necrosis

Chemico-Biological Interactions 139 (2002) 79 – 95

www.elsevier.com/locate/chembiont

Acrolein-induced cell death: a caspase-influenced decision between apoptosis and oncosis/necrosis Julie C. Kern *, James P. Kehrer Di6ision of Pharmacology and Toxicology, College of Pharmacy, The Uni6ersity of Texas at Austin, Austin, TX 78712 -1074, USA Received 1 October 2001; received in revised form 16 October 2001; accepted 3 November 2001

Abstract Due to the dominating roles that caspases play in the apoptotic cascade, their activities appear to be a primary factor in the death pathway (apoptosis versus oncosis/necrosis) decision. In murine FL5.12 proB lymphocytes, the cellular consequences of acrolein treatment included a lack of typical apoptotic features in preference to oncosis/necrosis. Oncosis/ necrosis was apparent by detection of a reduction in intracellular ATP concentration, increased plasma membrane leakage (measured by LDH release and flow cytometric detection of propidium iodide uptake) and morphological criteria. Analysis of acrolein-treated cell lysates or recombinant caspase enzymes showed overall dose-dependent decreases in caspase3, -8 and -9 activities. In addition to acrolein’s effect on intracellular caspases, it was also able to alter caspase-dependent apoptosis induced by secondary treatment with etoposide or following cytokine withdrawal. Acrolein at doses ] 20 mM circumvented etoposide or interleukin-3 withdrawal induced apoptosis. When acrolein was combined with mechlorethamine, another alkylating agent not dependent on caspases for its cell death signaling, necrosis was increased in a dose-dependent manner. Overall, these data suggest that caspase inhibition plays an important role in the cell death pathway decision, particularly with treatments dependent on the caspase cascade to induce apoptosis. © 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Acrolein; Apoptosis; Oncosis; Necrosis; Caspase

Abbre6iations: AO/EB, acridine orange/ethidium bromide; ATP, adenosine triphosphate; HN2, mechlorethamine; IL-3, interleukin-3; LDH, lactate dehydrogenase; PARP, poly (ADP-ribose) polymerase; PI, propidium iodide; PM, phosphoramide mustard; VP-16, etoposide. * Corresponding author. Tel.: +1-512-471-5188; fax: +1-512-471-5002. E-mail addresses: [email protected] (J.C. Kern), [email protected] (J.P. Kehrer). 0009-2797/02/$ - see front matter © 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 0 0 9 - 2 7 9 7 ( 0 1 ) 0 0 2 9 5 - 2

80

J.C. Kern, J.P. Kehrer / Chemico-Biological Interactions 139 (2002) 79–95

1. Introduction Caspases are cysteine aspartate proteases which normally exist as inactive proenzymes comprised of an amino terminal pro-domain, and large and small catalytic subunits [1]. Caspases are named for their ability to cleave caspase-specific tetrapeptide sequences at aspartate residues. Cleavage of caspases generates at least two active fragments by removing the inactivating pro-domain. These active fragments associate as a heterotetramer and, in some cases, are responsible for initiating cleavage of other caspases involved in apoptotic signaling. Caspase-3 is recognized as an apoptotic effector predominantly activated by caspases-8 and -9, which are classified as initiators [2]. Caspase-8 generally initiates apoptosis as a result of stimulation by death receptors on the plasma membrane. Through a compilation of death domain signaling, caspase-8 is activated followed by caspase-3, leading to the execution of apoptosis. Caspase-9 is predominantly activated via mitochondrial signaling. In an ATP-dependent process, the mitochondrion releases cytochrome c which complexes with Apaf-1 and cleaves procaspase-9 to activate caspase-9. Caspase-3 activation and the subsequent execution of apoptosis follow the activation of caspase-9. The result of an inhibition in caspase activity seems to be very cell line specific and involves either the maintenance of viability or a switch to death exhibiting apoptosis-independent characteristics, i.e. necrosis. For example, the addition of tumor necrosis factor and/or stimulation of its related ligands induce apoptosis via caspase activation in a number of cell lines. Pretreatment with a peptide caspase inhibitor allows necrotic morphology to replace apoptosis in IEC-6 intestinal epithelial, murine L929 fibrosarcoma, SW480 colon adenocarcinoma and H460 non-small cell lung cancer cells [3–5]. In contrast, the same pretreatment inhibited both death pathways and maintained viability in HCT116 human colon cancer and 293 human embryonic kidney cells [5]. In murine B lymphocytes, dexamethasoneinduced apoptosis was prevented by the caspase inhibitors zAsp.cmk or zVAD-fmk, although cells deviated to necrosis [6,7]. Biochemically, the a,b unsaturated aldehyde acrolein is well known for its strong electrophilic interactions, particularly with thiol-containing molecules, such as glutathione (GSH) [8]. Exposure to acrolein is unavoidable due to its presence in smoke (including cigarette smoke), automobile exhaust, overheated cooking oil and herbicides, as well as being a metabolic product of cyclophosphamide [8–11]. Its ubiquitous environmental presence indicates the need for understanding its mechanism of toxicity. At high doses, acrolein clearly causes massive disruption to cell structure and function. However, at lower doses, our understanding of the critical pathways that are affected is limited. In this study, interleukin-3 (IL-3) withdrawal, etoposide (VP-16) and mechlorethamine (HN2) were used as secondary treatments after acrolein in FL5.12 pro-B cells to investigate acrolein-mediated alterations in the cell death pathway, potentially through changes in caspase activity. Previous work in FL5.12 cells has established that withdrawal of IL-3 results in activation of caspase-3 and subsequent apoptosis [12]. VP-16 is a topoisomerase II inhibitor widely used in

