Alterations in the Thymic Selection Threshold Skew the Self-Reactivity of the TCR Repertoire in Neonates

The developing neonatal immune system is distinct from that of adults, with neonates displaying a more limited ability to develop cellular immunity in response to pathogen challenge (1, 2). Within the T cell compartment, in particular, cell-intrinsic differences in neonatal versus adult T cells account for many of their functional differences (3–7). These have often been attributed to the immaturity of the T cell repertoire in neonatal mice because their peripheral lymphoid organs are disproportionately composed of recent thymic emigrants (RTEs) (8, 9). However, gene expression, epigenetic, and functional differences have also been identified between both thymocytes and RTEs from neonatal and adult mice, suggesting the distinct functions of neonatal T cells may, in fact, arise during thymic development (6, 8, 10, 11). Such differences may be due, at least in part, to the unique stem cell populations that give rise to T cell progenitors at different stages of ontogeny (10, 12–14). Alternatively, it is increasingly appreciated that the quantity and quality of TCR signals that accompany thymic selection not only influence lineage fate choice during development, but also subsequent T cell function in the periphery (15–18). Thus, it is possible that variations in thymic selection impact the self-reactivity and, consequently, the function of the T cell repertoire during ontogeny.

A rather broad range of TCR affinities for self-peptide presented by MHC (self-pMHC) molecules promote positive selection to support the survival and differentiation of developing T cells (19, 20). This leads to significant variation in the subthreshold levels of self-reactivity between individual cells within the mature T cell population. The cell surface level of CD5, a negative regulator of TCR signaling, correlates with the strength of the signals obtained through the TCR that accompany thymic selection (21, 22). In the peripheral lymphoid organs, the relative levels of CD5 are maintained via subthreshold interactions between the TCR and self-pMHC. These trophic signals are important for T cell survival and trigger cell division in lymphopenic environments where CD5^hi T cells outcompete their CD5^lo counterparts (23–26). In addition, there is emerging evidence that the self-reactivity of a T cell significantly impacts its contribution to the T cell response during acute infection (15–18). Several attributes of CD5^hi T cell populations are characteristic of the neonatal T cell phenotype, including elevated effector differentiation molecule expression in CD5^hi CD8⁺ T cells (e.g., T-bet and Eomes) and the peripheral bias of CD5^hi CD4⁺ T cells to preferentially differentiate into regulatory T cells (T_regs) (6, 18, 27, 28).

A number of distinct cell-intrinsic and -extrinsic mechanisms may differentially influence thymic selection to shape the TCR repertoire during development in the neonatal and adult thymus. In addition to the aforementioned stem cell populations that seed the neonatal and adult thymus from the fetal liver and bone marrow, respectively, a paucity of TdT activity in developing T cells from neonatal mice restricts the diversity of the TCR pool by ∼90% because of a lack of template-independent nucleotide additions at the junctions of recombined TCR gene segments and results in a germline biased repertoire (29, 30). In addition, cell-extrinsic, microenvironmental cues have been linked to differences in T cell development throughout ontogeny. For example, age-dependent changes in the peptide repertoire, Ag processing, and Ag presentation are thought to influence negative selection and the development of a distinct T_reg population with unique tolerogenic properties in neonatal mice (31, 32). Therefore, multiple differences in neonatal and adult thymocytes and their thymic microenvironments may impact the TCR repertoire and its self-reactivity at different stages of ontogeny.

In this study, we sought to determine age-dependent differences in TCR repertoire selection. Our studies show that differences in thymic selection contribute to the altered self-reactivity of the T cell repertoire in neonates and adults, and that stronger TCR signals accompany both conventional T cell and T_reg selection in neonatal and young mice. Thymocytes bearing TCRs with low affinity for self-pMHC are inefficiently selected and, as a consequence, are ultimately depleted from the mature T cell population in neonatal and young mice. Thus, the CD5^hi T cell repertoire in neonatal mice is poised to undergo more robust homeostatic proliferation to fill the lymphopenic neonatal compartment and may influence T cell effector function.

C57BL/6, B6.SJL, and Nur77-GFP (33) mice were purchased from The Jackson Laboratory. The generation of Nur77-GFP (33), Rag2p-GFP (34), TdT^−/− (30), OT-I Rag1^−/− (35, 36), P14 TCRα^−/− (37, 38), and HY^cd4 (39) mice has been previously described. Both male and female mice were included in each experiment unless otherwise noted. Mice were maintained in a specific pathogen-free environment at the Maisonneuve-Rosemont Hospital Research Center, University of Alberta, or McGill University animal facilities. All animal protocols were approved by the local Animal Care Committee in accordance with the Canadian Council on Animal Care guidelines.

