Impaired lymphocyte function and differentiation in CTPS1-deficient patients result from a hypomorphic homozygous mutation

Emmanuel Martin,1 Norbert Minet,1,2 Anne-Claire Boschat,2,3,4 Sylvia Sanquer,4 Steicy Sobrino,1,2 Christelle Lenoir,1 Jean Pierre de Villartay,2,5 Maria Leite-de-Moraes,6 Capucine Picard,1,2,7,8 Claire Soudais,1,2 Tim Bourne,9 Sophie Hambleton,10 Stephen M. Hughes,11 Robert F. Wynn,12 Tracy A. Briggs,13,14 Genomics England Research Consortium,15 Smita Patel,16 Monica G. Lawrence,17 Alain Fischer,2,8,18,19 Peter D. Arkwright,9,12 and Sylvain Latour1,2


Introduction
T lymphocytes need a very active metabolism to facilitate rapid cellular proliferation and the production of effector molecules in response to antigen challenge (1)(2)(3). The production of purine and pyrimidine nucleotides has a critical role in the synthesis of nucleic acids and lipids, and these pathways are thus necessary for cell cycle progression of activated T lymphocytes (4). Activation of T lymphocytes with mitogens or in Cytidine triphosphate (CTP) synthetase 1 (CTPS1) deficiency is caused by a unique homozygous frameshift splice mutation (c.1692-1G>C, p.T566Dfs26X). CTPS1-deficient patients display severe bacterial and viral infections. CTPS1 is responsible for CTP nucleotide de novo production involved in DNA/RNA synthesis. Herein, we characterized in depth lymphocyte defects associated with CTPS1 deficiency. Immune phenotyping performed in 7 patients showed absence or low numbers of mucosal-associated T cells, invariant NKT cells, memory B cells, and NK cells, whereas other subsets were normal. Proliferation and IL-2 secretion by T cells in response to TCR activation were markedly decreased in all patients, while other T cell effector functions were preserved. The CTPS1 T566Dfs26X mutant protein was found to be hypomorphic, resulting in 80%-90% reduction of protein expression and CTPS activity in cells of patients. Inactivation of CTPS1 in a T cell leukemia fully abolished cell proliferation. Expression of CTPS1 T566Dfs26X failed to restore proliferation of CTPS1deficient leukemia cells to normal, except when forcing its expression to a level comparable to that of WT CTPS1. This indicates that CTPS1 T566Dfs26X retained normal CTPS activity, and thus the loss of function of CTPS1 T566Dfs26X is completely attributable to protein instability. This study supports that CTPS1 represents an attractive therapeutic target to selectively inhibit pathological T cell proliferation, including lymphoma.
response to TCR stimulation leads to increased purine and pyrimidine pools (5,6). Cytidine triphosphate (CTP) synthetase is the limiting nucleotide in cells, and its synthesis depends on 2 pathways: a salvage pathway allowing production of CTP from cytidine derived from nucleic acid degradation and a de novo pathway that relies on CTP synthase (CTPS) activity (7). CTPS activity is encoded by 2 genes, CTPS1 and CTPS2, and their proteins share 80% identity and homology (8)(9)(10). CTPS catalyzes ATP-dependent amination of uridine triphosphate (UTP) to CTP with ammonia (NH 3 ) transfer from hydrolyzed glutamine. In normal T cells, CTPS activity is increased only in the S phase of cell cycle progression, while in cancer cells, including in acute T cell lymphoblastic leukemia (T-ALL), CTPS activity is elevated irrespective of cell cycle progression (11)(12)(13). Thus, CTPS might be considered an attractive therapeutic target for the treatment of T-ALL (14).
Recently, we and others identified immunodeficient patients with a CTPS1 deficiency leading to mainly life-threatening herpes virus infections with varicella-zoster virus (VZV) and EBV. All patients characterized so far have been shown to carry the same homozygous mutation IVS18-1 G>C or NM_001905.3: c.1692-1G>C, p.T566Dfs26X (rs145092287) in the acceptor splice site at introns 17-18, with a likely common ancestor from northwest England (6,(15)(16)(17). In our initial study, we showed that CTPS1 is important for the expansion of activated T cells. In T lymphocytes, CTPS1 expression is upregulated in response to TCR stimulation. The capacity of activated CTPS1-deficient T cells to proliferate is impaired (6). However, in this previous study, analysis of immunological parameters and lymphocyte functions was carried out in only 1 patient available at that time, and no molecular and functional characterization of the mutation was done.
