As part of an unbiased screen for genetic control of the peripheral T cell compartment, N-ethyl-N-nitrosourea (ENU)-exposed C57BL/6 mice were bred to Foxp3^GFP females to generate a standard F2 intercross pedigree (22). The inclusion of Foxp3^GFP allowed changes to the regulatory T cell (Treg) compartment to be included in the screening protocol. The resulting ENU mutants were screened for altered ratios of CD4 and CD8 T cells in the peripheral blood and spleen. Within one pedigree, individuals were identified with a decreased CD4:CD8 ratio in the peripheral blood. Intercrossing of affected individuals resulted in a mutant strain which consistently demonstrated decreased CD4+ cells and a decreased CD4:CD8 ratio in the spleen at 5–6 weeks of age (Figures 1A–C). All-exon sequencing of affected individuals identified an A → G nucleotide substitution at nucleotide 1469 of the Nup210 gene, which was confirmed by Sanger sequencing (Figure 1D). This mutation in exon 11 (Figure 1E) resulted in a predicted isoleucine to threonine change at amino acid 476 (Nup210^I476T). The mutation was located in an invariant amino acid in a region of Nup210 highly conserved throughout vertebrates (Figure 1F).

Nup210 is thought to be a key component of the nuclear pore complex (NPC), forming a membrane ring around the NPC (15). Due to the lack of known biology linking Nup210 to T cell-specific processes, we sought to validate the mutation through an F2 phenotyping cross and a complementation cross. First, Nup210^I476T mice were crossed with wildtype mice to produce an F1 generation and then intercrossed to produce an F2 generation. Genotyping for the Nup210^I476T allele allowed identification of wildtype, heterozygous and homozygous mice, while other, unlinked, ENU-induced alleles would be randomly segregated across the pedigree. Phenotyping of the F2 cross found a replication of the original finding, a reduction in CD4+ T cells and the CD4:CD8 ratio in Nup210^I476T homozygous mice (Figures 1G,H Supplementary Figure 1). Second, for a complementation cross, mice bearing a Nup210KO allele were generated using EUCOMM ES cells, with an insertion of a LacZ/Neo cassette between exons 3 and 5 (Supplementary Figure 3). Mice bearing one copy of the Nup210^I476T allele and one copy of the Nup210^KO allele phenocopied the Nup210^I476T homozygous mice (Figure 1I). Together, these data validated Nup210 as a novel genetic control mechanism controlling CD4:CD8 ratio.

While no known functions of Nup210 are linked specifically to T cell biology, the rapid rate of proliferation of early stage thymocytes (double negative T cells) is known to sensitize this lineage to minor genetic insults in general cell biology components, such as the chromatin condensing unit kleisin beta (7). Based on the essential function of Nup210 in the nuclear pore complex, and the lethal phenotype that results from knock-down of Nup210 in cell lines (23), it was assumed that the Nup210^I476T allele was a mild hypomorph, with sufficient function maintained to prevent cell death in lineages beyond early T cell stages. The generation of mice bearing the Nup210^KO allele allowed direct testing of this hypothesis, by intercrossing Nup210^het mice to create Nup210^KO mice. Surprisingly, intercross of Nup210^het mice produced wildtype:Nup210^het:Nup210^KO mice at the Mendelian 1:2:1 ratio. Nup210^KO mice were viable, demonstrated no visual abnormalities or altered total bodyweight or the weight of key immunological organs (Figure 2A). Histological screening of the organs was unremarkable (data not shown). Complete knock-out of Nup210 in Nup210^KO mice was confirmed by Western blot (Figure 2B), leaving the perplexing finding that an evolutionarily conserved component of the nuclear pore complex is largely redundant for life. We therefore generated mouse embryonic fibroblast (MEF) lines from wildtype and Nup210^KO littermates to allow for nucleopore imaging. Fibroblasts were stained with mAb414, which detects a number of FXFG-repeat-containing nucleoporins, such as Nup62, Nup153, Nup214, and Nup358/RanBP2. Confocal imaging revealed that the frequency and distribution of nuclear pores between the two genotypes was similar (Supplementary Figure 2). Nup210 does not, therefore, have non-redundant functions in nucleopore formation.

