Alterations in molecular chaperones and eIF2α during lung endothelial cell apoptosis

Abstract

We have previously demonstrated that inhibition of CAAX carboxyl methylation with AGGC caused redistribution and condensation of the ER molecular chaperones, glucose-regulated protein (GRP)-94 and calnexin; an effect that was attenuated by overexpression of dominant active RhoA. We have also shown that AGGC decreased GRP94 protein level; an effect that was dependent on caspase activity. In the present study, we tested the effects of inhibition of posttranslational processing of CAAX proteins on localization and protein levels of molecular chaperones and phosphorylation and protein level of eIF2α. We found that both AGGC, which inhibits CAAX carboxyl methylation, and simvastatin, which inhibits CAAX geranylgeranylation, caused relocalization of GRP94, calnexin, and calreticulin, effects that were not seen during endothelial apoptosis induced by TNF-α or ultraviolet (UV) irradiation. These results suggest that posttranslational processing of CAAX proteins is important in maintaining localization of molecular chaperones normally found in the ER. We also noted that AGGC, but not simvastatin, TNF-α, or UV irradiation, decreased protein levels of most molecular chaperones. Increased eIF2α phosphorylation was observed in the early stages of apoptosis, which was independent of the cause of apoptosis. These results suggest that eIF2α phosphorylation is a common early response to apoptosis-inducing stimuli. Interestingly, eIF2α protein level was decreased in the late stages of apoptosis induced by AGGC, TNF-α, and UV irradiation: an effect that was prevented by caspase inhibition. Thus we speculate that caspase(s)-dependent proteolysis of molecular chaperones and eIF2α may be novel signaling pathways of apoptosis. We also speculate that increased eIF2α phosphorylation is a defensive response against endothelial cell apoptosis.

the cooh terminal CAAX motif (C: cysteine; A: aliphatic amino acid; X: any amino acid) of small GTPases is sequentially modified by isoprenylation of the cysteine residue by geranylgeranyltransferase or farnesyltransferase, followed by cleavage of the AAX residues by prenyl-CAAX protease and subsequent carboxyl methylation of the prenylcysteine by isoprenylcysteine-O-carboxyl methyltransferase (ICMT) (6). ICMT is an endoplasmic reticulum (ER) resident integral membrane protein (7, 33). Many substrates of ICMT have been putatively identified based on sequence characteristics; among these are small GTPases (6). Posttranslational processing of small GTPases is essential for their membrane localization and biological activation (4, 5). Inhibition of CAAX geranylgeranylation with the pharmacological inhibitor simvastatin has been shown to inhibit RhoA activity: an effect that was associated with endothelial apoptosis (12, 15, 40). Similarly, we have previously demonstrated that inhibition of CAAX carboxyl methylation with ICMT chemical inhibitor AGGC decreased carboxyl methylation and subsequent activation of RhoA and Ras GTPases, effects that were also correlated with endothelial apoptosis (14, 19, 20). To further investigate the mechanism underlying ICMT inhibition-induced endothelial apoptosis, we have shown that ICMT inhibition caused relocalization and condensation of ER molecular chaperones, glucose-regulated protein (GRP)-94, and calnexin: an effect that was ameliorated by overexpression of constitutively active RhoA (20). In addition, ICMT inhibition decreased GRP94 protein level: an effect that was prevented by a broad caspase inhibitor (20). Whether inhibition of posttranslational processing of the CAAX motif of small GTPases alters localization and protein levels of other molecular chaperones is not known.

