Constitutive autophagy: vital role in clearance of unfavorable proteins in neurons.

Investigations pursued during the last decade on neurodegenerative diseases have revealed a common mechanism underlying the development of such diseases: conformational disorder of certain proteins leads to the formation of misfolded protein oligomers, which subsequently develop into large protein aggregates. These aggregates entangle other denatured proteins and lipids to form disease-specific inclusion bodies. The failure of the ubiquitin-proteasome system to shred the protein aggregates has led investigators to focus their attention to autophagy, a bulk degradative system coupled with lysosomes, which is involved in non-selective shredding of large amounts of cytoplasmic components. Research in this field has demonstrated the accumulation of autophagic vacuoles and intracytoplasmic protein aggregates in patients with various neurodegenerative diseases. Although autophagy fails to degrade large protein aggregates once they are formed in the cytoplasm, drug-induced activation of autophagy is effective in preventing aggregate deposition, indicating that autophagy significantly contributes to the clearance of aggregate-prone proteins. The pivotal role of autophagy in the clearance of aggregate-prone proteins has been confirmed by a deductive approach using a brain-specific autophagy-ablated mouse model. In this review, we discuss the consequences of autophagy deficiency in neurons.

Regulation of FLRG expression in rat primary astroglial cells and injured brain tissue by transforming growth factor-beta 1 (TGF-beta 1).

Follistatin-related gene (FLRG) is a member of the follistatin family of proteins and interacts with transforming growth factor (TGF) superfamily proteins like follistatin. To understand the expression level of FLRG in brain tissue, we examined whether primary neurons and glial cells from rat embryos express FLRG mRNA and produce its protein product. FLRG and follistain mRNAs were mainly expressed in astroglial cells, while activin A mRNA was abundant in primary neurons. TGF-beta1 highly enhanced expression levels of FLRG mRNA in astroglial cells, compared with those of follistatin and activin A mRNAs. Particularly, TGF-beta1 facilitated the secretion of FLRG protein from primary astroglial cells in a dose-dependent manner. Moreover, changes in expression levels of FLRG mRNA and protein in brain tissue were also analyzed after a penetrating injury, using quantitative polymerase chain reactin (PCR) and immunohistochemical methods. Expression levels of FLRG mRNA were significantly increased in damaged regions after penetrating injury together with those of activin A and TGF-beta1 mRNAs. Immunohistochemical observations showed that positive signals of FLRG protein were colocalized in glial fibrillary acidic protein-positive reactive astroglial cells located in damaged regions after a penetrating injury. The expression of follistatin mRNA rather decreased in damage regions after the brain injury. These results suggest that FLRG is synthesized in and secreted from astroglial cells. In particular, FLRG, but not follistatin, may play a role in the regulation of activin A in brain wound healing in response to TGF-beta1.

Different distribution patterns of the two mannose 6-phosphate receptors in rat liver.

Two mannose 6-phosphate receptors, cation-dependent and -independent receptors (CDMPR and CIMPR), play an important role in the intracellular transport of lysosomal enzymes. To investigate functional differences between the two in vivo, their distribution was examined in the rat liver using immunohistochemical techniques. Positive signals corresponding to CIMPR were detected intensely in hepatocytes and weakly in sinusoidal Kupffer cells and interstitial cells in Glisson's capsule. In the liver acinus, hepatocytes in the perivenous region showed a more intense immunoreactivity than those in the periportal region. On the other hand, positive staining of CDMPR was detected at a high level in Kupffer cells, epithelial cells of interlobular bile ducts, and fibroblast-like cells, but the corresponding signal was rather weak in hepatocytes. In situ hybridization analysis also revealed a high level of expression of CIMPR mRNAs in hepatocytes and of CDMPR mRNA in Kupffer cells. By double immunostaining, OX6-positive antigen-presenting cells in Glisson's capsule were co-labeled with the CDMPR signal but were only faintly stained with anti-CIMPR. These different distribution patterns of the two MPRs suggest distinct functional properties of each receptor in liver tissue.

Cathepsin D deficiency induces lysosomal storage with ceroid lipofuscin in mouse CNS neurons.

