Antigen-presenting T cells provide critical B7 co-stimulation for thymic iNKT cell development via CD28-dependent trogocytosis (VSports手机版)

V体育ios版 - B7-CD28 co-stimulation is important for thymic development of all iNKT subsets

We first examined the role of B7-CD28 co-stimulation in iNKT subset development in multiple mouse strains. Thymic iNKT cells were detected by PBS57:CD1d-tetramer binding followed by sequential gating strategy (Figure S1A), and NKT subsets were defined by expression of PLZF, T-bet, and RORγt (Figure 1A). NKT1 is dominant in B6 thymus, and NKT2 is dominant in BALB, while CB6F1 showed an intermediate phenotype (Figures 1B and S1). For all genetic backgrounds, absence of B7 (Figure 1) or CD28 (Figure S1B) resulted in significant decrease of all iNKT subsets (Figure 1B). Cytokine-producing capability of the residual iNKT subsets that developed in B7-CD28-deficient mice was assessed after 4-h stimulation with PMA + ionomycin. The proportion of cytokine-producing cells within each iNKT subset was overall normal or slightly reduced relative to wild-type controls, with statistically significant decreases in frequency of IFNγ production by NKT1 and of IL-17 production by NKT17 (Figure 1C upper panel), while the number of cytokine-producing iNKT cells in all subsets, as a proportion of total thymocytes, was significantly and substantially decreased (Figure 1C lower panel).

We next determined whether, in addition to affecting the number of developing iNKT cells, B7-CD28 co-stimulation plays a role in selection of the TCR repertoire of iNKT cells. It is known that mouse thymic iNKT cells express an invariant TCRα in association with preferentially expressed TCR Vβs (Vβ8, Vβ2, Vβ7), and that iNKT subsets show substantially different patterns in usage of these Vβs.¹⁷ We assessed TCR Vβ expression by single-cell 5′ RNA-seq in BALB/c wild-type (WT) and B7 double knockout (DKO) iNKT cells. TCR Vβ usage was significantly altered in B7 DKO iNKT cells, and these effects were selective for individual iNKT subsets (Figure S1C). These results demonstrated a critical role of B7-CD28 co-stimulation for development of all iNKT subsets and for TCR repertoire selection in these subsets.

Strong B7-CD28 interaction occurs during early stages of iNKT cell development

To elucidate underlying cellular and molecular mechanisms of B7 co-stimulation, we first sought to assess B7-CD28 interaction over the course of iNKT cell development. To this end, we defined iNKT cell developmental stages using CD44, CD24, and CD69 expression (Figure 2A). Since higher CD44 expression is a hallmark of mature iNKT cells, we first posited that CD44^low cells represent developing iNKT cells, while CD44^high represents mature thymic iNKT cells.¹⁸ Within the CD44^low fraction, we analyzed CD24 (HSA), for which decreasing expression level correlates with increasing maturation of thymic iNKT cells, and CD69, which is an indicator of TCR signaling during thymic iNKT positive selection.¹⁸ We postulated a developmental sequence starting with (1) CD44^lowCD24^highCD69⁻, followed by (2) CD44^lowCD24^highCD69⁺, (3) CD44^lowCD24^intCD69⁺, (4) CD44^lowCD24^lowCD69⁻, and then (5) CD44^high. This sequence of developmental stages was confirmed by comparing expression level of Rag2-GFP as a molecular timer for residency time following positive selection and termination of Rag2 gene expression³⁰ (Figure 2B). We defined the stages analyzed in this study as (1) preselection (prior to TCR signaling), (2) stage 1, (3) stage 2, (4) stage 3, and (5) mature, as shown in Figure 2C. To evaluate thymic localization of each stage of iNKT, CCR7 expression was analyzed. CCR7 is induced on positively selected thymocytes, and its expression is important for cortical to medullary migration.^31–33 CD69 and CCR7 expression were inversely correlated during early stages of iNKT development. CD69 was upregulated in stage 1 relative to preselection iNKT and remained high in stage 2 but was downregulated in stage 3 (Figure 2B), while CCR7 was not expressed in stage 1, was upregulated in stage 2, and increased in expression at stage 3 (Figure 2D). Egr2, an early TCR signal response gene shown to be important for NKT cell development, showed kinetics of expression similar to that of CD69 (Figure S2A). These results suggested that stage 1 (CD69⁺CCR7⁻) and stage 2 (CD69⁺ CCR7^low) iNKT cells were under positive selection by TCR signaling and localized to the cortex, while stage 3 (CD69⁻CCR7^high) iNKT cells might be migrating and/or have migrated to the medulla as a result of higher CCR7 expression^31,33 (Figures 2B and 2D).