ER molecular chaperones, including GRP78, GRP94, calnexin, and calreticulin, are central regulators of ER homeostasis due to their critical roles in ER protein folding, ER calcium balancing, and controlling the activation of unfolded protein response (UPR) (36). Induction of ER molecular chaperones is a prosurvival signal in conditions of ER stress (36). We have previously shown that suppression of GRP94 by siRNA exacerbated ICMT inhibition-induced endothelial cell apoptosis (20). Although ER molecular chaperones are induced by ER stress and may protect against ER stress-induced apoptosis, there are reports that caspase-mediated proteolytic cleavage of molecular chaperones, such as GRP94 and heat shock protein (HSP) 90, occurs during apoptosis (24, 30, 32). In addition to the chaperone proteins, increased eIF2α phosphorylation is a hallmark of UPR in response to ER stress, contributing to restoration of ER homeostasis by inhibiting global protein translation and thus reducing ER client protein loading (36). In contrast, sustained eIF2α phosphorylation has been implicated in TNF-α-induced apoptosis due to prolonged inhibition of protein translation (35). In addition, eIF2α can be cleaved by caspase 3, 6, 8, and 10, resulting in diminished protein translation, thus contributing to apoptosis (23). Whether endothelial cell apoptosis occurs after inhibition of posttranslational processing of CAAX via effects on ER molecular chaperones or eIF2α is not known. We hypothesize that inhibition of posttranslational processing of CAAX proteins by AGGC and simvastatin causes endothelial apoptosis through relocalization and caspase-mediated proteolysis of ER molecular chaperones and eIF2α.

The overall goal of this study was to determine the effects of inhibition of posttranslational processing of CAAX proteins on ER molecular chaperones and eIF2α. Apoptosis occurs through two fundamental signaling pathways: intrinsic and extrinsic. TNF-α promotes apoptosis through activation of the extrinsic pathway (39, 43), whereas UV irradiation causes apoptosis through activation of the intrinsic pathway (26). Thus we utilized AGGC, simvastatin, TNF-α, and UV irradiation to cause endothelial apoptosis and compared their effects on molecular chaperones and eIF2α. We found that both AGGC and simvastatin, but not TNF-α or UV irradiation, caused relocalization of GRP94, calnexin, and calreticulin. However, we noted that only AGGC decreased protein levels of molecular chaperones. Phosphorylation of eIF2α was increased during the early stages of apoptosis, independent of apoptosis-inducing stimuli. However, eIF2α was ultimately cleaved/proteolysed; events that were dependent on caspase activity. These results suggest that posttranslational processing of CAAX proteins is important in maintenance of appropriate localization of ER molecular chaperones. Additionally, increased eIF2α phosphorylation is a common response to apoptosis-inducing stimuli, and cleavage/proteolysis of eIF2α may be a common pathway of apoptosis.

MATERIALS AND METHODS

Cell lines and reagents.

Bovine pulmonary artery endothelial cells (PAEC) were purchased from Vec Technologies (Renesselaer, NY). Antibodies directed against GRP94 and calnexin were purchased from Assay Designs (Ann Arbor, MI). Antibodies directed against phospho-eIF2α, calreticulin, pro-caspase-3, and cleaved caspase-3 were purchased from Cell Signaling Technology (Danvers, MA). Simvastatin and antibody directed against GRP78 were purchased from Sigma Aldrich (St. Louis, MO). Antibody directed against eIF2α was purchased from Biolegend (San Diego, CA). Antibodies directed against HSP70 and HSP90 were obtained from BD Biosciences (San Diego, CA). Antibody directed against lamin A/C was purchased from Santa Cruz (Santa Cruz, CA). N-acetyl-S-geranylgeranyl-l-cysteine (AGGC) was obtained from Biomol (Plymouth Meeting, PA). zVAD-fmk was obtained from Axxora (San Diego, CA). TNF-α was purchased from EMD (La Jolla, CA).

Assessment of apoptosis.

Endothelial cell apoptosis was assessed by DAPI staining of pycnotic nuclei, as previously described (10). Five randomized high-power fields per slide were used to examine the percentage of apoptotic cells. The data is presented as the mean of percentage of apoptotic cells over the total cells counted in each slide.

Gel electrophoresis and immunoblot analysis.

