Background
Hepatocellular carcinoma (HCC) ranks as the sixth most prevalent cancer and the third leading cause of tumor-related death [
1,
2]. Despite huge developments in HCC diagnosis and treatment, 5-year survival rates of most HCC patients remain dismal [
3]. Therefore, it is urgently needed to discover efficient markers and therapeutic targets for HCC. Emerging evidence has proven that abnormal glycosylation in HCC is closely associated with cancer progression. Some glycosylation modifications have been reported to promote HCC tumor growth, angiogenesis, and metastasis, and some of them are capable of predicting the prognosis of HCC patients [
4]. However, due to instability of this glycosylation and the specialized equipment required for detection, it is still hard to use these as rapid markers for clinical application. Abnormal glycosylation usually results from aberrant expression of glycosyltransferases that are relatively stable in tissues and easy to detect [
4,
5]. Therefore, some of them like
N-acetylglucosaminyltransferase V (GnT-V),
N-acetylglucosaminyltransferase III (GnT-III), and α1-6 fucosyltransferase (a1-6FT) have been used as tumor markers and therapeutic targets in HCC [
6‐
9].
Macrophages originate from immature monocyte in the bone marrow and migrate throughout the entire body through the circulation. Final differentiation occurs in tissues to form macrophages, including Kupffer cells which are found in the liver. It is already known that all kinds of macrophages coexist in tumors, but recruited macrophages may account for the majority of tumor-associated macrophages (TAMs). Peripheral blood monocytes from the bone marrow are recruited and differentiate into TAMs in response to chemokines and growth factors in the tumor microenvironment [
10,
11]. TAMs promote solid tumor development through providing factors which can establish a pre-malignant niche and enhance metastasis. TAMs may also play a role in forming pre-metastatic niches in organs where the tumor will eventually metastasize [
12]. There is also evidence that TAMs are closely associated with formation of stem-like cells in human cancers [
13]. Dysregulation of glycosyltransferases has been reported to regulate the functions of TAMs [
14], but whether some glycosyltransferases influence TAM recruitment remains to be elucidated.
β-1,3-
N-acetylgalactosaminyltransferase II (B3GALNT2) belongs to the β-1,3-glycosyltransferases (b3GT) family, consisting of β-1,3-galactosyltransferases (B3GALT), β-1,3-
N-acetylglucosaminyltransferases (B3GNT), and β-1,3-
N-acetylgalactosaminyltransferases (B3GALNT). B3GALNT2 efficiently adds
N-acetylgalactosamine (GalNAc) on both N-glycans and O-glycans by β-1,3-linkage and generates GalNAcb1 → 3GlcNAcb1-R structure [
15]. B3GALNT2 has been reported to regulate the development of both normal tissues and tumor tissues [
16]. Upregulation of B3GALNT2 in breast cancer predicts poor prognosis [
17]. Since knockdown of B3GALNT2 in zebrafish leads to degeneration of the extracellular matrix [
16], B3GALNT2 might also exert functions in cancer progression via altering secretion or remodeling the extracellular environment. However, studies on the functions of B3GALNT2 are quite limited. Whether and how B3GALNT2 functions in HCC remain to be elucidated.
In this study, we investigated expression of B3GALNT2 in HCC and analyzed its potential role in HCC progression. Our study also determined how B3GALNT2 remodels the tumor microenvironments to promote tumor growth.
Methods
Hepatocellular carcinoma patient samples
Usage of human pathological tissues and clinical data was approved by the Ethics Committee at the Shanghai Cancer Center of Fudan University (Shanghai, China; approval no. 050432-4-1212B). Written consent for all patients conformed to the ethical guidelines of the Helsinki Declaration. A total of 139 patients with primary HCC resected between 2010 and 2012 in the Department of Hepatic Surgery, Shanghai Cancer Center of Fudan University (Shanghai, China) were collected. None of the patients had received pre-operative therapy. Clinical tumor stages were determined according to the TNM classification system of International Union against Cancer. Follow-up was done until December 9, 2016. These patients were followed every 3 months. The median follow-up was 33.3 months (ranging from 0.8 to 60.4 months). Among all of the primary tumor specimens, 24 were used for Western blot and quantitative real-time PCR assays.
