Involvement of Endoplasmic Reticulum Stress in Hereditary Tyrosinemia Type I*

Hereditary tyrosinemia type I (HTI) is the most severe disease of the tyrosine degradation pathway. HTI is caused by a deficiency of fumarylacetoacetate hydrolase (FAH), the enzyme responsible for the hydrolysis of fumarylacetoacetate (FAA). As a result, there is an accumulation of metabolites such as maleylacetoacetate, succinylacetone, and FAA. The latter was shown to display mutagenic, cytostatic, and apoptogenic activities and to cause chromosomal instability. Herein, we demonstrate that FAA also causes a cellular insult leading to the endoplasmic reticulum (ER) stress signaling. Treatment of V79 Chinese hamster lung cells with an apoptogenic dose of FAA (100 μm) causes an early induction of the ER resident chaperone GRP78/BiP and a simultaneous phosphorylation of the eIF2α. FAA treatment also causes a subsequent induction of the proapoptotic CHOP (CEBP homologous protein) transcription factor as well as a late activation of caspase-12. Data obtained from fah–/– mice taken off the therapeutic 2-(2-nitro-4-trifluoromethylbenzoyl)-1,3 cyclohexanedione drug are similar. However, in this mouse model, there is also an increase in proteasome activity indicative of ER-associated degradation. This difference observed between the two models may be due to the fact that the murine model measures the effects of all metabolites accumulating in hereditary tyrosinemia type I as opposed to the cellular model that only measures the effects of exogenous FAA.

Hereditary tyrosinemia type I (HTI) 2 is a severe metabolic disease affecting mainly the liver and renal functions of patients inflicted with this disorder. The disease is caused by a deficiency of the last enzyme involved in the metabolic breakdown of tyrosine, fumarylacetoacetate hydrolase (FAH) (1)(2)(3)(4)(5). As a result, there is an accumulation of metabolites upstream of FAH, such as succinylacetone, maleylacetoacetate, and fumarylacetoacetate (FAA). Although the molecular basis of the disease has been extensively studied, much remains to be learned concerning the cellular events underlying the pathology of the disease (progressive liver damage, liver failure, cirrhosis, hepatocarcinoma, tubular renal dysfunction, and porphyria-like crisis). Several studies have shown that FAA is cytotoxic to cells. FAA is also shown to be mutagenic and to cause chromosomal instability as well as lead to cell cycle arrest and apoptosis (6 -9). Jorquera and Tanguay (8) recently observed a disruption of the Golgi apparatus following the treatment of mammalian cells with FAA. An increase in intracellular calcium levels of 60% was observed in HeLa human cells and a 25% increase was observed in V79 Chinese hamster lung cells treated with exogenous FAA (8). Because cellular calcium homeostasis is tightly regulated by the endoplasmic reticulum and this cellular compartment is closely related to another organelle of the secretory pathway, the Golgi apparatus, we decided to investigate the possibility that FAA also causes a deleterious effect on the endoplasmic reticulum function.
The ER is involved in many essential cellular functions such as protein folding, modification, and assembly, as well as lipids and sterols synthesis. The ER is also a major signal-transducing organelle sensitive to changes in cell homeostasis (e.g. calcium levels) and cellular stresses, which include glucose deprivation, mutations in the genes of secreted proteins, viral infection, heme deficiency, and oxidative stress (reviewed in Ref. 10). The ER stress can also be induced by the cytotoxic agents that disrupt protein folding or calcium homeostasis. In response to the ER stress and protein misfolding, cells activate a signal-transducing pathway known as the unfolded protein response (UPR) in an effort to deal with the protein load in the organelle (reviewed in Ref. 11). This response is mediated by kinases IRE1 and PERK, and by the activating transcription factor 6, all of which are regulated by the ER resident GRP78/BiP chaperone levels. The ER stress leads to an induction of these transducers that will result in the translational attenuation through eIF2␣ phosphorylation by PERK kinase (12), transcriptional activation of protein chaperones and folding catalyst genes (13), and the ER-associated degradation (ERAD) of misfolded proteins by the proteasome (14). Sustained ER stress may cause cells to undergo apoptosis. An ER-specific caspase activated by prolonged ER stress is caspase-12. This caspase may be activated by two different means. First, Ca 2ϩ release from the ER to the cytosol may activate calpain protease in the cytosol, which may cleave the procaspase-12 zymogen to produce its active form (15). Activated caspase-12, in return, initiates a caspase cascade leading to caspase-3 activation (16,17). Another means by which procaspase-12 may be activated is through the sequestration of tumor necrosis factor receptor-associated factor 2 by IRE1 enabling the clustering of the procaspase (18). Procaspase-12 may then be cleaved by caspase-7 (19). It is noteworthy that the ER stress may trigger both the extrinsic and intrinsic pathways of apoptosis. In fact, Ca 2ϩ released from the ER may be taken up by the mitochondria (20), where it can cause loss of the membrane potential of the organelle.