J.C. Kern, J.P. Kehrer / Chemico-Biological Interactions 139 (2002) 79–95

81

chemotherapy to stimulate apoptosis via DNA damage, stress-related and caspase signaling pathways [13]. In Jurkat T lymphocytes or its reconstituted organelles, VP-16 stimulates apoptosis signaling pathways via caspase-dependent and -independent mechanisms [14,15]. Phosphoramide mustard (PM) and acrolein are two active metabolites of cyclophosphamide, a widely used chemotherapeutic agent [16]. While cyclophosphamide’s antineoplastic effects are associated with PM, acrolein is linked with its toxic side effects. However, interactions between these species are possible and HN2 is an alkylating nitrogen mustard that was used as a model for PM. The data presented here demonstrate the ability of acrolein to inhibit the activities of caspase-3, -8 and -9 in cell lysate or with recombinant enzyme, while characterizing the resultant cell death as primarily oncosis/necrosis. This study also demonstrates that acrolein may vary cell death pathways when coupled with secondary treatments. Overall, acrolein’s inhibition of caspases may provide a better understanding of how apoptosis can be evaded, leading to the induction of oncosis/necrosis in lymphocytes.

2. Materials and methods

2.1. Materials RPMI-1640, EBSS, Trypan blue exclusion dye and penicillin/streptomycin were obtained from Gibco BRL (Grand Island, NY). Fetal bovine serum (FBS) was obtained from Summit Biotechnology (Fort Collins, CO). Acrolein (90% pure, 10% dimers and water), L-glutamine and other molecular biology supplies were obtained from Sigma (St. Louis, MO). Fluorogenic caspase substrates and human caspase recombinant enzymes were obtained from Alexis Biochemicals (San Diego, CA). Caspase antibodies were obtained from Stressgen Biotechnologies Corp. (Victoria, BC). Propidium iodide (PI) was obtained from Roche Molecular Biochemicals (Indianapolis, IN). Protein concentrations were measured using the BioRad DC protein assay (BioRad Laboratories, Hercules, CA).

2.2. Cell culture The IL-3 dependent FL5.12 wild type, murine, pro B lymphoid progenitor cells were incubated at 37 °C/5% CO2 and cultured in complete media consisting of RPMI-1640 media plus 10% FBS (v/v), 10% WEHI-3B conditioned media, 100 U/ml penicillin and 100 mg/ml streptomycin. WEHI-3B cells were grown under similar conditions (without IL-3 supplementation) to produce IL-3 containing medium, as described previously [17]. Cells were passaged every other day and refreshed with complete media. Cell counting was carried out with a hemacytometer and viability was assessed using the Trypan blue exclusion assay to ensure \ 95% viable cells prior to treatment. Cells were plated 106/ml in sterile 6-, 12- or 24-well dishes (Falcon, Corning or Costar). During treatments, cells were incubated in EBSS supplemented with 0.6 mg/ml L-glutamine and replenished with complete

82

J.C. Kern, J.P. Kehrer / Chemico-Biological Interactions 139 (2002) 79–95

RPMI media after 0.5 h. Glutamine supplementation maintained cell viability while cells were without serum, while EBSS use prevented the binding of acrolein or HN2 to serum components.

2.3. Caspase acti6ity assay Caspase activity was assayed using relatively specific fluorogenic substrates, as described by Stennicke and Salvesen [18]. Cells (106) were centrifuged at 200×g for 10 min at 4 °C and resuspended in 100 ml lysis buffer (50 mM Tris (pH 7.5), 150 mM NaCl, 0.5 mM EDTA, 0.5% (v/v) Nonidet P-40). Lysate (50 ml) were added to 140 ml reaction buffer (10 mM HEPES (pH 7.5), 50 mM NaCl, 2.5 mM DTT) and 10 ml 800 mM Ac-DEVD-amc (caspase-3), Ac-IEHD-amc (caspase-8) or Ac-LEHDamc (caspase-9) in a black walled 96-well plate and incubated in the dark at 37 °C for 2 h. Excess lysate was subject to a protein assay. Fluorescence of the cleaved aminomethylcoumarin (amc) was measured in a microplate reader at ex. 360 nm and em. 450 nm. Caspase activity was normalized to protein concentrations and calculated as a percentage of control.

2.4. Recombinant caspase acti6ity The quantitation of recombinant caspase activity was based on the methods used in cell lysate with modifications from Alexis Biochemicals. In a 96-well, blackwalled plate, 0, 2.5, 5, 10, 20 or 40 mM acrolein in PBS was directly added to 50 ng recombinant caspase-3 or -8, or 1 U caspase-9 and incubated for 15 min at 37 °C. Some 185 ml recombinant reaction buffer (20 mM HEPES, 100 mM sodium chloride, 10 mM DTT, 1 mM EDTA, 0.1% (w/v) Nonidet P-40, 10% sucrose, pH 7.2) and 5 ml 800 mM caspase-3, -8, or -9 fluorogenic substrate (as stated above) were added to the acrolein/enzyme solution and incubated for 1 h at 37 °C. Fluorescence was measured as described above and compared to the 0 mM acrolein-treated sample.

2.5. SDS-PAGE and Western blot Protein (40 mg) was separated on a 15% SDS-PAGE gel, followed by a semi-dry transfer to an Immobilon-P 0.45 mm PVDF membrane. The membrane was incubated overnight at 4 °C in 5% milk in TBS-T followed by 1 h incubation with 1:500 – 2000 caspase rabbit polyclonal primary antibody in TBS-T with 2.5% milk and 1:3000 donkey anti-rabbit HRP secondary antibody in TBS-T with 2.5% milk. The membrane was incubated in enhanced chemiluminescence solution and exposed to Hyperfilm ECL (Amersham).

