Fax: (604) 875-4497
Prognostic significance of RUNX3 expression in human melanoma†
Article first published online: 23 DEC 2010
Copyright © 2010 American Cancer Society
Volume 117, Issue 12, pages 2719–2727, 15 June 2011
How to Cite
Zhang, Z., Chen, G., Cheng, Y., Martinka, M. and Li, G. (2011), Prognostic significance of RUNX3 expression in human melanoma. Cancer, 117: 2719–2727. doi: 10.1002/cncr.25838
Zhizhong Zhang's current address: Department of Molecular and Genetic Toxicology, School of Public Health, Cancer Center of Nanjing Medical University, Nanjing, China.
- Issue published online: 7 JUN 2011
- Article first published online: 23 DEC 2010
- Manuscript Accepted: 26 OCT 2010
- Manuscript Revised: 25 OCT 2010
- Manuscript Received: 14 SEP 2010
- tissue microarray
RUNX3 is a tumor suppressor that plays important roles in cell proliferation, apoptosis, and metastasis. The authors investigated the role of RUNX3 in melanoma pathogenesis and analyzed the prognostic impact of RUNX3 expression in a large series of melanoma patients.
Two sets of tissue microarrays were constructed, including 440 cases of melanomas (202 for the training set and 238 for the validation set) and 88 cases of nevi (25 normal nevi and 63 dysplastic nevi). RUNX3 expression was evaluated by immunohistochemistry.
Positive RUNX3 expression was observed in 56%, 54%, 33%, and 24% of the biopsies in normal nevi, dysplastic nevi, primary melanoma, and melanoma metastases, respectively. Significant differences for positive nuclear RUNX3 staining were observed between dysplastic nevi and primary melanomas (P = .002, chi-square test), between dysplastic nevi and melanoma metastases (P < .001, chi-square test), and between primary melanoma and melanoma metastases (P = .045, chi-square test). Loss of RUNX3 expression was correlated with a worse 5-year survival of melanoma patients in both training and validation sets. Furthermore, loss of RUNX3 expression was also correlated with a poor 5-year disease-specific survival in primary melanoma (P = .001) and metastatic melanoma patients (P = .008). Multivariate Cox regression analysis revealed that positive RUNX3 expression is an independent prognostic factor to predict melanoma patient outcome.
Our findings indicate that RUNX3 plays an important role in melanoma pathogenesis and may serve as a promising prognostic marker for melanoma. Cancer 2011;. © 2010 American Cancer Society.
Malignant melanoma is an aggressive form of skin cancer with a rapidly increasing incidence in the Caucasian population worldwide. 1 In the United States, the number of new cases of melanoma in 2009 was estimated to be 68,720, and there were predicted to be 8650 deaths because of melanoma. 2 Melanoma accounts for only 4% of all dermatological cancers; however, it is responsible for >80% of deaths from skin cancers. 3 Although early localized melanoma is curable with surgical excision, the median survival time for patients with metastatic melanoma is 6 to 10 months. 4 Discovery of biomarkers and their application in conjunction with traditional cancer diagnosis, staging, and prognosis could to a large extent help improve early diagnosis and patient care. However, despite the efforts that have been made, reliable markers are still lacking, and the prognosis remains poor. Thus, a better understanding of the regulating factors contributing to melanoma initiation, progression, and metastases is needed.
The Runt family of transcription factors consists of 3 members, RUNX1, RUNX2, and RUNX3. All 3 RUNXs play important roles in both normal developmental processes and carcinogenesis. 5 RUNX family proteins contain the well conserved 128 amino-acids region (Runt domain) and forms a stable complex with PEBP2β/CBFβ to exert its transactivation ability. RUNX proteins regulate the expression of cellular genes by binding to promoters or enhancers of target genes related to cell-fate decisions, which become deranged in cancer cells. 6
Roles of the RUNX family genes have now been described in many different biological systems and animal models, and down-regulation of RUNX proteins has been associated with various types of cancers. 7 RUNX3, in particular, has been shown to play a tumor suppressor role in gastric cancer and intestinal cancer. 8, 9 Notably, the chromosome locus of RUNX3 (1p36) is among the most frequently affected regions in various types of cancers, including melanoma, 10, 11 which suggests that RUNX3 may be involved in the tumorigenesis of these cancers. To date, the expression profile of RUNX3 protein in melanoma has not yet been reported. To further investigate the role of RUNX3 in melanoma pathogenesis, we used tissue microarray (TMA) containing 528 melanocytic lesions and immunohistochemistry to evaluate the expression of RUNX3 in relation to clinicopathologic features and patient outcome. Our data demonstrated that loss of RUNX3 expression is significantly associated with melanoma progression and a worse patient survival. We also found that RUNX3 is an independent prognostic factor for melanoma.
