MTHFD2 Is A Metabolic Checkpoint Controlling Effector And Regulatory T Cell Fate And Function
Mar 02, 2022
Graphical abstract
Authors :
Ayaka Sugiura, Gabriela Andrejeva, Kelsey Voss, ..., Ashutosh K. Mangalam, Joshua D. Rabinowitz, Jeffrey C. Rathmell
Correspondence:
jeff.rathmell@vumc.org
In brief ：
Nucleotide synthesis is required to support rapid T cell proliferation. Sugiura et al. show that de novo purine metabolism signals direct T cell differentiation and function and identify MTHFD2 as a metabolic checkpoint and therapeutic target for inflammatory diseases.
Highlights
MTHFD2 is critical for activated CD4 T cells to maintain de novo purine synthesis
Insufficient MTHFD2 promotes Treg cell-like phenotypes and metabolism in Th17 cells d Inhibition of MTHFD2 suppresses mTORC1 signaling and alters histone methylation
MTHFD2 can be targeted to protect against inflammation and autoimmunity in vivo
MTHFD2 is a metabolic checkpoint controlling effector and regulatory T cell fate and function
Ayaka Sugiura,1 Gabriela Andrejeva,1 Kelsey Voss,1 Darren R. Heintzman,1 Xincheng Xu,5 Matthew Z. Madden,1 Xiang Ye,1 Katherine L. Beier,1 Nowrin U. Chowdhury,2 Melissa M. Wolf,2 Arissa C. Young,1 Dalton L. Greenwood,1 Allison E. Sewell,1 Shailesh K. Shahi,3 Samantha N. Freedman,3 Alanna M. Cameron,4 Patrik Foerch,4 Tim Bourne,4 Juan C. Garcia-Canaveras,5 John Karijolich,1 Dawn C. Newcomb,2 Ashutosh K. Mangalam,3 Joshua D. Rabinowitz,5 and Jeffrey C. Rathmell1,6,*
1Vanderbilt Center for Immunobiology, Department of Pathology, Microbiology, and Immunology, Vanderbilt University Medical Center, Nashville, TN 37232, USA
2Department of Medicine, Division of Hematology and Oncology, Vanderbilt University Medical Center, Nashville, TN 37232, USA
3Department of Pathology, University of Iowa, Iowa City, IA 52242, USA
4Sitryx Therapeutics Limited, Magdalen Centre, Oxford Science Park, Oxford, UK
5Department of Chemistry, Ludwig Cancer Research Institute Princeton Branch, Lewis-Sigler Institute for Integrative Genomics, Princeton University, Princeton, NJ 08544, USA6Lead contact
*Correspondence: jeff.rathmell@vumc.org
https://doi.org/10.1016/j.immuni.2021.10.011
SUMMARY
Antigenic stimulation promotes T cell metabolic reprogramming to meet increased biosynthetic, bioenergetic, and signaling demands. We show that the one-carbon (1C) metabolism enzyme methylenetetrahydrofolate dehydrogenase 2 (MTHFD2) regulates de novo purine synthesis and signaling in activated T cells to promote proliferation and inflammatory cytokine production. In pathogenic T helper-17 (Th17) cells, MTHFD2 prevented aberrant upregulation of the transcription factor FoxP3 along with the inappropriate gain of suppressive capacity. MTHFD2 deficiency also promoted regulatory T (Treg) cell differentiation. Mechanistically, MTHFD2 inhibition led to depletion of purine pools, accumulation of purine biosynthetic intermediates, and decreased nutrient sensor mTORC1 signaling. MTHFD2 was also critical to regulate DNA and histone methylation in Th17 cells. Importantly, MTHFD2 deficiency reduced disease severity in multiple in vivo inflammatory disease models. MTHFD2 is thus a metabolic checkpoint to integrate purine metabolism with pathogenic effector cell signaling and is a potential therapeutic target within 1C metabolism pathways.
INTRODUCTION
Effective mobilization of the adaptive immune response requires robust activation of CD4+ T cells to undergo rapid cell growth and proliferation. This involves metabolic reprogramming, driven by the nutrient sensor mTORC1, from a catabolic resting state to an anabolic growth state with increased biosynthetic, bioenergetic, and signaling demands. Depending on the cytokine milieu, CD4+ T cells differentiate into effector as well as regulatory subsets that have distinct metabolic programs driving their effector functions (Bantug et al., 2018; Buck et al., 2015). The balance of these subsets is critical for normal immunity while preventing inflammatory and autoimmune diseases. For example, multiple sclerosis (MS) is characterized by increased interleukin-17 (IL- 17)-producing T helper-17 (Th17) cells and decreased or ineffective suppressive regulatory T (Treg) cells (Dendrou et al., 2015). Targeting the specific metabolic programs of these T cell subsets offers an alternative approach to immunotherapy.
Cell metabolism-based therapy began in 1948 when antifolate chemotherapeutic agents were shown to be effective in children with acute lymphoblastic leukemia (Farber et al., 1948). Folate contributes to one-carbon (1C) metabolism for purine biosynthesis, generation of methyl donors, and maintenance of cellular redox balance (Ducker and Rabinowitz, 2017; Yang and Vousden, 2016). Many drugs, including methotrexate (MTX), fluorouracil (5-FU), and mercaptopurine (6-MP), have since been developed to target this pathway. Rapidly proliferating cells, including cancer cells and activated T cells, share dependencies on these pathways to synthesize DNA and RNA. MTX remains commonly used to treat autoimmune diseases like rheumatoid arthritis (RA) (Brown et al., 2016). However, because of the broad expression of the targeted enzymes, these therapeutic agents are associated with common and potentially severe adverse effects. Identifying metabolic enzyme targets that are selectively important in cell populations of interest could lead to the development of safer and more efficacious immunotherapies.
1C metabolism includes the folate and methionine cycles with transfer of single carbon units. Serine serves as a 1C donor to activate tetrahydrofolate (THF) to 5,10-methyleneTHF, with concomitant glycine production. 5,10-methyleneTHF can then be oxidized using NAD(P) to generate the purine precursor 10- formylTHF. Alternatively, 10-formylTHF can be synthesized from formate and THF in an ATP-dependent manner. In the cytosolic pathway, 10-fomylTHF production is mediated by methyleneTHF dehydrogenase 1 (MTHFD1). Loss of MTHFD1 starves cells for cytosolic 10-formylTHF, ablating purine biosynthetic capacity, and mutations in MTHFD1 cause severe combined immunodeficiency (SCID) (Field et al., 2015). The mitochondrial pathway relies on mitochondrial MTHFD2 and MTHFD1-like (MTHFD1L).
Figure 1. 1C metabolism and MTHFD2 are upregulated in activated CD4+ T cells and in the context of EAE
(A) Change in nucleotide metabolism species in in-vitro-differentiated CD4+ T cell subsets compared with naive cells measured by mass spectrometry (n = 3 biological replicates).
(B) Workflow for targeted CRISPR screening in the in vivo lung inflammation model.
(C) Change in gRNA abundance from 1C metabolism-targeted in vivo CRISPR screening in primary CD4+ T cells (statistical analysis performed by MAGeCK, n = 3 biological replicates).
(D) mRNA expression of genes identified in (C) during T cell development and activation (data from the ImmGen RNA-seq data browser).
(E) Protein expression of genes identified in (C) in naive and activated CD4+ T cells (data from the Immunological Proteome Resource [ImmPRes]).
(F) mRNA expression of genes identified in (C) in whole blood of individuals with the indicated inflammatory disorders relative to healthy control individuals (Aune et al., 2017). RA-MTX, RA under MTX treatment; SLE, systemic lupus erythematosus; MS-Tx naive, treatment-naive MS at time of diagnosis; MS-established, MS under treatment and in disease remission; n = 3–8 donors.
(G) IHC showing CD3 and MTHFD2 staining in the cauda equina segment of the spinal cord of mice with symptomatic EAE at 2003 (top; scale bar, 100 mm) and 4003 (bottom; scale bar, 50 mm) magnification (data are representative of two independent experiments).
(H and I) Relative (H) mean fluorescence intensity (MFI) and (I) mRNA expression of MTHFD2 in CD4+ T cells from the spleen and spinal cord of mice with symptomatic EAE and spleen of control mice (mean ± SD, one-way ANOVA, data are representative of two independent experiments with 6 total biological replicates).
(J) MTHFD2 mRNA expression in undifferentiated CD4+ T cells 0, 5, and 24 h after activation with anti-CD3 and anti-CD28 antibodies (mean ± SD, one-way ANOVA, n = 3 biological replicates).
(K) Relative MTHFD2 MFI over 5 days after activation in CD4+ T cell subsets, normalized to resting cells (mean ± SD, n = 3 biological replicates). See also Figure S1. Statistically significant results are labeled (* p < 0.05, ** p < 0.01, *** p < 0.001, **** p % 0.001).
