Maternal infections and maternal antibody
The maternal factors potentially affecting the success of neonatal immunisation are threefold: maternal-fetal transfer of antibodies, maternal-fetal transfer of pathogenic organisms and chronic maternal infections.
Maternal antibodies, transplacental or milk-derived, may potentially interfere with the neonatal adaptive immune response and subsequent vaccine responses [
91,
92]. This has already been shown in the context of maternally derived antibodies inhibiting immunogenicity of the measles vaccine in infants [
93]. The suppressive effect is thought to depend on the ratio between antibody titre and vaccine antigen dose, given that circulating antibodies hide the antigen from the host’s immune cells thereby masking B cell epitopes [
92]. This is exacerbated by the unique nature of the neonatal Fc receptor which increases the serum half-life of passively transferred antibodies [
5,
94]. By contrast, other studies have shown that the infant’s APC uptake and T cell responses appear to be largely unaffected by maternal antibody [
13]. Further research is currently ongoing to gain better mechanistic insights. Antibodies in breast milk may similarly limit the efficacy of oral vaccines at the mucosal surface, although this has been contested [
95‐
97]. A recent study by Ali et al. demonstrated that withholding breastfeeding around the time of rotavirus vaccine administration did not improve vaccine immunogenicity with no increase in anti-rotavirus IgA seroconversion measured [
98].
Transmission of pathogens across the placenta to the infant can also play a role in early life immunity with a potential impact on newborn vaccine responses: HIV-associated immune compromise, particularly cellular immune deficiency, exacerbates poor neonatal vaccine outcomes and some adverse effects. In a study carried out in South Africa between 2004 and 2006, the pooled incidence of disseminated BCG was estimated to be as high as 992 per 100,000 HIV-infected vaccinees, almost 1000 times higher than those who were HIV-uninfected [
99]. The mortality rate associated with disseminated BCG disease was > 70% [
99,
100]. Other pathogens that are vertically transmitted and may potentially affect neonatal immunity include toxoplasmosis, rubella, CMV and other herpes viruses, syphilis, enterovirus, parvovirus B19 and hepatitis viruses. A comprehensive discussion of the impact of each disease on vaccine responses is beyond the scope of this review.
Chronic maternal infections may shape neonatal vaccine responses by their impact on maternal antibody transfer across the placenta or their generalised effect on the developing neonatal immune system. This occurs independently of vertical transfer of pathogenic organisms and is secondary to exposure to microbial antigens, either free or complexed with immunoglobulins crossing the placenta. Subsequently, there is sensitisation in-utero and induction of adaptive, inflammatory and regulatory responses [
101]. The unique maternal disease profile during pregnancy may contribute to the observed differences in immune responses of infants in developed compared to developing countries [
102].
During pregnancy, prevalence of infection with at least one helminth species is 10–70% in endemic regions [
101]. Detection of specific immunoglobulins, particularly IgE, to various helminth species in cord blood reflects in-utero sensitisation of both B and T lymphocytes to microbial antigens, while parasites are rarely transmitted vertically [
103,
104]. Poor T cell responses to tetanus toxoid (TT) immunisation and reduced HiB vaccine-specific antibody responses have been reported in neonates born to mothers with lymphatic filariasis and/or schistosomiasis [
105]. Furthermore, Malhotra et al. found that Kenyan infants born to mothers infected with
Schistosoma haematobium and sensitised in utero had lower IFNγ responses to BCG than did unexposed infants [
106]. A similar pattern was seen in the context of maternal filariasis during pregnancy [
106]. By contrast, Elliot et al. concluded that maternal helminth infection may have little, if any, adverse effects on the outcome of infant immunisation. Of note, only maternal
Mansonella perstans infection was associated with significantly higher IL10 responses to BCG and tetanus immunisation but with no reduction in IFNγ, IL5 and IL13 responses [
107].
Maternal infection with
Trypanosoma cruzi and congenital Chagas disease induces a trend to type 1 polarisation of infant immune responses to vaccines. High levels of IFNγ were noted in response to BCG administered in exposed but uninfected infants compared with congenitally infected infants or unexposed and uninfected infants [
108].
Data on vaccine responses of HIV exposed but uninfected newborns are conflicting. On one hand, studies have found that their humoral and cell-mediated responses to vaccines, including to the measles vaccine and BCG, are similar to responses in unexposed infants [
109,
110]. Moreover, Jones et al. demonstrated that HIV exposed uninfected infants born in South Africa mounted higher increments of antibody responses to pertussis and pneumococcus vaccines [
111]. This may be because the reduced transfer of specific maternal antibodies noted in this cohort mitigated the poor vaccine-induced humoral responses seen in early infancy [
33]. The same research group also studied the impact of maternal HIV infection and demonstrated that HIV exposed, uninfected infants have normal responses to BCG vaccination administered at birth. T cell subsets are affected, however, which likely reflects priming of the developing immune system even in the absence of infection. Similarly, sensitisation to mycobacteria resulted in a positive correlation between maternal and newborn levels of BCG-specific CD4 T cells. This again suggests that maternal cellular responses may shape the infant response to BCG antigen. Following administration of BCG at 6 weeks of age, however, no difference in subsequent BCG-specific T cell proliferative responses or cytokine induction was documented at 16 weeks of age between those infants with either exposure to maternal HIV and/or latent TB [
112]. Therefore, any effects of maternal HIV and possible TB infection on infant immune profiles at birth are considered transient, as long as the infant remains uninfected. Given this, HIV exposed, uninfected infants are equally likely to respond to and be protected by BCG vaccination as HIV unexposed infants [
33,
112]. Not all studies, however, currently agree on this conclusion. Data from two Brazilian studies suggests that HIV exposed but uninfected infants have altered T cell responses to BCG with lower TT- and HBV-specific antibody levels [
113,
114].
