Molecular Medicine Research Center, Chang Gung University, Tao-Yuan, Taiwan
Department of Medical Research Development, Chang Gung Memorial Hospital, Tao-Yuan, Taiwan
Corresponding author: Lang-Ming Chi, PhD, Department of Medical Research Development, Chang Gung Memorial Hospital; No. 5, Fusing St, Gueishan Township, Taoyuan County 333, Taiwan; Fax: (011) 886 32118800-3533; firstname.lastname@example.org
A previous comparative tissue proteomics study by the authors of the current study led to the identification of caldesmon (CaD) as one of the proteins associated with cervical metastasis of oral cavity squamous cell carcinoma (OSCC). In the current investigation, the authors focused on the potential functions of CaD in patients with OSCC.
CaD expression was examined in tissue samples from 155 patients using immunohistochemical analysis. The expression of CaD variants was determined by Western blot analysis and reverse transcriptase-polymerase chain reaction. In addition, the specific effects of CaD gene overexpression and silence were determined in OSCC cell lines.
CaD expression was found to be significantly higher in tumor cells from metastatic lymph nodes compared with primary tumor cells, and was nearly absent in normal oral epithelia. Higher CaD expression was found to be correlated with positive N classification, poor differentiation, perineural invasion, and tumor depth (P = .001, P = .029, P = .001, and P = .031, respectively). In survival analyses, OSCC patients with higher CaD expression were found to have poorer prognosis with regard to disease-specific survival and disease-free survival (P = .003 and P = .014, respectively). Multivariate analyses further indicated that higher CaD expression was an independent predictor of disease-specific survival (P = .043). Serum CaD levels were found to be significantly higher in patients with OSCC, but this finding was not associated with clinicopathological manifestations. Data obtained from in vitro suppression, rescue, and overexpression of CaD in OEC-M1 cells indicated that CaD promotes migration and invasive processes in OSCC cells.
Oral cavity squamous cell carcinoma (OSCC) is the most common head and neck cancer, accounting for approximately 3% of all newly diagnosed cancer cases.[1, 2] Despite recent advances in surgical, radiotherapy, and chemotherapy treatment protocols, the long-term survival of patients with OSCC has remained almost unchanged over the past decade.[3, 4] These unsatisfactory results and lack of improvement, despite advancements in treatment modalities, are explained mainly by the finding that OSCCs are associated with a high probability of cervical lymph node metastasis, historically regarded as the major poor prognosticator for OSCC.[5-8] Because cervical metastasis presents a major obstacle in the management of OSCC, clarification of its molecular basis is critical. Currently, modern high-throughput molecular technology platforms are widely used to identify OSCC metastasis-related tumor markers. Previous cDNA microarray studies focusing on mRNA expression analysis of primary tumors have established the genetic expression profiles associated with OSCC lymphatic metastasis.[9-13] In the systemic proteomics field, we adapted a novel approach to integrate laser capture microscopy for sample preparation and isobaric tag labeling for comparative analysis to identify the potential markers of OSCC metastasis. From the differentially expressed protein database generated using this approach, caldesmon (CaD) was identified as a highly expressed member associated with OSCC metastasis.
CaD, encoded by the CALD1 gene, exists as 5 isoforms. The high-molecular-weight CaD protein (h-CaD; 120-150 kilodaltons) is restricted to visceral and vascular smooth muscle cells, whereas the low-molecular-weight isoforms (l-CaD; 70-80 kilodaltons) are present in non-smooth muscle cells. CaD is a major actomyosin-binding protein that is in addition able to bind actin, tropomyosin, Ca+-calmodulin, myosin, and phospholipids in the cell cytoplasm. Therefore, the protein is crucially involved in the regulation of the microfilament network, and acts as an important modulator of various cell functions, including motility.[16, 17] Although CaD expression in the early stages of tumor neovascularization in a wide variety of cancer types and high serum levels have been detected in patients with glioma,[18, 19] to the best of our knowledge its potential role in human cancers has rarely been addressed to date. We hypothesized that CaD is overexpressed in OSCC and plays a specific role in tumor cell pathophysiology modulation. Consistent with this theory, our experiments revealed significantly higher CaD levels in OSCC tumor cells and sera from patients with OSCC. Immunohistochemical (IHC) analyses provided preliminary evidence that CaD overexpression is associated with cervical metastasis and poorer prognosis after treatment in patients with OSCC. Finally, we demonstrated that expression of CaD in OSCC cell lines promotes cellular migration and invasion.