Thymocytes are highly susceptible to cellular stress, owing to the rapid proliferation at the double negative (DN) stage of development in the thymus, followed by the precarious double positive (DP) stage, at which “death by neglect” is the default outcome (1). We therefore tested whether the peripheral T cell phenotype observed in Nup210^KO mice is due to thymic defects. No difference was observed between wildtype and knockout mice in the DN or DP stages, or in single positive (SP) thymocytes which had diverged into either the CD4 or CD8 lineage (Figure 2C), or Foxp3+ regulatory sublineage (Figure 2D). Key transcription factors known to impact the CD4-CD8 lineage decision in the thymus, Tbx21, Runx3, and Thpok, remained unchanged throughout thymic differentiation (Figures 2E–G). In order to test the efficiency of negative selection, we utilized the endogenous MMTV self-superantigen, which recognizes both Vβ5 and Vβ11 TCR and I-E MHC molecules with a high affinity (24). Using a mixed bone-marrow chimeric system, of CD45.1 wildtype bone-marrow and CD45.2 Nup210^KO bone-marrow, we reconstituted B6.Rag^KO mice and B10.BR.Rag^KO mice. As B6 mice do not express I-E, MMTV can only drive TCR signaling weakly, allowing the Vβ5- and Vβ11-bearing T cells to pass through negative selection intact (Figures 2H,I). By contrast, in B10.BR mice, Vβ5- and Vβ11-bearing T cells receive a strong TCR signal from I-E, and undergo efficient negative selection (Figures 2H,I), a system which allows the detection of genetic modifiers of negative selection efficiency (25). In comparison to wildtype thymocytes, Nup210-deficient thymocytes demonstrated efficient negative selection (Figures 2H,I). Nup210^KO mice, in summary, demonstrate largely intact thymic development, through early differentiation, positive selection, CD4-CD8 lineage separation and negative selection. Together, these results demonstrate that the role of Nup210 in shaping the peripheral T cell compartment is a novel, peripheral-specific function, rather than a reflection of thymic sensitivity to basic cellular processes.

Intact thymic development in Nup210^KO mice indicates that the CD4-CD8 ratio skew initially observed in Nup210^KO mice results from peripheral abnormalities. Both the myeloid compartment (Supplementary Figure 4) and B cell compartment (Supplementary Figure 5) were largely unaltered. We therefore investigated the peripheral T cell compartment in greater detail (Supplementary Figure 6). Nup210^KO mice showed a near-50% reduction in total T cell numbers (Figure 3A), less notable in percentages (Figure 3B). γδ T cells were largely intact, with no alteration in the frequency (Figure 3C). Within the αβ T cell lineage, Nup210^KO mice demonstrated the greatest change in naïve CD4 T cells, which were reduced by 70% in absolute numbers in the spleen (Figure 3D). While other CD4 T cell subsets were intact in absolute numbers, the reduction in naïve T cells was reflected in a proportional increase in activated T cells (Figure 3E). CD8 T cells were affected in a similar way, with a 50% reduction in naïve CD8 T cell numbers (Figure 3F), and corresponding increases in proportion of central memory (TCM) cells (Figure 3G). Tregs (Figure 3H) and Treg subsets (Figure 3I) were largely intact. The relatively mild decrease in naïve CD8 T cells, compared to naïve CD4 T cells, is sufficient to explain the initial CD4-CD8 ratio skew observed (Figure 1I). The molecular basis for the differential CD4 T cell compared to CD8 T cell response was not due on differential expression of Nup210 (in wildtype mice) (Figure 3J), and compensatory upregulation of two other scaffolding proteins, Pom121 and Ndc1, was not increased in response to loss of Nup210 (Figures 3K,L). Together, these results demonstrate that the role of Nup210 in shaping the peripheral T cell compartment is a novel, peripheral-specific function, rather than a reflection of thymic sensitivity to basic cellular processes.

Further investigation into the function of Nup210 in T cells was led by compartment analysis. Mixed bone-marrow chimeras were set up, where wildtype and Nup210^KO haematopoietic stem cells reconstitute an irradiated mouse. This approach allows the competitive comparison of wildtype and Nup210^KO T cells in a context where the host environment is directly shared, thus excluding any effects of non-haematopoietic origin. Furthermore, the mixed bone-marrow chimera approach allows intrinsic versus extrinsic differences to be identified, as extrinsic differences act in trans and are shared across bone-marrow origin, and only cis-acting intrinsic differences allow an origin-dependent phenotype. Using this system, the CD4:CD8 ratio disturbance still developed in Nup210^KO cells in this context (Figure 4A), demonstrating that the function of Nup210 in driving this phenotype is intrinsic to T cells. The same result was also demonstrated when reconstituting Rag-deficient mice, where contamination from residual host-derived T cells can be excluded (Figure 4B). These results indicate a T cell-intrinsic phenotype driven by Nup210-deficiency.

