CKLF1 Enhances Inflammation-Mediated Carcinogenesis and Prevents Doxorubicin-Induced Apoptosis via IL6/STAT3 Signaling in HCC

Hepatocellular carcinoma (HCC), the most common form (70%–90%) of liver cancer, is ranked as the second leading cause of cancer-related deaths worldwide (1, 2). An estimated 788,000 deaths because of liver cancer occurred worldwide during 2015 (8.9% of the total), with China alone accounting for over 50% (422,100 deaths; ref. 3). The development and progression of HCC, including cellular transformation, proliferation, metastasis, invasion, and angiogenesis (4, 5), are inflammation-based carcinogenic processes because more than 90% of HCC in patients develop in the context of chronic liver damage and inflammation (6, 7). Chemokines play vital roles in the inflammation-mediated promotion of tumor invasion and metastasis by regulating the migration of immune cells into damaged or diseased organs in response to proinflammatory stimuli (8, 9). Chemokines are a class of small-molecular weight proteins ranging in size from approximately 8 to 13 kDa. Numerous CC chemokines, one group of chemokines, participate in the initiation of inflammation, progression, and maintenance of liver inflammation and fibrosis and have been shown to regulate tumor-dependent angiogenesis and several biologic actions in hepatocytes, including proliferation (10, 11).

Transcript variant 2 of the chemokine-like factor, named CKLF1, is a newly discovered CC chemokine. CKLF1 has three isoforms, CKLF2, CKLF3, and CKLF4, among which CKLF2 is the full-length cDNA product. All these four isoforms contain a CC motif characteristic for the CC chemokine family. Previous studies have verified that CKLFs have multiple activities, such as attract human neutrophils, lymphocytes, and monocytes and stimulate the proliferation and differentiation of murine skeletal myoblast cells. CKLF1 has shown broad chemotactic activities on leukocytes and plays essential roles in inflammatory and autoimmune diseases (12–17). However, the relationship between CKLF1 and cancer is uncertain, particularly HCC.

In this article, we discovered that levels of CKLF1 and STAT3 in HCC tissues were higher than that in normal liver tissues by Oncomine database and R software. The level of CKLF1 was related to vascular invasion and tumor size. Particularly the overexpression of CKLF1 and STAT3 correlated with poor overall survival (OS). Furthermore, clinical data revealed positive relationships between CKLF1 and HCC/STAT3. A series of experiments, including real-time cellular analysis (RTCA), wound healing assay, transwell, foci formation assay, subcutaneous xenograft experiment, and Western blotting, were used subsequently to explore the mechanism of CKLF1-mediated carcinogenesis in HCC. The results suggested that CKLF1 might promote the malignant transformation, as well as invasion and metastasis in the development of HCC by activating the IL6/STAT3 signaling pathway. Moreover, CKLF1 prevented doxorubicin-induced apoptosis via the IL6/STAT3 pathway.

In conclusion, this study suggested a new mechanism of HCC development and illustrated detailed information for further exploration of inflammation-mediated carcinogenesis, immune escape and so on. Our findings provided a novel potential biomarker and therapeutic target for HCC.

The Oncomine database (http://www.oncomine.org/resource/login.html) was used to predict the CKLF1 expression levels in HCC and noncancerous hepatic tissues. Then Gene Expression Omnibus (GEO) datasets were examined for specifically dysregulated genes in HCC. The mRNA profiling data of HCC samples were obtained from the GEO (http://www.ncbi.nlm.nih.gov/geo). The platform used for this work was GPL570 (Affymetrix Human Genome U133 Plus2.0 Array). With P ≤ 0.05 and fold change ≥ 2, all bioinformatics analyses were performed with R software. The R software was also used to analyze differential gene expression patterns between tumor and normal samples. Gene Ontology (GO) annotation and KEGG pathway analyses were performed on the basis of the target genes of R software. The GO annotation consisted of biological process, cellular component, and molecular function.

A total of 58 HCC samples and their corresponding nontumor tissues were collected for qPCR and Western blotting assay. Another 46 pairs of HCC samples were obtained for IHC staining assay. All the clinical samples were collected from Renmin Hospital of Wuhan University (Wuhan, China). The tumor–node–metastasis (TNM) classification of malignant tumors of the Union for International Cancer Control published in 2017 by John Wiley & Sons, Ltd was used to classify the samples. Sample collections were under consensus agreements and approved by the Ethics committee of Wuhan University, School of Medicine (Wuhan, China). The study was conducted in accordance with the International Ethical Guidelines for Biomedical Research Involving Human Subjects (CIOMS).

