Association of nuclear annexin A1 with prognosis of patients with esophageal squamous cell carcinoma

Abstract

Although recent progress has been made in the diagnosis and treatment of cancer, the prognosis of esophageal squamous cell carcinoma (ESCC) remains poor. The identification of biomarkers for ESCC prognosis is important for treatment decisions. The aim of this study was to evaluate the relationship between the expressions of Annexin A1 (ANXA1), three prime repair exonuclease 1 (TREX1) and apurinic/apyrimidinic endonuclease-1 (APE1) and clinical outcome of patients with ESCC. The expressions of ANXA1, TREX1 and APE1 in 93 pairs of ESCC and paracancerous tissues were tested using immunohistochemistry. ANX1, TREX1 and APE1 were dysregulated in ESCC. Nuclear expressions of ANXA1 and APE1 were significantly associated with pathologic type (P = 0.004 and 0.040, respectively). Patients with low expression of nuclear ANXA1 had a better prognosis than those with high expression of nuclear ANXA1 (HR = 0. 448, 95% CI 0.236-0.849, P = 0.014), especially for those with histologic grade 1 and 2 (HR = 0.303, 95% CI: 0.155-0.593, P < 0.001). In conclusion, nuclear ANXA1 may be potentially used as a prognostic biomarker for ESCC.

Introduction

Esophageal cancer is the fifth most common cancer and the fourth most cause of cancer-related death in China [1]. Esophageal squamous cell carcinoma (ESCC) is the most common histological type, accounting for > 90% of cases [2]. Despite the improvement achieved in the diagnosis and treatment, the prognosis of ESCC remains poor, with the five-year survival rate of ~30% [3]. Most ESCC patients who undergo curative resection will eventually relapse and die of treatment-resistant disease, with local and regional recurrence being the most prevalent pattern of failure. Treatment for ESCC remains one of the most challenging tasks for cancer clinicians. Therefore, there is an urgent need for safer and more effectively therapies to improve the prognosis of ESCC patients.

Annexins are a family of Ca²⁺-regulated phospholipid-binding proteins, with 12 members in mammals, which have been implicated in the regulation of several biological processes, including cell differentiation, proliferation and apoptosis [4]. There is increasing evidence that annexins play important roles in cancer incidence and progression [5-11]. Annexin A1 (ANXA1), the first characterized member of the annexin superfamily, is an intracellular protein which is aberrantly expressed in many types of cancer, such as gastric [6,9,12], breast [10,13] and esophageal cancer [14,15]. Dysregulation of ANXA1 is related to the incidence, invasion, metastasis and drug resistance of cancers [16].

DNA repair enzymes are vitally important for protecting cells against damage caused by endogenous and exogenous agents. Apurinic/apyrimidinic endonuclease-1 (APE1) is the major AP endonuclease, which is important for the base excision repair pathway. In addition to DNA repair activity, APE1 has important roles in protection against oxidative stress- and hypoxia-induced apoptosis and necrosis [17]. Overexpression of APE1 in cancer cells is closely linked to poor prognosis and chemo- and radio-resistance [18]. Three prime repair exonuclease 1 (TREX1), the major 3’ to 5’ exonuclease in mammalian cells, excises bases from 3’ end of single- and double-stranded DNA templates with a preference for mismatched nucleotides [19]. In human fibroblasts, TREX1 was predominantly localized in the cytoplasm and translocated into the nucleus upon ultraviolet light exposure [20]. However, TREX1 is overexpressed in the nucleus of cancer cells [20].

In the present study, we investigated the expression levels of ANXA1, TREX1 and APE1 in 93 patients with ESCC and examined their associations with clinicopathologic factors and overall survival.

Materials and methods

Patients and tissue samples

A total of 93 patients were recruited who were underwent surgery between August 2008 and February 2010. All patients had histologically confirmed primary ESCC and had no history of other malignancy. ESCC and matched adjacent normal tissues were collected before patients were treated with any anti-cancer therapy, including surgery, chemotherapy and radiotherapy. This study was approved by the ethical committees of Taizhou People’s Hospital and National Engineering Center for Biochip at Shanghai. Written informed consent was obtained from each patient before enrolling in the study. The clinicopathologic features of ESCC patients were summarized in Table 1.

Table 1.

