Safety and efficacy of neonatal vaccination
Newborns have an immature immune system that renders them at high risk for infection while simultaneously reducing responses to most vaccines, thereby posing challenges in protecting this vulnerable population. Nevertheless, certain vaccines, such as BCG and Hepatitis B vaccine, do demonstrate safety and some efficacy at birth, providing proof of principal that certain antigen–adjuvant combinations are able to elicit protective neonatal responses. Moreover, birth is a major point of healthcare contact globally meaning that effective neonatal vaccines achieve high population penetration. Given the potentially significant benefit of vaccinating at birth, availability of a broader range of more effective neonatal vaccines is an unmet medical need and a public health priority. This review focuses on safety and efficacy of neonatal vaccination in humans as well as recent research employing novel approaches to enhance the efficacy of neonatal vaccination.
Neonates and infants suffer a high frequency and severity of microbial infection resulting in millions of deaths worldwide 1. The same immune deficiencies that render newborns susceptible to infection also reduce their memory responses to most antigens, thereby potentially frustrating efforts to protect this high-risk population. As birth is the most reliable point of healthcare contact worldwide 1 and effective vaccination at birth would provide early protection for newborns and infants, expanding and improving the available means of neonatal vaccination is a global health priority.
Newborns have impaired immune responses due to a range of deficiencies in both adaptive immunity 2 and innate immunity 3 as well as the potentially suppressive effects of maternally derived Ab (MatAb) 4, 5. Newborns exhibit increased activity of suppressive Treg cells 6, 7 coupled with impairments in functional activity of APC 8, 9. Thus, study of neonatal vaccination is in part a quest for Ag–adjuvant (Aj) combinations that will be efficacious at birth. In addition, neonates and infants have a limited Ab repertoire and may produce suboptimal Ab in response to some Ag 10, 11.
This review summarizes clinical data on the safety and efficacy of human neonatal vaccination as well as translational studies aimed at developing novel approaches to effective neonatal vaccination. Throughout, our emphasis will be on safety and efficacy of approaches to neonatal vaccination, bearing in mind that basic aspects of neonatal immunity (with a specific focus on DC) are reviewed in an accompanying article by Willems et al.12.
Potential barriers to neonatal immunization
Concerns that have been raised regarding vaccination of neonates and infants include: (i) doubts about efficacy given the limited capacity of neonates to respond to many Ag and (ii) potential effects on immune system polarization, including potential for triggering autoimmunity via epitope mimicry or Aj effect 13, 14. From a theoretical perspective, these concerns are in part mitigated by: (i) the documented ability of newborns to respond to several vaccines including BCG and hepatitis B vaccine (HBV, as outlined below), which serves as proof of the concept that neonatal vaccination can be safe and effective and (ii) the presence of extensive immunologic mechanisms for central and peripheral tolerance that eliminates self-reactive T and B cells in newborns, coupled with (iii) evidence that multiple pediatric vaccines, including BCG, are not linked to allergy or autoimmunity 15. Nevertheless, despite these conceptual reassurances, novel vaccines, as any new drugs, do have the potential of inducing side effects and must certainly undergo rigorous and on-going safety analysis, including that provided in the US by the Vaccine Adverse Event Reporting System, a program of the US Food and Drug Administration and the Centers for Disease Control and Prevention (CDC). Indeed, safety concerns have prompted discontinuation and/or changes in some pediatric vaccines, with two examples discussed below.
In 1998, the measles–mumps–rubella (MMR) vaccine was the subject of controversy in the UK when Wakefield 16 reported on 12 children who developed symptoms of autism spectrum disorder soon after they had received MMR. The interpretation section of this study was later retracted in 2004 by ten of Wakefield's coauthors, and subsequent large studies concluded that there was no evidence of a link between MMR and autism 17. Early thiomersal exposure was also hypothesized to be associated with neuropsychological deficits in children, although this link was not supported in a study of 1047 children aged 7–10 years 18. Nevertheless, in 1999, the American Academy of Pediatrics and CDC requested removal of thiomersal from all pediatric vaccines, and this ethylmercury-containing preservative was no longer used in routine childhood vaccines in the US as of 2001. Although the autism link has been refuted, the need for stringent safety monitoring in the development of all vaccines remains, particularly those that may be given to newborns.
The live attenuated rotavirus vaccine RotaShield® (Wyeth-Ayerst) contained three rotavirus reassortants, with different genes encoding specific serotypes (VP4 or VP7) evoking virus-specific Ab, along with genes of Rhesus macaque-passaged rotavirus that attenuated virulence 19. After approval, 76 cases of intussusception, in which one segment of the bowel enfolds within another segment, causing obstruction, were reported to the Vaccine Adverse Event Reporting surveillance system; 70% of intussusception cases occurred after the first dose of vaccine. Due to this surveillance, the CDC recommended the suspension of the rotavirus vaccine until further studies could be performed. One study found one case in every 5000–9500 vaccinated infants, with the highest risk after the first dose. Due to the possible association with intussusception, RotaShield® was withdrawn from the market in 1999.