2.6. Propidium iodide PI uptake by cells was measured using methodology in a similar manner to the Annexin V-FITC/PI assay without Annexin present. Cells (106) were centrifuged at

J.C. Kern, J.P. Kehrer / Chemico-Biological Interactions 139 (2002) 79–95

83

200× g for 10 min at 4 °C and rinsed once with PBS. Cells were resuspended in 2.5 mg PI/ml PBS. Fluorescence in 20,000 cells per sample was measured using an argon laser at ex. 488 and em. 620 (FL 3, log scale) in a Coulter EPICS XL MCL flow cytometer. Cells stained positively for PI were quantitated as a percent of total cells.

2.7. ATP ATP was quantitated using the HPLC methodology described by Sellevold et al. [19]. The stationary phase contained a 7 mm guard column (Alltech) and a 5 mm, 25 cm long C18 Dynamax column (Varian). The mobile phase consisted of 0.2 mm filtered and degassed 215 mM KH2PO4, 2.3 mM tetrabutylammonium hydrogen sulfate and 3.5% acetonitrile at pH 6.25 (adjusted with KOH). The system was rinsed daily with 3.5% acetonitrile to remove salt build-up on the column. Cells (2×106) were incubated in 200 ml perchloric acid with shaking for 10 min, neutralized with 100 ml KOH and centrifuged at 16,000× g for 4 min. The supernatant was injected on the HPLC for detection of nucleotides at 254 nm with a retention time of 7–8 min for ATP. ATP concentration was quantitated according to peak area integrations of pure ATP samples and normalized to protein concentrations.

2.8. LDH release Cells (106) were centrifuged at 200×g for 10 min at 4 °C. The cell pellet was resuspended in PBS with 0.1% Triton in an equivalent volume to the collected media. Supernatant or cell lysate (80 ml) were combined with 920 ml reaction mixture (2 ml 1 M Tris – HCl, pH 7.4; 50 ml 0.1 M sodium pyruvate, freshly added 1.75 mg b-NADH and diluted to 10 ml total volume with H2O). A DA340/min was measured for 5 min and averaged. Samples were normalized to protein concentration and the percentage of released LDH was quantitated by dividing LDH present in the media by total cellular LDH (media plus lysate).

2.9. Acridine orange/ethidium bromide for flow cytometry Cells (106) were centrifuged at 200× g for 10 min at 4 °C. Cells were resuspended in 1 ml PBS with 2 ml each 100 mg/ml acridine orange and ethidium bromide diluted in PBS. Fluorescence in 20,000 cells per sample was measured using an argon laser at ex. 488 and log scale emissions 525 (FL1) and 620 (FL 3) in a Coulter EPICS XL MCL flow cytometer. Measurement and differentiation of cell populations was based on published descriptions and diagrams [20].

2.10. Cell morphology Cells (106) were centrifuged at 200× g for 10 min at 4 °C and all but 50 ml media was removed. The pellet was resuspended with 2 ml each 100 mg/ml acridine orange and ethidium bromide (which is diluted in PBS). Cells were immediately viewed

84

J.C. Kern, J.P. Kehrer / Chemico-Biological Interactions 139 (2002) 79–95

using a Nikon Eclipse E800 microscope, PlanFluor 20X objective and filter cube B-2A that incorporates ex. 450–490 nm, dichroic mirror at 500 nm and barrier filter at 515 nm to separate the green from orange-red fluorescence. Digital images were captured using Metamorph 4.1 software (Universal Imaging Corp., Downingtown, PA). Morphology was defined according to descriptions from Duke and Cohen [21] for general apoptotic versus oncotic/necrotic characteristics, e.g. shrunken versus swollen cells.

2.11. Statistics Multiple group comparisons were performed using an analysis of variance followed by Tukey– Kramer multiple comparison statistics to determine a mean9 S.E. and a p value (GB-STAT 6.5 program). Samples with a pB 0.05 were considered significantly different from control values.

3. Results As an electrophile, acrolein reacts with numerous cellular nucleophiles. Thus, its toxicity is highly dependent on dose per cell rather than concentration [22]. While doses have been expressed in terms of concentration for clarity, cell density was maintained at a constant level so that the dose per cell varied proportionally to concentration. For example, under our conditions, 2.5 mM acrolein translates to 2.5 fmol of acrolein per cell. Caspases-8 and -9 are well-known activators of caspase-3 via the receptor-mediated death domains and mitochondrial release of cytochrome c, respectively. Fig. 1 depicts the activities of caspases-3, -8 and -9 in cell lysates after cells recovered for 12 or 24 h from an initial 0.5 h treatment with various concentrations of acrolein. At 12 h, caspase-3 activity was non-significantly decreased : 40% following 2.5 and 5 mM acrolein and significantly decreased by 80% following 20 and 40 mM acrolein. The 10 mM dose showed a 2-fold increase in activity that was, however, not statistically significant. At 24 h after 2.5–10 mM acrolein treatments, caspase-3 activity recovered to levels that were statistically indistinguishable from control. Cells treated with 20 and 40 mM acrolein maintained low caspase-3 activity, possibly because of low ATP levels (a requirement for apoptotic processes) and low viability at these doses (Fig. 2). Acrolein decreased the activities of caspases-8 and -9 in a more consistent dose-dependent manner, declining to B 10% of control with the 40 mM dose at 12 h. Statistical significance was evident with ] 5 mM acrolein at 12 h. At 24 h, caspase-8 and -9 activities remained depressed at 10 and 20 mM, but recovered to control levels at doses B 10 mM acrolein. It is noteworthy that acrolein treatments never significantly increased caspase activity and thus these decreases were from the low basal levels of activity, suggesting apoptosis was not occurring. A lack of caspase-3, -8 and -9 activation was confirmed by SDS-PAGE and Western blots, which revealed no changes in pro-caspase proteins or appearance of cleavage products (data not shown).

