Heat shock factor 1 promotes invasion and metastasis of hepatocellular carcinoma in vitro and in vivo




Heat shock factor 1 (HSF1) is a powerful, multifaceted modifier of carcinogenesis. However, the clinical significance and biologic function of HSF1 in hepatocellular carcinoma (HCC) remain unknown.


Quantitative reverse transcriptase-polymerase chain reaction analysis, Western blot analysis, and immunohistochemical staining were used to detect expression levels of HSF1, and its correlation with clinicopathologic parameters and the prognosis for patients with HCC were analyzed. In addition, the biologic function and molecular mechanisms of HSF1 in HCC were investigated in vitro and in vivo.


HSF1 levels were elevated predominantly in HCC, especially in venous emboli from HCC (P < .05), and high expression levels of HSF1 were correlated significantly with multiple nodules, venous invasion, absence of capsular formation, and high Edmondson-Steiner grade as well as poor overall survival and disease-free survival in patients with HCC (P < .05). Multivariate Cox regression analysis revealed that high HSF1 expression was an independent prognostic factor for overall survival in patients with HCC (relative risk, 4.874; P < .001). Finally, HSF1 was capable of promoting HCC cell migration and invasion in vitro and in vivo by facilitating the expression and phosphorylation of heat shock protein 27.


Collectively, the current findings suggested that HSF1 may serve as a novel prognostic marker and therapeutic target for HCC. Cancer 2012;. © 2011 American Cancer Society.


Hepatocellular carcinoma (HCC) ranks as the fifth most common human malignancies and the third leading cause of cancer death worldwide, especially in China, East Asia, and South Africa.1, 2 Hepatic resection is the first choice for patients with HCC and has evolved into a safe procedure with low operative mortality, which is <2% in large liver surgery centers.3, 4 However, the long-time survival of patients with HCC after hepatic resection remains unsatisfactory, mainly because of the high incidence of recurrence and metastasis, with 5-year actuarial recurrence rates from 75% to 100% reported in the literature.5, 6 The recurrence and metastasis of HCC is a multistep process that involves complex biologic and pathologic events.7, 8 In past decades, various molecules have been reported to play a role in recurrence and metastasis of HCC, such as osteopontin; ras homolog gene family, member C (RhoC); epidermal growth factor-like domain, multiple 7 (Egfl1); micro-RNAs (miRNAs); and so on.9-14 Despite extensive clinical as well as basic research efforts, the molecular mechanisms for HCC metastasis largely remain unknown. Therefore, identifying novel metastatic factors and understanding the molecular mechanism underlying the progression of metastasis are important for HCC prevention and treatment.

Heat shock factor 1 (HSF1) is a transcription factor that is conserved in eukaryotic cells and is activated in response to heat shock and other forms of environmental and chemical stress.15-17 Upon activation, HSF1 trimerizes, translocates to nucleus, and binds to the heat shock element (HSE) sites of promoters of heat shock protein (HSP) genes in a hyperphosphorylated form that is competent to activate transcription.18, 19 Elevated levels of HSP27, HSP70, and HSP90 have been observed in many types of tumors, indicating the crucial roles of HSPs in tumor development.20 Thus, as a master regulator of HSPs, the roles of HSF1 in carcinogenesis and progression are now emerging.

Recently, Dai et al21 reported that Hsf1−/− mice were far more resistant to tumor formation driven by 7,12-dimethylbenz(a)anthracene (DMBA) plus tissue plasminogen activator (TPA), H-Ras, or mutant p53 (p53R172H) than Hsf1+/+ mice. Another in vivo study indicated that p53−/− mice developed spontaneous lymphomagenesis that was suppressed in Hsf1−/−/p53−/− mice, which lost Hsf1-dependent function.22 Meng et al23 also observed that HSF1 was required for human mammary epithelial MCF-10A cell transformation and tumorigenesis induced by the human epidermal growth factor receptor-2 (HER2) oncogene. However, HSF1 is not a classic oncogene, because the overexpression of HSF1 was unable to transform immortalized mouse embryonic fibroblasts (MEFs); instead, MEFs that lacked HSF1 were refractory to transformation induced by either oncogenic H-RasV12D or platelet-derived growth factor β (PDGF-B).21 These data suggest that HSF1 is a powerful, multifaceted modifier of carcinogenesis. Moreover, HSF1 plays a role in cell migration, and immortalized MEF cells derived from Hsf1−/− animals were deficient in both basal and epidermal growth factor (EGF)-induced migration.24 On the basis of the finding that the cancer recurrence and metastasis are attributable to a large extent to the ability of cells to migrate,25-27 HSF1 also may be involved in the regulation of cancer metastasis.