The demonstration that the function of Nup210 was T cell-intrinsic led to further fine phenotyping of Nup210^KO mice. Using the 1W1K immunization system, antigen-specific T cell responses were assessed (Figure 5A). While Nup210^KO mice showed an increased percentage of 1W1K-reactive T cells (Figure 5B), the absolute magnitude of the response was normal (Figure 5C), indicating that antigen-specific T cell responses were intact. Regulatory T cell responses were assessed through in vitro suppressive assays, where purified Tregs from Nup210^KO mice showed a normal level of suppressive function (Figure 5D). These results indicate that basic immunogenic and tolerogenic processes are intact in Nup210-deficient T cells.

In contrast to the basic immunogenic and tolerogenic processes, Th1 and Th2 responses were substantially altered in Nup210^KO mice. Compared to heterozygous littermates, Nup210^KO mice manifested increased numbers of IFNγ-producing CD4 and CD8 cells after stimulation, while Th2 and Th17 cells were unchanged (Figures 6A,B). These results at the cellular level were disparate from at the transcriptional level, where ex vivo T cells demonstrated normal Tbet (Figure 6C) and GATA3 expression (Figure 6D), the Th1 and Th2 master transcription factors, respectively. Using an in vitro differentiation system, the induction of classical Th1 and Th2 cells was normal, as defined by expression of Tbet and GATA3 (Figure 6E). However an unusual population was observed in Nup210^KO T cells, where Tbet⁺GATA3⁺ T cells emerged from Th2-polarizing conditions (Figure 6E). At the cytokine production level, a similar phenotype was observed whereby IL-4 production was normal under both Th1- and Th2-polarizing conditions, and IFNγ production was normal in Th1-polarizing conditions, however only Nup210-deficient T cells demonstrated IFNγ production in Th2-polarizing conditions (Figure 6F). Together, these data were suggestive of a pro-inflammatory status of Nup210-deficient mice. An immunological challenge to induce inflammation, using the collagen-induced arthritis model, did not, however, identify any susceptibility to inflammatory disease (Supplementary Figure 7). Together, these results demonstrate that Nup210 restrains IFNγ production under inappropriate, Th2-biased conditions, however the deficiency effect remains subclinical even under inflammatory conditions.

All animal experiments were approved by the Animal Ethics Committee of the KU Leuven and performed in accordance with the approved protocol. To generate the Nup210^ENUI476T strain, founder C57BL/6 male mice were treated with 100 mg/kg ENU and bred to Foxp3^GFP females (46). First-generation (F1) male offspring were bred back to WT females to produce the second-generation (F2) offspring, which were in turn inter-crossed to produce the third generation for phenotypic screening. Phenotypic screening involved flow-cytometric analysis for CD4 and CD8 in the blood. NUP210^KO mice were generated through the European Conditional Mouse Mutagenesis (EUCOMM) program, by insertion of a LacZ/Neo cassette between exons 3 and 5 (Supplementary Figure 1).

Surface staining was performed in RPMI containing 2% fetal bovine serum and anti-mouse CD16/CD32 Fc block (from hybridoma supernatant generated in-house). The following antibodies were used in this study: CD3 (145-2C11), CD4 (GK1.5 and RM4-5), CD8α (53-6.7), Foxp3 (FJK-16s), IFNγ (XMG1.2), IL-17 (TC11-18H10), IL-4 (BVD6-24G2), IL-2 (JES6-5H4), IL-10 (JES5-16E3), Ly5.1 (A20), Ly5.2 (104) (eBioscience, CA, USA). Nuclear staining for Foxp3 was performed according to the manufacturer's recommendations (eBioScience, CA, USA). Intracellular cytokine staining was performed after a 4-h stimulation with 50 ng/mL phorbol 12-myristate 13-acetate and 0.5 μg/mL ionomycin (Sigma) in the presence of GolgiStop (Monensin A, BD Biosciences, NJ, USA), using Cytofix/Cytoperm (BD Biosciences, NJ, USA). Dead cells were excluded from analysis by staining with Zombie dyes (BioLegend, San Diego, CA, USA) according to the manufacturer's instructions. Samples were analyzed using a BD FACSCanto II instrument (BD Biosciences) and FlowJo software (Treestar Inc., OR, USA).

