mTORC2 Deficiency Alters the Metabolic Profile of Conventional Dendritic Cells

In myeloid dendritic cells (DC), deletion of the mechanistic target of rapamycin complex 2 (TORC2) results in an augmented pro-inflammatory phenotype and T cell stimulatory activity; however, the underlying mechanism has not been resolved. Here, we demonstrate that mouse bone marrow-derived TORC2-deficient myeloid DC (TORC2−/− DC) utilize an altered metabolic program, characterized by enhanced baseline glycolytic function compared to wild-type WT control (Ctrl) DC, increased dependence on glycolytic ATP production, elevated lipid content and higher viability following stimulation with LPS. In addition, TORC2−/− DC display an increased spare respiratory capacity (SRC) compared to WT Ctrl DC; this metabolic phenotype corresponds with increased mitochondrial mass and mean mitochondrial DNA copy number, and failure of TORC2−/− DC mitochondria to depolarize following LPS stimulation. Our data suggest that the enhanced metabolic activity of TORC2−/− DC may be due to compensatory TORC1 pathway activity, namely increased expression of multiple genes upstream of Akt/TORC1 activity, including the integrin alpha IIb, protein tyrosine kinase 2/focal adhesion kinase, IL-7R and Janus kinase 1(JAK1), and the activation of downstream targets of TORC1, including p70S6K, eukaryotic translation initiation factor 4E binding protein 1 (4EBP1) and CD36 (fatty acid translocase). These enhanced TORC1 pathway activities may culminate in increased expression of the nuclear receptor peroxisome proliferator-activated receptor γ (Pparγ) that regulates fatty acid storage, and the transcription factor sterol regulatory element-binding transcription factor 1 (Srebf1). Taken together, our data suggest that TORC2 may function to restrain TORC1-driven metabolic activity and mitochondrial regulation in myeloid DC.

The function of mTORC1 in dendritic cells (DC) has been studied extensively using the immunosuppressive prodrug rapamycin (RAPA) (8)(9)(10). RAPA inhibition of mTORC1 in DC prevents their maturation, leading to decreased T effector cell proliferation and increased regulatory T cell (T reg ) differentiation (8,11,12). While little had been known previously about the function of RAPA-insensitive mTORC2 (referred to subsequently as TORC2) specifically in DC, we have shown recently that functional TORC2 deletion specifically in these antigen-presenting cells (APC) leads to both an enhanced proinflammatory DC phenotype and Th1/Th17 allogeneic T cell polarization and proliferation (13). Additionally, intratumoral delivery of TORC2-deficient DC delays melanoma progression in a CD8 + T cell-dependent manner (14), whereas skin grafts from donors lacking TORC2 in DC undergo enhanced CD8 + T cellmediated rejection (15). However, the mechanisms underlying these enhanced DC functions remain undefined.
These previous studies concluded that TORC2 did not play a role in regulating DC immunometabolism (34). However, the investigations were conducted using adenosine triphosphate (ATP)-competitive dual TORC inhibitors; therefore, the discrete function of TORC2 in linking metabolic programming and immune function in DC remains unclear. In addition, a recent study of tissue-resident peritoneal macrophages (26) demonstrated that TORC2 deficiency promoted their generation, and that TORC2-deficient macrophages and peritoneal resident macrophages had enhanced mitochondrial biomass, as well as an altered metabolic profile, compared to monocytederived macrophages.
In the present investigation, we have utilized DC generated from mice in which Rictor (an essential component for TORC2 assembly) is deleted specifically in CD11c + DC (35) to ascertain whether the augmented inflammatory phenotype of TORC2 −/− DC is a consequence of altered cellular metabolism. We demonstrate that TORC2 −/− DC are biased toward glycolytic metabolism, have an increased dependence on glycolysis to generate ATP, elevated lipid content and higher viability following TLR4 agonism with bacterial lipopolysaccharide (LPS). In addition, TORC2 −/− DC have augmented spare respiratory capacity (SRC), mitochondrial mass and mean DNA copy number, and mitochondria that fail to depolarize following TLR agonism, as well as differential Golgi apparatus dispersal compared to WT control (Ctrl) DC. We also show that TORC1 inhibition by rapamycin in TORC2 −/− DC abolishes their enhanced glycolytic activity and SRC. Finally, our data suggest a possible pathway via which TORC2 −/− DC display augmented TORC1 metabolic activity and through which enhanced integrin alpha IIb (Itga2b) and protein kinase 2 (Ptk2)/focal adhesion kinase (FAK) expression leads to increased hematopoetic cell signal transducer (Hcst; also known as PIK3AP) expression upstream of TORC1 activity, increased NFκβ activation, enhanced eukaryotic translation initiation factor 4E binding protein 1 (4EBP1), and subsequent increased nuclear receptor peroxisome proliferator-activated receptor gamma (Pparγ), sterol regulatory element-binding transcription factor 1 (Srebf1) and CD36 (fatty acid translocase) expression downstream of TORC1 activity. Our findings suggest that TORC2 may restrain TORC1-regulated metabolic function in myeloid DC.

