Overexpression of Survivin‑1, TAG‑72 and HERC5 in patients diagnosed with hepatocellular carcinoma in the Black Sea coast geographical area

Introduction

Despite the great achievements obtained in the early detection of hepatocellular carcinoma (HCC) through screening programs and application of targeted therapies with protease inhibitors such as sorafenib, the incidence and the poor survival rate reported are still high, especially in endemic areas for hepatotropic viruses (HBV, HCV and HDV) such as southeastern Europe (1-4). Panels of biomarkers are standardized to help clinicians in their efforts to improve knowledge in terms of better HCC characterization for more efficient therapies. A single biomarker still used for HCC diagnosis or follow-up is α-fetoprotein (AFP), which is considered as the gold standard of care. Yet, clinical evidence suggests that it does not help facilitate improvement in HCC progression, prognosis, or survival rates (5). In general, tumor cells present metabolic signatures compared to healthy cells, both at the tissue and bio-humoral levels. The detection of new tumor cell biomarkers, and their validation has presented new research goals for HCC characterization. Viral infections, alcohol abuse, dysmetabolic states [obesity, type 2 diabetes mellitus (T2DM), nonalcoholic steatohepatitis (NASH)], and other rare conditions causing subsequent chronic liver damage promote liver tumorigenesis through different mechanisms and this situation makes it difficult to standardize a panel of predictive biomarkers for HCC progression (6-9). Survivin-1, an anti-apoptotic protein modulated by the p53 gene, presents overexpression in 70% of Asian HCC patients with chronic viral hepatitis in whom mutations of p53-gene are apparent (10). The overexpression of this tissue biomarker has not been studied in other research. It represents an opportunity to study Survivin-1 due to the similarity between risk factors for HCC occurrence in our patients and Asian patients-a vast majority being infected by hepatitis B, C, and D viruses. Tumor-associated glycoprotein 72 (TAG-72) is a mucin-like membrane complex considered to be a feasible biomarker for unfavorable prognosis of adenocarcinomas in general, with potential applicability in HCC (11,12). HECT and RLD domain containing E3 ubiquitin protein ligase 5 (HERC5), a protein with a ligand role, activates the chemotaxis and the local infiltration with T lymphocytes, being considered a biomarker for predicting HCC recurrence, as well as the poor survival rates even for early stage HCC (13,14). In this scenario, we aimed to evaluate and validate the overexpression of Survivin-1, TAG-72, and HERC5 as tissue biomarkers for HCC characterization in patients from our geographical region and to standardize a local biomarker panel to be introduced in the current management guideline.

Materials and methods

Materials

Thirty liver specimens with a histopathological diagnosis of HCC (study group) and a similar number of liver tissue specimens of benign liver tumors (adenomas, HNF, regenerative nodules, and hemangiomas-control group) were selected from the Gastroenterology file database and Pathology Clinic registries from St. Apostle Andrew Emergency Clinical Hospital, Constanta and Fundeni Institute, Bucharest and compared in terms of Survivin-1, Tag-72, and HERC5 overexpression. All cases were registered in our databases in the last 3 years and 6 months (January 2017 to June 2020).

All cases histologically confirmed by pathologists from both clinics, were reinterpreted for the current study during the interval December 2019 to June 2020, and the immunohistochemistry study was conducted at the Research and Development Centre for the Morphologic and Genetic Study of Malignant Pathology (CEDMOG) and founded by ‘Ovidius’ University. The morphological features of the tumors were noted, establishing the histological type, the grade, and the stage of HCC based on the World Health Organization (WHO) Histological Classification for digestive tumors (15).

Demographic data of all patients (study and control group) providing the liver specimens, including age, sex, provenence, medical history, liver disease background-chronic viral infections with B, C, or D viruses, co-morbidities and laboratory parameters recorded at the time of hospital admission were obtained from the clinical files and are noted in Table I.

Table I Demographic, clinical, imaging and laboratory features of the HCC and control (benign tumor) group.

Table I

Demographic, clinical, imaging and laboratory features of the HCC and control (benign tumor) group.

