The MUTYH DNA glycosylase counteracts mutagenesis by removing adenine misincorporated opposite DNA 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodG). Biallelic germline mutations in MUTYH cause the autosomal recessive MUTYH-associated polyposis (MAP). The impact on genetic instability of the p.Tyr179Cys and p.Arg245His MUTYH variants was evaluated in lymphoblastoid cell lines (LCLs) derived from MAP patients and their relatives in comparison to wild-type LCLs. No difference in MUTYH expression was identified between wild type and LCLs with the p.Tyr179Cys, while the p.Arg245His mutation was associated with an unstable MUTYH protein. LCLs homozygous for the p.Tyr179Cys or the p.Arg245His variant contained increased DNA 8-oxodG levels and exhibited a mutator phenotype at the PIG-A gene. The extent of the increased spontaneous mutation frequency was 3-fold (range 1.6- to 4.6-fold) in four independent LCLs carrying the p.Tyr179Cys variant, while a larger increase (6-fold) was observed in two p.Arg245His LCLs. A similar hypermutability and S-phase delay following treatment with KBrO3 was observed in LCLs homozygous for either variant. When genetic instability was investigated in monoallelic p.Arg245His carriers, mutant frequencies showed an increase which is intermediate between wild-type and homozygous cells, whereas the mutator effect in heterozygous p.Tyr179Cys LCLs was similar to that in homozygotes. These findings indicate that the type of MUTYH mutation can affect the extent of genome instability associated with MUTYH inactivation. In addition, the mild spontaneous mutator phenotype observed in monoallelic carriers highlights the biological importance of this gene in the protection of the genome against endogenous DNA damage.
8-Oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodG) is one of the most common lesions produced by reactive oxygen species (ROS). The major function of MUTYH (mutY homolog), a monofunctional DNA glycosylase, is to counteract the mutagenic effects of 8-oxodG. It does this by excising adenine misincorporated opposite template 8-oxodG to initiate base excision repair (BER). BER-mediated resynthesis generates C:8-oxodG base pairs, which are the substrates for OGG1-mediated removal of the oxidized purine. The action of MUTYH prevents the formation of G:C to T:A transversion mutations (for reviews 1–4). The importance of preventing mutations associated with persistent DNA 8-oxodG is shown by the genetic association between defects in the human MUTYH gene (OMIM *604933) and development of colorectal cancer (CRC). Germline biallelic mutations in this gene are associated with a recessively heritable colorectal polyposis, named MUTYH-associated polyposis (MAP) (OMIM #608456), linked to an increased risk of CRC. The disease phenotype of MAP is usually relatively mild with regard to colorectal polyps burden/number, in most cases mimicking the attenuated form of familial adenomatous polyposis (FAP). In contrast to tumors from FAP patients, who have germline mutations in the adenomatous polyposis coli (APC) gene, APC in tumors from MAP patients contains distinctive somatic G:C to T:A transversions (5–7). As a consequence, it is generally regarded that biallelic MUTYH mutations drive genomic instability in colorectal epithelial cells and thereby increase colon cancer risk in MAP patients. Although a number of studies have confirmed that biallelic mutation carriers are at high risk of CRC, the degree to which carriers of a mutation inherited from only one parent are at increased risk of CRC remains uncertain (8–12).
MUTYH inactivation in Mutyh−/− mice is associated with an increased susceptibility to spontaneous and oxidative stress-induced intestinal tumorigenesis (13) and accumulation of steady-state level of DNA 8-oxodG (14). The same lesion accumulates in Mutyh−/− mouse embryo fibroblasts (MEFs) expressing human MUTYH variants (15) and lymphoblastoid cell lines (LCLs) derived from MAP patients (16). The biological effects of the different MUTYH variants are, however, largely undefined. Although the MAP patient LOVD database (http://www.LOVD.nl/MUTYH) contains >300 unique MUTYH variants, most of our understanding of the biological consequences of MUTYH inactivation comes from studies of the effects of c.536A > G (p.Tyr179Cys) and c.1187G > A (p.Gly396Asp), the most commonly documented mutations in Caucasian populations. We previously showed that the extent of the mutator phenotype in LCLs derived from MAP patients harboring different mutations is quite variable (16). To investigate whether these differences depend on the type of mutation and/or on inter-individual variability, here we present a functional characterization of LCLs harboring the same mutation in the MUTYH gene. We characterized LCLs carrying the p.Tyr179Cys or the p.Arg245His (c.734G > A) variants, the first one because is one of the most common mutations among MAP patients (>500 entries in the LOVD data bank) and the second one because of its previously described strong mutator phenotype (16). Neither variant protein has detectable DNA binding or DNA glycosylase activity toward 8-oxodG:A mismatches (17–20). Clinical information was collected and LCLs were established from patients and from heterozygous relatives. MUTYH expression, DNA 8-oxodG levels, spontaneous and oxidative stress-induced mutant frequency were examined in LCLs from MAP patients and monoallelic MUTYH mutation carriers and compared with wild-type LCLs.
