Impaired Mitochondrial Dynamics and Nrf2 Signaling Contribute to Compromised Responses to Oxidative Stress in Striatal Cells Expressing Full-Length Mutant Huntingtin

Huntington disease (HD) is an inherited neurodegenerative disease resulting from an abnormal expansion of polyglutamine in huntingtin (Htt). Compromised oxidative stress defense systems have emerged as a contributing factor to the pathogenesis of HD. Indeed activation of the Nrf2 pathway, which plays a prominent role in mediating antioxidant responses, has been considered as a therapeutic strategy for the treatment of HD. Given the fact that there is an interrelationship between impairments in mitochondrial dynamics and increased oxidative stress, in this present study we examined the effect of mutant Htt (mHtt) on these two parameters. STHdhQ111/Q111 cells, striatal cells expressing mHtt, display more fragmented mitochondria compared to STHdhQ7/Q7 cells, striatal cells expressing wild type Htt, concurrent with alterations in the expression levels of Drp1 and Opa1, key regulators of mitochondrial fission and fusion, respectively. Studies of mitochondrial dynamics using cell fusion and mitochondrial targeted photo-switchable Dendra revealed that mitochondrial fusion is significantly decreased in STHdhQ111/Q111 cells. Oxidative stress leads to dramatic increases in the number of STHdhQ111/Q111 cells containing swollen mitochondria, while STHdhQ7/Q7 cells just show increases in the number of fragmented mitochondria. mHtt expression results in reduced activity of Nrf2, and activation of the Nrf2 pathway by the oxidant tBHQ is significantly impaired in STHdhQ111/Q111 cells. Nrf2 expression does not differ between the two cell types, but STHdhQ111/Q111 cells show reduced expression of Keap1 and p62, key modulators of Nrf2 signaling. In addition, STHdhQ111/Q111 cells exhibit increases in autophagy, whereas the basal level of autophagy activation is low in STHdhQ7/Q7 cells. These results suggest that mHtt disrupts Nrf2 signaling which contributes to impaired mitochondrial dynamics and may enhance susceptibility to oxidative stress in STHdhQ111/Q111 cells.


Introduction
Huntington disease (HD) is a devastating inherited neurodegenerative disease caused by a CAG trinucleotide repeat expansion in exon 1 of huntingtin (Htt) gene. Although the mechanisms by which mutant huntingtin (mHtt) causes neurotoxicity have been widely studied, the pathological processes have not yet been fully elucidated. mHtt-induced impairment of the cellular responses to oxidative stress has been suggested as a crucial contributing factor in the progression of HD. Indeed, there is clear evidence of increased oxidative stress in HD. Further, defects in mitochondria, which are both a source of oxidative stress and a target, are apparent in HD and HD models [1,2].
Neurons are highly dependent on mitochondria for not only energy production but also Ca 2+ buffering, and reactive oxygen species (ROS) regulation. Studies using mouse and cell models for HD have shown mitochondrial impairment and bioenergetic deficits, reminiscent of the pathological characteristics of HD [2][3][4]. In addition, increased oxidative stress is apparent in HD cases [5][6][7]. Mitochondria from STHdh Q111/Q111 cells show impaired function and significant increases in ROS production compared to STHdh Q7/Q7 cells [8][9][10]. Our previous studies showed that the PPARc pathway, which regulates the expression of metabolically important genes, is severely compromised in STHdh Q111/Q111 cells which coincides with an increased sensitivity to thapsigargin induced loss of mitochondrial membrane potential (DYm), and increased cell death at higher concentrations of thapsigargin. Activation of the PPARc pathway attenuated thapsigargininduced DYm loss and cell death in STHdh Q111/Q111 cells [8,11].
Mitochondria are dynamic organelles that are constantly undergoing fission and fusion, which is essential for normal cellular function. Imbalances between mitochondria fission and fusion have been shown to negatively impact the physiology and viability of neuronal cells [12,13]. Key mediators of mitochondrial fission/fusion include the GTPases Dynamin-Related protein 1 (Drp1), which is essential for fission, and Optic Atrophy Type 1 (Opa1) and the Mitofusins (Mfn1 and Mfn2) which mediate fusion. mHtt has been reported to directly bind Drp1 and increase its activity, suggesting that this may contribute to the apparent mitochondrial fragmentation and dysfunction [14]. Indeed, fragmentation of mitochondria results in increased ROS production in cell models [15,16]. Mutations of Opa1 result in autosomal dominant optic atrophy [17]. In addition, the levels of Opa1, as well as Mfn1/2 were shown to be decreased in HD cases relative to controls [18]. There is also data to indicate that alterations in mitochondria morphology in HD enhance cellular susceptibility to apoptosis [14,19].
In the present study, we analyzed mitochondrial dynamics and how they are affected by oxidative stress in a striatal cell model of HD. The cells used in this study are homozygous for either the wild type or mutant Htt gene. Although HD is an autosomal dominant disease, the effects of mHtt are often studied in cell and animal models that are homozygous for mHtt to facilitate the dissection of the pathogenic mechanisms involved [1][2][3][4]. Although these knock-in models are imperfect, they are advantageous in that full-length mHtt, rather than truncated versions of the protein, is expressed and thus may be more informative about disease mechanisms in humans. Nonetheless the fact that they express two copies of the mutant allele, rather than just one as in HD, needs to be carefully considered when evaluating the data obtained with these models.
In this study we found that STHdh Q111/Q111 cells exhibit alterations in mitochondrial dynamics compared to STHdh Q7/Q7 cells. Determination of static mitochondrial morphology along with analysis of dynamic mitochondrial fusion processes revealed that STHdh Q111/Q111 cells exhibit an increase in fragmented mitochondria concomitant with a decrease in fusion. Further STHdh Q111/Q111 cells show impairment of the mitochondrial response to oxidative stress. Given that Nrf2 is not only a key regulator for antioxidant system but also an emerging target to counteract mitochondrial dysfunction in many disease models [25,26,[32][33][34], we investigated whether Nrf2 signaling was compromised in this cell model of HD and found that the presence of mHtt results in defects in Nrf2 signaling. Significant impairment in oxidative stress-induced Nrf2 activation was observed in STHdh Q111/Q111 cells. Our study provides important insights into mHtt-induced mitochondrial impairment, with activation of Nrf2 signaling being a potential therapeutic strategy.

