Use of the Comet Assay to Identify Cells Sensitive to Tirapazamine in Multiceli Spheroids and Tumors in Mice1

ICANCKR KlÃ¬SIiAKCM56.446(1-4463. Oclobcr I. IÂ«W6| Use of the Comet Assay to Identify Cells Sensitive to Tirapazamine in Multiceli Spheroids and Tumors in Mice1

Peggy L. Olive,2 Charlene M. Vikse, and Judit P. BanÃ¡th

Metileni /Â¿/fi/i/iv.v/'r.vDffitirniii'iii.BrÃ¬tÃ¬\liColumbiaCtintrr Ri'scturh Centre. 601 Wc\i 10th Avcniu', Vtinciniver. Rriiish Ctilitnihui V5Z IL3. Ctintulit

ABSTRACT lution to detect subpopulations that differ in damage by as little as a factor of 2 (4). Because the tumor cells that survive treatment with a DNA- Tii apa/aminc. a bioreductive ilnii; preferentially toxic to hypoxic cells, damaging drug are likely to be those that do not harbor extensive DNA produces significant numbers of I)N \ single-strand breaks that can be de damage following treatment, a DNA damage-based assay is appealing as tected using; the alkaline comet assay. Our goal was to determine whether single-strand breaks measured using this assay could act as a surrogate end a surrogate for cell killing. point tor cell killing in multiceli spheroids and solid tumors in mice. In Tirapa/amine (SR4233) is a hioreductive drag that is preferentially spheroids composed of Chinese hamster V79 cells. \ViI)r human colon car toxic to hypoxic mammalian cells (5. 6). Phase I and II clinical trials are cinoma cells, or Sil la human cervical carcinoma cells, histograms of tail currently underway to evaluate the use of this drug in combination with moments (indicators of DNA damage in the comet assay) could he used to radiation and chemotherapy. We have previously used the alkaline comet identify the percentage of cells that sustained sufficient DNA damage to cause assay to measure DNA damage in multiceli spheroids, murine tumors, cell death after treatment with tirapazaminc. The proportion of comets with and normal tissues exposed to tirapa/amine (7. 8). This drug produces tail moments Â£20(i.e.. with damage comparable to that produced In about large numbers of DNA single-strand breaks in hypoxic cells even at 1U (.\ i correlated with cell survival irrespective of cell type, dose of lira pazamine, time of treatment, or position of cell in the spheroid. Single-cell nontoxic drug concentrations. Both strand breaks and cell killing increase suspensions from squamous cell carcinoma Nil tumors in C3H mice or Sil la with dose and time of exposure, suggesting that strand breaks might xenografts in severe combined Â¡mmunodeficicnt mice were also analy/ed for correlate with (if not cause) cell killing. Within subpopulations ot sphe clonogcnicity and DNA damage. Again, the percentage of comets with tail roids and tumors, cell-sorting experiments indicated that cells that sus moments 220 was found to be a good predictor of cell killing for both tumor tained the most DNA damage were also the most likely to die following types, providing tumor samples were obtained no more than l h after i.p. treatment (7). However, if DNA damage is to be used as a surrogate for drug administration. Because tirapa/aminc is currently undergoing clinical cell killing, it is necessary to show that a specific amount of DNA damage trials, application of this procedure could provide an early indicator of is predictive for cell killing independent of the dose of the drug, hypoxic tumors likely to contain hypoxic. susceptible cells and also the extent of the status, time of exposure, or cell type. Therefore, we tested whether DNA response. damage by tirapa/amine. measured using the comet assay, could be an effective predictor of cell killing in multiceli spheroids and tumors in INTRODUCTION mice and might therefore have potential applicability in the clinic.

