Abstract
It is estimated that more than 40 different lysosomal storage disorders (LSDs) cumulatively affect one in 5000 live births, and in the majority of the LSDs, neurodegeneration is a prominent feature. Neuronal ceroid lipofuscinoses (NCLs), as a group, represent one of the most common (one in 12 500 births) neurodegenerative LSDs. The infantile NCL (INCL) is the most devastating neurodegenerative LSD, which is caused by inactivating mutations in the palmitoyl-protein thioesterase-1 (PPT1) gene. We previously reported that neuronal death by apoptosis in INCL, and in the PPT1-knockout (PPT1-KO) mice that mimic INCL, is at least in part caused by endoplasmic reticulum (ER) and oxidative stresses. In the present study, we sought to determine whether ER and oxidative stresses are unique manifestations of INCL or they are common to both neurodegenerative and non-neurodegenerative LSDs. Unexpectedly, we found that ER and oxidative stresses are common manifestations in cells from both neurodegenerative and non-neurodegenerative LSDs. Moreover, all LSD cells studied show extraordinary sensitivity to brefeldin-A-induced apoptosis, which suggests pre-existing ER stress conditions. Further, we uncovered that chemical disruption of lysosomal homeostasis in normal cells causes ER stress, suggesting a cross-talk between the lysosomes and the ER. Most importantly, we found that chemical chaperones that alleviate ER and oxidative stresses are also cytoprotective in all forms of LSDs studied. We propose that ER and oxidative stresses are common mediators of apoptosis in both neurodegenerative and non-neurodegenerative LSDs and suggest that the beneficial effects of chemical/pharmacological chaperones are exerted, at least in part, by alleviating these stress conditions.
INTRODUCTION
In the majority of more than 40 human lysosomal storage disorders (LSDs), neurodegeneration is a prominent pathological feature ( 1 ). Although apoptotic cell death is prevalent in many LSDs, especially in those that cause neurodegeneration, the molecular mechanism(s) of apoptosis in these diseases remains poorly understood. As a result, although the treatment options for LSDs have significantly increased in recent years, the development of rational and effective treatments for many LSDs awaits advances in our understanding of the molecular mechanism(s) of pathogenesis of these diseases. Recently, we and others have reported that endoplasmic reticulum (ER) and oxidative stresses mediate apoptosis in at least two neurodegenerative LSDs, G M1 gangliosidosis ( 2 ) and infantile neuronal ceroid lipofuscinosis (INCL) ( 3–5 ), caused by the deficiency of acid beta-galactosidase-1 ( 6 ) and palmitoyl-protein thioesterase-1 (PPT1) ( 7 , 8 ), respectively.
Accumulating evidence indicates that the eukaryotic cells have a ‘quality control’ system that scrutinizes whether the proteins in the ER are properly folded ( 9–13 ). Specific proteins such as inositol-requiring enzyme-1 monitor the ER for the balance between unfolded protein load and the capacity of this organelle to deal with that load in order to maintain homeostasis ( 9–13 ). If such efforts fail to achieve a balance, ER stress may activate the unfolded protein response (UPR), a signaling pathway to reduce ER stress ( 10–14 ). However, overwhelming ER stress that cannot be resolved by UPR may lead to programmed cell death or apoptosis ( 11–13 ). Recent reports indicate that an integrated stress response regulates the metabolism of amino acids and confers resistance to oxidative stress ( 14 , 15 ). A number of conditions including the depletion of Ca ++ storage from the ER, glucose starvation and treatment of cells with reducing agents that affect the ER environment activate the stress signaling pathways ( 9–13 ).
It is increasingly evident that ER stress may lead to acute disorders as well as chronic neurodegenerative diseases ( 10 , 16 ). However, it is not clear whether ER and oxidative stresses are unique manifestations of INCL and G M1 gangliosidosis, or these stress conditions are common to both neurodegenerative and non-neurodegenerative LSDs. Moreover, recent reports indicate beneficial effects of chemical/pharmacological chaperone treatment of patients with some LSDs, although the molecular mechanisms by which these agents exert their beneficial effects remain unclear.