STHdh Q111/Q111 Cells Show Altered Mitochondrial Dynamics
Mitochondrial morphology is an important component of cellular metabolism and pathogenic processes. Mitochondria are dynamic and display a range of morphologies. Maintaining the appropriate proportion of each mitochondrial morphological state is essential for appropriate functioning of the cell. To gain insight into mitochondrial dynamics in mHtt expressing cells, we transfected mitochondria-targeted GFP into STHdh Q7/Q7 and STHdh Q111/Q111 cells. Cells were divided into four groups based on the morphology of the mitochondria: tubular, fragmented, mixed, and swollen as previously described [35]. In most cells of both genotypes the mitochondria (about 60%) exhibited mixed forms with predominantly short tubular shapes. Approximately 20% of STHdh Q7/Q7 cells had predominantly tubular mitochondria while less than 10% of STHdh Q111/Q111 cells showed this mitochondrial morphology. Interestingly, more than 30% of the STHdh Q111/Q111 cells presented with fragmented mitochondria which was significantly greater than the STHdh Q7/Q7 cells. Both cell types exhibited very few cells with swollen mitochondria (Fig. 1A). Similarly, a significant increase in fragmented neuronal mitochondria in the striatum of R6/2 mice was also observed ( Figure S1). Next, we examined whether there was a difference in the expression of mitochondrial fusion or fission regulators between the two cell types. Immunoblot analyses revealed that Drp1 levels were significantly lower in STHdh Q111/Q111 cells than in STHdh Q7/Q7 cells. In addition Drp1 phosphorylated at serine 616, which is a more active form [36,37], was also lower in STHdh Q111/Q111 cells compared to SThdh Q7/Q7 cells. Of interest, Opa1 expression was also significantly decreased in STHdh Q111/ Q111 cells, while Mfn2 appears to be at a similar level between the two cell types (Fig. 1, B and C). Together these results indicate that proteins involved in the dynamics of mitochondrial fission/fusion are altered in STHdh Q111/Q111 cells.

Mitochondrial Fusion is Impaired in STHdh Q111/Q111 Cells
To gain further insight into dynamic changes in mitochondrial morphology, we investigated whether mitochondrial fusion processes were altered in STHdh Q111/Q111 cells compared to STHdh Q7/Q7 cells. To this end, striatal cells were fused using PEG after transfection with either mitoCherry or mitoGFP as described in materials and methods. Cell fusion allows two individual cells to share cellular components including mitochondria. If a mitochondrion expressing mitoCherry undergoes fusion with a mitochondrion expressing mitoGFP, the resulting fused mitochondrion becomes yellow. STHdh Q111/Q111 cells exhibited significantly lower levels of fused mitochondria compared to STHdh Q7/Q7 cells (Fig. 2, A and B). Since this cell fusion assay requires the use of cycloheximide to prevent de novo protein synthesis which would generate mitochondria containing both mitoCherry and mitoGFP without the fusion process, it can be speculated that inhibition of protein synthesis might differentially affect mitochondrial fusion process in the two cell types. Therefore we used a second method to monitor mitochondrial fusion, a photoconvertible GFP (Dendra) targeted to mitochondria. Dendra, which in the unconverted state fluoresces green, was stably converted in a given area of the cell by light illumination at 400 nm and thus showed red fluorescence. When mitochondria in the cell emitting red fluorescence fuse with mitochondria emitting green fluorescence, the fluorescence becomes yellow. Using this approach, more mitochondrial fusion events were again observed in the STHdh Q7/Q7 cells compared to STHdh Q111/Q111 cells (Fig. 2, C and D). Although neither method estimates the mitochondrial fission process, these results provide strong evidence that there is reduction in mitochondrial fusion events in STHdh Q111/Q111 cells.