In the present work, a detailed analysis of immune parameters was performed in 7 CTPS1-deficient patients. The consequences of the CTPS1 mutation at the protein and enzymatic levels were also characterized in normal cells and derived cell lines from patients and in a T cell leukemia cell line in which CTPS1 was inactivated by CRISPR/Cas9.

Results
Identification of new patients with CTPS1 deficiency. Since our original report (6), we have screened further patients originating from England and the United States with a clinical phenotype compatible with a CTPS1 deficiency that notably includes severe VZV and EBV infections (Supplemental Table 1; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.133880DS1). Five of them (P10 to P14) were found to carry the same homozygous mutation IVS18-1 G>C in CTPS1 (data not shown). In addition, a further 6 patients have been reported since our original report (6). Clinical manifestations appeared during the first 2 years of life in the majority of patients. Hematopoietic stem cell transplantation (HSCT) was performed in 15 of the 19 patients. Thirteen out of 15 are alive and well at the most recent follow-up (17). Two patients died from severe viral infection (P3.1 and P5) before HSCT could be performed, and 2 (P12 and P13) are alive and have yet to receive transplants. No extrahematopoietic manifestation (not related to their immunodeficiency or HSCT treatment) was reported in any of the 19 patients with a median follow-up extending to 14.6 years (with the oldest patient aged 24 years old). Routine laboratory investigations were carried out in the 5 newly identified patients (Supplemental Table  2 and Supplemental Table 3). Blood cell counts of the different subsets were within, or close to, normal ranges. IgGs were elevated but specific antibodies were low as well as T cell proliferation in response to phytohemagglutinin (PHA) and anti-CD3 (OKT3). These data were similar to those reported previously for CTPS1-deficient patients. (6,(15)(16)(17). Ranges and numbers of the different blood cell populations were compared to established healthy pediatric standards (18)(19)(20)(21).
CD4 + CD25 + CD127 lo FOXP3 + regulatory T cells (Tregs) were detected at normal ranges ( Figure 2A). In contrast, invariant T cell populations invariant natural killer T (iNKT) (CD3 + Vα24 + Vβ11 + CD161 hi or Tet-CD1d + CD161 hi ) and mucosal-associated invariant T (MAIT) (CD3 + Vα7 + CD161 hi or Tet-MR1 + CD161 hi ) cells were absent or markedly reduced ( Figure 2B and Supplemental Figure 1C). These defects did not result from impaired VDJ recombination, as percentages of mainstream T cells using Vα7.2 and Vα24 chains of the TCR (Figure 2B), Vα and Vβ gene segment usage, and PROMIDISα signature (23) were all comparable to controls (Supplemental Figure 2, A and B). In summary, CTPS1 deficiency is characterized by an almost complete absence of blood iNKT cell and MAIT cell subpopulations and decreased NK cells and memory B cells. This phenotypic analysis of 8 patients corroborates well the results of our first report (6).