J.C. Kern, J.P. Kehrer / Chemico-Biological Interactions 139 (2002) 79–95

85

Fig. 1. The effect of acrolein on caspase activity in cell lysate. Caspase activity was measured in cell lysates at 12 h (black bars) and 24 h (gray bars) after acrolein treatment (0.5 h). Cells (106) were lysed and incubated with Ac-DEVD-amc, Ac-IETD-amc or Ac-LEHD-amc substrate that fluoresces when cleaved by active caspase-3, -8, or -9, respectively. Treated cells are compared to 100% control. Data are expressed as the mean of three individual experiments 9 S.E. *Significantly different from 0 mM acrolein (p B 0.05).

86

J.C. Kern, J.P. Kehrer / Chemico-Biological Interactions 139 (2002) 79–95

ATP is considered important for the execution of apoptosis [23] and its absence may predispose cells to undergo oncosis/necrosis [24]. At 24 h after acrolein, ATP intracellular concentrations were decreased in a dose-dependent manner, reaching 0% of control following 40 mM acrolein (Fig. 2). Non-significant decreases in ATP concentration by 35% with 55 mM acrolein did not increase oncosis/necrosis compared to untreated cells, while at concentrations of acrolein at or above 10 mM the reduction of ATP ( \45%) correlated with an increase in oncosis/necrosis (see below). The relative amounts of apoptosis versus oncosis/necrosis are often measured using flow cytometric quantitation of Annexin V-FITC binding to externalized phosphatidylserine (PS), which is characteristic of early apoptosis, coupled with the uptake of propidium iodide showing the loss of plasma membrane integrity in oncotic/necrotic cells. Initial studies with acrolein using this assay indicated the induction of only minor amounts of early apoptosis (data not shown). However, the potential reactivity of acrolein with components on the plasma membrane [10], as well as the development of leaky plasma membranes typical of oncotic/necrotic cells, suggested Annexin has the potential to be blocked and/or to interact with PS on the interior of a permeable plasma membrane, thus becoming an unreliable indicator of apoptosis. This possibility was recently confirmed by a study showing that malondialdehyde can modify PS in red blood cells [25]. Thus, Annexin V should not be used to identify apoptosis induced by reactive aldehydes. Therefore, alternative assays were used to assess apoptosis.

Fig. 2. Acrolein’s effect on plasma membrane integrity and ATP at 24 h. At 24 h recovery, oncosis/necrosis was quantitated as the percentage of 20,000 cells positively stained with PI, measured by flow cytometry (bars). ATP (line) was measured by UV-HPLC detection of nucleotides and expressed as a percentage of control. Control ATP levels were 722 pmol/mg protein with a retention time of 7 – 8 min. Data are expressed as the mean of three individual experiments 9S.E. *Significantly different from 0 mM acrolein (pB0.05).

J.C. Kern, J.P. Kehrer / Chemico-Biological Interactions 139 (2002) 79–95

87

Fig. 3. Acrolein’s effect on VP-16 induced apoptosis at 24 h. One million cells/ml were treated for 0.5 h with EBSS or acrolein followed by 24 h in media 9 10 mM VP-16. Cells were collected and resuspended in 1 ml PBS and 2 ml each 100 mg/ml ethidium bromide and acridine orange. Flow cytometric readings were measured immediately and 20,000 cells per treatment were counted. The populations were differentiated, using log scales, according to their staining capabilities as previously described: live cells (I), apoptotic cells (J) and necrotic cells (K).

Initially, an ELISA was used to measure mono- and oligonucleosomal fragmentation, which is characteristic of apoptosis [26,27]. This assay did not reveal any fragmentation, consistent with the absence of apoptosis (data not shown). Acridine orange and ethidium bromide staining of cells was then measured by flow cytometry to more clearly differentiate live, apoptotic and necrotic cells [20]. An example of the flow cytometry output is shown in Fig. 3. Live cells (oval I) stained brightly for the cell permeable acridine orange (green), with low cell impermeant ethidium bromide influx (orange-red). Apoptotic cells (oval J) stained with a lower concentration of acridine orange and low ethidium bromide. Apoptotic cells are distinguished from live cells mainly due to apoptosis-induced DNA fragmentation, but may also reflect cell shrinkage or formation of apoptotic bodies, all of which will reduce cellular fluorescence. Necrotic cells (oval K) stained both low and high with acridine orange and high for ethidium bromide, indicative of deteriorating cells with permeable plasma membranes. As shown in Fig. 3, the viability of control cells is very high, with significant apoptosis evident after treatment with VP-16. The results show extensive necrosis, but only minimal apoptosis with acrolein alone,

