Large chromosomal deletions and impaired homologous recombination repairing in HEK293T cells exposed to polychlorinated biphenyl 153

Materials

All cell culture medium and components, such as Dulbecco’s modified Eagle’s medium (DMEM) and penicillin-streptomycin (PS) were provided by Invitrogen (Waltham, MA, USA). Fetal Bovine Serum (FBS) was obtained from Thermo Fisher Scientific (Waltham, MA, USA). Dimethyl sulfoxide (DMSO) was purchased from Sigma (St. Louis MO, USA). PCB153 (molecular weight, 361 g/mol; purity, 99.9%) (Fig. 1A) was provided by Accustandard (USA).

Cell culture

HEK293T cell line provided by Professor Feng Wang (Tianjin Medical University) was used in the present study. The HEK293T cell line was cultured in high-glucose DMEM with 10% FBS and 1% PS in a humidified apparatus at 37 °C with 5% CO₂ (Lu et al., 2016).

HEK293T cells exposure in PCB153

A stock solution of PCB153 was dissolved in DMSO and 0.05% DMSO is the final concentration after adding into cell culture medium. HEK293T cells were treated with PCB153 at different concentrations of 5 µM, 10 µM, and 15 µM (Eum et al., 2009; Xin et al., 2016). Control cultures only received the solvent. Cells were collected after 48 h and 96 h of culture. For genome and RNA sequencing, HEK293T cells exposed to 15 µM PCB153 for 96 h were used.

Cell viability examination

The cell viability was measured using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) test. The HEK293T cells were seeded in 96-well plates at 1  × 10⁴ cells per well. The gradient concentrations of PCB153 were added after 24 h. The HEK293T cells treated with PCB153 were continually cultured for 24 h. Totally 20 µL of MTT solution (0.5% MTT) were added and incubated for 4 h. 150 µL of DMSO was added to each well to dissolve formazan crystals. The light absorption value was measured at 490 nm.

Whole-genome sequencing (WGS)

293T cell precipitations were randomly interrupted by the High Performance Ultrasonic Sample Processing System (COVARIS), and fragments of about 350bp were obtained after fragment selection. The end of the DNA fragment was then repaired by adding “A” base to the 3 ’end and adding library connectors to both ends. Linear amplification (LM-PCR) was performed on the ligated libraries. After rolling circle amplification (RCA), the DNA Nano Ball (DNB) was generated from the library. After qualified quality control, the DNA Nano Ball (DNB) could be processed by computer. We used the BGISEQ-500 platform to perform high-throughput sequencing on each qualified library and ensure that the data volume of each sample was up to standard.

Total RNA extraction and RNA sequencing (RNA-seq)

Total RNA was extracted from cell precipitations using Trizol (Invitrogen, Carlsbad, CA, USA) according to manual instruction. Before detection by Fragment Analyzer, RNA samples were thawed on the ice, thoroughly mixed and centrifuged. Standard Sensitivity RNA Analysis Kit (15 nt) (DNF-471) was used to detect the samples, and the samples were diluted 3–8 times at a detection concentration of 400–1,000 ng/ µL. RIN values of all samples were greater than or equal to 7.0 and 28S/18S ratios were greater than or equal to 1.5, indicating that the extracted RNA was of good quality and met the requirements of sequencing. After sequencing data quality control and filtering, we compared the filtered clean reads to the reference sequence. After the alignment, whether the alignment results passed the second quality control (QC of alignment) was determined by statistical analysis of alignment rate and distribution of reads in reference sequences. In the case of secondary quality control, quantitative gene analysis and various analyses based on gene expression level (differential gene screening, etc.) were carried out, and differentially expressed genes were detected among the screened samples.

DNBSEQ platform was used for RNA-sequencing. Pathway enrichments were carried out by the phyper function in R software according to KEGG pathway annotation classifications. Then Q value was obtained by FDR correction of the P value, and Q value ≤ 0.05 was usually considered as significant enrichment. According to Heatmap, Bowtie2 was applied to compare clean reads to the reference gene sequence, and then RSEM was used to calculate the gene expression level of each sample. The DESEQ2 method is based on the principle of negative binomial distribution. The method described by Love et al. was used to conduct differentially expressed gene (DEG), which is a test for a gene whose expression level varies between samples (Love, Huber & Anders, 2014). According to the results of DEG detections, the differential genes were performed by using R-package pheatmap for hierarchical clustering analysis.

Immunofluorescent staining

Immunofluorescence tests were used to measure 53BP1 expression as described in previous study (Noordermeer et al., 2018). We fixed the cells with 2% paraformaldehyde for 10 min, permeabilized with 0.5% Triton X-100 in PBS for 20 min and incubated with blocking solution (1% bovine albumin V, 1% gelatin in PBS) for 1 h. The cells were incubated with 53BP1 antibodies (1:5000 dilution; Novus Biological, Littleton, CO, USA) for 2 h. After washing in 0.1% Tween 20 in Tris-buffered saline (TBST), the cells were incubated for 30 min with secondary antibodies (1:5,000 dilution; Life Technologies, Carlsbad, CA, USA) in the dark. After dehydrated in ethanol, the cells were mounted with 4′,6-diamidino-2-phenyl-indole (DAPI, D3571; Life Technologies, USA). The positive signals were observed and counted by using fluorescence microscopy (Nikon, Japan).