Inadequate immunogenicity of most vaccines at birth
Immunization in early life is a major public health imperative but remains a challenging field. The neonatal immunological milieu, skewed towards Th2 immunity to prevent recognition of the developing fetus as an allograft by the maternal immune system 20, represents an important obstacle that vaccination during neonatal period must overcome. In addition to the challenge posed by immaturity of the neonatal leukocyte compartment, effective neonatal vaccines must also overcome the potential inhibitory effect of MatAb 20. It is believed that inhibition of adaptive immune responses by MatAb depends on the ratio between MatAb titers and vaccine antigen dose and is due to determinant-specific masking of B-cell epitopes 21. Infant APC uptake and T-cell responses appear to be largely unaffected. For example, with respect to the Haemophilus influenzae type b (Hib)-conjugate vaccines, MatAb to the tetanus toxoid (TT) carrier protein inhibit infant responses to TT but do not inhibit Ab responses to the Hib polysaccharide moiety 22. Thus, MatAb result in specific masking of TT but not of Hib antigenic determinants to infant B cells, preserving APC uptake of MatAb:Ag immune complexes and allowing response to the Hib polysaccharide moiety. Overall, responses of human newborns to vaccines are not predictable from studies of older infants or adults. Nevertheless, several vaccines have been shown to elicit a clinically significant immunogenic response at birth, as reviewed below.
Of note, in assessing the potential efficacy of neonatal vaccines, although the prevention of infection is the ultimate goal and most important end-point, correlates of vaccine-induced immunity must be carefully considered, as recently reviewed by Plotkin 23. Both quantitative and qualitative (i.e. functional activity) of Ab can serve as “co-correlates” and surrogate markers for protection and are predominantly used in vaccine studies. Nevertheless, cell-mediated immunity is critical in protection against intracellular infections and, through the function of CD4+ cells, necessary to enhance B-cell development, as illustrated below in the case of BCG.
Early studies with whole-cell pertussis vaccine given alone or combined with diphtheria and tetanus vaccines within the first 24 h of life demonstrated safety, without any signs of erythema, infiltration, fever, irritability, vomiting or anorexia 24. However, pertussis immunization at birth resulted in serologically inadequate responses and blunting of booster responses to pertussis in 75% of study subjects until 5 months of age, suggestive of antigen-specific “immunologic paralysis” or tolerance induced by the immunization. This failure was believed to be independent of any effects of MatAb, as these were low or undetectable. In contrast, immunization at 3 wk of age resulted in adequate serologic response 24.
Purified polysaccharide vaccine (PRP), the first vaccine licensed to prevent Hib disease, was neither immunogenic in neonates nor consistently immunogenic in children older than 18 months 25. In contrast, the current Hib conjugate vaccine, diphtheria CRM 197 protein conjugate is given as a series of three injections starting at 2 months of age. Lieberman et al. 25 attempted to enhance Ab response to HbOC by administering the diphtheria–tetanus vaccine at birth, only to find that at 7 months, children exposed to at birth had a lower Ab response than those immunized beginning at 2 months of age. Impairment in neonatal Th1-cell response compared with that in adults may contribute to reduced neonatal responses to some vaccines. For example, after oral polio vaccination, young infants produce a relatively weak IFN-γ and cell-mediated response compared with adults, although they produce high titers of neutralizing Ab 26, thought to be essential for protective immunity against poliovirus 27.
In general, neonates mount impaired responses to T-independent polysaccharide antigens, and their Ab responses to T-dependent protein antigens are short-lived 5. Accordingly, the 23-valent Streptococcus pneumoniae polysaccharide vaccine (PPV23) is not immunogenic in children younger than 2 years 28. Although the pneumococcal protein–polysaccharide conjugate vaccine is safe and effective when administered to infants as a four-dose series (2, 4, 6 and ≥12 months), its efficacy at birth is unknown and currently under investigation 29.
Vaccines currently given at birth
Although newborns generally mount weaker responses than older persons to a wide range of vaccines, some vaccines do have a measure of efficacy when given at birth.
The BCG vaccine is a live attenuated Mycobacterium bovis vaccine administered within the first few days of life in most countries to prevent childhood tuberculous meningitis and miliary disease. With more than 3 billion people having received it, it is the most widely used vaccine worldwide 30. It is not generally recommended for use in the US due to a relatively low prevalence of tuberculosis and the variable effectiveness of immunization against adult pulmonary tuberculosis.
