A comprehensive study of immunology repertoires in both preoperative stage and postoperative stage in patients with colorectal cancer

2.1 Patients and controls

Whole blood samples from 19 CRC patients and five healthy controls were collected at The Second Medical College of Jinan University (Shenzhen People's Hospital), Shenzhen, China, and PBMCs were extracted. We collected the PMBCs of 10 colon cancer patients, who had taken the TIVA anesthesia pattern, at 1 day preoperative, 3 days’ postoperative, and 7 days’ postoperative time point, respectively. The PMBCs of nine CRC patients, who had taken the CEGA anesthesia, at 1 day preoperative, 3 days’ postoperative, and 7 days’ postoperative time point were collected. The Ethical Committee of the Department of Anesthesiology, Shenzhen People's Hospital, 2nd Clinical Medical College of Jinan University, Shenzhen, Guangdong, China, reviewed and approved our study protocol in compliance with the Helsinki Declaration.

2.2 T-cell isolation and DNA extraction

Informed consent was obtained from all the participants in our study. T-cell isolations were performed using superparamagnetic polystyrene beads (Miltenyi) coated with monoclonal antibodies specific for T cells. DNA was prepared from 0.5 to 2 × 106 T cells from each sample, which was sufficient for analyzing the diversity of TCRV in the T-cell subsets. DNA was extracted from PBMCs using GenFIND DNA (Agencourt, Beckman Coulter, Brea, CA, USA) extraction kits following the manufacturer's instructions.

2.3 Multiplex-PCR amplification of the TCR CDR3 region

The TCR CDR3 region was defined according to International ImMunoGeneTics collaboration, starting with the second conserved cysteine encoded by the 39 portions of the V gene segment and ending with the conserved phenylalanine encoded by the 59 portions of the J gene segment. To generate the template library for Genome Analyzer, a multiplex-PCR system was designed to amplify rearranged TCR CDR3 regions from genomic DNA using 45 forward primers, each is specific to a functional TCR V segment, and 13 reverse primers, each is specific to a TCR J segment. The forward and reverse primers contain, at their five ends, the universal forward and reverse primer sequences, respectively, which are compatible with GA2 cluster station solid-phase PCR. After amplification and selection, the products were purified using QIAquick PCR Purification Kit. The final library was quantitated in two ways: by determining the average molecule length using the Agilent 2100 bioanalyzer instrument (Agilent DNA 1000 Reagents) and by real-time quantitative PCR (QPCR; TaqMan Probe). The libraries were amplified with cBot to generate the cluster on the flow cell, and the amplified flow cell was pair-end (PE) sequenced using a Hiseq2500 instrument, with a read length of 100 as the most frequently used sequencing strategy.

2.4 High-throughput sequencing and data analysis

The PCR products were sequenced using an Illumina Genome Analyzer, and the sequencing quality of these reads was evaluated by the formula shown below. The quality of the HiSeq sequencing ranged from 0 to 40 and was used for filtering out low-quality reads. First, we filtered the raw data, including adapter contamination. Reads with an average quality score lower than 15 (Illumina 0–41 quality system) were removed, and the proportion of N bases was not more than 5% (sequences with higher values were also removed). Next, a few bases with low quality (lower than 10) were trimmed; the quality score was expected to be over 15 after trimming, and the remaining sequence length was expected to be more than 60 nt. After filtering, PE read pairs were merged into one contig sequence in two steps: (a) by aligning the tail parts of two sequences and assessing the identity (BGI developed software COPE v1.1.3), with at least 10 bases of overlap required and the overlapping section having 90% base match; (b) as different primers might result in sequences of different lengths, some might be very short (<100 bp) and such reads were merged by aligning the head part of the sequence (BGI developed software FqMerger). In this way, we obtained the merged contig sequences and the length distribution plot. Subsequently, we used miTCR, developed by MiLaboratory (https://mitcr.milaboratory.com/downloads/) for the alignment. This program has an automated adjustment mechanism for errors introduced by sequencing and PCR and will provide alignment statistical information, such as the CDR3 expression and INDEL. After alignment, we utilized the following method for the sequence structural analysis: (a) We calculated the number of each nucleotide and analyzed the proportion at each position; (b) according to the last position of the V gene, start site of the D gene, end site of the D gene, and start site of the J gene after alignment, we retrieve the INDEL (insertion and deletion) introduced during V–D–J recombination; (c) nucleotides were translated into amino acids. According to the identity of each sequence after alignment, the expression level of each clone was clear and calculated. The expression of each distinct DNA sequence, amino acid sequence, and V–J combination was also identified. In addition, to measure the diversity of each sample, we calculated the distinct clone number, Simpson coefficient, and Shannon–Waver coefficient based on different resolutions of distinct DNA sequences, amino acid sequences, and V–J combinations. The expression level of each sample was also calculated at different resolutions of distinct DNA sequence, amino acid sequence, and V–J combination. Moreover, we constructed the specific expression graph and plotted a heatmap according to the V–J combination profile. The diversity of the TCR repertoire was calculated based on the Simpson index of diversity (Ds) and the Shannon–Wiener index (H).