MATERIALS AND METHODS
Construction of TMA
We recruited 289 formalin-fixed, paraffin-embedded melanoma tissues as the training set from the 1990 to 1999 archives of the Department of Pathology, Vancouver General Hospital. To validate the findings from the training set, we recruited an additional 333 melanoma tissues from January 2000 to July 2009 as the validation set. Moreover, 40 normal nevi and 86 dysplastic nevi were also retrieved. The use of human skin tissues was approved by the clinical research ethics board of the University of British Columbia and was done in accordance with the Declaration of Helsinki guidelines. Patients who entered the case cohort were prospectively followed until death or the latest follow-up. The median follow-up time was 48.2 months to the last follow-up date, May 2010. During the follow-up period, 19 patients were lost to follow-up; 166 died of melanoma, and 24 died from other causes. Those lost to follow-up were considered as censored data. The most representative tumor area was carefully selected and marked on the hematoxylin and eosin-stained slide. The TMAs were assembled using a tissue-array instrument (Beecher Instruments, Silver Spring, Md). Duplicate 0.6 mm-thick tissue cores were taken from each biopsy specimen. By using a Leica microtome (Leica Microsystems, Bannockburn, Ill), multiple 4 μm sections were cut and transferred to adhesive-coated slides using regular histology procedures. One section from each TMA was routinely stained with hematoxylin and eosin. The remaining sections were kept at room temperature for immunohistochemical staining.
Immunohistochemistry of TMA
TMA slides were dewaxed by heating at 55°C for 30 minutes and by 3 washes, 5 minutes each, with xylene. The rehydration of tissues was done by 5-minute washes in 100%, 95%, and 80% ethanol and distilled water. Antigen retrieval was done by heating the samples at 95°C for 30 minutes in 10 mM sodium citrate (pH 6.0). Endogenous peroxidase activity was blocked by incubation in 3% hydrogen peroxide for 30 minutes. After blocking with the universal blocking serum (Dako Diagnostics, Glostrup, Denmark) for 30 minutes, the samples were incubated with a monoclonal mouse anti-RUNX3 antibody (1:200 dilution; Medical and Biological Laboratories, Nagoya, Japan) at 4°C overnight. The sections were then incubated for 30 minutes each with a biotin-labeled secondary antibody and then streptavidin-peroxidase (DAKO Diagnostics). The samples were developed using 3,3′-diaminobenzidine substrate (Vector Laboratories, Burlingame, Calif) and counterstained with hematoxylin. The slides were dehydrated and sealed with coverslips. Negative controls were included by omitting the RUNX3 antibody during the primary antibody incubation.
Evaluation of Immunostaining
The evaluation of nuclear RUNX3 staining was blindly and independently examined by 2 observers, including 1 dermatopathologist. In the few cases with discrepancy between the 2 observers, the immunostained slides were reviewed in a double viewing microscope so that the discrepancy was settled. The expression of RUNX3 was graded as positive when ≥5% of tumor cells showed immunopositivity. 12 Biopsies with <5% tumor cells showing immunostaining were considered negative.
Differences in demographic and clinical characteristics and expression levels of RUNX3 were evaluated by chi-square tests or Student t test between patient subgroups. Survival time was calculated from the date of melanoma diagnosis to the date of death or last follow-up. The Kaplan-Meier method and log-rank test were used to evaluate the effects of RUNX3 expression on the overall and disease-specific survival of patients. Univariate or multivariate Cox proportional hazards regression models were preformed to estimate the crude hazard ratios (HRs) or adjusted HRs and their 95% confidential intervals (CIs). A P value of <.05 was considered significant, and all tests were 2-sided. SPSS version 11.5 (SPSS Inc, Chicago, Ill) software was used for all analyses.