	Benign tumors (control group) (N=30) n (%)	HCC group (study group) (N=30) n (%)	P-value	r-value
Age, mean ± SD, years	42.09±9.4	59.2±5.9	0.002	0.96
Sex
Male	17 (56.66)	25 (83.33)	0.023	0.24
Female	13 (43.33)	5 (16.66)	0.033	-0.32
Urban	12 (40.00)	15 (50.00)	0.051	0.70
Risk factors
Viral infection	3 (10.00)	21 (70.00)	0.001	0.98
Alcohol abuse	15 (30.00)	20 (66.66)	0.048	0.33
NASH/NAFLD	5 (16.66)	11 (36.66)	0.042	0.34
Metabolic disease	2 (6.66)	3 (10.00)	0.7	0.11
Clinical features
Hepatomegaly	3 (10.00)	22 (73.33)	0.005	0.91
Jaundice	11 (36.66)	20 (66.66)	0.044	0.59
Ascites	1 (3.33)	16 (53.33)	<0.001	0.99
Weight loss	12 (40.00)	24 (80.00)	0.003	0.94
UDB	-	7 (23.33))	-	-
Imaging features
Multifocal	6 (20.00)	10 (33.33)	0.037	0.28
Portal vein thrombosis	1 (3.33)	6 (20.00)	0.032	0.26
Ascites	1 (3.33)	9 (30.00)	0.027	0.21
Laboratory results
ALT UI/ml, mean ± SD↑	30±11.67 10 (33.33)	37.3±14.11 23 (76.66)	0.067	0.58
GGT UI/ml ↑	34.33±8.22 16 (53.33)	71.54±6.99 24 (80.00)	0.029	0.30
Albumin g/dl↓	4.3±1.2 1 (3.33)	2.3±1.4 22 (73.33)	<0.001	0.99
Bilirubin mg/dl↑	1.9±0.5 11 (16.66)	4.3±2.7 24 (80.00)	0.002
INR↑	1.4±0.8 10 (33.33)	1.9±1.0 12 (40.00)	0.026	0.44
AFP ng/ml ↑	12 (40.00)	23 (76.66)	0.011	0.29
≥6 ng/ml	6.92±2.6 11 (18.60)	8.22±3.1 14 (46.66)	0.051	0.70
>180 ng/ml	204.11±17.77 1 (4.80)	308.56±44.01 9 (30.00)	0.015	0.18
BCLC classification
A	-	13	-	-
B	-	11	-	-
C	-	6	-	-

[i] HCC, hepatocellular carcinoma; NASH, nonalcoholic steatohepatitis; NAFLD, non-alcoholic fatty liver disease; ALT, alanine transaminase; GGT, γ-glutamyl transferase; AFP, α-fetoprotein; BCLC, Barcelona Clinic Liver Cancer; INR, International Normalized Ratio. Laboratory results are expressed as the mean increased/decreased level (top value) and the number of patients (%) with an increased/decreased level (lower data).

The Ethics Committee of Emergency Clinical Hospital St. Apostle Andrew of Constanta approved the study following European and local regulations (no. 32/22.11.2019).

Methods

For the immunohistochemical (IHC) assessment, the representative samples were chosen, and 4-µm sections of formalin-fixed, paraffin-embedded tissue blocks were obtained for each case enrolled. Epitope retrieval was conducted prior to incubation of tissue sections with a panel of three primary antibodies (ready-to-use) from Novus Biological: Survivin-1 (NB100-911 clone), Tag-72 (CC49 clone), and HERC5 (NBP-91985 clone). The immunostaining protocol for each antibody used was provided by the manufacturer. As chromogen, we used 3,3'diaminobenzidine (DAB), and brown staining was obtained. The final step was represented by counterstaining all slides with Mayer's Hematoxylin. Positive control was used for each antibody: Human testis for HERC5 antibody, malign melanoma for Survivin-1 antibody, and human prostate carcinoma for the Tag-72 antibody. Comparisons of the studied biomarker overexpression from HCC tissue samples with a matched non-HCC group of normal liver tissue specimens were made.

Statistical analysis

Quantitative variables such as mean ± standard deviation (SD) and categorical variables are presented as percentages. The Student t-test and Mann-Whitney U test were used to identify differences between the two studied groups. The Chi-square test facilitated the comparison of categorical variables. The correlation between Survivin-1, Tag-72, and HERC5 overexpression and different HCC variables was performed using the Spearman rank correlation test. A multivariate logistic regression analysis detected the independent variables of HCC. The discriminative power of Survivin-1, Tag-72, and HERC5 overexpression was assessed using ROC curves. The predictive performance of biomarker overexpression was classically evaluated by the area under the ROC curve (AUC), sensitivity (Se), specificity (Sp), positive, and negative predictive values (PPV, NPV). SPSS 16.0 software (SPSS, Inc.) was used for statistical analysis. P-values <0.05 were considered statistically significant.