Thymic and peripheral lymphoid tissues were harvested from neonatal and young mice from 3 to 21 d after birth and from adults that were 6–12 wk old. Single-cell suspensions were prepared manually with glass tissue homogenizers followed by RBC lysis with Ack lysis buffer (0.15 M NH₄Cl, 10 mM KHCO₃, 0.1 mM Na₂EDTA) before use.

Human peripheral blood and cord blood were obtained from donors after informed consent and in accordance with Research Ethics Committee guidelines at the Maisonneuve-Rosemont Hospital. Mononuclear cells were prepared with Ficoll (GE Healthcare) and cryopreserved until use.

Anti-murine CD4 (GK1.5), CD8α (53-6.7), TCRβ (H57-597), TCR Vα2 (B20.1), CD5 (53-7.3), CD25 (PC61), CD69 (H1.2F3), CD44 (IM7), CD62L (MEL-14), GITR (DTA-1), PD-1 (RMP1-30), CD11b (M1/70), and CD11c (N418), and anti-human CD4 (OKT4), CD8α (RPA-T8), TCRαβ (IP26), CD5 (UCHT5), CD45RA (HI100), CD62L (DREG-56), CD103 (Ber-ACT8), and CD31 (WN59) Abs, as well as Live/Dead Zombie, were purchased from BioLegend; anti-Foxp3 (150D/E4) and HY-TCR (T3.70) Abs were purchased from eBioscience; and anti-murine CD19 (1D3) was purchased from BD Biosciences. Murine splenocytes and human cells were first incubated with 2.4G2 supernatant or Human TruStain FcX (BioLegend), respectively, for 10 min at 4°C before cell surface staining. Cells were incubated with Live/Dead Zombie according to the manufacturer’s protocol followed by fluorescently labeled cell surface Abs for 20 min at 4°C. For intranuclear staining, cells were fixed and permeabilized using a Foxp3 Staining Kit (eBioscience) according to the manufacturer’s protocol. Cells were acquired on an LSRFortessa X-20 flow cytometer using FACSDiva software (BD Biosciences). Analyses were performed using FlowJo software (Tree Star) on live cells.

Spleens from neonatal mice 3–7 d after birth, as well as male and female adults, were harvested, and single-cell suspensions were prepared as described. The sex of neonates was confirmed by PCR of tail biopsies before cell sorting as previously described (40). Naive T cells (CD19⁻CD11b⁻CD11c⁻CD44⁻) were negatively enriched on a FACSAria III (BD Biosciences), and >95% purity was confirmed by subsequent flow cytometry analysis of TCRβ, CD4, and CD8 of sorted cells. Sorted, naive T cells were then labeled with 1 μM CFSE (eBioscience) according to the manufacturer’s protocol. A total of 5 × 10⁵ CFSE-labeled cells were transferred i.v. into the lateral tail vein of sex-matched adult mice that were sublethally irradiated (600 rad) 1 d before injection. Spleen and lymph nodes of adult recipient mice were harvested 7 d after reconstitution, and dilution of CFSE was assessed by flow cytometry. The proliferation index was determined using FlowJo software (Tree Star) and is defined as the average number of divisions of all cells that have divided.

Statistical analysis was performed using GraphPad Prism (GraphPad Software). An unpaired two-tailed t test was performed when comparing two conditions with normal distribution, and an unpaired two-tailed Mann–Whitney U test was used when comparing two samples with nonnormal distribution. A one-way ANOVA and post hoc Tukey test was applied when comparing neonates of different ages with adults. Statistical significance is indicated by p values: *p <0.05, **p <0.01, ***p <0.001, ****p <0.0001.