Real-time quantitative PCR (qPCR) analysis

The mRNA expression of BRCA1, RAD51B, and RAD51C were examined by qPCR. The total RNA was reverse-transcribed into cDNA with a reverse transcription system (Promega, Madison, WI, USA) consistent with the manufacturer’s protocol. The objective sequences of genes to be observed were amplified with SYBR green master mix (QIAGEN, Germany) using the Rotor gene Q 2000 (QIAGEN, Germany). The primers were listed in Table S1. The level of GAPDH served as control for the data analysis.

Statistical analysis

FACET software was used to detect Somatic CNV in Tumor and Normal pairs. Phyper function in R software was used for enrichment analysis to calculate p value, and then Q value was obtained by FDR correction of P value, Q value ≤ 0.05 was considered as significant enrichment. Bowtie2 was used to compare clean reads to the reference gene sequences, and then RSEM was used to calculate the gene expression level of each sample. The DESEQ2 method was used to conduct DEG for differentially expressed genes. The union of differential genes was performed by using R-package pheatmap for hierarchical clustering analysis.

All experiments were repeated at least three times. The results were expressed as mean ± standard deviation (SD). The student’s t-test was used for comparisons between two groups, and one-way ANOVA was used for comparisons among multiple groups by IBM SPSS Statistics 21. Differences with P < 0.05 was considered statistically significant.

PCB153 poisoning caused multiple large fragment deletions on many chromosomes

The cytotoxicity of PCB153 on HEK293T cells was analyzed by MTT. Compared with 0.05% DMSO-treated cells, PCB153 did not affect cell viability (Fig. 1B). To explore the genotoxic events caused by PCB153, copy number variations (CNVs), single nucleotide variants (SNVs), and insertions and deletions (Indels) were determined by WGS. The results indicated that PCB153 poisoning generated apparent change of CNVs, SNVs and Indels in the whole-genome (Tables S2–S6).

There were 34 large (>1 Mb) copy number variations (CNVs), including 4 duplications (282.6 Mb) and 30 deletions (245.4 Mb) were identified in PCB153-exposed cells (Fig. 1C, Tables S2–S4). It was shocking that CNVs covered multiple chromosomes and large fragment deletions could be observed on many autosomes in PCB153-exposed HEK293T cells (Fig. 1D). And more than half of the fragment deletion events were homozygote deletions (Table S4). For example, terminal deletions from the short arm of chromosome 9 and chromosome 18 generated about 18 Mb and 15 Mb length fragments loss, respectively, with lots of genes affected (Figs. S2, S3).

PCB153 poisoning induced missense mutations of 6 tumor suppressor genes

Germline mutation information was also analyzed through comparing the detected mutant genes with the CGC (Cancer Gene Censue) database using GATK software. And there were missense mutations located in 6 tumor susceptibility genes, including PTCH1, BRCA2, BLM, ERCC4, BRCA1, and SETBP1 (Table S7), whose function loss have been closely associated with several hereditary tumor susceptibility syndromes (Acuna-Hidalgo et al., 2017; Qian et al., 2017; Romagnolo, Romagnolo & Selmin, 2015; Zhang et al., 2018). Interestingly, BRCA1 and BRCA2 are key members participating in homologous recombination (HR) repair upon DNA damage and the missense mutations were likely to affect the HR repair potential.

PCB153 exposure inhibited HR repair in HEK293T cells

RNA-seq was applied to examined the altered pathways that might be related to the genotoxicity of PCB153 exposure. Figure 2A showed a VENN diagram of the overlapping genes between PCB153 exposed cells and control samples. The sequencing data showed 1051 up-regulated and 1622 down-regulated DEGs (volcano plot, Fig. 2B). The enriched pathways included ‘homologous recombination repair’ (KEGG Term ID 03440) and ‘mismatch repair’ (KEGG Term ID 03430) (volcano plot, Fig. 2C). And most of the genes involved in the two cascades were downregulated (Fig. 3, Fig. S1), suggesting the inhibited HR repair and mismatch repair responses. The reduction of DNA damage repair ability will probably lead to severe mutations.

PCB153 poisoning caused DNA double-strand break (DSB) and HR repair genes downregulation in the HEK293T cells

The content of 53BP1, a marker of DSBs, was examined to show the effects of PCB153 poisoning on DNA damage in HEK293T cells in vitro. The 53BP1 foci assay showed observably increased 53BP1 signals after PCB153 exposure compared to the control cells (Fig. 4A), indicating the aggravated DSBs. The mRNA expression of some important HR repair genes were determined by qPCR. Consistent with the RNA-seq results, the expression levels of BRCA1, RAD51B, and RAD51C that involved in HR repairing were all significantly decreased after 15 µM PCB153 exposure (Figs. 4B–4D). There were no obvious changes in the transcription level of LIG4, XRCC5, and XRCC6, which are responsible for non-homologous end-joining (NHEJ) (Fig. S4). These results demonstrated that PCB153 poisoning caused obvious DNA damage and apparently inhibited DNA HR repair.