Relationship of DNA ploidy to chemoresistance of tumors as measured by in vitro tests

Stratification of the biologic aggressiveness of non-small cell lung cancer using DNA flow cytometry and immunohistochemistry for the retinoblastoma protein

DNA ploidy pattern and S-phase fraction (SPF) measured by flow cytometry and expression of the retinoblastoma gene product (pRB) estimated by immunohistochemistry were correlated with outcome in 114 patients who received a curative operation for primary non-small cell lung cancer (NSCLC). One hundred ten tumors yielded an adequate DNA histogram, and all tumors exhibited an assessable immunohistochemical stain. DNA diploidy was detected in 31 tumors and DNA aneuploidy in 79 tumors. The mean SPF was 14.1 +/- 6.4%. Eighty tumors were positively stained, and 34 tumors were negative for pRB. Multivariate analysis clarified that both TNM staging and DNA ploidy were prognostic factors after surgery. In 39 recurrent cases, the SPF value was inversely correlated with disease-free interval. With only supportive care after recurrence, high SPF tumors and pRB-negative tumors progressed rapidly, whereas active treatment yielded an equivalent effect on recurrent tumors regardless of the SPF or pRB expression. DNA ploidy is related to the risk of recurrence, while SPF is related to tumor growth rate and the impact of active treatment on recurrence. The utility of pRB expression was limited. The combination of DNA ploidy and SPF allows practical stratification of the biologic aggressiveness of NSCLC.

S-phase accumulation precedes apoptosis induced by preoperative treatment with 5-fluorouracil in human colorectal carcinoma cells

BACKGROUND

In this study, the authors examined the correlations among enhancement of apoptosis, changes in cell cycle distribution, and expression of Ki-67 in human colorectal carcinoma cells during different durations of preoperative treatment with 5-fluorouracil (5-FU).

METHODS

Forty-one patients with advanced colorectal carcinoma were divided into 4 groups, which received intravenous 5-FU at 500 mg/body/day preoperatively for 3 days (n = 11), 5 days (n = 13), 7 days (n = 9), or 10 days (n = 8). Patients were further divided into two subgroups according to the DNA ploidy pattern of their tumors, i.e., diploid or aneuploid. Apoptotic cells were stained by the terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling (TUNEL) method. The expression of Ki-67 was examined by immunohistochemical staining. Flow cytometry was used to analyze changes in cell cycle distribution.

RESULTS

Apoptosis of cancer cells was mostly increased in the 7-day group (apoptotic index [Al]: 18.3+/-1.7%] compared with that of the 3-day, 5-day (AI: 11.3+/-2.3% and 13.2+/-2.8%, respectively) (P < 0.01), or 10-day (Al: 16.2+/-2.1%) (P < 0.05) groups. The expression of Ki-67 was reduced with increasing prolongation of 5-FU administration, from 55.2+/-2.1% in the 3-day group to 38.1+/-2.7% in the 10-day group. The authors also assessed the S-phase fraction (SPF) of cancer cells to evaluate changes in cell cycle distribution caused by 5-FU. All tumor samples from patients treated by 5-FU showed S-phase accumulation. The ratio of SPF (after 5-FU/before 5-FU) was higher in the 5-day group (2.35+/-0.78) than in the 3-day (1.71+/-0.48), 7-day (1.68+/-0.55), or 10-day (1.20+/-0.20) groups (P < 0.05). DNA ploidy pattern of the tumor had no influence on the enhancement of apoptosis, the increased ratio of SPF, or the decrease in proliferative activity of colorectal carcinoma cells induced by 5-FU.

CONCLUSIONS

S-phase accumulation preceded apoptotic cell death when intravenous 5-FU was administered continuously to human colorectal carcinoma cells.

Heterogeneity of chemosensitivity in six clonal cell lines derived from a spontaneous murine astrocytoma and its relationship to genotypic and phenotypic characteristics