MATERIALS AND METHODS
Mice C57BL/6 (B6) CD11c-CreRictor f/f (herein referred to as TORC2 DC−/− ) mice were generated (35) by crossing B6 mice in which rictor is flanked by loxP restriction digest sites (generously provided by Drs. Keunwook Lee and Mark Boothby, Vanderbilt University School of Medicine) with B6 mice expressing Cre recombinase on the CD11c promoter (CD11c-Cre; The Jackson Laboratory). The genetic background of crossed mice was verified by polymerase chain reaction (PCR) genotyping; CD11c-Crelittermates were used as negative controls.
ATP concentrations were determined using an ATP determination kit (ThermoFisher, Waltham, MA) as per the manufacturer's instructions. Where indicated, DC were stimulated with LPS (100 ng/mL) for 1 h.

NanoString Analysis
Total RNA was extracted from bead-purified CD11c + DC generated from the BM of Ctrl or TORC2 DC−/− mice using an RNeasy Mini Kit (Qiagen, Hilden, Germany) as per the manufacturer's instructions. NanoString analysis was performed using a Mouse Immunology Panel (NanoString Technologies, Seattle, WA) as described (38).

Quantitative PCR
cDNA was amplified using Platinum Quantitative PCR SuperMix-UDG (Invitrogen, Waltham, MA) in 10 µl volumes in quadruplicate with gene-specific primers and probed on the ABI Prism 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA) according to the manufacturer's instructions. Thermal cycling conditions were 50 • C for 2 min then 95 • C for 2 min, followed by 40 cycles of 95 • C for 15 s and 60 • C for 1 min. Data were analyzed using the Ct method with expression normalized to the housekeeping gene GAPDH.

Fluorescence Imaging and Quantitation of Lipid Droplets
Cell suspensions were fixed in 2% paraformaldehyde for 20 min at room temperature, washed with PBS and permeabilized in 0.1% saponin (Sigma-Aldrich, St. Louis, MO) for 10 min. The cells were then stained for 45 min with HCS LipidTox Red neutral lipid stain (Invitrogen), diluted 200-fold in PBS, according to manufacturer's instructions. Subsequently, nuclear staining was performed using 4 ′ , 6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich). Confocal images were obtained as described above. Lipid droplets were counted manually, yielding each time data from 60 cells using NIS Elements software (Nikon, Tokyo, Japan).

Transmission Electron Microscopy
Cell suspensions were fixed in 2.5% v/v glutaraldehyde then pelleted immediately in a 1.5 ml microfuge tube at 300 x G. Pellets were then post-fixed for 1 h in 1% OsO 4 , 1% K 3 Fe(CN) 6

Statistical Analyses
Results are expressed as means ± 1SD. Significances of differences between groups were determined using Student's 't'-test or one-way ANOVA Tukey's multiple comparisons test (GraphPad Prism) as indicated, with p < 0.05 considered significant.

