FUMAGILLIN-INDUCED CHROMOSOME ABERRATIONS IN MOUSE BONE-MARROW CELLS

The genotoxicity of fumagillin was evaluated assessing the mitotic index and chromosome aberrations in mouse bone-marrow cells. Mice were given fumagillin orally in doses 5, 10 and 20 mg/kg bw. All doses significantly (p<0.001) reduced the mitotic index. The medium and maximum doses led to significant (p<0.01 and p<0.001, respectively) increases in numeric aberrations. The highest dose induced both structural and numeric aberrations (p<0.001), and insertions on the first pair of autosomes that were amplified in the 1C and 1E regions. These results pointed to the genotoxic potential of fumagillin in the range of medium and maximum doses applied.


INTRODUCTION
Fumagillin is an antibiotic derived from the fungus Aspergillus fumigatus.It has proved most effective in suppressing cryptosporidiosis and microsporidiosis caused by Enterocytozoon bieneusi which can be fatal in HIV-infected persons (Molina et al., 2000;Conteas et al., 2000).Due to its antiparasitic efficacy, fumagillin has also been widely applied in veterinary medicine against microsporidiosis of bees and fish (Katznelson and Jamieson, 1952;Bailey, 1953;El-Matbouli and Hoffman, 1991;Morris et al., 2003).
Fumagillin is very stable in honey (Furgala, 1962) where it was detectable after having been kept at 80 o C for 35 days (Assil and Sporns, 1991).Mixed with syrup it is very effective in suppressing Nosema in hibernating honeybee colonies.However, it has not been effective against dormant spores of Nosema apis, or against Nosema ceranae.
To date some adverse effects of fumagillin have been discovered.After the treatment of Nosema apis-infected bees with fumagillin the electron-density of the mitochondrial matrix in corpora allata increased and the dimensions of the mitochondria diminished in comparison with untreated infected bees (Liu, 1990a).Fumagillin produced certain effects on the secretion granules of the hypopharyngeal glands of bees which increased and had a homogeneous structure, which is explained by the changes in the secretory activities of glands (Liu, 1990b).It was confirmed that fumagillin largely increased mortality in bees as well as the number of fungi (Rada et al., 1997).

R E T R A C T E D
Data on the genotoxic effects of fumagillin obtained from in vitro studies showed discrepancy, being either positive (Stoltz et al., 1970;Stanimirović et al., 1999;Stevanović et al., 2000 and2008;Kulić 2006) or negative (Purchase et al., 1978;Mortelmans et al., 1986;Heil, 1996).In addition, there are insufficient data on the genotoxic effects of fumagillin in vivo.
Having considered the aforementioned, there are plausible reasons for testing the genotoxic effects of fumagillin in vivo and in vitro; firstly, because the effects were different in various tests; secondly, they depended largely on the doses of fumagillin and, thirdly, there interactions of fumagillin with endogenous and exogenous factors have been reported (Ames, 1989;Albertini et al., 2000;Norppa, 2003;Stanimirović et al., 2005Stanimirović et al., , 2007;;Stevanović et al., 2008).
The aim of this work was to evaluate the possible in vivo genotoxic effects of fumagillin at doses lower than those used in bee-keeping.
Fumagillin does not dissolve readily in water.For that reason, it was stirred in a small quantity of water (≥32-35°C) until it turned into a paste and a water-sugar syrup was gradually added.The solution was administered over seven consecutive days.
Five groups of mice were tested: three groups treated with fumagillin, the positive and negative control.Each group was comprised of six sixmonth-old BALB/c male mice weighing approximately 20 g.The animals were kept in unchanged conditions under 12/12-h light-dark periods at a constant temperature (21°C) with free access to food and water.The animals in the negative control group were treated with water-sugar syrup (1:1).Cyclophosphamide, a well-known clastogene and mutagene (Anderson et al., 1995), was used as the positive control and administered i.p. at a dose of 15 mg/kg body weight for seven consecutive days.
Cytogenetic analyses were carried out on the bonemarrow cells obtained from the femur and tibia according to Hsu and Patton (1969), modified by Zimonjić et al. (1990).Preparations were flame-dried and stained with Giemsa solution (Sigma Chemical Co., St. Louis, MO).G-banding was carried out by the tripsin method of Seabright (1971) and Ronne (1991).Chromosomes and sets of chromosomes were identified on the basis of criteria established by the Committee on Standardized Genetic Nomenclature for Mice (1979) and Cowell's photoatlas of mice chromosomes (Cowell, 1984).

