Method for kinetic analysis of drug-induced cell cycle perturbations

Analysis of radiation-induced changes to human melanoma cultures using a mathematical model

OBJECTIVES

Tumour cells respond to ionizing radiation by cycle arrest, cell death or repair and possible regrowth. We have developed a dynamic mathematical model of the cell cycle to incorporate transition probabilities for entry into DNA replication and mitosis. In this study, we used the model to analyse effects of radiation on cultures of five human melanoma cell lines.

MATERIALS AND METHODS

Cell lines were irradiated (9 Gy) prior to further culture and harvesting at multiple points up to 96 h later. Cells were fixed, stained with propidium iodide and analysed for G(1)-, S- and G(2)/M-phase cells by flow cytometry. Data for all time points were fitted to a mathematical model. To provide unique solutions, cultures were grown in the presence and absence of the mitotic poison paclitaxel, added to prevent cell division.

RESULTS

The model demonstrated that irradiation at 9 Gy induced G(2)-phase arrest in all lines for at least 96 h. Two cell lines with wild-type p53 status additionally exhibited G(1)-phase arrest with recovery over 15 h, as well as evidence of cell loss. Resumption of cycling of surviving cells, as indicated by increases in G(1)/S and G(2)/M-phase transitions, was broadly comparable with results of clonogenic assays.

CONCLUSIONS

The results, combined with existing data from clonogenic survival assays, support the hypothesis that a dominant effect of radiation in these melanoma lines is the induction of long-term cell cycle arrest.

Cell cycle effects of gemcitabine

Gemcitabine (2',2'-difluoro-2'-deoxycytidine, or dFdC) is a promising anticancer agent with demonstrated clinical activity in solid tumours currently undergoing clinical trials. Despite extensive studies on the biochemical mechanism of action, cell cycle perturbations induced by dFdC have not yet been thoroughly investigated, apart from the expected inhibition of DNA synthesis. The aim of our study was to clarify whether cell population kinetics is a vital factor in the cytotoxicity of dFdC in single or repeated treatments and in the dFdC-cisplatin combination. Ovarian cancer cells growing in vitro were treated with dFdC for 1 hr in a range of concentrations from 10 nM to 10 microM. Cell kinetics was investigated by DNA-bromodeoxyuridine flow cytometry, using different experimental protocols to measure either the time course of DNA-synthesis inhibition or the fate of cells in G(1), S or G(2)M at the time of dFdC treatment or 24 hr later. A modified sulforhodamine B test was used to assess the growth inhibition caused by dFdC given alone or with cisplatin. Although dFdC promptly inhibited DNA synthesis, cytotoxicity on proliferating cells was not specific for cells initially in the S phase. DNA synthesis was restored after a G(1) block of variable, dose-dependent length, but recycling cells were intercepted at the subsequent checkpoints, resulting in delays in the G(2)M and G(1) phases. The activity of repeated treatment with dFdC + dFdC or dFdC + cisplatin was highly dependent on the interval length between them. These results suggest that the kinetics of cell recycling from a first dFdC treatment strongly affects the outcome of a second treatment with either dFdC itself or cisplatin.

Desynchronization rate in cell populations: mathematical modeling and experimental data

We characterize the kinetics of two cancer cell lines: IGROV1 (ovarian carcinoma) and MOLT4 (leukemia). By means of flow cytometry, we selected two populations from exponentially growing in vitro cell lines, depending on the cells' DNA synthesis activity during a preceding labeling period. For these populations we determined the time course of the percentages of cells in different phases of the cycles, sampling every 3 hr for 60 hr. Initially, semi-synchronous populations quickly converged to a stable age distribution, which is typical of the cell line (at equilibrium); this desynchronization reflects the intercell variability in cell cycle duration. By matching these experimental observations to mathematical modelling, we related the convergence rate toward the asymptotic distribution (R) and the period of the phase-percentage oscillations (T), to the mean cell cycle duration and its coefficient of variation. We give two formulas involving the above-mentioned parameters. Since T and R can be drawn by fitting our data to an asymptotic formula obtained from the model, we can estimate the other two kinetic parameters. IGROV1 cells have a shorter mean cell cycle time, but higher intercell variability than the leukemia line, which takes longer to lose synchrony.

Hematopoietic toxicity and cell cycle perturbations induced by new DNA minor groove-alkylating agents

Some new alkylating agents which bind to the minor groove of DNA and have sequence-specific patterns of alkylation have shown anti-neoplastic activity in pre-clinical systems. Two of them, carzelesin and tallimustine, are now in phase II. Considering the severe dose-limiting bone marrow toxicity of both these drugs in clinical use, it was of interest to investigate the mechanism of their myelotoxicity in a detailed pre-clinical study and compare it with a conventional alkylating agent, such as melphalan. The origin and progression of the myelotoxicity of carzelesin, tallimustine and melphalan were investigated comparatively in mice, combining data on bone marrow and peripheral blood cellularity with data on the proliferative activity of bone marrow cells, obtained by in vivo administration of bromodeoxyuridine. Significant differences were found between the hematopoietic response to the 3 drugs, though all caused severe leukopenia. Carzelesin induced a short-term increase in myeloid proliferative activity, which prevented the high leukocytopenia on day 3 observed with the other drugs. However, when this effect was exhausted, a second nadir was seen in peripheral blood, with a new wave of cell proliferation of all lineages in the bone marrow. Reconstruction of the lymphoid lineage was slow for all 3 drugs but particularly difficult with high-dose tallimustine. In general, the hematopoietic system response to tallimustine was dampened, with no overshoots, suggesting either lasting effects or extensive cytotoxicity from the early to late precursors of all lineages.

