3,4-Dideoxyglucosone-3-ene Induces Apoptosis in Renal Tubular Epithelial Cells

MCT cells are a cultured line of proximal tubular epithelial cells harvested originally from the renal cortex of SJL mice (22). The cells were maintained in culture in RPMI 1640 (Gibco, Grand Island, NY), 10% heat-inactivated FCS, 2 mmol/l glutamine, 100 units/ml penicillin, and 100 μg/ml streptomycin at 37°C in 5% CO₂ (23). For cell death experiments, cells were plated and then grown for 24 h in RPMI with 10% FCS. Then, the media was replaced with fresh serum-free RPMI (RPMI with 0% FCS), unless otherwise stated. RPMI with 0% FCS contained the same supplements as the serum-supplemented RPMI, with the exception of the FCS.

3,4-DGE was generously provided by Anders Weislander (Gambro, Lundia, Sweden) and was directly dissolved in cell culture medium (24). Endotoxin levels were below the detection limit as assessed by a limulus amebocyte assay (Sigma).

The following caspase inhibitory peptides were used: the caspase-3 inhibitor Z-Asp(OMe)-Glu(OMe)-Val-dl-Asp(OMe)-fluoromethylketone (DEVD-fmk) and the pancaspase inhibitor Z-Val-Ala-dl-Asp-fluoromethylketone (zVAD-fmk) from Bachem (Bubendorf, Switzerland); and the caspase-8 inhibitor Z-Ile-Glu(OMe)-Thr-Asp(OMe)-fluoromethylketone (IETD-fmk), the caspase-9 inhibitor Z-Leu-Glu(OMe)-His-Asp(OMe)-fluoromethylketone (LEHD-fmk), and the caspase-2 inhibitor benzyloxycarbonyl-Val-Asp-Val-Ala-Asp-fluoromethylketone (z-VDVAD-fmk) from Calbiochem (San Diego, CA). The caspase inhibitory peptides were used at concentrations previously shown to protect from apoptosis in MCT cells and other cell systems (23). DEVD-fmk, zVAD-fmk, IETD-fmk, LEHD-fmk, and z-VDVAD-fmk were dissolved in DMSO. Final concentration of DMSO was 0.05%, and it did not influence MCT cell apoptosis. Peptides (200 μmol/l) or vehicle were added to the cell cultures 3 h before addition of 3,4-DGE. Cells were cultured in the presence of stimuli for an additional 24–72 h.

Bax inhibitor peptide P5 and a negative control peptide were from Tocris (Ellisville, MO). P5 is a cell-permeable synthetic peptide inhibitor of Bax translocation to mitochondria. The peptide was designed from Ku-70, a protein that suppresses the mitochondrial translocation of Bax, and it inhibits Bax-mediated apoptosis (25). Cells were treated with 50 μmol/l 3,4-DGE for 72 h in the presence or absence of Bax inhibitor peptide P5 (200 μmol/l) or negative control (200 μmol/l).

For quantification of cell death, 10,000 cells were seeded in 12-well plates (Becton Dickinson, Franklin Lakes, NJ) under the experimental conditions. Apoptosis was quantified by flow cytometry assessment of DNA content in pooled attached and detached cells, permeabilized, and stained with propidium iodide (23). The percentage of cells with decreased DNA staining (A₀), comprising apoptotic cells with fragmented nuclei, was counted. To assess the typical nuclear changes seen in apoptosis, cells were fixed and stained with propidium iodide (23). After fixation, propidium iodide stains both live and dead cells.

Western blot was carried put according to previously described techniques (23,26). Primary antibodies were rabbit polyclonal anti–caspase-9 (1:1,000; Cell Signaling, Hertfordshire, U.K.), rabbit polyclonal anti–cleaved caspase-3 (1:1,000; Cell Signaling), or rabbit polyclonal anti-Bid (1:3,000; gift of S.J. Korsmeyer, Howard Hughes Medical Institute). Blots were then probed with anti-tubulin antibody, and levels of expression were corrected for minor differences in loading (23,26).

Cells were plated onto Labtek slides in RPMI with 10% FCS. After 24 h the media was changed to RPMI with 0% FCS, and then cells were incubated for 24 h with 3,4-DGE. Staining with rabbit polyclonal anti-Bax antibody or rabbit polyclonal anti–cytochrome c antibody (1:100; Santa Cruz Biotechnology) was carried out as previously described (23). Cell nuclei were counterstained with propidium iodide.

Caspase-3 activity was measured after the manufacturer’s instructions (Biomol, Plymouth, PA) (23). The assay is based on the cleavage of a DEVD-pNA peptide, yielding colored free p-nitroaniline.

