Address correspondence and reprint requests to Dr Jeffrey N. Keller, 205 Sanders-Brown Building, University of Kentucky, Lexington, KY 40536-0230, USA. E-mail: Jnkell0@pop.uky.edu
Increasing evidence suggests that proteasome inhibition plays a causal role in promoting the neurodegeneration and neuron death observed in multiple disorders, including Alzheimer's disease (AD) and Parkinson's disease (PD). The ability of severe and acute inhibition of proteasome function to induce neuron death and neuropathology similar to that observed in AD and PD is well documented. However, at present the effects of chronic low-level proteasome inhibition on neural homeostasis has not been elucidated. In order to determine the effects of chronic low-level proteasome inhibition on neural homeostasis, we conducted studies in individual colonies of neural SH-SY5Y cells that were isolated following continual exposure to low concentrations (100 nm) of the proteasome inhibitor MG115. Clonal cell lines appeared morphologically similar to control cultures but exhibited significantly different rates of both proliferation and differentiation. Elevated levels of protein oxidation and protein insolubility were observed in clonal cell lines, with all clonal cell lines being more resistant to neural death induced by serum withdrawal and oxidative stress. Interestingly, clonal cell lines demonstrated evidence for increased macroautophagy, suggesting that chronic low-level proteasome inhibition may cause an excessive activation of the lysosomal system. Taken together, these data indicate that chronic low-level proteasome inhibition has multiple effects on neural homeostasis, and suggests that studying the effects of chronic low-level proteasome inhibition may be useful in understanding the relationship between protein oxidation, protein insolubility, proteasome function, macroautophagy and neural viability in AD and PD.
The proteasome is a large intracellular protease that is responsible for the degradation of most misfolded, oxidized, and aggregated proteins (Goldberg et al. 1997; Kopito and Sitia 2000; Davies 2001; Ulrich 2002). Inhibition of proteasome activity has been demonstrated to occur in AD and PD (Keller et al. 2000; Lopez-Salon et al. 2000; McNaught and Jenner 2001; McNaught et al. 2003), with numerous in vitro and in vivo studies documenting the ability of proteasome inhibitors to induce neuropathology and neuron death similar to what is observed in AD and PD (Keller and Markesbery 2000; Qiu et al. 2000; Lee et al. 2001; Rideout et al. 2001; McNaught et al. 2002). Taken together, these data are consistent with proteasome inhibition playing a direct role in the neurodegenerative process in both AD and PD. Although these in vitro and in vivo studies have been very informative in elucidating the effects of severe and acute proteasome inhibition, at present the effects of chronic low-level proteasome inhibition on neural homeostasis has not been elucidated. Understanding the effects of chronic low-level proteasome inhibition may be particularly important in understanding the role of proteasome inhibition in age-related disorders such as AD and PD, where neuropathology and neuron death may require decades to become manifest.
In the present study we analyzed the effects of chronic low-level proteasome inhibition on neural homeostasis. Cumulatively, these data indicate that chronic low-level proteasome inhibition results in elevated levels of protein oxidation, elevated levels of protein aggregation, and activation of macroautophagy. Continued analysis of the effects of chronic low-level proteasome inhibition may therefore be useful in elucidating the role of proteasome inhibition in AD and PD.
Materials and methods
All cell culture medium, serum, and antibiotics were purchased from Gibco Life Technologies (Gaithersburg, MD, USA). Protein oxidation assay kit was purchased from Intergen (Ann Arbor, MI, USA), the ubiquitin antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA), and the proteasome substrates were purchased from Bachem (Torrance, CA, USA). The MG115 was purchased from Calbiochem (San Diego, CA, USA), and [3H]leucine purchased from NEN-Life Science Products (Boston, MA, USA). All other reagents and chemicals were purchased from Sigma Chemical (St. Louis, MO, USA).
