Radiotoxicity of 125I in mammalian cells

The radiotoxicity of 125I in Chinese hamster V79 lung fibroblasts has been studied following extracellular (Na125I), cytoplasmic [125I]iododihydrorhodamine (125I-DR), and nuclear (125IUdR) localization of the radionuclide. Exposure of the cells for 18 h to Na125I (less than or equal to 7.4 MBq/ml) had no effect on survival. A similar exposure to 125I-DR produced a survival curve with a distinct shoulder and with a mean lethal dose (D37) of 4.62 Gy to the nucleus. While this value compares well with the 5.80 Gy X-ray D37 dose, it is in contrast to the survival curve obtained with DNA-bound 125IUdR which is of the high LET type and has a D37 of 0.80 Gy to the nucleus. Furthermore, when the uptake of 125I into DNA is reduced by the addition of nonradioactive IUdR or TdR to the medium and the survival fraction is determined as a function of 125I contained in the DNA, a corresponding increase in survival is observed. This work demonstrates the relative inefficiency of the Auger electron emitter 125I when located in the cytoplasm or outside the cell. It indicates that a high dose deposited within the cytoplasm contributes minimally to radiation-induced cell death and that radiotoxicity depends not upon the specific activity of IUdR but upon the absolute amount of 125I that is associated with nuclear DNA.     

Induction of sperm head abnormalities by incorporated radionuclides: dependence on subcellular distribution, type of radiation, dose rate, and presence of radioprotectors

In contrast to the biological effects caused by exposure to external beams of radiation, the effects of tissue-incorporated radionuclides are highly dependent on the type of radiation emitted and on their distribution at the macroscopic, microscopic, and subcellular levels, which are in turn determined by the chemical nature of the radionuclides administered. Induction of abnormalities of sperm heads in mice is investigated in this work after the injection of a variety of radiochemicals including alpha emitters. When the initial slopes of the dose-response curves are used to compare the relative biological effectiveness (RBE) of different radiocompounds, the alpha particles emitted in the decay of 210Po are more effective than Auger electrons emitted by 125I incorporated in the DNA of the spermatogonial cells, and both emissions are more effective than X rays. It is also shown that the Auger emitters (125I, 111In) distributed in the cell nucleus are more efficient in producing abnormalities than the same radionuclides localized in the cytoplasm. These findings are consistent with our earlier observations, where spermatogonial cell survival is assayed as a function of the testicular absorbed dose. Further, chronic irradiation of testis with gamma rays from intratesticularly administered 7Be is about three times more effective in causing abnormalities than a single acute exposure to 120-kVp X rays. The resulting RBE values correlate well with our data on sperm head survival with the same radiocompounds. Finally, the radioprotector cysteamine, when administered in small, nontoxic amounts, significantly reduces the incidence of sperm abnormalities from alpha-particle radiation as well as emissions from 125I incorporated into DNA, the dose reduction factors being 10 and 14, respectively.     

Toxicity of DNA-incorporated iodine-125: quantifying the direct and indirect effects

To understand the biophysical mechanism(s) underlying the induction of cell death by the decay of the Auger electron emitter iodine-125 in DNA, Chinese hamster V79 lung fibroblasts were labeled with 5-[(125)I]iodo-2'-deoxyuridine ((125)IdU) for two doubling times and frozen and stored at -135 degrees C in the presence of 0.26-3.0 M dimethyl sulfoxide (DMSO), which acts simultaneously as a cryoprotector and a hydroxyl radical scavenger. After the accumulation of (125)I decays, the cells were defrosted and their survival was determined. Within the range of the number of decays examined (up to 470 disintegrations per cell), the survival curves are exponential. The dependence of the D(37) on DMSO concentration is triphasic and seems to reach a plateau at approximately 1.3 M. By extrapolating to infinite DMSO concentration, we estimate the D(37) for maximal hydroxyl radical scavenging to be 411 +/- 36 disintegrations per cell. To determine the D(37) in the absence of DMSO, we extrapolate the D(37) curve to zero concentration, and a D(37) of 54 +/- 5 disintegrations per cell is obtained. The maximal dose modification factor, calculated as the ratio of the D(37) at infinite DMSO concentration (i.e. direct effects only) to the D(37) at zero DMSO concentration (i.e. direct and indirect effects), is 7.6 +/- 1.0. By inference, approximately 90% of the radiotoxic effects of DNA-incorporated (125)I are due to indirect mechanisms.     

