Folate deficiency is associated with the formation of complex nuclear anomalies in the cytokinesis-block micronucleus cytome assay
Abstract
Chromosomal instability (CIN) is an important hallmark to oncogenesis and can be diagnosed morphologically by the presence of nuclear anomalies such as micronuclei (MN), nucleoplasmic bridges (NPBs), and nuclear buds (NBuds). We have identified additional nuclear anomalies formed under folate-deficient conditions, defined as “fused” nuclei (FUS), “circular” nuclei (CIR), and “horse-shoe” nuclei (HS) and investigated their suitability for inclusion as additional CIN biomarkers in the lymphocyte cytokinesis-block micronucleus cytome (CBMN-Cyt) assay. Although the morphological appearance of FUS, CIR, and HS suggested an origin from multiple NPB in the fusion region between the two nuclei, the very low frequency of dicentric chromosomes in metaphase spreads from these cultures did not support this model. Fluorescence in situ hybridization (FISH) analysis of cytokinesis-blocked binucleated (BN) cells with peptide nucleic acid probes for telomeres and centromeres (PNA–FISH) revealed a high proportion of fusion regions contained both centromeric and telomeric DNA. This suggests that folate deficiency may disrupt the process of sister chromatid separation and chromosome segregation during mitosis. It was concluded that the FUS, CIR, and HS morphologies represent promising biomarkers of CIN that are sensitive to folate deficiency, and further validation and investigation of the mechanisms responsible for their formation is warranted. Mol. Mutagen. 2012. © 2012 Wiley Periodicals, Inc.
INTRODUCTION
Folate is a micronutrient that is essential for maintaining genomic stability. This is because of its two roles; first, as the methyl donor in the formation of deoxythymidine triphosphate from deoxyuridine monophosphate for DNA synthesis and repair and second, in cytosine methylation. The second is achieved in concert with vitamin B12 via the synthesis of methionine and the common methyl donor S-adenosyl methionine [Fenech,2001]. Dividing human peripheral blood lymphocytes (PBL) show an inverse dose-dependent relationship between concentration of folic acid (FA) in culture medium (within the physiological range) and level of chromosomal damage biomarkers, as measured using the cytokinesis-block micronucleus cytome (CBMN-Cyt) assay [Crott et al.,2001; Kimura et al.,2004; Beetstra et al.,2005; Lindberg et al.,2007]. These biomarkers, including uracil incorporation into DNA, are usually minimized when cells are maintained in culture medium with a FA concentration above 120 nM [Duthie and Hawdon,1998; Crott et al.,2001; Fenech and Crott,2002; Kimura et al.,2004].
The CBMN-Cyt assay is an established, comprehensive method for measuring chromosomal instability (CIN), cell death, and cytostasis in cells exposed to a variety of environmental conditions, including altered micronutrient status [Fenech,2007]. In this assay, cytostasis and cell death measures include nuclear division index (NDI) and frequency of necrotic and/or apoptotic cells, respectively, while chromosomal stability and DNA damage are assessed by measuring frequencies of cytokinesis-blocked binucleated (BN) cells displaying one or more micronucleus (MN), nucleoplasmic bridge (NPB), or nuclear bud (NBud) [Fenech,2007].
We have often observed nuclear anomalies in cells grown under folate deficient conditions, in addition to those so far tested within the CBMN-Cyt assay. However, they had not been formally defined or validated as useful biomarkers of genome stability nor had their underlying mechanism been examined. Based on their appearance, these abnormal nuclear morphologies were designated as “fused” (FUS), “circular” (CIR), and “horse-shoe” (HS) cells, possibly arising from fusions between daughter nuclei in cytokinesis-blocked cells, perhaps due to multiple NPB formation. FUS, CIR, and HS were particularly evident in the cultured lymphocytes of a specific individual, although they were also observed in cells from other donors. Accordingly, this study was conducted using PBL isolated from this individual and grown in culture for 21 days at FA concentrations within the physiological range (20, 60, and 180 nM). At weekly intervals, measurements were made of cell growth, cell viability, NDI, apoptosis and necrosis, and conventional biomarkers of CIN, as defined in the CBMN-Cyt assay, were scored. Scoring criteria were defined for the three new categories of abnormal nuclear morphologies (FUS, CIR, and HS), so that direct comparisons of the frequencies of each could be made against the standard biomarkers of the CBMN-Cyt assay. Furthermore, fluorescent in situ hybridization (FISH) analysis using centromeric and telomeric peptide nucleic acid (PNA) probes was conducted on metaphase preparations and cytochalasin-blocked BN cells in order to gain insights into the mechanism(s) responsible for these abnormalities. We specifically tested the following hypotheses: (i) FUS, CIR, and HS are induced by FA deficiency in a dose-related manner and (ii) NPB (or other apparent fusion structures) between the nuclei of abnormal BN cells involve telomeres predominantly, which would suggest a telomere end-fusion mechanism. The regular presence of telomeres in these structures would support the hypothesis that NPB and other internuclear fusion structures induced by folate deficiency are consequences of fusion between dysfunctional telomeres [Gisselsson and Hoglund,2005; Gonzalo et al.,2006; Blasco,2007].
MATERIALS AND METHODS
Blood Donor
A single venous blood sample was collected into 9-ml lithium-heparin Vacuette collection tubes from a fasted (overnight) 44-year-old Caucasian male volunteer. Plasma micronutrient and homocysteine (Hcy) concentrations were within the normal clinical reference intervals (RI) specified by the Institute of Medical and Veterinary Science (IMVS, South Australia); plasma folate was 33.9 nmol/l (RI 6.5–45.0), vitamin B12 was 359 pmol/l (RI 100–700), and Hcy was 6.2 μmol/l (RI 4.0–14.0 μmol/l). Allelic discrimination analysis determined the donor to be heterozygous for methylene tetrahydrofolate reductase, a key enzyme required for the synthesis of 5-methyl tetrahydrofolate, the methyl donor used to convert Hcy to methionine. The study was approved by the Human Research Ethics Committee of CSIRO Human Nutrition, Adelaide, Australia, and informed consent was obtained.
