Tissue and CD4 T Cell Subset Dependence on the Amino Acid Transporter SLC38A1

SUMMARY

Amino acid (AA) uptake is essential for T cell metabolism and function, but how tissue sites and inflammation affect CD4⁺ T cell subset requirements for specific AA remains uncertain. Here we tested CD4⁺ T cell AA demands with in vitro and in vivo CRISPR screens and identify subset- and tissue-specific dependencies on the AA transporter SLC38A1 (SNAT1). While dispensable for T cell persistence and expansion in vivo in lung inflammation, SLC38A1 was critical for Th1 but not Th17 cell-driven Experimental Autoimmune Encephalomyelitis (EAE) and contributed to Th1 cell-driven inflammatory bowel disease. SLC38A1 deficiency reduced mTORC1 signaling and glycolytic activity in Th1 cells, in part by reducing glutamine uptake and disrupting hexosamine biosynthesis and redox regulation. Pharmacological inhibition of SLC38 transporters also delayed Th1-mediated EAE but did not affect lung inflammation. CD4⁺ T cells thus have subset- and tissue-specific nutrient transporter dependencies that may guide new metabolic approaches to selective immunotherapies.

Graphical Abstract

eTOC Blurb

Metabolic demands and mechanisms of nutrient uptake shape T cell function and offer new therapeutic opportunities but selective targeting remains challenging. Here in vivo CRISPR screens show that CD4 T cell metabolism and nutrient uptake vary based on both cell subset and the tissue and inflammatory site.

INTRODUCTION

An effective immune response requires appropriate T cell activation, proliferation, differentiation, and function. Each of these processes utilizes specific cellular metabolic programs with distinct nutrient requirements. In addition to glucose and lipids, coordination of appropriate amino acid (AA) uptake and metabolism is critical to CD4⁺ T cell activity^1,2. While essential AAs must be imported from the extracellular environment, non-essential AAs can be synthesized intracellularly. Cells may also become dependent on uptake of conditionally essential AAs from extracellular pools when synthesis fails to meet increased demands, such as may occur in rapid cell growth and proliferation following T-cell receptor (TCR) stimulation. Previous studies have shown that CD4⁺ T cell fate and function may be altered or impaired in the setting of specific AA insufficiencies, including glutamine^3–9, leucine^10,11, arginine^12,13, serine¹⁴, alanine¹⁵, and methionine^16–18. Cancer cells may also impair AA uptake to suppress anti-tumor immunity¹⁹. Of these, the conditionally-essential AA glutamine has been most clearly shown to differentially affect both effector and regulatory CD4⁺ T helper (Teff and Treg) cell subsets^8,20,21.

AA transport is mediated by the solute carrier (SLC) family of nutrient transporters. The SLC family is comprised of over 400 facilitative and secondary active membrane transporters that carry a variety of organic and inorganic substrates and includes approximately 60 family members that selectively transport AAs. With more than 100 monogenic diseases linked to SLCs to date, these transporters have attracted interest as potential therapeutic targets^22–24. The breadth of transporters with differential expression patterns and overlapping substrates provides transport redundancy and opportunities to fine-tune nutrient use and metabolism by expression of different transporter isoforms. Previous studies have shown that SLC7A5 (LAT1) and SLC1A5 (ASCT2) play critical roles affecting inflammatory functions in CD4⁺ Teff cells^4,11. SLC7A5 forms a heterodimer with the heavy chain SLC3A2 (CD98) and imports leucine and other large neutral AAs in exchange for glutamine. The intracellular glutamine pool is in turn primarily maintained by SLC1A5-mediated uptake and drives transport of essential AAs and support a wide array of metabolic pathways and regulatory mechanisms. Loss of SLC7A5 is associated with failure in mTORC1 and MYC signaling-driven metabolic reprogramming leading to impaired clonal expansion and Teff cell differentiation following TCR engagement¹¹. Similarly, loss of SLC1A5 is associated with mTORC1-mediated impairment in Th1 and Th17 cell differentiation and function⁴. Despite these clear roles, T cells express multiple glutamine transporters that may also contribute to T cell metabolism in specific settings.

How AA transport mechanisms and requirements of distinct CD4⁺ T cell subsets are shaped by specific tissue microenvironments remain poorly understood. Differential gene expression and transport redundancy of SLC family members together with variable nutrient access in distinct tissues suggests that individual nutrient transporters may have context-specific phenotypes. To test this, we assessed AA requirements and focused on the conditionally essential AA glutamine in vitro and across multiple in vivo disease models of autoimmunity and inflammation. Glutamine was required by Teff, but not by Treg in vitro. CRISPR screening of glutamine transporters and metabolizing enzymes showed that while dispensable in vitro, the sodium-coupled neutral AA transporter SLC38A1 (SNAT1) was specifically required in vivo in pathogenic Th1-, but not Th17-, driven Experimental Autoimmune Encephalomyelitis (EAE). SLC38A1-inhibition or deficiency delayed EAE symptom onset and lessened weight loss associated with Th1-cell driven inflammatory bowel disease (IBD). SLC38A1 was dispensable, however, in a T-cell driven model of allergic lung airway disease. These data identify cell subset- and tissue-specific AA transport dependencies that illustrate the role of the tissue microenvironment on T cell metabolism and provide a new tissue and disease-based mechanism for selectivity of potential metabolic immunotherapy targets.

RESULTS

T cell nutrient transporters and amino acid (AA) uptake.

Within the SLC family, over 60 transporters are predicted to transport AA²⁵. To better understand mechanisms by which T cells acquire necessary AAs, we analyzed a previously collected RNAseq dataset²⁶ examining primary CD4⁺ T cells following TCR stimulation. At 5- and 24-hours post activation, T cells upregulated multiple AA-carrying SLCs, with peak expression at 5 hours for Slc7a1, Slc7a5, Slc1a5, Slc3a2, Slc38a2, and Slc38a1, and Slc7a11 (Figure 1A). Notably, glutamine transporters were disproportionately represented among the upregulated SLCs, including the sodium-dependent small neutral AA transporters Slc1a5, Slc38a1, and Slc38a2.

Figure 1: CD4⁺ T cells are dependent on uptake of select essential and non-essential AAs through dynamic regulation of SLC transporters.

A. mRNA expression of SLC transporters in primary CD4⁺ T cells at 0, 5, and 24 hours post activation with anti-CD3/CD28 antibodies, referenced from published RNAseq dataset²⁶. Red = transports glutamine, blue = transports other AAs, gray = all other SLCs detected (mean, n=3 biological replicates).

B. Experimental design for in vitro CRISPR screening.

C-E. Volcano plots showing changes in gRNA abundancies from in vitro CRISPR screens performed with primary Cas9-transgenic (C) Th0, (D) Th17, and (E) Treg cells transduced with SLC gRNA library. Red = transports glutamine, blue = transports other AAs, gray = all other SLCs detected (n=3 biological replicates with 2 technical replicates each, statistical analysis performed using MAGeCK⁵⁴).

F-J. Characterization of Th0 cells cultured in vitro for 72 hours in either complete media or test media deficient in the indicated AA. (F) Cell viability normalized to complete media, (G-H) proliferation quantified as division index based on CTV dilution, and (I-J) expression of activation markers CD25 and CD44 as measured by flow cytometry (mean±SD, one-sample t-test, α=0.01, n=3 biological replicates representing at least three independent experiments compared to complete media).

K-L Expression of lineage-characterizing transcription factors RORγt and FoxP3 in Th17 and Treg cells cultured in complete or test media for 72 hours (mean±SD, one-sample t-test, α=0.01, n=4 biological replicates across four independent experiments compared to complete media).

Ns denotes p>0.05, * p≤0.05, ** p≤0.01, *** p≤0.001, **** p≤0.0001. See also Figure S1.

To test T cell requirements for AA transport and metabolism, we performed in vitro CRISPR screens using a custom guide RNA (gRNA) library targeting 362 SLCs with four gRNAs per gene and 20 non-targeting controls (NTCs) curated from the whole-genome Mouse CRISPR Knockout Pooled Brie²⁷ and GeCKO v2²⁸ libraries (Supplemental Table 1, Figure 1B). The pooled library was transduced into activated Cas9-transgenic CD4⁺ T cells²⁹ which were expanded in vitro by anti-CD3/CD28 stimulation. Comparison of gRNA abundancies before and after expansion revealed depletion of SLC7A5-targeted gRNAs, consistent with its known role in CD4⁺ T cell clonal expansion^10,11 (Figure 1C–E). To a lesser extent, gRNAs targeting the neutral AA transporters Slc1a5, Slc38a1, Slc38a2, Slc38a8 (Snat8), Slc38a10 (Snat10), and Slc7a8 (Lat2), the heavy chain subunit Slc3a2, the cationic AA transporters Slc7a1 and Slc7a9, the phosphate transporter Slc17a7, and the mitochondrial aspartate/glutamate (Slc25a12) antiporter were also depleted. Overall, RNAseq and CRISPR screen results were highly overlapping, showing enrichment of glutamine-metabolism associated SLCs with notable differences between T cell subsets.

CD4⁺ T cells are dependent on uptake of select AAs to support activation, proliferation, and differentiation.

Previous studies have shown that CD4⁺ T cells are dependent on the uptake of select essential and conditionally essential AAs²⁴. To evaluate relative AA dependencies during activation, proliferation, and differentiation, CD4⁺ T cells were cultured in complete RPMI media or equivalent media lacking a single AA. Nineteen AA-deficient media were tested, with alanine-deficient media excluded as it is absent in RPMI. Despite supplementation with 10% fetal bovine serum (FBS) nuclear magnetic resonance (NMR) confirmed that significant reduction was achieved in each of the AA-deficient media relative to complete media (Supplemental Figure 1A, 1B).