Impaired TCR-induced IL-2 secretion and proliferation of CTPS1-deficient T lymphocytes. In our first report we showed that CTPS1 was required for T cell proliferation but not for IFN-γ and TNF-α cytokine production in response to TCR activation (6). To extend these observations, PHA-expanded T cell blasts from 5 CTPS1-deficient patients and 10 healthy donors/controls were stimulated for 48 hours with anti-CD3 antibody, anti-CD3/CD28-coated beads, or PMA/ionomycin, and production of IL-2, IL-4, IL-17, IFNγ, TNF-α, and granzyme B was analyzed in culture supernatants. IFNγ, TNF-α, IL-17, granzyme B, and IL-4 release was comparable between patient and control cells ( Figure 3A). In contrast, IL-2 production was significantly reduced. To exclude the possibility that IL-2 was consumed faster by CTPS1-deficient T cells, intracellular IL-2 production from 6 CTPS1-deficient patients and 10 controls was examined after 12 hours of stimulation (Figure 3B). Intracellular IL-2 expression in T cells of patients was reduced following anti-CD3/CD28 or PMA/ ionomycin stimulation. In contrast, intracellular IFN-γ was slightly increased in T cells from patients ( Figure  3B). Degranulation of cytotoxic granules by CD8 + T cells was also normal from 2 examined CTPS1-deficient patients (Supplemental Figure 2C). The proliferation of CTPS1-deficient T lymphocytes was impaired in response to TCR stimulation by an anti-CD3 antibody or anti-CD3/CD28-coated beads ( Figure 3D). This was variably associated with increased activation-induced cell death (Supplemental Figure 2D). Remarkably, CTPS1-deficient T cells stopped dividing after a few divisions and expressed decreased levels of the activation marker CD25 (Figure 3, C and D). This defective proliferation was restored after addition of cytidine in the culture medium ( Figure 3E). To investigate whether impaired autocrine IL-2 secretion was contributing to the defective proliferation, exogenous IL-2 was added to CTPS1-deficient T cells or control T cells treated with a specific inhibitor of CTPS1 (3-deazauridine; 3-DAZ) (24) ( Figure 3F and Supplemental Figure 3A). Addition of IL-2 only slightly restored the proliferation of CTPS1-deficient cells or 3-DAZ-treated control T cells. Thus, we conclude that the defective capacity of CTPS1-deficient T cells to produce IL-2 did not contribute significantly to impaired proliferation. Interestingly, in contrast to CTPS1-deficient cells, 3-DAZ treatment did not impair IL-2 production while CD25 expression was decreased (Supplemental Figure 3, B and C), suggesting that CD25 expression but not IL-2 production is dependent on CTPS activity. These observations support the selective role of CTPS1 in promoting the expansion of human T cells following TCR-mediated activation, while T cell differentiation and most effector functions were spared and normal, with the notable exception of IL-2 production (which was significantly decreased, Figure 3A).
Residual CTPS1 mutant protein expression in CTPS1-deficient patients. We previously showed that the mutation in CTPS1 was associated with no detectable expression of CTPS1 in cell lysates from patients, suggesting that the mutation behaved as an amorphic (null) mutation (6). Using a newly available anti-CTPS1 monoclonal antibody, which gives a more sensitive and specific signal than the polyclonal antibody previously used (Supplemental Figure 4), CTPS1 expression profile and the effect of the CTPS1 mutation were reanalyzed by Western blot and intracellular FACS staining in PBMCs, T cell blasts, lymphoblastoid cell lines (LCLs), primary fibroblasts, transformed/immortalized fibroblasts, and T cells activated with PMA or anti-CD3 antibody ( Figure 4 and Supplemental Figure 5). PMA induced marked CTPS1 expression in T and B lymphocytes from healthy donors ( Figure 4A). Among the different T cell subsets, iNKT and MAIT cells (without stimulation) expressed the highest levels of CTPS1, while the induction of CTPS1 was the strongest in naive T cells following anti-CD3 stimulation ( Figure 4B). Expression of CTPS1 was also found to be high in cell lines such as the leukemia Jurkat cells, the LCLs, and SV-40-transformed cells compared with T cell blasts and primary fibroblasts ( Figure 4C). CTPS2 protein expression was also readily detectable in these cell lines, except in Jurkat T cells.
After stimulation with anti-CD3, CTPS1 protein expression in normal T cells was upregulated from 24 hours and declined after 96 hours ( Figure 4, C and D). CTPS2 protein expression was also strongly increased with a more rapid and transient kinetic, being detectable 12 hours after stimulation but diminishing after 24 hours ( Figure 4D). Of note, CTPS2 detection resolved in 2 protein products in T cells that are differentially expressed after CD3 stimulation, possibly indicative of posttranslational modifications. As expected, CTPS1 and CTPS2 protein expression was preceded by a peak in CTPS1 and CTPS2 mRNA expression, respectively ( Figure 4D). The upregulation of CTPS2 could not compensate for CTPS1 deficiency in patient cells.
