NAD-dependent deacetylase sirtuin-1 (SIRT1) is a class III histone deacetylase that positively regulates cancer-related pathways such as proliferation and stress resistance. SIRT1 has been shown to promote progression of colorectal cancer and is associated with cancer stemness, yet the precise mechanism between colorectal cancer stemness and SIRT1 remains to be further clarified. Here we report that SIRT1 signaling regulates colorectal cancer stemness by enhancing expression of CD24, a colorectal cancer stemness promoter. A novel miRNA, miR-1185-1, suppressed the expression of CD24 by targeting its 3′UTR (untranslated region) and could be inhibited by SIRT1 via histone deacetylation. Targeting SIRT1 by RNAi led to elevated H3 lysine 9 acetylation on the promoter region of miR-1185-1, which increased expression of miR-1185-1 and further repressed CD24 translation and colorectal cancer stemness. In a mouse xenograft model, overexpression of miR-1185-1 in colorectal cancer cells substantially reduced tumor growth. In addition, expression of miR-1185-1 was downregulated in human colorectal cancer tissues, whereas expression of CD24 was increased. In conclusion, this study not only demonstrates the essential roles of a SIRT1–miR-1185-1–CD24 axis in both colorectal cancer stemness properties and tumorigenesis but provides a potential therapeutic target for colorectal cancer treatment.

Significance:

A novel tumor suppressor miR-1185-1 is involved in molecular regulation of CD24- and SIRT1-related cancer stemness networks, marking it a potential therapeutic target in colorectal cancer.

Colorectal cancer is one of most common and lethal cancers in the world, with higher rates of cancer metastasis or recurrence at advanced stage (1). Cancer recurrence, metastasis, and drug resistance are considered to be associated with the properties of cancer stem cells (CSC; ref. 2). CSCs upregulate the expression of ATP-binding cassette transporters to increase the drug-resistance ability and escape from the chemotherapeutic treatment (3). Colorectal cancers with higher expressions of CD133, CD44, and CD24, so-called CSC markers, tend to have an aggressive behavior (4–6), and targeting these surface makers can reduce the cancer recurrence or metastasis (7–9). However, the regulatory network related to CSC stemness is still poorly understood.

CD24 is a small glycosylphosphatidylinositol (GPI)-anchored protein on the external side of cell membrane (10), and interacts with intracellular proteins via microdomain or lipid raft to mediate the signal transduction (11). CD24 is upregulated in some cancers (12), and correlated to the cancer stemness, invasion, metastasis, and the poor prognosis of patients with the bladder and breast cancer (13–16). Besides, CD24-Siglec10–mediated cancer cell–macrophage interaction signaling has been shown to become a target for cancer immunotherapy (17). CD24 activates several signaling pathways to promote colorectal cancer progression, such as ERK1/2, STAT3/VEGF, and Src-mediated signaling (18–20). CD24 is transcriptionally upregulated by TWIST2 in hepatocellular carcinoma, but downregulated by estrogen receptor and miR-34a in the ER+ breast cancer and colorectal cancer, respectively (19, 21). Nevertheless, the regulatory mechanism and the role of CD24 in colorectal cancer stemness associated networks still need to be explored.

Sirtuin 1 (SIRT1) belongs to the class III histone deacetylase (HDAC), consumes NAD+ to mediate the deacetylation of H3 lysine 9 (H3K9; ref. 22), and indirectly promotes the production of trimethyl H3K9 (23). SIRT1 has nonhistone deacetylating targets of different pathways in cancers and acts as a regulatory hub to affect the autophagy, stress resistance, proliferation, tumorigenesis, and inflammation signaling pathways through the targets as ATG7, HIF, PI3K/AKT, TGFβ, WNT, NFκB, and so on (24). Previous reports have revealed that HDAC3 is upregulated in colorectal cancer and loss of global acetylated H3K9 (H3K9ac) and H4K16 is a common event in colorectal cancer (25, 26). The patients with higher SIRT1-expressing colorectal cancer have been found to be correlated with poor prognosis (27). Furthermore, SIRT1 transcriptionally represses the tumor suppressor genes through the epigenetic regulation. For example, E-cadherin is usually inactivated epigenetically by SIRT1 during tumor metastasis and progression of colorectal cancer (28). SIRT1 also promotes colorectal cancer stemness (27, 29). The above mentioned indicates the critical influence of SIRT1 on colorectal cancer. However, how SIRT1 affects colorectal cancer aggressiveness and stemness still need to be further investigated.

In this study, we first showed that SIRT1 promoted CD24 expression to increase the stemness and aggressiveness of colorectal cancer cells through repressing the expression of a novel tumor suppressor miR-1185-1 by histone deacetylation, and also demonstrated that miR-1185-1 could reduce colorectal cancer tumorigenesis in vivo. Therefore, SIRT1–miR-1185-1–CD24 axis might be considered as one of the colorectal cancer stemness signaling pathways to modulate the cancer stemness, and provide the molecular evidence for cancer target therapy.

Cell culture

Two human colorectal cancer cell lines, HT29 and DLD-1 (BBRC) were cultured in RPMI1640 medium (Gibco) with 10% FBS (GeneDirex) and 1% penicillin-streptomycin-glutamine (PSG; Gibco). CCD841 CoN (ATCC) was cultured in EMEM (Gibco) with 10% FBS. HEK293T (ATCC) was cultured in DMEM (Gibco) with 10% FBS and 1% PSG. Cells were grown at 37°C in humid incubator containing 5% CO2. All experiments were performed on all cell lines within 12 passages. All cell lines were verified to be Mycoplasma-free using Myco-Blue Mycoplasma Detector (Zgenebio).

Lentivirus knockdown and overexpression system

pRSV-Rev, pMD2.G, and pMDLg/pRRE were used to generate lentivirus as package vector combined with pCDH-CMV-MCS-EF1α-copGFP (SBI) or pLKO vector system. pCMV Δ8.91 and pMD.G were used as the package vector combined with pLAS3w.Pneo system. These packaging plasmids were gifted from Didier Trono (Addgene plasmids, pMD2.G # 12259, pRSV-Rev # 12253, pMDLg/pRRE # 12251; ref. 30). The pLKO vector (Sinica RNAi core) was used to construct knockdown and miRNA overexpression (miROE) cell line. The knockdown plasmids chosen for SIRT1 or CD24 knockdown were TRCN0000018979, TRCN0000229630, TRCN0000245103, TRCN0000245104 (Sinica RNAi core). The pCDH-CMV-MCS-EF1α-copGFP (SBI) vector was used to construct SIRT1 overexpression cell line. The pLAS3w.Pneo vector (Sinica RNAi core) was used to overexpress CD24 in miROE cell line. pLKO1-shLuc is used as the control for miR-1185-1 overexpression (miROE). pLAS3w-empty vector is used as mock control for pLAS3w-CD24.We transduced these two vectors (pLAS3w-empty vector and pLAS3w-CD24) into miR-1185-1 overexpressed cells (MOCK-miROE vs. CD24-miROE) to perform rescue experiment. Lentivirus plasmid and package plasmids were transfected to HEK293T. Virus was harvested by centrifugation, and then introduced into HT29 or DLD-1 for 24 hours, followed by selection with puromycin for 48 hours.

RNA extraction and qRT-PCR

The mRNA was extracted by TriPure isolation reagent (Roche Life Science). The mRNA was reversely transcribed into cDNA with High-Capacity cDNA Reverse Transcription Kit (Applied Biosystem). The miRNA was extracted by miRNeasy Mini Kit (QIAGEN). The miRNA was reversely transcribed into cDNA with HiSpec system of miScript II RT Kit (QIAGEN). The qRT-PCR was performed using SYBR Green system (Bio-Rad), and the analysis was performed by Bio-Rad CFX96 Real-Time PCR Machine. Each combination of cDNA and primers was assayed in triplicate and normalized to internal control. GAPDH and β-actin served as mRNA internal control. U6 served as miRNA internal control. The primer used for qPCR were listed in Supplementary Table S1.

Western blot anaysis

The cells were lysed by RIPA lysis buffer supplemented with cOmplete Cock Tail Protease Inhibitor (Roche). The cell lysates were denatured with protein sample buffer at 100°C for 10 minutes. Protein samples were analyzed by SDS-PAGE, and then transferred onto polyvinylidene difluoride membrane (Roche). The membrane was blocked with 4% skim milk and probed with primary antibodies. The membrane was rinsed with Luminated Classico Western HRP Substrate (Millipore), and analyzed by Biospectrum MultiSpectral Imaging System (UVP BioSpectrum 800). The antibodies and their dilution condition were listed in Supplementary Table S2.