Histidine, isoleucine, leucine, lysine, methionine, phenylalanine, threonine, tryptophan, and valine are essential AAs and cannot be de novo synthesized. Deficiency in any essential AA impaired T cell activation, proliferation, and survival as expected (Figure 1F–J). Notably, deficiency of several conditionally essential and non-essential AAs - arginine, asparagine, cystine, glutamine, and tyrosine - also impaired activation and proliferation. In contrast, aspartate, glutamate, glycine, proline, and serine did not significantly affect the measured parameters. These data are consistent with published literature^{3–7,10,12,17}. When these experiments were repeated in the presence of, either IL-6, TGFβ IL-1β, and IL23 or TGFβ and IL-2 to promote Th17 cell or Treg cell differentiation, respectively, Th17 cells showed AA patterns similar to bulk T cells, with lysine, methionine, threonine, glutamine, and asparagine required to upregulate the Th17-lineage characterizing transcription factor RORγt (Figure 1K). Treg cells displayed similar AA requirements for FoxP3 upregulation, with the notable exception of glutamine (Figure 1L).

We next sought to better characterize the effects of extracellular glutamine deficiency on CD4⁺ T cell subsets. Glutamine deficiency reduced proliferation across Th1, Th17, and Treg cells while viability was reduced in Th1 and Th17 cells but not Treg cells (Supplemental Figure 1C, 1D). The activity of mTORC1, a central nutrient sensor, was significantly reduced as measured by phosphorylated ribosomal protein S6 (phosphor-S6) in both Th1 and Th17 cells as expected (Supplemental Figure 1E). Differentiation of Th1 and Th17 cells was again impaired under these conditions, while FoxP3 induction and maintenance were preserved in Treg cells (Supplemental Figure 1F). Additionally, cells cultured in glutamine-deficient media exhibited decreased effector functions with fewer IFNγ⁺ Th1 cells and IL-17⁺ Th17 cells, while the fraction of IL-2⁺ Th1 cells was increased (Supplemental Figure 1G). These data are consistent with published literature^3–7 and supported further investigation of subset specific glutamine transport and metabolism.

Glutamine transport and metabolism requirements vary depending on cell type, experimental setting, and tissue microenvironment.

There is growing evidence that the extracellular nutrient composition can impact T cell metabolism and function¹. For instance, different synthetic medias can cause large shifts in the cellular metabolome in vitro^30–32. To interrogate glutamine transport and metabolism pathway dependencies in vivo, we next synthesized a second gRNA library targeting 32 glutamine uptake and metabolizing genes (Supplemental Table 2). CRISPR screens were then performed in T cell-driven inflammatory disease models involving different T cell subsets in different tissue microenvironments.

We first tested a model of Experimental Autoimmune Encephalomyelitis (EAE) neuroinflammation model of multiple sclerosis³³. Primary 2D2/Cas9-transgenic CD4⁺ T cells expressing myelin oligodendrocyte glycoprotein (MOG)-specific TCR were activated with MOG_35–55 peptide, polarized to Th1 or Th17 cell lineage, and transduced with the glutamine gRNA library. Transduced cells were then adoptively transferred into Rag1^−/− mice to induce either Th1- or Th17-cell driven CNS demyelinating disease (Figure 2A)³³. The recipient mice developed characteristic clinical symptoms after several weeks, including ascending paralysis in the Th1 cell-driven model and ataxia in the Th17 cell-driven model. At the study endpoint, the adoptively transferred cells were recovered from the brain and spinal cord for gRNA sequencing to compare relative changes in gRNA abundancies between the Th1 and Th17 cells (Figure 2B, Supplemental Figure 2A). Positive control Tsc2 gRNAs were enriched in both subsets as expected with loss of negative regulation of mTORC1 activity. Similarly, Ppat and Cad gRNAs were depleted in both subsets, indicating shared requirement for de novo nucleotide biosynthesis pathways. Notably, Gfpt1 and Slc38a1 gRNAs were only depleted in Th1 cells, while those targeting Gclc were only depleted in Th17 cells. These data are consistent with previous findings of selective Gfpt1-dependency in Th1 cells and Gclc-dependency in Th17 cells^34,35. Th1 and Th17 cells may thus be distinguished by their differential dependencies on SLC38A1-mediated AA import, hexosamine biosynthesis catalyzed by glutamine-fructose-6-phosphate transaminase 1 (GFPT1), and glutathione synthesis mediated by glutamate/cysteine ligase catalytic unit (GCLC).

Figure 2: AA transporter requirements are subset-specific and dependent on culture conditions in vitro and tissue microenvironment in vivo.

A. Experimental design for in vivo CRISPR screening in Th1- and Th17-cell driven EAE model.

B. Changes in gRNA abundancies from in vivo CRISPR screens using a focused glutamine gRNA library in Th1 cell-mediated (x-axis) and Th17 cell-mediated (y-axis) EAE models. Green = statistical significance only in Th1 cells, blue = only in Th17 cells, red = in both subsets (statistical analysis performed by MAGeCK, representative plot from n=3 biological replicates with 2 technical replicates each across 2 independent experiments).C. Volcano plots showing changes in gRNA abundancies from in vivo CRISPR screens using glutamine gRNA library performed in Th1 cell- and Th17 cell-mediated inflammatory lung disease models following previously published methods²⁹ (n=3 biological replicates with 2 technical replicates each, statistical analysis performed using MAGeCK).

D. Relative protein expression of SLC1A5, SLC38A1, and SLC38A2 in CD4⁺ T cell subsets (Immunological Proteome Resource, ImmPRes)³⁶.

Ns denotes p>0.05, * p≤0.05, ** p≤0.01, *** p≤0.001, **** p≤0.0001. See also Figure S2.

To test the generality of this finding, a similar CRISPR screen of the glutamine library was performed on Th1 and Th17 cells in an inflammatory lung disease model²⁹. OTII/Cas9 double-transgenic primary CD4⁺ T cells expressing ovalbumin-specific TCRs were activated with OVA_323–339 peptide. T cells transduced with the gRNA library in Th1 or Th17 conditions were then adoptively transferred into the recipient Rag1^−/− mice by tail vein injection. The recipient mice were administered intranasal ovalbumin on days 1, 3, 5, and 7 after transfer to promote recruitment of the transferred cells to the lungs and induce inflammation. 24 hours after the last treatment, T cells were recovered from the lungs and associated lymph nodes for gRNA sequencing. Relative to the adoptively transferred population, the lung population was significantly depleted of Gfpt1, Cad, Ctps, Gclc, and Got2 gRNAs while Tsc2 and Cps1 gRNAs were enriched in Th17 cells (Figure 2C). Notably, Slc38a1 and Slc38a2 gRNA frequencies were not significantly changed in either Th1 or Th17-mediated lung inflammatory settings, indicating that unlike EAE, these transporters were dispensable for lung-infiltrating pathogenic T cells.

Based on differential requirements for Slc38a1 in Th1 and Th17 cells and in different disease models, glutamine transporter expression patterns were examined. Quantitative proteomics showed distinct patterns of SLC1A5, SLC38A1, and SLC38A2 expression across human T cell subsets (Figure 2D), T cell developmental stages (Supplemental Figure 2B), and various human inflammatory diseases (Supplemental Figure 2C)³⁶. SLC38A1 upregulation was also tissue-dependent, occurring only in the intraepithelial CD4⁺ T cells in a T cell transfer model of IBD (Supplemental Figure 2D). These data suggest that, while multiple SLCs can have overlapping substrate profiles, differential cell subset expression and tissue-specific regulation may provide selective and non-redundant functions.

Direct genetic testing validates SLC38A1 role in Th1 cells.

To validate the CRISPR screen findings, genetic loss of SLC38A1 was induced by transducing primary Cas9-transgenic Th1, Th17, and Treg cells with NTC (WT) orSlc38a1-targeted gRNA (ΔSlc38a1) marked with a Thy1.1 reporter. The frequencies of the Thy1.1⁺ gRNA-transduced populations were monitored over time to determine relative fitness. SLC38A1-deficiency significantly hindered Th1 cell expansion, mildly reduced Treg cell proliferation, but had no measurable effect on Th17 cell proliferation (Figures 3A). Immunoblotting confirmed reduced SLC38A1 protein levels in ΔSlc38a1 cells relative to WT in both Th1 and Th17 cells, with lower baseline expression in Th17 cells (Figure 3B) consistent with the proteome atlas data (Figure 2D). Because multiple AA transporters can contribute to glutamine uptake and SLC38A1 can also transport alanine, we next tested how SLC38A1-deficiency affected glutamine uptake. While SLC38A1-deficiency alone resulted in only a trend for decreased glutamine uptake, significant inhibition was measured upon simultaneous inhibition of SLC1A5 with V9302 (Figure 3C). Decreased fitness of SLC38A1-deficient Th1 cells was attributed to reduced glutamine rather than alanine uptake, as alanine supplementation failed to rescue and instead exacerbated impairment. (Figure 3D).

Figure 3: Genetic ablation of Slc38a1 selectively impairs Th1 cell proliferative capacity and reducesweight loss in a Th1 cell-mediated IBD model.

A. Change in Thy1.1⁺ frequency over time in primary Cas9-transgenic Th1, Th17, and Treg cells transduced with NTC (WT, black) or Slc38a1-targeted gRNA (ΔSlc38a1, red) (mean±SD, comparison of quadratic least squares fits, dotted lines show corresponding best fit curves, n=3 unique NTC gRNA sequences and n=2 unique Slc38a1-targeted gRNA sequences).

B. Immunoblot of SLC38A1 and ß-actin in WT and ΔSlc38a1 Th1 and Th17 cells (representative blot from n=3 independent experiments). Graph shows quantification of SLC38A1/ß-actin band intensity ratios from immunoblot (paired t-test, n=3 biological replicates).

C. Glutamine uptake in WT and ΔSlc38a1 Th1 cells cultured in 1mM Gln ± V9302 as detected by NMR (repeated measures two-way ANOVA with Tukey’s multiple comparisons test, n=3 biological replicates).

D. Change in WT and ΔSlc38a1 Th1 cells frequencies after 4 days of culture in media with 0.5 or 1mM Gln and 0 or 1 mM Ala. (repeated measures two-way ANOVA with Tukey’s multiple comparisons test, n=3 biological replicates).

E-G. Expression of (E) activation markers CD44, CD62L, and CD25, (F) transcription factors T-bet and FoxP3 in Th1 cells, and (G) RORγT in Th17 cells.

