Early-life human CD8⁺ T cells exhibit rapid, short-lived effector responses and a unique transcription factor landscape

Abstract

Neonates and infants are distinct in their clinical and cellular responses to viral infections, with neonatal CD8⁺ T cells displaying innate-like characteristics and a low threshold for T cell receptor activation. However, specific molecular programs that drive these unique responses are incompletely understood, particularly in humans, and targetable pathways to modulate viral illness in this vulnerable population remain to be elucidated. Early-life immune responses may be developmentally programmed to prioritize avoidance of tissue immunopathology, especially while maternal immunoglobulin provides passive immunity. We set out to define the unique response characteristics and transcription factor landscape of neonatal human CD8⁺ T cells. Here, we report evidence that naïve neonatal human CD8⁺ T cells are poised for an accelerated effector switch, with elevations of killer cell lectin-like receptor G1 (KLRG1), killer cell lectin-like receptor B1 (KLRB1/CD161), Fc epsilon receptor I-gamma (FCER1G), DNAX accessory molecule-1 (DNAM1/CD226), granzymes, tumor necrosis factor alpha (TNFα), interleukin 2 (IL-2), and glycolysis compared to naïve adult CD8⁺ T cells. Further, rapid proliferation and cell death occur upon activation of neonatal CD8+ T cells, with cell viability largely rescued by IL-2 or IL-7. These features are coupled with a unique transcription factor landscape, including high expression of thymocyte selection associated high mobility group box (TOX) and HELIOS (IKZF2), and these signatures continue in postnatal life until at least 2 mo of age. We conclude that early-life human CD8⁺ T cells maintain a unique transcriptional state associated with an accelerated effector switch and short-lived effector program, revealing key nodes of regulation relevant for the unique immunobiology of neonatal humans.

Introduction

The early-life neonatal and infant period is defined by formative biological transitions, important for the shift from life as a fetus to life outside the womb. The fetal immune system develops in a state devoid of encounters with live microbes and prioritizes critically important regulation mechanisms maintaining immunological tolerance to maternal tissues while supporting sufficient capacity for anti-microbial immune defense for survival after birth. In postnatal life, the immune system must adapt from birth through infancy to increasingly function independently for pathogen defense, especially as passive immunity conferred by maternal immunoglobulin wanes, and these pathogen defense mechanisms must be appropriately balanced to avoid immune-mediated tissue damage. Neonates and infants are uniquely vulnerable to infection, with respiratory viral infection-induced bronchiolitis being the leading cause of infant hospitalizations, contributing to millions of outpatient visits and over 80,000 U.S. hospitalizations annually in children under five y of age (1, 2). The toll of viral illness on infants around the world is exponentially higher (3, 4). CD8⁺ T cells function as a crucial determinant of anti-viral immunity and contribute to vaccine responses and pathological tissue damage. Prior work has established key cellular features of neonatal human and mouse CD8⁺ T cells, including quicker proliferation, susceptibility to cell death, and deficient immunological memory formation (5–10). At the same time, in humans, memory T cells are present in the intestine and lungs during infancy and accumulate over the first few months of life at sites of antigen exposure (9). Notably, elevated levels of airway CD8+ T cells with cytotoxic capacity have been identified in human infants with lung injury from viral respiratory tract infection, implicating the need for fine-tuned responses to protect mucosal tissue during antigen encounter (11). Still, the molecular and metabolic drivers of the unique effector response and the associated transcription factor landscape of early-life human CD8 T cells remain poorly understood.

Neonatal human CD8⁺ T cells comprise a significantly lower frequency of immune cells compared to their adult counterparts and are more prone to cell death (5, 12). Healthy adult CD8⁺ T cells differentiate during activation-induced proliferation from “naïve” (CCR7⁺CD45RA⁺) into effector and memory cells, each with unique roles in response to new or previously encountered antigens. Naïve neonatal human and mouse CD8⁺ T cells have a low threshold for T-cell receptor (TCR) activation, enhanced proliferation, and an accelerated program of differentiation into effector cells at the expense of forming memory cells (7, 13). Additionally, the cytotoxic features of neonatal CD8⁺ T cells differ from those of adults, with neonatal human and mouse CD8⁺ T cells exhibiting innate-like functions (14) and ability to become cytotoxic effectors in response to cytokines (IL-2, IL-12, and IL-18) alone without TCR engagement, referred to as “bystander activation” in the absence of recognizing a specific antigen (15, 16). Moreover, neonatal mouse CD8⁺ T cells transferred to adult mouse recipients that were later infected with LCMV and assessed at the peak of the response (day 8) have been shown to express elevated cytotoxic molecules such as granzymes A, B and K, perforin, interferon gamma (IFNγ) and KLRG1 compared to their adult cell counterparts (15). Since CD8⁺ T cells act as a double-edged sword by protecting from intracellular pathogens while also contributing to immunopathology during severe viral bronchiolitis in human neonates and infants (11, 17), careful regulation of these potentially damaging cells is critical for host survival.

The molecular features, including metabolic and gene regulatory biology, of neonatal CD8⁺ T cells are also distinct. The metabolic state of CD8⁺ T cells is integral to their functions, with glycolysis (and its associated biosynthetic precursors) supporting effector cell responses and oxidative phosphorylation supporting naïve and memory T cells (18, 19). Upon activation, neonatal mouse CD8⁺ T cells increase glycolytic capacity to a greater extent than adult CD8⁺ T cells (20), fitting with their increased proliferative and effector capacities. Unique underlying gene regulation mechanisms in neonatal human and mouse T cells include altered microRNA expression (21) as well as rapid downregulation of T cell factor 1 (TCF1; encoded by TCF7) in dividing neonatal T cells after activation, in line with decreased stemness, survival, and memory (13). Recent thymic emigrants (RTE) have been described to express transcription factors including forkhead box protein 3 (FOXP3), IKZF2 (HELIOS), and IKZF4 (22), thymocyte selection-associated high mobility group box family member 2 (TOX2), and SRY-box transcription factor 4 (SOX4) (23, Park, 2020 #23, 24). However, transcription factor programs in early-life naïve human T cells at baseline and after activation that may mediate metabolic and effector programs have not yet been fully elucidated.

Here, we set out to further define unique response characteristics and the coupled transcription factor landscape of early-life human CD8⁺ T cells, which may be developmentally programmed for rapid responses to hasten host defense and limit tissue immunopathology during this vulnerable growth phase. We found that in early-life, naïve human CD8⁺ T cells exhibit a rapid activation-induced switch to increased glycolytic capacity, proliferation, and cell death concomitant with heightened effector molecule production, which we associate with the unique combination of elevated TOX and HELIOS transcription factors.

