Determinants of early life immune responses to RSV infection

Abstract

Respiratory syncytial virus causes significant morbidity and mortality in both developed and developing countries, and a vaccine that adequately protects from severe disease remains an important unmet need. RSV disease has an inordinate impact on the very young, and the physical and immunological immaturity of early life complicates vaccine design. Defining and targeting the functional capacities of early life immune responses and controlling responses during primary antigen exposure with selected vaccine delivery approaches will be important for protecting infants by active immunization. Alternatively, vaccination of older children and pregnant mothers may ameliorate disease burden indirectly until infants reach about six months of age,when they can generate more effective anti-RSV immune responses.

Introduction

Respiratory syncytial virus (RSV) is a globally ubiquitous pathogen that commonly results in upper respiratory infection. It has a highly infectious nature, resulting in infection ofup to 70% of children in their first year of life, and almost all children by two years of age [1]. A large proportion of primary infection events result in lower respiratory tract infection (LRTI), resulting in hospitalization for approximately 2-3% of infected infants. LRTIcan be manifested as bronchiolitis, pneumonia, and wheezing and can ultimately result in respiratory failure and death. Most disease and death due to RSV-associated LRTI occur in resource-poor, developing countries. Globally, RSV is estimated to cause about 33.8 million new cases of LRTI in children under five[2]. Recent meta-analysis of global causes of death in children in the first year of life highlights the considerable role of LRTI and places RSV as the second most-likely single pathogen to cause death in children less than 1 year of age[3]. These recent epidemiological analyses illustrate the health and economic toll of RSV, andimplicate it as one of the most importantpathogens of early childhood.

Prematurity, immunosuppression, and genetic or congenital factors are among many risk factors known to predispose infants to RSV-associated LRTI, but the majority of children who develop a LRTIrequiring hospitalization do not have an identifiable risk factor other than young age[4]. Several key aspects of early life underlie the susceptibility of young infants to LRTI following RSV infection. Physical and immunological immaturity cause infants younger than three months to be the most prone to severe disease. Airway size, in particular, is a problem in neonates and young infants because obstruction from sloughed epithelium, mucus, fibrin, and inflammatory debris is a major factor in disease pathogenesis. The months following birth are a period of dramatic development of the lung architecture and cellular functions[5,6], including the awakening and education of immunological effector mechanisms that must learn to deal with a world laden with both commensal and pathogenic microorganisms[7].

Recent scientific advances have made the prospects of significantly diminishing the morbidity and mortality caused by RSV through vaccination a foreseeable possibility, and may lead to a change in the current landscape of RSV [8]. This is particularly significant for the neonate's evolving immunological repertoire because for many infants, RSV may be the first pathogen encountered in life. Therefore, RSV is one of the first tests of immune defenses and a significant priming event for future adaptive responses. Here, we discuss some recent findings on the limitations of early life immune responses to natural infection and vaccination, the likely requirements to counter these limitations,and the prospects for protecting infants from severe disease by either direct or indirect mechanisms.

Infant innate responses to natural RSV infection

Infants with no prior exposure to RSV must initially rely solely on innate defenses. These early defenses must combat the suppressive mechanisms of early life. The neonatal immune environment can be considered tolerance-inducingas a mechanism toprotect from the loss of maternal or self-tolerance and potential for immunopathological responses to developing organs. It has long been recognized that infants suffer disproportionately from infectious disease due to this suppressive environment and the qualitative and quantitative deficiencies of newly generatedimmune cells. More recently, these deficiencies are recognized as deficiencies by design, as studies in the ontogeny of early life immunity paint a picture of specialized early life immune function[7,9-12]. Early life immunity has been found to display aspects of layering, where cells derived from fetal hematopoietic stem-progenitor cells (HSPC) develop in uteroand can exert age-specific, innate-like, or regulatory functions and cells derived from “adult” HSPC following parturition possess distinct functional profiles. A comparative microarray analysis of human fetal and adult naïve CD4+ T cells demonstrated that while phenotypically similar, these subsets are functionally distinct. Fetal CD4+ T cells are biased toward immune tolerance and are more likely to differentiate into regulatory T cells (Tregs) after stimulation [13]. In this way, early life immunity may be considered distinct and specialized in the needs of early life rather than deficient.