In the present study, we used cultured cells derived from patients with neurodegenerative and non-neurodegenerative LSDs to investigate whether ER and oxidative stresses are the unique manifestations of neurodegenerative LSDs and whether the beneficial effects of chemical or pharmacological chaperones are mediated by alleviating these stresses that mediate apoptosis.
RESULTS
ER and oxidative stresses are common manifestations of NCLs
To determine whether ER and oxidative stresses are unique manifestations of INCL or these stress conditions characterize all forms of NCLs ( 17 , 18 ), we used cultured fibroblasts from patients with various forms of NCLs and determined the mRNA levels of several marker genes that are expressed in response to ER and oxidative stresses. We first tested the cultured fibroblasts from patients with the following NCLs: infantile (CLN1), late infantile (CLN2), juvenile (CLN3), variant late infantile (CLN6) and Northern epilepsy (CLN8) (Supplementary Material, Table S1). At least, two normal fibroblast lines were used as controls. To determine whether these cells suffer from ER stress, we quantitated the mRNA levels of ER stress marker genes, ATF6, calnexin, Grp78 and XBP1, by real-time RT–PCR and compared the results with those of the normal control fibroblasts. Unexpectedly, the results show that when compared with the normal control cells (Fig. 1 A, open bar), the mRNA levels of the ER stress marker genes are significantly elevated in virtually all NCL cells used in this study (Fig. 1 A, solid bars). As expected, the mRNA levels of those ER stress marker genes in G M1 gangliosidosis cells are also elevated (Fig. 1 A, solid bars). It should be noted that the mRNA levels of all ER stress marker genes are not equally elevated in cells from all NCL types, although these levels are consistently higher than those in normal control cells.
ER and oxidative stresses in normal, NCL and non-NCL LSD cells. The spontaneous apoptosis was detected in confluent cells after 4 days of growth without medium change. ( A ) The mRNA levels of ER stress marker genes, ATF6, calnexin, Grp78 and XBP1. ( B ) The mRNA levels of oxidative stress marker genes, SOD2, catalase and TTase1; * P < 0.05 and ** P <0.01. ( C ) Western blot analysis of ER and oxidative stress marker genes. At least, two cell lines are used for each LSD in determining the mRNA levels except for CLN8 and G M1 , in which only one cell line was studied. At least, two normal cell lines were used as controls. The fold change in the ratio of P-eIF2α/total-eIF2α is shown both graphically and numerically in (C-III). The cell lines used are listed in Supplementary Material, Table S1. NCL, neuronal ceroid lipofuscinosis; G M1 , gangliosidosis type 1; TSD, Tay-Sachs disease; GD II, Gaucher disease type 2; NPC2, Niemann-Pick C type 2; CTNS, nephropathic cystinosis; GD I, Gaucher disease type 1.
To determine whether the NCL cells also undergo oxidative stress, we determined the mRNA levels of representative oxidative stress marker genes, superoxide dismutase-2 (SOD2), catalase and thioltransferase-1 (TTase1). The results show that when compared with normal control cells (Fig. 1 B, open bars), those from all NCL types show elevated mRNA levels (Fig. 1 B, solid bars), suggesting that these cells are under oxidative stress. Taken together, our results show that all NCL cells tested suffer from ER and oxidative stresses.
Non-NCL LSD cells suffer from ER and oxidative stresses and activate the UPR
To determine whether the ER and oxidative stresses are common to non-NCL neurodegenerative and non-neurodegenerative LSDs, we tested cells from patients with representative non-NCL, neurodegenerative LSDs such as Tay-Sachs disease (TSD), Gaucher disease type II (GD II), Niemann–Pick C type 2 (NPC2) as well as non-neurodegenerative, GD I and nephropathic cystinosis (CTNS) (Supplementary Material, Table S1). Again, quite unexpectedly, our results show that in cells from different types of non-NCL LSDs, the mRNA levels of both the ER (Fig. 1 A, hatched bars) and oxidative stress (Fig. 1 B, hatched bars) marker genes are significantly elevated when compared with those of the normal control cells (Fig. 1 A and B, open bars). We then performed western blot analyses for representative marker proteins of ER (ATF6, Grp78/94) and oxidative (SOD2) stresses in the lysates from various LSD cells. The results show that when compared with normal controls ( n =2), the levels of both the ER (Fig. 1 C-I) and oxidative stress (Fig. 1 C-II) marker proteins are appreciably elevated in all LSD cells. These results show that the non-NCL, neurodegenerative and non-neurodegenerative LSD cells also express both ER and oxidative stress marker genes.