88

J.C. Kern, J.P. Kehrer / Chemico-Biological Interactions 139 (2002) 79–95

while the apoptosis induced by VP-16 is largely converted to oncosis/necrosis when treated after acrolein. Since noteworthy apoptosis was not evident in cells treated with acrolein, oncosis/necrosis was quantitated by flow cytometry with propidium iodide staining and LDH leakage. At 12 h, positive propidium iodide staining was statistically significant compared to untreated cells only with the 40 mM acrolein-treated cells (data not shown). However, by 24 h there was a significant loss of plasma membrane integrity with concentrations of acrolein = 10 mM (Fig. 2). Propidium iodide positive cells accounted for 43, 48 and 77% of total cells after 10, 20 and 40 mM acrolein, respectively. The LDH release data (Table 1) were consistent with the PI findings. Statistical significance was present 24 h after treatment with 20 and 40 mM acrolein, showing 30 and 53% LDH release into the supernatant. The differences in the percentages obtained by these two methods can be accounted for by differences in the nature of the assays themselves. In confirmation of oncosis/necrosis shown by PI and LDH, a qualitative assessment of cell death and morphology was completed using fluorescent microscopic analyses of acridine orange and ethidium bromide staining (Fig. 4). Live cells stain only with acridine orange, which show up green with a bright yellow-green circular organelle inside the cell. Apoptotic cells shrink and induce fragmentation of the yellow organelle, while oncotic cells are characterized by their cellular swelling. Necrotic cells stain for both acridine orange and the cell impermeant ethidium bromide, and is detected by their orange color. The images in Fig. 4 show that control and 2.5 mM acrolein induce no significant observable changes in cellular staining or shape at 24 h. With doses of 10 mM or more, there were increasing amounts of swollen and distorted cells coupled with organelle disintegration and a rise in necrotic cells. Acrolein’s high reactivity suggested it could affect caspase activity in intact cells via direct and indirect mechanism(s). To assess its direct effect, the activities of human recombinant caspase-3, -8 and -9 enzymes were determined after 15 min Table 1 Acrolein’s effect on cellular LDH leakage Acrolein (mM)

% LDH release into media

0 2.5 5 10 20 40

12.5 92 13.5 92 11.1 92 23.8 93 30.1 94* 52.7 92*

LDH release into the supernatant is a measure of plasma membrane integrity and is calculated spectrophotometrically by the conversion of NADH to NAD+ at 340 nm (DAbs340) over 5 min. LDH was measured in 106 cells 24 h after 0.5 h acrolein treatment. The average DAbs/min was normalized to protein concentrations and the percent LDH released was calculated by dividing LDH in media/total LDH (media plus lysate). Data are expressed as a mean of four individual experiments 9 S.E. *Significantly different from 0 mM acrolein (PB0.05).

J.C. Kern, J.P. Kehrer / Chemico-Biological Interactions 139 (2002) 79–95

89

Fig. 4. Acrolein’s effect on cellular morphology and death at 24 h. One million cells/ml were incubated for 0.5 h with EBSS or acrolein followed by 24 h in complete media. Cells were pelleted and resuspended in 2 ml each 100 mg/ml acridine orange and ethidium bromide. Cell death and morphology were immediately assessed using fluorescence microscopy. Cell populations were differentiated according to death characteristics and staining differences as described in Section 2.

incubation with 2.5–40 mM acrolein. Through fluorometric quantitation of specific substrate cleavage, acrolein was shown to decrease the activity of all recombinant caspases at concentrations of 10 mM or more, except with caspase-3, which required at least 20 mM to reach statistical significance (Fig. 5).

90

J.C. Kern, J.P. Kehrer / Chemico-Biological Interactions 139 (2002) 79–95

When the effect of acrolein on the toxicity of secondary treatments was examined using the acridine orange/ethidium bromide flow cytometry assay, it was evident that acrolein alone induced no significant apoptosis, while acrolein (\ 10 mM) combined with HN2 led to increased oncosis/necrosis (Table 2). Surprisingly, when 20 mM acrolein exposure occurred prior to IL-3 withdrawal or VP-16 treatment, there was substantially less apoptotic cell death and no increase in oncosis/necrosis. In accordance with Fig. 3, 40 mM acrolein9 VP-16 or IL-3 withdrawal led to substantial oncosis/necrosis, presumably due to the toxicity of this dose alone.

4. Discussion The end state of cellular toxicity falls into two main cell death pathways, apoptosis and oncosis [28,29], both of which progress into irreversible necrosis. Apoptosis is characterized by cell shrinkage, PS externalization on the plasma membrane, DNA cleavage into 180 base pair fragments, maintenance of ATP levels and a number of other distinct features. Oncosis was named for, and is best described by, its overall cell swelling, thus the designation ‘onco’. Cell or organelle swelling and marked reduction of ATP levels dissociate oncosis from necrosis, although the distinction between the two forms of cell death can be difficult. Thus, they have been described together. Caspase inhibition is receiving increasing attention due to its possible application in modulating apoptosis. While in some systems this approach seems to increase cell survival [5], this is not necessarily true in all cases, as increased death and a switch to oncosis/necrosis has also been observed [3–7]. Acrolein was recently shown to

Fig. 5. The direct effect of acrolein on recombinant caspase enzyme activity. A total of 50 ng of recombinant caspase-3, -8, or 1 U caspase-9 enzyme was incubated directly with increasing concentrations of acrolein. The enzyme/acrolein mixture was incubated with reaction buffer and caspase-specific substrate. Fluorescence was detected with a microplate reader. Data are expressed as the mean of three individual experiments 9 S.E. *Significantly different from 0 mM acrolein (p B 0.05).

