Clinical value of LHPP‐associated microRNAs combined with protein induced by vitamin K deficiency or antagonist‐II in the diagnosis of alpha‐fetoprotein‐negative hepatocellular carcinoma

Abstract Background Alpha‐fetoprotein (AFP) has received extensive attention in the differential diagnosis of hepatocellular carcinoma (HCC), especially for AFP‐negative HCC (AFP‐NHCC). The current study aimed to explore the value of targeted regulation of LHPP expression‐related microRNAs (miRs) and protein induced by vitamin K deficiency or antagonist‐II (PIVKA‐II) in the differential diagnosis of AFP‐NHCC. Methods A retrospective study was conducted on a testing set—including 214 AFP‐NHCC patients, 200 cirrhosis, and 210 controls, and a validation set—including 140 AFP‐NHCC patients, 134 cirrhosis, and 128 controls recruited from The First Affiliated Hospital of Hunan Normal University. Serum miRs were examined using quantitative real‐time PCR method. Serum PIVKA‐II was measured by enzyme‐linked immunosorbent assay. Results Compared with adjacent tissues, LHPP protein levels in cancer tissues were significantly decreased (P < .05). Predictive software and dual‐luciferase reporter assays showed that miR‐363‐5p and miR‐765 can target LHPP expression. Serum miR‐363‐5p, miR‐765, and PIVKA‐II levels were significantly higher in AFP‐HCC patients than in cirrhosis and controls. A logistic regression model combining miR‐363‐5p, miR‐765, and PIVKA‐II was performed. This model presented a high discriminating value (AUC: 0.930, sensitivity/specificity: 79.4%/95.4%) than any single indicator. In the validation set, this model still showed a high discriminating value (AUC: 0.936, sensitivity/specificity: 83.6%/94.7%). Conclusion Current model combining serum miR‐363‐5p, miR‐765, and PIVKA‐II has potential significance for diagnosis of AFP‐NHCC.


| INTRODUC TI ON
For decades, hepatocellular carcinoma (HCC) screening relied primarily on ultrasound imaging and alpha-fetoprotein (AFP).
Due to technical limitations, ultrasound images are often unrecognizable for HCC nodules, especially less than 1 cm. 1,2 Unexpectedly, AFP is measured separately in early HCC with a missed diagnosis rate of 40%. 3 AFP-negative hepatocellular carcinoma (AFP-NHCC) is an important type of HCC that currently causes many patients to lose early diagnosis and treatment, especially in patients with tumors less than 3 cm. 4 The clinical symptoms of AFP-NHCC patients are usually mild and lack specificity, and their clinical diagnosis relies mainly on other tumor markers or imaging. Protein induced by vitamin K deficiency or antagonist-II (PIVKA-II) is believed to be a suitable biomarker specific for HCC. 5 However, the sensitivity of PIVKA-II is still not satisfactory. 6 In addition, due to the small size of the AFP-NHCC tumor, imaging examination is prone to miss. It is reported that the diagnostic rates of AFP-NHCC patients by CT, MRI, and B ultrasound are about 50.9%, 50.0%, and 10.4%, respectively. 7 In addition, liver nodular lesions such as cirrhosis regenerative nodules, hepatic focal nodular hyperplasia, hepatic adenomas may also have HCC-like imaging findings, making AFP-NHCC easily misdiagnosed as benign disease, and thus lost the opportunity for early treatment. 2,8 In 2018, Hindupur et al 9 discovered a new HCC suppressor protein-LHPP, in the mouse HCC model, and they also revealed its potential anticancer mechanism. They found that (a) with the development of HCC tumors, the expression of LHPP protein in mouse HCC cancer tissues gradually decreased, while the level of LHPP in the adjacent tissues was normal; (b) The Cancer Genome Atlas (TCGA) data showed that the severity of HCC and the life expectancy of patients were significantly correlated with the level of LHPP in the tissues, and the patients with low LHPP expression had a lower median survival time than those with high LHPP expression for nearly 2 years; (c) when the expression of LHPP in cells is downregulated, the level of proteomic phosphorylation in the cells is significantly increased, thereby causing uncontrolled cancer cell proliferation. The above study suggests that the decreased expression of LHPP in tissues is an important factor in promoting the formation of HCC. However, since the above experiments are based on the protein quantification of LHPP in cancer tissues, it is not conducive to the spread of cancer screening. Therefore, searching for non-invasive markers involved in the regulation of LHPP expression is the focus of our study.
MicroRNA (miR) plays an important role in many biological processes. [10][11][12][13] In current study, firstly, we identified the miRs that are targeted for downregulating LHPP expression through bioinformatics software and luciferase reporter gene assay. Then, we evaluated the significance of miRs and PIVKA-Ⅱ in distinguishing AFP-NHCC.
In addition, logistic regression model was built for AFP-NHCC prediction.