Patient clinical phenotype
The phenotypic features of the six MAP patients and four heterozygous relatives are shown in Table 1. According to the average MAP features, 5 of 6 biallelic carriers were phenotypically similar to attenuated APC-polyposis patients and had a limited number of colorectal polyps (usually <100 in total), with the first diagnosis of colorectal disease at a late age (4th and 5th decade). Two p.Tyr179Cys and one p.Arg245His carriers also showed a CRC at the time of polyposis diagnosis. The FAP24 patient (p.Tyr179Cys), with a classical FAP phenotype (21 years at diagnosis, 300 total polyps), was also heterozygous for the p.Glu1317Gln variant in the APC gene. All MAP patients tended to develop variable numbers of polyps during the surveillance period. In view of wide differences in the follow-up time (1–22 years) the data on polyp numbers must be interpreted with caution. Notably, the pathology records of the two patients FAP07 and FAP483 with p.Arg245His variant reported co-existing adenomatous and hyperplastic polyps, although some adenomas of patient FAP24 also displayed partial hyperplastic/metaplastic features. Among the four heterozygous relatives, only FAP47 had a clinical history of multiple polyps, with development of three adenomas before the age of 67 years.
|Patient ID1||Allele 1||Allele 2||Sex2||Diagnosis age (years)||No. polyps at diagnosis||Polyps histology3||Years of follow-up||Total polyps no.4||CRC age (years)5||Colectomy age6||Other7|
|FAP07#||p.Arg245His||p.Arg245His||M||43||7||Ad, HP||14||60||43^^||43^||UGI Ad|
|Patient ID1||Allele 1||Allele 2||Sex2||Diagnosis age (years)||No. polyps at diagnosis||Polyps histology3||Years of follow-up||Total polyps no.4||CRC age (years)5||Colectomy age6||Other7|
|FAP07#||p.Arg245His||p.Arg245His||M||43||7||Ad, HP||14||60||43^^||43^||UGI Ad|
1FAP12 and FAP13 were sisters (*), FAP47 and FAP48 were relatives of FAP24 (**), FAP513 and FAP515 were relatives of FAP483 (***); § FAP24 and FAP47 were also heterozygous carriers of the p.Glu1317Gln variant of the APC gene (38); # identical-by-descent mutant alleles.
2F, female; M, male.
3Ad, adenoma, HP, hyperplastic.
4Approximate total number of polyps found at first and follow-up colonoscopies.
5In total, 3 malignant lesions were diagnosed: two synchronous CRC at diagnosis and one intraepithelial high-grade neoplasia at age 50 (^^).
6Total colectomy or hemi-colectomy (^).
7Other clinical manifestations: SB Ad, small bowel adenoma; UGI Ad, upper gastro-intestinal adenoma.
Expression of MUTYH in LCLs derived from MAP patients
MUTYH protein levels were measured by western blotting of LCL extracts from wild-type BR77 and homozygous and heterozygous MUTYH variants (Fig. 1A and B). Although some variability in the amount of MUTYH protein was observed among LCLs with the same biallelic p.Tyr179Cys genotype, no significant differences were identified between wild type and homozygous and heterozygous groups of p.Tyr179Cys LCLs (Fig. 1B). In contrast, in comparison to extracts from wild-type LCLs, MUTYH levels were reduced by 70% in both homozygous p.Arg245His LCLs. Since we previously reported a moderate decrease (30%) in the MUTYH transcript in LCLs with the homozygous p.Arg245His (16), these data indicate that the p.Arg245His mutation is associated with an unstable MUTYH protein. The 70–80% of protein levels found in the heterozygous p.Arg245His LCLs is consistent with a 50% expression of the wild-type protein and a residual 20–30% of the mutant one.
