Dissecting the defects in the neonatal CD8⁺ T Cell response

Abstract

The neonatal period presents a complex scenario where the threshold of reactivity toward colonizing microbiota, maternal antigens, autoantigens, and pathogens must be carefully moderated and balanced. CD8⁺ T cells are critical for the response against intracellular bacteria and viruses, but this immune compartment maintains altered function relative to adult counterparts because of the unique challenges which face infants. Here, we review our current understanding of the factors that may promote the attenuation and altered function of the neonatal CD8⁺ T cell response and potential avenues for future study. Specifically, we have focused on the neonatal CD8⁺ T cell ontogeny, memory formation, T-cell receptor (TCR) structure and repertoire, TCR inhibitory receptors, and the clinical implications of altered neonatal CD8⁺ T cell function. Special emphasis has been placed on examining the response of preterm neonates relative to term neonates and adults.

Summary Sentence:

Review of the factors that may promote the attenuation and altered function of the neonatal CD8+ T cell response and their clinical implications.

Graphical Abstract

Introduction

Generation of a functional immune repertoire is a developmental process that initiates during fetal life but matures several years after birth in humans[1]. Immediately following birth, neonates move from the sterile in utero environment to the external world, where they are exposed to a plethora of microorganisms, most of which are harmless [2]. This rapid and diverse exposure to organisms requires a fine-tuned tolerance to permit colonization, yet simultaneous clearance of pathogenic organisms [2]. It has been hypothesized that this tolerance leads to less robust cytotoxic CD8⁺ T lymphocytes (CTL) responses and reduced protection in this sensitive age group [3]. CTLs are critical for protection from viruses and intracellular bacteria [4, 5]. Prematurity may further impair CTL responses and increase susceptibility to infections. To compensate for a diminished effector T cell response during infant respiratory syncytial virus (RSV) and influenza virus infection, there is a concomitant increase in neutrophil and macrophage influx, which is likely the source of immunopathology and high rates of morbidity associated with respiratory viruses in this vulnerable population [6]. Despite the critical necessity of CD8⁺ T cell immunity, the mechanisms causing attenuated neonatal CD8⁺ T cell responses remain incompletely elucidated. Here, we discuss our current understanding of the potential reasons for the decreased neonatal CD8⁺ T cell response, the impact on the development of memory, and ultimately the clinical implications of a defective CTL response.

Neonatal T Cell Ontogeny

Neonates display an overall lymphopenia with decreased numbers of T cells. The relative lymphocyte frequencies vary based on gestational age and age after birth in both mice and humans [7–10]. The frequencies of lymphoid and myeloid populations are typically increased in a healthy full-term neonate compared to that of a premature infant [10]. In addition to decreased frequencies of these cell populations in premature neonates, there are also differences in the cell phenotype and functional development during the time when premature neonates should be in utero. Newborns born at an extremely low gestational age (<29 weeks) who do not develop an appropriate T cell phenotype (CD8⁺CD31⁺) by their term-equivalent corrected gestational age (the equivalent age of being born on their planned due date) have an increased likelihood of respiratory morbidity within the first year of life [11].

A potential rationale for T cell functional differences is the dominant presence of recent thymic emigrants (RTE) in the neonatal T cell pool [12, 13]. RTEs continue their post-thymic maturation for approximately 3 weeks, during which they exhibit distinct functions and chromatin architecture from those mature naïve T cells which have circulated longer [14]. Newly egressed human RTEs express complement receptors (CR1 and CR2) and produce IL-8 following activation and can be used to distinguish those RTEs which have undergone increased rounds of peripheral division [15]. Activated CD8⁺ RTEs have reduced effector functions such as proliferation, IFNγ production, and IL-2 production compared to mature naïve CD8⁺ T cells [12, 13, 16, 17]. Stimulation of human RTEs, similar to mice, results in decreased production of IL-2, IL-4 and IFNγ compared to mature naïve T cells [18]. Some of these reduced functions are likely attributed to defects in the aerobic glycolysis pathways required for complete activity and proliferation [19]. Recent evidence suggests the reduced metabolic activity of RTEs is regulated through reduced CD28 expression and signaling, thereby leading to reduced glutaminase and mitochondrial mass [20]. However, many of these changes are subtle and whether these alterations are sufficient to promote reduced function in vivo remains unclear [20].