Cathepsin D-deficient (CD-/-) mice have been shown to manifest seizures and become blind near the terminal stage [approximately postnatal day (P) 26]. We therefore examined the morphological, immunocytochemical, and biochemical features of CNS tissues of these mice. By electron microscopy, autophagosome/autolysosome-like bodies containing part of the cytoplasm, granular osmiophilic deposits, and fingerprint profiles were demonstrated in the neuronal perikarya of CD-/- mouse brains after P20. Autophagosomes and granular osmiophilic deposits were detected in neurons at P0 but were few in number, whereas they increased in the neuronal perikarya within days after birth. Some large-sized neurons having autophagosome/autolysosome-like bodies in the perikarya appeared in the CNS tissues, especially in the thalamic region and the cerebral cortex, at P17. These lysosomal bodies occupied the perikarya of almost all neurons in CD-/- mouse brains obtained from P23 until the terminal stage. Because these neurons exhibited autofluorescence, it was considered that ceroid lipofuscin may accumulate in lysosomal structures of CD-/- neurons. Subunit c of mitochondrial ATP synthase was found to accumulate in the lysosomes of neurons, although the activity of tripeptidyl peptidase-I significantly increased in the brain. Moreover, neurons near the terminal stage were often shrunken and possessed irregular nuclei through which small dense chromatin masses were scattered. These results suggest that the CNS neurons in CD-/- mice show a new form of lysosomal accumulation disease with a phenotype resembling neuronal ceroid lipofuscinosis.

Bi-directional trafficking between the trans-Golgi network and the endosomal/lysosomal system.

Protein transport in the secretory and endocytic pathways of eukaryotic cells is mediated by vesicular transport intermediates. Their formation is a tightly controlled multistep process in which coat components are recruited onto specific membranes, and cargo, as well as targeting molecules, become segregated into nascent vesicles. At the trans-Golgi network, two transport systems deliver cargo molecules to the endosomal system. They can be distinguished with regard to coat components that select cargo molecules. AP-1 assembly proteins mediate transport of MPRs and furin, whereas AP-3 adaptors mediate transport of lysosomal membrane glycoproteins to the endosomal/lysosomal system. The molecular basis for protein-specific sorting lies within sorting signals that are present in the cytoplasmic tails of cargo proteins and allow specific interactions with individual coat components. In order to maintain cellular homeostasis, some proteins are retrieved from endosomal compartments and transported back to the trans-Golgi network. Distinct points for protein retrieval exist within the endosomal system, retrieval occurring from either early or late endosomes. Whereas significant progress has been made in recent years in identifying anterograde and retrograde transport pathways, the molecular mechanisms underlying protein sorting and retrieval are only poorly defined. Recently, however, novel vesicle coats (e.g. AP-4) and proteins that might be involved in sorting (e.g. PACS-1 and TIP47) have been described, and the interactions between assembly proteins and sorting signals are becoming increasingly well defined.

Regulation of a novel pathway for cell death by lysosomal aspartic and cysteine proteinases.

PC12 cells undergo apoptosis when cultured under conditions of serum deprivation. In this situation, the activity of caspase-3-like proteinases was elevated, and the survival rate could be maintained by treatment with acetyl-DEVD-cho, a specific inhibitor of caspase-3. In a culture of PC12 cells treated with acetyl-DEVD-cho, where caspase-3-like proteinases are not activated, CA074, a specific inhibitor of cathepsin B induced active death of the cells. Cathepsin B antisense oligonucleotides showed a similar effect to CA074 on the induction of active cell death. By double staining of terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate-biotin nick end-labeling and activated caspase-3, the dying cells treated with CA074 were positive for terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate-biotin nick end-labeling staining but negative for activated caspase-3. Ultrastructurally, the cells were relatively large and had nuclei with chromatin condensation. The initiation of cell death by CA074 or the cathepsin B antisense were inhibited by the addition of pepstatin A, a lysosomal aspartic proteinase inhibitor, or by cathepsin D antisense. To examine whether this cell death pathway was present in cell types other than PC12 cells, we analysed dorsal root ganglion neurons obtained from rat embryos on the 15th gestational day, a time when they require nerve growth factor for survival and differentiation in culture. When cultured in the absence of nerve growth factor, the neurons survived in the presence of acetyl-DEVD-cho or acetyl-YVAD-cho. Under these conditions, CA074 reduced the survival rate of the neurons, which was subsequently restored by the further addition of pepstain A. These results suggest that a novel pathway for initiating cell death exists which is regulated by lysosomal cathepsins, and in which cathepsin D acts as a death factor. We speculate that this death-inducing activity is normally suppressed by cathepsin B.

