Mutagenesis, Vol. 18, No. 3, 221-233,
May 2003
© 2003 UK Environmental Mutagen Society/Oxford University Press
REVIEW |
What do human micronuclei contain?
Laboratory of Molecular and Cellular Toxicology, Department of Industrial Hygiene and Toxicology, Finnish Institute of Occupational Health, Topeliuksenkatu 41 a A, FIN-00250 Helsinki, Finland
 |
Abstract |
|---|
 Top Abstract IntroductionDetection of micronucleus...Fragments and chromosomes in...Specific chromosomes in...MN content after in...Effect of in vivo...ConclusionsReferences |
|---|
As micronuclei (MN) derive from chromosomal fragments and whole chromosomes lagging behind in anaphase, the MN assay can be used to show both clastogenic and aneugenic effects. The distinction between these phenomena is important, since the exposure studied often induces only one type of MN. This particularly concerns the use of MN as a biomarker of genotoxic exposure and effects, where differences in MN frequencies between exposed subjects and referents are expected to be small. A specific analysis of the induced type of MN may considerably improve the sensitivity of detecting the exposure effect. MN harbouring chromosomes can be distinguished from those harbouring acentric fragments by the presence of a centromere. The proportion of centromere-positive MN in human lymphocytes increases with age, which primarily reflects an age-dependent micronucleation of the X and Y chromosomes. The X chromosome especially tends to lag behind in female lymphocyte anaphase, being micronucleated more efficiently than autosomes. There is some evidence for an enhanced prevalence of fragments from chromosome 9 in spontaneous human lymphocyte MN and from chromosomes 1, 9 or 16 in MN induced in vitro by some clastogens; the breakage appears to occur in the heterochromatic block of these chromosomes. Although there are indications that centromere identification can improve the detection of clastogenic effects in humans in vivo, smokers have not shown an increase in centromere-negative MN in their cultured lymphocytes, although smoking is known to produce chromosomal aberrations. This may suggest that fragment-containing MN and chromosomal aberrations cover partly different phenomena. Understanding the mechanistic origin and contents of MN is essential for the proper use of this cytogenetic end-point in biomarker studies, genotoxicity testing and risk assessment.
|
Introduction |
|---|
|
TopAbstractIntroductionDetection of micronucleus...Fragments and chromosomes in...Specific chromosomes in...MN content after in...Effect of in vivo...ConclusionsReferences |
|---|
The analysis of micronuclei (MN) has gained increasing popularity as an in vitro genotoxicity test and a biomarker assay for human genotoxic exposure and effect. The main reasons for this development are obvious. In comparison with chromosomal aberrations (CA), the scoring of MN is simpler, requires shorter training and is less time consuming. In principle, the MN assay can be expected to be more sensitive than the CA assay, because of the increased statistical power brought about by the fact that the number of cells analysed can easily be increased to thousands when only a hundred or a few hundred cells are usually scored for CA.
In humans, most MN studies have been conducted using cultured peripheral lymphocytes, which lend themselves well to both genotoxicity testing and biomonitoring. The cytokinesis block assay, based on cytokinesis inhibition by cytochalasin B (Cyt-B), has facilitated MN analysis exclusively in (binucleate) cells that have completed their first in vitro division after treatment with the test agent or following culture initiation (Fenech and Morley, 1985
; Fenech, 1993, 1997, 1998, 2000). The identification of these cells is important, since MN are formed in cell division, and an accurate estimate on MN frequency can only be obtained from the first post-mitotic interphase after exposure. For biomarker studies, MN formed in vivo can also be examined from uncultured (or in vitro undivided) lymphocytes and from exfoliated epithelial cells collected, e.g. from buccal, nasal or urothelial mucosa (Moore et al., 1993; Surrallés et al., 1996a, 1997; Albertini et al., 2000; Kirsch-Volders and Fenech, 2001).
The two basic phenomena leading to the formation of MN in mitotic cells are chromosome breakage and dysfunction of the mitotic apparatus. MN are formed from acentric chromosome or chromatid fragments and whole chromosomes or chromatids that lag behind in anaphase and are left outside the daughter nuclei in telophase (Ford et al., 1988; Lindholm et al., 1991; Ford and Corell, 1992; Catalán et al., 2000; Falck et al., 2002). Laggards cannot move to the poles, because they are detached from the mitotic spindle or, as described by Cimini et al.(2002) for lagging chromatids, have bipolar (merotelic) orientation. Besides these fundamental mechanisms, some MN may have their origin in fragments derived from broken anaphase bridges (Cornforth and Goodwin, 1991; Saunders et al., 2000) formed due to chromosome rearrangements such as dicentric chromatids, intermingled ring chromosomes or union of sister chromatids.
MN formation is undoubtedly an important mechanism for chromosome loss (Ford et al., 1988), although it is not the only mechanism. Confusingly, chromosome loss is often used as a synonym for (chromosome-containing) MN, as MN are generated by chromosome loss from the nucleus. However, MN are not necessarily lost from the cell. As pointed out by Eastmond and Tucker (1989) and Schuler et al.(1997), MN accompany either the daughter nucleus they derive from or the other daughter nucleus. In the former case, neither of the daughter cells is aneuploid, and in the latter case the micronucleated cell has gained a chromosome, while its daughter cell has lost it. Assuming MN are segregated randomly between the daughter cells, the frequency of centromere-containing MN in binucleate cells (where both daughter nuclei are present) gives an indirect estimate of chromosome loss (or gain) due to micronucleation. Although it has been suggested that MN are eliminated from cells (Ford et al., 1988), the fate of the MN (e.g. expulsion, inclusion in nucleus in mitosis, cell death) remains unclear. It should be kept in mind that the segregation of both sister chromatids of a chromosome to the same daughter nucleus (Zijno et al., 1994, 1996a,Zijno et al., b,c), which may result from the failure of sister chromatids to separate properly (non-disjunction) or engulfment of a lagging chromatid to the same nucleus as its sister chromatid, is an important source of both chromosome loss and gain and does not result in MN (Schuler et al., 1997). Thus, although micronucleation contributes to chromosome loss, these two phenomena cannot be equated.