Culture Medium
Medium-containing either 20, 60, or 180 nM FA was prepared using FA-deficient medium RPMI 1640 (FDM; Sigma, R1145), supplemented with 10% fetal bovine serum (Thermo). The following were added to give final dilutions (v/v): 1% penicillin/streptomycin (Sigma, P4458) and 1% sodium pyruvate (Sigma, S8636). Interleukin-2 (5%) (Roche Diagnostics, 11147528001) and L-glutamine (1%) (Sigma, G7513) were added to medium immediately before use. To prepare FDM, the 10× stock solution was diluted 1:10 with sterile milli-Q water, and NaHCO3 (Sigma) was added as a sterile-filtered solution to a final concentration of 2 mg/l (as per manufacturer's instruction). To achieve the required FA concentration for each condition, the appropriate proportion of FDM was substituted with FA-replete RPMI 1640 (Sigma, R0883). After measurement of FA concentration by Abbott Architect Analyser at the Institute for Medical and Veterinary Sciences (IMVS Pathology, Adelaide) (and adjustment as required), each batch of medium was stored at −20°C and thawed at 4°C before use. Phytohemagglutinin (PHA; Oxoid; 1.3 μl/ml of 22.5 mg/ml) was added directly to each culture flask at day 0 to stimulate division of lymphocytes.
Cell Culture for CBMN-Cyt and PNA–FISH Assays
Lymphocytes were isolated by Ficoll-Paque™ gradient separation (Amersham Biosciences) and transferred into prewarmed complete medium containing either 20, 60, or 180 nM FA. All cultures were maintained for 21 days as 20 ml duplicates in upright 25-cm2 vented-cap culture flasks (Becton Dickinson), incubated at 37°C under humidified atmosphere with 5% CO2. Cells from each flask were harvested every 3.5 days by centrifugation and returned to culture in fresh medium incorporating 5% conditioned medium (spent medium obtained from the previous flask following centrifugation; Fig. 1). Total cell counts using a Coulter Counter (Beckman Coulter) and viability estimates using trypan blue (Sigma, T8154) exclusion were performed at days 0, 3, 7, 10, 14, 17, and 21 to establish and maintain a seeding density of 0.5 × 106 viable cells/ml (Fig. 1). Estimated growth over the complete culture period in each condition was calculated serially by counting the numbers of viable cells in the culture at each time of harvesting and extrapolating to numbers that would have been expected if all cells had been reseeded at each preceding split.
Experimental design for CBMN-Cyt assay, PNA–FISH, and metaphase-FISH, detailing when cell counts, cell viability estimates, and assays were conducted. For CBMN-Cyt assay and PNA–FISH, bulk cultures of PBL were stimulated with PHA at day 0 and maintained with biweekly changes of medium-containing IL-2 for 21 days. At days 7, 14, and 21, subcultures were established in 24-well plates, and after a 2 h incubation period, they were treated with cytochalasin B to block cytokinesis. Twenty-four hours later, cells were harvested onto slides by cytocentrifugation and fixed. Some slides were stained using Diff-Quik before scoring for the CBMN-Cyt assay, while others were stored at −20°C for PNA–FISH analysis. For metaphase preparations, 20 ml cultures containing 1,200 μl whole blood were established at day 0. Fresh medium (10 ml) was replaced at day 3. At day 7, 10 ml medium was removed, and cells were treated with colcemid for 4 h to impose mitotic arrest, followed by hypotonic treatment. Metaphase spreads were then prepared as described in the text.
CBMN-Cyt Assay
Chromosomal damage in PBL was measured using the CBMN-Cyt assay, as described in detail previously [Fenech,2007]. In brief, at each sampling time (days 7, 14, and 21), duplicate 500 μl subcultures were established in a 24-well plate from the main cultures, at a concentration of 0.5 × 106 viable cells/ml (Fig. 1). Following a 2-hr incubation period at 37°C in 5% CO2, cytochalasin B (4.5 μg/ml; Sigma) was added to block cytokinesis, and cells were harvested after a further 24 hr by cytocentrifugation using a Shandon Cytospin Cytocentrifuge (Shandon Scientific). Slides were air-dried for 10 min and some were fixed and stained using the commercial kit “Diff Quik” (Lab Aids). Other slides, for binucleate-fluorescence in situ hybridization (PNA–FISH) analysis, were fixed, air-dried for 10 min, and stored at −20°C. All CBMN-Cyt assay slides were scored blind by one person, using criteria described by Fenech [2007]. Cells cultured in 20 nM FA had low NDI and viability at days 14 and 21 resulting in reduced numbers of binucleates for analysis purposes. For this reason, 500 BN cells per duplicate culture (total 1,000 BN cells per treatment per sample point) were scored for frequency of BN cells containing ≥1 MN, ≥1 NPB, or ≥1 nuclear bud (NBud). Five hundred cells were also scored per duplicate culture for the percentage of cells displaying morphologies indicative of necrosis or apoptosis. NDI was calculated as NDI = (M1 + 2M2 + 3M ≥3)/N, where M1, M2, and M ≥ 3 represent the number of cells with 1, 2, or ≥3 nuclei, and N is the total number of viable cells scored (i.e., excluding necrotic and apoptotic cells).
Scoring Criteria for Novel Nuclear Anomalies
- 1.
“Fused nuclei” (FUS): The two main nuclei of FUS cells appear to be connected by multiple nuclear strands and/or structures. These are either too numerous (>3) and/or too broad, and/or too close together, to be clearly discernible as conventional NPB (width < one-fourth diameter of nucleus) for scoring purposes. The connecting nuclear material in the fusion region between the two nuclei may appear as a single, very wide structure (greater than one-fourth the nuclear diameter), or a combination of both wide and narrow fusion structures (Fig. 2i);
- 2.
