Antigen Presentation by Dendritic Cells in Tumors Is Disrupted by Altered Metabolism that Involves Pyruvate Kinase M2 and Its Interaction with SOCS3

Dendritic cells (DC) play a key role in the induction of tumor-specific immune responses, especially by cross-priming, which allows the transfer of antigens from tumor cells to DCs (1, 2). However, DCs in tumor environments often seem to be phenotypically and functionally defective, as reviewed by Gottfried and colleagues (3) and Bennaceur and colleagues (4). Indeed, tumor-associated factors, including cytokines transforming growth factor-β, granulocyte macrophage colony-stimulating factor (CSF), interleukin (IL)-6, macrophage CSF, and IL-10, as well as vascular endothelial growth factor, gangliosides, prostaglandin E₂, and some metabolic products, may affect the function and differentiation of DCs (4). Additionally, hypoxia, low nutrient levels, low extracellular pH, and high interstitial fluid pressure have also emerged as important components contributing to DC dysfunction (5). There is evidence showing that hypoxia can reduce the migratory capacity of human monocyte-derived DCs (6). Experimental hypoxia also alters the morphologic characteristics, expression of cell surface markers, viability, phagocytosis, metabolic activity, and release of cytokines by macrophages (7). Despite the wealth of information about factors that affect the differentiation and function of DCs, little is known about the molecular mechanism(s) responsible for DC dysfunction in cancer.

Suppressor of cytokine signaling 3 (SOCS3), which belongs to the SOCS family of proteins, can be induced and upregulated by not only cytokines such as IL-6 and lipopolysaccharide (LPS) but also tumor-associated factors in DCs (8). SOCS3 contains an NH₂-terminal region (12 residues of a kinase inhibitory region), a central SH2 domain, and a COOH-terminal motif called the SOCS box (9, 10). SOCS3 may use its SH2 domain to bind to phosphorylated tyrosine residues in its partners such as gp130 (Y759), leptinR (Y985), and EpoR (Y401; ref. 11) to regulate their activity. Herein, we found that SOCS3 can directly interact with the pyruvate kinase type M2 (M2-PK), a key enzyme in metabolic pathway of glycolysis, thereby suppressing M2-PK activity. Because SOCS3 can be induced by a wide variety of cytokines (12–14), especially tumor-associated factors (8), the upregulation of SOCS3 in DCs may play an important role in the path toward DC dysfunction, thus causing tumor immune tolerance.

Male C57BL/6 and BALB/c mice, 4 to 6 wk old (Beijing Experimental Animal Center), were maintained in a pathogen-free animal facility for at least 1 wk before use. Experiments were performed in accordance with institutional guidelines.

Mouse bone marrow–derived DCs (BMDC) were prepared as previously described (15). Monocyte/macrophage cell line RAW264.7, murine Lewis lung carcinoma (LLC) cell line, and B16 melanoma cell line were purchased from the American Type Culture Collection. The spontaneously transformed mouse ovarian surface epithelial cell line 1D8 was a gift of Dr. Roby (University of Kansas Medical Center, Kansas City, KS; ref. 16). All of the cell lines were grown in RPMI 1640 supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin.

Phycoerythrin (PE)–labeled or FITC-labeled anti-mouse CD11c (N418), CD80 (16-10A1), CD4 (L3T4), and CD8α (Ly-2) antibodies were purchased from Pharmingen. Rabbit anti-mouse SOCS3 antibodies (Transduction-Becton Dickinson), rabbit anti-mouse M2-PK antibodies (Abcam), goat anti-SOCS3 antibodies (Santa Cruz Biotechnology), PE-labeled donkey polyclonal antibodies against the rabbit IgG-H&L-F(ab)₂ fragment (Abcam), FITC-labeled rabbit anti-goat IgG antibodies (Santa Cruz Biotechnology), horseradish peroxidase–labeled goat polyclonal antibodies against rabbit IgG-H&L (Abcam), and FITC-labeled anti-V5 antibodies (Invitrogen) were also purchased.