Many genetic, pharmacological, and microenvironmental factors can influence the response of a tumor io a specific cytotoxic therapy. MATERIALS AND METHODS Although it is important to understand how all of these factors individually affect treatment outcome, a more immediate and practical Spheroid Culture. Chinese hamster V79-l71b lung fibrohlusts were main tained in exponential monolaver growth by subcultivation twice weekly in problem is to simply estimate the overall impact of all genetic and Eagle's MEM containing IO'* FBS. ' SiHa 1mm.m cervical carcinoma cells and epigenetic factors on tumor cell killing by a specific cytotoxic treat WiDr human colon carcinoma cells were purchased from the American Type ment. In an effort to develop a rapid assay to assess tumor suscepti Culture Collection and maintained in exponential growth by Iwice-weekly bility to treatment, we have examined the ability of the comet assay to Subcultivation in MEM plus l()r/r FBS. Spheroids were initiated by seeding act as a surrogate end point for cell killing. 2 X 104 cells/ml into BÃ©licospinner culture vessels containing MF.M plus \()ci In 1984, a single-cell gel electrophoresis method was developed that FBS. Spheroids were ted after 3 days and daily thereafter by replacing spent allowed measurement of DNA damage in individual cells (1). This medium with MEM plus 5*/i FBS supplemented with antibiotics. technique was later modified and adapted for image analysis, providing Spheroids were exposed to tirapa/amine in MEM plus 5'< F'BS in BÃ©lico greater sensitivity and resolution of subpopulations (2. 3). In this method, glass spinner culture vessels. Vessels containing spheroids were equilibrated single cells are embedded in agarose on a microscope slide, then lysed to with l()r/i oxygen. 5r/< carbon dioxide, and 859Ã•nitrogen at 37Â°Cfor l h prior remove proteins, exposed to low-voltage electrophoresis. and stained to and during incubation with the drug. Following incubation, spheroids were washed and disaggregated by exposure in an orbital shaker at 37Â°Cfor 8-10 with a fluorescent DNA-binding dye. The amount of DNA that migrates is proportional to the number of DNA breaks present in the cell. This min in 0.25Ã‡ÃŽtrypsindissolved in citrate saline. Single cells were centrifuged "comet" assay is versatile and able to detect DNA single-strand breaks, and resuspended in fresh cold medium. Cells were analy/ed lor clonogenicilv by plating cells in l(K)-mm tissue culture dishes containing 10 ml MEM plus double-strand breaks, or cross-links in virtually any cell type, provided 10% FCS. Colonies were counted X-14 days after treatment. Plating efficien that a single-cell suspension can be obtained (3). The method has also cies of cells from V79. SiHa. and WiDr spheroids were 0.64 Â± 0.04, been applied to human tumor samples obtained by tine-needle aspiration 0.66 Â±0.09. and 0.17 Â±0.06 (mean Â±SD). respectively. The same popula biopsy (4). a relatively safe and simple procedure, which is applicable to tions examined for clonogenicity were also examined for DNA damage using many human tumors. Not only is the comet method sensitive to low the alkali comet assay. levels of DNA damage, but more important, it provides sufficient reso- Tumor Cells. SCCVII squamous cell carcinoma cells were transplanted s.c. over the sacral region of inbred male C3H/HeN mice, approximately 30 g in weight. Tumors were used for experimentation approximately 2 weeks later Received 6/1X/96: accepted 7/30/96. The costs of publication of this article were defrayed in pan by the payment of page when they had reached a weight of 350-500 mg. SiHa human cervical charges. This article must therefore be hereby marked uJrenisemeni in accordance with carcinoma cells were transplanted s.c. in the back of CB-17 severe combined 18 U.S.C. Section 1734 solely to indicate this fact. 1This work was supported by a grant from the National Cancer Institute of Canada. 2 To whom requests for reprints should be addressed. Phone: (6()4) 877-6(XK). ext. 'The abbreviations used are: FBS. fetal bovine serum: SCCVII. squamous cell 3024: Fax: (604) 877-0743: E-mail: olive6Â»unixg.uhc.ca. carcinoma VII: CCD. charge coupled device. 4460