Clinicopathological characteristics of patients with ESCC

Characteristics	No. of Patients	%
Age (years)
median	63
range	49-85
Gender
female	22	23.7
male	71	76.3
Pathologic type
medullary	23	24.7
ulcerative	55	59.1
others	13	14.0
unknown	2	3.2
AJCC stage
stage I	7	7.5
stage II	52	55.9
stage III	29	31.2
stage IV	1	1.1
unknown	4	4.3
LNM
negative	55	59.1
positive	37	39.8
unknown	3	3.2
Tumor size (cm)
median	4.0
range	1.7-10
Histologic grade
well	15	16.1
moderate	63	67.7
poor	15	16.1

Tumor tissue microarray (TMA) and immunohistochemistry (IHC)

ESCC TMA was constructed as previously described [21]. Hematoxylin- and eosin-stained slides were reviewed by a pathologist, and representative areas that contained invasive tumor cells were marked on both the slides and corresponding paraffin block for TMA construction. TMA was constructed using an automated tissue arrayer (Beecher Instruments, Sun Prarie, WI). A single 600-μm tissue core was taken from each donor block and assembled into a recipient block. Subsequently, sections (4 μm thick) were cut from the array blocks and prepared for IHC analysis.

Sections were deparaffinized with xylene and rehydrated through graded alcohol. Antigen retrieval was performed using microwave treatment for 10 min in 10 mM sodium citrate buffer (pH 6.0). Endogenous peroxidase activity was quenched by incubating the sections in methanol with 6% hydrogen peroxide. Slides were then incubated with antibodies for 30 min (anti-ANXA1 antibody, 1:3000; anti-TREX1 antibody, 1:250; anti-APE1 antibody, 1:200). Staining was completed after a 10-min incubation with a freshly prepared substrate–chromogen solution (20 μL DAB chromogen per 1 mL of PBS), which results in a brown-colored precipitate at the antigen site. Slides were subsequently counterstained with hematoxylin, dehydrated, and mounted.

The slides stained by IHC were assessed by two pathologists who were blinded to clinical information. The staining intensity of cancer cells was scored as 0, negative; 1, weak; 2, moderate; 3, strong staining. For statistical evaluation, tumors were scored as 0, non-staining; 1, 1-10%; 2, 11-50%; 3, 51-80%; 4, 81-100% positive cells. The total histological score, which was the result of multiplication of intensity and percentage scores, was utilized to determine the result. The total histological score < 4 indicated as a low level of expression, whereas a total histological score ≥ 4 denotes a high level of expression [21].

Statistical analysis

The expression levels of ANXA1, TREX1 and APE1 between ESCC and paracancerous tissues were compared using the non-parametric Mann-Whitney U-test. The relationship between ANXA1, TREX1 and APE1, and clinicopathologic characteristics were tested by Chi-square test or Fisher’s exact test. The Kaplan-Meier method and the long-rank test were performed to compare the survival rates. Hazard ratios (HRs) and corresponding 95% confidence intervals (CIs) were calculated using Cox regression models for overall survival in uni- and multivariate analysis. A P-value < 0.05 was deemed statistically significant. All analyses were performed using the SPSS 19.0 software package (Chicago, IL, USA).

Results

The expression levels of ANXA1, TREX1 and APE1 in ESCC and paracancerous tissues

The expression levels of ANXA1, TREX1 and APE1 in 93 ESCC tissues were evaluated by IHC. The ANXA1-positive site was located in the cytoplasm and nucleus (Figure 1). The majority of ESCC showed negative or low expression of cytoplasmic and nuclear ANXA1 (68/93, 73.1%; 58/93, 62.4%, respectively), whereas a subset (25/93, 26.9%; 35/93, 37.6%) of ESCC showed high expression of cytoplasmic and nuclear ANXA1. The levels of cytoplasmic ANXA1 in cancer tissues were significantly higher than those in paracancerous tissues (P < 0.001), but nuclear ANXA1 showed a lower expression in ESCC tissues (P < 0.001). TREX1 was mainly localized in nucleus of both ESCC and adjacent non-cancerous cells. A weak positive cytoplasmic staining for TREX1 was observed in 26% (29/93) of cancer samples, whereas positive nuclear staining was found in 93.5% (87/93) of cancer samples. There were significant difference in cytoplasmic and nuclear expression of TREX1 between cancer and paracancerous tissues (P = 0.038 and < 0.001, respectively). APE1 was mainly localized in nucleus, whereas cytoplasmic APE1 expression was relatively rare. Nuclear expression of APE1 was observed in 100% (93/93) of cancer samples. The levels of nuclear APE1 in cancer tissues were also significantly higher than those in paracancerous tissues (P < 0.001).

Figure 1.