MTHFD2 is one of the most highly induced and overexpressed genes in all tumors (Nilsson et al., 2014). Although broadly upregulated during embryogenesis, MTHFD2 has little to no expression in most adult tissues (Nilsson et al., 2014). In anti-cancer therapy (Zhu and Leung, 2020), MTHFD2 inhibition may lead to increased oxidative stress (Ju et al., 2019; Wan et al., 2020), glycine dependency (Koufaris et al., 2016), and insufficient purine synthesis (Ben-Sahra et al., 2016; Pikman et al., 2016). MTHFD2 deficiency can also suppress mTORC1 activity through multiple mechanisms, including through guanine depletion and subsequent inhibition of the mTORC1-activating GTPase Rheb (Emmanuel et al., 2017). MTHFD2 deficiency also causes accumulation of purine synthesis pathway intermediates, including the adenosine monophosphate (AMP) analog 5-aminoimidazole carboxamide ribonucleotide (AICAR) (Ducker et al., 2016), which can activate AMP-activated protein kinase (AMPK) to inhibit mTORC1 (Su et al., 2019).
Because 1C metabolism integrates multiple nutrient inputs and is therapeutically tractable, we examined the role of 1Cmetabolism in CD4+ T effector (Teff) and Treg cell subsets. Through unbiased in vivo CRISPR-Cas9-based targeted screening in primary murine T cells, we identified MTHFD2 as a hit that was also upregulated consistently in individuals with inflammatory disorders. Notably, MTHFD2 deficiency impaired Teff cell proliferation and function. MTHFD2 deficiency also promoted aberrant FoxP3 expression and suppressive activity inTh17 cells while promoting Treg cell differentiation. These effects were associated with accumulation of AICAR, reduced purine concentrations, downregulation of mTORC1 signaling, increased mitochondrial metabolism, and altered DNA and histone methylation. In vivo, targeting MTHFD2 protected against multiple inflammatory disease models. These data show thatMTHFD2 serves as a metabolic checkpoint in Th17 and Treg cells and highlight the potential of this enzyme as a target for anti-inflammatory immunotherapy.
RESULTS
1C metabolism and purine synthesis is differentially active in CD4+ T cell subsets
We hypothesized that nucleotide synthesis may differ across CD4+ T cell subsets. CD4+ T cells were differentiated in vitro into Th1, Th17, and Treg cells and collected after 72 h for mass spectrometry. Nucleotide species amounts were generally elevated in all subsets compared with naive cells (Figure 1A). Notably, purine synthesis intermediates were differentially abundant, with the highest accumulation of glycinamide ribonucleotide (GAR), 1-(phosphoribosyl) imidazole carboxamide (SAICAR), and AICAR in Th17 cells (Figure S1A).
To test the dependence of CD4+ T cells on enzymes in 1C metabolism, we performed an in vivo CRISPR-Cas9-based screen in primary CD4+ T cells (Figure 1B). A custom guide RNA (gRNA) library was constructed to target enzymes in 1C metabolism, along with tuberous sclerosis complex 2 (TSC2) as a positive control and non-targeting negative controls (NTCs). CD4+ T cells isolated from ovalbumin (OVA)-specific T cell receptor (TCR) transgenic OT-II Cas9 double-transgenic mice were transduced with this library and intravenously transferred into Rag1__ /__ hosts. Mice were then immunized with intranasal OVA to induce lung inflammation. T cells recovered from the lungs of these mice were sequenced, and enrichment or depletion of gRNAs was established relative to input frequencies. TSC2 gRNA was enriched, supporting the inhibitory role of this protein in T cells. In contrast, PPAT, AHCY, MAT2A and MTHFD1 gRNAs were depleted signifificantly, and GART, DHFR, DNMT1, MTRR, MTR, SHMT2, and MTHFD2 gRNAs were reduced to a lesser extent, indicating that these genes contribute to T cell proliferation and inflammatory function in vivo (Figures 1C and S1B).
We next examined the expression profile of the genes identified from the screen. Although all were upregulated coordinately in developing thymocytes, MTHFD2 mRNA and protein were induced most strongly in stimulated peripheral T cells (Figures 1D and 1E). Moreover, in a separately published RNA sequencing (RNA-seq) dataset from whole blood of individuals with a variety of inflammatory and autoimmune diseases (Aune et al., 2017), MTHFD2 was overexpressed consistently across multiple conditions, including ulcerative colitis, Crohn’s disease, celiac disease, RA, systemic lupus erythematosus (SLE), psoriasis, psoriatic arthritis, Sjo¨ gren’s syndrome, and MS (Figure 1F). Notably, individuals with recently diagnosed MS had significantly elevated MTHFD2 expression relative to healthy donors or individuals with MS undergoing therapy with disease remission (Figure S1C). Thus, CD4+ T cell proliferation and survival are dependent on certain 1C metabolism genes, and expression of these genes differs through the stages of T cell development, activation, and differentiation as well as in the setting of pathological inflammation.
Figure 2. MTHFD2 deficiency impairs CD4+ T cell proliferation and function
(A–D) Proliferation measured by (A) CellTrace Violet (CTV) dilution, (B) viability, (C) CD25 expression, and (D) TF expression in CD4+ T cell subsets treated with the MTHFD2i for 72 h after activation (mean ± SD, one-way ANOVA, data are representative of three independent experiments with 9 total biological replicates).
MTHFD2 is highly upregulated in CNS-infiltrating CD4+ T cells in experimental autoimmune encephalomyelitis (EAE) and in activated CD4+ T cells in vitro
Regulation of MTHFD2 has not been well described in CD4+ T cells. To measure the expression of MTHFD2 in inflammatory lesions in vivo, CD4+ T cells from myelin-specific TCR transgenic 2D2 mice were activated and adoptively transferred into Rag1__ /__ mice to induce EAE. After mice began exhibiting hindleg paralysis, the cauda equina segment of the spinal cord was analyzed by immunohistochemistry (IHC). Mice that received 2D2 T cells showed enrichment of MTHFD2 and CD3+ cells that was absent in the control (Figure 1G). To directly determine whether CD4+ T cells upregulated MTHFD2 in inflamed lesions, T cells from the spleen and spinal cord of mice subjected to a myelin oligodendrocyte glycoprotein (MOG) and pertussis toxin (PTX)-induced model of EAE were collected for analysis. CNS-infiltrating CD4+ T cells overexpressed MTHFD2 compared with matched splenic CD4+ T cells from control and EAE mice, as measured by flow cytometry (Figure 1H) and qRT-PCR (Figure 1I). The kinetics of MTHFD2 expression in CD4+ T cells in vitro were measured next. Robust upregulation of MTHFD2 mRNA was detected after 5 h of stimulation and began to decrease by 24 h (Figure 1J), whereas differentiated CD4+ T cell subsets attained the highest expression of MTHFD2 protein 48 h after activation (Figure 1K). These data show that MTHFD2 is upregulated in CD4+ T cells with activation in vitro and at the site of inflammation in vivo.
CD4+ T cell subsets differentially require MTHFD2 for activation, proliferation, survival, and cytokine production
We next tested the role of MTHFD2 in CD4+ T cell subset activation, differentiation, proliferation, and function. CD4+ T cells were activated in vitro in the presence of cytokines for optimal Th1, Th17, and Treg cell differentiation and vehicle or an MTHFD2 inhibitor (MTHFD2i; DS18561882) (Kawai et al., 2019). After 72 h, all subsets had reduced proliferation (Figure 2A) and numbers of live cycling cells (Figures S2A and S2B). Viability was reduced in Th1 and Th17 cells (Figure 2B), and activation, as measured by CD25 expression, was reduced in all subsets (Figure 2C). Lineage-characterizing transcription factor (TF) expression was reduced in Th1 cells (T-bet+ ) but unchanged in Th17 cells (RORγt+ ) and changed minimally in Treg cells (FoxP3+ ) (Figure 1D). To measure Teff cell function, cells were re-stimulated with 12-myristate 13-acetate (PMA) and ionomycin. Signifificantly fewer Th1 cells expressed interferon-g (IFNg), and fewer Th17 cells expressed IL-17 when differentiated in the presence of the MTHFD2i (Figure 2E). This was not due to impaired initial activation because these phenotypes were maintained when drug exposure was delayed until 24 h after activation, although lineage-characterizing TF expression was unaltered in this setting (Figures S2C–S2F). MTHFD2 is thus required for maximal CD4+ T cell activation, proliferation, survival, and cytokine production.
To genetically validate these pharmacological effects, we generated a Mthfd2fl/fl mouse strain and crossed it with Cd4- cre transgenic mice to achieve conditional genetic deletion. As anticipated, activated CD4+ T cells from Mthfd2fl/fl Cd4-cre+ (CD4ΔMthfd2) mice had lower MTHFD2 expression compared with cells from Mthfd2fl/fl Cd4-cre+ (wild-type [WT]) littermates (Figure 2F). At baseline, CD4DMthfd2 mice had slightly fewer CD4+ T cells in the spleen compared with the WT (Figure 2G). Upon in vitro activation and differentiation, cell numbers and viability of CD4ΔMthfd2 Th17 cells and, to a lesser extent, Treg cells were signifificantly lower (Figure 2H). In addition, CD25 expression was decreased in Th17 and Treg cells and trended lower in Th1 cells (Figure 2I). Finally, CD4ΔMthfd2 Th1 and Th17 cells had lower expression of T-bet and RORγt, respectively. FoxP3 in Treg cells, however, was unchanged (Figure 2J). Despite changes in TF expression, Teff cells that did differentiate produced normal amounts of cytokines (Figure 2K). Differences in results between the pharmacological and genetic approaches may be attributed to the timing and completeness of enzyme function loss during development and potential compensatory mechanisms.