Finally, placental malaria, secondary to maternal infection with
Plasmodium falciparum or
Plasmodium vivax, may impact neonatal immunity and vaccine responses. Above all, placental malaria is associated with reduced transfer of specific maternal antibodies, including RSV, varicella zoster, herpes simplex and measles and TT [
115‐
117]. This may be due to pathological changes in the placental structure in endemic settings, with subsequent damage to Fc receptors [
89,
115‐
117]. Furthermore, similar antibody and T lymphocyte responses to TT were noted in infants born to mothers with or without placental malaria [
118]. Ultimately, further work is required to elucidate these effects. The mechanisms may be similar to those proposed in maternal HIV, in that the limited transfer of maternal antibody may overcome potential blunting of neonatal humoral responses to early vaccination.
Neonatal co-factors
In addition to maternal factors, genetic variation among hosts plays a key role in the observed heterogeneity of neonatal and infant vaccine responses [
119]. Marchant et al. measured antibody responses to tetanus toxoid, measles and total IgG in 210 Gambian twin pairs, estimating the environmental versus genetic components of any variation documented. They found that genetic determinants control the early phase of the vaccine antibody response, while the environment predominantly influences antibody persistence and avidity maturation [
120]. More recently, further studies have applied genome-wide association studies in order to attempt to identify the genetic determinants of different vaccine responses against hepatitis B, measles and rubella [
121].
The quantity and quality of innate and adaptive responses to newborn immunisation are further impaired in babies born prematurely. As such, the concerns regarding immunogenicity and safety are even more marked, particularly given the lack of detailed immunological studies in this age group, mainly limited by the small amounts of blood available [
122,
123]. For a number of antigens, the humoral response to initial vaccine doses may be lower than that of term infants, but the protective threshold is often reached and memory successfully induced, sometimes with an additional dose to achieve persistence of protection [
123]. Again, the lower titres of maternal antibodies seen in prematurity may enhance this effect. On the other hand, it is widely accepted that HBV is less immunogenic in preterm infants with birth weights < 2 kg than in term infants [
124,
125]. Nevertheless, unless they have underlying medical complications, these infants respond to HBV by 30 days of age regardless of birth weight or gestational age [
124].
Baxter et al. concluded that immunogenicity in this cohort is vaccine specific: robust responses are seen following TT and inactive preparations; however, subunit formulations are less optimal. As such, they recommend administering additional booster doses and/or serology testing for infants < 32 weeks [
122]. Further research in this area would give important insights into the ontogeny of the maturing fetal and neonatal immune system and should be encouraged, given that novel tools are now available potentially applicable to very small volume samples.
In parallel to the significant impact of maternal co-infections and maternal-fetal disease transmission on neonatal immune responses, other co-morbidities in the newborn may have equally profound effects. One example is congenital non-infective immunodeficiency such as severe combined immunodeficiency disease (SCID). BCG vaccine has a very high rate of complications in SCID patients, as in HIV disease, with increased vaccine-associated morbidity and mortality rates [
126]. Disseminated BCG infection in children is also linked to rare immunodeficiencies of the IFNγ and IL12 pathways [
50]. Furthermore, acquired infections may be important. Few studies have elucidated these interactions in the neonatal context although more work has been done in older children. Infant malaria is associated with lower IFNγ, IL5 and IL13 (i.e. Th1) responses to both BCG and tetanus immunisation [
107]. In vitro incubation of monocytes with
P. falciparum-infected red blood cells has been reported to lead to enhanced responsiveness to TLR ligands, which may shape neonatal immune responses [
127].
Environmental
It is well established that neonatal immunity and therefore vaccine responses vary considerably across different geographical settings [
33]. For example, Lalor et al. demonstrated that 3 months post-BCG immunisation, 100% of UK infants made an IFNγ response to
M.tb purified protein derivative (PPD), compared to 53% of Malawian infants in whom responses varied by season of birth and possibly influenced by early exposure to environmental mycobacteria, as earlier discussed. Furthermore, there was a higher Th2- and lower Th1-type response to PPD in Malawian infants compared with UK infants [
128]. Similarly, Hur et al. reported that both geographical location (Malawi, Gambia and UK) and season of birth (dry vs. wet) significantly affected the cytokine profile and immunogenicity observed in response to PPD in infants at 3 months post BCG vaccination [
129]. However, the type of assay used meant that it was not possible to show which cells were primarily involved in mediating these differences and the underlying mechanisms remain unclear [
33].
Finally, environmental factors combined with host genetics influence bacterial colonisation after birth and the establishment of gut microbiota, regarded as crucial for optimal host immune development. This is illustrated by the finding that germ-free mice are at increased risk for infectious as well as autoimmune disease [
2,
130,
131] Currently, the impact of the microbiome on early vaccine responses is still poorly understood. In one study, TLR5-mediated sensing of the microbiota was found to drive lymph node plasma cell differentiation and enhance antibody responses to inactivated polio and influenza vaccines [
132]. Moreover, it is already well known that immune responses to oral vaccines in developing countries are suboptimal, possibly related to gut dysbiosis. Huda et al. investigated if stool microbiota composition predicted infant oral (OPV) and parenteral (BCG, HBV, tetanus toxoid) vaccine responses at 6, 11 and 15 weeks of age in a small group of 48 Bangladeshi infants. They demonstrated that
bifidobacterium predominance may enhance thymic development and responses to both oral and parenteral vaccines early in infancy. By contrast, greater bacterial diversity may cause systemic inflammation (neutrophilia) and lower vaccine responses [
133]. There is now need for similar studies in the neonatal period.