Loss of Neurological Disease HSAN-I-Associated Gene SPTLC2 Impairs CD8⁺ T Cell Responses to Infection by Inhibiting T Cell Metabolic Fitness

SUMMARY

Patients with the neurological disorder HSAN-I suffer frequent infections, attributed to a lack of pain sensation and failure to seek care for minor injuries. Whether protective CD8⁺ T cells are affected in HSAN-I patients remains unknown. Here, we report that HSAN-I-associated mutations in serine palmitoyltransferase subunit SPTLC2 dampened human T cell responses. Antigen stimulation and inflammation induced SPTLC2 expression, and murine T cell-specific ablation of Sptlc2 impaired antiviral T cell expansion and effector function. Sptlc2-deficiency reduced sphingolipid biosynthetic flux and led to prolonged activation of the mechanistic target of rapamycin complex 1 (mTORC1), endoplasmic reticulum (ER) stress, and CD8⁺ T cell death. Protective CD8⁺ T cell responses in HSAN-I patient PBMCs and Sptlc2-deficient mice were restored by supplementing with sphingolipids and pharmacologically inhibiting ER stress-induced cell death. Therefore, SPTLC2 underpins protective immunity by translating extracellular stimuli into intracellular anabolic signals and antagonizes ER stress to promote T cell metabolic fitness.

In Brief

SPTLC2 mutations are associated with the neurological disorder HSAN-I, in which patients get frequent infections, attributed to loss of pain sensation and failure to seek treatment for minor injuries. Wu et al. show that protective CD8⁺ T cell responses are defective in HSAN-I patients and that CD8⁺ T cells require SPTLC2-mediated sphingolipid synthesis to promote T cell metabolic fitness.

Graphical Abstract

INTRODUCTION

Hereditary sensory neuropathy type 1 (HSAN-I) is a severe neurological disease that is characterized by neuron dysfunction and severe distal sensory loss. HSAN-I patients lose their sense of pain, and as a result, they often do not seek immediate medical treatment of minor injuries, which eventually develop into severe infections and ulcerations that may even necessitate amputations. HSAN-I has been shown to be associated with missense mutations in genes encoding the two subunits of serine palmitoyltransferase (SPT), SPT long chain base subunit 1 (SPTLC1) and SPTLC2 (Bejaoui et al., 2001; Dawkins et al., 2001; Rotthier et al., 2010). SPT catalyzes the first step of the de novo synthesis of sphingolipids by condensing L-serine and palmitoyl-coenzyme A into 3-keto-sphinganine (3-KDS). 3-KDS can be further converted to the simple sphingoid bases, ceramides and complex sphingolipids. Missense mutations of SPT reduce its enzymatic activity, shift SPT substrate specificity and generate neurotoxic products (Alecu et al., 2017; Penno et al., 2010; Rotthier et al., 2010). Although SPT mutations are accompanied by severe infections in HSAN-I patients, it remains largely unknown if SPT directly regulates anti-infection CD8⁺ T cell responses.

T cell responses to antigenic stimulation are accompanied by a metabolic reprogramming (Wang et al., 2011). Glucose and amino acids are metabolized to fuel bioenergetics and biosynthesis of macromolecules, such as lipids. Sphingolipids are an important group of structural lipids found in the plasma membrane with bioactive properties (Hannun and Obeid, 2008; Sandhoff, 1993). Accumulating evidence suggests that in addition to acting as building blocks for membrane biosynthesis, sphingolipids also regulate cellular signaling in immune cells. For example, sphingosine 1-phosphate (S1P) regulates T cell trafficking and modulates differentiation of regulatory T (Treg) cells and T helper 1 (Th1) cells (Kappos et al., 2010; Liu et al., 2009; Liu et al., 2010; Matloubian et al., 2004; Rivera et al., 2008). The ceramide synthase-6 deficiency protects mice from colitis development and T cell-induced graft-versus-host disease (Scheffel et al., 2017; Sofi et al., 2017). In addition, ceramides decrease mitochondrial membrane potential and induce apoptosis (Arora et al., 1997; Di Paola et al., 2000; Ghafourifar et al., 1999; Siskind et al., 2002; Zamzami et al., 1995). Ceramides also suppress dendritic cell antigen uptake and presentation (Sallusto et al., 1996). On the other hand, ceramides enhance Treg cell suppressive function by activating protein phosphatase 2A (PP2A) and protein phosphatase 1 (PP1) and dephosphorylating mTOR (Apostolidis et al., 2016). Moreover, acid sphingomyelinase has been shown to promote the cytotoxic cytokine secretion from CD8⁺ T cells (Herz et al., 2009). It remains incompletely understood how the simple sphingoid bases, such as sphinganine, regulate T cell responses to infectious diseases.

Here we analyzed SPTLC2-mutant T cells from HSAN-I patients and mouse models with T cell-specific deficiency of Sptlc2. Sptlc2-deficiency not only affected T cell sphingolipid anabolism but also led to a prolonged activation of the mechanistic target of rapamycin complex 1 (mTORC1), which caused endoplasmic reticulum (ER) stress and CD8⁺ T cell death. Antiviral CD8⁺ T cell survival and proliferation were restored in HSAN-I peripheral blood mononuclear cells (PBMCs) by supplementing sphingolipids and pharmacologically inhibiting ER stress-induced apoptosis. This study has provided an alternative explanation of the frequent infections observed in the HSAN-I patients, in which SPTLC2 mutations directly affect T cell responses and anti-viral immunity. Our study has also revealed that SPTLC2 mediates antigenic stimulatory and inflammatory signals to instruct T cell sphingolipid anabolism, tailors mTORC1 activation to antagonize ER stress, maintains CD8⁺ T cell metabolic fitness, and underpins protective immunity.

RESULTS

The SPTLC2 mutation affects HSAN-I patient CD8⁺ T cell effector cytokine production, proliferation, and survival

To determine if HSAN-I-causing mutations in SPTLC2 affected CD8⁺ T cell responses, we analyzed the PBMCs from HSAN-I patients bearing point mutations (G435V or G382V) in SPTLC2 and from age and gender-matched healthy subjects. The CD4⁺ and CD8⁺ T cell percentages of PBMCs were comparable between HSAN-I patients and healthy donors (Figure 1A). Using two surface markers CCR7 and CD45RA as previously reported (Romero et al., 2007), we found the percentages of naïve, effector and central memory T cell subsets were also similar between the two groups (Figure S1A–B). In addition, we detected no difference of the transcription factor Foxp3 protein expression or resting CD8⁺ T cell survival between the two groups (Figure S1C–D). Effector cytokine production by HSAN-I CD8⁺ T cells was significantly reduced (Figure 1B). Furthermore, HSAN-I CD8⁺ T cells proliferated more slowly than healthy subject CD8⁺ T cells upon T cell receptor (TCR) stimulation (Figure 1C). The reduction of T cell proliferation was associated with a significant increase of apoptosis (Figure 1D). Collectively, these results show that HSAN-I-causing SPTLC2 mutations dampen human T cell effector cytokine production, proliferation, and survival.

Figure 1. The SPTLC2 mutations affect HSAN-I patient CD8⁺ T cell effector cytokine production, proliferation, and survival.

(A) FACS dot plots and bar graphs show the percentages of CD4⁺ and CD8⁺ T cells in PBMCs from HSAN-I patients and healthy subjects.

(B) PBMCs from HSAN-I patients and healthy subjects were stimulated with PMA and ionomycin for 6 hours in the presence of brefeldin A before FACS analysis of IFNγ- and TNFα-producing CD8⁺ T cells.

(C-D) PBMCs from HSAN-I patients and healthy subjects were labeled with Celltrace Violet (CTV) and stimulated with anti-CD2, anti-CD3 and anti-CD28 for 3 days before FACS analysis. FACS plots and bar graphs show the percentages of proliferating (C) and apoptotic (D) CD8⁺ T cells.

Data are expressed as mean ± SD and cumulative of three independent experiments (ten pairs of samples for each experiment described in A-D). **p<0.01; n.s., not significant. See also Figure S1.