Downloaded from cancerres.aacrjournals.org on October 4, 2021. © 1996 American Association for Cancer Research. PREDICTINGTUMORRESPONSI;TOTIRAPAZAMINE immunodeficient mice. Tumors were used for experimentation approximately c 4-5 weeks later when they had reached a weight of 300-500 mg. O Tirapazumine (SR4233) was kindly provided by Dr. Martin Brown (Stan 10min ford University). Mice were injected i.p. with lirapazamine from a stock 30min solution of 1-2 mg/ml in PBS. One h after i.p. injection, mice were sacrificed, O) and tumors were rapidly excised and placed in ice-cold PBS. A single-cell C suspension was prepared from the entire tumor by finely mincing the tumor and incubating for 30 min with a mixture of trypsin. collagenase. and DNase 0.1 r as described previously (9). Cells were then filtered through 30-fim nylon 3 CO mesh, centrifuged. and resuspended in cold medium. Plating efficiency in MEM plus 10% PCS was 0.28 Â±0.04 for cells from SCCVII tumors and -+â€”' 0.22 Â±0.06 for SiHa cells. 0) 30 Hoechst 33342 Cell Sorting. During the final 20 min of drug exposure, 20 multiceli spheroids were incubated with 1 fig/ml Hoechst 33342 to selectively o recover cells from different depths within the spheroid (10). At the end of drug 10 exposure, spheroids were rinsed, disaggregated with trypsin. and sorted into 10 CO fractions according to Hoechst 33342 fluorescence concentration using a O Becton Dickinson FACS 440 cell sorter. O 20 40 60 Mice were injected i.v. with 0.1 ml of an 8-mg/ml stock solution of Hoechst 33342 in PBS; then 10 min later, animals were sacrificed, tumors were Tirapazamine removed, and a single-cell suspension was prepared for fluorescence-activated O cell sorting (II). Cells were sorted into eight fractions according to the Hoechst 1 fluorescence gradient. 1 DNA Damage Measured Using the Alkaline Comet Assay. Single cells LL 0.1 from spheroids or tumors were suspended in ice-cold PBS at a concentration O) of 2-4 X IO4 cells/ml. Then. 0.5 ml cell suspension (IO4 cells) was placed in C a5-ml disposable tube, and 1.5 ml of l<7rlow gelling temperature agarose (type 0.01 VII. Sigma Chemical Co.; prepared in distilled water and held at 40Â°C)were 0.001 added and mixed with the cells. Approximately 1.5 ml of this mixture were CO 10 20 30 quickly pipetted onto an agarose-coated. half-frosted microscope slide and allowed to gel. Slides were carefully submersed in an alkaline lysis solution Tail Moment containing 1 M NaCI. 0.03 M NaOH. and 0.29Ã•sarkosyl for 1 h. followed by Fig. I. Toxicity and DNA damage by tirapazamine in cells from SiHa spheroids. a l-h wash in two rinses of 0.03 M NaOH plus 2 m.MEDTA before electro- Multiceli spheroids containing approximately 50^ radiobiologÃa)]}1 hypoxic cells were phoresis in a fresh solution of 0.03 M NaOH plus 2 mst EDTA at 0.6 V/cm for exposed for IO min to 4 h io various concentrations of tirapazamine. Single cells from 25 min. Slides were rinsed and stained for 20 min in 2.5 (Â¿g/mlpropidium spheroids were analyzed for clonogenicity (u) and DNA single-strand breaks using the iodide. alkali cornel assay (h). c. cell survival as a function of the average tail moment using data in a and b as well as several additional experiments. Individual cells or comets were viewed using a Zeiss epifluorescence microscope attached to an intensified solid-state CCD camera and image analysis system. For viewing propidium iodide fluorescence, slides were illuminated with green light from a 100-W mercury source using a 580-nm population in which some of the cells are undamaged and some of the reflector and a 590-nm barrier filter. Individual comets were viewed using a cells are heavily damaged. The latter situation is the one of concern X25 objective, and 2(X) or more images were analyzed using a fluorescence when measuring tumor response to a cytotoxic agent, because undam image-processing system (2). As the number of DNA strand breaks increased, aged cells will survive treatment. The amount of DNA damage in cells the amount of DNA able to migrate away from the comet head increased in destined to die is largely irrelevant but will impact on the average tail proportion to the dose (7). The "tail moment," defined as the product of the moment. percentage of DNA in the comet tail and the distance between the means of the Therefore, to predict cell killing based on DNA damage accurately. head and tail distributions, and "DNA content." defined as the total fluores it is essential to examine the response of individual cells in the comet cence associated with an image, were the most informative features (2). assay. Heterogeneity in DNA damage is evident in tail moment histograms generated for SiHa spheroids exposed to tirapazamine (Fig. 3, a-d). If one assumes that all of the comets with tail moments RESULTS s2() represent cells that would survive treatment, whereas all of the Tirapazamine produces DNA damage and cell killing in SiHa. comets with tail moments >20 would die, then tail moment histo WiDr. and V79 spheroids, which increase with both time of exposure grams taken from the experiment in Fig. 1, a and /;, accurately predict and drug dose. Representative data are shown for SiHa spheroids in cell survival (Fig. 3). Unfortunately, however, the aver however, in these combined sets, the correlation between measured age amount of DNA damage in a population of cells is not a good and predicted surviving fractions is still remarkably good. The slope predictor of cell killing in a heterogeneous system such as a spheroid of the line is 0.92 for all cell lines. or tumor, because a single value could describe a population in which Promising results using multiceli spheroids prompted us to perform all of the cells show an intermediate level of DNA damage or a similar experiments using two solid-tumor models: SCCVII mouse 4461