The real-time effects of CKLF1 on cell proliferation were measured using the xCELLigence RTCA System (Roche) as described previously (18). First, the background measurements were recorded after adding 100 μL culture medium to the wells. Then, cells were seeded at a density of 5,000 cell/well in 16-well plate with electrodes for 18 hours to allow the cells to progress to the log phase. The cells were treated with different conditions and continuously monitored for up to 60 hours. The cell sensor impedance was expressed as cell index (CI). CI was recorded every 5 minutes using RTCA analyzer. The CI values were normalized to the value at the beginning of the treatment to eliminate variation between the wells.

Transwell without Matrigel was used to verify cell migration. Cells in serum-free medium were seeded on transwell chambers (Corning) without Matrigel at approximately 2–6 × 10⁴. Medium containing 10% FBS in the bottom chamber served as the chemoattractant. Following incubation of 48 hours, the cells that had migrated through the membrane were fixed and stained with crystal violet and counted under a microscope (Olympus CH-40; ref. 19).

Transwell was also used to assess cell invasion. The method was the same as above, but the Matrigel (BD Biosciences) diluted to 200 μg/mL was used.

Full-length human CKLF1 cDNA was PCR amplified and cloned into pcDNA3.1 (−) vector according to the manufacturer's instruction (Invitrogen, catalog no. V790-20). For knocking down CKLF1, one short hairpin RNA targeting CKLF1 (shCKLF1) and the scramble control short hairpin RNA (shCtl) were cloned into the pSilencer 2.1-U6 neo siRNA expression vector.

The level of IL6 in culture supernatants was measured by ELISA Technique (R&D Systems), following the manufacturer's instructions.

Western blotting was performed using the standard method. Anti-β-actin antibody from Sigma (A3854) was used as a loading control to normalize the levels of other proteins. Rabbit antibodies for CKLF1 (ab180512), STAT3 (ab68153), p-STAT3 (Y705) (ab76315), Bcl-xl (ab32370), MYC (ab32072), and Cyclin D1 (ab134175) were purchased from Abcam.

In vitro tumorigenicity was assessed by foci formation. Cells were seeded into 6-well plates at 1,000–2,000 cells/well. After 2–3 weeks, the colonies of cells were stained with crystal violet and the colonies consisting of >50 cells each were manually counted. The results were expressed as the means ± SD of three independent experiments.

The subcutaneous xenograft experiment was used to test the in vivo tumorigenicity of CKLF1 in HCC cell lines. Cells were subcutaneously injected into the dorsal flank of 4- to 5-week old BALB/c-nu (nude) mice, respectively. Mice were sacrificed by anesthesia and tumors were dissected over a 4–6 week period. All animal care and handling procedures were performed in accordance with the National Institutes of Health's Guide for the Care and Use of Laboratory Animals. Animal experiments were approved by the Animal Ethics Committee of Wuhan University, Wuhan University Center for Animal Experiment/A3 Laboratory.

Stattic, the inhibitor of p-STAT3 (Y705), was used to prevent the activation of STAT3. Cells were treated with 2 nmol/L stattic in a humidified incubator with 5% CO₂ for 24 hours at 37°C.

Doxorubicin, a chief chemotherapeutic agent in cancer treatment, was used to induce apoptosis in HCC cell lines. Cells were treated with 6 μmol/L doxorubicin for 48 hours, followed by flow cytometry and Western blotting.

The experiments were representative of at least three independent trials, each with three technical replicates. Statistical analysis was performed using SPSS 13.0. Pearson χ² test or Fisher exact test were employed to compare qualitative variables, whereas Student t test was used for quantitative variables. One-way or two-way ANOVA with Holm-Sidak correction or with Newman–Keuls correction were used for multiple comparisons. Kaplan–Meier analysis was used for survival analysis and the survival difference between subgroups was compared by log-rank test. The Cox proportional hazards model was used for multivariate analysis. We considered calculated probabilities of P < 0.001 to be highly significant, P < 0.01 to be very significant, and P < 0.05 to be significant.