In addition to the intrinsic reduction in effector functions of CD8⁺ RTEs, following egress from the thymus, RTEs are highly susceptible to extrinsic regulation. IL-7 is important for homeostatic cycling of naïve T cells and increases homeostatic proliferation during lymphopenia [12]. Interestingly, neonatal RTEs, but not adult RTEs, proliferate following IL-7 stimulation alone without TCR engagement. This is not due to increased or persistent IL-7Rα expression [12, 13]. Rather, neonatal IL-7 receptor signaling appears to be both more robust and faster in comparison to adults, as evidenced by increased phospho-STAT5, indicating a potentially different response to both homeostatic and lymphopenic cues [12]. Thus, the prevalence of RTEs within the neonatal pool, whose functions are reduced compared to mature naïve T cells, may drive reduced/altered CTL activity during infection.

Neonatal RTEs are also highly susceptible to the induction of Treg-mediated tolerance relative to mature naïve T cells, likely mediated through their sequestration of IL-2 [16]. TGFβ, a Treg associated cytokine, levels are increased in many neonatal infection models and may aid in the suppression of CD8⁺ T cell function during acute infections or microbial colonization [21, 22]. Increased numbers of Tregs are detected early in human fetal gestation (20 weeks) within cord blood relative to term neonates and adults [23, 24]. In addition, CD4⁺ T cells isolated from fetal samples have an increased propensity to differentiate into Tregs following stimulation [22]. In human infants, soluble adenosine levels, which promote a Treg response, are increased in comparison to adults [25]. Tregs from full-term neonates can suppress T cell function to the same capacity of adult Tregs in vitro [26]. However, preterm and term human neonatal Tregs are different in their suppressive abilities when exposed to inflammation. In the setting of intrauterine infection, late preterm (32–36 weeks gestation) Tregs are less suppressive to the function of conventional T cells, despite similar frequencies to term infants [26]. This reduction in Treg suppression could be explained by Toll-like receptor 2 and pro-inflammatory cytokine signals which induce divergent chromatin changes at the FOXP3 locus and downregulate function [27]. There is also evidence for a role of CD8⁺ Tregs in the developing neonatal immune system. CD8⁺ Tregs can be generated extrathymically by parenchymal cells and suppress neonatal CD8⁺ T cells in transplantation models and persist into adulthood to maintain tolerance [28]. Murine neonatal CD8⁺ Tregs have increased CD5 surface expression indicating stronger TCR signaling suggesting, there may be a preferential selection for only high affinity TCR clones for single positive thymocytes and Treg development [29]. In early murine life (pre-weaning) antigenic exposure is limited from the gut by microbial induction of epidermal growth factor which reduces the formation of goblet cell-associated antigen passages and promotes the development of Tregs. This promotes tolerization to specific microbes during a critical window of development [30].

In addition to increased frequency of Tregs, neonates have an increased frequency of myeloid derived suppressor cells (MDSCs) within peripheral blood, which decreases over the first months of life [31, 32]. The suppressive abilities of neonatal MDSCs are mediated through nitric oxide, PGE2, and S100A9/S100A8 proteins and are induced through lactoferrin [32]. The increased frequency of MDSCs within neonates may be a mechanism to promote tolerance to microbial colonization [32]. The role of these cells in the context of neonatal infection has not been explored. Therefore, neonates have decreased number of T cells, increased proportions of both conventional and CD8⁺ Tregs, and the cytotoxic CD8⁺ T cells are phenotypically and functionally distinct from adult CD8⁺ T cells. These quantitative and qualitative differences thus predispose neonates to inefficient clearance of viral infections.

