NBS1 I171V variant underlies individual differences in chromosomal radiosensitivity within human populations

Identification of NBS1 I171V variant in a Japanese ovarian cancer patient and generation of knock-in HCT116 cell clones using genome editing technology

To screen genetic variants underlying chromosomal radiosensitivity, we performed whole-exome sequencing of genomic DNA from the peripheral blood cells of 29 Japanese ovarian cancer patients. Average coverage for the exons was more than 100 ×. Since most of the mutations detected in HBOC patients are located in the DNA repair genes, we first extracted genetic variants in the top 10 HBOC genes, namely, BRCA1, BRCA2, ATM, CDH1, CHEK2, PALB2, TP53, Rad51C, Rad51D, and NBS1¹⁰. We also narrowed down the candidate variants on the basis of filtering criteria consisting of the ClinVar database evaluation, genomic position, function, and zygosity. As expected, of these patients, five cases harbored heterozygous mutations of either BRCA1 or BRCA2 (Table 1). In addition, heterozygous ATM missense variants (rs551411717 and rs587782298) in two ovarian cancer patients and a heterozygous TP53 missense variant (rs201382018) in two patients, which might be causative mutations of ataxia-telangiectasia (A-T, OMIM: 607585) and Li-Fraumeni syndrome (OMIM: 191170), respectively, were detected (Table 2). We also identified heterozygous variant c.511A>G, p.Ile171Val, in the NBS1 gene (NM: 002485.5, OMIM: 602667.0007), which encodes a component of the MRE11–RAD50–NBS1 (MRN) complex sensing DNA DSB sites for appropriate repair¹⁷, in one patient (Table 2, Fig. 1a,b). The NBS1 c.511A>G, p.Ile171Val variant was predicted to be “disease-causing” and “probably damaging” by MutationTaster and Polyphen2, respectively. In addition, it was previously reported that a Japanese homozygote of this variant showed aplastic anemia rather than the typical features of Nijmegen breakage syndrome, such as severe microcephaly, immunodeficiency, and cancer predisposition¹¹. These findings suggested that the NBS1 c.511A>G, p.Ile171Val variant might be involved in individual differences in chromosomal radiosensitivity within human populations.

Generation of NBS1 I171V knock-in HCT-116 cells. (a) Structure of human NBS1 protein. NBS1 contains FHA and BRCT1/2 domains at the N terminus and several DNA repair protein-binding regions at the C terminus. FHA and BRCT1/2 domains are involved in the IR-induced nuclear foci with phosphoproteins such as γ-H2AX. NBS1 I171V is located in the BRCT1 domain. (b) Sanger sequencing confirmed NBS1 I171V heterozygosity in ovarian cancer patient 18. (c) Targeting strategy for NBS1 I171V knock-in using the CRISPR/Cas9 system. Blue bases indicate silent mutations. (d–f) Sanger sequencing of NBS1^+/+ parental HCT116 cell (d), NBS1^+/+-edited HCT116 cell clone 1 (e), NBS1^I171V/I171V-edited HCT116 cell clone 1 (f). A single base substitution of the NBS1 I171V variant and a silent Sca I site are indicated by yellow boxes.