MATERIALS AND METHODS
Tumor specimens for IHC analyses were obtained from 155 consecutively enrolled patients (135 men and 20 women) with OSCC who were diagnosed and treated between 2002 and 2007. Patients with inoperable disease, synchronous cancers, recurrent cancers, distant metastasis, or a history of another malignancy were excluded from the analyses. All patients provided informed consent before study participation, and the study was approved by the Institutional Review Board. Patients underwent standard preoperative workups, treatment, and regular follow-up visits according to institutional guidelines. After discharge, all patients had regular follow-up visits every 2 months for the first year, every 3 months for the second year, and every 6 months thereafter.
Immunohistochemistry was performed on formalin-fixed and paraffin-embedded tissues sections incubated with homemade anti-CaD antibody as described previously.[14, 20] Expression of CaD was scored using a combined scoring method accounting for both staining intensity and the percentage of stained cells.[6, 14] Strong, moderate, weak, and negative staining intensities were scored as 3, 2, 1, and 0, respectively. For each intensity score, cells staining at that specific level were visually estimated and calculated as a percentage. The resultant combined score was calculated as the sum of the percentage of stained cells multiplied by the intensity scores. All specimens were independently evaluated by our pathologists (Y.L. and C.H.), who had no prior knowledge of the clinical origin of the specimen.
Western Blot Analysis
Extraction and quantification of proteins were performed as described previously. Briefly, equal amount of protein were separated on sodium dodecyl sulfate-polyacrylamide gel electrophoresis gels, transferred to polyvinylidene difluoride membranes, and probed using the homemade CaD antibody. The actin signal determined by β-actin antibody (NB600-501; Novus Biologicals, Littleton, Colo) was used as the loading control.
Human CaD Transcript Analysis Using Reverse Transcriptase-Polymerase Chain Reaction
Total RNA was extracted from culture or tissue cells using TRIzol, according to the manufacturer's protocol (Invitrogen, Life Technologies, Grand Island, NY). Primers for distinguishing between CaD isoforms were designed according to a previous report. The primers used in this study were as follows: sense primer Pn2, d(ATGGATGATTTTGAGCGTCG), which encodes the common amino terminal sequence of h-CaD and WI-38 l-CaD isoforms; sense primer Pn, d(ATGCTGGGTGGATCCGGATC), which specifically encodes the amino terminal sequence of HeLa l-CaD isoforms; and antisense primer Pm, d(GTTTAAGTTTGTGGGTCATGAATTCTCC) located at exon 5, which is common for all isoforms. The 1513-bp, 825-bp, and 747-bp fragments were amplified from cDNA of h-CaD, WI-38 l-CaD I and II, respectively, by paired Pn2 and Pm primers. Using paired Pn and Pm primers, 807-bp and 729-bp fragments were amplified for HeLa l-CaD I and II, respectively.
Cell Culture and Gene Silencing/Overexpression via RNA Interference and Ectopic Gene Delivery
The SCC4 (CRL-1624; ATCC, Manassas, Va) and SCC25 (CRL-1628; ATCC) oral cancer cell lines were cultured in a 1:1 mixture of Dulbecco modified Eagle medium/F12 medium supplement with 400 ng/mL of hydrocortisone and 10% fetal bovine serum. OEC-M1, an indigenous gingival epidermal carcinoma cell line in Taiwan, was maintained in RPMI medium containing 10% fetal bovine serum at 37°C in a humidified 5% carbon dioxide atmosphere. SMARTpool small interfering RNAs (siRNA) purchased from Thermo Fisher Scientific (Dharmacon Inc, Lafayette, Colo) were transiently transfected into culture cells using Lipofectamine RNAiMAX reagent (Invitrogen, Carlsbad, Calif) according to the manufacturer's recommendations. Homo sapiens WI-38 l-CaD I in pCMV6-Entry vector (OriGene, Rockville, Md) was transiently transfected for ectopic expression. After 24 hours (ectopic expression) or 48 hours (silence) of transfection, cells were harvested for proliferation, migration, and invasion analyses.