Results

Demographic, clinical and laboratory features of the HCC patients

According to demographics, we noted an evident predominance of male gender in the study group compared to the control group (17 vs. 25 male patients, P=0.023). The age of the HCC patients was significantly older than that of the control group (42.09±9.4 vs. 59.2±5.9 years, P=0.002). Laboratory results were various and without any statistical significance related to cytolysis (30±11.67 vs. 37.3±14.11 UI/ml, P=0.067) but significantly different related to parameters of liver insufficiency such as albumin (4.3±1.2 vs. 2.3±1.4 g/dl, P<0.001), bilirubin (1.9±0.5 vs. 4.3±2.7 mg/dl, P=0.002), and INR (International Normalized Ratio) (1.4±0.8 vs. 1.9±1.0, P=0.026). AFP had values slightly above the normal upper limit in both HCC and controls, but levels >180 ng/ml were more frequently encountered in HCC patients compared to the controls (204.11±17.77 vs. 308.56±44.01 ng/ml, P=0.015). Jaundice was more regularly present in the HCC patients compared with the control (11 vs. 20 patients, P=0.044) (Table I).

Comparison of IHC overexpression of Survivin-1, Tag-72, and HERC5 in liver tissue samples between HCC and controls

Statistical analysis using a multivariate linear regression tool was conducted to correlate the IHC overexpression of Survivin-1, Tag-72, and HERC5 and independent variables as age, sex, laboratory results, imaging and clinical features. The multivariate linear regression analysis revealed that Survivin-1, Tag-72, and HERC5 were significantly overexpressed upon IHC analysis in HCC samples in patients older than 50 years (P=0.003, P=0.006, P=0.004, respectively), male gender (P=0.031, P=0.004, P=0.020, respectively) patients with increased AFP over 180 ng/dl (P=0.012, P=0.004, P=0.029, respectively) and with low serum albumin <3 mg/dl (P=0.031, P=0.021, P=0.003, respectively) as well as in patients with imaging features of portal thrombosis (P=0.004, P=0.020, P=0.004, respectively) and ascites (P=0.002, P=0.004, P=0.019, respectively) and in BCLC B class (P=0.045, P=0.036, P=0.045, respectively) and C (P=0.033, P=0.001, P=0.027, respectively) classified patients (Table II). Overexpression of the studied biomarkers did not correlate positively with cytolysis and other cholestasis tests or with the remaining clinical or imaging features of HCC (Table II).

Table II Multivariate linear regression analysis to identify the correlation between overexpression of liver biomarkers and clinical, laboratory and imaging variables in HCC.

Table II

Multivariate linear regression analysis to identify the correlation between overexpression of liver biomarkers and clinical, laboratory and imaging variables in HCC.

	Liver tissue biomarker overexpression
Patient variables	Survivin-1 P-value (r)	Tag-72 P-value (r)	HERC5 P-value (r)
Age, mean ± SD, >50 years	0.003 (0.94)	0.006 (0.92)	0.004 (0.94)
Male gender	0.031 (0.55)	0.004 (0.96)	0.020 (0.70)
Bilirubin	0.076 (0.24)	0.064 (0.41)	0.060 (0.44)
INR	0.055 (0.54)	0.088 (0.27)	0.059 (0.52)
AFP >180 ng/ml	0.012 (0.83)	0.004 (0.96)	0.029 (0.80)
Albumin <3 mg/dl	0.031 (-0.66)	0.021 (0.72)	0.003 (0.94)
GGT	0.059 (0.72)	0.055 (0.51)	0.030 (0.48)
Portal thrombosis	0.004 (0.96)	0.020 (0.70)	0.004 (0.96)
Ascites	0.002 (0.98)	0.004 (0.96)	0.019 (0.82)
Hepatosplenomegaly	0.060 (0.74)	0.003 (0.94)	0.039 (0.59)
BCLC
A	0.076 (0.33)	0.086 (0.25)	0.055 (0.54)
B	0.045 (0.67)	0.036 (0.24)	0.045 (0.67)
C	0.033 (0.60)	0.001 (0.99)	0.027 (0.78)

[i] HCC, hepatocellular carcinoma; Tag-72, tumor-associated glyocoprotein; HERC5, HECT and RLD domain containing E3 ubiquitin protein ligase 5; GGT, γ-glutamyl transferase; AFP, α-fetoprotein; BCLC, Barcelona Clinic Liver Cancer; INR, International Normalized Ratio.

