Visual Abstract

The activation of T cells is accompanied by intensive posttranscriptional remodeling of their proteome. We observed that protein expression of enzymes that modify wobble uridine in specific tRNAs, namely elongator subunit 3 (Elp3) and cytosolic thiouridylase (Ctu)2, increased in the course of T cell activation. To investigate the role of these tRNA epitranscriptomic modifiers in T cell biology, we generated mice deficient for Elp3 in T cells. We show that deletion of Elp3 has discrete effects on T cells. In vitro, Elp3-deficient naive CD4+ T cells polarize normally but are delayed in entering the first cell cycle following activation. In vivo, different models of immunization revealed that Elp3-deficient T cells display reduced expansion, resulting in functional impairment of T follicular helper (TFH) responses, but not of other CD4+ effector T cell responses. Transcriptomic analyses identified a progressive overactivation of the stress-responsive transcription factor Atf4 in Elp3-deficient T cells. Overexpression of Atf4 in wild-type T cells phenocopies the effect of Elp3 loss on T cell cycle entry and TFH cell responses. Reciprocally, partial silencing of Atf4 or deletion of its downstream effector transcription factor Chop rescues TFH responses of Elp3-deficient T cells. Together, our results reveal that specific epitranscriptomic tRNA modifications contribute to T cell cycle entry and promote optimal TFH responses.

Engagement of the TCR and adequate costimulation induces naive T cells to exit quiescence and prepares their entry into the cell cycle and, ultimately, their differentiation into effector cells. The signaling events and transcriptional program of this transition have been extensively studied (1, 2). Likewise, the transcriptional regulation of T cell polarization is being characterized with exquisite detail (3). Although transcriptional control of gene expression is paramount to T cell activation and functional polarization, comparatively little is known of the impact of translation on T cell biology. Recent evidence suggests that T cells strongly engage translational control following activation. Indeed, within hours of TCR and costimulation engagement of T cells, almost 2000 proteins become differentially expressed compared with quiescent T cells (4). Strikingly, early changes in protein abundance in newly activated T cells correlate poorly with changes in mRNA levels, indicating that posttranscriptional regulation of protein abundance plays important roles in T cell activation (4).

An emerging level of translational control lies in modifications of the pool of cellular tRNA, which may convey epigenetic information as part of the tRNA epitranscriptome. Indeed, tRNAs are among the most heavily modified molecules in eukaryotic cells (5, 6). Base modifications in tRNAs are catalyzed by a wide array of enzymes that each target one or, more commonly, several positions in target tRNAs, sometimes acting one after another in a serial fashion (5, 6). Typically, modifications outside the anticodon region impact on the maturation, localization, stability, or ribosome binding of target tRNAs (5, 6). Modifications also occur in the tRNA anticodon on the base in position 34, which is most often called the wobble base. The wobble base is a hotspot of tRNA modifications (5, 6), mostly understood as playing a role in wobble pairing by improving or modifying the target tRNA decoding activity (710).

An interesting group of tRNA modifying enzymes consists in the enzymes that modify the wobble uridine (U34) in 11 cytosolic tRNA species in mammalian cells, namely the elongator and cytosolic thiouridylase (Ctu)1/2 complexes. Unlike other enzymes that modify the tRNA anticodon, which also target other positions in the tRNA, elongator and Ctu1/2 activities are restricted to the U34 base in cytosolic tRNAs (5). Elongator is composed of six subunits (Elp1–6), of which Elp1 is the scaffolding subunit and Elp3 the catalytic subunit, and catalyzes a first enzymatic reaction, leading to the formation of 5-methoxycarbonylmethyl or 5-carbamoylmethyl side chains on U34. The unrelated dimeric Ctu1/2 complex catalyzes a further 2-thioylation of U34 in specific tRNAs (ArgUCU, LysUUU, GlnUUG, and GluUUC) (5).

Inactivation of elongator and/or Ctu1/2 leads to a puzzling variety of cellular phenotypes in normal and cancer cells. In mice, full inactivation of elongator leads to embryonic death (11), as one may expect from the impairment of a vital housekeeping function. Arguing against this simple view though, inactivation of elongator in restricted mammalian cell lineages generates specific and most often discrete outcomes. In humans, mutations causing loss of activity of elongator in restricted cell types are responsible for familial dysautonomia, a neurologic disease (12). Elongator-mutant mouse models suggest that this condition is at least partly caused by death of peripheral sympathetic neurons and progenitors thereof (13). Likewise, deletion of Elp3 in murine cerebral cortex progenitors results in microcephaly by impairing the generation of intermediate progenitors of cortical neurons (14). In contrast, genetic invalidation of elongator in intestinal stem cells does not impact on intestinal epithelial cell development in the steady state, although it impairs tissue regeneration following irradiation and delays Wnt-driven tumorigenesis (15). Similarly, mouse breast development is unperturbed in the absence of elongator, but elongator-deficient breast tumors display delayed onset and decreased metastatic activity (16). Altogether, these data suggest that U34 modifications are usually dispensable in steady-state–differentiated cells but become important in specific transition states such as those imparted by cellular differentiation or adaptation to stress.

From a molecular standpoint, mammalian phenotypes linked with loss of U34-modifying enzymes have been mainly attributed to suboptimal peptide elongation rates, leading either to stress signaling through the unfolded protein response (14) or to loss-of-function phenotypes because of impaired expression of specific proteins whose mRNA is enriched in codons read by U34-containing tRNAs, especially LysUUU, GlnUUG, and GluUUC (1517). Yet, observations in yeast indicate that U34 modifications may also be sensed as metabolic cues of amino acid or sulfur availability (1820). Consequently, yeast deficient for U34 modifications react as if experiencing amino acid starvation even in amino acid–replete growth conditions and activate the Gcn4 metabolic stress response pathway, a functional ortholog of the mammalian metabolic adaptor activating transcription factor 4 (Atf4) (2022).

In this study, we report that protein expression of Elp3 and Ctu1/2 increases following mouse T cell activation. Building on this observation, we in this study investigate the impact of the inactivation of the elongator complex on murine T cell responses. We show that Elp3 activity is required for timely entry of naive T cells into the first cell cycle and that deletion of Elp3 or Ctu1 impairs T follicular helper (TFH) cell responses in vivo through activation of the Atf4/Chop stress response pathway.

All experimental procedures and protocols were reviewed and approved by the Institutional Animal Care and Use Ethics Committee of University of Liege. Elp3fl/fl mice have been described previously [Ladang et al. (15)]. CD4-Cre mice [Tg(Cd4-cre)1Cwi/BfluJ], OT-II mice [B6.Cg-Tg(TcraTcrb)425Cbn/J], wild-type CD45.2 mice (C57BL/6J), CD45.1 mice (B6.SJL-Ptprca Pepcb/BoyJ), and Rosa26fl-STOP-fl-Atf4/+ mice [B6;129 × 1-Gt(ROSA)26Sortm2(ATF4)Myz/J] were obtained from The Jackson Laboratory. Elp3fl/fl Cd4-CreT/+ (Elp3TKO) mice were obtained by crossing Elp3fl/fl mice with CD4-CreT/+ mice, and Elp3fl/fl littermates were used as controls. To generate OVA-specific T cell with conditional inactivation of Elp3, we crossed Elp3TKO mice with OT-II mice, and OT-II Elp3fl/fl littermates were used as controls. To generate Atf4TTG mice, CD4-CreT/+ mice were crossed with Rosa26fl-STOP-fl-Atf4/+, and Rosa26fl-STOP-fl-Atf4/+ littermates were used as control. Balanced groups of male and female littermates were randomly assigned to experimental groups and were used at 8–12 wk of age unless otherwise indicated. All mice were bred and housed in the specific pathogen-free facility of University of Liege, except in experiments involving influenza infection, which were housed in individually ventilated cages.

Naive CD4 lymphocytes were purified from spleen and lymph nodes with Naive CD4+ T Cell Isolation Kit (Miltenyi Biotec). A total of 2 × 105 naive CD4 lymphocytes were activated in an anti-CD3e (Invitrogen)–coated plate in RPMI 1640 medium (Lonza) supplemented with 10% of FBS, 0.1 mM of nonessential amino acid, 50 UI/ml of penicillin G, and 5 μg/ml streptomycin, 1 mM sodium pyruvate, and 10 μg/ml 2-ME (all from Life Technologies, Thermo Fisher Scientific). The following cytokine/Ab mixtures were added to culture to induce T cell differentiation: Th0, anti-CD28 (2 μg/ml) and IL-2 (100 UI/ml); Th1, anti-CD28 (2 μg/ml), anti–IL-4 (10 μg/ml), IL-12p70 (10 ng/ml), and IL-2 (100 UI/ml); Th2, anti-CD28 (2 μg/ml), anti–IFN-γ (10 μg/ml), IL-4 (10 ng/ml), and IL-2 (100 UI/ml); and Th17, anti-CD28 (2 μg/ml), anti–IFN-γ (10 μg/ml), anti–IL-4 (10 μg/ml), IL-6 (20 ng/ml), and TGF-β (3 ng/ml).