Results

Naïve neonatal human CD8⁺ T cells rapidly engage glycolytic metabolism and exhibit augmented proliferative and cell death programs

We first sought to assess cellular metabolism, proliferation, and cell survival in neonatal human CD8⁺ T cells. Through reanalysis of published single-cell RNA sequencing data from fetal, neonatal, and adult naïve CD8⁺ T cells (22), we observed elevation of a glycolysis gene signature (25) in naïve fetal and naïve neonatal CD8⁺ T cells compared to those from adults, reflecting prominent differentially expressed genes in the glycolysis pathway (SI Appendix, Fig. S1 A and B). Since these transcriptome changes were observed in naïve cells without prior activation, we next tested the hypothesis that naïve neonatal CD8⁺ T cells can robustly engage glycolytic metabolism more rapidly than adult cells, which are known to demonstrate low-level glycolysis on a timescale of minutes to hours after TCR activation (26). We used cord blood from healthy full-term C-section-delivered neonates and whole blood from young adult healthy donors between 18–30 years of age, to assess metabolic flux in freshly sorted naïve CD8⁺ T cells (CD8⁺CCR7⁺CD45RA⁺). Isolated naïve CD8 T cells were >97% CD8αβ+ and TCRαβ+ (SI Appendix, Fig. S1C), ruling out an influence of unconventional or innate T cell subsets. Freshly isolated naïve CD8 T cells were analyzed within 30 minutes of exposure to activation beads coated with anti-CD3 and anti-CD28 using an XF Agilent Seahorse Bioanalyzer. As expected, the extracellular acidification rate (ECAR) as a measure of glycolysis was detectable but low in adult naïve CD8⁺ T cells with this short stimulation; however, neonatal CD8+ T cells exhibited a significantly higher glycolytic capacity (Fig. 1 A and B and SI Appendix, Fig. S1D). Thus, naïve neonatal human CD8⁺ T cells have an augmented capacity to rapidly and robustly engage glycolytic metabolism.

Figure 1. Naïve neonatal human CD8⁺ T cells rapidly engage glycolytic metabolism and exhibit augmented proliferation and cell death in the absence of IL-7 or IL-2.

(A) Extracellular acidification rate (ECAR) measured in naïve CD8⁺ T cells stimulated with dynabeads (anti-CD3 and anti-CD28) for 30 minutes prior to addition of the indicated drugs for glycolysis stress testing (n=5 neonatal/adult pairs; maximal ECAR after oligomycin: P-value calculated using the unpaired T-test, p<0.0001). (B) Delta glycolytic capacity (maximal ECAR after addition of Oligomycin minus baseline ECAR prior to the addition of glucose). P-value calculated using the Mann-Whitney test. n=10 neonatal/adult pairs. (C) Proliferation after 72 hours activation of naïve CD8 T cells with anti-CD3 and anti-CD28 (representative histogram and delta division index; n=4 neonatal/adult pairs). Division index is the average number of cell divisions that a cell in the original population has undergone (including the undivided peak). (D) Delta percent (%) Ki-67 positive naïve CD8+ T cells at baseline. Representative histograms and quantified on the right. Mann-Whitney test. n=6 neonatal/adult pairs. (E) Representative histogram and delta mean fluorescence intensity (MFI) of Fas on day two to three of TCR activation of naïve CD8 T cells and restimulation with phorbol 12-myristate 13-acetate (PMA)/ionomycin. p-value calculated using the Mann-Whitney test. n= 6 neonatal/adult pairs. (F-H) Viability of naïve CD8 T cells after 48–72 hours in culture with the indicated exogenous cytokine added (20 ng/mL of IL-7 and 200 units/ml IL-2), measured by Live/Dead staining. Representative histogram and cumulative data on percent viable cells are shown. n=5–10 neonatal/adult pairs; p-value calculated using the Mann Whitney test. Delta MFI or delta % is the neonatal or adult value minus the adult value.*p ≤0.05, **p ≤0.01, ***p ≤0.001, ****p ≤0.0001, ns: not significant.

We next assessed cell proliferation and observed increased proliferation in naïve neonatal compared to naïve adult CD8⁺ T cells after 72 hours of stimulation with anti-CD3 and anti-CD28 (Fig. 1C and SI Appendix, Fig. S1E). In addition, we measured increased Ki-67 protein levels and an elevated mitotic spindle gene expression signature, consistent with rapid exit from quiescence, at baseline in naïve neonatal CD8⁺ T cells (Fig. 1D and SI Appendix, Fig. S1 F and G). Moreover, we observed increased Fas protein (Fig. 1E and SI Appendix, Fig. S1H) and a significant viability defect in neonatal CD8⁺ T cells with TCR-mediated activation (Fig. 1F and SI Appendix, Fig. S1 I and J), though no differences in protein levels of FASL were observed (SI Appendix, Fig. S1K). The predisposition to cell death may be influenced by the high circulating levels of factors such as the pro-apoptotic TRAIL protein that may poise neonatal naïve CD8⁺ T cells for reduced survival (SI Appendix, Fig. S1L). To quantify other proteins in circulation that may influence function and viability, we measured serum chemokines and cytokines in neonatal and adult blood and found several growth factors and chemokines, including CXCL10 and sCD40L, elevated in neonatal serum (SI Appendix, Fig. S1L). We next aimed to avoid confounding effects of increased cell death in neonatal CD8⁺ T cells by assessing the impact of exogenous cytokines. We observed that the addition of either IL-7 or IL-2 improved viability of neonatal CD8⁺ T cells 48–72 hours after TCR activation (Fig. 1 G and H and SI Appendix, Fig. S1 M and N); we, therefore, provided cytokine in subsequent functional assays. Together, these data demonstrate that naïve neonatal human CD8⁺ T cells rapidly and robustly engage glycolytic metabolism and exhibit augmented proliferative and cell death programs, consistent with a bias toward short-lived effector responses.

