Preapoptotic cell stress response of primary hepatocytes


  • Potential conflict of interest: Nothing to report.


Primary hepatocytes are an important in vitro model for studying metabolism in man. Caspase-9 and Bcl-2-associated X protein (Bax) are regulators of the apoptotic pathway. Here we report on the translocation of procaspase-9 and Bax from cytoplasm to nuclei as well as on dispersion of mitochondria; these processes occur after isolation of primary hepatocytes. The observed changes appear similar to those at the beginning of apoptosis; however, the isolated hepatocytes are not apoptotic for the following reasons: (1) cells have a normal morphology and function; (2) the mitochondria are energized; (3) there is no apoptosis unless it is induced by, e.g., staurosporine or nodularin. Staurosporine does not trigger apoptosis through activation of caspase-9, as its activity is detected later than that of caspase-3. We propose that the translocation of procaspase-9 and Bax into the nuclei reduces the ability to trigger apoptosis through the intrinsic apoptotic pathway. The shifts of procaspase-9 and Bax are reversible in the absence of the apoptotic trigger; the spontaneous reversion was confirmed experimentally for procaspase-9, whereas Bax shifted from the nuclei to the cytosol and mitochondria after the initiation of apoptosis. To distinguish this process from apoptosis, we call it preapoptotic cell stress response. It shares some features with apoptosis; however, it is reversible and apoptosis has to be induced in addition to this process. Conclusion: Knowledge on preapoptotic cell stress response is important for assessing the quality of the cells used in cell therapies, in regenerative medicine, and of those used for modeling metabolic processes. Hepatology 2010;51:2140–2151

Cell cultures, especially those of primary cells, are important models for studying biochemical and physiological phenomena; they are also used in cell therapies and in regenerative medicine. Ideally, the metabolism of isolated cells should not differ from the metabolism within the cells of intact tissues; therefore, the primary cell cultures are thought to be the closest models of in vivo processes. The primary cell cultures of isolated hepatocytes are being used as models to study drug metabolism or drug effects on the liver.1 Nevertheless, some metabolic changes occur upon culturing primary hepatocytes. Expression of cytochromes P450 (CYP) is especially well studied, because of their involvement in drug metabolism. The CYP messenger RNA (mRNA) levels of isolated hepatocytes are similar to those of liver2; however, they decline progressively during the first days in culture.3 Also, there are significant perturbations of genes encoding for antioxidant enzymes, heat shock proteins, nitric oxide synthase, and methionine adenosyltransferase following the isolation and culture of hepatocytes.2 As a consequence, the use of primary cultured hepatocytes in drug metabolism studies is confined to the first days in culture.4 Although the metabolic state of isolated hepatocytes is extensively studied, there are no data on modulation of apoptotic machinery after isolation.

Apoptosis is a mechanism for controlling cell numbers in hepatic tissue.5 It is a mechanism for regression of liver hypertrophy caused by numerous drugs, hormones, and environmental pollutants.6 Apoptosis is regulated by many proteins, among others by caspases, Bcl-2-related proteins (e.g., Bax, Mcl-1, Bcl-xL), and by p53. Caspase-9 is one of the initiator caspases, a part of the intrinsic apoptotic pathway, which is triggered by intracellular stimuli. Its activation is achieved by proteolytic cleavage of its precursor, procaspase-9. The resulting active caspase-9 then triggers the execution phase of apoptosis through the activation of caspase-3; its action accounts for many morphological and physiological features of apoptosis. Procaspase-9 is mainly cytoplasmic in normal cells, although it was localized also to the nuclei of rat brain7 and to the nuclei of some cultured cells.8, 9 In apoptotic cells, the activated caspase-9 was reported to shift to the nuclei of PC-12,10 SHEP neuroblastoma, and HeLa cells.11 Another feature of early apoptosis is that mitochondrial network fragments to a punctiform appearance through the processes of mitochondrial fission.12

p53 is a tumor suppressor important for cell survival and apoptosis. As a transcription factor it induces expression of Bax and represses expression of Bcl-2 and Bcl-xL (reviewed in Ref.13). In addition, it activates apoptosis through interactions with Bcl-2 family members independently of its transcriptional function. The hallmark of the transcription-independent pathways in p53-mediated apoptosis is the stress-induced accumulation of p53 in the cytosol and mitochondria that leads to direct activation of Bak and Bax.13 The presence of p53 in cytosol and mitochondria is necessary but not sufficient to promote apoptosis by transcription-independent pathways.