Mice were injected with OVA-CFA (see below). After 7 d, TFH were sorted from the draining lymph nodes as CD4+CD25CXCR5+ICOS+Viability Dye cells. Naive B cells were sorted from naive mice as CD19+GL7Viability Dye cells. Fifty thousand naive B cells were cocultured with up to 60,000 TFH cells in complete RPMI 1640 containing anti-CD3e (2 μg/ml) and Fab anti-IgM (5 μg/ml) for 5 d prior to analysis.

Blood was collected at sacrifice by retroorbital bleeding using sodium heparinized microhematocrit capillary tubes (Hirschmann). For serial blood analysis, blood was collected from the tail vain. Automated peripheral blood counts were obtained using a Cell-Dyn 3700.

For influenza A virus (IVA) infection, 8-wk-old mice received an intranasal instillation of 50 μl saline containing 5 PFU of IVA strain A/PR8/34 (H1N1) (kindly provided by F. Trottein [Institut Pasteur]). Mice were weighed daily, and the lung and bronchial lymph node were analyzed at day 10 postinfection.

For OVA-CFA vaccination, mice received in both hind legs via s.c. injection 50 μg OVA (Sigma-Aldrich) emulsified 1:1 with CFA (Sigma-Aldrich). On day 8 postvaccination, the inguinal lymph nodes were extracted for flow cytometry analysis. For OVA-alum vaccination, 10 μg of OVA (Sigma-Aldrich) alone or adsorbed on 4 mg of Imject Alum (Pierce Biochemicals) was injected i.p.

For SRBC immunization, SRBC (Innovative Research) were washed twice with sterile PBS, and then 1 × 109 SRBC were injected per mouse i.v. The spleen was assessed at day 7 and day 14 postimmunization.

For house dust mite extract (HDM)–induced model of asthma, lightly isoflurane-anesthetized mice were sensitized by three daily intratracheal (i.t.) instillations of vehicle (PBS) or HDM (Dermatophagoides pteronyssinus, 50 μg in 50 μl; Greer Laboratories) on days 0, 1, and 2. Mice were then challenged weekly by an i.t. instillation of vehicle or HDM (20 μg in 50 μl) on days 7, 14, 21, and 28 and were sacrificed at day 35.

The bronchoalveolar lavages and the draining lymph node were evaluated. Briefly, we catheterized the trachea and washed the lungs with 1 ml ice-cold Mg- and Ca-free PBS containing 0.6 mM EDTA. We assessed cell density in bronchoalveolar lavage fluid using a hemocytometer, and the differential cell populations were identified by flow cytometry (T cells as CD3+ cells, neutrophils as CD3 CD11b+ Ly-6G+ cells, eosinophils as CD3 CD11b+ Siglec F+ cells, and macrophages as autofluorescent F4/80+ Siglec F+ cells). The bronchial lymph node cells were cultured in vitro for 4 h in complete RPMI 1640 with PMA (50 ng/ml; Sigma-Aldrich), ionomycin (1 μg/ml; Sigma-Aldrich), and monensin (BD GolgiStop; BD Biosciences), then processed for intracellular staining as described below.

C57BL/6J (CD45.1) mice were lethally irradiated with two doses of 6 Gy 3 h apart using a Gammacell 40 Exactor (MDS Nordion). Two hours after the second irradiation, these mice received an i.v. injection with 5 million bone marrow cells consisting of a 1:1 mix of bone marrow cells obtained from C57BL/6J (CD45.1) and Elp3TKO (CD45.2) mice. Mice were treated with antibiotics (Baytril, 0.05 mg.ml–1 in drinking water; Bayer) from the day before irradiation. After 10 wk, the chimera mice were treated with OVA-CFA as described above.

Deletion of the exon 2 of Elp3 gene was measured by real-time PCR analysis. Genomic DNA was extracted from total bone marrow cells using Quick-DNA Universal Kit (Zymo Research) following the manufacturer’s recommendations. Quantitative real-time PCR was performed in duplicates with 20 ng genomic DNA using iQ SYBR Green SuperMix (Bio-Rad Laboratories) and iQ5 Multicolor Real-Time PCR Detection System (Bio-Rad Laboratories) under the following conditions: initial denaturation at 95°C for 15 min, followed by 40 cycles of 95°C 15 s and 60°C 1 min. Primer sequences were as followed: 5′-GACAGCAAGACTGCCCTTCT-3′ and 5′-CAGCTGGTTGAAGCTCATGA-3′ for Elp3, 5′-AGCCCAGTGTTACCACCAAG-3′ and 5′-ACCCAAGAACAAGCACAAGG-3′ for Ubc, and 5′-AGCAGGTAGACCACCTCAAG-3′ and 5′-CACCAGCTTCTCGTATTTCTC-3′ for Maf genes. Expression levels of Elp3 exon 2 were normalized relative to control genes (Ubc and MAF) using qBase+ software (Biogazelle).

For mRNA expression analyses, cells were put directly into TRIzol (Life Technologies) and extracted with chloroform–isopropanol. cDNA was obtained by reverse transcribing total RNA with a RevertAid H Minus First Strand cDNA Synthesis Kit (Thermo Fisher Scientific).

The quantitative PCR was carried out with iTaq UniverSYBR Green SMX (Bio-Rad Laboratories) on a LightCycler 480 System (Roche Diagnostics).

Primers for real-time RT-PCR were as follows: 5′-CGGCAAGGAGGATGCCTTTT-3′and 5′-GCTCATCTGGCATGGTTTCCA-3′ for Atf4, 5′-GAGGAAGAATCAAAAACCTTCACTACTCT-3′ and 5′-GACTGGAATCTGGAGAGCGAGG-3′ for Ddit3, 5′-CCTCACCCGGAACAAGGATG-3′ and 5′-TGTAGGACCTGGCATGTCCC-3′ for Cars, 5′-CGGCATCAGTGTGCTCGAAATG-3′ and 5′-GTTGGGAACAGCTAGCAATTCCC-3′ for Psat1, 5′-CTCTGAGATCACAGGGGACCTCA-3′ and 5′-GGCTTCGCCTTTAGCTTAGGAATC-3′ for Elp3, 5′-GCCTTGTGGGAGCTACACATATTGTG-3′ and 5′-CTCCCAGCATCACCACGCA-3′ for Ctu1, and 5′-CAACTACATGGTCTACATGTTC-3′ and 5′-CACCAGTAGACTCCACGAC-3′ for Gapdh.

Cells were washed twice with PBS then lysed in buffer (1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 25 mM Tris-HCl [pH 7.6], 150 mM NaCl, cOmplete Protease Inhibitor Cocktail [Roche Diagnostics], and PhosSTOP phosphatase inhibitors [Roche Diagnostics]), incubated on ice for 30 min, and treated by sonication (Bioruptor, Diagenode; five times at level 5 and duty cycle 50%). Proteins were extracted by centrifugation (16,100 × g) for 30 min at 4°C and quantified using bicinchoninic acid method (Micro BCA Protein Assay Kit; Thermo Fisher Scientific). The same amount of protein extracts was added to Laemmli buffer v/v, boiled, and run on an SDS-PAGE gel; after electrophoresis, gels were electrotransferred (wet transfer) onto a polyvinylidene difluoride membrane for Western blotting.

Nonspecific binding was blocked in 20 mM Tris (pH 7.5), 500 mM NaCl, 0.2% Tween 20 (Acros Organics) (TBST), and 10% dry milk during 1 h.

Membranes were then incubated with primary Abs (1:1000) in TBST with 5% BSA overnight at 4°C, rinsed with TBST, and further incubated with HRP-conjugated secondary Ab (conjugated anti-mouse and anti-rabbit, 1:2000; Cell Signaling Technologies) in TBST with 5% dry milk for 2 h at room temperature. Result was visualized using an ECL system (Bio-Rad Laboratories) or SuperSignal West Femto chemiluminescence kit (Thermo Fisher Scientific).

For OVA-specific IgG titer determination, flat-bottom, 96-well plates were coated with 50 μg OVA overnight. After blocking, diluted mouse serum and secondary Ab HRP-conjugated goat anti-mouse IgG (Dako) was successively incubated for 90 min at room temperature. Signal was revealed by Stabilized Chromogen TMB (Life Technologies).

For immunofluorescence assay, the spleen from OVA-CFA–challenged Elp3TKO or control mice were frozen, and 5-μm cryosection slices were prepared with a Crystat. Slices were fixed with 4% paraformaldehyde and blocked with 1% BSA and 2% FBS before successive staining with primary, secondary Ab and DAPI. After washing with PBS containing 0.5% Tween, slices were mounted with VectaMount Permanent Mounting Medium (Vector Laboratories). The confocal images were acquired with a Leica SP5 confocal microscope (Leica).

Flow cytometry analysis was carried on with a BD LSRFortessa or BD FACSAria IIIu (BD Biosciences). The data were analyzed with FlowJo software (Tree Star). Surface staining was performed in PBS containing 0.5% BSA and 0.1% sodium azide. Intracellular cytokine staining was performed with a Cytofix/Cytoperm Fixation/Permeabilization Kit (554714; BD Biosciences). Intranuclear staining was performed with the Foxp3/Transcription Factor Staining Buffer Set (00-5523; eBioscience, Thermo Fisher Scientific). For ex vivo EdU incorporation assay, a pulse of 10 μM EdU was added to lymphocyte culture 1 h before analysis. The click reaction was realized according to manufacturer recommendation (Thermo Fisher Scientific).

