Article pubs.acs.org/jpr
Comprehensive Proteomics Analysis Reveals Metabolic Reprogramming of Tumor-Associated Macrophages Stimulated by the Tumor Microenvironment Di Liu,†,‡,# Cheng Chang,‡,# Ning Lu,§,# Xing Wang,‡ Qian Lu,‡,∥ Xiaojie Ren,§ Peng Ren,§ Dianyuan Zhao,‡ Lijing Wang,⊥ Yunping Zhu,*,‡ Fuchu He,*,†,‡ and Li Tang*,‡,∥ †
School of Life Sciences, Tsinghua University, Beijing 100084, P. R. China State Key Laboratory of Proteomics, Beijing Proteome Research Center, National Engineering Research Center for Protein Drugs, National Center for Protein Sciences (Beijing), Beijing Institute of Radiation Medicine, Beijing 102206, P. R. China § Department of Orthopedics, PLA General Hospital, Beijing 100853, P. R. China ∥ Department of Biochemistry and Molecular Biology, Anhui Medical University, Hefei, Anhui Province 230032, P. R. China ⊥ Vascular Biology Research Institute, Guangdong Pharmaceutical University, Guangzhou, Guangdong Province 510006, P. R. China ‡
S Supporting Information *
ABSTRACT: Tumor-associated macrophages (TAMs) are major components of the tumor microenvironment. Although a role for TAMs in promoting tumor progression has been revealed, the differentiation mechanisms and intrinsic signals of TAMs regulated by the tumor microenvironment remain unclear. Here we constructed an in vitro TAMs cell model, TES-TAMs, which is from tumor-extractstimulated bone-marrow-derived macrophages. We performed a comparative proteomics analysis of bone-marrow-derived macrophages and TES-TAMs, which indicated that TES-TAMs possessed characteristic molecular expression of TAMs. Intriguingly, the signal pathways enriched in up-regulated differentially expressed proteins of TAMs demonstrated that glycolysis metabolism reprogramming may play an important role in TAM differentiation. We found that hexokinase-2, a key mediator of aerobic glycolysis, and the downstream proteins PFKL and ENO1 were remarkably increased in both TES-TAMs and primary TAMs from our MMTV-PyMT mice model. This phenomenon was then verified in human THP-1 cell lines stimulated by tumor extract solution from breast cancer patient. Taken together, our study provides insight into the induction of TAM differentiation by the tumor microenvironment through metabolic reprogramming. KEYWORDS: tumor-associated macrophages, proteomics, tumor microenvironment
■
differentiation of fetal monocytes into alveolar macrophages.7 Microglial cells, the resident macrophages in the brain, perform neurotoxicity in various central nervous system inflammations dependent on interaction with neurons.8 Tumor-associated macrophages (TAMs) are major components of the tumor microenvironment.8 There is strong evidence in both experimental and clinical studies that TAMs enhance tumor progression to malignancy,8,9 including supporting angiogenesis,10−12 promotion of tumor cell invasion and metastasis,13−15 as well as suppression of antitumor immune responses.16,17 Macrophage infiltration has been identified as an independent poor prognostic factor in several cancer types. The link between TAMs and tumor progression is
INTRODUCTION Macrophages are immune cells that ingest and degrade dead cells, debris, and foreign material and orchestrate inflammatory processes.1 Tissue-resident macrophages that are important in the maintenance of homeostasis can be maintained through local proliferation or differentiation in situ by circulating monocytic precursors.2 In addition, macrophages perform many tissue-specific functions, which is reflected in their phenotypic diversity.3 For example, Kupffer cells, the resident macrophages in the liver, exhibit a tolerogenic phenotype in the steady state, but address exposure to gut-derived antigens and bacterial endotoxin.4 Red pulp macrophages (RPMs) in the spleen are specialized in iron recovery and prominently express proteins involved in all recycling phases.5 Indeed, to develop tissue-specific phenotypes, macrophages are largely regulated by local tissue-derived signals including chemokines, cytokines, and endogenous signals.4,6 For example, respiratory epithelium is the major source of CSF-2, which is necessary for the © XXXX American Chemical Society