However, it is unclear whether and how HSF1 contributes to the carcinogenesis and development of HCC, and the role of HSF1 in metastasis of human malignancies has not been clearly defined. Therefore, we carried out the current study to determine the expression of HSF1 in human HCC tissues and to establish its clinical relevance. Furthermore, the function and molecular mechanisms of HSF1 in HCC invasion and metastasis were investigated by using in vitro and in vivo models.


Patients and Tissue Specimens

In the current study, HCC tissue specimens were obtained from 213 patients with HCC who underwent hepatectomy at the Department of Surgery, Xiangya Hospital of Central South University (CSU) from November 1998 to December 2005. These patients included 160 men and 53 women, and the median patient age was 47 years (range, 16-73 years). Among these 213 patients with HCC, matched fresh specimens of HCC tissues and adjacent nontumorous liver tissue (ANLT) from 30 patients were collected, immediately frozen in liquid nitrogen, and subsequently stored at −80°C for real-time quantitative reverse-transcription polymerase chain reaction (qRT-PCR) and Western blot analyses. All specimens were embedded in paraffin and stained with hematoxylin and eosin. The diagnosis in each patient was confirmed by histopathology that was blinded to examination. Previous informed consent was obtained from the patients for the collection of liver specimens in accordance with the guidelines of Xiangya Hospital, and the study protocols were approved by the CSU Ethics Committee.

Quantitative Reverse-Transcription Polymerase Chain Reaction

Real-time qPCR was performed as described previously.28 The following primers for HSF1 were used: forward, 5′-ACCCATGCTTCCTGCGTGGC-3′; and reverse, 5′-TGCTTCTGCCGAAGGCTGGC-3′. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) expression was determined as a control using the following primers: forward, 5′-GCACCGTCAAGGCTGAGAAC-3′; and reverse, 5′-TGGTGAAGACGCCAGTGGA-3′. The results were analyzed using the 2−(ΔΔCt) method according to the following formula: ΔΔCt = (CtHCC − CtGAPDH)/(CtANLT − CtGAPDH).

Western Blot Analysis

Total protein was extracted by using a lysis buffer (20 mmol/L Tris-HCl, pH 7.4; 10 mmol/L NaCl; 1 mmol/L ethylene diamine tetracetic acid, pH 8.0; 1 mmol/L MgCl2; 1% NP-40; 0.1% sodium dodecyl sulfate [SDS]; and 0.01% phenylmethylsulfonyl fluoride [Sigma, St. Louis Mo]) and protease inhibitor (Promega, Madison, Wis), and the protein was separated by SDS-polyacrylamide gel electrophoresis and then transferred onto polyvinylidine fluoride membranes (Millipore, Bedford, Mass). The blotted membranes were incubated with antihuman HSF1 antibody (Cell Signaling Technology, Danvers, Mass), anti-M5 antibody (Sigma), antihuman Hsp27 antibody (Santa Cruz Biotechnology, Santa Cruz, Calif), or antihuman phosphorylated Hsp27 antibody (Santa Cruz Biotechnology) and then secondary antibody (KPL, Gaithersburg, Md), in that order. β-Actin protein levels also were determined by using the specific antibody (Sigma) as a loading control.


Immunohistochemical staining of formalin-fixed paraffin sections (4 μm thick) was performed as described previously by using antihuman HSF1 antibody (1:250 dilution; Cell Signaling Technology).11 Immunohistochemical staining was scored according to the percentage of positive hepatocytes on a 4-point scale as follows: 0, <10% positive hepatocytes; 1+, 11% to 25% positive hepatocytes; 2+, 26% to 50% positive hepatocytes; and 3+, >51% positive hepatocytes.8 Thus, the protein expression of HSF1 in HCC specimens was divided into a low-expression group (0 or 1+ expression) and a high-expression group (2+ or 3+ expression). Immunohistochemical analysis and scoring were performed by 2 independent investigators.

