Advanced glycation end products (AGEs) are elevated in aged and diabetic individuals and are associated with pathological changes associated with both. Previously we demonstrated that the AGE Nε-(carboxymethyl)lysine (CML)-collagen induced fibroblast apoptosis through the cytoplasmic and mitochondrial pathways and the global induction of proapoptotic genes. In the present study we investigated upstream mechanisms of CML-collagen-induced apoptosis. CML-collagen induced activation of the proapoptotic transcription factor FOXO1 compared with unmodified collagen. When FOXO1 was silenced, CML-collagen-stimulated apoptosis was reduced by ∼75% compared with fibroblasts incubated with nonsilencing small interfering RNA, demonstrating the functional significance of FOXO1 activation (P < 0.05). CML-collagen but not control collagen also induced a 3.3-fold increase in p38 and a 5.6-fold increase in JNK(1/2) activity (P < 0.05). With the use of specific inhibitors, activation of p38 and JNK was shown to play an important role in CML-collagen-induced activation of FOXO1 and caspase-3. Moreover, inhibition of p38 and JNK reduced CML-collagen-stimulated apoptosis by 48 and 57%, respectively, and by 89% when used together (P < 0.05). In contrast, inhibition of the phosphatidylinositol 3-kinase/Akt pathway enhanced FOXO1 activation. p38 and JNK stimulation by CML-collagen was almost entirely blocked when formation of ROS was inhibited and was partially reduced by NO and ceramide inhibitors. These inhibitors also reduced apoptosis to a similar extent. Together these data support a model in which AGE-induced apoptosis involves the formation of ROS, NO, and ceramide and leads to p38 and JNK MAP kinase activation, which in turn induces FOXO1 and caspase-3.
- connective tissue
- mitogen-activated protein kinase
- reactive oxygen species
advanced glycation end products (AGEs) result from nonenzymatic reactions of carbohydrates and oxidized lipids with proteins (33). The formation of ketoamine (Amadori product) intermediates lead to protein cross links and other modifications (40). The Nε-(carboxymethyl)lysine (CML) structure is a prevalent AGE and is generated principally by oxidative cleavage of Amadori intermediates (40). AGE accumulation later in life is thought to contribute to pathological changes that occur during aging (36). These include cataract formation, Alzheimer’s disease, osteoarthritis and changes in myocardial dysfunction (17, 29, 34, 38, 41). AGE formation is greatly accelerated by hyperglycemia (18, 42). In animal models, several diabetic complications have been linked to AGEs and include diabetes-associated cataract formation, nephropathy, retinopathy, neuropathy, periodontal disease, and impaired dermal and osseous wound healing (6, 12, 24, 35, 37).
One potential mechanism through which AGEs may exert deleterious effects is through programmed cell death. AGEs are proapoptotic for cultured microvascular cells, neuronal cells, fibroblasts, and renal mesangial cells (2, 22, 25, 32, 39). These apoptotic effects have been associated with diabetic complications such as atherosclerosis, nephropathy, neuropathy, and retinopathy. We recently demonstrated that an AGE, CML-collagen, induced apoptosis in fibroblasts through receptor for AGE (RAGE) signaling that led to activation of both the mitochondrial and cytosolic apoptotic pathways (2). Furthermore, CML-collagen stimulated a global induction of proapoptotic genes that involved several classes of molecules: ligands, receptors, adaptor molecules, mitochondrial proteins, and caspases.
AGEs have been reported to stimulate several upstream signaling pathways. AGEs increase the formation of intracellular reactive oxygen species (ROS), NO, and ceramides as well as the mitogen-activated protein kinase (MAPK) cascade, which, through intermediate molecules, activates different targets including transcription factors such as NF-κB and AP1 (7, 8, 36). Activation of these transcription factors subsequently increases expression of different molecules such as inducible nitric oxide synthase or inflammatory cytokines such as TNF or IL-1 (7, 36). The MAPKs are a family of serine/threonine kinases. There are several MAPKs that can participate in apoptosis. JNK and p38 MAPK are activated by a variety of stress signals including TNF, IL-1, ionizing and UV irradiation, hyperosmotic stress, and chemotherapeutic drugs, and their activity correlates well with apoptosis induced by these stimuli (23, 28).