“Circular” (CIR): The nucleus of a CIR cell has an “O-shaped” nuclear structure. The hole at the center contains cytoplasm and may vary considerably in size but is typically no smaller than one-twentieth the area of the nucleus (Fig. 2ii);
- 3.
“Horse-shoe” (HS): Nuclei of a HS cell appear to be fused at one end by a wide, continuous nuclear structure, resulting in a single “horse-shoe”-shaped nucleus. The indentation in the “horse-shoe” is typically greater than one-fourth the diameter of a single nucleus, and the open ends should not be touching (Fig. 2iii).
Typical appearances of the additional nuclear anomalies observed in the CBMN-Cyt assay when PBL have been grown in complete medium-containing 20 nM FA. (For the sake of simplicity, the conceptual diagrams only show a representative fraction of chromosomes). The nuclear anomalies (i) “fused” (FUS), (ii) “circular” (CIR), or (iii) “horse-shoe” (HS) nuclei. In the proposed models, (i) FUS originate as a result of multiple nuclear strands occurring uniformly or centrally between the nuclei of a binucleated (BN) cell; (ii) CIR originate as a result of multiple nuclear strands occurring on opposite sides between the nuclei of a BN cell; and (iii) HS originate as a result of multiple nuclear strands occurring only on one side between the nuclei of a BN cell. In each of these models, the total combined width of the connections between the nuclei in a BN cell is typically larger than one-fourth of the nuclear diameter—the maximum width of a conventional NPB, as defined within the CBMN-Cyt assay [Fenech,2007]. This criterion, together with the additional descriptors detailed earlier, are used to distinguish these novel structures from conventional BN cells with single NPBs with a width less than one-fourth of the nucleus. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary. com.]
Metaphase Chromosome Preparations
Whole blood cultures were established using 1,200 μl blood in 20 ml of prewarmed complete medium containing either 20 or 180 nM FA. Duplicate cultures were maintained in upright 25-cm2 vented-cap culture flasks (Becton Dickinson, Australia) and incubated at 37°C in a humidified atmosphere with 5% CO2 for 7 days. At day 4, 10 ml of the original medium was replaced with fresh medium. At day 7, 10 ml of medium was removed, and cells were resuspended in the remaining 10 ml. Colcemid (Gibco; 0.15 μg/ml) was added into the 10 ml cultures, and cultures were incubated for a further 4 hr to impose mitotic arrest. The cells were then harvested, incubated in a hypotonic solution (0.075 M KCl) at 37°C for 15 min, fixed in Carnoy's solution (3:1 methanol:acetic acid), and stored at −20°C before use. Fixed cell suspensions were dropped onto dry slides at room temperature (22°C), heated on a hotplate at 60°C for 1 min, and “aged'' at room temperature for 24–48 hr before telomere and centromere detection using FISH with PNA probes [Newman et al.,2008].
Fluorescence In Situ Hybridization With Peptide Nucleic Acid Probes
Peptide nucleic acid-fluorescence in situ hybridization (PNA–FISH) was carried out using Cy3-conjugated pan-telomeric and FAM (5-carboxyfluorescein)-conjugated pan-centromeric PNA probes. Slide pretreatment consisted of an initial wash in phosphate-buffered saline (PBS, pH 7.4), fixation in formaldehyde (4%) in PBS (2 min), and a further three washes in PBS. The slides were then treated with acidified pepsin solution (1 mg/ml) containing 0.01 M HCl (pH 2) for 2 min at 37°C. Washing steps and formaldehyde fixation were then repeated. Slides were dehydrated with sequential washes in 70, 90, and 100% ethanol for 5 min each at RT, followed by air drying. A total of 20 μl hybridization mixture containing 70% deionized formamide, 0.5 μg/ml Cy-3 PNA-telomere probes (Panagene, Korea), 3.0 μg/ml FAM PNA-centromere probes (Panagene, Korea), 0.25% blocking reagent in maleic acid buffer, 10 mM Tris, and 5% 1 M MgCl2.6H20 was applied to each slide on 22 × 60 mm coverslips. The slides were then incubated at 80°C for 3 min in a hybridizer (Dako Cytomation) to denature DNA. Hybridization was carried out for 2 hr at RT in a humidified chamber, and coverslips were removed by gentle immersion in wash solution I at 37°C (70% formamide, 30% H2O, 0.1% bovine serum albumin, and 0.01 M Tris buffer). Two 15 min washes were carried out in wash solution I, followed by three washes for 5 min in wash solution II (1 × Tris-sodium chloride buffer and 0.1% Tween-20). Slides were dehydrated in ethanol series, air dried, and counterstained with 0.0375 μg/ml 4′-6-diamidino-2-phenylindole (DAPI) in Vectashield mounting medium (Vector Laboratories, USA) by incubation at RT for 30 min before storage at 4°C for a minimum of 12 hr before image capture and analysis [Newman et al.,2008].
Scoring of PNA–FISH Signals in Metaphase Chromosomes and in Fusion Regions of FUS, CIR, and HS Cells
Metaphase–fluorescence in situhybridization (FISH) and Cyt-B-blocked BN [peptide nucleic acid (PNA)–FISH] images were captured on a Zeiss Axioplan 2 imaging fluorescence microscope (Carl Zeiss) and analyzed using the in situ imaging software, Isis Analysis (Metasystems GmbH). Frequency of dicentric chromosomes in metaphases was scored manually using metaphase–FISH. PNA–FISH was used to determine the presence or absence of telomeric and/or centromeric DNA within the nuclear fusion regions connecting the two main nuclei of FUS, CIR, or HS cells.