Follow-Up and Prognostic Study

Follow-up data were obtained for all 213 patients. The follow-up period was defined as the interval between the date of operation and the date of the patient's death or the last follow-up. The median follow-up was 25 months (range, 12-98 months). Deaths from other causes were treated as censored events. Recurrence and metastasis were diagnosed by clinical examination, serial α-fetoprotein level measurements, and ultrasonography or computed tomography (CT) scans. Disease-free survival was defined as the length of time after hepatectomy during which a patient survived without signs of HCC. Data from 8 conventional clinical and pathologic variables also were collected for analysis, including age, sex, cirrhosis, Edmondson-Steiner grade, capsular formation, tumor size, the number of tumor nodes, and venous invasion.

Cell Lines and Cell Culture

The HepG2 and 293T cell lines were purchased from the American Type Culture Collection (Rockville, Md). The MHCC97-L and HCCLM3 cell lines were purchased from the Liver Cancer Institute of Fudan University (Shanghai, China). These cells were cultured in Dulbecco modified Eagle medium supplemented with 10% fetal bovine serum and antibiotics at 37°C with 5% CO2.

Vector Construction

The DNA fragment of HSF1 was amplified from genomic DNA and inserted into the EcoRI/SalI site of retrovirus expression vector pBabe-puro (Addgene, Cambridge, Mass). The following oligonucleotide PCR primers were used: forward, 5′-CAGTGAATTCACCATGGACTACAAGGACGACGATGACAAGATGGATCTGCCCGTGG-3′; and reverse, 5′-CAGTGTCGACCTAGGAGACAGTGG GGTCCT-3′. The Hsp27 expression vector pGCL-hsp27 was constructed by inserting its open reading frame sequence into the lentivirus expression vector pGCL (GeneChem, Shanghai, China). The short interfering RNA (siRNA)-expressing vector pLKO.1-puro was purchased from Addgene. Two putative candidate sequences were designed by using OligoEngine software (OligoEngine, Seattle, Wash), and their specificity was confirmed by nucleotide BLAST searches. The 2 putative candidate sequences for HSF1 were as follows (the 19 nucleotide sense or antisense strands are provided, and the stem loop sequences are indicated in italics): For HSF1 shRNA sequence 700, the sense sequence was 5′-CCGGGCAGGTTGTTCATAGTCAGAACTCGAGTTCTGACTATGAACAA CCTGCTTTTT-3′, and the antisense sequence was 5′-AATTAAAAAGCAGGTT GTTCATAGTCA GAACTCGAGTTCTGACTATGAA CAACCTGC-3′. For HSF1 shRNA sequence 1999, the sense sequence was 5′-CCGGCCAGCAACAGAAAGTC GTCAACTCGAGTTGACGACTTTCTGT TGCTGG TTTTT-3′, and the antisense sequence was 5′-AATTAAAAACCAGCAACAGAAAGTCGTCAA CTCGAGT TGACGACTTTCTGTTGCTGG-3′. For Hsp27, the sequences were as described previously29: The sense sequence was 5′-CCGGATCCGATGAGACTGCCGCCAACTCGAGTTGG CGGCAGTCTCATCGGATTTTTT-3′, and the antisense sequence was 5′-AATTAAAAAATCCG ATGAGA CTGCCGCCAACTCGAGTTGGCGGCAGT CTCATCGGAT-3′.

In Vitro Wound-Healing (Migration) and Invasion Assays

For the wound-healing assay, cells were seeded into 35-mm dishes (Corning Inc., Corning, NY) precoated with fibronectin. After cells reached 100% confluence, wound-healing assays were performed with a sterile pipette tip to make a scratch through the confluent monolayer. Then, after a wash in fresh medium, the cells were cultured for another 48 hours. The percentage wound closure was calculated for 5 randomly chosen fields. For the invasion assay, 1 × 105 cells in serum-free medium containing 0.1% bovine serum albumin were placed into the upper chamber of the insert with Matrigel (BD Biosciences, Woburn, Mass). After a 48-hour incubation at 37°C, the cells remaining in the upper chamber or on the upper membrane were removed with a cotton swab. After staining with a solution that contained 0.1% crystal violet and 20% methanol, the number of cells that adhered to the lower membrane of the inserts was counted. For each experimental group, the invasion assay was performed in triplicate, and 3 random fields in each replicate experiment were chosen for cell number quantification.