Clinicopathologic Features of TMAs
A total of 748 melanocytic lesions were used for TMA construction. Because of loss of biopsy cores or insufficient tumor cells present in the cores, 440 melanoma (202 cases in training set, and 238 cases in validation set), and 88 cases of nevi (25 normal nevi and 63 dysplastic nevi) could be evaluated for RUNX3 staining (Fig. 1). The distributions of selected demographic characteristics of melanoma patients in the training set and validation set are listed in Table 1.
|Variables||Training Set, No. (%)||Validation Set, No. (%)||Total, No. (%)|
|≤60||76 (55.9)||69 (43.9)||145 (49.5)|
|>60||60 (44.1)||88 (56.1)||148 (50.5)|
|Male||81 (59.6)||81 (51.6)||162 (55.3)|
|Female||55 (40.4)||76 (48.4)||131 (44.7)|
|Tumor thickness, mm|
|≤1.0||50 (36.8)||46 (29.3)||96 (32.8)|
|1.01-2.00||34 (25.0)||43 (27.4)||77 (26.3)|
|2.01-4.00||23 (16.9)||36 (22.9)||59 (20.1)|
|>4.00||29 (21.3)||32 (20.4)||61 (20.8)|
|Absent||113 (83.1)||124 (79.0)||237 (80.9)|
|Present||23 (16.9)||33 (21.0)||56 (19.1)|
|Lentigo maligna||31 (22.8)||22 (14.0)||53 (18.1)|
|Superficial spreading||48 (35.3)||57 (36.3)||105 (35.8)|
|Nodular||22 (16.2)||29 (18.5)||51 (17.4)|
|Unspecified||35 (25.7)||49 (31.2)||84 (28.7)|
|Sun-protected||110 (80.9)||113 (72.0)||223 (76.1)|
|Sun-exposed||26 (19.1)||44 (28.0)||70 (23.9)|
|≤59||36 (54.5)||42 (51.9)||78 (53.1)|
|>59||30 (45.5)||39 (48.1)||69 (46.9)|
|Male||45 (68.2)||52 (64.2)||97 (66.0)|
|Female||21 (31.8)||29 (35.8)||50 (44.0)|
|I||77 (38.1)||83 (34.9)||160 (36.4)|
|II||59 (29.2)||74 (31.1)||133 (30.2)|
|III||22 (10.9)||38 (16.0)||60 (13.6)|
|IV||44 (21.8)||43 (18.0)||87 (19.8)|
There were no significant differences in the distribution of the age, sex, American Joint Committee on Cancer (AJCC) stage, tumor thickness, ulceration, histology subtypes, and tumor site between the patients in the training set and those in the validation set. Thus, the 2 sets were combined to increase the study power. For the 293 primary melanoma cases, there were 162 men and 131 women, with age ranging from 7 to 95 years (median, 60 years). One hundred seventy-three primary melanomas were ≤2.0 mm thick, and the other 120 tumors were >2.0 mm thick. Ulceration was observed in 56 cases. For the histologic subtype, 53 tumors were lentigo maligna melanomas, 105 tumors were superficial spreading melanomas, 51 tumors were nodular melanomas, and 84 tumors were nonspecified. Seventy melanomas were located in sun-exposed sites (head and neck), and 223 were located in sun-protected sites (other locations). One hundred forty-seven melanoma metastases were available for RUNX3 staining evaluation, including 97 men and 50 women, with age ranging from 23 to 94 years (median, 59 years). We also applied AJCC criteria to RUNX3 evaluation in all melanoma patients; AJCC staging is done at the time of diagnosis. Among the 440 cases, 160 tumors were at AJCC stage I, 133 were at stage II, 60 were at stage III, and 87 were at stage IV (Table 1).
Loss of RUNX3 Expression Correlates With Melanoma Progression
Positive RUNX3 staining decreased from 56% and 54% in normal and dysplastic nevi to 33% in primary melanomas and further decreased to 24% in metastatic melanomas (Fig. 2). There was no significant difference in RUNX3 staining between normal nevi and dysplastic nevi (P = .863, chi-square test). However, significant differences for positive nuclear RUNX3 staining were observed between dysplastic nevi and primary melanomas (P = .002, chi-square test), between dysplastic nevi and melanoma metastases (P < .001, chi-square test), and between primary melanoma and melanoma metastases (P = .045, chi-square test).