Antigenic stimulation and inflammation induce SPTLC2 protein expression in CD8⁺ T cells

To study the potential role of SPTLC2 in anti-infection T cell responses, we used a mouse model infected with lymphocytic choriomeningitis virus (LCMV)-Armstrong and monitored SPTLC2 expression in the LCMV-specific CD8⁺ T cells. Briefly, TCR transgenic P14 CD8⁺ T cells (P14 T cells recognize the LCMV GP_33–41 epitope) were adoptively transferred into congenically mismatched C57BL/6 mice, which were subsequently infected with LCMV. Six days after infection, we detected significantly higher protein levels of SPTLC2 in donor P14 CD8⁺ T cells compared with those of naive P14 CD8⁺ T cells (Figure 2A). GP_33–41 peptide, and to a lesser extent, inflammatory cytokines, induced the SPTLC2 expression of in vitro cultured P14 CD8⁺ T cells (Figure 2B). These results suggest that both antigen exposure and inflammation promote SPTLC2 expression.

Figure 2. Sptlc2-deficiency impairs antiviral effector CD8⁺ T cell formation.

(A) Antigen-specific P14 CD8⁺ T cells were purified on day 6 (D6) after LCMV-Armstrong infection from the “P14 chimeric mice”. Naïve P14 T cells were also purified before LCMV infection (D0) to analyze the basal level of SPTLC2 protein by western blot. Sptlc2-deficient P14 CD8⁺ T cells were included as negative controls. Bar graphs show the densitometry quantification of the SPTLC2 immunoblot bands. GRP94 was used as a loading control.

(B) P14 TCR transgenic mouse splenocytes were stimulated with GP_33–41 peptide or various cytokines as indicated for 3 days for immunoblot. Each lane represents an individual mouse sample (A-B).

(C-F) Sptlc2^Flox/FloxCd4-cre (Fl/Fl) mice and wildtype littermates (+/+) were infected with LCMV-Armstrong and sacrificed 8 days later. Representative FACS plots and bar graphs (C-D) show the percentages and numbers of D^bGP_33–41 and D^bNP_396–404 tetramer-positive splenic Sptlc2-deficient or wildtype CD8⁺ T cells (C), the effector cytokine-producing CD8⁺ T cells after restimulation with or without the LCMV peptide _GP33–41 for 6 hours (D) and viral titers in mouse serum (E) and spleens (F).

Data are expressed as mean ± SD (error bars) and are representative of two (six (A) or three (B) pairs of mice in total) or three (C-F, eight pairs of mice in total) independent experiments (three in each experiment). *p<0.05; **p<0.01. See also Figure S2.

Sptlc2-deficiency impairs antiviral effector CD8⁺ T cell formation

Inspired by the observation that SPTLC2 protein was upregulated in effector CD8⁺ T cells, we went to explore if SPTLC2 was required for antiviral effector CD8⁺ T cell differentiation. We created mice with T cell-specific deficiency of Sptlc2 by breeding the Cd4-cre strain with the Sptlc2^Flox/Flox strain (Li et al., 2009). Sptlc2^Flox/FloxCd4-cre mice had similar numbers of total thymocytes and similar percentages of CD4 single positive (SP) cells, CD8⁺ SP cells, and CD4⁺CD8⁺ cells compared with wildtype littermates (Figure S2A–B). CD4⁺ and CD8⁺ T cell numbers in the inguinal lymph nodes were comparable between wildtype and Sptlc2-deficient mice. Splenic CD4⁺ T cell numbers were slightly reduced in Sptlc2-deficient mice but not statistically significant. Splenic CD8⁺ T cell numbers of Sptlc2^Flox/FloxCd4-cre mice were reduced compared with those of wildtype littermates (Figure S2A). The expression of CD69 and IL-7Rα was similar between Sptlc2-deficient and -sufficient CD8⁺ T cells, whereas CD62L was modestly reduced (Figure S2C). These results reveal that Sptlc2-deficiency does not affect T cell thymic development and moderately reduces CD8⁺ T cell numbers in the spleen but not lymph nodes under steady status.

To investigate if SPTLC2 protein regulated antiviral CD8⁺ T cell responses, we infected Sptlc2^Flox/FloxCd4-cre mice and wildtype littermates with LCMV. Eight days later, wildtype mice mounted a robust antiviral CD8⁺ T cell response, characterized by the differentiation of tetramer-positive CD8⁺ T cells (Figure 2C). Sptlc2-deficiency impaired the generation of LCMV-specific effector T cells, as manifested by the significant decrease of tetramer-positive CD8⁺ T cell numbers and effector cytokine production (Figure 2C–D). Moreover, in line with the impaired antiviral CD8⁺ T cell differentiation, the viral titers in serum and spleens were much higher in Sptlc2^Flox/FloxCd4-cre mice compared with those in the wildtype littermates (Figure 2E–F). Taken together, these results demonstrate that SPTLC2 is essentially required for the formation of robust antiviral effector CD8⁺ T cell responses.

SPTLC2 plays a CD8⁺ T cell-intrinsic role in effector T cell formation

Because the expression of Cre under the control of Cd4 deletes loxp-flanked genes in both CD4⁺ and CD8⁺ T cells (Lee et al., 2001; Sawada et al., 1994), we next addressed if SPTLC2 protein regulated antiviral T cell differentiation in a CD8⁺ T cell-intrinsic manner. We crossed Sptlc2^Flox/FloxCd4-cre mice to the TCR transgenic P14 mice. Then, we mixed the Sptlc2^Flox/FloxCd4-cre P14 CD8⁺ T cells with P14 CD8⁺ T cells from wildtype littermates before we adoptively transferred them into wildtype recipient mice (Figure 3A). We were able to distinguish between the congenically mismatched Sptlc2^+/+ and Sptlc2^Flox/Flox donor CD8⁺ T cells in the subsequent FACS analysis. Following LCMV infection, these P14 peripheral chimeric mice were sacrificed at different time points as indicated (Figure 3B). The Sptlc2-deficient and -sufficient P14 CD8⁺ T cell numbers were comparable in the first 4 days after infection, suggesting that the initial T cell activation was similar between the two groups. However, the Sptlc2-deficient CD8⁺ T cells were out-competed by their wildtype counterparts at day 6~8 (Figure 3C–D), accompanied by a significant increase of apoptosis (Figure 3E, F). In addition, SPTLC2 was required for the formation of the KLRG1^hiIL-7Rα^lo short-lived effector cells (SLECs) but not KLRG1^loIL-7Rα^hi memory precursor effector cells (MPECs) (Joshi et al., 2007) (Figure 3E, G). Furthermore, effector cytokine production was also affected by Sptlc2 deficiency (Figure 3E, H). Taken together, these data reveal that antiviral effector CD8⁺ T cell responses require SPTLC2 in a cell-intrinsic manner.

Figure 3. CD8⁺ T cells require cell-intrinsic SPTLC2 expression for robust effector T cell responses.

(A-D) Congenically mismatched Sptlc2^Flox/FloxCd4-cre and Sptlc2^+/+Cd4-cre P14 CD8⁺ T cells (10⁴ cells each) were mixed and adoptively transferred into B6 recipient mice, which were subsequently infected with LCMV-Armstrong (A). The percentages (B-C) and numbers (D, day 8) of Sptlc2-sufficient (WT) and -deficient donor CD8⁺ T cells are shown.

(E-H) FACS plots and bar graphs show the percentages or numbers of apoptotic CD8⁺ T cells (E and F), SLECs and MPECs (E and G), and cytokine-producing CD8⁺ T cells (E and H). Data (C, D, F-H) are expressed as mean ± SD (error bars) and cumulative of three independent experiments (six to eight mice in total). *p<0.05; **p<0.01; n.s., not significant.