Immunohistochemical analysis of ANAX1, TREX1 and APE1 in ESCC and paracancerous tissues. A: Negative ANXA1 expression in adjacent non-cancerous tissue. B: Positive ANXA1 expression in ESCC tissue (cytoplasmic and nuclear staining). C: Negative TREX1 expression in adjacent non-cancerous tissue. D: Positive nuclear expression of TREX1 in ESCC tissue. E: Negative APE1 expression in adjacent non-cancerous tissue. F: Positive nuclear expression of APE1 in ESCC tissue.

The association of ANXA1, TREX1 and APE1 with clinicopathologic characteristics of ESCC patients

Since the expressions of cytoplasmic TREX1 and APE1 were low and rare in ESCC, respectively, cytoplasmic expressions of TREX1 and APE1 were excluded from further analysis. Nuclear expressions of ANXA1 and APE1 were significantly associated with pathologic type (P = 0.004 and 0.040, respectively) (Table 2). No other difference between ANXA1, TREX1 and APE1 expression and clinicopathologic characteristics was found.

Table 2.

Associations between the expressions of ANXA1, TREX1 and APE1 and the clinicopathological features

Variables	Nuclear ANXA1			Cytoplasmic ANXA1			Nuclear TREX1			Nuclear APE1
	Low	High	P value	Low	High	P value	Low	High	P value	Low	High	P value
Age (years)			0.271			0.334			0.809			0.469
≤ 60	18	15		22	11		9	24		10	23
> 61	40	20		46	14		15	45		14	46
Gender			0.803			1			0.578			0.578
male	45	26		52	19		17	54		17	54
female	13	9		16	6		7	15		7	15
Size (cm)			0.83			0.636			0.472			0.811
> 4	25	14		30	9		12	27		11	28
≤ 4	33	21		38	16		12	42		13	41
Pathologic type			0.004			0.376			0.765			0.040
medullary	19	4		19	4		5	18		8	15
ulcerative	27	28		40	15		16	39		9	46
others	11	2		8	5		3	10		6	7
Histologic grade			0.923			0.626			0.349			0.677
well	9	6		10	5		3	12		3	12
moderate	39	24		48	15		19	44		18	45
poor	10	5		10	5		2	13		3	12
LNM			0.263			0.458			0.627			0.617
positive	25	10		28	7		10	25		7	28
negative	32	23		39	16		13	42		14	41
AJCC stage			0.764			0.606						0.937
I	4	3		4	3		3	4	0.293	2	5
II	32	20		38	14		10	42		12	40
III	20	9		23	6		10	19		7	22
IV	1	0		1	0		0	1		0	1

Survival analyses

Overall survival was calculated as the time from the date of ESCC diagnosis to death. Follow-up period ranged from 0.13 month to 52 months (median, 35 months). Among 93 patients with ESCC, 42 (45.2%) patients died as a result of disease progression during the follow-up. The overall survival rates at 1 and 3 years were 65.6% and 52.0%, respectively. Twenty patients of the 35 who had high expression of nuclear ANXA1 died from ESCC (20/35, 57.1%). In the group of patients with low expression of nuclear ANXA1, 22 died from the disease (22/58, 37.9%). ESCC patients with low expression of nuclear ANXA1 had a significantly longer survival time compared with those with high expression of nuclear ANXA1 (HR = 0.464, 95% CI 0.252-0.855, P = 0.014) (Table 3, Figure 2). Furthermore, TNM and LNM were significantly associated with shorter survival time in univariate analyses (Table 3). Upon multivariate analysis, only low expression of nuclear ANXA1 was significantly associated with better survival rate (HR = 0. 448, 95% CI 0.236-0.849, P = 0.014). Stratified analysis revealed that low expression of nuclear ANXA1 was significantly associated with better prognosis in ESCC patients with TNM stages III and IV (HR = 0.212, 95% CI: 0.050-0.899, P = 0.035) or histologic grade 1 and 2 (HR = 0.371, 95% CI: 0.193-0.713, P = 0.003). After adjustment for LNM and TNM, the association between nuclear ANXA1 and overall survival in patients with histologic grade 1 and 2 remained significant (HR = 0.303, 95% CI: 0.155-0.593, P < 0.001).

Table 3.

