Asymmetric inheritance of mTORC1 kinase activity during division dictates CD8(+) T cell differentiation

Abstract

The asymmetric partitioning of fate-determining proteins has been shown to contribute to the generation of CD8(+) effector and memory T cell precursors. Here we demonstrate the asymmetric partitioning of mTORC1 activity after the activation of naive CD8(+) T cells. This results in the generation of two daughter T cells, one of which shows increased mTORC1 activity, increased glycolytic activity and increased expression of effector molecules. The other daughter T cell has relatively low mTORC1 activity and increased lipid metabolism, expresses increased amounts of anti-apoptotic molecules and subsequently displays enhanced long-term survival. Mechanistically, we demonstrate a link between T cell antigen receptor (TCR)-induced asymmetric expression of amino acid transporters and RagC-mediated translocation of mTOR to the lysosomes. Overall, our data provide important insight into how mTORC1-mediated metabolic reprogramming affects the fate decisions of T cells.

Figure 1. mTORC1 activity is asymmetrically inherited in dividing CD8⁺ T cells upon TCR stimulation

(a) Flow cytometry analyzing adoptive transferred CD8^hi and CD8^lo eFluor450-labeled OT-I T cells gated on the first division from splenocytes of WT host mice at 48 h post LM-OVA infection. Histogram overlay of CD25, T-bet, CD62L, and p-S6 expression between CD8^hi and CD8^lo T cells. MFI, upper left corner. (b) Histogram overlay of mTOR pathway proteins between CFSE-labeled CD8^hi and CD8^lo T cells (gated as shown) that were stimulated in vitro for 36 h. MFI, upper left corner. (c) Immunoblot analysis of mTOR substrates of sorted in vitro activated CD8⁺ T cells. (d) Confocal images of dividing T cells that were activated in vitro. Quantification of the ratio of phosphorylated to total MFI of S6, n=39. (e) Quantification of MFI ratio of p-S6, β-tubulin, and CD90.2among dividing daughter pairs, n=132 (p-S6), 26 (β-tubulin and CD90.2). (f) Histogram overlay of CD8 and p-S6 expression from first-division adoptive transferred OT-I CD8⁺ T cells into either Rag2^−/− or LM-OVA infected WT host. (g) Confocal images of in vitro stimulated WT, T-Rheb^−/−, and T-Tsc2^−/− CD8⁺ daughter T cells. Statistics analysis of p-S6 MFI ratio between CD8^hi and CD8^lo daughter T cells, n=26 (WT) and 37 (T-Rheb^−/− and T-Tsc2^−/−). *P < 0.05; **P < 0.0001; NS, not significant (Wilcoxon rank test (d) or One-way ANOVA (e, g)). Data are compiled from 3 independent experiments (d, e) or one experiment representative of at least 3 independent experiments (a–c, f, g) (mean in e, g). Scale bars, 10μm (d, g).

Figure 2. Asymmetric inheritance of mTORC1 results in CD8⁺ T daughter cells with distinct metabolic capabilities

(a–b) CD8^hi and CD8^lo T cells from in vitro activation were assayed for ECAR (a) and OCR (b). Summary of basal ECAR, maximum ECAR, and SRC are shown on right, n=8 (a), 6 (b). (c) As in Fig. 1a, Histogram overlay showing MYC expression between CD8^hi and CD8^lo T cells. MFI, upper right corner. (d) Immunoblot analysis comparing expression of proteins involved in glycolytic pathway between in vitro stimulated CD8^hi and CD8^lo T cells. (e) Confocal images of in vitro stimulated daughter CD8⁺ T cells from GFPMYC mice stained with CD8 and p-S6. Statistical analysis of GFPMYC MFI between CD8^hi and CD8^lo T cells (right), n=18. (f) Quantification of mitochondrial and genomic DNA ratio between sorted in vitro activated CD8^hi and CD8^lo T cells, n=12. (g) Immunoblot analysis showing expression of mitochondrial associated proteins comparing between CD8^hi and CD8^lo T cells as in (d). (h) Confocal images of in vitro activated CD8^hi and CD8^lo T cells stained with MitoTracker dye and DAPI. CD8^hi T cells (green) are further stained with a different CD8 antibody clone post-sort to distinguish between CD8^lo T cells. Statistical analysis of MitoTracker MFI between CD8^hi and CD8^lo cells, n=60. *P < 0.05; **P < 0.005; ***P < 0.0005 (Mann-Whitney t test (a, b, f, h) or Wilcoxon rank test (e)). Data are from one experiment representative of at least 3 independent experiments (a–e, g) or compilation of 3 independent experiments (f) (mean ± s.d. in a, b left, mean ± s.e.m. in a, b, f). Scale bars, 10μm (e, h).

Figure 3. Asymmetric inheritance of mTORC1 results in CD8⁺ T daughter cells with different in vivo survival

PMA + Ionomycin-activated or sorted CD8^hi and CD8^lo T cells from in vitro activation were adoptively transferred into congenically distinct WT host, and then were infected with vaccinia-OVA the same day (a) or 21 days later (b). Six days after infection (day 6 or day 27), spleens were harvested and quantification of CD90.2⁺ CD8⁺ T cells was assessed by flow cytometry. Statistical quantification of percentage and absolute number of recovered CD90.2⁺ CD8⁺ T cells between CD8^hi, CD8^lo, and PMA + Iono activated T cells 6 days or 27 days after the adoptive transfer (six days after infection) are shown, n= 7,4,7 (a, CD8^hi, CD8^lo, PMA+Iono), 5,5,4 (b, CD8^hi, CD8^lo, PMA+Iono). *P < 0.05; **P < 0.005; NS, not significant (One-Way ANOVA). Data are from one experiment representative of at least 4 independent experiments (mean).