Cell Migration and Invasion Assay
Analysis of cell migration and invasion was performed in Boyden chambers. Serum-free medium containing 10 μg/mL of fibronectin (Sigma-Aldrich, St. Louis, Mo) was added to the lower chamber, and 2 × 104 and 2 × 105 cells in serum-free medium were seeded into the upper chamber with (invasion) or without (migration) coating with Matrigel (BD Biosciences, Billerica, Mass), according to the manufacturer's recommendations. Chambers were incubated at 37°C for 6 hours (migration) or 24 hours (invasion) and the membrane was fixed in methanol and stained with Giemsa solution (Sigma-Aldrich) for 30 minutes. Cells on the lower surface of the membrane were counted in 6 different visual fields under a light microscope with ×200 magnification. Two independent experiments were performed in triplicate. The percentage of migration and invasion decrease/increase was calculated by comparison with the control experiment and normalized with the average value of the cell proliferation rate by using the 3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide (MTT) assay.
All statistical data were expressed as the mean ± the standard deviation. The Wilcoxon signed rank test was used for comparison of the IHC staining scores between paired tumor and pericancerous normal mucosa samples. Cell proliferation, migration, and invasion experiments were compared using Student t tests for unpaired data. All patients received follow-up evaluations at the study institution's outpatient clinic until June 2012 or their death. Survival analysis was plotted using the Kaplan-Meier method, and differences were evaluated using the log-rank test. Multivariate regression analyses were applied to define specific risk factors for disease-specific survival (DSS). Statistical analyses were performed using SAS software (version 9.1; SAS Institute Inc Cary, NC). All P values were 2-sided, and statistical significance accepted at P < .05.
Patient age at the time of diagnosis ranged from 22 years to 84 years (mean, 51.3 years ± 12.0 years). The associated subsites were buccal mucosa (61 patients), gum (20 patients), hard palate (4 patients), lip (4 patients), floor of the mouth (5 patients), and tongue (61 patients).
Overexpression of l-CaD in Tumor Cells From OSCC Specimens
Variants of the human CaD gene transcribed in oral cancer cell lines and tissue lysates were first evaluated using Western blot analysis and reverse transcriptase-polymerase chain reaction (RT-PCR) with sense primers that specifically annealed to gene fragments encoding the amino terminal sequences of various CaD isoforms. As shown in Figure 1A, WI-38 l-CaD I and II (825-bp and 747-bp amplicons, respectively), but not h-CaD (1513-bp amplicon), was detected in oral cancer cell lines using RT-PCR with the sense primer Pn2 specific for nucleotide fragments encoding the amino termini of the WI-38 l-CaD and h-CaD isoforms. HeLa l-CaD isoforms were not observed in either oral cancer cell lines (Fig. 1A, lower panel) or tumor tissue (data not shown) using a HeLa-type specific sense primer Pn. The presence of l-CaD was further confirmed using Western blot analysis. As depicted in Figures 1B and 1C, the l-CaD protein was solely detected in OSCC cell lines used in this study, and dramatically increased in tumor tissues compared with adjacent nontumor parts. Next, tissue sections from the 155 patients were subjected to IHC staining. Notably, CaD was found to be highly expressed in the cytoplasm of tumor cells, with no nuclear staining observed (Fig. 1D). In contrast, CaD expression was mainly absent or extremely low in tumor-infiltrating lymphocytes and interstitial tissues. Moreover, paired normal oral epithelium samples disclosed relatively low or no expression of CaD. To ascertain whether this protein is specifically overexpressed in metastatic tumor cells, the corresponding resected metastatic lymph nodes were also subjected to IHC staining. CaD was found to be highly expressed in metastatic tumor cells compared with adjacent tissue and nontumor cells in metastatic lymph nodes. Statistical analysis of 136 paired samples available from the 155 patients with OSCC demonstrated significantly higher IHC scores of CaD in tumor tissue, compared with nontumor normal epithelium (143.1 ± 57.4 vs 6.6 ± 20.7; P < .0001) (Fig. 1E). Furthermore, CaD levels were markedly higher in 30 metastatic tumors from lymph nodes compared with the corresponding primary tumors (173.7 ± 34.8 vs 143.1 ± 57.4; P = .012). IHC experiments confirmed higher CaD expression in tumor cells of metastatic lymph nodes than in primary tumor cells, and nearly no expression in normal oral epithelia.