The radiation dose to cells in vitro from intracellular indium-111

Most of the radionuclides used in nuclear medicine emit low energy Auger electrons following radioactive decay. These emissions, if intracellular, could irreparably damage the radiosensitive structures of the cell. The resulting radiation dose, which is a measure of biological damage in the affected cell, could be many times the average radiation dose to the associated organ. In this series of experiments, the radiation dose to the nucleus of a chinese hamster V79 cell was determined for the intracellular radiopharmaceutical 111indium-oxine. Assuming the cell nucleus to be the radiosensitive volume, the radiation dose would be primarily due to the low energy Auger electrons. A much smaller dose would be absorbed from the penetrating X- and gamma-rays and internal conversion electrons released from other radiolabelled cells in the culture. The radiation dose to the cell from the intranuclear decay of 111In was empirically established from cell survival studies to be 3.5 mGy/decay, using cobalt-60 as a reference radiation. The average dose to V79 cells from extracellular 111In (i.e., from 111In located outside the target cell) was calculated to be 5.8 pGy/decay. This suggests that for an intracellular radiopharmaceutical, the radiation dose of consequence would be delivered by the low energy Auger electrons. In contrast, Auger electrons from an extracellular radiopharmaceutical could not directly damage the cell nucleus and therefore would not contribute to the radiation dose.     

Biological damage from the Auger effect,possible benefits

Summary Decay of radioactive isotopes by K-capture leads to the Auger effect and results in the loss of several orbital electrons and the emission of X-rays. Whereas radiation effects are produced from the emitted electrons, the consequences of the Auger effect are strictly localized to the site of the decaying nuclide.The paper reviews the biological consequences of the decay of¹²⁵I which produces the Auger effect. Nearly all data were obtained from DNA labeled with¹²⁵I-5-iodo-2-deoxyuridine (IUdR) in bacteria and mammalian cells. Parameters of effects were cell death, DNA strand breaks, and mutation induction. In order to recognize in a cell the contribution from the Auger effect and that of absorbed radiation, experimental data are analysed in terms of the specific energy for the nuclear volume which contains the isotope.The data indicate that decay of¹²⁵I is far more toxic than is expected on the basis of absorbed dose to the labeled nucleus. Moreover, it is emphasized that the toxicity of the¹²⁵I decay is largely determined by events immediately localized to the site of decay.Because the consequences of the Auger effect are strictly localized to the molecular site of the decay,¹²⁵I and perhaps other nuclides decaying by K-capture promise to be interesting tools in cell biology and molecular biology. First data on the Auger effect as a tool are summarized.It appears that recognizable biological damage is only observed when the Auger effect takes place in vitally important molecules, an example of which is DNA.Dedicated to Prof. Dr. H. Muth on the occasion of his 60th birthday.     

Using Hoechst 33342 to target radioactivity to the cell nucleus

We have explored the use of Hoechst 33342 (H33342) to carry radioactivity to the cell nucleus. H33342 enters cells and targets DNA at adenine-thymine-rich regions of the minor groove. Considerable membrane blebbing and ruffling occur in CHO cells within minutes after its addition to the culture medium in micromolar quantities. Blue vesicles are apparent in the cell cytoplasm, and by 30 min the nuclei are stained dark blue. Upon its binding to DNA, a visible emission shift of the dye can be observed with fluorescence microscopy. We have radioiodinated (125I) H33342 and specifically irradiated nuclear DNA by incubating CHO cells with 125I-H33342 at 37 degrees C and accumulating 125I decays at -90 degrees C. At various times, the cells are thawed and assayed for survival (clonogenicity) and DSB (gamma-H2AX) formation. 125I-H33342 decay leads to a monoexponential decrease in cell survival with a D0 of 122 125I decays per cell and a linear increase in DNA DSB induction (equivalent to 15 gamma-H2AX foci/cell). Cell death is not modified by the radioprotective effects of H33342 because we use considerably lower concentrations than those that provide a slight protection against gamma radiation. We conclude that cell killing by 125I-H33342 and the induction of gamma-H2AX foci are highly correlated.     

