Anti-AKT antibody (4691S), antibody to AKT phosphorylated at Ser473 (4060S), antibody to AKT phosphorylated at Thr308 (13038), anti-mTOR antibody (2983S), antibody to mTOR phosphorylated at Ser2448 (5536S), anti-PP2Cδ antibody (11901S), anti-FoxO1 antibody (2880), anti-Rictor antibody (9476S), anti-Raptor antibody (2280S), anti-H3K36me2 antibody (2901S), anti-H3K27me3 antibody (9733S), antibody to ACLY phosphorylated at Ser455 (4331S), anti-Phospho-p70S6Kinase (97596S), anti-p70S6 Kinase (9202s), anti-Phospho-4E-BP1 (2855), anti-4E-BP1 (9644S), anti-H3K9/14Ac antibody (9677S), anti-H3K27Ac antibody (8173S) were from Cell Signaling technology; anti-NSD2 antibody (ab75359), anti-Phospho-serine (ab9332), anti-ACLY (ab40793) were from Abcam; anti-CUL4B antibody (20882-1-AP) was from Proteintech; anti-NDRG1 antibody (abs136385) and antibody to NDRG1 phosphorylated at Thr346 (abs140323) were from Absin.

PP2Cδ^-/- mice on the C57BL/6J background were kindly provided by Pro. Zhenyu Ju (Institute of Ageing Research, Hangzhou Normal University). OT-II TCR-transgenic mice on the C57BL/6J background were from Pro. Zhijian Cai (School of Medicine, Zhejiang University). Wild-type control mice with same genetic background were used. All mice were housed under specific pathogen-free conditions. Animal experiments were carried out according to the National Institutes of Health Guide for the Care and Use of Laboratory Animal, and approved by the Animal Care and Use Committee of Nanjing University of Chinese Medicine.

Wild type and PP2Cδ^-/- female mice (8 weeks old) were immunized subcutaneously (s.c.) with 200 μg MOG_35-55 emulsified in complete Freund’s adjuvant (CFA). Animals were also intraperitoneally (i.p.) injected with 400 ng pertussis toxin (PTX; List labs) at 0 h and 48 h after immunization. The clinical score of EAE was rated using the following scale: 0, no clinical symptoms; 1, tail paralysis; 2, hindlimb weakness; 3, unilateral hindlimb paralysis; 4, complete hindlimb paralysis; 5, moribund or death (37).

BMDCs from WT and PP2Cδ^-/- mice were treated with 10 ng/mL LPS for 24 h and pulsed with 50 μg/mL MOG_35-55 for 12 h. 3 x 10⁶ cells were then administered subcutaneously into the flank region of C57BL/6 mice every 5 days for a total of 4 times. The mice were also intraperitoneally (i.p.) injected with 200 ng PTX at 0 h and 48 h after each cell immunization. The clinical scores were recorded after the last immunization (37).

Spinal cords from EAE mice were fixed in 4% paraformaldehyde for 2 days. Then, the specimen was embedded in paraffin for sectioning. The paraffin sections were stained with hematoxylin & eosin (H&E) and Luxol Fast Blue (LFB) using standard procedures to evaluate CNS inflammation and demyelination.

Bone marrow (BM) cells isolated from 6 weeks’ WT or PP2Cδ^-/- mice were cultured in complete RPMI-1640 medium supplemented with 20 ng/mL rmGM-CSF (PeproTech), and 10 ng/mL IL-4 (PeproTech). After 3 days, half of the medium was removed, and a fresh medium containing GM-CSF and IL-4 was added. On the sixth day, loose semi-adherent cells were collected, washed and subjected to positive selection with CD11c MicroBeads UltraPure (Miltenyi Biotec).

BMDCs were stimulated with 10 ng/mL LPS for 24 h and pulsed with 1 μg/mL OVA_323-339 for 12 h, and then co-cultured with OT-II CD4⁺T cells (labeled with CFSE) to a scale of 1:5 (DC/T cell) for 72 h. The differentiation was identified by flow cytometry and the supernatants were collected for cytokines examination by ELISA.