For cell division tracking, CellTrace Violet (Invitrogen, Thermo Fisher Scientific) was used as described previously (23). The division index was calculated in FlowJo by dividing the total number of cell divisions by the number of cells before proliferation.

The Abs and reagents used for flow cytometry staining are listed in Supplemental Table I.

A total of 9 × 104 cells was directly sorted in TRIzol (Life Technologies). Total RNA was purified using Direct-zol kit (Zymo Research) following manufacturer’s recommendations. Eluted RNAs were incubated with DNase for 10 min at room temperature in a total volume of 200 μl using the RNase-Free DNase Set from QIAGEN, followed by purification on RNeasy Micro columns (QIAGEN). RNAs were quantitatively and qualitatively verified using Agilent RNA 6000 Pico Kit run on an Agilent 2100 Bioanalyzer. RNA integrity numbers were above 9.2 for in vitro activated T cells and above 7.4 for TFH.

cDNA was synthesized with 10 ng of total RNA using the SMART-Seq v4 Ultra Low Input RNA Kit for Sequencing (Clontech Laboratories) and amplified by seven cycles of locked-DNA PCR. The sequencing library was generated using the Nextera XT DNA Library Preparation (Illumina) according to manufacturer’s instructions. Paired-end RNA sequencing was performed on a HiSeq 2000 (Illumina).

To overexpress Atf4, the open reading frame of Atf4 was cloned from lymphocyte cDNA pool with Expand High-Fidelity PCR System (Roche Diagnostics) and inserted into XhoI/EcoRI-opened pMIGR1 (MSCV-IRES-GFP) plasmid. For the short hairpin RNA–expressing plasmid, oligonucleotides were synthesized, annealed, and ligated into AgeI/EcoRI-opened pMKO.1 (U6-SV40-GFP) plasmid. pMIGR1 and pMKO.1 plasmids were a kind gift of Prof. L. Ye (Army Medical University, Chongqing, China).

Primers for cloning mouse Atf4 were as follows: 5′-ATCGCTCGAGCCACCATGACCGAGATGAGCTTCCTGAAC-3′ and 5′-ATCGGAATTCTTACGGAACTCTCTTCTTCCCCC-3′.

Primers for short hairpin RNA expression were as follows: Sh-Elp3, 5′-CCGGCCGTGCTAGATATGACCCTTTCTCGAGAAAGGGTCATATCTAGCACGGTTTTTG-3′ and 5′-AATTCAAAAACCGTGCTAGATATGACCCTTTCTCGAGAAAGGGTCATATCTAGCACGG-3′; and Sh-Ctu1, 5′-CCGGACCTGAGCTCTGCACTCTACACTCGAGTGTAGAGTGCAGAGCTCAGGTTTTTTG-3′ and 5′-AATTCAAAAAACCTGAGCTCTGCACTCTACACTCGAGTGTAGAGTGCAGAGCTCAGGT-3′.

Retroviruses were produced by cotransfecting 293T cells with the expression plasmid and pCL plasmid. The supernatant after a 2-d culture was used to transduce T cells.

For T cell transduction, OVA-specific OT-II CD4+ T cells were first purified from spleens of 6-wk-old naive OT-II or OT-II–Elp3TKO mice with Mouse CD4+ T Cell Isolation Kit (Miltenyi Biotec) and activated by culturing the cells in anti-CD3 (BioLegend)–precoated plates and soluble anti-CD28 (BioLegend). Then, activated T cells were spin transduced (500 × g at 37°C) in a centrifuge by incubating with retrovirus supernatant containing 8 μg/ml polybrene.

One hundred thousand transduced CD4+ T cells were adoptively transferred; receiver CD45.1+ mice and receiver mice were immunized 24 h later with OVA-CFA as described above. The overexpression or silencing of target protein were validated by RT-qPCR or Western blot in sorted OT-II (CD45.2+) cells at indicated time points.

Sequenced reads were aligned to the mouse genome (University of California, Santa Cruz mm10) and aligned and mapped with RNA-seq Alignment (v1.1.0) using STAR (STAR_2.5.0b) on BaseSpace (https://basespace.illumina.com). Uniquely mapped reads were used to calculate gene expression.

Differential gene expression analysis was calculated using DESeq2 in R. Ranked gene set enrichment analysis (GSEAR) analyses were performed on preranked list of genes ordered according to their average log2 fold change. Online GSEAR v6.0.10 (https://genepattern.broadinstitute.org/gp/pages/index.jsf) was used with the h.all.v6.1.symbols (Hallmarks) gene sets and default parameters, except for a minimal gene set size of 20.

All statistical analyses were performed using R (24). All experiments followed a randomized design. Sample sizes were determined by power analysis. Respect of tests assumptions and model fit were evaluated using diagnostic plots. Raw data were transformed when needed and back transformed for graphical presentation. Histograms with error bars in all graphs represent mean ± SD. Details on which analysis was performed in each experiment are provided in the figure legends. A p value < 0.05 was considered significant. For clarity of presentation, only results of intergroup comparisons of interest are displayed in figures.

The source data that support the findings of this study are available from the corresponding author upon request. Transcriptomic data of naive T cell activation and TFH cells of immunized mice have been deposited in the ArrayExpress database (https://www.ebi.ac.uk/arrayexpress) with accession codes E-MTAB-9057 and E-MTAB-9059, respectively.

To gain insight into the potential modulation of U34-modifying enzymes following T cell activation, we monitored the protein expression of Elp3 and Ctu2 in naive (CD4+CD44CD62L+) murine T cells activated by anti-CD3 and anti-CD28 Abs. We observed an increase in Elp3 and Ctu2 abundance starting ∼18 h postactivation (Fig. 1A). Of note, Elp3 and Ctu2 protein abundance appeared essentially regulated at the posttranscriptional level. Indeed, mRNA expression levels of the genes encoding Elp3 and Ctu1/2 subunits did not increase following T cell activation (Supplemental Fig. 1A). These results are in line with a recent report supporting the notion that protein expression changes in activated T cells heavily depend on posttranscriptional regulation (4). Data mining of this previous comparative proteomic dataset confirmed the upregulated expression of Elp3 and Ctu1/2 in activated T cells (Supplemental Fig. 1B). The above data hence suggested that U34-modifying enzymes could play a role in T cell activation.

FIGURE 1.

Inactivation of Elp3 in CD4 T cells delays cell cycle entry. (A) Experimental outline of ex vivo activation of naive (CD44 CD62L+) CD4+ T cells with immobilized anti-CD3 and soluble anti-CD28 (upper panel) and Western blot analysis of Elp3 and Ctu2 protein expression following CD4+ T cell activation (left, representative blots, actin as a loading control; right, quantification, ANOVA followed by Tukey honest significance test [HSD] tests). (B) Flow cytometric profile (left) and percentages of cells in the G0/G1, S, and G2/M phases of the cell cycle (right) based on EdU integration and 7-aminoactinomycin D signal intensity (Mann–Whitney U tests of percentages of cells in S phase at each time point, n = 9–16 per group, data pooled from four experiments). (C) Flow cytometric CellTrace Violet (CTV) dilution profile (left) and number of divisions (middle, Student t tests at each time point, n = 21–34 per group) and live cell number (upper right, ANOVA followed by Tukey HSD tests) and division index (lower right, Student t test) following CD4+ T cell activation (n = 21–34 per group, data pooled from three experiments). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 1.10−4. n.s., not significant.

FIGURE 1.

Inactivation of Elp3 in CD4 T cells delays cell cycle entry. (A) Experimental outline of ex vivo activation of naive (CD44 CD62L+) CD4+ T cells with immobilized anti-CD3 and soluble anti-CD28 (upper panel) and Western blot analysis of Elp3 and Ctu2 protein expression following CD4+ T cell activation (left, representative blots, actin as a loading control; right, quantification, ANOVA followed by Tukey honest significance test [HSD] tests). (B) Flow cytometric profile (left) and percentages of cells in the G0/G1, S, and G2/M phases of the cell cycle (right) based on EdU integration and 7-aminoactinomycin D signal intensity (Mann–Whitney U tests of percentages of cells in S phase at each time point, n = 9–16 per group, data pooled from four experiments). (C) Flow cytometric CellTrace Violet (CTV) dilution profile (left) and number of divisions (middle, Student t tests at each time point, n = 21–34 per group) and live cell number (upper right, ANOVA followed by Tukey HSD tests) and division index (lower right, Student t test) following CD4+ T cell activation (n = 21–34 per group, data pooled from three experiments). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 1.10−4. n.s., not significant.