Detailed analyses were executed using dataset GSE6764 by R software. Hierarchical clustering with 346 upregulated genes and 497 downregulated genes were identified (Fig. 1A). According to GO and KEGG enrichment analysis, the genes correlated with HCC were involved in the cell cycle, pathways in cancer (such as Wnt, Jak-STAT, PI3K-Akt and so on), P53 signaling pathway, miRNAs in cancer, FoxO signaling pathway and so on (Fig. 1A; Supplementary Tables S1 and S2). The results also showed that the level of CKLF1 in HCC tissues was higher than that in adjacent nontumor tissues (P < 0.05; Fig. 1B). Data of Oncomine GSE14520 also verified that the level of CKLF1 in HCC tissues was significantly higher than normal liver tissues (Supplementary Fig. S1A). Furthermore, high expression of CKLF1 predicted the risk of vascular invasion and linked to late-stage HCC (Fig. 1C and E). Survival analysis showed that high CKLF1 expression was substantially correlated with poor OS (P < 0.05; Fig. 1F). In conclusion, CKLF1 might be a prognostic factor for HCC and play a critical role in the development and progression of HCC.

The STAT3 copy number level was also markedly elevated in HCC tissues (GSE32649; Fig. 1D). More importantly, the patients with higher copy number units of STAT3 were more likely to be with poor OS (Fig. 1G).

HCC samples and their corresponding nontumor tissues were collected to verify the results of bioinformatics analysis.

qPCR indicated that CKLF1 was highly expressed in 37 of 58 (63.79%) of HCC tissues and only in 18 of 58 (31.03%) of adjacent nontumor tissues (P < 0.001; Supplementary Fig. S1B; Supplementary Table S3).

Western blotting also indicated increased expression of CKLF1 in 40 of 58 (68.97%) of HCC samples, but only in 14 of 58 (24.14%) of nontumor tissues (P < 0.001; Fig. 2A; Supplementary Table S4).

The expression of CKLF1 was also assessed via IHC staining in another 46 pairs of HCC samples (Fig. 2B; Supplementary Table S5). The result showed that 29 of 46 (63.04%) of HCC samples and 10 of 46 (21.74%) of the corresponding nontumor tissues were positively stained (P < 0.001; Fig. 2B).

To examine the potential relationship between CKLF1 and STAT3, we detected the protein expression of CKLF1 and STAT3 in 58 HCC tissues by Western blotting and used Pearson correlation coefficient. Interestingly, there was a significant positive correlation between the expression of CKLF1 and STAT3 (P < 0.01; r = 0.4117; Fig. 2E).

The expression level of CKLF1 was also detected in five HCC cell lines by qPCR (Supplementary Fig. S1C) and Western blotting (Fig. 2F). CKLF1 was found to be expressed in all the HCC cell lines. Its expression was significantly higher in HepG2 and HCCLM3 cell lines, with spontaneous metastatic potentials.

In summary, both mRNA and protein expression levels of CKLF1 were markedly upregulated in HCC tissues and HCC cell lines, suggesting a strong correlation between HCC and the expression level of CKLF1.

To further investigate the role of CKLF1 in the development of HCC, the data were reviewed and reanalyzed. Patients with stage II–IV liver cancer showed higher expression level of CKLF1 compared with stage I liver cancer detected by Western blotting (P < 0.01; Fig. 2C). In IHC, 6 of 16 (37.5%) patients with HCC at stage I showed positive expression of CKLF1, whereas 23 of 30 (76.67%) HCC at stage II–IV were CKLF1 positive (P < 0.01; Fig. 2D; Supplementary Table S6). In general, the expression of CKLF1 protein was significantly higher in advanced stage HCC (stage II–IV) compared with low-stage HCC (stage I). Given that HCC in late stage presented stronger invasive capacity and metastatic potential compared with earlier stage (20), CKLF1 might be involved in the vascular invasion. The data showed that the level of CKLF1 in HCC was also positively correlated with vascular invasion and tumor size. There was no significant association between the expression of CKLF1 and age/gender (Supplementary Table S6).

High CKLF1 and AFP level, large tumor size, vascular invasion, and advanced TNM stage were all associated with worse OS in univariate analysis (Supplementary Table S7). However, it was not accurate enough to assess the correlation between CKLF1 level and other risk factors by univariate analysis. Thus, a Cox proportional hazards analysis was performed, which indicated that high CKLF1 level is an independent risk factor for worse OS (HR = 1.61; P = 0.017).