CIA was induced in 6–10 week old mice as previously described (47). A total of 100 μL of chick type II collagen (CII, final concentration 1 mg/ml; Sigma, MO, USA) emulsified in complete Freund's adjuvant containing 5 mg/ml heat-killed M. tuberculosis H37RA (BD Difco, NJ, USA) was injected intradermally in two sites at the base of the tail. The injections were repeated 21 days later. Animals were monitored three times weekly for erythema and swelling of limbs, and a clinical score (0–3) was given for each paw. Serum was collected using serum separator tubes (Greiner, Vilvoorde, Belgium) and analyzed for anti-chick collagen type II IgG antibodies by ELISA as described (48). Standard curves were constructed from pooled sera of CII hyper-immunized DBA/1 mice, set at 100,000 units/ml.

Red cell-depleted bone marrow from donor mice was depleted for mature T cells by incubation with biotinylated antibodies to CD3, CD4, and CD8 (eBioscience, San Diego, USA) followed by streptavidin-coupled Dynabeads and magnetic separation according to the manufacturer's instructions (ThermoFisher, Gent, Belgium). Bone-marrow chimeras were generated by lethal irradiation (7 Gy for Rag^KO recipients, 9.5 Gy for lymphocompetent recipients) of recipients, followed by intravenous (i.v.) injection of 5 × 10⁶ donor cells in saline. Reconstitution was analyzed after 8 weeks by flow cytometry.

Wildtype (C57BL/6) and NUP210^KO mice were immunized subcutaneously with 50 ug of 1W1K peptide (EAWGALANKAVDKA, Cambridge Research Biochemicals) emulsified in Incomplete Freund's Adjuvant (F5506, Sigma). Eight days after immunization draining LNs were harvested, processed to single cell suspensions and stained with I-A^b1W1K tetramer (NIH tetramer bank) for 2 h at room temperature, followed by anti-CD4 (RM4-5, eBioscience) and anti-CD44 (IM7, BioLegend) for 45 min at 4°C. Data were collected on a FACSCantoII (BD Biosciences) and analyzed with FlowJo (Treestar).

Quantitative RT-PCR was performed on purified mRNA (Trizol reagent, Ambion, Belgium) from cell populations sorted by flow cytometry from thymi of 6-week old mice, using the GoScript Reverse Transcription kit (Promega, Wisconsin, USA) and FastSYBR green reagents (ThermoFisher, Belgium). The expression of both housekeeping genes hypoxanthine-guanine phosphoribosyltransferase (Hprt) and peptidyl-prolyl cis-trans isomerase A (Ppia) was used to normalize mRNA expression. PCRs were performed in triplicate. Statistical analysis was performed using Prism (GraphPad). A significance threshold of 5% using non-parametric Mann–Whitney U-tests was applied.

Western Blot was performed on thymocytes lysed by sonication in lysis buffer [200 mM NaCl, 50 mM Tris pH 7.5, 1% Triton X-100, 2 mM dithiothreitol, 1 mM EDTA, protease inhibitor (ThermoFisher, Gent, Belgium)]. Lysates (20 ug) were run on 8% NuPAGE BisTris gels and blotted to a polyvinylidene fluoride transfer membrane using the NuPage electrophoresis system (ThermoFisher, Gent, Belgium) according to the manufacturer's recommendations. After washing in NCP (147 mM NaCl, 40 mM Tris pH 8, 0.01% Tween), the membrane was blocked overnight at 4°C with 5% non-fat milk in NCP 0.01% Tween. Primary antibodies against Nup210 (Abcam, Cambridge, UK, ab15600, 1:500) or control vinculin (Sigma-Aldrich, St.Louis, USA, V9131, 1/2000), were incubated in NCP 0.01% Tween, 1% non-fat milk. The membrane was washed in NCP 0.01% Tween and the primary antibody was detected with horseradish peroxidase-conjugated anti-rabbit secondary antibody (ThermoFisher, Gent, Belgium, 1:40,000) for Nup210 and HRP-conjugated anti-mouse secondary antibody (Merck Millipore, Darmstadt, Germany, 1:10,000). After washing in NCP 0.01% Tween, blots were developed using the Amersham ECL Prime Western Blotting Detection Reagent (GE Healthcare, Buckinghamshire, UK). The Spectra multicolor high range protein ladder (ThermoFisher, Gent, Belgium) was used to determine the molecular weights of the visualized bands.