Neonatal Antigen APC:T Cell Connection

Although intrinsic differences in neonatal T cells can promote functional defects, neonatal APCs can also induce secondary defects in neonatal T cell responses. The neonatal response can be characterized by a Th2 phenotype [33]. This skewing is promoted at the level of antigen presenting cells (APCs), specifically dendritic cells (DCs) [33]. A strong connection has been drawn in mouse models between a deficit in the production of IL-12p70 by APCs and an increase in IL-4 receptor expression on T cells during early life (prior to day 6 of life) [33]. Delayed development of neonatal DCs or a decreased ability to become activated could attenuate an effective CD8⁺ T cell response. The frequency of neonatal conventional DCs (cDCs) are approximately ten-times less than that of an adult mouse [34] [35]. Furthermore, neonatal cDCs are predominantly cDC1s, whereas adults are dominated by cDC2s [34, 35]. Importantly, while cDC1s are typically a pro-inflammatory cell-type, neonatal cDC1s activated following L. monocytogenes infection also produce the suppressive cytokine IL-10 [36]. By day seven of life, neonatal mouse DCs are present and are proficient at the activation of T lymphocytes. Despite the ability of neonatal DCs to aggregate with T lymphocytes, they do display a decreased actin polymerization at the immunological synapse [34].

Typical surface markers of tissue-resident DCs, such as the CD103⁺CD11c⁺ cells, express lower levels of CD103 on their surface during an RSV infection in comparison to adults, despite CD103⁺ DCs being the dominant responders to RSV [37]. Neonatal CD103⁺ cells display a decreased ability to uptake and process antigen [37]. Additionally, these cells display low levels of costimulatory ligands required for proper T cell activation, specifically CD80 and CD86. Differences found between induction of neonatal and adult CD8⁺ T cell responses appears to be epitope-dependent, wherein specific epitopes induce a comparable level of activation while other epitopes do not [37]. Thus, these differences in epitope presentation may alter or skew the CTL response. However, there are inconsistencies in the literature whereby others have demonstrated sufficient numbers of DCs in neonates, normal surface protein expression, and fully functional cross-presentation skills [38, 39].

In humans, neonatal APCs are less mature [40–42] than adult APCs, associated with lower cytokine production in response to stimulation [43, 44]. Neonatal monocytes and macrophages show reduced sensitivity to inflammatory cytokines, via reduced STAT-1 phosphorylation [45] and deficiencies in phagocytic killing [46]. In addition, there is a significant reduction in phagocytosis in premature neonates [47, 48]. Neonatal DCs have no basal expression of co-stimulatory molecules [49] and do not upregulate these receptors upon activation to adult DC levels [50, 51]. Together, many features of DC biology may be context, infection or model dependent. There is a potential role of defective DCs function contributing to a defective neonatal T cell response. Alternatively, a sheer reduction in DC numbers could also be associated with reduced activation of CTLs [52].

T-cell receptor repertoire diversity and the impact on a functional CTL response

An important component of the antigen presenting cell and T cell interaction is at the level of the T-cell receptor (TCR). There is a direct link between the degree of protection elicited by a CD8⁺ T cell and the diversity of the TCR repertoire [8, 53, 54]. A diverse TCR repertoire is closely linked to the required clonal complexity of the T cell response against specific viral components [55]. Neonatal mice have diminished CTL responses to both dominant and subdominant epitopes for influenza virus [56] and RSV [57] associated with differences in the neonatal TCR repertoire compared to adults. Throughout the TCR, six hypervariable complementarity-determining regions (CDRs) have been identified which are responsible for the recognition of cognate peptide expressed on MHC-I. The CDR1 and CDR2 regions are encoded within the germ line TRAV and TRBV genes, while the CDR3 regions are constructed upon the formation of junctions of different V(D)J gene rearrangements. Within the CDR3 region, additional diversity is added by random nucleotide incorporation during recombination by Terminal Deoxynucleotidyl Transferase (TdT) [58–60]. As human fetal gestational age increases, the CDR3 length increases [61, 62]. Rechavi and colleagues interrogated the total T cell repertoire by next generation sequencing in a small number of fetal samples from selective fetal reduction cases, 12–26 weeks gestation, compared to children aged 9–48 months [63]. Fetal samples had fewer nucleotide additions and less junctional trimming compared to children. T cells generated early in gestation (at or after 12 weeks) display a skewing of their TRBV usage with limited polyclonality, but certain clones are enriched throughout the length of gestation [63].