We used a phosphorothionate-modified antisense oligodeoxynucleotide 5′-TGCTCCCCGGACCCGTCCAT-3′ directed against the translation initiation region of mouse Bax mRNA and a control scrambled oligodeoxynucleotide 5′-TCGTCCGGCCACCGCTCACT-3′, which has little complementarity with Bax mRNA but the same composition as the antisense oligodeoxynucleotide (Metabion, Martinsried, Germany) (23,27). Cells were treated with 50 μmol/l 3,4-DGE for 72 h in the presence or absence of Bax antisense oligodeoxynucleotide (20 μg/ml) or scramble oligodeoxynucleotide (20 μg/ml (27).

Results are expressed as the means ± SD. Significance at the P < 0.05 level was assessed by the nonparametric Mann-Whitney test for two groups of data and Kruskal-Wallis for three of more groups by means of the SigmaStat statistical software (Jandel, San Rafael, CA).

3,4-DGE induces renal tubular epithelial cell death in a dose- and time-dependent manner (Fig. 1). Morphologic studies in cells permeabilized and stained with propidium iodide demonstrated the presence of shrunk, bright, pyknotic nuclei characteristic of apoptosis among cells exposed to 3,4-DGE (Fig. 2A). Flow cytometry confirmed an increase in the number of hypodiploid apoptotic cells (Fig. 2B). 3,4-DGE–induced apoptosis was time and dose dependent (Tables 1 and 2). An increased rate of apoptosis was already evident at 48 h and peaked at 72 h (Table 1).

Next, we studied the molecular mechanisms of apoptosis induced by 3,4-DGE. In the mitochondrial pathway for apoptosis, Bax oligomerizes at mitochondria and promotes the release of caspase activators, such as cytochrome c (6). We observed that 3,4-DGE induced Bax oligomerization that preceded the nuclear apoptotic morphology (Fig. 3A). In addition, cytochrome c was released from mitochondria (Fig. 3B).

Release of cytochrome c from mitochondria promotes activation of caspase-9 (6). Caspase-9 is eventually cleaved and promotes activation by cleavage of other caspases, such as caspase-3. 3,4-DGE resulted in the appearance of cleavage fragments from caspases-3 and -9, as well as a cleavage fragment of the caspase-8 target Bid (Fig. 4A). Caspase-3 is the central executioner caspase. Functional evidence of caspase-3 activation was also observed (sixfold increase in caspase-3 activity) (not shown). A broad-spectrum caspase inhibitor (zVAD-fmk) prevented both caspase-3 activation (not shown) and the development of features of apoptosis (Fig. 4B, inset). However, zVAD-fmk did not prevent eventual cell death (Fig. 4B). Specific inhibitors of caspases -2, -3, -8, and -9 also decreased apoptosis induced by 3,4-DGE (Table 3), suggesting the involvement of multiple caspases in a positive feedback loop mode.

We then explored the role of Bax in cell death induced by 3,4-DGE in tubular epithelium. We used two independent approaches to inhibit or antagonize Bax. First, the level of expression of Bax was decreased by Bax antisense oligonucleotides (23). Western blot confirmed that antisense oligonucleotides decrease Bax protein levels (not shown) (23). Decreased Bax expression protected from both apoptosis and cell death (Fig. 5). Then, Bax function was antagonized by a Ku-70–derived Bax antagonistic peptide (25). This intervention also resulted in decreased apoptosis at 72 h (3,4-DGE: 23.7 ± 4.4%; 3,4-DGE/Ku70 peptide: 9.7 ± 2.5%, P < 0.05 vs. 3,4-DGE; 3,4-DGE/control peptide: 20.7 ± 2.9%) and cell death (not shown).

We have shown that the glucose degradation product 3,4-DGE promotes apoptosis in cultured tubular epithelial cells through the activation of a Bax-mediated pathway. 3,4-DGE had initially been isolated from a brown seaweed with immunosuppressive activity in vivo (28). 3,4-DGE had marked suppressive effects on T-cell proliferation, antibody production from B cells, and interleukin-1 production from macrophages (28). 3,4-DGE was also the main cytotoxic compound from glucose-containing peritoneal dialysis fluids (17). Induction of leukocyte apoptosis may underlie both the cytotoxicity observed in the peritoneal dialysis system and the immunosuppressive properties (24).