Immunomodulation in early life is due to both soluble and cellular mechanisms [7,14]. Immunosuppressive plasma factors such as adenosine and anti-inflammatory cytokines dampen inflammatory responses, and transforming growth factor beta (TGFβ) has been found to mediate suppression of natural killer (NK) cell function. Tregs, one potential source of TGFβ, have been found in higher numbers in newborns and preterm infants and possess immunosuppressive functions during early life [15,16]. Another cellular source of immunosuppression, CD71+ erythroid cells, are prevalent during early life and may compromise host defenses [17]. Neonatal neutrophils produce potentially anti-inflammatory and Th2-promoting IL-6, yet have dampened phagocytic and killing mechanisms [7]. Further work is necessary to understand the integrated contribution of these many immunomodulatorymechanisms onthe inability of infants to effectively combat respiratory infections,particularly RSV.

Dendritic cells (DCs), are critical orchestrators of both innate and adaptive immunity. DCs are specialized in recognizing and integrating signals conferred by pattern recognition receptors (PRRs), and RSV is recognized by toll-like receptor (TLR)4, TLR3, TLR2/6, TLR7/8 and retinoic acid-inducible gene I (RIG-I)[4]. Neonates have responses distinct from those of adults following stimulation by most TLR agonists, and TLR agonists as well as other adjuvants have been found to have age-dependent immune activity [18]. Studies of dendritic cells in the lungs of human infants is hampered by both ethical and technical hurdles. While there are several animal models used to study RSV[19-22], none completely recapitulate disease pathogenesis in infants. Additionally, most studies focus on infection of adult animals, which does not address the unique context of the infant respiratory tract and immune system. Neonatal mice have been used to help define basic immunological mechanisms underlying early life immune responses to RSV [23]. Limitations in plasmacytoid dendritic cell (PDC) function and type I interferon production in neonatal mice was shown to be one mechanism promoting Th2-biased immunopathology following RSV infection [24]. Studies in human neonates and very young children also demonstrated that early-life PDCs have limited type I interferon following exposure to RSV [25]. A thorough understanding of the limitations of dendritic cells in the lung must include local lung-migratory DC populations, namely CD103+ and CD11b+ DC subsets, known to induce adaptive T cell responses. We have described a dramatic early-life transition in the response of these two populations of dendritic cells following RSV infection in neonatal mice. In addition to the rapidly changing composition of these important dendritic cells subsets in the lung and mediastinal lymph node, both CD103+ and CD11b+ dendritic cells of neonatal mice are less efficient at antigen uptake and processing, and have lower costimulatory molecule expression than their adult counterparts [26]. These, and other limitations of early life DCs can affect their ability to induce adaptive immune responses.

For a more comprehensive examination of the ontogeny of early life innate immune responses and innate responses to RSV infection in infants, please see several recently published in depth reviews covering these topics[4,7,9,27,28].

Infant adaptive responses to natural RSV infection

The generation of effective anti-viral adaptive immunity relies on instruction by the innate immune response (elegantly reviewed in [29]). Antigen presenting cells (APCs), in particular, exert exquisite control, dictating both the induction and the downstream function of adaptive effector cells. Early life deficiencies in innate immunity, particularly in APCs, translate into altered or inferior T and B cell immunity. As one example of this, we have shown that the previously-described early life changes in the dynamics and function of lung-migratory DCs is directly linked to the distinct CD8+ T cell response established in neonatal mice. In particular, the decreased ability of neonatal DCs to provide co-stimulatory support alters the immunodominance hierarchy of responding T cells[26,30]. Every difference in early life DCfunction, such as how they take up and process antigens or produce cytokines, may affect adaptive immunity.

