Main

Alteration in cellular metabolism is one of the hallmarks of cancers. Cancer-associated metabolic changes include the deregulated uptake of glucose and amino acids, use of opportunistic modes of nutrient acquisition, use of glycolysis/tricarboxylic acid cycle intermediates for biosynthesis and nicotinamide adenine dinucleotide phosphate production, increased demand for nitrogen, alterations in metabolite-driven gene regulation, and metabolic interactions with the microenvironment.1 A common goal of these alterations is to maintain cell proliferation in the nutrient-poor environment. Cells with an adequate oxygen and nutrient supply obtain energy from mitochondrial respiration via the conversion of pyruvate to acetyl-CoA. However, cancer cells that exhibit a metabolic shift from mitochondrial respiration to aerobic glycolysis (Warburg effect) cannot efficiently obtain acetyl-CoA from pyruvate. Instead, cancer cells obtain acetyl-CoA from glutamine and short-chain fatty acids such as acetate, butyrate, and proprionate.2, 3

Acetyl-CoA is an important metabolite of intermediary metabolism. Acetyl-CoA is a substrate of tricarboxylic acid cycle, a key precursor of lipid synthesis, and a donor of the acetyl group for protein and histone acetylation.4 Furthermore, acetyl-CoA participates in the regulation of cell growth and proliferation.5 In aerobic conditions, acetyl-CoA is generated by the conversion of pyruvate. However, acetyl-CoA may be produced by the conversion of glutamine and short-chain fatty acids, such as butyrate and acetate, in a hypoxic or nutrient-deprived state. Acetyl-CoA synthetases convert acetate to acetyl-CoA and there are at least three acetyl-CoA synthetases in mammals, including acetyl-CoA synthetase 1, 2, and 3. Acetyl-CoA synthetase 1 and 3 are located in mitochondria, whereas acetyl-CoA synthetase 2 is located in the nucleus and cytoplasm.6

Acetyl-CoA synthetase-2 overexpression has been reported in several types of cancers, including hepatocellular carcinomas, gliomas, and breast carcinomas, and the overexpression of acetyl-CoA synthetase-2 has been associated with a worse prognosis in these malignancies.7, 8 It is of interest whether there are associations of acetyl-CoA synthetase-2 with colorectal carcinoma, because the primary source of acetate production is fermentation by intestinal microflora. However, the clinicopathologic characteristics and prognostic implication of acetyl-CoA synthetase-2 expressions in colorectal carcinomas have not been described previously.

In this study, we evaluated acetyl-CoA synthetase-2 expression alterations along the multistep colorectal carcinogenesis pathway including normal mucosa, premalignant lesions, and colorectal carcinomas, and also investigated the clinicopathologic characteristics and prognostic implication of acetyl-CoA synthetase-2 expression in colorectal carcinomas.

Materials and methods

Colorectal Carcinoma Cases

A total of 1527 colorectal carcinoma patients underwent curative surgery in Seoul National University Hospital, Seoul, South Korea, from January 2004 to December 2007. Initially, 1133 cases were subjected to clinicopathologic and molecular analyses following the exclusion of patients who refused to participate in the molecular study, or had non-invasive carcinomas, a neo-adjuvant treatment history, familial adenomatous polyposis, multiple tumors, or recurrent tumors.9 Among the 1133 cases, 1106 cases with complete clinicopathologic data and formalin-fixed, paraffin-embedded tissues were finally selected for the current study. This study was approved by the Institutional Review Board.

RNA Extraction and Reverse-Transcription Quantitative PCR

We prepared total RNA from paired tumors and normal mucosa in randomly selected 12 formalin-fixed, paraffin-embedded colorectal carcinoma tissues using RNeasy formalin-fixed, paraffin-embedded kit (Qiagen, Hilden, Germany). Five microliters of total RNA were reverse transcribed with First Strand cDNA Synthesis System (LeGene, San Diego, CA, USA) and diluted to 40 ng/μl, using 1 × TE Solution pH 8.0 (IDT). PCR reactions were performed using Power SYBR Green PCR Master Mix (Life Technologies, Carlsbad, CA, USA) according to the manufacturer’s recommendations. One hundred and twenty nanograms of cDNA in each sample was used for the PCR reaction and the cycling conditions were as follows: initial denaturation for 10 min at 95 °C, followed by 50 cycles of 45 s at 94 °C, 45 s at 53 °C, and 45 s at 72 °C in a CFX384 real-time PCR system (Bio-Rad). Primers used were as follows: acetyl-CoA synthetase-2 forward, 5′-GGATTCCAGCTGCAGTCTTC-3′ and acetyl-CoA synthetase-2 reverse, 5′-CAGCCAGCTCCTTCAGGTT-3′; and GAPDH forward, 5′-TGGTAAAGTGGATATTGTTGC-3′ and GAPDH reverse, 5′-GCCATGGGTGGAATCATA-3′. All experiments were performed in triplicate.