The ability of the MN assay to detect both clastogenic and aneugenic effects (leading to structural and numerical chromosome alterations, respectively) is an advantage of the MN technique (Kirsch-Volders et al., 1997; Fenech, 2000). The distinction between the two phenomena, by identifying the origin of MN, is important, whether MN analysis is used for genotoxicity testing or for biomonitoring of genotoxic exposure and effect in humans (Thomson and Perry, 1988; Eastmond and Tucker, 1989; Becker et al., 1990; Norppa et al., 1993a,b). The exposure studied often induces only one type of MN, and other types of MN may be uninformative for the assay. Figure 1 shows a theoretical example of a biomarker study on clastogenic exposure. Forty per cent of MN in the subjects represent MN deriving from chromosomal fragments. Unless these MN are specifically identified, the sensitivity of the assay is compromised. Looking at total MN frequency, a doubling of fragment-containing MN only results in a 1.4-fold increase and a tripling only in a 1.8-fold increase.
|
Centromere and kinetochore identification have successfully been used in numerous in vitro and in vivo studies to examine the contents of MN. Information on the more detailed ingredients of MN is more limited, but there are clear indications that different chromosomes are micronucleated non-randomly (see below). It is important to realise that we do not yet know enough on the contents of MN. Yet, a thorough understanding of what MN actually stand for is a basic requirement for the correct use of the MN assay. The present paper reviews the current knowledge on the contents of human MN and how this information may influence the use of the MN assay in short-term testing and biomonitoring.
|
Detection of micronucleus contents |
|---|
|
TopAbstractIntroductionDetection of micronucleus...Fragments and chromosomes in...Specific chromosomes in...MN content after in...Effect of in vivo...ConclusionsReferences |
|---|
As chromosomes contain a centromere associated with a kinetochore structure, the occurrence of an entire chromosome (or chromatid) in MN can be shown by the presence of centromere-specific DNA sequences (Ford et al., 1988
; Becker et al., 1990; Migliore et al., 1993; Norppa et al., 1993b) or centromeric (kinetochore) proteins (Hennig et al., 1988; Thomson and Perry, 1988; Eastmond and Tucker, 1989). MN with acentric fragments are not expected to contain centromeric DNA or kinetochore proteins.
The two main classes of MN can be distinguished from each other by using in situ hybridization (ISH) with DNA probes that identify 
-satellite DNA in all human chromosomes (Figure 2A and B). -Satellite DNA, characterized by a diverged 171 bp motif repeated in a tandem fashion, is found in all human centromeres (see Schueler et al., 2001). The pancentromeric DNA probes used for MN characterization have been cloned alphoid probes (Becker et al., 1990; Thierens et al., 1999a), oligonucleotides with an alphoid consensus sequence (Norppa et al., 1993b; Elhajouji et al., 1995; Darroudi et al., 1996), commercially available probes for all human centromeres (Migliore et al., 1993; Titenko-Holland et al., 1994; Doherty et al., 1996) or -satellite probes prepared by PCR (Huber et al., 1996). Fluorescence in situ hybridization (FISH) with digoxygenin- or biotin-labelled DNA probes detected in a fluorescence microscope by immunofluorescence has been used in most studies, but DNA probes directly labelled with a fluorochrome are nowadays available. Immunohistochemical detection by, for example, alkaline phosphatase or peroxidase has been utilized in some papers (Guttenbach et al., 1994; Nardone, 1997; Vral et al., 1997). In principle, centromeric DNA can also be detected by primed in situ labelling (PRINS) (Russo et al., 1996; Basso and Russo, 2000), but this technique appears not to have been applied for human MN. Immunofluorescence or immunohistochemistry have also been used for the detection of kinetochore proteins in human MN (Hennig et al., 1988; Thomson and Perry, 1988; Eastmond and Tucker, 1989; Fenech and Morley, 1989). Kinetochore proteins are identified by anti-kinetochore antibodies derived from the serum of scleroderma CREST (Calcinosis, Raynaud’s phenomenon, Esophageal dysmotility, Sclerodactyly and Telangiectasia) patients (Moroi et al., 1980).
|
Schuler et al.(1997)
have reviewed the advantages and disadvantages and technical aspects of the CREST and FISH methods in characterizing MN. Fragments with the break point at the centromere may form centromere-positive (C+) MN if the breakage destroys centromere function. True centric fragments with a complete centromere are expected to be rare in MN of non-cancerous cells, as they should normally be attached to the spindle, unless they lack a functional kinetochore. It is also possible that a MN contains both one or more fragments and one or more chromosomes; in this case, the detection of centromeric DNA would only suggest the presence of the latter. In general, MN deviating from the two simple categories (acentric fragment versus chromosome) should in most cases be relatively rare in non-cancerous cells. They would either require an inducer with a very specific mode of action or the simultaneous occurrence of two or several (more or less) independent events whose combined probability should be low.
MN formed from entire chromosomes with a disrupted or detached kinetochore may result in MN with no kinetochore signal (Schuler et al., 1997). This appears to be a real problem for the use of kinetochore identification in human biomonitoring and studies of spontaneous MN, since a considerable portion of MN harbouring the centromere of the X or Y chromosome seem to be kinetochore-negative (K-) (Hando et al., 1994, 1997; Nath et al., 1995; Tucker et al., 1996). Kinetochore defects may be common for the micronucleation of other chromosomes as well, but even the X chromosome alone is expected to give rise to errors in the assessment of MN contents using kinetochore antibodies, since this single chromosome may account for 70% of MN in binucleate lymphocytes of women (Hando et al., 1994, 1997). In practice, centromere and (with some reservations) kinetochore detection can be expected to be accurate enough in distinguishing the two main types of MN in experimental studies, but it appears that centromeric FISH should be recommended for studies of human exposure and spontaneous MN levels, due to the low prevalence of kinetochore label in MN harbouring entire sex chromosomes.