It has previously been posited that in adults, CD4⁺ T cells bearing TCRs with high sensitivity for self-pMHC, and thus displaying greater reactivity to foreign Ags, are depleted from the naive population over time as they are preferentially recruited to the memory compartment (16). Accordingly, it was observed that the peripheral naive CD4⁺ T cell pool in neonatal day 5 mice is enriched for TCRs with greater self-pMHC reactivity than their adult counterparts (16). Thus, we first sought to determine the extent of differences in the self-reactivity of the peripheral T cell pool in neonatal versus adult mice. To do so, we analyzed cell surface expression of CD5 on polyclonal wild type (WT) splenic T cells by flow cytometry. Both naive CD4⁺ and CD8⁺ T cells from neonatal (days 3 and 7) and young mice (days 14 and 21) consistently expressed higher levels of CD5 as compared with their adult counterparts (Fig. 1A, 1B), and no substantial sex bias in CD5 levels influenced these results (Supplemental Fig. 1). To complement these studies, we analyzed Nur77, an immediate early gene downstream of Ag receptor stimulation in which the level of induction correlates with TCR signaling, using Nur77-GFP reporter mice (33, 41, 42). We found consistently higher levels of GFP, indicating elevated Nur77 expression, in naive peripheral T cells from young mice as compared with adults, corroborating the analysis of CD5 (Fig. 1C). These observations suggest that the peripheral T cell repertoire in neonates and young mice is more self-reactive than that of adults.

In a functional context, CD5 levels also correlate with the extent of lymphopenia-induced proliferation of CD4⁺ and CD8⁺ T cells (23–26). Thus, to assess whether the small differences observed in CD5 and Nur77-GFP levels between the neonatal and adult T cell repertoires have a physiological relevance, we enriched naive T cells from neonatal (days 3–7) and adult mice, labeled them with a proliferation dye, and transferred them into sublethally irradiated, sex-matched adult recipients to assess their proliferation in a lymphopenic environment. We observed that neonatal naive CD4⁺ and CD8⁺ T cells underwent more robust proliferation in a lymphopenic environment as compared with adult naive T cells (Fig. 1D, 1E), consistent with the finding that neonatal T cells express higher levels of CD5 than their adult counterparts and suggestive of physiologically relevant differences in self-reactivity.

The neonatal peripheral naive T cell compartment is distinct from that of adults and consists mainly of RTEs with increased sensitivity to TCR stimulation (8, 9). Therefore, it was possible that the observed differences in CD5 levels in young versus adult mice not only reflected biased entry into the memory compartment, but also reflected the decreased RTE composition of the peripheral T cell repertoire with age. To examine this, we sought to directly compare CD5 levels on RTE populations in neonatal and young mice with those in adults. We used reporter mice in which GFP expression is driven by the RAG2 promoter during T cell development such that GFP expression in peripheral T cells indicates recent thymic export (34, 43). We observed an elevated pattern of CD5 expression on splenic GFP⁺ CD4⁺ and CD8⁺ RTEs from neonatal and young mice as compared with their adult counterparts (Fig. 1F). Together, these data suggest that the differences in the reactivity of the TCR repertoire for self-pMHC between young and adult murine T cells may, at least in part, derive from differences in T cell development.

We next sought to determine to what extent our observation of differences in CD5 levels on T cells from neonatal and young mice versus adults were conserved in humans. To do so, we isolated mononuclear cells from human cord blood and adult peripheral blood samples and assessed CD5 levels on CD4⁺ and CD8⁺ T cell populations. Indeed, we observed higher levels of CD5 on both mature naive CD4⁺ and CD8⁺ T cells as well as on cell populations enriched for RTEs in cord blood as compared with adult peripheral blood samples (Fig. 2) (44, 45). These results suggested conserved differences in the CD5 levels on T cells from young mice and humans as compared with their adult counterparts and prompted further evaluation of the thymic origin of these differences.

Because of the differences that were identified in CD5 levels on RTEs isolated from neonatal versus adult mice and humans, we sought to determine whether differences in TCR avidity for self-pMHC between neonates and adults originated during T cell development in the thymus. We assessed both CD5 and Nur77-GFP levels in thymocytes from neonatal and young mice in comparison with adults. Our results showed that the differences in CD5 levels on both CD4⁺ and CD8⁺ single-positive (SP) thymocytes were maintained for up to 3 wk after birth, with a peak after the first week (Fig. 3A, 3B). In addition, although still significantly different, CD5 levels decreased over time and, in 3-wk-old mice, began to approach levels observed in adult mice. Consistent with these data, CD4⁺ and CD8⁺ SP thymocytes from neonatal and young Nur77-GFP reporter mice displayed elevated GFP reporter signal as compared with their adult counterparts except for CD8⁺ SP thymocytes from day 3 neonates (Fig. 3C). Together, these observations suggest that differences in thymic selection between young and adult mice influence the reactivity of the TCR repertoire for self-pMHC in these populations.