The studies reported presently focused on upstream signaling to establish mechanisms that lead to CML-collagen-stimulated caspase-3 activity and apoptosis. The results indicate that CML-collagen compared with control collagen induces fibroblast apoptosis through a process that involves activation of the ROS, NO, or ceramide, leading to stimulation of MAPK pathways and the formation of the transcription factor FOXO1.
MATERIALS AND METHODS
CML-collagen was prepared by chemical modification of acid-soluble bovine skin collagen (Sigma) as previously described (19, 37). Briefly, 50 mg of collagen were dissolved in 25 ml of 1 mM HCl freshly made in sterile water and incubated at 37°C with occasional mixing. Sterile PBS (pH 7.8; 25 ml) was added, followed by sodium cyanoborohydride (1.42 g) and sodium glyoxyliate acid (0.715 g). Control collagen was prepared at the same time, except that no glyoxyliate acid was added. All samples were then incubated at 37°C for 24 h. AGE collagen and control collagen were then exhaustively dialyzed against distilled water. Both CML-collagen and control collagen were soluble at the concentrations stored and tested. In total, 3–8% of lysine residues in CML-collagen were converted to CML, as determined using the trinitrobenzenesulfonic acid assay (13). The percent modification of collagen that we have generated is 10-fold less than the amount used in a recent report to assess CML binding and activation of NF-κB (19) and only a small amount greater than that reported for the skin of aged or diabetic individuals (10). CML-collagen was highly reactive on Western blots with anti-CML monoclonal antibody 6D12 (Wako, Richmond, VA), whereas control collagen was not reactive.
Primary human adult dermal fibroblasts were purchased from Cambrex (Walkersville, MD). Cells were propagated and maintained in Dulbecco’s modified Eagle’s medium (Cambrex) supplemented with 10% fetal bovine serum, gentamicin (100 μg/ml), and amphotericin B (100 ng/ml) at 37°C in a humidified atmosphere of 5% CO2. Experiments with CML-collagen were performed in culture medium supplemented with 0.5% fetal bovine serum. Assays were performed when the cultures reached 75–85% confluency. In most experiments cells were incubated for 24 h with 200 μg/ml CML-collagen or unmodified control collagen based on our previous results (2). Apoptosis of fibroblasts was determined by performing ELISA to measure histone-associated DNA fragments (Roche Applied Science, Indianapolis, IN) normalized by the cell number per well. The statistical difference between samples was determined using analysis of variance followed by Tukey’s multiple-comparison tests. In some cases cells were preincubated for 2 h with or without the p38 inhibitor SB 203580 (10 μM), JNK inhibitor (H-GRKKRRQRRRPPRPKRPTTLNLFPQVPRSQDT-NH2; 10 μM), the MEK/ERK inhibitor PD-98059 (20 μM), the Akt inhibitor triciribine (1 μM), the phosphatidylinositol 3-kinase (PI3K) inhibitor LY-294002 (10 μM), the nitric oxide synthase (NOS) inhibitor NG-nitro-l-arginine methyl ester (l-NAME; 1 mM) (Calbiochem, La Jolla, CA), the ROS inhibitor N-acetyl-l-cysteine (NAC; 5 mM), or the ceramide synthase inhibitor desipramine (20 μM) (Sigma-Aldrich). The concentrations of inhibitors were selected based on a dose study (data not shown) and are similar to doses used in previously published studies (4, 8, 15, 16, 20, 21, 27, 31, 43). The extent of apoptosis was determined using ELISA.
Primary human adult dermal fibroblasts were stimulated for 30 min with 200 μg/ml CML-collagen or unmodified collagen. Total protein was extracted in the presence of phosphatase inhibitors, and concentrations were determined using the bicinchoninic acid assay (BCA; Pierce, Rockford, IL). The phosphorylation of selected kinases was analyzed by Kinexus Bioinformatics Services (Vancouver, Canada) using specific antibodies that recognize the activated epitope of respected kinases by immunoblot analysis.
Fibroblast cultures were incubated in assay medium for 1 h with 200 μg/ml of CML-collagen or unmodified collagen in the presence or absence of specific JNK, p38, PI3K, or Akt inhibitors. Nuclear proteins were extracted using a protein extraction kit (Pierce), and the protein concentration was measured using a BCA assay kit (Pierce). Interactions between nuclear proteins and FOXO1 DNA probe (CAAAACAA) were investigated using an EMSA kit from Panomics (Redwood City, CA) following the manufacturer’s instructions and as previously reported (26). Specificity was demonstrated for each by adding a 60-fold molar excess of unlabeled oligonucleotide. Each experiment was performed two to three times with similar results.