RESULTS
TORC2 −/− DC Display Augmented Glycolytic Capacity, Glycolysis-Dependent ATP Production and Viability Compared to Ctrl DC To investigate the impact of TORC2 deletion on DC metabolic function, we assessed glycolysis via extracellular flux, as measured by basal extracellular acidification rate (ECAR). Glycolysis was elevated significantly in TORC2 −/− compared with WT control (Ctrl) DC and in both Ctrl DC and TORC2 −/− DC following LPS stimulation ( Figure 1A). Interestingly, glycolytic capacity was also increased significantly in non-stimulated TORC2 −/− DC compared to Ctrl DC, and in LPS-stimulated TORC2 −/− DC compared to Ctrl DC. Glycolytic reserve was higher in TORC2 −/− DC, both without and with LPS stimulation.
While there was no significant difference in ATP production between Ctrl and TORC2 −/− DC following inhibition of OXPHOS with oligomycin, Ctrl and TORC2 −/− DC displayed significantly decreased ATP production when glycolysis was inhibited with 2-DG ( Figure 1B; non-stimulated; Figure 1C + LPS 1 h). TORC2 −/− DC stimulated with LPS also had significantly higher viability than Ctrl DC stimulated with LPS (representative histogram Figure 1D; quantified in Figure 1E); however, the immediate glycolytic response to LPS stimulation did not different significantly between Ctrl DC and TORC2 −/− DC (representative ECAR Supplementary Figure 1A; quantified in Supplementary Figure 1B).  Figure 2A; quantified in Figure 2B). We next assessed how TORC2 deletion in DC might influence mitochondrial phenotype. Mitochondrial mass, as determined by fluorescent mitochondrial labeling (MitoTracker Green), was significantly greater in non-stimulated TORC2 −/− DC compared with non-stimulated Ctrl DC (representative histograms Figure 3A; quantified in Figure 3B) and decreased significantly in both cell populations following LPS stimulation ( Figure 3B). Nonstimulated TORC2 −/− DC had a greater mean mt DNA copy number (although not statistically significant; Figure 3C) compared with Ctrl DC; mtDNA copy number decreased significantly in both Ctrl and TORC2 −/− DC following LPS stimulation ( Figure 3C). It has been reported that macrophages stimulated with LPS undergo mitochondrial depolarization (40). However, TORC2 −/− DC mitochondria failed to depolarize significantly following LPS stimulation (as opposed to Ctrl DC that did), as assessed via uptake of cationic TMRE fluorescent dye (representative histograms Figure 3D; quantified in Figure 3E).

TORC2 −/− DC Display More Compact Golgi Stacks With Less Perinuclear Localization Compared With Ctrl DC
It has previously been posited that the critical role of aerobic glycolysis in activated DC is to produce tricaboxylic acid cycle intermediates necessary for lipogenesis and subsequent Golgi apparatus and endoplasmic reticulum expansion to support de novo protein synthesis (30). As we observed enhanced glycolytic activity in TORC2 −/− DC, we next determined whether TORC2 deficiency in DC impacted the localization,  Frontiers in Immunology | www.frontiersin.org structure, and quantity of Golgi apparatus. To determine if TORC2 deficiency in DC impacted the localization, structure and quantity of Golgi apparatus, we first immunostained DC with anti-TGN38 Ab and assessed the location of the Golgi relative to cell nuclei and plasma membrane. In Ctrl DC, the Golgi complex was perinuclear, whereas in TORC2 −/− DC the Golgi showed less intense staining and were dispersed throughout the cell (Figure 4A, top panel). We next used TEM to assess the structure of the Golgi. In Ctrl DC, the Golgi cisternae appeared more dilated, and more Golgi were observed perinuclearly; in contrast, the Golgi in TORC2 −/− DC appeared more compact, with less Golgi visible perinuclearly ( Figure 4A, bottom panel). Finally, we quantified the MFI of TGN38 staining from 3-D stacks of confocal images to assess the total Golgi content of the DC. We observed no significant differences in total Golgi content as determined by TGN38 staining in Ctrl DC compared with TORC2 −/− DC, but localization was clearly affected (Figure 4B).