While CD8+ T cells have been shown to contribute to disease in mouse models, they have not been shown to be detrimental in RSV-infected infants. In fact, they have been found to coincide with convalescence in infants with severe disease [31] and a paucity of CD8+ T cells has been noted in a study involving autopsies of infants that died of severe disease [32]. CD4+ T cell responses in young infants undergoing disease have demonstrated a Th1 to Th2 imbalance, and Th2-driven pathology was associated with enhanced disease following formalin-inactivatedRSV (FI-RSV) immunization and may be a factor in airway reactivity and wheezing common in infants with severe RSV disease [4,8]. Th17 responses to RSV have also been documented [33], another potential result of the inability of dendritic cells to strongly polarize toward Th1 following RSV infection [4].

Neutralizing antibodies can effectively protect from severe RSV disease as demonstrated by the efficacy of Synagis, a monoclonal antibody (mAb) administered to premature and other high-risk infants. However, infants are known to have significantly diminished antibody responses compared to adults and several intrinsic and extrinsic B cell limitations have been described [7,34,35]. For the first several months of life, B cells have a limited ability to promote somatic mutation required for affinity maturation. A recent study on the development of RSV-specific antibody following natural infection in infancy demonstrated that RSV-neutralizing antibody was unlikely to develop in infants under four months of age. Infants began to show significant increases in neutralizing capacity after the age of four months, with a dramatic increase in infants infected after six months of age. These data suggest that vaccines that mimic natural infection may be ineffective if administered prior to four months of age[36]. A more detailed summaryof T and B cell responses to natural RSV infection, and their contribution to disease, is provided in recent reviews [4,27,37].

Recent studies in nonhuman primates have demonstrated how regulatory and suppressive mechanisms canimpact both the generation and function of adaptive immunity in the lung. Holbrook et al. studied antibody responses in infant African green monkeys (AGM) infected with influenza [38]. They found similar systemic influenza-specific IgG responses in infants and adults, yet infants had lower influenza-specific IgG and demonstrated less neutralizing activity in upper respiratory tract secretions. This was associated with higher viral titers and increased pulmonary damage in infants. Interestingly, while adult AGM showed robust generation of bronchus associated lymphoid tissue (BALT) following infection, infants demonstrated a near absence of these tertiary lymphoid tissues. This may be due to increased levels of Tregs, which may inhibit BALT formation, in the lungs of AGM infants[38]. Further studies on the dynamics of immune responses in the lung specifically will help elucidate potential mechanisms of suppression, and may suggest approaches for generating local immune responses to effectively counter respiratory pathogens.

Prospectsfor direct protection of the vulnerable infant population

A failed vaccine trial nearly 50 years ago has given the vaccinationof infants against RSVan unfortunate legacy, and many are wary of the potential to enhance disease in infant populations using protein-based or inactivated virus vaccines [39,40]. FI-RSVvaccine administered in the 1960s was an adaptation of an approach successfully usedforpreventing disease from pathogens such as polio, but in the case of RSV resulted in enhanced disease after natural infection and the deaths of two children [41]. FI-RSV elicited an abundance of undesired responses not predicted by previous experiences with formalin-inactivated vaccines. Antibody with low neutralizing activity was induced and associated with immune complex deposition. Another prominent feature of FI-RSV immunization was the presence of eosinophilia [41,42] and animal studies indicate that priming with FI-RSV inducesCD4+ T cells biased toward Th2 with low induction of CD8+ T cells [8].

Since the 1960s, relatively few attempts have been made to vaccinate infants under six months of age against RSV. Trials of live-attenuated virus vaccine candidates, known not to be associated with the FI-RSV vaccine-enhanced disease, have caused congestion and other side-effects that have so far been unacceptable for use in infants and it has proven difficult to strike the balance between attenuation and immunogenicity[39]. Several newer candidates for active vaccination of infantsare in the pipeline, and will be thoroughly evaluated for safety, generation of protective responses, and establishment of durable immunity in infantsif regulatory challenges can be overcome. These vaccines must also generate protective responses despite potential interference from maternal antibodies[43].

