Cancer-testis antigen lactate dehydrogenase C4 in hepatocellular carcinoma: a promising biomarker for early diagnosis, efficacy evaluation and prognosis prediction

Identification of isolated vesicles

The extraction process was summarized in Figure 1. TEM showed that the size of exosomes derived from serum ranged from 30 to 150 nm in diameter and featured small vesicles with membrane structures, which were consistent with the morphological characteristics of exosomes (Figure 2A). To confirm whether exosomes were successfully isolated and purified, expressions of exosome marker proteins CD9 and CD63 were determined by western blot. The results showed that CD9 and CD63 could be detected in the eluate containing serum exosomes of both HCC patients and healthy controls (Figure 2B). This suggested that exosomes extraction was successful.

Serum and exosomal LDHC acted as an early diagnostic biomarker for HCC

The qRT-PCR analysis showed that positive rates of LDHC mRNA expressions in serum and exosomes were 68% (34 / 50) and 60% (30 / 50) in early-stage HCC patients, and 18% (18 / 100) and 17% (17 / 100) in healthy controls, respectively. The results also showed that the average expression levels of serum and exosomal LDHC mRNAs in the preliminary diagnosis group were 4.64- and 14.28-fold higher than those in the healthy control group, respectively (Figure 3A, 3B, 3D, 3E). ROC curve analysis showed that the sensitivity, specificity and AUC of serum LDHC mRNA in distinguishing early-stage HCC patients from healthy controls were 83.3%, 76.5%, and 0.8382 (youden index = 0.598), respectively (Figure 3C); and the AUC of exosomal LDHC in distinguishing early-stage HCC patients from healthy controls reached 0.9451, under which the sensitivity and specificity were estimated to be 88.2% and 93.3%, respectively (youden index = 0.816) (Figure 3F). The above results suggested that serum and exosomal LDHC could be considered as a promising biomarker for early diagnosis of HCC.

The correlations between the expression of LDHC in serum and exosomes and the clinicopathological characteristics of HCC patients

The correlation analysis showed that serum LDHC levels were positively correlated with the levels of AFP (r² = 0.8295, P < 0.0001) and PIVKA-II (r² = 0.3639,P = 0.0011) (Figure 4A, 4B). Exosomal LDHC levels were correlated with the levels of AFP (r² = 0.7455, P < 0.0001) and PIVKA-II (r² = 0.2665, P < 0.0001) in HCC cases (Figure 4F, 4G). However, there were nonsignificant correlations between serum and exosomal LDHC levels and clinicopathological indicators such as gender, age, tumor size (all P > 0.05, data not shown), as well as serum levels of carcino-embryonic antigen (CEA), Golgi protein 73 (GP-73), and glycoprotein antigen 199 (CA19-9) (Figure 4C–4E, 4H–4J).

Values of serum and exosomal LDHC in efficacy evaluation and recurrence monitoring

To explore the effects of LDHC mRNA on efficacy evaluation and recurrence monitoring, pre- and post-treatment LDHC mRNA expressions in the serum and exosomes of HCC patients were quantitated. It was found that the expression level of serum LDHC in the treatment group was significantly lower than that in the preliminary diagnosis group (AUC = 0.7036), while the average expression level of serum LDHC in the recurrence group was significantly higher than that in the treatment group (AUC = 0.9839) (Figure 5A–5D). Similarly, serum exosomal LDHC levels in the preliminary diagnosis group were significantly higher than the levels in the treatment group (AUC = 0.7113), and serum exosomal LDHC levels in the recurrence group were significantly higher than those in the treatment group (AUC = 0.9538) (Figure 5E–5H). These results suggested serum and exosomal LDHC could be used as an effective indicator for the evaluation of the efficacy of HCC-related treatments and recurrence prediction of patients.