Specific aspects of a TCR which can affect T cell function are affinity which is the strength of a single binding interaction between a TCR and a peptide:MHC (pMHC) molecule and avidity which is the entirety of all binding affinities with the pMHC [64]. Neonatal mice infected with influenza virus at day of life 3 have a decreased affinity and avidity for the immunodominant peptide (NP_(366–374)) in comparison to adults [9]. In contrast, following vaccination of murine neonates with attenuated strains of Listeria monocytogenes, no differences in TCR avidity of antigen-specific T cells were noted in comparison to adults. The response generated was detectable after a single vaccination and resulted in a strong CD8⁺ T lymphocyte response as well as a CD4⁺ Th1 primary T cell response [65, 66]. Similar TCR avidity has also been noted by others in the context of DNA vaccination in neonatal mice immunized within 48 hours of birth [67]. Neonatal mice vaccinated with DNA within the first two weeks of life respond similar to adults and express a minimum of one effector function or experience comparable cytokine expression [67]. Differences in the measurement of TCR avidity between studies may explain the discrepancies noted.

The Role of the Human Public TCR Repertoire

CD8⁺ T cells responding to an epitope exhibit both public (TCRs shared among most individuals) and private (TCRs unique to individuals) specificities. Public clonotypes exhibit convergent recombination, where multiple CDR3 nucleotide sequences result in few amino acid CDR3 sequences [68]. Convergent recombination is more likely when particular TCR amino acid sequences can be encoded by nucleotide sequences that are efficiently produced or encoded by multiple available nucleotide sequences [69, 70]. This is largely determined by codon degeneracy of specific amino acids in the CDR3 sequence. It has been suggested that sequences are not uniformly probable, as opposed to non-random selection from all possible sequences [71]. The small effective size of the CTL CDR3 sequence repertoire is primarily attributable to the fact that CDR3 sequences with large numbers of junctional insertions have very low probability [71].

Extremely preterm neonates (gestational age 23–27 weeks) have a very even distribution of their repertoire, and therefore have a lower clonality and less richness, as compared to term (39–40 weeks gestation) neonates, young children and adults [62]. A developmentally regulated, even, less stringently-selected repertoire in preterm neonates leads to the inclusion of public TCR CDR3 sequences that overlap between unrelated individuals in the preterm repertoire [62]. Analysis of out-of-frame naive TCR repertoires from human twins, cord blood, and adults reveals certain public clonotypes are maintained from prenatal life through early adulthood, which reduces in number late in adulthood [72] and may be associated with the sex of the individual [73]. Ageing also has an impact on repertoire diversity; older individuals (>65 years of age) display a reduction in the influenza A viral-specific T cells Vα and Vβ usage with little diversity [74].

TCR inhibitory molecules

Inhibitory surface molecules modulate T cell function primarily through increasing TCR threshold or decreasing signaling downstream of the TCR or activating receptors [75]. The expression of inhibitory receptors on CD8⁺ T cells is generally associated with reduced function and clearance of pathogens [76]. Inhibitory receptors classically function through the phosphorylation of an immunoreceptor-based inhibitory motif (ITIM) following ligand binding which recruits cytosolic phosphatases such as SHP-1, SHP-2, and SHIP1 [75, 76]. During neonatal colonization, lung microbiota exposure induces PD-L1 expression, the ligand for inhibitory receptor PD-1, on CD11b⁺ conventional DCs presumably promoting tolerance to colonization [77]. Furthermore, CMV infected human infants express high levels of PD-1 and an exhaustion phenotype on T cells [78]. Late-onset sepsis (>48 hours post birth) is associated with increased expression of CEACAM1⁺CD4⁺ T cells as well as increased levels of soluble CEACAM1 [79].

Elevated expression of inhibitory receptors LAIR-1, CD31, and CD200 are found on all neonatal T cell subsets found within cord blood from healthy term infants compared to adult PBMCs [80]. CD31 (PECAM-1) is a known modulator of TCR signaling and is expressed highly on RTEs and naïve T cells [13, 81]. Following activation, CD31 is shed from the surface of T cells, mitigating any inhibitory effects [82]. T cells lacking CD31 expression are resistant to tolerance induction [83]. Interestingly, CD31 expression is reduced on late preterm infants (<36 weeks), but gradually increases as the fetus approaches full gestation [84]. In fact, there is a developmentally regulated increase in CD31 expression in preterm neonates over the first 2–3 months of life, as they approach their term-corrected age [11]. Premature neonates with a decreased abundance of CD31⁺CD4⁺ T cells and more CD31^-CD4⁺ cells at the time of discharge from the hospital are associated with increased morbidity due to respiratory illnesses potentially driven by excessive TNFα production[11].