- (i)
the nuclei must be connected by one or more continuous DNA-containing structure (i.e., nuclear strands within fusion regions), as indicated by DAPI staining characteristics similar to the main nuclei;
- (ii)
single and/or multiple nuclear strands may be present in the fusion region joining the nuclei of a cell, the width of which may vary considerably. Nuclear fusion strands, individually or in combination, typically exceed one-fourth the diameter of the main nuclei. Often, the fusion regions show indentations, such that the boundaries of the two nuclei can be clearly differentiated (Fig. 3);
- (iii)
for scoring purposes, the fusion regions in BN cells (including those with “circular'' nuclei) are the DNA-containing regions joining the two apparent nuclei (the boundaries of which are indicated by the pattern of DAPI staining) in a BN cell (Fig. 3); and
- (iv)
the presence of telomeric and/or centromeric DNA is recorded when the fluorescent signal is located within the fusion region, or immediately on the fusion region/nuclear boundary (Fig. 3).
Procedure and criteria used to identify the fusion regions connecting the nuclei of (A) FUS, (B) CIR, and (C) HS cells, as a preliminary to determining the presence of centromeric and/or telomeric DNA using PNA–FISH. Initially, the nuclear boundaries connected by the nuclear strands were identified by the continuous and uniform DNA staining of nuclei (dotted lines indicate these boundaries). Scoring for the presence of fluorescent PNA–FISH signals was then conducted, recording only those signals present within the defined fusion region (indicated by arrows) or immediately on the fusion/nuclear boundary. See text for detail. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
Statistical Analyses
Data are shown as mean ± SD for all figures and tables. Two-way analysis of variance (ANOVA) was used to compare the main effect for time, FA concentration in medium, and the interactive effect of these two factors. Pair-wise comparison of significance was determined using Bonferroni posthoc test. Significance was accepted at P < 0.05. Residuals for all data sets, or log transformed data sets, exhibited normal (Gaussian) distribution. Two-tailed Pearson correlations were used to analyze relationships between two variables. All statistical analyses were performed using Graphpad PRISM 4.0 (GraphPad, San Diego, CA) or SPSS for Windows 17.0 (SPSS, Chicago, IL).
RESULTS
Cell Growth, Cell Viability, NDI, Necrosis, and Apoptosis in PBL Cultured Long-Term (21 days) in Folate Replete and Deplete Medium
Viable cell numbers increased over the full 21 days in the 180 nM FA culture, whereas cells in 60 nM FA grew steadily through to day 14 but failed to increase in number beyond this point. Cell growth increased in 20 nM FA cultures to day 7, remained stable between days 7 and 17, and then declined from days 17 to 21 (Fig. 4A). FA concentration explained 16.6% of observed variance in cell growth over the 21-day study (P < 0.0001). These differences in cell growth were reflected by effects of FA concentrations on cell viability, with 31.7% of the variance in cell viability being attributable to FA status (P < 0.0001). A reduction in viability was observed in the 20 nM FA culture from >80% in the first 7 days, to 23% at day 21. In contrast, cells cultured in 180 nM FA showed excellent viability (>90%) throughout the study (Fig. 4B).
The effect of folic acid (FA) status on PBL cultured in medium containing 20, 60, or 180 nM FA over 21 days in vitro. A: Cell number; (B) cell viability (%); (C) nuclear division index (NDI); (D) necrotic cells (%); and (E) apoptotic cells (%). (N = 2 for each treatment at each time point. Error bars indicate SD. Points not sharing the same letter at each time point differ significantly, as measured by Bonferroni posthoc test).
There was a FA dose-dependent and time-dependent decline in NDI, with 40.3% of variance attributable to FA status (P < 0.0001; Fig. 4C). The percentage of cells with morphological evidence of necrosis showed an inverse relationship to FA concentration, with FA accounting for 49.4% of variance (P < 0.0001; Fig. 4D). Similarly, an inverse relationship was observed between apoptotic cells (%) and FA status (21.7% of variance due to FA, P = 0.02; Fig. 4E).
Frequency of BN Cells Displaying Conventional Biomarkers of Chromosome Damage (MN, NPB, or NBuds) in the CBMN-Cyt Assay
The frequencies of BN cells displaying one or more of the chromosomal damage biomarkers (MN, NPB, and NBuds) scored using the CBMN-Cyt assay standard criteria [Fenech,2007] increased significantly with decreasing FA concentration in medium (Table I). At day 21, the frequency of MN in the 20 nM treatment was sevenfold greater than in cells cultured in 180 nM FA. Cells grown in 60 nM FA displayed over fivefold greater frequency of MN compared to those grown in 180 nM FA for 21 days (Table I; i). FA concentration in medium accounted for 54.8% of the variance in the frequency of MN (P < 0.0001).
