The uptake and degradation of alkylating drugs by Yoshida ascites sarcoma cells in vitro. Biochem Pharmacol 19:209-217, 1970 (77) HILL BT, HARRAP KR: The uptake and utiliza-
- don of chlorambucil by-lymphocytes from patients with chronic lymphocytic leukemia. Br J Cance i 26:439-443, 1972 _ _ "(72fT6RRES-GAROA SJ, CAPLAN "S, PANASa L: Resistance to the nitrogen mustards in chronic lymphocytic leukemia correlates with enhanced removal of DNA crosslinks. Proc ASCO 8:197, 1989 (13) RAI KR, SAWTTSXY A, CRONHTE EP, ET AL:
Clinical staging of chronic lymphocytic leukemia. Blood 46:219-234, 1975 ( 7 0 BOYUM A: Isolation of lymphocytes, granulocytes, and macrophages. Immunology 5(suppl 5):9-15, 1976 (15) BEGLETTER A, GLAZER RI, ISRAELS LG, ET AL:
Induction of DNA strand breaks in chronic lymphocytic leukemia following treatment with 2'-deoxycoformycin in vivo and in vitro. Cancer Res 47:2498-2503, 1987 (16) STOUT DL, BECKER FF: Fluorometric quantita-
tion of single-stranded DNA: A method applicable to the technique of alkaline eluUon. Anal Biochem 127:302-307, 1982 (17) SEDLAK J, LINDSAY RH: Estimation of total,
protein-bound and non-protein sulfhydryl groups in tissue with Ellman's reagent. Anal Biochem 25:192-205, 1968 (IS) SUZUKAKE K, PETRO BJ, VISTICA D: Reduction
in glutathione content of L-PAM resistant L1210 cells confers drug sensitivity. Biochem Pharmacol 31:121-124, 1982 (79) HABIG WH, PABST MJ, JAKOBY WB: Glu-
tathione-5-transferase A: The first enzymatic step in mercapturic acid formation. J Biol Chem 249:7130-7139, 1974
References (7) GALE RP, FOON KA: Chronic lymphocytic leu-
kemia: Recent advances in biology and treatment. Ann Intern Med 103:101-120, 1985 (2) HILL BT: Studies on the transport and cellular distribution of chlorambucil in the Yoshida ascites sarcoma. Biochem Pharmacol 21: 495-502,1972 (3) BANK BB, KANGANIS D, LJEBES LF, ET AL:
(20) PANASCI L, HENDERSON D, TORRES-GARCIA SJ,
ET AU Transport, metabolism, and DNA interaction of melphalan in lymphocytes from patients with chronic lymphocytic leukemia. Cancer Res 48:1972-1976, 1988 (27)
JIANGB-Z,BANKBB,HSIANGY-H,ETAL:Lack
of drug-induced DNA cross-links in chlorambucil-resistant Chinese hamster ovary cells. Cancer Res 49:5514-5517, 1989
Chlorambucil pbarmacokinetics and DNA binding in chronic lymphocytic leukemia lym- (22) ROBSON CN, LEWIS AD, WOLF CR, ET AL: Reduced levels of drug-induced DNA phocytes. Cancer Res 49:554-559, 1989 crosslinking in nitrogen mustard-resistant Chi(4) FARMER PB: Metabolism and reactions of nese hamster ovary cells expressing elevated alkylating agents. Pharmacol Ther 35:301glutathione 5-transferase activity. Cancer Res 358, 1987 47:6022-6027,1987 (5) ROBSON CN, ALEXANDER J.HARRIS AL, ETAL:
Isolation and characterization of a Chinese hamster ovary cell line resistant to bifunctional nitrogen mustards. Cancer Res 46:6290-6294, 1986 (6~) ENDRESEN L, BAKKA A, RUGSTAD HE: In-
creased resistance to chlorambucil in cultured cells with a high concentration of cytoplasmic metallothionein. Cancer Res 43:2918-2926, 1983 (7) BULLER AL, CLAPPER ML, TEW KD: Glu-
tathione S-transferase in nitrogen mustard-resistant and -sensitive cell lines. Mol Pharmacol 31:575-578, 1987 (8) LEWIS AD, HKKSON ID, ROBSON CN, ET AL:
Amplification and increased expression of alpha class glutathione S-transferase-encoding genes associated with resistance to nitrogen mustards. Proc Natl Acad Sci USA 85: 8511-8515,1988
Detection of Etoposide Resistance by Measuring DNA Damage in Individual Chinese Hamster Cells Peggy L. Olive* Judit P. Bandth, Ralph E. Durand
(9) HARRAP KR, GASCOIGNE EW: The interaction
of bifunctional alkylating agents with the DNA of tumor cells. EurJ Cancer 12:53-59,1976 (70) HARRAP KR, HILL BT: The selectivity of action
of alkylating agents and drug resistance, d .
Vol. 82, No. 9, May 2, 1990
The comet assay, which measures DNA strand breakage in individual cells, was used to examine the relation between
DNA damage, cell survival, and resistance to the topoisomerase II inhibitor etoposide (VP-16). Chinese hamster V79*17H> cells and a VP-16-resistant subline (VP*) were exposed to VP-16 as monolayers or spheroids". The' comet" assay was comparable in sensitivity to the DNA precipitation and alkali unwinding assays for detecting DNA strand breaks induced by VP-16. However, unlike conventional DNA damage assays, the comet assay also indicated heterogeneity in cell response. For V79 multicell spheroids exposed to VP-16, the external cycling cells were 50 times more sensitive to killing and DNA damage than the internal noncycling cells; the comet assay indicated the fraction of cells resistant to the drug. V P cells, which were 10 times more resistant to killing and DNA damage by VP-16 than the parent cell line, could also be identified in mixed populations with the use of this method. These results suggest that the comet assay could be useful in predicting tumor cell response to DNAdamaging agents. [J Natl Cancer Inst 82:779-783,1990]
Development of new predictive assays for tumor response has attracted greater attention in recent years, perhaps as a result of an increased appreciation of the diversity of responses of individuals to cancer therapy (/). One of the most important properties of an effective predictive assay is its ability to detect a small proportion of resistant cells. Microelectrophoresis, or the "comet assay," is a rapid and relatively simple method for measuring DNA damage to individual cells, thus (theoretically) allowing identification of a small fraction of the cells that are resistant to treatment. This method was originally described in 1984 by Ostling and Johanson (2), and we subsequently adapted and
Downloaded from http://jnci.oxfordjournals.org/ at University of California, San Francisco on December 12, 2014
findings of Panasci et al. In our study, the CLL cells of the two patients who were chlorambucil-resistant showed a similar number of DNA cross-links to that of the cells from sensitive patients, when the resistant and the sensitive patients were treated with chlorambucil in vitro. However, the two resistant patients were receiving prednisone at the time of our study, and the influence of this agent on chlorambucil-induced DNA cross-linking in vivo, which may not have been observed in vitro, is unknown. Factors other than the formation of DNA cross-links, such as DNA repair, may contribute to alkylating-agent resistance in CLL. In this regard, the removal of melphalan-induced DNA cross-links in CLL cells in vitro appears to be more rapid in chlorambucil-resistant patients than in untreated patients (72). The area under the curve for DNA cross-linking may thus be the critical factor for cell kill, and the clinical responsiveness of CLL patients to chlorambucil should be correlated with the rate of formation and repair of DNA crosslinks.
Received October 27, 1989; revised January 26, 1990; accepted February 5, 1990. Supported in part by a grant from the British Columbia Health Care Research Foundation. British Columbia Cancer Research Centre, Vancouver, BC, Canada. Correspondence to: Peggy L. Olive, Ph.D., Medical Biophysics Unit, British Columbia Cancer Research Centre, 601 W. 10th Ave., Vancouver, BC, V5Z 1L3 Canada.