Close modal

To test this hypothesis, we generated transgenic mice with a conditional deficiency in Elp3, the catalytic subunit of elongator, in mature T cells by crossing Elp3fl/fl mice (15) with mice expressing the Cre recombinase under the dependence of the Cd4 promoter [Cd4-CreT/+ mice (25)] (Supplemental Fig. 1C). In the Elp3TKO mice, exon 2 of the Elp3 transcript was efficiently excised in CD4+ lymphocytes, leading to complete loss of Elp3 protein expression (Supplemental Fig. 1C, 1D). To assess the functional consequences of impairing Elp3 activity on the responses of naive CD4+ T cells to TCR engagement, we first compared the kinetics of entry into the cell cycle of Elp3TKO T cells with that of control Elp3fl/fl T cells using EdU incorporation assays. Elp3TKO T cells were delayed in entering the first S phase of the cell cycle (Fig. 1B). Percentages of Elp3TKO T cells in S phase yet were normalized after ∼36 h of activation (Fig. 1B). In line with these observations, Elp3TKO T cells also displayed a delay of approximately one cell division over 3 d after activation in dye dilution assays (Fig. 1C), resulting in a lower cell division index and total live cell number at the end of the culture (Fig. 1C). The ability of naive Elp3TKO T cells to polarize into Th1, Th2, or Th17 cells, however, was not affected (Supplemental Fig. 1E), even though a similar decrease in cell division index affected all polarization conditions (data not shown). These results indicate that loss of Elp3 activity delays entry into the first S phase following TCR activation in T cells.

We next sought to determine how inactivation of Elp3 would impact T cell biology in vivo. Elp3TKO mice had normal blood composition (Supplemental Fig. 2A). Abundance of thymocyte subsets was comparable to that of control mice (Fig. 2A). Moreover, Elp3TKO mice had normal abundance of CD3+CD4+ and CD3+CD8+ lymphocytes in their blood, spleen, and lymph nodes (Supplemental Fig. 2B). We also observed no significant difference in the relative abundance of naive versus memory T cell subsets in the steady-state lymph nodes and spleen of control and Elp3TKO mice (Fig. 2A). These results support the notion that Elp3 activity is not essential for T cell homeostasis in the steady state.

FIGURE 2.

Elp3 is dispensable for classical Th function. (A) Percentages of thymocytes into their four main developmental stages (left), percentages of naive (CD44 CD62L+), central memory T cells (TCM; CD44+ CD62L+) and effector memory T cells (TEM; CD44+ CD62L) in the spleen (central), and peripheral lymph nodes (right) of naive Elp3TKO and control mice (Student t tests comparisons for each indicated population, n = 9 per group, data pooled from three experiments). DN, CD4 CD8 double negative; DP, CD4 CD8 double positive; SP, single positive. (B) Experimental outline of the infection Elp3TKO and control mice by intranasal (i.n.) instillation of IAV (upper panel) and evolution of mouse body weight up to 10 d postinfection (two-way ANOVA followed by Tukey honest significance test [HSD] tests at day 10 postinfection [p.i.], n = 4–5 per group per experiment. Representative of two independent experiments). (C) Number of activated CD4+ and CD8+ T cells in the pulmonary draining lymph node (dLN) of mice in (B) 10 d postinfection (two-way ANOVA followed by Tukey HSD tests). (D) Experimental outline of the sensitization and challenge of Elp3TKO and control mice by i.t. instillation of (upper panel) and flow cytometric comparison of the immune cell populations in the bronchoalveolar lavage (BAL; lower left panel, two-way ANOVA followed by Tukey HSD tests on each cell type, n = 14 per group, data pooled from two experiments) and intracellular flow cytometric assessment of the number of IL-4– or IL-13–producing T cells in the bronchial lymph node (dLN, lower right panel, two-way ANOVA followed by Tukey HSD tests on each cytokine, n = 8 per group, data pooled from two experiments). ****p < 1.10−4. ns, not significant.

FIGURE 2.

Elp3 is dispensable for classical Th function. (A) Percentages of thymocytes into their four main developmental stages (left), percentages of naive (CD44 CD62L+), central memory T cells (TCM; CD44+ CD62L+) and effector memory T cells (TEM; CD44+ CD62L) in the spleen (central), and peripheral lymph nodes (right) of naive Elp3TKO and control mice (Student t tests comparisons for each indicated population, n = 9 per group, data pooled from three experiments). DN, CD4 CD8 double negative; DP, CD4 CD8 double positive; SP, single positive. (B) Experimental outline of the infection Elp3TKO and control mice by intranasal (i.n.) instillation of IAV (upper panel) and evolution of mouse body weight up to 10 d postinfection (two-way ANOVA followed by Tukey honest significance test [HSD] tests at day 10 postinfection [p.i.], n = 4–5 per group per experiment. Representative of two independent experiments). (C) Number of activated CD4+ and CD8+ T cells in the pulmonary draining lymph node (dLN) of mice in (B) 10 d postinfection (two-way ANOVA followed by Tukey HSD tests). (D) Experimental outline of the sensitization and challenge of Elp3TKO and control mice by i.t. instillation of (upper panel) and flow cytometric comparison of the immune cell populations in the bronchoalveolar lavage (BAL; lower left panel, two-way ANOVA followed by Tukey HSD tests on each cell type, n = 14 per group, data pooled from two experiments) and intracellular flow cytometric assessment of the number of IL-4– or IL-13–producing T cells in the bronchial lymph node (dLN, lower right panel, two-way ANOVA followed by Tukey HSD tests on each cytokine, n = 8 per group, data pooled from two experiments). ****p < 1.10−4. ns, not significant.

Close modal

To assess the potential role of Elp3 in effector T cell responses, we exposed Elp3TKO mice to prototypical type 1 or type 2 immunological challenges. First, as a model type 1 response, we infected Elp3TKO mice and control mice with IAV. We observed that Elp3TKO mice showed a pathological weight loss response to IAV infection similar to that of control mice (Fig. 2B). In addition, populations of effector CD4+ and CD8+ T cells in the draining lymph nodes were comparably increased in infected control and Elp3TKO mice (Fig. 2C). Similarly, in a model of type 2 response based on the repeated intranasal exposure to HDM, control and Elp3TKO mice developed comparable type 2 inflammation in their lung and Th2 responses in their lung draining lymph nodes (Fig. 2D).

To test a potential role of Elp3 in T cell–dependent humoral responses, we compared Ag-specific Ab titers in control and in Elp3TKO mice immunized with OVA-CFA. Serum OVA-specific IgG concentrations were significantly reduced in Elp3TKO mice (Fig. 3A). This lower Ab response correlated with lower frequency and total number of germinal center B (GCB) cells and plasma cells in the draining lymph nodes of Elp3TKO mice (Fig. 3B). Hence, Elp3 activity is required in T cells for the development of optimal Ab responses. The above results hinted at an impairment of TFH responses, which are instrumental in promoting GCB cell responses (26). Confirming this assumption, we observed that whereas OVA-CFA induced strong TFH responses in control mice, TFH (CXCR5+Foxp3PD1+CD4+) populations were reduced in Elp3TKO mice (Fig. 3C). Interestingly, numbers of TFH cells in the spleen of Elp3TKO mice were also reduced in the steady state (Fig. 3D). In another classical model of TFH cell responses consisting in the i.v. injection of SRBC, Elp3TKO mice also displayed reduced expansion of spleen GCB cells (Supplemental Fig. 3A). Immunofluorescence and histochemistry analysis revealed no difference in the general immune organization in the steady-state spleen of Elp3TKO mice. Yet, the spleen of Elp3TKO mice displayed much-reduced germinal centers (GC) following injection of SRBCs (Supplemental Fig. 3B).

FIGURE 3.

Elp3 is required for optimal TFH cell responses. (A) Experimental outline (upper panel) of the assessment of OVA-specific serum IgG1 titers in the blood of Elp3TKO and control mice immunized s.c. with OVA emulsified with CFA, 8 (lower left) or 15 (lower right) d postimmunization (n = 18 per group, data pooled from three experiments). (B) Flow cytometric profile (left) and numbers (right) of GCB (upper panel) and plasma cells (PC) (lower panel) in the inguinal lymph node (draining lymph node [dLN]), and (C) flow cytometric profile (left) and numbers (right) of TFH cell (n = 18 per group, data pooled from three experiments) of mice immunized for 8 d as in (A). (D) Flow cytometric profile (left) and absolute numbers (right) of TFH cells in the spleen of naive mice (n = 26–27 per group. Data pooled from six experiments). *p < 0.05, **p < 0.01, ****p < 1.10−4. ns, not significant.

FIGURE 3.

Elp3 is required for optimal TFH cell responses. (A) Experimental outline (upper panel) of the assessment of OVA-specific serum IgG1 titers in the blood of Elp3TKO and control mice immunized s.c. with OVA emulsified with CFA, 8 (lower left) or 15 (lower right) d postimmunization (n = 18 per group, data pooled from three experiments). (B) Flow cytometric profile (left) and numbers (right) of GCB (upper panel) and plasma cells (PC) (lower panel) in the inguinal lymph node (draining lymph node [dLN]), and (C) flow cytometric profile (left) and numbers (right) of TFH cell (n = 18 per group, data pooled from three experiments) of mice immunized for 8 d as in (A). (D) Flow cytometric profile (left) and absolute numbers (right) of TFH cells in the spleen of naive mice (n = 26–27 per group. Data pooled from six experiments). *p < 0.05, **p < 0.01, ****p < 1.10−4. ns, not significant.