Intracellular ROS level was detected using ROS Assay Kit (Beyotime) according to the manufacturer’s instructions. Briefly, DCs, with or without LPS stimulation, were incubated with 10 μM DCFH-DA at 37°C for 20 min. DCFH-DA was hydrolyzed into DCFH by intracellular esterase and further oxidized by ROS to produce fluorescent DCF. The DCF fluorescence was assessed by flow cytometry.

Commercially available ELISA kits (DAKEWE) were used to quantify supernatant levels of cytokines in BMDCs and T cells following the manufacturer’s instructions.

Plasmids or siRNA were transfected into cells using X-tremeGENE HP DNA Transfection Reagent (Roche) or X-tremeGENE siRNA Transfection Reagent (Roche), respectively, according to the manufacturer’s instructions. The siRNA sequences are in Table S1 .

For immunoblotting analysis, cells were lysed with RIPA buffer containing PMSF. Equal amounts of protein were separated on 10% SDS-PAGE and transferred onto Immobilon membranes (Millipore) followed by blocking in Tris-buffered saline with Tween-20 (TBST) containing 5% skim milk (BD). Membranes were then incubated with the indicated primary antibodies overnight, followed by the incubation of fluorescent secondary antibodies. The immunoblotted proteins were visualized with an ECL detection reagent (Yeasen).

For co-immunoprecipitation analysis, cells were lysed with NP-40 lysis buffer containing PMSF. 300 μg of cell lysis were incubated with Dynabeads (1001D, Invitrogen) that were pre-coupled with 3 μg target primary antibodies for 1h at room temperature. Dynabeads-Ab-Ag complex was then collected, washed 3 times and resuspended in elution buffer, and heated for 5 min at 100°C in SDS buffer for immunoblotting.

Total RNA was extracted using TRIzol reagent (Invitrogen), and cDNA was synthesized with the HI Script II One-Step RT-PCR kit (Vazyme). Quantitative RT-PCR was performed using qPCR SYBR Green Master Mix (Yeasen). The mRNA expression level was calculated by applying the 2^-△△Ct method. The primer sequences are in Table S1 .

For cell surface markers analysis, single-cells in the suspensions, after FcR blocking, were respectively incubated with FITC-anti-CD40 (11-0402-82, eBioscience), APC-anti-CD80 (17-0801-82, eBioscience), PE-anti-CD86 (12-0862-82, eBioscience), PE-anti-MHC Class II (12-5321-82, eBioscience), PerCP-Cyanine5.5-anti-CD11b (45-0112-82, eBioscience), FITC-anti-CD8a (11-0081-82, eBioscience), APC-anti-CD11c (85-17-0128-41, MultiSciences) for 30 min at 4°C. Cells were then washed and detected by flow cytometry.

For intracellular cytokines analysis, splenocytes were incubated in RPMI1640 containing PMA (50 ng/ml), ionomycin (1 μg/mL) (MultiSciences), and Golgi-stop (BD) at 37°C for 6 h. Cells were then washed and incubated with FITC-anti-CD4 (11-0041-81, eBioscience) and APC-anti-CD3 (17-0031-82, eBioscience) for 30 min. After that, cells were fixed and permeabilized with the Fixation/Permeabilization Kit, followed by staining with PE-anti-IFN-γ or PE-anti-IL-17A for 1 h. For regulatory T cell analysis, cells were incubated with FITC-anti-CD4 and PE-anti-CD25 for 30 min, followed by fixation, permeabilization, and staining with APC-anti-Foxp3 (Mouse Regulatory T cell, Invitrogen).

For proliferation assay, DCs, on day 3 and day 6 of culture, were labeled with BrdU for 3 h before analysis. For apoptosis assay, DCs were collected and performed using FITC Annexin V Apoptosis Detection Kit (BD) as per the manufacturer’s instructions.