It is preferable to evaluate the centromere/kinetochore content of MN by calculating the frequencies of signal-positive and signal-negative MN than by looking at their proportions (Fenech and Morley, 1989; Surrallés et al., 1995, 1996a; Falck et al., 1997; Schuler et al., 1997). Proportions do not well take into account the magnitude of the effect and can be misleading if the exposure induces both types of MN or actually decreases the frequency of one type of MN (Falck et al., 1997; Schuler et al., 1997).
Chromosome-specific centromeric DNA probes have been used to identify the presence of specific chromosomes in MN by FISH (Guttenbach et al., 1994; Hando et al., 1994, 1997; Richard et al., 1994; Catalán et al., 1995; Tucker et al., 1996; Acar et al., 2001; Bakou et al., 2002). The use of a pancentromeric probe, an X chromosome-specific centromeric probe and (in men) a Y chromosome-specific centromeric probe allows one to estimate the contribution of acentric fragments, X chromosomes, Y chromosomes and autosomes in MN (Catalán et al., 1995, 1998, 2000; Surrallés et al., 1996a; Falck et al., 2002).
The contribution of all human chromosomes in MN has been studied by chromosome painting (Fauth et al., 1998, 2000; Fauth and Zankl, 1999) and spectral karyotyping (Komae et al., 1999; Leach and Jackson-Cook, 2001); coupled with centromeric probes this approach can distinguish between MN harbouring centric and acentric DNA derived from specific chromosomes.
|
Fragments and chromosomes in spontaneous micronuclei |
|---|
|
TopAbstractIntroductionDetection of micronucleus...Fragments and chromosomes in...Specific chromosomes in...MN content after in...Effect of in vivo...ConclusionsReferences |
|---|
In various studies of spontaneous MN in cultured human lymphocytes, the average proportion of whole chromosomes in MN has usually varied between 30 and 80%, as measured by CREST antibody or FISH (Fenech and Morley, 1989
; Yager et al., 1990; Norppa et al., 1993b; Scarpato et al., 1996; Surrallés et al., 1996a,b; Migliore et al., 1997, 1999a, Migliore et al., b; Calvert et al., 1998; Carere et al., 1998; Davies et al., 1998; Thierens et al., 1999a, 2000; Maffei et al., 2000; Leach and Jackson-Cook, 2001; Bakou et al., 2002).
In exfoliated buccal, nasal and urothelial cells of non-smoking volunteers 23–42 years of age, the proportion of C+ MN was 56–57% (Moore et al., 1993; Titenko-Holland et al., 1994). In buccal cells of male benzene-exposed workers and controls (aged 19–61 years, mean 41 years, 45% smokers), C+ MN constituted 35% of identifiable MN (Surrallés et al., 1997). The same prevalence (35%) of C+ MN was observed in nasal cells of non-smoking male stainless steel production workers aged 34–62 years (Huvinen et al., 2002). A considerably lower level (6%) of C+ MN in buccal cells of hyperthyroidism and thyroid cancer patients before radiation therapy was reported by Ramírez et al.(1999a).
The prevalence and frequency of kinetchore-positive (K+) MN (Fenech and Morley, 1989; Hando et al., 1994, 1997; Odagiri and Uchida, 1998) and C+ MN (Catalán et al., 1995; Scarpato et al., 1996; Ramírez et al., 1997; Thierens et al., 1999a, 2000; Bakou et al., 2002) in human lymphocytes increases with age. Thierens et al.(2000) estimated that 80% of the effect of age on MN frequencies is due to C+ MN; women had higher frequencies of C+ MN than men. Also, Scarpato et al.(1996) noticed a clearer age-dependent increase in lymphocyte C+ MN in women than in men.
The effect of age on fragment-containing MN would appear to be less clear. An increase in K- MN with age was observed in cultured binucleate lymphocytes of male subjects (Nath et al., 1995) and in a pooled group of males and Turner syndrome patients (Hando et al., 1997). Also, Odagiri and Uchida (1998) reported that the age-dependent increase in lymphocyte MN frequency concerns both K+ and K- MN. In addition, they observed that serum vitamin C level was associated with an increased frequency of cultured lymphocytes with K+ MN and that serum folic acid level was negatively related to K+ MN after age adjustment. As the majority of MN harbouring X or Y chromosomes, the two chromosomes primarily responsible for the age effect on MN, appear to be devoid of kinetochore signals (Hando et al., 1994, 1997; Nath et al., 1995; Tucker et al., 1996), it is probable that kinetochore identification is not accurate enough in discriminating the two main groups of MN. Using FISH on binucleated lymphocytes, Bakou et al.(2002) reported a 1.7-fold higher frequency of centromere-negative (C-) MN in four women 47–50 years of age as compared with four women 22–26 years of age, but Thierens et al.(1999a), who studied 215 subjects, did not observe a significant age effect on C- MN.
In uncultured T lymphocytes of women, the proportion of C+ MN was higher (71.6%) than in cultured cells (53.3–55.2%) (Surrallés et al., 1996a). This particularly reflected an increase in C- MN in the cultures. In comparison with the uncultured T cells, the frequency of C- MN per 1000 nuclei was 2.3–2.9 times higher in the cultured cells. The increase in C- MN by cell culture probably reflected the expression, by the in vitro mitosis, of damage accumulated in vivo in the quiescent lymphocytes, i.e. conversion of pre-existing DNA lesions into chromatid breaks and appearance of in vivo chromosome type breaks.
