To date, apoptosis has been characterized biochemically by the production of 180-200 bp internucleosomal DNA fragments resulting from the activation of an endonuclease(s). The principal morphological feature of apoptosis is the condensation of chromatin and it has been assumed that this may reflect the oligonucleosomal fragmentation pattern. We have re-examined this dogma by comparing the biochemical and morphological features of cell death in several epithelial cell types (HT-29-I1 colon adenocarcinoma, CC164 mink lung, DU-145 human prostatic carcinoma and MCF-7 human breast adenocarcinoma) and one mesenchymal cell line (Hllras-R3 ras-transformed rat fibroblasts). Cell death was induced either by serum deprivation, TGF-,B1 or etoposide, or by leaving cells to reach confluence. Cell death was assessed with respect to detachment from monolayers, morphological changes and DNA integrity. The DNA-binding fluorophore Hoechst 33258 revealed chromatin condensation patterns consistent with apoptotic cell death in all cell types except MCF-7 cells. Using field inversion gel electrophoresis in conjunction wih conventional 2% agarose gel electrophoresis, cleavage of DNA to 50 kbp fragments was observed in all cases except MCF-7 cells. This preceded the appearance of oligonucleosomal fragments in HT-29-I1, CC164 and Hllras-R3 cells. Although the DNA of DU-145 cells fragmented into 50 kbp units, and although the cells exhibited classical apoptotic morphology, no subsequent internucleosomal cleavage was observed. These results suggest that changes in the integrity of DNA indicative of the release of chromatin loop domains occur before cleavage at internucleosomal sites is initiated and that the latter is not an essential step in the apoptotic process.

In previous studies hepatocytes undergoing cell death by apoptosis but not normal hepatocytes in rat liver showed immunostaining for transforming growth factor ,1 (TGF-13). Staining was much stronger with antibodies recognizing the pro-region of TGF-I31 than the mature peptide itself. Therefore we investigated the ability of both forms of TGF-31 to induce apoptosis in primary cultures of rat hepatocytes.Mature TGF-fi1 induced rounding up of the cells and fragmentation into multiple vesicles. As revealed by the DNAspecific stain H33258, the chromatin of these cells condensed and segregated into masses at the nuclear membrane; this was obviously followed by fragmentation of the nucleus. Ultrastructurally the cytoplasm was well preserved, as demonstrated by the presence of intact cell organelles. These features strongly suggest the occurrence of apoptosis. Quantification of nuclei with condensed chromatin revealed that mature TGF-131 was 30-fold more effective than the TGF-P1 latency-associated protein complex. Finally, we administered TGF-131 in vivo using an experimental model in which regression of rat liver was initiated by a short preceding treatment with the hepatomitogen cyproterone acetate. Two doses of TGF-P1, each 1 nM/kg, augmented the incidence of apoptotic hepatocytes 5-fold. Equimolar doses of TGF-fi1 latency-associated protein complex were ineffective. These studies suggest that TGF-P1 is involved in the initiation of apoptosis in the liver and that the mature form of TGF-fi1 is the active principle.

No abstract

Abstract. Chromatin condensation paralleled by DNA fragmentation is one of the most important criteria which are used to identify apoptotic cells. However, comparable changes are also observed in interphase nuclei which have been treated with cell extracts from mitotic cells. In this respect it is known that in mitosis, the lamina structure is broken down as a result of lamin solubilization and it is possible that a similar process is happening in apoptotic cells. The experiments described in this study have used confluent cultures of an embryonic fibroblast cell line which can be induced to undergo either apoptosis at low serum conditions or mitosis.Solubilization of lamin A+B was analyzed by immunoblotting and indirect immunofluorescence. These studies showed that in mitotic cells lamina breakdown is accompanied by lamin solubilization. In apoptotic cells, a small amount of lamin is solubilized before the onset of apoptosis, thereafter, chromatin condensation is accompanied by degradation of lamin A+B to a 46-kD fragment. Analysis of cellular lysates by probing blots with anti-PSTAIR followed by anti-phosphotyrosine showed that in contrast to mitosis, dephosphorylation on tyrosine residues did not occur in apoptotic cells. At all timepoints after the onset of apoptosis there was no significant increase in the activation of p34 ~2 as determined in the histone H1 kinase assay. Coinduction of apoptosis and mitosis after release of cells from aphidicolin block showed that apoptosis could be induced in parallel with S-phase.The sudden breakdown of chromatin structure may be the result of detachment of the chromatin loops from their anchorage at the nuclear matrix, as bands of 50 kbp and corresponding multimers were detectable by field inversion gel electrophoresis (FIGE). In apoptotic cells all of the DNA was fragmented, but only 14% of the DNA was smaller than 50 kbp. DNA strand breaks were detected at the periphery of the condensed chromatin by in situ tailing (ISTAIL). Chromatin condensation during apoptosis appears to be due to a rapid proteolysis of nuclear matrix proteins which does not involve the p34 ~c2 kinase.OPTOSIS waS originally defined as a program of morphological changes accompanying cell death during which the cells round up and the chromatin condenses, leading to the formation of crescent-shaped masses aggregating at the membrane. In parallel the nucleolus dissolves (Kerr, 1971). These initial changes are followed by cytoplasmic and nuclear compaction, and ultimately by fragmentation of the cell. The nuclear fragments formed in this process are still surrounded by the nuclear membranes (Kerr, 1971;Oberhammer et al., 1993a). As a common biochemical marker for apoptosis, an activation of a nonlysosomal, Ca2+/MgZ÷-dependent endonuclease has been suggested (Wyllie, 1980;Cohen and Duke, 1984). The response is so readily identifiable that endonuclease activation

No abstract