IHC staining

IHC was performed using SIRT1, CD24 antibodies (Supplementary Table S2) at the dilution of 1:200 and 1:100, respectively, on the human colorectal cancer tissue and xenograft tissue sections. Briefly, unstained tissue sections were deparaffinized and rehydrated. Antigen Retrieval Solution (sodium citrate buffer 10 mmol/L, pH 6.0) was used for antigen retrieval. Endogenous peroxidase was blocked by 3% hydrogen peroxide solution or peroxidase blocker (bio SB). Tissue sections were incubated with primary antibody overnight at 4°C. Ready-to-use solution with polydetector plus link anti-mouse/rabbit and anti-mouse/rabbit horseradish peroxidase (HRP)-labeled (Bio SB) or HRP-labeled anti-mouse antibody (1:200, Jackson ImmunoResearch) was applied at room temperature for 15–30 minutes. Peroxidase was detected by 3,3′-diaminobenzidine (DAB) chromogen solution (Bio SB). The IHC slides were counterstained by hematoxylin.

Establishment of 5-FuR colon cancer cells

To obtain stemness population with drug resistance, we treated DLD-1 cells with 5-fluorouracil (5-FU) to screen and generate the drug resistance population. 5-FuR population was established according to the paper of Izumiya and colleagues (31). Briefly, the DLD-1 cells were treated serially with different concentration of 5-FU (from low to high concentration; 0.1, 0.5, 1 μg/mL) for 30 days totally (each concentration for 10 days). After this screening process, 5-FU–resistant DLD-1 cells (5-FuR) were obtained.

Sphere-forming and extreme limiting dilution assay

In sphere-forming assay, colorectal cancer cells were cultured in sphere-forming medium (HT29: RPMI medium with 3% FBS; DLD-1: RPMI medium supplemented with 2% B27 supplement (Gibco), 20 ng/mL EGF (PeproTech), and 10 ng/mL basic FGF (bFGF; PeproTech) on 96-well ELISA plates (Greiner). After cultured for 9 days, the spheres were stained with 5 μg/mL Hoechst 33342 (Life Technologies), and then observed and analyzed by IN Cell Analyzer 2000 (GE Healthcare Life Sciences). In extreme limiting dilution assay (ELDA), cells were seeded at different densities in 96-well agarose-coated ELISA plates (Greiner) with sphere culture medium. These cells were cultured for 9 days. Spheres were detected and analyzed with In Cell 2000 Analyzer (GE Healthcare). Cancer stem cell frequency was calculated by calculating slope of cell number and logarithm of nonresponding fraction by web tool at http://bioinf.wehi.edu.au/software/elda/; P < 0.05 was considered to be significantly different (32, 33).

In vitro relapse assay

The in vitro relapse assay was used to evaluate the relapse capacity of cancer cells, which was designed to simulate the tumor recurrence after chemotherapeutic treatment in the clinical practice. For colony formation assay, colorectal cancer cells were seeded as 250 cells per well into 6-well plates with culture medium. After 12 hours, 5-Fu was added to the final concentration of 0.5 μg/mL for HT29 and 1 μg/mL for DLD-1 for 48 hours. After 5-Fu treatment, the remaining living cells were cultured in the culture medium without 5-Fu for 6 days until colonies formed (Fig. 2C, left). The colonies were fixed with methanol, and then stained with 0.5% crystal violet staining solution. Colony ratio was defined by the ratio of colony number of 5-Fu group to nontreated group. The relative colony ratio represented as the colony ratio of SIRT1-KD or SIRT1-OE divided by that of shLuc group or MOCK group, respectively.

Transwell migration assay

The cells (1 × 105 cells/well) were plated on the top of the filter membrane of the Transwell Cell Culture Insert (Falcon, 8 μm pore size) with serum free medium, and cultured medium with 10% FBS was added into the lower chamber in a 24-well plate. Following 12 hours incubation, the transwell insert was placed into methanol for 10 minutes to allow cell fixation. The nonmigrated cells were gently removed from the top of the membrane with a cotton tipped applicator, and the migrated cells were counted after 0.5% crystal violet staining.

Chromatin immunoprecipitation

Colorectal cancer cells were fixed with 1% paraformaldehyde at room temperature, and then collected by scraping with Farnham lysis buffer. Nuclei were then sonicated with Covaris S2 (Covaris) to break the DNA into 500 bp fragment. Fragmented soluble chromatin was immunoprecipitated with antibodies and Protein A/G PLUSAgarose (Santa Cruz Biotechnology). After incubated with proteinase K and RNase A, DNA was harvested by QIAquick PCR purification kit and the concentration was determined by Qubit (Life Technology). Input samples were prepared by incubation at 67°C overnight to perform reverse cross-linking. The information of antibodies was listed in Supplementary Table S2.

Flow cytometry analysis

Cells were stained with anti-CD24 primary antibody (BD Pharmingen) at 4°C for 1 hour. After washing with wash buffer (PBS containing 3% FBS) three times, cells were stained with anti-mouse IgG-conjugated Alexa Fluor 647 (Invitrogen), then incubated at room temperature for 30 minutes in dark place. Cells were washed with wash buffer twice, followed by passing through 40 μm Cell Strainer (Falcon) before analysis with BD FACSCanto 2 Cell Flow cytometry.

Side population assay

DLD-1 and HT29 cells were collected and resuspended with culture medium at the concentration as 5 × 106 cell/mL. The verapamil (Sigma) was added into cells to final concentration of 50 μmol/L to inhibit ABCG2 as the negative control. The cells were incubated at 37°C for 15 minutes in dark place, and Dye Cycle Violet (Invitrogen) was added to 5 μmol/L for 90 minutes. After harvesting by centrifugation, cells were resuspended in PBS containing 3% FBS, and then followed by passing through 40 μm Cell Strainer (Falcon) before analysis and sorted with BD FACSAria III Cell Flow sorter.

Dual-luciferase reporter assay

The psiCHECK2 vector (Promega) was used to evaluate interaction between 3′UTR (untranslated region) of CD24 and miR-1185-1. The 3′UTR of CD24 was cloned from HT29 cDNA and ligated to 3′end of Renilla luciferase gene in psiCHECK2 plasmid. The firefly luciferase in the same plasmid was used as internal control. The psiCHECK2 was co-transfected with shLacZ as negative control or miR-1185-1 vector into HEK293T cells for reporter assay. After 48 hours posttransfection, Dual-Luciferase Reporter Assay System Kit (Promega) was used to measure two luciferase activities simultaneously. Each combination of transfections was assayed for eight times and normalized to internal control.

In vivo xenograft model

The animal work was performed in accordance with a protocol approved by Institutional Animal Care and Use Committee, NTU (NTU 106-EL-00211). 6 weeks old Nu-Foxn1nu mice (BioLASCO) were used. HT29 and DLD-1 cells were counted and resuspended in PBS at a concentration of 5 × 106 cells/mL and 2.5 × 107 cells/mL, respectively. 0.5 mL Avertin (1.25% 2,2,2-tribromomethanol, 2.5% 2-methyl-2-butanol dissolved in ddH2O) was intraperitoneally injected into a 25 g mouse as anesthetic agent. 0.1 mL of the cell suspension was injected subcutaneously with 29G needle (BD, Ultra-Fine Insulin Syringes). The tumor diameter was measured every 2 days after injection. After 20 days, the mice were sacrificed and each tumor weight was measured.

Statistical analysis

Data were shown as mean ± SE. Statistical significance of results was evaluated using one- or two-way ANOVA and further analyzed by Tukey multiple comparison test. χ2 test and linear regression were used to calculate statistical significance of stem cell frequency, and error bar showed 95% confidence interval. Combined ANOVA was used to calculate group statistical significance of paired repeated measurements such as tumor size growth curve. Difference of tumor weight was evaluated with Wilcoxon signed-rank test as nonparametric statistics. The P value less than 0.05 were considered statistically significant.

Human colorectal cancer specimens

Deidentified specimens of colorectal cancer tissues, matched-normal colon tissues and unstained tissue sections were provided from Taipei Medical University Joint biobank, which obtained written informed consent from all patients. The samples were harvested by the pathologists in Taipei Medical University Hospitals. This study was conducted in accordance with the Belmont Report ethical principles and under TMU-Joint institutional IRB approval.

