Article Text

other Versions

Download PDFPDF
Loss of SIRT1 histone deacetylase expression associates with tumour progression in colorectal adenocarcinoma
  1. Si-Hyong Jang,
  2. Kyeung-Whan Min,
  3. Seung Sam Paik,
  4. Ki-Seok Jang
  1. Department of Pathology, College of Medicine, Hanyang University, Seoul, Korea
  1. Correspondence to Professor Ki-Seok Jang, Department of Pathology, College of Medicine, Hanyang University, 222 Wangsimni-ro, Seongdong-gu 133-791, Seoul, Republic of Korea; medartisan{at}hanyang.ac.kr

Abstract

Background The class III histone deacetylase SIRT1 is a nicotinamide adenine dinucleotide (NAD+)-dependent deacetylase, and has been reported to serve diverse roles in various biological processes, such as caloric restriction, apoptosis, neuronal protection, cell growth, differentiation and tumourigenesis. With respect to tumourigenesis, there have been conflicting data supporting whether SIRT1 act as a tumour promoter or as a tumour suppressor.

Methods SIRT1 protein expression, determined by immunohistochemistry, was investigated in human normal colonic mucosa, adenoma, adenocarcinoma and metastatic tissue samples.

Results All normal colonic mucosa showed SIRT1 expression with no exception, and 42 (80.8%) of 52 adenomatous polyps were positive for SIRT1. However, only 208 (41.9%) of 497 colorectal adenocarcinomas were positive. Moreover, 45 (35.7%) of 126 metastatic tissues were positive. Collectively, the SIRT1 expression was gradually decreased during carcinogenesis and tumour progression. The associations between SIRT1 expression and clinicopathological parameters revealed that loss of SIRT1 expression was associated with proximal tumour location, mucinous histology and defective mismatch repair protein expression. This suggests that loss of SIRT1 expression is associated with the microsatellite instability phenotype of colorectal adenocarcinoma. In survival analyses, the loss of SIRT1 expression was significantly associated with overall survival (p=0.027, log-rank test) in univariable analysis, but multivariable analysis failed to achieve significance.

Conclusions SIRT1 expression was gradually decreased during the normal–adenoma–adenocarcinoma–metastasis sequence, suggesting a possible role of SIRT1 in tumour suppression in the colorectum, and a probable link to the microsatellite instability pathway.

  • Colorectal cancer
  • SIRT1
  • tumour suppressor
  • prognosis
  • microsatellite instability
  • oncology
  • tumour biology
  • histopathology
  • cancer
  • colorectal cancer

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

The class III histone deacetylase SIRT1 is a mammalian orthologue of silent information regulator 2 (Sir2), a yeast nicotinamide adenine dinucleotide (NAD+)-dependent deacetylase that mediates heterochromatin formation.1 SIRT1 has been reported to fulfil diverse roles in various biological processes, such as caloric restriction, apoptosis, neuronal protection, cell growth, differentiation and tumourigenesis.2–9

In relation to tumourigenesis, there is conflicting evidence about whether SIRT1 acts as a tumour promoter or a tumour suppressor.10 On the one hand, there is evidence of involvement of SIRT1 in the epigenetic silencing of tumour suppressor genes and proteins with DNA damage repair functions, such as p53, SFRP (secreted frizzled-related protein), CDH1 (the E-cadherin gene), MLH1 and GATA.5 9 In addition, expression of SIRT1 is regulated by tumour suppressors, such as HIC1 (hypermethylated in cancer-1) and DBC1 (deleted in breast cancer-1).11 12 Additionally, it has been shown that SIRT1 is elevated in human prostate cancer, acute myeloid leukaemia and non-melanoma skin cancer.13–15 These results support a tumourigenic role of SIRT1. On the other hand, recent studies have provided evidence that SIRT1 serves as a tumour suppressor. Thus, overexpression of SIRT1, in a β-catenin-driven mouse model, significantly reduced tumour formation, cancer cell proliferation and animal morbidity. These effects were due to inhibition of the transcriptional activity and nuclear translocation of β-catenin.16 In breast cancer, the level of SIRT1 in BRCA1-associated cancers was lower than in BRCA1-wild type controls, and restoration of SIRT1 expression in BRCA1 mutant cancer cells inhibited their proliferation in vitro, and tumour formation by the cells in vivo.17 In addition, SIRT1 expression is lower in many human cancers, including glioblastoma, bladder cancer, prostate cancer and ovarian cancer than in the corresponding normal control tissues.18 These findings support the tumour suppressor potential of SIRT1 overexpression.