Tissue Microarray Construction

Through histologic examination, we marked portions that represent the tumor area for each formalin-fixed, paraffin-embedded tissue of 1106 colorectal carcinomas. A pair of 2 mm core tumor tissues was subsequently extracted from each formalin-fixed, paraffin-embedded tissues (donor block) and rearranged in a new recipient tissue microarray block using a trephine apparatus as previously described.10 To analyze acetyl-CoA synthetase-2 expression in normal mucosa, we generated an additional 2 mm core tissue microarray, which contains the normal mucosa of 54 randomly selected patients from the surgical margin. To evaluate acetyl-CoA synthetase-2 expression in premalignant lesions, we used a pre-constructed 2 mm core tissue microarray obtained from 157 colorectal adenomas/polyps.11 Of the total 1133 cases, 35 patients received concurrent metastasectomy for liver metastasis. Among them, 23 formalin-fixed, paraffin-embedded tissues were available. We generated 2 mm core tissue microarray from metastatic carcinoma in the liver.

Immunohistochemistry

Immunohistochemical analysis was performed with commercially available rabbit polyclonal anti-acetyl-CoA synthetase-2 antibody (Cell Signaling, cat# 3658S, 1:200), anti-KRT7 (CK7) (clone OV-TL 12/30, DAKO), anti-KRT20 (CK20) (clone Ks20.8, DAKO), anti-CDX2 (clone EPR2764Y ready-to-use, CellMarque), and anti-MKI67 (Ki-67) (clone MIB-1, DAKO).8, 12 After immunohistochemical staining, each tissue microarray slides were scanned using an Aperio Scanscope (Leica Biosystems, New Castle, UK). Acetyl-CoA synthetase-2 expression levels were independently assessed by two pathologists (JMB and JHK). Staining intensity of acetyl-CoA synthetase-2 was relatively homogenous within each tissue core and between a pair of tissue cores from the same cancers. We evaluated the staining intensity of the cytoplasmic acetyl-CoA synthetase-2 expression as 0 (no cytoplasmic stain), 1 (mild cytoplasmic stain), 2 (moderate cytoplasmic intensity), or 3 (strong cytoplasmic intensity). The levels of cytoplasmic acetyl-CoA synthetase 2 staining were subsequently categorized as cytoplasmic acetyl-CoA synthetase-2-low (cytoplasmic acetyl-CoA synthetase-2 intensity 0 and 1) and cytoplasmic acetyl-CoA synthetase-2-high (cytoplasmic acetyl-CoA synthetase-2 intensity 2 and 3) groups. The cutoffs for KRT7 expression, decreased KRT20 expression, and decreased CDX2 expression were 10, 50, and 20%, respectively.12 MKI67 labeling indices were measured using the Nuclear v9 algorithm in Aperio image analysis system.

Clinicopathologic Analysis

The clinicopathologic characteristics, including age, sex, tumor location, and TNM stage, were obtained from electronic medical records. Through microscopic examination of representative tumor section, two pathologists (JMB and GHK) without knowledge of the CpG island methylator phenotype, microsatellite instability, and KRAS and BRAF mutation statuses assessed each specimen for tumor differentiation, luminal necrosis, Crohn-like lymphoid reaction, the number of tumor-infiltrating lymphocytes, luminal serration, and extraglandular mucin production. The progression-free survival data were extracted from the patient’s medical records, direct interviews with the surviving patients or their family members, or from death registry offices. The baseline characteristics of 1106 colorectal carcinoma patients were as follows: the median age was 62 years (range: 20–90) and the male to female ratio was 1.57 (677 males and 456 females). Two hundred and seventy-nine patients had proximal colon carcinoma, whereas 441 and 413 patients had distal colon and rectal carcinoma, respectively. Seven hundred and eighty-five patients received 5-fluorouracil-based adjuvant chemotherapy.

