Unraveling CD8 lineage decisions reveals that functionally distinct CD8+ T cells are selected by different MHC-I thymic peptides

Shinzawa, Miho; Ramos, Nicole; Bui, Khanh; Hajjar, William; Crossman, Assiatu; Chen, Xiongfong; Cam, Margaret; Takahama, Yousuke; Singer, Alfred

doi:10.1038/s41590-025-02411-4

Download PDF

Article
Open access
Published: 19 January 2026

Unraveling CD8 lineage decisions reveals that functionally distinct CD8⁺ T cells are selected by different MHC-I thymic peptides

Nature Immunology (2026)Cite this article

6718 Accesses
22 Altmetric
Metrics details

Subjects

Abstract

Thymocytes signaled by T cell antigen receptors to undergo positive selection acquire different functional fates while migrating through the thymus, but how this occurs remains uncertain. We now report that encoding CD8 co-receptors in both Cd4 and Cd8 gene loci modulates major histocompatibility complex (MHC-I) class I T cell antigen receptor signaling duration to generate all potential CD8⁺ T cell subsets. Strikingly, such mice revealed that functionally different CD8⁺ T cells are selected by different MHC-I thymic peptides. Thymocytes signaled by β5t-peptides produced by thymoproteasomes exclusively expressed in the thymic cortex invariably become cytotoxic CD8⁺ T cells indicating their signaling ceases when thymocytes leave the cortex; whereas thymocytes signaled by nonβ5t-peptides expressed throughout the thymus become either helper or innate memory CD8⁺ T cells because their signaling persists or recurs outside the cortex. Thus, it is because of their different thymic distributions that different MHC-I peptides select functionally different CD8⁺ T cells, integrating peptide specificity and CD8⁺ T cell function during positive selection and thymocyte migration.

Reversal of the T cell immune system reveals the molecular basis for T cell lineage fate determination in the thymus

Article Open access 29 April 2022

The differentiation and function of heterogeneous thymic dendritic cell subsets require signals provided by distinct thymocyte cell types

Article 08 January 2026

Pre-T cell receptor self-MHC sampling restricts thymocyte dedifferentiation

Article 21 November 2022

Main

During T cell development in the thymus, immature thymocytes are signaled by T cell antigen receptors (TCRs) to undergo positive selection and to differentiate into CD4⁺ or CD8⁺ mature T cells that possess either helper or cytotoxic function^1,2,3. Thymocytes that are induced to express the helper factor ThPOK differentiate into helper T cells and those that are induced to express the cytotoxic factor Runx3d differentiate into cytotoxic T cells^4,5,6,7,8. The lineage factor that TCR-signaled thymocytes express depends on the ligand specificity of their TCR in that TCR–CD8 complexes that engage MHC-I ligands signal thymocytes to become cytotoxic T cells, whereas TCR–CD4 complexes that engage MHC-II ligands signal thymocytes to become helper T cells^9,10,11. How TCR/co-receptor signaling induces thymocytes to express one or the other lineage factor during positive selection remains controversial and incompletely understood.

T cell lineage fate determination in the thymus is currently best described by the kinetic signaling model, which is based on the fact that TCR signaling of immature CD4⁺CD8⁺ double-positive thymocytes initially terminates Cd8 gene expression, regardless of TCR’s ligand specificity^12,13. When Cd8 gene expression is terminated, TCR signaling dependent on Cd8-encoded co-receptors becomes disrupted, whereas TCR signaling that is independent of Cd8-encoded co-receptors persists^12,13,14,15. Consequently, the key concept of kinetic signaling is that T cell lineage fate is determined during positive selection by whether TCR signaling is persistent or disrupted: TCR signaling persistence induces ThPOK, which directs thymocyte differentiation into helper T cells^16,17, whereas TCR signaling disruption allows thymic cytokines to induce Runx3d, which directs thymocyte differentiation into cytotoxic T cells^18,19. Interestingly, six different thymic cytokines (interleukin (IL)-7, IL-15, IL-6, interferon gamma (IFNγ), thymic stromal lymphopoietin (TSLP) and transforming growth factor-beta (TGFβ)) have been identified as inducing Runx3d expression and CD8⁺ T cell differentiation when TCR signaling is disrupted during positive selection¹⁹.

Understanding of T cell lineage fate determination was substantially advanced by a recent study using ‘FlipFlop’ mice in which CD4 and CD8 co-receptors were encoded in opposite Cd4 and Cd8 gene loci²⁰. Remarkably, FlipFlop mice generated a functionally reversed T cell immune system in which CD8⁺/MHC-I T cells are helpers and CD4⁺/MHC-II T cells are cytotoxic effectors²⁰. Thus, this study revealed that T cell lineage fate is not determined by CD4/CD8 co-receptors themselves, but is instead determined by Cd4 and Cd8 gene loci that encode CD4/CD8 co-receptors and regulate the kinetics of co-receptor-dependent TCR signaling during positive selection^13,20,21. However, the possibility that factors other than co-receptor gene loci might also affect T cell lineage fate determination has not been addressed and remains unknown.

The present study was undertaken to specifically identify mechanisms underlying CD8⁺ T cell lineage fate determination during MHC-I positive selection in the thymus. To do so, we constructed CD8^Dual mice, which encode CD8 co-receptors in both Cd4 and Cd8 gene loci to focus specifically on functional lineage fate decisions by MHC-I TCR-signaled thymocytes. We found that CD8^Dual mice are unique in generating both helper and cytotoxic CD8⁺ T cells, and this made it possible to discover that functionally distinct CD8⁺ T cell lineage fates are induced by different MHC-I thymic selecting peptides. Surprisingly, we found that MHC-I thymic selecting peptides expressed exclusively in the thymic cortex only stimulate generation of cytotoxic CD8⁺ T cells indicating that TCR signaling ceases when thymocytes disengage from the cortex, whereas MHC-I thymic selecting peptides that are expressed throughout the thymus stimulate generation of helper and innate memory (IM) CD8⁺ T cells indicating that their TCR signaling continues or recurs outside the cortex. We conclude that it is because MHC-I TCR signaling persistence/disruption dictates CD8⁺ T cell lineage fates that different MHC-I thymic peptides select functionally distinct (helper, cytotoxic and IM) CD8⁺ T cell subsets. Thus, this study integrates peptide specificity, T cell function and thymic migration to indicate how developing CD8⁺ T cells acquire different functionalities during MHC-I positive selection in the thymus.

Results

Mice with Cd4 and Cd8 gene loci that both encode CD8 co-receptors

To identify underlying mechanisms responsible for different CD8⁺ T cell lineage fates, we constructed CD8^Dual mice whose Cd4 and Cd8 gene loci both encode CD8 co-receptors. The Cd4 gene locus was modified to encode CD8α.1 and CD8β proteins instead of CD4 proteins, while the Cd8 gene locus was left intact²⁰. In homozygous CD8^Dual mice, Cd4 gene loci encoded CD8 co-receptors (Cd4^CD8) composed of CD8α.1CD8β dimers (referred to as CD8.1 co-receptors) and Cd8 gene loci encoded CD8 co-receptors (Cd8^CD8) composed of CD8α.2CD8β dimers (referred to as CD8.2 co-receptors; Fig. 1a). CD8.1 and CD8.2 co-receptors differ only in a single CD8α amino acid that does not affect CD8 co-receptor function²².

**Fig. 1: Characterization of CD8^Dual mice.**

Expression of CD8.1 and CD8.2 co-receptors defined four thymocyte subsets in CD8^Dual mice that are analogous to those defined by CD4⁺ and CD8⁺ co-receptors in wild-type (WT) C57BL/6 (B6) mice (Fig. 1b). CD8^Dual thymocytes consisted of CD8.1⁻8.2⁻ double-negative cells, CD8.1⁺8.2⁺ double-positive cells and CD8.1⁺8.2⁻ and CD8.1⁻8.2⁺ single-positive cells (referred to simply as CD8.1 and CD8.2 cells; Fig. 1b,c and Extended Data Fig. 1a). Notably, CD8.1 and CD8.2 cells are distinct subsets of mature TCR^hi thymocytes that emigrate into peripheral lymph nodes (LNs) as CD8.1 and CD8.2 T cells (Fig. 1c and Extended Data Fig. 1b,c). We found that CD8.1 and CD8.2 T cells were both selected by MHC-I-specific TCRs, as generation of all CD8^Dual T cells was abrogated by MHC-I deficiency (β2m^KO) but their generation was not significantly reduced by MHC-II deficiency (MHC-II^KO; Fig. 1d and Extended Data Fig. 1d). Thus, CD8.1 and CD8.2 T cells both express MHC-I-specific TCR and CD8 co-receptors.

Distinct lineage fates of CD8.1 and CD8.2 T cells

Even though CD8.1 and CD8.2 T cells both express MHC-I-specific TCR and CD8 co-receptors, they expressed opposite T cell lineage factors in that Cd4-encoded CD8.1 T cells express helper factor ThPOK, and Cd8-encoded CD8.2 T cells express cytotoxic factor Runx3d (Fig. 1e and Extended Data Fig. 1e). Moreover, genome-wide RNA sequencing (RNA-seq) revealed that the transcriptional profile of CD8.1 T cells resembled that of B6 CD4⁺ helper T cells, while the transcriptional profile of CD8.2 T cells resembled that of B6 CD8⁺ cytotoxic T cells (Fig. 1f). In addition, the molecular expression pattern of CD8.1 T cells resembled that of B6 CD4⁺ helper T cells^17,23,24,25, while the molecular expression pattern of CD8.2 T cells resembled that of B6 CD8⁺ cytotoxic T cells (Extended Data Fig. 2a–c)^{26,27,28,29,30}. Thus, the different molecular expression patterns of helper lineage T cells and cytotoxic lineage T cells do not result from their expression of different co-receptor proteins but result from their expression of co-receptors encoded in different Cd4/Cd8 co-receptor gene loci, as helper lineage CD8.1 T cells express co-receptors encoded in Cd4 gene loci and cytotoxic lineage CD8.2 T cells express co-receptors encoded in Cd8 gene loci.