Immunofluorescence for Cell Skeleton

F-actin immunofluorescence staining was used to analyze the cell skeleton. The cells growing on cover slides were fixed and treated with cold methanol for 10 minutes and then incubated with fluorescein isothiocyanate-conjugated phalloidin (Beyotime Institute of Biotechnology, Jiangsu, China). F-actin filaments were observed and analyzed with a Nikon 108 fluorescence microscope (Nikon Corporation, Tokyo, Japan).

Hepatocellular Carcinoma Metastatic Mouse Model

An HCC model in mice was constructed as described previously.11 Briefly, 5 × 106 cells from different experimental groups were injected subcutaneously into the left upper flank regions of nude mouse (3-4 weeks of age, male, BALB/c). After 1 month, the subcutaneous tumor tissues were removed and implanted into the liver of nude mice (5 in each group). After 6 weeks, the mice were killed; and their livers and lungs were dissected, fixed with phosphate-buffered neutral formalin, sectioned serially, and stained with hematoxylin and eosin for standard histologic examination. The mice were manipulated and housed in the Animal Institute of CSU according to the protocols approved by the Medical Experimental Animal Care Commission.

Statistical Analysis

The software package SPSS 13.0 for Windows (SPSS Inc., Chicago, Ill) was used for statistical analyses. The Fisher exact test was used for statistical analysis of categorical data, and independent t tests were used to analyze continuous data. Survival curves were constructed using the Kaplan-Meier method and were evaluated using the log-rank test. P values < .05 (2-tailed) were considered statistically significant.


Heat Shock Factor 1 Levels Are Highly Elevated in Hepatocellular Carcinoma Tissues and Are Correlated With the Metastatic Potential of Hepatocellular Carcinoma

Messenger RNA (mRNA) and protein expression levels of HSF1 were detected in 30 paired HCC tissues and ANLTs. The real-time qRT-PCR analysis results indicated that HSF1 mRNA levels in HCC tissues were significantly higher than those in ANLTs, with a median up-regulation fold of 4.6 (range, 0.75-18.0) (Fig. 1A). Consistent with mRNA expression, HSF1 protein expression levels also were highly elevated in HCC tissues compared with levels in the corresponding ANLTs (relative protein expression levels: 1.50 ± 0.32 vs 0.38 ± 0.18; P = .006) (Fig. 1B). Even HSF1 protein levels in ANLTs were slightly higher than those in normal liver tissues, although the difference was not statistically significant (P > .05) (Fig. 1B). In addition, compared with levels in ANLTs and primary HCC lesions, a venous embolus of HCC had the highest HSF1 mRNA and protein expression levels (relative mRNA expression levels: 1.0, 5.1, and 10.2, respectively; relative protein expression levels [±standard error]: 0.1 ± 0.05, 0.5 ± 0.12, and 1.2 ± 0.25, respectively) (Fig. 1C), suggesting a positive correlation between HSF1 expression and metastasis from HCC.

Figure 1.

Heat shock factor 1 (HSF1) is up-regulated significantly in hepatocellular carcinoma (HCC), and especially in metastatic HCC tissues. (A) Real-time quantitative reverse transcriptase-polymerase chain reaction results are shown for analyses of HSF1 messenger RNA (mRNA) expression in 30 paired HCC tissues (T) and adjacent nontumorous liver tissue (ANLTs). Fold inductions were calculated using the formula 2−(ΔΔCt). NT indicates normal liver tissue. (B) HSF1 protein expression levels were detected by Western blot analysis. β-Actin served as a loading control. (C) The expression levels of HSF1 in ANLTs, a primary HCC lesion, and a venous embolus from HCC (M) are illustrated.

Correlations of Heat Shock Factor Expression With Clinicopathologic Characteristics and Prognosis of Hepatocellular Carcinoma

According to the immunohistochemistry results (Fig. 2A-D), the expression of HSF1 in HCC tissues was divided into a low-expression group (n = 108) and a high-expression group (n = 105), and correlations between HSF1 expression and clinicopathologic characteristics and the prognosis of patients with HCC were analyzed. We observed that HSF1 expression levels were significantly higher in HCCs with multiple nodules (P = .010), without capsular formation (P = .007), with high Edmondson-Steiner grade (P = .025), or with venous invasion (P = .005) (Table 1). In addition, patients who had HCC with high HSF1 expression had worse overall survival (median, 25.3 months vs 66.7 months; P < .001) (Fig. 2E) and disease-free survival (median, 25 months vs 56 months; P < .001) (Fig. 2F) than patients in the low-expression group. A multivariate Cox regression analysis indicated that high HSF1 expression (relative risk, 4.874; P < .001) was an independent prognostic factor for overall survival (Table 2).