Small interfering RNA experiments.
Experiments were carried out in primary adult fibroblasts that were ∼70–80% confluent. Cells were transfected by small interfering (si) RNA using Trans-messenger transfection reagent following the manufacturer’s instructions (Qiagen, Valencia CA). siRNAs specific for FOXO1 were designed by Qiagen to be nonhomologous with other genes and have been described previously (1). siRNA-890 GCCCUGGCUCUCACAGCAA silences FOXO1 strongly and was compared with scrambled siRNA AATTCTCCGAACGTGTCACGT. Apoptosis was measured by detecting cytoplasmic histone-associated DNA (Roche). Experiments were carried out three times. Statistical difference between samples was determined using analysis of variance followed by Tukey’s multiple-comparison test.
JNK and p38 activation assay.
Primary human adult dermal fibroblasts were stimulated for up to 30 min with 200 μg/ml CML-collagen or unmodified collagen. Total cellular protein was extracted and measured as described above. The p38 activation epitopes (Thr180 and Tyr182) and JNK1/2 activation epitopes (Thr183 and Tyr185) were quantified using ELISA (EMD Biosciences, San Diego, CA). A standard curve was run for each so that the data are expressed in activity units.
Cells were incubated with control collagen or CML-collagen (200 μg/ml) as described above in the presence or absence of p38 or JNK inhibitors. Caspase-3 activity was measured with fluorimetric kits purchased from R&D Systems.
FOXO1 activation mediates TNF-induced proapoptotic gene expression and is necessary for TNF to maximally stimulate fibroblast apoptosis (1). Experiments were undertaken to determine whether CML-collagen induced FOXO1 and whether its activation played a role in CML-collagen-stimulated fibroblast apoptosis. CML-collagen but not control collagen stimulated FOXO1 activation as measured by EMSA (Fig. 1A). Detection of FOXO1 was blocked by incubation with excess unlabeled competitive FOXO1 oligonucleotide but not by noncompetitive oligonucleotide. siRNA studies were then carried out to establish the functional role of FOXO1 activation (Fig. 1B). Cells were preincubated with siRNA that silences FOXO1 (1) and compared with results for scrambled siRNA. In the absence of CML-collagen stimulation, there was little basal apoptosis, which was not affected by transfection with FOXO1 siRNA. CML-collagen stimulated a more than threefold increase in apoptosis. Preincubation with FOXO1 siRNA reduced the level of apoptosis by 76% (P < 0.05). Transfection per se was not responsible for the reduced apoptosis, since nonsilencing scrambled siRNA had no effect.
To investigate potential pathways activated by CML-collagen, we carried out a screen of signaling molecules phosphorylated in response to CML-collagen and detected members of the MAPK pathway (data not shown). Therefore, experiments were performed to measure CML-collagen activation of p38 and JNK, proapoptotic arms of the MAPK pathway (Fig. 2). At 5, 20, and 30 min, CML-collagen stimulation increased p38 activity 3.3-, 2.1-, and 1.6-fold, respectively. The increase in p38 activity at 5 and 20 min was statistically significant (P < 0.05). At 5, 20, and 30 min, CML-collagen stimulation increased JNK(1/2) activity 2.2-, 4.1-, and 5.6-fold, which was statistically significant at all time points (P < 0.05). Western blot analysis demonstrated a 2.7-fold increase in p38 activity and a 5.9-fold increase in JNK activation (Fig. 2C).
To determine whether p38 and JNK mediated CML-collagen-stimulated activation of FOXO1, caspase-3, and apoptosis, we preincubated dermal fibroblasts without or with specific inhibitors for p38 (SB 203580) or JNK (H-GRKKRRQRRRPPRPKRPTTLNLFPQVPRSQDT-NH2) followed by incubation with CML-collagen or control collagen with or without inhibitors. CML-collagen induced FOXO1 activation that was substantially reduced by both p38 and JNK inhibitors as determined by EMSA, demonstrating that CML-collagen induces FOXO1 activation through both arms of the MAPK pathway (Fig. 3A). CML-collagen stimulated a 4.3-fold increase in caspase-3 activity (Fig. 3B). Sixty-five percent of CML-collagen-induced caspase-3 activity was blocked by treatment with JNK inhibitor (P < 0.05). The p38 inhibitor reduced CML-collagen-stimulated caspase-3 activity by 52% (P < 0.05). Apoptosis was next measured as the relative amount of histone-associated DNA fragments in the cytoplasm as determined by ELISA (Fig. 3C). CML-collagen compared with unmodified collagen induced a 4.1-fold increase in apoptosis, which was statistically significant (P < 0.05). p38 inhibition decreased CML-collagen-induced apoptosis by 48%, and JNK inhibition decreased apoptosis by 57%, both of which were statistically significant (P < 0.05). Doubling the concentration of p38 and JNK inhibitors did not significantly increase their inhibitory effect on CML-collagen-stimulated apoptosis (data not shown). However, when both inhibitors were used simultaneously, CML-collagen-stimulated apoptosis was reduced by 89%, which was greater than that reduced by either inhibitor alone (P < 0.05). Thus, both p38 and JNK arms of this signaling pathway are involved in CML-collagen-induced apoptosis.