From the above findings, the enhanced expression of CKLF1 might contribute to the progression of HCC.

The measurement of metastatic potential is through migration and invasion, which are common assays used in cancer cell biology (21). RTCA, CCK-8, wound healing assay, and transwell chambers with or without Matrigel were performed to further substantiate the role of CKLF1 in vascular invasion and prognosis.

RTCA showed that CKLF1 enhanced the proliferation rate of NIH3T3 cells significantly (P < 0.01; Fig. 3A; Supplementary Fig. S2A) compared with the controls, indicating that CKLF1 promoted the proliferation of cells. Down regulation of CKLF1 proved the above result. The proliferation rate of HCCLM3 cells was inhibited by knocking down CKLF1 (P < 0.01; Fig. 3B; Supplementary Fig. S2A) compared with the controls. The results of CCK-8 also verified it. The proliferation rate of NIH3T3 cells transfected with CKLF1 (NIH3T3-CKLF1 cells) was significantly higher than the controls in 24, 36, and 48 hours after the transfection (Supplementary Fig. S3A–S3C), whereas the proliferation rate was decreased in HCCLM3 after downregulating the level of CKLF1 (Supplementary Fig. S3D–S3F).

While the control required 36 hours for the wound to heal by 50%, NIH3T3-CKLF1 cells required 24 hours (P < 0.001; Fig. 3C and D). Conversely, knockdown of CKLF1 strongly inhibited wound closure in HCCLM3 cells (P < 0.001; Fig. 3E and F). The results of transwell without Matrigel verified the positive effect on cell migration of CKLF1. The number of NIH3T3-CKLF1 cells, which migrated through the pores to the underside of the membrane was three times higher than the controls (P < 0.001; Fig. 3G and I). However, number of HCCLM3 cells that migrated through the polycarbonate membrane was reduced by 50% after knocking down CKLF1 (P < 0.05; Fig. 3G and I).

Furthermore, transwell chambers with Matrigel were used to confirm the cell invasion. The results showed that cell invasion was enhanced in NIH3T3-CKLF1 cells (P < 0.001; Fig. 3H and J). In striking contrast, knockdown of CKLF1 reduced cell invasion in HCCLM3 cells dramatically (P < 0.01; Fig. 3H and J). All the experiments above, including RTCA, wound healing assay, and transwell, were repeated in HepG2 cells and Bel7402 cells and the results were similar with previous studies (Supplementary Fig. S2B–S2I).

The result above suggested that CKLF1 might enhance HCC cells’ proliferation and metastatic potential, including the promotion of both cell migration and invasion.

The gold standard for determining the tumorigenicity remains malignant transformation in culture medium and tumor formation (xenografts) in nude mice (22). So we discovered the ability of CKLF1 to become neoplastically transformed and then to form tumors in nude mice.

NIH3T3-CKLF1 cells formed colonies 3-fold higher than the control (P < 0.01; Fig. 4A), suggesting CKLF1 could efficiently transform cells in the foci formation assay. The fact that it formed fewer foci after knocking down of CKLF1 in HCCLM3 proved this from another side (P < 0.01; Fig. 4B).

Following an in vivo tumorigenesis test in nude mice, the results showed that 7 of 8 (87.5%) nude mice injected with NIH3T3-CKLF1 cells formed tumors (P < 0.001; Fig. 4G; Supplementary Table S8), whereas tumors did not grow in nude mice injected with pcDNA3.1-containing cells at 30 days. The rate of formation of the subcutaneously implanted tumors dropped to 50% (4/8) in nude mice injected with HCCLM3 cells transfected with shCKLF1 (HCCLM3-shCKLF1 cells), compared with the control (100%, P < 0.001; Supplementary Table S9). Further studies found knockdown of CKLF1 in HCCLM3 cells resulted in significantly smaller tumor volume compared with the positive control (P < 0.001; Fig. 4G; Supplementary Fig. S4A; Supplementary Table S9). Hematoxylin and eosin staining of the tumor samples was performed to show malignant morphologic changes, including nuclear pleomorphism, large nucleoli and so on (Supplementary Fig. S4B). Terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) staining and proliferating cell nuclear antigen (PCNA) staining showed that the reduction of tumor volume was a result of apoptotic cell death and reduced proliferation (Fig. 4H–K). These results implicated that CKLF1 might have oncogenic potential in HCC.