MitoTracker Green, Mito Tracker Red, and Mito SOX staining were performed according to the manufacturer’s instructions (Invitrogen).

The chromatin immunoprecipitation (ChIP) assay was carried out using SimpleChIP plus Sonication Chromatin IP Kit according to the manufacturer’s instructions (CST). In short, BMDCs were stimulated with LPS for 24 h and fixed with 1% formaldehyde solution. Solution chromatin was immunoprecipitated with anti-NSD2 (Abcam), H3K9/14Ac, or anti-IgG antibody (CST) respectively. Precipitated DNA and input DNA were assessed by real-time PCR.

Oxygen consumption rates (OCR) and extracellular acidification rates (ECAR) of BMDCs were analyzed with an XF-96 Extracellular Flux Analyzer (Seahorse Bioscience) as described (38). Intracellular ATP concentrations were measured with an ATP Bioluminescence Assay Kit (Beyotime). Briefly, cells were lysed, centrifuged at 12, 000 g at 4°C for 5 min and mixed with luciferase reagent. Luminescence was detected using a Multiscan Spectrum (PerkinElmer).

For mitochondrial DNA quantification, total DNA were extracted from BMDCs, with or without LPS treatment for 6 h, using the FlexiGene (QIAGEN) as per the manufacturer’s instructions. Three pairs of primers against mtDNA and one pair of primers against genomic DNA were used for amplification. The primer sequences are in Table S1 .

BMDCs were stimulated with or without LPS for 12 h, fixed with 2.5% glutaraldehyde for 2 h, and post-fixed in 1% osmium acid in 0.1 M phosphate buffer at room temperature for 2 h. cell samples were embedded for 48 h after dehydration and penetration. Ultrathin sections (60-80 nm) were then prepared using an ultramicrotome (Leica UC7), stained with 2% uranyl acetate and lead citrate for 15 min. The images were finally captured by the HT7700 transmission electron microscope (HITACHI Co.).

The proliferation capacity of BMDCs was detected using MTT assay (YEASEN). DCs were plated at a density of 4000 cells/well in 96-well plates and cultured in complete RPMI-1640 medium supplemented with rmGM-CSF and IL-4 (PeproTech) for six days. Plates of cells were added with 5 mg/mL MTT solution and dissolved in DMSO 3h later. The absorbance was measured at 490 nm.

The level of acetyl CoA was detected using PicoProbe Acetyl CoA Assay Kit (fluorometric) (ab87546) following the manufacturer’s instructions. BMDCs were suspended in 500 μl of the assay buffer and lysed using a glass homogenizer on ice for 20 passes. The free CoA in samples was quenched and acetyl CoA was converted to CoA, which then formed NADH to interact with PicoProbe. The fluorescence signal was detected at Ex/Em = 535/589 nm with a Multiscan Spectrum (PerkinElmer).

All of the data, from one representative experiment of three independent experiments, are presented as mean ± SD of three technical replicates. For the animal experiments, the representative results are from one of two or three independent experiments. The statistical significance of the differences between two groups was analyzed with Student’s t-test. Multiple group comparisons were performed by two-way ANOVA followed by Bonferroni’s post hoc t-test. All of the calculations were performed using the Prism software program for Windows (GraphPad Software). A p value of 0.05 or less was considered statistically significant.

Given the documented importance of PP2Cδ in the immune system (29–34), we sought to further explore its potential role in DCs development and function. For this, we firstly evaluated the impact of PP2Cδ on the in vivo generation of DCs. Flow cytometry analysis of splenocytes revealed that the frequencies of conventional DCs (cDCs, CD11c^hiMHC⁺) and plasmacytoid DCs (pDCs, CD11c^intB220⁺) were mildly affected by PP2Cδ loss, but their numbers were substantially decreased ( Figures S1A, B ). Within cDCs, PP2Cδ ablation caused a profound decrease in counts of CD8α^- cDC2 ( Figure S1B ). The results thus indicated that PP2Cδ was essential for DC development.