The potent anticancer activity of p53 has usually been linked to its ability to induce apoptosis through the intrinsic mitochondria-mediated pathway. The crucial event is the mitochondrial outer membrane permeabilization (MOMP) that is controlled by members of the Bcl-2 protein family and requires the activation of Bax or Bak.12

Bcl-2 related proteins have pro- and antiapoptotic functions. Proteins like Bcl-2, BclxL, and Mcl-1 are antiapoptotic. Bax is a proapoptotic protein with an important function in permeabilization of MOMP and an indirect role in mitochondrial fission. It is a monomeric cytosolic protein in healthy cells.24 Upon apoptotic signaling it is activated, which involves conformational changes and exposure of its C- and N-terminal parts and insertion into the mitochondrial outer membrane. The consequence of Bax activation is the formation of pores and MOMP, which is often considered as the point of no return in apoptotic signaling.

In normal cells Bax regulates mitochondrial fusion by regulating the assembly of the mitofusin 2 complex (Mfn 2) at the sites of mitochondrial fusion. In addition to its cytoplasmic location, Bax was also reported to reside in both cytoplasm and in nuclei14, 15 or in nuclei only.14 Translocation of Bax from the cytoplasm to nuclei was reported to occur also as a consequence of hyperthermia14 or induction of apoptosis.16 Another important role of Bax seems to be its translocation to mitochondria upon the initiation of apoptosis followed by an inactivation of mitochondrial elongation through inhibition of Mfn 2.12 Apoptotic fragmentation of mitochondria occurs in the same timeframe as Bax translocates to mitochondria, which leads to MOMP and release of cytochrome c (Cyt-c) across the outer mitochondrial membrane.17

Here we report that procaspase-9 and Bax move from cytoplasm to cell nuclei after the isolation of primary hepatocytes. The shift of procaspase-9 is reversible. In contrast, Bax remains in the nuclei and seems to move out from there only after the induction of apoptosis. The morphology of mitochondria changes after the hepatocytes' isolation as well: from dispersed in freshly isolated cells, to elongate by day 6 from isolation. The isolated hepatocytes are not apoptotic despite the changes in distribution and quantities of some apoptotic proteins, because apoptosis has to be induced by apoptotic inducers like staurosporine (STS) or hepatotoxins like nodularin. However, in the case of triggering apoptosis by STS, caspase-3 is active 3 hours before caspase-9. We propose that the function of nuclear translocation of procaspase-9 and Bax, described here, is to temporarily attenuate the cells' ability to trigger apoptosis through the intrinsic pathway. We describe here, for the first time, a novel mechanism that is activated in response to mild stress from cell isolation. To distinguish it from apoptosis, we call it preapoptotic cell stress response. Preapoptotic cell stress response is important for assessing the quality of cells used for modeling metabolic processes or the ones used in cell therapies and in regenerative medicine.


AI, apoptotic index; Bax, Bcl-2-associated X protein; Cyt-c, cytochrome c; CYP, cytochromes P450; Drp1, dynamin related protein 1; FCCP, carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone; Mfn 2, mitofusin 2 complex; MMP, mitochondrial membrane potential; MOMP, mitochondrial outer membrane permeabilization; STS, staurosporine.

Materials and Methods


All basic chemicals and materials were purchased from Sigma (Taufkirchen, Germany) and Merck (Darmstadt, Germany) if not stated otherwise.

Cell Cultures of Primary Rat Hepatocytes.