ings are in agreement with our previous data showing that CTPS1 cDNA from patients in which exon 18 is lacking encodes a CTPS1 protein with a smaller size when transiently expressed (6). This shorter mutated CTPS1 protein, hereafter denoted as CTPS1Δ18, was also detectable in LCLs and primary fibroblasts derived from patients ( Figure 5, B and C). The CTPS1Δ18 protein was also found at higher levels in activated T cells from 2 heterozygous maternal carriers of the mutation, but not in control cells ( Figure 5A and Supplemental Figure 5). These results suggest that the presence of WT CTPS1 stabilizes the CTPS1Δ18. This hypothesis was supported by transfection of WT CTPS1 cDNA in LCLs derived from 1 patient. In these conditions, expression of WT CTPS1 increased the expression of CTPS1Δ18 (Supplemental Figure  5B). Analysis of CTPS1 mRNA expression by qPCR showed no significant difference between cells from controls and patients (Supplemental Figure 5C), indicating that the mutation had no impact on the expression or the stability of CTPS1 mRNA as initially reported (6). Quantification of CTPS1Δ18 protein in T cell blasts, LCLs, and fibroblasts of patients showed that levels were 15%, 25%, and 12%, respectively, of the WT CTPS1 level in control cells ( Figure 5D), while CTPS2 protein expression in CTPS1-deficient cells from patients showed comparable levels to those of control cells ( Figure 5, A-C). Treatment of activated T cell blasts from 1 patient with the proteasome inhibitor MG-132 enhanced expression of CTPS1Δ18 protein after 4 hours of incubation, while WT CTPS1 expression in control cells was unchanged (Supplemental Figure 5D). After 8 hours of treatment, both WT and mutant CTPS1 protein were decreased because of MG-132 toxicity that induced cell death (data not shown). However, these data indicate that CTPS1Δ18 protein is highly susceptible to proteasome degradation.
CTPS activity was then measured in protein lysates extracted from cells of controls and patients. CTPS activity was found to be markedly decreased but detectable in lysates from T cell blasts, LCLs, and fibroblasts derived from CTPS1-deficient patients as compared with control cells ( Figure 5E). In control T cell blasts, CTPS activity was significantly increased following TCR-CD3 stimulation, while no or a weak increase in CTPS activity was detected in T cell blasts of patients. CTPS activity in CTPS1-deficient T cells was in the order of 20% of the activity found in control cells. These data demonstrate that expression of a mutant CTPS1 protein is detected, albeit at low levels, in CTPS1-deficient cells and is associated with a marked reduction of the CTPS activity.
Mutated CTPS1 protein is unstable but has a preserved CTPS activity. To determine whether the mutant CTPS1 protein was active and accounted for the residual CTPS activity observed in patients' T cells, we tested its capacity to rescue the proliferation of CTPS1-deficient Jurkat cells that do not express CTPS2 ( Figure 4A and data not shown). CTPS1-deficient Jurkat cells (CTPS1-KO Jurkat cells) were obtained by CRISPR/Cas9 technology through the introduction of deleterious mutations targeting exon 10 of CTPS1.   In these cells, CTPS1 expression was abrogated ( Figure 6, A and B) and proliferation profoundly impaired ( Figure 6C). The addition of cytidine to the culture medium, allowing CTP production through the salvage pathway, rescued proliferation of the CTPS1-deficient cells.
CTPS1-KO Jurkat cells were cultured with cytidine for their growth and then transduced with lentivirus to express WT CTPS1 or CTPS1Δ18 alongside GFP as a reporter gene ( Figure 6B). Cells were sorted for GFP expression to purify. Expression of WT CTPS1 and GFP were then analyzed by Western blot and FACS. If cytidine was removed from the culture, expression of GFP and particularly WT CTPS1 rapidly increased in cell lines transduced with CTPS1 ( Figure 6, D and E). In the cell lines transduced with CTPS1Δ18, the expression of the mutated protein was reduced compared with WT CTPS1 and relative to GFP (the expression of which was stronger than in cell lines transduced with CTPS1). These results suggest that in the absence of cytidine supplementation, the CTPS1Δ18-expressing cells that were selected integrated more copies of the lentivirus containing the transgene coding CTPSΔ18 and the GFP reporter (than WT CTPS1-expressing cells).
The proliferation of CTPS1-and CTPS1Δ18-expressing cells was examined in coculture experiments by mixing CTPS1-deficient cells with WT Jurkat cells at a 1:10 ratio ( Figure 6F). When cytidine supplementation was maintained, no selective proliferation advantage was observed, indicating that the proliferative rates of the different cell lines were similar. In contrast, when cytidine was removed, a rapid accumulation of GFP + cells expressing CTPS1 and to a lesser extent CTPS1Δ18 was observed ( Figure 6F). The accumulation of GFP + CTPS1Δ18-expressing cells was slower and plateaued at 30% compared with that of GFP + CTPSΔ-expressing cells that reached 80% of the whole culture. Taken together, these data indicate that CTPS1Δ18 has a diminished capacity to rescue proliferation of CTPS1-KO cells likely because of its instability.
