SLC15A4 mediates M1-prone metabolic shifts in macrophages and guards immune cells from metabolic stress

Abstract

The amino acid and oligopeptide transporter Solute carrier family 15 member A4 (SLC15A4), which resides in lysosomes and is preferentially expressed in immune cells, plays critical roles in the pathogenesis of lupus and colitis in murine models. Toll-like receptor (TLR)7/9- and nucleotide-binding oligomerization domain-containing protein 1 (NOD1)-mediated inflammatory responses require SLC15A4 function for regulating the mechanistic target of rapamycin complex 1 (mTORC1) or transporting L-Ala-γ-D-Glu-meso-diaminopimelic acid, IL-12: interleukin-12 (Tri-DAP), respectively. Here, we further investigated the mechanism of how SLC15A4 directs inflammatory responses. Proximity-dependent biotin identification revealed glycolysis as highly enriched gene ontology terms. Fluxome analyses in macrophages indicated that SLC15A4 loss causes insufficient biotransformation of pyruvate to the tricarboxylic acid cycle, while increasing glutaminolysis to the cycle. Furthermore, SLC15A4 was required for M1-prone metabolic change and inflammatory IL-12 cytokine productions after TLR9 stimulation. SLC15A4 could be in close proximity to AMP-activated protein kinase (AMPK) and mTOR, and SLC15A4 deficiency impaired TLR-mediated AMPK activation. Interestingly, SLC15A4-intact but not SLC15A4-deficient macrophages became resistant to fluctuations in environmental nutrient levels by limiting the use of the glutamine source; thus, SLC15A4 was critical for macrophage's respiratory homeostasis. Our findings reveal a mechanism of metabolic regulation in which an amino acid transporter acts as a gatekeeper that protects immune cells' ability to acquire an M1-prone metabolic phenotype in inflammatory tissues by mitigating metabolic stress.

Keywords: amino acid transporter; cytokine; immunometabolism; innate immune cell; macrophage.

Fig. 1.

Identification of SLC15A4-associated molecules by BioID. (A) Precipitation of SLC15A4-interacting proteins. Biotinylated precipitates were visualized by silver staining or by Western blot. (B) The enrichment of SLC15A4-interacting proteins involved in the mTOR pathway. Abundance ratio (AR) score: relative enrichment value over the control. (C) Pathway analysis of SLC15A4-interacting proteins by g:Profiler (https://biit.cs.ut.ee/gprofiler/gost). (D) Expression of glycolysis-related proteins in CAL-1-shCont or CAL-1-shA4 cells was analyzed by Western blotting. (E and F) Seahorse XF analysis of OCR and ECAR. (E) OCR and ECAR in CAL-1 cells were compared between the presence or absence of CpG stimulation in Seahorse XF basic RPMI medium containing 2 mM glutamine. Reagents were added at the indicated time points (arrows). (F) OCR (Upper) and ECAR (Bottom) were compared between CAL-1-shCont and CAL-1-shA cells in the presence of 10 mM glucose (Left) or 2 mM pyruvate (Right) by Mito Stress Test using Seahorse XF basic RPMI medium. (G) Glucose uptake assay was performed for CAL-1-shCont or CAL-1-shA4 cells treated with CpG-B (ODN2006) for the indicated periods. Uptake of 2-NBDG (a fluorescent D-glucose analog) was evaluated by fluorescence intensity within cells using flow cytometry. The results are representative of three independent experiments. *P < 0.05, as determined by one-way ANOVA followed by Tukey’s post hoc test.

Fig. 2.