Close modal

Reduced TFH numbers in Elp3TKO mice may result either from intrinsic defects in T cells or from impaired cross-talk with cognate B cells. To assess whether impaired GC responses in Elp3TKO mice resulted from T cell–intrinsic defects, we generated chimeric mice by reconstituting irradiated C57BL/6 recipient mice with a mixture of bone marrow cells from Elp3TKO (CD45.2+) and wild-type (CD45.1+) donor mice in a 1:1 ratio (Fig. 4A). Chimeric mice offered an environment in which Elp3TKO and wild-type TFH respond to the same stimuli and share the same B cell pool. We noted a very significant decrease in the percentage of activated CD62Llo CD44hiElp3TKO T cells in the chimeric mice (Fig. 4A). Yet, the percentage of TFH cells was comparable in the pool of activated control and Elp3TKO T cells (Fig. 4A). This observation indicated that TFH cell differentiation is not selectively impaired by inactivation of Elp3 and that Elp3-deficient TFH cells can be maintained by GCB cells. This notion is further supported by the observation that Elp3TKO and control CXCR5+ TFH cells displayed comparable levels of Bcl6 (Supplemental Fig. 3C). In addition, Elp3TKO TFH cell sorted from OVA/CFA-immunized mice had B cell helper activity in vitro that was highly comparable to that of their control counterparts (Fig. 4B).

FIGURE 4.

Elp3 is intrinsically required for the optimal development of TFH cell responses but does not impact on differentiated TFH B helper function. (A) Setup of bone marrow chimeras by the adoptive transfer of equal numbers of total bone marrow cells from control CD45.1+ and CD45.2 Elp3TKO donors to lethally irradiated CD45.1 recipient mice (upper panel) and subsequent flow cytometric assessment of each donor-derived T cell populations 10 d after s.c. immunization with OVA emulsified with CFA (OVA/CFA). Flow cytometric profiles (left) and percentages (right) of activated T cells (upper), and TFH cells out of activated T cells (middle) and total CD4 T cells (lower) in the inguinal lymph nodes are provided (n = 12 per group. Data pooled from two experiments). (B) Experimental layout of coculture experiments of increasing numbers of TFH cells sorted from the draining lymph nodes (dLN) of control and Elp3TKO mice 8 d after immunization as in (A) with syngeneic wild-type B cells from naive mice (upper panel) and number of B cells per well (left), percentage of IgG1-expressing B cells (center) and GC marker–expressing B cells (right) at the end of the coculture (n = 4 per group, data pooled from two experiments). (C) Experimental layout for the adoptive transfer of Elp3fl/fl or Elp3TKO CD45.2+ OT-II CD4+ T cells to naive CD45.1+ mice, followed by s.c. immunization with OVA emulsified with CFA (left) and number of transferred OT-II cells after 3 and 5 d in the dLN (n = 10 per group, data pooled from two experiments). (D) Flow cytometric assessment of the percentage of CD4+ CD44hi Cxcr5+ PD1+ TFH cells developing from adoptively transferred CD45.2+ CD4+ T cells transduced with retroviral silencing vectors against Elp3 and Ctu1 (or nontargeting control) into CD45.1+ naive recipient mice subsequently immunized against OVA as in (C) (n = 4 per group, one experiment). **p < 0.01, ***p < 1.10−4. ns, not significant.

FIGURE 4.

Elp3 is intrinsically required for the optimal development of TFH cell responses but does not impact on differentiated TFH B helper function. (A) Setup of bone marrow chimeras by the adoptive transfer of equal numbers of total bone marrow cells from control CD45.1+ and CD45.2 Elp3TKO donors to lethally irradiated CD45.1 recipient mice (upper panel) and subsequent flow cytometric assessment of each donor-derived T cell populations 10 d after s.c. immunization with OVA emulsified with CFA (OVA/CFA). Flow cytometric profiles (left) and percentages (right) of activated T cells (upper), and TFH cells out of activated T cells (middle) and total CD4 T cells (lower) in the inguinal lymph nodes are provided (n = 12 per group. Data pooled from two experiments). (B) Experimental layout of coculture experiments of increasing numbers of TFH cells sorted from the draining lymph nodes (dLN) of control and Elp3TKO mice 8 d after immunization as in (A) with syngeneic wild-type B cells from naive mice (upper panel) and number of B cells per well (left), percentage of IgG1-expressing B cells (center) and GC marker–expressing B cells (right) at the end of the coculture (n = 4 per group, data pooled from two experiments). (C) Experimental layout for the adoptive transfer of Elp3fl/fl or Elp3TKO CD45.2+ OT-II CD4+ T cells to naive CD45.1+ mice, followed by s.c. immunization with OVA emulsified with CFA (left) and number of transferred OT-II cells after 3 and 5 d in the dLN (n = 10 per group, data pooled from two experiments). (D) Flow cytometric assessment of the percentage of CD4+ CD44hi Cxcr5+ PD1+ TFH cells developing from adoptively transferred CD45.2+ CD4+ T cells transduced with retroviral silencing vectors against Elp3 and Ctu1 (or nontargeting control) into CD45.1+ naive recipient mice subsequently immunized against OVA as in (C) (n = 4 per group, one experiment). **p < 0.01, ***p < 1.10−4. ns, not significant.

Close modal

Hence, in line with our in vitro observations, the above results suggest that impaired TFH function in Elp3TKO mice originated at least in part from defective initial T cell expansion. To further test this possibility, we crossed OT-II mice, which carry an MHC class II–restricted TCR against a major epitope of OVA, with Elp3TKO mice (Elp3TKO-crossed OT-II [OT-IIElp3TKO)] mice. Littermates not carrying the Cd4-cre driver (OT-II not carrying the Cd4-cre driver [OT-IICtl]) were used as controls. We adoptively transferred dye-labeled OT-IICtl and OT-IIElp3TKO to CD45.1+ mice that we subsequently immunized with either OVA-CFA (Fig. 4C). Following immunization, compared with controls, OT-IIElp3TKO cells were present in reduced numbers (Fig. 4C). These results confirm that Elp3-deficient T cells underwent reduced initial proliferation in an otherwise wild-type environment.

To ascertain that the phenotype of Elp3-deficient TFH cells was indeed related to the loss of U34 modification in tRNAs, we silenced the expression of Ctu1, a subunit of the U34-modifying complex Ctu1/2, in OT-II cells and assessed the consequences thereof on TFH responses in recipient mice immunized with OVA-CFA (Supplemental Fig. 3D). We observed that silencing Ctu1 phenocopied the silencing of Elp3 in this setting (Fig. 4D), confirming the dependency of optimal TFH responses on U34 tRNA modification.

Loss of Elp3 has been repeatedly shown to elicit its main biological impacts in mammals by altering the expression of key transcription factors. This may happen as a consequence of hindered translation of said transcription factors or translational regulators thereof (1517) or more indirectly through induction of stress signaling responses (14). Therefore, we reasoned that transcriptomic profiling might reveal the key molecular pathways by which deficiency in U34 modification impacted on T cell responses. We therefore compared the transcriptome of control and Elp3TKO T cells 18 h after activation with anti-CD3ε and anti-CD28 Abs. We identified 20 differentially expressed genes (p < 0.05) between control and Elp3TKO T cells (Fig. 5A). GSEAR on the full transcriptome of these Elp3TKO T cells ordered by fold change compared with Elp3fl/fl control T cells indicated that these Elp3TKO T cells tended to harbor lower expression of cell cycle–related gene sets and upregulated expression of unfolded protein response genes (Fig. 5B). Downregulation of cell cycle–related genes is consistent with delayed cell cycle entry of Elp3TKO T cells. We noticed that most upregulated unfolded protein response genes consisted in Atf4-responsive genes, such as Ddit3/Chop, Psat1, and several aminoacyl-tRNA transferases (27). Indeed, the transcriptomic response of Elp3TKO T cells was highly enriched in Atf4 and Chop target genes (27) in GSEAR (Fig. 5C). Real-time RT-PCR confirmed the upregulation of Atf4 and its downstream target genes (Fig. 5D), and Western blot analyses confirmed the upregulation of Atf4 and Chop protein expression in activated Elp3TKO T lymphocytes (Fig. 5E).

FIGURE 5.

An Atf4/Chop response is induced by loss of Elp3 following T cell activation. (A) Heatmap of differentially expressed genes (p < 0.05) in control and Elp3TKO naive CD4+ T cells 18 h postactivation with immobilized anti-CD3 and soluble anti-CD28. (B) GSEAR analysis of Hallmark MSigDb gene sets upregulated and downregulated in the transcriptome of Elp3TKO T cells compared with their control counterparts. *p < 0.05, **p < 0.01, ***p < 1.10−3, for false discovery rate. NES, normalized enrichment score; ns, not significant. (C) GSEAR analysis for the enrichment in gene sets composed of Atf4, Chop or Atf4, and Chop targets. ***p < 1.10−3, for false discovery rate. (D) Relative expression by real-time RT-PCR of Atf4 and Atf4 target genes in naive CD4+ T cells activated in vitro for 0, 12, 24, and 36 h as in (A). *p < 0.05, **p < 0.01, ***p < 0.001, in pairwise t tests (n = 3 per group). (E) Western blot analysis of Atf4 and Chop protein expression following CD4+ T cell activation as in (A) (representative of two and one experiments, respectively).