a. Control a. 60 80 C 40 g â€¢¿SiHa o v V79 co 0 CO o WiDr 0) [b. 15 MM, 15 min en e o 20 O

O D (D re. 30 MM, 1 h CO O) CO C 5 0.1 -, O) O 01234 CD O â€¢¿IliÃ¨i l Ã»- cd. 60 MM, 4 h Incubation Time (h) b.

e 0 "og 0 10 20 30 40 50 60 co Tail Moment

O) C S? LO_e. Tail Moment < 20 'E co D CO T3 0.5 0.1 K 10 20 30 CO CO Tail Moment 0.0 Fig. 2. Comparison between the response of cells from three types of spheroids 0.0 0.5 1.0 exposed to tirapa/amine. The surviving traction {ni is compared with the average tail moment measured tor the same populations {b}. Predicted Survival

Fig. 3. Analysis of lail moment histograms as a method of determining surviving traction, a tl. reprÃ©sentaii\ e histograms (6(M) comÃ©iseach) tor SiHa spheroiils e\poseil to squamous carcinoma growing s.c. in C3H mice and SiHa human lirapa/.amine. e. fraction of comets with tail moment :=20 compared with the measured cervical carcinoma in severe combined immunodeficient mice. Mice surviving fraction for the same populations of spheroids. Data in e include only the results were injected i.p. with 4-60 mg/kg tirapa/.amine. and 60 min later, shown in Fig. I. u and h. the mice were sacrificed, tumors were removed and dissociated using enzymes, and single cells were analyzed for clonogenicity and DNA -Â»â€¢--SiHa C 1.0 O --v-- V79 damage. Heterogeneity in DNA damage was also observed in the cells "o of these tumors (Fig. 5, Â«-<â€¢).consistentwith previous experiments -O- WiDr using SCCVII tumors (7). A tail moment of <20 again provided a CO useful discriminator for cells that are clonogenic in vitro (Fig. 5il). A O) complication for exposure in vivo is that the plasma half-life of C tirapazamine in mice is estimated to be 15-30 min (12. 13). and the half-time for DNA strand break rejoining is 1-2 h (7). Therefore, it is 0.5 not surprising that tumor samples removed more than l h after CO injection showed significant DNA repair so that the same discrimi â€¢¿o nator was not accurate in predicting cell survival (Fig. 5i/). co DISCUSSION CO 0) DNA strand breakage appears to be a useful surrogate end point for cell killing by tirapazamine. However, the average amount of DNA 0.0 damage sustained by a heterogeneous population of cells is not 0.0 0.5 1.0 adequate to predict cell survival. Instead, the proportion of cells that Predicted Surviving Fraction demonstrate insufficient DNA damage for cell killing is a better l:Ã¬jl.4. Comparison between measured and predicted surviving fractions l'or multiple indicator of cell survival following exposure to tirapazamine. experiments performed using three types of spheroids. The fraction of cornels with tail The tail moment discriminator of 20 was reasonably accurate in moments ^20 was compared with the measured surviving fraction. l.inc\, linear regres describing the responses of three different rodent and human tumor sion lines with slopes of 0.9-0.92. 4462