H. IFNγ expression in WT and ΔSlc38a1 Th1 cells (repeated measures two-way ANOVA with Tukey’s multiple comparisons test, n=4–8 biological replicates across 2 independent experiments; flow cytometry histogram shows representative replicate).

Ns denotes p>0.05, * p≤0.05, ** p≤0.01, *** p≤0.001

Further, ΔSlc38a1 Th1 cells showed decreased expression of CD44 suggesting reduced activation, but CD25 was unchanged and CD62L was further downregulated relative to WT (Figure 3E). No changes were detected in Th1 cell expression of lineage-characteristic transcription factor T-bet, although ΔSlc38a1 Th1 cells showed minor increase in aberrant FoxP3 expression (Figure 3F). RORγt was also unchanged in ΔSlc38a1 Th17 cells (Figure 3G), Lastly, Th1 effector cytokine production was unaffected as measured by fraction of IFNγ⁺ cells among Thy1.1⁺ cells (Figure 3H). These data show that specifically ablating SLC38A1 results in impaired glutamine uptake and a proliferative defect in Th1 but not Th17 cells, while largely sparing lineage identity and effector functions.

SLC38A1 deficiency is associated with reduced intracellular glutamine availability with concurrent transcriptomic and metabolomic changes selectively in Th1 cells.

To identify potential mechanisms underlying the proliferative defect in ΔSlc38a1 Th1 cells, transcriptomic and metabolomic changes were surveyed by RNAseq and liquid chromatography-tandem mass spectrometry (LC-MS/MS) (Supplemental Tables 3, 4). WT and ΔSlc38a1 Th1 and Th17 cells were prepared by gRNA transduction followed by five days of expansion in culture to allow for sufficient protein turnover before enrichment by Thy1.1 positive selection. These cells were used for all further studies including RNAseq, metabolomics, NMR, and extracellular flux analyses. As anticipated, WT Th1 and Th17 cells had numerous gene expression and metabolite differences (Supplemental Figure 3A), with overrepresentation of phosphatidylinosityol signaling pathways (Supplemental Figure 3B). Analysis of the RNAseq dataset verified reduced Slc38a1 mRNA levels in both the ΔSlc38a1 Th1 and Th17 cells relative to their WT counterparts (Figure 4A, B). Most strikingly, while this was accompanied by altered expression of numerous other genes in ΔSlc38a1 Th1 cells, no other genes were significantly dependent on Slc38a1 in Th17 cells when corrected for false discovery. Among genes with altered expression in ΔSlc38a1 Th1 cells were multiple SLCs, including the glucose transporters Slc2a1 and Slc2a3, while Slc38a2 was not significantly changed (Supplemental Figure 3C). Given downregulation of glucose and AA transporters, these data suggest that anabolic metabolism was likely reduced in ΔSlc38a1 Th1 cells relative to WT. Comparing glycolysis and TCA cycle genes, multiple glycolysis genes were or nearly significant while TCA cycle genes were unchanged (Supplemental Figure 3D). Indeed, gene set enrichment analysis showed that SLC38A1-deficiency decreased glucose, fructose, and sucrose metabolism (Supplemental Figure 3E). Additionally, inflammatory pathways were also decreased, while multiple nucleic acid pathways were increased.

Figure 4: SLC38A1 deficiency is associated with altered metabolic activity mediated in part by reduced intracellular glutamine pool in Th1 but not Th17 cells.

A-B. Volcano plot showing fold change in mRNA expression between WT and ΔSlc38a1 (A) Th1 and (B) Th17 cells measured by RNAseq (data analyzed with DESeq2⁵⁵, n=3 biological replicates).

C-D. Volcano plot showing fold change in metabolite ion counts in WT and ΔSlc38a1 (C) Th1 and (D) Th17 cells from (A-B) as measured by mass spectrometry (multiple paired t-tests, n=3 biological replicates).

E. Glutamine ion counts in WT and ΔSlc38a1 Th1 and Th17 cells from (C-D) (repeated measures two-way ANOVA with Tukey’s multiple comparisons test, n=3 biological replicates).

F. Joint pathway analysis combining RNAseq data from (A) and mass spectrometry data from (C), identifying differential pathway dependencies between WT and ΔSlc38a1 Th1 cells (data analyzed with MetaboAnalyst v5.0⁵⁶, n=3 biological replicates).

G. Basal rates of glycolysis and oxidative phosphorylation as estimated by the extracellular acidification rate (ECAR) and oxygen consumption rate (OCR), respectively, in WT and ΔSlc38a1 Th1 cells (paired t-tests, n=4 biological replicates with 2–4 technical replicates per condition).

H. Acute change in extracellular flux of WT and ΔSlc38a1 Th1 cells following addition of vehicle or glutamine. (repeated measures two-way ANOVA with Tukey’s multiple comparisons test, n=4 biological replicates).

Ns denotes p>0.05, * p≤0.05, ** p≤0.01, *** p≤0.001, **** p≤0.0001. For multiple comparisons corrections using FDR, nd denotes no discovery and * q≥Q where Q=5% unless otherwise specified.

Metabolomic studies found more metabolites were altered in Th1 cells compared to in Th17 cells upon Slc38a1 deletion (Figure 4C, D). ΔSlc38a1 Th1 cells had increased amounts of citric acid and isocitric acid relative to WT Th1 cells, while having reduced glutamine, pyruvate, and aspartate. In comparison, ΔSlc38a1 Th17 cells had decreased citric acid. Notably, glutamine concentrations were reduced in SLC38A1-deficient Th1 but unchanged in Th17 cells (Figure 4C–E). These data show thatTh1 but not Th17 cells are dependent on SLC38A1 to import and maintain intracellular glutamine. Given previously established dependence of Th17 cells on glutaminolysis⁸, Th17 cells appear to instead rely on other mechanisms to acquire and maintain glutamine levels, such as import via a separate nutrient transporter, interconversion of other AAs, or scavenged from breakdown of macromolecules such from lysosomes or through autophagy. To identify specific metabolic pathways impacted by SLC38A1 deficiency and consequent reduction in intracellular glutamine availability in Th1 cells, genes with significantly changed mRNA transcript levels identified by RNAseq (Padj <0.05) and metabolites identified by mass spectrometry (P <0.1) were combined to perform joint pathway analysis (Figure 4F). This revealed that glycolysis/gluconeogenesis, pyruvate metabolism, amino sugar, glutathione, and the tricarboxylic acid (TCA) cycle were most significantly impacted by SLC38A1 loss in Th1 cells.

SLC38A1 deficiency is associated with altered glutamine-dependent glycolytic and mitochondrial flux.

To test SLC38A1 contributions to metabolic flux, a series of modified substrate oxidation tests were performed. WT and ΔSlc38a1 Th1 cells were plated in base assay media lacking both glutamine and alanine and equilibrated to establish baseline measures. Consistent with RNAseq and metabolomics data suggesting decreased glycolysis, baseline extracellular acidification rate (ECAR) but not oxygen consumption rate (OCR) was decreased with SLC38A1 deficiency (Figure 4G). Next, either vehicle or glutamine was added to the assay media to determine whether substrate availability elicited any acute changes in metabolic activity (Figure 4H). In WT Th1 cells, glutamine addition resulted in increased OCR and decreased ECAR relative to vehicle, consistent with previously published findings³⁷ and the established role of glutamine in anaplerosis to support mitochondrial metabolism. In comparison, ΔSlc38a1 Th1 cells had reduced response, with lesser decreases in ECAR and increases in OCR compared to WT cells. These data support the role of SLC38A1 in Th1 cell glutamine uptake and metabolism.

SLC38A1 supports protein synthesis, reactive oxygen species (ROS) management, hexosamine biosynthesis, and mTORC1 signaling.

In addition to serving directly as building blocks for protein synthesis, imported glutamine and alanine may participate in many different regulatory and metabolic pathways. Despite expressing multiple other potentially redundant transporters, ΔSlc38a1 Th1 cells were less proliferative (Figure 5A). The failure to accumulate in vitro appeared to be due to reduced proliferation and not apoptosis as cell death was unchanged (Figure 5B). Indeed, ΔSlc38a1 Th1 cells incorporated less puromycin per cell relative to WT over the same assay period to indicate reduced rate of total protein synthesis (Figure 5C). Moreover, SLC38A1 loss was also associated with reduced phosphor-S6 in Th1 cells indicating lower mTORC1 activity (Figure 5D), consistent with mTORC1 as a central nutrient sensor.

Figure 5: SLC38A1 supports protein synthesis and metabolism essential to maintain mTORC1 pathway activity.

A. Proliferation as measured by Ki-67 in WT and ΔSlc38a1 Th1 cells (paired t-test, n=4 biological replicates).

B. Fraction of live, early apoptotic, late apoptotic, and dead cells as measured by propidium iodide and Annexin V in WT and ΔSlc38a1 Th1 cells (paired t-test, n=4 biological replicates).

C. Rate of protein synthesis as measured by puromycin incorporation over one hour incubation period in WT and ΔSlc38a1 Th1 and Th17 cells (paired t-test, n=3 biological replicates).

D. mTORC1 activity as measured by phosphor-S6 expression in WT and ΔSlc38a1 Th1 and Th17 cells. mTOR inhibitor, rapamycin, added as control (paired t-test, n=4 biological replicates representative of 2 independent experiments).

E. Rescue by exogenous supplementation of potential downstream metabolites including nucleotide bases (adenosine + guanine, 1mM each), reduced glutathione (rGSH; 2mM), GlcNAc (10mM), dimethyl-ketoglutarate (DMK; 2mM), and sodium pyruvate (NaPyr; 1mM) in ΔSlc38a1 Th1 cells compared to WT on phosphor-S6 expression (mean±SD, one-way ANOVA with Dunnett’s multiple comparisons test, n=4 biological replicates across two independent experiments).

F. Mitochondrial superoxide levels measured by mitoSOX staining in WT and ΔSlc38a1 Th1 cells cultured in 0.1mM Gln with 0 or 1 mM rGSH (ordinary two-way ANOVA with Tukey’s multiple comparisons test, n=3 biological replicates).

Ns denotes p>0.05, * p≤0.05, ** p≤0.01, *** p≤0.001, **** p≤0.0001.