Inhibition of TORC1 Activity in TORC2 −/− DC Leads to Loss of Their Enhanced SRC and Glycolytic Capacity
While the exact mechanisms remain unclear, it has been reported that TORC1 and TORC2 may exert some regulatory influence on each other, Akt and insulin-dependent PI3K signaling (41). Therefore, to elucidate whether TORC2 in DC had TORC1independent or TORC1-dependent metabolic regulation, we incubated both Ctrl DC and TORC2 −/− DC with RAPA (10 ng/mL) for 18 h, then analyzed their SRC and glycolytic capacity via extracellular flux, as in Figures 1, 2. Inhibition of TORC1 in TORC2 −/− DC abolished the increase in SRC (representative OCR Figure 5A; quantified in Figure 5B) and glycolytic activity (representative ECAR Figure 5C; quantified in Figure 5D) in non-stimulated cells compared with Ctrl DC. Inhibition of TORC1 in both LPS-stimulated Ctrl DC and TORC2 −/− DC reduced their glycolytic activity significantly (Figures 5C,D).

TORC2 −/− DC Exhibit a Distinct Gene Expression Profile From Ctrl DC
To provide insight into possible mechanisms by which TORC2 −/− DC exhibit enhanced TORC1-dependent metabolic activity, we performed gene expression analysis to identify any differences in expression of genes between TORC2 −/− DC and Ctrl DC, as represented in the heat map in Figure 6A. Five genes with enhanced expression in TORC2 −/− DC (Ptk2, IL-7R, Jak1, Itga2b, and the PI3K component Hcst; Figure 6B) were identified as upstream mediators of augmented Akt/TORC1 activity. We then performed qPCR to determine the relative expression of the downstream targets of TORC1,-the transcription factor Srebf1, the nuclear receptor Pparγ (that regulates fatty acid storage and glucose metabolism) and the transcription factor Yin Yang 1 (Yy1) that impact cell metabolism (17,18,42) in TORC2 −/− and Ctrl DC. Both Srebf1 and Pparγ were expressed at significantly higher levels in TORC2 −/− DC compared to Ctrl DC (Figure 6C), suggesting a role for these transcription factors in the altered metabolic profile of TORC2 −/− DC.

TORC2 −/− DC Exhibit Enhanced Lipid Content
To ascertain the lipid content of TORC2 −/− DC, we stained lipid droplets (LDs) with HCS LipidTox Red neutral lipid stain. We found a significant increase in LDs in TORC2 −/− DC compared to Ctrl DC (Figures 7A,B).