Oxidative Stress Induces Differential Responses in Mitochondria Morphology
It has been shown that mitochondria dynamically change their morphologies in response to oxidative stress [38][39][40]. Therefore, we tested whether the two striatal cell lines showed different pattern changes in mitochondrial morphology in response to H 2 O 2 . In both cell types there were dramatic reductions in tubular and mixed forms of mitochondria in response to H 2 O 2 in a dose dependent manner (Fig. 3). Importantly, STHdh Q111/Q111 cells showed significant increases in swollen mitochondria while the percentage of cells with fragmented mitochondria slightly decreased as the concentration of H 2 O 2 was increased (Fig. 3). In contrast, STHdh Q7/Q7 cells showed a rapid increase in the population of fragmented mitochondria while swollen mitochondria only slightly increased with the increasing concentrations of H 2 O 2 . It is noteworthy that swollen mitochondria are shown to be functionally deficient and indicative of pathogenesis [41,42]. Therefore, this result suggests that mitochondria in STHdh Q111/Q111 cells show greater susceptibility to oxidative stress compared to STHdh Q7/Q7 cells.

Mitochondria in STHdh Q111/Q111 Cells are More Oxidized
Mitochondria play important roles in regulation of cellular redox state and defense against oxidative stress. In addition, our results show that oxidative stress dramatically affects mitochondrial morphologies. Given these data, we next determined if the two striatal cell displayed differences in in the redox states of Figure 1. Mitochondrial dynamics is altered in STHdh Q111/Q111 cells. A, Mitochondria-targeted GFP (mitoGFP), was transfected in the striatal cells to visualize the mitochondria and cells were then assigned to one of four groups based on the morphology of their mitochondria (tubular, mixed, fragmented or swollen). STHdh Q111/Q111 cells had more fragmented mitochondria and less tubular mitochondria compared to STHdh Q7/Q7 cells, while mixed and swollen mitochondria were at similar levels in the two cell types. More than 200 cells of each cell type were assessed for each independent experiment. n = 3 B, Immunoblots revealed that STHdh Q111/Q111 cells exhibit reduced levels of Drp1, its active form which is phosphorylated at Ser616, and Opa1 compared to STHdh Q7/Q7 cells, whereas Mfn2 appears to be similar in both cell types. C, Quantitative analysis of immunoblots shows a significant reduction in Drp1, phosphorylated Drp1, and Opa1. n = 4. Data shown are mean 6 SE. **P,0.01, ***P,0.001 vs. STHdh Q7/Q7 . doi:10.1371/journal.pone.0057932.g001 mitochondria in the basal condition, which could ultimately contribute to different distributions in mitochondrial morphologies. To this end, mitochondrial targeted GFP redox sensors, rosGFP1 or rosGFP2 were transfected into striatal cells [43]. The fluorescence emission at 525 nm was monitored at excitation wavelengths of 470 nm and 400 nm and the ratio of emission intensities (470/400) was used to evaluate mitochondrial matrix redox state as described in methods and materials. The results from these two redox sensors indicate that mitochondria in STHdh Q111/Q111 cells are in a more oxidized state compared to STHdh Q7/Q7 cells (Fig. 4). Mitochondrial fusion is significantly reduced in STHdh Q111/Q111 cells. A, Striatal cells expressing mitoCherry or mitoGFP were fused and mitochondrial fusion was analyzed as described in materials and methods. STHdh Q111/Q111 cells exhibit substantially lower levels of mitochondrial fusion compared to STHdh Q7/Q7 cells. B, Quantitative analysis reveals that STHdh Q7/Q7 cells exhibit ,60% mitochondrial fusion on average while STHdh Q111/Q111 cells display ,20% fused mitochondria. The data was acquired by analysis of more than 60 cells of each cell type from seven independent experiments. The white dashed line indicates the contact region between the two cells. C, To monitor mitochondrial fusion without cell fusion, mitoDendra was expressed in the striatal cells. The mitochondrial fusion process was observed after the photoconversion of green to red fluorescence in a small population of mitochondria as described in methods and materials. The arrow head indicates the yellowish mitochondria formed by the fusion between red and green mitochondria. D, The quantitative data shows that the mitochondrial fusion rate in STHdh Q111/Q111 cells is significantly reduced compared to STHdh Q7/Q7 cells. The data was acquired by analysis of more than 20 cells of each cell type from five independent experiments. Data shown are mean 6 SE. **P,0.01, ***P,0.001 vs. STHdh Q7/Q7 . doi:10.1371/journal.pone.0057932.g002 Nrf2 Signaling Pathway is Impaired in STHdh Q111/Q111 Cells Nrf2 is a key player in redox regulation and defense mechanisms against ROS [44,45]. Since STHdh Q111/Q111 cells show altered mitochondrial dynamics and increased susceptibility to oxidative stress, we investigated whether the Nrf2 signaling pathway was compromised in STHdh Q111/Q111 cells. Nrf2 activity was first assessed using an ARE luciferase reporter. STHdh Q111/ Q111 cells exhibited significantly reduced ARE activity at the basal level compared to STHdh Q7/Q7 cells (Fig. 5A). Next, we tested whether the Nrf2 signaling pathway was compromised in STHdh Q111/Q111 cells. The overexpression of Nrf2 resulted in a significantly greater increase in ARE activity in STHdh Q7/Q7 cells compared to STHdh Q111/Q111 cells, indicating an impairment in the Nrf2 signaling pathway in STHdh Q111/Q111 cells (Fig. 5B). To confirm that the impaired Nrf2 signaling is due to the presence of mHtt, we measured ARE activity in additional striatal cell lines. Two new striatal cell lines expressing mHtt (1A and 6L) exhibited a substantial reduction in ARE activity in the absence or the presence of the Nrf2 agonist sulforaphane compared to two new striatal cell lines expressing Htt (B3 and E4) (Fig. 5C). Furthermore, the ectopic expression of N-terminal mHtt led to a significant reduction in ARE activity in STHdh Q7/Q7 and HEK cells (Fig. 5D). These results demonstrate that the presence of mHtt leads to the disturbance of Nrf2 signaling.