Next, a series of rescue experiments were performed to test if the proliferation defect in ΔSlc38a1 Th1 cells could be restored by provision of metabolites representing other potential fates of glutamine and alanine metabolism. ΔSlc38a1 Th1 cell cultures were supplemented with vehicle, a solution of nucleotide bases, reduced glutathione (rGSH), the amino sugar hexosamine N-acetylglucosamine (GlcNAc), dimethyl-ketoglutarate (DMK) as the membrane permeable substitute for alpha-ketoglutarate, or sodium pyruvate (NaPyr) (Figure 5E). Provision of exogenous rGSH, GlcNAc, or NaPyr restored mTORC1 signaling in ΔSlc38a1 Th1 cell cultures to that of WT as measured by phosphor-S6. Consistent with rescue of mTORC1 signaling, ΔSlc38a1 Th1 cells also had elevated mitochondrial superoxide compared to WT Th1 cells and this was rescued by rGSH (Figure 5F).

SLC38A1 is essential in vivo for Th1 cells in IBD and vaccination.

Given CRISPR screening and in vitro mechanistic results pointing to a selective role for SLC38A1, we next interrogated its role in tissue microenvironments using multiple inflammatory disease models. First, in vivo nutrient availabilities were measured by mass spectrometry of interstitial fluid collected from the inflamed colons and enlarged spleens of IBD model mice. While there were no differences detected in glutamine or glutamate, glucose concentrations were significantly lower in the colon relative to spleen (Supplemental Figure 4A, B). Colonic interstitial fluid also showed decreased ratio of reduced to oxidized glutathione consistent with an oxidative inflammatory microenvironment (Supplemental Figure 4C) that may require T cells to maintain increased antioxidant capacity.

Primary Cas9-transgenic CD4⁺ Th1 cells were transduced with non-targeting (WT), Slc38a1 (ΔSlc38a1), or Slc38a2 (ΔSlc38a2) gRNAs marked by a Thy1.1 reporter. Successfully transduced cells were then enriched by Thy1.1 positivity and injected intraperitoneally into recipient Thy1.2⁺ Rag1^−/− mice to induce gut inflammation. Sham control mice were injected with PBS. Recipient mice were monitored for symptoms of disease progression over the next six weeks³⁸. All groups gained weight until 4 weeks post T cell transfer, after which WT and ΔSlc38a2 groups began to lose weight rapidly, while the ΔSlc38a1 group exhibited comparatively slower progression of weight loss and the sham group maintained weight (Figure 6A, B). These results indicate that SLC38A1 but not SLC38A2, contributed to pathogenic T cell activity in this model.

Figure 6: SLC38A1 deficiency impairs in Th1 cell expansion in vivo models.

A. Effect of SLC38A1 and SLC38A2 deficiency on IBD disease course as measured by percent change in body weight from day of adoptive transfer trended over 42 days. Colitis was induced in Rag1^−/− mice by intraperitoneal injection of WT, ΔSlc38a1, or ΔSlc38a2 Th1 cells generated using CRISPR/Cas9 or PBS control for sham.

B. Percent change in body weight at the end of study (mean±SEM, one-way ANOVA with Dunnett’s multiple comparisons test, n=6 biological replicates for WT, n=6 for ΔSlc38a1, n=5 for ΔSlc38a2, and n=2 for sham).

C. Fraction of transferred CD4⁺ T cells that are WT or ΔSlc38a1 re-isolated from the peripheral blood (PB), spleen, mesenteric lymph nodes (MLN), colonic lamina propria (LP), colonic epithelium (intraepithelial lymphocytes, IEL) in IBD model. Colitis was induced in Rag1^−/− mice by intraperitoneal injection of BFP⁺ WT and GFP⁺ ΔSlc38a1 Th1 cells at 1:1 ratio. (mean±SEM, multiple paired t-tests, n=8 biological replicates).

D-F. Expression of (D) T-bet, RORγt, and FoxP3, (E) IFNγ and TNFa, and (F) phospho-s6 in WT and ΔSlc38a1 Th1 cells re-isolated from the MLN in (C) (mean±SEM, paired t-tests, n=8 biological replicates).

G. Representative flow cytometry plot of the population of BFP⁺ WT and GFP⁺ ΔSlc38a1 Th1 cells mixed at 1:1 ratio and adoptively transferred into Rag1^−/− mice for in vivo vaccine model.

H. Fraction of WT or ΔSlc38a1 re-isolated from the peripheral blood (PB), spleen, and draining lymph nodes (dLNs) 10 days post transfer of (G) (mean±SEM, multiple paired t-tests, n=9 biological replicates).

I-K. Expression of (I) Ki-67, (J) T-bet and FoxP3, and (K) IFNγ in WT and ΔSlc38a1 Th1 cells re-isolated from the dLNs in (H) (mean±SEM, paired t-tests, n=9 biological replicates).

Ns denotes p>0.05, * p≤0.05, ** p≤0.01, *** p≤0.001, **** p≤0.0001.

Next, the same IBD model was performed in a competitive assay format to assess WT and Slc38a1 Th1 cells in vivo within the same tissue microenvironment. To do so, primary Cas9-transgenic Th1 cells were transduced with either NTC (WT) or Slc38a1-targeted gRNA with BFP or GFP markers, respectively, mixed to achieve equal fractions, and intraperitoneally injected into recipient Rag1^−/− mice. After 6 weeks, ΔSlc38a1 GFP⁺ cells were less abundant than WT BFP⁺ controls, particularly in the intraepithelial lymphocyte (IEL) fraction (Figure 6C). The ΔSlc38a1 GFP⁺ population also had reduced Tbet, but unchanged RORγT, and increased FoxP3 expression compared to the WT BFP⁺ population (Figure 6D). Consistent with impaired Th1 cell activity, IFNγ, TNFα (Figure 6E), and phosphor-S6 (Figure 6F) were all decreased with SLC38A1 deficiency. Despite the delay in weight loss, SLC38A1 deficiency ultimately did not protect from disease. WT, SLC38A1, and SLC38A2-deficient cells all developed colitis with marked edema grossly and abundant CD3⁺ cell infiltration within the colonic mucosa with crypt architectural distortion microscopically at the experimental endpoint (Supplemental Figure 5A). There were no statistically significant differences across groups at the experimental end point in the total number of CD4⁺ T cells recovered from the mesenteric lymph nodes (MLNs) or spleens (Supplemental Figure 5B). Nevertheless, the Thy1.1⁺ fractions within these MLN and splenic CD4⁺ T cell compartments contracted relative to the population transferred on day 0 in both the ΔSlc38a1 and ΔSlc38a2 groups when normalized to WT (Supplemental Figure 5C).

The role of SLC38A1 was next tested in a model of acute immunization using the same competitive assay approach as before. Primary OT-II/Cas9-transgenic Th1 cells were transduced to generate WT BFP+ and ΔSlc38a1 GFP⁺ cells, mixed 1:1, and adoptively transferred into recipient Rag1^−/− mice and stimulated in vivo by subcutaneous injection with OVA/CFA emulsion (Figure 6G). As in the chronic inflammation models, ΔSlc38a1 Th1 cells demonstrated competitive disadvantage, with reduced frequency and proliferation rates relative to WT (Figure 6H, I). Here, Tbet was unchanged, FoxP3 was increased, and IFNγ was decreased with SLC38A1 deficiency (Figure 6J, K), similar to in vitro and IBD model results. These effects may have been partly compensated by other glutamine transporters, as SLC38A2 was elevated in ΔSlc38a1 Th1 cells from peripheral blood, spleen, and lymph nodes (Supplemental Figure 5D).

Pharmacologic SLC38 inhibition selectively impairs Th1 cell expansion and function in the EAE model but has no measurable effect in the allergic-airway disease model.

To functionally establish the effects of SLC38 inhibition, primary CD4⁺ T cells activated with anti-CD3/CD28 antibodies and polarized to Th1, Th17, or Treg cells in the presence of 0, 1, or 5mM alpha-methylaminoisobutyric acid (MeAIB), a non-selective SLC38A1 and SLC38A2 competitive inhibitor^39,40. Concentrations were selected for MeAIB treatment to capture changes in cell count and activation as measured by CD44 without significantly impairing cell viability in Th0 cells (Supplemental Figures 6A–C). MeAIB treatment impaired proliferation across subsets in a dose-dependent manner (Supplemental Figure 6D, E) but selectively impaired Th1 cell viability (Supplemental Figure 6F). MeAIB treatment also reduced activation as measured by CD25 expression (Supplemental Figure 6G). Expression of lineage-characterizing transcription factors T-bet and RORγt were generally reduced with MeAIB treatment, though the expression of FoxP3 was mildly increased in Th17 cells and unchanged in Treg cells (Supplemental Figure 6H). Further, MeAIB-treated Th1 and Th17 cells had reduced expression of effector cytokines IFNγ and IL-17, respectively (Supplemental Figure 6I).

Next, Th1 and Treg cells treated with either vehicle or 5mM MeAIB were tested using metabolic extracellular flux assays. MeAIB treatment had no effect on basal or maximal ECAR in Th1 cells, while both the basal and maximal ECAR were significantly reduced in Treg cells (Supplemental Figure 7A, B). MeAIB treatment lowered basal OCR in Th1 cells, though this difference was extinguished under stressed conditions. Both basal and maximal OCR were unchanged with SLC38 inhibition in Treg cells (Supplemental Figure 7C, D). However, mitochondrial ROS was increased selectively in Th1 cells as measured by MitoSox (Supplemental Figure 7E).

Next, we sought to evaluate the therapeutic potential efficacy of SLC38A1/2 pharmacologic inhibition in Th1 and Th17 cell-mediated inflammatory diseases. To test a Th17 cell-mediated model, neutrophilic allergic-airway disease was induced in WT mice by intranasal administration of house-dustmite (HDM)/lipopolysaccharide (LPS) solution over two weeks to promote sensitization with Th17 cell recruitment⁴¹. The mice were treated daily starting one-week preceding disease induction with either vehicle or 3mg MeAIB subcutaneous injection daily until end of study. There were no differences in the total number of live cells Figure 7A), CD4⁺ cell frequencies (Figure 7B), or prevalence of CD4⁺IL-17⁺ cells (Figure 7C) in the lungs of control and treatment arms. Moreover, there were no statistically significant differences in the immune cell composition of the BALF used to characterize disease burden, including lymphocytes, neutrophils, eosinophils, and macrophages (Figure 7D).