The MTHFD2i induces FoxP3 expression in Th17 cells and enhances Treg cell differentiation
Given the dependence of Teff cells on MTHFD2, we tested MTHFD2i regulation of FoxP3 and Treg cells. Notably, MTHFD2i treatment-induced aberrant upregulation of FoxP3 in Th1 and more prominently in Th17 cells in a dose-dependent manner (Figure 3A). This also occurred when treatment was delayed until 24 h after activation (Figure 3B). A similar trend was observed in the CD4DMthfd2 Th17 cells (Figure 3C). This FoxP3 upregulation was functional because MTHFD2i-treated Th17 cells gained suppressive capacity when co-cultured with CD8+ T cells (Figure 3D). The MTHFD2i can thus promote FoxP3 and Treg cell-like phenotypes in Th17 cells. The MTHFD2i also enhanced differentiation of induced Treg cells. Treg cells were differentiated with a range of transforming growth factor b (TGF-b) concentrations in the presence of the MTHFD2i. At every concentration tested except the highest, FoxP3 expression increased with MTHFD2i treatment (Figures 3E and 3F). This phenotype was also maintained with delayed MTHFD2i treatment (Figure 3G).
Figure 3. MTHFD2 deficiency induces FoxP3 expression and promotes a shift toward oxidative phosphorylation
(A) FoxP3 expression in MTHFD2i-treated Th1 and Th17 cells 72 h after activation (mean ± SD, one-way ANOVA, data are representative of three independent experiments with 9 total biological replicates).
(B) FoxP3 expression in Th1 and Th17 cells activated for 24 h and then treated with vehicle or 500 nM MTHFD2i for 48 h (mean ± SD, unpaired t test, data are representative of three independent experiments with 9 total biological replicates).
(C) FoxP3 expression in Th1 and Th17 cells from WT and CD4DMthfd2 littermates 72 h after activation (mean ± SD, unpaired t test, data are representative of three independent experiments with 9 total biological replicates).
(D) Suppression assay measuring the aberrant suppressive capacity of MTHFD2i-treated Th17 cells. Pre-treated Th17 cells were co-cultured with CTV-stained CD8+ T cells to measure proliferation upon activation with anti-CD3 and anti-CD28 antibodies (mean ± SD, one-way ANOVA, n = 3 biological replicates).
(E) FoxP3 expression in Treg cells differentiated with a range of TGF-b concentrations and treated with the MTHFD2i (mean ± SD, unpaired t test, data are representative of three independent experiments with 9 total biological replicates).
(F) Flow cytometry plots for data tabulated in (E).
(G) FoxP3 expression in Treg cells activated and differentiated with low concentrations of TGF-b for 24 h and then treated with the MTHFD2i for 48h (mean ± SD, unpaired t test, data are representative of three independent experiments with 9 total biological replicates).
(H) Seahorse XF Cell Mito stress test performed on MTHFD2i-treated Th17 and Treg cells (mean ± SD, n = 3 biological replicates).(I) Basal ECAR, basal OCR, max OCR, and basal OCR/ECAR ratio measured in (H) (mean ± SD, unpaired t test, data are representative of two independentexperiments with 6 total biological replicates).Statistically signifificant results are labeled (* p < 0.05, ** p < 0.01, *** p < 0.001, **** p % 0.001).
Treg cells have a greater mitochondrial metabolism than Th17 cells (Michalek et al., 2011). To determine the metabolic consequences of MTHFD2i treatment, Th17 and Treg cells were assessed by extracellular flux analyses. The MTHFD2i increased the basal and maximal oxygen consumption rate (OCR) of Th17 cells to indicate a shift toward mitochondrial respiration (Figure 3H). Conversely, Th17 and Treg cells had a decreased basal extracellular acidification rate (ECAR) with MTHFD2i treatment, suggesting reduced glycolysis (Figure 3I). The basal ratio of OCR to ECAR was increased signifificantly in Th17 and Treg cells because the MTHFD2i shifted the metabolic program from glycolysis toward oxidative phosphorylation. These data show that the MTHFD2i enhances FoxP3 expression in Th17 cells and Treg cells differentiated under low TGF-b conditions and shifts Th17 cells to a more Treg cell-like metabolism.
MTHFD2 deficiency in human CD4+ T cells decreases Th17 cell proliferation and viability while increasing FOXP3 expression and dampening mTORC1 activity
To test the translatability of the above findings to human cells, CD4+ T cells isolated from peripheral blood of healthy donors were differentiated in vitro into Th1, Th17, and Treg cells and treated with vehicle or MTHFD2i from the time of activation. Proliferation was decreased signifificantly in MTHFD2i-treated Th17 and Treg cells (Figure 4A), and Th17 cells also exhibited increased apoptosis (Figure 4B). Decreased proliferation in Th17 cells was recapitulated with a genetic approach using small interfering RNA (siRNA) (Figures 4C and S2G). Similar to murine cells, FOXP3 expression was higher in MTHFD2i-treated human Th17 cells and trended higher in MTHFD2i-treated human Treg cells (Figures 4D and S2H). Given that the mTORC1-AMPK axis is critical for Th17 and Treg cell differentiation, phosphorylation of the mTORC1 target ribosomal protein S6 (phospho-S6) was also measured. This showed decreased mTORC1 activity in Th17 cells with MTHFD2i or siMTHFD2 (Figure 4E). Thus, the loss of proliferative capacity and induction of FOXP3 expression in Th17 and Treg cells can be recapitulated in primary human CD4+ T cells.
Figure 4. MTHFD2 deficiency impairs proliferation and induces FOXP3 expression in human Th17 cells
(A) Proliferation measured by CTV dilution in human CD4+ T cell subsets treated with vehicle or 2 mM MTHFD2i for 72 h after activation with anti-CD3 and anti-CD28 antibodies (paired t test, data are representative of three independent experiments with 5–6 healthy donors).
(B) Apoptosis measured by Annexin V in human CD4+ T cells treated with the MTHFD2i for 72 h (paired t test, data are representative of three independent experiments with 5–6 healthy donors).
(C) Ki-67 expression in human CD4+ T cells transduced with NTC or siMTHFD2 (paired t test, data are representative of three independent experiments with 4 healthy patient donors).
(D) FOXP3 expression in human CD4+ T cells treated with MTHFD2i for 72 h (paired t test, data are representative of three independent experiments with 6–10 healthy donors).
(E) Change in phospho-S6 expression in human CD4+ T cells treated with MTHFD2i for 72 h or transduced with siMTHFD2 (mean ± SD, one sample t test, data representative of three independent experiments with 4–6 healthy donors). Statistically signifificant results are labeled (* p < 0.05, ** p < 0.01, *** p < 0.001, **** p % 0.001).
T cells rely on MTHFD2 enzymatic function, and phenotypes can be rescued by product
MTHFD2 is required for maintenance of the mitochondrial formate pool for purine synthesis (Ma et al., 2017) but may also play non-enzymatic roles. Supporting essential enzymatic activity, MTHFD2i treatment led to compensatory uptake of formate from the medium in CD4+ T cell subsets (Figure 5A). The 1C unit used for formate synthesis is provided by conversion of serine to glycine, and intracellular serine accumulated whereas glycine was depleted with MTHFD2i treatment in all subsets (Figure 5B). Methionine concentrations were also increased slightly, suggesting inhibition of the methionine cycle. As expected, given that the serine-glycine conversion step is upstream of MTHFD2, these changes were not rescued by provision of formate in the medium.
Figure 5. MTHFD2i effects on T cell subsets are rescued by exogenous formate
(A) Formate uptake measured by 1 H-magnetic resonance spectroscopy (MRS) in CD4+ T cell subsets treated with vehicle or 500 nM MTHFD2i for 72 h after activation (mean ± SD, unpaired t test, data are representative of two independent experiments with 6 total biological replicates).
(B) Change in serine, glycine, and methionine concentrations in CD4+ T cells treated with 500 nM MTHFD2i or 500 nM MTHFD2i + 1 mM formate for 4–6 h compared with vehicle, measured by mass spectrometry (mean ± SD, n = 3 biological replicates).
(C) Live cell count, viability, and CD25 expression in MTHFD2i-treated CD4+ T cells rescued with 1 mM formate or 60 mM adenine and guanine purine solution for 72 h after activation (mean ± SD, one-way ANOVA, data are representative of three independent experiments with 9 total biological replicates).
(D) Live cell count, viability, and CD25 expression in CD4+ T cells from WT or CD4DMthfd2 littermates rescued with formate or purines for 72 h after activation (mean ± SD, one-way ANOVA, n = 3 biological replicates).