The induction of autophagic vacuoles and the unique endocytic compartments, C-shaped multivesicular bodies, in GH4C1 cells after treatment with 17beta-estradiol, insulin and EGF.

The mechanisms for the formation of autophagic vacuoles were investigated using GH4C1 cells, a rat pituitary tumor cell line, whose induction increases intracellular levels of lysosomal proteinases and their mRNA by treatment with a combination of hormones (17beta-estradiol, insulin and EGF). By ordinary electron microscopy, autophagic vacuoles containing various undigested structures with or without limiting membranes were abundant in the hormone-induced cells. These vacuoles, also containing numerous small vesicles, appeared to be derived from multivesicular bodies. In fact, there were also numerous C-shaped multivesicular bodies which enclosed cytoplasmic portions, suggesting that these unique structures are involved in the production of the autophagic vacuoles. Moreover, the cytoplasmic portions enlapped by the C-shaped multivesicular bodies were high in electron density and contained filamentous structures. By the cryothin-section immunogold method, the C-shaped multivesicular bodies in some cases contained lysosomal marker proteins such as cathepsins B and H, and Igp 120. Using an anti-actin monoclonal antibody, immunogold particles clearly labeled the cytoplasmic portions enclosed by the C-shaped multivesicular bodies. Pulse-chase experiments with horse radish peroxidase, a fluid-phase endocytic marker, revealed that the incidence of the C-shaped multivesicular bodies labeled with horse radish peroxidase peaked at 30 min after the beginning of chase incubation, whereas no C-shaped multivesicular body with horse radish peroxidase was detected in the cells by cytochalasin D treatment. These results suggest that the C-shaped multivesicular bodies occur in a transitional process from endosomes to lysosomes by the action of actin filaments, and that this morphological change may be essential for the production of autophagic vacuoles in the hormone-induced GH4C1 cells.

Overexpression of cation-dependent mannose 6-phosphate receptor prevents cell death induced by serum deprivation in PC12 cells.

PC12 cells express well cation-independent mannose 6-phosphate receptors (CI-MPR), but not cation-dependent (CD)-MPR as much. To examine CD-MPR dependency of transport of cathepsins B and D to lysosomes in PC12 cells, we prepared the cells overexpressing CD-MPR. Immunoreactivity for cathepsin B became more distinct and larger in size in the transfected cells than in wild-type cells. No difference in the distribution of cathepsin D was seen between these two cells. The viability of the cells following serum deprivation was significantly higher in the transfected cells than in wild-type cells. This increased viability of the transfected cells was blocked by CA074, a specific inhibitor of cathepsin B, while pepstatin A suppressed the action of CA074. The results suggest that CD-MPR preferentially transport cathepsin B in PC12 cells, and cathepsins B and D participate in the regulation of PC12 cell apoptosis.

Immunocytochemical localization of chromogranin A and secretogranin II in female rat gonadotropes.

Ultrastructures of pituitary gonadotropes are known to show a prominent sex-related difference: typical male rat gonadotropes contain both large- and small-sized granules, whereas typical female rat gonadotropes appear to exhibit uniformly small-sized granules. Our preceding studies have demonstrated that two representative granins, chromogranin A (CgA) and secretogranin II (SgII), are separately localized to each type of granule in male rat gonadotropes. To clarify whether or not there is a certain relationship between granin proteins and characteristic features of secretory granules in female rat gonadotropes, we examined the expression levels and immunocytochemical localizations of CgA and SgII in the cells. Northern blot and immunoblot analyses demonstrated that both CgA and SgII were synthesized and stored in the female pituitary, although the amount of CgA was much lower in the female than that in the male pituitary. Immunocytochemical observations clarified that gonadotropes in the female pituitary possessed intermediate secretory granules containing both CgA and SgII, in addition to solely CgA-positive and SgII-positive ones. However, secretory granules containing CgA in the female gonadotropes were much smaller in size and appeared less frequently than those in the male cells, whereas no sexual difference was discerned in SgII-positive granules. Moreover, the size and appearance of CgA-positive secretory granules varied depending on stages of the estrous cycle. These findings suggest that the size and appearance of secretory granules containing CgA are closely associated with the expression and storage levels of CgA in the pituitary.

Effects of sex steroids on secretory granule formation in gonadotropes of castrated male rats with respect to granin expression.