J.C. Kern, J.P. Kehrer / Chemico-Biological Interactions 139 (2002) 79–95

91

Table 2 Acrolein’s effect on secondary treatments: IL-3 withdrawal, VP-16 or HN2 Acrolein alone

Acrolein plus

Acrolein (mM)

EBSS/acrolein

-IL3

10 mM VP-16

250 nM HN2

(A) Li6e cells 0 10 20 40

979 1 749 11 839 1 5994*

26 9 3* 13 9 2* 75 94† 47 9 6*

359 4* 47 92* 86 91† 62 92*,†

87 9 4 57 93* 65 9 9 17 9 4*,†

(B) Apoptotic cells 0 10 20 40

290.3 229 9 1191 139 2

67 9 2* 82 9 5* 18 9 4† 16 93†

56 9 2* 44 9 3* 10 91† 13 9 1†

7 92 34 92*,† 19 9 3 11 9 4

(C) Necrotic cells 0 10 20 40

0.59 0.1 59 2 791 3193*

14 9 1 11 9 5 11 91 42 94*,†

18 92* 20 91* 7 9 1† 29 92*,†

15 94 22 92 23 9 7* 76 9 5*,†

* Significantly different from 0 mM acrolein (pB0.05). † Significantly different from -IL-3, VP-16 or HN2 treatment alone (pB0.05). One million cells were treated for 0.5 h with EBSS or acrolein. Secondary treatments included 0.5 h with 250 nM HN2 in EBSS followed by 24 h recovery in media, 24 h with IL-3 withdrawn media or 24 h with 10 mM VP-16 in media. Cells were collected and resuspended in 1 ml PBS and 2 ml each 100 mg/ml ethidium bromide and acridine orange. Flow cytometric readings were measured immediately and 20,000 cells per treatment were counted. The populations were differentiated as a percent of total cells according to their staining capabilities. Live cells (A) only incorporated the cell permeable acridine orange, fluorescing brightly green. Apoptotic cells (B) fluoresced a dimmer green without loss of plasma membrane integrity, while necrotic cells (C) took up both dyes and fluoresced orange-red. Results are expressed as a mean of three individual experiments 9 S.E.

reduce caspase-3 activity in neutrophils at similar doses, although cell death displayed mixed apoptotic and necrotic characteristics [30]. The current data demonstrate that acrolein alone induces only very modest levels of apoptosis at low doses (B 10 mM) while causing almost exclusively oncosis/necrosis at higher doses. Given that acrolein is highly electrophilic and reacts rapidly with thiol groups and that caspases have an essential thiol group, it was hypothesized that acrolein may inhibit this family of enzymes, thereby modulating the form of cell death that can occur, particularly in conjunction with secondary toxins. The current studies demonstrated the ability of acrolein to inhibit caspase activities both in intact cells and with active recombinant enzymes. The inhibition was clearly dose-dependent with the recombinant enzymes, but somewhat less so in intact cells, probably as a consequence of attempting to measure inhibition below the already low basal levels and an ill-defined balance between activation and inactivation and the complex nature of how acrolein interacts with cellular components.

92

J.C. Kern, J.P. Kehrer / Chemico-Biological Interactions 139 (2002) 79–95

Concentrations of \ 10 mM acrolein inhibited apoptosis induced by both IL-3 withdrawal and VP-16. Contrasting this, \ 10 mM acrolein enhanced the necrotic effect of HN2. This difference is probably a consequence of the differing modes of action of the secondary treatments. Specifically, since both acrolein and HN2 are electrophiles, an additive effect on toxicity occurred as expected. The protective effects seen against apoptosis induced by IL-3 withdrawal and VP-16 likely arose partially from the inhibition of caspase activity by acrolein, although replacement of acrolein with the general caspase inhibitor boc-asp.fmk had no effect on cytokine withdrawal or VP-16-induced apoptosis in this cell line (unpublished data). Other studies in FL5.12 and Jurkat cells support the findings that caspase inhibitors only partially block with IL-3 withdrawal and VP-16 invoked apoptosis-inducing cell signals [12,14,15]. Previous studies have noted that non-lethal doses of acrolein appear to exert some more subtle molecular effects, either directly or perhaps secondary to its effects on GSH [9,22]. In particular, acrolein is able to diminish the activation of redox sensitive transcription factors, specifically NF-kB [31] and AP-1 (unpublished data), within 0.5 h of treatment. Since these transcription factors can affect apoptosis [9,32], it is also possible that they play a role in the form of cell death induced in cells exposed to acrolein. It is evident with the recombinant caspase/acrolein interaction that the conserved nucleophilic amino acids found at the catalytic site in caspases provide targets for acrolein. If acrolein is able to reach these amino acids in intact cells, it could bind, thereby preventing activation, or cause the inactivation of activated caspases. This, in turn, would inhibit a primary apoptotic response and allow the secondary oncotic/necrotic pathway to dominate, consistent with the findings in the present study. Oncosis/necrosis appears to be the dominant death pathway for acrolein-treated cells. Only modest levels of apoptosis could be identified using a range of endpoints and various concentrations of acrolein. Examining endpoints more associated with oncosis/necrosis made it apparent that this is the primary response of cells to acrolein. The finding that ATP, which is required for apoptosis processes to terminate the cell in an ordered manner, is diminished in these cells further supported this conclusion. In addition, plasma membrane integrity, as shown by propidium iodide and ethidium bromide staining and LDH leakage, was compromised when ATP was depleted by : 50%. Interestingly, the form of cell death induced by other a,b-unsaturated aldehydes in lymphocytes appear to involve substantially more apoptosis [33,34]. The reason for this discrepancy is not known, but may be related to the greater reactivity of acrolein leading to more caspase inactivation and a more profound loss of ATP. ATP loss undoubtedly contributes to acrolein-mediated oncosis/necrosis. Since the use of the general caspase inhibitor boc-asp.fmk was unable to replicate acrolein’s effect on secondary treatments, ATP loss may play a stronger role in this form of cell death than expected. One possible explanation for ATP depletion may include poly (ADP-ribose) polymerase (PARP) over-activation [35]. PARP over-activation can lead to ATP depletion through its use in regenerating PARP’s substrate