It has previously been suggested that negative selection is inefficient in neonates and may therefore explain the higher levels of CD5 in young versus adult mice (46, 47). Thus, we reasoned that if differences in the fluorescence intensity of CD5 between young and adult mice were primarily influenced by defects in negative selection, then the breadth of the distribution of CD5 levels on the T cell population would increase to include T cells bearing TCRs of higher affinity. However, analysis of the coefficient of variation to measure the distribution of CD5 fluorescence suggested that CD5 levels on mature CD4⁺ and CD8⁺ SP thymocytes were not broader in neonatal and young mice versus adults (Fig. 3D). These results suggest that the neonatal repertoire lacks T cells bearing low-affinity TCRs, and the TCR signals accompanying thymic selection of conventional T cells are increased.

The strength of TCR signaling during thymic selection can influence T lineage differentiation. Thus, we examined whether the higher levels of CD5 on developing T cells impacted selection of T_reg populations. Although we did not identify significant differences in the proportion of thymic T_regs in day 7 postnatal mice versus adults (Fig. 4A), interpretation of this observation is complicated by the fact that T_reg development is delayed relative to conventional T cell development in neonates (48, 49). However, comparison of cell surface CD5 levels in the emerging CD4⁺CD25⁺Foxp3⁺ T_reg compartment in neonates with T_regs in the adult thymus revealed that neonatal thymic T_regs had significantly higher cell surface CD5 expression than their adult counterparts, suggesting that the strength of TCR signaling during T_reg selection may also be increased in neonatal mice relative to adults (Fig. 4B). In light of recent evidence that subpopulations of T_regs with nonoverlapping functions differ in their affinity to self-antigen as indicated by their CD5 levels (50), we sought to determine whether the neonatal thymic T_reg compartment also contained these subpopulations and whether differences in T_reg composition led to the observed differences in CD5 levels. The proportion of GITR^hiPD-1^hiCD25^hi (Triple^hi) cells, those with higher affinity for self-antigen, within the T_reg compartment was not significantly changed, whereas the GITR^loPD-1^loCD25^lo (Triple^lo) T_reg population was significantly higher in day 7 neonates as compared with adults (Fig. 4C, 4D). Therefore, changes in the frequency of T_reg subsets do not account for the differences in overall CD5 levels between day 7 neonates and adults. Instead, CD5 levels on both the Triple^lo and Triple^hi T_reg populations from neonatal mice were significantly higher than their adult counterparts (Fig. 4E). Together, these results suggest that TCR signals are increased in both conventional and T_reg populations from neonatal mice relative to adults.

The neonatal TCR repertoire is distinct from that of adults (51). This is attributable, in large part, to differences in the expression of TdT (30, 52). There is evidence of increased TCR cross-reactivity in TdT-deficient T cells, and greater cross-reactivity has been associated with higher self-pMHC reactivity (16, 53, 54). Thus, it was possible that the more germline biased TCR repertoire in neonates could account for the observed differences in the affinity of the TCR repertoire to self-pMHC between neonates and adults. To address this possibility, we compared CD5 levels on mature CD4⁺ and CD8⁺ SP thymocytes in adult WT and TdT^−/− mice (55). However, we did not observe an increase in CD5 levels on mature CD4⁺ and CD8⁺ SP thymocytes from adult TdT^−/− mice as compared with WT controls (Fig. 5). Therefore, higher CD5 levels observed on neonatal T cells do not appear to be solely due to their low levels of TdT expression and restricted TCR repertoire.