Figure 2.

Immunohistochemistry of heat shock factor 1 (HSF1) expression in hepatocellular carcinoma (HCC) tissues and its prognostic implications are illustrated. (A-D) In these representative images, HSF1 expression is observed (A) in >51% of cancer cells (scored as 3+), (B) in 26% to 50% of cancer cells (scored as 2+), (C) in 11% to 25% of cancer cells (scored as 1+), and (D) in a negative control, which was included to demonstrate the specificity of the antibody (original magnification, ×400 in A-D). On the basis of results from immunohistochemical staining, the expression of HSF1 was classified into a low expression group (0 or 1+; n = 108) and a high expression group (2+ or 3+; n = 105). Curves illustrating (E) overall survival and (F) disease-free survival were constructed using the Kaplan-Meier method and were evaluated using the log-rank test.

Table 1. Correlations Between Heat-Shock Factor 1 Expression and Clinicopathologic Variables in 213 Patients With Hepatocellular Carcinoma
  HSF1 Expression Level 
Clinicopathologic VariableNo. of PatientsLowHighP
  • Abbreviations: HSF1, heat-shock factor 1.

  • a

    Significant P value.

Age, y    
Tumor size, cm    
No. of tumor nodules    
 Multiple: ≥2773047.010a
Capsular formation    
Edmondson-Steiner grade    
Venous invasion    
Table 2. Multivariate Analysis With a Cox Proportional Hazards Regression Model
  Univariate AnalysisMultivariate Analysis
VariableNo. of PatientsRR (95% CI)PRR (95% CI)P
  • Abbreviations: CI, confidence interval; HSF1, heat-shock factor 1; RR, relative risk.

  • a

    Significant P value.

 Men1601.00 1.00 
 Women530.966 (0.637-1.465).8700.760 (0.468-1.233).266
Age, y     
 ≤601861.00 1.00 
 >60270.754 (0.469-1.211).2420.873 (0.519-1.470).611
 Presence1621.00 1.00 
 Absence510.867 (0.576-1.305).4930.791 (0.482-1.296).351
Tumor size, cm     
 ≤5901.00 1.00 
 >51231.160 (0.812-1.657).4161.253 (0.809-1.939).312
No. of tumor nodules     
 Solitary1361.00 1.00 
 Multiple: ≥2772.040 (1.324-3.142).001a2.287 (1.355-3.859).002a
Capsular formation     
 Presence1291.00 1 
 Absence841.660 (1.165-2.365).005a3.479 (1.966-6.157)<.001a
Edmondson-Steiner grade     
 1-21181.00 1.00 
 3-4950.945 (0.660-1.353).7571.107 (0.592-2.071).749
Venous invasion     
 Presence1111.00 1.00 
 Absence1020.557 (0.385-0.805).002a0.222 (0.135-0.364)<.001a
HSF1 expression     
 Low1081.00 1.00 
 High1053.720 (2.412-5.739)<.001a4.874 (3.089-7.691)<.001a

Heat Shock Factor 1 Promotes Hepatocellular Carcinoma Cell Migration and Invasion in Vitro

In view of the finding that HSF1 was overexpressed in HCC tissues and was highly correlated with the metastatic potential of HCC, we sought to determine the biologic functions of HSF1 in HCC metastasis. First, we compared the expression of HSF1 in 3 HCC cell lines (HepG2, MHCC97-L, and HCCLM3) with different metastatic potential.30 Among the 3 cell lines analyzed, HCCLM3 cells, which had the strongest metastatic ability, had the highest mRNA and protein expression levels of HSF1, followed by MHCC97-L cells, and then HepG2 cells (P < .05) (Fig. 3A). In addition, there was no statistically significant difference between the expression levels of HSF1 in MHCC97-L cells and HepG2 cells (P > .05) (Fig. 3A).