Nrf2 Activation in Response to Oxidative Stress is Impaired in STHdh Q111/Q111 Cells
Nrf2 is activated by oxidative stress. Therefore we next determined whether mHtt affects Nrf2 activity in response to oxidative stress. Striatal cells were treated with tert-butylhydroquinone (tBHQ) and Cu 2+ and ARE activity was measured as described in materials and methods. STHdh Q7/Q7 cells showed a dramatic increase in ARE activity in a dose dependent manner  in response to oxidative stress, while STHdh Q111/Q111 cells showed significantly less of an increase in ARE activity at low concentrations of tBHQ and did not exhibit further increases in ARE activity when higher concentrations were used (Fig. 6). This result suggests that the disturbance of Nrf2 signaling renders STHdh Q111/Q111 cells more susceptible to oxidative stress.
Disturbed Signaling of Nrf2 in STHdh Q111/Q111 Cells is not due to its Expression Level Although Nrf2 activity was compromised in the STHdh Q111/ Q111 cells, there were no detectable differences in protein expression levels between the two cell types (Fig. 7A). Since Nrf2 activation is dependent on translocation to the nucleus, we examined Nrf2 levels after subcellular fractionation. There were no obvious differences in Nrf2 levels in the cytosol or nucleus between the two cell types (Fig. 7C). In contrast, the basal expression of Keap1 and p62 in STHdh Q111/Q111 cells was significantly lower than that of STHdh Q7/Q7 cells (Fig. 7, A and B). Next, we measured mRNA levels by real-time PCR to test whether altered protein levels arise from the change in mRNA levels (Fig. 7D). Nrf2 levels were similar between the two cell types, suggesting that altered signaling of Nrf2 in STHdh Q111/Q111 cells is not caused by the change of its expression. However, the Nrf2 target genes, HO-1 and Nqo1, were significantly increased in STHdh Q111/Q111 cells as compared to STHdh Q7/Q7 cells, implying that STHdh Q111/Q111 cells are under oxidative stress which may in turn activate defense mechanisms. Upregulation of oxidative stress related genes, including Nrf2 targets, was previously reported in a PC12 cell model in which mHtt exon1 protein was inducibly expressed [46]. However, this PC12 cell model also showed increases in catalase, glutathione peroxidase (GPX1), superoxide dismutase 1 (SOD1), and SOD2, while other studies using different HD models, including our previous study Figure 5. mHtt expression compromises the Nrf2 signaling pathway. A, Nrf2 activity was measured in striatal cells using an ARE reporter. The basal activity of Nrf2 is significantly reduced in STHdh Q111/Q111 cells compared to STHdh Q7/Q7 cells. n = 4 Data shown are mean 6 SE. ***P,0.001 vs. STHdh Q7/Q7 . B, An ARE luciferase reporter and a normalizing plasmid phRL-TK were cotransfected with Nrf2, Keap1, or both. Exogenous Nrf2 expression increases the activity of ARE luciferase in STHdh Q7/Q7 cells to a much greater extent than in STHdh Q111/Q111 cells. The ectopic expression of Keap1 results in a reduction in the ARE activity at the basal level as well as in the presence of exogenous Nrf2 expression in both cell types. n = 4 C, ARE activity was examined in the two additional lines for each striatal cell type. STHdh Q111/Q111 cell lines (1A and 6L) show significantly reduced ARE activity at the basal level compared to STHdh Q7/Q7 cells lines (B3 and E4). Treatment with sulforaphane (SFN), an agonist of Nrf2, increases ARE activity in STHdh Q7/Q7 cells to a greater extent than in STHdh Q111/Q111 cells. n = 3 D, Transient transfection of mHtt with 63 N-terminal amino acids results in a significant reduction in the ARE luciferase activity in STHdh Q7/Q7 as well as in HEK cells. n = 3. Data shown are mean 6 SE. **P,0.01, ***P,0.001 vs. mCherry. doi:10.1371/journal.pone.0057932.g005 using the striatal cell lines, showed a decrease in these genes [8,[47][48][49]. This discrepancy is likely ascribable to the different forms of mHtt that were expressed (full-length vs highly truncated) as well as the different cell types that were used. Nonetheless, the findings in the different models all indicate that a pathological polyglutamine expansion results in increased oxidative stress. In addition, and unexpectedly, mRNA levels of Keap1 were significantly increased in STHdh Q111/Q111 cells, while mRNA levels of p62 were not different between the two cell types (Fig. 7D), despite a significant reduction in their protein expression (Fig. 7, A  and B). Since Keap1 and p62 are degraded by autophagy, the reduced protein levels may indicate increased signaling of autophagy in STHdh Q111/Q111 cells [50,51]. p62 is an adaptor molecule that targets polyubiquitylated proteins for degradation by either autophagy or proteasome [52]. p62 can also activate Nrf2 by inhibiting the binding of Keap1 to Nrf2 [53]. In addition, Nrf2 signaling negatively interacts with autophagy as inhibition of autophagy can increase Nrf2 activity [54]. Therefore, we examined whether the modulation of autophagy signaling can differentially influence Nrf2 activity in striatal cells. As previously shown, p62 expression results in a slight but significant increase in Nrf2 activity in both cell types (Fig. 7E) [53,54]. Inhibition of autophagy by treatment with 3-methyladenine (3-MA) or chloroquine (CQ) resulted in a significant increase in Nrf2 activity in both cell types, while the activation of autophagy by treatment with rapamycin (RP) or trifluoperazine (TFP) did not change Nrf2 signaling (Fig. 7E). However, the ectopic expression of p62 or inhibition of autophagy failed to recover Nrf2 signaling in STHdh Q111/Q111 cells to the levels observed in STHdh Q7/Q7 cells. Together, this result suggests that p62 and autophagy signaling are not the major contributing factors to impaired Nrf2 signaling in STHdh Q111/Q111 cells.