REPORTS
779
Materials and Methods Cell Lines and Culture Conditions Chinese hamster V79-171b cells were maintained as monolayers in exponential growth by twice weekly subcultivation in Eagle's minimum essential medium containing 10% fetal bovine serum (GIBCO Laboratories, Grand Island, NY). Single cells were prepared by trypsin treatment with the use of 0.1% trypsin (GIBCO) in phosphate-buffered saline. Spheroids were initiated by placement of 4 x 104 cells/mL in glass spinner-culture vessels (10). Spheroids were treated 10 days later when they had reached a diameter between 0.6 and 0.7 mm and contained approximately 103 cells. Cells resistant to VP-16 (VF cells) were obtained by treatment
780
with 10 jig of N-methyl-#'-nitro-#-nitrosoguanidine/mL followed by selection in ljigofVP-167mL(77). Cells were treated with VP-16 either as monolayers or in suspension when they were grown as spheroids. VP-16 was purchased from Bristol Laboratories, Belleville, ON, Canada, as Vespesid, a solubilized formulation in polysorbate 80 and polyoxyethylene glycol. It was diluted just prior to use in medium containing 5% fetal bovine serum. Following treatment, single-cell suspensions were prepared with the use of 0.1% trypsin for monolayers and 0.25% trypsin for spheroids. After 5 minutes, cells were resuspended in complete medium, centrifuged, and resuspended in phosphate-buffered saline purchased from GIBCO Laboratories.
590-nm band-pass filter. Individual images were digitized, and we calculated the tail moment, defined as the product of the percentage of DNA in the tail distribution and the displacement (in pixels) between the positions of the head and tail means. Data reduction was corrected for the background fluorescence as described in our previous publication (3).
Results
VP-16 caused significant numbers of DNA single-strand breaks, as detected with either the alkali unwinding assay or the DNA precipitation assay (fig. 1, panels A and B). For the former assay, DNA unwinding in a high-salt, weak-alkali lysis solution begins at sites of strand breakage; the rate of DNA unwinding has been found Measurements of DNA Damage to be proportional to the number of strand breaks (12). In comparison, the DNA preThe alkali unwinding assay of Ahncipitation assay measures the amount of strdm and Erixon (12) and the DNA precipitation assay were performed as previ- DNA precipitated when cells are lysed in ously described (13,14). For the comet alkaline 2% sodium dodecyl sulfate, folassay, 1.5 x 104 cells were suspended in lowed by precipitation with the use of 0.12 0.5 mL of phosphate-buffered saline on M potassium chloride (14). While based ice. Agarose (1.5 mL of 1% agarose; on different physical principles, both product No. 6013; Sigma Chemical Co., methods were equally sensitive in detectSt. Louis, MO) held at 50 °C was added to ing DNA damage by ionizing radiation the tube, and the suspension was rapidly (14) and, as shown here, were of similar pipetted onto a glass microscope slide. sensitivity in detecting VP-16-induced The sample was allowed to gel for 1 DNA damage. V F cells were about 10 minute, then placed in ice-cold lysing so- times more resistant to production of lution containing 0.03 M NaOH and 1 M strand breaks by VP-16 than the parent NaCl. After 1 hour in the dark at room V79 cell line (fig. 1, panel B). temperature, the slides were rinsed gently The comet assay was as sensitive as the in distilled water, then placed in a horizon- DNA precipitation and alkali unwinding tal gel electrophoresis apparatus (model assays were in detecting DNA damage by H4; Bethesda Research Laboratories, VP-16 (fig. 1, panel Q . In addition, since Inc., Gaithersburg, MD). Slides were cov- the cells need not be radiolabeled for the ered with 40 mAf Tris, 5 mAf edetic acid comet assay, results can also be obtained (EDTA), and 20 mAf acetic acid (pH 8.3). for Chinese hamster V79 spheroid cells, They were then exposed for 12 minutes to many of which are noncycling (10) and, 150 V (~4 V/cm) after which they were therefore, unable to incorporate radioacrinsed in distilled water and incubated for tive thymidine. Note that the average re10 minutes in the dark with 2.5 jig of sponse of cells from spheroids suggests propidium iodide/mL. After staining, greater resistance to VP-16 than that seen slides were rinsed twice in distilled water, with the V F cells, and the apparent nonplaced in a humidified air-tight container, linear shape of the dose-response curve and viewed within 48 hours. for spheroid cells could also indicate hetWe viewed 40-200 individual comets erogeneity in response. using a 25 x objective with a Zeiss epiAll three DNA damage assays were fluorescence microscope with attached sensitive to VP-16-induced damage at bihigh-sensitivity video camera and image ologically relevant doses, as indicated by analysis system (15). Cells were illumi- the cell survival data shown in figure 1 nated with a 546-nm light excitation from (panel D). The V F cells showed a change a 100-W mercury light. Emission was in both shoulder and terminal slope of the monitored using a 580-nm reflector and a dose-response curve; the slope of the cell Journal of the National Cancer Institute
Downloaded from http://jnci.oxfordjournals.org/ at University of California, San Francisco on December 12, 2014
modified it for image analysis (3). In the comet assay, a small number of single cells are embedded in agarose, lysed, subjected to an electric field, stained with a fluorescent DNA-binding dye, and visualized with the use of a fluorescence imageprocessing system. For cells containing DNA strand breaks, the electric current pulls the negatively charged broken and relaxed DNA away from the nucleus toward the anode (2,4). In undamaged cells, most of the supercoiled DNA remains within the nuclear cage, but a damaged cell resembles a "comet" with a brightly fluorescent head and a tail containing the fragments and relaxed supercoils of DNA. Previous results with this method indicate that (a) DNA single-strand breaks are recognized, (b) damage from radiation doses of 2.5 Gy or lower can be detected, and (c) heterogeneity in DNA damage within a population of cells is readily apparent (2-5). This assay thus appears to show promise as a means of identifying cells resistant to drug therapy. Etoposide (VP-16), an epipodophyllotoxin that traps topoisomerase II during cleavage of DNA, is an effective inducer of DNA damage, producing single- and double-strand breaks and DNA-protein cross-links (6-9). To evaluate the sensitivity and reliability of the comet assay to recognize drug damage, we compared this method with more conventional methods used to measure DNA damage. The usefulness of the comet assay for predicting cell survival in a mixture of VP- 16-resistant and -sensitive cells was also examined.
10
20
T"^-^^ •fc*. 1
30
40 1
D.
vs A;
10
20
Etoposide
30
40
(|jg/ml)
1
VP r
>, E)
•
0 10 20 30 40 Etoposide (pg/ml)
Figure 1. Panels A and B: DNA damage by VP-16 detected with the use of the alkali unwinding assay (o,«) or the DNA precipitation assay ( A , • ) . [ |i4C]Thymidine-labeled exponentially growing Chinese hamster V79 cell monolayers (V79m) or a resistant subline (VP1) were incubated for 1 hr with VP-16 at concentrations shown. Cells were analyzed for DNA damage as previously described (77). Mean and standard deviation for three determinations are shown. Note difference in drug concentrations used for drug-sensitive (panel A) and drug-resistant (panel B) cells. Panels C and D: Toxicity and DNA damage by VP-16 measured with the comet assay. Chinese hamster V79 cell monolayers (Vm), 0.6-mm-diameter V79 spheroid cells (Vs), and VP* cell monolayers (VP1) were exposed for 1 hr to VP-16. Panel C: DNA damage was measured in 80-200 individual cells per dose point by the comet assay. Mean and standard deviation for three separate experiments are shown for selected concentrations, x , cell monolayers; o, VF* cells; A, spheroid cells. Panel D: Cell clonogenicity is average of two to four experiments.
