Distinct translational control in CD4+ T cell subsets

Abstract

Regulatory T cells expressing the transcription factor Foxp3 play indispensable roles for the induction and maintenance of immunological self-tolerance and immune homeostasis. Genome-wide mRNA expression studies have defined canonical signatures of T cell subsets. Changes in steady-state mRNA levels, however, often do not reflect those of corresponding proteins due to post-transcriptional mechanisms including mRNA translation. Here, we unveil a unique translational signature, contrasting CD4(+)Foxp3(+) regulatory T (T(Foxp3+)) and CD4(+)Foxp3(-) non-regulatory T (TFoxp3-) cells, which imprints subset-specific protein expression. We further show that translation of eukaryotic translation initiation factor 4E (eIF4E) is induced during T cell activation and, in turn, regulates translation of cell cycle related mRNAs and proliferation in both T(Foxp3)- and T(Foxp3+) cells. Unexpectedly, eIF4E also affects Foxp3 expression and thereby lineage identity. Thus, mRNA-specific translational control directs both common and distinct cellular processes in CD4(+) T cell subsets.

Figure 1. Genome-wide analysis of translationally regulated mRNAs in primary CD4⁺ T cell subsets.

(a) Cytosolic mRNA was extracted and probed directly with DNA microarrays or processed using the polysome preparation technique where mRNAs are sedimented on a sucrose gradient and separated based on the number of ribosomes they associate with. Fractions containing mRNAs that engage ≥3 ribosomes were pooled and probed with microarrays to quantify mRNA levels. (b) Polysome UV-tracings from ex vivo and in vitro activated T_Foxp3+ and T_Foxp3− cells. Shown is the UV absorbance (254 nm) as a function of sedimentation. The large peak corresponds to the 80S ribosome peak and was used to align the polysome profiles so that fractions containing ≥3 ribosomes could be pooled from each sample. The part of the polysome profile that was pooled and used as the polysome-associated mRNA sample is indicated. (c) Assessment of data set quality. Shown is a dendrogram from a hierarchical clustering of all included samples (using Pearson correlations). Samples that are more similar cluster together. Cyto – cytosolic mRNA; poly – polysome-associated mRNA.

Figure 2. A translational signature that discriminates T_Foxp3+ and T_Foxp3− cells.

(a–b) Polysome-associated mRNA levels differ from cytosolic mRNA levels in primary CD4⁺ T cell subsets ex vivo and post-activation. Shown are density scatter plots of polysome-associated vs. cytosolic mRNA data (a blue scale from light to dark represents increasing local density of data points; outliers are indicated as dots) for T_Foxp3− cells (a) and T_Foxp3+ cells (b) at the ex vivo and the activated condition. The solid and dotted lines indicate a >3-fold and >2-fold difference, respectively, in the density scatter plot. The number of mRNAs that show a >3-fold difference in each direction is indicated. (c) Substantial differences in levels of polysome-associated mRNA between T_Foxp3+ and T_Foxp3− cells. Density scatter plots (as in a–b) compare polysome-associated mRNA data between T_Foxp3+ and T_Foxp3− cells in both the ex vivo and in vitro activated conditions. A few genes known to be differentially expressed between T_Foxp3+ and T_Foxp3− cells are indicated (Foxp3, Ctla4, Il2ra [CD25] and Tnfrsf18 [GITR]). As expected the differential expression of Il2ra is lost upon activation. (d) Differential translation in T_Foxp3+ vs. T_Foxp3− cells as identified with anota-RVM ex vivo and post in vitro activation. Significances (i.e. the −log10 p-value from the anota analysis used to identify differential translation) are compared to log2 translational fold changes (after correction for cytosolic mRNA levels).

Figure 3. Distinct modular translational control between activated CD4⁺ T cell subsets.

Graphical representation of the enrichment analysis within subsets of mRNAs identified as differentially expressed (up in T_Foxp3+ cells or down in T_Foxp3+ cells) in data from cytosolic mRNA, polysome-associated mRNA and as differentially translated by anota (after correction for cytosolic mRNA levels). The subsets are shown as columns and the rows represent cellular functions that were enriched. The colour scale represents −log10 p-values (adjusted for multiple testing) for the enrichment. All p-values that were <10e-7 were set to 10e-7.

Figure 4. Translationally regulated mRNAs encode proteins are involved in ubiquitination, chromatin modification, or cell cycle pathways.

Translational activity (from anota after correction for cytosolic mRNA levels) in T_Foxp3+ and T_Foxp3− cells ex vivo and post in vitro activation for individual mRNAs belonging to ubiquitination (a), chromatin modification (b) or cell cycle (c) pathways is shown. The colour scale represents translational activity in log2 scale.

Figure 5. Differential levels of eIF4E between T_Foxp3+ and T_Foxp3− cells partly explain their translational signature and correlate with CD4⁺ T cell subset proliferation.

