In rural Gambia, infants vaccinated during the rainy/hungry season had significantly higher antibody responses to their first DTP vaccination, compared with infants vaccinated during the dry/harvest season, despite presenting greater signs of undernutrition in early life. Of note, infants vaccinated during the rainy/hungry season spent the majority of their foetal life during the dry/harvest season months, at a point when their mother’s nutritional status was better as indicated by higher mean BMI, weight gain and mean haemoglobin levels. The timing of these effects suggests that the observed seasonal differences in infant’s antibody response to vaccination may be related to seasonal variations in maternal nutritional status during pregnancy. This is in line with the well documented critical role of prenatal nutritional exposures on the programming of the immune system and capacity to respond to vaccination in infancy ([10, 40).
Studies in both animals and humans have demonstrated that the complex immunological pathways involved in vaccine-induced antibody responses can be altered by both endogenous factors and exogenous exposures [35]. A number of factors may influence antibody responses to vaccination in infants including factors related to the vaccine itself (e.g. immunogenicity, adjuvant, dose and administration route), intrinsic host factors (sex, age, genetics and comorbidities) and extrinsic factors (diet, infections and microbiota) [39]. Moreover, maternal factors during pregnancy and lactation (e.g. nutritional status, immunisation status and infections) and perinatal factors (e.g. prematurity) may all modulate vaccine responses in infancy. The findings of our study highlight that, in rural Gambia, where many aspects of health and behaviour are dictated by a pronounced bimodal seasonality, seasonal exposures in early life may be a relevant factor in influencing antibody responses to vaccinations in infants.
Previous research highlights the importance of seasonality on vaccine responses. A systematic review published in 2015 identified 17 studies examining the impact of season of vaccination on antibody responses, including nine focusing on infant antibody responses to childhood vaccinations [26]. Seasonal variation in antibody response to the Rubella vaccine was examined in two studies: in Israel (n = 203) [19], the strongest Rubella antibody response was detected in infants vaccinated in the winter season, while in the Netherlands (n = 718), no seasonal differences were observed [8]. Four studies assessed seasonal differences in antibody responses to oral polio vaccine (OPV) in infants: in Israel (n = 121) [33] and India (n = 50) [25], the strongest antibody responses to OPV were observed in the winter season. In The Gambia (n = 679), however, OPV antibody responses were the highest in the dry/harvest season [20] while no seasonal variations were reported in Brazil (n = 730) [20]. Additionally, in The Gambia, a previous study reported seasonal trends in antibody responses to hepatitis B vaccine (HBV) in 121 infants of 52 weeks of age, and to diphtheria and tetanus vaccines included in the DTP vaccine in 138 infants at 16 weeks of age [22]. In another study from The Gambia, antibody responses to a 9-valent pneumococcal conjugate vaccine (PCV-9) were found generally higher in infants (n = 212) vaccinated during the rainy/hungry season [31]. Despite inconsistencies between settings and specific vaccines, the conclusions of this systematic review are in line with those of our study, indicating seasonal immunomodulation of vaccine responses in infants.
In the current study, infants vaccinated during the rainy/hungry season had significantly higher diphtheria, tetanus and pertussis antibody responses following their first DTP vaccination compared with infants vaccinated during the dry/harvest season. Tracking back to months spent in utero, all infants vaccinated during the rainy/hungry season were in mid-gestation across the dry/harvest season months. Mothers of these infants were less likely to be anaemic and underweight, had higher BMI, haemoglobin levels and gained more weight during pregnancy compared with mothers of infants vaccinated during the dry/harvest season who spent more of their pregnancy in the rainy/hungry season. These results support a possible role of seasonal variations in maternal nutritional status in shaping infant antibody responses to the DTP vaccines. Furthermore, infants with the highest antibody responses to the first DTP vaccination were the least exposed in utero to the rainy/hungry season; Their exposure was limited to a maximum duration of 2 months and split across the first and last of month gestation, which may be the least critical stages of foetal immune development.
A previously published analysis of data from the ENID trial demonstrated that maternal nutritional supplementation during pregnancy with multiple micronutrient and protein-energy improved the infant antibody response to DTP vaccinations [24]. In the current study, seasonal differences in DTP antibody responses were found to be attenuated by maternal nutritional supplementation during pregnancy. These results suggest that prenatal exposure to the rainy/hungry season during critical developmental stages may compromise immune development and the ability to mount an optimal immune response to vaccination in infancy, and that maternal nutritional supplementation during pregnancy may help to some extent reducing these detrimental effects. Together, these findings reinforce previous evidence of the pivotal role of maternal nutrition during pregnancy on the development and function of the immune system in the offspring.
Currently, the immune processes linking seasonal exposures to vaccination responses are not fully understood. A study conducted in rural Gambia showed that infants (n = 138) born during the rainy/hungry season, had higher levels of leukocytes and lymphocytes compared with infants born during the dry/harvest season [6]. Similarly, a study conducted in Denmark (n = 700) examining the influence of season of birth on 26 different immune cells subsets and 20 cytokines and chemokines, showed that at 1 month of age, infants born in winter had the highest levels of all immune cell types and mediators while infants born in summer had the lowest levels [34]. In most vaccine responses, including those elicited by the DTP vaccine, antigen-presenting cells (APC) such as B-lymphocytes, dendritic cells and macrophages are activated by vaccine antigens (e.g. diphtheria and tetanus toxoids or whole killed pertussis bacteria), initiating a cascade of processes involving many components of the immune system. Therefore, a seasonal effect on the development of immune cells involved in vaccine-induced antibody responses may provide a mechanistic explanation for the better DTP vaccine responses in infants born and vaccinated during the rainy/hungry season observed in our study.