We assessed the cellular function of CD8.1 and CD8.2 T cells by in vitro antibody stimulation and found that anti-CD3/CD28 stimulation induced only CD8.1 T cells to express the helper T cell molecule CD40L and that type 2 helper T (T_H2) cell stimulation cultures induced CD8.1 T cells to become IL-4⁺ T_H2 cells (Fig. 2a and Extended Data Fig. 3a). In contrast, anti-CD3/CD28 stimulation in IL-2 cultures induced only CD8.2 T cells to express the cytotoxic T cell molecules IFNγ and granzyme B (Fig. 2b). Moreover, injection of CD45.2 CD8^Dual T cells into sublethally irradiated lymphopenic CD45.1 host mice caused CD8.1 T cells to undergo limited lymphopenic proliferation like that of B6 CD4⁺ helper T cells, but caused CD8.2 T cells to undergo extensive lymphopenic proliferation like that of B6 CD8⁺ cytotoxic T cells, which correlated with their significantly higher GM1 staining of lipid raft components (Extended Data Fig. 3b–d)^31,32. Thus, CD8.1 T cells that express Cd4-encoded co-receptors possess helper function and CD8.2 T cells that express Cd8-encoded CD8 co-receptors possess cytotoxic function.

**Fig. 2: Effector functions of CD8.1 and CD8.2 CD8^Dual T cells.**

Finally, because CD4⁺ helper lineage T cells differentiate into Foxp3⁺ regulatory T (T_reg) cells^33,34 and PLZF⁺ invariant natural killer T (NKT) cells in B6 thymi^35,36,37,38, we asked if CD8.1 helper lineage T cells also differentiate into T_reg cells and NKT cells in CD8^Dual thymi. Indeed, CD8^Dual thymi contain Foxp3⁺CD8.1 T_reg cells (Fig. 2c and Extended Data Fig. 4a) and contain all three NKT subsets (NKT1, NKT2 and NKT17) of CD8.1 NKT cells that resemble those in WT thymi in that NKT1 cells are T-bet⁺IFNγ⁺; NKT2 cells are Gata3⁺IL-4⁺; and NKT17 cells are RORγt⁺IL-17⁺ (Fig. 2d,e and Extended Data Fig. 4b–f)^39,40. Notably, CD8^Dual thymi contained 2–3-fold higher frequencies of NKT2 cells than WT B6 thymi (Fig. 2e).

Thus, despite expressing MHC-I-specific TCR and CD8 co-receptors, CD8^Dual thymocytes differentiate into two distinct mature T cell subsets: CD8.1 T cells, which are ThPOK⁺ helper lineage cells, and CD8.2 T cells, which are Runx3d⁺ cytotoxic lineage cells. These two CD8⁺ T cell subsets display different gene profiles, different molecular expression patterns and different cellular functions. Thus, even when they encode the same CD8 co-receptor, Cd4 gene loci promote generation of helper CD8⁺ T cells, and Cd8 gene loci promote generation of cytotoxic CD8⁺ T cells.

Co-receptor kinetics and signaling duration during positive selection

In WT mice, Cd4 and Cd8 gene loci regulate MHC-I and MHC-II TCR signaling duration during positive selection by differentially regulating the kinetics of CD4 and CD8 co-receptor expression¹³. Consequently, we assessed if Cd4 and Cd8 gene loci regulate MHC-I TCR signaling duration during positive selection in CD8^Dual mice by differentially regulating the kinetics of expression of CD8.1 and CD8.2 co-receptors, respectively. To do so, we examined MHC-I signaled thymocytes expressing CD8.1 or CD8.2 co-receptors at five sequential stages of positive selection as defined by CD69 and CCR7 surface expression^33,41: CD69⁻CCR7⁻ stage 1 cells are TCR-unsignaled thymocytes; CD69⁺CCR7⁻ stage 2 cells are thymocytes that have just been TCR-signaled; and CD69⁺CCR7⁺ and CD69⁻CCR7⁺ cells are thymocytes at subsequent stages 3–5 of positive selection (Fig. 3a). We found that CD8.2 surface expression acutely declined on stage 3 thymocytes, whereas CD8.1 surface expression steadily increased (Fig. 3b), which are concordant with the kinetic signaling concept that Cd8 gene expression is specifically but transiently terminated in TCR-signaled thymocytes¹³. To determine how these different co-receptor kinetics affected TCR signaling, we normalized CD5 expression to TCR-signaled stage 2 thymocytes (Fig. 3c)⁴². Notably, CD5 expression between stages 2 and 4 increased on CD8.1 thymocytes but decreased on CD8.2 thymocytes (Fig. 3c), verifying that MHC-I TCR signaling that is dependent on Cd4-encoded CD8.1 co-receptors persisted without disruption, whereas TCR signaling that is dependent on Cd8-encoded CD8.2 co-receptors underwent disruption. Importantly, persistence of Cd4-dependent MHC-I TCR signaling induced ThPOK (Fig. 3c), while disruption of Cd8-dependent TCR signaling resulted in Runx3 expression (Fig. 3c). Note that CD5 expression always declined at stage 5 due to TCR signaling termination, which is necessary for thymocytes to express the chemotactic receptor S1P1 and exit the thymus (Fig. 3c)^43,44,45. Nevertheless, CD5 expression remained higher on CD8.1 than CD8.2 mature thymocytes even after they exited the thymus and became CD8.1 and CD8.2 LN T cells (Fig. 3d). Interestingly, we found that Nur77 mRNA was also more highly expressed in CD8.1 than CD8.2 mature thymocytes, and this difference also persisted in LN T cells (Fig. 3e)⁴⁶. Thus, CD5 and Nur77 expression on mature thymocytes and peripheral T cells reflect TCR signaling duration during positive selection.

**Fig. 3: Positive selection of MHC-I signaled CD8.1 and CD8.2 thymocytes.**

Our findings reveal that Cd4-encoded CD8 co-receptors promote TCR signaling persistence, which induces ThPOK expression and helper fate; whereas Cd8-encoded CD8 co-receptors promote TCR signaling disruption, which results in Runx3d expression and cytotoxic fate (schematized in Extended Data Fig. 5a).

Monoclonal TCR can induce different CD8⁺ T cell lineage fates

Having documented that CD8^Dual thymocytes are signaled during positive selection to adopt helper or cytotoxic lineage fates depending on whether MHC-I TCR signaling persists or undergoes disruption, we wanted to determine how monoclonal MHC-I-specific TCR would affect CD8⁺ T cell lineage fate decisions in CD8^Dual mice. To answer this question, we bred three different MHC-I-specific TCR transgenes (OT-I, P14 and HY) into CD8^Dual Rag^KO mice (Fig. 4a)^47,48,49. Expression of low-affinity HY and P14 TCR generated only CD8.2 cytotoxic T cells, which indicated that their signaling had been disrupted during positive selection (Fig. 4a,b). In contrast, expression of high-affinity OT-I TCR generated both CD8.1 T cells that expressed ThPOK and CD8.2 T cells that expressed Runx3 (Fig. 4a–c). That OT-I TCR generated CD8.1 helper T cells, but HY and P14 TCR did not, indicated that signaling persistence required higher-affinity TCR than signaling disruption, which was not surprising. However, the fact that monoclonal OT-I TCR generated both helper and cytotoxic CD8⁺ T cells indicated that OT-I thymocytes experienced different duration TCR signaling despite identical OT-I TCR and CD8 co-receptors, suggesting that another as-yet-unknown factor influenced MHC-I TCR signaling duration in the thymus.

**Fig. 4: Impact of MHC-I monoclonal TCRs on T cell lineage fates in CD8^Dual mice.**

Contribution of thymic selecting peptides to different CD8⁺ T cell lineage fates

Because MHC-I TCRs engage thymic MHC-I–peptide complexes to signal positive selection^50,51, we wondered if TCR engagement of different thymic MHC-I peptides might generate different duration MHC-I TCR signaling. Because TCR engagement of MHC-I–peptide complexes on cortical thymic epithelial cells (cTECs) signals thymocytes to undergo positive selection and to migrate to the corticomedullary junction and medulla^{52,53,54,55,56}, we thought that TCR signaling duration might depend on whether TCR–ligand engagements that are initiated in the cortex persist or become disrupted after thymocytes disengage from the cortex and migrate to the corticomedullary junction and medulla.

Notably, thymic MHC-I peptides are produced in cTECs by proteasomes with different β5 proteasomal subunits, in that ‘thymoproteasomes’ contain β5t subunits and produce β5t-peptides, whereas other proteasomes produce nonβ5t-peptides^57,58. Because thymoproteasomes are expressed exclusively in cTECs, the β5t-peptides they produce are only present on cTECs; and because other proteasomes are expressed in both cTECs and other thymic cellular elements, the nonβ5t-peptides they produce are expressed on both cTECs and other thymic cellular elements. Consequently, TCR engagement of β5t-peptide–MHC-I complexes must invariably be disrupted when thymocytes disengage from the thymic cortex and migrate, whereas TCR engagement of nonβ5t-peptide–MHC-I complexes might persist even after thymocytes disengage and migrate out of the thymic cortex (schematized in Extended Data Fig. 5b). Because disrupted TCR signaling during positive selection induces cytotoxic fate and undisrupted TCR signaling induces helper fate, we thought it possible that TCR engagement of β5t-peptides and nonβ5t-peptides on cTECs might induce different CD8⁺ T cell lineage fates (schematized in Extended Data Fig. 5b).

β5t-peptides exclusively select CD8⁺ cytotoxic T cells

To experimentally determine if β5t-peptides and nonβ5t-peptides induce different CD8⁺ T cell lineage fates, we compared CD8.1 helper and CD8.2 cytotoxic T cell generation in intact (β5t^WT) and β5t-deficient (β5t^KO) CD8^Dual mice (Fig. 5a). Quite unexpectedly, CD8.1 helper T cell generation was identical in β5t^WT and β5t^KO mice, which indicated that CD8.1 helper T cell generation was unaffected by the absence of β5t-peptides (Fig. 5a and Extended Data Fig. 6a). Absent β5t-peptides in β5t^KO mice also did not alter helper lineage features of CD8.1 T cells such as ThPOK expression, CD40L induction and Foxp3 expression (Extended Data Fig. 6b–d). In contrast, absent β5t-peptides reduced generation of CD8.2 cytotoxic T cells by ~50% in β5t^KO mice compared to β5t^WT mice, which indicated that CD8.2 cytotoxic T cells were selected by both β5t-peptides and nonβ5t-peptides (Fig. 5a). To more clearly appreciate the impact of β5t-peptides and nonβ5t-peptides on CD8⁺ T cells, we defined ‘MHC-I peptide selection index’ as the normalized percentage of T cells selected by β5t-peptides and nonβ5t-peptides, which illustrated that CD8.1 helper T cells were selected by nonβ5t-peptides, whereas CD8.2 cytotoxic T cells were selected by both β5t-peptides and nonβ5t-peptides (Fig. 5b).