Figure 7: SLC38A1/2 inhibition in vivo has no effect on Th17-cell driven allergic airway disease model as predicted by in vivo CRISPR screen.

A. Total live cell count measured from the lung homogenate collected at the end of study from mice with allergic airway disease Lung inflammation induced by intranasal administration of HDM/LPS over 2 weeks, treated daily with either vehicle or MeAIB from 7 days prior to initial sensitization until end of study (mean±SEM, unpaired t-test, n=8 biological replicates for HDM/LPS, n=2 each for PBS control).

B-C. Frequencies of (B) CD4⁺ T cell among total live cell population and (C) IL-17⁺ cells among CD4⁺ T cells in the lung homogenate from (A) (mean±SEM, multiple unpaired t tests, n=8 biological replicates for HDM/LPS, n=2 each for PBS controls).

D. Cell count of neutrophils, lymphocytes, eosinophils and macrophages in the BALF collected from (A) (mean±SEM, multiple unpaired t tests corrected for multiple comparisons by FDR, Q=5%, n=8 biological replicates for HDM/LPS, n=2 each for PBS controls).

E Average clinical score over time in mice with EAE induced by subcutaneous injection of MOG/CFA emulsion and pertussis, with daily treatment with either vehicle or 3mg MeAIB from 7 days prior to disease induction until end of study (mean±SEM, n=10 biological replicates for each condition). F Maximum clinical score attained in EAE mice from (A) (mean±SEM, unpaired one-tailed Mann-Whitney test, n=10 biological replicates for each condition).

G. Day of onset of clinical symptoms in EAE mice from (A) (mean±SEM, unpaired one-tailed Mann-Whitney test, n=10 biological replicates for each condition).

H-I. CD4⁺ T cell fraction and count in the (H) spinal cords and (I) brains collected at the end of study from mice in (E) (mean±SEM, multiple Mann-Whitney tests corrected for multiple comparisons by FDR, Q=5%, n=7 biological replicates for vehicle and n=6 for MeAIB).

J-K. Frequency of (J) T-bet⁺, RORγt⁺, and FoxP3⁺ and (K) IL-17⁺, IFNγ⁺, and IL-17⁺IFNy⁺ cells among total CD4⁺ T cell population recovered from the spinal cord of EAE mice from (A) (mean±SEM, multiple unpaired t-tests corrected for multiple comparisons by FDR, Q=5%, n=7 biological replicates or vehicle and 8 for MeAIB).

To test a Th1 cell-mediated model, EAE was induced in WT mice by subcutaneous injection of MOG/CFA emulsion augmented with pertussis. Parallel to the Th17 cell-mediated lung inflammation model, the mice were treated daily with vehicle or 3mg MeAIB daily starting one-week prior to disease induction. Mice were scored daily according to clinical symptomology to monitor disease progression until humane endpoints were reached at approximately 5 weeks post induction (Figure 7E). While there was no difference in the maximal disease severity (Figure 7F), disease onset was delayed in mice treated with MeAIB compared to vehicle (Figure 7G). At the end of the study, significantly fewer pathogenic CD4⁺ and CD8⁺ T cell were found infiltrating the spinal cords and brains (Figure 7H, I). Infiltrating CD4⁺ T cells showed similar expression of Tbet, FoxP3, IFNγ, and Il-17, but mildly decreased RORγt with MeAIB treatment compared to control (Figure 7J, K). Together, these data show T cell metabolic requirements are subset- and tissue microenvironment-dependent and SLC38A1 plays a selective role in Th1 cell proliferation or recruitment to diseased sites but not differentiation in vivo.

DISCUSSION

AA uptake and metabolism are critical to fuel and regulate T cell activities in health and disease. In this study, we investigated AA and nutrient transporter requirements in primary CD4⁺ T cell subsets to identify specific dependencies that may be exploited as an approach to immunotherapy. In vivo CRISPR screens revealed that pathogenic Th1 cells but not Th17 cells require the small neutral AA transporter SLC38A1 in vivo in the setting of EAE-associated CNS infiltration and inflammation. Notably, Th1 cells required SLC38A1 despite concomitant expression of multiple other transporters with overlapping substrate profiles. In contrast, SLC38A1 was dispensable for sustained expansion and persistence of Th0 cells both in vitro as well as in vivo in a model of inflammatory lung disease. Mechanistically, SLC38A1 loss in Th1 cells was associated with decreased intracellular glutamine, dampened mTORC1 signaling, alter glycolytic and mitochondrial metabolism, impaired ROS management and hexosamine biosynthesis, and slowed protein synthesis. Together, these results demonstrate that, while Teff cells require uptake of extracellular glutamine and alanine to meet biosynthetic and bioenergetic demands, specific transporter dependencies may vary between T cell subsets as well as across tissue microenvironments.

The family of SLC38A1/2 transporters have been described to preferentially carry glutamine^40,42,43 and or alanine^15,44 depending on experimental conditions. While both alanine and glutamine may be de novo synthesized intracellularly, additional supply may be needed from the extracellular environment when metabolic demands exceed synthetic capacity. Functionally, extracellular glutamine is required for Teff cell activation, expansion, fate, and pro-inflammatory functions³, whereas glutamine deprivation promotes generation of Treg cells with suppressive capacity⁵. This involves metabolic reprogramming mediated by TCR-induced upregulation of mTORC1 activity that is accompanied by rapid upregulation of multiple glutamine transporters including SLC7A5, SLC1A5, SLC38A1, and SLC38A2³. Upon import, glutamine is converted to glutamate by glutaminase (GLS), and GLS deficiency attenuates Th17 cell differentiation while enhancing Th1 cell differentiation and effector functions through altered epigenetic and ROS regulation⁸. Glutamine metabolism and CD4⁺ T cell subset activities are thus tightly intertwined. How SLC38A1/2 transporters fit into this intricate network of metabolic regulation in relation to T cell activity had not previously been directly established.

As active transporters, SLC38A1/2 transporters can drive steep concentration gradients and are responsive to microenvironmental factors including pH, hypoxia, tonicity, and nutritional deficiencies^45,46. One feature that distinguishes SLC38A1 and SLC38A2 from SLC1A5 is that the former are Na⁺-dependent symporters while the latter is a Na⁺-dependent antiporter⁴⁷. These transporters may contribute to cellular AA homeostasis in distinct ways based on their transport kinetics, with SLC1A5 acting in conjunction with SLC7A5/SLC3A2 to maintain a balanced intracellular pool of AAs and SLC38A1 and SLC38A2 functioning to specifically accumulate their preferred substrates glutamine and alanine⁴⁷. In the present study, ΔSlc38a1 Th1 cells had impaired proliferation and survival with associated depression in mTORC1 activity despite intact SLC1A5 expression. Notably, this phenotypic pattern is opposite of that observed for SLC1A5⁴. The same study also showed that SLC38A1 is similarly unable to compensate for SLC1A5 deficiency to support Th1 and Th17 cell differentiation and effector functions⁴. Further, SLC1A5 was required for mTORC1 signaling for naïve T cells but not activated Teff cells. These data suggest that these transporters are non-redundant even though they have shared substrates, and instead serve complementary functions depending on cellular expression, inflammation type, and tissue location.

In the absence of SLC38A1, extracellular GlcNAc partially rescued both proliferation and mTORC1 activity to suggest that glutamine uptake is crucial to support hexosamine biosynthesis. This pathway, like mTORC1, functions as a global nutrient sensor and integrator to calibrate cellular activities to reflect cellular nutrient supplies and demands. In the de novo synthesis pathway, glucose, glutamine, uridine triphosphate (UTP), and acetyl-CoA are combined to produce UDP-GlcNAc, whereas the salvage pathway may instead use imported or recycled glucosamine or GlcNAc. UDP-GlcNAc in turn serves as the substrate for N- and O-glycosylation reactions that can modify and regulate the activity of a wide array of proteins, lipids, and nucleic acids to induce global changes in cellular activities. This includes the expression and activity of essential T cell regulators including TCR, CD4, CD25, MYC⁴⁸, NFAT and NFkB⁴⁹, and PD-1⁵⁰. Previous studies have shown that adequate expansion of the intracellular UDP-GlcNAc pool upon TCR stimulation is dependent on the uptake of extracellular glucose and glutamine⁵¹, which aligns with our findings that SLC38A1 deficiency, associated with decreased intracellular glutamine and downregulation of key glucose transporters, can be partially compensated by GlcNAc supplementation. Consistent with a specific role for this pathway in Th1 over Th17 cells, a study of inborn errors of metabolism identified GFPT1, the first and rate limited enzyme in hexosamine synthesis, to be selectively essential for Th1 cells³⁴. Also, the lesser dependence than Th17 cells on GLS⁸ highlights the potential importance of glutamine processing into the hexosamine pathway for Th1 cells. However, hexosamine supplementation can suppress T cell receptor signaling and Th1 differentiation^52,53. This discrepancy may reflect a need for balance in de novo and salvage GlcNAc synthesis or a dosage effect relative to SLC38A1-dependent synthesis rates. Further investigation is warranted to better understand how immune cell dependencies on de novo hexosamine synthesis through targeting nutrient transport and salvage through dietary intake of glucosamine may be exploited in the context of anti-inflammatory and anti-tumor immunotherapies.

Findings from this study also underscore the utility of small-scale targeted in vivo CRISPR screens using the most relevant disease model to identify physiologically relevant drug target candidates. This is highlighted by the distinct transporter requirements identified in this study across in vitro and in vivo platforms as well as between disease models involving CNS, pulmonary, and enteric inflammation. In sum, these data suggest that, while having overlapping substrate profiles, nutrient transporters may serve distinct roles to regulate T cell activities. Therefore, pathogenic T cell populations may be distinguished and selectively targeted based on their cell-type and tissue site-dependent requirements for specific nutrient transporter as an approach to immunotherapy.