ta. Control amounts and types of DNA damage to be killed by this drug. As a caveat, it should be recalled that both spheroids and tumors were disaggregated immediately after treatment, therefore providing the *2 nutrient-deprived cells with a more favorable environment for DNA repair. E We previously attempted to use tirapazamine-induced DNA dam o b. 20 mg/kg age as an indicator of the presence of radiobiological hypoxia in M- 8 spheroids and tumors (7. 8). Unfortunately, there was no clear dis O tinction between the response of aerobic and hypoxic cells in tumors, O) which is the same conclusion from data shown in Fig. 5, b and c. CO 4â€”* Therefore, it is interesting that DNA damage predicts so well for cell Â§ o killing by this drug. This apparent contradiction (tirapa/amine pref E [ e. 60 mg/kg erentially kills hypoxic cells, and DNA damage predicts cell killing, 0) but DNA damage does not predict hypoxic fraction) may be explained n_ 4 by the ability of tirapazamine to induce large numbers of strand breaks that are not lethal and to cause DNA damage and cell killing in cells that border radiobiological hypoxia. Our other studies indicate that the excellent correlation observed using tirapa/amine is often, but not invariably, observed with other Tail Moment DNA-damaging agents. Moreover, tirapazamine is unique in that it induces significant numbers of DNA strand breaks without killing _ 1.0 CO cells. From the perspective of developing a predictive assay, this is important for two reasons: (Â«Iclinically relevant doses of this drug should produce sufficient DNA damage for accurate detection using i the comet assay: and (h) any "background" DNA damage induced by the processes of tumor disaggregation or fine-needle aspiration biopsy T3 0.5 2? becomes largely irrelevant. Clinical protocols using tirapa/amine. therefore, may benefit by application of the comet assay not only as an indicator of tumor or normal tissue response, but also as a method I of identifying those patients most likely to benefit from this drug.