**Fig. 5: Impact of different thymic MHC-I peptides on CD8⁺ T cell lineage fate.**

We then wished to determine the thymic peptides that selected monoclonal OT-I TCR thymocytes to become CD8.1 helper or CD8.2 cytotoxic T cells (Fig. 4). We found that both OT-I CD8.1 helper and OT-I CD8.2 cytotoxic mature thymocytes were selected by nonβ5t-peptides as neither was affected by the absence of β5t-peptides in β5t^KO mice (Extended Data Fig. 6e). It might also be noted that more OT-I CD8.1 helpers than OT-I CD8.2 killers appeared in the thymus, whereas the reverse was the case among peripheral LN T cells (Fig. 4 and Extended Data Fig. 6e), possibly because of the far greater proliferative potential in the lymphopenic periphery of CD8.2 cytotoxic than CD8.1 helper T cells (Extended Data Fig. 3c).

We conclude that different thymic MHC-I peptides induce different CD8⁺ T cell lineage fates, with β5t-peptides generating only CD8 cytotoxic T cells and nonβ5t-peptides generating either helper or cytotoxic CD8⁺ T cells.

IM CD8⁺ T cells are only selected by nonβ5t-peptides

To focus specifically on thymic selection of CD8 cytotoxic T cells, we compared CD8.2 cytotoxic thymocytes from β5t^WT and β5t^KO CD8^Dual mice (Fig. 5c–e). Most (~60%) CD8.2 thymocytes in β5t^WTCD8^Dual mice were conventional (CV) CD8⁺ T cells in that they were Runx3⁺Eomes⁻, whereas nearly all (>90%) CD8.2 thymocytes in β5t^KO CD8^Dual mice were IM CD8⁺ T cells that were Runx3⁺Eomes⁺ (Fig. 5c–e)^39,59,60. Generation of nonβ5t-selected IM CD8⁺ T cells in CD8^Dual mice resembled generation of IM CD8⁺ T cells in WT mice in being IL-4-dependent and abrogated in IL-4R^KO but not IL-15^KO mice, and in requiring NKT cells, which are absent in PLZF^KO and CD1d^KO CD8^Dual mice (Fig. 5c and Extended Data Fig. 7a–c)^39,59,60. Moreover, nonβ5t-selected IM CD8⁺ T cells in CD8^Dual mice are CD44^hi, CXCR3⁺, CD122⁺, Ly6C⁺ and CD49d⁻ and produce IFNγ after PMA + ionomycin stimulation (Extended Data Fig. 7d,e)^60,61. Importantly, because of their 2–3-fold greater frequency of NKT2 cells, CD8^Dual thymi express significantly more IL-4 mRNA than WT B6 thymi (Fig. 2e and Extended Data Fig. 7f). Notably, β5t deficiency affected neither thymic NKT2 cell frequencies nor IL-4 mRNA amounts (Extended Data Fig. 7g,h). We conclude that β5t-peptides select CV CD8⁺ T cells, whereas IM CD8⁺ T cells are selected by nonβ5t-peptides.

Because nonβ5t-selected CD8⁺ cytotoxic thymocytes could reencounter their selecting peptides outside the thymic cortex, differentiation into IM CD8⁺ T cells might be induced by late TCR signaling stimulated by nonβ5t-peptides on non-cortical cells. If this were the case, β5t-selected CV CD8⁺ T cells would be predicted to arise earlier than nonβ5t-selected IM CD8⁺ T cells. To assess this possibility, we utilized Rag-GFP expressing CD8^Dual thymocytes to determine the order of appearance in the thymus of CV and IM CD8.2 T cells (Fig. 5f,g)^41,62. We found that IM CD8.2 T cells appeared in the thymus ~10 h later and had significantly higher CD5 expression than CV CD8.2 T cells, suggesting that IM CD8⁺ T cells receive late TCR signaling stimulated by nonβ5t-peptides outside the cortex (Fig. 5f–h). We then wondered if expression of the nonβ5t-peptides that stimulated IM CD8⁺ T cell generation was dependent on Aire, which is specifically expressed in mTECs^63,64. However, contradicting this possibility, we found that IM CD8⁺ T cell generation was unaffected by Aire deficiency (Extended Data Fig. 7i). Based on these results, we suggest that late TCR signaling is induced by nonβ5t-peptides encountered outside the thymic cortex and delays CD8⁺ thymocytes from exiting the thymus, which increases exposure to intrathymic IL-4 and promotes differentiation into IM CD8⁺ T cells (schematized in Extended Data Fig. 7j).

Peptide selection of CD8⁺ T cells in WT mice

Finally, we wanted to determine if β5t-peptides and nonβ5t-peptides also select CV and IM CD8⁺ T cells in WT BALB/c and B6 mice, as they did in CD8^Dual mice. Because WT BALB/c mice resemble CD8^Dual mice in containing IL-4-producing NKT2 cells^39,59, we first examined the impact of thymic peptides on CD8⁺ T cell differentiation in BALB/c thymi. To do so, we bred the β5t^KO allele into BALB/c mice and compared CD8⁺ T cells in intact (β5t^WT) and β5t-deficient (β5t^KO) BALB/c littermates. Indeed, we discovered that absent β5t-peptides in β5t^KO BALB/c mice nearly abrogated generation of CV CD8⁺ T cells but did not significantly affect generation of IM CD8⁺ T cells in the thymus (Fig. 6a–c and Extended Data Fig. 8a). Thus, concordant with our findings in CD8^Dual thymi, β5t-peptides select CV CD8⁺ T cells and nonβ5t-peptides mostly select IM CD8⁺ T cells in WT BALB/c mice. Because of the greater number of IM CD8⁺ T cells in BALB/c thymi, which were selected by nonβ5t-peptides, β5t deficiency had little effect on total CD8⁺ T cell number in BALB/c compared to B6 thymi (Extended Data Fig. 8a,b).

**Fig. 6: Impact of different thymic MHC-I peptides on cytotoxic CD8⁺ T cells in WT mice.**

We then assessed CD8⁺ T cell selection by β5t- and nonβ5t-peptides in B6 thymi, which contain few IL-4-producing NKT2 cells and very few IM CD8⁺ T cells. We found that absent β5t-peptides in β5t^KO B6 mice nearly abrogated generation of CV CD8⁺ T cells, revealing that β5t-peptides select CV CD8⁺ T cells in B6 thymi, but that B6 thymi generated too few IM CD8⁺ T cells to allow us to confidently identify their peptide dependence (Fig. 6d–f and Extended Data Fig. 8b). Therefore, we analyzed peripheral CD8⁺ T cells in B6 LNs and found that β5t deficiency significantly reduced the peripheral number of total CD8⁺ and CV CD8⁺ T cells in B6 LNs, but did not reduce the number of IM CD8⁺ T cells in B6 LNs, which are dependent on IL-15 (refs. ^60,65,66; Extended Data Fig. 8c–e). Consequently, from our assessment of CD8^Dual, WT BALB/c and WT B6 mice, we conclude that CD8⁺ thymocytes consist mostly of β5t-selected CV T cells and nonβ5t-selected IM T cells, with the exception of WT B6 thymi, which specifically lack IM CD8⁺ T cells because of insufficient thymic IL-4 (Extended Data Fig. 9).

Our perspective on how different thymic peptides impact different CD8⁺ T cell lineage fates in the thymus is schematized in Fig. 6g.

Discussion

The present study documents that CD8⁺ T cell functionality is determined by MHC-I TCR signaling duration during positive selection and that different MHC-I thymic peptides select functionally different CD8⁺ T cells based on whether the signaling they induce is continuous or disrupted. Analysis of CD8^Dual mice reveals that Cd4-encoded CD8 co-receptors promote persistent MHC-I TCR signaling, which generates helper CD8⁺ T cells, and that Cd8-encoded CD8 co-receptors promote disrupted MHC-I TCR signaling, which generates cytotoxic CD8⁺ T cells. Importantly, CD8^Dual mice also reveal that different MHC-I thymic peptides generate functionally different CD8⁺ T cells, in that β5t-peptides (produced by the thymoproteasome) promote only disrupted TCR signaling, which generates cytotoxic CD8⁺ T cells, and that nonβ5t-peptides (produced by other proteasomes) promote persistent or recurrent TCR signaling, which generates helper and IM CD8⁺ T cells. Thus, different MHC-I thymic peptides stimulate generation of distinct subsets of helper, cytotoxic and IM CD8⁺ T cells. Importantly, generation of functionally distinct CD8⁺ T cell subsets by different MHC-I thymic peptides is not unique to CD8^Dual mice but is also true in normal WT mice as well.

The present study provides a new perspective on the role of MHC-I thymic selecting peptides in CD8⁺ T cell-positive selection that challenges the current understanding of their role in CD8⁺ T cell generation. It is currently thought that CD8⁺ T cell selection in the thymus is mainly mediated by a specialized subset of MHC-I thymic peptides (referred to as β5t-peptides) that are produced by thymoproteasomes and expressed exclusively in cTECs^57,58. Other MHC-I thymic peptides (nonβ5t-peptides) are produced by other proteasomes and are expressed throughout the thymus, including in cTECs, but these peptides are thought to be poor selectors of CD8⁺ T cells. Indeed, previous reports performed in B6-background mice have revealed that most CD8⁺ T cells are selected by thymic β5t-peptides and have revealed that β5t-selected CD8⁺ T cells are more TCR responsive and functional than nonβ5t-selected CD8⁺ T cells^67,68,69. Consequently, thymoproteasome-generated β5t-peptides are thought to express particular amino acid sequences that make them uniquely able to stimulate MHC-I TCR to signal thymocyte selection and differentiation into functional CD8⁺ T cells^70,71.

In contrast, our present study now documents that β5t-peptides and nonβ5t-peptides both stimulate CD8⁺ T cell-positive selection but that they generate functionally different CD8⁺ T cells, with β5t-selected CD8⁺ T cells acquiring only cytotoxic function and nonβ5t-selected CD8⁺ T cells becoming either helper and/or IM cells. Because β5t-peptides are expressed only in cTECs, TCR engagement of β5t-peptides would invariably become disrupted when thymocytes disengage from the cortex and invariably become cytotoxic CD8⁺ T cells; and, because nonβ5t-peptides are expressed throughout the thymus, TCR engagement of nonβ5t-peptides would either continue or recur when thymocytes disengage from the cortex and migrate through the thymus while becoming helper or IM CD8⁺ T cells. Thus, it is because MHC-I thymic peptides have different expression patterns in the thymus that they stimulate the generation of functionally different (helper, cytotoxic, IM) CD8⁺ T cells. Nevertheless, because MHC-I TCR clearly distinguish β5t-peptides from nonβ5t-peptides, the present study does not exclude the possibility that β5t-peptides may be intrinsically more stimulatory of MHC-I TCR than nonβ5t-peptides.