RUNX3 Expression in Melanoma and Clinicopathologic Parameters
In all 440 melanoma patients, we found that positive RUNX3 expression significantly decreased from 33% in early stage (AJCC I and II) to 24% in advanced stage (AJCC III and IV) (P = .045, chi-square test, Table 2). In primary melanoma, positive RUNX3 expression was found in 36% of patients without ulceration, compared with only 21% of patients with ulceration (P = .039, chi-square test; Table 2). Moreover, positive RUNX3 expression was found in 45% of lentigo maligna melanoma patients, compared with 30% of patients with other kinds of melanoma, and the difference was significant (P = .037, chi-square test; Table 2). For thickness, positive RUNX3 expression was observed in 36%, 31%, 34%, and 30% of the biopsies in AJCC T1 to T4, respectively. There is no significant difference in RUNX3 staining among different T stages (P = .803, chi-square test; Table 2). We did not find significant correlations between RUNX3 expression and other clinicopathologic variables in primary melanoma, including age, sex, or location. In addition, RUNX3 expression was not correlated with age or sex in metastatic melanoma (Table 2).
|Negative, No. (%)||Positive, No. (%)||Total||Pa|
|≤60||94 (64.8)||51 (35.2)||145||.457|
|>60||102 (68.9)||46 (31.1)||148|
|Male||108 (66.7)||54 (33.3)||162||.927|
|Female||88 (67.2)||43 (32.8)||131|
|Tumor thickness, mm|
|≤1.0||61 (63.5)||35 (36.5)||96||.803|
|1.01-2.00||53 (68.8)||24 (31.2)||77|
|2.01-4.00||39 (66.1)||20 (33.9)||59|
|>4.00||43 (70.5)||18 (29.5)||61|
|Absent||152 (64.1)||85 (35.9)||237||.039|
|Present||44 (78.6)||12 (21.4)||56|
|Lentigo maligna||29 (54.7)||24 (45.3)||53||.037b|
|Superficial spreading||76 (72.4)||29 (27.6)||105|
|Nodular||37 (72.5)||14 (27.5)||51|
|Unspecified||54 (64.3)||30 (35.7)||84|
|Sun-protected||151 (67.7)||72 (32.3)||223||.595|
|Sun-exposed||45 (64.3)||25 (35.7)||70|
|≤59||60 (76.9)||18 (23.1)||78||.825|
|>59||52 (75.4)||17 (24.6)||69|
|Male||77 (79.4)||20 (20.6)||97||.206|
|Female||35 (70.0)||15 (30.0)||50|
|I-II||196 (66.9)||97 (33.1)||293||.045|
|III-IV||112 (76.2)||35 (23.8)||147|
RUNX3 Expression and Patient Survival
A total of 421 patients (192 from the training set and 229 from the validation set) had complete follow-up and clinical information (Fig. 1). The Kaplan-Meier analyses revealed that negative RUNX3 expression was associated with poor overall (P < .001) and disease-specific 5-year survival (P < .001) in the training set (Fig. 3A), and this association was confirmed in the validation set (P = .001 for overall survival and P = .003 for disease-specific survival, Fig. 3B). When these 2 sets were combined, the correlation between RUNX3 expression and overall or disease-specific 5-year survival was more significant (P < .0001 for both; Fig. 4A). Among the 421 patients, there were 275 cases of primary melanoma and 146 cases of metastatic melanoma. The univariate analysis revealed that age, thickness, and ulceration were all significantly associated with overall and disease-specific 5-year survival outcome (Table 3).