For immunoprecipitation, V5-tagged SOCS3 vector–transfected or control plasmid vector–transfected cells were extracted in hypotonic lysis buffer as previously described (8). The nuclei were removed by centrifugation, and then the lysates were precleared with protein G-Sepharose and isotype control antibody. Immunoprecipitation was performed with the anti-V5 antibody and protein G-Sepharose with slow agitation. The immunoprecipitates were washed and resolved by 10% SDS-PAGE and then analyzed by Western blotting. The membranes were developed using the enhanced chemiluminescence (ECL) detection system (Amersham Biosciences).

For matrix-assisted laser desorption/ionization–time-of-flight mass spectrometry (MALDI-TOF-MS), protein bands from the immunoprecipitation were cut out from the SDS-PAGE gel and washed for 30 min in 100 μL of 50 mmol/L NH₄HCO₃ followed by 100 μL of 50% CH₃CN/50 mmol/L NH₄HCO₃. After evaporating the solution, 100 μL acetonitrile was used to shrink the gel piece for 10 min. The gel piece was reswollen with 0.2 μg trypsin in 20 μL of 25 mmol/L NH₄HCO₃ for 10 min and then covered with 20 μL of 25 mmol/L NH₄HCO₃. Digestion was performed overnight at 37°C. Tryptic peptides were dissolved in 0.5% trifluoroacetic acid and then desalted using ZipTipC18 (Millipore). All mass spectra using MALDI-TOF-MS were done by the Department of Biochemistry, Peking University of Medical School, and obtained on an Ultraflex MALDI-TOF/TOF (Bruker-Franzen) in positive ion mode with an accelerating voltage of 20 kV using α-cyano-4-hydroxycinnamic acid as the matrix. The spectra were internally calibrated using trypsin autolysis products. The obtained peptide mass fingerprints (PMF) were used to search the NCBInr database. TOF-MS/MS of selected polypeptides was performed when PMFs obtained from a particular band did not provide sufficient information for protein identification.

The glutathione S-transferase (GST)–SOCS3–green fluorescent protein (GFP) fusion protein was prepared according to our previous method (8, 17). Briefly, the SOCS3 gene cloned from RAW264.7 cells was subcloned into plasmid pGEX-6P-GFP to generate the GST-SOCS3-GFP or GST-GFP fusion proteins. GST-SOCS3-GFP was bound to a GST Trap FF column, and detergent extracts of 2 × 10⁶ DCs were passed over the column. After extensive washing, the compound was eluted with 50 mmol/L Tris-HCl and 10 mmol/L reduced glutathione at pH 8.0. Eluates were analyzed by Western blot.

Phosphotyrosine-containing peptides from M2-PK were synthesized and labeled using 5-carboxyfluorescein (FAM) by Shanghai Science Peptide Biological Technology Co. Ltd. GST-SOCS3 fusion proteins were covalently immobilized onto beads of glutathione agarose (Amersham Biosciences). Then, the immobilized protein was incubated with 5-FAM–labeled phosphopeptide solubilized in 1 mL of 50 mmol/L sodium phosphate (pH 7.5) containing 150 mmol/L NaCl, 2 mmol/L DTT, and 0.2% (v/v) Tween 20. Binding of 5-FAM–labeled phosphorylated and nonphosphorylated peptides was detected using fluorescence intensity, which was measured with a DyNA Quant 200 Fluorometer (Hoefer Pharmacia Biotech, Inc.) using an excitation filter of 365 nm and an emission filter of 510 nm.

pcDNASOCS3, pcDNAM2-PK, and the mutant form of M2-PK, pcDNAM2-PKΔ105, were prepared using the pcDNATOPOV5/His kit according to the manufacturer's protocol (Invitrogen). To produce lentiviruses containing SOCS3, M2-PK, or M2-PKΔ105, the genes for these proteins were first subcloned into PLenti6/V5/D vectors from pcDNA plasmids using the PLenti6/V5/D-Topo kit according to the recommended protocol (Invitrogen). Lentiviruses were generated by cotransfecting 293T cells together with packaging vectors based on the company's instructions (Invitrogen). These packaged lentiviruses expressing SOCS3, M2-PK, or M2-PKΔ105, as well as control scrambled lentiviruses, were used to transduce DCs. Primers used to clone SOCS3 include 5′-atggtcacccacagcaagtttcc-3′ (forward) and 5′-aagtggagcatcatactgatccagg-3′ (reverse). Primers used to clone M2-PK were 5′-atgccgaagccacacagtgaagc-3′ (forward) and 5′-aggtacaggcactacacgcatgg-3′ (reverse).