Naïve neonatal human CD8⁺ T cells are poised for heightened effector responses

Building from our cellular metabolism findings, we next sought to assess human neonatal naïve CD8⁺ T cell effector capacity at baseline and early after activation. We first utilized bulk RNA sequencing of freshly sorted naïve CD8⁺ T cells to assess neonatal versus adult cells at baseline (without stimulation) and after 48 h of activation with anti-CD3, anti-CD28, and IL-2 plus restimulation with phorbol 12-myristate 13-acetate- PMA- and ionomycin. We observed markedly different transcriptome profiles (SI Appendix, Fig. S2A), both at baseline with 544 upregulated and 549 downregulated transcripts and after activation with 388 upregulated and 324 downregulated genes in neonatal compared to adult CD8⁺ T cells (Fig. 2 A and B). Across the two datasets, we consistently observed increased expression of a core set of effector-related and innate-like cytotoxic genes, including KLRB1, KLRG1, CD226 (DNAM-1), FCER1G, and granzyme K (GZMK) in neonatal compared to adult cells (Fig. 2C), while other effector-related genes were increased in either baseline or activated neonatal cells (Fig. 2D and SI Appendix, Fig. S2B) (27–31). Moreover, we observe significant enrichment for effector genes among those upregulated in neonatal naïve CD8⁺ T cells at baseline using gene set enrichment analysis (GSEA) from published data on effector versus memory CD8⁺ T cells (Fig. 2E and SI Appendix, Fig. S2C)(32). Consistent with analyses of our bulk RNA sequencing, we again observed elevated effector signature scores and differentially upregulated effector genes in publicly available single-cell RNA sequencing data from naïve CD8⁺ T cells during fetal development and neonatal life compared to adult naïve CD8⁺ T cells (SI Appendix, Fig. S2 D and E) (22). Thus, the transcriptome of naïve neonatal CD8⁺ T cells from humans is biased toward an effector cell program.

Figure 2. Naïve human CD8⁺ T cells from neonates are poised for heightened effector differentiation.

Bulk RNA-seq analysis of naïve CD8⁺ T cells isolated from 11 neonates and 10 adults. Neonatal versus adult CD8⁺ T cell gene expression as volcano plot (A) baseline samples, adult n=6 and neonatal n=7 and (B) stimulated (anti-CD3 and anti-CD28 for 48 hours, followed by PMA/ionomycin restimulation for 4 hours) samples, adult n=4 and neonatal n=4. Significantly-up: log2FC ≥ 0.5 and p.adjust ≤ 0.05; Significantly-down: log2FC ≤ −0.5 and p.adjust ≤ 0.05. (C-D) Heatmap of select effector molecules differentially expressed in adult and neonatal naïve CD8⁺ samples at baseline or after stimulation as indicated. TPM ≥ 1, log2FC ≥ |0.5|, p.adjust ≤ 0.05. (E) GSEA demonstrating effector genes enriched in baseline neonatal CD8⁺ T-cells (32). (F-G) Protein levels were quantified by flow cytometry on naïve CD8+ T cells at baseline, after four hours of PMA/ionomycin stimulation (in cRPMI media with no IL-2 or TCR stimulation) on day zero (D0), and after TCR activation and restimulation with PMA/ionomycin for four hours and treatment with brefeldin/monensin on day two to three (D2–3) in IL-2 media. Shown are surface CD226 (DNAM-1), KLRG1, and KLRB1 (F) and intracellular TNFα and IL-2 (G) proteins. Representative histogram and graphs of cumulative delta mean fluorescence intensity (MFI) or delta percent (%). n=6–14; p-value calculated using Mann-Whitney test. Delta MFI or delta % is the neonatal or adult value minus the adult value. *p ≤0.05, **p ≤0.01, ***p ≤0.001, ****p ≤0.0001, ns: not significant.

We next sought to validate our gene expression analyses using surface protein staining and assessment of effector cytokine production by spectral flow cytometry. Naïve neonatal human CD8⁺ T cells expressed higher surface CD226 (DNAM-1), KLRG1, and KLRB1 at baseline, after 4 hours of PMA/ionomycin stimulation, and in the case of CD226 and KLRG1 after 48–72 hours of activation with anti-CD3, anti-CD28, and IL-2 followed by 4 hours of PMA/ionomycin restimulation (Fig. 2F and SI Appendix, Fig. S2F). To confirm functional relevance of the effector-like transcriptome profile and surface markers, we measured TNFα and IL-2 protein production by intracellular staining. While no detectable cytokine was produced in resting naïve CD8⁺ T cells, neonatal cells produced significantly more TNFα and IL-2 with just 4 hours of PMA/ionomycin stimulation in the naïve baseline state and also after 48–72 hours of activation with anti-CD3, anti-CD28, IL-2, and restimulation with 4 hours of PMA/ionomycin (Fig. 2G and SI Appendix, Fig. S2G), while no differences were observed in CD69, IFNγ, or CD107a (SI Appendix, Fig. S2H). To probe whether there may be a difference in clonally expanding CD8 T cells in neonates, TCRα/β diversity of adult and neonatal (term cord blood) naïve CD8 T cells was analyzed from single-cell RNA sequencing data (33). We found that TCR diversity is high and similar between naïve adult and neonatal human CD8 T cells (SI Appendix, Fig. S2I). In sum, naïve human CD8⁺ T cells from neonates are poised for heightened effector responses with capacity for effector cytokine production within hours of stimulation.

Naïve neonatal human CD8⁺ T cells express the transcription factors TOX and HELIOS

We next sought to uncover transcription factors that characterize the unique programming of naive neonatal human CD8⁺ T cells. We observed in our bulk RNA sequencing data of baseline and activated naïve CD8⁺ T cells significantly increased expression of the transcription factors thymocyte selection-associated high mobility group box (encoded by TOX) and HELIOS (encoded by IKZF2) in neonatal cells (Fig. 3A). Among all transcription factors assessed, these two had the highest fold change elevations in naïve neonatal human CD8 T cells at both baseline and after activation. We used spectral flow cytometry to assess protein levels of TOX and HELIOS and observed significantly higher expression of both factors in baseline and stimulated naïve neonatal CD8⁺ T cells compared to naïve adult CD8⁺ T cells (Fig. 3B and SI Appendix, Fig. S3A). Consistently, our re-analysis of publicly available single-cell RNA sequencing data revealed that both fetal and cord blood-derived naïve CD8⁺ T cells similarly expressed high TOX and IKZF2 (SI Appendix, Fig. S3B).

Figure 3. The transcription factor landscape of neonatal human CD8⁺ T cells is unique.