J.C. Kern, J.P. Kehrer / Chemico-Biological Interactions 139 (2002) 79–95

93

b-nicotinamide adenine dinucleotide. In a PARP − / − fibroblast model, the absence of PARP prevented ATP depletion and the induction of necrosis following treatment with N-methyl-N%-nitro-N-nitrosoguanidine, supporting this concept. The clinical implications of understanding acrolein’s molecular signaling are directly applicable to a number of disease states, including cancer, due to its presence as a metabolite of cyclophosphamide and in inflammation, where as a product of lipid peroxidation, the formation of acrolein-lysine adducts has been correlated with plaque deposits seen in atherosclerosis and Alzheimer’s patients [10]. Acrolein exposures may also result from a variety of industrial processes that use this aldehyde, such as the production of acrylate polymers and its presence in herbicides and as an active component of slimicides [8]. Furthermore, cigarette smokers may be exposed to at least 40 ppm acrolein, contributing to severe lung damage. In summary, the current results have shown that acrolein alone causes a dominant oncotic/necrotic response. This may be due to the failure of this electrophile to activate apoptosis pathways or, alternately, by its ability to inhibit such pathways. The data demonstrate that the latter possibility is more likely, since acrolein is able to diminish caspase activities needed for apoptosis induction. When acrolein treatment is combined with HN2 (an electrophilic nitrogen mustard similar to a metabolite found in cyclophosphamide treatment), an increase in oncosis/necrosis occurs principally due to an additive effect in toxicity. In contrast, acrolein had a protective effect against the apoptosis induced by VP-16 or IL-3 withdrawal, likely due to its inhibition of caspases. Although protective in this instance, the general caspase inhibitor boc-asp.fmk was unable to duplicate the protective effects, demonstrating that acrolein is utilizing more complex signaling pathways.

Acknowledgements This work was supported by Grant ES09791 and NIEHS Center Grant ES07784. J.P.K. is the Gustavus and Louise Pfeiffer Professor of Toxicology.

References [1] G.M. Cohen, Caspases: the executioners of apoptosis, Biochem. J. 326 (1997) 1 – 16. [2] N.A. Thornberry, Y. Lazebnik, Caspases: enemies within, Science 281 (1998) 1312 – 1316. [3] F.M. Ruemmele, S. Dionne, E. Levy, E.G. Seidman, TNF-a-induced IEC-6 cell apoptosis requires activation of ICE caspases whereas complete inhibition of the caspase cascade leads to necrotic cell death, Biochem. Biophys. Res. Comm. 260 (1999) 159 – 166. [4] D. Vercammen, R. Beyaert, G. Denecker, V. Goossens, G. Van Loo, W. Declercq, J. Grooten, W. Fiers, P. Vandenabeele, Inhibition of caspases increases the sensitivity of L929 cells to necrosis mediated by tumor necrosis factor, J. Exp. Med. 187 (1998) 1477 – 1485. [5] N. Ozoren, K. Kim, T.F. Burns, D.T. Dicker, A.D. Moscioni, W.S. El-Deiry, The caspase 9 inhibitor Z-LEHD-FMK protects human liver cells while permitting death of cancer cells exposed to tumor necrosis factor-related apoptosis-inducing ligand, Cancer Res. 60 (2000) 6259 – 6265.