KRAS/BRAF Mutations and Microsatellite Instability Analyses

Through histologic examination, the representative tumor portions of each formalin-fixed, paraffin-embedded tissue were marked and subsequently subjected to manual micro-dissection. The dissected tissues were collected in microtubes that contained lysis buffer and proteinase K, and were incubated at 55 °C for 2 days. Direct sequencing of the KRAS codons 12 and 13, and allele-specific PCR for the BRAF codon 600 were performed as previously described.13 The microsatellite instability status of each tumor and paired normal mucosa sample was determined using five Bethesda markers, including BAT25, BAT26, D2S123, D5S346, and D17S250. Microsatellite instability-high was defined as two or more markers associated with alleles of altered size in tumor DNA compared with DNA from normal mucosa and microsatellite instability-low was defined as one marker associated with allele of altered size in tumor DNA compared with DNA from non-tumor tissue. Microsatellite-stable was defined as the absence of instability.

Analysis of the CpG Island Methylator Phenotype

Bisulfite DNA modification and real-time PCR-based methylation assays (MethyLight) were performed as previously described.13, 14, 15 We quantified the methylation of eight CpG island methylator phenotype-specific CpG islands (CACNA1G, CDKN2A, CRABP1, IGF2, MLH1, NEUROG1, RUNX3, and SOCS1). The primer sequences and PCR conditions have been previously described.13, 16 M.SssI-treated genomic DNA was used as a reference sample. The percentage of methylated reference at a particular locus was calculated by 100 × (methylated reaction at the GENE/control reaction at the ALU ratio)sample/(methylated reaction at the GENE/control reaction at the ALU)M.SssI-treated placental genomic DNA. The MethyLight assay was performed in triplicate and the median value was calculated as the representative value of the methylation level of each marker. A CpG island locus with a median percentage of methylated reference >4 was considered to be methylated. CpG island methylator phenotype-high was defined as five or more methylated markers of the eight-marker CpG island methylator phenotype panel, CpG island methylator phenotype-low was defined as one to four of the eight markers methylated, and CpG island methylator phenotype-negative was defined as 0 methylated markers.

Genome-Wide Expression Analysis

To validate the results from our cohort, we analyzed two publicly available genome-wide mRNA expression data (GSE39582 and The Cancer Genome Atlas (TCGA) COADREAD data sets). For GSE39582, raw CEL files based on the platform of the Affymetrix Human Genome U133 Plus 2.0 Array (Affymetrix, Santa Clara, CA, USA) were downloaded from the Gene Expression Omnibus database. Next, raw data were normalized using the robust multi-array average. The probes were mapped to the gene symbols based on the microarray annotation information in R Bioconductor package hgu133a2. The differentially expressed genes were calculated via an empirical Bayes model using limma package in R. For TCGA data set, the normalized RSEM values from RNA sequencing of the normal mucosa and tumor tissues were downloaded from www.firebrowse.org. The differentially expressed genes were calculated via generalized linear model using edgeR package in R. To analyze the differentially expressed genes at the functional level, Gene Ontology Biological Process enrichment analyses were performed using the PANTHER classification system in the Gene Ontology Consortium (geneontology.org).

Statistics

SAS software (version 9.4 for Microsoft Windows; SAS Institute, Cary, NC, USA) was used for the statistical analysis in our cohort. Bioinformatic analysis and graph works were performed using R software. To compare the acetyl-CoA synthetase-2 expression between the tumor and normal mucosa, Wilcoxon’s signed-rank test was performed. To compare clinicopathologic characteristics to acetyl-CoA synthetase-2 expression, we performed Pearson’s χ2-test. The age of colorectal carcinoma group and Ki-67 labeling index according to cytoplasmic acetyl-CoA synthetase-2 expression was compared using Wilcoxon’s rank-sum test. For the survival analysis, 5-year progression-free survival was calculated using the log-rank test with a Kaplan–Meier curve. Hazard ratios were calculated using the Cox proportional hazard model. The assumption of the proportional hazards was verified by plotting the log(−log(S(t)) against the time of the study. In the modeling process, all variables that were associated with progression-free survival with a P<0.1 were entered into an initial model; these variables were subsequently reduced by backward elimination. All statistical tests were two-sided and statistical significance was defined as P<0.05 for survival analysis and P<0.005 for clinicopathologic analysis due to multiple comparison.