2.5 Statistical analysis

Because of the small sample size in this study, the analysis of differences among the data groups was performed with the t test. p values <0.05 were considered significant. The statistical analyses were conducted with GraphPad Prism software (GraphPad Software, San Diego, CA, USA).

3.1 Sequencing data summary

A total of 19 colorectal cancer patients and five healthy controls were recruited for this study. Blood sample from 1 day preoperative, 1 day postoperative, 3 days’ postoperative, and 7 days’ postoperative was collected. We obtained an average of 713,362 sequencing reads per sample (Figure 1). The mean unknown sequence number is 15,856, and the mean immune sequence number is 697,505. The productive sequence number and the nonproductive sequence number are 519,165 and 178,340, respectively. And In-frame sequences number and Out-of-frame sequences number are 539,825 and 151,114, respectively. The total CDR3 sequence number, unique CDR3 nt sequence number, and Unique CDR3 aa sequence number are 505,707, 54,921, and 47,892 respectively (Table 1).

3.2 CDR3 length distribution pattern

The length distribution of the TCR CDR3 is an important determinant of T-cell repertoire diversity. In this study, we first assessed the length distribution of TCR CDR3 sequences (aa) in the preoperative group and healthy control group (Figure 2). TCR CDR3 sequence length in preoperative group was significantly higher compared to those in the NC group with the amino acid length 1 (p = 0.028), 28 (p = 0.026), 29 (p = 0.0064), and 30 (p = 0.00078).

We draw the Gaussian distribution curve for each sample, and the goodness of fit was quantified by R², which ranges from 0 to 1 (from lowest fitness to highest fitness). R² values were calculated for each sample and compared between the preoperative CRC patients and healthy controls (p = 0.0016; Figure 3).

3.3 Highly expanded clones and TCR repertoires diversity

The expression level of each unique clone is another major measurement or index for immune diversity. After aligning to the human genome reference, the expression level of each clone is calculated. In this study, the TCR clones with frequency above 0.5% of total reads in a sample were defined as highly expansion clone (HEC). The comparison of HEC number between preoperative group and healthy control group showed significant higher HEC ratio in preoperative group (Figure 4a). In the TIVA group, HEC number of 3 days’ postoperative samples was higher than 1 day preoperative group (p = 0.021) and also higher than the value of 7 days’ postoperative group (p = 0.018; Figure 4b).

The HEC ratio of TIVA group also showed similar distribution of HEC number; the 3 days’ postoperative group showed higher HEC than 1 day preoperative group (p = 0.031), higher than the HEC ratio of 7 days’ postoperative group (p = 0.015). The HEC ratio of 7 days’ postoperative group was lower than 1 day preoperative group (p = 0.022; Figure 4c). To further study the difference in effect of TIVA and CGEA on immune repertoire, we also compared the HEC number and HEC ratio at the 3 different time period. The results showed there was no significant difference in the effect between the two anesthesia methods. (Figure 4d,e). The comparison of HEC number and HEC ratio at different time period in CGEA group showed no significant difference (Figure 5).

To further understand the percentage of shared HECs, we then analyzed the top 60 highly used amino acids and nucleotide sequences in 0.5% used clones of CRC patients and healthy controls. According to Figure 6, there were highly shared sequences in CRC patients than in healthy controls.

3.4 Comparison of TRBV and TRBJ gene repertoires between preoperative patients and healthy controls

To determine the disease-specific TCR repertoire characteristics, we compared the expression levels of TRBV and TRBJ genes of preoperative patients and healthy controls. In comparison with TRBV gene between preoperative patients and healthy controls, TRBV11-2 (p = 0.016), TRBV12-1 (p = 0.0068), TRBV16 (p = 0.0032), TRBV3-2 (p = 0.0096), TRBV4-2 (p = 0.03), TRBV4-3 (p = 0.048), TRBV5-4 (p = 0.011), TRBV6-8 (p = 0.038), TRBV7-8 (p = 0.042), and TRBV7-9 (p = 0.023) usage showed significant difference (Figure 7a). In contrast, the differentiation of TRBJ gene between preoperative and healthy control patients showed significant difference in the usage of TRBJ1-3 (p = 0.035), TRBJ2-2 (p = 0.00053), and TRBJ2-5 (p = 0.023; Figure 7b). Analysis of top 20 used TRBV genes was performed; TRBV7-8, TRBV7-9, and TRBV9 were well used in both CRC patients and healthy controls. TRBV2, TRBV12, TRBV19, TRBV20-1, and TRBV24-1 were poorly used in either CRC patients or healthy controls (Figure 8).