Primary hepatocytes were isolated from adult male rats (Wistar-Hannover, 200-300 g) by reverse two-step collagenase perfusion as described by Milisav et al.18 The viability of hepatocytes was 94% ± 1%, as determined by Trypan blue exclusion. Around 105 cells/cm2 were placed on collagen type 1 coated coverslips, incubated for 4 hours to permit adhesion in a humidified atmosphere with 95% air and 5% CO2 at 37°C. The cultures were then washed to remove dead or unattached cells and further incubated for the periods indicated overnight in William's medium E with penicillin and streptomycin (50 U/mL, each), insulin (0.1 U/mL) and 1 μM hydrocortisone hemisuccinate. Each experiment was performed at least three times on the cells from independent isolations. When indicated, 10 μM vinblastine was added to the cells 4 hours after the isolation and incubated for up to 24 hours. One μM STS was added to primary hepatocytes 24 hours after isolation and incubated further for 2-6 hours.

Immunocytochemistry and Immunohistochemistry.

Immunocytochemical and immunohistochemical analyses were performed using standard protocols as described by the suppliers. The following antibodies and dyes were used: anti-caspase-9 (Cell Signaling Technology, Beverly, MA), anti-Bax 6A7 (Sigma, St. Louis, MO), anti-Bax, anti-Bcl-xL (Bcl2L1), anti-Mcl-1, and anti-p53; all by Abcam (Cambridge, UK). They were detected by the appropriate secondary antibody conjugated to the fluorescent dyes AlexaFluor 488 or 546 (Invitrogen, Molecular Probes, Carlsbad, CA). Streptavidin was conjugated with Alexa Flour 546 (Invitrogen, Molecular Probes). The primary antibodies and streptavidin were added sequentially. The coverslips were mounted with Vectashield Hard Set mounting medium with DAPI (Vector Laboratories, Burlingame, CA). Nonspecific labeling by antibodies was tested by staining the cells with fluorescent secondary antibodies only. The cells were visualized using a Leica SP5 confocal microscope (LeicaMicrosystems, Wetzlar, Germany) with an oil immersion objective (×63 magnification and numerical aperture 1.25).

Immunoblotting and Measurements of Mitochondrial Membrane Potential.

One hundred μg of mitochondrial proteins were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto a polyvinylidene fluoride (PVDF) membrane. The same primary antibodies were used as for immunocytochemistry. They were detected by luminescence through the secondary goat antirabbit or goat antimouse antibodies conjugated to horseradish peroxidase (BioRad, Hercules, CA).

Mitochondrial membrane potential (MMP) was estimated by the accumulation of the MMP-dependent fluorescent dye as described by the manufacturer (ApoAlert Mitochondrial Membrane Sensor kit, ClonTech Laboratories, Mountain View, CA). The MMP ratios between the intact cells and carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (FCCP)-treated cells were used to compare the MMP of cells grown at different timepoints.

Detection of Apoptosis by Annexin V.

Cell apoptosis of hepatocytes treated with 0.02 μM nodularin was detected by Annexin V, Alexa Flour-568 (Invitrogen, Molecular Probes) according to manufacturer's instructions. The apoptotic cells were visualized using an Olympus Motorized Inverted Research Microscope IX81 (Tokyo, Japan).

Detection of Caspase Activity.

Primary hepatocytes, 1 day after isolation, were treated with 1 μM STS (Sigma) for 6 hours at 37°C, 5% C02. The controls were treated with an equivalent amount of dimethyl sulfoxide (DMSO). The cells were harvested, counted, and solubilized in cell culture lysis buffer (Promega, Madison, WI). Protein concentrations were determined by BCA Protein Assay Kit (Pierce, ThermoScientific, Rockford, IL). The activities of caspases-3 and -9 were deduced from formation of luminescent substrates by using Caspase-Glo 3/7 Assay and Caspase-Glo 9 Assay, respectively (both from Promega) as described by the supplier. Each sample contained 20 μg of protein.

Statistical Analyses.