FIGURE 5.

An Atf4/Chop response is induced by loss of Elp3 following T cell activation. (A) Heatmap of differentially expressed genes (p < 0.05) in control and Elp3TKO naive CD4+ T cells 18 h postactivation with immobilized anti-CD3 and soluble anti-CD28. (B) GSEAR analysis of Hallmark MSigDb gene sets upregulated and downregulated in the transcriptome of Elp3TKO T cells compared with their control counterparts. *p < 0.05, **p < 0.01, ***p < 1.10−3, for false discovery rate. NES, normalized enrichment score; ns, not significant. (C) GSEAR analysis for the enrichment in gene sets composed of Atf4, Chop or Atf4, and Chop targets. ***p < 1.10−3, for false discovery rate. (D) Relative expression by real-time RT-PCR of Atf4 and Atf4 target genes in naive CD4+ T cells activated in vitro for 0, 12, 24, and 36 h as in (A). *p < 0.05, **p < 0.01, ***p < 0.001, in pairwise t tests (n = 3 per group). (E) Western blot analysis of Atf4 and Chop protein expression following CD4+ T cell activation as in (A) (representative of two and one experiments, respectively).

Close modal

To assess whether Atf4 activation was also a feature of Elp3-deficient T cells in vivo, we compared the transcriptome of TFH cells sorted from the draining lymph nodes of OVA-CFA–immunized control and Elp3TKO mice 8 d after immunization. Transcriptomic differences were much more pronounced in this in vivo setting, with more than 2000 differentially expressed genes identified between control and Elp3TKO TFH cells (Fig. 6A). In line with our observations of delayed proliferation of Elp3-deficient T cells, Elp3TKO TFH cells displayed higher expression of genes belonging to the hallmark classes of E2F targets, G2M checkpoints, mitotic spindle, and Myc targets. In contrast, they displayed lower expression of genes belonging to hallmark classes associated with effector T cell functions such as inflammatory response, IL-6/JAK/STAT3 signaling, TNF-α/NF-κB signaling, KRAS signaling, or type I/II IFN response (Fig. 6B). Of note, upregulated genes in Elp3TKO TFH were also enriched in genes belonging to the mTORC1 signaling and unfolded protein response hallmark pathways, which overlap with the repertoire of Atf4 target genes. More importantly, GSEAR indicated enrichment in Atf4/Chop target genes among the genes more highly expressed in Elp3TKO TFH cells in vivo (Fig. 6C). The results above hence revealed increased Atf4/Chop signaling in Elp3-deficient CD4+ T cells, both in early in vitro activated T cells and in TFH cells in vivo.

FIGURE 6.

An Atf4/Chop response induced by loss of Elp3 hinders TFH cell responses. (A) MA plot of the changes in mRNA expression in TFH cells isolated from the draining lymph nodes of Elp3TKO and Elp3fl/fl mice 8 d after immunization with OVA emulsified in CFA. (B and C) GSEAR analysis of Hallmark MSigDb gene sets (B) and of gene sets composed of Atf4, Chop or Atf4, and Chop target genes enriched or depleted in the transcriptome of Elp3TKO TFH cells compared with their Elp3fl/fl counterparts. *p < 0.05, **p < 0.01, ***p < 1.10−3, for false discovery rate. NES, normalized enrichment score.

FIGURE 6.

An Atf4/Chop response induced by loss of Elp3 hinders TFH cell responses. (A) MA plot of the changes in mRNA expression in TFH cells isolated from the draining lymph nodes of Elp3TKO and Elp3fl/fl mice 8 d after immunization with OVA emulsified in CFA. (B and C) GSEAR analysis of Hallmark MSigDb gene sets (B) and of gene sets composed of Atf4, Chop or Atf4, and Chop target genes enriched or depleted in the transcriptome of Elp3TKO TFH cells compared with their Elp3fl/fl counterparts. *p < 0.05, **p < 0.01, ***p < 1.10−3, for false discovery rate. NES, normalized enrichment score.

Close modal

Having observed that inactivation of Elp3 in T cells activated Atf4 signaling as the most prominent transcriptional event, we aimed at determining whether deregulation of Atf4 was indeed responsible for the observed phenotype of Elp3TKO T cells. First, we generated mice overexpressing Atf4 in T cells (Cd4-CreT/+ Rosa26fl-STOP-fl-Atf4/+ Atf4TTG mice). We observed delayed cell cycle entry of naive Atf4TTG CD4+ T cells following activation with anti-CD3 and anti-CD28 Abs, phenocopying Elp3TKO cells in this assay (Fig. 7A, compare with Fig. 1B). Second, we retrovirally overexpressed Atf4 in OT-II CD4+ lymphocytes that we transferred to naive CD45.1+ host mice (Fig. 7B), which we subsequently immunized with OVA-CFA. Atf4 overexpression, like deficiency in Elp3, led to reduced numbers of OT-II T cells in the draining lymph nodes of immunized mice (Fig. 7C). Third, we tested whether reducing Atf4 activity could rescue the phenotype of Elp3-deficient T cells. To do so, we silenced Atf4 expression in OT-IIElp3-TKO or OT-IICtl CD4+ T cells that we transferred to naive CD45.1+ host mice before immunization with OVA-CFA (Supplemental Fig. 3E). We observed that whereas final numbers of OT-IICtl CD4+ T cells were not affected by silencing of Atf4, numbers of Atf4-silenced OT-IIElp3-TKO T cells were increased or normalized in a majority of recipient mice compared with those that received control OT-IIElp3-TKO T cells (Fig. 7D). Hence, overactivation of Atf4, alone or as a consequence of Elp3 inactivation, delays cell cycle entry in vitro and hinders CD4+ T cell expansion in vivo.

FIGURE 7.

Atf4/Chop upregulation is deleterious for T cell and TFH expansion. (A) Atf4TTG CD4+ T cells activation: flow cytometric profile (left) and percentages of cells in the G0/G1, S, and G2/M phases of the cell cycle (right) based on EdU incorporation and 7-aminoactinomycin D signal intensity (two-way ANOVA followed by Turkey honest significance test [HSD] test, n = 13–20 per group, data pooled from two experiments). (B) Experimental layout of the retroviral transduction of CD45.2+ OT-II T cells and their adoptive transfer to CD45.1+ naive recipient mice prior to immunization with OVA emulsified in CFA (OVA/CFA) and assessment of CD45.2+ T cell expansion in the draining lymph nodes (dLN) 7 d later. (C) Flow cytometric profile (left) and number of OT-II T cells normalized to the experiment mean (right) in OT-II T cells transduced with a control GFP or an Atf4 overexpression vector in recipient mice as in (A) (n = 11 per group, data pooled from three experiments). (D) Flow cytometric profile (left) and number of CD45.2+ T cells silenced or not for Atf4 with lentiviral vectors in recipient mice as in (A) (n = 8 per group, data pooled from two experiments). (E) Flow cytometric profile (upper panel) and numbers (lower panel) of activated T cell and TFH cell in the draining lymph node of control, Elp3TKO, Ddit3−/−, and Ddit3−/− Elp3TKO double-knockout mice 7 d after s.c. immunization with OVA/CFA (n = 12 per group, data pooled from three experiments). ***p < 0.001, ****p < 1.10−4. ns, not significant.

FIGURE 7.

Atf4/Chop upregulation is deleterious for T cell and TFH expansion. (A) Atf4TTG CD4+ T cells activation: flow cytometric profile (left) and percentages of cells in the G0/G1, S, and G2/M phases of the cell cycle (right) based on EdU incorporation and 7-aminoactinomycin D signal intensity (two-way ANOVA followed by Turkey honest significance test [HSD] test, n = 13–20 per group, data pooled from two experiments). (B) Experimental layout of the retroviral transduction of CD45.2+ OT-II T cells and their adoptive transfer to CD45.1+ naive recipient mice prior to immunization with OVA emulsified in CFA (OVA/CFA) and assessment of CD45.2+ T cell expansion in the draining lymph nodes (dLN) 7 d later. (C) Flow cytometric profile (left) and number of OT-II T cells normalized to the experiment mean (right) in OT-II T cells transduced with a control GFP or an Atf4 overexpression vector in recipient mice as in (A) (n = 11 per group, data pooled from three experiments). (D) Flow cytometric profile (left) and number of CD45.2+ T cells silenced or not for Atf4 with lentiviral vectors in recipient mice as in (A) (n = 8 per group, data pooled from two experiments). (E) Flow cytometric profile (upper panel) and numbers (lower panel) of activated T cell and TFH cell in the draining lymph node of control, Elp3TKO, Ddit3−/−, and Ddit3−/− Elp3TKO double-knockout mice 7 d after s.c. immunization with OVA/CFA (n = 12 per group, data pooled from three experiments). ***p < 0.001, ****p < 1.10−4. ns, not significant.

