Despite complete resection of the early stage of oral tongue cancer by partial glossectomy, late cervical lymph node metastasis is frequently observed. Gene amplification of ACTN4 (protein name: actinin-4) is closely associated with the metastatic potential of various cancers. This retrospective study was performed to demonstrate the potential usefulness of ACTN4 gene amplification as a prognostic biomarker in patients with stage I/II oral tongue cancer. Fifty-four patients with stage I/II oral tongue cancer were enrolled retrospectively, in accordance with the reporting recommendations for tumour marker prognostic studies (REMARK) guidelines. The copy number of ACTN4 and the protein expression of actinin-4 were evaluated by fluorescence in situ hybridization (FISH) and immunohistochemistry (IHC), respectively. The overall survival time of patients with gene amplification of ACTN4 was significantly shorter than that of patients without gene amplification ( P = 0.0010, log-rank test). Gene amplification of ACTN4 was a significant independent risk factor for death in patients with stage I/II oral tongue cancer (hazard ratio 6.08, 95% confidence interval 1.66–22.27). Gene amplification of ACTN4 is a potential prognostic biomarker for overall survival in oral tongue cancer.
Treatment planning for node-negative patients with T1 and T2 squamous cell carcinoma of the tongue remains an important dilemma when selecting a surgical strategy with or without elective neck resection. According to the clinical practice guidelines of the Japan Society of Clinical Oncology, the primary therapeutic strategy for stage I and II oral tongue cancer is a glossectomy alone, or glossectomy and elective neck dissection. Even in patients with oral tongue cancer that can be resected completely from the tongue by partial glossectomy, late cervical lymph node metastases after surgery are frequently observed, without local recurrence in 26.8–48.2% of cases . Late cervical lymph node metastasis is an adverse prognostic factor for patients with oral squamous cell carcinoma , and identifying patients who are at potential risk of late cervical lymph node metastasis is important when deciding on the therapeutic strategy and/or follow-up options. It is considered that the metastatic potential of the tumour at the primary site is involved in late cervical lymph node metastases, and that the evaluation of the metastatic potential of each tumour may serve as an indicator to predict late cervical lymph node metastasis.
Actinin-4, encoded by the gene ACTN4 , is an actin-bundling protein that was identified by the present research group. ACTN4 is a non-muscle-type alpha-actinin that is closely associated with cancer invasion and metastasis . The actinin-4 molecule is essential in the cellular processes for invasion and metastasis. Overexpression of actinin-4 has been detected in several cancers that have an invasive phenotype , such as breast cancer , pancreatic cancer , ovarian cancer , lung cancer , salivary gland carcinoma , and oral squamous cell carcinoma . Moreover, gene amplification of ACTN4 has been shown to be one of the causes of the overexpression of actinin-4 in pancreatic cancer , ovarian cancer , and non-small cell lung cancer (NSCLC) .
A retrospective study was performed to demonstrate the potential usefulness of ACTN4 gene amplification as a prognostic biomarker in patients with stage I and II oral tongue cancer.
Materials and methods
Patients and tissue samples
Fifty-four patients who underwent glossectomy alone with curative intent for stage I or II oral tongue cancer at the National Cancer Center Hospital (Tokyo, Japan) between 2001 and 2013 were enrolled retrospectively in this study. The study was performed in accordance with the reporting recommendations for tumour marker prognostic studies (REMARK) guidelines ( Fig. 1 ) . For the risk assessment of late cervical lymph node metastasis, those patients who underwent elective neck dissection at the first operation were excluded from this study. The duration of follow-up for all cases ranged from 40 to 3956 days; the median duration of follow-up was 2154 days. None of the patients had undergone adjuvant chemotherapy. This study was approved by the ethics committee of the National Cancer Center.
The anti-actinin-4 monoclonal antibody (13G9), which was originally established by the present study group, was purchased from Abnova Inc. (Taipei, Taiwan) . Actinin-4 was immunostained using the Ventana DABMap Detection Kit and automated slide stainer (Discovery XT) (Ventana Medical Systems Inc., Tucson, AZ, USA) .
