presence of maternal antibodies when alternate routes of immunization (other than subcutaneous or intramuscular) were used (134, 150–152). Experimentally, this approach could be combined with measures of IgG titers at the site of inoculation (e.g., mouth, eye, and nose) to correlate vaccination success with lower antibody titers. The concept of continuous expression of antigen is based on the finding that repeated immunization leads to an antibody response in the presence of maternal antibodies. If a vector system can be found that continuously expresses antigen, B cell responses should be stimulated when antibody titers fall to very low levels. An example of this concept is the expression of infectious bursitis virus (IBV) proteins through an attenuated Marek’s diseases virus (MDV) vector that persists in the chicken and leads to good immunity (153) in the presence of maternal antibodies. The induction of type I interferon is based on the ability of type I interferon to drive B cell responses in tissue culture and in experimental systems in the presence of maternal antibodies. An example of this concept is the immunization of pigs with adjuvants inducing type I interferon (154).

Because of their relative immunological immaturity, infants are usually not immunized before the age of 2–3 months (depending on country-specific immunization schedules) with the most common exception of neonatal BCG immunization in some countries. In order to provide protection earlier in life, two other vaccination strategies have been discussed and/or used: cocooning and maternal immunization. The method of cocooning relies on immunization of all people in contact with the infant to minimize the risk of pathogen transmission and infection. Although this method can be effective it is sometimes difficult to implement. Another concept that has received more attention recently is maternal immunization. Maternal antibodies have been proposed as a means of protecting the infant during a sensitive time in the development of their immune system. If protective high levels of maternal antibodies can be achieved in infants, they would be protected during the most immature phase of their immune system. In addition, the young organism becomes better adapted to cope with infectious diseases by improved non-immune host responses, e.g., faster regeneration of epithelial cells after gastro-intestinal infection. Maternal immunization has the added advantage that the mother’s immune system is fully mature and will respond well to vaccination, thus providing protection to the mother and high levels of maternal antibodies to the infant. Based on the above named studies about protective levels of maternal antibodies and their kinetics of decline, it should theoretically be possible to induce high levels of antibodies, which are protective for up to 6 months.

Many examples exist that demonstrate the inhibition of both human and veterinary vaccines by maternal antibodies. B cell inhibition is mediated through a cross-link between the BCR with the FcγRIIB by a vaccine–antibody complex. Knowledge of the underlying mechanisms helps us to understand the possibilities and limitations of our current clinical approaches and will help in the development and testing of new vaccines. This knowledge also will help us to evaluate and develop the concept of maternal immunization as a mechanism to protect infants against common infectious diseases. The potential benefit of maternal immunization is protection of the mother and a delayed susceptibility to infection of the child. However, by most estimates children will not be protected for more than 6 months of life whereas full immunological maturity seems to be accomplished only after 12 months. A final unanswered question about maternal immunization, however, is how to deal with immunization in the presence of maternal antibodies. In consequence, the question of vaccinating in the presence of maternal antibodies will remain and still has to be resolved.

The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.