Special Issue: The Immune System and the Proteome 2016 Received: July 1, 2016
A
DOI: 10.1021/acs.jproteome.6b00604 J. Proteome Res. XXXX, XXX, XXX−XXX
Article
Journal of Proteome Research especially well-established in breast cancer. CD68+ TAMs in patients18 or F4/80+TAMs in mouse breast cancer model19 correlate with poor outcome. Additionally, transcriptional analysis of TAMs derived from a mouse mammary tumor model indicates that an enrichment in macrophage transcripts is predictive of poor prognosis.20 Furthermore, genetic ablation of colony-stimulating factor-1 (Csf-1) or its direct effector Est-2 transcription factor results in a lower incidence of lung metastasis in both spontaneous and orthotopic transplant breast cancer models.20,21 Targeting TAMs with anti-Csf-1R antibody reveals a strategy for cancer therapy through impeding TAMs differentiation. As physiological tissue-resident macrophages, TAM phenotypes are controlled by environmental signals, including immune signals, cell death signals and metabolism signals in the tumor.22 For example, tumor-cellderived lactic acid has a critical function to induce the M2-like polarization of TAMs.23 Although a role for TAMs in promoting mammary tumor progression has been revealed, the mechanism and intrinsic signals of TAM differentiation programs regulated by the microenvironment remain to be defined. To assess the effects of the breast cancer microenvironment on macrophages, we employed a well-characterized transgenic mouse model of breast cancer, MMTV-PyMT mice, in which the tumors arise rapidly due to the expression of the polyomavirus middle-T antigen (PyMT) in mammary epithelium driven by the mouse mammary tumor virus (MMTV) promoter.24 This mouse model that develops multifocal mammary tumors with a high incidence of pulmonary metastasis is one of the most common and wellsuited models for TAM differentiation and promotion for tumor progression.25 We isolated tumor tissue from MMTVPyMT mice of advanced-stage disease to obtain a tumor extract solution to stimulate bone-marrow-derived macrophages (BMDMs) from normal mice. This approach has several advantages. First, compared with the analysis on TAMs isolated from mouse models, the study on BMDMs under different conditions has good reliability and reproducibility. Second, compared with mouse macrophage cell lines, BMDMs can better respond to external stimuli to mimic TAMs stimulated by the microenvironment. Finally, the tumor extract is closer to the true environment than conditional medium of tumor cell lines because there are many other cells contributing to TAM differentiation through secretion besides the tumor cells. In recent years, considerable progress in mass spectrometry (MS) techniques has made the precise characterization of the proteome possible. An unbiased proteomics analysis assessing the abundant change of thousands of proteins may therefore have a better chance to identify and quantify the response of macrophages to the tumor microenvironment. To date, highthroughput and quantitative proteomic analysis have not yet been applied to analyze TAM differentiation. Here we used the proteome responses of the mouse BMDMs treated with tumor extract to demonstrate the relative differential effects of TAMs stimulated by the tumor microenvironment. There were many differences in gene expression between BMDMs and TESTAMs identified by proteomics analysis. Our data indicated that TES-TAMs possessed characteristic molecular expression of TAMs and metabolic reprogramming, such as glycolysis and inositol phosphate metabolism, indicating that the tumor microenvironment played an important role in TAM polarization. Furthermore, we applied real-time PCR to evaluate whether primary TAMs isolated from MMTV-PyMT mice can
express the classical proteins identified by the proteome assay. The expression trend of these genes would be consistent with the primary TAMs from mice with advanced-stage disease.
