Tumor suppressive role of the antimicrobial lectin REG3A targeting the O -GlcNAc glycosylation pathway

Abstract

Background and aims: Antimicrobial proteins of the regenerating family member 3 alpha (REG3A) family provide a first line of protection against infections and transformed cells. Their expression is inducible by inflammation, which makes their role in cancer biology less clear since an immune-inflammatory context may preexist or coexist with cancer, as occurs in HCC. The aim of this study is to clarify the role of REG3A in liver carcinogenesis and to determine whether its carbohydrate-binding functions are involved.

Approach and results: This study provides evidence for a suppressive role of REG3A in HCC by reducing O -GlcNAcylation in 2 mouse models of HCC, in vitro cell studies, and clinical samples. REG3A expression in hepatocytes significantly reduced global O -GlcNAcylation and O -GlcNAcylation of c-MYC in preneoplastic and tumor livers and markedly inhibited HCC development in REG3A-c-MYC double transgenic mice and mice exposed to diethylnitrosamine. REG3A modified O -GlcNAcylation without altering the expression or activity of O-linked N-acetylglucosaminyltransferase, O-linked N-acetylglucosaminyl hydrolase, or glutamine fructose-6-phosphate amidotransferase. Reduced O -GlcNAcylation was consistent with decreased levels of UDP-GlcNAc in precancerous and cancerous livers. This effect was linked to the ability of REG3A to bind glucose and glucose-6 phosphate, suggested by a REG3A mutant unable to bind glucose and glucose-6 phosphate and alter O -GlcNAcylation. Importantly, patients with cirrhosis with high hepatic REG3A expression had lower levels of O -GlcNAcylation and longer cancer-free survival than REG3A-negative cirrhotic livers.

Conclusions: REG3A helps fight liver cancer by reducing O -GlcNAcylation. This study suggests a new paradigm for the regulation of O -GlcNAc signaling in cancer-related pathways through interactions with the carbohydrate-binding function of REG3A.

Graphical abstract

FIGURE 1

Cancer-free survival and overall survival of patients with cirrhosis or HCC, respectively, according to hepatic REG3A expression. (A) Kaplan-Meier curve of tumor-free survival (GSE15654; n = 216 patients with hepatitis C–related early-stage cirrhosis). Patients with high REG3A-expressing cirrhosis have longer tumor-free survival than those with low REG3A-expressing cirrhosis. (B) Overall survival of patients with HCC according to the level of REG3A expression in the tumor (TCGA; n = 286). (C, D) Overall survival of patients with HCC (GSE14520) according to the level of REG3A expression in (C) nontumor area adjacent to the tumor (HCC-NT; n = 209) and (D) in the tumor area (HCC-T; n = 221). No difference in overall survival was found in patients with HCC, whether or not the tumor area expressed REG3A. Abbreviations: NT, nontumor; REG3A, regenerating family member 3 alpha; T, tumor.

FIGURE 2

REG3A strongly reduced cancer development in 2 mouse models of HCC. (A, B) WT and transgenic mice overexpressing REG3A in hepatocytes (REG3A-TG) received a single injection of the chemotoxic agent DEN at 14 days of age, and then tumor growth was monitored in independent groups of mice monthly for 8 months. (A) Proportion of mice with tumors over time. n = 9–16 mice for WT group and n = 9–14 for REG3A-TG group. (B) Left: Representative macroscopic views. Scale bar: 10 mm. Right: Circle pie charts for counting mice with healthy (dark green) or tumor livers (color-coded by nodule size, measured by the sum of the largest diameters of the nodule). WT livers show multiple nodules of (sub)centimeter size from 6 months of age. Transgenic livers are macroscopically normal over a longer time period or contain much smaller and fewer nodules than WT livers. The value shown in each portion of the pie chart represents the percentage of mice with a given nodule size. Arrows: the smallest tumor nodules detected at 6 months. (C–F) Genetic model for HCC in single transgenic mice for MYC (MYC-TG) and double transgenic mice for MYC and REG3A (MYC/REG3A-TG). (C) Liver weight to body weight (LW/BW). n = 12–20 independent mice per time and per group of mice. (D) Percentage of mice with HCC tumors over time (n = 40). (E) Left: Representative macroscopic views. Scale bar: 10 mm. Dotted lines: delineated tumor mass when possible. Right: Proportion of mice with healthy (dark green) or tumor (other colors of the code) liver. (F) Kaplan-Meier curve of overall survival of mice. Data are means ± SEM. The 1-tailed Fisher exact test was performed for analysis except for (C) (Student t test). *p < 0.05, **p < 0.01. NS or no statistical significance, no significance. Abbreviations: DEN, diethylnitrosamine; REG3A, regenerating family member 3 alpha; WT, wild-type.

