The aberrant expressions of MACC1, ZEB1, and KLF4 in hepatocellular carcinoma and their clinical significance

Abstract

Background and objective: Metastasis-associated in colon cancer-1 (MACC1) is involved in the progression and metastasis of various cancers. Zinc finger E-box-binding homeobox 1 (ZEB1) is a key transcriptional factor of the epithelial-mesenchymal transition (EMT) that is involved in the migration and invasion of cancer cells. Kruppel-like factor 4 (KLF4) is a tumor suppressor that can inhibit tumor cell proliferation, migration, and metastasis. The purpose of this study was to investigate the expressions and clinical significance of MACC1, ZEB1, and KLF4 in hepatocellular carcinoma (HCC). Methods: We analyzed the expressions of MACC1, ZEB1, and KLF4 in 153 HCC specimens and their corresponding control specimens. The patients’ clinicopathological and follow-up data were also collected. Results: The rates of positive expression of MACC1 and ZEB1 were significantly higher in the HCC specimens than in the control specimens, and their expressions were positively associated with the number of tumors, grades of differentiation, lymph node metastasis (LNM), and tumor-node-metastasis (TNM) stages. Inversely, the rate of positive expression of KLF4 was significantly lower in the HCC specimens than it was in the control specimens, and its expression was negatively correlated with the number of tumors, grades of differentiation, LNM, and TNM stages. The patients who expressed MACC1 or ZEB1 had a reduced overall survival (OS) when compared with patients not expressing these proteins. However, the patients who expressed KLF4 had an increased OS when compared with patients who did not show any KLF4 expression. A multivariate analysis indicated that the expressions of MACC1, ZEB1, and KLF4 and tumor size, LNM, as well as the TNM stages were independent, prognostic factors for HCC patients. Conclusions: Therefore, positive expressions of MACC1, ZEB1, and KLF4 should be correlated with the duration of OS in patients with HCC and considered promising prognostic biomarkers, as well as potential therapeutic targets for HCC.

Introduction

New liver cancer cases were estimated at 783,000, and deaths were estimated at 746,000 worldwide in 2012 [1]. But in China, new liver cancer cases were estimated at 466,000 and deaths were estimated at 422,000 in 2015 [2], accounting for approximately 50% of the total number of new cases and deaths. The most common type of liver cancer is hepatocellular carcinoma (HCC). Many HCC patients are diagnosed at the advance stages in China because the disease does not show any apparent symptoms during the early stages.

Relapse and metastasis are the most common reasons for liver cancer treatment failure. Metastasis-associated in colon cancer 1 (MACC1) is considered an oncogene that was originally identified in the colon cancer cell line in 2009 [3]. Previous studies have demonstrated that MACC1 is able to induce the epithelial-mesenchymal transition (EMT) to promote tumor cell invasion and metastasis both in vitro and in vivo [3-6]. MACC1 is a key regulator of the hepatocyte growth factor (HGF)/MET signaling pathway that can bind to the promoter of the mesenchymal-epithelial transition (MET) gene to regulate its transcriptional activity [3,7]. It was reported that MACC1 is an effective and valuable biomarker in various types of cancer, especially in the prediction of metastasis and prognosis [8].

It is well known that EMT is one of the key mechanisms of carcinoma metastasis [9]. EMT occurs when epithelial cancer cells lose their epithelial features and acquire mesenchymal features that can promote invasion and metastasis [10,11]. Zinc finger E-box binding homeobox 1 (ZEB1), which is a key transcriptional regulator of EMT, consists of two zinc finger clusters and a centrally-located homeodomain responsible for DNA binding [12]. ZEB1 is able to induce tumor cell invasion and metastasis by promoting EMT [13]. The overexpression of ZEB1 is often found in various cancers and may be considered a metastatic and prognostic biomarker for many cancers [13-15].