Measurements of 8-oxodG levels
We previously demonstrated that MUTYH gene inactivation is associated with DNA 8-oxodG accumulation in LCLs from MAP patients (16). To examine the effects of specific MUTYH mutations, the steady-state DNA 8-oxodG levels were measured in LCLs from homozygous and heterozygous individuals and compared with wild-type BR77 LCL (Fig. 2A). The basal level of DNA 8-oxodG in LCLs carrying the p.Tyr179Cys or p.Arg245His variants in homozygousity was 1.4- to 1.7-fold higher than the 0.67 × 10−6 8-oxodG per dG in BR77 cells (P ≤ 0.01–0.0001, Student's t-test). Increases in the range of 1.25- to 1.5-fold were also observed in heterozygous p.Tyr179Cys LCLs with values of 0.85 and 0.99 residues × 10−6 dG (P = 0.07 and 0.005, respectively). There was no significant difference between the levels in p.Arg245His heterozygotes and wild-type LCL (0.75 and 0.66 residues × 10−6 dG in FAP513 and FAP515, respectively). The data show that in this particular test the p.Arg245His mutation is recessive, whereas the p.Tyr179Cys mutation is co-dominant with respect to the wild-type cells.
To examine the effect of p.Tyr179Cys and p.Arg245His mutations on 8-oxodG repair, cells were treated with the oxidant KBrO3. Following a 30 min exposure to 20 mm KBrO3 repair kinetics of p.Tyr179Cys (FAP12 and FAP13) and p.Arg245His (FAP483) LCLs were compared with that of wild-type BR77 cells. The rate of disappearance of 8-oxodG from genomic DNA was slower in the MAP variants than in wild-type BR77 cells and the half-life of the oxidized base increased from 3 to 6 h (Fig. 2B). The observation that removal of 8-oxodG occurs more slowly in MUTYH defective than in wild-type cells reflects the relatively small fraction of replication-dependent damage recognized by MUTYH, i.e. adenine misincorporated opposite template 8-oxodG. Thus, the increased steady-state DNA 8-oxoG levels in the homozygous p.Tyr179Cys and p.Arg245His LCLs reflects 8-oxodG:A mispair accumulation and consequent impaired removal of the oxidized purine by OGG1. The severity of this defect is similar in the two mutants.
Spontaneous mutagenesis in LCLs
The impact of p.Tyr179Cys and p.Arg245His mutations on spontaneous mutagenesis was analyzed by the PIG-A assay (Fig. 3). PIG-A mutations result in complete or partial deficiency of membrane glycosyl-phosphatidylinositol (GPI)-linked proteins. PIG-A mutant cells can be identified by fluorescence-activated cell sorting for CD48-, CD59- and CD55-negative cells (21). The mean frequency of GPI-negative cells in three wild-type cell lines (BR77, BR805 and BR806) was 5.0 × 10−6, ranging from 1.6 × 10−6 (BR806) up to 7 × 10−6 (BR77) (Fig. 3). The values for homozygous p.Tyr179Cys LCLs were 23.1, 15.8, 8.2 and 16 × 10−6 GPI-negative events for FAP12, FAP13, FAP599 and FAP24, respectively, yielding an average frequency of 16.3 × 10−6 and an increase over wild type of >3-fold. The mutant frequencies observed in two monoallelic mutation carriers (FAP47 and FAP48) was in a comparable range (8.2 and 22.2 × 10−6, respectively) (Fig. 3). Comparison by one-way ANOVA showed that the increase in spontaneous mutant frequency was statistically significantly in both homozygous and heterozygous p.Tyr179Cys LCLs (F = 9.18; P = 0.00085).