In general, the BCG vaccine exhibits an excellent safety profile. The main adverse events to vaccination are local reactions, including scarring (up to 92% of healthy neonates), pustule formation and drainage 31. These usually respond to conservative management. Axillary and cervical lymphadenopathies are the most common regional adverse effects and may persist for a few months, occasionally resulting in surgical drainage 32. Disseminated BCG infection is a rare complication, occurring in less than one per million individuals. It has been reported in children with congenital immune disorders, such as severe combined immunodeficiency, chronic granulomatous disease and the acquired immunodeficiency syndrome. About half of the cases of disseminated BCG infection in children are linked to rare immunodeficiencies of the IFN-γ and IL-12 pathways 33, including a report of fatal BCG infection in an infant with IFN-γ-receptor deficiency 34. A relatively high incidence of osteitis, osteomyelitis and disseminated BCG infection was noted upon use of the Danish 1331 BCG vaccine strain manufactured by Statens Serum Institut, involving several hundreds of children in Finland between 2000 and 2006. Retrospective analysis suggested that the increased reporting rate, although within the expected frequency of adverse reactions expected for the product, might be due to a combination of factors including heightened awareness surrounding use of the newly available BCG vaccine SSI following publicity associated with the withdrawal of the previously used product, the relatively higher potency/reactogenicity of the Danish 1331 strain and administration errors (incorrect dose or route of administration) 35, 36. The low and declining rate of tuberculosis in Finland prompted a change in vaccination policy in Finland from universal to risk-group targeting 37.
Studies of the efficacy of BCG vaccine have provided widely varying results. Efficacy has ranged from 0 to 80% in case–control studies using different BCG strains 32. This variability has been attributed to disparate exposure to environmental mycobacteria among study populations, strain variation in BCG preparations, genetic or nutritional differences and other environmental factors such as sunlight exposure and poor cold-chain maintenance 38. In a meta-analysis by Rodrigues et al. 39, a 75–86% protective effect was noted against miliary and meningeal tuberculosis. In measuring vaccine efficacy with relative risk or odds ratio for tuberculosis in vaccinated versus unvaccinated infants, the protective effect was 0.74 when estimated from four randomized controlled trials and 0.52 when estimated from nine case–control studies 40. In a meta-analysis of the effect of BCG vaccination on childhood tuberculosis meningitis and miliary tuberculosis worldwide, Trunz et al. estimated that the 100.5 million BCG vaccine doses given to neonates in 2002 prevented ∼30 000 cases of tuberculous meningitis and ∼11 500 cases of miliary disease during the first 5 years of life 41. The greatest beneficial BCG immunization was noted in regions where both the risk of tuberculosis and rates of vaccine coverage were highest, including Southeast Asia, sub-Saharan Africa and the western Pacific. Of note, the efficacy of neonatal BCG administration has been linked to its ability to effectively induce a Th1-polarized neonatal immune response 42. BCG also affects the immune response to unrelated Ag in early life, boosting both Th1- and Th2-type responses to other Ag (e.g. HBV and oral polio vaccine), probably through its influence on DC maturation 43. At the current cost of US$2–3 per dose, the global cost of BCG vaccination is approximately US$206 per year of healthy life gained.
Research directed at developing even more effective vaccines against tuberculosis continues, using two vaccination strategies. One strategy involves BCG priming at birth and introduces a booster dose to prolong immunity and protect the adult population. Heterologous boosting is also an option, employing one of the novel, more potent tuberculosis vaccines to replace BCG 44. Novel tuberculosis vaccines include live recombinant BCG vaccines, such as rBCG30, which express high amounts of Mycobacterium tuberculosis major secretory protein 45, modified vaccinia Ankara virus vaccine expressing protective AG 85A (MVA-85A) as well as Aj subunit vaccines, such as H1/IC31 given by parenteral delivery and H1/LTK63 by mucosal delivery 46. The DNA vaccine containing the heat shock protein 65 (hsp65) protein is a promising candidate both as a replacement for BCG and as a booster dose.
With over 2 billion individuals having serological evidence of HBV infection worldwide and suboptimal treatment provided by current antiviral therapy, primary prevention through immunization remains the most effective way of controlling the spread of HBV 47.
Safe and effective vaccines against HBV infection have been available since 1982. Three classes of vaccine are available, produced in plasma, yeast or mammalian cells. The vaccine prepared by concentrating and purifying plasma from hepatitis B surface antigen (HBsAg) carriers to produce subviral particles, although highly efficient and safe, is no longer used in most developed countries because of concerns for potential transmission of blood-borne infections. Yeast-derived recombinant HBV are produced by cloning the HBV S gene in yeast cells and contain thiomersal as a preservative. Mammalian cell-derived recombinant vaccine, in addition to the S antigen, contain either antigens from the pre-S2 region or both the pre-S1 and pre-S2 regions that assemble into a virus-like particle and produce an enhanced immunologic response 48. In 1991, the US CDC Advisory Committee on Immunization Practices recommended HBV vaccination for all infants, regardless of the HBsAg status of the mother 49. HBV is usually given as three intramuscular doses over a 6-month period, with the first dose given at birth. This vaccination schedule decreased the burden of HBV disease in the US, a protective effect also noted in many other countries. These guidelines were updated in 2005 to recommend implementation of universal vaccination of neonates before discharge from the hospital 50.