To investigate whether CML-collagen modulated antiapoptotic pathways, we examined two well-defined pathways, PI3 kinase/Akt and the MEK/ERK (Fig. 4). PI3K inhibitor (LY-294002) at 10 μM increased the apoptotic effect of CML-collagen by 41%, which was statistically significant. PI3K inhibitor had no effect when incubated with unmodified collagen (P > 0.05). The Akt inhibitor triciribine (1 μM) also increased the apoptotic effect of CML-collagen by 39% (P < 0.05) but did not enhance apoptosis when added to unmodified collagen (P > 0.05). In contrast, the MEK/ERK inhibitor did not modulate apoptosis when incubated with CML-collagen or unmodified collagen (P > 0.05). When the concentration of PI3K, Akt, and MEK/ERK inhibitors was doubled, there was no additional change in apoptosis (data not shown).
To investigate the effect of CML-collagen on FOXO1 activation in the presence of PI3K or AKT inhibitors, we measured the activation of FOXO1 using EMSA (Fig. 4D). FOXO1 activation in response to CML-collagen stimulation was slightly enhanced in the presence of PI3K or AKT inhibitors. The results are consistent with reports that the PI3K/Akt pathway inhibits FOXO1 activation (11).
To investigate upstream events that potentially lead to activation of p38 and JNK, we used inhibitors of ROS, NO, and ceramide formation. Fibroblasts were preincubated with inhibitors and then stimulated with CML-collagen, and p38 and JNK activity were measured using ELISA. The ROS inhibitor completely blocked CML-collagen-stimulated p38 activity, whereas the NOS inhibitor decreased p38 activity by 66% (P < 0.05) (Fig. 5). The ceramide inhibitor decreased p38 activity by 27% (P < 0.05). The ROS inhibitor also completely blocked CML-collagen-stimulated JNK activity. The NOS inhibitor decreased JNK activity by 54%, and the ceramide synthase inhibitor reduced CML-collagen-stimulated JNK activity by 33%, both of which were significant (P < 0.05).
To determine the significance of ROS in AGE-induced apoptosis, we treated human adult primary skin fibroblasts with an inhibitor that blocks generation of ROS (NAC), a broad NOS inhibitor (l-NAME), or a ceramide inhibitor (desipramine) for 2 h, followed by CML-collagen (200 μg/ml) or collagen stimulation for 24 h (Fig. 6). When generation of ROS was blocked, CML-collagen-induced apoptosis was reduced 92% (P < 0.05). Blocking formation of nitric oxide caused a 61% decrease, and blocking formation of ceramide resulted in a 33% reduction in CML-stimulated apoptosis, both of which were significant (P < 0.05). When the concentration of ROS, NOS, or ceramide inhibitors was increased twofold, there was no further effect on CML-collagen-induced apoptosis (data not shown).
We and others have shown that the AGE CML-collagen induces apoptosis in fibroblasts that is mediated through RAGE and activation of caspase-3 that is largely mediated by caspase-8 (2). In addition, CML-collagen induces expression of more than 35 proapoptotic genes (2). To investigate more upstream events, we examined the activation of the proapoptotic transcription factor FOXO1. We presently report that CML-collagen activates FOXO1 and that CML-collagen-induced apoptosis is reduced by ∼75% when FOXO1 is silenced. Thus FOXO1 plays an important role in apoptosis induced by both death receptor ligands (5) and, as shown in the present study, by an AGE.
