High-throughput T cell receptor sequencing reveals distinct repertoires between tumor and adjacent non-tumor tissues in HBV-associated HCC

Abstract

T lymphocytes, which recognize antigen peptides through specific T cell receptors (TCRs), play an important role in the human adaptive immune response. TCR diversity is closely associated with host immune response and cancer prognosis. Although tumor-infiltrating T lymphocytes have implications for tumor prognosis, few studies have performed a detailed characterization of TCR diversity in both tumor and non-tumor tissues in hepatitis B virus (HBV)-associated hepatocellular carcinoma (HCC). Here, we performed high-throughput sequencing of the TCRβ chain complementarity determining region 3 (CDR3) of liver-infiltrating T cells from 48 HBV-associated HCC patients. A significantly higher average number of CDR3 aa clonotypes (2259 vs. 1324, p < 0.001), and significantly higher TCR diversity (Gini coefficient, p < 0.001; Simpson index, p < 0.01; Shannon entropy, p < 0.001) were observed in tumor tissues compared with adjacent non-tumor tissues. The ratio of highly expanded clones (HECs) was significantly higher in non-tumor tissues than in tumor tissues when the HEC threshold was defined as 2% or greater (p < 0.05). Our analysis of the median Morisita-Horn index indicated weak TCR repertoire similarity between tumor and matched non-tumor tissues. The median number of shared clones in tumor tissue and matched non-tumor tissue from each patient was 360.5, representing 5.1-15.8% (10.6 ± 0.4%) of all clones in each patient. We observed extensive heterogeneity of T lymphocytes in tumors and higher HEC ratios in adjacent non-tumor tissues of HCC patients. The differential T cell repertoires in tumor and non-tumor tissues suggest a distinct T cell immune microenvironment in patients with HBV-associated HCC.

Keywords: Diversity; T cell receptor; hepatitis B virus; hepatocellular carcinoma; high-throughput sequencing.

Figure 1.

Numbers of TRBV gene segments used in tumor and adjacent non-tumor tissues. C, tumor tissues; NC, adjacent non-tumor tissues. *** p < 0.001, according to a pair-wise t-test. The error bars indicate the median with the interquartile range.

Figure 2.

Heatmap of Vβ and Jβ gene segments usage in tumors and adjacent non-tumor tissues of HBV-associated HCC. The heatmap bar indicates the usage frequency of Vβ or Jβ gene segments in each sample.

Figure 3.

Comparison of Vβ and Jβ gene segments used in tumors and adjacent non-tumor tissues. (A) Twenty-five Vβ gene segments with different usage frequencies between tumor and non-tumor tissues (all p-values were less than 0.05, Wilcoxon signed rank test). Fold change = the mean frequency in tumor tissues divided by the mean frequency in non-tumor tissues. (B) Two Jβ gene segments with different usage frequencies between tumor (C) and non-tumor tissues (NC). *p < 0.05, **p < 0.001.

Figure 4.

CDR3 aa clonotype comparisons. (A) Comparison of CDR3 aa clonotype numbers between tumors and adjacent non-tumor tissues. (B) Comparisons of CDR3 aa clonotype numbers between different age groups. (C) Comparison of CDR3 aa numbers between tumor and non-tumor tissues in different age groups. (D) Comparison of CDR3 aa numbers between the two gender groups. C, tumor tissues; NC, adjacent non-tumor tissues. *p < 0.05; ***p < 0.001; n.s, no statistical significance.

Figure 5.

Comparison between tumor and adjacent non-tumor tissues. (A) Comparison of the top five CDR3 aa ratios between tumor and non-tumor tissues. (B) Comparison of HEC numbers between tumor and non-tumor tissues. (C) The cutoff for HEC. C, tumor tissues; NC, adjacent non-tumor tissues. *p < 0.05; ***p < 0.001.

Figure 6.

TCR CDR3 diversity between tumor and adjacent non-tumor tissues. Comparison of TCR CDR3 diversity by the Gini coefficient (A), Simpson index (B), and normalized Shannon diversity entropy (NSDE) (C). C, tumor tissues; NC, adjacent non-tumor tissues. **p < 0.01; ***p < 0.001.

Figure 7.

Comparison of shared clones between different groups. (A) Comparison of percentage of shared clones in tumor and adjacent non-tumor tissues. (B) Comparison of numbers of shared clones between different frequency groups. (C) Comparison of percentages of shared clones between different frequency groups. C, tumor tissues; NC, adjacent non-tumor tissues. C > NC, the frequency in C was increased compared with NC; Cp < 0.01, ***p < 0.001, according to a non-parametric test.

Figure 8.

Multidimensional scaling. To distinguish tumor from adjacent non-tumor tissues, we used VDJ (A) and CDR3 aa (B) abundance to perform the MDS.

Figure 9.

Comparison of MHSI and NSDE in different groups. In total, 48 HCC patients were divided into three groups (stage I, stage II and stage III) according to the tumor TNM stage. In addition, patients were divided into three groups (poorly differentiated, moderate differentiated and well differentiated) according to the degree of differentiation. (A–C) comparison of MHSI; (D–F) comparison of NSDE. *p < 0.05; n.s, no statistical significance.