Limitations of the Study

In this study, the role of SLC38A1 was investigated by a combination of genetic and pharmacological approaches using CRISPR/Cas9-mediated gene editing and MeAIB treatment, respectively. The method employed for genetic ablation of SLC38A1 involved gRNA delivery into activated Cas9-transgenic primary T cells by retroviral transduction, which excluded the possibility of testing transporter requirements in naïve T cells and for activation events and requirements immediately following TCR-stimulation. In contrast, the pharmacologic approach provided the advantage of initiating treatment prior to T cell activation. However, this was limited by the lack of specificity in that MeAIB is a substrate for all system A transporters that include SLC38A1, SLC38A2, and SLC38A4. Thus, future studies are needed to develop models for inducible and conditional ablation of SLC38A1 to facilitate further characterization. Further, while the scope of the present study was limited to characterization of Th1 and Th17 cell subsets, the role of SLC38A1 in Th2 and Treg cell subsets as well as in CD8⁺ T cell populations in association to SLC1A5 and SLC38A2 is warranted given related findings in previous studies and the distinct transporter expression patterns observed among the cell types in health and in disease.

RESOURCE AVAILABILITY

Lead contact

Further information and request for resources and regents should be directed to and will be fulfilled by lead contact, Jeffrey C. Rathmell (rathmell@uchicago.edu).

Materials availability

Any materials may be available from the lead contact upon appropriate MTA.

Data and code availability

RNA-seq gene expression data are reported in the Gene Expression Omnibus (NCBI GEO: GSE190131). CRISPR screening data and detailed library information are available: https://functionalimmunogenomics.shinyapps.io/crispr/.³⁴ Original western blot images and values that were used to create all graphs in this paper are included in Data S1. Data S1 contains high-resolution western blot scans and unprocessed data underlying items in the manuscript related to Figures 1–7 and Supplemental Figures S1–S7.

STAR METHODS

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Mice:

All experiments were performed at Vanderbilt University animal facility in accordance with Institutional Animal Care and Utilization Committee (IACUC)-approved protocols and conformed to all relevant regulatory standards. Mice were housed in pathogen-free facilities with 12-hour light cycles in ventilated cages with ad libitum standard chow and water and at most 5 animals per cage. Eight- to sixteen-week-old male and female mice were used for all animal experiments. All mice were obtained from Jackson Laboratory and were treatment-naive until the start of study. 2D2 mice (C57BL/6-Tg(Tcra2D2,Tcrb2D2)1Kuch/J, JAX Strain#: 006912) were crossed to Cas9 mice (B6J.129(Cg)-Gt(ROSA)26Sortm1.1(CAG-cas9*,-EGFP)Fezh/J, JAX Strain#: 006912) to generate 2D2 Cas9 double-transgenic strain, and OT-II mice (B6.Cg-Tg(TcraTcrb)425Cbn/J, JAX Strain#: 004194) were crossed to Cas9 mice to generate OT-II Cas9 double-transgenic strain. Animals were genotyped for transgenic TCR and Cas9 allele.

Cell Lines:

Plat-E retroviral packaging cell line was maintained at 37°C with 5% CO₂ in DMEM media supplemented with 10% FBS, 100U/mL penicillin/streptomycin, 1μg/mL puromycin, and 10μg/mL blasticidin to maintain expression of viral packaging genes.

Method Details

In vitro mouse CD4+ T cell activation and differentiation:

Primary murine CD4⁺ T cells were isolated from the spleens and lymph nodes of mice using a CD4 negative isolation kit according to the manufacturer’s instructions. The cells were cultured at 37°C with 5% CO₂ in RPMI-1640 media supplemented with 10mM HEPES, 50μM 2-mercaptoethanol, 100U/mL penicillin/streptomycin, and 2mM AA, unless otherwise stated. AA subtracted media were produced by adding back all AAs except the subtracted one at concentrations specified in the RPMI-1640 media formulation to modified RPMI-1640 media lacking all AAs (200mg/L L-arginine, 56.82mg/L L-asparagine•H2O, 20mg/L L-aspartic acid, 65.2mg/L L-cystine•2HCl, 20mg/L L-glutamic acid, 300mg/L L-AA, 10mg/L glycine, 15mg/L L-histidine, 20mg/L hydroxyl-L-proline, 50mg/L L-isoleucine, 50mg/L L-leucine, 40mg/L L-lysine•HCl, 15mg/L L-methionine, 15mg/L L-phenylalanine, 20mg/L L-proline, 30mg/L L-serine, 20mg/L L-threonine, 5mg/L L-tryptophan, 28.83mg/L L-tyrosine•2Na•2H₂O, 20mg/L L-valine). The pH of the final media was adjusted to 7.2.

Primary CD4⁺ T cells from wildtype and Cas9-transgenic mice were activated using plate-bound anti-CD3 (3μg/mL) and anti-CD28 (5μg/mL) antibodies at 1 million cells/well in a 24-well plate. OT-II/Cas9 double-transgenic T cells were activated with splenocytes irradiated at 30 Gy and OVA_323–339 peptide (10μg/mL), and 2D2/Cas9 double-transgenic T cells with MOG_35–55 peptide (10μg/mL). T cells were cultured for 4 days with subset-specific cytokines and blocking antibodies to promote differentiation - Th1 cells: IL-12p70 (10ng/mL), IL-2 (100U/mL), anti-IL-4 (10μg/mL), anti-IFNγ (1μg/mL); Th17 cells: IL-6 (50ng/mL), TGFβ (1ng/mL), IL-23 (20ng/mL), IL-1β (10ng/mL), anti-IL-4 (10μg/mL), anti-IFNγ (10μg/mL); Treg cells: TGFβ (1.5ng/mL), IL-2 (100U/mL), anti-IL-4 (10μg/mL), anti-IFNγ (10μg/mL). MeAIB (SLC38A1/2 inhibitor) was dosed at 1mM and 5mM for in vitro studies.

Flow Cytometry:

Cells were first stained with viability dye and antibodies for pertinent surface markers for all experiments. For intracellular and transcription factor stains, cells were then fixed and permeabilized using appropriate kits. For cytokines, cells were stimulated with 1μg/mL 12-myristate 13-acetate (PMA), 750ng/mL ionomycin, and GolgiPlug for four hours before fixation. Unstimulated cells served as negative control. All dilutions and washes were performed in PBS with 10% FBS.

In vitro single gene knockout using CRISPR/Cas9:

MSCV-U6-gRNA scaffold-IRES-Thy1.1 was cloned by excising the U6 promoter and gRNA scaffold from the pMx-U6-gRNA scaffold-PGK-GFP vector⁵⁷ and inserting into the MSCV-IRES-Thy1.1 DEST vector (Addgene #17442)⁵⁸. NTC and targeted gRNA sequences were referenced from the Mouse CRISPR Knockout Pooled Libraries Brie (Addgene #73633)²⁷ and GeCKO v2 (Addgene #1000000052)²⁸, and cloned into the new MSCV-U6-gRNA scaffold-IRES-Thy1.1 vector following the protocol made publically available by the Zhang Lab. The gRNA sequences used are as follows: NTC gRNA#1 AAAAAGTCCGCGATTACGTC (Brie), NTC gRNA#2 ATTGTTCGACCGTCTACGGG (GeCKO v2, MGLibA_66412), NTC gRNA#3 ACCCATCGGGTGCGATATGG (GeCKO v2, MGLibA_66413), Slc38a1 gRNA#1 TGCATGGTGTATGAGAAGCT (Brie), Slc38a1 gRNA#2 AGATTGGCAGGACGGACGGG (Brie), and Slc38a2 gRNA#1 (CTCAAGACTGCCAACGAAGG, Brie). For experiments with single gRNA in multiple biological replicates, NTC gRNA#1, Slc38a1 gRNA#1, and Slc38a2 gRNA#1 were used to generate WT, ΔSlc38a1, and ΔSlc38a2 cells, respectively.

Briefly, Plat-E retroviral packaging cell line was transfected using Polyplus jetPRIME DNA and siRNA transfection reagent with the gRNA sequence-cloned expression. The viral supernatants were collected, filtered, and spun onto retronectin-coated plates at 2000xg for 2 hours at 32°C. In parallel, Cas9-transgenic CD4⁺ T cells were activated and cultured for 48 hours, transferred to the prepared virus-bound plates, and centrifuged at 2000xg for 15 minutes at 32°C. Transduced cells were identified by Thy1.1 positivity by flow cytometry, and enriched for using CD90.1 positive selection kit for downstream applications including RNAseq and metabolomics.

In vitro CRISPR screening:

Custom SLC transporter (Supplemental Table 1) and glutamine metabolism (Supplemental Table 2) gRNA libraries were curated by referencing the Brie and GeCKO v2 libraries. Pooled plasmid libraries were prepared following published methods^59,60 and the in vitro and in vivo CRISPR screen using the lung-inflammation model were performed in primary CD4⁺ T cells as previously described²⁹. Briefly, each gRNA library was synthesized as an oligonucleotide pool he following sequence: GGAAAGGACGAAACACCGXXXXXXXXXXXXXXXXXXXXGTTTTAGAGCTAGAAATAGCAAGTTAAAATAAGGC, where Xs denote the variable gRNA sequence. The oligo pool was bulk cloned into the pMx-U6-gRNA scaffold-PGK-GFP vector by PCR using Array primers (Supplemental Table 5) followed by Gibson Assembly. The resultant plasmid pool was amplified using ElectroMAX DH10B Cells, with greater than 50-fold coverage of the library, and packaged in retrovirus. Cas9-transgenic CD4⁺ T cells were activated for 48 hours and transduced at a multiplicity of infection (MOI) of 0.4. Pre- and post-selection cell samples were collected as specified for each screen.

Next, genomic DNA was isolated from cells and gRNA sequences were amplified by two rounds of PCR using Adapter primers followed by barcoded sequencing primers (Supplemental Table 5). Amplicons were sequenced to obtain 150bp paired-end reads on the Illumina NovaSeq 6000 platform. At least 1000-fold representation of the library was maintained throughout the assay. FASTQ files were analyzed using the Model-based Analysis of Genome-wide CRISPR/Cas9 Knockout (MAGeCK v0.5.0.3) method⁵⁴.