Establishment of clonal lines
Neural SH-SY5Y cells were obtained from the American Tissue Culture Collection and propagated as described previously (Ding and Keller 2001b). In order to establish individual clonal lines following chronic proteasome inhibition neural SH-SY5Y cells were maintained in normal growth medium containing 100 nm MG115. The medium was replaced weekly, with fresh MG115 added each week. Isolated colonies were evident after 12 weeks, and were individually selected for further propagation using ceramic cloning rings. Individual clonal lines were propagated in a 75-cm2 flask and utilized for all experiments described. Control cultures consisted of sister SH-SY5Y cells that were propagated alongside with MG115-treated cultures for the duration of selection period. Cells of fewer than 25 passages were utilized for all described studies.
Analysis of proteasome activity
Proteasome activity was determined as described previously (Ding and Keller 2001b; Ding et al. 2002). Briefly, cell lysates were collected and protein aliquots (1 µg/µL) generated in proteasome activity buffer (10 mmol/L Tris-HCl (pH 7.8), 1 mmol/L EDTA, 0.5 mmol/L dithiothreitol, and 5 mmol/L MgCl2). A standard curve of free 7-amido-4-methylcoumarin was utilized for quantification of proteasome activity. Fluorescence was monitored at 340 nm excitation and 440 nm emission.
Analysis of lysosomal activity
Lysosomal protein degradation was measured as described previously (Cuervo and Dice 1996; Martin et al. 2003). Briefly, confluent cells were labeled with 2.5 µCi/mL [3H]leucine for 48 h and, after extensive washes, the cells were plated in medium containing an excess (2.8 mm) of unlabeled leucine. Aliquots of the medium were taken at the indicated times and precipitated in acid. The radioactivity in the acid-soluble (amino acids) and acid-precipitable (proteins and peptides) fractions was converted to disintegrations per minute (d.p.m.) in a WinSpectral 1414 liquid scintillation analyzer (NEN-Life Sciences), by correcting for quenching using an external standard. Proteolysis was calculated as the amount of acid-precipitable radioactivity transformed to acid-soluble at each time. Where indicated, cells were supplemented with 15 mm ammonium chloride (NH4Cl) or 10 mm 3-methyladenine (3MA) at the beginning of the chase period. Analysis of autophagic, monodansylcadaverine (MDC)-labeled vesicles was conducted as described previously (Larsen et al. 2002). Briefly, cells were incubated for 1 h with 50 µm MDC and then washed. Cells were visualized using fluorescence microscopy, with random 100 × fields utilized to determine the presence of MDC-labeled vesicles.
Analysis of protein oxidation, protein aggregation, and protein ubiquitination
The levels of protein oxidation were determined by dot blot analysis using an oxidized protein detection kit as described previously (Lauderback et al. 2001; Aksenova et al. 2002). This assay is based on the immunochemical detection of protein carbonyl groups derivatized with 2,4-dinitrophenyl hydrazine, according to manufacturer's instructions. All samples were performed in duplicate. The amount of protein aggregation was determined as described previously (Kawarabayashi et al. 2001). Briefly, 1 mg aliquots of cell lysates were prepared in phosphate-buffered saline containing protease inhibitor cocktail (Sigma Chemical). Cells were lyzed by 10 consecutive freeze–thaw cycles, and brought to a final volume of 250 µL containing a final concentration of 1% sodium dodecyl sulfate (v/v). The lysate was centrifuged for 1 h at 100 000 × g at 4°C, and an aliquot of the protein pellet (sodium dodecyl sulfate-insoluble protein pellet) taken for protein determination. The remaining pellet was sonicated and resuspended in 70% formic acid and centrifuged for 1 h at 100 000 × g at 4°C. Protein ubiquitination was determined by western blot analysis as described previously (Ding et al. 2002), using 50 µg of total cell lysate protein and a commercially available ubiquitin antibody (Santa Cruz).