Heterogeneity in drug sensitivity must, in part, account for the relative lack of success with single agent chemotherapy for glioblastoma multiforme (GBM). In order to develop in vitro model systems to investigate this, clones derived from the VM spontaneous murine astrocytoma have been characterised with regard to drug sensitivity. Six clonal cell lines have been tested for sensitivity to a panel of cytotoxic drugs using an intermediate duration 35S-methionine uptake assay. These lines have previously been extensively characterised with regard to morphological, antigenic, kinetic, tumourigenic potential in syngeneic animals and chromosomal properties and display considerable heterogeneity. The present study indicates that heterogeneity extends to sensitivity to all classes of cytotoxic drugs. The greatest difference in sensitivity between the clones was seen in response to cell cycle-specific drugs like the Vinca alkaloids (14-fold and 20-fold for vincristine (VCR) and vindesine (VIND) respectively), while the nitrosoureas, CCNU and BCNU displayed a smaller fold difference in sensitivity (4.3 and 3.6-fold difference respectively). All the clones were considerably more resistant to the adriamycin (ADM), cis-platinum (C-PLAT) and the Vinca alkaloids than the parental cell line although the difference in sensitivity between the clones and parental cell line were less marked for the nitrosoureas and procarbazine (PCB). It has also been possible to examine the relationship between drug sensitivity and the phenotypic and genotypic properties of these clonal cell lines. There is a relationship between chromosome number and sensitivity of a wide variety of cytotoxic drugs including the nitrosoureas, Vinca alkaloids, PCB, C-PLAT, BLEO but not ADR or 5-FU. Clones with small numbers of chromosomes were more resistant than clones with gross polyploidy. Similarly, sensitivity to Vinca alkaloids and ADM, but not other classes of drugs, was greatest in cells with numerous cytoplasmic processes and which did not express large amounts of cell surface fibronectin. Preliminary experiments have been conducted on reconstituting clonal mixtures of cells with different sensitivity to Vinca alkaloids and results from these studies indicate that the drug resistance phenotype is dominant, with clonal mixtures of sensitive and resistant cell adopting the sensitivity of the more resistant partner. These cell lines should prove to be useful models for examining the cell biological basis of drug resistance in glioma and may lead to the identification and exploitation of novel cellular targets in new therapies for GBM.

Analysis of DNA content in multidrug-resistant cells by image and flow cytometry

Nuclear DNA content was assessed in multidrug-resistant (MDR) cells by image and flow cytometry. Two human MDR cell lines (K562-Dox and CEM-VLB) obtained by in vitro drug selection and overexpressing mdr1 gene were compared to their respective sensitive counterparts (K562 and CCRF-CEM) and to the MDR hamster LR73-R cell line obtained by transfection of mouse mdr1 cDNA. Both cell lines obtained by selection displayed a decreased DNA content, as measured by image cytometry after Feulgen staining, or by flow cytometry after staining with propidium iodide, ethidium bromide, or Hoechst 33342. This decrease was not accompanied by changes in cell cycle phase distribution of cells. Moreover, image cytometry of cells stained after various hydrolysis times in 5 M HCl indicated that MDR cells displayed the same hydrolysis kinetics and sensitivity as drug-sensitive cells with a well-preserved stoichiometry of the Feulgen reaction. LR73-R cells transfected with mdr1 cDNA exhibited only a very limited change in propidium iodide staining as compared with sensitive LR73 cells, suggesting that mdr1 gene overexpression alone could not account for the alterations in DNA content observed in the selected MDR cells.

Nuclear DNA content and chromatin texture in multidrug-resistant human leukemic cell lines

Nuclear morphological alterations associated with multidrug resistance (MDR) were evaluated by image cytometry in various human leukemic cell sub-lines: 3 cell lines with P-gp-mediated resistance (CEM-VLB, HL60/Vinc, K562-Dox), the non-Pgp-mediated MDR HL60/AR leukemic cell line with over-expression of MRP, and the at-MDR CEM-VMI leukemic cell line with alteration of topoisomerase II. All these MDR cell sub-lines were obtained by drug selection and were compared with their sensitive counterparts and with the hamster LR73-R cell line obtained by transfection of mouse mdrl cDNA. All MDR cell sub-lines obtained by drug selection displayed decreased DNA Feulgen stainability as compared with their respective sensitive parental cell line, a phenomenon not observed in the transfected LR73-R cells. Nuclear texture analysis on G0/G1-selected cell nuclei revealed 2 types of textural phenotype. The first phenotype was characterized by chromatin decondensation with small but compact chromatin clumps, and was observed in drug-selected P-gp-mediated MDR cells (CEM-VLB, HL60-Vinc, K562-Dox) and in the non-P-gp-mediated MDR HL60/AR cell line. The second phenotype was characterized by a condensed and homogeneous chromatin pattern, and was observed in the at-MDR CEM-VMI cell line. LR73-R cells transfected with mdrl cDNA did not display any significant changes in textural phenotype as compared with sensitive LR73 cells, suggesting that P-gp over-expression alone cannot account for the cytological modifications observed in MDR cells. These data suggest that multidrug resistance could be associated with specific nuclear morphological changes which appeared to be a consequence of alterations occurring during selection by cytotoxic drugs rather than of P-gp over-expression.

