Detection of DNA Strand Breaks in Individual Apoptotic Cells by the in Situ Terminal Deoxynucleotidyl Transferase and Nick Translation Assays1

DNA strand breaks which occur in HL-60 cells as a result of activation of endonuclease during apoptosis induced by cell treatment with the DNA topoisomerase I inhibitor camptothecin and topoisomerase II inhibitors teniposide, 4'-(9-acridinylamino)-3-methanesulfon-/n-anisidide, and fostriecin were labeled in situ, in individual fixed and permeabilized cells, with biotinylated dUTP (detected by fluoresceinated avidin), using the terminal deoxynucleotidyl transferase or nick translation assays. During the early stage of apoptosis, prior to nuclear fragmentation, the breaks were predominantly localized at the nuclear periphery, close to the nuclear envelope. In more advanced stages, all cellular DNA, then localized within the cell as dense, homogeneous granules of a variety of sizes, was strongly labeled, indicating extensive and more uniform distribution of breaks throughout genomic DNA. Divariate analysis of the incorporated biotiny lated <ll 'IT and cellular DNA content by flow cytometry made it possible to estimate the kinetics of the labeling reaction and relate DNA breaks to cell position in the cycle. The kinetics of biotinylated dUTP incorporation was faster, and the distinction of cells with DNA breaks was more pro nounced, using the terminal transferase rather than the nick translation assay. Camptothecin, teniposide, and 4'-(9-acridinylamino)-3-methanesulfon-m-anisidide induced DNA breaks preferentially in S-phase cells, having little effect on cells in the GI phase of the cycle. In contrast, fostriecin affected cells indiscriminately, in all phases of the cell cycle. The method of detection of DNA strand breaks (3'-hydroxyl termini) in indi vidual cells offers several advantages and can be applied to clinical ma terial (tumor biopsies) to study the induction of apoptosis in tumors during treatment, as a possible prognostic marker. The protein-associated DNA breaks in the "cleavable" DNA-topoisomerase complexes, which are the primary lesions induced by the inhibitors and precede apoptosis, were not detectable by the present methods. at least in some tumors, malignancy may be associated with the loss of the ability of the cell to undergo spontaneous apoptosis rather than with increasing cell proliferation rates (28, 29). There are several methods of detection of apoptosis. Morphological changes involve a characteristic pattern of condensation of chromatin and cytoplasm, which makes it possible to identify these cells by microscopy (1-3). The landmark of apoptosis is endonucleolysis, with nuclear DNA initially degraded at the linker sections to fragments equivalent to single and multiple nucleosomes (3). The latter are revealed as the typical "ladder" on agarose gels during electrophoresis. The flow cytometric methods of identifying apoptotic cells (for a review see Ref. 30) are based on measurement of cellular DNA content (31, 32), increased sensitivity of DNA to denaturation (33), or altered light scatter properties (34). In the present study, we describe and compare two methods which are based on detection of the exten sive DNA breakage which characterizes apoptosis. In these methods the 3'-hydroxyl termini of DNA breaks are labeled with b-dUTP,3 either by exogenous TdT assay or Escherichia coli DNA polymerase (NT assay). Both methods (35) were applied to detect apoptosis of HL-60 cells triggered by DNA topoisomerase I (CAM) and topoi somerase II inhibitors (TN, m-AMSA, and FST). The NT assay was originally proposed by Fehsel et al. (36) as a method to detect apo ptotic fibroblasts in histological specimens and was used by Jonker et al. (37) to identify apoptotic thymocytes following ionizing radiation. Since this paper was submitted for publication, TdT and NT assays were applied in studies on the effects of protease inhibitors on apop tosis induced by topoisomerase inhibitors (38) and used to monitor apoptosis during the treatment of human leukemias (39).