Suppressive impact of metronomic chemotherapy using UFT and/or cyclophosphamide on mediators of breast cancer dissemination and invasion

Abstract

Metronomic chemotherapy using the 5-FU prodrug uracil-tegafur (UFT) and cyclophosphamide (CTX) was previously shown to only modestly delay primary tumor growth, but nevertheless markedly suppressed the development of micro-metastasis in an orthotopic breast cancer xenograft model, using the metastatic variant of the MDA-MB-231 cell line, 231/LM2-4. Furthermore, a remarkable prolongation of survival, with no toxicity, was observed in a model of postsurgical advanced metastatic disease. A question that has remained unanswered is the seemingly selective anti-metastatic mechanisms of action responsible for this treatment. We assessed the in vivo effect of metronomic UFT, CTX or their combination, on vascular density, collagen deposition and c-Met (cell mediators or modulators of tumor cell invasion or dissemination) via histochemistry/immunohistochemistry of primary tumor sections. We also assessed the effect of continuous exposure to low and non-toxic doses of active drug metabolites 5-fluorouracil (5-FU), 4-hydroperoxycyclophosphamide (4-HC) or their combination, on 231/LM2-4 cell invasiveness in vitro. In the in vivo studies, a significant reduction in vascular density and p-Met[Y1003] levels was associated with UFT+CTX treatment. All treatments reduced intratumoral collagen deposition. In the in vitro studies, a significant reduction of collagen IV invasion by all treatments was observed. The 3D structures formed by 231/LM2-4 on Matrigel showed a predominantly Mass phenotype under treated conditions and Stellate phenotype in untreated cultures. Taken together, the results suggest the low-dose metronomic chemotherapy regimens tested can suppress several mediators of tumor invasiveness highlighting a new perspective for the anti-metastatic efficacy of metronomic chemotherapy.

Fig 1. Effect of UFT, CTX and UFT+CTX metronomic chemotherapy on localized primary tumors.

231/LM2-4 human breast metastatic variant cells were orthotopically injected into the right inguinal mammary fat pad of 6–8 week-old female SCID mice. When tumors reached volumes of ~160–190 mm³, treatment with vehicle control 0.1% HPMC orally daily (black), 15 mg/Kg/d UFT by gavage (red), 20 mg/Kg/d CTX through the drinking water (green) or the combination UFT+CTX treatment (blue) was initiated. Primary tumors were surgically removed when they reached an average size of approximately 1,300 mm³.

Fig 2. Tumor necrosis and vascular density are altered by the metronomic chemotherapy treatments.

(A) The percentage of tumor necrosis was assessed by H&E staining in sections of the central region and the invasive border of the untreated and treated tumors. A significant increase in tumor necrosis was observed in the invasive border sections of the UFT+CTX treated tumors (**p<0.01 vs. control); significance was analyzed by one-way ANOVA with Dunnett’s post-test. (B) Effect of the different treatments on vascular density as determined by double CD31/VEGR2 immunofluorescence labelling. The average vessel count of seven fields per tumor (x20 magnification), from the central region and the invasive border is presented. A significant reduction in vascular density was observed in the CTX (*p<0.05 vs. control) and the UFT+CTX (**p<0.01 vs. control) groups; significance was analyzed by one-way ANOVA with Dunnett’s post-test.

Fig 3. Analysis of local invasion by vimentin immunostaining in tumor xenograft from all mice.

The percent of mice that showed 231/LM2-4 cells invasion in the adjacent soft tissue (grey), in deep muscle (black) and no invasion (white), at the time of sacrificed, was analyzed by vimentin immunostaining of the tumors from the different treatment groups. The number of animals per group is indicated. The statistical analysis has shown no statistically significant difference in local invasion between groups. A logistic regression was applied for comparing local invasion to either soft tissue or deep muscle for the different treatment groups, and the estimated proportions were calculated using a generalized linear mixed model (GLIMMIX procedure).

Fig 4. Assessment of peritumoral and intratumoral collagen deposition in paraffin tumor sections.

The collagen content of the tumors was analyzed by Masson´s trichrome histochemical staining in control and treated groups. A grading system composed of mild/focal (grade 1), moderated (grade 2) and extensive (grade 3) collagen deposition was used. The columns represent the mean ± standard error of the mean (SEM), of the score values corresponding to peritumoral collagen (white) and intratumoral collagen (grey) of the tumor sections corresponding to each treatment group. The statistically analysis by One-Way ANOVA shows no statistical significant difference between groups. (One-Way ANOVA, Dunnett's multiple comparison test P>0.05).

Fig 5. Expression and cellular distribution of c-Met in untreated 231/LM2-4 tumor xenografts.

c-Met staining was relatively homogenous throughout the tumor tissue, and was both cytoplasmic and nuclear. Hyperplastic acini showed cytoplasmic and nuclear staining, with strong staining in the nucleus (down-right arrows).

Fig 6. Expression and cellular distribution of p-Met[Y1003] in untreated 231/LM2-4 tumor xenografts.