Diagnostic accuracy of Survivin-1, Tag-72 and HERC5 for HCC

Survivin-1 tissue biomarker had an AUC of 0.96 [95% confidence interval (CI), 0.90-1.00] for the diagnosis of HCC, with 90.0% sensitivity, 100% specificity, 100% PPV, and 94.2% NPV for an optimal cut-off value of 0.3 (Fig. 1A and Table III). Tag-72 tissue biomarker had an AUC of 0.88 (95% CI, 0.78-0.94), with 72.8% sensitivity, 95.6% specificity, 82.2% PPV, and 84% NPV for an optimal cut-off value of 0.3 (Fig. 1B and Table III). HERC5 had an AUC of 0.78 (95% CI, 0.66-0.84), with 65% sensitivity, 86% specificity, 77.4% PPV, and 72.3% NPV (Fig. 1C and Table III) for an optimal cut-off value of 0.3. AFP, still considered the gold standard biomarker used in clinical settings and recommended by the international guidelines for HCC management, had an AUC of 0.34 (95% CI, 0.28-48) for the diagnosis of HCC, with 38.0% sensitivity, 66% specificity, 64% PPV, and 68.8% NPV for an optimal cut-off value of 180 ng/dl (Fig. 2). The diagnostic performance of Survivin-1, Tag-72 and HER-C5 tissue biomarkers for HCC characterization was superior to that of AFP, considered the gold standard biomarker used in clinical guidelines (Survivin-1: Z statistic=2.911, P=0.0039; Tag-72: Z statistic=2.789, P=0.0049, respectively; HERC5: Z statistic=2.844, P=0.0043) and AFP assay alone (Z statistic=5.022, P<0.0001) (Table IV).

Figure 1 Area under the curve (AUC) for the liver tissue biomarker used for HCC diagnosis: (A) Survivin-1, (B) Tag-72, and (C) HERC5. HCC, hepatocellular carcinoma; Tag-72, tumor-associated glyocoprotein 72; HERC5, HECT and RLD domain containing E3 ubiquitin protein ligase 5.

Figure 2 Area under the curve (AUC) for AFP used for HCC diagnosis. AFP, α-fetoprotein; HCC, hepatocellular carcinoma.

Table III Accuracy parameters of Survivin-1, Tag-72 and HERC5 for HCC diagnosis.

Table III

Accuracy parameters of Survivin-1, Tag-72 and HERC5 for HCC diagnosis.

	Survivin-1	Tag-72	HERC5
AUC (95% CI)	0.96 (0.90-1.00)	0.88 (0.78-0.94)	0.78 (0.66-0.84)
Accuracy (%)	97.2	80.4	75
Sn (%)	90	72.8	65
Sp (%)	100.0	96.6	86
PPV (%)	100.0	82.2	77.4
NPV (%)	94.2	84.0	72.3

[i] Tag-72, tumor-associated glyocoprotein; HERC5, HECT and RLD domain containing E3 ubiquitin protein ligase 5; AUC, area under the ROC curve; Sn, sensitivity; Sp, specificity; PPV, positive predictive value; NPN, negative predictive value.

Table IV Diagnostic performance of Survivin-1, Tag-72 and HER-C5 tissue biomarkers for HCC compared to gold standard AFP.

Table IV

Diagnostic performance of Survivin-1, Tag-72 and HER-C5 tissue biomarkers for HCC compared to gold standard AFP.

Diagnostic performance	Survivin-1 z statistic (P-value)	Tag-72 z statistic (P-value)	HERC5 z statistic (P-value)
AFP	2.911 (0.0039)	2.789 (0.0049)	2.844 (0.0043)

[i] Tag-72, tumor-associated glyocoprotein; HERC5, HECT and RLD domain containing E3 ubiquitin protein ligase 5; AFP, α-fetoprotein.