To investigate whether SIRT1 is involved in the development and progression of colorectal cancer, we evaluated its expression by immunohistochemistry in normal human colonic mucosa, adenoma, adenocarcinoma and metastatic tissue samples. We also assessed the association between SIRT1 expression and the conventional clinicopathological parameters of colorectal cancer, and determined whether SIRT1 expression could predict patient outcomes.

Materials and methods

Patients and specimens

In this retrospective study we enrolled a consecutive series of 497 patients with colorectal adenocarcinoma. All the patients were diagnosed and treated at the Hanyang University Hospital (Seoul, Korea) between January 1991 and August 2001. There were 281 men and 216 women. Median age was 57.7 years. The tumours were located in the caecum (n=17), ascending colon (n=70), hepatic flexure (n=10), transverse colon (n=25), splenic flexure (n=4), descending colon (n=22), sigmoid colon (n=104) and rectum (n=245). Mean tumour size was 5.7 cm, and the mean follow-up period after surgery was 5.9 years. A total of 171 (34.4%) of the patients died and 326 (65.6%) remained alive. Twenty-four samples of normal colonic mucosa, 52 adenomas, 65 metastatic lymph nodes (LN) and 61 distant metastatic lesions were randomly selected to evaluate the role of SIRT1 in multi-step carcinogenesis.

All the tissue samples were formalin-fixed and paraffin-embedded. H&E slides, pathology reports, and other medical records were reviewed to confirm the diagnoses and clinicopathological parameters including age, gender, tumour location, tumour size, depth of invasion, LN metastasis, distant metastasis, AJCC stage, degree of differentiation, lymphovascular invasion and patient survival. Mismatch repair (MMR) protein expression was detected by immunohistochemistry, and classified into ‘intact’ and ‘defective’. Intact MMR protein expression was defined by the presence of both MLH1 and MSH2.

Tissue microarray construction and immunohistochemical staining

H&E-stained slides from the formalin-fixed, paraffin-embedded blocks were used to define the most morphologically representative and non-necrotic areas. Single tissue cores (2 mm in diameter) were sampled from each paraffin block and assembled into a recipient paraffin block using a tissue microarray instrument (AccuMax array, ISU ABXIS, Seoul, Korea). Tissue sections (4 μm thick) were cut from the tissue microarray blocks and deparaffinised; immunostaining was performed using a Bond Max automated immunostainer (Vision BioSystems, San Francisco, California, USA), and heat-induced epitope retrieval was performed with Bond Epitope Retrieval Solution. Endogenous peroxidase activity was blocked with 0.3% hydrogen peroxidase. Sections were stained with mouse monoclonal anti-SIRT1 antibody (sc-74465, Santa Cruz Biotechnology, California, USA) (dilution 1:100) as the primary antibody for 15 min at room temperature. The slides were then incubated with post primary reagent for 15 min at room temperature. Reactions were developed with BOND Polymer Refine Detection followed by colour development using 3,3′-diaminobenzidine tetrahydrochloride as chromogen.

Interpretation of SIRT1 expression

Immunoreactivity for SIRT1 was evaluated by two independent pathologists (KJ and SP) in a blinded fashion. Discordant cases were reviewed on a multi-headed microscope to achieve consensus. For each sample, a score was given for the percentage of positive cells: <5% of the cells (1 point), 6–35% of the cells (2 points), 36–70% of the cells (3 points), and >71% of the cells (4 points). Another score was given for the intensity of staining: negative staining (1 point), weak staining (2 points), moderate staining (3 points), and strong staining (4 points). A final score was then calculated by multiplying the above two scores. If the final score was ≥4, expression of SIRT1 was considered positive.