(A) Transcription factor expression in neonatal (n=11) and adult (n=10) naïve CD8⁺ T cells. TPM ≥ 1, log2FC ≥ |0.5|, p.adjust ≤ 0.05. (B) Protein levels for TOX and HELIOS were quantified by flow cytometry on naïve CD8+ T cells at baseline, after four hours of PMA/ionomycin stimulation within the first 24 hours after plating with cRPMI media on day zero (D0), and after two to three days (D2–3) of TCR activation in IL-2 media and restimulation with PMA/ionomycin for four hours, followed by treatment with brefeldin/monensin. Representative histogram and graphs of cumulative data on delta mean fluorescence intensity (MFI). n=4–9 adult/neonatal pairs; p-values calculated using Mann-Whitney test. Delta MFI or delta % is the neonatal or adult value minus the adult value. *p ≤0.05, **p ≤0.01, ***p ≤0.001, ****p ≤0.0001, ns: not significant.

In examining messenger RNA (mRNA) expression of major transcription factors in our bulk RNA sequencing, we found that additional transcripts were differentially expressed. Baseline naïve CD8 T cells from neonates expressed significantly more eomesodermin (EOMES) and significantly less RUNX3 (SI Appendix, Fig. S3C). Activated naïve neonatal CD8 T cells expressed significantly more RUNX3, ID2, and IKZF4, and significantly less TCF7 and KLF8 (SI Appendix, Fig. S3D). The transcript levels of stemness factor TCF1 (encoded by TCF7) signature was decreased in fetal and neonatal cells compared to adult naïve CD8⁺ T cells at baseline in a published single-cell RNA sequencing (scRNA-seq) dataset (SI Appendix, Fig. S3E) (34, 35). At the protein level, T-box expressed in T cells (TBET, encoded by TBX21) and EOMES, which are major regulators of the CD8⁺ T cell effector cell program, were significantly elevated at baseline and more so after activation in neonatal CD8⁺ T cells (SI Appendix, Fig. S3F). TCF1 protein levels were similar between adult and neonatal at baseline but exhibited a slight, statistically significant decrease after stimulation (SI Appendix, Fig. S3F). Additionally, the BLIMP1 repressor was not differentially expressed after TCR activation (SI Appendix, Fig. S3G). Although TOX expression is a hallmark of CD8⁺ T cell exhaustion, the canonical exhaustion marker TIM3 was not consistently different at the protein level in neonatal CD8⁺ T cells (SI Appendix, Fig. S3G). Programmed cell death protein 1 (PD-1), which is linked to cell exhaustion and driven by TOX but also elevated after cell stimulation, was elevated at baseline and especially after two to three days of TCR stimulation (SI Appendix, Fig. S3G). We found that cells expressing higher levels of TOX also had higher activation/effector phenotypes, including higher expression of cytokines TNFα, IL-2, and IFNγ and CD69 (SI Appendix, Fig. S3H). We also considered the possibility that elevated TOX and HELIOS expression is developmentally related to being recent thymic emigrants (RTE) in the neonate, and we did observe enrichment of the RTE gene set in naïve neonatal compared to adult cells at baseline and especially after activation (SI Appendix, Fig. S3 I and J). Thus, naïve neonatal human CD8⁺ T cells are unique in their high expression levels of TOX and HELIOS and have more robust upregulation of TBET, EOMES, and PD-1, and decreased TCF1, providing a unique transcription factor landscape shaping their response to stimulation.

The unique signatures of neonatal human CD8⁺ T cells persist in post-natal life

To assess if the characteristics we observed in perinatal CD8⁺ T cells persist postnatally beyond when the stress of birth would influence their cell state, we integrated baseline single-cell RNA sequencing data on naïve CD8⁺ T cells across fetal, cord blood-derived, and postnatal samples from both premature (33) and full-term infants up to 2 mo of age (36). Relative to adult naive CD8⁺ T cells, we still observed elevated expression of glycolysis- and effector-related transcripts, as well as increased TOX and HELIOS, after birth until at least two mo of age (Fig. 4 A–C). Of note, the activation phenotype in post-natal life had features of bystander cell activation (SI Appendix, Fig. S4 A and B) and a high RTE score (SI Appendix, Fig. S4 C–F) (22, 37). In sum, the neonatal features that distinguish CD8⁺ T cells from those of adults persist into the first two mo of life.

Figure 4. The neonatal human CD8⁺ T cell glycolysis, effector, and transcription signatures persist postnatally.

(A-C) Dot plot showing the expression of glycolysis (A), effector (B), and transcription factor (C) genes in naïve CD8⁺ T cells across age groups from two integrated datasets (36) and (33), adult (n = 3), preterm (n = 5) and full-term (n = 5) neonatal cord blood, preterm early-life at one week (n = 5), one month (n = 7), and two months (n = 4), as well as full-term infants at 2 months (n = 6).

Discussion

The neonatal life stage represents a particularly vulnerable period in development when the immune system must balance protection from infection with regulation of tissue damage by immunopathology. CD8⁺ T lymphocytes function to protect the host from intracellular infections, including viral infections, by leveraging their capacity for cytotoxicity to kill infected or damaged cells. How CD8⁺ T cells in neonates are uniquely programmed to achieve functional immunity without causing inflammatory tissue damage during this critical growth period remains incompletely understood, particularly in humans. Our findings demonstrate the capacity of naïve neonatal human CD8⁺ T cells to exhibit rapid, short-lived effector responses, accompanied by a distinctive transcription factor landscape when compared to adult CD8⁺ T cells. These insights significantly enhance our understanding of unique neonatal and infant human CD8⁺ T cell responses, highlighting three key concepts discussed below: (i) poising for an immediate effector switch, (ii) elevated TNFα effector cytokine balanced by increased cell death, and (iii) altered transcription factor landscape.

Studies of naïve adult CD8⁺ T cells have defined the kinetics of differentiation responses, with the earliest cellular responses of cell growth and amplification of protein synthesis beginning 8–12 h after activation in vitro (38). This is followed by activation marker upregulation, cell proliferation, and differentiation over time with peak effector responses typically 3–8 d after activation. Glucose transporter trafficking to the cell surface occurs upon specific signals such as those induced by hypoxia and IL-7 signaling (39, 40). We now demonstrate that resting naïve neonatal human CD8⁺ T cells are already poised to engage glycolysis rapidly and robustly. Hand in hand with this rapid engagement of glycolytic metabolism is rapid proliferation and increased cell death. The remarkably swift intrinsic effector switch upon stimulation of naive neonatal human CD8⁺ T cells may be driven at least in part by the increased growth factors present in vivo in growing neonates and infants, which can elevate signaling pathways such as MAPK and mTOR in T cells and other cell types.