Nrf2 Expression Attenuates the Fragmentation of Mitochondria
Given that STHdh Q111/Q111 cells display an increase in fragmented mitochondria and more oxidized mitochondria, we investigated whether Nrf2 expression modifies mitochondrial morphology (Fig. 8). Striatal cells were transfected with mito-mCherry and Nrf2, Keap1, p62, or GFP as a control. Nrf2 expression results in a significant reduction in fragmented mitochondria in STHdh Q111/Q111 cells and a similar trend of reduction in STHdh Q7/Q7 cells. Conversely, Keap1 expression appeared to increase the number of cells with fragmented mitochondria, although it was not statistically significant. The expression of p62 did not lead to any obvious alteration in mitochondrial morphology in either cell type. This result suggests that oxidative stress is a contributing factor to mitochondrial morphology which can be in part modulated by Nrf2 signaling. This observation is in line with the previous study showing that oxidative stress caused by 3-NP increases mitochondrial fragmentation, leading to neuronal cell death and that blockade of ROS abolishes mitochondrial fragmentation as well as cell death [55].
Autophagy Signaling is Highly Activated in STHdh Q111/ Q111 Cells We next examined whether autophagy regulation is altered in STHdh Q111/Q111 cells compared to STHdh Q7/Q7 cells. First, we looked at the conversion of LC3-I into LC3-II as an indicator of autophagy activation. Although the total level of LC3 is significantly reduced in STHdh Q111/Q111 cells, the ratio of LC3-II to LC3-I appears to be increased in STHdh Q111/Q111 cells compared to STHdh Q7/Q7 cells (Fig. 9, A and B). To further assess the autophagy activation, we transfected GFP-LC3 into striatal cells and monitored the cellular distribution of GFP-LC3. As the activation of autophagy increases, GFP-LC3 relocates from the cytosol into autophagosome, resulting in the punctate pattern. We found that STHdh Q111/Q111 cells show numerous GFP-LC3 positive punctae, while STHdh Q7/Q7 cells showed only a few GFP-LC3 positive punctae (Fig. 9C). Together, these results indicate that STHdh Q111/Q111 cells exhibit highly activated autophagy signaling.