The results from in vitro and in vivo studies, together with the analysis of patient specimens, collectively showed that CKLF1 could promote the development and progression of HCC. In recent years, evidence has accumulated that CKLF1 plays a crucial role in the immune and inflammatory responses (14–16). Therefore, whether CKLF1 might be involved in inflammation-mediated carcinogenesis of HCC was investigated in this study.

Interestingly, the results also revealed that STAT3, which participated in oncogenesis through upregulating the genes encoding apoptosis inhibitors and cell-cycle regulators such as Bcl-xl, MYC, and cyclins D1 (23), was positively related to CKLF1 in HCC tissues. The copy number of STAT3 in HCC tissues was significantly higher than that in normal liver tissues. Patients with HCC with a higher copy number of CKLF1 or STAT3 were more likely to relapse. The level of the STAT3-related cytokines, such as IL17A, TNFα, and VEGF, were upregulated in Bel7402-CKLF1 cells, whereas these genes were decreased after knocking down CKLF1 in HCCLM3 cells (Supplementary Fig. S5A and S5B).

The expression level of proteins involved in the STAT3 pathway was studied using qPCR and Western blotting (Fig. 5; Supplementary Fig. S5). It showed that CKLF1 increased the expression level of STAT3, p-STAT3 (Y705), Bcl-xl, and MYC in Bel7402 cells (P < 0.05; Fig. 5A), while downregulation of CKLF1 resulted in a decrease of these proteins (P < 0.05; Fig. 5B).

Investigative studies have shown that aberrant STAT3 activation in cancer cells is related to signal transduction from the IL6 (24–26). In this study, there was considerable upregulation of IL6 in Bel7402 cells transfected with CKLF1 (Bel7402-CKLF1 cells), whereas knockdown of CKLF1 in HCCLM3 cells inhibited the expression of IL6 (P < 0.05; Fig. 5C). After knocking down IL6 by IL6 siRNA, the proteins levels of STAT3, p-STAT3 (Y705), Bcl-xl, and MYC had been all downregulated in Bel7402-CKLF1 cells (P < 0.01; Fig. 5F).

Stattic, a small molecule, which selectively inhibited dimerization, activation, and nuclear translocation of STAT3, was used to inhibit the expression of p-STAT3 (Y705). Cells were exposed to Stattic for 24 hours, followed by Western blotting. The results showed that Stattic effectively decreased expression levels of p-STAT3 (Y705), Bcl-xl, and MYC, which were upregulated by CKLF1 (P < 0.01; Fig. 5E). All the above findings indicated that CKLF1 could activate the IL6/STAT3 signaling pathway.

Stattic was also used to investigate whether the STAT3 pathway was involved in CKLF1-mediated tumorigenesis. RTCA, transwell, foci-forming assay, and tumorigenesis test in nude mice were performed in Bel7402 cells transfected with CKLF1 after being treated with Stattic for 24 hours. We found that Stattic decreased cell survival and proliferation in Bel7402-CKLF1 cells. RTCA and transwell showed that Stattic inhibited cell proliferation, migration, and invasion in Bcl7402-CKLF1 cells (P < 0.05; Fig. 4L and M). As in Fig. 4C and F, foci formation assay demonstrated Stattic treatment resulted in markedly decreased clonogenic survival compared with control (P < 0.001). Tumor volume of Stattic-treated mice was smaller than control (P < 0.001; Fig. 4N; Supplementary Table S10).

Aberrant regulation of cell metabolism plays a central role in deterring the survival and growth features of malignant cells. Most tumor cells display a metabolic switch toward aerobic glycolysis, which is thought to favor the synthesis of essential cellular components required for fast cell duplication. STAT3 is found to act as a central regulator of cell metabolism, such as driving a metabolic switch toward aerobic glycolysis (27). STAT3 was regulated by CKLF1, suggesting that CKLF1 might also regulate aerobic glycolysis. The results demonstrated that the overexpression of CKLF1 increased the pyruvate kinase activity and lactate production (Fig. 5D and G), which were significantly impaired by Stattic in Bel7402 cell. In comparison, knockdown of CKLF1 inhibited the pyruvate kinase activity and lactate production in HCCLM3 cells.