To demonstrate that the NBS1 c.511A>G, p.Ile171Val variant indeed underlies chromosomal radiosensitivity, we attempted to generate cultured human cells with knock-in of this variant along with a uniform genetic background using the CRISPR/Cas9 system. We constructed a plasmid vector expressing both Cas9 protein and single guide RNA (sgRNA) for cutting NBS1 exon 5 including c.511A>G, p.Ile171Val (Fig. 1c). The plasmid vector (px459 provided by Addgene) contained the Cas9 gene and a puromycin resistance gene separated by a 2A peptide sequence, expressing the discrete proteins from a single open reading frame. As a targeting donor, we chemically synthesized single-strand oligonucleotides (ssODNs) harboring the NBS1 c.511A>G, p.Ile171Val variant and putative silent CRISPR/Cas9 blocking mutations in the PAM and sgRNA sequences (Fig. 1c). These silent mutations also functioned as an ScaI site for checking ssODN knock-in easily (Fig. 1c). We also designed ssODNs carrying the silent mutations alone in order to evaluate their effect on chromosomal radiosensitivity (Fig. 1c). Since the human colon cancer cell line HCT116 has two copies of the NBS1 allele and relatively high efficacy of ssODN knock-in^18,19, we transfected both the Cas9-2A-puromycin resistance gene plasmid vector and the ssODN targeting donors into HCT116 cells. After transient puromycin selection for 48 h post-transfection and subsequent culture for 2 weeks, the drug-resistant colonies were isolated, and their genotypes were analyzed by ScaI digestion and direct sequencing of the PCR amplicon of the target locus. Finally, two clones of the biallelic NBS1 c.511A>G, p.Ile171Val knock-in cells (NBS1^I171V/I171V cells) and one clone of the biallelic wild-type NBS1 with silent mutation knock-in cells (NBS1^+/+ cells) were generated (Fig. 1d–f). We also established two clones of the NBS1 null HCT116 cells (NBS1^−/− cells) with c.514_514insG, p.Val172fs and c.660_6678del, p.Gln220fs mutations. Western blotting analysis revealed that the NBS1^−/− clones had no signal of NBS1 protein, while the NBS1^I171V/I171V clones showed almost the same amounts of NBS1 protein as the NBS1^+/+ parental HCT116 cells and NBS1^+/+ cell clone, suggesting that the NBS1 c.511A>G, p.Ile171Val variant is not involved in the stability of NBS1 protein (Fig. 2a). It was reported that over extended culturing of HCT116 cells caused mutations in the MRE11 promoter thereby causing loss of expression of MRE11²⁰. These generated clones showed almost the same amounts of MRE11 protein and mRNA level (Fig. 2a,b). Thus, CRISPR/Cas9 system-mediated knock-in technology in the HCT116 cell line enabled the generation of an experimental system for comparing the biological effects among the NBS1 variants with a uniform genetic background.

Chromosomal radiosensitivity of NBS1 I171V edited HCT116 cell clones. (a) Western blotting analysis data showing expression levels of NBS1 and MRE11 protein in NBS1 I171V edited HCT116 cell clones. Full-length gel is presented in Supplementary Fig. 5. GAPDH antibody was used as a loading control. The intensity of NBS1 and MRE11 bands was normalized to that of GAPDH and shown as a percentage, regarding the score of NBS1^+/+ parent cell clones as 100%. (b) RT-PCR data showing mRNA expression levels of MRE11 in NBS1 I171V edited HCT116 cell clones. GAPDH was used as a control. (mean ± SEM; no significant change in each t-test parameter; n = 4) (c) Survival fractions for NBS1 I171V edited HCT116 cell clones for 11 days after irradiation (mean ± SD based on averages from triplicate samples; t-test; n = 3). (d–f) Metafer MN Search images showing the cytokinesis-blocked NBS1 I171V edited HCT116 cells stained with DAPI. Arrowheads indicate MN. BN cell without MN of NBS1^+/+ clone 1 (d); BN cell with one MN of NBS1^I171V/I171V clone 1 (e); BN cell with three MN of NBS1^−/− clone 1 (f). (g) Percentage of IR-induced MN formation in NBS1 I171V edited HCT116 cell clones (mean ± SEM; t-test; n = 3; > 1000 BN cells). (h–j) Representative metaphase after 4 Gy irradiation of NBS1 I171V edited HCT116 cells. Remarkable aberrations are enlarged. Arrows indicate chromosomal breakages. Metaphase without chromosomal breakages of NBS1^+/+ clone 1 (h), metaphase with one chromosomal breakage of NBS1^I171V/I171V clone 1 (i), and metaphase with two chromosomal breakages of NBS1^−/− clone 1 (j). (k) Frequency of IR-induced chromosomal aberration in NBS1 I171V edited HCT116 cell clones (mean ± SEM; t-test; n = 3; > 50 cells).