The mutant frequencies in the two homozygous p.Arg245His cell lines were comparable (33.2 × 10−6 and 27.5 × 10−6 GPI-negative events in FAP483 and FAP07, respectively) and were elevated 6.6-fold over wild-type levels (Fig. 3). In contrast a single mutant allele in FAP513 and FAP515 was associated with mutant frequencies of 18.3 and 16.8 × 10−6, values that are intermediate between those in wild-type and homozygous p.Arg245His cells (3.3-fold). One-way ANOVA again highlighted a significant genotype effect on spontaneous mutant frequency across the three groups (F = 29.44, P = 0.0005).
Thus, whereas there was no gene dosage effect for the p.Tyr179Cys variant, the mutator effect in heterozygous p.Arg245His LCLs was half of that seen in the homozygous group. In addition, there was a difference in mutant frequency between homozygous p.Arg245His and p.Tyr179Cys LCLs, with the mutator phenotype in p.Arg245His LCLs being significantly more pronounced (P = 0.0015 by Student's t-test).
We have previously shown that cells compound heterozygous for the p.Gly396Asp/p.Glu410Glyfs*43 MUTYH variants were more mutable by KBrO3 than wild-type cells (16). To investigate whether this phenotype extends to other MUTYH variants, we compared KBrO3-induced mutagenesis in wild-type LCLs (BR77 and BR806) and homozygous and heterozygous p.Tyr179Cys and p.Arg245His LCLs. Although hypersensitivity to KBrO3-induced killing has been observed in Mutyh−/− MEFs expressing human MUTYH variants (15), KBrO3 sensitivity is not a common phenotype among MUTYH-defective human cell lines (Fig. 4A). In contrast, both p.Tyr179Cys and p.Arg245His LCLs were hypermutable by KBrO3 in comparison to wild-type cells (Fig. 4B). Thus, in wild-type cells exposure to 60 mm KBrO3 introduced into the genome over background level, 18.7 × 10−6 mutations (range 16.6–20.8 × 10−6 for BR77 and BR806, respectively), while in p.TyrY179Cys and p.Arg245His LCLs this value reached almost 40 × 10−6 mutations (range 34–46 × 10−6 in FAP13 and FAP483 LCLs, respectively) (P < 0.05 by Student's t-test in comparison to either BR77 or BR806).
KBrO3-induced mutagenesis in the heterozygous p.Tyr179Cys or p.Arg245His LCLs was similar to that in BR77 and BR806 wild-type LCLs (Fig. 4B).
Thus, the spontaneous mutator phenotype associated with homozygosity for variant MUTYH is further accentuated under conditions of oxidative stress. A single wild-type allele provides sufficient protection and oxidant-induced mutational load is not increased in heterozygous variant LCLs.
Cell cycle analysis
FACS analysis indicated that KBrO3 treatment causes in wild-type BR77 LCLs a decrease in the number of cells in the S phase and an accumulation in the G2 phase. In contrast, no perturbation of the cell cycle was apparent either at 24 or at 48 h post-KBrO3 treatment of FAP12 or FAP483 (Supplementary Material, Fig. S1). Since these data suggested that the accumulation of the cell population in G2 depends on the wild-type MUTYH processing of oxidatively damaged DNA, cell cycle progression following KBrO3 treatment was analyzed further in BrdU-labeled cells. One hour after a 30 min treatment with 60 mm KBrO3 S-phase cells were BrdU labeled for 3 and 6 h and analyzed by FACS. Oxidant treatment produced a decrease in the number of BrdU-positive BR77 cells at 3 h, which was partially relieved at 6 h, together with a clear accumulation of cells in the G2 phase of the cell cycle (Fig. 5). Untreated FAP12 and FAP483 both showed a lower number of cells in the S phase. In contrast to wild-type cells, KBrO3-treated MAP cell lines displayed a persistent decrease in the number of S-phase cells at both 3 and 6 h incubation with BrdU. The previously observed minor accumulation of cells in the G2 phase was also confirmed (Fig. 5). In addition, a further subdivision of the S phase in early (S1), middle (S2) and late (S3) allows identifying a slower transit of FAP12 and FAP483 LCLs particularly evident in the S1 phase (Supplementary Material, Fig. S2).