It has been reported that to cope with ER stress, the cells have developed a system in which the general protein production in the ER is minimized in order to increase the production of the chaperone proteins, which can facilitate correct folding of the unfolded or misfolded proteins ( 10 , 19 ). This process is known as UPR, and one of the critical events in this signaling pathway is the inactivation of the translation initiation factor, eIF2α, mediated by its phosphorylation. Accordingly, we determined the levels of phosphorylated-eIF2α (P-eIF2α) in control as well as in LSD cells. We found that when compared with the control cells, those from various types of LSDs show significantly higher levels of P-eIF2α protein (Fig. 1 C-III). We also found that the ratios of P-eIF2α/total eIF2α are elevated in all LSD cells studied (Fig. 1 C-III). More specifically, the ratio of P-eIF2α/total eIF2α was elevated ranging from 2.5- to 3.8-fold for NCLs, from 3.8- to 4.8-fold for non-NCL neurodegenerative LSDs and from 3.3- to 4.3-fold for non-NCL, non-neurodegenerative LSDs (Fig. 1 C-III, bottom row). Taken together, these results suggest that in response to ER stress, the cells from both NCL and non-NCL neurodegenerative and non-neurodegenerative LSDs activate the UPR signaling.
Innate ER and oxidative stresses cause increased apoptosis in LSD cells
As the results discussed earlier indicate that the cells from patients with LSDs are under ER and oxidative stresses, we sought to further confirm this finding by treating these cells with brefeldin-A (BFA), a small hydrophobic compound derived from toxic fungi (reviewed in 20 ), which causes ER stress and apoptosis ( 21 ). We rationalized that if the LSD cells are naturally under pre-existing ER and oxidative stresses, then these cells would be more sensitive to BFA treatment providing strong evidence for the stress load of the LSD cells. Accordingly, we treated the cells derived from normal control, NCLs and non-NCL neurodegenerative and non-neurodegenerative LSDs with varying doses of BFA and first determined cell viability using Trypan blue dye exclusion test. The results show that when compared with the normal control cells, BFA treatment causes a significantly higher level of death in both groups of LSD cells (Fig. 2 A). To determine whether the death of these cells occurred because of apoptosis, we performed annexin-V-FITC staining (Fig. 2 B) and cleaved PARP-1 levels (Fig. 2 C). The results show that when compared with the control cells, those from both NCL- and non-NCL-LSDs are significantly more susceptible to BFA-induced apoptosis in a dose-dependent manner. These results strongly support the notion that cells from patients with both NCL- and non-NCL-LSDs suffer from innate ER stress, and BFA treatment further aggravates the existing stress condition.
Effects of BFA on the cell viability, apoptosis and PARP-1 cleavage. ( A ) Trypan blue dye exclusion test for cell viability. The cells were exposed to the indicated concentrations of BFA for 24 h and the Trypan blue dye exclusion test was performed. The cell death rate was expressed as the percentage of dead cells (mean of three experiments±SE). Note that the LSD fibroblasts are more sensitive to BFA-induced cell death, compared with normal human fibroblasts. ( B ) Detection of early stage apoptosis by fluorescence microscopy. Note that BFA-treated cells show signs of early stage of apoptosis, indicated by annexin V-FITC staining. ( C ) Detection of cleaved PARP-1 in cells treated with BFA by western blot analysis. The cell lines used are normal control, INCL (CLN1), CLN3, CLN8, G M1 , GD II, CTNS and GD I (see Supplementary Material, Table S1 for details).