|Variable||Patients, No. (%)||Overall Survival||Disease-Specific Survival|
|Deaths||Deaths Rate, %||HR (95% CI)||P||Deaths||Deaths Rate, %||HR (95% CI)||P|
|>60||142 (51.6)||56||39.4||2.72 (1.69-4.36)||32||22.5||2.41 (1.44-4.02)|
|Female||126 (45.8)||39||31.0||1.13 (0.73-1.74)||23||18.2||1.01 (0.62-1.64)|
|>2.0||118 (42.9)||52||44.1||3.29 (2.08-5.19)||40||33.9||4.49 (2.63-7.67)|
|Present||54 (19.6)||30||55.5||3.76 (2.38-5.93)||23||42.6||4.10 (2.49-6.75)|
|Sun-exposed||66 (24.0)||23||34.8||1.49 (0.92-2.42)||13||19.7||1.62 (0.96-2.75)|
|Positive||89 (32.4)||16||18.0||0.40 (0.23-0.70)||11||12.4||0.37 (0.20-0.69)|
|>59||69 (47.3)||49||71.0||0.88 (0.60-1.29)||42||60.9||0.79 (0.53-1.17)|
|Female||50 (34.8)||39||78.0||1.22 (0.82-1.80)||38||76.0||1.34 (0.89-2.00)|
|Positive||35 (24.0)||19||54.3||0.48 (0.29-0.79)||18||51.4||0.51 (0.30-0.84)|
In primary melanoma, negative RUNX3 staining correlated with both overall and disease-specific 5-year survival (P = .001 for both, log-rank test; Fig. 4B). Furthermore, negative RUNX3 expression was also associated with poor overall (P = .003) and disease-specific 5-year survival (P = .008) in metastatic melanoma patients (Fig. 4C). Cox proportional hazard regression analysis showed that positive RUNX3 expression was a significantly favorable prognostic factor in both primary melanoma (HR, 0.40; 95% CI, 0.23-0.70 for overall survival; HR, 0.37; 95% CI, 0.20-0.69 for disease-specific survival) and metastatic melanoma (HR, 0.48; 95% CI, 0.29-0.79 for overall survival; HR, 0.51; 95% CI, 0.30-0.84 for disease-specific survival, Table 3).
Next, we examined whether positive RUNX3 expression is an independent prognostic marker for melanoma patients' survival by multivariate Cox proportional hazard analysis. In primary melanoma, our results clearly indicate that similarly to tumor thickness and presence of ulceration, which have been widely accepted as independent prognostic factors for melanoma patient survival, RUNX3 expression is an independent prognostic factor for both overall (HR, 0.41; 95% CI, 0.23-0.71; P = .002) and disease-specific 5-year survival (HR, 0.37; 95% CI, 0.20-0.70; P = .002; Table 4). Furthermore, RUNX3 expression was also correlated with both overall (HR, 0.46; 95% CI, 0.28-0.76; P = .002) and disease-specific 5-year survival (HR, 0.47; 95% CI, 0.28-0.79; P = .004) in metastatic melanoma patients (Table 4).
|Variable||Overall Survival||Disease-Specific Survival|
|βa||SE||HR||95% CI||Pb||βa||SE||HR||95% CI||Pb|
The role of RUNX3 in tumorigenesis has been studied extensively in recent years. Reduced RUNX3 expression has been observed in numerous types of human cancers. In this study, to better understand the role of RUNX3 activity in melanoma progression, we used TMA technology and immunohistochemistry to investigate RUNX3 activity in 528 cases of pigmented skin lesions at different stages. Our results demonstrated that RUNX3 expression decreases as melanoma progresses (Fig. 2). Furthermore, RUNX3 expression is significantly correlated with tumor stage, ulceration, and subtype and is correlated with 5-year survival in melanoma patients. To our knowledge, this is the first study to analyze RUNX3 protein using TMA technology and immunohistochemistry in melanoma, and its association with tumor progression and patient survival.
Kitago et al 13 reported that RUNX3 mRNA level was down-regulated in melanoma cell lines compared with that in normal melanocytes. They also examined RUNX3 mRNA expression level in 12 cases of normal skin, 82 cases of primary melanoma, and 41 cases of metastatic melanoma, and found RUNX3 was significantly down-regulated in melanoma. This observation was confirmed and extended by our study, in which we constructed TMA with a larger series of melanocytic lesions and showed that the RUNX3 protein was significantly reduced from normal and dysplastic nevi to primary and metastatic melanoma.