We can conclude that, in comparison to wild-type cells, MUTYH-defective cells display a more pronounced slowing down of the S phase consistent with a role of the wild-type MUTYH protein in helping oxidant-damaged cells progressing through the S phase.
Here, we have examined the properties of LCLs expressing two pathogenic missense MUTYH mutations associated with MAP. The p.Tyr179Cys variant has been widely studied. It affects adenine glycosylase and DNA binding to 8-oxodG:A mismatches (17–20). The less frequent p.Arg245His mutation has been identified in several studies across Europe (22–25). Substitution of the Arg245 with histidine (17) or leucine (26) also compromises the DNA binding and DNA glycosylase activities of MUTYH.
Tyr179 corresponds to Tyr88 in Bacillus stearothermophilus MutY. Tyr88 has a role in binding to the 8-oxodG containing strand. It intercalates into the DNA duplex between 8-oxodG and the 5′ nucleoside (27). Replacement of Tyr with Cys reduces both stacking interactions and hydrogen bonding required for efficient catalysis. The Arg245 residue does not directly contact the DNA substrate. It participates in an extended network of hydrogen bonds with other aminoacids (Arg241, Ala248, Thr291, Pro292, Ala308) and interacts with Cys290, one of the four cysteines bound to the [4Fe–4S] cluster (28). A detail of the region of interest is shown in Figure 6. Replacement of guanidine group of Arg with the His imidazole group diminishes the interaction with the [4Fe–4S] cluster (Fig. 6A and B). We observed that MUTYH protein levels are reduced in p.Arg245His but not in p.Tyr179Cys LCLs. The 70% decrease in MUTYH protein level that we observed in p.Arg245His LCLs may reflect protein instability resulting from increased [4Fe–4S] mobility.
A consistent phenotype of MUTYH-defective cells from several patients carrying the p.Tyr179Cys or p.Arg245His variants is an increased steady-state level of DNA 8-oxodG. It is important to stress that MUTYH contributes only indirectly to repair this oxidized base. Thus following the removal of inaccurate A from 8-oxodG:A mismatches, incorporation of cytosine opposite the lesion by the MUTYH-associated polymerase λ creates a new DNA substrate on which other repair factors can operate for direct removal of the oxidized base (4,29). Thus, the accumulation of DNA 8-oxodG in cells with no expression of the MUTYH protein or expressing a mutant protein must be the consequence of interfering with a proper removal pathway for 8-oxodG. This possibility is consistent with the extended persistence of DNA 8-oxodG in KBrO3-treated p.Tyr179Cys or p.Arg245His LCLs. We have observed that the increased DNA 8-oxodG levels in these cells are not correlated with higher steady-state ROS levels (unpublished observation). These findings indicate that DNA 8-oxodG accumulation in MAP patients does not exclusively reflect an enhanced oxidative stress. DNA 8-oxodG levels in LCLs derived from patients with other MUTYH missense mutations (p.Gly396Asp/p.Arg245Cys, p.Gly396Asp/p.Tyr179Cys, p.Gly396Asp/ p.Glu410Glyfs*43, p.Gly264Trpfs*7/p.Glu480del variants) are in the same order of magnitude (16). Together with the present data this suggests that the type of residue substitution does not significantly affect this phenotype.
A spontaneous mutator phenotype and hypermutability by oxidant treatment are also consistent characteristics of MUTYH variants. In view of the acknowledged protection that MUTYH affords against G > T transversions derived from 8-oxodG:A mismatches, the spontaneous mutator phenotype associated with MUTYH inactivation is not surprising. More importantly, our data indicate that the type of MUTYH mutation can determine the magnitude of the mutator phenotype. We observed a significant difference in the degree of the mutator effect in p.Tyr179Cys and p.Arg245His LCLs. This difference may reflect a number of factors including differences in the expression of the variant proteins and altered interactions with proteins involved in DNA replication, DNA repair or in DNA damage sensing (30–32). Additionally, in the Arg24His variant, in which specific conformational changes involve the [4Fe–4S] region, alterations in a DNA-mediated charge transfer process that may regulate lesion recognition and MUTYH turnover (33) may contribute.