Next, we detected the role of PP2Cδ in DCs differentiation by exploiting the in vitro culture of bone marrow (BM) cells. Strikingly, the frequency and count of CD11c⁺ DCs developed from hematopoietic stem cells (HSCs) were reduced upon PP2Cδ deletion ( Figure 1A and Figure S1C ). Few quantities of pDCs were generated in this setting likely due to the inhibitory effect of GM-CSF as described previously (39). We also noted that DCs from PP2Cδ lacking mice, though in the diminished numbers, exhibited much larger cellular mass at the end of the 6d culturing period ( Figure 1B and Figure S2A ). The results implied that PP2Cδ^-/- DCs might be in highly biosynthetic and metabolic status compare with their counterparts.

Considering that cellular numbers were generally the net outcome of cell demise and replication, we therefore proceeded to examine cellular apoptotic and proliferative rates. Indeed, the apoptotic rate of PP2Cδ^-/- DCs, upon LPS stimulation, was markedly elevated compared with that of control cells ( Figure 1C ). This pro-apoptotic effect of PP2Cδ loss was likely due to the increased level of cellular ROS ( Figure S2B ). We further assessed the proliferative rate of DCs by applying bromodeoxyuridine (BrdU) incorporation assay and MTT assay. As shown in Figure 1D , PP2Cδ^-/- DCs displayed blunted proliferative capability compared with WT cells, particularly at the later development stage. In parallel, PP2Cδ lacking DCs demonstrated to increasingly express pro-apoptotic factors such as Bax and P53, while suppressing that of pro-survival factors like Bcl2 and XIAP, and proliferation-associated factors like CDK1, 2, and 4 ( Figure S2C ). Thus, our data indicated that loss of PP2Cδ led to compromised proliferative and surviving potential of DCs, which likely accounted for the diminished numbers of DCs developed from HSCs.

DCs have been recognized as a key player in regulating immune response or tolerance, which is largely determined by their maturation and activation status (1, 3). We therefore further examined the effect of PP2Cδ on the activation and maturation of DCs. Remarkably, PP2Cδ^-/- DCs, compared with WT cells, expressed a much higher level of co-stimulatory molecules including CD40, CD80, CD86, and MHC class II following LPS stimulation ( Figure 1E ). In parallel, the cells produced elevated amounts of proinflammatory cytokines IL-1β, IL-6, IL-12, and TNF-α, and suppressed the expression of the immune regulatory cytokine IL-10 ( Figure 1F ). The data thus revealed an important role for PP2Cδ in controlling DCs maturation and activation, loss of which shifted DCs differentiation into the hyperactivated and proinflammatory phenotype.

DCs are the most powerful APCs capable of presenting antigens to naïve CD4⁺ T cells and inducing T cells differentiation and reaction (2). To further define the role of PP2Cδ in DC priming activity, we then exploited the OT-II T cells that are characterized by OVA-specific CD4⁺ T cell receptor (TCR). DCs were firstly matured upon LPS stimulation and pulsed with OVA peptide, followed by co-culture with OT II CD4⁺ T cells. As a consequence, T cells stimulated with PP2Cδ^-/- DCs were more proliferative than those induced by WT DCs ( Figure 2A ). These cells secreted much higher levels of IFN-γ and IL-17 ( Figure 2B ). In line, PP2Cδ deficient DCs promoted development of Th1 and Th17 cells while repressing the differentiation of CD4⁺CD25⁺Foxp3⁺ Treg cells ( Figure 2C ). Associated with this, PP2Cδ^-/- DCs produced increased levels of IL-12, IL-6, and IL-23, the cytokines essential for instructing differentiation of Th1 and Th17. Conversely, PP2Cδ ablating DCs repressed the expression of tumor growth factor-β (TGF-β) and IL-10 favoring Treg generation ( Figure 2D ). The data thus indicated that loss of PP2Cδ boosted the antigen-presenting activity of DCs and promoted their ability to induce Th1/Th17 development while restricting Treg differentiation.