Adverse events to HBV are mild and most commonly include pain at the injection site (3–29%), mild fever >37.7°C (1–6%), malaise, headache, joint pain and myalgia. These effects were reported no more frequently among children receiving both HBV and diphteria/tetanus/whole-cell pertussis (DTP) vaccine than among children receiving the DTP vaccine alone. More serious adverse reactions have been described in the literature 49, but the strength of these associations remains unclear. The estimated incidence of anaphylaxis following HBV vaccination among children and adolescents is one case per 1.1 million vaccine doses 50. Although a retrospective case–control study suggested an association with multiple sclerosis in adults, and routine school-based vaccination was suspended in France in 1998, multiple sclerosis was not reported after immunization with HBV among children 50. Similarly, a possible association with Guillain–Barré syndrome that was proposed in adult recipients of the plasma-derived HBV was not confirmed 50.
Efficacy of the HBV is measured by its ability to induce hepatitis B surface Ab at a titer of >10 IU/L. In healthy infants, one dose provides ∼30–50% protection, two doses 50–75% protection and three doses >90% protection against HBV infection, thereby eliminating the need for booster doses 48. A remarkable degree of protection had been demonstrated in the 1980s, although this effect was not as extensive as that obtained when the vaccine was used in conjunction with passive immunization with multiple injections of hepatitis B immune globulin. Immunization was estimated to reduce the carrier state of infants born to HBsAg-positive carrier mothers by ∼90% 51, 52. Chang et al. have shown that universal vaccination in Taiwan was associated with >50% decline in the incidence of hepatocellular carcinoma in children 53.
Oral polio vaccine
Halsey and Galazka studied the efficacy of trivalent oral polio vaccine (TOPV) and DTP administered to human neonates 54. The authors noted that although MatAb may modify or block the serum immune response during the first few weeks of life, the first or priming dose of DTP could be given effectively by 4 wk of age. TOPV administered to infants during the first week of life resulted in intestinal infections and local immune responses in 50–100% of infants and induction of serum Ab in 30–70% of infants. By 4–8 wk of age, TOPV administration induced serum Ab response matching that induced in older infants. Although the WHO Program on Immunization recommended initiating DTP and TOPV schedules at 6 wk of age, the authors suggested considering administration of the first dose of TOPV at birth (or as close to birth as possible), for countries where poliomyelitis has not yet been controlled.
The severity of pertussis among young infants and the immunogenicity in newborn mice of acellular pertussis (aP), as opposed to the tolerogenicity of whole cell pertussis vaccine in human newborns 24, has prompted investigation of aP in human newborns. Knuf et al. compared aluminum-adjuvanted aP vaccine (containing pertussis toxoid, filamentous hemagglutinin and pertactin) or HBV given intramuscularly at 2–5 days of age followed by diphtheria-tetanus-acellular pertussis (DTaP)–HBV–inactivated poliovirus vaccine (IPV)/Hib at 2, 4 and 6 months 55. This study demonstrated that neonatal aP vaccination was safe (no significant differences in reactogenicity between groups), induced higher Ab responses to pertussis Ag by 3 months (i.e. did not induce immunologic tolerance) and resulted in earlier Ab responses to DTaP but did dampen Ab response to Hib and HBV. The authors speculate that the dampening of responses to Hib and HBV was due to strong secondary T-lymphocyte-specific pertussis responses after the first dose of DTaP–IPV–HBV/Hib potentially interfering with CD4+ T-cell help, a phenomenon known as “bystander interference”. Given that the risk of death due to pertussis infection is diminished by the first infant dose of aP given at the currently standard time-point of 2 months of age 56, the authors speculate that a birth dose would further reduce the risks of pertussis-related deaths during the current early window of vulnerability.
Vaccines given in infancy
Vaccines given early in life, during infancy but after the neonatal phase, include Rotavirus, DTaP at 2, 4, 6 and 15–18 months, Hib at 2, 4, 6 and 12–15 months, pneumococcal conjugate vaccine at 2, 4, 6 and 12–15 months, IPV at 2, 4 and 6–18 months, influenza (yearly from 6 months to 18 years), MMR vaccine (12 months), Varicella (12 months) and Hepatitis A (12–18 months) 57. Although a complete discussion of the safety and efficacy of all infant vaccines is beyond the scope of this review, rotavirus vaccine will be discussed as illustrative of safety and efficacy studies in vaccinating the very young.