0.0 ACKNOWLEDGMENTS 0 0.5 1.0 The helpful comments of Dr. Ralph Durand are gratefully acknowledged. Predicted Survival Fig. 5. Analysis of DNA damage and survival in SCCVII and SiHa tumors, ti-c, REFERENCES represeniative histograms of SCCVII tumors analv/ed for DNA damage l h following i.p. injection of tirapa/amine. tÃ.the fraction of comets with tail moments ^20 was used to 1. Ostling. O.. and Johanson. K. J. Microclectrophoretic studs of radiation-induced predict survival measured by a colony formation assay 14 days later. Triangles with dins, DNA damages in individual mammalian cells. Biochem. Biophys. Res. Commun.. response of tumor cells removed 2 h follow ing lirapazamine injection. For these tumors IX: 291-298. 1984. only, a discriminator of 10 was used to predict survival. The slope (0.94) and 95% 2. Olive. P. L.. Banath. J. P.. anil Durand. R. E. Heterogeneity in radialion-induced DNA damage and repair in tumor and normal cells measured using the "comet" assay. confidence limits are shown for all of the data, excluding triangles with il/ns. Radial. Res.. /22; 69-72. 1990. .V Fairhairn. D. W.. Olive. P. L.. and O'Neill. K. L. The comet assay: a comprehensive review. MutÃ¢t.Res.. 339: 37-59. 1995. cell lines and two solid tumors for exposure times from 10 min to 4 h. 4. Olive. P L.. Durand. R. E.. Le Riche. J.. Olivotto. !.. and Jackson. S. M. Gel eleclrophoresis of individual cells lo quantify hypoxic fraction in human breast drug concentrations varying by a factor of 4 or more, and subpopu cancers. Cancer Res., 53: 733-736. 1993. lations of cells obtained by cell sorting. This suggests that the same 5. Teman. E. M.. Brown, J. M.. Lemon. M. J.. Hirst. V. K.. and Lee. W. W. SR 4233: discriminator should be effective in measuring the human tumor a new bioreductive agent with high selective toxicity for mammalian cells. Int. J. Radial. Oncol Biol. Phys., 12: 1239-1241. 1986. response to this drug, providing tumor samples are taken 30-60 min ft. Brown, J. M. SR4233 (lirapa/amine): a new anticancer drug exploiting hypoxia in after treatment to avoid problems due to DNA strand break repair. solid tumors. Br. J. Cancer., 67: 1163-1167, 1993. Surprisingly, cells that contained a high number of single-strand 7. Olive. P. L. Detection of hypoxia by measurement of DNA damage in individual cells from spheroids and murine tumors exposed to bioreductive drugs. I. Tirapa/amine. breaks survived this treatment: a tail moment of 20 is equivalent to Br. J. Cancer. 71: 529-536. 1995. about IO4 breaks/cell. This indicates that most DNA strand breaks 8. Olive. P. L. Use of the comet assay to detect hypoxic cells in murine tumors and normal tissues exposed to hioreduclive drugs. Acta Oncol.. 34: 301-305. 1995. produced by tirapa/amine. like those produced by hydrogen peroxide, 9. Olive. P. L. Distribution, oxygÃ©nationand clonogenicity of macrophages in a murine are repaired easily and accurately. Even when cells were selected from tumor. Cancer Commun.. 2: 93-100. 1989. various positions within spheroids or tumors, the same discriminator 10. Durand. R. E. Use of Hoechst 33342 for cell selection of multiceli systems. J. His- tochem. Cytochem.. 30: 117-122, 1982. was effective in determining cell survival. This finding indicates that 11. Chaplin. D. J.. Durand. R. E., and Olive. P. L. Cell selection from a murine tumor the mechanism of cell killing by tirapa/amine is largely independent using the fluorescent probe Hoechst 33342. Br. J. Cancer. 51: 569-572. 1985. 12. Minchinton. A. I.. Lemon. M. J.. Tracy. M.. Pollar!. D.. Martinez. A.. Tosto. L.. and of the many factors that might influence tumor cell response or tumor Brown. J. M. Second generation l.2.4-ben/olriazine-1.4-di-/V-oxide bioreductive cell recovery (e.g.. oxygÃ©nation,cell proliferation status. pH, and anti-tumor acents: pharmacology and activity in vitro and in vivo. Int. J. Radial. ATP). Thus, although incubation under hypoxia greatly increases the Oncol. Biol. Phys.. 22: 701-705. 1992. 13. Walton. M. !.. and Workman. P. Pharmacokinetics and bioreducli\e metabolism of amount of damage produced by tirapa/amine, cells equilibrated at ihe novel benzoiriazine di-A'-oxide hypoxic cell cyloioxin WIN 59075 (SR 4233: different oxygen concentrations must apparently sustain similar NCS 130181) in mice. J. Pharmacol. Exp. Ther.. 265: 938-947. 1993.

4463

Downloaded from cancerres.aacrjournals.org on October 4, 2021. © 1996 American Association for Cancer Research. Use of the Comet Assay to Identify Cells Sensitive to Tirapazamine in Multicell Spheroids and Tumors in Mice

Peggy L. Olive, Charlene M. Vikse and Judit P. Banáth

Cancer Res 1996;56:4460-4463.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/56/19/4460

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/56/19/4460. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.