Mouse subrenal capsule assay chemosensitivity and DNA flow cytometry parameters of human melanoma metastases

Chemosensitivity was analyzed by mouse subrenal capsule assay (SRCA) in 103 human melanoma metastases 79 of which were also analyzed by DNA flow cytometry. The most effective combination was dacarbazine, vincristine, and carmustine (DOB), followed by cisplatin plus etoposide (Plat-VP16). By the original criteria of the assay, 35% of tumors were sensitive to DOB treatment while 15% responded to Plat-VP16. Chemosensitivity of DNA diploid and aneuploid tumors showed no significant difference; 35% of the aneuploid tumors were sensitive to DOB and 19% to Plat-VP16, whereas 41% and 13% respectively of the diploid tumors were sensitive. An association between DNA index and chemosensitivity was observed. Growth of the implant was observed in 44% of tumors with a DNA index above 1.5 receiving DOB treatment, whereas only 12% of tumors with an aneuploid DNA index < or = 1.5 grew. No significant difference was observed in the SPF of chemotherapy sensitive and insensitive tumors, though a tendency to lower SPF was observed in sensitive tumors.

Multidrug resistance in rat colon carcinoma cell lines CC531, CC531mdr+ and CC531rev

A rat colon carcinoma cell line, CC531, was exposed to stepwise increasing concentrations of colchicine. A cell line, CC531mdr+, which grows in the presence of 0.2 microM of colchicine was obtained. A revertant cell line, CC531rev was isolated upon colchicine withdrawal. The CC531mdr+ displayed a multidrug-resistant phenotype. Marked resistance to the selecting agent colchicine, was found (RF = 37.5) as well as to vinblastine (RF = 11.3) and actinomycin D (RF = 2.6). Cross resistance to doxorubicin (RF = 8) and daunorubicin (RF = 13.3) was demonstrated. Verapamil was able to reverse drug resistance to colchicine and daunorubicin. The revertant cell line, CC531rev, showed increased sensitivity to colchicine (RF = 0.43), vinblastine (RF = 0.13), doxorubicin (RF = 0.28) and daunorubicin (RF = 0.56). Marked cross resistance to cis-platinum (RF = 6.9) was also induced in CC531mdr+ and was maintained in CC531rev. We conclude that CC531 displays an intrinsic low-level multidrug-resistant phenotype, which was amplified in the CC531mdr+ variant. This correlates with a higher level of expression of P-glycoprotein. CC531rev lacks the multidrug-resistant phenotype and can be used as the drug-sensitive counterpart of the latter two cell lines. Furthermore, it has been shown that in these cell lines cis-platinum resistance is mediated through a mechanism independent of the multidrug-resistant phenotype, since the revertant cell line CC531rev has lost the multidrug-resistant phenotype while retaining the concomitantly induced cis-platinum resistance of the multidrug-resistant variant CC531mdr+.

Immunohistochemical study of P-glycoprotein distribution in lung cancer

Indirect immunoperoxidase was used to determine the reactivity of C219 (P-glycoCHEK C219, Centocor Diagnostics, Malvern, PA), a monoclonal antibody (Mab) with high affinity for an internal epitope of the P-glycoprotein encoded by the multidrug resistance (MDR1) gene, in 40 surgically resected primary lung tumours. C219 reactivity was qualitatively classified in seven small cell lung cancers (SCLC), 29 non small cell lung cancers (NSCLC), and four carcinoid tumours. Ploidy was analysed by means of static cytometry using a computer-assisted image processor following Feulgen staining of cytologic prints of 32/40 lung tumours. Indirect immunoperoxidase reactivities of Mabs S-L 11.14 and MOC-1 were also studied to characterize the expression of cluster 1 lung cancer antigens and hence to determine among the NSCLC those which expressed the neural cell adhesion molecule (NCAM). Eighteen (45%) lung tumours strongly expressed P-glycoprotein as an immunostaining of many islets of malignant cells or almost all malignant cells. In addition, 8/40 tumours (20%) showed a weak reactivity (few immunostained cells) and 14/40 (35%) no reactivity. There was no difference of reactivity when NSCLC were compared with SCLC. The expression of P-glycoprotein in NSCLC did not vary significantly when the stage of disease was considered. Among the 29 NSCLC, 10 (36%) expressed S-L 11.14 and MOC-1. The NCAM positive NSCLC did not show any difference of P-glycoprotein expression in comparison with NCAM negative ones. Finally, C219 immunoperoxidase reactivity did not significantly differ according to the ploidy status. In conclusion, the internal epitope of the P-glycoprotein encoded by the MDR1 gene is frequently expressed by lung tumours of any histological type. This expression is not higher in Stage III and IV lung cancers in comparison with Stage I and II ones, or in NSCLC in comparison with SCLC either. Thus, the C219 related epitope seems to have a weak implication in the lower chemosensitivity of both advanced stages and NSCLC.