monolayer survival curve was about 10 times steeper than the terminal slope of the VI* cells. The dose-response curve for the spheroid cells appeared biphasic, and a significant fraction of the spheroid cells was about 50 times more resistant to VP16 than were the exponentially growing cell monolayers. Extrapolating the terminal region of the spheroid cell survival curve indicated that about 50%-60% of the spheroid cells were resistant to VP-16; this percentage is consistent with the fraction of noncycling cells within these spheroids (10). The important advantage of the comet assay is its ability to detect damage in individual cells and, therefore, to measure heterogeneity. For V79 spheroids, the external cycling cells respond to treatment with VP-16, but the internal, noncycling cells are resistant (16). Measuring the average amount of DNA damage does not provide this information; a decrease in average DNA damage would be observed if all cells were somewhat moreresistantto the drug or if only a fraction of the cells Vol. 82, No. 9, May 2, 1990
was considerably more resistant. The comet assay indicated that more of the spheroid cells (fig. 2, panel A) were resistant to VP-16 than monolayers of either drug-resistant (fig. 2, panel B) or -sensitive cells (fig. 2, panel C). Correlating tail moment and surviving fraction indicated that one can also use the average amount of damage detected by the comet assay to predict cell clonogenicity (fig. 3). All of the results were fitted to the same linear-quadratic curve, although the results with the V79 cell monolayers were on the left of this curve and the results with the spheroid cells were generally on the right. From the data shown in figure 2 (panel A), one might expect the spheroid cells to show greater survival for the same average tail moment, since a larger fraction of the spheroid cells showed minimal damage by VP-16. We examined the possibility of dispensing with the video imaging system and simply measuring drug resistance by counting, with the use of only the fluorescence microscope, the fraction of comets
10
20 30 TAIL MOMENT
40
Figure 2. Heterogeneity in DNA damage by VP-16 detected by the comet assay. V79 spheroid cells, VP cell monolayers (VP1), and Chinese hamster V79 cell monolayers (V79m) were incubated for 1 hr with VP-16 at concentrations that produced average tail moments between 10 and 15 U. Surviving fraction (S.F.) is also shown for same population of treated cells.
without long tails (i.e., the undamaged or less damaged cells). Samples of V F cells were mixed with the parent V79 cells at various ratios. The mixtures were incubated with 2 jig of VP-167mL for 1 hour and were prepared for the comet assay on coded slides. Five observers familiarized
1.0
z o SURVI\/ING FRACT]
D
Vm
Downloaded from http://jnci.oxfordjournals.org/ at University of California, San Francisco on December 12, 2014
•0.01
o
: ^ \ 0. 1 i
A
\ 0.01
1
x
-
r
\
:
.001 x 10
0
TAIL
: 20
MOMENT
Figure 3. Comparison between tail moment and clonogenicity for V79 cell monolayers (x), V F cells (o), and multicell spheroids (A). Data are fitted to a linear quadratic equation.
REPORTS
781
Table 1. Visual identification of VP-16-resistant cells by the comet assay* Actual % resistant cells
% scored resistant by Observer 1
5
15
10 25 50 75 90 95
15.5 40.5 46.5
r2 Slope
81 74 92.5 0 94 0.84
Observer 2
Observer 3
Observer 4
12.5
30 29 72 53 60 82 94 071
11.5 17.5
22 44.5 47.5
65 77 92 0.94 0.74
0.55
44 57 48 78.5
96 0.85 0.75
Observer 5
18 15.5 26.5
26 75.5
69 75 0 88 0.69
themselves with the appearances of comets from 100%resistantand 100% sensitive cells. They then examined 200 comets on each coded slide and rated each comet as damaged Gong tail, dim head) or undamaged (short tail, bright head). As table 1 indicates, the ability to detect resistant cells visually, without using criteria developed for the imaging system, was surprisingly good. In addition, only 10-20 minutes was required for visual observation of 200 comets compared with about 200 minutes for analysis using our current software for the fluorescence image-processing system.
Discussion Several criteria must be met for an assay to be a useful method for detecting resistant cells within a heterogeneous population: (a) The agent used must cause damage of a type that can be detected by the assay, (b) The resistant cells must be several times more resistant to the agent than the sensitive cells; how many times will depend on the proportion of resistant cells present and the number of cells analyzed. (c) The dose of the agent used must be sufficient to allow distinction between damaged and undamaged (untreated) cells, (d) A sufficient number of cells must be analyzed; if only 5% of the population is resistant to the treatment, then analyzing 20 cells is statistically inadequate, (e) Finally, the method should be relatively rapid and simple to perform. The comet assay enabled the experienced observers to visually identify resistant cells in mixtures of V F and V79 cells.
782
Other methods can measure DNA damage to individual cells (18-21) and may provide comparable information. However, the speed and simplicity of the comet It shows promise as a rapid, inexpensive assay and the possibility of using it in the method for detecting VP-16-resistant cells absence of sophisticated image analysis in tumors, provided the comets arereadon hardware and software to screen samples coded slides. More accurate identification give it a decided advantage over these is possible as the degree ofresistanceto the other methods. drug increases. A major problem in detecting small fractions of resistant cells (aside References from the number of comets that must be (/) BROCK WA, BAKER FL, TOFILON PJ: Tumor analyzed) is the heterogeneity within even cell sensitivities to drugs and radiation. In the relatively homogeneous V79 cell Prediction of Tumor Treatment Response (Chapman JD, Peters LJ, Withers HR, eds). monolayers. The large variation in tail New York: Pergamon Press, 1989, pp 139-155 moment associated with a homogeneously (2) OSTUNG O, JOHANSON KJ: Microelectroresponding population is not easily exphoretic study of radiation-induced DNA damages in individual mammalian cells. Biocbem plained. When the average tail moment of Biophys Res Commun 123:291-298, 1984 cells treated with 2 p,g of VP-16/mL is 10 (3) OLIVE PL, BANATH JP, DURAND RE: Heterogetimes that of the untreated cells, 3%-6% of neity in radiation-induced DNA damage and repair in tumor and normal cells measured using the treated comets still have tail moments the comet assay. Radial Res. In press that overlap the control distribution. This (-0 OSTUNG O, JOHANSON KJ: Bleomycin, in consituation immediately places a limit on the trast to gamma irradiation, induces extreme ability to detect small fractions of resistant variation of DNA strand breakage from cell to cell. Int J Radiat Biol Relat Stud Phys Chem cells. Determining whether this variability Med 52:683-691,1987 is a biological or a technical problem with (5) SINGH NP, MCCOY MT, TKE RR, ET AL: A the assay is of primary importance. simple technique for quantication of low levels
Cycling cells in V79 spheroids appear to be about 50 times more sensitive to VP-16 at a surviving fraction of 0.1. Therefore, tumor growth fraction might be estimated using the comet assay by the determination of the percentage of cells that respond to VP-16. While topoisomerase II levels drop precipitously in noncycling cells (17), noncycling cells in V79 spheroids can still be killed by VP-16. The comet assay can predict cell clonogenicity using the average tail moment (fig. 3), but it is of limited value if only a small percentage of resistant cells is present. It is also possible to calculate a cutoff value for the tail moment. This value can be used to distinguish damaged
of DNA damage in individual cells. Exp Cell Res 175:184-191,1988 (6~) KALWINSKY DK, LOOK AT, DUCORE J, ET AL:
Effects of the epipodophyllotoxin VP-16-213 on cell cycle traverse, DNA synthesis, and DNA strand size in cultures of human leukemic lymphoblasts. Cancer Res 43:1592-1597,1983 (7) Liu LF, ROWE TC, YANG L, ET AL: Cleavage of
DNA by mammalian DNA topoisomerase D. J Biol Chem 258:15365-15370,1983 (S) Ross W, ROWE T, GUSSON B, ET AL: Role of
topoisomerase II in mediating epipodophyllotoxin-induced DNA cleavage. Cancer Res 44:5857-5860, 1984 (9) LONG BH, MUSIAL ST, BRATTAIN MG: Single-
and double-strand DNA breakage and repair in human lung adenocarrinoma cells exposed to etoposide and teniposide. Cancer Res 45: 3106-3112,1985 (10) SUTHERLAND RM, DURAND RE: Radiation re-
sponse of multicell spheroids—an in vitro tumour model. Curr Top Radiat Res Q 11: 87-139, 1976
Journal of the National Cancer Institute
Downloaded from http://jnci.oxfordjournals.org/ at University of California, San Francisco on December 12, 2014
•Mixtures of V79 cells and VPf cells were incubated with 2 /»g of VP-16/mL. Each observer "blindly" scored 200 comets per slide. Observers 1 and 2 were experienced with the assay, the others were not The correlation coefficient (r2) for the linear best-fit line and the slope of that line are shown for each data set
from undamaged cells. A line extrapolated from the terminal region of the curve shown in figure 3 crosses the 100% survival level at a tail moment close to 5. This means that cells with a tail moment of 5 or less might be expected to survive damage by VP-16. With this criterion, the calculated survival rates from figure 2 (panels A-C) are 0.34, 0.36, and 0.05, respectively, which can be compared with the measured survival rates of 0.13,0.20, and 0.01, respectively. The agreement between the calculated and measured values would, of course, be expected to improve as more comets are analyzed.