Cell culture
All HCC cell lines, human THP1 cells, and mouse RAW264.7 cells were obtained from the Cell Bank of Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China) and cultured in Dulbecco’s Minimum Essential Medium (DMEM) supplemented with 10% fetal bovine serum at 37 °C in a humidified atmosphere containing 5% CO2. Fetal bovine serum and DMEM culture media were purchased from Sigma (St. Louis, MO, USA). The THP1 cell line was maintained in RPMI 1640 medium supplemented with 10% FBS and 2 mmol/L L-glutamine. THP1 cells were differentiated using 200 nM phorbol-12-myristate-13-acetate (PMA, Sigma-Aldrich) for 3 days.
Plasmid construction
The cDNA encoding B3GALNT2 was obtained by PCR and was inserted into the pCMV-Flag vector (Sigma, St. Louis, MO, USA). The sequence of the shRNA inserted in the pENTR vector (Thermo, USA) was as follows: B3GALNT2: 5′-CACCGGTCATATAATTGTGTGTTTACGAATAAACACACAATTATATGACC-3′, BDH2: 5′-CACCGGAACAGTTGATACGCCATCTCGAAAGATGGCGTATCAACTGTTCC-3′, and MIF: 5′-TCGAGGACACCAACGTGCCCCGCGCTTCAAGAGAGCGCGGGGCACGTTGGTGTCTTTTTTA-3′. Transfections were performed with Lipofectamine 3000 (Life Technologies, CA, USA), according to the manufacturer’s instructions. Stable cell lines were generated with G418 (200 μg/mL) in the medium.
Cell viability assay
Cell viability was quantified with a Cell Counting Kit-8 (CCK-8) (Dojindo, Japan), according to the manufacturer’s instructions. The cells were plated at a density of 3000 cells per well in 96-well plates. The CCK-8 assays were assessed by measuring the absorbance at 450 nm.
Cell cycle and apoptosis assay
Cycle arrest and apoptotic cells were detected by flow cytometric analysis. Cells were collected by trypsinization and washed twice with PBS. For cell cycle assay, the collected cells were stained with propidium iodide (PI) using a Cell Cycle Staining Kit (Lianke Bio, Hangzhou, China). Cellular apoptosis was determined using PE Annexin-V Apoptosis Detection Kit I (BD Biosciences, CA, USA). Stained cells were assessed by flow cytometry and the data were analyzed by FlowJo software (TreeStar, Ashland, OR, USA).
Transwell assay
Transwell invasion was assessed using 8-μm transwell filters (Milliporem, Billerica, MA, USA) in a 12-well plate. The bottom of the transwell chamber was coated with BD Matrigel Basement Membrane Matrix (BD Biosciences, San Diego, CA, USA). Macrophages were added into the upper chamber containing basic culture medium without serum, and the lower chamber was filled with HCC tumor cell lines in serum-free culture medium. Macrophage infiltration was determined 48 h later. Cells on the upper side of the chamber were removed from the surface of the membrane by scrubbing, and cells on the lower surface of the membrane were fixed with 4% paraformaldehyde and stained with 0.1% crystal violet. The number of infiltrating cells was counted in five randomly selected microscopic fields of each filter.
Western blotting
HCC tissues and cells were homogenized in SDS sample buffer (10% glycerol, 2% SDS, 0.01% bromophenol blue, 1.25% 2-beta-mercaptoethanol, 25 mM Tris–HCl, pH 6.8) with ULTRA-TURRAX (IKA, Germany) at 4 °C. Protein concentration was determined using the Quick Start™ Bradford protein assay kit (Bio-Rad, USA). Ten micrograms of total protein extracts was loaded in 10% SDS-PAGE and transferred to 0.45-μm PVDF membranes (Millipore, USA) using an electro-blotting apparatus (Bio-Rad, USA). Anti-B3GALNT2, anti-BDH2, and anti-GAPDH antibodies were purchased from Proteintech. The Immobilon™ Western Chemiluminescence HRP substrate kit (Millipore, USA) was used for chemiluminescence. Images were obtained with the ImageQuant™ LAS-4000 (Amersham Biosciences, GE, USA) and quantified using the ImageQuant™ TL software (version 7.0, Amersham Biosciences, GE, USA).