Metabolic analyses in BMMϕ. (A) Energy map of Seahorse XF Glyco stress test. BMMϕ were stimulated with the indicated TLR agonists for 2 h in complete RPMI containing 10% fetal calf serum. Data showed OCR and ECAR 15 min after the addition of glucose. (B) Principal component analysis of metabolites in the steady state (0 h) 1 and 18 h after CpG stimulation (n = 3). (C and D) Summary of metabolic flow analyses using ¹³C₆-glucose and ¹³C₅-glutamine. BMMϕ stimulated with CpG for 1 or 18 h were incubated for 1 h in the presence of ¹³C₆-glucose (C) and ¹³C₅-glutamine (D), and metabolic intermediates in glycolysis and the TCA cycle were quantified by MS. Values in bar graphs represent the ratio between ¹²C and ¹³C, and the reference of graphs was shown in Bottom Left. Red and closed circles in individual metabolites denote the numbers of ¹³C and ¹²C in the metabolites. References of graphs were shown in the area surrounded by dotted lines. (E) Bar graphs of statistical differences in the indicated metabolites (n = 3). *P < 0.05, **P < 0.01, ***P < 0.001 as determined by Student’s two-tailed t test. (F) Measurement of acetyl-CoA. The ratio of ¹³C₆-glucose–derived acetyl-CoA was calculated and shown as bar graphs of statistical differences (n = 3). **P < 0.01. (G) Phosphorylation of PDHA at Ser293 was compared between wild-type (WT) and Slc15a4^−/− BMMϕ. WT or Slc15a4^−/− BMMϕ were cultured and stimulated with CpG1668 for the indicated periods. (H) The ratio of Itaconic acid and transaconitate derived from ¹³C₆-glucose and ¹³C₅-glutamine (n = 3). Data are representative of two independent experiments.

Fig. 3.

Effect of glutaminolysis on the production of IL-12 family cytokines in Mϕ. BMMϕ stimulated with CpG1668 in 10% fetal calf serum–containing medium were analyzed. (A) Gene expressions of IL-12 family members after CpG1668 stimulation in BMMϕ were analyzed by RT-PCR. (B) Gene expressions of Tnf or Il6 after CpG1668 stimulation in BMMϕ were analyzed by RT-PCR. (C) TNF-α and IL-12/23p40 productions by BMMϕ. Wild-type (WT) or Slc15a4^−/− BMMϕ were cultured and stimulated with CpG1668 for 18 h, and the culture supernatant was subjected to ELISA. (D and E) Gene expressions of Il12b or Ebi3 after CpG1668 stimulation for the indicated periods in the presence of Glutaminase inhibitor 968 (appeared as C968 in D) or Dimethyl-α-oxo-glutarate (appeared as DMoG in E) in BMMϕ were analyzed by RT-qPCR. (F) Effect of glutaminolysis modulation on IL-12/23p40 production by BMMϕ. WT or Slc15a4^−/− BMMϕ were stimulated with CpG1668 for 18 h in the presence or absence of C968 glutaminase inhibitor or DMoG, and the culture supernatant was analyzed by ELISA. (G) Western blot analysis of the CpG-triggered signaling in BMMϕ. *P < 0.05, **P < 0.01, ns, not significant, as determined by Student’s two-tailed t test. Data are representative of at least two independent experiments.

Fig. 4.

SLC15A4-mediated resistance to nutrient stress. (A) Energy maps of Seahorse XF Mito stress test. OCR and ECAR of BMMϕ in different nutrient conditions were measured. The assay medium used was Seahorse XF basic RPMI containing 10% or 1% fetal calf serum (FCS). (B) Data showed values of ECAR and OCR 6 h after CpG stimulation measured by Mito Stress Test. BMMϕ were treated with CpG1668 (open circles) or control medium (closed circles) for 6 h in the presence of 10% (red squares) or 1% (black symbols) FCS. Before starting assay, media were changed to the assay medium (Seahorse XF basic RPMI medium containing 10% or 1% FCS with 10 mM glucose and 2 mM glutamine). (C) Data showed values of ECAR and OCR of BMMϕ in the steady state. BMMϕ were cultured for 6 h in the presence of 10% (red squares) or 1% (black symbols) FCS without glutamine. Assay medium used was Seahorse XF basic RPMI containing 1% or 10% FCS and 10 mM glucose without glutamine. (D) Immunoblotting of BMMϕ stimulated with CpG1668 for the indicated periods in different nutrient conditions. (E and F) Association between SLC15A4 and AMPK. (E) Myc-tagged SLC15A4 was precipitated from WEHI231-SLC15A4 transfectants using anti-myc or isotype-matched antibodies, and precipitates were analyzed by Western blot using the indicated antibodies. (F) HEK293T cells transiently expressing Flag-tagged mouse AMPK-α in combination with the indicated HA-tagged SLC15A4 proteins (Wild type [WT], EK, or mWT for human WT SLC15A4, human E465K, or mouse WT, respectively) were collected. After lysis, samples were subjected to HA or Flag immunoprecipitation and immunoblotting for the HA-tagged proteins. Data are representative of at least two independent experiments.