Association Between CaD Expression and Various Clinicopathological Manifestations
Next, we evaluated the relations between increased CaD expression and various clinicopathological characteristics of patients with OSCC (Table 1). Higher CaD expression was found to be significantly associated with higher pT classification, positive cervical metastasis, higher overall pathological stage, poorer cell differentiation, positive perineural invasion, and greater tumor depth (P = .009, P = .001, P < .001, P = .029, P = .001, and P = .031, respectively) (Table 1). However, we observed no association between CaD overexpression in OSCC tumors and patient age; sex; or habitual behaviors of betel nut chewing, cigarette smoking, and alcohol consumption. Consistent with our hypothesis, CaD overexpression was found to be significantly (P = 0.001) associated with lymph node metastasis (pN classification).
Table 1. Association Between CaD Expression Levels and Clinicopathological Characteristics in 155 Untreated Patients With OSCC
Association Between CaD and Overall Survival, Disease-Free Survival, and DSS
Based on expression data obtained from IHC, patients were stratified into 2 groups (those with high expression vs those with low expression using 150 [the upper tertile value of the IHC scores in all OSCC tumors] of 300 as the cutoff value), and the possible association between CaD expression and overall survival (OS) was evaluated. Survival analysis revealed that the 5-year OS rates for patients stratified into subgroups with high and low CaD expression were 66.9% and 52.9%, respectively. These differences in OS were not found to be statistically significant when compared using the log-rank test (P = .081) (Fig. 2A). However, the Kaplan-Meier plots evaluated 5-year disease-free survival (DFS) rates for patients stratified by high and low CaD expression as 73.9% and 54.7%, respectively. These differences in DFS were statistically significant, as observed using the log-rank test (P = .014) (Fig. 2B). Moreover, the 5-year DSS rates for patients stratified based on high or low CaD expression were significantly different using the log-rank test (73.6% and 52.9%, respectively; P = .003) (Fig. 2C). In addition, CaD expression was found to be a significant predictor of DFS and DSS on univariate analysis with the Cox proportional regression model. To further ascertain whether CaD expression can be applied as an independent predictor of patient survival, multivariate analysis was performed using age, sex, pT classification, pN classification, overall stage, perineural invasion, bone invasion, tumor differentiation, tumor depth, and CaD expression as parameters in the Cox proportional regression model. Our results collectively indicated that pN classification and CaD expression were independent predictors of DSS (P = .035 and P = .043, respectively) (Table 2). Accordingly, we concluded that CaD expression is critical in lymph node metastatic processes and DSS in patients with OSCC.