The immunoreactivity of actinin-4 was classified into two groups according to criteria that have been described previously: positive and negative. The positive group (actinin-4-positive) was defined as cases in which the staining intensity of the anti-actinin-4 antibody in the tumour was higher than that in endothelial cells, in addition to cases in which the staining area of actinin-4 encompassed more than 30% of the tumour area. The negative group (actinin-4-negative) was defined as those cases that did not show actinin-4-positive staining, as described previously . The staining patterns were evaluated by two independent investigators (T.K. and K.H. or T.M.) who had no clinical information about the cases .
Fluorescence in situ hybridization (FISH)
The FISH probes used were a bacterial artificial chromosome clone (BAC) containing ACTN4 and chromosome 19 (a control clone; CEP19); these probes were originally established in the authors’ laboratory and were purchased from Abnova. The labelled BAC DNA was used in the FISH analysis as described previously. Thinly sliced pathological sections were hybridized with FISH probes at 37 °C for 48 h. The nuclei were counterstained with 4′,6-diamidino-2-phenylindole. The numbers of fluorescence signals corresponding to the copy number of ACTN4 and control signals in the nuclei of 20 interphase tumour cells were counted (T.K. and K.H.) .
FISH patterns were categorized as described previously. Briefly, the samples were grouped as ‘gene amplification’ (i.e., FISH-positive) if the ACTN4 /CEP19 ratio was ≥2.0, or ‘no gene amplification’ (i.e., FISH-negative) if the ACTN4 /CEP19 ratio was <2.0, as per the US Food and Drug Administration recommendations for HER2 tests (human epidermal growth factor receptor 2) .
Significant differences were detected using the Mann–Whitney U -test, Student’s t -test, Pearson’s χ 2 test, and Fisher’s exact test. Overall survival (OS) was measured as the period from surgery to the date of death or last follow-up and was estimated by Kaplan–Meier analysis. Differences between OS curves were assessed with the log-rank test. Univariate and multivariate analyses were performed with the Cox regression model. Data were analyzed using StatFlex statistical software package version 6.0 (StatFlex, Osaka, Japan) .
Protein expression of actinin-4 in stage I/II oral tongue cancer
The protein expression level of actinin-4 in stage I/II oral tongue cancer was determined using IHC ( Fig. 2 ). Protein expression of actinin-4 was not observed in the normal mucosa of the tongue ( Fig. 2 A). The 54 patients with stage I/II oral tongue cancer were classified as actinin-4-positive or actinin-4-negative: 31 cases (57.4%) were actinin-4-negative ( Fig. 2 B) and 23 cases (42.6%) were actinin-4-positive ( Fig. 2 C). Protein overexpression of actinin-4 was particularly recognized at the invasive front of the oral tongue cancer in comparison with the borderline between the cancer and normal mucosa ( Fig. 2 D–F).
The correlations between clinical findings and actinin-4 expression patterns are shown in Table 1 . There was no statistically significant difference between the protein expression groups for actinin-4 and clinical findings in terms of age (median 62 years), sex, histological differentiation (poorly/moderately differentiated and well-differentiated), lymphatic invasion, neural invasion, and vascular invasion. However, there was a statistically significant difference in the expression level of actinin-4 between stage I and stage II cancers ( P = 0.0282, Fisher’s exact test): the expression of actinin-4 was higher in patients with stage II than in patients with stage I ( Table 1 ).
|Number of cases (%)||Actinin-4 IHC||P -value a||ACTN4 FISH||P -value a|
|Positive||Negative||Amplification||No gene amplification|
Copy number of ACTN4 in stage I/II oral tongue cancer
Forty-eight patients showed a normal copy number for ACTN4 by FISH analysis ( Fig. 2 G, and Table 1 ). Six patients showed gene amplification of ACTN4 (i.e., FISH-positive) ( Fig. 2 H; Table 1 ). There was no statistically significant difference between the FISH-positive and FISH-negative ACTN4 groups in terms of age, sex, clinical stage, histological differentiation, lymphatic invasion, neural invasion, or vascular invasion.