To further analyze CTPS1Δ18 enzymatic activity, CTPS1Δ18-expressing cell lines were maintained in culture without cytidine for 2 months, to select cells expressing higher amounts of CTPS1Δ18. Of note, however, CTPS1Δ18 expression remained lower than that of WT CTPS1 ( Figure 7A). Proliferative capacity of these CTPS1Δ18-expressing cells was then tested in coculture experiments with CTPS1-KO Jurkat cells. In the absence of cytidine, there was a rapid expansion of GFP + cells expressing CTPS1 or CTPS1Δ18 with similar kinetics, indicating that the level of CTPS1Δ18 in these cells was reached following selection, enabling a proliferation rate similar to that of WT CTPS1 ( Figure 7B). Finally, to formally prove that the enzymatic activity of CTPSΔ18 is preserved, CTPS activity was quantified from cell extracts of CTPS1 and CTPS1Δ18-expressing cells. While CTPS activity was almost undetectable in extracts from CTPS1-KO Jurkat cells, in CTPS1-and CTPS1Δ18-expressing cell extracts, CTPS activity was comparable and much higher ( Figure 7C). These results indicate that CTPSΔ18 has enzymatic activity comparable to that of CTPS1.
In conclusion, these results demonstrate that the mutation in CTPS1 causing immunodeficiency in patients behaves as a hypomorphic mutation, characterized by 80%-90% lower expression of an unstable mutated CTPS1 protein (CTPSΔ18), which has close to normal enzymatic activity. This should be taken into account in the analysis of the immunodeficiency phenotype.

Discussion
This study provides an in-depth analysis of the CTPS1 deficiency phenotype and consequences of the mutation on protein expression and function based on data from 19 CTPS1 patients from 13 families. All patients exhibited a combined immunodeficiency phenotype that resulted in severe viral infections that developed early in life, justifying a requirement for HSCT in 15 patients. All patients carried the same mutation in CTPS1, which descends from a common ancestor in northwest England, as suggested by our previous studies (6). Patients from the United States all had ancestors who could be traced back to this geographical area.
The immunological phenotype was remarkably similar between patients. All patients showed a markedly reduced number of iNKT and MAIT cells, as well as reduced NK cells and memory B cells, whereas the other populations, such as naive and memory conventional αβ T cells, γδ T cells, Tregs, and myeloid populations (i.e., monocytes, mDCs, and pDCs), remained normal. CTPS1 was highly expressed in normal iNKT and MAIT cells, and therefore this enzyme may be critically important for the expansion of these innate lymphoid populations or their differentiation or maintenance. Of note, in contrast to conventional T cells, iNKT and MAIT cells undergo a proliferative boost in a late stage of development, when they acquire a memory and activated phenotype (25,26). Until now, concomitant, profound defects in iNKT and MAIT cells have been reported in only 3 other primary immunodeficiencies (PIDs) caused by mutations in RORC, RASGRP1, and GINS1 (27)(28)(29). Interestingly, in RASGRP1 deficiency, CTPS1 expression was found to be strongly diminished in activated T cells, further supporting the concept that CTPS1 is necessary for the development of these invariant T cell populations (28). In RORC-deficient patients (and RORγt-KO mice), the absence of iNKT and MAIT cells is associated with a decrease in thymocytes' survival at the double-positive stage, as shown by the underrepresentation of upstream Vα segments in the TCRα repertoire (27). It may appear surprising that CTPS1 deficiency seems to have little or no effect on the development of conventional T cells, given the massive production of immature thymocytes that occurs in the thymus (30). It may be that the nucleoside salvage pathway compensates for this defect in CTP synthesis (31). Alternatively, residual de novo CTP synthesis may be sufficient to enable normal thymocyte proliferation and differentiation, which may be less intensive than that of activated T cells. Although conventional T cells developed normally in CTPS1-deficient patients, they exhibited poor proliferative capacities when activated by TCR stimulants. T cell effector responses, including degranulation and production of different cytokines, were unaffected, except for IL-2. Interestingly, inhibition of CTPS activity by deazauridine in control T cells completely blocked proliferation but did not impair IL-2 production, suggesting a role of CTPS1 in IL-2 production independent of its enzymatic activity. Memory B cells were lower in patients, but the proportions of switched memory cells were preserved, suggesting a role of CTPS1 in their expansion rather than in their function, like T cells. We have previously shown that, like in T cells, CTPS1 is required for B cell proliferation in response to B cell receptor stimulation (6).