Kruppel-like factor 4 (KLF4) is a zinc finger transcriptional factor that was originally identified as being expressed in the epithelial cells of the skin and intestines [16]. KLF4 can bind to the GC or CACCC-rich DNA sequences to regulate cell proliferation, differentiation, and apoptosis [17]. However, KLF4 plays dual functions in tumorigenesis and development since it can serve as a tumor suppressor or an oncogene [18,19]. For example, KLF4 has been shown to function as an oncogene in many tumors, such as skin squamous cell carcinoma and breast carcinoma [20,21]. A suppressed role of KLF4 was found in HCC, lung cancer, and gastric cancer [16,22,23].

Although these biomarkers are widely recognized in tumor initiation, progression and metastasis, studies on the role of these biomarkers in HCC remain unclear. The purpose of this study is to evaluate the hypothesis that these biomarkers associate with HCC progression and prognosis.

Methods

Patients and specimens

We recruited 153 recorded patients who were diagnosed with HCC from January 2012 to December 2013 by the Department of Pathology of our hospital and collected samples of cancer tissue and the corresponding normal liver tissue from all patients. This study is retrospective. All patients who had any history of chemo- or radio-therapy or other anti-cancer therapy were excluded. All HCC patients provided written informed consents. This study was authorized by the ethics committee of Bengbu Medical University and performed in accordance with the Declaration of Helsinki before it started. The patient data included clinicopathologic, demographic, and follow-up data (follow-up by phone or social applications). The follow-up data was calculated from the date of surgery to his/her death date or to December 2017. The TNM stages were evaluated in accordance with the 8^th edition of guidelines issued by the AJCC (American Joint Committee on Cancer). Grades of differentiation were evaluated in accordance with the guidelines issued by the WHO (World Health Organization). The specific characteristics are shown in Table 1.

Table 1.

The patients’ characteristics

Patients characteristics	Frequency (n)	Percentage (%)
Age (years)
< 60	68	44.4
≥ 60	85	55.6
Gender
Male	117	76.5
Female	36	23.5
Alcohol
No	91	59.5
Yes	62	40.5
Size (cm)
< 2.0	47	30.7
≥ 2.0, < 5.0	71	46.4
≥ 5.0	35	22.9
HBSAg
No	86	56.2
Yes	67	43.8
Number of tumors
1	82	53.6
> 1	71	46.4
Cirrhosis
No	86	56.2
Yes	67	43.8
Grades
Well + moderate	77	50.3
Poor	76	49.7
Lymph node metastasis
N0	147	96.1
Yes	6	3.9
TNM stages
I	52	34.0
II	64	41.8
III	31	20.3
IV A	6	3.9

Immunohistochemistry

All the HCC tissues and normal liver tissues were fixed in 10% buffered formalin solution and then embedded in paraffin. Continuous 4-μm-thick sections were cut. Immunohistochemical staining was performed using the Elivision^TM Plus method. The immunostaining procedure was performed following the kit’s instructions. Endogenous peroxidase activity was blocked by methanol containing a 3% H₂O₂ solution. The antigen repair used a citrate buffer solution (pH 6.0). Then all the sections were blocked with goat serum. Rabbit polyclonal antibody against human MACC1 (Santa Cruz, CA, USA) and ZEB1 (Abcam, MA, USA) and mouse monoclonal antibody against human KLF4 (Abcam, MA, USA) primary antibodies were added, then all sections were incubated at 4°C overnight. All the sections were developed in a diaminobenzidine (DAB) substrate solution. Finally, all the slices were re-dyed with hematoxylin and mounted with gum.

Evaluation of staining

Ten high-power-fields (HPF) were randomly selected to avoid intratumoral heterogeneity of biomarker expression. The immunostaining was interpreted by two pathologists who were blinded to the patients’ data and evaluated by semi-quantitative points. The staining results were scored in accordance with the staining intensity and staining extent. The staining intensity scoring was done as follows: no staining was 1; weak staining was 2; moderate staining was 2; and strong staining was 3. The staining extent scoring was done as follows: < 11% positive cells was 1; 11-50% positive cells was 2; 51-75% positive cells was 3; and > 75% positive cells was 4. The final scores (range 0-12) were calculated by multiplying the intensity score by the extent score. The expression was considered positive when the score was > 2. For tissues that were positive for MACC1, ZEB1 and KLF4, an average of the final score of each was taken.