Rotavirus vaccine is the most recent addition to the panel of immunizations in early life and has been recently reviewed by the WHO Weekly Epidemiological Record 58 as well as by Dennehy 19. Protection against rotavirus infection is of major clinical interest, as it is the leading cause of severe diarrhea in children less than 5 years globally, with over 25 million outpatient visits and over 2 million hospitalizations yearly. Licensed in 2006, two live attenuated oral rotavirus vaccines, monovalent human rotavirus vaccine Rotarix® and the pentavalent bovine-human vaccine RotaTeq, replaced their counterpart RotaShield®, which was withdrawn from the market in 1999 because of a possible association with intussusception. The two new vaccines have a similar safety and efficacy profile but a different immunization schedule: Rotarix® is administered in a two-dose schedule between 6 and 12 wk (at least 4 wk apart) and RotaTeq as three doses at 2, 4 and 6 months (first dose between 6 and 12 wk and subsequent doses at 4–10 wk intervals, with the first dose given no later than 12 wk and the third dose given before the age of 32 wk). The first dose of these vaccines should not be given to infants older than 12 wk, as the safety has not been established, and this confers a potentially higher risk of intussusception. According to the Global Advisory Committee on Vaccine Safety and their data on post-licensure surveillance until June 2007, the use of these vaccines was not associated with an increased risk of intussusception or other serious adverse events 58. Rare complications included mild and transient symptoms from the respiratory or gastrointestinal tract. The vaccines are contraindicated in infants with a history of intussusception or anatomical malformations possibly predisposing to intussusception. Of note, neither of these vaccines contains thiomersal.
These rotavirus vaccines provide 74–85% protection against rotavirus diarrhea of any severity and ∼90–100% protection against severe rotavirus disease that extends to the second year of followup. Both vaccine dose and host factors (e.g. MatAb, interfering bacterial and viral agents, and malnutrition) are believed to determine the extent of the immune response. Although optimal surrogate markers for vaccine efficacy have yet to be clearly defined, intestinal virus-specific IgA has correlated with protection and serum IgA responses to the VP4 and VP7 surface structural proteins have been used as end-points, though cell-mediated immunity is believed to contribute to antirotaviral defense as well 59. As clinical efficacy has thus far been demonstrated mainly in the US, Europe and Latin America, WHO has not yet recommended global inclusion of rotavirus vaccines into national immunization programs until its potential is confirmed in all regions of the world.
Clinical studies of novel early-life vaccines
Malaria is a leading global health problem against which no effective vaccine has yet been introduced in clinical practice. The RTS,S/AS02D candidate malaria vaccine was found to be safe, well tolerated and immunogenic in infants up to 18 wk-old living in the highly endemic area of Mozambique 60. It is a hybrid recombinant protein consisting of tandem repeats from a Plasmodium falciparum protein and the S antigen of HBV, formulated with the Aj system AS02 (a mixture of the TLR agonist monophosphoryl lipid A (the active moiety of lipopolysaccharide/endotoxin) and the detergent saponin QS21) 60, 61. Candidate HIV vaccines capable of generating robust immunologic responses in breastfeeding infants are also being developed 62. Other novel early-life vaccines currently being studied include vaccines against Salmonella typhi, RSV, influenza and parainfluenza. Additional studies are assessing co-administration at birth of HBV in combination with hepatitis A or BCG, which may modify responses to other vaccines 43.
Need for novel approaches to enhance neonatal vaccination
The ability of certain vaccines such as BCG and HBV to exhibit some efficacy at birth provides proof of concept that despite generally impaired APC function and Th1 responses, neonatal vaccination is possible. The medical advantages inherent to neonatal vaccines effective at birth include: (i) early protection that would close the window of vulnerability inherent to vaccination schedules that start later in life (e.g. 2 months), (ii) the practicality of birth being a global point of contact with healthcare systems and (iii) potential advantages of novel vaccines that may require fewer doses to achieve efficacy. In this context, we review recent approaches to the development of neonatal animal models and recent in vitro work with human neonatal cells.
Animal models of neonatal vaccination
Applicability of neonatal animal vaccination models to humans
In assessing the relevance of animal studies to humans, it is important to recognize that mammalian species vary in the type of placentation and relative placental and colostral transfer of Ig to the fetus/newborn 63. For example, pigs, horses and ruminants have either epitheliochorial or syndesmochorial placentation and no placental Ig transfer, relying very heavily on colostral transfer 64. In contrast, rodents and primates have hemendothelial and hemochorial placentation, respectively, and both rely heavily on placental transfer with lesser colostral transfer. In general, species that allow early (placental) transfer of maternal Ig (e.g. mice and humans) demonstrate a slower rate of immune maturation. Of interest, B-cell and Ab repertoire development in rabbits requires gut-associated lymphoid tissues 65.
Another aspect to consider in interpreting animal models is the relatively high divergence of the innate immune system. For example, the innate immune system of mice is particularly divergent from that of humans 66. Thus, although murine models are absolutely critical for immunologic research and provide powerful insights, results in mice do not always translate directly to humans.