Sptlc2-deficiency causes aberrant mTORC1 activation and ER stress in antiviral CD8⁺ T cells

To study how Sptlc2-deficiency affected effector CD8⁺ T cell formation, we went to analyze the global gene expression profiles of Sptlc2-deficient and -sufficient CD8⁺ T cells. Briefly, we performed RNA sequencing (RNA-Seq) analysis of the CD8⁺ T cells isolated from Sptlc2^Flox/FloxCd4-cre or wildtype littermates before and 8 days after LCMV infection. There were only modest differences between Sptlc2-deficient and -sufficient CD8⁺ T cell gene expression profiles before infection. Intriguingly, the transcriptional differences were remarkably increased at day 8 after LCMV infection, as manifested by the increased numbers of genes distributed outside of the dash lines (Figure 4A). Effector CD8⁺ T cell hallmark genes, such as Klrg1, Prdm1, Id2, Tbx21, and Ifng were reduced in Sptlc2-deficient CD8⁺ T cells (Figure 4B), echoing the observations that Sptlc2-deficiency impaired effector CD8⁺ T cell responses (Figure 2). In addition, genes encoding ceramidases (Asah1, Asah2, and Acer3) and sphingosine phosphate lyase 1 (Sgpl1) were increased in Sptlc2-deficient cells, implying that the sphingoid base salvage pathways were enhanced presumably to compensate for the deficiency of the SPTLC2 protein-mediated de novo synthetic pathway (Kitatani et al., 2008; Tettamanti et al., 2003).

Figure 4. Sptlc2-deficiency causes aberrant mTORC1 activation and ER stress in antiviral CD8⁺ T cells.

(A) Volcano plots show the distribution of significantly up- and down-regulated genes in the Sptlc2-deficient CD8⁺ T cells before and 8 days after LCMV-Armstrong infection.

(B) A heat map shows the z-score of the mRNA levels of the indicated genes in CD8⁺ T cells purified from three pairs of Sptlc2^Flox/FloxCd4-cre (Fl/Fl) mice and wildtype littermates (+/+) at day 8 after LCMV-Armstrong infection. “Effector” block contains hallmark genes of effector CD8⁺ T cells. Genes regulating the sphingoid base salvage pathways and ER-stress are shown in the “SB” and “ER stress” blocks. Sptlc2 is included as a control.

(C) FACS plots show the percentages of phosphorylated S6 (pS6) in Sptlc2-sufficient (+/+) and -deficient (Fl/Fl) CD8⁺ T cells after in vitro stimulation with anti-CD3 and anti-CD28. pS6 MFI is expressed as mean ± SD (error bars) and cumulative of three independent experiments (one pair of mice in each experiment).

(D) The ER stress marker expression in CD8⁺ T cells of Sptlc2^Flox/FloxCd4-cre (Fl/Fl) and Sptlc2^+/+Cd4-cre (+/+) littermate mice infected with LCMV-Armstrong 8 days earlier was analyzed by western blot. Rapamycin was administered every other day (100 μg/kg body weight) as indicated. T cells of three mice in each group were collected separately from two independent infection experiments and analyzed together. Each lane represents an individual sample. Bar graphs show the densitometry quantification of the immunoblot bands. Data are expressed as mean ± SD (error bars). *p<0.05; **p<0.01. See also Figure S3.

In addition, a group of genes regulating ER stress, such as C/EBP homologous protein (CHOP, also known as Ddit3), X-box-binding protein (XBP)-1, and binding of immunoglobulin protein (Bip, also known as Hspa5) were significantly increased in Sptlc2-deficient CD8⁺ T cells (Figure 4B). Because Sptlc2-deficiency affected CD8⁺ T cell subset differentiation at day 8, we monitored the mRNA levels of these genes by qPCR at an earlier time point and got similar results (Figure S3A–E). The subsequent Ingenuity Pathway Analysis (IPA) revealed that one top pathway upregulated in Sptlc2-deficient CD8⁺ T cells was ER stress-induced apoptosis (Figure S3F). Western blot analysis confirmed that Sptlc2-deficient CD8⁺ T cells expressed higher protein levels of ER stress markers, such as CHOP, XBP-1, Bip, phosphorylated eIF2α (peIF2α) and phosphorylated protein kinase RNA-like endoplasmic reticulum kinase (pPERK) (Figure 4D). Furthermore, Sptlc2-deficient CD8⁺ T cells displayed significantly dilated ER compared with the wildtype CD8⁺ T cells (Figure S3G). Collectively, these data reveal that Sptlc2-deficiency causes ER stress in CD8⁺ T cells.

We next explored the underlying mechanisms through which Sptlc2-deficiency caused ER stress. It has been reported that prolonged activation of mTORC1 and phosphorylation of S6 ribosomal protein (S6) causes ER stress (Di Nardo et al., 2009; Ozcan et al., 2008). Upon anti-CD3 and anti-CD28 stimulation, wildtype CD8⁺ T cells rapidly increased S6 and 4E-BP1 phosphorylation (surrogate markers of mTORC1 activity), which dropped at a later time point (Figure 4C, S4A). In contrast to the short-term mTORC1 activation in wildtype CD8⁺ T cells, Sptlc2-deficiency sustained S6 and 4E-BP1 phosphorylation for three days (Figure 4C, S4A). One might expect the enhancement of mTORC1 activity was dependent on Akt. However, we did not detect a change of Akt signaling. One explanation is that the mTORC1 activity is not dependent on PI3K/Akt activation in CD8⁺ T cells (Finlay et al., 2012). Furthermore, Sptlc2-deficiency did not affect NFκB (p65), MAPK (p38, JNK, Erk) or NFAT1 signaling (Figure S2D–E). To dissect if the prolonged mTORC1 activation was required for ER stress in Sptlc2-deficient T cells, we treated the LCMV-infected Sptlc2^Flox/FloxCd4-cre mice with mTORC1 inhibitor rapamycin. Remarkably, rapamycin reduced the expression of the ER stress markers (Figure 4D), indicating a causal relationship between the prolonged activation of mTORC1 and ER stress in Sptlc2-deficient CD8⁺ T cells.

Sptlc2-deficiency reduces the sphingolipid biosynthetic flux and prolongs mTORC1 activation through protein phosphatases in CD8⁺ T cells

Next, we examined how Sptlc2-deficiency prolonged mTORC1 activation in CD8⁺ T cells. Sphingolipids have been shown to regulate mTORC1 activity (Apostolidis et al., 2016; Chung et al., 1997; Grey et al., 2002; Kluk and Hla, 2001). Thus, we hypothesize that the aberrant mTORC1 activation in Sptlc2-deficient T cells was due to dysregulated sphingolipid metabolism. To test this hypothesis, we analyzed the sphingolipid profiles of Sptlc2-sufficient and -deficient CD8⁺ T cells. Sphinganine, dihydroceramide, ceramide, and sphingomyelin were all significantly reduced in Sptlc2-deficient T cells (Figure 5A and Table S1–2). On the other hand, 3-KDS was below the limit of detection. These data show that 3-KDS is probably a transient metabolic intermediate in mouse CD8⁺ T cells. Sphingosine generation was unaffected in Sptlc2-deficient T cells (Figure 5A).

Figure 5. Sptlc2-deficiency reduces sphingolipid biosynthetic flux and prolongs mTORC1 activation through protein phosphatases in CD8⁺ T cells.

(A) Sptlc2^Flox/FloxCd4-cre (Fl/Fl) or Sptlc2^+/+Cd4-cre (+/+) P14 CD8⁺ T cells were cultured with the cognate peptide GP_33–41 for 3 days. The primed CD8⁺ T cells were then expanded using IL-2 for another 2 days before lipid extraction and mass spectrometry analysis of sphingolipids. The heat map shows the z-score of the indicated metabolites.

(B) The sphingolipid de novo synthetic pathway is depicted with tracers highlighted in red.

(C) P14 CD8⁺ T cells as described in (A) were expanded using IL-2. Then T cells were washed and cultured in HBSS plus 1% FBS in the presence of isotope-labeled L-serine for 0~120 minutes for mass spectrometry analysis of the tracer incorporation into various sphingolipids.