Chromosomal radiosensitivity is more enhanced in NBS1
^I171V/I171V-edited HCT116 cell clones than in NBS1
^+/+ HCT116 cells

To investigate whether the NBS1 c.511A>G, p.Ile171Val variant affects the cellular lethality after γ-ray irradiation, we performed the colony survival assay of the set of NBS1 genome-edited cell clones. Consistent with previous reports^21,22, the NBS1^−/− clone showed much higher lethal radiosensitivity than the others. The NBS1^I171V/I171V cell clone did not show significant cellular lethality after less than 2 Gy of γ-ray irradiation in comparison with the NBS1^+/+ cell line (Fig. 2c, S3a–c, Table S1), while their lethal radiosensitivity was clearly segregated from those of the NBS1^+/+ cell line in the higher dose exposure. These results implied that the NBS1 c.511A>G, p.Ile171Val variant might contribute to cellular radiosensitivity.

Next, to quantify effect of the NBS1 c.511A>G, p.Ile171Val variant on radiosensitivity at the chromosomal level, radiation-induced micronuclei (MN) in the binucleated (BN) cells were measured (Fig. 2d–f). We evaluated the average ratio of MN to BN cells scored on more than 1000 BN images acquired from Metafer system as a semiautomatic approach⁸, in three independent conditions along with the standard error for each point and create representative dose–response calibration curves. The NBS1^−/− cell clones showed extreme IR-induced micronucleus formation, while the NBS1^I171V/I171V cell clones demonstrated a highly radiosensitive phenotype in comparison with the NBS1^+/+ parental HCT116 cell line and NBS1^+/+ cell clone with silent mutations (Fig. 2g, S3d–f, Table S2). To compare the chromosomal radiosensitivity among these cell clones in a more quantitative manner, we used a linear-quadratic model (MN frequency = c + βD + αD², D: exposure dose) to analyze the dose–response curves of the ratio of MN/BN cells among the NBS1-edited HCT116 cell clones. α, β, and c coefficients were scored by the chromosomal aberration calculation (Cabas) software version 2.0 (http://www.pu.kielce.pl/ibiol/cabas)²³. Since a linear-quadratic model is converted to (MN frequency-c)/D = αD + β, the sum of α and β coefficients at D = 1 Gy accurately represents the radiosensitivity of cells to γ-ray irradiation²⁴. The mean data scores of the sum of α and β (D = 1 Gy) from the NBS1-edited HCT116 cell clones were obtained (Table 3), indicating that biallelic NBS1 null mutations and the c.511A>G, p.Ile171Val variant contributed to approximately 2.0- and 1.3-fold increases of chromosomal radiosensitivity, respectively, in the HCT116 cell genetic background.

To confirm the radiosensitivity of the NBS1 c.511A>G, p.Ile171Val variant at the chromosomal level more directly, we stained the chromosomes with Giemsa dye in the γ-ray-irradiated NBS1 edited cell lines, and then measured the chromosomal aberrations including chromatid and chromosome gaps and breaks²⁵. Consistent with the results of the semiautomatic CBMN assay, the NBS1^−/− cells showed the highest ratio of radiation-induced chromosomal aberrations, while the NBS1^I171V/I171V cells demonstrated an incidence of them intermediate between those of the NBS1^−/− cells and the NBS1^+/+ cell clones (Fig. 2h–k, S3g–i, Table S3). These findings revealed that the NBS1 c.511A>G, p.Ile171Val variant has potency for enhancing chromosomal radiosensitivity.

Chromosomal radiosensitivity increases in a manner dependent on the copy number of Nbs1 I171V variant

We attempted to generate heterozygous NBS1^I171V/+ HCT116 cell clones using the genome editing method mentioned above. However, the candidate cell clones obtained were all genetically mosaics consisting of NBS1^+/+ and NBS1^{I171V/ I171V} cells, and no heterozygous NBS1^I171V/+ single-cell clones were isolated.