| Ethics approval and consent to participate
This study was approved by the Ethics Committee of The First Affiliated Hospital of Hunan Normal University (L20180104).
Written informed consent was provided in accordance with the Declaration of Helsinki.

| Testing set
We recruited 214 patients AFP-NHCC between April 2016 and January 2018 at The First Affiliated Hospital of Hunan Normal University, Changsha, China. AFP-NHCC was confirmed by liver puncture or histopathology examination. Two hundred patients with cirrhosis and two hundred and ten controls were also recruited.

| Logistic regression model establishment
A regression formula for AFP-NHCC prediction was established.
The formula is as follows: Logit (P) = X 0 + X 1 Y 1 + X 2 Y 2 + X 3 Y 3 +… +X n Y n = ln[p/(1-p)], "p" means the incident probability (AFP-NHCC), "n" means the number of interference factor, "X" means the influence coefficient of each interference factor, and "Y" means the value of each interference factor.

| Validation set
One validation set from The First Affiliated Hospital of Hunan Normal University (Changsha, China) was used to assess the above logistic regression model including a total of 140 AFP-NHCC, 134 cirrhosis, and 128 controls between February 2018 and April 2019.

| Serum and tissue specimens
Peripheral blood was collected from AFP-NHCC and cirrhosis before receiving treatment and healthy controls at the time of admission to the Medical Examination Center. In addition, eight pairs of cancer and adjacent tissues (>3 cm from the edge of cancer tissue) from AFP-NHCC patients who underwent surgical treatment were enrolled. Small interference RNAs (siRNAs) targeting LHPP (si-LHPP) was obtained from GenePharma Co. ltd (Lot number: W-19-09602).

| Cell culture and cell transfection
The transfection group was divided into two groups, including control group and LHPP inhibitor (si-LHPP) group. Cell lines were seeded in a six-well plate, and when the cell confluence reached about 50%, the transfected cell line was immediately mediated with Lipofectamine 2000 (Invitrogen, Thermo, New York, USA, Lot number: 11668-027), and the medium was changed 6 h after transfection. After si-LHPP treatment for 24 hours, the cell lysates were immunoblotted with antibodies against LHPP.

| Quantitative Real-Time PCR (qRT-PCR)
Cell and serum total RNAs were extracted using Trizol (Invitrogen,
The kit was provided by Wuhan Boster Bioengineering Co., Ltd (Lot number: 233887), and the detection process was carried out in strict accordance with the operation instructions. The procedure was as follows: All serum samples were sequentially added to the microplates and incubated with the antibody for 40 minutes at room temperature (adding PIVKA-II standards at a concentration of 100 mAU/ mL, 50 mAU/mL, 10 mAU/mL, and positive and negative controls), washing the plate five times. The enzyme-labeled monoclonal antibody was incubated at room temperature for 40 minutes in the dark, washed for five times, and the substrate was reacted for 15 minutes to terminate the reaction. The absorbance (A) was measured using an American Thermo Fisher Scientific Nanodrop ND2000. The kit performance includes the following: (a) accuracy-the linear regression of the standard and the expected concentration correlation coefficient R value is greater than or equal to 0.9900; (b) sensitivity-the lowest detection concentration is less than 1.0 mAU/mL; (c) specificity-does not cross-react with other soluble structural analogs; and (d) repeatability-the coefficient of variation between the plate was <15%. In this study, we calculated the intra-and interassay coefficients of variation for serum PIVKA-II assays to assess the repeatability and precision of the experiments.