(D) Sptlc2^Flox/FloxCd4-cre (Fl/Fl) or Sptlc2^+/+Cd4-cre (+/+) CD8⁺ T cells were stimulated with anti-CD3 and anti-CD28 for 3 days in the presence or absence of sphingolipids (5 μM 3-KDS, 5 μM sphinganine, 50 nM C16 ceramide, 50 nM C18 dihydroceramide, 4 μg/ml palmitoyl sphingomyelin, and 1 μM sphingosine) or microcystin-LR (it inhibits both PP1 and PP2A at 25 nM) as indicated before FACS analysis of pS6.

Data (A and C, two pairs of mice in each experiment) are cumulative of two independent experiments or representative (D, one pair of mice in each experiment) of three experiments. Data are expressed as mean ± SD (error bars) (C). *p<0.05; **p<0.01. See also Figure S4, S5 and S6.

To examine how SPTLC2 protein regulated sphingolipid biosynthetic flux in Sptlc2-sufficient and -deficient CD8⁺ T cells, we performed the isotope labeling experiment (Figure 5B–C and Table S1–2). Briefly, we pulsed the wildtype and Sptlc2-deficient CD8⁺ T cells with ¹³C₃,¹⁵N-labeled L-serine and then monitored the ¹³C and ¹⁵ N-labeled metabolic intermediates over time. Similar to the quantification of the endogenous 3-KDS, ¹³C and ¹⁵ N-labeled 3-KDS could not be detected, supporting that 3-KDS probably only existed transiently in T cells. The incorporation of isotope tracers into sphinganine and other sphingolipids became detectable as early as 30 minutes after ¹³C₃,¹⁵N-labeled L-serine addition, and throughout the entire assay (Figure 5C). Furthermore, Sptlc2-deficiency reduced the biosynthesis of sphinganine, dihydroceramide, ceramide, and sphingomyelin. Sptlc2-deficiency did not affect the incorporation of isotope tracers into sphingosine during the entire period of observation, which helped to explain the comparable levels of endogenous sphingosine between Sptlc2-sufficient and -deficient CD8⁺ T cells (Figure 5A, C).

To test if HSAN-I-associated SPTLC2 mutation influenced the canonical SPT activity in human CD8⁺ T cells, we pulsed the resting and in vitro activated CD8⁺ T cells of HSAN-I patients and healthy subjects with ¹³C₃,¹⁵N-labeled L-serine for 60 minutes. Then we monitored the ¹³C and ¹⁵N incorporation into the downstream sphingolipids. Most of the sphingolipids examined in this assay, except ¹³C,¹⁵N-dihydroceramide and ¹³C,¹⁵N-ceramide, were only modestly affected by the SPTLC2 mutation in resting CD8⁺ T cells (Figure S1E). Compared with the resting healthy subject CD8⁺ T cells, the activated healthy subject CD8⁺ T cells incorporated much more ¹³C and ¹⁵ N into sphingolipids, suggesting that T cell activation significantly enhanced the sphingolipid biosynthesis (Figure S1E). In contrast, HSAN-I patient T cells were resistant to the anti-CD2, anti-CD3 and anti-CD28-induced sphingolipid increase. As a result, the originally minimal difference of the SPT activity between the HSAN-I patient and healthy subject resting CD8⁺ T cells was amplified by the T cell activation. Compared with the mouse CD8⁺ T cell metabolic flux analysis, ¹³C,¹⁵N-3KDS was detectable and ¹³C,¹⁵N-sphingosine was below the detection limit in the human CD8⁺ T cells, suggesting a species-specific difference in the sphingolipid synthesis. Despite this minor difference, the data suggest that both Sptlc2 genetic deficiency in mice and SPTLC2 point mutation in human reduced the SPT activity in CD8⁺ T cells.

To examine if the prolonged mTORC1 activation in Sptlc2-deficient T cells was due to the reduction of sphingolipid biosynthesis, we supplemented various sphingolipids into Sptlc2-deficient T cell culture. Supplementation of 3-KDS, sphinganine, dihydroceramide, ceramide, and sphingomyelin but not sphingosine reduced S6 phosphorylation and ER stress marker expression in Sptlc2-deficient CD8⁺ T cells (Figure 5D, S4A and S5). Intriguingly, this pS6 suppression was blocked by microcystin-LR (an inhibitor of PP1 and PP2A) (Figure 5D), and the PP2A activity was significantly reduced in the Sptlc2-deficient cells (Figure S6A), suggesting that SPTLC2 prevented mTORC1 hyperactivation in CD8⁺ T cells dependent on PP1 and PP2A. Ceramides have been shown to increase the PP2A activity through inhibiting the “Su(var)3–9, Enhancer-of-zeste and Trithorax” (SET)-domain containing protein inhibitor 2 of PP2A (I2PP2A, encoded by Set) (Dobrowsky et al., 1993; Li et al., 1996; Mukhopadhyay et al., 2009). To determine if Sptlc2-deficiency suppressed PP2A activity dependent on I2PP2A in CD8⁺ T cells, we generated Set-deficient CD8⁺ T cells through the electroporation-mediated delivery of CRISPR gRNA/Cas9 complex following a recently reported protocol (Seki and Rutz, 2018). The Set/Sptlc2 dual deficient CD8⁺ T cells had higher PP2A activity, reduced S6 phosphorylation and enhanced cell survival compared with the Sptlc2-deficient CD8⁺ T cells (Figure S6B–E). Collectively, these results support the hypothesis that SPTLC2-mediated sphingolipid synthesis is required to maintain appropriate PP2A and mTORC1 activity through inhibiting I2PP2A (Figure S6F).

Sphingolipid supplementation and inhibition of ER stress-induced cell death partially restore Sptlc2-deficient antiviral T cell survival and expansion

Next, we tested if supplementation of sphingolipids and pharmacological inhibition of ER stress-induced cell death rescued the Sptlc2-deficient T cell survival and proliferation. We established an in vitro assay in which Sptlc2-deficient CD8⁺ T cells underwent apoptosis and proliferated more slowly than wildtype control CD8⁺ T cells upon stimulation with anti-CD3 and anti-CD28 (Figure 6A–B). These results also echoed the observation that Sptlc2-deficiency impaired CD8⁺ T cell responses in the LCMV infection model (Figure 2C–F). Exogenous sphinganine restored cell survival and proliferation (Figure 6A–B). In addition, 3-KDS, sphingomyelin but not sphingosine partially rescued Sptlc2-deficient CD8⁺ T cell survival and proliferation (Figure 6A–B and data not shown). This result correlated with the observation that sphingosine was not reduced in Sptlc2-deficient CD8⁺ T cells (Figure 5A and 5C). In addition, salubrinal (an inhibitor of ER stress-induced cell death) (Boyce et al., 2005) fundamentally restored Sptlc2-deficient CD8⁺ T cell survival and proliferation, suggesting that ER stress caused Sptlc2-deficient CD8⁺ T cell death and proliferation defect. Furthermore, the pS6 protein level was still very high compared with that in the wildtype T cells (Figure S4B). These results suggest that the increased pS6 can be uncoupled from cell death in salubrinal-treated Sptlc2-deficient CD8⁺ T cells. In addition, rapamycin enhanced Sptlc2-deficient CD8⁺ T cell survival and but only moderately increased cell proliferation (Figure 6A–B), suggesting that although mTORC1 hyperactivation causes cell death, mTORC1 activity is still required for the optimal proliferation of Sptlc2-deficient CD8⁺ T cells.

Figure 6. Sphingolipid supplementation and inhibition of ER stress-induced cell death partially restore Sptlc2-deficient antiviral T cell survival and expansion.

(A-B) Sptlc2^Flox/FloxCd4-cre (Fl/Fl) or Sptlc2^+/+Cd4-cre (+/+) CD8⁺ T cells were labeled with (B) or without (A) CTV and stimulated with anti-CD3 and anti-CD28 for 3 days in the presence or absence of 3-keto-sphinganine (3-KDS, 5 μM), sphinganine (Spha, 5 μM), sphingomyelin (Sphm, 4 μg/ml), rapamycin (Rapa, 10 nM), and salubrinal (Salu, 5 μM) before FACS analysis of apoptosis (A) and proliferation (B).