Table 2. Multivariate Analysis of Disease-Specific Survival in Patients With OSCC After Treatmenta
l-CaD Promotes Oral Cancer Cell Migration and Invasiveness In Vitro
Expression of endogenous l-CaD was initially assessed from amplicon products of RT-PCR. In cultured oral cancer cells, WI-38 l-Cad variants, but not h-CaD or HeLa l-CaD, were detected using RT-PCR (Fig. 1A). The effects of knockdown, rescue, and overexpression of l-CaD in OEC-M1 oral cancer cells were further examined to evaluate its biological significance. The efficiencies of gene silencing, rescue, and overexpression of target CaD were determined via Western blot analysis using the anti-CaD antibody (Fig. 3A). A significant decrease in the endogenous l-CaD protein level was observed in cells transfected with CaD-specific siRNA compared with those transfected with the scrambled sequence siRNA. Knockdown of l-CaD expression was found to have a marginal effect (approximately15%) on the proliferative ability of OEC-M1 cells by the MTT assay, but led to a significant attenuation of cell migration (approximately 60% reduction; P < .001) and invasiveness (approximately50% reduction; P < .001) (shown as treatments 1 and 2 in Figs. 3B and 3C). Retarded motility was rescued by exogenous l-CaD expression (shown as treatment 3 in Figs. 3B and 3C). Moreover, ectopic expression of l-CaD in OEC-M1 cells clearly augmented cell migration and invasion activities (shown as treatments 4 and 5 in Figs. 3B and 3C). The findings of the current study collectively indicate that l-CaD is involved in the regulation of migration and invasive processes in oral cancer cells.
Dysregulation of cancer cell adhesion and migration is crucial for cancer progression and metastasis. Cell adhesion and migration are modulated by complex changes in the cytoskeleton, particularly the actin-based structure, a key regulator of cell adhesion, invasion, and migration. CaD is an extraordinary actin-binding protein that in addition has the ability to bind myosin, calmodulin, and tropomyosin, thereby contributing to cytoskeleton modulation. These multiple binding partners and diverse biochemical properties suggest that CaD is a pivotal regulatory protein in cell motility. The human CaD gene generates at least 5 species of mRNA from 13 exons via alternative splicing. Two classes of CaD proteins have been identified on the basis of sodium dodecyl sulfate-polyacrylamide gel electrophoresis analyses, specifically, h-CaD and l-CaD. Although CaD isoforms generated via alternative splicing exhibit similar biochemical characteristics, their cell-type distribution and cellular localization are mutually exclusive. h-CaD is predominantly expressed in smooth muscle cells as a component of contractile filaments. Conversely, l-CaD is ubiquitously expressed in non-smooth muscle cells as a cytoskeletal component found not only in stress fiber, but also dynamic structures, including membrane ruffles and lamellipodia extensions.[25, 26] l-CaD has long been considered pivotal in the assembly and stabilization of the microfilament network in nonmuscle cells and acts as a crucial regulator of various cell functions, including cell motility. However, to the best of our knowledge, the available literature regarding the association between CaD and tumorigenesis is limited. Using 3 different antibodies against l-CaD and/or h-CaD, Kohler demonstrated a reduction in the h-CaD level in the stroma of primary and lymph node metastatic tumors of colon cancer, whereas that in cancer epithelial cells ranged from weak to strong, compared with normal cells. CaD has been shown to be overexpressed in hepatocellular carcinoma compared with adjacent nonmalignant liver. In particular, HeLa l-CaD is widely expressed in the early stages of tumor neovascularization in a variety of tumors derived from several organs. This finding is supported by a previous report demonstrating that HeLa l-CaD is involved in the migration of endothelial and endothelial progenitor cells in human neoplasm, and contributes to tumor angiogenesis. In addition, l-CaD is upregulated in the tumor neovasculature of gliomas and increased in the serum of patients with glioma. Recently, we have also characterized the CaD level in 292 serum samples by enzyme-linked immunoadsorbent assay and found it was significantly higher in 151 patients with OSCC compared with 141 healthy individuals (2047 ± 289 pg/mL vs 1295 ± 789 pg/mL; P = .011). To our knowledge, the current study is the first to demonstrate high CaD expression in the epithelial cells of primary and lymph node metastatic tissues from patients with OSCC. Moreover, Western blot analysis revealed that l-CaD is predominantly overexpressed in OSCC tumor tissue. RT-PCR data obtained from lysates derived from OSCC cell lines and tumor tissues suggest that WI-38 l-CaD is the predominant key isoform associated with OSCC tumorigenesis. In addition, a previous study reported high CaD mRNA expression in patients with esophageal adenocarcinoma who respond to chemotherapy. IHC findings from the current study further confirm the potential usefulness of CaD as a predictive marker.