Figure 3.

Heat shock factor 1 (HSF1) has a positive correlation with the metastatic potential of hepatocellular carcinoma (HCC) cells. (A) Expression levels of HSF1 in HepG2 cells, MHCC97-L cells, and HCCLM3 cells were detected by real-time quantitative reverse transcriptase-polymerase chain reaction and Western blot analyses. mRNA indicates messenger RNA. (B) MHCC97-L cells were infected with pBabe-HSF1 retrovirus or control vector; then, after selection with puromycin (2 μg/mL) for 48 hours, HSF1 protein expression levels were detected. Flag indicates polypeptide protein tag. (C) HCCLM3 cells were infected with lentivirus containing HSF1 short-hairpin RNA sequence 700 (HSF1-700), HSF1-1999, or control virus. The expression of endogenous HSF1 was confirmed by Western blot analysis. The results indicated that HSF1-1999 had >80% inhibitory efficiency.

Next, MHCC97-L cells (with low metastatic ability) and HCCLM3 cells (with high metastatic ability) were selected for further study. The HSF1 stably overexpressed MHCC97-L cells and control cells were termed MHCC97-LHSF1 and MHCC97-Lcontrol, respectively, as indicated in Figure 3B. Western blot results revealed that shRNA sequence 1999 inhibited HSF1 protein expression >80%, whereas shRNA sequence 700 resulted in only 30% to 40% inhibition efficiency (Fig. 3C). Then, for convenience, HCCLM3 cells that were infected with a control sequence or with a lentivirus containing shRNA 1999 were termed HCCLM3control cells and HCCLM3HSF1RNAi+ cells, respectively.

The effects of HSF1 on HCC cell migration and invasion then were determined by using wound-healing (migration) and invasion assays in vitro. The wound-healing assay revealed that the closure of MHCC97-LHSF1 cells was significantly faster than the closure of MHCC97-Lcontrol cells (95% vs 55%; P = .003) (Fig. 4A), whereas HCCLM3HSF1RNAi+ cells closed much more slowly than HCCLM3control cells (43% vs 90%; P = .008) (Fig. 4A). Consistent with the wound-healing assay results, data from the invasion assay indicated that the numbers of MHCC97-LHSF1 cells and HCCLM3control cells that passed through the Matrigel (±standard deviation) were much greater than the numbers of MHCC97-Lcontrol cells and HCCLM3HSF1RNAi+ cells (234 ± 40 vs 96 ± 25 [P = .007] and 144 ± 30 vs 39 ± 18 [P = .006], respectively) (Fig. 4B). These results suggested that HSF1 can significantly promote HCC cell migration and invasion in vitro.

Figure 4.

Heat shock factor 1 (HSF1) promotes hepatocellular carcinoma (HCC) cell migration and invasion in vitro. (A) Wound-healing and (B) invasion assays are illustrated in MHCC97-Lcontrol cells, MHCC97-LHSF1 cells, HCCLM3control cells, and HCCLM3HSF1 RNA interference-positive (HCCLM3HSF1RNAi+) cells. (C) F-actin filaments were observed in cells using fluorescein isothiocyanate-conjugated phalloidin staining.

Considering the crucial role of the cell skeleton in cell motility, we examined the pattern and morphology of F-actin using phalloidin-fluorescein isothiocyanate. Enforced expression of HSF1, as indicated in Figure 4C, can stimulate the reorganization of actin, leading to the formation of stress fiber-like structures in MHCC97-LHSF1 cells, and HSF1 knockdown vanished these structures in HCCLM3HSF1RNAi+ cells compared with HCCLM3control cells. These findings indicate that HSF1 promotes HCC cell migration and invasion, possibly by regulating actin reorganization.