Analysis of cell proliferation, neural differentiation, and neural survival
Cell proliferation was determined by trypsinization of cells, and conducting actual cell counts, 48 h following serum stimulation. Cultures were seeded at equal density 5 × 104 onto 35-mm plastic culture dishes, and synchronized in 1% serum-containing medium 24 h prior to serum stimulation. At least 10 dishes from two separate experiments were utilized for all studies. For neural differentiation studies, cells were seeded at equal density 5 × 104 onto 35-mm plastic culture dishes, and treated with 25 µm retinoic acid in 1% serum-containing medium. The percentage of cells with neurites was quantified in random 32 × magnification fields, with at least 50 cells counted for each individual dish. A neurite was defined as a cellular process of at least two cell diameters in length. At least 10 dishes from two separate experiments were utilized for all studies. Cell viability was determined by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) conversion as described previously (Holtsberg et al. 1998; Keller et al. 1998). At least 12 cultures from two separate experiments were utilized for each time point. Results from MTT were confirmed using Hoechst 33258 staining as described previously (Holtsberg et al. 1998; Ding and Keller 2001b).
Statistical significance was determined using Student's t-test, with a p-value of < 0.05 required for significance.
Clonal cell lines exhibit altered levels of cellular proliferation and differentiation
In order to elucidate the effects of chronic low-level proteasome inhibition on neural homeostasis we conducted studies in which neural SH-SY5Y cells were continually exposed to low levels of the proteasome inhibitor MG115 (100 nm). Additional cultures of SH-SY5Y cells not exposed to MG115 were propagated and passaged alongside MG115-treated cultures and utilized as controls for each experiment. Exposure to MG115 (100 nm) resulted in an 8% inhibition of chymotrypsin-like proteasome activity that was not toxic during the first 24 h of treatment (Fig. 1). Similar results were obtained with analysis of trypsin-like and postglutamyl peptidase proteasome activity (data not shown).
Following a 12-week exposure to MG115, a total of six individual clonal cell lines were isolated. All of the clonal cell lines appeared to be morphologically similar to control cells (data not shown), possessing a similar size and shape, but proliferated at different rates (Fig. 2a). Clones 1 and 6 exhibited a statistically significant decrease in proliferation, whereas clone 4 exhibited a statistically significant increase in cellular proliferation, as compared to control cell lines (Fig. 2a). Under non-differentiated conditions, clone 6 exhibited a significantly higher level of neuritic processes as compared to all other cell lines, and differentiated to a similar degree as control cells following retinoic acid administration (Fig. 2b). In contrast, clones 1–5 exhibited significantly lower levels of differentiation following administration of retinoic acid (Fig. 2b).
Clonal neural cells possess increased levels of ubiquitin and oxidized proteins
Because the proteasome inhibition is believed to play an important role in promoting increased levels of ubiquitinated proteins and oxidized proteins observed in AD and PD, we next conducted studies to quantify the amount of protein ubquitination and protein oxidation in the individual clonal cell lines. Western blot analysis revealed that the gross amount of protein ubiquitination was similar in all cells (Fig. 3a). However, clonal cell lines did exhibit increased levels of free ubiquitin (∼9 kDa band) and increased levels of a ubiquitinated protein of ∼100 kDa (Fig. 3a). Quantification of protein carbonyls, an established marker of protein oxidation, revealed a statistically significant elevation in protein oxidation in each of the clonal cell lines as compared to control cultures (Fig. 3b). Identical results were obtained with western blot analysis of 4-hydroxynonenal-modified proteins (data not shown), consistent with clonal cell lines possessing increased levels of proteins that possess oxidative damage.