■
METHODS
Mice
All animals were handled in strict accordance with the ‘Guide for the Care and Use of Laboratory Animals’ and the ‘Principles for the Utilization and Care of Vertebrate Animals’, and all animal work was approved by the Institutional Animal Care and Use Committee (IACUC) at the Beijing Institute of Radiation Medicine. Specific pathogen-free (SPF) C57BL/6J, MMTVPyMT transgenic mice (B6), and female nude mice were bred in our SPF facilities. For tumor extract solution and TAMs, tumors of female MMTV-PyMT transgenic mice were measured weekly by calipers for calculation of tumor volumes(0.5 × length × width2) after week 15, and tumors larger than 1.0 cm3 were defined as advanced-stage disease (AD). In TES-TAM and MDA-MB-231 coinjection experiments, the amount of Matrigel was reduced by 6-fold to minimize any effects of growth factors from the Matrigel. Thus the cells were suspended in 10 μL of a 4:1 DMEM and Matrigel mixture and injected into the fourth mammary fat pads of 6 week old female nude mice. Tumor Extract Solution
Non-necrotic fresh tumor tissues isolated from MMTV-PyMT mice in advanced-stage disease were cut into small pieces. Tumor pieces, ∼4 each, were homogenized by five strokes with 2 mL of ice-cold RPMI medium in a round-glassed homogenizer on ice and passed through a sieve. Then, the homogenized solution was centrifuged, and supernatant were filtered through a 0.45 μm nitrocellulose filter. The effluent liquid was used to stimulate mouse BMDMs. Mouse Bone-Marrow-Derived Macrophage Isolation, Culture, and Treatment
Mouse BMDMs were obtained as described.26 6 week old C57BL/6J mice were euthanized by cervical dislocation, and both femurs were dissected free of adherent tissue. The ends of the bones were cut off and the marrow tissue was eluted by irrigation with RPMI medium. Mouse bone-marrow progenitor cells were suspended by vigorous pipetting, washed once in RPMI medium, and collected by centrifugation. The cells were cultured for 6 days in M-CSF (100 ng/mL) to induce macrophage differentiation. TES-BMDMs were generated by BMDM in tumor extract solution for 3 days. The BMDMs and TES-BMDMs were incubated for 20 min with trypsin before collection. After centrifugation (5 min; 500g), the collected cells were washed with PBS and counted using a TC10 cell counter (Bio-Rad, Hercules, CA) with trypan blue staining. After centrifugation (5 min; 500g), ∼5 × 106 cells were used to protein isolate and digest. Isolation of Primary Peritoneal Macrophages
Peritoneal macrophages (pMs) were isolated from normal female C57BL/6J mice as previously described.27 In brief, pMs were isolated after lavaging the peritoneal cavity with PBS for 10 min. Peritoneal fluid was collected and centrifuged at 500g for 5 min. Cells were washed twice and resuspended in RPMI medium. It was then plated in a Petri plate and left for 2 h at 37 °C at 5% CO2. The nonadherent cells were removed by B
DOI: 10.1021/acs.jproteome.6b00604 J. Proteome Res. XXXX, XXX, XXX−XXX
Article
Journal of Proteome Research
MS analyses was performed on an Easy-nLC 1000 (Proxeon, Odense, Denmark) system coupled to an Orbitrap Fusion via a nanoelectrospray ion source (Thermo Fisher Scientific). Fractions from the 1D RPLC were dissolved with loading buffer; 10% of total tryptic peptides (∼500 ng) were eluted from a 360 μm ID × 2 cm, C18 trap column and separated on a homemade 100 μm ID × 10 cm column (C18, 1.9 μm, 120 Å, Dr. Maisch) with a series of adjusted linear gradients according to the hydrophobicity of fractions with a flow rate of 500 nL/ min. Survey scan were acquired after accumulation of 5 × 105 ions in Orbitrap for m/z 300−1400 using a resolution of 120 000 at m/z 200. The top-speed data-dependent mode was selected for fragmentation in the HCD cell at normalized collision energy of 32%; then, fragment ions were transferred into the ion trap analyzer with the AGC target at 5 × 103 and maximum injection time at 35 ms. The dynamic exclusion of previously acquired precursor ions was enabled at 18 s.