During acute respiratory viral infection, 3-day old murine neonates have an increased frequency of activated viral specific CD31⁺CD8⁺ T cells within the lung at the peak of the lymphocyte response compared to 7-day old neonates and adults [85, 86]. CD31⁺ neonatal viral-specific CTLs exhibit reduced IFNγ production and proliferation [86]. CD31 transcript levels are dramatically elevated in sorted CD31⁺ effector CTLs from neonatally-infected mice, which indicates a preferential continued expression in neonatal CTLs, compared to naïve or adult effector T cell populations [86]. In support of a differential regulation of CD31 expression between neonates and adult mice, CD8⁺ T cells expressing a transgenic TCR specific for the ova peptide SIINFEKL(OT-I CD8⁺ T cells) were injected into hosts immediately prior to infection with ova-peptide expressing Listeria monocytogenes (Lm-ova). At 8 and 15 days post-infection, OT-I CD8⁺ T cells were sorted and found to have a significant reduction in Pecam-1 transcripts [87]. Interestingly, CD31 can colocalize with the TGFβ receptor on the cell surface to enhance the suppressive effects of TGFβ [88]. Given that TGFβ levels are elevated in multiple neonatal infection models, it is interesting to speculate that neonatal CTLs would be exquisitely sensitive to TGFβ/CD31 mediated suppression [21, 22]. Therefore, in addition to differences in TCR structure and avidity, there are inhibitory molecules that regulate the TCR and regulate effector function. However, beyond the studies presented here, surprisingly few studies have been published examining the role of inhibitor receptors on neonatal T cells and is a potential avenue of future research.

T cell Chromatin Architecture

In addition to inhibitory surface markers, defective clonal expansion and function of effector CD8⁺ T cells in neonates is potentially linked to differences in chromatin architecture and histone modification [14, 89–91]. Neonatal CD8⁺ T cells have specific epigenetic programming, with adult CD8⁺ T cells having higher levels of open chromatin marks and less repressive marks, which contribute to differences in gene expression levels between neonates and adults [89]. The transcriptional profile of genes enriched in the neonatal cells is associated with cell cycle and innate immune response, whereas adult cells have differential transcription of genes important for T cell activation, cytotoxicity and function [89, 90]. Neonatal CD8⁺ T cell transcriptomes cluster with neutrophils, which demonstrates a more innate-like function compared to antigen-specific function [92]. In contrast, others have demonstrated neonatal CD8⁺ T cells generated immediately following birth have a chromatin architecture which closely resembles that of adult effector cells preparing them to act as effector T cells prior to exposure to antigen [93].

Through unbiased epigenetic and transcriptomic analyses, Galindo-Albarrán and colleagues identified histone modifications to genes which play a role in CD8+ T cell toxicity, leading to under-expression and epigenetic silencing of these genes and differential expression of transcription factors [89]. Neonatal mice display a bias towards Th2 activation with IL-4 producing CD4⁺ T cells [94]. Neonatal CD4⁺ T cells bear epigenetic modifications within their Th2 locus which increases the production of Th2 cytokines. Importantly, while neonatal CD4⁺ T cells are hypermethylated on the IFNγ promoter, CD8⁺ T cells do not display this same phenotype [95]. CTLs isolated from cord blood, in contrast to CD4⁺, are not hypermethylated within the IFNγ gene and IFNγ promoter [14, 95].