It is interesting to consider that the present study suggests that transient termination of Cd8 gene expression as emphasized in the kinetic signaling model may be unnecessary for β5t-signaled thymocytes to become CD8 cytotoxic T cells because β5t-induced TCR signaling would be disrupted anyway when signaled thymocytes disengage from cTECs and migrate out of the thymic cortex. In contrast, transient termination of Cd8 gene expression is necessary for nonβ5t-signaled thymocytes to adopt cytotoxic fates. Consequently, while transient termination of Cd8 gene expression is a universal mechanism for disrupting CD8-dependent MHC-I TCR signaling during positive selection, TCR signaling disruption of positive selection by any mechanism induces cytotoxic fate.

The effect of MHC-I thymic peptides on CD8⁺ T cell lineage fates in the present study was not limited to CD8^Dual mice as it was also documented in normal mice, although BALB/c and B6 normal mice differed from one another in terms of thymic IL-4 amounts. BALB/c thymi contain IL-4-producing NKT2 cells and intrathymic IL-4 in amounts that support robust numbers of nonβ5t-selected IM CD8⁺ T cells. In contrast B6 thymi contain few IL-4-producing NKT2 cells and little intrathymic IL-4, as well as containing very few nonβ5t-selected IM CD8⁺ T cells. Thus, we think it is their deficient number of IM CD8⁺ T cells that explains why B6 thymocytes contain mainly β5t-selected CD8⁺ T cells and few nonβ5t-selected CD8⁺ T cells. As possible underlying mechanisms, we think that thymocytes receiving late TCR signaling by nonβ5t-peptides may require thymic IL-4 for survival as well as to become IM CD8⁺ T cells; or, alternatively, it is possible that very few CD8⁺ thymocytes are signaled by nonβ5t-peptides but these then extensively proliferate in response to abundant thymic IL-4. While preliminary experiments suggest that in vivo expression of an antiapoptotic transgene does not reveal additional nonβ5t-signaled thymocytes, we cannot yet definitively distinguish between these two possibilities.

In any event, by revealing that functionally different CD8⁺ T cell subsets are generated by different MHC-I thymic peptides, our study provides new insights into the different developmental requirements for different CD8⁺ T cell subsets. Helper CD8⁺ T cells are only generated by TCR engagement of nonβ5t-peptides, which can persist and continue undisrupted despite thymocyte migration out of the thymus. However, if TCR engagement of nonβ5t-peptides becomes disrupted, they can reengage nonβ5t-peptides on non-cortical thymic elements to stimulate late TCR signaling, which induce thymocytes to become IM CD8⁺ T cells. In contrast to helper and IM CD8⁺ T cells, conventional CD8 cytotoxic T cells are generated by TCR engagement of β5t-peptides, which would become permanently disrupted when TCR signaling terminates thymocyte expression of the chemokine receptor CXCR4, which causes signaled thymocytes to disengage from cTECs and to migrate out of the cortex without receiving additional late TCR signaling⁵⁶. A recent bioinformatics analysis suggested that late TCR signaling may be required for generation of all mature CD8⁺ T cells^3,72,73. However, our current study documents that late TCR signaling is only required for generation of IM CD8⁺ T cells but is not required for generation of any other CD8⁺ T cells. How late TCR signaling, together with thymic IL-4, generates IM CD8⁺ T cells requires further investigation, but we think that late TCR signaling may prolong the encounter of developing CD8⁺ T cells with thymic IL-4 either by delaying their exit from the thymus or by increasing thymocyte responsiveness to IL-4 in some other way. Notably, our present findings support the concept that IM CD8⁺ T cells are generated by a TCR-instructed process in the thymus that causes IM CD8⁺ T cells to express a different TCR repertoire than conventional CD8⁺ T cells, as has been reported⁷⁴. While it was the abundance of IL-4-producing NKT2 cells in CD8^Dual thymi that revealed that IM CD8⁺ T cells required late TCR signaling by nonβ5t-peptides, the reason CD8^Dual thymi contain an abundance of IL-4-producing NKT2 cells is not yet understood. Because CD8^Dual thymi contain Cd4-encoded CD8 co-receptors whose surface expression is maintained throughout positive selection, its basis may be related to the finding that constitutive expression of transgenic CD8 also results in an abundance of IL-4-producing NKT2 cells^75,76.

Finally, the concept that thymocyte migration affects CD8⁺ T cell lineage fate determination as proposed in the present study merits further comment. The fact that β5t-peptides expressed exclusively by cTECs signal CD8⁺ thymocytes to only become cytotoxic T cells, whereas nonβ5t-peptides expressed throughout the thymus signal thymocytes to instead become helper or IM CD8⁺ T cells provides strong evidence that thymocyte migration contributes to these CD8⁺ T cell lineage fates. Notably, future experiments will directly assess if disruption of thymocyte migration indeed alters CD8⁺ T cell lineage fates.

In conclusion, CD8⁺ T cell lineage fates are determined by MHC-I TCR signaling persistence or disruption during positive selection, with disruption of TCR signaling resulting in generation of cytotoxic CD8⁺ T cells. Concordant with this perspective, the expression pattern of different MHC-I thymic selecting peptides affects CD8⁺ T cell lineage fate decisions by causing persistent or disrupted TCR signaling. Thus, the present study integrates thymocyte peptide specificity with CD8⁺ T cell fate determination during MHC-I signaled positive selection.

Methods

Mice

B6 (CD45.1 and CD45.2) mice were purchased from Charles River Laboratory. Aire^KO (ref. ⁷⁷), BALB/cJ, β2m^KO (ref. ⁷⁸), CD1d^KO (ref. ⁷⁹), CD8α^KO (ref. ⁸⁰), IL-4R^KO (ref. ⁸¹) and IL-15^KO (ref. ⁸²) mice were purchased from The Jackson Laboratory. β5t^KO from Y.T.⁵⁷, PLZF^KO from D. Kovalovsky³⁵, MHC-II^KO, HY.Rag2^KO, P14.Rag2^KO, OT-I.Rag2^KO, Rag-GFP transgene (Tg) from M. Nussenzweig⁸³, Runx3d-YFP knock-in from D. R. Littman⁸⁴ and ThPOK-GFP Tg mice from R. Bosselut²⁵ were maintained in our own animal colony. To generate BALB/c.β5t^KO mice, B6.β5t^KO mice were back-crossed five times with BALB/cJ. Mice were housed on a 12-h light–dark cycle at 20–26 °C with 30–70% humidity in accordance with US National Institutes of Health guidelines. All mice were analyzed without randomization or blinding at 6–12 weeks of age and both sexes were used unless mentioned in the paper. All mouse experiments were approved by the National Cancer Institute Animal Care and Use Committee.

Generation of CD8^Dual mice

CD8^Dual mice contained altered Cd4 gene loci encoding CD8α.1 and CD8β proteins (Cd4^CD8) as previously reported²⁰, and contained intact Cd8 gene loci encoding CD8α.2 and CD8β proteins (Fig. 1a).

Homeostatic proliferation assays

Donor T cells (CD45.2) purified from LNs using the Pan T Cell Isolation Kit II (Miltenyi Biotec) were labeled with 0.5 μM Cell Trace Violet (Invitrogen) and were injected intravenously into host B6 (CD45.1) mice that were sublethally irradiated (600 R) the previous day. Mice were analyzed 7 days after injection.

Flow cytometry

Single-cell suspensions (1–5 × 10⁶) were first incubated with anti-FcR (clone 2.4G2; 1:250 dilution) and stained with fluorescence-conjugated antibodies at 4 °C for 40 min in HBSS (Thermo Fisher Scientific) with 0.5% BSA and 0.5% NaN₃. After washing, stained cells were incubated with fluorescence-conjugated streptavidin against biotin-conjugated antibodies at 4 °C for 15 min. Staining of CCR7 (4B12; 1:50 dilution) was performed at 37 °C for 30 min and staining of CD1d tetramer (CD1d/PBS-57; 1:100 dilution) was performed at 4 °C for 30 min before staining with other antibodies. GM1 amount was analyzed by staining with recombinant cholera toxin subunit B (Thermo Fisher Scientific; 1:200 dilution)³². For intracellular staining of transcription factors and cytokines, cells were fixed and permeabilized with the Foxp3 Staining Buffer Set (Thermo Fisher Scientific) or BD Cytofix/Cytoperm Fixation/Permeabilization Kit (BD Biosciences), and then stained with fluorescence-conjugated antibodies at 4 °C for 30 min. Gata3 (TWAJ; 5 μl), ThPOK (T43-94, 2 μl), Runx3 (R3-5G4, 5 μl), PLZF (R17-809; 1:50 dilution), RORγt (Q31-378; 1:50 dilution) and T-bet (4B10; 1:100 dilution) were stained at 4 °C for 1 h. For cytokine staining, cells were stimulated at 37 °C with PMA (50 ng ml⁻¹; Calbiochem) and ionomycin (1 μg ml⁻¹; Calbiochem) for 2 h and added GolgiStop (BD Biosciences) for 2 h before staining. Stained cells were analyzed on a FACS LSR II or FACS Fortessa flow cytometer (BD Biosciences). Electronic cell sorting was performed on a FACSAria II or a FACSAria FUSION (BD Biosciences). Dead cells were excluded by forward light-scatter gating and staining with propidium iodide or LIVE/DEAD Fixable Aqua Dead Cell Stain Kit (Thermo Fisher Scientific) for fresh and fixed staining, respectively. Data were analyzed using FlowJo software (TreeStar). Anti-CD8α.1 (BioXcell; HB129/116-13.1) and CD8α.2 (BioXcell; 2.43) were labeled with biotin by EZ-Link Sulfo-NHS-LC-Biotinylation Kit (Thermo Scientific) or with Alexa 647 (BioXcell). Detailed information on antibodies is provided in the Reporting Summary.