(11) GUPTA RS: Genetic, biochemical, and crossresistance studies with mutants of Chinese hamster ovary cells resistant to the anticancer drugs, VM-26 and VP-16-213. Cancer Res 43: _ . 1568-1574,1983 (72) AHNSTROM G, ERTXON K: Radiation induced
strand breakage in DNA from mammalian cells. Strand separation in alkaline solutionTlnt J Radial Biol Relat Stud Phys Chem Med 23:285-289, 1973 (13) OUVE PL, CHAN AP, CU CS: Comparison
between the DNA precipitation and alkali unwinding assays for detecting DNA strand breaks and cross-links. Cancer Res 48: 6444-6449, 1988 (14) OUVE PL: DNA precipitation assay: A rapid and simple method for detecting DNA damage in mammalian cells. Environ Mol Mutagen 11:487^195,1988 (15) JAGGI B, POON SS, MACAULAY C, ET AL:
(16) DURAND RE, GOLDIE JH: Interaction of etopo-
side and cisplatin in an in vitro tumor model. Cancer Treat Rep 71:673-679, 1987 (17) HSIANG Y, Wu H, Liu LF: Proliferation-depen-
dent regulation of DNA topoisomerase II in cultured cells. Cancer Res 48:3230-3235,1988 (18) Ron Ron JL, WRIGHT WD: Visualization of
DNA loops in nucleoids from HeLa cells: Assays for DNA damage and repair. Cytometry 8:461-467, 1987 (19) NOSE K, OKAMOTOH: Detection of carcinogen-
induced DNA breaks by nick translation in permeable cells. Biochem Biophys Res Commun 111:383-389, 1983 (20) RAZA A, MEHDI I, MAYERS G, ET AU In vitro
detection of single stranded DNA damage (SSD) in individual cells: A useful test for detecting reversal of multi-drug resistance (MDR). Proc Am Assoc Cancer Res 30:597, 1989 (21) DEEN DF, KENDALL LE, MARTON LJ, ET AL:
Prediction of human tumor cell chemosensitivify using the sister chromatid exchange assay. Cancer Res 46:1599-1602, 1986
vV-Nitrosodimethylamine Blood Levels in Patients With Chronic Renal Failure: Modulation of Levels by Ethanol and Ascorbic Acid S. R. Dunn* M. L. Simenhoff, P. S. Leie, S. Goyal, J. W. Pensabene, W. Fiddler
We measured levels of iV-nitrosodimethylamine (NDMA) in peripheral blood from 13 fasting male patients, 30-74 years old, who had chronic renal failure, and in five healthy control subjects (four males and one female) 31-50 years
Vol. 82, No. 9, May 2, 1990
precursors and 4-methylpyrazole (17). Study results have also indicated that bacteria may play an active role in nitrosamine formation (18-24), In addition, chronic renal failure in humans has been shown to be a favorable model for studying intestinal nitrosamine production (a) because NDMA and N-nitrosopiperidine (NPIP) have been found in duodenal aspirates in concentrations greater than those found in the blood (11,25,26) and (b) because bacterial overgrowth occurs in the duodenum (27). If NDMA, and possibly other N-nitroso compounds, contribute to the total human exposure to carcinogens, it would be prudent to reduce the endogenous component in the gastrointestinal tract. One approach would be to eliminate the intestinal bacterial overgrowth; another would be to block NDMA formation. Ascorbic acid (vitamin C) has already been shown to be an effective inhibitor of nitrosation in human and animal studies (17,28-31). In this study, we investigated NDMA formation in patients with chronic renal failure. We used ethanol to block first-pass hepatic metabolism of NDMA derived mainly from the portal circulation and then measured the circulating NDMA released to the peripheral blood and target organs from the liver. We then examined the effect of oral ascorbic acid on modulation of gastrointestinal NDMA formation under the same conditions.
The role of N-nitroso compounds as contributors to the etiology of human cancer has been under investigation for more than 30 years since the original findings by Magee and Barnes (1) of the hepatotoxicity and carcinogenicity of N-nitrosodimethylamine (NDMA) in the rat. As a Materials and Methods class, nitrosamines are remarkable carcin- Patient Selection and Study Design ogens, as shown by their potency, their Patients entered in the study were unaction in many species, and the variety of dergoing hemodialysis three times per target organs (2). Direct evidence of in vivo formation of NDMA in humans has been difficult to Received August 1, 1989; revised February 15, obtain, owing to its ubiquitous presence in very low concentrations. However, there 1990; accepted February 23,1990. Supported in part by Public Health Service grant is evidence that N-nitroso compounds may CA-26571 from the National Cancer Institute, Nabe formed in vivo (3,4). tional Institutes of Health, Department of Health and Theoretically, conditions for endoge- Human Services. S. R. Dunn, M. L. Simenhoff, P. S. Lele, nous formation of N-nitroso compounds, particularly NDMA, would be favored in S. Goyal, Department of Medicine, Division of Nephrology, Jefferson Medical College, Thomas the normal gastrointestinal tract (5,6), and Jefferson University, Philadelphia, PA. this hypothesis is supported by reports of J. W. Pensabene, W. Fiddler, U.S. Department of elevated levels of total N-nitroso com- Agriculture, Eastern Regional Research Center, Philpounds in human stomach aspirates (7) adelphia, PA. We thank Hyman Menduke, Robert Gates, and and in the large intestine of rats (8). Nancy Picarello for their assistance. NDMA has been found in saliva, blood, Correspondence to: Stephen R. Dunn, Departurine, and gastric and duodenal aspirates ment of Medicine, Division of Nephrology, Jefferson in healthy individuals (9-16) and, re- Medical College, Thomas Jefferson University, Philcently , in the urine of ferrets given NDMA adelphia, PA 19107.