Several studies on human lymphocytes have shown that MN frequency is increased with increasing culture time, with and without Cyt-B (see Falck et al., 1997). This effect appears to be mainly due to C+ MN (Falck et al., 1997; Sgura et al., 1997). The reason for the finding is not clear, but it could reflect differential growth rate of subpopulations of lymphocytes with different baseline MN frequencies, prolonged cell cycle of damaged cells, deteriorating culture conditions or higher X micronucleation at longer culture times (Richard et al., 1994; Falck et al., 1997). In cultures without Cyt-B, the culture time-dependent increase in C+ MN could just be due to the increase in mitotic activity. Prolonged incubation with Cyt-B can lead to the formation of binucleate cells that have actually divided twice in the presence of Cyt-B; such cells are expected to have a high MN frequency (Lindholm et al., 1991; Norppa et al., 1993b; Surrallés et al., 1994; Zijno et al., 1994). The second division in the presence of Cyt-B (division of a binucleate cell) is grossly irregular, resulting in high frequencies of lagging chromosomes in anaphase and formation of C+ and K+ MN, mostly containing autosomes (Lindholm et al., 1991; Norppa et al., 1993b; Falck et al., 2002).
Binucleate human lymphocytes produced by Cyt-B were found to have a higher proportion of C+ and K+ MN than cells cultured without Cyt-B, which reflected decreased frequencies of C- and K- MN per 1000 nuclei in binucleate cells (Norppa et al., 1993b; Surrallés et al., 1996a; Falck et al., 1997, 2002; Catalán et al., 1998). The diminished micronucleation of fragments was suggested to be due to reduced distance between the poles in cytokinesis-blocked cells, leading to increased engulfment of laggards into the daughter nuclei (Norppa et al., 1993b; Surrallés et al., 1996a; Falck et al., 1997; Minissi et al., 1999). Measurements in human lung fibroblasts subsequently showed that the pole-to-pole distance is, indeed, decreased in the presence of Cyt-B (Cimini et al., 1999). In female lymphocytes, the frequency of lagging acentric fragments in bipolar anaphases was lower in the presence of Cyt-B than in its absence (Falck et al., 2002). Presumably, less acentric fragments are left outside the poles due to the shortened pole-to-pole distance. The short distance between the poles and absence of a contractile ring also appears to result in an increase in C-anaphases where sister chromatids have not travelled apart enough to give rise to two daughter nuclei but result in a polyploid restitution nucleus (Minissi et al., 1999); this seems to occur for heavily impaired anaphases of colchicine-treated lymphocytes, but may also concern spontaneously damaged anaphases.
Apparently healthy subjects with exceptionally high MN frequencies are occasionally observed in human studies. Thierens et al.(1999a) reported a high frequency of C+ MN (53/1000 binucleate cells) in lymphocytes of a 50-year-old non-smoking man. No apparent reason for this finding was indicated. In another study (Thierens et al., 2000), two female radiological workers with similarly high frequencies of C+ MN were encountered, many of the micronucleated binucleate cells carrying two or more MN. Most of the MN turned out to contain the X chromosome.
Some diseases have been associated with an increase in specific types of MN in cultured lymphocytes. Systemic lupus erythematosus patients and systemic sclerosis patients with topoisomerase I antibodies (but not antibodies against centromeric proteins) showed an increase in C- MN, while systemic sclerosis patients with centromeric antibodies (but not antibodies against topoisomerase I) had an elevation in C+ MN (Migliore et al., 1999b). Parkinson’s disease patients, who appear to have elevated oxidative stress and DNA damage, showed an increase in binucleate lymphocytes with C- MN (Migliore et al., 2002). Both C+ and C- MN were increased in patients suffering from mitochondrial encephalomyopathies, possibly reflecting the consequences of increased radical production caused by altered mitochondrial respiratory chain (Naccarati et al., 2000).
Fragments in MN can be of the chromatid type (single) or chromosome type (double), but it is not known whether micronucleation occurs equally for both types of fragments. In metaphase, chromatid fragments are attached to the homologous area of the sister chromatid while acentric chromosome fragments are apart from their chromosome of origin. It might be envisaged that acentric chromatid fragments are not always separated from the sister chromatid in anaphase, in which case they would form MN less efficiently than acentric chromosome fragments. Some information on the involvement of both types of fragments in MN may be obtained by simultaneous identification of telomeres and centromeres in MN (Figure 2C and D
). Preliminary FISH data on binucleate lymphocytes of four human donors (Norppa et al., unpublished results) indicated that 60% of C- MN contain one telomere signal, suggesting no marked reduction of chromatid type break micronucleation. In general, acentric fragments lagging in female lymphocyte anaphase seemed to be micronucleated quite efficiently, with 41–65% of them apparently ending up in MN (Falck et al., 2002).
The contribution of centric fragments to human MN is poorly known. Such MN may be frequent in cells with genomic instability. In HeLa cells, the expression of human T cell leukaemia virus type I (HTLV-I) Tax oncoprotein increased the proportion of C+ MN without telomeres or with free 3'-OH ends (examined by terminal transferase-mediated in situ addition of digoxygenin-dUTP) from ~20 to ~37% but decreased the proportion of C- MN with telomere signals or without free 3'-OH ends (Majone and Jeang, 2000). The authors concluded that Tax interferes with protective cellular mechanisms that stabilize DNA breaks by adding telomeric caps.
|
Specific chromosomes in spontaneous micronuclei |
|---|
|
TopAbstractIntroductionDetection of micronucleus...Fragments and chromosomes in...Specific chromosomes in...MN content after in...Effect of in vivo...ConclusionsReferences |
|---|
If the micronucleation of each human chromosome is random, each single chromosome should appear in 1/46 (2.17%) of chromosome-containing MN (assuming only one chromosome in a MN) and chromosome-specific DNA should be found in acentric fragment-containing MN corresponding to the length of each chromosome. FISH studies have, however, shown that different human chromosomes are non-randomly involved in MN. In cultured human lymphocytes, the increase in C+ MN with ageing has primarily been attributed to an age-dependent micronucleation of the X and Y chromosomes (Guttenbach et al., 1994
; Hando et al., 1994; Richard et al., 1994; Catalán et al., 1995, 1998; Surrallés et al., 1996a; Zijno et al., 1996a,b; Bakou et al., 2002). These findings are in line with the well-known age-dependent loss of X and Y chromosomes (see Stone and Sandberg, 1995; Catalán et al., 1998).
