Phospholipase A2 Group IIA Is Associated with Inflammatory Hepatocellular Adenoma

Abstract

Although benign hepatocellular adenomas (HCA) are very rare, recent observations have shown their occurrence in patients with diabetes mellitus. Consequently, most of these cases are treated by resection due to concerns regarding their potential progression to hepatocarcinoma (HCC). This decision is largely driven by the limited number of studies on HCC subtyping and the lack of molecular and biological insights into the carcinogenic potential of benign tumors. This study aimed to comprehensively investigate the subtype classification of HCA and to compare and analyze gene expression profiling between HCA and HCC tissues. One fresh inflammatory HCA (I-HCA), three non-B non-C HCCs, two hepatitis B virus-HCCs, and one normal liver tissue sample were subjected to single-cell RNA sequencing (scRNA-seq). Comparative analysis of scRNA-seq among different tissues showed that phospholipase A2 group IIA (PLA2G2A) mRNA was specifically expressed in I-HCA, following RNA-seq analysis in formalin-fixed paraffin-embedded tissues from other HCAs. Immunohistochemistry using the PLA2G2A antibody in these tissues indicated that the positive reaction was mainly observed in hepatocytes of I-HCAs and stromal cells surrounding the tumor tissue in HCC were also stained. According to a clinical database, PLA2G2A expression in HCC does not correlate with poor prognosis. This finding may potentially help develop a new definition for I-HCA, resulting in a significant clinical contribution, but it requires validation with other fresh HCA samples.

Keywords: Pla2g2a; hepatocellular adenoma; hepatocellular carcinoma; single-cell RNA sequencing.

Figure 1

scRNAseq of HCA sample. (a) The tSNE plot of HCA cell clusters based on 1591 cells. It was classified into eight cell clusters; four macrophages (cluster #1, #2, #3, and #7 MAC1MAC4), two HCA (cluster #5 HCA1, #8 HCA2), one T-cell (cluster #4), and one endothelial (Endo, cluster #6). (b) Violin plots for representative 10 cell marker genes (gene symbol). (c) Feature plot and violin plot for conventional specific inflammatory HCA (IHCA) markers; SAA1, SAA2, and CRP. “5” or “8” means the cluster # shown in (a). (d) The bar graphs of the distribution for the top 10 specific genes determined to cluster #5 and cluster #8 through all clusters. The gene symbol corresponding to each color-coded bar graph is presented in the upper right corner of each graph. The x-axis shows averaged intensity × positive cell (a.u.) and the y-axis shows the cluster number. (e) Differential expressed genes (DEGs) between cluster #5 and cluster #8. The color in heatmap from yellow to violet reveals the gradual expression intensity differences from high to low. The letters to be listed on the right side of the heatmap indicate the representative gene groups or gene symbols. For example, there are eight “Enzymes” (green lines), such as CYP2C9, LDHA, and CYB5R3; two “Leukocyte antigen” (blue lines); five “Complement” (orange lines); and four “ITIH gene family” (sky blue lines). (f) Reclustering of two hepatocellular cell clusters #5 and #8. The heatmap shows the differential expressed genes among three subclusters. Additionally, there are three “Apolipoprotein” (pink lines) groups.

Figure 2

scRNAseq of one HCA and three NBNC-HCC samples. (a) The tSNE plot of the total for 7612 cells (NBNCHCC1; 1897 cells, NBNCHCC2; 2001 cells, NBNCHCC3; 2123 cells, and HCA; 1591 cells). The merged cells were divided into 19 clusters. (b) Cell types and sample-identified tSNE plots. The majority of the cells in NBNCHCC1 (red), NBNCHCC2 (green), and NBNCHCC3 (sky blue) were CD8A⁺ macrophages (MP, cluster #3, #6), GZMB⁺NKG7⁺ Tcell (cluster #1), or FN1^highFABP1⁺cancer cells (clusters #4 and #5). (c) Heat map analysis of the four samples. The color in the heatmap from yellow to violet reveals gradual differences in expression intensity from high to low. The letters listed on the right side of the heatmap indicate representative gene groups or gene symbols. MP1, macrophage cluster 1; Mes, mesenchymal cells; Endo, endothelial cells; Mast: Mast cells; and Pro, MYC^high proliferated cells. (d) Feature plots of representative I-HCA makers, SAA1, SAA2, and CRP. The color bar, from red to blue through yellow, reveals gradual expression intensity differences from high to low through the middle. (e) PLA2G2A features and violin plots. (f) Ratio of double-positive cells to ALB and each gene for total cells in HCA clusters #15 or #18. Violet bars: Cluster #15; red bars: Cluster #18. (g) Gene–gene correlation plots associated with PLA2G2A in clusters #15 and #18. The x-axis indicates the coefficient of determination (R²).

Figure 3

Immunohistochemistry (IHC) of three different IHCA, one bICA, and NBNC-HCC samples. (a) Hematoxylin and eosin (HE) staining, IHC for SAA1/SAA2 or PLA2G2A were performed on the IHCA sample. Scale bar = 100 μm. (b) IHC for PLA2G2A with high magnification. Red arrow shows a representative dotlike staining pattern of PLA2G2A. Scale bar = 50 μm. (c) Low magnification of IHC for SAA1/SAA2 or PLA2G2A in other IHCA tissue. Red arrows; positive reaction, black arrow; negative reaction. Scale bar = 200 μm. (d) IHC for SAA1/SAA2 and PLA2G2A in another IHCA sample. Red arrow; positive reaction, black arrow; negative reaction. Scale bar = 200 μm. (e) HE staining, IHC for SAA1/SAA2, or PLA2G2A were performed on the βHCA sample. Scale bar = 100 μm. (f) HE staining and IHC for PLA2G2A in NBNCHCC tissue. Red or black arrows; representative positive or negative staining. Scale bar = 200 μm.