Following three doses of the DTP vaccine, seasonal variations in vaccine responses at 24 weeks of infant age were opposite to those observed at 12 weeks; infants vaccinated during the rainy/hungry season who presented the highest diphtheria vaccine responses at 12 weeks had the lowest vaccine response at 24 weeks. Interestingly, although these infants had more favourable nutritional exposures during foetal development compared with infants vaccinated during the dry/harvest season, they were also the most exposed to the rainy/hungry season in early life and were more likely to present signs of undernutrition at 12 weeks of age, at the time of the second DTP vaccination.
A possible explanation for these results is that whilst infant nutritional status may have a limited effect on vaccine-specific antibody titres, it may have an impact on long-term immunity and recall responses. It has been shown that malnutrition may not impact the number of circulating lymphocytes and that the majority of undernourished children are capable of mounting an immune response to vaccination [27]. However, the thymus gland, which is critical for the maturation of T-lymphocytes and the formation of memory T-cells, seems affected by malnutrition even in mild cases of undernourishment, putting this organ forward as a hallmark of nutritional status [30]. Notably, in rural Gambia, a smaller thymus and lower numbers of thymus-derived T-lymphocytes were measured in infants during the rainy/hungry season compared with the dry/harvest season [5]. It has been shown in animal studies that malnutrition has a detrimental impact on the homeostatic proliferation of CD8+ memory T-lymphocytes and recall responses following vaccine challenges [14]. These findings suggest that the lower diphtheria vaccine responses observed in our study at 24 weeks, in infants vaccinated during the rainy/hungry season, may be explained by their exposure to the rainy/hungry season in early life. This exposure may have compromised their immune system and the induction of immune memory following vaccination challenge, resulting in an impaired recall response.
Conversely to diphtheria vaccine responses, pertussis vaccine responses at 24 weeks remained significantly higher in infants exposed to the dry/harvest season at mid-gestation and no differences in pertussis vaccine responses were observed by season of vaccination. This result may be explained by the rate of loss of antibodies taking longer to appear for pertussis vaccine responses. Antibody decay is influenced by the nature of the vaccine itself; while pertussis vaccine antigen consists of a whole killed pertussis bacteria, diphtheria and tetanus vaccines contain toxoids (inactivated toxins) secreted by the bacteria. Previous studies suggest that pertussis vaccine antigens can induce CD4+ and CD8+ T-cells responses which may result in a slower loss of antibodies and confer longer-term immunity [32].
Our results show that tetanus vaccine responses measured in infants at 24 weeks did not differ by season of gestation or vaccination. This may be related to the high immunogenicity of tetanus toxoid vaccine; after only one dose of tetanus vaccine over 95% of infants presented protective tetanus antibody levels compared to about half of infants for diphtheria and pertussis vaccines, and virtually all infants presented protective tetanus antibody titres after three doses of the vaccine. The prime-boost strategy is often implemented in vaccination programmes to increase vaccine responses and allow all subjects to reach protective antibody levels [1]. Therefore, the elevated tetanus antibody titres along with the effects of recall responses may conceal seasonal pattern in tetanus vaccine responses.
Although the results of this study suggest a strong link between seasonal variation in nutritional status and seasonal differences in DTP vaccine responses in infants, other seasonally-driven factors known to impact on immunity may be also involved. These include circulating vaccine antigens in the environment [37], sun exposure and vitamin D status [16], toxin exposure − notably aflatoxin exposure in this setting− [4] and air pollution [13], warranting further research. Further, it was interesting to note that, in spite of the known seasonality of infections in this environment [3], incidence of both maternal and infant morbidity did not vary by season (Table 1) and did not impact on the associations observed.
This study presents several strengths. Firstly, diphtheria, tetanus and pertussis antibody responses were measured in large groups of infants at 12 (n = 710) and 24 (n = 662) weeks of age. Therefore, 4 and 8 weeks after the first and third vaccinations, respectively, enabling the assessment of antibody responses within short periods following vaccinations. This minimised the effects of other potential factors which could influence vaccine responses in the intermediary period as done in some studies several years after the first initial vaccine challenge [26]. Secondly, the comparison by both seasons of mid-gestation and vaccination enabled the investigation of how seasonal exposures during pregnancy and early life may influence vaccine responses in infants. Finally, the analytical models presented were strengthened by the inclusion of carefully measured maternal and infant confounding factors.
Several limitations of this study should also be noted. This study by its observational design does not allow to draw causation but only to highlight associations between seasonally-driven factors and seasonal differences in infants’ vaccine responses. The lack of maternal vaccination status or infant pre-vaccination antibody titres are further important limitations, however, the randomisation procedure may mitigate the potential effects of these factors. Another limitation of this study is the interpretation of the effects of seasonality on immune responses measured after the three doses of the DTP vaccine, at 24 weeks of infant age, which as discussed may be concealed by the immunological effects of repeated challenges. Furthermore, at 24 weeks, infant antibody responses to DTP vaccine were measured 8 weeks following their last vaccination while they were measured 4 weeks after their first vaccination. This difference leaves a greater period for other factors such as antibody decay which may vary depending on the vaccine antigens, to mitigate and conceal the effects of seasonality on vaccine-induced antibody responses.