Hazard ratios for death in patients with stage I/II oral tongue cancer
The hazard ratios (HR) were calculated for some of the factors including, sex, clinical stage, histological differentiation, copy number status of ACTN4 , and protein expression of actinin-4. Univariate Cox regression analysis indicated that age (median 62 years) and the ACTN4 copy number status showed statistically significant HRs for death: HR 6.78 (95% confidence interval (CI) 1.48–30.98, P = 0.0136) for age and HR 6.36 (95% CI 1.85–21.91, P = 0.0034) for copy number status of ACTN4 . Multivariate Cox regression analysis indicated that both age and the ACTN4 copy number status had statistical significance for death: HR 6.61 (95% CI 1.42–30.77, P = 0.0161) for age and HR 6.08 (95% CI 1.66–22.27, P = 0.0064) for ACTN4 copy number status. Thus, age and the ACTN4 copy number status were found to be independent prognostic factors for death in stage I/II oral tongue cancer ( Table 2 ).
|Univariate analysis||Multivariate analysis|
|HR||95% CI||P -value||HR||95% CI||P -value|
|>62 vs. ≤62||6.78||1.48–30.98||0.0136 a||6.61||1.42–30.77||0.0161 a|
|Female vs. male||0.7||0.23–2.16||0.536|
|II vs. I||1.52||0.32–7.23||0.6|
|Poor and moderate vs. well||1.74||0.59–5.20||0.318|
|Present vs. absent||2.86||0.61–13.38||0.182|
|Positive vs. negative||6.36||1.85–21.91||0.0034 a||6.08||1.66–22.27||0.0064 a|
|Positive vs. negative||2.31||0.76–7.00||0.14|
The prognostic significance of protein expression of actinin-4 and gene amplification of ACTN4
Kaplan–Meier analysis showed that there was no statistically significant difference in the OS of patients with stage I/II oral tongue cancer between the two groups separated by protein expression of actinin-4 ( Fig. 3 A) ( P = 0.133, log-rank test). On the other hand, the OS of patients who were FISH-negative for ACTN4 was significantly longer than that of patients who were FISH-positive ( Fig. 3 B) ( P = 0.0010). The OS median survival time (MST) of patients who were FISH-positive was 826 days. However, the OS MST of patients who were FISH-negative could not be calculated, because the mortality of patients with a normal copy number did not reach 50% ( Fig. 3 D). Thus the copy number of ACTN4 was a better prognostic indicator than the expression of actinin-4 for OS in patients with stage I/II oral tongue cancer.
The prognostic impact on OS after surgery of late cervical lymph node metastasis
The prognostic significance of the protein expression level of actinin-4 for disease-free survival (DFS) was investigated in the 54 patients with stage I/II oral tongue cancer who underwent a partial glossectomy without local recurrence after the first surgery. The difference in DFS between the actinin-4-positive and actinin-4-negative subgroups was statistically significant ( P = 0.0081, log rank test): the DFS of patients who were classified into the actinin-4-positive subgroup was significantly shorter than that of the patients in the actinin-4-negative subgroup (data not shown).
The patients were then further classified into three subgroups groups based on IHC and FISH data. These groups were (1) patients who were actinin-4-negative and FISH-negative for ACTN4 (actinin-4-negative/ ACTN4 FISH-negative) ( n = 31), (2) patients who were actinin-4-positive and FISH-negative for ACTN4 (actinin-4-positive/ ACTN4 FISH-negative) ( n = 17), and (3) patients who were actinin-4-positive and FISH-positive for ACTN4 (actinin-4-positive/ ACTN4 FISH-positive) ( n = 6); none of the patients in this study were actinin-4-negative and FISH-positive for ACTN4 .