X chromosome
Richard et al.(1994) observed by centromeric FISH that the proportion of X chromosome-positive (X+) MN in cultured lymphocytes of five female donors was dependent on age, ranging from 7.4% in the youngest (27 years) to 20.7% in the oldest (80 years); the frequency of X+ MN varied from 0.1 to 1.3 per 1000 cells in the same persons. The X chromosome was clearly over-represented in the MN, while chromosomes 11 and 22, detected (by chromosome painting and with a non-centromeric cosmid probe, respectively) in the oldest of the donors, were very rare. Guttenbach et al.(1994) found that the proportion of X+ MN among all lymphocyte MN was 20% in five older women (aged 71–84 years) and 8% in five girls (aged 1–10 years). As the frequency of total MN was more than 4 times higher in the older age group, there appeared to be an almost 10-fold age difference in the frequency of X+ MN. In six women over 50 years of age, the proportion of X+ MN among all MN was 24.0%, while it was 14.0% in six women below 30 years of age (Catalán et al., 1995); an age dependency was also observed for the proportion of autosome-containing MN (28 versus 20%).
Much higher proportions of X+ MN have been observed in binucleate lymphocytes produced by Cyt-B. Hando et al.(1994) observed an X signal in 72.2% of all MN in a group of female donors aged 0–77 years, indicating that X micronucleation is far from being random. About half of the X+ MN contained more than one signal. Interestingly, the majority of both K+ (77.3%) and K- MN (64.9%) contained an X signal; 36.9% (290/785) of the X+ MN did not appear to have a kinetochore, indicating that many micronucleated X chromosomes tend to have faulty kinetochores. The frequency of X+ MN was shown to be age dependent and no X+ MN were observed in newborns, who also otherwise exhibited a low MN frequency (Hando et al., 1994). In cultured binucleate lymphocytes of 12 women (25–56 years of age), X+ MN accounted, on average, for 62.3% of all MN and their frequency was 10.5 per 1000 binucleate cells (Zijno et al., 1996a,b); an age dependency was observed for the frequency but not for the proportion of X+ MN, suggesting that other types of MN also contribute to the age-dependent MN increase. Bakou et al.(2002) observed that binucleate lymphocytes of two women 47 and 49 years of age showed clearly higher frequencies of X+ MN and MN with autosomes than lymphocytes from two younger women 22 and 23 years of age.
In uncultured T lymphocytes of five women aged 47–60 years, the proportion of X+ MN of all MN was 28.5%, corresponding to 39.6% of C+ MN (9 times higher than expected by chance), which indicated that the X chromosome is also over-represented in MN in vivo (Surrallés et al., 1996a). The percentage of X+ MN did not change much on culturing the T lymphocytes for 72 h (without Cyt-B), although the frequency of X+ MN was increased from 2.6 to 3.9 per 1000 nuclei. However, in binucleate cells produced by Cyt-B, the proportion of X+ MN was 42% of all MN (79.8% of C+ MN). This high percentage was explained by an increase in the frequency of X+ MN (per 1000 nuclei) and reduction (to almost 1/4 of the frequency in cultures without Cyt-B) of the calculated frequency of MN harbouring autosomes. An increase in the proportion and frequency of X+ MN in binucleate cells and a decrease in the frequency of autosome- or fragment-containing MN was also observed in other studies comparing MN contents with and without Cyt-B (Falck et al., 1997, 2002; Catalán et al., 1998). The clearly higher percentages of X+ MN reported in series where Cyt-B was used as compared with those where Cyt-B was not used (Hando et al., 1994; Tucker et al., 1996; Zijno et al., 1996a,b; Hando et al., 1997; Catalán et al., 1998; Bakou et al., 2002) also support the view that Cyt-B favours a high prevalence of the X chromosome in MN.
In anaphase lymphocytes of a 62-year-old woman, the X chromosome frequently lagged behind, constituting 26 (without Cyt-B) and 12% (with Cyt-B) of all lagging chromosomes (Catalán et al., 2000; Falck et al., 2002). X laggards more often contained both sister chromatids (Figure 2E and F) and were more distally located than autosome laggards (Catalán et al., 2000). The high proportion of the X+ MN in binucleate cells appeared to be due to the fact that all X laggards were micronucleated in the presence of Cyt-B, while this efficiency was ‘only’ 49% without Cyt-B (Falck et al., 2002). It may be that X laggards, detached from the spindle due to faulty kinetochores, are not forced from the periphery nearer the daughter nuclei in cytokinesis-blocked cells which are not elongated and lack the central spindle and contractile ring.
The high loss and micronucleation of the X chromosome in women has been suggested to be primarily due to the inactive X (see Surrallés et al., 1996b, 1999; Tucker et al., 1996; Catalán et al., 2000). The inactive X chromosome was observed to show a higher age-dependent telomere shortening than the active homologue or all chromosomes in metaphases of cultured lymphocytes from newborn, middle-aged and elderly females (Surrallés et al., 1999); the accelerated shortening may affect the segregation of the inactive X chromosome. Tucker et al.(1996) observed that the inactive X is more often involved in female MN than the active homologue. They studied the cytokinesis-blocked lymphocytes of two healthy females who had a translocation between chromosome 9 and the active X homologue. The active X in MN was identified by the presence of both chromosome X and 9 paints. The X chromosome was found in 44.2% of all MN (in 52.3% of all MN in a karyotypically normal woman studied in parallel) and 83.3% of the X+ MN contained the inactive X but no chromosome 9 signal. Kinetochore label was absent from 73.8% (59/80) of MN harbouring the untranslocated (inactive) X and from 56.3% (9/16) of MN harbouring the translocated (active) X.