(C-E) Sptlc2^Flox/FloxCd4-cre (Fl/Fl) and wildtype P14 CD8⁺ T cells from littermates were mixed (1×10⁴ cells from either group) and adoptively transferred into C57BL/6 mice, which were infected with LCMV and sacrificed 8 days later. Sphinganine (100 μg/kg), rapamycin (100 μg/kg), and salubrinal (1 mg/kg) were peritoneally injected every other day (day 0~8 after LCMV infection). Bar graphs show the ratio (D) and total numbers (E) of Sptlc2-deficient and -sufficient donor CD8⁺ splenic T cells at day 8 after LCMV infection.

(F) FACS plots and bar graphs show the percentages or numbers of apoptotic CD8⁺ T cells, SLECs and MPECs, and cytokine-producing CD8⁺ T cells. Data are representative (A-B, three pairs of mice in total) and cumulative (D-F, six pairs of mice in total) of three independent experiments. Data are expressed as mean ± SD (error bars) (D-F). *p<0.05; **p<0.01; n.s., not significant. See also Figure S4 and S5.

To test if sphingolipid administration and ER stress-induced cell death inhibition rescued Sptlc2-deficient antiviral CD8⁺ T cell responses in vivo, we adoptively transferred Sptlc2^Flox/FloxCd4-cre and wildtype P14 CD8⁺ T cells into wildtype recipient mice. These “P14 peripheral chimeras” were infected with LCMV-Armstrong and treated with sphinganine, rapamycin, salubrinal, or vehicle control. Intriguingly, sphinganine, and to a lesser extent, rapamycin and salubrinal treatment increased the numbers of Sptlc2-deficient CD8⁺ T cell donors (Figure 6C–E). The administration of the three compounds also partially restored Sptlc2-deficient CD8⁺ T cell survival, SLEC differentiation and effector cytokine production (Figure 6F). Collectively, these results indicate that Sptlc2-deficient effector CD8⁺ T cell formation can be at least partially rescued by supplementing exogenous sphingolipids and inhibiting ER stress-induced cell death.

Sphingolipid supplementation and inhibition of ER stress-induced cell death enhances the SPTLC2-mutated HSAN-I patient CD8⁺ T cell cytokine production, proliferation, and survival

To study if the findings made using the mouse models were relevant to the immunodeficiency observed in the HSAN-I patients, we supplemented sphinganine, rapamycin and salubrinal to CD8⁺ T cells of the HSAN-I patients or healthy subjects. Sphinganine and salubrinal restored HSAN-I CD8⁺ T cell effector cytokine production, cell proliferation, and survival (Figure 7). Rapamycin also increased HSAN-I CD8⁺ T cell effector cytokine percentages and promoted cell survival. Compared with sphinganine and salubrinal, rapamycin only modestly but still significantly enhanced HSAN-I CD8⁺ T cell proliferation (Figure 7). Taken together, these results suggest that sphingolipid supplementation and ER stress inhibition at least partially correct the HSAN-I-associated immunodeficiency.

Figure 7. Supplementing with sphinganine and inhibiting ER stress-induced cell death corrects the immunodeficiency in SPTLC2-mutated HSAN-I patient PBMCs.

(A) PBMCs from HSAN-I patients and healthy subjects were stimulated with PMA and ionomycin for 6 hours in the presence or absence of the indicated compounds (upper panel). FACS plots show the production of IFNγ and TNFα by CD8⁺ T cells. Alternatively, PBMCs were labeled with CTV (middle panel only) and stimulated with anti-CD2, anti-CD3 and anti-CD28 for 3 days (both middle and lower panels). FACS plots show the percentages of proliferating (middle panel) and apoptotic (lower panel) CD8⁺ T cells.

(B) Bar graphs show the cumulative results obtained from ten pairs of samples. Data are expressed as mean ± SD. **p<0.01. See also Figure S7.

Deoxysphingolipids suppress CD8⁺ T cell proliferation and cytokine production at high concentrations

In addition to reducing the SPT enzymatic activity, the HSAN-I-associated SPTLC2 mutations also shift the substrate preference from L-serine to L-alanine and L-glycine, resulting in the production of neurotoxic 1-deoxysphinganine (m18:0), 1-deoxymethylsphinganine (m17:0) and downstream deoxysphingolipids (Alecu et al., 2017; Penno et al., 2010; Rotthier et al., 2010). To address if the neurotoxic deoxysphingolipids suppressed human CD8⁺ T cell proliferation and cytokine production, we cultured healthy human CD8⁺ T cells with two neurotoxic deoxysphingolipids, 1-deoxysphinganine (m18:0) and 1-deoxymethylsphinganine (m17:0) (Penno et al., 2010). We observed that the neurotoxic deoxysphingolipids suppressed human CD8⁺ T cell proliferation and effector cytokine production when added at high concentrations (>2.4 μM) (Figure S7A–D). The previous studies suggest that the concentrations of 1-deoxysphinganine (m18:0) and 1-deoxymethylsphinganine (m17:0) in the plasma of HSAN-I patients are lower than 0.8 μM (Murphy et al., 2013; Penno et al., 2010). Thus, these data suggest that the deoxysphingolipids, at the concentrations observed in the HSAN-I patient plasma, do not exert cytotoxicity on CD8⁺ T cells.

DISCUSSION

Investigations of rare human diseases, such as severe combined immunodeficiency (SCID), have revealed the importance of several key immunological pathways (Kovanen and Leonard, 2004; Leonard, 1996). Similarly, our finding of SPTLC2 as an immunometabolic checkpoint is a lesson learned from the rare neurological disorder HSAN-I. The current explanation for the frequent infections associated with HSAN-I is that the patients have a complete or partial loss of pain, and therefore do not seek immediate medical treatment of originally minor injuries. Our study has proposed a different mechanism that HSANI-associated genetic mutations in SPTLC2 directly affect T cell responses. Besides SPTLC2, HSAN-I neuropathy has been revealed to be genetically linked to mutations other genes, such as SPTLC1 (Bejaoui et al., 2001; Dawkins et al., 2001). Our unpublished data suggested that SPTLC1, similar to SPTLC2, was upregulated in CD8⁺ T cells upon LCMV infection, implying that SPTLC1 might also regulate antiviral T cell responses in the form of a heterodimeric complex with SPTLC2. Thus, our study has provided a molecular mechanism linking HSAN-I neuropathy to the impaired T cell metabolism, survival, and function. This link will help us to better understand the causes of the HSAN-I clinical symptoms, and provide a different view of HSAN-I and also other ulceromutilating neuropathies from an immunological perspective.

The HSAN-I-associated SPTLC2 mutations not only reduce SPT enzymatic activity but also shift the substrate preference from L-serine to L-alanine and L-glycine, resulting in the production of neurotoxic deoxysphingolipids (Alecu et al., 2017; Penno et al., 2010; Rotthier et al., 2010). It has been shown in a pilot study that L-serine ameliorates clinical symptoms in HSAN-I patients (Garofalo et al., 2011). A clinical trial has been carried out by the same group (NCT01733407). The rationale is to increase the availability of L-serine and also to reduce the relative abundance of alanine and glycine for the synthesis of the neurotoxic products. Our study suggests sphinganine supplementation as another potential option to rescue the anti-infection T cell-intrinsic defects associated with HSAN-I. One future direction is to explore if sphinganine alone, or in combination with L-serine, ameliorates the HSAN-I clinical symptoms.

Sphingolipid metabolism has been implicated in the regulation of cell death and survival. For example, ceramides can induce cell cycle arrest and apoptosis by promoting the formation of large lipid raft signaling platforms, CD95 clustering, phosphatase activation and increasing the mitochondrial membrane permeability (Cremesti et al., 2001; Grassmé et al., 2001; Nickels and Broach, 1996; Siskind et al., 2002, 2006). On the other hand, sphingosine-1-phosphate induces Erk activation and antagonizes ceramides-induced SAPK/JNK pathway to promote cell survival (Cuvillier et al., 1996). Our current study has revealed another layer of complexity that sphinganine and other sphingolipids promoted effector CD8⁺ T cell proliferation not only by fueling lipid anabolism but also through moderating mTORC1 activity and suppressing ER stress-induced cell death. mTORC1 activation plays an essential role in regulating cell growth and T cell metabolic reprogramming (Albert and Hall, 2015; Laplante and Sabatini, 2012; Wang et al., 2011; Zeng et al., 2013). Our study suggests that in spite of its importance in underpinning cell survival and growth, mTORC1 is tightly regulated to reduce ER stress and effector CD8⁺ T cell death. Induced by TCR stimulation and inflammation, SPTLC2 instructs sphingolipid biosynthetic flux, fine-tunes infection-induced mTORC1 activity, and antagonizes ER stress-induced antiviral CD8⁺ T cell death. Thus, SPTLC2 promotes T cell metabolic fitness and helps to maintain the sustainability of mTORC1-dependent metabolic reprogramming of antiviral effector CD8⁺ T cells.