Using newly available anti-CTPS1 and anti-CTPS2 monoclonal antibodies, we also showed that CTPS1 and CTPS2 expression are differentially regulated following TCR activation. CTPS2 expression is more rapidly upregulated after stimulation than CTPS1. CTPS1-deficient T cells can initiate cell division, but this is rapidly limited by the lack of CTP. Initial cell divisions might rely on CTPS2 activity, while CTPS1 takes over after 12-24 hours. Analysis of CTPS1/2 expression in different cell lines suggests a direct correlation between respective level of expression and CTPS activity and rate of proliferation. This is particularly clear in fibroblasts, in which transformation by SV-40 or telomerase ectopic expression is associated with increased CTPS1 expression and CTPS activity. Complemented cells were preselected for high CTPS1 expression by cytidine starvation. Data from 3-5 independent experiments. Each symbol corresponds to the CTPS activity of an independent biological replicate. The horizontal bars represent the median ± SEM. Values were compared 2 by 2 using Mann-Whitney U tests. **P < 0.01; ***P < 0.001.
Importantly, we now demonstrate that the mutation causing CTPS1 deficiency behaves as a hypomorphic mutation. The mutation does not affect RNA stability but results in the expression of an unstable protein lacking the last exon (exon 18). Its expression is estimated to be 10%-20% of the WT CTPS1. In heterozygous carriers, the mutated CTPS1 protein is better expressed because the stability of the mutated CTPS1 is supported through its association with the WT CTPS1, known to form active tetramers (32). In complementation experiments, we showed that the activity of the mutated CTPS1 is similar to that of the WT, thereby accounting for most of the residual CTPS activity detected in cells of patients.
Of note, we recently obtained conditional Ctps1 -/-KO mice that exhibit embryonic lethality (Claire Soudais and Sylvain Latour, unpublished observations). One may thus envisage that in humans too, null mutations in CTPS1 will be not viable and that the residual CTPS1 expression and activity found in patients may be sufficient to allow normal development (including lymphocyte development) and cellular function except for lymphocytes. None of the patients developed extrahematopoietic manifestations, the oldest patient having now reached the age of 20 years (following HSCT). However, we cannot exclude the possibility that CTPS1 and CTPS2 are nonredundant in mice while they may be in humans in nonlymphoid tissues. Alternatively, the pyrimidine salvage pathway may compensate for diminished CTPS activity in humans but not in the mouse.
Finally, we have generated a CTPS1-deficient Jurkat cell line obtained by CRISPR/Cas9 technology. The Jurkat cells have been derived from a patient with an acute T cell leukemia (33). We showed that Jurkat cells are completely dependent on CTPS1 for their proliferation. In the absence of CTP or cytidine supplementation, cells rapidly enter apoptosis (Norbert Minet and Sylvain Latour, unpublished data). We previously proposed that CTPS1 may represent an interesting target for selective therapy of pathologies caused by excessive T cell responses (6). The observation that CTPS1 is mandatory for this T cell leukemia cell proliferation indicates that such inhibitors of CTPS1 could also be used for the treatment of T cell lymphoma/leukemia. We previously showed that CTPS1 is involved in B cell proliferation (6); hence CTPS1 might also represent an interesting target for B cell leukemia/lymphoma. However, CTPS1 may be less important for proliferation in B cells than in T cells. Indeed, we showed that EBV-transformed tumor B cell lines expressed high levels of CTPS2 ( Figure 4C).
In conclusion, the detailed description of the consequences of this hypomorphic CTPS1 mutation in the human immune system sheds light on a critical checkpoint of antigen-triggered adaptive immune responses supporting rapid cell amplification, which is critical for the control of viral infections, notably EBV and VZV.