We next tested whether addition of formate or an adenine and guanine purine solution could rescue deficient enzymatic function and reverse downstream effects of MTHFD2 deficiency. Indeed, 1 mM formate restored proliferation, viability, and activation in MTHFD2i-treated cells (Figure 5C). Purine supplementation rescued proliferation in all subsets and activation in Th1 and Th17 cells but not viability in all subsets and activation in Treg cells. Formate also rescued proliferation, viability, and activation in CD4ΔMthfd2 T cells (Figure 5D). Impaired cytokine production in MTHFD2i-treated Th1 and Th17 cells was reversed to vehicle amounts with formate or purine addition (Figure 5E). Notably, aberrant upregulation of FoxP3 in Th17 cells was rescued by purine addition but not formate, whereas formate and purine addition reduced FoxP3 expression to baseline in Treg cells differentiated with low concentrations of TGF-b (Figure 5F). MTHFD2i phenotypes appear to be dependent on MTHFD2 maintenance of the mitochondrial formate or purine pools.
MTHFD2 regulates mTORC1 activity and DNA and histone methylation in CD4+ T cells
In immortalized cancer cells, MTHFD2 deficiency can lead to accumulation of the purine synthesis intermediates GAR, SAICAR, and AICAR, which are upstream of 10-formylTHF-mediated formylation steps (Ducker et al., 2016). CD4+ T cell subsets were cultured for 72 h before exposure to MTHFD2i for 4–6 h. This short time frame was chosen to avoid any secondary effects, including compensatory changes. Transient MTHFD2i treatment resulted in accumulation of GAR, SAICAR, and AICAR in all subsets (Figure 6A). This effect was fully rescued by addition of 1 mM formate, supporting an enzymatic role of MTHFD2. Nucleotides and nucleobases, including guanine, were also depleted by the MTHFD2i and rescued by formate in Th17 and Treg cells (Figure 6B).
Figure 6. MTHFD2i results in accumulation of purine synthesis intermediates dampened mTORC1 activity and altered DNA and histone methylation
(A) Change in GAR, SAICAR, and AICAR concentrations in CD4+ T cells treated with 500 nM MTHFD2i or 500 nM MTHFD2i + 1 mM formate for 4–6 h relative to vehicle, measured by mass spectrometry (mean ± SD, one-way ANOVA, n = 3 biological replicates).
(B) Change in nucleotide metabolism species in Th17 and Treg cells treated with MTHFD2i with or without formate for 4–6 h relative to vehicle, measured by mass spectrometry (n = 3 biological replicates).
(C) Immunoblot of phospho-S6, S6, Rheb, phospho-ACC, ACC, HIF-1a, and b-actin in CD4+ T cells treated with MTHFD2i with or without formate for 72 h after activation (data representative of three independent experiments with n = 3 biological replicates).
(D) Change in TCA cycle metabolites in Th17 cells treated with MTHFD2i with or without formate for 6 h relative to vehicle, measured by mass spectrometry (mean ± SD, one-way ANOVA, n = 3 biological replicates).
(E) Heatmaps of H3K27me3 to IgG control ratio in ±10-kb regions around promoters of UCSC known genes in Th17 and Treg cells treated with MTHFD2i, measured by CUT&RUN (data pooled from 3 biological replicates).
(F) Heatmap of median DNA methylation frequency of Foxp3 locus by CpG site in Th1 and Th17 cells treated with vehicle or 500 nM MTHFD2i and in nTreg cells from WT or CD4DMthfd2 littermates (Mood’s test, n = 3 biological replicates).
(G) Tabulation of (F) by proximal promoter, distal promoter, and TSDR (paired t test). See also Figures S3 and S4. Statistically significant results are labeled (* p < 0.05, ** p < 0.01, *** p < 0.001, **** p % 0.001).
AICAR and nucleotides are established regulators of the mTORC1-AMPK axis (Emmanuel et al., 2017; Kim et al., 2016). Indeed, phospho-S6 expression was decreased with MTHFD2i treatment in all subsets and fully restored by addition of formate (Figures 6C and S3A). The GTP-binding protein Rheb, an obligate activator of mTORC1, can be sensitive to intracellular guanine availability (Emmanuel et al., 2017), and Rheb expression was dampened with MTHFD2i treatment (Figures 6C and S3B). AICAR is an adenosine analog and AMPK activator (Rae and Mairs, 2019; Su et al., 2019). However, 72 h after activation, phosphorylation of AMPK or its targets, acetyl-coenzyme A(CoA) carboxylase (phospho-ACC) and Unc-51-like autophagy activating kinase (phospho-ULK1), did not change with MTHFD2i treatment (Figures 6C, S3C, and S3D). mTORC1 activity promotes Th17 and attenuates Treg cell differentiation in part through induction of HIF-1a-dependent glycolytic metabolism (Shi et al., 2011). Consistent with this, HIF-1a expression was decreased in MTHFD2i-treated and CD4DMthfd2 Th17 cells (Figures 6C, S3E, and S3F). These changes in mTORC1 activity were not associated with any deficits in TCR signaling, as measured by Nur77 expression with and without re-stimulation (Figure S3G). Signal transducer and activator of transcription 3 (STAT3) activity, which is critical for Th17 cell development, was unchanged or increased modestly with MTHFD2i treatment (Figure S3H). Decreased mTORC1 activity and a shift from glycolysis toward oxidative phosphorylation with MTHFD2i treatment suggested changes in relative abundance of tricarboxylic acid (TCA) cycle metabolites. Succinate and fumarate concentrations were reduced signifificantly by the MTHFD2i but rescued by formate (Figure 6D).
TCA metabolites are known inhibitors of DNA and histone demethylases (Su et al., 2016), and depletion of succinate and fumarate pools may support a hypomethylated state. H3K27 tri-methylation (H3K27me3) across the genome was measured by CUT&RUN (cleavage under targets and release using nuclease) sequencing (Skene and Henikoff, 2017). Indeed, MTHFD2i-treated Th17 cells had reduced H3K27me3 compared with vehicle (Figures 6E and S4A). DNA methylation at the Foxp3 locus was also measured by bisulfate sequencing to determine whether differential methylation contributed to MTHFD2iinduced FoxP3 upregulation. Although several CpG loci differed in methylation frequency between vehicle and MTHFD2i conditions, there was no discernable pattern in the Treg cell-specific demethylated region (TSDR). There were also no differences between nTreg cells (CD4+ CD25+ ) isolated from WT and CD4DMthfd2 mice (Figure 6F), although the purity of each enriched population was low (Figure S4B). However, there was a significant decrease in DNA methylation frequency over the Foxp3 proximal promoter region in MTHFD2i-treated Th17 cells compared with vehicle (Figure 6G), indicating a possible enhancement of transcription with MTHFD2i.
In addition to maintaining the intracellular formate pool, MTHFD2 activity contributes to the cellular redox balance through conversion of NAD+ to NADH and NADP+ to NADPH (Shin et al., 2017) and support of the transsulfuration pathway and subsequent glutathione synthesis. MTHFD2i-treated Th17 cells had decreased concentrations of NAD+ , NADH, NADP+, and NADPH, which was rescued with formate (Figure S4C). However, NAD+ /NADH and NADP+ /NADPH ratios were unchanged across conditions. This, combined with the lack of change in cellular reactive oxygen species (ROS) and mitochondrial superoxide (Figure S4D), suggest that changes in the redox state are unlikely to be a major contributor to the MTHFD2i phenotype.
MTHFD2 deficiency in vivo reduces the severity of multiple inflammatory diseases
The efficacy of MTHFD2 as a therapeutic target was next tested in an in vivo T cell-dependent delayed-type hypersensitivity (DTH) model. Mice were immunized and challenged on the ear with keyhole limpet hemocyanin (KLH) to induce local inflammation and treated with oral vehicle or MTHFD2i. Treatment showed no overt toxicity because animals maintained their weight (Figure S5A). Importantly, ear thickness and weight were increased in control DTH mice but not in inhibitor-treated animals, indicating protection from inflammation upon MTHFD2i treatment (Figures 7A, 7B, and S5B). KLH-specific immunoglobulin G (IgG) was also reduced at the highest dose of inhibitor, indicating potential effects extending to B cell function (Figure 7C).
Figure 7. Targeting MTHFD2 in vivo reduces disease severity in DTH, EAE, and IBD models
(A and B) Ear (A) relative thickness and (B) biopsy weight 10 days after KLH with CFA immunization and 3 days after KLH challenge to induce DTH (one-way ANOVA). Mice were treated twice daily with oral 0, 100, or 300 mg/kg MTHFD2i (mean ± SEM, n = 8 biological replicates).
(C) KLH-specific IgG concentrations on day 10 of KLH-induced DTH (mean ± SEM, Kruskal-Wallis test, n = 8 biological replicates).
(D) Average clinical score over time in WT and CD4DMthfd2 littermates immunized with MOG with CFA and PTX to induce EAE (mean ± SEM, multiple Mann-Whitney tests, data are representative of two independent experiments, each with 5 biological replicates).
(E) Cell count and frequency of CD4+ T cells in the spinal cord of EAE mice at peak disease severity (mean ± SEM, unpaired t test, data are representative of two independent experiments with 9 total biological replicates).
(F) CD25 and CD44 expression in CD4+ T cells from the spinal cord of EAE mice (mean ± SEM, unpaired t test, n = 4 biological replicates).
(G and H) Cell count of (G) T-bet+ , RORgt+ , FoxP3+ , FoxP3+ /T-bet+ ratio, FoxP3+ /RORgt+ ratio and (H) IL-17+ , IFNg+ , and IL-17+ IFNg+ CD4+ T cells from the spinal cord of EAE mice (mean ± SEM, unpaired t test, data are representative of two independent experiments with 9 total biological replicates).