Statistical analysis

All data were analyzed using SPSS 19.0 software (Chicago, IL, US). The countable data were subjected to a Chi-square test or Fisher’s exact test for comparisons between two groups. Univariate OS time analysis was performed using the Kaplan-Meier method with a log-rank test. Multivariate OS time analysis was performed using a Cox regression model test. P < 0.05 was defined as being indicative of statistically significant differences.

Results

The expressions of MACC1, ZEB1, and KLF4 in HCC, and their relationships to clinicopathologic characteristics

As shown in Figure 1A and 1B, the MACC+ expression was mainly situated in the cytoplasms. The MACC1+ expression in the HCC cells (60.8%, 93/153) was significantly higher than it was in the control cells (6.5%, 10/153; P < 0.001). The MACC1 expression was positively related to the number of tumors, grades of differentiation, and TNM stages. In contrast, there were no relationships between MACC1+ expression and patient age, gender, tumor size, alcohol status, HBSAg status, lymph node metastasis (LNM), or cirrhosis (P > 0.05; Table 2).

Figure 1.

Immunostaining of MACC1, ZEB1, and KLF4 in HCC and the control tissue (400 magnification). A. Negative staining MACC1 in the control tissue; B. Positive staining of MACC1 in the cytoplasms of cancer cells; C. Negative staining of ZEB1 in the control tissue; D. Positive staining of ZEB1 in the nuclei of cancer cells; E. Positive staining of KLF4 in the cytoplasms and nuclei of the control cells; F. Negative staining of KLF4 in the cancer cells.

Table 2.

The associations between the expressions of MACC1, ZEB1, and KLF4 and the clinicopathological characteristics of hepatocellular carcinoma (HCC)

Variables	MACC1		P	ZEB1		P	KLF4		P
	-	+		-	+		-	+
Age (years)			0.120			0.141			0.856
< 60	22	47		32	36		35	33
≥ 60	38	46		30	55		45	40
Gender			0.224			0.075			0.753
Male	49	68		52	65		62	55
Female	11	25		10	26		18	18
Alcohol			0.916			0.706			0.848
No	36	55		38	53		47	44
Yes	24	38		24	38		33	29
Size (cm)			0.246			0.206			0.160
< 2.0	22	25		24	23		30	17
≥ 2.0, < 5.0	28	43		25	46		34	37
≥ 5.0	10	25		13	22		16	19
HBSAg			0.671			0.168			0.196
No	35	51		39	47		41	45
Yes	25	42		23	44		39	28
Number			< 0.001			< 0.001			< 0.001
1	54	28		58	24		30	52
> 1	6	65		4	67		50	21
Cirrhosis			0.078			0.703			0.196
No	39	47		36	50		41	45
Yes	21	46		26	41		39	28
Grades			< 0.001			< 0.001			< 0.001
Well + moderate	44	33		44	33		29	48
Poor	16	60		18	58		51	25
LNM			0.082			0.082			0.029
No	60	87		62	85		74	73
Yes	0	6		0	6		6	0
TNM stages			< 0.001			< 0.001			< 0.001
I	44	8		44	8		14	38
II	7	57		7	57		46	18
III	9	22		22	20		14	17
IV A	0	6		0	6		6	0

As shown in Figure 1C and 1D, ZEB1+ expression was mainly situated at the nuclei. Similar to MACC1, the ZEB1+ expression in the HCC cells (59.5%, 91/153) was significantly higher than it was in the control cells (3.3%, 5/153; P < 0.001). Furthermore, the ZEB1+ expression positively related to the number of tumors, grades of differentiation, and TNM stages. And at the same time, there were no relationships between ZEB1+ expression and patient age, gender, tumor size, alcohol status, HBSAg status, LNM, or cirrhosis (P > 0.05; Table 2).

As shown in Figure 1E and 1F, KLF4+ expression was mainly situated in the nuclei and cytoplasms. Inversely to MACC1, the KLF4+ expression in HCC cells (47.7%, 73/153) was significantly lower than it was in the control cells (81.0%, 124/153; P < 0.01). Moreover, the KLF4+ expression was positively related to the number of tumors, grades of differentiation, LNM, and TNM stages. And there were no relationships between KLF4+ expression and patient age, gender, tumor size, alcohol status, HBSAg status, or cirrhosis (P > 0.05; Table 2).