In vitro stimulation of LN T cells

T cells were purified from LNs with the Pan T cell Isolation Kit (Miltenyi Biotec). For CD40L induction, LN T cells were stimulated with or without plate-bound anti-CD3 (1 μg ml⁻¹) and CD28 (1 μg ml⁻¹) antibodies at 37 °C for 24 h. For expression of cytotoxic lineage-related factors, sorted naive LN T cells (TCRβ⁺CD62L⁺CD44⁻) were stimulated with plate-bound anti-CD3 (10 μg ml⁻¹) and CD28 (5 μg ml⁻¹) antibodies for 3 days followed by stimulation with human IL-2 (100 U ml⁻¹) at 37 °C for 2 days.

T_H2 cell differentiation assay in vitro

Naive (TCRβ⁺CD62L⁺CD44⁻) T cells were electronically sorted from LNs. Sorted T cells (1 × 10⁶) were stimulated with soluble anti-CD3 (2 μg ml⁻¹) in the presence of irradiated (2000R) splenocytes (5 × 10⁶) in each well of a flat-bottom 24-well plate (Corning) at 37 °C for 2 days with hIL-2 (200 U ml⁻¹), mIL-4 (10 ng ml⁻¹) and anti-IFNγ antibody (20 μg ml⁻¹). The cultured T cells were further incubated with hIL-4 (hIL-2 (200 U ml⁻¹), mIL-4 (10 ng ml⁻¹) and anti-IFNγ antibody (20 μg ml⁻¹) in each well of a flat-bottom six-well plate (Corning) 37 °C for 3 days.

RT–qPCR

Total RNA was isolated with the RNeasy Plus Mini Kit (Qiagen) and cDNA was prepared with superscript III First-Strand Synthesis System for RT–PCR kit (Invitrogen). RT–qPCR was done with TaqMan PCR system (Thermo Fisher Scientific) or QuantiTect SYBR Green PCR system (Qiagen). TaqMan probes for Cd40lg (CD40L; Mm00441911_m1), Cd8a (Mm01182107_g1), Cish (Mm01230623_g1), Eomes (Mm01351985_m1), Il-4 (Mm00445259_m1), Nr4a1 (Nur77; Mm01300401_m1), Prf1 (Perforin; Mm00812512_m1), Rpl13a (Mm01612986_gH), Socs1 (Mm00782550_s1), Socs3 (Mm00545913_s1), Tbx21 (T-bet; Mm00450960_m1) and Zbtb7b (ThPOK; Mm00784709_s1), were from Thermo Fisher Scientific. The primer sequences for SYBR green PCR system were as follows: Runx3d forward (5’-GCGACATGGCTTCCAACAGC-3’) and reverse (5’-CTTAGCGCGCTGTTCTCGC-3’); Rpl13a forward (5’-CGAGGCATGCTGCCCCACAA-3’) and reverse (5’-AGCAGGGACCACCATCCGCT-3’). Samples were analyzed on a QuantStudio 6 Flex Real-time PCR System (Applied Biosystems). Gene expression values were normalized to those of Rpl13a expression in the same sample.

RNA-seq analysis

Naive (TCRβ⁺CD62L⁺CD44⁻) T cells were electronically sorted from LNs. Total RNA was prepared from sorted cells with the RNeasy Plus Mini Kit (Qiagen). The quality of RNA was assessed by Bioanalyzer (Agilent), and RNA samples with an RNA integrity number >9 were used. The library was made by using the SMARTer Universal Low Input RNA Kit (Clontech) for sequencing. The sequencing was performed with a paired-end sequencing length of 125 base pairs by using HiSeq 2500 equipment (Illumina). Reads were aligned to the mouse genome (mm10) with STAR aligner⁸⁵, and raw count data were produced using RSEM⁸⁶. Differentially expressed genes (DEGs) were genes whose fold change was more than 5-fold and P value was less than 0.05. DEGs between CD4⁺ and CD8⁺ T cells from B6 mice (631 genes) were analyzed in CD8.1 and CD8.2 T cells from CD8^Dual mice. Visualization of DEGs is shown in a heat map generated with Partek.

Differentiation time of double-positive thymocytes into CV and IM CD8⁺ T cells

To determine the time for CV and IM CD8⁺ T cell development from double-positive thymocytes, Rag-GFP Tg was introduced into CD8^Dual mice and Rag-GFP expression (MFI) on CD8^Dual thymocytes was analyzed. Rag-GFP expression (MFI) in CV and IM CD8.2 T cells relative to CD24⁺CD8α.1⁺CD8β⁺ double-positive thymocytes were used to calculate the differentiation time based on a Rag-GFP half-life of 54–56 h as: time (h) = (100 – relative Rag-GFP expression)/0.9 (ref. ⁶²).

MHC-I peptide selection index

We calculated the frequencies of β5t-selected and nonβ5t-selected cells based on cell numbers from β5t^WT and β5t^KO mice in the following way:

$${\rm{Non}}\beta 5{\rm{t}}_{\rm{selected}}\,{\rm{cells}}\;({\rm{ \% }})=({{\rm{No}}.\;\beta 5{\rm{t}}}^{{\rm{KO}}}\!/{{\rm{No}}.\;\beta 5{\rm{t}}}^{{\rm{WT}}})\times 100$$

$$\beta 5{\rm{t}}_{\rm{selected}}\,{\rm{cells}}\;({\rm{ \% }})=[1-({{\rm{No}}.\;\beta 5{\rm{t}}}^{{\rm{KO}}}\!/{{\rm{No}}.\;\beta 5{\rm{t}}}^{{\rm{WT}}})]\times 100$$

Values greater than 100 were set to 100.

Statistical analysis

Statistical analyses were performed using an unpaired Student’s t-test. P values < 0.05 were considered significant.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

RNA-seq of LN T cells from CD8^Dual and B6 mice are deposited in the Gene Expression Omnibus under accession no. GSE297710.