Clonal cells exhibit increased levels of insoluble and aggregated protein
In order to determine if chronic low-level proteasome inhibition altered the amount of intracellular protein solubility or protein aggregation, we conducted studies to determine the percentage of sodium dodecyl sulfate-insoluble protein in clonal cell lines and control cells. Previous studies have demonstrated that the sodium dodecyl sulfate-insoluble protein pool contains primarily aggregated proteins (Campbell et al. 2000; Nagao et al. 2000). Analysis revealed that clone 6 cells possessed a statistically significant elevation in the amount of protein that was sodium dodecyl sulfate-insoluble (Fig. 4), with 4.8% of total protein present as a sodium dodecyl sulfate-insoluble protein pellet. Subsequent analysis of sodium dodecyl sulfate-insoluble protein revealed that 0.3% of total protein present in clone 6 was formic acid-insoluble (Fig. 4), suggesting the presence of highly cross-linked protein aggregates in clone 6 cells. Identical results were obtained in the other clonal cell lines that were examined (data not shown).
Clonal cells are more resistant to cell stress
Because the clonal cell lines exhibited increased levels of protein oxidation and protein aggregation, we next sought to elucidate if clonal cell lines were more vulnerable to subsequent stress. Surprisingly, 24 h following serum withdrawal (Fig. 5a) or exposure to hydrogen peroxide (50 µm) (Fig. 5b), clonal cell lines were observed to exhibit enhanced viability, as determined by quantification of MTT conversion. Identical results were obtained using Hoechst 33258 analysis of nuclear morphology to determine neural viability (data not shown), confirming that clonal cell lines were more resistant to cellular injury.
Clonal cell lines exhibit enhanced macroautophagy and altered lysosomal activity following stress
To determine the effect of chronic low-level proteasome inhibitor treatment on the rate of protein degradation by the lysosomal system, we analyzed the breakdown of long-half-life proteins as described previously (Cuervo and Dice 1996; Martin et al. 2003). Analysis of lysosomal activity under basal conditions (serum present), revealed no significant difference between the rates of protein degradation in clone 6 cells and in control cultures (Fig. 6a). These data also indicate that the concentrations of MG115 utilized for cellular selection in the present study do not indirectly inhibit lysosomal activity. Following cellular stress (serum removal), lysosomal activity was increased, as expected, in control cells. In contrast to control cells, clone 6 cells exhibited a significantly reduced lysosomal activation following cellular stress (Fig. 6b). Identical results were obtained in clone 5 (data not shown).
To further analyze the contribution of the lysosomal system to the bulk of long-lived protein degradation, we performed similar experiments in the presence of NH4Cl, a classical inhibitor of the proteolytic activity of lysosomes. Under basal conditions (serum present), we found that NH4Cl exerted a similar amount of inhibition in control and clone 6 cells (Fig. 7a). In contrast to basal conditions, following cellular stress (serum removal), clone 6 cells were observed to be less responsive to NH4Cl inhibition (Fig. 7a). These results strongly indicate that chronic low-level proteasome inhibition is capable of causing a defect in the normal activation of the lysosomal system.
To determine whether or not the contribution of different lysosomal pathways to the total rates of lysosome-dependent protein degradation was altered in the clonal cell lines, we conducted studies utilizing 3MA. This inhibitor, at the concentrations used in this study, has been shown to efficiently block degradation of proteins by macroautophagy (Holen et al. 1995). Under basal conditions, macroautophagy contributed less than 12% to the breakdown of long-lived proteins in control cells (Fig. 7b). In clonal cells the inhibitory effect of 3MA, and consequently the contribution of macroautophagy to total protein breakdown, was significantly higher (Fig. 7b). In agreement our findings describing the inability of clonal cells to increase their rates of protein degradation following cellular stress (Fig. 6b), clonal cells displayed a lower level of 3MA inhibition following stress (Fig. 7b).
To further elucidate the presence of macroautophagy in control and clonal cells we conducted studies using MDC, which has been utilized previously to visualize macroautophagic compartments in cultured cells (Larsen et al. 2002). Analysis of control cells revealed that under basal conditions control cells possessed diffuse cytoplasmic MDC staining, with intensely labeled MDC vesicles observed in control cells only following cellular stress (Fig. 7c). In contrast to control cells, clone 6 cells exhibited numerous intensely labeled MDC vesicles under basal conditions, with the number of MDC-labeled vesicles apparently increasing in number following cellular stress (Fig. 7d).