CD8⁺ T cells mediate their function not only through cytolytic granule delivery to target cells but also via effector cytokine secretion. Key among these is TNFα, which we report is produced at higher levels from neonatal CD8⁺ T cells upon activation. Additionally, we observe that neonatal cells exhibit higher expression of several molecules, including CD226 (DNAM-1; an adhesion molecule), KLRG1 (with roles in effector function and regulation), and FCERG1, which are typically associated with natural killer and innate-like T cells with high cytotoxic potential (ILTCKs) (31). Moreover, FCERG1 has the potential to reduce the TCR activation threshold and may play a role in this manner in neonatal cells, which are known to exhibit fast TCR signaling (41). Importantly, this phenotype is not associated with CD8αα polyclonal unconventional T cells (42), TCRγδ cell phenotypes, or differential TCR diversity or repertoire. Notably, apart from responses to TCR activation, naïve CD8 T cells of neonatal mice and humans have been found to exhibit increased TCR-independent effector responses after stimulation with innate cytokines (IL-12 and IL-18), which likely contribute to their role in pathogen defense through bystander responses (16). The heightened effector activity in neonatal CD8⁺ T cells in our studies was accompanied by considerable cell death, particularly in the absence of the STAT5-activating IL-7 and IL-2, which are known to support increased glucose uptake. This finding suggests that neonatal CD8⁺ T cells may require a nutrient-rich environment and appropriate antigen stimulation for sustained viability. In contrast, adult CD8⁺ T cells typically exhibit slower yet more enduring responses. The rapid effector responses observed in neonates may reflect an evolutionary adaptation that provides immediate defense against pathogens during a critical phase of immune maturation while limiting immunopathology via programmed cell death. However, the transient nature of these responses raises important questions regarding the durability of CD8⁺ T cell immunity in neonates, particularly in the context of persistent infections and vaccination strategies. Immunological memory is known to be impaired in neonatal peripheral T cells, although mucosal tissues accumulate significant memory cells during infancy and the first couple years of life (9). Further investigation is needed to understand the connection between these systemic and mucosal memory T cell responses as well as their mediators in early life.

Our analysis of key transcription factors known to influence CD8⁺ T cell differentiation revealed several key differences in neonatal CD8⁺ T cells compared to adult counterparts. Notably, key transcription factors such as TOX and HELIOS were found to be significantly elevated in neonatal cells, both at baseline and after activation. Importantly, the differences in TOX and HELIOS expression between neonatal and adult cells became more pronounced after activation, potentially implicating functional roles during antigen encounter. Notably, TOX is a key transcription factor in the exhausted CD8⁺ T cell lineage; however, although the elevated TOX expression in neonatal CD8⁺ T cells did correlate with elevation of some traditional markers of T cell exhaustion such as PD-1 (induced by TOX), other canonical and more specific exhaustion markers such as TIM-3 were not elevated (43). Moreover, elevated TOX was correlated with increased activation/effector phenotypes, including higher CD69 levels and enhanced production of TNFα, IL-2, and IFNγ. Of note, TOX can be expressed by potent polyfunctional effector (non-exhausted) T cells and has also been shown to be associated with lung immunopathology (44,(Kim, 2024 #358, 45, 46). Our data provide evidence that TOX-positive cells in neonates are polyfunctional and produce robust effector responses. Given that TOX and HELIOS are expressed at higher levels in recent thymic emigrants (RTEs), we aimed to mitigate the influence of RTE variability by utilizing samples from healthy adults under 30 y of age (47); nonetheless, the recent emigration from the thymus may imprint neonatal cells for unique responses, although, in contrast to our findings reported in neonatal CD8 T cells, RTEs have been described to be less potent effectors (48, 49). Although a functional role for HELIOS in neonatal CD8⁺ T cells remains to be further defined, recent research has identified a regulatory phenotype of HELIOS⁺ human CD8⁺ T cells that play a role in downregulating autoreactive CD4⁺ T cells and are characterized by the expression of killer cell immunoglobulin-like receptors (50), (51). HELIOS has also been found to enhance naïve fetal human CD4⁺ T cell differentiation into regulatory T cells (52). Other transcription factor expression levels were less robust and/or context-dependent, though we recognize small expression changes in transcriptional regulators can lead to larger effects on the transcriptome and cellular behavior. In particular, we observed that TCF7 was downregulated more in neonatal CD8⁺ T cells after activation than the adult counterparts. Previous studies of activated mouse and human CD4⁺ T cells have demonstrated a similar pattern, indicating a quick shift towards effector-like cell differentiation and away from naïve-like and memory cell differentiation in both CD4⁺ and CD8⁺ T cells of neonates (13, 53, 54). Moreover, TBET and EOMES showed significant elevation in neonatal human CD8 T cells, further supporting the characterization of a short-lived effector phenotype and a possible role for EOMES and TBET in driving the effector phenotypes observed. The unique transcriptional profile of neonatal CD8⁺ T cells underscores the potential roles of these fluctuating transcription factors in facilitating efficient yet transient pathogen defense responses.

Taken together, the observed activation and transcriptional profiles of early-life CD8 T cells may play crucial roles in modulating immune responses to pathogens during the phase when passive immunity confers an additional layer of defense from infection. Investigating how these profiles evolve in relation to both favorable and unfavorable outcomes of viral infections in infants could yield valuable insights for developing improved therapeutic strategies for this vulnerable population. A comprehensive understanding of the regulatory networks governing T cell responses in early life may also inform strategies to enhance vaccine efficacy and develop treatments for infectious diseases that disproportionately affect neonates.

While our study provides critical insights, it is important to acknowledge its limitations. The experimental work with primary human CD8⁺ T cells was conducted in a controlled laboratory setting. We did not assess T cell responses to cytokine-induced, TCR-independent activation as has been addressed in recent reports, as our focus here was on TCR-driven responses. While we observed trends toward increased CD107a/LAMP1 and IFNγ levels, our sample size may have limited our ability to detect significant changes in these activation markers. Moreover, the immune cell populations from cord blood are known to rapidly change postnatally as a whole, but naïve CD8⁺ T cells have features that remain persistent (55), and we therefore did all our studies on purified naïve CD8 T cells isolated from full-term neonatal cord blood and compared to the same population sorted from 18–30-y-old adults. Future studies should aim to validate these findings in larger cohorts and diverse populations, with an emphasis on longitudinal assessments during the first months and years of life. Additionally, it is crucial to consider genetic and environmental factors that may influence immune development, including the important contributions of the microbiome (56), metabolism, and signal integration pathways, which all have implications for health, disease prevention, and treatment in early life.