Downloaded from http://jnci.oxfordjournals.org/ at University of California, San Francisco on December 12, 2014
Imaging system for morphometric assessment of absorption or fluorescence in stained cells. Cytometry 9:566-572, 1988
old. In the patients, we found significant (P < .01) levels of NDMA (mean ± SD; 201 ± 111 ng/kg of blood), which is known to be carcinogenic in animals. Five minutes after oral administration of ethanol (0:4 g/kg of body weight) j all patients exhibited a significant (P < .01) rise in blood NDMA levels (338 ± 125 ng/kg), suggesting continuous endogenous formation of NDMA that was unmasked by ethanol's ability to inhibit first-pass hepatic metabolism of NDMA. In five of six patients, pretreatment with oral ascorbic acid resulted in a blunting, but not statistically significant, effect on maximum blood NDMA levels after consumption of ethanol. Mean levels were 340 ± 100 ng/kg before treatment with ascorbic acid and 237 ± 127 ng/kg during treatment. Ethanol administration unmasks increased gastrointestinal formation of NDMA in patients with chronic renal failure. Further studies are required to confirm a possible link between endogenous NDMA formation and the increased incidence of cancer in these patients. [J Natl Cancer Inst 82:783-787,1990]
REPORTS 78;
- don of chlorambucil by-lymphocytes from patients with chronic lymphocytic leukemia. Br J Cance i 26:439-443, 1972 _ _ "(72fT6RRES-GAROA SJ, CAPLAN "S, PANASa L: Resistance to the nitrogen mustards in chronic lymphocytic leukemia correlates with enhanced removal of DNA crosslinks. Proc ASCO 8:197, 1989 (13) RAI KR, SAWTTSXY A, CRONHTE EP, ET AL:
Clinical staging of chronic lymphocytic leukemia. Blood 46:219-234, 1975 ( 7 0 BOYUM A: Isolation of lymphocytes, granulocytes, and macrophages. Immunology 5(suppl 5):9-15, 1976 (15) BEGLETTER A, GLAZER RI, ISRAELS LG, ET AL:
Induction of DNA strand breaks in chronic lymphocytic leukemia following treatment with 2'-deoxycoformycin in vivo and in vitro. Cancer Res 47:2498-2503, 1987 (16) STOUT DL, BECKER FF: Fluorometric quantita-
tion of single-stranded DNA: A method applicable to the technique of alkaline eluUon. Anal Biochem 127:302-307, 1982 (17) SEDLAK J, LINDSAY RH: Estimation of total,
protein-bound and non-protein sulfhydryl groups in tissue with Ellman's reagent. Anal Biochem 25:192-205, 1968 (IS) SUZUKAKE K, PETRO BJ, VISTICA D: Reduction
in glutathione content of L-PAM resistant L1210 cells confers drug sensitivity. Biochem Pharmacol 31:121-124, 1982 (79) HABIG WH, PABST MJ, JAKOBY WB: Glu-
tathione-5-transferase A: The first enzymatic step in mercapturic acid formation. J Biol Chem 249:7130-7139, 1974
References (7) GALE RP, FOON KA: Chronic lymphocytic leu-
kemia: Recent advances in biology and treatment. Ann Intern Med 103:101-120, 1985 (2) HILL BT: Studies on the transport and cellular distribution of chlorambucil in the Yoshida ascites sarcoma. Biochem Pharmacol 21: 495-502,1972 (3) BANK BB, KANGANIS D, LJEBES LF, ET AL:
(20) PANASCI L, HENDERSON D, TORRES-GARCIA SJ,
ET AU Transport, metabolism, and DNA interaction of melphalan in lymphocytes from patients with chronic lymphocytic leukemia. Cancer Res 48:1972-1976, 1988 (27)
JIANGB-Z,BANKBB,HSIANGY-H,ETAL:Lack
of drug-induced DNA cross-links in chlorambucil-resistant Chinese hamster ovary cells. Cancer Res 49:5514-5517, 1989
Chlorambucil pbarmacokinetics and DNA binding in chronic lymphocytic leukemia lym- (22) ROBSON CN, LEWIS AD, WOLF CR, ET AL: Reduced levels of drug-induced DNA phocytes. Cancer Res 49:554-559, 1989 crosslinking in nitrogen mustard-resistant Chi(4) FARMER PB: Metabolism and reactions of nese hamster ovary cells expressing elevated alkylating agents. Pharmacol Ther 35:301glutathione 5-transferase activity. Cancer Res 358, 1987 47:6022-6027,1987 (5) ROBSON CN, ALEXANDER J.HARRIS AL, ETAL:
Isolation and characterization of a Chinese hamster ovary cell line resistant to bifunctional nitrogen mustards. Cancer Res 46:6290-6294, 1986 (6~) ENDRESEN L, BAKKA A, RUGSTAD HE: In-
creased resistance to chlorambucil in cultured cells with a high concentration of cytoplasmic metallothionein. Cancer Res 43:2918-2926, 1983 (7) BULLER AL, CLAPPER ML, TEW KD: Glu-
tathione S-transferase in nitrogen mustard-resistant and -sensitive cell lines. Mol Pharmacol 31:575-578, 1987 (8) LEWIS AD, HKKSON ID, ROBSON CN, ET AL:
Amplification and increased expression of alpha class glutathione S-transferase-encoding genes associated with resistance to nitrogen mustards. Proc Natl Acad Sci USA 85: 8511-8515,1988
Detection of Etoposide Resistance by Measuring DNA Damage in Individual Chinese Hamster Cells Peggy L. Olive* Judit P. Bandth, Ralph E. Durand
(9) HARRAP KR, GASCOIGNE EW: The interaction
of bifunctional alkylating agents with the DNA of tumor cells. EurJ Cancer 12:53-59,1976 (70) HARRAP KR, HILL BT: The selectivity of action
of alkylating agents and drug resistance, d .
Vol. 82, No. 9, May 2, 1990
The comet assay, which measures DNA strand breakage in individual cells, was used to examine the relation between
DNA damage, cell survival, and resistance to the topoisomerase II inhibitor etoposide (VP-16). Chinese hamster V79*17H> cells and a VP-16-resistant subline (VP*) were exposed to VP-16 as monolayers or spheroids". The' comet" assay was comparable in sensitivity to the DNA precipitation and alkali unwinding assays for detecting DNA strand breaks induced by VP-16. However, unlike conventional DNA damage assays, the comet assay also indicated heterogeneity in cell response. For V79 multicell spheroids exposed to VP-16, the external cycling cells were 50 times more sensitive to killing and DNA damage than the internal noncycling cells; the comet assay indicated the fraction of cells resistant to the drug. V P cells, which were 10 times more resistant to killing and DNA damage by VP-16 than the parent cell line, could also be identified in mixed populations with the use of this method. These results suggest that the comet assay could be useful in predicting tumor cell response to DNAdamaging agents. [J Natl Cancer Inst 82:779-783,1990]
Development of new predictive assays for tumor response has attracted greater attention in recent years, perhaps as a result of an increased appreciation of the diversity of responses of individuals to cancer therapy (/). One of the most important properties of an effective predictive assay is its ability to detect a small proportion of resistant cells. Microelectrophoresis, or the "comet assay," is a rapid and relatively simple method for measuring DNA damage to individual cells, thus (theoretically) allowing identification of a small fraction of the cells that are resistant to treatment. This method was originally described in 1984 by Ostling and Johanson (2), and we subsequently adapted and
Downloaded from http://jnci.oxfordjournals.org/ at University of California, San Francisco on December 12, 2014
findings of Panasci et al. In our study, the CLL cells of the two patients who were chlorambucil-resistant showed a similar number of DNA cross-links to that of the cells from sensitive patients, when the resistant and the sensitive patients were treated with chlorambucil in vitro. However, the two resistant patients were receiving prednisone at the time of our study, and the influence of this agent on chlorambucil-induced DNA cross-linking in vivo, which may not have been observed in vitro, is unknown. Factors other than the formation of DNA cross-links, such as DNA repair, may contribute to alkylating-agent resistance in CLL. In this regard, the removal of melphalan-induced DNA cross-links in CLL cells in vitro appears to be more rapid in chlorambucil-resistant patients than in untreated patients (72). The area under the curve for DNA cross-linking may thus be the critical factor for cell kill, and the clinical responsiveness of CLL patients to chlorambucil should be correlated with the rate of formation and repair of DNA crosslinks.
Received October 27, 1989; revised January 26, 1990; accepted February 5, 1990. Supported in part by a grant from the British Columbia Health Care Research Foundation. British Columbia Cancer Research Centre, Vancouver, BC, Canada. Correspondence to: Peggy L. Olive, Ph.D., Medical Biophysics Unit, British Columbia Cancer Research Centre, 601 W. 10th Ave., Vancouver, BC, V5Z 1L3 Canada.