Figure 4. The differential RagC mediated translocation of mTOR to the lysosome contributes to the asymmetric division of CD8⁺ T cells

(a) OT-I Rag2^−/− splenocytes were stimulated in vitro with OVA-I peptide for 0, 4, or 24 h. RagC and LAMP-2 localization was assessed by confocal microscopy. Statistical analysis of co-localization of RagC and LAMP-2 (r) was assessed, n=30, 22, 23 (0, 4, 24 h). (b) Same as (a), except co-localization analysis for mTOR and LAMP-2, n=22, 9, 18 (0, 4, 24 h). (c) Statistical analysis of co-localization of mTOR and LAMP-2 (r) in CD8⁺ T cells from T-Rheb^−/− and T-Tsc2^−/− at indicated time point after TCR activation, n=40 (T-Rheb^−/−), 39 (T-Tsc2^−/−). (d) Confocal images of in vitro activated CD8⁺ T cells showing localization of LAMP-2 and mTOR between CD8^hi and CD8^lo daughter T cells. Statistical analysis of co-localization of mTOR and LAMP-2 (r) between paired daughter cells are shown on the right, n=14. *P < 0.005; **P < 0.0005; ***P < 0.0001; NS, not significant (One-Way ANOVA (a, b), One-Way ANOVA (c) or Wilcoxon rank test (d)). Data are from one experiment representative of at least 2 independent experiments (mean in a–c). Scale bars, 10μm (a, b, d).

Figure 5. TCR-induced mTOR localization to lysosome is dependent on differential amino acid transporter expression

(a) qRT-PCR analysis of Slc7a5 mRNA expression in sorted CD8^hi and CD8^lo T cells activated in vitro, n=8. (b) As in Fig. 1a, histogram overlay of CD98 expression in CD8^hi and CD8^lo T cells. MFI, upper left corner. (c) Confocal images of CD8 and CD98 expression between daughter cells activated in vitro. Statistical analysis of CD98 MFI between CD8^hi and CD8^lo daughter T cells, n=13. (d) Confocal images of mTOR and LAMP-2 localization between daughter CD8⁺ T cells from in vitro activation. Statistical analysis of co-localization of mTOR and LAMP-2 (r) between CD98^hi and CD98^lo CD8⁺ T cells, n=19. (e) Flow cytometry of p-S6 and CD98 expression in CD8⁺ T cells activated in vitro ± amino acid transporter inhibitor (BCH) and unstimulated (US). MFI, upper right corner. (f) Statistical analysis of mTOR and LAMP-2 co-localization in CD8⁺ T cells under US, no treatment, and BCH (top) or rapamycin (bottom) treatment by confocal microscopy, n=10 (BCH), 15 (rapamycin). (g) Confocal images of daughter CD8⁺ T cells showing localization of LAMP-2 and mTOR among no treatment, BCH, and rapamycin treatment. Statistical analysis of mTOR and LAMP-2 co-localization (r) is shown on the right, n=20. *P < 0.05; **P < 0.005; ***P < 0.0005; ****P < 0.0001; NS, not significant (Mann-Whitney t test (a, b), Wilcoxon rank test (c, d, g) or One-Way ANOVA (f)). Data are from compilation of 3 independent experiments (a), one experiment representative of at least 3 independent experiments (b–g) (mean ± s.e.m. in a, mean in f). Scale bars, 10μm (c, d, g).

Figure 5. TCR-induced mTOR localization to lysosome is dependent on differential amino acid transporter expression

(a) qRT-PCR analysis of Slc7a5 mRNA expression in sorted CD8^hi and CD8^lo T cells activated in vitro, n=8. (b) As in Fig. 1a, histogram overlay of CD98 expression in CD8^hi and CD8^lo T cells. MFI, upper left corner. (c) Confocal images of CD8 and CD98 expression between daughter cells activated in vitro. Statistical analysis of CD98 MFI between CD8^hi and CD8^lo daughter T cells, n=13. (d) Confocal images of mTOR and LAMP-2 localization between daughter CD8⁺ T cells from in vitro activation. Statistical analysis of co-localization of mTOR and LAMP-2 (r) between CD98^hi and CD98^lo CD8⁺ T cells, n=19. (e) Flow cytometry of p-S6 and CD98 expression in CD8⁺ T cells activated in vitro ± amino acid transporter inhibitor (BCH) and unstimulated (US). MFI, upper right corner. (f) Statistical analysis of mTOR and LAMP-2 co-localization in CD8⁺ T cells under US, no treatment, and BCH (top) or rapamycin (bottom) treatment by confocal microscopy, n=10 (BCH), 15 (rapamycin). (g) Confocal images of daughter CD8⁺ T cells showing localization of LAMP-2 and mTOR among no treatment, BCH, and rapamycin treatment. Statistical analysis of mTOR and LAMP-2 co-localization (r) is shown on the right, n=20. *P < 0.05; **P < 0.005; ***P < 0.0005; ****P < 0.0001; NS, not significant (Mann-Whitney t test (a, b), Wilcoxon rank test (c, d, g) or One-Way ANOVA (f)). Data are from compilation of 3 independent experiments (a), one experiment representative of at least 3 independent experiments (b–g) (mean ± s.e.m. in a, mean in f). Scale bars, 10μm (c, d, g).