washing with RPMI medium. The adherent macrophages were collected for RNA extraction. Flow Sorting of Primary Tumor-Associated Macrophages
Single-cell suspensions of mammary glands or tumors were prepared as previously described.28 In brief, tumor tissues were dissected, minced into small pieces, and digested for 1 h at 37 °C in RPMI medium containing 2% FBS supplemented with 300 U/mL type IV collagenase (Sigma, St. Louis, MO) and 100 U/mL DNase I (Sigma). For sorting TAMs, single-cell suspensions from breast tumors were stained with anti-CD45 FITC, anti-CD11b PECy7, anti-MHC II APC-Cy7 in PBS containing 0.2% BSA. Before staining, the cells were preincubated with Fc inhibitor (eBioscience, San Diego, CA). Quantitative Real-Time PCR
Gene expression levels were calculated as a ratio of the mRNA level relative to the RNA level for 18S rRNA in the same cDNA. For the regulation of gene expression of BMDM, TES-BMDM, TAM, or THP-1, total RNA prepared by RNeasy Micro Kit (Qiagen, Hilden, GER) was reverse-transcribed using the Reverse Transcription System (Promega, Madison, WI,) according to the protocol provided by the manufacturer and then amplified by PCR. SYBR Green real-time PCR (Toyobo, Japan) was done using Bio-Rad iQ5 cycler. Real-time PCR data analysis was performed using Bio-Rad software. The mRNA levels of mouse Vcam1, Ym1, Il4rα, Vegfa, Il27, Il1B, Pfkl, Hk2, Eno1, and 18scrRNA were determined by qPCR using the following primers: Vcam1, 5′-tgaacccaaacagaggcagagt-3′ and 5′ggtatcccatcacttgagcagg-3′; Ym1, 5′-ccagcatatgggcatacctt-3′ and 5′-cagacctcagtggctccttc-3′; Il4rα, 5′-cctctgtgggctgtctgatttt-3′ and 5′-gggctcacccaggacctt-3′; Vegfa, 5′-atgccaagtgggaaaatctg-3′ and 5′-tgtagcagtggcctgcatag-3′; Il27, 5′-ggccaggtgacaggagacc-3′ and 5′-cagcttgtaccagaagcaaggg-3′; Il1B, 5′-gtgttgctgaaggagttgcc3′ and 5′-ctggataagcagctgatgtg-3′; Pf kl, 5′-tgcagcctacaatctgctcc3′ and 5′-gtcaagtgtgcgtagttctga-3′; Hk2, 5′-caacatcctgatcgatttcacaa-3′ and 5′-gcagtcactctcgatctgagaca-3′; Eno1, 5′-tgcgtccactggcatctac-3′ and 5′-cagagcaggcgcaatagtttta-3′; 18S rRNA, 5′gtaacccgttgaaccccatt-3′ and 5′-ccatccaatcggtagtagcg-3′.
Protein Identification and Quantification
In this study, all Thermo raw data were directly processed using MaxQuant software (version 1.5.4.1).29 For identification, the acquired MS/MS spectra were searched by Andromeda30 against the Swiss-Prot mouse database (release 2014_11) with 245 commonly identified contaminants. Target-decoy strategy was applied to keep both protein and peptide FDRs below 0.01.31 The detailed search parameters were: a precursor ion mass tolerance of 20 ppm and a product ion mass tolerance of 0.5 Da. Tryptic cleavage was selected and two missed cleavage sites were allowed at most. Carbamidomethylation of cysteine was set as a fixed modification, and oxidation of methionine and acetylation of protein N-terminal were set as variable modifications. For quantification, the routine label-free quantification parameters including a minimum ratio count of 2 and “Match between runs” in MaxQuant were set for accurate analysis. The protein intensities in two biological replicates were median normalized at first to reduce the biases between experiments and then merged averagely as the final protein abundances following our previous procedure.31 All of the raw data can be freely accessed from iProX (http://www.iprox.org) with the identifier IPX00078200.