Results

Acetyl-CoA Synthetase-2 Expression is Downregulated During Colorectal Carcinogenesis

To explore whether there are alterations in acetyl-CoA synthetase-2 expression in colorectal carcinomas, we compared acetyl-CoA synthetase-2 expression in 54 paired carcinoma tissues and normal tissues using immunohistochemistry. The normal mucosa exhibited moderate-to-strong cytoplasmic staining, whereas the carcinoma tissues exhibited the absence of cytoplasmic staining to moderate cytoplasmic staining (Wilcoxon’s signed-rank P<0.001; Figures 1a and c). To investigate whether acetyl-CoA synthetase-2 downregulation occurs at invasive carcinomas or at the premalignant stage, we performed immunohistochemistry in 157 colorectal polyps, including hyperplastic polyps, conventional adenomas, sessile serrated adenomas, and traditional serrated adenomas. We obtained the results of immunohistochemistry in 50 hyperplastic polyps and 107 premalignant lesions (42 conventional adenomas with low-grade dysplasia, 18 conventional adenomas with high-grade dysplasia, 31 sessile serrated adenomas, and 16 traditional serrated adenomas). All 50 hyperplastic polyps and 107 premalignant lesions exhibited moderate-to-strong cytoplasmic staining regardless of the histologic type or grade of dysplasia (Figure 1b). Therefore, we concluded that the downregulation of acetyl-CoA synthetase-2 occurs at the invasive cancer stage.

Figure 1
figure 1

Acetyl-CoA synthetase-2 expression alterations in colorectal carcinogenesis. (a) High acetyl-CoA synthetase-2 expression in normal colonic mucosa. (b) High acetyl-CoA synthetase-2 expression in a premalignant lesion (conventional adenoma with low-grade dysplasia). (c) Low acetyl-CoA synthetase-2 expression in a colorectal carcinoma. (d) Low acetyl-CoA synthetase-2 expression in liver metastasis. (e) Comparison of acetyl-CoA synthetase 2 mRNA expression levels between normal mucosa and tumor tissues of our colorectal carcinoma samples (n=12) measured by reverse-transcription quantitative PCR. (f) Volcano plot for acetyl-CoA synthetase-2 mRNA expression levels in the GSE39582 data set. (g) Comparison of acetyl-CoA synthetase-2 mRNA expression levels between normal mucosa and tumor tissues in the TCGA data set.

Full size image

To test whether acetyl-CoA synthetase-2 expression is downregulated in metastatic colorectal carcinomas, we performed immunohistochemistry in 23 metastatic colorectal carcinomas in the liver. Twenty-one (91.3%) out of 23 metastatic colorectal carcinomas showed an absence of cytoplasmic staining or mild cytoplasmic staining (Figure 1d). Therefore, we concluded that the expression of acetyl-CoA synthetase-2 decreases along colorectal carcinoma progression.

To validate the downregulation of acetyl-CoA synthetase-2 in invasive cancers, we evaluated acetyl-CoA synthetase-2 mRNA expression using reverse-transcription quantitative PCR in 12 paired carcinoma and normal mucosa tissues. The reverse-transcription quantitative PCR results showed that carcinoma tissues exhibited lower mRNA expression than normal mucosa (Wilcoxon’s signed-rank P<0.001; Figure 1e). To validate this finding with independent data sets, we compared acetyl-CoA synthetase-2 mRNA expression between normal mucosa and tumor tissues with GSE39582 and TCGA data sets. In GSE39582, tumor tissues exhibited significantly lower acetyl-CoA synthetase-2 expression than normal mucosa (log2 fold change= −1.59, adjusted P=5.36 × 10−14) (Figure 1f). In TCGA data set, the tumor tissues exhibited significantly lower RSEM values than normal mucosa. (log2 fold change= −0.99, P=3.57 × 10−33; Figure 1g).