We found that Sptlc2-deficiency reduces the biosynthesis of sphinganine, dihydroceramide, ceramide, and sphingomyelin, but does not completely inhibit sphingolipid synthesis and barely affects sphingosine synthesis. This is in accordance with previous findings of residual sphingolipids in targeted cells and organs of other cell-specific Sptlc2-deficient mice (Chakraborty et al., 2013; Lee et al., 2012; Lee et al., 2017; Li et al., 2018; Li et al., 2009). One possibility is that the residual SPT activity of SPTLC1 suffices for low levels of sphingolipid synthesis. Another possibility is that sphingolipid metabolism is highly plastic. The influence of a single enzymatic defect in sphingolipid biosynthesis could be compensated by the changes of other metabolic steps; especially, if cells could take up sphingolipids from their environment, e.g. from exosomes. In addition, sphingolipid salvage pathways can also compensate for the deficiency of the de novo synthetic pathway. For example, sphingoid bases can be generated through degradation of the phosphorylated sphingoid bases by sphingosine phosphate lyase 1, which is encoded by Sgpl1. Remarkably, Sptlc2-deficient CD8⁺ T cells expressed higher levels of Sgpl1 mRNA. On the other hand, sphingosine can be generated through ceramide deacylation by ceramidases. Ceramidase-encoding genes such as Asah1, Asah2, and Acer3 were increased in Sptlc2-deficient T cells. This altered gene expression pattern is expected to desensitize sphingosine biosynthesis from the shrinking of ceramide intracellular pool in Sptlc2-deficient T cells. This metabolic plasticity is particularly important for T cells, because the sphingosine-derived metabolite sphingosine-1-phosphate plays an important role in T cell migration, differentiation, and proliferation (Cyster and Schwab, 2012; Jin et al., 2003; Liu et al., 2009; Liu et al., 2010; Mendoza et al., 2017).

Sptlc2-deficiency reduced antiviral CD8⁺ T cell numbers more potently in the Sptlc2^Flox/FloxCd4-cre mice than in the “P14 peripheral chimeras”. One explanation is that both CD4⁺ and CD8⁺ T cells lack functional SPTLC2 in Sptlc2^Flox/FloxCd4-cre mice, whereas in the “P14 peripheral chimeras”, SPTLC2 remains intact in the host CD4⁺ T cells. Our unpublished data suggest that SPTLC2 also regulates CD4⁺ T cell subset differentiation. Because CD4⁺ T cell help plays an important role in CD8⁺ T cell primary response (Bennett et al., 1998; Bennett et al., 1997; Oh et al., 2008; Ridge et al., 1998; Schoenberger et al., 1998; Sokke Umeshappa et al., 2012; Wilson and Livingstone, 2008), the dual deficiency of Sptlc2 in both CD4⁺ and CD8⁺ T cells is expected to influence antiviral CD8⁺ T cells through both CD8⁺ T cell-intrinsic and -extrinsic mechanisms in Sptlc2^Flox/FloxCd4-cre mice. To study the CD8⁺ T cell-extrinsic mechanisms, one future direction is to address the potential role of SPTLC2 in CD4⁺ T cell subset differentiation and function.

Overall, our study has revealed that antiviral CD8⁺ T cell responses required SPTLC2-mediated sphingolipid biosynthesis to support T cell anabolism and to reduce ER stress. These findings have also provided an immunological perspective to revisit the causes of human ulceromutilating neuropathies.

STAR ★ METHODS

CONTACT FOR REAGENTS AND RESOURCE SHARING

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Guoliang Cui (g.cui@dkfz.de).

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Mice

Sptlc2^Flox/Flox mice were kindly provided by Professor Xian-cheng Jiang (SUNY Downstate Medical Center, New York) via Professor Vishwa Dixit and Professor Susan Kaech (Yale University, New Haven) (Li et al., 2009). Mice were maintained in the German cancer research center (DKFZ) specific pathogen-free facility. We used sex-matched 6~10-week-old mice (both female and male) for the experiments and subsequent comparisons. All the studies were performed in accordance with DKFZ regulations after approval by the German regional council at the Regierungspräsidium Karlsruhe.

Human samples

The HSAN-I study was approved by the ethics committee of DKFZ and Heidelberg University. Informed consent from all patients was obtained before the analysis. HSAN-I human samples (4 female and 6 male at the age of 52.0±12.4) were collected by Drs. Nathalie Bonello-Palot, Beate Schlotter-Weigel, Michaela Auer-Grumbach, Pavel Seeman, Wolfgang Löscher, Markus Reindl, Eric Mah, David Boyle, Andre A Matti and Carla Grosmann. The patients did not have recent medical events related to their diagnosis.

METHOD DETAILS

LCMV infection

In infection-related experiments, mice were intraperitoneally infected with 2×10⁵ pfu LCMV-Armstrong. Where indicated, we generated the “P14 chimeric mice” by adoptively transferring 10⁴ LCMV-specific Sptlc2-sufficient and -deficient P14 TCR transgenic CD8⁺ cells to C57BL/6 mice 1 day before LCMV-Armstrong infection. Mice were intraperitoneally treated with sphinganine (100 μg/kg body weight, every other day), rapamycin (100 μg/kg body weight, every other day), and salubrinal (1 mg/kg body weight, every other day).

Primary cell cultures

Primary T cells derived from gender-matched human subjects or mice (both male and female) were cultured in complete RPMI 1640 medium (plain RPMI 1640 medium with 10% fetal bovine serum, Penicillin and Streptomycin, 2-mercaptoethanol, L-Glutamine and non-essential amino acids). Antibodies, cytokines, peptides, and chemicals were supplemented as indicated. For the in vitro treatment with ceramides, C16 ceramide and C18 dihydroceramide were first dissolved in ethanol. Then 2% dodecane was added to increase the solubility as reported (Chalfant et al., 1999; Novgorodov et al., 2008). The final concentrations of dodecane and ethanol did not exceed 0.001% and 0.05% in the cell culture. 3-ketosphinganine (final concentration is 5 μM), sphinganine (5 μM), sphingosine (1 μM), and salubrinal (5 μM) were dissolved in DMSO. Sphingomyelin was dissolved in ethanol (10mg/ml) and the final working concentration was 4 μg/ml. Microcystin-LR dissolved in DMSO was added to cell culture medium at 25 nM. The stock solutions were warmed to 37°C prior to supplementation into the medium. Vehicle controls (DMSO or ethanol) were added with the final concentrations lower than 0.05%.