Bioinformatic analysis by the UCSC Xena platform

Bioinformatic analyses were performed on the basis of The Cancer Genome Atlas (TCGA) TARGET GTEx cohort by the UCSC Xena platform (34). Only cases of colorectal samples were filtered and included for further analysis of the RNA expression of SIRT1 and CD24. Expression heatmaps of genes were generated online. The expression levels of SIRT1 or CD24 between tumor and normal tissue samples were compared with statistical analysis. The correlation between the expression of SIRT1 and CD24 in the colorectal samples was also analyzed. UCSC Xena platform website: https://xena.ucsc.edu/

SIRT1 was upregulated in colorectal cancer cells with higher stemness

We used sphere-forming assay to enrich CSCs with higher self-renewal ability (35). Compared with the cancer cells with two-dimensional (2D) culture, those sphere cells exhibited higher expression of colon CSC markers, such as CD24, CD133, BMI1, and β-CATENIN (Fig. 1A and B). SIRT1 and CD24 were upregulated on mRNA and protein levels in the sphere cells of HT29 and DLD-1 (Fig. 1A and B). Compared with colorectal cancer cells, CCD 841 CoN, a normal colon cell line, showed undetectable or much lower expressions of SIRT1, CD24, β-CATENIN, and CD133 (Supplementary Fig. S1A). CSC properties can be observed in chemoresistant cells. Here, 5-Fu–resistant population of DLD-1, designated as 5-FuR, was obtained by serially screening cells in different concentrations of 5-Fu for 30 days totally (0.1, 0.5, and 1 μg/mL for each 10 days screening; ref. 31). The protein and mRNA expressions of SIRT1 were upregulated in the 5-FuR cells that increased the expressions of CD24 and multidrug transporter ABCG2 (Fig. 1C; Supplementary Fig. S1B). The sphere-forming rate of 5-FuR group was higher than that of nontreated control group (WT; Fig. 1D). The side population of cancer cells is considered to have stemness (2, 36). The mRNA expressions of SIRT1, CD24, β-CATENIN, and ABCG2 were upregulated in side population–positive cells of HT29 and DLD-1 cells (Supplementary Fig. S1C and S1D). On the basis of above findings, SIRT1 was upregulated in sphere cells, 5-FuR cells, and side population–positive cells that possessed CSC property with self-renewal ability and higher drug resistance.

Figure 1.

SIRT1 was upregulated in CSC-like cells. A, The mRNA expressions of SIRT1 and colorectal cancer stem cell markers in colon cancer cells with 2D and sphere cultures were measured by qRT-PCR. Compared with the 2D culture cells, the sphere cells had much more increased expressions of CD24, CD133, BMI1, CTNNB, and SIRT1. Results were normalized to GAPDH and β-ACTIN. n = 3. B, The protein expressions of SIRT1, CD24, NANOG, BMI1, CD133, β-CATENIN, CD44, and ALDH1 in sphere cells were higher than those in 2D culture cells by Western blot analysis. C, The protein expressions of SIRT1, CD24, ABCG2, BMI1, CD133, and β-CATENIN in nontreated and 5-FuR DLD-1 cells were measured by Western blot analysis. The 5-FuR cells had higher expression levels of SIRT1, BMI1, CD133, ABCG2, and CD24 compared with the nontreated group. D, Self-renewal ability of 5-Fu–resistant DLD-1 (designated as 5-FuR) was much more enhanced than that of nontreated cells, analyzed by sphere-forming assay. n = 60; *, P < 0.05; **, P < 0.01; ***, P < 0.005.

Figure 1.

SIRT1 was upregulated in CSC-like cells. A, The mRNA expressions of SIRT1 and colorectal cancer stem cell markers in colon cancer cells with 2D and sphere cultures were measured by qRT-PCR. Compared with the 2D culture cells, the sphere cells had much more increased expressions of CD24, CD133, BMI1, CTNNB, and SIRT1. Results were normalized to GAPDH and β-ACTIN. n = 3. B, The protein expressions of SIRT1, CD24, NANOG, BMI1, CD133, β-CATENIN, CD44, and ALDH1 in sphere cells were higher than those in 2D culture cells by Western blot analysis. C, The protein expressions of SIRT1, CD24, ABCG2, BMI1, CD133, and β-CATENIN in nontreated and 5-FuR DLD-1 cells were measured by Western blot analysis. The 5-FuR cells had higher expression levels of SIRT1, BMI1, CD133, ABCG2, and CD24 compared with the nontreated group. D, Self-renewal ability of 5-Fu–resistant DLD-1 (designated as 5-FuR) was much more enhanced than that of nontreated cells, analyzed by sphere-forming assay. n = 60; *, P < 0.05; **, P < 0.01; ***, P < 0.005.

Close modal

SIRT1 enhanced CSC properties

To investigate the function of SIRT1 in colorectal cancer, SIRT1-knockdown and SIRT1-overexpressed stable clones of HT29 and DLD-1 were constructed, and the efficiency was validated by Western blot analysis (Fig. 2A). The ELDA was performed to evaluate the self-renewal ability and CSC frequency. As shown in Fig. 2B, the CSC frequency was significantly decreased in SIRT1-knockdown groups but increased in overexpressed group. Besides, the expressions of drug-resistance gene, ABCG2, and CSC-related genes, such as BMI1 and CD133 substantially and positively correlated with the expression of SIRT1 (Fig. 2A). In addition, in vitro relapse assay that proceeded as 5-Fu treatment in the beginning of colony formation assay was performed to examine the relapse capacity of cancer cells (Fig. 2C, left). This assay was designed to simulate the tumor recurrence after the chemotherapeutic treatment in the clinical practice. In both HT29 and DLD-1 cells, SIRT1-knockdown groups were found to have lower colony ratios, while overexpression groups significantly had higher colony ratios, compared with the control groups, respectively (Fig. 2C). These data indicated SIRT1 substantially enhanced the tumor relapse ability of colorectal cancer cells. The transwell migration assay demonstrated overexpression of SIRT1 enhanced the migration ability of colon cancer cells, while knockdown of SIRT1 impaired the migration ability (Fig. 2D). Taken together, SIRT1 could enhance the CSC frequency, tumor relapse ability, and migration ability of colorectal cancer cells.

Figure 2.

SIRT1 promoted CSC properties. A, The SIRT1-knockdown and overexpression systems were verified by Western blot analysis. shLuc was served as knockdown control. pCDH-infected cells were served as MOCK control group. SIRT1-OE denoted the SIRT1-overexpressed colorectal cancer cell group. The protein expression level of CD24, BMI1, CD133, and β-CATENIN was upregulated in SIRT1-overexpressed colorectal cancer cell group and downregulated in SIRT1-knockdown colorectal cancer cell group. B, CSC frequency of SIRT1-KD and SIRT1-OE colon cancer cells was measured by ELDA. SIRT1 promoted CSC frequency. C, Left, the experimental scheme of in vitro relapse assay is shown. This assay was designed to evaluate the tumor relapse capacity. Right, the relative colony ratios of SIRT1-KD and SIRT1-OE colon cancer cells are shown. The result revealed that SIRT1-KD group had lower relative colony ratio, while SIRT1-OE group had the higher ratio, compared with the corresponding control groups. n = 3. D, The migration ability of SIRT1-KD and SIRT1-OE DLD-1 cells was measured. Migration rate was normalized to shLuc or MOCK control group. n = 3; *, compared with shLuc as the control; #, compared with MOCK as the control; NS, nonsignificant; * or #, P < 0.05; ** or ##, P < 0.01; *** or ###, P < 0.005.

Figure 2.

SIRT1 promoted CSC properties. A, The SIRT1-knockdown and overexpression systems were verified by Western blot analysis. shLuc was served as knockdown control. pCDH-infected cells were served as MOCK control group. SIRT1-OE denoted the SIRT1-overexpressed colorectal cancer cell group. The protein expression level of CD24, BMI1, CD133, and β-CATENIN was upregulated in SIRT1-overexpressed colorectal cancer cell group and downregulated in SIRT1-knockdown colorectal cancer cell group. B, CSC frequency of SIRT1-KD and SIRT1-OE colon cancer cells was measured by ELDA. SIRT1 promoted CSC frequency. C, Left, the experimental scheme of in vitro relapse assay is shown. This assay was designed to evaluate the tumor relapse capacity. Right, the relative colony ratios of SIRT1-KD and SIRT1-OE colon cancer cells are shown. The result revealed that SIRT1-KD group had lower relative colony ratio, while SIRT1-OE group had the higher ratio, compared with the corresponding control groups. n = 3. D, The migration ability of SIRT1-KD and SIRT1-OE DLD-1 cells was measured. Migration rate was normalized to shLuc or MOCK control group. n = 3; *, compared with shLuc as the control; #, compared with MOCK as the control; NS, nonsignificant; * or #, P < 0.05; ** or ##, P < 0.01; *** or ###, P < 0.005.