The p.Tyr179Cys and p.Arg245His LCLs share a similar hypermutability by KBrO3. Thus, the subtle differences in spontaneous mutation rates are overwhelmed under oxidative stress conditions. Cell cycle progression following KBrO3 is also similarly affected in these two variants. We note that the fraction of S-phase cells is lower in untreated MUTYH variant than in wild-type LCLs. This suggests that a fully effective MUTYH protein is required for efficient transit through the cell cycle in the face of both endogenous and exogenous oxidative DNA damage.
Measurements of spontaneous mutagenesis in LCLs homozygotes and heterozygotes for the p.Arg245His variant identified a clear gene dosage effect. In contrast, heterozygous p.Tyr179Cys was dominant indicating that the mutant protein interferes with the function of the concomitantly expressed wild-type MUTYH. Nevertheless, mutation induction under conditions of extreme oxidative stress is similar in heterozygous and wild-type cell lines. It is possible that this depends on the upregulation of the wild-type allele(s) following oxidant exposures (29,34).
The mutator phenotype of MUTYH-defective cells is modest. Whether the 3- or 6-fold increase in spontaneous mutant frequencies observed in p.Tyr179Cys and p.Arg245His LCLs directly translates into a differential risk of CRC is difficult to envisage. Clinical data suggest that the p.Tyr179Cys variant is associated with a severe colorectal disease (35). The low number of p.Arg245His MAP patients described in the literature, most of which are compound heterozygotes (22 entries in LOVD database), does not allow to clarify this point (22–25,36). However, our findings are in agreement with the hypothesis that this variant is associated with a severe phenotype (24,25). In addition, our two p.Arg245His homozygous patients presented with important colorectal disease manifestation (high polyp burden/CRC) (Table 1).
Carriers of monoallelic MUTYH mutations may represent between 1.4 and 2% in the general population (8,9,12). Although heterozygous carriers of MUTYH variants are considered to have an increased risk of developing CRC, meta-analysis identifies only a small increased risk (8–12). Because of the prevalence of the p.Tyr179Cys mutation in the cohorts of MAP patients, the risk associated with this variant may skew the analysis. Indeed, the pooled odds ratio for the association between a monoallelic mutation and CRC was 1.35 (95% CI = 0.99–1.85) for the p.Tyr179Cys, while the p.Gly396Asp variant showed a lower value of 1.06 (95% CI = 0.88–1.28) (12). It is tempting to speculate that the mutator phenotype associated with expression of the p.Tyr179Cys (as well as of the p.Arg245His) variants might underlie, at least in part, this risk.
MATERIALS AND METHODS
The present study was carried out on blood cells from six patients affected by MAP, who were identified following genetic counseling and genetic testing at two National Cancer Institutes in Italy, the Centro di Riferimento Oncologico, Aviano and AOU San Martino-IST, Genova. Four patients were p.Tyr179Cys homozygotes and two patients were p.Arg245His homozygotes (identical-by-descent mutant alleles in one case due to consanguinity). Four heterozygous family members and three control wild-type individuals were also investigated. Mutations were named according to the RefSeq NM_001128425.1. Following the procedures dictated by the local Ethical Committees, each individual provided within the counseling session an informed consent to the use of blood samples for research purposes.
LCLs cultures and MUTYH expression
Methods for LCLs establishment and MUTYH protein analyses were essentially as previously described (15). LCLs were grown in RPMI supplemented with 15% fetal bovine serum and 1% penicillin–streptomycin (standard medium) at 37°C and 5% CO2. For oxidative treatment, cells were suspended in PBS 1×, 20 mm HEPES at a concentration of 106 cells/ml and treated for 30 min at 37°C with 20 or 60 mm KBrO3. After treatment, cells were washed once in PBS 1× and suspended in standard medium at a concentration of 5 × 105 cells/ml for further analyses.
To quantify the MUTYH protein by western blotting, cells were lysed in 50 mm Tris–HCl (pH 7.5), 150 mm NaCl, 1% Triton X-100, 1 mm EDTA and proteins (100 μg) were loaded on 12% SDS–PAGE gels. A mouse monoclonal antibody specific for the C-terminal epitope of the human MUTYH protein (Ab 55551; Abcam, Cambridge, UK) and the HRP-conjugated secondary antibody (Perkin Elmer, Waltham, MA, USA) were used at 1:500 and 1:3000 dilutions, respectively. MUTYH signals were normalized to the β-tubulin protein.