The relationships among MACC1, ZEB1, and KLF4 in HCC

The spearman correlation coefficient analysis indicated a negative association between KLF4+ expression, and the MACC1+ or ZEB1+ expression was negative (r = -0.466; r = -0.304; respectively; both P < 0.001). There was a positive association between MACC1+ expression and ZEB1+ expression (r = 0.482, P < 0.001) (Table 3).

Table 3.

The correlation among the expressions of MACC1, ZEB1, and KLF4 in HCC

Variable	MACC1		r	P	ZEB1		r	P
	-	+			-	+
MACC1							0.482	< 0.001*
-					42	18
+					20	73
KLF4			-0.466	< 0.001@			-0.304	< 0.001@
-	14	66			21	59
+	46	27			41	32

^*positive association;

^@negative association.

Survival analysis

As shown in Figure 2A, a Kaplan-Meier analysis demonstrated that the OS time of the HCC patients who expressed MACC1 (26.7 ± 13.6 months) was significantly lower than that of the patients who did not express the protein (50.1 ± 14.7 months; log-rank = 56.588, P < 0.001). As shown in Figure 2B, the OS time for the ZEB1-positive patients (29.0 ± 14.7 months) was significantly lower than the OS time of the ZEB1-negative patients (46.0 ± 17.9 months; log-rank = 38.518, P < 0.001). As shown in Figure 2C, the OS time of the patients who expressed KLF4 (47.0 ± 14.3 months) was significantly higher than the OS time of the patients did not express protein (25.7 ± 14.9 months; log-rank = 43.181, P < 0.001). As shown in Figure 2D, the OS time of the patients who expressed a combination of KLF4-negative, MACC1-positive, and ZEB1-positive was significantly lower than the OS time of the patients who expressed KLF4-positive, MACC1-negative, and ZEB1-negative (log-rank = 111.165, P < 0.001).

Figure 2.

A Kaplan-Meier analysis of the survival rate of patients with HCC. (A) Overall survival of all patients in relation to MACC1 expression (log-rank = 56.588, P < 0.001); (B) Overall survival of all patients in relation to ZEB1 (log-rank = 38.518, P < 0.001); (C) Overall survival of all patients in relation to KLF4 (log-rank = 43.181, P < 0.001); In (A-C) analyses, the green line represents the positive staining of factors and the blue line represents the negative staining factors. (D) Overall survival of all patients in relation to the combination of KLF4, MACC1, and ZEB1 (log-rank = 111.165, P < 0.001). The green line represents the positive expression of KLF4 and the negative expression of MACC1 and ZEB1 and the blue line represents the negative expression of KLF4 and the positive expression of MACC1 and ZEB1. The red line represents other positive or negative expressions of the proteins.

A multivariate analysis demonstrated that the expressions of MACC1, ZEB1, and KLF4 and tumor size, LNM, as well as TNM stages were independent prognostic factors for HCC (Table 4).

Table 4.

Results of the multivariate analyses of overall survival (OS) time

Covariate	B	SE	P	HR	95% CI
MACC1	0.500	0.247	0.043	1.648	1.016-2.675
ZEB1	0.594	0.274	0.030	1.812	1.060-3.097
KLF4	-1.705	0.229	< 0.001	0.182	0.116-0.285
Tumor size	0.631	0.186	0.001	1.879	1.305-2.706
LNM	1.163	0.562	0.039	3.199	1.063-9.628
TNM stages	1.054	0.210	< 0.001	2.869	1.903-4.326

Discussion

HCC is a common cancer of the digestive system and is a highly heterogeneous cancer. The heterogeneity of HCC makes it difficult to evaluate the effectiveness of biomarkers. It has been demonstrated that MACC1 not only stimulates cell proliferation, mobility, migration, and metastasis in vitro, but it also promotes cell growth, migration, and metastasis in vivo [3,6,7,24]. In this study, we detected MACC1 expression in HCC and the corresponding normal liver tissues and found that the HCC tissue expressed higher levels of the protein than the control tissues. Furthermore, MACC1 expression positively related to the number of tumors, grades of differentiation, and TNM stages. The OS analysis showed that HCC patients expressing MACC1 survived for less time than patients who did not express the protein. The above results suggest that MACC1 expression in this study should be similar to previous studies in HCC [25,26], and they also suggest that MACC1 should be considered a usefully prognostic biomarker for HCC.