Clinicopathologic and Molecular Implications of Acetyl-CoA Synthetase-2 Downregulation in Colorectal Carcinoma

Detailed clinicopathologic and molecular features according to acetyl-CoA synthetase-2 expression status are summarized in Tables 1 and 2. As the normal mucosa showed moderate-to-strong acetyl-CoA synthetase-2 expression, we subdivided the colorectal carcinomas to colorectal carcinomas with acetyl-CoA synthetase-2-low expression (no to weak acetyl-CoA synthetase-2 expression) and colorectal carcinomas with acetyl-CoA synthetase-2-high expression (moderate-to-strong acetyl-CoA synthetase-2 expression). The colorectal carcinomas with acetyl-CoA synthetase-2-low expression exhibited an advanced T category (P<0.001), N category (P<0.001), M category (P=0.001), and stage (P<0.001) compared with colorectal carcinomas with acetyl-CoA synthetase-2-high expression. The colorectal carcinomas with acetyl-CoA synthetase-2-low expression exhibited tendency of ulcerative gross morphology (P=0.003), poor differentiation (P=0.007), and frequent tumor budding (P=0.015) compared with the colorectal carcinomas with acetyl-CoA synthetase-2-high expression. In the molecular analysis, the colorectal carcinomas with acetyl-CoA synthetase-2-low expression were marginally associated with KRT7 expression (P=0.008) and decreased KRT20 expression (P=0.031), and strongly associated with decreased CDX2 expression (P=0.001). The colorectal carcinomas with acetyl-CoA synthetase-2-low expression exhibited lower MKI67 labeling indices (median: 10.4, min–max: 0.1–71.0) than colorectal carcinomas with acetyl-CoA synthetase-2-high expression (median: 21.9, min–max: 0.1–81.6; P<0.001). Microsatellite instability status, and KRAS and BRAF mutation rates were not significantly different between colorectal carcinomas with acetyl-CoA synthetase-2-low expression and colorectal carcinomas with acetyl-CoA synthetase-2-high expression.

Table 1 Clinicopathologic characteristics of colorectal carcinomas according to acetyl-CoA synthetase-2 expression status
Full size table
Table 2 Molecular characteristics of colorectal carcinomas according to acetyl-CoA synthetase-2 expression status
Full size table

To understand the functional difference between the colorectal carcinomas with acetyl-CoA synthetase-2-low expression and the colorectal carcinomas with acetyl-CoA synthetase-2-high expression, we performed genome-wide expression analysis using the GSE39582 data. In GSE39582, 562 colorectal carcinomas were classified into the acetyl-CoA synthetase-2-low group (375 colorectal carcinomas that exhibited a lower 2/3 of the acetyl-CoA synthetase-2 expression level) and the acetyl-CoA synthetase-2-high group (187 colorectal carcinomas that exhibited an upper 1/3 of the acetyl-CoA synthetase-2 expression level). Fifteen genes were upregulated and 35 genes were downregulated in the acetyl-CoA synthetase-2-low group compared with the acetyl-CoA synthetase-2-high group (Figure 2a and Supplementary Tables 1 and 2). The Gene Ontology analysis results indicated that genes associated with cell motility, chemotaxis, and epithelial–mesenchymal transition were upregulated in acetyl-CoA synthetase-2-low group, whereas genes associated with digestive function and metabolism were upregulated in the acetyl-CoA synthetase-2-high group (Figure 2b and Supplementary Tables 3 and 4). In the TCGA data set, 104 genes were upregulated and 58 genes were downregulated in acetyl-CoA synthetase-2-low group compared with the acetyl-CoA synthetase-2-high group. The Gene Ontology analysis of the TCGA data set demonstrated Gene Ontology terms similar to the GSE39582 (Supplementary Tables 5).

Figure 2
figure 2

Gene expression analysis in the GSE39582 data set. (a) Heatmap of differentially expressed genes between the acetyl-CoA synthetase-2-low group and acetyl-CoA synthetase-2-high group. (b) Gene Ontology analysis.