Quantitative real-time PCR
Total RNA of the samples was purified using TRIzol (Invitrogen, Carbad, CA, USA) and then reverse-transcribed to cDNA using the PrimeScript RT reagent kit (Takara, Tokyo, Japan). Real-time PCR was performed with cDNA production using SYBR Premix Ex Taq (Takara, Tokyo, Japan) on an ABI StepOne Plus (Applied Biosystems, USA) instrument. GAPDH was used as an internal control. Primers used in this study are listed as follows: B3GALNT2 forward: GATGTGGTAGTTGGCGTGTTG, reverse: CGTTGACTTAATGTGGGATGCTG, GAPDH forward: GAGTCAACGGATTTGGTCGT, reverse: TTGATTTTGGAGGGATCTCG, BDH2 forward: GCTTCCA GCGTCAAAGGAGTT, reverse: CAGTTGCGAATCTTCCCGTC, MIF forward: TACACCCAGACCAAATGATG, reverse: TTCTCCTAATGCTCCAATACTG.
Immunohistochemistry
Immunohistochemistry (IHC) tests on tissue microarray and paraffin sections were performed using a Dako REAL EnVision Detection System (Dako, Denmark) following the protocol recommended, and hematoxylin was used for counterstaining. Anti-B3GALNT2, anti-CD206, anti-F4/80, and anti-CD68 antibody were used to quantify relative expression levels. Immunohistochemical scoring was determined as previously described [
18]. The staining intensity was scored as 0 for negative, 1 for weak, 2 for moderate weak, 3 for moderate strong, and 4 for strong. The score for the stained area was set as 0 for 0–33%; 1, 33–66%; and 2, 66–100%. The final staining score was obtained by multiplying the staining intensity score by the staining area score, and the results are a series of numbers ranging from 0 to 8.
ELISA
MIF, CSF1, CCL2, VEGF, and MIP-1α in the culture supernatants of HCC cell lines were measured using ELISA kits (R&D Systems). The culture supernatants of the cells were collected and centrifuged at 500×g for 5 min to remove cellular debris. The ELISA was performed according to the manufacturer’s instructions.
Protein and small molecule components were separated using a 3-kDa ultrafiltration tube purchased from Millipore. Fraction containing small molecules (< 3 kDa) was collected from the residual liquid at the first ultrafiltration. The concentrated fraction (> 3 kDa) was further washed three times using PBS and finally concentrated in PBS.
MIF tautomerase activity assay
The tautomerase activity of MIF in the medium was detected as previously described with minor modifications [
19]. Phenylpyruvate could be conversed from enol- to keto- type by MIF, and this reaction was monitored by the decrease of absorbance at 288 nm on a spectrophotometer at room temperature. The assay mixture contained 50 mM sodium-phosphate buffer (pH 6.5) and a series of diluted MIF-containing culture medium. The assay was initiated by addition of ethanol-diluted phenylpyruvic acid with final concentration of 100 μm. Absorbance values for each group were normalized with the control group that contained buffer and fresh DMEM medium, thus yielding the relative percent of enzyme activity. For acetoacetate addition, 10 μM lithium acetoacetate was used. All the chemicals were purchased from Sigma (St. Louis, MO, USA).
For MIF inhibition,
N-acetyl-p-benzoquinone (NAPQI) was purchased from Sigma (St. Louis, MO, USA), and it could inhibit 96% of MIF by incubating with cells for 5 min at 200 μM as previously reported [
20].
Animal models
All animal experiments were approved by the research medical ethics committee of Fudan University (Shanghai, China; approval no. 170013-0056) and were performed in accordance with the approved guidelines. Nude mice were purchased from the Shanghai Laboratory Animal Center of Chinese Academy Sciences (Shanghai, China) and were housed in individual ventilated cages. All of the mice were randomly grouped (n = 6 in each group).