Lysates were solubilized in Laemmli buffer, and proteins were resolved by SDS-PAGE. The resolved proteins were then transferred to Immunobilon PVDF membranes and immunoblotted with indicated antibodies, as we previously described (18, 20).

Immunofluorescence microscopy.

Endothelial cells grown on coverslips were treated as described, fixed with 4% paraformaldehyde, and permeabilized with 0.1% Triton X-100. Cells were then stained with primary antibodies followed by Texas red-conjugated (red) or FITC-conjugated (green) species-specific secondary antibodies. Images were visualized using a Nikon Eclipse E400 fluorescence microscope at ×1,000 magnification and recorded, as we previously described (20).

Data analysis.

All experiments were performed at least in triplicate. Data are presented as means ± SE. ANOVA was performed, and significance among the groups was determined using Fisher's PLSO post hoc test. Differences among means were considered significant when P < 0.05.

RESULTS

Relocalization and condensation of ER molecular chaperones during endothelial apoptosis induced by AGGC and simvastatin.

We have previously shown that inhibition of CAAX carboxyl methylation by AGGC caused endothelial apoptosis: an effect that was associated with subcellular relocalization and condensation of ER molecular chaperones GRP94 and calnexin (20). Thus we investigated whether inhibition of CAAX geranylgeranylation with simvastatin has similar effects. As shown in Supplemental Fig. 1, A and B (see supplemental material available online at Am J Physiol Lung Mol Cell Physiol website), pulmonary artery endothelial cells (PAEC) exposed to either AGGC or simvastatin demonstrated a significant increase in cleavage of pro-caspase-3 and lamin A/C and an increase in apoptotic nuclei. Thus we confirm that inhibition of posttranslational processing of CAAX proteins causes pulmonary endothelial cell apoptosis. Similar to our previously reported results regarding GRP94 and calnexin (20), AGGC caused loss of perinuclear localization of another ER molecular chaperone, calreticulin (Fig. 1A). We then examined the effect of simvastatin on localization of ER molecular chaperones. As previously shown (20), GRP94, calnexin, and calreticulin displayed a perinuclear staining in controls, indicating an ER localization (Fig. 1B, top). However, PAEC exposed to simvastatin for 24 h displayed a dramatic relocalization and condensation of GRP94 and calnexin, as well as diffusion of calreticulin (Fig. 1B, bottom). To determine whether such a relocalization of ER molecular chaperones is a common event during the process of apoptosis, we examined the effect of TNF-α and UV irradiation on localization of these ER molecular chaperones. PAEC exposed to either TNF-α or UV light also demonstrated a significant increase in cleavage of pro-caspase-3 and lamin A/C (supplemental Fig. 1, C and D, available online at Am J Physiol Lung Mol Cell Physiol website). However, neither TNF-α nor UV irradiation affected subcellular localization of these ER molecular chaperones (Fig. 1, C and D). These results suggest that redistribution and condensation of ER molecular chaperones is unique to inhibition of posttranslational processing of CAAX proteins.

Fig. 1.

Relocalization of endoplasmic reticulum (ER) molecular chaperones during endothelial cell apoptosis. PAEC were incubated with vehicle or 10 μM AGGC in serum-free MEM medium for 3 h (A), with vehicle or 25 μM simvastatin in MEM medium containing 10% FBS for 24 h (B), with vehicle or 20 ng/ml of TNF-α in HEPES buffer without CO₂ for 3 h (C), or were exposed to 256-nm wavelength UV light for 0 or 3 min and then incubated in serum-free MEM medium for 3 h (D). Subcellular localization of GRP94, calnexin, and calreticulin were assessed by immunofluorescence microscopy. Representative pictures from three independent experiments for each data set are shown. Arrows indicate perinuclear staining, and arrow heads indicate condensed staining.

Protein levels of molecular chaperones were diminished during endothelial cell apoptosis induced by AGGC.