Immunohistochemistry for p-Met[Y1003] shows intense widespread staining in the vital areas of the tumor, with both nuclear and cytoplasmic localization. In certain tumor areas, a large amount of reactive stroma was present, which showed strong cytoplasmic and nuclear staining (down-right arrows).

Fig 7. Expression of p-Met[Y1003] in control and in treated groups.

(A) Tumor section showing peripheral areas with preserved morphology and vast areas of central necrosis. H&E stain. (B) The tumor cells show high pleomorphism, hyperchromasia and increased mitotic activity. H&E stain (C) Immunohistochemistry for p-Met[Y1003]. Untreated controls demonstrate uniform intense immunopositivity in the peripheral vital areas of the tumor (lower left). The necrotic central parts of the tumor are immunonegative (upper right). (D) Untreated controls reveal intense immunopositivity in both nuclear and cytoplasmic distribution. (E, F) UFT+CTX treatment group shows significantly decreased staining intensity with only focal immunopositivity p-Met[Y1003] in a predominantly nuclear pattern. Vast areas of viable tumour are not staining (immunonegative).

Fig 8. Immunohistochemistry for p-Met[Y1003] H-score results.

There was a statistically significant decrease in the expression of p-Met[Y1003] in the UFT+CTX treated group, compared to untreated control (Student´s t test, 0.009576 two-tailed p-value).

Fig 9. Effect of continuous exposure to 5-FU, 4-HC and 5-FU+4-HC on viability and invasiveness of 231/LM2-4 cells.

(A) Effect of prolonged continuous exposures (six days) to different concentrations of 5-FU (0.1, 0.5, 1 and 10μM), 4-HC (0.01, 0.05, 0.1 and 1μM), and to the combination 1μM 5-FU + 0.01μM 4-HC, on in vitro viability of 231/LM2-4 cells. The relative absorbance (A_490nm-A_650nm) to control is represented; A_490nm is directly proportional to the number of living cells and A_650nm eliminates any kind of background. (B) Assessment of the invasive capacity of 231/LM2-4 cells, under untreated and treated conditions, in a transwell chemoinvasion assay using collagen type IV as coating. The protracted treatments with low dose and non-toxic doses of 1μM 5-FU, 0.01μM 4-HC and the 1μM 5-FU + 0.01μM 4-HC combination are detailed in Materials and methods. Cells that invaded and passed through membrane pores were stained with 0.1% crystal violet and counted. The numbers are the mean ± standard error of the mean (SEM), of the cell counts of eight random fields (x20 magnification) per sample of a representative experiment. The experiment was repeated three times. A significant reduction in invasive capacity of the cells was observed in all treated groups (***p<0.0001 vs. control). Significance was analyzed by one-way ANOVA with Dunnett’s post-test.

Fig 10. Invasive capacity of 231/LM2-4 cells under control and 5-FU+4-HC conditions using a 3D lrECM “on-top” assay.

(Control): Mass structures (A): round morphology (1–3), irregular morphology showing collective cell migration as chains of few cells with smooth borders (Δ9), buds (Δ6) or as disorganized masses (Δ5). Single cell protrusions (Δ12,14,16). Multicellular streaming with no apparent junction contacts (Δ7,8). Dissemination of single tumor cells (pink Δ5,16). Pseudo-Stellate Mass structures (B): multicellular collective protrusive migration with leading cells with invadopodia (Δ2,4,5). Dissemination of single tumor cells (pink Δ1) and collective dissemination (pink Δ7). Contacts between several structures (pink Δ6). Stellate structures (C): several multicellular chains with a collective cell migration pattern of invasion and protrusive leading front with one (Δ9) or multiple leading cells (Δ15) with invadopodia. Multicellular invasive chains with 1–2 cells in diameter (Δ13) or broad masses of cells (Δ18). Secondary branches at the lateral margin of the invasive chains (Δ12). Cell dissemination (pink Δ7,14,17). Contacts between several structures (pink Δ22,24). (1μM 5-FU + 0.01μM 4-HC): Mass structures (A): round morphology (1–3), irregular morphology exhibiting collective cell migration as chains of few cells with smooth borders (Δ5), buds (Δ4), and structures with single-cell protrusions (Δ14,15). Multicellular streaming (Δ7,9). Dissemination of single tumor cells (pink Δ11) and group of cells (pink Δ12). Pseudo-Stellate Mass structures (B): multicellular collective protrusive migration with leading cells with invadopodia (Δ3) or leading buds (Δ9). Dissemination of group of cells (pink Δ17). Stellate structures (C): multicellular chains with a collective cell migration pattern of invasion and protrusive leading front with invadopodia (Δ5,7), or leading buds (Δ10). Multicellular invasive chains with few cells in diameter (Δ13) or broad masses of cells (Δ13,20). Single and collective cell dissemination (pink Δ27 of the latter). Multicellular chains with an uncoordinated arrangement of the component cells (pink Δ11), contacts (pink Δ15,26) mainly fusions (pink Δ24,25,27). The experiment was repeated at least three times with similar 231/LM2-4 invasiveness inhibitory effect by the treatments. The best multicellular structures were obtained with Matrigel^TM matrix basement membrane (BD BioSciences, Cat. 354234-Lot 88482) with a high concentration (10mg/ml).