Lipid extraction from T cells

Sptlc2-sufficient and -deficient P14 CD8⁺ T cells were activated with GP_33–41 peptide for three days and then stimulated with IL-2 for another two days. CD8⁺ T cells were then pulsed with 500 μM ¹³C₃, ¹⁵N-LSerine (Sigma) in HBSS plus 1% FBS at 37°C for 0–120 minutes. Then 2×10⁷ cells were pelleted and used for lipid extraction. 100 μL of a methanolic solution containing a mix of internal lipid standards [C12–3-ketosphinganine, C20-sphinganine, C20-sphingosine, Cer(d18:1/14:0), Cer(d18:1/19:0), Cer(d18:1/25:0), Cer(d18:1/31:0), 1-desoxy-Cer(d18:0/12:0), GlcCer(d18:1/14:0), GlcCer(d18:1/19:0), GlcCer(d18:1/25:0), GalCer(d18:1/31:0), SM(d18:1/12:0), SM(d18:1/17:0), SM(d18:1/py), and SM(d18:1/31:0)] was added and further suspended in 400 μL chloroform:methanol (1:1). The suspension was incubated for 15 min at 37 °C in an ultrasound bath, which was turned on for 3 min in the beginning, after 5 and after 10 min. The suspension was centrifuged at room temperature at 13000g for 3 min and 450 μL of the clear supernatant was collected with a Hamilton syringe into a clean glass vial. The residual pellet was mixed with 450 μL of chloroform:methanol:water (10:10:1) treated like before and the second supernatant was added to the first. This extraction was repeated with the residual pellet and the third supernatant added to the other two. The pooled extract was dried with a gentle nitrogen gas stream at 37 °C and suspended in 2mL 100 mM aqueous potassium acetate using ultrasound for subsequent desalting via C18-reversed phase cartridges. The lipid eluate from the cartridges was dried and finally taken up in 200 μL 95% methanol and subjected to UPLC-ESI-MS² analysis. To detect sphingomyelins in the absence of phosphatidylserines an aliquot of the samples was dried and treated with 25% aqueous ammonia:methanol (1:1) at 80 °C for 5 h. After drying samples with a gentle nitrogen stream at 40 °C, samples were taken up in 200 μl 95% methanol for LC-MS² analysis.

LC-MS² analysis of sphingolipids

Extracted lipids were separated on a Waters I class UPLC equipped with a Waters CSH C18 column (length 100 mm, diameter 2.1mm, particle size 1.7 μm) using a gradient starting with 57% solvent A (50% methanol, 50% water) and 43 % solvent B (99% 2-propanol, 1% methanol), both solvents containing 10 mM ammonium formiate, 0.1% formic acid and 5 μM sodium citrate. For details see Table S1. The UPLC was coupled to an ESI-(QqQ)-tandem mass spectrometer (Waters Xevo TQ-S) for compound detection in +ESI MRM mode. De novo synthesized SLs were discriminated from steady state SLs by incorporation of ¹³C₃,¹⁵N₁-stable isotope labeled L-serine leading to an n+3 mass shift of the corresponding molecular ions and a corresponding n+2/n+3 mass shift of the product ions in MRM mode. For details see Table S2. Samples were injected and processed using MassLynx software, whereas mass spectrometric peaks were quantified according to their peak area ratio with respect to the internal standard using TargetLynx software (both v 4.1 SCN 843) both from Waters Corporation (Manchester, UK). Subsequently, quantification of ceramides, hexosylceramides, and sphingomyelins was adjusted to the length of the acyl-chain and dihydro(hexosyl)ceramide quantification was further adjusted by a factor calculated between the intensities external ceramides and dihydroceramidstandards of the same concentration.

Flow Cytometry

For Annexin V/propidium iodide (PI) staining, cells were washed with Annexin V buffer before staining with anti-Annexin V and PI. Cells were directly analyzed using LSR II without fixation. To stain surface antigens, cells were washed with FACS buffer before adding the fluorescently-conjugated antibodies. After incubating with antibodies for 20~30 minutes on ice, cells were washed with FACS buffer and fixed with 4% PFA. To stain cytokines, cells were permeabilized using the eBioscience permeabilization buffer. To stain phosphorylated antigens, cells were fixed with 4% PFA first. Then cells were treated with cold 70% methanol on ice for 30 minutes. After washing three times using FACS buffer, cells were incubated with staining antibodies on ice for 30 minutes before LSR II analysis. To do Celltrace Violet dilution assay, cells were labeled with Celltrace Violet in complete medium at 37°C for 20 minutes. Then cells were washed with RPMI 1640 medium with 1% FBS for 3 times. After 3 days, Celltracer Violet fluorescence was analyzed in the PacBlue channel using an LSR II.

Immunoblot

Cells were lysed using RIPA lysis buffer (150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS and 50 mM Tris pH 8.0) and boiled in 5x SDS sample buffer (10% SDS, 10 mM DTT, 20% glycerol, 200 mM Tris-HCl, pH 6.8 and 0.05% bromophenol blue) for 5~10 minutes. Then cell lysates were loaded and resolved using 12% SDS-PAGE (120 V, until the blue indicator runs to the edge of the gel). Proteins were then transferred onto PVDF membranes (100 V, 1 hour, on ice). The membranes were then blocked with 5% milk in PBS supplemented with Tween-20 (PBST) for 1 hour at room temperature, followed by incubation overnight at 4°C with primary antibodies. The PVDF membrane was washed 3 times (5~10 minutes each time) with PBST and then incubated with HRP-conjugated secondary antibodies at room temperature for 1 hour. After extensive washing with PBST, the membrane was developed using the ECL method. We quantified the band intensities using the NIH ImageJ program.

RNA sequencing

CD8⁺ T cells were MACS-purified from naïve or LCMV-Armstrong-infected (2×10⁵ pfu, day 8 after infection, i.p.) Sptlc2^Flox/FloxCd4-cre mice and wildtype littermates. The purity of CD8⁺ T cells was checked using FACS and it was higher than 96%. RNA was prepared using the QIAGEN RNeasy Mini Kit. DNA was removed using the QIAGEN RNase-Free DNase Set. Library was prepared by German cancer research center High Throughput Sequencing Unit. Then, libraries were pooled with six samples in each lane and sequenced (Illumina HiSeq 2000 v4 Single-Read 50 bp). For all samples, low quality bases were removed with Fastq_quality_filter from the FASTX Toolkit 0.0.13 (http://hannonlab.cshl.edu/fastx_toolkit/index.html) with 90% of the read needing a quality phred score > 20. Homertools 4.7 (Heinz et al., 2010) were used for PolyA-tail trimming, and reads with a length < 17 were removed. PicardTools 1.78 (https://broadinstitute.github.io/picard/) were used to compute the quality metrics with CollectRNASeqMetrics. With STAR 2.3 (Dobin et al., 2013), the filtered reads were mapped against mouse genome 38 (mm10) using default parameters. Count data were generated using featureCounts (Liao et al., 2014)(parameters --minReadOverlap 3 -T 3 -M -O) for the genes annotated in the gencode.vM16.gtf file. For the comparison with DESeq2 (Love et al., 2014), the input tables containing the replicates for groups to compare were created by a custom perl script. For DESeq2, DESeqDataSetFromMatrix was applied, followed by estimateSizeFactors, estimateDispersions, and nbinomWald testing. The result tables were annotated with gene information (gene symbol) derived from the gencode.vM8.gtf file. We then further analyze the RNA sequencing results using the Ingenuity Pathway Analysis (IPA, QIAGEN). The accession number of the RNA sequencing data is GSE112715 and available at Genome Expression Omnibus.

QUANTIFICATION AND DATA ANALYSIS

Statistical Analysis

All the data were presented as mean ± SD (error bar) unless otherwise specified. Where indicated, p values were determined by a two-tailed Student’s t-test. We used GraphPad Prims (v 7.0.3) to do the statistical analysis. p<0.05 was considered statistically significant.

DATA AND SOFTWARE AVAILABILITY

The GEO accession number for the RNA sequencing is GSE112715.