We evaluated the presence of ongoing B7-CD28 interaction at each developmental stage by measuring B7-dependent downregulation of cell surface CD28, as previously reported.³⁴ In WT mice, CD28 was upregulated upon positive selection (stage 1), and its level subsequently decreased during maturation (Figure 2E). To assess B7-dependent CD28 downregulation, CD28 expression level at each stage of iNKT cell development was measured in B7 DKO mice. After CD28 upregulation at stage 1, its expression level was further upregulated in B7 DKO thymocytes through stage 2, after which expression decreased (Figure 2E). At each stage of differentiation, CD28 expression was greater in B7 DKO thymocytes than in WT. As a quantitative index of B7-CD28 interaction, the difference in mean fluorescence intensity (MFI) between B7 DKO and WT was plotted. This evaluation indicated that B7-CD28 interaction, as reflected by B7-dependent downregulation of CD28,³⁴ had already been initiated at stage 1, peaked at stage 2, and decreased at later stages of development (Figure 2F). These results indicated that strong CD28 upregulation and B7-CD28 interaction occurs during positive selection, likely in the thymic cortex.

B7-CD28 co-stimulation is critical for the developmental transition from stage 1 to stage 3 of iNKT cells at an early stage of thymic selection

To determine the effect of B7-CD28 co-stimulation on iNKT cell development, cell numbers at each developmental stage were analyzed in WT and in co-stimulation-deficient thymus. iNKT cell numbers were comparable in WT and co-stimulation-deficient thymus at preselection (CD24^highCD44^lowCD69⁻) and stage 1 (CD24^highCD44^lowCD69⁺), while cell numbers were significantly reduced in co-stimulation-deficient thymus at stage 2 (CD24^intCD44^lowCD69⁺), and markedly reduced in stage 3 (CD24^lowCD44^lowCD69⁻) and mature (CD24^lowCD44^high) iNKT cells (Figure 3A). These results indicated that B7-CD28 co-stimulation is critical for the developmental transition from stage 1 to stage 2 and stage 3 of iNKT cell development.