Proteasome inhibition has been proposed to directly contribute to the neuropathology and neuron death that occurs in several neurodegenerative conditions including AD and PD (McNaught et al. 2001; Ding et al. 2002; Hyun et al. 2002). Such a belief is based largely upon the observations that proteasome activity is inhibited in AD and PD brain tissue (Keller et al. 2000; Lopez-Salon et al. 2000; McNaught et al. 2001, 2003), and the fact that severe and acute inhibition of proteasome activity is capable of recapitulating neuropathology and neuron death in vitro and in vivo similar to that observed in AD and PD (Keller and Markesbery 2000; Pasquini et al. 2000; Qiu et al. 2000; Lee et al. 2001; Rideout et al. 2001, 2003; McNaught et al. 2002). While it is highly likely that proteasome inhibition plays a causal role in AD and PD, several important questions remain over the direct role severe and acute proteasome inhibition plays in mediating neuropathology and neuron death observed in each of these disorders. For example, the neuropathology and neurotoxicty of severe proteasome inhibition in vitro and in vivo is very rapid. The increases in protein aggregation and protein oxidation observed in AD and PD would be expected to occur over a prolonged time-frame of decades, and as such, the contribution of proteasome inhibition to this process would be expected to be mediated by a chronic low-level proteasome inhibition that does not result in acute neuron death. The ability of severe proteasome inhibition to rapidly induce neuron death suggests that it may play a direct role in causing the neuron death observed in AD and PD, possibly serving as a penultimate step in the neuron death process. We therefore hypothesize that chronic low-level proteasome inhibition is likely responsible for mediating deleterious neurochemical alterations, in particular the elevated levels of protein oxidation and protein aggregation observed in AD and PD, with toxic levels of proteasome inhibition directly contributing to the actual neural death observed in those conditions.
Despite the tremendous progress made in understanding the effects of lethal and acute proteasome inhibition, little is currently known about the effects of chronic low-level proteasome inhibition on neural homeostasis. In the present study we sought to begin establishing the effects of chronic low-level proteasome inhibition on neural homeostasis, utilizing clonal cell lines of neural SH-SY5Y cells that were derived following a 12-week exposure to low-level proteasome inhibition. Chronic low-level proteasome inhibition was observed to have dramatic and diverse effects on both neural cell proliferation and neural differentiation. These data suggest that for some neurophysiological aspects, the same degree of chronic low-level proteasome inhibition is capable of producing more than one neurophysiological outcome. As such, these data indicate that even very similar neuron populations within the brain may have dramatically different outcomes following similar degrees of low-level proteasome inhibition. Such differences in outcome may contribute to the selectivity of neuron death that is observed in AD and PD, although this has yet to be determined. Previous studies have demonstrated that the proteasome may have an important role in regulating neural cell cycle components (Azuma-Hara et al. 1999; Boutillier et al. 1999; Naujokat and Hoffman 2002), with proteasome regulation of cell cycle playing a direct role in mediating neural death (Rideout et al. 2003). In future studies, it will be important to identify potential alterations in cell-cycle components within clonal cell lines that have been chronically treated with low concentrations of MG115.
All cells exposed to chronic low-level proteasome inhibition exhibited enhanced levels of protein oxidation and protein aggregation. Surprisingly, these same cells exhibited enhanced survival following administration of two different neural stressors. Taken together, these data indicate that the elevations in protein oxidation and protein aggregation following low-level proteasome inhibition are not necessarily deleterious to neural viability. Recent studies have demonstrated that specific forms of protein aggregation, referred to as aggresomes, may mediate cytoprotection (Kopito 2000; Taylor et al. 2003). In future studies it will be important to determine if chronic proteasome inhibition can result in the formation of aggresomes in neural cells.