The apoptotic index (AI) was calculated as the ratio between the number of apoptotic cells and the number of all cells in the sample. The percentage of relative activity of caspases in a sample was calculated by dividing the luminescence values of treated or untreated cells with the average of luminescence values of untreated cells from each independent experiment. The data from at least tree independent experiments were plotted by Sigma Plot 11.0 (Systat Software, San Jose, CA). Statistical analyses were performed using Statistical Package for the Social Sciences, v. 15.0 (SPSS, Chicago, IL). An unpaired two-tailed Student's t test was used to compare two groups; two-way analysis of variance (ANOVA) and Kruskal-Wallis rank sum test to compare more than two groups (for equal and unequal variances, respectively). When indicated, post hoc analyses were performed by Dunnett T3. We considered values of samples as statistically significant when P < 0.05.


Subcellular Localization of Procaspase-9 Before and After Hepatocyte Isolation.

The location of procaspase-9 changed within a day of preparation of primary hepatocytes. The normal distribution of procaspase-9 was deduced from tissue sections (Fig. 1A). The same results were obtained from liver slices prepared from paraffin-embedded and from snap-frozen tissues; procaspase-9 was only in the cytoplasm. The distribution of procaspase-9 was unchanged in freshly isolated hepatocytes (Fig. 1A,B). Procaspase-9 seemed to be distributed all over the cells 8 hours postisolation (Fig. 1C). After that some procaspase-9 accumulated in the nuclei, whereas some of it remained in the cytoplasm (Fig. 1A,D).

Figure 1.

Subcellular location of procaspase-9 and mitochondria in liver and in isolated hepatocytes. (A) Confocal microscopy images of a paraffin-embedded liver slice and of primary hepatocytes, which were prepared for immunocytochemistry after the indicated times postisolation. Procaspase-9 (Casp-9, green) was detected by polyclonal antibodies, followed by the secondary antibodies conjugated to AlexaFluor-488. Mitochondria (red) were detected by streptavidin conjugated to AlexaFluor-546. Scale bars = 20 μm. (B-D) Images of single cells: immediately after hepatocyte isolation (B), 8 hours after the isolation (C), and 24 hours postisolation (D). Chromatin is stained by DAPI (blue). Scale bars = 5 μm.

The hepatocytes of day 1 did not appear apoptotic, even though some procaspase-9 shifted from cytoplasm to the nuclei. This was tested by the ability of apoptotic inducers to trigger apoptosis.

Apoptosis Is Triggered by STS and Nodularin in Isolated Hepatocytes.

Apoptosis was induced by two different agents: an apoptotic inducer STS and a hepatotoxin nodularin. The activities of caspase-3 and caspase-9 were measured 6 hours after the induction of apoptosis by 1 μM STS on primary hepatocytes' cultured for the following time periods: immediately after isolation (0 hours), 8 hours, 1 day, 3 days, and 6 days (Fig. 2A). Activities of STS-induced caspase-3 differed from mock-treated controls (P = 2 × 10−12, Kruskal-Wallis rank sum test). In contrast, no statistically significant difference was observed in development of luminescent substrate between the cells cultured for different timepoints.

Figure 2.

Induction of apoptosis by STS and nodularin in primary hepatocytes. (A) Activities of caspase-3 and caspase-9 were measured by increased luminescence of their cleaved substrates in parallel from the same samples. Bar charts: Apoptosis was induced by exposing the cells to 1 μM STS for 6 hours at five different timepoints after isolation of primary hepatocytes (right after isolation, 8 hours, 1 day, 3 days, and 6 days). The difference between the activity of caspase-3 in treated cells and controls is statistically significant (P = 2 × 10−12, Kruskal-Wallis rank sum test). The difference between the activity of caspase-9 in treated cells and controls is statistically significant (P = 5 × 10−8, two-way ANOVA). There is no statistically significant difference in either caspase-3 or caspase-9 activity among the STS-treated hepatocytes cultured for different time periods. Line plot: The difference between the activities of caspase-3 and caspase-9 in response to apoptosis induction in hepatocytes 1 day after isolation by 1 μM STS is statistically significant (P = 4 × 10−11, two-way ANOVA); there is no statistical significance between the activity of each caspase at different timepoints. (B) Apoptosis induction by 20 ng/mL of nodularin in hepatocytes cultured for different time periods. Top chart: Apoptotic cells were detected by fluorescently labeled annexin V. The difference between the AIs of treated and untreated cells is statistically significant (P = 1 × 10−12, Kruskal-Wallis rank sum test). There is a statistically significant decrease in AIs between the samples from the first two and the last two timepoints (0 hours to 3 days: P = 0.01; 0 hours to 6 days: P = 0.01; 1 to 3 days: P = 0.02; 1 to 6 days: P = 0.02; Dunnett T3 post hoc test for nonequal variances). Bottom chart: The difference in activity of caspase-3 between the treated cells and controls is statistically significant (P = 0.01, two-way ANOVA). No statistically significant difference was observed between the timepoints.