The underlying molecular mechanism(s) mediating CD28 co-stimulation for thymic iNKT cell development are not well understood. For global analysis of developing thymic iNKT cells, we performed transcriptomic characterization by scRNAseq of B7 WT and DKO thymic iNKT cells. Since B7-CD28 deficiency resulted in larger reduction in NKT cell number in BALB/c compared with B6, and BALB/c has more balanced composition of NKT1/2/17 subsets compared with B6 (dominated by NKT1), BALB/c mice were used for scRNAseq analysis. To allow concurrent analysis of sufficient numbers of both the quantitatively predominant CD44^high mature iNKT cells and the much less frequent CD44^low immature iNKT cells, we flow cytometrically sorted mature (CD44^high) and immature (CD44^low) iNKT cells separately; then we mixed these two populations in equal proportion for scRNAseq analysis. Unsupervised clustering analysis identified 13 clusters based on differential patterns of gene expression. These corresponded to a preselection cluster, a stage 1/2 signature (CD28^highCD69^highEgr1/2^high) cluster, proliferating clusters, and NKT1, NKT2, and NKT17 clusters (Figures 3B and 3C). B7 WT and B7 DKO showed similar cluster composition (Figure S3A). Clusters 5, 8, and 9 were classified as proliferating, with cluster 5 expressing an S phase cell cycle signature and clusters 8 and 9 expressing G2/M signatures (Figure 3C). The cell cycle signatures of the clusters presented in Figure 3B are shown in Figure 3D. Since a clonal burst of developing iNKT cells occurs after positive selection and before the mature iNKT cell stage,³⁵ and since we observed that a decrease in B7 DKO versus WT cell numbers became apparent in the transition from stage 2 to stage 3, we hypothesized that lack of B7 co-stimulation might impact gene expression related to proliferation and cell cycle in developing iNKT cells. In cluster 5 (S phase cell cycle signature), pathways including T cell proliferation (GO: 0042129) and negative regulation of apoptotic signaling pathway (GO: 2001234) were upregulated in WT relative to B7 DKO (Figure 3E), suggesting a role for B7-CD28 co-stimulation in regulating proliferation and apoptosis in proliferating thymic NKT cells. Of note, in differentially expressed gene analysis, Ckb (creatine kinase B) expression was consistently lower in B7 DKO NKT, while Crip1 (cysteine rich protein 1) was higher in B7 DKO NKT (Figures S3B and S3C), suggesting these as potential candidate genes in future analysis.

CD1d-expressing thymocytes acquire B7 via CD28-dependent trogocytosis (VSports手机版)

We next sought to determine the cell type(s) involved in B7 presentation to iNKT cells during development. Since CD1d-expressing DP thymocytes are uniquely required as antigen-presenting cells during iNKT cell selection,^4–7 two possible mechanisms of B7 co-stimulation appeared possible, both unconventional. It was possible that B7 co-stimulation is provided in cis by the same CD1d-expressing DP T cells that function in positive selection of iNKT cells. This would be unconventional in that B7 expression and co-stimulation by DP thymocytes has not been previously identified. Alternatively, it was possible that co-stimulation is provided in trans, i.e., by a cell distinct from the CD1d-expressing DP thymocytes that are required to provide the first or TCR-mediated signal for iNKT cell development. Potential sources of B7 co-stimulation in this case might include TECs, DCs, B cells, and single-positive (SP) T cells, noting that there is no requirement for CD1d expression on any of these cells for iNKT cell development.^5,7,35 Either case would represent a quite non-conventional mode of B7-CD28 co-stimulation.

Flow cytometric analysis demonstrated that cell surface B7 is indeed expressed on thymic DP and SP T cells (Figures 4A and S4A), in addition to the expected expression of B7 on TECs, DCs, and B cells (Figure S4B). To address cellular requirements for B7 co-stimulation in iNKT cell development, we utilized a B7 conditional knockout (KO) strategy.³⁶ Thymic iNKT cell development in B7^flox mice is comparable to that of WT B6 mice (Figure S4C). B7 was deleted from specific cell types of B7^flox mice, including TECs, DCs, B cells, or T cells by using Foxn1-Cre, CD11c-Cre, CD19-Cre²⁶ (Figure S4B), or proximal-Lck-Cre, respectively. However, none of these conditional knockout (cKO) strains showed a defect in thymic iNKT cell development (Figures 4B and 4C). In contrast, B7 deletion from all TECs, DCs, and B cells in triple-Cre mice resulted in abolishment of B7-dependent iNKT cell development to a similar level as that observed in B7 complete KO mice (Figure 4B). This result might indicate that B7 on TECs, DCs, and B cells has a redundant capacity to mediate B7 co-stimulation in trans for iNKT cell development. However, on careful evaluation of the cell type specificity of B7 deletion, we noted that B7 expression on peripheral T cells was eliminated in TEC/DC/BC-B7 triple-Cre cKO mice, while B7 expression on T cells was not eliminated by genomic deletion of B7 in T cell-specific B7cKO mice (Figure 4D). It has been reported that B7 can be acquired on peripheral T cells as a result of interaction of T and antigen-presenting cells (APCs) through CD28-dependent trogocytosis.^37,38 We therefore tested the possibility that this mechanism could be involved in the observed B7 expression on T cells, using a mixed bone marrow (BM) chimera strategy. Indeed, B7-gene-deficient peripheral T cells derived from B7 DKO BM acquired B7 molecules in vivo when B7-expressing cells co-existed in these mixed chimeras (Figure 4E). This B7 acquisition was CD28 dependent, as B7 was not acquired by CD28 KO T cells (Figure 4F). Notably, CD28-dependent B7 trogocytosis was not restricted to peripheral T cells, since B7 expression on thymocytes, including DP and SP cells, was also observed and was similarly CD28 dependent (Figure 4A). Of interest, truncation of the CD28 cytoplasmic tail did not affect B7 expression on thymocytes, indicating that the CD28 intracellular domain was not required for B7 acquisition (Figure 4G).