Heat Shock Factor 1 Facilitates Hepatocellular Carcinoma Cell Migration and Invasion Through Heat Shock Protein 27

Because of the important role of Hsp27 in the regulation of actin filament dynamics1, 32 and as a downstream target of HSF1, we investigated whether Hsp27 involved in HSF1 induced HCC cell migration and invasion. Our results demonstrated that HSF1 can facilitate the expression and phosphorylation of endogenous Hsp27 in both MHCC97-L cells and HCCLM3 cells (Fig. 5A). To verify whether the function of HSF1 in migration and invasion was Hsp27 dependent or independent, we silenced and restored Hsp27 expression in MHCC97-LHSF1 cells and HCCLM3HSF1RNAi+ cells, respectively (Fig. 5B), and performed the wound-healing and invasion assays again. The results indicated that the silencing of Hsp27 in MHCC97-LHSF1 cells abrogated an HSF1-induced increase in cell migration and invasion abilities, whereas the reintroduction of Hsp27 into HCCLM3HSF1RNAi+ cells restored the migration and invasion abilities of these cells (Fig. 5C). In addition, the F-actin cytoskeleton remodeling of these 2 cell lines also was abolished by silencing or reintroducing Hsp27 (Fig. 5D). These results suggested that HSF1 facilitates HCC cell migration and invasion in an Hsp27-dependent manner.

Figure 5.

Heat shock factor 1 (HSF1) facilitates hepatocellular carcinoma (HCC) cell migration and invasion through heat shock protein 27 (Hsp27). (A) Endogenous Hsp27 and phosphorylated Hsp27 (p-Hsp27) expression levels after HSF1 overexpression or knockdown were measured by Western blot analysis. (B) MHCC97-LHSF1 cells and HCCLM3HSF1 RNA interference-positive (HCCLM3HSF1RNAi+) cells were infected with lentivirus containing an Hsp27 short hairpin RNA (shRNA) sequence or an Hsp27 open reading frame sequence, respectively. Western blot analysis was performed to detect expression levels of Hsp27 and p-Hsp27. (C) Wound-healing and invasion assays and (D) F-actin staining were performed on the cells indicated in A and B.

Heat Shock Factor 1 Promotes Hepatocellular Carcinoma Cell Invasion and Metastasis in Vivo

To validate the observations obtained from in vitro studies, we examined the in vivo role of HSF1 in HCC metastasis by using the HCC metastatic mouse model described above (see Material and Methods) (Fig. 6A). Consistent with the in vitro results, the incidence of intrahepatic and lung metastasis in MHCC97-LHSF1 cells was significantly higher than the incidence in MHCC97-Lcontrol cells (P = .032 and P = .02, respectively) (Fig. 6B,C), and HCCLM3HSF1RNAi+ cells had a lower intrahepatic and lung metastasis rate than HCCLM3control cells (P = .003 and P = .032, respectively) (Fig. 6B,C). Together, these data support an important role for HSF1 in HCC metastasis in vivo.

Figure 6.

Heat shock factor 1 (HSF1) promotes invasion and metastasis of hepatocellular carcinoma (HCC) in vivo. (A) The HCC metastatic mouse model in mice was constructed by using MHCC97-Lcontrol cells, MHCC97-LHSF1 cells, HCCLM3control cells, and HCCLM3HSF1 RNA interference-positive (HCCLM3HSF1RNAi+) cells as described in the text (see Materials and Methods). (B) The percentage of mice with or without metastatic nodules in the livers or in the lungs was calculated and compared. (C) Intrahepatic (a,d) and lung (b,c,e,f) metastases of MHCC97-L cells (a-c) and HCCLM3 cells (d-f) are shown. Black arrows indicate metastatic foci in the liver or lung (hematoxylin and eosin stain; original magnification ×200 in a, b, d, and e; ×400 in c and f).


HSF1 originally was cloned and identified as a master controller of heat shock response (HSR), which is 1 of the most evolutionarily conserved defensive mechanisms against acute exposure to extreme environmental and pathologic conditions.33 HSF1 has a key role in the cellular response leading to the expression of HSP genes under stress conditions.34 This response is mediated by binding of the activated trimer of HSF1 to the heat shock element (HSE), which consists of 3 inverted repeats of the sequence NGAAN in the promoter region of HSPs.35

Although recent studies have linked HSF1 to human cancers, the precise relations between HSF1 and HCC remain unknown. In the current study, we demonstrated that HSF1 is up-regulated significantly in HCC, especially in metastatic HCC tissues. Moreover, elevated HSF1 expression was correlated with the clinicopathologic characteristics of multiple nodules, an absence of capsular formation, and the presence of venous invasion, which indicate the high metastatic potential of HCC.5 These findings suggest the involvement of HSF1 in the pathogenesis of HCC, including metastasis. We also demonstrated that HSF1 is overexpressed in human prostate tumors cell lines, and there is a close correlation between the expression of HSF1 and the aggressiveness of the tumors.36, 37 In addition, HSF1 was associated with overall survival and disease-free survival in patients with HCC and was an independent prognostic factor. These lines of evidences suggest that HSF1 plays critical roles in tumor initiation and progression.38