Having excluded the more germline TCR repertoire as the main factor leading to the global increase of CD5 on T cell populations in neonates and young mice, we hypothesized that their threshold for positive selection was increased relative to adults. Thus, we would predict that thymocytes bearing TCRs with relatively low reactivity for self-pMHC that are positively selected in adults would die by neglect in younger mice. To test this, we reasoned that transgenic mouse models in which developing T cells express a defined TCR with known, high self-pMHC reactivity would be efficiently selected in neonates, whereas those with low self-reactivity would not. MHC class I–restricted OT-I transgenic thymocytes express a TCR with high affinity for endogenous, positive selecting self-peptide relative to polyclonal adult CD8⁺ T cells (18), and a population of mature transgenic TCR^hi OT-I CD8⁺ SP thymocytes was detected in young OT-I transgenic mice (Fig. 6A, 6B). Although there was a lower proportion of mature CD8⁺ SP OT-I thymocytes in day 7 mice as compared with their adult counterparts (∼2-fold difference), this difference was less than observed in WT polyclonal mice at this age (∼3-fold lower than adult) and may be explained by the emergence of T cells that begins at birth and does not reach adult levels until after 3 wk of age (Supplemental Fig. 2). In addition, no significant differences in CD5 expression levels were observed on OT-I CD8⁺ SP thymocytes between neonates and adults (Fig. 6C). To confirm these observations, we examined a second MHC class I–restricted TCR transgenic mouse model, P14, with relatively high self-peptide reactivity (18). Again, we observed development of mature P14 CD8⁺ SP thymocytes in day 7 mice (∼1.2-fold lower than their adult counterparts) and similar CD5 levels between the P14 CD8⁺ SP cells in neonates and adults (Fig. 6D–F). Collectively, these data suggest efficient selection of thymocytes bearing TCRs with relatively high reactivity for endogenous, positively selecting self-ligands in young mice and that CD5 levels on neonatal T cells are not intrinsically higher than on corresponding adult T cell populations. In contrast, in the MHC class I–restricted HY^cd4 transgenic model in which the transgenic TCR has low affinity for positive selecting self-peptides (18, 39), there was a striking absence of mature HY TCR^hi (T3.70^hi) CD8⁺ SP thymocytes in female mice at 3, 5, and 11 d after birth relative to adults (∼20-, ∼12-, and ∼45-fold fewer mature HY TCR^hi CD8⁺ cells in 3-, 5-, and 11-d-old mice as compared with adults, respectively) (Fig. 6G, 6H). Notably, the few mature HY CD8⁺ SP thymocytes that did develop expressed similar levels of CD5 as their adult counterparts (Fig. 6I). However, negative selection of male Ag-specific HY^cd4 transgenic cells appeared normal in neonates, as evidenced by double-positive (DP) dulling, as well as upregulation of CD69 on DP thymocytes that was similar in both neonatal and adult male mice (Fig. 6J, 6K), confirming the absence of generalized defects in negative selection in neonates and young mice. Together, these data support the idea that thymocytes bearing TCRs with high affinity for selecting ligands undergo positive selection in neonates, whereas those expressing TCRs with low affinity for self-peptide are not efficiently selected, leading to a skewed TCR repertoire in neonatal mice with higher self-reactivity.

Despite well-characterized differences in neonatal versus adult T cell development, differences in thymic selection throughout ontogeny have not been thoroughly examined. We report in this article that apparent changes in the thymic selection threshold in neonatal and young versus adult mice significantly influence the self-reactivity of the TCR repertoire. Our observations that thymocytes bearing low-affinity TCRs drop out of the polyclonal T cell pool in neonatal mice, coupled with evidence of normal negative selection and the observation that CD5 levels on conventional T cells and T_regs is higher relative to that of adults, suggests a global shift in thymic selection thresholds with significant implications in the contribution of neonatal T cells to an immune response.

The affinity threshold for thymic selection has been meticulously defined using TCR transgenic models with variable affinity for altered peptide ligands in the context of fetal thymic organ culture (19). However, consistent with our results that suggest thymic selection thresholds may be increased in adults relative to neonatal and young mice, recent studies have demonstrated evidence of negative selection of adult thymocytes in response to low-affinity Ags that had been reported to support positive selection in the fetal thymic organ culture model (19, 56, 57). In the later studies, however, low-affinity peptides were restricted to dendritic cells or presented in vitro via tetramers, conditions that do not support positive selection. Thus, it will be important to further define thymic selection boundaries within a model of adult thymopoiesis that supports both positive and negative selection.