Using genome editing technology in mouse fertilized embryos, it is possible to establish heterozygous cell clones for the ssODN knock-in allele. We thus attempted to establish Nbs1 I171V knock-in mice by co-electroporation of CRISPR/Cas9 RNP, which constitutes recombinant Cas9 protein, chemically synthesized crRNA and tracRNA, and ssODN containing the variant, into mouse fertilized eggs (Fig S1a). Two mouse lines were generated from distinct electroporated embryos, in which the Nbs1 p.Ile171Val variant was introduced into the genome. Sanger sequencing confirmed correct knock-in of the variant in both lines. Since the two mouse lines did not exhibit different phenotypes, we mainly addressed data obtained from one line.

No obvious phenotypes were observed in heterozygous mutant mice up to 1 year of age. When they were inbred, wild-type (Nbs1^+/+), heterozygous (Nbs1^I171V/+), and homozygous (Nbs1^I171V/I171V) mice were born at the expected Mendelian ratio. Nbs1^I171V/I171V mice developed and grew normally in the laboratory environment and showed normal reproductive ability, and did not show any hematological profile such as aplastic anemia, which was previously reported in a human homozygote of the NBS1 c.511A>G, p.Ile171Val variant¹¹ (Fig. S2).

To compare chromosomal radiosensitivity among the Nbs1^+/+, Nbs1^I171V/+, and Nbs1^I171V/I171V mice, we intercrossed the Nbs1^I171V/+ mice to generate MEFs with each Nbs1 genotype (Fig S1b–d). Western blotting analysis revealed that both Nbs1^I171V/+ MEFs and Nbs1^I171V/I171V MEFs showed almost the same amount of Nbs1 protein as the Nbs1^+/+ MEFs (Fig. 3a). Next, to quantify the effect of the copy number of the Nbs1 I171V allele on chromosomal radiosensitivity, we measured the MN frequencies post-IR in these MEF lines using the semiautomatic CBMN assay (Fig. 3b–d). The Nbs1^I171V/I171V MEFs showed the highest rate of IR-induced micronucleus formation, and the Nbs1^I171V/+ MEFs demonstrated a moderately radiosensitive phenotype in comparison with the Nbs1^+/+ clones (Fig. 3e, S4a-c, Table S4). We also used Cabas software to evaluate the dose–response curves of the Nbs1 mutant MEFs (Table 4). The mean data scores (D = 1 Gy) suggested that one copy of the Nbs1 I171V allele contributed to an approximately 1.1-fold increase and two copies an approximately 1.7-fold increase of radiosensitivity compared with that of the biallelic wild-type Nbs1. Cytogenetic analysis also demonstrated that IR-induced chromosomal aberrations increased in a manner dependent on the copy number of the Nbs1 I171V allele (Fig. 3f–i, S4d–f, Table S5). Taken together, these findings suggest that the NBS1 c.511A>G, p.Ile171Val variant might be a genetic factor underlying individual differences in chromosomal radiosensitivity within human populations.

Radiation-induced chromosomal instability of Nbs1 I171V edited MEF clones. (a) Western blotting analysis data showing the expression levels of Nbs1 protein in Nbs1 I171V edited MEF clones. Full-length gel is presented in Supplementary Fig. 5. The GAPDH antibody was used as a loading control. The intensity of Nbs1 bands was normalized to that of GAPDH and shown as a percentage, regarding the score of Nbs1^+/+ clone 1 as 100%. (b–d) Metafer MN Search images showing the cytokinesis-blocked Nbs1 I171V edited MEFs stained with DAPI. Arrowheads indicate MN. Nbs1^+/+-BN cell without MN (b), Nbs1^I171V/+-BN cell with one MN (c), Nbs1^I171V/I171V-BN cell with three MN (d). (e) Percentage of IR-induced MN formation in Nbs1 I171V edited MEF clones (mean ± SEM; t-test; n = 3; > 1000 BN cells). (f, g) Representative metaphase of Nbs1 I171V-edited MEFs after 4 Gy irradiation. Remarkable aberrations are enlarged. Arrows indicate chromosomal breakages. (i) Frequency of IR-induced chromosomal aberrations in Nbs1 I171V edited MEF clones (mean ± SEM; t-test; n = 3; > 50 cells).