Discussion

Immunohistochemistry represents a research tool with large applicability of monoclonal and polyclonal antibodies in detecting a specific antigen with a high potential for a positive diagnosis (16). This method is largely used to diagnose malignant tumors, but it presents an area of interest for numerous applications useful not only for positive and differential diagnosis but also to provide a research field in the attempt to identify prognostic markers for cancer evolution, to confirm the positive diagnosis of tumors with uncertain histogenesis, to predict the response to treatment and to identify or confirm certain infections (7,17,18). Such biomarkers facilitate the efforts of researchers for a better understanding of HCC pathogeny, to provide more efficient therapies, and to improve disease prognosis. Epidemiology data confirm the persistent high global morbi-mortality rates for HCC, which are more consistent in endemic areas for chronic viral infections, such as Asia, the Middle East, Mediterranean countries, South America, and Africa (19-22). The current guidelines use a single biomarker for HCC follow-up, this being α-fetoprotein (AFP). The literature reports a low accuracy for AFP, and declines its role as a potent prognostic tool, thus new liver tissue biomarkers are being explored worldwide to improve HCC management (5). The difference between various etiopathogenic mechanisms involved in liver primary carcinogenesis makes it difficult to standardize a global panel of liver tissue biomarkers to simplify disease management. Using this scenario, having literature models from other studies conducted in different populations, we evaluated the expression of three liver tissue markers, aiming to confirm their overexpression in HCC tissue in patients from the Dobrogea area and to provide the background for further research for prognostic predictability or patient classification in risk groups. Our study demonstrated the same pattern of demographic, clinical, and laboratory features as the majority of published data, with a high prevalence of viral infections and alcohol abuse leading the risk factor background, and male gender and middle age, being the main characteristics of our patients. All liver tissue biomarkers explored provided a good accuracy for HCC diagnosis: 97.2% for Survivin-1, 80.4% for Tag-72, and 75% for HERC5, similar to the literature data (10,12,13). AFP over the upper limit did not prove to have diagnostic or predictive value for HCC, similar to other literature articles (23-26). Still, values over 180 ng/dl were highly predictive, and they were found to be positively correlated with all biomarkers studied, Survivin-1, Tag-72, and HERC5 (r=0.83, r=0.96, r=0.80, respectively). Our study results confirm the background hypothesis, indicating the overexpression of Survivin-1, Tag-72, and HERC5 as feasible biomarkers with which to diagnose HCC. Larger IHC studies should sustain the accuracy of the proposed tools before their introduction to international HCC management guidelines. Comparative study with other methods of diagnosis and prognostic evaluation used and other types of carcinomas, such as the study of neoangiogenesis or nuclear morphometry, can bring new useful data both in regards to diagnosis and in the prognosis of the disease (27,28). Despite the importance of our study results, our work had limits due to the increased costs of study materials, a fact that influenced the number of samples evaluated by immunohistochemistry, a problem partially solved by the funds gained through university research grant competition.

In conclusion, our study results validate the overexpression of Survivin-1, Tag-72, and HERC5 as tissue biomarkers for HCC characterization in patients from our geographical region and could be standardized in the current HCC management guideline.

Acknowledgements

The present research was performed at the Research and Development Centre for the Morphologic and Genetic Study of Malignant Pathology from the ‘Ovidius’ University of Constanta, with the precious support of Professor Dr Mariana Aschie.

Funding

Financial support was obtained after attaining the ‘Ovidius’ University contest-July 2019 for research grants [POSCCE 2.2.1; Project ID: 1844; code SMIS: 48750; CEDMOG, contract 627/11.03.2014].

Availability of data and materials

The data were obtained from the HIPOCRATE file archive of the Emergency Hospital St. ‘Apostle Andrew’ of Constanta and CEDMOG archive (http://www.spitalulconstanta.ro/). Further information about the present study is available from the corresponding author upon reasonable request.

Authors' contributions

AIS, APS, GIB, VH, CN, LCP, FV, ISM and LM conceived and designed the study; ND, LCP, CB and FB collected the data. LCP, ND, CB, and FV analyzed the data; AIS, APS, CN and ISM validated the results. FB, CN, LCP and ND were responsible for original draft preparation; AIS, GIB, LCP, ND, ISM and VH were responsible for the final manuscript editing. AIS, APS, GIB and CN supervised the manuscript publication. All authors read and approved the final manuscript.

Ethics approval and consent to participate

The Ethics Committee of Emergency Clinical Hospital St Apostle Andrei Constanta approved the study following European and local regulations. Approval no. 32/22.11.2020. Emergency Clinical Hospital St. Apostle Andrew from Constanta is a university hospital and all admitted patients sign an informed consent by which they agree that their data are available for academic and scientific purposes.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References