In vivo CRISPR screening:

The Th1- and Th17-cell driven EAE models were developed by modifying previously published methods for EAE induction by adoptive transfer of differentiated 2D2-transgenic CD4⁺ T cells³³. First, CD4⁺ T cells were isolated from the spleen and lymph nodes of 2D2/Cas9 double-transgenic mice and activated with splenocytes irradiated at 30 Gy and MOG_35–55 peptide (10μg/mL). Cells were cultured for 3 days with cytokines and blocking antibodies to promote either Th1 or Th17 cell differentiation. Three days post activation, cells were split in fresh media with IL-2 for Th1 cells and IL-23 for Th17 cells and cultured for an additional three days. Six days post activation, cells were re-stimulated with plate-bound anti-CD3 (3μg/mL) and anti-CD28 (5μg/mL) antibodies and transduced with the gRNA library as above. 24 hours post transduction, a sample of the cells was collected for the early timepoint and approximately 10 million live cells were adoptively transferred into separate Rag1^−/− mice for each subset by tail vein injection. On day 0 and 1 post adoptive transfer, the recipient mice were i.p. injected with 150ng PTX. Mice were monitored daily for clinical symptoms of disease progression. Once mice began exhibiting signs of bilateral hind leg paralysis and/or severe ataxia, they were sacrificed for brain and spinal cord collection. The tissues were mechanically dissociated and digested with 300U/mL Collagenase IA and 50U/mL DNase I at 37°C for 45 minutes, and then filtered through a 70μm filter to obtain a single-cell suspension. The cells were subsequently layered on a 18.6%/62.4% Percoll gradient and centrifuged at 2400rpm for 30 minutes at room temperature to partially purify the T cells for the late timepoint collection. Cells were processed, sequenced, and analyzed in the same manner as described for in vitro screening.

The in vivo CRISPR screen in the inflammatory lung disease model was performed following published methods²⁹. OT-II/Cas9 double transgenic CD4⁺ T cells were activated with irradiated splenocytes and OVA_323–339 peptide, transduced with the gRNA library, and adoptively transferred into Rag1^−/− mice. Recipient mice were intranasally administered ovalbumin protein on days 1, 3, 5, and 7 post adoptive transfer, and sacrificed on day 8 for lung collection. Lungs were mechanically dissociated using a gentleMACS Dissociator (Miltenyi Biotec), digested with 300U/mL Collagenase IA and 50U/mL DNase I at 37°C for 45 minutes, and then filtered through a 70μm filter to obtain a single-cell suspension. CD4⁺ T cells were isolated using a positive selection kit according to the manufacturer’s instructions for the late timepoint collection.

RNAseq analysis:

Bulk RNA was isolated using the RNeasy Mini Kit. mRNA enrichment and cDNA library preparation was performed using the stranded mRNA (polyA-selected)library preparation kit (NEB). cDNA was sequenced to obtain 150bp paired-end reads on the Illumina NovaSeq 6000 platform targeting an average of 50 million reads per sample. Demultiplexed FASTQ files were analyzed as follows. Adapters were first trimmed by Cutadapt (v2.10). Reads were then mapped to the mouse genome mm10 using STAR (v2.7.3a) and quantified by featureCounts (v2.0.0). DESeq2 (v.1.24.0)⁵⁵ was used to detect differential expression between the two groups.

Extracellular flux analysis:

Extracellular flux analyses were performed with the Seahorse XFe96 Analyzer (Agilent). Assay media were prepared by supplementing Seahorse XF RPMI Medium pH 7.4 with 10 mM glucose, 1 mM sodium pyruvate, and 2 mM glutamine, unless otherwise stated. Cells were seeded on 96-well cell culture microplates coated with Cell-Tak (1mg/mL) at 150,000 live cells per well with technical replicates for each biological replicate. The Glycolytic Stress Test was performed The Mito Stress Test was performed with 1.5μM oligomycin A, 1.5μM FCCP, and 0.5μM rotenone/antimycin A final concentrations. Data were analyzed in Agilent Wave software v2.6.

Immunoblotting:

Cells were lysed on ice for 30 minutes with base lysis buffer containing 1% IGEPAL CA-630, 200mM NaCl, 50mM Tris pH8.0, and supplemented with the protease inhibitors aprotinin (5ug/mL), leupeptin (5ug/mL), sodium fluoride (0.9mM), dithiothreitol (DTT, 1mM), sodium vanadate (1mM), and ß-glycerophosphate (20mM). Lysates were centrifuged for 15 minutes at 4°C to recover supernatant and quantified for protein concentration using Protein Assay Dye Reagent Concentrate. 40ug of protein was loaded per well in Mini-PROTEAN Precast Polyacrylamide Gels (Bio-Rad) for electrophoresis using. Western blotting was performed using low fluorescence PVDF membrane (Bio-Rad). Transfer was accomplished using 1X Towbin Transfer Buffer Containing 20% methanol at 300mA for 1 hour. Blots were blocked for 1 hour using Intercept (TBS) Blocking Buffer (LI-COR Biosciences) and incubated overnight at 4°C with primary antibody overnight at 4°C (1:1000 for anti-ß-actin, 1:1000 for anti-SLC38A1). Blots were incubated for 1 hour at room temperature with IRDye Secondary antibodies and visualized by near infrared fluorescence via Li-COR Odyssey CLx imager.

Mass Spectrometry:

Metabolites were extracted from cells by adding cold 80% methanol and incubating at −80°C for 10 minutes, followed by centrifugation at 10,000xg for 10 minutes at 4°C. Supernatants were collected and lyophilized by speedvac. The quantity of metabolite fraction analyzed was normalized to cell number. Liquid chromatography-based targeted tandem mass spectrometry (LC-MS/MS)-based metabolomics were performed and the data analyzed as previously described^61–63. The raw ion counts were median-normalized for each metabolite, and differences in metabolite concentrations were determined by calculating the Log2 fold change for each biological replicate (ΔSlc38a1/WT). The volcano plots were generated by taking the average of the biological replicates, and significance testing was performed using multiple paired two-tailed t-tests.

Glutamine use:

Glutamine uptake and consumption was measured by calculating glutamine depletion from culture media (Promega, J8021).

EAE model with in vivo MeAIB treatment:

Female C57BL/6 mice aged 8 weeks were i.p. injected with either PBS or 3mg MeAIB in PBS daily from day 0 until end of study. All mice were injected subcutaneously with 0.2mL MOG/CFA emulsion (Hooke Laboratories) on day 7, and i.p. with 100ng PTX on days 7 and 8 to induce EAE. Mice were monitored and scored daily for clinical symptoms according to the following criteria: 0- no symptoms, 0.5- partial loss of tail tonicity, 1-complete loss of tail tonicity, 2- unilateral hind limb paresis, 2.5- unilateral hind limb paralysis, 3-partial bilateral hind limb paralysis, 3.5- complete bilateral hind limb paralysis (humane endpoint), 4- forelimb paresis, and 5- moribund/death. At the end of study, surviving mice were sacrificed for spleen, spinal cord, and brain collection for further analysis by flow cytometry. Tissues were dissociated, digested, and purified as described above.

Allergic airway disease model with in vivo MeAIB treatment:

Parallel to the EAE model, female C57BL/6 mice aged 8 weeks were treated with daily injection of i.p. PBS or 3mg MeAIB in PBS from day 0 until end of study. On days 7, 14, and 21, mice were intranasally administered 50 μl PBS or 100 μg HDM Dermatophagoides pteronyssinus extract and 0.1 μg LPS from Escherichia coli 0111:B4 in 50 μl PBS. On day 22, mice were sacrificed by fatal dose of i.p. phenobarbital. BALF was collected by instilling and withdrawing 800μL of saline through a tracheostomy tube into the lungs. Cells from the BALF were then adhered to a slide, stained using Epredia Richard-Allan Scientific Three-Step Stain Kit, and identified as eosinophils, neutrophils, lymphocytes, and macrophages using light microscopy for quantification as previously described⁶⁴. Lungs were dissociated to obtain single-cell suspensions as above for analysis by flow cytometry

Th1-cell driven IBD colitis model with CRISPR/Cas9-mediated Slc38a1 knockout:

WT and ΔSlc38a1 Th1 cells were prepared following in vitro single-gene knockout using CRISPR/Cas9 protocol using NTC gRNA#1, Slc38a1 gRNA#1, and Slc38a2 gRNA#1. On day 7 post T cell activation, transduced cells were isolated by CD90.1 positive selection and 400,000 viable cells were transferred into male Rag1^−/− mice aged 8 weeks by i.p. injection to induce IBD colitis. Mice were weighed twice weekly to monitor disease progression and sacrificed at 7 weeks post transfer to collect spleen and mesenteric lymph nodes for T cell phenotyping by flow cytometry. Colons were also collected and fixed in 10% formalin for histology.

Th1-cell OVA immunization model with CRISPR/Cas9-mediated Slc38a1 knockout:

Purified CD4 T cells from Ovalbumin (OVA) specific OT-II Cas9-transgenic were activated in Th1-skewing conditions and prepared following in vitro single-gene knockout using CRISPR/Cas9 protocol using NTC gRNA#1 with BFP or Slc38a1 gRNA#1 with GFP. Cells were mixed 1:1 on day 3 and injected intravenously into Rag1^−/− recipients, which were immunized with an OVA/CFA emulsion (Hooke Laboratories EK-0301). T cells were then analyzed 10 days after transfer.

Quantification and Statistical Analysis

Statistical analyses were performed with Prism software (v10). Significance is indicated as follows: ns denotes p>0.05, * p≤0.05, ** p≤0.01, *** p≤0.001, **** p≤0.0001. For multiple comparisons corrections using FDR, nd denotes no discovery, * q≥Q where Q=5% unless otherwise specified. Error bars show mean ± standard deviation (SD) unless otherwise indicated for standard error of the mean (SEM). Sample sizes were chosen based on previous studies. Flow cytometric plots shown are representative of biological replicates.