Pituitary gonadotropes show sex-related differences in their ultrastructure. Typical gonadotropes of male rats exhibit both large granules, which contain chromogranin A (CgA), and small granules, which contain secretogranin II (SgII). In contrast, typical female rat gonadotropes show only a very few large granules among the numerous small granules. To clarify the nature of the biogenesis of these secretory granules and the effects of sex steroids, the ultrastructural and immunocytochemical changes in gonadotropes were examined in castrated male rats supplied with a testosterone or estradiol implant. In castrated rats, pituitary expression and plasma levels of LH increased drastically, but the pituitary content of CgA decreased. The majority of gonadotropes then showed features of "castration cells" containing many small secretory granules. A testosterone implant to castrated rats remarkably suppressed the expression and circulating levels of LH and increased the CgA content in the pituitary to near-normal levels. In this situation, immunocytochemical studies demonstrated that gonadotropes again exhibited large and small secretory granules with the respective localization of CgA and SgII. On the contrary, in castrated rats supplied with an estradiol implant, the expression and content of CgA in the pituitary were remarkably suppressed, and large secretory granules disappeared from gonadotropes. These results suggest that the expression of CgA in gonadotropes is regulated differently by male and female sex steroids. These different effects of androgen and estrogen on the expression level of CgA are closely associated with the sex-related differences in the ultrastructure of secretory granules within gonadotropes.

Prevention of hypoxic liver cell necrosis by in vivo human bcl-2 gene transfection.

Prevention of hypoxic cell death is a key to successful liver transplantation. We developed a new method for preventing liver hypoxic cell death by introducing an anti-cell death gene directly into rat livers. When the human bcl-2 gene (hbcl-2) was directly transfected into rat livers together with non-histone chromosomal protein high mobility group 1 (HMG1) by the hemagglutinating virus of Japan (Sendai virus; HVJ)-liposome method, human Bcl-2 protein (hBcl-2) was efficiently expressed. Electron microscopy and fluorescence microscopy revealed that hepatocytes expressing exogenous hBcl-2 were almost completely protected the hypoxic cell necrosis. The expression of the hBcl-2 also inhibited activation of caspase-3 (-like) proteases and liver dysfunction. Thus, we conclude that transfection of the hbcl-2 gene through HVJ-liposome method is useful to prevent liver cell necrosis induced by hypoxia. This finding could lead to new strategies to avoid the hypoxic cell death, the major problem in liver transplantation.

Immunocytochemical localization of lysosomal cysteine and aspartic proteinases, and ubiquitin in rat epidermis.

To analyze the degradation system in epidermal cells during their generation, differentiation, and cell death, immunocytochemical localization of lysosomal cysteine and aspartic proteinases, an endogenous cysteine proteinase inhibitor, cystatin beta, and ubiquitin were examined using rat sole skin. By confocal laser microscopy, granular immunodeposits for lysosomal proteinases were well demonstrated in epidermal cells; immunoreactivity for cathepsins B and C was prominent in the lower spinous and basal layers, while that for cathepsins L and D was intense in the upper spinous and granular layers, although immunoreactivity for cathepsin D was also detected in the lower epidermal layers. Immunoreactivity for cathepsin H was weakly detected only in the spinous layer, where there were some intensely immunopositive cells with processes which were also immunopositive for S-100 alpha, indicating that these cells were Langerhans cells. Diffuse immunoreactivity for cystatin beta was intense in the spinous and granular layers and weak in the basal layer. In addition, we also examined the localization of ubiquitin, which is a signal peptide for cytosolic proteolysis; clear-cut granular immunodeposits for ubiquitin were detected in spinous and granular cells, and some were co-localized with cathepsin B immunoreactivity. In the basal layer, mitotic cells were strongly immunopositive for ubiquitin. These results suggest that cysteine and aspartic proteinases are involved in the lysosomal system of the epidermis, showing different distributions in the epidermal layers depending on the enzymes examined. Moreover, ubiquitin may be associated with the cell cycle-dependent degradation in basal cells while it also participates in the non-lysosomal proteolysis and probably, lysosomal proteolysis in the spinous and granular cells.

Lysosomal cysteine and aspartic proteinases and ubiquitin in rat and human urinary bladder epithelium.