For mutant M2-PKΔ105, four primers were used: 5′-atgccgaagccacacagtgaagc-3′, 5′-aggtacaggcactacacgcatgg-3′, 5′-gcatctgatcccattctcGGGgtcctgttg-3′ and 5′-gagccaccgcaacaggacgCCCgtagagaatg-3′ (capital letters, mutated 105 amino acid residues).

Small interfering RNA (siRNA) sequences used for silencing murine SOCS3 (mSOCS3) were designed using the protocol described by Ambion.⁴

M2-PK activity was measured according to the reported method (18) with a M2-PK activity detection kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) using a continuous assay coupled to lactate dehydrogenase. The change in absorbance at 340 nm caused by the oxidation of NADH was measured using a spectrophotometer (Perkin-Elmer, Inc.). M2-PK expression of cells was measured using a commercially available mouse pyruvate kinase M2-PK ELISA kit (USCN Life Science & Technology). The absorbance of each well was determined using a microplate reader at 450 nm. The M2-PK concentration was derived from the standard curve.

Total cellular RNA was prepared using the Trizol Reagent (Invitrogen) followed by RNA cleanup with the RNeasy Mini kit (Qiagen) as recommended by the manufacturer. Reverse transcription-PCR (RT-PCR) was performed using SuperScript one-step RT-PCR with Platinum Taq according to the provided protocol (Invitrogen). The following primers were used: SOCS3, 5′-ATGGTCACCCACAGCAAGTTTCC-3′ (forward) and 5′-AAGTGGAGCATCATAAACTCCC-3′ (reverse); glyceraldehyde-3-phosphate dehydrogenase (GAPDH), 5′-ATGGTGAAGGTCGGTGTGAACGGATTTGGC-3′ (forward) and 5′-CATCGAAGGTGGAAGAGTGGGAGTTGCTGT-3′ (reverse).

Cells were harvested and washed twice with cold PBS, and 1 × 10⁶ pelleted cells were disrupted on ice in lysis buffer (50 mmol/L HEPES, 150 mmol/L NaCl, 5 mmol/L EDTA, 1 mmol/L NaOV₄, 0.5% Triton, 0.1 μg/μL pepstatin, 0.1 μg/μL antipain, 0.1 μg/μL aprotinin, 2 mmol/L phenylmethylsulfonyl fluoride). Proteins were separated on a 10% denaturing SDS polyacrylamide gel and transferred to a polyvinylidene difluoride membrane (Millipore). The membrane was saturated for 1 h at room temperature in PBS/0.05% Tween 20 supplemented with different concentrations of nonfat dried milk (Sigma-Aldrich). Hybridizations with primary antibodies were carried out for 1 h at room temperature in blocking buffer. The protein-antibody complexes were detected using peroxidase-conjugated secondary antibodies (Boehringer Mannheim) and ECL (Amersham).

Cells were incubated at 37°C under hypoxia (5% CO₂, balanced N₂) in a modular incubation chamber according to a reported method (7). ATP levels in the cells were assayed using ATP Bioluminescent Assay kit (Sigma-Aldrich) according to the protocols described by the company. The light emission was measured using the Kodak Imaging System. ATP levels were determined from the standard curve.

In vivo experiments were designed to investigate effect of interaction of SOCS3 with M2-PK on DC-based immunotherapy against tumors according to the reported protocol (19). LLC cells (0.5 × 10⁶ in 100 μL of saline) were inoculated s.c. into the back of C57BL/6 mice. Tumors were allowed to develop for 9 d and mice were then randomly divided into different experimental groups (six mice per group) and then treated with intratumoral injections of the transduced DCs, respectively. The groups treated with only PBS or nontreated DCs were used as controls.