Considering that CKLF1 activated the IL6/STAT3 pathway and the role of CKLF1 in the malignant transformation was inhibited by Stattic, we demonstrated that CKLF1 could enhance the carcinogenesis via the IL6/STAT3 pathway.

Doxorubicin, a common apoptosis inducer, is used as an antitumor drug. Of interest, the level of CKLF1 was downregulated by 60% significantly whereas the cell apoptosis rate was increased from 6.7% to 27.5% after treatment with doxorubicin (P < 0.001; Fig. 6A). HepG2 cells transfected with CKLF1 (HepG2-CKLF1 cells) were treated with doxorubicin for 24 hours to investigate the function of CKLF1 in doxorubicin -induced apoptosis. HepG2 cells transfected with pcDNA3.1 served as the negative control. Interestingly, the apoptotic rate of HepG2-CKLF1 cells was reduced by 30% compared with negative control (P < 0.05; Fig. 6A), suggesting that overexpression of CKLF1 promoted cell survival after doxorubicin treatment. On the basis of the above observations, CKLF1 could protect HCC cells from apoptosis.

We measured the rates of doxorubicin-induced apoptosis in HepG2-CKLF1 in the absence and presence of Stattic and found that the rate of doxorubicin-induced apoptosis in HepG2-CKLF1 cells increased after treatment with Stattic (P < 0.005; Fig. 6B). The results further indicated that STAT3 was involved in the doxorubicin-induced apoptosis, which was prevented by CKLF1. Considering that Stattic could inhibit doxorubicin-induced apoptosis prevented by CKLF1, we concluded that CKLF1 prevented doxorubicin-induced apoptosis through IL6/STAT3 signaling.

Then we detected the cell cycle of HepG2-CKLF1 cells after the treatment with doxorubicin (Fig. 6C). The data showed that HepG2-CKLF1 cells increased in the G₂–M-phase and decreased in the S phase compared with control cells (Fig. 6C). After the treatment with Stattic, the cells decreased in the G₂–M-phase and increased in the S phase (Fig. 6D), indicating that CKLF1 decreased the cells blocked in the S phase, and promoted proliferation in the G₂–M-phase through IL6/STAT3 signaling. In addition, the expression of IL6, STAT3, and p-STAT3 (Y705), as well as Bcl-xl and MYC, the downstream genes of the pathway, was downregulated by doxorubicin (P < 0.01; Fig. 6E and F). The expression of these proteins increased in HepG2-CKLF1 cells after doxorubicin treatment (P < 0.01; Fig. 6G and H). The results revealed that CKLF1 prevented the apoptosis induced by doxorubicin through the IL6/STAT3 pathway (Fig. 6I).

CKLF1 is a recently discovered chemokine with broad-spectrum biological functions in inflammation and autoimmune diseases. According to the presence of cysteine motifs near their amino-termini, chemokines are classified into four groups: CC chemokines containing two adjacent cysteine (C) residues, CXC chemokines containing two cysteines separated by one amino acid (X), CX3C chemokines containing three amino acids separating their cysteine residues, and C chemokines containing only one cysteine residue near the NH₂ terminus (28). Many chemokines have been demonstrated to be involved in hepatobiliary carcinogenesis. CCL2-CCR2 chemokine system has been reported to be a central activated pathway for the proneoplastic environment in HCC (28). CCL22 predisposes patients with HCC for the development of portal vein tumor thrombus (29). A recent study has confirmed that CCL17 enhances the stemness, epithelial–mesenchymal transition (EMT) process, and the TGF-β1 and Wnt/β-catenin signaling in MHCC97L cells and promotes the tumorigenesis of HCC (30).

CCL17 and CCL22, as well as CKLF1 have a CC motif in its C-terminal region and belong to the CC chemokine family (31). It is a potent chemoattractant for neutrophils, monocytes, and lymphocytes and is located in a cluster on chromosome 16. Three other splice variants of the CKLF1 gene, CKLF2, CKLF3, and CKLF4, have also been isolated (12). CKLF1 can stimulate the proliferation of human bone marrow hematopoietic progenitor cells and colony formation, whereas CKLF2 accelerates myoblast proliferation (32). Several studies have demonstrated that the expression of CKLF1 increases in inflammation and autoimmune diseases like rheumatoid arthritis and asthma (31, 33). Furthermore, CKLF1 displays chemotactic activities in neuroblastoma SHSY5Y cells and promotes the migration of rat primary cortical neurons (34, 35). However, there is no report about the relationship between CKLF1 and cancers, including HCC.