To determine whether there are common pathways in CML-collagen-induced FOXO1 activation and apoptosis, we investigated the MAPK pathway. The functional role of p38 and JNK in mediating CML-collagen-induced FOXO1 activation was demonstrated by using inhibitors of p38 and JNK, consistent with the observation that CML-collagen activates p38 and JNK. Likewise, inhibition of p38 and JNK reduced CML-collagen-stimulated apoptosis by 48 and 57%, respectively, and by 89% when used together. The role of p38 and JNK in AGE-induced apoptosis was further supported by significantly reduced caspase-3 activity when CML-collagen was incubated with p38 and JNK inhibitors. Furthermore, when the PI3K/Akt pathway was inhibited, there was a small but significant increase in CML-collagen-induced apoptosis and enhanced FOXO1 activation. PI3K and its target Akt play an important role in survival of cells, in part, through inhibition of FOXO1 activation (5). Thus CML-collagen simultaneously activates this antiapoptotic pathway, which is overridden by proapoptotic p38 and JNK signaling.
JNK and p38 MAPK respond strongly to a variety of stress signals including TNF, IL-1, ionizing and UV irradiation, hyperosmotic stress, and chemotherapeutic drugs and have been implicated in mediating apoptotic responses (23, 28). In contrast, activation of the ERK pathway is associated with survival induced by antiapoptotic signals such as growth factors (9). We found that inhibition of MEK/ERK had no effect on CML-collagen-induced apoptosis, suggesting that in contrast to the PI3K/Akt, the ERK pathway does not play an inhibitory role.
The mechanism through which CML-collagen activates p38 and JNK was shown to be through the generation of ROS, since inhibition of ROS formation blocked most CML-collagen-stimulated p38 and JNK activity. In addition to activating the MAPK pathway, ROS generation was also critical in CML-collagen-stimulated fibroblast apoptosis. Our results are consistent with previous reports that generation of ROS are important intermediates in apoptosis signaling by other stimuli ranging from UV irradiation to cancer therapeutics (30) and are consistent with the pervious report that AGEs induce ROS (19). Given that RAGE signaling mediates CML-collagen-induced apoptosis in fibroblasts (18), it is possible that RAGE leads to the production of ROS through activation of NADPH oxidase (44). Although ROS played the dominant role, two other proapoptotic intermediates, NO and ceramide, contributed to CML-collagen-induced stimulation of p38 and JNK activity and apoptosis. These results are in agreement with reports demonstrating that AGEs stimulate production of NO and ceramide, which are proapoptotic (3, 9, 14). Thus it is possible that a sequence of events occurs in which activation of RAGE leads to increased NADPH oxidase activation, which enhances formation of ROS, which leads to activation of the p38/JNK arms of the MAPK pathway and FOXO1 activation. FOXO1 activation can promote fibroblast apoptosis through induction of proapoptotic or inhibition of antiapoptotic gene expression (1).
On the basis of these results, we suggest the following pathway as one of the possible mechanisms through which apoptosis can occur. After AGEs-RAGE interaction, ROS increases inside the cells that activate both NOS and ceramides, which in turn activates p38 and JNK. Activated p38 and JNK activate a cascade that leads to enhanced cascade-3 activity, whereas activation of FOXO1 potentiates the likelihood of a cell undergoing apoptosis, most likely through enhanced expression of proapoptotic genes. Given that healing in diabetic mice is associated with higher levels of fibroblast apoptosis, the formation of AGEs may impair diabetic healing, in part, by stimulating fibroblast apoptosis. The identification of signaling pathways that lead to apoptosis of these cells may be useful in designing strategies to improve healing under conditions where AGEs accumulate, e.g., in diabetes and aged individuals.
In summary, the advanced glycation end product CML-collagen induces apoptosis in fibroblasts through a process that is dependent on FOXO1. FOXO1 activation is induced by p38 and JNK, which in turn, depend on the formation of ROS. Thus p38 and JNK are important mediators of CML-collagen-induced FOXO1, and FOXO1 activation represents an important mechanism through which the MAPK signaling pathway induces cell death. These events are critical, since inhibition of each significantly reduces CML-collagen-induced apoptosis. Given that apoptosis may contribute to pathologies associated with diabetes and aging, an understanding of the signaling mechanisms may provide therapeutic targets that are ultimately of clinical benefit.
This work was supported by National Institutes of Health Grants R01 DE-07559, PO1 AR-49920, and R01 DE-14066.
We thank Alicia Ruff for help in preparing this manuscript.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society