Statistical analysis

Statistical analysis was performed using SPSS V.13.0. The χ2 test was used to examine the associations between SIRT1 expression and various clinicopathological characteristics, including age, gender, tumour location, growth pattern, histological subtype, tumour grade, TNM category and AJCC stage. The relationships between mean SIRT1 expression and normal mucosa, adenoma, adenocarcinoma, and metastatic lesions were evaluated using the Mann–Whitney U test and paired sample t-test. The Kaplan–Meier method was used to calculate overall and disease-free survival curves. Univariable survival analysis with the log-rank test was used to compare differences between the survival rates of the patient subgroups. Multivariable survival analysis with Cox's proportional hazard regression model was used to evaluate independent prognostic factors. A difference of p<0.05 between groups was considered significant.

Results

SIRT1 expression in human colorectal tissue

SIRT1 expression was evaluated in normal mucosa, adenoma, adenocarcinoma and metastatic lesions. SIRT1 immunostaining was localised to the nucleus of epithelial and tumour cells. All 24 normal colonic mucosa samples (100%) were positive for SIRT1; 42 (80.8%) of 52 adenomatous polyps were positive for SIRT1 and 10 (19.2%) were negative. However, 208 (41.9%) of 497 colorectal adenocarcinomas were positive for SIRT1 and 289 (58.1%) were negative. Moreover, only 45 cases (35.7%) of 126 metastatic tissue samples were positive for SIRT1 and 81 cases (64.3%) were negative (table 1). Figure 1 shows representative sections of SIRT1 positive immunostaining. It is evident that SIRT1 expression gradually decreased during the successive transitions from normal mucosa to adenoma, adenocarcinoma, and metastatic lesion. Unfortunately, we had no complete batch of samples consisting of normal, adenoma, carcinoma, and metastatic tissues, which were obtained from the same patient. However, we found the same results in 22 paired normal and carcinoma tissues (p=0.003, paired sample t-test), and 72 paired carcinoma and metastatic tissues (p=0.003, paired sample t-test).

Table 1

Expression of SIRT1 in normal colonic mucosa, adenoma, adenocarcinoma and metastatic tissue

Figure 1

Representative sections of SIRT1 immunostaining. Normal colonic mucosa (A) and adenocarcinoma (B), showing positive nuclear staining.

Associations of SIRT1 expression and clinicopathological factors in colorectal adenocarcinomas

To assess the associations between SIRT1 expression and clinicopathological parameters, we evaluated SIRT1 expression in 497 primary colorectal adenocarcinomas. As shown in table 2, loss of SIRT1 expression was frequently observed in cases with microsatellite instability phenotypes, such as proximal tumour location, mucinous histology and defective MMR protein status. Loss of SIRT1 expression was associated with regional LN metastasis and higher AJCC stage. These results indicate that loss of SIRT1 expression is associated with colorectal tumour progression.

Table 2

Associations of SIRT1 expression with clinicopathological parameters in 497 colorectal adenocarcinomas

Associations of SIRT1 expression with overall survival and disease-free survival

We examined the impact of loss of SIRT1 expression on patient survival (figure 2). SIRT1 expression was significantly associated with overall survival (p=0.027, log-rank test). However, disease-free survival in univariable analysis failed to achieve significance. Univariable survival analysis was performed on the conventional prognostic factors and as a result, histological grade and AJCC stage were the factors considered statistically significant. Therefore multivariable analysis was performed on SIRT1 expression, histological grade and AJCC stage (table 3). However, SIRT1 expression had no independent prognostic influence on overall (p=0.107) or disease-free survival (p=0.339).

Figure 2

Kaplan–Meier curves for overall survival and disease-free survival for colorectal adenocarcinoma, according to SIRT1 expression.

Table 3

Multivariable Cox regression analysis of survival in colorectal adenocarcinoma

Discussion

Previous studies have demonstrated both oncogenic and tumour suppressor roles of SIRT1. We found that immunohistochemical expression of SIRT1 decreased during colorectal carcinogenesis. Moreover, loss of SIRT1 was associated with frequent regional LN metastasis and advanced tumour stage. Although these results do not constitute direct evidence of a role of SIRT1 in colorectal cancer, they support the evidence that SIRT1 is a tumour suppressor.