R E T R A C T E D
The mitotic index was determined in 1000 cells per treatment.Six-hundred well spread metaphases were analyzed for the presence of chromosome aberrations per each treated group.Statistical analysis was performed by Statistica 6.0 software programme, ANOVA, Student's t-test and LSD-test.

RESULTS
The genotoxic effects of fumagillin in the mouse bone-marrow cells were analyzed assessing the mitotic index (MI) and numeric (CNA) and structural chromosome aberrations (CSA).The doses tested were 5.0, 10.0 and 20.0 mg/kg bw.
All doses of fumagillin (Table 1) induced a significant decrease (p<0.001) in the mitotic index (MI=3.41±0.04,3.11±0.01,and 2.21±0.02%,respectively) in comparison with the negative control (MI=5.94±0.01%).In addition, it was shown that in certain doses fumagillin is capable of provoking both numeric and structural chromosome aberrations.
At the lowest dose of 5 mg/kg bw, fumagillin did not seem to influence either structural or numeric chromosome aberrations in mouse bonemarrow cells, since the mean number of gaps, acentrics and insertions remained similar to the numbers in the negative control; the same was true for aneuploidy and poliploidy (Table 2).
Significantly increased frequencies (p<0.01 or p<0.001) of numerical chromosome aberrations (aneuploidies and poliploidies) were observed both in the medium (10 mg/kg bw) and the highest (20 mg/kg bw) dose of fumagillin.
Twice the minimum dose of fumagillin (10 mg/kg bw) resulted in aneuploidies only.Their average number rose to 6.37±0.92,which was significantly more (p<0.01)than in the negative control (5.00±0.76).Poliploidy was not more frequent than expected.
Fumagillin administered to BALB/c mice in the dose of 20 mg/kg for seven consecutive days had major consequences on the frequency of both numerical and structural chromosome aberrations.The mean number of numerical chromosome aberrations, both aneuploidies and poliploidies rose considerably (p<0.001) to 31.75±1.28and 5.37±0.74,respectively.
Structural chromosome aberrations (gaps, acentrics and insertions) were noticeably more frequent in comparison to the negative control only in the highest experimental dose of fumagilline (Table 2).
The average number of gaps more than doubled, increasing from baseline 2.00±0.74 to 5.75±0.89(p<0.001).The frequency of acentric chromosomes rocketed from a negligible (0.75±0.71) in the negative control to 4.37±0.74(p<0.001).The increase in insertions was dramatic (p<0.001)having reached 3.25±0.71on average.After completing the G-band analysis and chromosomal identification, the presence of one large chromosome, unusual for the mouse karyotype, was observed (Figure 1).It was clear that the aberrant chromosome had a surplus of two insertions when compared to the normal autosome from this pair (Figure 2).These insertions were located in the region 1C5 (one segment) and 1E (three segments, between 1E3 and 1E4 bands) (Figure 3).
When taking into consideration total cytogenetic changes, a significant increase (p<0.01) in comparison to the negative control (8.00±1.20)was noticed in the medium (11.25±1.83)and in the highest (50.50±1.93)dose group (p<0.001).