Discussion
Mitochondrial impairment is emerging as a contributing factor to the pathogenesis of HD. A growing number of studies using mouse and cell models for HD, as well as tissues from HD patients, have shown that mHtt causes mitochondrial functional impairment such as hypersensitivity to Ca 2+ induced permeability transition pore opening, reduced Ca 2+ buffering capacity, deficits in the electron transport chain complexes and impaired mitochondrial bioenergetics [2][3][4]. An important factor mediating mitochondrial function is mitochondrial dynamics. An appropriate balance in mitochondrial fusion and fission is essential for cells to maintain metabolic states and homeostasis, and to respond to cellular stresses [12]. Recently, it demonstrated that the presence of mHtt leads to mitochondrial fragmentation in in vitro and in vivo HD models including HeLa cells and rat cortical neurons expressing mHtt, fibroblasts and lymphoblasts from HD patients, striatal progenitor cell lines, and YAC128 mouse [14,19,56]. Our finding of an increase in fragmented mitochondria with decreased level of tubular mitochondria in STHdh Q111/Q111 cells, as well as increased fragmented mitochondria in the striata of R6/2 mice, is in line with these previous studies. In addition, studies from two recent papers indicate that enhanced activation of Drp1 by mHtt causes mitochondrial fragmentation. Song et al. suggested that mHtt directly interacts with Drp1, which in turn increases its activity, whereas Costa et al. proposed the increased activity of calcineurin, a phosphatase which dephosphorylates Drp1, as a cause of increased Drp1 activity [14,19]. Increased steady state levels of mitochondrial fragmentation can be induced by increased fission, reduced fusion, or both. Their studies strongly support the idea of increased fission as one possible route to mitochondrial fragmentation. However, these studies do not rule out reduced fusion as an alternative pathway that contributes to mitochondrial fragmentation. To our knowledge we provide the first evidence that the expression of mHtt in striatal precursor cells negatively impacts mitochondrial fusion. Furthermore, we found that the overall levels of Opa1 are significantly reduced in STHdh Q111/ Q111 cells compared to STHdh Q7/Q7 cells, suggesting that reduced Opa1 could be causally involved in the reduced fusion process. Moreover, it has been shown that Opa1 expression attenuates the apoptosis in HD models, but Mfn1 expression did not despite correction of mitochondrial fragmentation, suggesting deficits in Opa1 may be a contributing factor to the mitochondrial fragmentation in HD [19].
It has been speculated that enhanced oxidative stress may play a significant role in HD pathogenesis [5,57]. Mitochondria play important roles in the defense against oxidative stress. A previous study showed that oxidative stress increases mitochondrial fragmentation in HeLa cells expressing mHtt [56]. In contrast, we observed a rapid increase in the population of STHdh Q111/ Q111 cells containing swollen mitochondria rather than fragmented mitochondria in response to oxidative stress. STHdh Q7/Q7 cells actually display an increase in fragmented mitochondria in response to oxidative stress, followed by a slow increase in swollen mitochondria. This result implies that upon cellular stresses, mitochondrial fragmentation would be a prerequisite for mitochondria swelling, an event that leads to the pathogenesis. Dysregulation of transcriptional processes that govern mitochondrial biogenesis and antioxidant system would result in mitochondrial impairment that would be exacerbated by cellular oxidative stress. PGC-1a is a transcriptional coactivator that plays important roles in mitochondrial biogenesis as well as defense mechanism against oxidative stress and was discovered to be repressed in models of HD [58]. In addition, we previously found that STHdh Q111/Q111 cells exhibit a deficit in the PPARc pathway [8,11] that regulates the expression of genes involved in mitochondrial function, ROS regulation, and fatty acid metabolism [59,60]. Given that transcriptional dysregulation is tightly linked with mitochondrial impairment as well as defects in antioxidant response, we sought to find additional transcriptional regulator that play important roles in these pathways and may be compromised in HD. Nrf2 has been considered as a key regulator in ROS defense mechanism and mitochondrial function. We discovered that Nrf2 signaling is severely compromised by the presence of mHtt. More importantly, STHdh Q111/Q111 cells fail to activate Nrf2 in response to oxidative stress. This result provides important insight into a potential contributing factor to the apparent impaired ability of cells to cope with oxidative stress when mHtt is present.
Two well known regulators of Nrf2 signaling are Keap1 and p62. Keap1 is an inhibitor that binds to Nrf2, leading to degradation through the ubiquitin-proteasome system. p62 is an activator that competes with Nrf2 for Keap1 binding and sequesters Keap1 into aggregates [53,54]. To gain insights into the mechanism of how Nrf2 signaling is compromised in STHdh Q111/Q111 cells, we examined the protein levels of Nrf2, Keap1, and p62 in the two striatal cell lines. We did not detect any difference in Nrf2 levels between the two cell types. Of interest, Keap1 and p62 were significantly downregulated in STHdh Q111/ Q111 cells. This result indicates that the reduced activity of Nrf2 is not due to decreased levels of Nrf2 or increased levels of Keap1. p62 has been reported to be regulated by binding of Nrf2 to an ARE in the p62 promoter [61], but this is unlikely to be responsible for the reduction in p62 protein in the STHdh Q111/ Q111 cells as mRNA levels for p62 were the same for both cell lines. It is noteworthy that Keap1 and p62 are selective substrates for autophagy [50,53]. It is plausible therefore, that reduced levels of Keap1 and p62 in STHdh Q111/Q111 cells may stem from enhanced activation of autophagy. We tested if STHdh Q111/Q111 cells show an increase in autophagy activity. As expected, autophagy signaling is highly activated in STHdh Q111/Q111 cells. Consistent with this finding, inhibition of autophagy by treatment with 3-MA or CQ resulted in a significant increase in Nrf2 activity in both cell types. Autophagy inhibition may lead to increased level of p62 which in turn inhibits Keap1 binding to Nrf2. Nonetheless, Nrf2 activity in STHdh Q111/Q111 cells was not increased by inhibition of autophagy to the same extent as in STHdh Q7/Q7 cells. This result suggests that hyperactivation of autophagy signaling in STHdh Q111/Q111 cells is likely a partial, but not a major, contributing factor to compromised Nrf2 activity.
The remaining question is what could be other contributing factors to disrupted Nrf2 signaling in STHdh Q111/Q111 cells? We sought to understand the underlying mechanism by which mHtt results in impairment in Nrf2 signaling. GSK3b has been reported to inactivate Nrf2 signaling by direct phosphorylation of Nrf2, a target of an E3 ubiquitin ligase adaptor b-TrCP that leads to proteasomal degradation, or by phosphorylation of Fyn, a kinase that leads to nuclear export and degradation of Nrf2 by phosphorylation of Nrf2 [62][63][64]. These studies suggest that GSK3b activation could be a causal factor reducing Nrf2 signaling. However, a previous study showed enhanced Akt activation which results in specific inactivation of GSK3b in STHdh Q111/Q111 cells and striatum of Hdh Q111/Q111 knockin mouse [65]. This led us to rule out GSK3b as a possible player that negatively influences Nrf2 signaling in STHdh Q111/Q111 cells. Previous studies demonstrated that CBP/p300 directly binds Nrf2, resulting in acetylation of Nrf2, a state that enhances Nrf2 transcriptional activity [66,67]. CBP/p300 has been shown to interact with mHtt, which results in inhibition of their acetyltransferase activity [68,69]. Therefore, it is conceivable that interference of CBP/p300 by mHtt could be a crucial factor that leads to disrupted Nrf2 signaling in STHdh Q111/Q111 cells. Investigation of this hypothesis in the future may be fundamental to understanding the exact molecular mechanism by which Nrf2 signaling is compromised in STHdh Q111/Q111 cells and to provide insights into development of therapeutic strategies of HD.
In summary, for the first time we provide evidence that mitochondrial fusion is impaired in an HD striatal precursor cell model. As a result, STHdh Q111/Q111 cells display an increased population of cells with fragmented mitochondria compared to STHdh Q7/Q7 cells. In addition, STHdh Q111/Q111 cells are more susceptible to oxidative stress, leading to increased levels of swollen mitochondria. We also identified disrupted Nrf2 signaling in STHdh Q111/Q111 cells which is likely to contribute increased mitochondrial fragmentation and functional impairment and thus increased susceptibility of STHdh Q111/Q111 cells to oxidative stress. This study provides important insights into involvement of defects in mitochondrial dynamics in the pathogenic mechanisms of mHtt.