References

Littman, D. R. How thymocytes achieve their fate. J. Immunol. 196, 1983–1984 (2016).
Article PubMed CAS Google Scholar
Ashby, K. M. & Hogquist, K. A. A guide to thymic selection of T cells. Nat. Rev. Immunol. 24, 103–117 (2024).
Article PubMed CAS Google Scholar
Steier, Z., Kim, E. J. Y., Aylard, D. A. & Robey, E. A. The CD4 versus CD8 T cell fate decision: a multiomics-informed perspective. Annu. Rev. Immunol. 42, 235–258 (2024).
Article PubMed CAS Google Scholar
He, X. et al. The zinc finger transcription factor Th-POK regulates CD4 versus CD8 T-cell lineage commitment. Nature 433, 826–833 (2005).
Article PubMed CAS Google Scholar
Sun, G. et al. The zinc finger protein cKrox directs CD4 lineage differentiation during intrathymic T cell positive selection. Nat. Immunol. 6, 373–381 (2005).
Article PubMed CAS Google Scholar
Taniuchi, I. et al. Differential requirements for Runx proteins in CD4 repression and epigenetic silencing during T lymphocyte development. Cell 111, 621–633 (2002).
Article PubMed CAS Google Scholar
Egawa, T., Tillman, R. E., Naoe, Y., Taniuchi, I. & Littman, D. R. The role of the Runx transcription factors in thymocyte differentiation and in homeostasis of naive T cells. J. Exp. Med. 204, 1945–1957 (2007).
Article PubMed PubMed Central CAS Google Scholar
Setoguchi, R. et al. Repression of the transcription factor Th-POK by Runx complexes in cytotoxic T cell development. Science 319, 822–825 (2008).
Article PubMed CAS Google Scholar
Teh, H. S. et al. Thymic major histocompatibility complex antigens and the alpha beta T-cell receptor determine the CD4/CD8 phenotype of T cells. Nature 335, 229–233 (1988).
Article PubMed CAS Google Scholar
Robey, E. A. et al. Thymic selection in CD8 transgenic mice supports an instructive model for commitment to a CD4 or CD8 lineage. Cell 64, 99–107 (1991).
Article PubMed CAS Google Scholar
Itano, A. et al. The cytoplasmic domain of CD4 promotes the development of CD4 lineage T cells. J. Exp. Med. 183, 731–741 (1996).
Article PubMed CAS Google Scholar
Brugnera, E. et al. Coreceptor reversal in the thymus: signaled CD4⁺8⁺ thymocytes initially terminate CD8 transcription even when differentiating into CD8 T cells. Immunity 13, 59–71 (2000).
Article PubMed CAS Google Scholar
Singer, A., Adoro, S. & Park, J. H. Lineage fate and intense debate: myths, models and mechanisms of CD4- versus CD8-lineage choice. Nat. Rev. Immunol. 8, 788–801 (2008).
Article PubMed PubMed Central CAS Google Scholar
Bosselut, R., Guinter, T. I., Sharrow, S. O. & Singer, A. Unraveling a revealing paradox: why major histocompatibility complex I-signaled thymocytes “paradoxically” appear as CD4⁺8^lo transitional cells during positive selection of CD8⁺ T cells. J. Exp. Med. 197, 1709–1719 (2003).
Article PubMed PubMed Central CAS Google Scholar
Sarafova, S. D. et al. Modulation of coreceptor transcription during positive selection dictates lineage fate independently of TCR/coreceptor specificity. Immunity 23, 75–87 (2005).
Article PubMed CAS Google Scholar
Sarafova, S. D. et al. Upregulation of CD4 expression during MHC class II-specific positive selection is essential for error-free lineage choice. Immunity 31, 480–490 (2009).
Article PubMed PubMed Central CAS Google Scholar
Luckey, M. A. et al. The transcription factor ThPOK suppresses Runx3 and imposes CD4⁺ lineage fate by inducing the SOCS suppressors of cytokine signaling. Nat. Immunol. 15, 638–645 (2014).
Article PubMed PubMed Central CAS Google Scholar
Park, J. H. et al. Signaling by intrathymic cytokines, not T cell antigen receptors, specifies CD8 lineage choice and promotes the differentiation of cytotoxic-lineage T cells. Nat. Immunol. 11, 257–264 (2010).
Article PubMed PubMed Central CAS Google Scholar
Etzensperger, R. et al. Identification of lineage-specifying cytokines that signal all CD8⁺-cytotoxic-lineage-fate ‘decisions’ in the thymus. Nat. Immunol. 18, 1218–1227 (2017).
Article PubMed PubMed Central CAS Google Scholar
Shinzawa, M. et al. Reversal of the T cell immune system reveals the molecular basis for T cell lineage fate determination in the thymus. Nat. Immunol. 23, 731–742 (2022).
Article PubMed PubMed Central CAS Google Scholar
Adoro, S. et al. Coreceptor gene imprinting governs thymocyte lineage fate. EMBO J. 31, 366–377 (2012).
Article PubMed CAS Google Scholar
Liaw, C. W., Zamoyska, R. & Parnes, J. R. Structure, sequence, and polymorphism of the Lyt-2 T cell differentiation antigen gene. J. Immunol. 137, 1037–1043 (1986).
Article PubMed CAS Google Scholar
Grewal, I. S. et al. Requirement for CD40 ligand in costimulation induction, T cell activation, and experimental allergic encephalomyelitis. Science 273, 1864–1867 (1996).
Article PubMed CAS Google Scholar
Salomon, B. et al. B7/CD28 costimulation is essential for the homeostasis of the CD4⁺CD25⁺ immunoregulatory T cells that control autoimmune diabetes. Immunity 12, 431–440 (2000).
Article PubMed CAS Google Scholar
Wang, L. et al. Distinct functions for the transcription factors GATA-3 and ThPOK during intrathymic differentiation of CD4⁺ T cells. Nat. Immunol. 9, 1122–1130 (2008).
Article PubMed PubMed Central CAS Google Scholar
El-Asady, R. et al. TGF-beta-dependent CD103 expression by CD8⁺ T cells promotes selective destruction of the host intestinal epithelium during graft-versus-host disease. J. Exp. Med. 201, 1647–1657 (2005).
Article PubMed PubMed Central CAS Google Scholar
Kagi, D. et al. Cytotoxicity mediated by T cells and natural killer cells is greatly impaired in perforin-deficient mice. Nature 369, 31–37 (1994).
Article PubMed CAS Google Scholar
Pearce, E. L. et al. Control of effector CD8⁺ T cell function by the transcription factor Eomesodermin. Science 302, 1041–1043 (2003).
Article PubMed CAS Google Scholar
Intlekofer, A. M. et al. Effector and memory CD8⁺ T cell fate coupled by T-bet and eomesodermin. Nat. Immunol. 6, 1236–1244 (2005).
Article PubMed CAS Google Scholar
Heusel, J. W., Wesselschmidt, R. L., Shresta, S., Russell, J. H. & Ley, T. J. Cytotoxic lymphocytes require granzyme B for the rapid induction of DNA fragmentation and apoptosis in allogeneic target cells. Cell 76, 977–987 (1994).
Article PubMed CAS Google Scholar
Bourgeois, C. & Stockinger, B. T cell homeostasis in steady state and lymphopenic conditions. Immunol. Lett. 107, 89–92 (2006).
Article PubMed CAS Google Scholar
Cho, J. H., Kim, H. O., Surh, C. D. & Sprent, J. T cell receptor-dependent regulation of lipid rafts controls naive CD8⁺ T cell homeostasis. Immunity 32, 214–226 (2010).
Article PubMed PubMed Central CAS Google Scholar
Tai, X. et al. How autoreactive thymocytes differentiate into regulatory versus effector CD4⁺ T cells after avoiding clonal deletion. Nat. Immunol. 24, 637–651 (2023).
Article PubMed PubMed Central CAS Google Scholar
Dikiy, S. & Rudensky, A. Y. Principles of regulatory T cell function. Immunity 56, 240–255 (2023).
Article PubMed CAS Google Scholar
Kovalovsky, D. et al. The BTB-zinc finger transcriptional regulator PLZF controls the development of invariant natural killer T cell effector functions. Nat. Immunol. 9, 1055–1064 (2008).
Article PubMed PubMed Central CAS Google Scholar
Savage, A. K. et al. The transcription factor PLZF directs the effector program of the NKT cell lineage. Immunity 29, 391–403 (2008).
Article PubMed PubMed Central CAS Google Scholar
Engel, I. et al. Co-receptor choice by V alpha14i NKT cells is driven by Th-POK expression rather than avoidance of CD8-mediated negative selection. J. Exp. Med. 207, 1015–1029 (2010).
Article PubMed PubMed Central CAS Google Scholar
Constantinides, M. G. & Bendelac, A. Transcriptional regulation of the NKT cell lineage. Curr. Opin. Immunol. 25, 161–167 (2013).
Article PubMed PubMed Central CAS Google Scholar
Lee, Y. J., Holzapfel, K. L., Zhu, J., Jameson, S. C. & Hogquist, K. A. Steady-state production of IL-4 modulates immunity in mouse strains and is determined by lineage diversity of iNKT cells. Nat. Immunol. 14, 1146–1154 (2013).
Article PubMed CAS Google Scholar
Pellicci, D. G., Koay, H. F. & Berzins, S. P. Thymic development of unconventional T cells: how NKT cells, MAIT cells and gammadelta T cells emerge. Nat. Rev. Immunol. 20, 756–770 (2020).
Article PubMed CAS Google Scholar
Kimura, M. Y. et al. Timing and duration of MHC I positive selection signals are adjusted in the thymus to prevent lineage errors. Nat. Immunol. 17, 1415–1423 (2016).
Article PubMed PubMed Central CAS Google Scholar
Azzam, H. S. et al. CD5 expression is developmentally regulated by T cell receptor (TCR) signals and TCR avidity. J. Exp. Med. 188, 2301–2311 (1998).
Article PubMed PubMed Central CAS Google Scholar
Matloubian, M. et al. Lymphocyte egress from thymus and peripheral lymphoid organs is dependent on S1P receptor 1. Nature 427, 355–360 (2004).
Article PubMed CAS Google Scholar
Shiow, L. R. et al. CD69 acts downstream of interferon-alpha/beta to inhibit S1P1 and lymphocyte egress from lymphoid organs. Nature 440, 540–544 (2006).
Article PubMed CAS Google Scholar
Carlson, C. M. et al. Kruppel-like factor 2 regulates thymocyte and T-cell migration. Nature 442, 299–302 (2006).
Article PubMed CAS Google Scholar
Moran, A. E. et al. T cell receptor signal strength in Treg and iNKT cell development demonstrated by a novel fluorescent reporter mouse. J. Exp. Med. 208, 1279–1289 (2011).
Article PubMed PubMed Central CAS Google Scholar
Hogquist, K. A. et al. T cell receptor antagonist peptides induce positive selection. Cell 76, 17–27 (1994).
Article PubMed CAS Google Scholar
Pircher, H., Burki, K., Lang, R., Hengartner, H. & Zinkernagel, R. M. Tolerance induction in double specific T-cell receptor transgenic mice varies with antigen. Nature 342, 559–561 (1989).
Article PubMed CAS Google Scholar
Kisielow, P., Bluthmann, H., Staerz, U. D., Steinmetz, M. & von Boehmer, H. Tolerance in T-cell-receptor transgenic mice involves deletion of nonmature CD4⁺8⁺ thymocytes. Nature 333, 742–746 (1988).
Article PubMed CAS Google Scholar
Hogquist, K. A., Gavin, M. A. & Bevan, M. J. Positive selection of CD8+ T cells induced by major histocompatibility complex binding peptides in fetal thymic organ culture. J. Exp. Med. 177, 1469–1473 (1993).
Article PubMed CAS Google Scholar
Santori, F. R. et al. Rare, structurally homologous self-peptides promote thymocyte positive selection. Immunity 17, 131–142 (2002).
Article PubMed CAS Google Scholar
Ueno, T. et al. CCR7 signals are essential for cortex-medulla migration of developing thymocytes. J. Exp. Med. 200, 493–505 (2004).
Article PubMed PubMed Central CAS Google Scholar
Takahama, Y. Journey through the thymus: stromal guides for T-cell development and selection. Nat. Rev. Immunol. 6, 127–135 (2006).
Article PubMed CAS Google Scholar
Cowan, J. E. et al. Differential requirement for CCR4 and CCR7 during the development of innate and adaptive alphabetaT cells in the adult thymus. J. Immunol. 193, 1204–1212 (2014).
Article PubMed PubMed Central CAS Google Scholar
Hu, Z., Lancaster, J. N., Sasiponganan, C. & Ehrlich, L. I. CCR4 promotes medullary entry and thymocyte-dendritic cell interactions required for central tolerance. J. Exp. Med. 212, 1947–1965 (2015).
Article PubMed PubMed Central CAS Google Scholar
Kadakia, T. et al. E-protein-regulated expression of CXCR4 adheres preselection thymocytes to the thymic cortex. J. Exp. Med. 216, 1749–1761 (2019).
Article PubMed PubMed Central CAS Google Scholar
Murata, S. et al. Regulation of CD8⁺ T cell development by thymus-specific proteasomes. Science 316, 1349–1353 (2007).
Article PubMed CAS Google Scholar
Murata, S., Takahama, Y., Kasahara, M. & Tanaka, K. The immunoproteasome and thymoproteasome: functions, evolution and human disease. Nat. Immunol. 19, 923–931 (2018).
Article PubMed CAS Google Scholar
Weinreich, M. A., Odumade, O. A., Jameson, S. C. & Hogquist, K. A. T cells expressing the transcription factor PLZF regulate the development of memory-like CD8⁺ T cells. Nat. Immunol. 11, 709–716 (2010).
Article PubMed PubMed Central CAS Google Scholar
White, J. T., Cross, E. W. & Kedl, R. M. Antigen-inexperienced memory CD8⁺ T cells: where they come from and why we need them. Nat. Rev. Immunol. 17, 391–400 (2017).
Article PubMed PubMed Central CAS Google Scholar
Ju, Y. J. et al. Self-reactivity controls functional diversity of naive CD8⁺ T cells by co-opting tonic type I interferon. Nat. Commun. 12, 6059 (2021).
Article PubMed PubMed Central CAS Google Scholar
McCaughtry, T. M., Wilken, M. S. & Hogquist, K. A. Thymic emigration revisited. J. Exp. Med. 204, 2513–2520 (2007).
Article PubMed PubMed Central CAS Google Scholar
Anderson, M. S. et al. Projection of an immunological self shadow within the thymus by the Aire protein. Science 298, 1395–1401 (2002).
Article PubMed CAS Google Scholar
Klein, L. & Petrozziello, E. Antigen presentation for central tolerance induction. Nat. Rev. Immunol. 25, 57–72 (2025).
Article PubMed CAS Google Scholar
Sosinowski, T. et al. CD8alpha⁺ dendritic cell trans presentation of IL-15 to naive CD8⁺ T cells produces antigen-inexperienced T cells in the periphery with memory phenotype and function. J. Immunol. 190, 1936–1947 (2013).
Article PubMed CAS Google Scholar
Hussain, T. et al. Helminth infection-induced increase in virtual memory CD8 T cells is transient, driven by IL-15, and absent in aged mice. J. Immunol. 210, 297–309 (2023).
Article PubMed CAS Google Scholar
Xing, Y., Jameson, S. C. & Hogquist, K. A. Thymoproteasome subunit-β5T generates peptide-MHC complexes specialized for positive selection. Proc. Natl Acad. Sci. USA 110, 6979–6984 (2013).
Article PubMed PubMed Central CAS Google Scholar
Takada, K. et al. TCR affinity for thymoproteasome-dependent positively selecting peptides conditions antigen responsiveness in CD8⁺ T cells. Nat. Immunol. 16, 1069–1076 (2015).
Article PubMed PubMed Central CAS Google Scholar
Tomaru, U. et al. Restricted expression of the thymoproteasome is required for thymic selection and peripheral homeostasis of CD8⁺ T cells. Cell Rep. 26, 639–651 (2019).
Article PubMed CAS Google Scholar
Sasaki, K. et al. Thymoproteasomes produce unique peptide motifs for positive selection of CD8⁺ T cells. Nat. Commun. 6, 7484 (2015).
Article PubMed CAS Google Scholar
Ohigashi, I. et al. The thymoproteasome hardwires the TCR repertoire of CD8⁺ T cells in the cortex independent of negative selection. J. Exp. Med. 218, e20201904 (2021).
Article PubMed PubMed Central CAS Google Scholar
Steier, Z. et al. Single-cell multiomic analysis of thymocyte development reveals drivers of CD4⁺ T cell and CD8⁺ T cell lineage commitment. Nat. Immunol. 24, 1579–1590 (2023).
Article PubMed PubMed Central CAS Google Scholar
Kappes, D. & Wiest, D. L. Doubling down to make killer T cells. Nat. Immunol. 24, 1407–1408 (2023).
Article PubMed Google Scholar
Miller, C. H. et al. Eomes identifies thymic precursors of self-specific memory-phenotype CD8⁺ T cells. Nat. Immunol. 21, 567–577 (2020).
Article PubMed PubMed Central CAS Google Scholar
Kojo, S. et al. Constitutive CD8 expression drives innate CD8⁺ T-cell differentiation via induction of iNKT2 cells. Life Sci. Alliance 3, e202000642 (2020).
Article PubMed PubMed Central Google Scholar
Nomura, A. & Taniuchi, I. The role of CD8 downregulation during thymocyte differentiation. Trends Immunol. 41, 972–981 (2020).
Article PubMed CAS Google Scholar
Kuroda, N. et al. Development of autoimmunity against transcriptionally unrepressed target antigen in the thymus of Aire-deficient mice. J. Immunol. 174, 1862–1870 (2005).
Article PubMed CAS Google Scholar
Koller, B. H., Marrack, P., Kappler, J. W. & Smithies, O. Normal development of mice deficient in beta 2 M, MHC class I proteins, and CD8⁺ T cells. Science 248, 1227–1230 (1990).
Article PubMed CAS Google Scholar
Smiley, S. T., Kaplan, M. H. & Grusby, M. J. Immunoglobulin E production in the absence of interleukin-4-secreting CD1-dependent cells. Science 275, 977–979 (1997).
Article PubMed CAS Google Scholar
Fung-Leung, W. P. et al. CD8 is needed for development of cytotoxic T cells but not helper T cells. Cell 65, 443–449 (1991).
Article PubMed CAS Google Scholar
Noben-Trauth, N. et al. An interleukin 4 (IL-4)-independent pathway for CD4⁺ T cell IL-4 production is revealed in IL-4 receptor-deficient mice. Proc. Natl Acad. Sci. USA 94, 10838–10843 (1997).
Article PubMed PubMed Central CAS Google Scholar
Liou, Y. H. et al. Adipocyte IL-15 regulates local and systemic NK cell development. J. Immunol. 193, 1747–1758 (2014).
Article PubMed CAS Google Scholar
Yu, W. et al. Continued RAG expression in late stages of B cell development and no apparent re-induction after immunization. Nature 400, 682–687 (1999).
Article PubMed CAS Google Scholar
Egawa, T. & Littman, D. R. ThPOK acts late in specification of the helper T cell lineage and suppresses Runx-mediated commitment to the cytotoxic T cell lineage. Nat. Immunol. 9, 1131–1139 (2008).
Article PubMed PubMed Central CAS Google Scholar
Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013).
Article PubMed CAS Google Scholar
Li, B. & Dewey, C. N. RSEM: accurate transcript quantification from RNA-seq data with or without a reference genome. BMC Bioinformatics 12, 323 (2011).
Article PubMed PubMed Central CAS Google Scholar