Close modal

CD24 was required to maintain the CSC properties and was regulated by SIRT1

To investigate the influence of CD24 on CSC properties, CD24-knockdown cell lines were constructed, and the efficiency was validated by Western (Supplementary Fig. S2A). CD24-knockdown cancer cells substantially reduced the CSC frequency, tumor relapse ability, and migration ability (Fig. 3A–C). In addition, we found that knockdown of SIRT1 in colon cancer cells resulted in the downregulation of CD24, while overexpression of SIRT1 caused CD24 upregulation (Fig. 2A; Supplementary Fig. S2B–S2D). To see the expression pattern of SIRT1 and CD24 in patients, we performed the IHC study on three human colorectal cancer specimens. The result revealed that both proteins seemed to be expressed in three colorectal cancer tumors (Fig. 3D; Supplementary Fig. S2E). We further used UCSC Xena platform to analyze the human colorectal samples of TCGA GTEx database (n = 686). On the basis of the transcriptome analysis results, the expressions of CD24 and SIRT1 were upregulated in colorectal cancer samples (Fig. 3E). The result also revealed a positive correlation between the two genes (Fig. 3F). Taken together, CD24 might be a downstream target of SIRT1 and seemed to be correlated with SIRT1's expression in colorectal cancer cells and the transcriptome database. Nevertheless, how to regulate CD24 by SIRT1 needs to be further clarified.

Figure 3.

CD24 was required to maintain the CSC properties and regulated by SIRT1. A, CSC frequency of CD24-knockdown HT29 and DLD-1 cells was measured by ELDA. B, The relative colony ratios of CD24-knockdown HT29 and DLD-1 cells were lower than those of the shLuc groups. n = 3. C, The migration ability of CD24-knockdown DLD-1 cells was impaired. n = 3. D, Human colorectal cancer tissue sections were immunostained with anti-SIRT1 or anti-CD24 antibody. The pictures of the representative case (Patient case No.1) that expressed both SIRT1 and CD24 are shown. E, TCGA GTEx transcriptome analysis of CD24 and SIRT1 expression in human colorectal tissues (n = 686) by UCSC Xena platform. The expressions of CD24 and SIRT1 were upregulated in colorectal cancer samples, compared with normal colorectal samples. The data were statistically analyzed by Welch t test. Compared with normal samples, the P value of CD24 and SIRT1 are 1.773 × 10−73 and 8.047 × 10−27, respectively. F, The correlation of CD24 and SIRT1 expression in colorectal samples (n = 686) of TCGA GTEx database was measured by UCSC Xena platform. The correlation was significantly positive (Pearson r = 0.3914 and P = 1.55 × 10−26). *, P < 0.05; **, P < 0.01; ***, P < 0.005.

Figure 3.

CD24 was required to maintain the CSC properties and regulated by SIRT1. A, CSC frequency of CD24-knockdown HT29 and DLD-1 cells was measured by ELDA. B, The relative colony ratios of CD24-knockdown HT29 and DLD-1 cells were lower than those of the shLuc groups. n = 3. C, The migration ability of CD24-knockdown DLD-1 cells was impaired. n = 3. D, Human colorectal cancer tissue sections were immunostained with anti-SIRT1 or anti-CD24 antibody. The pictures of the representative case (Patient case No.1) that expressed both SIRT1 and CD24 are shown. E, TCGA GTEx transcriptome analysis of CD24 and SIRT1 expression in human colorectal tissues (n = 686) by UCSC Xena platform. The expressions of CD24 and SIRT1 were upregulated in colorectal cancer samples, compared with normal colorectal samples. The data were statistically analyzed by Welch t test. Compared with normal samples, the P value of CD24 and SIRT1 are 1.773 × 10−73 and 8.047 × 10−27, respectively. F, The correlation of CD24 and SIRT1 expression in colorectal samples (n = 686) of TCGA GTEx database was measured by UCSC Xena platform. The correlation was significantly positive (Pearson r = 0.3914 and P = 1.55 × 10−26). *, P < 0.05; **, P < 0.01; ***, P < 0.005.

Close modal

miR-1185-1 was repressed by SIRT1 via epigenetic regulation

Because SIRT1 is known as a HDAC to repress the transcriptional activity, one possibility of how CD24 is regulated by SIRT1 is through miRNA. Therefore, we did cross-comparison of miRNA-targeting prediction by miRDB (target score greater than 75) and deep-sequencing data from miRBase (reads number greater than 100) to see whether there is a CD24-targeted miRNA repressed by SIRT1. The result revealed that two miRNAs, miR-1185-1 (miR-1185-1-3p) and miR-1185-2 (miR-1185-2-3p) were fitted our screening criteria (Fig. 4A; Supplementary Table S2). Further examination was performed to investigate whether these two miRNAs are the targets of SIRT1. As shown in Fig. 4B, the expression of miR-1185-1 was increased in SIRT1-knockdown cells and decreased in SIRT1-overexpressed cells, which was negatively correlated with the expression of SIRT1. miR-1185-2 was not further investigated because of the inconsistent results in the SIRT1-knockdown experiment. To examine whether miR-1185-1 was transcriptionally repressed by SIRT1, chromatin immunoprecipitation coupled with qPCR (ChIP-qPCR) was performed to detect the histone modification on the promoter region of miR-1185-1. Compared with the IgG control group, SIRT1 was shown to recruit on the sequences near the transcription start site (TSS) of miR-1185-1 (Fig. 4C). While the expression of SIRT1 was suppressed by short hairpin RNA, the H3K9ac, a substrate of SIRT1, was enriched around the TSS of miR-1185-1 (Fig. 4D). Therefore, SIRT1 could decrease H3K9ac on the promoter of miR-1185-1 and thus cause chromatin remodeling to repress miR-1185-1′s expression.

Figure 4.

SIRT1 suppressed the expression of miR-1185-1 via H3K9 deacetylation. A, The Venn diagram illustrated the screening strategy of miRNA target prediction. B, The expression of miR-1185-1 was upregulated in SIRT1-KD and downregulated in SIRT1-OE colon cancer cells. Results were normalized to RNA U6. n = 3. C, SIRT1 was recruited on the promoter region of miR-1185-1 by ChIP-qPCR analysis. Results were normalized with input as the internal control, and the IP SIRT1 group was compared with IgG group as the negative control. n = 4. D, H3K9ac was enriched around the TSS of miR-1185-1 in SIRT1-KD HT29 and DLD-1 cells by ChIP-qPCR analysis. Results were normalized using input as the internal control, and the IP H3K9ac of SIRT1-KD group was compared with IP H3K9ac of shLuc group. n = 4; NS, nonsignificant; *, P < 0.05; **, P < 0.01; ***, P < 0.005.

Figure 4.

SIRT1 suppressed the expression of miR-1185-1 via H3K9 deacetylation. A, The Venn diagram illustrated the screening strategy of miRNA target prediction. B, The expression of miR-1185-1 was upregulated in SIRT1-KD and downregulated in SIRT1-OE colon cancer cells. Results were normalized to RNA U6. n = 3. C, SIRT1 was recruited on the promoter region of miR-1185-1 by ChIP-qPCR analysis. Results were normalized with input as the internal control, and the IP SIRT1 group was compared with IgG group as the negative control. n = 4. D, H3K9ac was enriched around the TSS of miR-1185-1 in SIRT1-KD HT29 and DLD-1 cells by ChIP-qPCR analysis. Results were normalized using input as the internal control, and the IP H3K9ac of SIRT1-KD group was compared with IP H3K9ac of shLuc group. n = 4; NS, nonsignificant; *, P < 0.05; **, P < 0.01; ***, P < 0.005.

