The study was approved by the institutional review board of Huashan Hospital Affiliated to Fudan University (HIRB), Shanghai, China. All patients provided written informed consent. All in vivo experiments in the study protocol were strictly in accordance with the National Institutes of Health Guide for Care and Use of Laboratory Animals and were approved by the Animal Care and Use Committee of Shanghai Medical College of Fudan University, Shanghai, China.

CRC specimens including adjacent non-tumor colorectal tissues were obtained from patients who have undergone radical surgery for CRC in the Huashan Hospital, Shanghai, China. None of the patients received chemotherapy or radiotherapy before operation. The samples were surgically obtained at the following time points: 23 cases between 10:00 and 12:00, 8 cases between 12:00 and 14:00, 3 cases between 14:00 and 16:00, 3 cases between 16:00 and 18:00, and 1 case at 22:00.

Four human CRC cell lines SW480, SW620, HT29, LoVo, and human embryonic kidney cell line 293T were cultured in Dulbecco's modified Eagle's medium (DMEM) with 10% (v/v) newborn calf serum, 100 U/ml penicillin and 100 ng/ml streptomycin at 37°C in 5% CO₂. All cell lines were obtained from the Cell Bank of the Chinese Academy of Sciences, Shanghai, China.

Detailed materials and methods performed were according to our previous study (28).

Total cellular proteins were extracted and separated in SDS-PAGE gel, and western blot analysis was performed according to the standard procedures. GAPDH was used as a loading control on the same membrane. The antibodies used included anti-hClock, anti-VEGF (Abcam, Cambridge, UK), anti-GAPDH, anti-HIF-1α, anti-ARNT, anti-E-cadherin, anti-N-cadherin, anti-vimentin, and anti-fibronectin (Cell Signaling Technology, Danvers, MA, USA). Band densitometry was performed using Scion Image software (Scion Corp., Fredrick, MD, USA).

Full-length cDNA of hClock (Genbank accession number: NM_004898) was amplified and inserted into pcDNA3-Flag vector (Invitrogen, Carlsbad, CA, USA) to generate pcDNA3-Flag-hClock expression plasmid. The cDNA fragment of Flag-hClock was inserted into the pGV186 retroviral vector to generate pGV186-Flag-hClock. All the constructs were confirmed by DNA sequencing.

Detailed materials and methods used in this section were performed according to our previous work (27).

Retroviruses carrying the hClock or hClockRNAi were generated by cotransfection of recombinant pGV186 or pGV113 plasmids, respectively with pHelper plasmid into 293T cells using Lipofectamine 2000 (Invitrogen). After 48 h of incubation, the culture medium containing recombinant virus was harvested and purified by a 0.45 μm filter. Target cells were seeded (5×10⁵/well) into 6-well plates and incubated with recombinant virus was supplemented with 5 μg/ml polybrene for a spin infection procedure.

The migration ability of the cells was measured in transwell chambers (8 μm pore; BD Biosciences). The bottom chamber was filled with 600 μl of DMEM containing 10% FBS. For the migration assay, tumor cells (5×10⁴ cells in a total volume of 100 μl) were placed in the upper chamber and incubated at 37°C in 5% CO₂ under humidified air conditions. After 24 h of culture, non-migrating cells on the upper surface of the membrane were removed, and cells that migrated to the underside of the polycarbonate membrane were fixed with ethanol and stained with 1% crystal violet for 10 min. The number of migrating cells was then determined from 5 independent microscopic fields. The mean of triplicate assays for each experimental condition was used for the analysis.

Female, 5–6 weeks old, Balb/c nude mice were obtained from Shanghai Experimental Animal Center and provided with standard laboratory chow and tap water ad libitum under special pathogen free (SPF) conditions in the department of laboratory animal science of Shanghai Medical College of Fudan University. The mice were then separated into 2 groups randomly and were injected with SW620-hClock-shRNA 2# or SW620-shRNA-control infected cells via the caudal veins (10 mice/group, 1×10⁶ cells/mouse). Six weeks after injection, all mice were euthanatized by cervical dislocation and their lungs were sectioned and stained with hematoxylin and eosin (H&E), and then observed by light microscopy for the formation of lung metastasis. All the animal studies were conducted in accordance with 'Animal Research: Reporting In vivo Experiments' (ARRIVE) guidelines and the guidelines of Institutional Animal Care and Use Committee. All the mice were treated humanely throughout the experimental period.

The data are presented as mean ± SD unless stated otherwise. A paired t-test was used to test for the differences in hClock expression between matched tumor and benign mucosa. χ²-test or Fisher's exact test was performed to analyze the correlations of hClock protein levels with clinical and pathologic parameters. The statistical significance of the in vitro and in vivo studies was analyzed using Student's t-test. All P-values were two-sided, and P-values of <0.05 were considered to be statistically significant. All statistical analyses were conducted using SPSS software package, version 19.0 (SPSS Inc., Chicago, IL, USA).

In our previous studies, we found that hClock was highly expressed in CRC tissues when compared with the peritumoral tissues in CRC patients, and strongly associated with late TNM stage and positive lymph node metastasis (26). Similar results can be found in the present study since another additional 8 cases were added to the current study and were further analyzed.

