Serum PCSK9 is a novel serological biomarker for the diagnosis and prognosis of pancreatic cancer

Study design and ethical considerations

The study was conducted in March 2020 and July 2023 in West China Hospital of Sichuan University (Chengdu, China). The Pathology Department of West China Hospital made the diagnosis and graded the pancreatic cancer, and the American Joint Committee on Cancer (AJCC) criteria standard defined TNM staging. The enrollment criteria of the pancreatic cancer group were: (1) pathologically confirmed pancreatic cancer; (2) no history of other primary malignancies; (3) have not received any treatment. The exclusion criteria were: (1) incomplete medical records; (2) history of other cancer-related therapies; and (3) current inflammatory diseases. The control population was comprised of benign disease patients and normal individuals. The benign disease patients were clinically diagnosed with benign disease of the pancreas, including chronic pancreatitis, benign tumor of the pancreas, etc. Normal individuals were conducted with no signs of diseases.

The Ethics Committee at Sichuan University’s West China Hospital gave its approval for the study, which was carried out in compliance with the Declaration of Helsinki (1189, 20220325). Before being recruited, all participants were apprised of the ethical information relevant to this study and written informed consent was obtained. Patients diagnosed with pancreatic cancer were followed up via telephone interviews or outpatient clinic visits, allowing for the acquisition of overall survival (OS) information from all participants. All PC patients were followed up until October 2023.

Samples collection

The study involved the enrollment of 289 blood samples: the PC group consisted of 118 patients, while the control group consisted of 171 patients, which included 97 benign disease groups (BDG, including chronic pancreatitis (CP), n = 62; pancreatic neuroendocrine tumor, n = 10; intraductal papillary mucinous neoplasm (IPMN), n = 9; cystadenoma, n = 7; solid pseudopapillary tumor of pancreas, n = 6; others, n = 3) and 74 healthy controls (HC). All blood samples were collected before surgical resection, chemotherapy, or radiotherapy. Any hemolyzed blood samples were excluded. A total of 5 ml of fasting blood obtained from each participant was collected into tubes containing heparin sodium and then stored at 4 °C for transfer to a laboratory. After centrifuging serum samples for 10 min at 3,000 × rpm at 4 °C, they were split into 2 to 4 aliquots and stored at −80°C until analysis. Additional clinical data were gathered from the West China Hospital case management system, including age, sex, TNM stage, grade, serum lipid level, and serum level of CA19-9.

Measurement of sPCSK9 levels

sPCSK9 levels were quantified using enzyme-linked immunosorbent assay (ELISA) kits (Cat#EH384RB, Thermo Fisher Scientific, Waltham, MA, USA) following the manufacturer’s instructions. Absorbance values were measured at a wavelength of 450 nm using a 96-well microplate reader. Subsequently, the sPCSK9 concentration was determined by comparing the obtained result to the standard sample and constructing a standard curve.

Immunohistochemistry

The study involved the enrollment of 48 tissue samples, which were obtained from the pathology platform of West China Hospital and have been reviewed by the ethics department. After paraffin embedding, the thickness of the tissue samples was sectioned into 3 mm sections deparaffinized in xylene and rehydrated through different concentrations of alcohols (100%, 95%, 90%, and 80%). Endogenous peroxidase activity was blocked with hydrogen peroxide (3%) at 37 °C for 15 min. The slides were subjected to antigen retrieval by pressure cooking in citrate buffer (pH 6.0) for 16 min. A polyclonal anti-PCSK9 antibody (Cat. No 27882-1-AP, Proteintech, Chicago, IL, USA) was used as the reaction at a dilution of 1:20 at 37 °C for 1h. After washing the sections in PBS, we incubated them with biotinylated secondary antibody for 40 min at 37 °C. Diaminobenzidine (DAB) solution as the chromogen and hematoxylin solution as the counterstain. For the negative control, 1% BSA/PBS was used in place of the primary antibody and was processed in the same manner.

The slides were captured digitally, stored as high-resolution images files using the Slideview VS200 slide scanner (Olympus, Tokyo, Japan) and analyzed using the OlyVIA visualizer version 2.4 (Olympus, Tokyo, Japan). The IHC score of these areas: (% area of weak staining) + (2 × % area of moderate staining) + (3 × % area of strong staining). Then, we received a score between 0 and 300.

Statistical analysis

All variables were subjected to descriptive statistics, and the Shapiro–Wilk test assessed the data’s normality. Based on the results, the statistical methods were selected as follows. The χ² test was used to classify data. For normally distributed count data, the independent samples T-test was employed. For abnormally distributed count data, the non-parametric Mann-Whitney U test was applied. The study compared the expression levels of sPCSK9 and CA19-9 in PC patients and the control group using non-parametric Mann–Whitney U tests. The study utilized the receiver operating characteristic (ROC) curve to assess the sensitivity and specificity of sPCSK9 and CA19-9. The respective areas under the curves (AUCs) with 95% confidence intervals (CI) were also evaluated to determine the cutoff line that provides the optimal diagnostic accuracy and likelihood ratios. We also compare the performance of different ROC curves by using the DeLong test. The study assessed the correlation between serum levels of sPCSK9 and CA19-9 using Spearman rank correlation, and the correlation between serum levels of sPCSK9 and HDL using Pearson rank correlation. Survival was contrasted using log-rank tests by creating survival curves through the Kaplan-Meier method. We used the SPSS version 23.0 (IBM Corp, Chicago, IL, USA) to analyze all statistics. A statistically significant difference was defined as a P-value < 0.05. Finally, we used the GraphPad Prism V.9.5 software (GraphPad Software, La Jolla, CA, USA) to draw figures.