Close modal

The miR-1185-1 targeted 3′UTR of CD24

On the basis of the results of bioinformatics prediction, putative target genes of miR-1185-1 participated in different cancer-associated biological processes and pathways mapping in Kyoto Encyclopedia of Genes and Genomes (KEGG; Supplementary Table S3). To further analyze whether these target genes were regulated by miR-1185-1, we overexpressed miR-1185-1 in HT29 and DLD-1, which denoted miROE (Fig. 5A). The result revealed that CD24 was significantly decreased in miROE cells (Fig. 5B; Supplementary Fig. S2F). However, the expression changes of some other putative target genes were not shown as expected and not consistent between two colon cancer cells (Supplementary Fig. S3A). The bioinformatics prediction showed miR-1185-1 might target to 3′untranslated region (3′UTR) of CD24 mRNA with the target sequence as 5′-CTGTATA-3′ (Fig. 5C). The dual-luciferase reporter system was performed to validate whether miR-1185-1 could repress the translation of CD24. Compared with the control group, miR-1185-1 could significantly inhibit the Renilla luciferase activity in HEK293T cell line (Fig. 5D). Furthermore, when the predicted binding site was mutated, the repressive efficacy became insignificant (Fig. 5D).

Figure 5.

miR-1185-1 targeted 3′UTR of CD24. A, miR-1185-1 expression in miROE HT29 and DLD-1 was measured to validate the overexpression efficiency. Results were normalized to RNA U6. n = 3. B, CD24 was downregulated in miROE HT29 and DLD-1 cells. The expression of CD24 in miROE HT29 and DLD-1 cells was measured by flow cytometry analysis. C, The sequence of miR-1185-1 and the illustration of dual luciferase reporter plasmid with WT or mutant sequences of CD24 3′UTR are shown. D, Luciferase analysis of different reporters cotransfected with shLacZ as the control or the overexpressed miR-1185-1 was performed. n = 8; NS, nonsignificant; ***, P < 0.005.

Figure 5.

miR-1185-1 targeted 3′UTR of CD24. A, miR-1185-1 expression in miROE HT29 and DLD-1 was measured to validate the overexpression efficiency. Results were normalized to RNA U6. n = 3. B, CD24 was downregulated in miROE HT29 and DLD-1 cells. The expression of CD24 in miROE HT29 and DLD-1 cells was measured by flow cytometry analysis. C, The sequence of miR-1185-1 and the illustration of dual luciferase reporter plasmid with WT or mutant sequences of CD24 3′UTR are shown. D, Luciferase analysis of different reporters cotransfected with shLacZ as the control or the overexpressed miR-1185-1 was performed. n = 8; NS, nonsignificant; ***, P < 0.005.

Close modal

The miR-1185-1 reduced CSC properties and tumorigenesis through repressing CD24

To realize the relationship between the biological functions of miR-1185-1 and CD24, rescue experiment was performed (Fig. 6A; Supplementary Fig. S3B). The results demonstrated that the CSC properties such as CSC frequency, tumor relapse ability, and migration ability were all decreased in the miROE groups, and all were partially rescued in the CD24-overexpressed groups (Fig. 6B–D). Besides, the tumor suppressor potential of miR-1185-1 was investigated using in vivo xenograft model. The tumor size, weight, and the expression of CD24 in miROE groups were reduced compared with the shLuc control groups (Fig. 7A and B; Supplementary Fig. S4A). The data suggested miR-1185-1 suppressed the tumorigenesis of colorectal cancer. To understand the expression pattern between miR-1185-1 and CD24 in patients with colorectal cancer, we measured the gene expression by qPCR. Compared with the normal colonic tissues, miR-1185-1 was dramatically decreased while CD24 was increased in the colorectal cancer tumors of 3 patients (Fig. 7C; Supplementary Fig. S4B).

Figure 6.

miR-1185-1 suppressed colorectal cancer stem cell properties through repressing CD24. A, Flow cytometry analysis of CD24 expression was performed to validate the overexpression efficiency in miROE HT29 or DLD-1 cells. Fluorescence intensities of each cell are plotted in histogram. B, The CSC frequency of miROE was lower than that of the control group and could be rescued by CD24 overexpression in both HT29 and DLD-1 cells. C, The relative colony ratio of miROE was lower than that of the control group and could be rescued by CD24 overexpression in both HT29 and DLD-1 cells. n = 3. D, The migration ability of miROE was impaired and could be rescued by CD24 overexpression in DLD-1 cells. n = 3; *, compared with shLuc group as the control; #, compared with MOCK group as the control; NS, nonsignificant; * or #, P < 0.05; ** or ##, P < 0.01; *** or ###, P < 0.005.

Figure 6.

miR-1185-1 suppressed colorectal cancer stem cell properties through repressing CD24. A, Flow cytometry analysis of CD24 expression was performed to validate the overexpression efficiency in miROE HT29 or DLD-1 cells. Fluorescence intensities of each cell are plotted in histogram. B, The CSC frequency of miROE was lower than that of the control group and could be rescued by CD24 overexpression in both HT29 and DLD-1 cells. C, The relative colony ratio of miROE was lower than that of the control group and could be rescued by CD24 overexpression in both HT29 and DLD-1 cells. n = 3. D, The migration ability of miROE was impaired and could be rescued by CD24 overexpression in DLD-1 cells. n = 3; *, compared with shLuc group as the control; #, compared with MOCK group as the control; NS, nonsignificant; * or #, P < 0.05; ** or ##, P < 0.01; *** or ###, P < 0.005.

Close modal
Figure 7.

miR-1185-1 suppressed the tumorigenesis in mice xenograft model and was downregulated in human colorectal cancer tissues. A, The tumor sizes of shLuc (open circle) and miROE (solid circle) groups upon transplantation into Nu/Nu mice were measured. n = 5. P value was calculated by combined ANOVA. In this experiment, two group cells (miR-overexpression vs. shLuc control) were injected into bilateral sides of the trunk, respectively, in the same Nu/Nu mice. One side is control (shLuc, --, open circle) group and the other side is miR-1185-1 overexpressed group (miROE, -•-, solid circle). The diameters of tumors between miROE group and shLuc group were measured and compared in the same mouse every 2 days since 8th day after injection. One color with circle and line represents one mouse. Because of different tumorigenesis ability among different mice, the diameter changes of the tumors (miROE vs. shLuc group) were compared in the same mouse, represented by the same color. B, The gross pictures of tumors that were removed from mice after transplantation 20 days. The length of lattice was equal to 1.5 cm. Right, tumor weights of shLuc and miROE HT29 or DLD-1 cells. P value was calculated by Wilcoxon signed-rank test. C, The expressions of miR-1185-1, CD24, and SIRT1 in normal colonic and colorectal cancer tumor tissues of 3 patients were measured by qPCR. Results were normalized to RNA U6 for miR-1185-1 and to GAPDH and β-ACTIN for CD24 and SIRT1. n = 3; *, compared with the normal part group; *, P < 0.05; **, P < 0.01; ***, P < 0.005. D, Scheme of SIRT1–miR-1185-1–CD24 axis in the regulatory network of colorectal cancer progression and stem cell properties.

Figure 7.

miR-1185-1 suppressed the tumorigenesis in mice xenograft model and was downregulated in human colorectal cancer tissues. A, The tumor sizes of shLuc (open circle) and miROE (solid circle) groups upon transplantation into Nu/Nu mice were measured. n = 5. P value was calculated by combined ANOVA. In this experiment, two group cells (miR-overexpression vs. shLuc control) were injected into bilateral sides of the trunk, respectively, in the same Nu/Nu mice. One side is control (shLuc, --, open circle) group and the other side is miR-1185-1 overexpressed group (miROE, -•-, solid circle). The diameters of tumors between miROE group and shLuc group were measured and compared in the same mouse every 2 days since 8th day after injection. One color with circle and line represents one mouse. Because of different tumorigenesis ability among different mice, the diameter changes of the tumors (miROE vs. shLuc group) were compared in the same mouse, represented by the same color. B, The gross pictures of tumors that were removed from mice after transplantation 20 days. The length of lattice was equal to 1.5 cm. Right, tumor weights of shLuc and miROE HT29 or DLD-1 cells. P value was calculated by Wilcoxon signed-rank test. C, The expressions of miR-1185-1, CD24, and SIRT1 in normal colonic and colorectal cancer tumor tissues of 3 patients were measured by qPCR. Results were normalized to RNA U6 for miR-1185-1 and to GAPDH and β-ACTIN for CD24 and SIRT1. n = 3; *, compared with the normal part group; *, P < 0.05; **, P < 0.01; ***, P < 0.005. D, Scheme of SIRT1–miR-1185-1–CD24 axis in the regulatory network of colorectal cancer progression and stem cell properties.