The graded downregulation of CD5 on CD4⁺ and CD8⁺ SP thymocytes over time from the peak at day 7 may reflect either an increasing ratio of adult- versus fetal-derived progenitor cells contributing to development or an evolving thymic microenvironment. Indeed, fetal liver- versus bone marrow–derived progenitors that seed the thymus appear to significantly impact lymphoid development and function (13, 14, 58–60). Our observations that CD5 levels are similar on neonatal versus adult OT-I, P14, and HY TCR transgenic cells, but that HY TCR transgenic cells fail to develop efficiently early in life, suggest that the differences observed in the affinity of the neonatal TCR repertoire for self-antigen are a consequence of altered thymic selection rather than global, cell-intrinsic increases in CD5. We cannot rule out, however, progenitor cell–intrinsic modulation of the thymic selection threshold. Alternatively, differences between the neonatal and adult thymic environments, particularly as they relate to altered peptide processing and presentation machinery, may also play a role in neonatal versus adult TCR repertoire selection. Given the suggested role of the thymus-specific proteasome subunit, β5T, in supporting the development of MHC class I–restricted T cells bearing weakly self-reactive TCRs, it is possible to imagine that differences in the peptide repertoire alone could selectively eliminate thymocytes bearing low-affinity Ag receptors (61–63). However, this would not necessarily explain the overall increases in TCR signal strength that accompany selection of both conventional T cells and T_regs. Resolution of this issue awaits further investigation into the cell-intrinsic or -extrinsic mechanisms that skew the sensitivity of the neonatal TCR repertoire.

Our findings demonstrate a shift in the sensitivity of a TCR for self-pMHC required for thymic selection in neonates that may provide advantages to a neonatal TCR repertoire that is severely restricted in terms of both T cell number and diversity. In the CD4⁺ T cell lineage, in particular, increased subthreshold self-reactivity has been linked to higher affinity for foreign Ag, as well as more promiscuous binding to distinct foreign pMHC complexes (16, 18, 64). Thus, an increase in the thymic selection threshold during neonatal T cell development may endow the limited TCR repertoire with a greater ability to respond to multiple foreign Ags. A similar observation has been made at the other end of the life spectrum, where aged mice contain a naive CD4⁺ T cell population enriched for a higher-affinity TCR repertoire (64). Thymic output in the aged environment is increasingly limited, and thus, similar to neonates, there is a restricted naive T cell repertoire capable of responding to new foreign Ags. However, whether similar mechanisms are at play in generating the higher-affinity T cells during thymopoiesis remains to be determined.

Given the emerging correlation between CD5 levels and T cell function, it is tempting to speculate that the increased affinity of the neonatal TCR repertoire for self-antigen would play a role in shaping the unique responses of neonatal T cells to infection. For example, similar to CD5^hi CD8⁺ T cells in adult mice, neonatal CD8⁺ T cells express higher levels of both T-bet and Eomes and have robust, proliferative responses during primary infection (5, 6, 18). However, unlike their adult CD5^hi counterparts, whether neonatal CD8⁺ T cells efficiently generate protective memory remains controversial (5, 65, 66). In addition, we noted that CD5 levels were similar on OT-I TCR transgenic neonatal and adult T cells, whereas OT-I T cells from neonatal mice retain a microRNA signature that influences T-bet and Eomes expression, as well as downstream effector function (6). Thus, it is clear that skewing of thymic selection thresholds to generate T cells with higher affinity for self-pMHC correlates with several characteristics of the neonatal T cell repertoire but does not, alone, explain their unique responses to infection. Therefore, the extent to which the higher neonatal selection threshold impacts neonatal T cell function will need to be further investigated.

We thank Dr. Sylvie Lesage, Dr. Ivan Dzhagalov, Dr. Benjamin Haley, Dr. Salix Boulet, and Dr. Jean-Sébastien Delisle for critical reading of the manuscript, as well as members of the Lesage and Melichar laboratories for experimental advice and comments on the manuscript. We are grateful to Patricia D’Arcy at McGill University and the staff of the animal facilities at Maisonneuve-Rosemont Hospital Research Center and the University of Alberta Faculty of Medicine and Dentistry, Health Science Laboratory Animal Services for the maintenance of the mouse colonies. We thank Martine Dupuis at Maisonneuve-Rosemont Hospital Research Center for assistance with flow cytometry and cell sorting, the University of Alberta Faculty of Medicine and Dentistry Flow Cytometry Core, and the Cell Vision Core Facility for flow cytometry and single-cell analysis of the McGill Life Sciences Complex. Dr. Denis Claude Roy’s laboratory kindly provided cord blood samples and Dr. Ann Feeney (Scripps Institute) shared the TdT-deficient strain with us. C57BL/6-Tg(OT-I)-Rag1<tm1Mom> (OT-I Rag1^−/−) mice were obtained through the National Institute of Allergy and Infectious Diseases Exchange Program, National Institutes of Health.

The authors have no financial conflicts of interest.