Download references

Acknowledgements

We are grateful to D. Singer, M. Kimura, J.-H. Park, X. Tai and J. Dicarlo for critical reading of the paper; K. Bisht and T. Zhang for expert flow cytometry; B. Tran and J. Shetty from CCR Sequencing Facility for sequencing; D. Kovalovsky for PLZF^KO mice; M. Nussenzweig for Rag-GFP Tg mice; D. R. Littman for Runx3d-YFP knock-in mice; R. Bosselut for ThPOK-GFP Tg mice; and the National Institutes of Health Tetramer Core Facility for providing mouse CD1d tetramer. This work was supported by the Intramural Research Program of the National Cancer Institute, Center for Cancer Research, National Institutes of Health.

Author information

Authors and Affiliations

Experimental Immunology Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA
Miho Shinzawa, Nicole Ramos, Khanh Bui, William Hajjar, Assiatu Crossman & Alfred Singer
CCR Collaborative Bioinformatics Resource, Office of Science and Technology Resources, Office of Director, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA
Xiongfong Chen & Margaret Cam
CCR-SF Bioinformatics Group, Advanced Biomedical Computational Science, Biomedical Informatics and Data Science Directorate, Frederick National Laboratory for Cancer Research, Frederick, MD, USA
Xiongfong Chen
Thymus Biology Section, Experimental Immunology Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA
Yousuke Takahama

Authors

Miho Shinzawa
View author publications
Search author on:PubMed Google Scholar
Nicole Ramos
View author publications
Search author on:PubMed Google Scholar
Khanh Bui
View author publications
Search author on:PubMed Google Scholar
William Hajjar
View author publications
Search author on:PubMed Google Scholar
Assiatu Crossman
View author publications
Search author on:PubMed Google Scholar
Xiongfong Chen
View author publications
Search author on:PubMed Google Scholar
Margaret Cam
View author publications
Search author on:PubMed Google Scholar
Yousuke Takahama
View author publications
Search author on:PubMed Google Scholar
Alfred Singer
View author publications
Search author on:PubMed Google Scholar

Contributions

M.S. designed and performed the experiments, analyzed data and wrote the manuscript; N.R., K.B., W.H. and A.C. performed experiments; X.C. and M.C. analyzed RNA-seq data; Y.T. provided materials and valuable discussions; A.S. conceived the study, designed the experiments, analyzed data and wrote the manuscript.

Corresponding author

Correspondence to Alfred Singer.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Immunology thanks Juan Carlos Zuniga-Pflucker, David Wiest and Stephen Jameson for their contribution to the peer review of this work. Primary Handling Editor: S. Houston, in collaboration with the Nature Immunology team.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 T cell development in CD8^Dual mice.

a. Flow cytometry analysis of CD8^Dual (n = 16) and B6 (n = 15) thymocytes (representative of 15 independent experiments). b. Flow cytometry analysis of CD8^Dual (n = 16) and B6 (n = 15) LN T cells (representative of 15 independent experiments). Numbers (mean±s.e.m.) of total LN cells are shown above top histogram. c. Quantification of mature TCR^hi thymocytes and LN T cells, related to Fig. 1c. ***P < 0.001, **P < 0.01 (two-tailed unpaired t-tests); mean±s.e.m. d. CD8β expression of CD24⁻TCR⁺ mature thymocytes and TCR⁺ LN T cells from CD8^Dual mice with the indicated MHC deficiencies, related to Fig. 1d. e. Runx3d-YFP and ThPOK-GFP reporter expression among TCR^hi thymocytes and TCR⁺ LN T cells in CD8^Dual and B6 mice, related to Fig. 1e. Numbers within profiles and histograms indicate frequency of cells in each box or division (a,b,d).

Extended Data Fig. 2 Characterization of CD8^Dual T cells.

a. MFI of surface proteins on TCR^hi thymocytes from CD8^Dual and B6 mice (n = 5/strain, 5 experiments). b. MFI of Gata3 in LN T cells from CD8^Dual and B6 mice (n = 3/strain, 3 independent experiments). c. QPCR analysis of mRNA expression in electrically sorted LN T cells from CD8^Dual and B6 mice. Results are normalized to control gene Rpl13 (n = 3/strain, 3 independent experiments with technical triplicates). ***P < 0.001, **P < 0.01, *P < 0.05 (two-tailed unpaired t-tests); mean±s.e.m (a-c).

Extended Data Fig. 3 Functional assessment of CD8^Dual T cells.

a. Intracellular staining (ic) of IFN-γ and IL-4 on CD8^Dual CD8.1 T cells and B6 CD8 T cells in Th2 culture conditions. Electrically sorted naïve T cells (CD62L⁺CD44⁻) from pooled LNs were cultured in Th2 conditions for 5 days. At day5, cells were stimulated with PMA+Ionomycin for 4 h with Golgi stop (n = 3/strain, representative of 3 independents). b. CD45.2 donor LN T cells from CD8^Dual (n = 7) and B6 (n = 6) mice were injected into sublethally irradiated CD45.1 B6 WT host mice and their LNs were characterized on day 7 (representative of 3 independent experiments). c. Lymphopenia-induced homeostatic proliferation of CD8^Dual (n = 7) and B6 (n = 6) LN T cells. CD45.2 donor LN T cells that were injected into sublethally irradiated CD45.1 B6 WT host mice were assayed at day 7 for proliferation based on CTV dilution profile (representative of 3 independent experiments). Right, frequency of CTV dilution was analyzed (right, D0: no division, D1: one division, D2: two divisions, D3: three divisions, and >D4: more than four divisions). d. MFI of GM1 in LN T cells from CD8^Dual and B6 mice (n = 3/strain, 3 independent experiments). Numbers within profiles indicate frequency of cells in each box (a,b). ***P < 0.001, **P < 0.01, *P < 0.05 (two-tailed unpaired t-tests); mean±s.e.m (c,d).