Characterization of chemotherapy-induced morphonuclear modifications in the P388 leukaemia and the MXT mammary tumour models of the mouse

Chemotherapy-induced morphonuclear modifications were monitored in vivo by means of the digital cell image analysis of Feulgen-stained nuclei. Two experimental models were used, i.e. the P388 mouse leukaemia and the MXT mouse mammary carcinoma. The drugs used were doxorubicin, etoposide and cyclophosphamide. The results indicate that the chemotherapy induced a significant decrease in the MXT tumour growth and a significant increase in the survival of the P388 leukaemic mice. These effects were accompanied at the morphonuclear level by an increase in the nuclear area, by modifications in the DNA content in accordance with the effects of the drugs on the cell cycle and by several modifications in the chromatin texture in accordance with the model or the drugs studied. While there were neither homogeneous morphonuclear changes in all treatment groups nor clearcut correlations between the morphonuclear changes and tumour growth or the survival of the animals, the present study nevertheless shows that it is possible, at least partly, to monitor in vivo certain chemotherapy-induced effects occurring at the morphonuclear level, and subsequently to obtain information on the mode of action of the drugs.

Combination of computerized morphonuclear and multivariate analyses to characterize in vitro the antineoplastic effect of alkylating agents

The influence of 13 anticancer alkylating agents on cell proliferation, cell cycle parameters, and morphonuclear characteristics was monitored in vitro on three neoplastic cell lines. This monitoring was carried out by means of the digital cell image analysis of Feulgen-stained nuclei. This computer-assisted microscope analysis of chromatin texture made it possible to assess 15 morphonuclear parameters. These 15 parameters were submitted to multivariate analyses, that is, principal-components analyses followed by the canonical transformation of the data. The 13 alkylating agents included four nitrogen mustards (chlormethine, chlorambucil, melphalan, and cyclophosphamide), two nitrosoureas (carmustine and lomustine), two platinum analogues (cisplatine and carboplatine), two ethyleneimine derivatives (thiotepa and investigational PE1001), one antibiotic (mitomycin C), one alkylsulfonate (busulfan), and one triazene (dacarbazine). The mouse MXT mammary and the human J82 and T24 bladder tumor cell lines were used in this study. The results show that these alkylating agents induced specific modifications to the chromatin pattern according to the subclass to which they belong. In other words, the multivariate statistical analyses of the 15 parameters made it possible to identify, at least partly, distinct subclasses of alkylating agents according to their mechanisms of action. As a validation of the methodology, the results also show that most of the alkylating agents induced an increase in the percentage of cells in the G2 phase, while some sometimes induced an increase in the percentage of cells in the S phase of the cell cycle.

Insulin-like growth factor-1 is not mitogenic for the Walker-256 carcinosarcoma

This study was designed to determine whether intravenous infusion of recombinant human insulin-like growth factor 1 (IGF-1) stimulates tumor growth. In order to determine the potential interaction between nutrition and IGF-1 administration the study was conducted in fasting rats and during continuous feeding by total parenteral nutrition. Tumor cell cycle kinetics including labeling index, DNA synthesis time, cell cycle time in Go/G1, and G2/M in the total cell cycle, and potential doubling time were determined by flow cytometry after in vivo pulse labeling the rats bearing the Walker-256 Carcinosarcoma with 5'-bromo-2'-deoxyuridine (BrdUrd). The results show that IGF-1 treatment has no significant effects on the proliferative characteristics of the tumor model regardless of the feeding status of the animal. This study provides preliminary cell-cycle kinetics data on the short-term effect of IGF-1 on tumor growth. Failure to show a significant effect of IGF-1 on the proliferative characteristics of the tumor suggests that IGF-1 may be given to cancer patients in amounts sufficient to promote weight gain without deleterious stimulation of tumor proliferation.

A mathematical model of the effect of aging on bone marrow cells colonizing the thymus

The process of T cell generation in the thymus involves complex cell-cell interactions between the various types of thymic stromal cells, thymocyte progenitors, thymocytes at different stages of differentiation and external factors. We applied the tool of mathematical modelling to analyze hypotheses and direct experiments concerning mechanisms underlying the observed developmental inferiority of bone-marrow thymocyte progenitors from old mice. Previous experimental data showed that lower cell numbers were obtained from old bone marrow-derived thymocyte progenitors, compared to young bone marrow-derived progenitors, when colonizing simultaneously the same fetal thymus. In this study, simulations based on the mathematical model indicate that the developmental inferiority of old bone marrow-derived progenitors cannot be explained by a change in a single parameter, such as the observed differences in progenitor frequency, an increase in cell cycle duration, a reduction in the fraction of proliferating cells in old age, and/or an increase in the rate of cell death. We have performed experimental measurements of the fractions of cycling cells. No significant difference was found between these fractions in young and old bone marrow-derived thymocytes. The difference in developmental patterns of young and old bone marrow-derived thymocytes may be due to a combination of more than one mechanism, possibly including interactions between competing thymocytes of old and young bone marrow origin.