| Category | Sample point | [FA] in medium (nmol/l) | Analysis by two-way ANOVA | |||||
|---|---|---|---|---|---|---|---|---|
| 20 nM | 60 nM | 180 nM | Source of variation | % of total variance | P value | |||
| (i) | MN | Day 7 | 21.5 ± 2.1a | 5.0 ± 2.8b | 8.5 ± 0.7b | FA: | 54.82 | <0.0001 |
| Day 14 | 20.0 ± 4.2a | 16.5 ± 3.5a,b | 6.0 ± 1.4b | Time: | 19.24 | 0.002 | ||
| Day 21 | 35.0 ± 8.5a | 26.5 ± 0.7a | 5.0 ± 3.5b | Int: | 19.21 | 0.010 | ||
| (ii) | NPB | Day 7 | 11.5 ± 2.1 | 6.5 ± 0.7 | 4.5 ± 0.7 | FA: | 45.79 | 0.001 |
| Day 14 | 17.0 ± 2.8 | 5.5 ± 3.5 | 10.5 ± 12.0 | Time: | 17.89 | 0.024 | ||
| Day 21 | 33.5 ± 4.9a | 12.5 ± 0.7b | 5.0 ± 4.2b | Int: | 22.54 | 0.050 | ||
| (iii) | NBud | Day 7 | 5.0 ± 2.8 | 3.0 ± 4.2 | 0.5 ± 0.7 | FA: | 21.56 | 0.001 |
| Day 14 | 9.5 ± 3.5 | 6.0 ± 4.2 | 5.0 ± 2.8 | Time: | 43.76 | 0.0001 | ||
| Day 21 | 66.5 ± 19.1a | 29.0 ± 8.5b | 5.5 ± 2.1c | Int: | 28.11 | 0.003 | ||
| (iv) | Any biomarker | Day 7 | 38.0 ± 2.8 | 14.5 ± 6.4 | 13.5 ± 0.7 | FA: | 38.04 | <0.0001 |
| Day 14 | 46.5 ± 10.6 | 28.0 ± 11.3 | 21.5 ± 13.4 | Time: | 34.24 | <0.0001 | ||
| Day 21 | 135.0 ± 22.6a | 68.0 ± 8.5b | 16.0 ± 9.9c | Int: | 23.22 | 0.001 | ||
- Frequency of BN cells that contained one or more (i) micronucleus (MN), (ii) nucleoplasmic bridge (NPB), (iii) nuclear bud (NBud), and (iv) any damage biomarker (MN and/or NPB and/or NBud). Results shown are mean ± SD per 500 BN cells. N = 2. Data not sharing the same superscript letter within each time point differ significantly from each other, as measured by the Bonferroni posthoc test. Int, interaction of FA with time).
The frequency of NPB at day 21 increased sevenfold in the 20 nM culture and threefold in the 60 nM culture, when compared with the FA replete (180 nM) culture; 45.8% of the NPB frequency variance was due to FA concentration in medium (P = 0.001; Table I; ii). The difference between treatments for the frequency of NBud was also significant, with 21.6% of variance attributable to FA status (P = 0.001; Table I; iii). After 21 days, the frequency of NBud in the 20 and 60 nM cultures showed 12-fold and 5-fold increases, respectively, relative to 180 nM cultures (Table I; iii).
By day 21, there was an eightfold increase in frequency of BN cells exhibiting one or more of the above three classical DNA damage markers in the cells grown in 20 nM FA, compared to those grown in 180 nM FA; 38% of observed variance was attributable to FA treatment (P < 0.0001), 34.2% to time (P < 0.0001), and 23.2% of the variance was explained by the interaction of FA concentration with time (P = 0.001; Table I; iv).
Frequency of BN Cells Displaying Novel Nuclear Morphologies
Table II provides frequency data for the novel additional nuclear morphologies, “fused” (FUS), “circular” (CIR), or “horse-shoe” (HS), for each FA concentration at days 7, 14, and 21 of culture. The proportions of BN cells exhibiting these morphologies were increased only at the lowest FA concentration (20 nM). At this concentration of FA, the frequency of each morphology in BN cells was greatest at day 7 and reduced thereafter (Table II; i–iii). The frequency of FUS at day 7 was 24-fold that observed in the 60 nM FA treatment and 40-fold that in cells cultured in 180 nM FA. FA status accounted for 22.7% of this variance (ANOVA, P = 0.005). For CIR morphologies, the frequency at day 7 in the 20 nM culture was 14-fold greater than in the 60 nM culture and 19 times greater than that observed in the 180 nM FA condition; FA concentration explained 26.3% of this variance (P < 0.0001). Similar differences were also recorded for HS morphologies, with FA concentration accounting for 17% of observed variance (P = 0.003); at day 7, the frequency of HS in the 20 nM FA condition was ninefold that observed in 60 nM FA and 22-fold greater than in the 180 nM treatment.
| Category | Sample point | [FA] in medium (nmol/l) | Analysis by two-way ANOVA | |||||
|---|---|---|---|---|---|---|---|---|
| 20 nM | 60 nM | 180 nM | Source of variation | % of total variance | P value | |||
| (i) | FUS | Day 7 | 241.5 ± 77.1a | 10 ± 1.4b | 6.0 ± 5.7b | FA: | 22.71 | 0.005 |
| Day 14 | 21.0 ± 8.5 | 22.5 ± 7.8 | 44.5 ± 4.9 | Time: | 10.16 | 0.05 | ||
| Day 21 | 46.5 ± 14.8 | 82.5 ± 61.5 | 13.0 ± 5.7 | Int: | 56.80 | 0.001 | ||
| (ii) | CIR | Day 7 | 86.5 ± 12.0a | 6.0 ± 0.0b | 4.5 ± 0.7b | FA: | 26.30 | <0.0001 |
| Day 14 | 12.0 ± 9.9 | 12.5 ± 4.9 | 7.0 ± 0.0 | Time: | 19.79 | <0.0001 | ||
| Day 21 | 6.0 ± 0.0 | 7.5 ± 3.5 | 7.5 ± 6.4 | Int: | 51.11 | <0.0001 | ||
| (iii) | HS | Day 7 | 44 ± 8.5a | 5.0 ± 5.7b | 2.0 ± 2.8b | FA: | 17.08 | 0.003 |
| Day 14 | 3.0 ± 0.0 | 13.0 ± 5.7 | 5.0 ± 7.1 | Time: | 22.07 | 0.001 | ||
| Day 21 | 1.0 ± 1.4 | 3.5 ± 2.1 | 1.0 ± 1.4 | Int: | 54.58 | 0.0002 | ||
- Data represent mean frequency ± SD per 500 BN cells for each treatment at each time point. N = 2. Data not sharing the same superscript letter within each time point differ significantly from each other, as measured by Bonferroni posthoc test. Int, interaction of FA with time.
Correlations Between Validated Biomarkers of Chromosomal Damage and Novel Nuclear Morphologies
Frequencies of novel morphologies were correlated against those of standard (validated) biomarkers over the full 21-day time course (n = 18). Significant, positive correlations were recorded between frequencies of standard biomarkers; MN and NPB (r = 0.72, P = 0.001), MN and NBud (r = 0.82, P < 0.001) and between NPB and NBud (r = 0.81, P < 0.001). Novel morphologies correlated positively and highly significantly with each other (all P < 0.001); FUS versus CIR (r = 0.86), FUS versus HS (r = 0.79), and CIR versus HS (r = 0.95), but were not significantly associated with any of the standard biomarkers when data was analyzed across all time points.