Supplementary Material

1

Data S1. Unprocessed source data underlying all blots and graphs. Related to Figures 1–7 and Supplemental Figures 1–7

2

Supplemental Table 1: SLC gRNA library (related to Figure 1)

3

Supplemental Table 2: Glutamine metabolism gRNA library (related to Figure 2)

4

Supplemental Table 3: Metabolomics data - positive mode (related to Figure 4)

5

Supplemental Table 4: Metabolomics data - negative mode (related to Figure 4)

6

Supplemental Table 5: Primers for CRISPR gRNA library prep (related to STAR Methods)

7

Document S1. Supplemental Figures and Legends S1-S7

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Anti-IFNγ	Thermo Fisher Scientific	Cat# 16-7311-38, RRID:AB_2637490
Anti-IL-4	Thermo Fisher Scientific	Cat# 16-7041-95, RRID:AB_2573101
Mouse Anti-CD3e	Thermo Fisher Scientific	Cat# 16-0031-86, RRID:AB_468849
Mouse Anti-CD28	Thermo Fisher Scientific	Cat# 16-0281-86, RRID:AB_468923
Anti-IFNγ APC	BD Biosciences	Cat# 554413, RRID:AB_398551
Anti-IL-17A PE	Thermo Fisher Scientific	Cat# 12-7177-81, RRID:AB_763582
Anti-CD25 e450	Thermo Fisher Scientific	Cat# 48-0251-82, RRID:AB_10671550
Anti-CD44 PE	Thermo Fisher Scientific	Cat# 12-0441-82, RRID:AB_465664
Anti-CD62L APC	Thermo Fisher Scientific	Cat# 17-0621-82, RRID:AB_469410
Anti-CD4 eFluor 450	Thermo Fisher Scientific	Cat# 48-0041-82, RRID:AB_10718983
Anti-CD4 PE	BioLegend	Cat# 100512, RRID:AB_312715
Anti-CD8a PE	Thermo Fisher Scientific	Cat# 12-0081-82, RRID:AB_465530
Anti-FoxP3 APC	Thermo Fisher Scientific	Cat# 17-5773-82, RRID:AB_469457
Anti-T-bet PE	Thermo Fisher Scientific	Cat# 12-5825-82, RRID:AB_925761
Anti-RORγt PE	Thermo Fisher Scientific	Cat# 12-6988-82, RRID:AB_1834470
Anti-Ki-67 eFluor 450	Thermo Fisher Scientific	Cat# 48-5698-82, RRID:AB_11149124
Anti-SLC38A1	Cell Signaling Technology	Cat# 36057, RRID:AB_2799092
Anti-ß-actin	Cell Signaling Technology	Cat# 4970 RRID:AB_2223172
IRDye 800CW Secondary	LI-COR Biosciences	Cat# 926-32211, RRID:AB_621843
IRDye 680LT Secondary	LI-COR Biosciences	Cat# 926-68020, RRID:AB_10706161
Bacterial strains
ElectroMAX DH10B Cells	Thermo Fisher Scientific	Cat#: 18290015
Chemicals, peptides, and recombinant proteins
Recombinant murine IL-12p70	Thermo Fisher Scientific	Cat# 14-8121-62
Recombinant murine IL-6	Miltenyi Biotec	Cat# 130-096-683
Recombinant murine IL-23	Miltenyi Biotec	Cat# 130-096-676
Recombinant murine IL-1ß	Miltenyi Biotec	Cat# 130-101-681
Recombinant human IL-2	NCI	Cat# Ro 23-6019
Recombinant human TGFß1	Peprotech	Cat# 100-21
MeAIB	Sigma-Aldrich	Cat# M2283
GolgiPlug	BD Biosciences	Cat# 555029
Retronectin	Takara Bio	Cat# T100A
OVA323–339 peptide (chicken, Japanese quail)	Sigma-Aldrich	Cat# O1641
MOG35–55 peptide	GenScript	Cat# RP10245
EndoFit Ovalbumin	InVivoGen	Cat# vac-pova
House dust mite	Greer Laboratories	Cat# XPB70D3A25
LPS	Sigma-Aldrich	Cat# L5293
Ovalbumin/CFA Emulsion	Hooke Laboratories	Cat# EK-0301
Critical commercial assays
EasySep Mouse CD4 T Cell Isolation Kit	STEMCELL Technologies	Cat# 19852
EasySep Mouse CD90.1 Positive Selection Kit	STEMCELL Technologies	Cat# 18958
Fixation/Permeabilization Solution Kit	BD Biosciences	Cat# 554714
Foxp3/Transcription Factor Staining Buffer Set	Thermo Fisher Scientific	Cat# 00-5523-00
KAPA Mouse Gentoyping Kits	Roche Diagnostics	Cat# 07961804001
Seahorse XFe96 FluxPaks	Agilent Technologies	Cat# 102601-100
Herculase II Fusion DNA Polymerase	Agilent	Cat# 600675
Gibson Assembly Master Mix	NEB	Cat# E2611
Polyplus jetPRIME DNA and siRNA transfection reagent	VWR	Cat# 89129-922
EAE induction kit	Hooke Labs	Cat# EK-2110
Epredia Richard-Allan Scientific Three-Step Stain Kit	Thermo Fisher Scientific	Cat# 22-050-272
Glutamine/Glutamate-Glo Assay	Promega	Cat# J8021
Deposited data
SLC transporter in vitro screen in Th0 cells, replicates 1&2	This paper	Functional ImmunoGenomics resource (FIGS; https://figs.app.vumc.org/)
Glutamate metabolism in vitro screen in Th0 cells	This paper	FIGS
Glutamate metabolism in vivo EAE screen in Th1/17 cells	This paper	FIGS
Glutamate metabolism in vivo lung inflammation screen in Th0 cells	This paper	FIGS
ΔSlc38a1 Th1/Th17 cells bulk RNAseq	This paper	GEO# GSE190131
ΔSlc38a1 Th1/Th17 cells Metabolomics	This paper	See Supplemental Table 3 for positive mode, Supplemental Table 4 for negative mode
Original western blot images and values used to create all graphs	This paper	See Data S1
Experimental models: Cell lines
Plat-E retroviral packaging cell line	Cell Biolabs	Cat# RV-101, RRID:CVCL_B488
Experimental models: Organisms/strains
Mouse: C57BL/6J	Jackson Laboratory	Strain#: 000664 RRID:IMSR_JAX:000664
Mouse: Rag1−/− (B6.129S7-Rag1tm1Mom/J)	Jackson Laboratory	Strain#: 002216 RRID:IMSR_JAX:002216
Mouse: OT-II (B6.Cg-Tg(TcraTcrb)425Cbn/J)	Jackson Laboratory	Strain#: 004194 RRID:IMSR_JAX:004194
Mouse: 2D2 (C57BL/6-Tg(Tcra2D2,Tcrb2D2)1Kuch/J)	Jackson Laboratory	Strain#: 006912 RRIDIMSR_JAX:006912
Mouse: Cas9 (B6J.129(Cg)-Gt(ROSA)26Sortm1.1(CAG-cas9*,-EGFP)Fezh/J)	Jackson Laboratory	Strain#: 026179 RRID:IMSR_JAX:026179
Oligonucleotides
NTC gRNA#1 - AAAAAGTCCGCGATTACGTC	Mouse Brie CRISPR knockout pooled library, Control sgRNAs27	Addgene #73633; RRID:Addgene_73633
NTC gRNA#2 - ATTGTTCGACCGTCTACGGG	Mouse CRISPR Knockout Pooled Library, GeCKO v2, Mouse library A gRNA sequences28	Addgene #1000000052 (NonTargetingControlGuideForMouse_0007, MGLibA_66412)
NTC gRNA#3 - ACCCATCGGGTGCGATATGG	GeCKO v2	Addgene #1000000052 (NonTargetingControlGuideForMouse_0008, MGLibA_66413)
Slc38a1 gRNA#1 - TGCATGGTGTATGAGAAGCT	Brie library target genes	Addgene #73633; RRID:Addgene_73633
Slc38a1 gRNA#2 - AGATTGGCAGGACGGACGGG	Brie library target genes	Addgene #73633; RRID:Addgene_73633
Slc38a2 gRNA#1 - CTCAAGACTGCCAACGAAGG	Brie library target genes	Addgene #73633; RRID:Addgene_73633
SLC gRNA library	This paper	See Supplemental Table 1
Glutamine metabolism gRNA library	This paper	See Supplemental Table 2
Primers for CRISPR gRNA library prep	Modified from Shalem et al., 201459	See Supplemental Table 5
Recombinant DNA
pMx-U6-gRNA scaffold-PGK-GFP plasmid	Toffalini et al., 200956	N/A
MSCV-IRES-Thy1.1 DEST plasmid	Addgene	Addgene #17442; RRID:Addgene_17442
MSCV-U6-gRNA scaffold-IRES-Thy1.1 plasmid	This paper	N/A
Software and algorithms
Prism v10	GraphPad Software	https://www.graphpad.com; RRID:SCR_002798
FlowJo v10	FlowJo	https://www.flowjo.com; RRID:SCRJD08520
MAGeCK v0.5.0.3	Li et al., 201453	http://liulab.dfci.harvard.edu/Mageck
TopSpin v3.6	Bruker	https://www.bruker.com/en/products-and-solutions/magnetic-resonance/nmr-software/topspin.html; RRID:SCR_014227
MetaboAnalyst v5.0	Pang et al., 202164	https://www.metaboanalyst.ca; RRID:SCR_015539
Agilent Wave software v2.6	Agilent	https://www.agilent.com/en/products/cell-analysis/software-download-for-wave-desktop; RRID:SCR_014526
DESeq2 v1.24.0	Love et al., 201455	http://www.bioconductor.org/packages/release/bioc/html/DESeq2.html
Cutadapt v2.10		N/A
STAR v2.7.3a		N/A
featureCounts v2.0.0		N/A

HIGHLIGHTS.

• SLC38A1 contributes to glutamine uptake to aid Th1 but not Th17 cell proliferation

• Th1 cell redox and hexosamine pathways selectively depend on SLC38A1

• SLC38A1 plays tissue selective roles in gut and brain but not lung inflammation

• T cell nutrient transporter needs vary based on subset, disease, and tissue site