The activities of caspase-9 (Fig. 2A, second panel) also differed from those of controls (P = 5 × 10−8, two-way ANOVA). The time interval of cell culturing did not result in statistically significant differences in STS-induced activity of caspase-9 when measured 6 hours post-STS addition. This was similar to the activation of caspase-3 presented above. In contrast, the activities of both caspases were strikingly different at shorter timepoints after STS activation, even though they were measured from the same samples (P = 4 × 10−11, two-way ANOVA, Fig. 2A, third panel). Although caspase-9 was hardly activated even after 4 hours, caspase-3 was active even 1 hour after the STS treatment. Therefore, the activation of caspase-3 did not result from the caspase-9 activity in this case. This is in contrast to published observations that STS triggers apoptosis through the mitochondrial pathway.17

The treatment of cells by 20 ng/mL nodularin for 12 hours also resulted in an increase of apoptotic hepatocytes (Fig. 2B, top panel). The AIs of nodularin-treated samples differed statistically significantly from the controls (P = 1 × 10−12, Kruskal-Wallis rank sum test). AIs of treated cells were not equal among the cells cultured for different time periods; a decrease in the number of apoptotic cells was detected at days 3 and 6 after culturing hepatocytes. This was statistically significant between the samples from time 0 and 1 day to days 3 and 6 (0 hours to 3 days: P = 0.01; 0 hours to 6 days: P = 0.01; 1 to 3 days: P = 0.02; 1 to 6 days: P = 0.02; Dunnett T3 post hoc test for nonequal variances). Similar differences were also observed in the activity of caspase-3 in response to nodularin treatment; however, these differences were not statistically significant in hepatocytes incubated for different time periods (Fig. 2B, bottom panel).

Migration of Procaspase-9 into Nuclei Is Reversible and Depends on the Assembly of Microtubules.

We investigated whether the changes in location of procaspase-9 were reversible. By localizing procaspase-9 over several days we found that procaspase-9 was in the nuclei only transiently. It started to shift back from nuclei to cytoplasm 4 days after isolation; then the intense fluorescent nuclear signal of procaspase-9 started to appear patchy (Fig. 3A). Procaspase-9 was in the nuclei of about 50% of the cells 6 days after isolation (50% ± 6%). This decreased further to 39% of the nuclei 7 days postisolation (39% ± 21%). Considerably smaller numbers of cells remained after immunocytochemistry in subsequent days. As this occurred in several repeated experiments, the hepatocytes were not investigated further.

Figure 3.

The migration of procaspase-9 into nuclei is reversible and depends on microtubules. (A) Immunocytochemistry of procaspase-9 (green) and nuclei (blue) within the liver (liver slice) and in isolated hepatocytes cultured for the indicated times. Scale bars = 5 μm (liver slice, 4-5 days), 20 μm (1-3 days and 6 days). (B) Immunocytochemistry of procaspase-9 (green) and nuclei (blue) in normal (− vinblastine) and 10 μM vinblastine-treated (+ vinblastine) primary hepatocytes. Scale bars = 5 μm.

The movement of procaspase-9 into and out of the nuclei appeared to occur in an organized manner. Whether this transport depended on the assembly of microtubules was tested by the microtubule-disrupting agent vinblastine. Ten μM vinblastine was added to the cell culture medium 4 hours after the isolation of hepatocytes (Fig. 3B). Its addition prevented the transport of procaspase-9 into cell nuclei. Therefore, procaspase-9 is transported along the microtubules from the cytoplasm to the nuclei.

Changes of Mitochondrial Morphology.