94

J.C. Kern, J.P. Kehrer / Chemico-Biological Interactions 139 (2002) 79–95

[6] C. Lemaire, K. Andreau, C. Sidoti-de Fraisse, A. Adam, V. Souvannavong, IL-4 inhibits apoptosis and prevents mitochondrial damage without inducing the switch to necrosis observed with caspase inhibitors, Cell Death Diff. 6 (1999) 813 – 820. [7] C. Lemaire, K. Andreau, V. Souvannavong, A. Adam, Inhibition of caspase activity induces a switch from apoptosis to necrosis, FEBS Lett. 425 (1998) 266 – 270. [8] R.O. Beauchamp, D.A. Andjelkovich, A.D. Kligerman, K.T. Morgan, H.A. Heck, A critical review of the literature on acrolein toxicity, CRC Crit. Rev. Toxicol. 14 (1985) 309 – 380. [9] J.P. Kehrer, S.S. Biswal, The molecular effects of acrolein, Toxicol. Sci. 57 (2000) 6 – 15. [10] K. Uchida, Current status of acrolein as a lipid peroxidation product, Trends Cardiovasc. Med. 9 (1999) 109 –113. [11] P.K. Rudra, H.E. Krokan, Acrolein cytotoxicity and glutathione depletion in n3 fatty acid sensitive- and resistant human tumor cells, Anticancer Res. 19 (1999) 461 – 470. [12] H.K. Bojes, X. Feng, J.P. Kehrer, G.M. Cohen, Apoptosis in hematopoietic cells (FL5.12) caused by interleukin-3 withdrawal: relationship to caspase activity and the loss of glutathione, Cell Death Diff. 6 (1999) 61 –70. [13] S.H. Kaufmann, Cell death induced by topoisomerase-targeted drugs: more questions than answers, Biochim. Biophys. Acta 1400 (1 – 3) (1998) 195 – 211. [14] J.D. Roberston, V. Gogvadze, B. Zhivotovsky, S. Orrenius, Distinct pathways for stimulation of cytochrome c release by etoposide, J. Biol. Chem. 275 (42) (2000) 32438 – 32443. [15] X. Sun, M. MacFarlane, J. Zhuang, B.B. Wolf, D.R. Green, G.M. Cohen, Distinct caspase cascades are initiated in receptor-mediated and chemical-induced apoptosis, J. Biol. Chem. 274 (8) (1999) 5053 –5060. [16] S.M. Ludeman, The chemistry of the metabolites of cyclophosphamide, Curr. Pharm. Des. 5 (1999) 627–643. [17] J.C. Lee, A.J. Hapel, J.N. Ihle, Constitutive production of a unique lymphokine (IL-3) by the WEHI-3 cell line, J. Immunol. 128 (1982) 2393 – 2398. [18] H.R. Stennicke, G.S. Salvesen, Biochemical characteristics of caspases-3, -6, -7, and -8, J. Biol. Chem. 272 (1997) 25719 –25723. [19] O.F.M. Sellevold, P. Jynge, K. Aarstad, High performance liquid chromatography: a rapid isocratic method for determination of creatine compounds and adenine nucleotides in myocardial tissue, J. Mol. Cell. Cardiol. 18 (1986) 517 –527. [20] T.J. Liegler, W. Hyun, T.S. Benedict Yen, D.P. Stites, Detection and quantification of live, apoptotic, and necrotic human peripheral lymphocytes by single-laser flow cytometry, Clin. Diag. Lab. Immunol. 2 (1995) 369 –376. [21] R.C. Duke, J.J. Cohen, Morphological and biochemical assays of apoptosis, in: J.E. Coligan, A.M. Kruisbeak (Eds.), Current Protocols in Immunology, Wiley, New York, 1992, pp. 3.17.1 – 3.17.16. [22] N.D. Horton, B.M. Mamiya, J.P. Kehrer, Relationships between cell density, glutathione and proliferation of A549 human lung adenocarcinoma cells treated with acrolein, Toxicology 122 (1997) 111 –122. [23] P. Nicotera, M. Leist, E. Ferrando-May, Intracellular ATP, a switch in the decision between apoptosis and necrosis, Toxicol. Lett. 102 – 103 (1998) 139 – 142. [24] J.A. Walisser, R.L. Thies, Poly (ADP-ribose) polymerase inhibition in oxidant-stressed endothelial cells prevents oncosis and permits caspase activation and apoptosis, Exp. Cell Res. 251 (1999) 401– 413. [25] K. Balasubramanian, E.M. Bevers, G.M. Willems, A.J. Schroit, Binding of Annexin V to membrane products of lipid peroxidation, Biochemistry 40 (2001) 8672 – 8676. [26] J.D. Robertson, K. Datta, J.P. Kehrer, Bcl-xL overexpression restricts heat-induced apoptosis and influences hsp70, bcl-2 and bax protein levels in FL5.12 cells, Biochem. Biophys. Res. Comm. 241 (1997) 164 –168. [27] K. Datta, S.S. Biswal, J.P. Kehrer, The 5-lipoxygenase-activating protein (FLAP) inhibitor, MK886, induces apoptosis independently of FLAP, Biochem. J. 340 (1999) 371 – 375. [28] C. Majno, I. Joris, Apoptosis, oncosis, and necrosis, Am. J. Pathol. 146 (1) (1995) 3 – 15. [29] S. Levin, T.J. Bucci, S.M. Cohen, A.S. Fix, J.F. Hardisty, E.K. LeGrand, R.R. Maronpot, B.F. Trump, The nomenclature on cell death: recommendations of an ad hoc committee of the Society of Toxicological Pathologists, Toxicol. Pathol. 27 (1999) 484 – 490.

J.C. Kern, J.P. Kehrer / Chemico-Biological Interactions 139 (2002) 79–95

95

[30] E.I. Finkelstein, M. Nardini, A. Van Der Vliet, Inhibition of neutrophil apoptosis by acrolein: a mechanism of tobacco-related lung disease?, Am. J. Physiol. Lung Cell Mol. Physiol. 281 (2001) L732–L739. [31] N.D. Horton, S.S. Biswal, L.L. Corrigan, J.C. Bratta, J.P. Kehrer, Acrolein causes inhibitor kB decreases in nuclear factor-kB (NF-kB) activation in human lung adenocarcinoma (A549) cells, J. Biol. Chem. 274 (1999) 9200 –9206. [32] R. Ravi, G.C. Bedi, L.W. Engstrom, Q. Zeng, B. Mookerjee, C. Gelinas, E.J. Fuchs, A. Bedi, Regulation of death receptor expression and TRAIL/Apo2L-induced apoptosis by NF-kB, Nat. Cell Biol. 3 (2001) 409 –416. [33] J.F. Kalinich, R. Ramakrishnan, D.E. McClain, N. Ramakrishnan, 4-Hydroxynonenal, an endproduct of lipid peroxidation, induces apoptosis in human leukemic T- and B-cell lines, Free Rad. Res. 33 (2000) 349 –358. [34] W. Zhang, Q. He, L.L. Chan, F. Zhou, M.E. Naghy, E.B. Thompson, N.H. Ansari, Involvement of caspases in 4-hydroxy-alkenal-induced apoptosis in human leukemic cells, Free Rad. Biol. Med. 30 (2001) 699 –706. [35] H.C. Ha, S.H. Snyder, Poly (ADP-ribose) polymerase is a mediator of necrotic cell death by ATP depletion, Proc. Natl. Acad. Sci. USA 96 (1999) 13978 – 13982.