Disruption of lysosomal homeostasis in normal cells causes ER stress
How might an abnormality in lysosomes lead to ER stress? Although we do not have a clear answer to this question, it has been reported that intraventricular administration of leupeptin, a lysosomal protease inhibitor, causes accumulation of lipofuscin-like intracytoplasmic granules in neurons of the hippocampal dentate gyrus, similar to the ones seen in NCLs ( 22 ). Moreover, in mice, targeted disruption of the gene-encoding cathepsin D, a lysosomal hydrolase, causes NCL-like lysosomal storage ( 23 ). Although our results show that ER and oxidative stresses cause apoptosis in LSDs, to our knowledge, a clear link between lysosomal dysfunction and ER stress has not been established. Therefore, to understand whether disruption of lysosomal homeostasis may cause ER stress, we treated normal control fibroblasts with ammonium chloride, which is known to disrupt lysosomal homeostasis by altering the pH in this organelle, and determined the levels of mRNAs as well as the ER stress marker proteins. The results show that the disruption of lysosomal homeostasis in normal cells causes elevated expression of ER and oxidative stress marker (Supplementary Material, Fig. S1A–C). These results also suggest a cross-talk between the lysosomes and the ER.
Chemical chaperones alleviate ER and oxidative stresses in LSD cells
It has been reported that chemical and/or pharmacological chaperones, a group of small-molecular-weight compounds that stabilize the conformation of proteins, increase the protein folding capacity of the ER and facilitate the trafficking of mutant proteins ( 24 ). Examples of chemical chaperones include 4-phenyl butyric acid, trimethylamine N -oxide dihydride (TMAO) and dimethyl sulfoxide. Similar to these compounds, endogenous bile acids and derivatives such as ursodeoxycholic acid as well as its taurine-conjugated derivative, tauroursodeoxycholic acid (TUDCA), are strong modulators of ER function ( 25 ). Moreover, it has been recently reported that these chaperones are capable of reducing ER stress and restoring glucose homeostasis in a mouse model of type 2 diabetes ( 26 ). Further, chemical chaperones appear to correct the lysosomal storage in Fabry's disease, also an LSD ( 27 ). Accordingly, we first treated representative NCL- and non-NCL-LSD cells and determined the mRNA levels of ER stress marker genes, ATF6, calnexin, Grp78 and XBP-1 and compared the results with those from untreated control cells. Recently, it has been reported that HERPUD1 (homocysteine-responsive ER-resident ubiquitin-like domain member-1) is also an excellent indicator of ER stress. Thus, we also determined the mRNA levels of HERPUD1 in chemical chaperone treated and untreated control cells. The results show that when compared with the untreated controls, the expression levels of all the ER stress markers, including HERPUD1, in chemical chaperone-treated LSD cells are markedly lower (Fig. 3 A–C) (Supplementary Material, Fig. S2).
Chemical chaperones alleviate ER and oxidative stresses. The mRNA levels of ER and oxidative stress markers, ATF6, calnexin, Grp78, XBP1, catalase, SOD2 and TTase1 in LSD fibroblasts. ( A and D ) INCL (CLN1); ( B and E ) TSD; ( C and F ) CTNS. Semi-confluent (70–80%) cells were treated with 150 m m TMAO or 0.2 m m TUDCA for 14 h before the cells were harvested for RNA extraction and real-time RT–PCR analysis. The mRNA levels in parallel experiments without the drugs were used as controls. Data are presented as the mean values of three sets of experiments±SEM. * P < 0.05.