In conclusion, early-life human CD8⁺ T cells exhibit a unique profile characterized by rapid, short-lived effector responses and a distinct transcription factor landscape. These findings deepen our understanding of neonatal and infant immunity and underscore the necessity for tailored immunization and therapeutic strategies for this vulnerable population.

Methods

Human subjects and sample collection

This study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board at Yale. Informed consent was obtained from all participants prior to sample collection. Peripheral blood was collected from healthy young adult volunteers less than 30 y old at Yale collected in Acid Citrate Dextrose (ACD) blood collection tubes and matched with cord blood obtained from full term C-section-delivered neonates collected in ACD blood collection tubes. Exclusion criteria included any neonates born prematurely or with major maternal/fetal comorbidities or maternal fever. No samples were used from neonates whose placentas needed to be examined by clinical pathology for any reason. Adult healthy donors were excluded if they were over 30 y old, if they had any serious illness, immune system related illness, or need for immune system modifying medications, such as steroids. The Yale University Reproductive Sciences Biobank provided cord blood samples. Neonatal and adult blood sample pairs were matched by the same date of blood collection, as well as simultaneous processing, stimulation, staining and analysis. Each experiment included one neonatal-adult sample pair, or at most four pairs used simultaneously when several fresh samples were available on the same day. Whenever possible, samples were sex matched. Additional paired young adult and full-term cord fresh blood samples were obtained, respectively, from the New York Blood bank shipped overnight in citrate phosphate dextrose, and from the New York Cord Blood bank shipped overnight in phosphate dextrose solution. Post-natal blood was collected by capillary blood drawn into ethylenediaminetetraacetic acid (EDTA) tubes and paired with matched samples from peripheral adult and cord blood collected into EDTA tubes. Blood samples were processed within 24 hours of collection.

Blood processing

Peripheral blood mononuclear cells (PBMCs) were isolated using density gradient centrifugation. Blood samples were diluted 1:1 with phosphate-buffered saline (PBS) and carefully layered onto lymphoprep medium (StemCell) in a 50 mL conical tube. Following centrifugation for 20 min at 20°C at 824 Relative Centrifugal Force (RCF/G) with no deceleration, the PBMC layer was collected and washed twice with complete RPMI media. For the isolation of CD8⁺ T cells, PBMCs were subjected to positive bead selection using CD8⁺ microbeads (Miltenyi Biotec) according to the manufacturer’s instructions. Isolated naïve CD8⁺ T cells were then sorted using a BD fluorescence-activated cell sorting (FACS) Aria flow cytometer (BD Biosciences) to obtain live naive cells defined as DAPI⁻CD8⁺CD45RA⁺CCR7⁺. For a subset of experiments, including the dye dilution assays, the Miltenyi naïve CD8⁺ T cell isolation kit (130–093-244) was used to isolate naïve CD8⁺ T cells from ficoled PBMCs from adult and cord blood pairs. Cells were resuspended in complete Roswell Park Memorial Institute medium (cRPMI)-1640 supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin, 1x GlutaMAX (Gibco) and counted using a BioRad TC20 cell counter.

For scRNA sequencing of cord blood and postnatal infant blood samples, whole blood samples were collected in BD Microtainer^® MAP K2EDTA tubes and stored at 4°C until processing. Within 12 hours of collection, serial RBC lysis was performed as previously described (33). Up to 250 microliters of blood is placed in a 50 mL conical and 10 mL of red blood cell lysis buffer (eBioscience) was added for 15 minutes at room temperature. Next, 40 mL of Dulbecco’s Phosphate Buffered Solution (DBPS, Gibco) was added, samples were centrifuged at 1500 RPM at 20C for 5 minutes and the supernatant was discarded. RBC lysis steps were repeated until there were no visible red blood cells. The pellet was cryopreserved in 10% Dimethyl Sulfoxide (DMSO) in Fetal Bovine Serum (FBS) at −80°C Celsius for 24 hours and then in liquid nitrogen until batch processing/analysis.

CD8⁺ T cell culture

CD8⁺ T cells were cultured in 96-well plates coated with anti-CD3 antibodies (Clone: OKT3, BioLegend, Catalog no. 317326) at a concentration of 5 μg/mL (coated overnight at 4C and washed twice with PBS prior to cell plating). Soluble anti-CD28 (BioLegend, Catalog no. 302943, 1 μg/mL) was added to the cultures to provide a second signal for T cell activation and proliferation. Cells were maintained in complete RPMI-1640 media supplemented with 10% FBS and 1% penicillin-streptomycin and 200 units/mL IL-2 (Prometheus) for most experiments. At specified time points, cells were restimulated with PMA (500 ng/mL) and ionomycin (1 μg/mL) for four hours and cytokine secretion was blocked with 5 μg/mL brefeldin A (BioLegend, catalog no. 420601) and 2 μM monensin (BioLegend, Catalog no. 420701) two hours prior to intracellular cytokine quantification by flow cytometry. The cultures were incubated at 37°C in a humidified atmosphere containing 5% CO₂.

Viability Staining

Cells were stained with fixable viability stain 780 (FVS 780) (BD Biosciences, 1:2000 in PBS) and Human Trustain Fc block (BioLegend; 1:50) at a concentration of 100,000 cells in 50 microliters antibody mix for 10 minutes at room temperature prior to flow cytometry staining to detect viable cells and complete fc block.

Flow cytometry

For flow cytometric analysis, isolated naive CD8⁺ T cells were stained with a panel of fluorochrome-conjugated monoclonal antibodies specific for surface markers. The antibodies used included CD8-BV650 (BioLegend, catalog no. 344730), CCR7-PE (BD Biosciences, catalog no. 560765), CD45RA-FitC (BioLegend, catalog no. 304106), CD226-APC (BioLegend, catalog no. 338312), CD25-AF647 (BioLegend catalog no. 302618), CD69-PE (BioLegend catalog no. 310906), CD107-PerCP/Cyanine5.5 (BioLegend, catalog no. 328616), CD95/Fas-APC (BioLegend, catalog no. 305612), CD178/FasL (BioLegend, catalog no. 306417), CD161 (ThermoFisher Scientific, catalog no. 1–1619-42), PD1-Pacific Blue (BioLegend, catalog no. 329915), KLRG1-PE/Cy7 (BioLegend, catalog no. 368614), CD366/TIM-3 (ThermoFisher Scientific, catalog no. 78–3109-42). CCR7-PE was used at a concentration of 1:20 dilution and all other surface antibodies were used at a concentration of 1:100 dilution. Following surface staining for 30 minutes, cells were washed with FACS buffer (500mL PBS, 2.5g bovine serum albumin--BSA, 2 mL EDTA, 2mM, pH 8.5) and then fixed with 4% paraformaldehyde (BioLegend, catalog no. 420801). Cells were then permeabilized using perm/wash buffer (BioLegend, catalog no. 421002). Intracellular staining was performed using antibodies against intracellular cytokines (e.g., TNF-α, IFNg, and IL-2) according to the manufacturer's protocol. The following antibodies were used: TNFα-BV421 (BioLegend, catalog no. 502932), IL-2-Alexa Fluor 700 (BioLegend, catalog no. 500320), IFNγ -PE/Cy7 (BioLegend, catalog no. 506518). All intracellular antibodies were used at a concentration of 1:100 dilution. After staining, cells were washed and resuspended in FACS buffer. Analysis was performed on MACSQuant, and Cytek Northern Lights flow cytometer, and data were analyzed using FlowJo software (Tree Star, Inc.). Gates were set based on appropriate isotype and FMO controls.