The mitochondrial morphology changed over time in cultured hepatocytes (Fig. 4A). The same types of changes were observed when the mitochondria were labeled by a MitoTracker, fluorescently labeled Tim23 (an integral mitochondrial inner membrane protein) and by fluorescently labeled streptavidin, which specifically labels mitochondria by binding to biotinylated proteins of mitochondrial matrix.21 Mitochondria were labeled by streptavidin in this study because the strengths of streptavidin fluorescent signals did not vary during the incubation time of primary hepatocytes.

Figure 4.

Mitochondrial morphology and MMP in liver and in isolated hepatocytes. (A) Mitochondria were detected by streptavidin conjugated to AlexaFluor-546. Scale bars = 10 μm. (B) Left panel: MMP is presented as a ratio between the normal cells (untreated) and those treated with FCCP. The increase between the MMP ratios of freshly isolated cells compared to those cultured for 24 hours is statistically significant (P = 3 × 10−7, unpaired two-tailed Student's t test). Right panel: Immunoblot of cell fractionations; Cyt-c was detected from 100 μg of protein loaded per lane. T, total; C, cytoplasm; M, mitochondria.

Mitochondria appeared circular in liver slices and in freshly isolated cells for up to 8 hours. They appeared dispersed after 24 hours postisolation (Fig. 4A), whereas mitochondria formed longer tubules after 3 days in culture. It seemed that mitochondrial fission predominated immediately after the isolation of primary hepatocytes. As in the case of the nuclear shift of procaspase-9, the fission of mitochondria was reversible too.

Despite the changes in mitochondrial morphology, there was neither Cyt-c leakage from dispersed mitochondria nor was there a decrease in MMP (Fig. 4B). The ratio between the potentials of the energized mitochondria and when MMP was dissipated by FCCP was statistically higher at 1 day of hepatocyte culture compared to immediately after isolation (P = 3 × 10−7, unpaired two-tailed Student's t test). The difference between the MMPs of hepatocytes immediately after isolation and those cultured for 1 day is relatively small and could be due to the presence of some cells that were damaged during isolation in the sample that was assayed immediately thereafter.

Localization of Apoptotic Modulators.

The change in cellular location of Bax may be another feature of early apoptosis. We localized Bax to the cytoplasm of hepatocytes in rat liver sections (Fig. 5A). In contrast, only minor amounts of it were cytoplasmic, whereas most of it was in the nuclei of the primary hepatocytes cultured for 24 hours. Bax remained predominantly in the nuclei throughout the culturing of primary hepatocytes; it shifted to cytosol and mitochondria whenever apoptosis was induced by STS. The antibody used for labeling of Bax detected a single band of 22 kDa (Fig. 5B). This proves that Bax was labeled specifically. We conclude that Bax is located predominantly in the nuclei of nonapoptotic cultured primary hepatocytes. None of Bax was activated, unless the apoptosis was triggered by STS. The antibody specific for its active conformation anti-Bax 6A7 recognized Bax only in the cases of STS-treated cells (Fig. 5C). The activated Bax shifted to mitochondria, whereas some of it was still in the cytoplasm and only small amounts were in the nuclei (Fig. 5C).

Figure 5.

Location of Bax in liver and in normal and apoptotic primary hepatocytes. (A) Confocal microscopy images of a paraffin-embedded liver slice and of primary hepatocytes, which were prepared for immunocytochemistry after indicated times postisolation. Bax (red) was detected by monoclonal antibodies to Bax, followed by the secondary antibodies conjugated to AlexaFluor-546. Nuclei (blue) were detected by DAPI. Scale bars = 20 μm. (B) Western blot of total proteins from hepatocytes was probed with the same anti-Bax antibody as above. A single band of 22 kDa corresponding to Bax was detected from 50 μg of proteins loaded per lane. (C) Immunocytochemistry of Bax and mitochondria in normal (− STS) and STS-treated (+ STS) primary hepatocytes. Bax (green) was detected by polyclonal antibody to Bax or monoclonal antibody to activated Bax (Bax 6A7), followed by secondary antibodies conjugated to AlexaFluor-488. Mitochondria (red) were detected by streptavidin-AlexaFluor-546. Arrows point to examples of colocalization signal of Bax and mitochondria (yellow dots). Scale bars = 5 μm.