Full size image

Prognostic Implication of Acetyl-CoA Synthetase-2 Downregulation in Colorectal Carcinoma

In the univariate survival analysis, the colorectal carcinomas with acetyl-CoA synthetase-2-low expression showed worse 5-year progression-free survival (P<0.001) compared with the colorectal carcinomas with acetyl-CoA synthetase-2-high expression (Figure 3a). To verify whether acetyl-CoA synthetase-2 expression has an independent prognostic implication, we performed a multivariate survival analysis. In the multivariate survival analysis, colorectal carcinomas with acetyl-CoA synthetase-2-low expression exhibited an independently poor 5-year progression-free survival (hazard ratio, 1.39; 95% confidence interval, 1.08–1.79; P=0.01; Table 3). There was no significant interaction or co-linearity among the variables in the final model. To validate this finding, we analyzed progression-free survival in the GSE39582 data set. In GSE39582, the acetyl-CoA synthetase-2-low group showed stage-independent poor progression-free survival (hazard ratio, 1.49; 95% confidence interval, 1.07–2.09; P=0.02) compared with the patients in the acetyl-CoA synthetase-2-high group (Figure 3b).

Figure 3
figure 3

Kaplan–Meier survival curves for progression-free survival. (a) Our original study cohort (SNUH cohort). (b) Validation cohort (GSE39582 cohort).

Full size image
Table 3 Univariate and multivariate Cox proportional hazard models for 5-year progression-free survival
Full size table

Discussion

Acetyl-CoA synthetase-2 mRNA is highly expressed in the colon, adipose tissue, and skeletal muscle (www.proteinatlas.org). In the intestinal lumen, microflora ferment dietary fibers and produce short-chain fatty acids. Thus, acetate can be used by acetyl-CoA synthetase-2 in normal colonocytes. However, acetyl-CoA synthetase-2 expression is markedly reduced in colorectal carcinomas, because cancerous colonocytes rely on glycolysis as their primary energy source.17 Furthermore, the results from our current study indicated that acetyl-CoA synthetase-2 expression is decreased in a specific portion of colorectal carcinomas. The colorectal carcinomas with acetyl-CoA synthetase-2-low expression exhibited a tendency of gastric differentiation (increased KRT7 expression and decreased KRT20/CDX2 expression), whereas the colorectal carcinomas with acetyl-CoA synthetase-2-high expression were associated with intestinal differentiation (KRT7-negative expression and retained KRT20/CDX2 expression; Table 2). Moreover, the Gene Ontology analysis results demonstrated that the genes associated with digestive function and metabolism were downregulated in the colorectal carcinomas with acetyl-CoA synthetase-2-low expression compared with the colorectal carcinomas with acetyl-CoA synthetase-2-high expression (Figure 2b).

Recently, a series of studies have elucidated an oncogenic role and prognostic implications of acetyl-CoA synthetase-2 expression in several types of cancers. Mashimo et al8 reported that acetyl-CoA synthetase-2 expression in gliomas was increased along the increment of the WHO grade. Moreover, patients with acetyl-CoA synthetase-2-high expression showed worse overall survival than patients with acetyl-CoA synthetase-2-low expression in grade II or III gliomas.8 Comerford et al7 reported that acetyl-CoA synthetase-2-high expression was associated with shorter overall survival compared with acetyl-CoA synthetase-2-negative cases in triple-negative breast carcinomas. Schug et al18 reported that acetyl-CoA synthetase-2 expression was correlated with the stage of breast carcinomas and an increment of acetyl-CoA synthetase-2 expression in metastatic breast carcinomas than primary carcinomas. However, in our present study, acetyl-CoA synthetase-2-low expression was associated with clinicopathologic characteristics associated with aggressive behavior such as an advanced TNM stage, poor differentiation, and frequent tumor budding (Table 1). In the survival analysis, the colorectal carcinomas with acetyl-CoA synthetase-2-low expression exhibited worse progression-free survival compared with the colorectal carcinomas with acetyl-CoA synthetase-2-high expression (Figure 3 and Table 3). These features suggested that decreased acetyl-CoA synthetase-2 expression in colorectal carcinomas is associated with aggressive clinical behavior in contrast to other types of malignancies. The adverse prognostic effect of the loss of acetyl-CoA synthetase-2 expression in gastric carcinoma suggested that the prognostic implication of acetyl-CoA synthetase-2 expression in gastrointestinal tract malignancy may differ from malignancies in other organs.19