REPORTS
779
Materials and Methods Cell Lines and Culture Conditions Chinese hamster V79-171b cells were maintained as monolayers in exponential growth by twice weekly subcultivation in Eagle's minimum essential medium containing 10% fetal bovine serum (GIBCO Laboratories, Grand Island, NY). Single cells were prepared by trypsin treatment with the use of 0.1% trypsin (GIBCO) in phosphate-buffered saline. Spheroids were initiated by placement of 4 x 104 cells/mL in glass spinner-culture vessels (10). Spheroids were treated 10 days later when they had reached a diameter between 0.6 and 0.7 mm and contained approximately 103 cells. Cells resistant to VP-16 (VF cells) were obtained by treatment
780
with 10 jig of N-methyl-#'-nitro-#-nitrosoguanidine/mL followed by selection in ljigofVP-167mL(77). Cells were treated with VP-16 either as monolayers or in suspension when they were grown as spheroids. VP-16 was purchased from Bristol Laboratories, Belleville, ON, Canada, as Vespesid, a solubilized formulation in polysorbate 80 and polyoxyethylene glycol. It was diluted just prior to use in medium containing 5% fetal bovine serum. Following treatment, single-cell suspensions were prepared with the use of 0.1% trypsin for monolayers and 0.25% trypsin for spheroids. After 5 minutes, cells were resuspended in complete medium, centrifuged, and resuspended in phosphate-buffered saline purchased from GIBCO Laboratories.
590-nm band-pass filter. Individual images were digitized, and we calculated the tail moment, defined as the product of the percentage of DNA in the tail distribution and the displacement (in pixels) between the positions of the head and tail means. Data reduction was corrected for the background fluorescence as described in our previous publication (3).
Results
VP-16 caused significant numbers of DNA single-strand breaks, as detected with either the alkali unwinding assay or the DNA precipitation assay (fig. 1, panels A and B). For the former assay, DNA unwinding in a high-salt, weak-alkali lysis solution begins at sites of strand breakage; the rate of DNA unwinding has been found Measurements of DNA Damage to be proportional to the number of strand breaks (12). In comparison, the DNA preThe alkali unwinding assay of Ahncipitation assay measures the amount of strdm and Erixon (12) and the DNA precipitation assay were performed as previ- DNA precipitated when cells are lysed in ously described (13,14). For the comet alkaline 2% sodium dodecyl sulfate, folassay, 1.5 x 104 cells were suspended in lowed by precipitation with the use of 0.12 0.5 mL of phosphate-buffered saline on M potassium chloride (14). While based ice. Agarose (1.5 mL of 1% agarose; on different physical principles, both product No. 6013; Sigma Chemical Co., methods were equally sensitive in detectSt. Louis, MO) held at 50 °C was added to ing DNA damage by ionizing radiation the tube, and the suspension was rapidly (14) and, as shown here, were of similar pipetted onto a glass microscope slide. sensitivity in detecting VP-16-induced The sample was allowed to gel for 1 DNA damage. V F cells were about 10 minute, then placed in ice-cold lysing so- times more resistant to production of lution containing 0.03 M NaOH and 1 M strand breaks by VP-16 than the parent NaCl. After 1 hour in the dark at room V79 cell line (fig. 1, panel B). temperature, the slides were rinsed gently The comet assay was as sensitive as the in distilled water, then placed in a horizon- DNA precipitation and alkali unwinding tal gel electrophoresis apparatus (model assays were in detecting DNA damage by H4; Bethesda Research Laboratories, VP-16 (fig. 1, panel Q . In addition, since Inc., Gaithersburg, MD). Slides were cov- the cells need not be radiolabeled for the ered with 40 mAf Tris, 5 mAf edetic acid comet assay, results can also be obtained (EDTA), and 20 mAf acetic acid (pH 8.3). for Chinese hamster V79 spheroid cells, They were then exposed for 12 minutes to many of which are noncycling (10) and, 150 V (~4 V/cm) after which they were therefore, unable to incorporate radioacrinsed in distilled water and incubated for tive thymidine. Note that the average re10 minutes in the dark with 2.5 jig of sponse of cells from spheroids suggests propidium iodide/mL. After staining, greater resistance to VP-16 than that seen slides were rinsed twice in distilled water, with the V F cells, and the apparent nonplaced in a humidified air-tight container, linear shape of the dose-response curve and viewed within 48 hours. for spheroid cells could also indicate hetWe viewed 40-200 individual comets erogeneity in response. using a 25 x objective with a Zeiss epiAll three DNA damage assays were fluorescence microscope with attached sensitive to VP-16-induced damage at bihigh-sensitivity video camera and image ologically relevant doses, as indicated by analysis system (15). Cells were illumi- the cell survival data shown in figure 1 nated with a 546-nm light excitation from (panel D). The V F cells showed a change a 100-W mercury light. Emission was in both shoulder and terminal slope of the monitored using a 580-nm reflector and a dose-response curve; the slope of the cell Journal of the National Cancer Institute
Downloaded from http://jnci.oxfordjournals.org/ at University of California, San Francisco on December 12, 2014
modified it for image analysis (3). In the comet assay, a small number of single cells are embedded in agarose, lysed, subjected to an electric field, stained with a fluorescent DNA-binding dye, and visualized with the use of a fluorescence imageprocessing system. For cells containing DNA strand breaks, the electric current pulls the negatively charged broken and relaxed DNA away from the nucleus toward the anode (2,4). In undamaged cells, most of the supercoiled DNA remains within the nuclear cage, but a damaged cell resembles a "comet" with a brightly fluorescent head and a tail containing the fragments and relaxed supercoils of DNA. Previous results with this method indicate that (a) DNA single-strand breaks are recognized, (b) damage from radiation doses of 2.5 Gy or lower can be detected, and (c) heterogeneity in DNA damage within a population of cells is readily apparent (2-5). This assay thus appears to show promise as a means of identifying cells resistant to drug therapy. Etoposide (VP-16), an epipodophyllotoxin that traps topoisomerase II during cleavage of DNA, is an effective inducer of DNA damage, producing single- and double-strand breaks and DNA-protein cross-links (6-9). To evaluate the sensitivity and reliability of the comet assay to recognize drug damage, we compared this method with more conventional methods used to measure DNA damage. The usefulness of the comet assay for predicting cell survival in a mixture of VP- 16-resistant and -sensitive cells was also examined.
10
20
T"^-^^ •fc*. 1
30
40 1
D.
vs A;
10
20
Etoposide
30
40
(|jg/ml)
1
VP r
>, E)
•
0 10 20 30 40 Etoposide (pg/ml)
Figure 1. Panels A and B: DNA damage by VP-16 detected with the use of the alkali unwinding assay (o,«) or the DNA precipitation assay ( A , • ) . [ |i4C]Thymidine-labeled exponentially growing Chinese hamster V79 cell monolayers (V79m) or a resistant subline (VP1) were incubated for 1 hr with VP-16 at concentrations shown. Cells were analyzed for DNA damage as previously described (77). Mean and standard deviation for three determinations are shown. Note difference in drug concentrations used for drug-sensitive (panel A) and drug-resistant (panel B) cells. Panels C and D: Toxicity and DNA damage by VP-16 measured with the comet assay. Chinese hamster V79 cell monolayers (Vm), 0.6-mm-diameter V79 spheroid cells (Vs), and VP* cell monolayers (VP1) were exposed for 1 hr to VP-16. Panel C: DNA damage was measured in 80-200 individual cells per dose point by the comet assay. Mean and standard deviation for three separate experiments are shown for selected concentrations, x , cell monolayers; o, VF* cells; A, spheroid cells. Panel D: Cell clonogenicity is average of two to four experiments.