Close modal

Cancer stemness relates to the tumor aggressiveness and recurrence. Some stemness-related markers correlate with the poor prognosis of patient with colorectal cancer such as CD133 (5) and CD24 (4). Thus, eliminating the stemness of cancer cells may be a novel way for cancer therapy. The self-renewal and properties of cancer stem cells are maintained by complicated regulatory networks such as the genetic and epigenetic pathways. Here, we demonstrated SIRT1 could promote colorectal CSC properties via histone deacetylation to decrease the expression of a novel tumor suppressor miR-1185-1. Our finding also revealed that overexpressed miR-1185-1 in colorectal cancer cells significantly decreased the tumor formation ability in xenograft model and the CSC properties by targeting CD24. Therefore, we suggest that SIRT1–miR-1185-1–CD24 axis may facilitate the colorectal cancer tumorigenesis and participate in maintaining CSC properties of colorectal cancer cells (Fig. 7D).

SIRT1, a mammalian homolog of yeast silencing regulator 2 (sir2), is the most widely studied member of sirtuin. It can deacetylate histones and many nonhistone substrates, such as p53 and FOXO, to regulate cellular signalings and functions (37, 38). The role of SIRT1 in cancers has been studied over the past decades. However, the controversy over the role of SIRT1 in cancers still needs to be clarified. For instance, SIRT1, as an oncoprotein, deacetylates several tumor suppressors, such as p53 and p73, to suppress their functions. SIRT1 can reduce p53-mediated apoptosis and negatively regulate p53-induced cellular senescence in several types of cancer cells (38, 39). On the other hand, as a tumor suppressor, SIRT1 has an inhibitory effect on NFκB in TNFα signaling that sustains cancer cell survival of non‐-small cell lung cancer (40). In the colorectal cancer, SIRT1 has been reported not only to take part in the tumorigenesis (27), but also that its expression is positively correlated with the cancer progression in the clinical setting (41, 42). Our finding suggests SIRT1 is a oncoprotein through the promotion of colorectal cancer stemness (Fig. 2), which is consistent with the previous study (27). In SIRT1-knockdown colorectal cancer cells, the CSC frequency, tumor relapse ability and migration ability were reduced. We also found SIRT1 suppressed the expression of miR-1185-1, a novel colorectal cancer suppressor-miRNA, by histone deacetylation to promote the colorectal cancer aggressiveness. However, one study by Sun and colleagues has shown that SIRT1 can suppress colorectal cancer metastasis (43). The different status of the regulators to modulate SIRT1's enzymatic activity in different colorectal cancer cell lines may account for the discordant finding. Because SIRT1 is a NAD+-dependent class III HDAC, NAD/NADH metabolic status could influence its activity. NAD+ plays essential roles in metabolism, cellular signals, and energy utility of cancers. Recently, NAD(H) pool size (NAD+ and NADH) and NAD+/NADH ratio are shown to increase during colorectal cancer progression due to the activation of nicotinamide phosphoribosyltransferase (NAMPT), a rate-limiting enzyme in the NAD+ salvage pathway (44). Moreover, NAMPT has been revealed as one potent oncoprotein during colorectal cancer progression to increase the stemness properties through SIRT1 (29). These reports may imply the function of SIRT1 not only depends on the expression level, but is modulated by the ratio of NAD+/NADH and the activity of NAMPT. Therefore, understanding how to regulate the function of SIRT1 and its critical deacetylated targets is essential to clarify the role of SIRT1 in cancers.

CD24 has been found to be expressed in several kinds of human cancers, such as small-cell and non–small cell lung cancer, renal cell carcinoma, hepatocellular carcinoma, and colorectal cancer (21, 45–47). In colorectal cancer, the expression of CD24 is highly related to carcinogenesis (45, 48). CD24 is also considered as one of colorectal cancer stem cell surface markers, but the exact role of CD24 in colorectal cancer stemness properties is still unraveled. Our data revealed that knockdown of CD24 reduced stem cell properties of colorectal cancer cells (Fig. 3). These data suggest that CD24 is not only a colorectal cancer stem cell marker but involved in maintaining the network of colorectal cancer stemness properties. The downstream signaling of CD24 in cancer stemness is still unclear. CD24 could trigger STAT3 activation in colorectal cancer (20). STAT3/NANOG pathway has been reported to involve in colorectal cancer stemness (49). CD24 also modulates NOTCH signaling to promote breast cancer stemness (15). NOTCH signaling has been found to enhance stemness of colorectal cancer (50). These clues may suggest the regulatory network of CD24 in colorectal cancer stemness. On the other hand, CD24 is a GPI-anchor membrane protein that lacks intracellular domain, and may interact with membrane receptor neighborhoods to modulate the downstream signaling. Therefore, realizing the regulatory mechanism of the CD24-driven signaling is one of the important issues for cancer research advance.

We found that SIRT1 regulated the expression of CD24, and both SIRT1 and CD24 were upregulated in the colorectal cancer cells. As a result, it was implied CD24 might be one of the key factors involving in the SIRT1-related stemness regulatory pathway in colorectal cancer. By means of miRNA database screening, we found a novel CD24-targeted tumor suppressor miRNA, miR-1185-1, which could be suppressed by SIRT1-mediated histone deacetylation. To date, only few literatures address the function of miR-1185-1. Deng and colleagues reveal miR-1185-1 regulates the expression of VCAM-1 and E-selectin to promote arteriosclerosis, and induces apoptosis of endothelial cell by targeting UVRAG and KRIT1 (51, 52). In colorectal cancer, miR-1185-1 is found to be downregulated in colorectal cancer cell lines (53). However, the function and regulatory mechanism of miR-1185-1 in colorectal cancer were completely unknown until we found some clues in the current study. Here, we demonstrated that miR-1185-1 was substantially downregulated in three colorectal cancer tumors. Moreover, overexpression of miR-1185-1 suppressed the tumorigenesis of colorectal cancer in the xenograft model (Fig. 7). Overexpression of CD24 in miR-1185-1–overexpressed colorectal cancer cells rescued the tumor-suppressing effect of miR-1185-1 (Fig. 6), indicating the tumor suppressor role of miR-1185-1 was through targeting CD24 (Fig. 5). In addition to CD24, some predicted target genes of miR-1185-1 were related to the tumorigenesis. mTOR was one of miR-1185-1 predicted targets and downregulated in miR-1185-1–overexpressed colorectal cancer cells (Supplementary Fig. S3A). mTOR has been reported to maintain the stemness in cancer stem cells (54). Therefore, miR-1185-1 might also suppress the tumor formation via mTOR pathway. Besides, based on KEGG results (Supplementary Table S3), miR-1185-1 target gene–related pathways were highly associated with metabolic pathway, PI3K–Akt signaling pathway, and pathways in cancer. However, the comprehensive pathophysiology of miR-1185-1 and the complicated network between miR-1185-1 and the tumorigenesis remain to be elucidated further.

Immune checkpoint and associated proteins have become the therapeutic targets for cancer treatment (55, 56). Recently, Barkal and colleagues have indicated CD24-Siglec10 antiphagocytic signal can be a target for cancer immunotherapy (17). In this study, we revealed CD24 facilitated colorectal cancer stemness and aggressiveness. Thus, to eliminate the function or the expression of CD24 may demonstrate the therapeutic potential to deal with both tumor immunity and cancer stemness. Our work uncovered a novel tumor suppressor miRNA, miR-1185-1, which could downregulate the expression of CD24 and inhibited the tumorigenesis of colorectal cancer. Therefore, the RNA mimic of miR-1185-1 may be considered as one therapeutic option for CD24-targeted therapy against both the immunosuppressive tumor microenvironment and cancer stemness. In summary, our findings provide not only the evidence on the essential role of the SIRT1–miR-1185-1–CD24 signaling in the molecular regulation of stemness networks in the colorectal cancer but the potential therapeutic agent against CD24.

No disclosures were reported.

T.-W. Wang: Conceptualization, data curation, software, formal analysis, investigation, methodology. E. Chern: Conceptualization, resources, data curation, funding acquisition, validation, investigation, writing-original draft, writing-review and editing. C.-W. Hsu: Conceptualization, methodology. K.-C. Tseng: Conceptualization, supervision. H.-M. Chao: Supervision, funding acquisition, writing-original draft, project administration, writing-review and editing.