Flow cytometric analysis of multidrug-resistance-associated antigen (P-glycoprotein) and DNA ploidy in human colon cancer

In many cell systems, resistance to cytotoxic drugs is acquired by the amplification and/or overexpression of the multidrug resistance (mdr) gene, which codes for the glycoprotein, p170 (P-glycoprotein). Moreover, in a variety of malignant tumours there is increasing evidence of the relationship between the DNA ploidy pattern of patients and their prognosis. In this study we aimed to evaluate these two potential indicators of constitutive drug resistance in human colorectal tumours. We employed a method to quantify simultaneously, on a per cell basis, mdr gene expression (using the C219 monoclonal antibody for P-glycoprotein) and nuclear DNA content with high-resolution bivariate flow cytometry. The study was performed on a human colon-carcinoma-derived cell line (LoVo) and its doxorubicin-resistant variant (LoVo/Dx) and on tumour samples and adjacent normal mucosa from 35 untreated patients with colon cancer. The P-glycoprotein was found in both LoVo and LoVo/Dx cells with levels slightly lower in the parental than in the resistant subline (P, NS). A multi-drug-resistant specific probe for mRNA expression and Western blot assay confirmed the specificity of p170 expression. All of the colon cancer with unimodal diploid DNA distribution and all the normal colonic mucosa samples showed P-glycoprotein expression, without a statistically significant difference in median values between tumours and normal samples. Tumours with bimodal DNA distribution showed median values of P-glycoprotein expression of their hyperdiploid cell clones significantly higher than those of their diploid clones and of the tumours with unimodal DNA distribution (P less than 0.005). Our results show the feasibility of bivariate flow-cytometric analysis of P-glycoprotein expression and DNA content on clinical material and support the hypothesis that the MDR phenotype and DNA ploidy together may influence the biological behaviour of colon cancer in vivo.

Digital morphonuclear analyses of sensitive versus resistant neoplastic cells to vinca-alkaloid, alkylating, and intercalating drugs

We tested 12 resistant cell lines in vitro in order to evaluate common morphonuclear characteristics induced by various cytotoxic drugs on cell lines of different origins. We used the MXT mouse mammary cancer and the neoplastic J82 and T24 human bladder cell lines, whose variants are either sensitive or resistant to a vinca alkaloid derivative (Navelbine, NVB), to an investigational alkylating agent (PE1001), and to Adriamycin (ADR). We tested cell population variants resistant to NVB + PE1001 + ADR. The level of chemoresistance was evaluated by a colorimetric assay assessing the 50% concentration-induced inhibition of cellular growth (IC50) brought about by each drug on the growth of each cell variant under study. We show that resistant neoplastic cell nuclei present common morphonuclear characteristics, independent of cell origin (neoplastic mouse mammary versus human bladder cells) and the drug used (vinca alkaloid, alkylating, and intercalating derivatives). Our results further indicate that the phenotype of resistant versus sensitive cells corresponds to cell nuclei populations with smaller nuclei and less nuclear DNA content and, as a consequence, a chromatin texture showing large pale areas with some hyperchromatic clumps.

Flow cytometric double labeling technique for screening of multidrug resistance

We investigated the capabilities of flow cytometry in the analysis of a multidrug resistant (MDR) human ovarian cancer cell line 2780AD and its drug sensitive parental A2780. A functional assay using daunorubicin (DNR) as a fluorescent probe was combined with an immunofluorescence assay of P-glycoprotein (P-gp) using the monoclonal antibody MRK-16. Functionally MDR could be demonstrated by the lower DNR-content of MDR cells compared to DNR-content of drug sensitive cells. When incubation was performed with DNR in the presence of verapamil, DNR-content increased in the MDR cells. However the content of the A2780 cells was never attained. Differences in DNR-content were not related to differences in DNA-content. In experimental cell lines immunofluorescence data were inversely related with those of DNR-content: MDR cells had high levels of P-gp expression and low levels of DNR-content (and vice versa in drug sensitive cells). Both assays can be easily combined in a multiparametric flow cytometric procedure to evaluate both parameters simultaneously in the same cells. Analysis of clinical samples demonstrates the existence of aberrant subpopulations which would not be detected by using a single parameter assay.