However, Surrallés et al.(1996b), who identified the inactive X chromosome through histone H4 underacetylation, observed no significant differences in the micronucleation of the active and inactive X chromosomes, suggesting that both homologues are preferentially micronucleated. The X chromosome was highly over-represented (15.5–68.1% of all MN, average 24.4%) among binucleate lymphocyte MN of the two older (57 and 60 years) and two younger (24 and 27 years) women studied, and the older women had 2.5 and 3.4 times higher frequencies of MN with (respectively) active and inactive X than the younger women. When the late replicating inactive X in female lymphocyte anaphase (without Cyt-B) was labelled with bromodeoxyuridine (Figure 2F), the inactive and active X chromosomes seemed to lag behind in equal proportions (Catalán et al., 2000).
Although the role of inactive X as the source of the high X+ MN frequencies remains unclear, it is obvious that the phenomenon is not entirely explained by the inactive homologue. This is because the X chromosome is also over-represented in lymphocyte MN of men, who have only one (active) X chromosome (Hando et al., 1997; Catalán et al., 1998; Carere et al., 1999). In a combined group of 43 males (0–79 years of age) and seven Turner syndrome (45,X) patients (11–39 years), the X chromosome was found in 6.6% of all MN, which was higher than would be expected by chance, but still only 1/10 of the percentage found in females of similar age range (Hando et al., 1997). A total of 68.2% of the X+ MN did not show kinetochore signals, again suggesting a high prevalence of kinetochore defect in the micronucleated X chromosomes. The frequency of X+ MN in the pooled group of males and 45,X females was age dependent. The results primarily reflected increased X micronucleation in the males studied; conclusions could not be drawn for the Turner syndrome patients, as only the oldest one of them (the others were ~20 years old or younger) had X+ MN among the MN found in 2000 cells scored per individual (Hando et al., 1997).
In another study (Catalán et al., 1998), the X chromosome was included in 7.0 (without Cyt-B) and 15.2% (with Cyt-B) of all lymphocyte MN in five men >50 years of age and in 4.8 and 6.2% of all MN (respectively) in five men <30 years of age, showing over-representation among C+ MN in both age groups. Four women (aged 26–58 years), studied for comparison, had much higher rates of X micronucleation than the men, with 26.0 (without Cyt-B) and 41.8% (with Cyt-B) of their MN being X+. The sex difference in the frequency of MN was mainly due to the X chromosome. In 24 male subjects (age range 31–62 years), the frequency of chromosome X in binucleate lymphocytes was 2.33 per 1000 cells, much higher than that of chromosomes 7, 11 and 18 (0.20, 0.16 and 0.08, respectively) (Carere et al., 1999). Correlation between age and the frequency of X+ MN was almost statistically significant (P = 0.06).
The high micronucleation of the only (active) X in men suggests that the X chromosome has a general tendency to be micronucleated, regardless of the inactive X. Kinetochore defects may be behind this phenomenon (Hando et al., 1994, 1997; Tucker et al., 1996). The reason for the extremely high X micronucleation in women is still, however, unknown, although the inactive X remains an attractive explanation.
Y chromosome
Guttenbach et al.(1994) observed that the proportion of Y+ MN in cultured lymphocytes was 14% of all MN in five boys 0.5–4 years of age and 20% in five older men 73–91 years of age, the older men showing a higher mean frequency of Y+ MN (1.0/1000 cells) than the children (0.3/1000 cells). An over-representation of Y in MN was also seen by Nath et al.(1995) in binucleate lymphocytes of 35 adult men (22–79 years of age) and 18 newborns, 13.5% of all MN being Y+. A total of 87.8% of the Y+ MN did not show kinetochore signals, indicating that kinetochore damage is an important cause of Y micronucleation. The frequency of Y+ MN showed age dependency. The newborns had a very low MN frequency and no Y+ MN among the 17 MN characterized from them. In the study of Catalán et al.(1998), five men above 50 years of age were observed to have the Y chromosome in 10% of MN of their cultured lymphocytes (with and without Cyt-B), indicating a clear over-representation (18.1–23.6% of C+ MN) of this chromosome. The frequency of Y+ MN was 8–11 times higher in these men than in a group of five younger men aged below 30 years, who showed expected rates of Y micronucleation.
Autosomes
The involvement of other chromosomes than X and Y in human MN is poorly known. In anaphases of female lymphocytes, autosomes comprised the majority of laggards (Catalán et al., 2000; Falck et al., 2002). However, autosomes seemed to form MN clearly less efficiently than the X chromosome or acentric fragments, at a rate of only 11 (without Cyt-B) and 8% (with Cyt-B) (Falck et al., 2002). They were probably engulfed in the nuclei due to their proximal location (Catalán et al., 2000) or possibly because they were only delayed in their anaphase movement (Falck et al., 2002) or still had some connection to the spindle (perhaps merotelic orientation; Cimini et al., 2002).
There is some evidence for a preferential inclusion of DNA from chromosome 9 in human lymphocyte MN. Tucker et al.(1996) studied binucleate lymphocytes of two t(X;9) females and observed that 16.5% of the X- MN were positive for a chromosome 9 paint signal, which appeared to be higher than expected. Kinetochore signal was absent from 85% of the chromosome 9-positive (9+) MN. In a karyotypically normal woman, studied for comparison, 9.5% of X- MN showed chromosome 9 paint, but this finding was only based on two 9+ MN (Tucker et al., 1996). Fauth et al.(1998, 2000) observed by chromosome painting that besides X (11.5% of all MN in two female and one male donor) and Y (11.0% of all MN in a male donor), DNA from chromosome 9 was also over-represented (9.8% of all MN) in lymphocyte MN of two female and one male donor. This effect was suggested to be due to the large heterochromatic block of chromosome 9. Chromosomes 12 and 19 were not found in any MN (Fauth et al., 1998, 2000). The painting technique did not reveal whether the findings concerned an entire chromosome 9 or a fragment. As MN with chromosome 9 could also be highly induced by clastogen treatment (Fauth et al., 1998, 2000; Fauth and Zankl, 1999), fragments were probably involved. It may be of interest in this context that patients with the ICF (immunodeficiency, centromeric heterochromatin instability and facial anomalies) syndrome, who show under-condensation of the heterochromatic blocks of chromosomes 1, 9 and 16 in a portion of their PHA-stimulated lymphocytes, also have a high frequency of micronucleated lymphocytes, and chromosome 1 and (to a lesser extent) chromosomes 9 and 16 appear to be over-represented in the MN (Sawyer et al., 1995; Stacey et al., 1995).