Mouse embryonic fibroblasts were plated onto collagen-coated glass coverslips and cultured overnight to 80% confluency in DMEM high glucose, GlutaMAX, with pyruvate (Thermofisher), supplemented with 15% hyclone fetal calf serum (Thermofisher). Cultures were fixed for 10 min at room temperature with 4% paraformaldehyde in 1 X PBS and washed three times with 1 X PBS. Cells were then blocked for 1 h with 10% normal donkey serum diluted in 1 X PBS also containing 0.25% TritonX-100, followed by overnight incubation at 4°C with Monoclonal Antibody 414 (Mab414; AbCam ab24609) diluted 1:50 in the same solution. The following day coverslips were briefly rinsed in 1 X PBS, washed 3 x 10 min with 1 X PBS, and then incubated for 1 h at room temperature with Alexa Fluor 488 anti-mouse (Jackson Immunoresearch) in the 10% normal donkey serum solution. A second set of washes were applied and then coverslips were inverted and mounted onto slides with Vectashield containing DAPI (Vector Laboratories). For imaging a Nikon (Tokyo, Japan) C2 confocal scanhead attached to a Nikon TiE inverted microscope outfitted with a APO 60x 1.4 NA oil objective lens was used. Images of different genotypes were processed identically using ImageJ and Photoshop.

The conditions for in vitro suppression of CD4 T cell proliferation by Tregs were adapted from Collison and Vignali (49). Briefly, CD4⁺CD25⁺ Treg were flow sorted from WT and Nup210^−/− mice on a BD Aria I. Conventional CD4⁺Foxp3⁻ T cells (Tconv) were sorted from CD45.1 Foxp3-Thy1.1 reporter mice (50). Tconv (10⁵) were labeled with CellTrace Violet (Invitrogen), and were co-cultured with varying ratios of unlabeled Treg. Proliferation was stimulated by the addition of Rag1^−/− splenocytes (5 × 10⁴) and anti-CD3 (0.25 μg/ml, clone 145-2C11, eBioscience). After 5 days, the cells were stained for CD4-PE-Dazzle594 (clone GK1.5, BioLegend), Thy1.1-APC (clone HIS51, eBioscience), CD45.1-PE-Cy7 (clone A20, eBioscience), and fixable viability dye eFluor780, and the proliferation of the viable CD45.1⁺CD4⁺ Tconv was assessed by measuring the dilution of the CellTrace Violet. Proliferation was calculated by comparing the median intensity of the CellTrace Violet in the test conditions with the median for undivided cells.

For Th1/Th2 induction, naïve T cells were purified by negative selection using the Naïve CD4+ T cell Isolation Kit (Miltenyi Biotec) according to manufacturer's recommendation. Cells were cultured with 10 μg/ml anti-CD3 and 2.5 μg/ml anti-CD28 (eBioscience) for 6 days, under either Th1 (10 ng/ml 1L-12 and 10 μg/ml anti-IL-4; eBioscience) or Th2 conditions (10 ng/mL IL-4, Biolegend, and 10 μg/mL anti-IFN-γ, eBioscience). Cells were probed for transcription factor expression at 3 days and cytokine expression at 6 days. Transcription factor staining was performed after fixation and permeabilization (Foxp3/Transcription Factor Staining Buffer, eBioscience), with anti-GATA3 (eBioscience, AB_1963600), and anti-T-bet (eBioscience, AB_2686976). Cytokine expression staining was performed after culture with Phorbol-12,13-dibutyrate (1 μg/mL, Tocris Bioscience), Ionomycin calcium salt (1.5 μg/mL Tocris Bioscience) and Brefeldin A (4 μg/mL, Tocris Bioscience), and fixation with Foxp3/Transcription Factor Staining Buffer (Thermofisher). Cells were stained with anti-IL-4 (eBioscience, AB_469494) and anti-IFNγ (eBioscience, AB_469680). Samples were analyzed using a BD FACSCanto II instrument (BD Biosciences) and FlowJo software (Treestar Inc., OR, USA).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.