(I) H&E, anti-CD3 IHC, and Luxol Fast Blue (LFB) staining for myelin in the spinal cord of a WT mouse with no immunization and WT and CD4DMthfd2 littermates with EAE at peak disease (2003 magnification, representative of 5 biological replicates).
(J) Change in body weight over time of Rag1__/__ mice injected i.p. with WT or CD4DMthfd2 naive CD4+ T cells to induce IBD colitis (mean ± SEM, multiple t tests, n = 8 biological replicates).
(K and L) Cell count of (K) total CD4+ T cells and (L) T-bet+ , RORgt+ , and FoxP3+ CD4+ T cells from MLNs of IBD mice (mean ± SEM, unpaired t test n = 8 biological replicates). See also Figures S5 and S6. Statistically signifificant results are labeled (* p < 0.05, ** p < 0.01, *** p < 0.001, **** p % 0.001)
Similarly, the dual MTHFD1/2 inhibitor LY345899 (Gustafsson et al., 2017) was efficacious in ameliorating disease in an EAE model. EAE was induced with MOG and PTX, and mice were treated daily with intraperitoneal (i.p.) injection of DMSO or LY345899. Inhibitor treatment led to signifificantly reduced disease severity and cumulative clinical score compared with vehicle (Figures S5C and S5D). Mice were sacrificed on day 26, and spinal cord-infiltrating cells were collected for immune profiling. Notably, signifificantly fewer CD45+, CD4+, and CD8+ cells infiltrated the spinal cord of LY345899-treated mice (Figures S5E–S5G). There were no changes in the frequency of RORγt+ or FoxP3+ cells among the infiltrating CD4+ T cells (Figure S5H). However, there was a decrease in the frequency of IFNg+ IL-17+ cells and an increase in IL-17+ cells (Figure S5I), suggesting a possible switch to a less pathogenic phenotype.
To directly test T cell-intrinsic dependence on MTHFD2 in vivo, WT and CD4DMthfd2 littermates were subjected to three different models: EAE, inflammatory bowel disease (IBD), and allergic airway disease. EAE was again induced with MOG and PTX, and, similar to the inhibitor experiment, CD4DMthfd2 mice had signifificantly reduced disease severity over 26 days (Figure 7D). A separate cohort was euthanized at peak disease on day 15 after induction for T cell phenotyping (Figures S6A and S6B). Although CD4+ or CD8+ T cell counts were unchanged in the spleen (Figure S6C), CD4+ T cell infiltration into the spinal cord was decreased signifificantly (Figure 7E). Spinal cord-infiltrating CD4+ T cells in CD4DMthfd2 mice had similar CD25 expression as the WT but reduced expression of CD44, pointing to partly impaired activation (Figure 7F). The numbers of T-bet+, RORγt+, and FoxP3+ cells (Figure 7G), as well as IL-17+ and IFNg+ cells (Figure 7H), were also reduced in the spinal cord of CD4DMthfd2 mice relative to the WT. Notably, the ratios of FoxP3+ to T-bet+ and RORgt+ cells were elevated in CD4DMthfd2 mice compared with the WT at peak disease, mirroring the in vitro findings of increased FoxP3 expression with MTHFD2 deficiency (Figure 7G). Although the frequency of TF expressing cells among infiltrating cells overall trended higher, the frequency of cytokine expression was unchanged (Figures S6D and S6E). Histology of the spinal cord of control mice without EAE and WT and CD4ΔMthfd2 mice with EAE showed increased cellularity and CD3 positivity in the spinal cord in WT mice that was associated with demyelination, as shown by Luxol Fast Blue staining (Figure 7I). These changes were absent in no-EAE control and CD4ΔMthfd2 mice.
The role of MTHFD2 was next tested in IBD and allergic airway disease models. WT and CD4DMthfd2 naive CD4+ T cells were transferred to Rag1 / recipients to initiate IBD. Although mice that received WT cells began losing weight as IBD progressed, recipients of CD4DMthfd2 T cells continued to gain weight (Figure 7J). Signifificantly fewer numbers and frequencies of CD4DMthfd2 T cells were found in spleens (Figure S6F) and mesenteric lymph nodes (MLNs) draining the colon (Figures 7K and S6G). Numbers of T-bet+, RORγt+, and FoxP3+ CD4DMthfd2 T cells were also decreased in MLNs, but the frequency of FoxP3+ cells was increased (Figures 7L and S6H). However, there were no differences in FoxP3/T-bet and FoxP3/RORgt cell count ratios in MLNs at this late-stage time point (Figure S6I). Similar to findings of MTHFD2 as a modest hit in the in vivo airway inflammation CRISPR screen (Figure 1C), CD4DMthfd2 mice subjected to Alternaria-induced allergic airway disease had a trend of decreased neutrophil abundance in the bronchioalveolar lavage fluid (BALF), although no changes were observed in BALF lymphocyte, eosinophil, and macrophage counts (Figure S7A).
Finally, the effect of the MTHFD2i was tested on general immune activity. OVA-specific OT-II CD45.2+ CD4+ T cells were adoptively transferred into CD45.1+ mice, which were then immunized subcutaneously with an emulsion of OVA with complete Freund’s adjuvant (CFA) and treated daily with oral vehicle or the MTHFD2i. Bodyweight was unchanged under all conditions (Figure S7B). Expansion of the transferred OT-II CD45.2+ cells was delayed slightly in immunized MTHFD2i-treated mice but reached vehicle numbers by day 7 (Figures S7C–S7E). However, the frequency of IFNg+ cells was reduced in draining LNs (Figure S7F). These data show that in vivo MTHFD2i treatment allows antigen-specific T cell expansion following acute stimulation, although with dampened inflammatory cytokine production.
DISCUSSION
T cell activation, differentiation, and function require appropriate metabolic reprogramming to meet the cells’ increased demands for energy, biosynthetic building blocks and signaling molecules. In this study, we investigated the function of 1C metabolism in primary CD4+ T cells. Using a combination of in vivo primary T cell CRISPR-based screening and gene expression data, MTHFD2 was identified as a potential target for anti-inflammatory therapies. MTHFD2 inhibition or deficiency generally led to decreased CD4+ T cell proliferation. Notably, this was associated with induction of FoxP3 expression and suppressive function in Th17 cells as well as enhanced Treg cell differentiation. These data suggest that MTHFD2 may function as a metabolic checkpoint in the Th17-Treg cell axis, with MTHFD2i treatment sewing the balance from a pathogenic to a more anti-inflammatory phenotype. Indeed, MTHFD2 inhibition or genetic deficiency ameliorated disease severity in multiple in vivo models of autoimmunity and hypersensitivity.
1C metabolism consists of serine-glycine metabolism, the folate cycle, and the methionine cycle and is central to several processes, including de novo purine synthesis, methyl donor generation, and redox regulation (Yang and Vousden, 2016). 1C metabolism is engaged robustly with TCR stimulation (Tan et al., 2017) and required to support T cell expansion (Ma et al., 2017). The folate cycle is compartmentalized into cytosolic and mitochondrial components. In the cytosol, MTHFD1 interconverts 5,10-methyleneTHF, 10-formylTHF, and formate. In mitochondria, the same reactions are accomplished by MTHFD2 and MTHFD1L. These resulting intermediates are used to support formylation steps in de novo purine synthesis. MTHFD1 and MTHFD2 can also modulate the redox state through NAD(H) and NADP(H) generation (Ducker and Rabinowitz, 2017). Although some cell types display flexibility in switching to the cytosolic source under settings of mitochondrial pathway dysfunction (Ducker et al., 2016), T cells have been proposed previously to depend predominantly on the mitochondrial pathway to provide 1C units and reductive species (Ron-Harel et al., 2016). T cells have also been shown to rely on serine uptake, and in vivo immune responses can be modulated by dietary serine amounts. The effects of serine starvation are mediated by limiting purine biosynthesis even in the presence of intact salvage pathways and can be bypassed by provision of glycine and formate (Ma et al., 2017). Our data add to these findings and point to MTHFD2 as a critical regulator that influences CD4+ T cell proliferation and differentiation.
Echanicoside
Although 1C metabolism plays a broad role in cell metabolism, we found that some functions were distinct to select T cell subsets. All human and mouse T cells proliferated to a lesser extent with MTHFD2 deficiency, indicating a shared function for MTHFD2 to support T cell growth and division. The effects of MTHFD2 deficiency on T cell differentiation and effector function, however, differed in each tested subset. Th1 cells showed impaired differentiation with reduced induction of T-bet and decreased cytokine production. Th17 cells also had altered differentiation and decreased cytokine production. Although RORγt was not altered, MTHFD2-deficient Th17 cells upregulated the Treg cell TF FoxP3 and gained an ability to suppress proliferation of activated CD8+ T cells. Treg cells exhibited enhanced differentiation under low TGF-b conditions. It is now well established that each of these subsets can have distinct metabolic requirements (Bantug et al., 2018; Buck et al., 2015). Our data show that MTHFD2 is also selectively required for Teff cells while promoting Treg cell fate and function. It is notable that these outcomes differ from GLUT1, GLS, or ASCT2 deficiencies that impair or alter Teff cell differentiation while having modest effects to directly promote Treg cell differentiation or transdifferentiation of Th17 to Treg cells (Johnson et al., 2018; Macintyre et al., 2014; Nakaya et al., 2014).