As the current infant immunization schedule suggests, immune responses to effectively counter disease can be elicited in the young by repeated immunizations. Nearly all infant vaccines, including thehepatitis B vaccine first administered at birth, rely on repeated exposure to generate sufficiently protective memory responses. BCG, a potent immune stimulator given to infants outside of the US, can stimulate immune responses in infants similar to those of adults after a single dose [34]. The ability of infants to respond to vaccination appears to depend on nature and strength of the stimulus, with the responses to different vaccines or vaccine components in infants approaching those of adults at different ages[18,34]. Immunosuppressive features of infant immunity, as well as the inherent ability of the lung itself to negatively regulate effector cell function [34], make the choice of antigen and the context in which it is administered critical choices in the design of infant vaccines against RSV.

A better understanding of RSV structure has improved our ability to tailor vaccine antigens to elicit protective responses in young infants. Recent studies have solved the structure of the pre-fusion conformation of the RSV F trimer (pre-F)[44]. Antibodies have led to the definition of at least three sites unique to pre-F surfaces; D25, AM22 and 5C4 mAbs to site Ø[44,45], MPE8mAb to a site common to RSV and human metapneumovirus[46], and AM14 to a quaternary epitope on the pre-F trimer[47]. Antibodies to pre-F-specific surfaces have much higher neutralization potency than antibodies to the surfaces shared between the pre-and post-fusion (post-F) forms like palivizumab (site II) or 101F (site IV). More than 90% of neutralizing activity in samples from individuals ranging from 7-93 years of age can be adsorbed by pre-F, whereas only about 20% is adsorbed by post-F [48]. Delivery of pre-F as a subunit protein, virus-like particle (VLP), chimeric virus, or other gene-based approaches may facilitate the induction of these potently neutralizing antibodies in young infants [49].

Equally as important as antigen choice, the use of adjuvants in infants should be based on age-specific information about PRR responsiveness[18]. For induction of robust T and B cell responses and high-affinity antibodies, the inclusion of TLR agonists is one attractive approach[34]. Targeting DCs, particularly DC populations located in the lung, is critical to promoting the generation of more mature adaptive responses (Figure 1). We have found that TLR3, TLR4 and TLR9 agonists administered at the time of RSV infection promote the generation of adult-like epitope hierarchy in neonatally infected mice, as well as promote the maturation of lung-migratory CD103+ and CD11b+ DC subsets(Malloy, A. et al., abstract 1354, IDWeek, Philadelphia, PA, October 2014).The recently demonstrated efficacy of the RTS,S/ASO1 malaria vaccine in young infants, which was adjuvanted with the TLR4 agonist monophosphoryl lipid A and the saponin QS21 formulated in liposomes, shows promise for combining TLR agonists with non-TLR adjuvants [50]. Engagement of multiple TLRs has been shown to have synergistic effects on neonatal dendritic cells, and may be the key to the efficacy of BCG in neonates, which simultaneously stimulates five TLRs [34]. Little is known about the function of follicular dendritic cells (FDC) in early-life. FDC express an array of TLRs [51], and TLR ligation could potentiallypromote FDC maturation, help organize and maintain germinal centers, and support the development of high-affinity antibodies.Despite technical hurdles, this is an important area for future investigation. For an infant vaccine, special consideration of the potential of TLR agonists to stimulate Tregs should be given, as different TLR agonists have been shown to preferentially expand effector T cells over Tregs, and vice versa [16].

Figure 1. Optimal antigen design coupled with selected age-specific adjuvants can alter DC function during infant vaccination and promote effective anti-viral memory responses.