DISCUSSION
We and others have reported previously (13,43) that murine myeloid DC lacking functional TORC2 display an augmented pro-inflammatory phenotype and enhance Th1/Th17 and CD8 + effector T cell responses in vivo (13)(14)(15). Similar observations have been reported for mouse macrophages (44). We now demonstrate that TORC2 −/− DC have an altered metabolic profile, whereby they exhibit enhanced glycolytic activity and spare respiratory capacity, dependence on glycolysis for ATP production, enhanced mitochondrial mass and lipid production, and increased viability following TLR4 stimulation. These changes may contribute to/support the enhanced stimulatory function reported for TORC2 −/− DC (13)(14)(15)43). Earlier studies have shown that while quiescent, immature DC have relatively low metabolic needs, these needs increase upon cell activation and maturation due to bioenergetic pressure to upregulate co-stimulatory molecule and pro-inflammatory cytokine production (29). The metabolic process that permits DC to meet these enhanced metabolic demands occurs via a "switch" from oxidative phosphorylation to aerobic glycolysis. There is evidence (33) that this initial "switch" toward increased glycolytic metabolism is not dependent on TOR signaling. However, TORC1 (but not TORC2) has been described as essential for DC glycolytic commitment (32). We found that in this study, immature (non-stimulated) myeloid DC lacking functional TORC2 (Rictor) had significantly increased glycolytic function compared to immature Ctrl DC, and that the increase in glycolytic capacity and reserve was also observed following stimulation of the DC with the TLR4 agonist LPS. These findings, together with our previous observation (13) of enhanced co-stimulatory CD86, together with decreased coinhibitory programmed death ligand-1 (PD-L1) expression on TORC2 −/− DC compared with Ctrl DC, are consistent with an intermediate maturity phenotype and supported by the increased dependence of non-stimulated TORC2 −/− DC on glycolysis for ATP production that we observed.
In addition to its importance for nascent protein production, glycolytic commitment by mature DC is critical for their survival (32). Indeed, in conjunction with increased glycolytic activity, TORC2 −/− DC exhibited enhanced viability following exposure to LPS that can induce programmed cell death in DC (45). Thus, the enhanced pro-inflammatory function of TORC2 −/− DC that we have documented (13)(14)(15) in vitro and in vivo may also be attributed, in part, to their more robust viability. While T cells can trigger DC apoptosis through Fas and perforin as a mechanism to self-limit T cell activation (46,47), such enhanced viability of activated TORC2 −/− DC may augment/extend their capacity to interact with and stimulate responder T cells.
In addition to augmented glycolytic activity, TORC2 −/− DC also displayed increased SRC, in conjunction with enhanced mitochondrial biomass, increased mean mitochondrial DNA, and mitochondrial failure to depolarize following LPS stimulation. Mitochondrial SRC has been described as the extra capacity of mitochondria to produce energy under increased cell stress and is correlated with prolonged cell survival and function (48). Given the enhanced viability of stimulated TORC2 −/− DC, an increase in SRC compared with LPS-stimulated Ctrl DC corroborates our findings. As CD8 + memory T cells have enhanced SRC due to increased mitochondrial biomass (48), the increased mitochondrial content of TORC2 −/− DC that we observed compared with Ctrl DC is consistent with our other findings. Failure of the mitochondrial membrane of TORC2 −/− DC to depolarize significantly following LPS stimulation was however surprising, as LPS has been shown to collapse mitochondrial membrane potential ( m ) in macrophages (40). On the other hand, there is evidence that, in macrophages, that disproportionately utilize ATP generated via glycolysis as protection against cell death (as we observed in TORC2 −/− DC), high m is maintained via reverse functioning of F(o)F(1)-ATP synthase and adenine nucleotide translocase (40).
Enhanced glycolytic activity has been described as a means by which activated DC augment lipogenesis for increased production and transport of co-stimulatory molecules and cytokines (33). Indeed, we observed enhanced lipid production in the TORC2 −/− DC. We also assessed whether, in TORC2 −/− DC, there might be alterations in the localization, dilation and amount of Golgi apparatus that, in conjunction with the endoplasmic reticulum (ER), is critical for protein processing and transport. Surprisingly, given the enhanced glycolytic activity of TORC2 −/− DC, we did not observe any significant increase in the overall Golgi content of these cells. However, we did find that the Golgi of TORC2 −/− DC were less dilated than those in Ctrl DC, with less perinuclear localization. Recent studies (49) have favored a cisternal progenitor model of protein transport, whereby proteins are transported along tightly-compacted Golgi subcompartments that undergo fission and fusion events that allow stable protein transport to the membrane. Thus, the tightly (B) Blots were quantified using Image-Lab software and normalized to β-actin (n = 3-4 independent experiments with at least 2 mice per experiment; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). One-way ANOVA test Tukey's multiple comparisons test. NS, not significant. FIGURE 9 | Proposed mechanism by which TORC2 may restrain TORC1 activity in DC via the transcriptional suppression of upstream TORC1 activators. Based on our observations, TORC2 activity may lead to transcriptional suppression of integrin (ItgA2 and FAK) and cytokine receptor (IL-7R) and JAK1 signaling in conjunction with PIK3AP upstream of Akt/mTORC1 activity, and ultimately the metabolic regulators SREBP1 and Pparγ downstream of TORC1 activity (upregulated genes labeled in red). Itga2b, integrin alpha II b; FAK, focal adhesion kinase (Ptk2, protein kinase 2), Hcst, hematopoietic cell signal transducer; Pparγ, peroxisome proliferator-activated receptor gamma; Sregf1, sterol regulatory element transcription factor 1.