monolayer survival curve was about 10 times steeper than the terminal slope of the VI* cells. The dose-response curve for the spheroid cells appeared biphasic, and a significant fraction of the spheroid cells was about 50 times more resistant to VP16 than were the exponentially growing cell monolayers. Extrapolating the terminal region of the spheroid cell survival curve indicated that about 50%-60% of the spheroid cells were resistant to VP-16; this percentage is consistent with the fraction of noncycling cells within these spheroids (10). The important advantage of the comet assay is its ability to detect damage in individual cells and, therefore, to measure heterogeneity. For V79 spheroids, the external cycling cells respond to treatment with VP-16, but the internal, noncycling cells are resistant (16). Measuring the average amount of DNA damage does not provide this information; a decrease in average DNA damage would be observed if all cells were somewhat moreresistantto the drug or if only a fraction of the cells Vol. 82, No. 9, May 2, 1990
was considerably more resistant. The comet assay indicated that more of the spheroid cells (fig. 2, panel A) were resistant to VP-16 than monolayers of either drug-resistant (fig. 2, panel B) or -sensitive cells (fig. 2, panel C). Correlating tail moment and surviving fraction indicated that one can also use the average amount of damage detected by the comet assay to predict cell clonogenicity (fig. 3). All of the results were fitted to the same linear-quadratic curve, although the results with the V79 cell monolayers were on the left of this curve and the results with the spheroid cells were generally on the right. From the data shown in figure 2 (panel A), one might expect the spheroid cells to show greater survival for the same average tail moment, since a larger fraction of the spheroid cells showed minimal damage by VP-16. We examined the possibility of dispensing with the video imaging system and simply measuring drug resistance by counting, with the use of only the fluorescence microscope, the fraction of comets
10
20 30 TAIL MOMENT
40
Figure 2. Heterogeneity in DNA damage by VP-16 detected by the comet assay. V79 spheroid cells, VP cell monolayers (VP1), and Chinese hamster V79 cell monolayers (V79m) were incubated for 1 hr with VP-16 at concentrations that produced average tail moments between 10 and 15 U. Surviving fraction (S.F.) is also shown for same population of treated cells.
without long tails (i.e., the undamaged or less damaged cells). Samples of V F cells were mixed with the parent V79 cells at various ratios. The mixtures were incubated with 2 jig of VP-167mL for 1 hour and were prepared for the comet assay on coded slides. Five observers familiarized
1.0
z o SURVI\/ING FRACT]
D
Vm
Downloaded from http://jnci.oxfordjournals.org/ at University of California, San Francisco on December 12, 2014
•0.01
o
: ^ \ 0. 1 i
A
\ 0.01
1
x
-
r
\
:
.001 x 10
0
TAIL
: 20
MOMENT
Figure 3. Comparison between tail moment and clonogenicity for V79 cell monolayers (x), V F cells (o), and multicell spheroids (A). Data are fitted to a linear quadratic equation.
REPORTS
781
Table 1. Visual identification of VP-16-resistant cells by the comet assay* Actual % resistant cells
% scored resistant by Observer 1
5
15
10 25 50 75 90 95
15.5 40.5 46.5
r2 Slope
81 74 92.5 0 94 0.84
Observer 2
Observer 3
Observer 4
12.5
30 29 72 53 60 82 94 071
11.5 17.5
22 44.5 47.5
65 77 92 0.94 0.74
0.55
44 57 48 78.5
96 0.85 0.75
Observer 5
18 15.5 26.5
26 75.5
69 75 0 88 0.69
themselves with the appearances of comets from 100%resistantand 100% sensitive cells. They then examined 200 comets on each coded slide and rated each comet as damaged Gong tail, dim head) or undamaged (short tail, bright head). As table 1 indicates, the ability to detect resistant cells visually, without using criteria developed for the imaging system, was surprisingly good. In addition, only 10-20 minutes was required for visual observation of 200 comets compared with about 200 minutes for analysis using our current software for the fluorescence image-processing system.
Discussion Several criteria must be met for an assay to be a useful method for detecting resistant cells within a heterogeneous population: (a) The agent used must cause damage of a type that can be detected by the assay, (b) The resistant cells must be several times more resistant to the agent than the sensitive cells; how many times will depend on the proportion of resistant cells present and the number of cells analyzed. (c) The dose of the agent used must be sufficient to allow distinction between damaged and undamaged (untreated) cells, (d) A sufficient number of cells must be analyzed; if only 5% of the population is resistant to the treatment, then analyzing 20 cells is statistically inadequate, (e) Finally, the method should be relatively rapid and simple to perform. The comet assay enabled the experienced observers to visually identify resistant cells in mixtures of V F and V79 cells.
782
Other methods can measure DNA damage to individual cells (18-21) and may provide comparable information. However, the speed and simplicity of the comet It shows promise as a rapid, inexpensive assay and the possibility of using it in the method for detecting VP-16-resistant cells absence of sophisticated image analysis in tumors, provided the comets arereadon hardware and software to screen samples coded slides. More accurate identification give it a decided advantage over these is possible as the degree ofresistanceto the other methods. drug increases. A major problem in detecting small fractions of resistant cells (aside References from the number of comets that must be (/) BROCK WA, BAKER FL, TOFILON PJ: Tumor analyzed) is the heterogeneity within even cell sensitivities to drugs and radiation. In the relatively homogeneous V79 cell Prediction of Tumor Treatment Response (Chapman JD, Peters LJ, Withers HR, eds). monolayers. The large variation in tail New York: Pergamon Press, 1989, pp 139-155 moment associated with a homogeneously (2) OSTUNG O, JOHANSON KJ: Microelectroresponding population is not easily exphoretic study of radiation-induced DNA damages in individual mammalian cells. Biocbem plained. When the average tail moment of Biophys Res Commun 123:291-298, 1984 cells treated with 2 p,g of VP-16/mL is 10 (3) OLIVE PL, BANATH JP, DURAND RE: Heterogetimes that of the untreated cells, 3%-6% of neity in radiation-induced DNA damage and repair in tumor and normal cells measured using the treated comets still have tail moments the comet assay. Radial Res. In press that overlap the control distribution. This (-0 OSTUNG O, JOHANSON KJ: Bleomycin, in consituation immediately places a limit on the trast to gamma irradiation, induces extreme ability to detect small fractions of resistant variation of DNA strand breakage from cell to cell. Int J Radiat Biol Relat Stud Phys Chem cells. Determining whether this variability Med 52:683-691,1987 is a biological or a technical problem with (5) SINGH NP, MCCOY MT, TKE RR, ET AL: A the assay is of primary importance. simple technique for quantication of low levels
Cycling cells in V79 spheroids appear to be about 50 times more sensitive to VP-16 at a surviving fraction of 0.1. Therefore, tumor growth fraction might be estimated using the comet assay by the determination of the percentage of cells that respond to VP-16. While topoisomerase II levels drop precipitously in noncycling cells (17), noncycling cells in V79 spheroids can still be killed by VP-16. The comet assay can predict cell clonogenicity using the average tail moment (fig. 3), but it is of limited value if only a small percentage of resistant cells is present. It is also possible to calculate a cutoff value for the tail moment. This value can be used to distinguish damaged
of DNA damage in individual cells. Exp Cell Res 175:184-191,1988 (6~) KALWINSKY DK, LOOK AT, DUCORE J, ET AL:
Effects of the epipodophyllotoxin VP-16-213 on cell cycle traverse, DNA synthesis, and DNA strand size in cultures of human leukemic lymphoblasts. Cancer Res 43:1592-1597,1983 (7) Liu LF, ROWE TC, YANG L, ET AL: Cleavage of
DNA by mammalian DNA topoisomerase D. J Biol Chem 258:15365-15370,1983 (S) Ross W, ROWE T, GUSSON B, ET AL: Role of
topoisomerase II in mediating epipodophyllotoxin-induced DNA cleavage. Cancer Res 44:5857-5860, 1984 (9) LONG BH, MUSIAL ST, BRATTAIN MG: Single-
and double-strand DNA breakage and repair in human lung adenocarrinoma cells exposed to etoposide and teniposide. Cancer Res 45: 3106-3112,1985 (10) SUTHERLAND RM, DURAND RE: Radiation re-
sponse of multicell spheroids—an in vitro tumour model. Curr Top Radiat Res Q 11: 87-139, 1976
Journal of the National Cancer Institute
Downloaded from http://jnci.oxfordjournals.org/ at University of California, San Francisco on December 12, 2014
•Mixtures of V79 cells and VPf cells were incubated with 2 /»g of VP-16/mL. Each observer "blindly" scored 200 comets per slide. Observers 1 and 2 were experienced with the assay, the others were not The correlation coefficient (r2) for the linear best-fit line and the slope of that line are shown for each data set
from undamaged cells. A line extrapolated from the terminal region of the curve shown in figure 3 crosses the 100% survival level at a tail moment close to 5. This means that cells with a tail moment of 5 or less might be expected to survive damage by VP-16. With this criterion, the calculated survival rates from figure 2 (panels A-C) are 0.34, 0.36, and 0.05, respectively, which can be compared with the measured survival rates of 0.13,0.20, and 0.01, respectively. The agreement between the calculated and measured values would, of course, be expected to improve as more comets are analyzed.