The use of adjuvants with known responsiveness in neonates couldenhance the function of early-lifeplasmacytoid dendritic cells (PDC), follicular dendritic cells (FDC), and lung-migratory CD103+ and CD11b+ DC populations to support the development of anti-viral B and T cell responses. Effectively targeting and using DCs to promote Th1 and T_FH may assist with the generation of cytotoxic T lymphocytes (CTL) and high-affinity, neutralizing antibody production by B cells, respectively. Promoting the maturation of CD103+ DCs in the lung could enhance costimulatory signals and cytokine production, and lead to the induction of anti-viral CTL.

Prospects for indirect protection of the infant population

While infants are the highest priority target for prevention of severe outcomes of RSV infection, difficulties in eliciting protection in this group through active immunization haveled to strategies to diminish the impact of RSV on this age group indirectly by boosting adaptive immune responses of pregnant women or older children.

Promoting protection via enhancing passive transfer of maternal antibody is not a strategy unique to RSV. Immunization of pregnant mothers has proven an effective approach against tetanus, pertussis and influenza [52]. Maternally-transferred RSV antibody has been found to be associated with a lower risk of infectionin infants when antibody titers are maintained above levels that are considered protective [53]. It is estimated that transfer of high concentrations of RSV-specific antibody from vaccinated mothers could extend protection up to six months of age. This approach could be coupled with active vaccination in infants over 6 months of age, at which point immune responses to vaccination are more likely to elicit neutralizing antibody and prevent disease followingnatural infection.

In addition to vaccinating mothers, school-aged children are a promising target population to limit transmission to infants. In-depth household studies in rural Kenya have estimated the average duration of viral shedding as 11.2 days, longer than previously estimated, and more than 10% of infection episodes resulted in shedding for 21 days or more. Shedding was shown to be strongly associated with age and the severity of infection [54]. Additional studies in these same cohorts revealed that older, school-aged siblings are most often responsible for transmitting RSV to within-household infants [55]. A vaccine that would induce a level of protection equivalent to natural infection (around 6 months) in school-aged children is predicted to reduce RSV occurrence in those less than one year old by 35% within 10 years. This is comparable to the estimated 41% protection that might be achieved by directly vaccinating 3-month old babies [56]. In addition to the beneficial impact on infant infection, targeting reduction of shedding in school-aged children may help mitigate disease in the elderly and other high-risk groups.

Conclusions

RSV remains a serious pathogen for which there is no vaccine. While the virus is contained in the upper airways in the majority of RSV-infected infants, millions of children are hospitalized annually as a result of LRTI. Limitations in both innate and adaptive immunity in early life contribute to disease, and infection at a young age may prime infants for suboptimal and rapidly waning immune responses to subsequent exposures of RSV. This may be one reason why natural infection can provide protection against LRTI, but not preventupper respiratory tract infections that recur throughout life.

Immunization of pregnant women and school-aged children is expected to significantly diminish the morbidity and mortality of RSV in the very young. In order to achieve efficacy following direct active immunization of infants, fundamental changes will need to be made to the way that the immune system experiences RSV antigens. An increased understanding of infant immunity will be necessary, particularly in the respiratory tract. While generating protective responses against RSV appears to be very easy in adult, immunocompetent animals, few studies involve testing vaccine candidates or evaluating the response to infection in neonatal animals. A better understanding of responses in the lung, and the lung draining lymph nodes will help inform vaccine design for infants.

Age-specific vaccine strategies will need to counter the suppressive nature as well as harness the strengths of neonatal immunity. This may be accomplished by careful selection of TLR agonists most likely to promote adult-like responses and select for effector rather than regulatory responses. Increasing the potential of APCs to induce mature adaptive responses will be a critical function of an infant RSV vaccine. Coupled withimmunogen design targeted at promoting the most potent antiviral responses, direct protection of infants from RSV may be possible in thefuture.

Highlights.

• -Physical and immunological immaturity contribute to RSV disease in early life

• -Infants exhibit distinct, not defective, responses to inflammatory stimuli

• -Infant innate and adaptive immune responses demonstrate ineffective viral control

• -Age-specific approaches will be necessary to generate effective immunity in infants

• -Vaccination of pregnant mothers and young children may confer indirect protection