We then sought to determine whether the treatment of the cells with chemical chaperones alters the oxidative stress levels. Accordingly, we determined the mRNA levels of catalase, SOD2 and TTase1 in untreated and chemical chaperone-treated three representative LSD cell types, INCL, TSD and CTNS. The results show that in INCL cells, TMAO treatment reduces the level of catalase-mRNA, whereas TUDCA treatment appears to cause no significant change (Fig. 3 D, left panel). In contrast, SOD2-mRNA level remained virtually unaltered when the cells are treated with TMAO or TUDCA (Fig. 3 D, middle panel). However, TTase1-mRNA level was markedly reduced when these cells were treated with TUDCA, but not with TMAO (Fig. 3 D, right panel). Interestingly, none of the chemical chaperones appears to suppress the expression of all three oxidative stress markers in TSD cells (Fig. 3 E). Although in CTNS cells, both TMAO and TUDCA are effective in suppressing the catalase and TTase1-mRNA levels (Fig. 3 F, left and right panels); SOD2-mRNA level was suppressed by TMAO only (Fig. 3 F, middle panel). Taken together, these results show that chemical chaperones are efficient suppressors of oxidative stress markers depending on the type of LSD cells used.
Chemical chaperones reduce the level of phosphorylated eIF2α
It has been reported that in response to ER stress, UPR signaling suppresses the translation of housekeeping genes, facilitating the production of selected chaperone proteins, which help proper folding of the misfolded proteins. If the chaperone proteins fail to correct the folding, then these misfolded proteins are channelized to the ER degradation pathway. This response is considered to be a homeostatic mechanism for survival of cells that are under ER stress ( 9–15 ). Translational initiation is inhibited by the phosphorylation of translation initiation factor, eIF2α. Thus, we sought to determine whether chemical chaperone treatment of the LSD cells reduces the level of P-eIF2α. Accordingly, we pre-treated the INCL, TSD and CTNS cells with TMAO or TUDCA for 14 h followed by BFA treatment to induce ER stress and apoptosis. We then determined the levels of both phosphorylated and total eIF2α by western blot analysis. In addition, we determined the levels of ER stress markers, Grp78/94 and apoptosis marker, cleaved (active) caspase-3. The results indicate that in each of these cell lines, TMAO and TUDCA treatments appreciably reduce the levels of P-eIF2α as well as cleaved caspase-3, a biochemical marker of apoptosis (Fig. 4 A). Notably, the ratio between P-eIF2α and total eIF2α is also markedly decreased in cells treated with the chemical chaperones, suggesting the alleviation of ER stress.
Effects of chemical chaperones on BFA-induced ER stress and apoptosis. ( A ) Western blot analyses of ER stress and apoptosis markers. The bar graph shows the ratio of P-eIF2α/total eIF2α quantified by densitometric analysis, indicating decreased phosphorylation on serine 51 of eIF2α. The levels of Grp78/94 and cleaved caspase-3 are decreased, suggesting that TMAO and TUDCA pre-treatments alleviate the BFA-induced ER stress and apoptosis in the treated cells. ( B – D ) Effects of chemical chaperone pre-treatment on LSD cell apoptosis, monitored by annexin V-FITC staining. Cell nuclei were stained with DAPI to indicate cell density. Cell lines used were (B), INCL (CLN1); (C), TSD and (D) CTNS.
Protective role of chemical chaperones against apoptosis
Previous reports using cells from G M1 gangliosidosis and INCL as well as PPT1-KO mice that mimic INCL have shown that ER stress activates the UPR signaling leading to apoptosis ( 2–5 ). Thus, we sought to determine whether the suppression of ER stress by chemical cheparones also reduces the levels of apoptotic cell death. Accordingly, we compared the levels of apoptotic cells in untreated and chaperone-treated LSD cells representing both neurodegenerative and non-neurodegenerative LSDs. The results show that treatment of both the NCL and non-NCL LSD cells with chemical chaperones markedly reduces the levels of apoptotic cells (Fig. 4 B–D). These results suggest that chemical chaperone treatment not only suppresses the expression of ER and oxidative stress markers, but also is effective in reducing the levels of apoptotic cell death in both NCL and non-NCL LSDs.
DISCUSSION
In this study, we have demonstrated that cell death in both neurodegenerative and non-neurodegenerative LSDs by apoptosis is mediated by ER and oxidative stresses. We also demonstrated that chemical disruption of lysosomal homeostasis induces ER stress in normal cells, suggesting the existence of a cross-talk between the lysosomes and the ER. These results clearly demonstrate a link between the lysosomes and the ER and show that abnormality in one organelle can adversely affect the other. Most importantly, we demonstrated that alleviation of ER and oxidative stresses by chemical chaperones protects the LSD cells from apoptosis.