LDH-C4 as a prognostic marker of HCC

It was found that LDH-C4 was mainly localized in the cytoplasm of HCC cells, which was significantly up-regulated in HCC tissues (Figure 6A). The proportion of LDH-C4 high expression (+ + / + + +) in HCC tissues accounted for 55.84% (43 / 77), which was significantly higher than that in para-carcinoma tissues (26.32%, 20/76) (P < 0.0001). The correlation analysis showed that LDH-C4 levels were positively correlated with clinical stage (P = 0.002) and mass size (P = 0.023) of HCC (Table 1), but not related to KI-67, p53 and CD34 levels (Figure 6B). Moreover, LDH-C4 staining score was higher in stage 0-II HCC patients than that in stage III-IV HCC patients (Figure 6C). The Kaplan-Meier analysis showed that the overall survival of patients in the LDH-C4 low expression group was significantly prolonged compared with the LDH-C4 high expression group, while the prognosis of patients in the LDH-C4 high expression group was worse than that in the low expression group (P < 0.001) (Figure 6D). The COX regression model analysis showed that LDH-C4 was an independent risk factor for HCC prognosis (P = 0.016) (Table 2).

Expressions of LDHC in HCC tissues analyzed based on GEPIA data

We also assessed the clinical significance of LDHC expressions in HCC based on GEPIA data that from TCGA. We found that LDHC levels were elevated in HCC tissues compared with non-cancer controls (Figure 7A), and also highly expressed in stage I-IV HCC patients (Figure 7B). The correlation analysis showed that LDHC expressions were positively associated with AFP levels (P = 0.048, Figure 7C), but not correlated with EGFR expressions in HCC patients (Figure 7D). The survival analysis showed that LDHC in HCC tissues was a nonsignificant predictor for OS and DFS of HCC patients (Figure 7E, 7F).

Clinical data

A total of 112 serum samples from HCC patients—including 50 cases of preliminarily diagnosed HCC (stages I-II), 32 treated cases and 30 recurrent ones—who were admitted to Fujian Cancer Hospital from December 2018 to December 2019 were collected. Another 100 serum samples from healthy volunteers were enrolled in the control group. All serum samples were collected with the approval of the ethics committee at our hospital (ethical grant No. SQ2015-049-01). Commercial high-throughput HCC microarray (HLivH180Su10, contained 93 HCC cases and 87 adjacent non-cancerous tissues) was provided by Shanghai Outdo Biotech Co., Ltd. (Shanghai, China), and only 77 samples from HCC cancer tissues and 76 from adjacent non-cancerous tissues were used. Among the 77 cancer patients consisting of 70 males and 7 females aged from 21 to 75 years with an average age of 45.32 ± 5.18 years, 33 cases were diagnosed as stage T1 + T2 (or stage I + II) HCC and 37 as stage T3 + T4 (or stage III + IV) HCC (unavailable data from 7 cases were not included). The tumor stage classification was guided by the TNM stage criteria of primary hepatocellular carcinoma (7^th edition) in the American Joint Committee on Cancer (AJCC) / the Union for International Cancer Control (UICC) staging in 2010. All cancer tissues were confirmed by pathological examination. The enrolled patients had complete clinical data and follow-up information. The operation time ranged from January 2007 to November 2009, and the follow-up period varied between 4 and 7 years from September 2013.

Vesicle isolation and identification by western blot

Serum exosomes were extracted using an exoRNeasy Serum/Plasma Midi Kit (QIAGEN, Catalog No.77044) (Part I: vesicle isolation). The extracted exosomes were identified by transmission electron microscope (TEM). The detailed processes were described in our previous study [26]. Briefly, after total proteins extracted from the exosome eluate were lysed using RIPA lysis buffer, they were separated by gel electrophoresis (SDS-PAGE, 12% gels) at 80 V constant pressure and transferred to PVDF membranes at 250 mA constant current. TBST containing 5% skimmed milk powder was used for blocking the PVDF membranes overnight. Rabbit anti human CD9 monoclonal primary antibody (Abcam, Catalog No.ab92726; 1 : 300), rabbit anti human CD63 monoclonal primary antibody (Abcam, Catalog No.ab217345; 1 : 100), and mouse anti human GAPDH monoclonal primary antibody (Beyotime, Catalog No. AF0006; 1 : 1000) were added for incubation for 4 h at room temperature. Goat anti mouse / rabbit IgG (1 : 10000) with horseradish peroxidase (HRP) was added before a 2-h incubation at room temperature. ECL luminescent solution was employed for visualizing the protein bands which were subsequently photographed. Each experiment was repeated three times.