Our results showed that RUNX3 protein expression was reduced in primary melanomas compared with normal nevi and dysplastic nevi, and further suppressed in metastatic melanomas compared with primary melanoma. This indicated that decreased RUNX3 activity might be a common requirement for the transformation from benign neoplasia to malignancies, as well as in tumor progression from primary tumors to metastatic disease in melanoma.
In our study, the 5-year survival of melanoma patients with negative RUNX3 expression in the training set was significantly worse than that of melanoma patients with positive RUNX3 expression, which was confirmed in the validation set. In addition, RUNX3 expression was also correlated with 5-year survival in both primary and metastatic melanoma patients. Cox proportional hazard analysis further indicated that positive RUNX3 expression was an independent prognostic marker for melanoma. Moreover, reduced RUNX3 expression significantly correlates with advanced stages of melanoma and tumor ulceration. One of the most important factors of melanoma prognosis is AJCC stage at diagnosis, which confirms that early detection is crucial for the patient outcome. Ulceration was also identified as an adverse prognostic factor for survival. The importance of ulceration has led to its inclusion in the most recent AJCC staging classification for malignant melanoma. 14, 15
Our findings indicate that RUNX3 is an important prognostic factor in melanoma and are concordant with the results on other cancers, such as gastric cancer and esophageal cancer. 16, 17 The expression level of RUNX3 may provide the opportunity to identify subgroups of patients with melanoma who are at high risk of unfavorable survival. Moreover, studies showed that RUNX3 expression could induce apoptosis of tumor cells, 18 and the low therapeutic efficacy in melanoma treatment is largely because of a relative inability to induce apoptosis. 19 Thus, the RUNX3 expression signature may be used in addition to investigate drug response and chemosensitivity of melanoma, leading to more personalized treatment.
Transforming growth factor-β (TGF-β) is a ubiquitously expressed cytokine that plays very important roles in many cellular functions, and RUNX3 is a target of the TGF-β-mediated tumor suppressor pathway. 6 Malfunctions of TGF-β signaling are implicated in serious human diseases such as cancers. 20 TGF-β causes cell cycle arrest at the mid and late G1 phases of the cell cycle by inactivating cyclin-dependent kinases. RUNX3 and other signal transducers such as Smad are collectively required for the tumor suppressor activity of the TGF-β pathway. 6, 21 Recently, TGF-β was also found to be a novel diagnostic and prognostic marker for melanoma. 22 RUNX3, being a downstream effector of TGF-β, might provide an explanation for its broad involvement in human tumorigenesis, including melanoma.
Moreover, a recent study found that RUNX3 forms a ternary complex with β-catenin/TCF4 and attenuates Wnt signaling, 9 linking the Wnt and TGF-β signaling pathways. Aberrant expression of Wnt signaling has been implicated in various human cancers, including melanoma. 23 Furthermore, restoration of Wnt inhibitory factor-1 expression can suppress melanoma tumor growth through down-regulation of the Wnt signaling pathway. 24 These functional studies demonstrated that RUNX3 may play its tumor suppressor roles through the Wnt and TGF-β pathways. Because RUNX3 is a transcription factor, identification of tissue- and stage-specific target genes of RUNX3 functioning as a tumor suppressor is an important step forward, and combination of these genes would improve the potential of RUNX3 as a biomarker for cancer diagnosis, prognosis, and personalized therapy.
In conclusion, our study shows that RUNX3 expression is significantly decreased with progression of human melanoma. Strikingly, our data indicate that positive RUNX3 expression correlates with a better 5-year survival of melanoma patients and is an independent prognostic factor for melanomas. Moreover, the prognostic significance of RUNX3 expression in the validation set was consistent with that in the training set, greatly increasing the reliability of the study and reducing the probability of a false-positive result. These data suggest that RUNX3 plays an important role in melanoma pathogenesis and it may serve as a promising prognostic marker for malignant melanoma.
CONFLICT OF INTEREST DISCLOSURES
This work was supported by the Canadian Institutes of Health Research (MOP-84559 and MOP-93810) and Canadian Dermatology Foundation (G. Li), Michael Smith Foundation for Health Research Postdoctoral Fellowship (G. Chen), and Canadian Institutes of Health Research Skin Research Training Centre PhD Scholarship (Y. Cheng).