Herein, we show that exogenous FAA or conditions leading to the HTI phenotype in vivo cause a cellular insult that elicits the ER stress response as demonstrated by an early induction of GRP78/BiP and eIF2␣ phosphorylation, an increase in proteasome activity, and late acti-vations of CHOP and caspase-12. We also show increased levels of stress-related proteins, heme oxygenase-1 (HO-1), Hsp27, and Hsp70, in vitro and in a murine model of HTI.

EXPERIMENTAL PROCEDURES
Cell Culture and Treatments-Chinese hamster V79 lung cells (from the American Type Culture Collection) were grown in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 5% fetal bovine serum (Medicorp) at 37°C with 5% CO 2 . Cells were plated at a density of 2.5 ϫ 10 4 cells/ml in 15-cm flasks the day before treatment or at 1 ϫ 10 5 cells/ml in 35-mm culture dishes the day before transfection (luciferase assay). FAA, an accumulating metabolite in HTI, was synthesized as described by Edwards and Knox (21) from a porcine homogentisate dioxygenase extract using homogentisic acid (Sigma) as a precursor substrate for the reaction. Cells were treated with FAA (100 or 200 M) in Hanks' balanced salt solution (HBSS, Invitrogen) for 2 h and then left to recover in fresh medium for the indicated time periods. Cell treatment with the N-linked glycosylation inhibitor, tunicamycin (5 M, Sigma), was performed in regular culture medium for 1 or 16 h.
Luciferase Activity Assay-Transfection of cells was carried out by the calcium phosphate method using the CalPhos mammalian transfection kit (Clontech, BD Biosciences). Briefly, V79 cells were transfected with 5 g of firefly luciferase reporter plasmids pGL3-promoter vector (Promega) or pGL3-GRP78P(-132)-luc vector (kindly provided by K. Mori, Graduate School of Biostudies, Kyoto, Japan) carrying the ER stress response element (ERSE) and 100 ng of the reference vector pRL-CMV (Promega) carrying the Renilla luciferase gene in 35-mm culture dishes. Cells were then treated with FAA (200 M) in HBSS for 1 h or with tunicamycin (5 M) for 16 h in regular medium prior to harvest 48 h following the transfection. Cells were collected in 200 l of Passive Lysis Buffer (Promega). Both firefly and Renilla luciferase (internal control) activities of cell lysates (40 l) were measured using the dual luciferase reporter assay system (Promega) on a 1250 luminometer (LKB Wallac). The relative luciferase activity was defined as the ratio of firefly luciferase to Renilla luciferase activity following normalization of the experimental activity of the reporter with the internal control to minimize variability caused by transfection efficiency.
Isolation of Liver Proteins-Frozen liver tissues of NTBC-treated and untreated mice were homogenized on ice in a phosphate-buffered saline solution (pH 7.3) containing 1 mM phenylmethylsulfonyl fluoride using a Teflon pestle. The homogenized extracts were centrifuged at 10,000 ϫ g for 30 min at 4°C. Protein concentrations of the collected supernatants were determined using the Bio-Rad protein assay according to the manufacturer's recommendations.
Western Blot Analysis-The protein extract samples were resolved by electrophoresis on a 12% SDS-PAGE gel, transferred onto a nitrocellulose blotting membrane (BioTrace NT, Pall Life Sciences) for Western blot analysis (23), and visualized by the chemiluminescent detection method (Western lightning chemiluminescence reagent, PerkinElmer Life Sciences). Antibodies against KDEL peptide (see Fig. 2A), phosphospecific eIF2␣, HO-1, and protein disulfide isomerase (PDI) were purchased from Stressgen. The monoclonal antibody raised against CHOP was obtained from Santa Cruz Biotechnologies. The procaspase-12 and procaspase-7 antibodies were purchased from Calbiochem. The GRP78/BiP antibody (see Fig. 4A) was kindly provided by L. Hendershot (St. Jude Children's Research Hospital), and the Hsp27 and HC8 antibodies were a gift from J. Landry (Hôtel-Dieu Research Center, CHUQ, Quebec). The Hsp70 antibody was described in Tanguay et al. (24). The ␤-actin antibody was purchased from NeoMarkers.