FIGURE 3

REG3A is associated with a decrease in O-GlcNAcylation of MYC in mouse liver carcinoma. (A) qRT-PCR analysis of MYC transcript levels in liver samples at the indicated time points and pathological conditions. MYC-TG, transgenic mice homozygous for MYC; MYC/REG3A-TG, transgenic mice double homozygous for MYC, and REG3A. Healthy, normal liver sample from 3-month-old mice; preK, estimated precancerous liver sample from 6-month-old mice, livers in which the absence of tumor nodules was assessed macroscopically and microscopically; nontumor areas (NT); tumor areas (T); n = 6–16 per group. (B) Immunoblots for MYC and REG3A proteins and quantification of MYC from total protein extracts in REG3A-negative and positive tumors. Each dot represents a tumor sample from an independent mouse. (C) Immunoblots for the indicated proteins after nucleo-cytoplasmic fractionation and quantification of nuclear (Nuc) to cytosolic (Cyt) MYC ratio. (D) Quantification of total MYC protein levels and pT58 to total MYC ratio. (E) Immunoblots of phospho-GSK3β (Ser9), total GSK3β, phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204), and total Erk1/2, and related densitometry. Each dot represents a sample from an independent mouse. (F) Succinylated WGA lectin pull down to determine the level of O-GlcNAcylated MYC in MYC tumors expressing or not REG3A. n = 3 independent experiments. The arrow indicates the REG3A signal. Asterix: nonspecific signal. (G) Immunoblots for O-GlcNAc and MYC proteins following immunoprecipitation of MYC from liver extracts of MYC and MYC/REG3A transgenic mice. The arrow indicates the MYC signal. n = 3 independent experiments. Data are averages ± SEM. The Mann-Whitney U test was performed for analysis except for (E) (paired Wilcoxon test with correction for multiple testing). NS or no statistical indication, no significance. Abbreviations: qRT-PCR, quantitative real-time PCR; REG3A, regenerating family member 3 alpha; WGA, wheat germ agglutinin.

FIGURE 4

REG3A alters the MYC-dependent transcriptional program in mouse HCC. (A–F) Differentially expressed genes identified by microarray analysis of mouse liver specimens expressing MYC (MYC-TG) or MYC and REG3A (MYC/REG3A-TG). n = 5 mice per group. Enrichment plots of genes related to the S3 subclass (A) or S1 subclass (B) of the molecular classification of Hoshida et al in murine MYC tumors expressing REG3A or not, respectively. (C, D) Pathway enrichment analysis showing the different molecular pathways from enrichment analysis of differentially expressed genes in nontumor area (C) and tumor area (D) of MYC-REG3A mouse livers compared with MYC mouse livers. (E) Enrichment plots of the 100 most expressed genes in human HCCs that express endogenous REG3A (TCGA data set). The top 100 signatures is enriched in HCCs from MYC/REG3A-TG mice compared to MYC-TG mice. (F) Enrichment plot of an MYC core signature (n = 77) described by Ji et al that is strongly attenuated in MYC tumors expressing REG3A compared to MYC tumors not expressing REG3A. Data are averages ± SEM. The Mann-Whitney U test was performed for analysis. NS or no statistical indication, no significance. Abbreviation: REG3A, regenerating family member 3 alpha.