Univariate and multivariable Cox regression analysis of overall survival (n = 93)

Features	Univariate analysis		Multivariable analysis
	HR (95% CI)	P value	HR (95% CI)	P value
Age (years), > 60 vs ≤ 60	1.016 (0.540-1.914)	0.961
Gender, female vs male	0.782 (0.374-1.635)	0.513
Pathologic type, medullary vs others	0.932 (0.564-1.540)	0.784
Histologic grade, 1+2 vs 3	0.648 (0.255-1.649)	0.362
Tumor size (cm), > 4 vs ≤ 4	1.313 (0.703-2.449)	0.393
LNM, positive vs negative	2.181 (1.092-4.357)	0.027	1.425 (0.443-4.591)	0.552
TNM, III+IV vs I+II	2.677 (1.235-5.802)	0.013	3.515 (0.973-12.700)	0.055
Nuclear ANXA1 (low vs high)	0.464 (0.252-0.855)	0.014	0.448 (0.236-0.849)	0.014
Cytoplasm ANXA1 (low vs high)	1.129 (0.554-2.301)	0.738
Nuclear TREX1 (low vs high)	1.138 (0.560-2.316)	0.721
Nuclear APE1 (low vs high)	1.328 (0.635-2.777)	0.451

Figure 2.

Kaplan-Meier survival curves of ESCC patients according to nuclear expression of ANXA1. A: All ESCC patients. B: Patients with TNM stages III and IV. C: Patients with histologic grade 1 and 2.

Discussion

In the present study, we investigated the relationship between the expressions of ANXA1, TREX1 and APE1 and clinical outcome of patients with ESCC. ANXA1, TREX1 and APE1 were dysregulated in ESCC. In addition, decreased expression of nuclear ANXA1 in ESCC was correlated with a favorable prognosis. These findings indicate that nuclear ANXA1 may have an influence on the progression of ESCC.

The role of ANXA1 in cancer is complicated by the fact that ANXA1 is downregulated in some cancers, including gastric [6,12], breast [10,13], prostate [22], cervical [23] and thyroid cancer [24], but upregulated in other types of cancer, such as pancreatic cancer [25]. A controversy exists regarding the expression of ANXA1 in gastric cancer. Cheng et al. [6] and Yu et al. [12] reported that ANXA1 was downregulated in gastric cancer, but Jorg et al. [9] showed overexpression of ANXA1 in gastric cancer. TNM stage may be the one of the main reasons for the inconsistent and contradictory findings. ANXA1 appears to play multifaceted roles in cancer, and acts as context-depending tumor suppressor or oncogene. Inhibition of ANXA1 facilitates the growth of prostate cancer cells [5], whereas overexpression of ANXA1 facilitates the migration and invasion of gastric cancer cells [6]. Previous studies revealed that ANXA1 was downregulated in both ESCC and esophageal adenocarcinoma [14,15]. In the current study, the majority of ESCC showed negative or low expression of ANXA1. However, there was no association between both cytoplasmic and nuclear expression of ANXA1 and histologic grade, which was inconsistent with previous study that ANXA1 was predominantly in well-differentiated ESCC [15]. In this study, most cases were TNM stages I and II, while cases were almost TNM stage III in previous study [15,26], which may partly explain the discrepant results. Furthermore, recent studies have demonstrated that the levels of ANXA1 may influence the survival in cancer patients [6,27-30]. Although no association between cytoplasmic expression of ANXA1 and overall survival of ESCC patients was observed, patients with low expression of nuclear ANXA1 had longer survival time, which was in agreement with previous study [27]. Further studies are required to fully understand roles of ANXA1 in ESCC.

Although TREX1 knockout mice does not show an increase in cancer incidence [31], inhibition of TREX1 promotes cell death in malignant glioma and melanoma cells treated with anti-cancer agents [20]. Furthermore, Dong et al. reported that polymorphism in TREX1 [32] was associated with survival in patients with pancreatic cancer. APE1 are frequently overexpressed in some types of cancer, such as gastric cancer [33,34]. The levels of APE1 have previously been shown to correlate with survival in cancer patients [33]. In the present study, TREX1 and APE1 were upregulated in ESCC tissues, which were consistent with previous studies [20,34]. However, the level of cytoplasmic TREX1 was significantly lower than those of nuclear TREX1 in paracancerous tissues, which differ from fibroblasts [20]. Tomicic et al. [20] found that nuclear TREX1 seems to be associated with replication. The details of the mechanism of nuclear TREX1 in esophageal cell and ESCC need further studies. However, nuclear expression of TREX1 and APE1 did not correlate with survival. It was amazing that nuclear expressions of both ANXA1 and APE1 correlated with pathologic type. Are there different mechanisms in carcinogenesis between pathologic types? Further studies are warranted to verify our findings and to determine molecular mechanisms underlying the pathogenesis of ESCC.

In summary, high expression of nuclear ANXA1 was found to be a strong risk factor for the overall survival of ESCC. Even though further studies are required to precisely elucidate the role and significance of nuclear ANXA1 in the setting of ESCC, it may be a valuable biomarker for the prediction of ESCC prognosis.

Disclosure of conflict of interest

The authors declare that they have no conflict of interest.