Determination of DNA 8-oxodG
8-OxodG was measured by high-performance liquid chromatography with electrochemical detection (HPLC-EC) as described in (37). To determine 8-oxodG repair kinetics, cells were treated for 30 min with 20 mm KBrO3 and at each time point, 3 × 106 cells were removed from dishes, DNA was extracted and 8-oxodG levels measured as previously described.
Determination of mutations at the PIG-A gene
Mutant frequency was measured as previously described (15). Briefly, LCLs were stained on ice with a mixture of phycoerythrin mouse anti-human antibodies specific for three GPI-linked proteins, anti-CD48 (BD Pharmingen, BD Biosciences, San Jose, CA, USA), anti-CD55 (BD Pharmingen) and anti-CD59 (AbD Serotec, Oxford, UK). Staining with the fluorescein isothiocyanate (FITC) mouse anti-human leukocyte antigen-DR (HLA-DR) (BD Pharmingen), a non-GPI-anchored transmembrane protein, was used as a lineage marker. Mutant cells (GPI-negative) are detected as HLA-DR-positive and CD48-, CD55- and CD59-negative events. The mutant frequency is calculated as the number of GPI-negative cells divided by the number of total events analyzed (20). FACSAria (BD Biosciences) was used for this analysis.
Cell cycle analysis
After KBrO3 treatment, 1–3 × 106 cells were centrifuged, washed once in sample buffer (PBS 1×, glucose 1 g/L) and suspended by vortexing and slowly adding 1 ml of ice-cold 70% ethanol drop-by-drop to the pellet. Cells were fixed O/N at 4°C, then vortexed for few seconds and centrifuged. Pellets suspended in 1 ml of sample buffer, 50 μg/ml propidium iodide (PI), 10 μg/ml RNAse A were incubated for 30′ at room temperature (RT) and analyzed by flow cytometry. For BrdU (Sigma Chemical Co., St. Louis, MO, USA) incorporation assays, cells were treated with KBrO3 as described above and, following 1 h recovery in standard culture conditions, 30 μm BrdU was added to the medium. Three and 6 h later, 106 cells were harvested, the pellet was suspended in 300 μl PBS 1× with an addition of 700 μl of ice-cold ethanol and cells were fixed at least for 1 h at 4°C. Ethanol was eliminated by centrifugation, cells were washed twice in PBS 1× and incubated in 500 μl of 2 m HCl, 0.5% Triton X-100 solution for 30′ at RT. After centrifugation, cells were incubated in 500 μl of 0.1 m sodium tetraborate (pH 8.5) for 2′, washed in PBS 1×, 1% BSA and resuspended in PBS 1×, 1% BSA, 0.5% Tween-20 and 1 μg of Anti-BrdU-FITC antibody (eBioscience, San Diego, CA, USA) for 1 h at RT. After centrifugation, cells were washed in PBS 1×, 1% BSA, incubated in 500 μl PBS 1×, 20 μg/ml PI, 10 μg/ml RNAse A for 30′ at RT in the dark and then analyzed with FACScan (BD Biosciences).
The effect of mutations in each cell line was compared with the wild-type BR77 by Student's t-test for MUTYH expression, 8-oxodG levels and spontaneous PIG-A mutant frequency. Analysis of mutant frequency was also performed by one-way ANOVA having as groups the wild-type LCLs, the homozygous and heterozygous variant LCLs.
This work was supported by Associazione Italiana Ricerca sul Cancro (grant number 11755), and by Ministry of Health, Project “Malattie Rare” to M.B., F.M., L.V. and A.V.
We thank Michele Quaia (CRO Aviano) for MUTYH genetic testing, Mara Fornasarig (CRO Aviano), Daniela Barana and Cristina Oliani (ULSS 5, Montecchio Maggiore) for contributing blood samples and clinical data, Alessandro Giuliani for statistical advice and Pietro Pichierri for data discussion.
Conflict of Interest statement. None declared.