The results of the current study demonstrated that CaD is not only overexpressed in OSCC cells but is also significantly associated with several clinicopathological manifestations, such as pT classification, cervical metastasis, perineural invasion, and tumor depth, suggesting a relation between high CaD expression and OSCC tumor progression and metastasis. Furthermore, migration and invasion activities of OSCC cells were found to be significantly affected by l-CaD transfection and RNA interference knockdown. This association is supported by the finding that CaD is closely related to liver carcinogenesis and the metastatic potential of hepatocellular carcinoma cells. CaD is known to be a regulatory protein of the actin cytoskeleton, which may contribute to the regulation of cell morphology, migration, and invasion.[24, 26] It is interesting to note that the underlying mechanism promoting cell motility and the progression of M-phase mitosis is postulated to be phosphorylation of CaD by p34cdc2 (CDK1).[29, 31] In addition to p34cdc2, MAPK/ERK and p21-activated kinase (PAK1) also regulate the disassembly of actin stress fibers, affecting cell morphology and motility through l-CaD phosphorylation,[32, 33] thereby highlighting the necessity to examine the CaD phosphorylation level as well in the future.
The findings of the current study clearly demonstrate that CaD is overexpressed in OSCC tumors. Moreover, high CaD expression might be associated with several clinicopathological manifestations in terms of OSCC tumor progression and poorer prognosis. Western blot analysis and RT-PCR have further indicated that high CaD expression is attributable to l-CaD overexpression, with WI-38 l-CaD being the predominant isoform in OSCC tumors. Finally, transfection and RNA interference of l-CaD demonstrate the involvement of this isoform in OSCC cell proliferation, migration, and invasion.
Supported by grant NSC99-2314-B-182A-051-MY3 from the National Science Council; grant DOH99-TD-C-111-006 from the Department of Health Taiwan; and grants CMRPG381113, CMRPG391422, CMRPG34 0481, CMRPG380793, and CMRPG392062 from Chang Gung University and Chang Gung Memorial Hospital, Taiwan.
CONFLICT OF INTEREST DISCLOSURES
Dr. K-P Chang received grants from Chang Gung Memorial Hospital (CGMH) and Chang Gung University (CGU) (CMRPG381113, CMRPG391422, and CMRP G380793), a grant from the National Science Council of Taiwan (NSC99-2314-B-182A-051-MY3), and a grant from the Department of Health Taiwan (DOH99-TD-C-111-006). Dr. Wang received a grant from the National Institutes of Health (R01-HL-92,252). Dr. Kao received CGMH and CGU grants (CMRPG 381113, CMRPG391422, AND CMRPG380793), a grant from the National Science Council of Taiwan (NSC99-2314-B-182A-051-MY3), and a grant from the Department of Health Taiwan (DOH99-TD-C-111-006). Dr. Liang received grants from the Ministry of Education (EMRPD1A0391 and EMRPD1B0041). Dr. Liu received a CGMH grant (CMRPG381113). Dr. Huang received grants from CGMH (CMRPG392062 and CMRPG392063). Dr. Hseuh received grants from the Ministry of Education (EMRPD1A0391 and EMRPD1 BO041). Dr. Hsieh received grants from the Ministry of Education (EMRPD1A0401 and EMRPD1B0051). Dr. Chien received grants from CGMH and CGU (CMRPD180292 and CMRPD180293). Dr. Y-S Chang received grants from CGMH and CHU (CMRPD180 292 and CMRPD180293). Dr. Yu received grants from CGMH and CGU (CMRPD180302 and CMRPD180 303). Dr. Chi received grants from CGMH and CGU (CMRPG392062 and CMRPG392063).