Close modal

Finally, because we observed that loss of Elp3 triggered upregulation of the Atf4 target transcription factor Chop, we tested whether the Atf4/Chop pathway was involved in the phenotype of Elp3TKO TFH cells. To do so, we crossed Elp3TKO mice with Ddit3−/− mice and compared TFH responses in OVA-CFA–immunized Elp3TKO, Ddit3−/−, and Elp3TKO Ddit3−/− mice. Although deletion of Ddit3 alone did not affect TFH responses, it partially restored T cell expansion and TFH cell numbers in Elp3TKO Ddit3−/− mice (Fig. 7E). These results hence confirmed the involvement of the Atf4/Chop pathway in the impairment of TFH cell responses following loss of Elp3.

The U34-modifying enzymes Elp3 and Ctu1/2 are attracting increasing attention, notably as potential therapeutic targets in cancer (17, 28). Our results show that loss of conserved tRNA U34-modifying enzymes delays T cell cycle entry and hinders TFH cell responses. We document that expression of the key U34-modifying enzymes Elp3 and Ctu1/2 is regulated at the posttranscriptional level. These observations further illustrate the extent to which T cell activation involves posttranscriptional regulation (4). Identifying the molecular mechanisms that regulate the translation or stability of Elp3 and Ctu1/2 subunits might thereby improve our understanding of the posttranscriptional regulation of T cell activation.

In line with the notion that U34 tRNA modifications influence cell-specific processes, we observed that inactivation of Elp3 had limited impact on CD4+ T cell responses in vitro. Indeed, we observed no significant impact on T cell polarization, and T cell proliferation rates appeared largely unaffected. Yet, entry into the first S phase of the cell cycle was significantly delayed, irrespective of the cytokinic environment provided, which resulted in slightly lower expansion of Elp3-deficient T cells following activation in vitro. These observations indicate that the activity of Elp3 is not essential for T cell cycling, but rather impacts on the transition between the G0 quiescent phase and entry into the S phase. In this regard, we foresee that the phenotype of Elp3-deficient naive T cells may share similarities with the Gcn4 metabolic response of yeast deficient for U34 modifications (21, 22). Indeed, along their progression toward cell cycle entry, Elp3-deficient T cells progressively mount an exaggerated Atf4 response, which is the functional ortholog of the yeast Gcn4 response. In yeast, the 5-carbamoylmethyl, 5-methoxycarbonylmethyl, and 2-thioylation U34 modifications, because they require sulfur-containing amino acids (cysteine and methionine) and s-adenosylmethionine as intermediates, have been co-opted into sensors of methionine availability (18). Thereby, yeast deficient in U34-modifying enzymes display a metabolic phenotype reminiscent of starvation (1820). We hence foresee that loss of U34 modifications in T cells transmits a starvation signal to newly activated T cells. This signal would delay their entry into the S phase as part of a yet-to-identify restriction point and result in the compensatory activation of the metabolic adaptor Atf4. Identifying the molecular signaling pathways downstream of U34 modifications relaying this putative metabolic signal could open interesting perspectives into the metabolic regulation of T cell exit from quiescence and, more generally, into a potentially overlooked biological function of U34 modifications in mammalian cells.

Among effector T cells responses that we tested, we found TFH responses to be the most significantly impacted by deficiency in U34-modifying enzymes. The origin of this higher requirement for U34-modifying enzymes in TFH responses compared with Th1 or Th2 responses remains elusive. We may, however, point toward two likely contributing factors. First, TFH responses require the initial stimulation of T cells by dendritic cells before they eventually reach the GC, where they receive further stimulation by GCB cells (2932). We may hence envision that delaying T cell cycle entry, as happens in Elp3-deficient T cells, may desynchronize TFH and GCB cell responses, leading to suboptimal cross-talk between these two codependent cell types (33). In line with this notion, Elp3TKO TFH cells displayed higher expression of cell cycle–related genes than their wild-type counterparts, which could be a consequence of delayed cell cycle entry. Second, we showed in this study that genetic deletion of the Atf4-responsive transcription factor Chop partially restored GC TFH cell numbers following immunization in vivo. Hence, activation of the Atf4/Chop stress response pathway also contributed to the impairment of TFH responses in Elp3TKO mice.

Atf4 is an important regulator of cellular metabolism, involved in antioxidant defenses, amino acid uptake, and biosynthesis of essential metabolic intermediates (34). It is also involved in the balance between autophagy and cell adaptation versus cell death induction in stressful situations as a converging target of the integrated stress response (34). Perhaps not surprisingly, complete genetic invalidation of Atf4 has severe consequences on T cell activation, proliferation, and differentiation (35). In this study, we report that, although to a lesser extent, excessive expression of Atf4 also has detrimental consequences on T cells, negatively affecting TFH cell development and function in vivo by driving the expression of Chop. Together, our report and that of others (35) support the notion that Atf4 activity acts as a rheostat-conditioning T cell responses.

Altogether, our results reveal an original role of U34-modifying enzymes in T cell responses and call for a deeper understanding of the role of these epitranscriptomic modifiers and potential drug targets in T cell biology and immunity in general.

We thank the flow cytometry platform, the mouse facility, and the Genomics platform of the Interdisciplinary Group for Applied Genoproteomics, University of Liege. We also thank R. Fares, C. François, and I. Sbai for excellent technical and secretarial assistance.

This work was supported by the Fonds de la Recherche Scientifique, National Fund for Scientific Research (Credit de Recherche Grant J00641F and Walloon Excellence in Life Sciences and Biotechnology FRFS-WELBIO Grant CR-2015S-01), the Fonds pour la Formation à la Recherche dans l’Industrie et dans l’Agriculture, National Fund for Scientific Research (to P.L. and C.L.), the Léon Frédéricq Foundation of the University of Liege, and the University of Liege (Action de Recherche Concertee: tRAME). Funders took no part in data collection, analysis, or manuscript preparation.

The RNA sequencing data presented in this article have been submitted to ArrayExpress database (https://www.ebi.ac.uk/arrayexpress) under accession numbers E-MTAB-9057 and E-MTAB-9059.

The online version of this article contains supplemental material.

Abbreviations used in this article:

Atf4

activating transcription factor 4

Ctu

cytosolic thiouridylase

Elp3TKO

Elp3fl/fl Cd4-CreT/+

GC

germinal center

GCB

germinal center B

GSEAR

ranked gene set enrichment analysis

HDM

house dust mite extract

IAV

influenza A virus

i.t.

intratracheal

OT-IICtl

OT-II not carrying the Cd4-cre driver

OT-IIElp3TKO

Elp3TKO-crossed OT-II

TFH

T follicular helper

U34

wobble uridine.