Relative biological effectiveness of accumulated 125IdU and 125I-estrogen decays in estrogen receptor-expressing MCF-7 human breast cancer cells

The therapeutic potential for delivering a cytotoxic dose of radiation (using the decay of Auger-electron emitters) to the cell nucleus of cancer cells that express estrogen receptors (ERs) by radiolabeled estrogen was investigated in the ER-expressing human breast cancer cell line, MCF-7. The radiolabeled estrogen/ER complex irradiates the cell nucleus by binding specific DNA sequences called estrogen response elements (EREs). Cell clonogenicity and induction of DNA double-strand breaks (DSBs) by gamma radiation or accumulation of (125)I-iododeoxyuridine ((125)IdU) or E-17alpha[(125)I]iodovinyl-11betamethoxyestradiol ((125)IVME2) decays were determined. MCF-7 cells were efficiently killed by accumulation of (125)IdU (D(0) = 30 decays per cell) and (125)IVME2 decays (D(0) = 28 decays per cell). DNA DSBs were induced by the accumulation of (125)IdU (approximately 3750 decays per cell required to reduce the mean value of the elution profile to 50%) or (125)IVME2 decays (approximately 465 decays per cell required to reduce the mean value to 50%). For survival of MCF-7 cells after gamma irradiation, the D(0) was 1 Gy, and approximately 65 Gy was required to reduce the mean value to 50% for induction of DSBs. The RBE values for cell killing and induction of DSBs by (125)IVME2 relative to gamma radiation were 4.8 and 18.8, respectively. The RBE values for cell killing and induction of DSBs by (125)IdU relative to gamma radiation were 4.5 and 2.3, respectively. Cell killing in a manner similar to that induced by high-LET radiation and the high RBE for induction of DSBs by (125)IVME2 in the ER-expressing MCF-7 cells provide a biological rationale for the use of Auger electron-emitting radionuclides covalently bound to estrogen to deliver a cytotoxic dose of radiation to ER-positive cancers.     

A Comparison of the Killing of Cultured Mammalian Cells Induced by Decay of Incorporated Tritiated Molecules at -196°C

The killing efficiency of tritium disintegrations in frozen mammalian cells labeled with tritiated uridine, histidine, and lysine was compared with the killing efficiency of incorporated tritiated thymidine. In each case, the distribution of tritium in the cells was determined by chemical fractionation as well as by radio-autography. Of all tritium disintegrations, by far the most effective were those occurring in DNA molecules within frozen cells; such incorporated tritium has a killing efficiency of 0.006. When cells were incubated with tritiated uridine for 10 min to label nuclear RNA, the killing efficiency was 0.0015. When the cells were pulse labeled with tritiated uridine and permitted to grow in nonradioactive media for 10 hr before freezing in order to incorporate tritium into cytoplasmic RNA, the killing efficiency was reduced to 0.0010. The results suggest that decay of tritium in nuclear RNA is more effective than that in cytoplasmic RNA. When the cells were labeled with tritiated histidine or lysine for 30 min, tritium atoms were found mainly in the acid soluble rather than in the protein fraction and the killing efficiency in each case was approximately 0.0007. The results of these suicide experiments indicate that the killing efficiency of tritium disintegrations depends on where tritium is located within the cells. Tritium disintegrations in the nucleus are more effective in killing the cell than that in cytoplasm; and tritium disintegrations on DNA in the nucleus is more effective in killing the cell than that of nuclear RNA.     

Kinetics of uptake, retention, and radiotoxicity of 125IUdR in mammalian cells: implications of localized energy deposition by Auger processes

The kinetics of uptake, retention, and radiotoxicity of 125IUdR have been studied in proliferating mammalian cells in culture. The radioactivity incorporated into the DNA is directly proportional to the duration of incubation and to the extracellular concentration of 125I. The rate of proliferation of cells is related to the intracellular radioactive concentration and is markedly reduced at medium concentrations greater than or equal to 0.1 mu Ci/ml. At 37% survival the high LET type cell survival curve is characterized by an uptake of 0.035 pCi/cell, and the cumulated mean lethal dose to the cell nucleus is about 80 rad compared to 580 rad of X-ray dose for this cell line. The strong cytocidal effects of the decay of 125I correlate with localized irradiation of the DNA by the low energy Auger electrons.     