Human Neonatal CD8⁺ T Cells Display Innate-like Features

Neonatal CD8⁺ T cells produce more antimicrobial peptides and reactive oxygen species and display reduced IFNγ production [89]. In isolated PBMCs from cord blood, neonatal CD8⁺ T cells produce less perforin than adults. Although both adults and neonates express similar levels of degranulation marker CD107, neonates do not have a detectable amount of granzyme B [89]. Neonatal CD8⁺ T cells proliferate at a higher rate under homeostatic conditions, but not upon antigenic stimulation [89], which indicates less efficient clonal expansion. Levels of IL-7 and IL-7 receptor on neonatal cells are much lower, especially in premature neonates, which contributes to this diminished maturation of T cells and V(D)J recombination [96]. Upon stimulation, neonatal CD4⁺ and CD8⁺ T cells produce more IL-8 (CXCL8) compared to adults, and there is a correlation of increased IL-8 production in more premature neonates [97]. This shift in effector cytokine production gives neonatal CD8⁺ T cells an innate-like role during infection because IL-8 recruits and activates neutrophils and γδ T cells [98]. This neonatal innate-like phenotype supports an inflammatory environment that is less antigen driven. Increased T cell production of reactive oxygen species, coupled with neonates’ diminished free radical scavenging capabilities, damages the lung [99]. Prolonged recruitment and survival of neutrophils due to increased IL-8 from neonatal T cells plays a role in tissue damage in the lungs [97, 100]. Infants who develop acute lung injury have been shown to have a higher level of IL-6 and increased CD8:CD4 ratio in the lung during injury and infection [100].

CD8⁺ T cell function in viral infections

Despite a reduction in death rates attributed to respiratory distress syndrome and bronchopulmonary dysplasia in extremely premature infants over the past 20 years [101], death caused by infection has remained the same or has increased in rate [101, 102]. Specifically, respiratory viral infections contribute substantially to global fetal and infant losses and disproportionately affect preterm neonates. Of all deaths less than 12 months of age secondary to a respiratory viral infection, 55% occur in a neonate born before 30 weeks gestation [103]. Influenza virus, respiratory syncytial virus (RSV), herpes simplex virus, human immunodeficiency virus (HIV), human cytomegalovirus, hepatitis B virus, Mycobacterium tuberculosis, and Plasmodium falciparum cause more severe infections and worse clinical outcomes in infants, suggesting a deficiency in CD8⁺ T cell mediated immune responses [104–109]. Neonates, particularly premature neonates, have much lower frequencies and absolute numbers of NK cells, T cells, B cells [110, 111]. However, more important than decreased absolute numbers of cells, neonatal CD8⁺ T cell function is diminished and varies depending on the pathogen and context of infection [1, 110, 112–115].

A longitudinal study of 61 Kenyan infants and 65 infants from New York revealed the age of HIV infection to be a primary outcome-determining factors. Infants infected with HIV in utero, peripartum or postpartum have varying magnitude and function of CD8⁺ T cells strongly dependent on the age of infection, independent of their anti-retroviral therapy use [116, 117]. In contrast to HIV-infected adults, HIV-specific CD8⁺ T cell numbers in neonates do not correlate with reduced viremia or clinical symptoms [116], which demonstrates reduced effector function of these antigen-specific CD8⁺ T cells in the infant. Gag-specific CD8⁺ T cells have impaired IFNγ production, which is lowest in early infancy and increases with age [117]. Other infections such as rotavirus [118] and HSV-1 [4] also have delayed and reduced IFNγ responses in human neonates during primary infections [4]. However, the mechanism of reduced IFNγ has not yet been fully elucidated.

Samples collected from the upper respiratory tract and lower respiratory tract from human infants infected with a respiratory virus demonstrate a propensity toward expansion of terminally differentiated effector cells and reduced mature tissue resident memory (TRM) formation based on CD103 expression [119]. This increased effector T cell infiltration is associated with lung injury [119]. Two-week old mice infected with influenza virus demonstrate comparable numbers of viral-specific CD8⁺ T cells in the lungs at day 12–15 post infection adult mice [120]. Despite the prevalence of the viral-specific CD8⁺ T cells during acute infection, mice infected at 2 weeks of age have reduced TRM in adulthood, and, more importantly, this reduced TRM population provides less protection compared to adults re-challenged with a heterosubtypic virus mice [120]. In addition, more immature mice at three days of age have greatly diminished CTL responses to both dominant and subdominant epitopes for influenza virus. In contrast, 7-day old mice have normal CTL responses to the dominant PA_(224–233) and subdominant NS2_(114–121) epitopes. These 7-day old mice mounted a NP_(366–374) response that was intermediate between that of 3-day old and adult mice[56]. Such altered patterns of neonatal CTL hierarchy have also been demonstrated by others in a 7-day old mouse model of RSV infection [57]. Human infants have a reduced and delayed production of antigen-specific CD8⁺ T cells, despite the abundance of antigen, which indicates defective expansion of effector T cells [121, 122]. IL-4, a classic Th2 cytokine, can suppress the activity of Th1 CD4⁺ cells which are required to fully activate and differentiate cytotoxic T cells [123]. Despite the prevalence of Th2 skewing in murine studies, there has been some difficulty demonstrating the Th2 skew within human neonates [110]. The primary response within the human infant lung is dominated by Th2-like responses and is associated with severe bronchiolitis and the development of asthma [94].