For Transcription Factor staining, the cells were first stained with fixable viability stain 780, Fc block, and surface markers and then fixed/permeabilized with Foxp3 Fixation/Permeabilization (eBioscience, catalog no. 00–5523-00) buffer as per the manufacturer protocol. Cells were stained with 50 uL of conjugated antibody mixture at a concentration of 1:100 dilution with 1X Permeabilization Buffer for detection of transcription factors and incubated for 60 minutes at room temperature in the dark. Cells were washed with permeabilization buffer and resuspended in FACS buffer for flow cytometric analysis using Cytek Northern Lights flow cytometer. Antibodies used included: TOX-eFluor 660 (ThermoFisher Scientific, catalog number 50–6502-82), T-bet-PE-Cy5 (ThermoFisher Scientific 15–5825-82), EOMES-PE-CF594 (BD Biosciences, catalog no. 567167), TCF1-RB705 (BD Biosciences catalog no. 570635), Ki-67-Alexa Fluor 532 (Thermo Fisher Scientific, catalog no. 5699–82), Blimp-1-Alexa Fluor 647 (BD Pharmingen, catalog no. 565002), Helios eFluor 450 (ThermoFisher Scientific, catalog no. 48988341).

Proliferation Assay

Freshly isolated naïve CD8⁺ T cells from neonates and adult healthy donors (isolated with Miltenyi naïve CD8⁺ T cell selection magnetic beads) were incubated in two micromolar tag-it violet proliferation dye (BioLegend) in PBS for 20 minutes in a 37C incubator and then quenched with cRPMI prior to plating on anti-CD3 coated plates (coated overnight in 4C with 5 micrograms/mL of anti-CD3-OKT3 clone and washed twice in PBS) in complete RPMI media with soluble anti-CD28 (1 micrograms/mL) and 200 units/mL of IL-2 (Prometheus) for three days. Cells were plated at 100,000 cells per 100 microliters in flat 96 well plates. Flow cytometry evaluation of dye dilution was done using Cytek Northern Lights flow cytometer, and data were analyzed using FlowJo software (Tree Star, Inc.).

Bulk RNA sequencing

For RNA sequencing, total RNA was extracted from freshly isolated naïve CD8⁺ T cells (baseline) and naïve CD8 T cells that were activated with plated anti-CD3 and soluble anti-CD28 for 2 days and then re-stimulated with PMA and ionomycin for four hours (stimulated). The RNeasy Plus Mini Kit (Qiagen) was used following the manufacturer’s protocol. RNA quality was assessed using a Bioanalyzer (Agilent Technologies), and only samples with an RNA Integrity Number (RIN) > 7 were used for library preparation. Libraries were constructed using the TruSeq RNA Sample Preparation Kit (Illumina) and sequenced on NovaSeq. Sequencing read length used was 100 base pairs. 25 million read pairs were pooled. 150 paired-end fragments were trimmed using fastp software (57) and were mapped to the human transcriptome with Gencode annotation GRCh38 (58) using STAR (59) and in parallel transcript level expression was quantified using SALMON (60). Gene-length normalized counts were then used to generate batch corrected counts using ComBat-seq (61)package in R. Gene level differential expression analysis was performed using DESeq2 (62) package in R. Subsequent plots were generated using DEGs derived from DESeq2: GSEA (63) calculated using fast gene set enrichment analysis (FGSEA) (64) and clusterProfiler (65) in R, and heatmaps using pheatmaps (66) in R.

Single-cell RNA sequencing analyses

Analyses were performed on processed published data as indicated in the figure legends. Data were analyzed using the Scanpy toolkit (67) following published recommended standards. Before downstream analysis, low quality cells and genes were filtered out, only genes expressed in more than 3 cells and cells with more than 200 genes and less than 20% mitochondrial reads were included in the analysis. Afterwards, data were normalized (scanpy.pp.normalize_per_cell, scaling factor = 10⁴) then log-transformed (scanpy.pp.log1p). Batch correction was performed using bbknn (v1.5.1) (68). Highly variable genes were used to produce the principal component analysis (PCA). Dimensionality reduction uniform manifold approximation and projection (UMAP) and Leiden clustering was performed for clustering analysis.

For signature scoring, published gene sets from the MSigDB were used to compute average scores using the scanpy.tl.score_genes function across cells in each age group as specified in the figure legends. Scores were plotted as box plots across independent biological samples/donors in each age group.

For differential gene expression analyses, the limma package (69) was used to identify genes with false discovery rate (FDR) < 0.05 between the compared age groups as specified in the figure legends. Top genes were displayed using dot plots showing normalized gene expression across the compared donors or groups.

Seahorse glycolytic stress assay

Metabolic profiling of naïve CD8 T cells was performed using the Seahorse XF Analyzer (Agilent Technologies). A XF96 cell culture plate was coated with poly D lysine overnight. Freshly isolated naïve CD8⁺ T cells were plated at a density of 500,000 cells in 180 microliters per well in a XF96 cell culture plate and stimulated with Dynabeads (Dynabeads^™ Human T-Activator CD3/CD28 for T Cell Expansion and Activation, Gibco) at a 2:1 ratio in glucose-free Dulbecco's Modified Eagle Medium (DMEM) (D5030 Sigma), supplemented with 2mM glutamine, 1mM sodium pyruvate, 5mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) and 0.2% free fatty acid-free BSA. Within 30 minutes after adding Dynabeads, cell plates were loaded on the XF analyzer. Following baseline measurements, glucose (final 10 mM) was added to assess glycolytic capacity, followed by the injection of oligomycin (final 5 μM) to evaluate adenosine triphosphate (ATP)-linked respiration, 2-deoxyglucose (2-DG; final 50 mM) to assess glycolytic inhibition, and finally, rotenone and antimycin A (5 μM and 10 μM, respectively) to inhibit cellular respiration. Extracellular acidification rate (ECAR) was measured throughout the assay. Data were analyzed using Wave software (Agilent Technologies) to assess mitochondrial and glycolytic function.