Increased levels of protein oxidation and protein aggregation occurred in the absence of detectable increases in ubiquitinated proteins. Previous studies have demonstrated that proteasome-mediated degradation of oxidized proteins does not require protein ubiquitination (Shringarpure et al. 2003). These data suggest that low-level proteasome inhibition may preferentially increase intracellular protein aggregation and protein oxidation in a manner that is ubiquitin independent. All clonal cell lines exhibited enhanced levels of free ubiquitin, which is presumably mediated through the enhanced expression of ubiquitin, with previous studies documenting elevated levels of free ubiquitin expression in neurons following a variety of stressors (Massa et al. 1996). In future studies it will be important to identify the proteins which are aggregated and oxidized following low-level proteasome inhibition.
Evidence for alterations in the lysosomal system, in the form of increased number of autophagic features, are observed in both AD and PD brain (Anglade et al. 1997; Nixon et al. 2000; Bahr and Bendiske 2002; Larsen and Sulzer 2002), although the mechanisms responsible for mediating these alterations have not been elucidated previously. It is interesting to note that previous reports have indicated that elevations in lysosomal components selectively occur in the neuron populations most vulnerable to neurodegeneration (Cataldo et al. 1996). Our results demonstrate for the first time that chronic low-level proteasome inhibition induces macroautophagy in neural cells. As such, chronic low-level proteasome inhibition may directly contribute to the manifestation of macroautophagy in the AD and PD brain. It is likely that the macroautophagy observed in clonal cell lines occurs as a means of guaranteeing that the sufficient rates of protein degradation, at least the degradation of long-lived proteins, continues to occur. Chronic low-level proteasome inhibition appeared to significantly inhibit the ability of neural cells to further increase lysosomal-mediated proteolysis in response to stress. Following serum deprivation, macroautophagy and other lysosomal pathways of protein degradation become maximally activated (Cuervo and Dice 1996), with clonal cell lines in the present study exhibiting an inability to further stimulate lysosomal activity over basal levels. The inability of the clonal cell lines to increase their rates of protein degradation in response to stress could be explained if the macroautophagic system is already maximally activated under basal conditions. Such an hypothesis is consistent with the observation that clonal cell lines exhibited a higher content of autophagic vesicles under basal conditions. However, we cannot discard some degree of compromise in the macroautophagic process under stress conditions, since despite the detected increase in the number of autophagic vesicles in the clonal cells in response to serum removal, rates of protein degradation remain unchanged. The function of other nutrient-activated lysosomal pathways, such as chaperone-mediated autophagy (Cuervo and Dice 1996), may also be deleteriously affected in clonal cell lines and thereby contribute to the observed impairment in lysosomal plasticity. This possibility is supported by the fact that following serum withdrawal the inhibitory effect of NH4Cl, which blocks all lysosomal pathways by inhibiting the degradation of the proteins inside all lysosomal compartments, was significantly lower in the clonal cell lines. In future studies, it will be important to fully elucidate the effects of chronic low-level proteasome inhibition on the expression and function of all lysosomal components.
It is almost certain that proteasome inhibition plays a role in mediating protein and lysosomal alterations observed in both AD and PD. Data from the present study indicate that the relationship between protein aggregation, protein oxidation, proteasome activity, macroautophagy, and neural viability may be more complex than previously anticipated. Additionally, these data suggest that the utilization of our model of chronic low-level proteasome inhibition may provide a useful tool in developing a more complete understanding of this complex research area.
The authors wish to thank Dr William R Markesbery for his continual support. The present work was supported by the American Heart Association (JNK), and grants from the National Institutes of Health [AG018437 (JNK), AG005119 (JNK), AG021904 (AMC)] and from the Ellison Medical Foundation [AG-NS-0163 (AMC)].