Tumor tissue samples from LLC-bearing mice were harvested and placed in the OCT compound (Tissue-Tek) and stored at −80°C. Cryostat sections (5 μm) were fixed in cold acetone for 15 min. Slides were washed with PBS and incubated for 1 h at room temperature with the appropriate dilutions of PE-labeled anti-CD4, anti-CD8, or anti-CD11c antibodies or stained with 4′,6-diamidino-2-phenylindole and then observed under a fluorescence microscope.

The Student's t test was used for the statistical analysis and a 95% confidence limit was taken to be significant and was defined as P < 0.05.

We have previously reported the binding of SOCS3 with TYK2 by immunoprecipitating the lysates of SOCS3-transfected RAW264.7 cells (8). To look for additional potential partner(s) of SOCS3, we performed another immunoprecipitation of the lysates of SOCS3-transfected 1D8 ovarian carcinoma cells. At least five bands between 44 and 66 kDa were precipitated from the lysates, whereas these bands were not observed in the lysates of control 1D8 cells transfected with an empty vector or with the isotype-matched antibody control (Fig. 1A,1; data not shown), implying that new SOCS3 partner(s) was found. Unfortunately, these bands were not clearly identified by peptide mass fingerprinting. However, when the fragmentation patterns of the tryptic peptide molecular ions were further analyzed by TOF-MS/MS, seven peptides were identified as having a protein sequence that was consistent with M2-PK (Fig. 1A,2). Combining the protein accession number from National Center for Biotechnology Information, molecular mass, MASCOT protein score, and confidence index, we finally identified a novel SOCS3 partner, M2-PK. The association between SOCS3 and M2-PK was further confirmed by Western blot analysis also (Fig. 1A 3).

To investigate the interaction of SOCS3 with M2-PK in DCs, the lysates of BMDCs were passed over glutathione-Sepharose beads that had been precoated with GST-SOCS3-GFP. The bound proteins were separated and analyzed by Western blotting using a M2-PK monoclonal antibody. As illustrated in Fig. 1B, M2-PK from the lysates of BMDCs was bound to GST-SOCS3-GFP but not to GST-GFP. Both proteins displayed a similar, mostly perinuclear staining as well as showed an overlap in subcellular distribution, as seen by the appearance of yellow dots when the staining was superimposed (Fig. 1C). Thus, these results suggest that SOCS3 may bind to M2-PK in tumor cells as well as DCs.

Because SOCS3 has a central SH2 domain that targets the protein to phosphorylate tyrosine residues, thus causing it to interfere with signaling (20), we hypothesized that an interaction between SOCS3 and M2-PK might occur via the SOCS3 binding site of M2-PK, which contains tyrosine residues (Supplementary Fig. S1). To find the SOCS3 binding sequence in M2-PK, we first designed and synthesized peptides containing phosphorylated tyrosine residues based on the murine M2-PK protein sequence (Supplementary Fig. S2). In all of the phosphorylated peptides examined, only Y105 (pYRPVAVAL), with the amino acid residue valine at the +3 position from the phosphotyrosine, exhibited a better optimized interaction with SOCS3 (Fig. 2A,1). Several other phosphorylated peptides, such as Y161 and Y466, did not show similar binding. Additionally, consistent with other reports (21), SOCS3 did not bind to the nonphosphorylated peptides (Fig. 2A,2). As the positive control, peptide pYSTVVHSG (21), a sequence containing the Y757 docking site of murine gp130, also showed a remarkable tendency to bind SOCS3 (Fig. 2A 1).

The interaction between M2-PKpY105 and SOCS3 was also confirmed by pull-down analysis. Binding of SOCS3 to M2-PK was remarkably inhibited by Y105 (pYRPVAVAL) peptides but not by control pY148 peptides (Fig. 2B). To further determine the role of the pY105 tyrosine residue, we used glycine to replace the M2-PKp105 tyrosine residue to produce mutant M2-PK (M2-PKΔ105). As shown in Fig. 2C, M2-PK, but not M2-PKΔ105, was bound by the SOCS3 fusion protein. Furthermore, the SOCS3 fusion protein remarkably inhibited the M2-PK activity of the lysates from the M2-PK–transfected RAW264.4 cells but not from M2-PKΔ105–transfected RAW264.7 cells (Fig. 2D,1). As a control, the GST-GFP fusion protein did not significantly suppress M2-PK activity (Fig. 2D 2). Thus, our results suggest that the p105 tyrosine residue of M2-PK (M2-PK pY105) may be part of the SOCS3 binding site.