Methods
Genomic DNA sequencing. Genomic DNA was sequenced for CTPS1 as previously described (6). The diagnosis of patient P10 was made via whole-genome sequencing and confirmed by Sanger sequencing. Whole-genome sequencing data sets were created through the UK 100,000 Genomes Project main/ pilot program, using Illumina X10 sequencing chemistry. Sequencing reads were aligned to build GRCh37 of the human reference genome using Issac (Aligner) and variants within the PID panel version analyzed. All collected sequences were analyzed using 4peaks software (version 1.8; https://nucleobytes. com/4peaks/index.html).
Gene expression analysis. Total RNA was isolated from resting and activated T cell blasts, EBV-transformed LCLs, and fibroblasts of patients and control donors using the RNeasy Mini kit (QIAGEN). The samples were depleted of genomic DNA, and reverse transcription was performed using Superscript III First-Strand Synthesis System (Thermo Fisher Scientific). cDNAs were used as a template to quantify CTPS1, CTPS2, and GAP-DH gene expressions by qPCR using universal reaction mixture (Thermo Fisher Scientific). Gene expression assays were performed with custom and primer Assays-on-Demand probes (CTPS1, Hs01041858; CTPS2, Hs00219845; GAPDH, Hs027558991) from Applied Biosystems, Thermo Fisher Scientific. All probes were labeled with dye (6-carboxy-fluorescein [FAM] or 2′-chloro-7′phenyl-1,4-dichloro-6-carboxy-fluoroscein [VIC]) and Quencher (Black Hole Quencher-1 [BHQ-1]). Real-time qPCRs were performed in triplicate using a Light-Cycler 1536 System (Roche). Expression levels were determined by relative quantification using the comparative threshold cycle method 2 ΔΔCt , in which ΔΔCt is determined as follows: (Ct target gene -Ct reference gene ) target tissue − (Ct target gene -Ct reference gene ) calibrator tissue. The results were normalized to GAPDH gene expression and are presented as the relative value in gene expression.
Cell cultures. PBMCs collected from patients and healthy donors were isolated by Ficoll-Paque density gradient (Lymphoprep, Proteogenix) from blood samples using standard procedures. Expansion of T cell blasts index of proliferation was calculated using FlowJo software and corresponds to the average number of cell divisions per cell including the undivided peak.
Degranulation assay. T cell blasts were washed and stimulated for 3 hours with different concentrations of anti-CD3 antibody in the presence of anti-CD107a and anti-CD107b antibodies (H4A3 and H4B4, BioLegend). After 3 hours of stimulation, CD107a/b membrane protein expression was quantified by flow cytometry using an LSRFortessa cytometer.
CTPS activity assay. The assay method for measuring CTPS activity is based on the transamidation of UTP by cellular extracts containing the enzyme and the simultaneous quantitative determination of UTP and CTP by liquid chromatography coupled to tandem mass spectrometry. Glutamine was added as a source of nitrogen and ATP and guanosine-5′-triphosphate (GTP) as positive regulators of the enzymatic reaction. Cell pellets were sonicated on ice and centrifuged for 30 minutes at 9300 g at 4°C. Protein concentration was determined on supernatants using the BCA protein quantitation kit (Interchim). Thirty micrograms of proteins were used for measuring CTPS activity in reaction mixture containing 1.7 mM EDTA, 13 mM MgCl 2 , 1.3 mM ATP, 0.2 mM GTP, 13 mM glutamine, 1.3 mM phosphoenolpyruvate, 10 mM NaF, 1.3 mM UTP, and 10 μM stable CTP isotope as an internal standard in 10 μL HEPES buffer, 70 mM, at pH 8.0. The reaction mixtures were incubated at 37°C for 90 minutes, and then the enzymatic reaction was stopped by the addition of 2 volumes of HClO 4 . Five microliters of sample was injected onto an Acquity HSS T3 column, 1.8-μm particle size, 2.1 × 100 mm (Waters), connected to an Acquity H-Class ultra-performance liquid chromatography interfaced with a Xevo TQ-S triple-quadrupole mass spectrometer (Waters), both controlled by MassLynx software (Waters). Aqueous mobile phase A was water at pH 6.5, and organic mobile phase B was acetonitrile, both containing 5 mM dibutylamine acetate ion pair. A programmed mobile phase gradient was used for 12-minute runs: 0 minutes, 11% B; 8 minutes, 13% B; 8.1 minutes, 100% B; 9.1 minutes, 100% B; 9.2 minutes, 10% B; and 12 minutes, 11% B. The CTP identification and detection were performed in the electrospray positive ion mode with multiple reaction monitoring mode (MRM). The most intense MRM transitions used for integration and quantification were m/z 484 > 112 and 484 > 208 for CTP and m/z 495 > 119 and 495 > 102 for the internal standard. Quantification was performed using TargetLynx software. The linearity and intraassay variability were determined for the technical validation of this assay and gave a correlation coefficient of the linear regression curves greater than 0.99 and a variation coefficient less than 15%. The threshold of CTPS activity detection corresponds to the mean of activity detected in Jurkat cells deficient for CTPS1 (that did not express CTPS2) from 4 independent experiments (51 ± 5 nmol of CTP/g protein/min).