To examine localization of cysteine and aspartic proteinases, and ubiquitin in rat and human urinary bladders, immunocytochemistry was applied to the tissues. In semi-thin sections, immunoreactivity for cathepsins B and D was densely localized throughout epithelial layers of rats and humans, while that for cathepsins H and L was mainly localized in rat superficial and human intermediate cells. Immunoreactivity for cathepsin C was relatively high in rat and human epithelia, especially in humans. Immunoreactivity for ubiquitin was detected in rat and human epithelial cells. By electron microscopy, vesicular or heterogeneously dense lysosomes labeled with immunogold particles indicating cathepsin B were seen in rat and human epithelial cells; particularly, they often appeared near fusiform vesicles in rat superficial cells and in human intermediate and superficial cells. By double immunostaining, lysosomes with or without vesicular structures were co-labeled with immunogold particles showing both cathepsin B and ubiquitin. The results suggest that cathepsins B, C, H, and L, and cathepsin D are involved in the lysosomal system of rat and human bladder epithelia. Moreover, considering that ubiquitin is a cofactor in the soluble ATP-dependent proteolysis, the results may also indicate that epithelial cells actively form autophagolysosomes.

Bcl-2 blocks loss of mitochondrial membrane potential while ICE inhibitors act at a different step during inhibition of death induced by respiratory chain inhibitors.

Bcl-2, Bcl-xL, CrmA and tetrapeptide ICE inhibitor reduce the extent of necrotic cell death induced by cyanide, which primarily damages mitochondria. Although none of them affects the drastic decrease in ATP levels induced by cyanide, Bcl-2 and Bcl-xL but not CrmA or ICE inhibitor inhibit the cyanide-induced decrease in mitochondrial membrane potential. A similar blocking effect is observed on necrotic cell death induced by other respiration inhibitors, rotenone and antimycin A, and on apoptotic cell death induced by etoposide or calcium ionophore. These results indicate that Bc1-2 and Bcl-xL protect mitochondria against the loss of function during both apoptosis and at least some forms of necrotic cell death. The ICE family proteases act at a different step other than the loss of mitochondrial membrane potential.

Retardation of chemical hypoxia-induced necrotic cell death by Bcl-2 and ICE inhibitors: possible involvement of common mediators in apoptotic and necrotic signal transductions.

Inhibition of the respiratory chain reaction by cyanide, rotenone or antimycin A (chemical hypoxia) induces necrotic cell death characterized by apparently intact chromatin, remarkable mitochondrial swelling with loss of crista structure, and loss of plasma membrane integrity. The treatments induce no apoptotic cell death, as defined by fragmented nuclei with condensed chromatin, fragmented or condensed cytoplasm. The anti-apoptotic proteins Bcl-2 and Bcl-xL effectively retard the chemical hypoxia-induced necrotic cell death. The necrotic cell death is also retarded by inhibitors of ICE(-like) proteases, including interleukin-1beta converting enzyme (ICE), which are common mediators of apoptosis. These results indicate that Bcl-2/Bcl-xL and ICE(-like) proteases modulate apoptotic and at least some forms of necrotic cell death. Both cell death pathways appear to involve some common mediators; however necrotic or apoptotic cell death signals might be transduced through multiple pathways, because Bcl-2/ Bcl-xL or inhibitors of ICE(-like) proteases are relatively less potent in blocking necrotic cell death than in preventing apoptosis.

Cysteine proteinases in GH4C1 cells, a rat pituitary tumor cell line, are secreted by the constitutive and regulated secretory pathways.

Secretory granules of GH4C1 cells, a rat pituitary tumor cell line, are known to be induced by the treatment of estradiol (E2), insulin, and epidermal growth factor (EGF). We examined changes in the localization of cathepsins B, H, and L, lysosomal cysteine proteinases, in GH4C1 cells before and after hormonal treatment. Northern blotting and immunofluorescence microscopy showed that both mRNAs and intracellular protein concentrations of these enzymes were increased in the hormone-induced cells. By immunoelectron microscopy, immunogold particles indicating cathepsins B, H, and L were localized not only in lysosomes but also in some secretory granules. To further examine the molecular forms of these proteinases in secretory granules, radiolabeling and immunoprecipitation methods were applied to the media of the cells incubated with or without secretagogues (100 nM 12-O-tetradecanoylphorbol-13-acetate and 50 microM forskolin); the proforms of cathepsins B, H, and L were secreted from the cells by the constitutive pathway, whereas the mature forms of cathepsins B and H, and the proform and mature form of cathepsin L were secreted by the regulated pathway. These results suggest that in hormone-induced GH4C1 cells, cathepsins B, H, and L are sorted from the Golgi complex not only into lysosomes but also into secretory granules, in which proforms of cathepsins B and H, and a part of procathepsin L are processed into mature forms.

The fate of effete epithelial cells at the villus tips of the human small intestine.