The current study expands the range of cellular targets of 3,4-DGE to renal tubular epithelial cells and unravels the molecular mechanisms underlying its lethal potential. Information was gathered on the molecular pathways of 3,4-DGE–induced apoptosis. Bax appears to be a central mediator of cell death in this system. A stimulus-specific role of Bax in tubular epithelial cell apoptosis had previously been described. Bax is required for cyclosporine A–induced apoptosis, but it is not implicated in death promoted by paracetamol (23,29). In addition, Bax had previously been implicated in apoptosis found in a diabetic milieu. Indeed high glucose–induced blastocyst apoptosis does not take place in the absence of Bax (10). Our results suggest that either high glucose and 3,4-DGE share similar intracellular death pathways, or that high ambient glucose increases intracellular 3,4-DGE levels, and this leads to apoptosis. After recruitment of Bax to the mitochondria, cytochrome c is released and the caspase cascade activated. Inhibition studies suggest that in apoptosis induced by 3,4-DGE, multiple caspases are activated and that individual inhibition of any of them is enough to prevent apoptosis. Previous experience with caspase inhibitors in this cell line argues against nonspecific inhibition of multiple caspases by a single caspase inhibitor because different patterns of protection by caspase inhibitors were observed in a stimulus-specific fashion (23,29). The pattern of protection offered by caspase inhibitors is similar to that observed in another form of Bax-mediated cell death induced by cyclosporine A (23). However, contrary to Bax inhibition, caspase inhibition did not prevent eventual cell death induced by 3,4-DGE. This suggests that Bax may be a more significant target for eventual therapeutic intervention and that once the mitochondria has been injured, caspase-independent lethal pathways are activated (30).

The current experimental data may be relevant to two clinical situations: peritoneal dialysis and diabetic nephropathy. Glucose degradation products are thought to be pivotal mediators of the poor biocompatibility of peritoneal dialysis solutions. Peritoneal dialysis solutions may have local (peritoneal) and systemic adverse effects as a consequence of poor biocompatibility. There is evidence that glucose degradation products from peritoneal dialysis fluid enter the systemic circulation, where they promote an increase in plasma advanced glycation end product compounds (13,14) and may influence appetite (12). More recently, a randomized clinical trial reported that glucose degradation product–poor peritoneal dialysis solutions better preserve residual renal function in peritoneal dialysis patients than traditional glucose degradation product–rich peritoneal dialysis solutions (14). 3,4-DGE is the most cytotoxic glucose degradation product in peritoneal dialysis fluids, so it is conceivable that it plays a role in different aspects of the bioincompatibility of these solutions. We hypothesize that the nephrotoxic potential of 3,4-DGE may have contributed to the results observed by Williams et al. (14). The concentration of 3,4-DGE in peritoneal dialysis bags may be as high as 125 μmol/l in 1.5% glucose peritoneal dialysis solutions and even higher in 4.25% glucose bags (17,21). Although the plasma concentration of 3,4-DGE after infusion of bioincompatible peritoneal dialysis solutions is not currently known, the toxic range in cultured cells is relevant to the concentrations found in peritoneal dialysis fluids and the expected passage of the molecule to the organism from the dyalisate.

A role for 3,4-DGE in diabetes complications should also be explored. The concentration of several glucose degradation products (e.g., 3-DG) is increased in diabetic individuals (31,32). 3-DG levels are well correlated with plasma glucose and HbA_1c levels (18). In addition, fasting serum 3-DG levels were higher in diabetic patients suffering from more severe evidence of diabetic kidney microangiopathy (19). 3-DG concentration is severalfold higher inside the cell than in the extracellular milieu (20). Intracellular 3-DG is also correlated with glycemic control in diabetes (20). 3,4-DGE is generated nonenzymatically from dehydration of 3-DG (Fig. 6). The reaction is reversible, and equilibrium between these two compounds exists. Heat moves the equilibrium toward 3,4-DGE. For example, raising the temperature from 30 to 40°C increases 3,4-DGE concentration by 50% (21). It could be speculated that in vivo factors such as fever may tilt the balance toward generation of 3,4-DGE. Although the serum or tissue levels of 3,4-DGE in diabetic patients are not known, the hypothesis that high 3,4-DGE levels may be generated systemically or locally in tissues in diabetic patients with poor glycemic control and/or fever merits further testing. Indeed, there are reports that acute renal failure may be more frequent in diabetic patients with bacteriemia than in nondiabetic individuals (33)

In summary, 3,4-DGE, a glucose degradation product that may be absorbed from the peritoneum during peritoneal dialysis or produced locally inside the cell in diabetic individuals, promotes apoptosis in renal tubular epithelial cells. The intracellular molecular pathways require Bax in a similar manner as glucose-induced apoptosis in cells such as those in the blastocyst (10). The hypothesis that 3,4-DGE may participate in the pathogenesis of diabetic renal injury should be further explored.

This work was supported by grants from the Comunidad de Madrid (08.2/0030.1/2003) and the Ministerio de Ciencia y Tecnología (2003/884) and by European Sixth Framework Program Grant LSHB-CT-2003-503364 (ADDNET). P.J. and A.B.S. were supported by Fondo de Investigaciones Sanitarias.

We also thank María M. González for advice and assistance with confocal microscopy.