Finally, the timing of vaccine administration is also an important and, at times, controversial aspect of neonatal animal vaccination models. In particular, multiple studies have focused on mice that are 1 wk of age to model neonatal responses 2. However, given the importance of developing vaccines active on the first day of life, and growing evidence of distinct perinatal physiology at birth, including high levels of immunosuppressive adenosine at birth 67, it will be important to also study vaccination of animals in the first day of life.
Safety and efficacy of neonatal vaccination in animal models
Multiple studies have documented that certain vaccines are apparently safe and effective when administered in utero or to newborn animals. Although serious side effects due to vaccination of neonatal animals are generally rare, passive surveillance in the UK of dog vaccinations has demonstrated a relatively high prevalence of vaccine-associated adverse effects in very young animals 68. The most common adverse event appears to be facial edema and pruritis, believed due to immediate (type I) hypersensitivity reaction triggered by degranulation of mast cells sensitized by maternal IgE. These potential adverse effects may be secondary to high bovine serum albumin content in canine vaccines and are most prevalent in small breed dogs, suggesting that dose reduction may be in order. Alum- or lipid-adjuvanted vaccines induce greater tissue inflammation than non-adjuvanted vaccines after s.c. administration in 14–16 wk-old kittens 69.
Examples of efficacy of neonatal vaccination in animal models include avian studies of the live herpes virus of turkeys vaccine, aimed at preventing the α-herpesvirus neoplastic Marek's disease of chickens, demonstrate protection even when administered in ovo or at day 1 70. Beagle puppies have been vaccinated s.c. with modified live canine parvovirus at 1 day of age 63. Both kinetics and magnitude of Ab response were similar to those of older puppies. In this model, vaccination after colostral ingestion or of puppies of convalescent dams with high anticanine parvovirus titers was unsuccessful, illustrating the potential inhibitory role of maternal Ab. However, under certain circumstances the hurdle of maternal Ab can be overcome. Puppies born to dams boosted during pregnancy with killed adjuvanted rabies vaccine and who received colostral immunity nevertheless mounted protective Ab responses after immunization with RABISIN vaccine comprising rabies virus glycoproteins and aluminum hydroxide Aj 63. The authors speculate that either greater antigenic content and/or vector properties may allow more efficient Ag presentation. Thus, under certain conditions murine and human neonates can mount effective adaptive immune responses. Although these studies do not define the mechanisms by which the vaccine studied overcame impairments in neonatal immunity, they do illustrate the possibility of effective vaccination at birth.
Novel approaches to enhancing efficacy of neonatal vaccines
Multiple novel approaches are being explored in an effort to overcome deficiencies in neonatal immune responses and thereby allow effective neonatal vaccination 5. We provide examples of such approaches below, selecting recent examples from the published literature.
Intracytoplasmic delivery of antigens
Several murine studies suggest that a key requirement for induction of effective neonatal adaptive response is entrance of Ag into the cytoplasm of APC. Chen et al.71 studied adult and neonatal (1 wk old) BALB/c mice immunized i.p. with inactivated split-product influenza vaccine followed by a booster dose after 3 wk or with intramuscular injection and in vivo electroporation of plasmid DNA. Vaccination of neonates with hemagglutinin or neuraminidase DNA protected mice against influenza infection in the presence of MatAb. The authors concluded that in order to overcome potential inhibition of adaptive immune responses by MatAb, mothers and their offspring should be immunized with different influenza vaccines targeting distinct Ag (e.g. inactivated vaccine versus DNA vaccine or use of DNA vaccines targeting different influenza products). If the same Ag is to be used, a study by Pertmer suggests that maternal Ab do not blunt DNA vaccine-based responses to intracellularly expressed Ag 72.
Study of neonatal C57BL/6 and BALB/c mice immunized (i.p. and s.c.) within 24 h of birth with disabled infectious single cycle HSV-1 variant reveals that a single round of viral replication dramatically enhances protective responses 73. CD4+ and CD8+ T cells from neonatally vaccinated mice transferred to naïve recipients conferred protection against lethal viral challenge. UV-inactivated viral particles at up to 104-fold higher doses were not able to achieve this response, suggesting that cytoplasmic delivery of Ag can enhance neonatal immune responses.