Previous studies have shown that SIRT1 overexpression leads to inactivation of tumour suppressor proteins and promotes cell growth, angiogenesis and resistance to chemotherapy.19 However, those studies were performed on cell lines and tumour tissues. Recently, a tumour suppressor function of SIRT1 has been consistently observed in genetically modified mouse models. SIRT1 transgenic mice had fewer spontaneous carcinomas and sarcomas, and were markedly protected from metabolic syndrome-driven liver carcinogenesis as a consequence of reduced DNA damage and inflammation.20 Ectopic expression of SIRT1 in a β-catenin-driven mouse model reduced the number of intestinal adenomas and their proliferation index by inhibiting the transcriptional activity and nuclear translocation of β-catenin.16 A tumour suppressor function of SIRT1 was also evident in a whole-body SIRT1-deficient mouse model in which Sirt1+/−;p53+/− mice developed tumours such as sarcomas, lymphomas, teratomas and carcinomas.18 These recent studies with genetically modified mice support a tumour suppressor role for SIRT1, and, up to now, no animal cancer model has provided definitive evidence for an oncogenic role.21

Jang and collaborators have published three consecutive papers evaluating the expression of SIRT1 by immunohistochemistry in several human cancer tissues.22–24 They suggest that SIRT1 overexpression is associated with poor clinical outcomes in diffuse large B-cell lymphoma and gastric cancer.22 23 However, the study of ovarian serous carcinoma showed that frequent SIRT1 expression was associated with better patient survival.24 In a study of colorectal cancer samples, positive SIRT1 expression was significantly associated with high tumour grade, but had no impact on patient prognosis.25 On the other hand, Kabra et al have shown that immunohistochemical expression is high in normal colonic mucosa and adenoma tissue, whereas adenocarcinomas show a heterogeneous profile with reduced expression in stage IV tumours as compared with lower stage tumours.26 In addition, SIRT1 knockdown by shRNA increases the size of the tumour xenograft by the HCT116 colon carcinoma cell line.26 It is possible that SIRT1 has different functions in different tissue contexts, leading to conflicting data on clinical outcomes. SIRT1 can deacetylate a variety of substrates and the list of its substrates is constantly expanding. Therefore, SIRT1 must be involved in a broad range of signalling pathways affecting carcinogenesis and tumour progression; that is possibly why the role of SIRT1 in cancer biology is so complex and controversial. The decisive factor may be which substrate is the primary target of SIRT1 deacetylation in a given type of cancer. Further studies with well designed cohorts of various cancer types are required to clearly identify the functions of SIRT1 in the different types of cancer.

The association observed between SIRT1 expression and the molecular changes in colorectal cancer suggested that SIRT1 overexpression is significantly more common in microsatellite instability-high (MSI-H) tumours than in microsatellite stable tumours.25 This is consistent with its role in inactivating p53 and in defective DNA damage repair.11 In contrast, our findings suggest that loss of SIRT1 expression is associated with a microsatellite instability phenotype leading to proximal tumour location, mucinous histology and defective MMR protein expression. Possible explanations for the different findings are a difference in the antibody used, different ethnicity of the subjects, and false-positive/negative results in immunohistochemistry.

In conclusion, SIRT1 expression decreases stepwise during colorectal carcinogenesis (normal–adenoma–adenocarcinoma–metastatic lesion) and as a function of adverse prognostic factors such as regional LN metastasis and tumour stage. Our data also suggest that loss of SIRT1 is associated with the MSI-high phenotype. These results are consistent with some previous findings, and in conflict with others. A detailed understanding of the molecular mechanisms of SIRT1 action in animal cancer models is needed to resolve these contradictions.

Take-home messages

  • SIRT1 is a histone deacetylase, and there is conflicting evidence about whether it acts as a tumour promoter or a suppressor.

  • In this study, SIRT1 expression is significantly reduced during colorectal carcinogenesis and its loss is associated with lymph node metastasis and advanced tumour stage.

Acknowledgments

We thank our laboratory colleagues for technical support.

References

Footnotes

  • Funding This work was supported by a grant from the research fund of Hanyang University (HY-2010-N) to KJ.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.