In this study, we found that the level of CKLF1 was significantly higher than that in normal livers and the increased CKLF1 expression was associated with poor survival. The expression of CKLF1 in HCC tissues was related to HCC stage, vascular invasion, and tumor size. These findings suggested a strong correlation between HCC and the expression level of CKLF1. Clinical data further confirmed that the level of CKLF1 in HCC tissues was considerably higher than that in the adjacent nontumor tissues and the expression of CKLF1 was much higher in advanced stage tumors compared with low-stage tumors. In addition, HCC patients with a higher level of CKLF1 were more likely to be diagnosed with worse clinical outcomes. Taken together, it suggested that CKLF1 might have functions in the development of HCC.

Previous studies have reported that CKLF1 promotes the migration of rat primary cortical neurons (34, 35), indicating the chemotactic activity of CKLF1. Overexpression of CKLF2 induces the proliferation and survival of skeletal muscle cells (32). To investigate the effect of CKLF1 in HCC, we explored whether CKLF1 could enhance the malignant properties of HCC cells. The data suggested that the proliferation rate of cells significantly increased after transfection with CKLF1. The results of wound healing assay and transwell assay demonstrated that CKLF1 enhanced migration and invasion of HCC cells distinctly. Furthermore, the tumorigenic ability of NIH3T3 was promoted by CKLF1, whereas knocking down CKLF1 inhibited these malignant properties in the HCCLM3 cells. All these revealed that CKLF1 might promote carcinogenesis and progression in HCC.

One of the mechanisms that accounts for the tumor promotion is low-grade inflammation (36). A substantial body of evidence supports the conclusion that chronic inflammation can promote cancer. Current estimates suggest that chronic inflammation accounts for approximately 25% of cancer (37). The persistent presence and amplification of inflammation can predispose cells to oncogenic transformation and contribute to tumor development (38, 39), including colon cancer, gastric cancer, and HCC (40). Recent researches have shown that the tumor inflammatory microenvironment plays an essential role in the progression of HCC, such as the process of liver fibrosis, hepatocarcinogenesis, EMT, tumor invasion, and metastasis (41).

Various chemokines, which are involved in the inflammatory processes, orchestrate the interaction among parenchymal liver cells, Kupffer cells, hepatic stellate cells, endothelial cells, and infiltrating immune cells. Consequently, these cellular interactions result in the remodeling of the hepatic microenvironment toward a proinflammatory, profibrotic, proangiogenic, and thus preneoplastic milieu (28, 38). As a member of chemokines, CKLF1 plays a crucial role in the immune and inflammatory responses. Taken together with the results that CKLF1 was involved in the progression of HCC and had oncogenic potential in HCC, it was highly plausible that CKLF1 might function as a tumor promoter in the inflammation-associated carcinogenesis and tumor progression in HCC.

IL6, one of the tumor-promoting cytokines, causes hepatic inflammation, positively correlates with the progression of HCC, and activates the oncogenic factor STAT3 (42–44). STAT3 is the most crucial molecules in the IL6 signaling pathway and is recognized as a pivotal link between inflammation and cancer (24–26). IL6 exerts the proliferative effect via STAT3, which further synergizes with NF-κB to increase the expression of survival genes. STAT3 is constitutively active in many tumors, participates in tumor progression, and regulates the expression of genes involved in cell survival, proliferation, invasion, and angiogenesis (45). Simvastatin suppresses HCC cell growth by reducing the Skp2 expression to cause p27 accumulation and induces G₀–G₁-phase arrest via STAT3 inhibition (46), and STAT3 can suppress the function of P53 (47). Bioinformatics prediction and analysis revealed that the upregulated genes in GSE6764 might be involved in cell cycle, P53 signaling pathway, miRNAs in cancer, and other signaling pathways in cancer, such as Wnt, Jak-STAT, PI3K-Akt pathways and so on. Interestingly, STAT3 is an influential partner in all the signaling pathways above, and the correlation between CKLF1 and STAT3 was significant. Therefore, as a result of this bioinformatics analysis, it could be concluded that there was a significant positive relationship between CKLF1 and IL6/STAT3 signaling pathway.