Development of the memory CD8⁺ T cell compartment

Neonatal CD8⁺ T cell memory has been intensely investigated because of the implications for vaccine response and efficacy. Compared to other lymphocyte responses, the efficacy of neonatal CD8⁺ T cell memory has been a challenge given that most vaccination platforms provide minimal CD8⁺ T cell-mediated protection and determination of antigen-specific memory in a human population is highly variable [124]. In murine acute infection models, neonatal CD8 T cells terminally differentiate and form a reduced memory pool, whereas adult CTLs demonstrate a bifurcation of their response representing both a terminally differentiated and memory pools. There are two primary theories regarding why neonatal CD8⁺ T cells assume different end results from adults during infection: the proliferation model and the origin model [125]. The proliferation model dictates that variations in the CD8⁺ T cell response during development are the result of differences in homeostatic proliferation prior to infection. In comparison, the origin model states that neonatal and adult CD8⁺ T cells are the result of two discrete hematopoietic stem cell lineages.

To decipher the contribution of naïve T cells generated during the early neonatal period in response to adult infection, Smith and colleagues performed elegant fate-mapping studies and multiple adoptive transfer experiments to identify a population of neonatally-derived naïve CD8⁺ T cells which become Virtual memory (VM) cells. When murine adults were challenged with Listeria Monocytogenes, neonatally-derived VM cells responded within three days to pro-inflammatory cytokines and antigen stimulation but acted with innate-like function, as compared to classical memory cells [93]. This activity is dictated by a distinct transcriptional profile which is maintained through a unique landscape and specific transcription factors (Tbx21, Eomes, and Runx1) which maintain an effector-like profile [93]. In contrast, CD8⁺ T cells produced after the neonatal period are more likely to respond only to their specific cognate antigen and contract into long-lived memory cells.

During acute infection in the neonatal period, murine neonatal CD8⁺ T cells rapidly expand and differentiate into short-lived effector cells, but remain incapable of contracting into memory cells following Listeria Monocytogenes infection [126]. In memory pool formation, decreased numbers of clonotypes are found within mice who are vaccinated early in life (day seven of life) and display a lower TCR binding avidity compared with adults [55]. During a secondary response, the dominant CD8⁺ T cell populations are adult clonotypes responding to the infection and not the clonotypes derived from the neonatal memory compartment [127]. In fact, a larger number of adult naïve CD8⁺ T cells are recruited in a neonatal secondary response than in an adult secondary response. This is attributed to the inability of neonatal memory CD8⁺ T cells to outcompete adult naïve or memory T cells due to their lower avidity TCRs [55]. Altogether, under a precise set of conditions, neonates can produce a measurable albeit reduced or altered memory CD8⁺ T cell pool and this pool retains an innate-like effector functional capacity.

This lack of robust antigen-specific memory in neonates poses a great challenge in vaccine design and necessitates multiple booster doses during infancy to develop appropriate protection. In addition, children under the age of two have little memory T cells in lymphoid or mucosal sites, as determined using deceased organ donor samples [128]. Formation of CD8⁺ T cell memory is imperative for protection against intracellular pathogens. Studies in neonatal mice have shown selection of proper adjuvants [129] or use of DNA vaccines [130–132] can increase CD8⁺ T cell memory formation and increase functional memory recall response during a secondary challenge [133]. Antigens introduced via DNA vaccinations persist longer than subunit or attenuated vaccines. This antigen persistence could explain improved formation of more robust memory CD8⁺ T cells, as more mature naïve T cells can respond to the antigen. Nevertheless, newborns safely mount responses and form B cell and Th2 memory pools following the hepatitis B and diphtheria-tetanus-whole-cell-pertusis-HepB-Haemophilus influenza type b vaccines [134, 135]. Both vaccines promote a Th2 adaptive response via stimulation of an intracellular inflammasome pathway [136]. These studies highlight the role of neonatal memory CD8⁺ T cells as “first responders” until more mature naïve T cells can respond. These mature naïve T cells from later in life replace the neonatal memory cells and form the long-term memory CD8⁺ T cell compartment. Moving forward, determination of the different age-associated layers of memory in humans would be challenging, if not impossible.