Given the high frequency of novel morphologies at day 7, data for this time point were also analyzed in isolation (n = 6). The novel morphologies were significantly and positively associated with each other; FUS versus CIR (r = 0.92, P = 0.01), FUS versus HS (r = 0.89, P = 0.02) and HS versus CIR (r = 0.99, P < 0.001), as well as with conventional biomarkers, MN and NPB; MN versus FUS (r = 0.89, P = 0.02), CIR (r = 0.96, P = 0.002), and HS (r = 0.96, P = 0.003); and NPB versus FUS (r = 0.96, P = 0.002), CIR (r = 0.88, P = 0.02), and HS (r = 0.85, P = 0.03).
Fluorescence In Situ Hybridization Analysis of Fusion Regions Between Nuclei of FUS, CIR, and HS Cells
Metaphase–fluorescence in situ hybridization (FISH) was used to detect the presence of dicentric chromosomes. This analysis was undertaken to test the hypothesis that the high frequency of nuclear anomalies observed at day 7 in the 20 nM condition had arisen by chromosome end-to-end fusions. Of the 153 metaphase spreads scored, only two contained a dicentric chromosome (1.3%). No telomeric signals were detected between the centromeres of these dicentric chromosomes. This low frequency of dicentric chromosomes is consistent with the frequency of NPB scored using the CBMN-Cyt assay, where 11.5 ± 2.1 NPB were detected in 500 cells at the same time point (a frequency of 2.3%; Table I). In contrast, the frequency of FUS, CIR, and HS was 48, 17, and 9%, respectively, at day 7 (Table II). Although dicentric chromosomes are classically associated with NPBs [Thomas et al.,2003], their formation cannot account for the FUS, CIR, and HS nuclear abnormalities observed in these cells.
Further analysis was conducted using PNA–FISH to determine whether centromeric and/or telomeric DNA is present within the fusion regions between nuclei of FUS, CIR, and HS cells. Scoring was conducted on 132 cells, comprising 101 FUS, 16 CIR, and 7 HS. Representative images are provided in Fig. 5 . In the fusion regions between the nuclei of these 132 cells with abnormal nuclear morphologies, 75% contained centromeric DNA (C+), with 57% displaying >1 centromere signal (>1C+). Ninety-three percent of cells displayed telomeric DNA in the fusion region (T+), with 83% containing >1 telomere signal (>1 T+). Fusion regions in 70% of these cells exhibited both centromere and telomere signals (C+T+), while only 5% displayed neither (C−T−).
Of fusion regions in FUS, 75% were C+, 97% were T+, and 72% were positive for both (C+T+) while only 3% contained neither signal (C−T−; Fig. 6). Of the fusion regions in CIR, 93% were centromere positive (C+), 93% were T+, and 86% contained signals for both (C+T+). In 44% of CIR cells, there was a predominance of centromeric signal clustered around the inner nuclear boundary. In the seven HS cells scored, 86% of fusion regions were C+, 86% were T+, with the same 86% being C+T+.
Representative examples of nuclear anomalies in binucleated (BN) PBL cultured for 7 days in FA-deficient (20 nM) medium. Black and white images indicate DAPI staining of DNA. The corresponding color image indicates the location of telomeric (red) and centromeric (green) DNA sequences visualized using PNA–FISH. Dashed lines indicate the nuclear boundaries on each side of the fusion region that were used for scoring purposes (630× magnification). A: (i) In this “fused” (FUS) cell, there are two narrow and one wide fusion between the two nuclei, the locations of which are indicated by arrows on the black and white image. The upper narrow bridge contains neither centromeric nor telomeric signal (C−T−). The central wide strand contains both centromeric and telomeric DNA (C+T+) within the fusion region, while the lower narrow strand contains no centromeric signal, but is positive for telomeric DNA (C−T+). A (ii) This FUS cell also displays three fusions between the nuclei, all of which are C+T+; however, the upper fusion contains only one, faint centromeric signal (red arrow), while the central and lower strands display multiple, strong centromeric signals (>1 C+). B: (i) and (ii) are examples of “circular” (CIR) morphologies, both of which are C+T+ within the fusion regions or at the immediate interface of the fusion region and the nuclear boundary. Of specific note is the predominance of centromeric DNA at the inner circular nuclear boundary in both these examples. Images (C) (i) and (ii) depict “horseshoe” (HS) cells, both of which are C+T+ in the fusion region defined between the nuclei for scoring purposes, with C(i) displaying a strong clustering of centromeric signal in the fusion region, while the C+ signals in C(ii) are more evenly distributed throughout the two nuclei, without obvious clustering in the fusion region. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
Frequency (%) of fusion regions between nuclei containing PNA–FISH signals for centromeric (C) or telomeric (T) DNA in cytochalasin-B blocked binucleated PBL after culture for 7 days in folic acid deficient (20 nM) medium. Cells with fusions between nuclei were classified as having “fused” (FUS), “circular” (CIR), or horse-shoe (HS) morphologies. A fusion region was scored as being C+ or T+ when the fluorescent signal was located within the region or immediately on the fusion region/nuclear boundary. (C+, centromere signal present; >1C+, 2 or more centromere signals present; T+, telomere signal present; >1T+, 2 or more telomere signals present; C+T+, fusion region positive for at least one of each signal; C−T−, neither signal present in fusion region).