In contrast to the changed location of caspase-9 and Bax, the positions of Bcl-xL and Mcl-1 appeared unchanged after hepatocyte isolation (Fig. 6). There was an increase in synthesis of both proteins; however, their locations remained unchanged in the cytosol and mitochondria (Fig. 6B). Similarly, p53 remained distributed between the nuclei and the cytosol; the relative amounts of nuclear and cytosolic protein differed among the adjacent cells, from many nuclear p53 to none at all (Fig. 7). The cytoplasmic fraction of p53 appeared somewhat stronger on immunocytochemistries of 1 day compared to the earlier timepoints.

Figure 6.

Subcellular location of Bcl-xL and Mcl-1. (A) Immunocytochemistries of hepatocytes cultured for 1 day. Bcl-xL or Mcl-1 (green), mitochondria (red), DAPI (blue). An arrow points to an example of colocalization signal of Bcl-xL and mitochondria (yellow dots). Scale bars = 5 μm. (B) Immunoblots of cell fractionations from liver and primary hepatocytes cultured for 1 day. T, total; C, cytoplasm; M, mitochondria.

Figure 7.

Subcellular location of p53. Immunocytochemistry of hepatocytes cultured for 8 hours and 1 day. p53 (green), DAPI (blue). Scale bars = 5 μm.

On the basis of the results presented we propose the following model: stressors too mild to trigger apoptosis cause the shifts of procaspase-9 and Bax from cytosol into the nuclei. This sequestration of Bax and caspase-9 is cytoprotective, as it decreases the possibility of triggering apoptosis through the intrinsic apoptotic pathway by these two proteins. In the case of an additional apoptotic signal, apoptosis is initiated possibly through other apoptotic pathways. In the absence of an apoptotic trigger the process reverses to its original state. To distinguish the reversible shifts of procaspase-9 and Bax in nonapoptotic cells from early apoptosis, we named this process preapoptotic cell stress response (Fig. 8).

Figure 8.

The functional relationship between preapoptotic cell stress response and apoptosis.


To our knowledge, this is the first report of changes in intracellular locations of procaspase-9 and Bax and in mitochondrial morphology in primary hepatocytes as a consequence of tissue disruption and isolation. All of the changes described occur within the first 24 hours of isolation: procaspase-9 and Bax move from cytoplasm into nuclei and mitochondria seem to disperse into smaller units. These changes do not occur as a consequence of apoptosis for the following reasons: (1) the cells survive in cultures in seemingly unchanged numbers without replating for at least 6 more days; (2) dispersed mitochondria are fully energized and there is no leakage of Cyt-c; (3) apoptosis can be induced by STS and nodularin; and (4) the changes in location of procaspase-9 and in mitochondrial morphology reverse within 4-6 and 3 days, respectively. As the process observed differs from apoptosis and is triggered by cell isolation, we named it preapoptotic cell stress response.

There are at least two reports on the nonapoptotic cells with nuclear localization of caspase-9. Like in this study, the high levels of caspase-9 were detected in nuclear fractions of brains of normal Wistar rats when the tissue was isolated by perfusion.7 Procaspase-9 was in the nuclei of nontumorigenic mouse mammary epithelial cell line 31 D.8 Caspase-9 was also in the nuclei of some hippocampal neurons after ischemia/reperfusion and of PC-12 cells induced with tamoxifen for 12 hours.10 The nuclear location of caspase-9 in the last report was triggered by apoptosis. Therefore, the caspase-9 relocation has been reported in apoptotic and in normal cells; however, no functional significance of this relocation was described.

Mitochondrial fission may occur at the beginning of apoptosis, as the consequence of inhibition of fusion, because it temporally coincides with mitochondrial clustering of Bax.17 Alternatively, the inhibition of mitochondrial fission through dynamin-related protein 1 (Drp1) blocks apoptosis.22 The mitochondrial fission is reversible and occurs independently of Bax's shift to mitochondria in the case of the primary hepatocytes examined in this study. Namely, Bax is in the nuclei of primary hepatocytes when the mitochondrial fission is observed. It seems, therefore, that the mitochondrial fission occurs independently of the action of Bax on mitochondria in primary hepatocytes.