Chromosome 9 over-representation in MN was not, however, observed by Leach and Jackson-Cook (2001), who used spectral karyotyping (SKY) technology to probe the contents of lymphocyte MN in three female donors. Besides a general high contribution of the X chromosome in MN, two of the donors showed a high prevalence (22.2 and 13.1%) of MN with DNA from chromosome 19. Pancentromeric FISH showed a single C+ signal in most of the MN, but these data were not reported for MN harbouring individual chromosomes. When the accuracy of the SKY data was checked by using chromosome-specific DNA probes for 11 chromosomes, the assignments agreed in 71.1% of cases (detailed data not shown). Thus, the SKY technique, developed for metaphase chromosomes, appeared to require some refinements for a more accurate application to interphase nuclei.
Migliore et al.(1995) observed that 33% of lymphocyte MN of a male donor hybridized with a ß-satellite DNA probe identifying group D (13, 14 and 15) and group G autosomes (21 and 22). This suggested that acrocentric chromosomes could be responsible for the majority of spontaneous chromosome-containing MN. However, when the occurrence of acrocentric chromosomes in MN was further investigated by simultaneous use of the ß- and -satellite probes from cultured lymphocytes of 20 donors, acrocentric chromosomes were found among C+ MN at a rate only slightly higher than the expected value (Scarpato et al., 1996). The micronucleation of acrocentric chromosomes was not age or sex dependent. Richard et al.(1994) did not find chromosome 22 in any of the 53 MN they examined from an 80-year-old donor.
Although the non-disjunction of chromosome 21 in cultured lymphocytes was reported to be increased by age, there was no age or sex effect on chromosome 21 micronucleation in healthy persons (Shi et al., 2000). Migliore et al.(1997) observed that Alzheimer’s disease patients have an increased frequency of C+ MN and suggested that the disease may involve microtubule impairment. When the MN were further characterized with centromeric DNA probes for chromosomes 13 and 21 coupled with a single cosmid for region 21q22.2, the patients had, in comparison with controls, a higher frequency of MN with chromosome 21 signals (Migliore et al., 1999a). This led the authors to propose that mosaicism for 21 trisomy could underlie the dementia phenotype of Alzheimer’s disease.
Peace et al.(1999) described a healthy woman who had exceptionally high frequencies of chromosomal aberrations (12%) and MN (119 per 1000 cells) in her lymphocytes. Chromosome painting revealed that more than half of the MN contained chromosome 2. The reason for the aberrations and the high micronucleation of chromosome 2 was left unclear.
Obviously, data on preferential micronucleus formation by other than sex chromosomes are presently conflicting. The micronucleation of fragments from chromosomes with large heterochromatic blocks, especially chromosome 9, is an interesting possibility that has to be checked in further studies. If it turns out to be true, the causes and mechanisms of formation of such MN and their possible influence on the interpretation and conduct of the MN assay would have to be worked out.
|
MN content after in vitro treatment of human cells |
|---|
|
TopAbstractIntroductionDetection of micronucleus...Fragments and chromosomes in...Specific chromosomes in...MN content after in...Effect of in vivo...ConclusionsReferences |
|---|
In genotoxicity testing it is crucial to understand whether an MN inducer acts via a clastogenic or an aneuploidogenic (or both) mechanism. The mode of action will influence the interpretation of the positive test result and the use of the data in risk assessment. While both structural and numerical chromosome alterations occur in carcinogenesis, there is still much more information on carcinogen-induced chromosome breakage and clastogenic carcinogens than carcinogen-induced numerical chromosome abnormalities and aneugenic carcinogens. Evidence is presently limited in favour of agents purely modulating chromosome number, such as spindle poisons, cytokinesis inhibitors or disrupters of the kinetochore or centrosome, being (complete) carcinogens. As both numerical and structural chromosome aberrations are involved in birth defects and miscarriages, testing for both clastogenic and aneugenic activity is, naturally, highly justified.
During the last 12 years, numerous studies have applied kinetochore or centromere detection to assess the mechanism of action of agents tested for MN induction in human cell cultures (see, for example, Hennig et al., 1988; Eastmond and Tucker, 1989; Fenech and Morley, 1989; Becker et al., 1990; Yager et al., 1990; Robertson et al., 1991; Rudd et al., 1991; Weissenborn and Streffer, 1991; Antoccia et al., 1993; Crofton-Sleigh et al., 1993; Migliore et al., 1993, 1996, 1997, 1998, 1999c, Migliore et al., d; Slavotinek et al., 1993; Stopper et al., 1993; Fenech et al., 1994; Kolachana and Smith, 1994; Silva et al., 1994; Surrallés et al., 1995; Van Hummelen et al., 1995; Huber et al., 1996; Kirsch-Volders et al., 1996; Parry et al., 1996; Dopp et al., 1997; Fimognari et al., 1997, 1999, 2001; Schuler et al., 1997; Sgura et al., 1997; Vlachodimitropoulos et al., 1997; Vlastos and Stephanou, 1998; Digue et al., 1999; Gonzáles-Cid et al., 1999; Murg et al., 1999a,b; Andrianopoulos et al., 2000; Buckvic et al., 2000; Nesti et al., 2000; Laffon et al., 2001; Bakou et al., 2002; Decordier et al., 2002). The overall outcome from such exercises has been very promising. Clastogens have preferentially produced K- or C- MN, while aneugens have induced K+ or C+ MN. As pointed out by Schuler et al.(1995), there are relatively few genotoxic agents that induce solely one type of MN. For instance, several clastogens also appear to induce some K+ or C+ MN, although the mechanisms behind this apparent aneugenic activity are not known in detail (see, for example, Eastmond and Tucker, 1989; Slavotinek et al., 1993; Schuler et al., 1997; Vlachodimitropoulos et al., 1997; Vral et al., 1997; Murg et al., 1999a; Ramírez et al., 1999b; Touil et al., 2000; Ponsa et al., 2001).