STAT3 activation requires phosphorylation of a critical tyrosine residue (Y705), which mediates its dimerization, a prerequisite for nucleus entry and DNA binding. STAT3 participates in oncogenesis through upregulation of genes encoding apoptosis inhibitors and cell-cycle regulators such as Bcl-xl, MYC, and cyclins D1 (23). In this study, data showed that CKLF1 regulated the expression of IL6, STAT3, p-STAT3 (Y705), STAT3-related cytokines, and the downstream genes of STAT3, indicating that CKLF1 could activate the IL6/STAT3 pathway.

Stattic, the inhibitor of p-STAT3 (Y705), is a small molecule shown to selectively inhibit the function of the STAT3 SH2 domain and inhibit STAT3 activation, dimerization, and nuclear translocation. In this article, the proliferation rate of Bel7402-CKLF1 cells markedly decreased and the number of Bel7402-CKLF1 cells which migrated through the pores to the underside of the membrane significantly reduced after treatment with Stattic. The results also showed that after treatment with Stattic, Bel7402 cells transfected with CKLF1 almost failed to form foci, and the average of tumor volumes in nude mice was smaller than the control. Moreover, the upregulation of the downstream genes of STAT3, induced by CKLF1, was suppressed by Stattic. Similar results occurred in Bel7402 cells cotreatment of CKLF1 and IL6 siRNA. In summary, this study indicated that CKLF1 might promote carcinogenesis and progression in HCC through the IL6/STAT3 signaling pathway.

Doxorubicin, a chief chemotherapeutic agent in cancer treatment, can inhibit the progression of topoisomerase II, induce the DNA damage, and trigger apoptosis (48, 49). After the treatment with doxorubicin, the apoptosis rate of HepG2 cells increased dramatically, accompanied by significantly downregulating the level of CKLF1 and suppressing the IL6/STAT3 pathway. On the contrary, the percentage of apoptotic HepG2-CKLF1 cells reduced and the suppression of IL6/STAT3 signaling was weakened comparing with the HepG2 control cells after being treated with doxorubicin. Interestingly, the rate of doxorubicin-induced apoptosis in HepG2-CKLF1 cells increased after treatment with Stattic. The data indicated that CKLF1 protected HCC cells from doxorubicin-induced apoptosis through the IL6/STAT3 signaling pathway. Combined with the previous result that CKLF1 promoted HCC cells proliferation, migration, and invasion, CKLF1 might influence the curative effect of anticancer drugs, and even cause relapses.

As CKLF1 shares amino acid sequence similarity with CCL17 and CCL22, two of cognate ligands for CCR4, it is thought as a novel functional ligand for CCR4 (31). We hypothesized that CKLF1 initiated the series of functions through CCR4.

Up till now, the diagnosis and prognosis for HCC are still very poor, owing to the high metastasis rate and the limitations of current treatment modalities. Therefore, a better understanding of the molecular mechanisms involved in the initiation and development of HCC and discovering novel therapeutic targets are necessary and particularly important. Due to the pivotal role in the initiation and development of HCC, CKLF1 may be a promising target for further study.

In conclusion, we reported for the first time that high expression of CKLF1 was related to advanced-stage HCC and predicted vascular invasion and poor prognosis in patients with HCC. Studies both in vivo and in vitro showed that CKLF1 inhibited apoptosis and promoted proliferation, migration, invasion, metastasis, and malignant transformation in the carcinogenesis of HCC by activating the IL6/STAT3 signaling pathway. Due to the essential role of CKLF1 in inflammation-mediated hepatocarcinogenesis, it is attractive and useful to be exploit for the improvement of early diagnosis and better prognosis of HCC. The results from this study are eagerly awaited to enrich the mechanism of inflammation-mediated carcinogenesis and the treatment of HCC.

No potential conflicts of interest were disclosed.

Conception and design: F. Zhu

Development of methodology: Y. Liu

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y. Liu, Y. Zhou, X. Chen, F. Zhu

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Liu, P. Zhou, Q. Yan, S. Ding

Writing, review, and/or revision of the manuscript: Y. Liu, F. Zhu

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y. Liu, L. Liu, F. Zhu

Study supervision: F. Zhu

Many thanks to Zhi Ren and Anushesh Dhakal for improving the writing style and facilitate clarity of the article. This work was supported in part by the National Natural Science Foundation of China (81772196, 81502418, and 31470264) and Chinese Foundation for Hepatitis Prevention and Control (TQGB 20170068).

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.