The question of relative biological effectiveness and quality factor for auger emitters incorporated into proliferating mammalian cells

The problem of determining RBE values for Auger emitters incorporated into proliferating mammalian cells is examined. In general, the reference radiation plays a key role in obtaining experimental RBE values. Using survival of cultured Chinese hamster V79 cells as the experimental model, new data are provided regarding selection of a reference radiation for internal Auger emitters. These data show that gamma rays delivered acutely (137Cs) are more than twice as lethal as gamma rays delivered chronically with an exponentially decreasing dose rate (99mTc). The results confirm that the reference radiation should be delivered chronically in a manner consistent with the extended exposure received by the cells in the case of incorporated radionuclides. Through a direct comparison of the radiotoxicity of Auger emitters and alpha emitters, the high RBE values reported for DNA-bound Auger emitters are confirmed. These studies reveal that the DNA binding compound [125I]iododeoxyuridine (125IdU) is about 1.6 times more effective in killing V79 cells than 5.3 MeV alpha particles from intracellularly localized 210Po-citrate. In addition, toxicity studies with the radiochemicals 125IdU and [125]-iododeoxycytidine (125IdC) establish the equivalence of the radiosensitivity of thymine and cytosine base sites in the DNA. In view of these results, and information already available, the question of establishing quality factors for Auger emitters is considered. Finally, a method for calculation of the dose equivalent for internal Auger emitters is advanced.     

Inhomogeneity of the nucleus to 125IUdR cytotoxicity

Synchronized suspension cultures of Chinese hamster ovary (CHO) cells were used to determine the lethal effects produced by the decay of 125I incorporated into different subfractions of the nuclear genome. Such a shift in nuclear incorporation pattern was achieved by using the drug aphidicolin, which inhibits 95% of all nuclear DNA synthesis, is nontoxic to cells in a colony-forming assay, and does not modify the radiation response of CHO cells to X irradiation. In addition to shifting incorporation of 125I to only 5% of the nuclear genome, both nuclease digestions to characterize the molecular location of 125I and electron microscope autoradiography show an inhomogeneous distribution of sites of 125I incorporation in the presence of 5 micrograms/ml aphidicolin. These data in combination with survival curves of CHO cells labeled with 125I-iododeoxyuridine (125IUdR) either with or without aphidicolin showed a dramatic change in the survival response (DO: 30 decays/cell and 96 decays/cell, respectively). It is concluded, therefore, that the nucleus is not a homogeneous target for radiation-induced cell death because when subfractions of the nuclear genome are labeled, radically different levels in cell survival are obtained.     

Mammalian cells: Damage in late replicating DNA as the most efficient cause of reproductive death

Chinese hamster cells were synchronized by mitotic shake off, labeled in the first S period with ¹²⁵IUdR, cooled to 4 °C in the G2 stage and then stored up to 4 days to accumulate damage due to ¹²⁵I disintegrations in cell DNA. There was a large difference in the efficiency of induction of reproductive death when damage accumulated in the DNA, which replicated in the second half of the DNA synthesis period, was compared with damage accumulated in the DNA, which replicated in the first half of the DNA synthesis period. Damage accumulated in the late replicated DNA appears to be the most critical. This result suggests that the mammalian cell nucleus is not homogeneous with respect to the damaging events leading to reproductive death and may stress the importance for cell survival of the integrity of the late-replicating, heterochromatic DNA near the nucleus membrane.     

Fragmentation of chromatin with 125I radioactive disintegrations.

The DNA in Chinese hamster cells was labeled first for 3 h with [3H]TdR and then for 3 h with [125I]UdR. Chromatin was extracted, frozen, and stored at -30 degrees C until 1.0 X 10(17) and 1.25 X 10(17) disintegrations/g of labeled DNA occurred for 125I and 3H respectively. Velocity sedimentation of chromatin (DNA with associated chromosomal proteins) in neutral sucrose gradients indicated that the localized energy from the 125I disintegrations, which gave about 1 double-strand break/disintegration plus an additional 1.3 single strand breaks, selectively fragmented the [125I] chromatin into pieces smaller than the [3H] chromatin. In other words, 125I disintegrations caused much more localized damage in the chromatin labeled with 125I than in the chromatin labeled with 3H, and fragments induced in DNA by 125I disintegrations were not held together by the associated chromosomal proteins. Use of this 125I technique for studying chromosomal proteins associated with different regions in the cellular DNA is discussed. For these studies, the number of disintegrations required for fragmenting DNA molecules of different sizes is illustrated.     