The reason why decreased acetyl-CoA synthetase-2 expression exhibited poor clinical outcome in colorectal carcinomas remains unknown. In vitro study results have suggested that alterations in acetate metabolism may differentially contribute to cancer cell survival. Acetate induces apoptosis by acting as an histone deacetylase inhibitor. Soliman et al20 demonstrated that acetate supplementation increases histone acetylation and inhibits histone deacetylase activity. In the study by Marques et al21, acetate induced apoptosis and decreased cell proliferation in colorectal carcinoma cell lines including HCT-15 and RKO. Furthermore, acetyl-CoA synthetase-2 induces cancer cell proliferation and reduces apoptosis by consuming acetate to generate acetyl-CoA. Schug et al18 suggested that the inhibition of acetyl-CoA synthetase-2 may have a therapeutic effect in colorectal carcinomas, because inhibition of acetyl-CoA synthetase-2 may suppress tumor cell growth and proliferation. However, many studies have demonstrated that colorectal carcinomas with high proliferation index are paradoxically associated with a lower TNM stage and better clinical outcome.22, 23, 24, 25 In our present study, the colorectal carcinomas with acetyl-CoA synthetase-2-high expression showed higher Ki-67 labeling index, lower TNM stage, and favorable clinical outcome compared with the colorectal carcinomas with acetyl-CoA synthetase-2-low expression. Based on the results of our present study and in vitro study performed by other groups, we propose a model of roles of acetyl-CoA synthetase-2 downregulation in colorectal carcinogenesis (Figure 4).

Figure 4
figure 4

Acetyl-CoA synthetase-2 downregulation in multistep colorectal carcinogenesis. (a and b) A case of colorectal carcinoma with pre-existing high-grade dysplasia. (a) H&E stain (upper: high-grade dysplasia, lower: invasive carcinoma). (b) Acetyl-CoA synthetase-2 immunohistochemistry. The retained acetyl-CoA synthetase-2 expression in high-grade dysplasia components (upper glands) and the decreased acetyl-CoA synthetase-2 expression in invasive carcinoma components (lower glands) are noteworthy. (c) A model of roles of acetyl-CoA synthetase-2 downregulation in colorectal carcinogenesis.

Full size image

Our present study has several limitations. First, our study is a hospital-based retrospective study. This can be a source of selection biases including referral nature, preferences, reputations of the hospital and physicians, and physical and psychological distances. Second, molecular alterations in rectal carcinomas were underrepresented because of the exclusion of colorectal carcinomas with preoperative neoadjuvant chemotherapy and/or radiotherapy history. Third, our study showed the lower fractions of KRAS mutation, BRAF mutation, microsatellite instability-high, and CpG island methylator phenotype-high compared with those of Western countries. These discrepancies might originate from selection bias of hospital-based study, low analytical sensitivity, or ethnic difference. In general, fractions of those molecular alterations in colorectal carcinomas of Eastern Asia are lower than those of Western countries.26, 27, 28, 29 To rule out the possibility of analytical problem, we validated our technologies for KRAS mutation and BRAF mutation using pyrosequencing and targeted resequencing, and all methods reported similar fractions of KRAS mutation and BRAF mutation (data not shown).

To the best of our knowledge, this investigation is the first to demonstrate the clinicopathologic characteristics and prognostic implication of acetyl-CoA synthetase-2 expression in colorectal carcinomas. We showed that most of colorectal carcinomas exhibit decreased acetyl-CoA synthetase-2 expression, whereas normal colonocytes and premalignant lesions express acetyl-CoA synthetase-2 at a moderate-to-high intensity (Figure 1). Furthermore, we demonstrated comprehensive clinicopathologic and molecular characteristics of colorectal carcinomas depending on acetyl-CoA synthetase-2 expression status using our original cohort and other validation data sets. According to our current study, unlike in other malignancies, acetyl-CoA synthetase-2 exhibited tumor suppressor properties in colorectal carcinoma. However, the precise roles and mechanisms of downregulation of acetyl-CoA synthetase-2 expression in colorectal carcinogenesis remain to be elucidated. Further in vitro and in vivo studies on molecular basis and biological functions of acetyl-CoA synthetase-2 alterations in colorectal carcinoma should be conducted.

In conclusion, downregulation of acetyl-CoA synthetase-2 expression is a metabolic hallmark of malignant transformation and aggressive behavior in colorectal carcinoma.