These distinct T cell fates appear to be dependent on the enzymatic activity of MTHFD2 because the effects of MTHFD2 deficiency were rescued with formate or purine nucleobases. Although we found no significant effect of altered redox balance with MTHFD2 deficiency, altered de novo purine synthesis appears to be critical. MTHFD2i treatment led to decreased purine concentrations, particularly guanine while inducing accumulation of the purine synthesis intermediates GAR, SAICAR, and AICAR. These intermediates require formylation for further biosynthesis, and the effects were rescued when the MTHFD2- derived metabolite formate was added to the medium, supporting an on-target biochemical mechanism for the MTHFD2i. De novo nucleotide synthesis to support DNA and RNA synthesis is essential for T cell proliferation (Que´ me´ neur et al., 2003, 2004). Thus, a component of the mechanism by which the MTHFD2i impairs T cell expansion appears to be insufficient generation of nucleotides.
Flavonoid
Failure to synthesize adequate nucleotides may suppress effector T cells through multiple mechanisms. AMPK and mTORC1 are major drivers of metabolic reprogramming with generally opposing roles in Teff and Treg cells (Bantug et al., 2018; Buck et al., 2015). We show that MTHFD2i treatment in T cells led to acute accumulation of AICAR, a metabolite notable as an adenosine analog and AMPK activator (Rae and Mairs, 2019; Su et al., 2019). However, there was no evidence of increased AMPK activity at the selected time points. This may reflect additional regulation of AMPK, but AICAR can also have AMPK-independent effects (Dembitz et al., 2019). mTORC1 is also sensitive to purine concentrations because decreased guanine availability suppresses activity of the mTORC1 activator Rheb (Emmanuel et al., 2017; Hoxhaj et al., 2017). AICAR accumulation and reduction in mTORC1 activity may thus contribute to impaired generation of MTHFD2i-treated Th1 and Th17 cells and enhanced generation of Treg cells. Suppression of mTORC1 signaling may also contribute to the shift in the metabolic program from glycolysis to mitochondrial respiration and altered TCA metabolite abundance in MTHFD2i-treated Th17 cells. The mTORC1 pathway plays a central role in promoting anabolic metabolism and can drive accumulation of the TF ATF4, which, in turn, can induce MTHFD2 expression (Ben-Sahra et al., 2016).
Metabolic flux is tied to epigenetic regulation because metabolites contribute to many epigenetic modifications (Sharma and Rando, 2017), including 1C metabolism to generate the universal methyl donor S-adenosyl methionine (SAM) for protein and DNA methylation. Carbon tracing experiments in activated T cells, however, suggest that 1C units derived from glucose-derived and exogenous serine do not contribute meaningfully to the methionine cycle and methylation (Ma et al., 2017). Metabolic flux may also affect methylation patterns through changes in the abundance of TCA metabolites, including alpha-ketoglutarate, fumarate, and succinate. Specifically, fumarate and succinate inhibit DNA and histone demethylases (Su et al., 2016). Consistent with depletion of these metabolites in MTHFD2itreated Th17 cells, H3K27me3 was reduced widely across the genome, and DNA methylation was reduced specifically at the Foxp3 proximal promoter, which may be contributing to the transdifferentiation phenotype. The Th17 cell transdifferentiation phenotype was not rescued by formate supplementation, indicating that additional pathways are likely contributing to FoxP3 regulation, such as through activity of other nucleotide-concentration-sensing mechanisms
Acteoside(Verbascoside)
MTHFD2 has also been reported to play non-enzymatic roles that may contribute to the pro-inflammatory actions of this enzyme. In cancer cells, MTHFD2 can have nuclear functions,co-localizing with DNA replication sites to promote cell cycle progression (Gustafsson Sheppard et al., 2015). In murine stem cells, MTHFD2 has been shown to modulate DNA repair to maintain genomic stability (Yue et al., 2020). In renal cell carcinoma, MTHFD2 has been found to be crucial for metabolic reprogramming via mRNA methylation (Green et al., 2019). Although these roles for MTHFD2 are not mutually exclusive with our findings, the ability of formate to rescue many of the phenotypes of MTHFD2 deficiency in T cells suggests that MTHFD2 enzymatic activity is the primary driver of T cell proliferation and fate.
MTHFD2 has been largely considered a drug target in anti-cancer settings. It may also be a promising anti-inflammatory target and offer fewer adverse effects compared with currently available anti-folates, given low expression in most adult tissues. For instance, one target of MTX, DHFR, is expressed extensively in adult tissues (Nilsson et al., 2014). Therefore, MTX can be associated with a variety of adverse effects, including gastrointestinal toxicity. Moreover, the mechanism of action of MTX remains poorly understood despite its extensive history (Cronstein and Aune, 2020). The highly regulated expression of MTHFD2 and potential for redundancy with the cytosolic MTHFD1 pathway may result in selective dependency of specific cell populations on MTHFD2. Our findings identify MTHFD2 as a critical metabolic checkpoint in CD4+ T cells. Although there is a broad overlap between cancer cell biology and the biology of rapidly proliferating T cells, we find that CD4+ T cell subsets display an additional sensitivity through modulation of cell differentiation and function. It is likely that targeting MTHFD2 in cancer therapies may restrain anti-tumor immunity. However, settings such as colorectal carcinoma, where Th17 cell-mediated inflammation contributes to disease, may be well suited to MTHFD2is. T cell sensitivity to MTHFD2is may be an effective form of immunotherapy in settings of CD4+ T cell-driven inflammation beyond cancer and lead to fewer adverse effects than currently available therapeutic agents.
Limitations of study
The MTHFD2i used in this study has an IC50 value of 0.0063 mMfor MTHFD2 and 0.57 mM for MTHFD1; therefore, minor off-target effects of MTHFD2 and MTHFD1 dual inhibition can not be excluded. Given that the aberrant FoxP3 upregulation inTh17 cells was not rescued by provision of formate but mTORC1activity was, the mechanism underlying this phenotype requires further investigation. Additionally, all in vivo models were performed with MTHFD2i treatment or genetic ablation from the time of disease induction, and T cell phenotyping was performed at a single time point. To better characterize the effect ofMTHFD2 deficiency on disease pathogenesis, further studiesshould include time-course experiments with serial characterization of T cell populations to more definitively determine whetherCD4+ T cell subsets are affected differentially in vivo. Finally, toimprove the relevance of this study to clinical applications, theeffificacy of MTHFD2i treatment started after disease onset should be tested.
STAR★METHODS
Detailed methods are provided in the online version of this paper and include the following:
KEY RESOURCES TABLE
RESOURCE AVAILABILITY
B Lead contact
B Materials availability
B Data and code availability
EXPERIMENTAL MODEL AND SUBJECT DETAILS
B Mice B Human T cells
B Cell Lines
METHOD DETAILS
B In vitro mouse CD4+ T cell activation and differentiation
B Th17 cell suppression assay
B In vitro human CD4+ T cell culture conditions
B CRISPR screening B Mass spectrometry metabolomics
B Proton nuclear magnetic resonance spectroscopy (1H-MRS)
B Immunoblotting
B CUT&RUN B EpigenDx Targeted NextGen Bisulfifite Sequencing
B In vivo EAE models for IHC and flow cytometry
B In vivo LY345899 treatment in EAE model
B In vivo DS18561882 treatment in DTH model B In vivo IBD model with WT and CD4ΔMthfd2 littermates
B In vivo allergic airway disease model with WT and CD4ΔMthfd2 littermates
B In vivo DS18561882 treatment in OVA immunization model
QUANTIFICATION AND STATISTICAL ANALYSIS
SUPPLEMENTAL INFORMATION
Supplemental information can be found online at https://doi.org/10.1016/j. immuni.2021.10.011.
ACKNOWLEDGMENTS
We thank members of the Rathmell lab for contributing to this project. We thank Thomas Aune for providing RNA-seq data and J. Cools (VIB) for providing the pMx-U6-gRNA-GFP construct. We thank Max R. Van Belkum for designing the MTHFD2 qPCR primers. We thank Nello Mainolfifi, Vipin Suri, Adam Friedman, and Mark Manfredi from Raze Therapeutics, Inc. (Boston, MA) for providing the Raze 1459 compound, which was used to corroborate findings (data not shown). Diagrams were created with BioRender. We acknowledge the Translational Pathology Shared Resource, supported by NCI/NIH Cancer Center support grant 5P30 CA68485-19 and shared instrumentation grant S10 OD023475-01A1, for the Leica Bond RX. This work was supported by the William E. Paul Distinguished Innovator Award for the Lupus Research Alliance (to J.C.R.), R01s DK105550 (to J.C.R.), HL136664 (to J.C.R. and D.C.N.), CA217987 (to J.C.R.), AI153167 (to J.C.R.), AI137075 (to A.K.M.), T32 DK101003 (K.V.), and T32 GM007347 (to A.S.).