For the subcutaneous xenograft model, Huh7-luc cells were resuspended in PBS (5 × 106/mouse) and subcutaneously inoculated into the axillaries of 4-week-old male nude mice. The mice were sacrificed after 4 weeks, and tumor tissues were harvested and weighted. For the orthotopic translation model, tumors from xenograft models were separated and chopped in PBS at 4 °C. The diameter of each fragment was modified to 1 mm. The fragments were then transplanted into nude mice in the left lobes of the liver. Bioluminescent imaging was performed with an IVIS200 (Xenogen, Caliper, CA) 10 min after intraperitoneal injection of luciferin (3 mg/mouse) (Promega, WI, USA). The intensity of luciferase signals was quantified using ROI analysis.
For metabolomics analysis, 7402-B3GALNT2 and 7402-control cells were cultured to 80% confluence and the medium was removed, followed by washing the cells with ice-cold PBS. The cells were then collected into tubes with PBS by scraping. For targeted metabolomics analysis, the culture medium was changed to serum-free medium 2 h before metabolite collection. Metabolite fractions of the culture media were collected and analyzed by targeted LC-MS/MS. LC-MS analysis was performed as described previously [
21].
An ACQUITY UHPLC System (Waters Corporation, Milford, USA) coupled with an AB SCIEX Triple TOF 5600 System (AB SCIEX, Framingham, MA) was used to analyze the metabolic profiling. An ACQUITY UPLC BEH C18 column (1.7 μm, 2.1 × 100 mm) was employed with a binary gradient method. Data acquisition was performed in full scan mode (m/z ranges from 70 to 1000) combined with information-dependent acquisition (IDA) mode. For IDA analysis, range of m/z was set as 50–1000 and the collision energy was 30 eV. The QCs were injected at regular intervals (every eight samples) throughout the analytical run to provide a set of data from which repeatability could be assessed.
The raw data were converted to common data format (mzML) files using a conversion software program MSconventer. Metabolomics data were acquired using software XCMS 1.50.1 version, which produced a matrix of features with the associated retention time, accurate mass, and chromatographic separation. The positive and negative data were combined to get a combined data set which was imported into the SIMCA software package (version 14.0, Umetrics, Umea, Sweden). Principal component analysis (PCA) was carried out to visualize metabolic alterations among experimental groups, after mean centering and unit variance scaling. The differential metabolites were selected on the basis of p values from a two-tailed Student’s t test on the normalized peak areas, where metabolites with p values less than 0.05 were included. A reference material database built by the Dalian Institute of Chemical Physics, Chinese Academy of Sciences and Dalian ChemData Solution Information Technology Co., Ltd., HMDB, and METLIN was used.
Statistical analysis
All analyses were performed with SPSS 13.0 (Chicago, IL, USA). Results were presented as the mean ± standard deviation with at least three replicates for each sample. Optimal cut-off values for B3GALNT2 expression were determined by ROC curve analysis. Pearson’s chi-square test was used to identify the correlation between B3GALNT2 expression and other factors. Survival probability was determined by the Kalan-Meier curve, and the differences between groups were assessed by Log-rank test. Univariate and multivariate survival analyses were applied using Cox regression. Differences between groups were determined with Student’s t test. Statistical significance was set at two tails p < 0.05.
Discussion
In this study, our data reveal that B3GALNT2 is upregulated in HCC, and this upregulation is associated with tumor growth and poor prognosis. Mechanistically, B3GALNT2 reduced the expression of some metabolic enzymes and thus downregulated the secretion of acetoacetate levels, which relieved the activity of MIF and enhanced macrophage recruitment. Finally, recruited macrophages promoted tumor growth.
HCC is closely related with inflammation. A chronic inflammatory state is required for initiation, and the development of HCC and tumor cells also promote the construction and assist with the maintenance of the inflammatory extracellular environment [
24,
33]. Consistent with this, our results prove that HCC cells and their inflammatory microenvironment are mutually beneficial for each other. HCC cells recruit macrophages to maintain the inflammatory environment, and recruited macrophages promote tumor growth in return. Notably, continuous recruitment of inflammatory cells is commonly observed in inflammatory environment. Among these cells, tumor-associated macrophages (TAMs) occupy a major fraction, especially in HCC [
24,
25]. Liver-resident Kupffer cells and TAM in HCC are also polarized from peripheral blood mononuclear cell (PMBC), for which monocyte recruitment is indispensable [
34].