To further test the in vivo functional relevance of these observations, we then exploited experimental autoimmune encephalomyelitis (EAE), a well-appreciated mice model of multiple sclerosis (MS) that are presumably driven by DC-primed Th1 and Th17 cells (4, 40). Remarkably, compared with WT control littermates, PP2Cδ^-/- mice displayed accelerated EAE onset and exaggerated disease symptoms upon immunization with myelin oligodendrocyte glycoprotein (MOG)_35-55 peptide ( Figure 2E ). Exacerbated inflammation and demyelination were observed in lesions of these mice ( Figure 2F ). Consistently, PP2Cδ deficient mice exhibited augmented counts of cDCs and CD4⁺ T cells, along with the elevated ratios of Th1 (IFN-γ⁺) and Th17 (IL-17⁺) cells in spine cords ( Figures 2G–I ). We thus provided the in vivo evidences to substantiate the significance of PP2Cδ in controlling DC-driven T cell responses. To exclude the possibility that PP2C might act through other cells than DCs in this setting, we additionally performed passive EAE experiments wherein DCs were pre-activated, loaded with MOG peptide, and transplanted for inducing antigen-specific immune response in mice. Expectedly, mice receiving PP2Cδ^-/- DCs displayed earlier onset of EAE with more severe disease symptoms when compared with animals taking WT DCs ( Figure 2J ). Taken together, our data indicated that PP2Cδ played a pivotal role in regulating DCs’ antigen-presenting and T cell stimulatory properties, which was crucial for self-limiting immune response and preventing pathogenic Th1/Th17-driven diseases.

Cellular metabolic programs, especially glycolysis and mitochondrial oxidative phosphorylation (OXPHOS), are essential for shaping DC phenotype and functionality (8, 12, 41). Therefore, we further investigated the effect of PP2Cδ on DC metabolism by initially examining the physiology and function of mitochondria, the primary energy supplier as well as the nexus of multiple signaling pathways (42). As shown in Figure 3A , while WT control DCs were featured with smaller and round mitochondria dispersed in the cytosol, PP2Cδ lacking DCs displayed larger and elongated mitochondria mostly linked to endoplasmic reticule (ER), indicative of a higher level of metabolic activity (43). In accordance, PP2Cδ^-/- DCs exhibited increased mitochondrial mass and less dysfunctional mitochondria, along with the elevated mtDNA level, mtROS release, and ATP generation particularly following LPS stimulation ( Figures 3B–E and Figure S3A ). These cells consistently demonstrated an elevated oxygen consumption rate (OCR), the measurement of mitochondrial respiratory capability ( Figure 3F ). Supportively, the expression of genes representative of mitochondrial respiratory chain complex was also augmented in PP2Cδ^-/- DCs relative to their counterparts ( Figure 3G ).

Anaerobic glycolysis, though energy inefficient, is critical for synthesizing intermediates and macromolecules to bolster cellular activation and function (41). Compared with WT cells, PP2Cδ^-/- DCs displayed significantly enhanced glycolytic activity, as evidenced by elevated extracellular acidification rate (ECAR), increased expression of glycolysis-associated molecules, and augmented lactate production ( Figures 3H–J ). Taken together, our data indicated that PP2Cδ appeared to serve as a metabolic checkpoint to confine mitochondrial respiration and glycolytic metabolism of DCs, loss of which caused exaggerated cellular metabolism compatible with hyperactivated DCs.