Participants characteristics

The baseline clinical characteristics of the patients and benign disease groups are presented in Table 1. Between the two groups, demographic traits were well-balanced. Among the 118 PC patients, stratification based on AJCC stages resulted in four groups: stage I (n = 33), stage II (n = 22), stage III (n = 33), and stage IV (n = 30).

Serum level of PCSK9 in PC patients and control group

Figure 1 depicts the sPCSK9 in both the PC and control groups. The PCSK9 serum concentration was considerably lower in the PC group compared to the healthy control (HC) group (Z = −2.546, P < 0.05). Conversely, the sPCSK9 levels in the PC group were significantly higher than in the benign disease group (BDG) (Z = −5.457, P < 0.001). Furthermore, sPCSK9 in the BDG group was significantly lower compared to the HC group (Z = −6.853, P < 0.001) (Fig. 1A). Patients with early PC had considerably lower sPCSK9 levels than those with advanced PC (Z = −2.751, P < 0.01) (Fig. 1B). In addition, the sPCSK9 levels of patients diagnosed with early PC were significantly lower than those of the HC group (Z = −3.674, P < 0.001), and sPCSK9 levels of patients with advanced PC were much higher than those in the BDG group (Z = −5.845, P < 0.001) (Fig. 1B).

Correlations between clinicopathological characteristics and sPCSK9 levels in PC patients

We performed the nonparametric Mann–Whitney U test to compare sPCSK9 levels among clinical subgroups. Our analysis indicated that the group without lymph node invasion had significantly lower sPCSK9 levels compared to the positive groups (Z = −2.799, P < 0.01). However, the sPCSK9 levels had no significant differences among the remaining clinical factors, such as age, sex, diabetes, sex, hypertension, tumor invasion, or metastasis (Fig. S1).

Further, we separated PC patients into two groups based on the sPCSK9 cut-off value (54.83 ng/ml), with high and low sPCSK9 levels, and then used the χ² test to interpret the correlations between sPCSK9 levels and Clinicopathological features (Table 2). We found that the level of sPCSK9 had a strong correlation with drinking (P = 0.008), lymph node invasion (P = 0.009), and TNM stage (P = 0.022). However, the sPCSK9 level did not significantly correlate with other clinical features, such as age, sex, smoking, diabetes, hypertension, tumor invasion, or metastasis (P > 0.05).

Diagnostic value of sPCSK9 and CA19-9

We assessed the diagnostic value of sPCSK9 in detecting PC (Table S1), particularly early-stage PC (stage I-II) by receiver operating characteristic (ROC) curve analysis. ROC analysis determined the cutoff values, sensitivity, and specificity of CA19-9 and sPCSK9 levels. When comparing patients with early-stage PC to the HC group, the best cut-off levels for CA19-9 and sPCSK9 were 18.6 U/mL and 54.4 ng/mL, respectively. The area under the curve (AUC) values for CA19-9 versus sPCSK9 were 0.725 (95% CI [0.604–0.846], sensitivity 71.7% and specificity 82.4%) and 0.689 (95% CI [0.598–0.781], sensitivity 65.5% and specificity 67.6%), as illustrated in Fig. 2A. When comparing patients with early-stage PC to the BDG, the best cut-off levels for CA19-9 and sPCSK9 were 42.70 U/mL and 27.4 ng/mL, respectively. The AUC values for CA19-9 versus sPCSK9 were 0.706 (95% CI [0.595–0.818], sensitivity 52.2% and specificity 92.6%) and 0.651 (95% CI [0.563–0.738], sensitivity 92.7% and specificity 35.1%), as illustrated in Fig. 2B. Then, we evaluated the diagnostic performance of the combination of CA19-9 and sPCSK9 in early-stage PC patients and controls. When comparing the AUC of the combination between early-stage PC and HC, the AUC value for the combined group was statistically significantly higher than the CA19-9 group (0.875, 95% CI [0.802–0.949], sensitivity 78.3% and specificity 84.3%; Fig. 2A) (Z = 3.176, P < 0.05). When comparing the AUC of the combination between early-stage PC and BDG, the AUC value for the combined group also was higher than that of the CA19-9 group (0.762, 95% CI [0.668–0.855], sensitivity 65.2% and specificity 81.9%; Fig. 2B). But it’s not statistically significant (Z = 1.697, P = 0.11). The sensitivity and specificity of the sPCSK9 and CA19-9 in detecting early-stage PC are shown in Table 3.

In addition, we explored the relationship between sPCSK9 and CA19-9. The test kit has limitations, and any result higher than or equal to 1,000 U/ml will be disregarded for CA19-9 measurement. After that, we used Spearman’s correlation analysis to determine the relationships between sPCSK9 and CA19-9 levels in each patient. The results showed a negative level trend (r = 0.257, P < 0.05, Fig. 3).

Prognostic value of sPCSK9 levels before treatment for PC patients

To evaluate the prognostic significance of sPCSK9 in PC patients, we identified a sPCSK9 cutoff of 54.83 ng/ml, which approximates the optimal threshold observed when comparing PC patients with HC. Utilizing this cutoff, we stratified PC patients into two cohorts: high sPCSK9 and low sPCSK9. Interestingly, the low sPCSK9 group (n = 41) exhibited a significantly prolonged overall survival compared to the high sPCSK9 group (n = 15) (27 months (95% CI [17.59–36.41]) vs. 11 months (95% CI [7.21–14.79]), P = 0.022), as depicted in Fig. 4.

Immunohistochemistry staining of PCSK9 in human pancreatic cancer tissues

To assess PCSK9 protein expression in pancreatic cancer tissues, immunohistochemistry (IHC) was carried out in pancreatic tissues containing normal pancreatic tissues and tumor tissues. Statistical analysis indicated a significant increase of PCSK9 expression in tumor tissues (t = 7.693, P < 0.01) compared to normal tissues (Fig. 5).