To investigate the effect of intracellular SOCS3 on M2-PK activity, we first produced lentivirus-based structures to transduce monocyte/macrophage RAW264.7 cells or BMDCs. The high efficiency of SOCS3 transduction could be shown by immunofluorescence with an anti–V5 tag-specific monoclonal antibody (Supplementary Fig. S3). As expected, ectopic SOCS3 in both transduced RAW264.7 cells and BMDCs remarkably decreased M2-PK activity as compared with nontransduced and control lentivirus–transduced cells (Fig. 3A and B). Because the NH₂-terminal SH2 domain of SOCS3 plays a critical role in determining the interaction of SOCS3 with its phosphotyrosine binding site (21), we deleted the SH2 region of SOCS3 (NH₂-terminal 1–126 region, SOCS3Δ1-126) and observed the effect of SOCS3Δ1-126 on M2-PK activity. Indeed, M2-PK activity in the SOCS3Δ1-126–transduced cells did not exhibit a significant change as compared with control (Fig. 3A and B), suggesting that the SOCS3 SH2 region plays a critical role in mediating the suppression of M2-PK activity.

We next used knockdown techniques to investigate the effect of silencing SOCS3 on the activity of M2-PK in both RAW264.7 cells and BMDCs. SOCS3-targeting siRNA, but not the control siRNA, reduced the levels of SOCS3 transcript and protein without affecting GAPDH transcript or actin protein levels (Supplementary Fig. S4). M2-PK activity in SOCS3-targeting siRNA–transfected RAW264.7 cells or BMDCs was remarkably higher than that in control-transfected RAW264.7 cells or BMDCs (P < 0.05; Fig. 3A and B), further supporting a role for SOCS3 in regulating M2-PK activity. Notably, the levels of M2-PK expression were similar in the differently treated cells (Supplementary Fig. S5), indicating that SOCS3 does not affect the expression of M2-PK.

We previously reported that, in addition to IL-6 and LPS, tumor-associated factors could also upregulate SOCS3 expression in murine BMDCs (8). The isolated DCs from the spleens of mice bearing a 1D8 ovarian carcinoma (1D8), B16 melanoma, or LLC also expressed higher levels of SOCS3 than DCs from the spleens of disease-free mice (Fig. 4A). Thus, the DCs in tumor environments may also have reduced pyruvate kinase activity. As predicted, the DCs isolated from tumor tissues or spleens of mice bearing 1D8, B16, or LLC showed reduced pyruvate kinase activity as compared with the DCs isolated from the spleens of disease-free mice (Fig. 4B). Because DCs use the glycolytic pathway as the main source of energy in the absence of oxygen (22), we further hypothesized that the upregulated SOCS3 could be involved in the induction of altered DC function by interacting with M2-PK. Indeed, DCs isolated from tumor tissues lost the typical morphology seen in DCs isolated from healthy tissue (Fig. 4C). However, the levels of M2-PK expression were similar in all of the DCs (Supplementary Fig. S6), indicating that reduced M2-PK activity in the tumor-associated DCs was not caused by change of M2-PK expression.

Because M2-PK plays a critical role in ATP production in the absence of oxygen, interaction of SOCS3 with M2-PK may decrease ATP production to affect the function of DCs. SOCS3 transduction indeed declined the generation of M2-PK–mediated ATP in DCs under hypoxic conditions (Fig. 5). Importantly, the interaction between SOCS3 and M2-PK affected DC-based immunotherapy against tumor. As shown in Fig. 5, M2-PK alone–transduced DCs significantly improved DC-based immunotherapy against LLC as compared with control scrambled lentivirus–transduced or M2-PKΔ105 lentivirus–transduced DCs (P < 0.05), whereas, consistent with our previous finding (8), SOCS3-transduced DCs remarkably reduced DC-based immunotherapy against tumor (Fig. 5A and B). However, although both SOCS3-cotransduced and M2-PK–cotransduced DCs were used to immunize mice (Supplementary Fig. S7), SOCS3 remarkably impaired the role of M2-PK in DC-based immunotherapy against LLC. LLC tumor grew remarkably faster in mice vaccinated with SOCS3-cotransduced and M2-PK–cotransduced DCs than that in mice vaccinated with only M2-PK–transduced DCs (P < 0.05; Fig. 5B). The interaction of SOCS3 with M2-PK also affected the production of tumor-specific T cells. Lymph node cells isolated from mice intralesionally vaccinated with M2-PK–transduced DCs released higher levels of IFN-γ, whereas lymph node cells from mice immunized with DCs cotransduced with both SOCS3 and M2-PK only produced a low level of IFN-γ (P < 0.05; Fig. 5C). Decreased CD4⁺ and CD8⁺ T-cell infiltration in these tumors was also observed in the group of mice intralesionally vaccinated with DCs cotransduced with SOCS3 and M2-PK as compared with those vaccinated with only M2-PK–transduced DCs (Fig. 5D).