Full-length CTPS1 and CTPS1Δ18 cDNAs were obtained by real-time PCR from donor and patient blasts, respectively, using the same forward 5′-CGGGATCCCACCATGAAGTATATTCTGGTT-3′ and different reverse 5′-CCGCTCGAGTCAGTCATGATTTATTGA-3′ (CTPS1) or 5′-CCGCTCGAGTTA-AAGAAAGTCTCCAAAGC-3′ (CTPS1Δ18) primers. The cDNA was verified by sequencing and inserted into a bicistronic lentiviral expression vector encoding GFP as a reporter (pLenti7.3/V5-TOPO, Invitrogen, Thermo Fisher Scientific). Viral particles for infection were obtained by cotransfection of the lentiviral vector containing CTPS1 and the plasmids containing Gag-Pol, Rev, and the G protein of the vesicular stomatitis virus into HEK293T cells (ATCC). Viral supernatants were collected every 12 hours on 2 consecutive days, starting 48 hours after transfection, and viral particles were concentrated by ultracentrifugation at 49,000 g for 1.5 hours at 12°C. CTPS1-KO Jurkat cells were infected with viral particles at a minimal titer of 10 7 TU/mL, and 48 hours after infection, cells were sorted on GFP expression and starved in cytidine. The percentages of GFP + cells were determined by flow cytometry.
Jurkat coculture assay. All Jurkat cells were maintained in RPMI plus 10% fetal bovine serum plus 1% penicillin/streptomycin. CTPS1-KO Jurkat cells and CTPS1-KO Jurkat cells infected with a pLenti 7.3 lentivirus vector either empty or containing the CTPS1 or mutant CTPS1Δ18 sequences along with a GFP reporter gene were supplemented with 200 μM of cytidine. After an initial expansion, part of the CTPS1 or Δ18CTPS1-complemented cells were cytidine starved, leading to the loss of uninfected cells (GFP -). The number of GFP + cells was assessed with FlowJo software using Super Enhanced Dmax. This quantification was used to adjust the number of cells needed to obtain a mix of 10% of GFP-expressing cells with 90% of WT Jurkat or CTPS1-KO Jurkat cells. The cells were then seeded at a density of 150 × 10 3 cells/ mL with or without 200 μM of cytidine. The initial and subsequent percentages of GFP + cells during the culture were assessed by flow cytometry.
Statistics. P values were calculated with Mann-Whitney U test with a 2-tailed distribution. A P value less than 0.05 was considered significant.
Study approval. Informed consent was obtained from donors, patients, and families of patients. The study and protocols conform to the 1975 Declaration of Helsinki as well as to local legislation and ethical guidelines from the Comité de Protection des Personnes de l'Ile de France II, Hôpital Necker-Enfants Malades, Paris, France. Study protocols were approved by the Comité de Protection des Personnes de l'Ile de France II.

Author contributions
EM and NM designed and performed experiments and analyzed the data. EM and SL wrote the manuscript. CP, JPV, NM, ACB, S Sobrino, S Sanquer, CL, and MLDM performed experiments and/or analyzed the data. SP, PDA, SMH, RFW, TAB, SH, and MGL identified the patients, provided and analyzed clinical information and data, and reviewed the manuscript draft. CS, TB, and AF interpreted data and reviewed manuscript draft. SL interpreted data, designed and supervised the research, and supervised the writing of the manuscript.