ZEB1, which consists of two zinc finger clusters and a centrally-located homeodomain, is a critical transcriptional regulator of EMT [12,13]. Previous studies have demonstrated that the overexpression of ZEB1 should promote tumor cell migration, invasion, metastasis, as well as the EMT [26,27]. In this study, ZEB1 overexpression was significantly related to the number of tumors, grades of differentiation, and TNM stages. Furthermore, the OS analysis indicated that the patients who expressed ZEB1 survived for less time than those who did not express ZEB1. These results are similar to other studies suggesting that ZEB1 should be a usefully prognostic biomarker for HCC [13,28,29].

It has been demonstrated that KLF4 plays a dual function as either a tumor suppressor or an oncogene in tumorigenesis [18,30]. In this study, our results demonstrated that KLF4 expression was inversely related to the number of tumors, grades of differentiation, LNM, as well as the TNM stages. In addition, the OS analysis indicated that patients expressing KLF4 lived longer than those who did not express KLF4. Our results are similar to the previous studies in HCC [22,31,32] and suggest that KLF4 should be considered a useful and valuable biomarker for the prediction of progression and prognosis.

In this study, a Kaplan-Meier analysis showed that patients who expressed MACC1, ZEB1, KLF4, or the co-expressions of MACC1, ZEB1, and KLF4 survived less or longer than those who did not express these proteins. A multivariate analysis showed that the expressions of MACC1, ZEB1, and KLF4, and tumor size, LNM, as well as TNM stages were independent prognostic factors for HCC patients. Our findings also suggested that the expressions of MACC1, ZEB1, and KLF4 should be considered useful and valuable biomarkers for predicting the progression and prognosis of HCC. A normal expression of KLF4 can inhibit the expression of the EMT-related proteins ZEB1, snail, and slug by activating the expressions of miR-153, miR-506, and miR-200b [33]. Down- or lost-regulation of KLF4 should be involved in the initiation of tumorigenesis and increase cell proliferation, invasion, and metastasis [17]. Meanwhile, the overexpression of MACC1 should be involved in tumorigenesis by promoting cell proliferation and also promote tumor cell invasion and metastasis by the activation of EMT via the HGF/MET signaling pathways [8,34]. ZEB1 can act as a driver of EMT through the inhibition of E-cadherin expression and the activation of the Fak/Src signal pathway [35,36]. Thus, EMT and the aberrant expression of KLF4 should promote cancer cell mobility, migration, and metastasis. The OS time of patients who expressed the combination of being KLF4-negative, MACC1-positive, and ZEB1-positive was significantly lower than the OS time of the patients who expressed KLF4-positive, MACC1-negative, and ZEB1-negative. This result suggests that the combined detection of KLF4, MACC1, and ZEB1 expression in HCC may predict patients’ prognosis at the early stages.

Conclusions

This study found that positive expressions of MACC1, ZEB1, and KLF4 are associated with time of OS among patients with HCC. Therefore, MACC1, ZEB1, and KLF4 should be thought of as useful and valuable biomarkers to predict metastasis and prognosis in HCC patients.

Disclosure of conflict of interest

None.

Abbreviations

MACC1

metastasis-associated in colon cancer 1

HCC

hepatocellular carcinoma

ZEB1

Zinc finger E-box-binding homeobox 1

KLF4

Kruppel-like factor 4

EMT

epithelial-mesenchymal transition

HGF/MET

hepatocyte growth factor/mesenchymal-epithelial transition

LNM

lymph node metastasis

TNM

tumor-node-metastasis

OS

overall survival

HPF

high-power-field