Prediagnostic plasma proteomics profile for hepatocellular carcinoma

Abstract

Objective: Proteomics may discover pathophysiological changes related to hepatocellular carcinoma, an aggressive and lethal type of cancer with low sensitivity for early stage diagnosis.

Design: We measured 1305 prediagnostic (median = 12.7 years) SomaScan proteins from 54 pairs of healthy individuals who subsequently developed hepatocellular carcinoma and matched non-hepatocellular carcinoma control individuals from the Nurses' Health Study (NHS) and the Health Professionals Follow-up Study (HPFS). Candidate proteins were validated in the independent, prospective UK Biobank Pharma Proteomics Project (UKB-PPP).

Results: In NHS and HPFS, we identified 56 elevated proteins in hepatocellular carcinoma with an absolute fold change of more than 1.2 and a Wald test P value less than .05 in conditional logistic regression analysis. Ingenuity pathway analysis identified enrichment of pathways associated with cell viability, adhesion, proteolysis, apoptosis, and inflammatory response. Four proteins-chitinase-3-like protein 1, growth differentiation factor 15, interleukin-1 receptor antagonist protein, and E-selectin-showed strong positive associations with hepatocellular carcinoma and were thus validated by enzyme-linked immunosorbent assay (odds ratio = 2.48-14.7, all P < .05) in the NHS and HPFS and by Olink platform (hazard ratio = 1.90-3.93, all P < .05) in the UKB-PPP. Adding these 4 proteins to a logistic regression model of traditional hepatocellular carcinoma risk factors increased the area under the curve from 0.67 to 0.87 in the NHS and HPFS. Consistently, model area under the curve was 0.88 for hepatocellular carcinoma risk prediction in the UKB-PPP.

Conclusion: However, the limited number of hepatocellular carcinoma patients in the cohorts necessitates caution in interpreting our findings, emphasizing the need for further validation in high-risk populations.

Figure 1.

Box-whisker plots for differentially expressed proteins in hepatocellular carcinoma patients and non-patient control participants in the Nurses’ Health Study and Health Professionals Follow-up Study (sample size = 54 pairs). Expression pattern of 8 representative proteins among the top dysregulated proteins (P < .05, t test; absolute fold change > 1.2), using relative fluorescence units derived from SomaScan. The boxes represent the interquartile range (IQR). The solid line in the box is the median, and the “+” sign is the mean. The bottom and top edges of the box are the lower and upper quartiles, which is known as the IQR. The whiskers range from the minimum to the lower quartile and from the upper quartile to the maximum. The extreme values (within 1.5 times the IQR from the upper or lower quartile) are the ends of the lines extending from the IQR. Points at a greater distance from the median than 1.5 times the IQR are plotted individually as circles. HCC = hepatocellular carcinoma; RFU = relative fluorescence units; CHI3L1 = chitinase-3-like protein 1; GDF15 = growth/differentiation factor 15; THBS2 = thrombospondin-2; LGALS3BP = lectin galactoside-binding soluble 3 binding protein; SELE = E-selectin; IL1RN = interleukin-1 receptor antagonist protein; ENPP7 = ectonucleotide pyrophosphatase/phosphodiesterase family member 7; PIGR = polymeric immunoglobulin receptor.

Figure 2.

Plausible pathophysiological pathways linked to hepatocellular carcinoma based on the 56 statistically dysregulated proteins. Ingenuity pathway analysis identified top biological functions (A) including cell viability and apoptosis (B), inflammatory response (C), activation of myeloid cells (D), and synthesis of reactive oxygen species (E). Protein names are abbreviated according to the standard Entrez Gene symbol. Network and cluster analysis using the STRING database of functional and physical protein associations (F). Highlighted proteins include CHI3L1, GDF15, IL1RN, and SELE. CHI3L1 = chitinase-3-like protein 1; GDF15 = growth differentiation factor 15; IL1RN = interleukin-1 receptor antagonist protein; SELE = E-selectin.

Figure 3.

Plausible upstream regulators of the 56 statistically dysregulated proteins associated with hepatocellular carcinoma. Ingenuity pathway analysis identified top upstream regulators (A) including tumor necrosis factor (B), transforming growth factor-beta1 (C), interleukin-1 (D), and myeloid differentiation primary response 88 (E). Protein names are abbreviated according to the standard Entrez Gene Symbol. Highlighted proteins include CHI3L1, GDF15, IL1RN, and SELE. CHI3L1 = chitinase-3-like protein 1; GDF15 = growth differentiation factor 15; IL1RN = interleukin-1 receptor antagonist protein; MYD88 = myeloid differentiation primary response 88; SELE = E-selectin; TGFB1 = transforming growth factor-beta1; TNF = tumor necrosis factor.

Figure 4.

Receiver operating characteristic curves comparing the logistic regression models to distinguish hepatocellular carcinoma from healthy individuals in the (A) Nurses’ Health Study and Health Professionals Follow-up Study (sample size = 108) and the (B) UK Biobank Pharma Proteomics Project (sample size = 47 911). A) Basic model included age at blood collection, sex (cohort), smoking status (never, past, current), alcohol intake (g/day, continuous), leisure time physical activity (metabolic equivalent hours per week, continuous), body mass index (kg/m², continuous), history of diabetes (yes, no), and aspirin use (yes, no). Protein model included enzyme-linked immunosorbent assay quantified chitinase-3-like protein 1 (CHI3L1), growth differentiation factor 15 (GDF15), interleukin-1 receptor antagonist protein (IL1RN), and E-selectin (SELE). B) Model included age (year), sex (male, female), race (Black or Black British, other, White), smoking status (never, past, current), alcohol intake (never, past, current), leisure time physical activity (low, medium, high), body mass index (<25, 25 to <30, ≥30 kg/m²), history of diabetes (yes, no), aspirin use (yes, no), hepatitis B virus positivity (yes, no), hepatitis C virus positivity (yes, no), Olink quantified CHI3L1, GDF15, IL1RN, and SELE. HPFS = Health Professionals Follow-up Study; NHS = Nurses’ Health Study; ROC = receiver operating characteristic; UKB = UK Biobank Pharma Proteomics Project.