We have previously shown that AGGC decreased protein level of GRP94: an effect that was dependent on caspase activity (20). To determine whether other molecular chaperones were altered during endothelial apoptosis, we examined protein levels of ER molecular chaperones, including GRP78, GRP94, calnexin, and calreticulin, as well as cytoplasmic molecular chaperones, including HSP90 and HSP70, in PAEC undergoing apoptosis after treatment with AGGC, simvastatin, TNF-α, and UV irradiation. With the exception of HSP70, the protein levels of other molecular chaperones were significantly diminished in PAEC exposed to AGGC for 24 h, effects that were blunted by caspase inhibitor, zVAD-fmk (Fig. 2A). These results suggest that AGGC-activated caspase(s) may cause cleavage/proteolysis of molecular chaperones. We next assessed the effect of apoptosis induced by simvastatin on protein levels of these molecular chaperones. Interestingly, none of these molecular chaperone protein levels were altered in PAEC exposed to simvastatin for 24 h (data not shown) or 48 h (Fig. 2B). Similarly, none of these molecular chaperone protein levels were significantly affected by TNF-α (Fig. 2C). Although the protein levels of GRP94, calnexin, calreticulin, HSP90, and HSP70 were not altered, GRP78 protein level was significantly enhanced by exposure to UV light: an effect that was independent of caspase activity (Fig. 2D).

Fig. 2.

Alterations in protein levels of molecular chaperones during endothelial apoptosis. PAEC were preincubated with vehicle or 100 μM of zVAD-fmk for 1 h and then incubated with vehicle or 10 μM of AGGC in HEPES buffer without CO₂ in the absence or presence of 100 μM zVAD-fmk for 24 h (A), incubated with vehicle or indicated concentrations of simvastatin in serum-free MEM medium for 48 h (B), preincubated with vehicle or 100 μM of zVAD-fmk for 1 h and then incubated with vehicle or 20 ng/ml of TNF-α in HEPES buffer without CO₂ in the absence or presence of 100 μM of zVAD-fmk for 24 h (C), or preincubated with vehicle or 100 μM of zVAD-fmk for 1 h and then exposed to 256-nm wavelength UV light for 0 (vehicle) or 3 min and then incubated with serum-free MEM medium in the absence or presence of 100 μM of zVAD-fmk for 24 h (D). Protein levels of GRP78, GRP94, calnexin, calreticulin, HSP90, and HSP70 were assessed by immunoblot analysis using equal amounts of total proteins in lysates. Equal loading of proteins was confirmed by staining the blots using Ponceau S solution (data not shown). Representative immunoblots of multiple independent experiments for each data set are shown. A: n = 3–7; B: n = 3; C: n = 3–7; D: n = 3–7.

Phosphorylation of eIF2α was increased in all models of endothelial cell apoptosis.

We next elucidated the effect of AGGC and simvastatin on eIF2α phosphorylation. PAEC exposed to AGGC for 1 and 2 h demonstrated a significant increase in eIF2α phosphorylation compared with vehicle-treated cells (Fig. 3A). However, eIF2α phosphorylation was dramatically reduced after 8-h exposure to AGGC (Fig. 3A). PAEC exposed to simvastatin for 4 and 8 h also demonstrated a significant increase in eIF2α phosphorylation (Fig. 3B). TNF-α and UV irradiation have been shown to increase eIF2α phosphorylation (35, 46). As expected, eIF2α phosphorylation was significantly enhanced in PAEC exposed to TNF-α for 2 and 4 h (Fig. 3C). Interestingly, eIF2α phosphorylation was significantly decreased after 8-h exposure to TNF-α (Fig. 3C). Similarly, increased eIF2α phosphorylation was also seen in PAEC at 2 to ∼8 h after exposure to UV light (Fig. 3D). These results suggest that increased eIF2α phosphorylation is a common event during endothelial cell apoptosis.

Fig. 3.