compacted Golgi cisternae and diffuse TGN localization may reflect enhanced protein transport in these cells.
Given the mitochondrial dysregulation observed in TORC2 −/− DC, we investigated possible underlying mechanisms. Cells maintain mitochondrial homeostasis by controling both the number and quality of mitochondria via mitophagy (50). Interestingly, TORC1 is an important regulator of the nuclear transcription of genes necessary for mitochondrial biogenesis through peroxisome-proliferator-activated receptor coactivator-1α (PGC-1α) and yin-yang 1(YY1) (17). In addition, TORC1 regulates mitochondrial activity via phosphorylation of the mitochondrial membrane protein Bcl-X L (51). TORC1 has also been reported to positively regulate glucose uptake (52) and mitophagy (53). As there is possible interplay between TORC1 and TORC2 through their interactions with Akt and the tuberosclerosis complexes 1/2 (TSC1/2), we assessed whether the mitochondrial dysregulation observed in TORC2 −/− DC could be attributed to compensatory augmented TORC1 activity. Indeed, as we show, inhibition of TORC1 with rapamycin abolished the enhanced glycolytic function and SRC of TORC2 −/− DC, suggesting that compensatory effects mediated through TORC1 might represent the key mechanism accounting for the enhanced mitochondrial activity and glycolytic function of TORC2 −/− DC.
In this study, we performed gene expression analysis that identified five genes of interest (upstream mediators of augmented Akt/TORC1 activity) that were differentially expressed between Ctrl DC and TORC2 −/− DC. These genes identified a signaling pathway in which the integrin subunit Itga2b, Ptk2 (also known as focal adhesion kinase; FAK) and Hcst (PIK3AP) are upregulated in TORC2 −/− DC. Integrin clustering has been shown to mediate intracellular signaling via the catalytic kinase Ptk2 (54), that leads ultimately to activation of the PI3K/Akt/TORC1 signaling pathway (55). Another possible mechanism for upregulation of TORC1 activity in TORC2 −/− DC based on these gene expression data, is upregulation of the IL-7R and Janus kinase 1 (JAK1), as IL-7R/JAK1 signaling can activate the PI3K signaling pathway (56).
We identified upregulation of two genes downstream of TORC1 signaling in TORC2 −/− DC that are known to regulate metabolic function: Pparγ and Srebf1. Activation of these transcription factors downstream of TORC1 leads to expression of lipogenic genes (57). It has been demonstrated (33) that glycolysis in DC drives lipogenesis and subsequent Golgi/ER expansion upon DC activation and more recently (42), that Pparγ signaling can also promote glycolysis in hematopoietic stem cells. Interestingly, loss of TORC2 function in yeast results in enhanced lipogenesis (58). Taken together, our data suggest that TORC2 may play a role in transcriptional control of TORC1-regulated metabolic processes, as outlined schematically in Figure 9.
Our findings suggest a novel role for mTORC2 in the negative regulation of DC metabolism. Its absence equips these APC to function with enhanced pro-inflammatory/T cell stimulatory activity, as we (13)(14)(15) and others have described (43). Myeloid DC lacking functional TORC2 are more glycolytically active and have increased dependence on glycolysis for ATP production. TORC2 −/− DC also have abnormalities in mitochondrial regulation, characterized by enhanced mitochondrial biomass and mitochondria that fail to depolarize following DC activation. The metabolic phenotype of TORC2 −/− DC is lost upon inhibition of TORC1, suggesting a significant compensatory effect of the mTORC1 pathway in TORC2 −/− DC. In addition, we have identified several genes upstream of TORC1, and the transcription factors Pparγ and Srebf1 downstream of TORC1, that are upregulated in TORC2 −/− DC. The data strongly suggest that TORC2 may function to restrain TORC1-driven anabolic metabolism in myeloid DC.

ETHICS STATEMENT
All animal procedures were performed according to an Institutional Animal Care and Use Committee-approved protocol (#16058226) in accordance with National Institutes of Health (NIH) guidelines.

AUTHOR CONTRIBUTIONS
AW, HD, YZ, and RN performed experiments, analyzed that data, and wrote the manuscript. AG provided conceptual input and analyzed data. AM performed experiments. DS provided conceptual insights and supervised the cell imaging analyses. GD and AT provided conceptual insights and edited the manuscript. AT supervised the project. All authors provided feedback on the manuscript.

FUNDING
This work was supported by National Institutes of Health grants R01 AI67541 and R01 AI118777 and T32 AI74490. HD was supported by the National Natural Science Foundation of China (81800664) and the Natural Science Foundation of Hunan Province of China (2019JJ50842). GD was supported by a Stand Up 2 Cancer Innovative Research Grant (SU2C-AACR-IRG-04-16), the NIH Director's New Innovator Award (DP2AI136598), the Alliance for Cancer Gene Therapy/Swim Across America, and the Sy Holzer Endowed Immunotherapy Fund.