AUTHOR CONTRIBUTIONS
A.S., G.A., and J.C.R. designed the research. A.S., G.A., K.V., D.R.H., X.X., M.Z.M., X.Y., K.L.B., N.C., M.M.W., A.C.Y., D.L.G., A.E.S., S.K.S., S.N.F., A.M.C., P.F., T.B., and J.C.G.-C. performed the research. A.S., G.A., X.Y., J.K., D.C.N., A.K.M., J.D.R., and J.C.R. analyzed data. A.S. and J.C.R. wrote the paper with contributions from the other authors.
DECLARATION OF INTERESTS
J.C.R. is a founder, scientific advisory board member, and stockholder of Sitryx Therapeutics; a scientific advisory board member and stockholder of Caribou Biosciences; a member of the scientific advisory board of NirogyTherapeutics; has consulted for Merck, Pfizer, and Mitobridge within the past 3 years; and has received research support from Incyte Corp., CalitheraBiosciences, and Tempest Therapeutics. J.D.R. is a co-founder and stockholder in Raze Therapeutics, Toran, Serien Therapeutics, and Farber Partners and an advisor and stockholder in Agios Pharmaceuticals, Kadmon Pharmaceuticals, Bantam Pharmaceuticals, Colorado Research Partners, RafaelHoldings, the Barer Institute, and L.E.A.F. Pharmaceuticals; he has received consulting fees and research funding from Pfizer and Rafael and is the inventor of patents held by Princeton University. A.M.C., P.F., and T.B. are employees of Sitryx Therapeutics.
INCLUSION AND DIVERSITY
We worked to ensure sex balance in the selection of non-human subjects. One or more of the authors of this paper self-identifies as an underrepresented ethnic minority in science. While citing references scientifically relevant for this work, we also actively worked to promote gender balance in our reference list. The author list of this paper includes contributors from the location where the research was conducted who participated in the data collection, design, analysis, and/or interpretation of the work.
REFERENCES
Anderson, G.R., Winter, P.S., Lin, K.H., Nussbaum, D.P., Cakir, M., Stein, E.M., Soderquist, R.S., Crawford, L., Leeds, J.C., Newcomb, R., et al. (2017). A Landscape of Therapeutic Cooperativity in KRAS Mutant Cancers Reveals Principles for Controlling Tumor Evolution. Cell Rep. 20, 999–1015.
Aune, T.M., Crooke, P.S., 3rd, Patrick, A.E., Tossberg, J.T., Olsen, N.J., and Spurlock, C.F., 3rd (2017). Expression of long non-coding RNAs in autoimmunity and linkage to enhancer function and autoimmune disease risk genetic variants. J. Autoimmun. 81, 99–109.
Bantug, G.R., Galluzzi, L., Kroemer, G., and Hess, C. (2018). The spectrum of T cell metabolism in health and disease. Nat. Rev. Immunol. 18, 19–34. Ben-Sahra, I., Hoxhaj, G., Ricoult, S.J.H., Asara, J.M., and Manning, B.D. (2016). mTORC1 induces purine synthesis through control of the mitochondrial tetrahydrofolate cycle. Science 351, 728–733.
Brown, P.M., Pratt, A.G., and Isaacs, J.D. (2016). Mechanism of action of methotrexate in rheumatoid arthritis, and the search for biomarkers. Nat. Rev. Rheumatol. 12, 731–742.
Buck, M.D., O’Sullivan, D., and Pearce, E.L. (2015). T cell metabolism drives immunity. J. Exp. Med. 212, 1345–1360.
Cantor, J.R., Abu-Remaileh, M., Kanarek, N., Freinkman, E., Gao, X., Louissaint, A., Jr., Lewis, C.A., and Sabatini, D.M. (2017). Physiologic Medium Rewires Cellular Metabolism and Reveals Uric Acid as an Endogenous Inhibitor of UMP Synthase. Cell 169, 258–272.e17.
Cronstein, B.N., and Aune, T.M. (2020). Methotrexate and its mechanisms of action in inflammatory arthritis. Nat. Rev. Rheumatol. 16, 145–154.
Dembitz, V., Tomic, B., Kodvanj, I., Simon, J.A., Bedalov, A., and Visnjic, D. (2019). The ribonucleoside AICAr induces differentiation of myeloid leukemia by activating the ATR/Chk1 via pyrimidine depletion. J. Biol. Chem. 294, 15257–15270.
Dendrou, C.A., Fugger, L., and Friese, M.A. (2015). Immunopathology of multiple sclerosis. Nat. Rev. Immunol. 15, 545–558.
Ducker, G.S., and Rabinowitz, J.D. (2017). One-Carbon Metabolism in Health and Disease. Cell Metab. 25, 27–42.
Ducker, G.S., Chen, L., Morscher, R.J., Ghergurovich, J.M., Esposito, M., Teng, X., Kang, Y., and Rabinowitz, J.D. (2016). Reversal of Cytosolic OneCarbon Flux Compensates for Loss of the Mitochondrial Folate Pathway. Cell Metab. 23, 1140–1153.
Emmanuel, N., Ragunathan, S., Shan, Q., Wang, F., Giannakou, A., Huser, N., Jin, G., Myers, J., Abraham, R.T., and Unsal-Kacmaz, K. (2017). Purine Nucleotide Availability Regulates mTORC1 Activity through the Rheb GTPase. Cell Rep. 19, 2665–2680.
Farber, S., Diamond, L.K., Mercer, R.D., Sylvester, R.F., and Wolff, J.A. (1948). Temporary remissions in acute leukemia in children produced by folic acid antagonist, 4-aminopteroyl-glutamic acid. N. Engl. J. Med. 238, 787–793.
Field, M.S., Kamynina, E., Watkins, D., Rosenblatt, D.S., and Stover, P.J. (2015). New insights into the metabolic and nutritional determinants of severe combined immunodeficiency. Rare Dis. 3, e1112479.
Fuseini, H., Yung, J.A., Cephus, J.Y., Zhang, J., Goleniewska, K., Polosukhin, V.V., Peebles, R.S., Jr., and Newcomb, D.C. (2018). Testosterone Decreases House Dust Mite-Induced Type 2 and IL-17A-Mediated Airway Inflammation. J. Immunol. 201, 1843–1854.
Govindaraju, V., Young, K., and Maudsley, A.A. (2000). Proton NMR chemical shifts and coupling constants for brain metabolites. NMR Biomed. 13, 129–153.
Green, N.H., Galvan, D.L., Badal, S.S., Chang, B.H., LeBleu, V.S., Long, J., Jonasch, E., and Danesh, F.R. (2019). MTHFD2 links RNA methylation to metabolic reprogramming in renal cell carcinoma. Oncogene 38, 6211–6225.
Gustafsson, R., Jemth, A.-S., Gustafsson, N.M.S., F€ arnega˚rdh, K., Loseva, O., Wiita, E., Bonagas, N., Dahllund, L., Llona-Minguez, S., H€ aggblad, M., et al. (2017). Crystal Structure of the Emerging Cancer Target MTHFD2 in Complex with a Substrate-Based Inhibitor. Cancer Res. 77, 937–948.
Gustafsson Sheppard, N., Jarl, L., Mahadessian, D., Strittmatter, L., Schmidt, A., Madhusudan, N., Tegne´r, J., Lundberg, E.K., Asplund, A., Jain, M., and Nilsson, R. (2015). The folate-coupled enzyme MTHFD2 is a nuclear protein and promotes cell proliferation. Sci. Rep. 5, 15029.
Hoxhaj, G., Hughes-Hallett, J., Timson, R.C., Ilagan, E., Yuan, M., Asara, J.M., Ben-Sahra, I., and Manning, B.D. (2017). The mTORC1 Signaling Network Senses Changes in Cellular Purine Nucleotide Levels. Cell Rep. 21, 1331–1346.
Johnson, M.O., Wolf, M.M., Madden, M.Z., Andrejeva, G., Sugiura, A., Contreras, D.C., Maseda, D., Liberti, M.V., Paz, K., Kishton, R.J., et al. (2018). Distinct Regulation of Th17 and Th1 Cell Differentiation by Glutaminase-Dependent Metabolism. Cell 175, 1780–1795.e19.
Ju, H.-Q., Lu, Y.-X., Chen, D.-L., Zuo, Z.-X., Liu, Z.-X., Wu, Q.-N., Mo, H.-Y., Wang, Z.-X., Wang, D.-S., Pu, H.-Y., et al. (2019). Modulation of Redox Homeostasis by Inhibition of MTHFD2 in Colorectal Cancer: Mechanisms and Therapeutic Implications. J. Natl. Cancer Inst. 111, 584–596.
Kawai, J., Toki, T., Ota, M., Inoue, H., Takata, Y., Asahi, T., Suzuki, M., Shimada, T., Ono, K., Suzuki, K., et al. (2019). Discovery of a Potent, Selective, and Orally Available MTHFD2 Inhibitor (DS18561882) with in Vivo Antitumor Activity. J. Med. Chem. 62, 10204–10220.
Kim, J., Yang, G., Kim, Y., Kim, J., and Ha, J. (2016). AMPK activators: mechanisms of action and physiological activities. Exp. Mol. Med. 48, e224.