Ethics Statement
All animal protocols have been approved by the UCAR at the University of Rochester (UCAR# 2006-102R).

Cell Fusion Assay
To monitor mitochondrial fusion, cell fusion was carried out as previously described with modification [7,75]. In brief, STHdh Q7/ Q7 or STHdh Q111/Q111 cells were plated in two of 35 mm dish. The next day, mitoGFP or mito-mCherry was transfected in each dish using Lipofectamine 2000 following the manufacturer's protocol. After 24 h, cells in both dishes were trypsinized, combined, and plated together onto 2-3 wells of 12 well plate containing a glass-coverslip in each well. The next day, cells were treated with 50 mg/ml of cycloheximide for 30 min prior to cell fusion to prevent protein synthesis. To fuse cells, prewarmed 50% w/v polyethylene glycol (PEG) 4000 in G-HBSS (20 Hepes, pH 7.4, 137 NaCl, 5 KCl, 0.5 KH 2 PO 4 , 0.5 Na 2 HPO 4 , 10 NaHCO 3 , 0.01 glycine; in mM), 2 mM CaCl 2 , 0.6 mM MgCl 2 , and 10 mM glucose was added to cells for 90 s. The cells were then washed gently four times with G-HBSS, placed back in culture medium containing cycloheximide, and returned to the incubator at 33uC for 3 h. Cells were fixed with fresh 4% paraformaldehyde containing 0.25% glutaraldehyde for 20 min at room temperature. The percentage of mitochondrial fusion was calculated by colocalization of mitoCherry and mitoGFP using image J software.

Mitochondrial Fusion Assay with MitoDendra
Striatal cells were plated on 12 well plates containing a glass coverslip in each well. MitoDendra was transfected 1-2 d after plating. The next day, a glass coverslip was placed on the microscope (Observer D1, Zeiss) and the green/red fluorescence of mitochondria was observed with a 636 oil objective. To photoconvert a small area in mitochondria was chosen and the aperture was minimized. The green fluorescence was photoconverted into the red fluorescence by illuminating mitochondria using a 400 nm LED with 100% power controlled by Colibri system (Zeiss) for 6 min. The mitochondrial images were taken by a digital CCD camera (ORCA-ER, Hamamatsu Photonics) every 30 s for 30 min with 2 digital gain factor. The green fluorescence was obtained by using a 470 nm LED with 3-5% power for 20-30 ms exposure and a 525/20 nm emission filter. The red fluorescence was imaged with a 530 nm LED with 100% power for 10-20 ms exposure and a 630/75 emission filter. The time-lapse merged images from green and red channels were generated using image J. A green mitochondrion which encounters a red mitochondrion resulting in colocalization (yellow mitochondrion) was considered as a mitochondrial fusion event.

Ratiometric Measurement of Mitochondrial Matrix Redox State using Mitochondrial Targeted GFP Sensors
Cells were plated on glass coverslips and transfected with rosGFP1, or rosGFP2 [43] using Lipofectamine 2000 according to manufacturer's instruction. Th next day, media were removed and replaced with G-HBSS for 30 min at 37uC. The fluorescence from mitochondria was imaged at 636 magnification on an Observer D1 microscope (Zeiss) coupled with a digital CCD camera (ORCA-ER, Hamamatsu). The emission was detected through a 525/20 nm filter with excitation consecutively from a 470 nm LED and a 400 nm LED controlled by a Colibri system (Zeiss) with the same exposure time. Fluorescence images were background corrected, mitochondrial images were selected, and emission intensities from 470 nm and 400 nm were acquired using Image J software. The ratio of emission intensities (470/400) was calculated in Microsoft Excel software. To generate a standard curve for the measurement of matrix redox state, cells were treated for 30 min in G-HBSS containing 5 mM H 2 O 2 or 10 mM DTT to obtain the maxima of oxidation and reduction, respectively. The standard curves were generated using Sigmaplot (Systat Software).