Kollmann et al.74 have demonstrated a novel approach to neonatal vaccination, employing an attenuated strain of the intracellular pathogenic bacterium Listeria monocytogenes to deliver Ag to the cytoplasm of APC. Importantly, this approach appeared to be safe in neonatal mice in that they survived high-dose infection with the ΔactA strain of L. monocytogenes without any sign of disease or any recoverable bacteria in spleen or liver 7 days post-vaccination. Neonatal mice vaccinated a single time with attenuated L. monocytogenes strain ΔactA mounted strong CD8+ and CD4+ T-cell responses and were protected against subsequent challenge with wild-type L. monocytogenes. Moreover, ΔactA served as an effective vehicle for delivery of heterologous Ag resulting in a strong CD8 and CD4 Th1-type memory response, suggesting that this strain may serve as an effective vaccine vehicle for neonatal immunization. Of note, recombinant attenuated strains of L. monocytogenes induce specific immunity even in the presence of pre-existing immunity, potentially overcoming the hurdle of pre-existing maternal immunity that might interfere with neonatal vaccine responses 75.
Pelizon et al.76 vaccinated 5-day-old neonatal BALB/c mice by the intramuscular route with a cytomegalovirus intron-based plasmid containing an inserted fragment encoding the Mycobacterium leprae hsp65. pVAXhsp65 was transcribed at 2–7 days in the muscle tissue of newborn mice; 15 days after the last of a three-series dose (5, 12 and 19 days of age), an increased ConA-induced spleenic production of Th2-polarizing cytokines (IL-4 and IL-5) and inconsistent increases in anti-hsp65 IgG1 and IgG2a serum levels were noted. pVAXhsp65 appeared to be safe, in that Southern blot analysis did not reveal an evidence of integration in a range of organs, including spleen, liver, thymus, and regional lymph nodes. Moreover, similar to BCG, pVAXhsp65 when given as a single dose, was able to prime 5-day-old mice for a mixed Th1 and Th2 immune response to pVAXhsp65 boosting later during adulthood.
DNA-based vaccines have also shown promise in the effort to protect newborns against malaria. Neonatal BALB/c mice (7 days) were immunized with a Plasmodium yoelii circumsporozoite protein DNA vaccine mixed with a plasmid expressing murine granulocyte macrophage-colony stimulating factor then boosted at 28 days with pox virus expressing P. yoelii circumsporozoite protein 77. Immunized neonates, including those born to immune mothers, were noted to mount CD8+ T-cell-mediated protection similar to adults.
A measles virus (MV) DNA vaccine consisting of measles H, F and N genes was administered via the intradermal route with an IL-2 Aj to neonatal Rhesus macaques (4–5 days) that had received passive immunization with measles immunoglobulin (to mimic the presence of MatAb) 78. All macaques were boosted with the same regimen at 2 months after vaccination. Although it did not enhance MV-induced Ab responses, MV DNA vaccine did prime MV-specific T-cell responses as measured by MV-induced IFN-γ production by PBMC. Moreover, MV vaccine protected infant R. macaques from subsequent MV challenge-induced rash and immunosuppression. Overall, DNA-based immunization represents a viable option in developing novel neonatal vaccines.
Intranasal administration of live attenuated vaccines
Mucosally delivered live attenuated Salmonella enterica vector vaccines have been studied as a platform to deliver the model antigen tetanus toxin fragment C in neonatal mice immunized by the intranasal route at days 7 and 22 of life 79. Salmonella live vectors colonized and persisted primarily in nasal tissue and induced high (adult level) titers of Frag C-specific Ab, mucosal and systemic IgA- and IgG-secreting cells, T-cell proliferative responses, and IFN-γ secretion; 1 wk after the boost, a long-term mixed Th1- and Th2-type response to Frag C was established. Such effects were evident even in the presence of high levels of maternal Ab.
Mielcarek et al.80 have developed a live attenuated strain of Bordetella pertussis, the causative agent of whooping cough. Attenuated by deletion of genes encoding tracheal cytotoxin, pertussis toxin and dermo-necrotic toxin, the strain BPZE1 was given to infant (3 wk old) and adult BALB/c mice as a single intranasal dose. This attenuated intranasal vaccine induced stronger neonatal anti-Bordetella IgG responses than the aP vaccine, demonstrating sterilizing immunity to subsequent intranasal challenge with B. pertussis and Bordetella parapertussis. BPZE1 induced a reduced Th2-polarized response as measured by antifilamentous hemagglutinin IgG1/IgG21 ratio. The authors speculate that BPZE1 could also represent a platform for delivery of heterologous antigens. This study focused on a 3-wk-old infant mouse model, and no data were provided about potential efficacy of this attenuated strain in newborn mice. However, a different study of intranasal administration of live Bordetella bronchiseptica to 2-day-old neonatal piglets demonstrated efficacy against subsequent atrophic rhinitis challenge 81, suggesting that live attenuated Bordetella strains may induce effective immunity in newborns upon intranasal administration.
Recent studies have explored intranasal administration of Escherichia coli-expressed rotavirus VP6 protein and the Aj E. coli labile toxin (LT-R192G) to neonatal (7 days old) and adult mice, and protection against fecal rotavirus shedding following challenge with the murine rotavirus strain EDIM 82. In contrast to adult mice that developed both CD8+ T-cell responses (rotavirus-inducible, Th1-cytokine producing splenocytes) and Ab within 10 days, neonatal mice did not show protection until 28 days, at which point they possessed memory rotavirus-specific T cells, but did not produce antirotavirus Ab. These studies highlight the potential of intranasal immunization of newborns with live vaccines.