Koufaris, C., Gallage, S., Yang, T., Lau, C.-H., Valbuena, G.N., and Keun, H.C. (2016). Suppression of MTHFD2 in MCF-7 Breast Cancer Cells Increases Glycolysis, Dependency on Exogenous Glycine, and Sensitivity to Folate Depletion. J. Proteome Res. 15, 2618–2625.
Li, W., Xu, H., Xiao, T., Cong, L., Love, M.I., Zhang, F., Irizarry, R.A., Liu, J.S., Brown, M., and Liu, X.S. (2014). MAGeCK enables robust identification of essential genes from genome-scale CRISPR/Cas9 knockout screens. Genome Biol. 15, 554.
Ma, E.H., Bantug, G., Griss, T., Condotta, S., Johnson, R.M., Samborska, B., Mainolfifi, N., Suri, V., Guak, H., Balmer, M.L., et al. (2017). Serine Is an Essential Metabolite for Effector T Cell Expansion. Cell Metab. 25, 345–357.
Macintyre, A.N., Gerriets, V.A., Nichols, A.G., Michalek, R.D., Rudolph, M.C., Deoliveira, D., Anderson, S.M., Abel, D.E., Chen, B.J., Hale, L.P., et al. (2014). The Glucose Transporter Glut1 Is Selectively Essential for CD4 T Cell Activation and Effector Function. Cell Metab. 20, 61–72.
Michalek, R.D., Gerriets, V.A., Jacobs, S.R., Macintyre, A.N., MacIver, N.J., Mason, E.F., Sullivan, S.A., Nichols, A.G., and Rathmell, J.C. (2011). Cutting edge: distinct glycolytic and lipid oxidative metabolic programs are essential for effector and regulatory CD4+ T cell subsets. J. Immunol. 186, 3299–3303.
Nakaya, M., Xiao, Y., Zhou, X., Chang, J.-H., Chang, M., Cheng, X., Blonska, M., Lin, X., and Sun, S.-C. (2014). Inflammatory T cell responses rely on amino acid transporter ASCT2 facilitation of glutamine uptake and mTORC1 kinase activation. Immunity 40, 692–705.
Nilsson, R., Jain, M., Madhusudhan, N., Sheppard, N.G., Strittmatter, L., Kampf, C., Huang, J., Asplund, A., and Mootha, V.K. (2014). Metabolic enzyme expression highlights a key role for MTHFD2 and the mitochondrial folate pathway in cancer. Nat. Commun. 5, 3128.
Palmer, L.D., Maloney, K.N., Boyd, K.L., Goleniewska, A.K., Toki, S., Maxwell, C.N., Chazin, W.J., Peebles, R.S., Jr., Newcomb, D.C., and Skaar, E.P. (2019). The Innate Immune Protein S100A9 Protects from T-Helper Cell Type 2-mediated Allergic Airway Inflammation. Am. J. Respir. Cell Mol. Biol. 61, 459–468.
Pikman, Y., Puissant, A., Alexe, G., Furman, A., Chen, L.M., Frumm, S.M., Ross, L., Fenouille, N., Bassil, C.F., Lewis, C.A., et al. (2016). Targeting MTHFD2 in acute myeloid leukemia. J. Exp. Med. 213, 1285–1306.
Que´ me´ neur, L., Gerland, L.-M., Flacher, M., Ffrench, M., Revillard, J.-P., and Genestier, L. (2003). Differential control of cell cycle, proliferation, and survival of primary T lymphocytes by purine and pyrimidine nucleotides. J. Immunol. 170, 4986–4995.
Que´ me´ neur, L., Beloeil, L., Michallet, M.-C., Angelov, G., Tomkowiak, M., Revillard, J.-P., and Marvel, J. (2004). Restriction of de novo nucleotide biosynthesis interferes with clonal expansion and differentiation into effector and memory CD8 T cells. J. Immunol. 173, 4945–4952.
Rae, C., and Mairs, R.J. (2019). AMPK activation by AICAR sensitizes prostate cancer cells to radiotherapy. Oncotarget 10, 749–759.
Ron-Harel, N., Santos, D., Ghergurovich, J.M., Sage, P.T., Reddy, A., Lovitch, S.B., Dephoure, N., Satterstrom, F.K., Sheffer, M., Spinelli, J.B., et al. (2016). Mitochondrial Biogenesis and Proteome Remodeling Promote One-Carbon Metabolism for T Cell Activation. Cell Metab. 24, 104–117.
Shahi, S.K., Freedman, S.N., Murra, A.C., Zarei, K., Sompallae, R., GibsonCorley, K.N., Karandikar, N.J., Murray, J.A., and Mangalam, A.K. (2019). Prevotella histicola, A Human Gut Commensal, Is as Potent as COPAXONE in an Animal Model of Multiple Sclerosis. Front. Immunol. 10, 462.
Shalem, O., Sanjana, N.E., Hartenian, E., Shi, X., Scott, D.A., Mikkelson, T., Heckl, D., Ebert, B.L., Root, D.E., Doench, J.G., and Zhang, F. (2014). Genome-scale CRISPR-Cas9 knockout screening in human cells. Science 343, 84–87.
Sharma, U., and Rando, O.J. (2017). Metabolic Inputs into the Epigenome. Cell Metab. 25, 544–558.
Shi, L.Z., Wang, R., Huang, G., Vogel, P., Neale, G., Green, D.R., and Chi, H. (2011). HIF1alpha-dependent glycolytic pathway orchestrates a metabolic checkpoint for the differentiation of T H17 and Treg cells. J. Exp. Med. 208, 1367–1376.
Shin, M., Momb, J., and Appling, D.R. (2017). Human mitochondrial MTHFD2 is a dual redox cofactor-specific methylenetetrahydrofolate dehydrogenase/ methyltetrahydrofolate cyclohydrolase. Cancer Metab. 5, 11.
Skene, P.J., and Henikoff, S. (2017). An efficient targeted nuclease strategy for high-resolution mapping of DNA binding sites. eLife 6, e21856.
Su, C.-C., Hsieh, K.-L., Liu, P.-L., Yeh, H.-C., Huang, S.-P., Fang, S.-H., Cheng, W.-C., Huang, K.-H., Chiu, F.-Y., Lin, I.-L., et al. (2019). AICAR Induces Apoptosis and Inhibits Migration and Invasion in Prostate Cancer Cells Through an AMPK/mTOR-Dependent Pathway. Int. J. Mol. Sci. 20, 1647.
Su, X., Wellen, K.E., and Rabinowitz, J.D. (2016). Metabolic control of methylation and acetylation. Curr. Opin. Chem. Biol. 30, 52–60.
Tan, H., Yang, K., Li, Y., Shaw, T.I., Wang, Y., Blanco, D.B., Wang, X., Cho, J.-H., Wang, H., Rankin, S., et al. (2017). Integrative Proteomics and Phosphoproteomics Profifiling Reveals Dynamic Signaling Networks and Bioenergetics Pathways Underlying T Cell Activation. Immunity 46, 488–503.
Toffalini, F., Kallin, A., Vandenberghe, P., Pierre, P., Michaux, L., Cools, J., and Demoulin, J.-B. (2009). The fusion proteins TEL-PDGFR and FIP1L1-PDGFR escape ubiquitination and degradation. Haematologica 94, 1085–1093.
Wan, X., Wang, C., Huang, Z., Zhou, D., Xiang, S., Qi, Q., Chen, X., Arbely, E., Liu, C.-Y., Du, P., and Yu, W. (2020). Cisplatin inhibits SIRT3-deacetylation MTHFD2 to disturb cellular redox balance in colorectal cancer cell. Cell Death Dis. 11, 649.
Wang, L., Xing, X., Chen, L., Yang, L., Su, X., Rabitz, H., Lu, W., and Rabinowitz, J.D. (2019). Peak Annotation and Verification Engine for Untargeted LC-MS Metabolomics. Anal. Chem. 91, 1838–1846.
Yang, M., and Vousden, K.H. (2016). Serine and one-carbon metabolism in cancer. Nat. Rev. Cancer 16, 650–662.
Yue, L., Pei, Y., Zhong, L., Yang, H., Wang, Y., Zhang, W., Chen, N., Zhu, Q., Gao, J., Zhi, M., et al. (2020). Mthfd2 Modulates Mitochondrial Function and DNA Repair to Maintain the Pluripotency of Mouse Stem Cells. Stem Cell Reports 15, 529–545.
Zhu, Z., and Leung, G.K.K. (2020). More Than a Metabolic Enzyme: MTHFD2 as a Novel Target for Anticancer Therapy? Front. Oncol. 10, 658.
STAR★METHODS
KEY RESOURCES TABLE
RESOURCE AVAILABILITY
Lead contact
Further information and requests for resources and reagents should be directed to and will be fulfilled upon completion of appropriate MTA by the Lead Contact, Dr. Jeffrey Rathmell (jeff.rathmell@vumc.org).
Materials availability
Plasmids and mouse lines generated in this study will be made available upon completion of appropriate MTA upon request.
Data and code availability
CUT&RUN sequencing data have been deposited at GEO and are publicly available as of the date of publication (GSE180356). Microscopy data and original western blot images reported in this paper will be shared by the lead contact upon request. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.