Similar to other cells, antigen-presenting cells such as DCs also use glycolysis as a metabolic pathway to generate ATP, and the use of glycolysis is considered one of the most fundamental metabolic alterations in the absence of oxygen (22). When M2-PK is inhibited by phosphorylation, phosphoenolpyruvate (PEP) is prevented from being converted to pyruvate, a process that is used to convert ADP to ATP (23). In normal cells, M2-PK is mainly high active tetrameric form, whereas M2-PK isoenzyme found in tumor cells is usually less active dimeric form of M2-PK to accumulate precursors for synthetic processes (24, 25). In this study, we found that M2-PK is a partner of SOCS3 and show that M2-PK activity can be inhibited by the direct interaction of SOCS3 with M2-PK. Importantly, we also found that DCs isolated from tumor tissues have upregulated SOCS3 expression and downregulated M2-PK activity. Moreover, the interaction of SOCS3 with M2-PK remarkably impairs DC-based immunotherapy against tumors. Thus, our results suggest a new mechanism for DCs in the tumor microenvironment by which SOCS3 may induce DC dysfunction (Fig. 6). SOCS3 expression in DCs is upregulated by tumor-derived factors. The upregulated SOCS3 may interact with M2-PK to decrease ATP production via the downregulation of M2-PK activity, which ultimately causes DC dysfunction. These dysfunctional DCs have a reduced ability to present antigens.

SOCS3 plays an important role in inducing DC dysfunction and causing immune tolerance. It is a critical regulator of cytokine signaling and the Janus-activated kinase/signal transducer and activator of transcription pathway (26–29). SOCS3 can act as a negative regulator of inflammatory responses (30–32). Recently, we (8) and others (33) found that the intracellular delivery of SOCS3 significantly reduces the production of inflammatory cytokines. DCs transduced with SOCS3 exhibit a tolerogenic/DC2 phenotype (34). Injection of SOCS3-transduced DCs significantly suppresses experimental autoimmune encephalomyelitis. SOCS3-transgenic mice exhibit enhanced Th2 differentiation (30, 31). SOCS3 proteins have been shown to modulate macrophage function and negatively regulate LPS-induced macrophage activation (35).

M2-PK, a key enzyme in the glycolysis metabolic pathway, undergoes very complex regulation. M2-PK activity may be downregulated by the direct interaction of SOCS3 with M2-PK. Other studies have shown that M2-PK activity in tumor cells is also regulated by the direct interaction of M2-PK with several oncoproteins. The transforming factor of Rous sarcoma virus, pp60v-src kinase, has tyrosine kinase–specific activity and phosphorylates M2-PK at a tyrosine residue to reduce M2-PK activity (23, 36). The serine/threonine kinase A-Raf directly binds to M2-PK to regulate the pyruvate kinase (25). Furthermore, pyruvate kinase is also regulated indirectly by insulin and glucagons, which control protein kinase activity.

No potential conflicts of interest were disclosed.

Grant Support: National Natural Science Foundation of China grants 30771967 and 30872315; Ministry of Science and Technology of the People's Republic of China grants 06C26211200695, 2008AA02Z129, and 2006AA020502; and Tianjin Municipal Science and Technology Commission grants 07JCZDJC03300 and 06ZHCXSH04800.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.