This study was supported by grants from the Ministry of Science and Technology (MOST), Taiwan (MOST 104-2320-B-002-047-; 109-2321-B-002-044-; 109-2320-B-002-051-MY2). We also thank Yi-Chun Liao who provided us the CCD 841CoN cells and the excellent technical support of Technology Commons, College of Life Science, National Taiwan University.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Siegel
RL
,
Miller
KD
,
Jemal
A
. 
Cancer Statistics, 2017
.
CA Cancer J Clin
2017
;
67
:
7
30
.
2.
Pardal
R
,
Clarke
MF
,
Morrison
SJ
. 
Applying the principles of stem-cell biology to cancer
.
Nat Rev Cancer
2003
;
3
:
895
902
.
3.
Dean
M
,
Fojo
T
,
Bates
S
. 
Tumour stem cells and drug resistance
.
Nat Rev Cancer
2005
;
5
:
275
84
.
4.
Yeung
TM
,
Gandhi
SC
,
Wilding
JL
,
Muschel
R
,
Bodmer
WF
. 
Cancer stem cells from colorectal cancer-derived cell lines
.
Proc Natl Acad Sci U S A
2010
;
107
:
3722
7
.
5.
Weichert
W
,
Denkert
C
,
Burkhardt
M
,
Gansukh
T
,
Bellach
J
,
Altevogt
P
, et al
Cytoplasmic CD24 expression in colorectal cancer independently correlates with shortened patient survival
.
Clin Cancer Res
2005
;
11
:
6574
81
.
6.
Huang
R
,
Mo
D
,
Wu
J
,
Ai
H
,
Lu
Y
. 
CD133 expression correlates with clinicopathologic features and poor prognosis of colorectal cancer patients: An updated meta-analysis of 37 studies
.
Medicine
2018
;
97
:
e10446
.
7.
Sagiv
E
,
Starr
A
,
Rozovski
U
,
Khosravi
R
,
Altevogt
P
,
Wang
T
, et al
Targeting CD24 for treatment of colorectal and pancreatic cancer by monoclonal antibodies or small interfering RNA
.
Cancer Res
2008
;
68
:
2803
12
.
8.
Zaytseva
YY
,
Rychahou
PG
,
Gulhati
P
,
Elliott
VA
,
Mustain
WC
,
O'Connor
K
, et al
Inhibition of fatty acid synthase attenuates CD44-associated signaling and reduces metastasis in colorectal cancer
.
Cancer Res
2012
;
72
:
1504
17
.
9.
Sato-Dahlman
M
,
Miura
Y
,
Huang
JL
,
Hajeri
P
,
Jacobsen
K
,
Davydova
J
, et al
CD133-targeted oncolytic adenovirus demonstrates anti-tumor effect in colorectal cancer
.
Oncotarget
2017
;
8
:
76044
56
.
10.
Sahlberg
SH
,
Spiegelberg
D
,
Glimelius
B
,
Stenerlow
B
,
Nestor
M
. 
Evaluation of cancer stem cell markers CD133, CD44, CD24: association with AKT isoforms and radiation resistance in colon cancer cells
.
PLoS One
2014
;
9
:
e94621
.
11.
Suzuki
T
,
Kiyokawa
N
,
Taguchi
T
,
Sekino
T
,
Katagiri
YU
,
Fujimoto
J
. 
CD24 Induces Apoptosis in Human B Cells Via the Glycolipid-Enriched Membrane Domains/Rafts-Mediated Signaling System
.
J Immunol
2001
;
166
:
5567
77
.
12.
Lim
SC
,
Oh
SH
. 
The role of CD24 in various human epithelial neoplasias
.
Pathol Res Pract
2005
;
201
:
479
86
.
13.
Smith
SC
,
Oxford
G
,
Wu
Z
,
Nitz
MD
,
Conaway
M
,
Frierson
HF
, et al
The metastasis-associated gene CD24 is regulated by Ral GTPase and is a mediator of cell proliferation and survival in human cancer
.
Cancer Res
2006
;
66
:
1917
22
.
14.
Hofner
T
,
Macher-Goeppinger
S
,
Klein
C
,
Schillert
A
,
Eisen
C
,
Wagner
S
, et al
Expression and prognostic significance of cancer stem cell markers CD24 and CD44 in urothelial bladder cancer xenografts and patients undergoing radical cystectomy
.
Urol Oncol
2014
;
32
:
678
86
.
15.
Lim
J
,
Lee
KM
,
Shim
J
,
Shin
I
. 
CD24 regulates stemness and the epithelial to mesenchymal transition through modulation of Notch1 mRNA stability by p38MAPK
.
Arch Biochem Biophys
2014
;
558
:
120
6
.
16.
Kwon
MJ
,
Han
J
,
Seo
JH
,
Song
K
,
Jeong
HM
,
Choi
JS
, et al
CD24 Overexpression is associated with poor prognosis in luminal a and triple-negative breast cancer
.
PLoS One
2015
;
10
:
e0139112
.
17.
Barkal
AA
,
Brewer
RE
,
Markovic
M
,
Kowarsky
M
,
Barkal
SA
,
Zaro
BW
, et al
CD24 signalling through macrophage Siglec-10 is a target for cancer immunotherapy
.
Nature
2019
;
572
:
392
6
.
18.
Su
N
,
Peng
L
,
Xia
B
,
Zhao
Y
,
Xu
A
,
Wang
J
, et al
Lyn is involved in CD24-induced ERK1/2 activation in colorectal cancer
.
Mol Cancer
2012
;
11
:
43
.
19.
Muppala
S
,
Mudduluru
G
,
Leupold
JH
,
Buergy
D
,
Sleeman
JP
,
Allgayer
H
. 
CD24 induces expression of the oncomir miR-21 via Src, and CD24 and Src are both post-transcriptionally downregulated by the tumor suppressor miR-34a
.
PLoS One
2013
;
8
:
e59563
.
20.
Wang
X
,
Zhang
Y
,
Zhao
Y
,
Liang
Y
,
Xiang
C
,
Zhou
H
, et al
CD24 promoted cancer cell angiogenesis via Hsp90-mediated STAT3/VEGF signaling pathway in colorectal cancer
.
Oncotarget
2016
;
7
:
55663
76
.
21.
Liu
AY
,
Cai
Y
,
Mao
Y
,
Lin
Y
,
Zheng
H
,
Wu
T
, et al
Twist2 promotes self-renewal of liver cancer stem-like cells by regulating CD24
.
Carcinogenesis
2014
;
35
:
537
45
.
22.
Vaquero
A
,
Scher
M
,
Lee
D
,
Erdjument-Bromage
H
,
Tempst
P
,
Reinberg
D
. 
Human SirT1 interacts with histone H1 and promotes formation of facultative heterochromatin
.
Mol Cell
2004
;
16
:
93
105
.
23.
Liu
T
,
Liu
PY
,
Marshall
GM
. 
The critical role of the class III histone deacetylase SIRT1 in cancer
.
Cancer Res
2009
;
69
:
1702
5
.
24.
Lin
Z
,
Fang
D
. 
The roles of SIRT1 in cancer
.
Genes Cancer
2013
;
4
:
97
104
.
25.
Wilson
AJ
,
Byun
DS
,
Popova
N
,
Murray
LB
,
L'Italien
K
,
Sowa
Y
, et al
Histone deacetylase 3 (HDAC3) and other class I HDACs regulate colon cell maturation and p21 expression and are deregulated in human colon cancer
.
J Biol Chem
2006
;
281
:
13548
58
.
26.
Nakazawa
T
,
Kondo
T
,
Ma
D
,
Niu
D
,
Mochizuki
K
,
Kawasaki
T
, et al
Global histone modification of histone H3 in colorectal cancer and its precursor lesions
.
Hum Pathol
2012
;
43
:
834
42
.
27.
Chen
X
,
Sun
K
,
Jiao
S
,
Cai
N
,
Zhao
X
,
Zou
H
, et al
High levels of SIRT1 expression enhance tumorigenesis and associate with a poor prognosis of colorectal carcinoma patients
.
Sci Rep
2014
;
4
:
7481
.
28.
Pruitt
K
,
Zinn
RL
,
Ohm
JE
,
McGarvey
KM
,
Kang
SH
,
Watkins
DN
, et al
Inhibition of SIRT1 reactivates silenced cancer genes without loss of promoter DNA hypermethylation