Proteasome Activity Measurements-Frozen liver tissues of NTBCtreated and untreated mice were homogenized on ice in a Tris-HCl (0.1 mol/liter, pH 7.5) and sucrose (0.25 mol/liter), protease inhibitor-free buffer solution, using a Teflon pestle. The homogenized extracts were centrifuged at 10,000 ϫ g for a period of 30 min at 4°C. Protein concentrations of the collected supernatants were determined as described above. Assays were carried out using 50 g of liver homogenate proteins diluted in a Tris-HCl buffer (0.1 mol/liter, pH 7.5) to a final volume of 100 l, containing the succinyl-LLVY-7-amino-4-methylcoumarin amide fluorogenic substrate (40 M) (Boston Biochem), which measures the chymotrypsin-like activity of the 20 S proteasome. Assays were performed at 37°C with shaking. Fluorescence was measured at 1-min intervals at excitation and emission wavelengths of 390 and 460 nm, respectively, on an Ascent FL Fluoroskan (Thermo Labsystems). Lactacystin (Boston Biochem), an irreversible inhibitor of the chymotrypsinand trypsin-like activities of the proteasome, was added at a concentration of 25 M prior to the measurements.
Statistical Analyses-Statistical analyses were carried out following the methods outlined in Ref. 25, except when otherwise mentioned. Homogeneity of variances was verified using an F max test (26). Accordingly, comparisons between the two samples were done using Student's t test or its nonparametric equivalent. A p value comparison (a posteriori) was done using the Bonferroni sequential technique (27). Data are presented as mean Ϯ S.D. p values less than 0.05 were deemed statistically significant.

RESULTS
FAA Treatment of V79 Cells Induces the Transcriptional Activity of the ERSE-Previous work done in our laboratory showed that in addition to causing Golgi disruption, exogenous FAA (100 M) also caused an increase in intracellular Ca 2ϩ levels in V79 Chinese hamster lung cells (8). These observations prompted us to examine the effects of exogenous FAA on the ER function. Among the early events to occur upon the ER stress is the induction of the GRP78/BiP chaperone, the key regulator protein of the UPR of the ER stress, which is transcriptionally regulated. Transcriptional regulation of GRP78/BiP is mediated by the binding of transcription factors to the ERSE, a cis-acting element present in the promoter region of GRP78/BiP and other ER chaperones such as GRP94 and calreticulin (28). To determine whether FAA induced an ER stress response in V79 cells, we measured the transcriptional activity of the ERSE in V79 cells transfected with a luciferase reporter vector carrying the ERSE upstream of the SV40 promoter following FAA treatment.
As can be seen in Fig. 1, FAA (200 M) treatment of V79-pGL3-GRP78P(-132)-luc vector-transfected cells causes an early (1 h) (ϳ 2.1fold) increase in relative luciferase activity compared with the control cells. Tunicamycin, an inhibitor of protein N-linked glycosylation that induces the UPR as a result of the ER stress, resulted in a ϳ3.6-fold increase in relative luciferase activity compared with the control cells. These results show that FAA induces the ERSE transcriptional activity and the ER chaperone GRP78/BiP, which is the key regulator of the UPR pathway.
Exogenous FAA Elicits the ER Stress Response Pathway in V79 Cells-Next, we examined the effects of exogenous FAA on the ER stressregulated proteins (Fig. 2). V79 cells were treated for 2 h with FAA (100 M) in HBSS and left to recover in fresh medium for the indicated time periods. Fig. 2A shows an early induction of GRP78/BiP (from 1 to 4 h following the treatment) accompanied by a simultaneous eIF2␣ phosphorylation which is at a maximum at 1 h post-treatment with an increase in its phosphorylation state compared to control cells (Fig. 2B). The expression of the oxidoreductase PDI was also analyzed. PDI is involved in the rearrangement of incorrect protein disulfides (reviewed in Refs. 29 and 30) and has been reported to be induced by the UPR. However, no induction of the PDI was observed ( Fig. 2A). Likewise, there was no visible induction of the HC8 protein from the ␣-subunit of the 20 S proteasome suggestive of the ERAD. FAA treatment also resulted in late activation of the proapoptotic CHOP transcription factor ( Fig. 2C) with peak expression levels at 10 and 22 h following the treatment. Another late event was the activation of the ER resident caspase-12 (Fig. 2C). At 22 h post-treatment, the level of uncleaved or inactive procaspase-12 was approximately one-third of the control, a level similar to that of the cells treated with tunicamycin, a potent inducer of the ER stress which also causes apoptosis (31,32). There was no evidence of ER stress associated with HBSS alone for the recovery periods studied (data not shown). Taken together, these results indicate that V79 cells exposed to FAA respond by activating the ER stress response pathway leading to apoptosis signaling. A late caspase-3-dependent cell death of FAA-treated V79 cells was previously demonstrated (7).