Interaction of measles virus vectors with Auger electron emitting radioisotopes

A recombinant measles virus (MV) expressing the sodium iodide symporter (NIS) is being considered for therapy of advanced multiple myeloma. Auger electrons selectively damage cells in which the isotope decays. We hypothesized that the Auger electron emitting isotope 125I can be used to control viral proliferation. MV was engineered to express both carcinoembryonic antigen and NIS (MV-NICE). Cells were infected with MV-NICE and exposed to 125I with appropriate controls. MV-NICE replication in vitro is inhibited by the selective uptake of 125I by cells expressing NIS. Auger electron damage is partly mediated by free radicals and abrogated by glutathione. In myeloma xenografts, control of MV-NICE with 125I was not possible under the conditions of the experiment. MV-NICE does not replicate faster in the presence of radiation. Auger electron emitting isotopes effectively stop propagation of MV vectors expressing NIS in vitro. Additional work is necessary to translate these observations in vivo.     

Radiotoxicity of an 125I-labeled DNA intercalator in mammalian cells

To explore the effect of the Auger electron emitter 125I attached to a DNA intercalator, we have synthesized 125I- and 127I-labeled 3-acetamido-5-iodoproflavine (AIP) and have examined the uptake, intracellular distribution, and radiotoxicity of A125IP in Chinese hamster V79 cells. After incubation with AIP, the nuclei of V79 cells become fluorescent. Uptake of A125IP is directly proportional to its extracellular radioactive concentration and reaches a plateau at about 10 h. Of the cell-associated radioactivity, 60% is retained by the cells after extensive washing. When the survival of V79 cells is plotted as a function of radioactive cell content, the curve has no shoulder with a mean lethal dose (DN) of about 1.3 Gy to the cell nucleus. Because the DN of these cells when irradiated with 250 kVp X rays is 5.8 Gy, the relative biological effectiveness (RBE) of A125IP is about 4.5. The dependence of the RBE values on the localization of the Auger emitter is discussed on the basis of our extended studies on the same cell line.     

Foreign protein can be carried into the nucleus of mammalian cell by conjugation with nucleoplasmin

In studies on the specific migration of macromolecules across the nuclear envelope, a karyophilic protein was injected into the cytoplasm of cultured cells and its subsequent location in the cell was examined. Nucleoplasmin of frog nuclear protein was used for this experiment. When [125I]nucleoplasmin was introduced into the cytoplasm of mammalian cells (human and mouse) by red blood cell-mediated microinjection, it rapidly accumulated in the nucleus. When nucleoplasmin conjugated with [125I]IgG against chromosomal protein was introduced similarly, it also accumulated rapidly in the nucleus, and reacted with its antigen inside the nucleus. On the contrary, when IgG alone or IgG conjugated with BSA were introduced, they did not migrate from the cytoplasm into the nucleus. These findings imply that the migration of macromolecules from the cytoplasm to the nucleus does not depend only on their molecular size but also on a specific transport mechanism, and that karyophilic proteins may act as useful carriers in the transfer of exogenous proteins into the nucleus.     

Efficient mutation induction by 125I and 131I decays in DNA of human cells

To examine the role of radiation energy deposition in DNA on cellular effects, we investigated the ability of 125IdUrd and 131IdUrd to kill cells and induce mutations at the hprt locus. We employed human lymphoblastoid cells proficient (TK6) or deficient (SE30) in the ability to incorporate a thymidine analog into DNA by way of the thymidine kinase (TK) scavenger pathway. Iodine-125 releases a shower of low-energy Auger electrons upon decay which deposit most of their energy within 20 nm of the decay site, whereas 131I is a high-energy beta/gamma emitter that is generally considered to emit sparsely ionizing radiation. Although 125IdUrd incorporated into cellular DNA was very effective at producing toxic and mutagenic effects in TK6 cells, virtually no effect was seen in TK-deficient cells incubated with similar levels of 125IdUrd in the extracellular medium. In response to 131IdUrd treatment, 0.45 X 10(-6) mutants were induced per centigray dose deposited within the nucleus in TK-proficient cells, whereas few mutations were induced in TK-deficient cells at doses up to 38 cGy from 131I decays occurring in the medium. The differences in biological response between TK6 and SE30 cells cannot be explained by differential radiosensitivity or IdUrd sensitization of the cell lines involved. We conclude that both 125I and 131I decays occurring while incorporated into DNA are more effective at inducing cell killing and mutations in human cells than either nonincorporated decays or low-LET radiations. These results suggest that localized energy deposition is an important factor in producing biologically important damage by both of these isotopes, and that residual lesions following the decay of DNA-incorporated radioisotopes may contribute to the toxic and mutagenic effects observed in TK-proficient cells. Furthermore, they emphasize that certain beta/gamma-emitting isotopes such as 131I may be particularly hazardous when incorporated into DNA.     