The effects of apoptosis-inducing stimuli on eIF2α phosphorylation. PAEC were incubated with vehicle (V) or 10 μM AGGC (A) in HEPES buffer without CO₂ for indicated times (A), incubated with vehicle (0), 10 and 25μM simvastatin in serum-free MEM medium for indicated times (B), incubated with vehicle (V) or 20 ng/ml of TNF-α (T) in HEPES buffer without CO₂ for indicated times (C), or exposed to 256-nm wavelength UV light for 0 (vehicle, V) or 3 min (UV) and then incubated with serum-free MEM medium for indicated times (D). Phosphorylated eIF2α was assessed by immunoblot analysis using equal amounts of total proteins in lysates. Equal loading of proteins was confirmed by staining the blots using Ponceau S solution (data not shown). The data is presented as the ratio of treatment with apoptotic stimuli (AGGC, simvastatin, UV, and TNF-α) to respective vehicle at each time point. A: n = 4–5; B: n = 4; C: n = 3; D: n = 3. In B, the statistical data represent 10 μM simvastatin. *Significant difference vs. respective vehicle (P < 0.05).

The protein level of eIF2α was ultimately decreased during endothelial cell apoptosis induced by AGGC, TNF-α, and UV irradiation.

Caspase-dependent eIF2α cleavage and proteolysis has been shown to mediate TNF-α-induced apoptosis (23). Thus we examined whether eIF2α cleavage and proteolysis occurred during endothelial apoptosis induced by AGGC, TNF-α, and UV exposure in endothelial cells. Exposure of PAEC to AGGC for 24 h significantly decreased both total and phosphorylated eIF2α levels; an effect that was prevented by a broad caspase inhibitor (Fig. 4A). As expected, PAEC exposed to TNF-α for 24 h demonstrated a significant reduction in both total and phosphorylated eIF2α levels; an effect that was also attenuated by caspase inhibition (Fig. 4B). Similarly, we noted that a band with a faster mobility (cleaved eIF2α) appeared, whereas a band with a slower mobility (eIF2α) disappeared in PAEC at 24 h after exposure to UV light, a change that was abolished by caspase inhibition (Fig. 4C). These results suggest that caspase-mediated eIF2α cleavage/proteolysis may be a common pathway of endothelial cell apoptosis.

Fig. 4.

The effects of apoptosis-inducing stimuli on eIF2α protein level. PAEC were preincubated with vehicle or 100 μM zVAD-fmk for 1 h and then incubated with vehicle or 10 μM AGGC in HEPES buffer without CO₂ in the absence or presence of 100 μM of zVAD-fmk for 24h (A), preincubated with vehicle or 100 μM zVAD-fmk for 1 h and then incubated with vehicle or 20 ng/ml of TNF-α in HEPES buffer without CO₂ in the absence or presence of 100 μM zVAD-fmk for 24 h (B), or preincubated with vehicle or 100 μM zVAD-fmk for 1 h and then exposed to 256-nm wavelength UV light for 0 (vehicle) or 3 min (UV) and then incubated with serum-free MEM medium in the absence or presence of 100 μM zVAD-fmk for 24 h (C). Protein levels of phosphorylated and total eIF2α were assessed by immunoblot analysis using equal amounts of total proteins in lysates. Equal loading of proteins was confirmed by staining the blots using Ponceau S solution (data not shown). Representative immunoblots of multiple independent experiments for each data set are shown. A: n = 4–8; B: n = 3; C: n = 5–7.

DISCUSSION

Using diverse stimuli to induce endothelial cell apoptosis, we demonstrated that both AGGC and simvastatin, but not TNF-α or UV irradiation, caused relocalization of ER molecular chaperones. AGGC dramatically decreased the protein levels of most ER and cytoplasmic molecular chaperones; effects that were attenuated by a broad caspase inhibitor. Although simvastatin caused similar changes in the subcellular distribution of ER molecular chaperones, it did not alter protein levels of any chaperones. eIF2α phosphorylation was increased in the early phase in all models of apoptosis examined in this study. However, eIF2α protein levels were dramatically reduced 24 h after initial apoptotic stimulation: an effect that was prevented by a caspase inhibitor. These results suggest that eIF2α proteolysis is a common pathway of endothelial cell apoptosis downstream from caspases.