(11) GUPTA RS: Genetic, biochemical, and crossresistance studies with mutants of Chinese hamster ovary cells resistant to the anticancer drugs, VM-26 and VP-16-213. Cancer Res 43: _ . 1568-1574,1983 (72) AHNSTROM G, ERTXON K: Radiation induced
strand breakage in DNA from mammalian cells. Strand separation in alkaline solutionTlnt J Radial Biol Relat Stud Phys Chem Med 23:285-289, 1973 (13) OUVE PL, CHAN AP, CU CS: Comparison
between the DNA precipitation and alkali unwinding assays for detecting DNA strand breaks and cross-links. Cancer Res 48: 6444-6449, 1988 (14) OUVE PL: DNA precipitation assay: A rapid and simple method for detecting DNA damage in mammalian cells. Environ Mol Mutagen 11:487^195,1988 (15) JAGGI B, POON SS, MACAULAY C, ET AL:
(16) DURAND RE, GOLDIE JH: Interaction of etopo-
side and cisplatin in an in vitro tumor model. Cancer Treat Rep 71:673-679, 1987 (17) HSIANG Y, Wu H, Liu LF: Proliferation-depen-
dent regulation of DNA topoisomerase II in cultured cells. Cancer Res 48:3230-3235,1988 (18) Ron Ron JL, WRIGHT WD: Visualization of
DNA loops in nucleoids from HeLa cells: Assays for DNA damage and repair. Cytometry 8:461-467, 1987 (19) NOSE K, OKAMOTOH: Detection of carcinogen-
induced DNA breaks by nick translation in permeable cells. Biochem Biophys Res Commun 111:383-389, 1983 (20) RAZA A, MEHDI I, MAYERS G, ET AU In vitro
detection of single stranded DNA damage (SSD) in individual cells: A useful test for detecting reversal of multi-drug resistance (MDR). Proc Am Assoc Cancer Res 30:597, 1989 (21) DEEN DF, KENDALL LE, MARTON LJ, ET AL:
Prediction of human tumor cell chemosensitivify using the sister chromatid exchange assay. Cancer Res 46:1599-1602, 1986
vV-Nitrosodimethylamine Blood Levels in Patients With Chronic Renal Failure: Modulation of Levels by Ethanol and Ascorbic Acid S. R. Dunn* M. L. Simenhoff, P. S. Leie, S. Goyal, J. W. Pensabene, W. Fiddler
We measured levels of iV-nitrosodimethylamine (NDMA) in peripheral blood from 13 fasting male patients, 30-74 years old, who had chronic renal failure, and in five healthy control subjects (four males and one female) 31-50 years
Vol. 82, No. 9, May 2, 1990
precursors and 4-methylpyrazole (17). Study results have also indicated that bacteria may play an active role in nitrosamine formation (18-24), In addition, chronic renal failure in humans has been shown to be a favorable model for studying intestinal nitrosamine production (a) because NDMA and N-nitrosopiperidine (NPIP) have been found in duodenal aspirates in concentrations greater than those found in the blood (11,25,26) and (b) because bacterial overgrowth occurs in the duodenum (27). If NDMA, and possibly other N-nitroso compounds, contribute to the total human exposure to carcinogens, it would be prudent to reduce the endogenous component in the gastrointestinal tract. One approach would be to eliminate the intestinal bacterial overgrowth; another would be to block NDMA formation. Ascorbic acid (vitamin C) has already been shown to be an effective inhibitor of nitrosation in human and animal studies (17,28-31). In this study, we investigated NDMA formation in patients with chronic renal failure. We used ethanol to block first-pass hepatic metabolism of NDMA derived mainly from the portal circulation and then measured the circulating NDMA released to the peripheral blood and target organs from the liver. We then examined the effect of oral ascorbic acid on modulation of gastrointestinal NDMA formation under the same conditions.
The role of N-nitroso compounds as contributors to the etiology of human cancer has been under investigation for more than 30 years since the original findings by Magee and Barnes (1) of the hepatotoxicity and carcinogenicity of N-nitrosodimethylamine (NDMA) in the rat. As a Materials and Methods class, nitrosamines are remarkable carcin- Patient Selection and Study Design ogens, as shown by their potency, their Patients entered in the study were unaction in many species, and the variety of dergoing hemodialysis three times per target organs (2). Direct evidence of in vivo formation of NDMA in humans has been difficult to Received August 1, 1989; revised February 15, obtain, owing to its ubiquitous presence in very low concentrations. However, there 1990; accepted February 23,1990. Supported in part by Public Health Service grant is evidence that N-nitroso compounds may CA-26571 from the National Cancer Institute, Nabe formed in vivo (3,4). tional Institutes of Health, Department of Health and Theoretically, conditions for endoge- Human Services. S. R. Dunn, M. L. Simenhoff, P. S. Lele, nous formation of N-nitroso compounds, particularly NDMA, would be favored in S. Goyal, Department of Medicine, Division of Nephrology, Jefferson Medical College, Thomas the normal gastrointestinal tract (5,6), and Jefferson University, Philadelphia, PA. this hypothesis is supported by reports of J. W. Pensabene, W. Fiddler, U.S. Department of elevated levels of total N-nitroso com- Agriculture, Eastern Regional Research Center, Philpounds in human stomach aspirates (7) adelphia, PA. We thank Hyman Menduke, Robert Gates, and and in the large intestine of rats (8). Nancy Picarello for their assistance. NDMA has been found in saliva, blood, Correspondence to: Stephen R. Dunn, Departurine, and gastric and duodenal aspirates ment of Medicine, Division of Nephrology, Jefferson in healthy individuals (9-16) and, re- Medical College, Thomas Jefferson University, Philcently , in the urine of ferrets given NDMA adelphia, PA 19107.
Downloaded from http://jnci.oxfordjournals.org/ at University of California, San Francisco on December 12, 2014
Imaging system for morphometric assessment of absorption or fluorescence in stained cells. Cytometry 9:566-572, 1988
old. In the patients, we found significant (P < .01) levels of NDMA (mean ± SD; 201 ± 111 ng/kg of blood), which is known to be carcinogenic in animals. Five minutes after oral administration of ethanol (0:4 g/kg of body weight) j all patients exhibited a significant (P < .01) rise in blood NDMA levels (338 ± 125 ng/kg), suggesting continuous endogenous formation of NDMA that was unmasked by ethanol's ability to inhibit first-pass hepatic metabolism of NDMA. In five of six patients, pretreatment with oral ascorbic acid resulted in a blunting, but not statistically significant, effect on maximum blood NDMA levels after consumption of ethanol. Mean levels were 340 ± 100 ng/kg before treatment with ascorbic acid and 237 ± 127 ng/kg during treatment. Ethanol administration unmasks increased gastrointestinal formation of NDMA in patients with chronic renal failure. Further studies are required to confirm a possible link between endogenous NDMA formation and the increased incidence of cancer in these patients. [J Natl Cancer Inst 82:783-787,1990]
REPORTS 78;