Invasion and metastasis, 2 of the most important hallmarks of cancer, are the leading lethal factors in HCC.14 The pathogenesis of invasion and metastasis is a multifactorial, multistep, and complex process in which cell migration plays a fundamental role.39 There is growing evidence that actin polymerization and the dynamic reorganization of the actin cytoskeleton play important roles in the regulation of cell migration.40 We observed that HSF1 can promote HCC cell invasion and metastasis in vitro and in vivo by regulating cell migration in an Hsp27-dependent manner. Hsp27, a member of the small HSP family, is a major downstream target of HSF1.41 Available information suggests that Hsp27 plays an important role in cell migration as a sole effector of actin remodeling.42, 43 Studies have demonstrated that unphosphorylated Hsp27 binds to the barbed end of actin filaments and, thus, inhibits the polymerization of F-actin; whereas phosphorylated Hsp27 reverses the inhibition and increases the dynamics of F-actin assembly.44, 45 Indeed, the overexpression of HSP27 and the capacity of HSP27 to regulate actin reorganization and cell migration have been reported.13, 46 Our findings that HSF1 may facilitate the expression and phosphorylation of Hsp27, which are correlated with the pattern and morphology of F-actin, further confirmed these results.

Because HSF1 has no kinase activity, it is unclear how HSF1 facilitates the phosphorylation of Hsp27. It has been reported that Hsp27 phosphorylation is catalyzed by the mitogen-activated protein kinase (MAPK) superfamily.46 Usually, the p38 MAPK pathway phosphorylates Hsp27 through MAPK-activated protein kinase 2 (MAPKAP2), which is 1 of the substrates of p38 MAPK.47 A previous study also demonstrated that HSF1 can regulate the migration of MEF cells through c-Jun N-terminal kinase (JNK) and extracellular regulated kinase (ERK) signaling.24 Taking these findings into account, it is most likely that HSF1 regulates Hsp27 phosphorylation through the p38 MAPK pathway. However, how HSF1 regulates p38 MAPK remains unclear. One study demonstrated that the impairment of JNK and ERK signaling in Hsf1−/− MEF cells was caused in part by the reduced expression of EGFR1.24 Furthermore, the involvement of Wip1, a serine/threonine phosphatase, and protein kinase C-β has been implicated in the regulation of p38 MAPK in HCC cells.13, 46 However, the detailed mechanisms by which HSF1 regulates p38 MAPK need further investigation.

Current strategies for anticancer drug development involve identifying novel molecular targets that are crucial for tumorigenesis.48 On the basis of the finding that HSF1 inhibition suppresses HCC cell invasion and metastasis in vitro and in vivo, silencing HSF1 expression can be used as a potential treatment strategy for HCC metastasis. Recently, Westerheide et al49 reported that triptolide, a diterpene triepoxide from the plant Triptergium wilfordii, could abrogate the transactivation function of HSF1, suggesting the great value of HSF1 inhibition in cancer treatment. Screening the small-molecular compound library to identify the specific inhibitors of HSF1 that have high efficiency will help to develop new anticancer drugs for HCC.

In conclusion, our study demonstrates that HSF1 is up-regulated in HCC, and its overexpression is correlated significantly with a poor prognosis in patients with HCC. Furthermore, we have demonstrated that HSF1 can promote invasion and metastasis of HCC by enhancing cell motility through Hsp27. To our knowledge, this study is the first report about the clinical relevance and biologic function of HSF1 in HCC. Collectively, these findings suggest that HSF1 may serve as a novel prognostic marker and therapeutic target for HCC.


This research was supported by National Key Technologies R&D Program of China (No. 2001BA703B04, No. 2004BA703B02), National Keystone Basic Research Program of China (No. 2004CB720303, No. 2009CB521801), National High Technology Research and Development Program of China (No. 2006AA02Z4B2), Clinical Subjects' Key Project of Ministry of Health (2010-2013), and National Science and Technology Major Projects (2008ZX10208, 2009ZX09103-681).


The authors made no disclosures.