Kaplan–Meier analysis revealed statistically significant differences in DFS between the actinin-4-negative/ ACTN4 FISH-negative group and both the actinin-4-positive/ ACTN4 FISH-negative group ( P = 0.036; log-rank test) and the actinin-4-positive/FISH-positive group ( P = 0.008; log-rank test) ( Fig. 3 C). The DFS time was significantly shorter in the actinin-4-positive/ ACTN4 FISH-negative patients and in the actinin-4-positive/ ACTN4 FISH-positive patients than in the actinin-4-negative/ ACTN4 FISH-negative patients. The MST of actinin-4-positive/ ACTN4 FISH-negative patients was 1964 days and for actinin-4-positive/ ACTN4 FISH-positive patients was 744 days. The DFS rate of the actinin-4-negative/ ACTN4 FISH-negative group did not reach 50%. Univariate Cox regression analysis indicated that the HRs of the actinin-4-positive/ ACTN4 FISH-negative and actinin-4-positive/ ACTN4 FISH-positive patients in comparison with actinin-4-negative/ ACTN4 FISH-negative patients were 2.82 (95% CI 1.04–7.64) and 2.19 (95% CI 1.17–4.09), respectively ( Fig. 3 C).
In contrast, there was no statistically significant difference in OS time between actinin-4-negative/ ACTN4 FISH-negative patients and actinin-4-positive/ ACTN4 FISH-negative patients ( Fig. 3 D). There was, however, a statistically significant difference in OS time between the actinin-4-negative/ ACTN4 FISH-negative and actinin-4-positive/ ACTN4 FISH-positive groups ( P = 0.0008; log-rank test): the HR for death in the actinin-4-positive/ ACTN4 FISH-positive group was 7.60 (95% CI 1.95–29.6) compared to the actinin-4-negative/ ACTN4 FISH-negative group ( Fig. 3 D). Moreover, there was also a statistically significant difference in OS time between the actinin-4-positive/ ACTN4 FISH-negative and actinin-4-positive/ ACTN4 FISH-positive groups: the OS time of the actinin-4-positive/ ACTN4 FISH-positive group was significantly shorter than that of the actinin-4-positive/ ACTN4 FISH-negative group ( P = 0.035; log-rank test) ( Fig. 3 D).
In spite of the fact that the DFS time of the actinin-4-positive/ ACTN4 FISH-negative patients was significantly shorter than that of the actinin-4-negative/ ACTN4 FISH-negative patients, there was no statistically significant difference in OS time between the actinin-4-positive/ ACTN4 FISH-negative and actinin-4-negative/ ACTN4 FISH-negative subgroups. The reason for this discrepancy between the two subgroups in terms of the prognosis for DFS and OS was unclear. Thus, to identify the reason for this discrepancy, the influence of second surgery for late cervical lymph node metastases after the first glossectomy was investigated. Twelve patients underwent neck resection or late cervical lymph node metastasis after glossectomy in the National Cancer Center Hospital. Table 3 shows the correlations between the clinical characteristics of patients who underwent neck resection after glossectomy and actinin-4/ ACTN4 FISH status. Ten of the 12 patients were actinin-4-positive and three of the 12 patients were FISH-positive for ACTN4 . There were no statistically significant differences between the actinin-4 status groups (actinin-4-positive vs. actinin-4-negative; ACTN4 FISH-positive vs. ACTN4 FISH-negative) for the clinical characteristics of age, sex, stage, histological differentiation, lymphatic invasion, neural invasion, or vascular invasion ( Table 3 ). However, the difference in OS time between FISH-positive and FISH-negative groups was statistically significant ( P = 0.0145) ( Fig. 3 E): the OS time of the FISH-positive subgroup was significantly shorter than that of the FISH-negative subgroup. The 5-year survival rate of the FISH-negative group was 88.9% and that of the FISH-positive group was 0% ( Table 3 ). The OS MST of the FISH-positive subgroup was 626 days; however, the OS rate of the FISH-negative subgroup did not reach 50% ( Fig. 3 E).
|Number of cases (%)||Actinin-4 IHC||P-value a||ACTN4 FISH||P-value a|
|Positive||Negative||Gene amplification||No gene amplification|
|5-year DFS rate||0%||0%||0%||0%||0%|
|Median DFS (follow-up time), days||187 (38–744)||176 (38–744)||187 (187–407)||350 (64–744)||187 (38–486)|
|5-year OS rate||61.70%||52.50%||100%||0%||88.90%|
|Median OS (follow-up time), days||NR (218–3562)||NR (218–2648)||NR (2223–3562)||626 (403–879)||NR (218–3562)|