"VSports手机版" CD28 expression on CD1d-expressing antigen-presenting T cells is required for NKT cell development

Based on these observations, we hypothesized that CD28-dependent B7 acquisition on T cells is a critical mechanism for providing B7 co-stimulation to developing thymic iNKT cells. This possibility was tested in mixed BM chimeras. One donor bone marrow (BM1) was CD1d KO (CD45.1), and cells derived from this BM were used as a readout for iNKT cell development, since these cells alone would not be able to support iNKT cell development due to lack of CD1d. The second donor bone marrow (BM2) in mixed chimeras was CD1d WT (CD45.2), to test the APC capability of this CD1d-expressing population for supporting iNKT cell development of BM1-derived cells. We tested the requirement for CD28 expression on BM2 antigen-presenting T cells by comparing CD1d WT/CD28 WT (group 2) and CD1d WT/CD28 KO (group 3) bone marrow as BM2. T cells from CD1d WT/CD28 KO bone marrow (group 3, BM2) express CD1d but are not able to acquire B7 due to lack of CD28, while DC and B cells derived from the same BM express both CD1d and B7 (Figures 5A–5C). When BM2 was CD1d KO, iNKT cells did not develop in the BM1 compartment (group 1), while wild-type (CD1d WT/CD28 WT) BM2 cells did support iNKT cell development in the BM1 compartment (group 2). When CD1d WT/CD28 KO BM was tested for APC capability to support iNKT cell development (group 3), the efficiency of iNKT cell development was markedly decreased, to a level similar to that seen in the complete absence of CD28 (group 4) (Figure 5A). We confirmed CD28 KO DP cells express CD1d and SLAM family molecules at a comparable level to those of CD28 WT DP (Figure 5D). It should be noted that in the group 3 BM chimera mice, T cells from BM1 were CD1d KO but CD28 WT, and thus able to acquire B7 through trogocytosis, and that BM2-derived cells other than T cells, including DCs and B cells, expressed both CD1d and B7 (Figure 5C); nevertheless, those cell populations were not sufficient to support B7-CD28 co-stimulation-dependent iNKT cell development (Figure 5A). Consistent with reported results,⁵ CD1d-expressing SP cells were not required for iNKT cell development, since TCRα KO cells, in which T cell development was arrested at DP stage, were still capable of providing both TCR signal and co-stimulation (Figure S5A). The requirement for CD28 expression on antigen-presenting T cells for iNKT cell development was similarly observed in BALB/c BM chimera mice (Figure S5B). Of interest, the CD28 cytoplasmic tail was dispensable for both B7 acquisition (Figure 4G) and for the ability of CD1d⁺ T cells to function as APCs in support of B7-dependent iNKT cell development (Figure S5C). Taken together, these results indicated that co-expression of CD1d and CD28 on antigen-presenting DP thymocytes is required for B7 trogocytosis and for providing B7 co-stimulation for iNKT cell differentiation.