1
Smith-Garvin
,
J. E.
,
G. A.
Koretzky
,
M. S.
Jordan
.
2009
.
T cell activation.
Annu. Rev. Immunol.
27
:
591
619
.
2
Chapman
,
N. M.
,
H.
Chi
.
2018
.
Hallmarks of T-cell exit from quiescence.
Cancer Immunol. Res.
6
:
502
508
.
3
Zhu
,
J.
,
H.
Yamane
,
W. E.
Paul
.
2010
.
Differentiation of effector CD4 T cell populations (*).
Annu. Rev. Immunol.
28
:
445
489
.
4
Tan
,
H.
,
K.
Yang
,
Y.
Li
,
T. I.
Shaw
,
Y.
Wang
,
D. B.
Blanco
,
X.
Wang
,
J.-H.
Cho
,
H.
Wang
,
S.
Rankin
, et al
.
2017
.
Integrative proteomics and phosphoproteomics profiling reveals dynamic signaling networks and bioenergetics pathways underlying T cell activation.
Immunity
46
:
488
503
.
5
El Yacoubi
,
B.
,
M.
Bailly
,
V.
de Crécy-Lagard
.
2012
.
Biosynthesis and function of posttranscriptional modifications of transfer RNAs.
Annu. Rev. Genet.
46
:
69
95
.
6
Krutyhołowa
,
R.
,
K.
Zakrzewski
,
S.
Glatt
.
2019
.
Charging the code - tRNA modification complexes.
Curr. Opin. Struct. Biol.
55
:
138
146
.
7
Karlsborn
,
T.
,
H.
Tükenmez
,
A. K. M. F.
Mahmud
,
F.
Xu
,
H.
Xu
,
A. S.
Byström
.
2014
.
Elongator, a conserved complex required for wobble uridine modifications in eukaryotes.
RNA Biol.
11
:
1519
1528
.
8
Johansson
,
M. J.
,
A.
Esberg
,
B.
Huang
,
G. R.
Björk
,
A. S.
Byström
.
2008
.
Eukaryotic wobble uridine modifications promote a functionally redundant decoding system.
Mol. Cell. Biol.
28
:
3301
3312
.
9
Rezgui
,
V. A.
,
K.
Tyagi
,
N.
Ranjan
,
A. L.
Konevega
,
J.
Mittelstaet
,
M. V.
Rodnina
,
M.
Peter
,
P. G.
Pedrioli
.
2013
.
tRNA tKUUU, tQUUG, and tEUUC wobble position modifications fine-tune protein translation by promoting ribosome A-site binding.
Proc. Natl. Acad. Sci. USA
110
:
12289
12294
.
10
Nedialkova
,
D. D.
,
S. A.
Leidel
.
2015
.
Optimization of codon translation rates via tRNA modifications maintains proteome integrity.
Cell
161
:
1606
1618
.
11
Cheishvili
,
D.
,
C.
Maayan
,
R.
Cohen-Kupiec
,
S.
Lefler
,
M.
Weil
,
G.
Ast
,
A.
Razin
.
2011
.
IKAP/Elp1 involvement in cytoskeleton regulation and implication for familial dysautonomia.
Hum. Mol. Genet.
20
:
1585
1594
.
12
Slaugenhaupt
,
S. A.
,
A.
Blumenfeld
,
S. P.
Gill
,
M.
Leyne
,
J.
Mull
,
M. P.
Cuajungco
,
C. B.
Liebert
,
B.
Chadwick
,
M.
Idelson
,
L.
Reznik
, et al
.
2001
.
Tissue-specific expression of a splicing mutation in the IKBKAP gene causes familial dysautonomia.
Am. J. Hum. Genet.
68
:
598
605
.
13
George
,
L.
,
M.
Chaverra
,
L.
Wolfe
,
J.
Thorne
,
M.
Close-Davis
,
A.
Eibs
,
V.
Riojas
,
A.
Grindeland
,
M.
Orr
,
G. A.
Carlson
,
F.
Lefcort
.
2013
.
Familial dysautonomia model reveals Ikbkap deletion causes apoptosis of Pax3+ progenitors and peripheral neurons.
Proc. Natl. Acad. Sci. USA
110
:
18698
18703
.
14
Laguesse
,
S.
,
C.
Creppe
,
D. D.
Nedialkova
,
P.-P.
Prévot
,
L.
Borgs
,
S.
Huysseune
,
B.
Franco
,
G.
Duysens
,
N.
Krusy
,
G.
Lee
, et al
.
2015
.
A dynamic unfolded protein response contributes to the control of cortical neurogenesis.
Dev. Cell
35
:
553
567
.
15
Ladang
,
A.
,
F.
Rapino
,
L. C.
Heukamp
,
L.
Tharun
,
K.
Shostak
,
D.
Hermand
,
S.
Delaunay
,
I.
Klevernic
,
Z.
Jiang
,
N.
Jacques
, et al
.
2015
.
Elp3 drives Wnt-dependent tumor initiation and regeneration in the intestine. [Published erratum appears in 2017 J. Exp. Med. 214: 1199.]
J. Exp. Med.
212
:
2057
2075
.
16
Delaunay
,
S.
,
F.
Rapino
,
L.
Tharun
,
Z.
Zhou
,
L.
Heukamp
,
M.
Termathe
,
K.
Shostak
,
I.
Klevernic
,
A.
Florin
,
H.
Desmecht
, et al
.
2016
.
Elp3 links tRNA modification to IRES-dependent translation of LEF1 to sustain metastasis in breast cancer.
J. Exp. Med.
213
:
2503
2523
.
17
Rapino
,
F.
,
S.
Delaunay
,
F.
Rambow
,
Z.
Zhou
,
L.
Tharun
,
P.
De Tullio
,
O.
Sin
,
K.
Shostak
,
S.
Schmitz
,
J.
Piepers
, et al
.
2018
.
Codon-specific translation reprogramming promotes resistance to targeted therapy.
Nature
558
:
605
609
.
18
Laxman
,
S.
,
B. M.
Sutter
,
X.
Wu
,
S.
Kumar
,
X.
Guo
,
D. C.
Trudgian
,
H.
Mirzaei
,
B. P.
Tu
.
2013
.
Sulfur amino acids regulate translational capacity and metabolic homeostasis through modulation of tRNA thiolation.
Cell
154
:
416
429
.
19
Gupta
,
R.
,
S.
Laxman
.
2020
.
tRNA wobble-uridine modifications as amino acid sensors and regulators of cellular metabolic state.
Curr. Genet.
66
:
475
480
.
20
Gupta
,
R.
,
A. S.
Walvekar
,
S.
Liang
,
Z.
Rashida
,
P.
Shah
,
S.
Laxman
.
2019
.
A tRNA modification balances carbon and nitrogen metabolism by regulating phosphate homeostasis.
Elife
8
: e44795.
21
Zinshteyn
,
B.
,
W. V.
Gilbert
.
2013
.
Loss of a conserved tRNA anticodon modification perturbs cellular signaling.
PLoS Genet.
9
: e1003675.
22
Chou
,
H.-J.
,
E.
Donnard
,
H. T.
Gustafsson
,
M.
Garber
,
O. J.
Rando
.
2017
.
Transcriptome-wide analysis of roles for tRNA modifications in translational regulation.
Mol. Cell
68
:
978
992.e4
.
23
Quah
,
B. J. C.
,
H. S.
Warren
,
C. R.
Parish
.
2007
.
Monitoring lymphocyte proliferation in vitro and in vivo with the intracellular fluorescent dye carboxyfluorescein diacetate succinimidyl ester.
Nat. Protoc.
2
:
2049
2056
.
24
R Core Team
.
2016
. R: a language and environment for statistical computing.
R Foundation
,
Vienna, Austria
. Available at: https://www.R-project.org/. Accessed: October 1, 2017.
25
Lee
,
P. P.
,
D. R.
Fitzpatrick
,
C.
Beard
,
H. K.
Jessup
,
S.
Lehar
,
K. W.
Makar
,
M.
Pérez-Melgosa
,
M. T.
Sweetser
,
M. S.
Schlissel
,
S.
Nguyen
, et al
.
2001
.
A critical role for Dnmt1 and DNA methylation in T cell development, function, and survival.
Immunity
15
:
763
774
.
26
Vinuesa
,
C. G.
,
M. A.
Linterman
,
D.
Yu
,
I. C. M.
MacLennan
.
2016
.
Follicular helper T cells.
Annu. Rev. Immunol.
34
:
335
368
.
27
Han
,
J.
,
S. H.
Back
,
J.
Hur
,
Y.-H.
Lin
,
R.
Gildersleeve
,
J.
Shan
,
C. L.
Yuan
,
D.
Krokowski
,
S.
Wang
,
M.
Hatzoglou
, et al
.
2013
.
ER-stress-induced transcriptional regulation increases protein synthesis leading to cell death.
Nat. Cell Biol.
15
:
481
490
.
28
Rapino
,
F.
,
S.
Delaunay
,
Z.
Zhou
,
A.
Chariot
,
P.
Close
.
2017
.
tRNA modification: is cancer having a wobble?
Trends Cancer
3
:
249
252
.
29
Goenka
,
R.
,
L. G.
Barnett
,
J. S.
Silver
,
P. J.
O’Neill
,
C. A.
Hunter
,
M. P.
Cancro
,
T. M.
Laufer
.
2011
.
Cutting edge: dendritic cell-restricted antigen presentation initiates the follicular helper T cell program but cannot complete ultimate effector differentiation.
J. Immunol.
187
:
1091
1095
.
30
Benson
,
R. A.
,
M. K.
MacLeod
,
B. G.
Hale
,
A.
Patakas
,
P.
Garside
,
J. M.
Brewer
.
2015
.
Antigen presentation kinetics control T cell/dendritic cell interactions and follicular helper T cell generation in vivo.
Elife
4
: e06994.
31
Krishnaswamy
,
J. K.
,
U.
Gowthaman
,
B.
Zhang
,
J.
Mattsson
,
L.
Szeponik
,
D.
Liu
,
R.
Wu
,
T.
White
,
S.
Calabro
,
L.
Xu
, et al
.
2017
.
Migratory CD11b+ conventional dendritic cells induce T follicular helper cell-dependent antibody responses.
Sci. Immunol.
2
: eaam9169.
32
Eisenbarth
,
S. C.
,
D. A.
Piggott
,
J. W.
Huleatt
,
I.
Visintin
,
C. A.
Herrick
,
K.
Bottomly
.
2002
.
Lipopolysaccharide-enhanced, toll-like receptor 4-dependent T helper cell type 2 responses to inhaled antigen.
J. Exp. Med.
196
:
1645
1651
.
33
Baumjohann
,
D.
,
S.
Preite
,
A.
Reboldi
,
F.
Ronchi
,
K. M.
Ansel
,
A.
Lanzavecchia
,
F.
Sallusto
.
2013
.
Persistent antigen and germinal center B cells sustain T follicular helper cell responses and phenotype.
Immunity
38
:
596
605
.
34
Wortel
,
I. M. N.
,
L. T.
van der Meer
,
M. S.
Kilberg
,
F. N.
van Leeuwen
.
2017
.
Surviving stress: modulation of ATF4-mediated stress responses in normal and malignant cells.
Trends Endocrinol. Metab.
28
:
794
806
.
35
Yang
,
X.
,
R.
Xia
,
C.
Yue
,
W.
Zhai
,
W.
Du
,
Q.
Yang
,
H.
Cao
,
X.
Chen
,
D.
Obando
,
Y.
Zhu
, et al
.
2018
.
ATF4 regulates CD4+ T cell immune responses through metabolic reprogramming.
Cell Rep.
23
:
1754
1766
.

A.C. is director of research, P.C. is senior research associate, and C.J.D. is research associate of the Fonds de la Recherche Scientifique, National Fund for Scientific Research (Belgium). The other authors have no financial conflicts of interest.

Supplementary data