Until recently, little has been known about the morphological features of dying enterocytes at the villus tips of the human small intestine. The present study aimed to show the exfoliating processes of effete enterocytes at the villus tips. Cellular elements of the duodenal lumen and jejunal tissue in humans were fixed and processed for DNA nick end labeling (TUNEL), and transmission and scanning electron microscopy (TEM and SEM). Most cellular elements in the duodenal lumen were enterocytes having TUNEL-positive nuclei. By SEM, protruding enterocytes were discerned at the villus tips. Using the SEM samples embedded in epoxy resin, protruding enterocytes were observed at the villus tips by TEM; they were shrunk by forming numerous clear and autophagic vacuoles, took dome-like profiles, and possessed nuclei with chromatin condensation. The intercellular spaces beneath these protruding or effete enterocytes were often occupied by large lymphocytes. By TUNEL reaction, positive stainings appeared in the epithelium not only at the tip of the villi but also around the site. The results suggest that effete enterocytes at the villus tips of human small intertine are first shrunk by forming clear and autophagic vacuoles, and showed that their nuclei exhibit chromatin condensation immediately before being exfoliated into the lumen.

Delayed neuronal death in the CA1 pyramidal cell layer of the gerbil hippocampus following transient ischemia is apoptosis.

The CA1 pyramidal neurons in the hippocampus are selectively vulnerable to transient ischemic damage. In experimental animals, the CA1 pyramidal neurons undergo cell death several days after brief forebrain ischemia. It remains, however, unknown whether this delayed neuronal death is necrosis or apoptosis. To investigate the degenerating processes of the CA1 pyramidal neurons in gerbil hippocampus after brief ischemia, lysosomal and nuclear alterations in the cells were examined using immunocytochemistry, in situ nick-end labeling, and Southern blotting. By light and electron microscopy, immunoreactivity for cathepsins B, H, and L, representative lysosomal cysteine proteinases, increased in the CA1 pyramidal neurons 3 d after ischemic insult, which showed cell shrinkage. By morphometric analysis, the volume density of cathepsin B-positive lysosomes markedly increased 3 d after ischemic insult, while that of autophagic vacuole-like structures also increased at this stage, suggesting that cathepsin B-immunopositive lysosomes increasing in the neurons after ischemic insult are mostly autolysosomes. Nuclei of the CA1 neurons were nick-end labeled by biotinylated dUTP mediated by terminal deoxytransferase 3 and 4 d after ischemic insult, but not in the prior stages. Simultaneously, dense chromatin masses appeared in nuclei of the neurons. By Southern blotting, laddering of DNA occurred only in CA1 hippocampal tissues obtained 4 d after ischemic insult. Confocal laser scanning microscopy demonstrated that the fragmented DNA in the CA1 pyramidal layer was phagocytosed by microglial cells. The results suggest that delayed death of the CA1 pyramidal neurons after brief ischemia is not necrotic but apoptotic.

Neuronal differentiation of PC12 cells as a result of prevention of cell death by bcl-2.

Rat pheochromocytoma PC12 cells die when cultured in serum-free medium. Neurotrophic factors can rescue PC12 cells from cell death, and induce neuronal differentiation. To further investigate the relationship among cell death, survival, and differentiation, the bcl-2 cDNA, which is known to prevent apoptosis in various types of cells, was transfected into PC12 cells. Six monoclonal bcl-2-transfected cell lines were isolated and confirmed to express mRNA and protein product of bcl-2. The wild-type and bcl-2-transfected PC12 cells were kept to adhere to collagen-coated dishes at the initiation of serum-free experiments to avoid cellular damage due to detachment of the cells by trituration. Even under the conditions, the control PC12 cells mostly died within 24 h, when cultured in serum-free medium, whereas those expressing Bcl-2 survived even for 7 days in serum-free medium. Moreover, outgrowth of long processes in the bcl-2-transfected cells was only observed under the condition to keep the cells attached to the dishes in serum-free medium without any additive neurotrophic or growth factors. Neurofilament medium protein, which is a neuron-specific cytoskeletal component, was also expressed in the differentiated cells, suggesting that the long processes in bcl-2-transfected PC12 cells are neurites. However, neuronal differentiation of PC12 cells expressing Bcl-2 was not observed when cultured in serum-containing medium. Accordingly, survival of PC12 cells expressing Bcl-2 under the condition which cells usually die may be accompanied with neuronal differentiation.