The shift of Bax into nuclei was observed in nonapoptotic and apoptotic cells. The examples of healthy cultured cells with nuclear Bax are human breast cancer cells MCF-7,23 rat colon carcinoma cells CC531,23 and human tumor cells, like COLO 205 cells, PA-1, U-373 MG15, and various human lung cancer cells.14 Interestingly, out of the 10 lung cancer cell lines examined, six had lower sensitivity to hyperthermia and four had higher. Only cells with high sensitivity to hyperthermia had Bax localized in the nuclei.14 As in the case of primary hepatocytes, Bax may have shifted to the nuclei of hyperthermia-sensitive cells as the result of an unidentified cell stressor. Then the second stimulus (hyperthermia) triggered apoptosis more efficiently.

There are some reports on nuclear localization of Bax in apoptotic cells. Examples are: in dexamethasone-treated or gamma-irradiated mouse thymocytes24; in STS-treated HL-60 promyelocytic leukemia cells24; in etoposide-treated tumor cells with wildtype p53, but not in those without it25; in cisplatin-treated human melanoma cell lines26; and in a human colorectal carcinoma cell line treated with the antibody against epidermal growth factor receptor.27 All of these are malignantly transformed cells, with the exception of mouse thymocytes; therefore, they are likely to be more resistant to apoptosis than normal cells. If so, the apoptotic trigger for the normal cells may induce only preapoptotic cell stress response of the more resilient cells. The nonnuclear distribution of Bax in many apoptotic cells and the redistribution of Bax out of the nuclei upon the induction of apoptosis in primary hepatocytes reported here as well as in the seemingly normal cells with nuclear distribution of Bax support the hypothesis that Bax moves out of the nuclei upon induction of apoptosis.

The shifts in locations of caspase-9 and Bax reported in this article may result in different sensitivities of cells for apoptosis. The activation of caspases-3 and -9 did not differ significantly between the freshly isolated cells and those cultured for 1 day. This does not necessarily mean that the shifts in distribution of caspase-9 and Bax have no role in apoptosis sensitivity. Apoptosis was induced with a relatively high concentration of STS (1 μM). This concentration is often used for apoptosis triggering in different cell types.11, 17, 18 Other STS concentrations are reported in the literature as well.10, 22 The concentration of STS used here was possibly high enough to trigger apoptosis even when Bax was in the nucleus. A comparison of STS dose-response curves in hepatocytes at time 0 and 24 hours postisolation may determine whether the shifts in locations of caspase-9 and Bax are linked to apoptosis sensitivity.

We propose a two-step mechanism that is in agreement with all the data on Bax localization: a mild stressor induces the shift of Bax into the nuclei; it needs a second hit or persistence of an inducer to trigger apoptosis. This agrees also with the observation that apoptosis is triggered through a different pathway when procaspase-9 and Bax are in the nuclei.

The proposed relation between the preapoptotic cell stress response and apoptosis is depicted in Fig. 8. Strong apoptotic triggers induce apoptosis immediately. Cell stressors or weaker apoptotic triggers may induce a preapoptotic cell stress response. The cells subsequently undergo apoptosis in the case of the prolonged stress and of another (or persistent) apoptotic trigger. Otherwise, the cells may recover back to a normal state. Judging from the similarities of responses from so many different cell lines described in the literature, the preapoptotic cell stress response is a general process. It is important to investigate it further because discovering the mechanisms of preapoptotic cell stress response may lead to a novel way to presensitize tumor cells so that apoptosis can be triggered efficiently by the second hit. Knowledge of the preapoptotic cell stress response is important also for being able to assess the well-being of cells, especially of primary hepatocytes, which are used to model biochemical processes within liver; the same is needed for the cells used in cell therapies and in regenerative medicine.


We thank Prof. Nina Zidar for assistance with tissue sections of liver and Andrej Vovk and Rok Blagus for advice with statistical analyses.