R E T R A C T E D
there is very little information on the genotoxic effects of fumagillin in vivo, which was obtained with much higher doses (Stevanović et al., 2006;Stanimirović et al., 2007).Having analyzed the aforementioned it is clear that further evaluation of the genotoxic effects of fumagillin is necessary.
In our in vivo study the MI and CA were monitored in bone-marrow cells of BALB/c mice treated with selected doses of fumagillin (5, 10, and 20 mg/kg bw).The results showed that all experimental doses induced a significant decrease (p < 0.001) in MI in comparison with the negative control.
As expected, treatment with cyclophosphamide resulted in an increase in the MI.These results are in accordance with the findings of many other authors who examined the antiproliferative effects (anti-angiogenic effects) of fumagillin (Ingber et al., 1990;Wang et al., 2000;Mazzanti et al., 2004).The molecular target of fumagillin and its analogue TNP-470 is methionine aminopeptidase-2 (MetAP-2) (Griffith et al., 1998;Sin et al., 1997;Liu et al., 1998).Fumagillin binds MetAP-2 on His-231, inactivating the enzyme.MetAP-2 removes the N-terminal methionine from most proteins involved in cell-cycle regulation as part of the translocation process, so its inhibition results in cell-cycle arrest and apoptosis (Fardis et al., 2003).These findings provided a starting point for the rational design of novel fumagillin analogues with fewer side-effects (Fardis et al., 2003).
The results of the present study point to the capability of fumagillin at certain doses to cause numerical and structural chromosome aberrations in vivo.The highest tested dose (20 mg/kg bw) resulted in the increase in the frequencies of aneuploidies, poliploidies, gaps, acentrics and insertions (p<0.001).These results are in accordance with the previous in vivo findings of Stanimirović et al. (2007), although, the doses of fumagillin investigated by him were much higher (25, 50 and 75 mg/kg bw) than those in the current work (5, 10 and 20 mg/kg bw).Our in vivo findings of chromosome aberrations agree with the previous in vitro results of Stanimirović et al., (1999), Stevanović et al. (2000 and2008) and Kulić (2006), which claimed that fumagillin significantly increased the frequencies of structural chromosome aberrations.
Indeed, in vitro research on fumagillin has not revealed indicators of its genotoxic potential in the Salmonella test using Salmonella typhimurium types TA 98, TA 100, TA 1535 and TA 1538 (Purchase et al., 1978;Mortelmans et al., 1986).On the other hand, in the research carried out by Stoltz et al., (1970) human lymphocytes incubated in the presence of dicyclohexylamine sulfate for 5 and 24 hours showed a marked dose-dependent increase in aberrations.
The National Toxicology Program (2006a, b) has reported positive genotoxic effects of dicyclohexylamine nitrate in the Salmonella test (National Toxicology Program 2006a) while dicyclohexylamine did not exhibit genotoxicity (National Toxicology Program 2006b).Also, there are certain data on the genotoxic effects of secondary metabolites (gliotoxin and verruculogen) of Aspergillus fumigatus which fumagillin is derived from.Gliotoxin causes changes in the DNA T E D (Golden et al., 1998) and it appeared to be genotoxic in in vitro test systems (Niemien et al., 2002); meanwhile, verruculogen produced effects in Salmonella/microsomal mutagenicity assays (Sabater-Vilar et al., 2003).
The results of our research clearly showed an increase in the frequency of chromosome aberrations provoked by fumagillin, especially at the dose of 20 mg/kg bw, which significantly influenced the amplification of certain chromosome regions on one of the chromosomes of the first pair of autosomes in a certain number of the animals tested.By G-band analyses we showed that these insertions were located in the region 1C5 (one segment) and 1E (three segments, between 1E3 and 1E4 bands).These results are in accordance with the findings of Kulić (2006), who investigated a synthetic analogue of fumagillin (dicyclohexylamine-ET).Also, the same type of insertion was found when investigating some other substances, e.g.levamisole hydrochloride in vivo in rats (Stanimirović et al., 1998).These authors claim that the insertions provoked by levamisole hydrochloride were observed on the q arm of the first pair of autosomes between q21 and q22, where the q21a derived from endogenous duplication or amplification of DNA region was inserted.These findings point to the instability of the first pair of autosomes in mice and rats which was already described by Agulnik et al. (1988Agulnik et al. ( , 1990) ) and Stanimirović et al. (1995Stanimirović et al. ( , 1998)).For a more certain explanation of the reason for the insertion further investigations are necessary.
Finally, although there is no reliable information regarding fumagillin residue levels in food except those of Mladjan and Jović (2000) and Kulić (2006), our results concerning the increased frequencies of chromosome aberrations (poliploidy, aneuploidy, gaps, acentrics and insertions) induced by fumagillin lead to the conclusion that fumagillin residues in food may have genotoxic effects that could increase the risk for chromosome aberrations and cancer (Stanimirović et al., 2006(Stanimirović et al., , 2007;;Steva-nović et al., 2006Steva-nović et al., , 2008)).Moreover, beekeepers, who are occupationally exposed to fumagillin, may also be at genotoxic risk.There is an urgent necessity for their compulsory education concerning consumers' safety.Similar caution should be taken with patients treated with fumagillin against microsporidia.

R E T R A C
T E D

Figure 2 .
Figure 2. G-band karyogram of a BALB/c mouse treated with fumagillin (20 mg/kg bw) with heterozygous insertions in 1C5 and 1E regions of the first pair of autosomes

Figure 3 .
Figure 3. Idiogram of the first pair of autosomes of a BALB/c mouse (A-normal chromosome, B-chromosome with insertions in 1C5 and 1E region)

Table 2 .
Cytogenetic parametres in the cells of bone marrow in BALB/c mice in control and experimental groups of animals treated by fumagillin.