Next, we sought to identify the molecular mechanism underlying PP2Cδ-mediated DC regulation. Given its role in controlling both activity and metabolism of DCs, we wondered whether mechanistic target of rapamycin (mTOR), a master factor integrating cellular metabolism and immunological signaling (6–8), might participate in the action mode of PP2Cδ. Our initial data demonstrated that mTOR was increasingly activated in PP2Cδ^-/- DCs relative to WT cells following LPS stimulation. Phosphorylation of Akt at Ser473, indicative of mTORC2 activation, was consistently increased in PP2Cδ^-/- DCs. However, phosphorylation of Akt at Thr308 and p70S6K, which was presumably mediated by mTORC1, was not significantly altered. We also observed increased level of NDRG1 and decreased expression of Foxo1, the major signaling events downstream of mTORC2 (21, 44), in PP2Cδ^-/- DCs relative to WT cells on LPS stimulation ( Figures 4A, B ). The results thus indicated that loss of PP2Cδ induced the enhanced activation of the mTORC2/Akt pathway in DCs.

We next examined the functional relevance of mTORC2 pathway during DCs activation. Due to the lack of specific mTORC2 inhibitors, we used the mTORC1/2 inhibitor, Torin1, in subsequent studies ( Figure S4A ). Notably, Torin1 treatment largely abolished the effect of PP2Cδ loss on DCs activity, as the key parameters such as the elevated levels of cell size and activation markers in PP2Cδ^-/- DCs were substantially reduced to the level comparable to WT cells ( Figures S4B, C ). Also, the augmented production of IL-1β, IL-6, and IL-23 by PP2Cδ^-/- DCs was abated while the reduced IL-10 level was resumed upon Torin1 treatment ( Figure S4D ). Significantly, the polarized Th1/Th17 cell differentiation primed by PP2Cδ^-/- DCs was reversed following Torin1 treatment, and the generation of Treg cells were restored ( Figures S4E, F ). Collectively, our data indicated that unbridled activation of mTOR, specifically mTORC2 pathway, contributed substantially to deregulated DCs activation and function under PP2Cδ ablation.

As is known, Raptor and Rictor constitute the essential component of mTORC1 and mTORC2 respectively to drive distinct pathways and exert differential effects. In line with the observation that PP2Cδ specifically suppressed mTORC2 pathway in DCs, our data demonstrated that PP2Cδ ablation remarkably elevated Rictor level but marginally affected Raptor expression ( Figure 4C ). The findings prompted us to further investigate how PP2Cδ regulated the level of Rictor. Considering that PP2Cδ was able to co-operate with the epigenetic modulators such as methyltransferase or acetyltransferase to mediate its effector function (21, 45), we then embarked the study on the potential epigenetic regulation by PP2Cδ on Rictor. Of interest, we noted that the expression of NSD2, a histone methyltransferase essential for immunoregulation and tumorigenesis, was remarkably enhanced in PP2Cδ^-/- DCs relative to WT cells. Along with this, the enrichment of H3K36me2, an active epigenetic mark catalyzed by NSD2 (46, 47), was increased in PP2Cδ lacking DCs. Conversely, the abundance of H3K27me3, a repressive methylation addition, was shown to be repressed ( Figure 4D ). By performing chromatin immunoprecipitation (ChIP) assay, we further revealed that NSD2 specifically bound to the genomic locus of Rictor in PP2Cδ^-/- DCs particularly following LPS stimulation ( Figure 4E ). Accordingly, the expression of Rictor was increased in PP2Cδ^-/- DCs, which however was abolished upon NSD2 silencing ( Figure 4F and Figure S5A ). The data thus indicated that the methyltransferase NSD2, controlled by PP2Cδ, contributed to the transcriptional induction of Rictor. To be functionally relevant, interference of NSD2 expression largely abrogated the effect of PP2Cδ ablation on DC activity, as demonstrated by a profound reduction in cellular apoptosis, expression of activation markers, secretion of proinflammatory cytokines, as well as by the reversed Th1/Th17 and Treg differentiation in PP2Cδ^-/- DCs upon NSD2 knockdown ( Figures 4G-K ).