Limitations of the study

This study identified a requirement for CD28 expression on DP thymocytes for acquisition and presentation of B7 ligand and for providing B7-CD28 co-stimulation to developing thymic iNKT cells in mouse. Since we utilized CD28 KO mice to demonstrate the importance of CD28 expression on DP cells to acquire B7, we cannot completely exclude the possibility that CD28 KO DP cells as APCs have cell intrinsic defect(s) other than B7 acquisition to support iNKT cell development, despite our inability to identify any such differences. We showed that thymic DCs, B cells, and TECs have redundant roles as sources of B7 for trogocytosis by utilizing B7.1 cKO (B7.2 KO) analysis. This B7.1 cKO system can test the cell type-specific role of B7.1 but cannot address the cell type-specific role of B7.2, if any. It is known that iNKT cell subset development and composition vary among mouse strains. In this regard, we incorporated three mouse strains, C57BL/6, BALB/c, and CB6F1, in this study, but in some analyses, we were unable to analyze all strains due to technical and/or resource limitation.

RESOURCE AVAILABILITY

EXPERIMENTAL MODEL AND SUBJECT DETAILS

C57BL/6 (B6) and BALB/c mice were purchased from Charles River. CB6F1 mice were bred in our facility. B6 CD45.1 congenic (B6.SJL-Ptprca Pepcb/BoyJ), BALB/c CD45.1 congenic (CByJ.SJL(B6)-Ptprca/J), B6 CD1d KO (B6.129S6-Del(3Cd1d2-Cd1d1) 1Sbp/J), BALB/c CD1d KO (C.129S2-Cd1tm1Gru/J), B6 CD28 KO (B6.129S2-Cd28tm1Mak/J), B6 B7.1/B7.2 double KO (DKO) (B6.129S4-Cd80tm1Shr Cd86tm2Shr/J), B6 CD19-Cre-KI (B6.129P2(C)-Cd19tm1(cre)Cgn/J), B6 CD11c-Cre-Tg (B6.Cg-Tg(It-gax-cre)1-1Reiz/J) and B6 Lck-Cre-Tg (B6.Cg-Tg(Lck-cre)548Jxm/J) mice were purchased from Jackson Laboratory. B6 Foxn1-Cre-Tg mice were generously provided by Dr. Georg Hollander. B6 CD28-tailless (TL) mutant mice²³ were provided by Dr. Alfred Singer and Dr. Xuguang Tai. B6 Rag2-GFP-Tg mice⁴⁵ and B6 CD28 mutant Knock-In (KI) mice²⁴ were described previously. B6 B7^flox mice were generated and described previously.³⁶ BALB/c background CD28 KO and B7 DKO strains were generated at our facility by backcrossing B6 CD28 KO and B6 B7 DKO to BALB/c at least eight generations as previously described.⁴⁶ Bone marrow chimera mice were generated by reconstitution with 10⁷ total T cell–depleted BM cells from donor mice to irradiated (950 rad one time and 450 rad two times for B6 and BALB/c hosts, respectively) host mice i.v. Chimerism of radiation BM chimeras was analyzed 10 weeks after reconstitution with CD45.1 and CD45.2 donor cells and generally ranged between 20% and 80%. Chimeras that were more highly skewed to either CD45.1 or CD45.2 donor were excluded from analysis. 6- and 10-week-old female and male mice were used for all experiments. Mice were maintained in accordance with National Institutes of Health guidelines. All animal experiments were approved by the National Cancer Institute Animal Care and Use Committees.

V体育安卓版 - METHODS DETAILS

QUANTIFICATION AND STATISTICAL ANALYSIS

For statistical analysis, Prism GraphPad 8 software was used. Student’s t test with two-tailed distribution was performed for statistical analyses with a single comparison. For multiple comparisons, statistical analysis was performed with one-way ANOVA followed by Dunnett’s multiple comparison. p values <0.05 were considered statistically significant.