Monocyte recruitment depends on some cytokines, such as chemokine (C–C motif) ligand 2 (CCL2), macrophage colony-stimulating factor (M-CSF), macrophage inflammatory protein 1a (MIP-1a), vascular endothelial growth factor (VEGF), CCL4, CCL5, CCL8, angiopoietin-2, and MIF [
26,
27]. Usually, tumor cells elevate the secretion of cytokines to promote TAM recruitment. In our study, instead of increasing cytokine levels directly, B3GALNT2 elevated the activity of one cytokine, MIF. Although MIF was first identified as an inhibitor of macrophage migration [
35], later studies revealed that it has pleiotropic effects on cell migration and chemotaxis [
36,
37]. Actually, MIF can induce macrophage recruitment through CCL2 and its receptor CCR2 [
30]. Sometimes, MIF acts as a counter-regulation factor against anti-inflammatory and immunosuppressive machinery by overriding the glucocorticoid inhibition to immune response like T cell proliferation and cytokine production [
38,
39]. MIF has also been identified as a key cytokine for TAM polarization in melanoma-bearing mice. MIF deficiency or treatment with a MIF antagonist attenuated tumor-induced TAM polarization and reduced expression of angiogenesis-related genes in TAMs [
40].
Different from most cytokines, MIF exhibits perplexing tautomerase activity which is conversion of d-dopachrome and phenylpyruvate [
41], but its natural substrate is still not clear. Early studies considered that the tautomerase active site is vestigial with no true physiological function [
42]. But some researchers revealed that MIF interplays with CD74 as a cytokine and initiates signaling [
43,
44]. This interplay could be disrupted when the tautomerase pocket of MIF is mutated or occupied by some molecules [
31,
32]. NAPQI that we used in this study is one of the inhibitors targeting the tautomerase pocket of MIF. Our data verify that suppressing the tautomerase activity of MIF attenuates its promotion of macrophage recruitment.
Some natural small molecules, like ketone bodies, also show the ability to inhibit tautomerase activity of MIF [
19]. As one of the ketone bodies, acetoacetate inhibited MIF activity in our study and we proved that acetoacetate was the key molecule by which B3GALNT2 regulated MIF activity and macrophage recruitment. Although acetoacetate was reported to promote tumor growth in melanoma [
45,
46], ketone bodies including acetoacetate inhibit tumor progression in other cancers [
47‐
50]. It is reported that cancer cell lines grown in glucose plus acetoacetate medium show tightly coupled reduction of growth and ATP concentration [
50]. And an acetoacetate-related ketogenic diet decreases tumor cell viability and prolongs survival of mice with metastatic cancer [
48]. Meanwhile, since ketone bodies could serve as energy sources in tumor cells, the consumption of these ketone bodies might result in their decrease in extracellular microenvironment. In liver cancers, the consumption of acetoacetate was elevated for lipogenesis to compensate the energy from truncated TCA cycle [
51,
52]. Our data suggested that this persistent consumption decreases the extracellular acetoacetate levels and preserves an acetoacetate-low microenvironment benefit for tumor development. Our data also indicated that besides the direct inhibition on cell proliferation, acetoacetate could suppress tumor growth via inhibiting TAM recruitment.
The chronic inflammatory microenvironment is required for both tumor initiation and tumor progression, and increased cytokine secretion in the tumor microenvironment promotes recruitment of immune cells including TAMs [
24,
25]. TAMs play important roles in different cancers including HCC [
24,
33]. Targeting TAMs is becoming a promising strategy in treating tumors. Some studies suggest that suppressing TAM recruitment via targeting chemokines could also inhibit tumor progression. However, blocking chemokines directly with antibodies sometimes leads to unpredictable tumor growth and distant metastasis. Here, our study provides a new approach to treat HCC by increasing the small molecule acetoacetate. Due to its endogenous derivation, this strategy might be safer and have more efficiency.