Then the question to be addressed was how PP2Cδ modulated NSD2 level during DCs activation. As is known, PP2Cδ belongs to the serine/threonine phosphatase family, that has an important role in post-translational regulation of protein stability. We thus set to examine whether PP2Cδ would modulate the stability of NSD2 protein through regulating its phosphorylating status. Strikingly, the level of phosphorylated NSD2 (p-Ser) was shown to be higher in PP2Cδ^-/- DCs relative to WT cells following LPS-stimulation ( Figure 5A ). Co-IP assay confirmed that PP2Cδ interacted with NSD2 in WT but not PP2Cδ^-/- DCs. This interaction was further corroborated in 293T cells with enforced expression of PP2Cδ and NSD2 ( Figures 5B, C ).

To give a more detailed picture of the PP2Cδ-NSD2 interaction, we then constructed a series of plasmids containing the truncated mutants of PP2Cδ gene. The results showed that, besides the intact fragment, the PP2Cδ truncate lacking B domain (Δ77-361) remained to associate with NSD2. However, the mutants deficient in the N-terminal (77-598) or C-terminal (1-361) failed to bind to NSD2, indicating that the N- and C-terminal but not the catalyze domain (B) of PP2Cδ were required for its association with NSD2 ( Figure 5D ). Nevertheless, the integrity of catalyzing domain proved to be essential for fulfillment of PP2Cδ phosphatase activity, because introduction of the phosphatase-dead mutant (D317A) into PP2Cδ^-/- DCs failed to de-phosphorylate NSD2 as the intact PP2Cδ constructs did ( Figure 5E ). Together, the data indicated that PP2Cδ was able to physically associate with NSD2 and mediate the de-phosphorylation effect.

Next, we examined whether PP2Cδ-mediated de-phosphorylation of NSD2 would affect its expressive level. Indeed, NSD2 level was reduced in WT but not PP2Cδ^-/- DCs upon LPS stimulation, and this reduction was alleviated by treatment of the proteasome inhibitor MG132 ( Figure S5B ). The result suggested that PP2Cδ might act through the ubiquitin-proteasome pathway to promote NSD2 degradation. Supportively, we observed that K48-linked ubiquitination of NSD2, the prerequisite for proteasomal degradation, was enhanced in WT DCs relative to PP2Cδ^-/- cells following LPS stimulation ( Figure 5F ). To further understand this regulatory pathway, our attention was then turned to CUL4^DCAF2, an E3 ubiquitin ligase that was critically involved in epigenetic regulation and DC modulation (48, 49). Impressively, we confirmed the interaction between CUL4^DCAF2 and NSD2 using Co-IP assay. This association was shown to be compromised upon PP2Cδ deletion, implying that PP2Cδ-mediated de-phosphorylation of NSD2 was required for this binding ( Figure 5G ). In support of this, the expression of NSD2 was elevated in PP2Cδ^+/+ DCs upon the deletion of the essential component of CRL4 ligase, DCAF2 ( Figure 5H ). Moreover, the restoration of the intact PP2Cδ but not phosphatase-dead (D317A) expression in PP2Cδ^-/- DCs promoted the ligation of CUL4^DCAF2 to NSD2, leading to the consequent K48-ubiquitination of this protein ( Figure 5I ). The data thus indicated that the phosphatase activity of PP2Cδ was required for the association of CUL4^DCAF2 with NSD2, and hence the ubiquitination and degradation of NSD2. Of functional relevance, the resumption of wild-type PP2Cδ but not phosphatase-dead mutant (D317A) was shown to abrogate hyperactivated phenotype of PP2Cδ^-/- DCs and the deregulated Th1/Th17 and Treg differentiation ( Figures S6A, B ). Together, our data demonstrated that PP2Cδ phosphatase promoted the ubiquitination and degradation of NSD2 via the CUL4 E3 ligase, thereby limiting the Rictor/mTORC2 pathway to control DC activation and functionality.