Block copolymer micelles target Auger electron radiotherapy to the nucleus of HER2-positive breast cancer cells

Intracellular trafficking of Auger electron emitting radionuclides to perinuclear and nuclear regions of cells is critical to realizing their full therapeutic potential. In the present study, block copolymer micelles (BCMs) were labeled with the Auger electron emitter indium-111 ((111)In) and loaded with the radiosensitizer methotrexate. HER2 specific antibodies (trastuzumab fab) and nuclear localization signal (NLS; CGYGPKKKRKVGG) peptides were conjugated to the surface of the BCMs to direct uptake in HER2 expressing cells and subsequent localization in the cell nucleus. Cell uptake and intracellular distribution of the multifunctional BCMs were evaluated in a panel of breast cancer cell lines with different levels of HER2 expression. Indeed cell uptake was found to be HER2 density dependent, confirming receptor-mediated internalization of the BCMs. Importantly, conjugation of NLS peptides to the surface of BCMs was found to result in a significant increase in nuclear uptake of the radionuclide (111)In. Successful nuclear targeting was shown to improve the antipoliferative effect of the Auger electrons as measured by clonogenic assays. In addition, a significant radiation enhancement effect was observed by concurrent delivery of low-dose MTX and (111)In in all breast cancer cell lines evaluated.     

Estrogen receptor-mediated cytotoxicity using iodine-125

Auger effects from 125I decay are singularly damaging if localized in DNA as the thymidine analogue 125I-iododeoxyuridine (125IUdR). Recent experience with steroid sex hormones extends these observations by demonstrating cytotoxicity in sites other than the DNA backbone. We have compared the cytotoxicity in human MCF-7 breast cancer cells of 125IUdR, 125I-iodotamoxifen, a nonsteroidal antiestrogen that is translocated from the cytoplasm to the nucleus of receptor containing cells, and 125I-iodoantipyrine, a biological indicator of the body water space. Cytotoxicity is critically dependent upon subcellular localization.     

Comparison of cell inactivation by Auger electrons using the two reagents 4-[123I]iodoantipyrine and [123I]NaI

The Auger electron emitter ¹²³I was examined in the form of 4-[¹²³I]iodoantipyrine and as [¹²³I]NaI for its effectiveness in killing cells of different sensitivity to photon irradiation. Micronucleus assays showed that 4-[¹²³I]iodoantipyrine is 2–3 times more effective in cell inactivation than [¹²³I]NaI. This can be attributed to the fact that antipyrine, for reason of its lipid solubility, can enter cells and can reach the nucleus, whereas [¹²³I]NaI is excluded from the cytoplasm. In the nucleus Auger decay is conceivably located on the DNA where it may invoke high-LET irradiation damage. Irradiation damage by [¹²³I]NaI is by long range Auger and internal conversion electrons and hence less densely ionising. Results of the present study demonstrate, however, that the enhancement of micronuclei frequency (MNF) seen with 4-[¹²³I]iodoantipyrine as compared to [¹²³I]NaI is similar for all cell lines and that the ratio of 4-[¹²³I]iodoantipyrine/[¹²³I]NaI MN response remains the same. Experiments with the free radical scavenger DMSO, indicated nearly identical dose reduction factors for both ¹²³I carriers. These two observations strongly suggest that the cell inactivation by 4-[¹²³I]iodoantipyrine is not by direct high-LET ionisation of DNA, but is due to an indirect effect. The indirect radiation effect of Auger decay in the nucleus could arise because 4-[¹²³I]iodoantipyrine is not incorporated into the DNA, but is only associated with chromatin where the DNA is shielded by histones. Received: 24 May 2000 / Accepted: 1 November 2000