DISCUSSION
The findings of this study confirm and extend previous research [Fenech and Crott,2002; Wang and Fenech,2003; Kimura et al.,2004; Beetstra et al.,2005; Fenech et al.,2005; Lindberg et al.,2007] by demonstrating a more extensive impact of FA insufficiency on chromosomal stability in PHA-activated PBL (T cells) maintained in division for 21 days in the presence of IL-2. Additional abnormal nuclear morphologies associated with FA deficiency have been identified and characterized. A key aim for this study was to define and test descriptive criteria for scoring the novel complex nuclear morphologies observed in BN cells, with a view to validating them as chromosomal damage biomarkers. The three main morphological abnormalities identified were named FUS, CIR, and HS. It was found that the frequency of cells meeting the criteria for FUS, CIR, or HS was in each case inversely and significantly associated with FA status. Each abnormality reached maximal frequency early in the culture period (at day 7) and in cultures with the lowest FA concentration (20 nM). At day 7, FUS, CIR, and HS were more abundant than the conventional CBMN-Cyt biomarkers (MN, NPB, and NBud). The marked reduction in frequency of FUS, CIR, and HS recorded later in culture (days 14 and 21) suggests that they may represent a high degree of CIN, rendering the affected cells nonviable or resulting in mitotic arrest. This possibility is supported by the sharp rise in the percentage of both necrotic and apoptotic cells recorded at day 14 in the 20 nM FA cultures. The decline in apoptotic cells beyond day 14 in this treatment may, in turn, reflect the early loss of the most severely damaged cells, that is, those which displayed FUS, CIR, or HS morphologies at day 7. However, this can only be confirmed by tracing the fate of individual cells by live cell imaging [Huang et al.,2011]. Although the frequency of FUS, CIR, or HS peaked at day 7 in PBL cultured in the lowest concentration of FA (20 nM), maximum numbers were recorded in the mid-range of FA concentration (60 nM) at days 14 or 21, possibly reflecting a slower and/or delayed deterioration of genomic stability in the presence of marginal levels of FA.
When absolute frequencies of the novel FA-sensitive biomarkers are compared in 20 nM FA cultures, the abnormalities leading to the FUS defect are by far the most sensitive to FA deficiency, both in terms of fold-increase relative to FA-replete (180 nM) cultures and in absolute terms. The frequency of FUS at day 7 in 20 nM FA increased 40-fold compared to the frequency in 180 nM FA cultures at the same time point. The frequency of BN cells displaying ≥1 NPB at day 7 in the 20 nM FA condition, on the other hand, was only 2.5-fold greater than in the 180 nM condition. As such, FUS may represent a useful biomarker of chromosomal damage in short-term acute studies, whereas for longer term studies (beyond 7 days) scoring, the frequencies of clearly differentiated NPBs, NBud, and MN may be preferable as these markers increased more gradually over the course of the 21 days. These findings also suggest that cells with a small number of individual conventional NPBs, NBuds, and MN remain viable for longer, with delayed crisis. In contrast, the novel morphologies described herein may be reflective of more severe damage. As discussed earlier, it may lead to an early peak in the frequency of these abnormalities and to cell death. It is noteworthy that complex nuclear morphologies, such as FUS, have not been reported previously. One explanation may be that studies of genotoxic effects of folate on PBL in the CBMN-Cyt assay are typically performed after longer than 7 days of culture [Bull et al.,2011].
NPBs, which originate from anaphase bridges, have been reported in diverse medical and experimental settings [Fenech,2007; Fenech et al.,2011]. Mechanistically, they can be attributed to misrepair of DNA breaks, or to fusions between abnormally shortened, dysfunctional, or uncapped telomeres. This leads in turn to complex chromosomal aberrations and the potential for breakage-fusion-bridge (BFB) cycling [van Steensel et al.,1998; Thomas et al.,2003; Gisselsson,2005; Tusell et al.,2008]. We have previously speculated that FA depletion and resulting hypomethylation may lead to shortened or compromised telomeres, causing uracil incorporation and/or disruption of heterochromatin. Either of these events could cause telomere end fusions, with the attendant potential for BFB cycling and elevated CIN [Bull and Fenech,2008]. Accordingly, we hypothesized that a possible mechanism underlying the novel morphologies observed in this study may be the formation of multiple dicentric chromosomes and NPBs. Analysis of the frequency of dicentric chromosomes in metaphase preparations from PBL cultured for 7 days in 20 nM FA failed, however, to support this model. The frequency of dicentric chromosomes was found to be consistent with the frequency of conventional NPBs and insufficient to explain the high frequencies of the FUS, CIR, or HS morphologies.
Additional analysis of the novel morphologies, using PNA–FISH, revealed a high frequency of centromeric DNA within the nuclear strands/structures of fusion regions as well as at the inner nuclear perimeter of CIR cells. These unexpected observations suggest that the abnormal morphologies may be due to disruption of chromatid separation during mitosis. Given FA deficiency induces hypomethylation and uracil incorporation into DNA, it is plausible that conformational changes of chromatids may occur that lead to reduced binding affinity of key proteins required for sister chromatid separation during mitosis. This may result in stalling of mitosis at late metaphase or early anaphase and reconstitution of the nuclear membrane around the abnormally distributed, or incompletely segregated, chromosomes (Fig. 7). In support of this notion, altered DNA methylation status at the centromere has been shown to distort the topology of the binding site for topoisomerase II (Topo II), resulting in a significant reduction in decatenation activity, and consequent defective or stalled chromatid separation at early anaphase [Boos and Stopper,2001]. This in turn results in a reduction in integrity of Topo II activity and the failure of sister chromatid separation [Coelho et al.,2003; Nasmyth and Haering,2005]. Furthermore, evidence of fusions has been demonstrated in instances where decatenation failure occurs due to disruption of the “structural maintenance of chromosomes'' protein complexes, including cohesin and condensin [Coelho et al.,2003]. The latter proteins work in concert with Topo II and are pivotal in the metaphase–anaphase transition. Condensin is essential for the pairing and alignment of sister chromatids at the metaphase plate, while cohesin forms a loop structure around the chromatid arms [Hirano,2006]. Degradation of these structures is an ATP-dependent process leading to the subsequent release of chromatids at anaphase.