Because ER stress may be induced by different cellular stress stimuli including oxidative stress and previous studies suggested that oxidative stress may occur in HTI (33,34), we also examined the effects of FAA on other cellular stress proteins that may confer protection against the oxidative stress. FAA treatment of V79 cells caused a slight and late (10 and 22 h) induction of HO-1, a marker of oxidative stress (Fig. 3). Both the 70-kDa and 27-kDa heat shock proteins (Hsp70 and Hsp27, respectively) were also slightly induced following the FAA treatment. Their induction occurred late (10 -22 h post-treatment) compared with the induction of the proximal ER stress proteins (1-4 h post-treatment) (Fig. 3).

Retrieval from NTBC Treatment of fah Ϫ/Ϫ Mice Induces an ER Stress
Response in Liver Cells-Because the assays described so far involved an in vitro cellular model using exogenous FAA as the sole metabolite tested, we next used a HTI murine model (22) to investigate the implication of the ER stress response in vivo. To mimic the HTI phenotype, we used fah knock-out mice taken off the therapeutic NTBC drug used in human HTI patients. Homogenates of liver tissue obtained from mice taken off NTBC for up to 5 weeks were used to determine whether the ER stress response was elicited in vivo as well. Fig. 4A shows an induction of GRP78/BiP from 3 to 5 weeks following NTBC withdrawal.   There was a slight elevation of the eIF2␣ phosphorylation in liver samples from mice recently taken off NTBC followed by a pronounced increase in phosphorylation at 5 weeks. In contrast to FAA-treated V79 cells, livers of mice taken off NTBC showed an induction of PDI after NTBC withdrawal. We next examined the expression levels of proapoptotic proteins induced by the ER stress. As shown in Fig. 4B, there is an induction of CHOP following NTBC withdrawal. Further evidence supporting the activation of apoptotic signaling in the liver of untreated fah Ϫ/Ϫ mice is the reduced levels of the uncleaved forms of procaspase-7 and -12 as early as 1 week (procaspase-7) or 2 weeks (procaspase-12) after discontinuing NTBC treatment (Fig. 4B).
fah Ϫ/Ϫ Mice Taken Off NTBC Treatment Display Increased Levels of Proteasome Activity in Liver Cells-Because the results obtained from fah Ϫ/Ϫ mice taken off NTBC treatment suggested that hepatic cells from these mice were subjected to the ER stress and that the ER stress may lead to the ERAD, we measured the proteasome activity of liver homogenates from these mice. Data from Fig. 5A demonstrate that NTBC withdrawal caused an ϳ2-fold increase in proteasome activity as compared with wild-type mice. Furthermore, the proteasome activity measured in wild-type mice was not significantly different from that measured in NTBC-treated fah Ϫ/Ϫ mice despite increased levels of HC8 expression in these mice (Fig. 5B). Lactacystin, an inhibitor of the chymotrypsin-and trypsin-like activities of the proteasome, was used as a control for activity measurements. Fig. 5B shows that the HC8 protein of the 20 S proteasome is only slightly induced in the liver of fah Ϫ/Ϫ mice taken off NTBC treatment. Interestingly, fah Ϫ/Ϫ mice treated with NTBC also showed some increase in the levels of HC8 expression although higher levels of HC8 expression were observed progressively from 3 to 5 weeks after NTBC treatment was discontinued. The origin of the minor band appearing below HC8 in fah Ϫ/Ϫ NTBC-off mice (Fig.  5B) is unknown. However, because the 20 S proteasome HC8 subunit is reported to possess two phosphorylation sites (35), the minor band below HC8 may represent its dephosphorylated form. Therefore, mice with the tyrosinemic phenotype seem to undergo ERAD as a result of the ER stress.