ER molecular chaperones, including GRP78, GRP94, calnexin, and calreticulin, are normally localized in the ER due to two types of specific ER retention signals: a NH₂-terminal hydrophobic signal sequence and a COOH-terminal KDEL sequence (1, 38, 47). GRP78, GRP94, and calreticulin have been shown to be redistributed through an unknown mechanism from ER to cytoplasm after ER stress (1, 38, 47). We have previously shown that inhibition of CAAX carboxyl methylation with AGGC caused relocalization and condensation of GRP94 and calnexin: an effect that was associated with decreased RhoA carboxyl methylation and activity (20). In this study, we found that AGGC also caused diffusion of calreticulin to the cytoplasm. Furthermore, relocalization and condensation of these ER molecular chaperones were also seen in endothelial cells exposed to simvastatin, which inhibits RhoA geranylgeranylation and activation (12, 40), but not to TNF-α or UV light. Additionally, we have previously demonstrated that AGGC-induced relocalization and condensation of GRP94 was attenuated by overexpression of constitutively active RhoA (20). These results suggest that posttranslational processing of RhoA and resultant RhoA activation are important for maintenance of proper localization of ER molecular chaperones. There are several possible mechanisms for relocalization of ER molecular chaperones on inhibition of posttranslational processing of RhoA. RhoA family member Cdc42 has been implicated in blocking protein trafficking from Golgi to ER through regulation of actin nucleation/polymerization (21). It is possible that RhoA prevents ER molecular chaperone mislocalization through a similar mechanism. We have previously shown that a broad caspase inhibitor prevented AGGC-induced relocalization and condensation of GRP94 (20). Thus it is also possible that ER retention signal KDEL sequence is cleaved by caspases, leading to cytoplasmic diffusion of ER chaperones. Furthermore, since caspases cleave ER membrane structural proteins, such as lamin A/C (as noted in Supplemental Fig. 1, A, C, and D available online at the Am J Physiol Lung Cell Mol Physiol website), it is possible that caspase-induced ER membrane dysfunction causes diffusion of ER chaperones from ER to cytoplasm.

ER molecular chaperones are central regulators of ER homeostasis and cell survival. GRP78 controls activation of UPR sensors, whereas GRP94 is important in ER protein folding and export (25). Calnexin and calreticulin are necessary components of ER protein folding machinery. It has been shown that induction of GRP78 and calreticulin prevents oxidative stress and Ca²⁺ disturbances-induced apoptosis of renal epithelial cells (17). Overexpression of GRP94 suppresses neuronal cell death induced by perturbated calcium homeostasis and by ischemia/reperfusion injury (2, 31). The role of calnexin in regulation of apoptosis is controversial. Although calnexin-deficient cells are resistant to ER stress-induced apoptosis (49), others reported that calnexin suppresses ER stress-induced apoptosis (41). HSPs are important molecular chaperones of the protein folding machinery in the cytosol, thereby preserving the functionality of the cells under stress. HSP70 is the cytoplasmic homolog of the ER chaperone GRP78 (8), whereas HSP90 is the cytoplasmic counterpart of the ER chaperone GRP94 (9). Both HSP90 and HSP70 have been demonstrated to protect against cell apoptosis (13, 48). Although molecular chaperones are induced by stress, it has been reported that GRP94 and HSP90 are targets of caspase- and proteosome-mediated proteolysis during the process of apoptosis (24, 30, 32). We found that AGGC decreased protein levels of most molecular chaperones examined in this study: an effect that was prevented by a broad caspase inhibitor. This result suggests that caspase(s) activated by AGGC possibly decreased the levels of molecular chaperones through proteolysis, hence contributing to endothelial cell apoptosis. Although AGGC and simvastatin had similar effects on relocalization and condensation of ER chaperones, simvastatin did not alter protein levels of these ER chaperones, suggesting that AGGC may also act through other mechanisms.