| Western blot analysis
Total protein was extracted using RIPA buffer (Beyotime, Lot number: P0013B) containing protease inhibitors, and protein levels were detected using the BCA reagent (Beyotime, Lot number:

| Immunohistochemical staining
All tissue samples were fixed with 4% formaldehyde, and then, the samples were dehydrated and sectioned. The sections were blocked at room temperature for 1 hour (5% serum), and the monoclonal an-   SPSS 19.0 was used. Differences among normally distributed data were evaluated by t test or ANOVA; otherwise, Mann-Whitney U test and

| Statistical analysis
Kruskal-Wallis H test were used. P < .05 was considered significant.

| Expression of LHPP protein in AFP-NHCC patients and prediction and validation of miRs regulating LHPP expression
The results of immunohistochemical staining are shown in Figure 1A.  Figure   S2. Our results showed that the anti-LHPP antibody does not have cross-reaction with other proteins.

| The relationship between serum miR-363-5p, miR-765, and PIVKA-II levels and clinical features of AFP-NHCC and their differential diagnosis value for AFP-NHCC
The main baseline characteristics of the studied subjects are illustrated in Table 1. No significant difference was observed (P > .05).
The melting peaks of miR-363-5p and miR-765 were single, indicating that the primers did not form primer dimers and there was no non-specific amplification (Figure 2A,B).
To estimate the diagnostic value of miR-363-5p, miR-765 and PIVKA-II in AFP-NHCC, ROC was constructed using the following model: AFP-NHCC vs. non-AFP-NHCC (controls + cirrhosis), Figure 2F-H and Table 2. We found that the combination of the three indicators possessed a higher specificity (95.4%) for differentiating AFP-NHCC from non-AFP-NHCC.

| Validation of the logistic regression model
The validity of the logistic regression model was assessed in one external validation set from our hospital. A total of 140 AFP-NHCC, 134 cirrhosis, and 128 controls were recruited. The main baseline characteristics of the studied subjects are illustrated in Table 4

| D ISCUSS I ON
Although there are great developments in the current treatment of HCC, including surgical resection, liver transplantation, adjuvant therapy, and interventional therapy, many HCC patients are diagnosed after the occurrence of relevant clinical symptoms. [14][15][16] Therefore, identification of an effective diagnostic model for HCC is of great importance for patients, particularly for AFP-NHCC patients. Recently, LHPP is reported to play an essential role in inhibiting human HCC progression by regulating phosphatidylinositol-3-kinase/protein kinase B (PI3K/AKT) signaling pathway, and the loss of LHPP expression is also associated with reduced survival in HCC. 9 Zheng et al 17     serum levels of miR-363-5p and miR-765 in patients with AFP-NHCC were significantly higher than those in cirrhosis and controls and were related to differentiation, tumor size, and TNM stage, confirming that they are oncogenes in HCC. In recent years, PIVKA-II is considered to be a novel serological marker for HCC. 6,23 Our results indicated that PIVKA-II was significantly higher in AFP-NHCC patients and were useful for distinguishing AFP-NHCC from cirrhosis and controls, and the sensitivity was 65.4%, the specificity was 84.6%, which was consistent with previous research results. 23  potential significance for the non-invasive differential diagnosis for AFP-NHCC.

CO N FLI C T O F I NTE R E S T
The authors declare that there is no conflict of interest.

AUTH O R CO NTR I B UTI O N S
NN researched literature and conceived the study. ZT and HW were involved in protocol development, gaining ethical approval, patient recruitment, and data analysis. ZT and TY wrote the first draft of the article. All authors reviewed and edited the article and approved the final version of the article.

E TH I C A L A PPROVA L
The ethics committee of The First Affiliated Hospital of Hunan Normal University approved this study (L20180104).