Proposed models for the formation of cells displaying FUS, CIR, or HS nuclear morphologies, based on the observations using PNA–FISH for centromeres and telomeres. For the sake of simplicity, only a small number of sister chromatids are shown for each. A: Sister chromatids aligned normally on the metaphase plate; (B) chromatids separate successfully to form (C) two daughter nuclei with an equal complement of chromosomes; (D) a dicentric chromosome arising from misrepair of a double-strand break or chromosome end-to-end fusion, resulting in (E) a single nucleoplasmic bridge (NPB). (F) and (J) depict possible defective chromatid structures, each of which would be stalled at the metaphase plate, and possibly result in the FUS, CIR, and HS nuclear anomalies. F: Chromatids fail to completely separate at anaphase due to dysfunctional degradation and removal of “structural maintenance of chromosome” (SMC) protein complexes (i.e., cohesin and condensin) at chromatid arms; (G) FUS morphologies may arise when a large proportion of chromatids fail to separate completely at anaphase, resulting in multiple fusions containing centromeric and telomeric DNA centrally oriented between the daughter nuclei. H: CIR or (I) HS morphologies may result from fusions occurring laterally between the nuclei. J: Complete separation failure at the chromatid arms and at the centromere; (K and M) HS or CIR morphologies may also result from complete failure of all chromatids to separate at the centromere and/or chromatid arms at early anaphase, resulting in stalling at the metaphase plate (depicted here as viewed down the central axis), and reconstitution of the nuclear membrane around the resulting unseparated chromosomes. (Possible mechanisms for chromatid separation failure are discussed in greater detail in the text). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
Another process that may be detrimentally affected by altered methylation status at the centromere is kinetochore formation. CENP-B is one of a group of co-dependent proteins required for the formation of a functional kinetochore, and depletion of any one of them leads to chromosome missegregation and disruption of mitosis [Decordier et al.,2008]. CENP-B is the only protein in this process that binds directly and specifically to centromeric DNA, at a 17 bp motif called the CENP-B box [Decordier et al.,2008]. Interestingly, centromeric and pericentromeric repeating regions, including CENP-B box motif-enriched α-satellite regions, are highly sensitive to altered methylation status, even more so than subtelomeric repeats [Decordier et al.,2008; Jaco et al.,2008]. Hypomethylation of centromeric repeats has been shown to result in centromeric elongation as well as increased recombination [Jaco et al.,2008]. For this reason, it has been speculated that hypomethylation may disrupt binding of key centromeric proteins (including CENP-B) and lead to mitotic dysfunction and genomic instability [Jaco et al.,2008]. Another possibility that cannot be excluded is that the multiple bridges observed in FUS cells are due to defects in proteins, such as those responsible for Bloom's syndrome (BLM) and/or Fanconi anemia (FANC). These proteins are known to associate in a multienzyme complex that is required for sister chromatid decatenation. The FANC genes are involved in rescuing abnormal anaphase and telophase cells through FANC-dependent targeting of BLM to noncentromeric abnormal structures induced by replicative stress, such as fragile sites, that are likely to be induced when folate is deficient [Jacky et al.,1983; Chan et al.,2009; Naim and Rosselli,2009]. Therefore, it is plausible that FUS, CIR, and HS structures may arise as a result of folate deficiency-induced failure of sister chromatid separation during anaphase/telophase. Figure 7 illustrates how defects in sister chromatid segregation could lead to the FUS, CIR, and HS nuclear anomalies (Fig. 7).
A limitation of this study is that the observations reported are based on lymphocytes from a single donor. Although concentrations of folate, B12, and Hcy in the plasma of this donor were in the clinically healthy range, it is possible that other factors such as genetic background may have predisposed this individual to a greater frequency of rarer, abnormal nuclear morphologies when challenged by folate insufficiency. Further study is warranted with a larger pool of donors to verify whether these observations are more generally applicable, or whether they occur specifically in individuals whose genetic background make them susceptible to chromosome segregation errors when the cells are deficient in folate or other methyl donors.
In conclusion, new data generated from this study confirms and extends previous findings [Fenech and Crott,2002; Wang and Fenech,2003; Kimura et al.,2004; Beetstra et al.,2005; Fenech et al.,2005; Lindberg et al.,2007] by (i) defining novel nuclear anomalies (FUS, CIR, and HS) that are associated with FA deficiency in vitro; (ii) establishing scoring criteria to allow the frequencies of these novel nuclear anomalies to be scored; (iii) demonstrating a strong inverse, dose-dependent effect of FA concentration on the frequency of these abnormal nuclear morphologies in vitro; and (iv) providing preliminary evidence that suggests mitotic dysfunction as the main underlying mechanism responsible for these novel nuclear anomalies. We speculate that mitotic dysfunction may stem from the effects of FA-deficiency on DNA hypomethylation, with reduced methylation leading to altered DNA topology. A result of this may be reduction in binding of key mitotic enzymes that are required for chromatid decatenation, separation, and segregation. Given the rapid generation of these morphologies, and their subsequent loss beyond day 7, we propose that they may represent useful biomarkers for short-term, acute genotoxic studies caused by demethylating conditions. Our study suggests that these novel abnormal nuclear morphologies generated by FA deficiency may be useful additions to the existing biomarkers of CIN defined in the CBMN-Cyt assay. Further studies are required to confirm these findings and to test the hypotheses offered herein to explain the mechanism(s) responsible for generating these nuclear abnormalities.
Acknowledgements
The paper was written, the experiments conceived and designed, and data analyzed by C. Bull, G. Mayrhofer, and M. Fenech. All experimental work was conducted by C Bull. D. Zeegers, G.L.K. Mun, and M.P. Hande contributed technical expertise, advice, and training for the FISH method, including interpretation of data. All authors critically revised and approved the final manuscript.