.
PLoS Genet
2006
;
2
:
e40
.
29.
Lucena-Cacace
A
,
Otero-Albiol
D
,
Jimenez-Garcia
MP
,
Munoz-Galvan
S
,
Carnero
A
. 
NAMPT is a potent oncogene in colon cancer progression that modulates cancer stem cell properties and resistance to therapy through sirt1 and PARP
.
Clin Cancer Res
2018
;
24
:
1202
15
.
30.
Dull
T
,
Nguyen
M
,
Trono
D
,
Mandel
RJ
,
Zufferey
R
,
Kelly
M
, et al
A third-generation lentivirus vector with a conditional packaging system
.
J Virol
1998
;
72
:
8463
71
.
31.
Izumiya
M
,
Kabashima
A
,
Higuchi
H
,
Igarashi
T
,
Sakai
G
,
Iizuka
H
, et al
Chemoresistance is associated with cancer stem cell-like properties and epithelial-to-mesenchymal transition in pancreatic cancer cells
.
Anticancer Res
2012
;
32
:
3847
54
.
32.
Hu
Y
,
Smyth
GK
. 
Extreme limiting dilution analysis for comparing depleted and enriched populations in stem cell and other assays
.
J Immunol Methods
2009
;
347
:
70
8
.
33.
Chao
HM
,
Huang
HX
,
Chang
PH
,
Tseng
KC
,
Miyajima
A
,
Chern
E
. 
Y-box binding protein-1 promotes hepatocellular carcinoma-initiating cell progression and tumorigenesis via Wnt/beta-catenin pathway
.
Oncotarget
2017
;
8
:
2604
16
.
34.
Goldman
MJ
,
Craft
B
,
Hastie
M
,
Repecka
K
,
McDade
F
,
Kamath
A
, et al
Visualizing and interpreting cancer genomics data via the Xena platform
.
Nat Biotechnol
2020
;
38
:
675
8
.
35.
Cao
L
,
Zhou
Y
,
Zhai
B
,
Liao
J
,
Xu
W
,
Zhang
R
, et al
Sphere-forming cell subpopulations with cancer stem cell properties in human hepatoma cell lines
.
BMC Gastroenterology
2011
;
11
:
71
.
36.
Golebiewska
A
,
Brons
NH
,
Bjerkvig
R
,
Niclou
SP
. 
Critical appraisal of the side population assay in stem cell and cancer stem cell research
.
Cell Stem Cell
2011
;
8
:
136
47
.
37.
Kobayashi
Y
,
Furukawa-Hibi
Y
,
Chen
C
,
Horio
Y
,
Isobe
K
,
Ikeda
K
, et al
SIRT1 is critical regulator of FOXO-mediated transcription in response to oxidative stress
.
Int J Mol Med
2005
;
16
:
237
43
.
38.
Yi
J
,
Luo
J
. 
SIRT1 and p53, effect on cancer, senescence and beyond
.
Biochim Biophys Acta
2010
;
1804
:
1684
9
.
39.
Dai
JM
,
Wang
ZY
,
Sun
DC
,
Lin
RX
,
Wang
SQ
. 
SIRT1 interacts with p73 and suppresses p73-dependent transcriptional activity
.
J Cell Physiol
2007
;
210
:
161
6
.
40.
Yeung
F
,
Hoberg
JE
,
Ramsey
CS
,
Keller
MD
,
Jones
DR
,
Frye
RA
, et al
Modulation of NF-kappaB-dependent transcription and cell survival by the SIRT1 deacetylase
.
EMBO J
2004
;
23
:
2369
80
.
41.
Lv
L
,
Shen
Z
,
Zhang
J
,
Zhang
H
,
Dong
J
,
Yan
Y
, et al
Clinicopathological significance of SIRT1 expression in colorectal adenocarcinoma
.
Med Oncol
2014
;
31
:
965
.
42.
Shen
ZL
,
Wang
B
,
Jiang
KW
,
Ye
CX
,
Cheng
C
,
Yan
YC
, et al
Downregulation of miR-199b is associated with distant metastasis in colorectal cancer via activation of SIRT1 and inhibition of CREB/KISS1 signaling
.
Oncotarget
2016
;
7
:
35092
105
.
43.
Sun
LN
,
Zhi
Z
,
Chen
LY
,
Zhou
Q
,
Li
XM
,
Gan
WJ
, et al
SIRT1 suppresses colorectal cancer metastasis by transcriptional repression of miR-15b-5p
.
Cancer Lett
2017
;
409
:
104
15
.
44.
Zhang
H
,
Zhang
N
,
Liu
Y
,
Su
P
,
Liang
Y
,
Li
Y
, et al
Epigenetic Regulation of NAMPT by NAMPT-AS drives metastatic progression in triple-negative breast cancer
.
Cancer Res
2019
;
79
:
3347
59
.
45.
Sagiv
E
,
Memeo
L
,
Karin
A
,
Kazanov
D
,
Jacob-Hirsch
J
,
Mansukhani
M
, et al
CD24 is a new oncogene, early at the multistep process of colorectal cancer carcinogenesis
.
Gastroenterology
2006
;
131
:
630
9
.
46.
Zheng
Y
,
de la Cruz
CC
,
Sayles
LC
,
Alleyne-Chin
C
,
Vaka
D
,
Knaak
TD
, et al
A rare population of CD24(+)ITGB4(+)Notch(hi) cells drives tumor propagation in NSCLC and requires Notch3 for self-renewal
.
Cancer Cell
2013
;
24
:
59
74
.
47.
Arik
D
,
Can
C
,
Dundar
E
,
Kabukcuoglu
S
,
Pasaoglu
O
. 
Prognostic significance of CD24 in clear cell renal cell carcinoma
.
Pathol Oncol Res
2017
;
23
:
409
16
.
48.
Ahmed
MA
,
Al-Attar
A
,
Kim
J
,
Watson
NF
,
Scholefield
JH
,
Durrant
LG
, et al
CD24 shows early upregulation and nuclear expression but is not a prognostic marker in colorectal cancer
.
J Clin Pathol
2009
;
62
:
1117
22
.
49.
Yao
C
,
Su
L
,
Shan
J
,
Zhu
C
,
Liu
L
,
Liu
C
, et al
IGF/STAT3/NANOG/slug signaling axis simultaneously controls epithelial-mesenchymal transition and stemness maintenance in colorectal cancer
.
Stem Cells
2016
;
34
:
820
31
.
50.
Fender
AW
,
Nutter
JM
,
Fitzgerald
TL
,
Bertrand
FE
,
Sigounas
G
. 
Notch-1 promotes stemness and epithelial to mesenchymal transition in colorectal cancer
.
J Cell Biochem
2015
;
116
:
2517
27
.
51.
Deng
H
,
Chu
X
,
Song
Z
,
Deng
X
,
Xu
H
,
Ye
Y
, et al
MicroRNA-1185 induces endothelial cell apoptosis by targeting UVRAG and KRIT1
.
Cell Physiol Biochem
2017
;
41
:
2171
82
.
52.
Deng
H
,
Song
Z
,
Xu
H
,
Deng
X
,
Zhang
Q
,
Chen
H
, et al
MicroRNA-1185 promotes arterial stiffness though modulating VCAM-1 and E-selectin expression
.
Cell Physiol Biochem
2017
;
41
:
2183
93
.
53.
Alqurashi
N
,
Hashimi
SM
,
Alowaidi
F
,
Ivanovski
S
,
Farag
A
,
Wei
MQ
. 
miR-496, miR-1185, miR-654, miR-3183 and miR-495 are downregulated in colorectal cancer cells and have putative roles in the mTOR pathway
.
Oncol Lett
2019
;
18
:
1657
68
.
54.
Matsumoto
K
,
Arao
T
,
Tanaka
K
,
Kaneda
H
,
Kudo
K
,
Fujita
Y
, et al
mTOR signal and hypoxia-inducible factor-1 alpha regulate CD133 expression in cancer cells
.
Cancer Res
2009
;
69
:
7160
4
.
55.
Jiao
X
,
Nawab
O
,
Patel
T
,
Kossenkov
AV
,
Halama
N
,
Jaeger
D
, et al
Recent advances targeting CCR5 for cancer and its role in immuno-oncology
.
Cancer Res
2019
;
79
:
4801
7
.
56.
Tacconi
C
,
Ungaro
F
,
Correale
C
,
Arena
V
,
Massimino
L
,
Detmar
M
, et al
Activation of the VEGFC/VEGFR3 pathway induces tumor immune escape in colorectal cancer
.
Cancer Res
2019
;
79
:
4196
210
.