Serum collection and protein measurements

Serum samples were collected from full term healthy human neonatal cord blood and from adults following standard venipuncture procedures. Blood was allowed to clot for 30 minutes at room temperature, followed by centrifugation at 1,500 × g for 10 minutes to separate the serum from cellular components. The serum was then aliquoted and stored at −80°C until analysis. Human serum concentrations of cytokines and chemokines were measured using the human cytokine chemokine 71 plex discovery assay (Eve Technologies).

Transmission Electron Microscopy (TEM)

400,000 naïve CD8⁺ T cells from each donor (adult <30 years old and full term neonatal) were used at baseline or after 96 hours of culture in 24-well cell culture plates coated with anti-CD3 (Clone: OKT3, BioLegend, Catalog no. 317326) at a concentration of 5 μg/mL (coated overnight at 4C and washed twice with PBS prior to cell plating). Soluble anti-CD28 (BioLegend, Catalog no. 302943, 1 μg/mL) was added to the cultures to provide a second signal for T cell activation and proliferation. Cells were maintained in complete RPMI-1640 media supplemented with 10% FBS and 1% penicillin-streptomycin. Cells were pelleted and then fixed with 2.5 % Glutaraldehyde, 2% Paraformaldehyde, buffered in 0.1 M Sodium cacodylate at a pH of 7.4. After rinsing with 0.1 M Sodium cacodylate, the cells were post-fixed in 1% Osmium tetroxide, 0.8% Potassium cacodylate, buffered in 0.1 M Sodium cacodylate buffer. The T-cells were pelleted and resuspended in 2% molten Agarose. Cells underwent en bloc staining in 2% Uranyl acetate, rinsed in water and dehydrated via an Ethanol series. The T-cells were embedded in EPON. Once polymerized, 60 nm-thick sections were cut on a Leica EM UC7 Ultramicrotome (Leica Microsystems Inc., Deerfield, IL, USA), mounted on 200-mesh Nickel grids and post-stained with 2% Uranyl acetate followed by Reynold's Lead citrate. Samples were imaged on a Tecnai BioTwin G2 Spirit (ThermoFisher Scientific, Hillsboro, OR USA) operating at 80 kV. Images were acquired at varying magnifications using a SIS Morada 11-megapixel CCD camera.

TCR data analysis

Single-cell TCR sequences were aligned and quantified using cellranger vdj (10x Genomics) pipeline against the human GRCh38 VDJ reference genome. Filtered annotated contigs of TCRs were analyzed using the python Scirpy package (v0.10.1) (70) after which TCR data were integrated with gene expression data. Cells without a full TCR αβ chain pair were excluded. TCR clonotypes were defined based on CDR3 nucleic acid sequence identity using scirpy.tl.define_clonotypes. Clone diversity curves that measured Hill’s diversity metric across diversity orders (q) 0–4 was created using the R package alakazam (v1.2.1) with the alphaDiversity function using the same minimum depth across samples (71). The general diversity index (qD) proposed by Hill, which includes a range of diversity measures as a smooth curve over a single varying parameter q. Special cases of this general index of diversity correspond to the most popular diversity measures: species richness (q = 0), the exponential Shannon-Weiner index (as q =1), the inverse of the Simpson index (q = 2), and the reciprocal abundance of the largest clone (as q→∞). Mean ²D represents the mean Simpson diversity index (q=2) across all bootstrap realizations within each dataset. Higher Mean ²D values suggest greater clonal diversity. SD: Standard Deviation. Mean ²D ± SD is the mean clonal diversity for an individual sample, with error bars indicating the variation in diversity across the sample.

Statistical analyses

Each experiment was done on one or more pairs of adult and cord blood specimens processed fresh (within 24 hours of blood collection), tested and analyzed at the same time. Sex matched samples were used, when possible, for each experiment, and statistical analysis was performed by quantifying the delta mean fluorescence intensity (MFI) or delta percent. Supplemental figures also depict actual MFI or percent values to complement the normalized delta comparisons in the main figures. Delta MFI or delta % is the neonatal or adult value minus the adult value for each paired sample. Neonatal and adult blood samples were paired based on their matched processing with the same date of blood collection, as well as simultaneous processing, stimulation, staining and analysis. Comparisons were conducted using the Mann-Whitney test. Statistical analyses were performed using GraphPad Prism (GraphPad Software). A p-value of <0.05 was considered statistically significant. Bar graph data are presented as mean ± error of the mean or mean ± standard deviation as labeled. For scRNA-seq analyses, boxes in the box plots denote the interquartile range (IQR), horizontal bars represent the median, whiskers extend to 1.5 × IQR, and dots show the values of each sample. For signature scores, statistical significance was determined using the unpaired two-sided Wilcoxon rank-sum test. Differentially expressed genes were filtered using FDR < 0.05. For bulk RNA-seq analysis, DEGs are filtered using FDR ≤ 0.05, log2FC ≥ |0.5|, and gene-length normalized mean counts >10.

Significance:

During the trajectory from fetal to neonatal to infant life, naive CD8⁺ T cells are developmentally tuned through incompletely understood mechanisms. Here, we report that naïve neonatal human CD8⁺ T cells are poised for an immediate effector switch within hours of activation, including significantly increased glycolytic capacity, proliferation and cell death, differentiation, and production of effector molecules such as TNFα. Coupled with the distinctive transcriptomes of naïve neonatal CD8⁺ T cells is a unique set of transcription factors they express, which includes elevated TOX and HELIOS in both the resting and activated states. These distinctive features persist into infancy, underscoring their potential implications for shaping early-life responses to infections and vaccination while avoiding tissue immunopathology.

Data and materials availability:

Newly generated bulk RNA sequencing data are available as GSE289647 at Gene Expression Omnibus (GEO) (72). The raw scRNA-seq, TCR-seq, and processed data matrices are available at GEO accession GSE271413, and data from full-term infants at 2 mo were extracted from GSE204716 (73–74). All code used in analysis of scRNA-seq and TCR data is available at 10.5061/dryad.pk0p2ngxg (75).