NTBC Treatment Withdrawal in fah Ϫ/Ϫ Mice Induces Cell Stress Proteins in Liver Cells-Results from FAA-treated V79 cells showed that FAA induces different stress proteins such as HO-1, Hsp70, and Hsp27. This led us to examine if these proteins were also induced in fah Ϫ/Ϫ mice removed from NTBC treatment. Results shown in Fig. 6 demonstrate that there is an induction of Hsp27 in mice taken off NTBC for the period of 2 to 5 weeks following NTBC withdrawal. The 70-kDa heat shock protein is also induced in these mice with a maximum induction at 2 weeks. However, even fah Ϫ/Ϫ mice under NTBC treatment show a slight induction of Hsp70. HO-1 is also induced as soon as 1 week following NTBC withdrawal, implying that FAH deficiency may cause oxidative stress.

DISCUSSION
We previously reported that FAA displayed mutagenic and apoptogenic activities in vitro (6,7). In addition, FAA causes genetic instability and Golgi apparatus disruption (8). We now report that FAA elicits an ER stress response.
Results obtained from V79 cells treated with exogenous FAA support the hypothesis that FAA causes a disruption of the ER function. The first evidence of the ER stress in these cells is the early activation of GRP78/ BiP, a proximal signaling protein, functioning as the key regulator of the three transducers of the UPR. Phosphorylation of eIF2␣ inhibits general protein synthesis by preventing the assembly of the 80 S ribosome. Since eIF2␣ is a direct substrate of the PERK transducer kinase, it is phosphorylated at an early stage of the ER stress response. eIF2␣ phosphorylation and translational repression were reported to occur as early as 30 min following exposure to a stress-inducing stimulus (36). However, phosphorylation of eIF2␣ can be induced by many other cellular stresses such as heme deficiency, viral infection, and oxidative stress (reviewed in Ref. 37). In fact, the response elicited by these stimuli is referred to as the integrated stress response by some authors. This response converges to induce the ER stress response pathway by cross-talk between the UPR and integrated stress response pathways although the molecule involved in the cross-talk has not yet been identified (38). Our results show an early eIF2␣ phosphorylation, which is concomitant to the induction of GRP78/BiP. Among the late events after FAA treatment of V79 cells is the induction of CHOP, a proapoptotic transcription factor reported to be induced by the ER stress (39), and the activation of the ER-specific caspase-12 (reviewed in Ref. 40). Thus it appears that FAA accumulation causes an induction of the ER stress response that is ensued by apoptosis.
Further evidences that FAH deficiency can lead to the ER stress are the results obtained in a murine model of HTI. Indeed, NTBC withdrawal in these mice causes the induction of both GRPs, GRP78/BiP and PDI (a catalyst of protein folding in the ER), and suggests that some protein misfolding does occur in the ER of HTI mice. This is explained by the fact that maleylacetoacetate and FAA possess electrophilic properties that may interfere with sulfhydryl reactions by forming adducts with glutathione and other sulfhydryl groups (4). In fact, we previously showed that FAA depleted cells of GSH (6). Inhibition of translation is also observed in mice taken off NTBC treatment as evidenced by the eIF2␣ phosphorylation. Our results show that some ER apoptosis signaling occurs in HTI mice through the activation of caspase-7 and its target, caspase-12. As opposed to the treatment of V79 cells with exogenous FAA, proteins in the liver of HTI mice are subject to ERAD as demonstrated by the induction of the HC8 ␣-subunit protein of the proteasome and ϳ2-fold increase in proteasome activity. This type of degradation is an essential response to alleviate the stress caused by an accumulation of misfolded proteins.
We also analyzed the expression of other stress proteins in FAAtreated cells and in the livers of HTI mice. One such protein was HO-1, which was induced in both the models. HO-1 is usually associated with oxidative stress. HO-1 promoter activity may be stimulated by the bind-ing of Nrf2 transcription factor to the antioxidant response element. Nrf2 promotes the transcription of phase II genes of xenobiotic metabolism such as thiol reacting substances (41)(42)(43)(44). In fact, some of the genes activated by Nrf2 including HO-1, glutathione S-transferase, and NAD(P)H:quinone oxidoreductase were up-regulated in HTI mice (33,34,45). However, a recent study showed that the ER stress stimulates HO-1 promoter activity through the antioxidant response element in vascular smooth muscle cells (46). Thus, HO-1 induction in HTI may be a consequence of the ER stress. The induction of CHOP may also be responsible for HO-1 induction through activation of the ER oxidase gene ERO1 (47). ERO1 causes an accumulation of reactive oxygen species in stressed vertebrate cells (48).