FIGURE 5

Preneoplastic and tumor livers of HCC expressing REG3A show substantial reduction of O-GlcNAcylation in mice and humans. Anti-RL2 immunoblots for O-GlcNAc. Proteins in nuclear (Nuc) and cytosolic (Cyt) fractions of (A) preneoplastic (preK) livers (n = 6), (B) in the nontumor area (n = 6) and (C) the tumor area (n = 12) of HCC from transgenic mice homozygous for MYC (MYC-TG) and transgenic mice double homozygous for MYC and REG3A (MYC/REG3A-TG). Quantification by densitometry below western blots. Each dot represents a sample from an independent mouse. (D) Western blots for O-GlcNAc proteins in livers of WT (n = 6) and transgenic mice overexpressing REG3A in the liver (REG3A-TG; n = 4) with DEN-induced HCC. (E) Western blots for O-GlcNAc proteins and endogenous REG3A protein in human HCC. Quantification of O-GlcNAc proteins in 18 NT and T matched samples. P1 = patient 1. Increased O-GlcNAcylation in 13 T samples (red line) and no detectable real change in O-GlcNAcylation in 5 T samples (green line) compared with matched NT samples (p = 0.0047). (F) O-GlcNAcylation in human HCC samples as a function of whether they express or not endogenous REG3A in the NT and T areas. Ponceau S and Coomassie blue staining were used as loading controls. Data are averages ± SEM. The Mann-Whitney U test was performed except for (E) (paired Student t test) and (F) (Student t test). NS or no statistical indication, no significance. Abbreviations: DEN, diethylnitrosamine; NT, nontumor; REG3A, regenerating family member 3 alpha; T, tumor; WT, wild-type.

FIGURE 6

REG3A bound to glucose and glucose intermediates is a regulator of protein glycosylation. (A) Quantification of HS and CS disaccharides by ion-pair reversed-phase chromatography in HuH7 cells expressing REG3A or EXTL3 alone, or both, or appropriate empty vectors. (B) Representative sugar slot blot of mono- and polysaccharides deposited at the indicated doses and hybridized with recombinant REG3A lectin. Lactose, mannan: positive control for REG3A binding. (C) Slot blot of indicated sugars in the presence of a full-length human recombinant REG3A protein (rcREG3A) preincubated with BSA or BSA + lactose (Glc-Gal). (D) Slot blot of indicated sugars in the presence of rcREG3A or a mutant recombinant REG3A protein (rcREG3A^EPN/GPG) proteins. (E) Immunoblots for O-GlcNAc (anti-RL2) and REG3A in nuclear fractions of HuH7 cells expressing REG3A, REG3A^EPN/GPG or empty vector. Densitometry quantification (n = 3). (F) Enzymatic quantification of cellular UDP-GlcNAc content in preneoplastic (preK), nontumor (NT adjacent to a tumor), and tumor (T) liver samples from MYC/REG3A and MYC-TG transgenic mice. Each dot represents a sample of an individual mouse. (G) Rates of cellular ATP production in HuH7 cells expressing REG3A, REG3A^EPN/GPG, or the empty vector, showing significant changes in total, mitochondrial, and glycolytic ATP production under the action of the full-length REG3A protein (n = 3). (H) Lectin blot for biotin-labeled PHA-L (phaseolus vulgaris leucoagglutinin) lectin on preneoplastic liver extracts from mice transgenic for MYC (MYC-TG) and for MYC and REG3A (MYC/REG3A-TG). Densitometry quantification (n = 6). (I) Glycogen concentrations in preneoplastic liver extracts (preK; left) from the mice shown (n = 6) and (J) in HuH7 cells expressing full-length REG3A or the mutant REG3A^EPN/GPG (right) (n = 4). (K) Lectin blot for biotin-labeled WGA in extracts of preneoplastic (PreK) liver from MYC and MYC/REG3A transgenic mice and quantification by densitometry (n = 6). (L) A schematic view of carbohydrate metabolism from glucose and its intermediates. Gray = de novo synthesis of UDP-GlcNAc from glucose through the hexosamine synthesis pathway; Blue: salvage of GlcNAc through the metabolic processes that produce UDP-GlcNAc. *sugars bound to REG3A. Red arrows: carbohydrate pathways downregulated by REG3A. Equal symbols: glucose pathways not modulated by REG3A. Data are averages ± SEM. The Mann-Whitney U test was used for analysis, except for (E), (F), (G), and (J) (ANOVA test followed by a post hoc test). NS or no statistical indication, no significance. Abbreviations: CS, chondroitin sulfate; EXTL3, exostosin-like glycosyltransferase 3; HS, heparan sulfate; PHA-L, phytohemaglutinin-L; REG3A, regenerating family member 3 alpha; WGA, wheat germ agglutinin.