Protein Isolation and Digestion
Bioinformatics and Statistical Analysis
The collected BMDMs and TES-TAMs were sonic disrupted with 8 M urea. Cell lysates were centrifuged, and supernatants were collected. Protein was reduced by adding 5 mM dithiothreitol for 0.5 h at 56 °C and alkylated by adding 20 mM iodoacetamide for 0.5 h at room temperature in the dark; then, 5 mM dithiothreitol was added and kept in dark for 15 min. The 50 μg protein sample was finally transferred into 10 kDa filter unit, washed with 8 M urea and 50 mM ammonium bicarbonate twice, digested using trypsin at a mass ratio of 1:50 enzyme/protein overnight at 37 °C, and stopped by the addition of 1% formic acid (FA).
In this study, differentially expressed proteins (DEPs) were determined using the significant B algorithm in MaxQuant32 with the protein bin size of 500. Proteins with a p value less than 0.05 were considered as DEPs. All of the pathway and GO enrichment analyses including biological process, cellular component, and molecular function were performed using DAVID33 (version 6.7). GraphPad Prism 5.0 software was used for qPCR data analysis. Statistical significance was determined by t tests (and nonparametric tests); p < 0.05 was considered statistically significant.
■
Two-Dimensional Reverse-Phase Liquid Chromatography (RPLC) and MS Analysis
RESULTS
Assay Workflow
The first dimension RP separation by using a homemade Durashell RP column with 2 mg packing (3 μm, 150 Å, Agela) in a 200 μL tip. Mobile phases A (2% acetonitrile, adjusted pH to 10.0 using NH3H2O) and B (98% acetonitrile, adjusted pH to 10.0 using NH3H2O) were used to develop a elution buffer. The tryptic peptides were sequential separated with elution buffer at the gradient of 6, 9, 12, 15, 18, 21, 24, 30, and 35% B and then were combined into six fractions and dried under vacuum. The 2D low-pH RP chromatography coupled to MS/
The aim of the prototype assay was to compare the proteomes of BMDMs treated with M-CSF and tumor extract solution and the “reference” proteomes derived from BMDMs treated only with M-CSF. The work flow of the assay consists of six major steps, as depicted in Figure 1. We first isolated bone marrow cells from the C57BL/6J mice and induced with M-CSF for BMDMs. As a rule, nonadherent cells were discarded after 6 days of stimulation by M-CSF. With new medium with or C
DOI: 10.1021/acs.jproteome.6b00604 J. Proteome Res. XXXX, XXX, XXX−XXX
Article
Journal of Proteome Research
Figure 1. Proteomics-based TAM assay workflow.
without tumor extract solution, adherent cells were concluded to have started their differentiation into TAM-like macrophages or normal macrophages. We collected two types of macrophages by trypsin digestion and centrifugation. Then, we employed label-free proteomics approach to profile the proteomes of these cells and made comprehensive bioinformatics analysis. We finally isolated resident pMs from normal mice and primary TAMs from MMTV-PyMT mice and confirmed the proteomic data of DEPs in TES-TAMs.
Figure 2. Construction of the TAM model and characterization of TES-TAMs. (A) Morphology of BMDMs and TES-TAMs. 200× magnification. (B) Expression of primary TAM indicators in BMDMs and TES-TAMs by real-time PCR. (C) Functional verification of TESTAMs. MDA-MB-231 cells were orthotopically injected alone (left) or with admixed TES-TAMs (right) at a limiting dilution into Nude mice. Each group: n = 4. Significance: t test; *p < 0.05;**p < 0.01.
proteins quantified in total under the protein FDR 1%. As shown in Figure 3A, the overlap percentage between BMDM and TES-TAM is as high as 97.5%. Furthermore, there were 6024 (91.9%) proteins quantified with more than one unique peptide (Figure S1). The median value of protein sequence coverage was 29.8% (Figure S2). There were 1410 (21.5%) proteins with sequence coverage larger than 50%. As shown in Figure 3B and Table S1, 600 up-regulated and 479 downregulated DEPs were determined with p value