ER stress causes increased eIF2α phosphorylation, which contributes to restoration of ER homeostasis by inhibiting global protein translation (36). UV irradiation has been shown to activate UPR, as indicated by increased eIF2α phosphorylation (27, 46) and GRP78 induction (28). We similarly demonstrated that UV light exposure time-dependently increased eIF2α phosphorylation: an effect that was associated with elevation of GRP78 protein level. We also noted an increase in eIF2α phosphorylation in the early phase of apoptosis induced by AGGC, simvastatin, or TNF-α. These data suggest that increased eIF2α phosphorylation is a common event in response to apoptotic stimuli. UV irradiation-induced eIF2α phosphorylation suppressed apoptosis (11, 27). Thus we speculate that enhanced eIF2α phosphorylation may represent activation of UPR, a defensive mechanism against endothelial cell apoptosis. However, it has been reported that increased eIF2α phosphorylation plays an important role in mediating TNF-α-induced apoptosis (35). Thus the role of eIF2α phosphorylation in endothelial cell apoptosis after different stimuli needs further investigations.

Another important finding of this study is that both total and phosphorylated eIF2α levels were dramatically decreased in endothelial cells exposed to apoptotic stimuli for 24 h: an effect that was prevented by a broad caspase inhibitor. Our results are consistent with a previous report that eIF2α is a target of caspase-mediated proteolysis during the process of apoptosis (23). In addition, TNF-α-induced eIF2α cleavage enhanced apoptosis in Saos-2 cells (34). Loss of eIF2α protein may contribute to apoptosis through both prolonged inhibition of global protein translation and dysfunction of UPR. Taken together, our data suggest that caspase-induced degradation of molecular chaperones and eIF2α may be novel pathways of apoptosis downstream from caspases.

ER molecular chaperones have been proposed to interfere with the pathogenesis of a variety of diseases, particularly degenerative diseases associated with protein aggregation due to misfolding (22). Decreased expression of chaperones may accelerate such diseases due to decreased protein refolding. Our results demonstrate that ICMT inhibition with AGGC decreased several ER and cytoplasmic molecular chaperones. Intriguingly, decreased expression of ICMT has been observed in spinal cerebellar ataxia, caused by misfolding and aggregation of mutant ataxin-1 (16, 44). This raises a question whether suppression of ICMT leads to decreased molecular chaperone expression in this disease, thus contributing to the pathogenesis of protein toxicity. The expression and function of ICMT in diseases caused by protein misfolding may be of future interest. Inhibition of molecular chaperones has been proposed as a therapeutic target in the treatment of cancer (29, 37). Inhibition of ICMT has also been of interest as a therapeutic target in cancer (3, 42, 45). ICMT regulation of molecular chaperones would add to the potential benefits of ICMT inhibition in the treatment of malignancy.

In summary, our data demonstrated that inhibitors of either carboxyl methylation or geranylgeranylation of COOH-terminal CAAX motif caused relocalization of ER molecular chaperones (Fig. 5). ICMT inhibition with AGGC decreased protein levels of most molecular chaperones. Increased eIF2α phosphorylation was a common early event during endothelial cell apoptosis, which may represent activation of UPR, a defensive mechanism against apoptosis. However, the level of eIF2α protein was decreased by activated caspase(s) during the late stages of apoptosis. We speculate that caspase-mediated degradation of molecular chaperones and eIF2α may be novel pathways of endothelial cell apoptosis.

Fig. 5.

Proposed mechanisms: AGGC caused endothelial apoptosis through alterations in molecular chaperones and eIF2α. Solid arrows indicate defined pathways, and dashed arrows indicate speculative pathways.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grants HL-64936 (S. Rounds) and HL-67795 (E. O. Harrington), and ATS/PHA research grant (Q. Lu).