KEY RESOURCES TABLE

REAGENT RESOURCE	SOURCE	IDENTIFIER
Antibodies
PerCP anti-human CD8a	BioLegend	BioLegend Cat# 300922, RRID:AB_1575072
APC/Cy7 anti-human CD4	BioLegend	BioLegend Cat# 357416, RRID:AB_2616810
PE anti-human CD197 (CCR7)	BioLegend	BioLegend Cat# 353204, RRID:AB_10913813
APC anti-human CD45RA	BioLegend	BioLegend Cat# 304112, RRID:AB_314416
PE/Cy7 anti-human IFN-γ	BioLegend	BioLegend Cat# 502528, RRID:AB_2123323
PE anti-human TNF-α	BioLegend	BioLegend Cat# 502909, RRID:AB_315261
PerCP/Cy5.5 anti-mouse CD8a	BioLegend	BioLegend Cat# 100734, RRID:AB_2075238
APC/Cy7 anti-mouse/human CD44	BioLegend	BioLegend Cat# 103028, RRID:AB_830785
Brilliant Violet 421™ anti -rat CD90/mouse CD90.1	BioLegend	BioLegend Cat# 202529, RRID:AB_10899572
APC/Cy7 anti-rat CD90/mouse CD90.1 (Thy-1.1)	BioLegend	BioLegend Cat# 202520, RRID:AB_2303153
FITC anti-mouse CD8a	BioLegend	BioLegend Cat# 100804
PE anti-mouse CD8b	BioLegend	BioLegend Cat# 126608, RRID:AB_961298
Rat Anti-Mouse CD90.2 Monoclonal, FITC	Miltenyi Biotec	Miltenyi Biotec Cat# 130-091-602, RRID:AB_244295
APC anti-mouse CD45.1	BioLegend	BioLegend Cat# 110714, RRID:AB_313503
PE anti-mouse CD45.2	BioLegend	BioLegend Cat# 109808, RRID:AB_313445
PE/Cy7 anti-mouse KLRG1	BioLegend	BioLegend Cat# 138416
APC anti-mouse CD127 (IL-7Rα)	BioLegend	BioLegend Cat# 121122
Annexin V-FITC Kit	BioLegend	Miltenyi Biotec Cat# 130-092-052
PE/Cy7 anti-mouse IFN-γ	BioLegend	BioLegend Cat# 505826, RRID:AB_2295770
PE anti-mouse TNF-α	BioLegend	BioLegend Cat# 506306, RRID:AB_315427
Rabbit Anti-eIF2alpha, phosphor (Ser51) mAb	Cell Signaling Technology	Cell Signaling Technology Cat# 3597, RRID:AB_390740
Phospho-S6 Ribosomal Protein (Ser235/236) (D57.2.2E) mAb	Cell Signaling Technology	Cell Signaling Technology Cat# 4858, RRID:AB_916156
BiP (C50B12) Rabbit mAb	Cell Signaling Technology	Cell Signaling Technology Cat# 3177P, RRID:AB_10828008
CHOP (D46F1) Rabbit mAb	Cell Signaling Technology	Cell Signaling Technology Cat# 5554S, RRID:AB_10694399
XBP-1s (D2C1F) Rabbit mAb	Cell Signaling Technology	Cell Signaling Technology Cat# 12782, RRID:AB_2687943
Phospho-PERK (Thr980) (16F8) Rabbit mAb	Cell Signaling Technology	Cell Signaling Technology Cat# 3179S, RRID:AB_2095853
SPTLC2 (C-term) antibody	OriGene	Cat# TA319780
Phosphor-NF-kB p65 Ser536 (93H1) Rabbit mAb antibody	Cell Signaling Technology	Cell Signaling Technology Cat# 3033
Phospho-p38 MAPK (Thr180/Tyr182) (D3F9) XP® Rabbit mAb	Cell Signaling Technology	Cell Signaling Technology Cat# 4511
Phospho-SAPK/JNK (Thr183/Tyr185) (81E11) Rabbit mAb	Cell Signaling Technology	Cell Signaling Technology Cat# 4668
Phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) (D13.14.4E) XP® Rabbit mAb	Cell Signaling Technology	Cell Signaling Technology Cat# 4370,
Phospho-Akt (Ser473) (D9E) XP® Rabbit mAb	Cell Signaling Technology	Cell Signaling Technology Cat# 4060,
Phospho-Akt (Thr308) Antibody	Cell Signaling Technology	Cell Signaling Technology Cat# 9275
Phospho-4E-BP1 (Ser65) Antibody	Cell Signaling Technology	Cell Signaling Technology Cat# 9451
GRP94 antibody	Cell Signaling Technology	Cell Signaling Technology Cat# 20292,
HRP Donkey anti-rabbit IgG	BioLegend	Cat# 406401
LEAF™ Purified anti -mouse CD3	BioLegend	Cat# 100223
Ultra-LEAF™ Purified anti -mouse CD28	BioLegend	Cat# 102116
LEAF™ Purified anti -human CD2	BioLegend	Cat# 309212
LEAF™ Purified anti-human CD28	BioLegend	Cat# 302923
LEAF™ Purified anti -human CD3	BioLegend	Cat# 317315
Chemicals, Peptides, and Recombinant Proteins
Agarose, low melting	Thermo Fischer	Cat# 10377033
Rapamycin	Enzo Life Sciences	Cat# BML-A275–0005
Salubrinal	Santa Cruz	Cat# sc-202332
Microcystin LR	Focus Biomolecules	Cat# 10–2719
3-keto-sphinganine	Abcam	Cat# ab144107
D-erythro Sphinganine	Cayman Chemical	Cat# 10007945–50
C18-dihydro-ceramide (n-stearoyl-)	Promochem	Cat# LA 56-1077-4
C16 Ceramide (N-Palmitoylsphingosine, D-erythro)	Santa Cruz	Cat# sc-201379
Sphingosine	Echelon Biosciences	Cat# S-1000
Sphingosine-1-phosphate	Echelon Biosciences	Cat# 10–4526
Palmitoyl Sphingomyelin	Cayman Chemical	Cat# 10007946–1
Ficoll	Sigma-Aldrich	Cat# F2637
1-deoxysphinganine	Avanti Polar Lipids	Cat# 860493
1-deoxymethylsphinganine	Avanti Polar Lipids	Cat# 860473
Percoll	Sigma-Aldrich	Cat# 17-0891-01
Clarity Western ECL Substrate	Bio-Rad Laboratories	Cat# 1705061
Fixation Buffer	BioLegend	Cat# 420801
HBSS	GIBCO	Cat# 14175129
PBS	GIBCO	Cat# 20012–068
L-Glutamine	GIBCO	Cat# 25030–024
RPMI 1640 Medium	GIBCO	Cat# 21875034
Non-essential amino acids	GIBCO	Cat# 11140–035
Fetal bovine serum	Sigma-Aldrich	Cat# F0804
CellTrace ™ Violet	Thermo Fischer	Cat# C34557
13C3, 15N-L-Serine	Sigma-Aldrich	Cat# 608130
GP (33–41) peptide	GenScript	Cat# RP20091
Recombinant Mouse IL-2	BioLegend	Cat# 575408
Recombinant Mouse IL-12 (p70)	BioLegend	Cat# 577004
IFN-gamma	PeproTech	Cat# 315-05-20
IFN-alpha	PeproTech	Cat# 300–02A
DbGP33–41 tetramer	NIH tetramer core facility	Task# 31755
DbNP396–404 tetramer	NIH tetramer core facility	Task# 31756
Other resources
PVDF membrane	Santa Cruz	Cat# sc-358811
RNeasy Mini Kit	QIAGEN	Cat# 74106
RNase-Free DNase Set	QIAGEN	Cat# 79254
Experimental Models: Organisms/Strains
LCMV-Armstrong	(Dutko and Oldstone, 1983)	Strain# CA1371
Mouse: C57BL/6N	Charles River	Strain# 027
Mouse: SPTLC2tm2.1Jia	(Li et al., 2009)	MGI:4414752
Mouse: Tg(TcrLCMV)327Sdz (P14)	(Pircher et al., 1989)	MGI:2665105
Mouse: Tg(Cd4-cre)1Cwi/BfluJ	(Lee et al., 2001; Sawada et al., 1994)	JAX Strain# 017336
Software and Algorithms
Flowjo	FlowJo, LLC	https://www.flowjo.com/
ImageJ	NIH	https://imagej.nih.gov/ij/
Deposited data
RNA-seq data		GEO: GSE112715

HIGHLIGHTS.

• CD8⁺ T cell function and survival is impaired in HSAN-I patients with SPTLC2 mutation

• Mouse CD8⁺ T cells require SPTLC2 to protect against viral infections

• SPTLC2-mediated sphingolipid synthesis prevents mTORC1 hyperactivation and cell death

• Sphingolipid supplementation restores SPTLC2-deficient CD8⁺ T cell effector function