We analyzed hClock mRNA levels in 38 pairs of human CRC samples, including primary CRC tissues and paired normal colorectal tissues in this study. Significantly increased expression of hClock mRNA was observed in 26 of 38 (68.4%) CRCs compared with the paired normal colorectal tissues (P<0.01). Relative expression levels of hClock in the CRC compared with the non-cancerous components were 2.75:1 (Fig. 1).

hClock expression at the protein level was further examined in these 38 pairs of CRC tissues by immunohistochemical staining. hClock protein was predominantly detected in the nuclei of tumor cells. In all, there were three different hClock protein expression patterns: type I (50.0%, 19/38): staining for hClock was much stronger in the cancerous cells than in the paired non-cancerous cells (Fig. 2A); type II (44.7%, 17/38): there were no visually significant differences between the staining of cancerous and non-cancerous cells (Fig. 2B), although subtle differences in the hClock protein expression could not be excluded; type III (5.3%, 2/38): staining of hClock was stronger in non-cancerous cells than in the paired cancerous cells (Fig. 2C). Furthermore, 26 cases showed upregulated expression of hClock mRNA in the CRC tissues, 18 specimens demonstrated a higher expression of hClock protein. Thus, the expression of hClock protein and mRNA revealed a positive correlation (P<0.01).

We next examined the possible intrinsic relationship between hClock expression and clinicopathological features. As shown in Table I, hClock upregulation did not correlate with sex, age, and tumor sites, but increased expression of hClock was strongly associated with positive lymph node metastasis (P=0.003) and advanced TNM staging (P=0.003). These results suggested that higher hClock expression was associated with the degree of malignancy for CRC.

To validate our findings from the clinical CRC samples, we examined cellular hClock expression in four different CRC cell lines (HT-29, LoVo, SW480 and SW620). Results of hClock expression in the four CRC cell lines are shown in Fig. 1. Noteworthy, hClock was highly expressed in SW620 both at the mRNA and protein level (Fig. 3A and B). Since SW620 has the highest potential of metastasis among these CRC cell lines (28), our result again suggested that hClock overexpression significantly correlated with CRC metastasis.

To evaluate the role of hClock in CRC, we next tested whether overexpression of hClock promoted the migration of human CRC cells. We used the CRC cell line SW480 which relatively expressed a lower level of hClock. hClock was overexpressed in SW480 cells when infected with pGV186-hClock lentiviruses (Fig. 4A and B). Transwell migration assay showed that overexpression of hClock enhanced the migration ability of SW480 CRC cells (Fig. 4C, P<0.01). These findings suggested that hClock plays an important role in the CRC cell migration.

To further investigate the relationship between hClock upregulation and CRC progression, endogeneous hClock expression in the CRC cell line SW620 was knocked down and cell migration was examined. Transfection of SW620 cells with hClock hRNA significantly inhibited the hClock expression (Fig. 5A and B). To further examine whether the targeted knockdown of hClock in SW620 cells affected their migration, we performed an in vitro Transwell migration assay. The result showed that the number of hClock-shRNA cells migrated through the filter decreased markedly by contrast to the control group (Fig. 5C, P<0.01). These results together indicated that hClock was critical for CRC cell proliferation and migration in vitro.

To further explore the role of hClock in CRC metastasis in vivo, SW620 CRC tumor cells transfected with hClock-shRNA or with control hRNA were injected into the nude mice via the caudal vein (10 mice/group, 1×10⁶ cells/mouse). Six weeks after injection, none of the mice had died or behaved abnormally, and they were euthanatized. Their lungs were sectioned and stained with H&E, and then observed by light microscopy. Notably, in contrast to the control group, a significant reduction in the number of metastatic nodules in the lungs of nude mice injected with hClock-shRNA-transfected SW620 cells was observed (Fig. 5D). These results showed that targeting hClock by shRNA effectively suppressed the metastatic ability of SW620 CRC cells in nude mice and suggested that hClock knockdown exerted a strong antitumor effect in vivo.

Taken together, both our in vitro and in vivo experiments provided significant evidence that hClock was strongly associated with CRC metastasis.

It has been previously reported that EMT was the driving force of invasion in the epithelial cells (29,30). In recent years, changes in the level of expression of different cadherins, so-called cadherin switches have been increasingly used to monitor EMT. The cadherin switch from E-cadherin to N-cadherin, which is expressed in mesenchymal cells, fibroblasts, cancer cells, and neural tissues has often been used to monitor the EMT progress during the embryonic development and cancer progression (31). Thus, to investigate the mechanism underlying the effects of hClock overexpression on CRC progression, we examined the expression of EMT markers E-cadherin and N-cadherin. As shown in Fig. 6A and B, expression levels of E-cadherin were significantly decreased in SW480 cells which overexpressed hClock compared with those of the vector only-treated SW480 cells (P<0.05), whereas upregulation of hClock resulted in a significant increase in N-cadherin expression (P<0.05). In contrast, silencing of endogenous hClock expression resulted in a marked decrease of N-cadherin expression compared with that of the shRNA-transfected controls, whereas E-cadherin expression levels were significantly increased (P<0.05) (Fig. 6A and C).

In parallel, increasing evidence demonstrates that circadian genes participate in the angiogenesis and growth of tumors (32,33). Thus, we analyzed the expression of tumor angiogenesis-related genes, such as HIF-1α, ARNT, VEGF in the CRC cell lines with hClock knockdown or over expression, to see whether ectopic expression of hClock contributes to tumor angiogenesis and malignancy. Western blot analysis indicated that expression of HIF-1α, ARNT and VEGF were decreased in SW620 cells transfected with hClock shRNA, but increased in SW480 cells with hClock overexpression (Fig. 7).