Figure 4

scRNA-seq of normal and adenoma tissue from one IHCA sample. (a,b) The tSNE plot of the total of 4292 cells (HCA; 1601 cells, normal liver; 2691 cells). Merged cells were divided into 14 cell clusters. The majority of cells in normal liver (red) was HLA-DRA⁺ or ITGAX⁺ macrophages (cluster #1 and #2), CD8A⁺ T-cells (cluster #5 and #6,) or ALB^high hepatocyte cell clusters (cluster #4 and #5). Black dots: adenoma (HCA), red dots: normal liver. (c) Feature gene expression plots of ALB, SAA1, PLA2G2A, and APOA4 genes. The color bar indicates the intensity of gene expression from high (red) to low (blue) through the middle (green). Red circle: higher PLA2G2A⁺ cells and red dots circle: a weaker PLA2G2A⁺ cells. (d) The heat map analysis of adenoma (HCA, cluster #7) and normal liver (Hepatocyte, cluster #4 + cluster #5). The color in the heatmap from yellow to violet reveals the gradual expression intensity differences from high to low. The letters to be listed on the right side of the heatmap indicate the representative gene groups or gene symbols. Six representative genes shown in red color are I-HCA specific genes. (e) Violin plots of PLA2G2A, SAA1, CRP, and ALB genes. x-axis; clusters, y-axis; the intensity of gene expression (arbitral unit).

Figure 5

scRNA-seq of two HBV-HCC samples. (a,b) The t-SNE plot of total of 3900 cells (HCA; 1601 cells, HBV-HCC1; 631 cells, or HBV-HCC2; 1668 cells). Merged cells were divided into 14 cell populations: five cancer cells, four macrophages, one hepatocyte, one adenoma, one T-cell, one endothelial cell (Endo), and one fibroblast (Fibro). Red dots in adenoma (HCA), blue dots: HBV-HCC1, green dots: HBV-HCC2 sample. (c) Feature gene expression plots of SAA1, SAA2, CRP, and PLA2G2A genes. The color bar indicates the intensity of gene expression from high (red) to low (blue) through the middle (green). Red circle: higher PLA2G2A⁺ cells. (d) The violin plots (gene expression plot) of SAA1, SAA2, CRP, and PLA2G2A genes in each cell cluster. (e) The violin plots of ALB, SAA1, CRP, and PLA2G2A genes in each sample. Red: I-HCA, blue: HBV-HCC1, and green: HBV-HCC2. The plot indicates the gene expression of all cells in each sample. (f) The plot indicates the intensity of SAA1 or PLA2G2A gene expression in all ALB-positive cells in each sample.

Figure 6

Bulk RNA-seq of two different FFPE-blocked I-HCA samples. (a) Visualization of RNA-seq results with a Volcano plot. Left: two normal livers (Conts) vs. two I-HCAs, right: four HCCs vs. two I-HCAs. The dotted blue lines indicate the value of fold changes (±1.5-fold, x-axis) and the significance p-value (p < 0.01, y-axis). (b) GSEA enrichment plot for “Inflammation” (TNF-α, interferon-α, inflammatory response) and “Proliferation” (G2M checkpoint, MYC targets) with a comparison of two HCAs vs. two Conts or two HCAs vs. four HCCs. The x-axis shows genes (vertical black lines) represented in different pathways and the y-axis indicates the enrichment score. The color bars at the bottom represent the degree of correlation of the expression of these genes (red to blue through violet, high gene expression to low through middle gene expression). The nominal p value is shown under the color bars. Not significant: N.S. The bottom of the plot shows the value of the ranking metric which measures a gene’s correlation with a phenotype. (c) Heat map of the gene sets containing all genes found significantly up-regulated or down-regulated (±1.5 fold and p value < 0.05) when comparing two I-HCAs vs. two normal livers or two I-HCAs vs. four HCCs. Color ranges from dark red to dark blue representing the highest and lowest expression of a gene, respectively. (d) The Venn diagrams of three different conditions; I-HCAs vs. NBNC-HCC, I-HCAs vs. two Conts, and I-HCAs vs. two HBV-HCCs, are shown. The numbers reveal the genes which are considered to be differentially expressed in I-HCA as compared to each condition. For instance, a total of 1008 genes were differently (significance p < 0.01 and log₂ fold change >2.5) expressed in I-HCAs in comparison with the NBNC-HCC sample. The size of the circle reflects the number of genes. (e) The graph visualizing the expression levels of selected genes of interest in each sample. The y-axis indicates the normalized value (a.u.) by averaged normal livers (Conts). Black bars: normal liver tissues, red bars: I-HCAs, blue bar: NBNC-HCC, green bars: HBV-HCCs, violet bar: HCV-HCC sample. (f) Gene ontology analysis by Metascape. The heatmap of enriched terms across the input differently expressed gene lists, colored by p-value.