(a) eIF4E is translationally more active in activated T_Foxp3− cells as compared to T_Foxp3+ cells. Shown is the cytosolic mRNA level (x-axis) vs. the polysome-associated mRNA level (y-axis) for each condition; T_Foxp3+ N (blue) and T_Foxp3− N (red) – ex vivo cells; T_Foxp3+ 36 h (green) and T_Foxp3− 36 h (black) – in vitro activated cells. The lines indicate the regressions used by anota to correct the polysome-associated mRNA level for the cytosolic mRNA level. (b) Activated T_Foxp3− cells express higher protein levels of eIF4E, cyclin-E1, cyclin-D3, and Anapc4 as compared to activated T_Foxp3+ cells. Shown are western blots from T_Foxp3+ and T_Foxp3− cells activated for 36 hours. Densitometry was used to quantify protein levels and obtained levels were normalized to β-actin (the normalized values were related to T_Foxp3− 36 h which was set to 1 and are indicated above each lane). (c) Identification of an eIF4E responsive module in the activated T cell translational signature. Fold changes from differentially translated mRNAs from the activated T cell translational signature that also showed a fold change difference for translation in lungs from 4E-BPdko mice are plotted. The number of mRNAs in each quadrant is shown. (d) High IL-2 concentration induces proliferation in T_Foxp3+ cells. Cell numbers were counted when plated and after 72 h of culture with low (100 U/ml) or high (1000 U/ml) IL-2 concentrations. The fold increase in cell number was calculated and associated means and standard deviations (n = 3) are shown. Welch's two sample t-test was used to compare T_Foxp3+ cells cultured under different IL-2 concentrations. (e) High IL-2 concentration induces eIF4E expression in T_Foxp3+ cells. Shown are western blots of total protein extracts probed with antibodies for eIF4E, cyclin-E1, cyclin-D3, and Anapc4 in T_Foxp3+ and T_Foxp3− cells activated as described in (d). Densitometry was used to quantify protein levels and obtained levels were normalized to β-actin (the normalized values were related to T_Foxp3− 72 h IL-2 100 U/ml which was set to 1 and are indicated above each lane; lanes between lanes 3 and 4 in (e) were spliced out but all shown lanes are from the same gel).

Figure 6. eIF4E controls proliferation in T cell subsets.

(a) Inhibition of eIF4E activity suppresses T_Foxp3− cell proliferation. eFluor 670-labeled T_Foxp3− cells were IL-2/TCR-activated for 72 h in the presence of increasing concentrations of the eIF4E inhibitor 4ei-1 (K_d = 0.80 µM). Proliferation was determined under each condition by eFluor 670 dilution assessed by flow cytometry (upper panel). The effect on proliferation was also assessed by comparing cell counts after 72 h under each condition (lower panel; the control was set to 100%). (b) Inhibition of eIF4E activity abrogates IL-2-mediated reversal of anergy in T_Foxp3+ cells. IL-2/TCR-activated eFluor 670 labelled T_Foxp3+ cells were cultured in the presence of 4ei-1, and proliferation was determined as described in (a). (c–d) IL-2/TCR-activated eFluor 670 labelled T_Foxp3− cells (c) or T_Foxp3+ cells (d) were cultured in the presence of 4ei-1 or 4ei-4. Proliferation was determined under each condition as described in (a). (a–d) Representative histograms from 4 independent experiments are shown (upper panels; the percentages of proliferating cells are indicated). Means and standard deviations of cell counts from 4 independent experiments are shown (lower panel). (e) Induction of T_Foxp3+ cell proliferation occurs independently of signalling through 4E-BPs. 4E-BPdko T_Foxp3+ and T_Foxp3− cells were plated and counted as described in (Figure 5d), and the fold increase in cell number was calculated and associated means and standard deviations (n = 2) are shown. Welch's two sample t-test was used to compare 4E-BPdko T_Foxp3+ cells cultured under different IL-2 concentrations. Also shown is a western blot of total protein extracts probed with antibodies for eIF4E in 4E-BPdko T_Foxp3+ and T_Foxp3− cells. Densitometry was used to quantify protein levels and obtained levels were normalized to β-actin (the normalized values were related to T_Foxp3− 72 h IL-2 100 U/ml which was set to 1 and are indicated above each lane). (f) Ki-67 and eIF4E co-expression in total CD4⁺ T cells isolated directly ex vivo from lymph nodes (left panel). Quantification of eIF4E expression is shown as Δ (eIF4E vs. isotype control) mean fluorescent intensity (MFI). Filled histograms represent staining with an isotype control. Quantification of eIF4E expression (ΔMFI) in Ki-67^+/− T_Foxp3− and T_Foxp3+ cells isolated directly ex vivo (right panel, mean and standard deviation is indicated, n = 3). (g–h) eFluor 670-labeled T_Foxp3− or T_Foxp3+ cells adoptively transferred into separate TCR β−/− mice were isolated from mesenteric (mes) and peripheral (per) lymph nodes (LN) followed by measurement of eFluor 670 and eIF4E expression four days post transfer. (g) Representative dot plots (n = 3) of T_Foxp3− and T_Foxp3+ cell proliferation relative to eIF4E expression in mesLN. Staining with an isotype control are shown as contour plots. (h) Quantification of eIF4E expression (ΔMFI) in cells that have (eFluor 670 low) or have not (eFluor 670 high) undergone cell division (means and standard deviations are indicated after per experiment normalization to T_Foxp3+ cells, n = 4–6). P-value (Welch two sample t-test) is indicated.

Figure 7. Inhibition of eIF4E activity results in spontaneous induction of Foxp3 expression in activated T_Foxp3− cells.

T_Foxp3− cells were IL-2/TCR-activated for 72 h in the presence of increasing concentrations of 4ei-1 or the control pro-drug 4ei-4 in undifferentiating conditions, and Foxp3 expression (i.e. GFP) was assessed by flow cytometry. (a) Representative density plots from experiments using T_Foxp3− cells cultured in the presence of 4ei-1 from 4 independent experiments are shown. (b) Percentage Foxp3⁺ cells following treatment with 4ei-1 or 4ei-4 (shown are means and standard deviations, n = 4).