An induction of Hsp70 and Hsp27 was also observed in the cell and murine models. Noteworthy was the observation of a visible increase in the expression of Hsp70 in the liver of NTBC-treated fah Ϫ/Ϫ mice. This may be due to the fact that NTBC treatment does not completely inhibit the hydroxyphenylpyruvate dioxygenase enzyme of the tyrosine catabolic pathway, allowing some accumulation of the toxic metabolites causing the stress. Heat shock proteins are reported to play an important role in oxidative stress protection (49). In fact, Yan et al. (49) reported that the transcription of several heat shock proteins through the Hsf1 promoter is essential for renal homeostasis and protects kidney cells against oxidative stress. Therefore, the increase in both Hsp70 and Hsp27 expression in FAA-treated cells and in HTI mice may be the result of oxidative stress stimuli. Nonetheless, the protection conferred FIGURE 5. Proteasome activity in liver homogenates of fah ؊/؊ mice following NTBC withdrawal. A, the proteasome activity was measured using 50 g of 10% liver homogenate from wildtype, fah Ϫ/Ϫ NTBC-treated, or fah Ϫ/Ϫ untreated mice. Values for the proteasome activities, which are significantly different from their lactacystintreated counterparts, are illustrated by an asterisk on top of columns on the chart (p Ͼ 0.05). The proteasome activity of the liver homogenates from fah Ϫ/Ϫ mice taken off NTBC treatment was significantly different from the activity measured in wild-type mice (p Ͼ 0.05) (represented by two asterisks on the chart). B, fah Ϫ/Ϫ mice or control mice (first lane) were withdrawn from NTBC treatment, and livers obtained from the mice killed at 1-5 weeks post-NTBC withdrawal were collected and homogenized in a phosphate buffer. Liver proteins (20 g) were loaded on a SDS-PAGE gel and immunoblotted with an antibody raised against the HC8 20 S proteasome subunit. FIGURE 6. Expression of cell stress proteins in liver homogenates of fah ؊/؊ mice taken off NTBC. Wild-type (first lane) mice or fah Ϫ/Ϫ mice withdrawn from NTBC treatment for 1-5 weeks were killed, and their livers were collected and homogenized in a phosphate buffer. Liver proteins (20 g) were loaded on a SDS-PAGE gel and immunoblotted with antibodies specific to HO-1, Hsp70, and Hsp27. by these heat shock proteins in response to apoptotic stimuli is also documented (reviewed in Ref. 50). Thus, we cannot dismiss the possibility that Hsp70 and Hsp27 inductions are part of an anti-apoptotic response. In fact, the cleavage of procaspase-7 suggests that some cell death is occurring as early as 1 week following NTBC withdrawal in fah Ϫ/Ϫ mice. Recently, several studies showed evidence of a connection between Hsps and the protein quality control system in mammalian cells. Parcellier et al. (51) reported that Hsp27 binds to polyubiquitinated proteins and the 26 S proteasome, and causes an increase in phosphorylated IB␣ degradation. Ito et al. (52) recently found that the ER stress caused by tunicamycin or thapsigargin induces early phosphorylation of Hsp27 in U373MG, U251MG, and HeLa cells. In this study, sustained tunicamycin treatment caused a second rise in Hsp27 phosphorylation and accumulation of Hsp27 in aggresomes of U251MG cells. Another function of Hsp27, which may be relevant to this study, is its involvement in proteasome regulation; phosphorylated small heat shock proteins were recruited to protein aggresomes in muscle atrophy (53).
Because the mechanism of FAA cell toxicity is poorly understood and some reports suggest that oxidative stress may occur in HTI (33,34), it remains unclear if HO-1 and Hsps are induced as a consequence of the ER stress response or rather by an additional stress such as oxidative stress initiated by FAA.
In summary, both exogenous FAA and the induction of the HTI phenotype by withdrawal of NTBC treatment in fah Ϫ/Ϫ mice have a deleterious effect on the ER function. However, the stress response observed in the HTI murine model seems more severe as it induces a greater number of stress-related proteins. Another difference between the in vitro and in vivo models is the presence of ERAD. A possible explanation is that the in vivo model of FAH deficiency measures the effects not only of FAA but also of the other metabolites accumulating in FAH deficiency such as maleylacetoacetate and succinylacetone. Because liver cells possess a high secretory capacity, the ER stress observed in fah Ϫ/Ϫ liver cells may have important implications in the disease. Further work will be necessary to evaluate the implication of different HTI metabolites on the induction of the ER stress response, notably concerning the proximal kinases involved in the eIF2␣ phosphorylation activated by the metabolites. This should allow a better comprehension of the cellular events underlying the pathology of HTI.