Real-time qRT-PCR

Total RNAs were extracted from serum and serum-derived exosomes respectively using the MiRNeasy Kit (QIAGEN, Catalog No.217184) and exoRNeasy Serum/Plasma Midi Kit (Part II: RNA isolation, Figure 1), according to the manufacturer’s instruction. The serum and exosomal RNAs were reversely transcripted into cDNAs using the Transcriptor first strand cDNA synthesis Kit (Roche) according to the manufacturer’s instruction. The primer sequences of LDHC and GAPDH were previously described [26]: LDHC-F: 5'-TCATTCCTGCCATAGTCCA-3', LDHC-R: 5'-CAATTACACGAGTTACAGGTA-3'; GAPDH-F: 5'-TCGACAGTCAGCCGCATCTTCTTT-3'; and GAPDH-R: 5'-ACCAAATCCGTTGACTCCGACCTT-3'. Levels of LDHC mRNAs were determined using the ABI7500 fluorescence quantitative PCR detector and SYBR green real time PCR master mix (ROX) kit under the following conditions: denaturation at 95 °C for 10 min, then 95 °C for 15 s, 40 cycles, and finally 60 °C for 1 min. Subsequently, relative expressions of LDHC mRNAs were calculated using the 2^−ΔΔCt method normalized to GAPDH.

Immunohistochemical analysis

The immunohistochemical analysis was performed using an EliVision^TM Plus Two-Step Test Kit. The process was described in our previous study [37]. Briefly, wax-embedded tissues were dewaxed in xylene and rehydrated in a series of ethanol solutions. They were washed and treated with H₂O₂ for 10 min. T The antigen was extracted by heating citric acid in a boiling-water bath at 100-120 °C and high pressure for 3 min. Sections were incubated in the diluted (1 : 50) rabbit anti human LDHC monoclonal antibody (Abcam, Cat.no.ab52747) at 37 °C for 90 min, washed with PBS, and incubated with goat anti-rabbit antibody for 30 min. The sections were stained with DAB for 2 min, and hematoxylin was used for contrast staining.

The proportion of LDH-C4-positive cells in each sample was scored as follows. A positive rate of less than 5% was scored as 0, the rate of more than 5% and less than 25% was scored as 1, the rate of more than 25% and less than 50% as 2, and that of more than 50% as 3. The staining severity was graded according to the final score for LDH-C4 expressions based on the positive rates: 0 point was rated as “−”, 1-2 points as “+”, 3-5 points as “+ +”, and 6-9 points as “+ + +”. Among them, “− / +” was defined as negative or low expression and “+ + ~ + + +” as high expression. All sections were judged by two experienced pathologists using a double-blind method.

Bioinformatics analysis of available databases

LDHC mRNA expressions in HCC were also analyzed using available TCGA data in GEPIA (http://gepia.cancer-pku.cn/). A Log₂FC cutoff value (Log₂FC = 2, P = 0.01) was used to distinguish a high LDHC expression group from a low expression group. For the survival analysis, high and low LDHC expression groups were identified using a group cutoff of the median.

Statistical analyses

Data were analyzed using SPSS 16.0 software. Measurement data were presented as mean ± standard deviation (SD). We studied the fulfillment of the normality and homogeneity of variance assumption prior to the application of Student's t-test for comparisons between groups. Kaplan-Meier survival curve was used to analyze the overall survival of HCC patients. The correlations between LDH-C4 expressions and clinicopathological features and between the expression and the prognosis of HCC patients were analyzed using chi-square test and Spearman’s rho test. A Cox proportional hazard regression model was established to analyze independent prognostic risk factors of HCC. A P value of less than 0.05 was considered statistically significant.