Impaired responses of neonatal APC to many stimuli are a key hurdle to overcome in developing effective neonatal vaccines 12. One approach to overcoming deficits in neonatal APC is to exogenously administer co-stimulatory signals whose endogenous production is deficient, such as IL-12 83. Co-administration of IL-12 and influenza subunit vaccine within 24 h of birth elevated splenic expression of IFN-γ, IL-10 and IL-15 mRNA and the protective efficacy of antiviral vaccination 84. In addition, IL-12 co-administration also increased IFN-γ-, IL-2- and IL-4-secreting cells, and IgG2a Ab levels and enhanced survival in a B-cell-dependent manner after adult lethal challenge with infectious influenza virus.
Discovery of novel innate immune pathways and agonists that engage them has opened the door to assessment of novel vaccine Aj. Activation of TLR, transmembrane proteins that mediate recognition of microbial products, activates APC, including enhancement of DC maturation. Therefore, TLR agonists represent potential vaccine Aj, several of which (e.g. lipid A that signals through TLR4) are currently in clinical use 85, 86. The synthetic dsRNA polyriboinosinic:polyribocytidylic acid is a TLR3 agonist that induces type I/II IFN production and enhances primary anti-TT immune response of neonatal mice, increasing production of anti-TT IgG1, IgG2a and IgG2b isotypes 87. Enhancement of the secondary anti-TT IgG response was noted when polyriboinosinic:polyribocytidylic acid was combined with retinoic acid/Vitamin A, a combined immunological/nutritional intervention that represented an effective vaccine Aj in neonatal mice. CpG oligonucleotides that activate TLR9 have also been shown to enhance neonatal Th1 responses in neonatal murine models 88, although they appear to induce relatively weak responses in human newborn cord blood plasmacytoid DC tested in vitro89.
With respect to in vitro studies of human neonatal APC, TLR8 agonists, including certain synthetic imidazoquinolines and single-stranded viral RNA, are particularly effective at activating human neonatal APC in vitro, correlating with strong activation of the p38 MAPK and NF-κB signaling pathways 90, 91. TLR8 (and TLR7/8) agonists were remarkably more effective in inducing production of TNF and IL-12 p40/70 as well as enhancing up-regulation of the co-stimulatory molecule CD40. In addition to their ability to effectively activate APC, TLR8 agonists may also contribute to enhancing neonatal adaptive immune responses by their ability to reverse the inhibitory effects of Treg cells that suppress adaptive immune responses 92, and that are particularly potent and abundant at birth 6, 7. Importantly, TLR7/8 agonist R-848 is an effective vaccine Aj when covalently linked to HIV Gag protein in a Rhesus macaque model in vivo93, suggesting that these promising Aj merit further study, including assessment in neonatal animal models wherein transient and selective local amplification of APC and Th1-function including reversal of Treg function might safely and effectively enhance local neonatal adaptive immune responses to vaccines without effecting overall central and peripheral tolerance 13 or the systemic skewing of responses against Th1 2, 3. There are thus theoretical grounds that coupled with emerging evidence of the apparent safety of this approach in adult non-human primates (R. macaques) in vivo93, 94 and the efficacy of these agonists towards human neonatal APC in vitro90 suggest that such an approach might be both safe and efficacious 91. Nevertheless, as with all novel drug development, all novel neonatal vaccines will need to undergo rigorous safety evaluation, to ensure that doses and routes of administration avoid any harmful side effects, including potentially over-exuberant inflammatory responses/reactogenicity 95 or risk of autoimmunity 14.
Worldwide infectious diseases cause death of more than 2 million newborns and infants less than 6 months of age. Significant reduction of this burden will require development of early-life vaccination, including vaccines effective when given at birth, the most reliable point of global healthcare contact. Based on animal and human studies, neonatal vaccination is feasible but requires strong immune signals such as those provided by in vivo replication of attenuated agents, and perhaps by certain Aj. Advances in manipulating attenuated microbial strains and recent characterization of innate immune recognition pathways provide opportunities for developing novel delivery systems and/or Aj to meet this crucial challenge. Safety considerations will be paramount, but the large burden of early-life infections coupled with the practicality of immunizing at birth provide strong motivation to pursue effective neonatal vaccines.
Research by O.L. is supported by NIH grant RO1 AI067353-01A1. We acknowledge the mentorship and support of Drs. Michael Wessels, Richard Malley and Raif Geha.
Conflict of interest: O.L. has received research support from 3M Pharmaceuticals, Dynavax, and Idera Pharmaceuticals, companies that develop TLR agonists as vaccine adjuvants.