Concluding Remarks

Immediately following birth, neonates move from the sterile in utero environment to the external world, where they must quickly adapt to exposure to foreign organisms, most of which are harmless, but some are pathogenic. This carefully balanced response favors tolerance. This tolerance dampens a robust adaptive immune response and promotes an innate-like, pro-inflammatory response, which can cause bystander damage. Neonatal T cells harbor less-specific TCRs that can recognize a wider range of antigen than adult TCRs. These more promiscuous T cells are also programmed to respond to inflammatory cytokines and react in a less-antigen specific, more innate-like phenotype than adult CD8⁺ T cells. The antigen presenting cells responsible for activating T cells are also much less mature in neonates which contribute to T cell defects as well as have broader implications in fighting infections. Preterm neonates are particularly at risk because they do not have the benefit of the third trimester to prepare the fetal immune system for perinatal transition from in utero development to postpartum protection. The activation and function of neonatal CD8⁺ T cells appears to be antigen and context dependent. There remain significant knowledge gaps in the kinetics of APC maturation and its impact on T cell function in both the preterm and term neonate. Knowledge of the neonatal APC:T cell interaction can help to identify better adjuvants to increase vaccine efficacy during the vulnerable infant period. More importantly, longitudinal studies are needed to determine how early life infection impacts the development of the immune system and imprints the adult immune system. Prudent vaccine design should attempt to induce the expansion of a large breadth of TCR clones to improve recall responses. To dissect key components of the neonatal immune response, use of age-appropriate animal models and human samples are critical. Careful consideration of age-dependent differences is germane to the development of efficacious therapeutics and preventative strategies to protect this vulnerable population.

Figure 1. CD8⁺ T cell Differences in Neonates and Adults.

Neonates lack CD8⁺ T cell Receptor diversity. Receptor diversity, formation of memory, and response to antigen increase with age while the number of thymic emigrants decrease with age. Neonatal CTLs display an innate-like phenotype; they have a higher activation threshold, decreased effector function.

Figure 2. Phenotypic Differences of Neonatal and Adult CTLs.

Neonatal CTLs have lower affinity for their cognate antigen, fail to shed the inhibitory receptor CD31 upon activation. These cells produce more of the innate cytokine IL-8 and less IFN-γ upon activation as compared to adult CTLs. These differences are controlled at the chromatin level.

Abbreviations

APC

Antigen presenting cell

cDC1

Conventional dendritic cell 1

cDC2

Conventional dendritic cell 2

CDR1

Complementarity-determining region 1

CDR2

Complementarity-determining region 2

CDR3

Complementarity-determining region 3

CMV

Cytomegalovirus

CR1

Complement Receptor 1

CR2

Complement Receptor

CTL

Cytotoxic T Lymphocytes

DC

Dendritic cell

DNA

Deoxyribonucleic acid

HIV

Human immunodeficiency virus

HSV-1

Herpes simplex virus 1

IFNγ

Interferon gamma

IL-10

Interleukin 10

IL-2

Interleukin 2

IL-4

Interleukin 4

IL-6

Interleukin 6

IL-7

Interleukin 7

ITIM

Immunoreceptor-based inhibitory motif

LAIR-1

Leukocyte-associated immunoglobulin-like receptor 1

MDSC

Myleoid derived supressor cells

MHC

Major histocopatibility Complex

p:MHC

peptide:Major histocopatibility Complex

PBMC

peripheral blood mononuclear cell

PGE2

Prostoglandin E2

RSV

Respiratory syncytial virus

RTE

Recent Thymic Emigrants

STAT1

Signal transducer and activator of transcription 1

STAT5

Signal transducer and activator of transcription 5

TCR T

cell Receptor

TdT

Terminal deoxynucleotidyl transferase

TGFβ

Transforming growth factor beta

TLR2

Toll-like Receptor 2

TNFα

Tumor Necrosis Factor alpha

Treg

T regulatory cell

TRM

Tissue resident memory

VM

Virtual memory