More than 40 human LSDs have been characterized ( 28 ), and the phenotypic manifestations of these diseases include neurologic and skeletal abnormalities, enlargement of affected organs and pre-mature death ( 29 ), which are thought to be due to lysosomal storage of undegraded substrates. Several therapeutic approaches including inhibition of substrate production and the replacement of the defective enzyme are currently used ( 30 , 31 ). The results of our present study demonstrate for the first time that ER and oxidative stresses are common manifestations of both neurodegenerative and non-neurodegenerative LSDs and that chemical/ pharmacological chaperones ( 24 ) can alleviate these stresses as well as protect the cells from undergoing apoptosis. Currently, potential treatment modalities in LSDs include enzyme replacement therapy, substrate reduction therapy and chaperone therapy ( 31 ). As the ER and oxidative stresses are deleterious to the cells, small molecules that alleviate these stresses may have beneficial effects on all LSDs when used in conjunction with the treatment regimens currently available. Thus, chemical chaperones provide an attractive therapeutic approach for these diseases. Recently, novel methods for screening small molecular weight compounds as chemical chaperones have been reported for GD ( 32 , 33 ). It is likely that such techniques will be applied to identify effective chemical chaperones for other LSDs.
MATERIALS AND METHODS
Cell lines and detection of ER and oxidative stresses
A list of the cell lines used in this study is provided in Supplementary Material, Table S1. To determine the levels of baseline oxidative and ER stresses, exponentially growing cell cultures (70–80% confluent) were trypsinized and plated into six-well plates with a seeding density of 1–1.2×10 4 cells/cm 2 . The cells were allowed to grow for 2–3 days until they reached 100% confluence at which time the medium was changed. The cells were allowed to grow for an additional 4 days without medium change, and the cells were used for the detection of ER and oxidative stresses using real-time RT–PCR and western blot analyses.
BFA treatment
Details are provided in Supplementary Material and Materials and Methods.
Cell viability test
Cell viability was determined by the Trypan blue dye exclusion test, as described in details in Supplementary Material and Materials and Methods.
Induction of ER and oxidative stresses in normal cells by ammonium chloride treatment
Details are provided in Supplementary Material and Material and Methods.
Treatment of cells with chemical chaperones
The chemical chaperone drugs, TMAO (Acros Organics) and TUDCA (Calbiochem), were used to examine their effects on ER and oxidative stresses. Detailed procedures were provided in Supplementary Material and Materials and Methods.
Real-time RT–PCR
The primers for the real-time RT–PCR are listed in Supplementary Material, Table S2. All reactions were performed in triplicate and repeated in at least two independent experiments using real-time RT–PCR, as previously described ( 4 , 5 ).
Western blot analysis
Cell lysates were prepared using PhosphoSafe extraction reagent (EMD Biosciences). Western blot analyses were performed according to the procedure described previously ( 4 , 5 ). The primary antibodies used in the present study are anti-β-actin (1:2000; US Biological), anti-ATF6 (1:300; Imgenex), anti-eIF2α and P-eIF2α (1:1000 and 1:750, respectively; Cell Signaling Technology) and anti-SOD2 (1:1000; Santa Cruz Biotechnology). The sources for other primary antibodies (i.e. anti-caspase-3, anti-Grp78 and anti-PARP-1) and those of the secondary antibodies are as previously described ( 4 ).
Determination of apoptosis by annexin V staining
Details are provided in Supplementary Material and Materials and Methods.
SUPPLEMENTARY MATERIAL
Supplementary Material is available at HMG Online.
FUNDING
This research was supported in whole by the Intramural Research Program of the NICHD/NIH.
ACKNOWLEDGEMENTS
We thank Drs J.Y. Chou, I. Owens and S.W. Levin for critical review of the manuscript and helpful suggestions.
Conflict of Interest statement . None declared.