The activation of DCs is a coordinated process that integrates immunological signaling and metabolic program. Since mTOR pathway plays a central role in regulating cellular metabolism and differentiation, we thus further explored how the PP2Cδ/mTORC2 axis orchestrated the metabolic and gene programs that specified DC identity. We initially tested whether the Rictor/mTORC2 pathway was integrated into the DC regulatory program. Indeed, deletion of Rictor was shown to rectify the hyperactivated phenotype of PP2Cδ^-/- DCs, as revealed by the substantial alteration in cellular proliferation, apoptotic rate, activation markers levels, and Th1/Th17-stimulatory activity ( Figures 6A–F and Figure S7A ). Moreover, knockdown of Rictor decreased mitochondrial OXPHOS and glycolytic metabolism of PP2Cδ^-/- DCs to the level comparable to that in WT cells ( Figures 6G, H ). The data thus indicated that the Rictor/mTORC2 pathway served as the signaling nexus to link metabolic and effector programs of DCs, which prompted us to further explore the mechanism involved, particularly the metabolic-epigenetic pathway.

Recent studies have demonstrated that glycolytic byproducts such as pyruvate may through mitochondrial pyruvate carrier (MPC) enter mitochondria and be further catalyzed by ATP citrate lyase (ACLY) to generate acetyl-CoA (Ac-CoA). The generated Ac-CoA would be then transported to the nucleus for histone acetylation and genes transcription regulation (50, 51). This metabolic-epigenetic flux was thought to be driven by the mTORC2 and essential for genes expression that specified distinct subsets of immune cells (18, 22). Of interest, we observed a higher level of acetyl-CoA and increased phosphorylation of ACLY in PP2Cδ lacking DCs relative to WT cells, which, however, was reduced upon Rictor knockdown ( Figures 6I, J ). Along with this, the acetylation at H3K9/14 and H3K27, the transcription-permissive histone modifications (52), was augmented in PP2Cδ^-/- DCs, and Rictor deletion abolished this increment ( Figure 6J ). Consistent with this, boosted binding of H3Ac to the targeted genes was detected in PP2Cδ^-/- DCs relative to WT cells ( Figure 6K ). Thus, our data indicated that the Rictor/mTORC2 pathway activated ACLY and facilitated histone acetylation subsequently, which would contribute to PP2Cδ-mediated gene program in DCs. To further confirm this mechanistic linkage, we then applied UK5099 and SB204990 to block MPC or ACLY respectively. The results showed that either blocking mitochondrial pyruvate entry or inhibiting catalytic activity of ACLY corrected, though to different extent, the expression of genes characterizing hyperactivated DCs upon PP2Cδ deletion ( Figure 6L ). The results showed that a network of signature genes including the activation and metabolism-related genes (il-6, il-1β, cd80, h2-iab, and ccr7), mitochondrial and glycolytic metabolism-related genes (pgc-1α, hk-2, and pkm2), and DC-specifying transcription factors (irf4, pu.1, c-myc, nr4a1, and nr4a3) (53–55)were substantially altered upon these treatments. We thus proposed that ACLY-driven metabolic-epigenetic flux, downstream of the PP2Cδ/mTORC2 pathway, was critically involved in the gene program for DC identity. In addition, although previous study showed that lipid-derived acetyl-CoA could also fuel histone acetylation for gene resetting (56), our data indicated that PP2Cδ ablation did not affect the expression of genes essential for fatty acid metabolism ( Figure S7B ), implying that lipid metabolism might not the essential source of acetyl-CoA for epigenetic modification in our setting. Taken together, we showed that PP2Cδ acted through the Rictor/mTORC2/ACLY pathway to mediate the metabolic-epigenetic program, which was essential for activation and function of DCs ( Figure S7C ).