Extended Data Fig. 4 Agonist-selected T cells in CD8^Dual mice.

a. Foxp3 (ic) and CD25 staining among mature CD8.2 and CD8 thymocytes from CD8^Dual and B6 mice, related to Fig. 2c. b. CD1d-TET and TCRβ staining of whole thymocytes from CD8^Dual and LM control WT mice, related to Fig. 2d. c. Flow cytometry analysis of CD1d-TET⁺CD24⁻ thymic mature NKT cells from CD8^Dual (green) and LM control WT (gray) mice (n = 3/strain, representative of 3 independent experiments). d. Intracellular staining (ic) of T-bet in NKT subsets from CD8^Dual and LM control WT mice (n = 3/strain, representative of 3 independent experiments). e. Intracellular staining (ic) of Gata3 in NKT subsets from CD8^Dual and LM control WT mice (n = 3/strain, representative of 3 independent experiments). f. IFN-γ, IL-4, and IL-17 expression in CD1d-TET⁺CD24⁻ thymic mature NKT cells from CD8^Dual and LM control WT mice. Whole thymocytes were cultured with medium or PMA+Ionomycin in the presence of golgi stop for 4 h and assessed for cytokine production (n = 3/strain, representative of 3 independent experiments). Numbers within profiles indicate frequency of cells in each box (a-c,f).

Extended Data Fig. 5 Schematic of CD8 T-lineage fate decisions in the CD8^Dual thymus.

a. CD8 T-lineage fate decisions in the CD8^Dual thymus. TCR signaling transcriptionally terminates Cd8 gene expression (not Cd4 gene expression) which reduces surface expression of Cd8-encoded CD8 coreceptors but upregulates expression of Cd4-encoded CD8 coreceptors. Consequently, MHC-I TCR signaling that becomes disrupted allows thymic cytokines to induce the cytotoxic factor Runx3d that silences Cd4 gene expression, re-initiates Cd8 gene expression, and promotes thymocyte differentiation into cytotoxic T cells. In contrast, MHC-I TCR signaling that persists undisrupted induces the helper factor ThPOK that promotes thymocyte differentiation into helper T cells. b. MHC-I peptides affect CD8-T lineage fate decisions in the CD8^Dual thymus. cTECs in the thymic cortex express two classes of peptides which we refer to as β5t- and nonβ5t-peptides which display different patterns of expression. β5t-peptides are exclusively expressed on cTECs in the thymic cortex, whereas nonβ5t-peptides are expressed throughout the thymus. Consequently, TCR engagement of β5t-peptides on cTECs in the thymic cortex becomes permanently disrupted when thymocytes disengage from cTECs and migrate through the thymus, which induces Runx3d expression and results in generation only of cytotoxic CD8 T cells. In contrast, TCR engagement of nonβ5t-peptides on thymic cells in the cortex or CMJ can persist when thymocytes disengage from cTECs and migrate through the thymus, which induces ThPOK expression and results in generation of helper CD8 T cells.

Extended Data Fig. 6 Selection of cytotoxic CD8 T cells by β5t-peptides.

a. Flow cytometry of CD24⁻TCR⁺ mature thymocytes and TCR⁺ LN T cells from β5t^WT and β5t^KO CD8^Dual mice, related to Fig. 5a, b. Numbers (mean±s.e.m.) of mature thymocytes and LN T cells are shown above profiles. b. MFI of ThPOK and Runx3 in CD8.1 LN T cells from β5t^WT and β5t^KO CD8^Dual mice (n = 5/strain, 4 independent experiments). c. CD40L and CD69 expression on in vitro stimulated CD8.1 LN T cells from β5t^WT and β5t^KO CD8^Dual mice. MACS-purified T cells from pooled LNs were cultured with plate-bound anti-CD3 + CD28 mAbs for 24 h. Dot line indicates CD8.1 T cells cultured with medium for 24 h (n = 3/group, representative of 3 independent experiments). d. Foxp3 (ic) and CD25 staining and numbers of Foxp3⁺CD25⁺ mature Tregs among CD24⁻TCR⁺ mature CD8.1 thymocytes from β5t^WT and β5t^KO CD8^Dual mice (n = 5/strain, 4 independent experiments). e. Numbers of OT-I CD8.1 and OT-I CD8.2 T cells among Vα2⁺CCR7⁺ mature thymocytes from β5t^WT (n = 4) and β5t^KO OT-I. CD8^Dual Rag^KO mice (n = 3, 3 independent experiments). Numbers within profiles indicate frequency of cells in each box (a,d). ***P < 0.001, **P < 0.01, *P < 0.05 (two-tailed unpaired t-tests); mean±s.e.m (b,d,e).

Extended Data Fig. 7 Selection of IM-CD8 T cells by nonβ5t-peptides.

a. Numbers of total, CV ((Runx3⁺Eomes⁻), and IM (Runx3⁺Eomes⁺) CD8.2 T cells among mature thymocytes from LM control IL-4R heterozygous (IL-4R^HE: n = 5) and IL-4R^KO (n = 6) CD8^Dualβ5t^KO mice, related to Fig. 5c, third profile (5-6 independent experiments). b. Numbers of total, CV, and IM CD8.2 T cells among mature thymocytes from LM control IL-15^HE (n = 6) and IL-15^KO (n = 8) CD8^Dual β5t^KO mice, related to Fig. 5c, fourth profile (5 independent experiments). c. Runx3 and Eomes expression on mature CD8.2 thymocytes from indicated CD8^Dual mice (IL-4R^KO: n = 7, CD1d^KO: n = 3, PLZF^KO: n = 5, IL-15^KO: n = 3, 3-4 independent experiments). d. Flow cytometry of mature CD8.2 thymocytes from LM control β5t^WT and β5t^KO CD8^Dual mice (n = 3/strain, 3 independent experiments). e. IFN-γ expression by TCR^hi CD8.2 thymocytes from LM control β5t^WT and β5t^KO CD8^Dual mice. Whole thymocytes were cultured with medium or PMA+Ionomycin for 4 h in the presence of golgi stop (n = 3/strain, 3 independent experiments). f. QPCR analysis of IL-4 mRNA in whole thymocytes from indicated mice. Results are normalized to control gene Rpl13 (n = 3/strain, 3 independent experiments with technical triplicates). g. PLZF and RORγt expression of thymic mature NKT cells from LM control β5t^WT (n = 4) and β5t^KO (n = 5) CD8^Dual mice (3 independent experiments). h. QPCR analysis of IL-4 mRNA in whole thymocytes from LM control β5t^WT and β5t^KO CD8^Dual mice. Results are normalized to control gene Rpl13 (n = 3/strain, 3 independent experiments with technical triplicates). i. Runx3 and Eomes expression on mature CD8.2 thymocytes from LM control Aire^WT and Aire^KO CD8^Dual mice (n = 8/strain, 4-5 independent experiments). j. IM-CD8 T cell generation requires late-TCR signaling by nonβ5t-peptides. If TCR engagement of nonβ5t-peptides on cTECs is transiently disrupted by reduced CD8 coreceptor expression, thymocytes express Runx3d and become cytotoxic CD8 T cells. Because Runx3d re-initiates Cd8 gene expression and upregulates surface CD8 coreceptor expression, TCR can subsequently re-engage nonβ5t-peptides when they encounter them on non-cortical thymic elements, stimulating late-TCR signaling. Late-TCR signaling may delay thymocytes from leaving the thymus which prolongs their exposure to IL-4 produced by NKT2 cells in the thymic medulla. Numbers within profiles indicate frequency (c-e,g,i). ***P < 0.001, **P < 0.01 (two-tailed unpaired t-tests); mean±s.e.m (a,b,f,h).

Extended Data Fig. 8 IM CD8 T cells in WT mice also require nonβ5t-peptides.

a. Flow cytometry of CD24⁻TCR⁺ mature thymocytes and numbers of mature CD8 thymocytes from LM control β5t^WT (n = 5) and β5t^KO BALB/c (n = 8) mice (representative of 4-5 independent experiments). b. Flow cytometry of CD24⁻TCR⁺ mature thymocytes and numbers of mature CD8 thymocytes from LM control β5t^WT (n = 4) and β5t^KO (n = 6) B6 mice (representative of 4-6 independent experiments). c. Flow cytometry of TCR⁺ LN T cells and numbers of CD8 LN T cells from LM control β5t^WT (n = 3) and β5t^KO (n = 4) B6 mice (representative of 3-4 independent experiments). d. Intracellular staining (ic) of Runx3 and Eomes in CD8 T cells among TCR⁺ LN T cells from LM control β5t^WT (n = 5) and β5t^KO (n = 10) B6 mice (representative of 5-8 independent experiments). e. Numbers of CV (Runx3⁺Eomes⁻) and IM (Runx3⁺Eomes⁺) CD8 T cells among TCR⁺ LN T cells from LM control β5t^WT and β5t^KO B6 mice (d). Numbers within profiles indicate frequency of cells in each box (a-d). ***P < 0.001, **P < 0.01, *P < 0.05 (two-tailed unpaired t-tests); mean±s.e.m (a-c,e).

Extended Data Fig. 9 IM CD8 T cell development in B6 and BALB/c WT mice.

Nonβ5t-selected cytotoxic CD8 T cells re-encounter nonβ5t-peptides after migrating outside of the cortex and receive late TCR signaling in both BALB/c and B6 thymi. In BALB/c thymi with an abundance of IL-4-producing NKT2 cells, late-TCR signaled CD8 T cells respond to intra-thymic IL-4 which promotes Eomes expression, proliferation, and IM CD8 T cell fate. In contrast, in B6 thymi with few IL-4-producing NKT2 cells, most late-TCR signaled CD8 T cells do not acquire Eomes expression, do not proliferate, and do not acquire IM fate in the thymus. However, after late-TCR signaled CD8 T cells leave the B6 thymus and become peripheral CD8 T cells, we think they encounter IL-15 in the periphery which induces Eomes expression and conversion to IM CD8 T cells.

Supplementary information

Reporting Summary

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Shinzawa, M., Ramos, N., Bui, K. et al. Unraveling CD8 lineage decisions reveals that functionally distinct CD8⁺ T cells are selected by different MHC-I thymic peptides. Nat Immunol (2026). https://doi.org/10.1038/s41590-025-02411-4

Download citation

Received: 04 September 2025
Accepted: 18 December 2025
Published: 19 January 2026
Version of record: 19 January 2026
DOI: https://doi.org/10.1038/s41590-025-02411-4