Nuclear Fractionation
Nuclear fractionation was carried out as previously described [76] with slight modifications. Briefly, STHdh Q7/Q7 and STHdh Q111/Q111 cells were washed twice, harvested in ice cold PBS and the cell pellets were resuspended in lysis buffer (10 mM Tris, pH 7.5, 10 mM NaCl, 3 mM MgCl 2 , 0.05% Nonidet P-40, 1 mM EGTA) with protease inhibitors (0.1 mM phenylmethylsulfonyl fluoride, and 10 mg/ml of each of aprotinin, leupeptin, pepstatin A). A sample was taken and used as whole cell lysate. The cells were triturated followed by centrifugation at 380 6 g for 5 min at 4uC. The supernatants were collected and used as the cytosolic fractions. The pellets were washed once in lysis buffer and twice in wash buffer (30 mM sucrose, 10 mM Pipes, pH 6.8, 3 mM MgCl 2 , 1 mM EGTA, 25 mM NaCl) with protease inhibitors. The crude nuclei were overlaid on the top of 0.8 M sucrose with protease inhibitors, and spun at 1200 6 g for 10 min at 4uC. The pellets were collected and resuspended in buffer B (300 mM sucrose, 10 mM Pipes, pH 6.8, 3 mM MgCl 2 , 1 mM EGTA, 25 mM NaCl, 0.5% Triton X-100) with protease inhibitors and used as the nuclear fractions. 20 mg of protein from each sample was visualized by immunoblot analysis.

Dual-luciferase Reporter Assays
Cells were plated in 24 well plates. The next day, a reporter plasmid pGL3-ARE and a normalizing plasmid phRL-TK (Promega) were transfected using Lipofectamine 2000. The next day, cells were treated with vehicle control or drug. After 24 h, cells were lysed with Passive lysis buffer (Promega), and the reporter activity was measured using the Dual-Luciferase Reporter Assay System (Promega). The reporter activity from Firefly luciferase was normalized with the Renilla luciferase activity.

RNA Isolation, Reverse Transcription, and Real-time PCR
Cells were plated in 60 mm dish and total RNA was extracted using TRIzol (Invitrogen) according to the manufacturer's instruction. Extracted total RNA was treated with RNase free-DNaseI Amplification Grade (Invitrogen) to remove contaminating DNA. Two micrograms of total RNA was reverse transcribed using SuperScriptIII reverse transcriptase and random hexamers (Invitrogen). The reaction mixture was diluted with 500 ml of DEPC-treated H 2 O. The PCR reaction was prepared in triplicate containing 10 ml of diluted cDNA, 2.5 ml of 2.5 mM primer mixture (forward and reverse), and 12.5 ml of SYBR GreenER qPCR SuperMix (Invitrogen) in 96 well optical PCR reaction plate (Bio-Rad). PCR reactions were performed in MyiQ real-time PCR system (BioRad). Amplification conditions consisted of an initial hot start at 95uC for 10 min followed by amplification of 45 cycles (95uC for 15 s, 60uC for 20 s, and 72uC for 40 s). Melting curve analysis was performed immediately after amplification. The relative amount of mRNAs was calculated by using the DDCt

Measurement of Mitochondria in R6/2 and Wild Type Mouse Brain
Approximately 13 weeks old transgenic R6/2 mice overexpressing exon 1 of mHtt and their wild type littermates were transcardially perfused with paraformaldehyde/glutaraldehyde. Their brains were subsequently processed and sectioned at 100 mm thickness. Striatal sections were incubated with an antibody (EM48, Chemicon) recognizing huntingtin followed by a secondary antibody conjugated to 1.4 nm gold particles. Sections were silver intensified, osmicated in 1% OsO4, dehydrated and embedded in Eponate 12. Ultrathin sections (50 nM) were cut and counterstained with 5% aqueous uranyl acetate followed by Reynolds lead citrate prior to being examined and captured by the Transmission Electron Microscope Hitachi 7650. Each mitochondrion was measured using Image-Pro. All mitochondria (20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30) in a given cell and approximately 20 cells/animal from several striatal sections were analyzed.

Statistical Analysis
Data were expressed as mean 6 SE (standard error) and analyzed using Student's t test. Statistical significance was considered when a P value was ,0.05. Figure S1 Quantitative measurement of mitochondrial size and shape. Electron micrographs of striatal neurons from non-transgenic (A) and transgenic (Tg) R62 mice (B) were captured using electron microscopy. Nuclear aggregates in Tg animals were identified by an antibody against huntingtin protein (arrow). Each individual mitochondrion (red arrowhead) was measured using Image-Pro. All mitochondria (20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30) in a given cell and ,20 cells/animal from more than one striatal sections were analyzed. Data represent % of total mitochondria (, 600 from 2 mice/genotype) 6SEM. Quantitative measurements of mitochondria size were grouped into different size bins (C) and or expressed as aspect ratio (D) (major/minor axes, an index of roundness with values approach 1 as the structure becomes more circular) and analyzed using t-test. (TIF)