For some agents assayed in vitro, there is evidence for a chromosome-specific effect. Migliore et al.(1995) used a DNA probe identifying the ß-satellite region (pericentric heterochromatin) in cultured male lymphocytes to show that vanadium compounds preferentially induce the micronucleation of acrocentric group D and G chromosomes. Although sodium orthovanadate also specifically induced loss and gain of the X chromosome rather than chromosome 2, there were no clear differences in the occurrence of the two chromosomes in MN (Migliore et al., 1999d). The X chromosome and chromosome 2 also did not differ from each other for MN induction by griseofulvin, but estramustine favoured the micronucleation of chromosome 2 instead of the X chromosome (Migliore et al., 1999d). The X chromosome did not seem to be over-represented in lymphocyte MN induced by vinblastine or colchicine (Parry et al., 1996). Colchicine was, however, reported to induce mostly C-band-positive K+ MN, indicating preferential micronucleation of acrocentric chromosomes (Caria et al., 1996). In another study with colchicine (Wuttke et al., 1997), chromosome 7 paint signal was observed 1.5 times more frequently than would be expected on the basis of random distribution of chromosomes; chromosome 2 seemed to be micronucleated at the expected rate. With diethylstilbesterol, another aneugen, material from chromosomes 14, 19 and 21 was significantly more frequently micronucleated in comparison with MN from control cultures (Fauth et al., 2000). Bentley et al.(2000) used chromosome-specific centromeric probes (1, 8, 11, 18, X and 17) to study aneuploidy events in female lymphocytes treated with benomyl and carbendazim. Although the X chromosome was slightly more sensitive than the other chromosomes to non-disjunction induction by both chemicals, no differences between the six chromosomes were reported with respect to micronucleus induction. 1,2,4-Benzenetriol, tested in cultured human lymphocytes, was observed to induce more MN with the chromosome 8 centromere than the chromosome 7 centromere (Chung et al., 2002).
The clearest case of chromosome-specific MN induction concerns some clastogens, particularly base analogues preferentially inducing MN harbouring DNA from chromosomes that contain large heterochromatic blocks. 5-Azacytidine is known to induce C- and K- MN and a specific dose-dependent under-condensation of the heterochromatic regions in chromosomes 1, 9, 15, 16 and Y (Guttenbach and Schmid, 1994; Cimini et al., 1996). After an in vitro treatment with 5-azacytidine, a significant fraction of male lymphocyte MN showed hybridizations with these chromosomes, but not with chromosomes 11, 17 and X (Guttenbach and Schmid, 1994). Similar results were obtained by Fauth et al.(1998), who observed by chromosome painting that an in vitro treatment of cultured human lymphocytes with 5-azacytidine resulted in the micronucleation of DNA from chromosomes 1, 9 and 16. It was suggested that under-methylation of the large heterochromatic region in these chromosomes may be associated with their micronucleation. As 5-azacytidine cannot be methylated, its incorporation in DNA leads to changes in the extent and pattern of cytosine methylation, which may influence the conformation of DNA and impair the binding of DNA-interacting proteins such as topoisomerase II and the kinetochore complex (Stopper, 1997). Smith et al.(1998) studied the micronucleation of chromosome 16 in human lymphoblastoid TK6 cells by 2,6-diaminopurine, another nucleotide analogue known to enhance the under-condensation of the paracentromeric heterochromatin in chromosomes 1, 9 and 16. Chromosome 16 classical satellite DNA (together with pancentromeric DNA) was preferentially found in the induced MN. The MN were likely to harbour a fragment resulting from breakage within the centromeric area of chromosome 16. Likewise, DNA from chromosomes 1 and 9 was over-represented in human lymphocyte MN after an in vitro treatment with idoxuridine, which induces a specific decondensation at chromosome 9q12 and (less so) at 1q12 (Fauth and Zankl, 1999). Similar findings were obtained in studies with mitomycin C, which also induces under-condensation of chromosomes 1 and 9; 62–69% of all MN in mitomycin C-treated lymphocytes contained DNA from chromosome 9 (Fauth et al., 2000). The MN induced by mitomycin C were devoid of the -satellite of chromosome 9, suggesting that most of the breaks were induced in the pericentromeric heterochromatin (Kusakabe et al., 1999).
It is unclear how the under-condensed chromosome regions would break and form MN. Smith et al.(1998) suggested that the breakage results from torsional strain generated by movement of the chromatids towards opposite poles during anaphase. 5-Azacytidine-treated mouse L5178Y cells examined by ultravital UV microscopy showed normal metaphase arrangements, but chromatid separation in anaphase was disturbed with the formation of thin chromatin bridges which were occasionally ruptured, generating MN (Stopper et al., 1993; Stopper, 1997).
When chromosome painting (chromosomes 1, 7, 11 and 14) was used to characterize MN induced by 
-radiation in various human cell lines, MN with DNA from chromosome 7 appeared to be under-represented (Slavotinek et al., 1996). In human diploid skin fibroblasts, the incorporation of DNA from chromosomes 2 and 7 in MN after -irradiation was significantly greater than expected by chromosome size, while smaller chromosomes (11 and 16) were micronucleated as expected (Walker et al., 1996). In cultured human lymphocytes, painting of chromosomes 1, 7, 11, 14, 17 and 21 suggested that DNA from these chromosomes is found in -ray-induced MN according to the DNA content of each chromosome, supporting a random model for radiation-induced chromosome damage (Fimognari et al., 1997). Wuttke et al.(1997) could not see any difference between the expected and observed inclusion of chromosome 2 and 7 material (detected by chromosome painting) in human lymphocyte MN after X-ray treatment. Thus, there does not presently seem to